RATAN TATA LIBRARY

Call No.
\

Accession No. Date of release for loan

This book should be returned on or before the date last stamped
below* An overdue charge of 10 Patse will be collecied for each
day the book is kept overdue. ^
ANALYTIC GEOMETRY
AND CALCULUS
ANALYTIC GEOMETRY

AND CALCULUS
By

H. B. PHILLIPS, Ph.D., LL.D.

Professor of Mathematics
Massachusetts Institute of Technology

SECOND EDITION

New York: JOHN WILEY & SONS, Inc.

London
 Copyright, 1940 1941 1942
, ,

BY
Henry B. Phillips

Copyright, 1946
BY
John Wiley & Sons, Inc.

All Rights Reserved

This })ook or any part thereof must not

be reproduced in any form without
the written permission of the publisher.

SECOND EDITION
Tenth Printing, September, 1962

PRINTED IN THE UNITED STATES OF AMERICA

 PREFACE

It is the object of this book to present analytic geometry and

calculus in the formand order in which these subjects are required
for courses in scienceand engineering. In these courses calculus
is the subject that is most used. After a mere introduction of
graphical representation the book therefore begins with calculus.
In fact the first three chapters provide a complete elementary
course in the subject, including integration as a process of summa-
tion. Since these chapters contain the analysis that is most needed,
the physics courses taken simultaneously are enabled to make
free use of calculusalmost from the beginning. The result is a
saving of time which is becoming of great importance because of
the pressure to include additional subjects as well as more ad-
vanced material in the undergraduate program.
Because of the convenience of complex numbers and vectors in
the treatment of alternating currents, mechanics, and the analysis
of electromagnetic and fluid fields, an elementary treatment of
complex numbers and vector analysis is included.
I am indebted to Professor Philip Franklin for numerous prob-
lems and to W. S. Loud for many of the answers.

H. B Phillips
Cambridge, Mass.
February 19Jfi
 CONTENTS

6. CHAPTER I

LIMITS AND CONTINUITY

SECTION PAGE
1. Real Numbers 1

2. Points on a Line 1
3. Function 2
4. Independent and Dependent Variables 2
Functional Notation 3
6. Rectangular Coordinates 4
7. Limit of a Sequence 6
8. Limit of a Function 9
9. Angle 11
10. Limits of Trigonometric Functions 12
11. Properties of Limits 14
12. Infinity 16
13. Continuity 17
14. Some Fundamental Theorems 19

CHAPTER II

26. DERIVATIVE AND DIFFERENTIAL

15. Speed 27
16. Slope 28
17. Derivative 30
18. Differential 33
19. Significance of Differential 34
20. Approximate Value of the Increment 36
21. Differentiation of Algebraic Functions 37
22. Differentiation of Sine and Cosine 41
23. Small Errors 42
24. Derivatives of Higher Order 43
Implicit Functions 45
26. Sign of the Derivative 46
27. Sign of the Second Derivative 48
28. Maxima and Minima 49
29. The Second Derivative Test 62
30. Method of Finding Maxima and Minima 63
31. Other Types of Maxima and Minima 63
vii
viii CONTENTS
SECTION PAGE
32. Velocity, Acceleration, Rate of Change 66
36.
33. Related Rates 68
34. Mean Value Theorem 69
Indeterminate Forms 61

CHAPTER III

INTEGRATION AND SUMMATION

36. Integral 73
37. Constant of Integration 74
38. Integration Formulas 75
46. Variable with Given Rate of Change
39. 77
40. Curve with Given Slope 78
41. Area as Limit 80
42. Definite Integral 82
43. Evaluation of a Definite Integral 83
44. Area under a Curve 85
Volume of a Solid of Revolution 87
46. Volume of a Solid with Given Area of Section 89
47. Pressure 91
48. Work 92
49. Properties of Definite Integrals 94
60. Mean Value Theorem 96
61. Integral with Variable Limit 95

CHAPTER IV

ALGEBRAIC EQUATIONS AND GRAPHS

62. Equation and Locus 102
66.
53. Point of Division 103
64. Equation of a Straight Line 103
66. First-Degree Equation 105
56. Angles 106
67. Tangent Line and Normal to a Curve 108
68. Newton's Method 110
69. Angle between Two Curves Ill
60. Distance between Two Points 112
61. Equation of a Circle 113
62. Parabola 115
63. Ellipse 119
64. Hyperbola 123
Conic Sections 127
66. Transformation of Coordinates 129
67. Invariants 130
68. General Equation of the Second Degree 131
69. Graphs of Algebraic Equations 132
 CX)NTENTS

CHAPTER V
DETERMINANTS
SECTION PAGE
70. Determinants of the Second Order 146
71. General Definition of Determinant 147
72. Properties of Determinants 148
73. Minors and Cofactors 150
74. Solution of n Linear Equations in n Unknowns *.
152
75. Homogeneous Linear Equations 154

CHAPTER VI

TRIGONOMETRIC FUNCTIONS
76. Sine Curves 159
77. Periodic Functions 161
78. Differentiation of Trigonometric Functions 163
79. Inverse Trigonometric Functions 164
80. Formulas for Differentiating Inverse Trigonometric Functions . . 166
81. Angular Velocity and Acceleration 167
82. Simple Harmonic Motion 168

CHAPTER VII

EXPONENTIAL AND LOGARITHMIC FUNCTIONS

83. Exponential and Logarithm 175
1

84. Limit of (1 + h)^ 176

85. Differentiation of Logarithms 178
86. Differentiation of Exponentials 179
87. Differentials of Exponential and Logarithmic Functions .... 180
88. Logarithmic Differentiation 181
89. Integration 181

CHAPTER VIII

PARAMETRIC EQUATIONS
90. Parameter 187
91. The Cycloid 189
92. The Epicycloid 190
93. Length of Arc 190
94. Ratio of Arc and Chord 192
95. Differential of Arc 193
96. Curvature 194
X CONTENTS

CHAPTER IX
POLAR COORDINATES
SECTION PAGE
97. Polar Coordinates 202
98. Graphs 203
99. Intersection of Curves 206
100. Change of CoSrdinates 206
101. Equations of Certain Curves 207
102. Differential Relations 209
103. Area in Polar Coordinates 212

CHAPTER X
VECTORS
104. Scalar and Vector 218
105. Addition and Subtraction of Vectors 219
106. Multiplication by a Scalar 221
107. Points on a Line 223
108. Center of Gravity, Centroid 224
109. Components 225
110. Vector Function of a Scalar Variable, Derivative 227
111. Velocity and Acceleration 229
112. Tangential and Normal Components 231
116.
113. Polar Coordinates 234

CHAPTER XI
FORMULAS AND METHODS OP INTEGRATION
114. Elementary Functions 243
Formulas 243
126.
116. Integration by Substitution 245
117. Powers of Trigonometric Functions 246
118. Products of Sines and Cosines 247
119. Quadratic Expressions 248
120. Rational Fractions 250
121. Change of Variable 263
V
122. Integrals Containing (aa: -f 5)^ 254
123. Trigonometric Substitutions 266
124. Integration by Parts 257
Reduction Formulas 269
126. Infinite Values 260
127. Tests for Convergence 263
128. Numerical Integration 266
129. Prismoid Formula 266
130. Simpson’s Rule 267
 CONTENTS xi

CHAPTER XII

FURTHER APPLICATIONS OF INTEGRATION

SECJTION PAGE
131. Area of a Surface of Revolution 275
132. Area of a Cylindrical Surface 278
133. Mean Value of a Function 279
134. Moment 280
135. Center of Gravity of a Plane Distribution, Centroid 282
136. Center of Gravity in Space 286
137. Theorems of Pappus 288
138. Moment of Inertia 290

CHAPTER XIII

SERIES WITH REAL TERMS

139. Convergence 300
140. General Condition for Convergence 301
141. Positive Series 301
142. Integral Test 302
143. Comparison Test 304
144. Absolute and Conditional Convergence 305
145. Ratio Test 306
146. Alternating Series 307
147. Series of Functions 309
148. Power Series 309
149. Taylor’sand Maclaurin’s Series 312
Remainder
150. Taylor’s Series with 315
151. Operations with Power Series 317

CHAPTER XIV
COMPLEX NUMBERS
152. Complex Numbers 326
153. Geometrical Representation 328
154. Product and Quotient 330
155. Powers and Roots 331
156. Limit, Function, Continuity 333
157. Power Series 334
158. Exponential Function 335
159. Plane Motion 336
160. Logarithm 338
161. Hyperbolic Functions 339
162. Geometrical Representation of Hyperbolic Functions 342
163. Inverse Hyperbolic Functions 344
xii CONTENTS

CHAPTER XV
SPACE COORDINATES AND VECTORS
SECTION PAGE
164. Rectangular Coordinates 351
165. Vectors 352
166. The Scalar Product of Two Vectors 353
167. Work 356
168. Angle between Two Lines 357
169. Rotation of Axes 359
170. The Vector Product 360
171. Moment of a Force 364
172. Products of Three or More Vectors 365
173. Space Loci 367
174. Equation of a Plane 370
175. Equations of a Straight Line 372
176. Cylindrical Coordinates 373
177. Surface of Revolution 374
178. Quadric Surfaces 376
179. Length of Arc 379
180. Derivative of a Vector 381
181. Tangent to a Curve. Principal Normal 382
182. Derivatives in Cylindrical Coordinates 385
183. Angular Velocity 387

CHAPTER XVI
PARTIAL DIFFERENTIATION

184. Functions of Two or More Variables. Continuity 398

185. Partial Derivative 398
186. Dependent Variables 399
187. Geometrical Representation 400
188. Increment 400
189. Total Differential 402
190. Derivative of a Composite Function 404
191. Homogeneous Functions 406
192. Properties of the Differential 407
193. Tangent Plane and Normal to a Surface 410
194. Spherical Coordinates 411
195. Maxima and Minima of Functions of Two or More Variables . . 414
196. Higher Derivatives 415
197. Taylor^s Series 418
198. Exact Differentials 419
199. Directional Derivative. Gradient 422
200. Line Integral. Work 425
201. Case of Line Integral Independent of the Path 427
 CONTENTS xiii

CHAPTER XVII
MULTIPLE INTEGRATION
SECTION PAGE
202. Repeated Integration 437
203. Multiple Integration 441
204. Evaluation of a Multiple Integral by Repeated Integration . . . 442
205. Double Integrals, Rectangular Coordinates 444
206. Polar Coordinates 446
207. Triple Integrals, Rectangular Coordinates 460
208. Cylindrical Coordinates 461
209. Spherical Coordinates 454
210. Mean Values 466
211. Area of a Surface 457

Answers to Problems 464

Tables 492

Index 499
 CHAPTER I

LIMITS AND CONTINUITY

1. Real Numbers. The process of counting gives whole num-
bers. Division and subtraction give fractions and negative num-
bers. Whole numbers and fractions, whether positive or negative,
are called rational numbers. A number, like \/2, which can be
expressed to any required degree of accuracy, but not exactly by
a fraction, is called irrational. Rational and irrational numbers,
whether positive or negative, are called real.
The absolute value of a real number is the number without its
algebraic sign. The absolute value of a; is represented by the
notation I
xj. Thus
I -2l = 1+21 = 2.

2. Points on a Line.
Real numbers can be represented graphi-
cally by points on a straight line. To do this choose a unit of
length and take a point 0 on the line as origin. Upon the point 0
mark the number 0. On one side of 0 mark positive numbers,
on the other negative, making the number at each point equal in
0
absolute value to the distance from

•
-4
At
a -3
•
-2
# ». M
-1
OB
m
0
to the point.

16 2
»
3
%
4
Figure 1.

The line with its associated numbers is called a scale. Proceed-

ing along the scale in one direction (to the right in Figure 1) the
numbers increase algebraically. Proceeding in the other direction
the numbers decrease. The
which the numbers in-
direction in
crease is which they decrease negative.
called positive, that in
To each real number corresponds a point on the scale, and to two
numbers a, 6 correspond two points A, B whose distance apart is

U - «!•
To each relation between numbers corresponds a relation between
points, and conversely. Consequently in describing relations
1
2 LIMITS AND CONTINUITY [Chap. I

between numbers we often use geometrical language when the

intention is obvious.
3. Function. If two variables x and y are so related that when
a value is assigned to x a value of y is determined, then y is called
a function of x.
Thus, the area of a circle is a function of its radius; for, the
radius being given, the area can be calculated. The temperature
at a given point is a function of the time; for, the time being given,
the temperature has a definite value.
The function may not be defined for all values of the variable
but only for certain values. The values which a variable can
actually take constitute its range, and the function is said to be
defined on that range. That is,
A quantity y is called a function of a quantity x on a given range,
if to each value of x on that range corresponds a definite value of y.
In particular, the range consisting of all real numbers between
two values a and 6 is called an interval, and a variable which takes
all values in an interval or all real values is called continuous.

The interval consisting of the numbers a, b and all real numbers

between a and b is sometimes indicated by the notation (a, 6).
For example, if x and y are the sides of a right triangle with
hypotenuse a,
y = y/ x^.

Since, however, the side cannot be negative or as great as the

hypotenuse, y is defined only for values of x between 0 and a.
The range of the variable is the interval
0 < a: < a,

and on that range y is a function of x.

More generally,
A quantity y is called a function of any number of quantities Xi,
X2 ••- jXn on given ranges if to each set of values, xi, X2 , •, Xn, on
• •
f

those ranges correspondsa definite value of y.

Thus, the area of a triangle is a function of the lengths of its
three sides. The pressure of a given mass of gas is a function of
its temperature and volume.

4. Independent and Dependent Variables. In many problems

there occur a number of variables connected by equations. Arbi-
trary values are assigned to some of these variables and the others
are determined as functions. Those taking arbitrary values are
§ 6] FUNCTIONAL NOTATION 3

called independent variables. Those determined are called depend-

ent variables.
Which are chosen as independent and which as dependent is

often arbitrary. Certain choices may, however, be much more

convenient than others. Thus if x and y are real and

y = + X,
either x oy y could be taken as independent variable. If x is inde-
pendent variable, values of y are, however, easy to calculate,
but if y is independent the corresponding values of x are much
more difficult to determine.
6. Functional Notation. A particular function of a: is often rep-
resented by the notation f(x)j which should be read function of
x, or / of Xj not / times x. For example,

fix) = Vx^+1
means that f(x) is a symbol for + 1. Similarly

2/=/(x)

means that y is some definite (though perhaps unknown) function

of X,
If it is necessary to consider several functions in the same dis-
cussion, they are distinguished by subscripts or accents or by the
use of different letters. Thus /i(x), /2 (^), g{x) (read
/-one of X, /-two of x, /-prime of x, /-second of x, g of x) represent
(presumably) different functions of x.

a function of two variables x and y is represented by

Similarly,
the notation /(a:, y), a function of three variables a:, y, z by the
notation f(x, y, «), etc. Thus the equations

u = F{x, y), V = G{x, y, z)

express that u is a function of x and y and that t; is a function of

X, y, and z.

The / in the symbol for a function may often be interpreted as

representing an operation performed on the variable or variables.
Thus, if

fix, y) = Vx^ + y\
f represents the operation of squaring the variables, adding, and
extracting the square root of the result. If x and y are replaced
4 LIMITS AND CONTINUITY [Chap. I

by any other quantities the same operation is to be performed on

those quantities. For example,

VP + 22 = V5,
/(l, 2) =
fix, y + 1) = VP + (y + = VP + p + 1)^ 2j/ + 1.

6. Rectangular Coordinates. Graphically a function y of a con-

tinuous variable x can be represented by a curve. One way to do
this is to consider x and y as rectangular coordinates of a point

in a plane.
For this purpose take two perpendicular scales Z'X, F'F with
their zero points coincident at 0 (Figure 2). It is customary to
draw X'X, called the a:-axis,

horizontal with positive direc-

tion to the right, and F'F, called
the 2
/-axis, vertical with positive
direction upward. The point 0
is called the origin.
From any point P in the
plane drop perpendiculars PM,
PN to the axes. Let the num-
ber at M in the scale X'X be x
and that at N in the scale F'F
be y. The numbers x and y
are called the rectangular coordinates of the point P with respect
to the axes Z'X, F'F. The number x is called the abscissa, the
number y the ordinate, of P.
The point with coordinates x, y is represented by the notation

{x, y). To signify that P

has the coordinates x, y the notation
P(x, y) is often used. For example, (1, —2) is the point x = 1,
2/
= — 2. Similarly, A (—2, —3) signifies that the abscissa of A is

— 2 and its ordinate —3.

is

The process of locating a point with given coordinates is called

plotting. To obtain the curve representing a given function

y = fix)
we substitute values for x, calculate the corresponding values of
y, and plot the resulting points (x, y). The locus of such points is
the required curve. It is often called the graph of the equation
y = fix)-
Conversely, any locus (such as that in Figure 3) which cuts
each line parallel to OY
in a single point determines y as a. func-
§6] RECTANGULAR COORDINATES 6

tion of X. The value of the function which corresponds to a given

value X of the variable is merely the ordinate at the point of
abscissa x.

We thus have two ways of presenting a function f{x). One is

to give an expression by means of which its values can be calcu-
lated. The other is to give a graph on which its values can be
measured.

Example L Plot the curve

y — X
for values of x between —2
and 3.
By substituting values of x and calculating the corresponding values
of y we obtain the pairs of coordinates indicated in the following table:
x = -2, -1, 0, 1, 2, 3
2/
= 6 ,
2 ,
0 ,
0 ,
2 ,
6

We plot the points represented by these pairs of coordinates and draw

a smooth curve through the resulting points. The curve is shown in
Figure 4.

Figure 4. Figure 6.
6 LIMITS AND CONTINUITY [Chap. I

Example 2, When x is between —2 and 2 a function has the value

/(x) = — 4

and outside that range it has the value

/(*) = I
a:
I
- 2.

Plot the graph of y = /(x).

We substitute for x and calculate ?/, using 2/
= x^ — 4 if x is between
— 2 and 2, and t/ = |
x |
— 2 for all other values of x. Pairs of correspond-
ing values are indicated in the following table:

X = -4, -3, -2, -1, 0, 1, 2, 3, 4

y = 2, 1, 0, -3, -4, -3, 0, 1, 2

The graph is shown in Figure 5.

Limit of a Sequence. In many investigations direct calcula-

7.
tion does not give the exact value of a quantity that is under
consideration but does give a never-ending sequence of approxi-
mate values
*1 ,
a;2 , ara, •••, a:„, ••• (1)

which as n increases so approach a number c that the difference

\xn — c\

ultimately becomes and remains less than any given positive

number. We express this by saying that the sequence converges
to c as limit, or that the variable Xn tends to the limit c as n in-
creases.
For example, to obtain \/2 we use a process which gives the
value approximately as a decimal. To six places we find

V2 = 1.414214.

If Tn is the value to n decimal places, we have an infinite sequence

of approximate values

n, ^2, ^3 , • •
•, ^n,
• • •

such that
lr„- V2l < 10-".
( 2)

As n increases, the difference ]

r» — \/2 |
ultimately remains
less than any fixed positive number. The sequence of values
tn converges to as limit.
§71 LIMIT OF A SEQUENCE 7

A quantity is often determined as the sum of an infinite series.

For example, by division we find

1 -
Sn = = 1 +r+ +• •+ r^ •
(3)
1 — r

n being any positive integer. If j

r ]
< 1, the number

1 — r

can be made as small as we please by making n sufficiently large.

As n iiKToases, therefore tends to the limit

and the limit approached by the right side of (3) is indicated by

the notation
l + r + r2+...+ r"+*-*.
Thus, if
I
r I
< 1, wc have

= 1 -f r + H h r” + • * • • f4)
1 — r

To determine the circumference and area of a circle we inscribe

a regular polygon of n sides and determine its i)erimeter pn and
area We thus have two sequences
’ * ’
PSf PAi P5) *> Pny *

A3, A4, A5, • •

•
y A-jif • • •,

which converge to limits p and A respectively. These limits are

by definition taken as the circumference and area of the circle.
In each of these cases we have a variable (approximation to
n decimal places, sum of n terms of a series, perimeter or area of
a polygon of n sides) which, for the various positive integral values
of n, takes an infinite sequence of values

and these so about a number c that, as n increases, the

cluster
difference c j
—
remains less than a given positive number €
|

from some value of n onward. If that value is n = A, the above

definition is equivalent to the following:
8 LIMITS AND CONTINUITY [Chap. I

An infinite sequence of numbers xi, X2 X3, ,

• •
•
, Xn, • • • is said to

converge to the number c as limit if for each positive number e, no

matter how smalls there is a positive integer N such that
I
x„ - c I
< 6 (5)

for all values of n greater than N.

This behavior of the sequence, or of the variable Xn, is indicated
by the notation Xn'-^c or
lim Xn = c, (6)
n —» 00

Example 1, Find the limit approached by the decimal

0 12121212
.
- •

as the number of places is indefinitely increased.

Expressed as a series, this decimal is equivalent to

100 L ^ 100
IOC
+ ‘
(100)>*
+ (100)3

By equation (4), its value is therefore

12 1 12
100 1 — YJPS
99*

Example 2 , Find the value of

n - 00
2 n^
IV
— n
6
—
+ 1

Dividing numerator and denominator by n^, we obtain

2n2 - n + 1

n2 +3 1+4
As n increases, ~ and \ tend to zero and the right side of this equation
n TV
tends to the value 2. The limi t is th^efore 2.

Example 5.
n
lim (\/n
-* to
+ 1 — y/n),
Writing the given expression in the form

\/n + — 1 y/n
1

and rationalizing the numerator, we obtain

1

\/n +1— y/n —

§8] LIMIT OF A FUNCTION 9

As n increases; the right side of this equation tends to zero. The limit is

thus zero.
Example 4- Show that the area of a circular sector bounded by an arc s
and two radii is

A = (7)

Divide the arc into n equal parts and draw radii to the points of division.
These and chords between consecutive points determine a set of n triangles
having equal bases bn and equal altitudes
hn. The sum of lengths of the chords is

Sn =
and the sum of the areas of the triangles is

An = 'a(^hnbn) = ^hnSn*

As n increases, Sn approaches the arc s as

limit, hn approaches the radius r, and An
approaches the area of the sector. Thus

A = lim ^hnSn = irsy

n — » ao

which was to be proved.

8. Limit of a Function. Our discus- Figure 6.

sion of limits has hitherto been re-

stricted to variables that take only discrete values. Suppose now
that X a continuous variable and f{x) a function defined at all
is

points of an interval except possibly x = c. We are interested in

values of f(x) at points near x = c. If these differ arbitrarily little
from a number I at all points sufficiently near a: = c, we say f{x)
approaches, or tends to, the limit I as x tends to c. This is indi-
cated by the notation
lim/(x) = I (1)
X -*c

As an example consider the function

*2-4
*2 + 4

for values of x near 2. If we place x = 2, the function becomes

zero, and, if a: is near 2, the function is evidently near zero. Thus

x^ — 4:

lim = 0.
X 2 x^ +4
10 LIMITS AND CONTINUITY [Chap. 1

As a second example take

- 1
X —1
for values of x near 1. If we place a; == 1, the function assumes
the form
0
o'
which has no meaning. By division, however, we obtain

If X is nearly equal to 1, the right side of this equation is nearly

equal to 2 and the left side has the same value. Consequently

lim = 2.
X -»! X — 1

As a final example we take

\/X +1— 1

for values of x near zero. Since division by zero is meaningless,

we cannot place x = 0, If, however, we multiply above and below
by Vx + 1 + we have 1,

Va: +1 - 1 _ (Vx + 1 - l)(Va; +1+ 1)

X x(Vx +1+ 1)

X
x{\/ +1+ 1)

1
" Vx + l + l'
,

In the expression on the right we can now let x tend to zero and
so obtain
Vx + 1 - 1
=
1
“•
lim
X -0 X 2

In each of these cases we have a function which at points near

a value x == c is approximately equal to a number I, the error
being less than any stated amount € at all points of some interval
§9] ANGLE 11

of center x — c. If this interval extends a distance 5 on each side

of c, the above definition is equivalent to the following:
The function f(x) is said to approach the limit I as x tends to c,

if for each positive number € there is a positive number d such that

\f(x)-l\<e (2)

at all points of the interval

1
a: — c 1
< 5 (3)
except possibly x c.

From this definition it is clear that ifxij X 2

f
*
*
*,
•

is any sequence of values different from c, but tending to c as limit,

lim f(xn) = ly

n -+ 00

for Xn will belong to the interval (3) from some value of n onward
and then , ,

\f(Xn)

In the above definition x has been restricted to have values

differentfrom c. The reason for this is partly that the value x — c
is not needed in the determination of the limit and partly that

/(c) may not be defined or may not have the proper value if it is

defined.
9. Angle. In work that involves calculus angles are usually
expressed in circular measure.

If AOB (Figure 7) is a given angle and AB = s is the length of

arc it intercepts on a circle of radius r and center 0, the circular
measure of the angle is defined as the ratio
12 LIMITS AND CONTINUITY [Chap, I

Frequently no unit is mentioned, but when it is desired to

emphasize that circular measure is used we say the angle is one
of 6 radians, the radian being defined as the angle subtended by
an arc equal to the radius.
When the radius r and angle 0 are given, the length of the inter-
cepted arc is determined by the equation

s =* re. (2)

10. Limits of Trigonometric Functions. Inspection of a table

of sines will show that when 0 (measured in radians) is small

sin 0

is approximately equal to unity. Thus, if 0 is an angle of 5°,

sin 0 = .08716

0 - .08727
and
sin 0
.9987.
0

To prove that the limit of this ratio is unity, let 0 be the angle
AOP at the center of a circle of unit radius (Figure 8). Since
r = 1, equations (2), §9, and (7), §7, give

arc AP = 0, sector OAP = ^0. (1)

The inequality
AOMP < sector OAP < AOAQ (2)
§10] LIMITS OF TRIGONOMETRIC FUNCTIONS 13

is thus equivalent to

i sin ^ cos ^ ^ tan 6, (3)

whence division by ^ sin 6 gives

e 1
cos d < < (4)
sm d cos 6
and so
1 sin 6
> > cos d. (5)
cose d

1
As 6 tends to zero, cos 6 and both tend to the limit 1, and
cose

sin e

e
is between the two. Therefore

sin e
lim = 1. (6)
e-*o e

Another limit which is frequently used is that of

1 — cos e

e
14 LIMITS AND CONTINUITY [Chap. I

tends to the limit 1. Thus (8) gives as limit

1 — cos ^
lim = 0. (9)
6 -*0 6

Example 1, Find the value of

sin 5x
lim
x-*0 X
By algebra we have
sin sin 5x
g ( 10)
X 5x

As X tends to zero, 5x tends to zero and, by equation (6),

sin 5a;

ox

tends to the limit 1. From (10) we thus have

sin 5a;
lim = 5.
*-0 X
Example 2, lim
TT — 2a;

Placing
TT
-*= 2/.
2
we have
cos X __
sin y
TT — 2a; 2y
TT
As X tends to ~ y ,
tends to zero and
2
siny

to unity. Therefore
cos a;
= smy 1
lim lim -- *

» TT “ 2a; y-*o 2y 2

11. Properties of Limits. We often have several functions of

the same variable each of which approaches a definite limit when
the variable approaches a certain value. In such cases extensive
use is made of certain simple properties which follow almost
immediately from the definition of limit.
1. The limit of the sum of a finite number of functions is equal to
ihe sum of their limits.
Suppose, for example, that X, Y, Z are three functions of the
same variable x and that they approach the limits A, C
§11] PROPERTIES OF LIMITS 16

respectively when x approaches the limit c. The differences

X~A= €1, r-B= e2, Z-C ^

and their sum
(X + F + Z) - (A + B + C) = €1 + €2 + €3

then all tend to the limit zero as x tends to c. Thus

lim (X + F + Z) = A +B+C= lim X+ lim F + lim Z,

which was to be proved.
2. The limit of the product of a finite number of functions is equal
to the product of their limits.
Suppose, for example, that X, F are two functions approaching
the limits A,B respectively. Writing
X = A + Cl, F= B+ €2,

we have
XF - AB = A€2 + (B + C2)ci.

Since A, B are fixed numbers and €i, C 2 approach zero, the right
side of this equation will ultimately become numerically less than
any fixed positive number. Consequently

lim XF = AB = lim X lim F.

3. If the limit of the denominator is not zero, the limit of the

quotient of two functions is equal to the quotient of their limits.
Using the same notation as in the preceding discussion, if B is

not zero,
X
A A €i A Bei -- A 62 +
F ”B ” B+ ”B ""
C2 B(B + C2)

Since ei and €2 approach zero and B is not zero, the limit of the
right side is zero. Hence
X=—
— A = lim X
—
lim
F B lim F
If lim F= 0, the expression

lim X
limF
has no meaning.
16 LIMITS AND CONTINUITY [Chap. I

Sometimes a variable increases without bound.

12. Infinily.
If the increase such that, no matter how large
is may be, there M
is a stage in the process beyond which x remains numerically

larger than M, we say that x becomes infinite, or tends to infinity.

If, in addition, x ultimately has a fixed algebraic sign, we write

a;—>00 orx -^— 00 according as that sign is positive or negative.

Similarly a function f(x) of a continuous variable x is said to
become infinite as x tends to c if, no matter how large M may be,
there is some interval of center c at all points of which (different
from c) f{x) is numerically greater than M. If, in addition, f{x)
has a fixed algebraic sign at all points of a sufficiently small inter-
val of center c, we sometimes write

lim/(a;) = ]imf(x) = — oo
X -*C X —*c

according as that sign is positive or negative. These, however,

are not proper limits, and theorems concerning limits may not
apply to such cases.
In the definition just given it should be noted that nothing is
said about the value of the function at the point x — c. The
function could have any value at that point, but usually no value
is assigned. We often say the function is infinite at the point,
but this should be considered as merely a brief way of describing
a certain behavior of the function in the neighborhood of the point.
Consider, for example,

/(x) = -•
X
When X is near zero, /(x) is very large. In fact, no matter how
large the positive number M may be, /(x) is numerically greater

than M at all points of the interval

1 1
< X < H
M M
except X = 0 where it is not defined. The function becomes in-
finite as X tends to zero.
As an example of a different behavior consider the function

/(x) = 2i

When X tends to zero through positive values this function in-

creases indefinitely. But when x tends to zero through negative
§13] CONTINUITY 17

values it tends to zero. If both positive and negative values of x

are included the function neither tends to zero nor becomes infinite
when a; —> 0.
13. Continuity. If a function /(a:) is represented by a connected
curve, its value differs arbitrarily little from /(c) when x is suffici-
ently near c. This property is expressed by the word continuity.
That is,

A function f{x)is called continuous at x = c if f(c) has a definite

value and f{x) tends to f(c) as limit when x tends to c.

In view of the definition of limit (§8) the condition that f{x)
be continuous at a: = c is that for each positive number € there
exist a positive number 6 such that

l/(^)-/Wl<6 (1)

at all points of the interval

I
a; - c 1
< 5. (2)

From the analogous properties of limits it follows that ily = f{x)

is continuous at a; = c any positive power of y is also continuous.

and if two functions are continuous sum, product, and quo- their
tient are continuous provided the denominator is not zero. By
repeated application we conclude that any algebraic expression
in X is continuous for values of x that do not make any denominator
zero.
At a point where a function is not continuous it is called dis-
continuous. The most common discontinuities are those at which
the function becomes infinite and those at which it takes a finite
jump.
A simple illustration of finite jump is furnished by a step func-
tion. By this we mean a function which is graphically represented
by a set of horizontal line segments with abrupt rise or fall from
18 LIMITS AND CONTINUITY [Chap. I

one segment to the next, as in Figure 10. Many functions have

this form because fractional values are not used. Thus the cost
of first-class postage is a definite number of cents per ounce or
fraction of an ounce. The cost y therefore remains constant between
consecutive integral values of the weight x in ounces but changes
abruptly when x passes through a whole number. Similarly, the

Figure 10.

cost of life insurance is determined by the age of the applicant

at the nearest birthday. As a function of age it thus changes
abruptly midway between birthdays.
At a discontinuity a function may take the upper value, the
lower value, or any other. Frequently it is not defined. In the
case of postage, for example, the lower value is used, and in that
of life insurance no cost midway
between birthdays is defined.

Example 1, f{x)

If X is positive, \x\= x and f(x) Figure 11.

= 1. If X is negative, |
ic
|
= —x
and f{x) = —I. The graph y = /(x) is shown in Figure 11. As x passes
through zero from negative to positive
values the function jumps from — 1
to +1. No value of the function is

defined at x = 0.

Example 2. /(x) =
X J.

As X tends to 1, f(x) increases with-

out bound, being positive if x > 1

and negative if x < 1. The function

has an infinite discontinuity at x = 1

(Figure 12).

Figure 12.
§14] SOME FUNDAMENTAL THEOREMS 19

Example 3, f(x) = 2*.

As X tends to zero through positive values, f(x) increases without bound.

As X tends to zero through negative values, f(x) tends to zero. There is
an infinite discontinuity at a; = 0 (Figure 13).

14. Some Fundamental Theorems. In the development of cal-

culus certain nearly obvious results are needed exact proof of
which requires a more careful analysis of the nature of number
than is appropriate in a textbook of this kind. For convenience
of reference we state these results here, referring those who wish
proofs to books on advanced calculus or function theory.
(1) General Convergence Condition. A necessary and sufficient
condition that the sequence

^1 ; ^3 ;
* ’
*7 ^nj

converge is that for each positive number e there exist a positive integer
N such that
1
- a;„
1
< « (1)

for all values of m and n greater than N.

This amounts to saying that the condition for convergence is

that there be a place in the sequence beyond which the difference

of any two terms is in numerical value less than any preassigned
positive number €. That a convergent sequence has this property
is evident since there is a place in such a sequence beyond which

all terms differ from the limit by less than and so differ from
each other by less than c. What is important is that the converse
is also true, namely, that, if for each positive t there is a place in
20 LIMITS AND CONTINUITY [Chap. I

the sequence beyond which the terms all differ from each other
by less than €, there is a limit to which the sequence converges.
When we do not have an explicit expression for the limit this
theorem is the basis for inferring its existence. In the case, for
example, of regular polygons inscribed in a circle we know that
the areas of these polygons differ arbitrarily little when n is
sufficiently large, and from this wc infer that An tends to a limit
as n increases.

(2) Continual Increase or Decrease. A variable approaches a

limit if it always increases hut never becomes greater than some fixed
number or if it always decreases hut never becomes less than some
fixed number,
A variable that always increases or always decreases thus either
approaches a limit or becomes infinite. An illustration is the
decimal representation of an irrational number. If is the value
to n places, this increases as n increases but never exceeds the
next larger integer. It thus approaches a limit.

(3) Extreme Value Theorem. If a function f{x) is continuous in

the interval a ^ x ^ b, it has a greatest and a least value in that
interval.

By this statement we mean that there are numbers Xi, X2 within

the interval or at its ends such that

f(xi) g f(x) g /(xa) (2)

for all values of x in the interval. Graphically this means that

a finite portion of a continuous curve has a highest and a lowest
point.

(4) Intermediate Value Theorem. If f{x) is continuous in the

interval a ^ x ^by and N
is any number between f(a) and fQi),
there is some value c between a and b such (hat /(c) = N.
 PROBLEMS 21

This amounts to saying that between two given values a con-

all intermediate values.
tinuous function takes

Figure 15.

(5) Uniform Continuity. If f(x) is continuous in the interval

a ^ X ^ bj then for each positive number e there exists a positive
number 5 such that
I
/fe) - Kxi) < I
t (3)

for every pair of numbers Xiy X2 which belong to the interval and satisfy
the condition

1
a-2 - a-,
1
< 5. (4)

Graphically this theorem states that a function continuous at

all points of an interval
a ^ X ^ b

can be approximated by a step function in such a way that the

error is not greater than e at any point of the interval. To con-
struct such a step function it is sufficient to divide the interval
into parts each of length less than 8 and in each part replace f{x)
by the value it has at any selected point of that part.

PROBLEMS
1. Show that \xy\ = x |
?/ 1.
|
•
|

2. If X and y are both positive or both negative and |

a;
|
^ | y |,
show that

h+2/l=l-rl+|»/|,
|x-2/|=|x|-|j/|.
3. If X and y have different algebraic signs, show that

la:+3/l <1*1+13/1,
1*-J/1>|*1-13/1.
22 LIMITS AND CONTINUITY [Chap. I

4. Express the area A and circumference C of a circle as functions of the

radius r. By eliminating r express A as a function of C.
5. Express the volume of a sphere as a function of its area of surface.
6. From the definition of function show that a constant is a function of
any variable.
7. If 2 is a function of y and 2/ is a function of x, show that ^ is a function
of 10.
X.
8. Determine the interval formed by values of x that make = 4 —
positive.
9. Determine the range of the variable x if
|
x — 1 |
<3.
If X and y are real variables and

X
1 +x2

express x as a function of y, and determine the range of values y for which

it is defined.

11. If X and y are real numbers and — 22/ + x = 0, express y as & func-
tion of X, and determine the range of values of x for which it is defined.
12. If/(x) = x2 - X, find/(0),/(l),/(2),/(-2).
13. lifiy) = 2^ find/(0),/(l),/(2),/(-2).
14. If
X - y
y) =
x^y'
find /(I, 0) +/(0, 1) +/(1, 1).

16. If /(x) = 2 - X, find/(x) + 2 and/(x + 2).

16. If
19. X — 2
fix)
7+01 ’

find

and
fi:x)

17. If/(x) =X + 1, find/(x2) and [f(x)p.

18. If fix) = X + -X ,
show that

If

fix) =
ax +b
x — a
show that f(fix)) = x.
20 If/(x)
. = logx, show that /(x2/) =/(x) -j-fiy).
21. If/(x) = a* show that/(x +- y) ^ fix) fiy).
22. Plot the graph of 2/
= a? — 2 for values of x between —1 and 2.
23. Plot the graph of y = x^ -H x +- 1 for values of x between —2 and 1.
24. Plot the graph of y — x^ — x for values of x on the range —2 ^ x ^2.
26. Plot the graph of y = x a; for values of x on the range — 1 ^ x ^ 1.
+ | 1
 PROBLEMS 23

26. When x is between —1 and 4-1 a function has the value/(x) = 1 — a; | |,

and outside that range it has the value /(x) = a;^ — 1 Plot the graph of .

V = /W.
27. When x is between 0 and 2 a function has the value /(x) = x® and out-
side that range it has the value /(x) = 4x. Plot the graph of y = /(x).
Show that the following sequences converge, and determine their limits:

n
28. i f f , ,
• •

n + 1
(_i)«+i
29. 1, i
30. 1,1,1, •••,1, •••.

31. i (i)*, (i)», •••, (ir.

Find the values approached by the following decimals as the number of

places is indefinitely increased:

32. 1.11111 •••. 33. 3.131313- •*. 34. 2.010101

Determine the following limits, n being a positive integer:

n - 1 n 4 1
hm •
36. lim 4"3‘
n - 00 n 4 2 n -> 00 n2

1 - 1 4n4
2' 38. lim
1 _1_
n —» 00 2n3
1

lim 2» 40. lim 1

41.
44.
lim [\/2w 4 1 — V2w — 1 ].

42. lim [V (n
n “ 00
4 1 )^ 41 — Vn^-j-T ]•

43. If I
a;
I
<1, find the limit approached by the product
45.
fix) = (1 4a^)(l 4a;2)(l 4 x^Xl
.
4 a:
8
)(l 4xi«)-

as the number of factors is indefinitely increased. To do this form the product

(1 - x)/(x) = (1 - x)(l 4 x)(l 4 x 2

)(l 4 X^) - •

combine the first two factors on the right, then multiply by the third factor,
and finally solve for/(x).
etc.,
By factoring each term and then combining consecutive terms deter-
mine the value of the infinite product

Let An be the area and p the perimeter of a regular polygon of n sides.

If n increases and p remains constant, find
lim An.
n — 00 »
24 LIMITS AND CONTINUITY [Chap. I

46. The line AB oi length I is divided into n equal parts and on these parts
equilateral triangles are constructed as in Figure 16. If Sn is the sum of the
perimeters and An the sum of the areas of these triangles, find

lim Snt lim An.

n 00 n - CO

AAAAAA
A B
Figure 16.

47. In Figure 17 the triangles ABC, CDE, etc., are all equilateral, Z AOP
= e, AO = a. Find
and the total length of the polygonal line ABODE* • •

extending from A to O.

Figure 17.

48. Parts are cut from a line of length I in the following way. First take

f of the line, then \ of what is left, then | of what is left, etc. What is the
sum of all these parts?
49. A set of n circular cylinders with equal altitudes is inscribed in a cone

as shown in Figure 18. When the number of cylinders increases, their alti-

tudes tending to zero, does the sum of their volume approach that of the cone
as limit? Does the sum of the lateral areas of the cylinders approach that
of the cone as limit?
 PROBLEMS 26

60. A
cone with vertex at the center of a sphere of radius r intercepts an
area & on
the surface of the sphere. By considering & as limit of a sum of
polygons and the volume F within the cone and sphere as limit of the sum of
pyramids formed by joining these polygons to the center, show that F =
61. By expressing each term as a power of 2 show that the sequence

-v/2, V V2, F 2^ V2,

2 2
• •

converges, and find its limit.

62. Assume that a dropped upon the floor bounces to two-thirds the
ball
height from which If dropped from a height h find the total
it is dropped.
distance it moves in coming to rest.
63. If the time required to fall or rebound x ft. is \ \/x sec. and the ball in
the preceding problem is dropi)cd from a height of 16 ft., find the time required
to come to rest.

Determine whether the following functions tend to definite limits, and, if

they do, find the values:

lim
3a: +2 66. lim
x^ -f 1
64. ’

X —» X — 2 3 X + 1

x3 - 1 x^ - 2x - 3
66. lim 67. lim
X 1
1 X - 1 X -f 1

I
3x +4 X2 - 1
68. lim 3* 69. lim
2x - 00 X -j- 2

60. lim
Vx^ + 1
61. lim
xi -h 3
. x + l ’

x3 4- 2x6

vr+1 2 Vx 4-1 - 2
62. lim
X —*i3 X
63. lim ~ X - 3

Vx — Va
64. lim 66. lim 2^
X — a 0
1

66. lim 2 67. lim I

X - 1 1.

X -*0 X

\x-l\ sin 3x
lim 69. lim
1 X — 1 0 X
tan X
70. lim 71. lim (0 cos 0).
0 X 0 —* 0

sin^ 0
72. lim (e CSC 6). 73. lim
0-^0 0-^0 0'^

74. lim
sin 20
— :
76. hm ^ —
- 0 sm 40
—
sin 30
•

e -+ 0 sin ^ 0

X — sinx sm X
76. Urn 77. lim
26 LIMITS AND CONTINUITY [Chap. I

78. lim sin x.

X -* to
79. hm
X -*0
——
1
7-0
sin^
cosx
X
cosx _ 1 — sin X
80. lim 81. lim
2x r cos X

sin d — cos d
82. lim (esc x — cot x), 83. lim '
'

» 0

. sinx — cosx scc^ 0 — 2

84. lim 85.
tan X tan 0 — 1

86 . By writing
87.
sin X ^
= ^/x
sin a

find
sin X
lim — J=r •

\/x

Considering that the logarithm of a large number is approximately

equal to the number of its digits, determine

logx
lim
X - 00 X

88 Considering that the logarithm of a small number is approximately

equal to the number of zeros at the beginning of its decimal expression, find

lim X log X.
X -*o

89. By taking logarithms of both sides and using the result obtained in
1

Problem approached by y = x* as x increases without limit.

87, find the limit
90. By taking logarithms of both sides and using Problem 88, find the limit
approached by y = x^ when x tends to zero.

Determine the values of x, if any, for which the following functions are
discontinuous:

=
1
=
X® +1
91. f(x) 92. f(x)
x(x + 1) X* + l’
1

93. f(x) = 2 X. 94. f(x) = 2

95. f(x) = sec X. 96. f(x) = tan 2x.
X- Ixl
97 f(x)
. =
* + 1*1
 CHAPTER II

DERIVATIVE AND DIFFERENTIAL

16. Speed. Consider a particle moving in a straight or a curved

I)ath. we take the
If distance it moves during any interval of
time and divide by the time, the resulting ratio

distance moved
time

average speed during that interral.

is (tailed its

Thus, if an automobile travels 40 miles in 2 hours its average

speed during that period is

40
— = 20 miles per hour.
2

Let the time during which the motion is observed be made

shorter and shorter. As the interval tends to zero the average
speed during that interval usually tends to a limit. The speed
at a particular instant is defined as the limit approached by the
average speed when the interval of time measured from that
instant tends to zero. That is,

distance movedX
^
j.
( time /

Thus, if Si is the distance along the path from a fixed to a

moving point at time <i, and S2 is the distance at a later time t2 j

then S2 — Si is the distance it travels during the interval <2 — k and

^2 - Sj

h ~~ h
is its average speed during that interval. The speed at time k is

S2 ~ Si
Vi = lim ( 1)
ti t2 —k
27
28 DERIVATIVE AND DIFFERENTIAL [Chap. II

For example, if a body starts from rest and moves under the
action of gravity, the distance it falls in time t is

s = (2)

where gf, the acceleration of gravity, has approximately the value

g = 32.2 ft./sec.^

At time ti the distance fallen is

51 =
and at time t2 it is

52 =
During the interval from ti to t2 the average speed is

^2
— - <i)
—h
= hQ(f2 + k).
h ~~ h t2

The speed at time k is therefore

§2 ~ Si
Vi = lim gk- (3)
h t2 —k
Since k may have any value we can replace it by k thus obtaining

V = gt
as the speed at time t.

16. Slope. Let Pi(xij yi), P2 (x2 2/2 ) be two points on a line and ,

<l>
the angle from the positive end of the x-axis to the line. In mov-
ing from Pi to P2 the changes in x and y are

PiR = X2 — Xu RP2 = 2/2 — 2/1* ( 1)

The ratio of these changes

2/2 - 2/1 RP2

m= = = ^
tan
^
<!>
(2 )
X2 — xi PiR
is called the slope of the line.
Since it is equal to tan 0 the slope is evidently independent of ,

the positions of Pi and P2 on the line. If the coordinate axes have

the usual positions (a:-axis horizontal with positive direction to the
right, j/-axis vertical with positive direction upward) the slope
§16] SLOPE 29

of a line measures its steepness and is positive (Figure 19) or

negative (Figure 20) according as the right end of the line extends
upward or downward.

Figure 19 . Figure 20 .

If Pi(xij yi)y P2 (^ 2 ) 2/ 2 ) are two points on a curve (Figure 21)

- Vi _ RP2
2/2

X2 — xi PiR
is the slope of the chord P P2 Keeping1 . Pi fixed, let P2 move
along the curve toward Pi. If the slope of the chord tends to a
limit, that limit
2/2 - 2/1
mi = rhm (3)
-*Xi X2 — Xl

is called the slope of the curve at Pi.

Figure 21.

The line PiT with slope

tan <pi = mi (4)

is called the tangent to the curve at Pi. The angle 0i is measured

31 DERIVATIVE AND DIFFERENTIAL [Chap. II

from the positive end of the a:-axis to the tangent, being considered
positive when drawn in the direction that carries OX into OF by
rotation through 90°.

Example, Find the slope of the curve

2/
= 4x —
at the point A(l, 3).
Let P(Xj y) be a variable point on the curve. The slope of the secant
AP is
y-zl,
X — I

Replacing y by its value 4x — x"^ from the equation of the curve, this can
be written
~ 3 _ 4a; — — 3 _
2/ « _
a; — 1 a; — 1

When P moves along the curve toward A, x approaches 1 and the right
side of this eejuation approaclu^s 2 as limit. The slope of the curve at
A(l, 3) is therefore 2. The line AT with this slope is tangent to the curve
at A.

17. Derivative. To obtain the speed of a moving particle (§15)

we determine its distance s along the path from a fixed point as a
function of the time, divide the change in s by the corresponding
change in t, and take the limit as these changes tend to zero. To ..

find the slope of a curve (§16) we express the ordinate 2/ as a

function of the abscissa x, divide the change in y by the correspond-
ing change in a;, and take the limit.
§ 17 ] DERIVATIVE 31

When a variable changes value, the algebraic increase (new

value minus old) is called its increment and is represented by the

symbol A written before the variable.

Thus if X changes from 2 to 4 its increment is

Aa: = 4 — 2 = 2.

When X changes from 2 to —1

Ax = -1 - 2 = -3.
The increment is positive when there is an algebraic increase,
negative when there is a decrease.
Let
y •= Six). (1)

When X changes, y will also change, the new values being x + Ax,
and y + A?/. These new values satisfy the equation

i/ + Ai/ = /(x + Ax). (2)

By subtracting the preceding equation we obtain

^y = /(x + Ax) - /(x). (3)

The increment of y is thus a function of x and Ax.

If the ratio

—
Ay
=
/(x + Ax)-/(x)
(4)
Ax Ax
tends to a limit when Ax tends to zero, that limit is called the
derivative of y with respect to x. It is a function of x often repre-
sented by the notation /'(x). Thus

fix + Ax) - fix)

= lim (5)
Ax -»o Ax

In determining this limit x is held constant and Ax varies. If

for a particular value of x the limit exists, fix) is said to have a
derivative, or to be differentiable, at that value of x.
Our definitions of speed and slope are now equivalent to the
following:
If ^ = /(O is the distance along the path from a fixed point to a
moving particle at time t, then fit) is the speed of the particle at
time t; if y = fix) is the equation of a curve, fix) is the slope of the

curve at the point of abscissa x.

32 DERIVATIVE AND DIFFERENTIAL [Chap. II

To obtain the derivative of a given function f(x) it is merely

necessary to perform the operations indicated in equation (5).

Example 1. f(x) =
Replacing a: by+ Ax we obtain
a;

f(x + Ax) = (x + Ax)^ = x^ + 2x Ax + (Ax)^,

whence by subtraction

/(x + Ax) — /(x) = 2x Ax + (Ax)2

and so
/(x+Ax)-/(x)
Ax
2x + Ax.
When Ax approaches zero this gives as limit

f(x) = 2x.

Example 2. /(x) = - •

X
The increment of this function is

1 1 Ax
/(x + Ax) - /(x) = X + Ax X x(x + Ax)
whence
/(x + Ax) - /(x) 1

Ax x(x + Ax)
As Ax tends to zero this tends to the limit

/'(x) = -

Example 3, Find the slope of the curve

y = \/5
at the point (x, y).

Writing /(x) = -y/x, wehave

/(x + Ax) — fix) _ \/X + Ax — Vx 1

Ax Ax y/x + ^+y/x
If X is not zero this gives

as the slope at the point (x, y).

Example 4- The distance traversed by a moving point at time t is

s = 4^2 + 6«.
Find its speed.
§18] DIFFERENTIAL 33

Writing

we have
m= + 6t,

fit + At) - fit) = 8^ + 4 Ai + 6.

At
When At tends to zero this tends to the limit

fit) = 8t + 6

which is therefore the speed at time t.

18. Differential. It is customary to express the derivative of a

function as the ratio of two quantities called differentials. To
define these we must distinguish independent and dependent
variables (§4).
Let X be the independent variable and ^ = f(x) be a function
with derivative /'(x). The differentials of x and i/ are then defined
by the equations
dx = Ax, (1)

dy = /'(x) Ax, (2)

That is,
The differential of the independent variable is equal to its incre-
ment, and the differential of a function of the independent variable
is equal to the product of its derivative and the increment of the inde-
pendent variable.
From equations (1) and (2) we have
dy = /'(x) dx, (3)

This fundamental equation has been obtained on the assumption

that X is the independent variable. It is also valid, however, when
x is not the independent variable.
To show this let t be the independent variable and
y = /(a:), X = ^{t).

Since x is a function of t, f{x) is also a function of t, say

y = F{t),

When values are assigned to t and A^, values of x, y, Ax, and Ay

are determined. For the values considered we assume that f{x),
<f>{t), and F{t) all have derivatives. When At tends to zero the
equation
Ay Ay Ax
34 DERIVATIVE AND DIFFERENTIAL [Chap. II

then gives as limit

F\t) <#>'(0,
whence
F'(t) At = f'(x) At
Since t is the independent variable this is equivalent to

dy = f'{x) dxy
which was to be proved.
If dx is not zero we can divide by dx and so obtain

dy
dx
We thus conclude that
If dx y with respect to x is equal to the
is not zero the derivative of

quotient dy divided by dx.

Because of the convenience of this notation we shall nearly always
express derivatives in this quotient form.
Suppose, for example,

X = fy y = t^ +L
Eliminating t we get
y = x + Vx,
from which we could find

dx 2\/x
Instead of doing this we can, however, first determine

dx = 2t dty dy = (2t + 1) dt

and then by division obtain

dy
_ 2^ + 1
_ 1 _ 1
*

dx 2t 2t 2\/X
19. Significance of Differential. If

5 = f(t)

is the distance a body moves in time ty its differential

ds = f{t) dt

is the distance it would move in the time dt if during that interval

it continued to move with the speed

v^m
§19] SIGNIFICANCE OF DIFFERENTIAL 36

it has at time t. In general the speed does not remain constant,

and so the distance As it does move is different from ds. But if
the interval is short the speed is nearly constant and As is nearly
equal to ds.
Similarly, if

y == Six)
is any function,
= /'W dx

is the amount y would change while x is changing from xiox dx

if during that interval the ratio of the changes of the two variables
remained constantly equal to its limit

Ay
Urn
Ax -*0 ^
Ax
= fix).

In general this ratio is not constant, and so Ay (the actual change

in y) is different from dy. But when the changes are small this
ratio is nearly constant and Ay is nearly equal to dy.
Graphically, y = f(x) is the equation of a curve with slope

/'(x) = tan <t>

at the point P(x, y). In passing from P(x, y) to a second point

Q{x +
Axj y +
Ay) on the curve (Figure 23) the increments of x
and y are
Ax = PR, Ay = RQ,

If X is taken as independent variable,

dx = Ax = PR,
dy = f{x) Ax = (tan </>) Ax = RT.
36 DERIVATIVE AND DIFFERENTIAL [Chap. II

Thus Ax and Ay are the increments of x and y when we move

from P(x, y) to a second point Q(x + Ax, y + Ay) on the curve,
whereas dx and dy are the increments when we move from P(x, y)
to a point T{x dx, y + +
dy) on the tangent through P, The
equation
dy
= /'W
dx

merely expresses the obvious fact that the ratio of the sides of the
right triangle PRT is equal to the tangent of the base angle.
20. Approximate Value of the Increment. We have just seen
that when the values are small the increment and differential of a
function are approximately equal. To investigate the nature of
this approximation let x be the independent variable and y = fix)
a function with differential

dy = /'(x) Ax. ( 1)

Define a number 6 by the equation

Ay
(2)
Ax
Then
Ay = f\x) Ax + € Ax = dy + € Ax. ( 3)
Since

lim
Aaj -»
—
Ay
Ax
= /'(x)

equation (2) shows that when Ax tends to zero e tends to zero,

and so
Ay — dy = € Ax, (4)

ultimately becomes as small a fraction of Ax as you please.

A variable (such as Ax) which approaching zero as limit is
is

called an infinitesimaL Such variables, though ultimately as

small as you please, may have widely different relative sizes. If
cc and /3 are infinitesimals and

v ^ n
lim - = 0,
a -0 a

P is said to be of higher order than a.Both infinitesimals tend to

zero, but the one of higher order ultimately becomes indefinitely
smaller.
§21] DIFFERENTIATION OF ALGEBRAIC FUNCTIONS 37

Equation (4) is thus equivalent to the following:

At a value x where y = J{x) has a finite derivativey Ay differs from
dy by an infinitesimal of higher order than Ax.
To illustrate this numerically consider the function

y = /(a:) = X®.
The derivative is

fix) = 2x.

When X changes from 2 to 2.1 the differential of i/ is

dy = 2x Ax = 2(2) (0.1) = 0.40

and the increment is

Ay = (2.1)2 - 22 = 0.41.

The difference is thus

Ay — dy = 0.01 = Ax.

Similarly, when x changes from 2 to 2.01 the differential and

increment have the values

dy 2x Ax = 2(2) (0.01) = 0.04,

Ay = (2.01)2 - 22 = 0.0401.

In this case the difference is

Ay — dy = 0.0001 = yw
As Ax decreases, Ay — dy becomes a smaller and smaller, in fact
an infinitesimal, fraction of Ax.
The differential is usually simpler than the increment. When
it gives a sufficiently accurate approximation it is therefore fre-
quently used in place of the increment.
21. Differentiation of Algebraic Fimctions. The process of de-
termining derivatives or differentials is called differentiation. We
have hitherto obtained these by direct application of the limit
process. Instead of this direct method differentiation is usually
performed by means of formulas derived by that method.
The following formulas are sufficient for the differentiation of
functional combinations formed by the operations of ordinary
38 DERIVATIVE AND DIFFERENTIAL [Chap. II

algebra. In these, c, n are constants and u, v differentiable func-

tions of a single independent variable.

I. dc = 0.

II. d{u + v) = du + dt).

III. d{cu) = cdu,

IV. d{uv) ^ u dv + V du.

VI. d{u^) = du.

These formulas give the differentials of the various functions.

If the derivative with respect to a variable x is wanted it is merely
necessary to divide both sides by dx. Thus from IV we get

U— + V —
d(uv) dv du
=
dx dx dx

Formula II states that the differential of a sum can be obtained

by differentiating its separate terms and adding the results. An
operation that has this property is called distributive. This is a
most important property for it permits us to break a complicated
expression into simpler parts and operate separately on those parts.
Formula III states that differentiation and multiplication by a
constant are commutative operations. By this we mean that the
result independent of the order in which the operations are
is

performed.
To prove any one of these formulas it is convenient to represent
the expression to be differentiated by a single letter. To determine
the derivative with respect to an independent variable x we calcu-
late the increment of the expression, divide by Ax^ and take the
limit. To reduce the result to differential form we then multiply
by dx.
The only one of these proofs involving features different from
those already encountered in problems is the last. To prove this let

y = (1)

We must consider three cases according as n is a positive whole •

number, a positive fraction, or a negative rational number.

First suppose that n is a positive whole number. Let x be the
independent variable. When x changes to x Ax the functions +
§21] DIFFERENTIATION OF ALGEBRAIC FUNCTIONS 39

Uj y change to u + Au, y + Ay, these latter values satisfying the

equation
2/ + A?/ = (?/ + AuY.
Expanding by the binomial theorem, this becomes

, n(n — I)
2/ + Ai/ = + nu^ ^ Au -\ ^ (Au)^ H .

2 !

Since y = u^,
,
n(n — 1) _ _
Ay = nu^ ^ Au -\ u^ ^ (Au)^ H ,

2 !

whence
n(n — 1) Au
+ u n—2 Au +
Ax 2r Ax
For the values under consideration u is assumed to have a deriva-

tive with respect to x. Consequently Au tends to zero when Ax

tends to zero. Since each term in the bracket except the first
contains Au as factor, the above equation gives as limit

dy
-I
dx
= nu^-^
,
—
du
dx
Multiplying by dx we thus have

dy = nu^ ^ du (2)
which was to be proved.
V
Next suppose that n is a positive fraction - and
Q
p
y — u^ = (3)
Then
2
/" = (4 )

Since p and q are positive integers we can apply the formula just
proved to both sides of (4), obtaining
qy^~^ dy = pu^~^ du. (5)
Substituting
P\Q—1 V
( u^) = u
and solving for dy we obtain

dy = V—u^"-1 du = nu^
—1^
du, ( 6)
Q

which proves the formula when n is a positive fraction.

40 DERIVATIVE AND DIFFERENTIAL [Chap. II

Finally, if n is a negative rational number,

1
y = =-

By formula V we then have

dy =
u —2n
Since — n is positive we can apply equation (6) just proved and so
obtain
^
nu ^ du
dy = —2n = nvP' ^ du.
u
Therefore whether n is an integer or fraction, positive or negative,

du^ = du.

If irrational powers are properly defined this formula is valid even

when n is irrational. Proof of this, however, will be deferred to a
later section (§88 and Problem 50, Chapter VII).

Example L y= + 7,
Using formulas II, III, and VI in order we have

dy = d(x^) + d(—4x^) + d(—9x) + d(7)

= d(x^) - 4 d(x^) - 9 d(x) + d(7)
= (3x^ — 8x — 9) dx.

1
Example 2. y = y/x + y/x ’

This can be written

2/ = + x'~i.
Consequently, by II and VI

dy
— 1
= -X 1
a
I
X
3
a = 1
n:
1
*
dx 2 2 2y/x 2y/x^

Example 3, y — {x^ — a^)(x^ — 6^).

It is assumed that a and h are constants. Using IV with w =
t; = — b^y

- - + - - a*)
^= (x* a*)
^ (x* 6*) (** 6*)
^ (x*

= (x* — + (x* — 6®)2x

a®)2x

= 4x» - 2(0* + 6*)x.

§ 22 ] DIFFERENTIATION OF SINE AND COSINE 4]

Example 4- V = "S/x* — I.
Using VI with m = x* — 1,
dy = K®* - l)"i d(x^ - 1)

_ xdx
"S/x* — 1

2x - 1
Example 5. y =

Using V with m = 2x — 1, v = 2x -i- 3,

.
= (2x + 3) d(2x - 1) - (2x - 1) d(2x + 3)
(^+3?
{2x + 3) 2dx — {2x — 1) 2dx
”
+ 3)2^
_ ^dx
""
(2x + 3)2*

22. Differentiation of Sine and Cosine.

VII. d sin u = cos u du,
VIII. dcos i^ = — sin udu.
To prove the first of these formulas let

y = sin w. ( 1)
Then
y + Ay = sin {u + Au) = sin u cos Au + cos u sin Au,
Subtracting the preceding equation,

Ay = sin u (cos Aw — 1) + cos u sin Aw,

whence
—
—
Ay
Aw
= cos w
sin Aw
Aw
.

sin w
1 cos
Aw
;
Aw
.
( 2)

By equations (6) and (9) of §10 we have

sin Aw 1 — cos Au
lim = 1, lim = 0.
Au - 0 Aw Au -0 Aw
When Aw tends to zero equation (2) thus gives as limit

dy
— = cos w,
du
and so
dy = cos w du
which was to be proved.
42 DERIVATIVE AND DIFFERENTIAL [Chap. II

To prove the second formula we note that

cos « = sm I u]‘
\2 /
By the formula just proved we thus have

dcosu —'d sin { w )

= cos {
u]d( u) = — sin du.
\2 / \2 / \2 /
Example 1. y = 2 cos (3x + 4).
By formula VIII we have

dy = —2 sin (3® + 4) d (3a: + 4) = —6 sin (3a: + 4) dx.

Example 2. y = sin (3a:*).

Using formula VII

^
dx
= cos (3x*) ^
dx
(3x*) = 6a: cos (3a:*).

23. Small Errors. If the error in measuring a quantity y is Ay,

the ratio .

Ay
y
is called the relative error and
Ay
100 —
y
the percentage error in y.
When the relative error is small (as it usually is) a satisfactory
approximation is obtained by replacing Ay by dy. Thus
dy

is usually taken as the relative error and

dy
100 —
as the percentage error in y.
Example 1. The area of a rectangle is determined by measuring its
sides X and y. If the measurement of x is 1% too large and that of y
is }% too small, find the percentage error in the area.
From the expression
A= xy

for the area, by differentiation and division, we find

dA dx dy
A ~
^

X y
§24] DERIVATIVES OF HIGHER ORDER 43

The relative error in A is thus the sum of the relative errors in x and y
and the percentage errors, being 100 times as great, satisfy a similar equa-
tion. Thus
percentage error in A = 1 — ^ = \%.

Example 2, The time of vibration of a pendulum of length I is deter-

mined by the equation

r= 2ir -S--

Find the maximum error in the computed value of T due to errors of 1%

in the measurements of I and g.
The expression for T can be written

T= 2Tlig-i,
whence
dT = Tl^ig ~~2 dl — 'irlig'^i dg
and
dT _ Idl Idg
T"” 21“ 27*

The relative error in T is thus one-half that in I minus one-half that in g.

Since either measurement might be too large or too small the algebraic
signs could be such that the errors add. Thus

maximum error in T= | +§ = 1%.

24. Derivatives of Higher Order.

The derivative of
y = (1)

which we have written

(2)
dx

is also a function of x. Its derivative with respect to x, written

d^y d /dy\
(3)

is called the second derivative of y with respect to x. Similarly the

third derivative is

d^y

dx^
_ =
_ = dx\dxV d / d^y\
f 1
^ (4)

etc.
44 DERIVATIVE AND DIFFERENTIAL [Chap. II

For example, if

we find

and all derivatives of higher order are zero.

From the notation
d^y

dx'^

it might appear that the nth derivative is the ratio of two quantities
d^y and {dxy. These higher derivatives, however, do not have all
the properties this fractional notation suggests.
In the product
d^y /dx's^

dx^ \dt

one might think, for example, that (dx)^ could be canceled in

numerator and denominator, giving

cPy

'd^'

That this may not be true is shown by taking

y = X = t^.

Then
dH^y /dx'y
= 0
dx^Kdi,)
whereas
§25] IMPLICIT FUNCTIONS 45

Consequently when n is greater than unity we do not consider

dx^
as a fraction but merely as a notation for the nth derivative of y
with respect to x.
Since each derivative is the first derivative of the preceding one
wc can, however, write

the right side being considered a fraction.

Thus if

X = + 1, y = f— 1
we have
dy _ d{e - 1)
_
3^2
_ 3
^

dx ” d{e + 1) 2tdt “ 2^'

dx^ dx 2t dt 4it

25. Implicit Functions. If x and y satisfy an equation

/(x, y) = 0,

not solved for then y is called an implicit function of x. To

y,
a function we could solve for y and then differ-
differentiate such
entiate. It may be simpler, however, to differentiate the given
equation term by term and solve the resulting equation for the
derivative.
Suppose, for example.
2x2/ + 2/^=1. (1)

Differentiating with respect to x.

dy dy
2x~^ + 22 + 22/-
/
dx
= 0, (2)
dx
whence
dy y /o\
46 DERIVATIVE AND DIFFERENTIAL [Chap. II

To get the second derivative we differentiate again with respect

to Xy thus obtaining

^y
<ey
(.
+,^y\ ( , \

dx^ (a: + yf (a; + y)^

dy
Replacing — by its value from (3) and simplifying,
dx
dPy 2xy + y^
dx^ {x + yY

Since x and y satisfy (1), this can be reduced further to the form

d^y 1
_
dx^ {x + 2/)®

By differentiating again with respect to x we could find the third

derivative, etc.
dv
26. Sign of the Derivative. At a point where is positive the
dx
curve
y = /(^)
slopes upward on the right, downward on the left. For sufficiently
small changes. Ay then has the same sign as Ax and y increases

as X increases. At a point where —

dy
is negative the curve slopes
dx
downward on the right and y decreases as x increases. We thus
have the theorem

For a particular value of x, if ^

dx
is positivCy y increases as x in-

creases; tf —
•.dy
is negative, y decreases as x increases.
dx
The curve be divisible into arcs on which the slope
will usually
is positive and others on whichit is negative. For the curve in
Figure 24, for example, the slope is positive on the left of A,
negative between A and B, and positive on the right of B.
To determine for what values of x a given function /(x) increases
as X increases and for what values it decreases as x increases, we
thus determine the values of x for which /'(x) is positive and the
values for which /'(x) is negative.
§26] SIGN OF THE DERIVATIVE 47

If f(x) is a polynomial this is particularly easy. We merely

factor the polynomial into a product of real factors and use the
fact that a product is negative only if it contains an odd number
of negative factors.

Example, Determine the values of x for which

J{x) — x^ — + 1

increases as x increases and those for which this function decreases as x

increases.
Differentiating and factoring,

f{x)
— 4a:® — 12a:2 + 8a: = 4T(a: — l)(x — 2).

The values of x for which the right side is zero are a: = 0, 1, 2, These
divide all real numbers into four in- . ...
tervals. | \ \
—
1. If X < 0, all three factors in 0 1 2
the product are negative and f'{x) Figure 25.
is negative.
2. If X is between 0 and 1, two factors are negative and the product
positive.
3. If X is between 1 and 2, one factor is negative and the product
negative.
4. If X > 2, all factors are positive and the product positive.
The algebraic signs for all real values of x are indicated graphically in
Figure 25.
If
x<0 or l<x<2,
the derivative is negative and the function decreases as x increases. If

0<x<l or x>2,
the derivative is positive and the function increases as x increases.
48 DERIVATIVE AND DIFFERENTIAL [Chap, II

The graph of the function is shown in Figure 26.

dx^ dx \dx.
is the derivative of the slope. At a point where the second deriva-
tive is and
positive the slope therefore increases as x increases
decreases as x decreases, and so the curve bends upward, or is
concave upward, as at A in Figure 27. At a point where the sec-
ond derivative is negative the slope decreases as x increases and
increases as x decreases. The curve then bends downward, or is
concave downward, as at B, That is,
A curve
y =
is concave upward at points where the second derivative

d^y

dx^

and concave downward at points where it is negative.

is positive,
A point,
on one side of which the curve is concave upward and
the other concave downward, is called a point of inflection. On
d^y
one side of such a point
^ is positive

we
and on the other side nega-

tive. To find points of inflection thus determine positions at

which the second derivative changes sign. At such a point the
second derivative is usually zero, but it may be infinite or have a
finite discontinuity.
§28] MAXIMA AND MINIMA 49

Example. Find the values of x for which the curve

y = x^ — Zx^ +2
is concave upward and the values for which it is concave downward.
Also find the points of inflection.
In this case

When X > \ the second derivative is positive and the curve concave up-
ward. When X < I the second derivative is negative and the curve con-
cave downward. There is a point of inflection at x = l^y = 0. The curve
is shown in Figure 27.

28. Maxima and Minima. A function f{x) is said to have a

maximum at = a if when x = a the function
a; is greater than for
any other value immediate neighborhood of a. It has
of x in the
a minimum a ii when x = a the function is less than for
at x =
any other value of x sufficiently near a.
If we represent the function by y and plot the curve

y = Six),

a maximum occurs at the top,

a minimum at the bottom, of a wave.
Thus 28 the function has a maximum at A and a mini-
in Figure
mum at B. It should be noted that a maximum is not necessarily
the greatest and a minimum not necessarily the least value of the
function.

There are several types of maxima and minima. At present

we consider only the simplest, namely those represented graph-
60 DERIVATIVE AND DIFFERENTIAL [Chap. II

icallyby points such as il, jB, Figure 28, the curve extending on
both sides of the point and having a definite slope at the point.
At a maximum or minimum of this type the slope is zero. For, if
it were not zero the curve would rise on one side and fall on the

other, and the function could not have either a maximum or a

minimum at the point.
Hence in determining maxima and minima of a function f(x) we
first look for values of x such that

f(x) = 0.

If a: = a is such a value, the function f(x) may have a maximum

at a: = a, a minimum, or neither. Consideration of a diagram
makes evident the following rule for distinguishing the three cases:
If f(x) changes from positive to negative as x increases through a,
the function f{x) has a maximum at x = a.
If f(x) changes from negative to positive as x increases through a,

the function f{x) has a minimum at x = a.

If f{x) has the same sign on both sides of x = the function f(x)
has neither a maximum nor a minimum at x a.

Example 1. The sum of two numbers is 6. Find the maximum value

of their product.
Let one of the numbers be x. The other is then 6 — a:. The value of
X isto be found such that the product

y — x(Q — x) = Gx —
is a maximum. The derivative

^ = 6 - 2* = 2(3 -
dx
i)

is zero at x = 3. than 3 the derivative is positive and y de-

If x is less

creases as X decreases. greater than 3 the derivative is negative

If x is

and y decreases as x increases. If x starts from 3 and varies steadily in

either direction y steadily diminishes. The product therefore has the
greatest possible value
y = x(6 — x) =s 9
at X =
3 (Figure 29).
Example 2, Find the shape of the closed pint can which requires for
its construction the least amount of tin.
Let the radius of base be r and the altitude h (Figure 30). The areas
of bottom and top are each and that of the side wall is 2wrA. Hence
the area of tin used is
A = 27rr2 + 27rrA. ( 1)
§28] MAXIMA AND MINIMA 51

Let V be the number of cubic inches in a pint. Then

V = ( 2)
Consequently

^ = —2
irr^

and

(3)
r
Since tt and v are constants

dA 2irr^ - v\
— 47rr (4)

This is zero when 2Trr^ = v, that is, when

3 V
/
r = ( 6)
2ir

If a smaller value than this is substituted for r, 27rr® will be less than v and

—
dA
dr
negative. If a larger value is substituted, 27it® will be greater than v and

—
dr
positive. As r increases, —dr
changes from negative to positive and

so the above value determines a minimum.

Another method of showing that (5) really determines a minimum is

to observe that since the amount of tin cannot be zero there must be a

least value. At this least value — must

dA
dr
be zero (see exceptions in §31).

The derivative is zero for only one real value of r. That must then be
the value which gives the minimum area.
52 DERIVATIVE AND DIFFERENTIAL [Chap. II

Combining the equation

2Trr^ = V,

which determines the minimum, with the equation

V = Tr%
we see that h = 2r. The closed can requiring the least amount of tin
thus has an altitude equal to the diameter of its base.

29. The Second Derivative Test. Suppose that

/'(a) = 0.

If /"(a) is positive the curve y = f{x) is concave upward (§27) and

a: = a is a low point on the curve. If f"{a) is negative, the curve
is concave downward and x = a is a high point. Therefore
If f(a) = 0 and S"{a) > 0, the function f(x) has a minimum at
X — a, Iff(a) = 0andf\a) < 0, the function f{x) hasamaximum
at X = a.

Example. Find the largest right circular cylinder that can be inscribed
in a given right circular cone.
Let r be the radius and h the altitude of the cone, x the radius and y
the altitude of the inscribed cylinder.
From the similar triangles DEC and
ABC (Figure 31)

DE AB
EC BC ’

that is

I
r
—— y
;
X
h
r
» 2/
=
r
- \
^)-

The volume of the cylinder is

V = TTX^y = IT - (rx^ — x^).

r
Hence
dv Trh _
3x*).

This is zerowhen x = 0 or a; = |r. The value x = 0 obviously does not

give the maximum. Hence
X = fr

is the only value to try. At this value the second derivative is

dx^
= y{2r-6x)= -2irh.
§ 31 ] OTHER TYPES OF MAXIMA AND MINIMA 63

Since the first derivative is zero and the second derivative negative the
value of X found determines a maximum volume.

30. Method of Finding Maxima and Minima. The method we

have used in solving maxima and minima problems involves the
following steps:
(1) Decide what is to be a maximum or minimum. Let it be y,

(2) Express y in terms of a single variable. Let it be x.

It may be convenient to express y temporarily in terms of sev-
eral variables. If the problem can be solved by our present meth-

ods, there will, however, be relations enough to eliminate all but

one of these.
dy
(3) Calculate — and find for what values of x it is zero.
dx
(4) It is usually easy to decide from the problem itself whether
the corresponding values of y are maxima or minima. If not,
dy
determine the signs of — when x
,

is a
,

little less and little greater

than the values in question, and apply the criteria of §28, or find
the second derivative and apply the criteria of §29.
31. Other Types of Maxima and Minima. The methods of
§§28, 29 are sufficient for the determination of the maxima and
minima of a function which occur at interior points of an interval
within which the function and its derivative are continuous. In
Figure 32 are shown some types that do not satisfy these conditions.

At B the tangent is vertical and the derivative infinite. At C

the slope has a finite discontinuity. At A and D the curve ends.
This happens when values of the variable beyond a certain range
are impossible. According to our definition the function has max-
ima Sit Aj B and minima at C, D although the derivative is not
zero at any of these points.
64 DERIVATIVE AND DIFFERENTIAL [Chap. II

This diagram shows that in determining maxima and minima

special attention must be given to values at which the derivative
becomes infinite or discontinuous and values beyond which some
condition becomes impossible.

Example L Find the maximum and minimum ordinates on the curve

In this case
'
dx 3

No finite value of x makes the derivative zero but x = 0 makes it infinite.

Since y is never negative the value a; = 0, 2/

= 0 is evidently a minimum
(Figure 33).
Example 2, A man on one side of a river 1 mile wide wishes to reach
a point on the opposite side 2 miles down the river. If he can row 6 miles
per hour and walk 3, find the route he should take to make the trip in
least time.

Let Ay Figure 34, be the starting point and B the destination. Suppose
that he rows to C, x miles down the river. The total time is then

t = + K2 - x).
§31] OTHER TYPES OF MAXIMA AND MINIMA 55

Equating the derivative to zero, we get

X 1
0 ,

which reduces to
3^2 +4= 0.

There is no real solution.

The trouble is that
3(2 — a;)

is the time of walking only if C is above B, If C is below R, the time is

Ux ~ 2).

The complete value of t is thus

t = ^\/a:2 + 1 db J(2 -- x)^

the sign being + if ^ < 2 and — if x > 2. The graph of the equation
connecting x and t is shown in Figure 35. At a: = 2 the derivative is
discontinuous. Since he rows faster than he walks, the minimum obvi-
ously occurs when he rows all the way. That is, x = 2, ^ = i\/5*
Example 3. Find the point P{Xy y) on the circle == 2x at
shortest distance from Q( — 1, 0).

The circle is shown in Figure 36. The triangle QMP has vertical
side y and base a; + !• The distance QP is thus

I ^Qp =
Replacing y^ by its value 2x — from the equation of the circle, this

becomes
I = V(^ + 1)^ + 2a: - a:2 = V^ + h
whence
—= ^
dx y/4cx + I
56 DERIVATIVE AND DIFFERENTIAL [Chap. II

This does not become zero for any x. The explanation is

finite value of
that on the circle x lies The corresponding values of I
between 0 and 2.

are represented graphically by the curve AB in Figure 37. The minimum

value of I occurs at A where x = 0. The shortest distance is therefore

I = \/4x+ 1 = 1,

as is also obvious from Figure 36.

32. Velocity, Acceleration, Rate of Change. If t is the time and

2 is a function of t, the derivative

dz

dt

is called the rate of change of z. It is a function of the time which

measures how fast z is changing at time i. When the rate of change
is positive, z is increasing; when negative, z is decreasing.
Thus, if 5 = J{t) is the distance a body travels in time f, the rate
of change of s is its speed at time t. Whether the body moves one
way or the other along its path, the distance traveled always in-
creases and so the speed is always positive.
In general it is desirable to distinguish motion in one direction
from motion in another. At present we consider only a particle
moving along a straight line. Designate the position of the par-
ticle P at time t by means of its displacement from a fixed point 0
on the line. By this we mean

O
•
P
1
^ a coordinate

Figure 38. x = OP
numerically equal to the distance of P from O but positive when
P is on one side of 0, negative when on the other. The rate of
change of displacement
dx
dt

is called the velocity of P. It is numerically equal to the speed of P

but positive or negative according as the motion is in the direction
in which x increases or that in which x decreases.
The velocity of a moving particle is a function of the time. Its
rate of change
dv d^x
§321 VELOCITY, ACCELERATION, RATE OF CHANGE 67

is called the acceleration of the particle. When the acceleration is

positive the velocity is algebraically increasing; when negative,

the velocity is algebraically decreasing.
Displacement, velocity, and acceleration are usually considered
directed quantities. When positive they are drawn in the direc-
tion in which x increases; when negative, that in which x decreases.
When the velocity and acceleration have the same direction,
the change of velocity is in the direction of the velocity and so the
speed is increasing. When the velocity and acceleration are op-
positely directed the change of velocity is opposite to the velocity
and so the speed is decreasing.

Example 1 . The vertical height in feet of a ball thrown upward is

h= mt- 16^2

at the end of t seconds. Find its velocity and acceleration. Also find when
it is rising, when falling, and when its speed is increasing and W’hen de-
creasing.
The velocity and acceleration are

t; = = 100 - 32^,
(it

dv

The direction in which h increases is upward. The ball will be rising

and the velocity directed upward while v is positive, that is, until

“inr

It will be and the velocity directed downward after t = 3|. Since

falling
the acceleration is negative it is always downward. Before ^ = 3J the
velocity and acceleration are oppositely directed and so the speed de-
creases. After t = 3| the velocity and acceleration have the same direction
and the speed increases.
Example 2. A ship B sailing south 16
miles per hour is northwest of a ship A
sailing east 10 miles per hour. At what
rate are the ships approaching or separat-
ing?
Let X and y be the distances of the
ships from the point where their paths
cross. The distance between the ships is

then
8 = y/x^ + y\
68 DERIVATIVE AND DIFFERENTIAL [Chap II.

The rate of change of s is the rate at which the ships are approaching or
separating. By differentiation we find

dx dv
ds
dt

By hypothesis

= -16,
dt

X y 1
COS 45°
-s/x^ + y'^
V \/2
Therefore
ds _ 10 - 16
— 3-y/2 mi./hr.
dt '\/2

The negative sign shows that s is decreasing, that is, the ships are ap-
proaching.

33. Related Rates. In many problems the rates of change of

certain variables are known and the rates of others are to be cal-
culated. This is done by writing the equations connecting the
variables and The
differentiating these with respect to the time.
resulting equations will contain both the known and the unknown
rates. When values are substituted for the known rates, if the
problem is determinate, there will be equations enough to solve
for the unknown ones.

Example, The sides of a rectangle are a; = 5 in., y = 8 in. If a; is

increasing 0.1 in./sec. and the area is decreasing 0.2 in.Vsec., find the
rate of change of y.
From the expression
A — xy
for area we obtain

dt

Replacing x,y,
dx
—
at
^
and — by
dA.

at
the values stated, we have

- 0.2 = 5 ^+80 ( . 1 ),

whence
—0.2 in./sec.
dt
§34 ] MEAN VALUE THEOREM 59

34. Mean Value Theorem. Let AB

(Figure 40) be an arc of
the curve y = f{x). It is geometrically evident that there is some

point P on this arc at which the tangent is parallel to the chord AB.
Equating the slope at P to the slope of the chord, we have

/'(?) = m— -~/(a)

b a
whence
f(b) - /(a) = {b^a)m. (1)

This is known as the mean value theorem. The following is a

more precise statement:
Mean Value Theorem. If f{x) has a derivative for each value of
X in the interval a ^ x ^ by there issome value f between a and b
such that
m- f(a) == (b - a)m. (2)

To prove this analytically consider the function

~
n^) = /(x) - f(a) - 0 — a
(x - a) (3)

which has been so constructed that

F(a) = F(b) = 0.

The result to be proved is equivalent to showing there is some

value { between a and h such that

F'(^) =no - m— -/(«)

= 0.
b a

Now F(x) is continuous in the interval a ^ x ^ b and is equal to

zero at a: = a and x - b. Either this function is zero for all
60 DERIVATIVE AND DIFFERENTIAL [Chap. II

values of x between a and 6, and then F\o^) = 0 for all such

values, or there is some value { in the interval a ^ ^ 6 at which
is greatest or least. (Extreme Value Theorem, §14.) This
extreme value cannot occur at an end of the interval, for there
F(a;) is zero. Since F{x) has a derivative at a; = {, that derivative

F'(f) must be zero. For, if F'(5) were different from zero, F(x)
would increase when a:, starting from f varies in one direction and
,

decrease when x varies in the other direction and so could not

have a greatest or least value at a: = f
In particular, if
m
=/(&) = 0 ,

the mean value theorem shows that there is some value $ between
a and 6 such that
/'(€) = 0.

This is called Rollers theorem.

Rolle’s Theorem. 7/ /(a) = /(6) = 0 and f(x) has a derivative
at each point of the interval a ^ x ^ b, there is some value ^ between
a and b such that

m = 0 .

With two functions the following extension of the mean value

theorem sometimes useful:
is

If fix) and <l>ix) have derivatives for all values of x in the interval
a ^ X ^ bj and if 0'(a;) does not vanish in that interval except pos--
sibly at X ^ a or X = bf there is some value f between a and b such
that
m -m no

To prove this consider the function

Fix) = Uix) -fmm - <l>(a)] - I4>ix) - mmb) “/(o)]-

As this satisfies the conditions of Rolle’s theorem there is a num-
ber { between a and b such that

F'io = rmib) - <#•(«)] - muib) - /(a)] = o.

Since ^'(x) does not vanish between a and 6 the mean value
theorem shows that 0(6) — 0(a) is not zero. From the last equa-
tion, (4) therefore follows by division.
§36] INDETERMINATE FORMS 61

Example. Find ? in the mean value theorem if o = 1, 6 = 4, and

fix) = 2* -
X
For this function

/'(x) = 2 + Xi
and equation (2) becomes

= (2 + |) 3,
whence
? = ±2 .

Since J is between 1 and 4 the positive value must be taken, that is,

J = 2.

36. Indeterminate Forms. When a particular value a is substi-

tuted for Xj a function F(x) may take one of the forms
0 00

,
O-OO, 00 - 00, 0°, 1".
0 00

These meaningless expressions are called indeterminate. When x

tends to a, the function F{x) may, however, tend to a definite

limit. In that case we usually take that limit as defining F{a).

Thus, when x = 0,
sin x
Fix) =
X

becomes . When x tends to zero we have seen, however, that

Fix) tends to the limit 1, §10, (6). In a discussion involving this

function we would therefore probably take F(0) as 1.

At a point where a given function becomes indeterminate the

limit can often be found by writing the function in an equivalent
form that does not become indeterminate.
Thus, at j = 1,
x^-1
X — 1

becomes - . For values of x different from 1 the function can be

written
0:
2 - 1
—T
a: + 1,
X

and the right side obviously tends to the limit 2 when x tends to 1.
62 DERIVATIVE AND DIFFERENTIAL [Chap. II

If the limit cannot be obtained in this simple way the following,

known as VHospitaVs rnUj may be useful:

fix) 0 00
If assumes the form - or — when x = a,
<l)ix) 0 00

(1)
« a <t)ix) X -*a 4>\x)

provided the limit on the right exists.

To give a complete proof of this rule it is necessary to consider

several cases.
First suppose that a is finite and /(a) = <^(a) = 0. If

fix)
lim
x^a <t>\x)

exists theremust be an interval of center a in which fix) and 0(a:)

are differentiable and fi>\x) does not vanish except possibly at
X = a. For values of x in this interval, by §34, (4),

fix)
^ fix) - fja) ^ no
<tiix) <t>ix) — <t>ia) 0'(J)

{ beingbetween a and x. As x tends to a, ^ also tends to a and this

equation gives as limit
fix) fix)
lim ‘
lim ,
*a <t)ix) X -^a <l>'ix)

which was to be proved.

Next suppose that a is finite but fix) and <l>ix) become infinite
at X a. If mt f N
/(a:)
lim
x-^a (t)\x)

exists, there is an interval which fix) and 0(x) are

of center a in
differentiable and /(x), and
do not vanish except pos-
<l>'ix)

sibly at X — a. Take c in this interval and x between c and a.

By §34, (4), there is a value ^ between c and a such that
fix)-fic) no
(3 )
4>ix) - <l>ic) </>'(?)
whence
1 _
§36] INDETERMINATE FORMS 63

By taking c sufficiently close to a we can make the second factor

on the right as nearly equal to its limit as we please, for all posi-
tions of f between c and a. Then leaving c fixed and taking x still
closer to a the first factor can be made as nearly equal to 1 as we
please. Thus there is an interval within which the difference

/(^) /'(^)
lim
x-*a(i>'{x)

is numerically less than any assigned number, whence (1) again

follows.
Finally, by only slight modification of the above discussions
we can show the results also valid when a is infinite.

0
In applying THospitars rule, if
f'{x)
-
'v
also becomes - or — at
00

<j>'{x) 0 00

a: = numerator and denominator may again be replaced by

a,

their derivatives and this process repeated as many times as neces-

sary. At any stage of the work it may be helpful to replace the
fraction by an equivalent analytical form.
Other indeterminate forms, such as 0*oo, 00 -- 00 etc., can be ,

0
reduced to - or — by analytical transformation.
00

0 00

1 — cos 2a:
Example 1 . lim
*0

When a: = 0, numerator and denominator of this fraction are both zero.

Replacing both by their derivatives, we obtain

lim
1 cos
-„
2a:
= lim
0
—
2
-r
2a:
—=
sin 2a:
lim
sin 2a:

0
When a: = 0 the fraction on the right becomes We therefore again
^
replace numerator and denominator by their derivatives, thus finally
obtaining
1 — cos 2a: .. sin 2x 2 cos 2a:
lim -
= lim- = lim- = 2 .

X ->0 X X 0 1

1
Example 2,
sin (a: — a) \
When a: = a, the expression in brackets becomes 00 — 00 . Reducing
to a common denominator we get

sin (x — a) — {x — a)
1 1 _
X — a sin (x — a) (x — a) sin (x — a)
64 DERIVATIVE AND DIFFERENTIAL [Chap. II

which becomes ~ when a; = o. Replacing numerator and denominator by

their derivatives, we obtain

cos (a; — a) — 1
(x — o) cos (x — a) + sin (x — a)

Since this is also - when x = a, we again replace numerator and denom-

inator by their derivatives, thus finally obtaining

r - sin (x - a)
1 ^ ^
*-*012 cos (x — a) — (x — o) sin (x — o) J

PROBLEMS
8.
1. A particle moves the distance
5 = lOe + 16^2

in t seconds. Find its average speed during the first t seconds and its speed
at time ^ = 0.
.
2. A body thrown upward with a speed of 100 feet per second rises to the
height
= loot - let®

in t seconds. Find its average speed between t = 2 and t = 2.1, and its speed
at time t = 2.
A particle moves the distance « = t® feet in t seconds. Find its average
speed between t = 1 and t = 3. Also find its speed at t = 2.
4 . A particle moves the distance

a = t® - t® + 16t

in t seconds. Find its speed at time t = 0.

6 A body thrown downward with a speed of 60 feet per second falls the
distance
h -^50t + 16t®

feet in t seconds. Find its average speed during the first t seconds and its

speed at the end of t seconds.

6 . A particle, started with speed h and moving against an opposing force,
travels the distance , ^
8 \kt^

in t seconds, k being constant. Find when it comes to rest and its average
speed during the period of motion.
7. Find the slope of the line through the points Pi(l, 2) and P2(3, 6).
8 . Find the slope of the line through the origin and the point (1, —3).
Find the angle from the positive direction of the x-axis to the line joining
9.

Pi(0, 2) and P2(3, 6).

10 Find the angle from the positive direction of the x-axis to the line join-
ing (3, -1) and (-3, 6).
 PROBLEMS 65

11 . Find the slope of the curve y x the point (1,2).

12 . Find the slope of the curve ?/ = 4a; — at the origin.
13 . Find the slope of the curve y = + x + 2 at the point where x == 1.

14 . Find the slope of the curve t/ = x® at the point where x = x\.

16 . Find the slope of the curve 2/
== x** at x = 0, (a) if n > 1 (6) if n = 1.
16 . Find the slope of the curve y = sin x at the origin [§10, (6)].
17 . Find Ax when x changes from a to 6.
18 . If 2/
=
find At/ when x changes from 1 to 2.
19. Given y — x®, find At/ if x = 2 and Ax = —1.

In each of the following problems show that the derivative has the value
indicated:

20 f(x)
. 2x +3,/'(x) = 2.
21 fix)
. x^ — X, /'(x) = 2x — 1.

22 fix)
. x^ f\x) = 4x^

23 fix)
.

X 4- 1 (X + 1)2

24 fix)
. \.nx) =

26 fix)
. y/x - i,/'(x) =
2Vx
In each of the following problems find d?/, assuming that x is the independent
variable:

26 T/. = x^ — X.

27 . y = x2 -f 3x + 2.
28 y . = y/x.

30 If T/ = x2, X =
. and t is the indei)endent variable, find dx and dy,
and show that they satisfy the equation

dy = 2x dx.

31 If X = and is the independent variable, find dy and dx, and

.

p T/ t

show that their ratio is the derivative of y with rc'spcct to x.

32 Suppose that t is the indeptuident variable and s the number of miles

.

an automobile travels in t hours. Taking A/ = 1, make a stat/oment concern-

ing ds or As in each of the following cases:
(a) At time t the automobile is traveling 60 miles per hour.

(5) During the following hour the automobile travels 60 miles.

33 Let P be the population of a country at time t measured in years.
.

Taking t as independent variable and A^ = 1

(a) Find dP if the population is 1,000,000 and is increiising 1% per year.

(6) Find AP if the population is 1,000,000 and increases 1% in 1 year.

34 A coat of paint of thickness t is put on the surface of a sphere of radius r.
.

Make a common-sense formula for the volume of paint used. If v is the vol-
66 DERIVATIVE AND DIFFERENTIAL [Chap. II

ume
36. of the sphere and r is the independent variable, determine dv when
Ar = t,

A coat of paint of thickness t is put on a cube of side x and volume F.

Determine approximately the volume of paint used. If x is the independent
variable and x + Ax is the side after painting, find Ax and dV.
36. Let V be the volume of water in a reservoir, A the area of its surface,
and h the depth of water over a fixed level, all measured in feet. Determine
approximately the increase in supply when the water level rises 1 inch. Taking
h as independent variable, determine dV when A^ is 1 inch.
37. When a circular plate is heated its radius increases from 10 in. to 10.1 in.
Determine approximately the increase in area.
38. If X is the independent variable and /(x) = -\/x» A/’(^) and df(x):
(a) when x changes from 25 to 26; (b) when x changes from 25 to 25.01. In
each case express the difference A/(x) — d/(x) as a fraction of Ax.

39. If X is the independent variable and f(x)- find A/(x) — d/(x).

== ,
X
(a) when x changes from 1 to 1.1; (b) when x changes from 1 to 1.001. In
each case express the difference as a fraction of Ax.

Find^ in each of the following exercises:

dx
40. y = x^ — 4x^ + 2x^ -f- 4x — 6. 41. 2/ \x^ - f + 1.

42. y = 4(x3 - 3x2 ^ 3^)^ 43. y + 3x2 ^ 4).

44. 2/
== x^ + x^. 46. y

46. y = 2Vx3 —y/x

6
47. y (2x - 1)1

48. y = (x2 -f X + 1)^. 49. y Vx2 -- 2x + 3.

60. y = (x - 3
1 ) (3 x - 1 ). 61. y :

(x H- o)(x2 — ox + o2).
62. y = (X + 1)2(2x ~ 3)2 63. y (x - 2 )Vx~+T.
64. y =X V2 - x 2
. 66. y -
(3x2 _ 2a2)(x2 + a2)J.
-
3x +4 *
67. y
66. y
2x +3 1 - x*
\/4x2 - 1
68. y 69. y :

Vo*
(x» - 2)1 61. y
+1 x
60. y * =

Vx* + 2x + 2
x*
62. y 63. y =

( 2 x« + l)i
(3x - 1)«
64. y 66. y = '
3 sin ( 4x).
X*

66 y . .2cos(x+0. 67. y = '

X — sin X.

68. y X sin X + cos X. 69. y : '

sin2
(2xJ^^
 PROBLEMS 67

70 . By differentiating both sides of the equation

sin 2x — 2 sin x cos x

obtain a formula for cos 2x.
71 . Leaving y constant and differentiating both sides of the equation

sin (x + 2/)
= sin x cos y + cos x sin y
.

with respect to x, obtain a formula for cos (x -hi/).

72 . Find the angle between the x-axis and the tangent to the curve

at the origin.
73 . P"ind the points on the curve

y = (x - l)*(x + 2)*
where the tangent is parallel to the x-axis.
74 If X = cos df y = sin
. find dx and dy^ and show that their ratio is the
derivative of y with respect to x.
76 By differentiating both sides of the equation

cos^ 0 + sin^ 0 = 1

and equating derivatives, do we get a correct result?

76 . By differentiating both sides of the equation

x2 4* 2x = 1

and equating derivatives, do we get a correct result?

77 The area of a
. circle is calculated on the assumption that the radius is
8 inches. If the radius is actually 7.8 inches, determine the relative error and
the percentage error in the calculated area.
78 Determine the allowable percentage error
. in the diameter of a sphere
if the calculated volume
is to be correct to 0.3%.

79 The time of vibration of a simple pendulum of length

. I is

If a clock loses 1 second per day, determine approximately the error in the
length of its pendulum.

80 The volume of a cylinder is determined from its radius and altitude.

.

If themeasurements of radius and altitude are correct to within 1%, deter-

mine the maximum error in the calculated volume.
81 The radius of a spherical ball is calculated from the weight of the ball
.

and the density of its material. If an error of 0.6% is made in weighing the
ball and 1% in determining its density, what is the maximum error that may
result in the determination of the radius?
 DERIVATIVE AND DIFFERENTIAL [Chap. II

dpy ,

Find in each of the following problems:

82. y = \/2a: + 3. 83. y = sin 2x.

84. j/ - (2 - 86. X2/ 4- 32/ - X + 1

86. j/* = 2ax. 87. x^ +y'^ - 5.

s 1 li 89. x^ + xy + y^ - 5.

90. + 2xy — 2
/^ = 0. 91. xi + 2/4 = 1.

92. X = cos 0, y = sin e. 93. X == t

+ ^yy ^ t

94. Find the derivative of

dx
o
1 +
with respect to t.

96. Given y = x®, x = show that

d?y /dxX*
\¥/
is not equal to
A
dt^'

Find the values of x for which the following functions increase as x increases
and the values for which they decrease. Also find the values at which the
curves are concave upward and those at which they are concave (

96. y = x^ — 4tx + 4:. 91. y = 0? - 3x.

98. 2/ = + 3x2 + 4. 99. 2/ = ic

—
X
100. 2/
=* — 4x^ +L 101. ^ = x^ + 4x + 3.
102. Show that
x3 — 3x2 _|_ 3^ _l_ 2

has no maximum or minimum.

103. Show that

x + X-
has a maximum and a minimum but that the maximum is less than the mini-
mum.
104. A piece of wire 40 inches long is bent into a rectangle. Find the max-
imum area of the rectangle.
106. Find the minimum perimeter of a rectangle of area 100 square inches.
106. A box with a square base and open at the top is to be made from 300
square inches of material. If no allowance is made for thickness of material
or waste in construction, find the dimensions of the largest box that can be
made.
 PROBLEMS 69

107. A box is to be made from a piece of cardboard 6 inches square by cutting

equal squares from the corners and turning up the sides. Find the dimensions
of the largest box that can be made.
108. A tin can open at the top is to have a given capacity. Find the ratio
of height to diameter if the amount of tin required is the minimum.
109. Find the volume of the largest right circular cone that can be inscribed
in a sphere of volume 27 cubic inches.
110. Find the altitude of the largest cone that can be generated by rotating
a right triangle of hypotenuse 2 feet about one of its sides.
111. Among all circular sectors with a given perimeter find the one which
has the greatest area.
112 The same distance was measured four times, the results being ai, 02
.
»

azy 04 - By the method of least squares the most probable value for the cor-
rect distance is that which makes the sum of the squares of the four errors
least. What Ls that value?
113. A gutter of trapezoidal section is made by joining three flat strips
each 4 inches wide, the middle one being horizontal and the other two inclined
the same angle. How wide should the gutter be at the top to have maximum
at 116.
capacity?
114. To
reduce the friction of a liquid against the walls of a channel, the
channel sometimes so designed that the area of wetted surface is as small as
is

possible. Determine the ratio of width to depth for the best open rectangular
channel with given cross-section area.
A circular filter paper of radius 3 inches is to be folded into a conical
filter. Find the radius of the base of the filter if it has maximum capacity.
116. What are the most economical dimensions of an open cylindrical water
tank if the cost of the sides per square foot is twice that of the bottom?
117. If the top and bottom margins of a printed page are each of width a,
the side margins of width 6, and the printed matter covers an area c, what
should bo the dimensions of the page to use the least paper?
118. Assuming that the intensity of light is inversely proportional to the
square of the distance from the source, find the point on the line joining two
sources, one of which is 8 times as intense as the other, at which the illumina-
tion is least.

119. A ship B is 75 miles due oast of a ship A. If B sails west 12 miles per
hour and A south 9 miles per hour, find when the ships will be closest together.
120. A fence a feet high runs parallel to, and 6 feet from, a wall. Find the
shortest ladder that will reach from the ground over the fence to the wall.
121. A corridor of width h runs at right angles to a passageway of width a.
What is the longest beam that can be moved in a horizontal plane along the
passageway into the corridor? Show that the value obtained is a minimum.

Explain.
122. A and C are points on the same side of a plane mirror. A ray of light
passes from A to C by way of a point B on the mirror. Show that the length
of path will be least when the lines BA, BC make equal angles with the per-
pendicular to the mirror.
123. Let the velocity of light in air be Vi and in water F2 The path of a .

ray of light from a point A in the air to a point C below the surface is bent
at B where it enters the water. If di and O 2 are the angles that AB and BC
70 DERIVATIVE AND DIFFERENTIAL [Chap. II

make with the perpendicular to the surface, show that the time required for
light to pass from A to C will be least if B is so placed that

sin
_ Fi
sin 02 V2
124. Assuming that the cost per hour of propelling a steamer is proportional
126.
to the cube of its speed through the water, find the speed at which it should
be run against a current of 5 miles per hour to make a given trip at least cost.
In the preceding problem find the most economical speed if the steamer
moves with the current.
126. A wire of length L is cut into two pieces, one of which is bent to form
a circle, the other a square. Find the lengths of the two pieces when the sum
of the areas of the square and circle is greatest.
127. Find the point (x, y) on the curve 2/^ = 2 (a; — 1) at shortest distance
from the origin.
128. Find the point on the x-axis the sum of whose distances from (0, 1) and
(1, 0) is least.
129. Find the minimum value of the function

2/
= 1 + (x - 1)1

180. Find the maximum value of the function

2/=x- 2 la;~ll.
131. A log is 12 feet long and has diameters 4 feet and 5 feet at its ends.
Find the volume of the largest cylinder that can be cut from the log.
132. A particle moves along a straight line,

s = 32« -
being its displacement from a fixed point of the line at time t. Find its ac-
celeration at the point where its velocity is zero.
133. At time t the displacement of a particle from a fixed point of its line

of motion is

5 = ^3 _ 0^2 9^

() During what interval is its velocity negative? (6) During what interval
does its velocity decrease?

In each of the following problems a particle moves along a line, x being its
displacement from a fixed point of the line at time t. At the instant indicated
determine, (a) whether the particle is moving toward the fixed point or away,
() whether it is accelerated toward the fixed point or away, (c) whether the
speed is increasing or decreasing:

184. X = sin 3^, t « 136. a; = 3 -f 5^ - « 2.

4
186. X - <* - 4< - 1, i = 0. 187. x = 1 - cos = ^ •

3
188. X - «* + 6<, « - -1. 189. X - - 8<®, « = 1.
 PROBLEMS 71

140 . At time t the displacement of a particle from a fixed point of its line
of motion is
« = Vi^FTc,
k and C being constants. Show that the velocity of the particle approaches k
as limit as t tends to infinity and that its acceleration is inversely proportional
to the cube of s.

141 . A body, starting from rest, has the velocity

V = ^y/h ft ./sec.

when it has fallen h feet. By differentiating with respect to the time, and
eliminating
dh
dl
determine its acceleration.
.

142 . A particle, starting from rest at an indefinitely great distance and fall-
ing under gravity toward the earth, has the velocity v determined by the
equation

when at distance r from

. the center of the earth. By differentiating with respect
to the time find its acceleration, determining the constant k from the known
acceleration g at the surface of the earth.
143 . A particle moves along a straight line, x being its displacement from a
fixed point of the line, v its velocity, and a its acceleration. Show that
dv
v— — a.
dx

144 . A point moves along the parabola

= 2pz/

in such a way that its projection on the x-axis has constant velocity. Show
that projection on the 2/-axis has constant acceleration.
its
146 A ladder 60 feet long rests against a vertical wall. If the foot of the
ladder is pulled away from the wall at the rate of 3 feet per second, how fast
is the top moving when the foot is 30 feet from the wall?

146 A man 6 feet tall walks at the rate of 5 feet per second directly away
.

from a lamp 15 feet above the ground. Find the rate at which the end of his
shadow is moving and the rate at which his shadow is growing when he is
20 feet away from the lamp post.
147 . A ball falls vertically
5 = 16^2

feet in seconds from a point 64 feet above the ground. Find the speed of its
t

shadow on the ground just before the ball strikes if the sun's rays make an
angle of 60® with the ground.
148 A kite is 300 feet high, and 500 feet of cord are out. Assuming the
cord to stretch in a straight line, if the kite moves horizontally 10 feet per
72 DERIVATIVE AND DIFFERENTIAL [Chap. II

second directly away from the person flying it, how fast is the cord being paid
out?
149. The side of a square is increasing 10 feet per minute and its area 100
square feet per minute. Find the side of the square.
160. A basin has the form of an inverted cone of altitude 10 inches and
diameter of base 10 inches. If water runs in at the rate of 5 cubic inches per
second, how fast is the water level rising when the depth is 6 inches?
161. At what rate is the wetted surface increasing in the preceding problem?
162. A reservoir covers 100 acres. At what rate must water be pumped in
to raise the level 1 foot per day?
166.
In each of the following, find the value f in the mean value theorem:

163. /(a;) = a;2 - + 3,

4a; a = 1, 6 = 4.

164. f(x) = a;2 + a =

4a;, -3, 6 = 5.

fix) = a;3 -
+ 7a;2 36, a = -3, 6 = 2.

166. fix) = a - 0, 6 = 9.

167. /(x) = ® i ,
a = 1, 6 = 9.
X

168. fix) = —+^

a; 1
,
a = 0, 6 = 8.

169. If a =» -1, 6 = 1,

fix) = xi,
show that there is no value ^ that satisfies the mean value theorem. Explain.
160. Given fix) = </>(x) = a = ~3, 6 = 2, show that there is no
value ^ between a and 6 that satisfies equation (4), §34. Explain.

Determine the following limits:

A
x^o - 1 x» — a»A
161. \im-^ 162. lim
X - I X* — 1 -*a X
X -> t —a
X — sin X 1 — sin X
163. lim 164. lim
0 X® TT— 2x

166. lim
2 -X 2VT
* 0 X -*1 X — 1

167. lim sec X + *

IT
168. lim
X -*0 Lsin X xJ
2J
sin 2x — 2 sin x .. 2sinx + 8in2x
169. lim 170. hm
» 0 cos X + cos 2x
 CHAPTER III

INTEGRATION AND SUMMATION

36. Integral. In many problems the derivative or differential

of a function Ls known and it is necessary to find the function.
Thus, if the speed of a moving particle at time t is

ds
dt

the distance it moves in time t is a quantity s with differential

ds = f{t) dt.

To find the distance we must then find a function of t with differ

ential equal to/(0 dt.

The process of finding a function with given differential is called

integration. If
dF(x) = f(x) dx

then F{x) is called an integral of f{x) dx, and this is indicated by

the notation

F(x) = J fix) dx.

By differentiation we pass from a function to its differential. By

integration we pass from the differential to the function. Inte-
gration is thus the inverse of differentiation.
For example, since d{x^) = 2x dx.

2x dx = x^.

Similarly,

cos = sin x, sin xdx — —cos x.

J * a: da:

The test of integration is to differentiate the answer. If the

integration is correct, the differential of the answer must equal
the expression to be integrated.
73
74 INTEGRATION AND SUMMATION [Chap. Ill

It should be noted that we always integrate a differential and

not a derivative. There are several reasons for this. A very im-
portant one is the fact that the differential of a quantity is inde-
pendent of the variable in terms of which it is expressed. If the
same quantity y is expressible in terms of two different variables

y = y = /2(0,
has equal values whether the one or the other of
its differential

these expressions is used. But the derivatives of y with respect

to X and t are not in general equal. If integration were defined
as the process of finding a function with given derivative, there
would be a different integral for every variable.
37. Constant of Integration. In solving a problem by integra-
tion we need to know whether the given differential has more than
one integral; for, if there are several different integrals, we must
choose the one that represents the answer.
As a matter of fact each differential has an infinite number of
integrals but they are related in a rather simple way. If

dFix) = f{x) dx
and C is any constant,
d[F{x) + C] = d F{x) = f{x) dx.

Thus, if F{x) is one integral of a given differential,

F{x) +C
is another.
Conversely, if two functions have equal differentials at all points

of an interval, their difference is constant in that interval.

To prove this suppose that

dFi{x) = dF2 {x) = f{x) dx

at all points of the interval a ^ b. Let

^ x

= Fzix) - Fi{x).
By hypothesis
4>(x) = —
dF 2 (x)

dx
dFi{x)

dx
—= 0

at all points of the interval a ^ x ^b. For any value of a: in that

interval the mean value theorem [§34, (1)] then gives

4>(x) — <l>(a) — (x — a) <l>^(xi) = 0.

§381 INTEGRATION FORMULAS 75

Thus
tl>(x) = 0(a)
and so the difference
IMx) ~ F,(x) = 0(x)

has the same value <t>(a) at all points of the interval.

If one integral of a given
then F(x) is differential / (x) dx, any
other has the form

=
y fix) dx F(x) + C.

Any constant value can be assigned to C. It is therefore called an

arbitrary constant. In the statement of a definite problem there
is usually some information by which this constant can be de-
termined.
38. Integration Formulas. Integration is usually performed by
means which the following will be sufficient for the
of formulas of
applications made in this chapter. In these a, n, c are constants
and u, V functions of a single variable.

I. (du + dv) = J" du + J* dv.

J1

II.
1
a du ^ a
J
.n+l
III. 1 u^ du
= -
•>

IV. sin udu =

J

V. f cos udu -

Any one of these formulas can be proved by showing that the

differential of the right member is equal to the expression under

the integral sign.

Formula I states that the integral of a sum is equal to the sum
of the integrals of its parts. Integration like differentiation is

thus a distributive operation (§21).

Formula II states that a constant factor may be transferred
from one side of the integral sign to the other without changing
the result. Integration and multiplication by a constant are thus
76 INTEGRATION AND SUMMATION [Chap. Ill

commutative operations. It should be noted that a variable

cannot be transferred in this way. Thus

is not equal to

X / dx.

In applying formulas III, IV, and V it is important that du be

the differential of the quantity called u. Thus we cannot evaluate

cos 2x dx

by immediate application of V, for dx is not the differential of 2x,

We can, however, write
dx = ^ d{2x)
and so obtain

+
^ cos 2xdx —
^ J*
cos 2x d{2x) = \ sin 2x C.

Example L J'ix
— l)(x + 2) dx.
Expanding and integrating term by term, we obtain

y* (a; — l)(a; + 2) dx = J {x^ + x — 2) dx = \x^ + — 2a: + (7.

Example 2.
J y/2x + 1 dx.

Taking 2a: +1= w, we have

dx — \du
nd

j \/2a: + 1 dx = ^
du = +C= ^ (2x + 1)^ + C.

Example 3. sin (ax + b) dx.

J*
Taking oa: +6= w, we have
dx = - du
a
and
I r - — 1
+ b) dx = - + 6) + C.
J siaudu
ain (ax cos
/ - (oa:
§39] VARIABLE WITH GIVEN RATE OF CHANGE 77

39. Variable with Given Rate of Change. If the rate of change

of a variable 2 is
; a known function of the time

= /( 0 ,

integration gives

fit) dt +c
as its value at time t
The constant C can be determined if we know the value of z at
some particular time. Thus, if z has the value Zq at time we
find C by substituting
z = zoy t = to

after integration.

Example 1. From a point 60 feet above the ground a ball is thrown

upward with a velocity of 100 feet per second. If it is subject only to the
acceleration of gravity, find its height t seconds after the beginning of
motion.
Let y be the height (considered positive upward), and

the velocity of the ball at time t. Its acceleration is then

the negative sign being used because the acceleration is opposite to the
direction in which y is measured. By integration wo find

y = + Cl,

Cl being constant. Substituting the initial values

y = 100, e = 0,

we obtain
100 = Cl,
whence

^= y = + 100.

A second integration gives

y = ^hQt^ + 100< +C 2.
78 INTEGRATION AND SUMMATION [Chap. Ill

Substituting
!/ = 60, i = 0

we find the constant is equal to 60 and so

y = + ioo« + 60
at time L
Example 2. Suppose that, through condensation upon its surface, a
spherical drop of liquid grows at a rate proportional to its area of surface.
with radius ro, find its radius at time t.
If it starts
Let r be the radius, V the volume, and S the surface area of the drop
at time t By hypothesis

where k is constant. Substituting

V = ^7rr^ 6^ = 47it2,

we obtain
Ht
4:Trr^ “- = A; • 47rr®,

whence

Integration gives
r = kt + C,
C being constant. Substituting r — to, t — we have

ro = C,

and so
r ^ kt + To
is the value of the radius at time L

40. Curve with Given Slope. If the slope of a curve is a given

function of x,
dy
ax
integration gives

2/ = / /(^) dx +C
as the equation of the curve.
Since the constant may have any value there are infinitely many
§40] CURVE WITH GIVEN SLOPE 79

curves having the given slope. These are obtained from one,

y = dx,

by moving it through various distances C in a vertical direction.

If the curve is required to pass through a given point P, the value

of C can be found by substituting the coordinates of P after in-

tegration.

Example L Find the curve passing through (1, 2) with slope equal
to 2x.
In this case

Hence
y = J'2xdx = + C.
Since the curve passes through (1, 2) the values x — y = 2 satisfy the
equation; that is,

2 = 1 + 0 .

Consequently 0=1, and

y = x^ + I

is the required curve.

Example 2, On a certain curve

= 6x.

If the curve passes through (—2, 1) and has at that point the slope 4, find

its equation.
80 INTEGRATION AND SUMMATION [Chap. Ill

By integration we get

^ = Jf^x dx — Sx^ +
ax
Cl.

At (-2, 1), a: —2 and ^

ax
= 4. Hence

4 == 12 + Cl,
or Cl = —8. Thus
^ = 3x* - 8
and so
y = x^ — 8x +C 2.

Since the curve passes through (— 2, 1), C2 = —7, and

y = x^ Sx — 7

is the equation of the curve.

41. Area as Limit. The methods of elementary geometry de-

termine directly only the areas of regions with straight-line bound-
aries. To define the area of a plane region with curved boundary
it is customary to consider the region as limit of a set of regions

with rectilinear boundaries.

Thus to define the area ABQP (Figure 42) bounded by the

x-axis, a curve
y = fix)
above the a:-axis, and ordinates at A and 5, divide AB into equal
§41] AREA AS LIMIT 81

or unequal parts Ai, A 2 An, and on each part hi as base con-

>
* •

struct a rectangle of altitude equal to the ordinate yi at some point

of the arc above this part. The set of rectangles thus constructed
forms a polygonal region which is taken as an approximation for
the region bounded by the curve, and the sum of the areas of those
rectaniiles
>Sn = h\yi + A22/2 H— + An *
2/n

is taken as an approximation for the area between the x-axis and

the curve.
Now suppose that the number n of parts into which the base AB
is divided is made greater and greater but in such a way that the
lengths Ai, A 2 ,
• • *
An of all the parts tend to zero as n tends to in-
finity.If the sum of the areas of the rectangles tends to a limit

(as italways does in cases in which we are interested) that is

independent of the way the base AB is cut into parts hi and inde-
pendent of the choice of ordinates yi on those parts, that limit is
defined as the area of the region ABPQ,
This summation process provides not only a definition for an
area of the above type but also a practical method for calculating
its value to any desired approxi-
mation.

Example, Find approximately the

area bounded by the x-axis, the curve

and the ordinates at x = 1, x = 2.

To do this we shall divide the in- Figure 43.
terval from 1 to 2 into 5 parts. There
being no particular reason to do otherwise, we make all these parts equal.
The parts in question are then determined by consecutive values of x
in the following table:

X = 1, 1.2, 1.4, 1.6, 1.8, 2.

In the strip above each part we are to choose an ordinate to serve as alti-

tude of the approximating rectangle. Since the function

steadily decreases as x increases, the result will be too large if in each case
we take the left-hand value and too small if we take the right. As a
82 INTEGRATION AND SUMMATION [Chap. Ill

reasonable compromise we take the middle ordinate in each case. The

altitudes of the rectangles are then

JL _L JL _L
l.r 1.3’ 1.5’ 1.7’ 1.9’

and the sum of the rectangles is

The area correct to three decimals is 0.693.

42. Definite Integral. Many other quantities besides area are

represented by limits of essentially the same form as that just
described. A limit of this kind has therefore been given a special
name and notation defined as follows:
Let/(x) be a function defined for each value of x in the interval
(a, 5). Divide this interval into n equal or unequal parts by inter-

mediate values X2 x^, •f Xn arranged

y
•
in order from o = a*i to
•

b = Xn^i- In each part {xi, choose arbitrarily a value

which may be interior or at either end of this subinterval. Form
the sum

Sn =/({l)(a?2 — cl) +f(^2)i^3 — ^2) H |-/(fn)(6 Xn). (1)

§43] EVALUATION OF A DEFINITE INTEGRAL 83

This slun is often represented by the notation

n

Aa;,-.

Consider a sequence of such sums obtained by dividing (a, 6) into

a larger and larger number of parts but in such a way that the
increments
diXi = Xij^i - Xi (3)

all tend to zero as n tends to infinity. If these sums tend to a limit

independent of the way the interval (a, h) Ls cut into parts and inde-
pendent of the choice of values fi, *, fn in those parts, the
* •

function J{x) is said to be integrable in the interval (a, b) and the

limit

lim Sn = lim Axi

n —« 00 n 00

is called the definite integral of J{x) between the limits x = a and

X = 6.

It is represented by the notation

f f(x) dx. (5)

The number a is called the lower limit, b the upper limit, of the
integral.
If 6 > a the sum Sn is represented graphically by the area of a
set of rectangles, as in Figure 44, those above the x-axis being
considered positive, those below negative. The definite integral
is thus the difference of the areas above and below the axis.
43. Evaluation of a Definite Integral. To distinguish the defi-
nite integral

from the integral

/ /(x) dx

//w dx
defined in §36, the latter is often called an indefinite integral. The
reason for these terms is that the definite integral has a definite
value whereas the indefinite integral is any one of a set of func-
tions differing by constants.
84 INTEGRATION AND SUMMATION [Chap. Ill

The definite and the indefinite integral of a given function have

a relation indicated by the following theorem:
If fix) is continuous and

d Fix) = fix) dx (1)

at each point of the interval ia, 6),

/ fix) dx = Fib) - Fia). ( 2)

The difference Fib) — Fia) is represented by the notation

[F(a;)]t The above equation is thus written

= -
/ fix) dx = [F(x)]S Fib) Fia). (3)

To obtain the definite integral of a continuous function fix)

we thus determine any indefinite integral

Fix) fix) dx
-f
and substitute in the right side of (3).
To prove the theorem, divide the interval (a, b) into parts by
values X2 x^, ,
•
Xn arranged in order from a = a:i to 6 = Xn+i.
•
•,

Since fix) is the derivative of Fix), by the mean value theorem

there is some number between Xi and Xij^i such that

- Fix^) = fiki)ixiJ^i “ x^) = /(Ji) ^Xi,

There are n such equations given by i = 1, 2,

• •
•, n. Adding
these we get

Fix2 ) - Fia) + Fix^) - Fix2 ) + • •

• +F(6) - F^ n

i 1
which is equivalent to
^

Fib) -Fia) = V/tt.) Ax,-,

frl
Since the left side of this equation is constant the right side tends
to that constant as limit when the increments Ax,- all tend to zero.
Thus n
§44] AREA UNDER A CURVE 85

We have shown this, however, only when the numbers have the
values determined by the mean value theorem. If in each sub-
interval {xi, Xi^i) any other value is used, the sum has the form
n

Aa:.-.

Since S{x) is continuous and both belong to an interval of

length Aa:* we shall have

i/(^:) -/fe) I
< *

when all the increments are in numerical value less than some
number 5 [§14, (5)]. Consequently,

1 mi) Ax.- - S/(?.-) Ax.-

1
= mii) - S{k^)\ Ax.-
1

< «S Ax.- = 6 — o
I 1 «| |.

As the increments Axi all tend to zero, 5, and consequently e,

may be allowed to approach zero as limit. Thus, whatever posi-

tions the numbers have in their respective subintervals (x^,
Xi+i), we have
n n

lim y;/a:.)Axi= Um V/a,) Ax, = F(6)-/^(a),

which was to be proved.

Exam'ple, Find the value of

w
f
2 COS X dx,
J0
By formula V, §38, an indefinite integral of cos x is

=
Therefore
J cos X dx sin x.

cos xdx = [sin x]^ = sin — sin (0) = 1.

J"
44. Area under a Curve. We have seen that, if 6 > a and f(x)
isa positive continuous function, the region bounded by the x-axis,
the curve y = /(x), and the ordinates at x = a, x = 6 has the area
rh
A = I
ydx = I /(x) dx. (1)
86 INTEGRATION AND SUMMATION [Chap. Ill

If y = f(x) has negative values between a and 6, this formula

may still be used but the integral will then be equal to the area
above the aj-axis, minus that below.
An
area which does not have boundaries of exactly the form
stated can usually be expressed as a sum or difference of such areas.

Example 1, Find the area bounded by the x-axis, the curve y = I +

and the ordinates at a; = — l,a; = 1.
The required area is shown in Figure 45. By equation (1) it is

A dx = [x + i - (-i) = f.

Example 2, Find the area bounded by the curve x = 2 2/

“* 2/^ +
and the ^-axis.
The curve (Figure 46) crosses the 2/-axis at x = 2 + 2/ — 2/^ = 0, that
is, at 2/ = —1, 2/ = 2. Since the area is bounded by the curve and the
y-axis it is the limit of a set of rectangles of sides x and Ay extending from

2/
= —1 to 2/
= 2. Its value is

therefore

lim y^xA = •'“1

/* (2
2/ + 2/
— 2/^)
d2/
Ay -»0

= [2j/ + W-
Example 3, Find the area
bounded by the curves

y = 2/
= 4x — X®.

Solving the equations simultane-

Figure 47. ously, we find that the two curves
§46] VOLUME OF A SOLID OF REVOLUTION 8V

intersect at a: = 0 and x = 2. Taking yi = 1/2 = 4a; — a;^,the area

isthe limit of a set of rectangles of base Aa; and altitude ^2 — Vh extending
from a: = 0 to a; = 2. Thus

^ ^^ ““ ^ t*

46. Volume
of a Solid of Revolution. The region bounded by
the a curve y = f{x), and two ordinates a; = a, a; = 6 is
a:-axis,

rotated about the a;-axis. We wish to determine the volume of the

solid generated.
To do this divide the interval (a, h) into equal or unequal parts
and on each part (a:, x + Ax) construct a rectangle of base Ax and
altitude equal to the ordinate /(J) at some one of its points x =

When rotated about the x-axis this rectangle generates a cylinder

of radius /({), altitude Ax, and volume

Am)?
The volume generated by the set of rectangles is

If /(x) is continuous between a and 6, when the number of divisions

increases and all the increments Ax tend to zero, this sum tends
to the limit ^ ^

y= f ^mx)fdx= f wy^dx, (1)

•/a •/o

which is defined as the volume of the region generated.

This formula shows that the volume of the solid is numerically
equal to the area between the x-axis and the curve

y = V [f(x)]^.
88 INTEGRATION AND SUMMATION [Chap. Ill

Similarly, each application of the definite integral is made essen-

tially by showing that the quantity in question is a limit of the
same form as that which defines a certain area. For this reason
the determination of a quantity by definite integration is often
called quadrature.
If a solid is generated by rotating an area, not about the a:-axis
but about some other line, and all cross sections perpendicular to
the axis are circles, its volume is

y- •Ja
pirRUh,

where h is distance measured along the axis, R the radius of the

cross section at position A, and a, h limiting values of A.

Example 1. The area bounded by the a;-axis and a curve

y — 2x —
is rotated about the a;-axis. Find the volume generated.
The curve crosses the a;-axis at a; = 0 and x = 2. The volume required
is therefore

jT iry^ ^ dx = 7r[|x* —

Example 2. The area bounded by the curve y = x^ and the line y = 1

is revolved about the line 2/
= — 1. Find the volume generated.
The area between the curve AOB (Figure 50) and the line AB is re-

volved about the line CD. This area is cut into strips by lines parallel
to the 2/-axis. The strip between x and x +
Ax is approximately a rectangle
of base Ax and altitude I — y. Rotated about the line CD this generates
a hollow cylinder, or washer of inner radius y + I, outer radius 2, thick-
ness Ax, and volume

7r[2^ — {y + 1)*1 Ax = 7r(3 — 2y — if) Ax.

§46] VOLUME OF SOLID WITH GIVEN AREA OF SECTION 89

The entire volume is the limit of the sum of such cylinders extending
from a; = —1 to a; = 1. It is therefore

7r(3 — 2y — y^) dx — (3 — — x^) dx

J' ^
= irlSx -

46. Volume of a Solid with Given Area of Section. Consider

a solid cut by planes perpendicular to an axis along which a co-
ordinate h is measured. Let A(h) be the area of section in the

plane which crosses the axis at position h. Divide the portion of

the axis between h a and h = b into equal or unequal parts,
=
and pass planes perpendicular to the axis through the points of
division. The slice between the planes at /i, /i Ah is approx- +
imately a plate (not necessarily circular) of face area A(h)j thick-
ness Ah, and volume A(h) Ah, The entire solid is approximately
a sum of such plates of total volume
b

y^Ajh) M.
a

When all the increments Ah tend to zero this sum approaches the
volume of the solid

y=
f aqi) dh
as limit.
go INTEGRATION AND SUMMATION [Chap. Ill

To determine the volume of such a solid we thus merely express

the area of section AQi) in terms of h and integrate between the
appropriate limits.

Example, Theaxes of two equal right circular cylinders intersect at

right angles. Find the common volume.
Let a be the radius, OY and OZ the axes of the cylinders, and OX a line
perpendicular to both axes. In Figure 52, OABC is one-eighth of the

X
Figure 62.

common volume and PMQ is a section perpendicular to OX at distance x

from the center. OP, OQ are radii and OMP, OMQ right triangles
Since

MP = MQ= Va' -
Thus the section PMQ is a square of area
A(x) = MP-MQ = o2 -
The volume common to the two cylinders is therefore

V = s£\{x) dx = SjC V -X^)dx =

§47] PRESSURE 91

47. Pressure. A fluid at rest in contact with one side of a flat

plate exerts a force perpendicular to the surface of the plate.
If Af is the magnitude of this force on an area AA, the limit

AF
V = lim ) (1)
AA -*0 AA
as AA shrinks to zero about a point P, is called the force per unit
area, or 'pressure, at P. The pressure at a point is independent of
the direction of the surface across which it is measured and in a
liquid of constant weight w per unit volume
has the magnitude

p = wh (2)

at depth h below the surface.

We wish to determine the force exerted
by liquid pressure against one side of a
vertical plate such as a gate in the face of Figure 63.
a dam. To do this divide the plate into
strips by horizontal lines and
be the width and p the pres-
let I

sure at depth h below the surface. The area between the lines at
depths A, A + A/i is approximately I Ah and the force on this ap-
proximately pi Ah, these approximations becoming more accurate
as Ah diminishes. The total force is thus

F = lim 'Epl Ah =
Afc —0

a, b being the values of h at the top and bottom of the submerged

area.
To determine we express I in terms
the force in a particular case
of h, or I and h a third variable, and integrate (3) be-
in terms of
tween the limiting values of the variable used in the integration.
In the case of water
w = 62.6 Ib./cu. ft. (4)

= ton/cu. ft.,

approximately.

Example 1, A vertical gate in a dam has the form of a square of side

4 feet, the upper edge being 2 feet below the surface of the water. Find
the force it must withstand.
92 INTEGRATION AND SUMMATION [Chap. Ill

In this case Z = 4 and h varies from 2 feet to 6 feet. Hence

F wh-4:dh = [2wh^]l = 2 ions.

Example 2, Find the total pressure on one side of a triangle of base

6 feet and altitude 4 feet if its altitude is vertical and base in the surface
of the water.
At distance x above the vertex the width of the triangle (Figure 55) is

I = -f-x = ^x.

Also X varies from 0 to 4, ^ = 4 — x. Thus

F w(4: -- x)-^xdx = w[dx^ — = 1000 lb.

Work. Let a force F be applied at a fixed point of a body.

48.
When the body moves, work is done by the force. At present we
consider only motion along a straight line under the action of a
force parallel to the line; motion in a curve will be considered
later (§200).
If the force is constant, the work w done by the force is the
product
F = Fs
where F is the force and s the distance moved along the direction
of the force. The work done by a force is positive when the dis-
placement is in the direction of the force, negative when in the
opposite direction.
If the force is variable, represent the position of the body by
itsdisplacement x from a fixed point of its line of motion. Con-
sider the force positive when in the direction in which x increases,
and let F{x) be its value when the displacement is x. In a small
motion Ax the force is nearly constant and the work approximately
§48] WORK 93

F{x) Ax. As Ax diminishes, this approximation becomes more

accurate. The exact work done by the force F{x) is thus

W = Axlim— XFix) Ax = f F(x) dx (1)

0 d

when the body moves from x =

a to x = 6.
When the force is expressed in pounds and the displacement in
feet the work is given in foot-pounds.

Example 1. The amount a helical spring is stretched is proportional

to the force applied. a force of 100 pounds stretches the spring
If 1 inch,
find the work done in stretching it 4 inches.
Let X be the number of inches the spring is stretched. The force is then

F= kx,

k being constant. When x = 1, F= 100. Hence

100 = A;

and
F = lOOx.

The work done in stretching the spring 4 inches is thus

F dx lOOx dx = 800 in.-lb. = 66| ft. -lb.

Example 2. A hemispherical tank of radius 2 feet is full of water

weighing 62.5 pounds per cubic foot. How much work is required to
pump the water from the tank?
Divide the water into slices by horizontal planes (Figure 57). The
volume of waterbetween the planes at depths x, x + Ax below the center
is approximately a cylinder of radius

r = •\/4 — x^,

altitude Ax, volume

Trr^ Ax = 7r(4 — x^) Ax
and weight
62.57r(4 — x^) Ax lb.
94 INTEGRATION AND SUMMATION [Chap. Ill

The work done in lifting this slice the distance x to the top is thus approx-
imately
a:[62.57r(4 — x^) Ax] ft.-lb.

The total work required to empty the tank is therefore

62.57rx(4 -- x^) dx = 2507r ft.-lb.

49. Properties of Definite Integrals. If f{x) is continuous for

the values of x involved, the following equations are almost
immediate consequences of the definition of definite integral:

The first of these equations states that interchange of limits

changes the algebraic sign of a definite integral. To prove this
we notice that if the same intermediate values are used in both
cases the increments Ax when x varies from b to a merely differ
in algebraic sign from those obtained when x varies from a to b.

Thus
o b

'^mAx= -2^/(f)Ax,
b a

from which (1) follows by taking limits.

If c is between a and 6, equation (2) expresses graphically that the
area between the ordinates at a and 6 (Figure 58) is equal to the

Figure 68. Figure 69.

area between a and c plus that between c and 6. It should be

noted, however, that this equation is also valid when c is not be-
tween a and 6. In Figure 59, for example.
§51] INTEGRAL WITH VARIABLE LIMIT 95

dx

is negative and so

dx

is equal to the area from a to c minus that from b to c, the result

being

60. Mean Value Theorem. If f{x) is continuous in the interval

(a, 6), there is some value f between a and b such that

f(x) dx^{b^ a)m- ( 1)

To prove this we consider b > a. '^I'hc case b < a can be handled

in a similar way. Let m be the number such that

Then
/ f{x) dx = (b — a)m. (2)

/ [f{x) — m] dx = 0. (3 )

This shows that f{x) — m

cannot be positive for all values of x
between a and 6, for the integral would then be positive. Similarly
f{x) — m
cannot be negative for all these values. This function
must then either be zero for all such values or be negative for some
and positive for others. In the latter case there will be at least
one value between a place where f{x) — m is positive and a place
where f(x) — m is negative at which this function is zero (inter-
mediate value theorem, §14). In any case there is then at least
one value x = J between a and b such that

fiO -m= 0,

and (1) follows from (2).

61. Integral with Variable Limit. If f(x) is continuous for the
values considered and a is constant, the integral

Pix) = Cm
%/a
dt ( 1)
96 INTEGRATION AND SUMMATION [Chap. Ill

is a function of the upper limit x. Graphically it represents the

area (Figure 60) bounded by the a:-axis, the curve y = /(x), a fixed

Figure 60 .

ordinate at a: = a, and the variable ordinate of abscissa x. We

shall show that this function has the derivative

-
U/X
/
a
mdt=f(x),

To prove this we first determine the increment

ac+Ax
AF(x)
/ f(t) dt-
a
m
,
dt m dt.

If /(<) is continuous at < = a:, by the mean value theorem (§50)

/ f(t) dt = AxfiO,

where { is between x and x + Ax. Thus

AF(x)
Ax
When Ax tends to zero, f tends to x, and this equation gives as

JF(.)
ax
which was to be proved.
PROBLEMS
Find the values of the following integrals:

1
.
y’( 3 r* - 4* +2 ) 2. y'(8®« + 6®* - 2x + 6) dx.
 PROBLEMS 97

6. 6.

/• *» - 2a;* - 3
7.
,
8. y*(x — l)(a; — 2) cte.
J ^
du
9.
J x{x + 1)2 dx. 10.
/ + (« 1)"

f dx
11.
.
y (3a: + 2)*
12.
J V +S<U. 2t

f dx r ydy
13. 14.
J V3z + 2 J \/y2 _ 1

f xdx
j * V Z3? + 2
"
16. 16. dx.
J (3a:* + 2)*‘
f 3? dx
17. 18.
J (a* +
19.
r + 1) dx
(2x
20.
J Vx^ + X + 1
21. cos20d0. 22.
J*
23. r sin t cos t dt. 24.

26 A body, started vertically downward with a velocity of 30 feet per

second, falls under the acceleration of gravity. How far will it fall in t seconds?
26 From a height ho above the ground a body is started upward with a
.

speed vq. If it is subject only to the acceleration g of gravity, find its height
at the end of t seconds.
27 If the velocity of a body at time t is
.

v = (2i + ft. /sec.,

find the distance traversed between t — 2 and t = 5.

28 . A body, starting at the origin and moving along the x-axis, has the

—=
dx
TT sin
.
— ,
ft. /sec.
dt 2

at the end of t seconds. Find its position when t = 2.

29 . A particle, starting from rest, has the acceleration

(4i2 -2t + S) ft./sec.2

at the end of t seconds. Find its average velocity during the first 3 seconds.
30. The velocity of a body is

v = TT cos^t ft. /sec.

4
at the end of t seconds. Find its average velocity during the first 2 seconds.
31 . Assuming that the brakes of an automobile produce a constant negative
acceleration —k feet per second per second, find what k must be if when trav-
98 INTEGRATION AND SUMMATION [Chap. Ill

eling.60 miles per hour the automobile is brought to rest in a distance of

100 feet.
32 At each point (x, y) of a curve the slope is
.

dx
Ifthe curve passes through (2, 3), find its equation.
38 At each point (x, y) of a curve the slope is sin x. If the curve passes
through (0, 1), find its equation.
34 At each point (x, y) of a curve the slope is kx. Find the value of A; if
.

the curve passes through (0, 0) and ( — 1,2).

36 On a certain curve
.

d^y
.
- 1.

If the
. curve passes through (
— 1, 1) and has at that point the slope 2, find its
equation.
36 At each point
. (x, y) of a curve

cPy
= 6x.
dx^

If the curve passes through (1,2) and (0, — 1), find its equation.
37 .By dividing the interval into 6 parts determine approximately the area
bounded by the x-axis, the curve y = x^, and ordinates at x = 0, x = 1. The
exact area is f
88. By dividing the interval into 5 parts determine approximately the area
bounded by the x-axis, the curve
1

and ordinates at x = 0, x = 1. The exact area is jt.

39 By
dividing the area into 6 parts determine approximately the area
bounded by the x-axis, the curve

y = Vx(12 — x).

and ordinates at x = 0, x = 12. The exact area is IStt.

40 By dividing the area into 6 parts determine approximately the area
TT
bounded by the x-axis, the curve y = 2 sin x, and ordinates at x = 0, x = - .

The exact area is 2.

Find the values of the following definite integrals:

41 . — 2x + 2) dx. 42 . — x^) dx.

T
43 . + Idx. 44 .

X sin X dx.

46 . cos X dx. xVa* — x^ dx.

 PROBLEMS 99

47. Find the area bounded by the a:-axis, the line y = 2a; — 1, and the
ordinates at a; «= 1, a; = 2.
48. Find the area bounded by the coordinate axes, the curve

y = 3a;2 ^ 4^. 4.
and the line x = 2.
49. Find the area bounded by the x-axis, the curve y == sin x, and the ordi-

60. Find the area bounded by the x-axis and the curve

y — 4: — x^.

61. Find the area bounded by the x-axis and the curve

1/ = 4x — x^ — 3.

62. Find the area bounded by the y-axis, the curve

66. X = 3i/2 4. 2,

and the y = — 1, t/ = 2.
lines
63. Find the area bounded by the 2/-axis and the curve

X =2y —
64. Find the area bounded by
y^ = X
60.the lines x = 1, / = — 1.
and 2

Find the area bounded by the curve

2/
= 2x - x2
and the lino y = —3.
66. Find the area bounded by 1/ = 4 — x^ and y = S — 2x^.
67. Find the area bounded by 2/ — x^ x^j y == x® + 1.
68. Find the area bounded by y = x — x^ and y — —x.
69. Find the area bounded by the straight line x + 2?/ =3 and the curve
X = 2/® 32/ + 1
The triangle bounded by the coordinate axes and the line

y = 2 -X
is rotated about the x-axis. Find the volume generated.
61. The area bounded by the x-axis and the curve

2/
= 4 — X®

is rotated about the x-axis. Find the volume generated.

62. the line x = a is rotated
The area bounded by the parabola y® = 4ax and
about the Find the volume generated.
x-axis.
63. The area above the x-axis, bounded by the curve y^ = x and the line
X = 1, is rotated about the x-axis. Find the volume generated.
64. A sphere of radius a is generated by rotating the circle

y^ = a®

about the x-axis. Find its volume by integration.

100
66. INTEGRATION AND SUMMATION [Chap. Ill

The
area within the circle is cut into two parts by the line

X = h. Find the volumes of the spherical segments generated by rotating

these parts about the x-axis.
66. Find the volume of the ellipsoid generated by rotating the ellipse

about the x-axis.

67. The area bounded by the i/-axis and the curve
X = 2y —
isrotated
.
about the 2/-axis. Find the volume generated.
68. The area bounded by the curve y — 2 — x^ and the line y = 1 is rotated
about the x-axis. Find the volume generated.
69 The area bounded by the curve y^ = 4ox and the line x = o is rotated
.

about the line x = a. Find the volume generated.

70 The area bounded by the parabola y^ = 4ax and the line x = a is
.

rotated about the y-axis. Find the volume generated.

71 The area bounded by the parabola y^ = 4ax and the line x == a is
.

rotated about the line x = 2a. Find the volume generated.

72 The area bounded by the parabola y^ = 4ax and the line x = a is
.

rotated about the line y = —2a. Find the volume generated.

73 The area bounded by the hyperbola xy = 4 and the line x + 2/ = 5
.

is rotated about the y-axis. Find the volume generated.

74 Sections of a solid parallel to a fixed plane are circles, the section at
.

distance x from the plane having the radius r == x^. Find the volume of the
solid between the sections at x = a and x = &.
76 The base of a solid is a circle of radius r and diameter AB^ and each
cross section perpendicular to A B is a square. Find the volume of the solid.
76 The base of a solid is the area bounded by the parabola y^ = 4ax and
.

the line x = a, and each cross section perpendicular to the x-axis is an equi-
lateral triangle. Find the volume of the solid.
77 The base of a solid is the area bounded by the x-axis and the lines
.

y = 2x, X = 2, and each cross section perpendicular to the y-axis is a triangle

of altitude 3. Find the volume of the solid.
78 A square of side a revolves about a line perpendicular to its plane while
.

the point of intersection of line and plane moves the distance h along the line.
Find the volume of the twisted solid which is generated.
79 Two circles have a common diameter and lie in perpendicular planes.
.

A square moves in such a way that its plane is perpendicular to this diameter
and its diagonals are chords of the circles. Find the volume generated.
80 The base of a solid is a circle of radius a and diameter AB. Each sec-
.

tion perpendicular to AB is a triangle of altitude h. Find the volume of the

solid. To evaluate the integral that is obtained observe that it is a constant
times the area of the circle.
81 . A wedge is cut from the base of a right circular cylinder by a plane
passing through a diameter of the base and inclined at the angle a to the base.
Find the volume of the wedge.
82 . The generators of an oblique cylinder make an angle of 46® with the
base. A wedge is cut from the base of the cylinder by a plane through the
 PROBLEMS 101

center of the base and perpendicular to the generators. Find the volume of
the wedge.
86. Find the force due to water pressure on a rectangular flood gate 10 feet
83.
broad and 8 feet deep, the upper edge being in the surface of the water.
84. Find the force on the lower half of the gate in the preceding problem.
Find the force due to water pressure on one side of a triangle of base 6
and altitude h submerged so that its vertex is in the surface of the water and
its altitude vertical.
A vertical dam is in the form of a trapezoid 300 feet long at the surface,
86.
200 feet at the bottom, and 30 feet high. Find the horizontal force on the
dam when the water is 20 feet deep.
87. The end of a pool is a rectangle inclined 45° to the vertical. The edge
in the surface of the water is 12 feet long, and the submerged edge, 10 feet.
Find the force due to fluid pressure on the rectangle.
88. According to Hooke’s law the force required to stretch an elastic rod of
natural length o to the length a -f a; is
kx
>
a
where k is constant. Find the work done in stretching the rod from the length a
to the length 6.
89. A
cylindrical cistern, of diameter 4 feet and depth 8 feet, is full of
water. Find the work required to pump the water out.
90. A tank, 20 feet in diameter and 16 feet deep, is half full of water. Find
the work required to pump the water out.
91. A pyramid, of altitude h feet and base A square feet, is built of rock
of weight w pounds per cubic foot. Find the work required to lift the stones
from the ground into place.
92. Assuming that the force of gravity on a mass m at distance r from the
center of the earth is
km

where k is constant, find the work done by gravity on a 10-pound body when
it moves from an Take
indefinitely great distance to the surface of the earth.
the radius of the earth as 4000 miles and use the known force of gravity at
the surface of the earth to calculate k,
93. Inside the earth a particle is attracted toward the center with a force
proportional to the distance from the center. Find the work required to lift
a 10-pound body from the center to the surface of the earth.
94. A weight of 500 pounds hangs at the end of a 200-foot chain which
weighs 2 pounds per foot. Find the work done in winding in the chain.
96. A gas is compressed in a cylinder by a movable piston. Assuming that
it is compressed adiabatically (that is, without receiving or giving out heat),

its pressure and volume satisfy the equation

py'y = constant,

where y for air is about 1.4. Find the work done in the adiabatic compres-
sion of 100 cubic inches of air from atmospheric pressure (14.7 pounds per
square inch) to a volume of 10 cubic inches.
 CHAPTER IV

ALGEBRAIC EQUATIONS AND GRAPHS

62. Equation and Locus. To each equation

f(x, y) = 0

corresponds a locus consisting of all points with rectangular coordi-

nates X, y which satisfy the equation. The locus is called the graph
of the equation, and the equation is said to represent the locus.
Since rectangular coordinates were first used by Descartes, the
rectangular equation of a locus is often called its Cartesian equation.
To
each analytical property of such an equation or set of equa-
tions corresponds a geometrical property of the locus or loci rep-
resented. The study of this correspondence between analytical
and geometrical properties and, in particular, the deduction of
geometrical relations by means of the corresponding analytical
characteristics constitutes a subject called analytical geometry.
In this chapter we shall consider certain algebraic equations in
X and y and the plane loci they represent. To describe briefly
what we mean by an algebraic equation, consider first a single
term

where a is constant and m, n positive whole numbers. This term

is said to be of degree m -t- n in x and y. The sum of a finite
number of such terms, whether of the same or different degrees,
is called a polynomial. The degree of the polynomial is defined
as that of its terms of highest degree. The equation obtained by
equating to zero an nth degree polynomial is called an algebraic
equation of the nth degree.
Thus
x^ — xj/-l-3y -1-7 =0
is an equation of the second degree and

^ — 2x = 0
is one of the third degree.
102
§64] EQUATION OF A STRAIGHT LINE 103

63. Point of Division. Along a line not parallel to a coordinate

axis no positive direction is usually assigned. In handling dis-
tances along such a line it is convenient, however, to consider a
segment P\P2 as described from Pi to P2 and to consider two
segments as having the same algebraic sign when described in the
same direction, different signs when described in opposite direc-
tions.
To find the point dividing a line segment in a given ratio we
project on the coordinate axes and use the fact (which follows from
similar triangles) that the projections have the same ratio.

Example L Find the point P{Xyy) on the line through Pi(l, — 2 ),

P 2 O, 4) way from Pi to P2
two-thirds of the .

Let Ml, My M
2 be the projections of Pi, P, P2 on the x-axis (Figure 61).

Then Mis two-thirds of the way from Mi to 2 The coordinate x of M . M

is the displacement in moving from the origin to M. This may be obtained
as the sum of the displacements in going from the origin to Mi and then
two-thirds of the way from Mi to M 2. That is,

X = 1 + 1(7 ~ 1) = 5.

Similarly,
2/
= -2 + !(4 + 2) = 2.

Example 2, Find the point on the line through Pi (2, 3), P2 ( 8 , 5) such
that
PiP = -IP P2 1 .

By this we mean that the points are related as in Figure 62. The x-
coordinate of P may be obtained as the displacement from the origin to
Ml followed by a displacement — IM M 2 Thus1 .

X = 2 - i(8 - 2) = -1.
Similarly,
2/
= 3 - K5 - 3) = 2.

64. Equation of a Straight Line. Let Pi{xiy yi) be a fixed point

and P(x, y) a variable point on a given line LM.
104 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

If the line is not perpendicular to the a:-axis its slope is

y - yi
m= ,

X Xi
whence
y - yi = rn{x - Xi). ( 1)

This equation is satisfied by the coordinates of any point {x, y)

on the line and by those of no other point. It is therefore the equa--
tion of the line through (xi, yi) with slope equal to m.

In particular, if we take (xi, i/i) as the point (0, b) where the

line crossesthe i/-axis, (1) becomes

y = mx + h. (2)

The number h is called the intercept of the line on the 2/-axis. Thus
(2) is the equation of the line with
slope m and intercept 6.
If the line is perpendicular to
X Piie.V)
the a:-axis, its slope is infinite and
equations (1) and (2) cannot be
Pi (®i»l/i)
used. In that case a diagram
(Figure 64) shows, however, that

"o X xi (3)

Figure 64. is the equation of the line.

Example, Find the equation of the line through the points (1, 2) and
(3, -4).
The slope of the line is

m = 3-1
§66] FIRST-DEGREE EQUATION 106

Since the line passes through (1, 2) and has this slope, its equation is

2/
- 2 = - 1 ),

which is equivalent to
y + ^x = 5.

66. First-Degree Equation. Equations (1), (2), (3), of §54 are

of the first degree in x and y. Thus any straight line is represented

by an equation of the first degree in rectangular coordinates.

Conversely, any equation of the first degree in rectangular co-

ordinates represents a straight line. For any equation of the first

degree in x and y has the form

Ax + By +C = Q, (1)

A, fi, C being constants. If B is not zero, this is equivalent to

y =
B B
which represents the line [§54, (2)] with slope —^ - and intercept
C
— on the 2/-axis. If B is zero the equation is equivalent to
B
C

which represents a line perpendicular to the a:-axis [§54, (3)].

Because of this relation to a straight line a first-degree equation
in X and y is called linear. By analogy the range of this term has
been extended to include all first-degree equations.
A linear equation in x and y appears to contain three constants
Ay By C. Since, however, division by a constant does not change
the equation, it really contains only two significant constants, for
A C
— and ~
instance, . This means algebraically that the constants
B B
in the equation are essentially determined by two equations and
geometrically that the line may
be made to fulfill two conditions.
For example, a line can be made to pass through two points, to
pass through one point and be perpendicular to a second line, to
be tangent to two circles, etc.
If the coefficient B is zero, the line is perpendicular to the x-axis
and its slope is infinite. If B is not zero, the equation can be solved
for y and so reduced to the form,
y — mx + h.

The slope is then the coefficient of x.

106 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

66. Angles. An angle is regarded as generated or described by

rotation around its vertex from an initial to a terminal side, the
direction of description often being indicated by an arrow. The
angle isconsidered positive if its direction is that which carries
OX through 90° into OF. With the usual positions of the axes
this amounts to saying that an angle is positive when it is described
in the counterclockwise direction, negative when in the clockwise
direction. In Figure 65, for example, the angle A extends from

AB to AC and is positive, whereas B extends from BA to BC and

is negative.
The direction of a line is usually determined by means of the
angle <l>
extending from the positive end of the z-axis to the line.

This angle may be drawn in either direction and its terminal side
may be either end of the line. Any one of these angles uniquely
determines the direction of the line and any one of them satisfies
the equation
tan 0 = m, (1)

m being the slope of the line.

In the following diagrams taken as the smallest positive one

is

of these angles. But in some cases, as for instance when the line
is variable, this restriction may not prove convenient.

Let 01 , ^ be the angles which two lines make with the x-axis,
and
mi = tan 0i, m2 = tan 02 (2)
their slopes.
the lines are parallel (Figure 67) the angles 0 i, 02 are equal
If
and the slopes are equal. Conversely, if the slopes are equal the
lines are parallel. Thus, two lines are parallel if and only if their
slopes are equal.
 ANGLES 107

If the two lines are perpendicular, 02 and 0i differ by 90® and

1
tan 02 = cot 01 =
tan 01
Thus
1
m2 (3)
mi
Conversely, if this equation is satisfied, tan 02 = — cot 0i and the

angles differ by 90®. Thus, two lines are 'perpendicular if and only if
the slope of one is the negative reciprocal of the slope of the other.
Let
^ = 02 — 01 (4)

be the angle from the line of slope mi to that of slope m2. Then
tan 02 — tan 0i
tan p = tan (02 — 0i)
1 + tan 01 tan 02
-

Figure 69.
108 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

Since tan = mi, tan 02 = this is equivalent to

m2 — mi
tan /3
= (5)
1 + mim2
The angle P in this equation may be any angle extending from
m
the line of slope mi to that of slope 2 To find the tangent of a .

particular angle take mi as the slope of its initial side and 2 as m

that of its terminal side.
Example L Show that the lines
3a; — 4^/ + 1 = 0, 4x + 3?/ — 5 = 0,
are perpendicular.
Solving for y, the equations be-
come

y = \x + \, y= -^x +
The two slopes are

mi = f, m2 =
Since

m2 = —
mi
the lines are perpendicular.
Example 2, The vertices of a triangle are A(l, —1), R(3, 1), C(— 1, 2).

Find the interior angle A,

If the angle A is considered positive, Figure 70 shows that it extends
from AB to AC, The slope of AB is 1 and that of AC is — f. Hence
3
tan A = -
1

A = tan~^ 5.

67. Tangent Line and Normal to a Curve. Let (xi, yi) be a

dy
fixed point on a curve, the derivative obtained from the equa-
ax
tion of the curve, and

the value this derivative takes when x = xi, y — yi.

The tangent to the curve at Pi{xt, y{), being a line through Pi
with slope mi, is represented by the equation

y-yi {x - Xi). (1)

§67] TANGENT LINE AND NORMAL TO A CURVE 109

The line PiN perpendicular to the tangent at its point of con-

tact Tvith the curve is called the normal to the curve at that point.

Since the slope of the perpendicular is the negative reciprocal of

the slope of the tangent (§56), the equation of the normal at
yi) is

y -yi= - 777“ (* - Xi). (2)

(-)
\dx/i

In equations (1) and (2) it should be noted that xi, yi are co-
ordinates of a point on the curve but that (x, y) is in general not
on the curve.

Example. Find the equations of the tangent and normal to the curve
xy = 2 at the point (1, 2).
From the equation xy = 2 we find by differentiation

x^ + y-o.
whence
^ - 1?.
dx X
At (1, 2) the slope of the tangent is

The equation of the tangent is therefore

y - 2 = -2{x - 1)
and that of the normal
V-2 = Ux- 1 ).
no ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

68. Newton’s Method. For a short distance the tangent to a

smooth curve forms ahnost a continuation of the curve. This is
the basis of a method due to Newton, for the approximate solution
of equations.
Let
fix) = 0

be the equation. A root of this equation is the abscissa of a point

at which the curve

y = fix)
meets the x-axis. By trial sup-
pose /(xo) is found to be small.
The tangent at xq is

y - fixo) = fixo)ix - xo). (1)

This crosses the a:-axis at

fjxo)
Xi = Xo (2 )
fixo)

If the curve and tangent are nearly coincident over the range
ixof Xi) the value Xi should be an approximate root of the equation.
If necessary the process may be repeated, giving

fjxi)
X2 = Xi ' (3 )
fixi)

etc. Whether and how fast the process will converge depends on
the function /(a;) and on the value
Conditions favorable
initial xq-
to convergence are evidently that/(a:o) be small and/'(xo) large.

Example. Find an approximate solution of the equation

x* + 4x — 6 = 0.

Writing
/(x) = x* + 4a; — 6
we find by trial
/(l.l) = -0.269.

By Newton’s method we obtain

/(LI) 0.269
x\ = 1.1 - 1.1 + 1.1363.
/'(LI) 7.63

The value correct to four decimals is 1.1347.

§69] ANGLE BETWEEN TWO CURVES 111

69. Angle between Two Curves. A point of intersection of two

curves has coordinates that satisfy both equations. To find the
intersections of two curves we thus solve their equations simul-
taneously.
By the angle between two curves at a point of intersection we
mean the angle between their tangents at that point. If nti and

m2 are the slopes at a point of intersection and fi is the angle

from the curve with slope mi to that with slope m 2 by equation (5), ,

§56,
—
tan /3 = m2 mi
1 + mim2
Two curves are said to be orthogonal they intersect at right
if

angles. The condition for this is that either slope at a point of

intersection be the negative reciprocal of the other.

Example L Find the angle between the line y = x and the curve
y ^ x^ dX each intersection.
Solving simultaneously, we find that the line and curve intersect at
(1, 1) and (0, 0). The slope of the line is

mi = 1.

The slope of the curve at any point x is

At (1, 1) the slope of the curve is 2 and the angle /3i from line to curve is

determined by
2 - 1 1
tan/3i
1+2 s’
112 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

At (0 0) the
,
slope of the curve is zero and

0 - 1
-1
tan 182 =
1+0 ,

whence
182
= - 45 °.
The negative sign in this last case signifies that the acute angle is meas-
ured clockwise from the line to the curve.

Example 2. Show that if a and h are any constants the curves

= a, xy h
intersect orthogonally.
From the first equation we obtain

'
dx y
and from the second
^
dx X

At a point of intersection x and y have the same values in both equations.

The second slope being the negative reciprocal of the first, the curves are
orthogonal at all intersections.

60. Distance between Two Points. To determine the distance

between two points Pi{xu 2/i), -Pzfe, 2/2 ) construct the triangle
(Figure 76) with sides Pifl, RP2 parallel respectively to OX and
OF. The lengths of these sides are

PlR = ±(X2 — Xl)y RP2 = ±(2/2 “ Vl)}

the algebraic signs depending on the positions of Pi and P2 . Since

the expression
P1P2 = VPiB* + BPa*
§61] EQUATION OF A CIRCLE 113

contains only the squares of J\R and IiP2 it follows, however, in

every case that

P1P2 = V(a:2 - xx)" + {y2 - yi)". ( 1)

Example. Find the point equidistant from the three points A (9, 0)
R(-6, 3), C(5, 6).
Let P{x^ y) be the point required. Since PA = PB and PA = PC, we
have
- 9 )» +
\/{x — 9)* + •>/ = -s/ix — 5)* + ( 2/
- 6)*.

Squaring and canceling, we obtain

5x - 2/
= 6,

2x — = 5.

By solving these last equations simultaneously, we find x = 1, !/

= —L
The required point is therefore (1, —1).

61. Equation of a Circle. A circle is the locus of points at con-

stant distance, equal to the radius, from a fixed point called the
center.
Let CQi, k) be the center of a circle (Figure 77) and r its radius.
If P(x, y) is any point on the circle.

r = CP = V(x - hf + (y - kf.
Thus
(X - A)2 +(y-k)^ = ,^
(1)

is the equation of the circle.

114 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

When expanded (1) becomes

a;* + j/® — 2hx — 2ky + A* + A* — = 0.

This is an equation of the form

+ + Bx + Cy + D = 0, (2)

in which A, B, C, D are constants.

Every by an equation of the form (2).

circle is represented
Conversely, if 4 is not zero any equation of this kind represents
a circle if it represents any curve. To show this divide by A and
complete the squares of terms containing x and those containing
y separately. The result is

B^ + C^- iAD
(3)

If the right side of this equation is positive, comparison with (1)

~
shows that it represents a circle with center
^
radius
/52 + c* - AAD
44 *
If the right side of (3) is zero, the radius is zero and the circle
shrinks to a point. If the right side is negative, since the sum of
the squares of real numbers is positive, there is no real locus.
The equation of a circle contains three constants, for example,
h, k, r. These constants can in general be determined to satisfy
862] PARABOLA 115

three equations and so the circle can be made to fulfill three con-
ditions. Thus a circle can be passed through three points not on
a line, can be made to touch three lines not parallel or through a
point, etc.

Example 1. Find the center and radius of the circle

2x^ + 22/2 - ar + 42/ = 3.

Dividing by 2 and completing the squares, we have

{X - f)2 + (2/ + 1)2 =

Comparison with (1) shows that the center of the circle is (f, —1) and its
radius is -j.

Example 2. Find the equation of the circle which passes through the
origin and through the two points in which the circle a:2 + 2/2 = 3 inter-
sects the straight line x +y — 1.

To do this we use the equation

x2 + 2/" - 3 + + - 2/ 1) = 0, (4)

k being constant. If a;2 + 2/2 —

3 and x y — 1 are both zero this equa-
+
tion is satisfied. Thus it represents a locus which passes through the
intersections of the circle and line. Since (4) is of the form (2) this locus
is a circle. To make the circle pass through the origin we place x = 0,

2/
= 0, thus obtaining
-3 + A;(-l) = 0

and k = —3. When k has this value equation (4) satisfies all the condi-
tions of the problem. Thus

^2 -I- 2/2
- 3 - 3(a; + - 2/ 1) = 0,
or its equivalent
a;2 + 2/^ - ac - 32/ = 0,

is the equation required.

62. Parabola. A parabola is the locus of points equally distant

from a fixed point and a fixed straight line. The fixed point is called
the focuSy and the fixed line the directrix.
Take the a;-axis through the focus and the origin midway be-
tween the focus and directrix. In Figure 78, F is the focus, RS
the directrix, and
DO ^ OF ^ p, (1)
116 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

The focus is thus the point F{p, 0). If P(x, y) is any point on the
parabola and NP is perpendicular to RS, by definition

FP = NP, ( 2)
that is

y/Jx + 2/^ = ^ + P- (3)

This is the equation of the parab-

ola. By squaring and canceling
it becomes
(4)

Lines, x = constant, perpendicu-

lar to the a:-axis cut the curve in
pairs of points

y = zbV^4px

Figure 78. at equal distances above and below

the x-axis. We express this by
an axis of symmetry of the curve. Since it is the
calling the x-axis
only axis of symmetry, it is called the axis of the parabola. The
point 0 where the axis intersects the curve is called the vertex of
the parabola.
In equation (4), 2/ is the perpendicular from a point on the
curve to the axis, and x is the distance along the axis from this
perpendicular to the vertex. If
the vertex is not the origin but
some point AQi, fc). Figure 79,

these distances are

MP =: y -- k, AM = X — A

and so the equation of the

parabola is

(y
— = 4p(x — h). (5) Figure 79.

If the axis of the parabola is parallel to the 2/-axis, the variables

X and y are interchanged giving

(x - A)2 = 4p(2/ - k) (6)

as the equation of the curve.

§62] PARABOLA 117

Equations (6) and (6) may be written

A2/2+ Bx + Cj/ +D= 0, (7)

Ax^ + Cx + By D= 0, (8)

A, B^CjD being constants. Conversely, by completing the squares

it can easily be shown that any equation of the form (7) or (8)
represents a parabola if A and B arc not zero.
Since division by a constant has no effect, equations (7) and
(8) contain three essential constants. A parabola with axis parallel
to a coordinate axis can thus be made to satisfy three conditions,
for example, to pass through three points not on a line.
When rotated about its axis a parabola generates a surface
called a paraboloid of revolution. A mirror of this form has the
property that rays of light parallel to the axis, striking its inner
surface, are reflected to the focus, and rays issuing from the focus
are reflected parallel to the axis. Such a mirror is consequently
used in a reflecting telescope to focus light from a distant object,
and in an automobile headlight to reflect light from a lamp near
the focus in a substantially parallel beam.
When carrying a load proportional to the horizontal distance,
the cable of a suspension bridge hangs in a parabola, and a para-
bolic arch experiences only tangential compression.
If air friction is neglected a projectile describes a parabola with
vertical axis.

Example 1. Find the vertex, focus, and directrix of the parabola

2/2 - 2y - 2x = 5.

Completing the square of terms in this becomes

{y
- 1)2 = 2{x + 3).
Comparison with (5) shows that this is a parabola with vertex (—3, 1), and
that 4p = 2. The axis being parallel to OX, the focus is at distance p — \
to the right of the vertex and the
directrix at the same distance on A
the left. Thus the focus is ( — |, 1)
and the directrix is a; = —
Example 2, An arch has the
form of a parabola with vertical
axis. If the arch is 10 feet high at
the center and 24 feet wide at the
base, find the width QP (Figure 80)
at a height 6 feet above the base. Figure 80.
118 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

If P is any point on the parabola, by equation (4)

Thus
HP = 4p AM.
= 4MP = k-AM
where h = 16p is constant. At the base QP = 24, AM = 10, and this
becomes

whence by division
^= ifc*10,

AM
24* 10
At height 6 feet, AM = 4 and
PQ = 24\/S = 15.18 ft.

Example S. Parabolas with the same axis and focus are called confocal.
The parabolas with focus at the origin and axis OX are obtained by assign-
ing values to p in the equation

2/2 = 4px + 4p2. (9)

Show that two of these curves pass

through any point, not on the axis,

and that these intersect at right

angles. We express this by saying
the system of confocal parabolas is

self-orthogonal.
The two values obtained by solv-
ing (9) for p determine two parab-
olas through the point (x, y). Dif-
ferentiation gives

( 10 )
dx y
for the slope of either curve at (a:, y).
Eliminating p from (9) and (10) we
obtain

Figure 81. ..+^- 1.0

as an equation the two slopes must satisfy. If mi and m2 are the two roots,
this equation can be written

(m — mi)(m — m2 = ) 0.

Comparison with the preceding equation indicates that

mim2 = — 1,
and this shows that the two curves are orthogonal [§56, (3)].
563] ELLIPSE 119

63. Ellipse. An ellipse is the locus of points the sum of whose

distances from two fixed points is constant. The two fixed points
are called foci of the ellipse.
Let F', F be the foci, take the a;-axis through these points, the

origin midway between them, and let 2c be the distance F'F.

The foci are then
F'(-c, 0), F(c, 0).

If P(xy y) is any point on the ellipse and 2a is the sum of its

distances from the foci, we have

F'P + FF = 2a, (1)

that is

V(x + cf + - cf + j/2 = 2a, (2)

as the equation of the curve.

Transposing one of the square roots, squaring both sides, and
solving for the remaining square root, we find

FP = y/ix - c)2 + 1/2 = a - -a;,

a
(3)

F'P = V(x + cf + = a + -x.

a

By clearing radicals either of these equations gives

2 + a2-c2
'
„2
= 1 . (4)

In the triangle F'FP the sum of two sides, F'P FP = 2a, + is

greater than the third side, F'F = 2c. Thus a is greater than c,
120 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

and is a positive quantity which we denote by 6^. Sub-

stituting
— b^y (6 )

the equation of the ellipse takes the final form

-+-= 1 .

A line x = constant or 7/ = constant perpendicular to either

coordinate axis cuts the curve in pairs of points

y dtz-Va^ - x^y
a

on opposite sides and equally distant from the axis. The curve is
thus symmetrical with respect to OX and 07, which are therefore
called axes of the curve. Their intersection 0 is called the center
of the ellipse.
When y — 0 these equations give x = ±a, and when a: = 0,
2/
= ±6. The numbers a, b are therefore the distances OAy OB
along the axes from the center to the curve. The lines A'Ay
B'B are called the major and minor axis,and a and b are called
semiaxes. The intersections A', A of the curve and major axis
are called vertices.
From the definition of the curve, or from equation
(5), it is seen
that the distance from the end of the minor axis to either focus is
equal to the major semiaxis. When a graph of the ellipse is given
the foci can therefore be located by describing a circle of radius a
(assuming that A 'A is the major axis) and center B and determin-
ing its intersections with the major axis.
The ratio

is called the eccentricity of the ellipse.

It is always less than unity.
a = b and the ellipse becomes a circle. As e increases,
If it is zero,
the ratio of major to minor axis increases and the ellipse becomes
more and more eccentric.
§63] ELLIPSE 121

Equations (3) can be written

FP = a — ex = = e-NPj

F'P = a + ex = e(^ + x^=‘ e-N'P.

The lines
a
x = ±- ( 8)
e

are called directrices of the ellipse. In Figure 83 they are the lines
RSj R'S\ The above equations show that

FP _ FT
(9 )

Thus, an ellipse is the locus of a point the ratio of whose distance from
a fixed point (a focus) to its distance from a fixed line (the correspond-
ing directrix) is equal to a constant less than 1 (the eccentricity).
In the above discussions it is assumed that the major axis is
A'A, If the major axis is B'B, appropriate changes must be made
in expressions for foci, eccentricity, and directrices.

N
B

In equation (6) x and y are the displacements from the axes of

the curve to the point Pix, y). If the center of the curve, instead
of being the origin, is the point (A, A;), these displacements are
x ^ h, y — k, and the equation of the ellipse becomes
(x-hf (y-k)^
O
,

•» 9
= 1. ( 10 )

If a deformed in such a way that the distances of its

circle is
points from a fixed diameter are all changed in the same ratio,

the resulting curve is an ellipse.

122 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

When rotated about its major axis an ellipse generates an ellip-

soid of revolution. If the inner surface of such an ellipsoid is
silvered, light issuing from one focus is reflected to the other focus.
If the exterior surface is silvered, light converging toward one
focus is reflected as if it came from the other focus.
Under the law of gravitation the orbit of a planet moving around
the sun is an ellipse with the sun at one of its foci.

Example 1. Show that

25x2 + 92/2 + 50x - ISy 191

represents an ellipse, and find its center, axes, foci, eccentricity, and
directrices.
The equation can be written

25(x2 + 2x)+ 9(2/2 - 2y) = 191.

Completing the squares in the paren-

theses, this becomes

25(x + 1)2 +9 - ( 2/
1)2 = 225,

or
(x + 1)2 (y
- 1)2
= 1.
9 25

Comparison with (10) shows that this

represents an ellipse with center (—1, 1),
axes X = — 1, 2/
= 1, and semiaxes a = 3,
6 = 5. Since 6 is larger than a the
foci are on the vertical axis at distances

c = \/62 — o2 = 4

above and below the center. Thus they are /^'(— 1, —3), F(--l, 5).

The eccentricity is

-v/62 — o2 4
^

'
6 5

and the directrices are horizontal at distances

6
^ ^
e 4

above and below the center. Consequently, their equations are

§64] HYPERBOLA 123

Example 2. Show that if Pi(a:i, y{) is on the ellipse

- + ^=1
a2 ^ 62

the equation of the tangent at Pi is

xxi yyi
62
“ ^ ( 11 )
a2

If Pi is not on the ellipse determine what line the last equation represents.
From the equation of the curve we obtain
dy _ b^x
dx ahj

If P\{xiy yi) is on the ellipse, the tangent at Pi is

62^1 . .

Simplifying and using the equation

2 2
5i
+
^ ?^ = l
a* 6*

this can be reduced to the form (11).

If Pi is not on the ellipse, let P'(x', t/O and P"(x", y") be the points in

passes through Pi. In a similar way we show that the tangent at P" also
passes through Pi. (11) represents the line P'P" through the points
Thus
of contact of tangents which pass through Pi (Figure 85).

64. H]rperbola. The locus of a point, the difference of whose dis-

tances from two fixed points is constant, is called a hyperbola. The
two fixed points are called foci.
Let P', F be the foci; take the x-axis through these points and
the origin midway between them; and let 2c be the distance P'P.
The foci are then
F\-c, 0), F{c, 0).
124 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

If P{x, y) is any point on the hyperbola and 2a the difference

of its distances from the foci,

F'P -FP= ±2a, (1)

that is,

V(a: + c)^ + — V(® — c)* + y^ = ±2a (2)

is the equation of the curve.

Transposing one of the square roots, squaring, and solving for

the other square root, we find

FP = Viz - c)^ + y" =

(3)

F'P = \/(x + c)^ + y^ =

Clearing radicals either of these equations gives

(4)

as in the case of the ellipse.

In the triangle F'FP, the difference of two sides 2o is less than

the third side F'F = 2c. Thus o is less than c and is a

negative quantity which we denote by —b^. Substituting

c® — a* = 6^, (6)

the equation of the hyperbola takes the final form

(6)
§64] HYPERBOLA 125

Since the equation contains only even powers of x and y the

hyperbola is symmetrical with respect to both coordinate axes.

They are therefore called the axes of the curve, and their inter-
section is called the center. The distances a and h are called
semiaxes.
When 2/
= 0, X = zta; but when a: = 0, 2/ is imaginary. Thus
the axis through the foci, called the transverse axis, intersects the
curve in two points A', A, called vertices, at distance a on each side
of the center, but the other axis, called the conjugate axis, does not
intersect the curve. The hyperbola thus consists of two separate
parts, namely two symmetrical branches, on opposite sides of the
conjugate axis.
The ratio
OF c Va^ + 6“
e = =-= (7)
OA a a

is called the eccentricity of the hyperbola. It is always greater

than unity. As the eccentricity decreases to unity the two branches
of the curve draw closer to the transverse axis. As the eccentricity
increases, the two branches tend toward lines perpendicular to
the transverse axis.

Figure 87.

Solved for y the equation of the hyperbola becomes

2/
= ± - y/x^ — c?,
a

Assuming P in the first quadrant, the vertical distance between P

and the line 2/
= - is
a

ab
PP' = -a; - 2/
= - [a:
- - a^] =
[x + y/x^ — a^]
126 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

As X increases, this tends to zero. Similarly, if P{x, y) is a point

of the ciuve in any quadrant, the distance between P and one of

the lines
h
y = ±-x ( 8)
a

tends to zero as |
x |
tends to infinity. These lines are called
asymptotes of the curve.
If o and b are equal the equation of the hyperbola becomes
2 2 = ( 9)
The asymptotes
b
y = ± -X = zfcx
a

are then perpendicular to each other. Such a hyperbola is called

equilateral or rectangular.
If the center of the hyperbola, instead of being the origin, is

the point (A, A), and the transverse axis is parallel to OX, the
equation of the curve is

(x - A)2 {y - k?
( 10 )

If the transverse axis is parallel to OF this is replaced by

iv-kf {x-h?
^ ^
62 o2

and appropriate changes must be made in the expressions for

eccentricity, vertices, etc.

Example, Show that

2^2 _ 2^2 — 2y — &r — 9 = 0

represents a hyperbola, and find its center,

eccentricity, and foci.

Completing the squares this equation be-

comes

(2/2 - 22/ + 1) - 2(a;2 + 4x + 4) = 2,

which is equivalent to

(y - 1)^ {x + 2)g 1.
Figure 88. 2 1
§66] CONIC SECTIONS 127

This represents a hyperbola with center (—2, 1), semiaxes

0=1, 6 = \/2,

and transverse axis parallel to 07. In the expression for eccentricity a

and h must be interchanged, giving

y/a^ + 62 Is
^
b \2’

The distance from the center to the foci is

c = + h^ = \/3.

The foci, being on a line parallel to the y-axis at

distance c above and below the center, are

Figttke 89. (_2, 1 dz V3).

66. Conic Sections. The circle, ellipse, parabola, and hyperbola

are often called conic sections. The reason for this is that all can
be obtained as sections cut from a right circular cone by a
plane.
If the cutting plane is parallel to the base of the cone (Figure
89) the section is obviously a circle.

If the plane is inclined to the base

of the cone but at an angle less than
that between the base and generators
(Figure 90) the section is an ellipse.
To show this we construct two
spheres touching the cone in circles
ABy A'B' and the plane in points
F, F', If P is any point on the
plane and cone and Q, Q' are the
points in which the generator through
P cuts the two circles, the lines PF,
PQ, PF', PQ' are tangent to the two
spheres. Since all tangents from a
given point to a given sphere are
equal in length,

PF = PQ, PF' = PQ'

and so

PF + PF' = PQ + PQ' = QQ'

128 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

is the distance along the generator between the two circles. Since
this distance has the same value for all points P on the section,
this proves that the section is an ellipse and that F, F' are
its foci.

If the plane has the same inclination to the base of the cone as
the generators (Figure 91), the section is a parabola. There is

then a single sphere tangent to the cone along a circle and tangent
to the plane in a point. This point is the focus of the parabola,
and the directrix is the line in which the plane of the circle cuts
the plane of the parabola.
Finally, if the plane is inclined to the base of the cone at an
angle greater than that between the base and generators, the
plane will also cut the cone formed by the extension of these
generators (Figure 92). The section consisting thus of two parts
is a hyperbola. By a discussion similar to that presented for the
ellipse it may be shown that the foci of the hyperbola are the
points where the plane touches the two spheres tangent to the
cone and plane.
§66] TRANSFORMATION OF COORDINATES 129

66. Transformation of Co5rdinates. It is sometimes desirable

to move the coordinate axes to a new position. The coordinates
of a given point are changed, and the equations connecting old
and new coordinates are
needed.There are two
simple cases a combination
ofwhich gives any motion. h p
Jr
These are translatioriy in
which the origin is moved y'

to a new position without

changing the direction of O' 1 'x‘
\k
the axes, and rotatioriy in 1

which the origin is left 0

fixed and the axes turned Figure 93.
like a rigid frame about it.

(1) Translation of the Axes, Let OX, be the axes in their OY

first position, O'X', O' 7' the axes after motion, and h, k the co-
ordinates of the new origin with respect to the old axes. If x, y
are the coordinates of a point P
with respect to the old axes, and
x'j y' the coordinates of the same point with respect to the new,
from Figure 93 it is seen at once that

X = x' + h, y ^ +k (1)

are the equations of transformation.

(2) Rotation of the Axes. Let OX, OF be the original axes and
OX', OF' the new axes into which they are changed by rotation
through the angle 0 about
the origin. Let x, y be the
coordinates of a point P
with respect to the original
axes, and x', y' the coordi-
nates of the same point with
respect to the new. In Fig-
ure 94, X and y are the
horizontal and vertical dis-
placements in moving from
0 to P, and these are the
sums of the corresponding
displacements in moving from 0 to Af' and from Af ' to P.

X = x' cos ^ — 2
/' sin 0,

2/
= x' sin <#> + 2/^ (2)
130 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

the algebraic sign of each displacement being determined by its

direction parallel to the corresponding axis.

Example 1, By translation of the axes reduce the equation

a;2 + 22/2 - 6x + 42/ = 7

to a form that has no terms of first degree in the new coordinates.

Substituting
a; = a;' + A, y ^ y' +k
the equation becomes

(a:')* + 2{yy + {2h - 6)a;' + (4A; + 4)2/' + + ^ -7 ^

The coefficients of a;' and 2/' will be zero if

2A~6 = 4A; +4 = 0.

Thus h — k —1 are the coordinates of the new origin with respect

to the old axes. Substituting these values, the equation of the curve
becomes
(a
;')2
+2 ( 2/')2
- 18 = 0 .

Example 2, Determine the new equation into which

xy = k

is changed when the axes are rotated 45®.

The equations for rotation through = 45® are

X = x' cos </) — sin 0 =. —

x^ ~~~
i/'

y = x' sin (f) + y' cos </> =

v/2
Substituting these values, the above equation becomes

(xy - (2/')2 = 2k.

Thus xy = k is the equation of a rectangular hyperbola with the coordi-

nate axes as asymptotes.

67. Invariants. A transformation from one set of rectangular

coordinates to another is called orthogonal. With respect to such
changes geometrical quantities are of two kinds, those dependent
on the position of the axes and those independent of the axes.
Quantities of the second kind are called invariant^ or, more accu-
rately, invariant under orthogonal linear transformations. For
example, the radius of a circle is invariant (does not depend on
§68] GENERAL EQUATION OF THE SECOND DEGREE 131

the position of the axes), but the coordinates of its center are not
invariant (change when the axes change position).
The equation of a curve is not invariant, but there are some
things connected with it which are invariant. One of the simplest
is the degree of an equation in rectangular coordinates. Wherever
the axes are placed this degree always the same. To see this
is

it is sufficient to note that the equations for any change of rectangu-

lar coordinates have the form

X = aix' + 022/' + 2/ = + hy' + hy

oi, a2 f 03, 6i, 62, 63 a substitution a term
being constants. By such
^niyn
jg replaced by a sum
no one of which is of higher
of terms
than the (m + n)th degree in x' and y\ Hence by change of
rectangular axes the degree of an equation cannot be increased.
Neither can it be diminished; for, then, changing back to the old
axes would increase it. Therefore the degree of an equation in
rectangular coordinates is invariant.
If an equationx and y can be factored, so can the new equa-
in
tion resulting through a change of axes, for each factor of the old
equation will be changed into a factor of the new. Since real
expressions transform into real, if either equation has real factors,
the other does also.
68. General Equation of the Second Degree. An equation of
the second degree in rectangular coordinates has the form

Ax^ + Bxy + Cy^ + Dx -\- Ey F= 0, (1)

Af By Cf Dj Ey F being constants. To determine the locus repre-

sented we simplify the equation as much as possible by change of
coordinate axes.
First, by rotating the axes through a proper angle we make the
coefficient of Then, by translating to parallel axes, as in
xy zero.
Example 1, §66, we reduce the coefficient of a first power to zero if
the second power of the same coordinate occurs, and we reduce
the constant term to zero if only a first power occurs. By proper
choice of the axes any equation of the second degree in rectangular
coordinates can thus be reduced to one of the following forms:

03^ + +c= 0, (2)

^ + cx = 0, (3)
a, hy c being constants.
132 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

If c is zero, equation (2) or (3) is a product of first-degree factors.

If c is not zero, the locus is immediately seen to be a circle, ellipse,
parabola, hyperbola, or entirely imaginary. The last case occurs
when a, 6, c in (2) have the same algebraic sign. We thus conclude
that.
If a second-degree equation in rectangular coordinates is not a
product of first-degree factors, it represents a circle, ellipse, parabola,
hyperbola, or an imaginary locus.
To
determine which of these loci a particular second-degree
equation represents we note that in a rotation of axes the second
degree part of equation (1), namely

Ax^ + Bxy + Cy^,

transforms into the second-degree part of the new equation. Also
the second-degree part

a^

in the standard equation of the hyperbola is the product of real

first-degree factors

(t - -L
\a b) \a^ b/'

whereas the second-degree part of the equation of a paiabola is a

complete square, and that of an ellipse

is not factorable into real first-degree factors. Thus we conclude

that if equation (1) represents a real curve that curve is a hyper-

bola, parabola, or ellipse according as

Ax^ + Bxy + Cy^

is the product of real and distinct first-degree factors,
is a complete

square, or not factorable into real first-degree factors.

is

69. Graphs of Algebraic Equations. The equation of a curve

being given, any number of points on the curve can be determined
by assigning values to either coordinate, calculating the correspond-
ing values of the other coordinate, and plotting the resulting
points. When enough points have been located a smooth curve
drawn through them may be taken as an approximation to the
§ 69 ] GRAPHS OF ALGEBRAIC EQUATIONS 133

required curve. We desire as quickly as possible to obtain a satis-

factory approximation. To some extent this is accomplished by
plotting points rather sparsely where the curve is nearly straight
and more closely where it bends rapidly. The following are features
it may be helpful to note:

(1) Points where the curve crosses the axes and, if possible,
the points between crossings at which it reaches a maximum
distance from the axis.
(2) Values of either coordinate for which the other coordinate
is real and values for which it is imaginary.
(3) Symmetry.
(4) Infinite values of the coordinates, asymptotes.

(5) Direction of the curve near a point.

(1) Intersections with the Axes, The points where a curve meets
the x-axis are found by letting ?/ = 0 in the equation and solving

for X. on the y-axis are found by letting x = 0

Similarly, points
and solving for y. The a:-coordinates of points on the x-axis and
the ^/-coordinates of those on the i/-axis are sometimes called
intercepts of the curve on the coordinate axes.

Example 1. y = a:® — 2x.

+
Factoring the right side results in

y = x(x — l)(a; + 2).

The curve crosses the x-axis at a; = —2, 0, 1. These values divide the
a;-axis and the curve into four parts. On the x = —2 all three
left of

factors of the product are negative and y is negative. Similarly y is found

134 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

to be negative between = 0 and

a: a; = 1 and positive in the other two
sections. By solving the equation

^
ax
= 3a:* + 2a:-2 = 0
and substituting in the value of y, maxima and minima of y are found at
(—1.21, 2.11) and (0.55, —0.63). Using these values and plotting a few
additional points the curve in Figure 95 is obtained.
Example 2, y — 2 — x\x — 3).
Here intersections with y — 2 should evidently be used. These occur
at a; = 0 and a;= 3. The ordinate has a maximum at (0, 2) and a mini-
mum at (2, —2). The curve is shown in Figure 96.

(2) Real and Imaginary Values. When certain values are sub-
stituted for one coordinate the other coordinate may be real and
when other values are substituted it may be imaginary. The plane
is thus divided into strips in which there is a part of the curve and
strips in which there is none. These being determined, the part
of the curve in each strip may be plotted separately.

Example, y^ = xix — l)(a: 2). +

The curve crosses the x-axis at its intersections with the lines x = — 2,
a; = 0, a; = 1. These lines divide the plane into four strips. On the left
of a: = —2, the product xix — 1) (a: +
2) is negative and y is imaginary.
Between a; = — 2 and a; = 0, the product is positive and y is real. Simi-
larly, between a; = 0 and a; = 1, 2/ is imaginary, and on the right of a; = 1,

y is again real. The curve thus consists of two pieces, one between
a^ =* —2 and a; = 0, the other on the right of a; = 1 (Figure 97).
§69] GRAPHS OF ALGEBRAIC EQUATIONS 136

(3)Symmetry. Two points P, P' are said to be symmetrical with

respect toa line if the segment PP' is bisected perpendicularly by
that line. In Figure 98, for example, P and P' are symmetrical
with respect to the a:-axis, P and P" with respect to the i/-axis.
Two points P and P'" are called symmetrical with reject to a
point 0 if the segment PP'" is bisected by 0.

A curve is called symmetrical with respect to an axis if it is gen-

erated by pairs of points symmetrical with respect to the axis.
A curve is called symmetrical with respect to a center if it is generated
by pairs of points symmetrical with respect to the center.
A curve f{Xf i/) = 0 is symmetrical with respect to the a;-axis if

fix, y) = fix, -y).

For then it is generated by pairs of points (x, y), (x, — y) symmetri-

cal with respect to OX. In particular, a curve is symmetrical with
respect to the x-axis if its equation contains only even powers of y.
Similarly j a curve is symmetrical with respect to the y-axis if its

equation contains only even powers of x, and symmetrical with respect

to the origin if all terms in its equation are of even degree or all are
of odd degree.

Example, = x^ ^ x^.

The graph of this equation is shown in Figure 98. Since the equation
contains only even powers of x and only even powers of the curve is

symmetrical with respect to both axes and with respect to the origin.

(4) Infinite Values. One coordinate may increase without limit

when the other approaches a certain value. In such a case the
graph can be traced only until it runs off the paper. It thus con-
sists of two or more parts not connected together. Sometimes
136 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

two adjacent parts go off in the same direction, as at A, J5 in Fig-

ure 99; sometimes they go in opposite directions, as at C, D; and
sometimes there is no immediately adjacent part.

Figure 99.

If a branch of a curve when indefinitely prolonged approaches

a straight line in such a way that the distance between the two
tends to zero as limit, the line is called an asymptote of the curve.

The graph is shown As x tends to 0 or 2, 2/ becomes

in Figure 99.
infinite. When x is a than zero or a little greater than zero, y
little less

is large and positive. As x increases through zero the curve thus goes up
one side of the y-axis and comes down the other. When x is a little less
than 2, y is negative, and when x is a little greater than 2, y is positive.
As X increases through 2 the curve thus goes down the left side of CD and
then reappears at the top. When x increases without limit, y tends to zero.
The two coordinate axes and the line CD are asymptotes of the curve.

(5) Direction Near a Point, To determine the shape of a curve

near a point it is often helpful to find the direction along which
the curve or a branch of the curve approaches that point.

(«) (*>) (c)

Figure 100.
 PROBLEMS 137

In Figure 100, for example, are shown three ways in which a

curve may approach a horizontal line. These are distinguished
by the fact that the slope is zero in (a), is finite and different from
zero in (6), and is infinite in (c).

Example 1. == x^. The curve meets the x-axis at the origin but does
not extend to the left of the ^/-axis. Since y = xi, the slope

^ J
dx 2
is zero when x = 0. The curve thus has a sharp point, or cusp, at the
origin as shown in Figure 101.

Example 2, y^ = x\x + 2).

The curve symmetrical with respect to the x-axis and passes through
is

the origin, the slope at that point being

dx
= ± \/2 .

At the origin the curve is tangent to the lines having these slopes.

PROBLEMS
1. Find the coordinates of the point P(x, y) midway between Pi(xi, yi)
andP2(x2, 2/2).

2. On the line through Pi(l, —-3), P2(4, 6) find the point one-third of the
way from Pi to P2 .
 .

138 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

8 On the line through Pi (—2, 3), P2(2, —1) find the coordinates of the
point P(x, y) such that
6. ppj = -^PiPj.
4 Given Pi (2, —3), P2(“l, 2), find the point on P1P2 which is twice as
.

far from Pi as from P2. Also find the point on P1P2 produced which is twice
as far from Pi as from P2.
On the line through Pi (3, —4), P2(6, 2) find two positions of the point
P(a;, y) such that the distance from Pi to P is three times the distance from

Pi to P2.
6. On the line through Pi(aJi, 2/1), P2(a?2, 2/2) find the point P{x, y) such that

PiP _ ri

. PP2 n
7. On the line through Pi(xi, 2/1), P2(x2, 2/2) find the point P{x, y) such that

PiP = rPiP2.

8 Given the three points A(~3,

. 3), B(3, 1), C(6, 0) on a line, find the
fourth pioint D(a;, y) such that DC = AD
BC, : —AB :

9. Find the equation of the straight line through (2, —3) with slope 4.
10 Find the equation of the straight
. line through (1, 2) parallel to the
x-axis.
11 . Find the equation of the straight line through (3, —4) parallel to the
2/-axis.

12 Find the equation of the straight line through

. (2, —1) and (3, 2).
13 . Find the equation of the straight line through (2, 3) and (2, —4).
14 . Show that

is the equation of a straight line with intercept o on the x-axis and h on the
y-axis.
16 If the slopes of AB and BC are equal, show that the points A, P, C
lie on a line. Show that A(4, 1), P(l, 2), C(— 5, 4) are on a line.

16 . Show that the locus represented by

(x - 32/ + 4)(2x + 42/ - 1) =0

consists of two straight lines.
17 . Show that the line through (1, —2) and ( — 1,4) is parallel to that
through (3, —6) and (0, 4).
18. Find the equation of the straight line through ( — 1, 2) parallel to the
line 3x 4* 42/ — 7 =0.
19 . Find the equation of the straight line through (2, 3) parallel to the line
through (—3, 1) and (1, —4).
20 . Show that the straight lines determined by the pairs of points (2, 3),
(3, 5), and (1, 7), (3, 6) are perpendicular to each other.
21 . Find the equation of the line through (4, —1) perpendicular to the line
3x -f y — 2 = 0.
22 . Find the equation of the line through (—2, 3) perpendicular to the line
through (1, —2) and (3, 4).
 .

PROBLEMS 139
.

28 Find the angle from the positive direction of the a;-axis to the line
y -fx\/Z —2=0.
24 The sides of a
. triangle have the slopes 1, 2. Show that the triangle
is isosceles.

26 Find the tangents of the interior angles of the triangle with vertices
^(1,1), B(3, 3), C(2, 4).
i/ = 1, 2?/ — a; = 2, and Sx —
+
26 The lines x
.
y = 3 form a triangle.
Find the tangents of its interior angles.
27 . By showing that the angles ACB and ADB are equal prove that the
points A{2^ 0), B(3, —3), C(l, 1), Z>(— 2, 2) lie on a circle.
28 . If k is constant, show that

aix + hiy +C1A- k(a2X + 622/ + C2) =0

is the equation of a line through the intersection of the lines represented by

aix + hy + Cl = 0 , 02^ + hy -f C2 = 0 .

29By using the result in the preceding problem find the equation of the
.

line through the intersection of the lines 2a:+y — 5=0, x — 2y--5=0

and perpendicular to 2/ 3x — 1 =0. +
30 . Show that
m r
- - I,

y = yi -^[x xi
I

\
x xi\]

represents a locus which for x < xi coincides with the horizontal line y = yi
and for x > xi coincides with

y = yi+ m(x — xi).

31 . Determine the equations of the two lines which bisect the angles between

X +y 7 = Of 7x — y — 1=0.
32 Determine the lines which bisect the angles between x — 3t/
. = 0,

2/ + 3x = 0.
33 Find the equations of the tangent and normal to the curve y
. = x®
at (2, 8).
34 Find the equations
. of the tangent and normal to the curve xj/ = 6 at
(2,3).
35 . Find the equations of the tangent and normal to the curve y^ = 4x at
(1 ,
-2 ).

36 Find the equations of the tangent and normal to the curve

.

x^ -h xjy + = 3
at (1, 1).
37 . Determine the equations of the two tangents to the curve

at the origin.
38 . Show that the tangent to the curve y^ = x® at (xi, yi) intersects the
curve a second time at (^xi, —^2/ 1 ).
39 . A tangent is drawn to the curve xy = o at (xi, 2/1 ) • Determine the area
of the triangle it forms with the coordinate axes.
140 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

40 . A tangent is drawn to the curve

at the point (xi, yi). Find the length of the part of this tangent intercepted
between the coordinate axes.
41 Find the equations of the two tangents to the curve y =
. which pass
through the point (2, 3).
42 By Newton’s method find a root of the equation
.

+ 3x = 16
between 2 and 3.
43 By Newton’s method
. find a root of the equation

X® -h 3x + 2 = 0
between — 1 and 0.
44 By Newton’s method find a root of the equation
.

3x^ - fix^ = 1
between 2 and 3.
45 Find the angle between the straight
. line y — 3x +2 = 0 and the
curve y = X® at (1, 1).
46 Find the angle between the curves y^
. = 4x and x^ = 4y at each inter-
section.
47 . Find the angles at which the straight line x + 2y = 5 and the curve
+ y^ = 5x intersect.
48. Show that the curves xy = a, y^ ^ z^ = h are orthogonal for all values
of the constants a and h,
49 . Show that the curves x^ 2y^ = o^, x^ = by are orthogonal for all
+
values of the constants a and b,
60. Find the perimeter of the triangle with vertices (2,3), (—3,3), and
(1, 1 ).
61 . Show that (6, 2), (—2, —4), (6, —6), and (
— 1, 3) are on a circle with
center (2, —1).
52 . Find the point on the x-axis equidistant from (0, 4) and (—3, —3).
53 . Find the equation satisfied by the coordinates x, y of a point equidistant
from (0, 2) and (—3, 4).
54 . Find the equation of a curve such that any point P(x, y) on it is equi-
distant from (2, 0) and the y-axis.
56 . Find the equation of a curve such that the distance of any point on it
from (4, 0) is twice its distance from the origin.
66 . Each point on a curve is twice as far from (0, 6) as from the x-axis. Find
the equation of the curve.
57 . Find the equation of the circle with center (3, 4) and radius 6.
68 . Find the equation of the circle having A (—2, 1), B(4, 3) as ends of a
diameter.
59 . Find the center and radius of the circle

x^ + y^ + 6x — 2y == 6.

60. Find the center and radius of the circle

4(x2 + 2
2/ )
- 16x + I2y - 11.
 PROBLEMS 141

61. Show that the equation

+ 2/2 - 4x + 6j/ + 13 =0
is satisfied by the coordinates of only one real point.
62. Show that there is no real point which satisfies the equation

+ 2/^ — 4- 4?/ +8 = 0.

63. Find the equation of the circle through (—2, 4) and having the same
center as the circle

66.
+ 1/2 - + 2?/ + 6 = 0.

64. Show that the circles

3j2 _|_ ^2 _ 2x + 4ty + C = 0

have the same center for all values of the constant C.
Find the equations of the circles through (1,2) and tangent to both
coordinate axes.
66. Find the equations of the circles with centers at the origin and tangent
to the circle
a;2 -1- 7/2 - 8x + 62/ + 9 = 0.

67. Given A( — 1, 0), J5(l, 0), find the equation satisfied by the coordinates
Xy y of each point P{Xy y) on the circular arc APB if the angle from AP to BP
is 45^
68. Find the equation of the circle through the points (4, 2), (5, —5), and
(-3, 1).
69. Find the equation of the circle which circumscribes the triangle with
vertices (2, —2), (—4, 6), (3, 5).
70. Two circles are represented by the equations

a:2 + 2/^ + 3x - 22/ - 9 = 0,

^2 + 2/^ — X + 22/ — 6 = 0.

By subtracting these the equation of a line

4a: — 4y — 4 = 0

is obtained.How is this line related to the two circles?

71. A
point P{Xy y) moves so that the ratio of its distances from two fixed
points is a constant k. If k is different from unity, show that the locus is a
circle.
72. Find the locus of points from which tangents to the circles a;2 -f 2/2 = 4,

x2 + •“ 2a; + 4y = 4 are of equal length.

73. If Pi(xi, 2/i) is outside the cirele

a;2 + 2/2 + Ax + % + C = 0,
show that
3^1 + 2/1 + Axi + Byi + C
is the square of the length of the tangent from Pi to the circle.
142 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

Plot the following parabolas and determine their axes, vertices, and foci:

74. = 8z. 76. =* —6a;.

76. j/2 « 4a; + 4y. 77. a;^ ^ 5^ _ 0
78. a;2 - 6a; - 42/ +1« 0. 79. a;^ _ 12a; + 122/ » 0.

80. Find the focus and directrix of the parabola represented by

2/2 - 62/ 4- 8a; - 7 = 0.

81. Find the focus and directrix of the parabola

a;2 + 2a; - 62/ + 13 = 0.

82. Show that the parabolas

2/2 = 4px -f 4p2

86. the same focus for all values of p.

have
83. An arch has the form of a parabola with vertical axis. If the arch is
12 feet high at the center and 30 feet wide at the base, find the height above
the base at which the width is 15 feet.
84. An arch in the form of a parabola is 20 feet wide at the bottom, and the
highest point is 16 feet above the base. What is the length of the beam placed
horizontally across the arch 4 feet from the top?
A cable of a suspension bridge hangs in a parabola with vertical axis.
The roadway, which is horizontal and 240 feet long, is supported by vertical
wires attached to the cable, the longest being 80 feet and the shortest 30 feet.
Find the length of the supporting wire 40 feet from the middle.
86. A parabolic arch has base 6 and altitude h. Determine the area bounded
by the arch and base.
87. A parabolic arch of base h and altitude h is rotated about the base.
96. the volume generated.
I^nd
88 The normal to the parabola 2/^ = 4pa; at Pi(a;i, yi) intersects the a;-axis
.

at N, Find the projection of P\N on the x-axis.

89. Show that the tangent at any point P(x, y) of a parabola makes equal
angles with the line parallel to the axis and the line from P to the focus.

Plot the following ellipses and for each determine the center, axes, foci, and
eccentricity:

90. a;2^ ^y ^ 4 2
91. 4^2 + gyS ^
92. + 42/2 + 4a; - 82/ + 4 » 0. 98. 16a;2 + 9y^ » 144.
94. 3x2 4. 4y2 _ 0^ + 4y + 1 ^ 0.
9x2 4. ^y2 _ 303. _ 30^ 4- 36 = 0.

96. Find the equation of the ellipse with center (—2, 4) tangent to both
coordinate axes.
97. Find the equation of the ellipse with foci (0, 0), (0, 4) and passing
through (3, 0).
98. Find the equation of the ellipse with center (—2, 1) and passing through
(0,4) and (4, 2).
99. Find the equation of the line tangent to the ellipse 2x2 4. 3^2 ^ 5
at the point (1, 1).
 PROBLEMS 143

100. If the ends of a bar move on perpendicular lines, show that a point P
of the bar at distances o, and h from its ends describes an ellipse with semiaxes
a and h.

101. A circle is deformed in such a way that distances of its points from a
fixed diameter are all changed in the same ratio. Show that the resulting
curve is an ellipse.
102. By using the result in the preceding problem show that the area within
the ellipse
r
:;2 + 52
^

is ira5.

103. Show that lines from a point P on an ellipse to its foci make equal
angles with the tangent at P.
104. If the sum of the lengths of the tangents from P(x, y) to two fixed
circles is a constant, show that the locus of P is an ellipse.

Plot the following hyperbolas and for each determine the center, axes, foci,
eccentricity, and asymptotes:
106 9 x2
.
_ 4^2 ^ 30 106 4x2 . - _ 83.
2/2 _|. 4y ^4
107. 2/^ - 2x2 + 8x + 22/ = 11. 108. x2 - ^2 _ 2x - 62/ = 18.
109. Find the equation of the hyperbola with foci (0, 0), (0, 4) if the differ-
ence of the distances from the foci to any point on the curve is 2.
110. Find the equations of two hyperbolas with center (2, —1) and semi-
axes 1 and 2 parallel to OX and OY respectively.
111. Find the equation of the hyperbola with center ( — 1, 1), axes parallel
to the coordinate axes, and passing through (0, 2) and (1, —4).
116.
112. A point moves so that the product of the slopes of the lines joining it

to (~a, 0) and (a, 0) is constant. Show that it describes an ellipse or hyper-

bola.
113. On a level plane the sound of a rifle and that of the bullet striking the
target are heard at the same instant. Find the locus of the hearer.
114. A point P(x, y) moves so that its distance from a fixed point divided
by its distance from a fixed straight line is a constant e. Show that the locus
is an ellipse, parabola, or hyperbola according ase<l,e = l, ore>l.

Two vertices A, P of a triangle are fixed and the third C(x, y) is vari-
able. Find the locus of C if ZP = 2ZA.
116. Show that the ellipses and hyperbolas obtained by assigning values
to the constant k in the equation

= 1
a2 ~ A;
52 - fc

all have the same foci.

117. If F, F' are the foci and P any point on a hyperbola, show that the
tangent at P bisects the angle F'FF.
118. What are the codrdinates of the points (2, 3), (—4, 6), and (6, —7)
referred to parallel axes through the point (3, —2)?
119. Find the equation of the curve
x2 + 42/2 - 2x + 81/ + 1 = 0

referred to parallel axes through (1, —1).

144 ALGEBRAIC EQUATIONS AND GRAPHS [Chap. IV

120. Find the equation of the curve

yZ _ 62/2 -f 3a:2 + 122/ - l&c + 35 = 0

referred to parallel axes through (3, 2).

121 . Transform the equations a; 2/
— 3=0, 2a; +3^/ —+ 0 to par- 4=
allel axes so chosen that the new equations have no constant terms.
122 . Transform the equation
.

4" — 2a; -f 42/ — 6 = 0

to parallel axes so chosen that the new equation has no terms of the first degree
in X and 2/.
123 Determine the position
. of a new origin such that

2/2 + 42/ - 2a; +3

has no constant term or first power of y,
124 Transform the equation
.

4- a;2/ 4- 2/^ = 1

to new axes bisecting the angles between the original axes.

126 Determine the equation of the curve

V3(2/^ — x^) - 2x2/ = 4

when referred to axes making an angle of 60® with the original axes.
126 . By rotation of the axes change
x^ — 2/^ = 4

to a form that does not contain x^ or 2

/^.

127 Through what angle must the axes be rotated so that

.

x^ xy — \

shall be changed to a form that has no xy term?

128 By rotation of the axes reduce
.

x^ — 2xy 4 - = 4(x 4- 7)

to a form which has no xy term.

129 Determine the form into which
.

4" /^ ss
2

is changed when the axes are rotated 45®.

130 Show that
.

dx^ +
is invariant under a change of rectangular axes.
131 . Show that
ydx -- xdy

is invariant under rotation of the axes about the origin.

 PROBLEMS 146

Make graphs of the following equations:

182. y=(x- l)(x + 2)(x - 3). 183. y = x*(x + 1).

134. X = y^(y - 2). 186. y = - 2)»
(x
186. y = x* - 16. 137. x< - 5x2 + 4.
=
138. j/ + 1 = x(x - D*. 189. x-2 = yHy- 1)*.
140. y^ = X*. 141. j/2= x* — 4x.
142. y* = 4x - X*. 143. x2 = - y*.
j/2

144. j/* — x^ — X*. 146. = x2.

146. 2/
= X + -X •
147. p = x* + i-
x^
148. xy = X® - 2. 149. xj/2 = 1 — X.
1
160 . 161. x^j/* + 36 = 42/2.
x + i‘

162. y^^x^- —+ -
a X
164. x» + = 2
/* 1.

166. x*®® + - 2/"’® 1.

168. 2/2 = x2.

160. x^ + 2/^ = oi
 CHAPTER V
DETERMINANTS

70. Determinants of the Second Order. If the two lines

aix + biy = Cl,

a2X + 1*22/ = C2

intersect, the coordinates of the point of intersection may be found

by solving the two equations simultaneously. To do this eliminate
the unknowns one at a time, thus obtaining

(O162 — 0,^i)X = C162 — C261,

( 2)
(0162 — ^2^1)2/ “ (I1C2 — 02^1.

If aih2 — a2hi is not zero, these equations give

C162 — ^2^1 ttlC2 — 0,2^1

(3)
CI162 — U2&I CLib2 — U2^1

as the coordinates of the point of intersection.

Because of the frequency with which such combinations occur,
the expression 0162"" (^2^1 is given a special notation.

ai bi
(X162 — U2&1 “ ^ (4 )
(I2 62

and a special name, determinant The numbers, such as Ui, 61, in a

horizontal line constitute a row, those, such as 61, 62? in a vertical
line constitute a column. The above determinant, having two
rows and two columns, is called a determinant of the second order.
In terms of determinants the solution of equations ( 1 ) can be
written

Cl 61 ai Cl
C2 62 02 C2
> y = "

ai bi Oi bi
02 62 O2 h
146
§71] GENERAL DEFINITION OF DETERMINANT 147

The use of determinants makes the solution of two equations in

two unknowns, and as we shall see later the solution of n equations
in n unknowns, analogous to the solution of one equation

ax ^ c ( 6)

in one unknown. In the latter case the coefficient a of the one un-
known may be regarded as a determinant of the first order. The
solution
c
(7)
a

is a fraction having this determinant as denominator, and with

numerator obtained from the denominator by replacing the coeffi-
cient of X in the equation by the constant term c. Similarly, each
denominator of (5) is the determinant formed from the coefficients
of all the unknowns in the equations, and each numerator is formed
from its denominator by replacing the corresponding coefficients
(ai and a2 in case of x, bi and 62 in case of y) by the constant
terms ci, C2 respectively.
To show how determinants are used in solving simultaneous
linear equations in more than two unknowns it will be necessary
first to define determinants of the nth order and obtain some of

their properties.
General Definition of Determinant. Consider an array of
71.
numbers arranged in a square of n rows and n columns. Such an
array is called a square matrix, and the individual numbers are
called its elements. For example, if n = 3 the matrix has the form

ai hi Cl
U2 ^2 ^2
U3 ^^3 C3.

The rows are considered as arranged in order from the top down-
ward, the columns from left to right.
With such a matrix is associated a determinant of the nth order
defined as follows:
( 1 ) Form all possible products of n elements so chosen that no

two belong to the same row or column.

(2) Consider the elements in each product as placed in the posi-
tions they occupy in the matrix and joined in pairs by straight
lines. If the number of those lines which slope upward to the
 148 DETERMINANTS [Chap. V
right is even or zero, precede the product by a + sign; if the num*
ber of such lines is odd, precede the product by a — sign.

(3) Each product together with its attached algebraic sign is

called a term, and the sum of all such terms is called the determinant
of the matrix.
The determinant is represented by the square array with a
vertical bar placed on each side.
As an example take the third-order determinant

a\ hx Cl
U2 ^2 ^2

^3 ^3 ^3

The following six products can be formed having as factors one

element from each row and one from each column:

^^l&2^^3> Cia2&3>

^1^2^31 ^^1^2^31 ^162^3*

By rule (2) above, the sign to be attached to the first three is +,

that to the last three — . Thus

Ul 61 Cl

^2 ^2 ^2 == ^162^3 ”f“ 6102^3 Cia^^ — axc^^ — 61U2C3 — Ci 62 ® 3 »

U3 63 C3

72. Properties of Determinants. The following properties of de-

(3)
terminants are almost immediate consequences of the definition.
(1) If all elements of a row or column are ^ero, the determinant is
zero,

(2) If all elements of a row or column are multiplied by the same

factor k, the determinant is multiplied by k.
Thus
kai kbi kci ax bi Cl

^2 &2 C2 = k as 62 C2

as h Cb as 63 C3

If two rows or two columns are inlerchanged, the determinant

merely changes algebraic sign.
Thus
61 ax Cl ai hx Cl

62 02 C2 = — 02 62 C2

h as Cs Os 63 C3
 ^
2]
(4)
PROPERTIES OF DETERMINANTS 149

If two rows or two columns are identical^ the determinant is zero.

Thus ,

Oi &2 Oi
02 b2 02
(5)
h «3

the first and third columns being identical.

If the corresponding elements of two rows or two columns are
proportional^ the determinant is zero.
Thus
(6) 1 2 3
k 2k 3k = 0,
7 5 4

the second row being k times the first.

If all elements of a row or column are binomials the determinant ^

is the sum of two determinants obtained by deleting one term or the

other in each binomial.
Thus
Oi Xi + 2/1 Cl Oi Cl Oi 2/1 Cl
O2 X2 + ^2 C2 = 02 X2 C2 + 02 2/2 C2
O3 xz + Vz C3 03 ^3 cs 03 2/3 C3

(7) The determinant is not changed in value if to each element of

one row is added k times the corresponding element of another row,
or to each element of one column is added k times the corresponding
element of another column.
By repeated application of this theorem it may be shown that
the determinant does not change value when several rows are
multiplied by constants and added to another row, or several
columns multiplied by constants and added to another column.
Thus
ai + mbi + nci bi Ci ai bi Ci

a 2 4“ mb 2 4” nc2 62 ^2 ~ ^2 ^2 ^2

03 4- mbs + nc3 63 C3 as 63 C3

(8) If the rows are changed into columns without change of order,
the value of the determinant is not changed.
Thus
ai 61 Cl ai 02 03
02 62 ^2 ~ ^2 ^3

03 63 C3 Cl C2 C3

By application of these properties the work of evaluating a

determinant can often be greatly reduced.
150 DETERMINANTS [Chap. V
Example 1. Find the value of

1 2 3
3 2 1-
111
Adding corresponding elements of the first and second rows, dividing
by 4, and subtracting from the third row results in

1 2 3
3 2 1 =0.
0 0 0

Example 2, Find the value of

3 5 7 1
4 3 12 *

2 0 3 1
6-9-3 3

Multiply the last column by 2 and subtract from the first, then mul-
tiply the second row by 3 and add to the fourth. In this way we obtain

1 5 7 1 1 5 7 1
0 3 1 2 0 3 1 2
0 0 3 1 0 0 3 1

0 -9 -3 3 0 0 0 9

The only non-vanishing term in the last determinant is

1 -
3 3-9
-
= 81 .

73. Minors and Cofactors. In a determinant delete the row and

column to which a given element e belongs. The determinant
formed from the remaining elements without change of order is
called the minor of e.
In the determinant
CLl bi Cl
(l2 62 ^2 f

dS ^3 C3

for example, the minor of ai is

62 ^
bs C3
and that of 62 is

ai Cl

03 C3
§ 73 ] MINORS AND COFACTORS 151

If a determinant contains an element e, certain of its terms con-

tain e as factor. If the sum of such terms is

eEy
E is called the cofactor of e.

The minor and cofactor of a given element e are related in a

simple way. To determine this relation suppose first is the that e
element in the upper left-hand corner of the determinant. The
definition of determinant shows that the minor and cofactor are
then equal. Suppose next that e is the element in the rth row
and sth column. By r s — 2 interchanges of two adjacent rows
+
or two adjacent columns it can be moved to the upper left-hand

position. These moves leave the minor unchanged but change the
sign of the determinant and so that of the cofactor r + s — 2 times.
For the element in the rth row and sth column we thus have

cofactor = (— l)^"^'’"^ (minor) = (— (minor)

In the third-order determinant above, for example, the cofactor

of ai is

62 ^2
’
hz C3
and that of 61 is
02 C2
U3 C3

In the applications of determinants the following properties of

cofactors are fundamental:

(1) The sum of the 'products of the elements in any row, or column,
by their cofactors is equal to the determinant
For, if On are the elements of a row or column, and
ai, 02, • •
-
j

•> -An are their cofactors, then aiAi is the sum of terms
• •

containing Oi, 02^2 the sum of terms containing 02, etc. Since
every term contains an element from the given row, or column,
the sum of all terms is

aiA\ + CI2-A2 + • •
•
+ UnAn.
For example.

Oi bi Cl

02 &2 ^2 = ai
62 ^2
— 02
<=1
Ll « 61 Cl

— i &3 ^3 O3 C3 1
62 ^2
^3 O3 C3
152 DETERMINANTS [Chap. V
(2) The sum of the products of elements in any row {column) hy
the cofactors of corresponding elements in another row {column)
is zero.
Thus, if di, 02, • •
•, On are elements in one column and Bi, B2,
• •
•, Bn cofactors of corresponding elements in another column

aiBi + 02^2 H h CLnBn = 0.

For, the sum on the left may be obtained from the expansion of
the determinant
biBi + 62^2 + • •
•
+ hnBn
by replacing 6i, 62? *, On, and so is a determi-
• *
*, bn by Oi, 02, • •

nant with two columns equal to Oi, 02, *, On (§ 72 4 ).

• •
,

74 Solution of n Linear Equations in n Unknowns. As an il-

.

lustration take three equations

®11^1 + ^122:2 + Ul3^3 =

d21^1 4” ®22^2 H“ <^23^3 ~ ^2? (1)
^31^1 4" ^32:^2 4” ^ 33^3 ~

in three unknowns xi, X2f <cs. Let Afs be the cofactor of an in the
determinant
On 012 Ol 3
A = 021 O22 023 ( 2)

031 O32 O33

formed from the coefficients of all the unknowns. Multiply the

first equation by An, the second by A21, the third by A31, and
add. By theorems 1 and 2, § 73 , the coefficient of xi in the sum is

oii^ii 4 -
021 4-21
-
4“ 031-^31 = (3 )

and the coefficients of X2 and X3 are zero. Also the right side

Cl An + C2A21 4“ C3A31 (4 )

is obtained from A by replacing on, 021, 031 by ci, C2, C3, respec-
tively. The sum is therefore

On 012 Ol 3 Cl 012 Ol 3
021 022 023 Xi = C2 022 023
031 032 033 C3 032 O33
§74] SOLUTION OF n LINEAR EQUATIONS 153

If A is not zero, we thus have

Cl ai2 ®13
C2 022 c^23

cs «32 c^33
Xi = (5)
an ai2 C^13

021 a22 (^23

C^31 a32 «33

and similar expressions are obtained for X2 and x^.

That these values actually satisfy equations (1) may be shown
by direct substitution. Thus if we substitute in

anxi + ai2X2 + aisXs — ci = 0

and clear fractions the left side will be found to be a determinant

of the fourth order with two rows equal to an, ai 2 , ai 3 , Ci, and
therefore zero (§72, 4).
A similar argument is applicable to n equations in n unknowns,
thus proving the following, known as Cramer^ s rule:
A set of n equations

+ Ul2^2 H •
^In^n — Ci^

021^1 + ^22^2 + * * ”1" 02nXn C2 ,

«nl^l + Un2^2 + * • *
~f“ ClnnXn ~ Cny

in n unknowns xij X2 j
• •
Xn has one and only one solution if the
determinant
an ai2 •

a2i 022 '02n

A = •

Ofil Un2* 'Ofin

is not zero. Each unknown is a fraction with denominator A and

numerator obtained from A by replacing the coefficients of that un--
known by the corresponding constant terms Ci, C2 •, Cn. ,
• •

Example, Show that the equations

x +y+z= 2y
2x - y - z = I,

X + Sy + 4tZ = Sf

have a unique solution, and find the value of y.

164 DETERMINANTS [Chap. V
The determinant of the coefficients is

1 1 1
2 -1 -1 = -3.
1 3 4

Since this is not zero, the equations have a unique solution. By Cramer’s
rule, the value of y is

1 2 1
2 1 -1

2 -1 -1
1 3 4

76. Homogeneous Linear Equations. Linear equations of the

form
+ ^^12^2 + *
h O^lnXn = 0,

®21^1 + ^22^2 H h 0.2n^n == 0, (1)

“f“ ^n2^2 “f"

* * *
“f" (^nn^n —
with constant terms all zero are called homogeneous. The values

xi = 0:2 = • • • = Xn = 0 (2)

constitute an obvious but usually trivial solution. The question

of importance is whether there is any other solution.

For a system of n homogeneous equations in n unknowns to have

a solution in which the unknowns are not all zero it is necessary
and sufficient that the determinant of the coefficients be zero.
To prove this we reduce the equations to simpler form by the
process of elimination ordinarily used in solving such equations.
For this purpose, take any one of the equations, which we call
equation (1), and designate by Xi any unknown in itwith co-
efficient not zero. By multiplying this equation by appropriate
constants and subtracting eliminate xi from all the other equations.
Next, take any one of the remaining equations, which we call
equation (2), designate by X2 one of its unknowns with coefficient
not zero, and by means of this equation eliminate X2 from all the
other equations, including number (1). Continue this process
until the remaining equations, if any, do not have any coefficients
different from zero. Suppose that this happens after r equations
have been treated in the above way. The final equations have the
§76] HOMOGENEOUS LINEAR EQUATIONS 155

same tho original ones, for in each change we can re-

solutions as
obtain the preceding equations by merely adding proper multiples
of one equation to the others. Also these operations do not change
the determinant of the coefficients (except possibly in sign, owing
to change of order), for in each case the change consists in sub-
tracting a multiple of one row from another row (§72, 7). The
above process can end in one or other of two ways. First, if
r = n, the final equations take the form

biXi = 0, 62X2 = 0,
• *
*
, bnXn = 0

with bij bn, •

•
•, 5n all different from zero. The only solution is

then
Xi = X2 = • • • = Xn = Oy

and the determinant

616263* *
*6n

is not zero. Second, if r < can be assigned to

n, arbitrary values

Xr+i, • •
'y Xn, and Xiy X2, can then be determined from the
' *
Xr
first r equations. In this case the determinant is zero since all
elements in the r + 1st row are zero. Thus there are solutions
differentfrom zero when the determinant is zero and none when
the determinant is not zero, which was to be shown.

Example 1, By means of a determinant express the condition that

three points (xi, yi), (^2, 2/2), (^3, Vz) He on a line.

If
Ax + By + C = 0

is the equation of a line through the three points,

Axi + Byi + C = 0,
Ax2 + +U= J5?/2 0,
A.T3 + Byz + C = 0.

These equations are homogeneous in the unknowns Ay By C. For a solu-

tion to exist with Ay By C not all zero, the coefficient determinant must be
zero. Thus
xi yi 1

X2 y2 I =0 (3)
X3 yz 1

is the condition required.

Example 2. Find the equation of the circle through three points
fe, 2/1), {^2, 2/2), {xzy yz) not on a fine.
Let
A{x^ + 2/^) + + Cy + D = 0 (4)
156 DETERMINANTS [Chap. V
be the equation required. Since the circle passes through the three points

A(xi + y\) + Bx\ + Cy\ + D = 0,

A{x\ + + Bx2 + Cy2 + D =
2/2 ) 0,
-^(^8 + + Bxz + Cyz + D =
2/1) 0.

These equations and (4), being homogeneous in four unknowns A, J5, C, D,

must have a coefficient determinant equal to zero. Thus

+ y^ X y 1
x\ + y\ xi yi 1
= 0 . (5)
xl 4- 2/2 X2 2/2 1

Xz + 2/3 ^3 2/3 1

When expanded this has the form (4) and so represents a circle through
the three points provided that
xi 2/1 1

A = X2 2/2 1

Xz yz 1

is not zero, that is, provided that (xi, 2/1 ) > fe, 2/ 2 ), (^ 3 Vz) arc not
,
on a line.

If, however, the points are on a line, A = 0 and (5) is the equation of the
line.

PROBLEMS
Solve by determinants:

-
jSx +y= If
2
\x +2y ^ 4,

lx - 22/ = 5 .
.
[2x - = 2/ 3.

O jx+ 31/ - 1 = 0, j3x — 22/ — 2 = 0,

4.
\3x-y + 7 ^0, 15x + 42/ + 4 = 0.

Evaluate the following determinants:

1 2 3 4
6. 2 3 1 2
3 1 2 10

2 4 2
7. -3 -1 -3
4 8 1

1 2 3 -1
-1 4 6 1
9.
3 1 7 -2
2 3 9 3

Solve the following systems of equations:

X + 22/ - z = 1, X — 22/ — z « 2,
11 . 2x — 2/ 3z = 2,
+ 12 . 2x — 32/ — 2z — 1,
3x — 22/ « = 4.
+ 3x — 2y — 2z « 4.
 PROBLEMS 167

a: - y +« = 4, [a;4-2/4-2~-9 =0 ,

13. 2x Zy 2z — 3, 14. x-2/--2 + 3=0,

4a; 4- 22/ H- 3« = 8. (a;4-2/-“2 — 6 =0.

a;4-t/H-2 + w=5, x4-.v+2-w = l,

a; — 22/ + 3« 4- w = 1, 2x — 2/ — 2 4“ 3w = 8,
16. 16.
3x 4“ 2/ 2 4* 2u = 4, 3x 4" 22/ H- z - M = 6,
2x 4" 32/ 4“ 2z — u = 3. X — 32/ — z 4” 4w = 7.

17. Show that the equations

x 4- 2/
= l, x 4- 2/=2
have no simultaneous solutions.
18. If
aix 4“ hiy 4- ciz = 0 ,

oux 4- h^y 4- C2Z = 0 ,

show that
61 Cl Cl ai ai 61

y \ z ^
62 C2 C2 fl2 a2 62
19. Given
X 4- 2/
— 2z = 0,
5x 4- 2/
4- 2 = 0,

find three numbers proportional to x, y, z.

Show that each of the following sets of equations has solutions in which the
unknowns are not all zero, and find tliree numbers proportional to x, 2/,
*•*

X 4- 2/
4- 2 = 0 ,
2x — 2/ + 3z = 0 ,

20 . a; 4“ 42/ 4- 3z = 0, 4x 4- 32/ 4- 62 = 0,
5x — 2/ 4- 2 = 0. i 6x
— 62/ 4- 9z = 0.

X 4 .
2/
- z = 0, [x 4- 37/ = 0,
22 . X — 27/ — 4z = 0, 23. 27/ 4- 2= 0,
2/
4- 2 = 0. I 2x — 3z = 0 .

24. Show that

X y \

2/1 1 = 0
a:2 ^2 1

is the equation of the straight line through (xi, 2/ 1 ) and (x 2 , 2/ 2 ).

26. Show that

2/^x1
4 1 1=0
9 -1 1

isthe equation of the parabola with axis parallel to the x-axis and passing
through ( 1 2 ), (-1,3), (- 1 -3).
, ,

26. Find the equation of the parabola with axis parallel to the x-axis and
passing through ( 2 3), (—3, 2), (—3, —2). ,
168 DETERMINANTS [Chap. V
27. Find the equation of the circle which passes through ( 0, 1 ), (4, 0 ), and
( 2 , 6 ).
28. Determine whether the four points (4, 2 ), (5, 1 ), (— 2 6 ), and (—3, —5)
,

are on a circle.
29. If the three lines
aix + hiy + Cl =0,
oua; + + C2 = 0,
622/

aaa; + 63^ + ca = 0
pass through a point, show that

ai bi Cl

02 62 C2 = 0.
U3 &3 C3

Is it true conversely that if the determinant is zero the lines pass through a
point? Consider the case of parallel lines.

80. Determine whether the lines

a; ~ 2 + 1 =
!/ 0 ,

2x + y - 1 == 0,
3a; + 42 — 3 /
= 0,
pass through a point.
31. If the equations
aix + biy 4- Cl = 0,
a2X 4" b2t/ 4- C2 = 0,
asa; 4- bsy 4- C 3 = 0 ,

have a simultaneous solution, show that constants ki, k2 , ks, not all zero, exist
such that

ki(aix 4- biy 4- ci) 4* A; 2 (a2a; 4- 622/ 4- C2 ) 4- ksiasx 4- bsy 4- C3 ) =0

for all values of x and y. Is the converse true? Take, for example, the equa-
tions X +y — 1 = 0, X y — 2 -0, x-hy-~3=0.
82. Determine numbers A;i, k 2 ks such that
,

A;i(2a; 3y — z) -j- k2(3x + 4y — 3z) 4* A;3 (a; 4- 3y 4- 4z) =0

for all values of a;, y, z.
 CHAPTER VI

TRIGONOMETRIC FUNCTIONS

76. Sine Curves. If a, b, c are constants the graph of the equa-

tion
2/
= a sin {bx + c) (1)
is called a sine curve.
Unless the contrary is assumed that the
explicitly stated it is
angle bx + cis measured in radians. x changes, When
this angle
changes. As the angle varies from 0 to 2t the sine runs through
its entire range of values and the point (x, y) traces a complete

wave AC of the curve (Figure 103). As the angle continues to

increase, the sine runs repeatedly through the same range of

values. Thus the entire curve consists of an number
infinite of
such waves.
The distance between points at which the angles differ by 2t is

called the wave length. In a displacement between two such points

A(]bx + c) = bAx = 2w,

whence
2ir
wave length =
I

|
Aa: = t^. . ( 2)

The sine has its maximum value 1 when the angle is - ,

and its

159
160 TRIGONOMETRIC FUNCTIONS [Chap. VI

minimum — 1 when the angle is — “ . The curve thus lies between

the lines y = one of these lines at the middle of each

dia, touching
half wave. The coefficient a measures the height of the wave
crest above the axis y = 0.

The curve crosses the x-axis at the point where

+c= 0.

A change in c merely changes the position

X =
6
at which this crossing occura
Since

cos {bx + c) = sin (b^ + c + ^,

the equation
1/ = a cos (bx + c)
also represents a sine curve, namely the curve obtained by moving
(1) a quarter wave length parallel to the axis.

Example, Plot the curve

2/
= 2 sin (2 — 3x).

This curve lies between the

lines y = ±2. It crosses the
a;-axis at the point A (Figure
104) where 2 — 3x = 0,
X = f

When X increases from this

value the sine becomes nega-
tiveand so the curve slopes
downward. A complete wave
length is

At the point R, halfway between A and C, the angle is — ir and y is zero.

The half waves AB, BC can now be sketched and the remainder of the
curve obtained by indefinite repetition.
§77] PERIODIC FUNCTIONS 161

77. Periodic Fimctions. If is constant and

fix + A) = f{x) ( 1)

for all values of x, the function f{x) is said to be periodic and the
number k is called its period. If the part of such a curve between
X = xi and x = Xi + k is constructed, the entire curve is the
infinite set of parts obtained by moving this one to the left or to

the right through the distances k, 2k, etc.

Thus sin (bx + c) is a periodic function of period

b ‘
The co-

sine, tangent, cotangent, secant, and cosecant of bx +c are

also periodic functions of x with the same period. In addition
tan (bx + c) and cot (bx + c) repeat their values when the angle
is increased by tt. These functions therefore have the shorter

period

A function of period k has also the periods 2k, 3fc, etc. Some-
times an expression consists of parts each of which is periodic but
with different periods. If these periods are commensurable, their
least common multiple is a period for the entire expression.
In the function
X X
fix) = sin - + tan -
2 3

for example, the sine is of period 47r and the tangent of period
The sum is periodic with period equal to the least common mul-
tiple of these, namely, 127r.

Example 1, y = tan 2x.

When the angle 2x increased by increased by the tangent of

is tt

^
2x is not changed. The function is thus periodic of period ^tt. The curve
therefore consists of a series of parts each of which is obtained by moving
the preceding one a distance Jir to the right. One of these branches
TT
passes through the origin. As x increases from 0 to 2x = y increases

from 0 to infinity. The line a: = is therefore an asymptote. As x

decreases to — Jtt, y decreases to — oo. The branch through the origin
thus extends from x = y = — «», to x = Jtt, y = <x>. The whole
curve is a series of such branches at distances ^tt apart (Figure 105).
162 TRIGONOMETRIC FUNCTIONS [Chap. VI
§78] DIFFERENTIATION OF TRIGONOMETRIC FUNCTIONS 163

the expression —00,00 being used to signify that on the left of this position
y becomes indefinitely large and negative, and on the right indefinitely
large and positive. The curve is a series of U-shaped branches alternately
above y — I and below y = — 1 (Figure 106).
Example 3. 7/ = 2 sin -f- 6 cos jx. To construct this curve draw
the curves
7/1 = 2 sin ^Xy 7/2 = 6 cos ixy

and on each vertical line determine the point P(x, y) such that y = yi + 7/ 2
,

that is (Figure 107) such that

MP = MPi -h MP 2.

To leave sin ^x and cos jx both unchanged x must be increased by Stt

or a multiple of Stt. Hence it is sufficient to plot the section AB from

X = — 27r to a; = Ott. The whole curve is a series of such sections placed
end to end.

78. Differentiation of Trigonometric Fimctions. Let u be the

independent variable or any differentiable function of the variable.

VII. d sin w = cos u du.

VIII. dcosu = ’-sin udu.
IX. d tan u = sec^ u du,
X. dcotu = —csc^udu,
XI. d sec w = sec u tan u du.
XII. d CSC w = —CSC u cot u du.

The negative sign occurs in the differentials of all cofunctions.

This is to be expected since for angles in the first quadrant the
cofunctions diminish as the angle increases.
Formulas VII and VIII were proved in §22. The other formulas
may be proved by expressing the functions in terms of sine and
cosine.
164 TRIGONOMETRIC FUNCTIONS (Chap. VI

Thus to prove IX we write

sin u
tan u
cos u
Differentiating this as a fraction and using VII and VIII, we have

cos u d(sin u) — sin u d(cos u)

d tan u
cos^ u
cos^ udu + sin^ u du
cos^ u
= sec^ u du.

Example 1, y = sec 2x + tan 2x,

By IX and XI,
dy = sec 2x tan 2x d(2x) + sec^ 2x d(2x)
= 2 sec 2x(tan 2x + sec 2x) dx.
Example 2, y = sin^ (3x — 2).
This is equivalent to
y = [sin (3a: — 2)]^ =
The derivative is therefore

dx dx

= 4 sin^ (Sx — 2) 4- sin (3a: — 2)

djy

= 12 sin® (3a: — 2) cos (3a: — 2).

Example 3, y = + cot — | a: a: cot® x.

By formula X the derivative is
~= 1 — + cot^ X csc^ X
csc^ X
dx
= — cot® X + cot® X CSC® X
= cot® a: (esc® a: — 1)

= cot^a:.

79. Inverse Trigonometric Functions. For values of x numeri-

cally less than 1 the equation

sin 2/
= a:

determines j/ as a function of x. This function is represented by

the notation
y = sin“^ X
§ 79 ] INVERSE TRIGONOMETRIC FUNCTIONS 165

and is called the angle whose sine is Xj or the inverse sine of x.

Since the angle is equal to an arc on a unit circle the equivalent

notation
sin""^ X = arc sin x
is also often used.
The relation of x and y is shown graphically in Figure 108.
For a given value of x (numerically less than 1) there are infinitely
many values of represented by the points at which the vertical
line of abscissa x intersects the curve. We express this by saying
the inverse sine of a; is a multivalued function of x^ in fact, an
infinitely multivalued function. To avoid ambiguity the notation
sin""^ X is therefore usually restricted to represent a particular one

of these values, namely the one between —^ and - . This is

2 2
called the principal value. In Figure 108 it is represented by the
part AB of the curve.

Figure 108 . Figure 109 .

Similar definitions apply to cos“^ x, tan"”^ Xj cot"”^ x, sec“"^ x^

and csc“^ x. Each of these notations usually represents a prin-
cipal value, selected partly for simplicity and partly to make the
formulas of differentiation take definite form.
166 TRIGONOMETRIC FUNCTIONS [Chap. VI

For sin *a: and tan the principal value is taken as the

angle in the interval In Figures 108 and 109 these

values are represented by the solid lines.

For cos~^ z and cot~^ x the principal value is the angle in the
interval (0, ir).
For sec“^ x and csc“^ x the principal value is taken in the inter-

val > 0 and in the interval if X < 0.

80. Formulas for Differentiating Inverse Trigonometric Func-

tions. The following formulas give the differentials of the prin-
cipal values. If other values of the inverse trigonometric func-
tions are used, the algebraic signs given in the formulas may not
be correct.
du
XIII. dsin-‘« =
Vl-u^‘
du
=
.

XIV. cicos-*M
vT^‘
du
XV. d tan ^ u =
1 +
du
XVI. d cot-^ u = ~
r+^*
du
XVII. d sec-^ u =
u'^/ u^ — 1

du
XVIII. d csc-^ u — *1
u\/u^

be noted that the differential of each cofunction merely

It should
from that of the corresponding function.
differs in algebraic sign
This follows since function and cofimction are complementary
an^es. Thus

cos * X = sm * X.
2

To prove any one of the formulas we change from inverse to

direct functions. Thus to prove XIII let

y — sin~* u.
§811 ANGULAR VELOCITY AND ACCELERATION 167

Then
smy — u.
Differentiation gives
cos y dy = du.

If y is the principal value cos y is positive and

cos y = V 1 — 2
sin^ y = vT-^.
Thus
du du
dy =
cos y Vl - u^
’

which was to be proved.

Example L = sin" L
^- V
y
X + 1
By formula XIII

2dx dx
dy =
(x + l)-v/x
- (M)’ + -
(M)’
"’V'
Example 2. y = \/x^ —
a sec“‘ -• —
a
Considering a as constant, we have
d /x\
dy d /-2 2
w

cVx* —
_ y/x^ —
X

81. Angular Velocity and

Acceleration. Consider a body
rotating about a fixed axis OA
(Figure 110). Let OM, OP be
lines perpendicular to the axis,
OM being fixed in space and L
OP rotating with the body, A
and let 6 be the angle from
OM to OP. This angle is con-
sidered positive when described Figure 110.
168 TRIGONOMETRIC FUNCTIONS [Chap. VI

in one direction about OA, negative when in the other. The

positive direction is often indicated by an arrow along the axis

pointing in the direction a right-hand screw advances when given

the positive rotation.
The angle ^ is a function of the time. Its derivative

de
__

dt

is called the angular velocity and its second derivative

doi d^S

the angular acceleration, of the body.

Rotation about a fixed axis is analogous to translation along a
fixed straight line; angle, angular velocity, and angular accelera-
tion corresponding to displacement, linear velocity, and linear
acceleration.

Example, A body set rotating by a constant torque about a fixed axis

has constant angular acceleration a about the axis. If it starts from rest,

find the angle turned through in t seconds.

In this case
d^e
a = constant.
dt^

Integrating and determining the constants so that 0 and co are zero when
^ = 0, we obtain

0 =

82. Simple Harmonic Motion. The motion of a particle along

a line is called simple harmonic if the displacement of the particle
from a fixed point of the line is represented by an expression of the
form
8 ^ Asm{bt + c), (1)

where A, b, c, are constants and t is the time.

As the time increases, the angle bt +c varies steadily, and s
varies periodically between —A and A, The motion is thus an
oscillation or vibration back and forth along the line, A being the
§82] SIMPLE HARMONIC MOTION 169

maximum displacement from the center s = 0. The numerical

value of A is called the amplitude of vibration.

The angle bt +c is called the phase^ or phase angle. As the

phase increases from 0 to 2ir the displacement s runs through its
complete range of values. Such a portion of the motion is called
a complete vibration. The time required for one vibration

T= -
2t
(2)

is the period, and the number of vibrations per unit time

&
^ _ 1 _
“ T ” (3 )
^
is the frequency of vibration.
The velocity of a particle in simple harmonic motion is

i; = — = Ab cos Q)t + c), (4)

dt

and its acceleration is

a = — sin Q)t + c). (5)

Comparison with (1) shows that the acceleration can be written

a = -bh. (6)

The acceleration is thus proportional to the displacement from

the center of motion and is oppositely directed.
A force F acting on a particle of mass m gives it an acceleration
determined by the equation
F = ma.

If such a particle is describing the simple harmonic motion (1)

it is then subject to a force

F = ma = — m6^s
proportional to the displacement s and directed toward the center.

Example 1. A particle P moves with constant speed around a circle

of radius r making n complete revolutions per second. Determine the
motion of its projection on a fixed diameter of the circle.
170 TRIGONOMETRIC FUNCTIONS [Chap. VI

Let OX (Figure 111) be the fixed diameter. If

e=-xop
is zero at time /b> its value at any other time t is

6 = 2nir(t — to).

Hie projection of P on the a:-axis is

a: = r cos 0 = r cos 2nir(t — k)

= r sin ^2n7r(^ — fe)

+0 *

The projection thus has simple har-

monic motion of frequency n and
amplitude r. The projection on the
^-axis is

2/
= r sin 2mr(t — to).

This also has simple harmonic motion of the same amplitude and fr(v

quency but with phase at each instant - less than the projection on the
2
x-axis.

Example 2. The displacement of a particle at time t is

5 = sin 6^ + R cos ht.

Show that its motion is simple harmonic.

Construct a right triangle of sides A, R, and let 6 be the angle opposite R.
The hypotenuse of the triangle is

C =
and
= C cos B ^ C sin B.
Consequently

8 = (7(sin ht cos 6 + cos ht sin 6)

= C sin ijht + 0).
The motion is therefore simple harmonic
with amplitude
c=
and period
2t A
T FrotmB 112.
h‘
 PROBLEMS 171

PROBLEMS
Plot the graphs of the following equations:

1. y = 2sin-- 2. 2/
~ 3 sin 2x.
3
3. y = 4 cos \x, 4. 2/
= 2 cos *

5. 2/
=» 3 sin (« — 1). 6. y = sin X + sin 2x.
1. y ^ sin^ X, 8 y
.
= tan X.
9. 2/
* CSC X, 10. y — cotx.
11. y * tan^ x, 12. ^ = X sin X.

18.
13. y = xsin-' 14. X® = sin y.
X
dy
Find -- in each of the following:
ax
16. y = 4 sin 3a; + 3 cos 4a;. 16. y = 3 8in*?-

17. y = 2 cos^ 3a;. y = |x — I sin 2x.

19. 2/
= ic — sin x cos a;. 20. 2/
= 2 cos |x + X sin Jx,
2j 35
21. y = sin 3a; — sin® a;. 22 y . = cos® ““ 3 cos ~ •

g
23. 2/
^ sin 6a; — J sin® 5a;.

24. y = a;(§ cos® a; — cos a;) + ^ sin^ X 4* f sin X.

2b. y ^2 tan 3a;. 26. 1/
— 3 cot (1 — 4x).

27. y = V tan x. 28. y =» tan x x. —

29. y *= 3a; 4* cot 3a;. 30. 2/ =
i tan® 2x.
31. y = sec® 3a; tan® — 3a!. 32. 2/ = sec^ X tan^ x. —
33. ^ = (sec X 4- tan x)®. 34. ^ = tan 2x 4- 1 tan® 2x.

36. y =“
I tan® ^ + tan® 3 36. y = sec® |x — 3 sec f x.
5 3
37. 2/ = i sec®—X f sec® x + sec x.
38. 2/ = i cot® X — J cot® X 4- cot X
39. xy 4- tan (a;2/) =* 0-

40. If 2/
^A cos nx + P sin nx, where A, n are constants, show that
(Py
4- n®y =• 0.
dx®

41. Find the angle at which the curve

2/
s f sin 2x
crosses the x-axis.
42. Find the angle between the x-axis and the curve y = cotx at each
intersection.
43. Find the angles at which the curves y = tan x, t/ = 2 sin x intersect.
44. Find the maximum height of the curve

y = 3 cos X — 4 sin X
above the x-axis.
172 TRIGONOMETRIC FUNCTIONS [Chap. VI

45 Find the maximum value of the function 2 sin a; + cos 2x.

.

46 By Newton’s method find an approximate solution of the equation

.

tan X ^ §x
near ^ •

^
47 Find the area bounded by the x-axis and one arch of the sine curve
.

y =A sin (bx + c).

48. A revolving light 5 miles from a straight shoreline makes 6 revolutions
per minute. Find the velocity along the shore of the beam of light at the
instant makes an angle of 60® with the shoreline.
when it

49 An airplane 10,000 feet above the earth is flying horizontally directly

.

away from an observer. If the angle of elevation is 60° and this angle is
diminishing 0.03 radian per second, determine the speed of the plane and the
rate at which its distance from the observer is increasing.
60. A right circular cone of vertical angle 26 is inscribed in a sphere of
radius a. For what value of 6 is the lateral surface area of the cone greatest?
61 The lower corner of a page is folded over so as just to reach the inner
.

edge. If the width of the page is 6 inches, find the minimum length of crease.
62 A rope with a ring at one end is looped over two pegs in a horizontal
.

line and held taut by a weight fastened to the free end. If the rope slips freely,
the weight will descend as far as possible. Find the angle formed at the bot-
tom of the loop.
63 . A gutter is made by bending into shape a strip of copper so that the
cross section is an arc of a circle. If the width of the strip is a, find the radius
of the circle when the carrying capacity is greatest.
64. A vertical telephone pole at a bend in the line is supported from tipping
over by a guy wire 40 feet long attached to a stake in the ground. At what
angle should the guy wire be inclined to make the tension in it least? Observe
that the moment (§134) of the tension about the foot of the pole is constant.

Plot the graphs of the following equations, indicating the principal value
by a solid line and dotting the remainder of the curve:
66 y .
=* cos ^ Jx. 66 .
1 O f on 1

67 . y =» cot~^x. 68 .

dy
Find -r- in each of the following problems:
ax

69 . 2/
B 3 sin ^ 4x. 60
60. .
2/
= 4 cos ^

61 y .
« cos~^ ^1
~ 62 . 1/ =» sin ^ (cosx).

63 y .
=» cos~^ -• 64 ^ .
.
= X cos ' X ~ Vl — x^.
X
66 ^ .
» X sin”^ X 4- Vl — x^. 66 . X cos ^ 2x — JVl — 4x2.

x2
Vt, y a sin”^ - -f- Va* — x^. 68 . y » sin
a x2 +2
 PROBLEMS 173

69. y = V 1 — + j 70. y = tan~^ ---•

2x . ^
.X
71. y =* tan
+ 2-

, a
73. y « cof
2 \z a)

76. y = tan ^ (tan 2x). 76. 2/

= tan ^

1 - x2

77. 2/
= Vx^ — 1 — tan'^Vx^ — 1.

78. 2/
= 3 sec*"^ (2x — 1). 79. 2/
= Vx^ — — a sec~
1
BO. y ^ sec~^ Vx. oi
81. 2/
= CSC
-1 -•
X
««
82. y =
V x2 - a2 1
CSC *
x
- •
= csc->l(x+l).
5 88. 2,
ar a a

84. 2/
= tan ^
86. 2,
= itan-i(ltan0.
3 + 5 cos X
Determine the following integrals:

86.
J Vl-f^^- 3? J xV^l
J"
sec**<ix. 89./sec X tan x <ix.

90. If 5 is the arc from the x-axis to the point (x, ?/) on the circle x^ + = a^,

show that
5 = a cos
.
*
X
- ,
—=
cfe a
a ax y
91. If A is the area bounded by the circle x^ + 2/^ = the y-axis, and the
vertical line through (x, 2/)i show that
.

A = X 2/ +
„
sin
. X
-
a
,
—=
dA
ax
^
22/.

92. By interpreting the integral as an area show that

f Va^ — x^ dx = - xy/a^ — x^ + ^2 sin~^ - •

A) 2 a

93. A tablet 7 feet high is placed on a wall with

its base 9 feet above the

level of an observer’s eye. How

from the wall should the observer standfar
so that the angle of vision subtended by the tablet may be a maximum?
94. What is the angular velocity in radians per second of the second hand
of a watch?
96. Assume that a 12-inch shell rotates once about its axis while moving
forward 40 feet. If its speed is 2400 feet per second, find its angular velocity.
96. A wheel of radius 2 feet rolls down an inclined plane, its center moving
the distance s = 5t^ in t seconds. Find the angular velocity of the wheel
about its axis.
174 TRIGONOMETRIC FUNCTIONS [Chap. VI

97 The flywheel of a watch turns through the angle

.

6 A sin ut
in t seconds. Find its angular velocity and acceleration at the instant when $
has its greatest value.
98 OA is a crank revolving with angular velocity « about 0. AB is a
.

connecting rod attached to a piston B which moves along a straight line through
O. Show that the speed of B is o)‘OC, where C is the point in which the line
BA intersects the line through 0 perpendicular to OB,
99. In a simple harmonic motion
8 = Asin (ht + c).
For what values of 6/ + c is the velocity greatest?
100. At time t the displacement of a particle from its equilibrium position is

8 = 6 sin 6^.

Find the amplitude and period of vibration.

101 . The displacement of a particle at time t is

s = 3 cos 2^ + 4 sin 2t.

Show that its motion is simple harmonic, and find the amplitude and fre-
quency.
102 . A particle moves along the a;-axis, its position at time t being

a; = 2 — 3 sin* 2t,

Show that the motion is simple harmonic, and find

its amplitude and period.

To do terms of the angle At,

this express sin* 2t in
103 A particle moves with simple harmonic motion along the x-axis, its
.

displacement from the origin at time t satisfying the equation

Find the period of vibration.

 CHAPTER VII

EXPONENTIAL AND LOGARITHMIC FUNCTIONS

83. Exponential and Logarithm. If a is positive and p, q are

-
integers, the notation a® is used to represent the positive root

When X is irrational a* is defined as the limit

a* = lim o'*
r -*x

as r tends to x through rational values.

If a is a positive constant, not equal to 1,

== a®
2/ (1)

is called an exponential function of x. When a is greater than 1, the

function increases steadily as x increases. When a is less than 1,
176 EXPONENTIAL AND LOGARITHMIC FUNCTIONS [Chap. VII

it decreases as x increases. In either case y takes each positive

value just once.
For positive values of y equation (1) therefore determines x
as a function of y. This function, written

X = loga 2/, (2)

is called the logarithm of y to the base a.

Forreal values of x and positive values of y equations (1) and
(2) are by definition equivalent. Elimination of x gives the im-
portant identity
a'*’*”*' = y (3)

which states that the logarithm of a number is the power to which

the base must be raised to equal the number.
It should be noted that logarithms are defined only for a positive
base a. If a is negative a® may not be real. Also logarithms of
only positive numbers are here considered. Later we shall extend
the definition to include negative numbers but will then find that
the logarithm of a negative number is not real.
1

84. Limit of (1 + h)\ In determining derivatives of logarithms

we shall need the limit approached by the function

(1 + h)^
when h tends to zero. In discussing this it will be convenient to
distinguish three cases depending on the range of values h is al-

lowed to take.
First suppose that

takes only positive integral values. Then, by use of the binomial

theorem.

(1 + h)i = (i
+ ^) (1)

<{an be expressed as a sum of n + 1 terms

!+«•- +
1

n
—-
n(ra
r:
2!
1)

n*
1
5 + »(»
- l)(n
r:
3!
- 2) 1
j
n®
+
’
•
+—
'
n"

= 1 +1+ 2
( )
§ 84 ] LIMIT OF (I +h)^ 177

Each term in this series is positive and not greater than the cor-
responding term in the series

1 1 1
e = 1 + 1 (3)
2! 3! n!

the convergence of which is discussed in Example 1, §145. Further-

more each term of (2) tends to the corresponding term of (3) as
limit. By taking n large enough we can thus make the first m+ I

terms of (2) as nearly equal to the corresponding part of (3) as

we please. For any fixed integer we shall therefore have m
1 1 1
e >

when n is sufficiently large. Since m may be as large as you please

this requires

lim
{
n-oo\
1 H —n/^ = e, (4)

Next suppose that h takes all positive values. For each value
h < I there is a positive integer n such that

1
n < - ^n + 1,
h
whence
iv»+l

This can be written

/ 1

\ n + 1/ i
/ 1\ / IV

‘+;rTT
When h tends to zero, n tends to infinity and the first and third
terms of this last inequality tend to e as limit. Since

(1 + h)^
lies between the two, it must have the same limit.
178 EXPONENTIAL AND IXIGARITHMIC FUNCTIONS [Chap. VII

Finally, suppose that

h= -k
is negative. Then
1 _1 / 1 V
which is equivalent to

k
( + r^)* + r^) (‘
+ r^)
When h tends to zero

1 -k
tends to zero and the last expression again tends to e as limit.
Thus whether h varies through rational or irrational, positive
or negative, values

lim
k -*o
(1 + A)* = c. ( 5)

By summing the series (3) the value of this constant to four places
of decimals is found to be
e = 2.7183. (6)

86. Differentiation of Logarithms. Let

y = logo M. (1)

When u takes the increment Au the corresponding increment in ^ is

(
—+ —
u Am\
1 = loga
/
(1
Am\
+ —y
Thus

Ay
logo
(•3
Au Au
If we write
Au
h,

this becomes
iky logo (1 + ^) 1 1
§86] DIFFERENTIATION OF EXPONENTIALS 179

As Au tends to zero, h tends to zero and

(1 +
tends to the limit e [§84, (5)]. Thus (2) gives as limit

^ _ log„e ’ (3)
du u
whence
logo e
dy = d(loga u) = du. (4)
u
Equation (4) is valid if a is any positive constant different from 1.

In particular, if a = e,
loga e = loge e = 1
and

d(loge u) = —
du
u
(6)

Because of the greater simplicity of this formula logarithms to

basee are called natural logarithms and are used almost exclusively
in theoretical work.
The natural logarithm is indicated by the symbol In. That is,

In M = log, u, (6)

and equation (5) is usually written

d In « = —
du
u
(7)

86. Differentiation of Exponentials. To derive a formula for

the differential of
y = o“ (1)

write the equation in the inverse form

whence

dy = ydu = In a-ydu.
logaC
180 EXPONENTIAL AND LOGARITHMIC FUNCTIONS [Chap. VII

Replacing y by a“, we obtain

d(a“) = a“ In a du, (3)

In particular, if a = c, In a = In e = 1 and (3) becomes

d{e^) = du. (4)

87. Differentials of Exponential and Logarithmic Ftmctions. In

the preceding sections we have proved the following formulas:

loga e
XIX. d\ogaU = -^du.
u
XX. dlnu = —
du
u
XXL d(a“) = a^lnadu.
XXII. d(6“) = 6^ du.

Sometimes the calculation of a derivative is simplified if a

logarithmis replaced by an equivalent form by means of the

following formulas:
In (uv) = In w + In v. (1)

u
In - = In w — In v. (2)
V

In = n In u. (3)

1 1
In w” = - In u. (4)
n
Example 1. y = ln(x2 + 1).

By formula XX,

2X
dx x^+l X* + 1

b - 1
Example 2. y =
v - In
--*
^
By equations (4) and (2) this can be written

y = |[ln(z - 1) - ln(a: + 1)],

whence
^^ 1 r_i ^1 ^ 1

dx 2 La; — 1 x + lJ x* — 1
§89] INTEGRATION 181

Example 3. = e“” *
j/

By formula XXII
* sec* X.
^ ^
88. Logarithmic Differentiation. In some cases the derivative
of a function
y = f(x)
can be found more easily by taking the natural logarithm of both
dv
sides, differentiating, and then solving for
dx
This process is — .

called logarithmic differentiation.

Example, If u is positive and u and v are functions of x, find the

derivative of u^.
Writing ^
^

and taking logarithms of both sides, we have

In y = t; In w.
whence by differentiation

1 dy _ dv V du
y dx dx u dx
Thus
dy r, dv V du~\
~ .

dx
Replacing y by this becomes
d du
— /
(u'^) = u'"
,
\n u
dv
-r- + ^

vu'’
.
— (1)
dx dx dx

The function u° is two ways, namely through variation of u

variable in
and through variation alone were variable the derivative would
of i;. If t;

be the first term in the above equations. If u alone were variable, the
derivative would be the second term. When the two vary simultaneously
the derivative is thus the sum of the results obtained by varying them one
at a time.

89. Integration. Since an indefinite integral is merely a function

with given differential, any differentiation formula gives by in-
version a formula of integration. Thus from

d In w = —
du
u
we obtain
du
182 EXPONENTIAL AND LOGARITHMIC FUNCTIONS [Chap. VII

Each formula obtained in this way contains a single variable u

and its differential. To integrate an expression which contains
more than one variable it must be expressed as a sum of terms
each of which contains only one variable and its differential.
Reduction of an expression to such a form is called separation of
the variables. If the variables cannot be thus separated, the expres-
sion cannot be integratedby our present methods.
In case of an equation, multiplication or division by a proper
factor may separate the variables. In the equation

X dy — y dx ^ 0,

for example, the variables x and y are not separated, but division
by xy gives
dy dx
^
y ^
in which they are.

Example 1. Find the value of

dx

Taking
A 2x +l
2x + I u
we have
dx ^ ^ du
and

Example 2. Find the curves such that the part of the tangent included
between the coordinate axes is bisected at the point of tangency.
In Figure 114 let AB be a tan-
gent to the curve and P(x, y) its
point of tangency. Since P is the
middle point of AB,

OA = 2y, OB = 2x.

The slope of the curve at P is

% OA t.
dx OB X

This can be written

^ ^I
0 .
§89] INTEGRATION 183

Since each term contains a single variable, we can integrate and so get

In !/ + In x = C.
This is equivalent to
\iLxy = C.
Hence
xy — — k.

Cy and consequently ky can have any value. The curves are rectangular
hyperbolas with the coordinate axes as asymptotes.
Example S. A tank with vertical axis and full of water
cylindrical
has a leak at the bottom. Assuming that the water escapes at a rate pro-
portional to the pressure at the leak and that 10% escapes during the first
hour, how long will it take to half empty?
Let a be the radius of the tank, h its altitude, and x the depth of water
at the end of t hours. The volume of water at time t is then ira^x, and its
rate of outflow

This isassumed to be proportional to the pressure at the bottom and so

to x. That is

Oft

where k is constant. Separating the variables, we have

Tra* ^ — kdt,
X
Integration gives
In X — kt + C.
When t = 0 the tank is full and x = h. Hence

Tra^ In = C.

Subtracting this from the preceding equation, we get

Ta^ In f = kU
ft

When t — 1, X = Consequently

ira^ In = A;.

WThen the tank is half empty, x = ^ h and

184 EXPONENTIAL AND LOGARITHMIC FUNCTIONS [Chap. VII

PROBLEMS
1. By summing the series

e-l + l+^ + |+--- + i+-

3.

determine the value of e correct to two decimal places.

2. By use of a logarithm table find the value of

(1 +
and compare the result with 6.

By writing

(1 +h = )'^
[(1 +
determine

lim (1 H- h)^.

Plot the graphs of the following equations:

4. y =» logic X. 6. y — In X.
1

6. y = e*. 7. y = c».
1 — a:
8. y « e* sin a;. 9. 2^
= In
l-hx
dy
Find ~ in each of the following:
dx
10. y - In (3x + 2). 11. y = In (3a;^ + 6a; -f 1).
12. y = In (** - 4a: + 4). 13. 7/ = a; In 05 — 35.

16. 1 h
Zx
. X
16. y = In sin x. 17. y = In tan-*
Jd

18. y = In (sec 2a; + tan 2a;). 19. y = In (a; + Va;^ — ).

1 — cos 2a;
20. y — In sin cos® 21. y j
^ ^ 1 + cos 2x
22. y » (In x)®. 23. y logic (x + 6).
24. = i(e* + e-*). 26. y

Binx
27. y = e

28. » - a:* + e*. 29. V = 10*.

90. y = a:"n®. 31. y -{3x + l)e-»».

32. » = (2ar* - 2a: + l)e**. 33. V “ 4 . e-tz

..
«* - e"* «“•* cos 2x.
“•J'-lV + c-*-
,
36. y =
86. y =» ^(sin 2a; — 2 cos 2a;).

Zl. y ^ X sec""^ x -• In (a; + Va5® — T).

 PROBLEMS 185

38. y ^ z tan ^ — In (a;* + 4). 39. t/ » In (In ax).

40. 2x +y = In (2z — y), 41. e* + e*' = 1.

42. In (xy) ^ x + y. 43. tan y « e*.

44. e*-" = In-- 46. 2/

=» x“®.

46. y = (tana;)* 47. y =

\a;2 + a2
48. y =
a;V 2a; +3 49. y =
a;^ V a;^ — 4

\/3a; +2 Vx^TT
60. By logarithmic differentiation prove the formula for the derivative of
x^ when n is irrational.

61.
62. Find the value of

*
lim (1
-0 + mx)*,
by taking the logarithm of

(1 4“ rnx)^

and finding the limit by P Hospital’s rule.

If 1/ =5 Ae’** + where A, B, n are constants, show that
66.

S-nV-O.
63. If y = ze where 2 is a function of x, show that

dpy dy „ cPz

64. Find the slope of the curve y = In (3x + 2) at the point where it crosses
the X- axis.

If 2 =* a*, find

dx^

66. If 2/
= show that

—
dy
sz m
du
u
.

\- n —
dv
V
y

67. A box has a square base. By applying the formula of the preceding
problem find the percentage change in the volume when the side of base is
increased 2% and the altitude decreased 3%.
68. Find the maximum value of the function x®e”*.
69. On the part of the curve
y » c* sin x

between x — ir and x « 2ir find the point at greatest distance from the
x-axis.
 186 EXPONENTIAL AND LOGARITHMIC FUNCTIONS [Chap. VII

60. Find the least value of the function a;* for positive values of «.
61. If 0 < a; < 1 and

*-*ln*- — + 2*
-,
show that

dz dz
From this and the fact that z and -- are zero at a? = 1 show that --- is negative
dx dx
and
62.z is positive. In this way show that
x^ 1
x\nx>---
for values of x in the interval 0 < a? < 1.

63.
By a method similar to that followed in the last problem, show that

In sec a; > •;-

A

TT ir
for values of x in the interval
2
<*< 2
-

A particle moves along the a^axis, its abscissa at time t being

X =» oc** +
a, by k being constants. Show that it is repelled from the origin with a force
proportional to the distance.

Evaluate the following integrals:

“A dx
+3
66.
/:x^
x dx
4~i

66 . J^^dx. 67. dx.

68. Find the area bounded by the ®-axis, the hyperbola xy = 4, and the
ordinates at x ** 1, x =» 4.

69. The area bounded by the y-axis, the curve x^y = 4, and the lines y » 1,
y » 4 is rotated about the y-axis. Find the volume generated.
70. When bacteria grow without restraint they increase at a rate propor-
tional to thenumber present. Express that number as a function of the time.
71. The
portion of the tangent to a curve between the x-axis and the point
of tangency is bisected by the y-axis. If the curve passes through (1, 2), find
its equation.
72. A way that its angular acceleration is
rotating wheel slows in such a
proportional to angular velocity. If the angular velocity is initially 200 rev-
its

olutions per minute and in 2 minutes this slows to 50 revolutions per minute,
when will it become 25 revolutions per minute?
 CHAPTER VIII

PARAMETRIC EQUATIONS

90. Parameter. Sometimes the coordinates x, y of the variable

point on a curve are expressed in terms of a third variable instead
of in terms of each other. Such a third variable is called a para-
meter and the equations expressing x, y in terms of the parameter
are called parametric equations of the curve.
Thus, if t is the parameter and/i(0, /2(0 given functions, the
locus of points
x=m), y=h{t) (1)

is a definite curve. To construct the curve we assign values to t,

calculate the corresponding values of x and y, and plot the resulting

points.
To obtain the Cartesian equation of the curve we eliminate the
parameter. Suppose, for example, that by eliminating t from equa-
tions (1) we get
fix, y) = 0. (2)

This equation, being a consequence of the parametric equations,

is satisfied by the coordinates of any point on the curve. If, con-
versely, for every pair of values Xy y satisfying (2) a value t can be
found such that x = fi{t), y = /2(0> then (2) is the Cartesian
equation of the curve.

Example 1 . Construct the graph represented by

y = <
+ j,
and find its Cartesian equation.
To construct the curve imagine t starting from — oo to increase steadily
to + When t is negative and very large, x and y are nearly equal, y
00 .

being slightly larger. The point (a;, y) is then below and near the line
y ^ X. Ab t increases to —1, (x, y) moves to A (Figure 115). As t con-
tinues to increase from —1 to zero, x becomes positive and (x, y) moves
away, approaching the asymptote y = — x. As t passes through zero,
the point (x, y) reappears in the second quadrant, crosses the x-axis at
187
188 PARAMETRIC EQUATIONS [Chap. VIII

< = 1, and then moves away, approaching the asymptote y = x as t

tends to infinity.
By squaring and subtracting we obtain

— X* = 4

• ^ the Cartesian equation. The curve is thus a rectangular hyperbola.

Figure 115.

Example 2. x = cos (2<^), y = sin 0.

TT TT
As <i>
varies from to - the point (x, y) traces the arc ABC (Fig-
2 2
ure 116). As continues to advance (x, y) oscillates back and forth along
this path. The arc ABC is therefore the entire locus.

Figure 116.

Since « = cos 2^ = 1 — 2 sin* ^ and y - sin elimination of gives

x = 1 — 2y*.

The locus of this equation is not, however, entirely equivalent to that of

§91] THE CYCLOID 189

the parametric equations, for the Cartesian locus is an infinite parabola,

whereas the parametric locus is only the arc ABC of that parabola.
Example 3, Determine parametric equations for the parabola

= 4p?/,

using as parameter the slope of the line from the origin to P(x, y).
The parameter to be used is

X
Solving this and the equation of the curve for x and y, we find

X = 4p^, y = 4p^2

as the parametric equations required.

91. The Cycloid. When a circle rolls along a straight line a

point of its circumference describes a curve called a cycloid.
Let a circle of radius a and center C roll along the x-axis; let N
be its point of contact with the x-axis and P(x, y) a point of its

circumference tracing a cycloid. Take 0 found

as origin thfe point
by rolling the circle to the left until P meets OX, and take as
parameter the angle
^ = NCP.
</> (1)

Since the point of contact moves the same distance along the circle
and along the straight line,

ON = arc NP = (2)
a<t>.

The coordinates x, y of P are the horizontal and vertical displace-

ments in moving from 0 to P, and these are the sums of correspond-
ing displacements from 0 to C and from C to P. Thus

X = — a sin y == a — a cos (3 )

are parametric equations of the cycloid described by P.

190 PARAMETRIC EQUATIONS [Chap. VIII

92. The Epicycloid. When a circle rolls on the outside of a

fixed circle, a point of its circumference describes a curve called
an epicycloid.
Let a circle of radius a and center C roll on the outside of a fixed
circle of radius 6 and center 0; let iV be the point of contact of the
two circles, P(a;, y) a point describing an epicycloid, and A the

point obtained by rolling

the circle backward until P
meets the fixed circle. Take
the point 0 as origin, OA
as x-axis, and take as para-
meter the angle AOC =
Since the point of con-
tact moves the same dis-
tance along both circles,

arc AN = arc NP.

If B m the angle OCP we
then have

Figure 118. oB = b(t>, (1)

and so the angle between PC and the a:-axis is

a +&
( 2)
a

The coordinates of P are the horizontal and vertical displace-

ments in moving from 0 to P. These displacements are the sums
of corresponding displacements from 0 to C and from C to P.
Since the displacements from C to P are the negatives of those
from P to C, we thus have
X — (a + b) cos ^ — a cos

/a + b \
y = (o + 6) sin 0 — a sin I <#> 1

as parametric equations of the epicycloid.

93. Length of Arc. A parameter frequently used is the length
of arc along a curve from a fixed to a variable point.
To define what we mean by the length of an arc AB divide it

into parts by intermediate points Pi, P2 ,

• •
•, Pn-i and join con-
§93] LENGTH OF ARC 191

secutive points by straight lines. The length of the resulting

polygonal line from A to B is
L = APi + P P2 +
1
• •
•
+ Pn-iB. (1)

Now increase the number a way that

of points of division in such
the distances between consecutive points tend to zero as n tends
all

to infinity. If under these conditions L tends to a limit independ-

ent of the choice of intermediate points, the arc AS is said to be
rectifiable and the limit is called its length.
Let y = /(x) be the equation of the curve and assume
dy
( 2)
ax
continuous on the arc AB, If P(x, y), Q{x Ax, y Ay) are con-+ +
secutive vertices of the polygonal line, the chord between these
points is

PQ = V(Aa:)2 + {Ayf = + Ax.

Since the slope is continuous, by the mean value theorem there is

some number { between x and x + Ax such that
Ay
(3 )
Ax
Thus
PQ = vTTl/W Ax.
192 PARAMETRIC EQUATIONS [Chap. VIII

If a, b are the values of x at A, B and b > a, the total length of

the polygonal line is

+ (4)
a

When the number of divisions is increased and the increments Az

tend to zero this sum tends to

+ U'ix)fdx
dy
as limit. Thus, if the slope is continuous at all points of an arcj
ax
the arc is rectifiable and its length from x = a to x = b is

(5)

the length L of the polygonal line, and so its

An arc being given,
limit s, is independent of the coordinates used. The integral (5)
therefore depends only on the arc AB and not on the coordinate
axes. If there is any set of coordinate axes with respect to which
the slope is continuous, the arc is thus rectifiable and its length is

given by when
(5) referred to those axes. More generally, if an
arc is the sum of a finite number of parts each of which is rectifi-

able, the whole arc is rectifiable and its length is the sum of the
lengths of the parts.
94. Ratio of Arc and Chord. We are often concerned with the
ratio of a small arc to the chord by which it is subtended. If the
slope is continuous this ratio tends to the limit 1 when the chord
tends to zero.
Since a sufficiently small arc with continuous slope is approx-
imately straight, this may be taken as intuitively obvious. To
prove it analytically let P(x, y), Q{x Ax, y +
J^y) be the ends +
of = f{x) the curve on which it is located. By the
an arc and y
mean value theorems (§§34 and 50) there are values fi, {2 between
X and X Ax +
such that

chord PQ = Vl +
(2)
^ “ Vl + Ax,

a«5^*0=jf Vl + [/'(x)]*da: = Vl +
§951 DIFFERENTIAL OF ARC 193

Thus
^ VI + [/(fe)]^
arcPQ
chord PQ Vl + [/'({i)]*
When Aa; tends to zero, fi and ^2 tend to x, and this gives

arc PQ
lim = 1.
Q p chord PQ
-

96. Differential of Arc. Let Po(^o, Vo) be a fixed point, P{z, y)

a variable point, on a given curve, and

s = PoP (1)

the arc from Pq to P, this arc being considered positive for points
of the curve on one side of Pq, negative for those on the other.
If the slope is continuous at all points of the arc, by §93 (5)

the algebraic sign depending on the direction along the curve

which has been taken as positive. This is a function of x with
differential

ds = dz 4” ^ dx^ “h dy^, (3)

Thus ds is the hypotenuse of a right triangle with sides dx and

dyy as indicated in Figure 120.
From this diagram it is also
seen that

dx dy
— = cos </>,
— = sin (4)
ds

<t> being the angle from the

positive direction of the x-axis
to the tangent drawn in the
direction in which s increases.
When
dx and dy are sufficiently small ds is approximately the
arc between the points P(x, y), Q{x AXy y + Ay) on the curve.+
By treating this small arc as straight the above relations can be
194 PARAMETRIC EQUATIONS IChap. VIII

read from a diagram of the type shown in Figure 121. Since the
ratios of differentials are the limiting ratios of increments it is clear
that on a curve with continuous slope this picture, although slightly
inexact, will always suggest correct relations.

Example. Find the arc of the curve

X — cos 6 + 6 &in 6, y = sin$ — 6 cos 0

from ^ = 0 to .

2
By differentiation we find

dx = 6 cos 0 ddf dy — 0 sin 0 dJB.

Therefore

de — y/dx^ + dy^ = 0 dB.

Thus the required arc is

96. Curvature. On a given curve the angle 4> froni the x-axis to
the tangent at the variable point P{x, y) may be regarded as a
fimction of the arc s from a fixed point Fq of the curve to P. As
we move along the curve the change in measures the change in <t>

the direction of the tangent, and the derivative of 4> with respect
to 8 measures the change in direction per imit distance moved.
This derivative
d<l>

(1)
de
is called the curvature at P.
§961 CURVATURE 195

At a point where the curvature is large the change in direction

per unit distanceis large and the curve sharply bent. At a point
where the curvature is small the change in direction and bending
are more gradual.

To express the curvature in terms of x and y we use the equations

dx \dx/
whence

curvature = —=
d<tt
2
( )

The radical in the denominator is usually taken positive. This

amounts to taking the curvature positive when the curve is concave
upward, negative when con-
cave downward.
In a circle of radius r (Fig-
ure 123)

0 = 0 +^, I

and so the curvature is

d(t> do

ds ds Figure 123.
196 PARAMETRIC EQUATIONS [Chap. VIH

Thus the reciprocal of curvature

ds

d(t>

is equal to the radius of the circle.

Similarly, the reciprocal of curvature

dx^

on any curve is called its radius of curvature. At a given point

this is the radius of the circle which passes through the point and
which has at that point the same slope and second derivative as
the given curve. It is therefore the radius of the circle that most
closely fits the curve in the immediate neighborhood of the point.
In the determination of curvature it is clear that the angle
instead of being measured from the a;-axis, might be measured from
any fixed direction. If we measure it from the y-axis, the coordi-
nates X, y will be interchanged in the above discussion, giving

dy^

as the radius of curvature. If the radicals are taken positive in

both cases the expressions (3) and (4) may, however, have different
algebraic signs.

Example 1, Find the radius of curvature of the parabola

= Apx
at the origin.
Since

dx

is not continuous at the origin the conditions assumed in deriving equa-

 PROBLEMS 197

tion (3) are not satisfied. If, however, we measure slope and angle from
the y-axis, we have
dx _ 2/ _ cPx __ 1
'
dy 2p *
dy^ 2p

as the radius of curvature at the origin.

Example 2. The curve formed by a flexible chain when held by its
ends and hanging under its own weight is called a catenary. With suit-
able choice of axes its equation is

a - --
2/ = 2 ®

where a is constant. Find its radius of curvature.

By differentiation and substitution we find

The radius of curvature of the catenary is therefore

= (ea +e «)* = ^.
4 a

PROBLEMS
Plot the loci represented by the following pairs of parametric equations,
and determine the corresponding Cartesian equations:
1. a; = 2 + 32, y = 1 - 22.

2. X = 2^, 2/ = 2 + 2.
3. X =* 2^ -t- 22, 22
= - 22.

^ „ 1 - 22 22

. »!+'’ 21
108 PARAMETRIC EQUATIONS [Chap. VIII

* “ +
1 “ i(i + <*)'
7. * - «*(« - 1), y - + 1).
8. * •« an 0, y « sm +0 •

9. « « sec 0, 3/
« tan <l>,

10. X ^ acoa<l>y y ^ h sin 0.

11. x » sin 3/
s sin 2^.

12. X ^ cos y ==> sec

tf

13. a; *= 2 4“ 3 sin 3/ = 3 — 2 sin t

14. a; »® e* sin =* 6* cos
3/

16. Are a: = < 4-~, /=»f—

3 and a: = e* + e~~*, y e* — c S parametric
t t

equations of the same locus?

16. Plot two turns of the curve

X = cos 0, y =»

Find the intersections of the following pairs of curves:

r rX = 2 -h rX = 1 +2 «,

“ x = 5co8fl,
17.
[y =
I* f
4<, [y = 5sm 9. J3
[y = 2 +j
1

[
2/
“ 1 -

19. A circle of radius a has its center at the origin. Find parametric equa-
tions of the circle, using as parameter the angle <f> from the x-axis to the radius
through P(x, y),
20. Find parametric equations of the ellipse

1,
62

using the parameter defined by the equation x = o cos 0.

<t>

21. Find parametric equations of the hyperbola

f! -
“ ’
o* 6*

using the parameter 4> defined by the equation x => a sec 0.

22. Find parametric equations of the curve

xi +yi ^ a§,

using the parameter defined by the equation x « a cos^

23. Find parametric equations of the curve

X* 4- ^ « 0:3/,

using as parameter the slope of the line through the origin and the point (x, y),

24. Find parametric equations of the curve

2a -x’
u^g as parameter the slope of the line through the origin and the point (x, y).
 PROBLEMS 199

Find parametric equations of the following curves, using as parameter the

slope m of the curve at the point (x, y)
25. = 4pa;. 26. + 2/i ai.

1. 28. —-— 1.
62 a2 62

29. Find the slope of the curve

80.
1
V =
(1 -
at the point (a;, p).
Find the angle from the a;-axis to the line tangent to the curve

X — a cos t(l — cos 0 » y ~ o> sin ^(1 + cos 0

at the point (x, y).

81. A rod AB moves with its end A in the a;-axis and B in the p-axis. Using
as parameter the angle between the rod and the x-axis, find parametric equa-
tions of the locus described by the point P(x, y) on the rod, if a and 6 are its
distances from A
and B. Also find the Cartesian equation of the locus.
82. A string held taut is unwound from a circle. Taking the origin at the
center of the circle, the x-axis through the point where the string begins to
unwind, and using as parameter the angle from the positive end of the x-axis
to the radius through the point of tangency, determine parametric equations
of the locus described by the end of the string. The locus is called an involute
of the circle.
86.
88. When a wheel rolls along a straight line the curve described by any
point of a spoke is called a trochoid. Let the wheel roll along the x-axis and
use as parameter the angle 4> in the equation of the cycloid (§91). Find para-
metric equations of the trochoid described by the point (x, y) at distance 6
from the center of the wheel.
84. A hypocydoid is the locus described by a point on the circumference
of a circle which rolls internally along the circumference of a fixed circle. Find
parametric equations of the hypocydoid when the radius of the moving circle
is one-fourth that of the fixed circle, using a parameter analogous to that in

the equations of the epicycloid. Also find the Cartesian equation of the curve.
OA is diameter of a circle of radius a, A is tangent, and A
intersects ON
the circle at M. Lines through M
and Nj parallel respectively to AiV^ and OA,
intersect in P. Taking 0 as origin, OA as y-axis, and the angle as para- AON
meter, find parametric equations for the locus of P. Also find the Cartesian
equation. The curve is called a mtch.
86.Using the equations of the preceding problem, determine the area
bounded by the witch and the x-axis.
87. OA is diameter of a circle of radius a, AiNT is tangent, and ON intersects
the circle at Af. On ON the point P is found such that OP = MN. Taking
0 as origin, OA as x-axis, and the angle AON as parameter, find parametric
equations of the locus described by P. Also find the Cartesian equation.
The curve is called a ciesoid.
Through the point A (—a, 0) on the x-axis a line is drawn intersecting
88.
the y-axis at B, and on this line two points P(x, y) are taken at distance OB
200 PARAMETRIC EQUATIONS [Chap. VIII

from B. Find parametric equations for the locus of P, using as parameter the
angle from the o^axis to the line AB. Also find the Cartesian equation of the
curve. The locus is called a sirophoid,
39.The angles of a triangle are A, B, C and the opposite sides o, 6, c. If
the vertex A moves along the x-axis and B along the y-axis, find the locus of
C, using the angle from the a;-axis to A B as parameter. If C is a right angle,
show that the is a straight line.
locus
40. Show that the normal to the cycloid at the point P (Figure 117) passes
through the point N
where the rolling circle touches the fixed line, and that
the tangent intersects the circle at the upper end of the diameter through N,
41. Show that the normal to the epicycloid at the point P (Figure 118)
intersects the fixed circle at and that the tangent intersects the rolling
circle at the end of the diameter through N,

Find the arc lengths of the following curves between the points indicated:

42. V - Mx - 1)3, (1, 0) to (2, f).

48. a? « 4t^, y = 0 to = 7. 2/

44. a;« a cos y =* a sin S * 0 to 0 = 27r.

45. X ^ a cos3 y => a sin®

<f>, ^ = 0 to
A
46. a;« sin y cos
ty « 0 to — x. <

47. X = Vt* + 1, y = In (< + Vi* + D, =0 i to i = 5.

X X
48. a; » a sec <, y =* a In (sec t — tan 0, i to ^
4 4
49. a; 2 sin® t, y = 2t — sin 2i, t
» 0 to t ir.

60. The length of arc on a certain curve from a fixed point to the variable
point (Xf 2/) is s * X® — Find the slope of the curve at x = 2.
61. If 8 is arc length on a cycloid, show that

ds® = 2o®(l — cos <t>) d4? = 4a® sm®

^ d^®.
Find the length of one arch of the cycloid.
62. If 8 is arc length on the epicycloid, show that

ds® « 2(a + 6)® f\ 1 — cos -

a /
d^® = 4(a + 6)® sin® ^
2o
d0®.

Find the length of one arch of the epicycloid.

68. If the arc 8 of the catenary is measured from the lowest point, show that

^ i
dx a

64. Show that the area bounded by the a;-axis, an arc of the catenary, and
two ordinates is proportional to the length of arc.

Find the radius of curvature on each of the following curves:

56. ^ » In sec x.
66 . « - iy® —i In y.
67. X — a(coB 6 -f ^ an 6), y — a(sin 9 — S cos e).
 PROBLEMS 201

68. X =» a cos* y ^ a sin* L

69. X «* 2 sin* y =* 2i — sin 2t,
60. X ^ asectf y ^ a In (sec t — tan t).

61. Find the radius of curvature of

- (* - 1)»
at (2, 1).
62. Find the radius of curvature of the ellipse

at the points where it crosses the a>-axis.

63. Find the radius of curvature of

y = sin 2x
at the origin.
64. Find the greatest and the least radii of curvature on the curve 2/
~ sin x.
66. Find the radius of curvature of the epicycloid.
66. On an arch of the cycloid show that
-I- - 16a2,

s being measured from the highest point of the arch.

 CHAPTER IX
POLAR COORDINATES

97. Polar Coordinates. The most natural way to indicate the

position of a point P in a plane is to give the distance and direction
of P from a fixed point of the plane. Let 0.be the fixed point, called
the origin, or poZe, and OX a fixed line, called the initial Zinc, or
axis. The distance r ^ OP and the angle B from OX to OP are
called the polar coordinates of P.
The angle B is any angle from OX to the line OP, this angle
being considered positive when measured in the counterclockwise
direction, negative when in the clockwise direction.
The radius r is considered positive when OP is the terminal side
of B (Figure 124, 1 and 3), and negative when B terminates on OP
produced (Figure 124, 2).

A given pair of polar coordinates r, B determines a definite point.

But, since B may be positive or negative and may wind any number
of times around the origin, the same point is represented by in-
finitely many pairs of polar coordinates. It is mainly because of
this multiple representation that polar coordinates are a little less
simple than rectangular.
The point with polar coordinates r, B is represented by the nota-
tion (r, B). To indicate that the point P has the polar coordinates
r, 6 the notation P(r, B) is used.

Example. Plot the point P(— 1, — i^r) and find its other pairs of polar
codrdinates.
The point is shown in Figure 125. Since r = — 1, the angle B = — Jir
terminates on OP produced. The angle XOP is fir. Any other angle
202
 GRAPHS 203

terminating on OP or OP produced differs from one of these by a multiple

of 27r. Any such angle has one of the forms 2mr — Jir or 2n7r Itt, +
where n is a positive or negative in-
teger. Consequently any polar co-
ordinates of P have one of the forms

— 1, 2w7r — +
^ 0 , (1, 2n7r fir).

98. Graphs. The locus of

points with polar coordinates
which satisfy a given equation

r^m
is a curve. To plot the locus make a table of pairs of values r, 6
satisfying the equation, plot the corresponding points, and draw
a smooth curve through them.
It may be convenient to look for any of the things mentioned
in connection with plotting in rectangular coordinates (§69). It is
usually sufficient, however, to imagine d to increase from some
definite value and determine at each point merely whether r is
increasing or decreasing,
draw a curve on which r
varies in the proper direction,
and mark accurately the
points where r a maximum,is

a minimum, or zero. Proceed

in a similar way with 6 de-
creasing from the initial

value.
The work is facilitated by
using paper ruled as in Fig-
ure 126. In the following
diagrams parts of the curves
Figure 126. where r is negative are dotted.

Example 1. r = a(l + cos B).

The curve calleda cardioid is shown in Figure 126. As 0 increases from
0 to TT, r steadily diminishes from 2a to 0. As 6 continues to increase, the
cosine, and so r, retraces its values in inverse order, again reaching the
maximum, cos ^ = 1, r = 2a, at ^ = 27r. Because of the periodicity of
cos By negative values of 6 and values greater than 2ir give points already
plotted.
 Fioxtbe 128.

increases. When the angle makes a complete turn about the origin, r
is increased by the amount 2ira. This part of the curve thus consists of a
§99] INTERSEC3TI0N OF CURVES 205

set of expanding coils each at distance 2Tra outside the preceding. Nega-
tive values of 6 give a similar set of coils crossing the former on the vertical
axis.
Example 4. = 2a^ cos 26, To each value of 6 correspond two values

r = dba\/2 cos 26
differing only in algebraic sign. The curve is therefore symmetrical with
respect to the origin. Also angles differing only in sign determine the same

value of r. Hence the curve is symmetrical with respect to the initial

line. As 0 varies from 0 to r varies from dbo\/2 to 0. When 0 is

between dblir and ±f7r, cos 26 is negative and r imaginary. Other values
of 6 give points already plotted. The curve, called a Umniscate^ is shown
in Figure 129.

99. Intersection of Curves. If the polar coordinates of a point

satisfy the equation of a curve, the point lies on the curve. A
point may, however, be on a curve al-
though its coordinates (as given) do not
satisfy the equation. This happens be-
cause a point has many pairs of polar
coordinates. One of these pairs may
satisfy a given equation while another

does not. Thus the point B\

ison the curve sin 0 (Figure 130)

although the coordinates given do not
satisfy the equation, for the coordinates

—a, - represent the same point and do

Ji

satisfy the equation. FiGxmE 130.

206 POLAR COORDINATES [Chap. IX

To find the intersections of two curves we solve their equations

simultaneously. The pairs of coordinates thus obtained represent
points on both curves. There may, however, be other intersections.
This happens when some of the pairs of polar coordinates repre-
senting a point satisfy one equation, other pairs satisfy the other
equation, but no pair satisfies both. In finding intersections of
loci represented by polar equations, the graphs should therefore
be constructed. Any extra intersections will then be seen.

Example. Find the intersections of the curves

sin r = a.

Solving simultaneously we get sin 0 = 1, ^ ~ . One. point of inter-

im

section

is also
is

an
therefore

intersection.
A
( 4 - From Figure 130 it is seen that B
(-i)
100. Change of Coordinates. The same point may be repre-
sented by rectangular or by polar coordinates. It is sometimes

desirable to use the two systems simultaneously. The as-axis is

then usually coincident with the initial line, and the origin of
rectangular cobrdinates is the pole. A point P (Figure 131) then
has fom coordinates x, y, r, 0. These are connected by the equa-
tions
X => r cos B, y — r sin ,

r » it V ** 0 = tan“^ -•
X

By means of these equations (or, better, by direct use of a

diagram) any expression in rectangular coordinates can be trans-
§1011 EQUATIONS OF CERTAIN CURVES 207

fonned into a corresponding expression in polar coordinates, and

conversely.

Example, By changing to rectangular coordinates determine the locus

represented by the equation

r(2 cos 0 + 3 sin 0) = 4.

Since
r cos d = a;, r sin 0 = 2/

the corresponding equation in rectangular coordinates is

2x + ^y = 4.

The locus is therefore a straight line.

101. Equations of Certain Curves. By transformation of coordi-

nates, polar equations can be obtained for the curves we have rep-
resented in rectangular coordinates. In this section we determine
directly some of these equations.

(1) Polar Equations of a Straight Line, Let LK (Figure 132) be

the straight line and ON the perpendicular upon it from the origin;
let

ON =XON = a. p,

If P(r, 6) is any point on the line, ON is the base of a right triangle

with hypotenuse OP = r and base angle NOP = 6 — a. Hence

r cos (0 — a) == p (1)

is the equation required.

(2) Polar Equation of a Circle, If the circle passes through the
origin (Figure 133), let a be its radius and a the angle from the
initial line to the diameter OA through the origin. If P{r, 6) is
any point on the circle, 0.4 P is a right triangle of hypotenuse
208 POLAR COORDINATES [Chap. IX

OA = 2a, side r = OP, and acute angle AOP = 0 — a. Thus

r = 2o cos (8 — a) (2)

is the equation required.

If the circle does not pass through the origin (Figure 134) let
its radius be a and its center C{b, ff). The triangle OOP has sides
a, b, r and angle — /3 opposite a. Thus

= 6® + r® — 2br cos {8 — 0) (3)

is the required equation.

Polar Equation of a Conic. Let F be a focus and RS the
(3)
corresponding directrix of an ellipse, parabola, or hyperbola (Fig-
ure 135). Take F as origin and
the axis through F as initial line.
If DF
= km the distance from the
directrix to the focus and P(r, 0)

is any point on the curve,

FP = r, NP = k r CO&0.

In Chapter IV it was shown that

FP is c times NP, where e is the

eccentricity. Thus

r = e{k -\-r cos 0),

whence

1 — e cos 0

is the equation of the curve.

§ 102 ] DIFFERENTIAL RELATIONS 209

102. Differential Relations. Let 0 be the origin of polar coordi-

nates, P{r, e) a point on a given curve, and s the arc length meas-
ured along the curve from a fixed point Po to P. The letter ^j/

is used to represent the angle from the direction along OP in which

r increases to the tangent PT drawn in the direction in which s

increases.

FiatTBE 136 .

To determine this angle let Q(r -f Ar, 0 + AO) be a neighboring

point on the curve (Figure 136). Draw QR perpendicular to PR,
and let As = arc PQ. Then

sin (RPQ) = —
RQ
PQ
=
(r + Ar) sin AO

PQ
= (r -f Ar)
sin

AO
AO AO As
AsPQ
,

PR (r + Ar) cos AO — r
cos (RPQ) = ~
PQ PQ
Ar As r(l — cos AO) AO As
= cos A«
As PQ AO As PQ
As AO tends to zero
sin AO 1 — cos AO As
lim (RPQ) = lim = 1, lim = 0, lim = 1.
AO AO PQ
The above equations therefore give as limits

sin ^ =
rdO
,
cos ^'
1
= —
dr

ds
(1 )
ds
210 POLAR CJOORDINATES [Chap. IX

These equations show that dr and r dB are the sides of a right

triangle with hypotenuse da and base angle ^ (Figure 137). From

Figure 137.

this triangle all the relations connecting dr, ds, dSy and ^ can be
obtained. In addition to (1) the most important of these are

rdS r
tan 4/ = ( 2)
dr 7r'

dB

ds^ = dr^ + r^ dB^, (3 )

The arc a is usually drawn along the curve in the direction in

which B increases. Thiswill be assumed in the problems.

Example 1, The logarithmic spiral

r =
The curve is shown in Figure 138. In this case

dr = kae^^ dB
and
rdB 1
tan^ =
dr k

The angle \f/ between the radius and tangent is therefore constant. The
equation
dr

dr
shows that — is also constant and so r and a increase proportionally.
da
§ 102 ] DIFFERENTIAL RELATIONS 211

Example 2. Find the slope of the logarithmic spiral at the point (r, B),
From the diagram (Figure 138) we have

^
The slope is therefore

i + tan 0
^
tan
,
= tan . ,
\f/ + tan B« =
,
.
k
-
1 + A; tan B
(t>
1 ~ tan ^ tan B . 1 ^ ^ k — tan B
1 — 7 tan ^
k
AtB — 60®, for example, the slope is

m=
l+ kVs
;=-•
k- Vs
Example 3. Find the angle between the cardioid r = a(l + sin B)
and the circle r = 3a sin B at each intersection.
The curves are shown in Figure 139. To find intersections we solve
simultaneously, thus obtaining

1 + sin 0 = 3 sin

whence sin ^ - and B

2
= -
6
or —6 If and ^2 are the angles on the

circle and cardioid respectively, we have at 0 ==

, ,
ri 3a sin ^ 1

^
^ 3a cos 0 Vs ’
212 POLAR COORDINATES [Chap. IX

Figure 139.

tan ^2 =
dr2
a(l + sin e) ^
a cos 6
^
Hd
The angle
/3 = ^2 —
from the circle to the cardioid thus satisfies the equation

tan ^2 ~ tan 1
tan p =
1 +
tan ^1 tan ^2
V3'
whence P = 30®. By symmetry the angle at 0 = —6 has
Stt,
the same magni-

tude. At the origin the cardioid is perpendicular to the x-axis and the
circle is tangent. At this point the angle between the curves is thus 90°.

103. Area in Polar Coordinates. We shall determine the area

of a sector AOB boimded by two radii 6 = a, 0 = P > a, and the
arc AB of a continuous curve

r=m.
Areas of a more general form can be expressed as sums or differ-

ences of areas of this kind.

Divide the angle AOB into parts by intermediate values Bi,
$2 •••, Bn-i arranged in order from B = atoB = p. In each inter-
val B, B +AB choose a value B' and construct a circular sector of
angle AB and radius r' = f(B'). The sum of the areas of these
circular sectors is

t Wf AB.
§103] AREA IN POLAR COORDINATES 213

When the number of divisions increases and the intervals AS

all tend to zero this sum approaches the area AOB as limit. There-
fore

A =

is the area bounded by the curve and the radii 6 = aj 6 = p.

To evaluate this integral r must be replaced by its value from
the equation of the curve.

Example. Find the area enclosed

by the lemniscate

= 2a2 cos 26.

One loop of the curve (Figure

141) extends from 6 = to

The area of this loop is

1 - — r 2 n-
- r* dO = COS 26 d6 = sin 2^
^
= aK

The entire area is therefore 20^.

214 POLAR COORDINATES [Chap. IX

PROBLEMS

1. Plot the following points: A ^2, , R^2, — 2,

2. Plot the following points: A(3, 120®), R(3, -120®), C(-3, 120®),
D(-3, -120®).
8. Determine all the pairs of polar coordinates of the point
(-!)
4. Determine all the pairs of polar coordinates of the point (4, —300®).

Sketch the graphs of the following equations:

6. r =a o sin 6. 6. r = a sin

7. r =a a(l — cos 6). 8. r = a(l — sin d).

9. r s a sin 2d. 10. r = a cos* d.
11. r a a sin* d. 12. r = a sin 3d.
18. r * a(2 + sin d). 14. r = a(l + 2 sin d).
16. r « a(l — 2 cos d). 16. r = a sin Jd.

17. r a cos Jd. 18. r = a sm ~ •

3
19. « a* cos d. 20. r* = 2a* sin 2d.
21. “ a\\ — cos d). 22. r = e®.
28. r » 4c”2. 11 1
26. rd = 1. 26. r = a sec d.
27. r “ a sec d + a esc d. 28. r » a tan d.

29. Show that (1, fir) is on the curve r = sin 2d.

80. Show that (1, ^ir) is on the curve r = -2 sin -

5
81. If a is constant, show that the equations
1 1

a cos d -|- 1
’
a cos d — 1

represent the same locus.

82. Show that the equations coa^ 2d,r — a cos 26 represent the same
locus.

Plot each of the following pairs of curves in one diagram, and find their
intersections:

88. r sin 0 =s o, r cos 6 ^ a, 84. r » 2a cos 0, r a.

86. r sin a, r sin

86. r — o cos ,
r — a cos
+0 *

87. r* - 2o* cos 2d, r - a.

 PROBLEMS 215

88. r - o(l + cos e), r - a(l + sin $).

89 r . - a sin 2$y r « o(l — cos 2$).
40 . sin Oy r*« sin 20,
By changing to rectangular coordinates determine the loci represented by
the following equations:

41 r cos 6 « —2.
. 42 . r s 4 sin 0.

48 r “ cos ^ -f sin 0.
. 44 . r - 2 esc ^ — 3 sec 0.

46 .— 2r(cos ^ + sin 0) -f- 1 =0. 46 . r(l — cos 0 ) = 3.

47 r(2 — 3 cos 0) = 6.
. 48 . r cos 20 = sin 0,

49 Find the Cartesian equation of the cardioid r

. = o(l + cos 0),
60 Find the Cartesian equation of the lemniscate
. == 2a^ cos 20,

Determine the polar equations of the following curves:

61 . = 4x. 62 . ^ 2x.
68. xy = 1 . 64 x^. — y^ =* 1 .

66 x^. + y^ 4ax + 4ay. 66. (x2 + = a\x^ - i/).

67 Find the polar equation of the

. line through the point (a, 0) perpendic-
ular to the initial line.
68 . Find the polar equation of the straight line LK (Figure 132) if N is the

69 Find the polar equation

. of the straight line with intercepts a, h on the
X and y axes.
60 Find the polar equation
. of the circle with center (a, 0) passing through
the origin.

61 . Find the polar equation of the circle with center r o, 0 “ passing

through the origin.

62 Find the polar equation of the
. circle through the origin with center

68 . Find the polar equation of the parabola with focus at the origin and

vertex

64 Find the polar equation of the parabola with focus at the origin and
.

vertex

66 A variable line through the origin 0 intersects a fixed line in Q, and the
.

point P is taken on this line such that OP-OQ = k^. Find the locus of P,
66. On the line OQ through the origin O and the point Q(r, 0) of the curve
r « a cos ^ two points P are taken at distance b from Q. Find the locus of P.
67 A variable line through 0 intersects a fixed circle in the points Qu Qa,
.

and on this line the point P is taken such that

20P - OQi -f OQ2 .

Find the locus of P.

216 POLAR COORDINATES [Chap. IX
68. With the notation of the preceding problem find the locus of P if

OP OQi OQt
69. A line of length 2a slides with its ends in the x and y axes. Find the
locus of the foot of the perpendicular from the origin to the moving line.
70. Find the locus
of a point if the product of its distances from two fixed
points equal to the square of half the distance between them.
is

71. Find the angle ^ on the curve r = a$.

72. If 0 < ^ < 2ir and r = a(l — cos $), show that ^ = Jd.
74.
73. If 0 <0 and * 2a^ cos 20, show that
4
^ » 20 +
Find the angles at which the following pairs of curves intersect:

r = 2a cos 0, r cos 0 = a. 76. r » 2a cos 0, r = a sin 0.

76. r = a cos 0, r *= a. 77. r = 1 + sin 20, r — 1.

78. r = a sin ^0, r = a sin 0. 79. sin 0, = a^(l — sin 0).

9 0 9 0
80. Show that the parabolas r = a sec^ r = b csc^ - intersect at right

angles.

81. Show that the curves r

86.
« oc*®, r — he ^ are perpendicular at each point
of intersection.
82.
87. Taking the initial line as o^axis, find the slope of the cardioid

r « a(l + cos 0).

83. Find the points on the cardioid r = a(l — cos 0) at the greatest distance
from the initial line.
84. Let P(r, 0) be a point on the hyperbola sin 20 = c. Show that the
triangle formed by the radius OP, the tangent at P, and the initial line is
90.
isosceles.
Find the circumference of the circle r =« a by integration.
86. Find the circumference of the circle r ^ 2a cos 0.
IT
Find the distance along the line r a sec from0 Oto0 = 2‘
88. Determine the arc of the logarithmic spiral r — a^^ from 0 =* 0 to the
point (r, 0).

89. Show that the arc length of the cardioid r =* a(l + cos 0) satisfies the
equation
=. 2a2(l + cos 0) dd^ * cos^ - dfi.
2
Determine the perimeter of the cardioid.

If r « a sin* - show that

,

ds a sin‘ - (1 - cos §0) dd.

Determine the complete perimeter of the curve.

 PROBLEMS 217

91. With suitable conventions as to the angles used, show that the angle
from the initial line to the tangent to a curve is

0 = = tan ^
“T*
•

dr

By calculating
d<t> d<f> dd
ds dd ds

and taking its reciprocal, show that the radius of curvature is

r « .

96.
Using the formula in the preceding problem find the radius of curvature of
each of the following curves:

92.
97. r = ae. 93. r = a(l + cos 0).
94. r = ae' 96. r = a sim - •

Find the radius of curvature of the lemniscate

= 2a^ cos 20

Find the area within one loop of the curve

98. Find the area within one loop of the curve

r“ = sin nO.

99. Find the area of the sector bounded by an arc of the curve r = and
two radii n, r2 .

100. Find the area bounded by an arc of the spiral

and two radii n, r2 .

101. By calculating and using the identity

cos^ ^ = §(1 + cos 20)

determine the area enclosed by the cardioid r = o(l + cos 6),
102. Find the area within the lemniscate

= 2a^ cos 20
and outside the circle r ^ a.
 CHAPTER X
VECTORS

104. Scalar and Vector. The simplest quantities of physics,

such as mass and temperature, are characterized by magnitude
only, or by magnitude and an algebraic sign. Such quantities are
called scalar. When units of measurement have been chosen a
scalar is represented by a real number and graphically by a dis-
placement along a straight line.
In the description of certain quantities direction as well as mag-
nitude is required. Thus to specify the velocity of a moving body
it is necessary to state the direction and the speed of motion.

A quantity that is characterized by magnitude and direction is

called a vector. Force, velocity, and acceleration are illustrations.
Such a quantity is represented graphically by an arrow of length
equal to its magnitude pointing in the assigned direction.

Thus if a point moves from P to Q its displacement is a vector

represented by the arrow from P to Q and in symbols by the nota-
tion
PQ.

Two vectors A, B are called equal

A=B
if they have the same length and direction. Thus if a rigid body is
moved without rotation all its points are moved the same distance
218
§106] ADDITION AND SUBTRACTION OF VECTORS 219

in the same direction. The displacement of each point is a vector,

and the displacements of different points are equal vectors.
The symbol ~A is used to represent a vector having the same
magnitude as A but the opposite direction. Thus the displacement
from Q to P is the negative of that from P to Q,

S
Associated with a vector A is a positive scalar equal to its length
or magnitude. We represent this by the notation A Thus, if a |
|.

is the length of the vector A,

a = I
Aj.

With respect to multiplication and division we shall find that

vectors and scalars satisfy different algebraic laws. It is therefore
desirable to have a notation that indicates clearly which quantities
in an equation are scalars and which are vectors. In most books
vectors are indicated by bold-faced type. As this is impossible to
indicate in writing, an arrow or a bar is often placed over a letter
—> —
to show that it represents a vector. Thus A or A would be under-
stood to represent a vector.
106. Addition and Subtraction of Vectors. Given any two vec-
tors A, B (Figure 144) we can construct a vector equal to B with

its origin at the end of A as shown in Figure 145. The vector

from the initial point of A to the terminal point of this vector B
is called the sum of A and B and is indicated by the notation
A+ B.
To obtain a physical interpretation of this sum consider a point P
Suppose all points of the box displaced the amount A
inside a box.
and at the same time P displaced with respect to the box the
amount B. The total displacement of P in space is then A + B.
220 VECTORS [Chap. X
Since the opposite sides of a parallelogram are equal and parallel,
it is clear from Figure 146 that

A+B == B + A, (1)

both sums being equal to the vector diagonal C. The sum is thus
independent of the order in which the vectors are added. This
is called the commutative law of addition.

Similarly the sum of three or more vectors is obtained by con-

structing a polygon (Figure 147) having those vectors as consecu-
tive sides and drawing a vector from the initial point of the first

to the terminal point of the last. By constructing the vector

diagonals A + B and B C it is also clear from this diagram that
-f-

A + B + C = (A + B) + C = A + (B + C). (2)

The sum thus independent of the way its terms are associated
is

into groups. This is called the associative law of addition.

If two vectors A and B are drawn from the same origin (Fig-
ure 148) the difference A — B is defined as the vector extending

from the end of B to the end of A. This diagram also shows that

B + (A - B) = A,

which states that A—B is the vector which added to B gives A.

§ 106 ] MULTIPLICATION BY A SCALAR 221

106. Multiplication by a Scalar. If m is a scalar and A is a

vector, the notation
mA = km
is used to represent a vector m times as long as A and having the
same direction if m is positive, the oj)posite direction m is nega- if

tive. If m, n are scalars, it follows that

(m + n)k = mA + nk (1)

since both sides represent a vector m+n times as long as A,

having the same direction if m+ n is positive, the opposite
direction if m + n is negative.
Similarly, m is a scalar and
if
mA
A, B are vectors.
/
/
"'f
/ /
/ /
m(A + B) = mA + mB. (2)
/
/

For this equation merely states

that, if the sides A, B, and the
diagonal A+B of a parallelo- B
^
gram (Figure 149) are all in- Figure 149 .

creased in thesame proportion,

the new vectors mA, mB, m(A + B) arc the sides and the
diagonal of a similar parallelogram.
Equations (1) and (2) express what is called the distributive law
for the multiplication of scalars and
vectors.

Example 1, If A, B are vectors from the

point0 to the points A, B, find the vector
from 0 to the middle point of AB,
Let M be the vector from 0 to the middle
point M. The vector from A to M is
M — A, and that from ilf to B is B — M.
Since tlicse are equal we have

M - A = B - M,
whence
M = i(A + B).
Example 2, AB = CD =
AC = v, DE = u, and DF = ^DA,
Find CE and CF, and show that 0, F, E are on a line.
222 VECTORS [Chap. X
The requu^ arrangement of points is shown in Figure 151. We have

C£ = CD + DE = u--,
^ ^ ^=u- |(u + v) = fu - iv.

Comparison of these expressions shows that

OT =
The vectors CB and CF thus extend from C in the same direction. The
points C, Ej F are therefore on a line.
Examph 3, The vectors Ai, A2, A3 form consecutive sides of a regular
octagon. Express A3 in terms of Ai and A2.

In Figure 152, PQ is parallel to A2 and the angles at P and Q are 45®.

Since Ai, A2, and A3 are of equal length

PQ = I
Ai cos 46®
I + 1
A2 1 + I
A3 cos 45®
1
* (1 + \/2) I
^2 !•

Thus

PQ = Ai + A2 + A3 = (1 + \/2)A2,
and so
<3 - V^2 - Al.
§ 107 ] POINTS ON A LINE 223

107. Points on a Line. Let A, B, P be the vectors from a point

0 to the points A, P. The vector from A to P is P — A, and
that from A to B is B — A. If P lies on the line AB and divides
it in the ratio
AP
= m, (1 )
AB
these vectors have the same ratio, that is,

P - A = m(B - A),
or
P = mB + (1 — m)A. (2)

The vector P is thus a linear function of A and B with coefficients

whose sum is
m+ 1 —m= 1.

Conversely, if

P = mA + nB (3)
and
m + n = 1, (4)

the point P lies on the line AB and

AP n
PB^ m
For (3), in virtue of (4), can be written

(m + w)P = mA + nB,
whence
m(P - A) = n(B - P).
224 VECTORS [Chap. X
Example L Find the point P on the line AB three-fourths of the way
from A to B,
The vector from 0 to P may be obtained from equation (2) with m = f
or from (3) with m= n = f the result in either case being
,

P = iA + |B.

Example 2, A, B, C, D are vectors from 0 to the points A, P, C, D, and

Ip = 2^. (5)

Determine the vector from 0 to the point P in which PC and AD intersect.

Equation (5) is equivalent to

B - A = 2(D - C).

Transposing B and C to one side, A

and D to the other, this becomes

B + 2C = A + 2D
whence
B + 2C A + 2D

The left side of this equation represents the vector from 0 to a point on
PC; the right side, the vector from 0 to a point on AD, Thus either of
these represents the vector from 0 to P.

108. Center of Gravity, Centroid. If masses mi, m2 ,

• •
•, rrin

are located at the points Pi, P2 ,

• •
•, Pn> the center of gravity of
these masses may be defined as the point C such that

miCPi + Tn2CP2 + • •
•
+ TTlnCPn = 0. (1)

Let Pi, P2 ,
• •
*, Pn be the vectors from a point 0 to the points
Pi, P2 ,
*
*
*, Pnj and C the vector from 0 to the center of gravity.
Equation (1) is then equivalent to

mi (Pi - C) + m2(P2 -€)+•• ~t"^»(Pw C) — 0,

whence
miPi + rn^2 + • •
•
+ WnPn
C= ( 2)
wi + m2 H Vrrin

is the vector from 0 to the center of gravity.

It should be noted that the vectors in (1) depend only on the
points C, Pi, P2 ,
• •
*, Pn- Thus the point determined by (2) is
independent of the point 0 from which the vectors C, Pi, P2 ,
• •
•,

P„ are drawn.
226 VECTORS [Chap. X

parallel to OX and OY (Figure 157). If Vx and Vy are the com-

ponents of V along the axes, these vectors have the form

and so
PR-=iVxy ^= iVy

V= iF. + iVy. (1)

Being the hypotenuse of a right triangle with sides Vx and Vyy

the length of V is
I
V = VfI
I + Vl. (2)

In particular, the vector r = OP from the origin to the point

P(x, y) has the components x, y (Figure 158). This vector

r = Lc + jy (3)

is often used to represent the variable point P{x, y).

If U and V are two vectors with components f7*, Vy and F*, Vy

U = iC7* + jl7„, V = iF, + jF*
and
U+V= + F,) + j(C7„ + Vy).
§110] VECTOR FUNCTION OF A SCALAR VARIABLE 227

The components of the sum are thus obtained by adding corre-

sponding components.

Example 1, Determine the vector from P(— 2, 3) to Q(4, —2).

The components are evidently the
differences of the coordinates (end
minus beginning) at the two ends of the
vector (Figure 159).
Thus
PQ = 6i - 5j.

Example 2. Find the vector from

the origin to the centroid of the tri-

angle formed by the points Pi(l, —2),

P2(3, 4), P8(5, 1).

The vectors from the origin to the three points are

Pi = i - 2j,

P2 = 3i + 4j,
Pa = 5i + j.

The vector from the origin to the centroid is thus

C = ^Pi + P + Pa)
2 = 3i + j.

The centroid is the point (3, 1).

110. Vector Function of a Scalar Variable, Derivative. If to

each value of a scalar variable t there corresponds a value of a
vector V, then V is called a function of t, a relation indicated by the
notation
V = F(0.

Thus, if t is the time and r = OP is the vector from a fixed point

0 to a variable point P, then r is a function of t
When t approaches to the vector F(0 is said to approach the
constant vector A as limit if the magnitude of the difference
F(0 — A tends to the limit zero. That is,

limF(0 =A
t

and
lim I
F(0 -A = i
0

are by definition equivalent equations. If A is not zero, this

means that when t is sufficiently near to the vector F(0 has almost
228 VECTORS [Chap. X
the direction as well as almost the magnitude of A. But if A = 0
the direction of ¥(t) may vary arbitrarily provided merely that

I
F(0 I
tends to zero.
When t changes to < + the change in F(0 is

AF = F(^ + AO - F(0.
Since At is scalar the ratio
AF
At

is a vector parallel to AF. If this ratio tends to a limit when At

tends to zero, that limit, written

dF
— = lim —
AF
dt At

is called the derivative of F with respect to t In particular, if t

is the time, —
dF
dt
.

is called the rate of change of F.

In proving the formulas of differentiation no direct use was made

of the fact that the expression differentiated was scalar, but merely
that it satisfied certain algebraic laws (such as the distributive law
of multiplication) and that it had certain continuity properties.
Formulas that involve only addition and multiplication by a scalar
are therefore also valid for vector functions. Thus, if u and V
are differentiable, u a scalar and V a vector.

— + V—
dV du
= u
dt dt

To prove this one would calculate A(wV), divide by A^, and take
the limit just as if V were scalar.

Example, If A and B are vector constants, determine the first and

second derivatives of
V = A cos ^ + B sin ^
with respect to 6,

Differentiating as if A and B were scalar constants, we get

—A sin ^ + B cos
dd

(PV
—A cos ^ — B sin 6,
dS^
§ 111 ] VELOCITY AND ACCELERATION 229

111. Velocity and Acceleration. If a particle P moves along a

straight or curved path, its vector displacement

r = OP

from a fixed point 0 is a function of the time L The rate of change

of displacement
dr
V = (1)
dt

is called the velocity and the magnitude of velocity |

v |
is called
the speed of P.
To determine these quantities let P be the position of the
particle at time t, Q its position at time t +
A^, and s the arc along

the path from a fixed point Pq to P. During the interval At the

changes in r and s are Ar = PQ and

As = arc PQ. Also

Ar As
—
Ar
( 2)
A^ As A^
The vector

As As
extends along the line PQ and is of
length
I
Ar I
chord PQ
I
As I
arc PQ

As As tends to zero, Q approaches P, the line PQ approaches the

tangent, and the ratio of chord to arc approaches unity. Thus

—
As
approaches a vector of unit length

(3)

tangent to the curve at P.

In virtue of (1) and (3), equation (2) gives

dr ds
V = (4)
dt dt
230 VECTORS [Chap. X
as limit. The velocity is therefore tangent to the curve at P, and
the speed

is equal to the rate of change of distance along the curve.

The rate of change of velocity

dv dh
( 6)

is called the acceleration of the particle.

If 0 is taken as origin and x, y are the rectangular coordinates
of P,
r = Lr + jy.
Since i and j are constants,

dt dx dy
V = —= 1
,

(6)
dt

a = —=
d^r
i
dPx ,
d^y
(7)

Thus the velocity and acceleration of a moving point are vectors

with components along OX and OY equal to the first and second
derivatives of x and y with respect to t
The speed of the moving particle is

|v| (8)

Example 1. At time t the coordinates of a moving point are

X t cost, y t Bint,

Find its components of velocity and acceleration along the coordinate

axes at time t = 0.

The components of velocity are

§112] TANGENTIAL AND NORMAL COMPONENTS 231

The components of acceleration are

d^x
—2 sin f f cos f = 0,

dPy
—
^ == 2 cos t tsint = 2.

Examvle 2, At time t the coordinates of a moving point P are

X = a cos (aty y — a sin wf,

a and co being constants. Find its velocity, acceleration, and speed.

The displacement of P from the origin is

T = ix + yy = ia cos o)t -f ja sin coL

The velocity and acceleration are

V = —
dr
= .
—low sin
.
o)t
.
+ ,
.
cos (at.

(Pr
* — iato* cos ust — jaw* sin wf.
dfi

Comparison of the expressions for a and r shows that

a = — w*r.

The acceleration is thus directe<l toward the origin and in magnitude

equals w* times the distance. The speed is

I
V I
= '\/ a*w* sin* (at + a*w* cos* (at = aw.

112. Tangential and Normal Components. If

T=OP
is the vector from a fixed point 0 to a variable point P on a curve,
and s is arc length measured along the curve from a fixed point Pq
to P, we have found that
dx
= t (1)
ds

isa imit vector tangent to the curve at P [§111, (3)]. Along the
curve (Figure 161) this vector may be considered a function of s.
232 VECTORS [Chap. X

Figure 161.

To determine its derivative construct a vector CA equal to t with

its origin at a fixed point C (Figure 162). As P moves along its

curve the end A of this vector de-

scribes a circle of unit radius. If
CB taken parallel to the a:-axis,
is

the arc BA on this circle is equal to

the angle from the x-axis to the
<l>

tangent at P. By equation (1;

dt

Figure 162. d<l>

is then a unit vector n tangent to the circle and so perpendicular

tot. Also -

ds p
is the curvature at P (§96). Thus
dt dt d<i> n
ds d4> ds p
extends along the normal at P and has a magnitude equal to the
curvature. We can therefore write
dh dt n

where n is the unit normal at P and p is the radius of curvature.

§ 112 ] TANGENTIAL AND NORMAL COMPONENTS 233

From a diagram it is furthermore clear that this vector is directed

toward the concave side of the curve. If the radius of curvature
is considered positive (as is usual), the unit vector n thus extends
along the normal toward the concave side of the curve.
Suppose that a particle moves along the curve and at time t

has the position P. Its velocity is

dr ds
v =
dr
—
dt
= =
ds dt
t —
ds

dt
(3)

Differentiating with respect to t and using the equation

dt dt ds n ds
dt ds dt p dt

we obtain for the acceleration

dv dh n /dsy
a = — = —r t "h
dt dt^ P \dt)

The vectors t and n are unit vectors along perpendicular directions.

The coefficients of t and n in (4) are therefore the components of
acceleration along those directions. Calling these components at
and Un, we thus have
d^s
tangential component — at — , (5)

1 /dsV
normal component (6)
p \dtJ
When the absolute value

of the acceleration is known the

equation

Figure 163.

as the normal component of acceleration. Since

234 VECTORS [Chap. X
this can also be written

l/d^x\^ /dW /dh\^

“* ‘
^'w) + 1) " (5?) ®
Example, A particle P moves in a parabola, its position at time t being

X = ht, y = ct — Igt^,

Find its tangential and normal components of acceleration.

The vector from the origin to P is
T = lx + iy = ibt + i(ct ~
The velocity, speed, and acceleration are therefore

V = ^ = + j(c - ife gt), (9)

|v| = ^= V6^+(c-ff0*, ( 10)

a = -iff- (11)

By equations (5) and (10) the tangential component of acceleration is

_ ^ _ g(c - g<)

Vfr* + (c - gO®
Since the absolute value of the acceleration is

1
a 1
= I -iff I
=
equation (7) gives for the normal component

W -
V6^'F-gO*‘
113. Polar Codrdinates. When the position of a moving point P
is represented by polar coordinates, components of velocity and
acceleration are usually taken along the radius vector r = OP
and perpendicular to it. Let Ur be the unit vector along r in the
direction of increasing r and u« the imit vector perpendicular to r
in the direction of increasing 0.

Along a curve (Figure 164) u, and Ujmay be considered func-

tions of 0. To
determine the derivatives of these functions con-
struct vectors equal to Ur and U; extending from a fixed center C
(Figure 166). The ends of these vectors lie on a unit circle, and
236 VECTORS [Chap. X
By differentiating with respect to the time, using equations (1)
and relations of the form

dVLr dUr dd
*
di do dt
we obtain
dr dr do
V (3)
dt

d^r r d^O dr d^l

a — r h 2 (4)
de dt dt\

as the velocity and acceleration of the particle.

Denoting components along r and perpendicular to r by the
subscripts r and we thus have

dr
II
(5)

d^r /deV d^e drdJB

ss: 7*
"t"
2 , (6)
dt^ dt dt

as the components of velocity and acceleration. It should be

noted that these are not merely the first and second derivatives
as in rectangular coordinates.

Example, A particle moves in the parabola

^
1 *- cos ^
(7)

with constant speed v. Find its acceleration.

Writing
0
1 — cos 0 = 2 sin^ -

we have

dr
-pcsc*-cot^^,
dt

. /dr\* . s «
•'-(*) +’^(5?) -’'““zb;-
§113] POLAR COORDINATES 237

If we assume that 0 is between 0® and 180° and is increasing, the last two
equations give
dd V ,

dr 6
^=-«C08 -. (9)

Since the tangent line makes equal angles with the radius vector and a
line parallel to the axis (Problem 89, page 142) these equations merely
dr dO
state that — and r — are the components of velocity along the radius vec-
dt dt
tor and perpendicular to it. By differentiating a second time and sub-
stituting in (6), we obtain
de\^ 0
Or ~ ( 10)
V^) ~2'i 2
’

m ,
dr dd e 0
( 11 )
cUr dt dt 2p 2 2

By use of (8) these components can be written

V d0 . 0
“^=-2d^^“‘2'
V dd 0
^‘-2dt^^2
TT 0
These equations show that the acceleration makes an angle with
2 2
the line PO (Figure 166) and is therefore normal to the curve, as it must
238 VECTORS [Chap. X
be since the speed is constant. The magnitude of the acceleration is

V dd

2W
This also follows directly from the fact that the speed is constant and
the vector v makes the angle ^6 with a fixed direction. For, if we draw
a vector

CA=v
from a fixed point C, the acceleration
1.

<U

is the velocity of the point A in a rotation with angular velocity — {\B)

di
around a circle of radius v.

PROBLEMS
6.

If A, B, C, D are vectors from a point O to the points A, B, C, O, and

B -A = C - D,

show that ABCD is a parallelogram.

AB « u and BC =* v are sides of a square ABCD, and E is the middle
2.

point of CD. Express AE in terms of u and v.

3 If AD = \AB and BE = |BC, express DE in terms of AB and BC.

.

4 If AB =* u, BC = V, and D and E are the middle points of AB and BC,

.

express AB, CD, and DE in terms of u and v.

If A, B, C are vectors from the origin to the points A, B, C, and D is the

point on the line BC one-third of the way from B to C, find AD.

6. A, B, C are vectors from a point O to the points A, B, C, and D is the
point on the line AB one-third of the way from A to B. If B is the middle

point of AC, find DE,

7. From
the vertices of a triangle vectors are drawn to the middle points
of the opposite sides. Show that the sum of these vectors is zero.
> — —^ — — — —
8. If AB CD « u, AC = V, AP
= Ju, and CQ =» |u, express PQ in
terms of u and v.

9. If AB * CD = u, AC = V, DE = ^DB, and DF «* iDA, express CE

and CF in terms of u and v, and show that the points C, B, B are on a line.
10. A,B are vectors forming consecutive sides of a regular hexagon. Find
the vectors which form the other four sides.
11. The vectors Ai, A2 , As form consecutive sides of a regular polygon of
n sides. Express As in terms of Ai and As.
 .

PROBLEMS 239

12 A, B, C, D are four points so placed that

AB = DC,
13.
M is the middle point of BC, and P is the point on BD one-third of the way
from B to D, Show that A, P, M are on a line.

If A, B, C, D are four points so placed that

14.
AB
E is on the line BC two-thirds of the way from B to (7, and BD = DF, show
that A, E, F are on a lino.
A, B, C are vectors extending from a point. If three numbers x, y, z
16.all zero) can be found such that
(not

xA yB + zC = 0,

X -hy -\r z = 0,

show
17. that the ends of A, B, C are on a line.
Three vectors A, B, C extend from points on a line. Determine the
condition that their ends lie on a parallel line.

16. If OPi = A, OP2 = B, and

18.
OP3 = A sin^ ^ 4- B cos^ t,

show that Pi, P2 P 3 are on a line.

,

If Pi, P2 P 3 are vectors from O

,
to points Pi, P2 P 3,
on a line and m\,
W 2 W 3 are any numbers with sum equal to
, 1, show that

P = miPi -f m2P2 +WP 3 3

is the vector from O to a point P on the same line.

A, B, C, D
are vectors from O to the points A, B, C, D, and

AB = mCD,
Find the vector from 0 to the point P in which BC and AD intersect.
19. A, B, C are vectors from 0 to the points A, B, C, and D, E are points
on the lines AB, BC such that

AD n BE p
DB m ’
EC n
Determine the vector from O to the point P in which the lines AE and CD
intersect.
20. M is the middle point of AB and P the point on the line MC one- third
of the M to C. Show that P
way from is the centroid of A, B, C.
21. Afi the middle point of AB, M
is 2 the middle point of CD, and M the
middle point of M M Show that M 1 2. the centroid of A, B, C, D,
is

22. If the masses mi, m2 ,

• •
•, mn are located at points Pi, P2, •,
• •
Pn on a
line, show that on the same line.
their center of gravity is

28. The length of a vector V is 4, and the angle from the x-axis to V is 60°.
Find the components of V along OX and OF.
240 VECTORS [Chap. X
26.
24. Given A (2, —3), B( — 1, 4), C(3, 6), express AB, BC, CA in terms of
i and and show that the sum
j, of these vectors is zero.
Show that the vectors
Vi - 2i - 4j,

V2 = i + 2j,
V3 2j - 3i *
form the sides of a triangle.
26. Find the slope of the line along which the vector

V = io -h j6
extends.
27. Show that the vectors

Vi = i + 3j, V2 = 2i + 6J
are parallel.
28. Show that the vectors

Vi = io + j6, V2 = ja — ih
are perpendicular.
29. Find the components of 3i and 4j along the vector

V= 3i + 4j.
30. Find the component of the vector 5i -f j along the vector 5i — 12j.
31.Show that the vectors Vi, V2 Vn have a sum equal to zero if and
,
• •
•
,

only if the sum of their components along every direction is zero.

32. The vector 3i — 2j extends from P(— 4, 6) to Q(x, y). Find x and y.
33. Given A(l, 3), B(5, —1), C(3, 2), find the coordinates of D if

AB = 2^.
34. Weights of and 5 pounds are placed at the points Pi(2, —3),
1, 3,
P2(4, 3), P3 (— 1, 6) Find the center of gravity of these weights.
respectively.
36. Let i, j and i', j' be unit vectors along two sets of rectangular axes OX,

OY and OX', OF'. If y and x', y' are coordinates of the same point P with
respect to these axes, show that

+ 32/
* iV + i'2/'-

36. If A and B are vector constants and

F(0 A sin 2t -f B(1 ~ cos 2t),

findF(O) and

linn
FW
t-*0 t

37. The point P(x, y) moves along the curve

y ^2x? -X,
If r — OP is the vector from the origin to P, find
 PROBLEMS 241

38. OP is the vector from the origin to the point P(r, 0) on the curve r$ ^ 1.

Find
lim (OP).
r —0
39. If 6 is the angle from the x-axis to a given direction, show that
u = i cos 5 4- j sin 0

is a unit vector along that direction. Find the first and second derivatives
of u with respect to 0, and interpret the results graphically.
40. Determine the vector constants Ci, C2 such that

V = Cl cos 0 C2 sin d
satisfies the equation

— V = A sin 0.

In the following problems r == OP is the vector from the origin to a moving

46.
point P. Find its velocity, acceleration, and speed at the time indicated.
41. r = “ 3j^, < = 2.

42. r = - 1) 4- - 1), t = 1.

fl-

43. r = 3i sin < 4- 2j cos f, ^ •

44. r == 3ie^ 4- 4je~^ t = 0.

r = i In sin f 4- ^ •

46. r = i tan ^ 4- j cot tj f = 7 .

In the following problems find the components of velocity and acceleration,

62.
and the speed, of the point P(x, y) at time t:

47. X = 24<, y ^ St -
48. a; = « 4- y = t
-
49. X = t cos t — sin t, y = t sin t cos t.

60. X * In cos if y — t.

61. At time t the coordinates of a moving point are

X = aiP 4- bit 4“ Cl,

y = a 2p + b2 + C t 2,

dh bi, Cl, 02 b2 , C2 being constants.

,
Show that the acceleration is constant.
At time t the coordinates of a moving point are

* =I (e*‘ + 2/ =I (e*‘ - «"“)•

Show that its acceleration is proportional to the distance from the origin and
directed away from it.

68 . A point P(x, y) moves in the parabola y^ — 4px in such a way that

dy
~ is constant. Show that its acceleration is constant.
dt
242 VECTORS [Chap. X
64. A point P(a;, y) moves in the hyperbola xy ^ c m such a way that
dx
~ u is constant. Find its acceleration.
at
66.
In each of the following problems find the tangential and normal components
of acceleration:

X « 2 cos y — sin L
t,

66 . x--at\y^ U\
67. X n a cos kt, y ^ a sin kt,
B8. +
69. X = a(4> — sin </), y = a(l — cos <t>), 0 « kt.

60. X = o(0 cos 0 — sin 0), y = a(0 sin 0 -f cos 0), 0 *

61. If a particle moves in a curve with speed v and acceleration a, show that
the radius of curvature p of the path satisfies the equation
62.

In the following problems find the velocity, acceleration, and speed of the
point P(r, 6) at time i:

r oBf d = <at.

68. r = + c)\ 0 = k\n(t + c).

o(f

64. r = o(t + c), 0 = —^—

66. r - I (e® - «~*), 9 = orf.

66. r = 9 = 6c“‘.

67. The point P moves with constant speed v along the radius vector OP
while
70. OP turns about 0 with constant angular velocity «. Find the accelera-
tion of P.
68. A projectile P moves in a parabola under the constant acceleration g
of gravity. If r is the distance from the focus of the parabola to P, show that
d^r

69. A particle moves with constant speed v in the hyperbola sin 26 » c.

Find its acceleration. Show that it is perpendicular to the path and of magni-
tude (cf. Problem 84, page 216)

A particle moves with constant speed v in the cardioid

r — a(l — cos 6).

Find its acceleration. Show that it is perpendicular to the path and of magnl
tude (cf. Problem 72, page 216)
3 dJB
 CHAPTER XI
FORMULAS AND METHODS OF INTEGRATION

114. Elementary Functions. The algebraic, trigonometric, in-

verse trigonometric, exponential, and logarithmic functions and
those expressible by finite combinations of these functions are
sometimes called elementary Thus

is an elementary function.
By means of the differentiation formulas (§§21, 78, 80, 87) the
derivative of any elementary function can be obtained, the result
being again an elementary function. The indefinite integral of an
elementary function may not, however, be elementary. Thus

dx

isnot expressible as an elementary function of x.

In the integrations we shall perform the result will be an ele-
mentary function. Since no such function may exist, the method
of finding it cannot be entirely systematic. The general procedure
is to try to express the given differential as a sum of parts each of
which can be integrated by a known formula. A list of such for-
mulas is given in the following section. More extensive lists are
found in integral tables,
116. Formulas. In the following formulas u is any variable or
function of a single variable, du is its differential, n is constant,
and C is the constant of integration. In evaluating a definite in-
tegral C may be omitted, but in all other cases it should be added.

vT’ du = h C, if n is not —1.

n + 1

II.
/* —= In u + C.
J u
I I

243
244 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

III. udu ^ — cos u C,

IV. w dw = w +
J* cos sin (7.

V. sec^ udu = tan u C,

J*
VI. csc^ udu = — cot u + C.

= + C.
VII.
^ sec u tan w dw sec w

VIII. u cot u du = — u + C.
J " CSC CSC

— + C.
IX.
^ tan udu == In |
cos u \

X. ^ cot w dw = In I
sin w 1
+ C.

= + tan u +
XI.
^ sec w dw In I
sec w \
C.

XII. w dw = w — cot u\ + C,
J' CSC In I
CSC

du u
+ - = ,

/V —vr sin-1 -
a
C.

/—^ = -tan-i~ +
du 1 ^ u
C.
a a

—
/V -w
=
^udu
+
u^
. -
\
-sec-^-
a
C.

du
“ = + Vm2 ± o2 + c.
JfVu^ ±
XVI. -
7 In I
« I

/—
du 1 u — a
1 - i = ;ri“ + C.
u^ a? 2a u +a
XVIII. y^c“dw = e“ + C.
§1161 INTEGRATION BY SUBSTITUTION 245

It should be noted that logarithms only of absolute values occur

in these formulas. In the applications the absolute value signs are
usually omitted, it being assumed that the work has been so
arranged that logarithms of negative numbers do not occur. Also
it is understood that all square roots are positive and inverse trigo-

nometric functions are represented by principal values (§79).

Any one of these formulas can be proved by showing that the
differential of the right member is equal to the expression imder
the integral sign. Thus to prove XI we note that

I
sec w + tan u = |
zt(sec u + tan w),
the sign on the right depending on the value of u. Hence
it (sec + sec^ u) du =
u tan u
d In
,

sec + tan u = I

sec u du,
it(sec u + tan u)
|

116. Integration by Substitution. When some function of the

variable is a given integral may assume the form of
taken as w,
one of the integration formulas or differ from such form only by
a constant factor. Integration accomplished in this way is called
integration by substitution.
In each integration formula the expression integrated is of the
form f(u) du. To reduce a given differential to this form it is
necessary to choose u such that one factor becomes f{u) and the
remainder becomes du. Failure to make this second factor exactly
du is responsible for most errors in integration.

Example 1,
Jxy/2x^—— 1 dx.

The differential of 2x'^ 1 differs only by a constant factor from x dx.

Thus we take
2^2 — 1 = Uy 4x dx = du.
Then
y/2x^ — 1 dx = J du — \u^ +C - l{2x^ — 1)^ + C,

„
Example
, ^
2, /
7
r cos
7—
1 +
— —
dx
2a;
:

sin 2a;
7-

We observe that cos 2a; dx differs only by a constant factor from the
differential of 1 + sin 2a;. Thus we let

1 + sin 2x = u.
Then
2 cos 2a; da; = dw, cos 2a; da; = | du,
and
cos 2x dx
+ 8m2a:) + C.
/ 1 + sin 2a;
(l
246 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

Example S,
Jf
— •

9
This suggests formula XV. To bring it to exactly that form let

u = 2ty o == 3.

Then
^
dt r
- f
2dt

J t V4<* - 9 J 2iy/At^ - 9
du
uy/u^ —
1
= - sec -1- +C
a a
1 ,2t
= -sec-‘- C.

"
Example * sec^ x dx.
J
If u = tan Xf by formula XVIII
e^ * sec^ xdx
J* = J'e'^du = e^ * + C.

117. Powers of Trigonometric Functions. A power of a trig-

onometric function multiplied by its differential can be integrated
by formula L Thus, if u = tan x,

/ tan® x-sec^a:da; J tan^ X + C.

Differentials can often form by trigonometric
be reduced to this
transformations. The u a trigono-
general procedure is to take as
metric function such that one factor of the given differential is du
and the remaining factors can be expressed in terms of u without
introducing radicals. This is illustrated by the following examples.

Example 1.
J sin* x cos* x dx.
If we take cos xdxas d(sin x) and use the relation cos* x = 1 — sin* x,
the other factors can be expressed in terms of sin x without introducing
radicals. Thus

xdx =
J sin* X cos*
J*
sin* x* cos* x- cos x dx

—
* J sin* x(l sin* x) d(8in x)

s=
i sin* X — i sin® x + C.
§118] PRODUCTS OF SINES AND COSINES 247

Example 2.
Jtan x sec^ x dx.

If we take sec^ x as d(tan x) and use the relation sec^ a; = 1 tan^ x, +

the other factors can be expressed in terms of tan x without introducing
radicals. Thus

xdx =
Jtan X sec^
Jtan a; sec* x-sec* x dx

= J tan a:(l +
" tan* a;) d(tan x)

= § tan* aj + J tan^ a; + C.
Example 5. tan® x sec x dx.
J'
If we take tan x sec x dx as d(sec x) and use the relation

tan* X = sec* a; — 1,

the integral takes the form

tan® X see X dx = J'tan* x •

tan x see xdx

—
= J (sec* a; 1) d(sec x)

= *3 sec® X — sec X + C.

Example tan® x dx.

If we
4-
J
replace tan* x by sec* x — the integral becomes
1,

j'tan® xdx = tan® x(sec* x — 1) dx = J tan^ ^ ^ J'

J'
The integral is thus made to depend on a simpler one

tan® X dx.
J'
Similarly

tan® xdx — tan x(sec* x — 1) dx = ^ tan* x + In |

cos x |.

J*
Hence finally

tan® X dx = J tan^ x — | tan* x — In |

cos x |
+ C.
J'
118. Products of Sines and Cosines. Integrals of the form

/ sin*” X cos” X dx (1)

can be evaluated by the methods of the preceding section if at least

one of the exponents m or n is an odd positive integer. If both
248 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

m and n are even, however, those methods fail. In that event we

can make the integral depend on others of simpler form by means
of the formulas
1 — cos 2u
sin^ u = y ( 2)
2

1 + cos 2u
COS^ u = (3)
2

sin 2u
sm u cos u = (4)

Products of sines and cosines of different angles can be expressed

as sums or differences by means of the formulas
sin A sin B= ^[cos (A — JB) -- cos (A + B)], (5)

sin A B= cos ^[sin (A — B) + sin (A + B)], (6)

cos A cos B = ^[cos (A — B) + cos (A + B)]. (7)

Example x dx.
By
1.

formulas
J cos^

(2), (3), (4)

x sin^

we have
cos^ X sin^ x = cos^ x • cos^ x sin^ x

_ 1 + cos 2x sin^ 2x
~
“ 2 4

= ilsin^ 2x + sin^ 2x cos 2x]

= i 4- sin2 2x cos 2a? •

J
Consequently
sin 4x sin® 2a; ^
cos^ X sin^ xdx = H
, ,

r C.
16 64 48

Example 3x dx.
2,
J sin 2a; cos

By formula (6)

sin 2a; cos 3a; = §[sin (—a;) + sin 5a;] = §[sin 5a; — sin a;].

Hence
y*sin 2a; cos 3a; dx = — iV cos 5x + ^ cos + C. a;

119. Quadratic Expressions. Formulas XIII-XVII contain

quadratic expressions in which only the second power of the
variable appears. Integrals containing the general quadratic
§119] QUADRATIC EXPRESSIONS 249

0^ + bx +c
can be reduced to this form by completing the
square of ax^ + bx.
Before doing this, however, it may be help-
ful to separate a part in which ax^ bx c can be used as new + +
variable.

/to'+t + is -

Completing the square of the quadratic in the denominator, we get

Sx^ + 6a; + 15 = 3(a;2 + 2a; + 1) + 12 = 3(a; + 1)^ + 12.

Taking = x + 1 we thus have

dx r du
/i3x2 + 6x + 15 -h
J 3ii2 + 12

du
-\fvF+l
^itan-l + C-

dx
Example 2.
f V2 + 1 2a: - 3x2

The coefficient of x^ being negative, we place the terms containing x in

parentheses preceded by —3. Thus

2 + 2x - 3x2 = 2 - 3(^2 _ la;) = - 3(x - \Y.

If w = X — we then have
dx
ax _ r du
/ \/2 + 2x-3x^ ~ J \/|^3m 2

— ^ r
¥
1 u
sin" fc
V3 VI
1 3x - 1
sm" c.

(2x — dx
V3 V7
1)
Example 5,
J-\/4x2 + 4x + 2
Since the numerator contains the first power of x, we resolve the integral
into two parts

f*
(2x — 1) dx _ 1
1
/• (8x + 4) dx ^ r dx
J “ 7
V4®* + 4x + 4 \/4x2 + 4x + 2 \/4x2 + 4x + 2
250 FORMULAS AND METHODS OF INTEGRATION [Chap. XI
In the first integral on the right the numerator
is taken equal to the dif-

ferential of + + 2.
In the second the numerator is dz. The out-
4a;

side factors J and —2 are chosen so that the two sides of the equation
are equal. The first integral has the form

^ = ^\/4a;2 + + 2.
Z f 4a:

The second integral is evaluated by completing the square. The final

result is

(2a; — 1) da; __
/ V'4a;^ + 4a; + 2

+ 4a; + 2 - In (2a; +1+ V^ + 4.x + 2 + C,)

120. Rational Fractions. A fraction^ such as

a;3 + 3x
— 2a; — 3
^

whose numerator and denominator are polynomials, is called a

rational fraction.
If the degree of the numerator is equal to or greater than that
of the denominator, the fraction should be reduced by division.
Thus

+ 2 H— —
4“ 3x lOx “h 6
= X
x^ — 2x — 3 x^ 2x — 3

A fraction with numerator of lower degree than its denominator

can be resolved into a sum of partial fractions with denominators
that are factors of the original denominator. Thus

lOx +6 __
lOx +6 __
9 1

x^ — 2x — 3 (x — 3)(x + 1) X — s'^x + l

These fractions can often be found by trial. If not, proceed as

in the following examples. The procedure is one in which the order
of steps can be inverted. That the given expression is resolvable
into fractions of the kind suggested is in each case thus shown by
finding the fractions.

Case 1. Factors of the denominator all of the first degree and none
repeated.
§ 120] RATIONAL FRACTIONS 251

+ 2x + ^6 dx.
+—x^ — 2x
XT I 1 -7— ,
Example 1.
J/ x^

Dividing numerator by denominator, we get

x^ + 2x + 6 ^
Sx^ + 6 3x^ +6
x^ + x^ — 2x ^ x^ + x^ — 2x ^ x(x — + 2)
l)(x

Assume that
3x^
x(x-l)(x
+6
+ 2)
^ABC x"^x-l‘^x + 2*

Ay B, C
being constants. The two sides of this equation are merely dif-
ferentways of writing the same function. If then we clear of fractions,
the two sides of the resulting equation

3x2 +6= A(x + 2) + Bx(x + 2) + Cx(x - 1)

-
= (A+B + C)x2 + (A + 2R - C)x - 2A, (2)

are identical. That is,

A+R+C= 3, A+ 2R-C = 0, -2A = 6. (3)

Solving these equations, we get

A = -3, R= 3, C= 3.

Conversely, if A, 5, C have these values, equations (3), and so (2) and

(1), are satisfied. Therefore

+ 2x-Q
/:**
x*

+ X* — 2a:
+ —I
x+ 2/
dx

= |x2 — X — 3 In X + 3 In (x — 1)

4“ 3 In (x “b 2) + (7

= - X + 3 In ^ ± + c.
The constants can often be determined more easily by substituting
particular values for x on the two sides of the equation after clearing frac-
tions. Thus the equation (2) above,

3x2 +6 = _ i)(a. + 2) + Bx{x + 2) + Cx(x - 1),

isan identity satisfied for all values of x. In particular, if x = 0, it be-

comes
6 = -2A,
whence A = — 3. Similarly, by substituting x = 1 and x = —2 we get

9 = 3R, 18 = 6C7
whence B = 3, C= 3.
262 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

Case 2. Factors of the denominator all of first degree but some

repeated.
„ 1 r 8®* + 7 ,
y'(i+l)(itr+l)> -

Assume that
+78^^ ^ ^
,
C D
(x + l)(2x + 1)8 a; + 1 (2a; + 1)3 (2a; + 1)2 2a; + 1

Corresponding to the repeated factor (2a; 1)® we introduce fractions +

with (2a; +
1)® and all lower powers as denominators. Clearing and solv-

ing as before, we find

A = 1, 5= 12, C= -6, D= 0.
Hence
8^+7 ^ ^
+
12 6 1
r +!)»“* r dx
J (a: + l)(2x (2a: + 1)* (2a: + 1)*J
== In (a; + 1) — (2a; + 1)2 '
2a; +1+
C.

Case 3. Denominator containing factors of the second degree but

none repeated,
_
Example
-

3.
J — + -+—
/• 4a;2 a; 1 ,
dx.

The denominator can be factored in the form

x3 — 1 = (a; — l)(a;2 + + a; 1).

Assume
4a;2 +X+1 _ A Bx +C
x^ — 1 X — 1 x2 + x+ l
With the quadratic denominator x2 +x+ 1 we use a numerator that is

not a single constant but a linear function Bx + C. Clearing fractions

and solving for A, B, C, we find

A = 2, B= 2, C= 1.

Therefore
4x2 +X+ 1 + 1 dx
/(A—
2j;
dx
/ x® — 1 + x + i)
2 In (x 1) + In (x2 + X + 1) + C.
Case 4. Denominator containing factors of the second degree^
some being repealed.

Example 4.

Aissume
1
^ A Bx +C Dx +E
X(x2 +1)2 X (x2 + 1)2 x2 +1
§ 121 ] CHANGE OF VARIABLE 253

Corresponding to the repeated second degree factor {x^ 1)^, we intro- +

duce partial fractions having as denominators {x^ 1)^ and all lower +
powers of +
1, all the numerators being of first degree. Clearing frac-
tions and solving for A, R, C, D, we find

A = 1, H=-l, C = 0, Z)=-l, ^ = 0.
Hence

I -m- (x* + 1)2 x2 +

— I
1 l_
n

121. Change of Variable. An integral

J fix) dx (1)

may not be in form for direct evaluation by the methods we have

described but may be reducible to such form by a change of vari-
able. When X is replaced by

X = 0(0 (2)
and dx by
dx = 0'(O dt (3)

the above integral becomes

y fi<t>it)) <t>'it) dt.

Suppose that this can be integrated and the result is

Jmt)) <>'it) dt = #(o + c. (4)

This is the indefinite integral expressed as a function of L To

express it in terms of x we solve equation (2) for t in terms of x
and substitute in the right side of (4). If in this way we obtain

^(0 = F{x), (5)

then

f{x) dx = F{x) + C. (6)

In the case of a definite integral it is usually simpler not to

express the final result in terms of x but to change the limits.
264 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

Suppose that <2 are the values of t that correspond to the

values Xij X2 of x. From (5) we have

^(ti) = F(xi), ^(12) = F{x2).

Hence
F{x2 ) - F{xi) = ^(^2 ) -
From (6) and (4) this is equivalent to

dt (7)

To change the variable in a definite integral we thus merely replace

X and dx by their values in terms of the new variable and replace the
limits by the corresponding values of the new variable. In the above
discussion it has been assumed that, when t varies from ti to t2 ,

X varies from Xi to X2 and that for all intermediate values < is a

,

one-valued function of x, and conversely. If these conditions are

not satisfied equation (7) may not be valid.
z
122. Integrals Containing (ax + b)«. If p and q are positive
integers, the substitution
ox +6= 2®

can be used to express x, dx, and

z
{ax + 6)fl

as rational functions of z and dz. If several fractional powers of

the same linear function ox + occur, the substitution
?>

(ox + 6) = z^

may be used, n being so chosen that all the roots can be extracted.

dx
7==*
Jf 1 +
Example 1,
Vx-l
Let X — 1 = 2 *. Then dx = 22 dz and

dx r2zdz
/ 1 + ^/x - 1
^J l +z
dz

2«-21n(l + *) + C
2\/*- 1 - 21n (1 + ) + C.
§123] TRIGONOMETRIC SUBSTITUTIONS 255

Example 2.

Let X = 2®. Then dx = 62® dz and

+ 22
1

= 62 — 6 tan~^ 2 + C
= 6x^ — 6 tan“^ (a?i) + C.
123. Trigonometric Substitutions. If a differential contains
Va^ — Va^ + x^j or — a^, one should first consider
using the square root as a new variable. If that substitution fails,

the differential can often be integrated by taking x, a, and the

square root as the three sides of a right triangle and using one of
its acute angles as new variable.

Example L f x(d^ — x^)^ dx,

Jo
Let (a2 — x^)i = 2. We then have

a* — x2 = 2^, — 2x dx — 2zdZf xdx = —z dz.

When X varies from 0 to a, 2 varies from a to 0. Thus

x(a2 — x2)^ dx = jT (a2 — x dx

= z^{—zdz)
J*

=
" ?!
5*
a^ —
Example 2,
J y/ x^ dx.

Take a as hypotenuse and x as one side of a right triangle (Figure 167).

266 FORMULAS AND METHODS OF INTEGRATION [Chap. XI
Then
a: = a sin dx — a c^qbB dd, y/ = a cos B,

and
j* y/a^ x^ dx = Ja^ cos* B dB

== y*
^ (1 + cos 2B) dB
~ 4“ i sin 2B] +C

= -jr- [^ + sin 0 cos + C. B]

Z
From the triangle
\/a* —
B = sin""^ - ,
sin 0 ,
cos ^ =
a a
Hence

y* y/a? — dx = \/a* — +
x?
^ sin~^ a 1
x* C.

dx
/
•
1
(x* — a*)>

Take x as hypotenuse and a as base of a right triangle (Figure 168)

Then
X = a sec dx = o sec ^ tan B dB, \/x* — d* = a tan
and
dx r a sec BU tan U dB
B

/ (a;2 __ ^2^} J a^tSLTL^B

-n CSC B cot B dB

5 CSC ^ +C

— a* +
C.
a^y/x^
 INTEGRATION BY PARTS

Examvki.
(a
;2
+ ^2)2
Take a and x as the two sides of a right triangle (Figure 169). Then
X — a tan dx = a sec^ 6 dd, + x^ = o? sec^

Figure 169.

When X varies from 0 to a, ^ varies from 0 to . Consequently,

^
dx
^ 1 n
{x^ + Jo

cos^ 0 dd

= sinO cos 0]^

2a^ 0

JLri+i].
2a3 14 ^ 2J
124. Integration by Parts. From the formula

d(uv) = udv + vdu

we get
udv = d{uv) — V dUj
whence
udv = uv — J*vdu. (1)
J*
du is known, this gives u dv. Integration by the use of
this formula is called integration by parts.

Example 1, fin X dx.

Let w = In X, dy = dx. Then

and
y*ln X dx = (In x) X —
= X (In X — 1) + C.
268 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

ExamvU2. Jx^sinxdx.
Let u = x^y dv = sin x dx. Then du — 2xdXyV — — cos x, and

^ cos x
J'x^Binxdx + J*2x cos x dx.
A second integration by parts with u = 2x, dv = cos x dx gives

cos X dx = 2x sin X — ^2 sin X dx = 2x sin X + 2 cos X + C.

Hence finally

= — x^ cos x + + 2 cos x + C,
/ x^ sin X dx 2x sin x

The method of integration by parts applies particularly well to

functions that are simplified by differentiation, like Inx, or to
products of functions of different classes, like x sin x. In apply-
ing themethod the given differential must be resolved into a prod-
uct u*dv. The part called dv must have a known integral, and the
part called u should usually be simplified by differentiation.
Sometimes after integration by parts a multiple of the original
appears on the right side of the equation. It can be
differential
transposed to the other side and the integral solved for algebrai-
cally. This is illustrated by the following examples.

-- x^ dx.
Example 3.
J y/a^
Integrating by parts with u = y/a^ — x^ dv = dx, we get

y/a^ = xy/a* — “ /*
Jf
x^ dx x^
•/ V — x^

Adding to the numerator of the integral and subtracting an equivalent

integral, this becomes

J
f \/a* — X* dx = xVa* — x* — Jf
y/a^ -
_ dx + Jf— - r-.-

y/a^ x^

= xy/a^ — 7? —
^f
\/a* — X* dx + — ^
Jf y/a^ -
= -

x2

Transposing the second term on the right and dividing by 2, we get

y/a? — x^dx * — + %r sin“^ - + C.

Jf ^ y/a^ x*
z A a

Example cos bx dx.

Integrating by parts with u = e% dv = cos bx dXf we get

6®* sin bx a r
§1261 REDUCTION FORMULAS 259

Integrating by parts again with u = e®®, dv = sin hx dx^ this becomes

6 sin hx + a cos hx\

c®* cos hx dx = e“*
)-%! e®* cos hx dx.

Transposing the last integral and dividing by 1 + gives

^ /h
h sin hx + a cos hx
e®* cos hx dx = (
\ a
;

)
+ c.
126. Reduction Formulas. Integration by parts is often used
to make an depend on a simpler one and so to obtain a
integral
formula by repeated use of which the given integral can be deter-
mined.
To illustrate this take

^ X dx,

where n is a positive integer. Integrating by parts with

u = sin” ^
X, dv — sin x dx,
we have

^y^sin” xdx = — sin” ^ X cos x + fi.n- 1) sin” ^ X cos^ X dx

= — sin” ^ X cos X

+ (n — 1)
^y^sin” ^ x{\ — sin^ x) dx

= — sin”""^ X cos X + {n — 1)
sin”~^ x (

J*
- (n - 1)

Transposing the last integral and dividing by n, we get

sin” ^ X cos X n — 1
sin” xdx — 1
^y^sin”'
J" n n

This formula expresses the integral of sin” x in terms of that of

sin””^ x. By repeated application of the formula the result will
finally be made to depend on
^y
* dx or ^ sin
si xdx according as n

is even or odd.
260 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

E«mpU. dn-.*.
By the formula just proved

Jo O 0 o yo

At the given limits the integrated part is zero. Thus

X X
n
Jo
sin* a: dx =I
Jo
H sin* xdx

= ^-1 y** sin*xdx

= dx

= -fyir.

126. Infinite Values. The definite integral

f f(x) dx

has been defined only for a function f{x) which is bounded in the
interval (o, 6). If the function becomes infinite at a or & or at
any value between the limits, the definition in §42 is not applicable.
First suppose that/(a:) becomes infinite at x = 6 but is integrable
in the interval (c, 2), where 2 is any number between a and b. If
the integral

£f{x) dx

tends to a limit when z—*b, the integral between c and b is defined

as the limit

/(x) dx = lim
ffix) dx.
z-^b Ja

Similarly, if fix) becomes infinite at x = a,

fix) dx =
t-*a
lim r fix) dx,
J,

provided that the limit on the right exists.

§ 126 ] INFINITE VALUES 261

In the integral
f\dx
Jo
for example, the integrand

fix) = -7=
\/x
becomes infinite at the lower limit. Since

dX y— ' y-

J.
tends to the limit 2 when 2 — 0, we have by definition

dx dx .

Jo Vi J, y/x
If /(x) becomes infinite at some value c between a and 6, the
integral from a to 6 is defined by the equation

J{x) dx =/ f{x) dx+ j fix) dx,

provided that both limits on the right exist.

Take, for example,
dx
da n
J0 (x
(X — D*
-
The function

m =
{x - l)i

is continuous except at x = 1 where it becomes infinite. If zi is

between 0 and 1, and 22 is between 1 and 2,

- 3 (1 - fe - i)*l
jf (TTIji
Since both of these tend to limits when Zi and tend to 1, the
integral from 0 to 2 is the sum of those limits, namely
dx
(*-!)*"
3 +3 = e.
Jq
262 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

In the above discussions it has been assumed that a and b are

both finite. If

fix) dx

tends to a limit when 6 ,

the integral with infinite limit is de-
finedby the equation

/ f(x) dx = lim

Similarly, the integral with lower limit negatively infinite

P / fix) dx.

is
(6)

de-
fined by the equation
b rh
—
/ .
fix) dx lim
a-*-’. Ja
/ fix) dx. (6)

For example,
,b

Thus
f e^^dx = [-6-*]o= 1

= — =
/ e * cte
6
lim (1
- 00
e 1.

In each of the above definitions, (2), (3), (4), (5), (6), the inte-
gral on the left is said to converge, or to exist, if the limit on the
right has a definite value, and to diverge if that limit does not
exist.
All the integrals in the above examples have been convergent.
As illustrations of integrals which diverge, take

sin X dx.

The first of these diverges since

— In2;

becomes infinite when 2 —» 0. The second diverges since

in X dx — 1 — cos 6

does not approach a limit when 6 — 00 .

§1271 TESTS FOR CONVERGENCE 263

127. Tests for Convergence. We shall be concerned only with

integrals of functions which between the limits of integration have
at most a finite number of discontinuities. From the definitions
it follows immediately that, if the indefinite integral

fm dx = F{x)

of such a function is continuous in the interval of integration and

at its limits, the definite integral is convergent and its value is

determined by the usual equation

dx = [f(x)]‘= F{b) - Fia).

When the indefinite integral of a function is not known the con-

vergence or divergence of its definite integral can usually be deter-
mined by comparison with that of a simpler function which can
be integrated. The principle employed is the following:
Theorem. Suppose that p{x) is positive and f{x) and p{x) are
continuous for all values of x in the interval (a, b) except x = 6.

lim
x-^hp{x)
—
/(*)
= A; 0, (1)

the integrals

dx

are both convergent or both divergent.

To prove this we note that because of (1) the function f{x) has
a fixed algebraic sign when x is sufficiently near b. As 2 approaches
b the integrals

dx

thus ultimately increase continually or decrease continually. They

therefore approach limits or become infinite (§14, (2)). When z is
near b any change in the first integral is approximately k times the
corresponding change in the second. Thus either integral remains
finite if the other does.
To express that two integrals with the same limits both con-
verge or both diverge we may sometimes say that the one integral
behaves like the other.
264 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

In appl3dng this theorem we can break the interval of integra-

tion into a sum of parts and use a different comparison function
on each part. The whole integral is then convergent if each part
is convergent and divergent if any part is divergent.

Example 1, f —
y/xiX - x2)
On the interval of integration the integrand

\/x{\ — x^)

becomes infinite only at the limits x = 0, x = 1. Near a: = 0 the inte-

gral behaves like

which is convergent. Near a: = 1 it behaves like

dx
/ V^l — X -2vT^,
which is also convergent. The given integral is therefore convergent.

dx
Example 2, ^
^0 (y.2
(x^ _
— l)i
This integral must be investigated at x = 1 and at the infinite limit.
The part of the integral near a; = 1 behaves like

— I (•-!)',
J {x l)i 2

which is convergent, The part corresponding to very large values of x

behaves like

which becomes infinite when x —» « . The given integral is divergent.

^ 7 ^ sina; ,
Example 3. I dx.
Jl X
We might try
sing
p{x) =
x
But, since
M
p{x)
= smx

does not tend to a limit as x tends to infinity, the above theorem is not
applicable. Whether this integral converges or diverges cannot be deter-
mined by our present methods.
§ 129] PRISMOID FORMULA 266

128. Numerical Integration. When the indefinite integral of a

function f(x) is unknown or too complicated for convenient use its
definite integral

dx (1)

between given limits can still be computed to any desired accuracy

by numerical methods.
One way to do this is to make direct use of the definition of a
definite integral as limit of a sum (§42).
Another way is to divide the interval of integration into parts
and on each part replace /(rc) by an approximating function which
can be easily integrated. The parts are usually taken of equal
length, and the approximating functions are usually polynomials.
129. Prismoid Formula. Given any second-degree polynomial,

fix) = A+Bx + Cx^, (1)

we can express the constants A, B, C, and so the integral

h

fix) dx,

in terms of the values the polynomial takes at three points. If we

take X = a, X == b, and the midpoint x = f (a 6) as the three +
points, the expression obtained for the integral is

as may
dx =
^ [/(a) + 4/ +
be directly verified by substituting the above value for/(x)
m ]
, (2)

on both sides. In fact by substituting

fix) = A + Bx + Cx^ + Dx^

this equation can be shown to hold even more generally when /(a:)
is any third-degree polynomial.
Since equation (2) can be used to determine the volumes of cer-
tain solids called prismoids, it is called the 'prismoid formula. When
fix) a polynomial of not more than the third degree it gives the
is

exact value of the integral on the left. When fix) is not of this
form it can be used to obtain an approximation for the integral,
namely the approximation that results when fix) is replaced by
the second-degree poljmomial which at a, b, and §(a b) has the +
266 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

same values as/(x). The accuracy of the approximation is deter-

mined by the accuracy with which /(a:) can be thus represented.
Example 1. Find the volume of a truncated cone, the radii of its bases
being ri, r2 , and its altitude h.

Cross sections of the cone perpendicular to the axis are circles of area

which is a quadratic function of the distance y along the axis. The volume

is thus determined exactly by the prismoid formula. Since the cross sec-
tions at the ends and middle are circles of radii ri, r2 , and | (ri +r 2) the
result is

» = ^
[*^1 + 4ir + jttI] = y [rf + (n + rj)* + ri].

Example 2, Find an approximate value for

by the prismoid formula.

Here
/(x) =
X

The values of this function at the limits of integration and the midpoint
are
= /(f) = i
The prismoid formula thus gives

0.694.

The answer correct to three decimals is 0.693 (compare §41, example).

§130] SIMPSON'S RULE 267

130. Simpson’s Rule. To obtain a better approximation for a

definite integral

dx

than that given directly by the prismoid formula we can divide

the interval of integration into any even number of equal parts,
apply the prismoid formula to successive pairs of parts, and add.

Let 2n be the number of parts, each of length

X\yX2 i
• ^2 n-i the points of division arranged
• in order between
Xq = a and X2 n = and

Vi — f(p^i^9 O 7
Ij * *
*)

the values of f{x) at those points. Since each of the intervals

(a, X2 ), (x 2 X4 )y etc., is of length 2fe, the prismoid formula gives
,

h
f{x) dx = - [2/0 + 42/1 + 2/2I ,

h
fix) dx = - [2/2 + 42/3 + 2/4]
I

/'*’ h
f(x) dx = - [j/2n-2 + 4y2 »-l + y2n] ,
I

whence by addition
* h
= + + + 4y3 + yan].
/ _
/(a:) da; -
3
[yo 2^2 H h 4j/2n-i
268 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

This is known as Simpson^ s rule. It gives an approximation for

the integral in terms of an odd number of ordinates, the distance
between consecutive ordinates being h. The bracket contains a
sum of multiples of these ordinates, the first and last coefficients
being unity and the others alternately 4 and 2. In general the
accuracy of the approximation increases with the number of divi-
sions.

Example. Find tt from the formula

dx
4 1 +
by dividing the interval into four parts and using Simpson^s rule.
The divisions are determined by
r
^ — n i i
4> 2>
5.
4) A*
1

The corresponding values of

y =
1 +
are
4,-1
2/
- 1, 16 4 16
tT> -5^)
1
2-

Substituting these and = i, Simpson^s rule gives

whence
I - S [‘ ® + <5) + + s] -

TT = 3.14157.

The value correct to 5 decimals is

IT = 3.14159.

PROBLEMS
Determine the values of the following integrals:

1
.
JVs* +2dx
dx. 2 .

J (2x - 3)*
*•/;va + — Vx dx.

X dx
j*Q?y/X* +
5.
I y/a^ — X*
6. 1 (ix.

7. yxiVa,! _ 1 dx.
^ f 2xdx
’
J (3x2 _ 2)2’
dx xdx
10
9.
+3
.

/ 2x2 - 1
 PROBLEMS 269

r (2x+Z)dx
J x^+Zx+Z
13. 2 cos (4aj -f 1) dx.
.A
14.
J
/*
Sin
sin
•
— -—
3a
-
5
2
dx.
J'
16. y* a; sin (x^ + 1) dx. 16.
/
X COS (ox^) dx.

« 9
scc
2 ~ (i0
/ .
'
y cos2 30

/-^
Jsin2 2d
sin^l
20. ^ X csc^ (x^) dx.

•
3a; +4 /*—;r*
22.
y tan 2x
cos 2a; ,

/ -——dx.sin 2a;
cos a; cte
24.
24.
J/ sec 50 tan 50
' dO.

sec 0 (sec 0 tan

/ X sin^
26.
J ' -f- 0) dJO.

AO rl + sin 2x
27. ^ — (esc 6 cot 0) CSC 0 do.

/—3x
dx /
;r' 30 .

COS J sin (3x 4- 2)

—J+ I sin X
dx.
,
32. + sec 0)^ do.
/ COSX dx
X csc^

/r
J X
1 -f 2 cot “•/r-
dt
36 /-====•
J ^2- V3
.

- 4x 2

^ r (2x +3)dx r dx
’

J V4 - ** x 2 4- 5
dx r dx
’

/ x^+1
'
-"
4 2x 2 4 3
r dx
41 /-^£=- 43 .
.

^ x^ X V — 4 y xV 4x^ — 9
dx
/xV 5x^ - 1 -/vfe-
-/vro- -/ot-
« /??« 1^3
•^

60 .
Je-*^dx.

-/?• 62 -
.
J (e*

sin 0 d0
e-*)*(to.

68. fxe-**dx. ^ ^
•/ V 1 — cos 0
270 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

« r 66.
f X dx
J V2 - W® J V — x^
r r c®dx
68.
y 1 + «** J 1
+«**
A /• COS 5 cto
60.
/• dt

J2+ > 1
61. ^ sin* X cos x dx. 62. y*cos^ 5x sin 6x dx.

63. (cos e + sin 6)^ dd. 64. y*sin® X dx.

J'
66. j*cos* 2x dx. 66. y*sin* X cos* X cte.

67. cos*a;dx. 68. y*sin* 4d cos* 4d dd.

1
•/

r sin* x dx f cos* X dx
69. / 70.
y 1 — cos X J sinx
—
71.
r cos* * cte 72. y*tan* X sec* X dx.
/ .
sinx
«/

y*sec* X tan* x dx. 74.

f sinx dx
73.
J cos* X
r sin* X dx •
76. y*sec^ X dx.
76.
y/ cos^
^
a;

y' (1
77.
J tan*x(ix. 78. +cotfl)*d9.

X tan* x dx. / csc^xdx.

79.
^ sec^ 80.
J
r cos* i dt
tan d CSC d
/ sm««
82.
J dd.

cos* 4x dx. y'Cl + cos d)* dd.

83.
^ 84.

86. ^(1 — sin x)* dx. 86. y*sin^xdx.

87. sin* 2x cos* 2x dx. 88. y*sin^ d cos* d dd.

89. y*cos* X dx. 90. y*cos X sin 2x dx.

91. y*sin X cos 3x dx. 92. y*sin 2x sin 3x dx.

98. y*cos 2x cos 4x dx. 94. y*sin* 2x cos 3x dx.

/• dx
96. y*seG X CSC X dx. 96.
y 1 — cos X

97. y* V 1 + cos 6 dd. 98. y*(l + cos d)i dd.

 PROBLEMS 271

1. 100.
J V2 + 2a; - •/ V1 4- 4a; - 4a;^

^ cte

^f V^ 2 + 6* - /
1
. 13'
3»* »* +6® +
I. f ^ MM. /•
y 2** - to + 6 J {x — 1) V a;^ — 2x — 3
^ /•
Jf {2x -
1. 100.
3)Va;2 - 3x + 1
r (2a; 4- 3) da;
3* 108
y a;2 4-2a; - y X* + 2x - 3
r (a; - 1) da; /
110.
J 4x^ — 4x + 2 y X* + 2x + 2
I
f
^ 112.
/•
•/ Vl 4- 4a; — 4a;‘'* Vx* + 2x + 3
r (2x — 1) cte
^ 114. r
’
^ V3a;2 - 6a; - 1 ^ (x* - 2x + 3)1
c® cZa;
lie
/ e2^ + 2e* + 3* ‘
y X* + X - 6'
/•(x»+x*)eir
“®-
y ?'-37+2'
y x» + 3x* + 2x
122. .
/*»-***'• y (X + D*

B ^ + 2) da;
. / a;^ — 4a; 4- 4
r Sdx
J X*- 23?'
- g*) (fa - 1) <fe
^ (1 /• (ae
128
y *(** + 1)
•
y (X + l)(x* + 1)’
4x dx
100. rit+llt..
/ J X® - 1

,
/• (x* + x)dx 182.
/
J 3?-^ y (X* + i)(x* + 2)
r Zdx
184 f
'
y X* + 6x* + 4 y (X* + l)(x* - 2x + 3)
r 3? dx

y (®*+4)*‘ * y x(x* + D*
r (x» + l)«fo
188. fl±^dx.
y (x* - 2x + 3)*’ J 1- Vi
r X dx
140. 1‘xV'x — a dx.
272 FORMULAS AND METHODS OF INTEGRATION [Chap. XI

f dx
141. 142.
J * +3 J X VX — 1

/• dx f x^ dx
143. 144.
J x+ 2Vx - 1 J 1 + xi
f dx
145. 146. \^1 4- Vx d:i..

J — x^ J*
r 0? dx f x^dx
147. 148.
J y/a^ — x^ J Vo® - *®

f dx f dx
149. 160.
J X Va? - x^ J xW o2 — x2
f dx f dx
161. 162.
J (o® — *®)^ J x2 V —
x2 a2

*=
163. f —
164.
J 7?y/ x^ J X

f dx
166. 166.
7 ** J xVx® - 0®
f dx
®Vx® +
167.
J a® dx. 168.
J x®Va® + X®
f dx /• X® dx
169. 160.
J (o® + *®)i J (a2 + x2)i
r x^ dx
161. 162. j*X cos X dx.
J (a2 + x2)2
163. j*X sin X dx. 164.
J* sec2 X dx.

166.
J*
X sec X tan x dx. 166.
Jxh^ dx.

167. x^ In X dx. 168. j'x®c®* dx.

J*
169. sin-^ X dx. 170. tan“^xdx.
J' J'
^In Vx* + a* X X
171. (x -f ) dx. 172.
J sec“"^ dx.

—
173.
^ (x l)2sinxdx. 174.
^Vx2 — a2 dx.

x^dx
176.
J' Vx^ -jra^dx. 176.
V Vx® - o®
177.
J*
sin 3x dx. 178. ^ e“* cos X dx.

179. fsin 3x cos 2x dx.

 180.
PROBLEMS 273

Derive the formula

„ sec"“^ X tan x n — 2 r „ ,
^xdx,
/ sec’* X =«
n —
:
1
1

n — 7
1 J
/ sec**

and use it to integrate

181.
sec^xdx.

Derive the formula

182.
and use it to integrate

f (a^ — x^)^ dx.

Jq

By solving the equation of the preceding problem for the integral on

the right and changing notation, derive the formula

x{a^ — 2n + 3 /(a*-
j' (a^ — x^)^ da; = — dx,
2a\n -f 1) 2a\n + 1)
and use it to integrate
da

/ (7^-
(a^ a;2)1

By finding the indefinite integral determine whether each of the following

integrals is convergent, and if convergent find the value of the integral:

r® X dx
8. 184.
Jq (a^ - a:*)^

dr
8. T--
J-ix^
186.
/-I ^6
/* 00

7. r^. 188. /
JQ
dx.

oo

190.
^ sin^ X dx.
/
«/i xixix
X In i •'0

Determine whether the following integrals converge:

dx dx
5-
h Vx * (** - l)s

/I " .
'
•'o
(x2 - l)y/x V sin e
196. r —-
•'-» (1
r* dx
j Vx^ —
1
x^)» 1

* dx
197.
JQ
+ 1
X x^ +X
274 FORMULAS AND METHODS OF INTEGRATION [Chap. XI
199. Show that the prismoid formula gives the correct volume for each of
the following solids: (1) sphere, (2) cone, (3) cylinder, (4) pyramid, (6) seg-
ment of a sphere, (6) ellipsoid generated by rotating an ellipse about its major
axis.

In each of the following examples compare the value given by the prismoid
formula with that obtained by integration:

/**
200. rVidx.
Jo
201 .
y-i ;
{x H- 4)2

r- r-*
X
202. sin dx, 203.
J“i* sec^ x dx,
Compute each of the following integrals by Simpson’s rule with four inter-
vals:
r2 dx
204.
J-2 1 4-
206.
J 1 + x^ dx.

206. j*
^
T
1 x^dx. 207. 1
Jo
Vsin a; dx.

208. V2 4- cosxda;. 209. V*’ dx.

210. Two parabolas with axes parallel to OF intersect and are tangent at
ic = + b). A curve y =» f(x) coincides with one of the parabolas between
« = a and x « |(a -f h) and with the other between x =» J(a + h) and x ^ h.
Show that

jT /(X) dx

is represented exactly by the prismoid formula.

 CHAPTER XII

FURTHER APPLICATIONS OF INTEGRATION

131. Area of a Surface of Revolution. The surface generated by

rotating an arc AR of a plane curve about an axis in its plane is
called a surface of revolution.
To define the area of such a surface divide the arc AB into parts
by points arranged in order from A to B, connect consecutive

pointsby chords, and determine the sum of the conical areas gen-
erated by rotating these chords about the axis. If this sum ap-
proaches a limit as the number of divisions is increased and all the
chords tend to zero, that limit is defined as the area of the surface.
Suppose, for example, that the arc of the curve

y = /(^)
between x = a and x = 6 is rotated about the x-axis. We assume
that y is positive and
dy
^f'ix)
dx

continuous on the interval (a, b).Divide the arc into parts, and
let P(x, y)f Q(x + Ax, y + Ay) be consecutive points of division.
276
276 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

The chord joining these points has the ordinate

J/m = y + i Aj/
at its midpoint. The area generated by rotating this chord about
the a;-axis is

2irymPQ = 2ir(y + | ^y)PQ-

Except for infinitesimals of higher order than Ax, this is equiva-
lent to
2Try As,

where As is the arc PQ. As the chords tend to zero the sum of the
areas they generate thus tends to the limit

S = lim 'E2iry As = (1)

A« -0
which is therefore the area of the surface of revolution.
If the curve is rotated about some other axis, this equation is

replaced by
S= J'2irr ds, (2)

where r is the distance of ds from the axis.

Since the limits of integration depend on the variable actually
used, no limits have been indicated in these formulas. To evalu-
ate the integrals, y and ds or r and ds must be expressed in terms
of a single variable t. The limits are then the values this variable
takes at the ends of the arc. Also it has been assumed that ds is

positive. The variable t and its limits must therefore be such that,
as t varies from the lower to the upper limit.

ds
ds = — dt
dt
never becomes negative.
Example 1, Find the area of the surface generated by rotating an ellipse
about its major axis.

Take as equation of the ellipse

and assume the major axis to be the a;-axis. Then a> b and
§131] AREA OF A SURFACE OF REVOLUTION 277

The entire surface is generated by the part of the curve above the a;-axis
between x = —a and x = a. Its area is therefore

= ~ —
^ = f ds h^)x^ dx.

Integrating as in Example 3, §124, and expressing the result in terms of

the eccentricity

® ^ y/a^
a
— —

we obtain

S = 27r6 [^ + “ sin~^ e j •

Example 2. Find the area of the surface generated by rotating the

curve
r = a(2 + cos B)
about the initial line.

In this case

da =
"v/ (^) +
^ ” a\/ 5 + 4 cos 0 dB^

2/
= r sin 0 = a(2 + cos B) sin B.

The entire area is generated by the part of the curve above the x-axis
between ^ = 0 and B — ir. Its area is therefore

S =J' 2iry da = (2 + cos B)\/ 5 + 4 cos 0 sin B dB.

To evaluate this we substitute

5 + 4 cos 0 = z^-

When B varies from 0 to ir, 2 varies from 3 to 1 and the integral becomes

S
4 •'1
(3 + z^)z^ dz = ^o
278 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

132. Area of a Cylindrical Surface. The locus of lines parallel

to a fixed direction and touching a fixed curve is called a
cylindrical surface.
We shall determine the
area of the cylindrical sur-
face generated by perpen-
diculars to a plane MN at
points of an arc AB, Let s
be the distance measured
along AB to the variable
point Py and z the length
of the perpendicular at P.
We assume that AB is a
Figure 174. smooth curve and z a con-
tinuous function of s. The
part of the cylindrical surface between the perpendiculars at s and
s + As is then approximately a rectangle of area z As, and the
required area is

S = lim As = z ds.
As -*0 JI
To evaluate this integral z and ds must be expressed in terms of a
and the integral
single variable
taken between the limiting
values of that variable.

Example, Find the area cut

from the lateral surface of a right
circular cylinder by a plane
through the center of the base
making the angle a with the
base.
We are to determine the area
cut from the lateral surface of the
cylinder (Figure 175) by the plane
ABC, Let P be any point on the
arc forming the base of the re-
quired area, 0 the center, and

e = AAOP,
The height of the perpendicular
at P is
PQ « MP tan a » a sin ^ tan a. Figure 175.
§133] MEAN VALUE OF A FUNCTION 279

where a is the radius of the cylinder. Along the arc AB,

ds = add,

and the whole arc is described when 6 varies from 0 to tt. The required
area is therefore

sin 6 tan adS — 2a^ tan a,

133. Mean Value of a Function. Between x = a and x = h let

y = fix)
be a continuous function of x. Divide the interval from a to 6
into n equal parts
h — a
Ax =
n
and let yi, y2 j
be the values of 2/ at x = o and the points
•
•
•> 2/n

of division. The average value of these n numbers is

yi + y2 + ys \-yn

n
Multiplying numerator and denominator by Ax, this fraction be-
comes
b b

Ax y^J(x) Ax
yi Ax + y2 Ax \- yn Ax a a

n Ax

As n increases, this average value tends to the limit

which is called the mean value or average value of /(x) with respect
to X between x == a and x = 6.

This may
be regarded as a sort of average of all the ordinates
of the curve y = /(x) between x = a and x = 6, but it should be
noted that the result depends on the variable x in terms of which
y is expressed. In the above discussion we have taken ordinates
at equal intervals along the x-axis and have obtained as final result
280 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

the average value of y with respect to x. If y is also expressible

in terms ofsome other variable t and we take ordinates at equal
intervals A/, the resulting average with respect to t need not have
the came value.

Example. Find the average value of

o*
y ~ + 3-2

between x = — o and x = o, (1) if the average is taken with respect to

Xf and (2) if
a;= a tan 6 and the average is taken with respect to d.
If the average is taken with respect to x, its value is

ry^
•/-a
_ X^ ^ TT

2a ”4 “
J-a^
If the average is taken with respect to 6^ the result is

r T

f\ydd a cos^ 6 dO
4 4 TT +2

134. Moment If a force F acts in a plane perpendicular to an

axis AB (Figure 176), its
momentf or torqite, about the
axis is the product

M= FI,

where I is the lever arm, or

perpendicular distance, from
the axis to the line of the
force. If a body is acted on
by several forces, the total
moment about a given axis is
the sum of the moments of
the separate forces, a moment being considered positive when
the force tends to produce rotation in one direction about the axis
and negative when it tends to produce rotation in the other di-
rection.
The term moment is also applied to quantities other than forces,
the moment being the product of the magnitude of the quantity
§1341 MOMENT 281

by the distance of its position or point of application from a line

or a plane.
To find the moment of a plane area, length, or mass, divide it

into small parts such that the points of each differ only infinitesi-
mally in distance from an axis in the plane. Multiply each part
by the distance of one of its points from the axis, the distance
being considered positive for points on one side of the axis, nega-
tive for points on the other side. The limit approached by the
sum of these products when the parts are taken smaller and
smaller is called the moment of the length, area, or mass with
respect to the axis.
moment of a length, area, volume, or mass
Similarly, to find the
in space with respect to a plane, we divide it into elements whose
points differ only infinitesimally in distance from the plane and
multiply each element by the distance of one of its points from
the plane, these distances being considered positive for points on
one side of the plane and negative for points on the other side.
The moment with respect to the plane is the limit approached by
the sum of these products when the elements are taken smaller
and smaller.

Example 1, A rectangular floodgate is 10 feet broad and 6 feet deep.

Find the moment of water pressure about the ))ase line when the water
level is at the top of the gate.
Divide the rectangle into strips
of width by horizontal lines.
Neglecting infinitesimals of order '

6
higher than A/i, the force due to A/^
water pressure on the strip be-
tween the depths /i. A -f- is
-10 -
AA,
Figure 177.
w being the weight of a cubic
foot of water. About the base line this force has the lever arm

Z = 6 - /i

and moment
10iy^(6 — h) Ah
approximately. Thus the total moment is

lim S10ii?/i(6 — h) Ah = f lOwhifi ^ h) dh = 360m?.

Ah -*o *'0

Example 2, In the preceding problem find the moment of the area

about the base line.
282 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

The strip between the depths A, A + has the area

AA = 10 Ah.

All points of this strip are at distance approximately

Z = 6 -
from the base line. Thus the moment of the area about the base line is

lim 210(6 - /i) A/i = dh = 180.

AA 0

136. Center of Gravity of a Plane Distribution, Centroid. The

center of gravity of a system of masses in a plane be defined may
as the point at which these masses could be concentrated without
changing their total moment with respect to any axis in the plane.
Proof that there is such a point may be left as an exercise. An
important consequence is the fact that in finding the moment of
a body about an axis it may be cut into parts and each part con-
sidered as concentrated at its own center of gravity.
Let C(Xj y) be the center of gravity of a continuous mass. This
may be a thin wire bent around a curve or a thin plate overlying
an area. Cut the mass into pieces, and let Xy y be the coordinates
of the center of gravity of the piece of mass Am. The moment
of this piece about the x-axis is y Am, and the total moment of the
whole body about the axis is

lim Am = dm.
J/
2/
Am -» 0

The mass of the body is

If this were concentrated at the center of gravity, its moment with

respect to the x-axis would be My. By definition we thus have

whence

( 1)
§136] CENTER OF GRAVITY OF A PLANE DISTRIBUTION 283

Similaxly,

dm
X = ( 2)
dm

These formulas show that the rectangular coordinates of the

center of gravity are the mean
coordinates, the average being taken <

with respect to the mass.

For a thin wire (Figure 178) we
take G
dm = p ds, (3)

where p
and ds
is the mass per unit length
the element of arc. In
f y

is 0 X
particular, if p is constant, it may
be canceled from numerator and Figure 178.
denominator, leaving

J*X ds (

, y = J* ~j- (4)

ds ds
I I

as the center of gravity of a thin wire of constant cross section

and density.

For a thin plate (Figure 179) we take

dm = pdA, (5)

where p is the mass per unit area and dA is the element of area.
284 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

In particular, if p is constant, it may be canceled, leaving

as the center of gravity of a thin plate of constant thickness and

density.
The term center of gravity has its origin in the fact that if the
parts of a body are subject to parallel forces proportional to their
masses (as under gravity) the resultant of these forces passes
through the center of gravity. Expressions (4) and (6), however,
also occur frequently in discussions that do not involve mass or
force. In such cases (4) is usually called the centroid of the arc
and (6) the centroid of the area, but the expressions center of grav-
ity of arc and center of gravity of area are also frequently used
even when there is no association with gravity.

In these integrals x, y are the coordinates of the center of gravity

or centroid of the element dm or dA, Before integration all quan-
tities must be expressed in terms of a single variable and proper
limits inserted.
If a body is symmetrical with respect to an axis, its moment
with respect to that axis is zero. The center of gravity is thus
on an axis of symmetry.

Example 1, Find the center of gravity of a thin wire of uniform cross

sectionand density bent into a quadrant of a circle of radius a.

Example 2. Find the centroid of the area of a quadrant of a circle of

radius a.
§136] CENTER OF GRAVITY OF A PLANE DISTRIBUTION 285

Take the element of area perpendicular to OX (Figure 181). If Xy y

are the coordinates of the point P at the top of this element, its centroid
is (x + \dXy \y) and its moment with respect to the x-axis is

if dx.
I ^
The moment of the entire area about the x-axis is then

Figure 181. Figure 182.

Example 3, A square plate of uniform thickness and density has a

quadrant of a circle cut from one corner. Find its center of gravity.
Let h be the side of the square, a the radius of the circular quadrant,
and take the coordinate axes as indicated in Figure 182. In the formula

^ =
f ydA
-7—
fiA
the denominator is the difference between the area of a square and a quad-
rant of a circle and the numerator is the difference of their moments about
the x-axis. Since the moment of the square is
286 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

and equation (8) of the preceding example gives

w
as that of the circular quadrant, we have
- §0
*

5 =
4

as the coordinates of the center of gravity of the plate.

Example Find the centroid of the area enclosed by the cardioid

Use the polar element of area

dA = dd.

This element is approximately a triangle with centroid two-thirds of the

way from the vertex to the base. The coordinates of this point are thus
approximately
a; = |r cos 6^ y = \r sin B,

r, B being the coordinates of the point P(r, B) on the cardioid (Figure 183).
Hence
fxdA 30*(1 + cos By cos B dB
^
^ ~ Z ” /•air ” A
fdA f io*(l + cos oy dB ®

By symmetry it is clear that y = 0.

136. Center of Gravity in Space. The center of gravity of a

system of masses in space may be defined as the point at which
these masses could be concentrated without changing their total
moment with respect to any plane.
§136] CENTER OF GRAVITY IN SPACE 287

Thus to find the center of gravity of a continuous mass cut it

into slices of mass Am. the center of gravity of such

If (x, y, z) is
a slice, its moment with respect to the xy-plane is z Am and the
moment of the whole mass is

Am
lim
0
Am =/ z dm.

If the whole mass were concentrated at the center of gravity

(^} V) ^)i its moment with respect to the xy-plane would be

z
J"dm.
By definition we thus have

z
J'dm = J*z dm.
whence

z dm
z = (1)

/-dm
Similarly,

X dm y dm
X = —
J*
,
y = —
J*
(2)

/ dm I
dm

To find the center of gravity, or centroid, of an arc, surface, or

volume merely necessary to replace the element of mass in
it is

these formulas by the appropriate element of arc, area, or volume

and insert the proper limits.

Example 1, Find the center

of gravity of a homogeneous
hemisphere of radius a.

By symmetry the center of

gravity on the radius perpen-
is

dicular to the plane face.Cut

the hemisphere into slices by
planes perpendicular to this
radius (Figure 184). The slice
between the planes at distances
288 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

2, z + dz from the plane face has approximately the volume

dv = TTX^ dz = 7r(o^ — z^) dz

and approximately the moment

zdv = 7rz(a^ — z^) dz

with respect to the plane 2 = 0. The center of gravity of the hemisphere

is thus on the radius perpendicular to the plane face at distance

J' zdv jT 7r2(a^ — 2 ^) <

from the center.

Example 2, Find the center of gravity of a thin hemispherical shell of
radius a and uniform thickness and density.
Divide the shell (Figure 184) into strips by planes parallel to its base.
Each of these strips may be considered as generated by rotating an arc As
of the circle
+ z"^ — (J?

about the vertical axis. Since the differential of arc on the circle is

ds = \/l + dz = - dz,
X
the area of the strip is

dS = 2wx ds = 27ra dz.

The center of gravity is on the axis of the shell at distance

J'zdS jT 27razdz

J dS £ 2Tradz

from the center.

137. Theorems of Pappus. Theorem I. If the arc of a plane

curve is revolved about an axis in its plane and not crossing the arc,
the area generated is equal to the product of the length of the arc and
the length of the path described by its centroid.

Let the arc be rotated about the rc-axis. The ordinate of the
centroid is

s
§1371 THEOREMS OF PAPPUS 289

where 5 is the length of the arc. Thus

2Trys = 2Ty ds.

J*
The right side of this equation represents the area generated, and
2Try is the length of the path described by the centroid. This
equation therefore expresses the result to be proved.

Theorem 1 1. If a plane area is revolved about an axis in its plane

and not crossing the area, the volume generated is equal to the product
of the area and the length of path described by its centroid.

Let the area A be revolved about the x-axis. Cut it into strips
dA by lines parallel to the x-axis. The ordinate of the centroid is

ydA
y =
I
whence
2xyA 2Try dA,
-f
The right side of this equation represents the volume generated,
and 2iry is the length of path described by the centroid. The
equation therefore expresses the result
to be proved.

Example 1, Find the area of the torus

generated by revolving a circle of radius a
about an axis in its plane at distance h (greater
than a) from its center.

Since the circumference of the circle is 27ra

and the length of the path described by its

center 2Trh, the area of the torus is

S = 27ra-27r6 = 47r^a&. Figure 186.

Example 2, Use the theorems of Pappus to determine the centroid of

a semicircular area.
When a semicircle of radius a is revolved about its diameter, it gener-
ates a sphere of volume If y is the distance from the centroid of
the semicircle to this diameter, and A = f Tra^ is its area, by the second
theorem of Pappus
•jTra® = 2TryA = rc^ya^,
iVhence
Atto? 4a
290 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

Example 3, Find the volume generated by revolving the cardioid

r = a(l + cos d)
about the initial line.

The polar element of area OPQ (Figure 186) is approximately a triangle

with area
Ir^dB
and centroid at distance
y — \r sin 0

from the initial line. By the second theorem of Pappus the volume gen-
erated by OPQ is then approximately

2Try-^r^ dB = |7rr^ sin B dS,

The entire volume generated is therefore

i; = fTrr* mi 6 dB = ^

(1 + cos B) Tra®.
=[-i Jo” 3

138. Moment of Inertia. The moment of inertia of a particle

about an axis is the product of its mass and the square of its dis-
tance from the axis.
To find the moment of inertia of a continuous mass, we divide it

into parts such that the points of each differ only infinitesimally
in distance from the axis. Let Am be the mass of such a part and
r the distance of one of its points from the axis. The moment of
inertia of this part is then approximately Am, and the moment
of inertia of the entire mass is
§138] MOMENT OF INERTIA 291

If the mass of a body is M and its moment of inertia about a

given axis is

/ = Mk^ (2)

the distance k is called the radius of gyration of the body about

that axis. It is the distance from the axis at which the body could
be concentrated without changing its mo-
ment of inertia about that axis.
As an illustration of these formulas we
take the moment of inertia of a homogene-
ous cylinder about its axis. To determine
this cut the cylinder into pipe-shaped sec-
tions (Figure 187). Let a be the radius, h
the altitude, and p the density. The sec-
tion with inner radius r and outer radius
r + dr has the volume
dv = 2Trrhdr
and mass
dm = pdv.

The moment of inertia of the cylinder is therefore

I = J"r^ dm = 2irphr^ dr = ^irpha!^,

J*
and its mass is

M= TTpa^h,

Thus
I = ^Ma^ (3 )

is the moment of inertia of a homogeneous cylinder of mass M

and radius a about its axis. Comparison with (2) shows that the
radius of gynation in this case is

In certain discussions quantities appear that are similar to mo-

ments of inertia but with mass replaced by length, area, or volume.
Such quantities are usually called moments of inertia although
they may have no connection with inertia. Thus we have moments
of inertia of length, area, or volume about an axis, these being
represented by integrals of the form (1) but with dm replaced by
an element of length, area, or volume.
292 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

Example 1. Find the moment of inertia of the area of a circle about a

diameter of the circle.

Let the radius of the circle be a, and take the a;-axis along the diameter
about which the moment of inertia is taken. Divide the area into strips

by lines parallel to the a;-axis. Neglecting infinitesimals of higher order

than the area of each strip is 2x Ay, and its moment of inertia is

2xy^ Ay. The moment of inertia of the entire area is therefore

I 2xy^dy = 2j^y^\/ — y^ dy =

Example 2. Find the moment of

inertia of a right circular cone of con-
stant density about its axis.

Let p be the density of the cone, h

its altitude, and a the radius of its base.

Divide the cone into hollow cylindrical

(pipe-shaped) slices by cylindrical sur-
faces having the same axis as the cone.
By similar triangles the altitude of the
cylindrical surface of radius r is

2/
= - (a - r).
a

Neglecting infinitesimals of higher order

than Ar, the volume between the sur-
faces of radii r, r + Ar is

Av = 2Try Ar = r(a — r) Ar,

 PROBLEMS 293

the mass of this slice is p Av, and its moment of inertia about the axis is

pr^ At;. The moment of inertia of the entire cone is therefore

I = lim Spr^ Av = f r^(a — r)^ dr = »

Ar-^o a 10

The mass of the cone is

M — pv — ^irpd^h.

I = ^Ma\
Example 3. The area bounded by the parabola
2/2 = 4a:,

the oj-axis, and the line a: = 4 is rotated about the a:-axis. Find the mo-
ment of inertia about the a:-axis of the
volume generated.
The element of area y dx generates a
cylinder of radius y and volume

dv = iry^ dx.

By equation (3) its moment of inertia

about the x-axis is 01

dv = dx = Stts:'* dx. Figure 190.

Z A
The moment of inertia of the entire volume is therefore

f\irx^dx = [fw’lj =

PROBLEMS
1. Find by integration the lateral area of a right circular cone of altitude h

and radius of base a.

2. Find the area of the paraboloid of revolution generated by rotating
about the x-axis the part of the parabola = 4px between x = 0 and x = p.
3. Find the area generated by rotating about the y-axis the part of the
parabola x^ = Zy between x = 0 and x = 2.
4. Find the area of a zone on a sphere of radius a betw^een two parallel
planes at distance h apart.
6. Find the area of the surface generated by rotating about the p-axis the
part of the curve
8y = 9 - at*

above the x-axis.

6. The arc of the curve
a? = iy

between x 0 and x = 1 is rotated about the x-axis. Find the area of the
surface generated.
 7.
294 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

Find the area of the surface generated by rotating about the ^-axis the
arc of the tractrix
8.
a,
In
a + Va^ — a;*
— \/a* — »*
V
2
9.
between the points (xi, 2/1), to, yi)-
Find the area of the sufrace generated by rotating the cardioid

r = a(l “h cos 0)
about the initial line.
Find the area of the surface generated by rotating the circle r =* o sin 0
about the initial line.
10. Find the area of the surface generated by rotating the lemniscate

= 2a^ cos 20
about the initial line.
11. Find the area of the surface generated by rotating the lemniscate

= 2o^ cos 26 about the line ^ ~ ^ .

a
12. Find the area of the surface generated by rotating the hypocycloid

X = o cos* y — a sin* (f>

about the x-axis.

13. Find the area of the surface generated by rotating one arch of the
cycloid
X =» a(<t> — sin <f>), y « a(l — cos <t>)

about the x-axis.

14. The arc of the hyperbola

X ~ a sec 0, y a tan

between 0 = 0 and 0 =7 is rotated about the y-axis. Find the area of the
4
surface generated.
18. The arc of the curve

X = a(cos 0 + 0 sin 0 ), y = a(sin 0 — 0 cos 0)

between 0 = 0 and 0 = is rotated about the x-axis. Find the area generated.
16. At points of the curve y^ = 4px, lines perpendicular to its plane of
length h y are drawn. Find the area of the surface formed by these lines
at points of the curve between (0, 0) and (p, 2p).
17. At each point of the circle r = a cos ^ a perpendicular to its plane is
erected of length equal to r. Find the area of the cylindrical surface formed.
18. At points of a circle of radius a perpendiculars to its plane are erected,
the perpendicular at each point P being of length ks, where s is the arc of the
circle from a fixed point A to P. Find the area of the surface formed by the
perpendiculars along the arc beginning at A and extending once around the
circle.

19. The axes of two right circular cylinders of radius a intersect at right
angles (Figure 52, page 90). Find the area of the solid common to the two
(^rlinders.
 PROBLEMS 295

20. A wedge
cut from a right circular cylinder of radius a by a plane
is

inclined 45® to the base. Find the area of the curved surface of the wedge if
the circular arc that forms its lower edge is an arc of 120®.
21. The
angle between the axis of a cone and its generators is 45®. If the
vertex of the cone is on the base of a cylinder of radius a and its axis is a

generator of the cylinder find the area of the cylindrical surface below the
cone.
22. One diameter of a circular cylinder of radius a is a radius of a sphere.
Find the area of the cylindrical surface inside the sphere.
23. Find the mean value of 2/ = sin x with respect to x between a; =* 0 and
rc = TT.

mean value of sin^ x from x = 0 to » = 27r, the average being

24. Find the
taken with respect to x,
25. Find the average value of sin 2a; sin 3x for a common period of sin 2a;
and sin 3a;, the average being taken with respect to x,
26. At points of a semicircle of radius a perpendiculars are drawn to the
diameter through its ends. Find the mean value of these perpendiculars (1)
if they are taken at equal intervals along the diameter, (2) if they are taken

at equal intervals along the semicircular arc.

27. A body falls from rest under the acceleration of gravity. Find its mean
velocity during the first t seconds if the average is taken with respect to the
time. Show that the distance it falls in t seconds is this mean velocity multi-
plied by t,

28. A body falls from rest.

Find its mean velocity during the first t seconds
if the average taken with respect to the distance.
is

29. From a point on a circle of radius a, lines are drawn to points equally
spaced along the circumference. Find the mean of the squares of these lines.
30. Plane sections of a sphere are taken perpendicular to a diameter at
points equally spaced along the diameter. Find the average area of section.
Show that the volume of the sphere is this average area multiplied by the
diameter.
31. A particle moves along a straight line in simple harmonic motion

x ^ a sin kt.

Find the mean square of the average being taken with respect to x,
its velocity,

32. An amperes flowing through a resistance of R ohms

electric current of I
produces heat at the rate of RP watts. Find the average rate at which heat
is produced by an alternating current

/ =s A sin id
2ir\
flowing through a resistance of R ohms during one cycle
I
li*0to^* — I.

Find the steady current / which produces the same total heat during one
cycle.
33. A cubic foot of air is compressed from atmospheric pressure (14.7
pounds per square inch) to one-tenth its initial volume. If the gas satisfies
Boyle^s law jw ** 4;, find the mean pressure, the average being taken with
respect to the volume. Show that the work required to compress the gas is

this mean pressure multiplied by the change in volume.

296 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

34. The wind produces a pressure ofp pounds per square foot on a door
6 feet wide and h feet high. Find the torque tending to turn the door on its
hinges.
86. Find the moment due to water pressure on a rectangular flood gate of
width h and height h about a horizontal line through its center when the water
level is at the top of the gate.
36. A dam has the form of a trapezoid with base 100 feet, width at top
400 feet, and height 30 feet. Assuming that the water is level with the top
of the dam, find the moment of water pressure about the base line.
37. The diameter of a horizontal water main is 6 feet. One end is closed
by a bulkhead and the other is connected with a reservoir in which the water
surface is 30 feet above the center of the main. Find the moment of water
pressure on the bulkhead about its horizontal diameter.
38. A triangular plate of constant thickness and density has a base h and
altitude h. Find the distance of its center of gravity from the base. Take
the element of area parallel to the base.
39. A trapezoid has two parallel sides of lengths 10 feet and 20 feet, and
their distance apart is 12 feet. Find the distance from the 20-foot side to
the centroid of its area.

40. Find the centroid of the area bounded by the parabola — 4x and the
line X — 4.
41. Find the centroid of the area above the z-axis bounded by the ellipse

46.

42. Find the centroid of the area bounded by the a;-axis and the arc of the
curve y = sin a; between a; = 0 and x — t.
43. Find the centroid of the area bounded by the coordinate axes and the
parabola
ill
x^

44. The two arms of a steel square are 2 inches and 1J inches wide. If the
outer edges are 24 inches and 12 inches long, find its center of gravity.
From a semicircle of radius b a semicircle with the same center and
radius a is cut. Find the centroid of the area left.
46. Find the centroid of the area bounded by the parabolas

y^ — ax, x^ — ay.

Take an element of area parallel to a coordinate axis and extending between

the parabolas.
47. Find the centroid of the area bounded by y 2x — x^ and y — x,
48. Find the centroid of the area on the right of the y-axis bounded by the
curves
y ^ 7? — 4x, y = 2x —
49. Find the centroid of the arc of the curve

aj§ 4-
above the :r-axis.
 PROBLEMS 297

50 Find the centroid of the arc of the curve

.

61.
X = cos y » sin <
between t « 0 and t * -J-tt.
Find the centroid of the area between the x-axis amd one arch of the
cycloid
X * a{<l> — sin </>), y *= a(l — cos

52. Find the center of gravity of a thin wire bent into the form of the arch
of the cycloid in the preceding problem.
53. Find the centroid of the area within the cardioid

r = a(l 4" sin 6).

64. Find the center of gravity of a thin wire bent into the form of the
cardioid in the preceding problem.
65. A circular sector has radius a and central angle 2a. Find the centroid
of its area.
66. A wire is bent into an arc of a circle of radius a and central angle 2a.
Find its center of gravity.
67. Find the centroid of the area within the curve

r = a(2 — cos B).

68. Find the centroid of the area bounded by the parabola

r = a sec^ -

and the lines 0 = .

tj

69. Find the centroid of the area within one loop of the lemniscate

= 2a^ cos 20,

60. Find the center of gravity of a right circular cone of altitude h and
constant density.
61. Find the center of gravity of a thin shell of constant thickness covering
the surface of a right circular cone of altitude h.
62. The area bounded by the curve = ax and the line x = a is rotated
about the x-axis. Find the centroid of the volume generated.
63. Find the centroid of the volume generated when the area above the
x>axis in the preceding problem is rotated about the y-axis.
64. The area in the first quadrant bounded by the ellipse

isrotated about the y-axis. Find the centroid of the volume generated.
66. The area above the x-axis bounded by the circle x^ y* = and the +
straight line x 4- y = a is rotated about the x-axis. Find the centroid of the
resulting volume.
66. A hemispherical shell of constant density has the inner radius o and
outer radius 6. Find its center of gravity. Consider it as the difference of
two hemispheres (compare example 1, §136).
208 FURTHER APPLICATIONS OF INTEGRATION [Chap. XII

67. By finding the limit as h approaches a in the preceding problem find the
centroid of a hemispherical surface.
68. If Boyle’s law holds, the density of air at height h is

P ^ Poe

po and k being positive constants. Assuming this, find the center of gravity
of a vertical cylinder of air reaching from the surface of the earth to infinity.
69. Under the adiabatic law the density of air at height h is
p ^ a(b - A)"*,

a, b, m being positive constants. Assuming this, find the center of gravity of

a vertical cylinder of air reaching from the earth’s surface to the top of the
atmosphere.
70. The center of pressure on a submerged plane area is defined as the
point at which the total force could be concentrated without changing its
total moment about any Find the center of pressure on a
axis in the plane.
vertical rectangle of height h and width b with its upper edge in the surface
78.
of the water.
71. Find the center of pressme on a triangle of base b and altitude h if its
base in the surface of the water and its altitude is vertical.
is
76.
72. Show that the force due to liquid pressure on a submerged plane area
isequal to the product of the area by the pressure at the center of gravity.
As an application determine the force on a triangle of base b and altitude h
with its vertex in the surface and altitude vertical.
By using the theorems of Pappus find the lateral area and the volume
of a right circular cone.
74. Find the volume generated by rotating a circle of radius a about an
axis at distance b (greater than a) from its center.
By the theorems of Pappus find the centroid of a semicircular arc.
76. A groove with cross section an equilateral triangle of side 1 inch is
cut around a shaft 6 inches in diameter. Find the volume of metal cut away.
77. A groove with cross section a semicircle of diameter 1 inch is cut around
a shaft 6 inches in diameter. Find the volume of metal cut away (cf. ex-
ample 2, §137).
78. The length of an arch of the cycloid

z — a(<t> — sin 0), y =» a(l — cos 4)

is 8a, and the area generated by rotating it about the a>-axis is

—-— . De-
o
termine the centroid of the arch, and then find the area generated by rotating
it about the tangent at its highest point.

79. Find the area generated by rotating the arc in the preceding problem
about the ^-axis.
80. A circular sector is bounded by the circle r — a and the radii ^ a,
8 * (a > jS > 0). By the method of example 3, §137, find the volume
generated by rotating this sector about the initial line.
81. Find the centroid of the volume in the preceding problem. Use as
elements of volume the hollow cones generated by rotating the polar elements
of area and the fact that the centroid of a cone is three-fourths of the way
from its vertex to its base (cf. problem 60).
 PROBLEMS 299

82. Find the volume generated by rotating about the initial line the area
within the curve r =» o(2 — cos B).
83. Find the volume generated by rotating the area within the lemniscate
84.
= 2a^ cos 2d
86.
about the initial line.
The sides of a rectangle are a and b. Find the moment of inertia of
its area about the side of length a.
Find the moment of inertia of the area of a triangle of base b and alti<
tude h about the axis through its vertex parallel to its base.
86. A thin circular plate has radius a and mass M. Find its moment of
inertia about a diameter.
87. Find the moment of inertia of the area within the ellipse

about the x-axis.

88. Find the moment of inertia of a uniform straight wire of length I and
mass M
about an axis perpendicular to the wire at one end.
89. A uniform wire of mass M
is bent into a circle of radius a. Find its
moment of inertia about a diameter of the circle.
90. The rim of a flywheel has inner and outer radius n, r2 and its mass ,

is M. Find its moment of inertia about its axis.

91. The area within the ellipse

1
96.

is rotated about the x-axis. Find the moment of inertia of the resulting
volume about the x-axis.
92. The area bounded by the x-axis, the curve y — a? and the line x = 2
is rotated about the x-axis. Find the moment of inertia of the resulting vol-
ume about the x-axis.
93. A thin conical shell of constant thickness has the altitude h, radius of
base a, and mass M, Find its moment of inertia about its axis.

94.
98. Find the radius of gyration of a uniform spherical ball of radius a about
a diameter.
Find the radius of gyration of a thin spherical shell of radius a and
constant thickness and density about a diameter.
96. A ring is generated by rotating a circle of radius a about an axis at
distance b > a from its center. Find the radius of gyration of the ring.
97. The kinetic energy of a moving mass is

J* dm,

where v is the velocity of the element of mass dm. Find the kinetic energy of
a homogeneous cylinder of radius a and mass Af rotating with angular velocity
CO about its axis.

Find the kinetic energy of a sphere of radius a and mass rotating M

with angular velocity co about a diameter.
 CHAPTER XIII

SERIES WITH REAL TERMS

139. Convergence. If mi, W2 »
•••»«»») • • • is a given infinite se-
quence of numbers, the expression
00

Un = Ui + U2 + Uz -\
V Un -\
, (1)
n^l
often abbreviated Swn, is called an infinite series and wi, U2, Uzj
etc., are called its terms. For the present we assume all the terms
real.
If the sum
Sn = ^^1 + ^2 + W3 + • •
•
+ Wn (2)

of the first n terms tends to a limit as n 00 ,

the series is said to
converge and the limit
S = lim Sn (3)
n —» 00

is called its sum. We use the notation

#S = Wi + W2 + ^3 + * •
*
+ + * • •

to indicate that the series converges and that S is the sum.

Thus, if r 1, the first n terms of the series

a + ar ar^ h ur” + • •
• (4)
have for sum
1 -
/Sn = u + ur + ar^ H f- ar” ^ = a (5)
1 — r
If I
r I
<1 this tends to the limit

a
S = (6)
1 — r

as n increases. The series (4), called a geometric series is thus

convergent when |
r |
< 1 and has for sum the value S determined
by (6).
A series that is not convergent is called divergent
300
§14il POSITIVE SERIES 301

Suppose, for example, that the constant a in (4) is not zero. If

r = 1 the series becomes

and the sum of n terms, Sn = na, obviously does not tend to a

limit as n increases. If r = —1 the series takes the form

a — a + a — a***dba=Fa-**,
and Sn has the value a if n is odd and 0 if n is even. In this case
Sn also does not tend to a limit. Finally, if r > 1, equation (5) | |

shows again that Sn does not tend to a limit as n — oo The geo- .

metric series therefore diverges when a is not zero and r ^ 1. | |

140. General Condition for Convergence. The condition that

the partial sum Sn tends to a limit as n increases is merely that
the differences Sm — Sn become as small as you please when m
and n are sufficiently large [§14, (1)]. Since

Sn-\-p Sn ^ -j" * *
* 4” ^n+p> (f)

the general condition for convergence is thaty however small the posi-
tive number e may bey there is an integer N such thaty if n > N and
p is any positive integer

I
UnJfl + Wn+2 H f" ^n-fp |
< (2)

This amounts to saying the condition for convergence is that

there be a place in the series beyond which the sum of any number
whatever of consecutive terms is in absolute value less than a pre-
assigned positive c. In particular this requires that any single
term beyond that point be in absolute value less than e. Thus,
The terms Un of a convergent series tend to zero as n—^^.
The converse of this is, however, not true. That is, the terms
of a series may tend to zero although it does not converge. Thus,
the terms of
1 1 1
1 + +
2 3 n
tend to zero as n increases, but we shall presently see [§142, (7)]
that this series does not converge.
141. Positive Series. The simplest series are those in which all
terms have the same algebraic sign. In particular we call the
series positive if each of its terms is positive or zero.
302 SERIES WITH REAL TERMS [Chap. XIII

The sum Sn n terms in a positive series increases, or

of the first
at least does not decrease, as n increases. This sum therefore tends
to a limit if it is bounded [§14, (2)]. As an alternative to the gen-
eral convergence condition [§140, (2)] we thus have:
A necessary and sufficient condition for the convergence of a posi--

tive series is that there exist some number A such that

Sn<A
for all values of n.
Whether a positive series converges or not is usually determined
by comparing it with an integral or with another series that is
known to converge or diverge. In making these tests it should
be noted that a finite number of terms in a series may be omitted
without influencing convergence, provided these terms have finite
values.
142. Integral Test. The nth term in a series is a function of n,
say
Un=f(n). (1)

The series can thus be written

/(I) + /(2) + /(3) + . .

+ fin) + • •
•. (2)

This function fin) is usually defined not only for integral values
but for all values of n. With the series we can therefore associate
an integral

dx. (3)

In many cases the convergence of the series can be deduced

from that of the integral, the basis for this deduction being the
following theorem:
If f(pd positive and always decreases as x increases^ the series

(2) and the integral (3) both converge or both diverge.

To prove this we note that since fix) is a decreasing function

n+l ^n+1
- Six)] d* > - Sin + >
/ l/(n) 0, /
*/n
[Six) 1)] dx 0,

whence
^n+1
Sin) > Six) dx > Sin + 1). (4)
§142] INTEGRAL TEST 303

Writing this inequality forn = 1, 2, 3,

• •
•
, p, and adding, we get
P+1

/ f(x) dx

>/(2)+/(3)+--- + /(p + l). (5)

Since f{x) is positive, the integral

^p+i
fix) dx
J
and the sum
/(l)+/(2)+... + /(p)

both increase as p increases and so both tend to limits if they are

bounded [§14, (2)]. The left side of (5) shows that the integral
isbounded if the sum is, and the right side shows that the sum is
bounded if the integral is. Thus the series and the integral both
converge or both diverge.
As an illustration take the series

1 1 1
( 6)
2P 3i> nP

where p is any positive number. The nth term

is positive and decreases as n increases. The series is thus con-

vergent if the integral

is convergent, that is if p > 1, and divergent if the integral is

divergent, that is if p ^ 1. This discussion shows, in particular,

that the series
1 1 1
1+ + (7)
2 3 n
is divergent.
304 SERIES WITH REAL TERMS [Chap. XIII

143. Comparison Test. When the convergence or divergence

of a certain series is known that of others can often be inferred,
the principle employed being the following:
The positive series

ai + a2 + + • •
•
+ ttn + • •
• (1)

is convergent if the positive series

fci + ?>2 + ^3 H — •
+ ?>n + • •
• (2)

is convergent and either

a>n ^ (3)
where k is constant, or
^n+l ^n+1
^ (4)
an bn
for all values of n.
The positive series (1) is divergent if the positive series (2) is di-
vergent and either

a^n ~ khny (5)

where k is a constant greater than zero, or

«nH-l
^ -7~ (6)
an t)n

for all values of n.

The proof of these statements is immediate. Thus, if the con-

dition (3) is satisfied,

ai + 02 + * •
•
+ On = h(bi + &2 + * •
•
+ bn)

and so the sum of n terms in the series (1) is bounded if the sum
of n terms in the series (2) is bounded. The condition (4) can be
written

bn-^l bn
whence by induction
On
S—
Oi
—
bn bi

Thus (4) is equivalent to (3) with

§144] ABSOLUTE AND CONDITIONAL CONVERGENCE 305

As we run through the series San the ratio

dn

measures the rate of growth of the terms. The parts (4) and (6)
of the above theorem thus state that a positive series is convergent
if the rate of growth of its terms is ultimately less than that of a

given convergent positive series, and is divergent if the rate of

growth is ultimately greater than that of a given divergent posi-
tive series.

Example L The series

2
- + — + — +•
2 2
n(2n — 1)

is convergent since the ratio of its nth term to that in the convergent
series [§142, (6)]
1 .
1 .
1 . .1
^2 + 22 32 ^ n2
’

IS
2n
2n - 1
^ 2 .

Example 2, The series

22-4 2-4---(2n)

is divergent since the ratio of consecutive terms

Un-n
^ 2n + 1
^ n +\
an 2(n +1) n +1
is greater than the corresponding ratio

hn+i n
bn n +
in the divergent series [§142, (7)]

1+1+1+.. .+i+.

and Conditional Convergence. In the series we

144. Absolute
have been discussing all the terms have been positive. If a series
contains both positive and negative terms, it is necessary to dis-
tinguish two kinds of convergence.
306 SERIES WITH REAL TERMS [Chap. XIII

The series

Wi + ^2 + ^3 H h Wn H (1)

is called absolutely convergent if the series of absolute values

I
Wl 1 + 1
^2 I + % I 1
H h I I
H (2)

is convergent. A series that converges, but is not absolutely con-

vergent, is called conditionally convergent.
It should be noted that an absolutely convergent scries is con-
vergent. By that we mean that the series (1) is convergent if (2)
is convergent. This follows since

I
Wn+l + Un^2 H + Wn+p ^ 1 \
'i^n+1 1 + |
Un^2 H |
h |
Un+p |,

and so the general convergence condition [§140, (2)] is satisfied by

the terms of (1) if it is satisfied by those of (2).
146. Ratio Test. The following, called the ratio test, is partic-
ularly useful when each term is formed from the preceding term
by a comparatively simple multiplication:
Let
+ ?/2 + W3 H — •
+ Wn + • •
•
(1)

he an infinite series and

‘^n+l
L = lim ( 2)
n -> so Un

If L < If the series is absolutely convergent; z/ L > 1, the series is

divergent; if L = 1, the series may converge or may diverge.
First suppose that L < 1 and let r be any number that satisfies
the condition
L < r < 1. (3)
As n increases,

Un

ultimately becomes as nearly equal to L as we please and so ulti-

mately remains between L and r. Disregarding a finite number of
terms, which has no influence on convergence, we thus have

Wn+l
< r.
Un
§146] ALTERNATING SERIES 307

Since r < 1, the geometric series

+ h/*^+***

is convergent [§139, (4)]. The series

1 I
+ I
^2 I
H h I I
H

then also converges since the ratio of consecutive terms

Wn+l
Un

is less than the corresponding ratio

|.n4-l

in the geometric series [§143, (4)].

If L > 1, the series (1) obviously diverges since the terms Un
do not then tend to zero as n increases. If L = 1, examples show
that the series may converge or may diverge.

Example 1. ^ + ^+ ^+ ^“1 ‘

This is the series used to define the number e [§84, (3)]. In this case

(n+l)I 1
L = lim lim = 0 .

n 00 1 n — 00 ?^
+ 1

n\

Since L < the series convergent.

Example
1,

2, 1 + 2r + 3r^ 4
is

h + l)r — •.

In this case

L = lim
(^ + 2) r I I

|r|.
71 —* 00 n+1
The series thus converges if
|
r |
< I and diverges if
|
r |
> 1. The series
also diverges when |
r [
= 1, for the nth term does not tend to zero as
n increases.

146. Altemating Series. A series with terms alternately positive

and negative an altemating series. The convergence of
is called
such a series and the error due to using n terms as an approxima-
tion can often be determined by the following theorem:
308 SERIES WITH REAL TERMS [Chap. XIII

A series

tti — a2 + aa — a4 + • •
*
, (1)

mth terms alternately positive and negativSy is convergent if each term

is numerically less than the preceding and

lim an = 0. (2)
n »

The error committed by stopping at the nth term is less than the first
term omitted.
To prove this we note that the sum of 2m terms can be written
in the two forms

S2m = (^^1 — ^2) + (ua — a4) + h fem-l — *

Ct2m) (3 )

= Ui — (^2 — Ua) — — (a2w—2 ” Ct2m—l)

• • •
(^2m- ( 4)

Since an diminishes as n increases, each expression in parentheses

is Equation (3 ) thus shows that S2m increases as m in-
positive.
creases and (4 ) that it always remains less than ai. Therefore
S2m approaches a limit as m increases [§ 14 (2)]. Since ,

52 m+l = S2m + Cl2m+1

and a2m+i tends to zero, >S2m+i tends to the same limit. Thus,
whether n varies through even or odd values, Sn approaches a
definite limit as n increases. The series is therefore convergent.
If we stop with the nth term, the error

Un+l — (ln+2 + Ctn+3 ““

«n+4 H
” ^n+1 (®nH-2 ^n+s) (®n+4 ^n+s) ’ * *

is obviously less than an+i.

Example. 1 — i +^— .

Since the terms are alternately positive and negative and steadily
diminish to the limit zero, the series is convergent. Since

is divergent, the given series is, however, not absolutely, but only condi-
tionally, convergent. If
1 — i i “ 4

is used as an approximation for the sum of the series, the error will be less
than 7.
§1481 POWER SERIES 309

147. Series of Functions. In the series we have hitherto dis-

cussed, the terms have been constant. If the terms are functions
of a variable Xj the series has the form

Ui{x) + U2 (x) + U^(x) H ['Unix) H . (1)

This is said to converge for a particular value x = c if the series

of constants

Ui(c) + U2 ic) + Usic) H VUnic) H

is convergent. The set of values x for which a series converges

constitute its region of convergence.
For values in its region of convergence the sum of the series
^Unix) is a function

Six) = Uiix) + U2 ix) h Unix) H . (2)

This can be expressed as the sum of two parts

Six) = Snix) + Rnix)y

where
Snix) = Uiix) +U2ix) h Unix) (3)

is the sum of the first n terms and

Rnix) = Un^lix) + UnJ^2 ix) ^ h Wn+pW H (4)

is the remainder after n terms. At a point where the series con-

verges Snix) tends to Six) as n increases and Rnix) tends to zero.
148. Power Series. A series of the form

ao + o>ix + a2X^ H h UnX^ H , (1)

where Uq, ai, 02 , etc., are constants, and x variable, is called a

power series. More generally

Uq aiix — a) + •
— h Unix — a)” + • •
•
, (2)

where a is constant, is also a power series which differs from (1)

only in the origin from which x is measured. To distinguish the
two forms we sometimes call (1) a power series in x and (2) a
power series in x — a.
With regard to convergence of power series the following is
fundamental:
If a power series in x converges when x = it converges absolutely

for every value of x numerically less than b.

310 SERIES WITH REAL TERMS [Chap. XIII

In fact, if the series

Oo + dih + a2&^ + * •
•
(3)

converges, the terms an6" tend to zero as n increases (§140). They

are therefore all numerically less than some number Af, that is

I
Clnf>^
I
< Mf 1
Un I
<
•
(4)
I
^
Thus the terms of the series

I
Oo I + I
ai* I + I
02®^ 1
H 1-
I
OnX^ I
H (5)

are respectively less than those of the geometric series

If then I
X 1
< I
6 |,
the geometric series, and consequently (6),
converges, that is, (1) converges absolutely.
If a power series in x diverges when x — b, it diverges for every
value of X numerically greater than b.
For it could not converge when x | ] > |
6 ]
since, by the proof
just given, itwould then converge at x = 6.
A power series in x obviously converges when x = 0. The above
theorems show that it either converges for all values of x or there
is a numerical value beyond which it everywhere diverges. If we
consider x = 0 as a zero interval and the totality of values an
infinite one, we can then say that the region of convergence of a
power series in x is an interval of center x = 0, and inside that in-
tervdl the series converges absolutely.
The limits of the convergence interval are usually determined
by the ratio test (§145). At the ends
of the interval the series
may converge or may diverge, or it may converge at one end and
diverge at the other.
A zero convergence interval is illustrated by the series
1+ X + (2!)®* + (31)®® +• •+ •
(n!)x» +• • •.
 POWER SERIES 311

The ratio of consecutive terms in this series is

= (n + l)x,
Un
and unless a; = 0 this is numerically greater than 1 when n is suf-
ficiently large. Thus the series diverges for every value except
X = 0 (§145).
A finite convergence interval is illustrated by the geometric
series
l+a: + ^^H
which we have already seen (§139) converges absolutely within the
interval
-1 < X < 1

and diverges at the ends of the interval and outside.

As a series with infinite interval of convergence take

1 +-+
2! 3! n!

In this case the ratio of consecutive terms is

Un+l _ X
Un n + 1

For any fixed value of x this has the limit

L= lim = 0.
n -> 00 71 1

Thus the series all values of x (§145).

converges absolutely for
Series in powers of —
a are discussed in a quite similar way,
a:

the essential difference being that the center of the convergence

interval is a; = a instead of a; = 0.

{X - 2Y {x - 2)”
Example.
32 3^

The ratio of consecutive terms is

Un+l a; - 2
Un 3
The series therefore converges if

Un^l
L = lim < 1,
n 00 Un
that is, if

i
a; - 2 1
< 3.
312 SERIES WITH REAL TERMS [Chap. XIII

The values of x that satisfy this condition extend the distance 3 on each
side of a; = 2 and therefore form the interval

-1 < a: < 5.

At the limiting values a; = —1 and a; — 5 the series is divergent since its

nth term does not then tend to zero as n increases.

149. Taylor’s and Maclaurin’s

For values of a; in its Series.
region of convergence a power series in or a? — a represents a a?

function of x. Conversely, most of the commonly used functions

can be represented as power series in x — a, at least for certain
values of x and a. Assuming that a given function can be so
represented, the coefficients in the series are readily determined
by differentiation.
Thus suppose the function f(x) expressible as a power series of
the form

fix) = do + aiix — a) + a2ix — a)^ H h Onix - a)” H , (1)

the representation being valid within the convergence interval of

the series. Differentiating both sides successively with respect to
X and assuming that the derivatives can be obtained by differen-
tiating the series term by term, we find

/'(*) = oi -f 202(3; — o) -f Sasix — o)^ -f 404(3; — o)® -|

,

f"ix) = 2o2 -|- 2‘3o3(3; — o) -f- 8*404(3; — o)^ !"»

f"'ix) = 1*2*303 -I- 2*3*404(3; -o)+***,
etc. For values of 3; in a sufficiently small interval of center x — a,

these equations are assumed to be identities. Putting x = a, they

give
/(o) = oo, /'(o) = oi, /"(o) = (21)02,

Solving for oq, oi, 02, etc., and substituting in (1), we finally obtain

ix — o)^
fix) =fia)+ria)ix-a)+ria)
2 \

ix - o)"
+ /”(o) nl
j
1
. (2)

This is known as Taylor’s series. It is valid in a given interval

of center 3; = o, provided that fix) is expressible as a power series
convergent in that interval. The final result will be investigated
later in another way (§150).
51491 TAYLOR’S AND MACLAURIN’S SERIES 313

When 0 = 0, Taylor’s series becomes

fix) = m+ /'( 0 )x + /"(O) ^

2!
+ • • •
+ r (0) ^
n!
+ • •
•
, (3)

which is called Maclaurin^s series.

Taylor’s series gives f(x) in terms of the values that f{x) and
its derivatives assume at a; = a. In using the series to compute
f{x)f the constant a may to a certain extent be assigned arbi-
trarily. The value chosen must, however, be one
for which /(a),
^tc., are all known.
Furthermore it should be taken
as close as possible to the value x at which f(x) is wanted. For,
the smaller a; — a, the fewer terms (x — a)^, (x — a)^, etc., need
be computed to obtain a desired approximation. In particular
Maclaurin’s series should be used when x is small.
Since Taylor’s series converges most rapidly near x = a, it is
often called the expansion of f(x) in the neighborhood of a: = a.
It is a series in powers of x — a, and the above discussion shows
that for a given function f(x) and value a there is only one such
series.
Similarly, Maclaurin’s series is the expansion of f{x) in powers

of X, or expansion in the neighborhood of a: = 0.

Example 1. Expand (1 + x)^ as a series in powers of x, and determine

its region of convergence.
In this case

/(X) = (1+a:)-, m= 1,

/'(x) = m(l + /'(O) = m,

/"(x) = m(m — 1)(1 + /"(O) = m{m — 1),

etc. Substituting these values, (3) becomes

(1 +X)”* = 1 + mx + —x^ +
n\

This is called the binomial series. To determine its region of con-

vergence we take the ratio of consecutive terms

i/n+i _ m—n ’

Un W + 1
and find the limit

L =
n
lim
•
^
Un
314 SERIES WITH REAL TERMS [Chap. XIII

By the ratio test (§145) the series converges if |

a;
|
< 1 and diverges if

1*1 > 1-
Example 2. Expand In x in powers of a; — 1, obtaining four non-
vanishing terms.
In this case, a — 1,

Six) = In a;, /(I) = 0,

1
fix) /'(I) = 1.

1
fix) /"(I) = -1,

f'ix) II
fV) = 2,

6
fix) /'"(I) = -6,

etc. Substituting these values, (2) becomes

~
Inx = (z - - ~ ^
1)
^ (5)

Example 3. Find correct to four figures.

The number nearest 33 with integral fifth root is 32. We therefore
write
= ^32 + 1 = 2^1 +
By the binomial series

V ^ 32/ ^6 32 ^ 2 (32)*
^

= 1+7^-
160 12,800
+ •

Since this is an alternating series, the value

= 2(1 + rb) = 2.012

evidently has the required precision.

Example 4. Find tan (46®) to four decimals.
The value nearest to 46® at which tan x and its derivatives are known
TT
is 46®. We therefore let o = - •

4
Six) = tan ®, = 1,
/ (0
/'(*) = sec®*, 2,
-^'(l)
§1601 TAYLOR^S SERIES WITH REMAINDER 315

/"(a?) = 2 sec^ x tan x, /" = 4,

/'"(x) = 2 sec^ X + 4 sec^ x tan^ x, /"' = 16.

Using these values, Taylor^s series assumes the form

tanx = 1 + 2 (a: - + 2 (x - ~
I
When X = 46°
TT TT
X .01745.
4 180

Powers beyond the second are evidently too small to influence the fourth
decimal. Hence

tan 46° = 1 + 2(.01745) + 2(.01745)2 = 1.0355.

160. Taylor’s Series with Remainder. To obtain Taylor’s series

we assumed that f(x) could be represented by a convergent series
in powers of x — a. We shall now derive a formula that frequently
can be used to justify that assumption. For that purpose let Rn
be the function of x and a such that

/(x) = /(o) +/'(o)(x - a) +f''(a) -i

+ n\
+ (1)

This function Rn is called the remainder in Taylor’s series. It

is the error committed when we drop all powers of x — a beyond
the nth. Rn tends to the limit zero as n increases, the series is
If
convergent and its sum is /(x).

To find an expression for Rn let Pn(^, ct) be a function of x and

a defined by the equation

Rn - Pn(x, o)(x - 0)"+' (2)

and

Fit) = Six) - Sit) - mix -t)—--

- m n!
- Pnix, o)(x - <)"+'. (3)

In this equation consider x and a as constants and t as variable.

Equations (1), (2), and (3) show that F(o) = 0. It is also obvious
that Fix) = 0. By the mean value theorem (§34), if Fit) is dif-
316 SERIES WITH REAL TERMS [Chap. XIII

ferentiable, that is, if f(t) has derivatives up to the (n + l)8t

order, there is then a value { between a and x such that

F'(^) = 0.

Differentiating F{t) with respect to t, and replacing ( by we get

(x - Q"
F'(0 = + (n + l)P„(a:, a){x - ?)“ = 0
n!
whence
P„(x, a) (4)
(n + 1)!*

By substituting this in (2) we obtain the value Rn in (6) below.

Thus we have proved:
If f(t) is continuous in the interval a ^ t ^ Xy and has a derivative

of order n + 1 between a and x, there is a value f

for each value of t

between a and x such that the remainder in Taylor^ s series is

(x —
(5)
(n + 1)!

This may be regarded as an extension of the mean value theorem.

If we stop the series with the nth term, a limit to the possible error
is obtained by assigning to ^ the value between a and x which
makes largest. Unless a: — o is very small this limiting
| |

error may, however, be considerably too large. A better estimate

for the actual error is usually given by taking { midway between
a and x.

Example 1, Prove that

(6 )

for all values of x.

By substituting in (1) with a — 0 we get

e* = H-X + -+. ••+- + «„

By equation (5)

Rn —
(n+l)I
where £ is between 0 and x. Thus
glarl |n+l
3;
iRnl^ I

(n + 1)!

For any fixed value of x this tends to zero as n increases. The series (6)
therefore converges to e* as limit for every value of x.
§161] OPERATIONS WITH POWER SERIES 317

Example 2, Find the maximum error in the equation

In (1 +2 = ) a:
- |-

if|x| <0.2.
The remainder in the Maclaurin expansion is

° 3(1 + {)»
where { is between 0 and x. For |
a: |, and consequently | f |,
less than 0.2,
the greatest value of this expression isgiven by

X = f
= -0.2.
Thus
(0.2)3
\R2\ ^ - 0.2)3
0.0052.
3(1

If we take for ^ the middle value —0.1 instead of the extreme value,
the maximum error thus estimated is

(0.2)3
= 0.0037.
3(1 - 0.1)3

The actual maximum is

- ( 0 2)2
.

(- 0 2 )
. In (0.8) = 0.0032.

161. Operations with Power Series. In books on advanced

calculus it is shown that within its convergence interval (including
an end point at which the series converges) the function

f(x) = ao + ai{x — a) + a2 {x — a)^ H

+ On(x - a)"" H (1)

represented by a power series is continuous and its integral

J'f{x) dx
— c + ao{x — a) + -^ (x — a)^ -] ( 2)

may be obtained by integrating the series term by term. Within

the convergence interval (but possibly not including the ends) the
derivative

f'(x) = ai + 2a2(x — a) H h mn{x — H (3)

is also obtained by differentiating term by term.

If two or more series of the form (1) converge in an interval of
center x = a, they may be added, subtracted, or multiplied and
318 SERIES WITH REAL TERMS [Chap. XIII

the resulting series in powers of a: — o will converge within that

interval. Provided that the denominator does not vanish, one
series of the form (1) can even be divided by another. The quo-
tient arranged inpowers of a; — a will converge at a; = a, but its
convergence interval may be smaller than that of either series.

Example 1. The series

(4)

is convergent if 1
1 1
< 1. By integrating both sides we obtain

tan“* X
n dt
(5)
Jo 1 +
convergent if [
x |
< 1. This series converges at x = 1, and SO by con-
tinuity we have
w
l=-=l--1.1-_-+.
1 ,
tan
,
+ ,

(6 )

although (4), which was used in its derivation, does not converge at < == 1.

Example 2, The series

+ x) = X -— y
In (1
+
converges in the interval
-1 < X g 1.

By differentiating both sides we obtain

-“4
1+x
—= 1 — X + x^ — X* + • •
•,

convergent if |
x |
< 1 but not at the limit x = 1.

Example 3. The series

nf»2 rt(»3 rffi’

e* = H-x + ^ + - + -..+ ^ + ...,

sinx = x-|y + |-| + ---,

converge for all values of x. By multiplication term by term, we obtain

6* sin X = X + x^ + Jx* — ^tjx® H

valid for all values of x.
 PROBLEMS 319

Example 4. The series

smz = .:-- + --.

C 08 X = 1 -- + --...
converge for all values of x. By division we get

tanx = —
sin X
= X + - + — +•
X® 2x® ,

convergent in the interval

TT
^ ^ TT

3.
PROBLEMS
1. The sum of n terms of a series is
n -f 1
+ W2 H h Mn

Show
4.
that the series converges, and find its sum.
2. The sum of n terms of a series is
1
+ W2 + * •
•
+ Wn = -5

Find the terms of the series and their sum.

Show that
1
_
+ 1) m m{m m+ 1

By writing this equation for m = 1, 2, ,

n and adding, find the sum of n
terms of the series
1 1 1 1
'^ •
+ n(n
1-2 2 3-
3-4 + 1)

and show that the sum of the series is 1.

Show that
1 1 1

(m -J- a){m 4-1+ o) m+a m + l+ a

By means

(1 4" u)(2
1.1
of this identity determine the

+ a) (2 + a)(3 + a)
sum of the scries

(w + o)(?i
1

+ 1 + a)
+ •

By finding the sum of n terms determine whether the following series con-
verge or diverge:

cos (n) cos (n + 1)

5-
Y, e-E n +1
n-l n*l
<0
W 1
7.
^
n-O
cos (n) — cos (n + 1).
a
8.
"
> 1
In
"I"

n
 9.
320 SERIES WITH REAL TERMS [Chap. XIII

Determine the sum of the first n terms of the series

and the sum of the series.

10. Determine the sum of n terms of the series

and the sum of the series.

11. If X is a positive number, determine the sum of n terms of the series

1 + C-* 4- e-2* + e-3* _|. e~4x .

and the sum of the series.

By the integral test determine which of the following series converge and
which diverge:

**'
S V»> + 1

'"SSTT-
n=l n»l
n — 1

"
n»l
n
17.
“ + 3w +
D. n- 2

n=l n-2 - 1

20. Determine the values of p for which the scries

00
j

n(ln ny
is convergent and the values for which it is divergent.
21. Determine the values of p for which the scries

y
00

n In n(ln
-

—
In n)^
i

is convergent and the values for which it is divergent.

Determine which of the following series converge and which diverge:

22 , — +^ — 4-—
1-3 ^ ^ 2-4 3*6
H ^ n(n + 2) ^
1

23, — ^2-4 ^3-8^

1-2
4“ ' '

'
4“ •
"
4“ * * * 4"
^n-2**^
4“ * * •
•

2 3 4 n4-l .

2 3 4 ^ .
 PROBLEMS 321

(2 n)*
^
2T.1+H+1+...4
2 6 10

W.\ + -^ + ~+-
1 V2 V3 Vn
29. -i -4- -L + . . .
-I
1- . .

11 12 n + 10

30. - H r +•••-• 3
4-
2 1+25 1 + n5
1 1-3 L3-5
2-32 2*4-52 2-4-6-72
1
•
+ 71”
In n

12
2^“^ 32 '^42'^* ’“^(71
3 n
+ 1)2
12
“•2+4+8+"'+?^+"-
3 n

+^ +
w .

3 9 27

^ ^-3 13-5 1-3 ••• (2n - 1)

87 I

2 ’^2-6 '^2-5-8
,
^I •
+ 2-5 -
OQ 12 I I
3 _L . . . J
n
(3n

+ •••
1)

2-2 3-2* 4-2* (» + l)2"

1-2 l- 2 --r
»®-i+r3+- 1.3 ... 2n - 1)
+ •

3 3* 3"
+•
3 3* 3“
"•T+2i+- + ;i!+"-
42.1+^+iL+... ^ i^+...
^ + ^
10 ^ 10* ^ 10* 10"

43. By use of the inequalities

determine approximately how many terms of the divergent series

1 + 0^ + w0 +***4 f"’**
71

must be added to make the sum greater than 20.

322 SERIES WITH REAL TERMS [Chap. XIII

44. Show that Sun is convergent if

46. If a positive series Xun is convergent, show that is also convergent.

46. If
(-ir
Vn
show that Zon is convergent and Zan divergent.

Determine the interval of convergence for each of the following series:

X x^
1+-+-+... + -+....
47.
12 n

48 fL»^
+ ^!
1-2 2 2 -
3 2* -
4 2*-

22
^ 22-42 2 ^* 4** 6*
+ •

a? x^
60-
3
+v- ,

“•*+3+6+7+-
“•^-2+4-6+-
n.i+^2 + ^4 + ^+.
6
1-3 1 -
3-5
“•^-^T + ^*‘ + 1.2.3

„
66 X .
1 »* .
1 - 3 *® 1 -
3 5 - *''
.
[-•••.
2 3 2-4 6 2-4 6 7
66. X — + X® —

2 2»*

aj j- 1 (x 4- 1)2 (x 4- 1)»*
68.
1
+
'
2 n
4..

1 -X+2 .
(X + 2)2 (x + 2)» .

3 32 3»
-3 - -
0. 1

61. 1
+
X
12!'
- 3(x - 2) + 32(x -
.
(X 3)2

2)2 -
.

3®(x
{X

n!
- 2)® +•
12 3

68. 1 + : - 1 - D* (* - 1)»
+
 PROBLEMS 323

Assuming that each of the following functions is expressible as a series in

powers of z, obtain three non-vanishing terms:

64. 66. secx.

1 — »
66. tanz. 67. Vl+x.
68 . 69. In (1 - *).
70. In cos a;. 71. sin^x.
72. sin-^aj. 73. e^sinx.

In each of the following problems expand /(z) in powers of a; — a, obtaining

four non- vanishing terms:

74. /(*) = c*, a = 2. 76. Ax) =* In X, 0 = 1.

76. /(») = Vi + *, a « 3. 77. /(x) = Vi, 0 = 4.

1
78. /(*) = a = -1. 79. Ax) = X®, a = 2.

TT
80. /(X) = x^, a » 1. 81. fix) = sin X,
“ 6*
T
82. Ax) » cosx, a “ 2* 88. Ax) = vr= •
X®, X = -2 .

By using an appropriate power series compute each of the following values

to three significant figures:

84. sin 6^ 86. tan 5®.

86. In (1.2). 87. sec 68®.

88. tan 43”. 89. Ve.

90. tan-i (1.1). 91. sin"* (0.1).
JS.

92. Show that x* cannot be expressed as a power series in z.

By expanding in Taylor’s or Maclaurin’s series and showing that the re-

mainder Rn tends to zero as n increases, prove the following expansions for
appropriately chosen intervals of convergence:

a;8 ^5
98.8inx-x--+--....
Z^ z^
94. cos z =“ 1 •••.
2! 4!

(x In 2)*
96. 2* = 1 + X In 2 -A-
2!

96. Inx (*-i) _L^ + L_^

(x- - 1)2 (x D*

1
97. -1 -fx+2) - (x->-2)^ -•
1 +X
98. Ina+xl +
99. Estimate the error in the approximation

(1.01)^ - 1+ ^(0.01).
 SERIES WITH REAL TERMS [Chap. XIII

100. Show that x, sin x, and tan x have the same value accurate to 1% if
I
<0.1 radian.
a;108.
I

101. For what values of x does cos x differ from unity by less than 0.0005?
For what values of x is the approximation

sm a; = a;
“
— X®
6

permissible if the allowable error is 0.0001?

103. If cos X is replaced by 1 — ^x^, for what values of x will the error be
less than 1%?
104. By expanding In (1 + a;) and In (1 — x) in powers of x and subtract-
show that
106.
ing,

By substituting x \ calculate In 2 to three decimals.

Obtain the Maclaurin series for sin^ x by using the identity

. « 1 — cos 2aj

and expanding the right side.

106. Given
1
« l+a; -ba:® +•

find an expression for

(1 - x)^
by difierentiating both sides.

109. By differentiating the Maclaurin series term

107. by term verify the fol-
lowing:
d d d
~
dx
.

sin a; =* cos x, -p cos x

ax
= — sm x, — .

dx
- e* « c®.

108.
110. By differentiating
X® 2x® 17x7
^—
. .

tan X ** X *4"
3 ^
H ^
"T“ “f"
• • *
*
'
15 315
find an expansion for sec® x.
Given
1
« 1 -X H-x® - X® -H*

find the Maclaurin expansion for In (1 -|- x) by integrating both sides and
determining the constant so that the value is correct when x » 0.
Expanding
- (1 -
Vi -
by the binomial theorem and integrating, show that

sm~* a “
AI
r* — '
dt
—«— *
1
,
1**
2 3
.
l-S*'*
2-4 6
4-.
• • ••
 PROBLEMS 325

111. Obtain

dx

as a series in powers of x.
112. Find
sin
-dx
113. Jo a

correct to three decimals by expanding sin a; in a Maclaurin series, dividing

by Xj and integrating term by term.
By use of series find a solution of the equation

tanx =

114. By series solve the equation

In (1 + x) = 1.1.

115. By taking the product of the Maclaurin expansions for and cos x
obtain four terms in the expansion of e~^ cos x.
116. By squaring
yS ^5

cosx = l-- + ---..,

and adding, verify the equation

sin^ X + cos^ X = 1.

117. Find an expression for

1
sec X =
cos X

by expanding cos x in a Maclaurin series and dividing.

118. By replacing x in

by the Maclaurin series for tan x find

gtanx

correct to third powers of x.

119. By writing
1 1 1

1 + x^ 2 1 4- (a? — 1) + - 1)^
^

performing the division on the right, and integrating term by term, obtain

tan
-1
^ X
r*
= Jo
/
—^— -

1 4- x^

as a series in powers of x — 1 correct to fourth powers of a? — 1.

 CHAPTER XIV
COMPLEX NUMBERS
162. Complex Numbers. In many branches of science and
engineering it is found convenient to introduce a symbol

i =
and to form combinations
d “f* ih

of this symbol with real numbers a and 6. Such a combination is

called a complex number a being its real part and ih its pure imagi-
,

nary part.
No value is assigned to but algebraic operations involving this
symbol are defined as those obtained by following the usual rules
of algebra as if i were a real number and replacing i^ by — 1 when-
ever it occurs.
To each complex number

c = a +
corresponds a conjugate complex number

c = a -- ib

obtained by replacing i by —i. The product

cd = {a + ib){a — ib) = + 6^
is real. The conjugate is therefore the multiplier we use to ration-
alize a complex number. Thus to express the fraction

5 — i

3 + 2f
as a single complex number we multiply above and below by
3 — 2i, obtaining

6 - f (5 - - 2%)
z)(3 13 - 13i
"
3 + 2i “ (3 + 2i)(3 - 2i) 13
326
§152] COMPLEX NUMBERS 327

By expanding and rationalizing denominators any rational com-

bination of complex numbers can be expressed as a single complex
number.
In an equation with complex terms real and imaginary parts on
the two sides are separately equal. Thus, if

CLi -f- ibi = (12 “f" ^^2;

we must have
di = d2) bi = 62*

For, if this were not so, we could solve for

(l\ — d2
i =
62 — bi

and thus find a real value for i, which is impossible.

In solving a problem where only a real result has meaning, if

allanswers are complex, there is no solution. For this reason com-

plex numbers are sometimes called imaginary. In subjects, like
alternating currents, where significance has been assigned to com-
plex values, these are, however, entirely real.

Example 1. Show that if

r = a + 16
is a root of a polynomial equation

fir) = do + air + a r^ H 2 h Onf" = 0

with real coefficients, the conjugate

f — a — ih
is also a root.
To show this let
fir) =p+ ig

be the complex number obtained by replacing r by a i6, expanding, and +

substituting = — 1. If r is replaced by a — i6, the same operations
give expressions which differ from the former only in having — i in place
of i. Thus
/(f) =p- iq.

If r is a root.
Sir) =p + ig = 0,
p and q are both zero, and therefore

/(f) = p - ig = 0,

which shows that f is also a root.

328 COMPLEX NUMBERS [Chap. XIV

Example 2, Determine whether the parabola ^ Ax and the straight

line y — X + 2 intersect.
Solving simultaneously, we find

2/
*= 2 dt 2i, X — ±2i.

Since there is no solution with real coordinates, the loci do not intersect.

153. Geometrical Representation. customary to repre- It is

sent the complex number a +

ih by the vector from the origin to
the point of rectangular coordinates a, 6. Real numbers (6 = 0)
are represented by points on the x-axis, called the axis of reals.

and pure imaginaries (a = 0) by points on the 2

/-axis, called the
axis of imaginaries. The numbers 1 and i are represented by unit
vectors along OX and OY respectively.
A complex number
z ^ x + iy (1)

has an absolute value, or magnitude,

1
2 I
= Vx^ + = r (2)

equal to the length of its representing vector, and an angle 6 equal

to the angle from the positive direction of the x-axis to this vector.
Since this angle may wind any number of times around the origin,
it has an infinite number of values differing by multiples of 2w.
In terms of magnitude and angle we have

X = r cos y — r sin 0, (3)

and therefore
2 = r (cos ^ i sin 0). (4)

We sometimes refer to this as the polar form of the complex num-

ber x.
§ 163 ] GEOMETRICAL REPRESENTATION 329

The coefficients xi, i/i and X2, y2 in the complex numbers

Zi = xi + iyif ^2 = ^2 + iy2
are the components of the vectors that represent these numbers.
Since corresponding coefficients are added and subtracted in the
combinations
zi + Z2 — xi + X2 + i{yi + 2/ 2 )>

Zi - Z2 Xi - X2 + i{yi - 2/2),

it follows that the sum and difference of two complex numbers

Figube 193 . Figure 194 .

are graphically obtained by adding and subtracting the vectors

that represent them.
Similarly, the vector representing 21 + 22 H h^n is the sum
of the vectors representing its parts. Since a straight line is the

Figure 196 .

shortest distance between two points, these constructions show

at once that

1
+ 22 H h 2n 1 ^ 1
2^1 1 + 1
Z2 1
H 1"
1
25n 1, (5)

that is, the absolute value of a sum of complex numbers is equal to or

less than the sum of the absolute values of its parts.
330 COMPLEX NUMBERS [Chap. XIV
164. Product and Quotient By direct multiplication of two
complex numbers
zi = ri (cos ^1 + i sin ^i),
Z2 = ^2 (cos $2 + i sin 62)
we find that their product

Z1Z2 = rir2[cos $1 cos 62 — sin di sin 62

+ z‘(sin cos $2 + cos ^1 sin ^2)]
= rir2[cos (^1 4” 62) 4" i sin (^1 4" ^2)] (2)

has the magnitude rir2 and the angle 4-^2* Thus the product
of two complex numbers is obtained by multiplying their magnitudes
and adding their angles.
To find the quotient of two complex numbers
Zi ri(cos 4- i sin ^1)

Z2 r2(cos O2 + i sin 62)

we rationalize the denominator and thus get

— = ~ [cos (^1 — ^2) 4- i sin (^i ~ ^2)]» (3)

Z2 r2

The quotient of two complex numbers is obtained by dividing their

magnitudes and subtracting their angles.
The above equations show that multiplication and division by a
complex number can be graphically interpreted as operations of
rotation and stretching. Thus (2) shows that multiplication by
Z2 has the effect of rotating z\ through an angle equal to the angle
of Z2 and stretching it in a ratio equal to the magnitude of Z2.
Similarly division by Z2 is equivalent to rotating backward through
an angle equal to the angle of Z2 and
shrinking in a ratio equal to the mag-
nitude of 22- Multiplication by —1,
for example, is equivalent to rotation
through 180°, and division by i pro-
duces rotation backward through 90°.

Example. If z represents the vector

from the origin to a point moving around
a circle with center at the origin, find

Figure 196 . z
§1661 POWERS AND ROOTS 331

Graphically dz is tangent to the circle and directed 90® ahead of z.

Also the magnitude of the above fraction is

dz dz ds
21
\ \

\z\.
\ __
r
~
_ ^
Thus
—= i do.
z

The same result is obtained if we write

2 = + 1 sin 6),
r(cos 6

dz = r(— sin ^ + i cos 6) dO,

and determine the quotient

—=
dz — sin ^ + i cos ^ dS = , * . ,

"w i dd.
^

166. Powers and Roots. If n is a positive integer, the nth power

of a complex number
2 = r(cos 0 + i sin 6) (1)

is the product of n factors z. Its magnitude is thus the product of

n factors r, and its angle the sum of n angles $, Therefore

_ 7'»^(cos nS + i sin nO), (2)

Elimination of z from (1) and (2) gives

(cos 0 + i sin 6)^ = cos n^ + i sin n^, (3)

which is called DeMoivre^s theorem. By expanding the left side

and separately equating reals and imaginaries, it determines ex-

pressions for cos nS and sin vB, For example, if n = 2, the equation
becomes
cos^ 6 + 2i cos 6 sin 6 ^ sin^ 6 = cos 2$ + i sin 26
whence
cos 26 = cos^ 6 — sin^ 6, sin 26 = 2 sin 6 cos 6,

An nth root of z must satisfy the equation

r(cos 6 + i sin 6),

1

Its magnitude is therefore and its angle is such that after multi-

plication by n the result is ^ or differs from ^ by a multiple of 2ir.

332 COMPLEX NUMBERS [Chap. XIV

Thus
1 1 r ^ "h d -f* 2A;7r"|
zn ^ pi] cos h t sin , (4)
L n n J

where k is any integer. The values determined by fc = 0, 1, 2,

• •
•
,
n — 1 are distinct, but other values of k give roots already
found.
A complex number different from zero therefore has n distinct
nth roots.These all have the same magnitude but their angles
2t
differ by multiples of — . If vectors representing these roots are
n
drawn from a common origin, their ends are vertices of a regular
polygon of n sides.

Example 1. Solve the equation 2® = 8.

Figuke 197.

One root is 2 = 2. The other two have the same magnitude and form
with this angles of 120® and 240°. The three roots are therefore

21 = 2,

22 = + i sin 120°) = — 1 + i\/3,

2(cos 120°

23 = 2(cos 240° + i sin 240°) = — 1 — iy/Z-

Example 2. Find the value of

V2 + 3i.

The number 2 + 3i is represented by a vector of length

r = Vl3
and angle
e = tan-i I = 56° 19',
§156] LIMIT, FUNCTION, CONTINUITY 333

Its square roots therefore have the magnitude

y/r = (13) i
and the angles
Bi = 28® 9.5', 02 = 208® 9.5'.

Thus
‘V^2 4""^ == dbr^[cos "I" ^ sin 0i] = db [1.674 -j- 0.896i].

166. Limit, Function, Continuity. A complex number z which

varies through a discrete or continuous range of values is said to
approach, or tend, to a complex number a as limit if

lim \
z — a \
=0,
and to have an infinite limit, or tend to infinity, or become infinite,
if

2 I
—> 00.

We write
w = J{z) (1)

and say that u; is a function of a complex variable 2 on a certain :

range, if to each value of z on that range corresponds one or more

values of w. The function is called continuous at 2 == a if it tends
to a limit as z tends to a.

With these definitions practically all that has been said about
limits and continuity for real variables is applicable to complex
variables. It is only relations involving the notions greater than
and less than that are not extensible.
334 CX)MPLEX NUMBERS [Chap. XIV
The derivative of f(z) is defined by the equation

f'(z) = lim ,

h-*o h

provided that the limit on the right exists. In case of algebraic

functions the expressions used,and so the final differentiation
formulas, have exactly the same form whether the variables are
real or complex. For example

— 2^ = (3)
dz

A similar remark is applicable to integrals expressible by algebraic

functions.
167. Power Series. A series of the form

do + diZ + a2Z^ H h dnZ^ + • •

•
> (1)

where Oq, ai, a2 ,

•
•
•
, Un* • • are constants and

2 = x + iy,
is called a power series. Similarly

Oq + — a) + 02(2 — a)^ + • • •
+ dniz — o)” + • •
• (2)

is also a power series which has values distributed about z = a

in the same way the values of (1) are distributed about 2=0.
If we consider the symbols complex, the proofs in §148 show that
if (2) converges at 2 — a = 6 it converges absolutely for all values

I
2 — a <6, and if it diverges at 2 — a = 6 it diverges for all
values 2 — a >6.
I
The region of convergence is therefore of
the form
I
2 — a I
< iB.

It is thus in general a circle with center 2 = a but in special cases

may shrink to a point or extend to infinity. This is called the
circle of convergence of the series. On the circle of convergence the
series may converge at all points, diverge at all points, or converge
at some and diverge at others. The radius of the convergence
circle is usually found by the ratio test.
§168] EXPONENTIAL FUNCTION 336

Figure 199.

and diverges if
|
2 — 1 |
> 1. The circle of convergence (Figure 199)
thus has the center and radius r = 1. The series converges at all
2 = 1

points inside this circle and diverges at all points outside. On the circle
it converges conditionally at 2 = 0 and diverges at 2 = 2.

168. Exponential Function. For any real value of z (§150,

example 1)
^
2^ 2^ 2
+~+ -+•
2! 3! n!

This series converges not only for all real values but consequently

also for allcomplex values of 2 When z is complex, e* cannot be.

defined as a power of e, for raising a number to an imaginary power

does not make sense. When z is complex we therefore define e*
as the value determined by this series.
By multiplying the series for by that for it can be easily

verified that the function thus defined has the exponential property

( 2)

which the notation suggests.

If 0 is a real number and z = id, we have

The series in parentheses represent cos B and sin B (Problems 93

and 94, page 323). We thus have the remarkable relation

= cos ^ + i sin B, (3)

called Euler^s formula.
 COMPLEX NUMBERS [Chap. XI\

A similar equation
= cos B — i sin 6 (4)

isobtained by replacing B hy —B. By taking the sum and differ-

ence of the two equations, we get

COS B = , sin ^ = (5)

2 2i

Thus the sine and cosine, and consequently all the trigonometric
functions, are expressible in terms of exponentials. From these
expressions and the properties of exponents all the identities con-
necting trigonometric functions can be obtained.
In virtue of Euler’s formula the polar expression for a complex
number
z = r(cos B i sin B)

can be written
z = re^^, (6)

When complex numbers are represented in this form the expres-

sions for product, quotient,and power become obvious conse-
quences of the exponent formula (2).

Example, Determine the components of acceleration of a moving

particle in polar coordinates.
Let

represent the vector from the origin to the particle at time i, Its velocity
is then
dz dr dB
s+" s
.

and its acceleration is

—
**
=
Left*
^1
dtU'

Since e’® represents a unit vector along the radius and le*® a unit vector
perpendicular to the radius and extending in the direction of increasing By

the expressions in brackets are the components required.

169. Plane Motion. A rigid body is said to have plane motion

when all points of the body move parallel to a fixed plane. Such
a motion is evidently determined by that of a section parallel to
 PLANE MOTION 337

the fixed plane. In Figure 200 is shown such a section moving in

its plane.
In general the motion involves rotation which turns all lines
of the figurethrough the same angle. If Q is the angle through
which the figure has rotated at
time ty

dt ,Po

is the angular velocity at that \ / /

time. Both B and co are con-
j
sidered positive in the counter-
i\y/^ % •p/
clockwise direction. Let Pq be a
definite point of the figure, P ^/
any other point, and let Zq and z
Figure 200.
be the complex numbers repre-
senting vectors from the origin to those points. As the figure

2 - 2o = PqP (2)

rotates with angular velocity w but does not change length. Thus

— (2 - Zq) = io)(z - Zq).

at

The velocity of P is

V = —
dz

dt

Substituting for — its value from (3), we have

dt

V + i(ji{z — Zq).
dt

If CO is not zero, this can be written

V = i(a{z — c),

where

+ -—
i dzQ
c = 2o
CO at

Equation (6) gives the velocity of any point in the body at

time t. At an instant when co is not zero it shows that the point
338 COMPLEX NUMBERS [Chap. XIV

/, where z ^ c, has zero velocity and that every other point has
the velocity
V = toIP
it would have in a steady rotation with angular velocity co about
I as center. This point I is called the instantaneous center of rota-
tion, It usually changes with the time. At an instant when
w == 0 all points in the body have the same velocity and there is
no instantaneous center of rotation.
The instantaneous center is usually found either by directly
/ocating the point / with zero velocity or by using the fact that

IP is perpendicular to the direction of motion of P,

Example. A body moves with one

point A in the rr-axis, another B in the

2^-axis. Find the instantaneous center of

rotation.
Since A moves along the x-axis, the
instantaneous center on the line per-is

pendicular to the x-axis at A, Similarly,

it is on the line perpendicular to the

2/-axis at B, It is therefore the point I

(Figure 201) where these perpendiculars
intersect.

160. Logarithm. The logarithm of z is defined as a number

w = In z (1)
such that
= z. (2)

When z is real and positive, a real value can be found for w.

When
z = r(cos B i sin B)

is negative or complex, the logarithm is a complex number

w—u iv, (3)

which satisfies the equation

= e“(cos v + isinv) = r(cos B + i sin B).
Thus
c" cos t; = r cos B, e" sin t; « r sin B, (4)

and, since u is real,

e" = r, w = In r, v = B + 2nir, (5 )

where n is any integer.

§1611 HYPERBOLIC FUNCTIONS 339

Therefore
In 2 = w + iv = In r + i{e + 2mr) (6)

has infinitely many values differing by multiples of 2irt. In partic-

ular, when 2 = — 1, ^ = IT and hence
In (-1) = (2n + l)7rt. (7)

By differentiating the series for e* term by term we find

Similarly, by differentiating both sides of (2) we get

dw
pw = 1

dz
and therefore
—d In 2 = -•
1

dz z

Thus the derivatives of 6* and In z are determined by the same

formulas, (8) and (9), whether z is real or complex.

Example, Find the value of y

^
dx
p
Jq X + i

Integrating as if the values were real, we 1 +^

have X /

= In (1 + i) - In t*. A9
•^0 -H t ^
^
To interpret this we must consider
0 X
In (x + i) = In Vx^ + 1 + id Figure 202.

as varying continuously while x varies from 0 to 1. If we take ^ -

*
TT
when a; = 0 (Figure 202), we must then take d = - when x = 1. Thus
4
.X .X 1, .X
t — t« = «ni2„ — It’
Jo X + i 4 2 2 4

161. H rperbolic
3 Functions. By omitting the symbol i in the
exponential expressions for sine and cosine [§158, (6)] we obtain
two functions
c* - e-* c® + C-*
sinhx = cosh X =

called the hyperbolic sine and hyperbolic cosine of x.

340 COMPLEX NUMBERS [Chap. XIV
In terms of these, by equations analogous to those in trigonom-
etry, we define the hyperbolic tangent, cotangent, secant, and co-
secant, as the functions
sinh a: e* - e-*
tanh X = ~ ’ (2)
cosh a; c* + e-*
cosh X e* + e-*
coth X = “ (3)
sinh a; e* - e-*
’

1 2
sech X = _ (4)
cosh a: e* + e-*
1 2
cscha; = (5)
sinh a; c® - e-®

Graphs of the most important of these functions are shown in

Figure 203. It is to be noted that sinha: takes all real values.

Figure 203.

cosh X is equal to or greater than 1, and tanh x is between —1 and

+ 1.
The same definitions are used when x is complex. The equations

sinh (ix) = = i sin x, (6)

2

=
^ ^-ix =
cosh (ix) cos X, (7)

express the hyperbolic functions of a pure imaginary as trig-

onometric fimctions of a real number, and the equations

§1611 HYPERBOLIC FUNCTIONS 341

6
“^ -
sin {ix) = i sinh Xy (8 )
2i

+ c*
cos (ix) = cosh x, (9)
2

make the corresponding transformations from trigonometric to

hyperbolic functions.
The hyperbolic functions satisfy identities similar to those con-
necting trigonometric functions. Thus, by squaring and subtract-
ing we find
cosh^ X — sinh^ a: = 1. (10)

The only difference between these and the corresponding trigono-

metric identities consists in certain differences in algebraic sign,
as in the equation just written. These differences in sign are due
to the factor on the right side of equations (6) and (8). When
i

two sines are multiplied together (or two functions, such as tan-
gents, which may be considered as having a sine as factor) the
factor is replaced by —1 and the corresponding hyperbolic

formula has a different sign. Thus

+ y)
cos {x == cos X cos ^ — sin a: sin y,

cosh (x + y) = cosh x cosh y + sinh x sinh y.

By direct use of the definitions we find

—d sinh u = cosh u du
— ( 11 )
dx dx

—d cosh u = sinh u du
— , ( 12)
dx dx
du
—d tanh u = sech*^
^
u — (13)
dx dx

du
—d coth w = — csch*^
^
u — (14)
dx dx

du
—d sech w =
,
— sech
^
u tanh u
^
— (15)
dx dx

—d csch w =
dx
,
— csch u coth u —
du
dx
(16)
342 COMPLEX NUMBERS [Chap. XIV

In engineering applications hyperbolic functions are analogous

to the trigonometric, the occurrence of the one type of function
or the other depending on the algebraic signs and magnitudes of
the physical constants.
For example, if a particle moving along the x-axis is attracted
toward the origin with a force proportional to the distance, its

position at time Hs
X = a sm Q)t + c),
where a, 6 , c are constant. If it is repelled with a force propor-
tional to the distance, its position is

X = a sinh Q)t + c).

162. Geometrical Representation of Hyperbolic Functions. The
name hyperbolic given to these functions because they are
is

related to the rectangular hyperbola

= I

in the same way the trigonometric functions are related to the

unit circle 9.0
+r = 1.

In case of the trigonometric functions,

X = cos B, y = sin 6 (1 )

are coordinates of a point P on the circle. The angle 0 associated

with this point can be defined in two ways. In terms of the arc
« =• AP, it has for differential

2
( )
r
§162] GEOMETRICAL REPRESENTATION 343

where r is the radius OP. Also the area of the sector AOP is

and hence
6 — 2 (area of sector AOP). (3)
Similarly,
X = cosh Uy y = sinh u (4)
satisfy the equation

— cosh^ u — sinh^ w = 1,

and are therefore coordinates of a point P on the hyperbola.

The arc 5 = AP of the hyperbola has for differential

ds = + dy^ = V sinh^ u + cosh^ u du — r du,

where r is the radius OP. Hence

ds
du — — ( 5)
r

as in the circle. Also the area of the sector AOP is

- ydx = — ydx ^ —
^xy
J |
d{xy) h{xdy ydx)

— J*|(cosh^ u — sinh^ u) du — \u.

Therefore
u — 2 (area of sector AOP). ( 6)

Thus, whether u is determined by using the differential of arc

or area of sector, sinh u and cosh u are related to the hyperbola
in the same way that sin 0 and cos 6 are related to the circle.
344 COMPLEX NUMBERS [Chap. XIV
163. Inverse Hyperbolic Functions. To the hyperbolic func-
tions correspond inverse functions defined by the equations

y = sinh*”^ x^ if X = sinh y.

y = cosh“^ X, if X = cosh y.

y = tanh“"^ x, if X = tanh !/,

and similar equations for the other functions.

These inverse functions are also expressible as logarithms. From
the equation
e*' - e”*'
X = sinh y = ,
2
for example, we have
- 2xey - 1 = 0 ,

whence
e" = a; + Va:*^ + 1,

the positive sign of the radical being used because e*' is positive.
Thus
y — sinh“* a: = In (a; + Va:^ + 1 ). (1)

In a similar way we find

cosh~^ ® = In (x + Vx^”—T ), X ^ 1, (2)

1 4”
" ^
tanh~‘ X = I In x < 1, (3)
1 — X
, | |

X + 1
coth~^ X = 5 In X > 1. (4)
X — 1
, I I

From these equations we obtain the following differentiation

and integration formulas:

d 1 du
sinh ^ u (5)
dx Vl + dx

1 du
£. cosh~^ tt ’
u> 1. ( 6)
dx y/v? — 1 dx

du
—^ tanhu-i u '

—
1
u 1
< 1. (7)
dx 1 u^ dx’
 PROBLEMS 345

d \ du . ,

( 8)
da; \-u^dx '
'

—d sech _ ^ M
\ du
— ,
0 < M < 1. (9)
dx u'v \ — %r dx

—d csch
dx
* tt -
uV 1 + 7
1
- —
dw
da;
( 10 )

. u
= sinh ^ C. ( 11 )
a

= cosh“^
1
—
^
a
h C, u > a. ( 12 )

= - tanh""^
a a
h — C, (13)

1 u
= coth""^ ” + C, > 0?, (14)
a a

1 u
“ sech + C, 0 < w < a. (15)
a a

1 u
- csch ,
^
+ c. (16)
a a

In these formulas it is assumed that the inverse hyperbolic cosine

and inverse hyperbolic secant are positive. Otherwise the algebraic
signs must be changed.
Numerical values of the hyperbolic functions and other inte-
gration formulas are found in mathematical tables.

PROBLEMS
Reduce each of the following expressions to a single complex number a + hi:
^
1 Lti. 2'
1 — i 3 —
3 . (2 - if. 4 . (2 - 3t)(3 + 4i).
6. Given
/(*) - ** +2+ 1,

find/(3 + 2t) and/(3 - 2*).

346 COMPLEX NUMBERS [Chap. XIV
6. Given 2 — a; -f iy,

m 1

1
+32
-2 ’

show that

7. Show that
m
I 22 I

1
2122 1
= 1
21 1*1 22 1 ,
I21I

8 . Show that 2 =* 1 + t is a solution of the equation

2^ - 32^ + 42 = 2

and that 2 * 1 — t is also a solution.

9. Show that 2 = 1 + 2 is a solution of the equation
jj8 _ 2^
-f 22 = 2i

but that 2 = 1 — t is not a solution. Explain.

10. If X, y satisfy the equation of a curve, and are not real, (x, y) is called
an imaginary point on the curve. If distance is calculated by the usual for-
mula, show that the imaginary points in which a parabola intersects its
directrix are at zero distance from the focus.
11. If il(l, —1), B(3, 1), C(2, 4) are the vertices of a triangle, determine

the complex numbers which represent AB^ BC and show that their sum
represents AC,
12. If 2i =* xi + iyit 22 = X2 + iy2 t
show that |
22 — zi |
is the distance be-
tween (xi, yi) and (X2 , ^2 ).
13. If
®
- •

= p + .
tgf

where a and 6 are complex, p and q real numbers, show that p\h \
and q\h\
are the components of a along and perpendicular to b.

Express each of the following in the polar form r(cos 0 + i sin 0)

14. -2. 15. 3*.

16. 2 + 2i, 17. 4i - 4.

18. 1 + Vli. 19. 3 V3 + 3i.

Given 21 =» 2(cos 60° + i sin 60°), 22 = 4(cos 120° + i sin 120°), express
each of the following as a complex number a + ib:
20 . 2f. 21. 2122 .

32. -• 28. *2 - zi.

Zl

Determine the two square roots of each of the following complex numbers

24. 2*. 26. 2 + 2Vli.

26. 1 - V&-. 27. 1 + i.
 PROBLEMS 347

28. If zi, Z2 , Z8, 214 are the complex numbers representing vectors from the
origin to four points on a circle, show that
(Z3 - 22 ) (g4 ~ gi)

(23 - 21) (24 - 22)

is a real number.
29. By DeMoivre^s theorem find sin 30 and cos 30,
30. By DeMoivre’s theorem find sin 40 and cos 40,

Solve the following equations:

31. +8 = 0. 32. z^ + 2 4" 1 = 0.

33. 2^ - 16 = 0. 34. z^ + 6z 4- 6 = 0.
2

36. - z2 + 1 = 0. 36. z® 4- 1 = 0.

u = cos
2ir

n
h * Sin —n
2ir
,

show that the n roots of the equation z" = 1 are 1, u, • •

•, and that

1 + M + M* +---+U"-* - 0.

38. The equation

z ^w
3w
makes two values of w correspond to each value of z. Show that the six values
of w that correspond to roots of the equation

z® 4- 02 4- 6 = 0

are graphically represented by two equilateral triangles.

Find the circles of convergence of the following series:

39. +

(z - 2)’*
L.
.4-^^

(2 - 1)2 (2 - 1)8
42. 1 - (z - 1) 4-

19 1 1
(^ + 2) (2 4-2)2 (z+2)«

44. 1 +^ + 2V +3®2» +--- + n"*» +•••.

,, a - 2 1 (« - 2)»
. ,
1 3 (* - 2)* .

+i-^+2-i~6“+"-
(*-«)*
46.1+-^+-^+-^+....
Ifl 1 I I I I
348 COMPLEX NUMBERS [Chap. XIV
47. Show that if a power series converges for all real values of a variable z
it converges for all complex values of z.
48. Show that since

-4-j - 1 - *8+24 -
1+2
becomes infinite at z = f the scries cannot converge when [
z |
>1.
49. Show that since the logarithm of 0 is infinite the series

In (1 +2 )

cannot converge when |

z |
>1.
60. Show that if m is not a positive integer a derivative of (1 + x)^ be-
comes infinite at a? = — 1 and therefore that the binomial series

(1 + x)”* = 1 + mx + H

cannot converge if
j
a;
|
>1.
61. Show that
_ e^(co8 y + i sin y).
62. By expressing the trigonometric functions in terms of exponentials
verify the equation

sin (x + y) = sin x cos y + cos x sin y

for complex values of x and y.

Express the following as complex numbers:

63. e’\ 64. e‘3*.

66. 66. e\

67. Find the numbers r and d such that

1 + 2t = re^.

68. A wheel of radius a rolls with angular velocity « along the x-axis.
Find the instantaneous center of rotation and the velocities of the two points
at the ends of a horizontal diameter. Note that the angular velocity is posi-
tive when the rotation is counterclockwise.
69.The center of a wheel moves with velocity v along the x-axis while
the wheel rotates with angular velocity w. Find the instantaneous center of
rotation.
60. A piston A moving along a line CA is connected by a rod AB to a crank
CB rotating about C, Find the instantaneous center of rotation of AB.
61. A rigid body is in plane motion with angular velocity w and angular
acceleration
d<a

Using the notation of §159, determine the acceleration of the point P in the
body. At an instant when w and a are not both zero show that all points in
the body have the accelerations they would have in a rotation with angular
velocity w and angular acceleration a about a certain axis. This axis is called
 PROBLEMS 349

the instardaneous axis of acceleration. In a plane figure the intersection of the

plane and axis of acceleration is called the instantaneous center of acceleration
62. A wheel of radius a rolls along a horizontal line with angular velocity w
and angular acceleration a. Find the instantaneous center of acceleration.
63. A
wheel of radius a and center C rolls on the outside of a fixed wheel
of radius b and center O, the line OC rotating with constant angular velocity.
66. the instantaneous centers of rotation and acceleration.
Find
64. A plane figure is in plane motion. From each point P of the figure a
vector is drawn equal to the velocity of P. Show that the ends of these
vectors at each instant form a similar figure, if w is not zero.
A rod rolls with constant angular velocity to around a circle of radius a.
A particle, starting at the point of contact, moves along the rod with speed v.

By means complex number represent the position of the particle at time t.

of a
Take the origin at the center of the circle and the axis of reals through the
starting point.
66. Show that
In (Z 1 Z2 ) = In zi -f In Z 2 + 2A;7rt,

where k is an integer or zero.

67. If we take a particular value of

In z = In r + i{d + 2kTr)
and let this vary continuously as z moves in the positive direction around a
circle with center at the origin, what does this value become when z returns
to its initial position?
68. By writing
= ^1
—
1 + 2 Lx +i X ij

and integrating, check the equation

70.

dx TT

Jo 1 +x “ 2 4’

69. If xi, X2 X 3 are the three roots

,
of the equation x® +1 = 0, show that

x^ + 1 3 Lx — xi X — X2 X — X3 J
Hence determine
dx
Jo x^ -HL’

Prove the following identities:

() 1 — tanh^ x = sech^ x.

() sinh 2x = 2 sinh x cosh x.

(c) 2 cosh^ X = cosh (2x) + 1.
(d) 2sinh^x = cosh (2x) — 1.

In the following problems find the other hyperbolic functions:

71. sinh X => i. 72. cosh x =f

73. tanh x = 74. sech x = 0.8.
350 COMPLEX NUMBERS [Chap. XIV
75. If sinh a? * I* and sinh y ^ i, find sinh {x + y).
76. If tanh a; = f find tanh 2x.
,

77. Show that

78. (cosh d 4- sinh d)” = cosh nB -f sinh rtjd.

Make graphs of the following equations:

y » coth X. 79, y = sech x,

80. y « csch X. 81. y = sinh^ x.

By direct use of the definitions find the derivatives of the following func-
tions:

82. sech a;. 83. cschx.

84. coth a;. 86. sech“^ X,
86. coth~^a:. 87. csch“^a;.

88. Derive equation (3), §163.

89. Derive equation (4), §163.

Integrate the following:

tanh^xdte.
91.
J
dx
93.
/X V + 1

By use of tables evaluate the following integrals:

94.
'0
'8

Vx^To
^
dx
r^ rs dx

dx
16
*

96.
'h x* - 16*

98. If $ belongs to the interval ^ o and

sinh u = tan B,

then
^ * gd w = tan“^ (sinh u)

the gudermannian of u. Show that cosh w « sec 0, tanh u » sin B,

is called

coth u = CSC B, csch u = cot B, sech u = cos B. By means of this function

work involving hyperbolic functions can be handled with trigonometric func-
tions and conversely.
99. In Figure 205 show that

2B gd (2u).
100. Show that

fBee xdx ^ gd“^ X.

 CHAPTER XV
SPACE COORDINATES AND VECTORS

164. Rectangular Coordinates. A rotation about a directed

axis is called positive, or right-handed, if it causes a right-handed
screw to advance in the positive direction
along the axis (Figure 206).
As a reference system for rectangular co-
ordinates in space we use right-handed axes.
By this we mean a set of three mutually per-
pendicular scales X'X, F'F, Z'Z with their
zero points coincident at 0 and their positive
directions so related that OX goes into OF
by a positive rotation of 90® about OZ.
If planes perpendicular to the axes are
passed through a point P, the numbers x, y, z
in the scales at the points of intersection are called the rectangu-
lar, or Cartesian, coordinates of P. The point with coordinates

X, y, z is represented by the notation {x, y, z), and the notation

P{x, used to indicate that P has the coordinates x, y, z.
y, z) is
Taken two at a time the axes determine three planes called the
a:y-plane, y^-plane, zx-plane in terms of the axes included. On each
of these planes one coordinate is zero, and on each axis two co-
ordinates are zero.
351
362 SPACE COORDINATES AND VECTORS [Chap. XV
166. Vectors. The vectors of unit length and positive direction
along OX, OY, OZ are denoted by i, j, k respectively. The symbol
i (bold-faced type) for a unit vector should be distinguished from
the imaginary unit i = y/ — 1.
The point P with coordinates x, y, z is often represented by the
vector
r = OP (1)

from the origin to P. This is the sum of three vectors OA, AB,
BP, Figure 208, equal to xi, yj, zk, respectively. Hence

T = i\ + y] A- zk. (2)

The numbers x, y, z are the components of r along the coordinate

axes.

Since OBP isa right triangle, the distance r from the origin
to P{x, y, z) is determined by the equation

= OP2 + + j/2 4.
whence
r = V^TTT^- (3)
Similarly,
A = AJ + Ayj + A^k (4)

isa vector with components Ag, Ay, A* along the axes. These
components are edges of a box having the vector as diagonal.
The length of A is therefore

A = n/a * "
Ay -|- A,. (5)
§166] THE SCALAR PRODUCT OF TWO VECTORS 363

In particular, the vector from Piixy, yi, zi) to Vz, Zz) has
the components X2 - Xi, - yi, Z2 - Zi (Figure 209). Thus

P P2 = 1 i(x2 - xi) + j(i/2 - 2/ 1 ) + k:(z2 - zi), (6)

and the distance between these points is

P1P2 = V 2:2 - xi)^ + ( 2/2 - yif + (22 - zif- (7)

Example. Given A(0, —1, 4), J5(2, 2, —2), find the vector AB and the
distance AB.
By equation (6) the vector is

AB = 2i + 3j - 6k.
The distance between the points is

AB = V22 + 32 + 62 = 7.

166. The Scalar Product of Two Vectors. The scalar^ or dot,

product of two vectors A and B (indicated by placing a dot between
the factors) is defined as the scalar

A-B = a] iBjcose, 1
(1)

where d is the angle between the two vectors,

The quantity

6 = B I I
cos 0

is the component of B along A. The

scalar product

A-B = I
A| IbI cos<7 = A|&
I

is thus equal to the product a/ |

A |
and the component of B along A.
364 SPACE COORDINATES AND VECTORS [Chap. XV
It is also equal to the product of |
B and |
the component of A
along B.
From the definition it is evident that

A-B = B-A, (2)

that is, the scalar product is commutative (§105).

If 6, c are the components of B, C along A, it is clear from a

diagram (Figure 211) that 6 + c
is the component of B + C

along A. The equation

I
A I
(6 + c) = A I 1
6 + AI I
c

is thus equivalent to

Figure 211. A- (B + C) = A-B + A-C. (3)

Since the product is commutative, this requires also

(B + C)-A = B*A + C-A. (4)

Equations (3) and (4) express the distributive law (§106) for scalar
multiplication of vectors.
If A and B are perpendicular, cos 0 is zero and

AB = 0.

Conversely, if the scalar product is zero, one of the factors is zero or

else the two are perpendicular.
Since the product may be zero when neither factor is zero, it
follows that division by a vector cannot be performed. Thus, if A
is not zero and
AB = AC,
we cannot conclude that
B = C
but merely that
A-(B - C) = 0,

and hence B — C is zero or perpendicular to A.

If A and B have the same direction, cos ^ 1, and A‘B is equal

to the product of the lengths of the two vectors. In particular,

A-A = A 1
1* (5)

is the square of the length of A. The scalar product of a vector

§166] THE SCALAR PRODUCT OF TWO VECTORS 366

with itself is sometimes written as the square of the vector. Thus

AA = A2, (6)

but no other powers are defined.

Applying the definition of scalar product to the unit vectors
i, j, k, we obtain
i.i = j.j = k.k = l, (7)

i-j = j-k = k-i = 0.

If
A= Axi 4“ Ay] + AJs,,
B= Bxi + By] 4- J5;jk,

by expanding and using (7) we find

A*B == AxBx 4" AyBy 4“ AzBz» (8)

Example 1. Show that the vectors

A= ~ j + 4k,
2i

B = 3i + 2j - k
are perpendicular.
The scalar product is

AB= 6 - 2 - 4 = 0.

Hence the vectors are perpendicular.

Example 2, Find the component of

A= 4i -j- 2k
along
B =i— 2j 4- 2k.

If B is the angle between A and B, the component of A along B is

lAi 2
lA|cosO = = -.

Example 3. The vector

i^
A = i 4- j
- 2k

is expressed as the sum of two vectors A' and A" respectively along and
perpendicular to

B = 2i - j - k.
Find A' and A".
The projection of A on B is

AB
B *
Fioure 212.
I I
356 space coordinates and vectors [Chap. XV
The vector A' is the product of this projection and the unit vector

B
IBf
Thus
A' = ^^^ = |(2i- j-k),
A" = A - A' = la - k).

167. Work. The simplest illustration of the scalar product is

furnished by mechanical work. When the point of application
of a force F experiences a displacement which in magnitude and
direction is represented by the vector s (Figure 213), the work

done by the force is defined as the product of the displacement

and the component of the force along the direction of the dis-
placement. That is,

Work = 1
s I I
F I
cos ^ = F-s, (1)

where 0 is the angle between F and s.

If F = Fi -|- F2 is the resultant of two forces Fi, F2 acting at

the same point, the distributive law

F"S = (Fi -1- F2 )-s = Fi-s -j- F2-S (2)

states that the work of the resultant is equal to the sum of the
works done by the separate forces.
§168] ANGLE BETWEEN TWO LINES 367

When the point of application of a force F experiences two

separate displacements Si = AB^ S2 = BC (Figure 215), the dis-
tributive law

F-Si + F*S2 = F*(si + S2 )

shows that the sum of the

works in the two displace-
ments is equal to the work
done by F in the single dis-
placement
S = Si + S2 Figure 215.

from A to C, In a similar way we show that the total work done

by a constant force (such as the force of gravity on a body) when
its point of application experiences any number of consecutive

displacements is equal to the work done by the force in a single

displacement from the beginning to the end of motion. In partic-
ular, the total work of a constant force is zero when its point of
application traverses consecutively the sides of a closed polygon.
Variable forces will be discussed later (§200).
168. Angle between Two Lines. By the angle between two
lines which do not intersect we mean the angle between parallels
to these lines which do intersect.
The angles a directed line makes with the coordinate axes OX,
OF, OZ are usually represented by the letters a, /3, 7 and the
,
co-
sines of these angles

Z = cos a, m= cos P, n = cos 7 (1)

are called the direction cosines of the line.

Figure 216.
358 SPACE COORDINATES AND VECTORS [Chap. XV
Since the components of a vector

V= + Vyj + F*k
are its projections on the coordinate axes, the direction cosines
of the vector are evidently

'vl + Vl + vl

V,
cosy = V|
p^ I Vvl + vl + vl
In particular, a vector of unit length has components equal to
its direction cosines. Thus
u = i cos a + j cos /3 + k cos 7 (2)

is the unit vector along the line with direction angles a, j8, 7.
If two directed lines make with the coordinate axes the angles
«i) Pi) 7i and «2, p2 72. the unit vectors along those lines are
)

Ui = i cos ai + j cos j8i + k cos 7i,

U2 = i cos 02 + j cos 182 + k cos 72.
The angle d between the lines is therefore determined by the equa-
tion

cos 8 = Ui •U2 = cos Oi cos 02 -|- cos Pi cos P2 + cos 7i cos 72. (3)

If the two directions coincide, 0 = 0, and we have

1 = cos® o -h cos® P + cos® 7 (4)

as a relation satisfied by the direction cosines of any line. This

is the space analogue of the equation cos® 0 -f sin® 8 = 1 in the
plane.
From the equation
A‘B = A I 1 1
B I
cos 0
we get
—AB
—
COS0 = ;
( 5)
1A||B|
as the cosine of the angle between any two vectors.
§169] ROTATION OF AXES 369

Example, The lines AB^ AC in Figure 217 make with the 2-axis the
angles a, respectively. Find the angle between the two lines.

The unit vectors along AB^ AC are

Ui = i sin o£ — k cos a,
U2 = j sin — k cos jS.

The angle B between the two is therefore given by the equation

cos B = Ui-U2 = cos a cos jS.

169. Rotation of Axes. As an illustration of direction cosines

we take new axes having the same origin.
the transformation to
Let 0-XYZ and 0-X'Y'Z' be two sets of rectangular axes, the
unit vectors along the former being i, j, k, those along the latter
i', j', k'. The axes of one set determine nine angles with those of
the other set. Let the numbers in the following table be the cosines

X y z

x' h mi m
y' h m2 712

z' h ms m
of those angles, the number at the intersection of any row and
column being the cosine of the angle between the axes indicated
at the end of the row and the top of the column. These cosines
are the scalar products of the corresponding unit vectors. Thus

h = i-k' (1)

is the cosine of the angle between OX and OZ\

360 SPACE COORDINATES AND VECTORS [Chap. XV
If X, y, z and x', y', z' are the coordinates of the same point P
0-XYZ and O-X'Y'Z'y the vector OP
with respect to is represented
by two equivalent expressions
Lr + ji/ + k = iV +
2;
j V + kV. (2)

This is the vector equation for transformation of coordinates.

To obtain any coordinate of one system in terms of those of the
other system we multiply both sides of (2) by the corresponding
unit vector and use the appropriate equations of the form (1).
Thus, multiplying by i, j, k respectively, we get

X = lix' + hy' +
y = mix' + m 2y' + m^z', (3)

z = riix' + n2 y' + n^z'.

Example, The coordinate axes are rotated about the origin to new
positions 0X\ OY\ OZ' along which the unit vectors are
= + - 8kl
ill

f = i[4t + 7j +
k' = i[Si - 4j — k].

Determine the new coordinates of the point P(a:, y, z).

The vector transformation is

i'x' + fy' + k'z' = tx+ jy + kz.

Multiplying the left side of this equation by /, k' and the right side
by the equivalent values given above, we get

x' = iia; + 42/ - 8^],

y' = i[4a; + 7y + 4^],

z' = il&r - 42/ - z],

170. The Vector Product. The vector two

or cross, product of
vectors A and B (indicated by placing a small cross between the two
factors) is defined as the vector

V = AxB (1)

perpendicular to both A and B, with magnitude

|V| = |A|lBl8mfl, (2)

and so directed that right-handed rotation about V through an

angle 0 of not more than 180° carries A into B.
§170] THE VECTOR PRODUCT 361

Since
/i = B I 1
sin ^

is the altitude of the parallelo-

gram with sides A and B, the
product A X B is equal in magni-
tude to the area of that parallelo-
gram.
The product B x A has the
same magnitude as A x B, but the
rotation which carries B into A is
opposite to that which carries A
into B. Hence

B X A = —A X B. (3)

Thus the vector product is not commutative.

If A and B are parallel, B is zero or 180° and

AXB = 0.

Conversely, if the vector product is zero, one of the vectors is zero or

two are parallel.
else the

The product Ax B can be considered as the result obtained by

projecting B
on the plane MN
perpendicular to A, rotating the
projection 90° in the positive direction about A, and multiplying
the resulting vector by A Each of these operations changes a
|
|.

closed polygon into a closed polygon. If they are performed on

the sides B, C, B + C of a triangle, the
resulting vectors

A X B, A X C, AX (B + C)
thus form the sides of a second triangle.
Therefore

AX (B + C) = A X B + A X C, (4)

whence

(B-f-C)xA = BxA + CxA, (5)

since each term of (5) merely differs in algebraic sign from the
corresponding term of (4). These equations express the distribu-
tive law of vector multiplication.
362 SPACE COORDINATES AND VECTORS [Chap. XV
In case of the unit vectors i, j, k the definition of vector product
gives immediately
ixi = jxj = kxk = 0,

i X j
= k = -j X i, (6)

j
X k = i = -k X j,

k X i = j
= -i X k.

By means of these relations the cross product of two vectors

can be expressed in terms of their components. Thus, if

A= Axi + Apj + -4*k,

B = Bxi + Byi + BM,
by expanding and using (6) we find

A xB = {AyBx - AxBy)i + {AxBx -

+ {AxBy - AyBx)^. ( 7)
As a determinant this can be written

1 J k
AxB = A-X -A y Ax ( 8)
Bx B« Bx

Example 1 . Find a vector perpendicular to the plane through A (0, 2, 2)

B(2, 0, -1), C(3, 4, 0).
The vectors

AJ5 = 2i - 2j - 3k,

AC = 3i + 2j - 2k
are in the plane, and consequently their product

AB X ic = KH - 5j + 10k
is perpendicular to the plane. Since division by 5 does not alter the
direction, we may take
N = a - + 2k j

as the required vector.

§170] THE VECTOR PRODUCT 363

Example 2. Find the area of the triangle formed by the points A, C

in the preceding problem.

AH X AC = lOi - 5j + 10k

has a magnitude equal to the area of the parallelogram with vertices A, H,

C. The area of the triangle is thus

^VlO* + 5^ + 10=* =

Example 3, Find the shortest distance between the line through

A(0, 1, 2), H(l, and that through C(2, 0, 1), D(2, 1, 2).
2, 2)

The shortest distance is the common perpendicular to the two lines.

To obtain this we determine a vector

N = AHxCD = i- + k j

perpendicular to both lines and project on this a vector, such as

AC = 2i —j— k,

which extends from a point on one line to a point on the other. The short-
est distance is thus
 364 SPACE COORDINATES AND VECTORS [Chap. XV
171. Moment of a Force. If F is a force applied to a body at
the point P, and r is the vector from 0 to P, the product

M = rXF (1)

is called the moment of F about the point 0.

The force tends to rotate the body about an axis through 0
perpendicular to the plane of 0 and F. The moment is a vector
extending along that axis in the direction a right-handed screw
advances under such a rotation. In magnitude it is equal to the
product of the force by its lever arm, or perpendicular distance
from the axis.
Moment is usually defined with respect to an axis and not with
respect to a point, the moment of F about OZ being defined as

Figure 222.

the product of PiQi on the plane OXY perpendicular

its projection
to OZ, by the 0 to the line PiQi (§134). But that,
distance from
being the area of the parallelogram OPiQi, is the projection of
OPQ on XY. Now the parallelogram OPQ has an area equal to
I
M and makes with XT an angle equal to that between
I
and M
OZ. Thus the projection of the parallelogram on XY is equal to
the projection of M
on OZ. Hence, the moment of F abottf any axis
through 0 is equal to the component of M along that axis.
The distributive formulas

r X (Fi -f F2 ) = r X Fi -f r X F2 , (2)

(ri + 12 )
X F = ri X F -}-
12 X F, (3)

state that the moment

of the resultant of two forces is equal to
the sum moments, and that the moment of F about O2
of their
can be found by taking the moment about 0i and adding the
moment about O2 of a force F at Oi. These formulas are to a con-
§172] PRODUCTS OF THREE OR MORE VECTORS 365

siderable extent responsible for the importance of moment and

the vector product in mechanics.

Example 1 . The force

F = 3i - 2j -k
is applied at the point P(2, —1, 3). Find its moment about the point
A(l, 2, 3).
The required moment is

M = AP X F = (i - 3j) X (3i - 2j - k) = 3i + + 7k.

j

The components 3, 1, 7 are the moments of F about axes through A par-

allel to OX, OY, OZ.

Example 2. Given A(l, 2, 3), P(3, 1, 1), find the moment of the force F
in the preceding problem about the axis AB»

The moment of F about the axis AB is the component of

M = ^PxF = + + 7k 3i
along the vector ^
AP = 2i -i- 2k.

The value required is therefore

MAP -3.
\AB\
The negative sign signifies that F tends to produce rotation in the negative

direction about AP.

172. Products of Three or More Vectors. By scalar and vector

multiplication of three vectors A, B, C, triple products of the fol-
lowing types are obtained:

(1) (A-B)C,

(2) A-(BxC),

(3) Ax(BxC).
In these expressions parentheses have been introduced to show
the order in which the multiplications are performed. Thus to
obtain
AX (B X C)

we first obtain B x C and then multiply by A. In general the vector

(A X B) X C does not have the same value. Since the product BC
366 SPACE COORDINATES AND VECTORS [Chap. XV
without a dot or cross has not been defined, the expression A-BC
can only mean (A -3)0. Also, since both factors of a cross product
must be vectors, A-B x C can only mean A* (B x C). In products of
types (1) and (2) the parentheses may therefore be omitted. For
added clearness, however, they are often inserted.
The product (A-B)C is merely A-B times the vector C. It is
thus a vector A B times as long as C and having the same direction
•

if A-B is positive, the opposite direction if A-B is negative.

The product _ ^
A-BxC
.

isa scalar equal in magnitude to the volume of the parallelepiped with

edges A, B, C. This is easily seen since B x C is a vector perpendic-
ular to the base, of magnitude equal to the area of the base, and

the projection of A on this vector is the altitude h of the parallele-

piped. Hence
A-BxC = li|BxCl
is equal to the volume.
Inspection of the diagram shows that

A-BxC =B-CxA = C-AxB. (1)

The product is thus not changed by a cyclical permutation of the letters

A, B, C. The other three products

A-CxB = B-AxC = C-BxA (2)

are easily seen to have the opposite algebraic sign.

Since
(AxB)-C = C-(AxB),
equation (1) shows that
AxB-C = A-BxC. (3)

In the triple scalar product the dot and cross can therefore be inter-
chcmged wUhovi changing the value of the result.
§1731 SPACE LOCI 367

A= ail + 02 + aak, ]

B = 6ii + 62 + ^3^, j

C = Cii + C2j + cak,

the identity
Oi 02 03
A B xC = fel 62 bs (4)
Cl C2 C3

may be proved by showing that the product and determinant are

formed by the same laws from the components of A, B, C.
The product
AX (B X C)

isa vector sometimes encountered in work involving repeated

multiplication. It has no very simple geometrical interpretation
but can be reduced to simpler form by means of the equation

A X (B X C) = (A-C)B - (A*B)C. (5)

This and the similar equation

(AxB)xC = (A-C)B ~ (B*C)A (6)

can be proved by expressing A, B, C in terms of i, j, k and expanding

both sides.
By means of the above identities products involving four or more
vectors can be reduced to a form in which each term is at most
a scalar factor times the cross product of two vectors.

Example. Expand (A xB)-(C xD). This can be considered as the

scalar triple product of A x B, C, and D. By the interchange of dot
and cross we thus have
(AxB)-(CxD) = [(AxB) xC]-D
= [(A-C)B - (B-C)A]*D

= (A-C)(B-D) - (B-C)(A-D).

173. Space Loci. The locus of points in space with coordinates

Xy y, z which satisfy an equation

fix, J/, «) = 0
is in general a surface. For example,
« = 0
represents the x^-plane, and
+ y* + “ 1
368 SPACE COORDINATES AND VECTORS [Chap. XV
is the locus of points at distance 1 from the origin, that is, the
surface of a sphere with center at the origin.
In particular cases real solutions may, however, occur only at
discrete points or at points on a line or curve. Thus the only
real points for which
x* j/2 = 0 +
are the points a: = 0, j/
= 0 on the 2-axis, and

+ /^
2 + 2^ = 0

is satisfied only at the origin.

If an equation contains only two rectangular coordinates, the
surface it represents is not changed by motion parallel to the axis

of the missing coordinate. Such

an equation thus represents a
cylindrical surface, that is, a sur-
face generated by a line parallel
to a fixed direction and cutting a
fixed curve. The equation

+ 2
/^ = 22/,

for example, represents a circle in

the xy-plane. In space the equa-
tion is satisfied by the coordinates
of all points vertically above or
Figure 224 .
below this circle. The locus is

therefore the cylinder with axis

parallel to OZ and intersecting the xy-plsme in the above circle.

Two simultaneous equations

fi(x, y, z) = 0, /2(x, 2/, z) = 0

usually represent a curve or straight line; for each equation repre-

sents a surface, and hence two simultaneous equations represent
the points common to two surfaces. Thus the equations

+ 2/* + 2* = 2, 2 = 1

represent the circle in which a sphere and plane intersect. In

particular cases, the equations may, however, be satisfied only at
discrete points, or may have a common factor and the locus include
a surface.
§173] SPACE LOCI 369

Three simultaneous equations are usually satisfied only

by the coordinates of a definite number of points, which may be
obtained by solving the equations simultaneously. In particular
cases, there may be no solution or the equations may be satisfied
by the coordinates of all points on a curve or surface.

Example 1, Show that

+ 2^ — 2a; + 4i/ + 22 + 2 = 0
is the equation of a sphere, and find its center and radius.
Completing the squares of terms in x, z separately, we get

{x - 1)2 + iy + 2)2 +{z + 1)2 = 4.

Thus the point (x, ?/, z) is at distance 2 from the point (1,-2, — 1).
The locus is therefore a sphere of radius 2 and center (1, —2, —1).
Example 2, Determine the locus represented by

X +2= 1.

In the x2-plane the locus is the line AB (Figure 225). In space the

Figure 225.

locus is the plane generated by lines which are parallel to the y-axis and
cut AB.
Example 3. The equations

x^ + + z^ - 1, x +y+z 0 (1)

represent a space curve. Find the projection of this curve on the xjz-plane.
To do this we eliminate z from the equations of the curve. The resulting
equation
x^ + y^ + {x + yY = 1 (2)

represents a cylinder with generators parallel to OZ. Since this equation

is obtained by combining the equations of the curve, it is satisfied by the
370 SPACE COORDINATES AND VECTORS [Chap. XV
coordinates of any point on the curve. The cylinder thus passes through
the curve, and its intersection with the xy-plane is consequently the
projection required. Considered as an equation in the xy-plane, (2) is the
equation of the intersection and therefore the equation of the projection.

174. Equation of a Plane. Let P\{xij yi, zi) be a fixed point in

a given plane and ^
MAT = iA + jB + kC (1)

a vector perpendicular to the plane. A point P{x, z) will lie in

the given plane if and only if

PiP = i(x - xi) + j(y - yi) + k(z - zj) (2)

is perpendicular to MN. By taking the scalar product of (1) and

(2) we thus obtain

A(x - Xi) + B(y - yi) + C{z - Zi) = 0 (3)

as the equation of the plane which passes through (xi, yi, Zi) and
is perpendicular to the vector with components A, B, C.
A line perpendicular to a plane at a point is called normal to
the plane at that point. If a, /3, y are the angles such a normal
makes mth the codrdinate axes,

i cos a + j cos /3 + k cos y (4)

is a unit vector along the normal. Using this instead of (1), we get
(x — xi) cos a + (y — Vi) cos /3 + (a — *i) cos 7 = 0 (5)

as the equation of the plane which passes through (xi, yi, Zi) and
has a normal with direction angles a, /3, 7.
Equations (3) and (5) are of first degree in x, y, z. Any plane,
therefore, has an equation of first degree in rectangular codrdinates.
§174] EQUATION OF A PLANE 371

Conversely, any equation of first degree in rectangular coordi-

nates represents a plane. For any such equation has the form
Ax + By -(- + Z) = 0, (6)

AfBjCjD being constants. If X\^y\^ Zi satisfy this equation,

Axi “h Byi -f- Czi -f- Z) = 0.

Subtraction gives

A(x - Xi) + B(y - yi) + C{z - zi) = 0,

which has the form (3). Thus (6) represents a plane perpendicular
to the vector with components A, B, (7.

Example 1. Find the angle between the planes

X + 2y + 2z Oy a; — 4y + 82 = 5.

The vectors with components 1, 2 2 and

,
1, —4, 8 are normal to the
planes. The angle between two planes is the same as that between their
normals. By equation (5), §168, we thus have

=
1-8 + 16 1
cos d
Vs 3

as the cosine of the required angle. The planes determine two angles
less than 180®, the smaller being cos“^
Example 2, Find the equation of the plane which passes through
Pi(l, -1, 2), P2 ( 3 2, 1), and P 3 ( 2 1, 3).
,
,

The equation of the plane can be written

Ax + Py + C2 + Z> = 0.

Since this is satisfied by the coordinates of Pi, P2, P3, we have

A-P + 2 C + Z> = 0,

3A + 2B + C + jD = 0,
2 A + P + 3C + P = 0.

The four equations being homogeneous in A, P, C, P, the condition

X y z I

1-1 2 ^
3 2 11 1

2 13 1

must be satisfied (§75). By expanding the determinant we get

5x — 3y 4- « = 10

as the equation of the plane.

372 SPACE COORDINATES AND VECTORS [Chap. XV
Another method of obtaining the same result is to use the fact that the
vector

PiP2>'PiP3 = 5i-3i+k
is normal to the plane. Since the plane passes through Pi(l, —1, 2) its

equation is

5(x - 1) -+ 1) + 2 - 2 = 0.
3(y

176. Equations of a Straight Line. A straight line is the inter-

section of two planes. In rectangular coordinates it is therefore
represented by two simultaneous equations of the first degree.
Thus
x+y+z ^
2a; — + 22 = 2

are equations of a straight line. Since there are infinitely many

pairs of planes through the line it is represented by infinitely many
pairs of first-degree equations.
To obtain equations of more definite form choose a positive
direction along the lineand let a, jS, 7 be the angles this direction
makes with the coordinate axes. If Pi(a;i, 2/1 ^i) is a fixed point >

on the line and P{x, y, z) a variable point, the vector

PiP = i(a; - yi) + k(2 - zi) (1)

is then parallel to the vector

i cos a + j cos P + k cos 7. (2)

Thus
X - xi _y - yi _z - zi

cos a cos P cos 7

are equations of the line which passes through (xi, yu ^1) and
makes with the coordinate axes the angles a, P, 7.
Instead of (2) we can use any vector

iA + jB + kC (4)

parallel to the line. We then get

X — xi _ y — yi z — zi

A " B ^
as equations of the line which passes through (xi, yi, Zi) and is

parallel to the vector with components A, BfC.

§1761 CYLINDRICAL COORDINATES 373

Example 1, Find the equations of the line which passes through the
points Pi(— 2, 0) and P2 (l, 0, 1).
1,

The line passes through (1, 0, 1) and is parallel to the vector

^2 = 3i - j + k.
Its equations are therefore

X— 1 _ y 0—1
3
” ^ “"I

Example 2, Find direction cosines of the line represented by

+ +2=
a; 2/ 6,

2x — 32/ + 22 = 2.

The vectors
ni = i + + k, j

n2 = 2i — 3j + 2k

are perpendicular to the two planes. Their product

ni X n2 = 5i — 5k

is therefore a vector parallel to the line. The direction cosines of this

vector are
1 1

V2’ ’ \/2’

If the positive direction along the line is taken as that of the vector, these
are also the direction cosines of the line. If the line has the opposite direc-
tion, its direction cosines are

^ 1

176. Cylindrical Coordinates. Any three numbers which de-

termine the position of a point in space can be used to represent it.
If M
is the projection of a point P upon the xy-plane and r, 0

are the polar coordinates of M in the xiz-plane, the numbers r, 0, z

are called the cylindrical coordinates of P. The name is suggested
by the fact that the locus of points

r = constant
is a cylinder.
The angle d is measured from the x-axis and is considered posi-
tive when its direction is that which carries OX into OF by rotation
through 90°. The coordinate r is positive when Af is on the termi-
nal side of 0.
374 SPACE COORDINATES AND VECTORS [Chap. XV
The equations connecting the rectangular and cylindrical co-
ordinates of the same point can be read from a diagram (Fig-
ure 227). The most important of these are
x » r cos y = rsin 6, r == \/x^ + (1)

Example. Find the equation of a right circular cone with vertex at the
origin, axis OZ, and vertical angle 2a,
The cylindrical coordinates of any point P on the surface of the cone are

MP = r, OM = z.

Hence
r = 2 tan a
is the equation of the cone.

X
Figure 228 .

177. Surface of Revolution. The surface generated by revolv-

ing a plane curve about an axis in its plane is called a surface of
revolution.
§1771 SURFACE OF REVOLUTION 376

Suppose, for example, that the curve

f(x , «) = 0

in the a:2 -plane is rotated about the 2 -axis. Any point P on the
resulting surface has cylindrical coordinates r, z equal to the co-
ordinates X, z at some point Q on the curve. These cylindrical
coordinates thus satisfy the equation

f(r, z) = 0,

which is therefore the equation of the surface generated.

Since r — Vx* + j/*, the corresponding Cartesian equation is

f(Vx^ + z) = 0 .

Similar equations are obtained for the surfaces generated by

rotating about OX or OY.
Example 1 Find the equation of the cone generated by rotating the line

X y = 1 +
about the x-axis.
The equation is obtained by leaving x unchanged and replacing y by
Vy* + Thus
X ± \/y* + «* = 1,

and consequently
y* + 2* = (1 - x)*
is the equation required.
376 SPACE COORDINATES AND VECTORS [Chap. XV
Example 2, The circle
,2 — 2ax

is rotated about the 2-axis. In cylindrical coordinates find the equation

of the surface generated.
The equation is obtained by replacing x by r. Thus

+2 = ^ 2or
is the equation of the surface.

178. Quadric Surfaces. The locus of an equation of the second

degree in x, y, 2 is called a qvxjulric surface. All plane sections of
such a surface are curves represented by Cartesian equations of the
second degree, that is, conic sections (§65). The appearance of
the surface can be determined by the sections in a suitably chosen
set of planes.
The ellipsoid

o r ,.2 +^= 1- (1)

The X2- and y2-planes cut this surface in ellipses

z*‘
j
^ =1
^ 6"
'

having a common vertical axis. The section in a horizontal plane

2 = A; is an ellipse

+ ~o =
» ,
1
* o
b‘

the axes of which are chords

of the two vertical ellipses.
The distances a, 6, c, called
semiaxesy are the intercepts on
the coordinate axes. If two
of the semiaxes are equal, the
ellipsoid is called a spheroid.
It then a surface of revolu-
is

tion obtained by revolving an ellipse about one of its axes. If all

the semiaxes are equal, the ellipsoid is a sphere.
The hjrperboloid of one sheet
X? ^
§178] QUADRIC SURFACES 377

The sections in the xz-> and 1/2 -planes are

hyperbolas

having a common vertical axis. The sec-

tion in a horizontal plane z — k is an
ellipse,

^
or
+^= + 1

with the ends of its axes on the two Figure 231.

hyperbolas.
For different values of k all these ellipses are similar, the smallest
being in the xy-plsme.
If a = 6 the surface is a hyperboloid of revolution generated
by revolving a hyperbola about its conjugate axis.
The hyperboloid of two sheets

^ ^
( 3)
c*

The section in a plane a; = ^ is imaginary if

|
a:
|
< a. The
surface therefore consists of two parts, one for values x > a, the
other for values x < —a. The surface cuts the xy- and a:2 -planes
in hyperbolas, and is generated by ellipses with axes terminating
on these hyperbolas.

If 6 = c, the surface is a hyperboloid of revolution generated

by revolving a hyperbola about its transverse axis.
378 SPACE COORDINATES AND VECTORS [Chap. XV
^
The elliptic paraboloid

z = 03^ + by^, (4)

a and b having the same algebraic sign.

f
Sections in the xz- and j/z-planes are
parabolas
z = 03^, z = by®.

The section in a horizontal plane z = ^ is an

ellipse

+ by® = A
X
_ k has the same sign as a and b, and is
if

imagmary if k has the opposite sign. The

surface is thus generated by ellipses, the axes of which are chords
of the vertical parabolas.
If o = b, the surface is a paraboloid of revolution generated
by revolving a parabola about its axis.

The hyperbolic paraboloid

z = by® — oa:®,
a and b being positive.
The a parabola z = — oa;® with axis
section in the za:-plane is

extending downward; that in the yz-plane, a parabola with axis

Fioubb 234.

extending upward. Near the origin the surface thus rises on two
sides and falls on the other two, and hence has the general shape
of a saddle.
If a s b the equation can be changed to the form

z = Ascy

by rotating the axes 45^ about OZ.

§179] LENGTH OF ARC 379

The cone

The characteristic of this equation is that all terms are of the

same degree. If Xi, j/i, Zi satisfy the equation, kxi, ky\, kzi also
satisfy it. The surface is thus generated by
straight lines through the origin.
Equations containing first powers of the
coordinates may frequently be reduced to
the above forms by completing the squares.

Example, — 2z^ — 2x IBt/ — Az — 21. +

Completing the squares and dividing by 4,
this equation becomes

(x ^ 1)^ (y - 2Y {z + 1)^
^ .

4 1 2

Comparison with equation (3) shows that this

equation represents a hyperboloid of two sheets
with center (1, 2, —1).

179. Length of Axe. If /i(0> /2(0> /sCO Figure 235.

are given a variable
functions and Ms
parameter, the locus of points determined by the equations

X = y = /2 (0 ,
Z = fz{t) (1)

is a curve. To construct the curve we assign values to <, calculate

the corresponding values of x, !/, z, and locate the resulting points.
As t increases, the point (x, y, z) may pass more than once
through the same position. By restricting the range of ^ it is usu-
ally possible, however, to obtain an arc on which only one value
of t corresponds to each point, and conversely. In general we
assume that this has been done. Along such an arc there is as-
signed a positive direction, namely the direction in which points
are described as t increases.
The length of such an arc is defined as in §93. We divide the
arc into parts by intermediate points and join consecutive points
by chords. If the sum of these chords tends to a limit as the
number of divisions increases and the chords all tend to zero, the
arc is called rectifiable and the limit is its length.
When the functions /i(0,/2(0, /a (0 have continuous derivatives
for the values of t under consideration, we readily prove that the
380 SPACE COORDINATES AND VECTORS [Chap. XV
arc is rectifiable. For, if P(x, y, z), Q(x + Ax, y + Ay, z + Az)
are consecutive points of division and t, t-\- At the corresponding
values of t, the length of the chord between these points is

PQ = V{Axf + {Ayf + (A2)2.

By the mean value theorem (§34)

Ax=fi(^i)At, Ay = fiih) £^t, A2=/3(«3)A<,

where {i, {21 h are values between t and t + At. Thus

SPQ = [/2(l2)]"+ {/3(€)3]" A<

is the sum of the chords, and, as M tends to zero, this tends to the
limit

a, h being the values of t at the ends of the arc.

dx dy dz
At a point where t” > T" >
continuous and not all zero, we
dt dt dt
prove, as in §94, that
o *»r*
arc PQ ^
lim = 1. (3)
0 - p chord PQ
dx
Since dx ^ — dt, etc., equation (2) is usually written
dt

Vdx^ + dy^ + dz^. (4)

U
To evaluate this we express x,y,zm terms of a single variable and
use as limits the values of this variable at the ends A, B of the arc.
Considering the integral as a function of its upper limit, we have

ds = Vd3^ + dy^ + (fo2. (5)

Example. Find the length of the ciuwe

X = t, y = \, z = \/2 In t
t

between the points for which t = 1 and i = 2.

In this case

dx = dl,

and
3
2
§1801 DERIVATIVE OF A VECTOR 381

180. Derivative of a Vector. The derivative of a vector func-

tion with respect to a scalar variable is defined in space as in the
plane (§110).
Thus, if ^
r = OP
is the vector from a fixed point 0 to a moving point P,
dt
V = (1)
dt
is the velocity and
dh
a = (2 )

is the acceleration of P.
Derivatives of sums and products of vectors are obtained by
formulas like those for scalars, the only essential difference being
that the order of factors must be 'preserved in a product if order has
significance.
Suppose, for example, that U, V, W are functions of a scalar
variable t and that
W == U X V.
When U, V, W take increments AU, AV, AW this becomes

W + AW = + AU) X (V + AV)
(U

= UxV + UxAV + AUx(V + AV).

By subtraction we find

AW = U X AV + AU X (V + AV),
whence

+ — x(V +
AW
A^
= Ux —
AV AU
A^ At
AV).

If U and V are differentiable, AU and AV tend to zero with At

and we have as limit

+ — xV.
d dV dU
-(UxV) =Ux_ dt dt
(3)
dt

In both terms on the right U and V have the same positions as on

the left. It would not be correct to replace the second term by
382 SPACE COORDINATES AND VECTORS [Chap. XV
It should be noted that, if a vector V is of constant lengthy its
derivative is zero or else perpendicular to V. This may be regarded
as geometrically obvious. To prove it analytically we differentiate

V V = V
•
I
1^ = constant,
thus obtaining

V —+
dV
dt
dV
dt
V = 2V
dV
dt
= 0.

If I
*
V I

‘
is constant and —
dt
is not zero, this requires that
'
^
V and —
dt
be perpendicular.
In particular, the derivative of a unit vector u is zero or perpendic-
ular to u.

181. Tangent to a Curve. Principal Normal. Let

r = OP = Lc + jy + fcsj
(1)

be the vector from the origin to a point P{xy y, z) on a curve, and

s the distance along the curve from a fixed point Pq to P. As in
§111 we show that
dt
t = ( 2)
ds

is a unit vector tangent at P and drawn in the direction in which

s increases.
The equation
dr dx dy dz
i— + j — + k — (3)
ds ds ds ds
shows that
dx dy
— = cos a, — — coafi,
dz
— = cos 7 (4)
ds ds ds

are the direction cosines of the tangent at P.

Along the curve, t is a function of s.

Since t is of unit length its derivative

dt
-7- is perpendicular to t. This deriva-
ds
tive extends along a line called the
principal normal at P. We usually
write
dt n
(6 )
Figubb 236. ds P
§181] TANGENT TO A CURVE. PRINCIPAL NORMAL 383

where n is a unit vector along the principal normal and p is a

positive scalar called the radius of curvature. The reciprocal of

the radius - is called the curvature at P.

P
If a particle moves along the curve and has the position P at
time t, its velocity and acceleration are

dt dr ds ds
V = — == = t — (6)
dt ds dt dt

d^t d^s n /ds\^

^~d^~ t—
dt^
H
— ('
p \dt/
t) ’
(7)

The vector equations for velocity and acceleration in space thus

have the same form as in the plane (§112), the principal normal
replacing the normal in the plane.

Equation (7) shows that is the hypotenuse of a triangle with

CLL

It should be noted that if r is expressed in terms of any param-

eter t the derivatives are determined by equations (6) and (7).

In equation (8) the variable t need not therefore be the time but
may be any parameter in terms of which r and s are expressed.

Example 1. Find the equations of the tangent line to the helix

X — a cos By y = asm By z — mJB

at the point where ^ = 0.

The point where ^ =0 is (a, 0, 0). The tangent vector at that point
has components proportional to
384 SPACE COORDINATES AND VECTORS [Chap. XV
The equations of the tangent line are therefore

X — a _y _ ^
0 am
These are equivalent to
a; = a, my = az.

Example 2, The coordinates of a moving point at time t are

X = t,
y = z =
Find the tangential and normal components of its acceleration when t = 1
The velocity and acceleration are

V = + + toS
i

a = + j 21rf,

and its speed is

= vT+T+l* = \/3.

The tangential component of acceleration is

< + = V3,
'! + +
<=* 1*

and its normal component is

a„ = y/ — <i?t = y/5 — 3 = '\/2.

Example 3, Find the radius of curvature of the helix

X = acosdy y — asm 9 j
2 = mS,

Taking derivatives with respect to 9, we have

= — ia sin 0 + ja cos ^ + km

= — ia cos 9 — ia sin 9
au^

Substituting in (8) with t replaced by 9, we get

(a^ + m^)
Figure 238.
§182] DERIVATIVES IN CYLINDRICAL COORDINATES 385

182 Derivatives
. in Cylindrical Coordinates. At a point
P(r, 6, z) three coordinate directions are determined, namely the
directions in which P moves when
two of the coordinates are kept con-
stant and the third increases. The
unit vectors Ur, VLe in the directions
of increasing r, By respectively, evi-
dently have the same values as in
the plane (§113). They are there-
fore functions of 0 with derivatives

dUr due
= VLe, = -Ur (1)
IF dd

The unit vector in the direction of

increasing z is the constant k. By
the components of a vector in cylindrical coordinates we mean its

components along these unit vectors.

The vector from the origin to P is

r = OM + MP = rUr + kzy (2)

where r is the cylindrical coordinate. Along a curve r is a function

of the arc length s from a fixed point of the curve to P. By differ-
entiating (2) and using equations (1) we obtain

dr dr dO dz
t = — = Ur [- Vier hk — (3)
ds ds ds ds

as the unit tangent vector. This shows that

dr dd dz

ds^ ds^ ds

are the cosines of the angles the tangent makes with the unit
vectors at P.
By squaring both sides of (3) we have

whence
= dr^ + d^^ + dz^ 4
( )

is the square of the differential of arc.

386 SPACE COORDINATES AND VECTORS [Chap. XV
By differentiating with respect to the time we find
dr dr dd dz
V h k— } (6 )
It dt dt dt

as the velocity and acceleration of a moving particle. These are

merely the polar expressions in the xy-plane plus the components
parallel to the «-axis.

Example 1. Find the length of the helix r = a, 2 = from ^ = 0

to ^ = 27r.

On the helix
ds^ = dr^ + r2 dO^ + dz^ = (a^ + m^) dB\
Thus
=
8
=J^ \/ dd 27r\/

is the length required.

Example 2, The conical helix

r = <, B — z - t

is described on the cone r *= 2 . Find the angle at which it intersects the

generators of the cone.
Along the helix the unit tangent vector has cylindrical components
proportional to
dr dd dz
~ ~
^ ’’di ^
Assuming the unit tangent drawn in the direction of increasing t, it is then

u, + <ug + k
’

V2 + <*

the denominator being determined by the fact that 1

1 1
= 1. Along a
generator of the cone r = 2, ^ = constant. The unit vector along the
generator thus has components proportional to

dr A
' ^ dz 1.
dz ^
dz

If drawn in the direction of increasing 2 , it is then

u =
Ur +k
V2
§183] ANGULAR VELOCITY 387

The angle 0 between the tangent and the generator is determined by the
equation

cos <^ == t-u

1 + 1 V2
\/2\/2 + t
^
\/2 + t
^

183. Angular Velocity. If a rigid body rotates with angular

speed about a fixed axis, its angular velocity is defined as the
CO

vector 0) with magnitude co which extends along the axis in the

direction a right-handed screw advances when subject to the
given rotation.
Let be a fixed point on the axis and r the vector from 0 to
0
the point P in the body. This point moves in a circle of radius
I
r sin 0 with speed co r sin 0, where 6
I | |

is the angle between r and co. Its ve-

locity is therefore

dt
V = — = (0 X r. (1)
dt

For this vector has the magnitude

1
V I
= CO
I
r I
sin 0

of the velocity,and the definition of ^

vector product shows that it has the Figure 240.

correct direction.
If Ti, 12 are the vectors from 0 to two points Pi, P2 in the body,
equation (1) gives

—=
dri

dt
CD X n, —=
dx2

dt
CO X r2

as the velocities of those points. By subtraction we get

d
-j(x2 - ri) = (OX (12 - ri);
dt
that is,

y P^2 = PiP2- (2)

dt
>

This equation is satisfied by any vector P P2 1 forming part of the

body.
In particular,if the rectangular axes 0-XYZ are attached to

the rigid body and rotate with it, the unit vectors along those
388 SPACE COORDINATES AND VECTORS [Chap. XV
axes are variable, their derivatives being given by the equations

di
= 0) xi,
dt

= <» X j, (3)
dt

dk
= CO X k.
dt

Example. At a point of latitude 6 on the surface of the earth, rectangu-

lar axes are used with OZ vertical, OX directed toward the east, and OY
toward the north. Find the rates of change of i, j, k due to the rotation
of the earth with angular speed o> about its axis.

Figure 241.

The angular velocity is a vector of magnitude w directed toward the

north along the axis of the earth. This vector is in the 2/2-plane and makes
with the 2-axis the angle 90® — 6. Thus

(0 = co(k sin 0 -h j cos d).

The required rates are therefore

—=»X = i w(j sin 0 — k cos d),

d]
3- = CO X
.

j
= —col
. .

sm «
0,
at

*=
— CO X k =
1 .
«i cos
z*
 PROBLEMS 389

PROBLEMS
1. Determine the distance between the point (x, y, z) and the x-axis.
2. Through the point P(x,
y, z) a line is drawn perpendicular to the y-axis.
Find the coordinates of the point in which it intersects the i/-axis.
3. Find the distance between the points Pi(l, —2, 3) and P2 ( — 1, —3, 4),

4. Find the unit vector having the same direction as

A = i ~ 2j + 2k.
6. The vectors
A « i 4- j
- k,

B = 3i - 2j - k

extend from the origin. Show that the line joining their ends is parallel to
the xy-plane, and find its length.

6. By showing that AB and BC are parallel prove that the points

A(l, 2, -1), B(0, 5, 1),C(-2, 11, 5) are on a line.
7. If A, B are vectors and x, y, scalars, show that A, B, and xA + 2/B are
parallel to a plane.
8. AE is the diagonal of a parallelepiped with edges AB, AC, AD. Given
A(l, 1, 0), B(2y 3, 0), C(3, 0, 1), D(2, 1, 4), find the coordinates of E.
9. Given A(3,4,5), B(l, 2, -1), C(3, 0, 2), D(5, 2, 6), find the middle
points ofAB and CD. Then find the middle point P of the line joining the
two middle points. Show that

PA + PB + PC + PD = 0.

10. The vectors from the origin to the points A, B are

A = 7i - 3j + 5k, B = 2i + -
j 6k.

Find the vector from the origin to the point on the line AB, one-third of the
way from A to B.
11. If three vectors A, B, C satisfy an equation

mA -|- nB -|- pC = 0

with coefficients -w, n, p not all zero, show that the plane parallel to two of
the vectors is parallel to the third, and consequently that the three vectors
are parallel to a plane.
12. If A, B, C, P are vectors from the origin to the points A, B, C, P and

P = mA -f nB -b pC,

1 =m 4“ n 4* P,

show that the vectors AP, BP, CP are in a plane, and consequently that the
four points A, B, C, P are in a plane.
390
18. SPACE COORDINATES AND VECTORS [Chap. XV
If A, B, C, D are vectors from the origin to the points A, B, C, D, and

AB « mCD,
show that
B ~ rnP _ A — mC
1 —m 1 —m *

and
16.consequently that this is the vector from the origin to the intersection of
AC and BD,
14. The points A, B,C are vertices of an equilateral triangle of side x,
— — > — —
Determine the products, (a) AB^BCy (b) AB‘AC.
Show that the vectors
A = i + 4j + 3k,
B = 4i + 2j - 4k
are perpendicular.
16. Show that the vectors
A = 2i-j + k,
B = i - 3j — 5k,

C = 3i - 4j - 4k
form the sides of a right triangle.
17. If Ui, U2 ua are mutually perpendicular unit
,
vectors, show that
20.
(Ui + U2 + Us)* = 3.

18. Show that

(B-B)A - (A-B)B
is perpendicular to B.
19. Given
21.
A = i + 3k, - 2j

B = 3i + j + 2k,
22.a vector in the plane of
find A and B perpendicular to the :r>axi8.

Given
A = - j + k,
2i

B = i + 2j - k,
C « i + j - 2k,
find a vector parallel to the plane of B and C, and perpendicular to A.
Given
A- i - 3j + 2k,
B - 3i - 4k,

find the component of A along B.

Find the component of the vector

A- 2i -j+k
along the vector
B - i +2 - j 2k,
 PROBLEMS 391

23 . By squaring both sides of the equation

a =b — c

and interpreting
. the result geometrically, prove the formula

— 2bc cos A
expressing the square of one side of a triangle in terms of the other two sides
and the included angle.
24 Determine the angles the vector
.

V « 2i + 3 + 6k j

makes with the coordinate axes.

26 A line through the origin makes angles of 60° with both OX and OY.
Determine the angle between this line and the 2-axis.
26 Find the angle between the vectors
.

A = i -h j + 2k,
B = 2i - j -h k.

27 Find the angle between the vectors

.

a = 4i + + k,
j

b = 5i 4- 2j + 5k.
28 Determine the interior angles of the triangle with vertices A(0, 2, —2),
.

B(4, 1, -1), C(5, 0, 3).

29 Determine the interior angles of the triangle with vertices A(l, —2, 1),
.

B(2,0, -1), C(2, -3, 5).

30. Show that
i' = |(i + 2j ~ 2k),

j' = |(2i -f- j + 2k),

k' - i(2i ~ 2j - k)

form a set of mutually perpendicular unit vectors. If new axes OX\ OY\ OZ*
have the directions of these unit vectors, find the equations expressing x, y, z
in terms of x', y\ z\
31 Given
.

A = - j + 2k,
i

B = 2i 4- j - k,
find a vector perpendicular to both A and B.
32 Find the area of the parallelogram two edges of which are formed by
.

the vectors A » i 4* j, B « 2j — 4k.

33 Find the area of the triangle formed by the points A(l,
. 0, 0), B(0, 2, 0),
C(0, 0, 3).
34 Find the area of the triangle with vertices
. A (1,2,0), J5(3, 3, 0),
C(5, -1,0).
392 SPACE COORDINATES AND VECTORS [Chap. XV
85. In problem 30 show that i' x j' = k' and that the vector

ixi'+jxj' + kxk'
makes equal angles with i and i', j and j', and k and k^
36. Show that A is zero if A‘B = 0 or Ax B = 0 for all values of B.
87. Show that A is zero ifA*B = A><B=*0 and B is not zero.
88. The vectors from the origin to the points A, R, C are

A =i+j ~ k,

B - 3i + 3i + 2k,
C = 3i -j- 2k.

Find a vector N perpendicular to the plane ABC. By projecting A upon N

from the origin to the plane.
find the distance
89. Given A(l, 0, 0), R(0, 1, 1), C(l, —1, 1), find the vector which extends
from the origin to a point on the plane ABC and is perpendicular to that
plane.
40. Given
A = i - j + 2k,
B = i ~ j,

C = j
- k,

find the vector which is the projection of A upon a plane parallel to B and C.
41. Given A(l, 2, 3), R(0, 1, 0), C(l, 1, 0), find the projection of the point
C on the plane OAB.
42. Find the distance between the line through A(l, 1, 0), R(l, 0, 1) and
46.through C(0, 1, 1), /)(1, 0, 0).
that
43. Find the distance between the line through A(l, —2, —6), B( — 1, 1, 1)
and that through C(4, 5, 1), D(l, —1, 3).
44. The force
F = ia + j6 + kc
is applied at P{x, y, z). Find its moment about the origin and its moment
about the rr-axis.

45. The force

F = 2i - 3j +k
is applied at the point P(3, —2, 0). Find its moment about the point (2, 3, 1).
The force
F = 3i - 2j + 6k
is applied at the origin. Find its moment about the axis through (1, 2, 3)
parallel to OX.
47. A force
F - i + + 2k
j

is applied at the point (1, 2, 3). Find its moment about an axis through the
origin directed toward the point (2, 3, 1).
 PROBLEMS 393

48. Show that a force is zero if it has equal moments about two points not
on a line parallel to the force.
49. Show that the identity

V « ilix(Vxi) + jx(Vxj) H-kx(Vxk)l

is satisfied by every vector V.

60.Find the volume of the parallelepiped three of whose edges are i + j,
j -f- k, and
k + i.
61. Given A(-l, 1, 2), B(0, 2, 3), 0(1, 1, 1), Z>(-1, 3, 3), find the volume
of the parallelepiped three of whose edges are AB^ AC^ AD.
62. Find the volume of the tetrahedron with vertices A(l, 1, 0), B(3, 2, —1),
66.
C(-2, 1, 1), and D(2, -1,0).
63. By means of products express the condition that three vectors A, B, C
be parallel to a plane.
64. By means of products express the condition that the plane containing
the vectors A and B be perpendicular to that containing C and D.
By means of products find a vector which is in the plane of B and C
and is perpendicular to A.
66. By means of products express the condition that four vectors A, B, C, D
be parallel to a plane, assuming that no two of the vectors are parallel.
67. If A B xC is not zero, show that

®xC-fwCxA + nAxB
A B xC
is 69. solution of the equations
the

V-A = I, V-B = m, V-C = n.

68. If A*B ~ 0 and A is not zero, show that

A XB -h pA
AA
is the solution of the equations

VxA = B, V A « p.

Assuming that A-B xC is not zero, find the vector V which has the
components i, m, n along A, B, C.
60. Given
A = i - 2j -f 2k,

B = 2i + - 2k,
j

C = 2i -I- 2j - k,

find the vector V with components 1 along A, 2 along B, and 3 along C.

61. If A *= PP', B = PQ, C = P'Q', and Bx C is not zero, show that

^
ABxC
“ IB xCl

is the perpendicular distance between the lines PQ and P'Q'.

394 SPACE COORDINATES AND VECTORS [Chap. XV
62 . Prove the formula

(A X B) X (C X D) - (A*B X D)C -- (A-B x C)D.

Determine the loci represented by the following equations:

63 . 2a: -f 3y = 6. 64 . « 4a:.

66. + z^ -- 4. 66. ^ 1.

67 . Find the equation of the sphere with center (2, 0, 1) and radius 3.
68. Find the equation of the circular cylinder with axis OY and radius r.
69 . Find the equation of the plane which is parallel to the a:-axis and passes
through
. the points (0, 1, 2), (0, 3, 4).
70 . Find the projection of the curve = o^, on the
xe-plane.
71 . Show that the curve

^2 _j_ ^2 _ 2^, 2 = X -f 4
is on the cylinder
(x - 1)* + 1/* = 9.

72 . Find the equation of the cylinder which passes through the curve

X y

and has generators parallel U) OY.

73 Find the equation of the plane which passes through
.
(
— 1, 2, 1) and is

perpendicular to the vector 2i + 3j — k.

74 Find the equation of the plane through the origin
. if a =* 60°, ^ = 45°,
7 = 60° are the direction angles of its normal.
76 Show that

-+?+-
a h c
1

represents the plane with intercepts a, 6, c on OX, OF, OZ, respectively.

76 . Find the equation of the plane which passes through (2, 1,-3) and is
parallel to the plane 3x — y -f 22 =» 4.
77 . Find the equation of the plane through the points A (0, 1, 2), B( — 1, 0, 1),
C(2, 3, 0).
78 . Show that the planes x — 2^/ +2 =» 3, 3x — 6y + 32 = 1 are parallel.
79 . Show that the planes x -f- 2y — 22 =* 6, 2x + y + 22 ~ 7 are perpen-
dicular.
80 Find the angle between the planes x + y = l, y4-«“2.
.

81 . Find the angle between the planes x + 2y -f 22 =* 0, x — 4y -f- 82 « 0.

82 . Find the angle between the vector 7j + 8k — 7i and the plane
a: + 8y + 42 « 9.
88 . Find the equations of the line which passes through the points
Pi(2, -1, 3), P2(4, 2, ~1).
84 . Find the equations of the line which passes through the points
(2, 3, ~1), and (3, 4, 2).
86 . Find the equations of the line which passes through (3, 1, '->2) and is
parallel to the vector 4i — 2j + k.
 PROBLEMS 395

86. Find the equations of the line which passes through (2, 1,-3) and is

perpendicular to the plane 4x — 3y -|- 2 = 5.

87. Find the equation of the plane wliich is parallel to the 2/-axis and passes
through the line x -f 2?/ 32 = 4, 2a; 1/ z == 2. + +
88. Find the equation of the plane which passes through the line x + y = 3,
22/ -f- 32 = 4 and is parallel to the vector 3i — j 2k. +
89. Find the angle between the line

X
~l
— l
~
y-k-l
“
z

I
and the line
x y-l " z-hl
” -2 ”5
5

90. Find the angle between the lines x-f-2/“*2*=0, x +2 = 0 and

a; — 2/
= - 32/ 2 = 0.

91. Find the angle between the lines 2a; — 22/ — z = 0, 2a; + + 22 =
2/ 3
and 3x — y — 2 = 0, 2x — 22/ — 2 = 2.

Determine the loci represented in cylindrical coordinates by the following

equations:
92. r = a cos 6, 98. r = o sec 6.

94. 0 = 7 * 95. r = 02 +6 .

Find the equation of the paraboloid generated by rotating the parabola

96.
2/2 4 pa; about the x-axis.
97. Find the equation of the surface generated by rotating the hyperbola
X2/ = 1 about the 2/-axis.
98. Find the equation of the spheroid generated by rotating the ellipse

about the x-axis.

99. The line 2x + 32 /
= 6 is rotated about the 2/-axis. Find the equation
of the cone generated.
100. Find the equation of the cone with vertex (1, 2, 3) and axis parallel
to OZy if the generators make angles of 45° with the axis.

101. In cylindrical coordinates determine the equation of the surface gen-

erated by rotating the circle (2 — 1 )^ + (x — 2 )^ = 16 about the 2-axis.

Determine the quadric surfaces represented by the following equations:

102. x2 + 22/2 + 32^ « 6.
108. (X - 1)2 + 2(2/ + 1)2 + 3(2 - 2)2 = 6 .

104. x2 + 4(2/2 + 2 « 12. ^) 106. x 2 - 2/2 ^ 22 « 1.

*2
107. x2 - 2/^ + 22 - 2x + 42
/ 4

108. x2 - 2/2 = 22 109. x2 + 22 « y,

110 . 2 = 2x 2/.
111. 2 = xy + X + 2/.

112. 2/2 « ;52 118. r » a sin 0,

114. r2 + 422 - 4 . 116. 2 r + 32 - 6 .

396 SPACE COORDINATES AND VECTORS [Chap. XV
116. Find the length of the curve
0? = sin 6, y
— 2 = 1— cos B

from ^ = 0 to 0 = 27r.

117. Find the length of the curve

X y ’
\/2
from f = 0 to < = 3.
118. Find the length of the curve

X — ty y — 3t^y 2 6/8

from ^ =
0 to < = 2.
119. Find the length of the curve

= In sec x, 2 = In (sec x + tan x)

from X = 0 to X = 7 .

4
d£
120. Find the derivative of F x
^ with
at
respect to L

121. Find the derivative of F • —

dF
dt
^ -—5-
dt
with respect to /.

122. Find the equations of the tangent line to the curve

X — iy
y — t^y z —
at the point where / = 2.

123. Find the equations of the tangent to the curve

126.
X = / — 1, y - + ty z =
at the point where it crosses the y2-plane.
124. The heUx
X = a cos By y = asinBy z = m$
winds around the cylinder x^ -f Find the angle at which it intersects
the generators of the cylinder.
On the curve y = sin x, 2 = cos x find the unit vectors t and n, and
the radius of curvature.
126. On the curve
y = In sec X, 2 = In (sec x + tan x)
find the unit vectors t and n, and the radius of curvature.
127. A particle moving in a curve has the coordinates

X = t,
y =
at time t. Find its velocity and acceleration, and the radius of curvature of
the path at the point where / = 1.
128. Find the length of the curve r ^ aZy z = from d = 0 to 0 = 1.

129. Find the length of the curve

r = o cos ty 0 = sec /, 2 = a sin /

IT
from / = 0 to /
’
6
 PROBLEMS 397

180.
131. The curve r =» sin 0, 2 = In sec $ is on the cylinder r = sin 0, Find
the angle at which it intersects the generators of the cylinder.
A point P moves with constant velocity
dr dz

in the r2-plane while that plane rotates about the z-axis with constant angular
velocity w. Find the acceleration of P.
132. A particle moving in a conical helix has the position

r = 6 = ty z — t

at time t. Find its velocity and acceleration and the radius of curvature of
the path at the point where ^ = 0.

133. A rigid body including the coordinate axes rotates with angular
velocity
(i> = i — 2j + 4k
about an axis through the origin. Find the velocity of the point (1, 2, —1)
and the rate of change of the unit vectors i, j, k.

134. If
dPk

dt
« X A, —=«
dB
dt
><
B, show that

^(A xB) = <0 X (A xB).

dt

136. The position of a moving point P is determined by coordinates

Zy 2/,
z with respect to axes which rotate as in Figure 241. The vector from
the center of the earth to P is

r = ro H- Lr + ji/ + kz,
<Pt
where ro = CO. The acceleration of P is ,
the differentiations being per-
dt
formed with ro, i, j, k, x, z all variable. At the instant when P passes through
tpTo
Oy express its acceleration in terms of the acceleration —r
(Ur
of 0, and the

velocity
,dx ,dy dz

and acceleration
,dPx .JPy ,

dt* •’dt* dt*

of P with respect to the moving axes.

 CHAPTER XVI
PARTIAL DIFFERENTIATION

184. Functions of Two or More Variables. Continuity. A

function f(x, y) is called continuous at x = xq, y = yo if f(xo, yo)
has a definite value and/(a:, y) tends to f{xo, yo) as limit whenever
X tends to Xq and y tends to yo- The analytic condition for this
is that to each positive number e there correspond a positive
number 5 such that

\f{x,y) -f(xo,yo) I
< e

whenever |
a: —
xo and y — yo are both less than 8.
| | |

Continuity of a vector function F(x, y) is defined in the same

way, vector F(a:, y) merely replacing scalar f{x, y) in the above
statement. Similar definitions may be given in space or on a
range of any finite number of variables.
186. Partial Derivative. Let /(x, y) be a function of two inde-
pendent variables x and y. When y is kept constant, /(x, y) is a
function of x only. The derivative of this function with respect
to X is called the partial derivative of /(x, y) with respect to x, and
is d^oted by
— or ft ^
/r(x,y).
dx

Similarly, if we differentiate with respect to y with x constant,

we get the partial derivative with respect to y,denoted by

—
9f
or /»(x,y).
dy
For example, if

fix, y) = + xy + 2y*,
§186] DEPENDENT VARIABLES 399

Likewise, if a function, scalar or vector, depends on any number

of independent variables, its partial derivative with respect to
one is obtained by differentiating with all the other variables
constant.
186. Dependent Variables. It often happens that some of the
variables are functions of others. For example, let

u = xyz^

and let 2 be a function of x and y. When y is constant, z is a func-

tion of X and
—
du dz
= xy~- + yz-
dx dx
Similarly

—
du dz
= xy-- + xz.
dy dy

If, however, we consider z an independent variable,

du du
— ^yz, — =
dx dy

The value of a partial derivative thus depends on what is constant as

well ason what is variable during the differentiation.
The variables kept constant during differentiation are sometimes
indicated by subscripts. Thus, in the above example

dz
xy h yz-
dx

Example, If a, h are sides of a right triangle

with opposite angles A, B and hypotenuse c, find

From the diagram (Figure 242) it is seen that

o = csinA. Thus

sin A,

Also a =* \/c^ — 6^, and so

b C
^da\ c
{-) “ A
Figure 242.
\dc/t sin
400 PARTIAL DIFFERENTIATION [Chap. XVI

187. Geometrical Representation. If x, y, z are taken as rec-

tangular coordinates of a point P(x, 2 ), the equation

2 = /fe y)

represents a surface. The points with given value of y form a

curve AB in which a plane y = constant cuts the surface. On
this curve z is the vertical
and X the horizontal coordi-
nate, and

dz
= fx{x, y)
dx

is the slope with respect to

the xy-plane. Similarly, the
locus of points with given x
is a curve CD with slope

dz
= y)-
dy

Through any point P{x, 2/, z) of the surface there pass two curves
y = constant, x = constant, and these curves have the slopes
y), fy(x, y) at P.
At a top or bottom point of the surface the tangents to the
curves AB, CD are horizontal and

dz dz
^
dx dy

Example. Find the highest point on the surface

z = 2a; + 4y — — y^.

At the highest point

^ ^ ^ ^ dz ^
__3-2..0,
2;
-.4-23-0. ^

Consequently x = 1, 2/
= 2. These values substituted in the equation
of the surface give z = 5. The point required is therefore (1, 2, 5).

188. Increment. Let/(x, y) be a function of x and y with partial

derivatives fx{Xj y) and /^(x, y). When x, y change to x Ax, +
y + A2/, the increment of /(x, y) is

y) =fix + ^,y-\- Ay) - fix, y). (1)

§188] INCREMENT 401

This can be written

A/(a:, y) = /(x + Ax, j/ + Ay) - f(x, y + Ay) + f(x, y + Ay)

- fix, y). (2)

By the mean value theorem (§34)

fix ->r Ax, y + Ay) - f{x, j/ + Ai/) == Ax/*(xi, y + Ay), (3)

where Xi is between x and x + Ax. Similarly,

fix, y + Ay) - fix, y) = Ay fyix, yi), (4)

where yi is between y and y + Ay. These values being substituted,

(2) becomes

A/(x, y) = Ax/*(xi, y + Ay) + Ay fyix, yi). (5)

As Ax and Ay approach zero, Xi approaches x and yi approaches

y. If the partial derivatives are continuous,

Uixi, y->r Ay) = /*(x, y) + «i,

fvix, yi) = fyix, y) + t2,

«i and «2 tending to zero with Ax and Ay. In virtue of (5) we have

thus proved:
At a point where the partial derivatives fxix, y) and fyix, y) are
continuous,

^fix, y) = —
df
Ax H
df
Ay + «i Ax + €2 Ay, (6)
dx dy

€i and 62 tending to zero when Ax and Ay tend to zero.

The expression
df
Ax H Ay (7 )
dx dy

is called the principal part of A/. It differs from Af by the amount

ei Ax + <2 As Ax and Ay tend
Ay. to zero, if the partial deriva-
tives are continuous, ci and <2 tend to zero, and so this difference
becomes an infinitesimal of higher order than the larger of the
increments Ax, Ay. When Ax and Ay are sufiBciently small the
principal part thus gives a satisfactory approximation for the
increment.
402 PARTIAL DIFFERENTIATION [Chap. XVI
Analogous results can be obtained for any number of independ-
ent variables. For example, if f(x, y, z) is a. function of three inde-
pendent variables x, y, 2 , the principal part of Af is

—
df
Ax -]
df
Ay H
df
Az,
dx dy dz

df df df
At a point where the partial derivatives — — — are continuous,
,

dx dy dz
,

Af = —
Of
Ax
Of
Ay H
df
A2 + €1 Aa: + €2 Ay + 63 Az,
dx dy dz

€ 1 , ^2 ) €3 tending to zero when Ax, Ay, and Az tend to zero.

189. Total Differential. If x and y are the independent vari-
ables, the principal part of Af{x, y), written

df(x, y) —
df
Ax + — Ay,
df
(1)
dx dy

is called the differential, or total differential, of /(x, y).

This definition applies to any function of x and y. The partic-
ular values / = x,f = y give
dx = Ax, dy = Ay. (2)

Thus (1) can also be written

df{x, y) = —
df
dx H
df
dy. (3)
dx dy
The quantities

are called partial differentials. Equations (2) and (3) state that
the differentials of the independent variables are equal to their incre-
ments, and the differential of a function is equal to the sum of the
partial differentials obtained by letting the variables change one at a
time.
Similar results are obtained for fimctions of any number of
variables. Thus, if x, y, z are independent variables,

df dj_
df(x, y, z)
dx
Ax -1

dy
Ay + Az.
dz
§189] TOTAL DIFFERENTIAL 403

The particular values f — x, f==y, f — z give

dx = AXf dy = A?/, dz = A2,

and hence

= —
df
dx -\
df
dy
df
dz.
dx dy dz

Example 1. Find the total differential of the function

u = xhj + — xyz — 2*.

By equation (4)

du = dudu
— dx + — dy + — dz
du ,
.

dx dy dz

= {2xy + y’^ — yz) dx + + 2xy — xz) dy — {xy + 3z^) dz.

Example 2. Find the increment and the differential of

r =
if X and y are the independent variables and x = 3, y = 4, Ax = 0.1,
Ay = —0.2.
The differential is

dr = — Ax + — Ay = —7=^^:==^ = -0.1.
dx dy y/ + y^
The increment is

Ar = V(3.1)2 + (3.8)2 _ v'sz + 42 -0.096.

Example 3. Find the maximum error in

due to errors of 1% in x, y, and 2.

The error in u due to small errors Ax, Ay, A2 in x, y, 2 is Aw. If the

increments are sufficiently small and 2 not too near zero, this will be ap-
proximately
y , X , 2xy ,
dw = -2 dx + ^
, ,

“2 dy dz.

Dividing the left side by w, the right side by »

we get
z*

dw _ dx
^
dy
__ 2^
w ~ X 2
y

Now — is the relative error in x. The equation just obtained thus states
X
that the relative error in u is the sum of the relative errors in x and y
404 PARTIAL DIFFERENTIATION [Chap. XVI
minus twice that in z. Since the algebraic signs could be such that all
the errors add, the maximum error in u due to errors of 1% in x, y, z is
then 4%.

190. Derivative of a Composite Function. Let /(x, y) be a func-

tion of two variables and let x,yhe functions of two variables
x, y,
s, t. Then /(x, j/) is a function of s and t. To determine the partial
derivative of this function with respect to t we leave s constant
and change < to < -f- At. The
variables x, y change to x Ax,
y -f Ay, and, assuming the partial derivatives continuous,

A/ = —
dS
Ax -I
df
Ay + eiAx + t2Ay (1)
dx dy

is the resulting increment in /(x, y). Consequently

A/
— = df Ax
1
df Ay
h ei
Ax
h *2
—
Ay
•

Ai dx At dy At At At

When At approaches zero, Ax and Ay, and so «i and «2 , tend to

zero, and this gives as limit

dt dx dt dy dt

or —
dx dy
If X or w is a function of t only, the partial derivative —
dt dt
dx dy
becomes a total derivative — or — . If both x and y are functions
dt dt
of t only, /(x, y) is a function of t with derivative

df df dx df dy
(3)
dt dx dt dy dt
Similar results are obtained for any number of variables. Thus
if /(x, y, z) is a function of three variables x, y, z and these are
functions of t and other variables,

dl dfdx dfdy dfdz

’ (4 )
dt dx dt dy dt dz dt

df df df
provided that
.
— — — are continuous functions of
,
dx dy dz
, x, y, z.

The term
dfdx
dx dt
§190] DERIVATIVE OF A COMPOSITE FUNCTION 406

is obtained by differentiating f(x, y, z) with respect to t, leaving

y and z constant. Equations (2) and (4) state that, if f is a func-

df
tion of several variables^ — is obtained by differentiating as if only one
dt

of these were variable at a time and adding the results.

Example 1. If /(x, 2/)

= — x^y + y^, x = t y = find ~
By equation (3)

dt dx dt dy dt

-
= (4x’ 2xy)
^ + (2y - X*)
^
= 2 «» + e-2t
= 4f*.

This may be checked by substituting x = t y = before differentiating.

In this way we obtain
/(x, 2/)
= = t^

and

as before.
I-
Example 2. If x and y are independent variables, 2 is a function of

X and 2/, and u = /(x, 2/,

2 ), find — .

dx
Since ^
ox
is zero, equation (4) gives

du_df df dz
dx dx dz dx

Example 3. If 2/
= find “•
CU/

The function x® may be considered as a function of two variables, the

lower X and the upper x. If the upper x is kept constant and the lower
allowed to vary, the derivative (as in x”) is

XX* ‘
= X*.

If the lower x is kept constant while the upper varies, the derivative (as
in a*) is
X® In X.

When both vary the derivative is the sum

~ = X* + X® In X.

Compare §88, example.

406 PARTIAL DIFFERENTIATION [Chap. XVI
Example 4 , If the substitution

X = 2w + 3t;, y = 3u — 2v

changes /(x, y) into F(w, y), determine —

du
in terms of derivatives of /(x, y)

with respect to x and y.

By assumption
F{Uy v) = /(x, y).

Because of the equations connecting x, y and % v, both sides of this equa-

tionmay be regarded as functions of u and v. Differentiating both sides
with constant, we get
t;

du dx du dy du dx dy

Example 5, If X and y are independent variables and u = /(2x — 3?/),

show that
du
3 = 0 .

dx

For convenience of notation we take 2x — 82/ = v as a new variable.

We then have u = /(«;),
^ — 2 “ *
dx dv dx dv

—— — —3 ^
dy dv dy dv *

whence
= 0 ,

which was to be shown.

191. Homogeneous Ftmctions. A function is said to be homo-

geneous of the nth degree in a given set of variables if multiplication
of each variable by k multiplies the function by
Thus /(x, y) is a homogeneous function of the nth degree in x
and V if
f(kx,ky) ==k”f{x,y) (1)

for all values of k. Since k is arbitrary we can differentiate both

sides of this equation with respect to k, leaving x and y constant.
For convenience in doing this let

u — kx, V = ky.

The equation then becomes

/(u, v) = k”f{x, y).
§192] PROPERTIES OF THE DIFFERENTIAL 407

By equation (2), §190, we have

df du df dv

that is,

3/ df
(2)
du dv

This holds for all values of k. Putting A: = 1, we get u = x,v = y,

^_df df _df
’
du dx dv dy
and (2) becomes

x —
df df
-\-y— = nf. (3)
ax dy

In a similar way we show that, if /(xi, X2, * •

•
,
x^) is a homoge-
neous function of the nth degree in Xi, X 2 ,
• *
•
, x„i,

xi V X2-
df
1
\r Xm-
df
—= n/. (4)
dXi dX2 OXm,

This is known as Euler^s theorem on homogeneous functions.

Example 1, The function

/ = \/x^ + sin

is homogeneous of first degree in x and y. Hence

as may be directly verified by calculating the partial derivatives and sub-

stituting in the equation.
?/
Example 2. If m = - ,
y = — ,
/(w, v) = 0, we can consider u and v as
X X
homogeneous functions of zero degree, and z of degree J, in x and y. Thus

dz . dz 1

dx dy 2

192. Properties of the Differential. The utility of the differ-

ential is due largely to the following properties, which we state
for three variables, the extension to n variables being obvious.
408 PARTIAL DIFFERENTIATION [Chap. XVI

(1) The differential of f(x, y, z) is

df = —
df
dx
df
dy -]
df
dz,
dx dy dz

whether x, y, z are the independent variables or not

Suppose, for example, that there are two independent variables
s, t and that x, y, z are functions of s and t By the definition of
differential we then have
dx dx
dx = — As H At,
ds dt

dy = — As H At,
ds dt

dz dz
dz = — As H At,
ds dt

df = — As H At. (1)
ds dt
Also [§190, (2)]
df _dfdx dfdy dfdz
ds dx ds dy ds dz ds

df df dx df dy df dz
_
dt dx dt dy dt dz dt

These values being substituted, (1) becomes

_ /dfdx dfdz
df
\dx ds
+ fL^
dy ds
+ ^-)As
dz ds.)
+
Vda: dt
^ dy dt dz dt)

df (dx dx df/dy dy df (dz

= — — As H
( At As + — Atl)\ + -(-As + -An\
dz

dx \ds dt dy \ds dt / dz\ds dt )

= —
df
Ax H
df
Ay ^
df
Az,
dx dy dz

which was to be proved.

(2) If f(x, y, z) is constant, Af = 0.
Suppose again that x, y, z are functions of two independent
variables s and t. Then
df df
Af = — ds H d<.
ds dt
 When the variables satisfy given equations this often enables us
to determine partial derivatives by algebraic elimination. The
procedure is to differentiate the equations, leaving constant the
same quantities as in the desired derivative, and then solve for
the appropriate ratio of differentials.

Example 1. The equation /(x, ?/, 2) = 0 determines z as an implicit

function of x and y. Find —

Sz
.

ox
By differentiating
2) = 0

with y constant and z a function of x, we get

0 ,

dx dz dx
whence

dz _ ^
dx ^
dz
410 PARTIAL DIFFERENTIATION [Chap. XVI

Example llu ^ v = xy, find ,

\dv/u,y
Differentiating the two equations with u and y constant, we get

0 = 2a; cte + 22 tfe,

dv — y dx.
Eliminating dx, we find
yzdz — —X dv.
Thus

193. Tangent Plane and Normal to a Surface. If the point

Vi describes a curve on the surface

F{x, y, z) = 0, (1)

the differentials of its coordinates satisfy the equation

dF
—
dx
dx
dF
By
dy
dF
dz
dz — 0. (2)

In particular, the differentials at the point Piixi, yi, Zi) satisfy the
equation
/dF\ /dF\ /dF\
dz = 0, (3)

{
\dx/i
, (—0
\dy/i
)
(—
\dz/i
being the values of the derivatives at Pi.

Along a given curve dx, dy, dz are proportional to the direction

cosines of the tangent [§181,(4)]. Equation (3) thus states that
the tangent at Pi to any curve on the surface is perpendicular
to the vector

N= (4)

All tangents at Pi, being perpendicular to this vector, lie in a plane

called the tangent plane to the surface. Since it passes through
Pi(xi, yi, Zi) and is perpendicular to the above vector, the equa-
tion of the tangent plane is (§174)

I
(x - ®i) + Q/" - + (z - Zi) = 0. (5)
1
§1941 SPHERICAL COORDINATES 411

Figure 244.

A line perpendicular to the tangent plane at its point of contact

is called normal to the surface. The above discussion shows that
the vector N extends along the normal.
Example, Find the tangent plane and normal to the ellipsoid

+ 2y^ + 322 = 3x + 12
at (2, -1,2).
The equation can be written

a;2 + 22/2 + 3z^ - - 12 = 0.

The normal vector at (x, y, z) is therefore

i(2x-3)+j(4t/)+k(62).

At (2, — 1, 2) this has the value

N= - i 4j + 12k.

A line through (2, —1, 2) and having the direction of this vector is normal
to the surface. The equation of the tangent plane is

(x - 2) - 4(2/ + 1) + 12(2 - 2) = 0.

194. Spherical Coordinates. The spherical coordinates of a point

P in space are its distance r = OP from the origin, the angle if>

from OZ to OP, and the angle 6 from the x-axis to the plane OPZ.
These coordinates are particularly convenient when distance from
the origin plays an important part.
The positive direction of 6 is that of right-handed rotation
about OZ. The positive direction of <i>
is that of rotation through
412 PARTIAL DIFFERENTIATION [Chap. XVI
90° from OZ to the terminal side of 9. The positive direction of r
is along the terminal side of 0.

Figure 245.

The equations connecting the rectangular and spherical coordi-

nates of the same point may be read from the diagram (Figure 245).
The most important of these are

X = r sin <t) cos df y — r sin 0 sin z — r cos 0. (1)

The locus of points r = constant is a sphere with center at the

origin;0 = constant is a cone
with axis OZ; 6 = constant is
a plane through OZ,
At a point P with spherical
coordinates r, 0, 6 are three
unit vectors Ur, u^, drawn in
the directions P moves when
two of the coordinates are kept
constant and the third r, 0,
or 9 increases. When r alone
varies these unit vectors do
not change. They are there-
fore functions of 0 and 9 only.
Figure 246.
When 0 alone changes ue
does not change, and r, 0 are plane polar coordinates. Thus
§ 194 ] SPHERICAL COORDINATES 413

When 0 remains constant and a change dB is made in Bj the sys-

tem of vectors rotates through the angle dB about the 2-axis. Thus
(§183)

—=k
dUr

dB
X Ur = u^ sin 0, (3)

—=
dB
k X = Ua cos 0. (4)

Finally, by differentiating = Urx and using the preceding

equations, we get

— = —Ur
du^
sin 0 — cos 0. (5)

These values of the derivatives give for the differentials of the

unit vectors
dur = U0 d0 + VLe sin <f> dB,

dvL^ = — Ur d0 + cos 0 dBy (6)

due = — (Ur sin 0 + u^ cos 0) dB.^

When spherical coordinates arc used, components of vectors

are taken along Ur, u^, u^. Thus, if

r = rur (7)

is the vector from the origin to the point P with coordinates r, 0, By

dx = Ur dr + u^r d0 + u^^ sin 0 dB (8)

is its differential, and ds = [

dr |
is the differential of arc. Since
Ur, U0, Me are perpendicular unit vectors, this last has the value

ds = V^dr^ 4“ d0^ + sin*^ 0 dB^, (9)

If P is moving its velocity is

dr dr d4> dB
V = Ur -7 + u^r —
dt
+ u^r sin 0
dt
( 10 )
dt dt

The components of velocity are therefore

dr d0 dB
r sin 0
Jt' ^It' dt

By differentiating v and using equations (6) an expression for

acceleration can be obtained.
414 PARTIAL DIFFERENTIATION [Chap. XVI

196. Maxima and Minima of Functions of Two or More Vari-

ables. A maximum value of a function is a value greater than
any given by neighboring values of the variables. A minimum
value is a value less than any given by neighboring values.
If w is a function of any number of independent variables, it is
clear that a maximum or minimum value of u remains a maximum
or minimum when only one variable changes. 7/ ihm a maximum
or minimum occurs at an interior point of a region in which all the
first derivatives with respect to the independent variables are continuous^
those derivatives must be zero.
At a maximum or minimum on the boundary these conditions
need not, however, be satisfied. On the section cut from a solid
by an inclined plane, for example, the maximum value of z occurs
at a boundary point and the partial derivatives of z with respect
to X and y are not in general zero at that point.
When all the partial derivatives are zero, the total differential
is zero. Thus, if is a function of independent variables x, y, Zy
and all the partial derivatives with respect to those variables are
zero.
du du du
du = — dx dy dz = 0
dx dy dz

for all values of dx, dy, dz.

To find maxima and minima of a function of several variables,
we equate its partial derivatives with respect to the independent
variables (or its total differential) to zero, and solve the resulting
equations. It is usually possible to decide from the problem
whether a value thus found is a maximum, a minimum, or neither.

Example 1. The vertices of a triangle are (xi, yi), (x2 y^, (xs, ys). ,

Find the point in the plane of the triangle such that the sum of the squares
of its distances from the vertices is least.
bet (x, y) be the required point. The sum of the squares of the dis-
tances is

(S = (x - xi)® + (y — yif + (x - XiY +iy - y^Y + (x - xs)*

+ (j/ - I/s)®.
The variables x and y are independent. Thus

== 2(x - xi) + 2(x — xj) + 2(x — xs) = 0,

— = 2(y - yi) + 2{y - j/j) + 2(j/ - yt) = 0,
whence
X = + X2 + Xs), y = \{yi + ys + Vs).
§196] HIGHER DERIVATIVES 415

The sum The partial derivatives must be

evidently has a least value.
zero at the minimum. The equations have only one solution. It must
give the minimum.
Example 2, Find the point in the plane

a? + 2y + 3z = 14
nearest the origin.
The distance from any point (a;, z) of the plane to the origin is

D=
If this is a minimum

OD =
x(lx + ydy
r
+ zdz
:=rr- =
_
0,
7
x^+ y^ + z"^

that is

X dx + y dy + z dz = 0-

From the equation of the plane we get

dx + 2dy + Sdz = 0. (2)

The differentials can have any values vsatisfying this equation. If these
are all to satisfy (1), that equation must be a mere multiple of (2). Cor-
responding coefficients must therefore he proportional. Hence

X z
^
” ^ 3*
1 2

Solving these and the equation of the plane simultaneously, we get x = 1,

t/ = 2, z = 3. There is a minimum and it must satisfy the above equa-
tions. Since there is only one solution, it gives the minimum.

196. Higher Derivatives. The first partial derivatives are func-

tions of the variables. By differentiating these partially we get
higher derivatives.
For example, the first derivatives of /(x, y) are

— = rr fxix, y),
\ — = ff SvKX, y).
^

dx By
By differentiating these we obtain the second derivatives

d^S d /d.
— 1 = fxx(.x, y),
dsr dx \dx

= fvxix, y),
dy dx dy \dx
416 PARTIAL DIFFERENTIATION [Chap. XVI

d^f
dxdy

|;(D
By differentiating again, we get third derivatives, etc.
There an order to partial differentiation, the operation indi-
is

cated by the symbol on the left being performed last. Two partial
derivatives of the nth order, which differ merely in the order of
operations, are, however, equal provided that all partial deriva-
tives of order equal to or lessthan n are continuous. This is a
consequence of the following theorem:
d^f d^f
At a point where and are both continuous they are equal,
dx dy dy dx
Let fxy(x, y) and fyx{xj y) be continuous at (a, 6). We are to
prove
fxy{a, h) = fy^{a, b). (1)

For that purpose we use the second difference

A = f(a + /i, 6 + fc) — /(a, 6 + A;) — /(a + A, 6) + /(a, b), (2)

This can be written

A = [/(a + /i, 6 + ft) - /(a, 6 + ft)] - [/(a + ft, &) - /(a, 6)]. (3)

The expression in the first bracket is obtained from that in the

second by replacing 6 by 6 ft. +
From the mean value theorem
(§34) we thus have

A = k\fy(a + h, bi) - fy(a, 6i)], (4)

bi being between b and 6 + By a second application

ft. of the
mean value theorem (4) becomes
A = hkfxyifli} fti), (5)

ai being between a and a + ft.

Equation (2) can also be written

A = [/(a + ft, 6 + ft) -/(a + ft, b)] - [/(a, 6 + ft) ~/(a, 6)]. (6)

By a procedure similar to that above, but considering first the

change from a to a +
ft, then that from 6 to 6 ft, we get +
A = khfyx(a 2 f 62), (7)

where a^ is between a and a + ft, and 62 is between b and 6 + ft.

§196] HIGHER DERIVATIVES 417

From (5) and (7) we have

fxy{o,l} fyx(fl2f ^^ 2 )- (8)

When h and k tend to zero, ai and a2 tend to a, and 6i and 62 tend

to 6 Since fxy and fyx are continuous at
. (a, fe), equation ( 8 ) thus

gives as limit
fxyiP'i ~ fyxiS^i b)}

which was to be proved.

Example 1. Assuming all derivatives of third or lower order continu-
ous, prove that
dy __
dy
dx^ dy dy dx^

Under the conditions stated we have

ay dy dy
^ ay
a / \ d / \
^ ^ ^
dx^ dy dx \dx dy) dx \dy dx) dx dy \dx) dy dx \dx) dy dx^

Example 2. If the substitution a: = r cos d, 2/

= r sin d changes /(x, y)
into F(r, d), express
d^F 1 aF 1 d^F
df2 y y2 ^Q2.

in terms of derivatives of /(x, y) with respect to x and y.

By assumption
F(r,d) =/(x,t/).
Hence
dF __dfdx . dfdy _ ^ ^

^
dr dx dr dy dr dx^ dy
’

dF _ dfdx df dy
— r sin d ^ + r cos d —•
'dd dx dy

By differentiating again and replacing the derivatives of x and 2 we obtain /,

d^F dH dH dy
= cos^ d
;^ + 2 cos d sin d + sin^ d dy^ ,
dr^ dx^ dx dy

d^F
^= r* sin*
«
d^f
^- 2r* cos » sin
0

dr
^ dH
cos^ d
,ay
dy*

— r cos d r sin d —
dx a2/

Substituting these values for the derivatives and combining terms, we

obtain
d^F idF 1 d^F d^f ay , .
418 PARTIAL DIFFERENTIATION [Chap. XVI

197. Taylor^s Series. The functions f(x, y) most frequently

used are for most values of a and b expressible as double series

m,n** 0

convergent when x — a and y — b are

] | \ \
sufficiently small.
Assuming such an expansion to exist and to be differentiable term
by term, we find
1 / d^+7 \
, (2 )
mini \dx^ dy / (a,b)

the subscript (a, b) signifying that after differentiation x, y are to

be replaced by a, b.

Unless the precise order of terms is specified (which is not usu-

ally the case), convergence of (1) can only mean absolute con-
vergence. By a discussion similar to that in §148 we show that if
the series converges at x = Xi, 2/ = 2/i it converges absolutely for
all values x, y such that

|x-a|<|xi-o|, U-b|<Ui-6|.
The region of convergence is therefore symmetric with respect to

(a, b) but may still be very complicated.

If we write
X — a + htj y b + kt ( 3)

and consider h and k constant, /(x, y) becomes a function F{t)

defined by the equation

F{t) = /(a + ht,b + kt), (4)

Assuming sufficient differentiability, this function can be expanded

by the Maclaurin formula

Fit) = F(0) + F'm + • •

. + n!
e+ Bn, (5)

the remainder Bn being determined as in §150. Now

dfdx ,
dfdy . df ^
df
F'it)
dx dt dy dt dx dy
§198] EXACT DIFFERENTIALS 419

Similarly,
/a ay
F"«) = (*- + *-)/, (6)
\ dx dy/

the notation signifying that the expression

h— + k —dY
d
]
( dx dy/

isto be expanded by the binomial theorem and applied to/(x, y).

Equation (5) can therefore be written

+ kt) = ^ ^
/(a htyh ^
ml \ dx ^ ^ ^ 7”
dy/

the subscript zero signifying that after differentiation t is to be

replaced by zero, that is, x and y are to be replaced by a and h.

In (7) t is arbitrary. Replacing ^ by 1, we have

/(a + /i,6 + fc) - V — (;*- + *-)

^ ^

^m!\
1 / d

dx dy/
d
/o + /2„. (8)

By a discussion similar to that in §150 the remainder can be

shown to have the form

1 / d d Y^
Rn = -{h— + k — ) /, (9)
(n + 1) ! \ dx dy/

the derivatives being evaluated at some point between (a, b) and

(a + hjb + k)y and on the joining line.
Equation (8) is called Taylor's formula with remainder. If the
remainder tends to zero as n increases, it gives Taylor's series.
Provided the partial derivatives used are continuous, there is an
analogous expansion for a function of any number of variables.
198. Exact Differentials. The expression

M dx + N dyj
where M and N are functions of x and y, may or may not be the
total differential of some function /(x, y). If there exists a function
/(x, y) such that
df = M dx N dy,
the expression M dx + N dym called an exact differential.
420 PARTIAL DIFFERENTIATION [Chap. XVI

Likewise, if P, Q, R are functions of x, y, z and there exists a

function /(x, y, z) such that

df — P dx + Q dy + R dZy
the expression P dx + Qdy + R dz is called an exact differential.
A similar definition may be given for any number of variables.
If X and y are independent variables, a necessary and sufficient
condition that
M dx + N dy
be an exact differential is
dM _ dN
( 1)
dy dx

To prove this suppose first that there is a function /(x, y) such

that
df = M dx + N dy.
Since

df — — dx + — dy
dx dy

for all values of dx and dy, we must then have

and hence
dM _ dN _ d^f

dy dy dx ^
dx dx dy

The two second derivatives being equal, this requires

dM _ dN
^
dy dx
which was to be proved.
Next suppose, conversely, that the condition (1) is satisfied. Let

w
-fM dx
be the function obtained by integrating with y constant. The
expression
dw
N-
dy
§198] EXACT DIFFERENTIALS 421

is a function of y only; for

d /dw\
_ a /dw\
_ dM ’
dx V dy) dy \ dx) dy
and hence
dw\ dN dM
—a In
/
) 0 .

dx \ dy/ dx dy
Let

fix, y) = w+ - dy-
J{^
Since the second term on the right is a function of y only, we have
dw dw
df = —
dx
dx -\

dy
dy
(
+ [N
\
dw\
dy/
)dy

M dx + N dyy
=
which shows that M dx + N dy exact. is

In a similar way we prove that

If Xj y, z are independent variableSy a necessary and sufficient
condition that
P dx + Qdy + Rdz
be an exact differential is that P, Q, R satisfy the equations

dQ
dx
= _
dP
_
dR
= _
dP
dx
dQ
-1 =
dR
(2)
dy dz dz dy

Example L Show that

(2x Sy) dx — (3x + 2y) dy
is an exact differential.
In this case
M = 2x — 3y, N = -(3x + 2y)

dy dx

The two partial derivatives being equal, the expression is exact.

Example 2, In thermodynamics it is shown that

dU ^ TdS-pdVy
V being the internal energy, T the absolute temperature, S the entropy,
p the pressure, and v the volume of a homogeneous substance. Any two
422 PARTIAL DIFFERENTIATION (Chap. XVI
of these quantities can be assigned independently, and the others are then
determined. Show that

The result to be proved expresses that

T dS + v dp
is an exact differential. That such is the case is shown by replacing T dS
by its value dU + pdv. We thus get

TdS + V dp = dU + pdv + vdp = d{U + pv).

199. Directional Derivative. Gradient. Let fix, y, z) be a
function with first derivatives continuous at points P(x, y, z) of a
certain region. When P moves along a given direction, fix, y, z)
may be considered a function of the distance s moved. The deriva-
tive of f with respect to s,

ds dx ds dy ds dz ds

is called the derivative of fix, y, z) in the assigned direction. It is a

function of the position Pix, y, z) and the direction of motion.
In particular, if P moves parallel to the x-axis in the positive
direction, dy — dz — Q, ds = dx, and

ds dx

The partial derivative of f with respect to a codrdinate is thus

merely the derivative of / in the positive direction parallel to the
corresponding axis.
df df df
At a point where — — — are continuous, the vector with com-
dx dy dz
, ,

ponents equal to these derivatives is called the gradient of / and is

represented by the notation

df df df
grad/ = i^ + j/ + k^. (2)
dx dy dz

To see the significance of this let

dx ,dy dz
§199] DIRECTIONAL DERIVATIVE. GRADIENT. 423

be the unit vector in a given direction. Equation (1) can then be

written
df
-- = u-grad/. (4)
as

This a function in any direction is equal

states that the derivative of
to the component of its gradient in that direction.
If components of a vector are taken along different directions,
the component will be greatest along the direction of the vector
itself. The gradient of f(Xj y, z) therefore has the magnitude and
direction of the greatest directional derivative of /(x, y, z) at (x, y^ z).
Suppose, for example, that r(x, y, z) is the temperature at the
point P(x, y, z) of a region. As we start out from P in different
directions there will be some direction in which the temperature
increases most rapidly. The gradient of T points in that direction
and is equal in magnitude to the most rapid increase per unit
distance moved.
Through the point P(x, ?/, z) in the region there passes a surface
/(x, y, z) = constant. In a motion along this surface the com-
ponent of grad / is

and so the direction of motion is perpendicular to the gradient.

Therefore, the gradient of /(x, y, z) is normal to the surface

f{Xj y, z) = constant

which passes through (x, i/, z).

If r = ix + 32/ + kz is the vector from the origin to P(x, y, z),

dr = idx + j dy + kdz (5)

and

dr*grad/ = —
df
dx H
df
dz/ d
df
dz — df, (6)
dx dy dz
If then
V = iX + jy + kZ ( 7)

is the gradient of a scalar function /(x, y, z),

dr- V = dr-grad/ = df (8 )

is an exact differential. Conversely, if

dr-V = Xdx + Ydy + Zdz

424 PARTIAL DIFFERENTIATION [Chap. XVI
is the differential of a function f{x, y, z), then

dr*V = df = dr-grad/ (9)

for all values of dr, and consequently

V= grad/. (10)

Therefore, a necessary and sufficient condition that

Xdx + Ydy + Zdz

be an exact differentud is that the vector

LY + jF + kZ
be the gradient of some scalar function.
Equation (6) can be used to determine an expression for gradient
when dr is known in a particular system of coordinates. Thus in
spherical coordinates

dr = Ur dr + u^r d<t> + u^r sin <f> dB.

To make the scalar product of this and grad / equal to

df df df
df = — dr H d<t> -\ dB
dr d4> dB
we must have
A A
grad/ = Ur—
df
dr
^df
+—
,

—H
r d4>
^
Me
^

r sin
—— df
dB
( 11 )

Example 1, Find the derivative of + 2/^ + in the direction of the

vector i + 2j + 2k at the point (2, 3, —4).
In the assigned direction

ds ds ds 3

By equation (1) the derivative of / = +2 ^ in that direction is

at the point (2, 3, —4).

Example 2. Find a vector normal to the surface xyz = 6 at the point
(1, 2, 3).
Hie gradient

+ + = iya + ja» -1- fcry = + 3j + 2k

(iei/a)
j
^ (*»*) (*y«) (H

is a vector normal to the surface at the given point.

§200 ] LINE INTEGRAL. WORK 425

200. Line Integral. Work. Suppose that a point P{x, y, z)

moves along a given curve under the action of a force

F = iZ + jy + kZ (1)

with components that are in general functions of x, y, z. To

estimate the work done by this force we replace the curve by a
polygonal line with vertices on the curve and represent P by the
vector
t = ix + }y + Vlz (2)

from the origin to the point P. When the consecutive vertices

Vi ^)f Qix + AXy y + Ay, z + Az) are sufficiently close

together the force has almost the value F(x, y, z) at all points of the
chord PQ. The work in describing this chord is therefore ap-
proximately
—
F-PQ = FAr
and the total work is approximately

The work in passing along the curve from A to B is thus defined

as the limit
426 PARTIAL DIFFERENTIATION [Chap. XVI

Substituting for F and dr from (1) and (2) the expression for
work becomes
r

Xdx+ Ydy + Zdz. (4)

An integral of the form (3) or (4) is called a line integral. To

evaluate such an integral along a curve with continuously varying
tangent we use the equations of the curve to express F and dr,
or Xy y, z, X, Yy Zy iTi toHus of a single variable t and its differential.
The integral then takes the form

f(t) dty

ti and <2 being the values of ^ at A and B. When the curve is

formed by sections with an abrupt change of direction at each
junction, the integral is the sum of integrals along the separate
sections.

Example 1. Find the work done by the force

F - i|^ jx

when the point of application moves in the xy-plane along the parabola
= Ax from (0, 0) to (1,2).
In this case r = Lr + jy and
jY’dt = J'ydx-‘xdy.

Along the given curve x = ly^ and

ydx — xdy — \y^ dy.

As the point moves along the curve, y varies from 0 to 2. Thus the work
done is

=y' \y^dy =

Example 2. Find the value of the integral

j' xdy + ydz + zdx

along the curve
X ^ ty y — 2ty z ^
from the point where < = —3 to the point where t = 3.
§2011 CASE OF LINE INTEGRAL INDEPENDENT OF PATH 427

Substituting for x, y, z the values from the equations of the curve, we get
xdy + ydz + zdx = (21 + 5t^)dt.
The integral from 1 = —3 to i = 3 is therefore

+ 5«2) dl = [<* + = 90.

201. Case of Line Integral Independent of the Path. In general

the line integral

depends on the curve from A to B along which it is taken. If two

different paths join the same end points, the integral along one
need not be the same as that along the other.
//, however F is the gradient of a one-valued functiony the line
integral depends only on the end points of the path.
Suppose, in fact,
F = grad0 (1)

whore 0(a:, ?/, z) is one-valued in a certain region. For points in

that region [§199, (0)]

Fdr = dr-grad = d<l>

and consequently

f%^dx^ rd4> = 4>{B)-4>{A\ ( 2)

Ja Ja
the notation being used to represent the value of <l>(Xy y, z)
at the point Ay and the value at B. The integral thus de-
pends only on the end points Ay B and not on the path connecting
them.
When the integral is independent of the path, the integral
around a closed path is zero. For, if A, are any two points on
a closed path the integral around the circuit can then be written
nA nB nB
f F.dr+ / F-dr= / F-dr - F-dr = 0. (3)
A Jb Ja J/
Conversely, if F(a;, 2/, z) is continuous and the integral
428 PARTIAL DIFFERENTIATION [Chap. XVI
around every closed path in a three-dimerisioml region is zero, there
is a function y, z) such that

F = grad 0.
To show this let

f ^-dr, (4)

where Pq is a fixed point and P(x, y, z) a variable point in the region.

Since the integral around every closed path is zero, the integral
from Po to P does not depend on the path. For any integral from
Pq to P added to a particular integral from P to Pq must give zero.
Thus is a one-valued function of x, y, z. Take a particular path
from Pq to P along which the unit tangent vector

t - *
ds
is continuous. Along this path
f*p P dt
f F-dt = /* F — (

Jpq Jpo ds

is a function of s with derivative (§51)

_ = F —dr
d<b
(5)
ds ds

at P. The curve could be chosen to have any direction at P.

Hence the derivative in any direction is given by (5). This shows
that y, z) has a derivative in each direction and that the
derivative in any particular direction is a continuous function of
X, y, z. Thus <^(x, y, z) has a gradient at P and consequently
d0 dt
— •grad</>. ( 6)
ds ds

Combining this with (5) we have

dr
(F - grad <l>)
•— = 0.
ds

dt
Since this holds for every value of —
ds

F— grad 0 = 0,
which was to be shown.
 PROBLEMS 429

When the work done by a force is zero around every closed path,
the force is called conservative. We have thus shown:
The necessary and sufficient condition for a force F{x, y, z) to be
conservative is that it be the gradient of a scalar function,
A conservative force is usually expressed as a negative gradient

F = — grad0.
The is then called a potential.
function 0 For example, the force
on a weight of w pounds is of magnitude w and directed
of gravity
downward. Taking the z-axis upward,

F = —kw = —grad (wz).

The force F is thus derived from a potential wz.

PROBLEMS
Find — and —
dZ dz
9. in each of the following problems:
ax ay
1. 2 = — 2xy + 4^^. 2. 2 = sin (xy),

3. « — In - •
4. 2 =
X +y
X X — y

In the following problems perform the indicated differentiations, x, y, z

being the independent variables:

6. — - Zxy +1/^4- 2x2). 6. — sin (2x + 32/4- z)*

ax dy

7. —
dz
In (zyV). 8 . —
dX
Vx2 + 2/^ + 2^.
If /(x, y, z) = + 3xy* + 3yz\ find/i(x, y, z) - /„(x, y, z).
X*
y^
10. If /(x, y) = X In — find the value of /*(4, 2) + /y(3, 1).
,
X
11. If 2 = In (x^ + X2/ + 2/^), find the value of

dz
+ 2/
—
dz
dx dy

12. If /(x, y) =s (x + a)(y + b)y a and h being constants, show that

fx(x, y)fy{x, y) = /(x, y).

18. If X =»u+v, 2/
= u— V, show that

\dX/ u \dx/ y
14. If X « r cos 2/
« r sin find

/£x\ /£x\
\dr)e\dr)y
430 PARTIAL DIFFERENTIATION [Chap. XVI
15. li X » r cos $t y ^ r sin d, find

\de/y\dd/r
(-) •

16. Let a, b be the sides of a right triangle, c the hypotenuse, and p the
perpendicular from the vertex of the right angle to the hypotenuse. Find

(-)•
\da/b
18.
17. If K isthe area of a triangle, a side and two adjacent angles of which
are c, Ay determine

The sides of a triangle are a, 6, c, and the opposite angles are Ay By C,

Find

19. Find the lowest point on the surface

2 = + 2/^ — icy — 4x + 5y + 10.

20. Find the highest point on the surface

2 =» 6y — 2aj — — y^.

21. Show that the partial derivatives oi z ^ xy are zero at the origin but
that the origin is not a high or a low point on the surface.

Find the total differential of each of the following fimctions:

22. 2 — xy + 3y^. 23. r »= Vx^ + y^ + 2 ^.

24. u = 25. w = In (xyz).

26. Calculate dz and Az if z « y^ — x^ — xy, x = 2, y = 3, cte =» Ax = 0.01,

and dy ^ Ay ^ —0.02.
27. Calculate du and Am if w « xyz, x*»l, y=s2, z = 3, dx«Ax = 0.1,
dy ^ Ay ^ —0.1, dz = Az = 0.2.
28. A box is 12 inches long, 8 inches wide, and 6 inches deep. Considering
the edges as independent variables, find the increment and the differential
of the volume when each edge is increased | inch.
29. If V is the volume of a cylinder of radius r and altitude h and dr *« f,

dh » 2ty show that

dV « Aiy

where A is the total surface area of the cylinder, including the two ends.
80. The hypotenuse and one side of a right triangle are found by measure-
ment to be c » 5 feet, 6 « 4 feet, and the other side a is calculated. Find the
maximum error in a due to errors of 0.01 foot in the measurements of c and h.
 PROBLEMS 431

81. The acute angle A in the triangle of the preceding problem is calcu-
lated from the data given. Find the maximum error in A.
32. The distance from the origin to the point P(6, 3, 2) is calculated by the
formula
r - V** + 2/® +
Find the maximum error in r due to errors of 1 % in y, z,
83. The volume of a cone is computed by the formula

V= \lCT^h.

Find the percentage error in the volume if the measurement of the radius is
\% too large and that of the altitude 1% too small.
34. Two sides and the included angle of a triangle are found to be c = 50
feet, 6 = 20 feet, A = 60°, and the third side a is calculated. Find the maxi-
mum error in a due to errors of 0.1 foot in the measurements of h and c, and
0.01 radian in the measurement of A.

In each of the following problems find the derivative of /(a;, y) with respect
toi by formulas (2) or (3), §190, and check by expressing /(x, y) in terms of t

and then differentiating:

36. f(x, y) = x^y, x = y = <*.

86. f(x, y) = Vx* + X = 3«, y = 4i.
87. f(x, y) = 3?-\-xy,x=s + y = s — t, t.

38. /(x, y) = In (x - 2/), X = |(e‘ + e"‘). V = - e“‘)-

In each of the following problems find —

dx
,
assuming that x and y are inde-

pendent variables and that 2 is a function of x and y:

39, u == z sin (x + y), 40. m =* — z\

41. u = ye*“*. 42. w = In (y + 2)*
43. Uu = f(x, y), y =» X, find —
ax
44. If « f(Xy y), y -2x - 8, find .

46. If w « /(«, y, 2), z « find .

\dX/y
du
46. Find -j- if w = F(a;, y, 2), 2 == /(x, y), y =* <i>{x). Check by using the

functions u = ocyz, z — xy^y ^ x^.

47. The substitution y = a; -f- changes 1; f(x, y) into F{Xy v). Determine
df
— in terms of derivatives of F(x, v) with respect to x and v. Check by using
dx
the function f(x, y) * xy.
48. If the substitution a; = r cos y ** r sin ^ changes /(«, y) into F(r, 6),

show that
432 PARTIAL DIFFERENTIATION [Chap. XVI
49. If the substitution x ^ r cosh d^y ^ r sinh d changes /(a:, y) into F(r, d).

show that
/dFy 1 ^

50. If *» /(x* + y^), show that

y —
dU
dx
« X—
dU
dy

51. If M = /(ox + hy)i show that

— = a—
, du du
h
dx dy

52. If w =s
/ ,
show that

du du
X— + j/— “^ 0.
dx dy

53. If w = /(sin X + sin y), show that

secx-
du
secy —
du
dx
4" y^
*

54. If w = :
,
show that
X +y
du du
x-—
dx
+ y—’=u.
dy

66 .
— -), show that

dz
3
. •
« 0.
dx dy

56. If w = - 4“ - 4- “ »
show that
z X y
du du du
dx dy dz

57. If z is determined by solving the equation

Vx X/
show that
dz dz
® :r + T"
“
dx dy
dy
58. If /(x, y) * 0 express 7- in terms of derivatives of /(x, y) with respect
,
dx
to X and y,
59. Let a, jS, 7 be the direction angles of a line, and consider 7 as a function

of a and 8.
j
Find the value of ~
dy
da
when a » 45® and jS «7 « 60®.
 PROBLEMS 433

60. If f(x, y,z) — Of z — X -j- y, find .

ax
61. If F(Xy yy z) = 0, show that
62.
/^\ f^\ /£f\ ^
\dy)Xdz)Xdx)y"
If a; = 1/ = uVy find

63. If a; + = uVy xy ^ xi — v, find

66.

64. If M == a; + + 2/ z, xyzy find

If M = a; 4" 2/ -f 2, V = -f 2/^ + 2% find

68.
66 . If a; = r cos 0, y = r sin 0, show that69.
72.
xdy — y dx — d0, dx^ +73.
dy^ == dr* + d0*.
74.
67. If X = r cosh 0 , y = r sinh 0, show that
xdy — y dx = r* d0, dx* — d?/* = dr* — r* d0*.

Bind the equations of the tangent plane and the normal to each of the fol*

lowing surfaces at the point indicated.

X* + 2/2 + 2* = 9; (1, 2, 2). X* - y* = 8; (3, -1, 1).

70. 2 * = x* + r; (3, 4, 5). 71. X* + 2/^ = 2z; (1, 3, 5).
2 =: xy; (2, -3, -6), 2* x* - 2
/^ = 1; (2, -2, 3).

Find the components of acceleration in spherical coordinates.

76. A point moves on the sphere r = a. Find the components of its accel-
eration in spherical coordinates.
76. If the equations of a curve are given in spherical codrdinates, show that

t = Ur
dr
—+ d<#>
3- H-
.

sm <#)
~
d0
ds ds ds

is the unit vector tangent to the curve.

77. In any motion of the point P (Figure 246) show that
dUr du^ du0

where
(4 (i9

"•3r+‘5-
The unit vectors thus rotate with angular velocity «.
434 PARTIAL DIFFERENTIATION [Chap. XVI

78. Find the angular acceleration *7- in the preceding problem,

at,

79. Find the points on the surface ** -f 1 at least distance from the
origin.
80. A tent having the form of a cylinder surmounted by a cone is to contain
a given volume. Find its shape if the canvas required is a minimum.
81. When an electric current of strength I flows through a wire of resistance
R the heat produced is proportional to PR, Two terminals are connected by
three wires in parallel of resistances Ri, R2, Rzt respectively. A given current
82.
flowing between the terminals will divide in such a way that the total heat
produced is a minimum. Show that the currents /i, /2, H in the three wires
will satisfy the equations
IiRi = /2R2 * IzRz-

Two adjacent sides of a room are plane mirrors. A ray of light starting
from a lamp A in the ceiling strikes one of the mirrors at Py is reflected to
a point Q on the second mirror, and is then reflected to a point B on the
floor. Determine the positions of P and Q so that the path APQB shall be
as short as possible. Show that the angles of incidence and reflection are
equal at each mirror.
88 Find the point on the curve
.

a? ^2zy

at greatest distance from the x^-plane.

dh
In each of the following problems calculate and and show that
dx dy dy dx*
they are equal:

84. « » X* - + xy^ - y®. 86. 2 = x In (x -f y).

X — 1/
86. 2 . 87. 2 * X*'.
X +y
88 . Show that
d / du dw\ d / dw duy d r du du\
X
dx V dz * dy) dy V dx Hz^)

X, y, z being the independent variables.

» B/(i, y) 0,findj.

90. If 2 /(x, y), - X*, find

^
91. If the transformation x . e” cos v, y — e” sin v changes /(x, y) into
F(u v), express

du*

in terms of derivatives of /(x, y) with respect to x and y.

 PROBLEMS 435

92. If the substitution u -f v ^ xy changes F(w, v) into/(x, y\

express

98.
dx^ dy^

in terms of derivatives of F(w, v) with respect to u and v.

If the substitution y«=u — changes /(x, y) into F(u,
t; v),

express
dV
dx^ dy^

in terms of derivatives of F(w, v) with respect to u and r.

96. If
94. the substitution x *= r cosh d, y =* r sinh 0 changes /(x, y) into
F(r, 0)f express
d^F 1 dF 1 d^F
dr^ ^ r dr r* dd^

in terms of derivatives of /(x, y) with respect to x and y.

Determine the coefficients Amn in the series

2:A„,nX”‘y^

O-DO-D
96. Determine the coefficients Amn in the series

Determine which of the following are exact differentials:

97. (2x + y) dx “ (x — 3y) dy.

98. (x^ - y2) dx + (y**^ - 2xy) dy.
106.
99. X dx — (y 4- 2 ) dy + y dz.
100. (x^ + y^) dx — 2x dy.
101. (2 + x) dx ~ (y + 2 ) dy 4- (x - y) d2 .

102. (sin y — y sin x -f x) dx 4- (cos x + x cos y 4- y) dy.

103. (2x6*' 4“ y) dx 4- 4- x — 2y) dy.

104. Find the derivative of xy 4- yz in the direction of the vector

2i ~ 3j 4- 6k
at the point (1, 2, 2).
Find the derivative of x +y 4- 2 in the direction of the vector

i - 2j - 2k.

106. Find the derivative of /(x, y, 2 ) in the direction — i.

107. Find the gradient of (x + y)e*.

108. If /(x, y, 2 ) — xy 4“ y2 4* 2x, find the maximum value of %

ds
at the

point (2, —1, 3).

436 PARTIAL DIFFERENTIATION [Chap. XVI
109. If B is the angle in cylindrical coordinates find the gradient of B.

110. Find the work done by the force

F= 4* y^) - 2]xy

when the point of application moves along the straight line y = 2a;, z = 3a:

from the origin to the point (1, 2, 3).

111. Find the work done by the force

F = 6j2 — 4ia:t/

when the point of application moves along the curve

y * 2x\ z = 2a;*
from X 0 to a; = 2.
=
112. Find the work done by the force F = jxi/* when the point of applica-
tion moves in the positive direction around the square formed by the lines
x = =tl, = d=lin the
Find the value of each of the following integrals along the path given:

113. jydx--zdy — X dz; X y = -P, z = 3^ 0 ^ t ^ 3.

114.
f xdz zdx X — At, = 3t, z = 5P, -1 ^ ^
y 1.
J x^ +y^ +z^
,
;

116. ysinxdx — cos xdy -\- dz; y * cos Xy z « sin 2x, 0 ^ X ^ -

J' 2
116. If F = iy, show that
rB

along an arc AR in the xy-plane is equal to the area bounded by the a;-axis,

the arc, and the ordinates at A and B.

117. If F = ia:^ show that

fp.*
isindependent of the path from A to B.
118. If w is constant, show that F = kw is a conservative force.
119. If r is the vector and r the distance from the origin to P(x, y, z), show
that F = rr is a conservative force.
120. For paths in the xjz-plane that do not go around the origin show that

iy -

is a conservative force.
 CHAPTER XVII
MULTIPLE INTEGRATION

202. Repeated Integration. We found in §46 that, if A{x) is

the area of section cut from a solid by the plane x = constant,

is the volume of the solid between the planes x = x = b. When

the area of section is known as a function of x we can therefore
determine the volume by integration.
If, however, the section has curved boundaries (as it usually
does), the determination of its area requires integration. In the
section x — constant of Figure 248, for example, y and z are plane

Figure 248.

coordinates and the section extends from 2 = 0 to the top surface

z = /(x, y). The area of section is therefore

/(*, v) dy, (2)

and ^(x) being the values of y at the boundaries of the

section.
437
438 MULTIPLE INTEGRATION [Chap. XVII

To determine the volume we thus integrate (2) wilh x constant^

substitute the resulting function of a: in (1), and integrate again.
The two integrations are indicated by the notation

y= / ( / S{x,y)dy\dx (3)

and the result is called a repeated integral. The first integration

is with respect to y with x constant between the limits (l>i{x) and
02 (a:). The second integration is with respect to x between a
and b.

It should be noted that there is a definite order to these opera-

tions. In (3) the order is indicated by parentheses, the inner
integration being performed first. The parentheses are, however,
usually omitted, and instead of (3) we write

y= / /
Kx,y)dydx, (4)

the symbols being arranged in the order they would have if paren-
theses were used. The variable in the first integration is thus
indicated by the differential on the left, and its limits are attached
to the integral sign on the right. It should be noted, however,
that this usage is not universal. Some writers arrange the symbols
differently.
Since /(x, y) is the value of z we could write

'/(*, y)

Six, y) = dz

and so determine the volume

r” r<h(.x) /•/(».»)
V= /
dzdy dx ( 5)
*/o a^lCx) a/O

by three integrations. In the present instance this is an unneces-

sary complication, but we shall encounter cases, moment of inertia,
for example, in which three integrations are needed.
all

Usually the limits in the earlier integrations are functions of the

later variables, only the last limits being constant.

Example 1. Find the value of the integral

cte dy.
§ 202 ] REPEATED INTEGRATION 439

Integrating with respect to x between the limits 0 and y, and then with
respect to y, we obtain

+ y") = X* [^ =^y"]}y

= XVdy = M = 27.

Example 2, Find the volume bounded by the coordinate planes and the
plane
x +y+z^ 1.

The section in the plane x = constant is a triangle (Figure 250) extend-

ing from y = 0 to 2/=l — x. The area of this triangle is

""Jo ""Jo (
1 -^- 2/) ^2/-

The volume is formed by sections of this kind extending from x = 0 to

X = 1. Thus
^

V=X £~\l-x-y)dydx
=X ^ ~^y ~ dx

-fja-xrdx
= [-K1 - ®)’]J

= i.
Example 3, Determine the region in the xy-plane over which the integral

extends.
££ fix, y) dy dx
440 MULTIPLE INTEGRATION [Chap. XVII
In the first integration y varies from the curve y — to the line y — x.
In the second integration x varies from 0 to 1. The region is that bounded
hyy = X and y ^ x^ (Figure 251).

Figure 261.

Example 4» The integral

/•2

is extended over a certain region in the ir^-plane. Determine an equivalent

integral with the order of integration reversed.
The integral extends over the region (Figure 252) bounded by the line
2/
= — 1, the curve y = and the line x ^ 2, With order of integration

Figure 252.

changed, x varies from the curve a: = 2yt to the line » = 2, and then y
varies from —1 to 1. Thus

XiXi
is the expression required. That the two integrals are actually equivalent
will follow from our later discussions (§204).
§ 203 ] MULTIPLE INTEGRATION 441

203. Multiple Integration. We now define a second method of

integrating, or summing, over a region, and in the following section
we will show that this is equivalent to repeated integration.
First consider a region A in the xy-plane and a function f(x, y)
defined at points of this region. Divide the plane into rectangles
by lines parallel to the coordinate axes, the distances between

consecutive parallel lines being equal or unequal. In each rectangle

which lies entirely within A choose a point ({, rj) and form the

product
fiiy v) AAy

where AA is the area of the rectangle. In case of a border rectangle,

containing points both inside and outside A, a product of this kind
may be included or omitted. Let

S/tt, v) AA (1)

be the sum of the products thus formed. When the number of di-
visions is increased in such a way that the edges of all the rectangles
tend to zero, if this sum approaches a limit independent of the
choice of dividing lines and of points ($, ry) in the rectangles, and
independent of which boundary rectangles are included and which
omitted, the function /(x, y) is said to be integrable in the region A
and the limit is called the double integral of /(x, y) over that region.
It is indicated by the notation

(2)
442 MULTIPLE INTEGRATION [Chap. XVII

In particular, if /(x, y) = 1, this integral defines the area of the

region.
Similarly, being given a function and a region V in
/(x, y, z)
space, we divide the region into rectangular elements of volume
by planes parallel to the coordinate planes, choose a point ($, ry, f)
in each volume element Av, and form the sum

2/(5, 17 , f) At;, (3)

which may or may not include boundary elements. When the

number of divisions increases in such a way that all the edges of
the volume elements tend to zero, if this sum approaches a limit
independent of the choice of dividing planes and points (5, Vy
and independent of which boundary elements are included and
which omitted, the function /(x, t/, z) is said to be integrable in
the region V and the limit is called the triple integral of /(x, y, z)
over that region. It is indicated by the notation

In particular, if /(x, y, z) = 1, this integral defines the volume

of the region.
The sum (1) or (3) is not essentially different from the corre-
sponding sum which defines the definite integral of a function of
one variable. The principal difference is the more complicated
nature of the boundary. In the integrals we shall consider we
assume that the sum of the boundary elements of area or volume
approaches zero as the process of subdivision proceeds.
204. Evaluation of a Multiple Integral by Repeated Integration.
It remains to determine how a multiple integral is to be evaluated.
For brevity we use the notation

to represent the repeated integral over a region A, the first integra-

tion being with respect to y, the second with respect to x.
If /(x, y) is continwus in a finite region A,

,y)dA = ,y)dxdy — , y) dy dx, (1)

provided the boundary of A is such that the integrals are defined.

§204] EVALUATION OF A MULTIPLE INTEGRAL 443

The multiple integral and the repeated integrals over a given

region are thus equal provided that they are all defined.

In proving this it is convenient to use two auxiliary functions.

First let g{x, y) be the function equal to /(x, y) at all points of the

region and equal to zero at all other points. If R is some rec-

A
tangle which contains A and has sides parallel to OX and OF,
we then have

dy dx f{x, y) dy dx. ( 2)
y)
,

If-
By lines parallel to OX and OY divide R into rectangles so small
that, if (xi, yi) and (x2 , ^2 ) are two points of A in the same rec-

tangle,
1/(X2, j/2) -/(*!> yi)l < *•

That each positive value of e is an extension

this is possible for
of theorem 6, §14, to continuous functions of two variables. Next
let G{x, y) be a function which is constant in each rectangle
of

R and equal to the value of g{x, y) at some point (5, rf) in that
rectangle. We then have

G(x, y)dydx = (4)

the summation being of the form (§203) used in defining the double
integral over A. By subtracting (4) from (2) we obtain

- G{x, y)] dy dx. (5)

444 MULTIPLE INTEGRATION [Chap. XVII

The right side of this equation is equal to a sum of integrals over

the rectangles of R. The difference

I
g(x, y) - G{x, y) |

is less than € in each rectangle interior to A and in the boundary

rectangles cannot exceed 2ikf , where M is the maximum value of
f{Xy y) \n A, In rectangles outside A this difference is zero. The
right side of (5) is therefore less than

ea + 2Mhy
where a is R and h is the sum of the areas
the area of the region
of the rectangles that containboundary points of A. As the sub-
divisions diminish in size € may be taken smaller and smaller,
and h tends to zero. Equation (5) thus gives as limit

lim

which is equivalent to

dA.

In a similar way we obtain

fj Six, y) dx dy dA,

thus completing the proof.

206. Double Double inte-
Integrals, Rectangular Coordinates.
grals are used not only in planesummations but also in the calcu-
lation of space quantities which can be represented as double
sums. Often it is possible to calculate the same quantity by one,
two, or three integrations, the reduction in the number of integra-
tions being accomplished by choosing sections such that part
of the integrations can be performed mentally.
Thus, if the sec-
tion of a solid a rectangle or a circle, it is unnecessary to deter-
is

mine its area by integration. In general it is advisable to use as

few integrations as possible.
§206] DOUBLE INTEGRALS, RECTANGULAR COORDINATES 446

Example 1, Find the centroid of the plane area bounded by the para-
bolas 2/^ = 4 — x, = 4 — 2a:.

By symmetry the centroid is seen to be on the a:-axis. Its abscissa is

dA
£ =
/dA
the integrals being taken over the
region between the parabolas. Ex-
pressing these as repeated integrals
it is evident from the diagram (Fig-
ure 255) that the first integration
should be with respect to x. For,
ify varied first, the limits on the left
of A would not be the same functions
of X as those on the right. Hence Figure 255.
*2 /•4-v*

f f xdxdy
£ = "" 5*
2 ^ 4— v2
/
Example S, Find the moment of inertia of the area bounded by the
coordinate axes and the line x y = 2 about the axis perpendicular to
+
its plane at the origin.
The distance from the point (a:, y) to the axis is 's/x^ + y^. Thus the
moment of inertia is

J (x^ + 2/2) dA

the integral being taken over the triangle (Figure 256) bounded by the
axes and the line x y — 2, Integrating first with respect to y we have
+
(x* + y^)dA =J^
(as* + j/*) dydx = %.

Figubb 266.
446 MULTIPLE INTEGRATION [Chap. XVII

Example 3,Find the volume bounded by the a^y-plane and the

paraboloid z 4 —
^ ^ 4xK
One-fourth of the required volume is shown in Figure 257, This is the
limit of the sum of prisms of height z with bases Ax Ay covering a quadrant
of the ellipse + 4x2 = 4. The volume is therefore
/•l /•2Vi-a;2
— — ^
/ zdA == (4 2/2 4a;2) dy dx 4ir.

206. Polar Coordixiates. When polar coordinates are used the

plane is cut into parts by the circles r = constant and the lines
0 = constant. The part bounded by two circles of radii r, r + Ar
and two lines ^ + A^ is approximately a rectangle of sides Ar,
r AS and area
AA = r A0 Ar,
the error in this approximation being of higher order than A0 Ar.
If /(r, 0) is continuous, the sum

2/(r,(?) AA
taken for all elements AA in a region thus tends to the limit

IP /(r, 6)r dr dB
§206) POLAR COORDINATES 447

as Ar and Aff tend to zero, the integral being taken over the
region.
The limits in the first integration are the values of r at the ends
A, B oi the variable strip (Figure 258). The limits in the second

Figure 258.

integration are the values of 6 determining the outside strips.

If it ismore convenient the first integration may be wth respect
to 6, The integral then has the form

If fiX} cW dr.

The first limits are now the values of d at the ends of the strip

Figure 259.

determined by the circles of radii r, r + Ar (Figure 269). The

second limits are the extreme values of r.
448 MULTIPLE INTEGRATION [Chap. XV-

Sxamjile 1. Find the Value of the integral

2a ry/2ax^
/ i ix^ + y^)dydx
by changing to polar coordinates.
The integral is taken over the semicircle y = V2ckc - (Figure 260).

In polar coordinates its equation is r = 2a cos S. Also = r*.

Hence

+ y^) dy dx = t»drde = lira\

Example 2, Find the centroid of the area outside the circle r a

and inside the circle r = 2a cos 6,
By symmetry the centroid is on the x-axis. Its x-co6rdinate is

X =
/ r cos 0 dA
y

the integration being taken over the area between the circles (Figure 261).

Fiqure 261 .

By solving simultaneously we find that the circles intersect at 0 » =b com

§206] POLAR COORDINATES 449

Since the area is S3nnmetrical with respect to the a;-axis, we have

/•2<Jco8d
2 1 I cos 6 dr dd
+ 3„ V 3
_ ,

Stt

/•J /•2(icoafl 47r + 6\/3

H Ja

Example 3. Find the moment of inertia of the area of the cardioid

r = a(l + cos 0) about the axis perpendicular to its plane at the origin.

The distance from any point P(r, 0) (Figure 262) to the axis of rotation
is OP = r. Hence the moment of inertia is

I =fr^dA =

Example 4, A homogeneous solid has the form generated by rotating

the cardioid r = a(l+ cos 0) about the x-axis. Find its moment of in-
ertia about the x-axis.

When rotated about the x-axis the element of area

dA = r dr d0

generates a ring of volume

dV = 2iry dA
and moment of inertia

fyy^ dV = 2Trpy^ dA = 27rpr* sin® 0 dAj

where p is the density. The total moment of inertia is therefore

^
^ =
J — ^ 2irpr® sin® 0 dA —
X 0 dr d0 fl^irpa®.
4B0 MULTIPLE INTEGRATION [Chap. XVII

207. Triple Integrals, Rectangular Coordinates. In a triple

integral

y, z) dx dy dz

the integrations can be arranged in six different orders. The ease

of performing these operations depends very much on the order
in which they are taken. In a particular instance all orders should
therefore be considered and the one chosen which involves the
simplest integrations. Since the first two integrations sweep over
the cross section determined by a constant value of the third
variable, it is mainly a question of choosing these so that the
resulting section is as simple as possible. By proper choice of
sections most problems that lead to triple integrals can be solved
by a single integration.

Example* Find the centroid of the solid bounded by the paraboloid

+ 2z^ = 42/ and the plane y = 2*

One-fourth of the solid is shown in Figure 263. By symmetry x and 2

are zero. The 2/-co6rdinate of the centroid is

y are the values y « \{x + 2z^) at P and 2/ « 2 at Q.

^
The limits for

At S, 2/
— 2 and x = y/^y — 2^ = V
S — 2z^. The limits for x are
therefore a; « 0 at and x = \/8 — 2^^ at S. The limits for z are z * 0
at A and z » 2 at B.
A simpler way to solve this problem is to note that the section y = con-
stant is an ellipse
§208] CYLINDRICAL COORDINATES 451

with semiaxes a = \/4y, 6 — and area

A = Tcah = 27r\/ 2 y.

The coordinate of the centroid is therefore

4
y — ___ = ___ 3*

X X y^y

208. Cylindrical Coordinates. If r, 0, 2 are cylindrical coordi-

nates, the surfaces r = constant, B = constant, z = constant di-
vide space into elementary regions (PQ, Figure 264) which for

small values of the increments are approximately rectangular

parallelepipeds with edges Ar, r A^, A2 and volume Ai; = r Ar M A2 ,

the error in this approximation being an infinitesimal of higher

order than Ar AS Az, When the increments all tend to zero, a
sum of the form
S/(r, By z) Av

taken for all elements At; in a region thus tends to the limit

z)r dJB dr dz,

the integral being taken over the same region.

The integrations can be arranged in six different orders. The
one should be chosen that involves the simplest integrations.
452 MULTIPLE INTEGRATION [Chap. XVII

Example 1. A cone is rotated about a diameter of its base. Find the

moment of inertia.

Take the origin at the center of the base, let the 2;-axis be the axis of the
cone,and take the moment of inertia about the x-axis. The equation of
the conical surface is

2 = - (a - r),
a
h being the altitude and a the radius of base. The distance from a point
{x, Pj z) to the x-axis is

\/^ + z^ = y/r^ sin^ d + 2^.

The moment of inertia about the x-axis is therefore

*Lv r
X X Jo ^

h
The limits in the first integration are the ends « - (a — r) of a = 0, « =
a
prism standing on the base r dd dr. The limits in the second integration
are the boundaries r = 0, r *= a of a wedge-shaped slice extending from
the axis to the perimeter of the base. The third integration sums these
slices around the axis from ^ = 0 to ^ = 27r. By evaluating the integrals
we find

where M = \vpa%
/ = ^is
(3a* + 2ft*)

the total mass of the cone.

= 1 M(3a* + 2ft*),
§208] CYLINDRICAL COORDINATES 463

Example 2, A hole of radius

h is bored through a sphere of
radius a, the axis of the hole
being a diameter of the sphere.
Find the moment of inertia of
the ring thus formed about the
axis of the hole.
In Figure 266 is shown one-
eighth of the volume. The
equation of the spherical sur-
face is

The moment of inertia is

/• 2 ir
/•\/a2-22
^ ~ pr^-r dr dz dS
^Jo Jo Jh

= (8a* + 40^6* - 126*) Vo* - b\

^
p being the density.
Example 3, Find the volume bounded by the coordinate planes, the
surface z — x^ +
and the plane x y — 1. +
In cylindrical coordinates the equation of the surface is 2 = r*, ^ind
454 MULTIPLE INTEGRATION [Chap. XVII

that of the plane is r

§V2 sec (-i) The volume is therefore

‘'-XX X '**<»-»•
209. Spherical Coordinates. In spherical coordinates the locus
r = constant a sphere with center at the origin,
is = constant <t>

is a cone with axis OZ, and 6 = constant is a plane through the

2;-axis making the angle B with OX.

In Figure 268, PQRS is the region bounded by the spheres r,

r + Ar, cones </>, ^ + A</), and planes ^ + AB. When Ar, A<^, AB

are all very small this is approximately a rectangular parallele-

piped with edges PQ = Ar, PR = rA<t>, PS = r sin <t>AB. The
volume of this region is approximately

PQ-PR-PS = r^ sin Ar A^ AB,

the error in this approximation being an infinitesimal of higher
order than Ar A<l> AB. When all the increments of the coordinates
tend to zero, a sum of the form

2/(r, 0, 0) At),

taken for all volume elements in a given region, thus tends to the
limit
itdrd^dB,
J'J'J'
the integral being taken over the same region.
52101 MEAN VALUES 455

Example. The inner and outer radii of a hemispherical shell of constant

density are r = r = h. Find its center of gravity.
Take the center of the sphere as origin, and let the xy-plane be the face
of the hemisphere. By symmetry the center of gravity is on the 2-axis.
Also
2 = r cos <l>.

Hence

rT/: r cos <!> sin 0 dr d<l> dS

Zb*-a*
Sb*- a*'
rr/: am <t> dr d(f> dB

210. Mean Values. In §133 we defined the mean value of

f(x) with respect to x between x — a and a: = 6 as the quotient

( 1)

Similarly,

( 2)

is called the mean value of /(x, y) with respect to area in the region
A of the xy-plane. If we sprinkle points over the region A, mak-
ing the density (number per unit area) constant, take the average
of the values of /(x, y) at those points, and determine the limit as
the number of points becomes infinite, the result will be the value
determined by (2).
Similarly,

dV
— (3)

is the mean value of /(x, j/, z) with respect to volume in the region
V of space.
 466 MULTIPLE INTEGRATION [Chap. XVII

The mean values detennined by (2) and (3) are the mean with
respect to area and mean with respect to volume. Other types of
means are sometimes used, but, unless the contrary is explicitly

stated, the term mean value is understood to signify mean with

respect to area in a two-dimensional distribution and mean with
respect to volume in a three-dimensional one.

Example. If p is the distance from a fixed point Q outside a sphere

to a variable point P inside, find the average value of - for all positions
P
of P within the sphere.

Take the origin at the center of the sphere, the 2-axis through Q, let a
be the radius of the sphere and q = OQ. Using spherical coordinates,

p = + r2 — 2gr cos (p.

The average value of - is

P

fUv
J p
rrr^ +
Jq Jq Jq y/
r2 sin

r2
dfi>

—
dr dB

2gr cos ^
/dr £ fo fo
^Bm4>d4>drdB

The average value of the function - is thus the value - which this func-
P q
tion takes at the center of the sphere.
§211 ] AREA OF A SURFACE 4£7

211. Area of a Surface. On the surface

2 = /(a:, y)
dz dz
let She a, region at all points of which — and — are continuous
dx dy
functions of x and y.
Let P be a point of aS, and Q its projection on the x^z-plane.
When Q describes an elementary area AA in the x!/-plane, P de-
scribes an area AS on the surface. When AA is sufficiently small,

Figure 270.

AS lies approximately in the tangent plane at P. The areas AA

and AS thus satisfy approximately the equation

AA = AS cos 7,

where y is the angle between the a:y-plane and the tangent plane
at Py that is, the angle between the 2:-axis and the normal at P,

The total area of the region on the surface is approximately

COS 7

As the elementary areas tend to zero, these approximations become

more and more accurate. Hence in the limit we have

Ja cos 7
as the area of surface, the integral being taken over the projection
on the xy-plane. To evaluate this integral we choose appropriate
458 MULTIPLE INTEGRATION [Chap. XVII

coordinates and determine cos 7 from the equation of the surface.

In a similar way we could obtain a formula for the area of surface
in terms of its projection on any other plane. The plane of projec-
tion may to a certain extent be chosen arbitrarily. It should,
however, be such that cos 7 does not become zero (at least does
not change sign) at points of S,

Example. Find the surface of the sphere ^ a? inside the

cylinder + y* = ox.

The projection on the xy-plane is the circle = ox, or in polar

coordinates, r = o cos d (Figure 271). From the equation of the sphere
we have
z
_ \/a^ — x^ — y^
_
--
cos 7 a
a a
The area required is therefore

ar dr dJd
S = (27r — 4)a2.

The factor 4 is introduced because the cylinder intersects the sphere below
as well as above the xy-plane, and the integral represents only half the
area above the plane.

PROBLEMS
Find the values of the following integrals:
/•s /•z /•
dydx
^ Ji Jo ^JJ (a:+y)*
a 0080
sm e dr dfi.
Jo Jo Jo

5. Determine the volume bounded by the codrdinate planes and the plane

X + 2y 4* 32? 6.
 PROBLEMS 459

6. A solid is bounded above by the surface

2 =* X — 2/^

and below by the xy-plane. By integration determine the area A{z) of the
section cut from the solid by the plane x = constant. By a second integra-
tion determine the volume of the solid between the planes x =« 0 and x =* 4.

In each of the following problems determine the region over which the
integral extends:
^2oo>z
X X%,y)dxdy. 7 Vi f(p^.v)dy^-

cos

*•7 Jo f(r,6)drd0.
7 7 7
4

In each of the following problems determine an equivalent integral with the

order of integration reversed:
r\ /•V2-.2y*

“•7 7
/•4

18. Jf f{x,y)dydx.
X £''f(.^'y)d^dy-
15 . Find the centroid of the area bounded by the parabolas = x -f 3,
= 5 ~ X.
16 . Find the centroid of the area bounded by the parabola = 2x -f 1
and the line y « 1 — x.
17 An area is bounded by the parabola y^ = 4ax and the line x = o.
.

Find the moment of inertia of this area about the axis perpendicular to its
plane at the origin.
18 . A thin plate of mass M
has the form of a right triangle of sides a and 5.
Find its moment of inertia about the axis perpendicular to its plane at the
vertex of the right angle.
19 . Find the moment of inertia of a cube of side a and mass about an M
edge.
20 . Find the volume of the pyramid bounded by the coordinate planes, and
the plane x 4- 2y — Sz « 6.
21 . Find the volume under the paraboloid z =* x^ -h y^ and over the square
bounded by the lines x = ±l, y^ztlin the xy-plane.
22 Find the volume under the plane 2x + 2y -h 2 = 8 and over the tri-
.

angle bounded by the lines x « 0, y « 2, x = y in the xy-plane.

23 A solid is bounded by the planes y = 0, z =» 0, x =» z, x -f y « 1.
.

Find its volume.

24 . A solid is bounded by the planes x » 0 y « 0 s 0 x +y 2, and , , ,

the surface z (y — 2)*. Find its volume.

26 Find the volume above the xy-plane bounded by the surface z « x^ — y^
.

and the planes x » 1, x « 3.

26 Find the volume bounded by the planes y = 0, y * x and the cylinder
.

2* - a* — ox.
460 MULTIPLE INTEGRATION [Chap. XVII
Find the values of the following integrals by changing to polar coordinates:

»>•
rr
/•'/•*£=.
M.
Jo Joo -v/^i
+ j,*

86.
88. Find the area inside a loop of the curve r = a cos 26 and outside the

circle r = ^ .

84. Find the area cut from the lemniscate = 2a* cos 26 by the lino

Find the centroid of the area within the curve r = 2 cos 0 4- 3.

86. Find the centroid of the area bounded by the circle r « a cos 6y the
x-axis, and the line y = x,

87.Find the centroid of the area bounded by the »-axis and the circles
r *a cos r = a sin 6,
88. Find the moment of inertia of a circular plate of radius a and mass M
about an axis perpendicular to its plane at a point of its circumference.
89. Find the moment of inertia of the area within one loop of the curve
r* « a* sin 6 about the axis perpendicular to its plane at the origin.
40. A circular sector has the radius a and central angle 2a. Find the
moment of inertia of its area about the bisector of the angle.
41. Find the moment of inertia of the area within one loop of the curve
r* as 2a* cos 26 about the axis perpendicular to its plane at the origin.
42. A circular sector of radius a and angle a is rotated about one of its
bounding radii. Find the volume generated.
48. The area outside the circle r « a and inside the circle r ^ 2a cos $ is
rotated about the initial line. Find the volume generated.
44. An anchor ring is generated by rotating a circle of radius a about an
axis at distance h (greater than a) from the center. If its mass is M, find its
moment of inertia about the axis.
46. A pyramid of constant density p is bounded by the coordinate planes
and the plane x + 2y Sz 6. Find its moment of inertia about the ^axis.
46. Find the volume bounded by the cylinder

and the coordinate planes in which x, y, z are positive.

 PROBLEMS 4tFl

47. Find the volume bounded by the cylinders ^ z^ ^ c? — ox,

and the plane x = 0.
48. Find the volume bounded by x = 0, 2 « 0, 2 *= 1 — xy, =» 1 — x.
49. Find the centroid of an octant of the ellipsoid
60.
1 .

Find the moment of inertia of the ellipsoid in the preceding problem

about the z-axis.
51. A solid of constant density is bounded by the xy-plane and the parabo-
loid x^ “f 2y^ = 4—2. Find its moment of inertia about the 2^axis.
66. A hole of radius a is bored through a sphere of radius 2a. If the axis
62.
of the hole passes through the center of the sphere, find the volume cut out.
63. Find the moment of inertia of a sphere of radius a and mass about M
a diameter.
64. A cone has altitude radius of base a, and mass M. Find its moment
of inertia about an axis through the vertex parallel to the base.
Find the moment of inertia of a sphere of radius a and mass about M
a tangent line.
66. A segment is cut from a sphere of radius a and density p by a plane at
distance |a from the center. Find its moment of inertia about a diameter
parallel to this plane.
67. A spherical shell of inner radius a, outer radius 6, and density p is cut

by a plane at distance c <a from the center. Find the moment of inertia of the
segment cut off about the diameter perpendicular to this plane. Consider
the solid as the difference of two spherical segments.
68. Find the volume bounded above by the sphere + 2^ = 5 and below
by the paraboloid 42 = .

69. Find the volume bounded by the xy-plane, the cylinder r = 2a cos
and the cone r = 2 .

60. Find the volumes of the two parts into which the sphere “ o*
is cut by the cone r + 2 = a.

61. Find the volumes of the parts into which the sphere + 2^ = 4a2 is

cut by the paraboloid = az,

62. Find the volume bounded by the xy-plane, the paraboloid az = x^ +

and the cylinder x^ -f- 1/^ = a^.
63. Find the volume of the cylinder x^ -f = 2ox intercepted between the
xy-plane and the paraboloid x^ + = 2 o2 .

64. Find the volume bounded by the cylinder r = a cos B and the planes

^“0, 6 = ^,z=0, z = x+y.

66.Find the volume bounded by the surface 2 = c“ (**+*'*^ and the xy-plane.
66 Find the volume cut from the paraboloid 2 = x^ + y^ by the plane
I » 2x.
Find the moment of inertia of a cylinder of radius a and mass
67. about M
density is proportional to the distance from its axis.
^ts axis if its

68. Find the center of gravity of a cone of altitude h if its density is pro-
portional to the distance from its base.
462 MULTIPLE INTEGRATION [Chap. XVII

69 Find the center of gravity of a cone of radius a and altitude ^ if its

.

density is proportional to the distance from its axis.

70 Find the centroid of the volume cut from a sphere of radius a by two
.

planes through the center including an angle of 60°. Use spherical coordinates,

and take the planes as ^

.
« ±7*
6
71 . Find the centroid of the region within a sphere of radius a and a cone
of vertical angle 2a, if the vertex of the cone is the center of the sphere.

72 . vertex of a cone of vertical angle 60° is on the surface of a sphere

The
of radius a, and its axis is a diameter of the sphere. Find the moment of inertia
of the volume common to the cone and sphere about the axis of the cone.
73 . In spherical coordinates the equation r = a sin 26 represents a surface.

Find the volume between d «* 0 and 0 = “ bounded by this surface.

74 . Find the volume enclosed by the surface r = o sin 4>,

76 From a spherical shell of inner radius a and outer radius b a segment is
cut by a plane at distance c (less than a) from the center. Find the volume
of the segment. Use spherical coordinates, and integrate first with respect to 4>.
76 . Find the center of gravity of a hemisphere of radius a if its density is
proportional to the distance from its center.
77 . Find the center of gravity of a right circular cone of altitude h and
radius of base a if its density is proportional to the distance from the vertex.
78. Find the mass of a cube of side 2a if its density is k times the square of
the distance from its center.

79 . Find the center of gravity of a cube of side a if its density is propor-

tional to the square of the distance from one corner of the base. Take the
edges through this corner as axes.
80 . The density of a sphere of radius a and mass M
is proportional to the

distance from a diametral plane. Find its moment of inertia about the diam-
eter perpendicular to that plane.
81 . Find the volume above the xp-plane bounded by the sphere r a, the
cone 0 » 6, and the plane ^ » 0.
82 . Let p be the distance from a variable point P in a circle of radius a to
a fixed point Q at distance q from the center of the circle. Find the mean
value of for all positions of P within the circle.
88 . Find the mean distance from one comer of a square of side a to points
inside the square.
84. Find the mean square of the distance from the center of a cube of side a
to points inside the cube.
86 . Find the mean distance from a point on the surface of a sphere of radius
a to points inside the sphere.
86. On a circle of radius a a large number of chords are measured. Assuming
the ends of the chords to be uniformly distributed around the circumference,
find the average length of the chords.
87 . Find the area of the triangle cut from the plane

X +2y +3z 6
by the codrdinate planes.
 PROBLEMS 463

88. Find the area cut from the plane by the cylinder
^ +y^ - a*.

89 Find the area above the

. a;i/-plane cut from the cone + y* ** ** by
the cylinder 4* 2/^ 2aa;.

90 . Find the area of the surface of the paraboloid +2 =

^ 2ax cut off
by the plane a; = a.

91 Find the area of the surface

.
— 2a;y above the xy-plane bounded by
= and
1 a; = 2.
y
92 Find the area on the conical surface a:^ + y^ = 2^ between the a;y-plane
.

and the plane a; + 22 = 3.

93 Find the area cut from the plane by the cone in the preceding problem.
.

94 Find the area above the a;y-plane cut from the cone a;^ + y^ == 2^ by
.

the sphere x® + y^ + 2 ^ = 2ax.

96 Find the mean distance from points on the surface of a sphere to a fixed
.

point on the surface.

96 If p is the distance from a hxed point Q inside a sphere of radius a to a
.

variable point P on the surface, find the mean value of “ for all positions of
P
P on the surface.
 ANSWERS TO PROBLEMS

Chapter I. Pages 21-26

4. A 6. V
4t* 6 \/ TT
8. I
a; I
< 2. 9. -2 < a; < 4.

10. X = 0 ^ p < 1. 11. = 1 — Xf X ^

If 1 — y
, 2/

1*. /(O) = 0,/(l) = 0,/(2) = 2,/(-2) = 6.

18. /(O) = 1,/(1) = 2,/(2) » 4,/(-2) - i.
14. 0.
16. /(*) + 2 =. 4 - x,f(x + 2) = -®.
_l-2x 1
/1\ x + 1
•'W l+a:7(x) x-2
17. /(**) - X* + (/(x)]* = x» + 2x + 1, 1.

28. 1. 29. 0. 80. 1.

31. 0. 32. 38. SU-
84. 2^. 36. 1. 86. 0.
87. -1. 88. §. 89. 1.
40. 0. 41. 0. 42. 1.

48./
1 — a;
.
44. i
46. limSn Slf lim An » . 0. 47. 2o.

48. 1 49. Yes, no. 61. 2.

62. 5A 68. 6 + 2 V6 = 9.899 sec.

64. 1. 66. 7. 66. 3.
67. -6. 68. f. 69. 0.
60. 1. 61. 1. 62. 0.

68. J. W- /f- 66. 1.

2y/a
66. 0. 67. 0. 68. No limit.
69. 3. 70. 1. 71. 0.
72. 1. 78. 1. 74. 2.
76. f 76. 0. 77. 1.
78. No limit. 79. i 80. -i
81. 0. 82. 0. 88. V2.
84. 0 86. 2. 88. 0.

464
 ANSWERS TO PROBLEMS 465

90. Chapter
91. I (Continued) 92.
87. 0. 88 0
. . 89. 1.

1. 0, -1. -1.
93. 0. 94. None. 96. tit ± —2 •

96. rnr ±— 97. X g 0.

4

Chapter 11. Pages 64-72

1. 10 + 16<, 10. 2. 34.4 ft./sec., 36 ft./sec.

3. 13 ft./scc., 12 ft./sec. 4. 16.

5. (50 -f 160 ft./scc., (50 -f CO

/sec.

6. II
C
l-se
c3
> CQ
1 II
b
2* 7. 2.

8. -3. 9. 45®. 10. 135®.

11. 3. 12. 4. 13. 3.
14. 3x2. 16. (a) 0, (b) 1. 16. 1.

17. b a. 18. 7. 19. -7.

_ Ax
26. (3x2 27. (2x + 3) Ax. 28.
2 Vi'
Ax
29. 32. (o) ds = 60. (b) As = 610.
(X - D*
33. (a) 10,000. (6) 10, 000. 34. dV = 4irr2«.

36. 6x2^, Ax ~ 2tf dV = QxH. 36. ft.3

37. 2ir in.2

38. (a) A/(x) = 0.09902,
df{x)= 0.10000,
A/(x) - df{x) = -0.00098 Ax.
(6) A/(x) = 0.0009999,
df(x) = 0.0010000,
A/(x) - <i/(x) = -0.00001 Ax.
39. (a) 0.00909 = 0.0909 Ax.
(6) 0.000000999 == 0.000999 Ax.

40. 4x3 - 12x2 -f 4x + 4. 41. - 2x2 -|- x.

- 3x +
42. 12(x2 2x + 1). 43. x2 -f 2x.
2\/x
2x — -
47. 6(2x - 1)*

48. 4(2* + !)(*** +X + 1)*. 49.

V X* - 2x +3
50. 6(2x - l)(x - D*. 61. 3x*.
3x
62. (8x - 2)(x + l)(2x - 3). W. --
7
2Vx + 1

66. 15x*Vx* + o*. 66 .

V2 - X* (2x + 3)*'
466 ANSWERS TO PROBLEMS
Chapter n (Continued)
vv. • .

(1 - »*)*• **^^41* - l' (a* - 3?)i

- 2)i 62.
(i* + 2* + 2)i’ + 1

4(3® - D* 66. 12 cos 4a;.

(2*« + 1)*

-2sin(x+0- 67.11
67. — cos X. 68. X cos X.

4 sin (2x + 3) cos (2a; + 3). 72. 45^

* = 1, -i -2. 76. Yes.
1 76. No.

0.06, 6%. 78. 0.1%. 79. —

43,200
-

81 . h%- 82. -(2a: + 3r*.

—4 sin 2x. 84. 10(2** - 1)(2 - 31?)K

1.
ix + 2y)^
_i_
92. -csc®^.
2xi
4t*

(«* - 1)*'

Increasing if a; > 2. 97. Increasing if > 1.

|
x ]

Decreasing if a; < 2. Decreasing if < 1.

|
x |

Concave upward for all x. Concave upward if x > 0.

Concave downward if x < 0.
Increasing < —2 or > 0.
if a; a; 99. Increasing for all x.
Decreasing if — 2 < < 0. a; Concave upward if x < 0.
Concave upward if > —1. a; Concave downward if x > 0.
Concave downward if < —1. a;

Increasing if a; > 3.
Decreasing if a; < 3.
Concave upward if « < 0 or x > 2.
Concave downward if 0 < x < 2.

101. Increasing x > —1. if 104. A - 100 in.*

Decreasing x < —1. if

Concave upward for all x.

106. P - 40 in. 106. Side of base 10 in., depth 5 in.

107. Side of base 4 in., depth 1 in. 108. h^r.

109. F - 8 in.* 110. h - I V3.
111. Arc is twice radius. 112. i(ai 4- 02 +
as -f 04).
118. 8 in. 114. Width is twice depth.
116. V5 in. 116. r - 2h.
 ANSWERS TO PROBLEMS 467

Chapter II (Continued)

117. Ratio of length to width = a/6.

118. One-third the way from the lesser to the greater source.

119. Four hours later. 120. (a^ -f 6^)^.

lai. (o* + 124. Speed through water 7.5 mi. /hr

126. Speed through water is zero. 126. All circle.
127. (1, 0). 128. (1, 0).
129. y= 1. 130. 2/
== 1.

136.
131. V * 48ir. 132. -48.

133. (a) 1 < ^ < 3. (6) t < 2.

134. (a) toward, (6) toward, (c) increasing.

(a) away, (6) toward, (c) decrcfising.

136. (a) away, (6) toward, (c) decreasing.

137. (a) away, (6) away, (c) increasing.
138. (a) toward, (6) away, (c) decreasing.
139. (a) away, (6) away. (c) increasing.
141. 32 ft./sec.^
Ry
142. —s-
J.2
,
where R is the radius of the earth.

146. 2j ft. /sec. 146. ft./sec., 31 ft./sec.

147. 148. 8 ft./scc.

O
5
149. 5 ft. 160. in. /sec.

161. in.^/sec. 162. 4,366,000 ft.='/day.

o
163. f- 164. 1. 166. 0.
166. 4. 167. 3. 168. 2.
1
161. 162. g-a^. 163. 6*
164. 0. 166. h 166. “i*
167. 0. 168. 0. 169. -1.
170. 0.

Chapter III. Pages 96-101

1. ** - 2** + 2x + C. 2, 2*^ + 2x’ - x* + &E + C.

- i + C. |(x + 3)Vx + C.
3.
^d y
4.

6. J^x)* - (2x)i + C. 6. i(2i + + C.

1)»

7. ^ - 2x + + C.
- 8. ix» -fx* + 2x + C.

»• + §** + §X» + c. 10.

U -f 1
+ c.
- + 3)J+C.
(^+ 2)+^-
3
12.i(2i

18. |V3x + 2 + C. 14. + C.

468 ANSWERS TO PROBLEMS
Chapter m (Continued)
16. ^(3x* 4- 2)i + C.
“ +
3(a»
18. it(4z* + 6)i + C.
H ^ 1 V 3

19. 2Va:* + + 1+ C. a:
”--3('+7)+‘'-
21. J sin 29 + C. 22. — f cos {2x + 3) + C.

2S. 1 + C.
ffln* < 24. 1(1 - cosy)* + C.

26. m= + ^gt\ 26. /k) + 27. 138 ft.

28. X 4. 29. 18 ft./sec. 30. 2 ft./sec.

81. 38.72. 82. 1/ = - 3a; + 6. 33. 2/
= 2 — cos a;.

34. 4. 36. 2/ = fa;® “ f a;^ 4- f a; + -V'*

86. y = ** + 2x - 1. 41. 6. 42. i
48. 12f. 44. 2. 46. 0.
46. ^0*. 47. 2. 48. 18.
49. i 60.^. 61. i
62. 16. 68. f 64. 2.
66. \\ 66. 67. i.
68. i 69. 4i 60. -f JT.
61. 62. 2va^. 68.

66. |x(2a» - 3a*A + h\ \-ir{2o? + Za?h - h^). 66. ^o6*.

67. Hn. 68. 69.

70. 71. W’TO*. 72.
78. 9ir. 74. ^(6* - o®). 76. -^r®.
76. 2oV3. 77. 6. 78. a^h.

79. fo’*, where a is the radius of the circles. 80. |7rO*ft.

81. fa® tan a, where a is the radius of the cylinder.

82. 83. 10 tons. 84. 7.5 tons.

86. 86. 1388f tons. 87. 13.3 tons.
*(6 - o)*
Jig 89. 12Sww. 90. oeoorw.
2a
91. -jls w-AA*. 92. (2.1) 10» fMb. 9& (1.06) 10* ft-lb.
94. 140,000 ftrlb. 96. 463 ft-lb.

Chapter IV. Pages 137-146

1. X - i(xi + »s),l/- i(yi + y8). 8. (2,0).

8. X — 4, 2^
B —3. 4. (0, i). ( -4,7).

6. (9, 14), (-3, -22).

nx* + rjxi ny*+ r*yi
6« * “ ,y =
+ rt
ri
,
n -fra
7. X - xi + r(x* - Xi), y “ W + r(yj — yO-
 ANSWERS TO PROBLEMS 469

Chapter IV (Continued)
8. a: = 15, ?/
= —3. 9. 7/
~ 4a; 4- 11 = 0. 10.
y = 2.
11. a; = 3. 12. 2/
- 3x +7 = 0. X
13. « 2.
18. 3a; -f 4y = 5. 19. 5x 4- 4p = 22. 21. X ~ 3y ~ 7 = 0.

22. a; + 3?/ - 7 = 0. 23. 120°.

26. tan A = tan C = 2, R = 90°. 26. 1, 2, 3.

29. X — St/ — 6 = 0.
31. X - 3t/ 4- 17 * 0, 3x 4- ^ - 9 = 0.
32. X 4- 22/ = 0, 2x -- y = 0.

33. Tangent: y — 12x 4- 16 = 0. Tangen

34. Tangent: 3x 4- 2y — 12 =* 0.
Normal: x 4- 122/ — 98 = 0. Normal: 2x — 32/ 4" 5 = 0.
Norma!
36. Tangent: x 4- 2/ 4- 1 =0. Tanger x 4- y — 2 = 0.
36. Tangent:
Normal: x — y — 3 0. Normal: x — ?/ = 0.
Norma
37. y = ztx. 39. 2a.

41. 7/
— 2x 4“ 1 = 0, 2/
— 6x 4" 9 = 0. 42. 2.0650.

43. -0.6328. 44. 2.0393. 46. 0°.

46. tan“^ f, 90°. 47. tan~* 2. 60. 6 + Zy/5.
62. (—1,0). 63. Cix - 4t/ + 21 = 0. 64. = 4a: — 4.

66. Z3?
76. + 3y* + 8x - 16 = 0. 66. 3? - 3r - \2y + 36 = 0.
67. a? + y^ - &x -9,y = 0. 68. + y* — 2x — 41/ = 6.
69. Center (—3, 1), radius 4. 60. Center (2, —-I), radius 3.
63. + y- - 6a; + 2«/ - 40 = 0. 66. X* 1/2+ 2x 23/ -
1 = 0, - +
x2 3,2+ lOx 103/ _25 = 0. - +
66. X- +r = 1, + y‘^ = 81. 67. x* + 3/’* - 23/ - = 0.
1
68 X-
. + y- - 2* + 4j/ - 20 = 0. 69. x^ + 3/* + 2x - 43/ - 20 = 0.

72. X — 2t/ = 0.
74. Axis 2/
= 0, vertex (0, 0), focus (2, 0).
Axis t/ = 0, vertex (0, 0), focus 0).
76. Axis 2/
= 2, vertex ( — 1, 2), focus (0, 2).
77. Axis X = 0, vertex (0, 1), focus (0, f ).
78. Axis X = 3, vertex (3, —2), focus (3, — 1).
79. Axis X = 6, vertex (6, 3), focus (6, 0).
80. Focus (0, 3), directrix x = 4. 81. Focus ( — 1, ^)y directrix t/ =
83. 9 ft. 84. 10 ft.

86. 35.56 ft. 86 . Ihh.

3Th%
87. 88. 2p.
15

90. Center (0, 0); axes x = 0, 2/

= 0; foci (db\/3, 0); e =
91. Center (0, 1); axes x = 0, 2/
= 1; foci (db^\/5, 1); e = 3\/5.
92. Center (—2, axes x
1); = —2, 2/
= ^oci (—2 4= \/3, 1); e = i\/3.
93. Center (0, 0); axes x = 0, t/ *= 0; foci (0, ±\/7)> ^ = 4 V^-
94. Center (1, axes x = 1, = —ij ^oci
t/ (|, — (f, e =»
96. Center (2, f); axes x = 2, 2/ = f foci (2, J
; [9 ± 3\/5l); « *• i'x/S-
470 ANSWERS TO PROBLEMS
Chapter IV (Continued)
96. -f _ gy ^ 16 « 0. 97. - Vly - 36.
106.
98. 0 ?+ -f 4x — 8y = 32. 99. 2x + 3y — 6 *» 0.
Center (0, 0); axes = 0, = 0; foci (±-\/l3>a; 2/ ® asymptotes
2p « ±3x.
106. Center (1, 2); axes a; « 1, y « 2; foci (1 ± \/5, 2); e ^ y/ 5; as3anptotes
y - 2 - db2(a; - 1).
107. CJenter (2,—1); axes a: « 2, 2/ «* — 1; foci (2, —1 db ^ “ ^'\/Sj
asymptotes 2^ 1 ** =t\/2(a? — 2). +
108. Center (1, ~3); axes a; =« 1, 2/ * —3, foci (1 d= 2\/5, —3); c * \/2,
asymptotes 2/ 4- 3 « ±(a; ~ 1).
109. 32/* - a:^ - 122/ +9 * 0. 110. 4(x - 2)^ - ( 2/ 4- 1)^ * ±4.
111. 8x2 _ ^ 103. 4. 22/ « 0.
113. A hyperbola with foci at the rifle and the target.
116. Hyperbola through A with one focus at B. Eccentricity =» 2.

118. (-1, 5), (-7, 7), (2, -5). 119. x^ + 41/2 - 4 = 0.

120. y* 4- 3x2 + 16 = 0. 121. X + 2/
= 0, 2x + 32/ = 0.
122. x2 4“ V* = 11. 123. (-i -2).
124. 3x2 4. 2/2 « 2, or x2 4- 3i/2 = 2. 126. X2/ = 1.

126. xy = 2. 127. -•

128. y^ « — 2/)
4- 14. 129. /^
2
= ox\/2 — 2
®^*

Chapter V. Pages 166-168

1. X « 1, 1/ = ~2. 2. X - 2, 2/
- 1 3. X — —2, y — 1.

4. X « 0, j/
« —1. 6. -18. 6. 0.
7. 0. 8. 42. 9. 63.

10. 0. 11. » = * “ -i-

12. X « 6, y « —3, z »» 10. 13. H u 1 ti

14. X 3, 1/ s« 4, =“ 2. 2; 16. X — — 1, y — 2,
« — 1, u — 3.

16. x=*3, 2/=* — 1, z=s -l,u = 0. 19. xiy'.z — 3: —11: —4.

20. x:2/:2 * 1:2: —3. 21. x:y:z — 3:0: —2.
22. x:y\z « 2: —1:1. 23. x:2/:2 » 3:-l:2.

2/2 X 1
26. 9 2 1 - 0.
4 -3 1
27. 9(x2 4. 2,2) _ 44^ _ 41y -1-32-0.
28. The points are not on a circle. 29. Converse not true.
80. X « ^ a 31. Converse not true.
32. kiik^ikz 6:— 3:— 1.

Chapter VI. Pages 171-174

X X
16. 12 cos 3x — 12 sin 4x. 16. 3 sin - cos -•
2 2
17. — 12co8 3xsin3x. 18. sin2x.
 ANSWERS TO PROBLEMS 471

Chapter VI (Continued)
^ a;
19. 2 sin* a;. OA
20. -a; cos-*
2 2
21. 3 cos 3a; — 3 sin* x cos x.
23. 5 cos* 5x. 24. a;sin*x.
26. 6 sec* 3a;. 26. 12 CSC* (1 - 4a;).

sec* X
27. —7==. 28. tan*x.
2v tan X
29. —3 cot* 3a;. 38.
30. tan 2x sec* 2x.
31. 0. 32. 4 sec* X tan x.
33. 3(sec X + tan a;)* sec 2 . 34. 2 sec^ 2x.

36. tan*^sec^^* 36. 2sec|tan»|.

3 3 3 3
37. sec X tan* x. — cot®x.
39 . 41. 45°.
X
42. 135®. 43. tan“^ tan”^ 3, tan”^ f
44. 5. 46. |.

46. 0.9673

48. 80 ir mi. /min. 49. 400 ft. /sec., 200 ft./sec.

60. sin ^
i V3* 61. 7.794 in. 62. 120®.

63. -. 64. 45®. 69.-^=iL=.

TT
Vl - lOx*
1
60. - 62. -1.
\/4 — a;* V2ax ~ x*
1
64. cos”^x. 66. sin““^ X.

2x
66. cos~^2x. 68 .

(x* 4- 2)V^TT‘
6 a
69. 4xVl - x\ 70. 71.
4x* -f 9 X* 4-
2a
72. ^ 73. 74.
2a®
(x* + 4)* a* 4* ar* (x* 4- a*)*’

2
76. 2. 76. 77.
1 4-g^ X
3 ~ c 1
78. 79. 80.
(2x — 1) V X* —X X 2*V* - l'

1 2o* -2
81. 82. 88.
!'
Vl -x** ** +
4 1
84. 86 . 86. sin“"^ X 4- C.
’
5 + 3 cos X 5 4- 3 cosx
87. sec~^ X 4- C. tan X C, + 89. sec X + C'.
472 ANSWERS TO PROBLEMS
Chapter VI (Continued)

5. 12 ft.

96. 1207r radians per second. 96. 5^ radians per second.

^ A 2
99. 2nv,
dt dt^

100. Amplitude 6, period -• 101. Amplitude 5, frequency

o TT

3 TT 2ir
102. Amplitude - ,
period -• 103. -7-
k

Chapter Vn. Pages 184-186

10 11
(S(x + 1) 12.
. . r-
3a; +2 3a;* 4- 6a; H- 1 a; - 2
2 6
13. In 14.
x ^-1
a;.
4 - 9a;*‘

16. cot x. 17. CSC X. 18. 2 sec 2a;.

1 cos^ (x/3)
19. 20 . ‘ 21. 4 esc 2x.
3 sin (a;/3)

21nx logic e
22 23. 24. §(e* - c-*).
.

x x+5
6“*“
26. 6xe***. 26. -6c-** 27. *008*.
28. ex’~^ + e*. 29. 10* In 10. *n*(n -f a;lnn).
12
31. 32. 4xV^. 33.
^-3x)2
(e3x

4
34.
- 36. — (2 sin 2x + 3 cos 2x)e“"*®.
e**
36. c* sin 2a;. 37. sec“^ X.
1
38.
X In (oa;)

40.
2(1 + y - 2x) 41. -e*-*'.
1 — + 2/ 2a;

y(i - x)
42. 43. e* cos* y.
xiy - 1)
y(l - a;e*-^)
44. 46. — x-*(l + In x).
a;(l - ye*-*')’
46. (tan a;)*(ln tan a; + 2a; esc 2x). 47.
(x* + a?)\/x* —
+ 21a; + 12
123? 2x(x + 4ar* - 16)
4
48. 49.
2(2* 4- 3)i(3* + 2)i‘ (a? -4)V+4)*'

61. 64. 3.
66 . o*(ln a)”. 67. 1% increase.
68 . 27e-*. 69. X» y « •Jit,

60. e”*. 64. J In (2a; + 3) + C.

 ANSWERS TO PROBLEMS 473

Chapter Vn (Continued)

66. ^ In (** + 1) + C. 66 . + C.
67. ie*’ + C. 68 . 4 In 4.
69. Sir In 2. 70. Ce*^ C and k being constants.
71. j/* = 4*. 72. 3 min.

Chapter VIII. Pages 197-201

1. 2x + 32/ - 7 « 0. 2. X = (2/
- 2)2
3. (x - yf = S(x + y). 4. x2 -f 42r* » 4.

6. — 4y^ = 4. 6. y2 = ^
4 — X
7. {y - x)® = 2(j/ + x)^. 8. 2(x2 + 2/2) - 2x2/V2 = 1.
2/2
9. _ 1. -2+^ = 1-
or b^
11. 2
/^ = 4x^(1 — x^). 12. xy = 1, I
X I
^ 1.

13. 2x H- 32/ = 13, -1 ^ X ^ 5. 14. In (x2 4-2/2) =2 tan“^ ~*

y
16. The second locus is part of first. 17. (3,±4).
18. (1, 1), (0, f). 19. X = a cos <i>f y — a sin 0.
20. X — a cos 2/
= 6 sin </>. 21. X = o sec 0, 2/
= 6 tan

22. X = a cos^ 4>y y — a sin^ <l>. -• = r^'»- rT7 -

2a^^ 2Qi^
24. X = =
1 +i12. 2/
1+7'
a
26. X =
(1 - m )2 ’

(1 - m)2
a^m
./2 62
27. X = =t = T
V -f- b'^
’
2/
Vo2w2 4- 62

a^m 62
28. 2/
\/ \/m2a2 - 62’

29. 30.
^ “ 5*
2 2
2 2
31. X = 6 cos 2/
= a sin 0; -T -f ^ = 1.
6^ a*
82 X
. » a cos 0 4- sin 4>,

2/
» a sin 0 — 00 cos 0 where a , is the radius of the circle.
33. X s 00 — 6.sin <l>yy a — b cos <f>.

34. X a o cos® 4>ty — a sin® 0 x^; 4- !/^ = o^.

8o®
36. X = 2a tan 0, - 2o cos® 0 ; =
2/ 2/
^^
36. 4iro2.
37. X a 2a sin® 0 , 2/
« 2o sin® 0 tan 0 ; (x® 4- y^)x « 2ay^.
x 2 (o 4” ^)
88. X » o sin 0 p =0 tan 0 (1 4- sin 0 ); 2/2
,
o —X
39. X » -o cos (0 4- 5), y “* & sin (0 — -4).
474 ANSWERS TO PROBLEMS
Chapter Vm (Continued)
42. |(2V2 - 1). 48. 37.9. 44. 2ira.
46. fa. 46. ^/2(e^ - 1). 47. 6.
48. 2a. 49. 8. 60. =fc\/TS.

61. 8a.
8a(a + 6) 66. secx.
h
u

67. aB. 68. fasin2t.

4t/

69. 4sin(. 60. asec^t. 61. J^VI3.

62.
6*
—
a
68. iVl. 64. 00, 1.

65.
4a(o + h) sin
2a -|“ b

Chapter IX. Pages 214-217

8. ^4, 2nT ““
g) »
( t)
4. (4, 360n - 300), (-4, 360n - -
120).

88. 34.

86. 86.

(--i) •(“-!)
87.

88.

89. (a,±0,(a,±|),(O,O).

40.

49. (a? + W*)* - 2o®(®* + »*) - aV.

60. (** + »*)* - 2o*(** - »*). 61. r » 4 cot 8 CSC 8.

62. r -* 2 COS 8. 68. r* sin 28 - 2.

64. r* cos 28 * 1. 66. r — 4a (cos 8 + sin 8).

66. f* « o® cos 28. 67. r >" a sec 8.

68. rco6(fl-|)=3. 69. r{h cos 8 + a sin 8) « a6.

60. r "> 2a COB 8. 61. r "• 2a sin 8.
/ 2a
62. r * 4 cos ~ g)* 68. ^
1 + sin 8
4
64. 66. Circle.
 ANSWERS TO PROBLEMS 475

Chapter IX (Continued)

66. r o cos 0 + 6. 67. Circle.

68. Straight line. 69. r » a sin 20.

70. Lemniscate r* — 2a^ cos 20, if the points are (=ba, 0).

71. tan-* A 74. 90®.

76. 90". 76. 0®.
Qo
77. tan-* 2. 78. tan-i (8^)^
1 — cos 0 — 2 cos* 0
79. tan-* (4V3), 90*. 82.
sin 0(1 + 2 cos 0)
8S. (io, ±|t). 86. 2ira, 87. |oV3.
„„ Vl+A:*,
88. (r - a). 89. 8a. 90. fro.

«(«* + !)» 93. ^acos^" 94. oVl +jkV*.

0* +2 3 2
3 « 0
~ 97. 0*.
96. 7a sin^ •
96. fa.
4 3

98. —
n
99.
ik
100. iA:(n — n).

101. fya*. 102. io*(3\/3 - x).

Chapter X. Pages 288-2242

,
y ^ —y
5. DE = i(.AB + BC).
A. AE ~ u + CD - - Ju -y,DE- |(u + ).
AD + -JC — A. DE — —
6. »> fB 6. >= JC JA.

8. V + Ju.
9. ^«u- iv, CF - fu - iy - Ice.
2ir
10. B - A, -A, -B, A - B. 11. As =2AiCOB— -Ai.
B +mC
15. Same as 14.

mA + nB + pC as. 2, 2V3.
wi +n+p
24. AB - -M + 7j, ic - 4i + j, CA - -i - 8j.

26 -• . *»• i,
a
SO. 1. 82. » - -l.tf -3.
8S. (6, 0). 84. (I, 4).
S6. 0, 2A. 8T. i - J.

88. 0 . 40. Cl - 0, Cl - -iA.

476 ANSWERS TO PROBLEMS
Chapter X (Continued)

41. »= 4i —
3j, a = 2i, speed « 5.

42. V = 3i 2j, a = 6i + 2j, speed =* y/ 13.

-|-

48. V = — 2j, a =« — 3i, speed = 2.

44. V = 3i — 4j, a * 3i + 4j, speed = 6.
46. V =» i-\/3 + j, a = “-4i, speed = 2.
46. V =* 2(i - j), a = 4(i -f j), speed = 2V2.
47. Vx = 24, Vy = 8 — S2tj ax =* 0, Oy * —32, speed = s Vq (1
— 40 -f-

48. Vx * +
1 Vy =* 1 —
I <,
»'> '»» =* 1, Oy
O*«'y
— A1, speed = V2 + 2t^.
i,
»'> f
**

49. Vx = —t sin Vy = i cos <, ax = -~(l cost

cos t -h
-f sin t), dv = cos — ^ sin
Of ay iy ^
i

speed *=1^1.
60. Vx = —tan ty Vy =» 1, a* = — sec^ ty ay = 0, speed = |
sec <

64. -
* X*

66. at =
3 cos
—
V 3 sin^ - +
/
t
—
sin

t
t

1
I On = / — 2
'Z
-

V 3 sin* + 1 «

66 . at = 2\/o* + b*, On = 0-
67. at ^ = fc*a.
= 0, an Vi.,

a(e*‘ - e-*‘) 2a
68 .

V2(e*‘ + e"*‘) V2(e*‘ + e"*')

69. Of = oA:® cos 5 On,
= oA* sin ~
2 Z
60. at =** A;^o, On —
62. Vr — atoy v$ => a<aHy ar = — «^ai, — 2aJ^y speed = oo>V 1 +
68 . Vr * 2a(t 4- c), yg ~ ak(t + c), Or = o(2 — k^)y ae = 3A»,
speed *= a V A;^ +4 (i -f c).
a^k^
V/-—+—
;

= ^ _
=
64. Vr = a,ve —aB, Or «- ,00 = 0 speed
, a 1

66 . — (e® + e-®), ttr = 0, = a«2(e® + e“®),

[e*® -f e'
speed = Oco^-
\ 2
41.2^2^4
66. »r * —kcuT^*, V0 = 2kab^^y Or = fcV jjg — ,
« 0,

speed =* ky/a*6“®** 4- 4aVe^^

67. ar » — «V, a$ « 2(;«.

Chapter XI. Pages 268-274

1
l. |(3x + 2)i + C. 2 .

2(3 - 2x)
+ C.
8. o + te + C. 4. ox —
+ Jx* + C.
o
8. -Vo* - ** + C. 6 . |(x» + l)i + C.
1
7. t(** - l)i + C. 8. + 0.
3(2 - 3x»)
 ANSWERS TO PROBLEMS 477

Chapter XI (Continued)

9. i + 3) + c.
In (4* 10. i In (2a? - 1) + C.
11. i In + *») + C.
(1 12. In (a? + 3a + 2) + C.
13. \ sin (4* + 1) + C. 18.

16. -I cos (a* + 1) + C. 16. ^

2a
sin (oa*) + C.
17. 2tan|+ C. I tan 39 + C.
19. -^cot2» + C. 20. -I cot (a?) + C.
21. --Incos— +
5
O
3z
^+
O
4
C. 22. 5 In sin 2a + C.

23. i In sin 2x + C, 24. i sec 69 + C.

25. —CSC X + C. 26. tan d -|" sec 0 C,
27. CSC 9 — + C. cot 9 28. ^(tan 2x + sec 2x) + C.
29. ^ In (sec 3x + tan Zx) + C.
30. I In [esc (3a; + 2) — cot (3a; + 2)] + C.
31. In (sec^ x + sec X tan -h C. a;)

32. d -|- tan 0 -I- 2 In (sec 0 -f* tan 0) -f- C,

33. In (1 + 2 cot x) -f C, 34. In (In x) -f C.

1
2x
36. § sin

37. 3 sin’ 2 V4 - a* + C.
2a
39. ^tan-^_ + C.

-^l
+ c. 42. gSec
-l + c.
43. sec”^ x\/5 + C, 44. In (x + Va;2 + 4) -f C.

46. I + V4a* + 3) + C.
In (2a

46. ^ln(x^/2 + \/2a? + 5) +

ne. 1 , 3; -- 4 . ^ 2a; - 3
-hC.
+3 2x
.. 1
^ 2a-v'6
, . ^ 60. -ie-3* + C.
2 V6 2a +
^^ - -
61. -§e- * + C. 2
62. i(e 2x + C.
68 . -ie-»^ + C. 64. 2 V 1 — cos 6 -f C.

66. sin”^ “-7=* + C. 66. I sin“^ “2 +

V2
67. I In (1 +«**)+ (7. 68. tan“^ (c*) + C.
69. — tan”^ — ^ + C. 7
60. — sin”^ (6“0 + C'.
\/2 V2
61. J sin^ X + C. 62. — ^ cos® 5x + C,
1
478 ANSWERS TO PROBLEMS
Chapter XI (Continued)
$ + sin* $ + C. 64. J cos* a? — cos * + C.
I sin 2» — i sin* 2x 4- C, 66. i cos* x — i cos* x C.
sin
—cos 05
05 — f
—
sin* x
cos*
sin*
4-
05 4- C'* 68. ^
70. In (esc
sin^ 4^
x
— TT sin* 4® 4-
— cot x) 4- cos
C.
4- C-
i 05 C'. 05

In sin a; — J sin* a? 4- C. 72. ^ tan* a? 4- C.

i tan^ a? 4- C. 74. | tan* x 4- C.
I tan* X 4* C. 76. tan X 4“ i tan* x 4* C.
tan X — X 4- C. 78. 2 In sin a — cot a 4- C.

i tan* X 4- i tan* x 4- C.
—cot X — f cot* X — cot* X -J-
4* C.

— J cot* X — i cot* X 4- C'. $ 4- tan e)

82. In (sec 0 0) 4“
4- C.
Jx 4- A
sin 8x 4- C. 84
84.. |0 4- 2 sin a
0 4- i 4-
J sin 20 4“ C'.
<

fx 4- 4 cos X — J sin 2x — J cos* x 4- C'.

fx — J sin 2x 4“ Tnr 4x 4-
gX — sin 8x 4" C.
sin 40 — -3^ sin* 20 4- C.
A* 4- i sin 2x — -3^ sin* 2x 4- sin 4x A -f C.

— J cos* X 4- C. 91. J cos 2x — cos 4x 4- C.

i sin X — A sin 5x 4- C'. 98. J sin 2x 4* A sin 6x 4- C'*

— J sin X 4- i sin 3x — A sin 7x + C'*

In (esc 2x — cot 2x) + C. 96. —cot -

~ 4- C.
4"

2V2 sin 4- C. 98. 4V^

4V2 ^sin
(sin ~ ^g sin* + C.
4-
^ ^

sm“^ —V37^ 4- C. 100. -

1

2
.

sm~^
,
—-7^
2*
-v/2
1
4- C.
^

1 (x - l)\/3 ^ 1
—
X 4 3 ^ -
—7= sin"^,
.
7=
.

h C. itan-‘^
102. “ tan“^
,

+h C.
VS V5 2 2

-—7= tan
2V2
^
—V27^ 4- C. 104. - sec""'
sec
2
— +h C.c.
r.*-^
2
^
^
V5
seo-‘
v^
+ C. 106. sec-*
see-* (2»
(2* - 1) + C.
iln(x4-3)4-fln(x-l)4-C.
I In (x* 4- 2x — 3) 4-
iln (2x* - 2x 4- 1) - i tan""' (2x - 1) 4- C.
I In (05* 4“ 2x 4" 2) — tan"' (x 4* 1) 4- C.
— enin“l
in"' 'v/l ^
Vl 4x — 4x* ^
4- Afr mm.
4“ C.

*^05* 4" 2x 4“ 3 — 5 In (x 4“ 1 4* “V^oi^ 4“ 2x 4* 3) 4*

'3** - 6» - 1 + -^ In [VS(* - I) + V3®* - 6* - 1] + C.

V3
, + C. 116. + C-
^
V»*-2x + 8 V5 V5
 ANSWERS TO PROBLEMS 479

Chapter XI (Continued)
116. X +iln(a: - 2) - |ln (x + 3) +C.
117. 21nx-ln(x + l) + C.
118. ix* + 4x - 2 In (x - 1) + 12 In (x - 2) + C.
119. i In (x* - 1) - In X + C.
120. 41n (x + 1) - f Inx - f In (x + 2) + C.

121. X + - + 2 In (x — 1) — In X + C.
X

122 . + ln(x + l) +C. 128. In (x - 2) - - + C.

x +1 _ 2
6
124. 21n
X — 1 X — 1
+ C. 126. -+4
X or
+ ln^^
X
+ C.
1 / X 4-1 2x \

127. In — X + tan * X + C.
Vl +x*

128. In
Vx*
x +^+ 1
1
+ C. 129.
x^ — 1
In^-r^+C.
x^
,
+ 1

*“ * + i Sfl + tan -xJL

^-^ + C.
5^^ V5
132. ^/2 tan“^ — tan”^ x 4- C.

133. tan"^ ® ^
2
tan""^ ?
2
+ C'.

136. Jln(x*+4)+-j^+C.
4 X 4“ ^

136. In x + x^2 4 1 4- C,

loi*
137.
lx -1 ® “ 1
“Tstan ^
\/5 \/2 - 2a; + 3 -hC.
138. —[a; + 4\/x + 4 In (1 — \/x)] + C.

189. f (x + 1)* -
+ C. 2(x + l)i

140. f o(x -
+ t(x o)* + C. - a)*
141. 2Vx + 2 — 2 tan~^ Vx + 2 + C.
142. 2 tan~* Vx — 1 + C.

+ 2 In (1 + >/x - 1) + c.
lU. |x* - + 4 tan-‘ (x^) + C.
4xi
146. 2xi + 3x* + 6x* + 6 In (X* - 1) + C.
480 ANSWERS TO PROBLEMS
Chapter XI (Continued)

146. id + Vi)* - id + Vi)* + C.

147. — + 2o*)Vo* — ** + C. 148. — ^Va* — x*

a — Va* — a;*
149. -In )+c. 150. -
a (
\ z a^z

151. + C. \/x^ — c?
152.
<»Va* - X* o?x

1 / aV z^ — a*'
153. I sec _i®
^ -
).c.
2o^ \ a r*
X
y/^ — — a sec”^ - + C.
1

154. d?
a

If ,
» oV X* — ar\ , ^
155.

156.
(o^ + 2x^) Vx^ - + C.
3a^®
157. ]^(3x* - 2o*)(x* + a*)* + C.
Vo* + X*
158.
o^
+ C.
159.
«Vo* + X*
160. In (x + V + x^) — -hC.
y/a^ 7?
ox
161.
2a \ ax
a
162. X sin X + cos X 4“ C.
168. sin X — X cos X + C.
164. X tan X + In cos x -H C.
165. z sec z — In (sec x + tan x) + O.
166. (x2 - 2x + 2)e* + C.
167. ^x®(5 In X — 1) + C.
168. i(x2 4-C.
169. X sin”^ X + Vl — a? + C,
170. X tan”*^ X — i In (1 4- X*) 4- C-
171. xln (x + y/^ + g^) ~ V x^ 4- a* 4- C.
— JV —
172. Jx^ sec“^ X X* 1 4- C.
178. 2(x — 1) sin X 4- (1 4- 2x — x*) cos x 4- 0.
174. JxV x^ — a* — ^a^ In (x 4- V ^ ~ a^) 4- C.
175. ixV a? + 4“ in* In (x 4- Va? 4- a^) 4- O'.
176. ixV a? — a* 4- i®* In (x 4- V x* — a*) 4- C.
177. 3^*(2 sin 3x — 3 cos 3x) 4- C.
178. (sin X — cos x) 4“ 0.
 ANSWERS TO PROBLEMS 481

Chapter XI (Continued)

179. — J^(3 cos 3® cos 2a; -f 2 sin 3a; sin 2a;) -f- C.
180. i sec a; tan X + ^ In (sec x + tan x) + C.
a;(3a2-2x2) ^
181. ^a«. 182. T -f
.

C. 183. 3.

184. Divergent. 186. Divergent. 186. 10.

187. 2. 188. i 189. Divergent.
190. Divergent. 191. Convergent. 192. Convergent.
193. Divergent. 194. Convergent. 196. Divergent.
196. Convergent. 197. Divergent. 198. Convergent.
204. 2.0392. 206. 8.040. 206. 9.079.
207. 2.285. 208. 1.760. 209. 0.836.

Chapter XII. Pages 293-299

1. ToVo* + h\ 2. W(2\/2 - 1 ). 8. ^ir.

2irah. 6. V’T. 6.
7. 2xa(x2 — xi). 8. 9. A*.
10. W(2 - VS). 11. 4ira*V2. 12. -^ira*.

18. ^ira*. 14. iro*lV3 + 5\/2 In (V2 + VS)].

18. 6A*. 16. 4p»(2V2 - 1). 17. 2a*.

18. 2fcir*a*. 19. 16a*. 20. a*(V3 - §»).

21. 8a*. 22. 16a*. 23. -•

IT

24. i
27. W- 28. igt. 29. 2a*.
30. firO*. 31. §A:*a*. 32. jKil*, I = ^Ay/2
38. 37.6 Ib./in.* 84. |p6*/>lb-ft. 86.
86. 70,312,000 Ib-ft. 37. Ib-ft. 38. ^h.

39. ft. 40. (2.4, 0).

43. (-Ja, ia).

in. from 12-in. edge, 2y in. from 24-in. edge.

..
46.
T.. . .
Distance from center
. = —
4(0^ + 06 +— 62)
3ir(o + 6)

46. (3 /^, ^a). 47. (i I).

49* (fi, —•^)- 49. (0, ia).

3^3 £ =
60. £ 61. 5 TO, p= |o.
2x 2t’
62. (to, to). 63. 2 »» 0, y a |o.
64. « 0, p « « io.

66 . On the bisector at distance -a ™ from the center of the circle.

a
482 ANSWERS TO PROBLEMS
Chapter XII (Continued)
sm a
56. On the diameter that bisects the arc at distance a from the center
of the circle.

67. i - 5 “ 0. 68, i - 9-0.

69. a — ixo, # — 0.

60. On the axis at distance from the base.

61. On the axis at distance jh from the base.
62. (|o, 0). 68. (0, Ao).
64 . (0 , |6). 66. (K 0).
3(6* - o<)
On
66. the radius perpendicular to
the center.
its plane face at distance
8(6» - o*) from
67. The midpoint of the radius perpendicular to its plane face.

68. At height k. 69.

88.
At height ——
m+ 2
70. At depth f A 71. At depth §A
87, \vobhK
72. 74. 2ir*o*6. 89.

76. At distance 2r/r from the center of the circle.

76. Jx(6-v/3 — 1). 77. - 2).

78. (xo, |n). Area — ^xo*. 79. 16xV.
80. f xo* (cos p — cos a). 81. 9 — 0, r •• fa (cos a + cos /}}.

82. 8.
|xo»[3 In (1+ V2) - VS).
84. io6». 86. \bh*. 86. iMo*.
ixo6>. 88. fMo*.
90. + f|). 91. ^ab*. 92. ifJlx.
98. jMa*. 94. iaVTO. 96. fo-y/S.

96. Vh* + f o*. 97. 98. fMoV.

Chapter xm. Pages 319-826

1. S - 1.

0 .

8. ,«-l. 4. S = .

n "T 1 1 +a
6. Divergent. 6. Convergent.
7. Divergent. Divergent.
9. 8n - id - 3~"), S- i 10. Sn - ill - (-4)~»‘], 8

11 . Sn

12. Divergent. 18. Convergent.

14. Divergent. 16. Convergent.
16. Divergent. 17. Divergent.
18. Convergent. 19. Convergent.
 ANSWERS TO PROBLEMS 483

Chapter Xm (Continued)
22. Convergent
20. if p > 1, divergent
23. if p ^ 1. 24.
21.
28.
Convergent if p > 1, divergent if p ^ 1.
29. 30.
Convergent. Convergent. Divergent.
86.
26. Convergent. 26. Convergent. 27. Divergent.
40. Divergent. 41. Divergent. 42. Convergent.
81. Convergent. 32. Convergent. 33. Convergent.
34. Divergent. Convergent. 36. Convergent.
37.
60. Convergent. 61. Convergent.
88. 62. Convergent.
39.
Divergent. Convergent. Divergent.
66. 67. 68.
48. About 400,000,000. 47. ^ < 1.

48. —2 <x ^2. 49. Convergent for all x.

-1 ^ X ^ 1. -1 <x <h -1 gx ^ 1.
63. -1 < X < 1. 64. I
a:
I
< ^\/2. 66. -1 g a: ^ 1.
-1 < X < 1. -1 <x <3. -2 gx <0.
69. -6 < a; < 1. 60. Convergent for all x.

61. ^ < as < -J. 62. X > 1. x-1 >1.

+
1 1

64. a:+ a? +x» +•••. 66. 1 + fx* + T^x*

66. X + fa:* + A®* + •
67. 1 +5®

X* X*
68 . —X — — §x® —
^+2i + i!+-
69. Jx^ •

70. —fa? - — -Ar** 71. a? - ^x^ + iV* “ *

lx* l-3x*
72, * + 23 73. x-fx2 + ^x«+*--.
e^(x - 2)2 e^x - 2 )®
74. e* + «*(x - 2) + 2! .

'
3!
.

76. (X - 1) - i(x - 1)2 -f i(x - 1)2 - i(x - 1)^ +....

87.
76. 2 + — 3) — -^(x —88.3)2
-f irhyCa? — 3)* — • • 89.

77. 2 + J(x — 4) — -a^Cx — 4)2 -f ^^(x — 4)2 — • • •.

78. -1 - (X + 1 - (X + 1)2 - (X + 1)2

) .

79. 8 + 12(x - 2) -f 6(x - 2)2 -f (x - 2)2.

80. l+i(x- 1 )+%(X - 1)2 - 1)2 +•••.

81.

82.

83. 3 - 2(x + 2) + i(x + 2)2 + A(x + 2)2 +. ...

84. 0.0872. 86. 0.0876. 86. 0.182.

1.887. 0.933. 1.649.
90. 0.833. 91. 0.100. 99. 0.00019.
101. X < 0.0316. 102. X < 0.413. 108. X < 0.66.

104. 0.693. 106. x2-ix^+Ax®+---.

484 ANSWERS TO PROBLEMS
Chapter Xm (Continued)
m
111. X
1

3
3 ,

H
lx®
5 2!
112. 0.493. 118. 0.464. 114. 2.004.

115. 1 — + ix®
» H —
117. 1 4- ^~2
+4 ^~
^ —
!
a;®
6
X® ^
+ !
-f-
!
• • •

118. 1 + x + ^x® -f ^x® +•••.

Chapter XIV. Pages 346-350

1. i. 2. t.

3. 2 - Ih*. 4. 18 - t.

6. 9 + 14t, 9 - 14t.

11. AB = 2i + 2, BC = 3t - 1, AC 5i 4 1.

14. 2 (cos ir 4* t sin v). 15.

3 ^cos ^
4 t sin ^ •

2
16. 2 V2 ( COS
^
t sin •

(
cos 4 ^ sii:

19.
6 ^cos ^
^ sin

20 . 2 . 21 .
~8 .

22. 1 + iV3. 23. -34 V3.

24. db(l 4 t). 25. 4:(\/3 4 0*
_
26. ±i\/2(V3 - 0. 27. ±(1.0987 4 0.46510.
29. cos 30 == cos® 0 —
3 cos 0 sin^ 0, sin 30 = 3 cos® 0 sin 0 — sin® 0.

80. cos 40 = cos^ 0 —

6 cos® 0 sin® 0 sin^ 0, 4
sin 40 =a 4 cos® 0 sin 0 — 4 cos 0 sin® 0.
31. z = —2, 1 ± iy/S, 32. z i( — 1 ±
i\/3),
33. a » ±2, ±2i. 34. z ±i, ±iy/5.

36. z *= i(\/3 ± i), — i(\/3 ± i),

86 . 2 “ §(v^ ^ Az i), ±i.

39. I
2 » 2.
I
40. 2-3 = 1. 41. I
2 - 2 I
- 3.
42. All values. 48. 2 4 2 = 1. 44. 2 » 0.

46. 1
2 - 2 « 1. 1
46. All values. 63. -1.
54. i - iiy/S. 55. -e®t. 56. 0.5403 4 0.8415t.
67. r - \/5, 0 - tan““i 2.
58. Point where circle touches x-axis. v «« w(±at — o).
i
59. C =» 2o 4“ (a
where 20 is the center of the wheel.

60. Intersection of CB and the line perpendicular to CA at A.

icu/^
62. 2o 4 -5 — ta
r* where zq is the lowest point on the rolling wheel.
 ANSWERS TO PROBLEMS 485

Chapter XIV (Continued)

63. Center of rotation at point of contact of the two wheels; center of accel-
0^
eration on at distance CO from C.
a -f h
66 . + i(v - ao))tl
67. In « + 2in.
69. In 2 -|- yir\/3.
71. cosh « f tanh x — coth = t, sech x = y, csch x = f
a; , as

72. sinhx = itf, tanhx = dbf, cothx = sechx = cschx =* db^. 3^,

73. sinh x = ]^, cosh x = coth x = sech x = yf csch x = ,

74. sinh x = ±0.75, cosh x = 1.25, tanh x = ±0.6, coth x = ±1.67,

csch X = ±1.33.

76. 76. M.
90. I cosh® X — cosh x + C. 91. X — tanh x + (7.
92. —sech“^ X + C. 93. — csch"^ X -f C.
94. 0.8814. 96. -0.1059.
96. 0.2747. 97. 0.2682.

Chapter XV. Pages 389-397

1. 2. (0,j^, 0). 3. v/6.

4. A(i - 2j + 2k). 6. Vl3- 8. (5, 2, 5).
10 . i(16i - 5j + 4k). 14. (a) -lx®, (6) |x®. 19. j
- k.
20 . j+k. 21 .
-1 . 22 . -I.

24. cos ^
I, cos ^
y, cos ' ?. 26. 45°, 135°.
26. 60°. 27. 30°.

28. A = 30°, B 120°, C = 30°

29. A 135°, B = cos-1 -7= ,

C
V5 Vio’
30. a; = |(x' + 22/' + 22'),
y = |(2x' + - 22'), 2/'

z = |(-2x' + 22/' - 2').

31. i - 5j - 3k. 32. 6.
33. 3i 34. 5.
38. N=i -f- 2j - 2k, distance 39. |(2i + j + k).
40. i(i - 5j + 4k). (tV> A)-
42. 4V6
6
"' 43. -S-.

44. M =• Hcy — hz) + j(02 — cx) + k(ba: — ay), Mx = cy — bz.

46. 7k- 81- 3j. 46.-18.

47 4VI4. 60. 2.
61. 4. 62. I
68. A BxC-0. 64. AxB CxD =>0.
66. Ax(BxC). 66. (AxB)x(CxD) - 0.
;|A|BxC + m|B|CxA + n|c|AxB
69. V
A BxC
486 ANSWERS TO PROBLEMS
Chapter XV (Continued)

60.
39i + 18j + 15k
11

67. aj* + P* + — 4a? — 22 — 4. 68. a* + ** - r*.

69. 2 — p * 1. 70. p » 0, 2 - zta?.

72. a?* « 82 . 73. 2a? + 3p — 2 « 3.

74. X + yy/2 +2 — 0 . 76. 3a? — p 4" 22 +1 "» 0.

77. p — a? « 1. 80. 60®.

81. cos~^ 82. 45®.
a? — 2 p +1 2 — 3 r +l
83. 84. a? — 2 = p — 3
2 -4 3
a; -3 y-i X -2 y-l
86 . ' a = 2 + 2. 2 4" 3.
4 -2
87. 3a? —2 = 0. 88 . 2a? — 32 2 .

89. 30®. 90. 60®.

91. 45®. 96. p^ 4“ 2* *“ 4pa?.

97. y*(x* + **)-!. B.

a?

^2 +
V4-

w j^2

99. 4(a!* + **) - 9(y - 2)*. 100. (2 3)* - (X - 1)* + (y - 2)>

101 . (* - 1)* + (r - 2)* - 16. 116. 2Ty/2.
117. 12 . 118. 50.
d*F
_
119. V2 In (1 + \/2). 120. Fx^.
121 .
„ dF d*F
12
y — 2 * — 1
—
123. a? 124. tan”*^
3 4 m
i 4- cos a? — k sin ®
n =
j
125. t — (j sin a? 4- k cos a?), p — 2.
V2
126. t
i cos a? 4- j sin a? +k n -
,,
J:(j cos — , .

sin
^ ^ ,

sec
,
, a? i a?), p 2 |
a? |.

127. -i + j+ k.«-j+2k,/i - l-v/5.

128. V2o* + 1 (e - 1). 129. Jo In 3.

130. ^— 9. 181. 2awu» — w*rUr.

132. v='Ur+k,a-2u«, p 1.
dk
133. T« _ 6 + 5j + 4k,^ i = 2k + 4j/^ k-41.---2i-J.
d*r d*ro . , „
136.

Chapter XVI. Pages 429-436

1. 2(a? - p), 2(4p - a?). 2 . p cos (a?p), a? cos (a?p).

2y 2x
3.
X y •
(x-y)*’(x-»)»‘
6. 2a? —^ 4- 22. 6. 3 cos (2x + 3v + s).
 ANSWERS TO PROBLEMS 487

Chapter XVI (Continued)

7. 8.
Vx* + y* + «*
9. 3(3? + v^ -2xy- «*). 10. 5.
11 . 2 . 14. 1.
6*
15. r*.

c* sin* B he
17. 18. -sin A.
2sin*(A +B) a
19. (1, -2, 3). 20. (-1, 3, 10).

22. {2x — dx (6p — x) dy, 23.

xdx + ydy zdz
y) -f-
V X* -f y* -f 2
*

24. 2xy*2^ dx + 3x*yV dy -f dz.

25. ~
—X + dv
(2x
—
d2
z
26. dz = -0.15, A2 = -0.1495.
y
27. du *= 0.7, Au = 0 688
. . 28. dv = 64 in.*, Av — 55.64 in.*

30. 0.03 ft. 31. 0.006 radian.

32. 0.07. 33. 0.
dz
34. 0.302 ft. 39. sin (x -I- y) ~ + 2 cos (x -f y).

40. 2x + —dz
2 (^5 22)
dx --"(-I)'
y + 2 ax dx dy

44.
^
dx
+ 2|^.
dy
46. —
dx
+ 2* dz
dF dF
+—
dFfdf d/
0' W + T" T" + T"
. ,, , .
,
46. .'(*)]•
dx dy d2 LdX dy

dF dF dx
47. B8.
dx dt^ 5/
dy

dy dx
69. -fVa. 60.

dy d2

62. 63.
u +v
p" ' e
«» + .» 1 + wy
2 y -X
64. 65.
* —X V
-2 *-2
68 . (x - 1) + 2(y - 2) + 2(* - 2) 0, X - 1

3x+ y - 8, normal x — 3y - 6, * • 1.

70. 3x + 4y - 52 = 0,
X— 3 y — 4 2 — 5
3 -5
488 ANSWERS TO PROBLEMS
Chapter XVI (Continued)

a;4.32/-« = 6, = -y-3
2/
—
= ——
z 5
j-

3* - 2y + ^ = 6,
——
X 2 t/ 4- 3
= * + 6.
X — + 2 —
2y + 3*-ar = l,—2
= 1/ 2 3

^drd(t> /d0\^

ddd<t> drdB
09 =r8.n0^ + ^2rco80-- + ^2sm0--- ,
.

/d0\* . , /da\^ <P>i> fde\^

rfd d<t>
o,=as.n«^ + ^2acos«---

^ rf? ^
•

(0, 0, ±1).
The altitude of the cone is twice that of the cylinder.
(3, 3, 9).

2
VV_£^ _
dx dy dx dy \dx/ dy^ \dy/ dx^
di*

df aV aV dH
2 —
dy
H 5 4*
dx^
4x
dxdy
1- 4x^
«
—
dy^
«

<*’+•=>
(3
^
2u
/d^F
— ,
4"
dl‘^F\
"Ti 4-
. ,
4f; -—— H" ^dF
a^F
2“
.

\au^^
(
dv^/) dudv du
dV_^
dx* dy*’
_. (m +n)!
Amn = 96. A„„ = o-^d-"

Not exact. 98. Exact. 99. Not exact.

Not exact. 101. Exact. 102. Exact.
Exact. 104. 1. 106. -1.

107. ie* + je* 4- k(x 4- y)e*.

109. -u^. 110 .

-1 .

r
112. f 118. 31.5.
116. 1.
 ANSWERS TO PROBLEMS 489

Chapter XVn. Pages 468-463

1. 2. 2. i In 2. 8. Ja.
4. 4. 6. 6. 6. W-
11.
£ £^'‘Kx,y)dxdy.

X £fi^.y)dxdy. X X*
16. 2 = 1, 2/ = 0. 16. 2 = jl
= -1. 17. M^a*.
18. + 6*). 19. 20. 6.
21. |. 22. 8. 28. 1
24. 4. 26. %*-. 26. T^*.
27. Jjral 28. ixa*. 29. ixa*.

80. MV2 + In (1 + V2)]. 31. ^x.

V2 -
32. 1. 33.
^ (2x + 3\/3).
84. iaV3. 86. 2 =f g = 0.

se. j = + « .

6ir + 12 ’
" X +2
„„ ,
87.*-
(x + 2)o . (14-3x)o
•

j ^
88. iMo®. 39. ixaX
40. fo^(2a — sin 2a). 41. 4xa*.
42. §xo®(l — cos a). 43. iixa».
44. iM(46* + 3o*). 46. 27p.
46. T^. 47. fxo^
48. i 49. (fa, lb, fc).
60. ^xa6c(o* -t- b^). 61. 4x-y'^2.
62. ^(8 - 3V3)a*. 63. |Ma*.
64. + 4A*)- 66. iMa\
66.
45

67. TP (b^ - (6* - «*) + 1 (8* -»)].

1

68. fx(5\/5 - 4). 69.

60. §xo*, xo*. 61. I-Tra*, ^-ira*.
62. ^xa». 68. Jira®.
64. :,»gn»(3x + 4). 66. TT.

67. fMa^

68. On the axis at distance from the base.

69. On the axis at distance from the base.
490 ANSWERS TO PROBLEMS
Chapter XVn (Coatiaued)

70. On the x-axis at distance 'fga from the center of the sphere.
71. On the axis of the cone at distance f a(l + cos a) from the vertex.
67>o»
72 ’
78. 4o*.
480
74. 75. J»{2(6* - o*) - 3c(6* - o*)].
76. On the radius perpendicular to the plane face at distance |a from the
center.
77. On the axis at distance from the vertex.

78. 8fai‘. 79. « - S - J - -fsa. 80. jAfo*.

81. ^a*. 82. 5* + io*.
88. ialV2 + In (1 + x/2)]. 84. ia*.

85. fi. 87. 3Vl4.

X
88. iro*\/3. 89. iro*V5. 90. fvo*(3\/3 — 1).

91. 4. 92. 2irV6. 98. T-s/lS.

94. i»o*V2. 95. ^

 TABLE OF INTEGRALS
u” du = —
n+ 1
,
if nis
is not
n —1.

2. f
J u
—= In u,

o f du 1 . . u
J u* 4- a* a a
. r du 1 , u —a
Ju* — a*”'2a ”u4-a
5 .
^e'^du = <}“.

« /“'“-I;-
10.
Integrals of Trigonometric Functions

7. J*sin u du =» —cos u.

8. sin* u du = J u — J sin 2 u = — sin u cos u).

J*

9. J* sin* u du = f u — i sin 2u + sin 4 u.

J* sin* u du = u — i sin 2 u -f sin* 2 u + A sin 4 u.

n. f COS u du =3 sin u.

12.
J*
cos* u du = § u + J sin 2 u = i (i* + sin u cos u).
13.
J*
cos* u du = i u + J sin 2 u + A sin 4 u.
14. J* cos^ u du = u + I sin 2 u — iV sin* 2 u + A sin 4u.
16. J^tanudu » —In cos u.

16. J*cotudu « Insinu.

492
 TABLE OF INTEGRALS 493

J*8ec 14 dti = In (sec u tan u) = In tan -f

—
J sec* udu tan u.

J sec* 14 di4 = i sec 14 tan t4 + | In (sec u -f- tan u).

CSC udu = In (esc 14 — cot u) = In tan

J"

J*C8C*14dl4 = —cot 14.

J* CSC* udu = — 1 CSC 14 cot 14 -f J In (esc 14 — cot 14).

Integrals containing y/a^ — u^

J
r Va* — 14* di4 = ^2 Va* — 14* + ^2 sin'i -•
a

= ~ (2 14* —
^ v?y/a^ — u^ du Va* — w* Is:
o*) s
^ Q
« -
J U + aln 14

du -t1
= Sm •
-•
/V a* — 14* ®
U^ du U ^/-r a* U
- = — Va* — ;
^,
.
1 _
/ Vo*
Va -— 2 1x2
14*
o
2
1^2
2
gin
a
.

dl4
—— = 1,0 —
- In
Vo* — 14*

/ 14 Vo*
--z L-.

14
-
* ® ^
,

du _
~ Vo* — 14*

/ 1^2 Vo* — 14*

(a* — 14*)^ di4 = (5 o* — 2 14*) Vo* — 14*

-f sin"‘
J* g
du 14

/ (a* — 14*)^ a* Vo* — 14*

Integrals containing Vu* — a*

32.
^ Vi4* — a* (ii4
=^ Vt4* — o* —
^ lo (u + Vi4* — o*).

— o*(2i = — a* —
^ In (li-f Vv*—
38. J*
u* Vi4* 4 (2 14*— a*) Vo*
g
494 TABLE OP INTEGRALS

34.
< §“ 1

35. r .

J Vu* - o*
— du "• In (u

r u* du ^ u
36. Vtt> -
J Vu* - o* 2
r du ^ 1
37.
u Vu* — a* a

38.
rdu
J U* “V^U* — G* o*u

— a*)*du 3oS
30. (u* (2 u^
J* ^
r du ^ u
40.
^ (u* — a*)*

Integrals containing Vu* + o*

aMu Vr* + a* -f ^ In (u + Vr* + a*).
41,
J -I-
^

42. J* Ti* Vti* + a* du " ^ (2 u* +o*) Vu* a* — ^In (u + Vu*+o*)«

43.
J
r y^
u
» Vn« + a* -h Q Ip
u
— ?.

44. f ;
.-a «= In (ti + Vr* -fo*).
Vtt* + o*
45. f - - ^^ Vw* -f- a* —^ In (tt + Vu* +o»).
•/ V u* + o* «

-
Jf u Vu»
46.
+ o* ® «

47. f - ,
dfi
« + «*
J ti* Vu* + o* ahi

iS. j’(u* + o*)*du - 1 (2 u‘+6 o«) Vu* + o» + («+ ViS+S),

du u
(tl^ + o‘)* a* Vu* + a*
 TABLE OF INTEGRALS 495

Other Ihteobalb

- (««n6* - 6 coete).
r«“rin6*<fa
' a* +0*

f
e cos 6® a® « (6 sin + a COB
r-rr^
ft®)

“•J
1 a* + 6*
496 NATURAL LOGARITHMS

N
0

1
0

2.8026
1

2.3979
2

2.4849
DD
1.0986

2.5649
1.3863

2.6391
5

1.6094
6

1.7918

2.7726
7

1.9459
8

2.0794
1-509

2.1972

2.8332 2.8904 2.9444

2 9957 3.0445 3.1355 3.1781 3.2189 3.2581 3.2958 3.3322 3.3673
8 4340 4657 4965 5264 5553 5835 6376 6636
4 6889 7136 7377 7612 7842 8067 8286 8712 8918
6
6
9120
4.0943
9318
4.1109 4.1271
>na
4.1431
9890
4.1589 1744
[»Ml
1897 2047 2195
cKiVw.!
2341
7
8
0
2485
3820
4998
2627
3944
5109
Ml2767

5218
4188
5326
3041
4308
5433
3175
4427
5539
4543
5643
3438
4659
5747
3567
4773
5850
3694
4886
5951

10 6052 6151 6250 6347 6444 6540 6634 6728 6821 6913

11
12
7005
7875
7095
7958
7185
8040
7274
8122 Ml Ml
7362 7449
8283
7536
8363
7622
8442
03 7791
8598
13 8675 8752 8828 8903 8978 9127 9200 9273 9345
14 9416 9488 0558 9628 9698 9767 9836 9904 9972
15 5.0106 5.0173 5.0239 5.0304 smMi 5.0499 5.0562 0689
16 0752 0814 0876 0938 1059 1120 1180 1240 1299
17
18
19
1358
1930
2470
1417
1985
2523
1475
2040
2575
1533
2095
2627
2149
2679
Ml1648

2730
1705
2257
2781
1761
2311
2832
1818
2364
2883
1874
2417
2933

20 2983 3033 3083 3132 3181 3230 3279 3327 3375 3423

21 3471 3519 3566 3613 3660 3753 3799 3845 3891

22 3936 3982 4027 4072 4116 4161 4205 4250 4293 4337
23 4381 4424 4467 4510 4553 4638 4681 4723 4765
24 4848 4889 4931 4972 5053 5094 5134 5175
25
26
27
915215

5984
5255
5645
6021
5294
5683
6058
5334
5722
6095
5373
5759
6131
5413
5797
6168
5452
5835
6204
5491
5872
6240
5530
5010
6276
5568
5947
6312
28 6348 6384 6419 6454 6490 6525 6560 6595 6630 6664
29 6609 6733 6768 6802 6836 6870 6904 6937 6971 7004

80 7038 7071 7104 7137 7170 7203 7236 7268 7301 7333

31 7366 7398 7430 7462 7494 7526 7557 7589 7621 7652
32 7683 7714 7746 7777 7807 7838 7869 7900 7930 7961
83 7991 8021 8051 8081 8111 8141 8171 8201 8230 8260
84 8289 8319 8348 8377 8406 8435 8464 8493 8522 8551
85
36
87
8579
8861
9135
8608
8889
9162
8636
8916
0189
8665
8944
9216
8693
8972
9243
8721
8999
9269
8749
9026
9296
8777
9054
9322
Ml
8805

9349
8833
9108
9375
88 9402 9428 9454 9480 9506 9532 9558 9584 9610 9636
89 9661 9687 9713 9738 9764 9789 9814 9839 9865 9890

40 9915 9940 9965 9989 6.0014 6.0039 6.0064 6.0088 6.0113 6.0187

41 6.0186
7^ Ml
iSMM SSI 0307 0331 0355
jttr
42 lltSS 0426 0474 0521 0544 0568 0591
43
44 91
0638 0661
0890
0684
0936
0730
0958 sn 0776
1003
0799
1026
0822
1048
2J i/n
45
46
1092
1312
1115
1334
1137
1356
1159
1377
1181
1399
1203
1420
1225
1442
1247
1463
1269
1485
^
S'
S' rTi
47
48
1527
1738
1549
1759
1570
1779
1591 1612
1821
1633
1841
1654
1862
1675
1883
1696
1903 *^
49 1944 1964 1985 2046 2066 2086 2106

50 2146 2166 2186 2206 2226 2246 2265 2285 2305 2324

8
 INDEX
The numbers refer to pages.

Abscissa, 4 Center of gravity, 224

Absolute convergence, 305 in space, 286
Absolute value, 1 of a plane distribution, 282
Acceleration, 66, 229 Centroid, 224
angular, 167 in space, 287
Adiabatic law, 298 of a plane distribution, 282
Algebraic equation, 102 Change of coordinates, 129
Alternating series, 307 Change of variable, 263
Angle, between two curves. 111 Circle, 113
between two lines, 106-108, 357 involute of, 199
Angular acceleration, 167 of convergence, 334
Angular velocity, 167, 387 polar equation of, 207
Arc, differential of, 193, 210, 380, 385, Circular measure, 11
413 Cissoid, 199
length of, 190, 379 Cofactors of a determinant, 150
Archimedes, spiral of, 204 Commutative law, 220
Area, as limit, 80 Commutative operations, 38
in polar coordinates, 212 Comparison test for convergence, 304
of a cylindrical surface, 278 Complex numbers, 326
of a surface, 457 conjugate, 326
of a surface of revolution, 276 geometrical representation, 328
under a curve, 85 polar form, 328
Associative law, 220 powers and roots of, 331
Asymptotes, of a curve, 136 product and quotient of, 330
of a hyperbola, 126 series in powers of, 334
Average speed, 27 Components of a vector, 225
Average value, 279 in cylindrical coordinates, 385
Axis, of a parabola, 116 in polar coordinates, 234
of imaginaries, 328 in spherical coordinates, 413
of reab, 328 normal, 231
tangential, 231
Binomial 313
series, Conditional convergence, 305
Boyle’s law, 295, 298 Cone, 379
Confocal ellipses and hyperbolas, 143
Cardioid, 203 Confocal parabolas, 118
differential of arc, 216 Conic, 127
Cartesian equation, 102 polar equation of, 208
Catenary, 197 Conjugate axis of a h3rperbola, 125
radius of curvature, 197 Conjugate complex number, 326
499
500 INDEX
Constant of integration, 74 Determinants, cofactors, 150
Continuity, 17 general definition, 147
function of two variables, 398 minors, 150
uniform, 21 nth order, 147
Convergence, of an integral, 262 properties of, 148
of a sequence, 6, 19 second order, 146
of a series, 300 Differentiable, 31
absolute, 306 Differential, 33
circle of, 334 exact, 419
comparison test, 304 partial, 402
conditional, 305 significance of, 34
general condition for, 301 total, 402
integral test, 302 Differential of arc, 193, 380
ratio test, 306 cylindrical coordinates, 385
region of, 309 polar coordinates, 210
Coordinates, cylindrical, 373, 451 spherical coordinates, 413
polar, 202, 234 Differentiation, 37
rectangular, 4, 351 logarithmic, 181
spherical, 411, 464 of algebraic functions, 37
transformation of, 129 of exponential functions, 180
Cross product of vectors, 360 of hyperbolic functions, 341
Curvature, 194, 383 of inverse trigonometric functions;
radius of, 196, 383 166
Cycloid, 189 of logarithms, 180
differential of arc, 200 of sine and cosine, 41
Cylindrical coordinates, 373 of trigonometric functions, 163
derivatives, 385 Direction cosines, 357
triple integrals, 451 Directional derivative, 422
Cylindrical surface, area of, 278 Directrix, of an ellipse, 121
equation of, 368 of a parabola, 115
Discontinuous function, 17
Definite integral, 82 Displacement, 56
change of variable, 264 Distance between two points, 112
evaluation of, 83 Distributive law, 221
limits of, 83 Distributive operations, 38
properties of, 94 Divergence of series, 300
with variable limit, 95 Dot product, 353
Degree of an equation, 102 Double integrals, 444
DeMoivre^s theorem, 331 polar coordinates, 446
Dependent variable, 2
Derivative, 30 Eccentricity, of an ellipse, 120
directional, 422 of a hyperbola, 125
of a composite function, 404 Elementary functions, 243
of a vector, 227, 381 Ellipse, 119
of higher order, 43 radius of curvature, 201
partial, 398 Ellipsoid, 376
sign of, 46, 48 Epicycloid, 190
Determinants, 146 differential of arc, 200
 INDEX 501

yipicycloid, radius of curvature, 201 Homogeneous functions, 406

Equations, algebraic, 102 Hyperbola, 123
Cartesian, 102 equilateral, 126
first-degree, 105 rectangular, 126, 130
linear, 105 Hyperbolic functions, 339
parametric, 187 geometrical representation, 342
second-degree, 131 inverse, 344
Equilateral hyperbola, 126 Hyperboloid, of one sheet, 376
Errors, 42 of two sheets, 377
Euler’s formula, 335 Hypocycloid, 199
Euler’s theorem, 407
Exact differential, 419 Imaginary, 326
Exponential functions, 175, 335 Imaginary values, 134
differentiation of, 180 Implicit functions, 45
Extreme value theorem, 20 Increment, 31
approximate value of, 36
Focus, of an 119
ellipse, function of two variables, 400
of a hyperbola, 123 Independent variable, 2
of a parabola, 115 Indeterminate forms, 61
Frequency of vibration, 169 Infinite series, 300
Functional notation, 3 Infinite values, 135
Functions, 2 Infinitesimal, 36
continuous, 17 of higher order, 36
discontinuous, 17 Infinity, 16
elementary, 243 202
Initial line,
exponential, 175 Instantaneous axis of acceleration,
homogeneous, 406 349
hyperbolic, 339 Instantaneous center of rotation, 338
implicit, 45 Integrable, 83
periodic, 161 Integral test for convergence, 302
step, 17 Integrals, 73
trigonometric, 159 definite, 82
double, 444
Geometric series, 300 indefinite, 83
Gradient, 422 infinite values, 260
Graphs, 4 line, 425
algebraic equations, 132 multiple, 441
hyperbolic functions, 340 triple,450
inverse trigonometric functions, 165 with variable limit, 95
parametric equations, 187 Integration, 73
polar coordinates, 203 by change of variable, 253
trigonometric functions, 159 by parts, 257
Gudermannian, 350 by reduction formulas, 259
Gyration, radius of, 291 by substitution, 245
by trigonometric substitutions, 255
Harmonic motion, 168 constant of, 74
Helix, 384, 386 formulas of, 243, 244
Homogeneous equations, 154 multiple, 441
502 INDEX
Integration, numerical, 265 Minor axis, 120
of quadratic expressions, 248 Minors of a determinant, 150
of rational fractions, 250 Moment, of a force, 280, 364
of trigonometric functions, 246 of inertia, 290
repeated, 437 Motion, plane, 336
Intercept, of a curve, 133 simple harmonic, 168
of a straight 104
line, Multiple integral, 441
Intermediate value theorem, 20 evaluation of, 442
Interval, 2
Invariants, 130 Natural logarithm, 179
Inverse hyperbolic functions, 344 Newton’s method, 110
Inverse trigonometric functions, 164 Normal, to a curve, 108, 382
differentiation of, 166 to a surface, 410
Involute of a circle, 199 Normal component, 231
Irrational number, 1 Number, complex, 326
irrational, 1
Lemniscate, 205 rational, 1
Length 379
of arc, 190, real, 1
L’Hospital’s rule, 62 Numerical integration, 265
limit, of a function, 9
of a sequence, 6 Orthogonal, 111
of trigonometric functions, 12
limits, of an integral, 83 Pappus, theorems of, 288
properties of, 14 Parabola, 115
line integral, 425 Paraboloid, 378
independent of path, 427 Parametric equations, 187
linear equations, 105 Partial derivatives, 398
homogeneous, 154 geometrical representation of, 400
solution of, 152 higher order, 415
Logarithmic differentiation, 181 Partial fractions, 250
Logarithmic spiral, 211 Period, 161
Logarithms, 176, 338 Periodic functions, 161
differentiation of, 180 Phase, 169
natural, 179 Plane, equation of, 370
properties of, 180 tangent, 410
Plane motion, 336
Maclaurin’s series, 312 Plotting, 4
Major axis, 120 Polar coordinates, 202, 234
Matrix, 147 area, 212
Maxima and minima, 49, 53 components, 234
functions of two variables, 414 differential relations,209
method of finding, 53 double integrals, 446
second derivative test, 52 Polar equation, of a circle, 207
Mean value of a function, 279, 455 of a conic, 208
Mean value theorem, for a function, of a straight line, 207
59 Positive series, 301
for an integral, 95 Power series, 309
Minima, 49, 53 operations with, 317
 INDEX 603

Pressure, 91 Series, divergence, 300

Principal normal, 382 in powers of complex numbers,
Principal value, 165 334
Prismoid formula, 265 Maclaurin’s, 312
Product, cross, 360 of functions, 309
dot, 353 power, 309, 317
scalar, 353 Taylor’s, 312
triple scalar, 366 Taylor’s with remainder, 315
vector, 360 Simple harmonic motion, 168
Simpson’s rule, 267
Quadric surfaces, 376 Slope, of a curve, 29
of a line, 28
Radian, 12 Speed, 27, 229
Radius, of curvature, 196, 242 average, 27
in polar coordinates, 217 instantaneous, 27
of gyration, 291 Sphere, 376
Range of a variable, 2 Spherical coordinates, 411, 454
Rate of change, 56, 58 Spheroid, 376
Rates, related, 58 Spiral, logarithmic, 211
Ratio test for convergence, 306 of Archimedes, 204
Rational fraction, 250 Step function, 17
Rational number, 1 Straight line, equation of, 103, 372
Real number, 1 polar equation of, 207
Rectangular coordinates, 4, 351 Strophoid, 200
Rectifiable arc, 191 Surface, area of, 457
Reduction formulas, 259, 273 of revolution, 275, 374
Region of convergence, 309 quadric, 376
Related rates, 58 Symmetry, 135
Remainder in Taylor’s series, 315
Repeated integral, 437 Tangent line, 108, 382
Rolle’s theorem, 60 Tangent plane, 410
Rotation of axes, 129, 359 Tangential component, 231
right-handed, 351 Taylor’s series, 312
two variables, 418
functions of
Scalar, 218 with remainder, 315
Scalar product, 353 Torque, 280
Scale, 1 Total differential, 402
Second derivative, 43 Tractrix, 294
sign of, 48 Transformation of coordinates, 129
Semiaxes, of an ellipse, 120 Translation of axes, 129
of a hyperbola, 125 Transverse axis of a hyperbola, 125
Separation of variables, 182 Trigonometric functions, 159
Sequence, 6 differentiation of, 163
general convergence condition, 19 inverse, 164
increasing or decreasing, 20 Triple integrals, 450
Series, alternating, 307 cylindrical coordinates, 451
binomial, 313 spherical coordinates, 454
convergence, 300 Trochoid, 199
im INDEX
Unit vectors, 225, 234, 352, 385, 412 Vectors, scalar product of, 353
triple scalar product of, 366
Variables, change of, 253 unit, 225, 234, 352, 385, 412
dependent, 2 vector product of, 360
independent, 2 Velocity, 56, 229
separation of, 182 angular, 167, 387
Vector functions, 227 Vertex, of an ellipse, 120
Vectors, 218, 352 of a hyperbola, 125
addition of, 219 of a parabola, 116
angle between, 225 Volume of a solid of revolution, 87
components of, 225 with given area of section, 89
cross product of, 360
derivatives of, 227, 381 Wave length, 159
multiplication by scalar, 221 Witch, 199
products of three or more, 365 Work, 92, 356, 425