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