Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
&lt;yr
WO&lt;6DS
AND BAILEY
II
VOLUME
APR 19 1909
GIFT OF
R. Tracv
A COURSE IN
MATHEMATICS
FOR STUDENTS OF ENGINEERING AND APPLIED SCIENCE
BY
FREDERICK
AND
S.
WOODS
BAILEY
FREDERICK
H.
PKOFKSSOKS OF MATHEMATICS IN THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY
VOLUME
II
INTEGRAL CALCULUS FUNCTIONS OF SEVERAL VARIABLES, SPACE GEOMETRY DIFFERENTIAL EQUATIONS
GINN & COMPANY
BOSTON
.
NEW YORK CHICAGO LONDON
COPYRIGHT,
1909,
BY
H. BAILEY
FREDERICK
S.
WOODS AND FREDERICK
ALL RIGHTS RESERVED
79.3
gtftenaum GINN & COMPANY PROPRIETORS
BOSTON
U.S.A.
PKEFACE
This volume completes the plan of a course in mathematics outLined in the preface to the
first
volume.
The subject
treated in the
of integration of functions of a single variable is
first
eight chapters.
Emphasis
is
here laid upon the
fundamental processes, and the conception of the definite integral and its numerous applications are early introduced. after the
Only
student
well grounded in these matters are the more special methods of evaluating integrals discussed. In this the student s way
is
interest
is
early aroused in the use of the subject,
and he
is
drilled
in those processes
which occur
is
in subsequent practice.
A
new
feature of this part of the book
a chapter on simple differential
equations in close connection with integration and long before the formal study of differential
equations,,
With the ninth chapter
variables
of the
is
the study of functions of two or more
begun.
This
is
introduced and accompanied by the use
elements of solid analytic geometry, and the treatment of
partial differentiation
and
of multiple integrals is careful
and rea
sonably complete.
integrals.
texts, is
is the chapter on line This subject, though omitted from elementary generally needed by most engineering students in their later work.
A
special feature here
The
latter part of the
work
consists of chapters on series, the
complex number, and
differential equations
differential equations.
In the treatment of
many things properly belonging to an ex tended treatise on the subject are omitted in order to give the student a concise working knowledge of the types of equations which occur most often in practice.
iii
MJ2987JJO
iv
PREFACE
In conclusion the authors wish to renew their thanks to the
Massachusetts
members
of the
mathematical department
of the
Institute of Technology,
for
and
especially to Professor
H. W. Tyler,
continued helpful suggestion and criticism, and to extend
thanks to their former colleague, Professor
W. H. Eoever
more
of
Washington
University, for the construction of the
difficult
drawings particularly in space geometry.
MASSACHUSETTS INSTITUTE or TECHNOLOGY
February, 1909
CONTENTS
CHAPTER
ARTICLE
12. Order of infinitesimals
3.
,
I
INFINITESIMALS
AND DIFFERENTIALS
PAGE
1 4
Q
Fundamental theorems on
Differentials
infinitesimals
4.
56. Graphical representation
7.
8.
8
9
Formulas for
differentials
Differentials of higher orders
10
CHAPTER
9.
II
ELEMENTARY FORMULAS OF INTEGRATION
10.
Definition of integration Constant of integration
.12
12 13 14
17
.
11.
Fundamental formulas
12. Integral of u n
Integrals of trigonometric functions 1415. Integrals leading to the inverse trigonometric functions
13.
.
19
16. Integrals of exponential functions 17. Collected formulas
18.
19.
25
26
27
Integration by substitution
Integration
by
parts
30
32
20.
Possibility of integration
Problems
33
CHAPTER
21. Definition
III
DEFINITE INTEGRALS
.39
41
22. Graphical representation 23. Generalization 24. Properties of definite integrals 25. Evaluation of the definite integral 26.
43
45
by integration
....
47
Change
of limits
49
51
v
27. Integration
by parts
vi
ARTICLE
28. 29. 30.
Infinite limits Infinite integrand
CONTENTS
PAGE
52
53 54 55 59
61
The mean value
s
of a function
and Maclaurin s series 31. Taylor 3233. Operations with power series Problems
CHAPTER
34. 35. 36. 37. 38. 39.
IV
APPLICATIONS TO GEOMETRY
64 64
67 69
bases
Element of a definite integral Area of a plane curve in Cartesian coordinates Area of a plane curve in polar coordinates
Volume Volume of a solid with parallel The prismoidal formula
Length Length
of a solid of revolution
72
.74
40.
41.
42.
of a plane curve in rectangular coordinates of a plane curve in polar coordinates
....
75
77
The
differential of arc
78
79 SO
43.
Area of a surface of revolution Problems
CHAPTER V
44.
APPLICATIONS TO MECHANICS
86 86 88
Work
45. Attraction 46. Pressure
47. Center of gravity
90 92 94 96 98
99
48. Center of gravity of a plane curve 49. Center of gravity of a plane area 50. Center of gravity of a solid or a surface of revolution of con
stant density
51. Center of pressure
Problems
CHAPTER VI
52. Introduction
INTEGRATION OF RATIONAL FRACTIONS
103
.
5354. Separation into partial fractions 5557. Proof of the possibility of separation into partial fractions
58. Integration of rational fractions
.
109 113
117
Problems
CONTENTS
CHAPTER
ARTICLE
59. Rationalization
vii
VII
SPECIAL METHODS OF INTEGRATION
PAGE
119
60. Integrand containing fractional powers of a 61. Integrand containing fractional powers of a 62. containing integral powers of Vo
+ bx + bx n
4
119
.
.
Integrand
bx bx
.
+
.
x2 x2
.
.
.
63. Integrand containing integral powers of 64. Integration of trigonometric functions
65. Integrals of the 66.
Va +
.
.
.
.
.
.
.120 .121 .122 .123
123
forms f
sin"
xdx and
x
cos"
J cos
n
xdx
..
..
Integrals of the form
f
1 sin"
x dx
...
nx
J"ctn
.
.
.124
126 127
67. Integrals of the 68. Integrals of the 69. Integrals of the
forms C
tan"
xdx and
x dx and
I
dx
forms
J
sec"
csc n xdx
csc n xdx
forms ftan m o; sec n xdx and Cctnm x
.
128
129
70.
The
substitution
tan =
z
.
.
;
.
71. Algebraic reduction formulas 7273. Proof of reduction formulas 74. Trigonometric reduction formulas
75.
130
131
134 136
137
Use of tables Problems
of integrals
CHAPTER
76. Definitions
VIII
INTEGRATION OF SIMPLE DIFFEREN TIAL EQUATIONS
141
77.
The equation
rated
Mdx + Ndy =
when the
\
variables can be sepa
v
...
144
7879.
80.
81.
The homogeneous equation Mdx Ndy = The linear equation of the first order
Bernoulli
s
145
146
equation
148
11 9
.
.
82. Certain equations of the second order
Problems
.154
CHAPTER IX
83.
FUNCTIONS OF SEVERAL VARIABLES
158
158
.
.
Functions of more than one variable
Rectangular coordinates in space 8586. Graphical representation of a function of two variables
84. 87. Surfaces of revolution
.
159
...
168
viii
CONTENTS
PAGE
m
ARTICLE
88. Cylinders
IQQ
170 172
8990. Space curves
91.
t
Ruled surfaces Problems
.
f
174
CHAPTER X
92. Projection 93. Distance
PLANE AND STRAIGHT LINE
176
between two points
. .
178
.
94. Length of a space curve in rectangular coordinates 9596. Direction of a straight line 97. Direction of a space curve
179
180
182
98.
99. Direction of the
Angle between two straight lines normal to a plane
of a plane
.183
.
184
100.
Normal equation
.
185
186
186
Angle between two planes 102. Determination of the direction cosines of any straight line 103104. Equations of a straight line determined by two points 105. Equations of a tangent line to a space curve
101.
.
.
.187
189
106. Equations of a straight line in terms of its direction cosines and a known point upon it .190
107. Problems on the plane and the straight line 108. Change of coordinates
190
193
Problems
....
.195
CHAPTER XI
109. Partial derivatives
PARTIAL DIFFERENTIATION
198
110. Increment and total differential of a function of several variables
111. Derivative of /(a, y) when x 112. Tangent plane to a surface
200
and y are functions
of
t
.
.
.202
204
t
.
113. Derivatives of f(x, y} when x and y are functions of 114. Property of the total differential
115. Implicit functions
s
and
206
209
210 212
of
116117. Higher partial derivatives
118.
Higher derivatives of /(ar, y) when x and y are functions t, or of s and t
Problems
214
216
.
119. Differentiation of a definite integral
.
.
217
CONTENTS
CHAPTER
ARTICLE
120.
ix
XII
MULTIPLE INTEGRALS
PAGE
.
Double integral with constant limits
222
121. Graphical representation
225
226
122123. Double integral with variable limits 124. Computation of a double integral
125.
228
229
*.
Double integral in polar coordinates
126. Triple integrals
230
231
127. Cylindrical and polar coordinates 128. Elements of volume in cylindrical
and in polar coordinates
.
232
129.
Change
of coordinates
234
Problems
234
CHAPTER
XIII
APPLICATIONS OF MULTIPLE INTEGRALS
236 239 241
130132. Moment of inertia of a plane area 133134. Area bounded by a plane curve
135.
Area
of
any surface
136. Center of gravity 137138. Center of gravity of a plane area
139. Center of gravity of a solid
245
246
248 249
251
140. 141.
Volume
Moment
of inertia of a solid
142. Attraction
252 253
Problems
CHAPTER XIV
143. Definition 144.
LINE INTEGRALS AND EXACT DIFFERENTIALS
258
261
Fundamental theorem
145. Line integrals of the first kind, dy
146. Line integrals of the second kind,
=
dy

263
dx
^
267
x
147. Exact differentials
269
271
148149.
The
integrating factor
s
150151. Stokes
theorem
274 276
Problems
CHAPTER XV
INFINITE SERIES
279 280
281
152. Convergence 153. Comparison test for convergence 154. The ratio test for convergence
X
ARTICLE
CONTENTS
PAGE
283
155. Absolute convergence 156. The power series
284
s series
157. Maclaurin
158. Taylor
s
and Taylor
287
s series
for functions of several variables
....
288 290
159160. Fourier
161162.
163.
s series
The indeterminate form .
The indeterminate form
Problems
5
.
.
295 298 300
.
164. Other indeterminate forms
301
CHAPTER XVI
165. Graphical representation 166. Addition and subtraction
167. Multiplication
COMPLEX NUMBERS
304
305
306
.
and division
and evolution 169. Exponential and trigonometric functions
168. Involution
306 307
170.
The logarithmic function
310
311
171. Functions of a
complex variable in general 172. Conjugate functions
Problems
.
313
314
CHAPTER XVII
174. Solution
DIFFERENTIAL EQUATIONS OF THE FIRST ORDER
316
173. Introduction
by
series
318
175. Equations not of the first degree in the derivative 176. Equations solvable for_p
177. Equations solvable for y 178. Equations solvable for x
....
318
319
320
321 322
179181. Envelopes
325 182. Singular solutions 183. Orthogonal trajectories 327 184. Differential equation of the first order in three variables.
The
185.
integrable case
first
328 332 335
.
Two
differential equations of the first order in three variables
186. Differential equation of the The nonintegrable case
order in three variables.
Problems
336
CONTENTS
CHAPTER
ABTICLK
187. Definitions
.
. . .
xi
XVIII
THE LINEAR DIFFERENTIAL EQUATION
PA E
188.
The equation
of the first order with constant coefficients
.
341
189.
190.
The
operator
of the second order
The equation
by
with constant
coefficients
.
343 347
348
191. Solution
192.
partial fractions
constant coefficients general equation with 193. Solution by undetermined coefficients constant 194. Systems of linear differential equations with
The
.....
coeffi
.
350
cients
195.
The
linear differential equation with variable coefficients
series
354
8
196. Solution
by Problems
36
CHAPTER XIX
197. Introduction
PARTIAL DIFFERENTIAL EQUATIONS
363
.
198. Special forms of partial differential equations order 199. The linear partial differential equation of the first
304
8
200. Laplace 201. Laplace
s
s
equation in the plane equation in three dimensions
#*.
.
368
37
Problems
ANSWERS
INDKX
.
A COURSE
IN
MATHEMATICS
I
CHAPTER
INFINITESIMALS AND DIFFERENTIALS
which Order of infinitesimals. An infinitesimal is a variable mathe limit. The word "infinitesimal" in the approaches zero as a
1.
minute." matical sense must not be considered as meaning "very the size of an atom of matter is not a mathematical For example, definite. That infinitesimal, since that size is regarded as perfectly it is essential that it can be a quantity should be infinitesimal made smaller than any assigned quantity. In operating with in
finitesimals
of them as not, however, necessary to think are finite quantities and microscopic or of negligible size. They division, etc., like finite all the laws of
it is
obey
multiplication,
It is generally the last step in a problem involving infinitesimals to determine the limit of some expression, usually
quantities.
a quotient or a sum,
when
the infinitesimals contained in
it
approach
Ex.
1.
zero.
To
find the velocity of a
,
moving body
(I,
106)*
it is
necessary to
find the limit of the quotient
A
as As and Ai approach zero as a limit, where
As
are infinitesimals and the space traversed in the time A*. Here As and the velocity is the limit of the quotient of two infinitesimals. Ex. 2. To find the area of a circle it is customary to inscribe in the into n tri circle a regular polygon of n sides and to divide the polygon When radii of the circle drawn to the vertices of the polygon. angles by
is is
M
n
the area of the circle
increased without limit, the area of each triangle is infinitesimal, and of is the limit of the sum of the infinitesimal areas
triangles.
I refer to
an indefinitely great number of
* References preceded
1
by
Vol.
I.
2
INFINITESIMALS AND DIFFERENTIALS
In a mathematical discussion involving infinitesimals there will usually be two or more infinitesimals so related that as one ap proaches zero the others do also. Any two of these may be coinpared by determining the limit of their the following definitions
:
ratio.
We have accordingly
the limit
Two infinitesimals are of the same order ratio is a finite quantity not zero.
when
of their
An
infinitesimal
ft is
of higher order than an infinitesimal a rt
is zero.
if the limit of the ratio
Ex.
3.
Let
ft
=
sin
a:
and 7
.
=
1
cos a,
,.
where a
is
an infinitesimal angle.
Then
T Lim
.
ft
a
r=
Lim
.
.
sin
a
a
1
=
1,
T Lim 7 = T Lim a
cos a
a
=
0.
(I,
151)
Hence
ft is
of the
same order as
1) be an arc of a a with center at 0, the chord the side of an inscribed regular polygon of n sides, and CD the side of a regular circumscribed polygon. Also let
a, and 7 is of higher order. Ex. 4. Let the arc AB (fig.
circle of radius
AB
a=
ft
7
= =
n
the area of the triangle OB, the area of the triangle COD, the area of the trapezoid ABDC.
is
A
Then
FIG.
1
if
infinitesimal.
To compute
indefinitely increased, a, ft, their values,
and 7 are
draw
OE
perpendicular to
AB
n
and CD.
Then
,
,
n
OE = acos,
#D=atan.
From
this it follows that

=
2
a2
sin
n
cos
n

=
(
sin
EF = (EB + FD) (OF \
OE)
/
i.
 +
n
tan
) (
1

cos 
)
=
COS
7T 
n/
\
n]
n
OliDEK OF INFINITESIMALS
R
Hence
Liui
a
= Lim
1
=
1,
cos2
TT
n
Lim ^ = Lim a
Therefore
2.
/3
cos 2

=
0.
O TT
n
is
is
of the
same order as
cr,
and y
of higher order.
of the ratio of
Particular importance attaches to the case in which the limit two infinitesimals is unity. Suppose, for example, that
a
Then, by the definition of a limit
(I,
53),
s+
where
e
approaches zero as a approaches
zero.
Hence
= a + ae.
Now
Lim
the term ae
is
an infinitesimal
that
j3
of higher order
than
a, for
= Lim e =
0, so
and a
differ
by an infinitesimal
of
higher order than each of them.
Conversely, let @ and a differ by an infinitesimal of higher order than either of them, i.e. let
/3
= a + 7, =
0.
where, by hypothesis,
7 Lim 
Then
We
and
"
infinitesimals /3
ft
have accordingly proved that the two statements, "Two and a differ by an infinitesimal of higher order"
Lim 1.
=1
"
are equivalent.
Ex.
order
The
151).
infinitesimals
a and
sin
a
differ
by an
infinitesimal of higher
(I,
Ex. 2. The areas of the triangles and 1) differ (Ex. 4, infinitesimal of higher order, namely, the area of the trapezoid ABDC.
AOB
COD
by an
4
3.
INFINITESIMALS AND DIFFEKENTIALS
Fundamental theorems on
infinitesimals.
There are two im
:
portant problems which arise in the use of infinitesimals, namely
1. quotient problem infinitesimals as each approaches zero.
:
A
to find the limit of the quotient of
two
of
2.
A sum
problem
:
to find the limit of the
sum
of a
number
infinitesimals as the
number
increases without limit
and each in
1
finitesimal approaches zero.
Each
there
is
of these
problems has been illustrated in
:
;
for
each
a fundamental theorem as follows
1. If the quotient of two infinitesimals lias a limit, that limit is unaltered by replacing each infinitesimal by another which differs
from
2.
by an infinitesimal of higher order. sum of n positive infinitesimals has a limit, as n increases indefinitely, that limit is unaltered by replacing each infinitesimal
it
If the
by another which differs
from
it
by an infinitesimal of higher order.
To prove theorem 1, let a and ft be two infinitesimals and let a and ft respectively by a, and ft, be two others which differ from infinitesimals of higher order. Then we have ( 2)
Lim
!
= 1,
l
Lim ~ =
Pi
ft
1,
whence
where
e1
a
= a + e av
l
= ft,+ e^ftv
ft
and
e2
approach zero as a and
approach
zero.
Then
a
Therefore
ft Lim 
al +
e1
1
a,
1+ e,
Lim
a
1 eA = Lim /ft, r+  )= Lim \a, l+ej

aL
+e =  Lim ft, l+e a,
l
2
1
Ex.
1.
Since the sine of an
angle differs
3
,
from the angle by an infinitesimal

of higher order,
Lim
To show
this directly,
sm 
sin 2
3 a _, = Lim 3 a = 2 2a a
a_ ~~ sin 2 a
sin 3
Therefore
we may write 4 sin 3 o: _ 3 sin a ~ 2 sin a cos a sin 3 a Lim sin 2 a
3
_
a
2 sin 2
a
2 cos 3 2
cos a
FUNDAMENTAL THEOREMS ON INFINITESIMALS
To prove theorem
2, let
5
, a n be n positive infini increases indefinitely, each of the infinitesimals approaches zero and their sum approaches a limit and let /3 V /3 2 /3 3 /3n be n other infinitesimals such that
a v av as
,
tesimals so related that, as
n
;
,
,
,
Lim^ = l, Lim^ = l,
*i
*
Lim
=
3
as
1,
,
Lim
^=
an
1.
We
wish to show that
Now (2)
i=*i+ii&gt;
/32
=tf 2 +e 2 tf 2
38
,
ft
+&gt;
whence
Lim (
+
a
+
8
+
.
=
Lim(a 1
+ /3j +a +a +
. .
2
s
+a
2
n)
+ Lim (e ^ + e
1 1
tf
2
+
is
s
as +
+e
n
a n ).
Now let 7 be a positive quantity which numerical value of the quantities e v e 2 e 8
,
,
equal to the largest
,
eB
.
Then
whence
since
7^
is
=s
l
7,
by hypothesis a l
positive
and may multiply the inequality
without change of
sign.
Similarly,
whence
6
INFINITESIMALS AND DIFFERENTIALS
As n
increases indefinitely,
!+,+ *,+
&lt;
+
.
approaches a
finite limit
zero. by hypothesis, and 7 approaches + en a n ) = lim (6^ + e 2 a 2 + e 3 a s + Therefore
0,
and hence
Lim
(&+
is
2
+
3
+
+ ft)
=
The theorem
Ex.
is 2.
thus proved for positive infinitesimals.
The
limit of the
sum
of the areas of
n
triangles such as
AOB
as
(fig. 1)
the same as the limit of the when n is indefinitely increased.
sum
of the areas of
n triangles such
COD,
The theorem
is
also true
if
the infinitesimals are
is
all
negative,
simply to change since to change the sign of each infinitesimal are not all the sign of the limit of the sum. If the infinitesimals
of the
same
3.
sign,
however, the theorem
is
not necessarily true.
Ex.
Let
ai
= JL, Vn
a
.
01
1
= 4=i
Vn
a
2
fl*
= 4=i
Vn
0,
&lt;*4
=
etc
7=1 Vn
&gt;
o
n
,
= a2 +
3

n
,
=
&lt;r
+n
or
1
ft
=
4
+

etc

n
Then
&lt;*i
+
&lt;?2
+
+
+
+
and
h
^ =
Pn
0,
=
Vn
Or
01
+
^2
&
;
+
=
1,
=1+
,
according as n
is
even or odd
Lim
Lim
n= oo
4.
(&lt;*i
+ a2 + a s + +
02
1
&lt;*)
= 0,
L
is
(3i
+
03
+
of
+
)
Differentials.
The process
differentiation
an
illus
tration of
3. y=f(x) is a the quotient problem of and continuous function of x which has the derivative f(x), A and Ay are the corresponding infinitesimal increments of
For
if
x and
y,
then,
by
definition,
T
^ Lim A?/ ?=f(x).
.
*,.
I
/i\ 1)
AX= o A^?
DIFFERENTIALS
It appears
7
from
(1) that
A# and Ay
which
are infinitesimals of the
is
same
(1)
order, except in the cases in
f (x)
or
GO.
Moreover,
may
be written
A
where Lim
e
=
Ax0
0,
and hence
Ay=/()A + eA.
It appears, then, that
(2)
differs from Ay by an infinitesimal and f(x)&x may therefore be used in of higher order than A#, and sums. place of Ay in problems involving limits of quotients The quantity f(x)&x is called the differential of y, and is rep
f (x)&x
resented by the symbol dy.
Accordingly
(3)
dy=f(x)*x.
Now
in the special case in
which y
= x,
formula
(3)
reduces to
(4)
dx
= A#.
Hence we may
write (3) as
dy=f(x)dx.
To sum
equal
to
(5)
this
the
up
:
The
differential
of
the
independent variable
is
increment of the variable; the differential of
equal
the
function
to the differential of the independent variable multiplied ly the derivative of the function, and differs from the increment of the function ly an infinitesimal of higher order. is
From
this point of
view the derivative
is
called the differential
coefficient.
The use
of differentials instead of
increments
is justified
by the
fundamental theorems
It is to be
of
3 in
many problems which
are eventu
ally to involve the limit of a quotient or the limit of a
sum.
ject to all the
emphasized that dx and dy are finite quantities, sub laws governing such quantities, and are not to be
(5)
thought of as exceedingly minute. Consequently both sides of may be divided by dx, with the result
8
INFINITESIMALS AND DIFFERENTIALS
That
is,
the derivative is the quotient of two differentials. the notation already chosen for the derivative. explains
Ex.
1.
This
Let y
=
Ay
x3
.
We may
On
increase x
by an increment Ax equal
to dx.
2
Then
=
(x
+
dx)
3

x
3
=
3 x dx
&gt;2
+
3 x (dx)
+
(dx)
3.
the other hand, by definition,
dy
=
3 x 2 dx.
3 which is 2 It appears that Ay and dy differ by the expression 3 x (dx) + (dx) an infinitesimal of higher order than dx. Ex. 2. If a volume v of a perfect gas at a constant temperature is under
,
the pressure p, then v
=
k 
,
where k
dp.
is
a constant.
Now
let the pressure
be
gas
increased by an is then the increment
amount Ap =
k k
The actual change
in the
volume of the
_ ~
kdp
_
Mp /
p2
I
1
p + dp
The
differential of v
is,
P
however,
P(p + dp)~
1
dp
V
dv
P
=
^
p2
,
differs from Av by an infinitesimal of higher order. The differential dv may, accordingly, be used in place of Au in problems which involve the limit of quotients or sums of this and other infinitesimals.
which
Graphical representation. The distinction between the increment
5.
and the
differential
may
be
illus
trated graphically as follows:
Let the function be
represented
(fig. 2).
by
the curve
y=f(x]
Let
P(x, y) be any point of the curve, and Q (x f A#, y + Ay) a neighboring
.
x
point.
Draw
the lines
PR
and
RQ
FIG. 2
parallel to the axes, the chord PQ, and the tangent PT. Then
and
dy =f(x) dx
=
(tan
RPT) PR = R T.
GRAPHICAL REPRESENTATION
Hence the increment and the
an infinitesimal
differential differ
9
by QT. That
QT is
:
of higher order
than
PR may
.
be shown as follows
_ lm QT
= Llm RTRQ
___
=
RT LunRQ T
= tan RPT  Lim (tan RPQ) = 0.
6.
The formula
dy=f (x^dx
has been obtained on the hypothesis that x
variable and
is
(I)
the independent
y = f(x).
Consider
now
the case
y=f(x), where x
substitution
=
&lt;j&gt;(t)
and
t
is
the
independent variable.
By
r
we have
/W dy=F (t)dt.
F (t) = f(x)
r
.
&lt;f&gt;
Then, by
(1),
(2)
(t).
But by
I,
96, (7),
Substituting in
(2),
we have
(3)
But since from 1)
&lt;
t
is
the independent variable and x
=
&lt;
(t),
we
have,
Substituting in
(3),
we have
dy
= f (x)dx.
Therefore
(1)
is
This
is
the same form as
is
(1).
always true
whether x
7.
the independent variable or not.
for differentials.
Formulas
By
virtue of the results of the
preceding article, from the formulas for the corresponding derivatives. n For example, since the derivative of u with respect to
n
the formulas for differentials can be derived
u
is
nu*~ l
,
d(u
}
= nu
n
~l
du,
whether u be an independent variable or the function of another
variable.
10
INFINITESIMALS AND DIFFERENTIALS
we
derive the following
Proceeding as in the above example, formulas
:
8.
"
i^
Differentials of higher orders.
By
definition
(1)
dy=f(x)dx.
But dy
is
itself
a function of
differential,
which
n
will be denoted
in general, d(d ~^y) d n y. called differentials of the second, the third,
respectively.
=
and may, accordingly, have a 2 2 by d y i.e. d (dy) = d y, and, 2 The differentials d y, d 3 y, d*y are
x,
;
,
,
the %th order
When x is the independent variable, its increment dx may be taken as independent of x. With this assumption, applying our definition of a differential to (1), we have
d*y
= d[f(x) dx] =
[f(x) dx] dx
=f"(x)
2
dx2
.
,
where dx2
is
written for convenience instead of (dx)
In like manner,
d*y=f&gt;
(x)dx\
n
and, in general,
dny
=f
(n
\x)dx
.
DIFFERENTIALS OF HIGHER OKDEKS
The reason
the third,
,
11
for the notation
used for derivatives of the second,
the nth orders
is
now
apparent.
not the independent variable, the expressions for the If be assumed as differentials become more complex, since dx cannot
x
is
independent of
x.
We
have, then,
=/"(*)
tfy
= d \f(x) dx]
dx
dx*
+f(x) d\
2 i
&lt;"W^
&gt;
(by
(4),
7)
where,
if
x
=
&lt;(*),
= * (0^
and
&x =
bein g the in ~
that the formula for the dependent variable. It appears, then, second differential is not the same when written in terms of the when written in terms of a function independent variable as it is This is true of all differentials except the first of that variable.
(see
6).
first,
Hence the higher differentials are not as convenient as the and the student is advised to avoid their use at present.
CHAPTEE
II
ELEMENTARY FORMULAS OF INTEGRATION
9.
Definition of integration.
its
The process of finding a function
is called integration.
This is evi dently the same as the process of finding a function when its derivative is known, called integration in I, 110. In that place integration was performed by rewriting the deriv ative in such a manner that we could recognize, by the formulas of the function of which it is the derivative. But this differentiation,
when
differential is
known
method can be applied only
complex cases
it is
in the simpler cases.
For the more
necessary to have formulas of integration, which can evidently be derived from the formulas for differentials (7).
Using the symbolic denote integration,
definition that
if
it is
evident from the
f(x) dx
= dF(x},
= F(x).
then
Cf(x) dx
is said to be under the sign of integration, called the integrand. and/() F(x} is called the integral of f(x) dx. 10. Constant of integration. Two functions which differ
is
The expression f(x) dx
only
by a constant have the same derivative and hence the same differ ential and conversely, if two functions have the same differential,
;
they
differ
if
only by a constant
(I,
110
;
II,
30).
(1)
&lt;7],
Hence
it
follows that
= dF(x), f(x) dx = d [F(x] +
f(x) dx
(2)
where
any constant. Eewriting (1) and (2)
is
C
as formulas of integration,
we have
(3)
x)dx
and
= F(x)
= F(x) +
C.
dx
Jf(x)
(4)
FUNDAMENTAL FORMULAS
It is evident that (3) is
13
and hence that
but a special case of
(4),
all integrals ought to be written in the latter form. The constant C is called the constant of integration and is independent of the form
of the integrand. For the sake of brevity it will be omitted from the formulas of integration, but must be added in all integrals evaluated by means of them. As noted in I, 110, its value is
determined by the special conditions of the problem in which the
integral occurs. 11. Fundamental formulas.
The two formulas
C c du = c Cdu
and
(1)
\(du
+ dv + dw+&gt;) = Cdu + idv + idw +

(2)
are of fundamental importance, one or both of
them being used
in
the course of almost every integration. as follows
:
Stated in words, they are
(1)
A
constant factor
may
"be
changed from one side of the sign
finite
of integration
(2)
to the other.
The integral of a sum of a
the integrals
(1),
number of functions
is the
sum of
of the separate functions.
note that since
To prove
we
I
cdu
= d(cu),
I
it
follows that
c
du
=
I
d (cu)
= cu = c
du.
In like manner, to prove
(2), since
du
+ dv + dw +
+
dv
= d (u + v + iv +
=
I
we have
I
(du
+ dw H
)
d(u
f
v
+w
\
=u+v+w+ =
The
/
du
+
I
dv
+
I
dw +
application of these formulas
is
illustrated in the following
articles.
14
ELEMENTARY FORMULAS OF INTEGRATION
12. Integral of
u".
Since
d
it
(%"
)
= mum ~
m~l
l
du,
.
follows that
/
mu
du
= um
But by
11(1),
if
m=0,
I
m
Placing
m = n f 1,
/u
u m ~^du
um
~l
du
=
I
mu m ~
1
du
= um
,
=
m
we have
u du
"
the formula
= ~T^
(1)
for all values of
If
n
=
1,
n except n 1. the differential under the integral sign in (1) becomes
=
du
&gt;
which
is
recognized as dQogu).
I

Therefore
^
= l og u
.
(2)
In applying these formulas, the problem is to choose for u some function of x which will bring the given integral, if possible, under one of the formulas. The form of the integrand often suggests the
function of x which should be chosen for u.
Ex.
1.
Find the value of /Yax2
J \
+
bx
Applying
11 (2)
and then
11 (1),
 f \dx. x x2/ we have
+
f (*+*+*+$+ fMt+f^+ftp = + afxtdx bfxdx + ^
c
The first, the second, and the fourth of these integrals may be evaluated by formula (1), and the third by formula (2), where u = x, the results being
respectively
ax 3 bx 2 , and
, ,
32
x
c logx.
Therefore
2 flax + \
bx
+

"
x
+ \dx =
x
/
ax^
32
+ ~bx2 +
c
logx

x
f
C.
INTEGliAL OF U*
Ex.
2.
15
Find the value of f (x 2
+
2)xdx.
If the factors of the integrand are multiplied together,
we have
f(x
2
f
2)xdx=
in
which may be evaluated by the same method as that used
4 being \ x
Ex.
1,
the result
+
x2
f
C.
Or,
we may
let
x2
+
2

w,
whence 2 x ox
=
du, so that x dx
=
dw.
Hence
f (x 2
f
2)xdx
r=
fJ
=
2
J (x
f 2)
2
+
C.
Instead of actually writing out the integral in terms of w, x dx = d (x 2 + 2), and proceed as follows
we may
note that
%
:
(V + 2 )^
d:c
=
f
(
x&lt;2
+
2)
2 ld(x +
2)
=
Comparing the two values
2
I (x
+
2
2)
+
C.
of the integral found by the two see that they differ only by the constant unity, integration, be made a part of the constant of integration.
we
methods of which may
Ex.
3.
Find the value of
2
J(ax
+
2 6x) 3 (ax
+
b)dx.
Let ax 2 + 2 bx
= u. Then (2 ax + 2 6) dx = du, so that (ax + 6) dx =
2 6x) 3 (ax
f b)
1 du.
Hence
f (ax2 +
dx
=
it
3
du
=
Or, the last part of the
2
 (ax
2
+ 2 6x)* f C.
as follows
:
work may be arranged
f(ax
+
2 6x) 3 (ax
+
6)
dx
=
J(ccx
2
+
2
2 6x) 8
d (ax 2
+
+
2 6x)
2 bx)
= =
f(ax
J (ax
2
+
2 6x) 8 d (ax 2
+
2 6x) 4
+
C.
16
Ex.
ELEMENTARY FORMULAS OF INTEGRATION
4.
Find the value of
2 3, let ax
J
/
ax 2
f
2 bx
As
in
Ex.
+2 bx=u. Then
r 4 (ax + b) dx J ax 2 + 2 bx
(2
ax+2b)dx=du,
"
so that (ax + b)dx=^du.
Hence
_ ~
r2du_rdu
J ~^T
2 log
J
~u
= =
u
+C
f
2 log (ax 2
2 bx)
+C
Or
r4(ax + b)dx_ r 2 d (ax2 + 2 bx) J ax* + 2bx ~J ax 2 4 2 6x + bx
_
r d (ax 2 + 2 bx) J ax 2 + 2 bx
2 log (ax 2
=
+
2 bx)
+C
Ex. Let
5.
Find the value of C(e nx
+
2
b)
e ax dx.
e ax
+
b
=
u.
Then
e nx
a dx
=
aw.
Hence
cfct
:
a
= (e^ + oa
Or,
6)3
+
2
C.
J
f (e* +
2
6)
e
a:
dx
= f J
i(e a
+
6)
d(e^ +
6)
Ex.
6.
Find the value of f
Let tan (ax
+
6)
+
c
=
u.
2
+ 6) + 6) + c Then sec 2 (ax + 6) a cfx =
sec 2 (ax
d"x
tan (ax
d*w.
Hence
(ax
+
b)dx
_
_ ~
rl du
/sec (ax tan
f 6)
+c
J a
1
it
/"(fat
a J u
INTEGRALS OF TRIGONOMETRIC FUNCTIONS
c
17
=
/

log [tan (ax
+
6)
+
&)
c]
+ C.
c]
Or
sec2(ox
+
b)dx
J
UixT&Tc
= rl
J
d [tan (ax d [tan (ax
tan (ax
+
+
a"
_ ~
1
/
a J

+ &) + c] + 6) + c
6)
=
The student
is
a
log [tan (ax
+
+
c]
+
C.
advised to use more and more the second method,
illustrated in the preceding problems, as
he acquires
facility in
integration. 13. Integrals
the trigonometric functions. By rewriting func formulas ( 7) for the differentiation of the trigonometric tions. we derive the formulas
of
cos
u du u du
= sin u =
t
(1)
sin
cos u,
(2)
2
^ du
u du
= tan u, =
ctn u,
(3)
esc
2
(4)
sec
u tan u du
= sec u, =
esc u.
(5)
esc
u ctn u du
(6)
In addition to the above are the four following formulas
I
:
tan
u du u du
= log sec u,
= log sin u,
rj
(7)
I
ctn
(8)
fsec u du
= log (sec u + tan u) = log tan
= log (esc u
ctn u)
+
7j
)
(
9)
jcsc
u du
= log tan 
(10)
18
ELEMENTARY FORMULAS OF INTEGRATION
(7) we note Then
To derive
that tan
u
=

and that
sin
u du
= d(cosu).
I
tan
u du
=
d (cos u)
/
cosu
In like manner,
/C
ctn
= log cos u = log sec u.
cos
:
udu =
I
J
smu
u du  =
log sin u.
Direct proofs of (9) and (10) are given in 68. At present they may be verified by differentiation. For example, (9) is evidently true since d log (sec u + tan u) = sec u du.
The second form
of the integral
may
u
be found by making a
to tan
(
trigonometric transformation of sec
+ tan u
\4
+ )
2/
Formula
Ex.
1.
(10)
may
be treated in the same manner.
Find the value of f cos (ax 2
Let ax2
+
bx
=
u.
Then
2
(2
ax
+ bx) (2 ax + + b)dx = du. +
b)dx
b) dx.
Therefore
j^cos (ax
+
bx) (2
ax
= =
Jcos (ax
sin (ax
2
2
+
bx)
d (ax 2
C.
+
bx)
+
bx)
+
Ex.
Let
2.
Find the value of Tsec (e^2
f 6)
tan
(e
**
+
aa? x 6) e
dx.
e a3?
+b=
u.
Then
ax2
(e
e ay?
2axdx =
(e ay?
du.
aa? b) e
Therefore
/sec
+
1
b)
tan
+
+
b)
xdx
(e
=
=
r sec (ef** / 2a J
b)
tan
ax*
+b)d (e ax +
*
6)
2a
sec (ea ^
+
+ C.
The integral may often be brought under one or more of the fundamental formulas by a trigonometric transformation of the
integrand.
INVERSE TRIGONOMETRIC FUNCTIONS
Ex.
3.
19
Find the value of
cos2
f
cos z xdx.
Since
x
=
(1 f cos 2 x),
we have
f cos 2 xdx
=
% f(l
+
cos2x)dx
*
+ =
Ex.
If
4.
f cos2x d(2x)
x
\x
+
I
sin 2
+
C.
Find the value of
let sec 6 2
fsec 6

2xdx.
we
x
=
x,
sec 4 2 x
2 tan 2 2 x
+
tan 4 2
and place sec 4 2 x the original integral becomes
2 tan 2 2 x
sec 2
sec 2 2 x,
=
(1
+
tan 2 2 x) 2
=
f (1 +
Place
+
tan 4 2 x) sec 2 2 x dx.
du.
simplifying,
tan2x = w. Then we have
2x 2dx =
i
Making
2 u2
this substitution
and
f sec 6 2 x dx
f (1 +
+
u4 ) du
= =
14.
l(U
+
Jtt&gt;+t*)+C
^ tan 2 x
+^
tan 3 2 x
+ TL tan5 2 x + C.
From
Integrals leading to the inverse trigonometric functions. the formulas (7) for the differentiation of the inverse trig
onometric functions
of integration
:
we
du
derive the following corresponding formulas
fVlt
du
du
= sm _, u
.
or
cos
_T
u,
= tan~
con"
1
w
/
or
1
ctn"
^,
i 1
1
These formulas are
replaced
much more
Making
serviceable, however,
if
u
is
by

(a&gt;0).
this
substitution
and evident
Cb
reductions,
we have
2
as our required formulas:
du
/Va / +u
or
= sm _,u
.
,u
or
cos"
&gt;
(1)
^tan" 2
1
a
or
1
eta"
*
a
a
a
(2)
(3)
20
ELEMENTARY FORMULAS OF INTEGRATION
I,
Referring to
in the
first
153,
we
see that in (1)
1
sin"
fit
 must
ct
be taken
in
or the fourth quadrant
and that
1
cos"
 must be taken
1
the
esc"
first
r\i
or the second quadrant.
In like manner in
(3)
sec"
and
1
must be taken
in the first or the third quadrant.
(1) differ
It is to
be noted that the two results in
only by a
is
constant.
first
For
let
1
sin"
a
=
d&gt;
and
1
cos"
~
a
^r,
where
&lt;h
in the
or the fourth quadrant
and
i/r
is
in the first or the second
quadrant.
Then
sin
d&gt;
=
Therefore cos
(&lt;
+
^r)
= 0,
whence
&lt;
+
= (2 k + 1)
&gt;
where
k
is
any
integer or zero.
&lt;/&gt;
Hence
= (2 k + 1)~ i 1
or
sm"
07  = (2 &
a
+ 1) ^
x
i 1
cos"
.
2
a
Similarly, the results in (2) or (3) constants.
Ex.
may
be shown to differ by
1.
Find the value
dx
of
V94x
Letting 2 x
/
I
2
=
it,
we have du
1
/
I
2 dx,
and
or
1  cos .2x \C.
1
J
dx ^^^^ =  J 2 
=
V9 (2x) 2
4x2
 sin 1 2
1..2x h C 
Ex.
If
2.
Find the value of
x
let
V3 x 2 
4
we
V3 x =
r
JI
M, then du
= V3 dx,
and we may write
dx r J xV3x 2
I
== = 4
d(V3x)
_
+ C orc S c
+
C.
V3
INVERSE TRIGONOMETRIC FUNCTIONS
Ex.
3.
21
Find the value of
dx
\
Since
V4 x 
x2
= V4 
(x
V4 x  x 2  2) 2 we have
,
/axV4
x
_
x
2
r
*
ax
_/*(*
2)
2
*
(x
V4 
^
or
(x
V4 

*)
=
sini
?
+C

cosi 

+
C.
Ex.
4.
Find the value of
fractions
f
/
2 x2
4
3x
4
5
To avoid
and
radicals,
we
place
dx
8dx
16x 2 424x
=
4dx
Therefore
J 2x 2
f
^
+ 3x46
=
r
2
"J
4dx
(4x
=
4
/
2
d(4x
(4
+ 3)
2
+
n 1
2
3)
31
4
J
2
tan
4x43
C
or
V31
The methods used
in Exs. 3
V31
V31
2
x
+
3)
4
31
1
ctn
,4x43 \/31
=
\C.
and 4 are often of value in dealing with
functions involving ax 2
+
bx
+
c.
Ex.
5.
Find the value of C J

5
4
4 x4
i.e.
Separating the integrand into two fractions,
544x*
and using
11 (2)
5
we have
f(x*
5
/ ax 4 x) ax 4x)dx _ r x 3 dx + 4x4 ~J 544x4
r
/
xdx
5
J
But
and
J
dx
+ 4x 4
1
.
r x 3 dx
J5 +
J
5
4
I
=
1
/&gt;16x
3
4X4
4x
16J5 + 4x 4
r
/
=
log (5
4
4 x4 )
:
16
r_xdx_^l 4
=
4xdx
4 J 5
1
+
(2
x2 ) 2
tan4
.2x 2 1
1
or
Cto"
1
.2x 2
V5
log(5
V5
4
4
V5
"
V5
Therefore
(x8
f J 5
+ 4X4
+ x)(fa =
10
1
+ 4x4) + L= tani
V5
1
**
V5
2
a 2
+ C,
or
log(5
+ 4x4 ) JctnilfC. V5
22
ELEMENTARY FORMULAS OF INTEGRATION
15. Closely resembling formulas (1) and (2) of the last article form of the integrand are the following two formulas
:
in the
and
C
I
J
To derive
2
u2
du  = 7^1og u
1
,
a
; 2
2 a
*u + a
a
&lt;f&gt;.
&lt;&gt;
(1)
1
a
u
or
2 a
r
log& a
*fat
+u
a
sec*&lt;f&gt;d&lt;f&gt;,
(2)
(1)
we
&lt;.
place
u
= a tan
Then
=
and
+
a2
=
a sec
Therefore
du
Vtr
,
.
= C sec
a
2
+
J/
= log (sec
kg
+ tan +
2
&lt;)
(by
(9),
1 3)
= log (u + Vw + a
But
log
2 )
log
a.
a
is
a constant,
the formula of
and may accordingly be omitted from integration. If retained, it would affect the con
of the fact that the fraction
stant of integration only.
Formula (2) is derived by means
be separated into two fractions, the denominators of which a and u + a i.e. are respectively u
may
;
if
a
2
2
a\u
a
l
u
+
l
(53)
\du
\
\
,
Then
r
du
u
2
r,
a"
J

rf ==.(/2 a
l
J \u a l / r du a 2 a \J u
[
u
+ a/
u
r du
J
)
+ a/
= 5
lo S
(
u
a

2 a
,
u
[Qff
u+
a
LOGARITHMIC FUNCTIONS
The second form
of (2) is derived
23
by noting that
.
/du = J au =
I
C
du 
u
a
log(a
u).
The two
results differ only
and hence
and log(
1) is
,
,
,
log b
a
by a constant,
i
for
a
+
u u
=
,
1
u u
a
f
)
a
a
a
+
u u
= log
(
1)
+ log
,
*u +
u
a
a
:
a constant complex quantity which can be ex
(
pressed in terms of i
170).
/: 3x 2 + 4x
To
avoid fractions
we
multiply both numerator and denominator by
V.
Then
g
V3 x 2 +
4x
= ^/3jfa_ = Vo x 2 + 12 x V(3x +
3 dx, and
^&lt;fa
2)
2
4
Letting 3 x
+
2
=
M,
we have du =
dx
=
3x + 4x
2
r 1 3dx J /iJfV3 V(3x + 2) 2 2
4
= ^log(3x +
V3
Ex.
2.
+ V(3a
Find the value of
/*
_*L_.
8,
Multiplying the numerator and the denominator by
we have
J 2x 2 +
x15~
=
J
1
(4
x
+
2
I)
(11)2
11
f
_ log
Iog
(4 x 1
+ !)_ Z_J
5
c.
This
may
be reduced to
11
2x ~ 2x +
6
+ C,
or
11
2a;
Iog
x
and the term
log 2, being independent of x, only affect the value of the constant of integration.

^
~5 +3
11
log 2
it
+ C,
will
may
be omitted, as
24
ELEMENTARY FORMULAS OF INTEGRATION
If in
the formulas for the differentiation of
1
sinh"
1
^,
cosh"
^,
and
1
tanh"
^
(I,
161)
we
replace
u by
ct
&gt;
they become
d dx
.
,
,
Sinn"
u a
=
du dx
,
and
d
,
tanli"
,u 
=
a
dx
ax
a
u
The corresponding formulas
du
of integration are evidently
= smlr
.
,
i 1
*
u
=
v&lt;
"
.
cosh
_,u
:
and
r du = J d if a
/
/
.,"
2
tanh
1

These forms of the integrals are often of advantage in problems where the resulting equation has to be solved for the value of u in
terms
of the other quantities in the equation.
By
formulas
(2),
14,
and
+lx + /dxax2
c
(2) of this article
we
can find the value
can also find the
We
c
value of any integral of the form
J
/
^+bxUL ax +
2
as
shown
in Ex.
3.
/P
(x)
dx
?
ax?+lx
+c
where P(x)
is
a
is polynomial, can always be found, since by division the integrand to a polynomial plus a fraction of the form just mentioned. equal
INTEGKALS OF EXPONENTIAL FUNCTIONS
Ex.3. Find the value
If 2
25
of
x2
+
x
4
15
=
Now
3x
+
may
M, du = (4x + 1) dx. be written as f (4 x +
1)
+ \3
.
_3 / (4x + l)ds ~4 J 2x2 +x15
The
of the
first
13
4
/
cte
J 2x2
+ xlast integral is
integral
is
f Iog(2x
2
+ x 15),
44
log
by
f
12
(2),
and the
form solved
in Ex. 2,
and
is
x
o
Hence the complete
integral
is
Ex.4.
Find the value of
3x2
+4x
The value of this integral may be made to depend upon that of Ex. 1 in the same way that the solution of Ex. 3 was made to depend upon the solution of Ex. 2. For let 3 x 2 + 4 x = u then du = (6x + 4) dx.
;
Now 2 +
a;
5
=

Gx +
4)
+
J.
Therefore
J
{
V3 x 2 +
== =
4x
J
I
v8 x 2 +
2
J&gt;x
4x
=
The
11
I
1
+
4x)*[(6z
+ 4)dx] +
y fis
first integral
cr(3
x
+
2
V3x 2 + 4x, by 12 (1), + Vox + 12 x), by Ex. 1. Hence
is
2
and the second integral
the complete integral
is
3V3

3V3
log (3 x
+
2
+ V9 x2 +
12 x)
+
C.
16. Integrals of exponential functions.
The formulas
(1)
e"du
= eu
j
and
are derived immediately
entiation.
I
&lt;("du
=
lS
a
au
(2)
J
The
proof
is left
from the corresponding formulas of
to the student.
differ
26
ELEMENTARY FORMULAS OF INTEGRATION
17. Collected formulas.
?
~
J
flii.
n
+l
(2)
I
cos
udu =
u du
2
sin u,
(3)
/
sin
cos u,
(4)
/
sec
^ du
^
rfw
= tan w, =
rfw
(5)
/
csc
2
ctn u,
(6)
/
sec
u tan w
= sec u,
(7)
/
esc
u
^6
ctn
w ew
=
esc u,
(8)
/w
tan
c?^
J
= log sec u,
= log sin u,
(9)
/
ctn
u du
(10)
h
sec
u du
J
= log (sec u + tan M) = log tan
ctn u)
(
)
,
(11)
esc
udu = log (esc %
^
1
= log tan
&gt;
2
(12)
f

or
1 cos .
^sin"
(13)
^
/
2
+%
9 2
= tan~
a
1
 or
a
or
ctn
1
&gt;
a
1
a
osc
(14)
71
i
/du
ttVt**
tfM
1 =  sec _,u
_i^
a
/}/
cx
d
2
2
a
a
2
a
(15)
Vit + /flni a
2
= 1nafa + V^T^
)
or sinh ,
1
a
or
(16)
V^ /dU
2
=
a
2
1
ng
(7/
+ v/
9
?/
ffl
y.
11^
cosh"
&gt;
/1 rrv
)
a
(17)
INTEGRATION BY SUBSTITUTION
27
,
^f^^^Ta /du
1
u
a
1
*
,
2^
g
^
a
u
1 r
~a
tanh
,u
a
,.,
ox
(1)
(19)
= e*
fc*d&lt;u,
9
Ca u du =
18. Integration
au

(20)
by
substitution.
integral forms.
it is
necessary to reduce it to
In order to evaluate a given one of the foregoing standard
very important method by which this may be done is that of the substitution of a new variable. In fact, the work thus
far
A
has been of this nature, in that by inspection
of
we have taken
some function
In
x as
u.
is
many
cases
where the substitution
it
not so obvious as in the
new
previous examples, possible by the proper choice of a variable to reduce the integral to a known form. The choice of the new variable depends largely upon the skill and the experi
is
still
ence of the worker, and no rules can be given to cover
all cases.
A
systematic discussion of some types of desirable substitution will be taken up in later chapters, but we shall in this chapter work a
few
illustrative examples.
2 flr
/i* 2x
Let 2 x
+
3
=
z2
.
Then x
=
%
2
(z

3)
and dx
=
z dz.
Substituting these val
ues in the original integral,
we have
1(1
25
_ 2z 3 +
9z)
+C
.
Replacing z by
its
value in terms of x,
2(?X
we have
Jf
f
=
I
o
V2x +
3 (x 2

2x
+ 6) +
C.
Ex.2.
Find the value of
/
=
^+
z dz,
a
"
dx.
and
,
dx
=
1
x dx
=
zdz
Then, after substitution,
z
we have
l
*~
ft2
/z ^^=J
dz
r(
(
+
a2
\
,
trr*)
dz
=*+
a.
2 log
za a + ^Ta
28
ELEMENTARY FORMULAS OF INTEGRATION
its
Replacing z by
value in terms of
x,
we have
a
9:
2
/Vx
Ex.
3.
+
1
a2
/^
Vx2 +
a2
a
Find the value
a sin z.
of
f Va 2  x 2 dx.
Let x
Then dx
=
a cos z dz and
Va2
x2
=
+
a cos
z.
Therefore
f Va 2  x 2 dx = a 2 f cos 2 z dz =
a 2 f (1
z)
cos 2 z) dz
=
But
z
1 a2
(z
+
1 sin 2
+
a2
0.
=
sin 1 
,
and
sin 2 z
=
2 sin z cos z
=
2
Va 2
x2
.
a
Finally,
by
substitution,
we have
J
C Va 2  x 2 dx
=
2
x

x2
+
a2 sini a
+
C.
Ex.
4.
Find the value
of
f x 3 Vx 2  a 2 dx.
Let x
=
a sec
z.
Then dx
a 2 dx
=
a sec z tan z dz, and a5 Ttan 2 z sec 4 zdz a5

Vx 2
a2
a tan z.
Therefore fx3 Vx 2
=
=
2 (tan z
+ +
tan 4 z) sec 2 z
cZz
tan z
3
1 tan 5 z)
+
by
C.
But
sec z
=
,
whence tan z
=
x2

a2
so that,
substitution,
we have
2

a 2 dx
= TL V(x 2 
a2 ) 3
(2
a2
+
3 x2 )
+
C.
/dx +
(x
2
a 2 )i
5
Let x
=
a tan
z.
Then dx

=
I
a
sec 2
z dz
and
I
Vx 2 +
cos z dz
2
=
a sec z.
Therefore
r
I
 dx
2
=
r dz
I
=
r
/
/
=
sin z
J
(x
+
a2 ) 2
a2
/
sec z
a2
+
(7.
But tan
z
=
,
whence
sin z
=
Vx 2 +
a2
so that,
by
substitution,
dx
2
x
a2
x2
(x
+
+
INTEGRATION BY SUBSTITUTION
Ex.
6.
29
Find the value of f 
,
and
V6x2
Therefore
r
dx,
~_
3
^
(2
x
+
1)
V6x 2 + 8x +
an d hence
dz r J Vz 2 + 6 z + 6
log (2
_
r ^
dz
V(z
5)
=_
But
z
+
3
+ Vz +
2
6z
+
+ 3) 2 + C.
=
2x +
f

log (2
3
+ Vz 2 +
Gz
+
5)
==
+4 +
log
2
8x + 3
2x +
log
log 2.
2
3a
dx
(2x
2+V5x
2x +
2
+ l)V5x + 8x +
2
3
3x +
+V5x +
2
8x +
3
log 2 having been
made
a part of the constant of integration.
refer freely to these examples as possibly a type of substitution desirable in the solution of a new suggesting for substitution in similar problem. From them the following hints
The student should
cases
may
be deduced
:
as in Ex. 1. In integrals involving ~Va 4 bx try a + bx = 2 2 2 2 In integrals involving Var + a try either x + a = z as in Ex.
2",
8
,
2,
or
x
= a tan z,
as in Ex.
5.
In integrals involving In integrals involving
In integrals of the as in Ex. 6.
It is
Va V#
2
x2 try #
a try
2
a;
a
= a sin = a sec
0, z,
as in Ex.
3.
as in Ex. 4.
form
not to be supposed that the above substitutions are desir able in all cases. For instance, in Ex. 2 the substitution x = a tan z
does not simplify the integral
of advantage,
letter for the
;
2 2 but the substitution x f a =
z
2
is
though
it
is
rare that the substitution of a single
square root of a quadratic polynomial leads to any
simplification.
30
ELEMENTARY FORMULAS OF INTEGRATION
by
parts.
19. Integration
Another method
the reduction of a given integral to a gration ly parts, the formula for which
for the differential of a product,
known
is
importance in type is that of inte derived from the formula
of
d (uv)
= u dv
f
v du.
From
this
formula we derive directly that
uv
=

u dv
+
I
^ du,
which
is
usually written in the form
I
u dv
= uv

v du.
In the use of this formula the aim
inal integration
is evidently to make the orig the evaluation of a simpler integral. depend upon
Ex.
If
1.
Find the value of Cxex dx.
let
we
x
=
u and
e x dx
=
dv,
Substituting in our formula,
we have du we have
dx and
v
=
ex .
jxe
x dx
=
xex xex
iex dx
ex
+C
It is evident that in selecting the expression for dv it is desirable, if possible,
to choose
an expression that
is
easily integrated.
Ex.
2.
Find the value of
let
Here we may
sinix
Substituting in our formula,
/r
sin 1 xdx.
J
=
u and dx = we have
dv,
whence du
=
V1
=
x2
and
v
= x.
sm~ l xdx = x sin 1 x
xdx
x2
I
=
x sinix
12
(1).
+ Vl 
+ C,
the last integral being evaluated
by
Ex.
3.
Find the value of
/
xcos 2 xcZx.
Since cos 2 x
/I
^
(1
=
+
cos2x), we have
2
x cos 2 x dx
r
I
(x
+
x cos 2 x) dx
=
x2
(
1 
/*

x cos 2 x dx.
/
4
2
t/
INTEGRATION BY PARTS
Letting x
31
v
=
u and cos 2 x dx
//
du,
=
we have du = dx and
1
^ sin 2 x.
x cos 2 x dx
=
 sin 2 x
2
f
/*
/
sin 2 x
dx
2
/
=
fx
cos2 x dx
 sin 2 x
2
 cos 2 x 4
+
C.
.*.
J
=
4
+ ( sin 2 x + 2
\2
2
 cos 2 x ) 4 /
+C
+
C.
=
Sometimes an integral
tion
(2 x
+
2 x sin 2 x
+
cos 2 x)
may
be evaluated by successive integra
by
4.
parts.
Ex.
Find the value of Cx 2 e x dx.
will let
Here we
x2
=u
and
e x dx
=
dv.
Then du
2
=
2 x dx
and
v
=
ex
.
Therefore
Cx 2 ex dx = x2 ex
Cxex dx.
The integral so that finally
J
x^dx may
be evaluated by integration by parts (see Ex.
2 (x
x 1) e
1),
Cx 2 ex dx
= x2^
+C=
e x (x 2
2x
+
2)
+
C.
Ex.
6.
Find the value of

eF* sin 6x dx.
Letting sin bx
=
u and
eax
dx
=
dv,
we have
* sin 6x
/&* sin 6x dx = a e
In the integral
ax
a J
le** cos 6x dx.
je
cos6xdx we
cos bx dx
let
cos&x
=
u and

T
e"
dx
=
du,
and have
/&**
=
 e?* cos 6x a
+
a J
fe^ sin 6x dx.
Substituting this value above,
we have
(
Co J
* sin bx
dx
=
a
e
sin bx
 eax cos bx
a\a
member
+
a Jfe

ax sin bx dx)
/
Now
bringing to the lefthand
of the equation. all the terms con
taining the integral,
1 H
we have
e"*
1
\
a2 / J
)
f
sin bx
dx
=
a
e x sin
bx
e"*
cos
a2
in te
fox,
whence
J
C ** sin 6x dx 
"(
a2
 6 cos te) + &2
32
Ex.
ELEMENTARY FORMULAS OF INTEGRATION
6.
Find the value of
eax
(
Ce^ cos bx dx.
)
,
The
it is
acosbx
a2
result is
+ bsmbx
62
5.
the
+
work being
left to
the student, an.
exactly like that of Ex.
Ex.7.
Placing
Find the value
of
f Vx 2 + a 2 dx.
du,
Vx 2 +
a2
= u and dx =
whence du
=
Vx 2 +
a2
and
v
=
x,
we have
J
Since x2
J
a2 ) a2 the second integral of
,
Vx2 +
(1)
a2 be written as
=
2
(x
+
may
/x
J which equals
dx
a2
a?
Vx2 +
J
f Vz 2 + a?dx 
J\
Vx2 +
(1),
a2
in Evaluating this last integral and substituting
we have
f Vx2
whence
+
a2 dx
=
x
Vx 2 +
a2
 f Vx 2 +
a2
a2 dx
+
a 2 log (x
+ Vx2 +
a2 ) ]
.
a2 ),
f Vx 2 + a 2 dx
=
[x
Vx 2 +
+
a2 log (x
+ Vx 2 +
20. Possibility of integration. In this chapter we have learned how to express the integrals of certain types of functions in terms
elementary functions, and the discussion of in Chaps. VI and VII. gration will be continued
of the
methods of inte But it should
be noted
not always possible to express the integral For of elementary functions in terms of elementary functions.
now
f
that
it is
example
J
= V1^1&V
it is
cannot be so expressed; in
fact,
kind. this integral defines a function of x of an entirely new of certain func Accordingly when it is said that the integration
tions is not possible,
for
meant that the
integration
is
not possible
which are in only the elementary functions, In this fact the functions generally used in applied mathematics. from differentiation, which can differs
one
who knows
respect integration
radically
functions. This fact is not always be performed upon elementary to what takes place in surprising, since it is closely analogous
PROBLEMS
33
connection with other operations which are the inverse of each other. For example, the addition of positive numbers can always
tion of subtraction tion of negative
is
be expressed in terms of positive numbers, but the inverse opera is made always possible only by the introduc
numbers
;
also the involution of rational
is
numbers
always possible, but evolution
always possible only after
irrational
numbers are introduced.
PROBLEMS
Find the values of the following integrals
:
2.
4.
IZUL^rfx.
O f*Ls
51
.

s
^dx.
19
mxdx
/sn &COSX a +
C
J
(I
20.
Xdx
.
+
x2 ) 3
7.
Jf(xl)
2
x2 dx.
21
C 3a J (2 + 3x2 ) 2
.
8
/^lT^ dx
x*dx J x +

22.
r. C
(
x
+ ^ dx
*2
J
9.
f~
I
Vo + 4x +
a + bx
a
U,
23.
I
C
v J
dx.
f\
I~L
_
+
~T b
X t
11.
C
l
24
?L.
e + C *Je2x + tan2x
l
/X
f/
1
25
/^KifcT
"
13.
(1+ x^tanix
1 X
.
.
,u. 26.
f ^ J
i
^ (tan (tanix1)^
(fa.
1
14
* C 1 + cos x J (x + sinx) 3
^
27
r J
34
28.
ELEMENTARY FORMULAS OF INTEGRATION
*
b)*0*xdx.
,
32.
r
J
2
x
a
+
x
33
&lt;**
I
r
^7= 5

Vx
Vx 4
S7=
34.
31
VxM^)
36.
J"[sin
35.
Jsin3xcosxdx.
6)
4
(ax
+
6)
+
cos 4 (ax
+
6)] sin (ax
+
cos (ax
+
6) dx.
37.
j*(cscbx
ctnbx)cscbxdx.
46.
f (sec3x f
/&gt;

tan3x)
2
dx.
38.
f
./
^
sin 2 (ax
2
+
47
6)
_
sec x
J
+
tan x
on
r sec2 (z
+
a2 ) x dx
42)
J
tan 3 (x 2
I
.sinmxsinrixdx, (m
cos
^ ^
6
n).
W.
ft
.
f
J
I
sec x 2 tan x 2
2 a5

x dx
.
_
r
I
+
sec x 2
f
49.
J
50.
J
/
mx
cos nx dx.
(m
ri).
41.
f (sec x
/
tan x
sec x) 2 dx.
sin (ox
+
b)
cos (a x
+
)
dx.
42.
Jsin J
2
2xdx.
51.

(tan
2x +
x2
i/
ctn 2x) 2 dx.
43.
f
Cos2xdx
,
smx
sin 2
52.
^f(sec
+

tan x2 ) 2 sec x 2 x dx.
.
44.
Tcos x
x dx.
53.
J
J
cos 2 (1
2 x) dx.
45.
2 x Jf(tan

ctn 2 x)dx.
54.
\
cosx
sin2xy
55.

(sec 2 x
+
tan 2 x

1) (sec
2x

tan 2 x

1) dx.
56
esc x
dx
ctn x
"/ cos e
.
sin e
/
sec
d0
J sec0
&gt;
+
tan0
rji!i
sin
dd
sin
J
x
i1
60.
1
+
cos x
35
36
ELEMENTARY FORMULAS OF INTEGRATION
dx.
e
+
Vl 
2x

x2
141
f^T~
PROBLEMS
37
x
159.
x2

a2
.
176.
ftaniaxcfx.
f
J
X5dx
Vx 2 + a2
/
38
ELEMENTARY FORMULAS OF INTEGRATION
178.
flog(x
+ Vx 2 +
a 2 )dx.
186.
J"x
3
sin
ax ax.
2
179.
fxsinaxdx.
187.
J (log ax)
ax.
180.
f sec 1 ax dx.
Cxsec~ l axdx.
188.
J(x log ax)
2 d*x.
181.
189.
Cxsm z axdx.
Te 2;r cos 2 xdx.
fe^sin 2xsinxcZx.
182.
fxtaniaxdx.
190.
183.
Cx 2 e nx dx.
f(x
191.
184.
+
2
l)
ea; dx.
192.
Te 3;c sm3xcosxdx.
185.
Tx2 cosaxdx.
193.
fX
sin
*
az dx.
CHAPTER
III
DEFINITE INTEGRALS
have said in 3 that there are two impor tant problems in the use of infinitesimals, the one involving the limit of a quotient, the other the limit of a sum. have also
21. Definition.
We
We
4 that the derivative of a function is the limit of the of two infinitesimal increments, so that the limit of a quotient quo tient is fundamental in the differential calculus and its applications.
noted in
Similarly, there exists a limit of a
sum which has fundamental
This limit
is
connection with the subject of integration.
definite integral
called a
and
is
defined as follows
:
If f(x)
for
as
is
all values
a function of x which is continuous and onevalued of x between x = a and x = b inclusive, then the
of f(x)dx between a and
I is defined as the limit,
definite integral
n
increases indefinitely, of the following
sum of n
terms,
w/iere
7
A A# = 1)
a

n
and x v x2 x s
,

,
&gt;,
xn _^ are values of x between
a and
b
such that
The sum
notation
in the above definition
is
expressed concisely by the
where
(sigma), the
"
Greek form
of the letter S, stands for the
and the whole expression indicates that the sum is to be taken of all terms obtained from /(#,.) A# by giving to i in succession the values 0, 1, 2, 3,  1, where XQ = a. ., n
sum,"
. .
word
40
DEFINITE INTEGRALS
Also the definite integral
is
denoted by the symbol
b
t
r f(x)dx Ja
I
where
f is
a modified form of
S.
Hence the
definition of the
definite integral
may
I
be expressed symbolically by the equation
i=
rb
Ja
= UmVf(x )kx. f(x)dx
l
=*&gt;
^
1
&lt;=0
The numbers a and
b are called
* the lower and the upper limit
definite integral. respectively of the of a definite Ex. The conception of mechanical work gives an illustration the work done in moving a body against a constant integral. By definition, which the body to the force multiplied by the distance through
force
from A(x = a) against a force which is not constant but a function of x and Let the line AB be divided into n equal intervals, each expressed by /(x). = 7.) fi g 3, n n (I s 2 equal to Ax, by the points MI, be /(a) Ax if the work done in moving the body from A to MI would Then the interval AMi. Conse the force were constantly equal to /(a) throughout done is small, /(a) Ax is approximately equal to the work quently, if the interval and 2 is approxi between A and MI. Similarly, the work done between MI to 3 approximately equal 2 and mately equal to f(xi) Ax, that between so on. Hence the work done between A and B is approximately 2 Ax, and
O

is
equal
jf
JL
4 j/
3
MI
4
Al&lt;
b
*
is
moved.
Suppose now that a body
is
moved along to B(x b)
OX
(fig. 3)
M M
,
,
,
M
\&gt;
M
M
M
/(x ) equal to
f
^ +f ^ ^ + f{^ Ax +
A
.
.
.
+f(Xn _ l ) Ax.
Hence we have,
The
if
W
the better is this approximation. larger the value of n, and B, represents the work done between
W = Lim V
The use
of the
i=n\
f(Xi)
Ax
= f
h
/(x) dx.
"^o
word
" "
integral
and
of the
symbol
/
suggests a
connection with the integrals of the previous chapter. This con 25 for the present, it is to be empha nection will be shown in
;
or sized that the definition is independent of either differentiation
integration as previously
*
known.
word
"
The student should
notice that the
it is
limit
"
ferent from that in
which
used
when a variable
is
is here used in a sense quite dif said to approach a limit (I, 53).
/U
Iii
GRAPHICAL REPRESENTATION
c
a
b
I
41
distinction to the definite integral
is
f(x)dx,the integral
f(x)dx
called the indefinite integral.
The definition assumes that the
limit of
V/(# A# always
f)
exists.
A
rigorous proof of this will not
will find it
of the
be given here, but the student obvious from the graphical representation geometrically
article.
next
22. Graphical representation.
Let
LK
(fig.
4) be the
graph of
&lt;
= a and OB = b. f(x), and let OA For convenience, we assume in the first place that a that f(x) is positive for all values of x between a and b.
n
and lay
off
b
and
find
We
the
2
equal lengths
MM
2
=MM =
1
(In
=*... = 3 n = 9.) fig. 4,
Let
,
OM = x OM _ = x
1
l9
n
l
n
_r
Draw
f(a),
3/5 3/6 3/7 3/8
the
ordinates
AD
the
,
Draw
also
lines
FIG. 4
J^K 8
,
P _^R
n
n
parallel to
OX.
Then
= the = the /(#i) A# = the f(x Aa?
/(a) A#
2)
area of the rectangle ^
area of the rectangle
area of the rectangle
area of the rectangle
(._!) Aaj
= the
M_
H
l
P _ Rn B.
n
l
The sum
is
then the
sum
of the areas of these rectangles,
and equal
is
to the
area of the polygon the limit of this sum as
ADR^R,
n
.
.
R
n
n _^P _
l
R
n B.
It is evident that
is
indefinitely increased
the area
bounded by AD, AB, BC, and the arc DC.
42
Hence
DEFINITE INTEGRALS
Cf(x)dx
Jo.
= the
area
ABCD.
if
It is evident that the result is
not vitiated
AD
or
BC
is of
length
zero.
The area
2
ABCD
etc.
is
exactly the
sum
of the areas
ADP^Mi, MiPiP 2 Ma
,
,
P2
P3M"3 ,
But one such
area, for
example lf1 PiP2 3f2
rectangle the area
differs
from
that of the corresponding
MiPiR 2 M2 by
of
A
JKi
MM
z
3
M&lt;
M,
M
G
M, M,
B
"
x
the
is
figure
less
PiR 2 P2
,
which
than that of the rectangle
Ax Ay, where Ay = J2 2 P2 The area of Pi# 2 P 2 is an
.
infinitesimal of higher order
than
MiPiR&lt;zM2
,
since
PiR 2 P 2 MtPiRzMz
whence Lin
L
Ax Ay
Ay
yAx
Therefore
of
(
3)
the areas
FIG. 5
affect the limit of the
the
as
triangular figures
such
PiE 2 P2 do
not
sum used
all
in finding the area of the entire figure.
If
f(x)
is
negative for
values of x between a and
is
b(a&lt;b),
the graphical representation
as in
fig. 5.
Here
= f(xj &x =
f(a) A#
so that
I
the area of the rectangle the area of the rectangle
etc.,
r*
Ja
f(x) dx
=  the
area
ABCD.
we
In case f(x) is sometimes positive and sometimes negative have a combination of the foregoing results, as follows
:
If
the
a&lt;b,
the integral
I
Ja
f(x)dx represents
the algebraic
sum of
areas bounded ly the curve y =f(x), the axis of x, and the ordinates x = a and x = b, the areas above the axis of x being
positive
and
those below negative.
GENERALIZATION
If
43
The only change
a
&gt;
I,
A#
is
negative, since
A# =
ba
necessary in the above statement, however, is in the algebraic and those signs, the areas above the axis of x being now negative
below
positive.
.
Ex. The work done in moving a body against a force may be represented is the work in moving a body along by an area. For if the force is f(x) and the axis of x from x = a, to x = 6, then (Ex., 21)
W
W=
of
b
f f(x)dx.
Consequently if the force is represented by the graph DEC (fig. 6), the equation which is y = f(x), the area ABCED represents the work W. This fact is taken advantage of in constructing an indicator diagram attached
to a
steam engine. Here
ordinate
the
pressure. travels from
to
represents the distance traversed represents the Then as the piston
AB
by the
piston,
and
A
to
B
and back
is
A
the
curve
DECFD
automatically drawn. The
area
ABCED
represents the
work done by the steam on the piston. The area ABCFD represents the work done by the piston on the steam. The
difference of these
which
curve
piston.
is
areas, the area of the closed
two
A
FIG. 6
B
represents the net work done by the steam in one stroke of the In practice this area can be measured by an instrument, called a planimeter, or the figure is divided into rectangles and the area computed approxi mately. The latter method illustrates the definition of the definite integral.
DECFD,
23. Generalization.
\
In the definition of
21,
where
x= a
and x
=
b.
The sum
may
accordingly be written
or,
more compactly,
t=0
(1)
44
This
DEFINITE INTEGRALS
sum may
be generalized in two ways a mere matter of convenience to take the
:
In the
first place, it is
i
increments xi+1
, ,
x equal
for all values of
x..
In fact the n
1
values x v x 2 xs ",xn _ l may be taken at pleasure between the values x = a and x = b without altering the limit approached by
the
sum
(1),
provided
all
the differences %i+1
x are made to
i
approach zero as
MM
geometrically obvious from the graphical representation, for the bases V ... of the rectangles of fig. 4 may be of unequal length. 3 2
n
increases without limit.
This
is
AM
,
D/
FIG. 7
the
In the second place, the factor f(x ) in each sum (1) may be replaced by /(&), where
t
of the
is
terms
of
.
any value of
x between x and xi+r The effect on the graphical representa tion of fig. 4 is to alter the altitudes of the rectangles without
i
in altering the limit of their sum, as exemplified
fig.
7.
It
may
be noted that the rectangles here
by
infinitesimals of higher order.
from those of fig. 4 Hence the sum theorem of
differ
is
3 applies. For a rigorous discussion of these points the student
to
referred
advanced
*
treatises.*
See, for example, GoursatHedrick, Mathematical Analysis, Chap. IV.
PBOPEBTIES OF DEFINITE INTEGRALS
24. Properties of definite integrals.
definite integrals are
s* b
45
The following
:
properties of
consequences
/&
of the definition
1.
/
cf(x)dx
=c
I
i/a
Ja
f(x)dx,
2
Ja
3.
fVi(*) +/&gt;(*)]
&lt;**
=
a Jf
f,(x)dx
+ f f (x)dx, Ja
2
n
Ja
4.
C f(x)dx =  f f(x)dx, Jb
J
Jf a rb
I
f(x)dx =
Jo,
f
f(x)dx
+ f f(x)dx, Jc
where
5.
Ja
tion,
f(
The truth of formulas 1 and 2 follows at once from the defini 22. We shall and that of formula 3 follows at once from show the truth of formulas 4 and 5 graphically. A fully rigorous
FIG. 8
21 without the use of proof could be based on the definition of diagrams, but would follow the outlines of the following graphical
discussion.
To prove formula Then
and
I
4,
consider
area
fig. 8,
where
OA = a,OC=
tliQ
c,
OB = b.
CBGF,
is
C f(x)dx = t\iQ
*J a
b
ACFE, C f(x)dx =
Uc
area
c f(x)dx = the Ja
area
ABGE. The
3.
truth of formula 4
ap
parent for any order of the points A, C, B, reference being had,
if
necessary, to formula
46
DEFINITE INTEGKALS
5,
b
To prove formula
.
consider
t
fig. 9,
where the area
ABCD
repre
sents the value of
C
I
Ja
f(x)dx and
let
m
and
M respectively be the
AB
Construct also the rec
smallest and the largest value assumed by /(a;) in the interval AB. with the base and the altitude Construct the rectangle
ABKH AH= M. Its area is AB AH (b tangle ABLN with the base AB area is AB AN= (b a) m.
Y
a)
M.
and the altitude
AN= m.
Its
CD
FIG. 9
B
X
Now it is evident that the area ABCD is greater ABLNsm& less than the area ABKH. That is,
a)m&lt;
than the area
r f(x)dx&lt;(b Ja
I
a)M*
Consequently
some quantity greater than m and less than M, and is represented on fig. 9 by AS. But since f(x) is a continuous function, there is at least one value f between a and b such that /( ) = and therefore
where
//,
is
//,,
c/a
//(*)&lt;**:
= (&ABCD
ABTS
.42V
*
is equal to a rectangle Graphically, this says that the area whose base is and whose altitude AS lies between
AB
and AH.
A
I
slight modification is here necessary
if f(x)
=
k,
a constant. Then
M= m = k
and
rb
Ja
f(x)dx=
(b
a)k.
EVALUATION BY INTEGRATION
25. Evaluation of the definite integral
47
by
integration.
When
f(x) the integral
limit b
is
a
known
is
function and a and b are constants, the value of
by a variable
(fig.
fully determined. Hence if we replace the upper x, the definite integral is a function of x.
Graphically
constant,
10),
and
OM = x,
^ (x =
)
UJa
f (x) dx = area, AMPC,
a variable.
where
OA =
a,
a
Let us place
area
rx
I
Ja
f(x) dx
= the
109,
6,
AMPC.
Now we
have shown in
I,
that
proof of this can be given by use of the properties of 24, and this proof has the advantage of being
really independent of the graphical representation, although for conven
A"
A new
ience
we
shall refer to the figure.
li.
FIG. 10
Take
MN=
Then
$(x
+ h) =
(/&gt;
I
Ja
(x)
f(x) dx
= the
area
ANQC,
s*x + h
whence
c/&gt;
(x f h)
=
I
JX
f(x) dx
area
= the =
where x
&lt; &lt;
MNQP,
(by 3 and
4,
5,
24)
()&gt;
(by
24)
x
+
h.
Therefore
Taking now the limit as
T
;
7t
approaches zero, and remembering that
.
Lim
=o
.
=
d
^(a?)
and Lim f
,
,
.
=
a;,
we have
48
Let
DEFINITE INTEGRALS
now F(x)
be any
known
&lt;/&gt;(#)
function whose derivative
(
is
f(x).
Then F(x)
differs
from
by a constant
30), that
is,
f*f(x)dx=F(x) + C.
*J a
To determine
C, place
x
= a,
r*
and notice that
I
Ja
whence
f(x) dx
F(a).
= 0.
Then
Therefore
= F(a) + C,
*J a
C=
t
f /(*) dx = F(x)  F(a)
Ja
whence
C f(x) dx = F(b)  F(a).
This result gives the following rule for evaluating a definite
integral
:
To find
x
the
value of
r f(x)dx, Ja
\
b
evaluate
c
I
J
f(x)dx, substitute
=
b
and x
=a
successively,
and
subtract the latter result
from
the
former.
It is to
be noticed that in evaluating
is
I
f(x)dx the constant
is
of
integration
ever,
if
to be omitted, since
is
F(a)
that constant.
How
the constant
[F(b)
added,
it
disappears in the subtraction, since
+ C]  \F(a) + C]= F(b)F(a).
F(b)F(a) by
the
In practice
bol [F(x)] a
b
,
it is
convenient to express
sym
so that
Ex.
2.
f 2 sinxdx =
*
4
[
cosx]
2
=
2
cos^ + cosO
= 1.
2 log 4
Ex.3, f
Jo x 2
+
x
2
+
dx
2
=
[2 log(x
4 x +
tani
2)]*
=
Ex.4, f Ji
=
1
+
x2
[tan x]
1
v?=
V3 
tan 1 (
1).
CHANGE OF LIMITS
There
is
1 here a certain ambiguity, since tan
49
1)
Vs and tani(
have each
If, an infinite number tan J x is composed of an infinite values of tan 1 ambiguity is removed by taking the
of values.
however, we remember that number of distinct parts, or
the graph of branches, the
V3
^
/*
the same branch of the function. For
if
we consider
if
and tani( 1) from dx  = tan * b  tan 1 a
and
select
1
any value
tan a, since the value of the integral with a to
its final
= a, tan 1 b must be taken equal to then zero. As b varies from equality tan 1 a to the nearest value of value, tanift will vary from
of tan a, then
1
b
is
tan 1 b.
The simplest way
them both between
to
1 1 choose the proper values of tan b and tan a
is
to take
2
and
2
.
Then we have
r Vs
dx
l
Ji
dx

+#
_*_( _ ^\ _ I5 12
9
\
4/
Ex.
5.
Jo
f*
Va 2
x2
= r s ini1 =8inHBiniQ. a
L
lo
J
The ambiguity
that sin  must
1
in the values of sini
lie
and siniQ
first
is
removed by noticing
in the fourth or the
quadrant, and that the two
other by continuous values must be so chosen that one comes out of the 1 to accomplish this is to take both sin ^ and change. The simplest way
7T
,
7T
iiQ between
2
and
2
/*2
dX
Then
7T =
=
is
7T
Jo
Va2
x2
^
convenient to evaluate
"
26.
Change
of limits.
In case
it
Cf(x)dx by
substitution, the limits of
f
f(x)dx may be replaced
by the corresponding values
this,
of the variable substituted.
To
see
suppose that in
I
f(x)dx the
t,
variable
x
is
replaced by a func
tion of a
new
I, t
variable
from a to
varies continuously
such that when x varies continuously from t to t r Let the work of
:
be indicated as follows finding the indefinite integral
x)
dx
=
C&lt;f&gt;
(t)
dt
=
&lt;$&gt;
(t)
= F(x),
4&gt;(Q
where F(x)
terms
of x.
is
obtained by replacing
t
in
by
its
value in
Then
F(b)F()=&lt;t&gt;(t l )&lt;t&gt;(t ).
50
But
and
DEFINITE INTEGRALS
F(b)
 F(a) =
*) a
f f(x) dx
4&gt;(^4&gt;(g=
Jt
f
&lt;l&gt;(t)dt.
Hence
Ex.
1.
Ja
Find
Jo
Place x
varies
[
f f(x) dx = f
Jt n
(t)
dt.
Va2 x 2 dx.
Then dx = a
Hence
cos
cZ0,
=
a sin 0.
to
.
and when x
varies
from
to a,
from
2
Jo
L 2
In making the substitution care should be taken that to each
value of x between a and b corresponds one and only one value of t between t Q and t lt and conversely. Failure to do this may lead to
error.
7T
/5 cos _
7T
d0, which
by
direct integration
r~
Let us place cos0
= x, whence
77
= cos 1 ^
is
and d0
=
/7/j
is
equal to
2.
,
Vix
where the sign
2
depends upon the quadrant in which
this substitution in
2
/
found.
lies in
We
two
n
cannot, therefore,
make
cos0cZ0, since
different quadrants; but
we may
write
?
n
/g cos _![
2
(i0
=
J_E
2
I
cos
d0
f
I
cos
Jo
(Z0,
and
in the first of the integrals on the righthand side of this equation place
=
cos 1 ^,
d(f&gt;
=
.
VI 
and
in the second
0= cos 1 ^, eZ0 =
_
^
x2
Vlz2
x2

Then
l
_
x2
^
Vl
x2
"v/1
/+!
Ex.
3.
Consider
Ji
Then, when x

dx and place x 2
=
t.
tute without care
x
is
negative
;
= 1, t = 1, and when x = 1, would lead to error. But x = and x = $ and dx = J*~*ctt when x
=
f*
/&gt;
1;
t* is
1
and the attempt to substi and dx = ^t~*dt when positive. Hence
.
/+!
Ji
f
1
dz=
Ji
/
dx+
Jo
(
dx
=
/^
/
/
1
(
jr*&lt;tt+
Ji
Jo
I
ir*(tt= J
Jo
i
INTEGRATION BY PARTS
27. Integration
51
by
parts,
by
parts.
If it is
desired to integrate
and a and
b are values of the
s*b
I
independent variable, then
s*b
I
u dv
Ua
To prove
this,
= \uv\
v du.
Ua
note that
s*b
it
follows at once from the equation
/
6
s*b
/b
Ex.
1.
(I
(uv)
=
I
(u dv
+ v du] =
I
/
u dv
a
+
I
^ du.
\J a
l
Ua
Find
r
/
xex dx.
Take x
=
w,
Jo ex dx r
I
dv
l
;
then
xe x dx
=
Jo
n
\xe Y
x
i
r
I
1
e xdx
=
e
Jo
\e
x1
1.
Ex.
2.
Discuss
1
2
Jo

sin"xdx.
Take
sin"
^
=
M,
sinxdx
=
dv; then
7T
7T
7T
2
Jo
I
sin n xdx
=
=
[
cosx
1)
sin n
~1
2
x]
+
2
(n
1)

sin"
2
x cos2 xdx
^
/o
(n
f
Jo
2
2 (sin x
sin"x)dx.
By
transposing
we have
7T
7T
n
2
/
sin n xcZx
=
2
(n
i
1)
c/
Jo
2
sin"
{
sin"
2
x(Zx,

whence
If, in (1),
C
Jo
2
sin n x
dx
=

n
Jo
f
 2 x dx.
(1)
we
place n
=
2,
we have
1
psii^xdx 2Jo Jo
If, in (1),
pdx = 2 .^. 2
7T
we
place n
7T
=
3,
we have
2
Jo
If, in (1),
/
sin 3 xdx
=

3Jo
/
we
place
n
=
4,
we have
3
/
C\ sin*xefx =
Jo
J
4Jo
sin 2 x(Zx
.
=
42
2
52
If, in (1),
DEFINITE INTEGRALS
we
place
n
5,
we have
4
2 T "sinozdzsi
Cl
Jo o
is wa] Continuing in this way, we find
42 63
13
5
f Jo
Jo
isi
(2
k

1)
TT
24.6..2A;
246. .2ft
f
3
5
7
(2
k
+
1)
28. Infinite limits.
It is possible to
have the upper limit
oc,
where by
definition
f.
f(x) dx
= Lim
I
f(x) dx.
The integral, then, is represented by the area bounded by the curve y =f(x), the axis of x, and the ordinate x = a, the figure being unbounded at the right hand. There is no certainty that
.
such an area
is
either finite or determinate.
/~
I
The
tests
by which
it
may
be sometimes determined whether
Ja
In
case,
f(x)dx has a meaning
possible to find the in
will not be given.
definite integral
however,
it is
F(x)
=
^
f(x)dx, the definite integral can be found Iw flip formula by the formula
Ja
where
F(ao)
= Lim F (b)
.
EX.
1.
^=.85
Ex.2.
Jl
r^ = [X2
i] X_\l
=l.
(fig.
12)
Ex .3.
FIG. 13
r
I
oo
Jo
Binxdx=[
coex]
(fig.
= indeterminate,
13)
of the definite integral Similarly, the lower limit, or both limits,
may
be
infinite.
INFINITE INTEGRAND
29.
Infinite integrand.
53
21
it is
According to the definition of
f(x)dx should become
x
infinite
between
= a and x =
Xt&gt;
b.
It
is,
however, possible to admit the case in which
f(x) becomes infinite
when x
=b
by means
of the definition
Ja
and
to
C f(x)dx =
Um C
h
+ oJa
/(,*)
dx,
compute the integral by the formula
.6
Ua
/
h).
where F(b) means lim F(b
The
value.
integral has not, however, necessarily a
is
finite or
determinate
Graphically, the integral the curve y =/(&lt;), the axis of
represented by the area between
x,
the ordinate x
tote
= a, and the
x
=
asymp
b.
Ex.2,
Jo
Jo
(fig.
15)
i
Similarly, f(x) may become in finite at the lower limit or at both
limits.
If it
o
FIG. 14
FIG. 15
becomes
infinite for
any value
c
between the
limits,
the integral should be separated into two inte grals having c for the upper and the lower limit
respectively.
Failure to do this
may
lead to error.
Ex.
3.
Consider
/*
Jl
X2
Since
becomes
infinite
when x =
:
(fig. 16),
we sepa
rate the integral into two, thus
1
O
FIG. 10
dx
l
dx
54
Had we
DEFINITE INTEGRALS
carelessly applied the incorrect formula
11+ 1
we should have been
30.
led to the absurd result
2.
The mean value
of a function.
We
have seen in
24 that
/
where f
lies
=
r^ JCf^dx,
^
(*
(1)
a
between a and
J.
The value
is
called the
meaw m/^e
of /(a;) in the interval
from a
to
5.
This
is
in fact
meaning , y n ^ correspond y Q) y lf y 2 to n values of x, which divide the interval from a to I into n equal parts, each equal to Aa?. Then the average of these n values of y is
value of n measurements.
an extension
of the ordinary
of the average, or

mean,
For
let
,
n
This fraction
is
equal to
a
As
7i
limit
is
indefinitely increased,
\
C
I
b
y dx

~\.
this expression approaches
as
a
C
I
b
fix) dx. j
Hence the mean value
"
of a
"
be considered as the average of an infinite number of values of the function, taken at equal distances between a and b.
function
may
Ex.
1.
The
velocity
1
is
Find the mean velocity of a body falling from rest during the time t\. is gt, where g is the acceleration due to gravity. Hence the mean
velocity
1
c i gtdt Jo

=
^gt\.
This
is
half the final velocity.
"
In using the indicator diagram ( 22) engineers use the mean effect ive pressure," which is defined as the constant pressure which will do the same
Ex.
2.
amount
and
is
indicator diagram.
as is done by the varying pressure shown by the found by dividing the area of the diagram by b a, accordingly an example of the mean value of a function.
of
It is
work per stroke
THE MEAN VALUE OF A FUNCTION
Formula
(1)
55
may
be written in another form, not involving the
integral sign.
Let us place if(x)dx
**
= F(x),
then f(x)
=F
(x),
and
(1)
becomes
ba
or
F(a)l
(%).
(2)
(3)
.F(a)
= (b
d)F
Formula
(fig.
For let (2) has a simple graphical meaning. 17) be the graph of F(x) and let OA = a and OB=b. Then
F(b)
LK
ba=AB,
 F(a) =BEAD= CE,
()
a
=:
and
_JJ
b
I
DC
 the
slope of the chord
is
DE.
If
now
f
is
any value
of x,
F (^)
the slope of the tangent at
the corresponding point of
LK.
some point between
to the
A
and
is
B
Hence (2) asserts that there is for which the tangent is parallel
chord DE.
This
evidently true if F(x) and (x) are continuous.
F
K
Formula
(3)
may be used to
prove the proposition, which we have previously used with
out
proof,
namely
:
If
the
derivative of a function is
always zero, the function is a constant. For let F (x) be always zero and let a and b be any two values of x. Then, by (3), F(b) F(a)=Q. That is, any two values of the function are
equal; in other words, the function is a constant. From this it follows that two functions which have the same
derivative
differ
3&gt;(x)]
ly
a
constant.
For
3&gt;
if
F
f
(x)
=
&lt;&
(x),
then
[F(x)
= 0;
whence
F(x) =
(x)
+ C.
31. Taylor s
special case of a
to obtain.
and Maclaurin s series. Formula (3), 30, is a more general relation, which we will now proceed
Let us take the equation
t/ a
f
f"(x)dx
=/ (*)_
56
DEFINITE INTEGRALS
x.
and integrate between a and multiply both sides by dx,
have
x, x
We
s*x
(
Ja Ja
\
/
(*) fa
=
n*
n
I
a,
x
J\a
f (x)dx f (a)dx *J
(a).
=f(x)f(a)(xa)f
.
Therefore
f(x) =f(a)
+ (x )/ () + f
C"f"(x)d3t*.
t/a i/a
But if m is the smallest and a and x, then (24), when #
M the largest value
&gt;
of
f"(x)
between
a,
whence
m (xaf
2~ *
Ua
C*r*ut
t/a
/
/a
/
^
^w
,
2
^M(x
~2
Hence
If
Fl*f
t/o
(x).datK
where
is
m
&lt;
/*
&lt;
Jf
:
/"(a)
,
is
a continuous function, there
aj,
at least
/*
one value of
56).
x,
say
between a and
for
which
f"
() =
(I,
Therefore
we
have, finally,
/(x) Again,
let
=/() + (x 
)/ ()
+
^2^/"(l^
c
(!)
us take
= r
Ja
and
/"
=/ () / ()(* )/"().
m*f
whence
where a
*
&lt;
"(x)d3t?=
ff Ja
(x)dx
Ja
Cf (a)dx
Ja
Clxa)f"(a)dx
+
2
&lt;
^"
+
/
"&&gt;*
(2)
03.
The symbol [2 means 1 2, 3_ means 1 means the product of the n integers from 1
1
2
to
3,
n
and, in general, n, read factorial n, inclusive.
TAYLOK
Similarly,
if
S
AND MACLAUKIN S SERIES
with
/
57
we
start
Ja
f (n+l) (x)dx, we
obtain
where 7
being
tt
= ^
_ 71 ~T~
^L
/ (n+1
is
&gt;
(f ),
where a
&lt;
f
&lt;
a,
the only restrictions
tliat
the
of
n
+\
n
derivatives of f(x) exist
and are continuous.
The value
but
it
R
in (3)
not
known
a
is
f 1 terms of n and this approximation (3), omitting approaches f(x) as a limit as % is indefinitely increased. In this case we say that f(x) is expanded into the infinite series
n= oo by the
Lim7?n =
usually happens 0. Then the value otf(x)
first
that, if
x
exactly, since f is unknown, a sufficiently small quantity,
may
be expressed approximately
n
R
,
+
For larger values
value of
of
^/
a,
+~.
it
.
.
,
.
(4)
x
however,
Hence the omission of R n in (3) leaves n + 1 terms, the sum of which differs more and more from f(x) the more terms are taken.
n
R
may happen
increases.
that the
increases without limit as
n
In such a case the series
function.
(4)
cannot be taken to represent the
The determination
is,
of the values of
x
a for which
(4) is valid
in general, a matter involving a knowledge of mathematics which lies outside the limits of this course. In the illustrative examples
and in the problems
for the student
we
shall simply state the facts
in each case without proof. Formula (4) is known as Taylor s series.
Here a is a known value of x for which and its derivatives are known. The series f(x) then enables us to compute the value of f(x) for values of x not too remote from a.
58
DEFINITE INTEGRALS
a
Another convenient form of (4) is obtained by placing x whence x = a f h. We have, then,
= k,
f(a
+ h) =/() + hf
is
(a)
+
!/"()
+
/
+


.
:
(5)
Here the remainder
n
71
1
usually expressed as
+ 1f
n+
\a +6h),
where
&lt;
&lt;
1.
A
special
form
of (4) arises
when a
/"(0)
=
0.
We
/"
have, then,
/(*) =/(0)
This
is
+ */ (0) +
Maclaurin s
+
(0)
+
.
.
.
(6)
known
as
series,
x.
and the function
is
said to
be expanded into a power series in
Ex.
1.
ex
.
By
Maclaurin
s series (6),
we
find, since /(x)
1, etc.,
=
ex ,
/
(x)
=
ex ,
f"(x]
=
&,
etc.,
and/(0)
=
1,
/
(O)

1,
/"(O)
=
This expansion is valid for all values of x. If we place x~= ^, we have l + J T^ + y i^ + T ? = 1.3956, correct to four decimal places. If x has a larger value, more terms of the series must be taken in the computation,
e*
=1+
so that the series, while valid,
is
inconvenient for large values of
x.
Ex.
2.
since.
By Maclaurin
s series,
which
is
valid for all values of
x.
change 15 to circular measure, which is T^^TT = TL TT = .2618. Then the first two terms of the series give sin 15 = .2588, cor rect to four decimal places. By Taylor s series (5), we have
find sin
first
&lt;r
To
15, we
sin (a
+
h)
=
sin
a
+
h cos a
h2
7i
3
sin a
.
cos a
+
.
Ex.
3.
cosx.
s series,
By Maclaurin
x2
x4
OPERATIONS WITH POWER SERIES
Ex.
4.
59
(a
+
z).
s series,
By Maclaurin
This
to 0.
series
is
the binomial theorem.
polynomial of
n
is
+
1
terms, since
a positive integer, the expansion is a and all higher derivatives are equal f&lt;*+(x)
If
n
is
But
if
n
is
which
6.
valid
a negative integer or a fraction, the expansion when x is numerically less than a.
is
an
infinite
Ex.
logx.
The function logx cannot be expanded by Maclaurin s series, since its We may use Taylor s series (4), derivatives are infinite when x = 0.
placing a
=
1.
We
have, then,
Or we may expand log(l
+
x)
by Maclaurin
s series
with the result
This expansion
is
valid
when x
is
numerically
less
than
1.
Operations with power series. When a function is expressed power series, it may be integrated or differentiated by integrat or differentiating the series term by term. The new series will ing be valid for the same values of the variable for which the original series is valid. If the method is applied to a definite integral, the
32.
as a
limits
must be values for which the series is valid. Similarly, if two functions are each expressed by a power
sum, difference, product, or quotient
is
series,
their
the sum, the differ
ence, the product, or the quotient of the series.
Ex.
1.
Required
to
expand sin 1 x.
dx x
We
have
= sinix=
rx
Jo

=
x2
2
rx
I *^
Vl
(lx2 )~*dx
by Ex 4
"
24
1

24
&lt;
1
x3
3
x5
1.3.5
60
Ex.
2.
DEFINITE INTEGRALS
To
find
Jo
fV
cannot be found directly.
The
Ex.
1,
indefinite integral
31.
We may expand
e* by
2
Then
Ex.
3.
To expand
x)
3
By
division,
By
successive differentiation,
_
Therefore
(l
2 (1
(1
+ jc) 2 =  1 + 2x  3x 2 + 4x 3  5x 4 + x) 3 = 2  6 x + 12 x 2  20 x 3 + + x) 3 = 1  3x + 6 x 2  10x 3 +
.
.
Ex.
4.
To expand
lo,
1x
x)
=
J.
X
log(l
+

log(l

x)
Ex.
5.
To expand
1,
,
sin 1 x
Vlx
2
By
by Ex.
Ex.
sm l x = x
H
1
1
!;
+
16 x 7
.
4,
31,
(lx 2 )s=l +
^+
2x 3
^+
8x 5
Hence, by multiplication,
sinix
_
33.
By means
I,
of Taylor s
theorem we can complete the rule
the
given in
62, for the
maximum and
;
minimum
f"(a)
values of a
first
function of one variable.
Let a be a value of x for which the
i.e.
n
0,
derivatives of f(x) are zero n+l but
.
.
let
0.
f (a) =
0,
=
s
0,
f
=
"(a)
.,
/&lt;*&gt;()=
0,
f&lt;
\a)=
Then, by Taylor
f
theorem
(5),
31,
/(+
)f( a )
fc+i/&lt;+
\
n
+i
+
PROBLEMS
If
61
h
is
sufficiently small, the
term
/
I
will be larger
i~ L
1
?!/
&lt;
contains h n+z as a factor. numerically than R n + v since tlie latter Hence the sign of f(a + h)f(a) depends upon the sign of
n
"
If
n
is
even,
sign of
of h.
h.
n + 1 is odd, and the sign Hence the sign of f (a + h)
is
of
n
7*,
+1
f (a)
changes with the changes with that
Therefore f(a)
is
neither a
maximum
is
nor a
minimum
value
of /(a).
If
7i
odd,
n
+1
is
even and h n+l
is
always
positive.
Hence the
the same as the sign of / (n+1) (a). There sign oif(ali)f(a) fore if / (n+1) (a)&gt;0, f(a) is a minimum value of /(#); and if (n+l is a maximum value of/().
f
\a)&lt;Q,f(a)
PROBLEMS
Find the values of the following
definite integrals
:
is.
Jo
r
(a
2
+
x 2 )*
"
19.
f
t/
tj
Va 2 
x 2 dx.
20.
/a
fVx
I
2
a2 o!x.
21.
xsinxdx.
Jo
22.
f x2
Ji
logxa?x.
23.
24.
Jo
f x 2 sin!xdx.
OK
&lt;&u.
f x 2 ^^^ r 6 tix.
I
Jo
62
7T
DEFINITE INTEGRALS
W
sin 6
26. Prove
f
Jo
2
0d0=
f
Jo
2
sin 4 0d0.
77
E
sin
2 ~ xdx = n C x n l
27. Prove
Jo
f x
2
Jo
cos x dx,
if
n
&gt;
1.
28 Show that f
29
Ji
*
has a definite value when, and only when, k
&gt;
1.
Show
that
f* J a (6 
x)
k
has a definite value when, and only when, k
&lt;
1.
30. if
/(_ /(
x)
=
/(x), show graphically that
J
a
f(x)dx
=
2
0.
31. If
x)
=/(&), show graphically that Jf
=/(x), show graphically that
/(x)dx f(x)dx
=
=
Jf
f(x)dx.
32. If /(a
33.

x)
2
j^
x) dx.
*/(*)*
Show
graphically that
f
"/(sin
x)dx
=
k
^
"/(sin
6, and /i(x), values of x between a and
34. If a
&lt;
/2 (x), /8 (x)
are three functions such that, for
2 (x) &lt;/
&lt;/
all
& inclusive, /i(x)
3 (x),
prove
Ja
35. Find the
f /!(x)dx
&lt;
^
f /2 ()dx
&lt;
f*
f /8 (a;)dx.
mean
value of the lengths of the perpendiculars from a diameter
of a semicircle to the circumference.
36. Find the
mean value
of the ordinates of the curve y
=
sinx between
x
=
37.
and x
=
TT.
A
s
a sin
kt.
motion defined by the formula particle describes simple harmonic Show that the mean kinetic energy during a complete vibration is
kinetic energy.
half the
maximum
is
Obtain the following expansions. valid are given in each case.)
(The values of x for which the expansion
40.
,
=!+
, ,og
42. 43.
PROBLEMS
44. sinhz
63
(co
&lt;
=
x
+

+ +
r
+
+
r
+ +
.
z
&lt;
oc)
15
12
12
45. cosho;
46.
=l+
2
(co&lt;x&lt;oo)
[4
[6
Compute
the value of tanh ^ to four decimal places.
47. Given log 2 = 0.693, log 3 = 1.099,  0.693)? log 24 = log 2 + (1.099 48. Assuming sin 60
places.
what error would be made in assuming
=
^
V3 =
.8660, cos 60
=
^, find sin 61
to four decimal
Verify the following expansions
:
r \ x al ^ Jo + x b ol
I
_
.
_
11 ___
a
1
I
a
+
b
a
+ 26
a
+ 36
CHAPTER IV
APPLICATIONS TO GEOMETRY
34.
Element
of a definite integral.
In this and the subsequent
chapter
integral.
we
shall give certain practical applications of the definite Here we return in every case to the summation idea of
of
21.
The general method
is
handling one of the various problems
proposed
to analyze
it
into the limit of the
of
sum
of
an
infinite
the form f(x)dx. The expression number well as the concrete object it represents, is called the f(x)dx, as element of the sum.
of
infinitesimals
35.
Area
of a plane curve in Cartesian coordinates.
It
has
that the area bounded by the axis of x, already been shown ( 22) lines x = a and x = b (a&lt;b), and a portion of the the straight which lies above the axis of x is given by the defi curve
y=f(x)
nite integral
/*&gt;
Ua
\
ydx.
of the
(1)
It has also
been noted that either
be replaced by
of integration
x
= b may
a point in
bounding lines x = a or which the curve cuts OX.
of a rec
Here the element
y dx represents the area
dx and the altitude y. tangle with the base the area bounded by the axis of y, the straight lines Similarly, and a portion of the curve x =f(y) lying c and
y
=
y
=
d(c&lt;d),
to the right of the axis of y
is
given by the integral
*xdy
9
(2)
where the element x dy represents a rectangle with base x and
altitude dy.
Areas bounded in other ways than these are frequently found
by expressing the required area
areas of the above type.
as the
sum
or the difference of
64
AREA OF A PLANE CURVE
X2
2
65
Ex.
1.
Find the area of the
ellipse
+
?/
=
1.
It is evident
required area
if
A"
is
from the symmetry of the curve (fig. 18) that one fourth of the bounded by the axis of y, the axis of x, and the curve. Hence,
is
the total area of the ellipse,
,.1
a
ini _
a Jo
= ^06.
Ex.
2.
axis of x, the parabola y 2
straight line y
Find the area bounded by the = Ipx, and the
+
2x4j&gt;
=
1
(fig. 19). ,
..
The
straight line
tersect at
and the parabola in the point C(p, 2p), and the
F IG
is
.
18
straight line intersects
OX at B(2,p, 0).
the
The
figure
shows that the required area
if
sum
of
two areas
OCD
and
CBD. Hence,
K
is
the required area,
 2x)dx
FIG. 19
Ex.
3.
Find the area inclosed by the curve
?/,
(y
x
2
3)
=
4
x2
.
Solving the equation for
we have
y
=
x
+
3
V4 
x,
showing that the curve
It is clear
(fig.
20) lies
between the straight
lines x
s&gt;2
=
2
and x
=
x2
2.
from the
~
2
y&lt;zdx, I
figure that the area
.
A CEDE =
x 2 and
I
ACFDB =
where
/_2
y\
=
J
2/0
y^dx and the area
x
+
3
+ v4
=
2
x
+
3
v4
.
66
Therefore,
if
APPLICATIONS TO GEOMETRY
K
is
the required area
CEDFC,
y*)fa
K= JC
yidx
2
J
C y z dx= C
2
"
(y\
2
= =F
L
4
\/4
2
x 2 dx
x + 4 sini T
2J
= 4?r.
X&
2/
d# by
its
value f(x) taken from the equation of the curve. More generally, if the equation of the curve is in the parametric form, we replace both x and y by their values in terms of the independent parameter.
This is a substitution of a new variable, as explained in the limits must be correspondingly changed.
Ex.
4.
26,
and
Let the equations of the
x
ellipse
&lt;/&gt;,
be
=
a cos
y
=
b sin 0.
:
Then
the area
K of Ex. 1 may be computed as follows
ft
K=4l
Similarly,
m
JO O
ydx =
/^
o
4\
1/7T
=
n
4ab
I
JQ
if
r
=f(0) and the area sought
the equation of the curve is, in polar coordinates, is one of the above forms, we may place
and obtain thus a para
metric representation of the curve.
In case the axes of
reference are oblique, the
method
FIG. 21
of
finding
the
easily as follows
:
area
is
modified,
Let the axes
let
OX and OY
=a
us find the area bounded
and intersect at the angle (fig. 21) the axis of x, by the curve y =/(#),
&lt;w,
and the ordinates x
of the
and x
= b.
The area
is
sum
of the areas
of the parallelograms
evidently the limit whose sides are
AKEA OF A PLANE CURVE
parallel to
67
is
OX
?/Asina).
and OY. The area of one such paraUelogram Hence the required area is
sm
36.
c
ft&gt;
Ua
I
y dx.
Let
(fig.
(3)
Area
of
a plane curve in polar coordinates.
22) be
the pole,
OM
the initial line of a
system
of polar coordinates (r, 0),
OA and OB two
respectively,
fixed radii vectors for
which 6
=
a and 6
is
= /3
and
AB
any curve
for
which the equation
r =/(#).
Required the area AOB. The required area may be divided into
n smaller
areas
into
by dividing the angle
AOB=j3a Q
which equals lines
1}
n
equaj. parts,
each of
a
= A0,
)
and drawing the
OP OP OP
Z)
3
,
OP _^ where n
n
(In the figure required area is the
= A0.
=
8.)
The
the
sum
of
areas of these elementary areas for all values of n. The areas of these
FIG. 22
small figures, however, are no easier to find exactly than the given area, but
we may
find
them approxi
mately by describing from
as a center the circular arcs
AR
PR.
lt
Let
Then, by geometry,
the area of the sector
the area of the sector the area of the sector
AOR = J r A0, P^OR = J r* A0,
2
l
Z
P _ OE = n
l
n
J
r^^
The sum
of these areas,
namely
is
an approximation to the required area, and the limit of this sum
as
n
is
indefinitely increased is the required area.
68 To show
APPLICATIONS TO GEOMETRY
this
we need
to
show that the
area of each of the cir
from the corresponding elementary area (7^0P4)by an infinitesimal of higher order than either. For inter that purpose draw from P an arc of a circle with center
cular sectors
(e.g.
4)
J0LR
differs
secting
OP
3
in S.
Then
area
P OE^
3
&lt;
area
P OP
3
4
&lt;
area
80%,
area
P OP^
Z
area
area
But area
P OR= \ r A0 and
2 Z 3
area
SOP= \rl^Q = \ (r + Ar)
3 2 1
2
A0.
Therefore
area&gt;SO^
(r = ^^ + Ar) = / 1 H
3
,
Ar\
r.
2
:
I
Now
as
%
increases without limit,
limit.
A0 and

consequently
Ar
approach zero as a
Hence Lim
area
to
^p
= 1,
and there
OR differs from the area 3 OP^ by an infini Z order than either ( 2), and therefore the limit tesimal of higher OE i+l equals the limit of the sum of of the sum of such areas as such areas as J?0JJ*+1 as n is increased indefinitely ( 3). But the
Hence the
area
P
P
P
t
,
latter limit is the area
AOB,
since
= the
i
area
AOB
=Q
for all values of n.
Hence,
finally,
the area
AOB = Lim &rf&0 = n
"
J
C r*d6 =
/
\
C
J*
[f(0)] d0.
2
=*&gt;
i
=o
The student should compare this discussion with that of I, 189. The above result is unchanged if the point A coincides with 0, but in that case OA must be tangent to the curve. So also B may
coincide with 0.
VOLUME OF A SOLID OF REVOLUTION
Ex. Find the area of one loop of the curve r
69
=
a sin 3
(I,
177).
The required area
K
is
given by the equation
K= a Jo 7 f 2
To
integrate, place
2
gi
36
=
0;
then
/72
/&gt;
K=
37.
TT
f OJo
si
12
Volume
plane.
of a solid of revolution.
l&gt;y
a solid generated
in
its
the revolution
is
The simplest case
^4 sofoW of revolution is a plane figure about an axis of that in which the plane figure is
bounded by the axis of revolution, two straight lines at right angles to the axis, and a curve which does not cut the axis. Such a solid is bounded by two parallel plane bases which are circles and by
a surface of revolution generated by the revolving curve. Each point of this curve generates a circle whose center is in the axis of revolution and whose plane is parallel to the bases and perpen dicular to the axis. Consequently, if planes are passed perpendicular
to the axis, they will divide the solid into smaller solids with par allel circular bases. shall proceed to find an approximate
We
expression for the
volume
of
one of these smaller
solids.
23) be the revolving curve, the equation of which is x the axis of revolution, the line y =f(y), c, and the line y the volume generated by the revo d(c d). Eequired
(fig.
Let
KL
OY
CK
=
DL
=
&lt;
lution of the figure
parts,
s
CKLD.

Divide the line
CD
into
n equal
each of which equals
N ,. .,jru ^ where ON, = y v ON, = y v ON = yv ^ON _, = y _ r Pass planes through the points N N N N _ perpen

n
= Ay,
lt
by the points
3
N N
lt
2,
n
n

2,
3,
,
n
lt
dicular to OY.
circles
They
will intersect the surface of revolution in
NJ&gt;,
with the radii x v xv x2 = 2 P , xn _ lf where x l = Z xz = z p xni = ^n !?! The areas of these sections, begin ^ ning with the base of the solid, are therefore TTX^ TTX?, TTX^ ., ^i* where XQ = CK.
N
N
,
"
&gt;

The solid is now cut into n slices of altitude Ay. may consider the volume of each as to that of a cylinder with approximately equal
We
70
its
APPLICATIONS TO GEOMETRY
base coincident with the base of the slice and its altitude equal The sum of the volumes of the n cylinders is
to that of the slice.
and the limit of this sum as n is indefinitely increased is, by defi revolution. nition, the volume of the solid of is seen to be reasonable and in accordance with This definition the common conception of volume as follows. Whatever the defini
Y
r
N
I
IS
FIG. 23
tion of volume,
of the
it is
evident that the volume of a solid
is
the
sum
volumes
is
of its parts,
and the volume
of a solid inclosing
another
solid inclosed. greater than that of the be the volume of the slice Let then vol. 4 4 3 (for example) S
N N PP
P
of the plane figure 4 3 generated by the revolution and S S parallel to OY. Then evidently the lines
NN
^
and draw
PR
vol.
N N SE
9
A
&gt;
vol.
N^N.PR
3
&gt;
vol.
vol.
or
&gt;
vol.
N N SP
S
4
vol.
N N,SP
8
9
But
and
vol.
Z 4
vol.
N N PR = 7rN lf N N = irx^y =

4
4
3
4
TT
(x s
+ Aa?)*Ay.
VOLUME OF A SOLID OF REVOLUTION
Hence
N^N.P.R Lim __**
vol.
_
71
= Lim (#, + kxf = 1,
v
a
vol.
and consequently
Lim
vol.
^.
48
= 1.
vol.
Therefore the volume of the slice differs from that of the cylinder by an infinitesimal of higher order than that of either (2), and
therefore the limit of the
sum
same
as that of the
finally,
sum
of the
of the volumes of the slices is the volumes of the cylinders ( 3).
Hence,
vol.
CKLD = Lim ir^xf^y = TT
I
x*dy.
(1)
It is evident that the result is not invalidated
lies
if
either
L
or
K
on OY.
Similarly, the volume generated by revolving about OX a figure bounded by OX, two straight lines x = a and x = b(a I), and curve not crossing OX is any
&lt;
TT
f
Ja
fdx.
(2)
To evaluate either of these integrals it is of course necessary express x in terms of y, or y in terms of x, or both x and y terms of a new variable, from the equation of the curve. The volume of a solid generated by a plane figure of other
than that just handled
to
in
shape
or
may
often be found by taking the
sum
the difference of two such volumes as the foregoing.
Ex. Find the volume of the ring solid gen erated by revolving a circle of radius a about an axis in its plane 6 units from the center (b a).
&gt;
Take the axis of revolution as OF (fig. 24) and a line through the center as OX. Then the 2 2 2 equation of the circle is (x b) + y = a The volume required is the difference between the volume generated by CLEKD and that generated by CLFKD. But the volume generated by CLEKD is 7r where xi = b + vV2  y*, and the volume generated j_ X\dy
.
t
by
CLFKD
is TT
J
C* xfdy where
a
z2
=
b
 Va2 
y\
72
APPLICATIONS TO GEOMETRY
is
Therefore the required volume
a
TT
Ja
C P x fdyTT J* x$dy = v J ~
C"
(x*
a
xftdy
38.
Volume
of
a solid with parallel bases.
Fig.
25 represents a
is drawn perpen line solid with parallel bases. The straight at A, where y a, and dicular to the bases, cutting the lower base of x may be any line I. base at 7?, where y (The axis
OY
=
the upper
=
but it perpendicular to OY, not shown in the figure and
is
is
not needed in the discussion.) Let the line AB be divided into
n
to parts each equal


= Ay,
and
let planes be passed through each point of division parallel to the bases of the solid. Let A Q be the area of the lower base
of
the
solid,
A^ the area
of the
first section parallel to the base, A the area of the second sec
2
tion,
being the area of the section next below
and
so on,
An _
l
the
upper
base.
Then
^4
Ay
of a cyl represents the volume inder with base equal to A Q and volume of a cylinder represents the altitude equal to Ay, as a base and extending to the section standing on the next section next above, and so forth. It is clear that
A^y
4, Ay
is
+ ^ Ay + 4, Ay +
of
+ ^ n _iAy
the
solid,
an approximation to the volume
and that the limit
of the solid.
of this
sum
as
n
is the indefinitely increases
volume
That
this is rigorously true
may
to that of the previous article.
be shown by a discussion similar That is, the required volume is
Ua
Ady.
VOLUME OF A SOLID WITH PARALLEL BASES
To
73
terms of
variable.
find the value of this integral it is necessary to express in or both and y in terms of some other independent y,
A
A
This
is
for eacli solid.
The
a problem of geometry which must be solved solids of revolution are special cases. It is clear
that the previous discussion is valid if the upper base reduces to a point, i.e. if the solid simply touches a plane parallel to its base. Similarly, both bases may reduce to points.
Ex. 1. Two ellipses with equal major axes are placed with their equal axes coinciding and their planes perpendicular. A variable ellipse moves so that its center is on the common axis of the given ellipses, the plane of the moving
ellipse
being perpendicular to those of the given
ellipses.
Required the volume
of the solid generated.
FIG. 26
Let the given ellipses be ABA B (fig. 26) with semiaxes OA = a and OB = 6, ACA C with semiaxes OA = a and 0(7 = c, and let the common axis be OX. Let be one position of the moving ellipse with the center P where OP = x. Then if A is the area of
and
NMN M
NMN M
,
=
But from the
TT
PMPN.
(by
35,
Ex.
1)
ellipse
ABA B
72
+
b2
and from the
ellipse
A CA C

PN
j
a2
~~c?~
Therefore
be PM.PN=(a x
2
2
).
Consequently the required volume
7T&C
is
a a2
The
solid
is
called an ellipsoid
(
86, Ex. 5).
APPLICATIONS TO GEOMETRY
Ex.
angles.
2.
The axes
of
two equal right circular cylinders intersect at right
Required the volume
common
to the cylinders.
be the axes and Let (fig. 27) their common per of the cylinders, pendicular at their point of intersection 0, and a the radius of the base of each
OA
OB
OF
cylinder.
Then
the
figure
represents
one eighth of the required volume V. A plane passed perpendicular to OF intersects at a distance 0^ y from the solid in a square, of which one side
A
is
NP = VoP  ON =
2
2
a2

y
2
.
Therefore
and
39.
The prismoidal formula. The formula
F=
a simple and important result in those cases in which A can be not greater than 3. Let expressed as a polynomial in y of degree
r 4% leads to Ja
I
b
us place
A=
a
of the solid.
and take 0, for convenience, in the lower base if h is the altitude of the solid,
Then
= p
J
(a Q y*
+a
a 8 ) dy
If
now we
place
base,
B
for the area of the lower base, I for the area
of the
upper tween the bases,
and we have
M
for the area of the section
midway be
The formula
for
V then
becomes
V = (B6
This
is
known
as the prismoidal formula.
LENGTH OF A PLANE CURVE
To show
above.
its applicability to
75
a given solid,
that the area of a cross section of the solid
The most important and frequent quadratic polynomial in y. In this way the student may show that the formula applies to frustra of pyramids, prisms, wedges,
we need only to show may be expressed as cases are when A is a
cones, cylinders, spheres, or solids of revolution in which the gen erating curve is a portion of a conic with one axis parallel to the
and also to the complete solids just named. The formula takes its name, however, from its applicability to the solid called the prismoid, which we define as a solid having for its two ends dissimilar plane polygons with the same number of sides and the corresponding sides parallel, and for its lateral
axis of revolution,
faces trapezoids.*
Furthermore, the formula
two of whose whose lateral
is applicable to a more general solid, faces are plane polygons lying in parallel planes and faces are triangles with their vertices in the vertices
of these polygons.
Finally, if the number of sides of the polygons of the last defined solid is allowed to increase without limit, the solid goes
over into a solid whose bases are plane curves in parallel planes and whose curved surface is generated by a straight line which
touches each of the base curves.
applies (see
To such a
solid the
formula also
91, Ex.
is
5).
extensively used by in computing earthworks. engineers
40. Length of a plane curve in rec
The formula
tangular coordinates. To find the length of any curve AB (fig. 28), assume n l
%_ lt between A and B and connect each pair of consecutive points by a straight line. The length
points, JJ, 7J,
,
FIG. 28
of
AB
is
then defined as the limit of the
sum
n
is
n chords AP ly limit and the length
the
(I,
P^ P
of
2 P^&gt;
.,
J^^B
as
of the lengths of increased without
each chord approaches zero as a limit
104).
*
The
ih:it
which connects the
definition of the prismoid is variously given by different authors. solid most closely with the prism.
We
adopt
76
APPLICATIONS TO GEOMETRY
Let the coordinates of J? be
(x.,
y^ + Ay).
Then the length
of the
y ) and those of J?+1 be (x.+ Aa;, chord 7^+1 is V(Aa;) 2 + (A?/) 2 and
z
,
the length of AB is the limit of the sum of the lengths of the n chords as n increases indefinitely. But V(A) 2 + (Ay) 2 is an infini
tesimal which differs
infinitesimal of higher order.
from the infinitesimal Vcta? f dy* by an For
Lim
&gt;l
Now
if a; is
the independent variable,
the independent variable,
_
Lim

7
=1
Hence Lim
= 1.
Therefore
length of the curve, we may replace V(A^) Therefore if s is the length of AB, we have
2
_
^x
(4).
(
dx (4);
if a; is
not
Also Lim ~=.j.
3),
in
finding
the
.
2
+ (Ay)
2
2 1 by ^/dx + dy
y
,
(1)
where (J) and (5) denote the values of the independent variable for the points A and B respectively. If x is the independent variable, and the abscissas of A and B are a and b respectively, (1) becomes
d y\ = \c \^ + / dx \dx/ Ja N
)

( v
2)
v
If
are c
y is the independent variable, and the values and d respectively, (1) becomes
of
y at
A
and
jR
If a? and y are expressed in terms of an independent parameter t, and the values of t for A and B respectively are t and t lt (I) be comes / \ 2 TTTi
,&gt;/
/
7
LENGTH OF A PLANE CURVE
Ex.
point
1.
77
to the
Find the length of the parabola y 2
= 4px
from the vertex
(A, k).
Formula
(2)
gives
s
= =
I
\
I
Jo
s
\
r
k
 dx. c
.
Formula
(3) gives
 2p Jo
vy
8
Either integral leads to the result
Ex.
2.
Find the length of an
equation of the ellipse
axis,
ellipse.
x2
If the
is

7/
2
\

=
1,
and we measure the arc from the
end of the minor
we have
where
then
e
=
;
s=
4
,
1
ra
I
2
Jo
\
\
a2
 && ^ dx.  x2
Let us place x
the eccentricity of the ellipse.
=
asin&lt;/&gt;;
i
s
=
a f
t/O
2
Vl
e2 sin 2
d0.
The
We
(
therefore expand
31,
Ex. 4)
Vl
_
thus
e2
indefinite integral
Vl
i
cannot be found in terms of elementary functions. into a series by the binomial theorem e2 sin 2
sin 2
0= 1
e 2 sin 2
&lt;
This series converges for
2
all
values of 0, since e 2 sin 2
&lt;
1;
then
e* f sin 2 0d0e4 2 Jo 24 Jo
f\\n*&lt;t&gt;d&lt;t&gt;
i^e6
246
Jo
f sm60d0  \
J
2
The length
accuracy.
of the ellipse
may now
be computed to any required degree of
41. Length of a plane curve in polar coordinates.
8
The formula
= r
I
(/?)
,
Jdof+dy*
J(A)
of
40 may be transferred to polar coordinates by placing
x
= r cos 6,
y
= r sin 6.
APPLICATIONS TO GEOMETRY
Then
dx
= cos 6 dr r sin 6 dd, dy = sin 6 dr f r cos dO
}
and
Therefore
=/
/3
(1)
J(A)
If 6 is
the independent variable, and the values of 6 for
respectively, (1)
A
and
B
are a
and
becomes
I
C
ft
2
=1
If r is the
NT
+
U
A
(r
dr.
/dr
(2)
are
a and
b respectively, (1)
independent variable, and the values of r for becomes
1 f
and
dr
)
(3)
42.
The
differential of arc.
From
it
follows
(
9) that
ds
=
t
(1)
This relation between the differentials of x y, and s is often rep resented by the triangle of fig. 29. This figure is convenient as a
device for memorizing formula should be borne in mind that
rigorously equal to dy
rigorously equal to
ds.
(
(1),
but
is
it
RQ
nor
not
5),
In
fact,
PQ RQ = Ay
is
and PQ = As, but since Ay and As differ from dy and ds respectively by infinitesi mals of higher order, the use of RQ as dy and of PQ as ds is justified by the
theorems
used as a plane right triangle, we have an easy method of recalling the formulas
of
3.
If
now
the triangle of
fig.
29
is
dx
as
= cos
&lt;/&gt;,
AREA OF A SURFACE OF REVOLUTION
Similarly, from
s
79
=
u
2
,
we
find
is
ds=^dr +r2 d02
(2)
which
is the arc where suggested by the triangle of fig. 30, the arc of a circle with radius OP=r. This curve and of any used as a straightline figure, also gives the formulas figure, if
PQ
PR
tan
^=
rdO
&gt;
cos
dr
. .
Y = as
dr
sin
= rdd
ds
of revolu
is
43.
tion.
Area of a surface
A
AB
surface of revolution
a
FIG. 30
surface generated
of
by the revolution a plane curve around an axis in
its
plane
(
37).
Let the
curve
, P _^ n P^ and connect each pair of consecutive points by between A and B, a straight line. These lines are omitted in the figure since they are so nearly coincident with the arcs. The surface generated by
revolve about (fig. 31) surface generated, assume of the
OY
n
as an axis.
1 points, J^,
To
find the area
P
2,
AB
the
is
then defined as the limit of the
sum
by
as
of the areas of the surfaces generated
n chords
AP
lt
J?JJ,
PP
2
3
,
,
P_ B n
v
n
of
increases without limit
and the length
lateral sur
each chord approaches zero as a limit.
Each chord generates the
face of a frustum of a right circular cone,
X
of Pi
t
the area of which
may
mentary geometry. be (x it y ) and those of Jf+l be (x + A#, y + Ay). Then the has for the radius of the frustum of the cone generated by i+l for the radius of the lower base N^, and for base upper i+l lf +l) is therefore equal to its slant height JP i+r Its lateral area
i
t
be found by ele Let the coordinates
%P
N
T
1^PP
But
=x
80
APPLICATIONS TO GEOMETRY
Therefore the lateral area of the frustum of the cone equals
\
This
is
an infinitesimal which
differs
from
= 2 Trx.ds
by an infinitesimal of higher order, and therefore the area gener ated by A B is the limit of the sum of an infinite number of these terms. Hence, if we represent the required area by S we have y
,
/(
\
xds.
(1)
J(A)
To evaluate the
x, or express x
integral,
we must
either express ds in terms of
and ds in terms
of y, or express both
x and ds in
terms of some other independent variable. Similarly, the area generated by revolving
yds.
AB
about
OX is
(2)
XCB)
.
[&gt;
Ex. Find the area of the ring surface of the Ex.,
37.
The equation
of the generating curve
(x
is
2

2
6)
+
2
z/
=
a2
,
and
ds
= Jl + (^} dy = V /
the
a
2

dy.
2
LFK (fig.
The required area is 24). Hence
sum
of the areas generated
by the arcs
dii y
LEK &nd
Sy
=
27T
va(xi +
ra
CL
a;*)
v//2
_ 1/2
dy
=
fa
4 irab
I
"
= 4 7T2 a6.
0/2
a
&gt;Jri^.
PROBLEMS
1
.
2.
Find the area bounded by the axis of x and the parabola x 2  16 x + 4 y = 0. Find the area included between a parabola and the tangents at the ends
(The equation of the parabola referred to these tangents
(I,
of the latus rectum.
as axes
3.
is
x*
+
y*
=
a*
69).)
3
Find the
total area
bounded by the witch y
= ~7 2
2 ^~
and
its
asymptote
PROBLEMS
a Find the area bounded by the catenary y  (e 4 e h. lines x = the 2 4 = a 2& 2 x 2 4 2 5. Find the total area of the curve a y + &
81
n
f
x
\
j,
4.
the axis of x,
and
.
6.
2 2 Find the area bounded by the curve x i/
x,
7. Find the area bounded by the curve and the axis of y.
+a 2 62 = a 2 y 2 and its asymptotes. 2 2  ), the axis of 2 y(x + a = a (a
)
8.
units
Find the area of a segment of a from the center.
Find the area contained
circle of radius
a cut
off
by a chord h
9.
in the loop of the curve
2 2
y2
2
= X 2 (a its
x).
10.
11.
Find the area bounded by the curve y
(x
+
a
2
)
=
a x2 and
asymptotes.
Show
that the area bounded
by the hyperbola
is
^ ^
log f
I

=
1,
the axis of x,
and the diameter through PI (Xi, y t ) on the curve
Find the area bounded by the tractrix y the axis of x, and the axis of y.
12. 13.
+
^j.
y2
=  log *y
y
_
J"2

^ 2z
i
2
&gt;
x
= +
y
i
a(&lt;f&gt;
Find the area between the axis of x and one arch of the cycloid  sin0), y = a (I cos&lt;).
14. Find
the areas of each of the two portions into which the circle
is
X2
=
8
divided
2 by the parabola y

2x
2
15.
Find the area bounded by the parabola x
= 
0.
4y
=
and the straight
line
3x2y
 4=0.
bounded by the parabolas
y*
16. Find the area of the figure
=
18 x
and
2 17. Find the area between the parabola x
18. Find the total area of the lemniscate
8o3 = 4 ay and the witch y = 4 a8 ga r2 = 2 a2 cos 2 6.
,
and two
19. In the hyperbolic spiral r6 = a show that the area bounded by the spiral radii vectors is proportional to the difference of the length of the radii.
20. Find the area traversed
r
by the radius vector
of the spiral of
Archimedes
from
=
ad in the
first
revolution.
r=asec&lt;9
21
Find the area described by the radius vector of
22. Find the total area of the cardioid r
23. Find the area of the Iima5on r 24.
25.
26. 27.
+ cos0). + 6 when b a. Find the area bounded by the curves r = a cos 30 and r = Find the area of a loop of the curve r2 = a2 cos nO. Find the area of a loop of the curve r = a sin n0. Find the area of a loop of the curve r cos0 = a cos 20.
a(l
=
=
a cos0
&gt;
a.
82
APPLICATIONS TO GEOMETRY
=
2 2
)
by
28. Find the area of the loop of the curve r2 the initial line. 29. Find the total area of the curve (x 2
a2 cos 2
cos
which
is
bisected
+
y
=
+
4 a 2x 2
+
4 6 2 ?/ 2
.
(Transform (Transform
to polar coordinates.)
2 30. Find the area of the loop of the curve (x 2 3
?/
)
=
4 ax 2 y 2
.
to polar coordinates.)
31. Find the volume generated by revolving about OY the surface bounded by the coordinate axes and the curve x* + yl = ai 32. Find the volume of the solid generated by revolving about OF the plane surface bounded by OF and the hypocycloid x* + y$ a*.
33. Find the volume of the solid formed
by revolving about
OX the
OF
figure
bounded by the catenary y
=

(e
+
e~
),
the axis of x, and the lines x
=
h.
34. Find the volume of the solid
figure
bounded by the witch y
=
x2

formed by revolving about
the plane
+
and the
line
4 a2
y
=
a
35. Find the volume of the solid formed
figure
bounded by the
cissoid
2
?/
=
2a 
by revolving about
OX the
plane
,
the line x
=
x
a,
and the axis
of x
36. Find the volume of the solid generated by revolving about of the circle x2 + y 2 = a2 cut off by the chord x = h.
OF a segment
37. Find the volume of the solid generated by revolving a semicircle of radius a around an axis parallel to the boundary diameter of the semicircle, (1) when the arc of the semicircle is concave toward the axis; (2) when the arc is convex toward the axis.
38. Find the volume of the solid formed
by revolving an
ellipse
around
its
major
axis.
39. Find the volume of the ring surface formed
by revolving the
ellipse
a2
+
=
62
1
around OY(d
&gt;
a).
40. Find the volume of the solid generated
face bounded
by revolving about
line
OX the
OF the
=
sur
by the hyperbola
^L
=
i
and the
a2
62
x
=
a
+
h.
41. Find the volume of the solid generated by revolving about
face
sur
y.
bounded by the hyperbola
a2
62
=
1,
the lines y
=
h.
and the axis of
line
,
42. Find the volume of the solid formed the figure bounded by the curve y
line
by revolving about the
x=
y
a
= sin x,
the lines
and x = 
and the
y=
a.
43. Find the volume generated by revolving about the axis of x the figure bounded by the parabola y 2 = 4 px and the line x = h.
44. Find the volume of the solid formed by revolving about OF the bounded by the parabola y 2 = 4_px, the axis of y, and the line y=h.
figure
PROBLEMS
the figure bounded
83
a
h.
45. Find the volume of the solid formed by revolving about the line x 1 by that line, the parabola y = 4 px, and the lines y =
"by
46. Find the volume of the solid formed the figure bounded by the parabola y 2 47.
= Ipx and
revolving about the line x the line x = h(a h).
&gt;
=
a
A right circular cone has
its
vertex at the center of the sphere.
Find the
volume of the portion of the sphere intercepted by the cone.
48.
A steel band is placed around a cylindrical boiler. A cross
a semiellipse,
of the
its
section of the
axes being 6 and Vc in. respectively, the greater being parallel to the axis of the boiler. The diameter of the boiler is 48 in. What is
band
the
is
volume
band
?
r
=
49. Find the volume of the solid generated a(l + cos0) about the initial line.
by revolving the cardioid
50. Find the volume of the solid generated by revolving about bounded by the axes of coordinates and the tractrix. 51. Prove
OY the
figure
by the method
of
38 that the volume of a cone with any base
is
equal to the product of one third the altitude by the area of the base.
52. Prove that the volume of a right conoid is equal to one half the product and its altitude. (A conoid is a surface generated by a moving straight line which remains parallel to a fixed plane and intersects a fixed straight line.
of its base
moving line is perpendicular to the fixed line, the conoid is a right conoid. The base is then the section made by a plane parallel to the fixed line, and the altitude is the distance of the fixed line from the plane of the base.)
If the
53.
On
the double ordinate of the ellipse
x2
\
y2 
=1
an
a2
ft
2
isosceles triangle is
constructed with its altitude equal to the length of the ordinate. Find the volume generated as the triangle moves along the axis of the ellipse from vertex to vertex.
54. Find the volume cut
the center of the base
from a right circular cylinder by a plane through making an angle with the plane of the base.
55. Find the volume of the wedgeshaped solid cut from a right circular cylinder by two planes which pass through a diameter of the upper base and are tangent to the lower base. 56. Two circular cylinders with the same altitude have the upper base in common. Their other bases are tangent at the point where the perpendicular from the center of the upper base meets the plane of the lower bases. Find the volume common to the cylinders. 57.
Two
parabolas have a
common vertex and
a
common
axis but
lie
in per
moves with its center on the common axis, its plane perpendicular to the axis, and its vertices on the parabolas. Find the volume generated when the ellipse has moved to a distance h from the com
pendicular planes.
ellipse
An
mon
vertex of the parabolas.
58. cylinder passes through great circles of a sphere which are at right angles to each other. Find the common volume.
A
84
59.
APPLICATIONS TO GEOMETRY
A
variable circle moves so that one point
is
always on
OY
its
center
is
always on the
x2
ellipse
\
y2

=
a2 b2 Required the volume of the solid generated.
1,
and
its
plane
is
always perpendicular to OY.
60. Two equal fourcusped hypocycloids are placed with their planes per pendicular and the straight line joining two opposite cusps of one in coinci dence with a similar line in the other. A variable square moves with its plane
perpendicular to this line and
of the solid generated.
its
vertices in the
two curves.
Find the volume
61.
Two
equal ellipses are placed with their major axes in coincidence and
their planes perpendicular. straight line moves so as always to intersect the ellipses and to be parallel to a plane perpendicular to their common axis. Find
A
the volume inclosed
62.
variable rectangle other in the circle x 2 + y2 =
,
by the surface generated. moves so that one side has one end in OY and the a?. The ratio of the other side of the rectangle to the one mentioned is  and the plane of the rectangle is perpendicular to OF.
A
Required the volume of the solid generated.
63. The cap of a stone post is a solid of which every horizontal cross section a square. The corners of all the squares lie in a spherical surface of radius 8 in. with its center 4 in. above the plane of the base. Find the volume of the cap.
is
64. Find the length of the semicubical parabola y 2 the point for which x = h.
65. Find the length of the curve y
=
x 3 from the vertex to
=
e
log
ex
from x
I
=
1 to
x
=
2.
66. Find the total length of the fourcusped hypocycloid x 5 67. Find the length of the catenary from x
68. Find the length of the tractrix y
+
y*
= at.
=
a
to
x
=
h.
=
2
log
~
\b,
x
= htox =
a.
 Va 2 
x2
69. Find the entire length of the curve
70. 71.
fV + (\*=1.
(aj
(x
Find the length of the curve
2
y"
=
2 p) 3
from x
= 2p
=
a(0
to
x
=
h.
Find the length from cusp
cos0).
to cusp of
the cycloid x
y
=
sin0),
a(l
72. 73.
Find the length of the epicycloid from cusp
to cusp.
y
=
Find the length of the involute of a = to = 0i. a sin a0 cos from
circle
x
=
a cos
+ a0
sin 0,
74.
If the
From a
thread
is
spool of thread 2 in. in diameter three turns are
unwound.
held constantly tight, what
is
the length of the path described
by
its
end
?
of a suspension bridge hangs in the form of a parabola. The lowest point of the cable is 50 ft. above the water, and the points of suspension are 100 ft. above the water and 1000 ft. apart. Find the length of the cable. 75.
The cable
PROBLEMS
is
85
76. Show that the length of the logarithmic spiral between any two points proportional to the difference of the radii vectors of the points.
f\
77.
Find the length of the curve r
=
a cos 4  from the point in which the
f\
curve intersects the
78.
initial line to the pole.
Find the complete length of the curve r
a sin 8 .
3
Find the length of the cardioid r = a (I + cos0). 80. Find the area of a zone of a sphere bounded by the intersections of the sphere with two parallel planes at distances ^i and h from the center.
79.
2 portion of the parabola y
81. Find the area of the curved surface formed by revolving about = 4 px between x = and x = h.
OX the
82. Find the area of the curved surface of the catenoid formed
by revolving
h and x
about
OX the portion of the catenary y =
. ,
(e"
+e
a
)
between x
=
=
h.
83. Find the area of the surface formed
=
log 2 a
.
u.
a
a
u,
x2 fr + Va   v a2 2  x2  Va
f *
i*

by revolving about
OY the
tractrix
2
/
;
x2 .
84
Find the area of the surface formed by revolving about
OY the hypocycloid
x
by revolving an arch of the cycloid sin0), y = a (I a(0 cos0) about OX. 86. Find the surface area of the oblate spheroid formed by revolving an ellipse about its minor axis.
85. Find the area of the surface formed
=
87. Find the surface area of the prolate spheroid formed
ellipse
by revolving an
initial line
around
its
major
axis.
88. Find the area of the surface formed the cardioid r
89.
by revolving about the
=
a
(I
+
cos 6).
initial line
Find the area of the surface formed by revolving about the
the lemniscate r 2
=
2 a 2 cos 2
0.
CHAPTEE V
APPLICATIONS TO MECHANICS
44. Work. The application of the definite integral to determine the work done in moving a body in a straight line against a force 21. Problems for the in the same direction has been shown in
student will be found at the end of this chapter. The case of a body moving in a curve, or acted on by forces not in the direction of the motion, is treated in Chap. XIV.
45. Attraction.
Two
particles of matter of masses
m
x
and
m
2
respectively, separated
(YY)
by
a distance
is
r,
attract each other with a
force equal to
Jc
^
wyi
&gt;
where k
a constant which depends upon
the units of force, distance, and mass. units are so chosen that k = 1.
We
shall
assume that the
Let
of
it
now
mass
m
be required to find the attraction of a material body upon a particle of unit mass situated at a point A. Let
the body be divided into n elements, the mass of each of which may be represented by Am, and let J? be a point at which the mass of
one element
may
be considered as concentrated.
Then the
&gt;
attrac
tion of this element
on the particle
at
A
of
is
where r
cos
{
= P A,
L
and
6[
its
component in the direction
OX
is
it
where
the angle between the directions JF*A and OX. The whole body, therefore, exerts upon the particle at A an attraction whose
is
component in the direction
OX is
equal to
*
{
Lim
7i
= co
V
i=\
Am.
If
now
cos 6[., r
.,
and
Am may
\
be expressed in terms of a single
independent variable,
.
we
^
have, for this
COS
i
C7.
Lim
W=C
.
tr=l
V ^ Am = J
80
/

component, COS0 7 v^v/o v am, 5
ATTRACTION
where the limits
87
of integration are the values of the independent variable so chosen that the summation extends over the entire
body.
The manner
in
which
this
may
be done in some cases
is
illustrated in the following example.
Find the attraction of a uniform wire of length I and mass on a par unit mass situated in a straight line perpendicular to one end of the wire and at a distance a from it.
Ex.
ticle of
m
Let the wire OL (fig. 32) be placed in the axis of y with one end at the origin, and let the particle of unit mass be at A on the axis of x where A Divide OL into a.
n
parts, 03fi, 3fi3f2 , 3f2 3f3
,
,
3fn _ i i, each equal to

Ay.
Then,
if
p is the
mass per unit
3/4
,
of length of the wire, the mass of each element is pAy A?M. shall consider
=
We
//.
the mass of each element as concentrated at
this
its
/
,
^
,
first
point,
and
shall
in
way
obtain
an approximate ex
_
pression for the attraction due to the element, this approximation being the
better, the smaller
is
A
l
Ay.
The
attraction of the element 3/ 3/, + i on
A
is
then approximately
The component
of this attraction in the direction
OX is
and the component
in the direction
Y is
i
m sin
Then,
the total
if
rMir = (JAM
(a
2
X
is
the total
component
parallel to
component of the attraction OF, we have i=nl
parallel to
OX, and
Y
y
To
verify this
^Ay
2
=
)
l
+
2 3
2/
p
r
^o(
we may show
that the true attraction of
MiM
i
+l
,
differs
from
the approximate attraction, which we have used, by an infinitesimal of higher order. Let I be the true xcomponent of the attraction of 7X the approx { Mi + 1 imate attraction found by assuming that the whole mass of 31,3/i .V,. 1 is at
M
+
88
APPLICATIONS TO MECHANICS
and J2 the approximate attraction found by assuming that the whole mass of I /i; that is, 2 1. But is at Jfi + i. Then, evidently, I2
&lt; &lt;
&lt;
&lt;
M
*1
Ay

pato
,
and
(a
2
+
Vi 1
Ay )2] 
^= [
Jj
2
+
*t
J + Ay )2]!
Therefore
La2
4"
(Vi
Lim
=
and Lim
=
1.
Hence
Ii differs
from I by an
infinitesimal of higher
order ( 2), and may therefore be used in place of I in finding the limit of a sum (3). A similar discussion may be given for the ycomponent. To evaluate the integrals for X and F, place y = a tan 6. Then, if
a = tan x  = OAL.
a
X
since
If
cos a Jo
i I
dd
=
 sin a
P
a
at
7
sin
or,
P T=a
Zp
=
is
ra sin0d0 Jo
(1 av
cos a)
=
wi
:(! ar
cos a),
m.
the magnitude of the resultant attraction and
/3
R
the angle
which
its
line of action
makes with OX,
/3
=
tan !
,
X
F=
tan 1
t
1

cos
a
=
sin
a
 a. 2
1
46. Pressure. in a liquid at a
uniform depth
Consider a plane surface of area A^4 immersed The of h units below the surface.
of liquid of
submerged surface supports a column the weight of which is wh&A, where liquid) is the weight of a unit volume is the total pressure on the immersed
unit of area
is
volume h&A,
w
(a
of the liquid.
constant for a given This weight
surface.
The pressure per
then
which
is
independent of the
size of
AA We
may
accordingly
think of
&A
as infinitesimal
and define
wli as the pressure at a
point h units below the surface. By the laws of hydrostatics this pressure is exerted equally in all directions. may accord
We
ingly determine the pressure on plane surfaces which do not lie parallel to the surface of the liquid in the following manner. Let be in the surface of the liquid, and OH, for which (fig. 33)
OX
the positive direction is downward, be the axis of li. Consider a surface bounded by a portion of OH, two horizontal
ABCD
PRESSURE
lines
89
AD (h = a)
Divide
and BC(h
= b),
x=f(h).
AB
and a curve with the equation
into

n segments,
AMV
M^M^
MM
Z
S,
,
M
n
_l
B
)
each equal to
n
= AA,
and
draw a Hue
parallel
to
M^ through
OX.
Consider
i
each point
M
i
now
the ele
i
ment
of area
M
{
I?J?+ l
M + v where M^ = x
i+l
=/(/*,).
Its area is
{
tangle with base
line
between
is
M P and M
t
MM
{
equal to that of a rec
and
altitude
some
24),
i+l l*+1 (5,
and
&lt;
therefore equal to (xt
+ ej AA, where
c Ax. The pressure on the element would be wh (x + e.) AA, if all points of FIG. 33 =h the element were at the depth and would be w(k \ Ah)(x je^AA, if all points were at the depth Consequently, the pressure on J^Jf+1 i} 1 is plus an infinitesimal of higher order. Therefore the total pressure P on the area ABCD is
i
i
i
M
i
t,
{
i
M
i
M
_
P = Lim Y
==
=nl
s*b
s*b
=
/
^ajrfA
=w
I
/
7if(h)dh.
*/a
it
The modification
of this result necessary to adapt
to areas
is
of slightly different shapes is easily
made by the student and
total pressure
exemplified in the following example.
Ex. Find the
on a verti
cal circular area, a being the radius of the circle and 6 the depth of the center.
In 34 let OC =6, CA = CK = a, OM = h, MN = Ah. Then OA = b  a, OB = b + a, LK=2 MK= 2 ^a?(h  &)
fig.
2
,
and the pressure on the strip LKIG is 2wh\/a 2 (h 6) 2 AA, except for an infin
itesimal of higher order. Therefore the total pressure on the circle is
_a
Va 2 
(h
b
2
6)
Mh.
&lt;f&gt;
To
integrate, place h
JT
=
a sin
;
then
= 2w C ^
90
If the
APPLICATIONS TO MECHANICS
equation of the curve
c
CD
(fig.
33)
is
x =f(y), referred to
an origin at a depth OY perpendicular parallel and on ABCD may be shown to be
pressure
below the surface
the liquid, being to the surface of the liquid, the
of
OX
of A and B respectively. being the y coordinates the circle formula is used in .the example, the center of If this of coordinates, we have being taken as the origin
a and
I
=2w f
J
a
(6
47
m
of masses z lt Center of gravity. Consider n particles P (%2 y z ), P (x 8 y s ), , 8 at the points P^(x v y^ 2 , n placed form 35 ) respectively. The weights of these particles fi 20 ( gof parallel forces equal to a system
,
m m
m
,
,
,
Pi
p
where g is , n g, s g, 2 g, mtf, the acceleration due to gravity. The resultant of these forces is the total
weight
m
m
W of the n
&gt;
particles,
where
Hh
mg
n
ff
=a
m...
FIG. 35
This resultant acts in a line which about tion that the moment of
is is
determined by the condi
equal to the
W
sum
of the
moments
weights. that gravity acts parallel to OY, and thatthe line in a point the abscissa of which is x. Then cuts of action of and the moment of one of the is about the moment of f
of the
first
n
Suppose
W
g
OX
W
gx^ m
i
n weights Hence
is
m^.
g*^&gt;i=*9
to OX, the line of action of the Similarly, if gravity acts parallel in a point the ordinate of which is y, where resultant cuts
OF
CENTER OF GRAVITY
These two lines
of
91
of action intersect in the point G, the coordinates
which
are
Furthermore,
to either
OX
gravity acts in the or OY, the line of action of
if
XO Y plane,
its
but not parallel
through G.
This
may
be shown by
resultant always passes resolving the weight of eacli
and respectively, two components parallel to particle into the resultant of each set of components in the manner just finding
shown, and then combining these two resultants. plane, it may still If gravity acts in a direction not in the acts through G, but the proof requires be shown that its resultant
OX
Y
XOY
a knowledge of space geometry not yet given in this course.
of gravity of the n imrticles. If it is desired to find the center of gravity of a physical body, the problem may be formulated very roughly at first by saying that of an infinite number of particles of matter is made the
The point
G
is called the center
body
up
each with an infinitesimal weight hence the formulas for the co ordinates of G must be extended to the case in which n is infinite.
;
More
in question
solution of the problem is as follows. The body precisely, the is divided into n elementary portions such that the
weight of each within it. If m
may
is
be considered as concentrated at a point the total mass of the body, the mass of each
t ,
element
coordi may be represented by Aw. Then if (x ?/,) are the nates of the point at which the mass of the iih element is concen the equations trated, the center of gravity of the body is given by
x
= Lim and
i
y
= Lim
Am
can be expressed in forms of a single inde these values become pendent variable,
In case x it
y.^
I
Am
L_,
I
x dm
I
dm
y^dm
I
y
dm
(2)
92
APPLICATIONS TO MECHANICS
of integration are the values of the independent so chosen that the summation extends over the entire variable body.
It is to
where the limits
be noticed that
it
i
is
not necessary, nor indeed always
3,
possible, to
determine xi9 y exactly, since, by
Lim 2* (#i +
.)
Am = Lim V x
j3
ns
zero.
i
Am,
if e^
approaches zero as
Am
approaches
The manner in which the operation thus sketched may be car ried out in some cases is shown in the following articles. Discus sion of the most general cases, however, must be postponed until
after the subjects of double
and triple integration are taken up. 48. Center of gravity of a plane curve. When we speak of the center of gravity of a plane curve we are to think of the curve as
the axis of a rod, or wire, of uniform small cross section.
Let
(fig. 36) be such a curve, and let p be the amount of matter per unit of
LK
length of the wire which surrounds LK. That is, if s is the length of the wire
and
o
m
is its
mass,
we have
for a
homo
Am = = a p
As
for a
constant
;
and
nonhomogeneous wire
Lim T
.
Am = dm = = a  p
As
ds
.
.
.
function of
s.
The curve may now be divided
each being As and
applied by placing
I
into elements of arc, the length of
(2) of
its
mass Am, and formulas
47 may be
whence
x = I
pxds
I
&gt;
y=
I
py ds
p ds
(1)
pds
The
for
limits of integration are the values of the independent variable
L
and K.
CENTER OF GRAVITY OF A PLANE CURVE
If
93
p
is
constant, these formulas
become
sx
=
I
x ds,
sy=
\
y ds,
(2)
where
Ex.
s is
the length of
LK.
circle
1.
Find the center of gravity of a quarter circumference of the
a 2 which
,
x2
+
y
2
=
lies in
the
first
quadrant.
dy*
We
have
ds
= Vdz2 +
x ds
= dx =
I
dy.
/r /n
=
(I
Jm
a dy
a2
,
yds=
s
a dx
=
a2
,
I
and
=
ira
i
a quarter circumference.
Hence
x
=
y
= 2a
of the
The problem may
circle
also be solved
by using the parametric equations
7T
x
=
a cos 0, y
=
a sin
0.
Then
fx ds
a2 f
2
cos
d0
= =
a2
,
f y ds = a2 f
Therefore
x
a2
.
=
y
=
7T
,
as before.
Find the center of gravity of a quarter circumference of a circle when of matter in a unit of length is proportional to the length of the arc measured from one extremity. We have here p = ks, where k is constant. Therefore, if we use the para
Ex.
2.
the
amount
metric equations of the circle,
__
/r
pxds
I
_
sxds = = = J__ _ a
!
c
J_o
2
a 3 0cos0d0
_
(* TT

8)
a
b
/.Csyds

/;
2
ipyds
a
**
fpds
fsds
94
APPLICATIONS TO MECHANICS
49. Center of gravity of a plane area. By the center of gravity of a plane area we mean the center of gravity of a thin sheet of
matter having the plane area as its middle section. We shall first assume
is of the form of fig. 37, bounded by the axis of x, the being curve yf(%), and two ordinates. Divide the area into n elements by n l ordinates which divide AB into n parts each equal to A#. We de
that the area
note by p the mass per unit area. That is, if A is the area and the
m
mass,
we have
for a
homogeneous sheet of matter
Am = = a p
and
for a
constant
;
nonhomogeneous sheet
.Lim
T
.
of
matter
.
Am = dm = = a p
dA
,,
variable.
Hence
in (2),
47,
we
place
dm = p dA = py dx,
since for the area in question
dA = y dx
((1),
35).
Consider
now any one
of the
elements
MNQP,
where
lines
OM = x,
PR
&gt;
ON = x + Aa?, MP = y, NQ = y + Ay,
QS
parallel to
and draw the
and
OX. The mass
of the rectangle
MNRP
x

may
y\ }
be con
+ (A
Accordingly the mass of
at a point
x

)
and the
A?/\
(A^c
MNQP may be considered
2&gt;
#H
 vH
f

)
as concentrated
nates
MP
which lies above G v below G and between the ordi and NQ. The coordinates of any such point may be
(
expressed as
x
f
p&x,
+
q
where
)&gt;
0=^&gt;=1,
0=^ =
/ \
?/
1.
But the coordinates
an infinitesimal
of this point differ
from those
of
G
(x,
~) by
of the
same order
as A#.
CENTER OF GRAVITY OF A PLANE AREA
Therefore
centrated at
it is
95
as con
sufficient to consider the
mass
of
MNQP
become
G
(x,
Y
Hence formulas
(2),
47,
r
I
pydx
If
p
is
a constant, these formulas become
=
\xydx

\
,
I
(2)
\
I
y dx
ydx
which can be written
Ax = Cx dA,
If it is
Ay =
\y dA.
(3)
of gravity of a plane area of required to find the center that just discussed, the preceding method may be other shape than modified in a manner illustrated by Ex. 2.
If
the area has a line of symmetry, the center of gravity evi
lies
dently
upon
it,
and
if
this line is perpendicular to
one of the coordinates
of the center of gravity
may
OX or OF, be written
down
Ex.
at once.
Find the center of gravity of the area bounded by the parabola the axis of x, and the ordinate through a point (h, k) of (fig. 38), ?/ the curve. Here
1.
= 4px
Ax = f xydx =
Jo
r Ay =i Jo *
I
h
y^dx
2p
c Jo
t
and
A=
Jo
f ydx = 2p* Jo x*dx= ^p^A =
\
r!
/
/i
Therefore
x
=
;3
h,
y
=

fc.
96
Ex.
(fig.
APPLICATIONS TO MECHANICS
2.
Find the center of gravity of the segment of the
ellipse
a2
+
^=1
b2
by the chord through the positive ends of the axes of the curve. Divide the area into elements by lines parallel to OF. If we let y 2 be the ordinate of a point on the ellipse, and y l the ordinate of a point on the chord, we
have as the element of area
39) cut off
dA =
may be
(y z

2/1)
dx.
The mass
of this element
considered as concentrated at (x
\
r
Hence
n
x = ^
I
/.a
/ Jo
(yz
"*"

yi)xdx
y\)
(2/2

dx
y
FIG. 39
=
From
2/i
the equation of the ellipse y 2
=

b
"^a
2

x 2 and from that of the chord
,
=

M
(a
 x).
nO,
The denominator
ellipse
J
(y z
 y$ dx
is
equal to the area of the quadrant of the equal to
minus that
of a right triangle,
i.e. is

.
Hence
.
E_l\
2/
3(:r2)
50. Center of gravity of a solid or a surface of revolution of
constant density. volving about
OF
Consider a solid of revolution generated by re the plane area bounded by OF, a curve x =/(y),
Y. The solid may be resolved into elements by passing planes perpendicular to at a distance dy apart (37). The same planes divide the surface of revolution into elements ( 43). If the density of the solid is uniform, it is
and two
lines perpendicular to
OY
evident from the symmetry of the figure that the mass either of an element of the solid or of the surface may be considered as Y at the point where one of the planes which fixes the lying in element cuts OY.
CENTER OF GRAVITY OF A SOLID
Consequently, the center of gravity
lies in
97
O Y,
so that
Now
face,
if
V is
the volume of the solid and
S
the area of the sur
we
have, by
37 and
7TX
2
43,
dV=
dy
and
dS=2Trx ds.
we have
of gravity of the solid Consequently, to find the center 47 (2) by pjrtfdy with the result to replace dm of
\
piracy
dy
7rI
=TT\
j
x*y dy
(1)
pnrfdy
x*dy
and to find the center of gravity 47 (2) dm = have to place in 2 Trpx ds with the result
I
of the surface of revolution
we
2 Trpxy ds
I
xy ds
(2)
I
2 Trpx ds
x ds
I
Find the center of gravity segment of one base generated by revolving the area BDE OF. (fig. 40) about
Ex.
1.
of a spherical
Let
tion of the circle
OB = a and OE = c. The is x 2 + y 2 = a 2
equa
,
and
Ex.
of Ex.
2.
1.
Find the center of gravity of the surface of the spherical segment
1,
Using the notation and the figure of Ex.
we have
ds
=
,
and therefore
_
JI c
xyds
Jc
\
ydy
ft
Xa
The center
of gravity lies half
way between
E
and
7&gt;.
98
APPLICATIONS TO MECHANICS
of a
The pressures acting on the elements area form a system of parallel forces perpen submerged plane dicular to the plane. The resultant of these forces is the total
51. Center of pressure.
pressure
P( 46) and
To
the point at which
it
acts is the center of
find the center of pressure of an area of the type pressure. considered in 46, we may proceed in a manner analogous to that
used in finding the center of gravity of a plane area, only we have now to consider, instead of the weight of an element, the weight
of the
column
of liquid
which
it
sustains.
be the coordinates of the center of pressure. Then (x, h) the moment of the total pressure about OX is Ph, and the moment Also the pressure on an is Px. of the total pressure about
Let
OH
force equal to wh^Ah plus an 33) is infinitesimal of higher order ( 46). By symmetry this force acts at the middle point of the element, the coordinates of which are
element J^^. +1 Jf+1 ^
(fig.
a"
and h if except
Zi
for infinitesimals of higher order
(
49).
Hence
the
moment
of this force
about
OX is whfx^h
Hence
s*l)
and the moment
about
Now
of higher order. ^wli^f^li, except for infinitesimals the momentj of the resultant must be equal to the sum of the
ON is
moments
of the
component
t
forces.
1
=M
i
Ph = lAm^whfXi&h = w I n=a = Ja
&gt;
h x dh,
,rb
2
j
i
= nl
t
Px = Lim Vl wh x?&h =
n
= oo =
i
J
w
I
hx2 dh.
46.
Ja
Find the center of pressure of the circular area of the Ex., By symmetry it is evident that x = 0.
Ex.
Erom
the discussion just given,
ra
TT
= 2w
C
t/
2
7T
a 2 cos2
7T
(6
+
a sin 0) 2
d(f&gt;
(where h
7T
b
=
a sin 0)
7T
2
C cos 2 J _n
""
2
dtp
+
4 a s bw
J _1L
C
2
cos 2
sin
dQ
H
 w
2
4
2
\
sin 2 2
d&lt;}&gt;
v_2E
But
P = vaPbw
(Ex.,
46).
Therefore h
=
b
+
a2
.
46
PROBLEMS
PROBLEMS
1.
99
A
positive charge
m
of electricity
is
is
fixed at 0.
The
repulsion on a unit
charge at a distance x from
Find the
work done in bringing a unit charge
from
infinity to a distance
a from 0.
2.
A rod is stretched from its natural length a to the length a + x.
is
Assuming
that the force required in the stretching
3.
proportional to
,
find the
work done.
A =
piston
is
on the piston
ing pv
4.
fc,
Find the work
(2)
free to slide in a cylinder of cross section S. The force acting p is the pressure of the gas in the cylinder. as the volume of the cylinder changes from vi to u 2 (1) assum
is
equal to pS, where
,
assuming pvy
=
k,
y and k being constants.
of area.
to
bag of radius a contains gas at a pressure equal to p per unit Assuming that the pressure per unit of area is inversely proportional the volume occupied by the gas, show that the work required to compress the
& is 4 7ra 3p log .
A spherical
bag into a sphere of radius
5.
The
resistance offered
by any conductor
to the passage of a current of
in the con electricity is proportional to the distance traversed by the current ductor and inversely as the area of the cross section of the conductor. If a
source of electricity is applied to the entire interior surface of a cylindrical shell, and the current flows radially outward, what resistance will be encountered ? The length of the shell is h, the right circular section of the interior surface is
of radius a
and
of the exterior surface
is
is
of radius
6,
and a unit cube
k.
of the
substance of which the shell
made
offers
a resistance
6. A particle of unit mass is situated at a perpendicular distance c from the and length 2 I. Find the force center of a straight homogeneous wire of mass of attraction exerted in a direction at right angles to the wire.
M
7. Find the attraction of a uniform straight wire of mass upon a particle of unit mass situated in the line of direction of the wire at a distance c from one end.
M
8.
of unit
lines
Find the attraction of a uniform straight wire of mass upon a particle mass situated at a perpendicular distance c from the wire and so that drawn from the particle to the ends of the wire inclose an angle 6.
M
9.
upon a
10.
Find the attraction of a uniform circular wire of radius a and mass particle of unit mass situated at a distance c from the center of the ring
M
in a straight line perpendicular to the plane of the ring.
upon a
Find the attraction of a uniform circular disk of radius a and mass particle of unit mass situated at a perpendicular distance c from the center of the disk. (Divide the disk into concentric rings and use the result of Ex. 9.)
radius of
M
Find the attraction of a uniform right circular cylinder with mass Jf, its base a, and length Z, upon a particle of unit mass situated in the axis of the cylinder produced, at a distance c from ono end. (Divide the cylinder into parallel disks and use the result of Ex. 10.)
11.
100
12.
circle
APPLICATIONS TO MECHANICS
Find the attraction of a uniform wire of mass bent into an arc of a with radius a and angle a, upon a particle of unit mass at the center of
M
the circle.
13. Prove that the total pressure on a plane surface is equal to the pressure at the center of gravity multiplied by the area of the surface. 14. Find the total pressure on a vertical rectangle with base 6 and altitude a, submerged so that its upper edge is parallel to the surface of the liquid at a distance c from it.
Find the center of pressure of the rectangle in the previous example. Find the total pressure on a triangle of base 6 and altitude a, submerged so that the base is horizontal, the altitude vertical, and the vertex in the surface
15. 16. of the liquid. 17.
lies in
Show that the center of pressure of the triangle of the previous example the median three fourths of the distance from the vertex to the base.
Find the
total pressure
is
18.
so that the base
19.
lies in
on a triangle of base 6 and altitude a, submerged in the surface of the liquid and the altitude vertical.
Show
the
that the center of pressure of the triangle of the previous example
2 b and alti submerged so that the base is horizontal, the altitude vertical, and the vertex, which is above the base, at a distance c from the surface of the liquid. 21. A parabolic segment with base 26 and altitude a is submerged so that its base is horizontal, its axis vertical, and its vertex in the surface of the liquid. Find the total pressure.
median half way from the vertex to the base. 20. Find the total pressure on an isosceles triangle with base
a,
tude
22. Find the center of pressure of the parabolic segment of the previous
example.
23.
its
A
base
is
parabolic segment with base 26 and altitude a is submerged so that in the surface of the liquid and its altitude is vertical. Find the
total pressure.
24. Find the center of pressure of the parabolic segment of the previous
example.
25. Find the total pressure on a semiellipse submerged with one axis in the surface of the liquid and the other vertical.
26. Find the center of pressure of the ellipse of the previous example.
27.
An
immersed
is
8
ft.,
with its base horizontal and vertex downward is Find the pressure on the triangle if the length of the base the altitude 3 ft., and the depth of the vertex below the surface 5 ft.
isosceles triangle
in water*.
28.
The centerboard
sides are 1
is
of a yacht
two parallel
and 2
3
ft.
dicular to these two
is in the form of a trapezoid in which the respectively in length, and the side perpen in length. Assuming that the lastnamed side is
ft.
parallel to the surface of the water at a depth of 2 sides are vertical, find the pressure on the board.
*
ft.,
and that the
parallel
The weight
of a cubic foot of water
may
be taken as 62
Ib.
PROBLEMS
101
29. Find the moment of the force which tends to turn the centerboard of the previous example about the line of intersection of the plane of the board with the surface of the water.
is
30. Find the pressure on the centerboard of Ex. 28 if the plane of the board turned through an angle of 10 about its line of intersection with the surface
of the water.
31.
A dam
ft.
is
in the
form of a regular trapezoid with
its
two horizontal
sides
respectively, the longer side being at the top and the height 20 ft. .Assuming that the water is level with the top of the dam, find the total pressure.
300 and 100
32. Find the
moment
its
of the force
line.
which tends
to overturn the
dam of
Ex. 31
by turning
it
on
base
33. A circular water main has a diameter of 6 ft. One end is closed by a bulkhead and the other is connected with a reservoir in which the surface of the water is 100 ft. above the center of the bulkhead. Find the total pressure on the bulkhead.
pond of 10 ft. depth is crossed by a roadway with vertical sides. whose cross section is in the form of a parabolic segment with horizontal base on a level with the bottom of the pond, runs under the road.
34.
A
A
culvert,
Assuming that the base
its
of the
altitude 4
ft.,
find the total pressure
segment of the parabolic segment is 6 ft. and on the bulkhead which temporarily
closes the culvert.
35. Find the
x2
+
y
2
=
a2 which
center of gravity of the semicircumference of the circle is above the axis of x.
x*
+ y* =
36. Find the center of gravity of the arc of the fourcusped hypocycloid a* which is in the first quadrant.
x$
+ y* =
37. Find the center of gravity of the arc of the fourcusped a 3 which is above the axis of x.
hypocycloid
38. Find the center of gravity of the arc of the curve 9 ay 2  x (x between the ordinates x = and x = 3 a.
39. Find the center of gravity of the area bounded
 3 a) 2 =
by a parabola and a
chord perpendicular to the
parabola ay 2
axis.
40. Find the center of gravity of the area bounded = x s and any double ordinate.
by the semicubical
41. Find the center of gravity of the area of a quadrant of an ellipse. 42. Find the center of gravity of the area between the axes of coordinates and the parabola x* + y^ = a*.
43. Find the center of gravity of the area contained in the upper half of the loop of the curve ay2 = ax 2 xs
.
44.
Show
that the center of gravity of a sector of a circle
lies
on the
line
2 bisecting the angle of the sector at a distance a is the angle and a the radius of the sector. 3
a Sm ~9
.
a
from the vertex, where
102
APPLICATIONS TO MECHANICS
= sinx
e h
45. Find the center of gravity of the area bounded by the curve y and the axis of x between x = and x = TT. 46 If the area to the right of the axis of y between the curve y =
^
?
VTT
and the
axis of x
is
,
what
is
the abscissa of the center of gravity of this area
47. Find the center of gravity of a triangle.
2 48. Find the center of gravity of the area between the parabola y

= 4px
and the straight line y 49. Find the center of gravity of the plane area bounded by the two parabolas = 4 px, and x 2 = 4 py. 7/2 50. Find the center of gravity of the area bounded by the two parabolas _ 6) = 0, x2 4py = 0, the axis of y, and the line x = a. X2 _ 4_p
=
mx.
^
51. Find the center of gravity of the plane area 2 2 32 p 2 = 0. X 2 _ 4.py = o and the circle x + y 52. Find the
common
to the parabola
center of gravity of the surface bounded by the ellipse
a2
+
y
62
=i
the circle x 2
+
y*
=
a2
,
and the axis
of y.
53.
Find the center
of gravity of half a spherical solid of constant density.
54. Find the center of gravity of the portion of a spherical surface bounded from the center. by two parallel planes at a distance hi and h^ respectively 55. Find the center of gravity of the solid formed by revolving about OX 2 line x = a. the surface bounded by the parabola y = 4px, the axis of x, and the 56. Find the center of gravity of the solid formed by revolving about 2 = 4px, the axis of y, and the line y plane figure bounded by the parabola y
OF the
=
k.
the 57. Find the center of gravity of the solid generated by revolving about the parabola line x = a the surface bounded by that line, the axis of x, and 58. Find the center of gravity of the solid formed by revolving about bounded by the parabola x 2 = 4py and any straight line through 59. Find the center of gravity of the solid formed
OY
the surface
the vertex.
by revolving about
lines
OF
the surface bounded by the hyperbola
60. Find
~jp
=
l
and the
y=
and
y=b
the center of gravity of a hemispherical surface.
61. Find the center of gravity of the surface of a right circular cone. 62. Find the center of gravity of the surface of a hemisphere when the
density of each point in the surface varies as the circular base of the hemisphere.
its
perpendicular distance from
CHAPTEE VI
INTEGRATION OF RATIONAL FRACTIONS
52. Introduction.
The sum
each term of which
is
a rational fraction in
its
lowest terms with
is
the degree of the numerator less than that of the denominator,
known by elementary
.
algebra to be a fraction of the form
^\&gt;
J?
where
F(x)
(X)
and
= (a^x 4 \) = A^(a x + 1 f(x)
z
+
(A 3
+ b x + c^+A^x + \) (a x + B,} (ap + \) (a x + &
2)
(a s x
2
s
s
x2 + l z x + c 3 )
2
2 ).
Again, consider the
sum
Here the linear polynomial a^x + & x appears both in the first and the second powers as denominators of fractions which have the same form of numerator, a constant also the quadratic polynomial 2 a^x 4 b 3 x + c 3 appears both in the first and the second powers as
;
denominators of fractions which have the same form of numerator,
a linear polynomial.
If this
sum
(a,x
is
denoted by
bz)
&gt;
then
F(x)
= (a,x + ^)
2
+
(aj*
+ b,x + c,)\
and f(x), when determined, will be of lower degree than F(x). In both examples, f(x) and F(x) have no common factor and
f(x)
is of
lower degree than F(x).
103
104
INTEGRATION OF RATIONAL FRACTIONS
proceed in the following articles to consider, conversely, the
fix] ^~
We
possibility of separating
is of
any rational fraction
in
which f(x)
F(x)
we have
53.
lower degree than F(x) into a just added.
sum
of fractions of the types
Separation into partial fractions.
fix) J v
&gt;
Consider
now any rational
fraction
where f(x) and F(x)
If
are
two polynomials having
no common
the degree of f(x) is not less than that of F(x), we can separate the fraction, by actual division, into an inte of the numera gral expression and a fraction in which the degree
factor.
tor is less
than that of the denominator.
division,
For example, by actual
2 x*+ x*
3
+ x +x  18 x 2
6
_
x*+ x*+
3
Accordingly,
of f(x) is less
we
shall consider only the case in
of F(x).
which the degree
than that
is
Now
(I,
always equivalent to the product of linear factors F(x) which are not necessarily real and if the coefficients 42),
;
of F(x) are real, it is
equivalent to the product of real linear shall limit ourselves in this and quadratic factors (I, 45). to polynomials with real coefficients and shall assume chapter
We
that the real linear and quadratic factors of F(x) can be found.
We
shall
make two
Case
I,
cases
:
where no factor where some
is
repeated.
Case
II,
of the factors are repeated.
CASE
I.
As an example
F(x)= (a^x
of this case let
& t ) (a 2 x
+
+
& 2 ) (a 3 x*
+
lzx
+c
3 ).
May we
then assume, as suggested by the work of the previous
/(a)
article, that
=
A
,
A
,
F(x)
where
A lt A v A and B
3
,
z
are constants
?
SEPARATION INTO PARTIAL FRACTIONS
It is evident that the
105
sum
is
of the fractions in the righthand
member
of (1) is a fraction the
denominator of which
is
F(x) and
t
the numerator of which
a polynomial, which, like /(#),
is
of
lower degree than F(x). It will be proved in 55, 56 that A v A 2 A s and B s exist. As suming this, we may multiply both sides of (1) by F(x) with the
, , t
following result
:
f( x )
= A(
V+
+ (A
z
& 2)
(
a if*?+ \K + C s )+ A i( a i x +
3)
b i)
(
a ^*+
\ x + c a)
(2)
x
+B
(a,x
all
+ bj (,* + 6
values of
x,
2 ).
As
(2) is to
hold for
the coefficients of like
powers of
lefthand
x on the two
is
The righthand member
sides of the equation must be equal.* of degree three, and by hypothesis the
member
of degree
x*,
the coefficients of
x2
,
x,
no higher than three. Hence, placing and the constant term on the two sides
of the equation respectively equal,
we have
four equations from
2, 3,
which
to find the four
unknown
constants
A v A A By
,
Solving these equations and substituting the values of A lt A 2 A 3 and 3 in (1), we have the original fraction expressed as the sum of three fractions, the denominators of which are the factors
,
B
of the denominator of the original fraction. said to be separated into partial fractions.
The
fraction
is
now
way
It is evident that the
number
of the factors of
F(x) in ho
affects the reasoning or the conclusion,
and there
will always be
the same
number of equations as the constants to be determined.
* If the
number
of the
unknown
two members of the equation
_jz
+ an = b
,
xn
+
1
bix"
+
+ & M _ ia; + b n
(1)
are identical, so that (1) is true for all values of x, the coefficients of like powers of x on the two sides of (1) are equal, i.e. a Q 6 a 1 an i n Writing (1) in the equivalent form
=b
,
,
=b
.
(a

6
)
xn
+
(&lt;*!
1
&!)*"
+
4 (a B
_1
 &_!) x + (a n  &) = 0,
(2)
we have an
,=&
ai
algebraic equation of degree not greater than n, unless a
=b
,
a^
= bi,
Then (2) is true only for a certain number of values of x, since the number of roots of an algebraic equation is the same as the degree of the equation. But this is contrary to the hypothesis that (1), and therefore 6 (2), is true for all values of x. Hence a
= &i
=
&gt;
n
&n
& s was stated.
106
Ex.
1.
INTEGRATION OF RATIONAL FRACTIONS
Separate into partial fractions
(x
is less
.
+
3) (x
4)
Since the degree of the numerator
than the degree of the denominator,
we assume
..**_.(x
A
x
+
3) (x
2

4)

2
+
B
x
r&gt;
.
+
2
x
n C +3
2
where A, #, and C are constants. Clearing (1) of fractions by multiplying by
x2
or
(x f 3) (x
4),
we have
(2) (3)
x2
+ +
11 x
11
a;
+
14
14
+
= A (x + 2) (x + 3) + B(x  2) (x + 3) + C(x  2) (x + 2), = (A +B+ O)x 2 + (5 A + B)x + (QA 6J54C).
all
Since
(3) is to
hold for
values of x, the coefficients of like powers of x on
the two sides of the equation
must be equal.
Therefore
A + J5 + C = 1,
whence we
find
A
2,
B=
1,
and C
(1),
=
2
2.
Substituting these values in
we have
1
x2
(x
+ 11 x + 14 _ ~ + 3) (x 2  4)
2
2
x

2
x
+
x
+
3*
If the factors of the
denominator are
all linear
and
different, as in this
ex
ample, the following special method is of decided advantage. In (2) let x have in succession such a value as to make one of the factors of the denominator of
the original fraction zero,
When
comes whence
= 4 = C=
x
2, (2)
4
2.
,
= i.e. x = 2, x = 2, x becomes 40 = 20^4, whence A whence B = 1; and when x =
3.
=
2
;
when x =
becomes
2, (2)
be
5 C,
3, (2)
10
just used may seem to be invalid in that (2) apparently holds for all values of x except 2, 3, since these values make the multi 2, and
The method
plier (x
is
+
2
3) (x
4),
by which
(2)
was derived from
(1), zero.
This objection
(2)
met, however, by considering that the two polynomials in and therefore equal for all values of x, including the values 2,
are identical
2
and
3.
Ex.
x3
2.
I
4 x2
f
x
1 x3 Since the degree of the numerator is not less than that of the denominator, we divide until the degree of the remainder is less than the degree of the
Separate into partial fractions
divisor,
and thus
find
x3
1
=1+
1
x3
2
x
1
The
real factors of
x
3

are x
1
x
and x
+
x2
+
x
1.
Hence we assume
Bx+ C
x3

1

1
+
I
1
SEPARATION INTO PARTIAL FRACTIONS
Clearing of fractions,
107
we have
4x2
+
x
+ 1 = A (x 2 4 x + 1) + (Bx + C) (x  1) = (A + 5)x2 + (A  B + C)x + (4 powers of x
in (3),
4,
C).
(3)
Equating
coefficients of like
we
obtain the equations
A +B=
whence
A=
4 X2
2,
1
5=
2
2,
C=
2x +
X2
l
1.
Hence
+
X
+
X3
l
2
Xl
x x
2
+
X
+ +
1
and
f4x + x3  1
2x4
1

x2
x
+
1
The values of A, 5, and C may also be found by assuming arbitrary values Thus when x = 1, (3) becomes 6 = 3 A when x = 0, (3) becomes l = A C; and when x = 2, (3) becomes 19 = 1A+2B+C; whence A = 2, C = 1, B = 2.
of x.
;
54.
CASE
II.
We
will
now
2
1)
consider the case in which some of
the factors of the denominator F(x) are repeated.
For example,
2
3)
.
let
F(x)
= (ajc + &
(a z x
+ 6 (a^ + \x + c
2)
We
assume
+
\
,
+
2
,
+
ff+VHy
(i)
by the work of 52. Multiplying (1) by F(x), we have an equation of the 6th degree in x, since the degree of f(x) is, by hypothesis, less than that of F(x). & b 2 Equating the coefficients of x x x*, x*, x x, and the constant term on the two sides of the equation, we have seven equations from which to determine the seven unknown constants, A v A[,
as suggested
,
,
,
It is evident that, granted the existence of these constants, the above method for determining them is perfectly general.
108
Ex.
1.
INTEGRATION OF RATIONAL FRACTIONS
X4
Separate into partial fractions
_
6 x2
}
16
we
Since the degree of the numerator find by actual division
We now
assume
X*6X2
is
not
less
than that of the denominator,
+
16
2X2
2)(x24)
2
3*2
(x
+
2) (x
2

=
 A
7?
(z
+ 2)(x2 4)
C
1
1
4)
(x
+
2)2
x
+
1
2
x

2
(2)
Clearing of fractions,
we have
2)
2x 2 =
A (x (E
=
Equating the
+
C)x2
+ B(x 2  4) + C(x + 2) 2 + (A + 4 C)x + ( 2 A  4 B +
powers of
x,
4
Q.
(3)
coefficients of like
we
obtain the equations
B+C
2,
whence
J.
=
2,
B=
2x2
f
,
C=
.
Therefore substituting in
(2),
we have
2
2 (x + 2) (x
4)
(x
+
2)2
+
x
2
2
^
T
x

2
so that finally
x*
6x 2 +
16
_^
3
2)2
1
2
(x
(x+2)(x24)
Ex.
2.
+
2(x
+
2)
2(x2)
3 x7
Separate into partial fractions
+
6_421
By
division
we
first find
_
3
~2
We now
assume
X
2
x3
4x2 7x +
x3l2
4
B
3
Cx
1
2
(X

2
I)
(X

2
I)
X

(X
+
X
+D + I) 2
(2)
X2
Ex + F +X+1
and
clear of fractions.
The
result is
E +
B 2C + D E)x*
)
(3)
SEPARATION INTO PARTIAL FRACTIONS
Equating the
coefficients of like
109
powers of x
in (3),
we
obtain the equations
B+#=
0,
A+BE + F=0,
2A+B+C F=l, 3AB2C + DE =  4,
A
(4)
whence
,# =
(Xl)2
0,
C=
3,
D=
4,
E=
0,
F=
jj.
Substituting these values in
(2),
we have
3(X1)2
so that finally
2 (x 3

2
I)
2
2
3 (x 2

2
+
(x
3x
2
I)
+
x
+
a
l)
55. Proof of the possibility of separation into partial fractions.
we have assumed that the given fraction can be separated into partial fractions, and proceeding on this as sumption we have been able to determine the unknown constants which were assumed in the numerators. We will now give a proof
In the last two articles
that a fraction can always be broken the types assumed in 53, 54.
&gt;
up
into partial fractions of
f(x) Let the given fraction be J ^ where f(x) and F(x) are polyF(x) nomials having no common factor.
Let x
and F^(x) be the product = (x r^F^x) and
r be a linear factor of F(x) which occurs ra times, of the remaining factors. Then F(x)
F(x)
Now
the equation
(x
is
 r) m F^x)
A
(x
 r) m
(x
 r)
identically true,
being any constant.
110
If
INTEGKATION OF EATIOXAL FE ACTIONS
we can determine A
so that
1
f(r)~AF (r)=0,
then f(x) noted by (x
(3)
AF (x)
l
=
is divisible
by x
r
(I,
40) and may be de
r)fv (x).
t
But by hypothesis neither f(x) nor F^(x) is divisible by (xr) and hence f(r) and FJr) 0. Therefore, from (3),
=
a constant,
which
is
not zero.
With
this value of
A we
have
Applying
this
same method
i()
to
 ^_
yi r
y
&gt;
yi^^
,
n
we have
A
(f
/.(^
where
A=
;
and
be zero
however, that A l may be zero, since f^(r) cannot be infinite since F^(r) = 0. t times in succession, we have Applying this method
It is to be noted,
may
but
A
m
r)
A
m
A
m1
(x

A
(x
*
1
A
2
F(x)
(x
y
r)
x
r
r)"
F^x)
where
A Av A
z,
Am
are all finite constants, of
which
A
is
the
only one which cannot be zero. By the above reasoning it is evident that corresponding to any times we may linear factor of the denominator which occurs
m
assume
fractions, the numerators of which are constant, and the denominators of which are respectively the mth, the (m l)st,
m
1st powers of the factor. After these fractions have been removed, the remaining frac,
tion,
i.e.
f (x} J mV
&gt;
may
be treated in the same way.
FAx)
SEPARATION INTO PARTIAL FRACTIONS
111
In the above discussion r and the coefficients of f(x) and F(x) may be real or complex. Consequently, the method may be applied successively to each factor of F(x), thus making a complete separa
tion into partial fractions.
coefficients
If, however, /(#) and F(x) have real and we wish to confine ourselves to real polynomials, we apply the method to real linear factors only, and proceed in
will
the next article to deal with the quadratic factors. 56. Proceeding now to the case of a quadratic factor of F(x) of 2 the form (x +b 2 which cannot be separated into real linear a)
y
,
factors,let
Then
/()
the equation
/(*)
2
Now
 a) + tifF^x) [(x
is
 of + b [(x
and
Ax + B
f(x)(Ax + B)F
2
m
]
"*"
[(x
 a) + VfF^x)
2
l
(x)
identically true,
If
A
B
being any constants.
we can determine
f(a
A
and
B
so that
0,
0,
and
then f(x)
(I,
 U)  [A (a  bi) + B} F^a  U) = f(a
40),
+ U) [A(a + bi) + B} F^a + li) =
a (Ax + B)Fl (x) is divisible by x and hence is divisible by their product
2
2
U and x
2
a
+U
and
(x
a)
+b
2
,
we can
place
f(x)
 (Ax + B}F (x) = [(x  a) + I ]f,(x).
l
By
hypothesis neither f(x) nor F^x)
=
is
divisible
2
by
(x
a)
+
b
2
;
hence f(a
bi)
and
F^a
by
bi)
^ 0.
and
Doting
shall
have
&+
^i( a
P+
Qi)
^)
=*$F^ali)
Qi,
by
P  Qi,
we
and
where
A(a
bi)
+
J&gt;=P
P
and
Q
zero at the
same
are finite quantities, both of time.
which may not be
112
INTEGRATION OF RATIONAL FRACTIONS
aA
Therefore
+ B = P,
two equations from which A and values which cannot both be zero.
B
are found to have real finite
With
these values for
A
2
and
B
we have
f(x _ l
F(x)
Ax + B
 a) + &}** [(x
 of + tf} m [(x
f,(x)
l
FJx)
and repeating have finally
/(a) _
this process as in the case of the linear factor
we
F(x)
Ax + B  af + tf] m [(x
Ax+B  of + [(x
l
l
*"
fm (x)
time,
added that A and B may not be zero at the same and that any or all of the other constants may be zero. The same method may evidently be applied to each one of the
It should be
quadratic factors of F(x).
To sum
up,
if
F(x)
= (x
m r 1 ) (x
r
n
z)
[( a) + tf]
2
1
.
.".,
and we apply the above methods to the linear factors in succession and then to the quadratic factors in succession, we have finally
f(*)
f(x)
Cx + D _ a )*+ V]
2
+D [(x a) + 6
C^x
2
l
2
1
]
(x
a)
+b
2
where / is either zero or an integral expression in x. But if the degree of f(x) is less than that of F(x), and we
always reduce the fraction to this case by actual division,
shall
f(x\ z*
and
INTEGRATION OF RATIONAL FRACTIONS
all
113
the fractions on the righthand side of the equation are zero when x = co hence I is zero and the fraction is separated into
;
the partial fractions noted. 57. In the discussion of the last article the quadratic factors of the denominator are only those which represent the product of two
conjugate imaginary factors,
all
the real linear factors having been
factors may be previously removed. But some of these real linear in which case the algebraic work of the determination of the surd,
numerators
tor is of the
(
53, 54)
is
burdensome.
If,
however, the surd fac
b are rational, this
form x
a
Vb, where a and
:
work
I,
may
(1)
be avoided in the following manner It may be shown by a method similar to that used in
if
is
a ^/b F(x) has only rational coefficients, and x a + V& is also a factor, and hence that a factor of F(x), then x
44, 45, that
F(x) contains (x
(2)
If
2
of
n
b as a factor.
is
[(#
a)
b]
a factor of the denominator F(x), the
other factor being F^(x), then
Then the
rational fraction
f(x]
"^
may
be proved equal to
F(x)
[(a
_ of b]
,
n
[(x
 of by^F^x)
to
The
proof, being similar to that of the last article, is left
the student.
Accordingly,
if all
the coefficients of the denominator are rational,
and the surd
type just noted, the fraction will be separated into partial fractions, the denominators of which 2 n 2 2 n n shall be of the forms (x and [(x a) b] a) + & ] r) [(x
factors, if any, are of the
,
,
.
58. Integration of
rational fractions.
rational fraction in general consists of ration of the fraction into partial fractions
The integration of a two steps: (1) the sepa
;
the integration of each partial fraction, and the subsequent addition of the integrals. There will then be four types of integrals to consider
(2)
:
Adx
CAdx
f(Ax + B}dx
(*
(Ax+B}dx
114
INTEGRATION OF RATIONAL FRACTIONS
Turning
The first three, however, have already been discussed. then to the fourth, we may put that in the form
which
is
equal to the
sum
of the
two
2
integrals
A r
and
(
_n
+ a^)f
J
\(
x
a
,
The
first of
these integrals
is
readily seen to be
1
A
2{
n+1) "[(*)+ Mp.i
The second may be evaluated by a placing x
= I tan 6.
Then
/dx +y )
[ (a!
=
I
y
3,
When n =
case
n&gt;2
2, this integral is
evaluated as in Ex.
practice,
13.
The
occur,
rarely
occurs
in
but
if
it
does
cos*
n
2
C
gral
6d0 may be evaluated by methods
dx
of
65, or the inte
a)*+l*\* of the reduction formula
\(x
J
may
be evaluated b )7 successive applications
f du J (u*+c?r
which
Ex.
!
\
2( n
__
73.
8
,,
la&lt;u&gt;+&lt;t&gt;^
will be derived in
1.
Find the value of f jgJL * J z 3 + 2x 2
+
3)
&lt;*
z2
l)(x
7?
.
Since x*
+
2x2
xf
2
=
3
(x

1) (x
+
+
2),
we assume
8x
+
x
2x2


~
2
x1 + aT+T
^L
INTEGRATION OF RATIONAL FRACTIONS
Determining A,
x2
72,
115
we have
and
C by
the methods of the previous articles,
2
4 1
"
2 + 8 x 4 3 + 2z2 x2~xl x + 8s + 8)cto = rl 2 f (* J x 3 + 2x2 x2 J \xl
z3
3
x42*
2
3 3
\
dx
2/
x
+
l
x
+
=
2 log(x

1)
+
2 log(z
+
1)

31og(x
+
2)
+ C
T, * Ex. o T i .1 value of T 2. Find the J
i
( J
8ic
4
2 x2
{
4 Ox + 18)dx 8x 3 427
By
8 x4
division,
4
2 x 2 4
6x4 18
27 27 are
8x 3
+
= x4
3
2 x2
 21x4
18
(1)
8x 3 + 27
and 4x 2
1
The
assume
real factors of
8x 3 +
2x+
 6x +
^t
9.
Therefore,
we
(2)
~
8x 3
+
t 27
=
2x +
3
4x
iJ
6x +
3z
9
Determining A, B, and
Cby
the methods of the previous articles,
18 2
we have
(8)
2z 2 21x + 8x 3 + 27
"
2x +
3
4x 2
2
J
4 /(8x*42z 2 + 6x + 18) dx (8x + 2z 46z4 18)dx_ rl ~ 8x3 427 \
V
2x +
___
3 9 r 4 4J
6x +
9
3z 4x 2  6x4
9
But
/
Cxdx =
/
\x*, z
r
/
/x
4
&lt;5
and
3 X dx x
_
3 r (8 x

6)
dx
9
J 4x 2
6x + 9~
=
J 4 x 2  6x44?log(4x
2
+
dx
4x 2 _
6x +
9) 4
4
V3
finally
3
V3
Substituting the values of the integrals in
(4),
we have
r J
(8
"
x4
4
2
x2
4
6x
4
18)
dx
8x 3 427
= x 2 41 = x2
,
log(2x
+ 3)?log(4x 2 6x +
2x +
3
9)
4

tan*
4
*"!
8
4
C
V3
4
3
C.
V3
4log

^tan ,4z3
3
1

(4x
2
6x49)
4V3
3V3
116
INTEGEATION OF BATIONAL FRACTIONS
of
Ex.3. Find the value
C
J
+
X*
~
6 **
~
3
~ Ux +
54,
2(x l)2
as that of Ex. 2,
The fraction being the same the work completed.
Hence
/
we have
the
first
step of
(3a?
7
+
x6
J
 6x*8x 2 llx 3 2 2(x I)
J2
Now
4
rrtan 1
the
first,
r dx c
l
J3xl2 + Jr
2dx
(8x+4)dx
x2
+ x+l2
+ Jr
uated by
the second, the third, and the fifth integrals are readily eval8 1 2 previous methods, their values being respectively x 2 x, ,
9^01
423
,
(x
1)
and
3
V3
V3
^
There remains the fourth integral C J
X
2
^
,
(x
+
x
+
2J
I)
2 2
which may be reduced
dx
to
(x
2
+x+
2
I)
The
first
integral
is
\
o
2 (x2
,
+
f
x
+
while the second integral
may
be
1)
,
written in the form 40
2x
v
[(2
X
and can be evaluated by placing
1)2
(
3]2
+ = V3 tan 0.
1
The
result will be
40
J
2x
^
2
+
3
,=
"
3
V3
^
3
V3
3\/3
V3
l)
6 x2
+
x
+
l
Hence
J
C
(x
^++ ^ +
2
dx
l)2
=
___tl
2(x2
6
x
+x+
2x
+ J! ton i?fLt! 3V3 V3
6 x2
+
x
+
1
Finally, substituting the values of the integrals in (1) and simplifying, have, as the value of the original integral,
we
Vs
PROBLEMS
PROBLEMS
Separate the following fractions into partial fractions
1
f x  4 6x3 + 6x 2 6x*
:
117
a2
x2
x +
2x 3 + x 2
1Q
2
x*
+ 2x * + x
1
3x 2
4x
4x 8 + 8x 2
4 x3
x8
x2
12
"
3x 4
2x
x5
8
2
2x 2 l
5x
 llx1
x +
4
+
4
2
+ 2x 2
x4
3x2
x2
5x6 x3
+
7
x4
+ 3x 2 +
x3
x
2
3 x2
2
x3
x4
+
+
6
2
x
+
2
x4
+ 5x2
x4
2x
4 x2
x3
x2 12x8
x(x
+ x3 +x+

6x
+
8 8
+
2)3
x5 x3
2x4 4x3 + 8x2 + 4xx2
3
8x4
4x3 2x2 + 7x
Find the values of the following integrals
1 vj
:
C (14x J 4x 2 + r (3x
lg
+ 3)cZx 4x15 + 4)dx
/
6
J
/
x3

x
27
2g
(612x)dx
2
19
f
(8
g 10) dg
/^(3x
10x
3x
22
j ^
32xx
2
23.
8
24.
J
25.
p
9x 2 412x + a + 2x 2 + 3x
x2
2
2xl
dx
r2*
Hx6^ 6x
+ 17x 3 _3x 2 Ox B J 2x 3 + 7x2 f 2x3 x + + L rJ 4x 3 + 8x 2 9x18 + * r J 4x 3 4x 2 + x 2 (x + 2x + 33 J x 3 + Ox 2 + 1 + 10x + 9)d ^(fix2
/*6x 4
xxx xx
^
/.
J 4x 3
+ 8x 2 3x
118
,*
INTEGRATION OF RATIONAL FRACTIONS
(32z4z )dz
+ 12 z 2 +
9z
2
J 4 z3
45
2
r (15z 2
+
J
(3
z2
+
 29z  17)dz 2  2 z + 3) 2) (z
J z3
37 38
+2z2 4z8*
x4
J
(x
2
+3)(2z 2 +
x
z+
5)
J
9
+ 2z 3 +
ic
2
^

47.
J4x4
~ dx
"4^4"

/ ^6^ r^ 2 ^^ ^ 39. J X
2&lt;to

4x3
48.
+
x2
+
13x

8
dz.
/
r r 9z 3
z(zl) 3 (xl)3
16z4
T 2_
49
i
J
(z (
X2
2
+ 20z 2 + 21z_ 3) (3 2 + 3z + 3)(3z
1
8z2 +
6
dx.
l

4i.
r
r
I
xZ
+ Qx ~
2
/i*
51.

K
(to.
9 42.
(*
13z)dz
J
j ^
^z +
z
i)^
z4
44.
12z +  16
(a
8^
ab
52.
/z
4
+

53.
6)
z2
CHAPTEE
VII
SPECIAL METHODS OF INTEGRATION
59. Rationalization.
By
a suitable substitution of a
new
is
vari
able an irrational function
may
sometimes be made a rational
said
is
function of the
new
variable.
In this case the integrand
discuss in
to be rationalized, of Chap. VI.
and the integration
shall
We
this
performed by the methods
now
is
6063 some
of the
cases in
which
method
possible, together
with the appropri
ate substitution in each.
60. Integrand containing fractional
powers x can be rationalized by assuming
a
sions involving fractional
of
a
powers of a + bx. Expres + bx and integral powers of
+
bx
= zn
,
where n
For
if
is
the least
common denominator
then x
of the fractional
expo
Also,
,
nents of the binomial.
if
a
+ bx =
one
z",
= \ (z*  a]
x,
and dx
=z
b
n~l
dz.
(a
+
p bx) is
is
of the fractional
where pn
an
integer.
Since
a bx can all be expressed rationally in terms of z, it follows that the integrand will be a rational function of z when the substitution has been completed.
Ex.
1.
+
p powers of a + bx, (a + bx) = z pn and the fractional powers of dx,
Find the value of
let 1
J
C
(1
X * dx
.
+
2 3)*
Here we
+
2x
=
z*
;
then x
=
J
(z

1),
and dx
= *z z dz.
Therefore
C
J
(1
X * dx
=
J
+
2x)
? f 27 ( %J
_
2 z*
+
z)
dz
~
uitf
z&lt;2
(
5 z6
16 2 3
+
20)
+
C.
Replacing z by
its
value (1
+
3
2x)* and simplifying, we have
x*dx
^J
320
119
120
Ex.2.
SPECIAL METHODS OF INTEGRATION
Find the value of
&lt;*
+
2
&gt;*
&gt;
&lt;*
+
2
*)*
f
(x
+
2,
2)^
+
Since the least
common denominator
z*
;
of the exponents of the binomial
is 4,
we assume x +
integral
2
=
then x
=
z4
and dx
4z 3 dz.
On
substitution, the
becomes
=
4
/(
4
z3

 2z +
2
2)
3
  taniV V2
+
C.
Replacing z by
its
value (x
+
2)*,
we have
(x
+
x
2)*
=
+
2
+2  * (x +
2)3

4(x
+
2)4
+
8 (x
+
2)*
+ 8 log(Vx +
2
+
2)
V2
61. Integrand containing fractional
powers of a
+ bxn
.
If
the
integrand
is
the product
where q and r are
ization
is
integers, there are
two cases
in
which
rational
possible.
CASE
I.
When
n

is
an
a
integer or zero.
+
Let us assume
lxn
=z
r
and observe the
x
result of the substitution.
I
Then
i
H
=
1
(
z
ra)
n
,
and
dx
=
r
(sra)
rib*
z
r
~l
dz.
l
n
Therefore
x
=
r
sf&gt;
+ r l
sr
a
5ili n
dz.
But
if
n
rib*
is
an integer or
2,
zero, this
new
integrand
is
is
a
rational function of
ive one.
and the assumed substitution
an
effect
RATIONALIZATION
Ex.
1.
121
Find the value of
fx 5 (l + 2x 3 )*dx.
,
k
Since
The new
= 2 we assume 1 + 2 x 8 = z 2 whence x 3 =: (z 2  1), and x 2 dx=zdz.  z*)dz, which reduces to integral is ^ z 3 (3z 2 _ 5) + C
^
&V
.
Replacing z by
its
value,
we have
1)
yV(l +
CASE assume
II.
JFfte^

2x 3 )*dx = ?i5 (i + 2x 3 )*(3x 3 (
+
C.
*
^
^^
n
integer or zero.
n
Here we
will
+ bx = x
*
z
r
.
Then
=
and
Therefore
This
Ex.
2.
new
expression
is
a rational function of
x2 )
z.
Find the value of
f (2 +
J
dx
let 2
x2
Here
stitution
^~n~ +
we
=
r
lj
and accordin S lv we
+
x2
= x 22 2
,
.
After the sub
is
have, as the
new
integral,
2
f
Z * dz
the value of which
Replacing z by
its
value,
we have
62.
Integrand containing integral powers of ^/ a + bx + x2 If the integrand contains only integral powers of x and of Va + bx + or it may be rationalized by the substitution
.
2
_
,
V +
for
fa?
+x =zx
2
;
from this equation
The
result of the substitution is
evidently a rational function of
z.
122
SPECIAL METHODS OF INTEGRATION
Ex. Find the value of
2 x, we have as the new integral 2 Letting Vl + x + x = z which, by the method of Chap. VI, is
\
j
dz,
Replacing
z
by
its
value x
+ Vl +
x
+
x
x2
,
we have
as the value of the
original integral
1
f(
+
x
+
x 2) 2

3 (x
+ Vl +
2
+
x2 )
(
_
log
2x +
l
+
2
Vl +
x
+
x2)
+
C.
is any positive than unity, we may factor it out, thereby bringing constant other the expression under the case just discussed.
If the coefficient
of
x2 under the radical sign
63. Integrand containing integral powers of
Va +
bx
x*.
In
this case the substitution of the previous article fails, as
x could not
be expressed rationally in terms of
z.
a
+
Ix
x
=
a
,
2
,
.
(x
lx)
We may now, however, write a +W

an expression which can be factored into two linear factors ^ when the form (7^+ x) (r 2 x), which are real except
of
0.
But then
values of
it is
evident that
x2
is
(x
}
is
negative for
all
x,
and hence
a
+ Ix
always imaginary.
Now
place
Va +
bx
x2
= vVi +
x,
x) ( r 2
x)
= z (r
z
x}.
Solving this equation for
we have
After the substitution the
function of
z.
new
integrand
is
evidently a rational
TRIGONOMETKIC FUNCTIONS
Ex. Find the value of f
J
Since 2
123
dx

x

x2
=
(2
V2 The
to
=
+ x) (1  x), we assume z  z 2 = V(2 + z) (1  x) = z(lx).
is
result of the substitution
the integral 2
f/
,
which
is
za
2
equal
2
iog^ +c. Z+V2
its
2
Replacing z by
value
V2 + Z+V2 2z
64. Integration of trigonometric functions. There are certain types of trigonometric functions for which definite rules of pro
cedure
may
is
be stated.
The simplest
3
case
is
that in which the
integrand
any power of a trigonometric function multiplied by
I
its differential, e.g.
sin
x cos x dx, which
may
be integrated by one
of the
fundamental formulas.
It is evident that this class requires
no further consideration here. To deal with the more complex cases
usually necessary to make a trigonometric transformation of the integrand, the choice of the particular transformation being guided by the formulas for the differentials of the trigonometric
it is
functions.
65.
Some
of these transformations are given in
6570.
Integrals of the forms distinguish three cases.
lsm n xdxand.
lcos n xdx.
We may
CASE
place
I.
n an odd
sin
integer.
In the integral
I
sin
n
xdx we may
nl
xdx =
d(cos
sin
x sin x dx.
1
sin""
7
Now
is
sin
x dx
=
x),
and
a rational function of cos x since
n_ 
^
=
is
(1
cos ^)
integer.
2
2
,
which
i
an
2
Then
/
8ii\
n
xdx
=
I
(1
cos
In like manner
/P
we may
n
prove
nl
(1
I
cos
xdx=
sin
2
#
2
124
Ex.
1.
SPECIAL METHODS OF INTEGRATION
Find the value of f sin 5 xdx.
fsm^xdx =
C(l
+
cos2 x) 2 d(cosx)
cos 3 x
=
Ex.
2.
cos x
1 cos5 x
+
C.
Find the value of
f*L.
J COS 3 X
/dx x = J cos
3
C
(1
d (sin x)
I

s
To
sin 2 x) 2
integrate, place sin x
=
1
z.
Then
1 1
J
/_d(smx)_ = r
dz
(1
 Bin**)* (1
J
x
=
1
g
1
22)2
2
1
1
z2
4
^
&

g
z
+
_ ~
CASE
II.
1 sin
~
sin
x
2 cos 2 x
I
g
1
+
sinx
n a positive
even integer. In this case
we may
evaluate
the integral by transforming the integrand by the trigonometric formulas Sm2 = J (1 cos 2 x),
cos x
as
2
=
1(1
4
cos2#),
shown
Ex.
3.
in the following example.
Find the value of Ccos*xdx.
Applying the second formula above, we have

cos4 x dx
=
\
I
dx
+
^
cos 2 x dx

+
^
/
cos 2 2 x a?;,
Applying
this
formula a second time, we have
Tcos 2 2 x dx
=
^ C(l
+
cos 4 x) dx.
finally
Completing
all
the integrations,
we have
x
+
1 sin 2 x
+ ^L sin 4 x +
C.
CASE
sin
III.
n a
negative even integer.
In this case we replace
as in
68.
x by
CSCX
and cos x by
SQCX
I
and proceed
x
cos"
66. Integrals of the
form
sin
x dx. There
are
two
cases
in
which an
integral of this type
may
be readily evaluated.
If
CASE
is
I.
Either ra or n a positive odd integer.
1
"
m,
for example,
a positive odd integer, we place 1 m n ^ cos n #(sin x dx) sin # cos # dx =
sin"
ml
2
=
(1
cos #)
2
cos
n
# d (cos
x).
TRIGONOMETRIC FUNCTIONS
Since by the hypothesis
that the
125
it is

is
a positive integer,
evident
new
integral can readily be evaluated.
if
Similarly,
n
x.
is
an odd integer we express the integrand as a
function of sin
Ex.
1.
Find the value of C Vsinx cos8 xdx.
\
Vsinx cos 3 xdx
= f Vsinx(l
sin 2 x)d(sinx)
C.
Sometimes when one of the exponents is a negative integer the same method is applicable, but it is apt to lead to functions the
integration of
Ex.
2.
which
is
laborious.
Find the value of C J cosx
(where z
2x rsm 2 xd(s mx) r z 2dz  ~dx= I I 1 J cosx J J 1 1 sm 2 x =/ z 2
/"sin
=
sin x)
=  sm x +
CASE
II.
is
 log
,
1
1
1
+ sin x
:
2
 sm x +
C.
integrand of the last article and the additional formula
sin
Ex.
even integers. In this case the transformed into functions of 2 x by the two formulas
Both
m and n positive
x cos x
= \ sin 2 x.
3.
Find the value of f sin 2 x
sin 2 x cos4 x sin 2 xcos4 x
cos* xdx.
Placing
= = =
(sin
x cosx) 2 cos2 x,
2
we have
Therefore
^sin 2x(l
+
cos2x).
f sin 2 x
cos 4 x dx
2 J f sin 2 x dx +
f sin 2 2 x cos 2 x dx.
Applying the same method again, we have
2
Jsin
Completing
all
2 x dx
=
J*(l
we have
T*g.
cos 4 x) dx.
finally
the integrations,
2
ysin
x cos4 x dx
=
x
+ \ sin 3 2 x 
^
sin 4
x
+
C,
126
SPECIAL METHODS OF INTEGKATION
forms
I
67. Integrals of the
tan n * dx and
/
ctn"
x dx.
Since tan x
and ctn x are reciprocals of each other, we need consider only the case in which n is a positive quantity. Accordingly, if n is a
positive integer
we may
tan
w
a?
proceed as follows.
Placing
= tanw ~
2
2
x tan 2 ^,
and substituting
for tan
#
its
value in terms of SQCX,
"
we have
tan #
Therefore
I
n
= tanw
= =
n
I
2
2
^(sec
n~2
ic
1).
tann x dx
tan
x sec x dx
w 1
2
I
tan
~2
n
~*xdx
tan
I
x
J

tan"
x dx.
It is evident that the original integral
may
ctn"
be completely evalu
ated by successive applications of this method.
is
The same method
evidently applicable to the integral
Ex.
1.
I
x dx.
Find the value of
tan 5 a;
ftan*x&lt;tc.
Placing
=
tan 3 x tan 2 x
=
tan 3 x (sec 2 x
1),
we have
ftan 5 xd5x
= f tan 3 xsec 2 xdx =
4 J tan x
ftan 3 xdx
2
ftan 3 xdx.
Again, placing
tan 3 x
=
=
tan x (sec x
2

1),
we have
ftan xdx
3
= Ttanxsec
2 \ tan x
xcZx
J^tanxdx
+
log cosx
+
C.
Hence, by substitution,
Ttan 5 xdx
Ex.
2.
=
1 tan 4 x
4
2 ^ tan x
log cosx
+
C.
Find the value
of
ctn
2xdx.
Placing
ctn 4 2 x
= =
= = =
ctn 2 2 x (esc 2 2 x

1) ,
we have
fctn 4 2xdx
fctn 2 x
2
esc 2
2xdx
fctn 2 2 x dx
3 ^ ctn 2 x
fctn 2 2 x dx.
Again, placing
ctn 2 2 x
esc 2 2 x
1,
we have
fctn 2 2xdx
T(csc
2
2x  l}dx
x
=
Hence
Tctn 4 2 x dx
l ctn 2 x
3 J ctn 2 x
+
C.
=
+
^
ctn 2 x
+
x
+
C.
TRIGONOMETRIC FUNCTIONS
68. Integrals of the
127
forms
/
sec"
x dx and lcsc n xdx.
We
an
shall
consider these integrals only for the case in which
n
is
integer.
CASE
I.
If n
/
is
a positive even integer,
n
we
place
SQc
xdx =
I
/
where the integrand
is
a rational function of tan x, since


is
a positive integer. In the same manner,
csc
n
/r
xdx=
I
we may show
(1f
that
n
~2
2
ctii ^)
2
d(ctnx).
Ex.
1.
Find the value of fsec 4 3xdx.
j*sec*3xdx
= =
i f(l
+
tan 2 3x)d(tan3x)
(3
tan 3 x
+
tan 2 3 x)
+
C.
CASE
is
II.
If n
is
any
evident that the integral falls under the case of
replaced by
or
integer oilier than a positive even integer, it 65 when sec x
is
when
esc
x
is
replaced by
sin
,
maybe.
Ex.
2.
*
as the case
Find the value of Csecxdx.
fsecxdx= J cosx J
f
dX
=
J
cos 2 x
1
d (sin x)
1
sin x
**
}
2
.
1
sin 2 x
sin x\ 2
/
1 /I = rlog(+
2
\
cosx
 +C
+
C.
=
In like manner
it
log (sec x
+
tan x)
may
be shown that fcscxdx
=
log (esc x

ctnx)
+
C.
128
69
.
SPECIAL METHODS OF INTEGRATION
Integrals of the forms
I.
/
tan x sec n x dx and
777
/
ctn m x esc" x dx.
CASE
If n
is
a positive even
integer,
we may
2
write
1
tan m x SQcn x
w w ctn x csc x
dx=
dx
I
m tan #(l m ctn x (1
+ tan x)~d (tan x)
t
and
where 
/
=
I
+ ctn
2
x) ~*~d (ctn x),
is
Zi
a positive integer.
Ex.
1.
Find the value of f tan* 2 x sec 4 2 x dx.
4 Placing sec 2 x
=
sec 2 2 x sec 2 2 x,
we have
f tan* 2 x
sec 4
2 x dx
= =
^ f tan^ 2 x (1 i tan*2x
+
tan 2 2 x) d (tan 2 x)
C.
+
tan2x +
CASE
II.
If
m
is
a positive odd
integer,
1
"
we
1
place
/
tanm # secn x
dx=
1
i
(tan"
^
sec""
^) (tan
2
x sec x dx)
and
/
m ctn x cscn x
dx
=
where
m2
2.
1
/ml /ml
sec
n~1
2
ic(sec
^
1)
d(sec#),
""
1
I
(ctn
^
1
csc""
^) (ctn
a;
esc
# c?a;)
csc
w~1
^(csc ^
2
2
1)
.
is
a positive integer.
f
Ex.
Find the value of Ttan 3 3 x sec* 3 x dx.
ftan 3 3 x
sec*
3xdx=
2 f(tan 3 x
sec"
*
3 x) (tan 3 x sec 3 x dx)

I)d(sec3x)
C.
CASE
III.
If
,
m
.
is
an even
sin x = 
integer
and n
is
an odd
integer,
the integral
sec
x
=
1
may
be thrown under the cases of
cos x
66 by placing
and tan x
cos
x
TRIGONOMETRIC FUNCTIONS
70.
129
The
1
tan"
substitution
z,
tan=z.
if
The
substitution tan 
= z,
or
x
=
2
is
of considerable value in the integration of trigo
nometric functions, since
powers of the trigonometric functions, the result
tion is a rational function of
z.
the integrand involves only integral of this substitu
For,
if
tanthen
,,
.
x
= 2,
= =
.
sin
x
2 sin
x  cos x
2 2
1
=
1
i
2z
&gt;
+z
t
2
cos x
2 cos
2
y&gt;
=
&gt;
m
and
2tan
, tan
x
=
i
=
2,
&gt;
When
it is
these values for sin
x, cos
a?,
tan
x,
and d# are substituted,
is
evident, as stated above, that the result Find the value of
dx
2 cosx
a rational function.
Ex.
Placing tan ?
*
= z,
or x
= 2 tan&lt;**
1
z,
we have
cos x
=
1
+
Z2
,
and dx
=
1
2dz
+
Z
Therefore
/
f1
=_
+
2
f_^.
J
z2
2 cos x
2V3
tan
= i,log
l
+v^
+ C.
This method
t
is
applicable to an integral of
es
C J a
dx
h b
cos
x
C J a+
dx
b
sinx
C J a cos x + b sin x
any one dx
of the three
130
SPECIAL METHODS OF INTEGRATION
It
71. Algebraic reduction formulas.
was shown
in
is
61 that
a rational
zero or an
an integral
of the
form
,.
I
xm (a
j
.
+
n
number, can be rationalized
integer.
m+1
l^ydx, where p
or
if
m+1
n
\
p
is
In general, however, an integral of this type is evaluated use of the socalled reduction formulas, by means of which the by original integral is made to depend upon another integral in which
the power either of x or of the binomial a decreased.
+
lxn
is
increased or
The four reduction formulas
x m (a
are
:
+ lxn
(np
)
p dx
a
r
\
+ m + l)b
(np
b + m + 1)1 J
)
xm
n
(a
+ o^ydx,
(1)
Cxm (a
&*)
npa
r
,
np
(a
H
+
n
1
np + m + 1 J
+
lx
ydx
(m
(m
+ 1) a
1
n m l + np + + + T + l)fl J W(^
;
be verified by differentiation their deriva tion will be given in the next article. but in that case Formulas (1) and (2) fail if np \f 1 =
These formulas
may
m
;
we proved
Formula
Formula
in
61 that the integrand can be rationalized.
(3) fails if
m+1=
;
;
and in that case
also the inte
grand can be rationalized.
(4) fails
if
p+1=
and in that case
it is
evident that
60.
the integration
may
be performed by the method of
ALGEBRAIC REDUCTION FORMULAS
we
72. Proof of reduction formulas. ~ nbxn l dx, bxn ) note that d(a
131
71,
+
=
To derive formula (1), and accordingly place
x m (a
+ bx
n
)
p dx
=

(a
nb
+ bx
n
)
p nbxn  1
dx
and integrate by
~,m
parts, letting
n
4 1
nb
= u,
and
(a
+ bx
n
)
p nbx" ~ l
dx
v
= dv.
+l
Then
du
1 = mn+ xm ~ n dx,
,
and
=
nb
p+
1
As
a result,
fxm (a
*_mn+l C^ n(a
nb(p+l)J
To bring
this result into the required form, we place ~ ~ n xm n (a bx ) (a bxn ) p+l xm n (a
+
=
+
ixm  n (a
=a
I
xm
~n (a
+ btf) p dx + b
f
xm (a
+ bxn p dx.
)
Substituting this value in (1),
I
we have
xm (a\bxn ) p dx
=x
m  n+l
s
(a
+ bx
n
)
mn+l{ p+l
J
p
a
C
I
nb(p+l)
I
nb(p+l) \ J
It?)
^
"(a
+
bxn ) p dsc
+b
x m (a
+
dxlp
(2)
Solving (2) for fat* (a
If
+
bx")
dx,
we have formula
(1),
71.
we
solve the equation defining formula (1),
71, for
and then replace
m
n by m, the
result is formula (3),
71.
132
and xm dx
SPECIAL METHODS OF INTEGRATION
(2),
To derive
71,
we
integrate
= dv.
du
by
parts, letting (a
+ bxn p =u,
)
Then
l
= pnbxn ~ +
(a
+ bxn
p ~l
)
dx,
and
v
=
Then
/
xm (a
bxn ) p dx
m+
To bring
1
m+ 1J
form we place
this result into the required
=J
x m (a
+ lxn
n
p
)
axm (a
whence
r m+n
p
_l
J

ir
i
/v&gt;
m
/
i
hyf } P dx
1
a
C
I
x
(a
I
bxn } p
~
dx
t&gt;J
bj
(3),
Substituting this value in
we have
/m
np
["
r
a
m+
a
1
+ \J
1
r
I
V
Solving (4) for
If
xm (a
+ bxn
)
p  l dx
~]

(4)
J
J

x m (a
+
bxn ) p dx,
we have formula
71, for
(2),
71.
we
solve the equation denning (2),
1
r
I
x m (a + bxn ) p
71.
~l dx,
and replace p
73.
by p, the result
apply
these
is
formula
reduction
of
(4),
We
will
of
now
a
formulas can
the
to also
the
evaluation
few
integrals.
Many
the
these
of
be
evaluated
formulas.
by
substitution, without
use
reduction
ALGEBRAIC REDUCTION FORMULAS
Ex.
1.
133
Find the value of f x 3 Va2
(1),
+ x 2 dx.
the given integral depend
Applying formula
71,
we make
upon
f x Va
mulas.
2
f
x dx, an integral which can be evaluated by the elementary for
2
The work
is
:
fx
3
VoM^dx =
i
x 2 (a 2
+
x 2 )*
)

2 I a
J"
x (a 2
44
+
C.
x 2 )* dx
= I x 2 (a 2 + x 2  ^ a 2 (a 2 = TV (3 x 2  2 a 2 (a 2 + z 2 )*
)
x2 )*
+C
Ex.
2.
Find the value of C
71,
* x2
l
Va2  x2
a2
x2 )*
Applying
(3),
we have
dx r J x 2 Va2  x 2
Ex.
3.
_ x~
(a
2
~
_
/*
dx
2
^Va
 dx
^
_ _ Va 2 a 2x
x2
2
Find the value of f
J
x2
Applying
(3),
71,
we have
a2 a2 J
righthand
J
Applying
x2
(2),
2
71, to the integral in the
member
of (1),
we have
(2)
/(a
Applying
J

x 2 ) 3 dx
=
2 i x (a

x 2 )^
+
1
a2 /(a 2

x 2 )* dx.
(2),
2
71, again,
we have
f (a r
I
x2
)*
dx
= =
x 2 (a 2
.
x
x2 )*
C.
+  a 2 /f
2
A/ a 2
== _
,
(3)
X2
and
dx
sini 
J
Va2 _
X 2\i
x2
a
+
Substituting back,
we have
dx
finally
=
(^2
_
2
X 2~j
//(j2 x 2
2
x
C.
a2 a
dx
/C/T
Applying
(4),
,
where n
is
a positive integer.
71,
we have
/dxa +
2
=
2
)"
x(x
"
2
4
a2
+1
)~"
(x
2(n+
1 _i
l)a
T r
l(x2
2
+
2(x
n
)
2n43
r
dx
n+
,
2 2 l)a J (x
+
2(nl)a2
This formula
fractions.
is
+ a2
1
(2n
_
&gt;x
3)
r J (x 2 4a2
a2 dx
1
)"
1
)"
.
the one which
is
sometimes
used in integrating rational
134
Ex.
5.
SPECIAL METHODS OF INTEGRATION
Find the value
of
J V2 ax 
^
xdx
x2
If
we
becomes
ing
(1),
V2ax
71,
divide both numerator and denominator of the integrand by x*, it x* 1 a form to which we can apply a reduction formula. Apply
=
we have
f
V
x/O
""
=f,&gt;(2g
T
02
/
Ex.
6.
Find the value of f ^
x as
2
&lt;
(2
ax
a)
2
,
x2 ) 5
Writing 2 ax
apply
(4),
a2
(x
and noting that dx
is
= d(x
a),
we
can
71, to
advantage.
The
2
result
dX
J
f^ ax (2
~ 3
x2 ) 2
= f ^
[

(x

a) ]
2
MX
r
^ [a
.
2 = _(xa)[a (xan^ + 2
2
a
_
(x
_ a)Tidx
&gt;
x a2
a x2
V2 ax
74. Trigonometric reduction formulas.
It
was proved in
2
67
that
n
^^ =
71
1
tan"
^
Ct&n n
xdx.
~2
/tan
Similarly,
1
J
ctn n
~1
(1)
/ctn n xdx=
x
fctn"
x dx.
(2)
(1) and (2) are evidently reduction formulas for the of this particular type of integrand. There are four integration
Formulas
others
which we
shall derive,
i.e.
TRIGONOMETRIC REDUCTION FORMULAS
3in
135
m
*
cos"
* dx
1
sin
7"*
*
1
cos""
*
n
l
//
sin
m
*cos n *d*
sin m+1 *cos"
+
i
2
r
n
sin
J
sin
^
* n+2 xd*,
(4)
m
*cos"*
sin
7""
1
* cos n+1 *
m1
s
m m
and n are
m
m+n
/
sin
m
* cos n xdx
sin m+1 a;cos n+1 ^
xcosx
(6)
These formulas are useful when
tive or negative, or zero.
integers, either posi
(3) and (5) fail if + % = 0; but in that case the can be placed under integral or (2), by (1) expressing the integrand in terms of tan x or ctn x. Formula fails when =
Formulas
formula
(5) fails
when
m+1=
(4)
n
+
1
and
;
tion can be performed by the methods of To derive (3) we place sm m x cos" xdx
but in these cases the integra
66.
= cos n 
l
and
let
1
cos"
*
= u and sm m x cos xdx = dv for
n
m x(sm x
cos x dx)
As
a result
/
integration by parts.
cos
xdx
1
cos""
in m+1 *
*
nl
form we place
2
To bring
14 2
"
this into the required
sin"
*
2
cos"
*
2
= sin m *(lcos
I hen
*)
cos"
*=sin m *
2
cos"
*sin w *cos n *
J
sin
m
*
2
cos"
* dx
jsin
m
x cos n xdx.
136
SPECIAL METHODS OF INTEGRATION
(1),
Substituting this value in
I
we have
siu
mx
cos
n
xdx
=
smm+l xcos*~ l x
nl
1
m+
^
1
^
s* /
m H 1J
n
/
r. m sm x
co$
n
~2
xdx
(2)
m+ 1J
Solving (2) for
If
I
smm x
cos xdx.
sin
m
x
cos* x dx,
I
we
obtain formula
(3).
we
solve formula
for (3)
r sin w ^ cos n ~ 2 ,^^
and then replace
n by n + 2, the result is formula (4). The derivations of formulas (5) and
Ex.
1.
(6) are left to the student.
Find the value of fsin 3 x cos2 xdx.
(5),
By
formula
sin 3 x cos2 x

dx dx
=
=
^
sin 2 x cos 3 x
+
f
(
sin
% cos 2 xdx,
and

sin x cos2 x
integrals.
3 ^ cos x,
by the elementary
Therefore
I
sin 3 x cos 2 x
dx
=
1 3 2 y ^ cos x (3 sin x
f 2) f
C.
Ex.2. Find the value
Applying formula
/
of
fcos4 xcZx.
(3),
and noting that
I
m=
0,
we have
f
cos 2 xdx.
/
cos4 x dx
sinx cos 3 x
+
Applying formula

(3)
again,
we have
cos 2 x dx
=
^
sin x cos x
+
^
I
dx
=
^
sin
x cos x
f
^ x.
Therefore, by substitution,
/cos
75.
4x
dx
=
i sin x cos 3 x 4
+
sin x
cosx
+
ft
8
x
4
C.
Use of tables of integrals. In Chap. II we have evaluated simple integrals by bringing them under the fundamental formulas collected in 17, and in Chaps. VI and VII methods of
many
dealing with more complex integrals have been discussed. But in the solution of problems involving integration it is found that
PROBLEMS
some
integrals occur frequently.
137
If these integrals are tabulated
with the fundamental integrals, it is evident that the work of inte gration may be considerably Lightened by reference to such a table.
Accordingly the reader
a table of integrals.
table of integrals will be inserted here, but the reader is referred to Professor B. 0. Peirce s Short Table of Integrals."
"
is
advised to acquire facility in the use of
No
PROBLEMS
Find the values of the following integrals
:
14.
l.fjj*L.
2.
C^dx
^ x
r
I
15
*
&gt;
1 Vx +
r
16.
1
I
dx
t/
*/~T
dx.
/~
4.
J
n2)*(,g)_&gt; fa
x2i
5.
7.
x
 Vl
25.
12.
/
.
/(2
3x
2x2 ) 3
.
x 2 (3
+
x 8)
26.
r

sin*x cos 3 xdx.

J
138
SPECIAL METHODS OF INTEGRATION
+ a2
PROBLEMS
87. Find the area
139
X2
2
bounded by the hyperbola
=
V2 y
= I and the chord x = h.
a) (x
x).
2
88. Find the area of the loop of the curve cy
89. Find the total area of the curve a y 2
2
(x
6)
,
(a
&lt;
6).
=
x 3 (2 a

90. Find the area of the loop of the curve 16 a 4 ?/2
91. Find the area of a loop of the curve y 2 (a 2
=
)
92. Find the area of the loop of the curve (x
+ + y) 2
x2
62x 2 (a2  2 ax). = x2 (a2 x 2 = y 2 (y f 1).
).
93. Find the area inclosed by the fourcusped hypocycloid x*
+
y%
=
cfi.
94. Find the area inclosed by the curve
95. Find
the
W+W=
(

~
)
(

1
.
area included
between the cissoid y 2
=
2 a
x3
.

and
its
asymptote.
96. Find the area of the loop of the strophoid y 2
97. Find the area
x
=
x*
^ + x)
a x
bounded by the strophoid y2
and
a cos n9
)
its
asymptote,
excluding the area of the loop.
98. Find the area of a loop of the curve r
=
+
(
b sin nd.
99. Find the entire area bounded by the curve (
W +W=
)
1.
100. Find the area of the loop of the Folium of Descartes, x 3 by the use of polar coordinates.
+ y3
3 axy
= 0,
101. Find the length of the end of the first revolution.
spiral of
Archimedes, r
= ad,
from the pole
to the
102. Find the length of the curve 8 a?y
point x
=
x4
+
6 a2x2
from the origin
to the
=
2 a.
103. Find the volume of the solid formed by revolving about OX the figure bounded by OX and an arch of the cycloid x = a(0 sin0), y = a (I cos0). 104. Find the volume of the solid bounded by the surface formed by revolving the witch y
=
x2
8 a3

about
its
4 a2
asymptote.
of the cissoid y 2
105. Find the volume of the solid generated by revolving about the asymptote 3 = x the plane area bounded by the curve and the asymptote.
first
A right circular cylinder of radius a is intersected by two planes, the which is perpendicular to the axis of the cylinder, and the second of which makes an angle Q with the first. Find the volume of the portion of the cylinder included between these two planes, if their line of intersection is tan gent to the circle cut from the cylinder by the first plane.
106.
of
107.
An
ellipse
distance between which
and a parabola lie in two parallel horizontal planes, the is ft, and are situated so that a vertex of the ellipse is
vertically over the vertex of the parabola, the major axis of the ellipse being parallel to and in the same direction as the axis of the parabola. trapezoid,
A
140
SPECIAL METHODS OF INTEGRATION
having for its upper base a double ordinate of the ellipse and for its lower base a double ordinate of the parabola,generates a solid, whose upper base is the ellipse, by moving with its plane always perpendicular to the two parallel
planes.
Find the volume of the
solid, the
semiaxes of the
ellipse being
.
a and
6,
and the distance from the vertex
to the focus of the parabola being 
4
108. Find the center of gravity of the arc of the cycloid x
=
a(0
sin0),
y
a(l
cos0), between the
first
two
cusps.
first
109. Find the center of gravity of the plane surface bounded by the arch of the cycloid and the axis of x.
circles,
110. Find the center of gravity of the plane surface bounded by the two x 2 + ?/ 2 = a 2 and x2 + 2/ 2  2 ax = 0, and the axis of x.
111. Find the center of gravity of that part of the plane surface bounded by the fourcusped hypocycloid x$ + y$ a, which is in the first quadrant. 112. Find the center of gravity of the surface generated by the revolution about the initial line of one of the loops of the lemniscate r2 = 2 a 2 cos 26.
CHAPTER
VIII
INTEGRATION OF SIMPLE DIFFERENTIAL EQUATIONS
A differential equation is an equation which con Such an equation can be changed into one which contains differentials, and hence its name, but this change is usually
76. Definitions.
tains derivatives.
not desirable unless the equation contains the
first
derivative only.
of y equation containing x, y, with respect to x, is said to be solved or integrated when a relation between x and y, but not containing the derivatives, has been
A
differential
and derivatives
found, which,
it
if
substituted in the differential equation, reduces
to
an
identity.
tice
The manner in which differential equations can occur in prac and methods for their integration are illustrated in the two
:
following examples
Ex.
1.
to its intersection
Required a curve such that the length of the tangent from any point with OF is constant.
(fig. 41) be any point on the required the equation of the tangent at is
Let P(x, y}
curve.
Then
P
where (X, Y) are the variable coordinates of a moving
point of the tangent, (x, y) the constant coordinates of a fixed point on the tangent (the point of tangency),
Q
FIG. 41
and
is
derived from the, as yet unknown, equa
tion of the curve.
The coordinates
of U,
where the tangent intersects OF,
are then JT
=
0,
Y=y
x, and the length of dx
PR
is
\/z
2
\
Representing by a the constant length of the tangent, we
~ Xf
xz ( \dx have
+
+
,
X
Jdy\* =a
(dx)
or
dy
Va2  x2
(1)
141
142
which
is
SIMPLE DIFFERENTIAL EQUATIONS
the differential equation of the required curve.
Its solution is clearly
y=
=
r
J
^
1

x2
f
a 2
.
a a
log
T v a2 Va2 
x2 x2
+
C.
(2)
The arbitrary constant C shows that there are an infinite number of curves which satisfy the conditions of the problem. Assuming a fixed value for C, we see from (1) and (2) that the curve
is
symmetrical with respect to OF, that x 2 cannot be greater than a2
that
,
and y = C when x = a, dx dy becomes infinite as x and that
=
dx.
approaches zero.
From these facts and
the defining
property the curve is easily sketched, as shown in fig. 42. The curve is
called the tractrix
(I, p.
299).
y=0
X
uniform cable is sus pended from two fixed points. Re quired the curve in which it hangs.
Ex.
2.
A
Let
point,
A
and
(fig.
43)
be the lowest
any point on the re quired curve, and let PT be the
tangent at P.
Since the cable
is
P
in
equilibrium,
we may
forces,
consider the
portion
t
AP
as a rigid
body acted
the
on by three
at
tension
P
acting along
PT,
the ten
sion
h at
A
and the weight
tically.
acting horizontally, of acting ver
AP
is ps, where s is the uniform, the weight of of and p the weight of the cable per unit of length. Equating the length horizontal components of these forces, we have
Since the cable
is
AP
AP
t
cos
=
=
h,
and equating the
vertical components,
t
we have
sin
ps.

From
these
two equations we have
P
A
or
dy dx
~
FIG. 43
where P
=
a,
a constant.
DEFINITIONS
This equation contains three variables, with respect to x we have (I, 105, (4))
x, y,
143
and
s,
but by differentiating
(1)
the differential equation of the required path.
To
solve (1), place
ax
= p. Then
(1)
becomes
dx
dp
*

whence
log
(p
+ Vl + p 2 ) =
+
C.
(2)
Since ^4
is
the lowest point of the curve,
Hence, in
(2),
C=
0,
and we have
p + Vl + p 2 = e,
*
_
v
we know
that
when x =
0,
p =
0.
X
or
p=
(
Ve
+
_\
e
"J,
whence, since
= *,
dx
of
=
2
\e"
e~
"J
4
C
.
The value
x
C depends upon
if
=
0.
We
can,
we
wish, so take
the position of OJT, since y that OA = a. Then
=
x
a
+ C when
0,
OX
x
C =
and we
have, finally,
2/
=
a( Ve
+
*\
e
/,
the equation of the catenary
(I, p.
281).
The order
first
of a differential
equation
it.
is
derivative of the highest order in
Hence
equal to that of the (1), Ex. 1, is of the
first
order, and (1), Ex. 2, is of the second order. The simplest differential equation is that of the
order and
is
of the first
degree in the derivative, the general form of which
or
Mdx + Ndy =
Q
t
(1)
where
M and N are functions
of
x and
y, or constants.
144
SIMPLE DIFFEKENTIAL EQUATIONS
shall consider three cases in
We
solved.
1.
which
this equation is readily
They
are
:
2.
When the variables can be easily When M and N are homogeneous
degree.
separated.
functions of x and y of the
same
3.
When
the equation
is linear.
77.
The equation Mdx
+ Ndy =
(1),
when
the variables can be
separated. If the equation
76, is in the
form
it is
said that the variables are separated.
The solution
is
then
evidently
where c is an arbitrary constant. The variables can be separated if If and can each be factored into two factors, one of which is a function of x alone, and the other a function of y alone. The equation may then be divided which contains y multiplied by the factor of by the factor of which contains x.
N
M
N
Ex.
1.
dy=f(x}dx.
this follows
From
y
= C f(x) dx +
may
c.
Any
indefinite integral
be regarded as the solution of a differential
equation with separated variables.
Ex.2.
Vl 
2
7/
dx
+ Vl 
x 2 dy
=
0.
This equation rnay be written dx
dy
x2
1
a;
VI whence, by integration,
sin
VI + sin\f/
y
1
2
?/
=
c.
(1)
This solution can be put into another form, thus: Let sin 1 x = sin 1 ?/ = + = c, whence sin(0 + ^) = sine Equation (1) is then
\j/.
and
;
that
is,
sin
cos
\f/
f
cos
,
sin
\f/
^
cos
= Vl
In
x2 cos
= Vl
&,
where k
2
?y
;
is
a constant.
But
sin
=
x, sin
\f/
=
y,
hence we have
2
xVl^ + T/Vlx ^.
(1)
(2)
and
(2)
we have not two
of the differential equation. ential equation Mdx + Ndy
solutions, but two forms of the same solution, It is, in fact, an important theorem that the differ
=
constant.
The student must be prepared, however,
solution.
has only one solution involving an arbitrary to meet different forms of
the
same
THE HOMOGENEOUS EQUATION
Ex.
5.
145
(1

x2 )
^ + xy
(1
=
ax.
This
is
readily written as
x 2 ) dy
or
+ x (y dy ~ r
a) dx x dx

= _ =
0,
_

2
"\
whence, by integration,
log (y

a)
 %
log (1
x2 )
c,
which
and
is
the
same as
be written
log
this
may
The homogeneous equation Mdx + Ndy = 0. A polynomial sum of the expo in ,c and y is said to be homogeneous when the 2 letters in each term is the same. Thus ax + bxy + cf nents of those
78.
+ bx*y cxf ef homogeneous of the second degree, in such a polynomial, we homogeneous of the third degree. If, = vx, it becomes x"f(v) where n is the degree of the poly place y nomial. Thus as * + C = 3*(a + bv + cv*),
is
ax*
+
+
is
+ lxy y* ax* + lx*y + cxf + ef =x*(a +
lv
+
cv
2
+
ev
s
).
to This property enables us to extend the idea of homogeneity functions which are not polynomials. Representing by f(x, y) we shall say that/(ar, y) is a homogeneous a function of x and
y,
function of x and y of the nth degree,
j^ y^ = x
since
2
n
F(v).
Thus
W+
2
if is
vx, place y of the first degree, homogeneous
if,
when we
=
V^ + if = #Vl
log
4 v
2
,
and log x
v.
is
homogeneous
of
degree
0, since

= log v = x
log
When
M and N are homogeneous functions
Mdx
of the
same
degree,
the equation
is
said to
be homogeneous and can be solved as follows
:
Place y
tion
or
If /!
(i?)
= vx.
Then
becomes
^
z
dy = vdx + ^^
and
the differential
0,
0.
equa
^ ^^ + x = ^.^ ^ + = [A (v) 4 t/ (r)] ^ + xf (v) dv
a
^
^
t
(1)
+ vf
(v)
^
0,
this can be written
,
146
SIMPLE DIFFERENTIAL EQUATIONS
77.
z
where the variables are now separated and the equation may be
solved as in
If /! (v)
+ vf
)
(v)
= 0,
(1)
becomes dv =
0.
0,
whence v = c and y = ex.
Ex.
(x
2
Place y
 y2 = vx.
dx
+ 2 xy dy =
v 2 ) dx
There results
(1
+
2 v (x dv
^+^ = +
X
1
+
v dx)
=
0,
0.
V2
Integrating,
we have
logx
+
log (1
+
v2 )
2
whence
or
x(\ + v ) x2 + y 2
(aix
= = =
c
c,
,
ex.
79 The equation
+
biy
it
+
Ci)
dx
+
(a 2 x
+
b2 y
f c 2 )
dy
=
:
(1)
is
not homogeneous, but
Place
can usually be made
so, as
follows
k.
x
(1)
=x +
A,
y
=
y
+
(2)
Equation
becomes
(aix
+ +
x
&i2/
(agx"
+ aih + &ifc + ci) dx + & 2 y + a^h + &2^ +
so that
&ifc
c 2 )dy
=
0.
(3)
If,
now, we can determine h and k
+ +
(3)
b2 k
+ +
ci c2
= =
+
(H
0j
62 2/0
becomes
is
(aix
+
if
b\y
}
dx
+
2 &i
(a2x
in
0.
^
r
= 0,
which
k
is
homogeneous and can be solved as
cannot be solved
a\b 2
(1) is
78.
Now
(4)
In this case,
ttl
=
1
= k, where
(5)
some constant.
(aix
Equation
then of the form
[k
so that,
if
we
place
+ biy + ci) dx + dx + biy = x
,
fax
+
biy)
+
c2 ]
dy
=
0,
(x
+
+
which
is
dx
 ~
ci)
(5)
dx
(61
+ c2 kx/ + Cg  aik)x + &!Ci (
+
becomes ^X/
)
ai(Zx
=
0,
Ol
dx
0,
aic 2
and the variables are separated. Hence (1) can always be solved.
80.
The
linear equation of the first order.
The equation
(1)
!;
where f^ (x) and
y, is called
+ /,(*)*=/,&lt;*),
first order.
2
/
2
(x)
may
reduce to constants but cannot contain
a linear equation of the
Q,
of
Mdx + Ndy =
where
M=f (x)yf
1
(x),
N=
It is a special case
1.
LINEAR EQUATION OF THE FIRST ORDER
To
solve the equation
147
we
will try the experiment of placing
y
= uv,
later
where u and v are unknown functions of x to be determined in any way which may be advantageous. Then (1) becomes
dv
du
f
,
.
_
av dv
,
.
.
au
or
\
,
t
\
v\
+/i(J5)^
+ ^ dx =/
i
P
i
2 (
a5
\ )
/o\ 2
( )
Let us
be zero.
now
deter mine
u
so that the coefficient of v in (2) shall

We
have
or
U
which the general solution
is
of
I
f
l
(x)dx
=
c.
Since, however, all
we need
is
a particular function
which
c
will
0.
make
the coefficient of v in (2)
equal to zero,
I
we may take
=
Then
log
u=
u
/, (x) dx,
or
=e
J
(3)
With
this value of u, (2)
becomes
and
Wlience, finally, since y
I
v
=
l
f
z
(x)dx
+
c.
(4)
= wv,
j
/*
ff^d
(5)
148
Ex.
(1
SIMPLE DIFFERENTIAL EQUATIONS
x2 )
^+
dx
x?/
=
ax.
Rewriting this equation as
dy
dx
r
+
,
x
T
y 1&*
t
 =
ax
1x2
we
recognize a linear equation in which
Then
Hence
f/i (x)dx
=y
log
Vl x2
z2
,
and
eJ
fl(x)&lt;lf
=
= Vl 
J
f
(1
2.
eZx
+ cVlx 2
_
e
 log
VT^ _
1
Vl 
x2
x2 )*
=
This example
is
a
+ Vl
c
x
3, 77, showing that the methods of solving an equation are not always mutually exclusive.
the same as Ex.
81.
Bernoulli! s equation.
The equation
while not linear, can be
+ /i (x) y = /2 (x) y n
made we have
so, as
,
follows
:
Dividing by
?/",
and placing y l ~ n
= z,
and multiplying by
1
n,
we have
a linear equation.
That
is,
y4_, 3 dx x
/
I
z,
x2
or
=
x
3^2
where
z
=
dx
y~ s
.
We
have now
JCfi(x}dx
= f J
= =
x*.
=
x
Iogx3,
whence
Jf^
dx
Hence
and
z
x3 J
C( 3x 5 )dx +
.
,
x8
v3
= x +
2
x3
CERTAIN EQUATIONS OF THE SECOND ORDER 149
82. Certain equations of the second order. There are certain equations of the second order, occurring frequently in practice, which are readily integrated. These are of the four types
:
We
&gt;
proceed to discuss these four types in order
:
3 A*
direct integration
By
This method
is
equally applicable to the equation
^
Ctd}
=/(#).
Ex. 1. Differential equations of this type appear in the theory of the bend ing of beams. Each of the forces which act on the beam, such as the loads and the reactions at the supports, has a moment about any cross section of the beam equal to the product of the force and the distance of its point of application from the section. The sum of these moments for all forces on one side of a given section is called the bending moment at the section. On the other hand, it is shown in
the theory of
beams that the bending moment
is
equal to
I,
its
,
where E, the
of inertia
elasticity of the material of the beam, of the cross section about a horizontal line
modulus of
and
the
moment
and
R is
through
the radius of curvature of the curve into which
center, are constants, the beam is bent Now
by
I,
191,
[
where the axis of x
is
O
in
horizontal.
But
most cases arising
:
in practice
~y
is
very small, and if we expand 
by the binomial theorem, thus
150
we may
SIMPLE DIFFERENTIAL EQUATIONS
neglect all terms except the
is
first
without sensible error.
Hence the
bending moment
taken to be
EI^
CIX
This expression equated to the bending
moment as defined above gives
the differential equation of the shape of the beam.
We
will apply this to find the shape of a
beam uniformly loaded and sup
ported at its ends. Let I be the distance
Take the
metry,
is
between the supports, and w the load per footrun. of coordinates at the lowest point of the beam, which, by sym origin at its middle point. Take a plane section C (fig. 44) at a distance x from O and consider the forces at the
right of C. These are the load on CB and the reaction of the support at B.
x
FIG. 44
the center of gravity of
The load on
CB
is
w I l__ x
2
x\
,
acting at
CB, which
is
at the distance of
2

from
C.
Hence
the
acts
moment
of the load
is

downward. The support B
taken negative, since the load wl The moment supports half the load equal to

\

which
is

of this reaction about
C
is
therefore
I
/
]
Hence we have
2 \2
EI Vy_wl(l rl
5?
TVS
/
/Z_ x SVS I
is
Y_^_ ~iU
The general
solution of this equation
But
in the case of the
beam,
since,
when x
=
0,
both y and
are 0,
we have
=
0, c 2
=
0.
is
Hence the required equation
2
y
dx2
fix ^
\
dx/
is
The essential thing here
but does not contain
if
?/
that the equation contains
dy
^
and
^
d 2y
&gt;
,
Hence except implicitly in these derivatives.
have
the
we
place
= p, we
which
^f
first
=
j
&gt;
and the equation becomes
^
=/(, p),
is of
order in
p and
76.
a?.
If
we can
find
p from
this equation,
we can then
find
2,
y from ^
= p.
method has been exemplified in Ex.
CERTAIN EQUATIONS OF THE SECOND OKDER 151
The
essential thing here is that the equation contains
x.
but does not contain
&lt;ry
As
before,
we
d
place
)
~ and dx
dp
dx2
^,
= dpdpdy =
is
dp
u
_*
^=p but now
write
which
of the first order in
p and
y.
If
we can
find
p
from this
equation,
Ex.
2.
we can
find
~ y from
= p.
to the length of the portion of the
Find the curve for which the radius of curvature at any point is equal normal between the point and the axis of x
The length of the radius of curvature equation of the normal is (I, 101)
is

^
X
X/ J
(I,
y
192).
The
This intersects
therefore
OX at the point
of the
,_
(x
+
y
,
oV
The length
of the
normal
y\ 1 + ( V. \ \dx/
problem are /JAO , 3
satisfied
The conditions
equations
by
either of the differential
/
/rfi,\ 2
(1)
or
(2)
Placing
dx
=p
and
dx 2
=p
dy
in (1),
we have
whence
dy
The solution
of the last equation
is
y
=
Cl
whence
p=
^
c ~^
"
152
SIMPLE DIFFERENTIAL EQUATIONS
,
Replacing p by
This
is
we have
Cl
y
c
2
dx
dx
Vy2~
most neatly solved by the use
Cl
of hyperbolic functions.
We
have
(
15)
coshiy
y
= xc 2
=
Cl
,
or
cosh
X
~
2
i
+
e
ei
)
.
This
If
is
the equation of a catenary with
its
vertex at the point
(c 2 ,
we place ^
cZz
= p,
and ^= p ^ dx 2 dy
1
in (2),
we have
,
dp + p2 _ _ py . dy y
whence
= ^pdp
The
solution of this equation
is
y
whence
Replacing
p
1
2/
p by
,
we have
dx
Vc
.
2

dx.
7/2
Integrating,
we have
.
V&lt;
or
(X
c2 ) 2
+
2/2
_
C
2
8
This
is
the equation of a circle with
its
center on
OX.
&gt;$
&lt;*
If
we multiply both
sides of this equation
t by 2 dx dx, we have /
Integrating,
we have
V/
(
)
=
/
&lt;
J
M
whence, by separa separating the variables, we have
J
f_ V2
dy O
i
CERTAIN EQUATIONS OF THE SECOND ORDER 153
Ex.
(fig.
3.
45) of
Let the angle AGP and let AP s. By
direction
Consider the motion of a simple pendulum consisting of a particle mass m suspended from a point C by a weightless string of length
P
I.
=
6,
I,
the vertical, 108, the force acting in the
is
where
AC
AP
is
equal to
m
;
but the only force
The weight
Hence the
acting in this direction is the component of gravity. of the pendulum being mg, its compo
nent in the direction
AP
=
is
equal to
mgs m6.
is
differential equation of the
motion
m
We shall
sis
mg sin 6.
on the hypothe
treat this equation first
that the angle through which the
pendulum swings
is
so small that
we may
place sin 6
=
6,
without sensible error.
Then
since 6
=
,
the equation becomes
ds
Multiplying by 2
at
dt,
and
integrating,
we have
where a2
is
a
new
arbitrary constant.
Separating the variables,
we have
whence
where
t
sini 
=
J.(tg
t
),
is
an arbitrary constant.
s
From
this, finally,
=
8in*/
(**,).
The physical meaning
maximum
value of s
t
;
it is
of the arbitrary constants can be given. therefore the amplitude of the swing.
For a
is
the
t
,
When t =
s=0;
hence
is
the time at
which the pendulum passes through the
vertical.
We
will next integrate the equation
m (ft =
mg sin 6
is
without assuming that the arc of swing
small.
Placing
s
=
10,
multiplying
by
2
dt,
and integrating, we have
H
=
2 g cos 6
+ GI.
\dt)
154
If,
SIMPLE DIFFEKENTIAL EQUATIONS
now, the pendulum does not make a complete revolution around the point of
support C, port
=
dt
for some v value of 6
which we will
call a.
Hence C
=
2 g cos a,
and our equation becomes
whence
/
2 (cos 6
t
cos a)
where
to is
the value of
for
which
,
=
cos
0.
To bring
1
iar form, place cos 6
=
e
1
2 sin 2 2
a=

2 sin 2
have, then,
^
r
Jo
22
 and
,
the integral into a famil
let
k
= sin
.
We
T
=
r^
,
2
/A:
sin 2 2
Place, now,
sin 
k sin 0.
There
results
^0
^]
If
V1
k2 sin 2
\Q = = V7 n
* \
*"
*&lt;&gt;)
(*
we measure
time from the instant in which
=
0,
we have
t
=
The integral can be evaluated by expanding (1 A; 2 sin 2 0)~^by the binomial 2 2 1, and a table for the value theorem, the expansion being valid since & sin of this series is found in B. O. Peirce s Tables, p. 118.
0&lt;
The pendulum makes one fourth
a, and
of a complete swing
when
varies
from
to
from
to
If 4
T is
the time of a complete swing, then
as
may
be verified by the use of Ex.
2,
27.
PROBLEMS
Solve the following equations
:
x dx
"
y dy
l
_
0.

4. sec 2 x
5. 6.
7.
dx
+
tan x tan y dy
0.
1
+y
+ x~
dx
8.
x*
(x
2
2.
2xsinydx + cosydy =
3.
x^l + y
2
+ yVl + x*dy = Q.
(
+ y)dx  xdy = 0. 2 dx  x 2 dy ) (y + Vz +
2/
0.
Vx2 
y2

y sini  dx
x/
)
+
\
x sini  dy x
=
0.
9. sin
x sin y dx
cos y cos xdy
=
0,
PROBLEMS
10.
(
155
\
x sin X
Q
y cos  ) dx
X/
x)
+
x cos  dy X
0.
11. \Q x
_j_
(x
dy
y dx) x2
+
y
\e x
2
e
2/ )
x) x
dx
0.
= 0.
12. (2 y
13.
(
2
2
3 xy) dx
?/
X
+

3)
dx
+ (3 + (2 x y

xy
+ dy  1) dy = =
0.
=
0.
14.
15. (x 16.
35. _^
_
2
?/
= x2
e&lt;
.
+
\
2/)
dx
+
(x
+
+
1)
dy
dx
36.
dy dx
y
1)
=
dy dx
sin x.
^  ay =
dx

17. (x
+

2y
=
37.
=
sin ax.
18.
dx 2
19.
x (xycos^i
x
ycos)dx + +
y
xcoscfy x
=
39.^ =
40
y ^ dx 2 dx
dy 20. x dx
21. (y
22.
=
xex
.
+ (x xy 2 dy = 0.  ydx + Vx 2 + y 2 dx = 0. xdy +
x2 y) dx
x
1
)
41
x
d *V
23.
dy
d.x
+
l
x2
y
=
^
_
^_
dx
2
x.
24.
1
+
x2
l
+
x2
25.
dx
26.
27. _fL
+
y tan x
=
sin 2 x.
_
dx
28.
x2

f
1
dy_tf
dx
S
dx 2
dx
29.
dx
30.
dx
31. 32. (x 2 \/z 3 4 y 3
3
7/
)dx
f x?/
2
dy = 0.
49
(
x !l^__Li)^
dx 2
dx/dx
33.
dy
dx
.
ex
e
+
1
dx 2
34. 3
dx
dx2
156
52. Solve
SIMPLE DIFFERENTIAL EQUATIONS

u&

=8
2
7/
,
under the hypothesis that when x
=
1,
y
=
1
and
dx
=
y
3
.
53. Solve
=
 sin 2
y,
under the hypothesis that when x
=
1,
=
and
^dx
1.
54. Solve
= sin 2
dx 2
?/,
under the hypothesis that when y
=
0,
dx
=
0.
55. Solve 2
dx 2
=
y
2

,
under the hypothesis that
=
dx
when y =
&lt;x&gt;.
56. Solve
dx 2
= y+ 1, under the hypothesis that when x = 2, y =
and
dx
is
1.
5 7. Find the curve in which the slope of the tangent at any point the slope of the straight line joining the point to the origin.
that the load on the chain
n times
58. Find the curve in which the chain of a suspension bridge hangs, assuming is proportional to its projection on a horizontal line.
59. Find the curve in which the angle between the radius vector and the is n times the vectorial angle.
tangent
of
60. Find the curve such that the area included between the curve, the axis a fixed ordinate, and a variable ordinate is proportional to the variable cc,
ordinate.
61.
Show
is
that, if the
circle.
normal
to a curve
always passes through a fixed point,
the curve
a
the curve and the axis of x
62. Find the curve in which the length of the portion of the normal between is proportional to the square of the ordinate.
tangent
63. Find the curve in which the perpendicular from the origin is equal to the abscissa of the point of contact.
upon the
64. Find the curve in
which the perpendicular upon the tangent from the
is
foot of the ordinate of the point of contact
constant.
65. Find the curve in which the length of the arc from a fixed point to any is proportional to the square root of the abscissa of P. point
P
66. Find the curve in which the area bounded
a fixed ordinate, and a variable ordinate arc which is part of the boundary.
67. Find the deflection of a
other.
is
by the curve, the axis of x, proportional to the length of the
at one end
beam loaded
and weighted
at the
68. Find the deflection of a 69. Find the deflection of a
ends.
70. Find the curve 71.
beam
fixed at one
end and uniformly loaded.
beam loaded
at its center
and supported at
its
whose radius of curvature
in
is
constant.
Find the curve
the cube of the length of
which the radius of curvature at any point varies as the normal between that point and the axis of x.
PROBLEMS
A
157
72. particle moves in a straight line under the influence of an attracting force directed toward a fixed point on that line and varying as the distance from
the.
point.
Determine the motion.
particle moves in a straight line under the influence of an attracting force directed toward a point on the line and varying inversely as the square of the distance from that point. Determine the motion.
73.
A
74.
friction,
Find the velocity acquired by a body sliding down a curve without under the influence of gravity.
75. Assuming that gravity varies inversely as the square of the distance from the center of the earth, find the velocity acquired by a body falling from infinity to the surface of the earth.
CHAPTEE IX
FUNCTIONS OF SEVERAL VARIABLES
more than one variable. A quantity z is said a function of two variables, x and y, if the values of z are determined when the values of x and y are given. This relation is
83. Functions of
to be
= F(x, y), etc. expressed by the symbols z =f(x, y), z u is a function of three variables, x, y, and Similarly,
values of
given.
z, if
the
u
are determined
is
when
the values of
x, y,
and
z are
y, z),
This relation
y, z), etc.
u
= F(x,
Ex.
1.
expressed by the symbols u =f(x,
If r is the radius of the
base of a circular cone, h
its altitude,
and
u
its
2 volume, v = ^ Trr h, and v is a function of the two variables, r and h. Ex. 2. If /denotes the centrifugal force of a mass m revolving with a velocity
v in a circle of radius r,
u,
/=
,
and
/ is
a function of the three variables, m,
gas,
t
and r. Ex. 3. Let
its
v denote a
volume of a perfect
k
,
its
absolute temperature,
and p
pressure.
Then
where k
:
is
a constant.
t
This equation
1
t
pt&gt;,
may
be
written in the three equivalent forms
of the quantities,
_p, v,
p = k,
v
=
t
k,
by which each
and
,
is
explicitly expressed as a function of the other two.
tion
function of a single variable is denned explicitly by the equa y =f(x), and implicitly by the equation F(x, y) = 0. In either case the relation between x and y is represented graphically by
Similarly, a function of
t
A
a plane curve.
denned explicitly
t
= 0. In either case the graphical representa equation F(x y, z) tion of the function of two variables is the same, and may be
made by introducing the conception
84.
of space coordinates. coordinates in space. To locate a point in Rectangular space of three dimensions, we may assume three number scales, ZZ mutually perpendicular, and having their
r
two variables may be by the equation z=f(x y), or implicitly by the
XX YY
,
(fig. 46), zero points coincident at 0.
,
They
158
will determine three planes,
GRAPHICAL REPRESENTATION
XOY, YOZ, ZOX,
two.
lines,
159
each of which
is
perpendicular to the other
The planes
XX YY
,
,
are called the coordinate planes, and the three and ZZ , are called the axes of x, y, and z
respectively, or the coordinate axes, origin of coordinates.
and the point
is
called the
and through P pass planes perpen and ZZ intersecting them at the points L, M, and respectively. Then if we place x OL, y = OM, and z = ON, as in I, 16, it is evident that to any there corresponds one, and only one, point and that to set of values of x, y, and z any set of values of x, y, and z there cor responds one, and only one, point. These values of x, y, and z are called the coordi nates of the point, which is expressed as
Let be any point in space, dicular respectively to
,
P
XX YY
,
,
N
;
P(x,
y, z).
From
and
the definition of x
it
follows that x
is
equal, in
magnitude
coordinate
It follows
direction, to the distance of the point
from the
z.
plane YOZ.
Similar meanings are evident for y and
that a point may be plotted in several different ways by construct ing in succession any three nonparallel edges of the parallelepiped (fig. 46) beginning at the origin and ending at the point.
In case the axes are not mutually perpendicular, we have a system of oblique In this case the planes are passed through the point parallel to the coordinate planes. Then x gives the distance and the direction from the plane
coordinates.
YOZ
to
to the point,
z.
measured parallel
to
OX, and
y and
It follows that rectangular coordinates are
similar meanings are assigned a special case of oblique
coordinates.
85.
Graphical representation of a function of two variables. y) be any function of two variables, and place
,
Then the locus of all points the coordinates of which satisfy (1) is the graphical representation of the function f(x, y}. To construct this locus we may assign values to x and y, as x x l and y y lt
=
and compute from
(1)
the corresponding values of
z.
There will
160
FUNCTIONS OF SEVERAL VARIABLES
z,
be, in general, distinct values of
and
if
(1) defines
an algebraic
points all
at the point
function, their
number
will be finite.
The corresponding
an d these points alone of this line are points of the locus, and the portions of the line between them do not belong to the locus. As different values are assigned to x and y new lines parallel to OZ are drawn on which there are, in general, isolated
i&gt;
on a pi( x
lie
line parallel to
OZ
and intersecting
XOY
2/i)&gt;
y
It follows that the locus has extension in only two dimensions, i.e. has no thickness, and is, accordingly, a surface. Therefore the graphical representation of a function of two variables is a surface.* If
points of the locus.
f( x y)
&gt;
is
corresponding line parallel to
indeterminate for particular values of x and OZ lies entirely on the locus.
y,
the
and
Since the equations z =f(x, y} and F(x, y,z)=Q are equivalent, their graphical representations are the same, it follows that
the locus
of any single equation in
x, y,
and
z is
if
a surface. we demand
that the
There are apparent exceptions to the above theorem, surface shall have real existence. Thus, for example,
x2
is satisfied
+
y
2
+
z2
=
1
by no
it
real values of the coordinates.
"
It is
convenient in such cases,
however, to speak of
imaginary
lie
surfaces."
Moreover,
tion
may happen
For
give points which example, the equation
may
that the real coordinates which satisfy the equa upon a certain line, or are even isolated points.
x2
+
_i_
v2
(0, 0, z)
is satisfied
in real coordinates only
by the points
2
which
lie
upon the
axis of z
;
while the equation
x2
,
zz
_
n
is satisfied, as far as real points go, only by (0, 0, 0). In such cases it is still con venient to speak of a surface as represented by the equation, and to consider the part which may be actually constructed as the real part of that surface. The
imaginary part is considered as made up of the points corresponding to sets of complex values of x, y, and z which satisfy the equation.
86.
its
The appearance of any surface can be thus determined from equation by assigning values to any two of the coordinates and
computing the corresponding values of the third. This method, however, has practical difficulties. In place of it we may study
* It is to be noted that this method of graphically representing a function cannot be extended to functions of more than two variables, since we have but three dimen
sions in space.
GRAPHICAL REPRESENTATION
any surface by means
of the sections of the surface
If, for example, we place z parallel to the coordinate planes. the equation of any surface, the resulting equation in x and
161
made by planes
=
in
y is the equation of the plane curve cut from the surface by evidently the plane XOY. Again, if we place z = z l9 where z l is some fixed finite value, the resulting equation in x and y is the equation of the plane curve cut from the surface by a plane parallel to the new plane XOY and z^ units distant from it, and referred to
axes
and O Y which are the intersections of the plane z = z with the planes XOZ and YOZ respectively for by placing z = z instead of z = 0, we have virtually transferred the plane XOY,
X
f
,
l
;
1
parallel to itself,
through the distance z r In applying this method it is advisable to find = plane sections made by the coordinate planes x
of the surface,
first 0,
the three
0, z
y=
=
0.
These alone will sometimes give a general idea of the appearance but it is usually desirable to study other plane sec tions on account of the additional information that may be derived.
The following
it is
surfaces have been chosen for illustration because
important that the student should be familiar with them.
1.
Ex.
Ax + By +
Cz
+D=
(fig.
0.
Placing z
=
0,
we have
47)
0.
Ax + By + D =
(1)
Hence the plane XOY Placing and then x = 0, we find the sections of this surface made by the planes and YOZ to be respectively the straight lines
cuts this surface in a straight line.
y=
ZOX
and
Placing z
Ax + By + =
Zi,
Cz Cz
+D= +D= +D=
0,
0.
(2)
(3)
we have
Czi
0,
(4)
Ax + By +
which
is
the equation of a straight line in the plane z = z\. The line (4) is parallel to the
line (1), since
they
make
the angle tan 1
/
j
and lie in parallel planes. To find the point where (4) intersects the plane XOZ, we place shows that this point is a point of the y = 0, and the result Ax + Cz\ + D = linn (2). This result is true for all values of z\. Hence this surface is the locus
with the parallel lines
&X
and
OX
162
FUNCTIONS OF SEVERAL VARIABLES
which moves along a fixed straight
;
of a straight line
line
always remaining
parallel to a given initial position
hence
it is
a plane.
Since the equation
three coordinates,
Ax + By+ Cz + D =
a
0, 6
we have proved
,
is the most general linear equation in that the locus of every linear equation is a plane.
Ex.
2.
z
Placing
= ax 2 + by z where z = 0, we have
&gt;
&gt;
0.
ax2
+
by
2
0,
(fig.
(1)
and hence the
XOY plane
cuts the surface in a point
48).
Placing y
=
0,
(2)
we have
which
rabola
is
z
=
ax ^
the equation of a pa with its vertex at
axis along OZ.
0,
and
its
Plac
ing x
=
is
we have
z
=
by
2
,
(3)
which
and
also the equation of a parabola with its vertex at
its
axis along OZ.
&gt;
Placing z = z\, where z\ 0, we may write the resulting equa
tion in the
form
which is the equation of an ellipse
i
with semiaxes
\
/
and
a
\
6
As
the plane recedes from the origin, i.e. as z\ increases, it is
in magnitude.
FlG 48

evident that the ellipse increases
If
we
place z
is
Zi,
the result
may be written
in the
form
Zl
x2
\
y2 =
Zi
1,
and hence there
no part of
this surface
The surface
moving an
ends of
its
is
called an elliptic paraboloid,
on the negative side of the plane XOY. and evidently may be generated by
.
ellipse of variable magnitude always parallel to the plane XOY, the axes always lying respectively on the parabolas z = ax2 and z = by 2 While the appearance of the surface is now completely determined, we shall, nevertheless, find it of interest to make two more sections. In the first place, we
note that both the coordinate planes through OZ cut the surface in parabolas and their axes along OZ, and hence arises the question, with their vertices at
Do
all
planes through
OZ
To answer
this question
cut the surface in parabolas ? we shall make a transformation of coordinates by
&lt;
through an angle revolving the plane , of transformation will be x = x cos _
XOZ
about
0&gt;
OZ as an axis. The
formulas
y, sin
y
=
x sin
&lt;
+
GRAPHICAL REPRESENTATION
for z will not be changed at all, and x and y will be changed in the ner as in I, 115. The transformed equation is
z
163
same man
=
a (x cos
&lt;f&gt;
y sin 0) 2
Placing y
=
0,
we have
z
=
(a
cos 2
+ b (x sin + + b sin 2 0) z 2
,
y cos 0) 2
.
(5)
which
a parabola with its vertex at and its axis along OZ. But by chang we can make the plane OZ any plane through OZ, and hence every ing and its plane through OZ cuts this surface in a parabola with its vertex at axis along OZ.
is
X
y
we will place and write the result ing equation in the form
Finally,
y\
=.
which
allel
is
parabola with
to
the equation of a its axis par
its
OZ and
vertex
at a distance by*
from the
plane
XOY.
a2
Ex .S.
Placing z
+
_
&2
0, c2
=
=
we have
(1)
FIG. 49
a2
which
y
is
the equation of an ellipse with semiaxes a and 6
2
(fig.
49).
Placing
=
0,
we have
22
which
its
is the equation of an hyperbola with its transverse axis along conjugate axis along OZ. Placing x = 0, we have
OX
and
62
(3)
which
its
is the equation of an hyperbola with its transverse axis along conjugate axis along OZ. If we place z = z\ and write the resulting equation in the form
,
OY and
.1
OM 1 +
we
see that the section
is
(4)
=.
an
ellipse
with semiaxes a
X
c2
&gt;
c2
which accordingly increases in magnitude as the cutting plane recedes from the origin, and that the surface is symmetrical with respect to the plane JTOF,
the result being independent of the sign of
z\.
164
of one sheet,
FUNCTIONS OF SEVERAL. VARIABLES
Accordingly this surface, called the unparted hyperboloid or the hyperboloid may be generated by an ellipse of variable magnitude moving and with the ends of its axes always lying always parallel to the plane
XOY
and
on the hyperbolas
If
=
a2
c2
1
=
62
c2
1.
now we
revolve the plane
XOZ
sin 0,
about y
OZ
as
an axis by making the trans
x
=
x cos
is
/
y
=
+
x sin
+
y cos 0,
the transformed equation
/
(x
cos0
y
a2
sin0)
2
/
(x
sin0
y cos0) 2
/
z2
"c
_
"
62
2
Placing ,
= 0, we
have
+
x
,_
=
1,
(6)
and have thus proved that the section made by any plane through OZ
hyperbola with
Finally,
its
is
an
we
conjugate axis along OZ. place y = y, and write the resulting equation in the form
a
a2
?_ 2ll_^l ~
c2
52
whence we
see that the surface
is
To
discuss the equation further the value of y im
If yi
&lt;
we
symmetrical with respect to the plane XOZ. shall have to make three cases according to
6,
we put
the equation in the form
=
*!
(I) (59
which
2 a
is
the equation of an hyperbola with
its
transverse axis equal to
^
1
 and
parallel to
OX. Hence
the vertices will approach coincidence
as the cutting plane recedes from the origin. If y\ = 6, the equation may be put in the form
which
is
If yi
&gt;
the equation of two straight lines which intersect on 6, we write the equation in the form
*"
OY.
=
1.
which
2 c
is
the
1
equation of an hyperbola with
to OZ. 0:
its
transverse axis equal to
Ca
Vl v
and parallel
Hence the
vertices separate as the cutting
plane recedes from the origin.
GRAPHICAL REPRESENTATION
.?*"*
This surface
cone, with
its
(fig.
165
50)
is
a
OZ
as
its
axis
and
vertex at 0.
This surface
ellipsoid.
(fig.
51)
is
the
Ex
.
2
62
(fig.
c2
This surface
52) is the
biparted hyperboloid or the hy
perbolaid of two sheets.
~Fio.
50
The discussions
of the last
three surfaces are very similar to that of the unparted hyperboloid, that reason they have been left to the student.
and for
Ex.
7.
z
=
ax 2

by
2
,
where a
&gt;
0, b
&gt;
0.
Placing z
= 0, we
obtain the equation
ax 2
i.e.
=
by
2
=
0,
(fig.
(1)
two straight
lines intersecting at the origin
53).
Placing y
=
0,
we have
(2)
z
ax 2
,
the equation of a parabola with
its
vertex at
and
its
axis along the positive direction of OZ.
Placing x
= 0, we
2
,
have
z
=
by
(8)
the equation of a pa rabola with its ver tex at
O and
its
axis
along the negative direction of OZ.
Placing x
=
by
2
,
Xi,
we have
z
=
ax?

FIG. 51
b
l
a parabola with
the plane
its
axis parallel to
OZ and
its
vertex at a distance ax* from
is
XOY.
It is evident,
moreover, that the surface
.
symmetrical with
respect to the plane FOZ, and that the vertices of these parabolas, as different values are assigned to xi, all lie on the parabola z = ax 2
166
Hence
always
FUNCTIONS OF SEVERAL VARIABLES
this surface
may
parallel to the plane
is
be generated by the parabola z=by2 moving 2 FOZ, its vertex lying on the parabola z = ax
.
The surface
called the
hyperbolic paraboloid.
Finally, placing z = Zi,
where
z\
&gt;
0,
2
we have
by 2
,
zi
=
ax
an
to
hyperbola
axis
with
its
transverse
OX and length as the cutting plane recedes from the
origin.
If z
parallel increasing in

zi,
we may
write the equation in the
form
JV
=
1,
(0)
an hyperbola with
its
as the cutting plane recedes
transverse axis parallel to from the origin.
OF and
increasing in length
/r\
GRAPHICAL REPRESENTATION
Ex.
8.
167
z
=
kxy.
As
XOY
(fig.
the algebraic sign of the constant k merely shows which side of the plane we take as the positive side, we will assume fc&gt;0, and discuss the surface
on that hypothesis. = 0, we have xy = 0, which is the equation of the axes OX and OF. Placing y = 0, or x = 0, we have z = 0, and gain no new information about
54)
Placing z
the surface.
Placing z
=
z\,
where
z\
&gt;
0,
we have
an hyperbola referred to we have the hyperbola
its
asymptotes as axes
(I,
117, Ex.).
Placing z
=
z,
Hence all sections made by planes parallel to are hyperbolas with OX and OF as their asymptotes, and lying in the first and the third quadrants or in the second and the fourth quadrants according as the cutting plane is on the
positive or the negative side of the
XOY
XOY plane.
z
Placing x
Xi,
where x\
&gt;
0,
we have
=
kx iy
,
(3)
which
the equation of a straight line intersecting plane FOZ. If x = Xi, we have the straight line
is
OX
and
parallel to the
z
=&gt;
kxiy.
(4)
Similarly, placing y
y\,
where y\
z
0,
we have
(6)
fc?yiz,
168
which
is
FUNCTIONS OF SEVERAL VARIABLES
If
ZOX.
the equation of a straight line intersecting we have the straight line y = ?/i,
OF and parallel to the plane
(6)
As
the cutting plane recedes from the origin
it is
evident that these straight
lines revolve
about OX and OY respectively as they move along them, but remain parallel to the planes YOZ and ZOX respectively. always
about
Finally, we will revolve the plane OZ as an axis, the transformed equa
XOZ
tion being
?/
cos0).
Placing y
0,
we have
z
kx
2
cos0
sin 0, (7)
the equation of a parabola with its vertex at and its
axis along OZ.
Hence
all
planes through OZ cut this surface in parabolas with
their axes along OZ.
The surface is a special case of the hyperbolic pa for if raboloid of Ex. 7
;
we keep OZ
swing OX
FIG. 55
in its original position
and
and
OY into
I,
new
the formulas of
117,
by and choose the
positions
assumes the form z
but
if
=
2 a.
4ab
a
6
=
+
xy.
b
as in the illustrative example of angle 2 that article, the equation z = ax2 by Here the coordinates are oblique unless b = a
;
a, the
coordinates are rectangular and
we
have the case just con
sidered,
where k
=
of this surface on which the coordinates are all positive shows the relations between the pressure, the volume, and the tempera graphically ture of a perfect gas ( 83, Ex. 3). This part of the surface is shown in fig. 55.
The portion
87. Surfaces of revolution.
If the sections of a surface
made
are circles with by planes parallel to one of the coordinate planes on the axis of coordinates which is perpendicular to the their centers
revolution (37) with that cutting planes, the surface is a surface of coordinate axis as the axis of revolution. This will always occur
= the equation of the surface is in the form F^x^+y*, z ) in the which means that the two coordinates x and y enter only
when
&gt;
combination
^x + y
2
2
;
for
if
we
place z
=z
l
in this equation to
CYLINDEES
169
find the corresponding section, and solve the resulting equation for 2 e have, as a result, the equation of one or more circles, aj f y*
t
W
the equation F(y, z) = 0, which is the equation of the generating curve in the plane YOZ. = 0, we have F(x z) = 0, which is the Similarly, if we place y XOZ. It should be equation of the generating curve in the plane noted that the coordinate which appears uniquely in the equation shows which axis of coordinates is the axis of revolution.
according to the
number
place
of roots of the equation in
0,
x2 + y 2
.
Again,
if
we
x=
we have
t
= Conversely, if we have any plane curve F(x, z) XOZ, the equation of the surface formed by revolving
an axis
2
it
in the plane about OZ as
is
F(^/x?+ y
,
z)
=
0,
which
by
is
2
formed by simply replacing
2
.
x in the equation
Ex.
,
.
of the curve
V# + y
a2
1.
Show
this
that the imparted hyperboloid
^L + 2
o
=
a2
1 is
a surface of
revolution.
Writing
equation in the form
a2
,
b*
we
see that
it is
a surface of revolution with
OF as the axis.
curve.
Placing z
=
0,
we have  
= 1,
its
an hyperbola, as the generating
The hyperbola was revolved about
Ex.
y
2
conjugate axis.
2.
Find the paraboloid of revolution generated by revolving the parabola
about
its axis.
2
= 4tpx
=
88.
Replacing y by ~Vy
+
z2
,
we have
as the equation of the required surface
4px.
If a
= 0, given equation is of the form F(x, y) involving only two of the coordinates, it might appear to represent a curve lying in the plane of those coordinates. But if we are
Cylinders.
dealing with space of three dimensions, such an interpretation would be incorrect, in that it amounts to restricting z to the value 2 0,
=
any simultane ous values of x and y satisfying the equation F(x, y} = may be anything whatever. Hence, corresponding to every point of the curve F(x, y} in the plane XO Y, there is an entire straight line,
whereas, in fact, the value of z corresponding to
parallel to
OZ, on the surface F(x y)
t
=
0.
Such a surface
is
a
170
FUNCTIONS OF SEVERAL VARIABLES
in the cylinder, its directrix being the plane curve F(x, y) 2 0, and its elements being parallel to OZ, the axis of the plane
=
=
coordinate not present.
= a* is the equation of a circular cylinder, elements being parallel to OZ, and its directrix being the circle its = a 2 in the plane XOY. x*+ 2 In like manner, z = ky is a parabolic cylinder with its elements
2 For example, x f
2
2/
f
parallel to
If only
OX.
one coordinate
is
present in the equation, the locus
is
a
2 =Q (a + l}x + ab planes. For example, the equation x = 0, which represents be written in the form (x I) a) (x may a = I = 0. Similarly, any equation and x the two planes x
number of
involving only one coordinate determines values of that coordinate only and the locus is a number of planes.
Regarding a plane as a cylinder
straight line,
of
which the
directrix is a
we may
say that any equation not containing all the
coordinates represents a cylinder. If the axes are oblique, the elements of the cylinders are not perpendicular to the plane of the directrix.
89. Space curves. The two surfaces represented respectively by the equations /x (x, y,z)=Q and /2 (x, y, z)=Q intersect, in general, in a curve, the coordinates of every point of which satisfy each of
the equations. Conversely, any point the coordinates of which satisfy these equations simultaneously is in their curve of inter
section.
in
x, y,
and
Hence, in general, z is a curve.
the locus
the locus
of two simultaneous equations
.
In particular,
of
the two
simultaneous linear equations,
is
a straight
line,
since
it
is
the line of intersection of the two
planes respectively represented by the two equations. = 0, fz (x, y, z) = 0, If, in the equations f^(x, y, z}
we
assign a
value to one of the coordinates, as x for example, there are two equations from which to determine the corresponding values of y and z, in general a determinate problem. But if values are
assigned to two of the coordinates, as x and y, there are two
SPACE CURVES
171
a single unknown z, a prob equations from which to determine Hence there is only one independent lem generally impossible.
variable in the equations of a curve.
the independent variable and place the equations of the curve in the form y = &lt;M#) z = $z( x by curve for y and z in terms solving the original equations of the
In general, we
may make x
}&gt;
of x.
The new
(
surfaces, y
are cylinders
tively.
= fa (x), z = 2 (x), determining the curve, with elements parallel to OZ and OY respec 88),
&lt;
The equation y
= ^(z)
=
&lt;
interpreted in the plane
XOY is
the equation of the projection ( 92) of the curve on that plane. in the plane ZOX, is 2 (x), interpreted Similarly, the equation z
the equation of the projection of the curve on that plane. In particular, if we solve the equations
for
y and
z in
terms of
x,
we have two
z
equations of the form
s,
y
= px + q,
same
= rx +
as the equations of the
straight line that
was represented by
the original equations.
any variable parameter, and we make x a function of t, as x=fl (t), and substitute this value of x in the equations y ==&(*), * = &(*) (89), we have
90. If
t
is
*=A(0&gt;
y =/,(
),
*=/.(
)&gt;
as the parametric equations of the curve.
More
generally, the three equations
in represent a curve, the equations of which may be generally put the form y = $ 1 (x),z eliminating t from the first and 2 (#), by
=
&lt;
the second equations, and from the
first
and the third equations.
Ex. The space curve, called the helix, is the path of a point which moves around the surface of a right circular cylinder with a constant angular velocity and at the same time moves parallel to the axis of the cylinder with a constant
linear velocity.
172
FUNCTIONS OF SEVEEAL VARIABLES
o&gt;
Let the radius of the cylinder (fig. 66) be a, and let its axis coincide with and the constant linear velocity OZ. Let the constant angular velocity be be v. Then if 6 denotes the angle through which the plane Z OP has swung from its
initial position
ZOX,
the coordinates of
any point P (x, y, by the equations
x
z) of the helix are
given
=
6
a cos
0,
y
=
a sin
0,
z
=
vt.
But
x
t
=
wt,
and accordingly we may have
y
r\
as the parametric equations of the helix
=
a cos u,
=
a
sin at,
z
=
vt,
being the variable parameter.
Or, since
t
,
w
we may
regard 6 as
the variable parameter, tions are
and the equa
z
x
=
a cos 0,
is
y
=
a sin
0,
=
kd,
where k
the constant 
91.
~
Ruled surfaces.
x
A
surface
generated by a mov ing straight line is called a ruled surface. The plane, the cone, and
\vhich
may be
FIG. 56
of ruled surfaces,
and
in
the cylinder are simple examples 8 6, Ex. 8, it was shown that the hyper
bolic paraboloid is a ruled surface.
Ex. 1. Prove that the ruled surface generated by a straight line which moves so as to intersect two fixed straight lines not in the same plane and at the same time remain parallel to a fixed plane is an hyperbolic paraboloid. Let the fixed straight lines have the equations y = 0, z = 0, and y = c, z = kx, and let YOZ be the fixed plane, the axes of coordinates being oblique. Then the = mz is a straight line which is parallel to the plane YOZ straight line x = a, y and intersects the line y = 0, z = 0, for all values of a and m. If this line
intersects the line
y
=
c,
z
a,
=
y
kx, evidently
m

Therefore the equations of
the required line are x
=
=
z.
ka k c equations satisfy their product xy = z, or z = xy. Hence the line always ^ k lies on the surface z = xy, an hyperbolic paraboloid ( 86, Ex. 8). z2 x2 y2 = lisa ruled sur Ex. 2. Prove that the unparted hyperboloid ~^+^ ^
face having two sets of rectilinear generators,
i.e.
Any
values of x, y, z which satisfy these
two straight
lines
may be drawn, each
of
which
shall lie entirely
that through every point of it on the surface.
RULED SURFACES
If
173
we
write the equation of the hyperboloid in the form
T2
72
? /2
&lt;
I? 1 *
it is
evident that (1)
is
the product of the two equations
any value of ki. But (2) are the equations of a straight line ( 89). More over, this straight line lies entirely on the surface, since the coordinates of every point of it satisfy (2) and hence (1). As different values are assigned to k\, we
for
obtain a series of straight lines lying entirely on the surface.
Conversely,
if
PI (xi,
yi z\)
t
is
any point
of
(1),
Therefore PI determines the same value of ki from both equations every point of (1) lies in one and only one line (2).
(2).
Hence
We may
also regard (1) as the product of the
two equations
(3)
whence
it is
evident that there
is
the surface, one the surface.
and only one
of
a second set of straight lines lying entirely on which may be drawn through any point of
Equations (2) and (3) are the equations of the rectilinear generators, and every point of the surface may be regarded as the point of intersection of one line from each set.
Ex. 3. The conoid (Ex. 52, p. 83) is evidently a ruled surface, since it is gen erated by a straight line moving always parallel to a given plane and at the same time intersecting a given curve and a given straight line.
Ex.
4.
Another example of a ruled surface
is
the cylindroid,
i.e.
the surface
generated by a straight line moving parallel to a given plane and touching two given curves.
Ex. 5. Show that the prismoidal formula ( 39) applies to the volume bounded by a ruled surface and two parallel planes. The equations x = pz + q, y = rz + s represent a straight line. If p, q, r, and
s are
functions of a single independent parameter
,
the line generates a ruled
surface.
174
If
FUNCTIONS OF SEVERAL VARIABLES
we
place z
=
g,
0,
we have x
rzi
q,
y
tion of the ruled surface
by the plane
= s, the parametric equations of XOY. Similarly, if we place z
the sec
Zi,
we
have x
the parametric equations of a section parallel to XOY. Suppose now when t varies trom t to ^, the straightline generator of the surface traverses the perimeter of the section z = z\. Then the area of this
= pzi +
y
=
+
s,
section
is
dp
t
dt
Jto \
dq ^+ ^
r
dt
r tl da
Jt
dt
dt
is a quadratic polynomial in z\. Hence the prismoidal formula holds for a portion of a ruled surface bounded by parallel planes. The prismoid itself is a special case of such a surface.
This
PROBLEMS
1.
Show
is
T/,
that the surface lfi(x,
?/,
z)
+ fc/2 (x,
?/,
z)
=
0,
where
I
constants, faces /i (cc,
a surface passing through
z)
all
the points
common
to the
and k are two sur
=
and /2
2. Discuss the surface 3.
its
Show
that the
= 0, and meeting them at no other points. = a 3 by means of plane sections. xyz surface z = a Vx 2 + y* is a cone of revolution, and find
(x, y, z)
vertex and axis.
4.
Prove that the surface ax
+
f
by
= =
2
cz 2 is a cylindroid,
and discuss
its
plane sections.
5.
6.
Discuss the surface x
+ y*
z*
a?
by means
cz
is
of plane sections.
its
Prove that the surface (ax
Discuss the surface x
+ +
6y)
=
eft
a cylindroid, and discuss
plane sections.
7.
f
y%
z%
=
by means
of plane sections.
is
Find the equation of a right circular cylinder which plane XOZ and has its axis in the plane YOZ.
8.
tangent to the
9.
planes
XOZ
Find the equation of a right circular cylinder which and XOY.
is
tangent to the
10. Describe the locus of the equation z 11. Describe the locus of the equation
+
(y
+
by)
2
I)
=
3
0.
2y 2
5y
=
0.
12. Describe the locus of the equation (ax 13.
+
2
c 2z 2
=
0.
an
ellipse revolving
Find the equation of the oblate spheroid, about its minor axis.
of
i.e.
the surface generated
by
14.
Find the equation
a biparted hyperboloid of revolution with
the surface generated
OY
by
as
its axis.
15.
revolving an ellipse about
Find the equation of the prolate spheroid, its major axis.
i.e.
PROBLEMS
16.
175
y
2
= 4px
17.
Find the equation of the surface generated by revolving the parabola about OF as an axis. Find the equation of the ring surface generated by revolving about + (y  b) 2 = a 2 where a 6.
,
&lt;
OX
the circle x2
18.
xy
=
c
Write the equation of the surface generated by revolving the hyperbola about either of its asymptotes as an axis.
Find the equation of the surface formed by revolving the fourcusped 8 an axis. hypocycloid x + y* =a* about OF as
19.
20. 21.
22.
What
surface
surface
is
is
represented by the equation x*
What What
Show
is
represented by the equation
represented
+ (y2 + z 2 )* = a* ? x 2 + z 2  y6 = ? +
z2
the
line
2 by the equations y
 6x = =
0,
z3=0?
23.
is 2 that the line of intersection of the surfaces x
+
y
2
=
a2 and y
z
an
ellipse.
(Rotate the axes about
OX through 45.)
24. Show that the projections of the skew cubic x = t, y = P, z = t* on the coordinate planes are a parabola, a semicubical parabola, and a cubical
parabola.
 a cos0, y = asinfl, z = kd 25. Prove that the projections of the helix x and YOZ are sine curves, the width of each arch of on the planes
XOZ
which
is
krr.
26. Prove that the projection of the curve x the plane
=
e{ ,
y=e~
f
,
z
= tV%,
on
XOY
is
is
an equilateral hyperbola.
f
27. Turn the plane XOZ about OZ as an axis through an angte of 45, and e~ z = t V2 on the new show that the projection of the curve x = e, y
,
XOZ plane
28.
a catenary.
that the curve x
2
,
Show
y
=
2
1,
z
=
t
is
a plane section of a para
bolic cylinder.
29. Prove that the skew quartic x = t, y = 3 z = t* is the intersection of an hyperbolic paraboloid and a cylinder the directrix of which is a cubical
,
parabola.
30.
its
The
vertical angle of a cone of revolution
is
90,
its
vertex
is
at 0,
and
starting from the vertex, moves in a spiral path along the surface of the cone so that the measure of the distance it has traveled parallel to the axis of the cone is equal to the circular measure of the
axis coincides with OZ.
A point,
angle through which
it
has revolved about the axis of the cone.
Prove that the
t
equations of its path, called the conical helix, are x
=
t
cos
,
y
=
sin
t,
z
=
t.
CHAPTEE X
PLANE AND STRAIGHT LINE
92. Projection. The projection of a point on a straight line denned as the point of intersection of the line and a plane through the point perpendicular to the line. Hence in fig. 46
is
are the projections of the point M, and and z respectively. The projection of one straight line of
L,
N
P on
the axes of
x, y,
second
straight
line
is
the
of the
part
of
finite length upon a the second line included
first line, its
between the projections
ends of the
direction
of the
being from the projection
initial
point of the first line to the
projection of the terminal point of the first line. In fig. 57, for ex
ample, the projections of
A
and
B
M
R
FIG. 57
_ N on
being A and B respec the projection of AB on tively, is A B and the projection of
MN
on
MN
BA
,
MN
is
A
.
If
MN and
AB
denote the positive directions
lows that
and
is
respectively of these lines, it fol is positive when it has the same direction as negative when it has the opposite direction to MN.
AB
MN
and
In particular, the projection on
OXoi the
z 2 ) is
I,
straight line
from P^x v y^ zj
OL 2 = x2
.
But
P (x L L =x
to
2 2
1
2&gt;
2
y2 x 1 by
,
L^L zt
13.
where
OL = x
1
2 P^P drawn
l
Hence the
projection
of
im
If
on
OX
is
x2
x1

and
and
OZ are
we
yl respectively y 2 define the angle between
parallel to
similarly, its projections and z2 z r
on
OY
any two
lines in space as the
angle between lines
them and drawn from
176
a
common
is
point, then the projection of one straight line on a second
the
PROJECTION
product of the length of
angle between
the first line
177
the cosine
lines.
and
between the positive directions of the two
of the angle Then if is the
&lt;/&gt;
AB
and
MN
(fig.
57),
To prove
by
I,
this proposition,
ing the plane
14,
ST
at C.
draw A C parallel Then A C = AB, and
f f
to
A B and meet A B = A C cos
&lt;/&gt;,
whence the truth
of the proposition is evident.
Defining the projection of a broken line upon a straight line as the sum of the projections
of
segments, we may prove, as in that the projections on any straight 15, line of a broken line and tht
its
I,
straight line joining its ends are the same.
We
will
now
*
^fe^o
show that
projection
the
of any
FIG. 58
plane area upon another plane
the cosine
Let
&lt;
X
is the product of that area and of the angle between the planes. OY (fig. 58) be any plane through OY making an angle
with the plane
XOY.
Let
AB
1
be any area in
intersects its
X OY such
boundary
that
any more than two
straight line parallel to
points,
and
OX let AB
f
in not
be
r
its
projection on
XOY.
(1)
Then
(
35)
area
AB =
C(x2
 x[) dy
t
the limits of integration being taken so as to include the whole area.
In like manner,
area
AB =
I
(x2
x^
dy,
(2)
the limits of integration being taken so as to cover the whole area.
178
PLANE AND STRAIGHT LINE
of y are the same in both planes, since they measured parallel to the line of intersection of the two planes and hence the limits in (1) and (2) are the same. Since the x
But the values
are
;
coordinate
is
&lt;,
x2 = x2 cos
measured perpendicular to the x 1 = x[ cos and (2) becomes
&lt;,
line of intersection,
area
AB =
I
(x2
I
x[) cos
&lt;
dy dy
).
cos
&lt;p
r
(x2
x^)
(/&gt;)
= (cos
93. Distance
(area
AB
yz
,
z 2 ) (fig.
between two points. Let P^x^ y v zj and P^(x2 be any two points. Pass planes through P^ and P 59) 2
,
parallel to the coordinate planes, thereby
z
forming a rectangular parallelepiped having the. edges parallel
to
a diagonal.
OX, OY, and OZ, and having Pfc as Then the edges are equal
x v yz
respectively to x2
(
y v and
z2
zl
92), being equal to the projections of
on OX, OY,
Hence
59 If the
2/ 2
PI
%=
(*2
,)
+ (y
as, for
two points have two coordinates the same, z2 = z v the formula reduces to yv
1.
example,
Ex. Let
Find a point Vl4 units distant from each of the three points
(1, 0, 3), (2,

1, 1),
(3, 1, 2).
P (x,
y, z)
be the required point.
(x (X (X
Then
 1)2 +  2)2 +  3)2 +
(y
(y
 0)2 + + 1)2 +  1)2 + (y
( (
Z Z
(
Z
_ 
3)2
1)2
=
14,
= 14j 2)2 = 14.
(0, 2, 0)
Solving these three equations,
we determine
the
two points
and
its
(4,
2, 4).
Ex.
If
2.
Find the equation of a sphere
y, 2) is
of
radius r with
center at
Pi(*i, y\, zi).
P (x,
any point
(x
of the sphere,

xi)2
+
(y

yi )2
+
(
Z
_
Zl )2
=
r2.
(i)
P
Conversely, if P(x, ?/, 2) is any point the coordinates of which satisfy (1), is at the distance r from P Therefore 1? and hence is a point of the sphere. (1) is the required equation of the sphere.
LENGTH OF A SPACE CURVE
94.
179
Length of a space curve in rectangular coordinates. The method of finding the length of a space curve is similar to that of 40 may finding the length of a plane curve, so that the proof of
be repeated. the given curve, we assume n l points between A and B, and connect each pair of , JJ_j JJ, JJ, JJ consecutive points by a straight line. The length AB is then
If
(fig.
AB
60)
is
defined
as
the
limit
of
the
sum
Al[,
^
J
of the lengths of the
n chords
is
~^~*^5^
Pf^

,
I^_ 1 B, as
n
increased
without limit
each chord
limit
(I,
and the length of approaches zero as a
(#.,
i
i
FIG 60
104).
Let the coordinates of If be
(# .+ A#,
t
y it
l
zt)
and those
of 7? +1 be
2
(
y

t
+ Ay,
z.f Az).
Then
PP +
&gt;
= ^/~Kx*+ Ay V Az
t
2 ( 2
93).
Now
if
x, y, z are
functions of a variable parameter
&gt;
90),
2
and
have the derivatives ~
cw
/
infinitesimal which differs by an infinitesimal of higher
then vAa? + Ay + Az is an a a 2 2 2 from the infinitesimal ^/dx + dy + dz
order.
^
For we have
A,
rf^y dt)
Since
t
/^y
\dt)
/dz*
(dt
(
is
the independent variable,
A = dt
,
4)
;
also
_ ~
dt
etc
Hence
Therefore
T Lim
.
y
=
1.
(
3), in
finding the length of the curve,
we may
denotes
Aa/^Hreplace the length of AB,
V
A^+A/
by
V^
2
+
2
cZ?/
+ dz
2
.
Hence,
if s
we have
=/
J(A)
180
PLANE AND STRAIGHT LINE
of
where (A) and (B) denote the values for the points A and B respectively.
follows
(
From
the independent variable this formula for s it
9) that
Ex. Find the length of an arc of the helix x corresponding to an increase of 2 TT in e.
=
a cos0, y
=
a sin0, z
=
kd,
Here dx
=
a sin e
dd,
dy
=
a cos Odd, and dz
f&gt;01
= k dd.
+ 2TT
Therefore
95.
Direction of a straight line.
The
direction of
line in space is
determined by means
of the angles
any straight which it makes
with the positive directions of the coordinate axes OX, OY, OZ, We denote these angles by a, /3, and
7 respectively
cosines,
i.e.
(fig.
61).
Then
of
their
cos a,
cos
/3,
cos 7,
are
called
line.
the
direction
cosines
the
It
is
to
be noted that the same
straight line av 7 X or TT
makes the angles a^, j31} 7r/3 lt TT 7 X with the
according to the di
line is drawn.
coordinate
rection in
.axes,
which the
coso^,
Hence
coscTj,
its
direction
cosines
cos
are
either
cos^,
7r
A
straight line in
cos^, cos 7 X or which the direc
tion
is
The
fixed has only one set of direction cosines. three direction cosines are not independent, for they satisfy
Cos
2
the equation
that
is,
+ cOS /3 + cos
of
2
2
7
=
1;
the
sum of
the squares
the
direction cosines of
any
straight line is always equal to unity. To prove this theorem, let the line
line
make the OZ respectively. Then
and
let it
angles
a, j3,
Jf^ (fig. 59) be any straight and 7 with OX, OY, and
PP **
cos
7
DIRECTION OF A STRAIGHT LINE
Squaring and adding these equations,
cos
2
181
we have
a
f cos /3
2
+
cos
2
7
=
But
2 Therefore cos * 4 cos
2
/S 4
cos
2
7
=
1.
(fig.
96. If Pt(x v
y v z x ) is
any
fixed point
59), the coordinates of
z any second point P may be denoted by where Arc, Ay, and Az are arbitrary.
(x^\ Arc,
y x 4 Ay,
z x 4
Az)
?
If cos *, cos
y8,
cos
7 are the direction cosines
article,
of the straight
line P^P^
we
have,
by the work of the last
Arc
n
~"
cos
*
Ay
for
and
Hence,
rection
if Aa?,
+A
Ay, and Aa; are given, the direction of a straight but the particular straight line having that di
line is determined,
is not determined, for the values of A#, Ay, and Az in no determine x v y v and z v the coordinates of the initial point of way the line. Moreover, the ratios of Are, Ay, and Az are alone essen tial in determining the direction of the line. Accordingly we may
speak of the direction A# Ay Az, meaning thereby the direction of a straight line, the direction cosines of which are respectively
:
:
Arc
Ay
Az
Furthermore, if A, B, and C are any three given numbers, we may speak of the direction A\B\C. For we may place Arc = A, Ay = B, Az = C, and thus determine a direction, the direction
cosines of
which are
+ B* + C
A + 7/ + C
2
2
2
A + B + C*
2 2
182
PLANE AND STEAIGHT LINE
If
97. Direction of a space curve.
we
point P (x, y, z)
of a curve as
determined by
regard the position of a its distance s, along
as in
I,
the curve from some fixed point of the curve, y, and z are functions of s, i.e.
105, x,
Then
if
s
is
Q(x+ A#, y
+ Ay,
increased by an increment As, a second point z f Az) of the curve is located, and the direc
tion cosines of the chord
PQ
2
are
VA + Ay + Az
2
2
2
vA + Ay + Az
As
A
2
2
VA^+ Ay + A/
are the direction cosines of
2
The
limits of these ratios, as
=
0,
the tangent to the curve at P. \
/yi
,yt
Now
whence
\
o
for
lim
_ way with
=L
the other two
(I,
104)
Proceeding in the same
ratios,
we have
point,
dx &gt;
dy
&gt;
dz ds
.
,.
as the direction cosines ot the curve at
ds
ds
any
since the directions of the tangent
and the curve
at
any point are
of the
the same.
Instead of giving the direction cosines,
direction of the curve as the direction
we may speak
dz, since
dx:dy.
dz
2
.
ds
= \^dx + dy +
2 2
(
94)
Ex. Find the direction of the helix
x
at the point for
=
a cos 0,
y
=
a sin 0,
z
=
A:0,
which 0=0.
asin0d0, dy
Here dx
for which
=
=
a cos0d0, dz
=
kde.
:
0=0,
the direction ^ a
is
the direction
add
Therefore, at the point A:d0, and the direction
:
cosines are 0,
=
Va 2 +
A;^
ANGLE BETWEEN TWO STRAIGHT LINES
: :
183
98. Angle between two straight lines. Let the directions of any two straight lines be respectively A^ A^ AjZ and A 2 # A 2 ?/ A 22, where the subscripts are used merely to distinguish the two direc tions. If two straight lines having these directions are drawn from any point P(x, y, z) they
: :
will
pass
z x y,
through the two points P^(x
4
A^,
2
y +A
(fig.
+A
X
2)
and ^(a;
+A
2 ^,
y+A
z 2 y,
+ A z)
FlG 62

62) respectively.
if
Then
is
the angle between these two lines,
we
have, by
trigonometry,
x

But
 A,*)
whence, by
substitution in (1) and simplification,
2
,
(
93)
a v cos ^, cos y l are the direction cosines of any straight first direction, and cos a cos /32 cos y 2 are the direc 2 tion cosines of any straight line with the second direction, the above formula becomes
or
if
cos
line
with the
,
,
cos 6
If the
(2)
= cos
1
cos
a2
+ cos /3
l
cos
/32
+ cos 7
:
1
cos
72
.
(3)
:
two directions are given as
A
l
:J^ i
Cv and A Z :BZ C2
,
evidently becomes
If the lines are
perpendicular to each other,
A^A 2 + B^B2
since cos 9
;
+&=
0,
(5)
and
if
they are parallel to each other,
since
cos a^
= cos a
2
,
cos
:H=t =
/3 1
cos
2,
cos7
1
=cos7
2
.
184
If
PLANE AND STBAIGHT LINE
6 represents the angle between any two curves,
97,
we
have,
from
99. Direction of the
P (x + A#,
2
l
yv +
A?/, z l
+
to a plane. Let l (x v y v zj and be any two points of the plane Az)
normal
.
P
Substituting their coordinates in
(1),
we have
(2)
+
Subtracting (2) from
(3),
Z&gt;=0.
(3)
we have

A
whence, by
tion
:
A^c
(5),
:
98,
+ B &y + C A = 0, the direction A B C is
: :
(4)
normal
is
to the direc
Ax Ay
A2.
But the
to the
latter direction
the direction of
any straight
line of the plane.
Hence
plane
the direction
direction of the
normal
Ax + By +
Cz
A B C is + D = 0.
:
:
the
that the equation of any plane is a linear For let the given plane pass through equation of the form (1). a fixed point JfJ (x v y v 2^ and be perpendicular to a straight line C. having the direction A
We may
now show
:
B
:
Now
the equation
Ax + By +
Cz
+D =
A B
: :
to the direction represents a plane perpendicular plane will pass through P^ if
C.
This
whence
Therefore
D= (Ax^ + By
A(x
x^)
l
+
CzJ.
zl
+ B(y
y^)
+C(z
)=Q
(5)
and to the direction represents a plane perpendicular But only one plane passing through the fixed point : (x v y v zj. these conditions hence the given plane has the equa can
A:B:C
P
satisfy
;
tion just determined.
Any
plane
however, and hence every plane
equation.
may be determined in this way, may be represented by a linear
NORMAL EQUATION OF A PLANE
Ex. Find the equation of a plane passing through the point 1. normal to the straight line having the direction 2 3
: :
185
(1, 2,
1)
and
The equation
or
is
2(x
_
l)
+
o
(y
_
2)
3y
1
(
z z

1)
2x
+
1
= =
0,
0.
100. Normal equation of a plane. Let a plane be passed through the point J^(x v y v 2^) perpendicular to the straight line OP (fig. 63). l 96, the direction By z
of
OP^
is
the
direction
x^. y^. z v
and hence the
by
99,
equation of the required
plane
is,
(5),
+ *,(*
*,)=&lt;),
(1)
which may be put form
in the
~( x i+yi+Zi)=Q
()
If the direction cosines of
r/
FIG. 63
n
01^ are denoted respectively by I, m, and n, i.e. / = cos a, = cos 7, and the length of OP^ is denoted by p, then
m = cos
and
Accordingly,
if
2 p = V#
2
f ?/
+ zf.
V,/2
we
divide equation (2) by
Ix
f
y*
+2
2
,it
becomes
(3)
+ my 4
nz
p=
0,
which
is
known
(3)
as the
normal equation
of the plane.
Equation
Let P(x,
is
may
y, z) (fig.
z.
LM=y,MP =
m&gt;
OP and the projection of the broken V lOL + LM+ n MP 92). Hence

be derived geometrically as follows: 63) be any point of the plane, and let OL = x, Draw OP. Then the projection of OP on OP l
also
line
OLMP
on
OP
l
is
(
I
OL + m
.
LM+ n MP  OP^= 0,
186
PLANE AND STRAIGHT LINE
of the plane
The general equation
may
evidently be
made
2
+^ equation by the plane is the direction
2
V^
assume the normal form by dividing the +(7 since the direction of the normal to
to
2
,
The sign of the radical must in order that the constant term be taken opposite to the sign of may be the negative of the distance of the plane from the origin,
A\B\C.
D
as in (3).
Ex Find the direction cosines of the normal to the plane 2x3i/ + 6z+14=0,
.
also its distance
from the
origin.
Dividing by
 V2 2 +
32
we have
Hence the direction
the plane
is
+ 6 2 i.e.  7,  fx + %y  f z ,
2
=
0.
cosines of the normal to this plane are
f
,
^
,
f and
,
2 units distant from the origin.
101. Angle between
two
planes.
Let the two planes be
&gt;0,
(1)
0.
(2)
is the same as the angle between the directions of which are respectively their respective normals, C2 Hence if is the angle the directions A 1 B l Cl and A 2 2
The angle between these planes
:
:
:
B
:
.
between the two planes,
cos
=
.i
.
VJ + Bl + C?
A
b
(4)
9g)
T/
The conditions
are respectively
for perpendicularity
and parallelism
of the planes
^
^L
+BB
l
+(7(7=0
and
*S l
A
^B
=
V.
c
any
straight line.
102. Determination of the direction cosines of
Let the equations of the straight line be in the form
= 0, =
(1)
0,
(2)
LINE THROUGH
and
let its direction cosines
TWO POINTS
n.
187
lies in
be
/,
m, and
Since the line
both planes (1) and (2), Therefore (by (5), 98),
it is
perpendicular to the
normal to each.
also
2
/
+m
2
2 +7i =l.
(95)
of
/,
Here
are three equations
from which the values
m, and n
may
be found.
From
the
first
two equations we have
(I,
8)
.
l:m:n
= B #
l
C\ C,
C,
A
l
2
B.
line
Ex. Find the direction cosines of the straight + yz + 7r=0.
2xf3y +
z
4
=
0,
The
three equations for
Z,
m, and n are
2
I
+
Z
4
m +n= +mn=
3
0,
0,
the solutions of which are
l=
V38
or
5
38
V38
5
r=~
=
9.
V38
V38
V38
Since cos (180 cos0, it is evident that if the angles corresponding 0) to the first solution are a\, /Si, 71, the angles corresponding to the second solu tion are 180 nrj, 180 y lt Since these two directions are each the ft, 180



negative of the other,
it is
sufficient to take either solution
and ignore the
other.
103. Equations of a straight line determined by two points. Let P(x, y, z) be a point on the straight line determined by
"
*( x v Vv z 2 so situated that P^P &(%%). Then, 18, there are three cases to consider according to the If is between and v P^P and JJJJ position of the point P. JFJ
?(#!
y\&gt;
z i) a nd
p
=
)&gt;
as in
I,
P
P
have the same direction, and P^P P^ accordingly k is a posi tive number less than unity. If P is beyond P from JJ, Pf 2 and P^ still have the same direction, but P P&gt;P P there 2 1 1
&lt;
;
;
fore
k
is
a positive
number
greater than unity.
Finally,
if
P
188
is
PLANE AND STRAIGHT LINE
beyond
is
P
l
k
a negative number,
to co.
fig.
from P^ P^P and P^ have opposite directions, and its numerical value ranging all the way
first case,
from
In
7^,
64,
which represents the
2
pass planes through
P, and
P% perpendicular to
the points
Mv M, and M
= k(PP
1
OX; and
let
respectively.
them intersect Then OM=
1
OX
at
and since 7JP
2))
by geometry
M^M = k(M M
z ).
whence, by substitution,
(1)
By
and perpendicular
passing the planes through the points perpendicular to to OZ, it may be proved in the same way that
OF
y
= 2/i+fe/ z = z +k(z zj.
2
1
2/i)&gt;
(2) (3)
z
The construction and the proof for the other two cases
are
FIG. 64
the
same,
sults are the
cases.
and the re same in all the
By assigning different values to k we can make equations (1), on the straight line determined (2), and (3) represent any point
by the points
P^PV
7J
and
three equations, the point
Hence
line
(1), (2),
Conversely, if x, y, and z satisfy these must be a point of the straight line and (3) are the parametric equations of the
P^.
straight
determined
by two
points,
k being the variable
parameter.
104.
By
eliminating k from the three equations
(1), (2),
and
(3)
of the last article,
we have
(1)
result proves
two independent equations in x, y, and z. This 89, that two linear equations a straight line for we have any straight line always represent
Here
are but
the converse of
;
TANGENT TO A SPACE CUKVE
represented by two linear equations. x 1 :^/ 2 y l z 2 z l is the direction x2
:
189
The
(
direction of the line
96).
be noted that, if in the formation of these fractions any denominator is zero, the corresponding direction cosine is zero, and
It is to
the line
is
perpendicular to the corresponding axis.
line
(1, 5,
Ex. Find the equations of the straight  1) and (2,  3,  1).
determined by the points
x\ _
y

5
2_1~
_35"
z + 1 1 + 1
is
Hence the two equations of the line are z + 1 = 0, since the line the JTOFplane and passes through a point for which z = 1, and 8 x formed by equating the first two fractions.
+y
parallel to 13 0,
=
is is
any
105. Equations of a tangent line to a space curve. If (x v y v z^ of a space curve, and P^ (x 1 A#, y l Ay, z^ Az) any given point are second point of the curve, the equations of the secant
%
+
+
+
7^
xl
y
A#
or Aa?
Ay
A//
(
104)
A3
As
where As
tions
is
As
As
the length of the arc P^P V
at P^ as in
I,
Denning the tangent
59,
we have
as its equa
/dy\
ds /i (dx\
\ ds A
/dz"
\
dx
dy
dz
at the point
where dx, dy, dz are to be computed
Ex. Find the equations of the tangent
x
at the origin.
P r
to the helix
0,
=
a cos
0,
y
=
a sin
z
=
k
Since dx
=
a sin
0(Z0,
dy
2/
?
=
? a
= ?, or(104)z =
A;
0,
.
= a cos0d0, = ?.
k
dz
=
kd6, the required equations are
a
190
PLANE AND STRAIGHT LINE
line
106. Equations of a straight
in
terms of
its
direction
and a known point upon it. a known point of the line, and let
cosines
cosines.
Let
/,
P (x v
l
y v zj
(fig.
m, and n be
line.
its
65) be direction
as a
Let P(x, y, z) be any point of the diagonal construct a parallelepiped as in denote by r, we have
On P^P
Then
if
93.
we
Pf
= mr,
But
= nr.
z
whence
x
=x
l ~\
lr,
y
= y + mr,
l
= z + nr.
l
(1)
These are the parametric equations
of the line, the variable
param
eter being r, r being positive if the point is in one direction from P^ and negative if it is in the other direction
from
j^.
By
eliminating r
we have
n
/
m
FIG. 65
which are but two independent linear
equations.
Problems on the plane and the straight line. In this article we shall solve some problems illustrating the use of the equations of the plane and the straight line.
107.
Plane determined by three known points. Let the three given and ?J(#8 y 3 zs ), and let the points be J^(xv y v zj, I(x v y v ), of the plane determined by them be equation
1.
, ,
Ax + By +
Since
Cz
of
+D =
0.
(1)
P P v
2,
and
P
%
are points
the plane, their coordinates
satisfy (1).
Therefore
Ax2 +
%+
2
Cz 2 +
D = 0,
the ratios of the
(1),
(3)
We may now solve
(2), (3),
and
(4) for
unknown
eliminate
constants A, B, C, and D, and substitute iu
or
we may
PROBLEMS
A, B, C, and
either
191
D from
method the equation
the four equations (1), (2), (3), and (4). By of the required plane is found to be
y
2/1 2/2 2/3
Ex.
1.
(1, 1, 1),
Find the equation of the plane determined by the three points  3,  1). ( 1, 1, 2), and (2,
is
The required equation
1111 = 21 1
x y
1
z
1
0,
2311
is
no plane given points are on the same straight line, and the above equation reduces to = 0. determined,
If the three
2.
Distance of a point
2 X ),
from a
P^(x v y v
and
at first let
Let the given point be the equation of the plane be in the
plane.
(1)
normal form
^ + my + ^ _ p = Q
of a plane
Then the equation
/
through
)
P^ parallel to (1)
is,
by
(5),
99,
(2)
(x
 x^ + m (y  yi + n(z zj = 0,
in the
which may be put
Ix
form
(Ix^
+ my + nz
+ my + nz^ =
l
0.
(3)
Since equation (3) is in normal form, it tance of this plane from the origin is lx^
distance between the two parallel planes
is
evident that the dis
+my^+ nz r
in
Hence the
is
magnitude
(4)
Ix^+my^+nz^
this result being positive
if
p,
the plane through l is beyond plane is on if the plane through origin, (1) the same side of plane (1) as the origin. But the distance between the two planes is the same as the distance of l from
P
from the
and negative
^
P
the given plane. Hence tance in magnitude.
lx v \my^\nz^
p
is
the required dis
192
If the
PLANE AND STKAIGHT LINE
equation of the plane
is in
the form
finding the values of
I,
m,
n,
and p, and substituting in
(4),
we have
magnitude of the required distance, being positive points on one side of the plane and negative for all points other side. If we choose, we may take the sign of the always positive, in which case we can determine for which
the plane the above result preferably the origin.
Ex.
2.
as the
for all
on the
radical
side of
point,
is
positive
by
testing for
some one
Find the distance of the point
(1, 2, 1)
from the plane 2x
+
6z
+
14
=
3y
0.
The required distance
is
7
Furthermore the point is on the same side of the plane as the origin, for if (0, had been substituted, the result would have been 2, i.e. of same sign as 2.
3.
0, 0)
tion.
Plane through a given line and subject Let the given line be
1
It/
to
one other condi
1
1
*
\
/
Multiplying the lefthand members of (1) and (2) by l\ and & 2 respectively, where \ and k z are any two quantities independent of x, y, and 2, and placing the sum of these products equal to zero,
we have
the equation
&gt;,)=0.
(3)
Equation (3) is the equation of a plane, since it is a linear equa tion, and furthermore it passes through the given straight line,
since the coordinates of every point of that line satisfy (3) by
virtue of (1) and (2). Hence (3) is the required plane, and it may be made to satisfy another condition by determining the values of
\
and &2 appropriately.
CHANGE OF COORDINATES
Ex.
3.
193
(0, 1, 0)
the line
4z + 3y + 2z 
Find the equation of the plane determined by the point 4 = 0, 2z  lly 4z 12 =0.
of the required plane
2
and
The equation
may
(2
be written
11 y
^ (4 x + 3 y + 2 z  4) + k
Since
(0, 1, 0) is
x
its

4z

12)
=
0.
(1)
a point of this plane,
ki
coordinates satisfy
(1),
and hence
+
23 k 2
=
0,
or ki
=
23 k 2
.
Substituting this value of k\ in equation
(1)
and reducing, we have as the required
Ex.
4
4.
= 0, + 1 = 0.
2z
Find the equation of the plane passing through the line 4z + 3?/ 12 = 0, and perpendicular to the plane 2 x + y 4z lly
be written
0,
+ 2z
2z
The equation
fci(4z
of the required plane
may
+ 3y + 2z 4) +
+
(3fc 1
fc 2
(2zlly4z12) =
(1)
or
(4fc 1
+
2fc 2 )z
_Hfc 2 )y +
(2fc 1
Since this plane
is
to be perpendicular to the plane
2x + y
2 z
+
1
=
0,
2 (4 ki
+
2 k2 )
+
1 (3
fc!

11
fc 2
)
2 (2
fci

4
fc 2
)
=
0,
whence
^3
=
7 &i.
(1)
Substituting this value of & 2 in
and reducing, we have as the required
equation
z8?y3z8 = 0.
108. Change of coordinates. 1. Change of origin without change of direction of axes. Let (x Q y z ) be taken as a new origin of coordinates, the new axes being parallel respectively to the original
, ,
axes,
to
i.e.
OX
parallel to
OX,
and
Y
parallel to
OY, and O
z
1
Z
parallel
OZ.
Let
x, y,
and
z
be the coordinates
let
of
any point
P
with
respect to the original axes,
of the
x y
,
t
and
be the coordinates
same point with respect
to the
new
,
axes.
Z
Then
Q
,
= Z +Z
(1)
the proof being similar to that of
2.
I,
112.
Change of direction of axes without change of origin. Let OX, OY, OZ, and OX OY OZ be two sets of rectangular axes and making angles with each other, whose cosines meeting at
,
,
194
PLANE AND STRAIGHT LINE
where
is
the cosine of the angle ^ the cosine of the angle between 2 etc. Let x, y, and z be the coordinates of any point with respect to the axes OX, OY, OZ, and let x y and z be the
are given in the following table,
between and OY
OX
and
OX
,
1
is
,
OX P
1
1
,
,
y
ra 3
It,
to the axes OX Y drawn, the projection of OP on OX is x, and the projection on OX of the broken line and P, formed joining by constructing the coordinates x y z is Ijc + I 2 y + I 3 z (
coordinates of the
same point with respect
,
,
OZ
.
Then
if
OP
is
f
f
,
,
,
92).
But these two projections are
equal.
Hence
in like
manner,
y
z
m^x
= n^x
+ m y f m z +ny+nz
r
f ,
2
s
(2)
j
r
z
s
.
In like manner
we may
derive the formulas
(3)
=
Formulas
l
s
x
+
(2) and (3) may be expressed concisely by the above table. Since both systems of coordinates are rectangular, we have, by
and
m + m + m = 0, m % + m + ^ = 0,
l
l
l
/
2
2
/
3
3
1
1
2
7i
2
??i
3
3
PROBLEMS
Also,
195
1,
by
95,
/*
+m +
2
n* =
and
2
1
J
+Z
2 2
+^l,
2 3
m +m +m
=l,
&lt;+&lt;+&lt;=!.
All these formulas are easily remembered by aid of the above table.
PROBLEMS
1.
Find a point of the plane
(0, 0,
three points

1), (3, 1, 1),
(2,
2x43^ + 22 =  1, 0).
z
1,
equally distant from the
2. Find a point on the line 3 x y distant from the origin and the point
3.
5
=
0,
x
y
+
z
5
=
(
equally
(2,

2).
1, 4, 4),
Find the equation of the sphere passing through the points
1,3), (4,0, 7),
(5,
4.
A
(1,1, 5). point moves so that its distances from two
k.
constant ratio
k
Prove that
its
locus
is
fixed points are in the a sphere or a plane according as
*
1
or k
=
1.
Prove that the locus of points from which tangents of equal length can be drawn to two given spheres is a plane perpendicular to their line of centers.
5. 6.
Find the length of the curve x
t
=
t
2
,
y
=
2
1,
z
=
t
from the origin
to the
point for which
7.
1.
for which
Find the length of the curve x t = and t 1.
=
e*,
y
=
=
e~*, z
= tVz
=
between the points
8. Find the length of the conical helix x the points for which t = and t = 2 TT.
tcost, y
tsint, z
=
t
between
9.
Prove that a straight line can make the angles GO
,
45, 00 respectively
of the
with the coordinate axes.
10.
Show
that the helix
it is
makes a constant angle with the elements
cylinder on which
drawn.
11. Find the angle between the conical helix x the axis of the cone.
12.
z
t
=
tcost, y
=
sin, z
=
=
t
and
Show
that the angle between the conical helix x
is
=
tcost, y
tsiut,
and the element of the cone
tan
1
=.
V2
13. Find the equation of a plane three units distant from the origin and per pendicular to the straight line through the origin and (2, 3, 0).
196
14.
PLANE AND STRAIGHT LINE
The equations
x y
z
7
and
intersection
and
32 3 ?/ + 5 z = 3, of three planes are x } 2y 1, 2 x Find the equation of the plane through their point of equally inclined to the coordinate axes.
=
=
2.
15. If the
cepts a,
1
6,
normal distance from the origin and c respectively on the axes
to the plane
which makes inter
z
is
of x, y,
and
p, prove that
_
=
1
+
1
+
1
^~&lt;*2
&2
c2
16.
z
Find the equation of the plane normal
to the helix x
a cos 0, y x
a sin
0,
kd at the point
=
0.
17.
Find the equation of the plane normal
,
to the conical helix
=
tcost,
y
t
sin
z
=
t
at the point for
which
t
=
to the curve
18.
Find the equation of the normal plane Find the equation of the normal plane which t = 1.
x
=
e*,
y y
e
( ,
19.
z
to the
skew cubic x
=
,
=
&,
=
ts
at the point for
20. Find the direction cosines of the line
f
4
21. Prove that the three planes = are the lateral faces of a triangular prism.
22. Prove that the two lines
+ 3z 7=0. 3x2?/l=0, 4y3z + 2 = 0, z 2x
2x +
3y
z
6,
=
x
x
2X
+ y 7z + 11 = 01 + 4y + z3 = 0j
a
+ 17 y + 26z48 = 01 \4x 2y 16z + 26 = OJ
f
x
are coincident. 23. Prove that the lines
+i = oi
intersect at right angles.
1
r
2z/3z +
5
=
o^i
=
OJ
\7x4y*10 = OJ
two given straight
24. If
lines,
Z]_,
mi, n\ and ^,
m2
,
n% are the direction cosines of
and
/,
m, n are the direction cosines of a normal to each of them, prove
I
that
=
m
n
25. Find the equation of the plane passing through
(1, 4, 1)
perpendicular to
26. Find the angle between the line
x2y8 = 0,
3^ +
z
+
8
=
and the
(0, 1, 0)
27. Find the coordinates of a point on the straight line determined by and (2, 3, 2) and 1 unit distant from (1, 1, 1). 28. Find the equation of a plane perpendicular to the straight line joining and (4, 3, 2) at a point one third of the distance from the first to the
(1, 3, 5)
second point.
29. Find the equations of a straight line passing through
(1, 2,

1) parallel
PKOBLEMS
30. Find the foot of the perpendicular
197
(1, 2,
drawn from
1)
x
3yz
7
=
to the plane
0.
z
=
31. Find the equations of the tangent to the skew quartic x 4 at the t point for which t = 1.
32. Find the equations of the tangent to the curve x
=
2
1,
,
y
z
=
t
3
,
=
ef,
t
2
,
y
=
at
the point for which
t
=
1.
33. Find the equations of the tangent to the curve x 34. Find the direction of the conical helix x
origin,
=
y
=
e~ f z
,
=
t
tV%.
at the
=
t
cos, y
=
tsin
t,
z
=
and the equations of the tangent.
35. Find the equation of the plane determined by the three points
(3,

1, 2), (2, 1,

(1, 2,
4),
2).
36. Find the angles made with the coordinate planes by the plane deter mined by the three points (1, 2, 0), (4, 1, 2), ( 2, 2, 2).
3 7. Find the point of intersection of the lines
(
2xy3 = 01
5
i37/2z +
= 0/
a
1
\
r3x2Vy5 = 01 2xzl = o
by the
lines
38. Find whether or not a plane can be determined
39. Find the equation of the plane determined by the two lines
r
+ +
1 l
= 01 = 0/
a
straight line passing through the point nate axes.
40. Find the equation of the plane determined by the point (2, 4, 2) and the (1, 2, 3) equally inclined to the coordi
2x + y + 3z =
x
41. Find the equation of a plane passing through the line x  y and perpendicular to the plane + 2z + 1=0.
+z=
0,
xy
+ 2y +
1)
42. Find the equations of the projection of the line x 2 = z 1 = 0. upon the plane 3x + y + 3z
(2,
+
y
+ z2 =
and (
0,
43. Find the equation of the plane passing through 3 y f 2 2 perpendicular to the plane 2x 6 = 0.

1, 2)
1, 2,
2x

44. Find the equation of a plane passing through the line + 3 y 2z 1 = and parallel to the line 3x + y + 2z z + 5 = 0. 45. Prove that the plane
(x
x2y + z3 = 0,
4
=
2x
3y

2)2
4zf3y + 5z = 47is tangent + (y  3)2 + Z + 4)2 = 50.
(
to the sphere
46. Find a point on the line xz + 3 = 0, from the planes 3x + 3z5 = and x + 4y + 47. Find the center of a sphere of radius
(2, 4,
4xy6 =
z=7.
7,
equally distant

4)
and
(3,

1,

passing through the points
4)
and tangent
to the plane
3x6y + 2z+51=0.
CHAPTER XI
PARTIAL DIFFERENTIATION
109. Partial derivatives.
Consider f(x,
if
may, independent variables. We thus reduce f(x, y) to constant. vary, holding y temporarily a function of x alone, which may have a derivative, defined and as for any function of one variable. This derivative is
We
we
where x and y are choose, allow x alone to
y),
computed
called the partial derivative of f(x, y} with respect to x,
and
is
denoted by the symbol
^
V ex
Tnus b 7
&gt;
definition,
_ Lim
ex
Similarly,
if
(1)
function of
y,
x is held constant, f(x, y) becomes temporarily a whose derivative is called the partial derivative of
to y,
f(x, y) with respect
denoted by the symbol
&gt;
dfix y}
Tnen
(2)
 Lim f(x
dy
Graphically,
tion
if
y
~
Ay
)
A/=O
z=f(x,
y) is represented
by a
surface, the rela
between z and x when y is held constant is represented by the curve of intersection of the surface and the plane y = const., and
is
the slope of this curve.
is
Also, the relation between z
and y
dx
when x
constant
is
of represented by the curve of intersection
the surface and a plane x
= const., and
dy
is
the slope of this curve.
Thus, in
z
fig.
66,
=f(x, y}, PQ
is
the curve y
Let
P
be the point
PQSR represents a portion of the surface = const., and PR is the curve x = const. and L K = PK = Are, LM = PM = Ay. (x, y, z),
if
r
198
PARTIAL DERIVATIVES
Then
199
y
LP=f(x,
,
y),
KQ=f(x + bx,
y),
y)f(x,
M R=f(x,
y),
MR=f(x,
+ Ay),
and
y+Ay)/&gt;, y),
FIG. 66
Ex.
1.
Consider a perfect gas obeying the law v
=
.
We may
change the
temperature while keeping the pressure unchanged. The relation between the volume and the temperature is then represented by a straight line on fig. 65. If A and Au are corresponding increments of t and v, then
A,
and
dt
=
Or, we may change the pressure while keeping the temperature unchanged. relation between the volume and the pressure is then represented by an hyperbola on the surface of fig. 55. If Ap and Aw are corresponding increments
The
of
p and
v,
then
ct
ct&p
P + Ap
and
dv
ct
dp
So, in general,
if we have a we may have ,),
function of any
number
of variables
#
y&gt;
a partial derivative with respect to
200
PARTIAL DIFFERENTIATION
These derivatives are expressed by the sym
or sometimes
each of the variables.
bols J
J&gt;
&gt;
y
&gt;
dx
&gt;
dy
"
&gt;
&gt;
dz
z)
by fix, y,,z), fy (x,
derivatives,
y,
,
z),
fz( x
y&gt;
To compute these
we have
to
of one variable, apply the formulas for the derivative of a function as constant all the variables except the one with respect regarding
to
which we
Ex.
2.
differentiate.
/=
dx
x*

3 x*y
+
y3
,
Ex.
3.
/=
d
sin (x 2
+
y 2 ),
2
^ = 3x2 _ Qxy

L
dx
.
=
2xcos(x
+
2
2/ ),
=  3x2 + 3y2
Ex.
4.
/=
log
Vz 2 +
y
2
+ z2
,
x2
+ y2 +
z2
110. Increment
variables.
and
total differential of a function of several
y),
Consider f(x,
ments
A# and
Ay.
Then /
x and y be given any incre takes an increment A/, where
and
let
A/ =/( + Aaj,
In
fig.
y
+ Ay) /(a;,
and
y).
66,
NS=f(x + kx,y + ky}
N S = bf.
If
a;
and y
are independent variables, A# and Ay are also independent. the position of S in fig. 66 depends upon the choice of
Thus LK and
LM, which can be taken
The function f(x,
if A/ approaches as a limit in any
at pleasure.
x and y y} is called a continuous function of a limit when A# and Ay approach zero zero as
whatever.
manner
if z is
Thus
in
fig.
66,
a continuous function of x and y, the point
S
will approach the point
P
as
LK
matter what curve the point
point
N traces
and approach zero, no on the plane XO Y or the
LM
S on
the surface.
INCREMENT AND TOTAL DIFFERENTIAL
The expression
for
201
A/ may
be modified as follows
t
:
A/ =f(x + &x,y + Ay) f(x y) =f(x + &x,y + Ay) /(a?, y + Ay) +f(x y + Ay) /(,
t
y).
But
x Lim f(
+ Aa;,
y
+ Ay) /(a;,
y
+ Ay) = 3/(a?,
y
+ Ay)
Therefore
or
/(a + Ax, y
e
+ Ay) /(, y + Ay) =
\
7)f
OX
a
+
J
fix,
where Lim
Ax=o
=
0.
Also,
&gt;
since
is
OX
,
continuous function,
= V( x ^ +
dX
,
where Lim
Ay=
e"
=
0.
Therefore
y
/A
.
+ Ay) f(x,
y
+ Ay) = =
where
el
=e +e
=
Similarly,
/(*, y
0.
+ Ay) f(x,
y)
i^^l +
e 2 Ay.
,\
Ay,
where Lim
Aj/=0
e2
Hence we have
finally
A/ = ^ A* +
In like manner, y, ..,, then
if
^ Ay + 6.A* +
(1)
/

is
a function of any
number
of variables
x,
A/= ex A^ +
Ay +
cy

+
cz
A^
+ A^ + e Ay +
l
2
..+
en Az.
(2)
tion of one variable
In a manner analogous to the procedure in the case of a func ( 4), we separate from the increment the
terms
e^x + e Ay +
2
+ en Az,
call
the remaining terms the dif
The differen ferential of df. tials of the independent variables are taken equal to the increments, as in 4. is a function of Thus, we have by definition, when
the function,
and denote them by
/
two independent variables x and
y,
202
and
if
PARTIAL DIFFERENTIATION
/
is
a function of the independent variables
..*
x, y,
&gt;
z,
+
d
dz.
ox
oy
dz
(4)
be given any values whatever. one equal to zero, we have the If, in particular, we place all but Thus, partial differentials, indicated by dj, d y f, etc.
In
(3)
and
(4) dx, dy, etc.,
may
dx
in the partial differential expresses approximately the change function caused by a change in one of the independent variables the total differential expresses approximately the change in the function caused by changes in all the independent variables. It
;
A
appears from (4) that the total differential
partial differentials.
is
the
sum
of the
Ex. The period of a simple pendulum with small oscillations
is
(Ex.
3,
82)
r=s
Small errors
dl
and
dg, in determining
I
and
g, will
make an
error in
T of
The
ratio of error is
dT
I dl
1
dg
r
111. Derivative of f(x, y)
i
i
2 ?
when x and y
are functions of
t.
We
have been considering f(x,
variables.
We
shall
now
y) as a function of two independent suppose that x and y are functions of a
of x and y are single independent variable t, so that the variations caused by the variations of t. Graphically, if z=f(x, y) represents
and x and y are independent, the point P(x, y) may tne entire surface. If, however, x = $$), y = $ 2 P is restricted to a curve on the surface, whose projection on point the plane XOY has the parametric equations x = ^(0* V = The equations of the curve on the surface are, therefore, x =
a surface
move over the
(0&gt;
&lt;MO
&lt;f&gt;i(t),
DERIVATIVE OF A COMPOSITE FUNCTION
203
In this way, f(x, y) is now a function of t and may have a derivative with respect to t, which may be found as follows
:
Give
Ay, and
t
an increment
Atf.
Then x and y take increments A# and
an increment A/, where
2
f
in turn receives
A/= ^ A^ + ^ Ay + e^ + e Ay. 0#
tfy
Then
A/ =
A&lt;
y^ + /^ + ^ +
4 4
&lt;teA&lt;
e2
^.
A&lt;
fyA&lt;
A&lt;
By
limits,
allowing
A
d*
to approach zero as a limit,
and taking the
we have
=y
&e
^ y^
ott
ay dt
of (1)
If
we multiply each term
dt dt
by the
differential dt,
we have
.
f dt= v^ dt+ dy d, v^ dx
dt
...
:
But
by
since /, x
4,
t
and y are functions
,
, ,
of a single variable
,
t,
we
have,
,
7
,
,
a/
df=^dt, dt
Hence we have
df = ^dx
dx
dx
,
= cfe dt,
7a
dt
dy
= fdt.
&lt;iy
7j
dt
+
dy
dy,
(2)
110, made on the hypothesis that showing that formula (3) of x and y were independent, holds also when x and y are functions
of
t.
In a similar manner, if we have f(x,y,,z), where are all functions of t, then
x, y,

,
z
=
dt
dx dt
dy dt
dz dt
**+**+!*
Ex. Let F(x,
field,
w
y, 2)
be the electrical potential at a point in an electrified
that
is,
let
^
=  X,
~=
F,
=
Z,
where X, F, Z are the com
ponents of force in the directions of the coordinate axes. Required the rate of change of F in a direction which makes the angles or, /3, 7 with the axes.
204
PARTIAL DIFFERENTIATION
straight line
106,
A
by
making the given angles with the axes has the equations,
x
(1),
= Xi + r cos a, y = yi + z = zi f r cos 7.
?/,
If these
r,
values are substituted for x,
and
is
z in V,
.
it
becomes a function
of
and
its
rate of change with respect to r
By
dz
(3),
dV__8Vdx
dr dx dr
cV
dy
dV
/3
cy dr
dz dr
=
X cos a Y cos
Z
cos 7.
By the principle of the composition of forces this force in the given direction.
is
minus the component
of
z
112. Tangent plane to a surface. Let x y) be a curve on the surface z =/(#, y), and let =/(#,
(f&gt;M&gt;
y
=
&lt;
2
(0&gt;
^ (x v y v zj
be a point on the curve. Then the tangent line to the curve at the point JJ has the equations ( 105)
x
xl
=y
y
t
=z
dz
zl
^
dx
or
dy
The
line (1) is perpendicular to the straight line
xx _yy1 = zz1 ~1 d
l
l\
dxl
m
}"
by
(5),
98,
and therefore
lies in
the plane
Equations
(2)
define the curve,
and (3) are independent of the functions which and therefore the tangents to all curves through
PI lie hi the plane (3). This plane is called the tangent plane to the surface, and the line (2), which is normal to it, is called the normal line to the surface at the point J^. The tangent plane
TANGENT PLANE TO A SURFACE
may simply touch
ellipsoid, or it
205
may
the surface, as in the case of the sphere or the intersect the surface at the point of tangency,
as in the case of the hyperboloid of
one sheet.
In order that the function f(x, y) shall have a maximum or a minimum value for x = x v y = y v it is necessary, but not suffi
cient, that the
tangent plane to the surface z =f(x, y) at the point should be parallel to the plane XO Y. This occurs when (x v y v z^
^
=
)
0.
These are therefore necessary conditions for
mum
maximum or a minimum, and in case the existence of a maxi or a minimum is known from the nature of the problem, it may be located by solving these equations.
a
z
=
Ex. 1. Find the tangent plane and the normal line ax 2 + by 2
.
to
the
paraboloid
Here
dx
=
2 ax
and
cy
(x
=
2 by.
Hence the tangent plane
(y
y\] 2 by {
(z
Zi)
is
Xi) 2 axi
+
=
=
0,
0.
or
But since 2 ax}
+ 2 byiy 2 ax} 2 by} z + + 2 by} = 2 z l5 this may be written
2 axix
z\
The normal
Ex.
in the
x
is
 Xi _
y

yi
_z1
2axi
2.
2 by i
It is required to construct
out of a given amount of material a cistern
top.
if
form of a rectangular parallelepiped open at the
dimensions in order that the capacity
may
be a
maximum,
Required the no allowance is
made
for thickness of the material or waste in construction.
x, y, z be the length, the breadth, and the height respectively. Then the 2 xz 2 yz, which may be placed equal to the given superficial area is xy
Let
+
+
amount
of material, a.
If v is the capacity of the cistern,
Then
dx
(a
~
2 xy
~
y)
2
2(x
+
dy
it is
x
=
For the maximum these must be zero, and since 0, y = 0, we have to solve the equations
a
2 2
not admissible to have
xy
x2
a
xy
y
2
= 0, = 0,
206
PARTIAL DIFFERENTIATION
positive solutions
which have for the only
It is
z=y=*p,
;
whence
z^
Consequently, very evident that a
if
there
is
maximum capacity, it must be for these dimensions. maximum does exist hence the problem is solved.
a
in a
The graphical interpretation of the differential can now be given manner analogous to that in the case of one variable ( 5).
fig.
In
67
let
PQ S R
z
be the tangent plane at P(x,
y, z).
Then,
FIG. 67
if
z
and LM=dy, the coordinates of S are (x + dx, y + dy, and the value of z corresponding to S is found by replacing x by x\dx and y by y + dy in (3). There results
LK=dx
+ A),
z,
for
*
z
+ /dz\
(
7
I
efcc
+
/02
(
Therefore
Vay
\0y
ZV fl
=
&lt;fe,
whereas
N S = A*
and and may have the
an increment As.
takes an incre
and and
113. Derivatives of /(*, y) when * and y are functions of s If # and y are functions of two t. independent variables s
t,
then /(a?, y)
is
Q/
1
also a function of s
CH /*
t
derivatives
~
ds
o^
and
dt
To
find
3s
we
will give s
Then x and y take increments Az and Ay, and ment A/. As in 111, we find
T As
/
=
a^c
As
h
T
^y As
r
1
+
,
As
+
2
As
A
DEKIVAT1VES OF A COMPOSITE FUNCTION
Now
let
207
As approach
df
&gt;
zero
and take the
&gt;
limit,
remembering that
A/ = Lun
As
Lim
Az =
As
dx
etc.
^, , We have
ds
ds
/
ds
= 3/^ + ^^.
d#
ds
dy ds

df
Similarly,
J
dt
dy = df dx + df ~ ff
dx
dt
dy
dt
So, generally, the partial derivatives of f(x, y, u t are z are functions of s, t, , , y,
,), where
x,
/
ds
df_
=
^
dx ds
dt
+^
d
a
l
+...
+
^,
dz ds
cy ds
= df_dx
dx
dfty
dt
tf^,
dz dt
/2)
dt
dw
d^c
du
dy
die
dz
du
A
t
x where x
special case worth noting is that in which u. Then is a function of s, t, ,

/
is
a function of
= ^??,
ds
cfo ds
5/
a5
= ^^,
c?aj
^ = *^.
du
a^
dx du
(3)
If we multiply the equations (2) by the differentials ds, dt, 110, we have du, add the results, and apply the definition of
,
df=*dx + &dy +.. + dz ex
dy

&;&lt;&
(4)
This shows that the expression for the differential df z are independent variables or not. whether x, y,
,
is the
same
Ex.
1.
if z
=f(xy,y x),
= u,y
x
prove
z
+
ex
=
dy
u),
0.
Place x
y
=
v.
Then
=/(M,
and
a2_^awa/aw_^_^
dx du dx
dv dx
du
dv
fo__tf_fa.tf_to____3f_
dy du dy
is
tf
dv
dv dy
du
By
addition the required result
obtained.
208
Ex.
2.
PARTIAL DIFFERENTIATION
Let
it
polar coordinates
be required to change from rectangular coordinates = r cos 0, y = r sin 0. (r, 0), where x
ax
dr
dr
(x,
y) to
Then
dy
dd
(1)
and consequently,
if
/
is
a function of x and
y,
dr
dx
dy
(2)
dd
dx
6
dy
Also, since r
= Vx 2
f
y 2 and
=
tan * X
,
= sin 0,
(3)
whence
Equations
It is to
(4)
may
also be obtained
from
(2)
by solving for
and
dx
in
.
dy
(3).
be emphasized that
in (1) is not the reciprocal of
In (1)
means
dr
the limit of the ratio of the increment of x to an increment of r
when
u
FIG. 69
e is constant.
Graphically
(fig.
68),
PR = Ax
Also
is
thus determined.
in (3)
Then
OP = r = Lim
is

increased by
PQ =
Ar,
and
=
cos
9.
means the
limit of the ratio of the increment of r to that of x
(fig.
when y
is
dx constant.
Graphically
69),
OM=x is increased
by MN=PQ=Ax,
PROPERTY OF THE TOTAL DIFFERENTIAL
,
T&gt;
209
r\
iii
RQ = Ar is thus = But that
id
determined.
in (1)
Then
dx
= Lim  =
cos
6.
It
happens here
and
is
in (3) are neither equal
nor reciprocal.
dr
dx
d6
In cases where ambiguity
in a partial derivative, the
likely to arise as to
symbol for the derivative
is
which variable is constant is sometimes inclosed in a
/
parenthesis and the constant variable
written as a subscript, thus
(x, y).
jfirst
Ex.
3.
Consider f(x, y,
z)
when z =
We may find
Then
two
1
Pf
from the
1.
of formulas (2) in that
we
place s
=
x and
t
=
pf ex
y.
=1 =
Direct
substitution in (2)
left of the
would yield the symbol
it
in
Cl CS different senses.
On
the
equation
means the
y
constant, and attention is derivative of / with respect to x on right of the equation it means the partial the assumption that both y and z are constant. Ambiguity is avoided by the
is
partial derivative of / with respect to x when given to the fact that z is a function of x. On the
use of subscripts as suggested at the close of Ex.
2.
Thus we have
\W\
dX/y
\dX/yz
+ Wte.
CZ dX
114. Property of the total differential.
An
of the total differential is expressed in the following
important property theorem
:
If f( x
then
1.
&gt;
&gt;
y&gt;
z)
= c for
first
a ^ va ^ ues
f
th e independent variables
t
df=
0.
Let us suppose
that x, y,
,
,
z are
the independent

variables.
Then f(x,
f(x
/(a?
y,
=c z)
,
for all values of x, y,
.
y&gt;

,
z.
Hence
+4^
+ Aa;,
y&gt;
...,*)

_ /(aj)
,
z)
z}
= 0,
y,
z)f(x, y,,
for all values of x.
Taking the limit as

A# =
0,
we have ir
fif
=
Similarly,
^=
dy
0,
.,
^=
dz
0,
and hence
dx
2.
dy
9
cz
y,

Suppose, secondly, that x

,
z are functions of the inde

pendent variables
f(x,y,
first
s,
t,
,
u.
Then
if x, y,
,
z are replaced in
,

,z)
case again,
by their values in terms %f and hence as before cs
of
s, t,
u,
0,
we have
,
the
=
%f
0, f
=
ct
%f ^cu
= 0.
210
Hence
PAKTIAL DIFFERENTIATION
df
= ^dx + ^dy +
dx
1
.
.
.
+ &dz
dz
dy
dt
=
ds It is to
& + * K...+gte0. du

be emphasized that the coefficients of dx, dy, not equal to zero.
115. Implicit functions.
defines
,
dz are
CASE
I.
/(#, y)
=
0.
The equation
f(x, y)=Q y as an implicit function of x, or x as an function of y, since if one of the variables is given, the implicit values of the other are determined. Then, by 114,
whence
dy /= dx
dx T?df
(1)
ay
Ex.
1.
Find the tangent and the normal
to the curve /(x, y)
=0.
By I,
100, 101, the equations of the tangent and the normal are respectively
and
y
(1)
_ 1/1 =
become
By use of
these equations
_X
\dx/i
+ (y y ,)
\dy/i
=0
and
CASE
II.
/(#, y,
2)
=
0.
The equation f(x, y,z)=Q
variables.
defines
any
one of the variables
will take
x, y, z as
a function of the other two.
We
114,
x and y as the independent
Then, by
dz=0. df=ydx + ydy + dz dx
dy
r,
But
Therefore
,
dz
7
=
dz
dx
7
\
dx
dz  dy.
,
K+
*WW
dy
+&
^W =
IMPLICIT FUNCTIONS
This
is
211
dx
true for all values of the independent differentials
and
dy.
Therefore
dx
+
*_
dz dx
and
dy
+ #**
dz dy
dz
whence
_ ~ =
dx
df
dz
(2)
~ty
dx
dz
Ex.
2.
Find the equations of the tangent plane and the normal
z)
line to the
surface F(x, y,
=
0.
By
112 the required equations are
~ (xxo Vi ~
and
7
+ (yi
i
yi )i^\
W/i
z
(**i) =
zi
o
y
y\
Using
(2)
and reducing, we have
and
(**.}
**.
i
=r
**
(**)
\d*Ji
vex/i
CASE
f^x,
III. f^(x, y,
y, z)
=
z)=Q,
0,
/
2 (a;,
y, z)
=
f (x,
The two equations 2 y, z)=Q. taken simultaneously define any two
of the variables as functions of the other one.
By
114,
dx
cy
cz
dx
Therefore
dx:dy:dz
=
(3)
212
Ex.
3.
)
PARTIAL DIFFERENTIATION
y,
/,(,
=
Find the equations of the tangent
0.
line to the space curve /i(x, y, z)
=0,
By
105 the required equations are
dx
dy
(3)
dz
where the values of
dec,
dy, dz given in
may
be substituted.
116. Higher partial derivatives. The partial derivatives of are themselves functions of x and f(x, y) y which may have
partial derivatives called the second partial derivatives of /(a?, y).
sslST/ j(/ 57(37] But it may be shown dx\dx/ dy\dx/ dx\dy/ dy\dy/ that the order of differentiation with respect to x and y is imma
They
1
are
%{)&gt;
)&gt;
terial
(
117), so that the second partial derivatives are three in
number, expressed by the symbols
dxdx
=
da?
,
Jm&gt;
x
"
Similarly, the third partial derivatives of
f(x
t
y)
are four in
number, namely
:
2
?
/2 /\
_ a/
a?
3
~
dy \dx*}
,
dx \dxdy)
a / a
8
dx \dy)
a
8
2
~
dx*dy
a
.
/
3
\
Sy \8xdy /
dfx
/a/\ =
/
}
dxcf
So, in general,
dxp dyq
signifies
the result of differentiating/(^, y)
p
times with respect to x and q times with respect to y, the order
of differentiating being immaterial.
The extension
to
any number
of variables is obvious.
HIGHER DERIVATIVES
117. To prove the relation ^^dxdy consider the expression
_
213
y=
ft,
=
^ for
any particular values x = a,
dycx
*)/(a,
(a, 6)
where for convenience h and
1
fc
are taken as positive.
to
^j
)
We
shall prove
I equal
to
on the one hand, and
on the other hand, where (1,
70.
771)
asedy/c.f,
\dydx/ x= $,
are two undetermined points within the rectangle of
fig.
and
(
2 , 172)
In the
first place,
let/(x
+
h, y)
/(x,
y)
=
F(x,
y).
Then
_
1
"
F(a,
ft
+
~~
fe)
F(a,
"
6)
whence, by
30, (2),
!
2*v(
)
fv (a +
h,
171)
/y (a,
where
ing
Fy (a,
171)
=
(
.
\a?//x=a
=iJi
)
Apply
30, (2), a second time,
we have
FlG 70

jfit
In the second place,
\
if
we
let/(x, y
+
k)
/(x, y)
=
&lt;I&gt;(x,
y),
we have
fc)/x(^6)_
Therefore /^(d, ^) ^ 2 ), since each is equal to 7. y;r ( 2 Now let A and fc approach zero as a limit. The points (&, ;!) and (^ 2 172) both approach the point (a, 6) as a limit; and since the functions are continuous
,
=/
,
,
77!)
= Lim/yx
(
2
,
772),
or
/*&lt;,&) =/(&lt;!, 6).
This result being proved,
differentiations.
it is
easily extended to
any number
of variables or
Thus, for example,
cx*cy
dxdydx
214
PARTIAL DIFFERENTIATION
of
118. Higher derivatives of /(*, y) when x and y are functions t, or of s and t. By 111, if x and y are functions of t,
d^&lt;^dx
dt
dx dt
3f_dy^ dy dt
t,
To
differentiate again
with respect to
we must
notice that
each term on the righthand side of (1) is a product and must be handled by the law of products (I, 96, (4)). We have,
then, in the
first
place
dy = df^xdxd_/df\,^fd^,dyd/df\
dt
2
dx dt
2
dt dt \dx)
2 dy dt
dt dt \dy)
Now
(
and
j
(
)
may
be found from
(1)
by replacing / by
fif
fif
and
dx
d
respectively.
Hence
d
2
dt
2
=
dx dt
dx/d 2fdx
2
/
dy \
dt \dx dt
2
dxdy dt)
dfd y dy df
2
dy/dy
=
dx\dt)
dx
d*fdy\
dy*
dt \dxdy dt
_,
dt)
.
&gt;
^_
if
/ON
dxdy dt dt
df\)
dx dt
2
dy dt
2
(
In a similar manner,
_
x and y are functions
_
2
of s
and
.
t,
=
s*
dx \ds)
2
2
__
dxdy dt
d
2
dxdy ds ds
dy \ds)
dx ds
2
dy
ds*
dt
2
dx \dt
2
dt
2
dy \dt
dx
d
2
dt
2
dy dt
2
d
2
/ = d fdxdx
dx ds
2
f
/dxdy
dt
dxdy\
dt
fdydy
2
dsdt
dt
dxdy\ds
ds)
dy ds dt
dx dsdt
dy dsdt
The extension
obvious.
of these formulas to
any number
of variables is
HIGHER DERIVATIVES
V
215
function of x and y,
i
Ex.
1.
Required to express
y
2
+
1
fix a
2
F y 
,
where
F is a
polar coordinates. Since r
8r
^y
= Vx 2 +
x
y 2 and
&r_
=
tan
(
)
,
_
_ ~
(x
2
y
2
a0
_ ~"
y y2
d26
dx
Vx 2 +
y
y2
Sx *
fPr
+
x2
y 2 )*
ax
x2
+
x
f
_ ~
(x
2
2 xy
ax 2
a 2^
+ +
y2 ) 2
^_
y
Plence,
a2
2
_
2
aj?
_
x2
_ (x
2
2 xy
Vx +
from
y
2
^2/
2
(x
+
y
2
)^
cy
y
2
ay
2
y2 ) 2
(3),
F
a2
F
+
x2
dx 2
cr2 x 2
+
y
2
2 n a r ara0
F
r?/
"^y
x2 (
+
y2 )*
dff2
(x
2
+
y2 ) 2
8r
(
x2
aF_^xy_
c6 (x 2
+
?/ )
2 2
a2
F_aF
2
y2
x2
a2
F
(
xy
x2
aF
x2
aF
x2
+
y* 2
drdB
aF
xy
Hence
ax 2
+
a?/
=
2
ar2
x2
i
+
"a^
y2
"
w
+
i
_a F
2
"
Vx^Ty2
aF
"a7
dr
a^F
2
ar2
r2
r
Ex.
2.
If z
=/i(*
+
.
o&lt;)
+/
2
(x
a&lt;),
where
/i
and
/2
are any two functions,
show that
dt 2
=
erf
a2
ax 2
Let x
+
=
11,
x

ai
=v
;
then
ax
 = = ^=a
1,
a,
ot
ax
1,
,
and
a
^2_^A du^dfo to_dfi ~ ax dw
ax
dfz
dv ex
du
du
~ =
dt
du
dt
dv
ct
du
_^
^dv
Differentiating these equations a second time,
&lt;Pz
we have
d2
=
d 2/! /aw\ 2
d%
/av\ 2
dx2
du 2 \ax/
=
^
JM
dw 2
dfc
dv 2
du 2 \ax/
du2
"
S s ffiM%.m/aiy dw \a/
a
2 2
M
d
2
du 2 \ae/
By
inspection the required result
is
obtained.
216
PARTIAL DIFFERENTIATION
Consider
119. Differentiation of a definite integral.
f(x, a)dx,
where a is a parameter independent of x. In the integration a is considered as constant, but the value of the integral is a function of a. Let a be given an increment Aa. Then u takes an increment AM, where
Aw =
Now
by
31, (1),
r.
s&gt;b
r*b
I
f(x, a f Aa) dx
I
f(x, a) dx
&lt;J
a
*J a
[f(z,
a
+ Aa)
f(x, a)] dx.
Hence
A, =
Ja
a
+
Dividing by Aa, and taking the limit as
limit,
Aa
approaches zero as a
we have
The proof assumes that a and
6 are finite.
It is not
always
infinite possible to differentiate in this way an integral with an limit. The discussion of this lies outside the scope of this book.
The
integral
25,
u
is
also a function of the
upper limit
6,
and we
have, by
rr
f(x,
).
(2)
a
a)dx
=
I
f(x, a)dx,
Jb
= /(,).
(3)
PKOBLEMS
Suppose now that Then, by 111, (3),
a,
b,
217
t.
a are
all
functions of a single variable
du
_ du db
\
du da
^a
du da
da dt
x
~di~~db~dt
dt^
s;
db
da
,
da C b df(x,
a}
,
Ex.
If
u
= flog (1 Jo
du da
2 a cos x
+
a 2 ) dx,
_ ~
P*
Jo
1
2 cos x
+
2a

2 a cos x
+
a2
+
a)
2
tan 2 
=
=
Therefore u
In this
TT
a
0.
2T + a tan X\IT 1 tan ,/1 al \i a 2/Jo
(
)
= const.
But when a
= 0,
w
=
/
f
o
(log 1)
dx
=
0.
Therefore u
= 0.
way the value of a definite integral can sometimes be found direct integration is inconvenient or impossible.
when
PROBLEMS
1.
Given z
= =
4
log(x
2
+
y
2
),
prove y
dx
x
dy
=
0.
2.
Given z
Given x 2 Given y
x4
+
2
x 2 y2
+
y4 prove x
,
5x
+y
Sy
ex
= +
4 z.
3.
y

.2
xy
+
2 z2
=
c,
prove
=
ay
.
4.
=
eax sin6x,
prove x
ax
=
a
aa
f
6
a&
5.
Given z
=
=
(x
2
+
+
y
2
)
tan 1 x
,
,
prove x
ax
+
y
ay
=2z.
6.
Given
z2
xy
ev
tan 1 x
1
prove zx
dx
+
f
zy
dy
= xy.
z
7.
Given z Given z
=
=
sin
(x
y),
prove
ax
=
ay
8.
y
2
f
2 ye*, prove x2
+
ax
y
ay
=
2 y2
.
218
9. If z
PAliTIAL DIFFERENTIATION
=
xy, illustrate the difference
y.
between Az and dz by constructing a
rectangle with sides x and
10.
A
is
angle
30.
triangle has two of its sides 6 and 8 in. respectively, and the included Find the change in the area caused by increasing the length of
each of the given sides by .01 in. and the included angle by 1, and compare with the differential of area corresponding to the same increments.
11.
A
angle
is
60.
triangle has two of its sides 8 and 12 in. respectively, and the included Find the change in the opposite side caused by making the given
12.1 in., the angle being unchanged,
sides 7.9
and
ential of area corresponding to the
and compare with the same increments.
differ
right circular cylinder has an altitude 12 ft. and the radius of its base Find the change in its volume caused by increasing the altitude by .1 ft. and the radius by .01 ft., and compare with the differential of volume corre sponding to the same increments.
12.
A
is
3
ft.
13.
The distance
of
measuring a base
in
line
BC =
an inaccessible object h and the angles
A
CBA = a
from a point B is found by and BCA = p. Find*
caused by errors of cZ/i, da, dp the expression for the error in the length of measuring h, a, /3, assuming that higher powers of the errors of measurement
AB
may
be neglected.
,
Verify the formula
14. z
dz
dt
=
y
,
dz dx
dx dt
h dz dy m each of the following cases .
&gt;
:
dy dt
15. z
16. z
= = = =
x2
+
2
?/
,
x
=
t,
sinxy, x
e2
= P. y = e st
cos
t.
.
V
ex2
x
= sin
,
,
y
=
17. z
+ v2 x
=
sin, y
=
cost.
18. Find the tangent plane to the cone z 19.
=
a
Vcc 2
f
y2
.
that the tetrahedron formed by the coordinate planes and any 3 tangent plane to the surface xyz = a is of constant volume.
Show
Show
20.
that
lines.
any tangent plane
to the surface z
=
kxy cuts the surface in
nearest the origin.
two straight
21
.
Find the point in the plane ax
+
by \ cz + d =
which
is
22. Find the points on the surface xyz
=
a which are nearest the origin.
3
23. Find a point in a triangle such that the from the three vertices is a minimum.
sum
of the squares of its distances
24. Of all rectangular parallelepipeds inscribed which has the greatest volume.
in
an
ellipsoid find that
25. Find the point inside a plane triangle from which the sum of the squares of the perpendiculars to the three sides is a minimum. (Express the answer in the area of the triangle, a, &, c, the lengths of the three sides, and terms of
JK",
x, y, z,
the three perpendiculars on the sides.)
PROBLEMS
26. If z
^./x\ =/(), prove x ax \y/
219
az
az
\
y
ay
=
0.
27. If f(lx
4 my
(ly
4 nz,
x2
4
4
y
2
2 4 z )
=
0,
prove
\
mx)
(ny
mz)
ex
(lz
nx)
= 0.
dy

28. If z
=
y2
4 2/( 4 \x
is
log
A
/
prove
 = 2y   ay
n,
y ax
a/ prove x dx

29.
If
/(x, y)
a homogeneous function of degree
4y
a/
=
nf.
dy
30. Given x
=
r
,
y
=
r
,
prove
/ax\
=
/8r\
\ar/
W/
\Sx/ v
2
,
(dy\
w/
\ce/ r
=_
/er\
W/
\dy
d0Jr
31. Given u

log
Vx +
2
y
u
=
tan
1
,
prove
32. Find the tangent plane to the ellipsoid
*
+
CL
?
/
^+
=
C
1
at the point
(*1,
2/1,
l)
33. Show that the sum of the squares of the intercepts on the coordinate a* is constant. axes of any tangent plane to x 4 y* +
$
34.
Show
that the
sum
gent plane to x^
4
if
y^
+
of the intercepts z* a is constant.
on the coordinate axes of any tan
=
35. Prove that the plane
?L
Ix
+ !^4.^
2
=
1
p = Va2
Z
2
2 4 b
+ my 4 nz = p m 2 + c%2
.
is
tangent to the ellipsoid
36. Prove that the plane Ix
4 my +
nz
=p
is
tangent to the paraboloid
ax
4
by
2
=
z
if
p =
bl 2
4 am 2
straight line
is
37. Find the cosine of the angle between the normal to the ellipsoid and the drawn from the center to the point of contact, and prove that it
,
equal to
where p
is
the distance of the tangent plane from the center
and
r the distance of the point of contact
from the
center.
38. Find the angle between the line drawn from the origin to the point = a 3 and the normal to the surface at that point. (a, a, a) of the surface xyz
(x

39. Find the angle of intersection of the spheres x 2
2
4
y
2
4
z2
=
a2 and
6)
4
y
2
4
2
=
c2 .
220
PARTIAL DIFFERENTIATION
+
z 2x 2
40. Prove that the families of surfaces x 2 y 2
intersect everywhere at right angles.
= c\
z}
and x2
y2
z2
= c^
41. Derive the condition that two surfaces /(x, y,
intersect at right angles.
=
and 0(x,
y, z)
=
42. If /(x, y,
z)
=
0,
show that
z)
(\ (^} (?*} = _ L \ax/j,\c?// W*
2
43. If /(x, y}
=
*
and 0(x,
= 0,
and z
is
taken for the independent variable,
how ttat*
=
g2
dx dy dz
2j2
?*. dx dz
44. Find the equations of the tangent line to the curve of intersection of the
ellipsoid
a2
y2
{
f
=
1
o2
c2
and the plane
Ix
+ my +
nz
=
0.
45. Find the equations of the tangent line to the curve of intersection of the
2 cylinders x
f
y2
=
a 2 y2
,
+
z2
=
b2
.
46. Find the highest point of the curve of intersection of ^
x2
and
xa
Ix
+ my +
nz
=
+ ^H
y2
z2
=
1
0.
f
47. Find the highest point of the curve of intersection of the hyperboloid z 2 = 1 and the plane x + y + 2z = 0. y%
48. Find the angle at which the helix x 2
the sphere x 2
+
y2
=
a2 z
,
=
=
k tan
1
 intersects
+
y2
+
z2
=
r 2 (r
&gt;
a).
49. Find the angle at which the curve y 2 the surface x2 4 2 zy = c.
z2
=
a,
x
b (y
+
z) intersects
Verify
 =
dxdy dydx
in each of the following cases
:
~
51. z
52. z
x
+
y
= =
log
Vx 2 +
_
y2
53. z
.
= =
8111!.
x
KA 54. z
ex
smy.
55. If z
=
Iog(x2
(e
+
y2 )
+ tani, x
prove
\
gx2
2
+
= 0.
cy
5
56. If z
57. If z
= =
x

e~ x ) cosy, prove
 = 0.
dy
dx2
at),
sec (x
58. If z
CO
rr
/ = Vx
y
at)
5
+
tan (x
3*
+
Z
prove
dZ Z
y 2 prove
,
& &  = dx dxdy
y),
^ = a 
2
.
dc 2
Sx4
dy dx*
59. If z
= =
sin
+
e~
cos (x
prove

H

(
= 0.
60. If z
4
61. Given x
+
.
logx

ev\ prove
=
cxcy
V with
= e cosu, y respect to x and y.
u
=
eM sinu, find

xy.
in terms of the derivatives of
PROBLEMS
62. Given x
221 +
=
c2
2 i
=
e
cos
,
y
=
e
sin
,
prove
a2
_L +
MD
63. Gi
,
F
F
c2
F
64. Given x
= re
o
jf.
e e
,
J
a2
y
e = r
e
e~Q
^
,
prove
F ~ a2F _ 82F ~ ~~
5^2
1 g2
F
1
8F
cr
"^2
~^2
v)
^
02
r
65. If x
tions
=/(w,
,
u)
and y
,
=
&lt;f&gt;(u,
are two functions which satisfy the equa
of
^ = ?*
aw
du
8u
and
F
is
du
any function
a2
x and
y,
prove
&v
"
ew2
+
d*v
av 2
=
/a
2
F
F
2
2 \ ax
a?/
.
67. If z
=
0(x
+
iy)
+
t(x

iy),
where
find
i
=
.
V/
1,
az
prove
+
a^ =
0.
68. If u =/(x, y) and y
= F(x),
d 2 ?/
d*u
_
69. If /(x, y)
n
=
0,
prove
_=
 ax 2 \ay/
axay ex Sy
2 ay \gx
\cy)
70. Differentiate
w
=
/**

log (1
o
+
a cos x) dx with respect to
a,
and thence
find
/
the value of the integral.
Q*+A
77
L&lt;r
~
71. Differentiate
of the integral.
f
Jo
lo %
x
dx with respect
to a,
and thence
find the value
M
CHAPTEE
XII
MULTIPLE INTEGRALS
120. Double integral with constant limits. Let f(x, y) be any function of x and y which is single valued and continuous and posi tive for all pairs of values of x and y for which a^x^b and c^y^d.
Divide the interval
Aa?,
b a into n equal parts, denoting each part by thus forming the series of values of x, x v x 2 xs ,xn _ l where
&gt;
,
,
,
c into m equal parts, Similarly, divide the interval d denoting each part by Ay, thus forming the series of values of y, y v y 2
&gt;
,
where
The above values
allel to
of
x determine a
of
OY, and the values
series of straight lines par determine a series of straight lines y
parallel to
OX
(fig.
71).
Every
line of either set intersects
every line of the other set, and any one of these points of intersection
ya
may
to
l
where
be represented by ^(^, y^), i has all integral values from
n
1
and j has
to
c.
all integral
1,
&lt;rx
values from
m
XQ being a
FIG. 71
and y being
Taking the value
each point of intersection,
f(a,
of f(x, y) at
we form
the series
c)kxky +f(a,
c)
+f(xv
,
A^Ay +/(^,
y,
c
.
(I)
222
DOUBLE INTEGRAL WITH CONSTANT LIMITS 223
This series can be expressed more concisely by the notation
~
where two
s
are used, since there are
two elements
i
and j
which vary. The limit
of (2) as
m
called the double integral of f(x, y) over the area lines x d. a, x b, y c, and y
and n are both increased indefinitely is bounded by the
=
=
=
=
of the terms in (2) may evidently be made in ways, but there are two which we shall consider in par ticular: (1) when the sum of the terms of each row is found, and
The summation
many
these
sums added together (2) when the sum of the terms in each column is found, and these sums added together. It will appear from the graphical representation ( 121) that these two methods
;
lead to the
same
result;
and
it
may
it is
be shown that the result
is
always
If
independent of the order of
summation.
to be noted that the value
the
is
of
x
method is followed, the same in all the terms
first
of
any one row, and hence each
row
is
exactly the series used in defining a definite integral
(21)
let
with y as the independent variable.
increase indefinitely, (2)
Accordingly,
when we
m
becomes
(jf
/(,
y)dy^
/to,
+
Qf f(x v
+ +
(jf
/("V*
&lt;
+
But
3
&gt;
( jf
y)dy)**
I).*)**
(3) is the series
used in defining a definite integral with x
Letting
n
increase indefinitely,
/af(x y)dy we have
}
as the function of
x.
which represents the double integral on the hypothesis that and then n is made to increase indefinitely.
first
m
224
Another way
MULTIPLE INTEGRALS
of writing (4) is
nd
where the summation
right to left,
i.e.
f(x,y)dxdy,
(5)
first
is made in the order of the differentials from with respect to y and then with respect to x,
limits of
and the limits are in the same order as the differentials, i.e. the y are c and d, and. the limits of x are a and b.* If the second method of summation is followed, we have
b
f(x )y )dydx.
(7)
Ex. The moment of inertia of a particle about an axis is the product of its mass by the square of its distance from the axis. From this definition let us determine the moment of inertia of a lamina of uniform thickness k about an
v
axis perpendicular to its plane. Let the density of the lamina be uniform and de noted by and let the plane coin cide with the plane of the lamina, the axis
/&gt;,
XOY
being perpendicular to the plane at 0.
Let
the lamina be in the form of a rectangle (fig. 72) bounded by the lines x = a, x = &,
y
.
=
c,
y
=
d.
Divide the lamina into reclines
y
.tangles
by the
Then the mass
of any element, as PQ, is pkAxAy. If this mass is regarded as concentrated at P(X;, %), its moment of inertia would be (x? f yj)pkbxby. If the mass is regarded as concentrated at Q(x t + Ax, 2fr + Ay), its moment of

* Still another form of writing
(4)
is
f
/a
dx
t/e
f f(x,
y) dy, in
x.
which the order of
summation
is first
with respect to y and then with respect to
f(x, y)dydx, which is merely (4) with the parenthesis removed. In this form it is to be noted that the limits and the differentials are in inverse orders, and that the order of summation is the order of the differentials from left to right, i.e. first with respect to y and then with respect to x. In this text this last form of writing the double integral will not be used. In other books the context will indicate the form of notation which the writer has
chosen.
nd
GKAPHICAL REPRESENTATION
inertia
225
would be

[(x t
+
Ax)
2
+
(j#
+ Ay) 2 ]pkAx&y.
we have
1
t
Letting
M
represent the
moment
of inertia of the rectangle,
=n
i=0
j=
m
1
i=0
.7
=
j=Q and
The
limits of these
sums as n=oo,
Lim
Hence we
define
[(*
= cc, + A*) 2 + (x? +
ra
2
Ax= 0, Ay =
are the same
(
3),
for
M by the equation M= vC JC
a
c
(x
+
y*)pkdxdy.
is
If p
and k are each placed equal
to unity, the result
often referred to as the
moment
of inertia of the plane area
bounded by the
lines
x=a, x=6, y=c, y = d.
121. Graphical representation.
the equation of a surface
which
is
Placing z=f(x, y), we have the graphical representation
of f(x, y) (fig. 73). d a, x b, y c, y Through the lines x Then the volume bounded by the pass planes parallel to OZ.
=
=
=
=
FIG. 73
the planes x a, x = b, y = c, y d, and the surface =f(x, y) is a graphical representation of the double integral of 120. For if the planes x xv x x2 x xv y y v y ya are constructed, they divide the above volume, which y = y3
plane
XOY,
=
=
z
=
=
,
=


,
=
=
,
,
we
will denote
by V, into columns such as
MNQP,
each of which
226
MULTIPLE INTEGRALS
in the plane XOY. If the stands on a rectangular base coordinates of are (xit y^, the corresponding term of (2), the term/(^, ^.)Aa!Ay is/(^, %.)A^Ay, and since f(x{) i/j ) =
M
AAy
120&gt;
MP
)
is
MNQP,
the volume of a prism standing on the same base as the column and the volume of this prism is approximately the volume
of the column. Hence (2), 120, is the sum of the volumes of such prisms, and is approximately equal to V\ and as m and n both increase indefinitely, the limit of the sum of the volumes of these prisms is evidently V.
The significance For if the integral
of the
is
two ways
(5),
of
summation
is
now
clear.
first
written as
120, the prisms are
added together, keeping x constant, the result being a series of slices, each of thickness A#, which are finally added together to include the total volume and if the integral is written in the form
;
120, the prisms are first added together, keeping y constant, (7), the result being a series of slices, each of thickness Ay, which are
finally
(5)
added together to include the total volume. and (7) are equivalent, as was noted before.
It follows that
122. Double integral with variable limits. may now extend the idea of a double integral as follows Instead of taking the integral over a rectangle,
:
We
as in
120,
\
we may
take
over an area bounded by any closed curve (fig. 74)
it
such that a straight line
parallel to either
OX or OY
more
D
intersects
it
in not
A
\
than two points.
Drawing
straight lines parallel to
Y
and
to
straight lines parallel
OX, we form rectangles
area Aa;Ay,
of
some
of
M
FIG. 74
which
are entirely within
the area bounded by the
curve and others of which are only partly within that area.
Then
(1)
DOUBLE INTEGRAL WITH VARIABLE LIMITS 227
where the summation includes
all
the rectangles which are wholly
or partly within the curve, represents approximately the
volume
bounded by the plane XOY, the surface z =f(x, y], and the cylinder standing on the curve as a base, since it is the sum of the volumes
121. Now, letting the number of these prisms of prisms, as in and A# increase indefinitely, while A# 0, it is evident that
=
volume described above. (1) approaches a definite limit, the
If
we sum up
first
with respect to
y,
we add
together terms of
.
Then if is of x, such as x (1) corresponding to a fixed value xit the result is a sum corresponding to the strip the line x ABCD, and the limits of y for this strip are the values of y corre
{
ME
sponding to x
=x
),
and
MB =/ (#
2
i.e. in the equation of the curve the limits of y are/^aj,.) and /2 (# ).
i
; f
if
MA =f (x^
l
As
different
integral values are given to i, we have a series of terms correspond ing to strips of the type ABCD, which, when the final summation
is
made with
respect to x,
if
must cover the area bounded by the
the least and the greatest values of x for the curve are the constants a and b respectively, the limit of (1) appears in the form
curve.
Hence,
I
I
Ja
where the subscript
f(x,y}dxdy,
(2)
Jfi(x)
On
to x,
i is no longer needed. the other hand, if the first summation is made with respect the result is a series of terms each of which corresponds to a
strip of the
4&gt;i(y)
type
(y),
A B C D*
,
and the
limits of
x are
of the
form
in
and terms of
&lt;
2
y.
found by solving the equation of the curve for x Finally, if the least and the greatest values of y for
c
the curve are the constants
and d
appears in the
form
.
^
/
respectively, the limit of (1)
J/c
f(*,y)dydx.
(3)
i/fi(r)
While the limits of integration in (2) and (3) are different, it is evident from the graphical representation that the integrals are
equivalent. 123. In
120122, f(x, y) has been assumed positive for all the values of x and y considered, i.e. the surface z =/(#, y} was If, however, entirely on the positive side of the plane XOY.
228
f(x, y)
is
MULTIPLE INTEGRALS
negative for
is
all
the values of x and y considered, the
reasoning
integral
is
exactly as in the first case, but the value of the
Finally, if f(x, y) is sometimes positive sometimes negative, the result is an algebraic sum, as in
negative.
and
22.
Furthermore, be equal and
it is
all
not necessary that all the values of A# should the values of Ay equal also in place of f(xit y^
;
we may use /((.,
The work
of
is
17,),
where x
i
&lt;
&lt;
xi+l and yj
&lt;
^
&lt;
yj+r
2223,
it
making these extensions being similar to that of not repeated here, but the student is advised to
The method
meaning
of
review those
articles.
124. Computation of a double integral. ing a double integral is evident from the
comput
of the notation.
n2
As
this integral
is
xydxdy.
nZt
/&gt;2
\
written,
it is
equivalent to
I
I
I
Jo
Vo
xydy)dx, the integral
.
/
in parenthesis being
computed
first,
on the hypothesis that y alone
varies.
Ex.
2.
Find the value of the integral
/
J
xydxdy over the
first
quadrant of
the circle x2
If
+
y2
=
a2
.
we sum up
first
with respect to
(fig.
y,
we
is
to strips of the type
ABCD
75),
and the
find a series of terms corresponding limits of y are the ordinates of the
points like
A
and B.
The ordinate
of
A
evidently 0, and from the equation
of the circle the ordinate of
B is Va2
x2 where
,
OA = x.
Finally, to cover the
x are
and
a.
quadrant of the circle the limits of Hence the required integral is
2
B
n"^
*2
xydxdy=
^ j
r a P x ?/ T
2
i[Yl
2
&gt;&lt;
dx
te
=fl"*
_
A D
FIG. 75
1
rq
x"
x 4
~2l
2
Since the above computation of a double integral is simply the repeated computation of a single definite integral, the theorems of
24 may be used
in simplifying the work.
DOUBLE INTEGRAL IN POLAR COORDINATES
125. Double integral in polar coordinates. the double integral of f(r, 6) over any area,
If
229
to find
we have
we
divide that area
radii vectors
up
into elements, such as
ABCD
of
(fig.
76),
by drawing
at distances
A0
apart,
and concentric
circles the radii of
which
increase
is
by
AT*.
The area
ABCD
if
the difference of the areas of the
sectors
OB C
and OAD.
Hence,
2
OA = r,
area
AB CD =
\ (r
+ Ar) A0 2
J
r
A0
Therefore any term of the sum cor 120, is, responding to (1), at first sight, of the form
f(r, #)
(rArA0+J Ar
2

A0).
FIG. 76
in taking the limit, r Ar A0 rArA0 ( 3), for
But
+
JAr
A0 may
be replaced by
_.
Lim /(r, 0).(rArA0
+ lAr
is
2
A0)_ Lim / =
;
Hence the required
integral
Lim
A0Zo
2}
V /(r,
0) r
ArA0 =
/
/
f(r, 0) rdrdd.
(1)
fp&lt;
If the summation in (1) is made first with respect to r, the result a series of terms corresponding to strips such as A 1 1 C1 V and the limits of r are functions of found from the equation of the
is
S
D
bounding curve.
all
The summation with respect
i.e.
to
will then add
these terms, and the limits of
taken so as to cover the entire
area will be constants,
the least and the greatest values of
is
on
the bounding curve. If, on the other hand, the summation
to 0, the result
is
A 2 B2 C2 D2
,
made first with respect a series of terms corresponding to strips such as and the limits of are functions of r found from the
equation of the bounding curve. The summation with respect to r will then add all these terms, and the limits of r will be the
least
and the greatest values
of r
on the bounding curve.
230
MULTIPLE INTEGKALS
Ex. Find the integral of r2 over the circle r = 2 a cos 0. and 2 a cos 0, found If we sum up first with respect to r, the limits are from the equation of the bounding curve, and the result is a series of terms cor
responding to sectors of the type A OB To sum up these terms so as (fig. 77).
to cover the circle, the limits of
are
and !
The
result
is
a cos
J
rsdedr
TT
=
2 a cos
de
JO
=11 4a*cos
FIG. 77
4
0d0
The graphical representation may be made by the use
drical coordinates defined in
of cylin
127.
(fig.
126. Triple integrals.
respectively to
Let any volume
78) be divided into
rectangular parallelepipeds of
lel
volume A,xAyA2 by planes paral Z the coordinate
planes,
some
of the parallelepipeds
extending outside the volume in a
manner
similar to that in
which
122 extend out the rectangles in side the area. Let yjt zk) be a
(x.&gt;
point of intersection of
any three
of
these planes and form the i=n j=m k = p
_[&gt;,
sum
.\
A.A.Ai
W
"\
X
i=0 j=Q k=0
as
in
122.
Then the
limit of this
sum
as n,
m, and
p
increase
in
== 0,
definitely,
while
0,
Ax
Ay =
f(x
t
0,
Az =
all
FIG 7g
is
so as to
called the triple integral of
include
y, z)
points of the volume, throughout the volume.
It is denoted
by the symbol
f(x, y, z}dxdydz,
CYLINDRICAL AND POLAR COORDINATES
the limits remaining to be substituted.
231
If the summation is made x and y remaining constant, the result is to first with respect to z, extend the integration throughout a column of cross section A^Ay if next x remains constant and y varies, the integration is extended so as to combine the columns into slices and finally, as x varies,
;
;
the slices are combined so as to complete the integration throughout the volume.
127. Cylindrical and polar coordinates. In addition to the 84, we shall consider two other rectangular coordinates defined in
systems of coordinates for space of three dimensions,
drical coordinates, (2) polar coordinates.
(1) cylin
Cylindrical coordinates. If the x and the y of the rectangular coordinates are
1.
replaced by polar coordinates r and 6 in the plane XOY, and the z coordinate is re tained with its original significance, the new
coordinates
coordinates.
r, 6,
and z are called cylindrical The formulas connecting the
two systems of coordinates are evidently x = r cos 0, y = r sin 6, z = z.
Turning to
to the plane
fig.
M
FIG. 79
79,
we
that
see that z
=z
1
determines a plane parallel
XOY,
=
l
determines a plane
MONP,
passing
through
and making an angle l with the plane XOZ, and that r = r l determines a right circular cylinder with radius r x and OZ as its axis. These three surfaces inter
sect at the point P.
2.
OZ
Polar coordinates.
In
are
fig.
80 the cylin
.
x
drical coordinates of
P
If
OM=r,
of
MP =
z,
and
Z L OM =
9.
instead
r,
as the polar sin
OM = r we place OP = angle NOP by $, we shall have coordinates of P. Then, since ON= OP cos
r,
&lt;/&gt;,
placing and denote the
(/&gt;,
and and
(f&gt;
OP the following equations evidently express the con nection between the rectangular and the polar coordinates of P:
z
OM=
= r cos
cf&gt;,
x
= r sin
&lt;f&gt;
cos 0,
y
= r sin
&lt;f&gt;
sin 0.
232
MULTIPLE INTEGEALS
The polar coordinates of a point also determine three surfaces which intersect at the point. For 6 = l determines a plane (fig. 81)
through OZ, making the angle = with the plane XOZ; l
&lt;/&gt;
^
determines a cone of revolution, the axis and the vertical angle of
which
2
are respectively OZ and and r r^ determines (/^
;
a sphere with
at
its
center
and radius r r
In both
tems
nates
of
sys coordi
varies
6
X
from
dinates
to
2
TT,
and in polar coor
varies
&lt;f&gt;
from
FIG. 81
to
TT.
The
r
is
coordinate
in both systems, but
off
usually positive be negative, in which case it will be laid on the backward extension of
may
the line determined by the other two coordinates, as in I, 177.
128. Elements of
cylindrical and
volume
in
in polar coordi
nates.
If it is desired to
express
the triple integral of
126 in
either cylindrical or polar coor dinates, it is necessary to know
the expression for the element of volume in those coordinates.
1.
The element
of
volume
(fig.
in
cylindrical
is
coordinates
82)
bounded by two cylinders of radii r and r+Ar, two planes corresponding and + A0. corresponding to
the volume
/
FIG. 82
to z
It
and z+Az, and two planes
is
accordingly, except
for
ELEMENTS OF VOLUME
infinitesimals of higher order, a cylinder with
233
altitude
is
&z and
base
rArA0(
125).
Hence the element
of
volume
dV=rdrd0dz.
2.
(1)
The element of volume in polar coordinates (fig. 83) is the volume bounded by two spheres of radii r and r + Ar, two conical
surfaces
correspond
ing to
and
&lt;f&gt;
&lt;/&gt;+Ac/&gt;,
and two planes cor and responding to
+ A0.
amid
its
The volume
pyr
is
of the spherical
0ABCD
equal to the area of
base
tiplied
ABCD mul by one third
altitude r.*
of
its
To
find the area of
ABCD
we
note
first
that the area of the
zone formed by com pleting the arcs
AD
&lt;
FIG. 83
and
BC
is
equal to
r cos
is to
its altitude,
area of
ABCD
vol
r cos + multiplied by 2 irr. Also the the area of the zone as the angle A0 is to 2 TT.
(&lt;
A&lt;),
Hence
and
area
ABCD = r A0 [r cos $ r cos 0A BCD = r A0 [cos $  cos
3 3
((/&gt;
+ A$)],
Similarly,
vol
0EFGH = 1 (r + Ar) A0 [cos  cos
Therefore
vol ^4 J?
CDEFGH = 1 [(r + Ar)  r
3
3
]
A&lt;9
[cos
 cos
&lt;/&gt;
But mal
this expression differs
of higher order.
2 from r sin
$ArA&lt;/&gt;A0
by an
infinitesi
is
Accordingly, the element of volume
(2)
*
its
The volume
is to
base
of a spherical pyramid is to the volume of the sphere as the area of the area of the surface of the sphere.
234
It
is
MULTIPLE INTEGRALS
to
be noted that
dV
is
equal to the product of the
are respectively
rd(f&gt; )
three dimensions
rsm(f&gt;d0,
AB, AD, and AE, which
and
dr.
129.
in the
Change
form
I
of coordinates.
When
a double integral
is
given
/
f(x, y) dxdy, where the limits are to be substituted
it
so as to cover a given area,
of the integral
if
may
be easier to determine the value
coordinates.
of r
the rectangular coordinates are replaced by polar Then f(x, y) becomes f(r cos 6, r sin 6), i.e. a function
the other factor, dxdy, indicates the element of area, of 121 and the work of
and
0.
As
in
view of the graphical representation
125,
we may
replace
are not equivalent, but the
lent,
dxdy by rdrd0. These two elements of area two integrals are nevertheless equiva
provided the limits of integration in each system of coordi nates are taken so as to cover the same area.
In like manner, the three
f(x,y, z)dxdydz,
triple integrals
f(r cos f(r sin
are equivalent
total
0,
r sin 0,
z)
rdrdddz,
2
(f&gt;
cos 0, r sin
&lt;p
sin 0, r cos $) r sin fydrdfydO
when
the limits are so taken in each as to cover the
volume
to be considered.
PROBLEMS
Find the values of the following integrals
/&gt;;:
:
,X2 /io;z
i.
r r
Ji Ji
7X dxdy. y
5.
Jo Jo
"
re
nn
/rt~vsin
*
L*r$ dXdy
X
+
6
f
**UOr.

a
r
j
n
3.
C
Jit
2
Joy
/~
C^sindydx.
7.
C f^ J
i/
"rdrdd.
i
cos1
4.
f T ^xlog
*/2 t/i
y
dydx.
8.
f C
t/
e
*C
/
y
*
e**y+ s dxdydz.
o
/
o
o
PROBLEMS
r a /Vazx2
dxdydz
Jo Jo
Va2
x2
y2
_
235
z2
11.
JJood()
C
2
drdZ
.
13.
r
2
a
f
/asin&lt;J&gt;
r sin
cos
cos
6d6d&lt;j&gt;dr.
Jo J
15. Prove that
f ff(*)
f(y)dxdy
=
CHAPTER
XIII
APPLICATIONS OF MULTIPLE INTEGRALS
130.
Moment
of inertia of
a plane area.
product of
of a particle about an axis of
its
is the
its
The moment of inertia mass and the square
inertia of a
of the
distance
from
the axis.
of particles about the
The moment of same axis is the sum
number moments of
inertia
of
we
derive
the particles about that axis. From this definition (120) a definition of the moment of inertia of a
homogeneous rectangular lamina of thickness k and density p about an axis perpendicular to the plane of the lamina. The
result
may
be written in the form
M =pk
where
If
(*
i/a
rd
I
(&+f)dxdy
of inertia.
t
(1)
c/ c
M represents the moment
Jc
p and
are both replaced
by
unity, (1)
becomes
M:=
Uf \J\(&+f)dxdy, a c
(2)
which, as was noted in 120, is called the moment of inertia of the rectangle about an axis perpendicular to its plane at O. Reasoning in the same way, we may form the general expression
(3)
where the integration
is
XOY. Then
(3) is
the
moment
to extend over a given area in the plane of inertia of that area about the
axis perpendicular to the plane at 0.
Ex. Find the moment of
the origin, of the plane area y = 2 a, and the axis OF.
inertia,
(fig.
84)
about an axis perpendicular to the plane at bounded by the parabola y 2 = 4 ax, the line
236
MOMENT OF INEKTIA OF A PLANE AREA
If the integration is
237
made
first
with respect to
is
x, the limits of that integra
tion are
and
due
y2
4a
,
since the operation
the
summing
of elements of
moment
of inertia
to the
strip corresponding to
is
found from the axis of y, and the limit found from the equation of the parabola. Finally, the limits of y must be taken so as to include all the strips parallel to OX, and hence
is
elementary rectangles in any a fixed value of y the limit w2
;
y
must be
and 2
a.
Therefore
M = Jo
C"
Jo
f
4a
(z
2
+
y*)dydx
192
FIG. 84
131.
If
the plane area
is
of polar coordinates, (3) of
more conveniently denned by means 130 becomes
(1)
for,
the
129, in place of dxdy as the element of area we take by z 2 element of area rdrdO, and the factor x + y evidently 2 becomes r
.
Formula
(1)
may
also be de
rived directly from the fundamen tal definition at the beginning of
130, and the student
to
is
advised
make
that derivation.
Ex. Find the moment of inertia, about an axis perpendicular to the plane at 0, of the plane area (fig. 85) bounded by one loop of the curve r = a sin 2 6.
M
FIG. 85
We
the
shall take the loop in
first
quadrant, since the
of inertia of all the
moments
loops about the chosen axis are the same by the symmetry of the curve. If the first integration is made with respect to r, the result is the moment of
inertia of a strip
hence the limits for r are
bounded by two successive radii vectors and the curve and and a sin 2 6. Since the values of 6 for the loop of
;
238
APPLICATIONS OF MULTIPLE INTEGRALS
to
the curve vary from

,
it is
evident that those values are the limits for 6
/*asin20
in the final integration.
Therefore
vo M = Jo Jo
I
\
In the two preceding articles we have found the moment of inertia of a plane area about an axis perpendicular to the plane, which, with the exception of a constant factor, is the moment
132.
of
inertia of
a corresponding
shall
homogeneous lamina about the
same
the
axis.
We
now
find
of
moment
of
inertia
a
homogeneous lamina about an
axis in its plane.
Let the lamina be bounded
\
by the closed curve (fig. 86), and let its density at any point be p and its thickness be k.
Let OX be the axis about which
~x
FIG. 86
the
moment
is to
be taken.
Divide the area into rec
tangles of area A^A^/. Then the mass of any corresponding element of the lamina, as PQ, is pk&x&y. If this mass is regarded as concentrated at P, its moment about is pky^&x&y and if the mass is regarded as concen 2 trated at Q, its moment about is pk (y + Ay) Aa;Ay.
OX
;
Therefore,
if
M
OX
x
represents the required moment,
(i)
the summations to cover the entire area.
Since
Um P k (y+y?y = l
(3), and
the double
sums
of
have
the same limit
accordingly
t
&*=ffphfdax!y
where the integration
is
(2)
to cover the entire area.
AREA BOUNDED BY A PLANE CUKVE
If
239 moment
manner
(3)
of inertia of the plane area
If
p and k are each replaced by unity, (2) defines the about OX.
of inertia
M denotes the moment
about
Y, in similar
Ex. Find the moment of inertia about
parabola y
2
OY of
the plane area bounded by the
4 ax, the line y
is
=
2 a, and the axis
OY.
Since the above area
integration will be the 2 2 changed in that x + y
the
same
is
130, the limits of as there determined, but the integrand will be
.
same as that
of the Ex. in
2 replaced by x
_
Hence
M
y
= 
C"
C
Jo
4a
Jo
1
x 2 dydx
"
__
I
r2
/
192a3Jo
y*(ly *
If it is desirable to
use polar coordinates, (2) becomes
M=
x
ph* sin
s
2
OdrdB,
(4)
and
(3)
becomes
M = CCpkr
y
cos
2
6drd0
to
9
(5)
the substitution being
133.
made according
129.
Area bounded by a plane curve.
The area bounded by an
arc of a plane curve, the axis of x t and the ordinates of the ends of the arc has been determined in 35 by a single integration. By tak
ing the algebraic sum of such areas any plane area may be computed. The area bounded by any plane curve may also be found by a
double integration as follows
:
Draw
straight lines parallel to
OX
and
of
to
OF respectively,
fig.
forming
rectangles of area A:rAy,
which, as in
74, will be entirely
some of within the curve, and others
which
will be only partly within the curve.
of these rectangles,
Form
the double
sum
^^A^Ay
and then
let their
increase indefinitely while
integral
A# =
and
Ay =
0.
number Then the double
(I)
rr
JJ
is
dxdy
the required area.
240
APPLICATIONS OF MULTIPLE INTEGRALS
2 = 4 Ex. Find the area inclosed by the curve (y x x 2 (fig. 87). 3) The element of area is the rectangle AxAy. If the first integration is made with respect to y, the result is the area of a strip like the one shaded in fig. 87, and the limits for y will be found by solving the equation of the curve for y in
v
terms of
x.
Since y
=
x
2
+
3
the lower limit
V4x
the upperjimit
is
?/ 2
and
=x+3
ViT^ x For the integra tion with respect to x the limits are 2 and 2, since
2.
+
the curve
lines
is
bounded by the
2 and x x = Therefore
=
2.
area
=
\
C *dxdy
J&lt;zJ

2/i)
dx
= 2 f V4 x2 dx
J2
/
This example
35.
is
Ex.
3,
Comparing the two
solutions, we see that the result of the first integra
tion here
"*
is
grand in
35.
exactly the inteIt is evident
FIG. 87
that this will be the case
in all similar problems, and hence many areas may be found by single integration. The advantage of the double integral consists in the representation of the area of a figure for which
the limits of integration cannot easily be found.
134. In like manner, the area
bounded by any curve
in polar
coordinates
may
be expressed by the double integral
rdrd6,
(1)
the element of area being that bounded by two radii vectors the angles of which differ by A0, and by the arcs of two circles the
radii of
If
which
first
differ
by Ar.
integration of (1) is with respect to r, the result 2 before the substitution of the limits is ^ r d@. But this is exactly
the
AREA OF ANY SURFACE
241
the integrand used in computation by a single integration. Hence many areas may be computed by single integration in. polar
coordinates.
135. Area of
any
surface.
Let
C
(fig.
88) be any closed curve
on the surface /(#, y, z) = 0. Let its projection on the plane XOY We shall assume that the given surface is such that the be C
.
at any point within the curve C perpendicular to the plane meets the surface in but a single point. In the plane and QF, draw straight lines parallel to
1
XOY
XOY
OX
forming rectangles of area A#Ay, which
in
lie
wholly or partly
lines
the area bounded by
to
C
.
Through these
z
in
pass planes
parallel
OZ.
These planes
the
surface
will
intersect
curves which intersect in points the projections of which on the
are the vertices of plane the rectangles for example, is the projection of P. At every
;
XOY
M
such point as
P draw
the tan
gent plane to the surface,
each tangent plane there
will be cut a parallelo
gram
*
by the
planes
drawn
OZ.
F
FlG 88

parallel to
We shall now
define the area
of the surface /(#, y, z) 0, bounded by the curve C, as the limit of the sum of the areas of these parallelograms cut from the tan
=
gent planes, as their number is the same time that A# = and
made
Ay = 0.
to increase indefinitely, at This definition involves
the assumption that the limit is independent of the manner in which the tangent planes are drawn, or of the way in which the small areas are made to approach zero. This assumption may be
proved by careful but somewhat intricate reasoning.
* This parallelogram is not
drawn
in the figure, since
it
coincides so nearly with the
surface element.
242
If
APPLICATIONS OF MULTIPLE INTEGRALS
A^4 denotes the area of one of these parallelograms in a tan gent plane, and 7 denotes the angle which the normal to the tangent plane makes with OZ, then ( 92)
Aa?Ay
since the projection of
= A J. cos 7,
/I)
A^
on the plane
are,
XO Y is A#Ay.
(2),
The
direc
tion cosines of the
dz
1
&gt;
normal
by
1
112,
proportional to
1
v,
;
hence
cos
7
=
and hence
and
According to the definition, to find = and Ay = that is
;
A we
must take the
limit
of (3) as Aa;
where the integration must
be extended over the area in
the plane XOFbounded
by the curve
C
1
.
Ex. 1. Find the area of an octant of a sphere of
radius
a.
If the center of
the sphere is taken as the origin of co
ordinates
(fig.
89),
the equation of the
sphere
is
= a2
Y
area on the plane
,
(1)
XOY
is
the area in the
2
first
and the projection of the required quadrant bounded by the circle
(2)
and the axes
OX and OF.
+
2
2/
=
2
AREA OF ANY SURFACE
From
(1),
243
by(i)
=
3.
na
iTra
7ra2.

dx
/o
Ex.
2.
The center
cular cylinder of radius by the sphere.
of a sphere of radius 2 a is on the surface of a right cir Find the area of the part of the cylinder intercepted a.
Z
FIG. 90
Let the equation of the sphere be
the center being at the origin
2
2/
(fig.
90),
and
let the
0,
equation of the cylinder be
(2)
+ z22a?/ =
the elements of the cylinder being parallel to OX. To find the projection of the required area on the plane it is necessary to find the equation of a cylinder passing through the line of intersection of (1)
XOY
and
(2),
and having
its
elements parallel to OZ.
2
2/
+
+
z2

4 a2 )
+ Ml/ 2 +
z2
Now  2ay) =
(3)
244
APPLICATIONS OF MULTIPLE INTEGRALS
represents any surface through the line of intersection of (1) and (2), and hence Accord it only remains to choose k\ and k% so that (3) shall be independent of z.
ingly let ki
=
1
and k 2
=
1,
and
x2
(3)
becomes
+
dx
2 ay

4 a2
=
0.
(4)
From
(2),
_
A
_
cy
by
(2)

A
f*2a
p
=L /
=
8 a2
.
adydx
limits of integration were determined from (4), which is the projection bounding line of the required area on the plane XOY. As an equal area is intercepted on the negative side of the plane JTOF, the above result must be multiplied by 2. Hence the required area is 16 a 2
The
of the
.
The evaluation
of (4)
may sometimes
be simplified by trans
forming to polar coordinates in the plane
XOY.
Ex. 3. Find the area of the sphere x2 + y2 + z 2 = a 2 included in a cylinder having its elements parallel to OZ and one loop of the curve r = a cos 2 6 (fig. 91) in the plane as its directrix.
XOY
Proceeding as in Ex.
1,
we
find the integrand
.
Va2  x2  y2
(
Transform
ing this integrand to polar coordinates,
we have
129)
~ a cos 20
A=
ardddr
for the first integration with respect to r covers a
sector extending
from the
origin to a point on the curve a cos 2 0, since the cjurve r
=
passes through and the final
the origin
6
is
;
integration
with respect
to
from
FIG. 91
radii vectors
to
4
is
,
since
the loop
4 chosen
bounded by the
is
=
4
is
and
=4
The
factor 2 before the integral
necessary
because there
an equal amount
of area on the negative side of the plane
XOY.
CENTER OF GRAVITY
Therefore
245
A=2 f Vj
20
If the required area is projected
on the plane YOZ, we have
where the integration extends over the projection of the area on YOZ] and if the required area is projected on the plane XOZ, we have
the plane
where the integration extends over the projection
the plane
of the area
on
XOZ.
In
all of
136. Center of gravity.
system of particles ing formulas being
of a
47 we denned the center of gravity which lie in the same plane, the result
If
the particles do not
all lie in
the same plane,
^\
we
are obliged to
add a third equation,
z
m 2i = ~^ &
2^ra.
to define the third coordinate of the center of gravity, the deriva tion of which is not essentially different from that given in 47.
To determine the center of gravity of a physical body, we divide the body into elementary portions, the mass of any one of which
be represented by Am. Then if (#., y., z.) that the mass of one of the elementary portions
may
is
any point such
be considered
may
as concentrated at that point, we define the center of gravity (x, y, z) of the body by the formulas
246
APPLICATIONS OF MULTIPLE INTEGRALS
The denominator of each of the preceding fractions is evidently M, the mass of the body. Formulas (1) can be expressed in terms of definite integrals, the shall here evaluation of which gives the values of x, y, and z.
We
take up only those cases in which the definite integrals introduced are double or triple integrals.
137. Center of gravity of a plane area.
of a plane area in the plane
The center
of gravity
From we have immediately that z = 0. To determine x and y we divide the area into rectangles of area A^Ay (fig. 86), and if we denote the density by p, Am = /aAzAy. If we consider the mass of an element, as PQ, concentrated at P, we have, by substituting in (1), 136,
been defined in
49.
XOY has
that definition
(1)
an expression which is evidently less than x and as concentrated at Q, we have the mass of
;
if
we
consider
PQ
an expression which
is
But the
Limits of (1)
...
.
evidently greater than x. and (2) are the same ( 3), for
4 Aa?

(x. Lim ^^
The
limit of (1) is
^I
\pxdxdy

,
both integrals being taken over the
entire area.
iipdxdy
\\pxdxdy
Therefore
x
=
~^~
~&gt;
^
=
\\pydxdy
"
(3)
"77^
tjpdxdy
y being derived in the same manner as
jjpdxdy
x.
CENTER OF GRAVITY OF A PLANE AREA
If
247
p
is
constant,
it
can be canceled
;
and in any problem in
which p
is not denned, it will be understood that it is constant. In that case the denominator of each coordinate is the area of the
plane figure.
Ex. Find the center of gravity of the segment of the
off
x2
ellipse

?/
1
2
=
1
cut
a2
62
n
I
f\
I
by the chord through the positive ends of the axes of the curve. is Ex. 2, 49, and the student should compare the two solutions. The equation of the chord is bx f ay = ab. To determine x and y we have to compute the two integrals CCxdxdy and JJ ydxdy over the shaded area of fig. 39, and also find that shaded area.
This
The area is the area of a quadrant of the ellipse less the area of the triangle formed by the coordinate axes and the chord, and is accordingly
For the integrals the
and y%
=  v a2
_
b
(7ra6) la&
=
a6(7r2). y are y
limits of integration with respect to
=
and
x2 y\ being found from the equation of the chord, and y 2 being
,
found from the equation of the
ellipse.
The
limits for
x are evidently
a.
Jo Jahbx
r/6 f r l^xtofr = Jo \a z va^  te +
_ 32
n
~\/2
 bx
ydxdy
=
1
/
(

6 2x 2
a 2 Jo
+
a62 x) dx
x
=
3
2a
(TT
26
2)

3
(TT

2)
138.
If the
equation of the bounding curve of the area
is
in
polar coordinates,
we
have, by transforming equations
(3),
137,
by
129,
rr
x
=
prcos0drd0
I
CprdrdO
(i)
2
pr
sin
OdrdO
prdrdO
248
APPLICATIONS OF MULTIPLE INTEGRALS
circles
(b
Ex. Find the center of gravity of the area bounded by the two
r
It is evident
=
a cos
0,
r =
b cos
6.
&gt;
a)
As
from the symmetry of the area (fig. 92) that y = 0. is the area, the denominator of the fractions, after canceling
/&gt;,
it
is
equal to
444
r2 cos edddr
=  TT (W 
a*).
The numerator
for x
becomes
1
3
(ft
=

a3 )
r2

cos*0d0
_ FIG. 92
_
62
+
ab
+
a)
a2
~2 (6 +
139. Center of gravity of a solid.
of
To
find the center of gravity
any
solid
we have merely
to express the
Am
of formulas (1),
136, in terms of space coordinates
and proceed as in
137.
For
example,
if
rectangular coordinates are used,
Am = /aA^AyAz,
and
I
I
\
pxdxdydz
I
\
\pdxdydz
I
I
\pydxdydz
JJJ_ y=
\\ \pdxdydz
\\\pzdxdydz
z
=
I
/
\
pdxdydz
137. the work of derivation being like that of If desired, formulas (1) may be expressed in cylindrical or polar
coordinates.
Ex. Find the center of gravity of a body of uniform density, bounded by one nappe of a right circular cone of vertical angle 2 a and a sphere of radius a, the
center of the sphere being at the vertex of the cone.
VOLUME
249
If the center of the sphere is taken as the origin of coordinates and the axis of the cone as the axis of z, it is evident from the symmetry of the solid that 0. To find z, we shall use polar coordinates, the equations of the sphere x = y
=
and the cone being respectively
r
a and
&lt;
=
.
a.
C
C
nllf
I
C
{
r cos
nO.
I
r2
sin
(j&gt;d6d&lt;f&gt;dr
Then
Jo
Jo Jo
n(X
r 2 sin
Jo
Jo Jo
({&gt;d6d(f&gt;dr
The denominator
of altitude a(l
is
the volume of a spherical cone the base of which
its
is
a zone
cos a); therefore
r3 cos
&lt;
volume equals
3
7ra (l

cosa)
(
128).
C
Jo
Jo Jo
f f
sin
&lt;pddd(f&gt;dr
=
1 a4 f
sin
Jo
4
Jo
(""cos
QdBdQ
dt
..
= a (l a) C Jo = l7ra4 (l cos2 a). z = ! (1 + cos a) a.
cos 2

140. Volume.
In
126, 128
of
we found
expressions for the
element of volume in rectangular, in cylindrical, and in polar
coordinates.
The volume
a solid
of these
bounded by any surfaces
elements as their number
will be the limit of the
sum
increases indefinitely while their magnitudes approach the limit zero. It will accordingly be expressed as a triple integral.
Ex.
1.
Find the volume bounded by the
(fig.
Jg2
y2
2
ellipsoid
From symmetry
is
a2
h
62
H
=
1.
c2
93)
it
volume
evident that the required is eight times the vol
in the first octant
ume
bounded by the surface and the coordinate
planes.
In summing up the
rectangular
parallel
epipeds AxAyAz to form a prism with
edges parallel to OZ, the limits for z are
the latter being found from the
equation of the
ellipsoid,
Y/
FIG. 93
250
APPLICATIONS OF MULTIPLE INTEGRALS
respect to y, to obtain the volume of a slice,
Summing up next with
as the lower limit of y,
we have
is
and 61/1


as the upper limit.
This latter limit
determined by solving the equation
;
+
^=
and
1,
found by letting z
=
in the
that the ellipsoid has the equation of the ellipsoid for it is in the plane z greatest extension in the direction OF, corresponding to any value of x.
Finally, the limits for x are evidently
a.
=
Therefore
Fsgf J
= 8C
"*
o
f* Jo
f Jo
T
a2
b
*dxdydz
pa, /&gt;6\/l^
/o/o
a
"V
a*
C
Jo
(l\ dx a2 \
/
=
It is to
 rrabc.
first
be noted that the
integration,
when
rectangular
coordinates are used, leads to an integral of the form
where
faces.
z2
and
z l are
It follows that
found from the equations of the bounding sur many volumes may be found as easily by
is
double as by triple integration. In particular, if z l = 0, the volume
senting the double integral
Ex.
z
2.
(
the one graphically repre
121).
z
Find the volume bounded by the surface
it
=
ae~( xZ + y) and the plane
=
0.
nates.
To determine this volume Then the equation of
is
(
will be
the surface becomes z
advantageous to use cylindrical coordi = ae~ 2 and the element of
,
volume
ae~ r2
128) rdrdddz.
first with respect to z, we have as the limits of integration and and oo, for in the integrate next with respect to r, the limits are plane z = 0, r = GO, and as z increases the value of r decreases toward zero as a limit. For the final integration with respect to 6 the limits are and 2 TT.
Integrating
.
If
we
Therefore
V= Jo
I
stZir
/oo
I
s*ae~ rZ
I
rdedrdz
Jo Jo
/.2rr
/&gt;oo
=
a
I
Jo
Jo
de
I
re~ r *dddr
MOMENT OF INERTIA OF A SOLID
251
In the same way that the computation of the volume in Ex. 2 has been simplified by the use of cylindrical coordinates, the com putation of a volume may be simplified by a change to polar coor dinates and the student should always keep in mind the possible
;
advantage of such a change.
141.
Moment
of inertia of a solid.
The moment
:
of inertia of a
an axis may be found as follows Divide the solid into elements of volume, and let Am represent the mass of such an element. Let h and h + AA represent the least and the greatest
solid about
distances of any particle of from the axis. Then if is as concentrated at the least distance, its moment of inertia regarded
Am
is
Am
would be
If the
2
7t
Am
;
and
its
if
Am
greatest distance,
moment
2
of
regarded as concentrated at the 2 inertia would be (h fA/t) Am.
moment
of inertia of the entire solid is
denoted by M,
^ A Am
the solid was divided.
.
&lt;
M&lt;
^(h + AA) Am,
2
where the two sums include
all
the elements of volume into which
Lim J
/i

Am
=
1
when
the
number
their
of
the elements of
volume increases
the limit zero.
indefinitely while
magnitude approaches
Hence we
define
M by the equation
(1)
It is to
be noted that the cases of
of
130132
are but special
cases of (1).
The computation
integral in terms of
M requires
us to express (1) as a definite
of coordinates, the choice of a
some system
particular system of coordinates depending
Ex. Find the
a diameter.
upon the
solid.
moment
of inertia of a
homogeneous sphere of density p about
We shall take the center of the sphere as the origin of coordinates, and the diameter about which the moment is to be taken as the axis of z. The problem will then be most easily solved by using cylindrical coordinates.
The equation
is
of the sphere will be r2
+
z2
=
a2 and
,
dm =
the density
;
also h
=
prdrdedz, where p
r,
so that
we have
to find the value of the triple integral
252
APPLICATIONS OF MULTIPLE INTEGRALS
Va Va
.
Integrating first with respect to z, we find the limits, from the equation of 2 2 the sphere, to be r2 and r2 Integrating next with respect to r, we have the limits and a, thereby finding the moment of a sector of the sphere. To include all the sectors, we have to take and 2n as the limits of 6 in the last
integration.
Therefore
M = p Jo C
27r
a
f
J
C
"
2
~ rZ
r*dddrdz.
r2
J  Va2 
As a
result of the first integration,
M = 2 p Jo C
stitution,
f
Jo
V Va
dd
2

f2 dedr.
Making the next we have
integration
by a reduction formula or a trigonometric sub
2n
M=
^Ar
pa5
I
=
,% Trpa
5
.
142. Attraction.
was
defined,
In 45 the attraction between two particles and the component in the direction OX of the attrac
particle
tion of
any body on a
was derived
as
LirnV
n=oc^ T*
i
^Am,
where
represents an element of mass of the body, r may be considered the shortest distance from any point of the element to the particle, and 6 is the angle between
i
Am
OX
and the
line r {
.
This expression
is
entirely
general, and similar expressions may be de rived for the components of the attraction in the directions
OY and
OZ.
use double and triple the application of these formulas integrals
that
is simplified.
Now
we can
Ex. Find the attraction due
to a
homogeneous
circular cylinder of density p, of height A, and radius of cross section a, on a particle in the line of the axis
^
of the cylinder at a distance 6 units one end of the cylinder.
from
Take the
dinates
FIG. 94
(fig.
94),
particle at the origin of coor and the axis of the cylin
der as OZ.
Using cylindrical coordinates,
f
we have dm = prdrdddz and n = Vz2
r2
.
From
the
symmetry
of the figure the resultant
components
of attraction in
the directions
OX
and
OY
are zero,
and cos 0;
=
for the resultant
component
in the direction
OZ.
PROBLEMS
Therefore, letting
253
we have
Az
g
represent the component in the direction OZ,
=
ft
P
.
Jo o
o
Jb
2
(z
+
r2 )*
fig.
dedrdz,
94.
where the
limits of integration are evident
from
.
A, = ,"
a2
2

= =
Zlt
P
f Jo
7jv&gt;
(h
+ V62 +
2
 V(6 +
2
/i)
+
a
a2) e*0
2
2
+ V& +
a
 V(6 +
A)
2
+
).
PROBLEMS
x
=
+
Find the moment of inertia of the area between the straight lines x + y = 1, and y = 1 about an axis perpendicular to its plane at 0. 2. Find the moment of inertia of the area bounded by the parabolas y 2 = 4 ax 4 6x + 4 62 about an axis perpendicular to its plane at 0. 4 a2 y2 =
1.
1,
,
3.
x2 (a
4.
Find the moment of inertia of the area of the loop of the curve x) about an axis perpendicular to its plane at 0.
Determine the moment of inertia about an axis perpendicular
62 ?/ 2
=
to the plane
at the pole of the area included between the straight line r a sec 6 and two straight lines at right angles to each other passing through the pole, one of these lines making an angle of 60 with the initial line.
5. Find the moment of inertia of the area of one loop of the curve r = a cos 3 6 about an axis perpendicular to its plane at the pole. 6. Find the moment of inertia of the area of the cardioid r = a(cos0 f 1)
=
about an axis perpendicular to its plane at the pole. 7. Find the moment of inertia of the area of one loop of the lemniscate r2 = 2 a 2 cos 2 about an axis perpendicular to its plane at the pole. 8. Find the moment of inertia of the total area bounded by the curve about an axis perpendicular to its plane at the pole. r2 = a 2 sin
9.
r2
= a2 sin 3 6
10.
Find the moment of inertia of the entire area bounded by the curve about an axis perpendicular to its plane at the pole.
Find the moment of inertia of the area of a circle of radius a about an any point on its circumference.
axis perpendicular to its plane at
11. Find the
moment of
inertia of the area of the circle r
=
a which
is
not in
cluded in the curve
12.
r=a sin 2 6 about an axis perpendicular to its plane at the pole.
the hyperbola xy
13.
Find the moment of inertia about the axis of y of the area bounded by = a 2 and the line 2 x + 2 y 5a = 0.
Find the moment of inertia about the axis of x of the area of the loop
of the curve 62 ?/ 2
14.
=
x 2 (a
x).
Find the moment of inertia of the area of one loop of the lemniscate about the straight line through the pole perpendicular to the r2 = 2 a2 cos 2 initial line as an axis.
254
APPLICATIONS OF MULTIPLE INTEGKALS
moment
line,
15. Find the above the initial
of inertia of the area of the cardioid r
initial line as
=
a(cos0
+
1)
about the
an
axis.
2
16. Find the moment of inertia of the area bounded by a semicircle of radius a and the corresponding diameter, about the tangent parallel to the diameter. 17.
V
5a
=
Find the area bounded by the hyperbola xy
0.
=
a2 and the
line
2x+2y
18.
2
8 a3
Find the area bounded by the parabola x 2
2
=
4 ay
and the witch
~x +4a
19.
r
=
2
Find the area bounded by the lima^on
r
=
2cos0
+
3 and the circle
2 cose.
y
=
20. Find the area bounded by the confocal parabolas y 2 4 bx + 4 &2
.
=
0.
4 ax
+
4 a2
,
21. Find the area bounded
by the
circles r
a cos
8,
r
=
+
a sin y
2
22. Find the areas of the three parts of the circle x 2 which it is divided by the parabola y 2 = ax.
e
2 ax
=
into
23. Find the_area cut off from the lemniscate r 2
line r cos
=
2
a 2 cos 2
by the
straight
=
2
a.
24. Find the area of the surface cut from the cylinder x 2
cylinder y
+
z2
=
+
y2
=
a 2 by the
a2
.
circular cylinder of radius center of the sphere.
25. Find the area of the surface of a sphere of radius a intercepted by a right a, if an element of the cylinder passes through the
26. Find the area of the surface of the cone x2
2 cylinder x
+
y2
z2
=
cut out by the
+
y2
2 ax
0.
27. Find the area of the surface of the cylinder x2 by the plane XOY and a right circular cone having along OZ, and its vertical angle equal to 90.
+
y2
2 ax
=
bounded
its
its
vertex at 0,
axis
+
y
28. Find the area of the surface of the right circular cylinder z 2 sin a) 2 = a2 included in the first octant.
+
(x cos
a
29. On the double ordinates of the circle x 2 + y 2 = a 2 as bases, and in planes perpendicular to the plane of the circle, isosceles triangles, each with vertical angle 2 a, are described. Find the equation of the convex surface thus formed,
and
x2
its total area.
30. Find the area of the surface z

= xy included
in the cylinder (x 2
+ y2 2 =
)
y
2.
31. Find the area of the sphere x2 + y 2 + z 2 = a 2 included in the cylinder with elements parallel to OZ, and having for its directrix in the plane a single loop of the curve r = a cos 3 6. x ~ y
XOY
32. Find the area of that part of the surface z
=
which on the plane
XOY
2
is
the projection of
(I
bounded by the curve
r
2
=
2
CQ0,
PROBLEMS
tion, of the sphere
255
33. Find the area of the sphere x 2 + y 2 f z 2 = 4 a2 bounded by the intersec and the right cylinder, the elements of which are parallel to
the directrix of which
is
OZ and
the cardioid r
a (cos0
+
1)
in the plane
XOY.
34. Find the center of gravity of the plane area bounded by the parabola #i + yl = at and the line x + y = a. 35. Find the center of gravity of the plane area bounded by the parabola
x2
= 4 ay
=
x3
and the witch y
=
2
36. Find the center of gravity of the plane area bounded
2
2/
by the
cissoid
2a
x
 and
its
asymptote.
curve a 4 ?/ 2
37. Find the center of gravity of the area of the part of the loop of the = a 2 x4 x6 which lies in the first quadrant.
first
38. Find the center of gravity of the area in the the curves x^
quadrant bounded by
+
y*
=
a 3 and x 2
+
2
?/
=
a2
.
39. Find the center of gravity of the plane area
the curve x^
bounded by OJT, OF, and
+
y%
=
a$.
40. plate is in the form of a sector of a circle of radius a, the angle of the sector being 2 a. If the thickness varies directly as the distance from the center, find its center of gravity.
A
41.
How
a loop of the curve r
far from the origin is the center of gravity of the area included in = a cos 2 ?
42. Find the center of gravity of the area bounded
by the cardioid
r
=
a(cos0
+
1).
sity in the
43. Find the center of gravity of a thin plate of uniform thickness and den form of a loop of the lemniscate r2 = 2 a 2 cos 2 6.
44. Find the center of gravity of a homogeneous body in the form of an
octant of the ellipsoid
a2
++= 2
b2
c
&gt;
1.
surfaces z
46.
45. Find the center of gravity of the homogeneous solid bounded = k&, z = A 2 x(fc 2 &i), x 2 + y 2 = 2 ax.
by the
varies
The density
of a solid
bounded by the
/p2
y2
h 62
Z2
ellipsoid
a2
H
=1
c2
directly as the distance from the plane FOZ. portion of this solid lying in the first octant.
Find the center of gravity of the
47. Find the center of gravity of the homogeneous solid bounded by the
surfaces z
=
0,
y
=
0,
y
=
6,
62 z 2
is
=
y (a
2
2

x 2 ).
48.
A
homogeneous
solid
bounded by a sphere of radius a and a right
which
is
the vertex of which is on the , 3 surface of the sphere, and the axis of which coincides with a diameter of the sphere. Find its center of gravity.
circular cone, the vertical angle of
sity of
49. Find the center of gravity of a right circular cone of altitude a, the den each circular slice of which varies as the square of its. distance from the
vertex,
256
APPLICATIONS OF MULTIPLE INTEGRALS
if
50. Find the center of gravity of an octant of a sphere of radius a, density varies as the distance from the center of the sphere.
the
51. Find the center of gravity of a homogeneous solid bounded by the sur faces of a right circular cone and a hemisphere of radius a, which have the same base and the same vertex.
52. Find the volume
bounded by the surface x*
+
y*
+
y
2
z*
=
2
a*
and the coor
dinate planes.
53. Find the volume of the part of the cylinder x 2 between the planes z k\x, z = k%x(ki kz).
&lt;
+
ax
=
included
54. Find the total volume bounded
by the surface
first
^+^+
=
)
= 1.
55. Find the volume included in the
octant of space between the coor
j
dinate planes and the surface y2
56. Find the volume in the
first
16
[
1
+
8z
[
1
0.
octant bounded by the surfaces z
x2
+
y
2
=
=(x +
y)
2
,
a2
.
5 7. Find the volume bounded by the surface b2 z 2
= y 2 (a2
=
z2
x 2 ) and the planes
y
=
and y
=
b.
58. Find the volume bounded
59. Find the
by the surfaces az
xy, x
+
y
+ z = a,
z
= 0.
volume bounded by the cylindroid
2.
f
=
x
+
y and the planes
x
=
0,
y
=
0, z
=
60. Find the volume of the paraboloid y 2
z2
=
8 x cut off
by the plane
y
= 2x
61
.
2.
62.
= ax 2 + by 2 y 2 = 2 ex x 2 z = 0. Find the volume bounded by the surfaces x 2 + y2 = ax, x2 + y 2 = bz, z = Q.
Find the volume bounded by the surfaces
z
, ,
63. Find the total volume bounded by the surface x^
64. Find the volume cut
+
y%
+
z*
=
a*.
from a sphere of radius a by a right circular cyl
inder of radius sphere.
,
one element of the cylinder passing through the center of the
65. Find the volume bounded by the surfaces z
66. Find the volume
bounded by the surfaces
z
= a(x f ?/), = 0, z = ar2
(x f
2
z r
=
=
a(x
2
+
6.
y2 ).
,
b cos
67. Find the total volume
bounded by the surface
y
2
+
z 2) 3
= 27 cfixyz.
(Change to polar coordinates.)
68. Find the volume bounded by a sphere of radius a and a right circular cone, the axis of the cone coinciding with a diameter of the sphere, the vertex being at one end of the diameter and the vertical angle of the cone being 60.
69. Find the total
70.
volume bounded by the surface
(x
2
+
y2
2 2 f z )
=
axyz.
Find the volume of the sphere x2 + y 2 + z 2 = a2 included in a cylinder with elements parallel to OZ, and having for its directrix in the plane one
XOY
loop of the curve r
=
a cos 3
9,
PROBLEMS
71. Find the
257
volume bounded by the plane JTOF, the cylinder x2 f y 2 2 ax and the right circular cone having its vertex at 0, its axis coincident with 0, OZ, and its vertical angle equal to 90.
=
72.
Find the
total
volume bounded by the surface r2
inertia of a
f
z2
=
of
73. Find the moment of density Pl about OX.
ar(cos0 f 1). x2 z2 y2
(\
homogeneous
ellipsoid
=
1,
74. Find the
moment
in the
its
of inertia about its axis of a
p,
homogeneous
right cir
cular cylinder of density
75.
height
A,
and radius
a.
A solid
is
form
angle 2 a.
ticle is
Find
moment
its
proportional to
of a right circular cone of altitude h and vertical of inertia about its axis, if the density of any par distance from the base of the cone.
76.
The density
diametral plane. to the above diametral plane.
77.
of a solid sphere of radius a varies as the distance from a Find its moment of inertia about the diameter perpendicular
shell, the
its
homogeneous solid of density p is in the form of a hemispherical Find inner and the outer radii of which are respectively TI and r 2 moment of inertia about any diameter of the base of the shell.
.
A
solid is bounded by the plane z = 0, the cone z = r (cylindrical coor and the cylinder having its elements parallel to OZ and its directrix dinates), one loop of the lemniscate r2 = 2 a2 cos 2 in the plane XOY. Find its moment of inertia about OZ, if the density varies as the distance from OZ.
78.
A
79.
altitude h,
Find the attraction of a homogeneous right circular cone of mass and vertical angle 2 a on a particle at its vertex.
Jf,
80.
is
A
bounded by a spherical surface of radius
portion of a right circular cylinder of radius a and uniform density p the center of which coin 6(6&gt;a),
cides with the center of the base of the cylinder. Find the attraction of this portion of the cylinder on a particle at the middle point of its base.
center of
82.
81. Find the attraction due to a hemisphere of radius a on a particle at the its base, if the density varies directly as the distance from the base.
of a hemisphere of radius a varies directly as the distance its attraction on a particle in the straight line perpendic ular to the base at its center, and at the distance a from the base in the direc
The density
from the base.
tion
Find
away from
the hemisphere.
homogeneous ring is bounded by the plane JTOF, a sphere of radius 2 a with center at 0, and a right circular cylinder of radius a, the axis of which coincides with OZ. Find the attraction of the ring on a particle at O.
83.
A
CHAPTEE XIV
LINE INTEGRALS AND EXACT DIFFERENTIALS
Let C (fig. 95) be any curve in the plane the two points L and K, and let and be two func connecting tions of x and y which are onevalued and continuous for all points on C. Let C be divided into n seg
143. Definition.
M
XOY
N
ments by the points P^P^P^..., P _ lt n and let Ax be the projection of one of
these segments on
jection on
Y.
it
OX
is,
and
That
x
FIG. 95
Ay = y i+1 y where the values of Ax and Ay are not necessarily the same
ior all values of
p
n
i
Ay its Ax = xi+l
pro
xif
o
T
M
v.
Let the value of
for each of the
n points
L,
P P, v


,
%_
l
corresponding value of Ax, and the value of by the corresponding value of Ay, and let the i=nl
N
be multiplied by the for the same point
sum be formed
The
limit of this
sum
as
n
is
increases without limit and
Ax and
Ay
approach zero as a limit
denoted by
(Mdx + Ndy),
/(C
is called a line integral along the curve C. The point may coincide with the point L, thus making C a closed curve. If x and y are expressed in terms of a single independent vari able from the equation of the curve, the line integral reduces to a
and
K
definite integral of the ordinary type;
but this reduction
is
is
not
always convenient or possible, and it properties of the line integral directly.
important to study the
shall give first a few examples, showing the importance of the line integral in some practical problems.
258
We
DEFINITION
259
Ex. 1. Work. Let us assume that at every point of the plane a force acts, which varies from point to point in magnitude and direction. We wish to find the work done on a particle moving from L to along the curve C. Let C be divided into segments, each of which is denoted by As and one of which is rep resented in fig. 96 by PQ. Let F be the force acting at P, PR the direction in which it acts, PT the tangent to C at P, and the angle
K
RPT. Then
tion
ticle
the component of
to
F in
the direc
PT is F cos 0, and the work done on
moving from P
Q is Fcos 6 As,
a par except for
infinitesimals of higher order.
in
The work done
is,
moving the
particle along
C
therefore,
W = LimVFcos0As = C
**
J C
Fcoseds.
O
FIG. 96
the angle between
sin a.
&lt;f&gt;
Now
let
a
be the angle between
PR
=
cos
and OX, and
cos
&lt;f&gt;
PT
and OX.
Then
=
&lt;j&gt;
a and
\
cos 6
cos
a
f sin
Therefore
W = Jcc\(F cos
&lt;j&gt;ds
a+
F sin
sin a) ds.
is"
But F cos a is the component of force parallel to OX and usually denoted by X. Also Fsin a is the component of force parallel to OY and is usually denoted = dx and sin = dy 42). Hence we have finally by Y. Moreover cos
&lt;pds
(
W=
Ex.
2.
C (Xdx + Ydy).
(C)
Flow of a liquid. Suppose a liquid flowing over a plane surface, the lines in which the particles flow being indicated by the curved arrows of fig. 97. We imagine the flow to take place in planes parallel to. JTOF, and shall assume
^/
C^^*^~

find the
the depth of the liquid to be unity. wish to amount of liquid per unit of time which
We
flows across a curve C.
FIG. 97
Let q be the velocity of the liquid, a the angle which the direction of its motion at each point makes with OX, u = q cos a the component of velocity parallel to OX, and v = q sin a the com ponent of velocity parallel to OF. Take an ele ment of the curve PQ = As. In the time dt the particles of liquid which are originally on PQ will flow to P Q/, where PP = QQ = qdt (except for
infinitesimals of higher order). The the amount in a cylinder with base
amount
PQQ/P and
of liquid crossing altitude unity.
PQ is
therefore
of this cylinder is PQsin0 qdtsm6As, where amount of liquid crossing the whole curve C in the time dt

PP
=
=P
is
The volume PQ. Hence the
Liin
V qdt **
sin 0As
=
is
q sin 6As
=
dt
I
q sin 6ds
J(C)
and the amount per unit of time
L
q sin eds.
260
To put
Then
this in the
LINE INTEGEALS
standard form,
sin 6
let
be the angle
cos
sin a.
=
a and
q sin 6ds
=
made by PQ with OX.
vdx
f
sin
cos
a
Hence
=
q sin
cos ads
q cos
sin
ads
=
udy.
is
Therefore the amount of liquid flowing across
C per unit
of time
L
(
vdx
+ udy).
Ex. 3. Heat. Consider a substance in a given state of pressure p, volume u, and temperature t. Then p, u, are connected by a relation /(p, = 0, so ) that any two of them may be taken as independent variables. For a perfect gas pv = kt, if t is the absolute temperature, and the state of the substance is indi
,
cated by a point on the surface of
fig.
55, or equally well
by a point on any one
of the three coordinate planes, since a point on the surface is uniquely deter mined by a point on one of these planes. shall take t and v as the inde
We
pendent variables and shall therefore work on the
(t,
v)
plane.
is
Now
if
Q
is
the
amount
of heat in the substance
v,
and an amount dQ
added,
there result changes dp, do, dt in p, itesimals of higher order,
and
t
respectively, and, except for infin
Cdt.
dQ = Adp + Bdv +
73,
From
the fundamental relation /(p,
t)
=
0, it
follows that
dp
cv
dt
whence we have
Hence the
total
amount
of heat introduced into the substance
is
by a variation
of its state indicated
by the curve C
Q=f
Ex. x
4.
(Mdt
+ Ndo).
Area.
6,
Consider a closed curve
=
Y
a,
x
=
y
=
d,
and y
=
e,
and
C (fig. 98) tangent to the straight lines of such shape that a straight line parallel to
it
either of the coordinate axes intersects
in
not more than two points.
E
through any point
M
Let the ordinate intersect C in PI and
P
if
2
,
where
is
A
2 = yz y and Then, the area inclosed by the curve,
MPi =
MP
h
.
A=
C\jodx Ja
s&gt;a
 f va
=A
I
M
FIG. 98
X
Jb
y 2 dx

=  f ydx, *JtT\
C
in
the last integral being taken around the hands of a clock.
a direction opposite to the motion of
FUNDAMENTAL THEOEEM
Similarly,
if
201
NQi =
Xi
the line
NQ =
2
NQ
Z
intersects
s*e
/
C
in Q!
r*e \
and Q 2
,
where
and
Xo,
we have
A=
Jd
x z dy

x^dy
J.i
~d
Xe%idy + Je
I
xidy
=
f J (C)
xdy,
hands of a clock.
the last integral being taken also in the direction opposite to the motion of the By adding the two values of we have
A
2A = C (ydx + xdy). J
(C)
If
we apply
(I,
this to find the area of
166).
an
ellipse,
we may
take x
=
a cos 0,
y
=
b sin
HH&gt;
Then
A_
integrals
144.
curves,
Fundamental theorem. In using
we need some means
in
of distinguishing
around closed between the two
directions
which the curve
is
may when
be traversed.
the curve
Accordingly, a portion of
//
the boundary of a specified area, we shall define the positive direction around the curve as that in
^
H
^ Jk
FIG. 99
is
^
JH
H*A
mil
which a person should walk in order to keep the area on his left hand. Thus in fig. 99, where the
shaded area
is
bounded by two
curves, the positive direction of each curve
indicated by the arrows.
of
With
this convention, the
:
fundamental theorem in the use
line integrals is as follows
If M, N, cij
,
and   are continuous and onevalued in a
dx
closed
area
A and
on
its
boundary curve
C, then
r r /dM
\\ \^^ jj(A)\vy
where the double integral is taken over taken in the positive direction around
A and
C.
the line integral is
262
To prove
this,
LINE INTEGRALS
we
fig.
shall first
4,
the form described in Ex.
that example and
98.
assume that the closed area is 143, and shall use the notation
of of
Then
1
,
dM
,
,
dxdy
= C ax C
I I
v iu
7
dy
b
X
= C M(x,
Ja
&gt;
=  C Mdx,
!/&lt;O
rc
2/ 2 )
dx
 C M(x,
Ja
b
I
y^ dx dx
M(x,
7/ 2 )
dx
M(x,
y^)
Ja
(1)
the last integral being taken in the positive direction around C.
Similarly, ^
it
(JJ\
,
.
/
^^
/c\\
(
2)
By
a subtracting (2) from (1) we have the theorem proved for closed curve of the simple type considered. The theorem is now readily extended to any area which can be cut
up
into areas of this simple type.
For example, consider the
area bounded by the curve C (fig. 100). the By drawing the straight line
LK
A
area
r
is
divided into three areas
,
A",
A"
,
and the theorem applied
to each
of these areas.
By
adding the three
the
the
equations lefthand side of
~
obtained
we have on the new equation
X
double integral over the area bounded by (7, and on the righthand side the
line
line integral along C and the straight traversed twice in opposite directions. The integrals therefore cancel, leaving only the integral along the line
LK
LK
around
C.
LINE INTEGRALS OF THE FIRST KIND
The theorem
curve.
is
263
also true for areas
bounded by more than one
Consider, for example, the area bounded by two curves
1
C
and C
(%
101).
By drawing
is
the line
into one
LK
the area
turned
bounded by a single and the theorem may be curve,
applied to it. It appears that the line integral is taken twice
in opposite directions, along and these two integrals cancel
LK
1.1
each other.
1.1
The
i.
result
i
the double integral over the area is equal to the line integral around each of the boundary curves in the positive direction of each. In the same way the theorem is
j
^
is
that
FlG 101

shown
Ex.
to hold for areas
If
bounded by any number
x,
of curves.
Jf
=
y and
N=
we have
C(ydx
ff 2 dxdy = agreeing with the result of Ex.
4,
 xdy),
93.
145. Line integrals of the first kind,
are two functions of x
sion is restricted
to
and
y,
such that
2p  =
dy
.
=
jj
&gt;
If
M and N
the discus
and
dx
a portion of the
..
.
(x, y]
,
plane in which M, N,
_
aar
&gt;
ana
,dN
dx
are continuous
I
ana
onevalued, t/ien
f
dy
(1)
The
line integral
(Mdx
x&gt;
Ndy) around any
between
closed curve
is zero.
(2)
The
line integral
I
(Mdx
+ Ndy}
any two points
is
is
a function of
of
(3)
the coordinates
of the points, and
y} such that
independent
the curve connecting the points.
There exists a function
&lt;f&gt;(x,
=M,
=N
.
Conversely, if
any one of the
conclusions (1), (2), or (3) is fulfilled,
then

=
dx
dy
264
To prove
is at
,
LINE INTEGRALS
(1),
consider the fundamental theorem of
if
144.
It
once evident that
,
closed curve.
dy
=
ox
&gt;
I
(Mdx+Ndy) = Q around any
j
be any two points, and C and C To prove (2), let L and any two curves connecting them. Let / be the value (fig. 102) of the line integral from L to along C, and I the value of the C Then I is the value of the line integral from L to along
K
K
K
.
line integral
from
K to L along
C
f
Now
by
(1)
Therefore
/=/
(3),
(x, y)
.
To prove
consider the line in
X
FIG. 102
tegral
x
I
(Mdx+Ndy),
a variable point. By (2) and is therefore fully
of a
/&lt;**)
where
(a?
,
y
)
is
a fixed point
and
(x, y)
this integral is independent of the path
determined when (x y}
t
is
given.
Hence, by the definition
write
function of two variables,
/&lt;,*&gt;
we may
(Max + Ndy) = 9 (x,
S*(.x+h, V )
yj.
Then
f(0.+ Ay)
I
&lt;
/(a*
2/0)
and since this integral is independent of the curve connecting the upper and the lower limits, we may take that curve as drawn first to (x, y) and then along a straight line to (x + h, y). Then
(a,
= C
**
,
y)
s*(x + h,y)
(Mdx
2/o)
+
Ndy)
Jt
+
I
(Mdx
.
+ ZV#y)
+
^( x
V)
Xx
since in the last integral y
is
M(x,y)dx,
constant and dy
=
0.
Then, by
30,
&lt;(#
f
h, y)
(f&gt;
(x, y)
= liM(%,
y),
d&gt;
whence
(x
+ h,
y}
rf
LINE INTEGRALS OF THE FIRST KIND
Letting
Ji
265
approach zero and taking the
limit,
we have
%*&lt;*&gt;
In like manner
we may show
that
and the third conclusion
noted that
is
(f&gt;
of the
theorem
is
proved.
It is to be
may
determined except for an arbitrary constant which be added. This constant depends upon the choice of the fixed
1
&lt;
differs from point L. If another point L is chosen, the value of in using L by the value of the line integral between that obtained
L
and
L
.
We must now show that, conversely, if any one of the conclusions = of the theorem is fulfilled, then dy Let us
first
is
dx
assume that the integral around any closed path
;
whatever
is
zero
we wish
i
4
to
show +1, f that
 = n
dy
0.
If

not zero at
all points, let
us suppose
it is
ox dx dy not zero at a particular
point K.
pothesis,
Then, since
dM dN   is
dy
dx
are continuous functions by hydx also a continuous function, and hence has, at
and
dy
points sufficiently near to K, the same sign which it has at K. It is therefore possible to draw a closed curve around K, so that
 has
dy
curve.
dx
the same sign for
all
points in the area bounded by the
this area is not zero,
Then
rr/dM
I I
(
dN\ 
1
JJ\dy
and therefore
/
dx ]
dxdy taken over
(Mdx + Ndy) taken around the bounding curve
is
is
not zero. But this
contrary to the hypothesis that
is zero.
C
I
(Mdx
+ Ndy)
at all
taken around any curve whatever
points.
Hence
*y
=
dx
Let us
now assume that

(
Mdx + Ndy) between any two points
is
is
independent of the curve connecting the points. This
equivalent
266
to
LINE INTEGRALS
is
assuming that the integral around a closed curve
zero,
and
hence, as already shown,
Finally, let
cd&gt;
dM = dN
dy
dx
&lt;
us
assume that a function
exists
such that
each
is
*
dx
=M M,
r
,
d& dy
= N.
Then
,
it
follows that
dM = dN
dy
&gt;
since
dx
equal to ^^. dxdy
The theorem
Ex.1.
x2
of this article is
now
I
completely proved.
Hence
Let3f=^,
=
0, if
N=^
then
then^=g= ff~^V
These conditions are met
if
y* their derivatives are finite
+
taken around a closed curve within which M. N. and
and continuous.
In fact,
if
the
curve does not inclose the origin.
placing x
we introduce
=
polar coordinates,
=
r cos
0,
y
=
r sin
0,
f
=
=
Cdd.
Now,
value
for
any path
to the
which does not
inclose the origin,
varies
from
its initial
a back
same value, and therefore
varies
Cdd = 0.
If the path winds once around the origin,
I
from a
to
a
f
2
TT,
and therefore
dd
2
TT.
The function
Ex.
,
of the general theorem
If
is,
in this example, equal to 6
=
tan * 
.
2.
~
Work.
,
X and Y are
components of force
in a field of force,
and
is
p
=
y
dy
dx
then the work done on a particle moving around a closed curve
zero, and the work done in moving a particle between two points is independent of the path along which it is moved. Also there exists a function 0, called a
ponents of force parallel to the axes of x and
of
in
force function, the derivatives of which with respect to x and y give the com y. It follows that the derivative
any direction gives the force
is
as this
called a conservative force.
of
forces
which are a function
111). Such a force Examples are the force of gravity and the distance from a fixed point and directed along
in that direction (Ex.,
straight lines passing through that point.
Ex.
3.
clear that the net
Flow of a liquid. If we consider a liquid flowing as in Ex. 2, 143, it is amount which flows over a closed curve is zero, since as much
enters, there being no points within the area at
off.
must leave the closed area as
which
liquid
is
given out or
;
drawn
Hence f (
vdx
+
udy), taken over
any
closed curve, is zero
and consequently
dy
d\f/
+
v,
= 0.
dx
There
exists also a function,
usually denoted by ^, such that

=
o4&gt;
=
u.
ox
dy
LINE INTEGRALS OF THE SECOND KIND
146. Line integrals of the second kind,
267
case
and
gral
N are such functions that
~
I
dM  =. dN
dy
dx
dy
=. j
dx
of the
n
M
&gt;
the value
line inte
(Mdx+Ndy)
depends upon the path, and there exists no func
tion of which Ex.
1.
M and N are partial derivatives.
(ydx
C
J
(ri
/l)

xdy).
(0, 0)
Let us
first
integrate along a straight line connecting
is
and P!
ydx
(fig.
103).
The equation
of the line
y=
z,
and therefore along
is
this line
 xdy = 0,
and hence the value
Next,
let
of the integral
zero.
us integrate along a parabola connecting
is
and PI, the equation of which
parabola
)
y
2
=
yl
x.
Xl
Along
this
(ydx
 xdy} = ^
y,
O
Next, let us integrate along a path consisting of the
two straight lines OM\ and J/iPi. Along OJ/i, y = and dy = 0 and along J/iPi, x = xi and dx = 0.
reduces to
/*9t
/
Hence the
line integral
Jo
xidy
=
x
x\yi.
Finally, let us integrate along a path consisting of the straight lines
and
NiP
lf
Along
OZVi,
=
ONi
and dx
=
C
Jo
0;
and along
JViPj, y
=
y l and dy
=
0.
Therefore the line integral reduces to
\fidx
= &\y\.
Ex.
2.
Work.
\
"Tp
If the
*s
T7"
components of force
X and
Y
in a field of force are
such that
^
,
then the work done on a particle moving between two
points depends upon the path of the particle, the work done on a particle moving around a closed path is not zero, and there exists no force function. Such a force is called a nonconservative force.
Ex.
3.
Heat
If a
substance
is
brought, by a series of changes of temperature,
pressure, and volume, from an initial condition back to the same condition, the amount of heat acquired or lost by the substance is the mechanical equiva lent of the work done, and is not in general zero. Hence the line integral
Q = J (Mdt +Ndv) around a
closed curve
is
not zero, and there exists no function
is
whose partial derivatives are and N. In fact, the heat Q t and TJ, not being determined when t and v are given.
M
not a function of
Ex. 4. Adiabatic lines. The line integral C(Mdt +Ndv)of Ex. 3 along a curve whose differential equation is
is
zero
if
taken
Mdt + Ndv =
0.
268
LINE INTEGRALS
t
We will change this equation by replacing the variable done by means of the fundamental relation /(p, 75, t) = 0.
It is
by
p,
which can be
shown
in the theory of heat that for a perfect gas the equation then
becomes
vdp
+
ypdv
=
0,
is
where 7
is
a constant.
The
solution of this equation
If, then, a gas expands or contracts so as to obey this law, no heat is added to or subtracted from it. Such an expansion is called adiabatic expansion, and
the corresponding curve
is
called
an adiabatic curve.
is in the concrete illustration of a line integral being zero along any portion of a certain curve or family of curves.
The mathematical
interest here
In the line integrals of the kind before
still
us,
/

^
)&gt;
there
is
a
meaning
to be attached to
M and N.
;
For, suppose the inte
c
gral to be
and
let
taken along the straight line y = Au be the value of the integral that
from x to x
+ Ax,
is,
Au =
f*(x
I
+ Ax,
y)
(Mdx
2/)
+ Ndy)
dx
30)
/(*
X
_
M(x,
,
y)
y),
(
where
Consequently
Lim
=M(x,

y).
Similarly,
Lim
= N(x,
y}.
We
shall write these
two equations
as
/rfu\
AfoA
\dx] y
It
is
\dy/ x
these
derivatives are
of
to
be
noticed
that
different
character from the .partial derivatives of Chap. XI, since u is not a function of x and y. The property proved in 117 does
not hold here.
EXACT DIFFERENTIALS
Ex.
6.
269
Heat. Returning to the notation of Ex.
3,
we have
dt/v
\dv
M
is
the limit of the ratio of the increment of heat to the increment of
tem
as sensibly perature when the volume is constant. Therefore, if we consider constant while the temperature changes by unity, may be described as the amount of heat necessary to raise the temperature by one unit when the volume
M
M
is
constant.
Similarly,
JVmay be
described as the
amount
is
of heat necessary to change the
volume by one unit when the temperature
constant.
147. Exact differentials.
We
dd&gt;
have seen that
two functions
of
x and y satisfying the condition
&lt;f&gt;
exists a function
such that
dx
=M
dd&gt;
y
= N.
= ^
Then
if
M and N are
&gt;
there
^x
Then
cy
d4&gt;,
Mdx + Ndy = ^dx + ^dy =
dx
dy
and
Mdx + Ndy
is
called
an exact differential*
(/&gt;
also
x
b, .vo)
+ Ndy) =
(x v y,)
c/&gt;
(XQ
,
y
),
/ex
c/&gt;
the integral being independent of the curve connecting (# yQ ) and may be found by computing the line inte (x v y^). The function
,
gral along a conveniently chosen path, but
Q
J
it is
usually more cony), it
venient to proceed as follows
:
Since
= M(x,
follows that
M(x, y)dx, where y
is
considered constant, will give that part
all
of
which contains
y.
&lt;/&gt;
x,
but not necessarily
the part which
contains
We may
write, therefore,
where f(y)
is
is
a function of y to be determined.
relation
This determination
made by using the
is
j^=N.
yy
Then
CMdx +f (y) = N,
fiyj be found.
which
*In
an equation from which f(y)
may
where the emphasis is on the fact that both x and y are varied, is Here the emphasis is shifted to the fact that is exactly obtained by the process of differentiation, and hence it is called an exact differential.
110,
d&lt;f&gt;
called a total differential.
d&lt;j&gt;
270
This method
is
LINE INTEGRALS
valuable in solving the equation
=
when
called
0,
the condition
=
dy
is
satisfied.
dx
Such an equation
is
an exact differential equation.
Its solution is
where
the
c is
an arbitrary constant and
just described.
is
&lt;/&gt;
the function obtained in
manner
Because of the importance of the method of obtaining
it
(/&gt;,
we
give
in a rule as follows
:
Integrate
I
Mdx, regarding y as constant and adding an unknown
;
function of y
the
new result to ; from unknown function of y.
If
differentiate the result with respect to y and equate the resulting equation determine the
N
more convenient, the above
:
rule
may
be replaced by the
following
Integrate
\
Ndy, regarding x as constant and adding an unknown
;
function of x
the
new result to ; from unknown function of x.
Ex.
1.
differentiate the result with respect to x and equate the resulting equation determine the
M
3
(4x
3
+
10x7/

3y*)dx
+
,
2 2 (15 x ?/

12 xy s
+
5y*)dy
is
= 0.
Pro
Here
ay
= 30 x?/2
12 y s
=
ax
and the equation
therefore exact.
ceeding according to the rule,
we have
=
(4x3
+
I0xy8 3y*)dx +f(y)
Then from
*
= N. we

have
XT/
3
15 X 2 ?/ 2
12
+ f (y) =
7/
15 X 2 7/ 2

12
XT/
3
+
5
7/4,
whence /
therefore
= (?/)
57/
4
,
and/(7/)
=
5.
The
solution of the differential equation
is
X4
+
5 X 2 7/ 3

3X7/4
+
2/5
=
C
.
THE INTEGKATING FACTOR
Ex.2,
i
271
g \ dx+
x
Here
dy
=
(1/2
xV2/ 2
x
r
2
/
_ X 2)2
=
sx
and the equation
is
therefore exact.
Following
the second rule,
we have
= f J
From
dy
2

= log(y + Vy 2  x 2 ) + /(x).
x2
= Jf, we
have
whence /
(x)
= 0,
to
which reduces
x  x2 (y + V?/ 2  x 2 ) x V?/ 2  x 2 and /(x) = c. Hence the solution of the differential equation is 2  x2 =c log (y + Vy ) 2 2 cy + c 2 = 0. x
,
148.
The integrating
factor.
We
=
have seen that
Mdx + Ndy
is
But in all cases there dx dy exist an infinite number of functions such that, if Mdx Ndy is one of them, it becomes an exact differential. multiplied by any
not an exact differential
when
+
Such functions are
of
fjL
the
called integrating factors. That integrating factors of Mdx Ndy, then,
is, if
//,
is
one
+
by
definition,
(Mdx
Ex.
1.
+ Ndy)
is
an exact
differential.
is
The expression ydx
ydx
xdy
not an exact differential. But
xdy
ydx x2
xdy
+
=
,1 x\ d tan i , [
.
y
2
)
\
y]
are exact differentials, and the functions
,
x2
x2
+
z/
,
2
are integratine factors
xy
shall
To show that integrating factors always exist, we assume ( 173) that the differential equation
need
to
Mdx + Ndy =
has always a solution of the form
(1)
f(x where
c is
t
y, c)
=
0,
(2) (2)
an arbitrary constant, and that
&lt;M#,y)
can be written
(3)
=
c.
272
LINE INTEGRALS
d(f&gt;
Now
HT
/ox
i
(3) gives
us
dv
dx = d(f&gt;
,
dx
8y
,
and
X1
.
.
(1) gives
us
M dy = dx
f
N
;
and since
the same.
(3) is
the solution of
is,
(1),
these two values of
y7/v*
must be
That
d&lt;
&lt;ty
M~N
where
ft is
dx

dy
^ ~
IL
some function
of
x and
y.
Consequently
and therefore /n is an integrating factor. To prove that there is an infinite number
of integrating factors,
multiply the last equation by /($), any function of
Then
(/&gt;.
x
+
/*/((/&gt;)
Ndy
=/(&lt;/&gt;)
ety.
But Hence
Ex.
path
is
is
d&lt;f&gt;
f(4&gt;)
the exact differential of the function
/
is
fJ/((f&gt;)
an integrating factor
has been noticed
(
of
Mdx + Ndy.
3) that
2.
jEeai.
It
146,
Ex.
TdQ around
Hence
a closed
not zero, and consequently
f
dQ
is
not an exact differential.
is
It is
found
is
in the theory of heat that
around a closed path
fin
zero.
an
is
t
exact differential, and called the entropy.
we may
write
=
d&lt;f).
The function
0, thus defined,
149.
No
general method
is
known
for finding integrating factors,
but the factors are
known
for certain cases.
We
give a
list of
the
by simpler cases, leaving it differentiation that each of the equations mentioned satisfies the condition for an exact differential equation after it is multiplied
as an exercise for the student to verify
by the proper
factor.
THE INTEGRATING FACTOR
d_M__d_N
*
=/(&lt;)
273
1.
If
JL
2.
If
JL
_
M.
X
then
e*
is
an integrating factor
of
f.
* =/(;y), then e
V
is
an integrating factor
of
3.
Mdx+Ndy=0.
If
M and N are homogeneous
is
and
of the
same degree, then
0.
xM+yN
4.
an integrating factor
;
of
Mdx + Ndy =
If
M = yf^xy), N = xf^xy)
Mdx+Ndy =
is
then
 is
an integrating
factor of
5.
Q.
e^
an integrating factor
of the linear
equation

+
As
factor only after
a practical point, the student should look for an integrating he has tried to integrate by other methods.
(4
Ex.
1.
x2 y

3 y 2 } dx
+
3
(x

3 xy)
dy
&lt;w
=
0.
dM
Here
^V
1^ =
r^
(4
I
.
X
After multiplication by the
x x is an integrating factor. Consequently e^ the equation becomes factor,
x 3 ?y

3 xy 2 } dx
the integral of which
is
x4 y
0.
+ (x4  3 x 2  %x 2 y 2 = c.
it
?/)
dy
=
0,
Ex.
2.
(x
2

y*)dx
+ 2xy% =
is
Since this equation
homogeneous,
1
has the integrating factor
1
xM + yN
x2
~
x3
+
x?/
2
After multiplication by the factor, the equation becomes

y
2
,
dx\
2xy 
dy
=
0,
\X 2
+
2
7/
X
the solution of which
is
log (x
2
+
y
2
)
log x
=
c
,
or
x
+
= c.
x
274
For
this equation
LINE INTEGRALS
we have
also
Hence
it
has also the integrating factor
x2
After multiplication by
this,
the equation becomes
the solution of which
is
X
+
=
C.
The equation may
also be solved
by the substitution y
of
=
vx(
78).
150. Stokes s theorem.
The theorems
144145, which are
:
limited to a plane,
may
be extended to space as follows
If
are three functions of x, y, line integral
I
and
P, Q, R z, the
(Pdx
+ Qdy + Rdz)
x
along a space curve is defined in a manner precisely similar to the
definition of
143.
Let the integral
be taken around a closed curve
FIG. 104
(fig.
C
be
104) and
let
a
surface
S
bounded by
cos a, cos
/3,
C.
Let
dS
be the element of area of the surface and
of its normal.
cos
7 be the direction cosines
Then
dR
dP
~
\cy
=  C (Pdx + Qdy + Rdz),
J(C)
where the double integral is taken over the surface S and the single integral is taken around C, and the direction of the line have the relation of integration and that of the normal to
fig.
104.
STOKES
To prove
this, let
S
THEOREM
275
z=f(x,
y)
be the equation of S, and consider
Then, since P(x,
y,
I
P(x,y,z)dx.
y)]
J(C)
z)=P[x, y,f(x,
on
= I}(x,
y},
the values of
P which
correspond to points
to points
C are
which correspond
on
C",
the same as the values of JJ the projection of C on the plane
XOY. Hence

r
I
Pdx = 
r
I
l}dx.
(1)
J(C)
J(C )
But by
where S
144
is
Cl&gt;dx=
ff
d
j(C )
the projection of
JJ(S)vy
fdxdy,
(2)
S on
the plane
XOY.
Now
d% = dP
~dy~ dy
dPdz
dz dy
is
where the right hand of this equation Hence from (1) and (2) we have
computed
for points
on
S.
^_rr( + JJu\ty
)
fa ty/
:
fW
and
(
(3)
But
(8
112)
:
:
 1 = cos a
(3),
:
cos
/3
dx cy
cos 7,
92) dxdy
=
cosy dS.
Substituting in
we have
(
cos
 Cpdx = ff J
(C)
7
dz
cos
JJw\ty
0} dS.
/
Similarly

J(o
(Rdz = ff
JJfa
(
cos
{3
dy
cos
a}d&
in these Strictly speaking, the differential dS is not the same three results, since the same element will not project into dxdy dydz, and dzdx on the three coordinate planes. But since in a
t
double integral the element of area may be taken at pleasure with out changing the value of the integral ( 129), we may take dS as
equal in the three integrals.
required result.
Adding the equations, we have the
276
In the above proof
of the surface
LINE INTEGRALS
we have tacitly assumed that only one point over each point of the coordinate planes, and 7 are acute angles. The student may show that these
S
is
that a,
/3,
restrictions are unessential.
151.
From
:
the preceding discussion
we
x, y,
derive the following
theorem
If P, Q, and
R
are three functions of
and
z,
such that
dy
dx
cz
dy
dx
dz
and
the discussion is restricted to a portion of space in which the functions and their derivatives are continuous and onevalued, then
1.
The
The
line integral ine
I
(Pdx + Qdy \Rdz] around any closed curve
(Pdx f Qdy f Rdz) between any two points
is zero.
2.
line integral
I
is
independent of the path.
3.
There exists a function
TT~
&lt;f&gt;(x,
y, z),
such that
i
dx
7T~
&gt;
~7~
***
dy
dz
and Pdx
+
Qdy
+ Rdz
Conversely, if
is an exact differential d$. any one of the above conclusions is
fulfilled, then
ap
dy
= a^
dx
3Q
dz
= dR
dy
is left
a^ = ap
dx
dz
The proof
is
as in
145, and
to the student.
PROBLEMS
1.
cycloid x
Find, by the method of Ex. a cos 3 0, y = a sin 3 0.
4,
4,
143, the area of the fourcusped hypo
2. Find, by the method of Ex. cycloid and the fixed circle.
3.
143, the area
between one arch of a hypobetween one arch of an
Find, by the method of Ex.
4,
143, the area
epicycloid and the fixed circle.
4.
Show
case, the
that the formula for the area in Ex. formula for area in polar coordinates.
4,
143, includes, as a special
PROBLEMS
5.
277
+
(2
Show that the integral
its
/(!)
/ ^(0, 0)
[2
x (x
+
2 y) dx
x2

2
?/
)
iZy] is
independ
ent of the path, and find
6.
value.
/&gt;(3,
Show
4)
find its
that the integral value.
/
(xdx
+ ydy)
is
independent of the path, and
t/(5 0)
7. Find the force function for a force in a plane directed toward the and inversely proportional to the square of the distance from the origin.
origin,
8. Find the force function for a force in a plane directed toward the origin and inversely proportional to the distance from the origin.
9.
Prove that any force directed toward a center and equal to a function of
is
the distance from the center
10.
conservative.
/(!,!)
Find the value of
(1) (2) (3)
(4)
I
[(y

x)
dy
+
ydx] along the following curves:
(O.O)
the straight line x = t, y the parabola x = t 2 ,y =
the parabola x=t, y the cubical parabola x
[(x
t
=
t;
;
t;
= 2 =
t,
y
=
t*.
11.
Find the value of f
^(0,0)
(1)
 y*)dx + 2 xydy]
along the following paths
;
:
(2) (3)
a straight line between the limits the axis of x and x = 1
;
the axis of y and y
=
)
1.
12. Find the value of
f
(O, 0)
[y*fa
+
(
Xy
+ x2
dy] along the following paths
(1) (2)
y = 2xy = 2x 2
.
13.
Find the value of f
/(0 2
&gt;
[(l+y z )dx + (l+x 2 )cfy] along
the following
paths:
(1)
(2)
y
y=6x + 2; =5x*+ 2;
l
+ 2x
them
:
Show
that the following differential equations are exact, and integrate
14.
(2xy +
l)dx
+ (2yxl)dy =
0.
2
xy
+
3
2
?/
)
dy
_
.
\
dy
=
0.
278
LINE INTEGKALS
:
Solve the following differential equations by means of integrating factors
19. (x 2
+
2
T3
y )dx
2
dy
= +
0.
20.
(1
+
_x
y
y
+
x2 y)dx
/
+
(x
\
x 3 ) dy
Q.
=
0.
21. ye
dx y
3
)
\xe
22. (x
23.
3
2
dx
+y)dy = + xy 2dy = 0.
y
2
_
r
(x?/
3
y)dx +
(x
?/
+
dx
x)
dy
(2
=
xy
0.
0.
24. (5x 25. dy

3
xy
+
2
2
7/
)
+

x 2 ) dy
=
0.
(y tan x
cos x) dx
=
26. sin (x
+
y) (dx
+
dy)
y cos
(x 4 y)
dy
27. (y
+
2 xy 2
+
x 2 y 3 ) dx
+
(2
x 2y
+
x)
dy
= 0. = 0.
28.^
dx
1
+
x
CHAPTP:R
xv
INFINITE SERIES
152. Convergence.
The expression
a 1 +a a +a 8 +a 4
+a
5
+.
.
..,
(1)
where the number
series.
of the
terms
is
unlimited,
is
called
an
infinite
An infinite series is
sum of the first n
Thus, referring to
said
to converge,
or
to be
convergent,
when
the
terms approaches a limit as n increases without limit.
(1),
we may
place
Then,
if
Lim sn = A,
n= x
is said to converge to the limit A. The quantity A is frequently called the sum of the series, although, strictly speaking, it is the limit of the sum of the first n terms. The convergence of
the series
(1)
may
be seen graphically by plotting
scale as in
I,
s p s2 ,
,,,*
on the
number
53.
A
series
which
is
:
happen
in
two ways
not convergent is called divergent. This may either the sum of the first n terms increases
fail to
without limit as n increases without limit; or s n may a limit, but without becoming indefinitely great.
Ex.
1.
approach
Consider the geometric series
a
+
ar
+
ar2
+
ar 3
+
.
Here
less
sn
=
1,
a
r"
+
ar
+
ar2
f
+
ar"
1
=
a
.
Now
if
r
is
numerically
;
than
approaches zero as a limit as n increases without limit 279
and
280
therefore
INFINITE SERIES
Lim
n
= oo
sn
=
1

If,
however, r
is
numerically greater than
limit;
1,
r n in
r
creases without limit as
n increases without
the series
is
and therefore
sn
increases
without
limit.
If r
=
1,
afa + afa +
and therefore
sn
,
increases without limit with n.
If
,
n
=
1,
the series
is
a
a
+
a
a
+
and
s n is
alternately a and
0,
and hence does not approach a
limit.
Therefore, the geometric series converges to the limit
less
when
r is numerically
than unity, and diverges
2.
when
r is numerically equal to or greater
than unity.
Ex.
Consider the harmonic series
1
+
HH
+
+
l
+
c
+
+
+
^"
;
+
&lt;
consisting of the
sum
of the reciprocals of the positive integers.
Now
4
+
=
J&gt;l
i,
4
+1
H+
i &gt;f+ 4
+ 1 H = i.
way the sum of the first n terms of the series may be seen to be than any multiple of l for a sufficiently large n. Hence the harmonic greater
and
in this
series diverges.
153. Comparison test for convergence. If each term of a given series of positive numbers is less than, or equal to, the correspond
series converges. ing term of a known convergent series, the given each term of a given series is greater than, or equal to, the If corresponding term of a known divergent series of positive numbers,
the given series diverges.
Let
!+
2
+
as +
4
+
a positive
(1)
be a given series in which each term
is
number, and
let (2)
\+\+\+\+be a
known convergent series such that a k = b k Then if s n is the sum of the first n terms of
.
(1), s n
the
sum
of
the
first
n terms
of (2),
and
B
the limit of
s ^ it
follows that
**&lt;&lt;*
since all terms of (1), and therefore of (2), are positive. Now as increases, s n increases but always remains less than B. Hence B. approaches a limit, which is either less than, or equal to,
n
sn
The
first
part of the theorem
is
now proved
;
the second part
is
too obvious to need formal proof.
THE EATIO TEST FOE CONVEEGENCE
term
281
In applying this test it is not necessary to begin with the first of either series, but with any convenient term. The terms before* those with which comparison begins, form a polynomial the
value of which
is
of course finite,
and the remaining terms form
is
the infinite series the convergence of which
Ex.
1.
to be determined.
Consider
Each term
after the third
is
less
than the corresponding term of the convergent
geometric series
Therefore the
first series
converges.
Ex.
2.
Consider
V2
Each term
harmonic
after the first
is
V3
V4
V5
+
1
Vn
series
+
Therefore the
first series
2
1111 + + + +
3
4
5
greater than the corresponding term of the divergent
n
+
diverges.
154.
The
ratio test for convergence.
If in a
the
numbers the ratio of the (n +l)st term to a limit L as n increases without limit :
converges; if
L&gt;1,
series of positive nth term approaches
L&lt;1,
then, if
the series
the series diverges; if
L=l,
\
the series
may
(1)
either diverge, or converge.
Let
a^ +
2
3
H
h an
+
an + l
be a series of positive numbers, and
three cases to consider.
1.
L&lt;1.
let
lim
= L. We
1.
have
Taker any number such
approaches
that
L&lt;r&lt;
Then, since
the ratio
~
n
L
as a limit, this ratio
must become and
n.
remain
less
than r for sufficiently large values of
all
Let the ratio
be less than r for the rath and
subsequent terms.
Then
282
INFINITE SEKIES
compare the
series
Now
*
with the series
+^*
+
a
is
++*++
3
.
(
2)
V + y+ V +
a convergent series since
it
(3)
Each term
term
series of (3),
of (2) except the first is less
(3)
than the corresponding
a geometric (2) converges by the
is
and
with
its ratio less
than unity.
Hence
comparison
2.
L&gt;1.
test,
and therefore
(1) converges.
Since
^
this
approaches
L
as a limit as
n
increases
without
terms.
limit, this ratio eventually
than unity.
Suppose
happens
for the
becomes and remains greater mth and all subsequent
Then
Each term
term
of the series (2) is greater
than the corresponding
of the divergent series
Hence
3.
(2)
and therefore
(1) diverges.
is
L = l.
Neither of the preceding arguments
in this case the series
valid,
and ex
amples show that
diverge.
Ex.
1.
may
either converge or
Consider
2345
^
1
n
The nth term
is
3n
and the (n
+
l)st
term is^
The ratio of the
3
(n
f l)st
term to the nth term
is
=
&gt;
,
and
3n
=
3
3
Therefore the given series converges.
ABSOLUTE CONVERGENCE
Ex.
2.
283
Consider
l
+

+
* + * + ...
(r
+
g
is
+ ....
^
The nth term
is
and the
\n
+
l)st
term
n
n
+
l
The
ratio of the
(n
+
l)st
term
to the nth
term
is
(
n
(n
+ ^ n+1 = + n
i )
_,
and
Therefore the given series diverges.
ber
155. Absolute convergence. The absolute value of a real num is its arithmetical value independent of its algebraic sign. Thus the absolute value of both f 2 and 2 is 2. The absolute
value of a quantity a is often indicated by It is evident that \a\. the absolute value of the sum of n quantities is less than, or equal to, the sum of the absolute values of the quantities.
A
series converges
convergent series,
when the absolute values of its terms form a and is said to converge absolutely.
ai
Let
be a given
series,
+
a z + az +
a^\
(1)
and
KI +
a,
+ K +
K+&lt;
(2)
the series formed by replacing each term of (1) by its absolute value. assume that (2) converges, and wish to show the con
We
vergence of
(1).
Form
the auxiliary series
1
K+
)
+ K+
ak
2 )
+
(
3
+

3 )
+ K+a
and a k
4
)+....
(3)
The terms of (3) of (2). For a k =
positive.
are either zero or twice the corresponding terms
\ \
when a k
is
negative,
=
a k when a k
\
is
Now, by
hypothesis, (2) converges,
8

and hence the
2
4
series
2N+2,+2a +
converges. But each term
+...
(4)
of (3) is either equal to or less
than the
corresponding term of
parison test.
(4),
and hence
(3)
converges by the
com
284
INFINITE SERIES
let sn
Now
the
first
be the
sum
of the first
s"
n terms
of (1), s n the
sum
of
n terms
of (2),
and
the
sum
,
of the first
n terms
of (3).
Then
and, since
s"
*&lt;&lt;
and
sn
r
approach
limits, s n also
approaches a
limit.
Hence the
series (1) converges.
We
shall consider in this
tests of
chapter only absolute convergence.
applied, since in testing for
Hence the
156.
153, 154 may be
all
absolute convergence
terms are considered positive.
The power
aQ
series.
A
power
series is defined
by
,
+
,
a :x
,
+
a z z? + a & x* \\ a n x n \are
where a
,
a lt a 2 a s
We
that
\x\
shall prove the following
verges for x
&lt;
=x
lt
it
numbers not involving x. theorem If a power series con converges absolutely for any value of x such
:
XT\.
For convenience,
the series
let
[a;
= X,
an
9
=A
\
n,
x{
\
= Xr By
hypothesis
(1)
aQ +
a^ + a^xl + a xf +
to
.
.
.
+
a n x? H
converges, and
Q
we wish
show
Z
that
A + A^X + A X* + A
Z
X+
Z



+A X
n
n
+
.
.
(2)
converges if r Since (1) converges, all its terms are finite. Consequently there must be numbers which are greater than the absolute value of
&lt;
X X
any term
A n Xi
&lt;
M
of (1).
Let
M be
one such number.
Then we have
for all values of n.
Then
Each term
sponding term
of the series
of
(2)
is
therefore less than the corre
the series
But (3) is a geometric Hence, by the comparison
series,
which converges when
converges
test, (2)
when
X X Xr
&lt;
&lt;
THE POWER SERIES
From the preceding discussion it may behave as to convergence in one
1.
285
follows that a power series
of three
ways only
:
It
It It
2.
3.
may converge for all finite values of x (Ex. 1). may converge for no value of x except x = (Ex. 2). may converge for values of x lying between two finite
numbers
(Ex. 3).
R
and
+R, and
diverge for all other values of
x
In any case the values of x for which the series converges are together called the region of convergence. If represented on a
number
the region of convergence in the three cases just enumerated is (1) the entire number scale, (2) the zero point only, (3) a portion of the scale having the zero point as its middle point.
scale,
Ex.
1.
Consider
1
The nth term
n of x and
n
is
rpn 1
+

+
__
,
+ ?!+... +
the (n
x.
^ + 711
is
.
.
.
.
npn
x Lim  = =
[re

+
l)st
term
,
and
their ratio
is
1
[n
n
/j*
for
its
any
finite
value of
Hence the
series converges for
any value
region of convergence covers the entire
number
scale.
Ex.
2.
Consider
1
+
x
+
[2
x2
+
1
,
_3
x3
+
h
[n

1
xn ~ l
+
,
.
The nth term
is
\n
Ix"
the (n
+
l)st
This ratio increases without limit for
all
term is nx n and their ratio is nx. values of x except x = 0. Therefore
the series converges for no value of x except x
=
0.
Ex.
3.
Consider
l
+ 2x + 3x 2 f4x 3 +
nx"
H
nx"
l
(n
+
+
l)x
.
( n n=m n = ao \ and diverges when x 1. x ber scale between 1 and + 1.
n
&lt;
The nth term
x.
Lim
=
is
1
,
the (n
1 f
+
l)st
term
is
n
,
and
their ratio
is
x
=
x Lim
)
n]
=
x.
Hence the
series converges
when
1
&gt;
The region
of convergence lies
on the
num
A
power
series defines a function of
region of convergence,
x for values and we may write
f
of
x within the
f(c)
it
a
+
avx
a 2 x2 + a s x s +.
+a
n
xn +
.
.
.
,
(4)
sum
being understood that the value of f(x) is the limit of the of the series on the right of the equation. We shall denote
286
INFINITE SEEIES
n(
%
s
x) the polynomial obtained by taking the
(4).
first
n terms
of
the series in
Thus
Graphically,
etc.,
if
we
plot the curves
y
= s^x)
,
y
=s
z (x),
y
=s
a (x),
we
shall
have a succession
of curves
which approximate
to
the curve of the function y=f(x). These curves we shall call the approximation curves, calling y = s^x) the first approximation curve, y = s z (x) the second approximation curve, and so on. The
grapli of
y
= f(x)
we
shall call the
limit curve.
Ex.
4.
Let
The
limit curve
(fig.
is
the portion of the
hyperbola
105)
between z 1 and z = 1. The first ap proximation curve is the straight line y = 1, the second approximation curve is the straight line y = I + z, the third approximation curve is the parabola y = 1 + z + z 2 etc. In fig. 105 the limit curve
,
=
drawn heavy and the first four approxi mation curves are marked (1), (2), (3), (4).
is
It is to
X=l
be noticed that the curves, ex (1), cannot be distinguished from each other for values of z near zero.
cept
The power
portant property, not possessed by all kinds of very similarly to a polynomial. In particular
:
series
has the im
series, of
behaving
1.
2.
The function defined by a power series is continuous. The sum, the difference, the product, and the quotient
of
two
functions defined by power series are found by taking the sum, the difference, the product, and the quotient of the series.
MACLAURIN
3.
S
AND TAYLOR
ajc
S
SERIES
n
287
If
f(x)=a +
sib
/*
+ a z? +
2
...
+
+
anx
+
.
.
then
b
s*b
s*b
Xb
4.
f(x) dx
=
/
a dx +
lie in
/
djXdx
+
I
a 2 x*dx
+
/
an
c/a
i/a
t/a
ya
provided a and &
If
the region of convergence.
3 3
f(x)
=
+ a^ + a^H a
2 az x
Hh a n a?\.
.
,
then
f
(x)
= 0,^+
+
3 asx
2
+
+ na n x
n
~l
+

.
.
,
and the
series for the derivative
has the same region of convergence
as that for the function.
For proofs
treatises.*
of these
theorems the student
is
referred to advanced
157. Maclaurin s
and Taylor
s
series.
When
a function
is
expressed as a power series, it is possible to express the coeffi cients of the series in terms of the function and its deriva
tives.
For
let
By
differentiating
l
we have
3
f (x)=a + 2a x+ 3a
2
^ 2 +4a 4 ^ 3 H na n xn
\
~l
f
=
"(x)
3

2
3
+
4 3 2
.
ap +
+ n(n  l)(n 2)a n xn  +
3
.

.
,
/&gt;(*)=[n(l)(n2)...3.2]a.
Placing x
+ ....
we
find
=
in each of these equations,
,=/(0),
o,=/
(0),
2
Consequently we have
^+...
This
is
+
^+....
(1)
Maclaurin
s series
(
31, (6)).
*
See, for example, GoursatHedrick, Mathematical Analt/six, Chap. IX.
288
Again,
if
INFINITE SERIES
in the righthand side of
f(x)
=
+ ajc + a^ + ***?+
2
r
+
to
&lt;*&gt;&
+
of
we
place x
=a+x
=
5
,
and arrange according
2
powers
x we have
,
f(x)
or,
+ &X + &X + &8
its
8
+
+ &X B +
by replacing x by
value x
a,
By differentiating this equation successively, in the results, we readily find
and placing x
=
a
Hence
Taylor s series ( 31, (4)). have here shown that if a function can be expressed as a the series may be put in the form or (2). In 31 power series, (1) we showed that any function which is continuous and has con tinuous derivatives can be so expressed. Usually when a known
is
This
We
expressed by either (1) or (2), the region in which the expression is valid as a representation of the function is coincident with the region of convergence of the series. can be
is
function
Examples
is
given in which this
is
not true, but the student
not likely to
meet them in
practice.
158. Taylor s series for functions of several variables. Con sider f(x, y) a function of two variables x and If we place y.
x
=a+
lt,
y
= b + mt
t
t
where
a, b,
I,
and
m
&gt;
are constants
y}
and
is
a variable,
we have
f( x
= f( a +
lt,1&gt;
+ mt)=F(t).
TAYLOR
Now, by expansion
S
SERIES
289
into Maclauriii s series,
(1)
and, by
111 and 118,
if
When ^ = 0, we have x = a and y = b. Hence ^(0) =/(, &), and, we denote by a subscript zero the values of the derivatives of f(x, y) when x = a,y = b,
By substitution we have
/(*. y)
in (1), noting that
lt
=x
a and
=/(,
)+(*  +(y )
*)
290
INFINITE SERIES
series
Another form of this b = k. Then and y
may
be obtained by placing x
a=
In a similar manner, we
may show
that
The terms
of the
nth degree in this expansion
may
be indicated
symbolically as
159. Fourier s series.
A
2
series of the
form
a n cos
w^c
an
2

+ cos + + \ sin +
ctj
as a;
cos 2 x
a?
5 2 sin 2
+ +

+
.
.
.
+ + b n sin nx +
(1)
where the
coefficients
P &2
,
do not involve x and are
determined by the formulas derived in 160, is called a Fourier s series. Every term of (1) has the period* 2 TT, and hence (1) has
that period. Accordingly any function denned for all values of x by a Fourier s series must have the period 2 TT. But even if a function does not have the period 2 TT, it is possible to find a
Fourier
of
s
series
x between
f(x)
is
TT
which and
will represent the function for all values TT to TT TT, provided that in the interval
if
called a periodic function, with period k,
f(x
\
k)
f(x).
FOURIER
the function
is
S
SERIES
291
singlevalued, finite, and continuous except for finite discontinuities,* and provided there is not an infinite number
of
maxima
160.
or
minima
in the neighborhood of
any
point.
try to determine the formulas for the coeffi TT cients of a Fourier s series, which, for all values of x between
will
We
now
and
TT,
shall represent a given function, f(x),
which
satisfies
the
above conditions.
Let f(x)
= ^ f
2.
&lt;x
a 1 cos x
+ &j
To determine
TT,
,
+ sin x +
2
cos 2 x
+
j
& 2 sin 2
x
+ +
a n cos nx
&n
+ sin wa; +
.
(1)
TT
multiply
result
(1)
is
by dx and integrate from
to
term by term.
The
/
whence
aQ
,
7 77
=
T f(x) dx, Vjn
1
7
"
I
(2)
since all the terms on the righthand side of the equation, except vanish. the one involving
To determine the
term.
coefficient of the general cosine term, as
an
,
multiply (1) by Since for
cos
all
and integrate from integral values of m and n
dx,
nx
TT
to
TT,
term by
f
u/
1
sin
mx cos nx dx =
mx cos nx dx =
I
0,
T
cos
(m
TT,
=
?i)
and
all
rcos nx dx = ;
2
C/7T
the terms on the righthand side of the equation, except the one involving a nJ vanish and the result is
r
f(x) cos
T
nx dx
=a
n 7r,
whence
an
=
I
C
I
n
f(x) cos nx dx.
(3)
* If e) and /(Zi+&lt;0 have different limits as e z, is any value of x, such that/fo approaches the limit zero, then/(z) is said to have a finite discontinuity for the value x = x\. Graphically, the curve y =f(x) approaches two distinct points on the ordinate x = Xi, one point beinj? approached as x increases toward cc,, and the other being approached as x decreases toward x\.
292
It is to
INFINITE SEEIES
be noted that (3) reduces to
,
(2)
when n
=
0.
In like manner, to determine b n multiply (1) by sin TT to TT, term by term. The result is integrate from
nx
dx,
and
=i
For a proof
formulas
f^Jn
;)
sin
nx dx.
(4)
of the validity of the
(2), (3),
and (4), the reader
is
above method of deriving the referred to advanced treatises.*
development
to hold for all values
of x
Ex. 1. Expand x in a Fourier TT and TT. between
s series, the
By
(2),
(:37r,0)
FIG. 106
Hence only the
coefficients being
1, 2, 3,
sine terms appear in the series for x, the values of the determined by giving n in the expression for b n the values in succession. Therefore 61 = 2, 6 2 = and f 63 =
,
,,
_
sin
x
sin 2
~~
x
sin 3 x
~~~
The graph of the function x is the infinite straight line passing through the origin and bisecting the angles of the first and the third quadrants. The limit curve of the series coincides with this line for all values of x be
tween
vanishes
points
(
TT
and
TT,
when x
TT,
=
for every term of the series TT but not for x = TT and x TT or x = TT, and therefore the graph of the series has the
;
0) as isolated points (fig. 106).
By
1, 2,
TT and TT, and giving k the values taking Xi as any value of x between in succession, we can represent all values of x by x\ 2 for. But the 3,
*
See, for example, GoursatHedrick,
Mathematical Analysis, Chap. IX.
FOURIEK
series
S
SERIES
293
2 kir as TT, and accordingly has the same value for Xi Hence the limit curve is a series of repetitions of the part between 2 &TT, 0). TT and x = TT, and the isolated points ( x = It should be noted that the function denned by the series has finite discon tinuities, while the function from which the series is derived is continuous.
has the period 2
for Xi.
not necessary that/(#) should be defined by the same law TT to TT. In this case the integrals throughout the interval from
It is
defining the coefficients break up into shown in the following examples:
Ex. 2. Find the Fourier s and TT, where f(x) = x + ?r if
two
or
more
integrals, as
series for f(x) for all values of
TT
&lt;
x between
TT
x
&lt;
0,
and /(x)
x)
= TT
\
xifO&lt;x&lt;7r.
Here
a
= =
2 irn*
;
f C
(1 v
(x
+ +
TT)
dx
+ f
V+ C
dx
= TT
x) cos
On
(x
TT)
cos nx dx
(IT
nx dx\
=
bn
cos mr)
;
=
y
(x
+
TT)
sin
nx dx
+
C\TT
 x) sin nx dx
A
O
A
FIG. 107
Therefore the required series
TT
is
9
+ ^l~l2~ +
4 /cos x
cos 3 x
Q2
+
cos 5 x
r,2
+
\
")
The graph
(fig.
of /(x) for values of
x between
2
TT
and
IT is
the broken line
ABC
K,
f or
72
o4 )* ! o limit curve of the series coincides with the
111 +^+
107).
When
x
=
0,
the series reduces to
Tj2
*
"^+
"~^
When
x
=
it is
TT,
( 32 52 7T\12 the series reduces to 0.
+
+ ^ +
+
.
.
\=
y
Plence the
broken
line
1,
ABC
at all points.
From
is
the periodicity of the series
fig.
seen, as in Ex.
that the limit curve
the broken line of
107.
*
Byerly, Fourier s Series,
p. 40.
294
Ex.
3.
INFINITE SERIES
Find the Fourier s series for /(x), for where f(x) = 0if?r&lt;x&lt;0, and f(x) =
a
an
1 =7T
all
TT
values of x between
if
&lt;
TT
and
TT,
x
&lt;
ir.
Here
/
r
I
dx
+
rn
I
TT
dx )
/
\
=
TT
;
\JTT
I
i/O
=
TT
cos nx dx
=
;
1 1 = /**IT sin nx dte =  (1
v
cos
r).
Therefore the required series
TT
is
+
/sinx
2
sin3x "~
sinSx
~
\
of x
The graph of from x =
the function for the values of x between
TT
TT
and
TT is
the axis
to
x
=
0,
and the straight
0.
line
AB
(fig.
108), there being a
finite
discontinuity
when x =
J
FIG. 108
The curves
(1), (2), (3),
and
(4)
are the approximation curves corresponding
respectively to the equations
^
=+
2 sin x,
{sin
(2)
x
hr
2
\
1
sin 3 x\
TT
sin 3
(3)
x
sin 5 x\
3
5
in
I,
1,
/
149, Ex.
5.
They may be
readily constructed
by
the
method used
It
is
to be noted that all the curves pass through the point fO,
which
is
midway
between the points
and 0, which correspond to the finite discontinuity, and that the successive curves approach perpendicularity to the axis of x at that point.
A
INDETERMINATE FORMS
161.
295
The indeterminate form 
Consider the fraction
/M,
4&gt;(
x)
a)
and
let
a be a number such that f(a)
=
and
cf&gt;
(a)
=
0.
If
we
place x
=a
in (1),
we
obtain the expression , which
is literally
meaningless.
when x=a,
n1
customary, however, to define the value of the fraction (1) as the limit approached by the fraction as x approaches a. In some cases this limit can be found by elementary methods.
It is
T2
.
Ex.
1.
a
x
a, this x,
When x =
fraction
by a
becomes and have
When
2
x
^
a,
we may
divide both terms of the
_
x
2
"
a
for all values of x except x
=
a.
Equation
(1)
is
true as x approaches a,
and hence
Ex.
2.
x
When
x
=
1
0, this
becomes

.
When
2
x
1 1
^
0,
we have
a;2
Vlx
Lim
2
lVlz
Vl
x2
.
+ Vl + Vl  X 2
x
1+ VlX 2
1
= Lim
=
Ex.3.
sn 2
x = 0, this becomes Maclaurin s series, and have
When
When
(
x
2
^
0,
3
we may expand each term by
sinx
II
iice
Lim
x = o sinx
=
1.
296
INFINITE SERIES
into series, used in Ex. 3, suggests a to determine the limit approached
will place
The method by expansion
general method.
Since
we wish
by
f&L
$(x)
a s x approaches a,
we
x
=a+h
and expand by
Taylor
s series.
We
+h
have
(
31, (5))
f(a)
x
a
+ hf (a) + /() + ,^/
[g
"(a)
+
.
.
.
II
But by hypothesis f(a)
/(a)
_
^^
Now
as
a;
__ _ _+/+
&lt;/&gt;()
=
and
= 0.
Therefore, since A
=
0,
/
.
(&gt;+
,/"(*)
[
[
(2)
&lt;#&gt;
() +
+
*"
^
,
&lt;)+

= a,
(2)
A
===
0,
and hence, unless
/ (a) =

and
&lt;f&gt;
(a)
=
0,
we have from
Lim
x
f(x} 9&gt;()
= f (a)
f
=a
(a)
(o)
cp
(a)
If,
however,
/
(a)
=
and
&lt;f&gt;
=
0,
the righthand side of (3)
becomes 
In this case
(2)
becomes
whence
unless
/"(a)
=
and
=
&lt;"()
0.
In the latter case we
may go back
again to (2)
and repeat the reasoning. Accordingly we have the rule
:
To find
x
the value the
 when of a fraction which takes the form
the
= a, replace
and
numerator and
x
denominator each ly
is also
its
deriv
ative
substitute
= a.
If
the
new fraction
&gt;
repeat
the process.
INDETERMINATE FORMS
pX
297
= 0.
Ex.
4.
To
g
X
find the limit
pX
approached by
p
when x
2 = =
1
sinx
X~\
X
By
Ex.
If
the rule,
Urnx=o
=
f~f&gt;X
_1_
p
sinx
\
2.
L
cosx
Jz=o
5.
To
find the limit
approached by
when x =
xsinx
e~ x ~\
0.
we
apply the rule once,
we have
Lim
x=o
e*
2 cos x
+
e~ x
Ve*
+
2 sin x
=

.
xsinx
L
:
sinx
+
x cosx
Jx=o
We
therefore apply the rule again, thus
Lim
x
6^
2 cos x
+
e~ x
Ver
=o
xsinx
= 2
L
+
2 cos x
f
e~ x ~\
Ja:=o
4 = = 2. 2
cosx
xsinx
\
162.
of the expansion of f(x)
ever,
161 assumes the possibility power series. It is, how sometimes necessary to consider functions for which this
of the rule in
The derivation
and
into
&lt;f&gt;(x)
assumption
is
not valid.
We
shall accordingly give the following
new
proof of the rule.
(3),
Consider formula
30, namely,
From this
where f
is
it follows that if F(b) = and F(a) = 0, then F some number between a and b. Let us apply
(^)
= 0,
this to
the arbitrarily formed function
where F(b)
tion.
=
and F(a)
=
0, as
may
be seen by direct substitu
Then
whe nce
Now,
in (1), let/(n)
= 0,
&lt;#&gt;()
= 0,
+ h)
and place
6
= a + h. We
have
f(a
/ ()
298
INFINITE SEKIES
As h
=
0,
f
= a,
since
&lt;&lt;
+ &.
Hence
which may be written
If
Lim
^ = Lim
L
f
*
v

(3)
f (a)
and
&lt;j&gt;
(a) are
not both zero,
(3) gives
Lim 4^f
however,
f(x)
f (a) = :L^L

If,
f
(a)
=
and
&lt;ft(a)
0, (3)
gives
Lim fix)
The
163.
rule of
= Lim fix)
results.
f.
fix)
161 thus
The indeterminate form
If
/(a)
=00 and
=
&lt;(a)
oo,
the fraction ii2 takes the meaningless form
fraction
is
.
The value
of the
then defined as the limit approached by the fraction as
limit.
x approaches a as a
We
shall
now
prove that the rule for
 holds also for a finding the value of a fraction which becomes which becomes ~fraction
To prove
this,
we
shall take first the case in
a?
which the
fraction
becomes
when
= 00. By placing
b = x, a = c
in (1),
162,
we have
which
is
/()/() _ / (?)
twwrw)
~
/.... ,.
v
equivalent to
/
Now we assume that
,
^
approaches a limit ^ as x
f (} and therefore ^y^
co.
Hence
from
we may
take
c
so great that
f (c) ^^
&lt;/&gt;
&gt;
differs
(6 )
INDETERMINATE FORMS
k by as
little as
299
take x so
we
c
please.
This fixes
c
;
we may then
each
c
much
by as
greater than
little as
that
 and
4&gt;(x)
~ may
f(x)
differ
from zero
then,
we
please.
By
\
proper choice of
and
x,
we
have, from
(1),
f/ ff
where
e1
and
e2
may be
as small as
we
please.
Hence
To extend
this result to the case in
a;
which
4rr
4&gt;(*)
becomes
^
when
= a,
WQ
place
= a Hy
1
Then
Then
becomes
g
when y
=
oo.
Therefore, from (2),
Lim
r ..
^=
4&gt;(7/)
Liin
r
..
&(
1
Now
lUld
J
p"(
2/
)=/
W g = i
=* (#)
= Lim
.*.
&lt;#
dv
*
* (y)
=
&lt;/&gt;
=
and Li m
r
Hence
,

Hence we have
(*)
()
(3)
From
(2)
and
(3) follows the proposition that
^
we wished
infinite.
to prove.
Ex. To find the limit approached by
1
as x
becomes
xn
T*
By
the rule,
x =*
Lim
iQff n
O*
= Lim
x=
x
xn

nx n ~ l
= Lim
=
0.
a=*nx
300
INFINITE SERIES
There are other indeterminate
164. Other indeterminate forms.
forms indicated by the symbols
Ooo,
oooo,
0,
00,
r.
&lt;f&gt;
The form and f(a) =
&lt;
oo arises
&lt;f&gt;(a)
(x),
= oo. we have f(a) = oo,
may
when, in a pro duct f(x) (x), we have oo arises when, in /(#) The form oo
&lt;f&gt;
(a)
oo.
/(#)&lt;/&gt;
These forms are handled by expressing
as the case
be, in the
(x)
OT/(X)
$(x),
or
H
when x
Ex.
1.
= a.
.
The
rule of
form of a fraction which becomes 161 may then be applied.
x 3 e*2
x
oo,
When
becomes
this
becomes
oo.
co
0.
We
z xZ have, however, x e~
=
Xs
,
which
when x =
Then
Lim
xs
Lim
3x2
3x = Lim  = Lim
3
=
0.
In the same manner
for
Lim x n e~ xZ =
X=
00
any value
Ex.
2.
of n.
secx
tanx.
this is oo
oo.
When
x
=
We
have, however,
sec x
tan x
=
1
sin
x
,
cosx
which becomes  when x
TT =
Then
2
Lim
;
(sec
x
tan x)
= Lim
c
_._
1
sin
x
= Lim
^_ ~
rr
 =
cos x
sin
_
0.
^_ ~ TT
TT
cos x
x
2
~2
2
The forms 0,
00
00, I
may
arise for the function
when x = a. If we place
Wf
M
log
=
[f(x)]* \
(x
we have
If
u
= $(x) logf(x).
by the previous methods,
found.
Lim
x=a
&lt;/&gt;(^)
log/(a;) can be obtained
the limit approached by
u can be
PROBLEMS
i
301
Ex.
3.
(1

x}*.
When
x
=
0, this
becomes
1*.
Place
then
Now, by
Hence
log
161,
u approaches the limit
1
and u approaches the
limit 
PROBLEMS
1.
Prove that the series
1
i 1 i L 2a2^4^4448
l l
,
where there are two terms of the form
terms of the form
l
,
four terms of the form
2
"
and 2^ terms
of the
form
1
,
4a
,
eight
when a
2.
&gt;
1.
(
2
a
= 1, (k
2, 3,
),
converges
)
By comparison
152),
with the series in Ex.
1
or with the harmonic series
(Ex. 2,
prove that the series
converges
when a
&gt;
1,
and diverges when a^l.
Test the following series for convergence or divergence
23
25
22
"
1
1
1
1
6
12
34
6.6
(2nl)2n
S
12
2.3
34
nn+
302
INFINITE SERIES
10.
*+++*+.
11.
. 1.23
2.3.43.4.5

n(n
+
1)
(
+
2)
Find the region of convergence of each of the following
series
:
14., +
15.
1
^+l
2 2a;
+
+S
2
3
24
5
+
+
2.4.6...(2n2)
2nl +
"
21. log(x
+ Vl +
x*)
=
x

X
.
I 2
lo
+ ll?
2
2
.
4
5
.
22.
a
^ =
7 bx
:
4
6
1
(2
n

2)
2nl
.
1
a
+ Ax + 2
a
^
a3
+
...
+
^,a"
+
....
Expand each
four terms
23. sees. 24.
er
of the following functions into a
Maclaurin
s series,
obtaining
1
*.
25. log(l
26. e sina;
.
+
sinx).
27.
e 8111
secx.
28. log cos
x.
Expand each
x between
29. x 2
.
of the following functions into a Fourier s series for values of
TT
:
TT
and
30.
e.
=
TT if
&lt;
31. /(x), where /(x)
32. /(x), where /(x)
33. /(x), where /(x)
34. /(x), where /(x)
=  TT if  TT = x if TT = 7rif = Oif
&lt;
x x
&lt;
0, 0,
and/(x)
x x x
&lt;
&lt;
TT.
&lt;
&lt;
and /(x)
and/(x)
if
&lt;
&lt;
TT.
7r&lt;x&lt;0,
=
x2
x
if
if
&lt;
&lt;
TT.
7r&lt;x&lt;0,
and /(x)

&lt;
x
TT.
PROBLEMS
approaches the given value
:
303
variable Find the limit approached by each of the following functions, as the
cosx
cos a
66.
cscx
CHAPTEE XVI
COMPLEX NUMBERS
165.
of the
Graphical representation.
A
complex number
real
is
a
number
defined
form x +
iy,
i*
by the equation
a complex
=
where x and y are
1 (I, 12).
z,
numbers and
i is
number by
the letter
z
In this chapter thus
we
shall denote
= x + iy.
;
is the real part and the number iy is the imagi nary part of z. When y = 0, z becomes a real number and when x When both x = and 0, z becomes a pure imaginary number. and z = when, and only when, a; = and y 0. y = 0, then 2 = To obtain a graphical representation of
;
The number x
o
FlG 109
complex number, construct two axes of OX and OY (fig. 109), take any point P, and draw OP. Then the complex number x + *y is said to be represented
coordinates
either
a
* by the point P or by the vector OP. For if a complex number, z x\ iy, is known, x and ?/ are known, and there corresponds one and only one point P and one and only one vector OP. Conversely, to a
point
of
x and
this
real
OP corresponds one and only one pair of values and therefore one and only one complex number. In y, connection the axis of x is called the axis of reals, since numbers are represented by points upon it and the axis of
a vector
;
P or
y
is
called the axis of imaginaries, since pure imaginary
numbers
system
are represented
of
by points upon it. If we take as the origin and OX as the initial polar coordinates, we have x = r cos 0, y = r sin
z
line of a
6,
and therefore
=
x\iy
= r (cos + i sin 0).
304
*A
vector
is
a straight line fixed in length and direction hut not necessarily in
position.
ADDITION AND SUBTRACTION
The number
r
105
is
= ^/x + y\
2
which
is
always taken positive,
is
called the modulus of the complex
number and
equal to the
length of the vector OP.
The angle 6
=
1
tan"

is
called the
argu
made ment, or amplitude, of the complex number, and is the angle added to with by the vector OP. Any multiple of 2 TT may be
OX
the argument without altering the complex number, since
r [cos (6 f 2 for)
+ i sin (6 +
2 for)]
= r (cos + i sin
7
fife
(9).
The modulus is also called the number and is denoted by \z thus
,
absolute value of the
:
complex
166. Addition
iy^
and subtraction.
If
z^_
=x
l
and
*i
22
=
2
a?
2
+ iy^
( ^i
t
then,
by
(2/i
definition,
+* =
110
+
x z)
+*
+
2/2)
FIG. 110
In
fig.
let I{
and
I* represent the
two complex numbers,
vectors.
and
let
0%
and OP^ be the corresponding
Construct the
has the 3 parallelogram OJ(P^. Then it is easy to see that x 2 y l + y 2 ) and therefore represents the complex coordinates (x l
P
+
,
number z^+z y The
addition of complex
numbers
is
thus seen to
be analogous to the composition of forces or
velocities.
Since 0/f=^,
(*!&gt;
OP =
2
z2
,
and
OP =\z + z
3
l
2
\,
*i+*,N*il+K
the equality sign holding only OP have the same direction. Z
when
and
To subtract z2 from 2 P we first change the sign of z 2 and add the result to z r Graphically,
Fi.
Ill
the change of sign of z 2 corresponds to replacing
Ill) by 1J, symmetrical to 2 with respect to 0, or to turning the vector OP^ through an angle of 180. The corre parallelogram OPJIP^ is then completed, giving the point &
2
P
(fig.
P
P
sponding to
zl
z2 .
306
COMPLEX NUMBERS
and
division.
If
z^
167. Multiplication
then,
= x^+ iy^ and z = x + iy^
2 2
by
definition,
If
we
2
~r ^2
use polar coordinates,
*
we have
zl
=r
l
(cos
X
+
i
cos ^2 (
+ sm ^2)* an(^ = r^ [(cos cos sin sin 2^2 + * (sin cos + cos sin = r,r [cos (^ + + i sin (0 + ^
2 X
sin 0J,
2
l9
1
2)
X
2
X
2 )]
2
6&gt;
2)
l
2 )].
Hence, in the multiplication of two complex numbers, the
uli are multiplied together
ally,
mod
and the arguments are added. Graphic the vector corresponding to a product is found by rotating
the vector of the multiplicand in the positive direction through an angle equal to the argument of the multiplier, and multiplying the length of the vector of the multiplicand by the modulus of the multiplier. In particular, the multiplication by i is represented
by
and the multiplication rotating a vector through an angle of 90 1 is represented a vector through an angle of by by rotating
;
180,
as noted in
166.
is
The quotient
Hence,
if z
1
^
Zo
a
l
number which multiplied by
X)
z
gives 2
2 ),
= r^cos O + i sin
and
z2
=r
z
(cos
2
+ i sin
Graphically, the vector corresponding to a quotient is found by rotating the vector of the dividend in the negative direction through an angle equal to the argument of the divisor, and dividing the length of the vector of the dividend by the modulus of the divisor.
168. Involution
positive integer,
itself.
is
and evolution.
The value
of z
n
,
where n
is
a
obtained by successive multiplication of 2 by Therefore, iiz = r (cos 6 + i sin 0),
z
i
n
=r
is
n
(cos
nO
+ i sin n6).
is
(1)
The
root 2,
where n
a positive integer,
2.
raised to the nth
power gives
number which we have, at first Accordingly
a
INVOLUTION AND EVOLUTION
_
307
165) that
skrht, z
z
e\ = r cos  + i sin 
But
if
we remember
where k
is
(S
= r [cos (0 + 2 &TT) + i sin (0 + 2 &TT)],
we have
2
zero or an inte
ger,
also
1
= rco8((6
if"
I
2 kir\ /0 ift8in
. .
2
I
forY]
\n
n f
\n
i
n yj
by giving
z
n
(2)
There are here n distinct values
the
of z, obtained
to
k
i
n values
0, 1, 2,
,
(n
1).
Since the equation
=c
has only
n roots (I, means the
In this result r 42), (2) gives all the values of z. numerical root of the number r, such as may positive
be found by use of a table of logarithms. By a combination of (1) and (2), we have
= r P\ cos(
L

+^
/
1).
/p0 + i sm[* +
2/hwr

\f
(3)
?
where ^
=
0, 1, 2,
,
(//
1
cos
Finally,
z z
7.n [
+ i sin
m6)
C os(
+ i sin(
w0)].
The nth
roots of unity can be
found by placing r
h
*
=
1
and
=
= cos
2 far
sm
tices of a regular
The points which represent these roots graphically are the ver polygon of n sides, inscribed in a circle with and radius unity, the first vertex lying on the axis of center at
Similarly,
T*pO 1 Q
V/ 1 = cos (2 + l)7T + i sm (2/
n
l
+l)7T n
169. Exponential and trigonometric functions. The exponential and the trigonometric functions are defined in elementary work in a manner which assumes that the variable is real. For example,
the definition of sin x requires the construction of a real triangle with an angle equal to x, and the definition of ex involves actual
308
COMPLEX NUMBERS
In order to be able to regard the inde
2
involution and evolution.
pendent variable as a complex number, we adopt the following
definitions
:
Z
3
2
6
When
When
place z
z is real,
each of these functions
is
the corresponding real
31.
this,
function, since these series are those found in
z is
=
/
complex, each of the series converges. To show r(cos 6 4 i sin 0) in (1), for example. We have
r cos 6
(I
\
+
+
r
.
2
cos 2
.
+
r
.
3
cos 3 6
2
\
r
.
If
12
+ i (r sin + nr sin 2 +
V
If
/
r
3
sin 3
+
l
Each term of the two series in parentheses is less in absolute value than a corresponding term of the known convergent series
1
+f+
f
,
+
r
,
s
H
.
Therefore each of these series converges
(
153) and hence (1) converges.
In the same manner
It follows that
it
may
be
shown
that (2) and (3) converge.
each of the
series (1), (2),
and
(3)
may
It is necessary,
however, to give
properly be used to define a function. new proofs of the properties of
the functions, based upon the
new
definitions.
From
which
(1)
we have
fundamental properties
of
are the
the exponential function.
From
(1), also, if
we
replace z
by
iz,
we have
z 1
[2
[3
[4
EXPONENTIAL AND TRIGONOMETRIC FUNCTIONS 309
From
this,
with reference to
e
iz
(2)
and
(3),
we
2,
obtain the formula
(4)
=
cos z
+ i sin
which establishes a
relation
between the exponential and the
trigonometric functions. By changing the sign of z in (4)
e~
lz
we have
2,
whence
= cos 2 i sin ~ g2 _ e sin z =
2*
iz
:
,
i
(5)
cos z
iz e = e +iz
(6)
From (4), (5), and (6) it is easily shown that the trigonometric formulas obtained in elementary work for real variables are true for complex variables also (see Exs. 1 and 2).
We
and
shall use these relations to separate the functions (1), (2), (3) into real and imaginary parts, as follows
:
e
* + iV
_ e x e _ ^ cog y
iy
iy)
=
A
i
ix
~y
(e
=
e
y
v
[e~
(cos
x
+
i
sin x)
e
v
(cos
x
i
sin x)]
+e~ y 
sm x +
f
.
.e y
i

e~ y
cos
x
= cosh y sin x
Similarly,
i
sinh
.e
i
v
y
cos x.
cos (x iy) s(*Hy)
,
+
,
.
v
y v = e + e~ cos x
e~ v
.
sin
x
2
= cosh y cos x
If
i
sinh y sin
x.
x
0,
we have
sm
.
.
.e
i
v
e~ y
ly
= i sinh y,
. .
,
cos xy
=
2
= cosh y.
310
Ex.
1.
COMPLEX NUMBERS
Prove
(5)
sin 2z
(6),
+
cos 2 z
=
1.
From
and
sin*.
+
cos*.
=
=
sin z\ cos z 2
Ex.
2.
Prove sin (zi
(5),
+
z2)
+
cos z\ sin z 2
.
From
_
e
=
2

2 ?2 2 [e* ie
e~ iz ie~ iz z]
i
= =
170.
2
_ [(cos z\ +
i
i
sin z\) (cos z 2
i
+
i
sin z 2 )
i
(cos Zi
sin z\ cos z 2
sin Zi) (cos z 2

sin
2)]
+
cos z\ sin z 2
The logarithmic
function.
If 2
z.
= ew
,
then, by definition,
w = log
The
properties of the logarithmic function,
namely
log 2
are deduced
n
= w log
of a
2,
log
1=0,
log
=
co,
from the
us place
z
definition, as in the case of real variables.
The logarithm
ber.
complex number
is itself
a complex
num
For, let
= x + iy
*=
r (cos 9
+
i
sin 0)
= re
ie
.
Then
log z
= log r + log
log r
+ *   log (a? + /) + * tan2i
1
y *&

log r is the logarithm of the positive 154 be found as in I,
Here
number
r,
and may
if
We may now find the logarithm of a real negative number. For, _ a s such a number, we may write a = a (cos TT + i sin TT) = e whence = log a + log ( a)
l7r
i
,
r.
In particular,
log
(
1)
= *V.
ANALYTIC FUNCTIONS
It is to
311
a logarithm
be noticed that, in the domain of the complex numbers, is not a unique quantity. For
where k
is
zero or an integer.
Therefore
log z
=w+
:
2
kiir.
We may
infinite
express this as follows
The exponential function
is
a
periodic function with the period
2 ITT, and
the logarithm
has an
number of
values, differing by multiple* of
2 ITT.
171. Functions of a complex variable in general. have seen that functions of a complex variable obtained by operating on x 4 iy with the fundamental operations of algebra, or involving the elementary transcendental functions, are themselves complex
We
numbers of the form u + iv where u and v x and y. Let us now assume the expression w what conditions it must satisfy in order that of z = x \ iy.
t
are real functions of
= u + iv, and inquire
it
may
be a function
In the
of the
first place, it is
"
to be noticed that in the broadest sense 20),
word
function"
(I,
w
is
always a function
of
z,
since
x and y are determined and therefore u and v are determined. But this definition is too broad for our present purpose, and we shall restrict it by demanding that the function shall have
z is given,
when
a definite derivative for a definite value of
restriction is seen as follows
z,
:
z.
The
force of this
we may
respectively and obtain
the point
ical
&z = Ax + iky. The direction in which which corresponds to z + kz in the graph Q (fig. 112), representation, lies from P, which corresponds to z, depends
l\x
In order to obtain an increment of at pleasure increments A# and assign Ay to x and y
then on the ratio
 which may have any value whatever.
,
Cor
responding to a given increment Az,
w
takes an increment
Aw,
where
(110)
312
Dividing by Az Ay = 0, we have
COMPLEX NUMBERS
= Ax
+
iky and taking the limit
/
&gt;dv
as
Ax =
and
3u
du
dy
7
mdv \dy
Lrm
_

dy/ dx
*
V
/
Unless special conditions are imposed upon u and v the exprest
nil
sion on the righthand side of the above equation involves
.
&gt;
and
ctsc
the value of
Lim
depends upon the direction in which the
point
Q
(fig.
112) approaches the point P.
(1),
Now
the value of
,
the righthand side of
when
j
= =
0, is

+i
.
V
^
.
and
its
value
du
M dv
:
when
^^

oo is
2
&
Equating these
two values, we have
"
A
=0.
(2)
This, then,
is
the necessary condition
7
that
7
Lim AZ =
 should
It is also
all
A0
be the same for the two values / CtOG
the sufficient condition that
7
=
.
and ^ 6t^C
=
oo.
Lim
A =
"
A*
should be the same for
d?/
values of
^&gt;
for
if
from
dx
(1) is simplified
by aid
of (2), f disappears ^x
it.
Now
(2) is
equivalent to the two conditions
du _2_v_ dx dy dv du _
dy
.
(3)
dx
the equations (3) are the necessary and sufficient condi iv should have a derivative with respect tions that the function u
Hence
x
+
to
+ iy
which depends upon
the value
satisfies
of x
+ iy
only.
is
A
function u
+
iv
which
conditions (3)
called
an
analytic function
of
x
+ iy.
CONJUGATE FUNCTIONS
313
172. Conjugate functions. Two real functions u and v, which 171, are called conjugate functions. By satisfy conditions (3), 171, with respect to x, differentiating the first equation of (3),
the second with respect to
y,
and adding the
results,
we have
Also, by differentiating the first equation of (3), respect to y, the second with respect to x and
t
171, with
taking
the
difference of the results,
we have
Bx*
df
is
That
is,
each
of a pair of conjugate functions
a solu
tion of the differential equation
^+
,
*
2&lt;
a?
^~
_
n
Let us construct
now
is
the two families
a point of intersection of two of these y^ curves, one from each family, the equations of the tangent lines at fo, yj are (115, Ex.1)
2
.
and v
=c
(
173) of curves
u
=c
l
If (x v
But from
(3),
171,
intersect at right angles; that is, every curve of one family intersects every curve of the other family at right angles. We express this by saying that the families of curves corresponding to two conjugate functions form an orthogonal system.
Hence the two curves
314
Examples
curves
COMPLEX NUMBERS
may
of conjugate functions and of orthogonal systems of be found by taking the real and the imaginary parts of any function of a com
plex variable.
Ex.
1.
(x
+
2
= x 2  y 2 + 2 ixy.
iy)
From
x2
x2
this it follows that
2 y and 2 xy are conjugate functions, and that the curves 2 Ci and xy = c 2 form y an orthogonal system (fig. 113). In fact, the system consists of two families of rectangular hy perbolas, the one having the
coordinate axes for axes of the
curves and the other having the
coordinate axes for asymptotes.
FIG. 113
Ex.
2.
log(x
2
=
logVz
+
y
2
+ iy) + tani
i
.
Hence
log
tan 1  are conjugate funcX 2 2 tions, and the curves x + y Ci and y = c^x form an
orthogonal system (fig. 114). In fact, one family of curves
consists of circles
with their
centers at the origin, and the other consists of straight lines
through the
origin.
FIG. 114
PROBLEMS
Find the sums of the following pairs of complex numbers and the difference obtained by subtracting the second from the first, and express the results
graphically
1.
:
3
+
2i, 4
+
4
5i.
3.
4.
+ 8i
010 8+
i,
3+21
i,
12
6
+
9i.
PROBLEMS
of the first
:
315
Find the products of the following pairs of complex numbers and the quotient by the second, and express the results graphically
TT
5.
cos

+
.
7T
7T
i
sin
i cos
4
7T
6
6.
Msm 4
. .
7T
7.
8.
co8+tn ,
g
7T
.
27T
coB^
+ .Bin.
27T
9.
^
+ 2 + 1 + x +
1
I
\/3,
i
V3 +
i.
L
._
i
2
i,
.
V.3, 3
3
V2.
?
x
1+ _
.
Find the following powers and express the results graphically
11. (2
:
+
3
1)2.
12.
(1
 V3) 2
i
.
13. (1
+
i)3.
14. (1
+
i)
4
Find the following roots and locate them graphically
15. ^1. 16.
&lt;CTI.
:
21.
22.
20. ^8. V^S. 1S+V^1. cos^i cosz 2 sinz\ sinz2Prove, for complex numbers, CQS(ZI + z 2 ) for complex numbers, sin z\ + sin2 2 = 2sin^(zi+ z 2 cosl(zi z2 Prove,
17. ^T. 19.
)
)
23. Prove sinh iy
=
i
sin y.
24. Prove cosh iy
=
cos y.
25. log(2).
26. log(l
+
i).
27.
log(l
+
i).
28. log
i.
Find the orthogonal systems of curves defined by the
parts of the following functions
:
real
and the imaginary
29.1.
z
30. log &
i^l z a
+
31.
log(a*).
real
:
Find the orthogonal systems of curves defined by the
parts of the following functions, using polar coordinates 32. z
.
and the imaginary
33.
V.
CHAPTEE XVII
DIFFERENTIAL EQUATIONS OF THE FIRST ORDER
173. Introduction.
Consider the equation
/(s,y,c)0,
in
is
(1)
which
c is
an arbitrary constant.
;
If c is
the equation of a certain &lt;curve and if c values, the totality of the curves thus represented by (1)
a
given a fixed value, is supposed to take
is
(1)
all
called
family of curves. To determine the curves
of the family
fixed point
(1),
P(x v
y^),
we may
substitute x 1
which pass through any and y l for x and y in
thereby forming the equation
f(x v y v c)=Q.
(2)
The number of roots of (2), regarded as an equation in c, is the number of curves of family (1) which pass through P and their v
equations are found by substituting these values of c in (1). The direction of any curve of family (1) is given by the equa
tion(115,(l))
dy
8/ ex
3
&lt;
&gt;
~&x=~Jf
Yy
which, in general, involves c. In general, however, we may elimi nate c from (1) and (3), the result being an equation of the form
ft 1J
If
we
substitute the coordinates of P^ in (4), the values of /ctx
which
satisfy the resulting equation are the slopes of the respec tive curves of (1) which pass through r Hence (4) defines the
P
same family
of curves that is defined
by
(1),
but by means of
the directions of the curves instead of the explicit equations of
316
INTRODUCTION
the curves themselves.
tion of the
317
equa
Hence
(4) is called the differential
(1).
We
family of curves represented ~by have seen how an equation of type
(1) leads to
a differential
Conversely, to an equation of form (4) there a family of curves which may be represented always corresponds by an equation of form (1). For if the coordinates of a point I[
equation of type (4).
(fig.
directions through
115) are assigned to x and y in (4), (4) determines one or more Jf. Following one of these directions, we may
to a second point
move
in (4), a direction is
the coordinates of P^ are substituted determined in which we may move to a third
ZJ.
If
point y Proceeding in this way, we trace a broken line such that the coordinates of every vertex and the direction of the following segment at that point sat
isfy (4).
P
The
limit of this
broken
each
line, as the length of
segment approaches
is
the limit zero,
a curve
such that the coordinates
of
each point and the direc
tion of the curve at that
^ ]G
point satisfy (4). Since in this construction JJ may be any point of the plane, there is evi dently a family of curves represented by (4), as we set out to
prove.
The constant
c
in the equation
of
the family
may
be
taken, for example, as the ordinate of the point in which a curve x r Hence of the family cuts the axis of y or any other line x
=
every differential equation of
form
(4)
has a solution of form
(1).
The problem
solution (1)
is,
of proceeding
from a
differential equation (4) to its
however, a
difficult one,
the simpler cases.
Some
of these cases
which can be solved only in have been discussed in this
first
volume for equations in which These are the following
:
j
appears to the
power
only.
I.
Variables separable.
(
77)
la.
Homogeneous
Equation
equation.
(78)
2
Ib.
of the
form x
(ap
+ 1$ + cj dx + (a
+ bjj + r,) dy =
0.
(
79)
318
II.
DIFFERENTIAL EQUATIONS OF FIRST ORDER
The
linear equation
s
(80)
(
Ila.
Bernouilli
equation.
(
III.
The exact
equation.
81) 147\
Ilia.
Solution by integrating factors.
(
149\
serviceable, the student should appreciate that they may all fail with a given differential equation, since the above list does not contain all possible differen
tial
While the above methods often prove
equations which are of the
first
solution will
ax in general involve integrations of which the
^ degree in
Moreover, the
results
can be expressed as elementary functions only in the simpler cases. 174. Solution by series. The solution of a differential equation may usually be expressed in the form of a power series.
EX.
^=X
dx
i
2
+
y2.
Assume y = a + a& + we have
a2x2
\
+
a3z 3
\
.
Substituting in the given equation,
ix
+
2 a2 x
+ 3 a3
x2
= x 2 + (a +
+
a 2 x2
+ a 3 x3 +
x.
2
)
,
which the coefficients of like powers of x on the two must be equal, since the equation is true for all values of
in
sides of the equation
Equating
coefficients,
we have
ai
=
a2
2
,
2a 2 =
3 a3
aoi,
af
=I+ = =
2
,
+
whence
a:
2
a
8
,
Hence the required solution
y
is
=
a
+
a 2x
+
a 3x 2
+
^ (1
+
3 a 4 x3 )
+
.
175.
Equations not of the
first
degree in the derivative. If the
differential equation of the first order is of higher degree
first
than the
in
&gt;
new methods
make
1.
of solution are necessary.
:
Denoting
by p, we
shall
three cases
2.
3.
Equations solvable for p. Equations solvable for y.
Equations solvable
for x.
EQUATIONS SOLVABLE FOK
1&gt;
319
176. Equations solvable for p. Let the given equation be an equation of the nth degree in p, and let the roots of the equation, where p v , p n regarded as an algebraic equation in p, be p v p 2
, ,
P&gt;
tion
P n are functions of x and y, may be written in the form (I,
&gt;
or constants.
Then the equa
42)
(pPi) (PP
zero,
Z)
"(P
~Pn)
=
0.
(1)
But (1) is satisfied when and only when one of the n factors is and hence the solution of (1) is made to depend upon the solutions of the n equations
Let the solution
of
p
p^
=
be f^(x
t
y,
c
cj
v
c
=
2,
0,
the solution of
are arbitrary
p
p =
2
be
f
z
(x, y, c 2
)=
0, etc.,
where
constants.
Since each of these constants
is
arbitrary, however,
and
no necessity of distinguishing among them, we them all by the same letter c. Now form the equation
there
is
A(*&gt;
will denote
c
y&gt;
) M*&gt;
c
y&gt;
)
/&gt;&gt;
y,
=
&lt;0
o.
(3)
of x and y which make any factor of (3) zero satisfy by making the corresponding factor of its lefthand member zero. Hence (3) is a solution of (1), since the values of x and which satisfy (3) are all the values of x and y necessary to y and since (3) contains an arbitrary constant, it is the satisfy (1)
The values
(1)
;
general solution.
Ex.
1.
_p3
_
2 p2
+
y^\ Uy _ x _ z p _ 2 (2 y  x2  z \ = 0.
i
\
2
/
\
2
/
Solving this equation for p,
we have
,
p=
The
2,
y p = JO  x
and
y p =  %, +
x.
solution of the first equation
is
evidently y
is
2zfc =
0.
The second
equation,
lion,
when
its
written in the form
dx
solution
is
 y =  x,
x
in the
seen to be a linear equa
+ x 2 ex = 0. The third equation may also be written its solution is z 3 3 xy + c = 0.
and
y
form of a linear equation, and
Hence the
solution of the original equation
(y
is

2x
+
c) (y f
x2

ex) (z
3

3 xy
+
=
&lt;)
0.
320
Ex.
DIFFERENTIAL EQUATIONS OF FIEST OKDEE
_
2.
/dy\
y[ ) \dx/
2
f
2x dy
n
dx
y
0.
Solving for p,
we have
p=
The
is
equationits
a homogeneous equation, and The equation
solution
is
Vz 2 +
2
2
y"
x
+
c
=
Q.
x
Vx* + y
/
~y~ ~^
is
also a
homogeneous equation, and its solution is Vsc 2 Hence the required solution is y 2 + 2 ex c 2 = 0.
177. Equations solvable for y.
+
y2
+
x
c
=
0.
Let the given equation be
u&lt;
j
\
&gt;
y P)
&gt;
(J)
Solving this equation for y,
we have
the equation
(2)
y=F(x,p).
Differentiating (2) with respect to x,
and replacing ^ by p, we
have the equation
in
which
^&gt;
and
a;
are the variables.
Let the solution of
(3)
be
(4)
}
^(x,p,c)=Q.
^&gt;
between (1) and (4), we have a function of #, y Eliminating and an arbitrary constant which is, in general, a solution of (1). But the process of solution may bring in extraneous factors or
otherwise lead to error, and the solution should be tested by
substitution in
(1).
If the elimination
may
(I,
cannot be performed, equations (1) and (4) be taken simultaneously as the parametric representation
163) of the solution,
Ex.
1.
p
y
being the variable parameter.
xp
2

2 yp
+
ax
=
0.
Solving for y,
we have
=
??
+
.
(1)
By
differentiating (1) with respect to x,
we
obtain
or
(HK8*
EQUATIONS SOLVABLE FOR
The
for
first
X
321
found on
p
in the given equation,
factor placed equal to zero gives we have 2 ax
p=
2
Va.
y Va
If this value is substituted
0,
=
which
is
trial
to be a solution of the equation.
constant, and hence 182. discussed in
is
This solution, however, involves no arbitrary of a different type from that already considered. It will be
we
Placing the second factor equal to zero, and solving the resulting equation, find p = ex. Substituting this value of p in (1), we have, as the general
solution,
Ex.
2.
Clairaufs equation, y
= px
+f(p).
y
in
As
this equation already expresses
terms of x and p, we proceed imme
diately to differentiate with respect to x, with the result
As
in
Ex.
1,
placing the
first
factor equal to zero cannot give us the general
solution.
Neglecting that factor,
we have
dj)
(ZX
=
0,
whence p
=
c.
Substituting
this value for
p
in the original equation,
we
have, as the general solution,
Hence the general solution
solution
immediately by merely replacing p by
of Clairaut s equation may be written down c in the given equation. The ease of this
makes
it
desirable to solve
s
any equation for
y, in the
hope that the new
equation
may
y
be Clairaut a
equation.
.
Ex.
3.
= px +
is
Vl + p 2
in the
Since the equation
its
is
form of Clairaut
y
s
equation, with/(p)
c2 .
= a Vl
i
f
p2
,
solution
=
ex
+
a
Vl +
If
178. Equations solvable for x.
the given equation can be
solved for x, with the result
we may form
a
new
equation,
r *(!)
by differentiating
with respect to Let the solution of (2) be
(1)
y,
and replacing
dx
j
by
1 
p
(3)
,c)=0.
322
DIFFERENTIAL EQUATIONS OF FIRST ORDER
Then (1) and (3) may be taken simultaneously as the parametric representation of the solution of (1). Or p may be eliminated from (1) and (3), the result being a function of x, y, and an arbitrary constant, which is, in general, a solution of (1), but which should
be tested by substitution in
Ex. x
(1).
2p
logp
=
0.
Solving for x,
we have
x
= 2p +
?/,
logp.
(1)
Differentiating with respect to
we
obtain the equation
(2)
(3)
dy
the solution of which
is
y
= (2p + I) dp, = p 2 + p + c.
we
take
(1).
and
Since the result of eliminating p from (1) and (3) is complicated, (3) as the parametric representation of the solution of
(1)
179. Envelopes.
Let
f(x,y,
c)=0
(1)
be the equation of a family of curves formed by giving different values to the arbitrary parameter c. If any particular value of c is increased by Ac, the equation of the corresponding curve is
f(x,
y, c
+ Ac) =
0.
(2)
The limiting positions of the points of intersection of (1) and (2), as Ac = 0, will be called limit points on (1). We wish to discuss
the locus of the limit points.
One method
Ac
of
is
x and y in terms
of c
= 0,
c.
evidently to solve (1) and (2) simultaneously for and Ac. The limiting values of x and y, as
will be the coordinates of a limit point expressed in terms
c is eliminated from these values of x and y, the result is the Cartesian equation of the locus of the limit points. second method is as follows: Any point of intersection of
If
A
(1)
and
(2) is a
point of
f(x,y,
c
+ Ac)
Ac
/(a, y,
c)
A
so that we may use (3) in place of (2). As it is only the limiting positions of the points of intersection of (1) and (3) that are to be considered, we may take the limit of (3) as Ac 0, i.e.
=
c
ENVELOPES
Then
c
323
(4) is a (1)
between
and
curve passing through the limit points. Eliminating obtain the equation of the required locus. (4), we
lines represented
y
Ex. Find the locus of the limit points on the straight mx a Vl + m 2 = 0, m being the variable parameter.
First method.
by
We
first
solve the equations
with the results
+ m 2 = 0, a Vl + (m + Am) 2 = 0, y _ m 4. A?n) x + m 2  Vl + (m + Am) 2 x = a Am
y
mx
a Vl
(1)
(
(2)
(3)
y
am
and
(4), as
(4)
Taking the
limits of (3)
Am =
0,
we have
(5)
Vl +
terms of m.
?n 2
the coordinates of any limit point expressed in
Eliminating m,
we have
x2
+
y
2
=
a2
.
(7)
FIG. 116
It is thus evident
(fig.
straight lines represented
116) that the locus is a circle tangent to by the given equation.
(1),
&lt;f
each of the
Second method.
From
=
0.
dm
Eliminating
VT+ m2
(8),
(8)
m
from
(1)
and
we have
y
2
x2
the locus of which
is
+

a2
=
0,
the circle found by the
first
method.
180. In the illustrative example of the last article, the locus of
the limit points of the family, as those curves approach coincidence, is a curve tangent to every curve of the family. Hence the question is suggested, Is the locus of limit points always tangent to every
curve of the family ? To answer this question, we proceed as follows Let (x v T/J) be a limit point on one of the curves represented by
:
/(,.
/.
&lt;)=&lt;&gt;.
(i)
324
DIFFERENTIAL EQUATIONS OF FIRST ORDER
its
Then
coordinates satisfy (1) and
?=dc
The tangent
to the curve of the family at (x v y^) is
where the meaning
of
I
}
is
as in Ex. 3,
113.
The equation of the locus of the limit points may be found theo of c in terms of x and y from (2) retically by substituting the value in (1). Then the equation of the tangent to the locus of the limit
points at (x v
y^
is
or
(
113, Ex. 3)
+
But since
~
=
dc
0, (5)
reduces to
which
is
is
the same as
(3).
Hence, in general, the locus
of the limit
tangent to every curve of the family. points There may be limit points, however, which lie
on a locus that
let
i.e.
For is not tangent to every curve of the family. of the family have one or more singular points,
/P/"\
each curve
points for
Z.JSB 0,
(rl~f\9 fa/
(
)=
0.
Then such
from
(1),
points will be a part of the
\cyjx locus of the limit points
;
for,
we have
But
at a singular point the first terms vanish,
y, c)
and hence the coor
dinates of any singular point satisfy f(x,
are limit points.
=
ftf
and
c/c
=
0,
and
gent becomes
at a singular point the equation of the tan indeterminate, and hence the locus of the limit points
But
SINGULAR SOLUTIONS
may
is
325
Accord which
or
ingly,
may not be tangent to each curve of we shall separate that part of the locus
the family.
of limit points
tangent to each curve of the family, and give it the special name * is that part of the locus of the limit envelope. That is, the envelope a family of curves which is tangent to every curve of the points of
family.
Hence, in finding the equation of the envelope,
it is
neces
sary to find the locus of the limit points,
throwing out any extra
neous factor brought in by the elimination, and also discarding any part of the locus which is not tangent to each curve of the family.
The second method of finding the locus of limit points the method of determining the condition that f(x, y, c) = exactly
181.
if it is
is
0,
an algebraic equation in c, shall have equal roots (I, 64). Hence, if we form the discriminant of f(x, y, c) = 0, regarded as an equation in c, and place it equal to zero, the locus of the resulting
loci,
equation will contain the envelope. If there are any additional they are the loci of singular points, or correspond to extrane ous factors brought in by the elimination.
Ex. The equation of the example in
(x
2
179
may
(y
2
be written in the form
a2 )

a2 )
ra 2
 2 xym +
a 2 ) (y 2
is
2
?y

=
+
0.
The discriminant
of this quadratic equation in
m
is (I,
2
?/
37)
_
(
2 XT/) 2

4 (x 2
x2

a2 )
=
4 a2 (x 2

a 2 ).
Hence the condition for equal
and
this
roots
+
a2
=
0,
is
the equation of the envelope, since there are no extraneous factors.
182. Singular solutions.
Let
(1)
first
/&lt;**4)0
be the general solution of a differential equation of the
&lt;l&gt;(x,y,p)=0.
order,
(2)
Then every curve of the family represented by (1) is such that the coordinates of every point of it and the slope of the curve at that point satisfy (2). If the family of curves has an envelope, the slope of the envelope at each point is that of a curve of the familyIt follows that
the envelope
is
a curve, such that the coordinates
* Some writers call the whole locus of the limit points the envelope, while other writers define the envelope as a curve tangent to every curve of the family.
326
DIFFERENTIAL EQUATIONS OF FIRST ORDER
and the slope
of
of every point of it isfy (2).
the curve at that point sat
a solution of (2). of It is not a particular case of the general solution, since it cannot be obtained from the general solution by giving the constant a
Hence
the equation
the envelope is
particular value,
and
is
called the singular solution.
find the singular solution, if one exists, by Accordingly, the envelope of the family of curves represented by the finding general solution. This method requires us to find the general solu
we may
tion
first
;
but we
may
find the singular solution, without
:
knowing
the general solution, as follows
Let (1) and (2) (fig. 117) be two curves of the family represented by (1), intersecting at %( x v ^i) an ^ navm g the respective slopes p 1 and p 2 Then x v y v p v and x v y v p 2 satisfy (2). As curves (1) and (2) approach coincidence, in general, JJ
.
approaches a point of the envelope as a limit, and p l and p 2 become equal. Hence the locus
/(2)
FIG 117
of points for which (2), regarded as an equation in p, has equal roots must include the envelope, The equation of this locus may if one exists.
be found by placing the discriminant of the equation, regarded as
an equation in p, equal to
zero.
As
may
most
in the determination of envelopes, so here, extraneous factors appear in the course of the work, and they can be eliminated
easily
by
it.
testing
them
in the differential equation, to see
if
they satisfy
Ex.
1.
Find the singular solution,
y
if
one
exists, of the differential
.
equation
= px + a Vl + p 2
=
ex
First method.
The general
solution has been found to be
(
177,
Ex.
3)
y
+
a Vl
+
c2
;
and the envelope
x2
of this family of straight lines
is
(
179, Ex.)
the circle
+
y
2
a2
=
0.
Hence there
is
a singular solution, y
i
i.e.
xz
+
_
a2

o.
tion in p,
Second method. Writing the differential equation as a rational algebraic equa we have a _ _ 2 xyp + (y2 _ fl2) = 0? (a fl2) p2
.
the discriminant of which
is
4 a2 (x2
4
y2
a2 ).
it
Since x 2
solution.
+
y2
a2
=
satisfies the differential equation,
is
the singular
ORTHOGONAL TRAJECTORIES
Ex.
x
_f
327
2.
f
(p).
if
derived
tion
In solving Clairaut s equation ( 177, Ex. 2), we neglected the factor The equation x f / (p) = 0, however, is the equation which would be Clairaut s equation were differentiated with respect to p. Hence the
p between this and Clairaut s equation would give us an equa which would include the singular solution, if one exists. In Ex. 1 of 177 we found the solution 2 ax 2 y Va = 0. This is now seen to be a singular
elimination of
solution of the given equation.
183. Orthogonal trajectories.
A
curve which intersects each
curve of the family represented
by the equation
/(*,y,)
at a given angle
is
=
o
(i)
called a trajectory. In particular, if the given angle is a right angle, the curve is called an orthogonal trajectory. It is only this special class that we shall consider.
To determine the equation
we
first find
of the family of orthogonal trajectories, the differential equation of the family represented by
(1) in
the form
(173)
,
;
(2)
Since the trajectory and the curve of the family intersect at right angles, the slope of the curve of the family is minus the reciprocal
of the slope of the trajectory.
Hence,
if
we
is
replace j in (2)
by
dx
&gt;
the resulting equation
the differential equation of the family of orthogonal trajectories. The solution of (3) is the equation of the orthogonal trajectories.
Ex. Find the orthogonal trajectories of the family of hyperbolas xy
= a2
.
The
differential equation of this family of hyperbolas is
Hence the
differential equation of the orthogonal trajectories is
(!)
the solution of which
is
+ =
y2
x2
=
c.
Hence the orthogonal trajectories are hyperbolas, concentric with the given hyperbolas and having their common axis making an angle of 45 with the
common
axis of the given hyperbolas
(fig.
113).
328
DIFFERENTIAL EQUATIONS OF FIEST ORDER
case.
184. Differential equation of the first order in three variables.
The integrable
Any
family of surfaces
/(,
satisfies
y, *,
c)=0
(1)
a differential equation of the form
Q.
(2)
and the elimination
gives
(2).
of
c
from
this
equation by means of (1)
Conversely, we ask solution of the form
(1)
if
(1).
an equation of the form (2) always has a To answer this question, we notice that
may
be written
,
whence
which
is
 dx +
&lt;
_
&gt;
,
&gt;
dy
+
_
_&gt;
dz
=
()&gt;
(3)
an exact
it
differential equation
c, it
As
differ
(1)
(3)
does not contain
from
by some
it is
factor.
(151). must either be the equation (2) or Hence equation (2) has an integral
We
made exact by means of a an integrating factor. A special case of an exact differential equation is one in which the variables are separated.
exact or can be
factor, called
only when
shall accordingly consider three cases of equation (2),
I,
namely
:
Case
Case
II,
equations in which the variables can be separated. exact equations.
Case III, equations having integrating factors.
CASE
I.
If
the variables can be separated so that the equation
may
be written in the form
f, (x) dx
+ / (y) dy + / (z) dz =
2 3
0,
(4)
is
where any
coefficient
may
reduce to a constant, the solution
evidently of the form
dx
+/,($ dy +
/,() dz
=
c.
(5)
EQUATION IN THREE VARIABLES
Ex.
1.
329
=
0.
(x
+
a)
yzdx
+
(x

a) (y
+
b)
zdy
+ (xa)(z +
z
c)
ydz
Dividing the equation by (x
a) yz,
,
we have
6
7
x
+
a
a
dx
x
+
?/ 
+
y
dy
+

+
z
c
dz
7
=
0.
Hence
the solution
is
.
I J x
rx + a
a
dx+ J
a)
I
ry 
\b
y
dy+lJ
y
.
rz + c. dz =
z
c
k,
or or
x
+
2 a log (x
+
y
y
+
z
6 log
+
z
+
a)
log z
k,
x
+
+
+
log [(x
Za
y z
b
c
]
=
k.
CASE II. The necessary and sufficient conditions that (2) shall be an exact differential equation are ( 151)
8P_
=
3Q
dx
3Q_8fi
cz
8fi_dP
ex
cz
is
dy
cy
These conditions being
fulfilled,
equation (2)
of the
form
and the problem is If we omit from
to find
&lt;/&gt;.
(2)
one term, say Rdz,
we have
the equation
(8)
Pdx + Qdy =
0,
which, because of (6), is an exact differential equation ( obtained from (7) by considering z as constant. Therefore,
147)
if
we
integrate (8), holding z constant, we shall have all that part of $ which contains either x or y. The arbitrary constant in the solu
tion of (8)
"constant"
must be replaced by an arbitrary function of z, since in this connection means "independent of x and
y"
If,
then, the solution of (8)
is
^(x,
y,
)
= &(),
where
(f) l
is
a
known and
&lt;f&gt;
2
an unknown function, we have
(x y,
&gt;
(f
)
=
(f) l
z)fa(z).
Substituting in (7)
with
(2) will give
we should have equation (2), and comparison an equation from which to determine (z).
&lt;f&gt;
330
Ex.
DIFFERENTIAL EQUATIONS OF FIRST ORDER
2.
2
(?/
2 f z )
xdx
+
2
(z
+
x 2 ) ydy
+
(x
2
+
2
?/ )
zdz
=
0.
This equation
is
exact.
Omitting the
last term,
2
we have
the exact equation
(y
+
2
?/
z2 )
2
xdx
+
2
(z
+
z2
x 2 ) ydy
the solution of which
is
x
+ +
z 2x2
+
^
2
+
F(z)
= 0, = 0.
+^ =
Forming an exact equation from
2
2
(?/
this solution,
we have
x2 ) z
(2)]
+
z2 )
xdx
+
2 (z 2
x 2 ) ydy
+
[2 (y* f
dz

Comparing
F(z)
=
c.
this equation with the given equation, Therefore the general solution is
we have
F
(z)
=
0,
whence
x 2 ?/
+
z 2x2
+
2/
2 z2
=
k.
CASE
III.
If
equation (2) has an integrating factor
ft,
then
(9)
pPdx + pQdy + pRdz =
is
an exact
differential equation,
and therefore
P)
=
(M&gt;),
z
(M?)
=
(^),
(M) =
(^P).
(10)
Equations (10)
may
be placed in the form
dy
fji
[
dx/
ex
I
dy
ty
dQ  dz
dR\ dy ]
=
li,
 dp Sfji
dy
dz
_
dx
=
dz]
dz
_.
dx
equation by R, the second equation by P, and the third equation by Q, and adding the three resulting
Multiplying the
first
equations,
we have
dx
dy
\dx
dz]
\dy
This
is
that (9)
is
then a necessary condition that must be fulfilled in order may be an exact differential equation or that (2) may
factor.
have an integrating
(11)
It
may
be shown that the condition
is fulfilled,
also sufficient; that
factor.
is, if
(11)
equation
(2)
has
an integrating
EQUATION IN THREE VARIABLES
Let us
331
now
form
ft,
(2).
Then
suppose that (11) is satisfied for a given equation of if it were possible to find the integrating factor
we should form the equation (9), and, omitting the last term, should solve the exact differential equation
fiPdjc
+ fiQdy =
0,
as in Case II.
But since
//,
is
not known,
we may
solve the equiv
alent equation
Pdx+Qdy =
0,
= 0, where F(z) takes writing the solution in the form /(a?, y, F(z)) the place of the arbitrary constant. From this point on, the work
is
similar to that in Case II.
Ex.
3.
yz*dx
+
2 (y z

xz 2 ) dy

5 y dz
=
0.
This equation has an integrating factor. Regarding y as constant, we have the equation
yz*dx
the solution of which
is
+
y*dz
= =
0,
0.
x
+
F(y)
From
this solution
we form
dx
the differential equation
+
+ F/(2/)
by
?/z
dy
~ ~ dz =
*
If
we
divide the given equation
2
,
we have
whence, by comparison,


 F
(
(y)
0.
so that
F
(y)
F(y)
= =
0,
whence
cy.
is
Therefore the general solution of the given equation
J x +  + cy = 0,
?/
2
or
X
V
+
y
z
+
c
=
0.
The student should
and
notice the difference between the equations
Mdx + Ndy = Pdx + Qdy + Rdz =
Q
0.
332
DIFFERENTIAL EQUATIONS OF FIEST ORDEE
The former has always an integrating factor and a solution ( f(x, y, c)= 173). The latter has an integrating factor and a
solution/^,
185.
ables.
y, z, c)
=
only
when
condition (11)
first
is satisfied.
differential equations of the Let the two equations be
Two
order in three vari
P^dx
+ Q^dy + R^dz =
0,
where
7J,
I,
Qv
8,
R v P^
:
Q 2 R^
,
are functions of x, y, and
z,
or constants.
By
we have
:
dx dy dz
or
=
Q,
R,
K,
P,
P!
Q,
dx 
P
=
dy
= dz R Q
,
(2) only. gives a direction in space ( 97), (2) assigns a specific direction at each point in space. Moving from point to point in the direction determined by (2), we trace a curve in space.
Q2 R V Q = Q^R Z R^, and we shall consider equations in form Accordingly,
where
P=
R^
R=
.
Since
dx:dy:dz
Hence the
since
it
solution of (2) consists of a family of space curves, and requires two simultaneous equations to represent such
89),
it
a curve
(
follows that the solution of (2)
CM K*
is
a pair of
simultaneous equations. If the first of the equations
it is
(2), i.e.
P
*
=
Ct rJ
&gt;
is
Q
independent of
z,
an equation in two variables, the solution of which written in the form * \_n
.
may
V
be
/
/ 1
\
^
&
)
I/
Similarly, if the remaining equations of (2) are independent of x or of y, their solutions are respectively
/.l*
and
* 4) 0,
z, c 8 )
(4)
/
8 (aj,
=
0.
(5)
Any two
of the three equations (3), (4), (5) taken simultaneously
constitute the solution of (2).
EQUATIONS IN THEEE VARIABLES
dx
JJ/X. 1.
333
dy_dz
y
z
xy
z
From
y
=
we have y =
c\z.
The equation
=
xy
y
may
be written x
=
x
c 2 ev.
dy,
whence
Therefore the complete solution consists of the equations
y
taken simultaneously.
If
= ci,
x
=
c z ev
only one of the equations
(2)
can be solved by the above
:
Suppose, for example, that we first of equations (2) with the result (3) may then solve (3) for either x or y and substitute in one of the two remaining equations, thus forming an equation in y, z, and c l
method, we may we have solved the
proceed as follows
;
or x,
z,
and
c
v
which can be
solved.
This solution taken simul
(.2).
taneously with
Ex.2.
d x
the solution of (3) constitutes
dz
^ = ^=
y x
first
xyze
3
*
The
solution of
Equating the
= is y c& 0. y and the third fractions, we have dx
we have
d(x
Cixe?dx
x 1) e
=
dz

Substituting
d#
for
y
in this equation,
=
log
?
whence
=
c 2 z.
Therefore the complete solution consists of the equations
y
CiX
=
0,
Ci (X
X 1) 6
log C 2 Z
=
taken simultaneously.
If
both of the previous methods fail, we may proceed as follows By the theory of proportion we may write
:
dx
_ dy _ dz _ ~~~"
,
k^dx
+ k^dy + k dz
s
where k v k2 k s are arbitrary functions There are three cases to consider
:
of
x
t
y,
and
z,
or constants.
1.
kv k
,
k3
may
of
3).
fraction
and one
be so chosen as to give between the fourth the original fractions an equation which can be
solved (see Ex.
334
2.
DIFFEKENTIAL EQUATIONS OF FIEST OEDEK
k v kz k3
,
may
be so chosen as to
Then \dx
+
k^dy +
k 3 dz
=
make k^P
+kQ+kR=
2
s
0.
0,
and
if
this equation falls
under one
of the cases of
184,
its
solution
is
one of the equations of the
solution of the given differential equations (see Ex. 4). 3. may form a new equation
We
\dx +
+ k dz _ k(dx + k^dy + k dz == k[p + k& + kfi ^P + ^C + AVR
k^dy
s
8
f so choosing the multipliers Jc v k 2 ks , k[, k2) k3 as to equation solvable by previous methods (see Ex. 5).
,
make the new
Ex.3.
Let
^= x * y x +
dZ x
+
z
ki
=
0,
k2
=
1,
&3
=z
1.
Then
From
x
=
y
z
is
dx x
dy
x
,
dz
d y~ d z y y
z).
+
y
find
x
+z
x
z
we
=
Ci (y
Substituting this value of x in the equation
x
+
=
y
x+z
,
we have
dy
dz
ciz
(1
(1+
the solution of which
ci)y
log (z
y)
=
c2
Therefore the general solution consists of the two equations
x
=
ci(y

z),
log (z

y}
=
c2

taken simultaneously.
Ex
Jjg
4
dX
y
+
z
dz dy =  = x x + y + z
1
].
T.
1
J.
1
Then
and hence
ki(y
+
z)
+
k2 (
x)
kidx
+
k 2 dy
+ k s (x + + k 3 dz
dz
z
y
+
0.
z)
=
0,
But
this equation is
is
dx
evidently
 dy
x
=
=
0,
c\.
the solution of which
y
Substituting the value of y
+
z
from
is
this equation in
=
,
we have
X
=
Ci
^
X
,
the solution of which
Hence the general solution
x
consists of the equations

y

z
=
ci,
x
+
ci
log (x

ci)
=
e2

y
EQUATION IN THREE VARIABLES
Ex.
335
dx
5.
dy
z z
dz
y Let
+
+x
x
+
y
fci
=
k2
*^
=
fc 3
=
+
1.
dx
T^lioti

dy
dz
d (x
x
y
Again,
let k\
+
z
1
z+x
ks
x
y
2(x
+ +
y
y
+ +
z) r _ z)
.
=
1,
fc 2
= =
1,
,
=
0,
and we obtain the equal fraction
;
also, letting ki
=
0,
kz
ks
=
y y
y
1,
we
obtain the equal fraction
y)
.
d (x
"2(x
+ +
+ z)
_ d (x ^
,
_ d (y

z)
+ z)~
yx
zy
whence
Vx +
+z=
186. Differential equation of the first order in three variables.
The nonintegrable
case.
Consider again the equation
(1)
Pdx + Qdy+fidz=Q.
is
Geometrically, the equation asserts that the direction dxidy: dz Q R. Consequently, the geo perpendicular to the direction metrical solution of (1) consists of loci perpendicular to the curves
P
:
:
denned by the equations
^=^
~P
~Q
= ^z
fi
first
We may
surfaces
therefore seek the solution of (1)
in a family of
(3)
f(x,y,z,c)=0,
This is the form of will be orthogonal to the curves (2). the solution discussed in 184, and does not always exist. This leads to the geometric theorem that it is not always possible to
which
find a family of surfaces orthogonal to a given family of curves. the solution of (1) in the form (3) does not exist, it is
When
still
possible to find curves
which
satisfy (1)
and hence cut the
curves (2) at right angles. In fact, we may find a family of such curves on any surface assumed at pleasure. For let
&lt;(&gt;
y,*)=o
Then, from
(4)
(4),
be the equation of any arbitrarily assumed surface. dz dx
d
=
0.
dz
(5)
336
DIFFERENTIAL EQUATIONS OF FIRST ORDER
(1)
Equations
and
(5)
may then
solution will be a family of curves condition (1).
Ex.
be taken simultaneously. Their which lie on (4) and satisfy
xydx
+
ydy
+
zdz
=
0.
(1)
This equation cannot be satisfied by a family of surfaces. It may, however, be satisfied by curves which lie on any assumed surface and cut at right angles , the curves ^
^=
=
.
=
,
xy
or (Ex.
1,
y
z
(2)
185),
Let us assume the sphere
Then
and from
(4)
and
(1),
= c 2 ev. Cl z, y 2 2 2 = x + y + z a2 xdx + ydy + zdz = 0, dx = 0, whence x = c.
x
.
(3) (4)
Hence the
satisfy
1).
circles cut
from the sphere x 2
+
y2
+
z2
=
a 2 by the planes x
=
c
Again,
let
us assume the hyperbolic paraboloid
z
Then
and, from
(6)
ydx
f
xdy
dy
dz
= =
xy.
0,
(5)
(6)
and
(1),
y xz
is
=
xy
=
zy
dz
+
x2 y

tf\
y2
One
solution of (6)
known
to be z
xy.
Using
this,
we have
xdx
dy
whence
(1
+
y)
Vl +
x2
=
c.
(8)
Then
the curves defined
by
(5)
and
(8) also satisfy (1)
PROBLEMS
Express the solution of each of the following equations in the form of a
series
:
Solve the following equations
4.
5. 6.
7.
:
p 2  3p zi/(p
2
10
=
2
0.
10.
2
?/
py 2  2p*xy + p*x 2 =
)
1.
+
l)(3:
+
)p
=
0.
H.
12.
0.
8. 9.
 2 2 = 0. p3 + 2 yp2 _ X2p2 _ 2 x 2 yp = p 2 (x 2  a2 2  4 a2 = 0.
x 2p 2
+
xyp
?/
13. 14.
)
p*
+ Zpy ctn x 
y2
=
0.
15.
2 2px = 0. T/(! + p y = yp* + 2px. 2 2  2 xyp 3 + x 2 (1 + )p 2 y  2p = p*. p 3  4 xyp + 8 y 2 = 0.
7/
PROBLEMS
16. x2p 2 17.
337
20.
21. 24.
25. 27.
28.
29.
 p 2 + 1 = 0. 18. p 2 + 2 p X 3 = 0. 19. p(x + y)* = l. y(p + y)p*x 2 = p 2x + p. 22. x = 4p + 4 p. 2 = 0. 2 2 23. y = p 2  2 p. xfyp  x (x + y)p + (x  y) (x + 2/)  s 2 2 p 3  (x 2 + xy + y 2 )p + xy (x + xy + y*)p x y* = 0.  2p 2 x = 0. 26. p 2  2xp + x 2 + 4y 2 = y + 2py  4xy 8p + 4p 2 y 2 (x 2  1) + 1 = 0. (Let y 2 = z.) y* e 3 (p  1) + p 3 e 2 ^  0. (Let & = v, e* = u.) x = Q. (Let y z = v, x 2 = w.) py*(px y) +
?/

0.
Find the general and the singular solutions of the following equations
30.
31.
:
p
32. 33.
+ x 2xyp + y*l= 0. 3x  4xp + p 2 = 0. 2 2  a2 2  2 xyp  x = 0. )p (x =  xp + x4p 2 y
2 2
(l
2
)
34.
2
?/
35.
a2
+ (1  y* = p*y*. +
x 2 yp

2 xy)p
+
x 2p 2
=
0.
36. x 3p 2
+ +
a3
=
0.

37. Find the singular solution of the equation
38. Find the singular
/
p3
4 xyp
8 y2
= 0.
= px
f
and the general solutions of the equation y
62 f
a 2p 2 and interpret them geometrically.
,
:
Solve the following equations
39. (xy*z
+ (x yz + yz) dy + (x 2 2  x 2 + y 2  1) dz = 0. 40. (y + z}(2x + y + z)dx+(z + x)(1y + z + x)dy + (x + y)(2z+x + y)dz = Q. 41.3 x^i/zdx + (2 2 z + z) dy + 2 (y + yz 2 dz = 0. 42. (y + 2  6  c) dx + (z + x  c  a) dy + (x + y  a  b) dz = 0. 43. (y 2 + yz) dx + (z 2 + zx) cZy + (y*  xy) dz = 0.
xz)
dx
2
?/
?/
)
45.
(1
?/z
+
2
x
46.
dx
47. ?/zdx
+ 2/)dx + (1 + x + y)dy + (x + y)d = + (y 2 z  xz 2 dy  2 (y + z) dz = 0. zxdy + (x2 + 2 )dz = 0.
)
0.
?/
?/
Solve the following systems of simultaneous equations
:
48
dx
yz
dx
dy
dz
dx
=
_dy
x
dy
zx
dz
dy
dz
zx =
1
xy
^
dx
338
KQ
DIFFERENTIAL EQUATIONS OF FIRST ORDER,
dx
z
_ _
xdy x2 + z 2
dy
dz
_
o
dx
i.
dy
y
2
dz
x
dz
x2 5
_
+
2xy
dy
z
(xy)z
dz
54
KK
dx
dx
y
x~2xy~zxy
dx
 z~
dx
x~ xy
dy
c
00.
x+yz
x
= dy = az.
,
.
oti.
CQ
z
xyz
x2
yzx
z2
z
56.
__ = 2y
y
=
fc.
60.
y2
2
xy
2 xz
61. Find the envelope of the family of lines y
= 2 mx +
ac 2
?n 4 ,
m
being the
variable parameter.
62. Prove that each line of the family
c 2x
+
=
a 2 ?/
=
0,
a being the vari
able parameter, forms with the coordinate axes a triangle of constant area, and find the envelope of the family.
63. Find the envelope of the parabolas y 2 64.
a(x
a).
straight line moves so that the sum of its intercepts on the coordinate axes is always equal to the constant c. Express the equation of the family of lines in terms of c and the intercept on 0Z", and find their envelope.
A
65.
The semiaxes
of the ellipse
x2
h
2
v2
b2
=
1
a
vary so that ab
=
c 2,
where
c is
a constant. Express the equation of the ellipse in terms of a as a variable parameter, and find the envelope of the family of ellipses thus defined.
66. Find the envelope of the family of straight lines formed B in the equation y = mx 2 pm
by varying
the
slope
m
pm
.
y
=
67. Find the envelope of a family of circles having their centers on the line 2 x and tangent to the axis of y.
of the parabola
68. Find the envelope of a family of circles which have the double ordinates y 2 = 4px as diameters.
69. Find the envelope of a family of straight lines which move so that the portion of each of them included between the axes is always equal to the constant c.
70.
parabola y
Find the envelope of a family of circles which have their centers on the 2 = &px and pass through the vertex of the parabola.
71. Find the equation of a curve such that the tangent cuts off from the co ordinate axes intercepts the sum of which is always equal to the constant k. 72.
Find a curve
in
the origin upon any tangent
which the projection upon OY of the perpendicular from is always equal to the constant a.
73. Find the equation of the curve in which the part of the tangent included between the coordinate axes is always equal to the constant a.
74. Determine the equation of a curve such that the portion of the axis of x intercepted by the tangent and the normal at any point of the curve is always equal to the constant k.
PROBLEMS
75.
339
pole upon any tangent
76.
Find the polar equation of the curve in which the perpendicular from the is always equal to the constant k.
k times the radius vector of the point of contact.
Find the polar equation of a curve such that the perpendicular from the
is
pole upon any tangent
77.
Find the orthogonal trajectories of the family of parabolas y 2

=
2
78. Find the orthogonal trajectories of the family of ellipses X being the variable parameter. 79.
_ =
+
^
1,
fl2
Find the orthogonal trajectories of the family of
ellipses
a2
+
a2
=
1
80. Find the orthogonal trajectories of the family of parabolas 2 y =4ax+4a2 .
81. Find the orthogonal trajectories of a family of circles each of which tangent to the axis of y at the origin.
82. Find the orthogonal trajectories of the family of circles each of
is
which
passes through the points
83. If fir,
\
0,
(
1, 0).
rH
Uv/
=
is
the equation of a family of curves, prove that
f
(
r, 0,
r2
\
=
is
the equation of the orthogonal trajectories.
84. Find
2 a 2 cos 2
0.
the
orthogonal trajectories of the family of lemniscates r2
=
85. Find the orthogonal trajectories of the family of cardioids r=a(cos d
+ 1).
86. Find the orthogonal trajectories of the family of logarithmic spirals r = e0.
CHAPTEE XVIII
THE LINEAR DIFFERENTIAL EQUATION
187. Definitions.
The equation
JA
where
a)
y, is a
p v p^

,
pn _
lt
pn and
,
f(x) are independent of
,
linear differential equation. If the coefficients p v p 2 , p n _ lt p n are constants, the equation becomes the linear differential equation
with constant
coefficients,
where a v a z
In both


,
,
a n _ v a n are constants.
(2)
(1)
and
f(x)
is
a function of x,
which may reduce
this, it is
to a constant or even be zero.
We
shall begin
with the study
jctx
of (2).
To do
con
venient to express
by Dy,
D*y, \ ax by
~
,
ax
by
D y,
n
and to
rewrite (2) in the form
D y + a^^y +n
+a
n
_^Dy
+a
n
y =f(x),
or,
more compactly,
a,)y =f(x).
(3)
in parenthesis in (3) is called an operator, and we are said to operate upon a quantity with it when we carry out the indicated operations of differentiation, multiplication, and addition.
The expression
Thus,
3
(Z&gt;
if
we
2
operate on sina? with
2
3
Z&gt;
2
2Z&gt;
+ 3D +
x.
5,
we have
5 sin
D + 3D
5) sin
x
= =
cos x
+
2 sin x 3 sin
3 cos x
x
2 cos x
340
EQUATION OF FIRST ORDER
Also, the solution of (2) or (3)
is
341
expressed by the equation
where the expression on the right hand
of this equation is not to be considered as a fraction but simply as a symbol to express the solution of (3). Thus if (3) is the very simple equation Dy =f(x),
then
(4)
becomes
In this case
means
integration with respect to
x.
What
the
more complicated symbol (4) may mean, we are now to study. 188. The equation of the first order with constant coefficients. The linear equation of the first order with constant coefficients is
or,
symbolically,
is
(D
a)y =f(x).
(1)
This
we have
a special case of the linear equation discussed in 80, and a, f2 (x) =f(x) in formula (5) of only to place f^x)
=
80 to obtain the solution.
We
have in this way
y
=
two
parts.
The
is ax
solution (2) consists of
The
is
first part, ce
if
ax
,
contains
an arbitrary constant, does not contain f(x), and,
not a solution of
e
(1)
unless f(x)
zero.
taken alone, The second part,
is
e
 ax
I (1),
f( x ^d X) contains /(#), and, taken
is satisfied
alone,
a solution of
0.
since (1)
e
by
is ax
(2)
when
c
has any value including
Hence
ax
I
e~
ax
f(x)dx
called & particular integral of (1), and,
is
in distinction to this, ce
The sum
gral
is
of the
called the complementary function. complementary function and the particular inte
be written
the general solution (2). The complementary function can down from the lefthand member of equation (1), but the determination of the particular integral requires integration.
342
Ex.
1.
THE LINEAR DIFFERENTIAL EQUATION
Solve
dx
+ 3y= 5x
3
.
The complementary function
5 e~ 3x
is
ce~ 8x
.
The particular
integral
is
f
e3
*x dx
is
3
=

f
x3
{
x
2
+
3
1 
J x
\&lt;.
Hence the general solution
Ex.
2.
y
ce~ s *
+
jj
3
*2
;
+
Solve
dx
+
y
=
sin x.
is
The complementary function
e
cer*.
The
L s in
particular integral
cos x.
is
 x C ex s i n xdx
is
= =
xf
(
19,
Ex.
5)
Therefore the general solution
y
ce~ x
^ sin x

cos x.
189.
The operator
The
solution of linear equations of
coefficients depends upon the solution higher order with constant Hence a knowledge of the oper first order. of the equation of the
ator
D
is of
a
prime importance.
i
J_s
"~~~~
We
upon
give a few of the results
certain elementary func
obtained by operating with
tions
Ct
which occur frequently
in practice.
is ce
In writing these formulas
ax
the complementary function, which
in all cases,
is
omitted.
D
cu
=c
a
D
u.
a
(1)
ss
a =0.
(3)
s
e
te
f+r
_^
k
MX
nn^a
,
(4)
unless k
=
a.
a
*
a
(5)
Da
k
unless k
(6)
a
3
(k
cCf
(k
a)
= a.
i
~m + 1
m
(8)
EQUATION OF SECOND ORDER
1
.
343
,
a sin kx
k cos
lex
~
6SS
a
=
U
1
,
&lt;*
a,
cos kx
^
1
J+JT
+ k sin kx unless a =
K
(11)
These formulas may all be proved by substituting the special functions concerned in the general formula 188, (2)). For (3), ( (7), (9), and (11) the student may refer to Exs. 4, 5, and 6, 19. The derivation of (10) is as follows: 188,
By
(2),
i
D  ki sin kx
.
=e
ux
r
\
e~
kix
sin
kx dx.
J
If we attempt to use the method of Ex. 5, 19, it fails to work, but by replacing sin kx by its value in terms of the exponential functions ( 169, (5)) we have
Formula
190.
(12)
is
derived in a similar manner.
The equation of the second order with constant coeffi cients. The symbol (D a)(D l)y means that y is to be operated on with D b and the result a. Now operated on with D
(
D ~^}y = ~~
ax
l&gt;y,
and hence
(I)
ab. b),q This result, obtained by considering the real meaning of the operators, is the same as if the operators a and l had
where
p=
(a
+
=
D
D
344
THE LINEAR DIFFERENTIAL EQUATION
been multiplied together, regarding
Similarly,
D
as an algebraic quantity.
we
find
(2)b)(Da)y
= [D  (a + b) D +
2
ab] y
= (D  a) (D  I) y.
That is, the order in which the two operators used does not affect the result.
2
D
a and
D
I are
Moreover,
b so that (1)
if
(Z&gt;
h
pD +
q)
y
is
given,
it is
is satisfied.
In
fact,
we have simply
possible to find a to factor
and
2
Z&gt;
+
pD +
q, considering This gives a method order with constant coefficients.
D
as an algebraic quantity. of solving the linear equation of the second
For such an equation has the form
or,
what
is
the same thing,
(& + pD +
where
g)y=f(x),
is
(2)
q are constants and f(x) p reduce to a constant or be zero.
and
a function of x which
may
Equation
whence, by
(1)
may
(2),
be written
188,
"
ax
)
dx
Again applying
188,
(2),
we have
\
e/
/
".\
/
\
= cebx+ e C e b*( c e
x
&gt;
x
+e"
J
\
J
Ce axf(x)dx)dx.
:
(3)
There are
I.
now two
cases to be distinguished
If
a
3= b, (3)
becomes
(a
II.
If
a
~ bx x Ce~ axf (x) dx\ dx. y = c / + c/** + e C(e = b, (3) becomes ax ax ax y = (c + Cl x)e +e CCe~ f(x)dx JJ
b)x
2
.
(4)
2
(5)
In each case the solution consists
of
two
parts.
e"*,
The one
is
the
complementary function c^+c/* or (c 2 + c^) It can be written down arbitrary constants but not involving f(x).
involving two
EQUATION OF SECOND ORDER
345
from the lefthand member of the equation, and is, in fact, the solu 0. The other part of the tion of the equation (D a)(D solution is the particular integral, and involves f(x). Its general computation by (4) or (5) necessitates two integrations.
=
l&gt;)y
Formula
(4)
when a and
b are
holds whether a and b are real or complex. But conjugate complex, it is convenient to modify
:
the complementary function as follows
Let us place
in.
a
= m + in,
b
is
=m
Then the complementary function
where
= ^"[^(cos nx + i sin nx) f c (cos nx = e mx C\ cos nx + C sin nx), C = c + c C = i(c^ c Since c and c
2 (
z
1
i
sin nx)]
(6)
l
2,
2
2 ).
l
2
are arbitrary con
stants, so also are C\
and C2 and we obtain
,
all real
forms of the
.
complementary function by giving real values to C^ and C2 The form (6) may also be modified as follows Whatever be the values of Cl and C2 we may always find an angle a such that
:
cos
a
=
C
*
&gt;
sin
a
=
C2
Then
(6)
becomes
where a and k = V (7 2 + C 2 are new arbitrary constants.
find
Or,
we may
Then
an angle ft, such that sin (6) becomes
ft
=
^L
l
,
cos
ft
=
This equation
may
be written
(D + 2) (D + 3) ?y = e*. The complementary function is therefore Cie~ 2x +
ticular integral,
c 2 e
3:r .
To
find the par
we proceed
(D
as follows
:
+
3)
y
=
D f
&=
2
e~
*
J
C
Therefore the general solution
is
346
Ex
.
THE LINEAR DIFFERENTIAL EQUATION
2
*&gt;L
.
dx2
+
a
*+
dx
1/
=x
.
This equation may be written (D + l) 2 y = x. Therefore the complementary function is (ci
ular integral,
+
c 2 x)
er x
.
To
find the partic
we proceed
(D
as follows
1
:
+
1)
y
D+1
=
1
x
= 
r e * xex dx
I
=
(x
J
x
1.
y
D
(x
\
1)
=
e x
r
I

1)
e?dx
=
x

2.
I
is
*/
Therefore the general solution
Ex.
3.
Consider the motion of a particle of unit mass acted on by an attract
ing force directed toward a center and proportional to the distance of the par ticle from the center, the motion being resisted by a force proportional to the velocity of the particle.
If
we
attracting force is positive constants.
take s as the distance of the particle from the center of force, the ds ks and the resisting force is h where k and h are
,
Hence the equation
d2s
dt 2
of motion
is
=
ks
h
dt
,
ds
,
The
factors of the operator in (1) aref
\
DH
2
:
2
We
have therefore to consider three cases
I.
W _ 4k
&lt;
o.
The
e
2
solution of (1)
is
then
s
=
(
Ci cos
1
+ C 2 sin
&gt;
or
The graph of s has the general shape of that shown in I, 155, fig. 161. The particle makes an infinite number of oscillations with "decreasing amplitudes, which approach zero as a limit as t becomes infinite.
II.
h2

4k
&gt;
0.
The
solution of (1)
rrn&gt;
is
then
The
III.
particle
h*
makes no
oscillations,
but approaches rest as
is
t
becomes
infinite.

4k
=
0.
The
solution of (1)
The
particle approaches rest as
t
becomes
infinite.
SOLUTION BY PARTIAL FRACTIONS
191. Solution
347
of solving
(1)
by
partial fractions.
Another method
the equation
tfy =/&lt;*),
is
when
We may
the factors of the operator are unequal, express the solution in the form
as follows
:
^
~
2)
1
2
f
pD + q
=
~~
1
(D
a)(D
b)
have seen that in many ways the operator may be handled as if it were an algebraic quantity. This raises the ques
tion
Now we
D
whether
it
is
proper to separate
into partial
(D
fractions.
a]
(D
b)
Algebraically
1
we
have, of course,
=
l)
1
l
(
(D
and the question
a)(D
is,
a
l)\D
a
D
JL
l
Is
a solution of equation (1)
?
The way
in (1)
of (1),
to
answer
this question is to substitute this value of
result.
and observe the
We
y have then, on the lefthand side
ab
~
a
1
1
"
b
""
^
a x)l f\
=/(*)
Consequently
(2) is a solution of (1).
Writing
(2)
out in
[
full,
we have
C&lt;T~f(x)dx
f"
e
a
l
e
b*
u
,_/
o
C&lt;T*f(x)fa (3)
,_/
348
THE LINEAR DIFFERENTIAL EQUATION
complementary function
is
It is to be noted that the
the same
as in
but the particular integral appears in another form. b. This method fails if a
190,
(4),
Vy
dx*
dy_
dx
Since this equation
may
be written
the complementary function we proceed as follows
:
(D2)(D+ B)y = e**, is Cie 2x + c 2 e~ 3x To find
.
the particular integral,
y
=
D2
5
j
e +D_6
2x
ID 
2
D+3
Therefore the general solution
is
192.
The general equation with constant
of
coefficients.
The
solving a linear equation of the second order with constant coefficients are readily extended to an equation of the nth order with constant coefficients. Such an equation is
methods
or,
symbolically written,
n
(D
The
factors
first
+ ai D
is
n
l
+
+
a^D + a^y =f(x).
(2)
step
to separate the operator in (2) into its linear
(2) as
and to write
(D
where r v rf ,
,
 r,) (D  r
l
.
. .
2)
(D rj =/(*),
n
(3)
r n are the roots of the algebraic equation r
n
+ a^~ +
+ a.^r + a =
0.
It
may
be shown, as in
(3) are equivalent, (2) is immaterial. (3)
and
and that the order
190, that the lefthand members of of the factors in
GENERAL EQUATION
The general
solution of (1) consists
349
two
parts, the
now
of
com
plementary function and the particular
integral.
The complementary function
form
is
written
down from
,
the factored
the solution of (1) in the If r v r 2 ., rn are all distinct, the complementary function consists of the n terms
is
of the lefthand side of (3), special case in which f(x) is zero.
and
c/ *+c/^+
where c v
If,
+
/*,
(4)
c


2,
,
cn
however,
(4)
D
c
are arbitrary constants. r appears as a &fold factor in (3), k of the
i
terms of
must be replaced by the terms
+ cx h
(4)
*
k

Also,
if
two
factors of (3) are conjugate
complex numbers, the
corresponding terms of
sines
may
be replaced by terms involving
and
cosines, as in (6),
is
190.
The particular integral
found by evaluating
This
1.
f(X )

may
be done in two ways
(5)
&gt;
:
The expression
tors

may
be evaluated by applying the opera
left.
,
in succession from right to
This
leads to a multiple integral of the form
(*&*..
2.
.
C
(6)
side of (5) may be separated into partial frac the factors of (5) are all distinct, this leads to an integral of the type
tions.
The lefthand
When
C
r
*
f (x) dx +
/(*)&.
.
.
(7)
If
some of the factors must be combined with
of (3) are repeated, the previous
this.
method
350
THE LINEAK DIFFERENTIAL EQUATION
(6)
In evaluating
omitted,
function.
since
and
(7)
they are
taken
the constants of integration may be care of in the complementary
of the
The general solution is the sum and the particular integral.
193. Solution
complementary function
by undetermined
coefficients.
The work
of find
ing the particular integral may be much simplified when the form of the integral can be anticipated. The particular integral may
then be written with unknown, or undetermined, coefficients, and the coefficients determined by direct substitution in the differential
equation.
Since both (6) and
(7),
,
192, consist of successive appli
cations of the operator
.LJ
ct
we may apply
it
the formulas of
if
189
in
many
cases.
From
(2),
189,
follows that
/(a?) consists of
an aggregate
of terms, the particular integral is the
sum
of the
parts obtained by taking each follows that the coefficient of a
term by itself. From (1), 189, it term oif(x) affects only the coeffi
cient of the corresponding part of the particular integral. 189 we can deduce the following: the other formulas of
1.
From
a m _ l x{a m (m a positive + the particular integral is of the form Ajt m + AjP^ + integer), d m i x + A m unless the lefthand member of the differential equa
.
When /(#)=
,
a Q x m + a 1 (tfn
~1
+
+
tion
contains the factor
is of
r
D
l
r
.
solution
the form x (A xm
terms of the form
A m+l xr ~ +
In the latter case the particular +A l xm ~ l + + A m _ l x +A m ), while

+ A m+r _ x+A m+r
l
occur in the
complementary function, and hence need not be assumed as part
of the particular integral.
ae kx the particular integral is of the form Ae*x unless the lefthand member of the differential equation contains
2.
, ,
When f(x) =
r
a factor
(D
r kx
k)
,
.
In the latter case the particular integral
r
is
of the
kx
form
A r _ l x A r)e while terms of the form (A l x + occur in the complementary function, and hence need not be
Ax
e

~l
+
+
assumed
3.
in the particular integral.
the
When f(x) = a sin kx or,a cos kx, the particular integral is of form A sin kx f B cos kx, unless the lefthand member of the
2
(Z&gt;
differential equation contains the factor
+k
2 r
)
.
In the latter
SOLUTION BY UNDETERMINED COEFFICIENTS
while terms of the form x
case the particular integral is of the form x (A sin kx r~l
r
351
}
(A l sin kx +
B
l
coskx)
+
+ B cos kx) + (A smkx
r
+B
r
cos kx) occur in the
complementary function, and hence need
not be assumed in the particular integral.
4.
When f(x) = axme kx (m
member
gral is of the
form (A xm + A 1 xm
the lefthand
tor
r
.
a positive integer), the particular inte ~l fcx Am _ lx unless m )e of the differential equation contains a fac
+

+
+A
,
In the latter (D k) m form x r (A Q xm + Ajc *l
case the particular integral
kx
is
of the of the
form (A m+l x r
ular integral.
occur in the comple l function, and hence need not be assumed in the partic mentary
\
~1
+A m+r _ x+A m +
M^^+AX ^ while terms
r )e
The above statements may all be summed up in the following rule, which may also be sometimes used when f(x) is not one of
the forms mentioned above
:
If u
is
a term of f(x), and if u^
t&
8,
,
u k are
all the distinct
functions (disregarding constant
coefficients)
which can
he obtained
from u
by successive differentiation, then the corresponding part \A k u k of the particular integral is of the form Au + A l u l + unless u is a term of the complementary function, or such a
,
term multiplied by an integral power of
corresponding part
x.
In
the latter case the
particular integral is of the form where r is the number of times the + A^u^ ), which gives in the complementary function the term u, or factor, u divided by an integral power of x, appears in the lefthand
of
the
xr (Au

+ A ku t
member of
the differential equation.
,
The above rule is of course valueless unless the functions u v u 2 uh are finite in number. In applying it to functions other than
,
those already discussed, the student should consider that he is mak ing an experiment. If a function assumed in accordance with the
rule
is is justified.
found to satisfy the differential equation, the use of the rule When it fails, recourse may always be had to the gen
eral formulas (6)
and
kr
(7),
192.
When/(#)=
may
tion.
e
the
&lt;j&gt;(x),
work
of finding the particular integral
be lightened by substituting y
= e kx z
e
kx
in the differential
e
equa
kx
,
Then, since all derivatives of
z
contain the factor
the
352
THE LINEAR DIFFERENTIAL EQUATION
differential equation
for z in
new
may
z,
be divided by
is
e
kx
,
and there
is
&lt;t&gt;(x).
is left
an equation
which the lefthand member
y
When
this equation is solved for
readily found (see Ex.
2).
..
This
.
dx s
dx 2
dx
may
be written
D(D
2
I)
?/
=
xe 2x
+
e*,
whence the complementary function
is c\
+
x (c? 4 c^x) e .
The term xe 2x
,
,
if
suc
2 cessively differentiated, gives only the new form e *. Hence the corresponding part of the particular integral has the form Axe 2x + Be 2x The term ex gives no
new form
if
differentiated
;
but since
it
appears in the complementary function
2
corresponding to the double factor (D
ticular integral
is
I)
,
the corresponding part of the par
Cx 2 ex
.
Substituting
y
in the differential equation,
= 4xe 2x + Be zx + Cx 2 e x
we have
24xe 2 * +
Then
whence
(5
A+
2 B) e 2x
+
2 Cex
= xe2x +
2
e*.
24 =
1,
54 +
27^
^=j,
= 0, = , #
C= C=
1,
1.
The general
solution of the differential equation
is,
accordingly,
y
=
=
ci
+
(c 2 f
c 3 x) e^
+
l xe 2a:
%&*
Ex.
2.
d y 2
&lt;Zx
2
\
y
xe Sx sinx.
We
place y
=
e Sx z.
There results
The complementary function integral we assume
z
is
e 3 r

(c 1
sinx
4 c 2
cosx).
For the particular
= 4x sin x + .Bx cos x + C sin x + E cos x,
and, on substitution, have
+ (64+ 9E)xcosx + (64  27? + 9C e^sinx + (24 + 6U + 6C + 9^)cosx = xsinx. Therefore 4  6 B = 1, 64 + 95 = 0, 6425+ 9C 6^ = 0, 24 + 65 + 6(7 + 9^ = 0, 4  ^, 5= whence C = ~ T! 7, ^ = T$fy
(QA  6B)xsinx
A
Therefore the general solution of the original equation
is
y
=
Cisinx
+
c2
cosx
+
es x
(y ^ x sin x
1
^
x cosx
T F sinx + T f  r cosx).
SIMULTANEOUS EQUATIONS
194.
coefficients.
353
in
Systems of linear differential equations with constant The operators of the previous articles may be employed solving a system of two or more linear differential equations
with constant coefficients, when the equations involve only one independent variable and a number of dependent variables equal
to the
number
of the equations.
The method by which
this
may
be done can best be explained by an example.
S+
We may now
f+
may
be written
These equations
(D
3
2
+
4)
x
Dx + (D2 +
 3 Dy = = 4) y
sin 2
t,
1.
(I) (2)
cos 2
eliminate y from the equations in a
manner analogous
(1)
to that
used in solving two algebraic equations. the coefficient of y in (2), and have
We
first
operate on
with
D +
2
.
4,
16)x3(D 3 + 4D)y =
since
(
0,
(3)
D2 + 4) sin 2t =
coefficient of
4 sin 2
(1),
1 f
4 sin 2
1
0.
We
then operate on
(2)
with
3D, the
y in
and have
= 6sin2,
since 3
(4)
D cos 2
1
=
sin 2
1.
By adding (3) and (4) we have 17D 2 + 16)x = 6sin2,
cos 4
(5)
the solution of which
is
x
=
ci sin 4
1
+
c2
+
c 3 sin t
+
c4
cos
+
J sin 2
2
.
(6)
Similarly, by operating on (1) with tracting the result of the former operation
3D
and on (2) with D + 4, and sub from that of the latter, we have
(D
the solution of which
is
4
+
cc
17
D2 +
1
16)
y
=
6 cos 2
,
(7)
y
=
c 5 sin
4
1
+
cos 4
+
c 7 sin
+
c 8 cos
+
^ cos 2
$.
(8)
The constants
of x
in (G)
and y given
in
((&gt;)
and (8) and
are, however, not all independent, for the values (8), if substituted in (1) and (2), must reduce the
latter equations to identities.
Making
these substitutions,
c2 )
we have
t
12
(c

ci) sin
4
1
 12 (c 5 +
12
(C B
cos 4
1
+ +
3
(c 8 f c s ) sin
t
3
(c 7
c 4 ) cos
+
sin 2
c 8)
1
=
t
sin 2
1.

12
(c e
ci) cos 4 1
+ +
c 2 ) sin
4
1
3
(c 8
3
(c 7
c 4 ) sin
+ + cos 2 =
cos
cos 2
.
354
THE LINEAR DIFFERENTIAL EQUATION
may
are
In order that these equations
C6
be identically
,
satisfied,
C7
we must have
=Ci,
C5
= C 2
(2)
C8
= C 3 +
,
=
C4 .
Hence the solutions
x
of (1)
and
= = y
ci sin
4
+
4
c2
cos 4
Ci
+
c 3 sin t
c4
cos
c3
i
f
^ sin 2
Z,
(9)
.
c% sin
+
cos 4
+
c\ sin
cos
:
+
^ cos 2
(10)
The method
of solving
(6),
may
4
1
be modified as follows
substitute this value in
Having found
as before
the value of x in
we may
Ci sin
(1).
We
t
have then
l sin 2
1
,
.whence
Dy = =
?/
4
Ci
4
c2
cos 4
2
1
+
cos
c 3 sin
t
+
c4
cos
cos 4
c 2 sin
4
c3
+
c 4 sin t
+
l cos 2
+
C.
The constant C
is
found to be zero by substitution
in (2),
and we have again
the solution (10).
195. The linear The equation
differential equation
with variable
coefficients.
^coefficients p v p 2 p n _ lt p n are functions of x, can be solved in terms of elementary functions. In fact, such rarely an equation usually defines a new transcendental function. We
where the
,
,
may, however, easily deduce certain simple properties
tion of (1).
of the solu
Consider
first
the equation
which
differs
from
yn
(1) in that
\&gt;
the righthand side
is zero.
V Vv y^
(2),
Vn are
n
linearly independent
is
* solutions of
then the general solution of (2)
y = c iVi+ c 2/2+
2
+c
i2/i+
c n yn
&gt;
(
3)
where
c
v
c

2,
,
cn _
v
cn
are arbitrary constants.
The fact that
substitution in
(3) is a solution of (2)
(2).
may easily be verified by direct
That
(3)
is
upon the
fact that it contains
n
the general solution of (2) depends arbitrary constants, and the number
of constants in the general solution of a differential equation is equal
to the order of the equation.
* The n functions t/,, y 2 no relation of the form
,
This statement
we
shall not prove.
if
,
yni, yn are said to be linearly independent
H\a.n\y n
there
exists
i2/i
+
2 2/ 2
\
+ a n yn = 0,
"
where a v a 2
,
,
a n i, a n are constants,
and = means
identically
equal."
VARIABLE COEFFICIENTS
Keturning now to
(1),
355
If y
i
and u
is is
l/nv Vn y* any particular
&gt;
we may say are n H near ty independent
:
solutions of (2),
solution of (1), then the general solution
c,y,
f
W
The
y
c
= c,y, +
+
+c
n
yn +
u,
(4)
where
v
c
2,
,
cn
are arbitrary constants.
fact that (4) is a solution of (1)
may
be verified by substi
tution.
The
fact that
it
is
the general solution
we
shall accept
without proof.
It appears
now
ular integral of
that the complementary function and the partic 192 are only special cases of (4). There exists,
(4)
however, no general method of finding the solution coefficients of (1) are not constant.
when
the
Methods
of solution
may, however,
exist in special cases,
and
we
shall notice especially the equation
a n are constants. This equation has the , a n _ v that each derivative is multiplied by a power of x peculiarity It can be reduced to a linear equal to the order of the derivative.
where a v a 2
,
equation with constant coefficients, by placing
For
dx
..,. dx
dz
dz
dz
dz~
dz&gt;
dz
dz\
dz ]
2
dz\
_dz
~
~dz*
d?
+
"
C
~dz
where
D=

dz
356
Hence
THE LINEAK DIFFERENTIAL EQUATION
x
ctx
= Dy,
dx s
Placing x
dx 2
dx
=
ez
,
and making the substitution as above, we have
whence
y
=
GI
+
c2 z
+
c s e~ 2s H
e2z
=
GI
+
c2
(logx)
+
H
:
z2
.
196. Solution
by
series.
The solution
of a linear differential
equation can usually be expanded into a Taylor s or a Maclaurin s series. This is, in fact, an important and powerful method of
shall limit investigating the function defined by the equation. ourselves, however, to showing by examples how the series may be obtained. The method consists in assuming a series of the form
We
where
and the coefficients a Q a v a 2 are undetermined. This then substituted in the differential equation, and and the coefficients are so determined that the equation is identically satisfied.

m
,
,
series is
m
We assume a series of the form given above, and write the expression for each term of the differential equation, placing like powers of x under each other. We have then
d 2y
(mtr)a,rX
m+ r +
3^ = dx

Adding these results, we have an expression which must be identically equal to zero, since the assumed series satisfies the differential equation. Equating to
zero the coefficient of x
1 "
1
,
we have
m (m
4)
a
=
0.
(1)
SOLUTION BY SEKIES
Equating
to zero the coefficient of
357
xm
,
we have
(m
+
1)
(m

3) ai
+
(m

2)
a
,
=

0.
(2)
m+r Finally, equating to zero the coefficient of x
relation
(
we have
r
the
more general
(3)
m+r+
1)
(m
+
r
 3)a r + 1 + (m +
2)a r
=
0.
We
shall gain nothing
as the first term of the series.
m is assumed by placing a = in equation (1), since Hence to satisfy (1) we must have either
a&lt;&
m=
Taking the
first
or
m=
4.
of these possibilities,
ai
namely
2
m=
7T a ro)
0,
we
have, from
(2),
=
jj
ao,
and from
(3),
r
+
i
=
(r
r
+
(4)
1) (r
This last formula
(4)
the previous one, a,.. Thus we find a 2 coefficients after as equal to zero.
enables us to compute any coefficient, a r + i, when we know = J a 1 = 1 a a 3 = 0, and therefore all
,
Hence we have as one solution
of the differential equation the polynomial
take
Returning now to the second of the two possibilities for the value of m, we m = 4. Then (2) becomes 6 ai + 2 a = 0,
(3)
and
becomes
ar + i
=
(r
42
+
5)(r
+
r
(6)
l)
Computing from
have the solution
this the coefficients of the first four
terms of the
series,
we
\
5
(7)
66
567
/
We
have now in
Hence, by
(5)
and
two independent solutions of the
y
differential
equa
tion.
195, the general solution is
= 
Ciyi
+
C2 y2 .
Ex.
2.
Legendre
s equation.
(1
x2)
ax 2
2x
ax
+ n (n +
1)
y
=
0.
Assuming the general form
d 2y dx 2
2
of the series,
f
we have
1
f (in f
= m (m
1)
aox m
~2
Ow
+
1)
wiaix"
2)(m
+ 1) a 2 x m +
,
d y ~^~=
2x
m(ml)aox
2
,
=
dx
waox m
,
n(n
358
THE LINEAR DIFFERENTIAL EQUATION
to zero the coefficients of x m ~
2
,
Equating
x
"
1
,
and x m we have
,
m (m
(m

1)
a
= =
0,
(1)
(2)
(m
+
2)
(7/1
+
1)
a2
+ 1) mai 0,  (m  n) (m +
n
+
1)
o
=
0.
(3)
To find a general law for the coefficients, we will find the term containing x m + r2 i n each of the above expansions, this term being chosen because it con tains a r in the first expansion. We have
zjfss
dx 2
\(m
+
r}
(m
+
r

1)
a,.x"
+r2
+
,
_
X
2^ =  (m + r  2)(w + r 3) a
dx 2 dx
r
_2X"
+r2
,
,
_2x^ =
n(n
 2(7M + r2)a r _ 2 x +
"
2
+ l)y=
+
n
(n
+
l)a r _ 2 x
nt
+r2
+
.
The sum
(m
of these coefficients equated to zero gives
+
r)
(m
+
r

1)
ar
 (m  n +
placing
is
r

2)
(m
+
n
+
r

1)
ar 
2
=
0.
(4)
take
We may satisfy (1) either by m = 0. Then, from (2), a\
m=
;
arbitrary
or by placing from (3)
m = 1. We
shall
n(n
+
r
1)
lf.\
and from
(4)
ar
=
(n
+
2)(n
+
r
1)
_
ar _ 2
.
(6)
By means
of (6)
we determine
the solution
j
^
[3
(nl[
^
[6
\
we have in (7) the general solution of the dif In fact, the student will find that if he takes the value the second series in (7). ?7i = 1 from (1), he will obtain again Particular interest attaches to the cases in which one of the series in (7) reduces to a polynomial. This evidently happens to the first series when n By is an even integer, and to the second series when n is an odd integer.
Since
and
a\ are arbitrary,
ferential equation.
giving to a
or a\ such numerical values in each case that the polynomial
SOLUTION BY SERIES
is
359
(7)
equal to unity
when x equals
unity,
we
obtain from the series in
the
polynomials
42
PS
each of which
coefficients.
42
2
=
07 5 x, 42
75 s x, 42
+
H
53 42
x,
satisfies a Legendre s differential equation in which n has the value indicated by the suffix of P. These polynomials are called Legendre s
Ex.
3.
BesseVs equation,
series for
x2
^
+
x
(x
2
 n2
)
y
=
0.
Assuming the d 2y
dy
y
in the usual form,
we have
+
(m
+
l) ai x
+i
+ m
Equating to zero the sum of the coefficients of the have
first
three powers of x,
[(m
+
2)
2

n 2 ]a 2
+
a
=
0.
(3)
To
obtain the general expression for the coefficients,
x2
we have
m+r
^2
dy
2
?/
=
+
(m
+
r)
(m
f
r
l)a,.x
+
,
*=
n
".
+
&lt;m
+
r;
=
Equating to zero the sum of these
[(m
coefficients,
we have
+
r)2 n 2]a r
7/1
+
ar _ 2
=
0.
(4)
Equation
(1)
from
(2), (3),
be satisfied by and (4) we have
may
Oi
=
n.
We
flt
will take first
m=
n.
Then
=
0,
a2
=
^
2(2 n
.
+
2)
r(2n
^a +
r)
360
By
THE LINEAR DIFFERENTIAL EQUATION
we
obtain the series
use of these results
Similarly,
by placing
1 4.
m=
2(2
(
If,
n_
T2
2)
n,
we
\
obtain the series
y4 *
2
4 (2n

2)(2n
4)
\
.
)
(6)
/
now, n is any number except an integer or zero, each of the series (5) and converges and the two series are distinct from each other. Hence in this case the general solution of the differential equation is
(6)
If
n
=
0,
the two series
(5)
and
(6)
are identical.
If
n
is
a positive integer,
series (6) is meaningless, since some of the coefficients become infinite. If n is a negative integer, series (5) is meaningless, since some of the coefficients become
Hence, if n is zero or an integer, we have in (5) and (6) only one par ticular solution of the differential equation, and another particular solution must be found before the general solution is known. The manner in which this may
infinite.
be done cannot, however, be taken up here. The series (5) and (6) define new transcendental functions of x called BesseVs functions. They are important in many applications to mathematical physics.
PROBLEMS
Solve the following equations
1.
:
8^ + 2 V = 0.
dx
10.^ + 8^ + 16^ = 0. dx dx
2
2.
j
3.
+ 8y = a;* +
Jsinx.
11.
+ 9y =
dx*
0.
^2y = e^+ecosx. dx
12.
^  6^ +
dx
13y
=
0.
4.
^ + y = 8e* +
dx
dx
aj^.
13.
? + 2^8y=*dx dx*
dx 2
y
5.
+
4y
= G sm 2x.
+
14.
=
8.
dx
.!,
= _?__ x x
e
15.
e
16.
8.
dx2
+
2VSy = dx
5
0.
17.
9
^&gt;_
^=
dx
0.
18.
^ + 8^100 = ^ + 3^47/ = ^ + 3^ ^ + ^dx 2 dx
6
dx2
dx
dx 2
dx
dx 2
dx*
dx
PROBLEMS
361
21.
(X3)2
5e
23. *UL
r o 2:c
dX 3
o
dx2
dx
.
sm3x.
An d y 42. 3
j
d y ~
2
= x8
24.
^ +
d 2 ?/
y
=
2sin6zsin3x.
fa
nf,
26
dx 2
+
3

x sin 2
dx
d7
+
dy
di
rrt
(ia.
d"V
f
2
^C??/
^
+o
?/
x 2 e~ x 4
,1
e 3 *.
/7r
46
dy
=2x
on
^ 2 ?/ d 3^
.
^2/
O1 1
o
d 2y
^rO.
47
/7r
OT
?/
IS
oft 32
^ + dx =
?/
^
3
^y
dv

= 2a: +
23

10

(
dx4
J
=
dx 2
2
dQ
dy
37.
1
y
=
te*
+
fl^rnsrr
gQ
"
^
_
q2
__
362
51.
THE LINEAR DIFFERENTIAL EQUATION
#Sx dx + dx
2
a
V
=
0.
54. x
dx 3
+ 8*
dx 2
+
*
dx
52

*+"
+
v="&lt;**.
56. x
53
*.ft + *&lt;**
means
0.
+ 3x2
:

Solve the following equations by
57.
of series
a*^ + (x2x2)^0y =
dx 2
60.
dx
^ x^ + dx + dx
x2
2
(x

6)y
=
0.
59.
(xx
dx2
+
5
dx
+
6^0.
62.
(
l
+
dx 2
+
x W =
dx
0.
63. particle of unit mass moving in a straight line is acted on by an attracting force in its line of motion directed toward a center and proportional to the distance of the particle from the center, and also by a periodic force equal
to a cos
kt.
A
Determine
its
motion.
particle of unit mass moving in a straight line is acted on by three an attracting force in its line of motion directed toward a center and proportional to the distance of the particle from the center, a resisting force proportional to the velocity of the particle, and a periodic force equal to a cos kt. Determine the motion of the particle.
64.
A
forces,
65.
Under what conditions will
?
oscillations the amplitudes of
the motion of the particle in Ex. 64 consist of which increase without limit as the time increases
without limit
CHAPTER XIX
PARTIAL DIFFERENTIAL EQUATIONS
197. Introduction.
tion
A partial
differential equation is
an equa
which involves partial derivatives. The general solution of such an equation involves one or more arbitrary functions. Thus
z
f(x
y,
y
x)
is
a
solution
is
of
the
equation
h
vx

=
2
(
113, Ex. l),no matter what
f(x\at)\f^(x
cy the form of the function/. Also
z=
(
at) is a solution of the
equation
ct
=a

ex
.
no matter what are the forms of the f unctions /x and/2 Only in comparatively few cases can the solution of a partial In general, the differential equation be written down explicitly.
1 18,
Ex.
2),
nature and the properties of functions denned by such equations must be studied by the methods of advanced mathematics. In a
practical application, the problem is usually to determine a func tion which will satisfy the differential equation and at the same
time meet the other conditions of the practical problem.
198. Special forms of partial differential equations.
differential equations
Partial
sometimes occur which can be readily solved successive integration with respect to each of the variables, or by which can be otherwise solved by elementary methods. No general discussion can very well be given for such equations, but the fol
lowing examples will illustrate them.
Ex.
1.
5L =
dxcy
0.
By
integration with respect to y,
we have
,
where
X
is
an arbitrary function.
z
Integrating with respect to x,
we have
=
02 (z)
+
03 (y),
where both
2
and
8
are arbitrary functions.
303
364
Ex.2.
If
PARTIAL DIFFERENTIAL EQUATIONS
ax 2
?Ua%.
this equation
x were the only independent variable, the solution of
z
would be
=
GI sin
ax
+
c2
cos ax.
This solution will also hold for the partial differential equation if we simply impose upon Ci and 2 the condition to be independent of x but not necessarily independent of the other variables. That is, if z is a function of x and y, we have for the solution of the differential equation
c,
z
=
0i (y ) s in ax
+
2
(y) cos ox,
where 0i(y) and
2 (y) are arbitrary functions.
te.a^^=o. ax
a?/
2
2
Placing x
+
ay
=
M,
and x
ay
=
2
u,
we have
(
***_.*.*;+ aw
a?/
2
awai?
A
118)
au2
e*z_s2z
az 2
"^
a 2z
2
2
and the
differential equation
becomes
the solution of which (Ex.
1) is
Z
=
01 (U)
+ + +
1
2 (fl).
Hence the solution
of the given equation
z
is
= =
0! (Z
&lt;f&gt;i(x
ay)
iy)
+
2
(x

ay).
When
a
2
=
1,
we have
z
+
0.
2
as the solution of the equation
  =
ax2 ay
199.
The
linear partial differential equation of the first order.
Consider the equation
*
where P,
Q,
+
*
more
of
of (1) is a function (2)
and
x, y,
R
are constants or functions of one or
z.
the variables
and
The solution
*
=/(,
it
y),
which, substituted in
(1),
reduces
to an identity.
LINEAR EQUATION OF FIRST ORDER
Now
of
365
equation (2) represents a surface, the normal at any point
: :
which has the direction
normal
1
(
112).
Equation
is
(1) there
fore asserts that the
to (2) at
any point
perpendicular to
P\Q:R (98, (5)). Consequently we may start from any point on (2) and, moving in the direction P:Q\R, remain always on the surface. That is, the surface (2) is covered by a
the direction
family of curves each of which
is
a solution of the simultaneous
equations
dx
_dy _dz
and hence we reach the conclu
~P~~Q~ R
Now
(2) is
any solution
of (1),
sion that the solution of (1) consists of all surfaces which are cov ered by a family of curves each of which is a solution of (3).
We may
(3),
obtaining, as in
proceed to find these surfaces as follows 185, the solution
:
Let us solve
u (x,
Then,
if
y, z)
= cv
v
(x, y, z)
=c
2
.
(4)
we form
the equation
&lt;l&gt;(u,v)=Q,
(5)
where
any function whatever, we have a surface which is covered by curves represented by (4). For if in (4) we give c x and = 0, the corresponding curve lies on (5). c such values that c 2 (c v 2)
&lt;
is
&lt;
by means of (5) we have assembled the curves (4) into surfaces, and have therefore the solution of (1). We may formu That
late
is,
our result into the following rule
solve the equation
:
To
P
dx
+Q
dy
=R
y
solve first
7
n
.
,
7
.
dx
tlie
equations
P
u
= dy = dz
Q
v
R
for
the solution
= cv
$ (u,
Then
where
the equation
is
&lt;f&gt;
= cr = 0, v)
solution of the partial dif
an arbitrary function,
is the
ferential equation.
366
Ex.
PARTIAL DIFFERENTIAL EQUATIONS
(ny
mz)
C 
ex
+
(Iz
nx)^ = mx cy
ly.
(1)
We
form the equations
ny
mz
=
Iz
dy nx
=
_^_ mx
ly
,
(2 )
the solutions of which are
x2
+
y
2
+
z2
=
cb
Ix
+ my +
nz
=
c2
.
(3)
Hence
the solution of (1)
2
is
(x
+
2
?/
+
2 2 , te
+ my +
712)
=
0.
(4)
Geometrically, the first equation of (3) represents all spheres with their cen ters at the origin, and the second equation of which represents all
(3)
planes
are normal to the line
(3) taken simultaneously represent all circles whose centers are in the line (5) arid whose planes are perpendicular to Equation (5). (4), then, represents all surfaces which can be formed out of these circles that is, all surfaces of revolution which have the line (5) for an axis. These surfaces
;
Hence the two equations
of revolution
form the solution
of
(1).
200. Laplace s equation in the plane. equation in the plane,
Solutions of Laplace
s
have already been found in 172, 198. Another method of Let us place V = XY, dealing with this equation is as follows where is a function of x alone and Y is a function of y alone, and ask if it is possible to determine and Y so that s
:
X
equation
may
be
satisfied.
Laplace Substituting in the given equation
X
and dividing by XY, we have
X
which may be put
L** + 1^1=0
"
dx 2
Y df
in the
form
~
X
dz?
Y dy
2 
According to the hypothesis, the lefthand member of contain y and the righthand member cannot contain
(2)
x.
cannot
Hence
LAPLACE
S
EQUATION IN THE PLANE
367
.
a2 they are each equal to some constant, which we will denote by Then (2) breaks up into two ordinary differential equations,
a*Y=0.
ay
(4)
By
190, the solution of (3)
is
and the solution
of (4) is
Y= B
Hence
l
cos
ay
f
I&gt;
2
sin ay.
V e ax cos ay, V = e~ ax cos ay,
V = e ax sin ay, V = e~ ax sin ay,
a2 we should
,
are particular solutions of the given equation.
the value of the constant had been denoted by have obtained the particular solutions
If
V = eay cos ax, V e~ ay cos ax,
V = eay sin ax, V = e~ ay sin ax.
are, in
fact,
iaz
,
The particular solutions thus obtained
the imaginary parts of the functions
e
az
,
the real and
e
iaz
(
e~
az
,
e~
and
172).
The sum
plied
by an
of (1), each multi constant, is also a solution of (1), as is easily arbitrary
of
two
or
more particular solutions
verified.
Hence we may form the
particular solutions
F=A
, f
VA
m=l
e~ mv sin
mx + 2,Bme~ my cos mx
m=l
t
(5)
m=l m= oo
Q
m= oo
V = A + ^ A memy sm mx +
solution (6) becomes infinite with y.
Bm e my cos mx.
rn=l
(6)
Solution (5) has the property of reducing to
A when y =
oc,
while
Ex. Find the permanent temperature at any point of a thin rectangular plate TT and of infinite length, the end being kept at the temperature unity and the long edges being kept at the temperature zero.
of breadth
If
u
is
the temperature,
it is
known
+
that u satisfies the differential equation
368
If
PARTIAL DIFFERENTIAL EQUATIONS
take the end of the plate as the axis of x, and one of the long edges as we have to solve (1) subject to the conditions
:
we
the axis of y,
if
x X
if
if if
= =
0,
TT,
u
=
0,
&lt;&gt;,
(2)
(3) (4)
*
y
=
=
0,
,
= U = 0,
u
=
oo
1.
(5)
Condition
(4) is satisfied
by the solution
m= oo
M
m=
my sin ?nx
f
=
m=l
^T A m e
^
7^ m e~
W cos rax.
2/
=1
By
condition
(2),
=
m=l
and hence
for all values of y,
Bm =
,
for all values of m.
Our
solution
is
now
reduced to
= *,
m ^ sin
rax,
w
which
satisfies (2),
=
^V 4 m e~ m=l
(3),
and
(4).
In order that
it
determine the coefficients
A m that
m sin mx.
may
satisfy (5),
we must
so
(6)
m=l
But
.
(6) is
a special form of a Fourier s series
(
159, 160).
to. TT.
multiply
TT (
2 /I
 (6)
Accordingly,
result,
by
sin
mx dx, and
whence
)
integrate
from
2
cos m7r\
m
A =
I
4
n
=
AS = *
5
41 ~
As a
we we have
Therefore
 e~^v
sin3x
3
is
+ ~e~ 5 v sin 5x +
the conditions.
)
/
the solution of our problem, since
it satisfies all
201. Laplace s equation in three dimensions.
The general form
of Laplace s equation in rectangular coordinates is
If cylindrical coordinates are used, (1)
becomes
"
VV_
8r*
+
r~fr
(1)
\W +
ld*V
7
W +,W_ d?d&lt;f&gt;
and in polar coordinates
becomes
dr
dr
sm&lt;j&gt;ci&lt;l&gt;\
LAPLACE
of this text.
(3) in
S
EQUATION IN THKEE DIMENSIONS 369
of these equations is
The general discussion
beyond the limits
We
the special
0.
of shall, however, consider particular solutions in which V is independent of the diedral case
angle
In that case
a /
( 2
=
0,
and
(3)
.
reduces to
,
r 0r
Letting V function of
r
"
ar ~
0r
i
a /
sm
dv
a
=
7t&lt;,
where
R
is
a function of r alone and
4&gt;
is
&lt;/&gt;
alone,
we may
dif replace (4) by the two ordinary
ferential equations

 a?R =
)
0,
dr /
sin
2 where a
(5)
(6)
(j)
d(j)
\
d(f&gt;
an arbitrary constant. Expanding (5), we have
is
the solution of which
is,
by
195,
where
m=
J
+ V^ + iand
(6)
2
From
this
value
of
m
we have
a =
2
m (m + 1),
becomes
sin
(j)
d(f&gt;
Changing the independent variable from we have Legendre s equation ( 196, Ex. 2),
to
(f&gt;
t,
where
t
= cos
&lt;,
In the particular case in which
4&gt;
m
is
an
integer,
we may
choose for
Legendre
s coefficient
JJ(Q=.(C08^).
Therefore the particular solutions of (4) are
370
PARTIAL DIFFERENTIAL EQUATIONS
coordinates,
s
If the center of the ring is
Ex. Find the potential due to a circular ring of small cross section and radius a. taken as the origin of and the axis
is
OZ
perpendicular to the plane of the ring, Laplace
*
/
equation assumes the
(1)
form
*v\
i
a  /sin0 aFX _)=0,
since,
from the symmetry
of the problem,
V is
independent of
6.
This equation
is satisfied
by
V=
where
(2)
Am
and
At any
Bm are arbitrary constants. point on the axis OZ distant r from the origin
VOa
where
2 f r
(3)
is the mass of the ring (see Ex. 9, p. 90). = 0, cos0 = 1, and by 196, Ex. 2, Pm (cos0) = Then, when same time the righthand members of (2) and must be i.e.
M
1.
At
the
(3)
equal
m=
and the
coefficients
Am
and
Bm =
must be chosen
so as to satisfy this equation.
If r&lt;a(31,
Ex.4),
w
Va2 + r 2
_
\
2 a2
2
4 a4
andifr&gt;a,
V2 + r
we
6tC
J^ = ^
a
a
(
_
1
a \r
s
2 r3
g + LJ g? _
2
.\
/
4 r5
Hence
if
r&lt;a,
place all the
equal to zero,
A 
,
2
^=
0,
^4 2
=
~a a
an
and
"
if
r&gt;a,
we
,
place all the ^4
s
equal to zero,
B =
.a,
B =
l
 
0,
B2 =
a3
,
and obtain the solution
PROBLEMS
Solve the following equations
1.
:
=
a* 2
a*z.
3.
^a dx = W
d
ez
0.
5
^ + 4^w ^

Z
x
^ = 2 sx 9
a
ax
^
4

?x 2
5
ax
+ 62 =
0.
6.
=
w. y
axay
PKOBLEMS
7.
371
t
dxdy
= +
ez
x*
+
!.
7/2.
13.
ex
ly
ez _ 8. ex
,
_=
ly
,,
14.x
15. x
ez
ex
0.
\y
z
=
z.
cy
9.
x
y = ex cy
ex
*xy ex
+
ey
2
?/
=
0.
10.
2ZX + 2Z2/ =
ay
Z2.
16. XZ
K
ex
+y Z EZ =Xy.
cy
ey
(x
cy
ex
ex
to
10
&lt;
z
12.
y
 x
cz
eiy
=
CZ
1.
ex
18. y
ex
+
+
^
CZ
z)
=
y.
cy
19.
Find particular solutions of the equation
t.
nometric functions of x and
^= ^
a2
in
terms of trigo
20. Find particular solutions of Laplace s equation in the plane in polar
,.
coordinates
d*V cV F    =
1
1
c2
1
1
0.
cr2
r dr
r2 c^ 2
21. Find the permanent temperature at any point in a semicircular plate of radius unity, the circumference of which is kept at the temperature unity and the bounding diameter of which is kept at the temperature 0, given that the temperature u satisfies the differential equation of Ex. 20.
where u is the ex 2 and x is measured parallel to the direction in which temperature at any time the heat flows. If a slab of thickness TT is originally at the temperature unity
linear flow of heat
is
,
22.
The equation for the
,
=
a2
ct
ture at
throughout, and both faces are then kept at a temperature 0, find the tempera any point of the slab, the slab being so large that only the flow of heat
to its
normal
bounding faces need be considered.
ANSWERS
(The answers to some problems are intentionally omitted.)
Page 33
CHAPTER
II
12. log(logx).
13. log(tanix).
19.
_
log
(a
+
6cosz).
24. log
20
4(1
:
25.
x2)2
+
 _
1
(1
Ve 2 * + tan2x.
n) [log(x

+ o)]i
21
26

tanixl!.
27.
Page 34
28.
Oa
29.
30. x
(
2 e *
+
6)3.
37.  (esc bx b 38.


ctn bx).
 2 Va 2  x 2

.
ctn (ax a
+
6).
31. 32. 33.
,
34.

=(VoxVte)*

=
x  sin 4x. 43. log(cscx ctnx) 44. C os3x.
42.
+
2cosx.
2
&lt;
35
36.
? sin4x
45.
6)
tan x
+
ctnx.
x.
L + 6a [sine (ox
 cos
(ax
+
46&gt;
6)1
47.
 sec 3 x) 2tanxsecxx.
tan 3 x
ANSWERS
98.
i
375
99.
~
V2
log(2x+V4x23).
log (2 x
101.
1
102.
*2
+ 2V2
a  Vz* 
100.
3xV2
.
103. log (x

tan
2 x tan
a
l)
Page 36
104. lo
105. log (x 106.
+
sin
a + Vz 2 +
2 x sin
a+
l).
2
107.
V3
3x

3
+ V3
2
V5
3
108. ilo
x5
 6x +
3).
109.
\/3
110.
111.
v2
112. 
log(2x
+
2

10).
lo
113. i
114.
115. 2 116. 3
117.
+ x f 2  4 log(2x + 1 + 2 Vx2 + x + 2). Vx2 + 4x  1 + log (a + 2 + Vx^T^^).
Vx2

2
V2
2+V3
119. 5 sini ?
V5
120.
+ 2 Vl 
4x

x2.

121. ~ Iog(4x2 122.
+
4xsec
2
+
tan a)
_
^
Vs
log
2^+sec^
3Vl2xx
376
123.
ANSWERS
1
,2x43
124. i log(x 125. log(x
log a
2
+ Vx 4
a
4
)
2V5  i sec ~
1
V5
.
f
Vx 2 
l)
+
secix.
131.
132.
)
4 log a
127.
1
128. 4 log a
(
a2x
2
_
a 2 *
133.
c(l
129.
+
log a)
134.
2
n log a
~l
&lt;
2m log 6
n log a
+ m log 6
135.
1
136.
137.
log a
log (a*
4 1
+Va + 2 a* 4 sec
2a:
2
a).
138.
140.
2x alog(e4l).
x
(e
141. 2 log
+
1)

x.
pax
152.
sec 1
.
ab
153. x
154.
155.
63
_b
1).
4
sV
+ 4 Vx 4 41og(Vx 2  6 x 46) (2 x (5 x
log(a
3)1
+
bx)
+
156
157. 3 158. isec 1 .
a
a
159. T
160.
161.
V Vx2 + a2 (3 x4  4 a2x 2 + 8 a4 ^x4  3V 14 x 3  5
)
.
49(5
162.
7x
+
3 2
)
4(1
2x2
x2 ) 2
4
4
163. log
151. J (x 2
Vx 2 +

2 a2 )
Va2 + x2
.
164.
179.
(sin
ax
ax cos ax).
183
181.
.
2 2 * (a x sec ax
 Va 2x 2 a3 ax
l).
x2
.
sin
185.
1 86.
ax

2 x cos ax
2 sin ax
.
a
a2 a2x 2
2

[(3

6) sin

3 3 (a x

6 ax) cos ax].
187. x [(log ax)
188. 27 * 3 [9 (log ax) 2
189.
190. 191. 192.

8 a2
2 log ax + 2]. G log ax + 2 J2x2  2 ax sin 2 ax  cos 2 axl. [2 a
(2
^e
^e

2:r
+ cos2x +
sin2x).
^ex [5(cosx + sinx)  (cos3x + 3 sin 3 x)].
sx 4 a2
193.
[^(3sm4x  4 cos 4 x) + T 3 (3sin2x 2 2  a 2 x 2 ]. [(2 a x 1) sin ax + ax Vl
1 1
2 cos 2 x)]
378
2
ANSWERS
a 2
a2 log
x
+ Vz 2 
a2 \
)
23. 24.
21

*
2  log
22.
.
36.
.6366.
25
2
5

e
Page 62
35.
IT.
Page 63
46. .4621. 47. .016. 48. .8746.
CHAPTER IV
Page 80
1.
170 f.
2.
a
2
.
3.
22.
67r

23.
24.
2
7ra
.
25. 5?.
27
(3
a
/
00 1 Q 33. i7ra d \e a

I
e
+
Tra 2 h.
97
34. 47ra 3 (21og2
35.
1).
f
7ra 3 (31og2
2).
36. ITT (a 2
4_

ft
2
)*.
^^3
.
772^25
_
4
Tja3
^
i
s
the distance from the diameter of
semicircle to the axis).
38.
4 Trafr2
.
42.
39. 27T 2 a6cL
73
40.
3a2
41.
^
43.
(3
a
2
+
ft).
44 44.
_
T^ 5
^^(^
ANSWEKS
Page 83
45.
37 J
_5A_ (3 h* +
4Q pah*
is
+ 240 pW).
46.
]
vpW (5 a  3
is
ft).
47. 1 7ra 3 (l
of the cone).
cos a:) (a
the radius of the sphere
and 2 a
the vertical angle
* a&2
48. 07r(l
54. 55.
56.
+
OVOTr).
z
49.
rra 8
.
50.
?,
Tra 3
.
53.
.
a3
(TT
tan
(a is the radius of the cylinder).
is
%)ha (a
2
the radius of the cylinder
and h
is
is
the altitude).
4 /m (a
4
/
1
is
the radius of the cylinder
and h
the altitude).
57. 2rrh 2 58.
Vpip 2 (pi and p z
are the respective parameters of the parabolas).
a
8
3
\
)
+
7T\

1
(a is
the radius of the sphere).
2/
Page 84
59.
6
7ra 2 6.
is
60. T Q4j a 3 (a
the parameter of the hypocycloids).
67.
,
61.
a& 2
.
a } Uej.
(

71

8a

78
68
63. 1152 cu. in.
64. 65.
.
l
73&gt;
73
2
1
^[(4 + 0A)8].
log(e
4 (a 2 
+
a
a&
2 f 6 )
74
177
r
in
ft.
11+6
+
75. 1007
ei).
66. Ga.
Page 85
77.
* a.
78.
7ra.
79. 80.
80. 27ra(h 2

hi) (a
is
the radius of the sphere).
84.
81.
irVA +
a
.
&gt;**.
iTra 2
.
82
x
,

gg
83. 27ra 2
87. 27T& 2
+
2.


siuie
(e is
the eccentricity of the ellipse).
89. 47ra 2 (2
88.
3/7ra
V2).
CHAPTER V
1

7
(Hs the length of the wire).
C(C+f)
380

ANSWERS
r sinC
(I
is
the length of the wire) in a direction bisecting the angle
10.
0.
9.
^/I a
2
\c
v^Ttf/
L=Y
11.
^ 2  V(c +
O
2
+
a 2 ).
Page 100
12.
,
sm
2
.
21. 22.  a below the surface. 2 23. w.
14.
2ac).
3 (a
^ba
15
+
below the
2
c)
I a below the surface. 25. 2 a6 2 iy (the axis 2 a being in the

24
surface.
16. 18.
surface). 26. fir nb below the surface. 27. 2250 Ib.
20.
28. 781.25 Ib.
Page 101
29. 2234 ft.lb. 30. 769.4 Ib.
31. 33. 88.4 tons.
36.
(fa, fa).
(0,
34.
35.
41 tons
37.
38.
fa).
104 If tons.
32. 8333 _L foottons. 39.
On
(
40.
\5 4 / the axis, three fifths of the distance from the vertex to the chord. k, 0)(x = k is the equation of the ordinate).
42.
7T
k ^}.
V
2
(If,
^.
43

(t
,
\3
3
7T,
,5
57
Page 102
48.
\5
**
*} m2 m I
9 9
51.
176 ?
fp,
V
TT)/
5 (2
+
3
46. 47. Intersection of the
49.
52.
/i^,ii^&)
r
3?r
50.
medians.
53.
(, \2 2
+ ^1.
12p/
the radius perpendicular to the base of the hemisphere and three eighths of the distance from the base. 54. On the diameter of the sphere perpendicular to the planes, at a distance \ (hi + hz) from the center.
55. x
On
58.
= y = =
a.
56.
(y
is
y=
fc.
57.
x
=
5 T^ 6
2
(6
=
4 pa).
^ yi
b.
the ordinate of the point of intersection of the line and the
parabola).
59. 60.
y
fir
At
the middle point of the radius of the hemisphere perpendicular to
the base.
ANSWERS
61.
base.
62.
381
On
the axis of the cone, two thirds of the distance from the vertex to the
On
of the distance
the radius of the hemisphere perpendicular to the base, two thirds from the base to the vertex.
CHAPTER VI
Page 117
21. I log(4x 2 22.
23.
+ 4x +
log [(x
2)
+
3
zJx2
1x2
Jx
tani(2x
3)6].
+
1
1).
!)( +
+
1
Iog(9x2
6 Iog(x2
+
12x
+
8)
+
4
24.
+ 4x +
 2x 
1)
+
22 ~^
tani^t?.
V2 log^
x
~1  1 + V2
9.
Sx
loff
(X
1

2)
)*.
382
30.
ANSWERS
x2
2
 2x +
log (3
1
+
1)
+
1
1 log(2x

1)
+ V lo S(x +
3 )
31.
^ (x
+
x)
+
+
3)
32. Iog[x 3
(2xl)]
X
33. log(x
+
2x
2) f
(x
+
2)2
34.
log[(xl)V2x
2 (2 x
+
3)
Page 118
35.
2x
log (x
+
2).
38. x
+

oX
+
o
log(3x2).
39.
x&gt;
+
logx.
i^ z + 8 ^rij
Vx 2 43.
1
.
4x
V2
2+V2
44.
a
.
+
lOg
6
A
2 /x \x 2 +
6
a
1
h
x2 5 45.  log 2 3x 2
46.
2x+3
x2
+ +
2
= tan *
2
^
,
*
1
3
05
2
3
V2
V2
.
VO
^
5
V6
\
47. x
1
5 2 x
2
;
f
x
+
H
tan
1
V3
+
2)
V3
3 tan1
V^9 1)
(X
2
g log (x


(x

tan
1
(x
+
1).
48.
2X)+1
/
(2x
+
4
3)
4x3
3V3
49.
V3
50.
/
o
x
2
+ V3
/
/
V61
V51
tan
x
2V3
51.
V3
+x 2 (x 2 + 3)
4
52. 3 log
2V2
+
17
x
2x +
1
3
3V3
g
13
53. log
4V2
xV2 xV2 +
_V3
~
3
+ 7x
2
l
2(2x l)
ANSWERS
CHAPTER
137
VII
383
1
+
x
+ 4 V1 +
/
x
+
2 log (x
f
,  V1 +

N
.
x)
+

= log
14.
(8x61) (3x6 +
1)8.
17.
18.
+ x 4 + log (x 2 + \/4 + x 4 ). + 3x  2 /V2 tanV2
4
,
1
19.
V5
1
log
x
.
20.
lo
22.
V3 Vx 2 + 2x +
2(1
V2 + V2 x 2 + 3 x + V3 + 3XV3X V3 + 3x+V3x
3
5
+ V5
21.
2(2x
+
h3).
24.
x
+
1

log(l
+
x
23.
+
3x
6x)
/l3x
2(28x
3
12 x
8)
243(1 x2x 2 )^
17x6 4=tani Jt2f?. 25 V2 3x2x 2 \l2x V2
^
sin 6 x
7  sin x.
27. cos^3x( T1T cos8 3x

cos3x).
Page 138
28. I cos x

cos x
+
cos x

cosx.
x
+
gin
x\ /.
n
2x _
\
29. sinx
30.
31.
^ 2 v Icos(2x+l)[cos (2x+l) 31.  ] ros4x x
,
sin 8 x
+
sin 6 x.
1
33.
a \4
(sm 4 ox
/I
.
\ sin 6 ox).
.
1
384
ANSWERS
38.
A Vco
8/
39.log
X
2*(cos*2z7). 1+ sinrp
S
sin82sin. n
*
f\
41
"i .
1sin2
Sln 8 sin
8
1 l
*! lOg
lcos2x l+cos2x
 3
cos2x _
4sin22x
42. 43.
i
\
12cos*3x
tan 2 3 x
+
j*JL JL] S 8cos 2 3x 16 i log cos 3 x.

sin 3
x
44.
45.
1.
log sin 3 x
+
l ctn 2 3 x
 TL
ctn* 3 x.
^an3_2tan
46.
47.
50.
o1
+
x.
ctn*+ctn8ctnaf. ^tan^x  ctn^x) + 2 log tan*.
^
4
"
Ctn
3
l (
+
.
4cos22x
sin2x
1
^ctn3x(l+
8
ctn23x
+
ctn*3x+
lsin2x
l
rf
sin2 sin2x
Q
52
5 cos5
15 cos !?
4sin 4 5 53.


~
x
5
+ ^io
16

ctrt
54.
(sinax a\
_
+ 2cscaxicsc 3 axV
8/
55.
52
sec 5 .
ANSWERS
66.
385
i [x Vx
2
+
a2

a 2 log(
+ Vx 2 + a 2)].
x
2
67.
2 a3
Vx +
,
a2
+
a
68.
2a 2x 2

1
x
8
70.
x(2 x
2
+
a2 )
Vx 2 +
a2
8
log(x
+ Vx 2 +
a2 ).
8
71. 72.
2 2 2 2 x(2x  a ) Va  x
+
8
sini. a
75.
 TV(3x2 + 2a 2 )(a2 v
2 (a2
i

(1
+
x2 )
+
2
a
76.
x
73.
77.
78.
 V2 ax
V2ax x2
x2 x a
74.
+
a sin 1
79.
J
(2
x2
(x
+
x
1
ax

3 a2 )
V2 ax x2
x2
+ sini^^.
smi*^. + y ^ sin x cos x +
1
80.
81.
82.

i
3 a)
cos5

V2 ax x
+
^
sin
+
TjL
si n
x cos 3 x
.
x.
sin
x cos 2 x
+
3 sin x
2cos2 x
+
3 log (sec x 2
+
tan
x).
83.
84.
86.
tan 3
2 cos 3
3x+
2x3cos2x 3. .
2 sin 2 x
1
85.
4 cos4 x
4sin 2 2x
4
log (esc 2 x
ctn 2 x).
Page 139
/
87.
15 h?
2 a5
+
2 a. i log
")92.
88.
97.
Tra 2
.
15
89.
CJ
93.
06
90.
94. f 7ra6. 95. 3 Tra 2
.
4n
99.
30
91.
96.
2.
i4
100.
^(7r2)a
101.
102. 4 A/2
a.
105. 106. Tra 3 tan
0.
103.
104. 4
Tr 2 a 3 .
107.
386
Page 140
108.
(Tra,
(Tra,
ANSWERS
*
a).
111
/
256a
256a \
315
TT/
109.
a).
\315Tr
112.
110.
2(47r3V3),
CHAPTER
Page 154
1.
VIII
3 x2
e x8
+
2 x3
c.
3
2
?y
y2
2 ?/
=
c.
2.
3.
siny
=
f
6.
x x2
Vl
tanx
Vi +
=
c.
7.
= ce*. = 2c?/ +
c2.
4.
=
c
cosy.
8.
21ogx + (sin 1 )
cosx sin y
=
c.
9.
=
c.
Page 155
10.
xsin
=
e
c.
25. y
=
c
cosx  2 cos 2 x.
x
x)
26.
11. x \ex
=
c.
12. (x 14.
15.
13. x 2
16.
17. 18.
+ y) 5 + 4x?/  2  Gx  2y = (y + 3)(x + y + 1) = c. (x + y)* + 2y = c. 2y = sinx cosx + ce~ x = (x + l)2(ex + c). y = ^(x + l) + c(a; + l).
?/
.
28. y 2
c.
=
(2 x
J
2
3) f
29.
?/
2
=
=
i /I
^21
2
30.
+
x
ex
1
gtan"
a;
7/
31.
32.
33.
19. x
=
sin^
ce
x
.
34.
35.
36. 21. logx?/ 22. x 2 c2
23. 24.
+ l(x 2 y2) = = + 2cy. 2 y = 1 + x + cVl + x = l + c xVl x
y=
 X
Cix
2 2
+c X 2 (l 2 = C(lfX 2 ). x 2 logx 2 + (x 3 + y 3 )5 = ex 2 y (ex + 1) = e x + x + c. 3 (secx + tanx)(l + 2/4 y = ce 2 *  (2 x 2 + 2 x + 1). y = (x+ c)eax
e tan
?/
)
.
)
.
37.
38.
39.
c
2
sin ax + + c2 y = y = y = J x 3 logx  ^ x 3 + cix + c 2
.
.
40.
41.
42.
Cl

x2
+
sin
1
+
c2
.
?/
=
+
c2
.
V
r=^W(X.
43. (y + I) 2 = Cl x + c 2 44. y + Cilog(y  ci) f x
47.
y
T/
=
x
=
X
+
c2
.
c2
.
48.
=
(1X2
+ Vc
51.
2
l)]+C 2
.
y
=
Cicosh[fc(x
+
c 2 )].
ANSWEKS
Page 156
52. y(x
53.
387
4)2
y
=
9.
.
56.
?/
=l+
= cx n = ax 2 + =
.
e*
tan
=
ex 1
ce
x
57. y
54
y tan?/3
"^2
r
C
58. y 59.
r"
c.
2~
=
55. 4
62.
64. 65.
Q(k x
+
2.
60
c)]. is
fc
x=
y
+
Ce
y
= = =
cosh [A: (x
a cosh
2
(& is
the constant ratio.)
y y
1
(a
the constant length.)
1
sino
8x
K*
2
Vk*x 
_
^
63. x 2
f
y2
=
ex.
4 x2
66. y
=
k cosh

Z
+
c.
(k
is
the constant ratio.)
(k is the constant ratio.)
*"=
70. (x

2
Ci)
+
(x
(y

c2 ) 2
=
1.
c2
.
(c is
the given constant.)
71. ciy 2

c
2
1
fc
+
c2 ) 2
=
(k is the constant ratio.)
Page 157
72.
Harmonic motion.
73.
=
2
and x
74.
=
c+*/^fVaxx
=
if
 sini
~
a
,
where
A:
is
j
the constant ratio,
a when v 0. Same velocity as
the
body
fell freely.
75.
About
7
miles per second.
CHAPTER X
Page 195
1.
(1,
2, 2).
2.
3.
6.
x2
+
y
2
+
z2

(
L9,
_
iy&gt;,
J
2x

4y

2z

49).
43
=
0.
11.
cos
e
13
2x37/ +
ay
6z21 = 0.
Page 196
14. x 17. TTX
+ y + z  6 = 0.  2y  2z + 2?r =
ef )
16.
0.
(
+
0.
kz
=
0.
18. e (x 19. x

e f (y
e0.
)
+ 2y +
9
3z
7
0 =
+ V2(z
v7 2) =
25. x
26. 27.

2 y
+
z
+
G
=
0.
20.
3
sini].*.
(i, 2, i), (i, 4, ^).
Viiio
28.
Viso Viso
29.
xz + 2 = 0.
llx+ 13y37 =
0,
3x+13z+
10
=
0.
388
Page 197
30. (2,
31.
ANSWERS
1, 2).
xl = yl = 2l
3
37. (1,
39.

1, 1).
4
32.
x1 ~2~
x
e*
2
40.
41.
~T~
y
e~
f
z
33.
34.
&
x
z
=
z
0,
ey =
0.
V2 V2
t
42.
0.
43. 44.
a;
35.
2y +
1
=
+ I6y + 7z + 8 = 0. 3x2?/z + 4 = 0. z + lly + 52 = 0. z6y + 22 = 0, 3z + y + 32l = 0. 2 = 0. 93 x  46 y + 13 z  179 = 0.
x
36. cos
2^,
VlS
,
2
46.
c
(1,
2,4),
(f,
y,y).
Vl3
47.
(0, 1, 2),
CHAPTER XI
Page 218
10.
AA =
12.
AF=
sin sin (a
.396 sq. in., 5.11 cu. ft.,
ft
dA =
dV=
ft
.398 sq. 5.09 cu.
in.
ft.
11.
AL =
.057 in.,
dL =
.057 in.
h sin
j8)
cos (a
+
/3)
/
+
sin 2 (a
(zi
18. xix
+
21
/_
/
yi?/ n, ad
a) (z
+ a) =
+
0.
~
a2
cd
\
+
62
c2
a2
+
52
+
C 2/ a).
22. (a, a, a), a, a, a, a), ( a), (a, 23. Point of intersection of the medians.
24.
(a,
a,
V=
8 abc
2
aK
a2
3V3
Page 219
32
a2
+
b2
+
c2
a2
+
_
6*
+
c2
_+
!l*L J Ii
1
62
c2
39. cos
1
52
2ac
49.
0.
ANSWERS
Page 221
\,
389
cx 2
1^ + 2^cycx
1
70. Trlog
+V1 a2
71. log(a
+
l).
CHAPTER
Page 234
XII
CHAPTER
Page 253
1.
XIII
2. 4.
fs Va6 (a +
3
2 6) (5 a
64a 4 Va/l
1056
\3
+
2 ab
8. 9.
+
5 6 2 ).
.
+
a
\
lUp)
if
a*
V3.
iTra 4
12.
5
iTra 4
.
6
10.
11.
f
Tra4
13
.
3465
4
.
63
T%7ra
14.
Page 254
15.
21.
24.
8 a2
.
16.
17.
(107T /)a 4
.
22.
25.
2a 2
.
26. 2
18.
(27T
2 ) a
.
27. 8 a 2 23.
19. OTT.
20. 29. x 2
a
28. 2
2
.
a2 esc 2
or.
+
(y
+
z tan a) 2
=
a2
;
2 a2 (ctn
a+ a
csc 2 a).
30.
31.
32.
37T).
Page 255
33.
10a 2 [7rV2log(l+V2)]. /32a (*,a).
35. /O
\
3a
^
57r
+ 8 )\
37
H5a\
39
357T/
,
/ 26Ga
256 a\
V 15 *
\3157r
315W
5(67r4)/
36.
Q0 /288a 50. \1757T
I
288 a\  I.
1757T/
40.
(
)u the axis of the sector,
4
(
(
;
a
0).
distant from the center of the circle
41.
128V2a
105
TT
42.
]a,
43.
] Tra, 0).
390
5a
iO,
48. 49.
ANSWERS
5a(k,+ko)\
*=:
)
46.
,
/8a
(
106 10c\ ,_,).
its
47.
,
/4a
26
.
8a
On
On
the axis of the cone,
~
distant from
vertex.
the axis of the cone,  of distance from vertex to base.
Page 256
50. (^ a, 51.
being in the
On
the origin, and the octant a), the center of the sphere being at octant bounded by the coordinate planes. the axis of the cone, midway between the vertex and the base.,

a,
first
74.
} 2
irpa*h.
75.
76.
^L
^
fcTrA6
,
tan*a.
(fc
fora6
is
the constant ratio.) (fc the coefficient of variation.)
is
77. firirprfrf). of variation.) 78.  ka & (k is the coefficient
.
QM(l /i
cos a)
8Q
TTpq (2 6

a)
2
tan 2 a
(JW
is
&
81.
82
3a i3^
2
the mass of the hemisphere.)
2
7
"
25 a 2
\
(3f
is
the mass of the hemisphere.)
83.
^
.
(M is
the mass of the ring.)
CHAPTER XIV
Page 276
3
1.
7ra 2
.
%
2.
^(36
.
7).
2 a).
Page 277
5.
8. 2. 6.
0.
7.
(k is the constant ratio, r is the distance.)
;
k logr
;
&lt;j
(fc

and
r as in
Ex.
14. x 2 15.
16.
i
n
i i I
;
+ ?/  xy + x + ^ 2 + xv  C x 2
y
=
c.
.
5. 3.
_L
(x+
llA;Ll 5
17. x 2 18. lo
=
ANSWERS
Page 278
19. X*
391
_
+
3 X7/2
=
xy
r7/2.
24
25.
,
20.
tanix
+
=
c.
x3y
+
x2y2
f
2/ )
=
+
1
c
c.
c
_x
21. e ;+log
2x + sm
2x
y
=
4y cosz =
26
.
22

10
^ + 3V
log C os2(z
+
C
27.
28.
log* l
^
y2
=c
.
23. xy
+ log^c. x
Page 302
14.
l&lt;x&lt;l.
CHAPTER XV
18. All values of x.
19.
21.
l&lt;x&lt;l.
15.
16.
l&lt;x&lt;l. l&lt;x&lt;l.
l&lt;x&lt;l.
All values of
l&lt;x&lt;l.
x.
20.
22
a
a
&lt;X&lt;
~b
b
17.
no
1
+
,
X2
5a;4
24.
l
+
x
+
x2
+
3
+
+
22
....
27.
1
+
2
x
2
+
+ 23 +
45
.
25.
x3
+
\
_
I2
...
28
cos 3 x
_^
_^_^_17x
12
8
2520
29
30
cosx 4i
cos 2 *
32
sinh air
_
+
2
a sinh
TT /
~^
2 sinh
TT
CTTT
\rM^oTT7 +
2
cos x
cos 2 x
\
"7
/
sinx
~ _
2 sin 2 x 22
f
\1 2
+
a2
a2
+
3 sin 3 x
32
\
+
a2
/
31
4 / ^if
\
+ !l^x
I2
sin5x
1
K~ +
cos^x
\
32 30
r
!
_2/cosx
TT
cos3x
4
3jr 4
\
~32~
cos3x
~3^~"
~52~
cos5x ~~52~
~
/sinx
(,"1
sin 2
a;
7
~2~
+ ~Q
sin3x
_
+
2 /cos x
^l
/
I2
3 sin x
_ sin 2 z
22
234
3 sin 3 x
cos 3 x
32
~
sin 4
x
I2
Page 303
40
38.],,,".
2.
2

43.5.
44. 0.
45. 0.
37
~
2
39
f

46.0. 47.0.
48
0.
42.
392
ANSWERS
CHAPTER XYII
Page 336
1.
y
=
a
+ a*x +
2.
3.
y
y
= =

a
a
+
x
+
2
3 a: 3 ao 3
111 ^
3
(a

l)x
2
+
/a 4
^jx
3
+
.
5
(a


4a
+ rrf^
5a
5
4. (y
5.
8.
9.
10. 11. 12.
+ a x + (a + )x* + +a x+a 6. (x 2 y  c) (y  ex) = 0. + c) (y + 2 x + c) = 0. 2  x 2  c 2 = 0. 7. (yc)(yce* )(3yx 3 c)=:0. ex) (y (y = 0. c (x + a) ec (x + a) e] [x a a [x 2 2 2 y sin x + 2 cy + c = 0. 2 13. c 2 (1 f y 2  2 c 3 xy + c 4 x2 = 1. 2 c 2 xy + c 3 x2 = 1. cy 2 14. x = logp + p + c, y = _p + c2 p2 y = 2cx 2 2 = 2 ex 15. y = c (c + c2 x) y
x2
3
.
4
5x
2
)
]
)
.
.
.
.
Page 337
16. x 2
y 18. y
19. x
17.
= sin 2 (y +_c). = p log Vcp, x = (l + = ex 2 + c 2 + y = cti\(cy).
.
log
Vcp) log vep.
4 (x 2 y + c) 2 20. 9 (x x Iogcx)(2x 2 y 2 21. (y
+
=

+ I) 3 x 4 c) = 0.
.
22
s
=
(P

2
I)
2
23. (l
Vy
+
l)
=
ce
_(P1) a2v^Ti.
c) (XT/
2
24. (cy
25. y 26. 27. 28.
30.
31.

e 2")(x 3

3?y
+
+
cy
+
1)
=
0.
= cpW,
2
x
=
(1
+
2 2 x]
.
2 (!4?y) + 4x 2c[(l4?/)cos2x + 2x sin
(Z/
+
c2
=
0.
32.
 ex) 2  c 2 + 1 = 0. e = cex + c^. Vl  c 2 2  x 2 = 1. y = ex  3 x 2 + c) (2 y  x 2 + c) = 0, (2 y x 2 = 2 cy + c 2 + a 2 x 2 + 2 = a 2
2
_
29
^=
ca a
+ l. c
,
?/
x
=
0.
,
2/
.
2 = 34. (y ex) 38. &2 X 2 + a 2^2
39.
41.
42.
 1 = 0. 37. 27 y  4 x 3 = 0. c, 4 xy _ a 25 2 _ 0, y = CX + Vb 2 + a C 2 = c. 2 2 40. (x + y) (y + z) (z + 1) z (x + 1) (y x 3 + y 2 + z 2 + logyz 2 = c. xy + yz + zx (a + b) z = k. (c + a) y (& + c) x
+
2
2
.
x)
=
c.
ANSWERS
43. 2/(z
393
x
+
V
z
z)
+
x
c(7/
+
z)
=
0.
^
47.
c.
44
X
+ V_ Z
lQnz
=
Cf
y
+
45. log(x
+
y)
+
x
+
y
+
z
=
loz =
48. x 2
49. 50. 52.
+ 2 = M2/ 2 + &2 ), tan1 5 ~ x 2  2/2 = ci, 7/2  z 2 = C2 2 x = = (x + ci) Vx 2 f 1, 2 = y
.
c
51
e
*+
e
s
=&lt;*
e
:
Page 338
53. x 2
+
z2
=
c
2
y
+
c2
=
2
log
54. x 2 55. x 56. x
57. x
 xy = ci, z + x 2  c 2 x + v=c *,2y = * +
l
~ Cl + Ci + ci = 0.
* z
&lt;*.
!!
58.
59.
60.
61. 62.
63.
74.
 c 2 ^. = cie 2 66. 27 p^2 = 4(x  2p) 3 = ciz, x 2  y a = c 2 y. 6T. 3 2  4y = 0. y 2 = 68. X + T/+ zCi, x 2 + 2/2 + z 2 = c. 4px + 4p 2 =c 69. x* + 2 z = ci, x y = c 2 e 22 x + + 2 2 70. x 3 + (x + 2p) y* = 0. + z 2 = c 2 z. y = ciz, x + 2  2 71. (x 10 y 3 + 27 x = 0. k(x + y) + 2 72. x 2  4 a (a  y) = 0. 4x2/ = c 73. x* + y = a x 2  4 2/2 = 0. 2 (x  c) = k log (A; V^  y) T Vfc2  4 y2
y
,
^
_!
2
65.
yl ? += 2xy
=
Cl
,
.

c2
z/
e"
.
?/
.
s
2/
.
2/
.
7/
4
.
_
i
2/)
A;
2
=
0.
8
.
.
Page 339
75. r
=
k.
77.
2x 2 +
?/
=
c2
.
To. 81.
^=^0
7*
78. x 2
CC
+
?/
_
79. y
2 a2 logx
+
c.
80. y 2
= ex 4 = 4 ax +
.
4 a2
.
A
A
82. x 2 84.
family of circles tangent to + y 2  ex + 1 = 0.
OX at 0.
=
f or
c
family of lemniscates having the line e
a
common
axis.
85. r
=
c(lcos0).
86. r
= e^
 o\
CHAPTER
Page 360
1.
XVIII
y
y y y y y y
=
= = = = = =
T*.
ce 3 * + J x 2 ce2x + e3x +
a;
2.
3.

 x
1
+
^e
(sin
(3 sin x 5r + x cosx).
j
.
2
^

cosx).
4.
5.
6.
7.
+ 3x)e +(2xl)e ce 4 x + 5  ^V (2 cos 2 x f sin 2 x). e [c + 2x Iog(e 2 r + 1)]. sin 5 x + 5 cos 5 x) + ce 2x (2 sin x + Jg (2
(c

^
cos x).
394
8.
ANSWERS
= Cl e 2 * + c 2 e4* 10. y = Cl + c 2 x) e** = cie5x +c 2 11. y = cisinSx + c 2 cos3x. y = e 3x (c!cos2x + C 2 sin2x). x ~ Wy = cie* + c 2 e 3 *  J x 4  f x 3  y z 2 x y = GI(? + c 2 e  2 + f cos2x. y = cie 2 * + c 2 e~ 5a; + ^j(9sin3x  19 cos 3 x). y= cie* + c 2 e 4 * f. y = ci + c 2 e 3 * + TL x4 + 1 x 3  x 2 + x. y = (ci + J x 3 x 2  f x)e 2 *+ c 2 e~ 8a:
y
7/ (
.
9.
12. 13. 14. 15. 16. 17.
W
18.
.
Page 361
19.
y
20.
y
21. y 22.
= = =
+ c 2 x + l x 2 ex + 4 e 2 *. 3  6x 2 + (ci + Caxje^ + x + c 2 x  log (x  3)] e 3 *. [d
(ci
)
18 x

25.
y=(ci +
c 2 x)e 2 *H
e 2x
125
(7 sin
3x

24 cos 3
x).
23. y 24. y
25.
=
GI sin ci sin
c\
2x
+
c2
cos 2 x H
g2x
(sin 2
x
2 cos 2 x).
y
26. y 27.
28.
= = =
y.=
c 2 cos x J cos 2 x + ^3 cos 8 x. + (c 2 + x) sin 2 x. 1 x + x cos 2 x GI cos x V3 + c 2 sin x V3 + ^ 1 x) cos3x]. e 2x [cisin3x + (c 2
x
+
cos 2 x
2 sin 2 x.
l
+
1
e,
+
e
2
29. y 30.
= =
e~ x (ci cos 2 x
e2
(
+
c 2 sin 2 x) f ( h c 2 sin
x2
3 x2
x ^) e~
+
+
]

g
y
ci
cos
2
C 2x c2
\
2
.
J
+
+
10 x
3.
/
31. y
32.
33.
?/
=d+ = GI +
.
+
C 3 e 3x
cos x
e
+
c 3 sin x.
y
?/
cie^
+
*/ 2
\
(c 2 cos
xV3
2
hc 3 sm
.
xV3
2
34. 35.
y
= =
(ci
x2
+ c 2 x) cos x + (c 3 + c 4 x) sin x. + C!X + c 2 + c 3 cosx + c 4 sin x.
y
_;
36. y
= (e^e^+e^/2 / c^os 4
V
+
c2
sin^
)
V2
V2/
c4
^ +
_
37.
V2
sin^Y V2/
y
=
Cie~
x
+
e2
X
f
c2
cos

(2 sin
x
cos x)
+
1 x y ^ e~
(2 sin
x
+
3 cos x).
ANSWERS
38. y
395
c4
39.
?/
40.
41. 42.
= &^*(ci cos2x 4 c 2 sin 2 x) + e~ * (c a cos 2 x 4 e 8:r + &x*?nT x 3  2 x2 7 x 4 C 3 e*. = Ci 4 c 2 x 4 (1 x 4  ^ x 2 sin 2 x. y = (ci + c 2 x) cos 2 x 4 (c 3 4 c x x 2 )e 2 *. y = ci 4 x 4 (c 2 4 c 3 x +  3 x 2 + x 3  } x4 + ^ x5 x y = cie~ + c 2 + c 3 x
^

sin2x)
^
4
)
)
.
*
43. x
=
ce 2 ,
6_
y
=
_
ce2 .
L&lt;
44. x
45.
46. 47. 48.
49.
= ce 2 4f e5&lt;2e 2 y = ce 2  fe^e2 + 1 e + ^V (5cos2 + sin2 ), x = c est + c 2 e=  3 cie  ^ c 2 e 2 + ^ e  ^V (17 cos 2 + 19 sin 2 y  2 c 2 e4 + ^. x = cie + cae4 + U, y = Cie 4 3 x = c^ 3 + cae4 + 3, y = 2 Cie + Cae + 2. 2 +^ 3 x = ci + c 2 e 2 + c 3 e~ + ^    ^2 + 3  2c 2 e2* + C8 e + f y = 2ci + I 2 + i 3 x = d + c 2 e 2 + c 3 e 4   J
.
,
2&lt;
i
).
&lt;
&lt;
&lt;
,
*
^

&lt;
.
50
3.
_
=
e
c
cos
^ +
V2

03 sin
^ V2/
]
4 e
W
V
\
c3
cos
^ + V2
V2
c 4 sin

V2/
c 4 cos
J
,
ec
CiX
2
cos
V2
 d sin ^^ +
V2/
e
^2 /c 3 sin ^ 
V2/
)
Page 362
51. 52.
?/
=
+
c 2 x2
+ ^x 3
4
2
.
?/
=X
5
[ci
cos (log x2 )
4 c 2 sin (log
x 2 )]
+
]
^
(5
logx
2
.
2).
53. y
54.
?/
=
=
ci
I
j
cos (2 logx)
C 2 sin (2
logx)
+
.
(logx)

x2
4 4
ti
(log x)
4
c2
log x
c 3 sin
4
c3
55. y
56.
=
CiX
ca
cos (logx)
4
(logx)

.
z/
= =
Cl
4c 2 x 2
2
.
4^^(logx)
23
22 3
4
2

3
4
5
(r
+
57.

2)
y
x3
S8.
j,
=
59. y
=
ri
(3542x +
21 x 2

4 x 3 ) 4
(3
 14x421x 2 ).
396
ANSWERS
4
5
6
3.8
+
.
.
,
/
[8.6.7.8*
61
.
^FT^T
ir
4.6.6...(r
y
=
Cl
l
 i (nz3) + L4
_
62.
= cjl +
L
"x.
+ ^LZlil^ + n(4)(,18)
4
_6
[2
...
+ + +
63. s
+ *(n4)(n16)...rn(2r2)1
[2r
"
C2
xFl +
.
^^ + (!L^H^i^)
[3
x
^4
,
(n

1) (n
_7
9)
(
n
n
25)
.
.
+
(
n
~
[6
)(
n

9 )(^

25)
.
.
[n

(2r
1)2]
2
=
Cl
cos
^+
Atf
c 2 sin
^+
+
_
2r
C os
+
W. ^
l
J
(^
is
the constant ratio.)
s
=
ci
cos
+
c 2 sin kt
sin A:, if
=
yfc.
+
a (A 2

jp) cos kt
+
did sin kt
a(h
2
a(h*k*)coskt
"
+
2 akh sin to
2
65. h
=
(A
k.
and
Z
are the constant ratios.)
Page 370
CHAPTER XIX
4

z
=
ANSWERS
Page 371
5.
6.
397
z
z
= e 2 2/[0i(x)cosy f 2 (x)sin?/]. = ^x 2 y 2 f 0i(y) + 02(z).
(x
2
I
3
&lt;t&gt;[e
8z
(2x
&lt;
8.
y,
?/
2)
2)
=
0.
0.
9.
(x
+
y
2
,
=

1^
0(^2/5
xyz\
= 0.
16.
\~
/
11.
0(x
+
2
y
?y
+
2
,
2,
x2
+
2
?/

z
2
)
=
0.
0.
17. 18.
2
(?/\
x
(x
/

2
?/
,
2
?/

z 2)
I
=
0.
+
z
+
taniJ
 r 3 sin 3
=
x
(x

z,
xz

?/ 2 \
0.
21.
u
=

Ir sin x
+
+
2&lt;
 r5 sin 5 x
+).
25
2
22.
u
=77
2
(e\
sinx
+
 e3
9
sin3x
+
 e5
sin5x
+
V
/
INDEX
(The
Roman numerals
refer to the volume, the Arabic to the page.)
Abscissa,
I,
36
Arc, differential in space rectangular
length defined,
II, 77
Absolute convergence, II, 283 Absolute value of complex number,
II,
coordinates, II, 180 I, 195
305
II,
length in plane polar coordinates,
Absolute value of real number, Acceleration, I, 202
Adiabatic
line, II,
283
207
length in plane rectangular coordi nates, II, 75 length in space rectangular coordi nates, II, 179
limit of ratio to chord,
I,
Algebraic functions, classification, 1, 43 graphs, I, 121
differentiation,
I,
178
195
equations
implicit,
(see
I,
Equations) 188
Archimedes, spiral of, I, 332 Area, center of gravity by single in
tegration, II, 94, 96
integration, II, 26, 113, 119 Amplitude of complex number, II, 305
Analytic function,
of ellipse,
II,
311
radii
center of gravity by double inte gration, II, 246
derivative in polar coordinates,
I,
Angle between normal and focal
I,
191
348
derivative in rectangular coordi nates, I, 204 II, 47
;
between between between between
57
plane curves, I, 211 planes, II, 186
space curves,
II,
184
I,
straight lines in plane,
determination by double integra tion, II, 239, 240
determination by line integral, 260
II,
between straight
183
lines in space, II,
between tangent and radius vector,
1,345
eccentric, of ellipse,
vectorial,
I, I,
determination in oblique coordi nates, II, 66
determination
304
nates, II, 67
in
polar
coordi
329
Arc, derivatives in plane polar coordi nates, I, 347
derivatives in plane coordinates, I, 196
nates, II, 79
differential in plane rectangular
determination in rectangular co ordinates, II, 64
moment
of
of
of inertia, II, 236
II,
rectangular
any surface,
ellipse,
I,
241
II,
304;
65,
66,
differential in plane polar coordi
261
of lemniscate, I, 348 of parabolic segment,
I,
216
;
II,
coordinates, II, 78
95
399
400
Area
of rose of three leaves, II, 69 of sphere, II, 242 of surface of revolution, II, 79
INDEX
Center of gravity, of parabolic seg ment, II, 95
of plane area
by
single integra
projection of, II, 177
tion
II,
II,
94
representing definite integral, 41
of plane area
tion, II,
by double integra
246
of complex number, II, 305 Asymptote, defined and illustrated, I,
Argument
128
of plane curve, II, 92 of quarter circumference, II, 93 of solid, II, 248 of solid or surface of revolution,
II,
of hyperbola,
I,
145
of hyperbolic spiral, I, 333
96
Attraction,
by multiple
integration, II,
252
of spherical segment, II, 97 Center of pressure, II, 98 of pressure of circle, II, 98
by
single integration, II, 86 of cylinder, II, 252 of wire, II, 87
Change of coordinates (see Coordinates)
Chord
of limits in definite integral, II, 49 of contact, I, 248
Auxiliary circle of ellipse, I, 304 Axes of coordinates (see Coordinates)
of ellipse, I, 141 of hyperbola, I, 145
supplemental,
I,
264
Circle, auxiliary of ellipse, I, 304 center of gravity of quarter cir
Axis of parabola,
I,
147
cumference,
definition
II,
93
I,
of revolution, II, 69 of symmetry, I, 121 radical,
I,
and general equation,
134
involute of, I, 311 of curvature, I, 354
of, II,
175
Beams, bending moment
149
Bernouilli s equation, II, 148 Bessel s equation, II, 359
parametric equations, polar equation, I, 342
I,
303
BessePs functions, Binomial theorem,
II,
II,
360
59
pressure on, II, 89, 98 special case of conic, I, 149
special case of ellipse, I, 142 tangent to known line, I, 136
Bisection of straight line, I, 39 BoyleMariotte s law, I, 43 Cardioid, I, 337 radius of curvature, Cassini, ovals of, I, 338
through known point, I, 136 through three known points, 1, 138 with center on known line, I, 137
Cissoid,
I,
I,
362
281
151
Catenary, equation and graph, equation derived, II, 143
property of, Center of circle,
of conic,
I,
I,
polar equation, I, 341 Clairaut s equation, II, 321
Coefficient, differential, II, 7
II,
I,
152
of element of a determinant,
I,
8
134
238
undetermined, in differential equa tions, II, 350
undetermined, in partial fractions,
II,
of curvature, I, 356 Center of gravity, defined in plane, II,
104
I,
90
defined in space, II, 245 of elliptic segment, II, 96, 247
Collinear points in a plane, in space, II, 188
38
Comparison test for convergence, II, 280
INDEX
Complex numbers, argument,
conjugate,
I,
401
I,
II,
305
Coordinates, Cartesian,
224
32
;
denned,
I,
31
II,
304
functions of,
II,
307
change from plane rectangular to oblique, I, 224 change from plane rectangular to
polar,
I,
304 graphical representation, II,
341
to
modulus
of, II,
305
;
change from space rectangular
II,
operations with, I, 32 Complementary function,
305
cylindrical, II, 231
II,
349
change from space rectangular to
polar, II, 231
Components
of force,
I,
I,
34
112
of velocity,
200
I,
change in multiple integrals,
234
II,
Concavity of plane curve, Conchoid, I, 334
Cone,
II,
165
I,
change in partial derivatives, 208
221
II,
Conies, classification, definition, I, 148
237
change of direction of plane axes,
I,
diameter,
I,
252
I,
change of direction of space axes,
229
II,
general equation,
limiting cases,
polar,
I,
193
I,
in oblique coordinates, I, 244
I,
234
change of plane origin, change of space origin,
cylindrical, II, 231 oblique in plane, I, 223
217
193
II,
247
I,
polar equation,
tangent,
I,
343
246
through five points, I, 241 treatment of numerical equations, I, 240
See also Ellipse, Hyperbola, rabola
oblique in space, II, 159 polar in plane, I, 329 polar in space, II, 231
rectangular in plane,
I,
35
Pa
rectangular in space, II, 158 Curvature, center of I, 356
,
Conjugate axis of hyperbola, I, 145 complex numbers, I, 32 diameters of conic, I, 258
functions, II, 313
circle of,
I,
354
I,
definition,
353
radius
of, in
parametric form,
polar coordinates,
I,
360
radius
of, in
I,
Conoid,
hyperbolas, II, 173
I,
262
40
361
I,
I,
Constant, defined,
of integration,
radius of, in rectangular coordi
;
206
II,
12
Contact, chord
of, I,
248
point of, I, 105 Continuity of function of one variable, I, 101
of function of
nates, I, 354 Curve, Cartesian equation of plane curve, I, 44
center of gravity,
166"
II,
92
two
variables, II,
degree, I, direction of space curve, II, 182
200
Convergence, absolute,
II,
equations of space curve,
II,
170
283
comparison
test, II,
280
parametric equations curve, I, 302
of
plane
definition, II, 279
polar equation of plane curve,
I,
ratio test, II, 281
330
seconddegree curves,
I,
region of, II, 285
229
402
Curve, slope of plane curve,
I,
I,
INDEX
99 104
189
tangent to plane curve, tangent to space curve, with given slope, I, 207
Curves, family of,
Derivative of rectangular coordinates with respect to plane arc, I, 196
of rectangular
II,
coordinates with
See also Arc, Area, Curvature II, 316
I,
respect to space arc, II, 182 partial, II, 198
second,
I,
110
intersection of,
161
with
common
points of intersec
sign of first, I, 106 sign of second, I, 1 1 1
tion, I, 171
theorems on,
I,
I,
179
132
Cycloid, I, 305 radius of curvature,
See also Differentiation
361
Descartes folium,
rule of signs,
I,
tangent
Cylinders,
to, I,
315
I,
87
I,
II,
169 173
Cylindroid,
Definite
II,
Determinants, defined, elements, I, 4
expansion,
I,
4
8
integrals,
applications
to
minors,
I,
4
geometry,
II, 64,
236
applications to mechanics, II, 86,
properties, I, 6 solution of equations,
I, 1,
1223
236
change of coordinates, II, 234 change of limits, II, 49 defined, II, 39
differentiation, II, 216
Diameters, conjugate, of conic, 1, 252
of ellipse,
I,
I,
258262
256
of hyperbola, I, 257 of parabola, I, 254
Differential, defined for one independ ent variable, II, 6
double,
II,
222
64
41
element,
II,
evaluation, II, 47
graphical representation,
limits of, II, 40
II,
defined for several independent variables, II, 201
exact, II, 269
properties, II, 45
triple, II,
230
formulas, II, 10 graphical representation for one
with with
infinite integrand, II, 53 infinite limits, II, 52
I,
independent variable,
II, 8
Degree of plane curve,
166
I,
graphical representation for two independent variables, II, 206
Derivative, applications,
202, 203
I,
application to velocity, defined, I, 102
higher,
I,
198
higher order, II, 10 of plane arc, II, 78 of space arc, II, 180
partial, II,
111, 187
II, 212,
I,
202
higher partial,
illustrations of,
214
total, II,
200
203
illustration of partial, II, 199 of f(x, y) when x and y are func
when equal to zero, II, 209 Differential equations, (aix + biy
dx
+
(a%x
+
b2 y
+
c2 )
dy
+ Ci) = 0, II,
tions of s
and
t,
II, 206,
214
146
Bernoulli
Bessel
s, II,
of f(x, y) tions of
y are func t, II, 202, 214 of polar coordinates with respect to plane arc, I, 347
when x and
148
s, II,
359
321
Clairaut
s, II,
defined, II, 141
INDEX
Differential equations, first order, ex istence of solution, II, 316
first
403
10
defined,
I,
Differentiation, collected formulas, II,
order not of
first
degree, II,
102
I,
318
general formulas,
s,
s,
184
Laplace Laplace 368
in plane, II, 366 in three dimensions,
II,
of algebraic functions, I, 178 of definite integrals, II, 216 of exponential functions, I, 284 of hyperbolic functions, I, 290 of implicit functions, I, 188
;
Legendre
linear, II, linear,
s, II,
357
340
II,
constant coefficients, any order, II, 348
210
of inverse hyperbolic functions,
I,
linear, constant coefficients, first
292
of
order, II, 341
linear, constant coefficients,
inverse
tions, I,
second
trigonometric 276
func
order, II, 343
constant coefficients, solu tion by partial fractions, II, 347 linear, constant coefficients, solu
linear,
of logarithmic functions, I, 284 of polynomials, I, 103 of trigonometric functions, I, 272 of w, I, 185
partial, II, 198
tion
by undetermined
350
coeffi
cients, II,
partial,
independent of order,
II,
linear, variable coefficients, II,
354
213
successive, I, 187 use of logarithm, I, 288 See also Derivative, Differential
Mdx + Ndy = 0, exact, II, 270 Mdx+ Ndy = 0, homogeneous, II,
145
Mdx + Ndy  0, Mdx + Ndy = 0, Mdx + Ndy = 0,
rable, II, 144
linear, II, 146
re sume
,
II,
317
Direction, cosines defined, II, 180 cosines found from two equations
of line, II, 186 normal to plane, II, 184 of plane curve in polar coordi
variables
sepa
of
order of,
II,
143
partial, II.
363
partial, linear of first order, II, 364
Pdx + Qdy + Edz =
case, II,
nates, I, 345 of plane curve in rectangular co
0,
integrable
328
0,
Pdx + Qdy + Edz =
grable case, II, 335
noninte
ordinates, I, 197 of space curve, II, 182 of straight line in space, II, 180 Directrix of circle, I, 149
of conic, I, 148 of ellipse, I, 149
problems in geometry,
151, 327
II,
141,
problems
in mechanics, II,
142,
of hyperbola,
of parabola,
I,
149
101
149, 153, 346, 367, 370
I,
146
I,
second order, special cases, II, 149 singular solutions, II, 325 simultaneous of first order, II,
332
simultaneous,
solution
linear,
Discontinuity defined,
examples
283
;
of, I, 41, 128, 268, 282,
II,
finite defined, II,
292, 294 291
I,
constant
Discriminant defined,
117
i
coefficients, II,
353
:;:r,
by
series, II, 318,
of cubic equation, I, 114, 117 of quadratic equation, I, 73, 1 17
404
INDEX
Equations, fractional roots, I, 90 first degree in two variables, I, 52
first
Discriminant of quadratic equation in two variables, I, 236
Distance between two points in plane, I, 36
degree in three variables, II, 161
I,
between two points in space, II, 178 of point from plane, II, 191 of point from straight line, I, 63
Divergence,
II,
irrational roots,
92
multiple roots,
I,
116
279
Newton s method of solution, 1, 1 14 number of roots, I, 80
rational roots, I, 89 second degree in two variables, 229
I,
e,
the number,
I,
280
Eccentric angle of ellipse, I, 304 Eccentricity of conic, I, 148
Elasticity, I, 204 Element of definite integral, II, 64 of determinant, I, 4
simultaneous, I, 161 simultaneous linear,
I,
12
simultaneous linear homogeneous, 1,21
solution
Eliminant,
Ellipse,
I,
23
I, 1
by
Elimination,
sum and product
and
transcendental,
factoring, I, 77 of roots, I, 82
I,
focal radii,
area,
I,
angle between normal I, 191
293
I,
304
;
II, 65, 66,
261
with given roots, Evolute, I, 357
II,
79
center of curvature, I, 357 center of gravity of segment, 96, 247
definition
I,
Exact
of ellipse, I, 358 differential in three variables,
II,
276
II,
and simplest equation,
358
in
two variables,
269
139
I,
evolute,
Expansion, coefficient, I, 204 Exponential equations, I, 295
expressed by general equation of second degree, I, 231, 235
length, II, 77
Exponential function, differentiation, I, 284
expansion,
II, 58,
308
parametric equations,
I,
303
special case of conic, I, 148 radius of curvature, I, 355
integrals, II, 25 of complex number, II, 307 of real number, I, 279
referred to conjugate diameters,
I,
259
Factoring of quadratic polynomial ,1,79
solution of equations by,
I,
Ellipsoid, equation, II, 165
77
volume,
II, 73,
249
Factors, integrating,
II, 271,
328
Energy, kinetic, I, 203 Entropy, II, 272 Envelopes, II, 322
Epicycloid,
I,
list of integrating, II, 273 of polynomial, I, 81, 83 Families of curves, II, 316
307
Epitrochoid,
I,
309
I,
Equations, complex roots, depression of, I, 79
tions)
82
II, 314, 327 of liquid, II, 259, 266 Focus of conic, I, 148
orthogonal,
Flow
differential (see Differential
equa
of ellipse, I, 139 of hyperbola, I, 142
discriminant,
I,
117
Folium
of parabola, I, 146 of Descartes, I, 132
INDEX
Force,
I,
405
linear
differential
202 266
General solution,
function, II,
equation, II, 350
Forms, indeterminate, II, 295 Fourier s series, II, 290
Fractions, differentiation,
integration, II,
partial, II, 104
I,
Generators, rectilinear, II, 172 Graph, I, 40
182
Gravity
(see
Center of gravity)
I,
113
.
Harmonic
II,
division of line,
I,
250
Functions, analytic, Bessel s, II, 360
classes, I,
311
motion,
275
I,
property of polars,
249
267
43
Heat, as line integral, II, 260
complementary, II, 349 conjugate, II, 313 continuous, one variable,
200
decreasing,
I,
dependent on path, derivatives, II, 269
I,
II,
101
II,
entropy,
II,
272
continuous, several variables,
Helix, direction, II, 182 equation, II, 172
106
defined, one variable, I, 40 defined, several variables, II, 158
Homogeneous
145
length, II, 180 differential equation, II,
defined by equation of second de gree, I, 127
graphical representation, one vari
able, I, 44
Homogeneous equations, I, 21 Homogeneous functions, II, 145 Homer s method referred to, I, 92
Hyperbola, conjugate,
definition
I, I,
262
graphical representation, two vari
ables, II, 159 implicit, I, 188; II, 210
and simplest equation,
I,
142
equilateral,
146
increasing,
I,
106
I,
increment, one variable,
100
increment, several variables, II, 200 involving fractions, I, 128
expressed by general equation of second degree, I, 231, 235 referred to asymptotes, I, 224
referred to conjugate diameters,
I,
mean
value, II, 54
I,
259
I,
notation,
44
200
special case of conic,
149
I,
periodic, II, 290
total differential, II,
Hyperbolic functions, defined, differentiation, I, 290
expansion,
inverse,
II,
288
See also Algebraic, Exponential,
differentiation of inverse, I, 292
Hyperbolic, Logarithmic, Trans cendental, Trigonometric func
tions
63
integrals leading to inverse, II, 24
I,
291
163
Gas, graph of BoyleMariotte s law,
Hyperbolic spiral, I, 332 Hyperboloid of one sheet,
of
II,
1,43
illustrating differential, II, 8
illustrating function, II, 168
two
sheets, II, 165
I,
Hypocycloid,
309
I,
fourcusped,
132 31
illustrating partial derivative, II,
199 temperaturepressurevolume sur face, II, 168
Imaginary unit,
I,
numbers
(see
Complex numbers)
surfaces, II, 160
406
210
INDEX
See also Definite, Indefinite, and
II,
Implicit functions, general discussion,
special case, I, 188 Increment of function of one variable,
I,
Line integrals
Integrand, defined, infinite, II, 53
Integrating factors,
II,
12
100
II,
273
of functions of several variables,
II,
200
Integration, by parts, II, 30, 51 by substitution (see Substitution)
Indefinite integral, defined, II, 41
constant
of, II, 12
Vq + 6x, II, 29, 119 2 containing Va + bx + x II, 121 x 2 II, 122 containing Va + bx 2 x 2 II, 29 containing Va 2 2 containing Vx + a II, 29 Vx 2 a2 II, 29 containing
containing
, , ,
,
definition, II, 12
elementary formulas, II, 26 fundamental formulas, II, 13
of simple
II,
of rational fractions, II, 113 differential equations,
141
,
fundamental formulas, dx
of
4 J3)
II,
26
,
possibility of, II, 32
_
+
c
II,
29
preliminary discussion,
special
I,
205
Vax 2 +
x
bx
methods
of, II,
119
dx
of
dx
See also Definite and Indefinite
a
+
6 cos
a
+
b sin
II,
x
integrals Intercepts,
I,
53
129
6 sin x of e ax sin 6x dx, e ax cos 6x dx, II,
a cos x
+
Intersection,
curves
with
169
common
31
of rational fractions, II, 113 of sec M xdx, cscxdx, II, 127 of
points, I, 171 number of points,
I,
of plane curves,
I,
161
sinxdx,
cos w xdx, II, 123
Involute, I, 357 of circle, I, 311
Isolated points, examples,
1,
of sin m xcosw xdx, II, 124, 135 of tanxdx, ctn w xdx, II, 126
of tan
126 11,292
;
xsecxdx,
II, II,
ctn m x
Kinetic energy,
I,
203
csc"xdx,
II, 128
Laplace
28
of
of of
Va2 x2 dx, Va2 + x2 dx,
M",
equation, in plane, II, 366 in three dimensions, II, 3(58
s
32
130
II,
14
II, 120, II, 130,
of x
(a
+ 6x)^dx,
reduction formulas, See also Integration
135
Latus rectum, I, 211 Legendre s coefficients, II, 359 Legendre s equation, II, 357 Lemniscate, I, 340
area,
I,
348
I,
Indeterminate forms,
Indicator diagram,
Inertia (see
II,
295301
54
Cartesian equation,
341
II, 43,
Moment of inertia) Infinitesimal, defined, II, 1
fundamental theorems
on, II, 4
Length (see Arc and Distance) Lima9on, I, 336 Limit, approached by variable, defined,
I,
Infinitesimals,
97
I,
order
Infinity,
I,
of, II, 1
of ratio of arc to chord, 1 cos h sin h , and Limits of
195
I.
270
29
of (1
h
Inflection, point of, II, 112, 194

1
I,
Integral, particular, II, 349
+
h}*
and
283
INDEX
Limits, theorems relating to, I, 178 of definite integral, change of,
II,
407
Logarithmic function, differentiation of, I, 284
expansion,
of
II,
49
52
59
defined, II, 40
infinite, II,
II, 267 Line integral, defined, II, 258 dependent on path, II, independent of path, II, 203
complex number, II, 310 of real number, I, 279
Maclaurin
s series, II, 55,
Line, adiabatic,
287
I,
2(&gt;7
Maxima and minima
defined,
s
108
II,
related to double integral, II, 2G1
discussed by Taylor 60
for functions of
theorem,
Line,
straight,
points in plane, 187
determined by I, 60
two
two
variables, II,
205
determined by two points in space,
II,
problems,
test
test
I,
192, 275
determined by point and direction
in plane,
I,
by first derivative, I, 108 by second derivative, I, 112
II,
58
Mean
value of function,
54
determined by point and direction
in space, II, 190 direction cosines, II, 186 direction in plane, I, 55
Mechanics and physics problems, ac
celeration,
I,
202
adiabatic lines, II, 267 attraction of wire, II, 87
attraction of cylinder, II, 252 beams, bending moment, II, 149 BoyleMariotte s law, I, 43
direction in space, II, 180 equation in plane, I, 50
equations in space,
II,
170
61
I,
intersection in plane,
I,
normal equation
in plane,
64
parallel lines in plane, I r 56 parallel lines in space, II, 183
catenary, II, 142 center of gravity, II, 90, 245 center of pressure, II, 98
coefficient of expansion,
I,
204 200
parametric equations
in plane, I,
I,
302
perpendicular lines in plane,
57
II,
components of components of
force,
I,
34, 35
I,
velocity,
perpendicular lines in space, 183
damped
vibrations, II, 346 deflection of girder, I, 110
I,
polar equation in plane, I, 342 satisfying two conditions in plane,
elasticity,
204
II, 259,
flow of liquid,
force,
I,
266
1,58
tangent to plane curve,
I,
202
Gas)
I,
104
gas
(see
tangent to space curve, II, 189 Linear algebraic equations, I, 1 Linear differential equations, II, 146, 340
harmonic motion,
heat
(see
275
Heat)
Linear partial differential equations,
II,
:
&gt;&lt;it
height of atmosphere, I, 287 indicator diagram, II, 43 kinetic energy, I, 203
lever,
I,
See also Equations and Differential equations
Liquid, flow of, Locus, I, 45
II, 259,
192
266
mean velocity, II, 54 moment of inertia, II, momentum, I, 203
motion of particle
(see
236, 251
problems,
I,
316
I,
Motion)
Logarithm, Naperian,
280
pendulum,
II,
153
408
203
INDEX
Numbers,
rational, I, 28
tential, II,
real, I,
Mechanics and physics problems, po
potential due to ring, II, 370
pressure, II, 88
29
See also Complex numbers
Operator, differential,
pressuretemperaturevolume sur
face, II, 168
projectile,
I,
II,
,
340
II,
314
I,
formulas for
192
D a
342
refraction of light,
Order of
II,
differential equation, II, 143
1
saturated steam,
I,
42
of infinitesimal, II,
temperature in long plate, uniform motion, I, 199
velocity,
I,
367
Ordinarte, I, 36
198
Orthogonal trajectories, Ovals of Cassini, I, 338
II,
327
work (see Work) Minima (see Maxima)
Minors of determinant, I, 4 Modulus of complex number,
II,
Parabola, area of segment, 1, 216; center of gravity of segment,
II,
II,
95
95
305
Moment, bending,
II,
149
Moment
cubical, I, 74 definition and simplest equation,
I,
of inertia, defined, II, 236
146
area, II, 236 rectangle, II, 224
any plane
solid, II,
expressed by general equation of second degree, I, 231, 235
length, II, 77
251
sphere, II, 251
Momentum,
Motion of
I,
203
292
particle, falling in resisting
I, I,
path of projectile, I, 314 referred to diameter and tangent, I, 255
referred
to
medium,
harmonic,
in circle,
I,
I,
275 314
203
latus rectum,
tangents at end of I, 132
148
314
in ellipse,
semicubical, I, 131 special case of conic,
I,
path
of, I,
313
I,
Paraboloid,
elliptic, II,
II,
162
uniform,
346
projected upwards, I, 199
hyperbolic,
II,
166
vibrating in resisting medium,
tangent plane, II, 205 Parametric equations, I, 302
differentiation,
I,
315
with given velocity and accelera
tion, I, 207
Partial fractions, II, 104 derivative, II, 198
differential, II,
202 349
Newton s solution of equations, I, 114 Normal equation of plane, II, 185
equation of straight to plane, II, 184
line, I,
differential equations, II, 363
Particular integral,
II,
64
Parts, integration by, II, 30, 51
Pedal curves,
I,
319 290
to plane curve, I, 191 to straight line, I, 59 to surface, II, 204, 211
classification, I, 28
I,
Pendulum,
II,
153
II,
Periodic function,
Numbers,
complex,
31
;
II,
304
Physics (see Mechanics) Plane determined by normal and point, II, 184
irrational,
I,
28
determined by three points,
II,
190
INDEX
Plane, direction of normal, II, 184 distance of point from, II, 191 equation, II, 161
409
Revolution, center of gravity of sur face or solid, II, 96
equation of surface,
II,
168
normal equation,
II,
185 183
192
volume of
solid, II,
II,
69
parallel planes, II, 183
perpendicular planes, tangent, II, 204, 211
II,
Ringsurface, area, volume, II, 71
80
Roots, complex,
line, II,
I,
82
through given straight Polars, I, 247
reciprocal, I, 251 Pole of straight line, I, 247
fractional,
I,
90
irrational, I, 92, 114
location,
I, I,
86
116
multiple,
I,
of system of coordinates,
329
number
of, I,
80
Polynomial, defined,
derivative of,
factors, I, 81,
first
I,
I,
43
of unity, II, 307
rational, I, 89 relation to factors,
97
83
50
I,
78
degree,
I,
sum and
Rose of three
product,
I,
82
nth degree, I, 74 second degree, I, 70
square root
of, I,
leaves, I, 331 area, II, 69
II,
121
Ruled surfaces,
Segments of
172
Potential due to ring, II, 370 rate of change, II, 203
Pressure, II, 88 center of, II, 98
line,
addition of,
I,
32
Series, convergent, II,
279
divergent, II, 279
on
circle, II, 89,
98
173
Fourier
s, II,
290 279 280
55, 287
II, 59,
Prismoid, II, 75 Prismoidal formula,
Products, graphs of,
Projectile, path of,
geometric,
II, 74,
I,
II, II,
83
harmonic, Maclaurin
s, II,
I,
314
Projection in plane, I, 34 in space, II, 176
operations with, power, II, 284
by,
286
solution of differential equations
II,
318, 356
65, 287
Radius of
curvature
in
parametric
Taylor
s, II,
form, I, 360 of curvature in polar coordinates,
Sine, graphical construction, I, 267 Singular points of curve, II, 324
1,361
of
solution of differential equation,
II,
curvature in rectangular co ordinates, I, 354
325
I,
Slope of plane curve,
99
vector, I, 329 Rate of change, I, 203
Ratio test for convergence, II, 281 Rationalization of integrals, II, 119
of straight line, I, 64 Solid (see Center of gravity, Moment of inertia, and Volume)
Reduction formulas, algebraic,
trigonometric,
II,
II,
130
Space coordinates (see Coordinates) Sphere, area, II, 242
equation,
center of gravity of segment, II, 178
of inertia, II, 251
I,
134
II,
II,
97
Region of convergence,
285
79
Resultant, I, 23 Revolution, area of surface,
moment
II,
Spiral of Archimedes,
332
410
Spiral, hyperbolic,
I,
INDEX
332
Transformation of coordinates
ordinates)
(see
Co
333 Statistics, graphs, I, 41 Stokes s theorem, II, 274
logarithmic,
I,
Trigonometric equations,
I,
293
Straight line (see Line) Strophoid, I, 152
Trigonometric functions, differentia tion, I, 272 differentiation of inverse, I, 276
expansion,
II, 58, 62, 308 integrals leading to inverse, II, 19
Subnormal,
polar,
I,
I,
210
351
119
Substitution, integration by, II, 27
integration, II, 123
inverse,
I,
general discussion,
II,
269
in definite integrals, II, 49
suggestions for, II, 29
inverse, expansion, II, 59, 62 of complex number, II, 307
Sub tangent,
polar,
I,
I,
210
of real
number,
306
I,
266
351
Trochoid,
I,
Successive differentiation, I, 187 Supplemental chords, I, 264
Surface, normal line, II, 204, 211
of second order, II, 162168 of revolution, II, 168
Undetermined
efficients)
coefficients
(see
Co
Value
(see
Variable, defined,
Absolute and I, 40
Mean
value)
ruled, II, 172
tangent plane to, II, 204, 211 See also Area, Center of gravity, Moment of inertia, Volume
Sylvester s
Variation of sign in polynomial, Vector, II, 304
radius,
I,
I,
87
329
I,
method
axis of,
of elimination,
I,
1,
24
Vectorial angle, Velocity, I, 198
329
Symmetry,
Tangent
121
components, I, 200 Vertex of ellipse, I, 141
of hyperbola, of parabola,
I,
I,
to circle,
I,
190
144
to conies, simplest equations, 1, 190 to conies, general equation, to plane curve,
I,
I,
147
246
Volume, any
solid, II,
249
104;
II,
210
common
to
two
cylinders, II, 74
to space curves, II, 189, 212
to surface, II, 204, 211
cylindrical coordinates, II, 232
ellipsoid, II, 73,
249
II,
with given slope, I, 163 Taylor s series for one variable,
55, 287
polar coordinates,
II,
233
74
prismoidal formula,
solid
II,
solid of revolution, II,
69
II,
for several variables, II, 288 Temperature of long plate, II, 367
with parallel bases,
149
I,
72
Time
as parameter,
I,
313
Witch,
I,
Total differential, Tractrix, II, 142
II,
200
327
points of inflection,
194
Work,
II,
defined, II,
40
II,
dependent on path,
267
Trajectories, orthogonal,
Transcendental equations, I, 293 Transcendental functions of complex number, II, 307311
Transcendental functions of real ber, I, 266
expressed by indicator diagram,
II, 43 expressed as line integral, II, 259 independent of path, II, 266
num
Zero,
I,
29
14 DAY USE TO DESK FROM WHICH BORROWED OMY, MATHEMATICSThis book
is due on the last date*s1tamped below, or on the date to which renewed. Renewed books are subject to immediate recall.
,HfTUNED TO
?m
BOOK tJtKAKIMhNT
LD
2150m6 60
)
(B1321slO)476
General Library University of California Berkeley
MS298790
This action might not be possible to undo. Are you sure you want to continue?