Irving Strin^han

*th.
-4^ .C

f
 NOTES
ON

ELEMENTS
OF

(ANALYTICAL)

SOLID GEOMETRY

BY

CHAS. S. VENABLE, LL.D.,

PROFESSOR OF MATHEMATICS, UNIVE-RSITY OF VIRGINIA.

NEW YORK:
UNIVERSITY PUBLISHING COMPANY.
1891.
 Math. 0ot.

IN preparing these Notes I have used the treatises of Gregory,

Hymers, Salmon, Frost and Wostenholme, Bourdon, Sonnet et
Frontera, Joachimsthal-Hesse, and Fort und Schlomilch.
C. S. V.

COPYRIGHT, BY
UNIVERSITY PUBLISHING COMPANY
1879.
 NOTES ON SOLID GEOMETRY.

CHAPTER I.

1. WE have seen how the position of a point in a plane with ref

erence to a given origin O is determined by means of its distances

from two axes (Xr, Oy meeting in O. In space, as there are three
dimensions, we must add a third axis Oz. So that each pair of axes
determines a plane, CXr and Oy determining the plane xOy ; O.v
and O2 the plane xOz Oy and Oz the plane yOz. And the posi
;

tion of the point P with reference to the origin O is determined by

its distances PM, PN, PR from the zOy, zOx, xOy respectively, these
distances being measured on lines parallel to the axes CXv, Oy and
Oz respectively. This system of coordinates in space is called The
System of Triplanar Coordinates, and the transition to it from the
System of Rectilinear Plane Coordinates is very easy. We can best
conceive of these three coordinates of P by conceiving O as the
corner of a parallelopipedon of which OA, OB, OC are the edges,
and the point P is the opposite corner, so that OP is one diagonal of
the parallelopipedon.
2. If PM = OA = a, PN = OB = b, PR = OC = c, the equations
of the point P are x = a, y = b, z =
c, and the point given by
these

equations may be found by the following construction Measure on :

OX the distance OA = a, and through A draw the plane PNAR

parallel to the plane yOz. Measure on Oy the distance OB = 3,
and draw the plane PMBR parallel to xOz, and finally lay off OC
c and dnuv the plane PMCN parallel to xOy. The intersection
of these three planes is the point P required. (Fig. i.)*
3. The three axes Or, Oy,
Oz are called the axes of x y, and z }

respectively ;
the three planes xQy, xOz, and yOz are called the
*.

* For and end of book.

Figures see Plates I. II. at

814023
4 NOTES ON SOLID GEOMETRY.

planes xy, xz and yz respectively. The point whose equations are

x a, x b, x
= =
c is called the point (a, b, c).

4. The coordinate planes produced indefinitely form eight solid

angles about the point O. As in plane coordinates the axes Ox

and .cUvicieUtte plane considered into four compartments, so in
O>

space coordinates the planes xy. xz and yz divide the space con-
slcjttfed Compartments four above the plane xy, viz.
ia.;o .eight :

Q-xyz, Q-xy Q-x y z, Q-x yz and four below it, viz.

z, Q-xyz ;
:
,

f f

Q-xy z, O-x y z Q-x yz By an easy extension of the rule of

,
.

signs laid down in Plane Coordinate Geometry, we regard all x s

on the yz as + and on the left of yz as
right of the plane ally s ;

xz as -f- and those behind it as

in front of the plane all z s above ;

the plane xy as + and those below it as We can then write the .

points whose distances from the coordinate planes are a, b and c in

the eight different angles thus :

In the first Octant, Q-xyz P l

is (a, b, c)
In the second Octant, P 2 is (a, b, c)

In the third Octant, P3 is ( a, b, c]

In the fourth Octant, P 4 is
( a, b, c)
In the fifth Octant, P 5 is (a, b, c)
In the sixth Octant, P 6 is (a, b, c]
In the seventh Octant, I\ is ( a, b, c)
In the eighth Octant, P8 is ( a, b, c).

The signs thus tell us in which compartment the point falls,

and the lengths of a, b and c give us its position in these compart

ments.

1. Construct the points I, 2, 3 ; o, i, 2 ; 0,0, i

; 4, o, 3.

2. Construct the points i, 3, 4 ; 2, 3, o ; 3, o, i ; 2, o, o.

5. The points M, N
and R are called the projections of P on the
three coordinate planes, and when the axes are rectangular they are
its orthogonal
projections. We
will treat mainly of orthogonal pro

jections. For shortness sake when we speak simply of projections,

we are to be understood to mean orthogonal projections, unless we
state the contrary.
We will give now some other properties of orthogonal projections
which will be of use to us.
 NOTES ON SOLID GEOMETRY. 5

6 DEFINITIONS.

The
projection of a line on a plane is the line containing the

projections of its points on the plane.

When one line or several lines connected together enclose a plane

area, the area enclosed by the projection of the lines is called the

projection of the first area.

The
idea of projection may be in the case of the straight line thus
extended: if from the extremities of any limited straight line we draw

perpendiculars to a second line, the portion of the latter intercepted

between the feet of the perpendiculars is called the projection of the
limited line on the second line.

From this we see that OA, and OB OC

(coordinates rectangular)
are the projections of OP
on the three axes, or the rectangular coordi
nates of a point are the projections of its distance from the origin on the
coordinate axes.

7. FUNDAMENTAL THEOREMS.
I. The length of the projection of a finite right line on any plane is

equal to the line multiplied by the cosine of the angle which it makes with
tht,
plane.
Let PQ be the given finite straight line, xOy the plane of pro
jection draw PM, QN perpendicular to it then
; ;
MN is the
projec
tion of PQ on the plane. Now the angle made by PQ with the plane
is the angle made by PQ with MN. Through Q draw QR parallel
to MN meeting PM R, then
in QR = MN, and the angle PQR
= the angle made by PQ with MN. Now MN = QR = PQ cos
PQR. (Fig. 2.)

II. The projection on any plane

of any bounded plane area is equal to
thut area multiplied by the cosine
of the angle between the planes.
i. We shall begin with a triangle of which one side BC is parallel

to the plane of projection. The area of ABC = - BC x AD, and the

area of the projection ABC = - BC xA D. But BC = BC and

A D = AD ccs ADM. Moreover ADM = the angle between the
planes. Hence A B C= ABC x cos angle between the planes. (Fig. 3.)
2. Next take a triangle ABC
of which no one of the sides is
pa
rallel to the plane of projection. (Fig. 4.)
i*
6 NOTES ON SOLID GEOMETRY.

Through the corner C of the triangle draw CD parallel to the

plane of projection meeting AB in D. Now if we call 6 the angle
between the planes, then from i C A D = ACD cos 6 and B C D
= BCD cos 0. . . A B D -B C D =(ABD - BCD) cos or ABC
= ABC cos d.

3. Since every polygon may be divided up into a number of

triangles of each of which the proposition is true it is true also of

the polygon, i. e., of the sum of the triangles.

Also by the theory of limits, curvilinear areas being the limits of
polygonal areas, the proposition is also true of these.
8. The projection of a finite right line upon another right line is

equal to the first line multiplied by the cosine of the angle between the

lines.

Let PQ be the given line and MN its projection on the line CXv, by
means of the perpendiculars PM and QN. Through Q draw QR
parallel to MN and equal to it. Then PQR is the angle made by
PQ with Ox, and MN = QR = PQ cos PQR. (Fig. 5.)

9. If there be three points P, P ,

P"
joined by the right lines PP ,

PP" and P P",

on any line will be equal to
the projections of PP"

the sum of the projections of PP and P on that line. Let D, D P"

,

be the projections of the points P, P

D" on the line AB. ,
P"

Then D will either lie between D and or between D and D" D"

D In the one case DD" = DD + D

. and in the other DD"= D"

DD - D"D = in both cases the algebraic sum of DD and D D".

The projection is -f or according as the cosine of the angle above

is -f or .

In general if there be any number of points P, P etc., the pro , P",

jection of PP" on any line is equal to the sum of the projections of

PP P , etc., or, the projection of any one side of a closed po
P",

lygonal line on a straight line is equal to the sum of the projections

of the other sides on that line.

10. USEFUL PARTICULAR CASE.

The projection of th? radius vector of a point P on any line is OP
equal the sum of the projections on that line of tlie coordinates OM, MN,
NP of the point P. For is a OMNP
closed broken line, and
the projection of the side OP
on a straight line must be equal to
the sum of the projections of the sides OM, MN, and NP on that
line.
 AZOTES OA> SOLID GEOMETRY. 7

11. DISTANCE BETWEEN Two POINTS.

Let P and Q, whose rectangular coordinates are (x,y, z] and (x ,

y, ), be the two points. (Fig. 6.)

z
We have from the right parallelopipedon PMNRQ of which PQ
is the diagonal, PQ = PM +
2 2
MN 2
+ QN 2
. PM = x - x, MN
But
yy y NQ= z z.

- xj+ - z )\
Hence PQ =
2
(x (y -y Y + (*
If one of the points P be at the origin then x o, y = o, 2=0,
PQ = * + /
2 2 2 2
and +z" .

12. TO FIND THE RELATIONS BETWEEN THE COSINES OF THE ANGLES

WHICH A STRAIGHT LINE MAKES WITH THREE RECTANGULAR
AXES.
Take the line OP through the origin. Let OP r, the angle
POjt: = of, POy = /3, POz y, and x
{

, y ,
z the coordinates of P.

Then by Art. 1 1, r* = ,v
2
+y +
2
z -.

But, Art. 8, x = r cos a y = r cos ft ; ;

z = r cos y.
r = r (cos a -f cos ft
2
Hence 1 2 2
-\- cos 2 ^) or

cos
2
<?
+ cos
2
ft + cos 2 y = i. (i) A very im
portant relation.

Cos #, cos /?, cos y determine the direction of the line in rectan
gular coordinates, and are hence called the direction cosines of the line.
We usually call these cosines /, m and respectively. So the equa
tion (i) is usually written P m^ n? i, (i), and when + + = we wish to

speak of a line with reference to its direction, we may call it the line

(/, m, ii). Only two of the angles a, ft, y can be assumed at pleas
ure, for the third, y, will be given by the equation

cos y = A/ 1 cos
2
a cos
2
ft.

13. We can use these direction cosines also for determining the
position of any plane area with reference to three rectangular coordi
nate planes. For since any two planes make with each other the
is made
same angle which by two lines perpendicular to them respec
tively,the angles made by a plane with the rectangular coordinate

planes are the angles made by a perpendicular to the plane with the
coordinate axes respectively. Thus if be the perpendicular to a OP
plane, the angle made by the plane with the plane xy is the angle y ;

with xz the angle ft

is ;
and with yz is the angle a. So cos a, cos
ft, cos y, are called also the direction cosines of a plane. That is, the
g iVOTES ON SOLID GEOMETRY.
direction cosines of a plane with reference to rectangular coordinates are
the direction cosines of a line perpendicular to this plane.
The a + y=
2 2 2
14. relation cos cos ft+ cos i enables us to prove
an important property of the orthogonal projections of plane areas.
For let A be any plane area, and A x A A z its projections on the ,
y>

coordinate planes yz, zx and ;ry respectively. Then Art. 7, II., Az

= A cos a A y A cos ft A a A cos y.
;
= ;
=
Squaring and adding we have

A Z2 + A,,
2
+ = A (cos
A. 9 2 2
a + cos ft
2
-f cos y)
2

or A/+ A/ +A =A 3
9 2
.

That is, the square of any plane area is equal to the sum of the squares
of its projections on three planes at right angles to each other.

15. To FIND THE COSINE OF THE ANGLES BETWEEN Two LlNES IN

TERMS OF THEIR DIRECTION COSINES (cos cos ft, cos y) <*,

AND (cos a ,
cos ft ,
cos y ).

Draw OP, OQ through the origin parallel respectively to the given

lines. They will have the same direction cosines as the given lines,
and the angle POQ will be the angle between the given lines. (Fig. 7.)
Let POQ = 6, OP = r, OQ r ,
coordinates of P(x,y, z\ coor
dinates of Q (xy z }.

Now by Art. n,
PQ* = (
x -x>y + (y -yj + (Z
- Zj = X* + f + Z* + X* + /
+ 2ZZ }.
2}>/

And from triangle POQ,

PQ =
2 2 *
r -f r 2rr cos 0,

hence r 9
+ r 2 - 2rr cos B = x* +/+ z* + x* +/ 2
+ z*

-f 2yy + 2zz ).

But r*= .r
2
+j +
2
^
2
and r 2
= ,r
2
+y + a
^
2
.

Therefore
;
rr cos 6 = xx + yy + zz ,

x y y
or
a
cos B = x .+.--+-..
r
z z
r r
r r r

Hence cos 6 = cos cos <*

+ cos ^cos ft + cos y cos y (i)

which we write cos B // + ww + # .

(2)
 NOTES ON SOLID GEOMETRY. g

Cor. i. If the lines are perpendicular to each other cos 6 o or =

// + mm + nri = o (3). (3) is called the condition of perpendicu

larity of the two lines (/, m, ), (/ ,

m , ).

Cor. 2. From expression for cos 6 we find a convenient one

2
for sin 6.

Thus sin
2
0=1- (// + mm + nn Y = (/
2
+ m*+ n*} (l
z
+m*
+ n 2
) (// + mm + nn Y
whence sin 6
2
= (Irti
I
m}*-}- (ln
r
/ )
2
+ (mn m n)\ (4)

1 6. To express the distance between two points in terms of their

oblique coordinates.
Let P (xyz) and Q (xy z )
be the two points. (Fig. 8.)
The parallelopipedon MPQN is oblique. Let the angle xOy
= A, xOz
//, yOz
= =
v, and the angles made by PQ with the
axes respectively a, ft and y. Project the broken line PMNQ on
PQ. This projection is equal to TQ itself. Hence we will have
PQ = PM cos a + MN cos ft + NQ cos y. (a]

Now project the broken line PMNQ on the axes xyz respectively.
We obtain thus the three equations

PQ cos = PM + MN cos + NQ cos

a A.
/*
j

PQ cos ft
= PM cos A + MN + NQ cos v V (b)
PQ cos y PM cos /* + MN cos v + NQ )

Now multiply the of equations (3) by PM, the second by

first MN
and the third by NQ and add them taking (a) into account and we
have

PQ
2
= PM + MN 2 2
+ NQ + 2PM 2
. MN cos A + 2 PM NQ . cos
/* + 2MN,NQcos v (c)

or PQ 2
= (
X - XJ + (y -/) + (z
- z Y+ 2(x
- x ) (y -y )

cos /I + 2(x x )(z z ) cosfii + 2(y -y )(z z )

cos v. (5)

Cor. If one of the points as Q be at the origin then

PO 2
x* + y* + z* -f 2xy cos A+ 2x2 cos jn + 2zy cos v. (6)

17. Direction Ratios. In oblique coordinates the position of a line

PM MN NQ
PQ is determined by the ratios
-j-y ;
-^
-
;
~p^-,
and these we

call direction ratios. We may name these /, m, n respectively,

IO NOTES ON SOLID GEOMETRY.

taking care to note that \ve are using oblique coordinates and call

the line PQ, the line (/, ;//, n}. To find a relation among these
direction ratios, we divide equation (c) Art. 16, by PQ
2
. We thus
have
i = I
2
+ m* + ;/
2
+ 2//T2 cos A + 2ln cos /* + 2mn cos v, (7) the
desired relation.

1 8. The coordinates of the point (xyz) dividing in the ration : n

1

the distances between the two points (x y z ) are (x"y"z")

!
+ nx
mx"
= + ny + nz
z= -
my" mz"
x =.
Jy . .
(8)
m + n ,
7?i + n m + n

The proof of this is precisely the same as that for the correspond
ing theorem in Plane Coordinate Geometry.

19. POLAR COORDINATES.

Theposition of a point in space is also sometimes expressed by
the following polar coordinates :

The radius vector r, the angle PO0 OP =

6 which the radius =
vector makes with a and the angle CO^tr
fixed axis Oz, cp which
the

projection OC of the radius vector on a p ane yQx perpendicular to

O0 makes with the fixed line Ox in that plane. (Fig. 9.)
We have OC = r sin
the formulae for transforming from
6. Hence
rectangular to these polar coordinates are

x r sin 8 cos <p

}

Y r sin 8 sin <p\ (9)

z = r cos 8 j

2
and these give r1 x^ +.V +

tan cp =
(10).

a
cos 8 = *
= z

Conceive a sphere described from the centre O, with a radius a =

and let this represent the earth. Then, if the plane zOx be the
plane of the first meridian and the axis of z the axis of the earth,

Q latitude, cp
= longitude of a point on the earth s sur

face.
 NOTES ON SOLID GEOMETRY. Ir

20. Distance between two points in space in polar coordinates.

r
Let P be (r ,
6 , q> }
and Q (r, 6, <p). Project PQ on the plane xy,
MN is this projection, draw OM and ON the projections of OP and
OQ respectively on that plane. Through P draw PR parallel to MN,
then PR = MN. (Fig. 10.)
And we have

PQ = PR + RQ
2 2 2
= MN + 2
(QN - RN)
2
.

But in triangle MON

MN == OM + ON - 2OM ON cos MON,
2 8 2
.

or MN = r2 2
sin
2
&+ 8
r sin 2
6 2rr sin sin d cos (tp <p ).

Moreover QN r cos and RN = PM = r cos 6 .

Hence =
PQ r
2 2
sin
2
+ r* sin
2 - 2rr sin sin cos (q> <p )

+ (r cos
- r cos )*
or

PQ
2
=r 2
+r 2 - 2 rr (cos 6>cos^ + sin sin 6 cos (<p
- q> )). (n)
 CHAPTER II.

INTERPRTTATION OF EQUATIONS.

TRIPLANAR COORDINATES.

21. LET us take F (x y, z] =

o, that is any single equation con
y

taining three variables x, and z. This may be considered as a

y
relation which enables us to determine any one of the variables when
the other two are given. Let these be x and j . So the equation
may be written
*=/(*, jOi
in which we may attribute arbitrary and independent values to x and
y. And to every pair of such values there is a determinate point in the
plane xy ;
and if through each of these points we draw a line parallel
to the axis of z, and take on lengths equal to the values of z given
it

by the equation, it is clear that in this way we will get a series of

points the locus of which is a surface, .and not a solid since we take
determinate lengths on each of the lines drawn parallel to z. Hence
F (x, y, z)
=o represents a surface in triplanar coordinates.

22. If the equation contains only two variables as F (x,y) =o then

it represents a cylindrical surface.
For F (x, y] =
o is satisfied by certain values of x and y inde
pendently of 0, and x and y are no longer arbitrary but one is given
in terms of the other to each pair of values corresponds a point in
;

the plane xy, and the locus of these points is a curve in that plane.
If through each point in this curve we draw a coordinate parallel to

2, every point in that coordinate has the same coordinates x andj/ as

the point in which it meets the plane xy. Hence F (x,y) = o repre
sents a surface which is the locus of straight lines drawn through

points of the curve F(.r, v) = o in the plane xy and parallel to the

12
 NOTES ON SOLID GEOMETRY. 13

axis of z. This locus called a cylindrical surface

is either what is

with axis parallel to z or a plane parallel to the axis of z according

as the equation F (x,y) o in the plane xy represents a curve or a
straight line.
For example, x
1
y* r
2
=
o in rectangular coordinates is a
4-

right cylinder wiih circular base in plane xy (since jv

a
r* is a +y =
circle in plane xy) and its axis coincident with the axis of z.
And ax + by c =
o is a plane parallel to the axis of g, intersect
ing the plane xy in the line ax + by c. =
Similarly F (x, z) =
o represents either a cylindrical surface with
axis parallel tojy or a plane parallel .
to_>

F (y, z) =
o represents either a cylindrical surface with axis parallel
to the axis of x or a plane parallel to this axis.

23. An
equation containing a single variable represents a plane or
planes parallel to one of the coordinate planes.
Thus x =
a represents a plane parallel to the planejyz.
And as _/"(.#) = o when solved will give a determinate number of
values of x, as x = a, x x= I, c, etc., so it
represents several
planes parallel to the coordinate planers.
Thus also F(jy) =
o represents a number of planes parallel to the
plane xz.
And F (z)
= o, a number of planes parallel to xy.

24. Thus we see that in all cases when a single equation is inter

preted it
represents a surface of some kind or other.
The apparent exceptions to this are those single equations which
from their nature can only be satisfied when several equations which
must exist simultaneously are satisfied. As for example
(x of + (y b)* + (z c)
z
= o. This equation can only be
satisfied when (x a)*
= o, (y b}*
= o, (z c}
z
= o, or x= a,

y b, z = c.

Now these represent three planes, but being simultaneous they

represent the point a, b, c.

So also (x a)
z
-f ( y b}*
= o is only satisfied by x = a, y = b,
and hence though x = a is a plane, and y = b is a plane, the two
together must represent a line common to both of these planes, that
is their line of interseciion, which must be parallel to z.

25. In general two simultaneous equations as

f(*,y, *)
= F (x,y, z) =o
14 NOTES ON SOLID GEOMETRY.

represent a curve or curves, the intersections of the two surfaces

represented by the two equations.

Thus _ 7 r taken simultaneously we have seen

represent a straight

line parallel to the axis of z, the intersection of these two planes.

F (x) =o )

_ represent a number of straight lines parallel to the

p) \
f

axis of the intersections of the several planes parallel

z,
respectively to
the planes yz and xz.

F (x) =o )
re P resent a numbe r of straight lines parallel to the axis
p/j) of
of y, e;c.

represent the curves of intersection of the two cylin-

P (.v, z)
= o
f.

j
*

ders F (x, y) = o and F (x, z)

= o, e c., etc.

26. Three simultaneous equations

F (x, y, z}=o\ F (x, j>)

= o\
as f(x, y, z)
=o I or F (A-, z)
= o I etc .,

represent points in space or the intersections of the lines of intersec

tion of the surfaces.
The simplest case is,

y ^ >

representing the point (a, b, c],

z = c]
So also
2z \

x +y = 2z v.
represent points which can be found by

solving the three equations which themselves represent different sur

faces.

Interpretation of Polar Equations.

27. i. r =a
represents a sphere having the pole for its centre.
Hence the equation (r)
=
o which gives values for r as r a,

r = b, r c, etc., represents a series of concentric spheres about the

pole as centre.
 NOTES ON SOLID GEOMETRY. jcj

2. 6 =
a represents a cone of revolution about the axis of z with
itsvertex at the origin of which the vertical angle is equal to 2a.
Hence the equation F (6) o giving values 6 = a, 6 ft, etc.,
= =
represents a series of cones about the
axis of z having the origin for

a common vertex.

3. cp =
(3 represents
a plane containing the axis of z whose line
of intersection with the plane xy makes an angle /3 with the axis of
x. Hence the equation F (cp) o which gives values q) (3, cp
=
= ft , etc., represents several planes containing the axis of z inclined
to the plane zOx at angles ft, ft etc. ,

4. If the equation involve only rand 6 as F (r, 6) o, since =

F (r, 6)
=o gives the same relation between r and 6 for any value
of cp,
ii
gives the same curve in any one of the planes determined by
assigning values to (p. Hence it represents a surface of revolution
traced by this curve revolving about the axis of z.

Example, r =
a cos 6 is the equation of a circle in the plane
xz, or in any plane containing the axis of z. Hence r a cos 6

represen s a sphere described by revolving this circle about the axis

of*.

5. If the equation be F(<p, 0) o for every value of cp there

are one or more values of 6 corresponding to which lines through
the po e may be drawn, and as cp changes or the plane fixed by it
containing Oz revolves, these lines take new positions in each
new position plane, of the and thus generate conical surfaces
(a conical surface being any surface generated by a straight line

moving in any manner about a fixed straight line which it inter

sects. )

6. If the equation be F(r, (p)

= o, for every value of (p there are
one or more values of r, thus giving several concentric circles about
the pole in the plane determined by the assigned value of cp. As (p
changes, or the plane through Oz revolves these values of r change,
and the concentric circles vary in magnitude. The equation thus

represents a surface generated by circles having their centres at the

pole, which vary in magnitude as their planes revolve about the axis
of z which they all contain.
7. If the equation be F (r, 6, cp)
=
o, it represents a surface in

general. For if we
assign a value to as cp ft, then
F (r, 8, ft)
<p
=
= o will represent a curve in the plane (p ft. And as cp changes =
or the plane revolves about Oz this curve changes, and the equation
will represent the surface containing all these curves.
!6 NOTES ON SOLID GEOMETRY.

28. Two simultaneous equations in polar coordinates represent a

line, or lines the intersections of two surfaces. And three simulta
neous equations represent a point or points the intersections of three
surfaces.
Thus
r =a
j
6 a v taken simultaneously represent points determined

9 = ft)
by the intersection of a sphere, cone and plane.
 CHAPTER III.

EQUATION OF A PLANE.

COORDINATES OBLIQUE OR RECTANGULAR.

To find equation of a plane in terms of the perpendicular from the

29.
origin and its direction cosines.

Let OD p be the perpendicular from the origin on the plane,

and let it make with the axes O,r, Oy and Qz the angles a, ft and y
respectively. Let OP be the radius vector of any point P of the

plane ; OM, MN and NP the coordinates of P. (Fig. n.)

The projection of OM + MN + NP on OD is equal to the pro
jection of OP on OD.
The projection of OP on OD is OD itself, and the projection of
OM + MN + NP on OD is x cos a +y cos ft + z cos y.

Hence we have x cos a +y cos ft + z cos y p. (12)

30. To find the equation of a plane in terms of its intercepts on the

coordinate axes (coordinates oblique or rectangular).

Let the intercepts be OA a, OB = b, OC = c. The equation

(12) may be written
x y z

p sec a p sec ft p sec y

But since ODA, ODB and ODC are right-angled triangles, we have

p sec a = OA a, p sec ft = OB
= b, p sec y = OC = c.
Therefore the equation becomes

X V Z

the equation of the plane in terms of its intercepts.

 1 8 NOTES ON SOLID GEOMETRY.

31. Any equation Ax + By + Cz =D (14) of the first degree in x,

y and z is the equation of a plane*
For we may write (14)

x j^ _z_
Jl _D/
X J>/

"B" ~C~

D D D
And putting -^= a, -g-
= ^ --= A We have the form (13).

Hence (14) is the equation of a plane in

oblique or rectangular
coordinates.
Hence to find the intercepts of a
plane given by its equation on
the coordinate axes, we either put it in the form (13) or simply
raakej>
= o and z = o to find intercept on x z = o and x = o to ;

find intercept onj/ ;

x= o andjy o to find intercept on z.

"Example. Find the intercepts of the plane 2x + $y 52 = 60.

32. It is useful often to reduce the equation AJC + By + Cz =D

to the form x cos a +y cos ft + z cos y =/ in rectangular coordi
nates. We derive a rule for this.

Since both of these equations are to represent the same plane, we

have

cos a
_ cos (3 _ cos y _ p _ Vcos + cos (3 + cos 2
a.
2 2
y
A B C D yT7 A + B + C 2 2
""

Hence cos <*

=

+ B +C2 2
-v/A
2
+B +C 2

A/A + B +C
2 2 !
+ B +C 2

= (I5)
-vWWc*
it is in the perpendicular form (12).
 NOTES ON SOLID GEOMETRY. ig

Hence the Rule: If we each term of the equation Ax + By

divide
= D, by the square root
of sum the
of the squares of the coefficients
ofx, y and z, the new coefficients will be the direction cosines of the per

to the plane from the origin, and the absolute term will be the
pendicular
length of this perpendicular. Give the radical the sign of D.

Example. Find the direction cosines of the plane 2x + $y 40

= 6 and the length of the perpendicular from the origin.

Result.

cos a =
A/4 +
2

9
===
+ 16 v
7=,
2

29
cos ft
=
V
3
T=,
2 9
cos y =, V 29
4

P =
V 29
33. To find the angle between two planes (coordinates rectangu
lar).
If the planes are in the form

x cos a + y cos ft + z cos y =p

x cos a +y cos ft + z cos y p ,

then since this angle is equal to the angle of two perpendiculars from
origin on the planes the cosine will be (Art. 15) cos cos a cos a V=
1

+ cos /3 cos /3 -f cos y cos y .

If they are in the form

A.v + By + =DCs
A .v + B>
+ Cz = D .

Then cos a = ==. r, cos /?

=
B2 4- C2 A/A
2
+ B + C
2 2

C
cos =

= A B
COS flf
~, COS
al
P =
A/A
2
+B + /f

cos y = =
AA + BB + CC
And cos V
AA 2
+ B + C VA + B + C
2 /f 2 2
2Q NOTES ON SOLID GEOMETRY.

From this

B2 + C + B +C -(AA + BB + CC
2 2 2 2 2 2
. Q _ (A -f ) (A ) )

(A
2
+ B* + C )(A 2 2
+B 2
+ 2
C"
)

d . v- (AB -A B) 2
+(AC -A C) 2
+(BC -B C) 2

(A
2
+ B + C ) (A 8
+ B +C a 2
)

Cor. i. If the planes are perpendicular to each other, then cos V=o.
. . AA + BB + CC = o (18) is the condition of perpendicularity of the

planes.
Cor. 2. If the planes are parallel sin V= o. Hence

(AB + (AC - A C) + (BC - B C) = o

- A B) 2 2 2

AB - A B = o AC - A C = o BC - B C = o
or

ABCC
A7== B 7=
or the condition that the two planes shall be parallel, is that the coefficients

ofx,y and z in the two equations shall be proportional.

Ex. i. Find the angle between the planes

x+ 2y + 32 = 5 and $x 4y 4- z = 10.

2. Show that the planes

x + 3y 5Z 20 anc* 2X + V + z I0 are perpen

dicular to each other.

3. Write the equation representing planes parallel to the plane 3*

+ 2y 6z = ii.

34. To find the expression for the distance from a point P (x y z )

to a

plane (coordinates rectangular).

i. Let the equation of the plane be of the form

x cos a +y cos ft + z cos y = / when p OD.

Pass a plane through P parallel to the given plane,

and produce
OD to meet it in D .

The equation of this plane will be

x cos a +y cos ft + z cos y =p when OD =/ .

Now let PM be the perpendicular from P on the given plane.

Then PM = OD - OD = p - p. 1
 NOTES ON SOLID GEOMETRY. 21

Hence PM = x cos a 4- y cos /3 + z cos y p.

And 1

(x cos a + y cos /? + cos y p) (20) is the expression

for the perpendicular from the point xy z on the plane x cos or + _y
cos ft + cos y =p, the sign being + or according as P is or is

not on the side of the plane remote from the origin.

2. Let the plane be in the form Kx + By + Cz = D.

A
Then cos a = - etc., etc. (15) Art. 32.
V A + B + C* 2 s

Hence the expression

(y cos -FJF cos /3 + z cos y p] becomes

-D

Ex. Find the length of perpendicular from the point (3, 2, i) on

the plane
3Jt- + 4_y 6z = 24.

Result. p
9 4- 8
-
A/9+ 16+
- 6

36
24
-
13

35. The equation of the plane in the form x cos a + y cos (3 4-

cos y =p may be used to demonstrate the following theorem in
projec ions.
7% volume of the tetrahedron which has the origin for its vertex and
the triangle ABCy2?r its base is
equal to the three pyramids which have any
point (x, y, z) in the plane ABC for their common vertex and for bases
the projections of the area ABC on the three rectangular coordinate planes
respectively.
For let A be the area of the triangle ABC and
x cos a +y cos /3 + z cos y =p
the equation of its plane.

Multiply this equation by A.

Then A cos a x . -f A cos /3 .
y+ A cos y . z = Ap
or -JA cos a x + ^A
. cos ft .y + -JA cos y. s

But A cos a, A
cos /3, A cos y, are the projections of on the A
planes yz, xz, and xy respectively, and x, y and z are the altitudes of
the tetrahedrons which have these projections as bases and the point

(x, y, z) as common vertex, and is the volume of the pyramid -JA/>

V
22 NOTES ON SOLID GEOMETRY.

which has the origin for vertex and A for base. Hence the theorem
is true.

Calling these projections Ax Ay

, ,
and A x we may write the equa
,

tion of the plane A^v +A y y + A^ = 3V. (22)

36. To find the polar equation of a plane .

Let OP = r, POS = 6, P OM = cp be the polar coordinates of a

point P of the plane. (Fig. 12.)
Let OD =
a =
perpendicular on plane ; angle DOS a, D OM
= jff,
and POD = GO.

Then yyp - = cos POD = cos GO, or = cos GO. Now in order to

express GJ in polar coordinates conceive a sphere about O as centre

with OP =
r as radius. Prolong to D" on the sphere. Draw OD
the arcs of great circles SPP , SD"D MP D
,
and D"P.

The triangle SD"P has for its sides SD" = a, SP = 6, D"P = GJ

and angle D"SP = D OP = /3 <p.

But

cos D"P = cos SD" cos SP + sin SD" sin SP cos D"SP.

Or
cos GO = cos a cos 6 + sin a sin 6 cos (fi cp).

Therefore = cos a cos 6 + sin a sin 6 cos (f3 cp) (23) is the

polar equation of the plane.

37. The general equation of the plane Ax + By 4- Cz =D may

be reduced to the form
Ax 4 By
f

+ C2= i
(24) by dividing by the absolute term D.
And also to the form

z mx + ny+ c (25) by dividing by C transposing and putting

= m, = n and -^ = c. These two forms are very useful in

V^/ V^ v_/

the solution of problems and in finding the equations of the plane

under given conditions.

Plane tinder Given Conditions.

i. The equation of a plane through the origin will be of the

38.
Ax + By + Cz = o, for ihe equation must be satisfied
general form
by ^ = 0,^ = and 2 = 0.
2. The equation of a plane which contains the axis of z is of the
 NOTES SOLID GEOMETRY,

form Ax + By = o a plane containing the axis of y

;
is Ax + Gar

o ;
one containing the axis of x is By + Cz o. =
3. The equation of a plane parallel to the axis of z is Ax + By
= D; of one parallel to the axis ofjy is AJI* + Cz one parallel =D ;

to the axis of x is By + Cz D. =
4. The equation of a plane parallel to the plane j/s is A^t: =D ;

parallel to ^0 is By = D ; parallel to xy is Cs = D.
These equations we have had already in the forms x= a, y 3,

39. To find of a plane containing a given point (a, b, c)

the equation

and parallel a given plane Ax -f By + Cz

to D. i
( )
=
First, since the required plane is to be parallel to (i ) it may be writ
ten A* + Bj/ + (2) when G&=D
is undetermined. Secondly, the D
coordinates (a, b, c) must satisfy (2). Therefore Aa + ~B& + Cc =D .

Hence by subtraction we eliminate D and obtain

Ax + By + C0 = Aa + B3 + O (26)
the required equation.

Example. Find the equation of the plane passing through the

point (i, 2, 4) parallel to the plane 2x + 4y 32 6. =
40. To find the equation of a plane passing through three given points

(x , y ,
z ), (x", y", z")
and (x y z ", ",
").

Let the equation of the plane be of the form Ax + By + Cz = i,

A, B and C to be determined by the given conditions.

Since the plane is to contain each of the points, we must have

Ax + B

Hence
X, I, ^

JP", i, z

I,/", x i, 2
B=
",

C=

Substituting these values in the equation Ax + By + Cz = i

 NOTES ON SOLID GEOMETRY.
we have
#, I, *

I,/ , *, I, *
r
x+ x"
, i,
(27)
But from plane coordinate geometry the coefficients of x, y and z
in these equations are the double areas of
triangles in the planes jy0,
xz and xy respectively. Moreover these triangles are the projections
of the triangle of the three given points, on these planes. Hence
comparing this equation wi h the equation (22)
,z = 3V

we see that x ,y ,ss = 6V. That is = 6 times the volume of the

x">y",
2

pyramid which has the origin for vertex and the triangle of the three
given points for base.This equation fully written out is

^(yv--yv )+^ (yv-y3 )+y (jV -yv)-6v, (

2 8)

41. To find the equation of the planes which contain the line of intersec
tion + By + Cz = D and Ax 4- By + Cz = D
of the two planes Ax .

ThisequationisA# + By + Gs D + K(A jc + + C 0-D )=o(29)

/

B>

when K is arbitrary. plane when K takes a

For this represents a

particular value and when A^ + By + C-s: it D = o and

is satisfied

K x 4- B^ + C z D = o are satisfied simultaneously. Hence it

is a plane containing their line of intersection. Hence as K is arbi
trary it
(24) represents the planes containing the line of intersection
of the two given planes.

42. When the identity KU + K^Uj + KaUs = o (30) exists between

the equations U=o, U 2 o, U = o of three planes, then these planes
3

intersect each other in one and the same straight line. This is an
easy corollary of Article 41. Also when the equation of the first
degree in x, y and z contains a single arbitrary constant all the
planes which it expresses by assigning particular values to this con
stant intersect each other in one and the same straight line. This
line of intersection may be at infinity and then the planes are all

parallel.

Example I. The planes represented by the equation 6x+T&y + 2z

=. 3 all contain the line of intersection of the two
(M arbitrary)
planes 6x+ 22 3
= and y o.
 NOTES ON SOLID GEOMETRY. 25

Example 2. The planes represented by 2x + ^y 42 = n (n

arbitrary) are parallel.

Example. The planes $x + \v + 60 = 2 }

X+ 2V + 3 I ?

4.v + 4T + &z = 2 J

intersect in one and the same straight line because

2 =o
is an identity.

43. Wfo between the equations of four planes in any form U o, U,

= o, U U o the identity
2 o, 3

KU + K U + K U -hK U = o
1 1 2 a 3 3 (31) ^/j/j, then these four planes

intersect each other in one and the sam* point. For then any coor
dinates which satisfy the first three U= o, U, = o and U 2 = o will
satisfy the fourth 3 o. U =
44. Example
i. Find the equation of the plane passing through
the origin and containing the line of intersection of the two planes
Ax + By + C0 = i and A x + + C z = i. B>

First we have Ax + By + Cz i + K (A .r + B> + Cz 1)= o

for all the planes containing the line of intersection of the two given
planes. But as the required plane must contain the origin, the
equation must be satisfied by (o, o, o). Hence we have i K=o.
.-. K=- i.

The required equation is therefore

Ax + By -f Cz i
(A x + B>
+ Cz i)
= o

or (A
- A x )
4-
(B
- B )^ + (C - C )
z = o.

Ex. On the three axes of x, y and z take OA = a, OB

2. b,

OC = and construct on these a parallelopipedon having MP as the

c

edge opposite parallel to OC, and AR in the plane xz the edge

opposite and parallel to BN.
Find the equation to the plane containing the three points M, N
and R.
\? V
Now NR is the line of intersection of the two planes +-7- = i and
a b

-= i. Hence the plane containing this line must be of the form

26 A OTES ON SOLID GEOMETRY.

OC V
"*~
^ I
-- l
)
- To determine K \ve impose the
condition that this plane shall pass through the po nt M (a, b, o).

+K = o. K=
Hence we have
at
+ -r i
(
\c
i
J
) . . i

Therefore the required plane is

x v z x v z
--t-V
a b
H---- c
i o or -- (-^-_j-__
a b c
2.

Ex. 3. Find in like manner the equation of the plane containing

the points P, B, and C, in the same figure.

Result,

45. If two given planes be in the normal form as

x cos a +>
cos ft +z cos y=p and.v cos a +y cos /3 + zcos y = p .

The plane containing their line of intersection is

x cos a+y cos /3 +z cos y p-\-K (x cos a +y cos ft +z cos y

-/)=o
And if K = i the equation becomes
x cos a -\-y cos /3 +z cos y p (x cos a +y cos ft +z cos ;/

-/> )= o

which represents ihe two plane bisectors of the supplementary angles

made by the given planes.
That is to find the equations to the plane bisectors of the supplementary

angles made by two given planes, put their equations in the normal form
and then add and subtract them.
Example. Find the two planes which bisect the supplementary
angles made by the planes 2.v + 3.y Vz = 5 and 3^ + 4^23 =-- 4-

Result,
.
.. = .

A/14 V29
Remark. If we place A = x cos (Y + y crs
ft + z cos y p and
A =x cos a + v cos /? + cos y />
.

Then A A= o isthe plane bisector of one of the angles be

tween the planes A and A and A -f A o is that of the supple =
mentary angle.
46. The three planes which bisect the diedral angles of a triedral have.
a common line of intersection. Let A = o, A = o and = o be three A"
 NOTES ON SOLID GEOMETRY.

planes in the normal form, and let the origin be within the triedral angle
formed by the three of which P their point of intersection is the vertex.
Then the plane bisectors of the angles made by these planes is
A A o, A o, A A"o. And as these when added A" =
together vanish simultaneously, it follows that these three planes
have a common line of intersection.
We can give this theorem another form by conceiving a sphere to be
described about the vertex of the triangular pyramid as a centre. The
three planes o, A A= =
o, A"=o cut the surface of the sphere in
arcs of great circles which form a spherical triangle and the three

planes A A =o, A" A = o and A A" =o cut the sphere in

three arcs of great circles which bisect the angles of this spherical

triangle and their common line of intersection pierces the sphere in

the common intersection of these arcs. Hence the above demon
strates the following theorem, namely, The arcs of great circles which
bisect the angles of a spherical triangle cut each other in the same point
(the pole of the inscribed circle of the triangle).

47. To find the point of intersection of the planes AJC + By + Cs

= D, Ax + B>
+ Cz =D ,
A",r + B >
+ C"z = D".

We have by elimination

D, B, C A, D, C A, B, D
D ,
B ,
C A ,
D ,
C A ,
B ,
D
D". C" C" D"
B", A", D", A", B",

(32)
A, B, C [A,
B, C A, B, C
A ,
B ,
C A ,
B ,
C A ,
B ,
C
C" C" C"
A", B", A", B", A", B",

Hence the condition that one of these shall be parallel to the line
of intersection of the other two, or that the planes shall not meet in
a point, is

A, B, C
A ,
B , C
A", B",
C" = o, that is

A(B C"-B"C ) + A (B"C

- BC") + A"(BC
- B C) = o.
47. The condition that four planes

Ax -fBy +Cz +D = o 1

Ax + + Cz + D =o B>

A"x +B >
+C"z +D" =o
A "x+B "y +C "z + ~D
"

= o shall meet in a point is

28 NOTES ON SOLID GEOMETRY.

A, B, C, D
A ,
B ,
C , D
A", B", C", D"

|
A ",
B ", C",
D "

= o. (33)

49. We have seen that the equations of two planes Ajr

D = o and A. x+ Bjy + Cz D= o added together one or both
of them multiplied by any number give the equation of a plane
which contains the two given planes.
line of intersection of the If

we combine these two equations so as to eliminate x we shall obtain

a plane parallel to the axis of A\ containing this line of intersection.
If we eliminate y we obtain a plane parallel to the axis of y contain

ing the same line ;

and finally if we eliminate z we obtain a plane

parallel to the axis of z containing the same line.

 CHAPTER IV.

THE STRAIGHT LINE.

50. The equations of any two planes taken simultaneously represent

their line of intersection.
re P resent a strai S ht lme the c -

ordinates of every point of which will satisfy the two equations.

If we eliminate alternately x and y between these equations we
obtain equations of the form
x m p\ , x
^ to the
y ~ nz + q
planes parallel respectively
\

axes Oy and Ox which represent the same straight line as equations

(35). These non-symmetrical forms (35) are very useful. The

planes x =. mz + a, y = nz + b are called the projecting planes of the
line on the planes of xz and jar, and these equations are also the
equations of the projections of the line on those planes respectively.

If we eliminate z we get - ~= -- or Jy = x+g -- p the equa-

n m m m
tion of the projection of the line on the coordinate plane xy.
The equations (35) of the straight line contain four arbitrary con
stants, m, to which we can give proper significance by com
n, p, q,

paring these equations with the equation y mx + b in plane coor

dinate geometry.
The equations (35) may be thrown in the form

^_^ =z _^ = ,
m n i

which gives us an easy choice of fixing the line by the equations

of any two of its projecting planes.

51. To find the equations of a straight line in terms of its direction cosines
and the coordinates a, b, c of a point on the line : (axis rectangular.}
29
30
A T
OTES ON SOLID GEOMETRY.

Let a, ft, y be the angles made by the line with the coordinate axes
respectively. Let / be the portion of the line between
any point
(x, y, z) on the line and the point (a, b, c). Then / cos a = a- a ;

/ cos ft yb ;
/ cos y zc
and eliminating / we have

(37)
cos ex cos ft cosy
This form (37) of the equation of a straight line is
symmetrical
and therefore very useful.
is It contains six constants but in
reality
2
only four independent constants, since the relation cos o cos 2 ft +
2
4- cos y i =
holds, and of the three a, b, c one of them may be
assumed at will, leaving only two independent.

We have seen that the equation (35) may be thrown into the form
(37). So also (37) may be thrown into the form (35) by finding
from them expressions forjy and x in terms of z.

52. To find the direction cosines of any straight line given by its equations.
If the equations be in the form

x
=
a
= vbzc r^ .
T
L,
HT
M and
i XT
N are proportional
i

to the

direction cosines of the line.

So that we have
a + cos 2
~ a + cos
2
cos _ cos ft _ cos y __ -\Xcos- ft y
L ~M~~ N
Hence
N

Hence to find the direction cosines of any straight line

A.r + By + Cz = D
MX 4- + Cz=D
B>

we throw the equations into theform

.va _ yb _ z c
L ~N~ "IT"

cosines as above
by eliminating y and x, and. then write out the direction
equal to each denominator divided by the square root of the sum of the
squares of all three.
 NOTES ON SOLID GEOMETRY.

Thus to find the direction cosines of the line

y = nz + q
1
\

X p V Q Z
we
"

write it
m n
Hence
m
cos
=^+*+r. cos *>

(39)
Ex. i. Find the direction cosines of the lines

x 5 y~ 2 z +

,, _
= 24 . v

., T ^ 3^-1 j
\ >
.

>

3x 4 y+2z - 10 U;

53. To find the cosine of the angle between two lines given by the equa
tions

xa yb z c
and
,
x
ri
a
= vb
:
zc
L M N L M N
We have shown (Art. 15)

cos V= cos a cos a -f cos /3 cos /? + cos ^ cos y.

LL + MM + NN
Hence cos V= ~

*+ N VL *+ s
M"
2
+N 2

TT.U i- u
If the lines be m i r .\ mz-\-p] x = mz-\-p |

the form
y nz + q } y = n z+q .
, \
\

Then cos V = wm+nn +i

^= =..
2/ 2 ,
(41)

Ex. i. Find the cosine of the angle between the lines

X** 2Z +6

Ex. 2. Find the cosine of the angle between the lines

2
xy z =4) x+y + z =2p
32 NOTES ON SOLID GEOMETRY.

These equations may be put in the forms

= = _ ___ =
4-3 _
(I )
and
-5

. . cos V=
,T
-+
5 12

V26X38
54. The condition of perpendicularity of two lines given by the

equations in last article is LL + MM + NN ^ o.

(42)
The condition that they shall be parallel (see Art. 15)

is
(LM
/

-L M) + (LN -L N)
2 / 2
+(MN
/
-M N) 2
i=o

or -=^=-(43).
1 A IV J. IN
These two conditions when the lines are in the

x = mz+p )
x = m z+p \
forms
y nz +q \
y n z+q f

become mm + nn +i =. o,
(44) and m == m ,
n = ri (45) respec
tively.

55. To find the condition of the intersection of two lines

x = m,+p }

V nz +q )

This derived by eliminating x, y and z from the four equations.

is

Subtracting the third from the first we have o (m m }z +pf> .

_P~P ~~
Similarly from the second and fourth
/0 z z
n r^_ n
^ ,

two values of z are equal. There

and since the

fore we have -^^=

m m
lines intersect these

^
n n
.
(46)

~
Ex. Find / so that the lines i
^
tersect.

If the two lines are in the form

xa yb zc , x a __ y
=
b
_z c

~TT TT~ "N~

(I)
~T7" ~W~ N r
 NOTES ON SOLID GEOMETRY. 33

the elimination of x, y and z can be effected more readily by writ

ing (i)= K and (2)

=K .

.-. x-a ^LK )

=I*K ,
LK.
r Vif
=L . <I

x a K j

Similarly b- b MK MK
f-^
r
^ N K -NK.
Therefore eliminating K and K we have

L, -L a-a ,

M, -M - ,

(
N, -N ,
r-/ =ro
or

(47)

The Straight Line under Given Conditions.

56. The equations of a straight line parallel to one of the coordi

nate planes as xy are z c, v mx -\-p. = =
The equations of a straight line parallel to one of the coordinate
x=
axes as z. are
y
5 7. 7 o find the equations of a straight line, passing through a given point.
If (x, y, z )
is the point

we have seen the equation is

xx =-vy
== = z -^ z ,
(4)
ft
.

or if the equations are in

, r
the form
x = mz+p)* then
xx m(zz\} , , TT
Hence
.,.
if
.

the
y = nz + q = n(zz
\ } ,/ \ v(49).
v
j ^r-y ) j

equations of a straight line contain only two arbitrary constants, all

the lines obtained by assigning values to these arbitraries pass through
a single point.

58. To find the equations of a straight line passing through two given
points (x , y ,
z ) (x", y", z") using (48) we have

~ "

= , or dividing (48) by this to eliminate

.L IVl JN

3*
 34 NOTES ON SOLID GEOMETRY.

L, M, and N we have
xx yy zz .

:?r=P=-77=7w- <5>

If one of the points as (x",y, z")

be at the origin then the equa
tions become

59. To find the equation of a straight line passing through a given

point (x ,y, z }
and parallel to a given straight line

xa yb z c

L M
From the first condition we must have -W~ ~ =
-L M /

N -,

, ,
and Irom the second condition
L M N
=-r -.=-:.
LT MT NTr
-=r .

Hence the required equation is

x x __ yy _ zz 1

w2
.

L M ~~N )

If the equation passing through the point x ,y z be ,

of the form
A--JV = m (z-z ) \
x == mz+p
~ n (z z /! an d the given line be
vy i //
) ) y nz + q.

Then = n and / = /, and the line will be

To find
^2$$Y
60. the equations of a straight line passing through a given
r

point x , j ,
z and perpendicular to and intersecting a given right line

x a _ y b
_z c
I m n

The required line by the first condition will be of the form

where L, M, and N are to be determined by the conditions

= o (Art. 54)
 NOTES ON SOLID GEOMETRY.

and

61. Ex. i. Find the equation of the line joining the points (b, c, a)
and (a, c, It)
and show that it is perpendicular to the line joining the
origin and the point midway between these two points ;
and that it

is also perpendicular to the lines x =y = z and

a
=-=
b
-.
c

Ex. 2. The
straight which join the middle points of thelines

opposite sides of a tetrahedron all pass through one point.

Take O one of the vertices as origin and OA, OB, OC as the

axes of x, j ,
s.

Let M, M
be the middle points of BC, AC and
,
M"
respec OC
tively, N, N
the middle points of the edges OA,
,
N" and AB OB
opposite to those respectively. Then to find the equations of the
lines MN, MN ,
M"N".

We apply the equation ;

-= , ,, , -77- to the points
x x y y z z

(M, N) (M ,
N ) (M", N") respectively.
Let OA = 20, ;
OB = 2b ;
OC = 2c.

Then M is (o, b, c) and N is

(a, o, o).

Hence the equation of MN is

J
~ ^~^
(l
_.^_ ~^ /
T
\

a~~ b c
^

Similarly the equation of MN is

x __yb _z .

a b c

And the equation of M"N" is

X V Z C

(i) and (2) give x , y =, z = and these values satisfy (3).

Consequently these lines pass through the point (-, ,

-
-).
36 A 07^S OA SOLID GEOMETRY.

Straight Line, and Plane.

62. To find the conditions that a line shall be perpendicular to a plane

its
given by equation.
If the plane be of the form .v cos a+y
cos fi + z cos y (i) =/
we know that cos <v,
cos /?, cos y are the direction cosines of the

perpendicular from the origin on the plane.

And the equation of this perpendicular will be

cos a cos // cos y

If any plane A^ + Br + Cs= D be parallel to the plane (r) we must
have
A B C
-
~~
cos a cos// cos y
C
and if the line -~= -
be parallel to the line

x v z
-, we must have
cos at cos ft cos y
L M N
cos a cos/3 cosy

Hence the conditions that the line =

M = -^ shall be

perpendicular to the plane Ax + Ey + Cz = D will be

ABC
== (54)

If the line be in the form

)
v we write
.

it-
xp ="
\~q -
= -z
)
m p i

A
= B
= C A = ;;/C)
And the conditions are - -
or \ (55)
m n i
_,
B = C ,

The equation of a line passing through the point .r ,

r ,
z
1
and
perpendicular to the plane Ax + fy + Cz =D will then be

xx yv zz
A B
 NOTES ON SOLID GEOMETRY. 37

or in the unsymmetrical form

*-* = ~ (z-z)

yy = ~(z-z).
Ex. Find the equation of a line passing through the point (i, 2,

3) and perpendicular to the plane $x+2j \z = 5.

63. To find the condition that a straight line shall be parallel to a

given plane. Let the plane be Ax + l$y + Cz =D and the line of

^- ~
x a y b z c
the form .

Now if this line is parallel to the plane it will be perpendicular to

the normal to the plane. Hence the required condition will be

AL + BM+CN = .
(56)

64. To find the conditions that a straight line shall coincide with a given
plane Ax + By + Cz = D.
x a _y b
_z c
L, TT~ ~N~
:

The line must fulfil the condition (56) of parallelism above,

AL -f BM -f CN = o. And also any point on the line as (a, b, c)
must satisfy the equation of the plane. Hence we must have the
additional condition Aa + + Cc o. (57)
Bl> D=
2. Let the equations of the line be of the form x=mz +p )

y=nz + q }
. Sub
stituting these values of ,r and y in the equation of the plane, we
have

A(ms +/>)
+ K(nz -f q) + Cz = D,

whence z = 52 JT t And for coincidence this value of

Am + B/* + C
z must be indeterminate, and therefore A/ + B^ D = o )
(58)
Am + B + C = o f are the
conditions of coincidence.

NOTE. This last method is a general one of determining the con

ditions of coincidence of a straight line and any surface given by its
equation. That is substitute x and^ of the line in the equation of
the surface and since the z in the resulting equation mu^t be inde-
4
38 NOTES ON SOLID GEOMETRY,

terminate if there be coincidence we treat this equation as an iden

tity and make the coefficients of the different powers of z separately
equal to zero. ,

65. To find the expression

for the length of the perpendicular PD from
any point P(x , y ,z ) on a straight lin.> AB given by its
equation.

i. Let line be - -= : -

where a, b, c are the coordi-

cos a cos p cos
y
nates of any point A on the line. Now PD 2
= PA AD 2 2
.
(Fig. 13.)
But PA 2
= (x
- a y + (v -by + (z -cY and AD being the projec
tion of PA on AB, we have
AD (x a) cos a+(yb) cos /S+(z 6) cos y.

Hence

+ (z -b)cosy)\ (59)
2. If the given line be of the form

x a
__ y b
_zc :

A B C
Then
= A
cos a ____ etc., etc.

And therefore PD 2

3. If the given line be x mz + p )

y= nz + q )

Then PD 2

66. To find the expressionfor the shortest distance between two straight
lines given by their equations,
This shortest distance is a straight line AB perpendicular to both
the given lines PB and SR. (Fig. 14.)
Let the given equations

x a
= vb = ---
z c
and -
. xa yb-z -- 7 -,=
z c
7
and 6 = the
co 5 a cos// cos;/ a cos
cos cos ft y
angle between the lines.

And L, M, N the direction cosines of the perpendicular AB.

 NOTES ON SOLID GEOMETRY. 39

Then we must have

L cos a + M cos ft -f N cos y o }

L cos a + M cos ft + N cos ;/ = o. j

Whence
L M
cos /? cos T
-cos / cos y cos <* CO3 y cos a cos

N
cos <* cos ft cos M cos ft

M 2
-|-

[(COS /S COSy - COS /3 Cos y)*-t ^COS a COS >

- CO* a COS >)
2
-j- (COS a COS ^ -COBa COS 0)
2
1

sn
(Art. 15).
Now let P be the point (a, b, c) on the line PB and Q be the point
(a , b\ c) on the line SR, Then as the projection of PQ on AB is

AB itself, we have

AB = (a-a )L + (6-6 )M + (c-c )N =

(62)
6 ) (cosa cosy - cosacosy )+(c -c ) (cosacos/3 -cosa cos^S)
/
(a- a )(cos^cosv _co!ycos/3 )4-(6

If the given lines are expressed in other forms we can find cos a,
cos ft, etc, from the given equations and substitute them in
(62).
 CHAPTER V.

TRANSFORMATION OF COORDINATES.

67. To transform to parallel axes through a new origin the coordi

nates of which referred to the old axes are a, b, c.
Let OA= = r, NP = z be the coordinates of P referred to
.r, AN

the origin O and the axes O,r, Qv and Oz. Also let O be the new
origin, and OA
= a, A N = b, N O = c be its coordinates and let
O H = x HK =_/ and KP = z be the coordinates of P referred to
,

O as origin and axes parallel to the original axes. (Fig. 15.)

Then .r == OA = OA + A A
or x =a+ x )

Similarly y = b -f r !
(63)
and z c + z \

Substituting these values in the equation of a surface we obtain the

equation referred to the new origin and axes.
68. To pass from a rectangular system to another system tlie origin
remaining tlie same.

Let Ox, Or, Oz be the old axes at right angles to each other O_v ; ,

O/, Oz the new axes inclined to each other at any angle. (Fig. 16.)

OM = x, MN =.y, NP = z
OM M N =.r NT =
=.< , * .

Now the projection of the broken line -f + P OM MN N

on the
axis Ox is
equal to the projection OM
of the radius vector on OP
Ox. Let cos a, cos # cos ,
a" be the cosines of the angles which the
new axes make with the axis O.* ;
then

x =x cos a +y cos a + 1
z cos a 1

40
 NOTES ON SOLID GEOMETRY. 41

If cos cos ft ,
cos be the cosines of angles which the new
ft, ft"

with the axis Qy, and cos y, cos /, cos the cosines
axes make y",

we have similar values

of the angles which they make with Oz, shall

of transformation are
and z. Hence the three equations

x = x cos a +y cos a + z cos a"

j/
= ^ cos ft +y cos /? + cos /S"
>

(64)
z = x cos y +y cos / + z cos 7".
)

We have of course

cos
2
a + cos 2
ft +cos 2 y i }

cos
2
a + cos 2
ft + cos* y = i >

(B)
cos

For the angles A, /*, v between the new axes ofy and z ,
of 3 and
.r ,
of ..v and y respectively we have

cos A = cos a cos + cos /J cos

a"
1

ft"
+ cos y cos ;/
J

cos fA
= cos r"cos or + cos // cos /? +cos y"cos y I (C)
cos r = cos cos a + cos y5 cos ft +cos ^ cos y .
}

To pass from one system of rectangular coordinates to another also

69.
rectangular.
The formulae in this case are the same as those in the last with the

exception that since the new axes are also rectangular cos A
= o, cos //

= o, cos v =o
and formulae (C) give

cos a cos a" + cos ft cos ft"

+ cos y cos y"=
o )

cos a" cos a + cos ft"cos +cos ^"cos y

ft = o V (D)
cos <* cos + cos ft cos /T +cos y cos y o.)

Since between the nine quantities there are six equations of con
ditions, (B) and (D) there are only three of the quantities, cos a,
cos ft, etc., independent.
70. In changing from rectangular axes to rectangular, there is
another set of equations of condition among the quantities, cos a,
cos ft, etc., equivalent to the preceding which result from the fact
that the new axes are rectangular. For a, a being the angles ,
a"

made by the old axis of .v with the new rectangular axes, etc., we
must have
cos 2 tf + cos 2
+ cos
<*
2
a" = i }

cos
2
/? + cos
2
yS + cos
2
ft"
= i (
(E)
cos 2 ;/ + cos
2
7/ + COS
2
y"
= i )
42 A 07v<:s av SOLID GEOMETRY.

cos a cos /3 + cos rr cos ft + cos <*" cos y#"

=o i

cos a cos 7 + cos <? cos ;/ + cos "

cos 7"
=o >-

(F)
cos /? cos 7 + cos ft cos 7 + cos ft"
cos 7"
o ;

and the new coordinates expressed in terms of the old are

.v = .vcos a +jcos ft +scos y \

y ^cos a +ycosft +s cos y (F)

z x cos "

+y cos/?" +s cos 7"

)

71. In the study of surfaces by sections made by planes it is often

necessary to transform the coordinates in space to coordinates in

the cutting plane. To do this we must fix the plane with reference
to the old coordinate planes. Let the equation of the plane be given
as z= A^ + By. Then the angle 6 which this makes with the plane

xy is determined bv the equation cos 6 = and the

angle cp which it traces on that plane makes with the axis of x

= ^
by the equation tan cp ,
the trace being Ajv + Bj>
= o.

Let x Oy be the given plane, cutting the plane xy in the line

Ox which take for the axis of x and let a line perpendicular CV
to it in the given plane be the axis of v and OR = .r RM =y ,

the coordinates of any point M in the plane referred to the axes Ox\
(V also let
; .v, QP OQ = = r, P.AI =a z be the coordinates of M
referred to the old axes O v, Or, Os. Then the angle MRP = 9
and xOx = (p. (Fig. 17.)
Then PR =y cos ft PM =j/ sin 8.

OQ = OR cos (p + RPsin <p

y QP = OR sin <p-RP
co^ (p.

.-. z-y sin 8 \

x =x cos q>+y
cos 8 sin gj V (65)
y= .r sin <py
cos 8 cos cp )

And if these values be substituted in the equation of any surface

F(.v, r, z) o the result will be a relation between x andy, coor

dinates of the curve cut from the surface by the plane.

72. If the cutting plane contain one of the coordinate axes, the
formulae are simplified and in many cases sufficiently general.
Let x Oy (Fig. 18) be the cutting plane containing the axis of y ;

Ox its trace in the plane ; zx the axis of x .v _r ;

PM = ,
OM = ,
 .VOTES ON SOLID GEOMETRY. 43

the coordinates of any point P in the section =. x, =.r, ;

ON NQ
QP = 2, the coordinates of P referred to the old axes. Then angle
PMQ = 0, and MQ =
x cos PQ x sin 0. = , The formulae .

of transformation are

x =x cos 6* )

^=/ (66)
z x sin }

That is, we have only to make x = x cos 0, z x sin 0, y -=.y in

the equation of any surface, in order to find the equation of the sec
tion of this surface by a plane containing the axis ofj and making an
1

angle with the plane xy.

 CHAPTER VL

THE SPHERE.

73.To find the equation of the sphere.

i. In rectangular coordinates.
Let , <,
c be coordinates of the Centre, and Radius = R.

The equation then (Art. 1 )(x is 1

aY + (y-l>Y + (z c)
z
R z
(67)
or if the origin be at the centre

^+/ + 2 = R 2 2
.
(68)

2. Tn oblique coordinates.
Let A, v be the angles of the axes then
/<,
the equation is (Art. 16)

x-a)(y-b) cos A-f

2(xa)(zc) cos )A+2(yb)(zc) cos r =R 2

(69)

or if the origin be at the centre

x* + .)
2
+ -s
2
+ 2xy cos A -f 2AT2 cos yu + 2yz cos r =R 2
.
(70)

3. In polar coordinates
Let r , a, ft be the polar coordinates of the centre then the equa
tion is

2
2rr (cos ^ cos rt + sin 8 sin ^ cos (q> >5))
=R tf

.
(71)

If the pole be at the origin and the centre on the axis of 0, the
is
equation
r=2Rcos#. (72)

Since that is the equation of the generating circle in any one of its

positions.
44
 NOTES ON SOLID GEOMETRY. 45

74. The Sphere under conditions (coordinates rectangular).

The equation (67) may be written

JV
a
+>* + *
2CLX 2by 21-2 + 0* + &* + *
R 2
O

or .T
2
-f /+ ,s
2
+ DA- + Ey + FZ + G = o.
(73)

And since this equation contains four arbitrary constants, the

sphere may be made to fulfil four conditions (which are compati

ble)and no more. Four given conditions give four equations for

determining the constants D, E, F, G, and with these determined
we know the radius and centre of the sphere, for we have only by
completing the squares to throw the equation (73) into the form

/ DV / ^\ / F\ D E --
F-
2 2 2 2

+ y + - +(* + -) =
\x+-\
\ 2/ \ 27 V a/
+ + C* ( )
444
/ D --
-- E --
F\
to see that the centre is ( , , )
and the radius is
V 2 2 2j
^
F~~
7"

1. The equation of a sphere passing through a givtn point d, e, f, is

v2 4-y +
2
+ D(^-^) + E(.y-^) + F(0-/)-^-^-/ = 2
o. (74)

If the given point be the

origin the equation is

= o.
(75)

2. The equation of a sphere cutting the axis of z at distances c and c

from the origin is

1
x*+^+(z-c)(z-c ) + I)x + Ey= o (76) for"
^

must give two values for z, c and c ,

and this equation fulfils that

condition.

3. The equation of a sphere touching the axis of z at a distance cfrom

the origin is

y =o (77) for this gives two coinci

dent values of z = c when

"

=o
> .

J f
46 NOTES ON SOLID GEOMETRY.

4. 7 he equation of a sphere touching all three axes at distance a,

from
origin.
To meet these conditions the equation must be of such a form as
je=o )
to give equal roots for z when \ the same equal roots for y
*

when >
and the same equal roots for x when -
I . Let
z=o }
z=o j

the distance of points of contact from origin be a, then the equation

will be

= o (78)

as this fulfils the above conditions.

5. The equation of a sphere passing through the origin and having its

centre on the axis of x is

* s +y + **= 2R.V. (79)

6. The equation of a sphere tangent to the plane xy at the point (a, b)

is

(;t--rt) -f
2
-) 2
+ * + F*=o
2
(80)

for then z=o gives x=a, and y=b t

a point (a, b)
in the plane xy.

75. Interpretation of the expression

(x-a? + (.y-b)* + (z-c)*-W. (i)

i. Let (x, y, z) be the coordinates of a point P without the

sphere whose centre O is and radius = R and let

(a, b, c]
PM be

tangent to this sphere at the point M. Then PM 2 = OP 2 OM 2

.

Now

and hence PM = 2
(x-a)*+ (jy-) +
2
(z-<)
2
-R 2
.

Therefore the expression (i) is the square of the tangent from the

point P to the sphere.

2. Let P (x, y z)
}
be a point within the sphere. Join OP and
erect a perpendicular PM to OP meeting the sphere in M, and join

OM.
Then PM =OM*-OP =
8
R* -((x-ay + (y-by + (z-c)*}
 NOTES ON SOLID GEOMETRY. ^
That is the expression (i) becomes negative and represents the

square of the half chord through P perpendicular to the radius

through P.

76. .Radical plane of two spheres.

Def. The radical plane of two spheres is the plane the tangents
drawn from any point of which to the two spheres are equal.
If the equations of the two spheres are

the equation of their radical plane is

o
For this expresses (Art. 75) that the squares of the tangents from
point y, z) to the two spheres are equal, and moreover it is an
(_r,

equation of the first degree in x, y and z and therefore the equation

of a plane. If the spheres intersect their radical plane is their

plane of intersection. It may be easily proved that the radical

plane of two spheres is perpendicular to the line joining their cen

tres.

77. ,The six radical planes of four spheres intersect in a common

point.

Let S = o, S =o ;
S" =o ;
S
"

o be the equations of the four

spheres. Then the equations of their radical planes are

S-S =o S -S" =o
=
S-S" S - S
"

= o

S-S = o "

S"-S"f:==o

These may be arranged in groups of four equations, which added

vanish simultaneously and therefore the planes intersect in a common
point. This point of intersection of the six radical planes is called
the radical centre of the four spheres.

78. Examples :

i. Find the centres and radii respectively of the spheres

48 NOTES ON SOLID GEOMETRY.

40 = o.

\x + 5 jy = o.

2. F ind the equation of a sphere passing through the origin and

the points i, 2, 3, i, 4, 5, 3, >
i-

CHAPTER VII.

CYLINDERS, CONES, AND SURFACES OF REVOLUTION.

79. CYLINDERS. Def. A cylinder is a surface generated by the motion

of a straight line which always intersects a given plane curve, and is
always parallel to a fixed straight line. The moving straight line is

called \\\Qgeneratrix; the plane curve which it always intersects is

called the directrix or guiding curve.

80. To find the general equation of a cylinder (coordinates rectangular}.

Let m, n, i be the direction cosines of the axis.

And let >

(i) be the equations of the generatrix in

y nz + q \

which m and n are constant since the generatrix remains parallel to

the axis. For convenience take the guiding curve in the plane xy,

its equations will then be

^ z
~

o \
.
(2) Now making z =o in
5

the equations (i) we obtain x=p y q for the point in

which the

generator pierces the guiding curve y) in the plane xy.

(x,
Hence we have F(/>, q)
= o, (3) and eliminating the arbitrages/
and q between (i) and (3) we obtain

?(x-mz,y-nz) =o (82)

the general equation of cylinders.

If the cylinder be a right cylinder with its guiding curve in the

plane xy and the axis of z for its axis, then

in equation (82) m o, =
and o, =
and the required equation of the cylinder is

F(*, y) = o. (83)
 NOTES ON SOLID GEOMETRY. 49

8 Cylinders of second order.

1. We shall confine ourselves to cylin
ders whose equations are of the second degree.
i. To find the equation of the oblique cylinder with circular base.
Here F(x,y) = x 2 + R 2 = o. Hence (xmz, ynz) = o
/
gives (x mzY + (y nzf R =
2
o (84) the required equation.
2. To find the equation of the right cylinder with circular base. If

the axis be the axis of z, the equation is

J?(xy)
=o that is

3. To find the oblique cylinder with elliptical base. Let the guiding

curve in plane xy be +^ = i, z = o.
a2 b-

x* v*
Then (Jft jf)
= ^ + J an ^ the equation is

4. The equation of the right cylinder with elliptical base whose

x*
axis is the axis of z is F(x, y) = o,
v*
that is, s
+^= i.

5. The equation of the right parabolic cylinder whose guiding

curve is f = ^dx\ z o, is f $dx.

82. CONES.

Def. A cone is a surface generated

by a straight line which passes
through a fixed point and always intersects a given plane curve.
The fixed point is called the vertex, the moving line the generator,
and the given plane curve the directrix or guiding curve.

83. To find the general equation of a cone.

Let the coordinates of the vertex be b, the equation of the
(a, c)
x a y b z c
generator
^ ~~n~ ~T~ ^ and take the dircctrix in the

~
^1 r ) Now we
equation being then
- L
plane (xy) its
(2). if
z o )

eliminate w and by means of the definition of cone and the equa

tions (i) and (2), the resulting equation will be the
equation to the
cone, the locus of the right line (i).

z =o in (i) the values of* and

~ a ~ mc
x
Making namely, 1
y,
=
y b nc )

which result will be the coordinates of the point in which the

gene-
5
50 NOTES ON SOLID GEOMETRY.

rator meets the plane xy and these will consequently satisfy (x,j )
=o the equation of the directrix. We have therefore

F(a me, bnc) = o (3). But from (i) m =^~

2 C
,
n = -
Z C
,

and therefore (3) becomes

zc z c

az ex bz cy
or
zc
the general equation of cones. If vertex be on axis of z, then a o

and b = o and equation (85) becomes F( c

,
z c
)
= o. (86)
\z )

84. Cone with vertex at origin.

If the vertex of the cone is at the origin and the directrix in a

plane parallel to the plane xy, and at a distance c from it then the

equation of the generatrix will be = = ,

(i) the vertex (o, o, o)

~
and the directrix will be v*^J I
(2)
=
.

z c )

To find the point in which the generator meets the directrix we

i). We thus get

Hence we have (mc, nc)

= o, but m= ,
and n - from (i).

Therefore
r
(B7)
<?,7 )=
is the equation required.

The equation (87) is a homogeneous equation in x,y and z.

85. Cones of second degree. 1. The equation of an oblique cone with

circular base.

The equation of the directrix is

~F(x, y)
= x +y R = 2 2
o.

Hence

zc zc zc
 NOTES ON SOLID GEOMETRY. 5!

or (az -ex)* + (bzcy}^ K\z-c}\ (88)

2. To find the equation of a right cone with circular base, the axis of
z being the axis of the cone and vertex being (o, o, c). The equation

of the directrix
.

is
v
F(*,.> )=-* +y-R =o s

z c.

Hence
W ^ )=o -^ 7^-R
zc =s <7
is
f
2 f

R 2

or jt-
2
+y= (2 r)
2
(89). This is a cone of revolution about the

axis of z.

3. The equation of a right cone with vertex at the origin and circular,
elliptical, or hyperbolic bases.

The equations of the circular base (directrix) are

Hence
/ ex cy\ . fix
1
c^y
1

""
2
R 2

2 ^ 9
\T" ~z)~ *~z* s ""?"

The equations of the elliptical and hyperbolic directrices are

_ =o and - = 1 o
^ +- i )
\
| --|
2
respective]y .

z=c] z= c)

Hence the cones are

~
x f z 9 2
. .

i
=oor--+--=-5 (91)

c*x* W
86. SURFACES OF REVOLUTION.

To find the general equation of a surface generated by the revolution of

a plane curve generator about the axis of z.

Let SPi=r be an ordinate of the point P to the axis of z of the

52
NOTES ON SOLID GEOMETRY.

plane curve and OM = x, MN =y, NP = z the coordinates of P.

Then SP 2 = ON 2
OM + MN 2 2
,
or r
2
.r
2
+y
That is, the distance from any point of revolving curve (gen

erator) from the axis of z r==vC?4j? (i). But r being an ordi-

is

nate of the generating curve to the axis of z we must have by the

equation of the curve in any position r = (z) (2). Therefore

eliminating the arbitrary r between (i) and (2) we have

V^+y = F(s) (93)

the required equation of surfaces of revolution about axis of z.

If the curve revolved about the axis of .r the equation is

Vy +s 2
F(JC). (94)

87. Surfaces of revolution of second order.

i. Equation of Cylinder of revolution about the axis of z. The
equation of the revolving line is r = a.
V. -**+ = gives x-+f = a\

2. Equation of a Cone of revolution about the axis of z, vertex at

(o, o, c).
The equation of the generating line is r = m(zc}.
Hence .#
2
+y 2
= m*(zc)* (95) the required equation where m is

the tangent of the angle made by side of cone with axis of 2.

3. Equation of the Sphere. The equation of the generating curve

2
a* or r =\/a* z .

Hence

4. Equation of the Surface generated by the revolution of an ellipse

about its
conjugate axis.
9 2

The generator is - + Z-=

r*
i or r~ = ~(b /70
z
OX
2
).

Hence the equation of the surface is

+=- (96)

This is one of the ellipsoids of revolution called the oblate spheroid.

 NOTES ON SOLID GEOMETRY, 53

5 .
Equation of the Ellipsoid generated by the revolution of an ellipse
about its transverse axis the {Prolate spheroid].

Take the axis of x as the axis of revolution. Then the equation

jv r
of the generator is + 75 =

or r
2
=
a
>(#
2
x*)>
Hence vj + s*
_
*
F(,r) gives

v _ x
r**;
i

(97)

the required equation.

88. Hyperboloids of revolution. Definitions. When the Hyperbola

revolves aboutits conjugate axis it generates the Hyperboloid of revo
lution of one sheet. When it revolves about the transverse axis it gen
erates the hyperboloid of revolution of two sheets.
i. Equation of the Hyperboloid of one sheet. Let the axis of z be
2 2
r z a2
the conjugate axis then -273 = i or r 2 = j3(s + 8 2
).
Hence

2. 7%f equation of the Hyperboloid of revolution of two sheets. Take

the axis of x as the axis of revolution. Then the equation of the
2
a: r2 b*
generator is
^
= i or r
2
= (j\r
2
^ 2 ).

Hence for the equation of the surface we have

A- ,- 3
?--?-= (99)

89. Equation of the Paraboloid of revolution about the axis of\.

The equation of the generator is r 2 \dx.
Hence the equation of the Surface isjy 2 + 2 2
\dx. (100)
5*
 CHAPTER VIII.

ELLIPSOIDS, HYPERBOLOIDS, AND PARABOLOIDS.

89. To find the equation to the surface of an Ellipsoid.

Def. This surface generated by a variable ellipse which always

is

moves parallel to a fixed plane and changes so that its vertices lie on
two fixed ellipses whose planes are perpendicular to each other and to
the plane of the moving ellipse, and which have one axis in common.
Let BC, CA (Fig. 19) be quadrants of the given fixed ellipses traced
in the planes ;

> 0, zx ;
OC c their common semi-axis along the axis
of z, OA =
a (on the axis of x), and OB b (on the axis of y) the =
other semi-axes QPR quadrant
a of the variable
; generating ellipse
in any position, having i s centre in OC and two of its vertices in the

ellipses AC, BC, so that the ordinates QN, RN are its semi-axes ;

also let ON = z, NM = x, MP =y be the coordinates of any point

P in it :

3? y*
Then -^ + .- = i. And since Q is on the ellipse AC we

have . = i -.
Similarly
-- = i .

a c u c

Hence eliminating RN 2
and QN 2
we have
*
^L
-r +-
7,-+--r== ,
(loi)

the equation to the surface.

90. To determine the form of the ellipsoid from its equation. Since in
2 2
.v y z*
the equation y -f- y- -f
- = i ,
x can only receive values between a

54
 NOTES ON SOLID GEOMETRY. 55

and a, y between b and b, and z between c and c, the surface

is limited in all directions.
2
A2 V
If we put z = o we obtain -f r= i, for the equation to the

trace on xy, which is therefore the ellipse AB.

x* z1
If we put y o we have ~--\ ^- =i, the ellipse AC.

If we put x = o we have 1
r -f = i, or the ellipse BC.

These three sections by the coordinate planes are called the princi-
pal sections, and their semi-axes a, 6, c, are the semi-axes of the ellip
soid and their vertices the vertices of the ellipsoid, of which it has six.
;

If we make z=h we have

the equations of any section parallel to xv, which is an ellipse similar

to AB, since its axes are in the ratio of a to b, whatever be the value
of h,and which becomes imaginary when h c. In the same man >

ner sections parallel to xz and yz are ellipses respectively similar

all 9

to AC and BC. The whole surface consists of eight portions pre

cisely similar and equal to that represented in the figure.
z
yA ^_ z
Cor. If =0 the ellipsoid becomes ^
j,2
t-- = i the ellip

soid of revolution about the axis of z, Art. (87), all the sections of

which by planes parallel toyz, are circles. Hence the spheroids may
be generated by a variable circle moving as the variable ellipse, in
Def. Art. (89).

91. To find the equation to the hyperboloid of one sheet.

Definition. generated by a variable ellipse, which

This surface is

moves parallel to a fixed plane, and changes so that its vertices rest
on two fixed hyperbolas, whose planes are perpendicular to each
other, and to the plane of the moving ellipse, the two hyperbolas
having a common conjugate axis coincident with the intersection of
their planes.
(Fig. 20.)
Let AQ and BR bethe given hyperbolas traced in the planes zx,yz ;
OC = c their common semi-conjugate axis coinciding with the axis
of z ; OA = a, OB = b the semi-transverse axes ; QPR the generating

ellipse in any position having its plane parallel to xy, its centre in
56
NOTES ON SOLID GEOMETRY. 1

OC, and its vertices in the hyperbolas AQ, BR, so that the ordinates
NQ, NR, are its semi-axes. Also, let MN = x, MP = y, ON 2,
be the coordinates of any point P in the generating ellipse ;
then the
ellipse PQR gives

2 + NlT = 2
I
"NQ

NQ 2
z*
Also from hyperbola AQ 2
"~=i.

NR 2

BR
z"

And from hyperbola

--
^ i.

Hence,
f

8
A- V
2
Z
2

or + ^- -
i
(102) the equation to the surface.

92. To determine the form of the hyperboloid of one sheet from its

equation.

Since the equation (102) admits values of x, y and z positive and

negative however large, the surface is extended indefinitely on all

sides of the origin. If we put z = o we obtain

~z- + ~i = i for the trace on xy which is the ellipse AB. Similarly

2 2
r s
the sections by the planes xz and j 2 are respectively ^ - i the

2
v z
hyperbola AQ, and 1
-^ ^ = i the hyperbola BR. The ellipse AB
and the hyperbolas AQ and BR are the principal sections. The sections
parallel to xy are all ellipses similar to and greater than AB. The
sections parallel to xz and yz are hyperbolas similar to the principal
sections.
The semi-axes a and b are called the real semi-axes of the surface
and c the imaginary semi-axis, since x = o and y =o give z =
c\/ i. The extremities of the real axes are called the
vertices of the surface. The surface is continuous and hence is

called the hyperboloid of one sheet. The hollow space in the inte
rior of the volume of this hyperboloid of which the ellipse AB is the
smallest section has the shape of an elliptical dice-box.
 NOTES ON SOLID GEOMETRY. 57

Cor. If b = a the equation becomes =

2
_
j
= i that of the
fl c

hyperboloid of revolution of one sheet. Its sections parallel to xy

are all circles.

93. To find the equation to the hyperboloid of two sheets.

Definition. This surface is generated by a variable ellipse which

moves parallel to itself, with its axes on two fixed planes at right an

gles to each other and to the plane of the generating ellipse and ver
tices in two hyperbolas in those planes having a common transverse
axis.

LetAQ and AR be the given hyperbolas traced in the planes zx,

xy, OA = a their common semi-transverse axis along the axis of x,
OB = b OC = c the semi-conjugate axes along the axes ofy and z ;

QPR the generating ellipse in any position having its plane parallel
loyz its centre in Ox, and its vertices AQ, AR so that the ordinates
t

QN, RN are its semi-axes. Let ON = x, MN MP = z be the =j>,

coordinates of any point P in the ellipse. (Fig. 21.)

"

also from hyperbola AQ c a

i

and from hyperbola AR "-

= i.
l> a1

Hence

- >-?
-
the equation to the surface.

94. To determine the form of the hyperboloid of two sheets from its
equa
tion.

The equation shows that all values of x between -f# and a give

imaginary results, therefore no part of the surface can be situated be

tween two planes parallel toyz through A and A the vertices of the
common transverse axis ; but the equation can be satisfied by values
58 NOTES ON SOLID GEOMETRY.

of x, > , z, indefinitely great, therefore there is no limit to the distance

to which the surface may extend on both sides of the centre.
V* Z*
If we make x-= o we have ~- -\
=
o c

for the principal section by the plane yz. For x = h and h >
a
2
v* z* //
we have -
-\
^-
-
2
i which represents similar ellipses. The

principal sections by the planes xy and zx are and AQ, respec AR

tively. For the sections parallel to xy and putting z I we =
have ^
= = + i a hyperbola similar to AR with its vertices

in AQ andthe opposite branch of that hyperbola and conjugate axis

parallel to Oy. In the same way the sections parallel to zx are hy

perbolas similar to with vertices in AQ
and its opposite branch AR
and conjugate axes parallel to Qz, 2a is the real axis of the surface
and its vertices the vertices of the surface. The axes 2b and 2c
are the imaginary axes of the surface as it cuts neither y nor z. The
whole surface consists of two indefinitely extended sheets perfectly
similar and equal, separated by an interval. Hence its name.

Cor. \ib-c the equation becomes ~ ^

= i the equa

tion to the hyperboloid of revolution about its transverse axis.

95. Asymptotic cones to the two hyperboloids.

i. The hyperboloid of one sheet has an interior asymptotic cone.

2
x* r z*
Putting its equation T + ^ -j-
= i
(i) in the form

4. -=s /I ^--- .
(2) Now when z is very great

,-
is very small, and hence the limiting form of (2) for z increased

without limit is

v
r + ~- = r (3) the equation of an elliptical cone having

its vertex at the origin and its elliptical section parallel to xy.

Moreover, this elliptical section is always within the corresponding

 NOTES ON SOLID GEOMETRY. 59

section of the surface by the same plane. For putting z h in

(i) and (2) respectively we have

~-?-
1
+ ~rr = + l
r f r tne section of the surface
a // c*

x *
/ = /r

a
^-
2 + 4^-
b
1
c
7-
2
for the section of the cone.

This cone asymptotic to the hyperbola.

is

2. The hyperboloid of two sheets has an exterior asymptotic cone,

x~ V* z~
=
rr
Putting the equation ,
^-r i (i) under the form
a2 (j

we have as a limiting form of

this equation when jy and z increase without limit,

2 2
jv z~
= j
-----
-|
---
(2) an elliptical cone with vertex at the origin

and with an elliptical section parallel to the plane \z. Moreover,

this elliptical section is greater than the corresponding section of
the surface by the same plane. For putting x .h in (i) and (2)
2 2
h*
respectively we have
i

-,-
*
H---
-- = 5
2
i

^ "T"
55
?"

This cone is asymptotic to both branches of the hyperboloid.

96. To find the equation to the elliptic paraboloid.

Definition. This surface is generated by the motion of a parabola

whose vertex lies on a two parabolas
fixed parabola, the planes of the

being perpendicular to each other, their axes parallel and their con
cavities turned in the same direction.

Let OR be a parabola in the plane xy, its vertex at the origin, its
axis along the axis of .r, and / its latus rectum RP the generating ;

parabola in any position with its plane parallel to zx, vertex in OR,
and axis parallel to O.v, and let / denote its latus rectum. Also let
ON NM = MP
x, r, zbe the coordinates of any point P in it :

also draw RM parallel to Oy. (Fig. 22.)

60 NOTES ON SOLID GEOMETRY.
Then z* = I RM = / .M N and / = .OM
. I ;

. .

j-
+ -j- = x (
IO 3) the equation to the surface.

97. To determine the form of the elliptic paraboloidfrom its

equation.
Since only positive values of x are admissible, no part of the sur
face is situated to the left of the planers. But the surface extends
indefinitely in the positive direction of x. If we makej o, 2
2
Ix = =
is the equation to the principal section OQ, and all sections parallel
to zx are parabolas equal to OQ, with vertices in OR ; similarly, all
sections parallel to xy are parabolas equal to the other principal sec
tion OR, with vertices in OQ. If we make x h we have

21 + Jl-,
~
Ih Ih

Therefore the sections parallel to zyare similar ellipses, and hence its

name.
Cor. If /= / the equation becomes j
2
+ = 2
/.v, the paraboloid of
revolution.

98. To find /he equation to the hyperbolic paraboloid.

Definition. This surface is generated by the motion of a parabola

whose vertex lies on a fixed parabola, the planes of the two parabolas
being perpendicular to each other, their axes parallel, and their con
cavities turned in opposite directions. (Fig. 23.)
Let OR be a parabola in the plane of xy, vertex at the origin, and
axis along with the axis of x, and / its latus rectum, RP the generat

ing parabola in any position, vertex in OR, axis parallel to O.v,

and let t denote its latus rectum, and ON = NM =y, MP

x, z,

the coordinates of any point P in it ;

draw RM parallel to Oy.
Then
z-=! MR and/=/.OM
.
;

but OM - MR = ON = AT.

2 2
V
Hence :
---2- = x (104), the equation of the surface.

99. To determine the form of the hyperbolic paraboloidfrom its equation.

The surface cuts the coordinate axes only at the origin, and since
the equation admits positive and negative values of.r, v, s, as great
 NOTES ON SOLID GEOMETRY. 6f

as we please, the surface extends indefinitely both ways from the

origin.
we makejy = o we have z~ = I x the principal section, the para
If
bola OQ, with its concavity turned towards the left ofyz, and all sec
tions parallel to zx are parabolas equal to OQ with their vertices in
OR. Making z = o we have y 2 = Ix the parabola OR, and sections
parallel to xy are parabolas equal to OR with vertices in OQ.
If we make x o we have the principal section in yz,

z^ I = y +Jl or two straight lines through the origin; and for sec
tions parallel toyz making x = h we have

- -- f
= i a hyperbola with its vertices in OR, and con

jugate axis parallel to Oz. For h negative the section becomes

2 2
y
In
7-
Ik
-
i a hyperbola with its vertices in OQ, and conjugate

axis parallel to O.v.

The surface has but one vertex, and consists of one sheet and one
infinite axis.

100. Asymptotic planes to the hyperbolic paraboloid.

v* z*
The equation --- = x may be written
-jr

j
2
z* / l x\
z
-j~
==
~jr (
T + ~7 )
which has for its limiting form
/ / \ z J
l
Z
<1

y
when y and z become infinitely great with regard to x,
- = 7- ,

or -:
V
= = -. z
This represents two planes
v
=. = H--z and

-
V
= = --^ - .
through the origin and asymptotic to the surface.
V?
These planes contain the asymptotes to all the hyperbolic sections of
the surface parallel toj 2.

101. The elliptic and hyperbolic paraboloids are particular cases of the
ellipsoid and hyperboloid of one sheet respectively when the centres of these
surfaces are removed to
infinite distance.

Take the equation + ^- H =i, and transfer the origin to

62 NOTES ON SOLID GEOMETKY.
the left vertex of the axis 20,
(a, o, o). (New coordinates being
parallel to the primitive.)

-
-,. 1 . or
vJ1 o
2
a

r2
or multiplying through by a - -- t-
y*
=2jr (i),
-y^-

in which ,
and are the semi-latera recta of the principal sec-
a a
V 1
c*
tions in xy and zx. Now make a =. oo and put -
and ,

a a
which remain finite, equal to / and / respectively.

.*.
(i) becomes

=y- JT
2x, the equations to the paraboloids.

1 02. The equations of the surfaces of the second order which we have
been studying are of the two forms

=D (i)

=kA* (2)

and we will show hereafter that all the surfaces of the second degree

may by transformation of coordinates be included in these two forms.

The first form (i) includes the sphere, ellipsoid, hyperboloids,
cones of second order, elliptical and hyperbolic cylinders which
have centres. For if .r, -y, z be written for (x, y, z) in (i)
the equation is not altered, therefore for every point P (x, y, z) on
the surface there is a point P ( x, y, z) and PP passes through
the origin O
and is bisected in O.
Moreover, the coordinate planes bisect all the chords parallel to
the axes perpendicular to these planes respectively and are principal

planes of the surface.

The second form (2) includes the elliptic and hyperbolic parabo
loids and the parabolic cylinder which have a centre at an infinite
distance.
The planes ^.s: and zx are principal planes of the two paraboloids,
the other principal plane being at an infinite distance.
Also both families may be represented by the equation
 NOTES ON SOLID GEOME IRY. 63

the origin being at the vertex and A=o when the surfaces have
no centre.

EXAMPLES.

1. Construct the sphere whose polar equation is

r =a sin 6 cos cp.

2. Find the locus of the point the sum of the squares of the dis

tances of which from n fixed points is constant.

3. Find the locus of the point the ratio of the distances of which
from two fixed points is constant.
4. Find the equation of the surface generated by the motion of a

variable circle whose diameter is one of a system of parallel chords

of a given circle to which the plane of the variable circle is perpen
dicular.

5. The sphere can be represented by the simultaneous equations

x= a cos cp cos 6 }

y = a cos cp sin 6 >

z = a sin cp

6. The ellipsoid may be represented by the equations

x = a cos <p
cos 6 }

y b cos cp sin 6 Y

z = c sin cp

7. The hyperboloid of one sheet may be represented by the equa

tions
x a sec cp cos 6 \

y b sec cp sin 6 \ -

z =c tan gj

8. The hyperboloid of two sheets may be represented by the equa

tions
x a sec cp \

y = b sin 6 tan cp >

z c cos 6 tan cp )

9. A line moves so that three fixed points on it move on three

fixed planes mutually, at right angles. Find the locus of any other
point P on its line.
64 NOTES ON SOLID GEOMETRY.
Solution :

Let the three fixed planes be the coordinate planes (x,y,z) the coordinates
of P. A, B, C the points in which the line meets the coordinate planes of yz,
xz, xy, respectively. Take PA=<z, PB=/>, PC=c, ON=.r, NQ=,y, QP = z,
<ACA =:<p, <CB .r=:0 (CA being the projection of CA on the plane xy and B
the projection of B on the axis of x).
Then x=a cos cp cos Q,y=b cos q>
sin 6, z= c sin q>,
and therefore the sur
face is an ellipsoid.

10. Find the locus of a point distance of which from the plane xy
is equal to its distance from the axis of z
(coordinates rectangular).
11. Find the locus of the centres of plane sections of a sphere
which allpass through a point on the surface.
12. Find the equation of the elliptical paraboloid as a surface
generated by the motion of a variable ellipse the extremities of whose
axes lie on two parabolas having a common vertex and common
axis and whose planes are at right angles to each other.

13. Find the equation of the hyperbolic paraboloid as generated

in a similar manner by the motion of a variable hyperbola.
14. Construct the surface r sin 6 = a.
15. Find the equation to the surface B = JTT in rectangular coor

dinates.
 CHAPTER IX.

RIGHT LINE GENERATORS AND CIRCULAR SECTIONS.

103. SURFACES of the second degree admit of another division, viz.

into those which can be generated by the motion of a straight line

and into those This property which we have seen to

which cannot.
belong and cone
to the cylinder we shall now show to belong also to
the hyperboloid of one sheet and the hyperbolic paraboloid. The
ellipsoid being a closed finite surface does not possess this property ;

nor the hyperboloid of two sheets, since that consists of two surfaces
separated by an interval ;
nor the elliptical paraboloid, since that is

limited in one direction.

104. Straight line generators of the hyperboloid of one sheet.

The equation of the hyperboloid of one sheet

2 2
a2 z* .r z* r
v-
z + ^y
5- i mav be written -
-r- = i ~-~
a o* c a2 c b
1

fx z\ fx z\ _ f
y\( y\ - ( \\
\a c ) \a c ) \ b ) \ b)

Now (A) is satisfied by the pair of equations

(B)
Z f V\
+ 7
=K I +
i)J
and also by the pair
66 A OTES ON SOLID GEOMETRY.

And m being arbitrary equations (B) represent a system of straight

lines, and all of these lie on the hyperboloid as the two equations

together satisfy the equation to the hyperboloid.

Similarly equations (C) represent another and distinct system of
straight lines which also lie on the hyperboloid which is the locus of
both systems, and we shall see the lines of either system may be used
as generators of the surface.

105. No tivo generators of the same system intersect one another,

For example take two of the system (B),

m / /-v
i
A )
= -
i -.
v 1
\a c J b
(0
x
a

(2)

a c \
j
j

Combining the first equation of (i) with the first of (2) we obtain

(in m")
( I
J
=

Combining the second equation of (i) with the second of (2) we

have

or v ~ .

These values for y being incompatible the lines do not intersect.

1 06. Any generator of the system (B) will intersect any generator of
the system (C).

Take

,
(*--*-}=
c /
t-Zb
of system (B)
z \
(x + =w ,
(
+ y\
t r) ( i) J
 NOTES ON SOLID GEOMETRY. 67

U (4) of system (C).

Eliminating x, y, and a; we obtain the identity m m" m m",

therefore the lines intersect.

Hence, through any point of an hyperboloid of one sheet two
straight lines can be drawn lying wholly on the surface.

107. No straight line lies on an hyperboloid which does not belong to

one of the systems of generating lines (B) or (C).

For, if suppose a straight line H to lie entirely on the

possible,
hyperboloid, must meet an infinite number of generating lines of
it

both systems (B) and (C). Let two of these (one of B and one of

C) intersect H in two different points, \ve could then have a plane in

tersecting the surface in three straight lines, which impossible since

is

the equation is of the second degree. Hence no such line as can H

lie on the surface.

1 08. The hyperboloid of one sheet may be generated by the motion of a

straight line resting on three fixed straight lines which do not intersect, and
which are not parallel to the same plane.
In the first place it is
necessary that the motion of a right line
which is to generate a surface should be regulated by three condi
tions. For, since its equations contain four constants, four condi
tions would position absolutely ; with one condition less the
fix its

position of the line is so far limited that it will always be on a certain

locus whose equation can be found.
Take then three fixed generating lines of the system (B), these do
not intersect, nor are they parallel to the same plane. Now, if a
straight line move in such a manner as always to intersect these three
straight lines, it will trace out the hyperboloid of which they are the
generating lines.
For the moving line meets the hyperboloid in three points (one
on each of the fixed straight lines), and hence must necessarily lie
wholly upon the surface. For the equation of intersection of a line
and this surface being a quadratic equation, if satisfied by more than
two roots, it is satisfied by an infinite number. The moving straight
line, therefore, in its different positions, will generate the hyper
boloid.
68 NOTES ON SOLID GEOMETRY.

109. Lines through the origin parallel respectively to generators of the

systems (B) and (C) lie on the cone

x*
r + y = =
z-
asymptotic to the
-73- hyperboloid.

For this equation of the cone may be put in the form

A A \a
\
A + A = _,vz,
a c J c t b b

which gives two systems of lines through the origin lying on the
cone, one system evidently parallel to the lines (B) and the other to
the lines (C).

no. The projection of a generating line of either system upon the

principal planes., is tangent to the traces of the surface on those planes.

The equation of the trace of the surface on the plane zx is

The projection of the line of system (B) on xz

f_ i\ + * +
\ a c
J a c
=2m .
or ^}.
2m
*
a
+ 1=*.
2m
i=
c
I
(I) .

P
Now. the condition that a line in the form
"V

~|
= i shall be
p q

tangent to the hyperbola

~
.r
2
z1
F =I IS
~^~
a* c*
T T -

This -condition is fulfilled by the projection (i), for

( +!) (i-m Y

Hence this projection is tangent to the hyperbola.

III. The straight line generators of the hyperbolic paraboloid.

The equation of the hyperbolic paraboloid
 NOTES ON SOLID GEOMETRY. 69

=y
--- = x may be written

And hence it is satisfied by the pair of equations

y
V7
=i
n + ^) _

or by the pair

y z }

HE).

Hence the surface has two systems of straight line generators (D)
and (E).
The lines of both systems are parallel to the asymptotic planes
of the surface respectively. The equations of these planes being

z
==o and

112. We can show in the same manner as in the Articles (34)

and (35) that no two lines of the same system intersect and that a ;

line of either system intersects all the lines of the other system, and
that no other line than the lines of these two systems can lie on the

hyperbolic paraboloid. And hence that through every point of the

surface two lines may be drawn which lie wholly on the surface.
And as in (108) that this paraboloid may be generated by the motion
of a straight line which rests on two fixed straight lines and is con

stantly parallel to a fixed plane also by a straight line which rests

;

on three fixed straight lines which are all parallel to the same plane.

113. The projections of the generating lines on the principal planes are
tangent to the principal sections of the paraboloid.
The principal section in xy is y* = Ix (i).
 NOJ^ES ON SOLID GEOMETRY
The projection of any line of the system (D) on xy is

j= mx ^ or y = - jp-fJ^-rw . 2 /)
/t / / m
-WJ O
2 2m (*

Now the tangent line to the parabola y>=

Ix is of the form

y tx + : and if / = then = -5 -.

4^ 2 4/ 2m

Hence the projection (2) is tangent to the section y = lx t

114. Distinctions of surfaces of second order generated by straight

lines.

All the
generators of the cone intersect in one point. All
the of the cylinder are parallel.
generators Hence cones and
cylinders are called developable ruled surfaces. In the case of the
hyperboloid of one sheet and the hyperbolic paraboloid, the gen
erators of neither system intersect or are parallel. These are
styled skew ruled surfaces. The distinction between these last two
surfaces is that the generators in the paraboloid are parallel to a fixed

plane.

115. Plane sections of surfaces of the second order.

If we intersect the surfaces represented by the general equation

A* + By + Os + 2h. xz +
2 2 2
2B>
+ 2Cxy + 2& x+2 B"y+2C"z - D

by the plane z = o we will obtain

A^ + B| + 2C A7
2 2
|-2A"jt: + 2B y =D (i) a conic section.

If we intersect it
by a plane z a we have for the curve of inter

section

Aa 2 -h B^ + 2C xy+ 2G x + 2H>=D
2
,

a conic similar to the conic (i).

Therefore sections of surfaces of the second order by parallel

planes are similar curves, and hence, in determining the form of these
sections we may confine ourselves to the discussion of sections through
the origin.
 NOTES ON SOLID GEOMETRY. ^
1 1 6. To determine the nature of the curve formed by the intersection of
a surface of the second order by any plane.

Take the equation

= 2A r
,r. And in order to get the equation of the

curve of intersection in its own plane

Make
x x cos q>
-f y cos 6 sin cp

y zz: x sin cp y cos 6 cos cp

f
z =y sin 6. See Art. (71).

Arranging the result we have

x *(K cos
2
r/>
+B sin
2
cp) + 2x y (K B) cos 6 sin cp cos cp

+/ ((A sin v
2 2
+ B cos 2 93) cos
2
+C sin
2
0) = 2 AV cos <p

-f 2Aj/ cos 6 sin (>,

the equation to a conic section which will be an ellipse, parabola or

hyperbola, (including particular cases of these curves,) according

as the quantity

(A
2 2
B) cos 6 cos
2
cp sin
2
cp (A cos
2
+ B sin
<p
2
<p)(A
2
cos 6 sin 2 cp
+ B cos #cos 2 2
<p
+ Csin 2
0)

or -AB cos
2
cp-AC cos
2
^ sin
2
BC sin
2
<p
sin
2
8, (i)

is negative, zero or positive.

Hence every section of an ellipsoid is an ellipse because A, B and

C are all positive.
The sections of the hyperboloids may be ellipses, parabolas or
hyperbolas since one or two of the quantities A, B and C will then
be negative.
For paraboloids A = o. Hence for the elliptic paraboloid in
which B and C have the same signs the section is an ellipse except ;

when B = o or cp = o in which cases it is a parabola.

For the hyperbolic paraboloid since B and C are of contrary signs
the section is a hyperbola except when 6= o or cp=o when it is a

parabola.

1 1
7. Circular sections. Since the section is referred to rectangular
axes it cannot be a circle unless the coefficient of xy vanishes
72 NOTES ON SOLID GEOMETRY.
or (A B) cos 6 sin cos cp o
cp

7T 7T
or #
2
or &= -
2
, or cp o

which shows thaty^r # circular section the cutting plane must be perpen
dicular
to one of the principal
planes of the surface,
1 1 8. Let us now examine the surfaces of the second order for cir
cular sections.
Take first the surfaces having a centre and therefore represented

by the equation
A* + B/ + Cz =
2 2
i.
(i)
Since every circular section must be perpendicular to a princi
pal plane, let the cutting plane contain the axis of y, and make
the angle 6 with the plane xy
To transform (i) to this plane make

x =x cos 6

y=y
z x sin 6. Art. (72).

Hence we have

.v"(A
cos
2
+C sin
2
0) + B/* = i
(2)

which represents a circle if

Acos #-fC 2
sin
2
B

or tan
2
9 = B-A .
(3)

We must now examine for each of the surfaces which axis it is

that coincides with the axis ofj>.

i. For the ellipsoid A= ,

C =
3=-^-,

Hence for a real B b must lie

(in value) between a and c or the
axis of the surface to which the cutting plane of circular sections is

parallel is its mean axis.

2. For the hyperboloid of one sheet since we cannot have B ne-

 NOTES ON SOLID GEOMETRY.

gative we must put A = ^ B = C -

. . b >
a or the cutting plane is parallel to the greater of the real
axes.

3. For the hyperboloid of two sheets since we cannot have A and

C negative, we must put

.*. b >
c or the cutting plane is parallel to the greater of the im
aginary axes.
Since tan 6 has two equal values the cutting plane may be inclined
at an angle 6 or 180 6 to the plane of xy. Hence there are two
sets of parallel circular sections of the surfaces having a. centre. If
the surface becomes one of revolution we have tan 6 = oo or o, and
the two positions of the circular sections coincide with each other,
and are parallel to the two equal axes.

119. Secondly. For the surfaces not having a centre, we take

equation B_/ + Cs = 2k x (i).
2

i. For the elliptic paraboloid, B and C have the same sign.

Transforming (i) we have B/ + Or sin #
2 2 2
2K x cos 6\ and hence
for circular sections we must have the condition C sin = B, or 2

sin 6 = \ - Therefore the

cutting plane is perpendicular to the
^
principal section whose latus rectum is least.

2. For the hyperbolic paraboloid, sinceB and C have different

signs, sin 6 is
imaginary, and no plane can be drawn which shall in
tersect it in a circle. This was evident, too, from the fact (Art.
116) that the hyperbolic paraboloid can have no elliptic sec
tions.

i 20. Then, to sum up, all the surfaces discussed with the excep
tion of the hyperbolic paraboloid admit of two sets of planes of cir-

7
74 NOTES ON SOLID GEOMETRY.
cular sections. Therefore they can be generated by the motion of a
variable circle whose centre is on a diameter of the surface.

121. The planes of circular section may be found directly from

the equations of the surfaces, as follows:

The equations of the central surfaces

may be written
or

which shows that either of the planes

<\/A
BJI +V^-^ =o (i) <v/A
B.r VB C.s = o (2)

cuts the surface in the same line in which it cuts the sphere

Hence the planes (i) and (2) and all planes parallel to them cut
the surface in circles.

The equation to the elliptic paraboloid may be treated in a similar

manner, thus showing its planes of circular section.

122. Sections of Cones and Cylinders.

i. The sections of the cones may be inferred from Art. 95. For
elliptic cones, sections of the hyperboloids by any plane are always
similar to the sections of the asymptotic cone to the surface made by
the same plane, as is evident from the equations respectively. Hence
the section of a cone of revolution by a plane will give an ellipse,

parabola, or hyperbola. But we will examine this case more par

ticularly.
In the equation of the cone of revolution
2

A. + /= L (
2 _, ), or .v>+y
= (*-<)

(when - -
J put x=x cos 0}
tan v J
 NOTES ON SOLID GEOMETRY, 75

And we have for the curve of intersection by the plane containing

the axis ofjy

jr
2
(cos
2
#tan 2 v sin
2
6) + _/* tan
2
v+2cx sin 6c = o 1
(i).

This equation (i) represents an ellipse, parabola, or hyperbola,

2 2
according as cos 6 tan v
2
sin 6 is o, that is according as >
= <

tan #<=:> tan v.

2. For the cylinder of revolution about the axis of z, we make

x=x cos 0,y y in its equation x*+y* = r*; /, the curve of in

tersection is a-
2
cos
2
04-/
2
= r* an ellipse.

EXAMPLES (Coordinates Rectangular).

1. Find the right line generators of the hyperboloid

*+>_:=,
9 i 4

for the point (2, 3 ?) on the surface.

2. Find the right line generators of the paraboloid 4jy

2
2$z
2
= icxxr
for the point (? 2. i) on the surface.

3. Find the planes of circular sections of the following surfaces :

36 (0
=i44 (2)
i2 (3)

(4)

In the hyperboloid of revolution of one sheet

^ ---g =
-f. j;
2 2

4. f r i

find the equations of the generating line whose projection on the

plane xz is tangent to hyperbolic section in that plane at its vertex.

2
5. Find the sections of the cone x*+y*=(z 2) by planes con
taining the axis of j/, at angles to the plane xy of 30, 45, and 60
respectively.

6. Find the curve of intersection of the surface

by a plane inclined at an angle of 30 to the plane xy, and whose

trace on that plane makes an angle of 45 with the axis O.v.
 CHAPTER X.

TANGENT PLANES, DIAMETRAL PLANES, AND

CONJUGATE DIAMETERS.

meeting surfaces of second

123. Straight line order.

We can transform the general equation

to polar coordinates by writing x = /r, y mr t

z nr, (when /, m, n
are in rectangular coordinates, direction cosines, and in oblique co

ordinates, direction ratios). The equation becomes

r*( A/
2
+ Bfl* + Cri + 2K ?nn + 2 Win + 2C7m)
2 2

+C" + F= o.

Hence a straight line meets the surface in two points, and if these
two points be coincident the line is tangent to the surface.

124. Tangent Plane to surfaces of second order.

Let the origin be on the surface (and therefore o) then one F=

of the values of r in (2) is r o. =
Now, in order that the radius
vector shall touch the surface at the origin, the second root must be
o, and the condition for this is A"/+B"w + C"n o. Multiplying
this by r and replacing Ir, mr, nr by x, j>, z, this becomes

= o. (3)

Hence the radius vector touching the surface at the origin lies in the
fixed plane (3); and as n are arbitrary, A",r + B j + C"2
/, ;;/,
c is =
the locus of all the radii vectores which touch the surface at the

origin, and is therefore the tangent plane at the origin.

Hence, if the equation of the surface can be written in the form

+ != o (where u.2 represents terms of second degree and u terms
{

76
 NOTES ON SOLID GEOMETRY. ^
of first degree in #, y, and z), then u l
= o is the equation of the tan

gent plane at the origin.

Therefore, to find the equation to the tangent plane to the surface at the
r

point x y z transfer the origin

,
to this point. l he equation may then be
writ/en u2 -f Uj = o, and Uj =o is the tangent plane referred to the

point of contact as origin ; then in Uj= o retransfer the origin to the

primitive one.

125. For the central surfaces (origin at centre) take the equation

and let (x ,y ,
z )
be the point of contact. Transferring the origin to
r
the point (x , y ,
z ) by the formulae
X = X+X \

y y +y V we have
2= Z + Z 1

Ax* + B/ + Os 8 + 2 Axx + 2 fyy + 2Czz =o.

Hence the tangent plane at the new origin is

Axx -f By/ + Czz = o. (

i
)

Now retransfer the origin for equation (i) to the centre by the
formulae
x =x x }

y =y y \ and we obtain
*=*-* )

+ Czz Ax *
Bi/
2
- CV a
= o,

or AAVT + BJ/ + Czz =i (2) the required equation of the tangent

plane, at the point xy z referred to centre.

i. For the sphere A =B = C = -^ .

a
Hence (2) gives xx +yy + zz = a\ (3)

2. For the ellipsoid A= -i

-, B = -1 C= i

xx yy zz
_

3. For the hyperboloid of two sheets A= ~ E~ ^ C= \

xx yy zz

7*
78 .VOTES ON SOLID GEOMETRY.

4. For the hyperboloid of one sheet, A= ,

B=: ,
C= ^.

xx yv

126. For the surfaces which have no centre (origin at vertex) by

equation By + Cz* = 2A. x in a similar manner we
2
treating the
obtain
Bj/y + Oss = K (x + x ) (7) for the equation to the tangent plane
to the elliptical paraboloid and
Bij/ Czz = A. (x + x ) (8) for the tangent plane to the hyper
bolic paraboloid.

Remark. The same method may be applied to cones and cy

linders.

127. Polar planes to surfaces of second order. The equations (3)

(4) (5) (6) (7) (8) are the equations to the polar planes to the sur
f

faces respectively with respect to the point y z ) and these polar (.v , ,

planes possess properties analogous to the polar lines to the conic

sections.

128. The length of the perpendicular from the centre on the tangent

plane to the ellipsoid is p = \Az cos

8 2
a + &* cos 2
ft + c* cos y 2
,
when
cos a, cos (3, cos y are its direction cosines.

The equation to the tangent plane is

"-^
+ "-jrjH
g-
= i. It may

also be written x cos a +y cos ft + z cos y p. Hence we must

have

/~ cos a cos /? cos y a cos a _ b cos ft _ c cos y _

i x y z x y _s/
~J ^~ ~^~ ~a~ ~T ~7

Hence calling the direction cosines /, z, w, the equation of the

tangent plane may be written

Ix 4- my -f .s = V^ + ^ ^ 2 2 2
+^ 9 2
(9 )
 NOTES ON SOLID GEOMETRY. 79

129. To find the condition thai the plane

3C \f 2
f-
-~ H =i (i) shall be tangent to
a ft y

the ellipsoid
x*
+ ~-r
/ z*
r
5- -\
a u c

xx yy 22
-
Comparing (i) with ^ + -^ -\

we must have
I
"
_ X I
_ y I
o~ 7T~ /o i

a 2 j

ft 6*
y
a x b c z
- = -^-, y - =
or
^f a
,

ft
r
by c
;
/. squaring and

adding r + ^ H r i is the required condi-

<^
p y
tion.

130. The sum of the squares of the perpendiculars \>, p , \>" , from the

centre of the ellipsoid on three tangent planes mutually at right angles is

constant.

Let cos a cos ft cos y\ cos <* cos ft cos j/, etc., be the direction
cosines.

Then p
2
= 2
cos
2
-f <$
2
cos
2
/5 +<;
2
cos
2
y
p
*
= a" cos 2
a +b~ cos
2
/? +t f
cos
2
y
p
2
= <z
2
cos
2
a" +^ 2
cos
2
ft"
}- c
1
cos T/
2
,

and adding we have

131. Cor. Hence the locus of the point of intersection of three tangent

planes to the ellipsoidwhich intersect at right angles is a concentric sphere

of the radius V + b* + c V 1
.

For 2 the square of its distance from the centre is equal to

p*+p *+p and therefore to a* + tf + c\
"\

Remark. In the case of hyperboloids one at least of the quantities

a2 .
b~, c
2
is negative, and hence their sum may be negative or nothing ;

in the former case there through which three is no point in space

rectangular planes touching the hyperboloid can be drawn, and in

the latter case the centre is the only point which has that property.
8o NOTES OA SOLID GEOMETRY.

132. Diametral Planes. Definition. A diametral surface is the

locus of the middle points of a series of parallel chords of a given
surface. Diametral lines or diameters are the intersections of the
diametral surfaces.

133. To find the diametral surface corresponding to a given series of

parallel chords in a surface of the second order which has a centre.

Let the equation of the surface be

/, m, n the direction cosines of each the parallel chords, and # , /, z

the coordinates of middle point.
its

The equation of the chord will be

x y v z z
oc
m =: zn i*.
I m n

Then for the points in which it meets the surface (i) we shall

have

or

Imposing on this the condition of equal roots for r, we have

Atx + Bniy + Cnz = o (2) the equation of the diametral surface, a
plane passing through the centre.

134. The diameter = = is one of the series of parallel

/ m n
chords bisected by the plane (2),
and is called the diameter conju
gate to the plane, and conversely the plane lx + my + nz o is con-
A.v Bv Cz
luxate to the diameter
- = = -.
I m n
be chosen as a new plane of xy and its con
If a diametral plane

jugate diameter be taken as the new axis of z, the centre O being still
the origin; then, since every chord parallel to Oz is bisected by
the plane xy, the equation of surface will contain only the second

power of z. Hence, if there be three planes through the centre the

intersection of any two of which is conjugate to the third, the equa
tion of the surface referred to these planes will be of the form

AV + By + CV=i, (3)

that is of the same form as the equation referred to rectangular axes.

 NOTES ON SOLID GEOMETRY. 81

!35- To find the conditions that of three planes through the centre of a
diametral to the intersection of the
surface of the second order each may be

other two.
Let the planes be

/v _|_ mv + nz = 0, I x + my + n z = o, l"x 4- m y+!

n"z = o.

The equations of the diameters conjugate to the first plane are

and if this be parallel to the other two planes, we shall have

/ m n I ... m n
oandr + ,,
. .
, ,,
/ +M + A
"

c= 0;
A B "c= B+"

these with the third equation I +m + n o, found in

likemanner, are the required conditions.

These three planes are called conjugate planes, and their intersec

tions conjugate diameters.

Since we have only three relations between the six quantities there
will be an infinite number of systems of conjugate planes in each
surface.

136. Equations referred to conjugate diameters. If in (3) Art. 134 we

make
A _ >_

a v
B
i
,
_JL
- v
b
,
c--
^- -&
Then for the ellipsoid
2 2
X~ 2
+
I

-f = i will be the equation referred to conjugate di

ameters, and a, b ,
c will be the semi-conjugate diameters.
For the hyperboloids we shall have

a-
2
y z* , x 1
y z
2

^-^- 7T= i and

jr^-^bc r.

Remark. The tangent planes at the extremities (.* j/, z) of any ,

diameter to a central surface are parallel to the diametral plane

conjugate to the diameter so that the conjugate plane of the diameter
1

through the point (x , y ,

z }
on the ellipsoid is

xx vy zz
82 NOTES ON SOLID GEOMETRY.

137. The sum of the squares of three conjugate semi-diameters

of the
ellipsoid is constant.

In the first place, any point on the ellipsoid may be represented by

the equations x =. a cos A, y b cos //, z c cos Y, when cos A, = =
cos ju, cos v are the direction cosines of some line, for the condition
2
cos A + cos 2 //-fcos 2 v i cause these three
equations to satisfy the
equation of the ellipsoid.
Therefore if cos A, cos fa cos r, cos A cos //, cos v are the direc ,

tion cosines of two lines answering to the extremities of two conju

gate diameters, these will be at right angles to each other.

T, xx yy zz
ror the equation
+~j^- + g-= o will give

cos A cos A + cos yu cos //4-cos v cos v = o.

Now
the square of the length of any semi-diameter ~v
2
+y +V2 2

expressed in terms of A, /i, v, is

rt
2
= a* cos 2
A+ <
2
cos
2
JJL + C* cos
2
r,

and of the conjugates in terms of A , //, v\ A", // ,

y"

2
2
=rt cos A 2 2
+ 2
cos
2
yu+r cos r
2
// +
2 2 2
<:"*= a* cos A"+3 cos r> cos v".

Adding we have
2
a"
+ 3 /2 + c 2
= a + tf + 2
<r
2
,
since the lines A, ju, r, A , // ,
r x and
,

A", yu",
y" are mutually at right angles.

138. To find the locus of the intersection of three tangent planes at the

extremities of three conjugate diameters.

The equations of the three tangent planes are

cos A + vo cos u, -f - cos v i

a c

V Z
OC
cos A + cos v -\-
- cos v = i
(2 O C

x
- y ,,
z
COS A -f COS V -\ COS V I .

a b c

Squaring and adding, we get for the equation of the locus

H ~-f--r = 3 an ellipsoid with the semi-axes a\/~$~, b^/ 3 ,

>

tV~T-
 NOTES ON SOLID GEOMETRY. 83

139. The parallelepiped whose edges are three conjugate semi-diameters

of an ellipsoid lias a constant volume.

Let Qx, Oy, O.2 be the semi-axes of the surface a, b, c\ Oo; ,

Oy,
r
Oz any system of semi-conjugate diameters a b , , c\ let the plane
of x y intersect that of xy in the semi-diameter OjCi A, and let
Oy 2
= B be the semi-diameter of the curve ^# which is conjugate to
CXrj. Hence parallelogram a b = parallelogram AB.

.-. Vol (a, b\ f

)
= Vol (A, B,-O
for these figureshave the same altitudes and equal bases.
Let the plane z Oj 2 intersect xy in the semi-diameter Oi^ C,
then this plane must contain Oz for, being conjugate to OA*, in a ;

principal plane it must be perpendicular to that plane hence CXv,, ;

Or,. Oz form a system of semi-conjugate diameters, and any two of

them are semi-conjugate diameters of the plane section in which they
are situated.
.-. Vol (A, B, c
)
=Vol (A, C, c)
Vol (A, C, c) =Vol (a, b,-c)

. . Vol (a
1

,
b ,
c )
=Vol (a, 6, c).

140. To find the diametral plane bisecting a given system ofparallel

chords in the case of the surfaces which have not a centre.

Taking the equation of the surface

and one r
or the
,
chords -
, , xx = vy- z
=. ~
z
- =r
I m n

the equation of the diametral plane will be

;;/ BJ/ + W Cz= A 7.

Hence the diametral planes are parallel to the common axis of the

principal parabolic sections.

Wecannot, therefore, in these surfaces have a system of three con
jugate planes at a finite distance, but we can find an infinite number
such that for two of them each bisects the chords parallel to the other
and to a third plane,
by proceeding as in Art. (135).
By takingthe origin where the intersection of these two meets the

paraboloid, and referring to these three planes, the equation of the

surface will be of the form
84 NOTES OiV SOLID GEOMETRY.

And the third plane is evidently the tangent plane to the surface at
the new origin.

141. The tangent planes to the hyperboloid of one sheet and /he hyperbolic

paraboloid at a point x y z intersect the surfaces each in two right line

generators through the point of contact.

The equation of the hyperboloid of one sheet referred to any con
jugate diameters is

and the equation of the section made by any plane y = ft parallel

to the conjugate plane of xz, is

and it is evident that the value /3b gives us the section of the
tangent plane at the extremity (.v , y ,
z )
of the diameter b ;
or

x 3
z"*

~~^ 7? >
two right line generators.
Ct c

For the hyperbolic paraboloid

8
B>
-CV=2E".,v (i)
the tangent plane through the origin is x = o, and its intersection
with (i) is

B^
2
CV= o, two right line generators.
 CHAPTER XL

GENERAL EQUATION OF THE SECOND DEGREE IN

x,y, AND z.

142. In order to discover all the surfaces represented by the gen

eral numerical equation

2
Bj/

Gz 2
+2C ^ + 2C"a =D (E).

we will first transform the coordinates to a new origin by means of

the formulae
x =. a + x* \

and endeavor to determine the coordinates (a, /?, y) of the new

origin in such manner as to cause the terms of the first degree to dis
appear. If this can be effected the equation will be reduced to the
form
Ax 2
4- Br + Cs + 2A. zy + 2%
2
zx + 2C xy = F (F)

in which there is no change when x, z are substituted for

y,
+ x, +y, +z, and which therefore represents a surface having a
centre, and the new origin of coordinates is at this centre.
Now, several different cases may arise according to the numerical
relations among the coefficients A, B, C, A x
,
B ,
C , A", B",
C".

i. a, ft, y the coordinates of the centre may each have a finite

value found from the three equations determining the conditions of
the transformation.

2. a, /3, y may have infinite values.

3. a, /?, y may be indeterminate.

8
85
86 A 07^ES
T
ON SOLID GEOMETRY.
The surfaces corresponding to these three cases will be

(A) Surfaces having a centre.

(B) Surfaces having no centre (centre at an infinite distance).
(C) Surfaces having an indefinite number of centres.

143. Making the actual transformation of (E) by the formulae (i

we have

And in order that the terms of the first degree in x, y, z shall dis
appear, we must have

(C)

which are called the equations of the centre.

1. If these three equations give finite values for a. /?, y, then
the surface represented by the given equation has a centre.

2. If two of ihese equations are incompatible this shows infinite

values for a, /3, y, and the surface has no centre.

3. If the three equations reduce to two, then the surface has a

line of centres. For each one of the equations is the equation of a
plane, and two taken simultaneously represent a line, and the surface
is an elliptical or hyperbolic cylinder. For, cut the surface by the

planes P and Q, P cutting the line of centres (D)

and Q containing
that line. The section by P is a curve of the second degree having
its centre on the line D, and hence an ellipse or hyperbola. The
section Q will be two straight lines parallel to the line D, and as Q
may revolve about D positions giving two straight line
in all its sec

tions parallel to D, the surface is a cylinder.

surface
4. equations reduce to a single one, then the
If the three

has a plane of centres (i. e., the given equation represents coincident
or parallel planes).

Note. The equations of the centre can be found in any given

 NOTES ON SOLID GEOMETRY. 87

equation most readily by finding the derived equations with regard

to x, y, and z respectively (i.e., by differentiating with regard to x. y, z
respectively), the x, j ,
and 2 in the resulting equations standing for

a, ft, y.

144. Example i. Determine the class of the surface represented

2
by the equation ,v + ^y + 42 + 2yz + \zx + 6.17
1 2
2 6x 2$y 322= 26.
The equations of the centre are

6> + 22 + 6x 24 =O > . These give y =2 I

82 + 27 + 4^32 = 0) *=3J
and the surface has a centre.

Example 2. Determine the class of the surface

The equations of the centre are

+ 22+2 V 4 =
2JI- O )

y + 22 + 2X 2 =
f
x
2 O ,
-

4 Z + 2y + 2X + 2 = O )

the first two of which x+y+z = 2, x+y + z = I are incompatible,

hence the coordinates of the centre are infinite, and the surface has
no centre.

Example 3. Determine the class of the surface

2
2
2yz zx-\-^xy+2z = o.

The equations of the centre are

2x z +4y ~ o
8j/ 22 + 4^v = o
2Z 2V X+ 2=OJ
The firsttwo of these are identical, hence the three equations re
duce to two and the surface has a line of centres (/. e. is a
cylinder). }

Example 4. Determine the class of the surface

+ 75 =
2 2
8-v + 1 8v
2
+ 22 + 1
272 + 8zx + 2^xy 5o.v 757 252 o.

The equations of the centre are

1 6x + 82 + 247 50 = o
=
i2r+ 8.v 25 = o
88 NOTES OX SOLID GEOMETRY.

which are all three the same, each being 8.r+ 4-4.3 =25. Hence i2>

the surface has a plane of centres, and consists of a pair of parallel

planes.

145. Recurring to the general equations of the centre

A <*4-C7? + By + A"= o ]

+ ir=o( (c)

we may an easy rule for a relation among the coefficients in any

find

given equation by which we can distinguish the central surfaces from

those having no centre and those having an infinity of centres.
The common denominator of the values of a, /3, and y in these

equations is the determinant

A, C B ,

C B, A
,
2
+ CC -ABC-2A B C
2
.

B A C
, ,

Now, if R be different from zero, the surface has a centre ;

but if

R= o it
may either have no centre or an infinity of centres.
The value of R may be written out by the following mnemonic
A, B, C
the letters to be multiplied by columns for the first
form : A B C
three terms, and by rows for the two last.
A B C
146. To find an easy rule for F, the new absolute term in the trans
formed equation of the central surfaces when the origin is moved to the
centre.

This complete transformed equation is

A_v s
+ By + Or + 2 A zy + zR zx + 2C xy = F when
2

F=D-

Now, multiplying the first of the equations (C) of the centre by

a, the second by ft,
and the third by y, and adding them

we have
 NOTES ON SOLID GEOMETRY. 89

Hence F = D (A"
+ B"/? +C . Therefore the rule for F is

substitute for x, y, z in the terms of the first degree one-half the co

ordinates of the centre (i. e., ^a, |-/?, ^y respectively and take result

from D.
Example i. Taking the Example i, Art. (144), in which the coordi
nates of ihe centre are found to be x =-i,y 2, z = 3,
we have F=26 + 26 x J- + 24 x +32 i xf 1 1 1;

and the transformed equation is

+ 43 + 2yz + 4zx -f
X* + =
3>
2 2
6,r>
1 1 1 .

Ex. 2. 2^ + 3y* + 4Z + Syz + 6xz + 4xv 6x SyI4z = z

20.

Here the coordinates of the centre are x = = 2, s= \>y

i.

. . F==2o + 6 x-J + 8 x i + 14 x 1 = 17;

and the transformed equation is

3/ + 4-s + 870 + 6^2- + ^xy

2
2 A* 4- 1
7.

147. Removal of the terms in xy, xz, yz. Reduction of the equation of
the second degree to two forms.
Fora more complete discrimination of the surfaces represented by
the general equation, we will now remove the terms in^ry, xz, yz by
a transformation of coordinates. So far we have made no supposition
as to the direction of the axes. Henceforth, for convenience, we will
consider the axes rectangular.

Taking the equation (E) in rectangular axes we propose now to

transform it to a system also rectangular in such manner that the
terms in xy, xz, yz shall disappear. The disappearance of these
terms can only be effected by taking for coordinate planes either dia
metral planes or planes parallel to them.
We will therefore begin by finding a diametral plane conjugate to
a given diameter.

148. To find a diametral plane conjugate to a given diameter.

x x yy __ zz
I m n

Putting x-= x + lr, y y + mr, z=z -\-nr in the general equation,

and arranging with reference to r, we have for the coefficient of the
first deree in r
QO NOTES ON SOLID GEOMETRY,

and this placed equal to zero is the equation of the diametral plane,

namely

( A/+ B + C m)x + (C7+ Bm + A n}v + (B7+ A m + Cn)z + A 7

149. 70 determine a diametral plane perpendicular to the chords which

it bisects, that is, to find a principal plane.
In order that the diametral plane shall be perpendicular to the line
x a v b z c
we must have the conditions fulfilled
m n
,

I m
or putting each of these equal to s.

At +R n+C m =
(A)

and also the condition /

2
+w 2
-f #2 = i .

To determine /, w, and n in equations (A) we first find s. Writ

ing these equations

they give the result

A-J, C B ,

C B j, A
,
=o
B A C-.r
, ,

or

or

C *-ABC-2A B C f
=o (D).

This cubic has necessarily one real value for s, which substituted
in (A) gives one set of real values for /, m, n. Hence there is one
principal plane.
For convenience of discussion let us take this plane perpendicular
 NOTES ON SOLID GEOMETRY. gi

to the axis of z, then / = o, w = o, and n =i. And hence equa

tions (A) give B = o, A= and the general equation transformed
o,

to this principal plane as plane of xy is of the form

z = D.

Now we know from the like discussions in conic sections that one
transformation always possible, and but one to a system of rectan
is

gular axes in the plane xy which shall cause the term in xy to dis
appear. Hence there are three principal planes, and three sets of
values for /, m, n, and the cubic (D) has three real roots.

The general equation may then be always reduced in rectangular

coordinates to the form

* = D. (E )

which represents then all the surfaces of the second order.

1
50. The reduction of this equation Lx 2 + My 2 -f Nz + 2L 2
x + 2M y
+ 2N z =D to two forms.

i. If L, M, and N are different from o.

Then we may cause the terms of the first degree to disappear by

= L
= M N
transferring the origin to the point x ~r~>-J ivf
*
\f
1 4 i* L IN

The surface will then have (# ,.> ,

z )
for its centre, and the equation
will be of the form

Lr + My + Ns ^ 2
F. (I.)

2. If one of the three coefficients, L, M, N, for example L = o

and L be different from o.

We cannot then cause the term 2~L!x to disappear, but by trans

ferring the origin to the point

D
= --., M = --N
x
2L
y= M z ^p-
N
the equation will take the form

M/ + Ns = f
2V x. (II.)*

The forms I. and II., we have seen, belong to the surfaces of the
second order, which we have already discussed. Hence the general
equation of the second degree (E) represents these surfaces and no
others.
QO
NOTES ON SOLID GEOMETRY.

151. The form

I. we have seen represents the ellipsoid, the two
hyperboloids and cones of second degree, and includes the elliptic
and hyperbolic cylinder, Mv 2 + Ns 2 F and parallel planes N2 2 =F. -

The form II. represents the elliptic and hyperbolic paraboloids,

and the parabolic cylinder.

152. The complete reduction of the equation of the second degree to the
simple forms I. and
Use of the discriminating cubic (D).
II.

The resolution of the equations (A) furnishes for each value of s

in the cubic (D), one system of values of/, m, n. We have then three

systems, /, m, n] /,
which are the direction cosines
,
??i ,
n /", m",
n
of the three rectangular axes {principal axes ] to which the surface
must be referred in order to cause the products .%T, xz, yz to dis

appear ;
the formulae of transformation are then

y mx + my + m"z

z = nx + n y + n"z .

If we take only the terms in x* in this substitution we find

L = A/ + Bw + C;* + 2h. mn + 2BW+ zC lm.

2 2 2

But if we multiply the equations (A) respectively by /, m, n and add,

remembering that /
2
+ 7
2
+ 7z
2
= i we have
A/ 2 + B;;;
2
+ CV + 2h. mn + 2 BW+ 2C lm s ;

Hence L is a root of the cubic (D) and M and N are the other two
roots.
For the values of L ,
M N ,
we will have

L=A7 +B"w +C"n \

M=A7 +B";;/
r

+C V (M).

The absolute term D does not change in this transformation since

the origin is not changed thereby.
For the surfaces having a single centre after solving the cubic, we
have only to calculate F, for which we have given a rule.
For the surfaces having no centre the coefficient designated by V
is equal to L and is computed by first finding in equations (A)
,

the values of/, m, n, which correspond to s o. Both in the cases =

of surfaces having no centre and one root of cubic a line of centres,
= o and we have only quadratic
a to solve to determine L and M.
 NOTES ON SOLID GEOMETRY. 93

For surfaces having a centre, if \ve wish only to discover the

153.
particular class of the surface, without making the complete trans
formation of the equation to its centre and axis, the sign of the roots
of the discriminating cubic willtell us whether the surface is an ellip

soid, hyperboloid of one sheet, or hyperboloid of two sheets. These

* without
signs we can ascertain from inspection by Descartes s rule
solving the equation.
2 2 2
Example. Find the nature of the surface 7# + 6y +5.s 4_> $xy
= 6. The cubic (D) gives

s
3
(7 + 6 + 5)s + (42 + 35 + 30 -4 4)5+28 + 20
a
210=0; or
3
s i8s* + 99^162=0.

.-. The row of signs is H 1

,
three changes of sign. Hence
all the roots are + and the surface is an ellipsoid.
So also for surfaces having a line of centres, the signs of the roots
of the quadratic into which the discriminating cubic degenerates,
serve to distinguish the elliptic from the hyperbolic cylinder.
And for surfaces having no centre, the signs of the roots distinguish
the elliptic paraboloid from the hyperbolic paraboloid.

154. Recapitulation of the method of reduction of numerical equations

of the second degree and of distinguishing the surfaces represented by them.
We
now propose to give the mode of distinguishing the nature of
the surface represented by any given numerical equation of the second

degree in x, y, and z, and of finding its principal elements.

I. Form the equations of the centre, and also the discriminating
cubic from the remembered form
s *- (A + B + C)s +
z
(
AB + AC + BC - A - B -C + AA + BB
2 2 2
).r
2 2

+ CC -ABC-2A B C = 2
o,

observing that the absolute term is equal to R, the denominator

of the values of the coordinates of the centre in the general equation,
ABC
and therefore can be formed by the mnemonic ABC (Art. 145).
ABC 1

Then
155. i. If R be different from o, the surface has a centre. Find

Note. the roots being real the number of positive roots is equal to the number of
"All

changes of sign row of signs of the terms, and the number of negative roots is equal to the
in the
number of continuations of sign."
94 NOTES ON SOLID GEOMETRY.
the coordinates of the centre and transform to the centre by the rule
in Art. (146). Determine the signs of the roots of the cubic by Des-
cartes s rule. Then calling these roots L, M, and N, and calling F
the new absolute term on the second side of the equation.
Then
a. If L, M, N all have the same sign as F, the surface is an ellip
soid.

b. If L, M, N all have a different sign from F, the surface is im

aginary.
c. If two only of the roots L, M, N have the same sign as F, the
surface is the hyperboloid of one sheet.

d. If only one of the roots L, M, N has the same sign as F, the

surface is the hyperboloid of two sheets.

e. If F = o and L, M, N all have the same sign, the locus is a

point.

f. If F =o and one of the roots L, M, N has a different sign

from the other two, the surface is an elliptic cone (Art. 85).

156. 2. If R= o the cubic has one of its roots s = o and is

degraded to a quadratic, the coefficient of s, namely AB + AC + BC

A 2
B 2
C 2
,
becomes the absolute term.
And if the equations of the centre are incompatible the surface has
no centre.
Then
a. If the roots M and N of the quadratic (degenerate cubic) have
the same sign (i. e.)
if AB + AC + BC A2 B 2
the surface *C >o

is the elliptical paraboloid.

b. If M and N have different signs (/. e.)

if AB + AC + BC-A"
B 2
C *<o the surface is the hyperbolic paraboloid.

c. If one of the roots M or N be zero (*

.
e.)
if AB + AC + BC-A 2

B 2
C 2
=o the surface is the parabolic cylinder.

3. If R = o and the equations of the centre can be reduced

157.
totwo equations, the surface has a line of centres. The cubic as in
(2) has one of its roots S = o and degenerates into the quadratic

j_(A+B + C)s + AB + AC + BC --A"-B -C" = o.

Then
a. If the roots M and N of this quadratic have the same sign (i. e.)
 NOTES ON SOLID GEOMETRY. 95

if AB + AC + BC A" B 2
C ~> o the surface is an elliptic cy
linder.

b. If the roots M and N have different signs (i. e.) if AB + AC

+ BC A 2
"

B 2
C 2
o the surface is the hyperbolic cylinder.
<

c. If in the reduced equation of the cylinder Mz 2

-f Ny 9 = H, H be
equal to o, and M and N both of same sign, the locus is a straight
=oz }
line
J=o
>
.

d. If H = o and M and N be of different signs the surface con

sists of intersecting planes.

158. 4. If R =
o and the equations of the centre become a
single equation, the surface has a plane of centres, and consists of
two parallel or coincident planes, which are readily found by solving
the equation with reference to any one of the variables.

159. 5. In the case of surfaces of revolution the cubic has equal

roots. To examine the cubic for equal roots in the case of central
surfaces of revolution, we simply look for a commoe root between it
and its first derived equation (differential).

1 60. GENERAL REMARK. In any of the above cases we may com

plete the reduction by solving the cubic to get the new axes and
thus obtain their direction by finding /, m, n from equations (A).
And in the case of the surfaces without a centre we may find V,
from equations (M).

161. REMARK I. In the cases of surfaces having a line of centres

and of those not having a centre, we can distinguish readily the sur
face represented by a given numerical equation through sections by
the coordinate planes.

i. If the equations of the centre show a line of centres, sections

by the coordinate planes will tell whether the surface is an elliptic or
a hyperbolic cylinder.

2. When the equations of the centre show no centre, then

a. If there are ellipses among these sections by the coordinate

planes, the surface is an elliptical paraboloid.

b. If there are hyperbolas among these sections, the surface is a

hyperbolic paraboloid.
c. If all these sections are parabolas, or one of them parallel
straight lines, the surface is a paraboMc cylinder.
96
NOTES ON SOLID GEOMETRY.
162. REMARK II. Again, if the terms of the second degree in the

given equation break up into unequal real factors, the surface must
be either the hyperbolic paraboloid or hyperbolic cylinder, and these
two surfaces are otherwise readily distinguished. We may note also
that if the terms of the second degree in the given equation form a

perfect square, the surface is either a parabolic cylinder or two

parallel planes.

163. We will now illustrate by a few examples :

Ex. i. yjt-
2
iy~ + 6z + 2^xy+ \2yz- 1202:
i
= 84.

As this is a central surface with the origin at the centre, we only

need the discriminating cubic, which is

j
8
343^4-2058 = o ;
or s*os* 343^4-2058 = o.
The
signs 4-
-- 1- show one continuation and two changes, and
hence the surface is a hyperboloid of one sheet, or two sheets, accord

ing to the sign of 84.

By trial we find that 7 and then by depress
is a root of the cubic,
two roots are 14 and
we find the other21. There
ing the equation
fore the equation of the surface referred to its centre and axes is

7* -f 14^-212*= 84; or.r

2
+ 2_y -3S =:i2.
2 2

Ex. 2. 2 5.r
2
+ 2 2_/ + i6s 2
4- i6yz4zx 2oxy z6x 40^442
= -46.
The equations of the centre are

250, ioy 22 = 13

I o:r 4- 2 2j 4- 82=20
-
2X+ 8y+ 162 = 22 ;

whence we find the coordinates of the centre x \,y =i, * =i.

Moreover

F = 46 4- 26 \ + 40 J + 44 I
= 9-

The discriminating cubic is

/ 63^+ 134^5832 1 = o.

Its signs give three changes.

Hence all the roots are positive.

The surface then is an ellipsoid.

 NOTES ON SOLID GEOMETRY. 97

By trial we find that 9 is one of the roots of the cubic. Hence the

other two are 18 and 36. The reduced equation is then

2 2 8
9jr
J

-fi8> -f 36s 9 ;
or A-

And the principal semi-axes are i

-
, __
V2 2

Ex. 3. 5 x* + of +i i
ys
2
+ 2 6yz + 1 8zx + i $xy + 6x + ?>
+ 1 02 = 64.
The equations of the centre are

5.r+ 7^+ 92 =-3

7A:+iqy+ 132 =4
Multiplying the first of these equations by i and the second
by 2, and adding, we obtain the third. Hence the equations are

only two independent ones. The surface is therefore a cylinder. In

tersecting it by the coordinate plane xy, i. e., making z = o, we obtain

5^
2
+ 1 4xy + i oy* + 6x + Sy 64,

which is an ellipse. The surface is therefore an elliptic cylinder.

To complete the reduction we transfer the origin to the point

where the line of centres >

pierces the plane

x, y, that is, to the point z o, y i, x 2,

and find F=64 +6 4=:66.

Also the discriminating cubic is

.r
3
32,r-f6 9 >
= o, which gives j
8
32^ + 6 = o,

the roots of which are 16 + 5^ 10 and 16 5\/ 10.

And the reduced equation of the cylinder is

7 2
io)jr + (i6 5V io)>
=66.

Ex. 4 .
5 jp
2
+ 5/
J
+ 8s + 4sv + \zx
2
Sxy + 6x + 6y 30 = o.

The equations of the centre are

^4-4*=
98 NOTES ON SOLID GEOMETRY.

Adding the two last of these equations we have

$x4y+2z 2j. An equation which is
incompatible with the
first. Hence the surface has no centre.

The cubic / iS>r

2
+ 8ij= O ;
or s*i 8^ + 81 = o,
which gives two roots equal to 9. The surface is therefore a para
boloid of revolution
2

To find V, we first determine /, m and n. For these we have the

equations
4/ ^ 272 = o
o

U- U
which 7
Rive / = 2
, m = 2
n -- !

3 3 3

Therefore (Eq. M) L = 3 .-+ 3 .-+-.- = -

and 2V 2L =9.
The reduced equation of the surface is therefore

v
Ex. 5. 2A70+2B ^v+2C .r^ 4-2A .v + 2B jv4-2C"2;
D.

The cubic is

s-(A + B
2 2
f-C
2
)s-2A B C =o .

The surface is a hyperboloid if A B ,

and C are all different from
o. If A BC
of the same sign as F in the reduced equation the
is

cubic will have two roots of the same sign as F and the surface will
be a hyperboloid of one sheet. In the opposite case it would be a hyper
boloid of two sheets.

If A = o the cubic becomes

j.
2
_(B
2
4-C *) =o, whose roots are of different signs. Hence
the surface 2H zx + 2C xy-\- 2A".r+ 2B V + 2C
r
z = o is a hyperbolic

paraboloid.
2 2 2
Ex. 6. jc +_y 4-9s + 6r0 6xz 2~\y+ 2.v 42 o.

The equations of the centre are incompatible and the terms of the
 O.Y SOLID GEOMETRY. 99

second degree form a perfect square, hence the surface is a parabolic

cylinder.

EXAMPLES.

164. i. Find the nature of the surfaces represented by the follow

ing equations.

(
i
).
i i~v -f
2
5_y 4-
2
2o\ z + 4zx 4- 6xv 4- 2 2.v + 6 v 4- 42 4-
2z
2
= o. 1 1 1 1

(
2 +/
).
.X
2
+ 2X2 4- 2AJ O,V
4- S* 4- OV 2F
IOZ-\-2$ = O. I I

( 3 3 jc). 37 2J2T +i2zx + Sxv

2
6x 2
+ 3^ = 0.
1
6>>

60jt- + 54^ =36.

2
(4). 4.i- + QV + 97s
2 2
1

(5). 3X*+2f2XZ 4VZ 4.V 80 8 = O. \-

2. The equation 7^- 4-8/ + 4s 7^0 nzx jxy = ^ represents 2 2

a hyperboloid of one sheet.

3. +32: + yz+zx+xy
The equation jx 1
252 12 d .r
2
+>
;2
2
4>
=
represents an ellipsoid, a point, or an imaginary surface according
as d is <
= >
67.

4. The equation _v
2
4-.v
8 8
4-2 -i-yz+zx + xy = a* represents an oblate
spheroid.

5. Find the nature of the surface (y z)*+ (z xY + (x v)

2
= a~.

2
6. Find the nature of the surface yz \ zx + xy # .

7. ^a-
4-9s
2
+ 4v 2 2
o re
4- I2yz + 6zx + 4xy+i4x+ 16^+2404-47 =
presents an elliptic, a parabolic or a hyperbolic cylinder according as
a >
= < i.
 CHAPTER XII.

PROBLEMS OF LOCI.

165. PROB. I. To find the

surface of revolution generated by a right
line turning around a fixed axis which it does not intersect.
Let the fixed line be the axis of z and let the shortest distance a
from the revolving line to the axis of z lie along the axis of x in the

original position of this line so that its equation is x a, y= mz.

Then the equation of the surface is

or

The hyperboloid of revolution of one sheet.

Prob. 2. To find the locus of a point whose shortest distances from two
given non-intersecting, non-parallel straight lines are equal.
Take the axis of z along the shortest distance between the two
lines, the plane xy perpendicular to z at the middle point of this
distance 2c, and the axes of x and y bisecting the angles between the
projections of the line on their plane. Then the equation of the lines
will be
z =c \
z c )

y mx f y mx

-
}

y mxY V + mxV
rL =
xo
and we have cY +
(
+ +
(
~ ~-
(z (* 8
I + m~ i + ///

or

cz(i + ?n*) -f mxy = o, a hyperbolic paraboloid since it has no

centre and its term of second degree breaks up into two real factors.

Prob. 3. 7 wo planes mutually perpendicular, contain each a fixed

straight line. To find the surface generated by their line of intersection.

Take the axes as in Prob. 2. Then the equations of the planes are
100
 NOTES ON SOLID GEOMETRY. IO I

K(z c) +y mx o ; (i) K (z + t)+y + Mx = G. (2)

The condition of perpendicularity of these planes is

KK +i m* o,and eliminating K+K between this equation

and equations (i) and (2) we have

y _ /;/V + _ m *Y (
i
(
i
- m*y
which represents a hyperboloid of one sheet.

Prob. 4. To find the surface generated by a right* line which always

meets three fixed right lines no two of which are in the^santy fflantj *++

For greatest simplicity take the origin at the Centre of & paraljek)--
piped, and let its faces be at the distances a, &, d respectively from-
the coordinate planes yz, xz, and xy. Then take three edges of this
parallelopipedon as the three fixed lines fulfilling the conditions.

Assume for the equations of the movable line

*^.
cos a
m 2Z*
cos
-_=*.>
cos
(4)
ft y
The conditions that the line (4) shall meet the lines (i) (2) and

(3) are respectively

y b
_ z +c z c
_ x + a x a _ y -f b
~ ~~~
cos ft cos y cos y cos a cos a cos ft

Eliminate the arbitrages a, ft, y by multiplying the equations to

gether, and we have for the surface

or reducing

ayz + bzx + cxy + abc o,

which the discriminating cubic shows to be a hyperboloid of one

sheet. The same surface will be generated by a straight line resting
x =a } y = b } x= a
on the other three edges -

z = C
>

]
,
z c \
,

y=
.

Prob. 5. To find the surface generated by a right line which alw ns

9*
102 NOTES ON SOLID GEOMETRY.
meets three fixed right lines, no two of which are in the same plane, but all

of which are parallel to the same plane.

Take one of the fixed lines as the axis of x, and then the other two
parallel to the plane of A:; . Then their equations are

Now, the equations of a moving line meeting lines (i) and (2) are

- .

,-
(4) (/ and k arbitrary), and the condition that this

jine shall also meet (.3)

is Ic mk (cb],
arid eliminating therV and k by means of equation (4), we have

cy mx(cb)
2 z b

or cyz -\-m(b c)xz (

by o,

a hyperbolic paraboloid, as its equation shows no centre, and the

terms of the second degree break up into two real factors.

Prob. 6. To find the surface generated by a right line which meets

two fixed right lines, always parallel and is to a fixed plane.

Since the two fixed lines must meet the fixed plane, we can take
"

-y 1
y and the
(2), as in 2, as the fixed lines,
1
Z = c \
(i),
z = c [
}

plane yz as the fixed plane.

Then the equation of the moving line parallel toyz
Z
is (3), /, A and k arbitrary.

The conditions that this line shall meet the lines (i) and (2)

mk = k+/>

-mk=-lc+t\
or mk = Ic and/>
= o ;

or eliminating /, k, and /,

mx = y
c -- ;
z

or mxz = cy, a hyperbolic paraboloid.

 NOTES ON SOLID GEOMETRY. Ic>
3

Prob. 7. Two finite non -

intersecting non-faralld right lines are
divided each into the same number of equal parts ; to find the surface
ivhich is the locus of the lines joining corresponding points of divi
sion.

Let the line which joins two corresponding extremities of the given
lines be the axis of z let the axes of x andjy be taken parallel to the
;

given lines and the plane of xy be halfway between them. Let the
lengths of the given lines be a and b.

Then the coordinates of two corresponding points are

z =. c, x ma, y = o ;
z = c, x = o, y mb ;

and the equations of the lines joining these points are

ma
+ ^-=1
mb
]

2X
---Z =1
?na c

whence eliminating m the equation of the locus is

a hyperbolic paraboloid.

Prob. 8. To find the locus of the middle points of chords of a surface

of the second order that has a centre, which all pass through a given fixed
point.
Take the given point for the origin and two conjugate diametral

planes which pass through it for the

planes of zx and xy, and a plane
parallel to the third conjugate plane for that of> 2; then the equation
to the surface will be of the form

ax* + by 1 + cz + 1
2a"x
+/ o.

Let x viz, y = nz be the equations of any chord. Combining

these with the equation of the surface, we have

y
(am* + bn^ + c
)z + 2ct mz + d o,

in which the values of z belong to the extremities of the chord.

104 VOTES ON SOLID GEOMETRY.

Therefore the z of its middle point is

a"m

and the other two coordinates of the middle point are

x = mz t (2) y = nz .
(3)

Hence eliminating m and n the required equation of the locus

a surface of the second order similar to the first, and passing through
its centre and through the origin.
 CHAPTER XIII.

SOME CURVES OF DOUBLE CURVATURE.

1 66. To find the equations to the equable spherical spiral.

Definition. meridian of a sphere revolve uniformly about its

If a

diameter PP while a point M

moves uniformly along the meridian
from P to P",
so as to describe an arc equal to the angle through
which the meridian has revolved, the locus of M is the equable

spherical spiral.
Taking PP as the axis of z, PAP the initial position of the plane
of the meridian as the plane of xz, the equation of the sphere is

ac*4y +* = <!*.

Let MON = 0, AON q), then, by definition 6 = cp,

and from polar coordinates

x= a cos cos cp, r a cos 6 sin cp ;

= a cos 6, y = a cos
2
/. x sin 0.

**+/ = cos 2 + sin* = ax.

2
Therefore (cos 8)

Hence the equations of the spiral are

.r+y + = 2
(i) #" + v* = ax; (2)

or the spiral is the curve of intersection of the sphere and a right

circular cylinder whose diameter is the radius of the sphere.
If we subtract (2) from (i) we obtain
2
z~ <? ax (3) a parabolic cylinder.

And the equations (2) and (3) also represent the curve, which is
therefore also the intersection of a right circular and right parabolic

cylinder at right angles to each other.

167. To find the equations to a spherical ellipse.
Definition. The spherical ellipse is a curve traced on the surface
105
I0 6 A OTES ON SOLID GEOMETRY.

of a sphere such that the sum of the distances of any point on it

from two fixed points on the sphere is constant.
Let S, H
be the t\vo fixed points on the surface of the sphere
whose radius is ;-, C the middle point of the arc of the great circle
which joins them. IfP be any point of the spherical ellipse, SP and
HP arcs of great circles, then

SP-fHP = 2a = a constant.

Through P draw PM, an arc of a great circle perpendicular to

SH, and let SH = 2;/, CM = y, PM = 0.

Then, in the right-angled spherical triangle SPM we have

= cos
cos SP (y + q>)
cos 0.

And in the triangle HPM

cos PH = cos f cos 6.

2 J cos a
/SP-HPN
- Y sin
sin A q>
cos 6
sin V ?B .
.

V 2 / sin r

Squaring and addini

2 2
cos Y s V 2/3
cos1 Cr-xsi;
9 /) 2
cos cos c H ;
sin <z>

cos r.f sin- (y

or if we transform from polar to rectangular coordinates

o o
cos*
+ -r-r^y = ^ (0
c

This equation and the equation of the sphere

A-
9
+.!* + ** nr r2 (2)

determine the spherical ellipse, as the intersection of a right elliptic

cylinder and the sphere.

 NOTES ON SOLID GEOMETRY. 107

1 68. To find the equations to the helix.

Definition.Whilst the rectangle revolves uniformly about ABCM

its side so that the parallel side
AB, generates the surface of a CM
right circular cylinder, the point P moves uniformly along CM, and
generates a curve called a helix.
Let AB be the axis of and when the rectangle is in the plane
z,

xz let P and M both be at D

on the axis of x, and let the velocity of
P =n times the velocity of M.

.-. PM = ;/. arc. DM.

Also let AN = jv, NM y, PM = z be the coordinates of P, and
AM = a the radius of the circular base of the cylinder in the plane xy.

. . z = na cos"
1

,
and j
2
+ x^ <r

(i)

are the required equations of the helix.

Or we may represent the curve by the two equations

1
z na cos" . z na sin"
1

(2):
a a

or the same in the forms

z z
x = a cos , y = a sin , n) and
na na

z / z \ z / z
since cos cos 2m 7t -\ and sin -
= sin 2m n-\
na \ na J na \ na

the same values of x and y correspond to an infinite number of

values of z. The equations (i) (2) and (3) show that the projec
tions of the helix on the planes xz, and yz give the curve of sines,
and the projection on xy is the circle.
PLATE T
PLATE II
 7 DAY USE
RETURN TO DESK FROM WHICH BORROWED
ASTRONOMY, &ATHEMATICS-
STATiSTICS LIBRARY
This publication is due on the LAST DATE
1
1
-
and HOUR stamped below.
,

General Library
RB 17-60m-3, 66 University of California
(G1105slO)4188 Berkeley
 BERKELEY LIBRARIES

814023

-y
<#

UNIVERSITY OF CALIFORNIA LIBRARY