COORDINATE

GEOMETRY
BY

LUTHER PFAHLER EISENHART

PROFESSOR OF MATHEMATICS
PRINCETON UNIVERSITY

DOVER PUBLICATIONS, INC.

NEW YORK NEW YORK

Copyright, 1939, by Ginn and

Company

All Rights Reserved

This new Dover edition

is

first

published in 1960

an unabridged and unaltered republication

of the First Edition.

book is republished by permission of

Ginn and Company, the original publisher of
This

this text.

Manufactured

in the

United States of America

Dover Publications, Inc.

180 Varick Street

New York

14,

New York

Preface

The purposes and

general plan of this

book are

set forth

a text has been tested

and proved through use during two years in freshman courses
in Princeton University. Each year it has been revised as a
result of suggestions not only by the members of the staff but
also by the students, who have shown keen interest and helpfulness in the development of the project.
An unusual feature of the book is the presentation of coordinate geometry in the plane in such manner as to lead readily
to the study of lines and planes in space as a generalization of the
geometry of the plane this is done in Chapter 2. It has been
our experience that the students have little, if any, difficulty in
handling the geometry of space thus early in the course, in the
in the Introduction.

Its practicability as

which it is developed in this chapter.

have
also found that students who have studied deterWe
minants in an advanced course in algebra find for the first time
in the definition and frequent use of determinants in this book
an appreciation of the value and significance of this subject.
In the preparation and revisions of the text the author has
received valuable assistance from the members of the staff at
in particular, Professors Knebelman and Tucker
Princeton
Messrs.
and
Tompkins, Daly, Fox, Titt, Traber, Battin, and
Johnson the last two have been notably helpful in the preparation of the final form of the text and in the reading of the
proof. The Appendix to Chapter 1, which presents the relation
between the algebraic foundations of coordinate geometry and
axioms of Hilbert for Euclidean plane geometry, is due in the
main to Professors Bochner and Church and Dr. Tompkins,
particularly to the latter, at whose suggestion it was prepared
and incorporated. The figures were drawn by Mr. J. H. Lewis.
It remains for me to express my appreciation of the courtesy
and cooperation of Ginn and Company in the publication of

way

in

this book.

LUTHER PFAHLER EISENHART

iii

Contents
PAGE

INTRODUCTION

ix

CHAPTER
Points

and Lines

in tKe

Plane

SECTION
1.

The Equation

2.

Cartesian Coordinates in the Plane

3.

of the First Degree in x

Distance between
rection Cosines

Internal

5.

An Equation

Direction

of a Line

Two

Direction Cosines of a Line

20

Angle

Lines

27

of a Line

33

The Slope

8.

Directed Distance from a Line to a Point

Two

11

17

Parametric Equations of a Line

7.

9.

Numbers and Di-

Angle between Directed Line Segments

Numbers and

Direction

between

Points

and External Division of a Line Segment

4.

6.

Two

and y

Equations of the First Degree in x and y

35
-

Determinants

of the Second Order

39

10.

The Set

47

11.

Oblique Axes

53

12.

The

54

13.

Resume

of Lines through a Point

Circle

Line Coordinates

CHAPTER
Lines and Planes in Space
14.

15.

62

Determinants

Rectangular Coordinates in Space

71

Two

Points
Distance between
Direction Numbers and Direction Cosines of a Line Segment Angle between Two Line

Segments

76

Contents
PAGE

SECTION

Equations of a Line Direction Numbers and Direction Cosines of a Line Angle of Two Lines

82

17.

An Equation

88

18.

The Directed Distance from a Plane

16.

of a Plane
to a Point

The

Dis-

tance from a Line to a Point

95

Two

19.
Equations of the First Degree in Three
Line as the Intersection of Two Planes

20.

Two Homogeneous

Unknowns A
98

Equations of the First Degree in Three

Unknowns
21.

104

Determinants of the Third Order

Degree in Three Unknowns

Three Equations of the

106

First
22.

23.

Three Homogeneous Equations of the First Degree

in

Three

Unknowns

114

Equations of Planes Determined by Certain Geometric ConShortest Distance between Two Lines

119

The Configurations

123

ditions
24.

of Three Planes

25. Miscellaneous Exercises

26.

Determinants of

27. Solution of

of

The Sphere

127

130

Any Order

Equations of the First Degree

of Four Dimensions

in

Any Number

Unknowns Space

CHAPTER

137

Transformations of Coordinates
28.

Transformations of Rectangular Coordinates

29. Polar Coordinates in the Plane

30.

Transformations of Rectangular Coordinates in Space

31. Spherical

and Cylindrical Coordinates

vi

149
154

160

166

Contents

CHAPTER

The Conies Locus Problems

PAGE
171

SECTION
32.

A Geometric

33.

The Parabola

174

34.

Tangents and Polars

177

35. Ellipses

36.

Definition of the Conies

and Hyperbolas

182

Conjugate Diameters and Tangents of Central Conies

The Asymptotes

37. Similar Central Conies

191

of a Hyperbola

196

Conjugate Hyperbolas
38.

The Conies

39.

Equations of Conies
ordinate Axes

as Plane Sections of a Right Circular

Whose Axes Are

40.

The General Equation

41.

The Determination
Form

Cone

^Parallel to the

201

Co203

of the Second Degree

Invariants

of a Conic from Its Equation in General

215

42. Center, Principal Axes,

and Tangents of a Conic Defined

by a General Equation
43.

208

221

Locus Problems

229
ti

CHAPTER
The Quadric
44. Surfaces of Revolution
45. Canonical

46.

Surfaces

The Quadric

Surfaces of Revolution

Equations of the Quadric Surfaces

243

The Ruled Quadrics

47. Quadrics

Whose

248

Principal Planes

Are

Parallel to the

Co252

ordinate Planes
48.

239

The General Equation of the Second Degree in x, y, and z

The Characteristic Equation Tangent Planes to a Quadric
vii

256

Contents
PAGE

SECTION
49. Centers
50.

The

Vertices

Invariants

Points of

7, /,

51. Classification of the

D, and

Symmetry

267
271

Quadrics

APPENDIX TO CHAPTER
INDEX

264

279

293

Vlll

Introduction

Coordinate geometry is so called because it uses in the treatof geometric problems a system of coordinates, which
associates with each point of a geometric figure a set of numso that the conditions which each point
coordinates
bers

ment

satisfy are expressible by means of equations or inequaliordinarily involving algebraic quantities and at times
trigonometric functions. By this means a geometric problem is

must

ties

reduced to an algebraic problem, which most people can handle

with greater ease and confidence. After the algebraic solution
has been obtained, however, there, remains its geometric interpretation to be determined for^ttigdPfpblem is geometric, and
4
algebra is a means to its solution, not the end. This method
was introduced by Rene Descartes in La Geometric, published
in 1636
accordingly coordinate geometry is sometimes called
Cartesian geometry. Before the time of Descartes geometric
reasoning only was used in the study of geometry. The advance
in the development of geometric ideas since the time of Descartes
is largely due to the introduction of his method.
Geometry deals with spatial concepts. The problems of
physics, astronomy, engineering, etc. involve not only space but
usually time also. The method of attack upon these problems
is similar to that used in coordinate geometry.
However,
geometric problems, because of the absence of the time element,
;

'

are ordinarily simpler,

student

first

become

and consequently
familiar with the

it is

advisable that the

methods of coordinate

geometry.

The aim

of this

book

is

to encourage the reader to think

mathematically. The subject matter is presented as a unified

whole, not as a composite of units which seem to have no
relation to one another.

and the various

Each

situation

is

completely analyzed,

possibilities are all carried to their conclusion,

for frequently the exceptional case (which often is not presented to a student) is the one that clarifies the general case
the idea epitomized in the old adage about an exception and a
ix

Introduction

Experience in analyzing a question fully and being careful

is one of the great advantages
of a proper study of mathematics.
Examples included in the text are there for the purpose of
illustration and clarification of the text there is no attempt to
formulate a set of patterns for the reader so that the solution of
exercises shall be a matter of memory alone without requiring
him to think mathematically. However, it is not intended that
he should develop no facility in mathematical techniques rather
these very techniques will have added interest when he understands the ideas underlying them. The reader may at first have
some difficulty in studying the text, but if he endeavors to
master the material, he will be repaid by finding coordinate geometry a very interesting subject and will discover that mathematics is much more than the routine manipulation of processes.
Since, as has been stated, coordinate geometry involves the
use of algebraic processes in the study of geometric problems
rule.

in the handling of all possibilities

and

also the geometric interpretation of algebraic equations,

it

important that the reader be able not only to use algebraic

processes but also to understand them fully, if he is to apply
is

them with confidence. Accordingly in Chapter 1 an analysis is

made of the solution of one and of two equations of the first
degree in two unknowns and in Chapter 2 are considered one,
two, and three equations of the first degree in three unknowns.
In order that the discussion be general and all-inclusive, literal
;

used in these equations. It may be that most,

not all, of the reader's experience in these matters has been
with equations having numerical coefficients, and he may at

coefficients are
if

have some difficulty in dealing with literal coefficients. At

times he may find it helpful in understanding the discussion to
write particular equations with numerical coefficients, that is,
to give the literal coefficients particular numerical values, thus
first

supplementing illustrations of such equations which appear in

the text. However, in the course of time he will find it unnecessary to do this, and in fact will prefer literal coefficients because
of their generality and significance.
In the study of equations of the first degree in two or more
unknowns determinants are defined and used, first those of the

Introduction

second order and then those of the third and higher orders, as
they are needed. Ordinarily determinants are defined and
studied first in a course in algebra, but it is a question whether
one ever appreciates their value and power until one sees them
used effectively in relation to geometric problems.
The geometry of the plane is presented in Chapter 1 in such
form that the results may be generalized readily to ordinary
space and to spaces of four and more dimensions, as

Chapter

is

done in

2.

Some of the exercises are a direct application of the text so

that the reader may test his understanding of a certain subject
by applying it to a particular problem and thus also acquire
others are of a theoretical
facility in the appropriate techniques
of
in
which
the
reader is asked to apply
the
solution
character,
the principles of the text to the establishment of further
theorems. Some of these theorems extend the scope of the
text, whereas others complete the treatment of the material
in the text.
Of general interest in this field are A. N. Whitehead's
Introduction to Mathematics, particularly Chapter 8, and E. T.
;

Bell's

Men

of Mathematics,

Chapter

3.

COORDINATE GEOMETRY
CHAPTER

Points and Lines in the Plane

1.

The Equation

of tKe First

Degree

in

x and y

In his study of algebra the reader has no doubt had experience in finding solutions of equations of the first degree in two
unknowns, x and y, as for example x
2y 3 = Q. Since we
shall be concerned with the geometric interpretation of such
equations, a thorough understanding of them is essential. We
therefore turn our attention first of all to a purely algebraic
study of a single equation in x and j>, our interest being to find
out what statements can be made about a general equation
of the first degree, which thus will apply to any such equation
without regard to the particular coefficients it may have. Accordingly we consider the eqtiation

ax

(1.1)

+ by + c = 0,

a, &, and c stand for arbitrary numbers, but definite in

the case of a particular equation. We say that a value of x and
a value of y constitute a solution of this equation if the left-

where

hand

side of this equation reduces to zero

are substituted.

when

Does such an equation have a

these values

solution what-

The question is not as trivial as may

appear at first glance, and the method of answering it will serve
as an example of the type of argument used repeatedly in later
sections of this book.
We consider first the case when a in (1.1) is not equal to
zero, which we express by a ^ 0. If we substitute any value
whatever for y in (1.1) and transpose the last two terms to
the right-hand side of the equation, which involves changing
tbeir signs, we may divide through by a and obtain the value

ever be the coefficients?

K -_

fkjg va i ue o f K ancj the chosen value of .y satisfy

hence any value of

the equation, as one sees by 'substitution
resulting value of x constitute a solution. Since y
may be given any value and then x is determined, we say that
when a ^ there is an endless number of solutions of the equation.
;

y and the

The above method does not apply when a

since division

= 0,

that

is,

to

x + by + c = 0,

the Aquation

by zero

may have been

is

told that

The reader
and that any number,

not an allowable process.

it is

allowable,

Points
for

and Lines

in the

Plane

[Chap. 1

but infinity defined

2, divided by zero is infinity
manner is a concept quite different from ordinary numand an understanding of the concept necessitates an ap-

example

in this
bers,

propriate knowledge of the theory of limits. If now

above equation may be solved for y that is, y

6^0,

the

c/b.

For

this equation also there is an endless number of solutions, for

all of which y has the value
c/b and x takes on arbitrary
c = 0,
Usually the above equation is written by
which does not mean that x is equal to zero (a mistake frequently made), but that the coefficient of x is zero.
If b 7* 0, no matter what a is, equation (1.1) may be solved
ax
c<
> the value of
for y with the result y =
y corresponding
~^~

values.

any choice of x being given by

have the theorem

to

[1.1]

An

this expression.

zero has

an endless number of

There remains

= 0;

unknowns in which
unknowns is not equal to

equation of the first degree in two

the coefficient of at least one of the

Hence we

that

is,

solutions.

for consideration the case

the equation

when a

and

Evidently there are no solutions when c ^ 0, and when c =

any value of x and any value of y constitute a solution. The
reader may say that in either case this is really not an equation
of the first degree in x and y, and so why consider it. It is true
that one would not start out with such an equation in formulating a set of one or more equations of the first degree to express in algebraic form the conditions of a geometric problem,
but it may happen that, having started with several equations,
and carrying out perfectly legitimate processes, one is brought to

an equation of the above type that is, equations of this type do

arise and consequently must be considered. In fact, this situation
arises in
9, and the reader will see there how it is interpreted.
However, when in this chapter we are deriving theorems concerning equations of the first degree in x and y, we exclude the
degenerate case when the coefficients of both x and y are zero.
;

The Equation

Sec. 1]

of the First

Degree

in

x and y

Any solution of equation (1.1) is also a solution of the

equation k(ax + by + c) = 0, where k is any constant different
from zero. Moreover, any solution of this equation is a solution of (1.1)
for, if we are seeking the conditions under which
the product of two quantities shall be equal to zero and one of
the quantities is different from zero by hypothesis, then we must
seek under what conditions the other quantity is equal to zero.
Accordingly we say that two equations differing only by a con;

stant factor are not essentially different, or are not independent,

two equations are equivalent. In view of this dis-

or that the

cussion

it

follows that a

common

factor,

if

any, of

all

the co-

an equation can be divided out, or the signs of

all the terms of an equation can be changed, without affecting the
solutions of the equation
processes which the reader has used
even though he may not have thought how to justify their use.
It should be remarked that the values of the coefficients in
efficients of

equation (1.1) are the important thing, because they fix y when
x is chosen and vice versa. This is seen more clearly when
we take a set of values xi, y\ and seek the equation of which
it is a solution
this is the inverse of the problem of finding
solutions of a given equation. Here the subscript 1 of Xi and
y\ has nothing to do with the values of these quantities. It is a
means of denoting a particular solution of an equation, whereas
x and y without any subscript denote any solution whatever.
;

and yi is to be a solution
must be such that

If xi

and

axi

(1.2)

On

is

stant for

When

of the

form

and substituting

in (1.1),

we

(1.1),

where now

c is

(axi

+ by\},

a con-

any values of a and b, x\ and y\ being given constants.

equation (1.3) is rewritten in the form

- *i) + b(y - yi) = 0,

and y = yi is a solution

a(x

(1.4)
it is

+ by - axi - byi = 0,

ax

which

b,

+ byi + c = 0.

solving this equation for

obtain the equation
(1.3)

of (1.1), the coefficients a,

seen that x

be a and

6.

= xi

of (1.3) whatever
(1.4), we see

Since a and b can take any values in

and Lines

Points

that an equation of the

determined (that
is

first

is, a, b,

in tKe

Plane

[Chap.

degree in x and y is not completely

c are not fixed) when one solution

and

given.

Suppose then that we require that a different set of quanx2 and >>2, be also a solution, the subscript 2 indicating
that it is a second solution. On replacing x and y in (1.4) by
#2 and y2 we obtain
tities,

a(x 2

(1.5)

Since
or

are dealing with two different solutions, either X2 ^ x\

we assume that x2 ^ x\ and solve (1.5) for 0, with

we
7*

y\
the result
3/2

- xi) + b(y2 - yi) = 0.

(1.6)

On

=
we

substituting this value of a in (1.4),

This equation

is satisfied if b

0,

obtain

but then from

= 0.

(1.5)

we have

Since #2 ^ #1, we must have a = 0, contrary to

Xi)
a(x2
the hypothesis that a and b are not both equal to zero. Therefore since b cannot be zero, it may be divided out of the above
equation, and what remains

may

Multiplying both sides by #2

be written
(y2

(1.8)

which is
numbers.

Thus

- yi)x -

of the

far

form

(x2

tion differs only

the

same

is

it

the resulting equation

(x2 yi

that x2

may

- Xiy2 = 0,
)

since *i, y\, x2 ,

we have assumed

cannot have
and consequently from
tion (1.4) becomes a(x

And

x\,

- xi)y +

(1.1),

also y 2 = y

be written

x\

and y2 are
if

now x2 =

fixed

xi

we

two solutions are different,

we have b = 0. In this case equa-

since the

(1.5)

#1)

= 0. We

by the constant

observe that this equa-

factor a from x

Xi

true of the form which (1.8) takes

= 0.

when

Sec. l]

The Equation

of the First

Degree

x and y

in

we put # 2 = Xi, namely, (jy 2 y\)(x *i) = 0. But we have

remarked before that a common factor of all the coefficients
does not affect the solutions of the equation.
Accordingly we have the theorem
[1.2]

Equation (1.8) is an equation of the first degree in x and y

which has the two solutions x\, y\ and * 2 y*.
,

We say "an equation" and not "the equation" because any

constant multiple of equation (1.8) also has these solutions in
this sense an equation is determined by two solutions to within
an arbitrary constant factor.
As a result of the discussion leading up to theorem [1.2] we
have the theorem
;

[1.3]

Although an equation of the

first degree in x and y admits

an endless number of solutions, the equation is determined
to within an arbitrary constant factor by two solutions, that
is, by two sets of values of x and y; for the solutions x\, y\
and X2, yz the equation is equivalent to (1.8).

NOTE. In the numbering of an equation, as (1.5), the number preceding the period is that of the section in which the equation appears,
and the second number specifies the particular equation. The same
applies to the number of a theorem, but in this case a bracket is used
instead of a parenthesis.

EXERCISES
1.

What

values

must be assigned

that the resulting equation

is

to
;

a, b,

and

so that

equation (1.1) so
equivalent to this

c in

it is

equation?
2.

Find two solutions of the equation

2x-3;y +

(i)

and show that equation

equation
3.

(1.8) for these

= 0,
two solutions

is

equivalent to

(i).

Show

that the equation

ferent solutions for both of

*-2;y +

which x

degree in x and y for which this

is

=2

true.

3
;

=
find

does not have two difan equation of the first

Points
4.

Show

that

it

where
5.

and Lines

follows from (1.1)

b

= y - y\
2

in the

and

*i

and x 2t y2 are solutions of

xi, y\

Criticize the following statements

x2

Plane

x 2yi

and

= 0,

this

2.

as
is

an equation

in

Xiy 2 ,

b.

(1.1).

In obtaining solutions of an equation ax

take any value if a
0, and only in this case.

= 0,

(1.8) that

a.

[Chap,

x and

y,

by

+c=Q

has the solution x

may

3,

the only solution.

Cartesian Coordinates in the Plane

Having studied equations of the first degree

in

two unknowns,

we turn now

to the geometric interpretation of the results of this

study. This is done by the introduction of coordinates, which
serve as the bridge from algebra to geometry. It is a bridge
with two-way traffic; for also by means of coordinates geo-

metric problems may be given algebraic form. This use of coordinates

was Descartes's great contribution
to mathematics, which revolutionized the study of geometry.
As basis for the definition of coordinates,

we take two

lines

perpen-

dicular to one another, as A' A and

B'B in Fig. 1, which are called the
x-axis and y-axis respectively their

A/

FIG.

intersection

is

called the origin.

Suppose that xi and yi are a pair of numbers. If x\ is positive,

we lay off on OA from a length OC equal to x\ units and draw
through C a line parallel to B'B if xi is negative, we lay off
a length equal to
from
x\ units on OA', and draw through
the point so determined a line parallel to B'B. Then, starting
we lay off on OB a length OD equal to y units if y l
from
is positive, or on OB a length
yi units if y\ is negative, and
;

'

through the point so determined draw a line parallel to A' A.

These lines so drawn meet in a point PI, which is called the
graph of the pair xi, yi\ we say that P\ is the point (x\, >>i),

Cartesian Coordinates in the Plane

Sec. 2]

and

also xi is called the

x\ y\ are called the coordinates of PI
abscissa of PI and y\ the ordinate. Evidently xi is the distance
of Pi from the j-axis (to the right if x\ is positive, to the left
9

and y\ is the distance of PI from the #-axis.

In specifying a line segment OC, the first letter
indicates
the point from which measurement begins, and the last letter
C the point to which measurement is made. Accordingly we
have CO =
OC, because for CO measurement is in the direc-

if

Xi is negative)

tion opposite to that for OC.

annoying, but in many cases

when

(The question of sign may be

is important
there are also

it

not important, and the reader is expected to

discriminate between these cases.) If we take two points
Ci(#i, 0) and C 2 (# 2 0) on the #-axis, we see that the magnitude
and sign of the segment dC 2 is x 2
xi, since CiC 2 = OC 2
OC\.
cases

it is

When

and C 2 lie on the same side of 0, OC 2

OCi is the diftwo lengths when they are on opposite sides of 0,
it is the sum of two lengths.
(The reader does not have to
x\ takes care of all of it.) CiC 2 is
worry about this, since x<2
called the directed distance from Ci to C 2
Ci

ference of

We

observe that the coordinate axes divide the plane into

four compartments, which are called quadrants. The quadrant

formed by the positive #-axis and positive ;y-axis is called the
the one to the left of the positive ;y-axis (and
first quadrant
;

the second quadrant; the ones below the

negative #-axis and the positive *-axis, the third and fourth

above the

x-axis),

quadrants respectively.

By definition the projection of a point upon a line is the foot

of the perpendicular to the line from the point. Thus C and D
in Fig. 1 are the projections of the point Pi upon the *-axis and
The projection of the line segment whose
end points are Pi and P 2 upon a line is the line segment whose
end points are the projections of Pi and P 2 on the line. Thus
in Fig. 1 the line segment OC is the projection of the line segment DPi upon the #-axis.
Two points Pi and P2 are said to be symmetric with respect
to a point when the latter bisects the line segment PiP2
sym-

>>-axis respectively.

metric with respect to a line

when

the latter

the line segment PiP2 and bisects

it.

is

Thus

perpendicular to
the points (3, 2)

Points

and

and

2)

(3,

The

in the

Plane

[Chap. 1

symmetric with respect to the origin, and

are
symmetric with respect to the #-axis.
2)

2) are

3,

and Lines

(3,

reader should acquire the habit of drawing a reasonably accu-

rate graph to illustrate a problem under consideration.

carefully
made graph not only serves to clarify the geometric interpretation

of a problem but also may serve as a valuable check on the accuracy

of the algebraic work. Engineers, in particular, often use graphical

methods, and for

many purposes numerical results obtained graphiare

cally
sufficiently accurate. However, the reader should never forget that graphical results are at best only approximations, and of
value only in proportion to the accuracy with which the graphs are
drawn. Also in basing an argument upon a graph one must be sure
that the graph is really a picture of the conditions of the problem
(see Chapter 3 of W. W. R. Ball's Mathematical Essays and Recrea-

The Macmillan Company).

tions,

EXERCISES
1. How far is the point (4,
3) from the origin? What are the
coordinates of another point at the same distance from the origin?
How many points are there at this distance from the origin, and how

would one find the coordinates of a given number of such points ?

2.

and
(

What are
6)

4,

4, 6) is
3.

the coordinates of the point halfway between the origin

What are the coordinates of the point such that

halfway between

it

and the

origin ?

are the points in the plane for which x > y (that is, for
2
y2 < 4 ?
greater than y) ? Where are the points for which x

Where

which x

is

Where are the points for which < x ^ 1 and < y ^ 1 (the
symbol ^ meaning "less than or equal to")? Where are the points
for which
< y < x < 1?
4.

Where

5.

are the points for which xy

= 0?

6. What are the lengths of the projections upon the #-axis and jy-axis
of the line segment whose end points are (1, 2) and (
3, 4)?

segment with the origin as an end point

is of length /.
projections upon the *-axis and ^-axis in terms of / and
the angles the segment makes with the axes ?

7.

What
8.

line

are

its

Given the four points P (l,-2), P2 (3, 4), P8 (- 5, 1), and

show that the sum of the projections of the line segments
1

P4(0, 3),

10

Distance between

Sec. 3]

Two

Points

z, P*Pz, and PaP4 on either the *-axis or the ;y-axis is equal to

the projection of P\P* on this axis. Is this result true for any four
points ; for any number of points ?

Given the point (2,

3), find
it
with
to
respect to the
symmetric

the

9.

three

origin,

points which are

and to the

*-axis

and

y-axis respectively.

3.

Distance between

Direction

Numbers and

Two

Points.

Direction Cosines.

Angle between Directed Line Segments

Consider two points P\(x\, y{) and

P 2 (# 2

jy 2 ), and the line

which the angle at
Q is a right angle. The square of the distance between the
points, that is, the square of the line segment PiP 2 is given by

segment PiP 2 joining them, as

in Fig. 2, in

(PiP 2 )

(PiQ)

(QP 2 )

- *0 2 +
(* a

(y a

- yi) 2

Hence we have the theorem

[3.1]

The distance between the points

(xi,

y\)

and

(# 2 , yz) is

V(* 2 -*i) +0> 2 -;Ki) 2

(3.1)

The length and

in Fig. 2 are given

sign of the directed segments

PiQ and PtR

by

= x2
PiR = JV2 -

R (XI

PiQ
(3.2)

no matter in which quadrant

Pi lies and in which P 2 as
,

the reader will see when he

draws other figures. These
numbers determine the rec-

PiP 2 is a
and consequently

tangle of which
diagonal,

Pi

FIG. 2

determine the direction of PiP2 relative to the coordinate

axes. They are called direction numbers of the line seg* 2 and y\
ment PiP 2 In like manner x\
y 2 are direction
numbers of the line segment P 2 Pi. Thus a line segment has
two sets of direction numbers, each associated with a sense
.

along the segment and either determining the direction of the

*
11

Points

and Lines

in the

Plane

[Chap.

segment relative to the coordinate axes. But a sensed line segment, that is, a segment with an assigned sense, has a single
set of direction numbers.
line segment parallel to PiP 2 and having the same
as PiP 2 has the same direction numbers as PiP 2
sense
and
length
new
this
for,
segment determines a rectangle equal in every
for PiP 2 This means that the differences of
one
the
to
respect
the *'s and the /s of the end points are equal to the corresponding differences for PI and P 2 Since one and only one line segment having given direction numbers can be drawn from a given
point, we have that a sensed line segment is completely determined by specifying its initial point and its direction numbers.
There is another set of numbers determining the direction
of a line segment, called the direction cosines, whose definition

Any other

involves a convention as to the positive sense along the segIf the segment is parallel to the #-axis, we say that its

ment.

positive sense is that of the positive direction of the x-axis.

If the segment is not parallel to the #-axis, we make the convention that upward along the segment is the positive sense on
the segment. This is in agreement with the sense already established on the ;y-axis. In Fig. 2 the distance PiP 2 is a positive
number, being measured in the positive sense, and the distance
P2 Pi is a negative number (the numerical, or absolute, value of
these numbers being the same), just as distances measured on
the #-axis to the right, or left, of a point on the axis are positive, or
negative. When the positive sense of a line segment is determined
by this convention, we refer to it as a directed line segment.

By definition the direction cosines of a line segment are the

cosines df the angles which the positive direction of the line segment makes with the positive directions of the x-axis and y-axis
respectively, or, what is the same thing, with the positive directions
of lines parallel

to

them.

They

are denoted respectively

Greek letters X (lambda) and n (mu). Thus,

line segment PiP 2
= cos B,
X = cos A,
(3.3)
fji

and also

= PiP2 cos A = PiP 2 X,

PiR = PiP 2 cos B = PiP 2

PiQ

ju-

12

by the

in Fig. 2, for the

Direction

Sec. 3]

If

we denote

(3.4)

and

Numbers and

the positive distance PiP2

*i

X and

since their values

JJL,

tion of the

segment

relative

we have from

x2

xi
is

the distance

in Fig. 2, this does not

depend only upon the

direc-

to the coordinate axes;

and

become

consequently equations (3.5)

where 3

d,

= d\,

we imagine PI and P 2 interchanged

alter

by

(3.2)

(3.5)
If

Direction Cosines

yiy* = dp,

d\,

P 2 Pi

in the

Hence equations (3.5) hold also

if d is negative, that is, when
PI is above P 2 on the line and
the distance PiP 2 is negative.

new

figure

and

is

positive.

P*(x*y t )

R(xi,y*)

We consider now the difference,

sults

if

any, in the above repositive sense

when the

of the segment makes an obtuse angle with the positive

sense of the #-axis.

situation

case

is

shown in

Such a
Fig. 3, where

FIG. 3

this angle is

180

C.

In this

we have

= QP = p p 2 cos C =
=
PiR
PiP 2 cos B = PiP 2
1

jtt,

same

and consequently equations (3.5)

in
with
the
hold
understanding that d is the directed
every case,
distance PiP2. Thus a line segment is completely determined
by an end point, its direction cosines, and its length and sense
relative to the given end point.
which are the

Two

as (3.4)

segments are parallel if their direction cosines are

equal, in which case their direction numbers are proportional,
as follows from (3.4). As a means of distinguishing line segments with the same direction but with opposite senses, some
writers reserve the term parallel for segments with the same
direction and the same sense, and use the term antiparallel for
the case when the senses are opposite; such distinction is
line

necessary in mechanics, for example.

13

Points

From
ment
is

it

and Lines

in the

Plane

[Chap. 1

the definition of the positive direction along a segwhen PiP 2 is not parallel to the *-axis /z

follows that

positive,

whereas X

take any value between

may

the segment is parallel to the *-axis, /* =

in fact, the direction cosines of the #-axis and

+ 1. When
X

are

and

1,

From

When
tion,

;y-axis

0, 1 respectively.

the definition of d and

Theorem

[3.1],

we have

the expressions from (3.5) are substituted in this equa-

we obtain

+ M2 =

X2

(3.6)

Accordingly
[3.2]

and
and

The

we have the

first

direction cosines X,

1.

part of the theorem

of any line segment satisfy the

2 =
2
1
X
never negative; when
is
> 0,
+ ju
equation
= 0, X = 1 ; and, conversely, any
1 < X < 1
when
two numbers X and
satisfying these conditions are direc;

JJL

JJL

/JL

fJL

jj,

tion cosines of a line segment.

To prove

the second part of the theorem we take any point

and an arbitrary number d, and determine numbers
# 2 and y2 from equations (3.5). Then X and /z are direction
cosines of the line segment PiP 2 where P 2 is the point fa, y^)
and d is the directed distance PiP 2 and the theorem is proved.
Consider now in connection with the line segment PiP 2 another line segment PI Pa, where P 3 is the point fa, y$). We
denote by 6 (theta) the angle formed at PI by these sensed
segments and we note that if they have the same direction,
or 180 according as the segments have the same or
9=
PI(XI, y\)

opposite sense. We consider the case when 6 is not equal to

or 180, and denote by Xi, jui and X 2 ^2 the direction cosines
,

of the segments PiP 2 and PiP 3 respectively, and by di and d2

the directed distances PiP 2 and PiP 3
draw, as in Fig. 4,
from the origin two line segments, OP\ and OFV, parallel to
.

We

and of the same directed lengths as PiP 2 and PiP3 respectively.

The direction cosines of these line segments are Xi, ^i and
14

Angle between Directed Line Segments

Sec. 3]

X2
2t and the coordinates of /Y and /Y are di\i, dipt and
d2 \ 2 d2 2 respectively, as follows from equations analogous to
= yi =
and the an(3.5) with Xi
,

fjL

/jL

gle of these

is equal to
the angle of the original line segments. By the Law of Cosines of

segments

plane trigonometry applied to the

triangle in Fig. 4, in

and

of the

sides

which

/,

/i,

denote the lengths of the

/2

triangle,

as distin-

from directed distances,

guished

we have
I

(3.7)

By means
I

(3.8)

2
/i

of

/2

-2

Theorems

(d 2 \ 2

12

FlG> 4
6.

and

[3.2]

(d 2fji 2

2 did 2 (\i\ 2

we have

the definition of k and 1 2 it follows that k is equal to

di according as d\ is positive or negative, and similarly

di or

for

[3.1]

- diXx) 2 +

= d2 2 + di 2
From

cos

/!/ 2

from

we equate

this

it

follows that

2
/i

= di 2

and

the expressions (3.7) and (3.8) for

the resulting equation,

we

/2
/

= d2 2 When
.

2
,

and simplify

obtain.

= did2 (\i\2 + MiM2)

/i/ 2 cos 6
and d2 are both positive or both negative, IJ2 = did2 and
Hence we have the
if di and d 2 differ in sign, hl 2 =
d\d2
theorem
If di

13.3]

The angle 6 between two line segments whose direction

jui and X 2 ju 2 is given by

sines are Xi,

cos e

(3.9)

where

e is

same sense

1 or

= ^(XiX 2 +

/xiM2),

1 according as the

two segments have the

(both positive or both negative) or opposite senses.

Here, and throughout the book, we use e in place of the sign

because it enables one to state more clearly when the sign

is

co-

+ and when

15

and Lines

Points

From Theorem

[3.3]

is

[Chap. 1

we have

(3.5)

given by

cos e

(3 10)

and equations

Plane

between the directed line segments P\P<2 and

The angle

[3.4]

in tKe

~ *0(*3 - *0 +

(* 2

CV2

- yi)(y*

y\) t

111?

where k and
the

As a

corollary

The
and

[3.5]

12

are the lengths (not directed distances) of

segments P\Pz and PiPa respectively.

we have

line segments from the point (x\, y\) to the points (x%, y%)
(#3,

jy 3 )

(x*

(3.11)

are perpendicular,

- xi)(* 3 - *i) +

0>2

if

and only

if

- yi)(ya - yi) = 0.

The phrase "if and only if" used in this theorem (and
throughout the book) is a way of stating that both a theorem
with "if" alone and its converse are true; that is, a statement

that

holds,

is

true

and B

if

is

and only

true

if

if

holds

means that A

is

true

if

holds.

EXERCISES
Show

1.

(5,

2.

and
3.

(6,

5), (2,

4),

Find the point on the x-axis which

is

equidistant from

(0,

Find

xi

and

;yi

such that PI(XI,

an equilateral

yi), (0, 0),

and

(3,

- 4)

are the

triangle.

Find the direction numbers and direction cosines of the

- 2)

to

P2 (4,

segment

segline

in Ex. 4.

For what value of a

is

the line segment from the point P\(2

to the point P(3, a) perpendicular to the line

is

line

2).

Find the coordinates of the point p(x, y) so that the line

ment P\P has the same direction and is twice as long as the
5.

Pa

1)

(3, 3).

segment joining Pi(l,

6.

and

1) is isosceles.

vertices of
4.

that the triangle with the vertices

the point

(5, 1) ?

16

1)

segment PiP2 where

,

Internal

Sec. 4]

7.

and External Division

of a Line

Segment

Find the lengths and direction cosines of the sides of the

whose vertices are (3,
also the
6), (8,
2), and (
1,
1)

triangle

cosines of the angles of the triangle.

8.

Prove that

(2, 1), (0, 0),

(-

1, 2),

and

(1,

3) are the vertices of

a rectangle.
9.

Find the condition to be

and

points PI(XI, y\)

^2(^2,

y2 )

subtend a right angle at the

shall

by the coordinates

satisfied

in order that the line

of the

segment PiP2

origin.

Derive equation (3.10) directly by applying the Law of Cosines

to the triangle with vertices at the points (x\ y\), (x 2 y 2 ), (xa, ya).
10.

Find the coordinates of a point P(x, y) so that the segment

has the same direction as the segment P\P* with the end points

11.

P\P
(xi,

y\)

and

(x 2 ,

How many

y2 ).

points

possess this property,

and

where are they ?

Find the coordinates of a point P(x, y) so that the segment

perpendicular to the segment P\P2 with the end points (*i, y\)
How many points P possess this property, and where
(#2, 3^2).

12.

P\P
and

is

are they ?

4. Internal

and External Division

In this section

is,

Segment

we determine the

P(x, y) which divides the

that

of a Line

line

coordinates of the point

segment PiP 2 in the ratio Ai/A 2f

such that

Since the triangles

PiQP and PRP 2

in Fig. 5 are similar,

we

have, by a theorem of plane geometry, that corresponding

sides of these triangles are
thus we have
proportional
;

(A
v

2^;

Since

rir

PP 2

ri **

PR

= P

RP 2

= x x\,
= Xz x,
= y- y\,
= y z - y>

~F
FIG. 5

17

Points and Lines in the Plane

it

follows from (4.1)

and

'

From

the equation of the

Xl

y2-y

two terms we have

first

fi2 (* 2

from which, on solving

(4 5)

xz-x

A2

(4.4)

x and y are such that

h = x-x i= y-yi.

(43)'
v

(4.2) that

[Chap, i

-*) = *-*,

for x,

we obtain

+ h2* 2

Proceeding in like manner with the equation consisting of the

and third terms in (4.3), we obtain

first

(4.6)

If the points PI,

P 2 and P are on a
,

or are on the #-axis, the situation

line parallel to the *-axis

expressed by (4.4), from

these points are on a line
is

which (4.5) follows; similarly if

parallel to the jy-axis or are on the y-axis,
hi

we have

(4.6).

In order to obtain the coordinates of the mid-point,

= h 2 = 1 in (4.5) and (4.6), and have the theorem

[4.1]

The coordinates

we put

of the mid-point of the line segment joining

and (x 2 y 2 ) are

the points (x\, y\)

(4.7)

Returning to the consideration of equations (4.3), we observe that they express the condition that the ratio of the segments PiP and PP 2 is equal to the ratio of the first direction
numbers of these segments, and also to the ratio of the second
direction numbers. Suppose then that we consider in connection with the line segment P\P 2 the line segment P 2 P, where
now P is the point (x, y) such that P*P has the same direction and sense as P P 2
then the ratio P \PJPP 2 is a negative number, since the segments P\P and PP 2 have opposite
{

18

and External Division

Sec. 4]

Internal

sense,

but their algebraic sum

of a Line

PiP 2

equal to

is

Segment
In this case

we again have equations

(4.3), in which, however, the quantity

a negative number, that is, either hi or A 2 may be taken

hi/h2 is
as negative and the other positive; but the numerical value
of hi is greater than the numerical value of A 2 since the length
of PiP is greater than the length of PP 2
With this understanding about h\ and A 2 the coordinates of P in terms of those
,

and P 2 are given by (4.5) and (4.6).

It is customary to denote the numerical, or absolute, value
thus 2 = 2,
of a number a by
a
2 = 2. Then the
h
above statement about hi and 2 is expressed by hi > h 2
On the other hand, if we have equations (4.5) and (4.6) with
< h% this means that
hi and /* 2 differing in sign, and
hi
x and y are the coordinates of a point P lying below Pi on the
line through PI and P 2 as the reader will see by considering
the above definition of P for the case h\ < A 2 and drawing
a figure. In accordance with custom we say that in these cases
P(#, y) divides the line segment PiP 2 externally in the ratio
Ai/A 2 and that when P is a point of the segment, it divides
the segment PiP 2 internally. Accordingly we have the theorem
of PI

|.

|,

1,

[4.2]

The equations

give the coordinates of the point

PiP 2

with end points PI(XI,

case

yi),

dividing the line segment

P 2 (* 2

and A 2 have

y%), in the ratio

same sign and

when hi
when h\ and h% have opposite signs; in either

hi/h2, internally

externally

lies

greater than

nearer PI or
A 2 |.

P2

the

according as \h\\

is

less or

EXERCISES
1. Find the coordinates of the points dividing the line segment
between the points (7, 4) and (5, - 6) in the ratios 2/3 and - 2/3.

2. Find the coordinates of the two points which

2) and (-3, 1).
segment between the points (3,
3.

In what ratio does the point

between the points

(5, 2)

and

(6,

(3,

4) ?

19

trisect

the line

2) divide the line

segment

Points and Lines in the Plane

[Chap,

Find the coordinates of the point P(x, y) on the line through

- 3) and P2 (2, 1) such that P2 bisects the segment PiP.

4.

the points Pi (5,

5.

Show

4),

and

that the medians of the triangle with vertices (- 1, 2),

6) meet in a point by finding for each median the
coordinates of the point twice as far from the corresponding vertex
(3,

(5,

as from the opposite side, distances being measured along the median.

P2 (x2

6. Is there a point on the line through P\(x\, y\) and

1 ?
dividing the line segment PiP2 in the ratio hi/h 2

7.

Where

are the points in the plane for which

UK 3?
8.

origin
9.

when

Show
if

>

y*)

2 and

that the points (x\,y\) and (x2t y2 ) are collinear with the
if their coordinates are proportional.

and only

Show

that the points

there are three

that ki

k2

(xi, y\), (x2 ,

numbers

+ kz = 0,

ktfi

k\ 9

k2

3,

yz ), and
all

+ k2 x2 -f fax* = 0,

10. Interpret the following

statements

(* 3

different

k&i

are collinear

jVa)

from

zero,

+ k2y2 +

fejvs

such

= 0.

and h 2 both positive x defined by (4.8) is the weighted

of
and x2 with the respective weights h 2 and hi, and accordx\
average
a.

For

hi

ing as h 2 is greater than hi or less than hi the weighted average is

nearer x\ than x2 or vice versa.
b. x and y given by (4.8) are the coordinates of the center of gravity
of masses h 2 at PI and hi at P2
.

11. What are the coordinates of the center of gravity of masses

of n masses
3 at the points (xi, ;yi), (x 2 y 2 ), and (* 3 jVa)
2 and
mi,
at n different points? (Use mathematical induction.)

5. An Equation of a Line.
Parametric Equations of a Line

We have

seen how any two numbers x and y define a point

the
in
plane, called the graph of the ordered pair of numbers.
Reversing the process, we have that any point in the plane

the graph of an ordered pair of numbers, the coordinates

of the point, obtained by drawing through the point lines
parallel to the coordinate axes and taking for these numbers
the magnitudes of the intercepted segments with appropriate
is

signs, as

OC and OD

in Fig.

a restriction on x and

An equation in x and y imposes

that only one of these may take

1.

y, so

20

An

Sec. 5)

Equation of a Line

and then the other is determined by the equaThe graph of an equation is the locus (place) of all points

arbitrary values
tion.

whose coordinates are solutions of the equation, in other words, the

locus of the graphs of ordered pairs of numbers satisfying the equation.
Each of the points of this locus possesses a geometric
property common to all the points of the locus, and no other
points this property is expressed algebraically by the equation.
Just what this geometric property is in the case of a particular
equation is one of the interesting questions of coordinate geom;

etry. The converse problem is that of finding an equation

whose solutions give the coordinates of every point of a locus
when the locus is defined geometrically. All of this will become
clearer as we proceed to take up particular equations and

particular

loci.

requires an ordered pair of numbers to define a

in
the
point
plane and any point in the plane is so defined, we

Since

it

say that a plane is two-dimensional. A line is one-dimensional,

since any point on the line is defined by one number, for
example, its distance from a fixed point on the line. Similarly,
since any point on a curve is defined by one number, for
example, its distance measured along the curve (say, with a piece
of string) from a point of the curve, we say that any curve is onedimensional.

The graph

of

an equation

in

x and y

is

some kind

for, as remarked above, only one of the

be chosen arbitrarily, the other being determined

of curve, or a line;

unknowns may
by the equation, and thus the equation picks out from the two-

dimensional set of points in the plane a one-dimensional set.

We consider now the graph of an equation of the first degree in % and y. When one takes such an equation, and, having
obtained a number of solutions after the manner discussed in
1, plots their graphs, one observes that they seem to lie on
a straight line. In fact, one may have been told that all one

has to do is to plot two solutions and draw a straight line

through the two points, and that this is the graph of the equation in the sense that the graph of every solution is a point of
the line and every point of the line is the graph of a solution.
In order to prove that this is a correct statement, we consider
first the inverse problem of finding an equation of a line.
21

Points

and Lines

in tKe

Plane

[Chap, l

Consider a line which is not parallel to either axis, and

choose upon it two particular points, PI of coordinates xi, y\
and P 2 of coordinates x 2 y 2 We denote by P of coordinates
x y any point on the line
meaning that it may be placed
line
while
in this
on
the
keeping PI and P2 fixed
anywhere
sense we speak of P as a general or representative point, and use
x and y without subscripts to distinguish a representative point
.

from a particular point. Through PI, P, P 2 we draw lines parto the axes, forming similar triangles, as shown in Fig. 6.

allel

Since the triangles are similar, corresponding sides are pro-

and we have

portional,

PiQ

P^

(5 1}
'

It

= QP

should be observed

that

in

each

ratio are taken in the

same

two segments

the

sense.

We

have PiQ

=x

/*

xi,

= y 2 yi.
(Although
6
has been drawn with all
Fig.
Q 2 P2

FIG. 6

the coordinates positive, the reader will readily verify that these
relations hold equally well when the figure is so placed with
reference to the axes that some or all of the numbers x\, y\, x,

y2 are negative.) Consequently (5.1)

equation
_
__

is

y, X2,

(5.2)

equivalent to the

"

Although this result has been derived for the case when P lies
on the segment PiP2, by drawing suitable figures the reader
can assure himself that equation (5.2) holds when P lies on the
line

but outside the segment PiP2. Equation

(5.3)

(y 2

This equation

yi)x
is

(x 2

satisfied

x\)y

(x 2 y\

(5.2) is reducible to

Xiy 2 )

= 0.

by the coordinates of every point on the

Moreover, from the form

(5.2) of the equation it follows

that any solution of the equation leads back to the ratios (5.1),
and consequently gives the coordinates of a point on the line.

line.

22

An

Sec. 5]

Equation of a Line

Although equation (5.3) was derived for a line inclined to

the coordinate axes, that is, when * 2 ^ xi and jy 2 ^ y\, it applies to the cases when the line is parallel to either axis. In
fact, if

in

the line

is

parallel to the #-axis,

which case equation

(5.3) is

we have ^2

= y\, x 2 ^

xi,

equivalent to

y-yi = 0.

(5.4)

Also,

when

x%

= XL,

y^

y\,

(5.5)

which

equation

(5.3) is

equivalent to

- xi = 0,

an equation of a line parallel to the jy-axis. These

from (5.2), if we adopt the principle that
an equation of two ratios either term of a ratio is equal to
other term also is equal to zero. Thus if y% = y\, we must

is

results follow also

when

in

zero, the

have y

= 0.

This does not say anything about the value

this value is determined by the values of
the terms of this ratio. Hence we have the theorem
y\

of the other ratio

[5.1]

(5.2), or (5.3), is

Equation

the points (xi, y\)

and

an equation of

the line through

(x%, yz).

Since x\ y\, X2, and y<z are fixed numbers, equations (5.2)
(5.3) are of the first degree in x and y, as are also (5.4)
9

and
and

[5.2]

(5.5).

Consequently we have the theorem

An equation of a straight line is of the first degree in x and y.

We say "an equation/' and not "the equation/' because, if

we have an equation of a line, so also is any constant multiple
of this equation

an equation of the

line, since

any solution

of

a solution of the other.

Having shown that in an equation of any line the coefficients of x and y are not both simultaneously equal to zero, we
remark that if, as stated in
1, in deriving theorems in this
chapter we exclude from consideration the degenerate case of
equations of the first degree for which the coefficients of both
x and y are zero, we are not thereby restricting the consideraeither equation

tion of

all

is

the lines in the plane.

23

Points

We

and Lines

in the

prove the converse of Theorem

shall

The graph of any equation of

a straight line.

[5.3]

We

Plane

(5.6)

[5.2],

the first degree in

consider the general equation of the

ax

[Chap,

first

namely,

x and y

is

degree

+ by + = Q
c

which not both a and b are equal to zero. For a = 0, b ^ 0,

equation (5.6) is reducible to the form (5.4) (with a change of
notation), which is an equation of a line parallel to the jc-axis.
For a T 0, b = 0, equation (5.6) is reducible to the form (5.5),
which is an equation of a line parallel to the ;y-axis. In 1 it
was shown that when a ^ 0, 6^0, equation (5.6) can be put
in the form (1.8) in terms of two of its solutions x\, y\ and #2, y*.
in

the same as equation (5.3). Hence any

equation (5.6) can be given one of the forms (5.3), (5.4), or
(5.5), and the theorem is proved.

But equation

(1.8) is

Returning to the consideration of equation (5.2), we rethat, when the line is not parallel to either axis, for each
point on the line the two ratios have the same value, this value,
say /, depending upon the values of x and y. If we put each

mark

of the ratios equal to

x and

y,

we

(5.7)

and solve the resulting equations

for

obtain

= xi + t(x 2 - *i),

= yi + t(y 2 - yi).

Conversely, for each value of / the values of x and y given by

(5.7) are such that equation (5.2) is satisfied, as is seen by sub-

and consequently these values

stitution,

of x

and y are the

coordinates of a point on the line. When the line is parallel to

the x-axis, equations (5.7) hold, the second equation reducing

= y\

and similarly for a line parallel to the ;y-axis. Thus

the single equation (5.2) of the line has been replaced by the
two equations (5.7) through the introduction of an auxiliary
variable /, called a parameter. Accordingly equations (5.7) are
called parametric equations of the line. Ordinarily, as will be
seen later, a curve in the plane is defined by one equation in
to

x and y, but sometimes it is convenient to define it by two

equations in x, y, and a parameter an example of this is found
;

24

Parametric Equations of a Line

Sec. 5]

in

mechanics when the coordinates of a moving particle are

expressed in terms of time as parameter. When the equations

of a curve are in parametric form, it may be possible to obtain
a single equation in x and y by eliminating the parameter from
the two equations. Thus, if we solve the first of (5.7) for / and
substitute the result in the second, we get an equation reducible
to (5.2). Again, consider the locus of a point whose coordinates
2
are given by x
2 /, as / takes on all values. Eliminating
t
y
2

we

=
= 4 x,

which the reader who has plotted curves in

get y
his study of algebra will identify as an equation of a parabola.
The use of a parameter in defining a line, or a curve, em/,

phasizes the fact that a line, or a curve, is one-dimensional in

that any point on it is specified by the appropriate value of a
single variable, a parameter. In fact, in plotting the graph of

line,

or a curve, defined by parametric equations one assigns

parameter and plots the points having

different values to the

as coordinates the values of x

and y thus obtained

it is

not

to eliminate the parameter from the two equanecessary

tions and then plot the resulting equation in x and y (see Ex. 11).
first

When

equations (5.7) are written in the form

(5.8)

(1

- 0*i + tx 2

(1

O^i

+ ty*>

we see that y is equal to the sum of the same multiples of y\

and y 2 as x is of x\ and x 2 We say that x and y have the same
linear and homogeneous expressions in xi, x% and y i9 y 2 respectively, meaning that every term in each expression is of the
first degree in these quantities. Accordingly we have
.

[5.4]

point of a line is expressible linearly and homogeneously in terms of two fixed points of the line.

Any

Any two points of

the line can be used for PI and

P2

that

is,

as the basis for writing the equations of the line in the form
(5.8). The above theorem is the geometric equivalent of the

algebraic statement
[5.5]

Any

solution of

linearly

an equation ax

and homogeneously
25

+ by + c =

in terms of

is

expressible

any two

solutions.

Points and Lines in the Plane

[Chap. 1

EXERCISES
Obtain equations of the

1.

a.

(-

Show

2.

1,

8)

and

(4,

that the points

3.

Find

4.

Where

(4,

3), (2, 0),

(c,

- 2),

are the points for which x

are the points for which 2 x

5.

Show

6.

Show

7.

In what ratio

(-

>

0,

0,

1).

on a

and (-

y>

(3,

2, 6) lie

(3, 1),

3y>

Where

and

b. (3, 2)

so that the points

collinear, that is, lie on a line.

is

determined by the points

lines

- 2).

line.

2, 4) shall

be

+y<

1?

and x

2?

that the equation of the second and third ratios in (4.3)

reducible to equation (5.3). What does this mean?
/

= hi/ (hi + h2

).

the line segment with end points (2, 3) and

the point of intersection of the line segment and
is

divided by
-2
6

1)

(3,

that (4.8) are of the form (5.8) with

y- = 0?

the line 3 x

Show

for each set

that the equation a(x
x\)
yi)
b(y
of values of a and b is an equation of a line through the point (xi, yi).
For what values of a and b is it an equation of a line parallel to the
8.

y-axis
9.

an equation of the

Show

line

through the point

that the line ax -f by

in the points Pi(

c/a, 0)

and

+c=

P2 (0,

(x2 ,

y2 ) ?

meets the *-axis and

c/b) respectively;

when

;y-axis

0,

c/a and
c/b, are called the
^-intercept and y-intercept respectively of the line.
the lengths OPi and OP*, that

10.

Show

that

is,

when an equation
g

of a line

+h=
*

is

written in the form

1
l'

where g and h are constants, g and h are the x- and ^-intercepts respectively. When can an equation of a line not be given this form, which
is

called the equation of the line in the intercept

form ?

curve with the parametric equations x = 2 t 2 y = 3 1;

also the curve x = 5 cos /, y = 5 sin /, using a table of natural sines
and cosines. What is the equation of the curve in each case when
the parameter is eliminated ?
11. Plot the

Show that,

the expressions (4.8) for x and y are substituted in

the value of hi/h 2 obtained from the
result is the "ratio in which the segment PiP* is divided by the point
of intersection of the line ax + by
c
and the line through PI and
12.

an equation ax

if

+ by + c = 0,

+ =

P2 The
.

reader should apply this method to Ex.

26

7.

Sec. 6]

Numbers and

Direction

6. Direction

Numbers and

Direction Cosines of a Line

Direction Cosines of a Line.

Angle between

Two

Lines

The equation

(xi, yi) and (x 2 y 2 ), obtained in

the
condition
that
the direction numbers of the
expresses
and
are
PiP
2
P\P
segments
proportional when P is any point
on the line. In fact, in Fig. 6 the sensed segments PiQ and QP

of the line through the points

5,

give the direction numbers of PiP, and PiQ 2 and Q 2 P 2 the direction numbers of P\Pz. Thus equation (6.1) is the algebraic

statement of the characteristic property of a line that any two

segments having an end point in common have the same direction.
If we say that direction numbers of any segment of a
line are direction numbers of the line, it follows that a line has

an endless number of sets of direction numbers, but that the

numbers of any set are proportional to those of any other set.

The equation
(6.2)

by x\ and y\, being of the first degree in x

an equation of a line through the point (x\, y\} and
with direction numbers u and v. In order to find a segment of
which u and v are direction numbers, we have only to find x 2
and y 2 from the equations

which
and y,

is satisfied
is

X2

Then PiP 2

is a
having u and v
equation (6.2)
have the same

[6.1]

//

y 2 -yt

= v.

segment with end points PI(XI, y\} and P 2 (x 2 y 2 )

as direction numbers and in terms of x 2 and y 2
becomes equation (6.1). Since parallel lines
,

by

definition, it follows that

y2 )

are

direction

(xi 9 y\)

and y 2
any line

- xi = u,

and

(x 2

any two points on a

y\ are direction numbers of the line,

parallel to

line,

and

x2

x\

also of

it.

3 we
In defining the direction cosines of a line segment in
assigned sense to a segment, making the convention that for

27

Points

and Lines

in the

Plane

[Chap. 1

a segment not parallel to the x-axis upward along the segment

and for a segment parallel
positive and downward negative
to the x-axis the positive sense is to the right along the segment. Accordingly the positive sense along a line is upward,
that is, y increasing, when the line is not parallel to the
x-axis; and to the right, that is, x increasing, when the line
Since all segments of a line have the
is parallel to the x-axis.
is

same
same

direction, the direction cosines of all segments are the

we call them the direction cosines of the line. Accord-

ingly the direction cosines X, ^ of a line are the cosines of the

angles between the positive direction of the line and the positive
directions of the x-

From Theorem
[6.2]

The

and y-axes respectively.

[3.2] we have

direction cosines X,

/x

is

never negative;

p = 0,

of a line satisfy the equation

+M

=1;
when ju>0
X

(6.3)

ju

<

<

when

and, conversely, any two numbers X, /x satisare direction cosines of a line.

conditions
these
fying

In Figs. 7 and 8

(6.4)

cos A,

fj,

cos B.

FIG. 8

FIG. 7

Since equations (3.5) hold for all segments of a line with the
point (xi, y\) as an end point of the segment, and consequently
for the line, we have
[6.3]

(6.5)

The equations
x

= xi + d\

= yi + dp

are parametric equations of the line through the point

28

(x\,

y\)

Angle between

Sec. 6]

and with

X and

direction cosines

distance

from
and is

Two

Lines

/z;

the point (%i, y\) to

the parameter d is the

a representative point

positive or negative according as the latter

above
or
below P\ along the line, or to the right
point
or left of PI when the line is parallel to the x-axis.
(x,

y),

is

With the understanding that for any quantity A the symbol

= a
means the positive square root of A, and that
in
defined
4, we shall prove the theorem

V^

as

[6.4]

The direction cosines of a line for which u and

of direction numbers are given by

|,

v are

any

set

(6.6)

1 or
where e is
eu is positive when

In

1 so that ev is positive
v

when

0,

and

= 0.

the quantities (6.6) satisfy the condition (6.3) of

[6.2]; and the requirement in Theorem [6.4] conis positive when not equal to zero, and
is such that

fact,

Theorem

cerning e
that when

/JL

ju

= 0,

then X

From Theorem

which proves the theorem.

using the Greek letter
(phi), we
1,

<

[3.3],

have
[6.5]

The angle

<f>

between the positive directions of two lines

whose direction cosines are

cos

(6.7)

From Theorems
[6.6]

[6.5]

and

Xi,

jui

XiX 2

[6.4]

and

\2,

^2

is

given by

+ MiM2.

we have

The angle </> between the positive directions of two lines whose
numbers are u\, v\ and w 2 v 2 is given by

direction

(6.8)

where

and

e\ is

1 or

e\u\ is positive

1 so that e\v\ is positive

when

vi

29

when

Vi 7* 0,

and similarly for

62.

and Lines

Points

As a

lines with direction

pendicular

numbers

one another,

to

Uiu 2

(6.9)

Consider

now an

[Chap,

if

u\, v\

and only

and u 2

v 2 are per-

if

+ ViV2 = 0.

equation

+ by + c = Q

ax

(6.10)

of a line.

Plane

we have

corollary of the theorem

Two

[6.7]

in the

If the point (*i, y\) is

on the

line, c

must be such that

+ byi + c = 0.

axi

Subtracting this equation from (6.10),

we have

as an equation

of the line

a(x-xi)

(6.11)

When

+ b(y-yi) = 0.

a and b are different from zero, and this equation

x _ Xi

written in the form

~~T~

= y yi
-a

is

'

a are direction numbers of the line, and

we see that b and
consequently of any line parallel to it. Hence we have

An

[6.8]

equation of any one of the endless

ax
is
c
by

allel to the line

(6.12)

+ =
ax + hy + d = 0;

a particular one of these

number of

lines par-

lines is determined by the value of

the coefficient d.

For example,
point

(3,

if

we wish

to find an equation of the line through the

2) parallel to the line

+ = 0,

d
(- 2)
4(3)
desired line is 4 x

that

14

is,

d=

4^-^ + 5 = 0,
-

14,

d is determined by
and hence an equation of the

= 0.

We

seek next an equation of any line perpendicular to the

with equation (6.10). If we denote by u and v direction
numbers of such a perpendicular, and note that &,
a are diline

numbers of the line (6.10),

u and v must be such that

rection

that

bu

av

30

it

follows from

= 0.

Theorem

[6.7]

Direction

Sec. 6]

This condition

Numbers and

Direction Cosines of a Line

by u = a, v
Hence we have

is satisfied

multiple of a and

6.

= &,

and by any constant

The geometric significance of the coefficients a and b of

an equation ax + by + c = Qofa line is that a and b are
direction numbers of any line perpendicular to the given line.

[6.9]

In particular, an equation of the line perpendicular to the

with equation (6.11) at the point (x\, y\) of the latter is

line

x-xi = y-yi_
a

This

may

be stated as follows

- *i) + b(y - y$ = 0,

^^ = y^=f

result, illustrated in Fig. 9,

[6.10]

Either of the lines

a(x

(6.13)

is the

perpendicular

to the other at the

point

(x\, y\).

FIG. 9

When

the second of equations (6.13)

(6.14)

we see that the coefficients of x and y are

Hence we have
[6.11]

is

written in the form

- *i) -a(y- yi) = 0,

b(x

Any
ax

line perpendicular

+ by + = Q
c

to

the

and

line

a respectively.

with the equation

has for an equation

bx-ay + d = Q;

(6.15)

a particular

line is

determined by the value of

31

d.

Points and Lines in the Plane

[Chap,

EXERCISES
Find an equation of the

1.

to the line

line

through the point

(3,

Find an equation of all lines perpendicular to 2x

one passing through the point

3.

in particular of the

Find

4.

-f

and

parallel

4jy+l =

2. What are the direction cosines of the line 3 x

of a line perpendicular to it?

and

0)

3x-y + 5 = 0.

equations

of

and 4#

tively to the line

the

lines

through

= parallel
2x + 5y-3 = Q.
1

the

+ 5y

0;

= 0,

1).

(1,

intersection

and perpendicular

of

respec-

5. Find the angle between the positive directions of the perpendiculars to each of the following pairs of lines
:

a.

5x + y+l =

Q.

Find the angles between the positive directions of the

6.

in

2x-7y + 3 = Q,

Ex.

lines

5.

Show

7.

that the coordinates of any point on a line through the

numbers of the line.

origin are direction

Show

8.

that equations (6.5) hold for a line parallel to either

if we make the convention that "the angle between"

coordinate axis

the positive directions of two parallel lines

9.

Show

zero.

is

that for the lines in Figs. 7 and 8 the equation (6.3)

A + cos 2 A = l.

is

2
equivalent to the trigonometric identity sin

10.

Show

that
(x

* i)( * 2

*3)

(y

~ y^* - *) = o

two lines through the point (xi, y\) parallel and perpendicular respectively to the line through the points (# 2 yz) and
are equations of

(*s, y*)-

Given the

11.

triangle

whose

vertices are A(a, 0),

B(0,

b),

and

prove that the perpendiculars from the vertices of the triangle upon the opposite sides meet in a point that the perpendicular
bisectors of the sides meet in a point. Does this prove that these
C(c, 0)

results are true for

12.

any

Discuss equation

direction

numbers

is

triangle ?
(6.2)

and Theorem

zero.

32

[6.4]

when one

of the

The Slope

Sec. 7]

7.

By
by m,

of a Line

The Slope

of a Line

definition the slope of a line, which is usually denoted

is the tangent of the angle which the positive direction

of the line, as defined in

6, makes with the positive direction
of the #-axis
it is the tangent of the angle measured from the
;

*-axis to the line in the counterclockwise direction (as is done

in trigonometry). In Fig. 7 the angle at PI is equal to A, and

consequently

(7.1)

In Fig. 8 the angle at PI

tan (180
v

which
It is

is

the supplement of

- A) =
'

^=

equivalent to (7.1), since tan (180

customary to write (7.1) in the form
is

y-yi = m(x-

(7.2)

consequently

A)

tan A.

*i),

as an equation of the line expressed in terms of the slope and a

point (xi, yi) on the line.

When

the line

is

parallel to the j-axis, the

denominator in

the right-hand member of (7.1) is zero. Hence we say that the

slope is not defined for a line parallel to the jy-axis some writers
;

say that in this case the slope

is infinite.

x\ and y
In equation (7.1) the quantities x
y\ are direction
and
line
of
the
numbers
consequently of the line
segment P\P
are
and
since P(x, y)
any points of the line, we have
P\(x\, y\)
;

[7.1]

The slope

m of a

line not parallel to the y-axis is equal to the

and first of any set of direction numbers

in particular,
(y 2
yi)/(x*
*i), where
(* 2 3^2) are any two points of the line.

ratio of the second

of the line;
(*i, y\)

Equation
(7.3)

and

(7.2) is

m=

a particular case of the equation

= mx+

h,

where h = y\
mxi. For a given value of m equation (7.3) is
an equation of all lines with the slope m, that is, a set of parallel
33

Points

and Lines

in the

Plane

[Chap. 1

line of the set being determined by a suitable value

In particular, equation (7.2) is an equation of the line
of the set through the point (*i, ;yi). An equation
lines,

of

any

A.

ax

(7.4)

for

which

6^0,

+ by + c = 0,

can be put in the form

we have

(7.3)

by

solving the

given equation for y. Hence

// an equation ax

[7.2]

y-axis

equation

is the

As a consequence

Two

[7.3]

+ by + c = Qofa

if

of theorems [7.1]

and only

The proof

the resulting

and

[6.7]

we have

m\ and m^ are perpendicular

to

if

miw 2 +

(7.5)
is, if

x in

slope of the line.

lines with the slopes

one another,

that

line not parallel to the

solved for y, the coefficient of

is

= 0;

either slope is the negative reciprocal of the other.

is left

to the reader.

We

have introduced the concept of slope because of its

and particularly because of its use in the application of the differential calculus to the study of lines and curves
in the plane, an interesting subject awaiting the reader. On
the other hand, we have emphasized direction numbers and
direction cosines because they may be applied to any line withtraditional use,

out exception, and, as we shall see in Chapter 2, are applicable

to the study of lines in space, whereas the concept of slope is

not used in this connection.

EXERCISES
1.

Find an equation of

all

lines parallel to

particular the one through the point

(1,

=3#

5,

and

2. Find an equation of the line through the origin whose slope

twice that of the line
+ 2 Q.

3x-y

in

3).
is

Find an equation of the line through the point (2, positive direction makes the angle of 60 with the positive
3.

34

3)

whose

*-axis.

Directed Distance from a Line to a Point

Sec. 8]

4. Find equations of the two lines through the point (1, 3)

forming with the x-axis an isosceles triangle with the point as vertex
and the base angles 30

For what value of

5.

joining the point

is

the line (7.2) perpendicular to the line

to the origin?

(x\ 9 y\)

Using the formula from trigonometry for the tangent of the

two angles, show that an angle between two lines whose
slopes are mi and m2 is given by
"* 2 "" mi
tan =
6.

difference of

1 -f
7.

How

8.

Find equations of the

make

is

Theorem

angles of 45

a consequence of Ex. 6?

[7.3]

(xi, y\) which

verify that they are per-

through the point

lines

with the line

and

(7.2),

pendicular to one another.

9.

Show

2
that the graph of the equation ax 2
2 hxy
is
by
lines through the origin if h 2 > ab. What are the

a pair of straight

slopes of these lines ?

10. Draw the graph of a line and indicate the various quantities
appearing in equations (6.1), (6.2), (6.5), and (7.2).

8.

When

Directed Distance from a Line to a Point

a point P\(x\

y\) is

line

+ by + c = 0,

ax

(8.1)

on the

c is equal to zero, as we have seen.

the expression ax\
by\
PI is not on the line, this expression has a value different

When

from zero. One might question whether perhaps this value has
something to do with the distance from the line to the point.
We shall answer this question by showing that this distance is
0*1

y\

_,

V0 2 + b 2

ancj gj ve a

mean n g
j

to the algebraic sign of the

resulting number.

When

and

b^

in (8.1), that

is,

when

the line

is

parallel to the *-axis, the value of y for each point on the

line is
c/6, and consequently the distance of Pi from the
line is y\
c/&, and this distance is positive
(
c/b), that is, y\

or negative according as PI

lies

35

above or below the

line

we

Points

and Lines

in the

Plane

[Chap. 1

in the
get this expression for the distance when we put a =
expression at the close of the preceding paragraph. When
6
and a 7* 0, that is, when the line is parallel to the
c/a, which is positive or negative
.y-axis, the distance is x\
according as P\ lies to the right or left of the line this expres-

sion

from the one of the preceding paragraph on

follows

putting b = 0.
We take up next the case when the line is inclined to the
6 we made the conven#-axis, in which case a ^ 0, b ^ 0. In
tion that the direction upward along a line inclined to the
axes is positive, and showed that the direction cosine JJL is

always positive, and that X

is positive or negative according as

inclined as in Fig. 7 or in Fig. 8. In consequence of
this result, and of Theorems [6.9] and [6.4], we have that the

the line

is

direction cosines of a line perpendicular to the line ax

are given

by
X

(8.2)

where

when

We

+
= 0.

e is

or

consider

the line (8.1)

is

=
1

first

+ by + c =

so that eb

the case

>

when

and ea

>

when

inclined as in Fig. 10,

where P\(x\, y\) is a point above the

line and P 2 (*2, y 2 ) is the point in
which the perpendicular through Pi
to the line meets it. We denote by
d the distance from P 2 to PI, d
being a positive number, since it is
in the positive direction

measured

of the line P2Pi.

line in Fig. 10,
line

P\P

FIG. 10

this

Comparing
whose direction cosines are X and

in Fig. 7,

we

see that in place of (6.5)

= d\,

(8.3)

xi

X2

from which and

(8.2)

we have

yiy2 = dp,
db

(8.4)

36

JJL,

with the

we have

Directed Distance from a Line

Sec. 8]

Since (* 2 j 2 )

is

to

a Point

a point on the line (8.1), on substituting the exand rearranging terms, we have

pressions (8.4) in (8.1)

axi

from which, on solving

When

+ b * \ = 0,
2
2
\e V0 + b /
a*

+ by + c - d (
for

d and noting that 1/e

et

we

obtain

below the line, equations (8.3) hold equally

d is a negative number, as the reader may verify by
drawing a figure and noting that when the line (8.1) is inclined
as in Fig. 10, X is positive.
PI

lies

well provided

We

consider the other case

the line (8.1)

and

this

is

when

inclined as in Fig. 11,

time take PI below the

line,

add variety to the discussion.

Comparing the line PiP 2 in Fig. 11
with the line PiP in Fig. 8, we have
so as to

from

(6.5)

_x ^
where d

is

dx

FlG<

=d

a positive number, the directed distance PiP 2

Com-

paring these equations with (8.3), we see that the left-hand

members of these respective equations differ only in sign. Con-

sequently equations (8.3) apply in this case also, with the

understanding that d is negative; it is the directed distance
P 2 Pi, that is, PI is on the negative side of the line. The reader
may verify that equations (8.3) hold also when PI is above the
line in Fig. 11
[8.1]

and d

is

positive.

Hence we have

The directed distance of a point PI(*I, y\) from

Qis given by
ax + by + c

the line

.__
(8.5)

where
if

e is
;

+1

or

1 so that eb

the distance

>

is positive

if

the point lies above or below the line.

37

0,

and ea

>

or negative according as

Points

and Lines

in the

Plane

[Chap. 1

is the distance of the point (2, 1) from the

Since the coefficient of y is a negative number,

As an example, what
x

line
e

=-

3.y-f2
1 in (8.5),

= 0?

and we have
2

- VI6
Since d
verify

We

is

positive, the point lies

VlO

above the

line,

may

as the reader

by drawing the graph.

observe from (8.5) that the origin, that
above or below the line according as ec

(0, 0), is

is,

is

the point
positive or

negative.

Consider

now

the equation

(8 6)
.

where k is a constant. Since the equation is of the first degree

and y, it is an equation of a line. If P\(x\, y\) is any point
on this line, on substituting x\ and y\ for x and y in the above
equation and comparing the result with (8.5), we see that PI is
at the distance k from the line (8.1). Since this is true of every
point on the line (8.6), it follows that the line (8.6) is parallel
to the line (8.1) and at the directed distance k from it; that
the two lines are parallel follows also from the fact that equations (8.1) and (8.6) satisfy the condition of theorem [6.8].
in x

EXERCISES
1.

Find the distances from the

-1) and (-

(1,

2, 4),

and

points parallel to the given

Find

the

distance

;y

+5=

to the points

line.

Find equations of the two

at the distance 2 from this line.
2.

3.

line

find equations of the lines through these

lines parallel to

between

the

lines

2*

4#

3 j>

;y-f6

-f

=
and

4*-2;y-3 = 0.
4.

4x

Find the points on the *-axis whose distances from the

+ 3y - 6=

5.

How

line

are numerically equal to 3.

far is the origin from the line through the point

parallel to the line

3x

+ y = 0?

38

(-

3, 2)

Two

Sec. 9]

Equations of the First Degree in x and y

Find the points which are equidistant from the points (2, 6)
50 = 0.
2, 5) and at a distance of 2 units from the line 7x + 24y

7.

Show

6.

and

sected

that the circle of radius 2 and center

by the

line

4*-?:v +

28

(3,

Find the area of the triangle with the vertices

8.

is

4)

not inter-

= 0.
(2, 3), (5,

1),

(3, 1).

be multiBy what factor must the equation ax -f by 4- c

0, where X and M are
plied so that it is of the form \x
p
the direction cosines of any line perpendicular to the given line?
Show that p is the directed distance from the line to the origin.
9.

10. Prove that the line 2 x

plane between the two lines

4y + & = Q bisects the portion of the

*-2j> + 2=0 and x-2y + 6 = Q.

Prove that the line 3 * -f 4 ;y

quadrilateral whose vertices are (
11.

9.

Two

w+ =

12

4, 6),

bisects the area of the

(7,

),

(1,

-f

),

- 3).

(8,

Equations of tke First Degree in x and

Determinants of the Second Order

y.

Having discussed in 1 the solutions of an equation of the

first degree in x and y, we now consider two such equations,
namely,
ai*

(9.1)

+ biy + ci = 0,

where a subscript

2*

+ b2y + c2 = 0,

attached to a coefficient (as, for example, 0i)

a coefficient of the first equation and has
do
with
to
the numerical value of the coefficient;
nothing
are
this
coefficients of the second equation
similarly #2, &2, 2
same
letter
for
the
the
enables
to
use
of
one
device
subscripts
distinction
the
x
in
of
the
two
coefficients
being inequations,
for
the
coefficients of y
dicated by the subscripts
similarly
with
the
discussion in
and the constant terms. In accordance
1, each of equations (9.1) has an endless number of soluindicates that

it is

tions

and

x and y in the

in the second x

first

equation represent any of

and y represent any of

its

its

solutions,

solutions.

When

the left-hand member of either equation is a constant multiple

of the left-hand member of the other, the two equations have
1
that is, they are not
the same solutions, as remarked in
;

39

Points

and Lines

in the

Plane

[Chap.

independent equations, and consequently x and y have the

same meaning in both equations. But when the equations are
essentially different, x and y have, in general, different mean-

two equations. To emphasize this fact, different

might more appropriately be used in place of x and y

ings in the
letters

but this is not the general practice;

to the customary practice of using the same letters
in both equations and expect the reader to bear the

in the second equation,

we adhere
x and y

distinction in mind.

not be that for given values of the coefficients

a solution of the first which is also a solution of the
second; that is, the two equations may or may not have a
common solution. We assume that there is a common solution,
which we denote by x\, y\, and substitute it for x and y in the
It

may or may

there

is

two equations, obtaining

+ biyi + ci = 0,

(9.2)

a 2 xi

+ b yi + c = 0.
2

Following the method with which the reader is familiar, we

multiply the first of these equations by b 2 and from the result
in the resulting equation
subtract the second multiplied by hi
of
is
and
we
obtain
coefficient
the
zero,
y\
;

(0i 6 2

By

- 02&i)*i +

(cib 2

c 2 bi)

= 0.

adopting the following shorthand notation for the quantities

in parentheses

(9.3)

02

the above equation

(9.4)

In like manner,

by

0i

if

may be

written

0i

Ci

02

c2

= 0.

we multiply the second

and from the

we obtain
(9.5)

of equations (9.2)

result subtract the first multiplied

0i

0i

02

02

40

= 0,
C2

by

2,

Determinants of

Sec. 9]

where, similarly to

(9.3),

trie

Second Order

we have put

ai

Cl

Q>2

C2

(9.6)

The square

arrays defined by (9.3) and (9.6) are called deteruse the sign of identity =, rather than the sign
of equality, to indicate that (9.3) and (9.6) are definitions.
We observe that the second determinants in (9.4) and in

We

minants.

(9.5) are

and the

obtained from the

fr's

by

c's

first

on replacing the

a's

by

c's

respectively.

We
first

consider equations (9.4) and (9.5) in detail. Suppose

that the 0's and &'s have such values that the determinant

ai

$2

U2

is

Then equations

not equal to zero.

(9.4)

and

(9.5)

can be solved at once for x\ and y\, the common solution;

this is the process with which the reader is familiar, although
maybe not in this notation. Moreover, since any common
solution of equations (9.1) must be a solution of equations (9.4)
and (9.5), there is only one common solution in this case. We
call the above determinant, that is, the determinant of the
coefficients of x and y in equations (9.1), the determinant of the
equations. Accordingly we have
[9.1]

When

the determinant of two equations of the first degree in

x and y

common

is

not equal to zero, there is one

and only one

solution of the equations.

two lines with equations (9.1) meet in one

one
and
only
point, that is, the lines intersect one anpoint
if
their equations have one and only one
if
and
only
other,
common solution, in which case the point of intersection is the
It is evident that

graph of the
equivalent of
[9.2]

common
Theorem

solution.

Accordingly the geometric

the following:

[9.1] is

The lines with equations (9.1) intersect one another, that

have one and only one point in common, if
**

(9.7)

41

^0.

is,

Points

and Lines

in tKe

Plane

[Chap. 1

We consider next the case when the determinant of the equations (9.1)

is

equal to zero, that

02^1

01 &2

(9.8)

If at the

same time

is,

Ci

or

C2

= 0.
01

c\

02

C2

0, it follows

that

cannot be true. This means that the assumption

a common solution of equations (9.1) is not valid,
not that one has proved that zero is equal to a number which
is not zero. Hence we have the theorem
(9.4) or (9.5)

that there

is

Two equations (9.1) do not have a common solution if the

determinant of the equations is equal to zero and one of the
determinants

[9.3]

0i

Ci

(9.9)

c2

is

62

Ci

c2

not equal to zero.

We shall show that in this case the equations have a form which
and to this end we consider separately
and
the case when a\ = 0.
a\
Case 1. ai ^ 0. We cannot have 02 = 0, otherwise it follows from (9.8) that b 2 = 0, and thus the second of equations
(9.1) is of the degenerate type which we have excluded from
is

readily distinguishable,

the case

when

our consideration (see

1). Since a 2 ^ 0, a number k (^ 0) is
defined by k = 02/01, from which we have 02 = ka\. When
this value for 02 is substituted in (9.8), we obtain the second
of the following equations

02

(9.10)

Case
wise

bi

2.
From (9.8) it follows that 2 = 0,
ai = 0.
= 0, and the first of equations (9.1) is degenerate.

other-

Since

a similar reason, the second of equations (9.10) is

satisfied by k = b 2 /bi, and the first of equations (9.10) is satis6 2 7*

for

fied identically, since 0i

Thus equations
when the

versely,

(9.10) are

= 0.
a consequence of

(9.8),

and, con-

coefficients of equations (9.1) are related as

42

Two

Sec. 9j

Equations of the First Degree in x and y

in (9.10) for some value of k, equation (9.8) holds, as is seen

by substitution. Furthermore, when (9.10) holds and either

and

of the determinants

must have

Two

[9.4]

c%

kc\.

Cl

02

C2

Accordingly

is

not equal to zero,

we

we have

equations (9.1) are of the form

aix

(9.11)

ai

+ b\y + c\ = 0,

where k

is

k(a\x

some constant

+ fay) + c 2 =

different

from

zero, if

(c 2

^*Ci),

and only

if

the determinant of the equations (9.1) is equal to zero and

one of the determinants (9.9) is not equal to zero; in this

case there is no
If
c 2 /k

we

= d,

common

divide the second of equations (9.11) by k and put

then d ^ d. If then we say that the second of
d
to within a constant factor, we may
a\x
b\y

+ =

(9.11) is
state Theorem [9.4] as follows
[9.5]

solution of the equations.

Two
mon

equations of the first degree in x and y have no comsolution if to within possible constant factors they are

of the forms
(9.12)

ax-\-by

+ c = Q,

ax

+ by + d = Q

(d^c).

In accordance with Theorem [6.8], two lines with equations

6 they have
that is, by the definition in
(9.12) are parallel
;

the same direction.

By Theorem

[9.5] this definition is equiva-

lent to the definition that they do not have a point in common.

The reader may have been told that two parallel lines meet in

a point at infinity. A point at infinity so defined is not like the

it is a concept sometimes
points with which we are dealing
he may make the general
so
that
introduced by the geometer
statement that any two noncoincident lines meet in a point.
;

We

consider finally the case when all three of the determinants in (9.4) and (9.5) are equal to zero, that is, when we have

equation (9.8) and

(9.13)

Cife

czbi

= 0,

0iC2

atfi

= 0.

In this case any value of xi and any value of y\ satisfy equaand (9.5). This does not mean that any value of

tions (9.4)

43

Points

and Lines

Plane

in the

[Chap.

x and any value of y is a common solution of equations (9.1),

1 that any value of x and any value
because we know from
of y is not a solution of either equation. We shall show that it
means that either equation is a constant multiple of the other.
We have seen that equation (9.8) is equivalent to the two
equations (9.10). When the values for a% and 6 2 from (9.10)

we find that the latter are satisfied,

From this result and (9.10) we have the

are substituted in (9.13),

and only
theorem

if

[9.6]

if c 2

kci.

All three of the determinants

(9.14)

0i

c\

02

are equal to zero,

if

and only
02

(9.15)

01
C2
if

01

with the understanding that if either term in any ratio

equal to zero so also is the other term; for example,
62

= 0,

so 0/so is bi

This theorem

is

is
if

= 0.

equivalent to the statement that

all

three

of the determinants (9.14) are equal to zero, if and only if

either of equations (9.1) is a constant multiple of the other, in

which case any solution of either equation

other. Hence we have the theorem

is

a solution of the

Equations (9.1) have an endless number of common solu-

[9.7]

tions, if

and only

if all three

of the determinants (9.14) are

equal to zero.

The geometric
[9.8]

is

the following

Two lines with the equations (9.1) are coincident, that is,
coincide at every point, if and only if all three of the determinants (9.14) are equal to zero.

The above
last

equivalent of theorem [9.7]

discussion

is

set forth in the following table, the

column giving the number

of

44

common

solutions of the

two

Two

Sec. 9J

Equations of the First Degree in x and y

equations (9.1), or, what is the geometric equivalent, the number of common points of two lines with these equations
:

One

a2 b2

None

not both zero

C2 b 2

02 C2

0i,bi,a2 b2 not
,

02 b2
C

&i|_|0i
b2

c2
Ici

I|

02 c2

zero

all

An endless

number

= QJ

both

Ci,c 2 not

zero

Ci

= c2 =

None

Any

x and

anyjy

The

case a\
b\
that is, the case

62

was not discussed

in the text,

when both

of equations (9.1) are degenerate,

but the situation in this case is given in the table so that the

algebraic treatment may be complete

the statements made.

the reader should verify

In illustration of some of the above results

ing equations

we consider

the follow-

22. Hence,
of these equations is 2(
5)
4(3) =
accordance with Theorem [9.1], these equations admit one and
only one common solution, given by equations (9.4) and (9.5), which

The determinant
in

in this case are

that
If

is,

xi

22

1/2, y\

now we

*i

take the

and the equation

11

= 0,

22 yi

+ 22 = 0,

1.

first

of the above equations, 2 x

3y 4 0,
3
0, we find that the determinant of

4x-h6> +

these equations is equal to zero, and the second determinants of (9.4)

- 33 and 22 ; consequently the
(9.5) are equal respectively to

and

given equations have no common solution, by Theorem [9.3].

When, however, we consider the equations 2* + 3;y 4 = and
4 * + 6 ;y
8 = 0, all three of the determinants in equations (9.4)
and (9.5) are equal to zero, and in accordance with Theorem [9.7] any
this is seen to
solution of either equation is a solution of the other
;

be the case when we note that the second equation

tiplied

by

2,

an

illustration of

Theorem

45

[9.6].

is

the

first

mul-

Points

and Lines

in

trie

Plane

[Chap. 1

EXERCISES
common solution of the equations 2x 3.y + 4 =
+ y + 2 = 0, using determinants, and check the result by solving

Find the

1.

and x

the equations

by the method previously known by the

reader.

y -f 4 =

and ax 2 y + c = 0, for
2. Given the equations 2x
what values of a and c have these equations one common solution
no common solution an endless number of common solutions?
;

Show

3.
I

gai

+ hdi

What

4.

that

vertices are

bi

5.

bi

b2

hdi
b2

b2

b2

are equations of the diagonals of the quadrilateral whose

(0, 0), (0, 0), (a, b\ (0, b) ? Find their point of intersection

and show that they

2x +

gai

gat

bisect

one another.

For what values of a and b are the lines ax + 8 jy + 4 = and

ay + b = Q coincident? For what values are they parallel?

6. Show that the angle <t> between the positive directions of the
perpendiculars to two intersecting lines (9.1) is given by

cos

where

and

is -f 1

or

similarly for e2

so that e\b\

>

if bi

0,

and

ciai

>

if bi

-2 (3 x + y- 5) + 5 (x
4) vanx and y take the values of the common solution
+ :y 5 = and x 2 y 4 = 0. Is it necessary
to find the common solution to verify the above statement? Would
the statement be equally true if the multipliers
2 and 5 were re7.

Show

-2y-

that the expression

ishes identically when

of the equations 3#

placed by any other numbers ?

8. Show that as a result of the whole discussion in
9 we may go
"
"
"
"
back and put if and only if in place of if in all the theorems of
9 in which "if" alone appears.

9.

Using Ex.

3,

show that
b\c 2
-f-

is

a\d\ 4- b\d%

b2 c 2

equal to the product of the determinants

46

ai

l
.

02

and

.*'

The

Sec. 10]

Set of Lines through a Point

The Set

10.

From Theorem

[9.2]

of Lines through a Point

and the other theorems

in

it

follows

that the lines with equations

aix

(10.1)

intersect, that

and only

+ biy + d = 0,
is,

meet

in

a2x

+ b 2y + c 2 =

one point and only one point,

their determinant

if

not equal to zero; that is,

if (9.7) holds.
Thus two intersecting lines determine a point. This
is called the dual of the theorem that two points determine a
line.
Equation (5.3) gives the relation between the coordinates x, y of any point of the line determined by the points
(xi, ;yi) and (x 2 y 2 ) and the coordinates of these two points. As
the dual of this result there should be a relation between an
equation of any line through a point and equations of two lines
determining the point. It is this relation which we now obtain.
In connection with the equations (10.1) we consider the
if

is

expression
(10.2)

ti

(ai*

+ biy + ci) +

t2

(a 2 x

+ b 2y + c 2

),

Since the leftt\ and t 2 are constants, not both zero.

hand members of equations (10.1) are equal to zero when x
and y are given the values of their common solution, it follows

where

that the expression (10.2) is equal to zero for this common

solution whatever be t\ and t 2 Hence the common solution of
equations (10.1) is a solution also of the equation
.

(10.3)

h fax

+ biy + ci) +

t2

+ b2y + c2 = 0.

(a 2 x

when

it is

first degree in x and y, which

written in the form

(hbi

This being an equation of the

seen more clearly
(10.4)

(hai

t2

a 2 )x

t2

b 2 )y

(hci

is

+ fac 2 = 0,
)

not both zero, is an

we have that (10.3) for any
t\
equation of a line through the intersection of the lines (10.1)
9, Ex. 7).
(see
We show next that equation (10.3) for suitable values of
t\ and / 2 is an equation of any particular line through the intervalues of

section of the lines (10.1).

In the

47

and

first

f2,

place

we remark

that

Points

and Lines

in the

Plane

[Chap. 1

of the lines (10.1) is given by (10.3) when /2 =

and
from
different
and
number
the
second
of
the
lines
zero,
any
and t 2 is any number different from zero.
(10.1) when t\ =
the intersection is determined by a
line
other
through
Any

the

/i

first

is

second point of the line, say (*i, y\). Equation (10.3) is an

equation of this line if t\ and t 2 are such that the expression
(10.2) is equal to zero when x and y are replaced by x\ and y\
that is, if ti and t 2 are such that
9

hAi

(10.5)

A2

where A\ and

is

numbers defined by

of these

zero and solve (10.5) for

(*i,

2 *i

from

different

we give
we have

If

t 2j

+ b2yi + c 2

zero, since (*i, y\)

any value other than

values of t\ and t 2 for

t\

an equation of the line of the set through

From the form of (10.5) it follows that the ratio

(10.3)

>>i).

A2 =

numbers are

not on either line (10.1).

which

= 01*1 + biyi + Ci,

Ai

Both

are the

+ 2A 2 = 0,

is

a fixed constant, so that, if we choose another value of

find the corresponding t 2 from (10.5), the equation (10.3)

ti/t 2 is
ti

and

for these values differs

only by a constant
responding value of
the theorem
[10.1]

from the equation

for the former set

Hence any choice of h and the corgive an equation of the line, and we have

factor.
t2

When the lines (10.1) intersect in a point, that is, when

the determinant of equations (10.1) is not equal to zero, the
equation (10.3), namely,
/i

(0i*

+ biy + ci) +

/2

(02*

for any values of the constants h

+ b2y + c 2 = 0,
)

and

t2 ,

not both zero,

is

an

equation of a line through the point of intersection of the

lines (10.1) ; and (10.3) is an equation of any line through
this point for suitable values of ti and h.

Ordinarily the simplest choice of

t\

in

When

an equation such as

these values are sub2 , and then t 2

AI.
(10.5) is
stituted in (10.3), we have the desired equation.

48

The

Sec. 10]

Set of Lines through a Point

For example, if we seek an equation'of the line through the point (2,1)
intersection of the lines * + 2>>
1=0 and 3^-^ + 2 = 0,
equation (10.5) is in this case fi3 + fe7 = ; and consequently an
- y + 2) = 0, which
equation of the line is 7(x -f 2 y - 1)
3(3 x
reduces to - 2 x + 17 y - 13 = 0.

and the

In order to find an equation of the line through the intersection of the lines (10.1) which is parallel to the jy-axis, t\ and
/ 2 in (10.3) must be chosen so that the coefficient of
y in (10.4)
b2 t2
shall be equal to zero. If we take ti
and
substifei,

tute these values in (10.4), the resulting equation

written in the form
0i

(10.6)

02

b<2

may

be

-0,

is the same as (9.4).

This is not surprising, since we
obtained (9.4) by multiplying the first of (9.2) by b 2 and subtracting from the resulting equation the second of (9.2) mul-

which

resulting equation (10.4)

fa

the

Thus we have shown that the

alge-

may

0i

(10.7)

which

02
is

the same as

(9.5).

= b 2 and
= #2, fe =

0i

tiplied by 61, which is equivalent to taking

in (10.4). In like manner, if we take t\

t\

/2

be written

y+

01

Ci

02

= 0,

problem of finding the common solution of two equations

of the first degree in x and y, when there is one, has as its geobraic

metric equivalent the finding of lines parallel to the >>-axis and

#-axis respectively through the point of intersection of the
graphs of the two equations.
Any two lines of the set (10.3) could have been used equally
well in place of the lines (10.1) as a basis for expressing an
equation of every line through the point of intersection in the

form

(10.3),

with the understanding, of course, that for a

particular line the values of h and

lines are used as basis. This result
[10.2]

t2

depend on which two

be stated as follows

may

through a point, each line is expressible

linearly and homogeneously in terms of any two particular

Given

all the lines

lines of the set.

49

and Lines

Points

in the

Plane

[Chap, i

This is the dual of Theorem [5.4].

We have remarked that when ti = in (10.3) the latter is
equivalent to the second of equations (10.1). With this exception ti is not equal to zero for any of the lines (10.3), and
consequently,
replaced
following
[10.3]

by

t,

if

equation (10.3) is divided by t\ and /2 //i is

as a consequence of Theorem [10.1] the
9

we have

When

the lines (10.1) intersect, the equation

(01*

(10.8)

where

is

+ hy + ci) + t(a 2 x + b2y + c2 = 0,

)

any

the intersection;

constant,

is

and any

an equation of a

an equation of
an
given
appropriate value.

cept the second line (10.1), has

(10.8)

when

is

line through

line through the intersection, ex-

the

form

This theorem may be proved independently by the same

kind of reasoning which established Theorem [10.1]. In the
case of Theorem [10.3] we have in place of (10.5) the equation
and thus / is completely determined except
AI + tA 2 =
when A 2 = 0, in which case we are dealing with the second of
the lines (10.1). This makes the application of Theorem [10.3]
somewhat simpler than that of Theorem [10.1] in some cases,
and the reader may choose which theorem is better to apply in
a particular problem.
From a theorem in plane geometry we have that the locus
of a point equidistant from two intersecting lines is the bisectors
of the angles formed by the lines.
Consequently this locus
;

consists of

two of the

set of lines

through the point of interIn order to find equations of the

bisectors, we take two equations (10.1). At least one of the
coefficients bi and b 2 is not equal to zero, otherwise the lines
are parallel to one another, and to the ;y-axis.
We consider first the case when both bi and b2 are different
from zero. Let (*i, y{) be a point of the locus. Then its distances from the lines (10.1) are by Theorem [8.1]
section of the given lines.

(10 9)

eiai
50

The

Sec. 10]

Set of Lines through a Point

where e\ and e 2 are + 1 or 1 so that e\b\ > 0, e 2 b 2 > 0. If (x\, y\)

above the two lines, both the numbers (10.9) are positive
and equal if (xi, yi) lies below the two lines, both the numbers
are negative and equal; if (x\, y\) lies above one of the lines
and below the other, one of the numbers (10.9) is positive and
the other negative, but they have the same numerical value.
Since these results hold for every point on the bisectors, we have
lies

The

[10.4]

bisectors of the angles

(10.1) for which

a\x

bi^O
b\y

-f-

-f-

(10.10)

whgre

e\

and

e 2 are

formed by two

and

b2

Ci __

such that

,
i

intersecting lines

have the equations

a2 x

+by

e\b\

>

-f-

0, e 2 b 2

c2

>

0.

The equa-

tion with the sign

is that of the bisector each of whose
points is above, or below, both of the lines (10.1) ; and the
is that of the bisector each of whose
equation with the sign
is above one of the lines (10.1) and below the other.

points

This theorem
with the signs

is

illustrated in Fig. 12,

and

are

where the

indicated

lines (10.10)

by (+) and

respectively.

When

bi

or b 2

= 0,

must be chosen so that

>

e\

e\a\

or e%

>

In this case the

equation (10.10) with the sign
is that of the bisector whose
points lie above one line and
or e 2 a 2

0.

to the right of the other, or

below the one and to the left of

the other and equation (10.10)

is that of the
with the sign
bisector whose points lie above, or below, one line and to the
left, or right, of the other, as the reader should verify.
;

For example, the bisectors of the angles between the

lines

have the respective equations

(2

V5

4)*

+ V5 T 3)y (

51

(3

V5

1)

= 0.

Points

and Lines

in the

Plane

[Chap, l

EXERCISES
Find an equation of the line through the intersection of the
x-2y + 7 = and 3x+.y-l=0 and the point (1, - 1)

1.

lines

of the line through the intersection

2, 7)

of the

through the intersection and parallel to the *-axis.

line

2.

Find an equation of the line through the intersection of the

= and x+y + 2 = and parallel to the line
:v + 8

2*

lines

5x

and the point

- y = 0.
3.

Show by means

of

Theorem

4.

What must

# + 5;y-4 =
+ y + I =0 and

[10.1] that the line

passes through the intersection of the lines 2 x

be the value of a in the equation ax

so that this line shall pass through the intersection of x

and2x + 3^4-10 = 0?
5.

Show

that the lines ax

+ b = kc,

the condition a

through a
6.

common

Show

that

by

where k

+ c = 0,

is

whose

+y+6=
+y+4=

coefficients satisfy

a constant not equal to zero, pass

point.

when

the lines (10.1) are parallel, (10.8) is an equato the former for any value of /, except

tion of a line parallel

/

01/02

b\ /b%.

Why does

this exception

not arise when the lines

(10.1) intersect? Is every line parallel to the lines (10.1) so defined?

7.

Find equations of the bisectors of the angles between the

lines

8.

Find equations of the bisectors of the angles between the line

5y + l = and the line perpendicular to it through the origin.

9.

Find an equation of the line through the point of intersection

such that the point
5x + 2y + 1 = and x- y
i s at the directed distance + 2 from the line.
4)

4 *

-5=0

of the lines
(5^

10. Find the point of intersection of the bisectors of the angles of

the triangle whose sides have the equations

3x + 4y +
that

is,

= 0, 4* + 3jy-21 = 0,

12

x-5y + 38 = 0,

the center of the inscribed circle; find the radius of this

circle.

11.

Draw

the graph of the lines in Ex. 10, extending the sides of

the triangle, thus dividing the plane into seven compartments. Show
that in three of these compartments outside the triangle there is a
point which is the center of a circle tangent to the three lines find
;

the center

and radius

of each of these escribed circles.

52

Oblique Axes

Sec. 11]

11.

At the beginning

of

Oblique Axes

we took the

coordinate axes perpen-

dicular to one another, in which case they are said to be rectangular, and defined the coordinates of a point by drawing
did not dethrough the point lines parallel to the axes.

We

fine the coordinates as the perpendicular distances of the point

and #-axis respectively, although we remarked

what they are when the axes are perpendicular to

from the

jy-axis

that this

is

one another. Consequently the definition of coordinates given

2 applies equally well when the axes are not perpendicular
in
to one another, in which case we say the axes are oblique.
The question of whether the axes are rectangular or oblique
does not arise in the case of lengths parallel to the axes, and
thus does not affect direction numbers, or equations of a line
not involving direction cosines satisfying equation (3.6)
in
for
from
the
derived
was
fact, equation (3.6)
expression (3.1)
the distance between two points (#1, y\) and (* 2 ^2), which
;

presupposes that the axes are rectangular. The same assumption was involved in deriving formulas for the angle between
lines and the distance from a line to a point. The reader will
find it instructive, and conducive to getting a clear picture of
the subject thus far, to take each theorem and determine

whether

The

proof depends upon the axes' being rectangular.

subject of oblique axes is introduced here not only to
its

enable one to understand how much of what precedes

pendent of the angle formed by the
coordinate axes, but also because
the use of such axes in certain problems leads to simpler algebraic treat-

is

inde-

/f<?(o,c)

ment of these problems. For example,

if
one wishes to prove that the
medians of a triangle meet in a point,

advisable to choose two sides of

the triangle as axes, and denote the

it is

"
O/A(O,O)

jpd.o)

by (0, 0), (ft, 0), and (0, c),

shown in the accompanying figure. Equation

2?(&,o)

vertices

as

because

it

(6.1) applies,

expresses the equality of ratios of direction numbers.

53

Points

The equations

If

we

subtract the

first

get an equation of a

by Theorem

[10.1]

Plane

[Chap. 1

2 ex

+ by - be = 0.

of these equations from the second, we

through the intersection of these lines

line

this equation is

cx-by =

(11.2)

But

in the

BE and CF are

of the lines

+ 2 by - be = 0,

ex

(11.1)

and Lines

Q.

an equation of the median AD, as the reader can

this is

and thus it is shown that the medians meet

readily verify
in a point. Another method of proof consists in solving (11.1)
;

common

solution; this is b/3, c/3, which evidently

Moreover, one shows readily that the point
c/3) divides each median in the ratio 1/2.

for their

satisfies (11.2).

(6/3,

12.

The

Circle

By definition a circle is the locus of points each of which is at

the same distance, the radius, from a fixed point, the center.
In order to find an equation of a circle, we denote by r its
radius, by PO(#O, ;Vo) its center, and by x, y the coordinates of a
representative point, that

of

Theorem

[3.1]

(x

(12.1)

which
(12.2)

may
x2

any point on the circle. By

as an equation of the circle

is,

we obtain

2
)

(y

- jo) 2 =

means

r2,

be written

+ y2 - 2 x x - 2 y y +

(x

+ yo 2 - r = 0.
2

Evidently any solution x, y of this equation (and there is an

endless number of them) gives the coordinates of a point on the
circle with center at P and radius r. On giving XQ, y
r suit,

able values, equation (12.1) is an equation of any

We remark that (12.2) is of the form

x2

(12.3)

Conversely,

form

we

circle.

+ y* + 2 fx + 2 gy + k = 0.
shall

determine whether an equation of this

and k is an equation of a circle. If

for given values of/, g,

we add/2
expression

to the left-hand

may

be written

(x

member of (12.3), the resulting

2
2
+/) + (y + g) + k. But then

54

The

Sec. 12]

we must add f2

Circle

+ g2

to the right-hand side of (12.3), so as to

continue to have an equivalent equation. Consequently we can
replace (12.3) by its equivalent
(12.4)

(x

+f) 2 +(y

+ g) 2 =/2 + g 2 - *.

this equation with (12.1), we see that (12.4), and

consequently (12.3), is an equation of a circle whose center is
2
2
k.
From the form
(
/>
g) and whose radius is V/ + g
(12.4) of equation (12.3) it follows that the latter equation does
2
not have any real solutions when f2
k < 0. But because
g
(12.3) is of the same form as in the case when there are real
2
2
k is negative (12.3) is
solutions, we say that when f + g

Comparing

of an imaginary circle
we cannot plot such a
2
2
k
0, it follows from
only talk about it. When/ + g
(12.4) that x =
f, y
g is the only real solution of (12.4),
and consequently of (12.3) sometimes this point is called a
point circle, that is, a circle of zero radius. Thus only when
2
2
k >
is (12.3) an equation of what the reader would
/ +g

an equation

circle,

call a genuine circle. But if one wishes to make a general statement about the geometric significance of equation (12.3), one may

say that
g,

and

it is

an equation of a

circle

whatever be the values of/,

admitting the possibility of imaginary and point circles.

important that the reader get clearly in mind the form of

k,

It is

equation (12.3). Note that there is no term in xy, and that the
2
2
1.
coefficients of x and y are equal, in this case both equal to
If they were both equal to some other number (for example 3),
we could divide through by that number (which does not affect
the solutions of the equation) before putting the equation in

the form (12.4) by the process used above, and called completing
the square of the terms involving x, and of the terms involving^.

For example,

if

we have
2x*

we

divide

by 2 and complete the

*2

When
it is

this

the equation

+ 2y 2 -5x + 4y-7 = 0,
squares, getting

-f*+(f) + r + 23'+l =
2

equation

seen that (f

is

1)

(f)

+l + f = f

written in the form

is

r. /

the center of the circle and

/

55

?V97

its

radius.

Points
Geometrically

[Chap. 1

say that a line meets a circle in two points

tangent to it. Let us see what this means
take the equation of the line in the form

We

which case

Plane

is

(12.5)
in

in the

we

or in no points or
algebraically.

and Lines

= mx + h,

m is the slope and h the ^-intercept.

Instead of tak-

ing a circle placed in general position with respect to the axes,

we take a circle of radius r with center at the origin (for the

sake of simplifying the calculations involved)

x2

(12.6)

+ y2 = r2

its

equation

is

The reader is familiar with the process of finding common

solutions of equations such as (12.5) and (12.6), namely, substituting the expression (12.5) for y in (12.6) and solving the
Making this substitution and collecting

resulting equation.

terms,

we obtain

(12.7)

(1

+m

)*

+ 2mhx+

(h

- r2 = 0.
)

The geometric interpretation of this quadratic in x is that its

two roots are the ^-coordinates of the points in which the line
meets the

circle.
Applying the formula for the solution of a
quadratic equation to (12.7) and reducing the result, we have

+ m2 >

2
If r (l

A2

(12.8) gives two real and distinct values

These and the corresponding values of y obtained from
(12.5) on substituting these values of x are the coordinates of
the two points of intersection.
)

of x.

+ m 2 ) < A 2 the two solutions (12.8) are imaginary,

the
as are
corresponding /s. In this case the line does not
meet the circle, or, as is sometimes said, it meets the circle in
If r 2 (l

two imaginary points; we cannot

we may
h2

plot such points, although

talk about them.

We consider finally the remaining possibility, namely,

= r2 (l + m2 ). When h = rVl + w2 the two solutions of (12.7)
,

56

TKe

Sec. 12]

Circle

are equal to one another, as follows from (12.8)

that is,
the two points coincide. In this case we say that the line is
tangent to the circle at the (doubly counted) point PI, whose
;

coordinates are

mh

as follows from (12.8) and (12.5).

The line joining the center of the circle (that is, the origin)
to the point PI has xi, yi for direction numbers, as follows

from Theorem

[6.1]

(see

follows that m,

it

(12.9)

Ex. 7); and consequently from

6,

are direction

numbers of

this

In consequence of
being proportional to x\ and yi.
Theorem [6.9], m,
1 are direction numbers of a line perpendicular to the line (12.5). Hence the tangent, as defined
line,

above, is perpendicular to the radius of the circle at PI this

agrees with the definition of a tangent to a circle, with which
the reader is familiar. However, the latter definition is limited
;

to the
its

circle,

whereas the definition we have used

application.
Since similar results are obtained

value

Vl + m 2 we

general in

in (12.5) has the

have

For each value of

[12.1]

when h

is

(12.10)

m the two equations

l + m2
7
y = mx

are equations of two parallel tangents, of slope m, to the

circle

This

is

x2

+ y 2 = r2

illustrated in the ac-

companying

figure.

Equations

(12.10) are equations of tangents

to a circle only in case the

equation of the circle has the

form "(12.6). However, the same
process may be applied to an
equation of the circle in the
general form (12.3) (see Ex. 5

and

13,

.13

Ex. 23).

57

Points

and Lines

in

tlie

Plane

[Chap.

a point PI(XI, y\) is on the circle (12.3), the expression

Xl 4. yi
2fxi + 2 gyi + k is equal to zero but when PI is
not on the circle, this expression is a number different from zero.
We shall give a geometric interpretation of this number when
Pi lies outside the circle, as in Fig. 14. Denoting the center by

When

C,

we have from

(12.4)

Hence, since PiT

= PiT',

we have

JF/^w a point P\(x\, y\) lies outside the circle (12.3), the
2
x\
y^ 2fxi 2 gyi k is equal to the square
distance
the
from P\ to the point of contact of each of the
of
two tangents to the circle from PI.

[12.2]

number

Since there are three independent coefficients, /, g, k, in equaindependent in the sense that each choice of these

tion (12.3)

an equation of a particular circle, and a change

of the coefficients chosen gives a different circle

coefficients gives

of one or
it

more

follows that a circle

conditions.

center

The

is

completely determined by three

definition of a circle involves the location of its

and the length of its radius when these are given, the
g, and k are completely determined, as shown in equa;

values of/,

tion (12.2). Also a circle is determined by three noncollinear

if the coordinates of these points are given
points (see Ex. 8)
and these are substituted in (12.3), we obtain three equations
to be solved for the appropriate coefficients/, g, and k in (12.3).
;

Consider

and

now

the two circles

in connection with

(12.12)

h(x

them the equation

+ y* + 2 fa +

The

Sec. 12]

Circle

where t\ and t 2 are constants not both zero. This equation

be written also in the form
(h

(12.13)

t 2 )(x

may

+ y 2 + 2(/i/i +
)

t2

g 2 )y

(ti^

/2

*2 )

= 0.

If xi, yi is a common solution of equations (12.11), it is a solution of (12.12) whatever be the values of ti and / 2 since the expression in each parenthesis in (12.12) becomes equal to zero
,

y are replaced by x\,y\. When t\ + fa ^ and equation

divided by (h + / 2 ), the resulting equation is of the
form (12.3). Consequently equation (12.12) for ti + fa?*Q is
an equation of a circle passing through the points of intersection
of the circles (12.11) if the circles intersect, that is, if the com-

when

x,

(12.13)

mon

is

solutions of (12.11) are real.

In order to find the common solutions, if any, of equations

(12.11), we subtract the second of these equations from the first

and obtain
(12.14)

2(/t

This does not

-/2 )* + 2(ft -

mean

g 2 )y

(*i

*2 )

= 0.
a comhave a
observe that

that any solution of this equation

mon solution of equations (12.11), but that, if the

common solution, it is a solution of (12.14). We

is

latter

t\ and divide
(12.14) follows from (12.13) when we take t 2
out the factor ti
consequently (12.14) is the reduced form of
;

In view of this fact, if we have a comt\.

(12.12) when t 2
mon solution of (12.14) and either of equations (12.11), it is a
solution of the other. In fact, if xi, ;yi satisfy the first of (12.11)
and (12.12) with t 2
h, then the quantity in the first paren-

when x, y are replaced by x\, y\ throughout

the equation, and consequently the expression in the second
parenthesis also. Hence the problem of finding the points of
intersection of the circles (12.11) reduces to that of finding the
thesis is equal to zero

points of intersection of either circle and the line (12.14). This

line is called the radical axis of the two circles. From considerations similar to those applied to equations (12.5) and (12.6) it
follows that this line and the circles have in common two points

no real points or one point (counted doubly), in which case

the two circles have the line for common tangent at the point.

or

59

Points

When

and Lines

Plane

in the

[Chap, l

the two points of intersection of the circles (12.11) coinwhich case the circles have a common tangent at this

cide, in

(doubly counted) common point, each of the circles (12.12)

has the same tangent at this point.
Whether the circles have a point in common or not, the
From Theorem [6.9] and the form of
radical axis exists.
follows that f\

it

(12.14)

numbers

82 are direction

/2, g\

of a line perpendicular to the radical axis. But by Theorem

[6.1] these are also direction numbers of the line joining the

points

circles.

[12.3]

/i
gi) and
Hence we have
f

Given two

/2

2),

that

is,

the centers of the

circles (12.11), equation (12.12)

for values of

ti

an equation of a circle through

/2
t\ + /2
the points, if any, in which the circles meet; when t\ + ^2 = 0,
it is an equation of a line perpendicular to the line through

and

such that

is

the centers of the two circles.

EXERCISES
1. Find an equation of the
whose radius is 2.

2.

circle

whose center

is

2,

Find an equation of the circle of radius 3 which

and lies below the #-axis.

is

and

3)

tangent to

the #-axis at the origin

3.

Find the center and radius of the

4.
if

Show

circle

whose equation

is

6x-5y = Q.

2x 2 + 2y 2 +

that a circle (12.3) is tangent to the x-axis, if and only

is the condition that it be tangent to the jy-axis ?

= / 2 What

5.

Find equations of the tangents to the

the slope 3
the slope

also the tangents to the circle x 2

circle

+ y2 +

x2

-f

4 x

=4
=

with
with

2.

6. Find equations of the lines with the slope 3 which are at the
distance 4 from the origin.
7. Find equations of the lines through the point
at the distance 2 from the origin.
8.

Show

(2,

3)

which are

that the radius and center of a circle are determined

by

three points on the circle.

9.

Prove that an angle inscribed in a semicircle

= r 2 as an equation of the circle.

2
2
taking x -f y

60

is

a right angle,

The

Sec. 12]

10.

Where

11.

What

are the points for which x 2

12.

a circle ?

circle

through the three points

Find the length of the tangent from the point

Show

(1, 1),

and

/2,

(2, 1)

to the circle

5x-2y + 2 = 0.

that an equation of any circle through the points of intwo circles (12.11) is given by (12.12) for suitable values

tersection of
/i

+ 2 xy + y 2 - 4 =

Find an equation of the

13.

of

x2

2), (2, 3).

3x 2 + 3y 2 +
14.

+ y2

the locus with the equation

is

Is the locus of

(1,

Circle

and that the centers

of

these circles

all

lie

on a

line.

Show that an equation of any circle through the points of

intersection of two circles (12.11), with the exception of the second
15.

of the circles (12.11),

x2

+y +
2

2 fix

is

+ 2giy + h +

for a suitable value of

16.

Show

given by the equation

l(x

+ y 2 + 2f2 x + 2

g2y

k2 )

=Q

/.

that equation (12.12)

is

an equation of a

circle

even

if

the circles (12.11) do not intersect, except when t\ -f t2

0, and that
in the latter case it is an equation of a line perpendicular to the line

through the centers of the two

17.

x2

Of the

+ y2 -

circles.

through the points of intersection of the circles

6x + 8y
and x 2 + y 2
+4
0, find

circles

4 x 4- 2

(a) the one which passes through the point (1,2);

(b) the one whose center is on the #-axis ;
(c)

an equation of the

radical axis.

Show

that the circles (12.11) are orthogonal to one another,

that their tangents at each point of intersection are perpendicular to one another, if and only if 2(/i/2 + gig*)
k\ + k 2
18.

that

is,

Find an equation of the circle which passes through the point

and is orthogonal to the two circles x 2 + y 2 + 2x-4y-3 = Q
and x 2 + y 2 -6x + 2y + 6 = Q.
19.

(0, 2)

20. Show that, if (xi, y\) is a point within the circle (12.3), the ex2
2
2 gyi 4- k is a negative number, and that its
pression xi
yi 4- 2fxi
absolute value is equal to the square of one half of the chord through

the point
21.
is

(x\,

y\) perpendicular to the radius

Determine whether each of the points

outside, on, or inside the circle

61

through
(1,

this point.

),

(2,

1), (1, 1)

Points
13:

and Lines

in the

Plane

[Chap. 1

Resume. Line Coordinates

As

stated in the Introduction, coordinate geometry involves

the application of algebra to the study of geometric problems.
The first step consists in setting up the algebraic equations (or

equation) which express the conditions of a geometric problem, and solving these equations or reducing them by algebraic
processes to their simplest form. The second step is the geometric interpretation of the result. Let us interpret some of the
results of this chapter in the light of this statement.

A line in the plane is determined either by two of its points or

by a point and the direction of the line. We have established
J

the following equations of a line

m(x

The first
(#2, JV2)

xi)=y

yi.

of these equations involves the coordinates (xi, yi) and

two points on the line, the second and third a point

of

and a direction, the latter being expressed either by direction

numbers u and v, that is, numbers proportional to the cosines
of the angles which the positive direction of the line makes
with the positive directions of the axes, or by the slope, that is,
the tangent of the angle which the positive direction of the
line makes with the positive direction of the #-axis. Since

x2

xi

and y 2

equation

is

y\ are direction numbers of the line, the first

in fact a special case of the second equation. Con-

sequently, when the above data are given for a line, an equation of the line may be written, an equation of a line being an
equation each of whose solutions consists of the coordinates of

a point on the

line,

and conversely.

Each of the above equations is of the first degree in x and y,

and it has been shown that every equation of the first degree is
an equation of a line. Any such equation is of the form
62

Reesume

Sec. 13]

Suppose then that we have such an equation and wish to

inter-

geometrically. If we find any solution of it, say (x\, y\),

pret
we can put the equation in either of the last two forms above,
it

provided

we adhere
if

in

term of a ratio

is

to the

an equal-

principle that,
ity of two or more ratios either

zero,

so

also

is

equal to
the other

term
a principle which has
been used several times in
Also by means

this chapter.

of a second solution the equation may be given the first

FIG. 15

of the above forms.

Since any multiple of the above equation has the same solutions as the given equation, any such equation is an equation
of the line ; and consequently we should expect that not only

but also quantities proportional to them have

geometric significance. To overcome this ambiguity we write
the above equation in one of the following forms, of which
the first is possible only when a, &, and c are all different from
zero, and the second when 6^0:
a,

ft,

and

JL + -2- =

(13.1)

in other notation, 2

+ \ = 1.

a
notation,

(13.2)

(13.3)

where

when

e^/a 2
e is

+1

or

1 so that eb

= 0.

>

= mx + h.

+ b2

= 0,

when 6^0, and ea>0

In (13.1) g and h are the x- and ^-intercepts respectively

in

m is the slope of the line and h its ^-intercept in (13.3)

2
2
2
2
a/e^/a + b and b/e^/a + b are the direction cosines of any
(13.2)

of the endless

and c/e-^a 2

number

of lines perpendicular to the given line,

+ b2 is the directed distance of the origin

line (see Fig. 15).

63

from the

Points

and Lines

in tKe

Plane

(Chap.

Since the coefficients a, b, and c of an equation ax + by + c =

determine the equation, and consequently the line of which it is

an equation, we call a, b, and c line coordinates of the line. Since

an equation of a line is determined only to within a constant
multiplier, it follows that if 0, ft, and c are line coordinates of a
line, so also are ka, kb, and kc line coordinates of the same line
for every value of k different from zero. The line coordinates
~ 1 in (13.1), m, - 1, h in (13.2), and a/e V0 2 + b 2
I/ft !/*.
2
2
2
2
b/e^/a + b and c/e^/a + b in (13.3), have in each case the
geometric significance mentioned in the preceding paragraph.
,

Now

we may state Theorem [10.1] as follows: Line coordinates of any line through the intersection of two lines with coordinates 0i, fti, Ci and 02, 2, 2 are of the form t\a\
/202, t\b\
/2&2,
/i^i

/2C2-

Thus

line coordinates of

each of the set of

all

the

through a point are expressible linearly and homogeneously in terms of line coordinates of any two lines of the set.
This is the dual of Theorem [5.4].
If the reader at some time takes up the study of projective
lines

geometry, he will find that in this subject line coordinates are

as fundamental as point coordinates, and that corresponding
to any theorem concerning points and lines there is a dual

theorem concerning lines and points.

The two steps involved in coordinate geometry mentioned
in the first paragraph of this section are involved in the solution
of any locus problem, that is, in finding the answer to the
question What is the locus of a point satisfying certain geometric conditions? First of all the reader is expected to set
up the equation (or equations) satisfied by the coordinates of
:

any and every point meeting the conditions of the problem.

Ordinarily these conditions involve the distance between
points, or the distance of a point from a line, or the angle between two lines, formulas for which have been derived in the
text. The next step, after the equations have been solved or
reduced to simple forms, is to interpret the result geometrically, stating whether the locus is a line, a circle, or some other
curve, as the reader can do when he knows the forms of equations of other curves, knowledge which he will acquire in later
parts of this book.
64

Line Coordinates

Sec. 13]

As an example of the foregoing we refer to equations (10.10),

obtained as answer to the question What is the locus of a point
equidistant from the intersecting lines (10.1)? We observed
that it consists of two lines, the bisectors of the angles between
:

the two given lines. Suppose again that we were asked to find
the locus of a point satisfying certain geometric conditions and
eventually obtained an equation of the form

x2

We

should

specify

its

+ y 2 + 2fx + 2 gy + k = 0.

know from

12 that the locus

is

circle

and how to

center and radius.

The resume given in the first part of this section is in no

sense a complete summary of this chapter. A good way for the
reader to get a picture of the chapter as a whole is by reading
successively the theorems given
to succeeding chapters.

the same suggestion applies

In this chapter use has been made of theorems of plane

geometry assumed to be known by the reader, and the results
of their use have been translated into algebraic form. There has
been no attempt to develop coordinate geometry systematically
from a set of axioms. However, in the Appendix to Chapter I
(p. 279) the reader will find an exposition of the relation between
a set of axioms for Euclidean plane geometry and coordinate

geometry as developed in the present chapter.

EXERCISES
1.

Find the locus of a point such that 3 times

its distance from the *-axis

the jy-axis plus 4 times

is

its
is

distance from

equal to 12.

2. Find the locus of a point whose distance from the

a constant times its distance from the ;y-axis.

line

3. Find the locus of a point the ratio of whose distances from lines
through the point (*i, y\) parallel to the .y-axis and *-axis respectively
is equal to u/v, where u and v are constants.

Given the isosceles right-angled triangle whose vertices are

(0, 0), (a, 0), and (0, 0), find the locus of a point the square of whose
distance from the hypotenuse is equal to the product of its distances
from the two legs of the triangle.
4.

65

Points and Lines in tne Plane

5.

[Chap. 1

Find the locus of the mid-point of a line of constant length

its extremities on the positive parts of the axes.

having

Find the locus of a point which is twice as

2*-3;y-fl = Oas from the line x-2y + 4 = Q.
6.

far

7. Find the locus of a point such that, if Q and R are

on the coordinate axes, the distance QR is constant.

from the

its

line

projections

Find the locus of a point such that the mid-point of the segment
2
2
2
joining it to the point (0, b) lies on the circle x + j =
8.

9. Show that the locus of a point whose distances from two fixed
points PI(XI, y\) and P2(*2, y%) are in constant ratio different from
unity is a circle. When this ratio is equal to unity, what is the locus ?

10.

sides

Find the locus of a point the sum of whose distances from the
the triangle ^ = 0, 3jy-4*=:0, 12^ + 5^-60 =
is

of

constant.
11.

What

is

any number n
12.

What

is

the locus of a point the

of lines

is

sum

of

whose distances from

constant ?

the locus of a point the sum of the squares of whose

number n of points is constant ?

distances from any

13.

3* +

Find two points on the

4

>

;y-axis four units distant

from the

line

y-12 = 0.

14. The base of an equilateral

and the opposite vertex is (4,

triangle lies on the line 3x

1).

2j + 5 =

Find equations of the other two

sides of the triangle.

15.

Show

that the four lines

ax

ax

by
by

+c=
+c=

0,

bx

0,

bx

form a square with center at the

origin.

ay
ay

What

+ c = Q,
-f c =
is

the situation

when

the axes are oblique?

16.

Find the distance of the center of the

circle

from the line 4x-3;y + 6 = 0, and determine thereby whether this

above or below the circle, intersects the circle, or is tangent

line lies

to

it.

17. Show that if a point P lies above or below each of two intersecting lines, the angle at P between the positive directions of the
perpendiculars upon the lines is the supplement of the angle in which

66

Resume

Sec. 13]

P lies that if P lies above one of the lines and below the other, the
angle between the positive directions of the perpendiculars is equal
to the angle in which P lies.
;

Find what relation must hold

18.

cients of the equation ax 4- by

(a)

(b)

in

each case between the

coeffi-

of a line in order that the line

have the intercept 2 on the *-axis

have equal intercepts

be perpendicular to the line 2#-3j>-f-l = 0;

through the origin
(e) pass through the point (5,
4)
3 from the origin
(/) be at the directed distance
(g) be at the directed distance 4 from the point (2, 3)
(h) pass through the intersection of the lines 3x + ;y
(c)

(d) pass

and

19. Find equations of the lines through the point

with the axes a triangle of area 16.

20.

Given the

x-2y + l=Q.

triangle

whose vertices are 0(0,

(2,

0),

3)

which form

A(a, 0),

(0, b),

prove that the line joining the mid-points of any two sides is parallel
to the third side and equal to one half of it
that, if on the line joining
to the mid-point C of AB we take any point P and denote by D
and E the points in which AP meets OB and BP meets OA respectively,
then DE is parallel to AB. Which, if any, of the foregoing results hold
;

when
21.

the axes are oblique ?

Given the

triangle

whose

vertices are

A (a,

0),

B(0,

ft),

and

prove that the perpendiculars from the vertices of the triangle

upon the opposite sides meet in a point; that the perpendicular
bisectors of the sides meet in a point.
C(c, 0),

Given the rectangle whose vertices are 0(0, 0), A (a, 0), C(a, b),
if E and F are mid- points of OB and AC, prove that the lines
and BF trisect the diagonal OC. Does this result hold for the

22.

B(0, b),

AE

figure

OACB when

23.

the axes are oblique?

Find the condition upon the

of a circle

and

coefficients in the equations (12.3)

be tangent to the

(12.5) of a line so that the line shall

circle.

24.

not

all

25.

Prove that the radical axes of three

on the same line meet in a point.

circles

whose centers are

Find an equation of the circle circumscribing the

= 0, y = 2x, 2 * + .y - 8 = 0.

sides are x

67

triangle

whose

Points and Lines in the Plane

26.

Find an expression for the area of a triangle whose vertices are

and (*3, jVs).

y\)> (*2, ^2),

(*i

27.

Prove that the

vertices are (28.

(f

[Chap. 1

line

29.

5 x

3),

(-

V3 y = bisects the
- V5), and (1, V3).

4-

1,

Find an equation of the

3V3,

line

), is

tangent to the line y

circle

triangle

whose

which passes through the point

= V3 x + 2, and whose center is on the

= 0.

Find the coordinates of the point equidistant from the lines

12 y
and 4 x
4 = 0, and at the distance 2 from
5 =
3j

the origin.
30.

Show
(2

that the curve with the equation

-y

4- 2) (3

4-

1) 4- (2
4- (3

*
#

- j 4- 2)(* 4- j>)
4- 4jy

l)(jt 4- .y)

passes through the vertices of the triangle whose sides have the

equations

31.

What

is

the significance of equations (10.10)

(10.1) are parallel?

68

when

the lines

CHAPTER

Lines and Planes in Space.

Determinants

a
14.

Rectangular Coordinates in Space

The reader will agree that the position of any point in a

room is fixed by its distances from the floor and from two adjacent walls. If the planes of the floor and the two walls are
thought of as extended indefinitely, the position of any point
is fixed by its distances from these three planes, provided that one has a means of indicating whether a point is
on one side or the other of a plane. A similar question arose
in space

in the case of the points in the plane and was met in

2 by
of
the
and
distances
from
two
using
concept
positive
negative

perpendicular lines, the coordinate axes. We generalize this concept to points in space, and take as the basis of a coordinate
system three planes meeting in a
point 0, every two planes being perpendicular to one another, just as in

>

f/\
*

'

the case of the floor and two adjacent

walls of a room. Fig. 16 is the cus-

tomary way of drawing three mutually

perpendicular planes, the line Oy

7
,

pointing in the general direction of

the observer. In Fig. 16 the lines of
intersection of the planes, namely,
Ox,,Vy, and Oz, are called the x-axis,

FIG. 16

and z-axis respectively. The plane of the two lines Ox

and Oy is called the xy-plane similarly we have the yz-plane
and the xz-plane. These are called the coordinate planes.
The figure corresponds to the case in which the eye of the

y-axis,

reader

is

in front of the %2-plane, to the right of the ;yz-plane,

xjy-plane. Distances measured upward from the

taken
to the
as positive, downward as negative
ry-plane are
as
in
negative
right of the jyz-plane as positive, to the left

and above the

front of the #2-plane as positive, and back of it as negative.

The reader will get a picture of this situation by looking at a

the floor lines through

lower corner of the room, and calling it
to the right and left correspond to the #-axis and j-axis respectively, and the line of intersection of the two walls through
;

to the z-axis.

The

three planes divide space into eight parts,

rooms meet in one point when the

called octants, just as eight

71

Lines and Planes in Space

room in which the reader

the two adjacent walls at

is

sitting is

on the second

[Chap. 2

floor

and

are inside walls.

Coordinate planes and coordinate axes having been chosen

as just described, the coordinates of a given point P are defined
as follows Draw through P three planes
parallel to the coordinate planes, and
:

PX9 P y and Pz the points in

which these planes meet the x-axis, j/-axis,
and z-axis respectively. A unit of length
having been chosen, the lengths of the
directed segments OPXJ OP y and OP Z are
by definition the x-, y-, and z-coordinates
denote by

Each of these coordinates is positive or negative according as the directed

of P.

segment on the corresponding coordinate axis is positive

or negative, as defined in the preceding paragraph.
Conversely, given any three numbers x, y, z, to find the point
on the coordiof which they are coordinates we lay off from
line

nate axes Ox, Oy, and Oz distances of x, y, and z units respectively, in the positive or negative direction in each case according as the respective numbers are positive or negative, and
denote the end points by Px P yy and P z respectively. Through
these points we draw planes parallel to the yz-, xz-, and ^-planes
respectively the point P of intersection of these planes is at the
,

x, y, z from the coordinate planes, since parallel

are
everywhere equally distant. Consequently P is the
planes
point with the coordinates x, y, z and is indicated by P(x, y, z).

directed distances

An

equivalent way of locating P(x, y, z) is as follows Lay

a length of x units on the #-axis in the appropriate direction
from
at the end point Px of this segment draw a line in the
:

off

and on this line lay off

xy-plane perpendicular to the #-axis
from P x in the appropriate direction a length of y units. At
the end point P xy of this segment construct a line parallel to
the z-axis, that is, perpendicular to the jcy-plane, and on this
;

line lay off

z units.

from

The end

Pxy

in the appropriate direction a length of

of this

is the point P(x, y, z).

as defined, being mutually perpendic2-coordinates of a point are the perpendicular

segment

The coordinate planes,

ular, the x-, y-,

and

72

Rectangular Coordinates in Space

Sec. 14]

distances of the point from the yz-, xz-

There

may

be times when

and ry-planes

respec-

advantageous to take
as coordinate planes three intersecting planes not mutually
perpendicular; in such cases the coordinates as defined in
tively.

it is

the preceding paragraph are not the perpendicular distances

of the point from the planes (see
11). Throughout this book
are
as coordinate planes.
used
mutually perpendicular planes
The point in which the perpendicular from a point P upon

a plane meets the plane is called the orthogonal projection of P

upon the plane thus, Pxy in Fig. 17 is the orthogonal projection
of P upon the ry-plane. The point in which the perpendicular
from a point P upon a line meets the line is called the orthogonal
;

P upon

the line
thus, Px is the orthogonal prothe
#-axis
of
the
jection upon
points P, Pxy and PIZ The proa
line
a
jection of
segment upon
line, or plane, is the line segment

projection of

whose end points are the orthogonal projections of the end

points of the given segment upon the line, or plane.
Two points Pi and P 2 are said to be symmetric with respect
to a point when the latter bisects the segment PiP2
symmetric
;

a line

when

perpendicular to the
with
respect to a plane
segment PiP2 and bisects it symmetric
when the latter is perpendicular to the segment PiP2 and bisects
with respect

to

the latter

is

it.

The reader can picture each of these

situations geometrically,

not in position yet to handle any one of these types of

symmetry algebraically for general positions of the points Pi
and P2. However, he is in position to discuss symmetry with
respect to the origin, the coordinate axes, and the coordinate

but he

is

planes (see Ex. 3).

It is evident that it is not possible to represent a spatial
figure accurately on a plane. However, it is convenient to use

what

In applying this method,

is called parallel projection.
figures in the xz-plane and in planes parallel to it are represented
for example, the point (*i, 0, 21) in the #2-plane is
as they are
;

placed at the distances x\ and z\ from the 2-axis and #-axis

respectively. On the other hand, the jy-axis and lines parallel
to it are drawn to make the angle 135 with the #-axis and
2-axis as in Figs. 16

foreshortened

by the

and

17,

factor

andjengths in

1/V2;
73

that

is,

this direction are

for

a length

in

Lines and Planes in Space

this direction the length

//V2

is

laid off

on the

[Chap. 2

jy-axis or

a line

parallel to it, as the case may be. Thus, if squared coordinate

paper is used, distances along or parallel to the #-axis or
z-axis are laid off to their full amount with the side of a square

as unit, whereas distances along or parallel to the >>-axis are

a diagonal of a small square. On the paper the

laid off along

length of a diagonal is V2, but since 1/V2 is the unit of

length along a diagonal in accordance with the described method
of representation, in such representation the length of the

diagonal

is

for,

V2

~
=
V2

2.

The angle 135 and

the re-

/
suiting factor 1/V2 are chosen because they are simple to
handle, and because they have been found to give a clear

conception of a spatial figure.

This chapter deals with configurations of points, lines, and
planes in space defined geometrically, these definitions being
then translated into algebraic form involving the coordinates
of a representative point of the locus under consideration. In
the definition of coordinates we have used the concepts of
parallel and perpendicular planes, and of a line as the intersection of two planes, and have asked the reader to visualize
the definition by considering the walls and floor of a room as
planes, and their intersections as lines.
fined a line as a locus determined by

In Chapter

any two

of

1
its

we

de-

points

and having the same direction throughout. Also we explained

what was meant by saying that a plane is a two-dimensional
locus, but we did not define a plane. When we look upon a
plane as lying in space, we use Euclid's definition that a plane
is a two-dimensional locus such that every point of a line
which has two points in the locus is in the locus. A sphere i
>

two-dimensional in the sense that any point on

it can be fixed
example, by latitude and longitude but
a sphere does not possess the above property, nor does any
surface other than a plane.
In the consideration of points, lines, and planes in general
position with respect to the coordinate axes, we are concerned
with such metric questions as the distance between two points,

by two numbers,

for

the angle between two lines or two planes, and the distance

74

Rectangular Coordinates in Space

Sec. 14]

of a point from a plane. These quantities having been expressed in algebraic form, the reader is enabled to convert a
geometric problem into an algebraic one, and by algebraic
processes, frequently by the use of determinants, to arrive at
the solution of a problem more readily than had he employed

purely geometric reasoning.

KXERCISES
1.

What

andP Z2
2.

are the coordinates of the points 0,

inFig. 17?

Show

P x P v P P xy P yt
,

t,

that in Fig. 17 the line segment OP X is the projection of

OP on the *-axis, and OP XZ the projection of OP on the

the segment
#2-plane.

3. What are the coordinates of the seven points whose coordinates

have the same numerical values as those of the point (1, 2, 3) but
with one or more of the coordinates negative ? Which pairs of these

eight points are symmetric with respect to the origin

to the y-axis ; with respect to the xz-plane?
4.

Where

which x
5.

=3

are the points for which x

2?

and y

Where

+ y* +

are

>

=y

for

with respect

which x

<y

for

the

points

for

which

x2

+ y2 = 4

for

which

9?

Given the points Pi(l, - 3, 4), P2 (4, 2, - 2), P3 (0, 1, 5), and
P4 (6, 5, - 3), show that the sum of the projections of the line segments PiP2 P2Ps, and P3 Pi upon the x-axis is equal to the projection
of PiP 4 on this axis, and that the same is true of the projections on
the jy-axis and the 2-axis. Is this result true for any four points what6.

ever?

true for

Is it

any number

of points?

cube of side 5 has one vertex at the point (1, 0, 0) and the
three edges from this vertex respectively parallel to the positive
Find the coordinates of the
A:-axis, and the negative y- and z-axes.
7.

other vertices and of the center of the cube.

8.

Plot to scale on a single sheet of squared paper the following

nine points
(3,

2,

(3, 2, 3), (3, 4, 0), (3, 0, 0), (3, 0, 3), (3,

1), (3, 0,

1), (3, 3,

4, 3), (3,

- 2, 0),

1).

Find an equation of the locus of a point which is twice as far

from the ry-plane as from the #2-plane. From geometric consider9.

ations

what

is

this locus?

75

Lines and PI anes in Space

15.

Distance between

and

Two Points.

Direction

[Chap. 2

Numbers

Direction Cosines of a Line

Two

Angle between

Segment.
Line Segments

In deriving the formula for the distance between two points

PI(XI, y\, zi) and P 2 (x 2 y* z 2 ) we make use of Fig. 18. The
reader may get a good idea of this figure if he holds a box in the
,

room with

its edges parallel to

the three lines of intersection

two adjacent walls and the

of

floor,

considered to be the co-

ordinate planes. We note that

the angle PiQP 2 is a right angle,

S(x 2t yi,zi)

and consequently
(PiP 2 )

(PiQ)

the angle QRP 2

being a right angle, we
Also,

have

FIG. 18

Since
/i

i \

f\p

we have the
tance PiP 2

p O

<p

/?P

following expression for the square of the dis-

(15.2)

=
=

+ Pl Q)2
- X,) 2 + (y 2

(QR)2
(X2

If the segment PiP 2 were placed in any other octant, some of

the coordinates of PI and P 2 would be negative, and also some
of the expressions in the above parentheses might be negative,
but the above formula would still hold true. When the line

to the ry-plane, z 2
z\ = 0, and the above formula
reduces to the one in
3, as we should expect. When the line
is parallel

parallel to either of the other coordinate planes, the corresponding term in (15.2) is equal to zero. Hence we have the
is

following theorem

76

Distance between

Sec. 15]

[15.1]

The distance between

V(* 2 -

(15.3)

Points

the points (x\, y\, z\)

2

*i)

Two

~ yi) 2 +

(y 2

(22

and

(x 2 yi, 22) is

- zi) 2

is a generalization of Theorem [3.1] in the plane.

The length and sign of the directed segments QR, P\Q, and
RP2 are given by (15.1) no matter in which octant PI lies and
in which P 2 lies.
These numbers determine the "box" of

This

which PiP2 is a diagonal, and consequently determine the direction of PiP2 relative to the coordinate axes. They are called
direction numbers of the line segment. In like manner x\
#2,
z 2 are direction numbers of the line segand z\
y<z>
y\
ment P*P\. Thus a line segment has two sets of direction
numbers, each associated with a sense along the segment and
either determining the direction of the segment relative to
the coordinate axes. But a sensed line segment, that is, a
segment with an assigned sense, has a single set of direction
numbers.
Any other line segment parallel to PiP2 and having the
same length and sense as PiP 2 has the same direction numbers
as PiP2i for, this new segment determines a "box" equal in
every respect to the one for P\P%. This means that the differences of the x's, /s, and z's of the end points of such a parallel
segment are equal to the corresponding differences for PI and
P 2 Since one and only one line segment having given direction numbers can be drawn from a given point, we have that
a sensed line segment is completely determined by specifying
its initial point and its direction numbers.
There is another set of numbers determining the direction of
a line segment, which are called the direction cosines of the line
segment, whose definition involves a convention as to the positive sense along the segment. This convention is that, when
.

is not parallel to the ry-plane, upward along the

the positive sense on the segment when a segment
is parallel to the ry-plane but not parallel to the x-axis, toward
the observer in Fig. 18 is the positive sense on the segment (this

a segment

segment

is

is

the convention used in

3)

when

the segment

is

parallel

to the x-axis, toward the right is the positive sense. In Fig. 18

the distance PiP 2 is a positive number, being measured in

77

Lines and Planes in Space

[Chap. 2

the positive sense, and the distance P 2 P\ is a negative number

(the absolute values of these two numbers are the same), just
as distances measured on the #-axis to the right, or left, of a
point on the axis are positive, or negative. When the sense on

a line segment is assigned

a directed line segment.

by

we

this convention,

refer to

it

as

By definition the direction cosines of a line segment are the

cosines of the angles of the positive direction of the segment and
the positive directions of the x-, y-, and z-axes respectively; that is,
of the angles made by the positive direction of the segment with
line segments drawn through any point of the given segment
parallel to and in the positive directions of the coordinate axes.
Thus in Fig. 18, if we denote by A, B, and C the angles SPiP 2
QPiP 2 and TPiP 2 the direction cosines X, /z, and v (nu) of the
,

segment P\P 2 are given by

= cos B,
X = cos A,
,u
(15.4)

line

If

from

we denote the
(15.1)

and

since

cos C.

PiP 2 by

positive distance

we have

d,

(15.4)

= d\, y-2 yi =
QR = PiS = P\P 2 cos A, and
x2

(15.5)

#1

22

rf/z,

so on.

21

If

= dv

now we imagine

PI and P2 interchanged in Fig. 18, this does not alter X, ju,

and v, since their values depend only upon the direction of the
segment relative to the coordinate axes consequently in this
;

case equations (15.5)

become

Xi-X 2 =d\

yiy2=dfl,

2i-2 2

=di>,

the distance P2Pi in the new figure and is positive.

Hence equations (15.5) hold also if d is negative, that is, when

where 5
PI

is

is

above

P 2 on

the

line, in

which case the distance P\P 2

is

negative.
consider

We

now the difference, if any, in the above results

the positive sense of the segment makes an obtuse angle
with the *-axis or the ;y-axis. If the angle A for the segment
PiP 2 is obtuse, then x 2 < x\ and the first of equations (15.5)
when

is satisfied,

since cos

is

negative in this case

similarly

when

Consequently equations (15.5) hold in every case,

with the understanding that d is the sensed distance PiP 2
is

obtuse.

78

Direction

Sec. 15]

Numbers and

Direction Cosines

Thus a

line segment is completely determined by an end point,

direction cosines, and its directed length measured from the
given end point. Parallel line segments have the same direction

its

cosines, and direction numbers of two such segments are equal

or proportional, as follows from (15.5).
From the definition of the positive sense of a line segment
it follows that when PiP 2 is not parallel to the xy-plane v is

whereas X and

positive,

and

when PiP 2

#-axis, v

= 0,

and

+1

/z
;

is

is

take any values between

may

/z

parallel to the ;ry-plane but not to the

and X may take any value between

positive,

when PiP 2

is

parallel to the #-axis, v

/x

= 0,

l.

When the expressions for x 2 Xi, and so on, from (15.5) are
substituted in (15.2), and we note that (PiP 2 ) 2 is d 2 we obtain
,

X2

(15.6)

+ M 2 +" 2 =1-

The converse of this result may be established as in

of Theorem [3.2]. Accordingly we have the theorem
The

[15.2]

direction cosines X,

the equation
*>

/*

>

0,

0, then

X2

+ M2 +

< X<
M>

and, conversely,

0,

v2

v of

/z,

<
< X<
1

1,

any numbers

any

line

the case

segment satisfy

v is never negative;

/i

<

X,

when
/z,

when
^

when
and

= 0,

v satisfying these con-

ditions are direction cosines of a line segment.

Consider in connection with the line segment PiP 2 another

line segment PiPs, where P 3 is the point (#3, jVa, 23). We denote
by 9 the angle formed at PI by the sensed segments PiP 2 and
and we note that if they have the same direction,
PiP 3
= or 180 according as the segments have the same or
;

opposite sense. We consider the case when the two segments

do not have the same direction, and denote by Xi, /ii, v\ and

segments PiP 2 and PiPa

and by d\ and d 2 their respective directed lengths.
If we draw from the origin two line segments OP\ and OP*
parallel to and of the same directed lengths as PiP2 and
Pi P 3 respectively, we have a graph like Fig. 4, with the difference that the points P\ and P 2 are (rfiXi, d\^ rfi^i) and
X2, M2, ^2 the direction cosines of the

respectively,

'

79

Lines and Pla nes in opace

Now

d2 v 2 ).

0/2X2, <?M2,

if

- </lX0 2 +
2
+ M2 2 +
(X

t/ 2

/2

+ (d2 V2 + </r(X! + Mi 2 + "i 2

2 did (\i\ 2 + MiM2
2 did 2 (\i\ + MiM2 + ^1^2),

(d 2 \2

again denote the numerical

(3.7), but in place

12

we have equation

lengths of 0/Y and 0/Y,

of (3.8) we have

P=

and

/i

[Chap. 2

U2

(fi? 2j

</lMl)

*>2

/i

the last expression being a consequence of Theorem [15.2] and

2
2
2
the fact that k 2
From (3.7) and the above
di and / 2 = d 2
2
expression for I we have

(15.7)
t/2

rfi

d\

[15.3]

did 2 (\i\ 2

juiju 2

j/u/ 2 ).

are both positive or both negative, IJ 2

differ in sign,

/i/ 2

d\d 2

if

Hence we have

^1^2.

T/ze ow^
between two directed line segments which have
one end point in common and whose respective direction

cosines are Xi,

//i,

cos 6

(15.8)

where

e is

same sense

From
[15.4]

cos

/i/ 2

and
and rf 2

If

v\

and X 2

e(Ai\ 2

1 or

^t 2 ,

i> 2

is

+ MiM2 +

given by
v\i>2\

according as the two segments have the

(both positive or both negative) or opposite sense.

equations (15.7) and (15.5)

The angle 6 between

(xi, ;vi, 21) to the

the

we have

line segments

points (x 2 y 2 z 2 )
,

and

(x3 ,

from
;y3 ,

the point

23) is

given by

(15.9)

COS

= fe-^l)fe-^l) + (^2-^l)(^3-^l) + fe--^l)fe--^l)^

/1/2

where

l\

and

12

are the respective lengths (not directed dis-

tances) of the segments

From

(15.8)

2
(15.10) sin

and

(15.6)

0=1- cos

PiP 2 and PjP3

we have

(\!

Mi

+ Ml 2 + ^ 2 )(X +
2

/x 2

y2 2 )

Mi

M2
as

may be

verified

by multiplying out and reducing the

and expanding the second.

of these expressions for sin 2 6

80

first

Angle between

Sec. 151

Two

Line Segments

EXERCISES
Find the vertices of the "box" of which the line segment from
2, 3) to P2(2, 4,
1) is a diagonal; find also the direction
numbers and direction cosines of this segment, and the directed dis1.

Pi(l,

tance PiP2

2. Find a so that the line segments from the point (3, 1, 5) to

the points (a,
2, 3) and (1, 2,
4) make an angle whose cosine is
4/5 ; so that the two segments shall be perpendicular.

Find the condition to be satisfied by the coordinates of the points

and (x2 y2 z 2 ) in order that the line segment with these as
end points shall subtend a right angle at the origin.
3.

(xi, y\> Zi)

4.

Show

the origin,

that the points

if

and only

if

(x\, y\, z\)

and

(x 2

y2

z 2 ) are collinear

with

their coordinates are proportional.

be the point on the line segment PiP2 with end

and fa, yz, z 2 ) which divides the segment in the
ratio h\/h 2
denote by Qi, Q, Q 2 the orthogonal projections of Pi, P,
P2 on the xy-plane; show that Q divides the line segment QiQ2 in the
ratio hi/h 2 and that similar results hold when Pi, P, P2 are projected orthogonally on the yz- and xz-planes. Derive from these results and equations (4.5) and (4.6) the first two of the following
5.

Let P(x,

points

y, z)

(x\, yi, z\)

\

expressions for the coordinates of

Pi and P2, and hi and h 2

in

-f

hiy2

terms of the coordinates of

*-

_ fayi
^"

and derive the third by projection upon the

6.

Show

that,

when

h\

_ h 2 Zi

'

or h 2 in equations

yz(i)

and

4-

*2-planes.

of Ex. 5

is

a negative

number, these equations give the coordinates of a point on an extenhi

sion of the line segment beyond PI or beyond P2 according as
h2
and that PiP and PP2 are in the ratio
is less or greater than
in this case P is said to divide the segment PiP2 externally,
hi/h 2
and when h\/h 2 is positive, internally.
|

|,

Find the coordinates of the mid-point of the line segment

5, and in particular of the line segment joining the points
- 2, 3) and (2, 2, - 3). Find the coordinates of the points where

7.

of Ex.
(3,

a segment

is

trisected.

81

Lines and Planes in Space

[Chap. 2

the line segment joining the points (2, 3, 1) and

intersection with each of the xy- yz-, and xzwhat are the coordinates of these points of
and
planes respectively,
intersection ?

In what ratio

8.

2) cut

(1, 5,

9. If

is

its

the line segment of Ex. 8

until its length

10.

by

is

trebled,

what

produced beyond the second point

is

will

be the coordinates of

Show that the points (x\, y\, zi), (x 2 y 2 z 2 ), and

when there are three numbers k\, k2 and 3, all
,

ki

kiyi

Show

11.

y%,

(*!

(* 3

extremity ?
,

ja, 23)

different

are

from

such that

zero,

(#2,

collinear

its

z2 )

+X2 +

k2 y 2

= 0,
=
+ fayz 0,

k2

+
+

kixi
k\z\

fe* 2

k2 z2

+
+

= 0,
=
0.
faz^
3*3

that the medians of the triangle with vertices (x\, y\ z\)

and (*3, j3, 23) meet in the point with coordinates
t

Xz),

JO'l

+ J2

4- jVs),

+Z2 + 23

^(2i

).

12. Explain why equations (i) of Ex. 5 may be interpreted as giving

the coordinates of the center of mass of masses h 2 and hi at Pi and P2
respectively. Obtain the coordinates of the center of mass of masses

at points Pi, P 2 PS. Do the

mathematical induction.
using
points,

m\

2,

mz

same

for

n masses at n different

Equations of a Line.
Direction Numbers and Direction Cosines of a Line.
16.

Angle

of

Two

Lines

Consider the line through the points PI(*I, y\, z\) and
^2(^2, JV2, 2 2 ) and denote by P(x, y, z) a representative point of
the line. We consider first the case when the line is not parallel
to any one of the coordinate planes, that is, x 2 ^ x\, yz ^ y\,
and 2 2 ^ z\. The segments P\P and PiP 2 have the same direction by a characteristic property of a line (see
6). Conse-

quently their direction numbers are proportional, as follows

from
(16.1)

(15.5), that

is,

-i
Xz

*i

y\

y<2

z2

Zi

These equations are satisfied by the coordinates x, y, z of any

point on the line, and by the coordinates of no other point;
82

Equations of a Line

Sec. 16]

P(x, y, z) is not on the line, the segments P\P and PiP 2

io not have the same direction. From (16.1) we obtain the
for, if

three equations

22

y\

zi

readily seen that, if given values of x, y, and z

two
of equations (16.2), they satisfy also the third;
satisfy any
that is, two at most of these equations are independent. The
17.
geometric significance of equations (16.2) will be shown in

However,

We

consider next the case

ry-plane.
ive

it is

when

= z\,

In this case z 2

the line

and

for

is

parallel to the

any point on the

line

have
z

(16.3)

- zt = 0.

The direction numbers of the segments P\P and P\P^ are

*
and x<2
%i, y
%i, y%
yi
y\ 0, and the proportionality
3f these numbers is expressed by the first of (16.2).
Consequently this equation and (16.3) are equations of the line.
9

Similar results hold

Dr

the #2-plane.
When the line

is

when

parallel to the >>2-plane

0,

z-z, =

and similarly for the cases

to the jy-axis or the 2-axis.

are equations of the line

line is parallel

is

parallel to the x-axis,

y-yi =

(16.4)

Accordingly
[16.1]

the line

when

the

we have

Equations (16.1) are equations of the line through the

points (xi, y\, z\) and (#2, 3^2, 2), with the understanding
that, if the denominator of any one of the ratios is equal to
zero, the numerator equated to zero is one of the equations
of the line;

these exceptional cases arise

when

one of the coordinate planes or

parallel to
coordinate axes.

Thus a

line,

lying in space,

which
which

when

is

one-dimensional,

is

three-dimensional,

83

is

to

the line is

one of the

considered as

defined

by two

Lines and Planes in Space

[Chap. 2

equations, whereas when considered as lying in a plane, which

5.
is two-dimensional, it is defined by one equation, as in
Since any two segments of a line have the same direction,

numbers are proportional. By definition the dinumbers of any segment of a line are direction numbers of
the line. Consequently there is an endless number of direction
numbers of a line, the numbers of any set being proportional
to the corresponding numbers of any other set. Thus, if u, v w
are direction numbers of a line through the point (x\, yi, zi),
their direction

rection

(16.5)

are equations of the line, since these equations express the proportionality of two sets of direction numbers of the line for

each point

(x, y, z) of the line.

Conversely, for each set of numbers u, v, w, not all equal to
zero, equations (16.5) are equations of the line through the
point (*i, y\ Zi) with direction numbers u, v, w. For, if we
define numbers x 2 y 2 z 2 by the equations
9

- xi = u,

*2

(16.6)

y 2 -yi

= v,

z2

-zi

w,

y 2 z 2 equations (16.5) are expressible in the form

and thus are equations of the line through the points
(*i, y\ 21) and (x 2 y 2 z 2 ). As a result of the above discussion
and the fact that any set of direction numbers of a line are
direction numbers also of any line parallel to it, we have
in terms of x 2

(16.1),

[16.2J

and

If (*i> y\> z i)
the quantities

y2
y2
and of any

x2

bers of the line,

(x 2

x\,

z 2 ) are

any two points of a

yi, z 2

zi

are direction

line parallel to

In defining direction cosines of a

line

line,

num-

it.

segment in

15,

we

assigned sense to a line segment. Since this applies to all segments of a line, we have that the positive sense along a line not
parallel to the ry-plane is upward, that is, z increasing when
;

and not parallel to the *-axis,

the positive sense is the direction in which y increases; and
when parallel to the #-axis, the positive sense is the same as on
a

line is parallel to the ry-plane

the x-axis.

84

Parametric Equations of a Line

Sec. 16]

segments of a line have the same direction, the diall segments are the same
we call them the
direction cosines of the line. Accordingly the direction cosines X,
V, v of a line are the cosines of the angles which the positive
direction of the line makes with the three line segments from
Since

all

rection cosines of

any point on the line and parallel to the positive directions of

the #-, y-, and 2-axes respectively. Hence Theorem [15.2] holds
for direction cosines of a line. As a consequence of this theorem
we have
[16.3]

__ _

// u,

v,

are direction numbers of a line, the direction

cosines of the line are given by

.

~~

e^/u

_x

,_

+v +w
2

"

where

e is

1 or

and eu which

is

'

"~

1 so that the first of the

not zero shall be positive.

numbers ew,

ev,

In fact, since e 2 = 1, the quantities X, /z, v given by (16.7) satisfy

the conditions of Theorem [15.2].
Since equations (15.5) hold for all segments of a line with

any particular point

[16.4]

The

(x\, y\, z\)

fji 9

v has the parametric equations

x = *i + d\
the
to

we have

line through the point (x\, y\, z\) with direction co-

sines X,
(16.8)

as an end point,

= yi + d,

= Zi + dp,

parameter d being the directed distance from

P(x, y,

Another

(x\, y\, z\)

z).

set of

parametric equations of a

when we observe

if

obtained from
is not parallel
on the line the

line is

the line

that,
equations (16.1)
to a coordinate plane, for each point (x, y, z)
ratios in (16.1) have the same value, say /, depending upon the
values of x, y, and z. If we put each of the ratios in (16.1) equal

to

/,

and solve the

resulting equations for x, y,

2,

85

and

y =(1-0* + #2,

z,

we obtain

Lines and Planes in Space

[Chap. 2

Conversely, the values #, y, z given by (16.9) for any value of

/ are coordinates of a point on the line with equations (16.1), as

one

by substitution. Equations (16.9) hold also when

parallel to a coordinate plane or coordinate axis, as

verifies

the line

is

the reader can show, using a method analogous to that used in

connection with equations (5.7). Hence we have

Equations

[16.5]

are parametric equations of the line

(16.9)

through the points

(x\ 9 y\, z\)

and

(#2, >>2, 22)

the coordi-

nates x, y, z of any point on the line are given by (16.9)

for a suitable value of /, and conversely.

Ordinarily two lines in space do not intersect, even if they

are not parallel. Two nonintersecting, nonparallel lines are
said to be skew to one another, and, for the sake of brevity,
they are called skew lines. If through a point of one of two

such lines one draws a line parallel to the other, each of the
angles so formed is called an angle of the two skew lines. Accordingly from Theorem [15.3] we have

The angle

[16.6]

of the positive directions of two lines with di-

<f>

rection cosines Xi,

COS

(16.10)

From

this

</>

cos

</>

e\ is

and

larly for

As a

e\u\

1 or

is

given by

2
Vi

+ Wi

)(W 2

+ V 2 + W2

numbers e\w\,
positive, and simi-

1 so that the first of the

which

is

not zero shall be

we have
numbers
and only if

lines with direction

are perpendicular,
(16.12)

w2

02,

2.

corollary

Two

[16.8]

=
2

e\v\ t

V\V^,.

we have

[16.3]

w\ and w 2

v\,

(Hi

where

^2 is given by

positive directions of two lines with

<f>

direction

(16.11)

jU2,

+ /Zi//2 +

XiX2

theorem and Theorem

The angle
of the
numbers u\,

[16.7]

and \^

v\

jui,

if

UiU 2

u\, v\>

+ v\V2 + w\W2 = 0.
86

w\ and w 2

#2,

w%

Angle

Sec. 16]

of

Two

Lines

EXERCISES
1.

Find the distance between the points

and the direction cosines of the

2.
(3,

(6,

4.

1,

3)

For what value of a do the points

1,
4) lie on a line?

What

and (-

3, 2, 5),

through these points.

Find equations of the line through the points

through the points (1,3,- 2) and (-2,

2, 1)

3.

line

(2,

(a,

(2,

3)

and

2, 3),

and

1,

3, 1).

3, 10), (2,

are the direction cosines of the coordinate axes?

What

are equations of the axes ? What are the direction cosines of the lines
through the origin bisecting the angles between the coordinate axes?

What
5.

are equations of these lines ?

Find the direction cosines of the

lines equally inclined to the

coordinate axes.
6.

Direction numbers of a line through the point (2,

1, 3) are
4 find the equations of the line in the parametric form (16.8).

7.

Show by means

3, 1,

(x\, y\, Zi)

and

x2

where k

is

of (16.9) that if the coordinates of

are in the relation

two points

(#2, ^2, 22)

y2

kxi,

some constant

different

z2

ky\,

kzi,

from zero, the line joining the two

How does this follow from geo-

points passes through the origin.

metric considerations without the use of (16.9)

Find the cosine of the angle of the positive directions of the

through the points (3, - 1, 0), (1, 2, 1) and (- 2, 0, 1), (1, 2, 0)

8.

lines

respectively.
9.

and

For what value of a are lines with direction numbers 1,

2, 2
a perpendicular? For what value of a is the cosine of the

2, 2,

angle of the positive directions of these lines equal to 4/9?

10.

two

Find the direction cosines of a line perpendicular to each of

whose respective direction numbers are 3, 5, 6 and 1, 3, 4.

lines

11.

Show

that equations (16.5)

where

is

a parameter.

distance from
12.

Show

(x\, y\, zi)

What
to

that equations

be written

may

y = y\ +

x = xi + ul,

vl,

relation does

= zi + wl,
/

bear to the directed

(*, y, z) ?
(i)

of Ex. 5,

15 are parametric equations

of the line through (xi, y\ *i) and (* 2 y^ 2 2 ), and find the relation
between hi and h 2 in these equations and / in (16.9).
9

87

Lines and Planes in Space

13.

and

Show

that the parameter

in (16.9)

is

d* are the directed distances of the points

respectively from the point

14.

(x\,

y\

[Chap. 2

equal to d/d2 where d

,

(x, y, z)

and

(* 2 ,

;y 2 ,

22)

z\).

Find the locus of a point equidistant from the three coordinate

planes.
15.

Find the locus of a point whose distances from the

xz-planes are in the ratio

16.

Find the locus of a point P(x,

to Pi(l,

17.
(1.

and

y, z) so that the line joining

perpendicular to the line through PI with direction

2, 3) is

numbers

xy-, yz-,

1:2:3.

3.

2, 1,

Find the locus of a point at the distance 2 from the point

- 2,

3).

17.

An

Equation of a Plane
14

In defining coordinates in
the reader understood what

we took

it

meant by a

for

granted that

and that
planes meet they intersect in a straight line. Euclid
proved the latter result by means of his definition of a plane
as a surface such that a straight line joining any two points
is

plane,

when two

of the surface lies entirely in the surface, as stated in

14.
By means of this characteristic property of the plane we shall

prove the theorem

[17.1]

Any

equation of the first degree in x, y,

tion of

and

z is

an equa-

a plane.

Consider the equation

ax

(17.1)

+ by + cz + d = 0.

It is understood that the coefficient of at least

one of the un-

knowns is different from zero, that is, that we are dealing with
a nondegenerate equation (see
this understanding ap1)
;

plies to all theorems concerning equations of the first degree in

x, y, z in this chapter.

Let PI(XI,
locus defined
(17.2)

yi, 21)

and

P 2 (x 2 y 2t

by equation

(17.1)

axi + byi + czi + d = 0,

88

22) be any two points of the

then we have

ax 2

+ by2 + cz2 + d = 0.

An

Sec. 17]

In

16

it

Equation of a Plane

was shown that the coordinates x, y, z of any point

and P2 are given by (16.9)

of the line through the points PI

for an appropriate value of the parameter /. When the expressions (16.9) are substituted in the left-hand member of (17.1),
the resulting expression may be written in the form
(1

- t)(axi + byi + czi + d) + t(ax 2 + by2 + cz 2 + d).

In consequence of (17.2) this expression is equal to zero for

every value of t. Hence the coordinates of every point of the line

through Pi and P 2 satisfy (17.1) that is, every point of this line
is a point of the locus defined by (17.1). Since this result holds
;

every pair of points PI and P 2 whose coordinates satisfy

(17.1), the theorem follows from Euclid's definition of a plane.
We shall consider several particular forms of equation (17.1)
and in the first place prove the theorem
for

[17.2]

When two and only two of the coefficients of x, y, and z

in equation (17.1) are equal to zero, the locus is one of the
coordinate planes or a plane parallel to it according as
d

or

0.

Consider, for example, the case

the equation

Qx + 0y +

(17.3)

cz

+d=

when a

= b = 0,

(usually written cz

that

is,

+ d = 0).

This equation is satisfied by z =

d/c and any values of x and
of
is
an
the *jy-plane when d = 0,
equation
y consequently it
d
of
the plane parallel to the ry-plane and at
and when ^
the distance
d/c from it, above or below it according as
d/c is positive or negative. In like manner,
;

(17.4)

ax

+ Oy + Qz + d = 0,

Qx +

by

+ Oz + d = Q

are equations respectively of the jz-plane or a plane parallel to

it and of the #z-plane or a plane parallel to it according as

= Qord^O.
Next we prove the theorem

[17.3]

When one and only one of the coefficients of x, y, and z

in equation (17.1) is equal to zero, the locus is a plane
through one of the coordinate axes or parallel to it according
as d

or

d^

0.

89

Liincs an dpi anes in opace

when

Consider, for example, the case

[Chap. 2

= 0,

that

is,

the

equation

ax

(17.5)

+ by + 0z + d =

+ by + d = 0).

(usually written ax

The points

(x, y, 0), where x and y are solutions of this equation,

upon a line in the ;ry-plane, namely, the line whose equation
in two dimensions, as discussed in Chapter 1, is ax + by + d = 0.
But any such pair x, y, say x\ and yi, and any z satisfy equation
hence the point (x\, y\, z) for any z
(17.5) in three dimensions
is on the line through (x\, y\, 0) parallel to the z-axis. Therefore
equation (17.5) is an equation of a plane perpendicular to the
;ty-plane and meeting the latter in the line whose equation in
when d = 0, this plane
two dimensions is ax + by + d =
lie

passes through the z-axis.

In like manner

+ by + cz

(usually written by

+ cz + d = 0)

+ Oy + cz

=Q

(usually written ax

+ cz + d = 0)

(17.6)

and
ax

(17.7)

are equations of planes perpendicular to the ;yz-plane

and the #z-plane respectively.

now

to equaobtained from
equations (16.1) of the line
through the points (xi,yi,zi),
(#2, y<2, z 2 ), we observe that

Referring

tions

(16.2),

these equations are equations

of planes, each containing

and parallel to the

and #-axes respecas shown in Fig. 19.

the

line,

z- f

y-,

tively,

Next we
[17.4]

shall

Any

FIG. 19

prove the converse of Theorem

plane

is

defined algebraically by

first degree in x, y,

and

[17.1],

namely,

an equation of

the

z.

In proving this theorem we remark that the ry-plane is deby (17.3) for d = 0, that is, by z = 0, and that any plane

fined

90

An

Sec. 17]

Equation of a Plane

parallel to the ry-plane is defined by (17.3) with c and d numbers

such that
d/c is equal to the directed distance of the plane
from the ry-plane. Any other plane intersects the ry-plane in
a line. In deriving an equation of such a plane we make use
of the property of a plane that it is completely determined by
three noncollinear points, that is, by three points not on the
same line, or, what is the same thing, by a line and a point not
on the line this property follows from Euclid's definition of
a plane, stated at the beginning of this section. The line in
;

which the given plane cuts the ry-plane is defined in this plane
by an equation ax + by + d = (in accordance with the results
5) when we are dealing with the geometry of the plane
but when we are dealing with the line in space, it is defined by
this equation and the equation z = 0. Consider then the equation

of

ax

(17.8)

+ by + d + cz = Q,

a, b, and d have the values from the above equation of

the line in the ry-plane, and c is as yet unassigned. Whatever
be c, by Theorem [17.1] this is an equation of a plane, in fact,

where

of a plane through the line, since the coordinates of

on the
(17.8).

line satisfy

z\

+ by + d =

Suppose now that

plane not on
9 0.

ax

If

and

(x\, y\, zi) is

0,

any point
and consequently

a point of the given

of intersection with the xy-plane;

found from the equation

its line

now

c is

0*i

(17.9)

then

+ tyi + + c*i = 0,
rf

and this value is substituted in (17.8), the resulting equation

is an equation of the given plane, since the plane with this
equation passes through the point (xi, y\, Zi), in consequence of
= Q in the plane
(17.9), and through the line ax + by + d
=
z
0; and the theorem is proved.
From Theorem [17.1] and equations (17.1) and (17.2) it
follows that
(17.10)

a(x

- xi) +b(y- yi) + c(z - *0 =

is an equation of a plane containing the point P\(x\, y\, z\).

Since x
Zi are direction numbers of the line
x\, y
y\, z
joining PI to any point (x, y, z) in the plane, it follows from
(17.10) and Theorem [16.8] that a, 6, c are direction numbers

91

Lines and Planes in Space

[Chap. 2

of a line perpendicular to every line in the plane which passes

through Pi ; hence the line through PI with direction numbers

We

follow the custom

a, 6, c is perpendicular to the plane at PI.
of saying that this line is the normal to the plane at PI. Since
PI is any point of the plane, a, b, c are direction numbers of

every normal to the plane;

Hence we have

parallel.
[17.5]

the normals are

all

evidently

Theorem

(a generalization of

[6.9])

The geometric significance of the coefficients a, b, c of an

of a plane is that they are
equation ax + by + cz + d =
the
normals to the plane.
each
direction numbers of
of

From Theorem

[17.5]

and equations

(16.5)

we have

The equation

[17.6]

a(x
is

an equation

- xi) + b(y - ;yi) + c(z - 21) =

of a plane through the point (xi, y\, z\),

and

*-*i=Z=J!l = L=li

(17.11)
'
^

are equations of the normal to the plane at this point.

theorem and Theorem [6.10],

a plane, or is parallel to a plane, it is perto
the
normals
to the plane, and conversely. Hence
pendicular
we
have
Theorem
[17.5]
by

Compare
If

this

line lies in

[17.7]

line with direction

in, the

numbers

u, v y

plane with equation (17.1),

au

(17.12)

if

is

parallel

and only

to,

or lies

if

+ bv + cw = 0.

When two planes are perpendicular, the angle of the normals

to the planes at a point on their line of intersection is a right
angle,
[16.8]

and conversely. Hence

and [17.5] we have

in consequence of

Theorems

The planes with equations

[17.8]

(17.13)

a& + biy + ciz + di = 0,

are perpendicular
(17.14)

to

a 2x

one another,

if

+ b y + c 2 z + d2 =
2

and only

aia 2 + bib* + CiC 2 = 0.

92

if

An

Sec. 17]

If

we wish

Equation of a Plane

to construct a plane through a point

to a given plane,

we draw through P two

parallel
lines parallel to the

given plane, and the plane determined by these two lines is

the plane desired. This construction is equivalent to that of
constructing the plane through P perpendicular to the line
through P which is normal to the given plane. Hence a line

normal to one of two

parallel planes is normal to the other

In consequence of this result and of Theorem [17.5] we

also.

have

An

[17.9]

equation of any plane parallel

ax

+ by + cz + =

ax

+ by + cz + e =

to the

plane

is

(17.15)

for a suitable value of the coefficient

We have

remarked that a plane

is

e.

determined by three non-

collinear points. Accordingly, if we wish to find an equation

of a plane through three noncollinear points, we have only to

substitute the coordinates of the points in equation (17.1) and

solve the three equations thus obtained for a, 6, and c. When
the resulting expressions are substituted in (17.1), d appears as

a factor, unless d = 0, which emphasizes the fact that an equation of the first degree in x, y, and z, and any constant multiple
of this equation, are equations of the same plane. More expeditious methods of finding an equation of a plane when the
coordinates of three of its points are given are developed in
21 and 23.

EXERCISES
1.

(2,
2.

(-

1,

Find an equation of the plane through the points

1,

(2, 3,

0)

and

0) parallel to the 2-axis.

Find an equation of the plane through the three points (1, 1, 1),
- 7, - 5). Is the plane parallel to one of the
1, 1), and (- 3,

axes?
3.
(1, 3,

Find an equation of the plane through the origin and the points
2)

and

(2,

1,

1).

93

Lines and Planes in Space

Show

4.

collinear,

that the points

and

find

1), and (1, 3,

3) are
set of planes containing these

(1, 2,

(1, 1, 1),

an equation of the

[Chap. 2

points.

Find the intercepts on the coordinate axes of the plane

5.

=i
?+2+
h ^k
g

When

can an equation of a plane not be put in

6.

twice

this intercept

form ?

Find the locus of a point whose distance from the ry-plane

distance from the ^z-plane.

is

its

7.

For what value of the

coefficient a are the planes

ax-2y + z + 7 = Q

and

3x + 4;y-22+l =

perpendicular to one another?

8. Find equations of two planes through the point
pendicular to the plane
2 % - 2 y + 5 z - 10 = 0,
(i)

1, 2,

0) per-

one perpendicular to the ry-plane and the other perpendicular to the

Using this result, find an equation of the set of planes through

>>z-plane.

the point perpendicular to the plane

9.

(i)

Find an equation of the plane through the

x-l = y+2 =

-35

y + z + 2 = Q.

perpendicular to the plane x

10.

Find an equation of the plane

7x

and containing the point

11.

lar to the plane

12.

3x +

(1,

For what value of a

is

line

3y

parallel to the plane

+z-5=

2, 3).

the plane

3x

4y-7z+l = 0?

2y +

az

= Q perpendicu-

Find an equation of the locus of a point equidistant from the

(8, 3, 4) and (3,

1,
2), and show that it is the plane perpendicular to the line segment joining the two points at the midpoint of the segment.

points

13. Show that the angle

of the positive directions of the normals
to the planes with equations (17.13) is given by

+ ClC2
2
fteaVfa* + M + d (02 2 + fe 2 + c2
under what conditions d is + 1 or
1, and e 2 is
cos0

ai

2+

blb*
3

and

state

94

-h 1

or

1.

The Directed Distance from

Sec. 18]

Show

14.

when the

that

expressions

(i)

a Plane to a Point

of Ex.

15 are substituted

5,

an equation of a plane, the resulting value of the ratio h\/h% gives

the ratio in which the segment P\P^ or the segment produced, is
divided by the point of intersection of the line of the segment and the
given plane. Find the ratio in which the line segment joining the
- 6, 2) is divided by the point of intersection
points (2, 3, - 1) and (5,
of the line of the segment and the plane 3^
2^ + 62 2 = 0; find
in

also the coordinates of the point of intersection.

18.

The Directed Distance from a Plane to a

The Distance from a Line to a Point

We

shall establish the following

The

[18.1]

directed distance

(18.1)
to the

point

from

+ by +

ax

theorem

the

cz

= axi

plane

+d=

(xi, y\, z\) is given

(18.2)

Point.

by

czi

yi

+ b + c2

e^/a 2

1 so that the first of the numbers

1 or
where e is
ea which is not zero shall be positive.

When
x

+ a- =

= c = 0,

that

is,

ec, eb,

equation (18.1) of the plane reduces to

the plane

is

parallel to the ;y2-plane

and at

d/a from this plane. Consequently the distance

= x\ + - which
to PI is
(
to x\

the distance

from the plane

result follows

from

equal

(18.2).

In this case

is

positive or negative

lies to the right or left of the given plane.

consider next the case when c = 0, in which case the
plane is perpendicular to the #y-plane and passes through the
It is evident that the distance
d
0.
line ax
0, z
by

according as PI

We

from the plane to P\(XI, y\, zi) is equal to the distance in the
d = Q to the point (x\, yi, 0).
by
ry-plane from the line ax
But as given by (8.5) this distance is the number obtained

from

(18.2)

on putting

= 0.

95

Lines and PI anes in Space

We

consider finally the case

when

[Chap. 2

that

0,

is,

when

the

plane cuts the 2-axis. From Theorems [17.5] and [16.3] we

have that the direction cosines of a normal to the plane are

given by
a, b, c

(18.3)

+ b + c2
2

1 according as c >
or c < 0. If we denote
in
the
which
the
normal to the plane
point
by P2(*2, JV2, 22)
we
meets
the
have
from
PI
plane,
through
(15.5)

where

e is

(18.4)

xi

or

x2

y\

/A,

y2

z2

z\

/M,

/*>,

the distance from Pz to Pi, positive or negative acas

Pi is above or below P 2 on the normal. Solving
cording
equations (18.4) for # 2 j2, and 2 2 and substituting the resulting

where

is

expressions in (18.1),
making use of (18.3),
axi

we

obtain,

+ by, + czi + d = l(a\ + b + cv)

= le Va + 2 + c 2
2

since

on rearranging the terms and

1/e = ^,

fr

"

-.

e\/a 2

+ b2 + c2

from which we derive the expression

(18.2)

and

the theorem is proved.

We turn now to the problem of finding the distance d from
a line to a point P\(x\, >>i, zi). Let X, ju, v be direction cosines
of the line, and P2(*2, y*> 22) a point on the line. Denote by
6 the angle between the line and the line segment P 2 Pi
then
;

rf

(18.5)

PTP2~sin0,

where PiP 2 denotes the length of the line segment. The direction cosines of the line segment PiP 2 are, to within sign at most,

From these expressions, (18.5), and (15.10)

desired result
(18.6)

d2

yi

fJ>

22-

%2

96

we have

as the

TKe Distance from a Line

Sec. 18]

to

a Point

EXERCISES
1. Find the distance from the plane 2 x
y + 2 z 6 = to the
1, 2), and the coordinates of the point in which the normal
point (1,
to the plane from the point meets the plane.

Find an equation of the plane through the point

2.

parallel to the plane x

+ 2y

2z +

= Q. What

is

(1,

2, 3)

the distance

between the planes?

Find equations of the two planes at the distance 5 from the
and perpendicular to a line with direction numbers 1,
3, 2.

3.

origin

What

4.

the distance between the planes 2 x

is

5. Given two intersecting planes and a point P in one of the four

compartments into which the planes divide space, that is, one of the
four dihedral angles of the planes, show that the dihedral angle in which
P lies, and the angle between the positive directions of the normals to
the two planes through P are supplementary when P lies'on the positive
or negative sides of the two planes, and that these angles are equal when
P is on the positive side of one plane and the negative side of the other.
</>

Determine whether the origin

angle between the planes

6.

dihedral

8x-j> +

+3y

42-5 = 0.

Find equations of the planes which are parallel to the plane

and at the distance 2 from- it (see equation
2 + 6 =

7.

2 x

2 x

an acute or obtuse
62 + 3 =
and

in

lies

+2y

(8.6)).

Find the locus of a point equidistant from the planes

a 2 x + b 2 y + c2 z + d2 =
and
+ b\y + Ci2 + d\

8.

a\x

0,

consists of the two planes which bisect the angles

and show that
formed by the given planes when the latter intersect; discuss also
it

the case

when

the given planes are parallel.

Apply the

result to the

planes in Ex. 6.
9.

Find the locus of a point which is twice as far from the plane
3 = Oas from the plane x-2;y + 22-6 = 0.

2x + 2;y-2 +
10.

lines

Show

that for the plane containing the two lines

JL

JL

Mi

V\

= JL,

JL

U2

W\

with direction numbers k\u\

any values of the constants

or are parallel to it.

k\

= JL = _L,
V2

k 2 u 2t k\v\

tV2

and k 2 not both

,

97

k2 v 2

k\w\

k 2 tuz for

zero, lie in this plane

Lines and Planes in Space

Find the angle which the

11.

jection
the line
12.

line

on the plane * + 2;y-3z

makes with the plane.

[Chap. 2

= = -~ makes
^

= 0.

This angle

Find what relation or relations must hold

the coefficients of the equation ax + by

order that the plane
(a) have the intercept 2 on the jy-axis

in

is

with

its

pro-

called the angle

each case between

a plane in

+ cz + d = Qof

(b)
(c)

(d)

have equal intercepts

+ 22-l=:0;

be parallel to the plane 2*-3;y

be perpendicular to the jyz-plane

contain the point (5, - 4, 2)

X
- = -=V = Zr
(/) contain the line
Z
i
o
the
to
x-axis
be
parallel
(g)
(e)

(h)

be at the distance

2 from the origin.

13. Show that if (x\, y\, z\) and (x 2 y^ 22) are points of two parallel
and ax -f by + cz -f- d2 0,
planes with equations ax + by + cz + di =
the distance between the planes is the numerical value of
,

V0 2 -f b 2 4- c 2
Under what conditions

is

this expression

a positive number ?

Show

that the numerator in the expression (i) of Ex. 13 is equal

d2 and that the number given by the expression (i) is equal to
to d\
the length of the segment between the two planes of any normal (the
end points of the segment being the projections upon the normal of
14.

any point

19.

in each plane).

Two Equations of the First Degree in Three Unknowns.

A Line as the Intersection of Two Planes

Consider the two equations

(19.1)

aix

+ bty + c\z + d\ = 0,

Since

a2x

+ b 2y + c 2 z + d2 = Q.

by hypothesis the coefficient of at least one of the unknowns in each equation is different from zero, two of the
unknowns in either equation can be given arbitrary values, and
then the other can be found. Thus, when a\ ^ 0, if we give
98

Two

Sec. 19]

y and

Equations of tKe First Degree

and solve
and the given values of y and z
constitute a solution of the equation. Although each of equations (19.1) admits an endless number of solutions involving
two arbitrary choices, it does not follow necessarily that the
equations have a common solution.
We assume that equations (19.1) have a common solution.
Instead of denoting it by xi, y\, z\ to make it evident that we
are dealing with a particular solution (after the manner followed in 9), we count on the reader's thinking of x, y, and z
in what follows as the same set of numbers in the two equations.
If we multiply the first of equations (19.1) by b 2 and from the
result subtract the second of (19.1) multiplied by bi, the final
z

any values

in the first of equations (19.1),

for the value of x, this value

result

may

be written
0i

(19.2)

This

02

b\

Ci

b2

Co

= 0.

the process followed in

9, leading to equation (9.4)

as eliminating y from the two equations, meaning
that from these equations we obtain an equation in which the

we

is

refer to

coefficient of

and

it

is

zero.

If,

in similar

manner, we eliminate x

respectively from equations (19.1),

0i

bi

a?

b2

(19.3)

y-

Ci

01

a\

C2

02

02

we obtain

and
Ci

01

02

(19.4)

If

y+
2

the determinant

is

= 0.

not equal to zero,

when

z in

02

equations (19.2) and (19.3) is given any value, these equations

can be solved for x and y, which values together with that
assigned to z constitute a common solution of equations (19.1).

Thus

in this case there is

tions of these equations.

an endless number of common

In like manner,

is

if

c2

not equal

to zero, equations (19.2) and (19.4) can be solved for

99

solu-

y and

Lines and Planes in Space

for

any given value of x

and

0i

if

is

02

[Chap. 2

not equal to zero,

equations (19.3) and (19.4) can be solved for x and z for any
given value of y. Hence we have

Equations (19.1) have an endless number of solutions when

[19.1]

any one of

the determinants
f

(19.5)

02

b2

b2

c2

c2

02

not equal to zero.

is

When

one at

least of the

determinants (19.5)

is

not equal to

common

solutions of equations (19.1) are coordinates

of points of the line of intersection of the two planes with the

zero, the

equations (19.1).

Two

[19.2]

Hence we have

equations (19.1) are equations of a line when one at

determinants (19.5) is not equal to zero; this

least of the

line is the intersection of the planes with equations (19.1).

be shown how equations (19.1) can be put in the

when
a common solution is known.
(16.5)
We turn now to the consideration of the case when all the
determinants (19.5) are equal to zero. By Theorem [9.6] this

20

In

it

will

form

condition

is

equivalent to the three equations

/ is some number, not zero.

determinants in equations (19.2),
to the respective values

where

- bi(d 2 - tdi),
If

not

all

ai(d>2

By means
(19.3),

of (19.6) the last

and

(19.4)

reduce

- /</i),

three of these quantities are equal to zero, one at least

and (19.4) cannot be true when all three of

of (19.2), (19.3),

the determinants (19.5) are equal to zero, and consequently

the assumption that equations (19.1) have a common solution
If d 2 ^ tdi, all three of the above expressions can be
is false.
equal to zero only in case 0i = bi = Ci = 0, which is contrary
to the hypothesis that one at least of these coefficients is not
= td\, all the above expressions are equal
If, however, d 2

zero.

100

Sec. 19]

to zero.

But the equations d 2

= td\

tions that the second of (19.1)

first.

[19.3]

Two

Line as the Intersection of

and

Planes

(19.6) are the condi-

a constant multiple of the

is

Hence we have

Two equations of the first degree in three unknowns neither

of which is a constant multiple of the other admit an endless
number of common solutions or none; the condition for the
latter in

terms of equations (19.1)

2

(19.7)'
v

is

&

= 62^02
1

with the understanding that

and similarly for

the b's

if

and

one a

is

zero so is the other a,

c's.

If we say that two planes are parallel when they do not have
a point in common, Theorem [19.3] may be stated as follows:

The planes with equations

[19.4]

#2

Another way of stating

[19.5]

plane

and only
e

d, or

is

if

(19.1) are parallel, if

_ ^2

this

^2

theorem

is

__

Cj2

parallel to the plane

its

equation

is

ax

any constant multiple

and only

if

+ by + cz + d = 0, if
+ by + cz + e = where
ax

of this equation.

This theorem is equivalent to Theorem [17.9], which was derived from another, and consequently an equivalent, definition
of parallel planes.
Next we establish the following theorem, which
zation of Theorem [10.1]

is

a generali-

[19.6]

When
when

(19.8)

/i

two equations (19.1) are equations of a

line, that is,

the coefficients are not proportional, the equation

(aix

+ biy + Ciz + d{) +

t2

for any values of the constants

(a 2 x
t\

+ b 2y + c z + d2 = 0,

and

t2 ,

not both zero,

is

an

and
equation of a plane through the line defined by (19.1)
(19.8) is an equation of each plane through this line for
;

suitable values of

t\

and

/2 .

101

Lines and Planes in Space

In order to prove the theorem,

we

[Chap. 2

rewrite equation (19.8) in

the form
(19.9)

+ 02/2)* +

(0i/i

+ b2

(61/1

t2

)y

(citi

+ c 2 2 )z
t

Since by hypothesis the coefficients of x, y, z in (19.1) are not

proportional, it is impossible to find values of /i and / 2 not
,

both zero, which

will

make

the coefficients of

all

x, y,

in

Consequently from Theorem [17.1] we

have that equation (19.9) is an equation of a plane for any
values of t\ and t 2 not both zero. Moreover, it is an equation
of a plane through the line with equations (19.1)
for, the two
(19.9) equal to zero.

expressions in the parentheses in (19.8) reduce to zero when

x, y, and z are given the values of the coordinates of any point
on the- line. Hence the coordinates of every point on the line
satisfy equation (19.8) whatever be ti and / 2
for each choice of t\ and t 2 equation (19.8)

is

and consequently
an equation of a

plane containing the line.

Conversely, any particular plane containing the line is determined by any point of the plane not on the line, say (x\ y\ Zi).
If the values %i, y\, z\ are substituted in (19.8), we obtain
9

tiAi

(19.10)

where A\ and

A2

are the

t2

A2 =

numbers

to

0,

which the expressions

in

(19.8) reduce when x, y,

replaced by
Moreover, not both A\ and A 2 are equal to zero,
since the point (x\, y\, z\) is not on the line. If the point
= 0, and
(x\ y\y z\) is on the first of the planes (19.1), then A\
=
from (19.10) it follows that t 2
and that t\ can take any
value. Also, if (*i, yi, Zi) is on the second of the planes (19.1),
we have A 2 = and t\ = 0. For any other plane both A\ and
A 2 are different from zero. If then we choose any value other
than zero for t\, substitute this value in (19.10), and solve
for /2, the resulting value of t 2 and the chosen value of ti are
such that when they are substituted in (19.8) the resulting
equation is an equation of the plane through the line with
equations (19.1) and the point (xi, y\, z\). Thus the theorem
z are

parentheses in
xi, y\, z\.

is

proved.

102

A Line as

Sec. 19]

the Intersection of

Two

Planes

If we take two sets of values of t\ and / 2 which are not proportional, the corresponding equations (19.8) are equations of
two different planes, and these two equations are equations of

Hence we have

the line (19.1).

Equations of any two of the set of all the planes through

a line constitute equations of the line.

[19.7]

EXERCISES
Find an equation of the first degree in x, y, and z which has
a solution but which has no solution in common with

1.

2, 3) for

(1,

the equation

5* + 2.y-3z+l =

0.

Find an equation of the plane through the point

2.

parallel to the plane

*-2.y + 7z +

Find equations of the

3.

are cut

by

the plane 3 x

lines in

2^-3^ + 2 + 2 = 0,
and the point

1,

(1,

1);

1,

2)

and

which the three coordinate planes

4y + 5z

10

= 0.

Find an equation of the plane containing the

4.

(3,

0.

line

3* + 2j>-z +

containing the line and parallel to the

z-axis.

Find an equation of the plane through the

5.

3x-2.y--2-3 = 0,

2^ + ^ + 42+1 =

which makes equal intercepts on the

What

6.

when

7.

Show

ingtheline

is

is

x-

and

;y-axes.

the locus of the equation (ax

k ^ ?

by

cz

+ d) = k 2
2

when

+2z=
4 = 0>

that the plane x

_ 2y + 4z +

and consequently
/2 is

line

is

is

parallel to

a plane contain-

x + y + 2 _ s = Q>

parallel to the line.

8. When the planes (19.1) are parallel, for what values of t\ and
equation (19.8) an equation of a plane parallel to the planes (19.1) ?

9. Find the condition upon ti and h in equation (19.8) so that it

an equation of a plane parallel to the #-axis to the .y-axis to the
;

z-axis.
10.

(19.3),

the y-,

Show

that,

when

(19.1) are equations of a line, equations (19.2),

and (19.4) are equations of planes through the

x-, and z-axes respectively.
103

line parallel to

Lines and Planes in Space

20.

[Chap. 2

Two Homogeneous

Equations of the First Degree in

Three Unknowns

We consider next the particular case of equations (19.1) when

= d2 =
we consider the equations
that
02* + b 2 y + c 2 z = 0,
0i* + biy + ciz = 0,
(20.1)

di

is,

with the understanding that they are independent, that is,

that not all of the determinants (19.5) are equal to zero. We
say that each of equations (20.1) is homogeneous of the first
degree in the unknowns, since every term is of the first degree
in (19.2), (19.3), and
in these unknowns. On putting d\ = d 2 =

we have

(19.4),

lb2

blC2

I
.

z>

Glb2

\y =
.

\biC2

= ""
I

aiC2

>

\a\C2 I*,

where for the sake of brevity we have indicated only the

ments of the main diagonal', that is, 0i& 2 stands for
01

bi

02

b2

We

have also made use of the fact that

z
These equations are satisfied by x = y
evident

common

satisfied also

(20.3)
for

Two

|ci0 2

\a\c%

= 0,

which

But they

solution of equations (20.1).

\.

an

is

are

by

biC 2

/
1

any value of

[20.1]

ele-

t.

=-

t
\

0iC 2

/
1

0i& 2

Hence we have

independent homogeneous equations of the first degree

unknowns admit an endless number of common solu-

in three

tions; for equations (20.1) the solutions are given by (20.3),

which
(20.4)

may

be written in the

x y
:

z =
\

biC 2

with the understanding that

zero the corresponding
solutions.

from

:
|

if

0iC 2

:
1

0162

any one of the determinants

unknown

The reader should observe

just as (20.4) follows

form

is

zero for all

is

common

that (20.3) follows from (20.4),

(20.3).

104

Two Homogeneous

Sec. 20]

We

Equations of the First Degree

return to the consideration of equations (19.1)

when

one at least of the determinants (19.5) is not equal to zero, and

denote by x\ y\> z { a common solution of these equations. Then
equations (19.1) can be written in the form
9

- *i) + bi(y- yi] + Ci(z- ii) = 0,

- xi) + b 2 (y - y\) + c 2 (z - 21) = 0.
a*(x
ai(x

These equations being homogeneous of the first degree in

z it it follows from Theorem [20.1] that
Xi, y
y\, and z

x
r

90

x-xi

CN

y-yi

__

These equations being of the form

[20.2]

__ ~

z-zi ~
we have

(16.5),

Two equations (19.1), for which one at least of the determinants

is different from zero, are equations of a line with these determinants as direction numbers.

EXERCISES
1.

Find the common solutions of the equations

2.

Verify

by substitution that

equations (20.1) for

3.

Show

that the point

2^ + 3^-22 +

(1,

Find equations of the

1).
through the point (2, 1,

on the

2, 0) is

solution of

line

2^-3^-52-8 = 0,

line in the

4.

common

/.

= 0,

and derive equations of the

5.

is

(20.3)

every value of

form

line parallel

(16.5).

to the line in Ex. 3

and

Find an equation of the plane through the origin and perpen-

dicular to the line

4x-;y + 32 +
6.

Show

plane,

if

plane,

if

is it

x-y-z = Q.

= 0,

that the line with equations (20.5) is parallel to the xy0ife

0; parallel to the jyz-plane, or the xz-

and only if
and only if

6iC 2

=
= 0,
1

or

tfiC 2

parallel to the >>-axis?

105

= 0.

Under what conditions

Determinants
7.
lel

Using Theorem

[Chap. 2

+2z=

show that the plane x

[20.2],

is

paral-

to the line

x-2y + 4z + 4 = 0,
19,

(see

Ex.

x+y+z-8=Q

7).

8. Given two nonparallel lines with direction numbers u\, v it w\

are direction
and U2 V2, tv2) show that VitV2 |,
u\w 2 |,
u\Vz
numbers of any line perpendicular to the given lines.
t

Show

that if two lines with direction numbers ui, Vi, w\ and

pass through the point (x\, y\, 21), any line through this
point and contained in the plane of the two lines has as equations
9.

^2, v 2 , tv 2

x- xi
t\U\ -f /2W2

for suitable values of

z-

/2

and

both zero, these are equations of a

z\

t\W\

tiVi -f /2^2

and

/i

- y\ _

any values of

for

line in this

ti

and

/2 ,

not

plane and through

the point.
10.

Show

that

if

aix

a\u

the planes (19.1) intersect in a line the equation

b\y

Ciz

biV -f

d\

a2 x

c\w

b 2 y -f

a?.u -f b%v

is an equation of a plane through this

with direction numbers u, v, w. Discuss

line

this

and

2^
parallel to

any

line

equation when the planes

(19.1) are parallel.

21. Determinants of the Third OrJer.

Three Equations of the First Degree

in Three Unknowns

The determination
equations

of the

common

solutions,

if

any, of three

+ b\y + Ciz + di = 0,
02* + b y + c z + d = 0,
=
0i*

(21.1)

is facilitated

by the use

of determinants of the third order.

Such

represented by a square array of 9 elements as

in the left-hand member of (21.2), and is defined in terms of

a determinant

is

106

Determinants

Sec. 21]

of the

Third Order

determinants of the second order as in the right-hand member

is called an expansion of the determinant:

of (21.2), which

(21.2)

We

01

01

a2

b2

b3

C2

= 01

Ci

03
3

define the minor of

an element

a determinant of the

in

third order to be the determinant of the second order obtained

on removing from the given determinant the row and column in

which the element lies; thus in the right-hand member of
(21.2) each of the elements of the first column is multiplied by
minor.

its

a means of reducing writing, to represent

the determinant (21.2) by \aib 2 Cz\, that is, by writing only
the elements of the main diagonal of the determinant, as was
It is convenient, as

done in

20

determinants of the second order.

for

notation equation (21.2)

= 01

(21.3)
It is

In this

is

02

+ 03

evident that this abbreviated notation cannot be used

when

the elements are particular numbers for, then there would be

no means of telling what the elements not on the main diagonal
;

are.
If

we

pick out the terms in the right-hand

involving

bi, b 2 ,

and

6 3 respectively,

of writing this right-hand

member

we

member

of (21.2)

see that another

way

is

and consequently another expansion of the determinant

given by

(21.4)

in like

get the expansion

On

63

01^2

.
|

manner, we pick out the terms involving

If,

we

(21.5)

a2c3

bi

aib 2 c-3

= Ci

a 2 b3

is

Ci,

2,

-c
|

the other hand, the following expansions of the determi-

nant are obtained according as we pick out from the

107

right-

Determinants

hand member
first,

[Chap. 2

of (21.2) the terms involving the elements of the

second, or third rows

= - 02

(21.6)

+ b2 0iC 3 - c 2
- 63 0lC2 + C3

biC 3

When we examine

the six expansions of the determinant

given by (21.3), (21.4), (21.5), and (21.6), we note that each
element multiplied by its minor appears twice: once in the

expansions in terms of the elements of a column and their

minors, and once in the expansions in terms of the elements
of a row and their minors. Also, in both cases the algebraic
and we note that if the element
sign of the term is the same
;

pth row and qth column the sign is plus or minus

according as p + q is an even or odd number. For example, a%
is in the second row and first column and, 2+1 being odd,
the sign is minus, as is seen to be the case in (21.3) and the
second of (21.6). If then we define the cofactor of an element
in the pth row and qth column to be the minor of the element
p
multiplied by (
l) +, all six of the above expansions are
to
the
theorem
equivalent
is

in the

[21.1]

A determinant of the third order is equal to the sum of the

products of the elements of any column (or row) and their
respective cofactors.

As
[21.2]

corollaries of this

// all the elements of

determinant
[21.3]

theorem we have

is

any column

(or

row) are zeros, the

to zero.

equal

// all the elements of

any column

factor k, the determinant

row) have a common

k times the determinant

(or

is equal to
obtained by removing this factor from all the elements of

this

column

(or row).

Consider now the determinant

01

02

03

c2

Ca

108

Determinants of tKe Third Order

Sec. 21]

is, the determinant obtained from (21.2) by interchanging

rows and columns, without changing the relative order of
the elements in a row or column. When this determinant is
expanded in terms of the first row, we have an expression equal
to the right-hand member of (21.2), since interchanging rows
and columns in a determinant of the second order does not

that
its

change

its

value

for

example,

"2

c2

Hence we have
[21.4]

The determinant obtained from a determinant of the third

by interchanging its columns and rows without
changing the relative order of the elements in any column or
row is equal to the original determinant.
order

If this theorem did not hold, we could not use

to
a\biCi
denote the determinant (21.2), since the interchange of columns
and rows does not change the main diagonal.
We consider next the result of interchanging two columns
(or rows) of a determinant, and begin with the case of two adjacent rows, say, the />th and (p + l)th rows, where p is 1 or 2.
An element of the qth column and pth row goes into an element of the #th column and (p + l)th row. This interchange
does not affect the minor of the element but the cofactor of
|

now

the product of the minor by (

whereas in the original determinant the multiplier is (
Consequently the expansion of the new determinant in terms
the element

is

of the elements of the (p

original determinant.

l)th

The same

row

is

equal to minus the

result follows

when

there

is

an interchange of adjacent columns. The interchange of the

first and third rows (or columns) may be effected by three
interchanges of adjacent rows (or columns), thus:
123

>

132

> 312

321.

Since three changes of sign result in a change,

[21.5]

we have

The determinant obtained by interchanging two rows (or

columns) of a determinant of the third order is equal to

minus

the original determinant.

109

Determinants

As a

corollary

[Chap. 2

we have

// all the elements of one row (or column) of a determinant

are equal to the corresponding elements of another row (or
column), that is, if two rows (or columns) are identical, the

[21.6]

determinant

is

to zero.

equal

these rows (or columns) are interchanged, the

determinant is evidently the same as before; and in consequence of Theorem [21.5] it is equal to that number which is

In

equal to

We
[21.7]

if

fact,

its

are

negative, that is, zero.

able to prove the theorem

now

The sum of the products of the elements of a row (or

column) and the cofactors of the corresponding elements of
another row (or column) is equal to zero.

such a sum is an expansion of a determinant with

rows (or columns), which by Theorem [21.6] is
two
This proves the theorem.
zero.
to
equal
aid
of these properties of determinants we are able
the
With
In

fact,

identical

common

to find the

if

solutions,

any, of equations (21.1).

Assuming that they have a common solution, and letting x, y,

and z denote the common solution, we multiply equations
bic 2
respectively and add.
l&iCalt and
(21.1) by Ifecsl,
In consequence of (21.3) the coefficient of x is a^c-s and the
and from Theorem [21.7] it follows
constant term is dib 2 C3
z are equal to zero. Hence we have
of
and
that the coefficients
y
\

(21.7)

(21.4)

(21.8)

if

a\c<2

\X+\

rfl& 2 C 3

we multiply equations
\

respectively

and Theorem

(21.1)

by

get, in

2c3

1,

consequence of

[21.7],

01^3

\y

0irf2C 3

if we use the multipliers

obtain, in consequence of (21.5)

(21.9)

= 0.
I

and add, we

Again,

we

i& 2 C 3

Likewise,

and

0i& 2 c 3

z
1

2 ft 3

= 0.
1

1,

+ 0iMs = 0.
1

110

#1*3

and Theorem
I

1,

and

[21.7],

a\bz

|,

Three Equations

Sec. 21]

of the First

Degree

We observe that in (21.7), (21.8), and (21.9) x,

the same coefficient, namely, aib 2 c 3
which we
|

y,

and

have

call the deter-

minant of equations (21.1). Moreover, the second determinants

in (21.7), (21.8), and (21.9) are obtained from the determinant
of the equations

by

by replacing the

0's, &'s,

and

c's

respectively

having the same subscripts.

Equations (21.7), (21.8), and (21.9) have one and only one
<f s

solution if a\b 2 cz ^ 0. In this case the values of x, y, and z

given by these equations satisfy equations (21.1), as the reader
can verify (see 26, Ex. 8). Hence we have
|

Three equations of the

[21.8]

one and only one

first degree in three

common

unknowns have

solution if their determinant is

not equal to zero.

The geometric

equivalent of this theorem

is

Three planes have one and only one point in

[21.9]

common

if

the determinant of their equations is not equal to zero.

22 and 24 we analyze the case when the determinant of

equations (21.1) is equal to zero.
Theorem [21.8] and the processes leading up to it may be
applied to the problem of finding an equation of the plane
In

through three points

the coefficients a,

(x {

b, c,

yi, 21), (x 2 ,

and d

y2

z 2 ), (*3,

JVs,

Thus

2 3 ).

in the equation

+ by + cz + d = Q
a xi + by + cz + d =
ax 2 + by 2 + cz + d = 0,
0*3 + by 3 + cz^ + d = 0.
ax

(21.10)

must be such that

(21.11)

These equations looked upon as equations in a, b, and

one and only one common solution if the determinant
(21.12)

xi

yi

x2

y*

z2

JV3

*3

*3

have

z\

not equal to zero. If this condition is satisfied, we solve (21.11)

for a, b and c in terms of d and substitute in (21.10) to obtain
an equation of the plane (see Ex. 10).

is

Ill

Determinants

[Chap. 2

EXERCISES
For what value of a

1.

is

the determinant

36
equal to zero ? Is it always possible to choose the value of one elein a determinant, all the others being given, so that the determinant shall be equal to a given number ?

ment

2.

Show

that
1

P
2
P
3.

Show

that
'

01

4.

Show

-I-

0i

bi

02 4" 02'

b%

03 ~h 03'

bz

fli'&aCs

I-

that the determinant

01

02

C2

/202
is

equal to zero for any values of

5.

Show

t\

and

/2 .

that whatever be the constants

k\a\ H~ k%b\

-f-

^i

Ci

b$

3^1

k\

*3

6.

Show

7.

Show

^lfcC3

that
I

points

0iC2
03Ci

that equation (5.3) of a line in the plane through the

and (x 2 j 2 ) can be written in the form

(jci, ;yi)

yi

112

= 0.

Three Equations

Sec. 21]

-6

-4

2),

points
9.

4,

Show

2x

= 0,

= 0,

-2
2

show that the

that

when

y2

b\y

z 2 ),

by these points

(21.9),

and

v,

c\z

b\v -f

12.

(see

result.

different from zero

is

(*3, jVa, 23),

an equation of

is

Xi

y\

x2

y*

d\

(21.1)

x\
z

X2

-f b 2y

aw -f

c\w

have one and only one point

c2 z

b%v -f

+ d? __

azx

c^w

are equations of a line through this point

M,

the coordinates of

and check the

Xz

11. Show that if the planes

common, the equations

and

the determinant (21.12)

23

2x-y+z-2=Q

common. Find

in

of (21.7), (21.8),

the plane determined

a\u

3x-y + 2z+l=Q,

for three points (xi, y\, z\), (x 2t

0i*

-1 = 0,

that the planes

by means

Show

10.

is

have one and only one point

this point

an equation of the line through the points

2, 3), and the second an equation of the line through the
6), (2, 4). Of what is the third an equation?
first

+ 2y-z + 3 = Q,

in

Degree

Without expanding the determinants in the following equations

8.

(1,

of the First

a^u

and with

b$y

c^z -f d*

fav -f

direction

c$w

numbers

20, Ex. 10).

= a n bn

Using the notation

4- 012^21

and so

on,

show that
2fli.fi

202

t fc

V
^03ttM
l

is

V
^03i0i2
I*

"V

^03i'

equal to the product of the determinants

011

012

013

021

022

023

031

032

033

hi
hi

b 22

The
13. Show that the result of Ex. 12 may be stated as follows
product of two determinants of the third order is equal to the determinant of the third order whose element in the ith row and yth column
is the sum of the products of corresponding elements of the zth row of
the first determinant and the yth column of the second determinant.
:

Discuss the effect of interchanging the two given determinants.

113

Determinants

in

14. Show that if the planes

common, the equation

(dix

ti

+ bi y -f Ci2 + di

-f /2 (a**

[Chap. 2

have one and only one point

(21.1)

+ fe;y + c2z +

fife)

4 fejv 4- c3 z 4 </a) = 0,

4- fe (#3*

and / 3 not all zeros, is an equation of a plane

that an equation of any plane through this
and
through
point is given by the above equation for suitable values of t\, /2, and / 3
15. Show that if the planes (21.1) have one and only one point in
common, an equation of the plane through this point and parallel to
any values of

for

/i, /2,

this point,

the plane

4 by 4

04*

(i)

c4 2

(/4

is
|

(0i*

biy

CiZ

4 di) 4 0Ac3
4 aib C4
|

when the plane (i) is

when it is parallel to the

Discuss the case

also

(21.1);

(a2 x

(3#

4 b2y 4 c2 z 4 efe)
4 fejv 4 c 3 2 4 ds
)

0.

one of the planes

two of

parallel to

line of intersection of

the planes (21.1).

22. Three

Homogeneous Equations
in

We
the

of the First

Degree

Three Unknowns

consider in this section three homogeneous equations of

first

degree

+ b y + c z = 0,
03* + fay + 32 = 0.
2*

(22.1)

Since these equations are of the form (21.1) with the d's equal
to zero, it follows from Theorem [21.2] that the second de-

terminants in (21.7), (21.8), and (21.9) are equal to zero. Conis the only common solution of equasequently x = y = z =
tions (22.1) when the determinant of these equations is not
equal to zero.

We

consider

now

the case

when

016203

= 0.

any two

If

of equations (22.1) are equivalent, that is, if either

stant multiple of the other, the determinant a ibzc^
|

is

a con-

is
\

equal

by Theorems [21.3] and [21.6]. In this case the common

solutions of two nonequivalent equations are common solutions
to zero

of all three equations.

Thus,

if

the third of (22.1)

114

is

equivalent

Three Homogeneous Equations

Sec. 22]

by Theorem

to either of the other equations,

solutions of (22.1) are

= t\ bic 2

(22.2)

= -t\ 0iC2

\,

for every value of the constant

If

now no two

of the First

common

[20.1] the

1,

Degree

/
1

0i&2

/.

of equations (22.1) are equivalent, the com-

mon solutions of the first two of (22.1) are given by (22.2).

When these expressions are substituted in the left-hand member
of the third of equations (22.1),
/

(03

bid

~
|

b3

we obtain

0lC2

+ C3

01*2

The expression in parentheses

equal to 01*2^3 by (21.6).
Since by hypothesis this is equal to zero, the above expression is equal to zero for all values of /. Consequently, when
is

= 0,

the expressions (22.2) for every value of / are

solutions of equations (22.1).
In consequence of Theorem [20.1] the common solutions of

0ik>C3

common

the second and third of equations (22.1) are

(22.3)

=r

bzc-s

for every value of

(22.4)

=-s
|

for every value of

=-r
\

and of the

r,

biC's

By

s.

a 2 c3

first

r\

and third of

= s\ 0ic3

2 63

(22.1)

s\ 0i& 3

the above argument these are

common

solutions of all three of equations (22.1) when |0i&2C 3

0.
The signs in (22.2), (22.3), and (22.4) are chosen so that the
|

determinants in these expressions are the cofactors of the elements in the third, first, and second rows respectively of (21.2.)
As a consequence of (22.2), (22.3), and (22.4) we have that
when 01&2C3 = the corresponding determinants in any two
1

of the sets

t)f

tional (see
[22.1]

equations (22.2), (22.3), and (22.4) are proporAccordingly we have

21, Ex. 6).

Three homogeneous equations of the first degree in three unsolutions other than zero, if and only

knowns have common

if the

determinant of the equations

equations (22.1) the

x
(22.5)

=
=

common

115

is equal to zero; for

solutions are given by

Lines and Planes in Space

From the

foregoing discussion

[Chap. 2

follows that the three planes

it

have the origin as the only common point if the determinant of the equations, that is, \aib 2 C3\, is different from
zero, and that if their determinant is equal to zero the three
planes have in common a line through the origin. Equations
(22.1)

of such a line are

v
Z

Comparing these equations with

7
..

[22.2]

and

(22.2), (22.3),

have as the geometric equivalent of Theorem

(22.4),

we

[22.1]

Three distinct planes whose equations are (22.1) have the

origin as the only common point if the determinant of their
equations is not equal to zero; and if the determinant is
equal

planes intersect in a line through the

numbers each set of three quan-

to zero the three

origin, having as direction

in equations (22.5).

tities

As a consequence
[22.3]

When a
quantities

of

Theorems

determinant
hi, h<2 , h$,

[22.1]

aib 2 c-3

is
\

and
equal

[21.4]

we have

to zero, there exist

not all zero, such that

(22 6}

= 0,
and

also quantities k\,

kiai

(22 7)

+ k2 2 + *

k3 not

2,

such that

=
klbl + k ^ 2 +
+ * 2 C 2 + *3C 8 = 0.

3 ^3

*lCi

all zero,

'

k 3b 3
'

'

'

In fact, the A's are proportional to each set of three quantities

in equations (22.5). In like manner the 's are proportional to
the cofactors of the elements in each column of aib2 Cz \.
|

Theorem
tions

may

[22.1] is very important, and

strike the reader as surprising.

some of its applicaFor example, if we

seek an equation of the line in the plane through the points

PI(XI, y\) and P 2 (* 2 ^2), we have for consideration three
,

equations
<*x

(22 8)

+ by + c = Q a*i + by\ + c = 0,
0*2 + by + c = 0,
9

116

Three Homogeneous Equations

Sec. 22]

the

first

of which

of the First

Degree

an equation of the line whose coefficients

and third hold. Applyto the last two of equations (22.8), we have

is

are to be determined so that the second

Theorem

ing

[20.1]

y\

On

substituting these values in the

common factor /, we obtain

first

of (22.8)

and dividing

out the
(22.9)

X2

However,

if

we

y*

consider (22.8) as three homogeneous equations

ft, and c, the coefficients being x, y, 1 ;

of the first degree in a,

xi, y\,

x2

y2, 1, it follows

from Theorem

[22.1] that, in

order that these equations shall have a solution other than

a= b= c=
(in which case we have no equation of the line),

we must have
(22.10)

This is the same equation as (22.9), as is seen on expanding

the determinant (22.10) in terms of the elements of the first
row. Consequently we have found an equation of the line by
means of Theorem [22.1], without finding directly the expressions for 0, &, and c (see
21, Ex. 7).
This is an interesting and subtle process, which may be
applied to some of the exercises below, and which is used in
the next section. Accordingly it is important that the reader
think

it

through so that he will have confidence in using

it.

EXERCISES
1.

Find the

coefficient

a in the

so that these equations shall have

find these solutions.

first

of the equations

common

solutions other than zeros,

and

2.

Show

triangle,

if

that three points (x, y), (xi, yi), (x2 y*) are vertices of a
if the determinant in (22.10) is not zero.
Express
,

and only

the result of

13,

Ex. 26 by means of a determinant.

117

Lines and Planes in Space

[Chap. 2

Show

3.

vanish

is

that the condition that the determinant in Ex. 2 shall

4 that the three
equivalent to the conditions of Ex. 9 of

points shall be collinear.

4.

Show

Qix

meet

that three distinct nonparallel lines

+ biy -f n = 0,

in a point,

if

and only

Compare

this result

Ex. 4 of

21.

5.

points

Show
(x\

02* 4- fay -f c 2

= 0,

a**

+ b*y + c3 =

if

0i

bi

ci

02

fe

C2

03

fe

= 0.

with the discussion of equation (10.4) and with

that an equation of the plane through the origin and the

yi, z\)

and fa,

2 2 ) is

;y 2 ,

=0.
6.

Show

that three points fa, y\,

a plane through the

origin,

if

21),

and only
yi

fa,

22),

fa,

y^

23) lie in

if

21

=0.
23
7.

Show

when

the three

Interpret this result

and Ex. 6

that the condition of Ex. 6

points are collinear (see

15,

Ex.

10).

satisfied

is

geometrically.
8.

For what values of

do the equations

admit solutions other than zeros ? Find the solutions.

9. Show that if MI, v\, w\ ; w 2 v2 tv 2 ; w 3 v3 w 3 are direction numbers of three lines through a point P, the lines lie in a plane, if and
only if the determinant
Vl
Ul
Wl
,

is

equal to zero (see Theorem

10.

Given three

lines

U2

t/2

M3

Va

M/2

[17.7]).

through a point P, and not in the same plane,

line through P are expressible

show that direction numbers of any

linearly and homogeneously in terms
given three

lines.

118

of the direction

numbers of the

Planes Determined by Geometric Conditions

Sec. 23]

23. Equations of Planes Determined by Certain Geometric

Conditions. Shortest Distance between Two Lines
If

we

desire to find

an equation of the plane determined by

three noncollinear points (xi, y it zi), (x 2 y 2 z 2 ), (* 3 jVa,

we may substitute these values for x, y, z in the equation
,

ax

(23.1)

23),

+ by + cz + d = 0,

solve the resulting equations for 0, b, and c in terms of d by

the method of
21, substitute the values of a, ft, and c so ob-

tained in (23.1), and get an equation of the plane, as in Ex. 10

21
or we may proceed as follows.

of

If

we

substitute xi, y\,

z\ for x,

the resulting equation from (23.1),

a(x

(23.2)

and subtract

z in (23.1)

we obtain

- *i) + b(y - y{) + c(z - z,) = 0.

Expressing the conditions that the points (* 2 y 2

(#3, Js, 2 3 ) are points of the plane (23.2), we have
,

and

^ 2 ~~ *^ + 6 ^ "" yi) + C ^ ~~ Zl) = 0>

~ Xl + b(y - yi) + c(z - zi) = 0.
a(x 3
a

(23 3^;
(

z2 )

'

2
3

and

(23.3) as homogeneous
Looking upon equations (23.2)
in
with Theorem
accordance
in
we
have
and
c,
a, &,
equations
solution
other than
a
common
admit
that
these
[22.1]
equations
zero
if
is
to
that is,
and
if
their
determinant
zeros,
equal
only
;

xi

y\

= 0.

(23.4)

When

this

determinant

is

expanded

in

terms of the elements of

seen to be an equation of the first degree in

and
x, y, and z, and consequently is an equation of a plane
it is an equation of the plane determined by the three points,
since the determinant is equal to zero when x, y, z are replaced
by the coordinates of each of the three points, in accordance
the

first

row,

it is

with Theorems [21.2] and

[21.6].

Another way of obtaining equation (23.4) is to solve equations

(23.3) for a, b, c in accordance with Theorem [20.1] and to substitute the result in (23.2), as the reader should verify. But the
leads to the result more immediately.

method we have used

119

Lines and Planes in Space

[Chap. 2

Equation (23.4) is an equation in x y, and z unless the

minors of (x
>>i), and (z
z\) are all equal to zero.
*i), (y
If they are equal to zero, the last two rows in (23.4) are pro9

portional, that

is,

*2

which

is

- xi

y2

- y\

the condition that the three points shall be collinear.

is an endless number of planes through a line, we

Since there

would not expect equation (23.4) to apply to this case. However,

since equations (23.3) are equivalent in this case, if we take
any values of a, b, and c satisfying either of equations (23.3)

and substitute these values in (23.2), we get an equation of a

plane through the three collinear points.
Accordingly we have
[23.1]

Equation (23.4)

is

an equation of

the

plane determined by

three noncollinear points (x\, y\, Zi), (x 2

When

y2

22), (*3, yz, 23).

the points are collinear, equation (23.2) for values of

satisfying either of equations (23.3) is an

equation of one of the endless number of planes containing
the points.
a,

We
(

6,

and

consider next two lines with the equations

23.5)

MI
(

23.6)

U2

V2

W2

and seek an equation of the plane through the

first line

and

Since the plane passes through the point

an equation of the plane is of the form

parallel to the second.

(*it y\, 21),

(23.7)

a(x

- *i) +

b(y

- yi) + c(z - zi) = 0.

If the line (23.5) is to be in this plane, and the line (23.6) is

to be parallel to the plane, any normal to the plane must be
perpendicular to both lines consequently a, b, and c, direction
;

numbers of such a normal, must satisfy the conditions

au + *i + cwi = 0,
i

au 2

+ bv 2 + cw<z =5 0,
120

Shortest Distance between

Sec. 23]

Two

Lines

by Theorems [16.8] and [17.5]. Looking upon these equations

and (23.7) as homogeneous equations in a, b, and c, we have that
these equations admit solutions not all zeros, if and only if

-xi y-yi z-z

w
HI
vi

(23.9)

= 0,

which

is

the equation sought.

When two

lines do not intersect and are not parallel, by

constructing a plane through each line parallel to the other we
have the lines lying in two parallel planes, like the ceiling and
floor of

The

a room.

between the

shortest distance

lines is the length of the

segment they determine on their common perpendicular that

is, in the above analogy it is the length of the normal to the
floor and ceiling which meets the two lines. This length is
the distance between the two planes. If then we wish to find
the shortest distance between the lines (23.5) and (23.6), and
we observe that (23.9) is an equation of the plane through the
;

line (23.5) parallel to the line (23.6), it is clear that this shortest distance is the distance of any point on the line (23.6) from

the plane (23.9), and in particular the distance of the point

Theorem [18.1]
by substituting
(23.9) and divid-

(* 2 , ^2, z 2 ) from this plane. In accordance with

this value of the directed distance is obtained

the left-hand member of

the
ing by
square root, with appropriate sign, of the sum of
the squares of the coefficients of x, y, and z in equation (23.9).
*2, ^2, 22 for x, y, z in

Hence we have
The directed

[23.2]

shortest distance

line (23.6) is given

Mi

(23.10)

D=

where
e

being

tive,

+1

and

or

from

the line (23.5) to the

by

v\w 2

Vi

2
|

WiU 2

2
1

1 according as \u\V2\ is positive or nega-

so on, as in [18.1].

121

Lines and Planes in Space

[Chap. 2

EXERCISES
Find an equation of the plane through the points (h, 1, 1),
3). For what values of h and k is there more than
one plane through these points ? Find an equation of one of the planes.
1.

(k, 2,

2.

- 1), and (1,3,

Under what condition is equation (23.9) satisfied by any values

and z ? What does this mean geometrically ?

of x, y,
3.

Show

that an equation of a plane containing the line (23.5)

and perpendicular to the plane

ax

(i)

is

Discuss the case

4.

xi

by

+ cz + d =

yi

zi

MI

Vi

w\

when the

line (23.5) is

= 0.

normal to the plane

Find the shortest distance between the

x

+l~
-Z-2

l__y
~~
2

-2

+ 2 _y _ * -

(i).

lines

3~~1~"5*

'

5. Show that an equation of the plane through the point

and normal to the line

aix

b\y

+ c\z +

di

is

6.

Show

and only

= 0,
r

a2 x
yi

-f b 2y

x2

Ci

b2

C2

y2

y\

y\ z\)
t

+ c?z + d2 =

z\

bi

= 0.

that the lines (23.5) and (23.6)

if

(xi,

lie in

the

same

plane,

if

z2

z\

= 0.
U2
7.

and

Show

that an equation of the plane through the point (x\, y\, z\)
with direction numbers u\ v\ w\ and u 2t v2 w2 is

parallel to lines

8.

is

W2

1/2

Show

- xi y-yi z-zi
MI

Vi

w\

M2

V2

W2

that for the points P\(x\, y\,

z\),

= 0.
P2 (x 2t y2

Xi

X2

#3

X4

Xs

yi

y*

yz

y*

y?>Q

z\

z2

23

Z4

an equation of a plane through Pi

122

z 2 ),

and so on

ZB

parallel to

P2Pa and to

P4 P6

The

Sec. 24]

Configurations of Three Planes

9. Find the coordinates of the point or points common to the

two planes 2 x + 3;y + z + 3 = 0, * - 2;y - 3 z - 2 = 0, and each of

the planes

= Q;

What

is

10.

+ 5y + 4 z + 6 = 0.

the geometric relation between each set of three planes?

When

u\

t/i,

tion cosines, how

quence of (15.10)?
11.

(d)

w\ and u 2

may

v 2 , tV2 in (23.5)

and

(23.6), are direc-

the expression (23.10) be written in conse-

Show by means

ing the line (23.5)

of Ex. 3 that an equation of the plane containand perpendicular to the plane (23.9) is

xi

ViW 2

- y\

vi

u\
I

zi

= 0.

w\

WlU 2

UiV 2

Show

that this equation, and the one obtained from it on replacing

t/i, w\
by x2 yi> z2 u2 v2 w2 respectively in the first
two rows, are equations of the common perpendicular to the lines

Xi, y\, z\\ u\,

(23.5)

and

(23.6).

What

equations replace (23.9) and the equation of Ex. 11

the lines (23.5) and (23.6) are parallel?
12.

24.

Tne Configurations
we

In this section

when

Three Planes

of

return to the consideration of three

+ biy + ClZ + ^ = 0,
02* + b 2 y + c^z + d 2 = 0,
03* + &3JV + c$z + dz =

equations

aiX

(24.1)

and seek the conditions upon the

coefficients in these equations

corresponding to the various types of configurations of three
21 it was shown that a common solution or soluplanes. In
tions, if any, of these equations satisfy the equations

= 0,
01&2C3 \y+\ #1^3 = 0,
\aibzC3\z + \aib2d 3 = 0;

(24.2)

tfifeca

dib 2 cz

and from these equations followed Theorem [21.9], that three

planes with equations (24.1) have one and only one point
123

Lines and Planes in Space

[Chap. 2

common

if the determinant of these equations, namely,

not equal to zero.
It is evident geometrically that there are the following configurations formed by three planes, other than that of planes having
one and only one point in common (1) two of the planes are
(2) the
parallel, and the third intersects them in parallel lines
three planes are parallel
(3) the three planes have a line in
common, or intersect in three parallel lines forming a triangular
prism. In accordance with Theorem [21.9] these cases must
correspond to the various ways in which the equation

in
|

a\biCz

is

0i

(24.3)

is satisfied.

We

and observe

first

mon

02

02

C2

03

b%

shall consider these various

of

all

that

if

ways

successively,

(24.3) is satisfied there

solution of equations (24.1) unless

all

is

no com-

the determinants

in (24.2) are equal to zero.

Case 1. The condition (24.3) is satisfied when any two

rows are proportional. If the first two rows are proportional,
that

if

is,

0i&2

=
1

biC 2

=
1

01^2

= 0,

by Theorem

[19.4] the

two planes are

parallel or coincident according as di/d 2

that is,
is not or is equal to the ratio of the other coefficients
as one can show, when the minors of the elements
3
fc,

first

and

the determinants (24.4) are not or are equal to zero ;

there
are six of these minors, only three are distinct.
although
If the third row is not proportional to the other two, the third
plane meets the other two planes in two parallel lines or in
c 3 in

two coincident
allel

lines

or coincident.

common

solution or

according as the first two planes are parIn these cases equations (24.1) have no

an endless number of

common

solutions

respectively.

Case

rows in (24.3) are proportional, that

the elements in (24.3) are equal to zero,
the three planes are parallel, or two are coincident and parallel
to the third, or all three are coincident. These cases are disis, if

2.

If all three

the minors of

all

tinguished from one another

by the values
124

of the ratios of the

Tne

Sec. 24]

d's.

Thus,

are not

if

Configurations of Three Planes

the minors of
the

all zero,

first

i n the determinants
(24.4)
(24.1) are parallel (and not

3 , fe, c 3

two planes

coincident), and similarly for other pairs of planes. Only in

case all three planes are coincident do equations (24.1) have

common
Case

solutions.

We

3.

consider finally the case

when no two rows

are

proportional. From Theorem [22.3] we have equations (22.7),

in which all the &'s are different from zero, since otherwise one

two rows are proportional. If we

and c 3 and put
ki/kz = /i,

finds that

(22.7) for 03, #3,

solve equations
k 2 /kz
t2
we

have
(24.5)

= tldi +

b3

t 2 Ct 2 ,

= tibi +

t2

b2

C3

= tiCi +

t2C2 .

When

these expressions are substituted in the determinants

(24.4), the latter reduce respectively to

k 6iC 2

(24.6)

-k\aiC 2

|,

where the common factor k

\,

given by

is

k = d* - Mi -

(24.7)

k\aib 2

\,

t2

d2

We

consider first the case when the three quantities (24.6)

are equal to zero, that is, when the determinants (24.4) are
equal to zero. Since by hypothesis the first two rows of (24.3)
that is,
are not proportional, we must have k =
;

d-3

(24.8)

From

this result

and

(24.5)

tidi

d2

we have

+
Hence by Theorem
common.

t2

t2

(a 2 x

+ b y + c z + d2
2

[19.6] the three planes (24.1)

have a

).

line

in

When
zero,
(24.9)

7* 0,

from
3*

that

(24.5),

is,

and

when

the determinants (24.4) are not

(24.7) solved for d 3

we have

+ fay + c 3 z + d3 = h (a\x + b\y + dz + d{)

+ (a 2 x + b 2y + c z + d2 + k,
t2

from which it is seen that the three equations (24.1) do not

have a common solution for, if there were a common solution,
;

125

Lines and Planes in Space

[Chap. 2

the expression on the left in (24.9) and the expressions in parentheses would be equal to zero for the values of x, y and z of
t

this solution,

which

is

impossible for k

means that the three planes

that

0.

The

intersect in three parallel lines;

results of the foregoing discussion

follows

Geometrically this

they form a triangular prism (see Ex.

is,

[24.1]

2).

may

be stated as

When
planes

the determinant (24.3) of equations (24.1) of three

is

equal

(1) If the
zero,

to zero,

minors of

the planes

it

follows that :

the elements of

corresponding

any row are equal

to the other

to

two rows are

coincident or parallel according as the minors of the elements

of the corresponding rows in the determinants (24.4) are

equal to zero or not; if only two of the planes are parallel

and all three are distinct, the third plane intersects the other

two in parallel lines.

(2) If the minors of
to

the

zero,
equal
coincident and parallel

(3) If the

elements of all the rows are

or two are

all the

three planes are parallel,

to the third, or all three

minors of

are coincident.

all the elements in the

determinant

(24.3) are not all zero, the planes meet one another in one
line or in three parallel lines according as all the determi-

nants (24.4) are equal

We
[24.2]

to zero

or not.

have also the algebraic theorem

Three equations (24.1), not all of which are equivalent and
whose determinant is equal to zero, have an endless number
of common solutions or none according as all three determinants (24.4) are equal to zero or not; in the former
case either two of the equations are equivalent or any one is
a linear combination of the other two, that is, a sum of constant multiples of the other two.

Theorems [21.8] and [24.2] constitute a complete statement about the common solutions of three nonhomogeneous
equations of the first degree in three unknowns, and Theorems
[21.9] and [24.1] give a geometric picture of this algebraic
problem.

126

Miscellaneous Exercises

Sec. 25]

EXERCISES
1.

Discuss Ex. 9 of

2.

Show
t\(a\%

is

23 in the light of the above analysis.

that for any nonzero values of

biy

an equation of a plane

(See
3.

Ci2 -f d\)

t2

(a2 x

b2 y

/i,

/2 ,

c2 z

and

/3

d2 )

the equation

+k=

parallel to the line

0i*

+ hy + Ciz + di = 0,

19,

Ex.

2*

b2 y

c2 2

</2

= 0.

7.)

Using Theorem

[20.2],

show that when

(24.1) are equations of

three distinct planes and equation (24.3) is satisfied, the cofactors

of any row in the determinant of the equations are direction numbers
of the line, or lines, of intersection of the planes
are not all zero.
4.

ax

For what values of

+ 2y + 3z-l=Q,

a, b,

3* +

and

c in

cofactors

the equations

+ 2 + 2 = 0,

fry

when such

llx

+ 8>> + C2-3 =

are the planes with these equations mutually perpendicular? For

what values do the planes meet in one line? For what values do they
meet in three parallel lines ?

25. Miscellaneous Exercises.

1.

(2,

2.

Find an equation of the

1,

through the point

Find an equation of the plane through the origin normal to

3 x

_^ + 42 + 5 =: o,

Prove that the

x

meets the

4.

set of all planes

3) parallel to the line

the line
S.

The Sphere

line

+ 2>>-2 + 3 = 0,

3x-

line

Find the point where the

plane x

+ y + 2 = 0. How

line

= ^= =

far is this point

127

^- meets the

from the point

(3, 4,

5) ?

Lines and Planes in Space

Show

5.

that

What

6.

[Chap. 2

is

any number n

(P

</)(?

r )(r

- P)(P + Q + r).

sum

the locus of a point the

is a constant?

of

whose distances from

of planes

7. Show that the projection (see

14) of a line segment of length /
upon a line L is equal to / cos 0, where 6 is the angle of the line L and
where
the segment and that the projection upon a plane is / sin
is the angle of the segment and a normal to the plane.
;

Find the projection of the

8.

(2,

</>,

5, 1)

and

upon the plane

Show

9.

and 2-axes

(4,

5)

1,

segment between the points

line

upon the

<f>

line

-=~

= ~=

also

2x-y + 2z = Q.

that

k are the intercepts of a plane on the x-, y-,

is the distance of the origin from the

if g, h,

and p

respectively,

plane,

I^1
.

A2

" ."
1

1 = 12
k2

'

10. Given a fixed point P on a line in space through the origin and
equally inclined to the three coordinate axes, show that for every
plane through P meeting the three axes the sum of the reciprocals

same

of the intercepts has the

value.

Find the condition that the three

11.

_*.

Q\

2-

= A.

b\

Ci

lines

XL

2.

#2

^2

'

y.

jL

03

b$

'

shall lie in a plane.

Show

12.

lines

that the bisectors of the angles between perpendicular

through the origin and with direction cosines

X2, M2,

Xi,

AH

v\

and

respectively have equations

v<i

13.

Find an equation of the sphere of radius

point (x
14.

Show

+ y2 +

22

2/x

an equation of a sphere.

(See

and center

at the

that

x2

(i)
is

yo, Z Q ).

+2

gy

What

12.)

128

+2
are

fe 4- e

the center and radius?

Miscellaneous Exercises.

Sec. 25]

Show

15.

that the points

common

The Sphere

two intersecting spheres

to

lie in

a plane which is perpendicular to the line through the centers of the

spheres. This plane is called the radical plane. Discuss this question
when the spheres do not intersect (see 12).

Show

16.

that the square of the length of any tangent to a sphere

(i) of Ex. 14 from a point (x\
y\ z\) outside the
2
2
2 fx\ + 2gyi + 2hzi
e (see
-f *i
equal to xi +
12).

with an equation
sphere

is

Find an equation of the plane normal to the radius of the

2
2
2
sphere x + y + z -2x + 4y + 2z + 2 = Q through the point
2, 1) of the sphere. This plane is called the tangent plane to the
(1,
17.

sphere at this point.

18. Find an equation of the sphere which passes through the origin
and the points (1, - 2, 3), (2, 0, - 1), (4, 4, 0) of the sphere when
;

the last point

all

replaced

by

(3, 2,

5).

and S2 = be equations of two spheres in the form

Discuss the equation tiSi + / 2 S 2 = when t\ and fe take
fe.
values, not both zero, and in particular the case t\

19.
(i)

is

Let

Si

of Ex. 14.

20. Find an equation of the sphere inscribed in the tetrahedron

whose faces are the coordinate planes and the plane x
2y + 2z 4.
21.

origin
22.

Find the locus of a point the square of whose distance from the
2 y + 2 z = Q.
equal to its distance from the plane x

is

Find the locus of a point the sum of the squares of whose

- 1) and (2, 1, - 3) is 10.
(1, 0,

dis-

tances from the points

23.

What

is

distances from

the locus of a point the

any number n

of points

sum
is

of the squares of whose

a constant ?

24. Find an equation of the right circular cylinder of radius

for axis.
with the line x - xi = 0, y - y\ =
25.

Show

that ax*

have the same

sign, is

and

+ cz 2 = 0,

where a, b, and c do not all

an equation of a cone with vertex at the origin,

-f

by

by showing that if P\(x\, y\, z\) is a point of the locus, so also is every
point on the line joining the origin and Pi. Could the locus be one
or more planes?
26.

to

its

Find the locus of a point whose distance from the 2-axis is equal
distance from the ry-plane. For what part of the locus are the

directed distances equal ?

129

Determinants
27.

twice
28.

Find the locus of a point whose distance from the origin

distance from the ry-plane.

is

its

What

where f(x,

z)

the character of the locus with the equation f(x, z) =

denotes an expression in x and z of the locus f(y) =

is

21

we

0,
?

Any Order

26. Determinants of

In

[Chap. 2

defined determinants of the third order in terms

first column and their minors, these being

determinants of the second order, and derived various theorems
concerning determinants of the third order. In this section we
define determinants of the fourth and higher orders, and show
21 apply equally well to these deterthat the theorems of
minants.
We begin with a determinant of the fourth order, represented
by a square array of 16 elements and defined as follows in
terms of determinants of the third order

of the elements of the

(26.1)

0i

bi

Ci

di

#2

b'2

C'2

do

03

b'3

di

04

c^

<L\

b'2

C'2

b'3

-02
64

+ 03

d'2

-04

b'2

the determinants in the right-hand member of

terms of the elements of the first columns, and collect
the terms in b\, 62, 63, and 6 4 we have, using the main diagonal
to represent a determinant,
If

we expand

(26.1) in

(26.2)

04
04

02

64

130

Cid'3

+ 03

Any Order

Determinants of

Sec. 26]

This is an expansion of the determinant in terms of the

ments of the second column and their minors.

ele-

we expand the determinants of the third order in (26.1)

terms of the elements of the second columns, and again in
terms of the elements of the third columns, and proceed as
above, we obtain
If

in

(26.3)

C/2

In the foregoing expansions it is seen that the element of

the pih row and qth column, for any values of p and q from 1
p+ times the minor of the
to 4, is multiplied by (
element,
l)
which by definition is the determinant of the third order obtained on removing from the determinant (26.1) the row and
column in which the element lies. For example, b-s is in the
3+2
third row and second column, in which case (
!)?+ = (
1)

1, and we see that this checks with (26.2).

As in 21, we define the cofactor of the element in the pth
row and qth column to be ( l) p+q times the minor of the element. Accordingly, although the determinant was defined as
the sum of the products of the elements of the first column and
their respective cofactors, it is shown by (26.2) and (26.3)
that the determinant is equal to the sum of the products of the
elements of any column and their respective cofactors.
Since, as we have just seen, the terms involving any element

consist of the products of this element and its cofactor, it follows also, as in 21, that the determinant is equal to the sum

of the products of the elements of

cofactors (see Theorem [21.1]).

any row and

their respective

Just as determinants of the third and fourth orders have

been defined to be the sum of the products of the elements of
the first column and their respective cofactors, generalizing the
notation and terminology, we define a determinant of any order
to be the

and

their

sum

of the products of the elements of the first

respective

cofactors.

Just as

131

column

we have shown

that

Determinants

[Chap. 2

follows from the definition of determinants of the third

and fourth orders that the following theorem holds, so by proceeding step by step with determinants of the fifth order, and
so on, we can establish the following theorem for determinants
it

of

any order

[26.1]

determinant of any order is equal to the sum of the products of the elements of any row (or column) and their
respective coj'actors.

We consider in connection with the determinant (26.1) the

determinant obtained from (26.1) by interchanging rows and
columns without changing the relative order of the rows and
columns, that is, the determinant
a\

a<2

#3

#4

b\

62

#3

64

Ci

C2

C3

C4

d\

d2

d't

(26.4)

When
the

this

first

determinant

row, we have

is

expanded

d*
in

terms of the elements of

in place of the right-hand

member

of

(26.1) an expression obtained from the latter when the rows

and columns in each of the four determinants of the third order

are interchanged. But by Theorem [21.4] these determinants

of the third order are equal respectively to the determinants
from which they were obtained by the interchange. Hence the

determinant (26.4)

is

equal to the determinant (26.1). Proresult holds for a

ceeding step by step, we can show that this

determinant of any order. Hence we have
[26.2]

The determinant obtained from a given determinant by inits rows and columns without changing the
relative position of the elements in the rows and columns is
terchanging

equal to the given determinant.

Consider now the effect of interchanging two adjacent colof a determinant. An element in the new determinant
is in the same row as originally, but the number of its column
is one less, or one greater, than in the given determinant. Con-

umns

sequently,

if (

l)

is

the multiplier of

132

its

minor yielding

its

Determinants of

Sec. 26J

Any Order

cofactor in the original determinant, the multiplier in the new

determinant is (_ l)*+-i or (- l) p + +1
Since (- l)p+-i
+1 it follows that in either
case the
(- l)p+*= (- l)p+
sign of the cofactor is changed. If then we expand the two
.

=-

determinants in terms of the same elements, we have that the

1 times the original determinant.
new determinant is
Consider next the determinant resulting from a given determinant by the interchange of any two nonadjacent columns,
say the rth and the 5th, where s > r. The elements of the 5th
column can be brought into the rth column by 5
r interchanges of adjacent columns. This leaves the elements of the
original rth

column

in the

(r+l)th column, and

then,

by

interchanges of adjacent columns, these elements can

be brought into the 5th column. Since this interchange of the
rth and 5th columns can be accomplished by 2 (s
1 interr)
5

changes of adjacent columns, and since each such interchange

1 as a multiplier, the result is to multiply the
introduces
1 raised to the odd power 2(s
1
original determinant by
r)
1.
that is, to multiply the original determinant by
Since similar results are obtained when two rows are inter;

changed,
[26.3]

we have

the theorem

The determinant obtained by interchanging two rows (or

columns) of a determinant is equal to minus the original
determinant.

As a
[26.4]

corollary

When
tical,

we have

two rows (or columns] of a determinant are idendeterminant is equal to zero.

the

on interchanging the two rows (or columns) the sign is

changed, but we have the same determinant over again and
zero is the only number which is equal to its negative.
From Theorems [26.1] and [26.4J we have
For,

[26.5]

The sum of

the products of the elements of any row (or

and the cofactors of the corresponddeterminant
a
column) of
row
another
elements
(or column) is equal to zero.
ing
of
133

Determinants

[Chap. 2

sum of such products is an expansion of a determinant

with two identical rows (or columns).
Since each term in the expansion of a determinant contains
one element, and only one, from each column and each row,
it follows that
For, the

The multiplication of each element of a row

[26.6]

of a determinant by a constant k

is

(or

column)

equivalent to the multi-

plication of the determinant by k.

Accordingly,

same

factor

if all

k,

the elements of a row (or column) have the

is equal to k times the determi-

the determinant

nant which results on removing this factor from each element

of this row (or column).
As a consequence of Theorems [26.4] and [26.6] we have
[26.7]

When the elements of two rows (or columns) of a determinant are proportional, the determinant is equal to zero.

In consequence of Theorem [26.1]

[26.8]

we have

(see

21, Ex. 3)

// each element of any row (or column) of a determinant

expressed as the sum of two quantities, the determinant
may be written as the sum of two determinants.
is

If one wishes to write a determinant of any order n, it is

convenient to designate an element by one letter having two
subscripts, for example, by a ljt where i denotes the row and j
the column in which the element occurs. In this notation from

Theorem

[26.8]

we have

012022^33

Since the second of these determinants, having two columns

identical, is equal to zero, we have by similar procedure applied
to any two rows (or columns) the following theorem
:

134

Determinants of

Sec. 26]

[26.9]

Any Order

elements of any row (or column) of a determinant

added equal multiples of the corresponding elements
of any other row (or column), the determinant is unaltered.

//

to the

there be

is frequently used to replace a determinant by an

one
with one or more zero elements, an operation
equivalent
which reduces the calculation involved in evaluating the determinant. For example, consider the following determinant and
the equivalent one obtained by multiplying the second row by
2, subtracting the result from the first and fourth rows, and

This result

adding the result to the third row

-4

= -1 -4

-4

-4

determinant we subtract the

If in the last third-order

from the second and third rows, the

8-41
-3

12

-4
-4

-100

result

= -3 -12
-1

first

row

is

= -15.

EXERCISES
1.

Evaluate the following determinants

1231-2
21321
01221
-2

-1 -1
2

2.

For what value of a

is

a
7
3
4

the determinant

5
2

-2 -3

4-1
724

equal to zero ?

135

3-1-1

Determinants
3.

Show with

determinants

6.

is

the aid of Theorem [26.9] that each of the following

equal to zero
:

Show by means

Find the

Theorems

of

the determinant (26.1)

6.

[Chap. 2

is

[26.6]

and

[26.9] that

if

d4

equal to
I

010*4

020*4

0304

fad*

cid

M4

C2 0*4

^30*4

ratio of the determinants

0i/g

big

cigk

a2 fh

b2 h

c 2 hk

a\g

a2 k

c3fh
a3 h

big

b2 k

b$h

c2fk

7. Show that a determinant is equal to an algebraic sum of all

terms each of which consists of the product of one element, and only
one, from each row and each column, and that every such product is
a term of the sum.
8. Show that, if the determinant
of equations (21.1) is
aib2 c3
not equal to zero, and we multiply any one of these equations by
a\b2 cz and substitute in this equation the values of x, y, and z from
(21.7), (21.8), and (21.9), the result is expressible as a determinant of
the fourth order with two identical rows. Does this prove that these
|

values of

x, y,

and

2 constitute the solution of

equations (21.1)

ijXiXj ~p

where A

ai\a^a^

=
\

l}

|,

and A tJ

136

is

the cofactor of a l} in A.

Sec. 27]

10.

Unknowns

Equations of the First Degree in Four

Show

that the determinant in Ex. 9

011 4" #1

012

equal to

is

'

021 4~ #2*1

022 4~ #2

031 4- #3*1

032 -f #3*2

and that consequently the

023 4~

0;

2
033 4" #3

latter

determinant

is

equal to

1,2,3

z, AijX\Xj

=a

4~

A.

11.

Show

that

12.

Show

that the result of Ex. 10 holds for

is

any positive

if

in

Ex. 9 a lt

]V

for all values of i

i,

and

/,

then

!>..., n where
,

integer.

13. Show that the rule for the multiplication of two determinants
of the third order stated in Ex. 13 of 21 applies to two determinants
of any order, both determinants being of the same order.

27. Solution of Equations of the First Degree in Any

Number of Unknowns. Space of Four Dimensions

Consider the four equations

0i* 4- biy 4-

dz 4- d\w 4-

e\

= 0,
0,

= 0,
=
4
0,
3

4-

by 4-

42

+ dw 4-

with the understanding that not all the coefficients of x, y z,

and w in any equation are equal to zero. These equations
may or may not have a common solution. We assume that
they have at least one common solution and that x, y, z, w
in equations (27.1) denotes a common solution. We multiply
t

equations (27.1) by the cofactors of a\ a 2 03, and 4 in the

determinant
a\b^c^d^
respectively and add the resulting
equations. In this sum the coefficient of x is a\b 2 czd^ |, the
9

determinant of equations (27.1), in consequence of Theorem

[26.1], and the coefficients of y, z, and w are equal to zero, in

137

Determinants
consequence of Theorem
(27.2)

Hence we have the equation

[26.5].

0162^3^4

-f-

[Chap. 2

= 0,

#i 62^3^4

a\b<zd on
with the same subscript.
If in like manner we multiply equations (27.1) respectively by the cofactors of b\, b 2 6 3 and 64 in the determinant
and add the results, we obtain the equation
0i&2C3</4

where e\b^zd\
|

is

the determinant obtained from

by an

replacing each

(27.3)

Similarly

0162^3^4

= 0.

0102^3^4

d\b^^d\

01&2C304

we have
|

(27.4)

-\-

= 0,
= 0.

not equal to zero, these equaMoreover,

this solution is a solution of equations (27.1). In fact, if we
multiply the left-hand member of the first of (27.1) by
If the

tions

determinant

a^c^d^

is

have one and only one common solution.

and substitute from

(27.2), (27.3),

and

(27.4),

we have

0i

When

this

is

rewritten in the form

+ ei

-d,
it is

seen to be the expansion in terms of the elements of the

first

row of the determinant,

Ci

02

C2

03

04
first two rows identical, and consequently is equal to
Since similar results follow for the other equations (27.1),
we have that equations (27.1) have one and only one common
solution when their determinant
a^Czdt is not equal to

with the

zero.

138

Equations of the First Degree in Four

Sec. 27]

Unknowns

zero. Evidently this process may be applied to any number n

of equations of the first degree in n unknowns. As a consequence
we have the theorem
[27.1]

n equations of the first degree in n unknowns have one

and only one common solution when the determinant of
the equations is not equal to zero.

When

the determinant

equal to zero, there is

(27.1)
any one of the second
determinants in equations (27.2), (27.3), and (27.4) is not equal
to zero. A similar statement applies to n equations in n unknowns whose determinant is equal to zero.
We consider next the case when the equations are homogeneous in the unknowns, that is, when all the e's in equa|

aibzCzd*

is

no common solution of equations

if

We write them thus

=
ajc + bj + c>z + d,w =
(i

tions (27.1) are equal to zero.

(27.5)

1, 2, 3, 4).

In this case all the second determinants in (27.2), (27.3), and

and, in consequence of Theorem [27.1],
(27.4) are zero;
=
=
z
w
x
is the only common solution of equations
y=.
of these equations, that is,
determinant
(27.5) when the
I

is not equal to zero.

|,
In order to consider the case when the determinant aibtfsd^
equal to zero, we denote by A\ the cofactor of a\ in this

aibzCzdi

is

determinant, by Cs the cofactor of

for

for

[26.5].

1,

2,

and 3 are

equations (27.5) and (27.6) as i

and 4, we see that a common solution
which afaczd* = is given by

1, 2, 3,

of equations (27.5) for

(27.7)

a^czd* is equal to zero,

three equations (27.6)
The
[26.1].
identities in consequence of Theorem

Theorem

When we compare

takes the values

for

= 4 states that the determinant

in consequence of

The equation

+ b B 4 + c,C 4 + d D 4 =

a %A

(27.6)

c 3 , etc.

= *A 4

each value of

zeros are the only

/.

= tB 4

= /C4

w=

tf>4,

This result would seem to indicate that

common

solution of equations (27.5)

139

when

Determinants

A = B4 =

= D4 =

but we

shall

[Chap. 2

show that

this is not the

correct conclusion.

From

equations (27.7)

we obtain

D 4 y = B 4 tu, D 4 z = C 4 w.
B 4 C 4 and D 4 are the cofactors

D 4 x = A 4 w,

(27.8)

of the
Since by definition A 4
in
row
the
determinant
elements of the last
(26.1), they are
,

=
D4 =
C4

(27 9)

B4 =

aiC 2 d3

aib 2 d 3

Accordingly equations (27.8) are those which one obtains when,

using the method of 21, one solves for x, y, z in terms of w
the three equations (27.5) as i takes the values 1, 2, 3. From
this point of

view

it

follows that

when A 4

=B =
4

C4

= D 4 = 0,

either one of the three equations under consideration is a constant multiple of one of the others, or any one of the equations

equal to the sum of certain constant multiples of the other

(see Ex. 7). In either case a common solution of two of
the equations (not any two in the first case) is a solution of the

is

two

third equation,

and the three equations are

Accordingly we have

not independent.

established the theorem

Three independent homogeneous equations

[27.2]

+ biy + dz + d\w = 0,
a x + b 2y + c 2 z + d w = 0,
03* + b y + c z + d*w =

a\x
2

(27.10)

admit an
(27.11)

endless

x :y :z :w =

and

number

of

\biC 2 d3

common

:\aiC 2 d3

\:

solutions given by

aib 2 d3

\aib 2 c 3

1;

these are solutions also of the equation

ax + b 4 y + ctz + d4 w = 0,
if

and only

if the

determinant of the four equations

is

equal

to zero.

Observe that this theorem is a generalization of Theorem [20.1].

Returning to the consideration of equations (27.5), we remark that if we replace A 4 B 4 C 4 D 4 in (27.6) by any one of
,

140

Equations of the First Degree

Sec. 27]

in

n Unknowns

A J9 B,, C; Z), respectively, as j takes on the values 1, 2, and 3, the resulting equations are satisfied, one of them
because a^c^d* = 0, and the other three in consequence of

the three sets

Theorem

Hence not only does

[26.5].

(27.7) give a solution

of (27.5), but also

= tjA

x
for j =

J9

= tjBj,

tjCJ9

w=

tjDj

A =B =

C} = Dj = 0, it
any j we have }
}
follows from the above discussion that the three equations
(27.5), as i takes on values different from the particular value

of

j,

1, 2, 3.

If for

are not independent.

The preceding arguments apply to a set of n homogeneous

equations of the first degree in n unknowns and we have
[27.3]

When

n homogeneous equations of the

n unknowns is equal to zero and the cofactors
of the elements of any row are not all equal to zero, these
cofactors multiplied by an arbitrary constant constitute a
the determinant of

first degree in

common

solution of the equations, in addition to the solu-

tion consisting only of zeros.

As a consequence of Theorem
Theorem [22.3]

generalization of
[27.4]

When a

we have

the following

determinant of the nth order is equal to zero, there

not all zero, such that the sum
hi,
-, h n

numbers

exist

[27.3]
:

and the corresponding ele-, h n

of the products of hi,
ments in the 1st to nth columns of each and every row is
-

equal

to zero;

and

likewise

numbers

k\,

sum

kn

not all

-,k n and the

of the products of k\,
corresponding elements of the 1st to nth rows of each and

zero,

every

such that the

column

is

equal to zero.

We

return to the consideration of equations (27.2), (27.3),

when the determinant aibzCzd* is equal to zero.
In accordance with Theorem [27.4] there exist numbers k\, 2,

and

(27.4)

3
4, not all equal to zero, such that when equations (27.1)
are multiplied by k\, 2, fa, k respectively and added, the
coefficients of x, y, z, and w in the sum are zero
consequently
,

141

Lines and Planes in Space

the assumption that there

in

kiei

This and
|

dibtfzdi

+k
=

2 e2

is

common

[Chap. 2

solution

valid only

is

+ k 3 e3 + kei = 0.

are the conditions that the four equa-

in other words, common solutions (27.1) are not independent

tions of three of the equations are solutions of the fourth. But
;

three equations in four

unknowns admit an

endless

number

of

solutions of at least one degree of arbitrariness for, one at least

of the unknowns may be chosen arbitrarily, and then the others
;

are fixed

Since

by the equations.
all

of the foregoing discussion applies equally to n

equations of the
[27.5]

first

degree in n unknowns,

n equations of the first degree

and only one common solution,
minant of
nant

is

endless

in
if

we have

n unknowns have one

and only if the deter-

the equations is not equal to zero.

equal

to zero, there

number of one

or

are no

common

more degrees of

If the determisolutions, or

an

arbitrariness.

If

in the latter case the equations are homogeneous, there

an endless number of solutions.

is

In the preceding section and the present one we have shown

the theory of determinants and of first-degree equations
which was developed in Chapter 1 in the study of the geometry
of the plane and in Chapter 2 in the study of the geometry of
space may be generalized algebraically to determinants of
any order and to linear equations in any number n of unknowns. In order to speak of these generalizations geometrically, we introduce the concept of spaces of four, five, and any
number n dimensions. Although we may not be able to visualize the geometry of such spaces, we may speak about it and
deal with it.
Since in two-dimensional space an equation of the first degree defines a line, such an equation is sometimes called linear
and a line a linear entity. These terms are used in three dimensions to designate an equation of the first degree in three

how

unknowns and

the plane represented by such an equation reIn spaces of four, five, and higher dimensions it
customary to call an equation of the first degree in the

spectively.
is

142

Space

Sec. 27]

of

Four Dimensions

corresponding number of unknowns a linear equation, and the

locus defined by this equation a linear entity. Thus in twodimensional space a linear equation defines a linear entity of
a line. In three-dimensional space a
one lower dimension
linear equation defines a two-dimensional linear entity
the
plane; and two independent linear equations define a line.
Likewise in space of n dimensions one linear equation defines
1 dimensions, in the sense that each
a linear entity of n
solution of the equation gives the coordinates of a point in
1 of the unknowns may be chosen arbithis entity, and n
We say that the
trarily, and then the other is determined.
1 linear entity, and
space of n dimensions envelops such an n
that this entity is embedded in the space; for example, the
ry-plane is a linear entity of two dimensions embedded in space
of coordinates x, y, z. Similarly, in n-space, that is, space of n
dimensions, two independent linear equations define a linear
2 dimensions, and so on
and, in particular,
entity of n
In fact, if
1 independent linear equations define a line.
n
x n the coordinates of a representative
we denote by x
;

n
2
l
and
#i
*i
point of ^-dimensional space, and by x\
n
2
x 2 the coordinates of two particular points, equaX2\ x 2
tions of the line through these points are
,

X1
X2

Xi
1

Xi

X2

*2

Xi

Xi

Xn
#2

which are a generalization of equations

(5.2)

Xi

Xi

n
n

and

(16.1).

Just as the concepts of direction cosines and direction numbers of a line introduced in Chapter 1 have been generalized
to space of three dimensions in this chapter, so they may be
generalized to space of any number of dimensions, and there-

from the measure of angle between lines. Thus the analogue

of Theorems [6.9] and [17.5] is that the coefficients a\,
-, a n
in the equation

+ <*nX n + 6=0
1 linear entity
numbers of the normals to the n
by the above equation, there being one of these normals

are direction

defined

at each point of this entity.

143

Lines and Planes in Space

We

[Chap. 2

speak also of a generalized sphere of radius

n
its equation being
xo

at the point xo 1
l

(x

- xo

and center

(x

-x

2 2
)

(x

-*

n 2
)

= r2

a generalization of (12.1) (see 25, Ex. 13).

These are only suggestions of the manner in which the geometric concepts of three-dimensional space may be generalized
to a space of n dimensions. The subject is a fascinating one,
which the reader may pursue further either by himself (see
Exs. 9-15) or in consultation with books and articles dealing with
the subject (see the reference list which follows the exercises).

EXERCISES
Solve by means of determinants the equations

1.

3x-2y + 6z + 5w = -l,
Show

2.

x-lQy-3z-7w = 2.

that an equation of the plane determined by three non-

collinear points

(x\, y\, 21), (*2 , y*> 22), (*3 , J>3, 23) is

=0

^ &

y$

23

Show

that this equation is the same as the one in Ex. 10 of 21. Discuss this equation when the three points are collinear (see 15, Ex. 10).
3. Show that a necessary condition that the four planes whose
0, as i takes the values 1, 2, 3, 4,
equations are a t x + b ly + c,2 + d t
shall have at least one point in common is
a\biCzdi = 0. In what
|

manner can the above condition be satisfied without the four planes'
having a point in common ? Under what condition have the four planes
a

line in

common?

Given the tetrahedron whose vertices are 0(0, 0, 0), A(a, 0, 0),
0), and C(0, 0, c), show that the six planes each passing through
an edge of the tetrahedron and bisecting the opposite edge meet in a
4.

5(0,

b,

point. Is this result true when the axes are oblique, that is, when they
are not mutually perpendicular? Is it true for any tetrahedron?
5.

Show

that for the tetrahedron of Ex. 4 the six planes each biand perpendicular to this edge meet in a point.

secting one edge

144

Space

Sec. 27]
6.'

Show

of

Four Dimensions

that in the plane an equation of the circle through the

three noncollinear points

(x\ 9 yi), (x2 ,

y2 ),

(*s, yz) is

= 0.

x2

Discuss this equation when the three points are collinear. What is the
corresponding equation of a sphere through four noncoplanar points?
7.

(see

Show by means of Theorem

[22.3] that

(27.9)) there is a linear relation

(27.5) for

if

= C4 =

Z>4

between the three equations

1, 2, 3.

Show

that when all the determinants in equations (27.2),

and
(27.4) are equal to zero there is a linear relation between
(27.3),
two or more of the equations (27.1).
8.

9. Show that the coordinates of any point of a line in w-dimensional

space are expressible linearly and homogeneously in terms of the coordinates of two fixed points of the line (see Theorem [5.4] and

equations (16.9)).
10.

In four-dimensional space of coordinates

x, y, z,

and w the en-

+ =

is called a hyperplane. Show

e
tity defined by ax -f- by + cz -f dw
that it possesses the property used by Euclid to characterize a plane

in 3-space (see
11.

Show

a\%

17).

that in four-dimensional space two equations

b\y 4- ciz -f d\w

e\

a^x

0,

+ b2y -f c2 z -f d w + e2 = 0,
2

such that the coefficients of the unknowns are not proportional, are
equations of a plane (see page 74).

in

12. Show that in four-dimensional space two planes ordinarily meet

one and only one point. Discuss the exceptional cases.
13.

Where are
1, y = z =

0< x<

the points in space of four dimensions for which

iv
w Q;
x< 1,
0;
1, z
tv
Q; x, y, z, w are all greater than zero and

Q<y<l Q<z<l
t

less

than

14.
[17.6],

15.

0<
=

0<y<

0<*<1,

1?

Generalize to

and

-space

Theorems

[15.1],

[15.2],

[16.3],

[17.5],

[18.1].

How is a linear entity of r (<

n) dimensions defined algebraically

in n-space ?

145

Lines and Planes in Space

[Chap. 2

REFERENCES
1'

JORDAN, C. "Essai sur la geometric & n dimensions, Bulletin de la

Soctitt Mathimatique de France, Vol. 3 (1875).
MANNING, H. P. Geometry of Four Dimensions. The Macmillan

Company,

1914.

An Introduction to the Geometry of n

SOMMERVILLE, D. M. Y.
Dimensions. Methuen, 1929.
VEBLEN, O., and WHITEHEAD, J. H. C. The Foundations of Differential Geometry, Chapters 1 and 2. Cambridge University Press, 1932.
CAIRNS, S. S. "The Direction Cosines of a /?-Space in Euclidean
w-Space," American Mathematical Monthly, Vol. 39 (1932), pp. 518523.

146

CHAPTER

Transformations of Coordinates

it

28. Transformations of Rectangular Coordinates

In defining rectangular coordinates in the plane in 2 we chose

a point of the plane for origin, one line through it for the
#-axis, and the line perpendicular to the latter and through the
origin for the jy-axis, and defined the x- and ^-coordinates of
any point of the plane so as to be the directed distances of the

point from the j^-axis and #-axis respectively. Since the origin
and axes may be chosen arbitrarily, there is no such thing as the
coordinate axes for a plane, in the sense that a plane has a
If then we set up two
definite set of axes predetermined.
different sets of axes, it is evident that a given point of the
plane will have different coordinates with respect to the two

Since an equation of a locus has for its solutions

the coordinates of every point of the locus, we should expect
the equations of a given locus for two different sets of axes
to be different. Just what the difference is will be revealed if
sets of axes.

we know

the relation between the coordinates of each point

with respect to the two coordinate systems. Knowing this relation and because the choice of a coordinate system is arbitrary,
we are able at times to choose a coordinate system with respect
to which an equation of a locus is in simple form (see
32). It
is this relation which we shall obtain in what follows.

We consider first

the case

when

the two coordinate systems have

different origins but the %- and

^y

^y

two systems are paras shown in Fig. 20.

is

jy-axes of the
allel,

(*',yy

the origin of the system of coordinates x and y, and 0' of the

system of coordinates

%'

and

y'

furthermore the coordinates of

0' in the ^-system are XQ and yo.
FIG. 20
Then for a representative point P
of coordinates x, y and x', y' in the respective systems
'

TS + SP

we have

We

use parentheses with a prime, as in Fig. 20, to denote

a point in the * '/-system.

149

Transformations of Coordinates

[Chap. 3

An equation of any line or curve referred* to the ry-system is

transformed into an equation of the line or curve referred to
the *y-system by the substitution
x

(28.1)

= x' + XQ,

= y + yQ.

Although these equations have been derived for the case when
P is in the first quadrant of each system, the reader can easily
verify that they are valid for any position of P.
When equations (28.1) are solved for x' and y', we obtain
x'

(28.2)

= x-x

y'

=y

yo,

which could have been obtained directly from Fig. 20 on

relative to the ^'/-system are
noting that the coordinates of
~ XQ, yo. Also equations of the #-axis and jy-axis with respect
and x' + XQ = respectively.
to the ^'/-system are y + y =

The transformation
which

is

of coordinates

given by (28.2),

is

(28.1),

the inverse of

sometimes called a translation or

parallel displacement of the axes.

We

consider next the general situation when the

of axes are not parallel, as shown in Figs. 21 and 22.

two

sets

In each

\
FIG. 21

FIG. 22

case the direction cosines of the *'-axis relative to the ry-system

are cos 6 and sin 0, and an equation of the *'-axis is

cos

sin 6

or in other form
(28.3)

(y

jy )

cos

0-(x- XQ) sin = 0.

150

Sec. 28]

Transformations of Rectangular Coordinates

In each case by Theorem [6.11] an equation of the /-axis,

is the perpendicular to the #'-axis at 0', is

which

(y

(28.4)

Since cos

sin

(x

- xo) cos 9 = 0.

since sin 2

and

[8.1] that for

yo) cos 6

(y

the coefficient of y in equation (28.3),

9,

tive for Fig. 21,

Theorem

- yQ

+ cos

any values

of x

1,

it

is posifollows from

and y the expression

XQ) sin 6 is the directed distance of the point

(x

P(x, y) from the *'-axis, and that it is a positive or negative number according as P is above or below this line. By definition this
is the coordinate y' of P in the ^'/-system. Hence we have the
relation

/=

(28.5)

(y

Moreover, since sin

for Fig. 21,

by

0,

- yo) cos 0-(x- x

sin 0.

the coefficient of y in (28.4),

is

positive

the same kind of reasoning applied to (28.4)

we have
x'

(28.6)

Consider

(y

now Fig.

-y

22.

sin

0+(x- x

Since sin

is

cos

0.

positive also in this case,

the right-hand member of (28.6) for any values of x and y is the

directed distance of the point P(x, y) from the /-axis, that is,
the line (28.4), the distance being positive or negative according
as

P lies above or below

of the #'-axis

is

the /-axis. Since the positive direction

(28.6) holds for this

above the /-axis, equation

is negative, when the coordinates x and y

of a point P above the line which is the #'-axis are substituted
in the left-hand member of equation (28.3), the resulting num-

case also. Since cos

minus the directed distance from the line to the point, as

8. In the *'/-system P lies
follows from the considerations of
and
the
below the #'-axis,
/ of P is negative, and its
consequently

ber

is

equal to the absolute value of the left-hand

member of (28.3). Hence equation (28.5) holds also for Fig. 22.
Similarly it can be shown that (28.5) and (28.6) hold when the
absolute value

^'/-system

is

is

in such position that

lies

Solving equations (28.5) and (28.6)

write the result in the form

between 180 and 360.

for x - XQ and y
y

we

= x' cos - / sin + XQ

y = x' sin + / cos + yQ
x

151

Transformations of Coordinates

[Chap. 3

Hence an equation of any line or curve referred to the ry-system

transformed into an equation in the *y-system when the
expressions (28.7) are substituted for x and y in the given
is

equation (see
(28.7) when
When, in

40).

= 0.

Equations (28.1) are the special case of

particular, the origin 0' coincides with 0, equations (28.7) reduce to

(28.8)

= x' cos 6 - y' sin 0,

= x' sin + y

cos

6.

This transformation is sometimes referred to as a rotation of the

coordinate axes. In this case equations (28.6) and (28.5) are
(28.9)

x'

= x cos + y sin 0,

y'

= - x sin 6 + y cos 6.

These equations may also be obtained from (28.8) by interchanging x and y with x' and y respectively and replacing
8, which is what we should expect from the fact that the
by
6 with the #'-axis.
*-axis makes the angle
and
(28.9) of rotation of the axes may be
Equations (28.8)
written in the condensed form
1

with the understanding that each equation (28.8) is obtained

by equating x (or y) to the sum of the products obtained by
multiplying the element in the square which is in the same row
as x (or y) by the coordinate x' or y' directly above the element.
Equations (28.9) are obtained by taking the sum of the elements in the same column as x' or y' after multiplying each of

them by the coordinate on

its left.

Consider, for example, the transformation of coordinates when the

is the line x + 2y
4
Q and the origin 0' is the point (2, 1)

*'-axis

is 2 %
y 3 = 0, it being
the perpendicular to the *'-axis through the point (2, 1). The slope
of the *'-axis is negative and if we take as the positive sense of the
*'-axis that of the line x + 2 y - 4 = 0, we are dealing with the type

on

this line.

Then an equation of the jy'-axis

;

152

Sec. 28]

Transformations of Rectangular Coordinates

represented in Fig. 22 in which the positive direction of the y-axis is

downward. Accordingly we have the second of the following equations, the first following from the fact that the positive direction of
the *'-axis

is

upward

__

~_
V5

V5

EXERCISES
1.

Find an equation of the

line

2 x

3y

*y-system with origin at the point

the x- and >>-axes.

the
to

2.

Show

that

it

5
1)

(1,

follows from (28.8) that x 2

with reference to

and with axes

parallel

+ y 2 = x' 2 + y' 2

and

2
2
explain why the quantity x -f y should be invariant under a rotation
of the coordinate axes.

2 x
3. Find an equation of the curve x
4jy
y
ferred to a coordinate system with axes parallel to the x-

and with

origin at the point

5.

2,

when
1)

re-

and

4#-f3;y

1=0.

Find the transformation of coordinates of oblique axes into

axes parallel to
7.

Find the transformation of coordinates to axes with origin at

1=0 for the new jy-axis, the
1) and with the line 4* + 3;y

positive sense along the latter being that of the line

6.

re-

.y-axes

1, 2).

Find an equation of the curve y 2 + 2y

&x 15
ferred to a coordinate system with origin at the point (
with axes parallel to the x- and ;y-axes.
4.

(1,

when
and

Two

them

(see

11).

systems of rectangular coordinates,

related so that the points

(0, 0),

(1, 0), (1, 1) in

y) and (*', /)', are

the ry-system are the

(x,

1)', (1, 0)', (0, 0)' respectively in the *'y-system. Draw

points (1,
a diagram showing the relative positions of the two sets of axes, and
find the equations of the corresponding transformation.

8.

Show

that equations (28.5), (28.6), and (28.7) for

a negative
180 and
are the same as those for an appropriate

angle between
e between 180

and 360.

Determine the translation of the axes such that the equation

3 y2
4 x
12 y =
is transformed into one in which there
are no terms of the first degree in x' and y'.
9.

2 x2

153

Transformations of Coordinates

[Chap. 3

Into what equation is the equation 9 x 2 + 2 V3 ry

transformed when the axes are rotated through 60?
10.

-f 11

y-4=

11. Justify the following statement without carrying through the

In terms of coordinates x'
transformation of coordinates involved
and /, referring to a coordinate system with origin at the center of
the first of the circles (12.11), and with the line of centers of the two
:

circles for x'-axis,

and from this

lie on a line.

equations of the circles are

result it follows that the centers of the circles (12.12)

12. Show that equations (28.5) and (28.6) may be interpreted as

the result of a translation of the ry-system to the point (x yo) for
new origin, and then a rotation of axes through the angle 0. What
,

are the equations

if first

there

is

a rotation of the axes through the

jvo) as new origin ? Compare

and then a translation to (x

angle
the resulting equations with (28.7).

13. Find the transformation of coordinates from a rectangular xysystem to oblique axes with the equations ax + by = 0, ex + dy = 0.
What is the area of the triangle whose vertices are (1, 0)', (0, 0)',
(0, 1)' in

the

new system?
29. Polar Coordinates in tRe Plane

Rectangular and oblique coordinate systems are not the only

kinds of coordinate systems which may be employed in the treatment of geometric loci. Another system frequently used is one
in which the coordinates of a point P are its distance r from a
called the origin, or pole,
fixed point
p
and the angle which the line segment,
or vector, OP makes with a fixed vector
r
OM, the polar axis of the system, the
\~^\

angle
being measured from the axis in
the counterclockwise direction, r and
6 so defined, and called the radius vector

and

"p IG

23

vectorial angle respectively, are polar coordinates of P,

and

denotes the point with these coordinates. The reader in

drawing graphs of algebraic equations in x and y may have
used graph paper with two sets of parallel lines, each set perpendicular to the other, which form an array of small squares.
(r,

6)

154

Sec

Polar Coordinates in the Plane

There exists also graph paper for polar coordinates ruled with
a set of concentric circles and with lines radiating from the

common

We

center of the circles.

observe that the point

+ n 360
have

P may

be defined also by

and

any positive integer n. In this sense a point may

sets of polar coordinates, all referred to the same

for

many

may be used to assign

pole and axis. Also negative values of
a
to
a
coordinates
point,
negative angle being described
polar
from the axis
in the clockwise direction. Thus far we have

OM

understood r to be a positive number, but it is advisable to give

a meaning to polar coordinates when r is negative. By definithe length r not on
tion, if r is negative, we lay off from
with OM but on the vector makthe vector making the angle
|

180. With this understanding

and
2,
ing the angle
of
are
the
same
The
coordinates
neces180
2,
polar
point.
sity for a convention concerning negative values of r arises when

one seeks the graph of an equation in polar coordinates, that is,

the locus of all points with solutions of the equation as polar
coordinates.

Consider, for example, the equation

r

(29.1)

cos

(a>

0).

say 0\ in the first quadrant r is positive, and

has the same numerical value but is negative
one obtains the same
consequently, for two such values of
is in the second quadpoint. Similar reasoning applies when
takes the
rant. Another way of stating this result is that as
to 360 the curve is described twice. Clearly
values from

For a value of
r for 0i

0,

+ 180

is no necessity in this case of taking values of

greater
than 180, because of the periodic property of cos 0. This does

there

not apply to such an equation as

Consider next the equation
(29.2)

r2

= cos ~

(see

Ex.

6).

= 4 cos 0.

In order that r may be real, cos cannot be negative conse90 to +90. For
are from
quently admissible values of
2Vcos 0. Coneach such we have two values of r, namely,
with the axis OM
sequently, on the line making such an angle
155
;

Transformations of Coordinates
there are

two points of the curve on opposite

[Chap. 3

sides of

and

equidistant from it (see Fig. 24). The curve is symmetric with

respect to the origin, to the polar axis, and to the line 6 = 90.
In Fig. 24, on each line there is noted the angle which it makes

with the

axis,

expressed in degrees and in radians.

1ST

FIG. 24

When
ment

enters in an equation directly, and not as the argu-

of a trigonometric function, as, for example, in the case of

the spiral of Archimedes with the equation

r

(29.3)

ad,

necessary to express the angle in terms of radians, since

both members of the equation must have the character of length.
In such cases
takes on all possible values. Thus, if in (29.3)
it is

= 2,

^ + 2nw,

(2n

we take

and

takes, for example, the values

+ I)TT, ^ +

(2

+ I)TT,

n being any

+ 2 nir,

positive or

negative integer, then the values of r are twice these values, and
the reader on plotting the curve will see that it is a double spiral.
Although the same point has the two sets of coordinates
180), it may be that if one of these sets
(r, 0) and (- r,

156

Polar Coordinates in tne Plane

Sec. 29]

of coordinates satisfies

an equation and thus determines a point

This situation did not arise

equations (29.1) and (29.2), but it does arise for the equation r = cos2 0. For, if r, 6 is a solution of this equation,
- r, 6 180 is not a solution.
Polar coordinates are necessarily related in some way to recof the locus, the other set does not.
for

tangular coordinates. A simple form of the relation is obtained

when the origins of both systems coincide and the axis of the
polar system is the positive *-axis of a rectangular system. In
this case for r positive the relation is

(29.4)

= r cos 0,

= r sin 0,

the x- and ^-coordinates of a point P being the projections of

the vector OP (Fig. 23) upon the respective axes. Suppose
now that r is negative ; then, as one sees by drawing an appropriate figure,

= r cos 0,

cos

r
|

and equations

sin 6

r
\

= r sin 6,

(29.4) hold in this case also.

When

polar coordinates of a point are given, we obtain

rectangular coordinates of the point directly from (29.4). If
rectangular coordinates are given, we must solve (29.4) for r

and

9.

(29.4)

The

first

of these

and adding the

(29.5)

where e = +

or

1.

is

results

obtained by squaring equations

from this we obtain

= eVx2 + y2

And

must

satisfy

any two of the equa-

tions
(29.6)

sin0

>

cos

1 in (29.5) and (29.6), we

According as we take e = + 1 or
have two sets of polar coordinates, (r, 0) and ( r, + 180)
for any point whose rectangular coordinates are given. If we are
transforming an equation in polar coordinates into one in rectangular coordinates, we must substitute directly from (29.5)
and (29.6). Then one of two things will happen: either e will
enter only as e 2 and thus be replaced by + 1, or e itself will
appear. In the second case, if a polar equation admits both

157

Transformations of Coordinates

[Chap. 3

positive and negative values of r, there are, in fact, two equaone for e
1
tions of the locus in rectangular coordinates

=+

1
giving the points for which r is positive the other for e =
r
which
is
in
for
If
second
the
case
the
negative.
giving
points
2
one solves the equation for e^/x 2
y and squares the resulting
equation, one obtains an equation in x and y whose graph is the
complete locus of the given equation in polar coordinates.
;

(29.5) and (29.6) in (29.1), one

with center (a/2, 0) and passing
through the origin this equation is an equation of the locus of (29.1).
For the equation r = sin + % we have x 2 + y 2 = y + ^ e^/x 2 + y 2
Consequently, when e = + 1 this is an equation of the part of the curve

For example, on substituting from

obtains x 2

+ y 2 = ax,

that

is,

circle

for
If

r is positive ; when e ==
solve the equation for e

which

1,

the part for which r is negative.

and square the result, we ob-

V* 2 -f y 2

we

an equation in x and y of the complete locus.

For the equation r = cos 2 6 we have (x 2 + y 2 )? = x 2 since r cannot
be negative and hence e = + 1. This equation in x and y is the equacos 2 6. In fact, the two polar equations are equation also for r =
tain

tions of the

same

2
curve, since cos (0 4-

180)

= cos 2

0.

When an

equation of a locus is given in rectangular coordinates,

equation in polar coordinates is obtained on
the
expressions (29.4) for x and y. Thus in polar
substituting
is
coordinates the equation of a line ax + by + c =
its

r(a

(29.7)

When

cosO

+ b sin 6) + c = 0.

in (29.7), that is, when the line passes through

c
the origin, we have for all values of r other than zero that
tan 6
a/b ; and when 6 satisfies this condition the equa-

tion

is

satisfied

by

all

values of r

that

is,

= const,

is

an

equation of a line through the origin. It should be remarked

that in polar coordinates an equation of a line is not of the first
degree in r and 6, as is the case when rectangular coordinates
are used ; in fact, (29.7) is not an algebraic equation in r and 9.

Since any trigonometric function

in terms of x

and y by means of

is

expressible algebraically

(29.6), it follows that

when an

equation in polar coordinates is an algebraic function of r and

of one or more trigonometric functions of 0, the corresponding
equation- in rectangular coordinates

158

is

algebraic.

Polar Coordinates in tKe Plane

Sec. 29]

EXERCISES
1. Draw the graph of each of the following equations for a positive,
using a table of natural trigonometric functions or radian measure, as
the case requires
:

b.

r = a sin 6.
r = 0(1
cos 0)

c.

a.

a sin 2

(the cardioid).

0.

= a sin 3 0.
= asec2 0.
r 2 = 2 a 2 cos 2

i.

d. r

j.

e.r

k.

/.

= 0(cos
= 2 a cos
r = cos
r0 = a.
r 2 = a0.

sin 0).

g. r

h. r

sec 2

(the limaQon).
0.

(the lemniscate).

Plot separately the portions of the curve h in Ex. 1 for which

and r is negative; derive equations in rectangular coordinates of each portion, and an equation of the complete locus.
2.

r is positive

3.

For what values of k

is r

for

what values of k

is r

always positive for the curve

= sin 2 +

always negative ? Plot a curve of each

set.

For which of the curves in Ex. 1 for a positive does r have posizero, and negative values; for which no negative values; for
which no positive values?
4.

tive,

5.

Find equations

6.

What range

in rectangular coordinates of the curves in Ex. 1.

f\

for

must be used

for the

equation

= sin - in order

to obtain the complete graph of the equation, a being

some integer?

7. Let AB be a fixed line perpendicular to the axis of a polar cothe

ordinate system meeting the axis in the point D ; denote by
The
the
of
meets
line.
locus
the
in
which
vector
through
any
point

points

P and

length,

is

P' such that r

= OM +

called the conchoid.

Find

/,

its

r'

= OM

/,

where

/ is

a fixed

equations in polar and also in

rectangular coordinates.

and has the polar axis for

a passes through
in Q, and the tangent
the
circle
a
line
meets
diameter
;
through
to the circle at A in R. The locus of P on the line OR such that
8.

circle of radius

OA

PR = OQ

is

called the cissoid.

Find

its

equations in polar and also in

rectangular coordinates.
9.

Find the points of intersection of the curves r

aQ for
also of the curves r = a0 and r =

= a sin

159

= a cos
^ 0.

and

Transformations of Coordinates
30. Transformations of Rectangular Coordinates in

[Chap. 3

Space

In this section we derive the equations connecting the coordinates of a point in space with reference to two different
sets of rectangular axes.

We derive first the equations

connecting the coordinates of
a point with reference to two
sets of axes

respectively parto one another, as shown

in Fig. 25. The origin 0' of

allel

the x'/z'-system

is

the point

) with respect to the

xyz-system, and hence the origin

(XQ, yo, z

of the latter system

point

(#0,

is

the

ZQ)' of the

yo,

28, we
#'/2'-system. As in
obtain the following equations :
x'

FIG. 25

=x

= x' +

(30.1)
Z

=Z

ZQ,

Z'

+ ZQ.

If the expressions for x, y, and z in the second set are substituted in an equation in x, y, and z, one obtains the corre1
sponding equation in x', y', z .

transformation (30.1)
translation of the axes.

We

is

called

consider next the general

shown in Fig. 26, for

which the origin 0' of the x'y'z'-

case,

as

system is the point (XQ, yo, z 0t ) with

respect to the ryz-system, and the
direction cosines of the axes O'x',
O'y', O'z' with respect to the xyzsystem are

A3

FIG. 26

160

Sec. 30]

Transformations of Rectangular Coordinates

Being direction cosines, they must satisfy the

equation (15.6) hence we have the three equations

respectively.

X,*

(30.2)

+ VL? +

v?

(i

1, 2, 3).

This notation means that one gives i the value 1 and gets one
equation then the value 2 and gets a second equation and
so on. Moreover, since the axes are mutually perpendicular,
in accordance with Theorem [16.8] we have
;

Xi\ 3

(30.3)

= 0.
Equations (30.2) and (30.3) are the conditions that the
quantities involved are direction cosines of three mutually
perpendicular lines, but do not determine the mutual orientation of positive directions on these lines. Consequently we must

impose a condition to ensure that the mutual orientation of

the axes shall be the same in both systems, mutual orientation
being as defined in 14 and illustrated by the edges of a room
meeting in a corner of the floor, the corner being the origin
of the system. If we start with an ryz-system, it is evident
geometrically that, once the positive axes O'x', O'y' of a second
system are chosen, the positive axis O'z' is completely determined by the prescribed mutual orientation of the axes. Suppose then that Xi, jui, v\ and X2, ^2, ^2 are given; if we
apply Theorem [20.1] to the second and third of equations
(30.3),

we obtain
X3

H3

(30.4)

Substituting these expressions in (30.2) for

= 3,

we

obtain

(30.5)

Since the axes O'x' and O'y' are perpendicular by hypothesis,

it follows from equation (15.10) that the expression in brackets
in (30.5)

is

equal to

+ 1,

and consequently
161

is

+1

or

1,

Transformations of Coordinates

When we

substitute the expressions (30.4) in the determinant

(30.6)

Xi

MI

Z>= X 2

M2

X3

MS

and note that the expression

+ 1,

we

[Chap. 3

in brackets in (30.5)
is equal to t.

find that the determinant

is

equal to

When,

in

particular, the axes of the Jt'/z'-system are parallel to those of

the ryz-system, as in Fig. 25, the determinant (30.6) is
1
1
1

,0

1.
is
Consequently we impose the condition
that the determinant (30.6) be
1, to ensure that if the axes
of the j'/z'-system, as shown in Fig. 26, are rotated about 0',
r
so that the axes O'x' and O'y become parallel to Ox and Oy re-

whose value

then the axis

spectively,
sense.

Oz and has the same

O'z* is parallel to

When we put / = 1 in (30.4), we have that the right-hand

members of (30.4) are the cofactors of the corresponding lefthand members in the determinant (30.6). It is readily shown
that similar results hold for the other direction cosines, so that
the theorem

we have

When

[30.1]

the direction cosines of three lines satisfy the con-

and

ditions (30.3)

equal

With

D=

to its cofactor in the

in (30.6),

is

reference to the ^y^'-system the direction cosines of

Ox, Oy, Oz are Xi, X 2 X 3

MI, M2, M3
we have also the equations
,

v\ 9

1*2,

J>3.

Consequently

+ X 2 M2 + X 3 Ms = 0,
+ M2^2 + M3*>3 = 0,

+ X 22 + X 3 2 = 1,
2
2
2
Ml + M2 + MS = 1,
Xi

(30.7)

any one of them

determinant D.

= 0.
These equations are in keeping with the above theorem, as
readily verified.

162

is

Sec. 30]

Transformations of Rectangular Coordinates

Equations of the /2'-plane, the *'2'-plane, and the *'/-plane

in the sjyz-coordinate system are respectively

- *o) + MiCv - yo) + vi(* - *o) = 0,

X 2 (* - *o) + M2CV - yo) + 2(z - z = 0,
X 3 (* - X + /zaCv - ^o) + "3(* - *o) = 0.
Xi(x

(30.8)

When the axes O'x', O'y', O'z' are so placed that each of them
makes an acute angle with the positive direction of the axis Oz,
the direction cosines v\, P2, and 3 are positive; consequently
we can apply directly Theorem [18.1] and have that the disj>

tances of the point P(x, y, z) from the planes (30.8), that

the coordinates x', /, z' in the new system, are given by

=
=
y'
z' =
*'

(30.9)

- x + fjii(y - y +
X 2 (* - * + 2 (y - yo) +
Xi(*

/i

-z
- zQ
v*(z
v\(z

is,

),
),

When the new axes are not so placed, the above equations
hold just the same, as can be shown. For example, from the
18 it follows that if v in any of the equations
discussion in
is
negative, then below the corresponding plane is its
(30.9)
side
with respect to the ryz-system thus if v\ is negapositive
;

tive the positive #'-axis is directed downward. If for any position of the new axes the signs of the coefficients in two of

equations (30.9) are chosen in accordance with the positive

directions of the corresponding new axes relative to the xyzsystem, the appropriate signs of the coefficients of the third
equation are determined in accordance with Theorem [30.1],
which assures that the positive directions of the axes in the
*'/2'-system have proper mutual orientation. For example,
suppose O'x' lies in the first, or principal, octant, in which case
Xi, /zi, PI are all positive, and O'z' is tilted forward so that O'y'
is

directed downward.

Now

X3

is

negative, since O'z' makes an

v 3 are positive. In accordance

obtuse angle with Ox, and ju 3 and

with the above theorem, v 2 = Xa/ii

Xi/zs

the right-hand

member

of this equation is negative, and consequently 1/2 is

9
negative, as should be the case because of the position of O'y
.

The reader should

verify the statement following (30.9)

considering the other possible cases.

163

by

Transformations of Coordinates
If

[Chap. 3

we multiply equations (30.9) by Xi, X 2 X 3 respectively,

results, and make use of (30.7), we obtain the first of
,

add the

the following equations, the others being obtained similarly,

using jui, /z 2 MS and v\> ^2, ^3 respectively as multipliers:
,

+ x 3 z' + *o,
y = jui*' + M2/ + Ms*' + yo,
Z = ViX' + V 2 y' + V*Z' + ZQ.
x 2y

(30.10)

These are the equations of the inverse of the transformation

When, in particular, the center 0' of the x'/2'-system

of the ;ryz-system, the equations of
coincides with the origin
the transformation are
(30.9).

= Xi*' + X 2/ + X 3 *',
y = MI*' + M2/ + M3*',
z = v\x' +
2 y + vzz'.
x

(30.11)

j/

The transformation

(30.11)

is

sometimes referred to as that

corresponding to a rotation of the original axes.

When

the ex-

pressions (30.10), or (30.11), for x, y, and z are substituted in an

equation (or equations) of a locus referred to the #>>z-system,

the resulting equation (or equations) is an equation (or are

equations) of the locus in the *'/2'-system.
The inverse of a transformation (30.11) is
x'

(30.12)

Xi*

+ p,iy +

viz,

as follows from (30.9) on putting #

The reader should observe that in

equal to zero.
any of the equations (30.11)
and (30.12) the coefficient of any term on the right is the cosine
of the angle between the axis of the coordinate which is multiplied by the coefficient considered and the axis of the coordinate on the left-hand side of the equation. For example, from
,

;vo,

ZQ

the second of (30.11), and also from the third of (30.12) we have
is the cosine of the angle between Oy and Oz'.
This

that MS

164

Sec. 30]

Transformations of Rectangular Coordinates

observation and equations (30.11) and (30.12) are set forth in

the following table
f
x
z*
:

Any element

in the square

the axes in whose

is

the cosine of the angle between

row and column

it lies. Moreover, an equaobtained by multiplying each element in the

same column as #', /, or z' by the coordinate in the same row
on the left, adding the results and equating this to the coordinate at the top of the column. Equations (30.11) are similarly
obtained by using rows instead of columns.

tion (30.12)

is

EXERCISES
Transform by a suitable transformation (30.1) the equation
2 x - 16 y + 12 z + 28 =
so that in the resulting
equation there are no terms of the first degree. Is this possible for
3 = 0?
the equation xy + 2 z
1.

x2

+ 4 y2 + 3 z2 -

2.

Show

that for the second of transformations (30.1), and also for

(30.11), the expression (15.2) is transformed into an expression of the

should one expect this to be
same form in the new coordinates.

Why

the case?
3.

Show

that

the

2 y + z = Q may

planes

+ 2^ + 22 = 0, 2x + y-2z = Q,

be used as coordinate planes of a coordinate

system, and find the equations of the corresponding transformation.

2 x

that the three planes x

0, x-2
y + z- 4
be taken as the coordinate planes of a rectangular
and x
coordinate system, and find the corresponding transformation of
4.

Show

z =
may

coordinates.
5.

Find the equations of a transformation of coordinates so that

the plane x -f
6.

x2

-f

-f 2

is

the x'y '-plane in the

new system.

Transform by means of (30.11) the equation of the sphere

2
y 4- 2 2 4- 2 fx + 2 gy 4- 2 hz -f e = and discuss the result.
165

Transformations of Coordinates
7.

[Chap. 3

Interpret the transformation

= *' cos 6 - / sin 0,

y=x

+ / cos

sin

0,

= z'

as a particular type of rotation of the axes.

8.

How must

6 be chosen in a transformation of the type of Ex. 7

2 hxy
Q may be transd

+ =

2
cz 2
so that the equation ax 2
by
in
a
term
formed into one lacking
x'y'

9. Find equations of the line each point of which has the same
coordinates in two coordinate systems in the relation (30.11); inter-

pret the result geometrically.

10.

Show that three lines with direction numbers

1, 2,

2,

1,

are mutually perpendicular, and find the transformation of coordinates to an #'yz'-system having lines with these direc-

and

2,

2,

tion numbers and through the point (2, 0,

Apply this transformation to the equation

4 x2
11.

x = x'

+ 4 y 2 - 8 z 2 + xy -

Show

that the equations

(cos

cos

<

sin

sin

y (cos
y = x'

(sin

</>

cos

\f/

+ cos

</>

\f/

-f sin

</>

cos 0)
sin ^
cos

i/'

5yz

+ 9 = 0.

cos 6)

sin

sin

y (sin
z = x' sin ^ sin 6 + y' cos

\l/

5 xz

3) for coordinate axes.

cos

^ cos 6) + z' sin

<f>

sin

cos

<f>

sin 0,

0,

\f/

sin 6

cos

\[/

cos 0)

z'

+ z' cos 0,

any values of 0, 0, and ^, are equations of a transformation

(30.11). These equations are known as Ruler's formulas.

for

31. Spherical

and Cylindrical Coordinates

In the study of certain phenomena in space there are systems

of coordinates other than rectangular coordinates which are
found to be more useful.
define

We

one such system and establish

its

relation to a rectangular system,

as shown in Fig. 27. The position

of a point P is determined by its
distance r from a point 0, by the
angle <f> which the vector OP makes

with a fixed vector Oz, and by the

angle 6 which the plane of the
166

Sec. 31]

Spherical

and Cylindrical Coordinates

OP and Oz makes with a fixed plane through Oz. When

taken for origin, Oz as the positive 2-axis, and the fixed
plane through Oz as the *z-plane of a rectangular coordinate system, the angle 6 is the angle between the #-axis and the projection of OP on the #>>-plane. We make the convention that 6 is
measured from Ox toward Oy. Hence we have the equations
vectors
is

(31.1)

= r sin

<j>

cos

Squaring these equations

r

(31.2)

Then from

0,

r sin

<

sin 0,

and adding the

rcos0.

results,

we

find that

we have

(31.1)

as the other equations of the in-

verse transformation

The

coordinates

ordinates

From

cos

(31.3)

(31.2)

</>,

and 6 as defined are

writers call

it

= 2.

tan

called spherical cocoordinates

in space.
polar
k as the constant
follows that the surfaces r

r,

some

*
,

them

as their common
k takes different values are spheres with
as the center of the earth, of the line
center. If we think of
joining the center to the North Pole as the fixed vector of
reference Oz, and the plane through this line and Greenwich
as the fixed plane of reference, then 6 is the west longitude of a
</> is its latitude
</> is
point on the earth's surface, and 90
;

sometimes called the colatitude of the point.

Another system of spatial coordinates

is

defined

by the

equations
(31.4)

= rcos0,

;y

= rsin0,

= d,

from which we have as the inverse

(31.5)

r=\/x2 + y2

tan

0=2,

= z.

Ai

In this system the surfaces r = k as the constant k takes different values are cylinders with the z-axis as common axis. The
quantities r, 0, d defined by (31.5) are called cylindrical coordinates. These equations have for basis a plane with a polar
coordinate system, and an axis perpendicular to the plane at
the pole of the polar system.

167

Transformations of Coordinates

[Chap. 3

EXERCISES
1. What are the surfaces 6 = const, in spherical coordinates ? What
= const. ?
are the surfaces
<

2. Discuss the several loci defined by the equations obtained when

two of the spherical coordinates are equated to constants.

Find the direction cosines with respect to the ryz-system of the

from the origin to the point whose spherical coordinates are r, 0, 6.

3.

line

4. Find the expression for the distance between

spherical coordinates ; in cylindrical coordinates.

two points

in

Find the spherical coordinates of the points whose rectangular

- 2), (- 1, 2, - 2).
(2, 4, 3), (3, 3,

5.

coordinates are
6.

Find the cylindrical coordinates of the points

7.

Find equations

in

Ex.

in spherical coordinates of the loci

5.

whose equa-

tions in rectangular coordinates are

a.

b.

x 2 + y 2 = 5.
3 x - 4y + 5

c.4* 2 -.y a

+ 2 y* + 3 z 2 - 6 = 0.
xy + yz + xz = 0.

d. x*

- 1 = 0.

e.

= l.

168

CHAPTER

The

Conies. Locus Problems

a
32.

A Geometric Definition of the Conies

In the preceding chapters we have obtained in a number of

an equation of the locus of a point which satisfies a certain
relation to points and lines defined with reference to a given
coordinate system, and we have given certain exercises to find
a locus so defined. An equation of such a locus depends for
its form not only upon the geometric character of the locus
but also upon the particular coordinate system used. In view
of the results of
28 it is clear that in finding an equation of a
given locus one is free to choose a coordinate system with respect to which the fixed objects (points, lines, etc.) involved in
the definition of the locus have such position that the equation
of the locus, and the calculation involved, shall be as simple as
possible, provided that in so doing one is not imposing additional conditions upon the locus. For example, the results of
Ex. 21 of
13 hold for any triangle, because a rectangular
coordinate system can be chosen so that the coordinates of the
vertices of any triangle are as given in this exercise. On the
13 apply only to a rightother hand, the results of Ex. 20 of
angled triangle if we are using rectangular axes, but the results
would be true for any triangle if shown to be true for a triangle
with vertices 0(0, 0), A (a, 0), B(b, c). We remark that a coordinate system can be chosen so that any two points have
the coordinates (b, 0) and ( b, 0) by taking the line through
the points for the #-axis and the origin at the mid-point of the
segment connecting the two points accordingly the geometric
13 is not changed, but its
character of the locus in Ex. 9 of
cases

equation

two

is

simplified,

by taking

(b,

0)

and

b,

0) as the

points.

We make

use of this idea of choice of an advantageous co-

ordinate system in deriving equations of the conies from the

following geometric definition:
Let F be a fixed point, called the focus, and d a fixed line, not
through the point, called the directrix; then a conic is the locus of
a point P such that the ratio of the distance FP to the distance of P
the line d is a positive constant, e, called the eccentricity; it
a parabola when e = 1, an ellipse when e < 1, and a hyperbola

from
is

when

>

1.

171

The

Conies. Locus Problems

[Chap. 4

Distance as used in this definition is numerical, or absolute,

38 we show the equivalence of this definition
In
and the definition of the conies as plane sections of a right
distance.

circular cone.

From the above definition it follows that the quantities which

determine, and therefore can affect, the size and shape of a
conic are its eccentricity and the distance of the focus from the
directrix. Hence conies of all possible shapes and sizes are obtained by taking all values of the eccentricity, a given line d
for directrix, and a given line / perpendicular to d for a line of
a particular focus being determined by its distance from
two other lines d' and /' are chosen for directrix and line
of foci respectively, a conic defined with respect to these lines
can be brought into coincidence with one of the set of conies
defined with respect to d and / by bringing d' into coincidence
with d and the focus into coincidence with one of the points
foci,

d.

If

of/.

In order to obtain an equation of the conies, we take the

directrix as the jy-axis and the perpendicular to the directrix
through the focus as the #-axis, and denote the focus by (Jfe, 0).

The

distances of a representative point

and

directrix are

V(x

K)

+ y2

cordance with the definition,

and x
we have
|

y) from the focus

respectively. In ac-

(x,
\

-*)*

where

D denotes the

directrix.

From

upon squaring both

(32.2)

foot of the perpendicular from

P upon

the

the equation

(1

sides,

we

obtain

- e*)x> -2k

Suppose now that we consider in connection with equation

(32.2) an equation of a conic of eccentricity e with the y-axis
for directrix and for focus the point (tk, 0), where / is a positive
172

Sec. 32]

A Geometric

constant.

If

Definition of the Conies

we denote by

P'(x',

/) a

the conic, the equation of the conic

(32.3)
If

(1

* 2 )*' 2

is

representative point on

found to be

- 2 tkx' + y' 2 + tW = 0.

then we substitute
x'

(32.4)

= tx,

y'

= ty

we obtain equation (32.2) multiplied by /

Hence to each point on either curve there corresponds a point
on the other such that the line joining corresponding points
2

in this equation,

passes through the origin, that is, through the point of intersection of the directrix and the perpendicular upon the latter

through the focus.

Moreover, the distances of

the origin are related as follows

P and

P' from

OP'

= vV 2 + y' 2 = V*2 + y2 =
/

OP.

Thus if / > 1 the second curve is a magnification of the first,

and vice versa if / < 1. In both cases, as a matter of convenience, we say that either curve is a magnification of the other.
Two curves in a plane are said to be similar when there is a
in the plane such that if P is any point on one curve,
point
the line OP meets the other curve in a point P', and the ratio
OP I OP is a constant for all such lines. Also two curves are
said to be similar when either curve is congruent to a suitable
magnification of the other. Likewise in space two figures of any
1

their points are related to a point

congruent to a suitable magnification of

kind are said to be similar

as above, or
the other.

As a
and the
[32.1]

if

either

is

if

result of the discussion of equations (32.2)

definition of similar curves, we have

Two

conies having the

particular, all

same

and

(32.3)

eccentricity are similar;

in

parabolas are similar.

By a suitable translation of the axes we shall obtain in 33

an equation of the parabola, and in 35 equations of the ellipse
and the hyperbola, in simpler form than (32.2). When the
equations are in these simpler forms, the geometric properties
of the loci are more readily obtained.

173

The

Conies. Locus Problems

[Chap. 4

EXERCISES
1.

Show

that in accordance with the definition of a conic the line

through the focus perpendicular to the directrix is an axis of symmetry of

the curve that is, if P is any point on the conic, so also is the point
symmetric to P with respect to this axis (see 2). Show also geometrically that there is one and only one point of a parabola on this axis,
;

and that there are two and only two points of an ellipse or hyperbola
on this axis. Verify these statements algebraically from equation (32.2).
2.

Find an equation of the conies when the directrix

is

taken as

the x-axis, and the line perpendicular to the directrix through the
focus for the jy-axis, and the focus is denoted by (0, k).

Find an equation of the parabola whose directrix is the line

= and whose focus is (2, 3) also an equation of an ellipse of
with this directrix and focus.
eccentricity
3.

Find an equation of a hyperbola of eccentricity 2 whose

4^ 3^4-2 = and whose focus is (1, 1).

directrix

Find an equation of the conic of eccentricity e whose

+ by + c = and whose focus is the point (h
reduce the equation to the form

directrix

4.
is

the line
5.

is

the line ax

Ax 2

+2

Hxy

By

+ 2 Fx + 2

Gy

k),

and

+ C = 0.

Given two

triangles whose sides are proportional, show that a

can be found with respect to either triangle such that a
suitable magnification of this triangle with respect to gives a triangle
6.

point

congruent to the other.

33.
Since a parabola

+ 1, equation

is

(32.2) for

The Parabola

a conic for which the eccentricity e

a parabola may be written

^=2

(33.1)

In order to obtain an equation in simpler form,

translation of the axes defined
(33.2)

and equation

is

*'

(33.1) is

by

= *-

'=
y=y>

transformed into

/*

= 2 fee'.
174

we make

the

The Parabola

Sec. 33]

In this coordinate system the axis of the parabola, that is,

the line through the focus perpendicular to the directrix, is
the x'-axis, and the point of the parabola on its axis, called the
vertex, is at the origin. Moreover, the focus, which is (k, 0) in
the ry-system,
(33.2)

is

(k/2, 0)' in the ^'/-system, as follows

and similarly, an equation of the directrix, which

in the ry-system,

is x'

+5=

in the

is

from
x=

#y-system.

L4

new coordinate system in what

and
x
to
use
denote
the coordinates instead of
follows,
y
f
x and y'. If then, in order to avoid fractions, we put k = 2 a,
an equation of the parabola is
As we

shall use only this

we

(33.3)
its

focus

This

is

the point

shown

is

(a, 0),

= 4 ax,

and

its

directrix the line x

28 for the case when a

in Fig.

The perpendicular

is

+ a = 0.

positive.

to the axis at

a = 0,
the focus, that is, the line x
in
the
meets the parabola
points
2 a). The line seg(a, 2 a) and (a,

ment with

these points as end points

is called the latus rectum of the parabola
its length is the numerical
;

value of 4

a.

Since the >>-axis meets the parabola in the origin counted doubly,
p IG
it is the tangent to the parabola at
its vertex. From (33.3) it is seen that a parabola is characterized geometrically as follows
:

[33.1]

parabola

distance

is the

from a

locus of a point

whose
four times the
and the focus of the
the square of

line (its axis) is equal to

product of the directed distances of

tangent at the vertex, this tangent being

the perpendicular to the axis at the vertex.

parabola from

its

In consequence of this theorem an equation of a parabola is

determined by equations of its axis and of the tangent at the
vertex, and the distance of its focus from this tangent (see
Exs.

7, 8).

175

The

Conies. Locus Problems

[Chap. 4

EXERCISES
Find the focus and directrix of the parabola y 2
points in which it is intersected by each of the lines
1.

2x-y-l=Q,
Draw

2.

and

x,

2x-;y+l=0.

the graph of each of the following parabolas, and

directrix

and the

its

focus

a.

= 12 x.
= -4x.

y2

b.y*

c.

x2

d.x*

= 8 y.
= - 12 y.

3. Denoting by V and P the vertex and a representative point on

a parabola, show that if P is the point of the line containing VP such
that VP' = tVPj where / is any constant not equal to zero, the locus
of P' is a parabola and determine the relation between the foci and
1

directrices of the
4.

two parabolas.

Find an equation of the parabola whose directrix is the *-axis

focus is the point (0, 4), and find a transformation of co-

and whose

new

ordinates so that in the

is

;t'/-system the equation of the parabola

of the form (33.3).

5.

Prove that each of the following

with focus at the origin

a.

b.

=
=

4 a(x
4 a(y

is

an equation of a parabola

+
+

a).

c.

a).

d.

x2

= - 4 a(x = - 4 a(y -

a).
a).

6. Find an equation of the parabola with vertex at the point (1,2)

and focus at the point ( 1, 2).

7.

Given the

lines

and x

+ 3 = 0,

find

an equation of the

parabola of latus rectum 8 which has these lines for axis and tangent
at the vertex, when the first line is the axis and the curve is to the right
of the second line

below the
8.

when the second

line

is

the axis and the curve

Find an equation of the parabola with the

and 2

x+

is

first line.

lines x

>>

-f 1

and tangent at the vertex respectively,

y
with latus rectum of length 6, and which lies below the tangent at the
3

for axis

vertex.

Find an equation of the parabola whose directrix is the line

and whose focus is (1,
What are equations
2).
of its axis and of the tangent at its vertex ?
9.

3*-4.y-l=0

176

Tangents and Polars

Sec. 34]

34. Tangents

The

and Polars

coordinates of the points of intersection of the line

(34.1)

= mx + h

with the parabola (33.3) are found by solving their equations

Substituting the expression (34.1) for
we get

simultaneously.
(33.3)

and

in

collecting terms,

(34.2)

m 2x 2 + 2(mh -2a)x + h 2 = Q

as the equation

whose roots are the ^-coordinates of the points

When m = 0,

of intersection.

that

to the axis of the parabola, there

2

m^

-T'h)- When
(h

0,

when

we have from

latter case

two

we say

is

parallel

mh)

(34.2)

m2

According as the quantity under the radical

tive, there are

the line

one point of intersection,

(0

~~

(34 ' 3^'

^

is,

is

is

positive or neganone in the

real points of intersection or

that the points of intersection are conjugate

imaginary.
When the line (34.1) intersects the parabola in two real
points, the ^-coordinate of the mid-point of the segment joining
these points, being one half the sum of the two values of x in
(34.3), is (2 a
mh)/m2 and the ^-coordinate of the mid-point
;

found, on substituting this value of x in (34.1), to be 2 a/m.

Since this value does not depend upon the value of h in the equais

tion (34.1) of the line,

[34.1]

we have

The mid-points of a set of parallel chords of a parabola

on a line parallel to the axis of the parabola.

lie

When

the quantity under the radical in (34.3) is equal to

= a/m, the two points of intersection
is, when h
coincide and the line is tangent to the parabola at the point

zero, that

'

(see

12 following equation (12.8)). Hence

177

we have

The

Conies. Locus Problems

[Chap. 4

m different from zero the line

y = mx + m
'
the parabola y 2 = 4 ax, the point of tangency

For every value of

[34.2]

(34.4)
is

tangent to

being
6

(jL2

\m

\.
I

The tangent (34.4) meets the *-axis

2
T(- a/m 0), and the length of the

(see Fig. 29) in the point

segment TF of the axis is a( 1 H

i \

^J

distance from the focus F to

the point of contact P(a/m 2 2 a/rri)
is of the same length, as is readily

The

Hence

shown.

FTP

is

an

isosceles

triangle with vertex at F.

the

tangent makes equal

Since
angles

with the #-axis and any line parallel

to

[34.3]

it,

FIG. 29

we have

to a parabola at any point P makes equal

line joining the focus to P and with the line
with
the
angles
P
parallel to the axis of the parabola.
through

The tangent

When a parabola

is

revolved about

its axis,

the surface so generated

called a paraboloid of revolution, and the axis of the generating

parabola is called the axis of the paraboloid. Each plane section of
is

the surface by a plane containing the axis is a parabola, and the

parabolas have the same focus. A mirror with such a curved surface
is called parabolic.
Recalling from physics that the angle of reflection from a mirror is equal to the angle of incidence, we have from
Theorem [34.3] that rays of light parallel to the axis of a parabolic
mirror are reflected to the focus, and, conversely, when a light is placed
at the focus of a parabolic mirror, the rays emanating from it, upon
reflection by the mirror, emerge parallel to the axis. This phenomenon
is given practical use in reflecting telescopes and automobile headlights.
If

we multiply equation

in the

form

(34.4)

178

by 2 a/m and

write the result

Tangents and Polars

Sec. 34]

this equation of the tangent becomes in terms of the point of

contact Xi = a/m 2 yi
2 a/m

(34.5)

When

(xi, y\) is

equation of a line.

any point

The

of the plane, equation (34.5)

is

an

line so defined is called

the polar of
the point (x\, y\} with respect to the parabola, and the point
(xi, ;yi) is called the pole of the line. In particular, the polar of
a point on the parabola is the tangent at the point. We shall
now find the geometric significance of the polar of a point not
on the parabola.
If we solve (34.5) for x and substitute the result in (33.3),
we obtain the quadratic

2 yiy

+ 4 axi = 0,

whose two roots are the ^-coordinates of the points of

section of the line and the parabola. These roots are

= yi

Vyi

inter-

4 axi.

line meets the parabola in two real points or in two

2
4 axi is a positive
conjugate imaginary points according as yi
or negative number, that is, according as the point (x\, y\)
lies outside or inside the parabola. To prove the latter statement
we observe that when a is positive and x\ is negative, the quantity under the radical is positive and the point P\(x\, y\) lies

Hence the

to the left of the y-axis, which, as previously shown, is the

tangent to the parabola at its vertex, and consequently the point
For x\ positive the line x
lies outside the parabola.
x\

meets the parabola in two points (xi, y 2 ) and (xi,

^2), where
2
2
=
or nega4
is
and
4
ax\
axi
0,
2
positive
y\
consequently
y
2
2
that is, active according as yi is greater or less than y 2
,

cording as PI

outside or inside the parabola.

any point on the polar of (xi, y\),

is

If (#2, 1X2) is

y&2

we have

= 2 a(x 2 + *i).

In consequence of the fact that

xi, y\

and # 2 y 2 enter sym,

metrically in this equation, it follows that the point (x\ y\)

is on the polar of the point (* 2 ^2), namely, the line
9

y 2y

= 2a(x + x 2
179

).

The

Conies. Locus Problems

[Chap. 4

Hence we have
// the polar of a point (xi, y\) with respect

[34.4]

passes through the point

passes through the point

By means

of this theorem

we

yz)>

(*2,

to

a parabola

the polar of (*2, ^2)

(x\, yi).

derive the geometric significance

of the polar of a point (x\ y\) lying outside the parabola.

has been shown that the
9

It

polar of this point intersects the parabola in two

real points.

theorem

By

the above

(xi, y\) lies

on the

polars of these two points,

but these polars are the tan-

SK/

gents to the parabola at

these points. Hence we have
(see Ex. 13)

FIG. 30

The polar of a point outside a parabola

[34.5]

is the line

joining

the points of tangency of the two tangents to the parabola

through the given point.

We

just shown indirectly that through any point

outside a parabola two tangents to the parabola can
be drawn. In order to obtain equations of these tangents, we
substitute xi, y\ for x, y in (34.4) and obtain the quadratic

have

(*it y\)

m 2xi

my +
i

= 0,

whose two roots are the slopes of the two tangents.

These

roots are

which are real and distinct when the point (xi, y\) is outside
in (34.4), we
the parabola. Substituting these values of
obtain the desired equations of the two tangents.

for the parabola y 2 = 12 x and the outside point

values
of
the
are 1 and
3, and equations of the respec1, 2)
(
and 3* >>-fl = 0.
tive tangents are x - y + 3 =

For example,

180

Tangents and Polars

Sec. 34]

EXERCISES
1.

Find an equation of the tangent with the slope 3 to each of the

parabolas

= g Xt

_.

y2

_4x

Also find equations of the tangents to the first parabola at the points
for which x = 2, and of the tangents to the second parabola from the
point
2.
is

(3, 2),

and the angles between the tangents

in the latter

Find an equation of the tangent to the parabola y 2

perpendicular to the line

3* +

1=0;

2jf

two cases.

= 5 x which

also the coordinates

of the point of contact.

3. The line segment joining the focus of a parabola to any point P
of the parabola is called the focal radius of P. Find the coordinates
of the points of the parabola (33.3) whose focal radii are equal in

length to the latus rectum.

4.

Show

x2

that

= 4 ay

is an equation of a parabola with the ;y-axis for the axis of the parabola
and the *-axis for tangent at the vertex. What are the coordinates of
the focus and an equation of the directrix? Show that the tangent

with the slope

m is

and the point of contact

5.

For what value of

= 6y?

y
is

_.

(2

mx

am,

am 2

__

am

c is the line

2
).

2x + 3y + c = Q

tangent to the

2
parabola x

6. Show that a tangent to a parabola at a point P meets its axis

produced at a point whose distance from the vertex is equal to the distance of P from the tangent to the parabola at the vertex.
7.

The normal

to a curve at a point

pendicular to the tangent at P. Show that

to the parabola (33.3) at the point (xi, yi)

is by definition the peran equation of the normal

is

8. Show that the normal to a parabola at a point P meets the axis

of the parabola in a point whose distance from the projection of P upon
the axis is one half the latus rectum of the parabola. This segment of
the axis is called the subnormal for the point.
9.

Show

that the point of intersection of a tangent to a parabola

to the tangent through the focus lies on the

and the perpendicular

tangent to the parabola at the vertex.

181

The
10.

Show

Conies. Locus Problems

[Chap. 4

that the segment of a tangent to a parabola between the

point of tangency and the directrix subtends a right angle at the focus.
11. Show that the tangents to a parabola at the extremities of any
chord through the focus meet at a point on the directrix. What relation does this result bear to Theorems [34.4] and [34.5]?
12. Show that the chord joining the points of contact of any two
mutually perpendicular tangents to a parabola passes through the

focus.
13. Let /i and 12 be two chords of a parabola through a point P
within the parabola, and let P\ and P* be the points of intersection of
the tangents to the parabola at the extremities of l\ and /2 respectively.
Show that the line PiP2 is the polar of P.

35. Ellipses

and Hyperbolas

In accordance with the definition in 32 the locus of a point

distance from a fixed point, the focus, is equal to a
positive constant e times its distance from a fixed line, the
directrix, is an ellipse when e < 1 and a hyperbola when e > 1.
In
32, on taking the ;y-axis for the directrix and the point
on the *-axis for focus, we obtained the equation
(jfe, 0)

P whose

(35.1)
If

we

(1

- e2 )x2 -2kx + y 2 = -k2

divide this equation

by

(1

e2 )

and complete the square

by adding k2 /(l
the
result
equation,
may be written

e2 ) 2

of the terms in x

If,

to both sides of the

in order to obtain simpler forms of equations of ellipses and

we make the translation of the axes defined by

hyperbolas,
(35.2)

'=-rb' *-*

an equation of the conic in terms of

(35.3)

x'

and

y' is

Sec. 35]

Ellipses

and Hyperbolas

The
(k,

0)

focus and directrix, which in the ry-system are the point

and the line x = 0, are given in the ^'/-system by

(35.4)

respectively, as follows from (35.2).

From the form of equation (35.3)

it is

seen that the

/-axes are axes of symmetry of the conic

(see

2)

x'-

and

that

is,

a point of the conic, so also are (x',

y') and
This means geometrically that the two points in
*', y').
(
which a line parallel to the /-axis meets the conic are at equal
distances from the #'-axis, and on opposite sides of the latter
and likewise the two points in which a line parallel to the
#'-axis meets the conic are at equal distances from the /-axis,
and on opposite sides of the latter. Also it follows from (35.3)

if

(#',

/)

is

that the #'-axis meets the conic in the two points

and

>

__

ke
2

OJ

e
(

-z

^>

'

the points with these coordinates are the end

points of a line segment of which the origin of the ^'/-system

is the mid-point. Hence an ellipse or a hyperbola has two axes
of symmetry, called the principal axes, the one (the #'-axis)

being perpendicular to the directrix, and the other (the /-axis)

being the line parallel to the directrix and through the mid-point
of the segment of the former axis (the #'-axis) whose end points
are points of the conic.
f
It is seen also from (35.3) that if (x
y') is a point of the
conic, so also is (
x',
/) that is, any chord of the conic
through the origin of the ^'/-system is bisected by the origin.
,

Consequently the intersection of the principal axes of an

a hyperbola is a point of symmetry of the curve.
It is called the center of the conic. Since there is no center of
a parabola in this sense, ellipses and hyperbolas are called

ellipse or of

central conies.
If, in order to simplify the equation (35.3),
quantity a by

(35.5)

and drop the primes

we

define the

'-f^?'
in equation (35.3), that

183

is,

use x and y for

The

Conies. Locus Problems

[Chap. 4

new system (since only this new system

used in what follows), this Aquation may be written

coordinates in the

<

is

35 6>
-

From

this equation it follows that the curve meets the axis

through the focus (the x-axis) in the two points A'( a, 0) and

A(a, 0). Also from (35.4) it follows that the coordinates of the
focus and an equation of the directrix are respectively

x+-a = 0.

F'(-ae,0),

(35.7)

Since, as remarked before, an ellipse or a hyperbola is symmetric with respect to its principal axes, which in the present
coordinate system are the x- and ;y-axes, it follows from considerations of symmetry that for each of these curves there is

a second focus and a second directrix symmetric with respect to

the
to F' and the directrix x + - =
;y-axis

are given

respectively.

They

by
x

F(ae, 0),

(35.8)

respectively.

As a

- -=
c

further proof of this result, we observe that

this focus and directrix is

an equation of the conic with

which reduces to (35.6), as the reader should verify.

Thus far we have treated ellipses and hyperbolas simultaneously, but in proceeding further

it is advisable at times to treat

In such cases the treatment will be given in
parallel columns, that on the left pertaining to the ellipse and
that on the right to the hyperbola, while statements which
apply equally to both types of conies will not be separated.

them

separately.

The quantity
is

a 2 (I

e2 )

appearing in equation (35.6)

positive for an ellipse, since

<

1.

number
(35.9)

Consequently a
b

is

ft

defined

by

= a*(l - e2

real

negative for a hyperbola,

> 1. Consequently a
real number b is defined by
is

since e

(35.9')

),

184

ft

Sec. 35]

Ellipses

and Hyperbolas

in terms of which, as follows

from

(35.6),

equations of the re-

spective curves are

j^+=l,

(35.10)

in

which

it is

(35.10')

s-g-1,

understood that a and b are positive numbers.

t

FIG. 31'

From
an

(35.10)

ellipse

it

follows that

meets the

jy-axis

in the points B'(Q ~- b)

B(Q, b). Also it follows
r

and
from

(35.9) that b < a. Accordingly

the line segment A' A of the
principal axis containing the
foci is called

the major axis,

From

(35.10') it follows that

hyperbola does not meet the

>>-axis.

The segment A 'A of the

principal axis containing the

foci is called the transverse axis

of the hyperbola.

The segment

with end points '(0,

b) and
B(Q, b) of the other principal

and the segment B'B of the

axis,

other principal axis, that is,

the perpendicular to the major axis through the center of

the center perpendicular to the

transverse axis, is called the
conjugate axis of the hyperbola.

the

ellipse,

the minor axis.

The numbers a and

b are called

that

is,

the line through

The numbers a and

are

called the semi-transverse

and

the semi-major and semi-minor

semi-conjugate

axes respectively, in the sense

that they are the lengths of

tively.

these segments. The points

A' and A, the extremities of

axes
respecpoints A' and A y
the extremities of the transverse axis, are called the ver-

The

the major axis, are called the

vertices of the ellipse.
(See

tices of the hyperbola.

The
geometric significance of the
end points B' and B of the con-

Fig. 31.)

jugate axis

is

(See Fig. 31'.)

185

given in

37.

The

Conies. Locus Problems

[Chap. 4

From

equations (35.9) and (35.9') we have that in the respective cases the eccentricity is expressed in terms of the
semi-axes as follows
:

(35.H

(35.11)

which emphasizes the

fact that e

<

1 for

an

ellipse

and

>

a hyperbola. If we denote by c the distance of either focus

from the center of each curve, then c = ae as follows from
(35.7) and (35.8), and we have from (35.11) and (35.11')
for

(35.12)

When
and

a2

(35.12')

these results are considered in connection with Figs. 31

we see that

31',

a circle with center at either

a circle with center at the cen-

end of the minor axis of an

ellipse and radius equal to the
semi-major axis meets the

ter of a

major axis in the two

foci.

hyperbola and radius

the hypotenuse of

to

equal
a right triangle whose legs
are the semi-transverse and

semi-conjugate axes meets the

transverse axis in the two foci
(see Fig. 35).

We

shall

now

find

of the focal radii of

an important relation between the lengths

any point on these curves, that is, the

lengths F'P and FP.

The absolute distances of a point P(x, y) on either curve
from the directrices are the numerical values of

(35.13)

The
of

x-Z,

+ Z-

focal distances of P, being e times the absolute distances

from the

(35.14)

directrices, are the

numerical values of
ex

ex

186

Sec. 35]

Ellipses

and Hyperbolas

For an ellipse the first of (35. 13)

is negative and the second positive, whatever be P, as is seen
from Fig. 31. Hence

FP = a-

from which
(35.15)

F'P

ex,
it

= a + ex,

follows that

FP +

F'P

For a hyperbola both of the

quantities (35.13) are positive
when P is on the right-hand

branch of the hyperbola (see

Fig. 31'), and both are negative when P is on the left-hand
branch.

= 2a.

Hence

in

the

first

case (35.14) are the focal radii,

and in the second case they are

the negatives of these quantities.

Accordingly we have

(35.150

F'P-FP =

2a,

sign or the
sign applying according as P is on the
right-hand or left-hand branch.

the

Thus we have the theorems

[35.1]

The sum of the focal

radii of any point of an
ellipse is equal to
stant,

the

major

axis.

length

[35.1']

The numerical value of

the difference of the focal
radii of any point of a

a conof the

hyperbola

is

equal

to

constant, the length of the

transverse axis.

These results make possible a continuous construction of the

central conies, as distinguished from a point-by-point construction.
For the construction of an ellipse two

thumbtacks are fastened through a sheet

and a loop of string is placed
loosely around the tacks; the string is
then drawn taut by the point of a pencil.
As the pencil is made to move, the string
being held taut, it describes an ellipse with
of paper,

the tacks at the

foci.

In fact,

if

the dis-

FIG. 32

tance between the foci is denoted by 2 c and the length of the loop by
2 a H- 2 c, the sum of the focal radii is 2 a. Since c = ae, an ellipse
of given major axis 2 a and given eccentricity e is described when
the tacks are set at the distance 2 ae apart and the loop is of length

2 0(1

+ e).
187

The

Conies. Locus Problems

[Chap. 4

For the construction of a hyperbola one end of a stick of length d

pivoted at a point F' on the paper, and at the other end of the
stick is fastened one end of a string of
2 a
the other end of the
length d
string is then fastened by a thumbtack
is

at a point F on the paper. If the string

is held taut by pressing it with the point

P of a pencil against the stick, the pencil

describes a branch of a hyperbola as the
about the point F', since
Again, since F'F = 2 ae,

stick rotates

F'P

-FP = 2a.

a hyperbola of given transverse axis and

F
given eccentricity can be drawn by
choosing the length of the string suitably and by setting the points
F' and F at the proper distance apart.

As a result of the foregoing discussion we have that Theorems

and [35. 1'] state characteristic properties of ellipses and

[35.1]

hyperbolas, in the sense that these properties may be used to

define these central conies. The same is true of the following

two theorems, which are geometric statements of equations

(35.10) and (35. 10') respectively without regard to any coordinate system

and d% denote the distances of any point on an ellipse

major and minor axes respectively, which axes have
respective lengths 2 a and 2 b, then

// d\

[35.2]

from
the

its

f + f-

(35.16)

and any curve

'

whose points are so related to two per2a and 2b and each

is
an
other,
ellipse, whose eccentricity is

all of

pendicular line segments, of lengths

of which bisects the
2
given by e

[35.3]

(a

b 2 )/a 2

when

a>

b.

// d\ and d 2 denote the distances of any point on a hyperbola

the lines of its transverse and conjugate axes respec-

from

tively,

which axes have the respective lengths 2 a and 2

then

188

b,

Sec. 35J

Ellipses

and Hyperbolas

and any curve all of whose points are so related to two perpendicular line segments, of lengths 2 a and 2 b and each of
which

bisects the other, is

2
given by e

(a

+ b?)/a2

a hyperbola, whose

eccentricity is

For example, an equation of the ellipse whose major and

minor axes are of lengths 6 and 4, and are on the lines x + 1 =

and

>>

respectively,

is

The conies were studied extensively by the Greeks from a purely

geometric point of view, and many of their properties which now are
derived by means of coordinate geometry were discovered by geometric reasoning (see 38). Following the adoption of the Copernican
theory of the planetary system, Kepler by laborious calculations from
observational data showed that the orbits of the planets are ellipses
with the sun at one of the foci. This enabled Newton to discover his

law of gravitation, so that now the orbits of the planets are obtained
readily from Newton's law by the use of coordinate geometry and the
calculus. We have seen that for an ellipse each focus is at the distance ae from its center, and thus the eccentricity determines the
departure of the focus from the center. For the earth e is about 1/60,
so that its orbit is almost circular. For the recently discovered planet
Pluto e is about 1/4, the semi-major axis of its orbit is nearly 40 times

and

its period of revolution is approximately

of the comets are practically parabolas, some
of them being ellipses of eccentricity almost equal to 1. In fact, for
the celebrated Halley's comet e
.98, the semi-major axis is almost

that of the earth,

The paths

250 years.

18 times that of the earth's orbit, and its period is 75 years

to the earth's neighborhood has been recorded many times.
;

its

return

EXERCISES
1.

that equation (35.10) in which b > a is an equation of an

with semi-major axis b and semi-minor axis a, that (0, be) and
= b/e are the directrices, where
be) are the foci, and that y

Show

ellipse
(0,

e* =
2.

(b

a*)/b*.

Find the
a.

b.

vertices, center, foci,

+ 4 y 2 = 12.
2
9 x + 5 y 2 = 45.
3 x2

and

directrices of the following

c.

d.

189

- 4 y 2 = 20.
2
9 x - 16 y* = 12.
5 x2

The
3.

Conies. Locus Problems

Find an equation of an ellipse whose

(a) when its minor axis is 8

foci are

[Chap. 4
3, 0)

and

(3, 0),

(b)
(c)
4.
(0, 3),

5.

when
when

its

major axis

its

eccentricity

is

twice

is

minor axis

its

2/3.

Find an equation of a hyperbola with

and eccentricity 3/2.

directrix

l,

focus

Find an equation of a hyperbola whose transverse and conjugate

and

axes, of lengths 4

6,

are on the lines x

+3=

and y

respectively.
6.

Find an equation of the ellipse whose major and minor axes, of

and 6, are on the lines 3 # 4 jy + 1 = and 4*-f3jy-f-2 =

lengths 8

respectively.
7. Find an equation of the locus of a point the sum of whose distances from the points (c, 0) and ( c, 0) is 2 a \ also the locus when
the difference of these distances is 2 a.

8. By definition the latus rectum of an ellipse (or a hyperbola) is the

chord through a focus and perpendicular to the major (or transverse)
find its length in terms of a and b.
axis
;

9.

Show

that,

sponding directrix

What
and

is

if

the distance of a focus of an ellipse from the corree 2 ).

the semi-major axis is given by a
he/ (I

is h,

same

the semi-transverse axis of the hyperbola with the

focus

directrix?

10.

Show

that the quantity

^- + ^-

1 is

positive or negative

according as the point (xi, y\) lies outside or inside the ellipse (35.10).
What is the similar theorem for the hyperbola ?
11.

In what sense

is

circle

an

ellipse of eccentricity zero?

12. What is the eccentricity of a rectangular hyperbola, that is, one

whose transverse and conjugate axes are of equal length? What is
an equation of a rectangular hyperbola?

Prove that the distance of any point of a rectangular hyperbola

Ex. 12) from the center is the mean proportional to its distances

13.
(cf.

from the
14.

foci.

Find an equation of the

the line

3^-4^-5 =

ellipse

as directrix,

with the point

and

(2,

1) as focus,

eccentricity 1/2

also find

an equation of the hyperbola with this focus and directrix, and ecthe transverse
centricity 2. What is the major axis of this ellipse
;

axis of this hyperbola?

(See Ex.

9.)

190

15.

Conjugate Diameters and Tangents

Sec. 36]

Given the equation

by means of

(29.5)

and

tangular coordinates

and that

this is

- in
polar coordinates;

cos 6

show

(29.6) that an equation of the locus in rec-

is

an equation of a conic of eccentricity

+ -e =

the origin and the line x

rectum is of length 21.

Show

for directrix.

with focus at

that the latus

16. Consider in space of three dimensions a circle of radius a with

center at the origin, and lying in a plane which cuts the
ry-plane in
the x-axis and makes the angle 6 with this plane ; show that the or-

thogonal projection of the circle upon the ry-plane

the equation
2
^2
"a

a 2 cos 2

We

turn

now

an

ellipse

with

1'

Is the orthogonal projection of any circle

to the plane of the circle an ellipse?

36. Conjugate Diameters

is

upon a plane not

and Tangents

parallel

of Central Conies

to the consideration of sets of parallel chords

and to the finding of equations of tangents.

of a central conic

The

^-coordinates of the points in which the line

y = mx

(36.1)

+h

meets the curves with equations (35.10) and

(36.2)

(35.10') are given

by

(36.2')

- a 2 mh d= ab Vft 2 + a 2 m 2 - h 2
a2 m 2

+ b2

_ - a 2 mhab^/b 2 - a 2 m 2 + h'
a2 m2 - b2

these expressions and (36.1) we find as the coordinates

XQ, yo of the mid-point of the line segment connecting the
points of intersection

From

As

a*mh,

yo

'

a2 mh,
-2

- hb2

a*m*-,

h takes on different values for which the points of intersec-

tion are real,

remaining fixed, that is, for different parallel
chords, these expressions give the coordinates of the mid-point

191

TKe
of each chord.

By

on the respective

Conies. Locus Problems

substitution

we

[Chap. 4

find that these are points

lines

,_.

ft

(36.30

(36.3)

Each, passing through the origin, which is the center of the

curve, and hence being a line through the center of a central
conic, is called a diameter. If in each case we denote by m' the
slope of the line (36.3) or (36.3'),

mm'

(36.4)

we have

=-

FIG. 34

FIG. 34'

and m', it follows

Since these equations are symmetric in
that the mid-points of the chords parallel to the lines (36.3)
and (36.30 are on the line y = mx in each case. Two lines
through the center of an ellipse, or of a hyperbola, whose slopes
are in the relation (36.4), or (36.4'), are called conjugate diameters.
[36.1]

Hence we have
an ellipse, or of a
locus of the mid-points of chords parallel

Either of two conjugate diameters of

hyperbola,

is the

to the other.

do not apply when m = 0, that

when the chords are parallel to the major axis of an ellipse

Equations
is,

(36.3)

and

(36.3')

or the transverse axis of a hyperbola. But since the axes in

each case are axes of symmetry, for each curve either axis is
the locus of the mid-points of chords parallel to the other.
Consequently the axes are said to be conjugate.

192

Conjugate Diameters and Tangents

Sec. 36]

We leave to the reader the proof of the following statements,

with the aid of (36.2) and (36.20, calling his attention to the
34 of similar results for the parabola
derivation in
:

The
(36.5)

are

The

parallel lines

y=

y = mx

(36.5')

tangent

to

the

ellipse

(35.10) at the points

(36.6)

parallel lines

are tangent to the hyperbola

(35.100 at the points
(36.6')

On

denoting by x\, y\ the coordinates of either point, the equation of the tangent at the point may be written in the form
(36.7)

Si

*L_m
=
a

( 36.7')

b2

We now establish the following theorem, which shows that

the central conies possess a property somewhat analogous to
the property of the parabola stated in Theorem [34.3]
:

[36.2]

P of an ellipse, or of a hyperangles with the tangent to the conic ai P.

The focal radii of any point

bola,

make equal

We prove this theorem by using straightforward algebraic

methods, which illustrate the power of coordinate geometry.
In order to handle the ellipse and hyperbola at the same time,
we use equation (35.6), that is,
(36.8)

a2

'

a 2 (l

e2 )

1.

equations (36.7), (36.70 and (35.9), (35.90 it follows

that an equation of the tangent to either conic at the point

From

PI(XI, yi)

is

rr
xx
a

VVi

2
.

193

The
Hence

direction

Conies. Locus Problems

numbers of

this line are (see

[Chap. 4

page 30)

Since the foci are F'( ae, 0) and F(ae, 0), direction numbers
of the focal radii are (see Theorem [6.1])
W2

= xi

v2

ae,

= jyi,

+ sign holds for F'Pi and the sign for FPi. We

in the expression for 2, in order that the result
keep the sign
later obtained shall apply to both F'Pi and FPi.
In order to apply formula (6.8), we first make the following
where the

calculations

^^
2
Substituting x\ and y\ for x and y in (36.8), solving for yi and
substituting in the last of the above equations, we have
,

U2

+ v2 =
2

(xi

= a2

+ a 2 (l - e2 - *i 2 (l - e2
2 xiae + e Xi 2 = (a + Xie) 2
ae)

When

these expressions are substituted in (6.8), where now

the angle between the tangent at PI and a focal radius, we
obtain
is

+ *i 2 (l - e2
Since the factor

+ x^),

which the

2
)

sign refers to the

radius FP\, does
to
the
focal
sign
not appear in the result, it follows that the angles which these
focal radii make with the tangent to the conic at the point PI
focal radius F'Pi

in

and the

are equal, as was to be proved.

As a physical

application of this theorem we have that a ray of

emanating from one focus of an ellipse and meeting the ellipse
at a point P would be reflected by a mirror tangential to the ellipse
at P into a ray which would pass through the other focus.
light

194

Conjugate Diameters and Tangents

Sec. 36]

EXERCISES
1.

Find equations of the tangents to the hyperbola

5 x2

which are perpendicular to the

2.

4;y

line

10

3y

= 0.

Find equations of the tangent and normal to 3 x 2

at the point

(2,

4 y2

Find the extremities of the diameter of x 2 + 2 y 2

conjugate to the diameter through the point (1, 1).
3.

4.

point
5.

Find the two tangents to the

(2,

1).

ellipse

5 x2

=4

+ 9 y 2 = 45

which

is

from the

2).

Show

that equation (36.7)

when

the point

outside

is

y\)

(xi,

the ellipse is an equation of the line through the points of tangency

of the two tangents to the ellipse from (x\, y\). This line is called

the polar of the point

part of 34).
6.

Do

(x\ t y\)

the results of

with respect to the ellipse (see the latter

34, Ex. 13 hold also for

an

ellipse

and

for

a hyperbola ?
7.

Show

that any tangent to an ellipse meets the tangents at the

from the major axis

vertices in points the product of whose distances

is equal to the square of the semi-minor axis.
8.

Show

(35.10')

is

by a rotation

that

of the axes (28.8) for

= 90,

equation

transformed into

and that in the #';y '-system the coordinates of the vertices are (0, a)'
and equations of the direcae)', (0, ae)
(0, a)' and of the foci, (0,
9

trices are

vertices, foci,
9.

= 0.

and

Given an

Apply

this result to the determination of the

directrices of the hyperbola 3 x 2

ellipse (35.10), for

which

a>

b,

4 y2

and the

12.

circle

with

center at 0(0, 0) and radius a, through a point (x, y) on the ellipse

draw the line perpendicular to the *-axis, and denote by P the point
in which this line meets the circle, and by 6 the angle which the radius

OP makes

with the *-axis.

Show

that x

= a cos 8,

therefore are parametric equations of the ellipse.

195

y=

b sin

0,

which

The

Conies. Locus Problems

[Chap. 4

Find the relation between the lengths of two conjugate diaman ellipse and also between those of a hyperbola. Show
that when a hyperbola has a pair of conjugate diameters of equal
10.

eters of

length,

it is

rectangular.

Show

that the product of the distances of the foci from any

tangent to an ellipse is equal to the square of the semi-minor axis.
11.

of

12.

Show

an

ellipse

that the point in which a perpendicular from either focus

upon any tangent meets the latter lies on a circle with

center at the center of the ellipse

and radius equal

to the semi-major

axis.

that the sum of the squares of the reciprocals of two perof an ellipse is constant.
diameters
pendicular
13.

Show

14.

Show

x2

that

*rZ
for all values of

conies, all

less

than a 2 except

of which have the

same

foci.

b2

defines a system of central

This system is called a system
,

of confocal conies.
15. Show that through each point in the plane not on either coordinate axis there pass two conies of a confocal system, one being
an ellipse and the other a hyperbola, and that the tangents to these
curves at the point are perpendicular to one another.

37. Similar Central Conies.

The Asymptotes

of a

Hyperhola.

Conjugate Hyperbolas
Consider in connection with equation (35.10) the equation

%+ i = k

(37.1)

where k

is

some constant.

written

If
2

X
7
a *

= ka 2

is

>

positive, this equation

+ &'y22 = 1

may be

'

kb 2 and consequently is an ellipse with

the same center (the origin) as the ellipse (35.10). Furthermore
if (x, y) is a point of (35.10), then (Vkx, V&jy) is a point of the
ellipse (37.1), and corresponding points lie on a line through

where a 12

the center; that

b'

is,

the ellipses (37.1) for positive values of k

196

Similar Central (Jonics.

Sec. 37]

Asymptotes

are similar (see

35, Ex. 1 it follows
32). From (35.11) and
that the ellipses (37.1) for all positive values of k have the same
eccentricity (see Theorem [32.1]).
in (37.1), we say that
When k

ellipse, since

point

there

is

it is an equation of a
only one real solution of the equation,

namely, x = y = 0. However, in this case y =

that is,
V^T~x,
a

we have two conjugate imaginary lines. When k < in (37.1),

we say that it is an equation of an imaginary ellipse, since
there are no real solutions of the equation. Hence we have
[37.1]

>

is an equation of a family of
same center and respective principal axes; equation (37.1) is an equation of a point ellipse
when k = Q, and of a family of imaginary ellipses when

Equation (37.1) with k

similar ellipses having the

k<0.
Similarly we consider in connection with (35.10') the equation

(37.2)

=*

is some constant. Proceeding as in the case of equation

we have that for each positive value of k equation (37.2)

where k
(37.1),
is

an equation of a hyperbola

equation (35.10') arid

principal axes.

We

consider next the case

<>
it is

when

= 0,

that

is,

the equation

= 0.

(37.3)

When

similar to the hyperbola with

having the same center and respective

iT2-f-2
this

equation

is

written in the form

(M)(i

iH

seen that the coordinates of a point on either of the lines

are a solution of (37.3), and, conversely, any solution of (37.3)

is a solution of one of equations (37.5).
Consequently (37.3) is

an equation of the two

lines (37.5) (see

Ex.

7).

We

the relation of these lines to the hyperbola (35.10').

197

shall

study

The

Conies. Locus Problems

[Chap. 4

Since a and b are understood to be positive, the first of the

bx = 0, passes through the origin and
is, ay
lies in the first and third quadrants. In order to find the dis-

lines (37.5), that

tance d from this line to a point (x\, y\) on the hyperbola in

the first quadrant, we observe that it follows from (35.10') that
the ^-coordinate of this point

is

terms of XL Hence by Theorem

given by y\

=-

[8.1] the distance

a 2 in

is

given by

~
( 37.

rf

6)

-*i)

- a2 - *i)(V*i 2 - a2 +

As

Xi

becomes larger and

larger, that

is,

as the point

(xi, yi)

taken farther out on the hyperbola, this distance becomes

smaller and smaller, and we say that the hyperbola approximates the line for very large values of x. Since the center of
is

the hyperbola, that

is,

the origin in this coordinate system,

is

a point of symmetry of the hyperbola, it follows that the same

situation exists in the third quadrant for numerically large
negative values of x. The reader can show that the second of
the lines (37.5) bears a similar relation to the hyperbola in the
second and fourth quadrants.

The

lines (37.5) are called

the asymptotes of the hyperbola; they pass through the

center of the hyperbola and are
equally inclined to its axes. For

a rectangular hyperbola (see

35, Ex. 12) the asymptotes
FIG. 35
are mutually perpendicular,
and only for such a hyperbola.
The lines (37.5) by a similar argument are the asymptotes
also of the hyperbola
(37.7)

-5

~w=
198

1-

Asymptotes. Conjugate Hyperbolas

Sec. 37]

The two hyperbolas (35.10') and (37.7) are said to be conjugate

to one another. They lie with respect to their common asymptotes as shown in Fig. 35, and the transverse axis of either
is

hyperbola

the conjugate axis of the other. Equation (37.2)

is an equation of a hyperbola similar

any negative value of k

to (37.7). Hence we have
for

Equation (37.2) with k

[37.2]

an equation of two families

is

of hyperbolas, the one family for k

transverse axis of each hyperbola,

<

bola;

>

having the x-axisfor

and

the other family for

having the y-axis for transverse axis of each hyperall of these hyperbolas have the same asymptotes,

= 0.

given by (37.2) for k

The proof of the last part of this theorem

reader as an exercise.

is

left

to the

When

the values b/a and

in
b/a are substituted for
of
the
the
asymptotes. However, in
get
equations
(36.5'),
the expressions (36.6') for the coordinates of the points of con-

we

become equal

tact the denominators

m, and consequently

to zero for these values of

cannot propit
is said that
Sometimes
the
to
be
asymptotes.
applied
erly
the asymptotes are tangent to a hyperbola at infinity, but we
prefer the statement that the hyperbola approximates the
asymptotes as distances from the center become very large.
this treatment of tangents

EXERCISES
1.

Find the center,

2
conjugate to 3 x

2.

Show

vertices, foci,

4 y2

and

directrices of the hyperbola

12.

that the circle in Fig. 35 meets the transverse axis of each

its foci, and meets each hyperbola in points on the direc-

hyperbola in

trices of the other

3.

hyperbola.

Show by means

of a transformation (28.8) that when the rec2 a 2 is referred to its asymptotes as

2
2
y
tangular hyperbola x
2
=
a
axes its equation is xy

4.

Show

that the eccentricities e

jugate are in the relation

and

= !.
C

199

e'

of a hyperbola

and

its

con-

The

Conies. Locus Problems

[Chap. 4

Find the points where the diameter of x 2 - 2 y* = 2 conjugate

to the diameter through (2, 1) meets the conjugate hyperbola.
5.

Find equations of the tangents of slope +

and the points of contact
3 y 2 -f 12 =
the slope are there tangents to this hyperbola ?
6.

4 x2

7.

Show

the hyperbola
of

1 to

what values

for

that the graph of the equation

fax

b\y

Ci

+ b2y + c2 )
and a 2 x + b*y

Ci)(a2 x

= 0.

is

the two

is

8. Show that if the angle between the asymptotes of a hyperbola

denoted by 2 a, the eccentricity of the hyperbola is sec a.

9.

lines a\%

Show

b\y

-f c 2

that the portion of an asymptote of a hyperbola included

directrices is equal to the length of the transverse axis.

between the two

10.

Show

that the distance of a focus of a hyperbola from either

asymptote is equal to the semi-conjugate axis.

11. Find the bisectors of the angles between the lines joining any
point on a rectangular hyperbola to its vertices, and determine their
relation to the asymptotes.
12. Show that the point of contact of a tangent to a hyperbola is
the mid-point of the segment of the tangent between the points in
which it meets the conjugate hyperbola.

13. Show that if a line meets a hyperbola in the points P' and P",
and the asymptotes in R' and R", the mid-points of the segments
P'P" and R'R" coincide.

14.

Show

bola from
15.

that the product of the distances of a point on a hyperasymptotes is constant.

its

Show

that the lines joining either vertex of a hyperbola to the

its conjugate axis are parallel to the asymptotes.

end points of

16. Identify

each of the

loci defined

x2
kl

when each
graph

all

of the

's

a2

k*

v2

J*

takes the values

these loci for a

1,

by the equation

= fe
0,

1,

and

draw on one

= 2.

Show that two equations (37.2) for values of k equal numeribut of opposite sign are equations of conjugate hyperbolas.

17.

cally

18. Show that the ratio of the semi-axes for each family of hyperor k <
bolas (37.2) as k >
is the same for all members of the
family. What relation does this ratio bear to the eccentricity of each

hyperbola?

200

The Conies

Sec. 38]

38.

The Conies

as Plane Sections of a

Cone

as Plane Sections of a Right Circular

and a section

Consider a right circular cone with vertex

MVN of the cone by a plane as in Fig. 36

the curve MVN on the intersection DVKL of
TT,

Cone

V is
TT

the point of
and the plane

perpendicular to TT through the

is the intersection of TT
axis OA
;

and OA, and C the point

the

bisector

meets OA.

the

of

in

which

angle

KVO

The point C

same distance

at the

is

from the element

of
OF the cone and trom the line of
symmetry KV of the curve MVN
and the perpendicular from C upon
KV is normal to the plane TT, being
r

in a

plane perpendicular to TT.

With C as center and r as radius
describe a sphere denote by F the
;

point where it is tangent to the

plane of the section and by B
the point where it is tangent to the
0V. Since the cone is right

line

tangent to
each element of the cone, and all the points of tangency are
on a circle BE. DG is the line of intersection of the plane of
circular, this sphere is

the circle and the plane TT this line is perpendicular to ED.

Let P be any point of the curve in which the plane cuts the
;

and PHO the element of the cone through P,

being
point of tangency to the sphere. Since PF and PH are
tangents to the sphere from an outside point, they are equal,
as shown in solid geometry. Since all the elements of the cone
make the same angle with the plane EHB, the line PH makes

cone,
its

with this plane an angle equal to VBD. From P we draw PG

perpendicular to DG, which being parallel to the line KFD
makes the same angle with the plane of the circle EHB as the
If we denote by Q (not
latter line, that is, the angle VDB.
shown in Fig. 36) the foot of the perpendicular from P on the
plane of the

circle,

we have
201

The

Conies. Locus Problems

PQ

PH =

sin

VBD

PQ

PG =

[Chap. 4

sin

VDB

Hence we have

The

PH
PG

PF
PG

'

sin

VDB

sin

VBD

is the complement of the angle AOB, and the

the angle which the plane of the section makes
with the plane of the circle EHB, which is a plane normal to
the axis of the cone. Since these angles do not depend in any

angle

angle

VDB

is

way upon the

from

P on

position of the point

(38.1) that

PF =
;

r(j

const.,

the curve,

and thus the curve

it

is

follows

a conic,

being the focus and DG the directrix (see

32).
For the curve to be an ellipse, the plane must intersect all
the elements of the cone that is, the angle VDB must be less
than the angle VBD, in which case the ratio (38.1) is less than
As the angle VDB is taken smaller and
1, as it should be.

smaller, the eccentricity becomes smaller and approaches the

value 0, in which case the plane is normal to the axis of the

cone and the plane section is a circle. This is in agreement

when a = b.
with (35.11), from which it follows that e =
When the angle VDB is equal to the angle VBD, that is,
when the line D VL is parallel to the element OE, in which case
the cutting plane is parallel to OE, the ratio (38.1) is equal
to + 1, and the section by the plane is a parabola.
When the angle VDB is greater than the angle VBD, in which
case the ratio (38.1) is greater than +1, the plane intersects
only some of the elements of the cone and the section is one
branch of a hyperbola, the other branch being the section of
the cone obtained by extending the elements through 0. If we
take a plane through
parallel to the cutting plane, it interwhen these are projected
sects the two cones in two elements
orthogonally upon the cutting plane, the resulting lines are the
;

asymptotes of the hyperbola, and their point of

intersection,

the projection of 0, is the center of the hyperbola.

When the conic section is an ellipse, there is a sphere above
the cutting plane which is tangent to the plane at a point F' on
that

is,

the line

VK and is tangent to the cone along a circle E',

202

H', B',

Equations of Conies

Sec. 39]

these points being on the elements through the respective

points E, H, B. Since the planes of the two circles are parallel, the cutting plane meets the plane of the second circle in

a line D'G' parallel to Z)G, the points D' and G' being on the
As in the preceding case,
lines VK and GP respectively.
PH' = PF', being equal tangents from P to the second sphere.
By the argument used above we show that F is the other focus
of the ellipse and D'G' the corresponding directrix. Moreover,
for any point P the sum of PF and PF' is equal to the length
1

of the segment of
of the two circles
;

an element of the cone between the planes

and this length is equal to the length of the

axis of the ellipse, as

(the point in which

major

VK

seen when P is taken at V or at

meets the cone again) (see Theo-

is

rem [35.1]).
As previously remarked, when the conic section is a hyperbola one branch lies on each of the two cones which are a
prolongation of one another through 0. The focus within
each branch

is

the point of contact of a sphere in each cone

similar to the one first discussed in connection with Fig. 36.

leave it to the reader to show that the difference of the

We

focal radii is equal to the segment of an element of the cones

between the planes of the two circles of tangency of the spheres
with the cones, the length of the segment being equal to the
length of the transverse axis of the hyperbola (see Theo-

rem

[35.1']).

39. Equations of Conies

to the

Whose Axes Are

Parallel

Coordinate Axes

It was remarked in
32 that an equation of a curve is determined not only by the geometric character of the curve but
also by its position relative to the coordinate axes. The equation

(39.1)

(1

was derived

- e2 )*2 + y2 - 2 kx + k2 =

32 as an equation of a conic of eccentricity e,

is the jy-axis, and the perpendicular through
the focus upon the directrix the #-axis, the focus being denoted
33 and 35 it was shown that by a suitable
by (k 0). Also in

when

in

the directrix

203

The

Conies. Locus Problems

[Chap. 4

translation of axes equation (39.1) was transformed into (33.3)

for a parabola (e=l), and into (35.10) and (35.10') for an
ellipse (e

<

and a hyperbola

1)

(e

>

1) respectively.

the directrix of a conic is taken as the #-axis and the perpendicular through the focus upon the directrix as the jy-axis,
If

and the focus

denoted by

is

y?

(39.2)

(1

(0, &),

an equation of the conic

is

- e*)y* - 2 *y + A 2 = 0,

as the reader can verify directly, or can obtain from (39.1)

by

interchanging % and y.

Equations (39.1) and (39.2) are special cases of the equation

ax 2

(39.3)

in

+ by2 + 2 fx + 2 g y + c = 0,

which not both a and b are zero.

two cases when either a or b

for the

We consider
is

this equation

equal to zero, and

when

equal to zero.
or b
Case 1. a

neither

is

= 0. If a = and 6^0, on dividing

and completing the square of the terms

equation (39.3) by
involving y, we obtain
(39.4)

When/ 7*

0,

and

on applying the

equation (39.5)
y'

(39.7)

this equation is written

translation of axes defined

is

by

transformed into

= 4a'x',

where

a'

= --^r>
Zi

seen to be an equation of a parabola.

which
= 0,
In the ^'/-system the vertex is at the point x' = 0,
the
at
the focus
point (a 0), and equations of the tangent at
rea
the vertex and of the directrix are %' =
and x'
is

then one desires the coordinates of these points

and equations of these lines in the ry-system, he has only to
spectively.

If

204

Equations of Conies

Sec. 39]

substitute from (39.6) and (39.7) the expressions for *', /, and
terms of x, y, and the coefficients of equation (39.3).

a' in

When

= f = 0,

As a quadratic

in

is

equation (39.3)

it

has two solutions, say

=k

and consequently the locus

y = k2

lt

consists of

two

lines parallel to the

x-axis, distinct or coincident according as k\

fa or k\

fa

and fa are imaginary.

imaginary
we
have
Accordingly
if

[39.1]

k\

An

+ =

2
2 gy
c
is an equation of
equation by
2fx
a parabola with axis parallel to the x-axis when f ^
;

when / =

it

0,

x-axis, which

is

an equation of two

may

lines parallel to the

be coincident if real, or

which

may

be

imaginary.
Similar reasoning applied to equation (39.3)
yields the theorem

and b
[39.2]

when a

+ =

An

2
2 gy
c
is an equation of a
2fx
equation ax
to
the
when
axis
with
parallel
y-axis when g ^
parabola
two
lines
to
the
an
it
is
parallel
y-axis,
0,
equation of
g
;

which

may

be coincident if real, or

which

may be

imaginary.

Consider, for example, the equation

2y 2 +

3x-4y + 4 = 0.

On dividing the equation by 2 and completing

Hence the locus

the square in y,

(y-l)> = -i*-l = -i(* +

1 =
is a parabola with axis y

we have

).

(that is, its axis is

2/3, 1). Since the coefficient
of x is negative, the parabola lies to the left of the vertex, and the
focus is at the directed distance
3/8 from the vertex, that is, the
parallel to the *-axis)

and with vertex

focus

is

the point

25/24, 1)

and an equation of the

directrix is

* + -f = * + ^ = 0.
In obtaining these results we have used the processes which led to
Theorem [39.1]. The reader should adopt this method in the solution
of any exercise.
205

The
Case

and

2.

in the

7* 0,

/s

[Chap. 4

Conies. Locus Problems

On

7* 0.

in (39.3),

completing the squares in the x's

we obtain

(39.8)

where by definition

If

then

we

x'

new

-c.

effect the translation of the axes defined

(39.10)

in the

(39.9)

= x + ^>

y'

=y+

by

coordinates (39.8) becomes

r '2

2L

(39.11)

-,'2

+ 2L = *.

When we compare
(37.2), in

this equation with equations (37.1) and

consequence of Theorems [37.1] and [37.2J we have

An

[39.3]

equation

+ by2 + 2fx + 2gy + c = Q

ax 2

and b have the same sign is an equation of an

when a = ft), real or imaginary according
as k, defined by (39.9), has the same sign as a and b or a
different sign from a and b, and of a point ellipse when k = Q;
when a and b have different signs, it is an equation of
in which a

ellipse (or circle

or
a hyperbola or intersecting lines according as k ^
=
a
real
or
a
k
ellipse
for
hyperbola the principal axes
are on the lines
;

= 0,

= 0.

(39.12)

The

theorem follows from

last part of this

-f-

(39.10).

Consider, for example, the equation

Completing the squares

in the x's

and

206

in the

/s, we have

Sec. 39]

Equations

By Theorems
on the

is

Conies

and [39.3] the locus is a hyperbola whose semiis on the line %

1 =
and semi-conjugate axis
=
2
0.
Its
is the point (1,
center
y+
2), and

[37.2]

2/V3

transverse axis

V2

ot

line

equations of the asymptotes are

x-

V3Cv

4- 2) __

x-l

o
2

/^

V2

o
2

'

/;r

V2

'

For a hyperbola
e2

= (a 2 +b

)/b

Since the foci

(37.7) the equation analogous to (35. 11') is

hence for the hyperbola under consideration e=V5/2.
1 = 0, and above and below the
are on the line x
;

center at the distance

be,

they are the points

(l,

db

VlO/3). Equa-

tions of the directrices are

y>L =
EXERCISES
1.

Draw

vertex, axis,

the graphs of the following equations after finding the

and focus of each curve
:

4^ -32 x + 4^-63 =
2

0,

3 *2

+ 6* + 3y + 4 = 0.

Find equations of two ellipses with center at ( 2, 4), and prinaxes

cipal
parallel to the coordinate axes, the semi-axes being 4 and 3.
2.

Find an equation of a hyperbola with center at (2,

1), one end
- 1), and one end of the conjugate axis
(5,
at (2, - 4). What is an equation of the conjugate hyperbola?
3.

of the transverse axis at

Find an equation of the parabola whose axis is parallel to the

;y-axis and which passes through the points (0, 0), (- 1, 2), and (2, 2).
4.

5. Draw the graphs of the following equations, after finding the

center and principal axes of each curve
:

- 4 y2 -

- 8 y - 31 = 0,
4* -.y + 2x-3;y-2 = 0.
that the line y = 2 x + h shall be

9 x2

6.
first
7.

Determine h so
conic in Ex.

Show

18 x

that the eccentricity of an ellipse (39.3)

tangent to the

5.

according as the major axis of the ellipse

*-axis or to the .y-axis (see equation (39,11)).

207

is

is

\l
^

or

parallel to the

The

Conies. Locus Problems

Show

that an equation of a parabola with axis parallel to the

and which passes through the noncollinear points (x\ yi),
^2), and (* 3 yd is
2
1
x
y
y

8.

*-axis
(x2 ,

y*

Discuss this equation

Ex.

[Chap. 4

9).

when

What

its

axis

is

= 0.

*2

when

y-2

the three points are collinear (see

4,

an equation of the parabola through these points

parallel to the ;y-axis ?

is

9. How many points are necessary to determine an ellipse or a

hyperbola whose principal axes are parallel to the coordinate axes?
Using Theorem [27.2], derive for this case an equation analogous to
that of Ex. 8.

40.

The General Equation

of the

Second Degree.

Invariants

is

The most general equation of the second degree in x

of the form
ax 2

(40.1)

and y

+ 2hxy + by2 + 2fx + 2 gy + c = 0.

In this section we shall show that any such equation is an

equation of a conic, including the case of two lines, called a
degenerate conic. In consequence of the results of 39 it follows
that all we have to do is to show that by a suitable rotation
of the axes equation (40.1) is transformed into an equation in
%'

and
If

y' in

which there

now we apply

is

no term

in x'y'.

the transformation (28.8) to equation (40.1),

we obtain
(40.2)

a'x' 2

where the

+ 2 h'x'y' + 2 b'y' 2 + 2f'x' + 2 g'y' + c' = 0,

coefficients are given

by

= a cos2 + 2 h sin 6 cos 6 + b sin2 0,

h = (b - a) sin 6 cos + h(cos 2 6 - sin2 0),
V = a sin2 6 - 2 h sin cos 6 + b cos2 0,
+ g sin 0, g' = f sin B + g cos 0,
/' =/ cos
a'
f

'*

208

c'

c.

The General hquation

Sec. 40]

ot the

Second Degree

From

the above expression for h it is seen that for the

cient of x'y in (40.2) to be zero 6 must be such that

coeffi-

(40.4)

(b

- a)

cos 9

sin

b = a, this equation
= 45. When b ^ a, if equation

When
is,

+ h (cos2 - sin2 0) = 0.
= sin 9
is satisfied by cos

the resulting equation

that

divided by cos 2

0,

be written

may

h tan 2

(40.5)

(40.4)

is

(a

b)

- h = 0,

tan

from which we have, by means of the quadratic formula,

(40.6)
If for the moment we denote by 6\ and 2 the values of
correspectively before the radical,
responding to the signs + and
we find that
tan 0, tan 2 = - 1.

This means that the #'-axis for the angle 6\ and the one for 2
are perpendicular to one another (see Theorem [7.3]). How-

we

are interested in finding a rotation of the axes so that

be zero, and consequently either solution (40.6) yields
the desired result. Since such a transformation is always possible, we have shown that any equation of the second degree
in x and y is an equation of a conic.
ever,

h' shall

For example,

so that in the

if it is

required to transform the equation

new coordinates

equation (40.5)

is

2 tan 2

there

3 tan

is

no term
2

in x'y', in this case

= 0.

Solving this equation

obtain tan 6 = 2 or
1/2.

by means of the quadratic formula, we

Taking the first root, we have sin 6 = 2/\/5, cos
the desired transformation (28.8)

1/V5, so that

is

|-,
Substituting these expressions for x and y in the above equation and
collecting terms, we obtain as the new equation

209

The

Conies. Locus Problems

which, on completing the squares in x' and

y',

{Chap. 4

assumes the form

Consequently the curve is an ellipse, and with regard to the coordinate

system x', y' the center of the ellipse is at the point (1/VJ5, 1/2 Vs)',
its semi-major axis is V372, and its semi-minor axis 1/2. The major
axis has the slope 2 in the ^-system.

From

(40.3)

we

derive other results of importance. On addand &', and making use of the fun-

ing the expressions for a'

damental identity
(40.7)

sin 2

+ cos 2 =

a'

+ b' = a + b.

1,

we obtain
(40.8)

h' 2 from (40.3),

something of an exercise to calculate a'b
but the reader will feel repaid when he finds that with the aid
f

It is

of (40.7) the result

reducible to

is

a'V

(40.9)

- h' 2 = ab- h 2

Thus we have found that when the coordinates

in the general
of
the
second
which
is
member
the
left-hand
degree
expression
of equation (40.1) are subjected to any transformation of the

b and ab
h 2 are equal to
form (28.8), the expressions a
the same expressions in the coefficients of the transform (40.2),
that is, the equation into which equation (40.1) is transformed.
In this sense we say that a
b and ab
h 2 are invariants under
the transformation. Invariants are of fundamental importance in applications of transformations of coordinates, as we

shall see in

what

follows.

(40.1), we consider the

following three possible cases
h*
Case 1. ab
0.
From (40.9) it follows that in a coordinate system for which h'
either a' or V must be zero,

Returning to the general equation

:

but not both, otherwise (40.2) is of the first degree, and consequently (40.1) is of the first degree. This latter statement
follows also from the fact that if a' = V = 0, then from (40.8)

210

Invariants

Sec 40]

2
2
and ab
h2 =
(a + h ) = 0, which can
hold for real values of a and h only when both are zero. When
or b' = 0, we have from Theorems
h' = 0, and either a' =
[39.1] and [39.2] that the curve is a parabola or two parallel lines.
Conversely, when the curve is a parabola or two
1
0, as follows from (40.9). This is an
parallel lines, ob
example of the fact that when an invariant is equal to zero for

we have

0,

lfi

one coordinate system,

system.

Case

that a' and

- A2 >

ab

2.

it is

0.

equal to zero for every coordinate

When

= 0,

it

follows from (40.9)

are both positive or both negative, and in consequence of Theorem [39.3] it follows that (40.1) is an equation
of a real or imaginary ellipse or a point ellipse.

Case

b'

3.

that a' and

- h 2 < 0. When

ab

differ in sign,

[39.3] it follows that (40.1)

two intersecting

h' = 0, it follows from (40.9)

and in consequence of Theorem
is an equation of a hyperbola or

lines.

Gathering together these

theorem

results,

we have the

following

[40.1]

(40.1) is an equation of a parabola or two

h2
or
coincident lines when ab
parallel
of a real
h2 >
or imaginary ellipse or a point ellipse when ab
;

Equation

of a hyperbola or two intersecting real lines when ab

h2 <

0.

given another form. We consider first

from which it follows that a and b have
the same sign they can be taken as positive for, if they are
not positive, by changing the sign of every term in (40.1) we
make them positive. Consider now the terms of the second

This result

the case ab

may be

h2

= 0,

2
2
2hxy by According as h is
degree in (40.1), that is, ax
V06, the
positive or negative, when h is replaced by Va# or
2
that is, the
above expression may be written (Va x
j)
terms of the second degree form a perfect square.
Ifab
h2 7*Q and 6^0, the terms of the second degree are

equal to

| (by

+ hx + ^h2 -abx)(by + hx211

The

Conies. Locus Problems

[Chap. 4

as the reader can verify by multiplying these terms together.

These factors are real or conjugate imaginary according as

ab

h2

<

or

>

If b

0.

0, in

which case ab

h2

<

0,

the

2 hy,
factors of the terms of the second degree are x and ax
both real and distinct. From this result and Theorem [40.1]

we have
[40.2]

An

equation (40.1) is an equation of a parabola or of two

parallel or coincident lines when the terms of the second
degree are a perfect square; it is an equation of an ellipse,
may be real, imaginary, or a point ellipse, when the

which

factors of the terms of the second degree are conjugate imaginary; it is an equation of a hyperbola or two real inter-

when

secting lines

are real

We

and

consider

where

now

all

the equation

+ btf + d)(a 2 x + b 2y + c =

(a&

(40.10)

the factors of the terms of the second degree

distinct.

2)

In consequence of Theob 2 y of the terms

the coefficients are real.

the factors a\K

b\y and a 2 x
of the second degree are different, (40.10)
two intersecting lines according as k ^

rem

[40.2],

if

*,

a hyperbola or

is

or k

= 0. When

and the axes are rotated so that there is no term

and an appropriate translation of the axes is made,

7* 0,

x'y',

in
if

necessary, each of the factors in (40.10) is transformed into a

factor homogeneous of the first degree, and these are such that

the terms of the second degree consist of the difference of a

2
2
Hence we have
multiple of x' and a multiple of

(ai'x

(40.11)

From

+ fti'/X*!'*' - *,'/) = *.

this result, the discussion leading

and the definition of asymptotes in

lines whose equations are obtained

up to Theorem
37

it

on equating to zero the

member of (40.11), and consequently
are the asymptotes of the hyperbola. Hence

factors in the left-hand

those in (40.10),

we have
(40.3 J

(40.12)

An

[39.3],

follows that the

equation (40.10) for k

aix

+ btf + ci

= 0,
212

0,

such that the lines

a2x

+ b 2y + c =
2

The General Equation

Sec. 40]

of the

Second Degree

intersect, is an equation of a hyperbola for which these lines

are the asymptotes; and (40.10) for a suitable value of k

an equation of any hyperbola having

is

the lines (40.12) for

asymptotes.

The proof of the second part of this theorem is left to the

reader (see Ex. 4).
We consider next the case when the lines (40.12) are parallel
or coincident, in which case equation (40.10) may be written
(ax

(40.13)
If

we

+ by + Cl

(ax

+ by + c 2 = *.
)

a transformation of coordinates for which

effect

ax
is

+ by + Ci =

the #'-axis, this equation becomes

'

= Ci,

and

where

c2

this

an equation of two

is

if c 2

upon the values

An

[40.4]

of c 2 ',

'

if c 2

In either case

CL

parallel or coincident lines,

k, a,

equation (ax

c2

and

b.

depending

Hence we have

+ by + Ci)(ax +

fry

+ c 2 = & is an equa)

tion of two real or imaginary, parallel or coincident lines.

As a consequence

of this result

and Theorem

[40.2]

we have

When the terms of the second degree in an equation (40.1)

are a perfect square, the locus is two parallel, or coincident,
lines or a parabola according as the equation can or cannot
be put in the form (40.13).

[40.5]

Consider the equation

x*

(40.14)

+ xy-2y*-2x + 5y-2 = Q.

the second degree have the real factors x

y and x + 2 y,
and consequently we seek constants d and e so that the above equation
shall assume the form

The terms of

(40.15)

When
get

(x

-y + d)(x + 2y + e) = k.

the two expressions in parentheses are multiplied together,

+ xy _ 2y * +

(d + e)x + (2 d
213

- k = 0.
e)y + de

we

TKe
Comparing

this equation
</

from the

first

an

[Chap. 4

with (40.14), we see that we must have

+ 6 = - 2,

cf

- =
<?

de-k = -2;

5,

two of these equations it follows that d=l, e =

- 2, k must be - 1. Since k ^ 0,
and

k to be equal to
equation of a hyperbola

that for de
is

Conies. Locus Problems

_.
>>,

+1-

and

3,

so

(40.14)

+ 2y-3 =

its asymptotes.
From these results it follows that the equation
2 by
obtained from (40.14), on replacing the constant term
3, is

are

an equation of two straight

equal to zero.

lines, since in this

case k in (40.15)

is

In analyzing a general equation of the second degree for

which ab
h 2 < 0, it is advisable to use the above process,
which yields the asymptotes when the locus is a hyperbola, and
equations of the lines when the locus is two intersecting lines.

EXERCISES
1. Determine the type of conic defined by each of the following
equations either by effecting a rotation of the axes or by using the

theorems of

this section

- 4 xy + 8 y* + 18 x - 36 y + 9 = 0.
2
b. x 2 - 4 xy + 4 y + 5 y - 9 = 0.
2 2
11 x - 2 y + 12 = 0.
c. 2 x + 3 xy - 2 y
2 2
2 x - 2 y - 3 = 0.
d. x + 2 xy + y
2
2
e. x -4xy-2y -2x + 7y-3 = Q.
a.

2.

5 x2

Find the asymptotes of the hyperbola

2x 2

-3xy-2y + 3x-y + S = Q,
2

and derive therefrom equations

3. Find an equation of
axes for asymptotes.

all

of the principal axes of the hyperbola.

hyperbolas which have the coordinate

Show

that a hyperbola is completely determined by its asympand a point of the hyperbola, and apply this principle to find
3 y + 1 = 0,
an equation of a hyperbola whose asymptotes are 2 x
3 =
and which passes through the point (1,
x+y
2).
4.

totes

Show

that the centers of all hyperbolas whose asymptotes are

the coordinate axes and which pass through the points
to
parallel
0; also find an equation
(2, 5) and (3, 2) lie on the line 3x
5.

>~4 =

of the hyperbola of this set which passes through the point

214

2, 3).

The Determination

Sec. 41]

6. Find the eccentricity

making use of Ex. 8 of 37.

of

a Conic

of a hyperbola with equation

(40.10),

7. Verify the following statement

An equation of a conic involves
a term in xy, if and only if the directrices (or directrix) are not parallel
to one of the coordinate axes.
:

41.

The Determination

Conic from
General Form
of a

Its

Equation in

Having shown in 40 that the locus of any equation of the

second degree in x and y is a conic, or a degenerate conic, we
show in this section how one may determine whether a conic
is degenerate or not directly from the coefficients of its equation without effecting

any transformation of coordinates. The

method used to establish these criteria gives at the same time

a ready means of drawing the graph of such an equation.

We

consider then the equation

+ 2hxy + by + 2fx + 2gy + c = 0,

ax 2

(41.1)

and discuss
the form

first

(41.2)

by

the case

when

+ 2(hx + g)y +

(ax

0,

writing the equation in

+ 2 /* + c) = 0.

Considering this as a quadratic in y with x entering into the

coefficients, and solving for y by means of the quadratic formula, we have

y=-**

(41.3)

where A
(41.4)

is

defined

A=

&

\^A,

by
(hx
(h

+ g) 2 - b(ax2 + 2fx + c)
- ab)x2 + 2(hg - bf)x + &2 -bc.

For any value of x for which A is positive, the two values of y

given by (41.3) are the ^-coordinates of two points on the curve
having the given value of x for x-coordinate. For a value of x,
if any, for which A = 0, equation (41.3) reduces to

215

The
that

Conies. Locus Problems

[Chap. 4

the point having for coordinates a value of x for which

(41.5) is a point of intersec-

is,

A=

and the value of y given by

tion of the curve and the line

+ by + g = 0.

hx

(41.6)

We may

interpret (41.3) as follows: Draw, as in Fig. 37, the

at any point on the line whose
graph of the line (41.6)
^-coordinate, say x\ is such that A is positive, add and sub;

tract

from the ^-coordinate of the point on the

- VZi, where A\

the value of

is

A when

line the

quantity

x has been replaced

the two values thus obtained are the ^-coordinates of

xi
the two points on the curve for which the jc-coordinate is XL
In other words, these are the coordinates of the two points
If XQ is
in which the curve is met by the line x
x\ = 0.

by

such that for this value of

x the quantity A is equal
to zero, the two points
of

intersection

on the

in the point

which y

for

hx

-*fl.
i

1
/Z\
-t-^v^i;

J.

line

+6
j--

#o =

line x
tangent to the curve at

hence the
is

/r
AUi,

coincide

this point (see

12, after

equation (12.8)) if the

curve is not a degenerate
conic.

gestive

37 is only sugthe position of

Fig.
;

the line hx

+ by + g=

pIG

37

and

the position and shape of the curve depend upon the values of
the coefficients in (41.1).

From
if

an equation of two lines,

a constant or the square of an expression of
degree in x. In order that A shall be a constant, we
(41.3) it follows that (41.1) is

and only

the

first

if

is

must have

hg-bf=0,

(41.7)

216

The Determination

Sec. 41]

which case

in

of a

Conic

(41.3) is

(41.8)

which are equations of two parallel lines, real if g 2

bc>
2
be < 0, and of real and coincident lines
imaginary if g
g

-bc =

We

if

0.

when ab

consider next the case

fect square.

The

condition that

(kg

(41.9)

The

0,

bf)

(h

h2

and

is

be a perfect square

a peris

- ab) (g2 - be) = 0.

expression on the left reduces to

- b(abc - as2 - bf2 - ch2 + 2jgh),

which

is

bD where

equal to

D=

(41.10)

D is defined by

a
h

f
Since
to

b
0, we have that (41.9) is equivalent
this condition is satisfied, equations (41.3)

by hypothesis

D = 0. When

are equivalent to

by

(41.11)

+ hx + g^ (\fh2 -abx+e\fg2 -bc) = 0,

since the square of the expression in parenthesis is equal to A

in consequence of (41.9), e being
or
1 according as

- bf > or < 0.
hg
We consider now
as ab

bilities

= 0,

in

(41.12)

As a

h2

is

+1

the above results for the various possi-

zero, positive,

which case equation

ax 2

and negative, and

also

when

(41.1) is

+ 2hxy + 2fx + 2gy + c = Q.

result of this analysis the reader will

be able to determine

completely the character of the locus of any equation of the

second degree from the values of the coefficients of the equation. It is suggested that, as he proceeds, he make a table of
the results of the analysis.
Case 1. ab - A 2 = 0. When b ^ and D = 0, we have (417),
as follows from (41.9),

and consequently the locus is two parallel

When 6 = 0, then h = 0, and (41.12)

or coincident lines (41.8).

217

The
is

Conies. Locus Problems

an equation of two

[Chap. 4

parallel or coincident lines,

if

and only

if

g = 0, as follows from Theorem [39.2]. When b = h = g = 0, we

have D = 0. Accordingly we have
When ab h = Q, the locus of equation (41.1) is a
[41.1]
or D =
parabola or two lines according as D ^
when D = and b ^ 0, the lines are parallel and real or
2

2
imaginary according as g

coincident

if

be

bc>
;

when

or

D=

<

and real and

and b = h = 0,

0,

and real or imaginary according as

and real and coincident when/2 ac = 0.

the lines are parallel

f2

The part

ac>Qor<Q,

of this theorem for

When = h =
b

D = 0, b 5* follows from (41.8).

D = that g = 0. Hence we

follows from

it

have

which is obtained on solving (41.12) in this case.

h2 > 0. Since this case does not arise when
Case 2. ab
b = 0, we have that when D = 0, equation (41.1) is equivalent
to two conjugate imaginary equations of the first degree (41.11),
whose common solution is real, and consequently (41.1) is an
equation of a point ellipse, which is a degenerate conic. When
DT* 0, (41.1) is an equation of a real or imaginary ellipse. In
order to distinguish these two types of an ellipse, we observe
that for the ellipse to be real there must be real values of x for
which A as defined by (41.4) is equal to zero, as follows when
one considers Fig. 37 for the case of an ellipse. The condition
for this to be so is that the left-hand member of (41.9) shall be
bD > 0. Hence we have
positive, that is,
[41.2]

When ab
ellipse

h2

=
if D

according as

>

0, the locus of equation (41.1) is

D ^ 0, it is
bD > or < 0.
;

if

a point

a real or imaginary

ellipse

h* < 0. When b ^ 0, it follows from (41.11)

Case 3. ab
and Theorem [40.1] that (41.1) is an equation of a hyperbola
or D = 0. When
or two intersecting lines according as D ?
fr

= 0,

the factors of the second-degree terms in (41.12) are x

218

The Determination

Sec. 41]

and ax
lines,

is

+ 2 hy.

there

then (41.12)

If

must

of a

Conic

to be an equation of
e such that

is

two

numbers d and

exist

equal to the left-hand

member

of (41.12).

Multiplying these

expressions together and equating the coefficients of x and y

and the constant terms of the two expressions, we have
e

Solving the
the third,

first

+ ad = 2f,

two equations

we have

satisfied:

[41.3]

= 0,

+ ch

in fact

is

ed

g,

c.

and d and substituting in

following condition must be

for e

that the

a g2

which, since b

hd

2 __

2 fgh

D = 0. Hence we

have

When ab h 2 < 0, the locus of equation (41.1) is a hyperor

bola or two intersecting lines according as
0.

D^

From

the foregoing theorems

An

[41.4]

D=

we have

equation (41.1) is an equation of a degenerate conic,

if the determinant D is equal to zero.

if

and only

The
[41.5]

algebraic equivalent of this theorem

equal

to the

+ 2hxy + by2 + 2fx + 2gy + c

It follows

from
ax

product of two factors of the

=
if D

and only

is

the following

quadratic expression

ax 2
is

is

this

first degree, if

0.

theorem that a quadratic form

+ by + cz + 2hxy + 2fxz + 2 gyz

2

the product of two linear homogeneous factors in x, y, and

and only if D 0. (See page 267.)

2, if

When by means

of

Theorems

[41.4]

and

[41.2] the reader

is a degenerate conic
not a point ellipse, he should reduce the equation to
the product of two factors of the first degree in x and y, after
the manner of the exercise worked toward the end of 40, and
interpret the result geometrically.

finds that the locus of a given equation

which

is

219

The

Conies. Locus Problems

[Chap. 4

EXERCISES
1.

Using the method of

tions of Ex.
2.

For what value of

c is

an equation of two intersecting

3.

For what values of k

4.

Show

that

draw the graphs of the equa-

this section,

40.

1,

when

D=

is

(41.1)

is

Find equations of the

lines ?

and

for equations (40.10)

an equation of a

lines.

(40.13) ?

real ellipse, the values

of x for which A, defined by (41.4), is positive lie between the roots

when (41.1) is an equation of a hyperbola, the
of the equation A
;

values of x for which

roots

is

positive are less than the smaller of the two

if the roots are
larger of the roots of A

and greater than the

real, that
5.

is, if

Show

bD>

0.

Discuss the case

that the line (41.6)

is

bD <

when

0.

the locus of the mid-points of a set

of parallel chords of the conic ; and that when the conic is a parabola
it is a line parallel to the axis of the parabola (see Theorem [34.1]).
6. Show that when (41.1) with b
conic, the x-coordinate of the center

an equation of a central

is

one half the sum of the roots

is

of the equation A
0, where A is defined by (41.4),
coordinates x Qt y$ of the center are given by

_bfDiscuss the case

when

and that the

_ag- hf

hg

= 0.

in equation (41.1)
consider the equation as
7. Assume that a ^
a quadratic in x with y entering in the coefficients, and discuss the
solution for x in a manner similar to that developed in this section.
;

8.

form

Show

that,

when an equation

(40.10), equations of

its

+ biy -f Ci =

a\x

of a hyperbola

a2 x

<?

Apply

written in the

-f b2y

+ c2

V02 + fc 2
2

this result to find the principal axes of the hyperbola (40.14).

Show that when

a number c such that
9.

ax 2
is

is

principal axes are (see (10.10))

ab

h2

+ 2hxy +

for

by

the product of two factors of the

an equation

+ 2fx +
first

220

degree.

(41.1) there exists

Center, Principal Axes,

Sec. 42]

Show

10.

and Tangents

that in consequence of Ex. 9 and (41.11) equation (41.1)

can be written

[(A

-06)x +

V/* 2

where

e is

or

by

the hyperbola (40.14)

+ g + eVg 2 - be]

- VA 2 - ab)x + by + g
- e^/g^-bc] = b(c - c),

[(h

- bf > or < 0. Show that for

according as hg
c = - 3, and obtain by this method the results

following (40.15).
11.

an

Show

ellipse,

that x 2

and obtain

2 xy

its

2 x

in the

equation

+ 4y=

is an equation of
form (40.10) by the method

of Ex. 10.

and Tangents of a Conic

Defined by a General Equation

42. Center, Principal Axes,

In accordance with Theorem

x

(42.1)

[6.3],

= xi + tu,

= yi + tv

are parametric equations of the line through the point (x\ 9 y\)
v, t being proportional to the

and with direction numbers u and

directed distance from

(x\, y\)

to a representative point (x, y).

if any, for points in which the

In order to find the values of /,

line so defined meets the curve
(42. 2)

we

ax2

+ 2hxy + by2 + 2fx + 2gy + c = Q,

substitute the above expressions for x

and y and obtain the

equation
(42.3)

(au

+ 2huv + bv 2 )t2 + 2[(ax + hy +f)u + (kxi + by + g>]/

+ (axS + 2 hxiyi + by{2 + 2fxi + 2gyi + c) = 0.

We make

use of this result in the discussion of a

number

of

questions.

be the mid-point of the chord of the conic

with direction numbers u and v, that is, the line segment with
If (xi, y\) is to

the points of intersection of the line and the conic as end points,
the two solutions of (42.3) must differ only in sign, and hence the
that is,
coefficient of / in (42.3) must be equal to zero
;

(42.4)

(ax,

+ hy +f)u +
l

(h Xl

221

+ byi + g)v = 0.

The

Conies. Locus Problems

[Chap. 4

In order that (x\, y\) shall be the center of the conic, it must
be the mid-point of every chord through it, and consequently
(42.4) must hold for every value of u and v. This means that
the center (JCG ;yo), if the conic has a center, is given by
,

ax<>

(42.5)

+ hy +f=Q,

hx

+ byo + g = Q.

By Theorem [9.1] these equations have one and only one common solution if ab h2 ^ 0. When this condition is satisfied,
the coordinates of the center are given by (see

41, Ex. 6)

If the conic (42.2) is not degenerate, these are the coordinates

h 2 > 0, and of the center
of the center of an ellipse when ab
2
h < 0.
of a hyperbola when ab

h 2 ^ 0, and (*o,
and ab
of the two real or imaginary

If the conic is degenerate

now

yo) is

the intersection
lines,
then the expression in the last parentheses in (42.3) is equal
to zero. Also, if u and v are direction numbers of either of the
lines, equation (42.3) must be satisfied for every value of /.

Consequently we have (42.4) holding, and also

au2 + 2huv

+ bv2 = 0.

From this it follows, as shown before, that the lines are real
h2 <
or imaginary according as ab
or > 0. Since (42.4)
must hold for the direction numbers of the two lines, we again
obtain equations (42.5). Consequently the coordinates of the
vertex, that is, the point of intersection of the lines, are given
by

(42.7)
it

When equation (42.2) is written in the form

(ax + hy+f)x+(hx + by + g)y+fx + gy + c = 0,

(42.6).

foUows from (42.5) that

fxo

(42.8)

+ gyo + c = 0.

This equation and equations (42.5) have a common solution,

and only if D = 0, where D is defined by (41.10) this result
is in accord with Theorem [41.4]. Hence we have
if

[42.1]

When ab
center (42.6)
vertex (42.6)

h2

a central conic with

a degenerate conic with

?* 0, the conic (42.2) is

when
when

D ^ 0,
= 0.

and

222

is

Center, Principal Axes,

Sec. 42]

and Tangents

When ab-h 2 = 0, by Theorem [9.3] equations

not have a common solution unless hg
and hf
bf =

do

(42.5)

ag

= 0,

in which case the conic is degenerate, as follows from (41.9)

and (41.11). Thus, as stated in 35, a parabola does not have a
center. When the conic is degenerate, and consists of two parallel

or coincident lines, equations (42.5) are equivalent; they are

equations of a line of centers, that is, a line of points of symmetry.
Returning to the consideration of equation (42.4), we observe that for fixed values of u and

any point of the

line

(ax + hy +f)u + (hx + by + g)v = 0,

(42.9)
for

0,

which the values of

given by (42.3) are real

when

xi

and

y\ are the coordinates of such a point, is the mid-point of the

chord through this point and with direction numbers u and v .

Hence we have
Equation (42.9)

[42.2]

is

an equation of

u and

midnumbers

the locus of the

points of the set of parallel chords with direction

v.

In order that (42.9) shall be an equation of a principal axis,

is, a line of symmetry of the conic (42.2), the line (42.9)
must be perpendicular to the chords it bisects consequently
u and v must be such that they are direction numbers of the
perpendiculars to the line. When equation (42.9) is written
that

+ hv)x + (hu + bv)y + (fu + gv) = 0,

that au + hv and hu + bv are direction numbers

(au

we observe

perpendicular to (42.9) by
quently u and v must be such that

any

line

au

Theorem

[6.9].

of

Conse-

+ bv = rv,
proportionality r ^ 0,

+ hv = ru,

hu

otherwise there
where r is a factor of
written
are
is no line (42.9). When these equations
;

(42.10)

(a-r)u

+ hv = 0,

hu+(b-r)v =

we have

Q,

that a solution u, v, not both zero, of these equations is

for
each value of r, other than zero, satisfying the equation
given
r

(42.11)

b-r
223

= 0,

TKe

Conies. Locus Problems

as the reader can readily show.

it

[Chap. 4

When this equation is expanded,

becomes
r2

(42.12)

This equation

We

(42.2).

ab-h

Case

is

(a

+ b)r +

(ab

- A2 = 0.
)

called the characteristic equation of equation

when ab h 2 Q and

consider the two cases

9* 0.
1.

zero root,

ab
h* = 0. In this case
and from (42.10) we have

-f b is

the only non-

=-=v~ b~ h

In consequence of this result equation (42.9) reduces to

+ b)(hx + by) +fh + gb = 0.

(42.13)

(a

Hence we have

(see

When

[42.3]

41, Ex. 5)

an equation of a parabola, (42.13)

(42.2) is

is

an

equation of the principal axis of the parabola.

Case

2.

are found

(42.14)

ab-h*7

0.

In this case the solutions of (42.12)

by the use of the quadratic formula to be

r

and are always real numbers. The two roots are different, unless o = b and h = 0, in which case the conic is a circle (see
12). In any particular case, other than that of a circle, with
the two values of r from (42.14) two sets of values of u and v
may be found from (42.10), which when substituted in (42.9)
give the two principal axes of the conic.
When the left-hand member of (42.2) is the product of two
factors of the first degree, in which case by Theorem [41.5]
D = 0, and a transformation of coordinates is effected, the resulting expression in x y' is the product of two factors of the
first degree, and consequently D' =
for this expression, where
f

D' is the corresponding function (41.10). From this result it

does not follow necessarily that D = D' in general, but we shall
show that this is true, and consequently that D is an invariant

224

Center, Principal Axes, and Tangents

Sec. 42]

under any transformation of rectangular coordinates. We prove

this not by direct substitution, but by the following interesting
device. We denote by f(x, y) the left-hand member of (42.2),

and consider the expression

/(*,jO-r(*

(42.15)

that

+ y 2 +l),

is,

(a

r)x

By Theorem

+ 2hxy+(b- r)y 2 + 2fx + 2gy+(c- r).

[41.5] the condition that this expression shall be

first degree is that r shall be

the product of two factors of the

such that
h
a _ r
h

= 0.

b-r

c-r

This equation, upon expansion of the determinant,

r

(42.16)

/r

is

+ Jr - D = 0,

where
(42.17)

When

=a+b+

J=

c,

ab

+ be +

a rotation of the axes (28.8)

is

ac

h2

-f2 - g2

applied to (42.15),

we

obtain
(42.18)

/'(*',

/)-r(*'

+ / 2 +l),

where /'(*',

y') denotes the transform of f(x y), that is, the exinto
which f(x, y) is transformed. If (42.15) is the
pression
product of two factors of the first degree, so also is (42.18) and

we have

^_

//r2

_D =^
,

Jftf

where /', J', and ZX are the same functions of

Since r
c.
/, and D respectively are of a,
-

a',
is

c'

as

/,

not affected

by the transformation, it follows that / = /', / = /',

Consequently /, /, and D are invariants under any

D=

D'.

rotation

of the axes.

When a translation (28.1)

not changed, and we have
/'(*', y')

is

applied to f(x, y),

a,

ft,

and h are

= ax' 2 + by' 2 + 2 hx'y' + 2(ax + hy +f)x'

+ 2(hx + by<> + g)y' +/(*<>, JVo).
225

The
For

If,

Conies. Locus Problems

D'

this expression

[Chap. 4

is

+g

considering f(x Q yo) expressed in the form (42.7), we subfrom the last row the first row multiplied by XQ and the
,

tract

second multiplied by yQ we have

,

which

is

lation,

we have

axo

hx

fxo

+ by + g
+

seen to be equal to D. Since any transformation of

rectangular coordinates is equivalent to a rotation and a trans-

[42.4]

The function D of the

second degree in x and y

an equation of the
an invariant under any change

coefficients of
is

of rectangular coordinate axes.

h 2 are invariants, it
b and ab
40, a
Since, as shown in
follows that the roots of the characteristic equation are the
same in every coordinate system. When by a rotation of the

axes equation (42.2) is transformed into an equation without

a term in xy, it follows from (42.11) that the coefficients of
the terms in x 2 and y 2 are the roots of the characteristic equation. Hence, when one finds the roots of the characteristic

equation of an equation

which are the

(42.2), one has obtained the numbers

coefficients of the second-degree terms in an

equation without a term in xy into which (42.2) is transformable.

In accordance with the theory of algebraic equations it follows from (42.12) that

ab-h 2

(42.19)

where

r\,

r2

h2

when ab

are the roots of (42.11).

and ab
h2
0.

We

consider the cases

ab-/i 2 ^0.

From (42.19) it follows that both

roots are different from zero, and from the results of
40 that
(42.2) is transformable into an equation of the form
Case

1.

ax 2

+ by 2 + c = 0.
226

Center, Principal Axes, and Tangents

Sec. 42]

r\ = a, r 2 = b, and D = abc = rir 2 c. Consequently

an appropriate coordinate system an equation of the conic is

In this case
in

n*2

(42.20)

+ r2y2 + 7ir2 = 0.

Hence we have
an

ellipse or of

^
D^

h2

Equation (42.2) for which ab

[42.5]

a hyperbola, when

have the same or opposite signs;

roots

0w equation of
according as the

is

0,

is

it

D = 0-

when

Case 2. ab
one of the roots

7z

is

= 0.

zero,

From

it

(42.19)

degenerate

follows that at least

and the reader can show that the other

not zero (see Ex. 6). From the results of 40 it follows that
(42.2) is transformable into an equation of the form

is

ax 2

+ 2 by =

In both cases we may take TI

2
in the second case
rife
the conic is

D=

ri*

(42.21)

+ 2 x\

ax 2

or

+ c=* 0.

= a, r 2 = 0. In the first case

D = 0. Hence an equation of

n* 2 + c =

or

0.

r\

The second equation is an equation of two parallel or coincident lines according as c ^

or c = 0. In the latter case every
of
the
lines
is a point of symmetry and lies
coincident
point
on the locus. Consequently, if c in the second of equations (42.21)
is to be zero, the common solutions of equations (42.5) for a
given equation (42.2) must be solutions of the given equation.

The preceding

results are set forth in the following table, in

the
which
canonical, or type, form of an equation of a conic
in
is given
terms of the roots of its characteristic equation
:

r i* 2

......
=
a

~~

Ellipse or hyperbola

ri/2

Two

lines

Parabola

= 0* ....... Two parallel or coincident lines

*c =

if

(42.5)

and

(42.2)

have a common

227

solution.

The

Conies. Locus Problems

[Chap. 4

When

in a particular problem one has found the values of the

is able to determine completely the form of the conic,
but in order to find its position relative to the given coordinate

he

roots,

axes

it is

means

necessary that he find the principal axes, or axis, by

of (42.9)

and

(42.10).

We

consider finally the case when (x\, y\) is a point of the

conic and (42.1) are equations of the tangent to the conic at
this point. Since the point (x\ y y\) is on the conic, the expression in the last parentheses in (42.3) is equal to zero. Conse0, for which from (42.1)
quently one solution of (42.3) is /

the point (xi, y\). The other solution of (42.3) when

substituted in (42.1) gives the coordinates of the other point
in which the line meets the conic. If the line is to be tangent

we have

at (*i,

and

this other solution also

;yi),

must be

zero, that

is,

must be such that

+ hyi +f)u + (hx + by, + g)v = 0.

follows that hxi + by\ + g and
(ax\ + hy\ +/),

(axi

From

this

it

being proportional to u and v, are direction numbers of the

tangent, and consequently an equation of the tangent at the
(*i, ;vi) is

point

x-xi

y-yi
|

When

=0

equation is cleared of fractions, and the expression

2
- (2fxi + 2gyi + c), to
2
axi +
hxiyi + by, is replaced by
which it is equal, as follows from (42.2), we have
this

An

[42.6]

equation of the tangent

to the

conic (42.2) at the point

(xi, y\) is

(42.22)

EXERCISES
1.

Find the

axis, vertex,

and tangent at the vertex of the parabola

4x 2 -12xy + 9y 2
2.

Find the principal axes and center of the following

a. 8 x 2 - 4 xy + 5 y 2 - 36 x + 18 y + 9 = 0.

b.

Show

-3x-2y + 4 = Q.

2 x2

-4ry - y 2 + 7x-2>> + 3 = 0.

that the axes are perpendicular to one another.

228

Locus Problems

Sec. 43]

3.

Show

that

when

translation of axes

is

(42.2) is an equation of a central conic, and a

r
effected by the equations jc
%' -f * y
y
yo,

which the point (*

there are no terms of the
for

.yo)

first

is

the center, in the resulting equation

degree in %' and

y'.

Apply

this process

to the equations in Ex. 2.

4. Determine whether equations (34.5), (36.7), and (36.7') conform to Theorem [42.6], and formulate a rule for obtaining (42.22)
from (42.2).
5.

Find equations of the tangent and normal at the point

to the conic

6.

Show

2* a -4*> + 3j a

(2,

1)

-2* + 3j'-12 = 0.

that not both of the roots of the characteristic equation

(42.12) can be zero.

7. By finding the roots of the characteristic equation determine
completely the form of each of the conies in Ex. 2 and also in 40, Ex. 1.

8.

What
h2

ab

= 0,

is

the character of the locus of equation (42.2), for which

it is possible to derive by a translation of axes an

when

equation in which there are no terms of the

9.

first

degree in x and y' ?

Prove that the centers of the conies whose equations are

ax 2

as the parameter
the origin.

+ 2hxy +

by

+ 2 fix + 2 gty + c = 0,

takes different values,

lie

on a straight

line

through

43. Locus Problems

13 we explained what is
In the closing paragraphs of
finding the locus of a point satisfying certain geo32 it was stated that when a locus is
metric conditions. In
defined geometrically and without reference to a coordinate
system, the reader is free to choose any set of coordinate axes
in obtaining an equation of the locus, but that it is advisable
that axes be chosen in such manner relative to the geometric
configuration as to obtain an equation in simple form. In

meant by

some cases the choice of an axis of symmetry, if there is such,

as one of the coordinate axes tends toward simplicity in the
equation.
give below two examples to be studied before

We

the reader proceeds to the solution of the exercises at the end

229

The
of this section.

Conies. Locus Problems

[Chap. 4

Before doing so he should understand that in

certain cases only a portion of the graph of an equation obtained for a given problem is the locus, that is, some points
of the graph do not satisfy the geometric definition of the locus
;

and the reader must have this in mind as he interprets geometrically any equation he has obtained.
1. Find the locus of the vertex of a triangle whose base is fixed
in position and length, and whose
angles are such that the product of
the tangents of the base angles is a

constant, not zero.

We

choose the base of the

triangle for the #-axis and its
mid-point for origin. If we denote
by 2 a the length of the base, its

end points are A'(

A (a,

0).

0)

a,

and

In accordance with the problem (see Fig. 38),

tan PA' A
If,

A( a ,o)

(*.o)

^'(-a,o)

tan

PAA'

k (* 0).

as in the figure, the base angles are acute, the constant k is

If P lies so that one of the base angles is obtuse, k

positive.
is

negative.

In both cases

we have from

the above equation

When P lies below the #-axis, each y in the equation must be

y, but this does not change the equation. All
replaced by
such considerations as these must be taken into account in
studying a problem, if one is to be sure that one
every aspect of it.
When the above equation is written in the form
r2

a2

is

handling

+JL2 =
T
ka

1
'

seen that when k >

the curve is an ellipse, and that A' A
the major or minor axis according as k < 1 or k > 1
and
that when k < 0, the curve is a hyperbola with A 'A for transverse axis. In both cases all points of the curve except A' and
it is
is

satisfy the definition of the locus.

230

Locus Problems

Sec. 43]

2. The base of a triangle is given in length and position, and

one of the angles at the base is double the other; find the locus

of the vertex.

We
and

take the base as the *-axis

^ #>
x

its

mid-point as origin, with

the result that the coordinates

of the ends of the base are of

the form (- a, 0) and (a, 0). We

denote the angle at ( a, 0) by

"7^5

and the angle at

(a, 0) by 2 c/>
In accordance with
the definition of the locus we have
</>

(see Fig. 39).

tan

(43.1)

Dividing the
obtain

first

From

<f>

this result

MQTV
^
'

=
a

2>
x

tan 2

3Q

2
a

equation by the second and solving for

r
X

(A* ?\
(

B\x.o) &o)

- ^(tan 2
~
tan 2

and the

first

</>

of (43.1)

_2
" a tan
tan 2

- tan
+ tan

</>

x,

we

</>)
</>

we obtain

tan 2
'

+ tan

</>

as a parameter, and
are expressed in terms of
are
and
parametric equations of the locus.
(43.3)
equations (43.2)
In many locus problems, particularly those in which the

Thus % and y

point P is defined with respect to movable points and lines, it is

advisable to use a parameter, and in some cases several para1
meters which are not independent but connected by n

n being the number of parameters. By eliminating the

parameters from these equations and from the two expressions
for x and y in terms of these, we obtain an equation in x and
relations,

y of the

locus.

In the above problem the equation of the curve in x and y

from (43.2) and (43.3), or more
is obtained by eliminating
from
(43.1) by using the formula from trigonometry
readily

tan 2

= 2 tan </(!

tan2 0).

This gives the equation

231

The

Conies. Locus Problems

[Chap. 4

y = 0, the *-axis, which is an evident solution of the problem, and to the equation obtained by
equating to zero the expression in parentheses. When this
equation is written
this equation is equivalent to

(43.4)
it is

seen to be an equation of a hyperbola with center at the

point

0],

conjugate axis

vertices at

20/V3. Only

through the vertex (f

a,

0)

and

h|,

V and semi-

the branch of the hyperbola

satisfies
)

the conditions of the problem,

as one sees geometrically.

This problem

is

of historical interest because of

its

relation to

that of the trisection of an angle, a problem which goes back to the

Greeks, who endeavored to obtain a construction for trisecting an
angle, using only a ruler and a compass. Years ago it was shown
by algebraic considerations that this is impossible for a general angle,

but many people keep on trying to do it. We shall show that it

be done by means of a hyperbola constructed not point by point
continuously (see \ 35 after Theorem [35.1']).
Suppose that on a line A'( a, 0)A(a, 0) the hyperbola (43.4)
been accurately constructed, as shown in Fig. 40. Through
point A' and below the line A' A a
line is drawn making an angle of
90 - 6 with the line A' A, 6 being a
given acute angle. Denote by C the
A' (-0,0)

can
but
has
the

point of intersection of this line and

the perpendicular to A' A at its mid-

point 0. With C as center and CA'

as radius describe a circle, meeting
the branch of the hyperbola through

(|

(A in the point P.

A'CA

The

angle
FIG. 40

equal to 2 0; consequently
the angle A' PA is (360 -26) = 180 - 0. But
the angle A' PA is 180 -30. Consequently
is

by the above problem

= J 9, as was to be

shown.

From this result it follows that if a hyperbola could be constructed

continuously, instead of only point by point, by means of a ruler and
232

Locus Problems

Sec. 43]

compass, a given angle could be trisected by this means; since it

has been shown, as remarked above, that the latter is impossible, it
follows that a hyperbola cannot be constructed continuously by means
of ruler and compass, although it can be by other means (see
35).

EXERCISES
1. Find the locus of a point the sum of the squares of whose distances from two fixed points is constant.

2.

Find the locus of a point whose distances from two fixed points

are in constant ratio.

3.

Find the locus of a point the square of whose distance from a

is a constant times its distance from a fixed line not passing

fixed point

through the fixed point.

4. Find the locus of a point the sum of the squares of whose distances from two intersecting lines is constant.
5. Find the locus of a point such that the square of its distance
from the base of an isosceles triangle is equal to the product of its
distances from the other two sides.

6. Find the locus of a point such that the length of a tangent

drawn from it to one of two given circles is a constant times the length
of a tangent drawn from it to the other circle.

7.

Find the locus of a point the sum of whose distances from two
is equal to the square of its distance from

fixed perpendicular lines

the point of intersection of the lines.

8. Find the locus of a point the sum of the squares of whose distances from two fixed perpendicular lines is equal to the square of
its distance from the point of intersection of the lines.

9.

two

Find the locus of a point the product of whose distances from

fixed intersecting lines is constant.

10. Find the locus of a point the sum of the squares of whose distances from two adjacent sides of a square is equal to the sum of the
squares of its distances from the other two sides.

Find the locus of a point which is the center of a

through a fixed point and tangent to a fixed line.
11.

12.

Find the locus of a point which

and to a fixed circle.

to a fixed line

233

is

circle

passing

the center of a circle tangent

The

Conies. Locus Problems

[Chap. 4

13. Given two parallel lines LI and L%, and a third line L3 perpendicular to the first two, find the locus of a point the product of
whose distances from LI and L3 is a constant times the square of its

distance from

L3

14. Given a circle which is tangent to a given line L at the point A

denote by B the other end of the diameter through A
through A
draw a line and denote by Q and R its points of intersection with the
circle and with the tangent to the circle at B. The locus of the point
of intersection of a line through Q parallel to the line L and of a line
through R parallel to the diameter AB as the line through A is rotated
about A is called the witch of Agnesi. Find an equation of the locus
;

and draw

graph.

Through the point

15.

=3x

and y

AB

its

in the points

for all the lines

16.

(2,

through

variable line

x
0) a line is drawn meeting the lines y
B. Find the locus of the mid-point of

A and

(2, 0).

makes with two

fixed perpendicular lines a

triangle of constant area. Find the locus of the point dividing in

constant ratio the segment of the variable line whose end points are

on the two
17.

fixed lines.

variable line

ABC, meeting

AB

and

is

drawn

AC

the locus of the intersection of

18.

set of parallel line

two fixed perpendicular

them in the ratio h k.

parallel to the base

in the points

BE and

D and E

BC

of a triangle

respectively.

Find

CD.

segments are drawn with their ends on

Find the locus of the point dividing

lines.

19. One side of each of a set of triangles is fixed in position and

length, and the opposite angle is of fixed size. Find the locus of the
centers of the inscribed circles of these triangles.

20. Find the locus of the intersection of the diagonals of rectangles

inscribed in a given triangle, one side of each rectangle being on the
same side of the triangle.

a rectangle ABCD line segments EF and GH are drawn pai

AB and BC respectively and with their end points on the sides
rectangle. Find the locus of the intersection of HF and EG.

21. In
allel

to

of the
22.

meet

Find the locus of the mid-points of the chords of a

a point on the circle.

in

234

circle

which

Locus Problems

Sec. 43)

23.

Given the base

AB

of a triangle

ABC,

find the locus of the

vertex C,
(a)

(b)
(c)

(d)
(e)

when
when
when
when
when

CM 2 CB 2 is constant
2
2
CA + CB is constant;
CA/CB is constant;
;

the angle at C is constant

the difference of the base angles
;

A and B

is

constant.

ABC,

Given the base A B and the opposite angle C of the triangle

find the locus of the point of intersection of the perpendiculars

from

A and B upon

24.

the opposite sides.

25. AB is a fixed chord of a circle and C is any point of the circle.

Find the locus of the intersection of the medians of the triangles ABC.

26. AB is a line segment of fixed length and position, and C is any

point on a line parallel to AB. Find the locus of the intersection of
the three altitudes of the triangle ABC; also the locus of the inter-

section of the medians.

Through each of two fixed points P\(x\, y\) and PaOfe, ^2) lines
drawn perpendicular to one another denote by A and B the points
which these lines meet the j-axis and #-axis respectively. Find the
27.

are
in

locus of the mid-point of

A B.

(Use the slope of either line as parameter.)

Show

that the locus of a point the tangents from which to an

perpendicular to one another is a circle. This circle is called
the director circle of the ellipse.
28.

ellipse are

29. Find the locus of a point the tangents from which to the
2
parabola y = 4 ax include an angle of 45. (Use the slopes m\ and m 2
of the tangents as parameters.)

Find the locus of the extremities of the minor axes of the

ellipses which have a given point for focus and a given line for directrix.
30.

31. Find an equation of the locus of

a point the product of whose distances
from two fixed points is a constant k 2
Observe that as k 2 takes on different
.

values, the curve varies in form ; these

curves are known as the ovals of Cassini.
When k is equal to half the distance
between the points, the curve is the

lemniscate (see

FIG. 41

29, Ex. 1).

235

The
32.

Conies. Locus Problems

Find the parametric equations of the

by a point P on the circumference of a

scribed

cycloid,

[Chap. 4

the locus de-

circle as the circle rolls

at the center C of
along a fixed line, using as parameter the angle
the circle formed by the line CP and the perpendicular to the line on
which the circle rolls.
33. Through each of two fixed points PI and P2 lines are drawn so
as to intercept a constant length on a fixed line below PI and P2 find
the locus of the intersection of the variable lines. Can the line PiP2
;

be parallel to the fixed line?

34.

Find the locus of the points of contact of the tangents drawn

on the line of the major

to a set of confocal ellipses from a fixed point

axes of the ellipses (see 36, Ex. 14).
35.

Given two concentric ellipses one within the other, and with
on the same lines, if P is the pole with respect to

their principal axes

the outer ellipse of a line tangent to the inner ellipse, find the locus of P.
36.

Find the locus of the centers of the conies

ax 2

when

a, h,

and

+ 2hxy +

g are fixed

and

by

+ 2 gy = 0,

b varies (see (42.6)).

236

CHAPTER

The Quadric

Surfaces

it

44. Surfaces of Revolution.

The Quadric

Surfaces of Revolution

When one thinks of a surface, among the first examples which

to mind are the plane and the sphere. In
17 we showed
that any equation of the first degree in x, y, z is an equation of
a plane. In consequence of (15.2) an equation of a sphere of

come

radius a

two cases the

+ =

2
origin is x
y2 z2 a 2 In these
coordinates of any point of the surface are a

and center at the

solution of a single equation in the coordinates.

Whether we

use rectangular coordinates, or spherical coordinates, or any

other, we say that the locus of a point in space whose coordinates
satisfy a single equation is a surface. When rectangular coordinates are used and one of the coordinates does not appear
in the equation, the locus is a cylinder whose elements are
parallel to the axis of this coordinate. For example, if z does not
appear in the equation, the locus is a cylinder whose elements

are parallel to the 2-axis. In fact, if (xi, y\, z\) is any point on
the surface, so also is (*i, y\ 2 2 ), where z 2 is any z, because the
equation imposes a condition upon x and y and none upon 2,
y

and

and y have been chosen to satisfy the equation

two and any z are coordinates satisfying the equation.
A line in space may be thought of as the intersection of two
planes. In fact, in rectangular coordinates a line is defined by
two equations of the first degree in x, y, and 2, as was done in
19. Thus a line is a special case of a curve, which may be
defined as the intersection of two surfaces that is, a curve in
space is defined by two independent equations. In the ry-plane
the circle with center at the origin and radius a has as an
after x

these

2
2
2
equation x + y = a but when thought of as a curve in space
From the viewpoint
it has in addition the equation 2 = 0.
of space the former of these equations is a circular cylinder
with the 2-axis for its axis, and consequently the circle under
,

consideration
2

is

the intersection of this cylinder and the plane

= 0.

Ordinarily the curve of intersection of two surfaces, neither

of which is a plane, does not lie entirely in a plane. Such a
curve is called a twisted, or skew, curve, and one which lies
entirely in a plane

is

called a plane curve.

239

Consider,

for

The Quadric

Surfaces

[Chap. 5

example, the curve defined in terms of a parameter

equations

(44.1)

>>

/,

2
,

by the

3
.

This curve meets the plane

ax
in the points for

+ by + cz + d = Q

which the parameter / has values which

the equation

at

satisfy

+ ^2 + Ct + ^ _ Q
3

Consequently at most three of the points of the curve lie in any

one plane. The curve defined by (44.1) is called a twisted cubic.
When a plane curve in space is revolved about a line in its
plane, the surface generated is called a surface of revolution. For
example, when a circle is revolved about a diameter, the surface
generated is a sphere. If a curve lying in the ry-plane is revolved about the #-axis, each point of the curve describes a
circle with center on the #-axis. Let P(x, y, z) be a representafor all other points of the same
tive point of such a surface
circle as P the coordinate x is the same and y and z are differ2
2
ent, but they are such that Vy + z is the same, since it is the
;

radius of the circle.

When P

in the ry-plane, y is the radius

Hence an equation of the surface is obtained from
and
on replacing y by
an equation of the curve in the plane z
is

2 = 0.

We apply this process to the parabola with the

= 4 ax in the plane 2 = 0, and get the equation
2
2
y + z = 4 ax
(44.2)

equation

y*

the surface generated is called a paraboloid of revolution. If the

parabola is revolved about the .y-axis, by an argument similar
to the above we have

y2
Squaring both

sides,

= 4 a V*2 + z2

we obtain the equation of

y* =

l6a

(x

+z

and consequently the surface generated

of the fourth degree.
fact that

when a

about which

it is

the fourth degree

2
),

is

said to be a surface

The preceding examples

curve

illustrate the

symmetric with respect to the line

revolved, the degree of the surface is the same
is

240

TKe Quadric Surfaces

Sec. 44]

of Revolution

as the degree of the curve, but when the curve is not symmetric
with respect to this line, the degree of the surface is twice that
of the curve. Another example of the latter case is afforded by
the rotation of the line ax

+ by =

about the

#-axis.

The

sur-

face generated is a cone, consisting of what the reader might

consider as two right circular cones with the #-axis as common

an element of one cone being the prolongation through its

vertex of an element of the other. An equation of the surface is

axis,

a 2x 2

(44.3)

According as the
the #-axis or the

ellipse

;y-axis,

b*(y

+ z2

~+^ =

).

l, z

X 2 4*
-f

Qr

a
a

>

and

ball,

b,

is

the

first

revolved about

the surface generated has as an equation

4- z
Z2

When

= 0is
Z2
z

V2
JL

^fr2

somewhat

surface has a shape

called a prolate spheroid]

_i

like

a foot-

the second surface

is

discus-like in shape, particularly when b is much smaller than

when a < b, the situation
a, and is called an oblate spheroid
;

is

reversed.

= is revolved
1, 2
^1
about the *-axis or the y-axis, the surface generated has as
According as the hyperbola

an equation

The

first

surface consists of

two parts analogous

is called a hyperboloid of revolution of two sheets

to Fig. 45,

and

the second sur-

shaped somewhat like a spool of endless extent, and is

called a hyperboloid of revolution of one sheet. When the asymptotes are rotated simultaneously with the hyperbola, they
generate cones with the respective equations
face

is

--iy

,AAG\

(44.6)

In the

n
=0,

:v_A

-^----0.

case the two sheets of the hyperboloid lie inside the

in the second case the
of the cone, one in each part

first

two parts
cone

lies inside

the hyperboloid.

241

The Quadric

Surfaces

[Chap. 5

EXERCISES
1.

y=

Show

x* z
9

2.

that the curve (44.1)

= x3

Show

y*

= z2

that for each value of

meets the curve

is

the intersection of the cylinders

x -

3 ty

the corresponding plane

+z-

(44.1) in three coincident points,

and that ordinarily

through a point in space not on this curve there are three planes each
of which meets the curve in three coincident points.
3.

Show

2
cylinder x

that the curve x

+ y2 =

a2

= a cos

and that

it is

t,

a sin

t,

bt lies

on the

not a plane curve.

Find an equation of the cone generated by revolving the line

+ 6 = about the #-axis. Find the vertex of the cone and
the curve in which the cone intersects the .yz-plane.
4.

2#

3;y

5. Find an equation of the torus generated by revolving about

2
2
b 2 = 0, z = 0, where b < a.
the >>-axis the curve (x
a) + y

6.

Find the locus of a point equidistant from the point

and the plane x

7.

plane
8.

points

-f

Find the locus of a point equidistant from the

z

Find the points

in

a constant.

which the surface x 2

+ y2

+7=

is

line

Find the points

met by the

and

and the

is

by finding the values of / for which x, y, and z given

are coordinates of a point on the surface.
10.

2-axis

Find the locus of a point the sum of whose distances from the
also the locus when the
c, 0, 0) is a constant
(c, 0, 0) and (

met by the

is

0)

= 0.

difference of these distances

9.

(a, 0,

= 0.

line

in

which the surface x 2

2 xy

by equations

+ 3 z2

(i)

5 y + 10 =

a
3 __ ^ + 2 _ z- 3
~~T~~ -2 ~~~2~'

interpret the result.

11.

Show

that the curve of intersection

x2

C of

the quadric

+ y 2 - 5 z2 - 5 =

5 =
and
and the plane x = 3 z lies on the cylinders y 2 -f 4 z 2
45 = 0. What are the respective projections of C on
4 x 2 -f 9 y 2
the coordinate planes, and what kind of curve is C?

242

Canonical Equations of tKe Quadric Surfaces

Sec. 45]

12.

Show

surface z 2

that the line

x-l=y-2=z+l

-xy + 2x + y + 2z-l = Q.

lies entirely

on the

Find equations of the projections upon each of the coordinate

4=
planes of the curve of intersection of the plane *
.y + 22
and the surface x 2 yz + 3 x = 0.
13.

45. Canonical Equations of the Quadric Surfaces

On interchanging x and
a special case of
ax2
(45.1)

z in (44.2), the resulting

equation

is

in

+ by2 = cz,

which a and b have the same

sign.

A plane z = k, where k is a
whose equations are

constant, intersects the surface in the curve

ax

(45.2)

The

+ by

= ck,

z =

k.

of these equations is an equation of an elliptical cyland consequently the curve is an ellipse. In like manner
k intersects the surface in the curve
a plane y
first

inder,

ax*

(45.3)

a parabola a similar result follows for a plane x = k.

a and b in (45.1) differ in sign, the curve of intersection
(45.2) is a hyperbola, and the curve (45.3) is a parabola. The
surface (45.1) is called an elliptic paraboloid or a hyperbolic pa-

which

is

When

raboloid according as a and b have the same or different signs

these surfaces are illustrated in Figs. 42 and 43 respectively.

FIG. 43

FIG. 42

243

The Quadric
Equations
equation

(44.4)

(44.5)

a, b, c,

[Chap. 5

are particular cases of the

+ by2 + cz2 = d.

ax 2

(45.4)

When

and

Surfaces

and d have the same

sign,

which

may

be taken

as positive in all generality, the intersection of the surface

k, y
k,
any one of the planes x

or z

=k

is

an

shown by'
to that employed

ellipse, as

a process similar
above. In this case the surface
called

When

by

is

an ellipsoid (see Fig. 44).

a and d are positive and b

and c are negative, a plane x

k
intersects the surface in an ellipse,
2

d
imaginary according as ak
and a plane
is positive or negative
FIG. 44
y = k or z = k intersects the surface
in a hyperbola. The surface consists of two sheets, since when
real or

k2

<

- the
plane x

=k

does not intersect the surface in real

points but in an imaginary ellipse.

hyperboloid of two sheets (see Fig. 45).
special case of this type of surface.

The surface is called a

The first of (44.5) is a

When

a,

b,

and d are

k intersects the surface

positive and c is negative, a plane z
k intersects it in a hyperk or y
in an ellipse and a plane x
bola. The surface is called a hyperboloid of one sheet (see Fig. 46).

The second

of (44.5)

is

a special case of this type of surface.

FIG. 45

FIG. 46

244

Canonical Equations of the Quadric Surfaces

Sec. 45]

When d in (45.4) is equal to zero, if (xi, y\ zi) is a point of

the surface, so also is (txi, ty\, tz\) a point of the surface for
every value of /. Since all these points lie on the line through
t

the origin and the point (xi, y\, z\)>

is a cone with vertex at the origin.

equation (45.4) with d

b

and

are negative.

in all generality

(45.6)

l-0
(A.^7\

b are positive and c negative.

intersects
the cone in an ellipse,
(^ 0)
k (^ 0) in a hyperbola.

of canonical, or type, forms of equa-

considered

paraboloid

elliptic

cz,

hyperbolic paraboloid

-4"I-

2- 4-r
2

-a

rea '

//

+~-^=

>

~ + ^ - ^ = 0,

?2

-5

c^

ft

hyperboloid of two sheets

10)
(45.
v

r point ellip-

= + l - 1, or 0;

-3

ma gi narY

hyperboloid of one sheet

(45.9)

~V

soid as d

!'

a2

02,

(45.8)

r2

real solution of

is (0, 0,

assume that a and

^+^ = 2
~ ~=2
n

The only

0) unless one or two of a,

In order to consider real cones, we may

In this case a plane z = k

and a plane x = k or y =
The following is a list
tions of the surfaces just
(45.5)

follows that the surface

it

quadric cone.

Other surfaces of the second degree are

+ |s =

(45.11)

(45.12)

(45.13)
(45.14)

(45.15)

2L

*2

3L

dry

~
x2

elliptic cylinder

1,

hyperbolic cylinder

= 0,

= 0,
a2

= 0,

parabolic cylinder

pair of intersecting planes

pair of parallel planes.

245

TKe Quadric

Surfaces

[Chap. 5

the above equations contain a term in x 2 and none

in x, it follows that if (x\ y\> z\) is a point of any one of the
is a
xi, y\, z\}, that is, the plane x =
surfaces, so also is (
of
of
a
of
14). Any plane
symmetry
plane
symmetry (see
Since

all

surface

is

called a principal plane of the surface. From equait follows that ellipsoids, hyperboloids,

tions (45.7) to (45.10)

principal planes, each two perone

another
to
they are the three coordinate planes
pendicular
for these equations. From equations (45.5), (45.6), (45.11),

and quadric cones have three

;

(45.12), and (45.14) it is seen that the paraboloids and elliptic

and hyperbolic cylinders have two mutually perpendicular
principal planes, as do also two intersecting planes; they are
for these equations.
and y =
Parabolic
the planes x =
one
two
have
and
parallel planes
principal plane, as
cylinders
is
follows from (45.13) and (45.15), for which the plane x =

the principal plane.

As a result of this discussion

it follows that any quadric surof

surface
the
second degree, including the
is, any
of
two
planes, has at least one
degenerate quadrics consisting
is
It
evident
geometrically that any plane
principal plane.

face, that

through the axis of a quadric of revolution is a principal plane.

For the surfaces with equations (45.7) to (45.10) the origin
that is, if (x\, y\, z\) is a point of any
is a point of symmetry
one of these surfaces, so also is the point ( *i,
jyi,
z\). For
an ellipsoid or a hyperboloid the point of symmetry is the midpoint of every chord of the surface through it such a chord is
called a diameter. Consequently the ellipsoids and hyperboloids
;

are called the central quadrics, the origin being the center in the
coordinate system used in this section. The point of symmetry

of a cone

is

the

indefinitely in

vertex,

and each element of a cone extends

both directions.

For elliptic and hyperbolic cylinders with equations (45.11)

and (45.12), each point of the z-axis is a point of symmetry.
For, if we denote by P (0, 0, z ) a point on the axis, and if x\ y\
are solutions of (45.11) or (45.12), then (x\ y\, zo + t) and
are points of the cylinder and are symmetric
Xi,
yi, ZQ
(
9

with respect to Po for every value of /. This line of points of

symmetry is called the axis of the cylinder in each case.

246

Sec. 45]

Canonical Equations of the Quadric Surfaces

EXERCISES
1.

Find the

foci of the ellipses in

6 x2
is

which the

ellipsoid

+ 3 y2 + 2 z2 = 6

cut by the coordinate planes.

Find an equation of the cone obtained by revolving the line

2y + l = Q, z = about the #-axis. Effect a translation of the
axes so that the resulting equation of the cone in the j'^Y-system
is of the form (45.10).
2.

3.

Show

that a real ellipsoid (45.7), for which a

2
by the sphere x

4.

Show

+ y2 + z =
2

b 2 in

two

<

<

c,

is

that the semi-diameter r with direction cosines X,

of an ellipsoid (45.7)

is

cut

circles.
/*

given by

i2 = ^?4_H?
+ !^2
2
b2

and prove that the sum of the squares of the reciprocals of any three
mutually perpendicular diameters of an ellipsoid is a constant (see
(30.2)).
5.

Show

that for the surfaces (45.7) to (45.10) the coordinate

symmetry (see 14). Find lines of symmetry of the
'

axes are lines of

other quadrics.
6. Show that, with the exception of equations of the central
quadrics and of quadric cones, the terms of the second degree in
equations (45.5) to (45.15) are the product of two real or imaginary

factors of the first degree. Does it follow that the same is

true of equations of these surfaces in any rectangular coordinate system ?

homogeneous
7.

Show by means
ax 2

of the argument preceding equation (45.5) that

+ by + cz 2 + 2
2

hxy

+ 2fxz + 2gyz = Q

an equation of a cone with vertex at the origin, or of two planes

through the origin and that in consequence of Theorem [41.5] it is
an equation of two planes, if and only if the determinant
a h f
is

is

equal to zero.
8.

Show

that xy

at the origin,

and

which the surface

+ yz + zx = Q

is

find the cosine of

is

intersected

an equation of a cone with vertex

an angle between the two lines in

by the plane ax
247

-f by

+ cz = 0,

The Quadric Surfaces

Determine the value of

9.

2 x

by

12 z

b so that

3 xy

2 xz

11

jy*

be an equation of two planes through the origin

shall

and

[Chap. 5

(see

Ex. 7)

find equations of the planes.

10.

Show

that an equation of the cone whose vertex

of a real ellipsoid (45.7) and which passes through

intersection of the ellipsoid and the plane Ix
my

is

all

at the center
the points of

+ nz =

1 is

+ = r2

z2
that the curve of intersection of the sphere x 2 y 2
and a real ellipsoid (45.7) for which a > b > c, lies on the surface
1 1.

Show

I
~~2
2

--1\1* +
r
2

/I
IS2
\6

--1\
^ +
r2
2

/I
~52

\c

A
--1\
U 2 = 0.
2
r /

Determine for what values of r this is an equation of a cone, and for

what values it is an equation of two planes (see Ex. 7) and show that
;

when

degenerate and real, the curve of intersection

of the ellipsoid and the sphere consists of circles. What is the interthis surface

section

when

is

this surface is degenerate

46.

and imaginary ?

The Ruled Quadrics

The equation (45.6) of the hyperbolic paraboloid

written in the form

may

be

/5
a
\a

Consider in this connection the two equations

(46.2)

+ b = 2 c*

=f
--|
a
b
k

For each value of k other than zero these are equations of a

line which lies on the paraboloid. In fact, any set of values of
a given value of k also satisfies (46.1),
seen on multiplying together equations (46.2) member by
member. And conversely, through each point of the paraboloid (46.1) there passes one of the lines (46.2), the corresponding
x, y, z satisfying (46.2) for

as

is

248

The Ruled Quadrics

Sec. 46]

value of k being obtained from either of equations (46.2) on

substituting in the equation the coordinates of the point. Con-

sequently the paraboloid has lying upon it an endless number

of lines, or rulings. In like manner, the lines whose equations are

M>

f-f-T

f+f-*

for values of / other than zero, lie on the paraboloid (46.1),

and through each point of the surface there passes a line of the
set (46.3). A surface having an endless number of lines lying
upon it, one through each point, is called a ruled surface. For
example, cones and cylinders are ruled surfaces. Since a hyperbolic paraboloid has two sets of rulings, it is said to be doubly
ruled.

In like manner,
equations are
(46.4)

lie

'
a

it

can be shown that the set of

on the hyperboloid
*

(46 5)1
.

b/

(45.8), as

whose

+ WH-J),
c
\

lines

do

also the lines

Consequently a hyperboloid of one sheet is doubly ruled, a line

of each of the sets (46.4) and (46.5) passing through each point
of the surface. Hence we have
[46.11

Hyperbolic paraboloids and hyperboloids of one sheet are

doubly ruled surfaces.

b, 2k are direction
By Theorem [20.2] we find that a,
numbers of the line (46.2), and a, b, 2/1 are direction numbers
of the line (46.3). If we denote by k\ and /i the values of k
and / for the lines through the point (x\, y\, z\) on the paraboloid (46.1), we have from (46.2) and (46.3)

(466^
(4b b)
'

^4-^--2c,
+ b ~ Z Ckl
a

'

249

*I-2l-?-.
a

b"

I,

TKe Quadric
The plane through
rulings through

Xi

that

[Chap. 5

(x\, y\, Zi) and containing the two

has as an equation (see 23, Exs. 6

the point

this point

7)

Surfaces

y\

-b

z\

2*i

= 0,

is,

When

the values of k\ and /i given by (46.6) are substituted in

this equation, the resulting equation is

Since
tion

y\ Zi) is a point of the paraboloid (46.1), this equareducible to

(x\,

is

The direction numbers of any line in this plane are a linear

homogeneous combination of direction numbers of the lines
20, Ex. 9). Consequently we have as
(46.2) and (46.3) (see
parametric equations of any line through (xi, y\, z\) and contained in the plane (46.7)

= xi +
y = yi-t
x

(h\a

+ h2d)t,

'

(46.8)

hi and A 2 determine the line, and / is proportional to the

distance between the points (*i, y\, z\) and (x, y, z). When
these values are substituted in (46.1) in order to find the values
of / for the point in which the line (46.8) meets the paraboloid,

where

the resulting equation is reducible, by means of (46.6) and the

fact that x\ 9 y\ 21 is a solution of (46.1), to
9

*i* 2 /

Thus when

= 0.

h\ or h% is equal to zero, that

one of the rulings, every value of

250

is,

when

the line

is

satisfies this equation,

The Ruled Quadrics

Sec. 46]

meaning that every point of the ruling is a point of the paraboand h 2 ^ 0, the two solutions of the above
loid. When hi ^
=
t
are
0, that is, each line for any values of hi and
equation
from zero, meets the surface in a doubly
different
both
*2,
counted point. Consequently every line in the plane (46.7)
through the point (xi, y\, z\), other than the rulings, is tangent
to the surface at the point; accordingly (46.7) is called the
tangent plane to the paraboloid at the point (xi, y\, z\} (see

Theorem

[48.7]).

EXERCISES
Find equations of the rulings on the paraboloid x 2
through the point (2,
also an equation of the tangent plane at this point.
1.

and

in particular of the rulings

2.

2 y2

in particular of the rulings

3.

1,

1)

find

Find equations of the rulings on the hyperboloid

x2

and

2y 2 -f 2 z = 0,

4 z2

= 4,

through the point

Find direction numbers of the rulings

(0, 2, 1).

(46.4)

and

(46.5) of a

hyperboloid of one sheet (45.8), and show, using equations analogous

to (46.8), that an equation of the plane containing the rulings through
the point (xi, y\ z\)
t

is

*i*

a2

which consequently
4.

is

4.
^

__

b2

a?
c2

i'

the tangent plane at

(x\ t y\, Zi).

Given the equation

>

b > c, determine the values of / for which the surface is an

a hyperboloid of one sheet, a hyperboloid of two sheets;
2
Show also that each
discuss in particular the cases / = b 2 and / = c
of the coordinate planes intersects the set of quadrics (i) in confocal
conies (see 36, Ex. 14). This set of quadrics is called a set of confocal

where a

ellipsoid,

quadrics.
5. Show that through every point in space not on one of the coordinate axes there pass three of the confocal quadrics of Ex. 4, one an
ellipsoid, one a hyperboloid of one sheet, and one a hyperboloid of

two

sheets.

251

The Quadric
Show

6.

that

xz
~2

>

Surfaces

[Chap. 5

z2

y
=k
+ -fi--2

an equation of a family of hyperboloids of one sheet

k < 0, an equation of a family of hyperboloids of two
In what sense is
sheets ; for k
0, an equation of a quadnc cone.
this cone the asymptotic cone of the two families of hyperboloids ?
k

for

7.

is

32)

(see

f or

Show

that for a suitable value of

in

terms of k each of the

planes

_
b

passes through a ruling (46.4), and consequently equations

equations of the ruling. Show also that

be

are

a)

(i)

are equations of the rulings (46.5).

\VHose Principal Planes Are Parallel

47. Quadrics

to tKe

Each

Coordinate Planes

of equations (45.5) to (45.15)

is

a special case of an

equation

ax 2

(47.1)

+ by + cz + 2lx + 2my + 2nz + d = Q,

2

with the understanding that a, b, c in this equation are not

the same as in equations (45.5) to (45.15). For example, if
d~Q and we replace a by I/a2 b by I/ft2 and n
l = m
c
,

by

c,

(47.1),

and

we have equation

(45.5).

Now we

shall

prove that

whatever be the coefficients provided only that a, 6,

all zero, is an equation of a quadric or degenerate

are not

quadric, by showing that by a suitable transformation of coordinates equation (47.1) can be transformed into one of the

forms (45.5) to (45.15). We consider the three cases, when

are different from zero, when one of a, b, c is equal to
zero, and when two of them are equal to zero.
Case 1. a i 0, b ^ 0, c ^ 0. On completing the squares in
x, y, and z in (47.1), we have
a, b, c

252

Sec. 47]

Principal Planes Parallel to Coordinate Planes

where

= - + ?~ + --d.
a

If

then we effect the translation of axes defined by

'

(47.2)

x'

'

'

we obtain the equation

+ by' 2 + cz' 2 = k.

ax' 2

(47.3)

If k j*
and a, b, c have the same sign, the surface is a real
or imaginary ellipsoid. If k ^
and two of the numbers a/k,
and
one is negative, equation (47.3) is
b/k, c/k are positive
either of the form (45.8) or one of the forms

_^+^
+
a

?!

b2

*!_

a2

b2

=
+ ?_
c
2

Consequently the surface is a hyperboloid of one sheet. If k^

and one of the numbers a/k, b/k, c/k is positive and the other
two negative, equation (47.3) is either of the form (45.9) or
one of the forms

2
**.

-4-Z

b2

^=
c

==
_^!_^4.5?
2
2
2

'

ft

Consequently the surface is a hyperboloid of two sheets. In

every case, as follows from (47.2), the center of the surface is
I /a,
at the point (
m/b,
n/c).
and a, b, c have the same sign, the surface is a point
If k =
ellipsoid (an imaginary cone). If k~Q and one or two of a,

and the others negative, the surface is a quadric

at the point ( I /a,
the
vertex
with
cone
m/b,
n/c).
0. On completing the squares in
a
b
2.
Case
5* 0,
7* 0, c
x and y, we have
b c are positive
y

where
If
first

2
2
= d ----m
I

j-

n T
and we effect the transformation defined by the
two of equations (47.2) and z' = z + k/2 n, we obtain
ax'

+ by' + 2nz' = 0,
2

253

The Quadric

Surfaces

[Chap. 5

which is of the form (45.5) or (45.6) according as a and b have the

same or opposite signs. Consequently the surface is a paraboloid.
and we effect the transformation defined by the
If n =
first two of equations (47.2) and z' = z, we obtain
ax' 2

we have

cylinder
0.
k

if

with equations (45.11), (45.12), and

that the surface is an elliptic or hyperbolic

this equation

Comparing
(45.14),

if

+ by'2 + k = 0.

0,

and two intersecting planes,

real or imaginary,

If one of the coefficients a, b, c other than c is equal to zero,

on using the above methods we get an equation obtained from
one of the equations (45.5), (45.6), (45.11), (45.12), and (45.14)
on interchanging x and 2, or y and z, as the case may be, with

the corresponding interpretation.

0. When the equation
c
Case 3. a 5* 0, b
the form

= =

(47.4)

afx

is

written in

+ ^\ + 2 my + 2 nz + k = 0,
72

where
if

>

not both

defined

and n are zero and we

effect the

transformation

by
a

^/m 2

the resulting equation

ax' 2

(47.5)

+ n2

is

+ 2 Vm 2 + n / = 0.
2

Consequently the surface is a parabolic cylinder.

If m = n = 0, it follows from (47.4) that the surface

is

de-

generate, consisting of two parallel, real or imaginary, planes

if k 7* 0, and of two coincident planes if k
0.

When any two

of the coefficients

similar results follow.

[47.1]

a,

ft,

are equal to zero,

Hence we have

An equation ax 2 + by 2 + cz 2
in which not all the coefficients a, b, c are zero, is an equation
of a quadricj which may be degenerate; its principal plane
or planes are parallel to the coordinate planes.

254

Sec. 47]

The

Principal Planes Parallel to Coordinate Planes

last part of this

theorem follows from the

fact that in

each

case equation (47.1) was transformed by a translation (30.1)

into one of the forms (45.5) to (45.15), except in Case 3, where,
however, the y'z '-plane is parallel to the yz-plane.

EXERCISES
1.

Show

that a surface with an equation

a(x

which

in

as center
sign,
2.

and

is

an

2
)

+ c(z - z

ellipsoid

when

2
)

1,

with the point (XQ, y Q z )

do not all have the same
,

a, b, c

that a surface with an equation

*o)

b(y

-y

+ c(z - z = 0,
)

an

elliptic or hyperbolic paraboloid according

b have the same or different signs. Find equations of the

a, 6, c

is

What

is

the surface

when

by each of the following equations,

position relative to the coordinate axes

Identify the surface defined

and determine

its

9 x2
x2

c.

e.

+ 4 y 2 + 36 z 2 - 36 x + 9 y + 4 = 0.
- 4 y 2 - z 2 - 4 x - 24 y + 4 z - 32 = 0.
x 2 + 4y 2 + 4x-8y-6z+U = Q.
x 2 -4z 2 + 5y-x + Sz = Q.
2
2
y -4 z + 4;y + 4z + 3 = 0.

a.

b.

d.

4.

-y

b(y

find equations of the principal planes.

principal planes.
3.

discuss also the cases

Show

which
as a and

2
)

are positive,

a, b, c
;

a(x
in

For what values of

ax 2

(i)

and d

a, b, c,

by

cz

is

the surface

+ d=

a real ellipsoid for what values a hyperboloid of one sheet

values a hyperboloid of two sheets ?
;

for

what

44,

Find the projection upon each of the coordinate planes of the

and x -3y + 2 2 =
of x 2 - 2 y 2 - 3 z 2 - 6 =
(see
Ex. 11). Of what type is the curve of intersection?

6.

Find an equation of the projection upon each of the coordinate

5.

intersection

planes of the intersection of the central quadric (i) of Ex. 4 and the
plane Ix + my + nz = 0, and identify each of these curves.

255

The Quadric
7.

Show

that,

if

[Chap. 5

the line with equations

x = xi + ut,

(i)

Surfaces

y=yi +

vt,

= zi +

wt

meets a central quadric with equation (i) of Ex. 4 in two points, the
point (xi, y\ z\) is the mid-point of the intercepted chord if it lies in
t

the plane

aux

(ii)

-f

bvy

+ cwz

0.

Why is this plane appropriately called a diametral plane of

and under what conditions

is it

the quadric,

perpendicular to the chords which

it

bisects ?
8. Show that if (x\, y\ *i) is a point on the intersection of the
plane (ii) of Ex. 7 and the quadric (i) of Ex. 4, equations (i) of Ex. 7
are equations of a tangent to the quadric at the point (x\ y\, z\).
9

9. Find an equation of the plane which is the locus of the mid2

3 y2
2 z whose
points of the chords of the elliptic paraboloid x
direction numbers are 1, 2, 3.

Let Pi and P2 be two points on a real ellipsoid (45.7) such that

on the diametral plane of chords parallel to the line OP\, where
the center (0, 0, 0)
prove that PI is on the diametral plane of

10.

P2

is
is

chords parallel to

48.

OP2

(see

Ex.

7).

The General Equation of trie Second Degree

and z. The Characteristic Equation.

in x, y,

Tangent Planes

is

to

a Quadric

The most general equation of the second degree in x, y,

of the form

(48.1)

We

ax 2

and

+ by2 + cz2 + 2 hxy + 2fxz + 2 gyz

+ 2 IK + 2 my + 2 nz + d = 0.

show that any such equation,

for which not all the

terms of the second degree are zero, is an equation of a quadric. We prove this by showing that in every case
there exists a transformation of coordinates by which the equation is transformed into an equation of the form (47.1), which
was shown in 47 to be an equation of a quadric. In 45 we
observed that any one of the surfaces (45.5) to (45.15) has at
shall

coefficients of the

256

The General Equation

Sec. 48]

Second Degree

of tne

one plane of symmetry, and in Theorem [47.1] that the

plane or planes of symmetry for a quadric with an equation
(47.1) are parallel to the coordinate planes. Accordingly we seek
the planes of symmetry of the surface with an equation (48.1).
least

The equations

= xi + ut,

(48.2)

= yi + vt,

= zi + wt

are parametric equations of the line through the point (x\, y\, z\\
and with direction numbers u, v w. If we substitute these exy

pressions for x y,
t

and

z in (48.1)

and

collect the

terms in

and

/,

we have
(au

(48.3)

+ bv 2 + ctv2 + 2 huv + 2futv + 2 gvw)t2

+ 2 [(ax, + hy, +fz, + l)u + (hx, + byi + gz + m)v
l

where F(x\,
of (48.1)

denotes the value of the left-hand

y\, z\)

when

x, y,

and

z are

For fixed values of

tively.

replaced by xi,

y\ and
y

member

z\

respec-

equation (48.3) is
whose roots are proportional to the
xi, y\, z\

u,

v,

a quadratic equation in /,
distances from (jci, y\ z\) to the two points of intersection, if
any, of the surface (48.1) and the line with equations (48.2). In
order that (x\, y\, z\) shall be the mid-point of the segment
between these two points, the roots of this equation must differ
hence the coefficient of / must be equal to zero,
only in sign
9

that

is,

(48.4)

As a

(0*1

first

[48.1J

+ hyi +fzi + l)u + (hxi + byi + gzi + m)v

+ (fxi + gy\ + czi + ri)w = 0.

consequence of this equation we prove the theorem

The mid-points of any

with an equation (48.1)

set of parallel
lie

chords of the surface

in a plane.

the quantities u, v, w have constant values for a

on reassembling the terms in (48.4), we
have that the coordinates of the mid-point of each chord satisfy

In

fact, since

set of parallel chords,

the equation
(48.5)

(au

+ hv +fw)x + (hu + bv + gw)y

+ (fu + &v + cw}* + (lu + mv + nw) = 0,

which evidently

is

an equation of a plane.
257

The Quadric

Surfaces

[Chap. 5

If this plane is to be a plane of symmetry (see

45) of the
surface (48.1), w, v, and w must be such that the chords are
normal to the plane. By Theorem [17.5] the coefficients of
x, y, and z in (48.5) are direction numbers of any normal to

the plane.

Accordingly,

the chords are to be normal to the

if

we must have

plane,

+ hv +fw _ hu + bv + gw _fu + gv + cw __ r

au

r denotes the common value of these ratios.

write these equations in the form

when

we have by Theorem

[22.1] that

and only

v,

w, not

all zero, if

they admit a common solution

r is such that

b-r

c-r

/
r

if

!-r
(48.8)

Thus

When we

hu+(b- r)v + gw = 0,
= 0,
fu + gv + (c
r)w;

(48.7)

'

= 0.

must be a solution of the cubic equation

r3

(48.9)

the coefficients

7, /,

and

7r 2

+ Jr - D = 0,

being defined by

/ = a + b + c,

(48.10)

(48.1).

(48.8) is called the characteristic equation of equation

Equation
tion

We now

by means

of

derive properties of the characteristic equawhich we shall prove that (48.1) is an equa-

tion of a quadric.

Let

and

r\

denote by

r2

u\, v\,

be two different roots of this equation, and

w\ and w 2 #2, w 2 solutions of equations (48.7)
,

as r takes the values

(48.7)

we

replace w,

v,

and r 2 respectively. If in equations

w, and r by MI, v it u>i, and TI respectively
r\

258

TKe

Sec. 48]

Characteristic Equation

and form the sum of these equations after multiplying them by

2, 02, and tu 2 respectively, we obtain

ri

+ ViV + tVitV = 0.
2

2)

manner, we replace u, v, w, and r in (48.7) by

r 2 and form the sum of these equations after

in similar

If,

u2

(UiU 2

w 2 and

v2,

multiplying them by Wi, v\, and w\ respectively, we obtain

the above equation with the exception that r\ is replaced by
r2
Having obtained these two equations, if we subtract one
from the other and note that by hypothesis r\ ^ r 2 we have
.

UiU 2 + v\v 2 + w\w 2 = 0.

(48.11)

With the

we

aid of this result

shall establish the

important

theorem
[48.2]

The

roots of the characteristic equation of

whose

the second degree

an equation of

coefficients are real are all real

numbers.

= +

<r
ir,
For, suppose one of the roots were imaginary, say r\
V^l. From the thewhere o- and r are real numbers and i

ory of algebraic equations

root.

If

we

follows that r 2

it

ir also is

by r\ and solve the resulting

and w\ by means of Theorem [22.1], we

replace r in (48.7)

equations for ui, v\

have that at least one of these quantities
9

is

= +

imaginary;

for,

a
ir and we equate
in (48.7) everything is real but r
to zero the real and imaginary parts of each equation, we have

if

that r

=
MI

= or

where the
is

= v = w = Q. Accordingly we put
fi =
+ *02, wi = 71 + 172,
+ UK2

unless u
i

a's,

/3's,

= cti

real numbers. When r in (48.7)

a solution of the resulting equations is

and 7*5 are

replaced by a

u2

181

ir,

ia 2

v2

1*182,

M> 2

(48.7) for r 2 is

since the set of equations

above set on changing i to
it

j8i

i.

71

obtained from the

For these two

sets of solutions

follows from (48.11) that

c*i

2
/3i

7i

+ <*2 2 + fo 2 + 72 2 = 0,
259

172,

The Quadric
which evidently

is

equation are real

impossible, since

and not

As a consequence

proved.

Surfaces
all

[Chap. 5

the quantities in this

Hence Theorem [48.2] is

result and equation (48.11)

all zero.

of this

we have
[48.3]

Planes of symmetry of the surface (48.1) corresponding to

two unequal roots of the characteristic equation are perpendicular to one another.

Another theorem which we need in the proof that every

equation of the second degree is an equation of a quadric is the
following

Not

[48.4]

all the roots of the characteristic

equation of an equa-

tion of the second degree with real coefficients are equal

to zero.

In order to prove it, we observe that if all the roots are zero,
3
= 0. In
equation (48.9) must be r = 0, that is, I =
=
7*
2
and
this case
from (48.10) we have
/ 0,

J=D

2/

72

= a 2 + b 2 + c 2 + 2 /2 + 2

+ 2 h2

This expression can be zero only in case a, b, c, /, g, h are all

zero, in which case equation (48.1) is not of the second degree;
and the theorem is proved.
From Theorems [48.2] and [48.4] it follows that at least one
of the roots of the characteristic equation is a real number
different from zero. For this root equations (48.7) admit at
least one solution u, v, w, not all zero, since (48.8) is satisfied.
If then the principal plane having equation (48.5) with these
values of u, v, and w is taken as the plane z'
of a new reccoordinate
the
transform
of
tangular
system,
equation (48.1)
does not involve terms in x'z', y'z', or z', since, if (x\, y\ z\]
is on the surface, then fa', yi',
z\} is also. Hence the equation is of the form
y

(48.12)

a'*'

+ b'y' 2 -f c Y 2 + 2 h'x'y' + 2 /'*' + 2 m'y' + d' = Q,

where some of the

coefficient*

may be

260

equal to zero, but not

TKe General Equation

Sec. 48]

those of the terms of the second degree.

all
is

Second Degree

of the

of the form (47.1).

If h'

and we

If h'

= 0,

(48.12)

effect the rotation of the

axes about the z'-axis defined by

= K " cos 6 - y" sin 0, / = *" sin 6 + y" cos 0,

where 6

is

a solution of the equation (see

h'

tan

6+

(a'-

tan

[48.5]

= 0,

A'
is

of the form (47.1).

equation of the second degree in x, y, and z

equation of a quadric, which may be degenerate.

Any

As a consequence

is

an

we have

of the preceding results

When any plane cuts a quadric, the

a conic, which may be degenerate.

[48.6]

2",

40)

the resulting equation in #", y", and z"

Hence we have established the theorem

z' ==

curve of intersection

is

given any plane, in a suitably chosen coordinate systhe plane z = 0. Solving this equation simultaneously
with (48.1), we have an equation of the second degree in x and

In

fact,

tem

it is

y which

is

an equation of a

conic, as

shown

in

40.

We now make

use of equation (48.3) to derive an equation

of the tangent plane to a quadric (48.1) at a point (x\, y\, 21),
proceeding as was done in 46 for a hyperbolic paraboloid with
equation (45.6). When (xi, y\, z\) is a point of the quadric
in (48.3). In order that the line (48.2)
(48.1), F(XI, yi, zi)

be tangent to the quadric, / = must be a double root of

(48.3). Consequently we must have equation (48.4) satisfied

shall

by

u,

v,

and w

and

for

any

set of values of u,

v,

and w

satisfy-

ing (48.4) the corresponding equations (48.2) are equations of

a tangent line. This means that any line through (xi, y\ z\)
9

perpendicular to the line through this point

numbers

+ hyi +fzi +
hxi + byi + gz + m,
f*i + gyi + czi + n,
0*1

(48.13)

I,

261

and with

direction

The Quadric
is

tangent to the quadric at

siderations

we have

that

Surfaces

(#1, yi, *i).

all

ically

by
x

xi,y
Hence we have

lines,

yi,z

geometric conin a plane,

lie

We prove

this analyt-

any point on one of these

are direction numbers of the line.

if (x,

z\

From

these tangents

called the tangent plane to the quadric.

observing that

[Chap. 5

y, z) is

+ gz + m)(y-yi)
+ gyi + czi + ri)(z- zi) = 0,
1

which

is

an equation of a plane.

On

multiplying out the terms in this equation and making

use of the fact that x\, y\, z\ is a solution of (48.1), we have

An

[48.7]

equation of the tangent plane to the quadric (48.1) at

the point (x\> ji, Zi) is

axix

(48.14)

+ byiy + cz^z + h(x y + yix) +f(xiz + zix)

}

Since (48.13) are direction numbers of the normal to the

tangent plane at the point (x\, y\, 21), called the normal to the
quadric at this point, this normal has equations

_ __
x-xi

y-y\

hx\

fxi

+ by\ + gz\ + m

+ gyi + CZI + H

EXERCISES
1.

Find the roots of the characteristic equation of the quadric

2x 2 +

5jy

+ 3z 2 + 4*3>-3;c + 4>>-6z-3 = 0;

determine the corresponding planes of symmetry, and a system of coordinates in terms of which the equation has one of the canonical

forms of
2.

45.

Show

terms in xy,

when an equation of a quadric does not involve

and yz the roots of the characteristic equation are the

that
xz,

coefficients of the second-degree terms.

equations (48.7) in this case?

262

What

are the solutions of

Tangent Planes

Sec. 48]
3.

Show

to

a Quadric

that the characteristic equation is unaltered by a transis, by a transformation to parallel axes.

lation of the axes, that

Prove that the normals to a central quadric ax 2 + by 2 + cz* = 1

at all the points in which the quadric is met by a plane parallel to
one of the coordinate planes intersect two fixed lines, one in each of
4.

the other coordinate planes and parallel to the intersecting plane.

5.

Denote by Q the point in which the normal to a real ellipsoid

a point P meets the plane z = 0. Find the locus of the mid-

(45.7) at

point of PQ.
6.

point

Find equations of the tangent plane and of the normal at the

(1, 1,

2) to the quadric

+ y 2 - z 2 + 2 xy + xz + 4 yz - * + y + z + 4 = 0.
7. Prove that the line * - 2 = 0, 2-1=0 lies entirely on the
2 yz
6 z -f 6 = 0. Find an equation of the
3 x
quadric xy + 3 xz
tangent plane to the quadric at the point of this line for which y = y\
x2

and show that the

line lies in the

tangent plane, and that as yi takes

one gets a family of planes through the line. May

one conclude from this that the quadric is neither a cone nor a cylinder ?
different values

8.

Find the conditions to be

(48.2) is a ruling of the

that

is,

lies

entirely

satisfied

by

u, v,

so that the line

quadric (48.1) through the point

on the quadric

(x\, y\, *i),

and show that when these con-

ditions are satisfied, the line lies entirely in the tangent plane to the

quadric at the point

(x\, y\, z\).

9. Show that any of the quadrics (45.5) to (45.15) is cut by a set

of parallel planes in similar conies and that the principal axes of any
two of these conies are parallel. Is this true of a quadric with equa-

tion (48.1)?
10.

Show

that the normals at

through a point

to the three confocal quadrics

are mutually perpendicular (see

46, Ex. 4).

11. The plane (48.14) is called the polar plane of the point
Pi(*i, yif zi) with respect to the quadric (48.1), whether PI is on the
quadric or not ; and PI is called the pole of the plane. Show that if

the polar plane intersects the quadric and P3 is any point on the intersection, the tangent plane to the quadric at P2 passes through Pi,

and the

line

through Pi and P2 is tangent to the quadric at

all such lines through Pi is a quadric cone,

that the locus of

263

P3

also

The Quadric
Show, with the aid of

12.

(48.3),

Surfaces

that

if

[Chap. 5

an endless number of tan-

gents can be drawn to the quadric (48.1) from a point PI(XI, y\, Zi)
not on the quadric, the points of contact lie in the polar plane of the
point Pi.

Where do such points Pi

lie

relative to each type of quadric ?

Find the pole of the plane 3x-2y - 3 2

2 xz + 4 yz
to the quadric x 2 -f y 2 + z 2 + 4 xy
13.

=
= 0.

with respect

Show

that the poles with respect to the quadric (48.1) of the

planes through a line are collinear.
14.

15. Show that if PI(XI, y\, z\) is a point from which tangents cannot be drawn to a quadric (48.1), the poles of three planes through PI
and not having a line in common determine the polar plane of PI
34, Ex. 13).
(see

49. Centers.

Vertices.

Points of

Symmetry

Having shown that any equation of the second degree in

and z is an equation of a quadric, we shall establish in
this section and the next criteria which enable one to determine the particular type of quadric defined by a given equation.
As a step in this direction we seek the conditions upon the
x, y,

equation (48.1) in order that the quadric have a

that there be a point (x\,y\,z\) not on the quadric
the mid-point of every chord (48.2) through the point,

coefficients of
center,

that

is,

which is
whatever be

u, v, and w. Referring to equation (48.4) and the

discussion leading up to it, we see that x\ y\, z\ must be a
common solution of the equations
9

+ by + gz + m = 0,
fx + gy + cz + n = 0.
hx

(49.1)

In accordance with Theorem [21.8] these equations have one

and only one common solution when their determinant Z), defined by (48.10), is different from zero. Since the property that

a point
used,

is

a center

we may

is

independent of the coordinate system

when the coordinate system is

consider the case

such that / = g = h = 0. In this case D = abc, that is, the

2
2
2
Applying the above
product of the coefficients of x y and z
264
,

Centers. Vertices. Points or

Sec. 49)

test to equations (45.5) to (45.15),

boloids,

we

Symmetry

see that ellipsoids, hyperfor which

and
j*

and cones are the only quadrics

in the coordinate

system of these equations the point whose

coordinates satisfy (49.1) is the origin. In the case of a cone
the solution of (49.1) is a point of the cone, the vertex, so
that ellipsoids and hyperboloids are the only quadrics which
have a center.

In order to discuss the case D ^

more fully, and also the
D = 0, we denote by F(x, y, z) the expression which is the
left-hand member of equation (48.1), and note that it may be
written
case

F(x, y, z)

(49.2)

+ hy +fz + l)x + (hx + by + gz + m)y

+ (fx + gy + cz + n)z + lx + my + nz + d.

(ax

a solution of equations (49.1), and the point is

(48.1), it follows from (49.2) that x\, y\, z\ is a
solution also of the equation
If xi, yi, z\ is

on the quadric

Ix

(49.3)

From Theorem

+ my + nz + d = 0.

[27.2] for the case in

which

w = 1 it follows
common solu-

that this equation and equations (49.1) have a

tion,

if

and only

if

A = 0,

where by definition

so defined

is

we have

an
cone when

z).

From

the

the theorem

When D ^ Q for an
tion of

called the discriminant of F(x, y,

foregoing results
[49.1]

equation (48.1), the

latter is

an equa-

A^

0, and of a
ellipsoid or a hyperboloid when
0. In either case the point whose coordinates

A=

are the unique solution of equations (49.1)

symmetry of

is

a point of

the quadric.

theorem follows from the fact that if

for a value of / the values of x, y, and z given by (48.2) are coordinates of a point on the quadric, so also are the values when
/ is replaced by
/, as follows from (48.3) and (49.1).

The

latter part of this

265

The Qua dric

Applying the method of
have in place of equations

21

(p.

Surfaces

110) to equations (49.1),

(21.7), (21.8),

Dy - M = 0,

Dx - L = 0,

[Chap. 5

and

Dz

we

(21.9)

- N = 0,

M, and Af are the cofactors of /, m, and n respectively

and D is seen to be the cofactor of d. Hence if D = 0,
equations (49.1) do not admit a common solution unless
where
in

L,

= M=N = 0,

(49.5)

A = 0.

In accordance with the discussion in 24,

(49.1) admit a common solution,
either meet in a line or are cointhere
is
a
that is,
line or a plane all of whose points
cident
of
of
the quadric. Accordingly we have
are points
symmetry
in

which case

when D =

and equations
the planes defined by (49.1)
;

[49.2]

When

D=

for an equation (48.1) of a quadric, and

also;
equations (49.1) admit a common solution, A
and the quadric so defined has a line or plane of points

of symmetry.

Since in either case there is at least one line of points of

symmetry, on choosing an #yz'-system of coordinates in which
this line is the z'-axis, we have that the corresponding equations
= 0, / = 0, and any value of
(49.1) must be satisfied by x'
z', and consequently

If h'

by a suitable rotation
we obtain a coordinate system in

in the equation of the quadric,

of the axes about the z'-axis

terms of which an equation of the quadric

a'*' 2

Comparing
(45.14), and
[49.3]

is

+ b'y' 2 + d = 0.
f

equation with equations

(45.15), we have the theorem

this

When D

(45.11),

(45.12),

for an equation (48.1) of a quadric, and

equations (49.1) have a common solution, the quadric is an
elliptic or hyperbolic cylinder, or is degenerate in the case

when

this

common

solution satisfies equation (49.3).

266

The

Sec. 50]

Invariants

D, and A.

I, /,

All types of quadrics except paraboloids and parabolic cylinders are covered by Theorems [49.1] and [49.3]. Hence we have
[49.4]

Paraboloids and parabolic cylinders do

points of symmetry.

not

have any

From the fact that the above theorems discriminate between geometric properties of the quadrics, it follows that if
D is not or is equal to zero in one coordinate system, the
same is true for any coordinate system the same observation
;

applies

and

In the next section we shall show that D

are invariants under any transformation of rectangular
to A.

coordinates.

50.

The

Invariants

I,

/,

D, and Zl

we show

that 7, /, D, and A, as defined by

under any change of coordiare
invariants
(49.4),
(48.10)
a
when
in
the
sense
that
nates,
general transformation of coordinates (30.10) is applied to equation (48.1) each of the above

In this section

and

functions of the coefficients in (48.1) is equal to the same function of the coefficients of the equation in the new coordinates.

We
(48.1)

denote by f(x, y, z) the terms of the second degree

and consider the expression

in

(50.1)

that

is,

(50.2)

(a

- r)x2 +(b- r)y* + (c- r)z2 + 2hxy + 2fxz + 2 gyz.

From Theorem

[41.5], which applies also to a homogeneous

the
second degree in three unknowns, it follows
of
expression

that the equation

(50.3)

is

a-r

b-r

C-T

the condition that the expression (50.2) be the product of

267

The Quadric
two

linear

factors

homogeneous

Surfaces
that

is,

[Chap. 5

each root of the

for

cubic equation (50.3), written in the form

r3

(50.4)

Ir 2

+ Jr - D = 0,

the equation obtained by equating (50.2) to zero is the equation of two planes through the origin.
When a transformation (30.11) is applied to (50.2), the latter
is

transformed into
/'(*',

(50.5)

where /'(*', /,
f

(a

z') is

/,

r(*

the transform of f(x, y,

- r)x' 2 + (6' - r)/2 +

z)

that

is,

(50.5) is

- r)z' 2 + 2 h'x'y' + 2f'x'z + 2 g'y'z'.

f

(c'

If (50.2) is the product of two homogeneous factors of the first

degree in x, y, z each of these factors is transformed by a
transformation (30.11) into a factor of the same kind, that is,
(50.5) is the product of two such factors, and consequently we
y

have, analogously to (50.4),

r3

7'r 2

+ J'r - D = 0,
f

where 7', /', and D' are the same functions of a', &',, h' as
7, /, and D are of the corresponding coefficients without primes.
Since r is unaltered by the transformation, this equation and
(50.4) must have the same roots, and it follows that 7 == 7',

/=

/',

D=

D'.

Since a general transformation of axes may be

first a translation (30.1) and then a rota-

obtained by applying

and since the former transformation does not affect

the coefficients of the second-degree terms in (48.1), we have
established the theorem
tion (30.11),

[50.1]

7, /, and D of the coefficients of the secondterms

an
degree
of
equation of the second degree in three unknowns are invariants under any change of rectangular

The functions

coordinate axes.

We apply this theorem to the consideration of the case when

two of the roots of the characteristic equation (50.4) are equal
and not zero, that is, one of the roots, say r : is a double root,
,

268

The

Sec. 50]

and

ri ?* 0.

Invariants

I,

it first

Considering

/,

D, and

as merely a root of the charac-

we

find a solution of equations (48.7),

take the principal plane so determined for the plane *'
teristic equation,

and

of

a new rectangular coordinate system. Whatever be the other

coordinate planes, in the transformed equation we have
h' = /' = 0, so that the new characteristic equation is
a'

~r

V-

(50.6)

= 0.

g'

c' -

g'

In this coordinate system equations (48.7) are

(a'
'

Since

w + (&' - r)v + g'w; = 0,

+ w = 0,
Ott + g'i>+(c'-i> = 0.

1, 0,

numbers

are direction

= 0,

for this solution of (50.7) r

ri, the first of (50.7) is satisfied

by
and the

last

This

is

as

seen

If

- r)u +

is

two by

0, g', ri

the condition that

when

(50.6)

j\

any normal to the plane

If

we denote

by any values

of u,

this root
v,

and

w,

provided that

b' y

(=

of

= a'.

a'}

be a double root of

(50.6),

written in the form

is

a plane whose normals have direction numbers 0, g', r\

V
taken for the plane y = 0, without changing the coordinate
r

is

in this new jyz'-coordinate system, equations (50.7) must

be satisfied by 0, 1, 0, which are direction numbers of any
normal to the plane y
0, and consequently in this system
V = TI, g = 0, and c = r 3 If then any plane perpendicular to
= is taken for the plane z = 0, in
and
the planes x' =
this final coordinate system the transform of (48.1) is
x',

ri(x'

+ / 2 + r3 2'2 + 2 n'z' + d = 0,
f

there being no terms of the first degree in x' and y', since the
and y' = are planes of symmetry. Any plane
planes x' =
z'
const, which meets the quadric intersects it in a circle

269

The Quadric

Surfaces

[Chap. 5

(which may be a point circle or an imaginary circle). Consequently the quadric is a surface of revolution. Since the
geometric character of the surface is independent of the coordinate system, we have the theorem

When a

[50.2]

of

nonzero wot of the characteristic equation (48.8)

(48.1) is a double root, the quadric is a sur-

an equation

face of revolution; for this root equations (48.7) admit an

endless

number of

solutions defining principal planes, all

passing through the axis of the surface.

We

proceed

now

to the proof that the discriminant

of

F(x, y, 2), that is, of the left-hand member of (48.1), is an

invariant. To this end we consider the discriminant of the

equation
F(x,

(50.8)

yr

z)

- r(x2 + y2 + z2 +

= 0,

1)

namely,
(50.9)

When the coordinates are subjected to a

equation (50.8) is
(50.10)

',

where F'(x', y',

nant of (50.10)

As remarked

transformation (30.11),

transformed into

/,

z'} is

z'}

- r(x' 2 + y' 2 + z' 2 +

1)

the transform of F(x, y,

= 0,

z).

The

discrimi-

is

at the close of

49, if the discriminant of the

equa-

tion (50.8)
equal to zero, so also is the discriminant of equation (50.10). Equating these discriminants to zero, we have
is

two equations of the fourth degree in r. Since these two equations have the same roots, the equations can differ at most by
270

Classification of the

Sec. 51]

Quadrics

a constant factor. However, in each equation the coefficient of

r 4 is + 1
consequently corresponding coefficients are equal,
and thus A = A'.
When a transformation (30.1) is applied to F(x, y, 2), and
we denote the resulting expression by F'(x', /, z'), we have
;

*"(*', y', *')

= a*' 2 + by' 2 + cz' 2 + 2 hx'y' + 2fx'z' + 2 gy'z'

+ 2(ax + hy +fz + /)*' + 2(hx + by + gz + m)y'
+ 2(fx + gyQ + cz + n)z' + F(x y z
,

The

discriminant of F'(x'
a
h

fx Q

gy Q

If we subtract from the last row the

second by yo, and the third by ZQ,

F(XO, yo, 20)

terminant

is

+ hy + fz Q + 1
-f byo -f gz Q + m
fx Q + gy + CZQ + n

ax
hx

by Q

).

y', z'} is

hx

hy Q

multiplied by XQ, the

have, on considering
expressed in the form (49.2), that the above deequal to

a
h
I

axo

hx

Ixo

first

we

+ hyo + fzo +
+ byo + gzo + m
I

+ myo + nzo + d

which is readily seen to be equal to A. In view of these results

and the fact that a general transformation is equivalent to a
translation and a rotation, we have
[50.3]

The discriminant A of an equation of the second degree in

three unknowns is an invariant under any change of rectangular coordinate axes.

51. Classification of the

Quadrics

We are now in position to analyze any equation of the second

degree and determine the character of the surface defined by it
and the position of the surface relative to the given coordinate
axes.

271

The Quadric

Surfaces

[Chap. 5

In 48 it was shown that given any such equation a coordinate system can be found in terms of which the equation is
of the form
(51.1)

a'x' 2

+ b'y' 2 + c'z' + 2 /'*' + 2 m'y' + 2 riz' + d = 0.

2

In this coordinate system the characteristic equation (48.8)

a' -

is

b'-r

(51.2)

=0,

c'-r
of which the roots are

V,

a',

Thus when an equation

c'.

of a

in the general form (48.1), and one finds the roots of

quadric
the characteristic equation (48.8), one has obtained the numis

bers which are the coefficients of the second-degree terms in an

equation (51.1) into which the given equation may be trans-

formed by a suitable transformation of coordinates.

In accordance with the theory of algebraic equations
follows from (48.9) that

D = n7 2 r 3

(51.3)

where n,

72, ra

it

are the roots of the characteristic equation.

We

and D = 0.
Case 1. D ^ 0. By Theorem [49.1] the surface is an ellipsoid
or hyperboloid when A ^ 0, and a cone when A = 0. If we
consider the two cases

D^

a translation of axes to the center or vertex as new

it follows from equations (49.1) that l = m = n = Q in
the new system. Hence under such a translation equation (51.1)
is transformed into an equation of the form

effect

origin,

ax 2

For

this equation

A=

Hence when A

A=

abed

(49.4)

= rir2 r3

rf.

an equation of the quadric

rue

(51.5)

(51.6)

we have from

(51.4)

and when

+ by + cz + d = 0.

+ r2y2 + r3 z 2 + -A- = 0,
TlWs

an equation
r lX 2

is

+ r 2y2 + rz z2 = 0,
272

is

Classification or the Quadrics

Sec. 51]

where

each case

in

r\ 9 r 2 , r 3

are the roots of the characteristic

equation. Hence we have

When

for an equation of the second degree in x, y, and z

and A ^ 0, the quadric is an ellipsoid, real or imaginary, when all the roots of the characteristic equation
have the same sign, and a hyperboloid when the roots do not
have the same sign; when D ^
and A = 0, the quadric is

[51.1]

D 7*

a cone or a point ellipsoid (imaginary cone) according as

the roots have different signs or all have the same sign.
If all the roots are equal, the

a real or imaginary

is

quadric

and a point sphere when A = 0. If two

and only two of the roots are equal and D ^ 0, the quadric is
sphere

when A ^

0,

a surface of revolution.
Case 2. D = 0. One at least of

th'e roots of the characteristic

equal to zero, as follows from (51.3). For the equation (51.1) this means that either a', V, or c' is equal to zero.
and that a' and V are not zero. By a
Suppose that c' =
translation of the axes, as in
47, equation (51.1) can be
transformed into an equation of the form

equation

is

ax 2

(51.7)

+ by + 2nz = 0,
2

or

ax 2

(51.8)

For equation

(51.7)

A=-

(51.9)

and

for (51.8)

quadric
(51.10)

+ by 2 + d = Q.
abn 2

= - rir2

A = 0. Hence when A ^

an equation of the

is

n* 2

+ r y2 + 2

= 0,

the quadric is an elliptic or hyperbolic paraboloid according as the nonzero roots of the characteristic equation have
the same or opposite signs. When A = 0, we have (51.8) and
that

is,

an equation of the quadric

(51.11)

is

n* 2 +

r 2y

273

+ d = 0.

The Quadric

By Theorem

[49.3] the quadric

according as equations (49.1)

common

solution.

an

It is

imaginary planes when

Surfaces
a cylinder or

is

and

elliptic

[Chap. 5
is

degenerate

(49.3) have not or have a

cylinder or two conjugate

the nonzero roots of the characteristic

equation have the same sign; it is a hyperbolic cylinder or

two real intersecting planes when the nonzero roots have opposite signs.

When two

of the roots of (51.2) are equal to zero, we may in

b = c = 0. As shown in 47, the #'-, /-, z take
generality
coordinates can be chosen so that (51.1) is of the form
f

all

riX

(51.12)

+ 2my = 0,

or
r 1 x*

(51.13)

In both cases

+ d = 0.

A = 0, and by Theorem

[49.3] the quadric is

parabolic cylinder or two parallel or coincident planes according as equations (49.1) and (49.3) have not or have a common
solution.

Gathering these results together, we have

[51.2]

and
for an equation of the second degree D
the quadric is a paraboloid, which is elliptic or
hyperbolic according as the two nonzero roots of the char-

When

7^ 0,

acteristic equation have the

D=

and

A=

and

same

or different signs;

when

the characteristic equation has two

roots, the quadric is an elliptic or hyperbolic cylinder

or consists of two intersecting planes, real or imaginary;

nonzero

when

D=

and

A=

and

there is only one nonzero root

of the characteristic equation, the quadric is

a parabolic

cylinder or consists of parallel, or coincident, planes.

The

preceding results are set forth in the following table,

which a canonical form of an equation of a quadric is given

in terms of the roots of the characteristic equation. When in a
particular problem one has found the values of the roots, one
is able to determine completely the form of the quadric, but in
in

order to determine

its

position relative to the given coordinate

274

Classification of the

Sec. 51]

axes

it is

Quadrics

necessary to find the principal planes by the method

48:

of

i*

+ r2y 2 -f rz z 2 H

717273

Ellipsoid or hyperboloid

Cone
rix

>

=O

+ r2y 2 + 2~

f\x

+ r2 y 2 + d =
0,

or d

= 0,

7-=

Paraboloid
Elliptic or hyper-

bolic cylinder

4
[

A = 0\d *

d=

Two

planes

as (49.1), (49.3)

have not, or have, a common solution.

(X

EXAMPLE

1.

(i)

+ 2 my =
+d=

Parabolic cylinder

Two

planes

For the quadric

+ 2 y2 + 3 22 - 4 xy - 4 yz + 2 = 0,

the characteristic equation

is

- 6 r2 + 3 r + 10 = 0,
of which the roots are 2, 5,
1. Since A =
r3

20, it follows from

system with Respect to which

(51.5) that there is a coordinate

an equation of the quadric

2x

(ii)

from
of

is

+ 5 y2

z2

+2=

this equation it follows that the quadric is a hyperboloid

sheets. In order to find this new coordinate system, we

two

observe that it follows from (49.1) that (0, 0, 0) is the center

of the surface in the original coordinate system. Equations
(48.7)
f for
v

(i)

are

= A0,
2 u + (r - 2> + 2 w = 0,
2 v + (r - 3)w = 0.
/i

(1

r)u

o
2

2
1 are 2,
Solutions of these equations for the roots 2, 5,
1,
1, 2,
2; 2, 2, 1 respectively. The lines through the origin with these respective direction numbers are the x-, y-,
;

z-axes respectively of the coordinate system with respect to

275

The Quadric

Surfaces

CChap. 5

The three planes

(ii) is an equation of the hyperboloid.
through the origin and whose normals have as direction num2 2, 2, 1 in the original coordinate
2
bers 2,
1, 2,
1,
which

system are the principal planes of the hyperboloid.

EXAMPLE
2 x2

(i)

2.

For the quadric

+ 2 y2 -4 z2 -5 xy-2xz-2 yz-2x-2y+z=Q

the characteristic equation

is

and D = 0, A = 729/16.
which the roots are 9/2, - 9/2,
From (51.10) we have that an equation of the surface is

of

3* 2 -3;y 2

(ii)

so that the surface

9/2 and
1, 1, 4.

is

+ 2z = 0,

a hyperbolic paraboloid.

For the roots

and
1,
9/2 respective solutions of (48.7) are 1,
These are direction numbers in the original coordinate

system of the *-axis and jy-axis in the new system, and from
(48.5) we have that in the original system equations of the new
jyz-plane

and #z-plane are respectively

x

(iii)

-y = Q,

+ y + 4z = 0.

Since these two planes pass through the original origin, which
is a point of the surface, and the new origin lies on the surface,
as follows from

(ii),

the two origins coincide.

Since the

new

ry-plane is perpendicular to the planes (iii) and passes through

the original origin, an equation of the new ry-plane is

2x + 2y-z = 0.
EXAMPLE

3.

For the quadric

-2y + 3z 2 + 3xy + 7xz-yz + 3x-4y-z + d = Q

we have D = and A = whatever be d. In this case the
(i)

2x

characteristic equation

is

r ( T2

_3r_

ift

= o.

Since the two nonzero roots are different, the quadric is an

or hyperbolic cylinder or two intersecting planes in accordance with Theorem [51.2]. Since D = 0, we have by

elliptic

276

Classification of the

Sec. 51]

Theorem
sisting of

Quadrics

[41.5] that the portion of the above equation conterms of the second degree is factorable. Factoring

this expression,

we

consider the product

In order that this expression shall be the same as the left-hand

member
k

The

of equation

we must have

(i),

+ 2m = 3, 2A;-w = -4,
first

m = 2.

If

+w=-

1,

km =

d.

three of these equations are satisfied by k

then d
km
2 in equation (i), the latter

1,
is

an

equation of a degenerate quadric, namely, two intersecting

2, equation (i) is an equation of a hyperbolic
planes. If d ^
since
the
nonzero roots of the characteristic equation
cylinder,
differ in sign.

EXERCISES
Determine the form of each of the following quadrics and
tion to the coordinate axes

2.
3.
4.
5.

6.

10.
is

its rela-

+ y 2 + 4 z 2 - 2 xy + 4 xz - 4 yz + 4 x - 8 z + 7 = 0.
x + 2 y 2 - z 2 + 4 xy + 4 xz - 2 yz - 2 x - 4 z - 1 = 0.
x 2 -2y 2 + z 2 + 2xy + 2xz
x 2 + y 2 + z 2 + xy + xz + jy* = 0.
2 * 2 + 2 y + 3 z 2 + 2 xy-x + 6
x2
2

Discuss the quadric of which the surface

Vx

^/2~y

+ \Tz =

a part.
11.

For what value of d

is

an equation of a cone?
an
12. For what value of a is ax 2 + 6 ;y 2 + 7 z 3 4- 4 x* + 30 =
equation of a surface of revolution ? Determine the form and position
of the surface for this value of a.

277

The Quadric

Surfaces

[Chap. 5

Show

that when for a general equation of the second degree

the portion of the equation consisting of terms of the second
degree consists of two factors, distinct or not according as the characteristic equation has one zero root or two zero roots.
13.

D = 0,

14.

Show

15.

Show

that for the quadric (48.1) the quantity

an invariant for any rotation of the coordinate axes.
that

if

+ m2 + n2

is

three chords of a central quadric have the same

lie in a plane or the common mid-point

mid-point, either all the chords

is the center of the quadric.

Find the most general equation of a quadric cone with vertex

and having the *-, y-, and 2-axes for elements.

16.

at the origin

Show

17.

that

is an equation of an elliptic paraboloid, a hyperbolic paraboloid, or a

> 1, a < I, or a 1.
parabolic cylinder according as

Show

- y) 2 -f

+ c(y -

2
that a(x
d 2 is an equaz)
z)
b(x
or
as
ab + be + ca is
elliptic
hyperbolic cylinder according
positive or negative. Find equations of the line of points of symmetry
in each case. Discuss the case when ab + be -f ca = 0.

18.

tion of

19.

where

an

Find the locus of the centers of the quadrics

/,

m, and n are fixed numbers and / and g are parameters.

20. Three mutually perpendicular lines meet in a point P such that

two of them intersect the axes of x and y respectively, and the third
passes through a fixed point (0, 0, c). Find the locus of P.
21.

Consider equation (48.1) with

and y entering in the

after the manner of

coefficients

as a quadratic in z with x
and analyze the result

solve for z

41.

278

Appendix

to

Chapter

In Chapter 1 we made use of definitions and theorems of

Euclidean plane geometry to set up coordinate systems in which
a point is represented by an ordered pair of numbers, and
to prove that an equation of the first degree in two variables
represents a (straight) line, in the sense that the coordinates
of

any point of a given line are a solution of a particular equaand every solution of this equation gives the coordinates

tion

of a point of this line.

with which

The advantage

of this set-up

is

the ease

permits one to use algebra in the solution of

geometric problems. However, the reader may appropriately
raise the question: Do the methods of coordinate geometry
enable one to solve any problem in Euclidean plane geometry ?
This question will be answered in the affirmative if we show
that the processes which have been used imply a set of axioms
from which all the theorems of plane geometry can be derived
without making use (consciously or unconsciously) of any tacit
it

assumption in deriving these theorems.

Before listing and testing such a set of axioms we shall sketch
briefly the nature and development of the subject usually called
Euclidean plane geometry. This subject has been the object
of study from the eras of the intellectual glory of Egypt and
of Greece. In the fourth century before the Christian Era,
Euclid assembled in his Elements the results of the study of
plane geometry by his predecessors and added to these results.
topic of Greek geometry was the description and investigation of different figures and their properties in the plane
and in space. Some of the figures were encountered as shapes

The

by motion of bodies others

were produced by mechanical devices, and their properties
were suggested by and derived from experience with physical
phenomena.
It was the principal aim of Greek geometers to proceed with
logical rigor, and this implied for them the necessity of deducing
every geometric proposition from those previously established
of material objects or generated

279

Appendix

to

Chapter

without undue reference to physical phenomena. However,

every logical deduction must have a beginning somewhere.
Accordingly, out of the accumulated knowledge and experience
Euclid drew up a set of definitions and first propositions called
axioms or postulates, and from these all subsequent propositions
were derived by purely logical processes without further reference to physical intuition. (Euclid made a distinction between
axioms and postulates, which, however, will be ignored in our
discussion. We shall use the words interchangeably.) Euclid
does not discuss the origin of the postulates or the philosophical
means of deciding their validity. In his Elements he is concerned only with the possibility of deducing the known theorems of geometry from the postulates given, no matter what
the origin of the latter might be.
For about two thousand years Euclid's Elements were accepted in most part as the final word in plane geometry. However, one point in Euclid was challenged from the earliest times

his parallel axiom

"If a straight line falling on two
makes
the
interior
lines
angles on the same side of the
straight
line less than two right angles, the two straight lines if produced
this

was

indefinitely

two

meet on that

side

on which the angles are

less

than

right angles/'

Most

critics required that an axiom, since it is accepted without demonstration, should be sufficiently simple in its content
to be self-evident, and the parallel axiom did not appear to be
of that nature. For that reason many attempts were made to

divest the proposition of its character as an axiom

it to be a logical consequence of the other axioms.

by proving

The most notable attempt was made by Saccheri (1733),

who tried to prove the axiom by reductio ad absurdum. He
tentatively replaced the parallel axiom by hypotheses opposed
to it, with a view to deducing conclusions which would be
obviously contradictory. The attempt was unsuccessful, but
this type of argument prepared the way for geometries in
which the parallel axiom would be replaced by an axiom opposed to it. The first concrete proposal was made independently about 1825 by the Russian mathematician Lobachevsky
and the Hungarian Bolyai, namely, a system of geometry

280

Appendix

to

Chapter

equally valid with the Euclidean results when Euclid's parallel

postulate is replaced by the assumption that through a given
point A not on a given line / there pass at least two lines which

This non-Euclidean geometry was studied

/.
and proof of its logical self-consistency is implied in
the works of Cayley (1859). Consequently it is now known
that any attempt to prove the parallel postulate as a consequence of the other Euclidean axioms must necessarily be futile.
These discoveries aroused anew widespread interest in plane
geometry and its logical foundations, and new results led to a
thorough overhauling of the axiomatic method. It was found
that Euclid's set of first principles was in some respects incomplete and in other respects redundant. It was incomplete in so
far as the results obtained by Euclid involved unannounced
axioms because of the part played by intuition in obtaining
these results. On the other hand, some of Euclid's definitions
do not serve any useful mathematical purpose. For instance,
Euclid states in the form of definitions: (1) a point is that
which has no part
(3) a
(2) a line is a breadthless length
is
lies
with
its
These
line
a
line
which
straight
points.
evenly
definitions can only be meant to announce to the reader that

do not

intersect

in detail,

certain objects, as point, line, etc., will be studied henceforth.

The description of them has no mathematical value and, as a

matter of

fact, Euclid never refers to these definitions in subsequent definitions or theorems. All significant information
about these geometric objects is contained exclusively in the
axioms, as, for instance, in the axiom that any two points determine a straight line on which they lie. Consequently it seems
preferable to begin with the mere enumeration of certain mathematical objects by name and then to state axioms describing
properties of these objects. The objects given at the outset
are not defined explicitly, but the axioms state the relations
which the objects have to one another. For instance, if we say
that a point lies on a line, or that the line goes through the
point, then we have an intuitive situation in mind. However,
from the standpoint of mathematical rigor we must not refer
to this intuition explicitly, but must realize that this fundamental relation is to be introduced and given by the axioms

281

Appendix
announcing

it.

to

CKapter

Thus a modern

set

of axioms enumerates

undefined terms and enunciates axioms containing these terms.

(These undefined terms may designate either particular objects

or classes of objects or relations.)
At the close of the last century the German mathematician
Hilbert proposed a set of axioms, and since then has revised
Many other mathematicians have likewise proposed
sets of axioms. We give below a set of fifteen axioms for plane

them.

geometry as published by Hilbert in 1930 (Grundlagen der

Geometrie, seventh edition), with some modifications for the
sake of clarity.
The undefined terms in the axioms are six in number point,
line, on (a relation between a point and a line), between (a relation between a point and a pair of points), congruent (a relation
:

between pairs of points), and congruent

(a relation

between

angles).

As explanation but not as definition: The word "line" is

here used to mean a straight line not terminated but extended
both directions. A point B is said to be between the points A and C if it lies on the line AC and between
A and C in the order of points on that line. A pair of points
A, B is said to be congruent to a pair of points A', B' if the
straight-line distance from A to B is equal to the straight-line
distance from A' to B'. Two angles are said to be congruent
indefinitely in

they are equal in measure, for example, in degrees, or if

they are superposable one on the other. (As remarked, the
statements in this paragraph are not definitions, but are explanations for the benefit of the reader who may be more familiar
with these ideas under somewhat different names; from a
logical point of view they not only can but should be omitted.)
Although congruence of angles is taken as an undefined term,
the term "angle" is not itself undefined, but a definition is
given for it below.
if

Following usual geometric terminology, in order to express

that a point A is on (or lies on) a line / we may also say that
We shall also use the
/ passes through A or that / contains A.
AB"
to
"the
line
the
designate
unique line which, acphrase
cording to

Axiom

1,

passes through

282

A and

B.

Appendix

to

Chapter

We now proceed to the statement of the fifteen axioms, interthem with definitions as necessary. The reader should
draw a figure for each axiom as a means of clarifying its meaning.
spersing

Axiom 1. There is one and only one

any two given (distinct) points.

Axiom 2. Every
any line there is at

line contains at least

least

line passing

two

one point not on

points,

through

and given

it.

Axiom 3. If a point B lies between the points A and C, then

A, B, and C all lie on the same line, and B lies between C and A,
and C does not lie between B and A, and A does not lie between
B and C.

Axiom

4.

Given any two

Axiom

5.

If

A and C, there
between A and C,
and D.

(distinct) points

can always be found a point B which

and a point D such that C lies between

lies

A, B, C are (distinct) points on the same

lies between the other two.

line,

one of the three points

DEFINITION. The segment (or closed interval) AC consists of the

points A and C and of all points which lie between A and C. A point
B is said to be on the segment AC if it lies between A and C, or is

or C.

Two

DEFINITION.

lines,

a line and a segment, or two segments,

if there is a point which is on both

are said to intersect each other

of them.

DEFINITION. The triangle ABC consists of the three segments AB,

BC, and CA (called the sides of the triangle), provided the points A, B,
and C (called the vertices of the triangle) are not on the same line.

Axiom 6. A line which intersects one side of a triangle and

does not pass through any of the vertices must also intersect
one other side of the triangle.
Axiom 7. If A and B are (distinct) points and A' is a point
on a line /, there exist two and only two (distinct) points
B' and B" on / such that the pair of points A', B' is congruent
to the pair A, B and the pair of points A', B" is congruent to
the pair A, B moreover A lies between B' and B".
1

283

Appendix

to

Chapter

8. Two pairs of points congruent to the

are
congruent to each other.
points

Axiom

If B lies between A and C, and B' lies between

and A, B is congruent to A', B', and B, C is congruent
then A, C is congruent to A', C.

Axiom
A' and
B', C',

same pair of

9.

C',

DEFINITION. Two segments are congruent if their end points are

congruent pairs of points.
DEFINITION. The ray AC consists of all points B which lie between
A and C, the point C itself, and all points D such that C lies between
A and D. (In consequence of preceding axioms it is readily proved
that

if

C'

is

any point on the ray AC the rays AC' and

is said to be /row the point A.

AC are identical.)

The ray AC

DEFINITION. The angle BAC consists of the point A (the vertex of

the angle) and the two rays AB and AC (the sides of the angle).
DEFINITION. If ABC is a triangle, the three angles BAC, ACB
t

are called the angles of the triangle. Moreover the angle BAC
said to be included between the sides A B and AC of the triangle

CBA
is

(and similarly for the other two angles of the triangle).

Axiom 10. If BAC is an angle whose sides do not lie in the

same line, and B' and A' are (distinct) points, there exist two
and only two (distinct) rays, A'C and A'C" from A' such that
the angle B'A'C is congruent to the angle B AC, and the angle
B'A'C" is congruent to the angle BAC] moreover if E is any
point on the ray A'C and E" is any point on the ray A'C",
the segment E'E" intersects the line A'B'.
1

Axiom

11.

Every angle

is

congruent to

itself.

and the included angle of one triangle

are congruent respectively to two sides and the included angle
Axiom

12.

If

two

sides

of another triangle, then the remaining angles of the first triangle are congruent each to the corresponding angle of the

second triangle.

Axiom 13. Through a given point A not on a given line

there passes at most one line which does not intersect /.

Axiom 14. If A, B, C, D are (distinct) points, there exist on

A n such
the ray AB a finite set of (distinct) points A\, A^
A n _i, A n
that (1) each of the pairs A, A\ A\ A 2 A 2 A 3
is congruent to the pair C, D and (2) B lies between A and A n
-

284

Appendix

to

Chapter

The

points of a line form a system of points such

points can be added to the space and assigned to
the line without causing the line to violate one of the first

Axiom

that no

15.

new

eight axioms or

The

first five

Axiom

14.

axioms state the simplest properties of a

line

and of the order possessed by points on a line. They assure in

particular the existence of an endless number of points on a
line, that a line is not terminated at any point, and that the
order of points on a line is not like that of points on a closed
curve such as a

Axiom 6

also

circle.
is

concerned with order properties, but involves

points not all on one line, and thus gives information about the
plane as a whole in a way in which the previous axioms do not.
The congruence Axioms 7-12 are introduced so as to avoid
in the proof of a proposition the use of superposition, that

is,

picking up a geometric figure and placing it upon another, as

is done in the customary treatment of elementary geometry.
Euclid himself used superposition, but even to Greek mathematicians and philosophers the use of this process in proving
a theorem was open to question. Now mathematicians meet
the question by means of congruence axioms. This explains,
in particular, why Axioms 10 and 12 appear here as axioms

and not as propositions to be proved.

Axiom 13
ously stated.

is

the equivalent of Euclid's parallel axiom previIt should be noted that Axioms 1-12 (because

they lead to the essential properties of perpendicular lines)

enable us to prove the existence of at least one line which passes
through A and does not intersect /. Axiom 13 is required to
assure us that there are not two such lines.
Axiom 14 is known as the Axiom of Archimedes. It corresponds to the process of using a measuring stick to find the
distance from one point on a line to another, and insures that,
starting at one point and laying off equal distances (the length
of the stick) in succession towards the other point, the other
point will ultimately be passed.

Axiom 15
shall

an axiom of completeness, assuring that there

be on any line all the points necessary to constitute a
285
is

Appendix

to

Chapter

continuum, which means that the points of any line may be

brought into one-to-one correspondence with the set of all real
numbers. In this sense Axiom 15 is equivalent to an axiom of
continuity in the field of real numbers. For a discussion of
this question the reader is referred to Part First of Fine's
159, where the concept
College Algebra and in particular to
of a Dedekind cut is explained (but without calling it such).
The question of continuity is fundamental in the calculus, and
there the reader will find it fully discussed. Axiom 15 might
be omitted from a set of axioms for elementary geometry since
the usual elementary theorems follow without it, but it is nec-

numbers in coordinate geometry,

and especially for application of the calculus to geometry.
Granted these fifteen axioms, all further propositions of
Euclidean plane geometry can be derived from them by a
essary for the free use of real

rigorous process of inference without further appeal to intuition. To carry this out in detail is, of course, a long story.

The

reader may consult in this connection O. Veblen's The

Foundations of Geometry, Monographs on Topics of Modern
Mathematics, pp. 3-51, Longmans, Green & Co., 1911, and
H. G. Forder's Foundations of Euclidean Geometry, Cambridge
University Press, 1927 in the latter many theorems of plane
;

geometry are traced back individually to their axiomatic source,

but there is much additional material.
Returning now to the question with which we introduced
this

Appendix, consider the following composite proposition

(A), the six clauses of which (I- VI) correspond to the six undefined terms used in Axioms 1-15* :

Proposition A. It

is

in the plane such that

possible to construct a coordinate system

:

I. Every point has associated with it a unique pair of real

numbers (its coordinates) and every pair of real numbers is

associated with a unique point.

II. Every line has associated with it an equation ax+by+c=0
in which x and y are unknowns (variables), a, &, c are constant
*

See in this connection O. Veblen's The Modern Approach to Elementary

The Rice Institute Pamphlet, Vol. 21, 1934, pp. 209-222.

Geometry,

286

Appendix
coefficients,

and a and

to

b are not

Chapter

both

0.

This equation is unique

and every such

to within a possible constant multiplier (?* 0),

equation is associated with a unique line.
III.

A point

an equation
IV.

The

(*2,

less

than

2 ), if

is

on a

line, if

and only

point

(x,

and only

such that the following equations both hold

1,

(1

- 0*i + tx 2

y=(l- t)yi + ty2

V.

The
,

pair of points (*i,

jXs),

(* 2

(2)

VI.

The

angle (xj,

(*4, y*)> if

(x 2 ,y2 ) is congruent to the

>>i),

and only

if

- *!) 2 + 0*
angle (x 2 , y 2 )(xi, y\) (#3,

y 2 ')(xi',

(x 2

and
and

y) lies between the points (xi, y\)

there is a number /, greater than

if

,jx

pair (* 3

coordinates satisfy

if its

of the line.

yi')(*3,

JVs'), if

jy3 )

congruent to the

is

and only

if

(y*

V(* 2

- *i') 2 + CV2 - y
X

Proposition A,

remarked, this

necessitates

Axiom 15

since, as previously

axiom guarantees the existence upon any

line

of points in one-to-one correspondence with all real numbers. As a matter of fact, points corresponding to a suitable

subset of real numbers satisfy Axioms 1-14, and such a subset

suffice to give algebraic expression to the fourteen

would

However, we desire to deal with all real numbers and

thus have included Axiom 15.
The truth of Proposition A follows as a consequence of
familiar theorems of elementary Euclidean geometry (see Theo-

axioms.

and equations (5.8)).

conversely that Axioms 1-14 follow as
consequences of Proposition A entirely by the methods of algebra and without use of other propositions of geometry. We
call the left-hand member of equation (2) the square of the
rems

[5.2], [5.3], [3.1], [3.4]

We

shall

now show

287

Appendix

to

CKapter

length of the line segment (xi, y\) (x 2 y 2 ), and the left-hand member of equation (3) the cosine of the angle (x 2 yz)(xi, yi)(x 3 ;y 3 ).
These names are applied to algebraic expressions involving
Xi, yi, and so on, not to geometric objects.
,

Axioms 1 and 2 follow from the algebraic definitions of point

and line in A and the results of 1.
For any real value of t > I equations (1) determine a point.
On solving these equations for x 2 and y2 we obtain
,

it is seen that (x 2
y 2 ) lies between (x, y) and
accordance with the algebraic definition A, IV of
betweenness. Similarly, it can be shown that if / < 0, (x\ y\)
lies between (x, y) and (x 2 ^2). Thus Axioms 3, 4, and 5 fol-

from which

yi)> in

(*i

low from the algebraic definition A, IV.

Theorem [6.8] gives an algebraic definition of parallelism
satisfying Axiom 13.
definition of direction

From

nates of any two points of the

and the algebraic

of a line in terms of the coordi-

this definition

numbers

there follows the result that

line,

numbers of

parallel lines are proportional.

In the consideration of Axiom 6 we take on the side

direction

PiP 2 of
the triangle with vertices PI(*I, >>i), P 2 (x 2 y*\ P 3 (*3, y*) a
point P(x, y) other than PI and P 2 Its coordinates are given
< / < 1. If the line of the axiom is parallel to
by (1) for
P2 P3 it meets the side PiP 3 in the point P' of coordinates
,

t)y\ + ty 3 since direction numbers of the

0*i + txa, (1
- * 2 ), t(y3 - y 2 ), which are direction
PP'
are
segment
t(x&
numbers of P 2 P 3 Similar results hold for a line parallel to PiPa.
Suppose then that the line through P is not parallel to PiP3 or
P 2 P 3 and denote by P' and P" the points in which it meets the
(1

line

lines

PiP 3 and P 2 P 3

respectively.

The

coordinates of

are of

the form
(1

r)xi

+ rxs,

(1

r)yi

+ ry3

Since the line does not pass through PI or P 3 r is not equal to

or 1. If r lies between and 1, the conditions of the axiom are
,

288

Appendix

to

Chapter

or > 1, in which case P' lies outmet. Suppose then that r <
side the segment PiP 3 and write the coordinates of P" thus

Since the points P, P', P"

u such that (see page 25)
(1

- s)* 2 + s*3 =

(1

on a

lie

)[(!

0*i

number

there exists a

line,

+ to] + ul(l - r)xi + r*

P

3 ],

and similarly for the /s. Since PI, 2

3 are not collinear,
there can be no linear homogeneous relation of their coordinates,
,

and consequently we must have

(l~w)(l-/) + w(l~r)=0, (1 _ W )f - (1 - 5 ) = 0, ur-s = 0.

The first of these equations is satisfied if the other two are.
Eliminating u from the second and third of these equations, we
obtain

< <

Since

<

<

1,

that

and
is,

P"

<
is

or > 1, by hypothesis, we have that

on the segment P 2 P 3 and the axiom
;

follows.

Axiom 8 is satisfied by the algebraic definition A, V of congruence of pairs of points. If (jti, y\) and (x 2 yz) are the points
A and B of Axiom 7, (#3, jya) the point A' on the line /, and
an equation of the line /, the dey*) =
3) + b(y
a(x
,

B" is the algebraic problem of

solutions of the equations

termination of the points B' and

finding the

common

- *0 2 + (y 2 - ;vi) 2 =(x- *3 2 +
= 0.
* 3 + 6(y
a(x
JVs)

(*2

(y

- ^) 2

It is evident that these equations have two and only two soluand one can show that (# 3 yz) is the mid-point between
tions
,

the points B' and B" (see Theorem [4.1]).

In the consideration of Axiom 9, we denote
f

(xi

yi'}> fa'* y*')

Then B and

by

the coordinates of A, C, A',

B' are points

(x,

y) and

= (1 - 0*i + tx 2
x'=(l- /')*i' + t'xj,
x

for suitable values of

and

C respectively.

where

y = (l- t)yi + ty2,

y'=d- t'W + t'y 2

(x' 9 y'),

(#i,;yi), (^2,^2)*

/'

such that

289

<

<

1,

'
9

<

<

1.

to

Appendix

When now

in accordance with

Chapter

A,V we

express the congruence

and A'B', and of BC and B'C', we find that /' = /, and

then that AC and A'C are congruent and the axiom follows.
If for Axiom 10 the angle BAG is (* 2 y*)(xi, y\)(x*, jVa) and

of

AB

y 2 ), the left-hand member

of equation (3) is a fixed number k which can be shown to be
such that k < 1, and we have to determine solutions # 3 y* of
the points A', B' are (x\

y\), (x 2

',

this equation.

we put

If

then
x2

(4)

'

=x

and

for any positive value of d2 these equations give the coordinates of a point on the ray A'B' and for negative values of
d2 on the ray through A' in the opposite direction. If, in like
,

manner, we put

and require

#',

we obtain

y' to satisfy equation (3),

k = uu 2 + M2.
If this

equation

is

written in the form

vv 2 ==

(5)

and then squared, the

uu 2

resulting equation

- u 2 k) 2 =
u 2 2 + v 2 2 = 1,

is

reducible to

(u

we have from

U2k

w2

V2l,

(5)
f2k

1.

and

(6)

T W2/.

Consequently there are two and only two solutions of the prob-

and E" on the respective rays

from A' forming with A'B' an angle congruent to the given

lem.

The

coordinates of points

angle are given by

290

+ u2

l),

to

Appendix

1*3

and d. Since

for positive values of ds

\,

Chapter

d3

,
.

*4

+ "4
_,

03

= *i +
,

03

2t

by A, IV that the segment E'E" has a point

with the ray A 'B' when k is positive, and with the
opposite of the ray when k is negative. Thus Axiom 10 follows from A, V and A, IV as Axiom 11 does directly from A,V.
We denote by /2 2 the left-hand member of equation (2),

it

follows from (4)

in

common

which we have called the square of the length of the line segment (xi, 3>i) (#2, ^2), and similarly we denote by /i 2 and I 2 the
squares of the lengths of the segments (x\, yi)(xs, jys) and
also we denote by cos A the left(*2, ^2) (#3, JV3) respectively
;

hand member

of equation (3),

which we have called the cosine

of the angle (#2, ^2X^1, y\)(xs, ^3).

substituted in the equation
I

(7)
it is

2
/i

/2

When

these expressions are

- 2 /i/a cos A,

found that this equation is an identity, that is, the Law

is thus an algebraic identity.
If in Axiom 12 the

of Cosines

vertices of the
f

(xi

two

triangles are (x\,y\), (^2,^2), (*s,

'
,

yi'), (x 2

(*i, yi)(x2,

7
t

j>2')

y 2 ) and
^3) and

(*3

(xi' 9

jVs)

and

y*'), the expressions (2) for the lengths

yi')(x2, y/) are equal, and likewise for

(xi, jyiOfe', yz). The expressions (3) for

cos A' of the angles at (x\, y\) and (x\, y\) are
equal. Then from equation (7) we have the equality of the expressions for the lengths of the third sides. Having the expres(xi, yi)(xz,

cos

A and

sions for the lengths of corresponding sides of the triangles

equal, we obtain from equations of the form (7) the equality of

the expressions for the cosines of corresponding angles and by

A, VI the congruence of these angles.
If in Axiom 14 the coordinates of A and B are xi, y\ and
x 2 y2 respectively, and the coordinates of A\ are given by (1),
then t is equal to /i//2, where l\ and 12 are the lengths of the
segments CD and AB respectively. If then we replace / in (1)
by 2 1, 3 1, and so on, equations (1) give the coordinates of the
,

291

Appendix

to

Chapter

and so on, of the axiom. Consequently this axiom

a sufficiently large integer n as multiplier of
taking
by

points A2, ^3,

is satisfied
t

in equations (1).

Since

it

has

now been shown

that Axioms 1-15 follow as

entirely by the methods of algebra

consequences of Proposition
and without use of other propositions of geometry, the situation
is

as follows

Clauses I-VI of Proposition A provide algebraic correspondents for each of the six undefined geometric terms. Given any

geometric term, write out its definition from the six undefined
terms, and in the definition replace each undefined term by its
the result will be the algebraic coralgebraic correspondent
;

respondent of the given geometric term. Hence every geometric

term has its algebraic correspondent or representative. This
enables one to translate every geometric theorem into a corresponding theorem of algebra. And in view of the algebraic
proofs provided for the theorems of algebra corresponding to
Axioms 1-15, the geometric proof of a theorem can be translated into an algebraic proof of the corresponding theorem of
algebra (although, as indeed often happens, a shorter algebraic
proof may be found in another way).
Thus the question which we raised concerning the adequacy
of the methods of coordinate geometry in the study of any
question of Euclidean plane geometry is answered in the
affirmative.

292

Index

Abscissa, 9

equation of, 54, 145

52; imaginary, 55;
inscribed, 52; point, 55; tangent to, 57

Circle, 54

Absolute value, 12, 19

Agnesi, witch of, 234
Angle, between directed line segments, 15, 80; between nor-

Circles,

mals to a plane, 94; between

positive direction of lines, 29,
35, 46, 86, of parallel lines,

32

bisectors of, 51, 128; dihedral,

97, bisectors of,

with a plane, 98
232

97

Archimedes, axiom
ral of, 156

line

makes

trisection of,

of,

285;

spi-

radical

orthogonal, 61;

axis of, 59

system

of,

59

Cissoid, 159

Cofactor.

See Determinant

Collinear points, 20, 26, 82, 87

Completing the square, 55

Conchoid, 159
Cone, 129, 201, 245-248; asymptotic, 252
general equation of,
vertex of, 246, 265
265, 273
;

Asymptote. See Hyperbola

Asymptotic cone, 252
Axes, oblique, 53 rectangular, in
the plane, 8, 53, in space, 71
choice of, 171, 229
;

Conies, 171 ; as orbits of planets,

189; as plane sections, of a
cone, 201, of a quadric, 261;

222;

183,

of,

196; degenerdiameter of
ate, 208, 219, 222
directrix of, 191,
central, 192

of Archimedes, 285

confocal,

Axioms of

Hilbert, 282
Axis, polar, 154 of cylinder, 246
;

of surface of revolution,
240; of symmetry, 174

center

183;

central,

Axiom

escribed,

215, see also Parabola, Ellipse,

Hyperbola

178,

171, 202

eccentricity

equation

of,

of, 172,

204,

208, polar, 191

equations of,
canonical, 227
parallel chords
;

BOLYAI, 280

of,

Canonical equations.

See Conies

177,

191, 220,

cipal axis of,

and Quadric surfaces

Canonical form of equation, 227
Cardioid, 159
Cassini, ovals of, 235

223

223;

prin-

similar, 173,

196, 263;
tangent to, 228;
with axes parallel to coordinate
axes, 203. See also Equations
of the second degree, Parabola,

CAYLEY, 281
See Conies and Quadric
surfaces
See
Characteristic
equation.

Center.

Equations of the second degree

Ellipse, Hyperbola
Conjugate imaginary factors, 212
Conjugate imaginary lines, 197
Conjugate imaginary points, 177
Coordinate axes. See Axes

293

Index
Coordinate planes, 71
Coordinates, in the plane, Cartesian,

8,

oblique,

53,

polar,

154, rectangular, 8, transforma-

tions of, 149, line, 64;

choice

Directed line segment, 12, 78

Direction, positive, of a line, 28,
84 ; of a segment, 12, 77
Direction cosines, of a line, 28, 29,
85 ; of a line segment, 12, 77 ;
of normals to a plane, 96

171, 229; in space, cylin167, polar, 167, rec-

of,

drical.

tangular,

71,

perpendicular
three lines, 162

166,

spherical,

transformations of, 160

Curve, in the plane, 21 in space,
239, plane, 239, skew, 239,
twisted, 239
Cycloid, 236
Cylinder, 239, 266, 267, 275 axis
of, 246
Cylindrical coordinates, 167
;

pendicular to two

Determinant, the, of equations,

41, 111, 137; cofactorof, 108,
131; element of, 106, 131;
main diagonal of, 104, 107, 130
minor of, 107, 131
Determinants, of the second order,
41; of the third order, 106110; of the fourth and higher
orders, 130-137 evaluation of,
;

properties

132-136,

of,

141;

137;

108-110, 116,
reduction of,

135; sum of, 46, 112

Diameter, of central conies, 192
of central quadrics, 246

of

Distance, between points, 11, 76,

153, 165
directed, 9 ; from a
;

a point, 35, 96 from a

plane to a point, 95 shortest,
between two lines, 121
line to

Division, internal and external,

of a line segment, 17, 26, 81, 95

Dual, 47
e,

15

Eccentricity, of a conic, 171

conjugate diameters of, 192; construction

director circle of, 235
of, 187

Ellipse, 171, 188;

directrices of,

eccentricity of,

83; of
a linear entity, 143 ; of a plane,
74 of space, 83 of spaces of
higher order, 142, 143
;

118; in n-space, 143

Directrix of a conic, 171

Diametral plane of central quadrics, 256

of a curve,
Dimensionality, 21

21,25; of a

106

of, 46, 113,

lines,

a line segment, 11, 77; of normals to a plane, 92 of perpendicular lines, 30, 31, 86; of
three lines through a point,

Discriminant, 265, 271

A, 265, 271

135; product

of

143

Direction numbers, of a line, 27,

31, 84, 105, 127; of a line per-

A 217, 226, 258, 268

in w-space,

of

86;

lines,

line, 21, 25,

294

189;
186, 189, 202

183,

184,

equation of, 182, 184, 185, 206,

227, parametric, 195, see also
Equations of the second degree focal radius of, 186, 193
;

foci of,

nary,

190

183, 184, 189;

197;

major

latus

imagi-

rectum of,
minor

axis of, 185

Index
point, 197 polars
with respect to, 195 ; principal
axes of, 183 properties of, 187,
193, 195, 196;
tangents to,
193 vertices of, 185
Ellipses, similar, 197
of revoEllipsoid, 244, 255, 275
axis of, 183

Equations of the first degree, in

two unknowns, 3, 39, degenerate, 4, 23, determined by solu-

104,
114,
solutions of, 99, 104, 110, 115,
126; in four unknowns, 137,

142
in n unknowns, 139-142
Equations of the second degree,
;

two unknowns, 208, 229,

characteristic equation of, 224,
in

Elliptic cylinder, 245, 266, 274,

275, 278

Embedded

226, 229, determination of the

locus of, 208-229, invariants of,

entity, 143
Endless number of solutions, 3
Equation of a line in the plane,
general, 23, 24, 31 ; in polar coordinates, 158 ; intercept form,

225, 226; in three unknowns, 254, 256, characteristic

equation of, 258-260, 262, 263,
270, determination of the locus

210,

26; parametric equations, 24,

25, 28
point-direction number
form, 27 point-slope form, 33

of,

See

Equation of a plane, 88-93 containing, a line, 90, 102, one line

and parallel to another, 120, a
point, 91, two lines, 97; determined, by a line and perpendicular to a plane, 122, by a point
;

and a line, 91, 102, by a point

and normal to a line, 122, by a
point and parallel to two lines,

by three

122,

space,

145

100;

line,

EUCLID'S definition of a plane, 74,

88
EUCLID'S Elements, 279
EUCLID'S parallel axiom, 280, 285
EULER'S formulas, 166
Focal radius, 181, 186, 193
Focus, 171

Graph,

8,

10

of an equation, 21

of parametric equations, 25

in

w-space,

number

Hilbert, axioms of, 282

Homogeneous

143,

87 pointform, 84;
;

first

degree

Homogeneous expressions,
Hyperbola, 171

212

two-point form, 82

198, 200,

200;

295

See

equations.

Equations of the

general, in

parametric, 85,

direction

see

of,

A and A

points, 91, 93,

111, 113, 118, 119; intercept

form, 94, 128 ; in 4-space, 145

Equations of a

256-261, invariants

/, /,

two-point form, 22, 117.

also pages 62, 63

5,

homogeneous,

98,

equivalent,

see

Spheroid; general
equations of, 265, 273
point,
273; properties of, 247; tangent line to and tangent plane
to, see Quadric surfaces

25,

pendent, 5, 40, solutions of, 4,

39-45 in three unknowns, 88,

lution,

5-7,

tions,

essentially different, 5, 40, inde-

219

25,

asymptotes

of,

conjugate, 199,

conjugate axis

of,

185,

Index
199 construction of, 188 diameters of, 192, conjugate, 192,
199; directrices of, 183, 184,
195 eccentricity of, 186, 199,
;

Line, characteristic property of,

27, 28, 82;

200, 207; equation of, 182, 184,

185, 195, 206, 212, 214;
radii of, 186, 193

184;

foci of, 183,

rectum

of,

with respect

to,

195;

183,

220;

latus

polars

focal

axis

principal

of,

190;

188,

193,

195, 196, 199, 200, 232;

rec-

properties of,
tangular,

190,

tangents

to,

187,

196,

193;

199;
transverse
198,

axis of, 185; vertices of, 185, 195

Hyperbolic cylinder,
274, 275, 278

245,

266,

Hyperboloid, asymptotic cone of,

252 general equation of, 265,
273; of one sheet, 241, 244,
of revolution, 241
255, 275

nate plane, 83
Line coordinates, 64
Line segment, 9, 11 ; antiparallel,
13
division of, 17
mid-point

to,

polar planes with respect

see Quadric surfaces ; prop-

249

tangent line to
and tangent plane to, see Quaderties of,

parallel,

13,

79;

per-

Linear dependence, 25, 116, 141,

145
Linear entity, 142, 145
Linear equation, 142
Linear expression, 25
Lines, in a plane, 118, 122, 128;
parallel, 27, 30, 33, 38, 84, 85
;

perpendicular, 30, 31, 34, 86;

skew, 86

through a point, 47three, through a

51, 85, 106;

18;

pendicular, 16

of two sheets, 241, 244, 255,

275

of,

in the plane, see

Equation of a line in the plane

in space, see Equations of a line
in space
parallel to a coordi-

point, 118

LOBACHEVSKY, 280
Locus, 21, 64, 171

Locus problems, 229-236

Magnification, 173

ric surfaces

Mid-point of a line segment, 18

Minor. See Determinant

Hyperplane, 145
225, 258, 268
Intercept, 26, 94, 128
Invariant, 153, 210, 267. See also

7,

I, J, Z>,

and

w-space. See Space of n dimensions

Non-Euclidean geometry, 281

Normal, to a curve, 181 toa plane,
92 to a quadric surface, 262
;

/, 225, 258, 268

Octants, 71

Latus rectum. See Parabola, Ellipse, Hyperbola

Law

of Cosines, 15

Lemniscate, 159
Limagon, 159

Ordered pair of numbers, 20

Ordinate, 9
of a
Orthogonal projection, 73
;

circle,

191

Ovals of Cassini, 235

296

Index
the origin, 20, 87;

Parabola, 171, 175; axis of, 175,

177; directrix of, 175; equa-

Polar, line called the, 179.' See also

tions of, 174, 175, 205, 208, 213,

Equations of the second

see also

focus of,

degree;

rectum
polars

175;

Parabola, Ellipse, Hyperbola

Polar coordinates, 154; transformation of, to rectangular

latus

175 normal to, 181

with respect to, 179;

of,

properties of,

178,

subnormal of, 181

181,

coordinates, 157
Polar plane, 263, 264

182;

tangents

Parabola, Ellipse, Hyperbola

Positive direction. See Direction

Parabolas, similarity of, 173

Parabolic cylinder, 245, 267
Paraboloid, elliptic, 243, 255, 275,
general equation of, 274 ; hyperbolic, 243, 255, 275, general
equation of, 274 of revolution,

Positive sense.

See Sense

Principal axis.

See Conies

Principal plane. See Quadric surfaces

upon a

polar plane with respect to, see Quadric surfaces;

178, 240;

of, 249, 267; tangent

250 tangent plane to,251

Parameter, 24, 231, 235

Parametric equations, 24, 85, 231.
See also Equation, Equations
Perpendicular line segments, 16
Perpendicular lines, 30, 31, 34, 86
Plane, 74, 88 ; normal to a, 92
parallel to a line, 89, 90;
;

through three points, 111, 113,

119. Seea/soEquationofaplane
Planes, parallel, 89, 93, 98, 101,
103, 105, 114 ; parallel to a line,

segment,
128,

upon

Quadrant, 9
Quadratic expression, in two unknowns, 219; in three unknowns, 219
Quadratic form, 219
Quadric surfaces, 246 canonical
center
equations of, 245, 275
central, 246 clasof, 246, 264
sification of, 271-275 confocal,
251, 263 chords of, 278, parallel, 257
degenerate, 246, 273275; diameter of, 246; dia-

127; perpendicular,
through a line, 101, 106
three, configurations of, 123126, with a line in common, 116,
with one point in common, 111,
113, 114, 116, 119; four, with
a point in common, 144
Points, collinear, 20, 26, 82, with
106,
;

line

Parallel line segments, 13

Parallel lines, 27, 30, 33, 38, 84, 85

line, 9, 73, 75,

a plane, 73, 128; of a point,

upon a line, 9, 73, upon a plane,
73 parallel, 73

properties

103,

of

Projection,

90, 92

See also

Pole, 154, 179, 263, 264.

to,

175, 177, 180-182; vertex of, 175

line to,

conjugate

imaginary, 177

297

metral plane

of,

symmetry

266

256; line of
normal to,
262, 263; of revolution, 241242, 270; plane of symmetry
of, 266; plane section of, 261,
263; polar plane with respect
;

264; principal planes

246, 254, 260, 270; ruled,

to, 263,
of,

of,

Index
248-250

tangent line

to,

257

tangent plane to, 251, 252, see

also Paraboloid ; with principal
planes parallel to the coordinate
planes, 252. See also Equations
of the second degree, Paraboloid, Ellipsoid,

Hyperboloid

Surface, 239

degree

of,

240 ; of

revolution, 178, 240-242, 270;

ruled, 249. See Quadric surfaces

Symmetric, with respect to, a line,

9, 73 ; a plane, 73 ; a point, 9, 73
Symmetry, line of, 266 ; plane of,
266; point of, 267. See also

Symmetric
Radical axis of

circles,

59

Radical plane of spheres, 129

Radius vector, 154
Ratios with zero terms, 23

Tangent, 57. See also Parabola,

Ellipse, Hyperbola, Paraboloid,
Quadric surfaces

Representative point, 22, 82

Rotation of axes, in the plane,
152 ; in space, 164

Tangent plane,

Ruled

surface,

249

Ruling, 249, 263

SACCHERI, 280
Sense, along a segment, 11, 12, 77
positive, 12, 77, along a line, 28,
84
Similar curves, 173
Similar figures, 173, 174
Simultaneous equation. See Equa;

curve, 239
lines,

86 ;

Paraboloid,

coordinates, in the plane, 149,

into polar coordinates, 157 ; in
space, 160, into cylindrical coordinates, 167, into spherical

coordinates, 167

Translation of axes, in the plane,

150; in space, 160

tions

Skew
Skew

see

Quadric surfaces ; to a sphere,

129
Tetrahedron, 144
Torus as a surface of revolution,
242
Transform, 210
Transformations of rectangular

common

perpen-

dicular to, 121, 123

Slope of a line, 33

Triangle, condition that points are

vertices of, 117; perpendiculars to sides of,

32

Solutions. See Equation, Equations

Trisection of an angle, 232

Space, of four dimensions, 142,

plane in, 145 ; of n dimensions,

Twisted cubic, 240, 242

Twisted curve, 239

142, generalized sphere in, 144,

hyperplane
145

in,

145, line in, 143,

Sphere, 128, 129, 273;

general-

ized, 144

Vector, 154
Vectorial angle, 154
Vertex, of a cone, see

Cone

of a

conic, see Conies, Parabola, El-

Spherical coordinates, 166

Spheroid, oblate, 241 ; prolate, 241
Spiral of Archimedes, 156

lipse,

Hyperbola

Witch of Agnesi, 234

PRINTED IN THE UNITED STATES OF AMERICA

NON-EUCLIDEAN GEOMETRY
by Roberto Bonola
This

is

Italian

an excellent histoiical and mathematical view by a renowned

geometer of the geometries that have arisen from a study of
on parallel

Euclid's 5th postulate

ematicians
field

will

that has

lines.

and math-

Students, teachers

here a ready reference source and guide to a

find

now become overwhelmingly

NON-EUCLIDEAN GEOMETRY

important.

examines

first

the

various

attempts

by the Greeks, by the Arabs,

prove Euclid's parallel postulate
by mathematicians of the Renaissance. Then, ranging through the

to

17th,

and

18th

founders
Legendre,

19th

centuries,

non-Euclidean

of

W.

Bolyai,

considers

it

geometry,

such

Schweikart,

Gauss,

of

discussion

the
as

forerunners

Saccheri,

Taurmus,

Bolyai
the

later

and

Lambert,

and

author

developments,
Lobatschewsky
work of Riemann, Helmholtz and Lie, the impossibility of
proving Euclid's postulate, and similar topics. The complete text of
In

treats the

two of the founding monographs is appended to Bonola's study

"The Science of Absolute Space by John Bolyai and "Geometrical
1

'

Researches on the Theory of Parallels" by Nicholas Lobatschewsky.

"Firmly

recommended

inclination"

"Classics on

to

any

scientific

reader with some mathematical

JOURNAL OF THE ROYAL NAVAL SCIENTIFIC SERVICE

the subject," SCIENTIFIC AMERICAN

Translation with additional appendices by H.

liographic

Index.

181

footnote

diagrams

references

431pp.

Introduction

5%

S.

by

Carslaw. 256 bibFederigo

Ennques.

8.

S27 Paperbound $1.95

THE GEOMETRY OF RENE DESCARTES

an unabridged re-publication of the definitive English transone of the very greatest classics of science. Originally
published in 1637, it has been characterized as "the greatest single

This

is

lation

of

step ever made in the progress of the exact sciences" (John Stuart
as a book which "remade geometry and made modern geometry

Mill

possible," (Eric Temple Bell). It "revolutionized the entire conception

of the object of mathematical science," (J. Hadamard).
With this volume Descartes founded modern analytical geometry.

Reducing geometry to algebra and analysis, and conversely showing

that analysis may be translated into geometry, it opened the way
for modern mathematics.
Descartes was the first to classify curves
systematically, and to demonstrate algebraic solution of geometric
curves.
His geometric interpretation of negative quantities led to
later concepts of continuity and the theory of function. The third

book contains important contributions

This

edition

to the theory of equations.

contains the entire definitive

Smith-Latham translation

of Descartes' three books: Problems the Construction of which Requires

Only Straight Lines and Circles; On the Nature of Curved Lines; On

the Construction of Solid or Supersolid Problems. Interleaved

by page with the translation is a complete facsimile of the

page
1637

French text, together with all Descartes' original illustrations, 248

footnotes explain the text and add further bibliography.
Translated by David E. Smith and Marcia L. Latham. Preface. Index.

50

figures.

xiii

+ 244pp. 5%

x 8.

S68

Paperbound $1.50

GEOMETRY OF FOUR DIMENSIONS

by

H. P.

Manning's
clear
in

Manning
GEOMETRY OF FOUR DIMENSIONS

and concise

introduction to a branch of

unique

is

English as a

in

modern mathematics now

application as an indispensable part of mathematical physics (algebra,

analysis, relativity).

Proceeding by the synthetic method the author

geometry of the fourth

in

some

is

make

to

dimension, aiding you

dimensional figures. Treatment

although

will

clear to you the

reason about four-

based mostly on Euclidean geometry,

cases, as hyperplanes at infinity, non-Euclidean geometry

is

used. After a discussion of the history of dimensions, with references

to

Moebius, Riemann, Lobatchevsky, and others, the author discusses the

foundations of fourth dimensional geometry; perpendicular and simple

angles; angles of two planes

and

of higher order; symmetry, order, motion,

hyperpyramids; hypercones; hyperspheres; Euclidean geometry, figures

with parallel elements;

measurement of volume and hypervolume

in

hyper-

space; regular polyhedroids.

"Clearly written

... an

particular interest

excellent book," SCIENTIFIC AMERICAN. "Of

The author shows that the knowledge of the new

geometries has clarified and sometimes corrected men's understanding

even of Euclidean geometry," MODERN SCHOOLMAN. "A standard
treatise for students of mathematics,"

SKY

AND

TELESCOPE.

Complete unabridged reproduction of the first edition. Preface. Historical

introduction. 179 footnotes, mostly bibliographical. Glossary of terms.
Index. 76 figures, including 3 full-page plates.

ix

SI

+ 348

pp.

5%

x 8.

82 Paperbound $1.95

AN INTRODUCTION TO

THE GEOMETRY

OF N DIMENSIONS
by

D.

M. Y. Sommerville

is the only book in English devoted exclusively to the subject of

higher-dimensional geometry. Unavailable for many years, it has now
been reprinted in response to the current increased interest among

This

geometers,

geometry of

mathematicians,
n dimensions.

To introduce

this

physicists

increasingly

fruitful

and crystallographers

branch

of

in

the

investigation,

Dr.

Sommerville selects and demonstrates several representative topics of

n-dimensional geometry which not only illustrate the extensions of
three-dimensional geometry, but reveal results which are unexpected
and where analogy would be a faithless guide. Each topic is de-

veloped at length. Both the metric and, to a

projective properties of n-dimensional
In

the

first

slightly less extent, the

geometry are dealt with.

four chapters, Dr. Sommerville explains the fundamental

ideas of incidence, parallelism, perpendicularity, and angles between

linear spaces. (Chapter
also contains an excursus into enumerative
I

geometry.)

Chapter V presents analytical geometry from the pro-

jective point of view,

and contains some of the

simplest ideas relating

algebraic varieties, and a more detailed account of quadratics,

especially with reference to their linear spaces. Chapter VI treats

to

analytic geometry of n dimensions from the metric point of view,

and

and

addition to the ordinary cartesian formulae, some account

applications of the Plucker-Grassmann co-ordinates of linear

contains,

in

space, and applications to line-geometry. The remaining four chapters deal with polytopes, and contain, especially in Chapter IX, some
of the elementary ideas in analysis situs. Chapter VIII treats of the

content of hyper-spacial figures. The

chapter establishes the

final

regular polytope.

Bibliography. Index. 60 diagrams. 196pp. 5

/s

x 8.

Paperbound ($1.50

tent.)

THE ELEMENTS OF NON-EUCLIDEAN GEOMETRY

by

D.

M. Y. Sommerville

Almost immediately upon publication Dr. Sommerville's book became

the standard text in the field. It is renowned for its lucid yet meticulous
exposition and, unlike advanced treatises,

it

can, for the most part,

be understood by anyone who has a good knowledge of high school

algebra and geometry. The arrangement follows the traditional pat-

and solid geometry where theorems are deduced from

axioms and postulates. In this way, the student can follow the development of non-Euclidean geometry in strictly logical order, from a
fundamental analysis of the concept of parallelism to such advanced
tern of plane

and transformations.

topics as inversion

Elementary
Euclidean

hyperbolic geometry; elliptic geometry; analytic nongeometry; representations of non-Euclidean geometry in

Euclidean space; space curvature and the philosophical implications

of non-Euclidean geometry; the theory of the radical axes, homothetic
centres,

and systems of circles; inversion, equations of transformation,

and the classification of conies are developed with

groups of motions;

exceptional clarity.

Although

this

is

primarily an elementary text, Dr. Sommerville treats

such important and

and

parallelism,

difficult topics

the

absolute

as the relation

measure,

the

between parataxy

pseudosphere,

proof of the defect-area theorem, geodesic representation,

Gauss'

and others

with simplicity and ease. 126 problems at chapter endings give the
student practise in using the forms and methods developed in the text

and provide many important

133

figures.

corollaries.

126 problems. Index,

xvi

274pp.

5%

x 8.

S460 Paperbound $1.50

FAMOUS PROBLEMS OF
ELEMENTARY GEOMETRY
by

Felix Klein

This

expanded version

of

the

famous

1894

Easter

lectures

at

Gottingen University has been accepted as a modern mathematical

classic, and has been translated into four different languages. Using
techniques of modern mathematics, it examines three famous problems which were intensely investigated in premodern mathematics:

doubling the volume of

cube, trisecting an angle, squaring

circle.

Written

with

all

profundity, this

Felix

Klein's

mathematical

volume provides answers

to

breadth,

clarity,

and

modern problems con-

nected with these three problems of the past.

esting to the modern student in answering

It

is

especially inter-

such

questions

as:

Under what circumstances is a geometric construction possible? By

what means can a geometric construction be effected? What are
transcendental numbers, and how can you prove that e and pi are
transcendental? Treatment is simple, and no knowledge of higher
mathematics

is

required.

CONTENTS.
THE POSSIBILITY OF THE CONSTRUCTION OF ALGEBRAIC EXPRESSIONS. 1. Algebraic equations solvable by square
I.

roots.

The delian problem and the

trisection of the angle.

The

divi-

sion of the circle into equal parts. The construction of the regular
polygon of 17 sides. General considerations of algebraic constructions.

II.

TRANSCENDENTAL NUMBERS AND THE QUADRATURE OF

THE CIRCLE. Cantor's demonstration

of the existence of transcen-

dental numbers. Historical survey of the attempts at the computation and construction of pi. Transcendence of the number e. Transpi. Integraph and the geometric construction of pi.
Notes by R. C. Archibald discuss in detail Gaussian polygons, Fermat's theorem, the irrationality of pi, and similar topics.

cendence of

Translated by W. W. Beman, D. E. Smith from the second revised

edition. 16 figures, xi -f 92pp. 5 3/a x 8.

T298

Paperbound

$1.00

GUIDE TO THE LITERATURE OF

MATHEMATICS AND PHYSICS
Including Related

by Nathan
WHAT

literature

and
this

Brier

for the

Works on Engineering Science

Parke

III

available and

is

WHERE

time available

first

can

it

be found?

Fully revised

a handy, inexpensive edition

in

GUIDE TO THE LITERATURE OF MATHEMATICS

PHYSICS puts a comprehensive library catalogue at your finger

unique, up-to-date

AND
tips.

Over 5,000

entries

(more than double the number

in

the

first

edition)

are included under approximately 120 subject headings, further subdivided by an average of 6 subheadings. Every branch of physics,

mathematics, and related engineering science is fully represented. For

example, 28 books appear under the subhead, Projective Geometry;
78 under Geometrical Optics; 22 under Cosmic Rays, not including
cross-references and general works.

books were selected by a practicing mathematician and physicist

on the basis of current significance, fullness of treatment, and the
ability to give clues to further material on a subject. Citations are as

All

full

as possible

number

giving author,

of volumes

number

title,

edition, publisher, place, data,

Many recently available

Russian works are included. Discussions of the literature under each
or

of

pages.

heading define the subject matter and provide numerous cross-references and suggestions for further investigation. The GUIDE also contains an extensive listing of bibliographical aids- abstracts, indexes,
periodicals, reviews, bibliographies, directories, encyclopedias, docu-

mentary reproductions, guides, and library resources. A complete

Author and Subject Index allows instantaneous location of the bibliographical data on any book.

GUIDE will not only save you countless hours and help you locate
much difficult-to-find material; it will often help you shape your problem and prevent a duplication of work by showing you exactly what
has been done on a subject. With the literature grown to such huge
proportions, the GUIDE TO THE LITERATURE OF MATHEMATICS AND
PHYSICS has become an indispensable research assistant for every
physicist, mathematician, engineer, scientist, student, and researcher.
2nd revised edition. Over 5,000 entries. 71 -page introduction. Indexes.

This

464 pp. 5 3/s x 8.

Paperbound

$2.49

Catalogue of Dover

SCIENCE BOOKS
DIFFERENTIAL EQUATIONS
(ORDINARY AND PARTIAL DIFFERENTIAL)
INTRODUCTION TO THE DIFFERENTIAL EQUATIONS OF PHYSICS, I. Hopf. Especially valuable
Emphasizes intuitive rather than
engineer with no math beyond elementary calculus
Law of causality, energy theorem, damped
Partial contents
formal aspects of concepts
oscillations, coupling by friction, cylindrical and spherical coordinates, heat source, etc
48 figures 160pp 53/8 x 8
S120 Paperbound $1 25
to

F. Bowman. Rigorous, provides all necessary material

Bessel functions of zero order, of any
applications
asymptotic expansion, circular membranes, Bessel's solution
useful not only to students of physics and
problem, much more "Clear
engineering, but to mathematical students in general," Nature 226 problems Short tables
27 figures x + 135pp. 5% x 8
S462 Paperbound $1.35
of Bessel functions

INTRODUCTION TO BESSEL FUNCTIONS,

during development,
real order,
to Kepler's

definite

includes

practical

integrals,

DIFFERENTIAL EQUATIONS,

F.
R. Moulton. Detailed, rigorous exposition of all non-elemenChapters on practical problems,
solving ordinary differential equations
Includes analytic differential
than problems usually given as illustrations
a
of
differential
of
variations
equations, analytic implicit
parameter, integrals
equations,
functions, problems of elliptic motion, sine-amplitude functions, deviation of formal bodies,
with
linear
differential
periodic coefficients, much more
equations
Cauchy-Lipshitz process,
S451 Paperbound $2.00
10 figures 222 problems xv -f 395pp 5% x 8.
Historical notes

tary

processes

of

more advanced

EQUATIONS OF MATHEMATICAL PHYSICS, A. G. Webster. Valuable

compression theory, potential theory, theory of sound, heat conducContents include deduction of differential equation,
Cauchy's method, boundary problems,
tions, vibrations, normal functions, Fourier's series
method of Riemann-Volterra, spherical, cylindrical, ellipsoidal harmonics, applications, etc
S263 Paperbound $2.00
97 figures vu + 440pp 5% x 8
PARTIAL

sections

DIFFERENTIAL

on elasticity,

wave propagation, vibration theory

E. L. Ince. A most compendious analysis in real and

Existence and nature of solutions, continuous transformation groups,
an infinite form, definite integrals, algebraic theory Sturmian theory, boundary
problems, existence theorems, 1st order, higher order, etc. "Deserves highest praise, a
notable addition to mathematical literature," Bulletin, Amer. Math Soc. Historical appendix.
18 figures vni + 558pp 53/a x 8.
S349 Paperbound $2.55

ORDINARY DIFFERENTIAL EQUATIONS,

complex

solutions

domains
in

A. Erde~lyi. Only modern work available m English,

unabridged
Discusses various procereproduction of monograph prepared for Office of Naval Research
dures for asymptotic evaluation of integrals containing a large parameter, solutions of
S318 Paperbound $1.35
ordinary linear differential equations, vi -I- 108pp. 5% x 8.

ASYMPTOTIC EXPANSIONS,

LECTURES ON CAUCHY'S PROBLEM, J. Hadamard. Based on lectures given at Columbia, Rome,

discusses work of Riemann, Kirchhoff, Volterra, and author's own research on hyperbolic
case in linear partial differential equations. Extends spherical cylindrical waves to apply
to all (normal) hyperbolic equations. Partial contents: Cauchy's problem, fundamental formula, equations with odd number, with even number of independent variables, method of
S105 Paperbound $1.75
descent. 32 figures in
x 8.
316pp.

5%

CATALOGUE OF
NUMBER THEORY
INTRODUCTION TO THE THEORY OF NUMBERS, L E. Dickson. Thorough, comprehensive, witn
adequate coverage of classical literature Not beyond beginners chapters on divisibility,
congruences, quadratic residues and reciprocity, Diophantme equations, etc Full treatment
of binary quadratic forms without usual restriction to integral coefficients
Covers infinitude
of primes, Fermat's theorem, Legendre's symbol, automorphs, Recent theorems of Thue,
Much material not readily available elsewhere 239 problems 1 figure
Siegal, much more
VIM + 183pp 53/8 x 8
S342 Paperbound $1.65

ELEMENTS OF NUMBER THEORY, I. M. Vmogradov. Detailed 1st course for persons without
advanced mathematics, 95% of this book can be understood by readers who have gone
no farther than high school algebra Partial contents divisibility theory, important number
theoretical functions, congruences, primitive roots and indices, etc
Solutions to problems,
exercises
Tables of primes, indices, etc
Covers almost every essential formula in elereads smoothly," Bull of the Amer. Math
mentary number theory' "Welcome addition
Soc 233 problems 104 exercises VIM + 227pp 5% x 8
S259 Paperbound $1.60

PROBABILITY THEORY AND INFORMATION THEORY

SELECTED PAPERS ON NOISE AND STOCHASTIC PROCESSES, edited by Prof Nelson Wax, U of
6 basic papers for those whose work involves noise characteristics
Illinois
Chandrasekhar,
Included is Kac's ChauvenetUhlenback and Ornstem, Uhlenbeck and Ming, Rice, Doob
"
Extensive bibliography lists 200 articles, through 1953
Prize winning "Random Walk
21
x
S262 Paperbound $2 35
9V4
6Vs
figures. 337pp.

A PHILOSOPHICAL ESSAY ON PROBABILITIES, Marquis de Laplace This famous essay explains

without recourse to mathematics the principle of probability, and the application of probTranslated
ability to games of chance, natural philosophy, astronomy, many other fields
from 6th French edition by F W Truscott, F L Emory Intro by E T Bell 204pp 5% x 8
S166 Paperbound $1 25

MATHEMATICAL FOUNDATIONS OF INFORMATION THEORY,

Khmchin

For

mathematicians,

statisticians, physicists, cyberneticists, communications engineers, a complete, exact introduction to relatively new field
Entropy as a measure of a finite scheme, applications to
coding theory, study of sources, channels and codes, detailed proofs of both Shannon
theorems for any ergodic source and any stationary channel with finite memory, much more
"Presents for the first time rigorous proofs of certain fundamental theorems
quite

amazing expository

complete

ability,"

American Math

Monthly

+ 120pp 5Vs x 8
S434 Paperbound $1 35

VH

VECTOR AND TENSOR ANALYSIS AND MATRIX THEORY

VECTOR AND TENSOR ANALYSIS,
important subject
Cartesian vectors,
solid,

analytical,

formulae

listed,

G.

E.

Start with
simple
Chnstoffel symbols,
differential

demonstrated

Hay.

One

definitions,

solenoidal

of

clearest introductions to
with sure mastery

finish

tensors
chapters

geometry Separate
195 problems 66 figures

Complete
on
vni

increasingly
oriented
of plane,

of

breakdown

fundamental
All
193pp 5% x 8
S109 Paperbound $1 75

application

APPLICATIONS OF TENSOR ANALYSIS, A J. McConnell. Excellent text

methods to such familiar subjects as dynamics, electricity, elasticity,

for

applying

tensor

hydrodynamics

Ex-

plains fundamental ideas and notation of tensor theory, geometrical treatment of tensor
"The
algebra, theory of differentiation of tensors, and a wealth of practical material
variety of fields treated and the presence of extremely numerous examples make this
volume worth much more than its low price," Alluminio Formerly titled "Applications of the
xn 4- 381pp
Absolute Differential Calculus." 43 illustrations. 685 problems

S373

VECTOR AND TENSOR ANALYSIS,

Paperbound

$1.85

Wills. Covers entire field, from dyads to non-Euclidean

manifolds (especially detailed), absolute differentiation, the Riemann-Chnstoffel and RICCIEmstem tensors, calculation of Gaussian curvature of a surface Illustrations from electrical
A.

P.

Presupposes only working

relativity theory, astro-physics, quantum mechanics
knowledge of calculus. Intended for physicists, engineers, mathematicians 44 diagrams
S454 Paperbound $1.75
114 problems, xxxii + 285pp. 53/a x 8

engineering,

DOVER SCIENCE BOOKS

PHYSICS, ENGINEERING
MECHANICS, DYNAMICS, THERMODYNAMICS, ELASTICITY
MATHEMATICAL ANALYSIS OF ELECTRICAL AND OPTICAL WAVE-MOTION, H. Bateman. By one
of century's most distinguished mathematical physicists, a practical introduction to developments of Maxwell's electromagnetic theory which directly concern the solution of partial
differential equation of wave motion. Methods of solving wave-equation,
polar-cylindrical
coordinates, diffraction, transformation of coordinates, homogeneous solutions, electromagnetic fields with moving singularities, etc. 168pp 5 3/s x 8.
S14 Paperbound $1.60
Enrico Fermi. Unabridged reproduction of 1937 edition
Remarkable for
requires no knowledge of advanced math beyond calculus, only familiarwith fundamentals of thermometry, calonmetry. Partial Contents
Thermodynamic systems, 1st and 2nd laws, potentials; Entropy, phase rule, Reversible electric cells, Gaseous
reactions- Van't Hoff reaction box, principle of LeChatelier,
Thermodynamics of dilute
solutions- osmotic, vapor pressures, boiling, freezing point, Entropy constant 25 problems
x 8
S361 Paperbound $1.75
24 illustrations, x + 160pp

THERMODYNAMICS,

clarity, organization,
ity

5%

FOUNDATIONS OF POTENTIAL THEORY,

0. D. Kellogg. Based on courses given at Harvard,

suitable for both advanced and beginning mathematicians, Proofs rigorous, much material
here not generally available elsewhere Partial contents gravity, fields of force, divergence
theorem, properties of Newtonian potentials at points of free space, potentials as solutions
of LaPlace's equation, harmonic functions, electrostatics, electric images, logarithmic pox 8.
S144 Paperbound $1.98
tential, etc. ix -I- 384pp.

5%

DIALOGUES CONCERNING TWO NEW SCIENCES,

mechanics, engineering, as" enjoyable as

it

is

Galileo Galilei. Classic of experimental scierice,

Characterized by author as "superior

important

to everything else of mine

Offers a lively exposition of dynamics, elasticity, sound, ballistics,
strength of materials, scientific method Translated by H Grew, A de Salvio 126 diagrams
xxi -t- 288pp. 53/8 x 8.
S99 Paperbound $1.65

THEORETICAL MECHANICS; AN INTRODUCTION TO MATHEMATICAL PHYSICS, J

Murnaghan. A mathematically rigorous development for advanced students,

S.

Ames,

with

F.

D.

constant

Used in hundreds of advanced courses Unusually thorough coverage

practical applications
of gyroscopic baryscopic material, detailed analyses of Corilis acceleration, applications of
Lagrange's equations, motion of double pendulum, Hamilton-Jacobi partial differential equa159 problems 44 figures
tions, group velocity, dispersion, etc. Special relativity included
ix
S461 Paperbound $2.00
462pp. b3/8 x 8.

STATICS AND THE DYNAMICS OF A PARTICLE, W. D. MacMillan This is Part One of "Theoret"
Mechanics
ical
For over 3 decades a self-contained, extremely comprehensive advanced
undergraduate text in mathematical physics, physics, astronomy, deeper foundations of
engineering Early sections require only a knowledge of geometry, later, a working knowledge
of calculus
Hundreds of basic problems including projectiles to moon, harmonic motion,
340
ballistics, transmission of power, stress and strain, elasticity, astronomical problems
xvn + 430pp
5% x 8
practice problems, many fully worked out examples 200 figures
S467 Paperbound $2 00

THE THEORY OF THE POTENTIAL, W.

"

D. MacMillan. This is Part Two of "Theoretical MechanComprehensive, well-balanced presentation, serving both as introduction and reference
Assumes no prior
mathematicians
to specific problems, for physicists and
knowledge of integral relations, all math is developed as needed Includes Attraction of
Finite Bodies, Newtonian Potential Function, Vector Fields, Green and Gauss Theorems,
Two-la/er Surfaces, Spherical Harmonics, etc. "The great number of particular cases
should make the book valuable to geo-physicists and others actively engaged in practical
applications of the potential theory," Review of Scientific Instruments xn -4- 469pp 5% x 8
S486 Paperbound $2.25
ics

with

regard

DYNAMICS OF A SYSTEM OF RIGID BODIES (Advanced Section), E. J. Routh. Revised 6th ediPartial contents
tion of a classic reference aid
moving axes, relative motion, oscillations
about equilibrium, motion Motion of a body under no forces, any forces Nature of motion
given by linear equations and conditions of stability Free, forced vibrations, constants of
integration, calculus of finite differences, variations, procession and mutation, motion of
the moon, motion of string, chain, membranes 64 figures 498pp 5% x 8
S229 Paperbound $2.35
THE DYNAMICS OF PARTICLES AND OF RIGID, ELASTIC, AND FLUID BODIES: BEING LECTURES

ON MATHEMATICAL PHYSICS,

A. G. Webster. Reissuing of classic fills need for comprehensive

work on dynamics. Covers wide range in unusually great depth, applying ordinary, partial
differential equations. Partial contents
laws of motion, methods applicable to systems of
all sorts, oscillation, resonance, cyclic systems, dynamics of rigid bodies, potential theory,
stress and strain, gyrostatics, wave, vortex motion, kinematics of a point, Lagrange's equations, Hamilton's principle, vectors, deformable bodies, much more not easily found together in one volume. Unabridged reprinting of 2nd edition. 20 pages on differential
S522 Paperbound $2.35
equations, higher analysis. 203 illustrations xi + 588pp 5% x 8.

CATALOGUE OF
PRINCIPLES OF MECHANICS, Heinrich Hertz. A classic of great interest in logic of science.
Last work by great 19th century physicist, created new system of mechanics based upon
returns to axiomatic analysis,
understanding of formal, structural
space, time, mass,
Of great
aspects of science, taking into account logic, observation, a priori elements
Milne
20 page introduction by R. S
to
historical importance
Pomcar6, Carnap, Einstein,
analyzes implications of Hertz's thought and logic of science 13 page
Cohen, Wesleyan U
x 8
S316 Clothbound $3.50
introduction by Helmholtz. xln -I- 274pp
,

5%

S317 Paperbound $1.75

FOUNDATIONS

MATHEMATICAL

OF

STATISTICAL MECHANICS, A. I. Khinchin. A thoroughly

up-to-date introduction, offering a precise and mathematically rigorous formulation of the
Provides analytical tools to replace many commonly
problems of statistical mechanics
used cumbersome concepts and devices Partial contents- Geometry, kinematics of phase
space, ergodic problem, theory of probability, central limit theorem, ideal monatomic gas,
"Excellent
foundation of thermodynamics, dispersion, distribution of sum functions, etc
introduction
clear, concise, rigorous," Quarterly of Applied Mathematics. VIM + 179pp
S146 Clothbound $2.95
53/8 x 8
S147 Paperbound $1.35
.

MECHANICS OF THE GYROSCOPE, THE DYNAMICS OF ROTATION,

R. F. Deimel, Prof, of MeElementary, general treatment of dynamics of

Engineering, Stevens Inst of Tech
No knowledge of vectors
of
with
gyroscopic phenomena
rotation,
special
application
needed Velocity of a moving curve, acceleration to a point, general equations of motion,
motion
the
free
of
discs,
damped gyro, 103 similar topics Exergyro,
gyroscopic horizon,
x 8
S66 Paperbound $1.65
cises 75 figures 208pp.

chanical

5%

MECHANICS VIA THE CALCULUS,

P. W. Morris, W. S. Legge. Wide coverage, from linear motion

vector analysis, equations determining motion, linear methods, compounding of simple
harmonic motions, Newton's laws of motion, Hooke's law, the simple pendulum, motion of
a particle m 1 plane, centers of gravity, virtual work, friction, kinetic energy of rotating
bodies, equilibrium of strings, hydrostatics, sheering stresses, elasticity, etc
Many workedS207 Clothbound $3.95
out examples 550 problems. 3rd revised edition xn + 367pp

to

A TREATISE ON THE MATHEMATICAL THEORY OF ELASTICITY, A. E. H. Love. An indispensable

reference work for engineers, mathematicians, physicists, the most complete, authoritative
treatment of classical elasticity in one volume Proceeds from elementary notions of extension to types of strain, cubical dilatation, general theory of strains Covers relation between
mathematical theory of elasticity and technical mechanics, equilibrium of isotropic elastic
nature of force transmission, Volterra's theory of
solids and aelotropic solid bodies,
dislocations, theory of elastic spheres in relation to tidal, rotational, gravitational effects
on earth, general theory of bending, deformation of curved plates, buckling effects, much
more "The standard treatise on elasticity," American Math Monthly 4th revised edition
S174 Paperbound $2.95
76 figures, xvm
643pp 6Vs x 91/4.

NUCLEAR PHYSICS, QUANTUM THEORY, RELATIVITY

MESON PHYSICS, R. E. Marshak. Presents basic theory, and results of experiments with emmesons as virtual transitions
involving
phasis on theoretical significance. Phenomena
avoided, eliminating some of least satisfactory predictions of meson theory. Includes production study of TT mesons at nonrelativistic nucleon energies contracts between ir and u
mesons, phenomena associated with nuclear interaction of ir mesons, etc Presents early
evidence for new classes of particles, indicates theoretical difficulties created by discovery
x 8.
S500 Paperbound $1.95
of heavy mesons and hyperons. viu + 378pp.

5%

THE FUNDAMENTAL PRINCIPLES OF QUANTUM MECHANICS, WITH ELEMENTARY APPLICATIONS,

E. C. Kemble. Inductive presentation, for graduate student, specialists in other branches of
physics. Apparatus necessary beyond differential equations and advanced calculus developed
needed Though general exposition of principles, hundreds of individual problems fully
treated. "Excellent book ... of great value to every student
rigorous and detailed
has succeeded in keeping his presentation clear and undermathematical discussion
standable," Dr. Linus Pauling, J. of American Chemical Society. Appendices: calculus of
T472 Paperbound $2.95
variations, math, notes, etc. 611pp. 5Vs x 8%.

as

WAVE PROPAGATION

IN PERIODIC STRUCTURES, L. Brillouin. General method, application to

different problems: pure physics
scattering of X-rays in crystals, thermal vibration in
electrical engineering. Partial
metals; problems
crystal lattices, electronic motion
contents: elastic waves along 1-dimensional lattices of point masses. Propagation of waves
Energy flow. 2, 3 dimensional lattices. Mathieu's equation.
along 1-dimensional lattices
Matrices and propagation of waves along an electric line. Continuous electric lines. 131

illustrations,

xn

253pp.

5%

8.

S34 Paperbound $1.85

DOVER SCIENCE BOOKS

THEORY OF ELECTRONS AND

ITS

APPLICATION TO THE PHENOMENA OF LIGHT AND RADIANT

HEAT, H. Lorentz. Lectures delivered at Columbia Univ

by Nobel laureate Unabridged, form
historical coverage of theory of free electrons, motion, absorption of heat, Zeeman effect,
109 pages notes explain more advanced secoptical phenomena in moving bodies, etc
tions 9 figures 352pp. 53/8 x 8.
S173 Paperbound $1.85
,

SELECTED PAPERS ON QUANTUM ELECTRODYNAMICS, edited by

papers which established quantum electrodynamics, beginning

J.
Schwmger. Facsimiles of
to present position as part
of larger theory First book publication in any language of collected papers of Bethe, Bloch,
Dirac, Dyson, Fermi, Feynman, Heisenberg, Kusch, Lamb, Oppenheimer, Pauli, Schwmger,
Tomonoga, Weisskopf, Wigner, etc 34 papers 29 in English, 1 in French, 3 in German,
1
in
Italian
Historical commentary by editor, xvn
423pp 6Vs x 9V4

S444 Paperbound $2.45

FOUNDATIONS OF NUCLEAR PHYSICS,

edited by R. T. Beyer. 13 of the most important papers

on nuclear physics reproduced in facsimile in the original languages, the papers most often
cited in footnotes, bibliographies Anderson, Curie, Johot, Chadwick, Fermi, Lawrence, CockHahn, Yukawa. Unparalleled bibliography. 122 double columned pages, over 4,000
roft,
S19 Paperbound $1.75
6Va x 9V4.
articles, books, classified. 57 figures. 288pp

THE THEORY OF GROUPS AND QUANTUM MECHANICS, H. Weyl. Schroedmger's wave equation,
de Broglie's waves of a particle, Jordon-Hoelder theorem, Lie's continuous groups of transformations, Pauli
Unitary geometry,

exclusion

principle,

quantization

of

Mawell-Dirac

field

equations,

etc.

quantum theory, groups, application of groups to quantum mechanics,

symmetry permutation group, algebra of symmetric transformations, etc. 2nd revised ediS268 Clothbound $4.50
tion xxn + 422pp 53/8 x 8.
S269 Paperbound $1.95

PHYSICAL PRINCIPLES OF THE QUANTUM THEORY, Werner Heisenberg. Nobel laureate discusses quantum theory, his own work, Compton, Schroedinger, Wilson, Einstein, many
others. For physicists, chemists, not specialists in quantum theory Only elementary formulae
considered in text, mathematical appendix for specialists. Profound without sacrificing
18 figures 192pp 53/8 x 8
Eckart, F. Hoyt
clarity. Translated by C
S113

Paperbound

$1.25

INVESTIGATIONS ON THE THEORY OF THE BROWNIAN MOVEMENT, Albert Einstein. Reprints

from rare European journals, translated into English. 5 basic papers, including Elementary
Theory of the Brownian Movement, written at request of Lorentz to provide a simple
D Cowper Annotated, edited by R
Furth
Translated by A
33pp. of notes
explanation
124pp 5% x 8.
elucidate, give history of previous investigations. 62 footnotes
S304 Paperbound $1.25

THE PRINCIPLE OF RELATIVITY,

E. Einstein, H. Lorentz, M. Minkowski, H. Weyl. The 11 basic

papers that founded the general and special theories of relativity, translated into English.
Michelson
on
the
Lorentz
phenomena Mmkowski's
experiment, electromagnetic
by
papers
"
7 epoch-making papers by Em"Space and Time," and Weyl's "Gravitation and Electricity
stem- "Electromagnetics of Moving Bodies," "Influence of Gravitation in Propagation of
7 diagrams
Special
Light," "Cosmological Considerations," "General Theory," 3 others
S93 Paperbound $1.75
notes by A. Sommerfeld 224pp 5% x 8

STATISTICS
IN MEDICINE AND THE BIOLOGICAL SCIENCES,
E. Croxton. Based primarily on biological sciences, but can be used by anyone desiring
introduction to statistics. Assumes no prior acquaintance, requires only modest knowledge
of math. All basic formulas carefully explained, illustrated, all necessary reference tables
included. From basic terms and concepts, proceeds to frequency distribution, linear, nonContains concrete examples from medicine, biology. 101
linear, multiple correlation, etc
x 8.
S506 Paperbound $1.95
charts. 57 tables. 14 appendices. Iv -f 376pp.

ELEMENTARY STATISTICS, WITH APPLICATIONS

F.

5%

ANALYSIS AND DESIGN OF EXPERIMENTS,

H. B. Mann. Offers method for grasping analysis of

Partial
contents
Chi-square distribution, analysis of
variance, variance design quickly
variance distribution, matrices, quadratic forms, likelihood ration tests, test of linear
hypotheses, power of analysis, Galois fields, non-orthogonal data, interblock estimates, etc.
S180 Paperbound $1.45
15pp. pf useful tables, x
195pp. 5 x 73/8

FREQUENCY CURVES AND CORRELATION, W.

Elderton. 4th revised edition of standard

statistics. Practical, one of few books constantly referred to for clear
Partial
contents:
material.
basic
presentation
Frequency Distributions, Pearsons Frequency Curves, Theoretical Distributions, Standard Errors; Correlation Ratio Contingency,
Corrections for Moments, Beta, Gamma Functions, etc. Key to terms, symbols. 25 examples.
Clothbound $1.49
40 tables. 16 figures, xi
272pp. 5Vz x 8Vz

work on classical
of

P.

CATALOGUE OF
HYDRODYNAMICS, ETC.
HYDRODYNAMICS, Horace Lamb. Standard reference work on dynamics

of liquids and gases.

theorems, equations, methods, solutions, background for classical hydrodynamics. Chapters: Equations of Motion, Integration of Equations in Special Gases, Vortex
Motion, Tidal Waves, Rotating Masses of Liquids, etc. Excellently planned, arranged, Clear,
lucid presentation. 6th enlarged, revised edition. Over 900 footnotes, mostly bibliographS256 Paperbound $2.95
ical. 119 figures, xv
738pp. 6Vs x 91/4.

Fundamental

HYDRODYNAMICS, A STUDY OF LOGIC, FACT, AND SIMILITUDE,

Garrett Birkhoff. A stimulating

pure mathematics to an applied problem. Emphasis is on correlation of
deduction with experiment. Examines recently discovered paradoxes, theory of
modelling and dimensional analysis, paradox and error in flows and free boundary theory.
Classical theory of virtual mass derived from homogenous spaces; group theory applied
x 8.
to fluid mechanics. 20 figures, 3 plates, xui
186pp.
S22 Paperbound $1.85
application
theory and

of

5%

H. Dryden, F. Murhaghan, H. Bateman. Published by National Research

1932. Complete coverage of classical hydrodynamics, encyclopedic in quality.
contents: physics of fluids, motion, turbulent flow, compressible fluids, motion in
'stance of motion through viscous fluid
3 dimensions, laminar mo* ^n
ddy
2,
1,
"
ri
23
~c_ t;'
yjstacles, etc. Over 2900-item bibi <,&.**,. .
ifge of g*eS303 Paperbound "$2.75
figures.' 634pp. 5 3/s x 8.

HYDRODYNAMICS,

Council,
Partial

ACOUSTICS AND OPTICS

PRINCIPLES OF PHYSICAL OPTICS, Ernst Mach. Classical examination of propagation of light,
etc.
Historical,
color,
philosophical treatment unequalled for breadth and
polarization,
readability. Contents- Rectilinear propagation, reflection, refraction, dioptrics, composition
of light, periodicity, theory of interference, polarization, mathematical
representation of
x 8.
S170 Paperbound $1.75
properties, etc- 279 illustrations. 10 portraits. 324pp.

5%

THE THEORY OF SOUND, Lord Rayleigh. Written by Nobel laureate, classical methods here
will cover most vibrating systems likely to be encountered in practice. Complete coverage
of

mathematical aspects. Partial contents. Harmonic motions, lateral vibracurved plates or shells, applications of Laplace's functions to acoustical
friction, etc. First low-priced edition of this great reference-study work.
x 8
introduction by R. B. Lindsay. 1040pp. 97 figures.
S292, S293, Two volume set, paperbound $4.00

experimental,

tions

of

problems,
Historical

bars,

fluid

5%

THEORY OF VIBRATIONS, N. W. McLachlan. Based on exceptionally successful graduate

Discusses linear systems having 1 degree of freedom, forced
course, Brown University
vibrations of simple linear systems, vibration of flexible strings, transverse vibrations of
bars and tubes, of circular plate, sound waves of finite amplitude, etc. 99 diagrams. 160pp.
S190 Paperbound $1.35
x 8.

5%

APPLIED OPTICS AND OPTICAL DESIGN, A. E. Conrady. Thorough systematic

presentation of
"
Stresses practical
physical and mathematical aspects, limited mostly to "real optics
problem of maximum aberration permissible without affecting performance Ordinary ray
tracing methods; complete theory ray tracing methods, primary aberrations; enough higher
aberration to design telescopes, low powered microscopes, photographic equipment. Covers
fundamental equations, extra-axial image points, transverse chromatic aberration, angular
magnification, similar topics. Tables of functions of N. Over 150 diagrams, x + 518pp.
S366 Paperbound $2.98
5% x 8%.
RAYLEIGH'S PRINCIPLE AND ITS APPLICATIONS TO ENGINEERING, 6. Temple, W. Bickley.
Rayieigh's principle developed to provide upper, lower estimates of true value of fundamental period of vibrating system, or condition of stability of elastic system Examples,
Energy method of discussing vibrations, stability. Perrigorous proofs. Partial contents
turbation theory, whirling of uniform shafts. Proof, accuracy, successive approximations,
of
Rayieigh's theory. Numerical, graphical methods. Ritz's method 22 figures,
applications
ix

156pp.

5%

x 8.

S307 Paperbound $1.50

OPTICKS, Sir Isaac Newton. In its discussion of light, reflection, color, refraction, theories
of wave and corpuscular theories of light, this work is packed with scores of insights and
discoveries. In its precise and practical discussions of construction of optical apparatus,
contemporary understanding of phenomena, it is truly fascinating to modern scientists.
Foreword by Albert Einstein. Preface by I. B. Cohen, Harvard. 7 pages of portraits, facsimile
S205 Paperbound $2.00
pages, letters, etc. cxvi + 414pp. 5% x 8.

DOVER SCIENCE BOOKS

ON THE SENSATIONS OF TONE, Hermann Helmholtz. Using acoustical physics, physiology,
relation of music
experiment, history of music, covers entire gamut of musical tone
science to acoustics, physical vs. physiological acoustics, vibration, resonance, tonality,
progression of parts, etc. 33 appendixes on various aspects of sound, physics, acoustics,
music, etc. Translated by A J Ellis. New introduction by H. Margenau, Yale. 68 figures 43
musical passages analyzed. Over 100 tables, xix + 576pp
6Vs x 91/4.
S114 Clothbound $4.95

ELECTROMAGNETICS, ENGINEERING, TECHNOLOGY

INTRODUCTION TO RELAXATION METHODS, F. S. Shaw. Describes almost all manipulative resources of value in solution of differential equations. Treatment is mathematical rather
than physical. Extends general computational process to include almost all branches of
applied math and physics. Approximate numerical methods are demonstrated, although high
accuracy is obtainable without undue expenditure of time. 48pp. of tables for computing
irregular star first and second derivatives, irregular star coefficients for second order
"Useful.
exposition is clear, simple ... no
equations, for fourth order equations
previous acquaintance with numerical methods is assumed," Science Progress. 253 diaS244 Paperbound $2.45
grams. 72 tables. 400pp. 5% x 8.
.

THE ELECTROMAGNETIC FIELD, M. Mason, W.. Weaver. Used constantly by graduate engineers
Vector methods exclusively, detailed treatment of electrostatics, expansion methods, with
tables
tical

5%

converting any quantity

Discrete charges,

units
x 8.

absolute electromagnetic,
Maxwell
ponderable bodies

into

absolute electrostatic, pracequations, etc

416pp
S185 Paperbound $2.00

field

ELASTICITY, PLASTICITY AND STRUCTURE OF MATTER, R. Houwink. Standard treatise on

Theological aspects of different technically important solids crystals, resins, textiles, rubber,
Covers genclay, etc
Investigates general laws for deformations, determines divergences
eral physical and mathematical aspects of plasticity, elasticity, viscosity
Detailed examination of deformations, internal structure of matter m relation to elastic, plastic behaviour,
formation of solid matter from a fluid, etc Treats glass, asphalt, balata, proteins, baker's
2nd revised, enlarged edition
Extensive revised bibliography m over 500
dough, others
footnotes 214 figures, xvn
S385 Paperbound $2.45
368pp 6 x 9V4

DESIGN AND USE OF INSTRUMENTS AND ACCURATE MECHANISM, T. N. Whitehead. For the
instrument designer, engineer, how to combine necessary mathematical abstractions with
observations of actual facts
Partial contents
instruments and their parts,
of errors, systematic errors, probability, short period errors, erratic errors, design
human
precision, kinematic, semikmematic design, stiffness, planning of an instrument,
x 8
S270 Paperbound $1.95
xn
288pp
factor, etc. 85 photos, diagrams

independent
theory

5%

APPLIED HYDRO- AND AEROMECHANICS, L. Prandtl, 0. G Tietjens. Presents, for most part,
methods valuable to engineers Flow m pipes, boundary layers, airfoil theory, entry conditions, turbulent flow, boundary layer, determining drag from pressure and velocity, etc
An
"Will be welcomed by all students of aerodynamics," Nature
Unabridged, unaltered
Engineering Society Monograph, 1934. Index 226 figures 28 photographic plates illustrating
S375 Paperbound $1.85
flow patterns, xvi + 311pp 53/8 x 8

FUNDAMENTALS OF HYDRO- AND AEROMECHANICS,

L. Prandtl, 0. G. Tietjens. Standard work,

Prandtl's lectures at Goettmgen
Wherever possible hydrodynamics theory is
Presentapractical considerations in hydraulics, unifying theory and experience
tion extremely clear Though primarily physical, proofs are rigorous and use vector analysis
An Engineering Society Monograph, 1934 "Still recommended as an
to a great extent
186 figures, xvi
excellent introduction to this area," Physikalische Blatter
270pp
S374 Paperbound $1.85
53/8 x 8.

based

on

referred

to

GASEOUS CONDUCTORS: THEORY AND ENGINEERING APPLICATIONS,

Cobme. Indispensable
reference, to gaseous conduction phenomena, with engineering viewpoint prevailing
throughout Studies kinetic theory of gases, lomzation, emission phenomena, gas breakdown,
spark characteristics, glow, discharges, engineering applications m circuit interrupters, rectiDetailed treatment of high pressure arcs (Suits), low pressure arcs (Langmuir,
fiers, etc
Tonks). Much more. "Well organized, clear, straightforward," Tonks, Review of Scientific
Instruments 83 practice problems Over 600 figures 58 tables xx + 606pp
S442 Paperbound $2.75
53/fe x 8.
J.

D.

text,

PHOTOELASTICITY: PRINCIPLES AND METHODS, H. T. Jessop, F. C. Harris. For engineer, speproblems of stress analysis. Latest time-saving methods of checking calculations in

cific

2-dimensional design problems, new techniques for stresses in 3 dimensions, lucid descripof optical systems used in practical photoelectricity
Useful suggestions, hints based
on on-the-job experience included
Partial contents- strain, stress-strain relations, circular
disc under thrust along diameter, rectangular block with square hold under vertical thrust,
simply supported rectangular beam under central concentrated load, etc Theory held to
164 illustrations, vin + 184pp.
minimum, no advanced mathematical training needed
S137 Clothbound $3.75
6Ve x 9i/4.
tion

CATALOGUE OF
MICROWAVE TRANSMISSION DESIGN DATA, T. Moreno. Originally classified, now rewritten,
enlarged (14 new chapters) under auspices of Sperry Corp Of immediate value or reference
use to radio engineers, systems designers, applied physicists, etc Ordinary transmission
line theory, attenuation; parameters of coaxial lines, flexible cables, tuneable wave guide
impedance transformers, effects of temperature, humidity, much more "Packed with informatheoretical discussions are directly related to practical questions," U
tion
of Royal
Naval Scientific Service. Tables of dielectrics, flexible cable, etc. ix
248pp. 5% x 8.
S549 Paperbound $1.50
.

THE THEORY OF THE PROPERTIES OF METALS AND ALLOYS, H. F. Mott, H. Jones. Quantum
methods develop mathematical models showing interrelationship of fundamental chemical
phenomena wtih crystal structure, electrical, optical properties, etc. Examines electron
in applied field, cohesion,
heat capacity, refraction, noble metais, transition and
dt-valent metals, etc. "Exposition is as clear
mathematical treatment as simple and
reliable as we have become used to expect of ... Prof. Mott," Nature 138 figures, xin
x 8.
S456 Paperbound $1.85
320pp.

motion

5%

THE MEASUREMENT OF POWER SPECTRA FROM THE POINT OF VIEW OF COMMUNICATIONS

ENGINEERING, R. "B. Blackman, J. W. Tukey. Pathfmdmg work reprinted from "Bell System
Various ways of gettmg practically useful answers in power spectra
Technical Journal
measurement, using results from both transmission and statistical estimation theory. Treats.
Autocovanance, Functions and Power Spectra, Distortion, Heterodyne Filtering, Smoothing,
Decimation Procedures, Transversal Filtering, much more
Appendix reviews fundamental
Fourier techniques. Index of notation. Glossary of terms. 24 figures. 12 tables
192pp.
S507 Paperbound $1.85
5% x 8%.

TREATISE ON ELECTRICITY AND MAGNETISM, James Clerk Maxwell. For more than 80 years
seemingly inexhaustible source of leads for physicists, mathematicians, engineers. Total
1082pp. on such topics as Measurement of Quantities, Electrostatics, Elementary MatheTheory of Electricity, Electrical Work and Energy in a System of Conductors, General Theorems, Theory of Electrical Images, Electrolysis, Conduction, Polarization, Dielectrics,
Resistance, much more. "The greatest mathematical physicist since Newton," Sir James
S186 Clothbound $4.95
Jeans. 3rd edition. 107 figures, 21 plates. 1082pp. 53/8 x 8.
a
of

matical

CHEMISTRY AND PHYSICAL CHEMISTRY

THE PHASE RULE AND ITS APPLICATIONS, Alexander Findlay. Covers chemical phenomena of
1 to 4 multiple component systems, the "standard work on the subject" (Nature)
Completely
Smith New material on binary, tertiary
revised, brought up to date by A N. Campbell, N
liquid equilibria, solid solutions in ternary systems, quinary systems of salts, water, etc
Completely revised to triangular coordinates in ternary systems, clarified graphic representasolid models, etc 9th revised edition. 236 figures. 505 footnotes, mostly bibliographic
S92 Paperbound $2.45
xu + 449pp. 5 Ye x 8.

tion,

DYNAMICAL THEORY OF GASES, James

Jeans. Divided into mathematical, physical chapters for

Discusses mathematical theory of gas
convenience of those not expert in mathematics
in steady state, thermodynamics, Bolzmann, Maxwell, kinetic theory, quantum theory, expoas lucid and comprehensive
nentials, etc. "One of the classics of scientific writing
an exposition of the kinetic theory as has ever been written," J of Institute of Engineers
4th enlarged edition, with new material on quantum theory, quantum dynamics, etc 28 figures.
S136 Paperbound $2.45
444pp. GVa x 9V4.
.

MOLECULES, Pieter Debye. Nobel laureate offers complete guide to fundamental

electric
electrostatic field relations, polanzability, molecular structure. Partial contents
intensity, displacement, force, polarization by orientation, molar polarization, molar refracetc.
Special
tion,
halogen-hydrides, polar liquids, ionic saturation, dielectric constant,
coordination of experimental
chapter considers quantum theory. "Clear and concise
x 8.
results with theory will be readily appreciated," Electronics Industries
172pp

POLAR

5%

563 Clothbound $3.50

564 Paperbound $1.50

ATOMIC SPECTRA AND ATOMIC STRUCTURE,

G. Herzberg. Excellent general survey for chemphysicists specializing in other fields. Partial contents: simplest line spectra, elements
atomic theory, multiple structure of line spectra, electron spin; building-up principle,
periodic system of elements, finer details of atomic spectra, hyperfme structure of spectral
80 figures. 20 tables, xiii + 257pp.
lines; some experimental results and applications
S115 Paperbound $1.95
5% x 8.

ists,

of

TREATISE ON THERMODYNAMICS, Max Planck. Classic based on his original papers. Brilliant
concepts of Nobel laureate make no assumptions regarding nature of heat, rejects earlier
approaches of Helmholtz, Maxwell, to offer uniform point of view for entire field. Seminal
work by founder of quantum theory, deducing new physical, chemical laws. A standard
text, an excellent introduction to field for students with knowledge of elementary chemistry,
S219 Paperbound $1.75
physics, calculus. 3rd English edition, xvi + 297pp. 5% x 8.

DOVER SCIENCE BOOKS

KINETIC THEORY OF LIQUIDS, J. Frenkel. Regards kinetic theory of liquids as generalization,
extension of theory of solid bodies, covers all types of arrangements of solids, thermal
displacements of atoms, interstitial atoms, ions, onentational, rotational motion of molecules, transition between states of matter Mathematical theory developed close to physical
will serve as
subject matter. "Discussed in a simple yet deeply penetrating fashion
seeds for a great many basic and applied developments in chemistry," J of the Amer.
Chemical Soc. 216 bibliographical footnotes. 55 figures xi + 485pp 5% x 8
S94 Clothbound $3.95
S95 Paperbound $2.45
.

ASTRONOMY
OUT OF THE SKY,

H. H. Nininger. Non-technical, comprehensive introduction to "meteontics"

science concerned with arrival of matter from outer space By one of world's experts
meteorites, this book defines meteors and meteorites, studies fireball clusters and
processions, meteorite composition, size, distribution, showers, explosions, origins, much
x 8.
more vin
T519 Paperbound $1.85
336pp.

on

5%

AN INTRODUCTION TO THE STUDY OF STELLAR STRUCTURE,

S.

Chandrasekhar.

Outstanding

on stellar dynamics by one of greatest astro-physicists Examines relationship between loss of energy, mass, and radius of stars in steady state Discusses thermodynamic
laws from Caratheodory's axiomatic standpoint, adiabatic, polytropic laws, work of Ritter,
Emden, Kelvin, etc
Stroemgren envelopes as starter for theory of gaseous stars, Gibbs
statistical mechanics (quantum), degenerate stellar configuration, theory of white dwarfs,
Bulletin
etc. "Highest level of scientific merit,"
Amer. Math
Soc
33 figures. 509pp.
S413 Paperbound $2.75
5% x 8.
treatise

LES METHODES NOVELLES DE LA MECANIQUE CELESTE, H. Poincare. Complete French text

one of Pomcar6's most important works Revolutionized celestial mechanics- first use of
integral invariants, first major application of linear differential equations, study of periodic
orbits, lunar motion and Jupiter's satellites, three body problem, and many other important
"Started a new era ... so extremely modern that even today few have mastered
topics
his weapons," E. T. Bell. 3 volumes. Total 1282pp. 6Va x 9V4
1 S401
Vol
Paperbound $2.75
2 S402 Paperbound $2.75
Vol
Vol
3 S403 Paperbound $2.75
The set $7.50
of

THE REALM OF THE NEBULAE,

E. Hubble. One of the great astronomers of our time presents

concept of "island universes," and describes its effect on astronomy Covers velocityrelation, classification, nature, distances, general field of nebulae, cosmological
theories, nebulae in the neighborhood of the Milky way, etc. 39 photos, including velocitydistance relations shown by spectrum comparison. "One of the most progressive lines
New Introduction by A Sandage 55 illustraof astronomical research," The Times, London
S455 Paperbound $1.50
tions, xxiv + 201pp. 53/8 x 8.

his

distance

HOW

TO MAKE A TELESCOPE, Jean Texereau. Design, build an f/6 or f/8 Newtonian type
reflecting telescope, with altazimuth Couder mounting, suitable for planetary, lunar, and
Disstellar observation. Covers every operation step-by-step, every piece of equipment.
cusses basic principles of geometric and physical optics (unnecessary to construction),
comparative merits of reflectors, refractors. A thorough discussion of eyepieces, finders,
grinding, installation, testing, etc. 241 figures, 38 photos, show almost every operation
and tool. Potential errors are anticipated. Foreword by A Couder. Sources of supply, xin
T464 Clothbound $3.50
191pp. 6V4 x 10.

BIOLOGICAL SCIENCES
G. K. Noble, Late Curator of Herpetology at Am. Mus. of
most used text on amphibia, most comprehensive, clear, detailed. 19
page supplement: development; heredity, life history; speciation, adaptation;
instinct,
sex, integument, respiratory, circulatory, digestive, muscular, nervous systems;
"Nothing
intelligence, habits, economic value classification, environment relationships, etc
comparable to it;* C H. Pope, curator of Amphibia, Chicago Mus. of Nat. Hist. 1047 item
S206 Paperbound $2.98
bibliography. 174 illustrations. 600pp. 5% x 8.

THE BIOLOGY OF THE AMPHIBIA,

Nat.

Hist.

chapters,

Probably

85

LIFE, A. I. Oparin. A classic of biology. This is the first modern statement

of theory of gradual evolution of life from nitrocarbon compounds. A brand-new evaluation
of Oparm's theory in light of later research, by Dr. S. Margulis, University of Nebraska,

THE ORIGIN OF
xxv

270pp.

5%

x 8.

S213 Paperbound $1.75

CATALOGUE OF
THE BIOLOGY OF THE LABORATORY MOUSE, edited by

Prepared in 1941 by staff

Roscoe B
Jackson Memorial Laboratory, still the standard treatise on the mouse,
amount
of
material
for
which
enormous
otherwise
assembling
you spend hours of research.
Embryology, reproduction, histology, spontaneous neoplasms, gene and chromosomes mutatumor
of
tumor
of
spontaneous
formations,
tions, genetics
transplantation, endocrine secremilk influence and tumor formation, inbred, hybrid animals,
tion and tumor formation,
care
and
wealth
"A
of information of vital coninfectious
diseases,
recording
parasites,
recommended to all who could use a book on such a subject," Nature Classified
cern
items
172
128
ix
of
1122
+ 497pp 6Vs x 9V4.
figures,
including
photos,
bibliography
S248 Clothbound $6.00
G.

0.

Snell.

of

THE TRAVELS OF WILLIAM BARTRAM, edited by Mark Van

American anthropology, natural history, geography, is record
travels through wilderness of Florida, Georgia, Carolmas
descriptions of Indians, settlers, fauna, flora, it is one
ever written 13 original illustrations 448pp 5% x 8.

Doran. Famous source-book of

kept by Bartram in 1770's on
Containing accurate, beautiful
of finest pieces of Americana

T13 Paperbound $2.00

BEHAVIOUR AND SOCIAL LIFE OF THE HONEYBEE, Ronald Ribbands. Outstanding scientific
study, a compendium of practically everything known of social life of honeybee. Stresses
Extends von Frisch's experiments on communiindividual bees in field, hive
Covers perception of temperature, gravity, distance, vibration, sound
cation among bees
production, glands, structural differences, wax production, temperature regulation, recogni"This
communication, drifting, mating behaviour, other highly interesting topics
tion,
valuable work is sure of a cordial reception by laymen, beekeepers and scientists," Prof.
Karl von Frisch, Brit J of Animal Behaviour
Bibliography of 690 references 127 diagrams,
S410 Clothbound $4.50
graphs, sections of bee anatomy, fine photographs, 352pp
behaviour of

ELEMENTS OF MATHEMATICAL BIOLOGY,

apply

A.

J.

Lotka.

Pioneer classic,

modern mathematical techniques on large scale

to

1st

phenomena

of

major

attempt to
biochem-

biology,

Partial
contents
Statistical
psychology, ecology, similar life sciences
meaning of
irreversibility, Evolution as redistribution, Equations of kinetics of evolving systems, Chemical,
mter-species equilibrium, parameters of state; Energy transformers of nature, etc
Can be read with profit by even those having no advanced math, unsurpassed as study"
72 figures xxx -4- 460pp 53/a x 8
reference. Formerly titled "Elements of Physical Biology
S346 Paperbound $2.45
istry,

TREES OF THE EASTERN AND CENTRAL UNITED STATES AND CANADA, W. M. Harlow

Serious

middle-level text covering more than 140 native trees, important escapes, with information on general appearance, growth habit, leaf forms, flowers, fruit, bark, commercial use,
Keys within text enable you to locate various species
distribution, habitat, woodlore, etc
"Well
easily, to know which have edible fruit, much more useful, interesting information
Over 600
make
identification
to
illustrated
very easy," Standard Cat. for Public Libraries
T395 Paperbound $1.35
288pp. 5 5/s x 6V2
photographs, figures XIM

FRUIT KEY AND TWIG KEY TO TREES AND SHRUBS (Fruit key to Northeastern Trees, Twig key
to Deciduous Woody Plants of Eastern North America), W. M. Harlow. Only guides with photographs of every twig, fruit described Especially valuable to novice Fruit key (both deciduous
Twig key
evergreens) has introduction on seeding, organs invojved, types, habits
In keys proper,
identification is almost automatic
introduction treats growth, morphology
Exceptional work, widely used in university courses, especially useful for identification in
Index of
winter, or from fruit or seed only Over 350 photos, up to 3 times natural size
T511 Paperbound $1.25
common, scientific names, in each key. xvn + 125pp. 5 5/a x 8%
trees,

INSECT LIFE AND INSECT NATURAL HISTORY, S. W. Frost. Unusual for emphasizing habits, social
ecological relations of insects rather than more academic aspects of classification,
life,
morphology. Prof. Frost's enthusiasm and knowledge are everywhere evident as he discusses
insect associations, specialized habits like leaf-rolling, leaf mining, case-making, the gall
Examines matters not usually covered in general works insects
insects, boring insects, etc
as human food, insect music, musicians, insect response to radio waves, use of insects in
literature.
art,
"Distinctly different, possesses an individuality all its own," Journal of
x + 524pp
5 3/e x 8
Forestry. Over 700 illustrations. Extensive bibliography
T519 Paperbound $2.49
A WAY OF LIFE, AND OTHER SELECTED WRITINGS, Sir William Osier. Physician, humanist,
Osier discusses brilliantly Thomas Browne, GUI Patm, Robert Burton, Michael Servetus,
William Beaumont, Laennec. Includes such favorite writing as title essay, "The Old Humanities and the New Science," "Books and Men," "The Student Life," 6 more of his best
discussions of philosophy, literature, religion. "The sweep of his mind and interests embraced every phase of human activity," G L Keynes. 5 photographs Introduction by G L
T488 Paperbound $1.50
Keynes, M.D., F.R.C S xx + 278pp. 53/3 x 8.

THE GENETICAL THEORY OF NATURAL SELECTION,

R. A. Fisher. 2nd revised edition of vital

reviewing of Darwin's Selection Theory in terms of particulate inheritance, by one of
Theory stated in mathematical
greatest authorities on experimental, theoretical genetics
form. Special features of particulate inheritance are examined evolution of dominance, maintenance of specific variability, mimicry, sexual selection, etc. 5 chapters on man's special
circumstances as a social animal. 16 photographs, x + 310pp. 53/8 x 8

S466 Paperbound $1.85

10

DOVER SCIENCE BOOKS

AUTOBIOGRAPHY OF CHARLES DARWIN, AND SELECTED LETTERS, edited by Francis
Darwin. Darwin's own record of early life; historic voyage aboard "Beagle," furore surroundLetters to Henslow, Lyell, Hooker,
ing evolution, his replies, reminiscences of his son
and thoughts on religion, vivisection We see how he revoHuxley, Wallace, Kingsley, etc
lutionized geology with concepts of ocean subsidence, how his great books on variation
of plants and animals, primitive man, expression of emotion among primates, plant fertilizacame into being 365pp 5% x 8
tion, carnivorous plants, protective coloration, etc
T479 Paperbound $1.65

THE

ANIMALS

IN MOTION, Eadweard Muybridge. Largest, most comprehensive selection of Muy"

3919 high-speed
bridge's famous action photos of animals, from his "Animal Locomotion
shots of 34 different animals, birds, in 123 types of action, horses, mules, oxen, pigs,
cats
camels,
dogs,
guanacos,
lions,
elephants,
sloths,
goats,
tigers,
jaguars, raccoons,
baboons, deer, elk, gnus, kangaroos, many others, walking, running, flying, leaping Horse
alone in over 40 ways. Photos taken against ruled backgrounds, most actions taken from
3 angles at once
90, 60, rear Most plates original size Of considerable interest to
scientists as biology classic, records of actual facts of natural history, physiology
"Really
marvelous series of plates," Nature "Monumental work," Waldemar Kaempffert Edited by
L. S.
Brown, 74 page introduction on mechanics of motion 340pp of plates. 3919 photoDeluxe binding, paper (Weight 41/2 Ibs ) 7Va x 10%
graphs. 416pp

T203 Clothbound $10.00

THE HUMAN FIGURE

MOTION, Eadweard Muybridge New edition of great classic in history

~'ection ever made from original Muybridge photos of
of science and photogran*
*? 163 types of motion
human action 478
walking, running, lifting,
eds up to l/6000th of a second Men, women,
e*~ hi tir.^
cnno.ei,,

,.

IN

muscle positions against ruled backgrounds,

/ was this a great work of photography, acclaimed
genius, but it was also a great 19th century landmark
Robert Taft, U
of Kansas
Plates
Historical introduction by Prof
of detail
Over 500 action strips 407pp. 7 3/4 x 10% Deluxe edition

mostly taken ai ,
by contemporary crm*.
in
biological research
original

size,

full

7204 Clothbound $10.00

AN INTRODUCTION TO THE STUDY OF EXPERIMENTAL MEDICINE, Claude Bernard. 90-year

old
classic of medical science, only major work of Bernard available in English, records his
efforts to transform physiology into exact science
Principles of scientific research illustrated by specified case histories from his work, roles of chance, error, preliminary false
Much of modern
conclusion, m leading eventually to scientific truth, use of hypothesis
application of mathematics to biology rests on foundation set down here "The presentation
is polished
reading is easy," Revue des questions scientifiques New foreword by Prof
T400 Paperbound $1.50
B Cohen, Harvard U xxv
I.
266pp 53/8 x 8

STUDIES ON THE STRUCTURE AND DEVELOPMENT OF VERTEBRATES, E. S Goodrich. Definitive

Exhaustive morphological, phylogenetic
study by greatest modern comparative anatomist
skeleton, fins, limbs, skeletal visceral arches, labial cartilages, visceral
from fish to
vascular, respiratory, excretory, penphal nervous systems, etc
"For many a day this will certainly be the standard- textbook on Vertebrate
Morphology in the English language," Journal of Anatomy 754 illustrations 69 page biographical study by C C Hardy Bibliography of 1186 references Two volumes, total 906pp
Two vol. set S449, 450 Paperbound $5.00
53/8 x 8

expositions
clefts,

higher

of

gills,

mammals

EARTH SCIENCES
THE EVOLUTION OF IGNEOUS BOOKS,

N. L. Bowen. Invaluable serious introduction applies

to explain igneous rock diversity in terms of chemical
Discusses liquid immiscibihty in silicate magmas,
crystallization
crystal sorting, liquid lines of descent, fractional resorption of complex minerals, petrogen,
etc. Of prime importance to geologists, mining engineers, physicists, chemists working with

techniques of physics,
composition, fractional

chemistry

"Most
high temperature, pressures
82 figures, xviu + 334pp 53/8 x 8

GEOGRAPHICAL ESSAYS, M.

Davis.

important,"

Modern

Times,

geography,

London

263 bibliographic notes

S311 Paperbound $1.85

geomorphology

rest

on

fundamental

26 famous essays present most important theories, field researches

of this scientist
Partial contents
Geographical Cycle, Plains of Marine, Subaenal Denudation, The Peneplain,
Rivers, Valleys of Pennsylvania, Outline of Cape Cod, Sculpture of Mountains by Glaciers,
"Part of the very texture of geogetc "Long the leader and guide," Economic Geography
x 8
.
models of clear thought," Geographic Review 130 figures vi
777pp
raphy .

work

5%

S383 Paperbound $2.95

URANIUM PROSPECTING,

H. L. Barnes. For immediate practical use, professional geologist

uranium ores, geological occurrences, field conditions, all aspects of highly
easy-to-use, easy-to-fmd style," Geotimes.
"Helpful information
profitable occupation
T309 Paperbound $1.00
X
117pp. 53/8 x 8.

considers

11

CATALOGUE OF
DE RE METALLICA, Georgius Agricola. 400 year old classic translated, annotated by former
President Herbert Hoover. 1st scientific study of mineralogy, mining, for over 200 years
after its appearance in 1556 the standard treatise. 12 books, exhaustively annotated, discuss
history of mining, selection of sites, types of deposits, making pits, shafts, ventilating,
pumps, crushing machinery; assaying, smelting, refining metals, also salt alum, nitre, glass
making. Definitive edition, with all 289 16th century woodcuts of original. Biographical,
Indexes. A fascinating book
historical introductions. Bibliography, survey of ancient authors
for anyone interested m art, history of science, geology, etc. Deluxe Edition. 289 illustraS6 Clothbound $10.00
tions. 672pp. 6% x 10. Library cloth.
INTERNAL CONSTITUTION OF THE EARTH, edited by Beno Gutenberg. Prepared for National
Research Council, this is a complete, thorough coverage of earth origins, continent formanature and behaviour of earth's core, petrology of crust, cooling forces in core,
tion,
seismic and earthquake material, gravity, elastic constants, strain characteristics, similar
a high standard
there is no reader who
topics "One is filled with admiration
will not learn something from this book," London, Edinburgh, Dublin, Philosophic Magazine.
43 diagrams.
Largest Bibliography in print- 1127 classified items. Table of constants
S414 Paperbouno* $2.45
439pp. 6Vs x 9V4
.

THE BIRTH AND DEVELOPMENT OF THE GEOLOGICAL SCIENCES,

F.

Adams. Most thorough

D.

Geological thought from earliest times to end of

history of earth sciences ever written
19th century, covering over 300 early thinkers and systems, fossils and their explanation,
vulcamsts vs neptumsts, figured stones and paleontology, generation of stones, dozens of
similar topics 91 illustrations, including Medieval, Renaissance woodcuts, etc. 632 footnotes,
T5 Paperbound $2.00
mostly bibliographical 511pp. 5 3/a x 8.
E. Meinzer, prepared for the National Research Council
Detailed,
on precipitation, evaporation, snow, snow surveying, glaciers,
moisture, ground water, runoff, drought, physical changes produced
by water hydrology of limestone terranes, etc Practical in application, especially valuable
24 experts have created "the most up-to-date, most complete treatment of
for engineers
the subject," Am Assoc of Petroleum Geologists. 165 illustrations, xi
712pp. 6Vs x 9V4.

HYDROLOGY, edited by
complete
lakes,

reference

infiltration,

0.

library

soil

S191 Paperbound $2.95

LANGUAGE AND TRAVEL AIDS FOR SCIENTISTS

SAY

IT language phrase books

"SAY IT" in the foreign language of your choice! We have sold over Vz million copies of
these popular, useful language books. They will not make you an expert linguist overnight,
but they do cover most practical matters of everyday life abroad
Over 1000 useful phrases, expressions, additional

Modern! Useful!
tioned,"

Hundreds of phrases

not

variants,

available

in

substitutions

other

texts.

The ONLY inexpensive phrase book completely indexed. Everything

of

your

"Nylon,"

"air-condi-

etc.

finger,

ready to

is

available

at a

flip

use

Prepared by native linguists, travel experts.

Based on years of travel experience abroad.

May be used by itself, or to supplement any other text or course. Provides a living element. Used by many colleges, institutions- Hunter College, Barnard College, Army Ordinance
School, Aberdeen;

etc.

1 book per language:

Danish (T818) 75$
Dutch (T817) 750
English (for German-speaking people) (T801) 600

Available,

English (for Italian-speaking people) (T816) 600

English (for Spanish-speaking people) (T802) 600
Esperanto (T820) 750
French (T803) 600
German (T804) 600
Modern Greek (T813) 750
Hebrew (T805) 600

12

Italian (T806) 600

Japanese (T807) 75C
Norwegian (T814) 750
Russian (T810) 750
Spanish (T811) 600
Turkish (T821) 750
Yiddish (T815) 750
Swedish (T812) 750
(T808) 75C
Portuguese (T809) 750

**"

DOVER SCIENCE BOOKS

MONEY CONVERTER AND TIPPING GUIDE FOR EUROPEAN TRAVEL, C. Vomacka. Purse-size handbook crammed with information on currency regulations, tipping for every European country,
including Israel, Turkey, Czechoslovakia, Rumania, Egypt, Russia, Poland. Telephone, postal
rates, duty-free imports, passports, visas, health certificates, foreign clothing sizes; weather
tables. What, when to tip. 5th year of publication. 128pp. 3V2 x 5V4
T260 Paperbound 600
.

NEW RUSSIAN-ENGLISH AND ENGLISH-RUSSIAN

M. A. O'Brien. Unusually comprehensive guide to reading, speaking, writing Russian, for both advanced, beginning students. Over 70,000 entries in new orthography, full information on accentuation, grammatical
classifications. Shades of meaning, idiomatic uses, colloquialisms, tables of irregular verbs
Individual entries indicate stems, transitiveness, perfective,
for both languages.
im perfective aspects, conjugation, sound changes, accent, etc. Includes pronunciation instruction.
Used at Harvard, Yale, Cornell, etc. 738pp. 5% x 8.
T208 Paperbound $ 2.00

DICTIONARY,

PHRASE AND SENTENCE DICTIONARY OF SPOKEN RUSSIAN, English-Russian, Russian-English.

Based on phrases, complete sentences, not isolated words recognized as one of best
methods of learning idiomatic speech. Over 11,500 entries, indexed by single words, over
Shows accent
32,000 English, Russian sentences, phrases, in immediately useable form
changes in conjugation, declension, irregular forms listed both alphabetically, under mam
form of word. 15,000 word introduction covers Russian sounds, writing, grammar, syntax
15 page appendix of geographical names, money, important signs, given names, foods,
special Soviet terms, etc
x 8.

5%

Gov't Manual

Originally published as U.S

TM

30-944.

iv

573pp.

T496 Paperbound $2.75

PHRASE AND SENTENCE DICTIONARY OF SPOKEN SPANISH, Spanish-English, English-Spanish.

Compiled from spoken Spanish, based on phrases, complete sentences rather than isolated
words not an ordinary dictionary. Over 16,000 entries indexed under single words, both
25 page introduction
Castihan, Latin-American. Language in immediately useable form
provides rapid survey of sounds, grammar, syntax, full consideration of irregular verbs.
Especially apt in modern treatment of phrases, structure 17 page glossary gives translations
of geographical names, money values, numbers, natronal holidays, important street signs,
useful expressions of high frequency, plus unique 7 page glossary of Spanish, SpanishAmerican foods. Originally published as U.S. Gov't Manual TM 30-900. iv + 513pp. 5% x 8%.
T495 Paperbound $1.75

SAY

IT

CORRECTLY

language record sets

The best inexpensive pronunciation aids on the market. Spoken by native

ciated with major American universities, each record contains.
14 minutes of speech 12 minutes of normal,
normal conversational speed.

relatively

linguists

slow speech, 2 minutes of

120 basic phrases, sentences, covering nearly every aspect of everyday

introducing yourself, travel in autos,
restaurants, money, shopping, etc.

buses,

taxis,

etc.,

asso-

walking,

life,

sightseeing,

travel
hotels,

32 page booklet containing everything on record plus English translations easy-to-follow

phonetic guide.
Clear,

high-fidelity

recordings.

Unique bracketing systems, selection of basic sentences enabling you to expand use of
SAY IT CORRECTLY records with a dictionary, to fit thousands of additional situations.

Use this record to supplement any course or text. All sounds in each language illustrated
imitate speaker in pause which follows each foreign phrase in slow section,
perfectly
and be amazed at increased ease, accuracy of pronounciation. Available, one language per
record for

French

Japanese

Spanish
Dutch
Russian

Polish

Swedish

Italian

German
Modern Greek
Portuguese

Hebrew
English (for Spanish-speaking people)

English (for German-speaking people)

7" (33 1/3 rpm)

record, album, booklet. $1.00 each.

SPEAK MY LANGUAGE: SPANISH FOR YOUNG BEGINNERS. M. Ahlman,

Z.

Gilbert.

Records pro-

vide one of the best, most entertaining methods of introducing a foreign language to
children. Within framework of train trip from Portugal to Spam, an English-speaking child
is introduced to Spanish by native companion. (Adapted from successful radio program of
N.Y. State Educational Department.) A dozen different categories of expressions,, including
greeting, numbers, time, weather, food, clothes, family members, etc. Drill is combined
with poetry and contextual use. Authentic background music. Accompanying book enables
a reader to follow records, includes vocabulary of over 350 recorded expressions. Two
10" 33 1/3 records, total of 40 minutes. Book. 40 illustrations. 69pp. 5V* x 10V2.
T890 The set $4.95

13

CATALOGUE OF
LISTEN & LEARN

language record sets

LISTEN & LEARN is the only extensive language record course designed especially to meet
your travel and everyday needs. Separate sets for each language, each containing three 33 1/3
rpm long-playing records 1 1/2 hours of recorded speech by eminent native speakers
Queens College
who are professors at Columbia, New York U
,

Check the following features found only

in

LISTEN & LEARN

Dual language recording 812 selected phrases, sentences, over 3200 words, spoken first
Pause after each foreign phrase allows time to
in English, then foreign equivalent
repeat expression.

everything on records, plus simple transcrip128-page manual (196 page for Russian)
tion. Indexed for convenience Only set on the market completely indexed
Practical
No time wasted on material you can find in any grammar No dead words.
Covers central core material with phrase approach Ideal for person with limited time.
Living, modern expressions, not found in other courses
Hygienic products, modern
equipment, shopping, "air-conditioned," etc Everything is immediately useable
High-fidelity recording,

equal

in

clarity to

to

any costing up

$6 per record.

"Excellent
impress me as being among the very best on the market," Prof
of Romance Languages, Columbia U
"Inexpensive and well done
Pei, Dept
"More genuinely helpful than anything of its
present," Chicago Sunday Tribune
of
"All
the
Best"
author
travel books
well-known
Sidney Clark,
.

UNCONDITIONAL GUARANTEE.
refund,

if

Try

you are not satisfied

It

LISTEN & LEARN, then return it within

is guaranteed after you actually use it

modern languages FRENCH, SPANISH, GERMAN,

language to each set of 3 records (33 1/3 rpm).

Spanish
French
*

Available

Mario

the set $4 95
the set $4 95
Oct.

German
Italian

ITALIAN,

RUSSIAN,

128 page manual

the set $4 95
the set $4 95

10

days

or JAPANESE
Album

ideal

kind,"

for

full

one

the set $5 95
the set $5 95

Japanese*
Russian

1959

TRUBNER COLLOQUIAL SERIES

These unusual books are members of the famous Trubner series of colloquial manuals They
have been written to provide adults with a sound colloquial knowledge of a foreign language, and are suited for either class use or self-study Each book is a complete course in
itself, with progressive, easy to follow lessons
Phonetics, grammar, and syntax are covered,
while hundreds of phrases and idioms, reading texts, exercises, and vocabulary are included
These books are unusual in being neither skimpy nor overdetailed in grammatical matters,
and in presenting up-to-date, colloquial, and practical phrase material. Bilingual presentation
is
stressed, to make thorough self-study easier for the reader
H. Harley, formerly Nizam's Reader in Urdu, U
of London
30
phonetics and scripts (devanagan & Arabic-Persian) are followed by 29 lessons,
including material on English and Arabic-Persian influences Key to all exercises. Vocabufary.
5 x 7V2. 147pp
Clothbound $1.75

COLLOQUIAL HINDUSTANI, A
pages on

COLLOQUIAL ARABIC, DeLacy O'Leary. Foremost

Islamic

scholar

Syria, Palestine, & Northern Arabia Extremely clear coverage of

Vocabulary, xvni
plurals, also cultural aspects of language

covers

language

of

Egypt,

complex Arabic verbs & noun

192pp.

7V2.

Clothbound $1.75

COLLOQUIAL GERMAN,
form.
texts.

Intensive thorough coverage of grammar in easily-followed

with hundreds of colloquial phrases. 34 pages of bilingual

P. F. Doring.

Excellent for brush-up,

224pp 5 x 7V2

COLLOQUIAL SPANISH, W.

Clothbound $1.75

Patterson. Castilian grammar and colloquial language, loaded

with bilingual phrases and colloquialisms Excellent for review or self-study 164pp 5 x 7V2.
R.

Clothbound $1.75

COLLOQUIAL FRENCH, W. R. Patterson. 16th revised edition of this extremely popular manual.
Grammar explained with model clarity, and hundreds of useful expressions and phrases,
exercises, reading texts, etc. Appendixes of new and useful words and phrases. 223pp.
Clothbound $1.75

5 x 7V2.

14

DOVER SCIENCE BOOKS

COLLOQUIAL PERSIAN,

L. P. Elwell-Sutton. Best introduction to modern Persian, with 90 page

Clothbound $1.75
grammatical section followed by conversations, 35 page vocabulary. 139pp.

COLLOQUIAL CZECH,

J.
Schwarz, former headmaster of Lingua Institute, Prague Full easily
followed coverage of grammar, hundreds of immediately useable phrases, texts Perhaps the
best Czech grammar in print
"An absolutely successful textbook," JOURNAL OF CZECHOSLOVAK FORCES IN GREAT BRITAIN. 252pp 5 x 7V2
Clothbound $250

COLLOQUIAL RUMANIAN, G. Nandris, Professor of University of London Extremely thorough

coverage of phonetics, grammar, syntax, also included 70 page reader, and 70 page vocabulary
Probably the best grammar for this increasingly important language 340pp 5 x 7 Viz
Clothbound $2.50
Excellent self-study course m grammar, vocabulary,
ITALIAN, A. L. Hayward
idioms, and reading Easy progressive lessons will give a good working knowledge of Italian
the shortest possible time 5 x 7 Viz
Clothbound $1.75

COLLOQUIAL
in

MISCELLANEOUS
TREASURY OF THE WORLD'S COINS, Fred

Remfeld. Finest general introduction to numisCoins of Greece, Rome, modern counnon-technical, thorough, always fascinating
every continent, primitive societies, such oddities as 200-lb stone money of Yap,
nail coinage of New England, all mirror man's economy, customs, religion, politics, philosophy, art. Entertaining, absorbing study, novel view of history. Over 750 illustrations.
Table of value of coins illustrated List of U S com clubs. 224pp. 6V2 x 91/4
matics,
tries

of

T433 Paperbound $1.75

ILLUSIONS AND DELUSIONS OF THE SUPERNATURAL AND THE OCCULT, D.
tionally examines hundreds of persistent delusions
including witchcraft,

H.

Rawcliffe.

Ra-

trances, mental
Indian rope trick,
burial, auras,
healing, peyotl, poltergeists, stigmata, lycanthropy,
Explains, exposes mental, physical despiritualism, dowsing, telepathy, ghosts, ESP, etc
ceptions involved, making this not only an expos6 of supernatural phenomena, but a valuable
exposition of characteristic types of abnormal psychology
Originally "The Psychology of
14 illustrations
the Occult." Introduction by Julian Huxley
551pp 53/8 x 8
live

T503 Paperbound $200

MacDougall. Shows how art, science, history, journalism can be perverted
Hours of delightful entertainment, a work of scholarly value, often
for private purposes
shocking Examines nonsense news, Cardiff giant, Shakespeare forgeries, Loch Ness monster,
biblical frauds, political schemes, literary hoaxers like Chatterton, Ossian, disumbrationist
school of painting, lady m black at Valentino's tomb, over 250 others. Will probably reveal
truth about few things you've believed, will
help you spot more easily the editorial
and shrewd analysis,"
"gander" or planted publicity release "A stupendous collection
T465 Paperbound $1.75
New Yorker New revised edition 54 photog/aphs. 320pp 5% x 8

HOAXES,

C.

D.

YOGA: A SCIENTIFIC EVALUATION, Kovoor

T. Betran*m. Book that for first time gave Western

Author draws on laboratory
readers a sane, scientific explanation, analysis of yoga
experiments, personal records of year as disciple of yoga, to investigate yoga psychology,
to
plumb deepest human powers. In this
physiology, "supernatural" phenomena, ability
study under auspices of Yale University Institute of Human Relations, strictest principles
of physiological, psychological inquiry are followed Forevyord byiW A Miles, Yale University
T505 Paperbound $1.65
17 photographs xx + 270pp. 5Vs x 8

Write for free catalogs!

Indicate your field of interest. Dover publishes books on physics, earth
sciences, mathematics, engineering, chemistry, astronomy, anthropology, biology, psychology, philosophy, religion, history, literature, mathematical recreations, languages, crafts, art, graphic arts, etc.
Write
Science B

V_^| ^>

'

to

Dept. catr

Dover Publications,

__
'

180 Varick
15

St.,

Inc.

N. Y.

14,

N. Y.

(continued ftotn inside front covet)

Introduction to Differential Equations of Physics, L.

The Continuum and Other

Tyfces of Serial Order,

Hopf

Ordinary Differential Equations, E L Ince

Table of Functions, E Jahnke cr F Emde

of Integration,

$1.00

$2.55

$200

Theory of Sets, L Kamke $1 35

Foundations of Potential Theory, O D Kellogg

Modern Theories

$1.25

V Huntington

$1 98

Kestelnian

$1 75

A Khmchm

$1 35

Mathematical Foundations of Information Theory, A. Khmchm

Lectures on the Icosahedron, Felix Klein $1 85
$1.75
Infinite Sequences and Series, K Knopp

$1 35

Mathematical Foundations

of Statistical Mechanics,

Elements of Theory of t unctions, K Knopp $1 35

Theory of Functions, K. Knopp Four volume set, $5.40
Dictionary of Conformal representations, H Kober $200
of Differential Equations, H Levy cr E Raggott
Elements of Theory of Real Functions, J Littlewood $1.25
Linear Integral Equations, H' L Lovitt $1.60

Numerical Solutions

Higher Mathematics

for

Students of Chemistry and Physics,

Differential Equations,

F R Moulton

J.

$1 75

Mellot

$225

$200

Functions of a Complex Variable, James Pierpont $2J5

Lectures on the Theory of Functions of Real Variables, James Pierpont
Two Volume set $-1 90
Collected Works of Bernard

A Table

Riemann

(in

Elements of Non-Euclidean Geometry,

An

German),

Riemann

1!

$285

of the Incomplete Elliptic Integral of the Third Kind,

G Selfndge & J, E Maxfield ^Clothbound $7 50

Introduction to Geometry of

Elements of

Y Somnrerville

Dimensions, ,D

Number Theory/ 1 Mr'Tinogradov

Theory of Functional, Tito* Kolterra

and Tensor Analysis, Albeit P

Advanced Cahulus, E H Wilson

Trigonometrical Series, A

A G

Wills

7,ygmund

$1.50

$1 60

$175

Partial Differential Equations of Mathematical />/ty>ws

Vector

$1 50

Sommen'ille

Webster

$2.00

$1 75

$245
$1.50

Available at your book dealer or write for free catalogues to Dept. TF1,
Publications, Inc., 180 Varick St., N.Y. 14, N.Y. Please indicate
field of interest. Dover publishes over 75 new selections each year on

Dover

science, puzzles, art, languages, music, philosophy, etc.