Thomas Calculus

I.A.
Maron
Problems
in
Calculus
of
One Variabl
H. A. MAPOH
HHOOEPEHUHAJlbHOE
W MHTErPAJlbHOE
HCMHCJ1EHME
B nPMMEPAX H 3AiXAHAX
<DyHKUHH OJ^HOfl nEPEMEHHOM
H3/lATEJlbCTBO «H AY K A»
MOCKBA
I. A. MA RON
Problems
in Calculus
of One Variable
(With elements of theory)
TRANSLATED
FROM THE RUSSIAN

by
LEONID LEVANT
MIR PUBLISHERS
MOSCOW
UDC 517=20
First published 1973
Revised from the 1970 Russian edition
Ha aneAUucKOM R3biKe
0223-200
041(01)-73
Contents
From the Author 9
Chapter I. Introduction to Mathematical Analysis 11
§ 1.1. Real Numbers. The Absolute Value of a Real Number 11
§ 1.2. Function. Domain of Definition 15

§ 1.3. Investigation of Functions 22
§ 1.4. Inverse Functions 28
§ 1.5. Graphical Representation of Functions 30
§ 1.6. Number Sequences. Limit of a Sequence 41
§ 1.7. Evaluation of Limits of Sequences 48
§ 1.8. Testing Sequences for Convergence 50
§ 1.9. The Limit of a Function 55
§ 1.10. Calculation of Limits of Functions 60
§ 1.11. Infinitesimal and Infinite Functions. Their Defini-
tion and Comparison 68
§ 1.12. Equivalent Infinitesimals. Application to Finding
Limits 71
§ 1.13. One-Sided Limits 75
§ 1.14. Continuity of a Function. Points of Discontinuity
and Their Classification 77
§ 1.15. Arithmetical Operations on Continuous Functions.
Continuity of a Composite Function 84
§ 1.16. The Properties of a Function Continuous on a Clo-
sed Interval. Continuity of an Inverse Function 87
§ 1.17. Additional Problems 91
Chapter II. Differentiation of Functions 93
§ 2.1. Definition of the Derivative 98

§ 2.2. Differentiation of Explicit Functions 100
§ 2.3. Successive Differentiation of Explicit Functions.
Leibniz Formula 107
6 Contents
§ 2.4. Differentiation of Inverse, Implicit and Parametri-

cally Represented Functions Ill
§ 2.5. Applications of the Derivative 115

§ 2.6. The Differential of a Function. Application to App-
roximate Computations 122
Chapter III. Application of Differential Calculus to Investigation of
Functions 131
§ 3.1. Basic Theorems on Differentiable Functions .... 131
§ 3.2. Evaluation of Indeterminate Forms. L'Hospital's Rule 138

§ 3.3. Taylor's Formula. Application to Approximate Cal-
culations 143
§ 3.4. Application of Taylor's Formula to Evaluation of
Limits 147
§ 3.5. Testing a Function for Monotonicity 148
§ 3.6. Maxima and Minima of a Function 152

§ 3.7.Finding the Greatest and the Least Values of a
Function • • 159
§ 3.8. Solving Problems in Geometry and Physics .... 162
§ 3.9. Convexity and Concavity of a Curve. Points of
Inflection 166
§ 3.10. Asymptotes 170
§ 3.11. General Plan for Investigating Functions and Sket-
ching Graphs 174
§ 3.12. Approximate Solution of Algebraic and Transcen-
dental Equations 183
Chapter IV. Indefinite Integrals. Basic Methods of Integration .... 190
§ 4.1. Direct Integration and the Method of Expansion 195

§ 4.2. Integration by Substitution 199
* 4.3. Integration by Parts 202
§ 4.4. Reduction Formulas 211
Chapter V. Basic Classes of Integrable Functions 214
§ 5.1. Integration of Rational Functions 214

§ 5.2. Integration of Certain Irrational Expressions . . . 219
§ 5.3. Euler's Substitutions 222
§ 5.4. Other Methods of Integrating Irrational Expressions 224
§ 5.5. Integration of a Binomial Differential 228
§5.6. Integration Trigonometric and Hyperbolic Functions
of 230
§ 5.7. Integration of Certain Irrational Functions with the
Aid of Trigonometric or Hyperbolic Substitutions .... 237
Contents 7
§ 5.8. Integration of Other Transcendental Functions . . . 240

§ 5.9. Methods of Integration (List of Basic Forms of In-
tegrals) 242
Chapter VI. The Definite Integral 247
§ 6.1. Statement of the Problem. The Lower and Upper

Integral Sums 247
§ 6.2. Evaluating Definite Integrals by the Newton-Leib-
niz Formula 256
§ 6.3.Estimating an Integral. The Definite Integral as a
Function of Its Limits 262
§ 6.4. Changing the Variable in a Definite Integral . . . 275
§ 6.5. Simplification of Integrals Based on the Properties
of Symmetry of Integrands 288
§ 6.6. Integration by Parts. Reduction Formulas .... 294
§ 6.7. Approximating Definite Integrals 301
Chapter VII. Applications of the Definite Integral 310

§ 7.1. Computing the Limits of Sums with the Aid of De-
finite Integrals 310
§ 7.2. Finding Average Values of a Function 312
§ 7.3. Computing Areas in Rectangular Coordinates ... 317
§ 7.4. Computing Areas with Parametrically Represented
Boundaries 327
§ 7.5. The Area of a Curvilinear Sector in Polar Coordinates 331
§ 7.6. Computing the Volume of a Solid 336
§ 7.7. The Arc Length of a Plane Curve in Rectangular
Coordinates 345
§ 7.8. The Arc Length of a Curve Represented Parametri-
cally 348
§ 7.9. The Arc Length of a Curve in Polar Coordinates 351
§ 7.10. Area of Surface of Revolution 354
§7.11. Geometrical Applications cf the Definite Integral 360
§ 7.12. Computing Pressure, Work and Other Physical
Quantities by the Definite Integrals 367
§ 7.13. Computing Static Moments and Moments of Inertia.
Determining Coordinates of the Centre of Gravity .... 372
Chapter VIII. Improper Integrals 387

§ 8.1. Improper Integrals with Infinite Limits 387
§ 8.2. Improper Integrals of Unbounded Functions .... 397
§ 8.3. Geometric and Physical Applications of Improper
Integrals 409
Answers and Hints 418
In fond memory
of my parents
From the Author
This textbook on mathematical analysis is based on many years'

experience of lecturing at a higher technical college. Its aim is to
train the students in active approach to mathematical exercises,
as is done at a seminar.
Much attention is given to problems improving the theoretical
background. Therefore standard computational exercises are supple-
mented by examples and problems explaining the theory, promo-
ting its deeper understanding and stimulating precise mathema-
tical thinking. Some counter-examples explaining the need for cer-
tain conditions in the formulation of basic theorems are also in-
cluded.
The book is designed along the following lines. Each section
opens with a concise theoretical introduction containing the prin-
cipal definitions, theorems and formulas. Then follows a detailed
solution of one or more typical problems. Finally, problems with-
out solution which are similar to those solved but
are given,
contain certain peculiarities. Some of them are provided with hints.
Each chapter (except Chap. IV and V) closes with a separate
section of supplementary problems and questions aimed at reviewing
and extending the material of the chapter. These sections should
prove of interest to the inquiring student, and possibly also to
lecturers in selecting material for class work or seminars.
The full solutions developed in the text pursue two aims: (1)
to provide lecturers with a time-saver, since they can refer the
students to the textbook for most of the standard exercises of a
computational character and concentrate mainly on the solution
of more sophisticated problems, thus gaining time for more rewar-
10 From the Author
ding work; and (2) to meet the needs of those who are working
on their own or following correspondence courses, providing a sub-
stitute for the oral explanations given to full-time students.
The student will find the book most useful if he uses it acti-
vely, that is to say, if he studies the relevant theoretical material

carefully on to the worked-out solutions, and finally
before going
reinforces the newly-acquired knowledge by solving the problems
given for independent work. The best results will be obtained
when the student, having mastered the theoretical part, immedia-
tely attacks the unsolved problems without referring to the text
solutions unless in difficulty.
Isaac Mar on
Chapter
INTRODUCTION
TO MATHEMATICAL ANALYSIS
§ 1.1. Real Numbers.

The Absolute Value of a Real Number
Any decimal fraction, terminating or nonterminating, is called
a real number.
Periodic decimal fractions are called rational numbers. Every
rational number may be written in the form of a ratio, , of two
integersp and q and vice versa.
y
Nonperiodic decimal fractions are called irrational numbers.

If X is a certain set of real numbers, then the notation x£X
means that the number x belongs to X, and the notation x^X
means that the number x does not belong to X.
A set of real numbers x satisfying the inequalities a < x < b y
where a and b are fixed numbers, is called an open interval (a, b).
A set of real numbers x satisfying the inequalities is a^x^b
called a closed interval [a, b]. A set of real numbers satisfying
x,
the inequalities a^x<b or a<x^.b, is called a half-open in-
terval [a, b) or (a, b]. Open, closed, and half -open intervals are
covered by a single term interval.
Any real number may be depicted as a certain point on the
coordinate axis which is called a proper point. We may also intro-
duce two more, so-called improper points, +°° and oo infinitely —
removed from the origin of coordinates in the positive and nega-
tive directions, respectively. By definition, the inequalities oo — <
<x<+ oo hold true for any real number x.
The interval {a e, a — +
s) is called the e-neighbourhood of the
number a.
The set of real numbers x > M is called the M -neighbourhood
of the improper point +00.
The set of real numbers x < M is called the M- neighbourhood
of t e improper point —-00.
The absolute value of a number x (denoted \x\) is a number
that satisfies the conditions
\x\ = —x if x < 0;
\x\ =x if xÔ.
12 Ch. I . Introduction to Mathematical Analysis
The properties of absolute values are:

(1) the inequality j
a;
|
â means that — a^xâ;
(2) the inequality |x|>a means that xâ or jc<- a;
(3) x±y\K \x\+\y\;
(4) x± y\>\\x\ — \y\\;
(5) xy\=\x\\y\
x
(6)
y
1.1.1. Prove that the number

0.1010010001... 1000... 01...
is irrational.
Solution, To prove this, it is necessary to ascertain that the
given decimal fraction is not a periodic one. Indeed, there are n
zeros between the nth and (n+\)th unities, which cannot occur
in a periodic fraction.
1.1.2. Prove that any number, with zeros standing in all deci-
mal places numbered 10" and only in these places, is irrational.
1.1.3. Prove that the sum of, or the difference between, a ra-
tional number a and an irrational number P is an irrational
number.
Solution. Consider the sum of a and p. Suppose a P y is a + =
rational number, then P Y =
a is also a rational number, since
it is the difference between two rational numbers, which contra-
dicts the condition. Hence, the supposition is wrong and the number
a+ P is irrational.
1.1.4. Prove that the product a|J and the quotient a/[i of a
rational number a ^= and an irrational number P is an irrational
number.
1.1.5. (a) Find all rational values of x at which y=\/ x* + x +3
is a rational number.
Solution, (a) Suppose x and y=\/~x <iJr x Jr 3 are rational num-
bers. Then the difference y x — =q is also a rational number. Let
us now express x through q
y — x=\/ x + x-{-3 — x —
r 2
q,
Vx
2
+ x+3 = q + x,
x2 + x + 3 = q* + 2qx + x\
x ~ \-2q
•
By a direct check it is easy to ascertain that q^ 1

/^
§ LI. Real Numbers. The Absolute Value of Real Number 13
Prove the reverse, namely, y = ]/ x + x + 32

is a rational number
if x = i
where q is any rational number not equal to 1
/2 .
Indeed,
g4-2?» + 7g*-6? + 9 (^-^ + 3)2 _ ^_^ + 3 ,

q=*=
_1_\
K (1-2?) 2 K (l-2^)a ~|1-2<7| \ 2J'
The latter expression is rational at any rational q not equal to 1
/2 .
(b) Prove that \/~2 is an irrational number.

1.1.6. Prove that the sum V~3 + V2 is an irrational number.
Solution. Assume the contrary, i.e. that the number [/3 + |/*2
is rational. Then the number
is also rational, since it is the quotient of two rational numbers.

Whence the number
K2 = l[(K3 + K2)-(/3-K2)l
is rational, which contradicts the irrational nature of the number
V2 (see Problem 1.1.5). Hence, the supposition is wrong, and the
number K3 + 1/2 i s irrational.
1.1.7. Prove that for every positive rational number r satisfying

<
the condition r 2 2 one can always find a larger rational number
r+ h(h>0) for which (r + hf < 2.
Solution. We may assume h<\. Then h 2
h < and (r+ < /i)
2
<r + 2r/z + That

2
/i. is why it is sufficient to put r
2
+ 2rh + h = 2,
i.e. ft = (2 — r 2
)/(2r+\).
1.1.8. Prove that for every positive rational number s satisfying
the condition s
2
> 2 one can always find a smaller rational number
s — fc(fc>0) for which (s — k) >2
2.
1.1.9. Solve the following inequalities:

(a) —3 <
\
2x |
1;
(b) (x-2) 2 >4;

(c) x + 2x — 8<0;
2
\x — 7x+ 12 >x — 7x+

2 2
(d) 12.
Solution, (a)
|
The inequality |2# — 3|<1 is eqivalent to the

inequalities
— < 2x — 3 <
1 1,
whence
2 < 2* < 4 and 1 <*< 2.
(d) The given inequality is valid for those values of x at which

x2 — 7x+ 12 < 0, whence 3 <x< 4.
1.1.10. Find out whether the following equations have any

solutions:
(a) \x\=x b] (b) |x|=x +5? —

Solution, (a) we have x AtxÔ = x + Hence, there are no 5.
solutions. At x we have x=-x < — + 5, whence x== — 5/2. This
value satisfies the initial equation.
(b) At x>0 we have x = x — Hence, there are no solutions.
5.
At x <
we have — x = x — 5, whence # = 5/2, which contradicts
our supposition (#<0). Thus, the equation has no solution.
1.1.11. Determine the values of x satisfying the following equa-

lities:
x—\ X— 1 .
(a)
x+ 1
(b) 5* + 6| = — (x — 5*+6). 2
1.1.12. Determine the values of x satisfying the following equ-

alities:
(a) + 4x + 9) + (2x — 3) - x + 4x + 9 +
(x
2
| |
2
1 |
2x —3 |;
(b) |(x — 4) — (x + 2)| = |x — 4 — |x + 2|.

|
4 2 4 2
|
Solution, (a) The equality |a + 6| = |a| + |fe| is valid if and

only if both summands have the same sign. Since
x2 + 4*+9:=(.x:+2) + 5 > 2
at any values of x the equality is satisfied at those values 9

of x
at which 2x 3>0, i.e. at *>3/2. —
(b) The equality \a 6| = |a| \b\ holds true if and only — if a
and b have the same sign and
In our case the equality will hold true for the values of x at
which
xx — 4>jc + 2
2.
Whence __
x2 — 2>1; \x\^V$-
1.1.13. Solve the inequalities:
(a) 3* — 5| — + 3 > 2JC 0;
— 5x| > — |5x|.
| | |
2 2
(b) \x |jc |
1.1.14. Find the roots of the following equations.

(a) sin x sin x 1; = +
— 2\x\ — 3 = 0.
| |
(b) x2
Solution, (a) This equation will hold true only for those values
of x at which sin*<0, that is why we may rewrite it in the
following way:
— sinjc = sinA;+ 1, or sin a: =— 1

/a ;
whence x = nk — (—l) n/6 (k = 0, k

±1, ±2, . ..).
(b) This equation can be solved in a regular way by considering
the cases and xÔ x^.0. We may also solve this equation re-
writing it in the form
|a:|
2
— 2|jc| — 3 = 0.
Substituting y for |
x |, we obtain
whence #, = 3, y2 =— 1. Since y = \x\^0, the value y 2 =— 1 does

not fit in. Hence
y=\x\ = 3,
i.e. = — 3, x 2 ^=3.
§ 1.2. Function. Domain of Definition

The independent variable x is defined by a set X of its values.
If independent variable x(tX there corres-
to each value of the
ponds one definite value of another variable y then t) is called y
the function of x with a domain of definition (or domain) X or,

in functional notation, y — y(x), or y f(x), or y = q>(x), and so =
forth. The set of values of the function y(x) is called the range
of the given function.
In particular, the functions defined by the set of natural num-
bers 1, 2, 3, are called numerical sequences. They are written
in the following way: x Xi x 2y ... or {x n \.
1.2.1.
2
Given the function / (x) =(*+ \)/(x— 1). Find f(2x), 2f(x),
fix ), [f(x))>.
Solution.
f(20=§±i; 2f W _2£±};
/<«>-£!; tfW]'-(f±{)'.
1.2.2. (a) Given the function
Show that at x lt x 2 £(— 1, 1) the following identity holds true:
/(*>+/(*)=/ (t££-).
16 Ch. I. Introduction to Mathematical Analysis
Solution. At 1, 1) we have (1 — x)/(l + x) > and hence
On
f(x l)+ f
the other hand,

M= log
^ + log = log ;; ;g .
(
i,
^1 + ^2
^
*l + *2 \ __
j Qg
1 ~f- XXX 2 _ j Qg
1 —X —X x 2
^
\l-jrX1X2J -^l ~l~ ^2 1 -\~ x i x 2~i~~ x i ~h x 2
j
1 -j- XiX 2
(1— *i) (\—x 2 ]
log
(!+*!) (l+x 2 )
which coincides with the right-hand member of expression (1).

(b) Given the function / (x) = (a x + a' x )/2 (a > 0). Show that
f{x + y) + f(x-y) = 2f(x)f{y).

1.2.3. Given the function f (x) = (x+ l)/(x»— 1). Find /(— 1);
f(a+\); f(a)+\.
1.2.4. Given the function f(x) =x — 3
1. Find
/(»>-/(">
6 —a ( ^ a) and ^«±*.
1.2.5. Given the function
(
3-^—1, —\<Cx< 0,
= J tan(*/2), 0<*<jx,
[
x/(jc
2
— 2), jt<*<6.
Find /(-l), f(ji/2), /(2ji/3), /(4) f /(6).
Solution. The point x = — 1 lies within the interval [ — 1, 0).
Hence
^ ( — = 3-1)
(
-J
> — = 1 2.
The points x = n/2, x = 2n/3 belong to the interval [0, ji).
Hence
/ (ji/2) - tan (ji/4) - / (2n/3) = tan (jt/3) - 1 ;
\^3.
The points jc = 4, x = 6 belong to the interval [ji, 6]. Hence
f (
4) = 16^2 = T ;
= 36-2
= 17 '
1.2.6. The function f (x) is defined over the whole number scale
by the following law:
2x 3 + 1. if jc<2,
f(x)-{ l/(*-2), if 2<*<3,
2* — 5, if x> 3.
Find: f(V2) y /(|/8), f (/log, 1024).

1.2.7. the square
In with side ABCD
a straight line AB=2
MN is drawn perpendicularly to AC. Denoting the distance from
the vertex A to the line as x, express through x the area S MN
of the triangle AMN
cut off from the square by the straight
line MN. Find this area at x = ]^2/2
B
and at x =2 (Fig. 1).
Solution. Note that AC = 2 V~2 \
hence 0<*<2l/T. If*<K"2,

then
S (X) — S^ AMN —X 2
-
If x> V2 y
then
S(x) =4 — (2 K2~— -^)

2
=— jc
2
+
+4* j/y 4.
Thus, Fig. 1
S(*) = 0<x</2",
-x \\x\f2 —4,
2
V2< x^2\/2.
Since |/T/2 < K2~ = {V~2 j2) = V Since
, S (V2/2) 2
2 . 2 > J/
2~,
S (2) = — 4 + 8 V~2— 4 = 8 (K~2~— 1).
1.2.8. Bring the number a„, which is equal to the nth decimal
place expansion of V~2 into a decimal fraction, into cor-
in the
respondence with each natural number n. This gives us a certain
function an = tp (ri). Calculate q)(l), cp (2), q)(3), q)(4).
Solution. Extracting a square root, we find \f2 =1.4142... .
Hence
q>(l) = 4; cp(2)-l; q>(3) = 4; cp (4) = 2.
f (x) = 49/x + x at the points for which 7/x + x=3.
2 2
1.2.9. Calculate
Solution. f(x) = 49/x + x = (7/x + x) —\4,but 7/x + x = 3, hence
2 2 2
/ (a') =9 —
14 = — 5.
1.2.10. Find a function of the form / (x)
= ax' 1
-\- bx -f c, if it is
known that /(0) = 5; / ( — = 1) 10; / (1) = 6.

Solution.
= 5=a'0 + b-0 + c,
f(0)
2
f(—l)= I0â — b + c,
/(l)=6 = a + 6 + c.
Determine the coefficients a, b c from the above system. y

We have:
a = 3; b=r.—2; c =
5\ hence / (x) = 3x
2
2x 5. — +
1.2.11. Find a function of the form
f(x) = a + bc* (c>0),

if /(0)=15; f(2) = 30; f(4) = 90.
1.2.12. Find cp [xp (a:)] and [cp (jc)] if
cp (jc) =x 2
and ^(x) = 2x .
Solution.
i|)[<p(*)] = 2* < * ,
= 2*".
5x2+1
f(x)~ 2—x '
3
Find f(3x); f (* ); 3f (x); [f (x)]\
1.2.14. Let
(3*
at — < x<1 0,
4 at 0<*<1,
3x— 1 at l<x<3.
Find f(2), /(0), /(0.5), /(-0.5), /(3).
1.2.15. Prove that for an exponential function y

if a x (a>0; =
a=£l) the values the argument x ofx n (fl=l, 2, ...) form an =
arithmetic progression, then the corresponding values of the func-
tion y n =
aXn (n= 1, 2, ...) form a geometric progression.
1.2.16. f + 6, q)(#) = 5*. Solve the equation / (x) = |

q) (x) |
1.2.17. Find /(*) if
/(*+l) =* a
— +2 3jc .
1.2.18. Evaluate the functions
/ (
X) = x + l/x 2 2
and (p (x) = x* + 1 /*
4
for the points at which l/x + x = 5.

1.2.19. f (x) = x+l; = x — 2; solve the equation
y(x)
\f(x) + <f(x)\ = \f(x)\ + \<t(x)\.
A rectangle with altitude x is inscribed in a triangle

1.2.20.
ABC with the base b and altitude h. Express the perimeter P and
area 5 of the rectangle as functions of x.
1.2.21. Find the domains of definition of the following functions:

(a) / (
X) = \/It=l+V6=i;
(d) f(x) = Vs\nx— 1 ;
(e) log^^
(0 f(x) = \og x 5\
— 5x-\-6 x'2
(g) /W = log- -4x + 6 v
(h) /(*) = arc sin — log — (4 x)\
(j) /W = logcosx;
3
(k)/(x)ârccos
4 + 2 sin;c ;
(i) y
Solution, (a) The domain of definition of the given function

consists of those values of x at which both items take on real
values. To ensure this the following two conditions must be satis-
fied:
1 >0,
6 — xÔ.
By solving the inequalities we obtain x^l; x^.6. Hence,
the domain of definition of the function will be the segment [1,6].
(e) The function is defined for the values of x for which
log^^->0.
This inequality will be satisfied if
-5^1>1, or x 2 — 5x + 4<0.
Solving the latter inequality, we find 1â:^4. Thus, the seg-
ment [1,4] the domain of definition of the function.
is
(f) The function is defined for all positive x different from unity,
which means that the domain of definition of the function consists

of the intervals (0, 1) and (1, oo). +
(k) The function is defined for the values of x for which
-1 ^ 4+2 A ,
3
.
sin x
<
^ 1.
Since 4+2 sin x >0 at any x, the problem is reduced to solving

the inequality
4 + 2 sin x ^
Whence
3<4+-2sinx, i.e. sinx>-— 7 2
.
By solving the latter inequality we obtain
_£ + 2£:ri<x<^ + 2fcn; (k = 0, ±1, ±2, ...).
(1) The function is defined for the values of x for which

|*| — *>0, whence |
a: |
> x. This inequality is satisfied at x 0. <
Hence, the function is defined in the interval (— oo, 0).
(a) = j/arc sin (log x)

/ (x) 2 ;
(b) / (x) = log log log x;

2 3 4
(C) / W
= l + 2 a-in, + _J_.
(d) f (x) = log|4-x 2
|;
(e) f (x) =V cos (sin x) +- arc sin ~£

x
.
Find the ranges of the following functions:
f
( ) &
= 2- cos 3* ;
(g) y=j^ •
Solution, (a) For the function /(x) to be defined the following

inequality must be satisfied
arc sin (log 2 x) ^ 0,
whence ^
log 2 x^ 1 and l^x^2.
(b) The function log 2 log 3 log 4 x is defined for log 3 log 4 x>0,
whence log 4 x>l and x>4. Hence, the domain of definition is
the interval 4<x< +

oo.
(c) The given function is defined if the following inequalities are
satisfied simultaneously:
x=^=0; — < x< 1 1 and x>2,

but the inequalities — l^x^l and are incompatible, that x>2
is why the function is not defined for any value of x.
(e) The following inequalities must be satisfied simultaneously:

+ x2
cos (sin x)^0 and
1
2x
< 1.
The first inequality is satisfied for all values of x, the second,

for |*|=1. Hence, the domain of definition of the given function
consists only of two points x=±\.
(f) We have
cos 3x ~— —
Since
-1
< cos3x< 1, we have — —-< 1,
whence, taking into account that y > 0, we obtain

— y^2y— or y 1.
(g) Solving with respect to x y

we obtain
x - Ty
'
The range of the function y will be determined from the relation

1 — 4*/
2
>0.
Whence
1.2.23. Solve the equation
arc tanVx(x + 1)+ arc sin l^x 2 + x+ 1 = jx/2.
Solution. Let us investigate the domain of definition of the func-

tion on the left side of the equation. This function will be defined
for
x2 + *>0, 0< x + x+ < 2
1 1,
-\-x = 0.
2
whence x
Thus, the left member of the equation attains real values only
at x =0 and # 2
t
=— 1. By a directcheck we ascertain that they
are the roots of the given equation.
This problem shows that a study of domains of definition of a
function facilitates the solution of equations, inequalities, etc.
, 2x— 3
(a)
x
y= r « _ ;
(b) y = \og sin (a:— 3) + j/l6 — x\

(c) y = V 3 — x + arc cos—^—
1.2.25. The function f (x) is defined on the interval [0, 1]. What
are the domains of definition of the following functions:
(a) /(3r>); (b) f(x-5); (c) f(tan*)?
Solution. The given functions are functions of functions, or su-

perpositions of functions, i. e. composite functions.
a) Let us introduce an intermediate argument u 3x 2 Then the = .
2
function f(3x )j=f(u) is defined if i.e. 2
<l, 0<w<l, 0<3x
whence — 1/]/3<*<; l/|/3.
(c) Similarly: O^tanx^l, whence
fcrt< + (k = 0, +1, ±2, ...).
1.2.26. The function f (x) is defined on the interval [0, 1]. What
are the domains of definition of the functions
(a) f (sin*); (b) f(2x + 3)?
§ 1.3. Investigation of Functions
A function f (x) defined on the set X is said to be non-decreasing

on this set (respectively, increasing, non-increasing, decreasing), if
for any numbers x 19 x 2 £X, x ± <x 2y the inequality f (xj^ f (x 2 )
(respectively, fW<f(4 f(*i)>f(* 2 ). /(*i)>/(* 2 )) is satisfied.
The function said to be monotonic on the set
f (x) is if it pos- X
sesses one of the four indicated properties. The function f (x) is
said to be bounded above (or below) on the set if there exists a X
number M
(or m) such that for all f(x)^M
(or m^Lf(x) x£X
for all x£X). The function f (x) is said to be bounded on the set X
if it is bounded above and below.
The function f (x) is called periodic if there exists a number
T>0 such that / (x -f-
— f (x)
x belonging to the domain
T) for all
of definition of the function (together with any point x the point
x-\-T must belong to the domain of definition). The least number T
possessing this property (if such a number exists) is called the
period of the, function f (x). The function f (x) takes on the maxi-
mum value at the point x $X
if f(x )^f(x) for all x£X,
Q
and
the minimum value if f(x )^f(x) for all x£X. A function f (x)
defined on a set X which is symmetric with respect to the origin
of coordinates is called even if /( x) = f(x), and odd if f (— x) =
= -/(*).
In analysing the behaviour of a function it is advisable to de-
termine the following:
1. The domain of definition of the function.

2. the function even, odd, periodic?
Is
3. The zeros of the function.
4. The sign of the function in the intervals between the zeros.
5. Is the function bounded and what are its minimum and ma-
ximum values?
The above items do not exhaust the analysis of a function, and
later on their scope will be increased.
1.3.1. Find the intervals of increase and decrease of the func-
tion / (x) =
ax 2 -f- bx c +
and its minimum and maximum values.
y
Solution. Isolating a perfect square from the square trinomial,

we have
If a > 0, then the function f(x) will increase at those values of x

satisfying the inequality x + b/(2a) > 0, i. e. at x > b/(2a), and
decrease when x + b/(2a) < 0, i.e. at x<— b/(2a). Thus, if a > 0,
the function f (x) decreases in the interval (

— oo, — — ) and inc-
reases in the interval ( b/(2a), +oo). Obviously, at x = — b/(2a)

the function / (x) assumes the minimum value
At a >
the function has no maximum value.
Similarly, at a <
the function f (x) will increase in the inter-
val — oo, — ^ and decrease in

^
the interval ( b/(2a), oo); at
x = — b/(2a) the function takes f (x) on the maximum value

£ r / b \ 4ac — b*
whereas it has no minimum value.

1.3.2. (a) Find the minimum value of the function
y = 3x + 5x — 1.
2
(b) Find the rectangle with the maximum area from among all
rectangles of a given perimeter.
Solution, (a) Apply the results of Problem 1.3.1: a — 3 0, b=5,
>
c =— 1. The minimum value is attained by the function at the
point x = — 5/6
4ac — b* 37
4a 12'
(b) We denote by 2p the length of the perimeter of the required

rectangle, and by x the length of one of its sides; then the area 5
24 Ch. J. Introduction to Mathematical Analysis
of the rectangle will be expressed as

S = x(p — x) or S = px —x 2
.
Thus, the problem is reduced to the determination of the maximum

value of the function S (x) = x2 — +
px. Apply the results of Prob-
lem 1.3.1: a =—
1 < 0, b p, c = =
0. The maximum value is attai-
ned by the function at the point x b/(2a)=p/2. Hence, one =—
of the sides of the desired rectangle is p/2, the other side being
equal to p x — =
p/2, i.e. the required rectangle is a square.
1.3.3. Show that

(a) the function f (x) —x 3
+ 3x + 5 increases in the entire domain
of its definition;
(b) the function g(x) x/(l = +
x2 ) decreases in the interval (1, +oo).
Solution. The function is defined for all points of the number
scale. Let us take arbitrary points x x and x29 x x x 2 on the number <
scale and write the following difference:
/ (x t )-f (x x ) = (4 + 3x + 5) - (x\ + 3x, + 5) =

2
= (*2 — *l) (A + X X + X\ + 3) = 1 2
Since x 2 xx — >
and the expression in the brackets is positive
at allx 1 and jc 2 then / (x 2 )
,
f (xj 0, —
i.e. / (* 2 ) >f(Xj), which>
means that the function f (x) increases for all values of x.
1.3.4. Find the intervals of increase and decrease for the follo-
wing functions:
(a) / (x) = sin #-|-cos x;
(b) tan(*+n/3).
Solution, (a) Using the familiar trigonometric formulas, we find
f (x)
— ]/^ 2 cos (x— ji/4).
It is known that the function cos a: decreases in the intervals
2nn ^.x^(2n+\) n
and increases in the intervals
(2n—\)ji<^x^2nn (n=0 t ± 1, ±2, ...).
Hence, the intervals of decrease of the function / (x) are:
ji/4 + 2azji< n/A + (2n+ 1) ji (ai = 0, +1, .

..).
and the intervals of increase of the same function are:
n/A + (2n— 1) k< x< n/4 + 2nn (n = 0, ± 1, . ..).
1.3.5. Find the minimum and maximum values of the function
j (x)
= a cos x-\-b sin x (u b 2 > 0).
1
-J-
Solution. The given function can be represented as:
f (x) = Vcl 1
-V b cos (x
2
— a),
where cos a = a/)/ a + b 2 2
, sin a = 6/ Ka + b'\ 2
Since |cos(x — a)|^ 1,
cos(a: — a)-=l),
r
the maximum value of f (x) equals +\/ a 2
+ t)
z
(at
the minimum value of f (x) being equal to j/a 2 + fr

2
(at
cos (x—a) = — 1).
1.3.6. Find the minimum value of the function

f(x)
=3<*'- 2,,+ 8 .
Solution. We denote by (p (x) the exponent, i. e.
(p(*) = (*
a
— 2)
3
+ 8.
The function f(x)=-3w takes on (x)

the minimum value at the
same point as the function (p (#).
Hence
cp (x) =x« — 6jc 4 + 2x 2 1 =x 2
[(x
2
— + 3)
2
3].
Whence it is clear that the function (p (x) attains the minimum

value (equal to zero) at x 0. That is = why the minimum value of
the function f (x) is equal to 3° 1. =
1.3.7. Test the function
f (x) = tan x + cot x, where <x< ji/2,
for increase and decrease.
1.3.8. Given: n numbers a,, a2 ,

an . Determine the value
of x at which the function
f (x) = ( x - ai )
2
+ (x-a + 2)
2
. . .
+ (x-a n f
takes on the minimum value.
Solution. Rewrite the function f (x) in the following way:
f(x) = nx 2
— 2(a, + a + 2
. . . +a n )x+(a +al+ 2
l
. .
wherefrom it is clear that f (x) is a quadratic trinomial ax 2 + bx c, +

where a = /t>0. Using the results of Problem 1.3.1, we find that
the function assumes the minimum value at x = b/(2a) i. e. at — 9
x= (aj +a + 2
. . . +a n )/n.
Thus, the sum of the squares of deviations of the value of x
from n given numbers attains the minimum value when x is the
mean arithmetic value for these numbers.
1.3.9. Which of the given functions is (are) even, odd; and
which of them is (are) neither even, nor odd?
(a) f(x) = \og(x+V\+x*);

(b) = log^l+x'
/(x)
f(x) = 2x — x+l;
s
(c)
(d) /W = *S^T-
Solution, (a) It can be seen that /(+ x) + f (— x) ^=0. Indeed,
/(+ x) + f(- ^)-log(^ + KTT^ + log(-^+KTT^ = 2

)
2
)
= log ( 1 +* 2 -- x 2 )=0,
hence, / (x) = — / — *) for all ( a:, which means that the function is odd.
v =—
1
(b) m_ ^ = iog = log \+x)
io 6
g
1
1+jc'
Thus, f(—x) = — f(x) for all x from the domain of definition

(— 1, 1). Hence, the function is odd.
1.3.10. Which of the following functions is (are) even and which
is (are) odd?
(a) f(x) = 4 — 2x* + sm 2
x\
(b) f(x) = V \+x + x

2
— \r \— x + x 2
;
(d) f (x) = sin x + cos x;

(e) f (x) = const.
1.3.11. Prove that if f (x) is a periodic function with period T 9
then the function / (ax + b), where a > 0, is periodic with period T/a.
Solution. Firstly,
f[a(x+ T/a) + b]=f [(ax + b) + T] = f (ax + b) y
since T is the period of the function f(x). Secondly, let 7\ be a

positive number such that
f[a(x+T 1) + b]=f(ax + b).

Let us take an arbitrary point x from the domain of definition
of the function f (x) and put x' (# b)/a. Then = —
/ + = / (ai=^+ + T + b] = ft) =/(x) =f [a {x' l )
= f(ax' + b + aT = f(x + aT l) l ).
Whence follows that the period T<a7\, i.e. 7\>7> and

it
T/a the period of the function f(ax + b).

is
Note. The periodic function (x) = A sin (ayx + where 9 / cp), /I, co,
are constants, is called a harmonic with amplitude |j4|, frequency co

and initial phase (p. Since the function sin* has a period 2k, the
function A sin(o)A: + (p) has a period T 2n/(s>. =
1.3.12. Indicate the amplitude \A\, frequency co, initial phase (p
and period T of the following harmonics:

(a) / (x)
= 5 sin \x\
(b) /(*) = 4sin(3* + n/4);
(c) /(x) = 3sin(jt/2) + 4cos(x/2).
1.3.13. Find the period for each of the following functions:
(a) / (a;)
= tan 2x;
(b) f(x)=cot(*/2);
(c) \(x)^s\x\ 2nx.
Solution, (a) Since the function tanx has a period Jt, the function
tan 2x has a period n/2.

(a) f (x) = s\n* x + cos 4
x;
(b) /(a:) = |cosa:|.
Solution, (a) sin 4 x + cos x= (sin x + cos
4 2 2
x)
2
— 2 sin 2
x cos 2 x =
= 1— y sin 2
2x = 1 — cos 4x) = -| + -j-sin ^4x + y) ;
whence T = 2n/a = 2n/4 = n/2 .
(b) /(x)= |cosjc| = Kcos x^K(1 + cos2x)/2; but 2

the function
cos 2.x; has a period T = n\ hence, the given function has the same
period.
Prove that the function / (x) = cos x 2 is not a periodic one.

1.3.15.
Solution. Let us prove the contrary. Suppose the function has a
period T\ then the identity cos (x T) 2 cos x 2 is valid. + =
By the conditions of equality of cosines for a certain integer k
we have
x2 + 2Tx + T ±x = 2nk. 2 2
But this identity is impossible, since k may attain only integral

values, and the left member contains a linear or quadratic function
of the continuous argument x.
1.3.16. Find the greatest value of the function
2
f(x)= r = .
1.3.17. Which of the following functions are even, and which are
odd:
3
(a) /W=^(l-x) + /(l+x) 2 2
;
(b) f(x) = x*-\x\;

(c) / (x)= x sin x — x 2 3
;
(d) /(x) = (l + 2*) /2*? 2

(a) / (x) = arc tan (tan x)\
(b) f (jc)-2cos^.
1.3.19. Prove that the functions
(a) f(x) = x + s\nx; (b) f (x)
= cos Vx
are non-periodic.
§ 1.4. Inverse Functions

Let the function y =
f(x) be defined on the set X and have a
range V\ If for each y£Y
there exists a single value of x such that
f(x)=y then this correspondence defines a certain function x — g(y)
y
called inverse with respect to the given function y f(x). The suf- =
ficient condition for the existence of an inverse function is a strict
monotony of the original function y f{x). If the function increases =
(decreases), then the inverse function also increases (decreases).
The graph of the inverse function x g(y) coincides with that of =
the function y =
f(x) if the independent variable is marked off along
the (/-axis. If the independent variable is laid off along the x-axis,
i.e. if the inverse function is written in the form y = g(x) then y
the graph of the inverse function will be symmetric to that of the

function y = / (x) with respect to the bisector of the first and third
quadrants.
1.4.1. Find the inverse to the function y 3x 5. = +

Solution. The function y Sx = +
5 is defined and increases through-
out the number scale. Hence, an inverse function exists and in-
creases. Solving the equation y 3x 5 with respect to x we obtain= +
x = (y-5)/3.
1.4.2. Show that the function y k/x (k=^=0) is inverse to itself. =
Solution. The function is defined and monotonic throughout the
entire number scale except x =
0. Hence, an inverse function exists.
The range of the function is the entire number scale, except # 0. =
Solving the equation y =
k/x with respect to x, we get x k/y. =
1.4.3. Find the inverse of the function
y = \°ga(x + V*+l) (fl>0, 9

a¥= 1).
Solution. The function y = log a (x + }/~x + l) 2

is defined for all x,
since ]fx
2
+ >
1 \x\ y
and is odd [see Problem 1.3.9 (a)]. It increases
§ 1.4. Inverse Functions 29
for positive values of x, hence, it increases everywhere and has an

inverse function. Solving the equation
y=\og a (x + Vlfi+l)
with respect to x, we find
a? = x + Vx + 2
l; a-y= — x+Vx + 2
1,
whence
x = y (a^ — = sinh (*/ lna).
1.4.4. Show that the functions
f(x)=x 2 — x + 1, x^ 1/2 and cp (*) = 1/2 + j/x — 3/4

are mutually inverse, and solve the equation
x* —x+ 1= 1/2+ Vx — 3/4.

Solution. The function a; #=
x + 1 = (x— 1/2) 2 + 3/4 increases
2
—
in the interval 1/2â:<cx), and with x varying in the indicated
interval we have 3/4^*/<oo. Hence, defined in the interval
3/4<#<oo is the inverse function x = g(y), x^ 1/2, which is
found from the equation
x*-x + {l-y) = 0.
Solving the equation with respect to x, we obtain
x=g (y) = 1/2 + K^T? = q) (*/).
Let us now solve the equation
x* — x+l = l/2 + Vx — 3/4.

Since the graphs of the original and inverse functions can intersect
only on the straight line y x, solving the equation x
2
=
x+l = x —
we find x = 1
1.4.5. Find the inverse of y smx. =

Solution. The domain of definition of the function y s\nx is =
the entire number scale, the range of the function is the interval
—
[ 1, 1]. But the condition of existence of an inverse function is
not fulfilled.
Divide the x-axis into intervals tin n/2. If n is — n/2^x^nn +
even, then the function increases on the intervals tin — n/2^x^
^nn + n/2; if n is odd, the function decreases on the intervals
—
nn ji/2^ x^nn +
n/2. Hence, on each of the indicated intervals
there exists an inverse function defined on the interval [—1, 1].
In particular, for an interyal — jt/2 ^ ji/2 there exists an in-

verse function jt arcsinr/. =
The inverse of the function y =
s\x\x on the interval mi —
^*^wt+jt/2 is expressed through atcsin y in the following way:
x= (— l)
B
arcsin# + wr (ai = 0, +1, ±2, ...).
1.4.6. Find the inverse of the given functions:

(a) y = s\n(3x— 1) at — (jt/6+ l/3)< (jt/6+ 1/3);
(b) # = arcsin(*/3) at — 3<x< 3;
(c) # = 5 e*; lo
xix '
y=2
1}
(d) .
1.4.7. Prove that the function y = (l — x)/(l+x) is inverse to

itself.
§ 1.5. Graphical Representation of Functions
1.5.1. Sketch the graph of each of the following functions:

(a) f( x )=x* — 2x + 2
3;
(*>)
f(*)=TT#>
(c) f(x) = s\n x — 2sinx; 2
(d) / (x) = arc cos (cos x);
(e) f (x) = Vsm x;
f( X = X
(f)
°*x.x ' l
)
Solution, (a) The domain of definition of the function/^) is the

entire number scale. The function f (x) is even, hence its graph is
symmetrical about the ordinate axis and it is
sufficient to investigate the function at xÔ.
Let us single out a perfect square f(x) =
= (x 2 — +1)
2
2. Since the first summand
(x
2
— 1) ^0, the minimum value of the func-
2
tion, equal to 2, is attained at the points

x = ±l (see Fig. 2).
The function f (x) decreases from 3 to 2 on
the closed interval O^x^l and increases
unboundedly on the open interval 1 x oo. < <
Fig. 2 (b) The domain
func- of definition of the
tion fun-
f (x) is the entire number scale. The
ction / (x) is odd, therefore its graph is symmetrical about the origin
of coordinates and it is sufficient to investigate the function at
x ^0.
Since /(0) = 0, the graph passes through the origin. It is obvious
that there are no other points of intersection with the coordinate
axes.Note that |/(jc)|<1. Indeed, (1— U|) 2 >0or l+x 2 >2|*|,

whence
Since / (x) > at x >

and f (1) = 1, in the interval [0, oo) the
maximum value of the function f (x) equals 1, the minimum value
being zero (see Fig. 3).
-3 "2
I
-1
1
/
/
12
i i
3'
i
-/
Fig. 3
Let us prove that the function increases on the closed interval 0^

<*<1. Let 0^*,<*
2 <1.
Then
£ / x r / x 2a: 2 2x t 2x 2 -}- — 2x — 2x^5 "

2x 2 x\ x
Ml, ~l+4
,
fl,)
(l+*S)0 + *i)
^ 2 (x — s 2 t ) (1— x^)
(l+*I)(l+*f)
and f(x 2 ) >f(x l ).
Similarly, we can show that on the interval (1, oo) the function
decreases. Finally,
f (x) = 2x/( + x )<

1
2
2xi'x
2
= 2/x,
whence it is clear that f (x) tends to zero with an increase in x.
(c) The domain of definition of the function / (x) is the entire
number scale. The function has a period 2n, that is why it is quite
sufficient to investigate it on the interval [0, 2jx], where it beco-
mes zero at the points x = 0; x = n\ x = 2n.
Writing the given function in the form
f (*)
= (1— sin*) 2
— 1,
we note that it increases with a decrease in the function sin* and

decreases as sin* increases. Hence, the function / (*) decreases on
the intervals [0, n/2] and [3ji/2, 2n]> and increases on the interval
[jx/2, 3jx/2]. Since f(n/2) =—
1, and /(3ji/2)=3, the range of the
function is -!</(*)<
3 (Fig. 4).
(d) The domain definition of the function is the entire number

of
scale. Indeed, |cosx|^l at any x hence, arc cos (cos*) has a
y
meaning. The function f (x) is a periodic one with the period 2n y
hence, it is sufficient to sketch its graph on the interval [0, 2jx].

But on this interval the following
equality is true:
f(x)= l
*. 0<x< n ,
\ 2n — x, n^x^ 2jt.
Indeed, the first assertion follows

x from the definition of the function
arc cos x while the second one can be
9
proved in the following way. Let us

= 2n—x
put x' ji jk 2ji; then y < <
O^x'în and
/(*) = arc cos [cos (2n— x')] =arc cos (cos x') = x' = 2ji— x.
Taking all this into con sider ation, we draw the graph (see Fig. 5).
(e) The function y = ]/ s'mx is a periodic one with period 2n\
that is why we may confine ourselves to the interval [0, 2ji]. But
X
-2% -% Tt 2lt
Fig. 5
the function is not defined in the whole interval [0, 2ji], it is

defined only in the interval [0, ji], as in the interval (jx, 2n) the
radicand is negative. The graph is symmetrical about the straight
Fig. 6
line a: = ji/2, as well as the graph y =

s'mx (see Fig. 6). Here we
have an example of a periodic function which does not exist in
the infinite set of intervals.
(f) The domain of definition of the function is
<x< 1 and 1 < x < oo.

Reduce the formula to the form

f(x) = x xn °* x = x Xo** 10 = jo.
Hence, the graph of the given function is the half-line #=10

in the right-hand half plane with the point x = 1 removed (see Fig. 7).
1.5.2. Sketch the graphs of functions defi-
ned by different formulas in different inter-
vals (and in those reducible to them):
sin x at — jx<x<0, 10
(a) */=< 2 at 0< x< 1,
l/(x— 1) at < < 4;

1 a;
—2 at x>0,
(b)0= 1/2 at x= 0,
—x 3
at x < 0; Fig. 7
(c) y=x + V#;

_
(d) y=2/(x+\f?).
Solution, The domain of definition of the function is the
(a)
interval [
— The graph of the function consists of a portion
it, 4].
of the sinusoid y = s\x\x on the interval Jt< x 0, straight line — <
9,
2
1/2
{
1
-1 O 1
.1 1— -1
12
I i
3 4
-2
-i
Fig. 8 Fig. 9
i/ =2
on the interval (0, 1] and a part of the branch of the hyper-
bola y=\l{x—\)
on the interval (1,4] (see Fig. 8).
(b) The graph of the function consists of a portion of a cubic
parabola, an isolated point and a half-line (see Fig. 9).
(c) The function may be given by two formulas:
I 2*. if x>0,
y==
\ 0, if x<0.
Thus, the graph of our function is a polygonal line (see Fig. 10).
(d) From (c) it follows that the function defined only in the is
interval (0, +oo), y being equal Thus, the graph

to l/x (x > 0).
of our function is the right-hand part of an equilateral hyperbola
(see Fig. 11).
Fig. 11
1.5.3. Sketch the graphs of the following functions:

(a) ^ = cosx+|cosx|;
(b) y=\x + 2\x.
\ 2cosjc at cos* ^5 0,
Solution, (a) cosa; + |
cos*| =< q at COSA; <o.
Doubling the non-negative ordinates of the graph for the func-
tion y = cosx (the broken line in Fig. 12) and assuming y = at
X —
N
x
-7C Sic JL v % s'3n 2% Jax
' 2 2 ~T~ ~7T
Fig. 12
the points where cosx<0, we can sketch the desired graph (the
solid line in the same figure).
(b) The function + 2|x may be
|x given by two formulas:
=/ +
(x 2)x at — 2,
# \ —(x+2)x at — 2.
Plotting separately both parabolas: y = (x+2)x = (x+l) —l

2
y
and y = [(*— l)
2
+1], retain —
only the parts corresponding to
the above indicated intervals. Drawn in a solid line in Fig. 13 is
the graph of the given function, the broken line showing the de-
leted parts of the constructed parabolas.
1.5.4. Sketch the graph of the function
y = 2\x—2\ — x+ \
l\ + x.
Solution. At x ^2
y= 2(x—2) — (x + l) + x = 2x— 5.
At — l<x<2
y== — 2(x — 2) — (x+ + x = — \) 2x + 3.
19
\ 5
i /
-1
— \\V//'
1
'2/
1
3
>
-1
Fig. 13 Fig. 14
Finally, at x<.~ 1
y== -2(x-2) + (x+\) + x = 5.

Hence, the given function can be rewritten in the following way:
( 5, *<-l,
y=) — 2x + 3, —1 <x<2,
[
2x—5, x> 2.
Therefore the graph is a polygonal line
(see Fig. 14).

y^2 x —2~ x .
Solution. Draw graphs for the functions

yt =2 x and y2 = — 2~ x (broken lines in
Fig. and add graphically their ordina-
15), Fig. 15
tes. In doing so bear in mind that y 2 < y <
< y 19 and that y 2 tends to zero with an increase in x y

whereas y x
tends to zero with a decrease in x (the solid line in Fig. 15).

*/ = x sin x.
Solution. Being the product of two odd functions y

x
x and =
*/ 2 = sin*, the function y is an even one, that is why we shall
analyse it for xÔ.
We draw graphs for yx =x and y2 =smx (the broken lines in
Fig. 16).
At the points where r/
2
= sin# = 0, y = y 1 .y 2 = i
and at the
points where y 2 = smx = ± 1, y=± yi = ±x. The latter equality
Fig. 16
indicates the expedience of graphing the auxiliary function y 3 x. =—

Marking the indicated points and joining them into a smooth
curve, we obtain the required graph (the solid line in Fig. 16).
1.5.7. Sketch the graph of the function y x(x 2 = — 1) by multi-

plying the ordinates of the graphs y t and y 2 z=x 2 =x — 1.
1.5.8. Graph the following functions:

(a) y = x/(x 2 — (b) y = /arc cos x.
4), 1
Solution, Since the function

(a) odd, is it is sufficient to inves-
tigate it for x ^ 0.
Let us consider it as the quotient of the two functions:
y x
=x
and y = x —4. 2
2
Since at x =2 the denominator y = 0, the

2 function is not de-
fined at the point 2. In the interval [0, 2) the function y x increases
from to 2, the function y 2 is negative and|r/ 2 4 x 2 decreases |
=
from 4 to 0; hence, the quotient f(x)=yjy 2 is negative and in-
creases in absolute value, i.e. f(x) decreases in the interval [0,2)
from to — oo.
In the interval (2, oo) both functions are positive and increasing.
Their quotient decreases since from t 2
it follows that 2^x <x
x2 xt (*! —x x + 2 ) (a: 1 2 4) . n
The indicated quotient tends to zero as x — oo, since y = _^X 1

2
~^ 0-
The general outline of the graph is presented in Fig. 17 (three
solid lines).
(b) Denote y = arc cos x. The domain of definition of this func-
tion |
x
1
At x=\ we have y l = 0, hence, y=\/y l * oo at —
Fig. 17 Fig. 18
x—+ 1, i. e. x= 1 is a vertical asymptote. The function y decreases

on the entire interval of definition [
— 1, 1), hence y=\/y
l
l
incre-
ases. The maximum value y 1 =
n is attained at *--=— 1. Accor-
dingly, the minimum value of the function is 1/jx. The solid line
in Fig. 18 represents the general outline of the graph.
Simple Transformations of Graphs
I. The graph of function y

the f(x = +
a) is obtained from the
graph of the function y = f(x)
by translating the latter graph along
the x-axis by |
a |
scale units in the direction opposite to the sign
of a (see Fig. 19).
II. The graph y f(x) = +
b is obtained from the graph of the
function y =
f(x) by translating the latter graph along the y-ax\s
by \b\ scale units in the direction opposite to the sign of b (see
Fig. 20).
III. The graph of the function y f(kx)(k>0) is obtained from =
the graph of the function y f (x) by "compressing"
=
the latter graph
against the *y-axis in the horizontal direction k times at k 1 and >
by "stretching" it in the horizontal direction from the r/-axis \\k
times at k 1 <
(see Fig. 21).
IV. The graph of the function y =

kf (x) (k >
0) is obtained from
the graph of function y =
f(x) by "stretching" it in the horizontal
direction k times at k 1 >and "compressing" it against the #-axis
(i. e. vertically) \/k times at k 1 <
(see Fig. 21).
u>-0 / vr ,o yuf(x)
y=f(x)+l,l<0
Fig. 20
V. The graph of the function y= —

f(x) is symmetrical to that
of the function y —
f(x) about the x-axis, while the graph of the
function y = f( — x) is symmetrical to that of the function y f (x) =
about the y-axis.
Fig. 21 Fig. 22
VI. The graph of the function y = f(\x\) is obtained from the

graph of the function y = f(x) in the following way: for jc>0 the
graph of the function y =
f(x) is retained, then this retained part
of the graph is reflected symmetrically about the #-axis, thus de-
termining the graph of the function for jc<0 (see Fig. 22).
VII. The graph of the function y = \f(x)\ is constructed from
the graph y = f (x) in the following way: the portion of the graph
of the function y =
f(x) lying above the x-axis remains unchanged,
its other portion located below the x-axis being transformed sym-
metrically about the x-axis (see Fig. 23).
VIII. The graphs of the more complicated functions
y^'kf (kx+a) +b
are drawn from the graph of y = f(x) ap-
plying consecutively transformations I to V.
1.5.9. Graph the function
» = 3J/ -2(a; + 2.5) — 0.8

by transforming the graph y = ]fx.
Solution. Sketch the graph of the function
y=z\fx (which is the upper branch of the
parabola y 2 =
x) (Fig. 24, a), and transform
it in the following sequence.
Sketch the graph of the function y = 3]/2x Fig. 23
by enlarging 3 V 2 times the ordinates of

the points on the graph of the function y= \ x and leaving their
abscissas unchanged (see Fig. 24, b).
Then sketch the graph of the function y = —

3V 2x which will
be the mirror image of the preceding graph about the (/-axis (see
Fig. 24, c).
Fig. 24
By shifting the obtained graph 2.5 scale units leftward and then
0.8 u nit downwar d draw the desired graph of the function
y=3V— 2(jc + 2.5)— 0.8 (see Fig. 24, d).
1.5.10. Graph the function # = 3cos# — j/3sin;t by transforming

the cosine curve.
Solution. Transform the given function
y = 3 cos x— V 3 = 2 j/3~( — y sin

Q j-cosx
si n x '
jc
= 2 K3 cos
Thus, we have to sketch the graph of the function
#= 2VTcos (x + ji/6),
which is the graph of the function */ 2 K 3cosjc translated by ji/6 =
leftward. The function has a period of 2ji, hence it is sufficient to
draw its graph for — n ^Zx ^ ji
ZVT
\
// \ \
/ /
1
/' \ \
1
// \\ 1
%
> •X
/ fa/2 f\\7t/2
I l
\\
>^ -2V3
Fig. 25 Fig. 26
The graph of any function of the form */ = acosx + frsin x, where

a and b are constants, is sketched in a similar way.
1.5.11. Graph the following functions:
(a) y—î
 f-JTZ^
(c) ^ x2 + a:+ 1, if— l<x<0,
s sin 2 x if < x < k,
(x-l)/(x+l), if ft <x^ 5;
(d) = x+ \/x;
y
(e) y = x —x 2 3
;
(f) = sin
(g) y = 1/cos*;
(h) j/=3s in (2jc — 4);
(i) t/=2K — 3(x+ 1.5)— 1.2;
(j) y= |x« — 2x— 1|;

(k) y=\\x\-l\;
(1) j/ = cos(sinjc);
(m) y=\s'mx\ + s\nx on the interval [0,3k];
f 1 at x> 0,
(n) y = x sign x
2
y
where signx = V at x= 0,
— 1 at x < 0.
1.5.12. The function y f(x) is given graphically = (Fig. 26).
Sketch the graphs of the following functions:
(a) y = f(x+ 1);
(b) y = f(x/2);
(c) y=\f(x)\;
(d) » = (|/(*)|±/(*))/2;
(e) y = \f(x)\/f(x).
§ 1.6. Number Sequences. Limit of a Sequence

The number a is called the limit of a sequence x u x2y xn
... as n —
oo, a— lim x n if for any e>0 there exists a number
,
n -+ co
N(e)>0 such that the inequality |jt„ a\<& holds true for all
n> Af(e).
A sequence which has a finite limit is said to be convergent.
A sequence {x n } is called infinitely small if lim.*^ 0, and infi- —
nitely large if lim^ôo.
1.6.1. Given the general term of the sequence {xn }:
sin (nn/2)
Write the first five terms of this sequence.

Solution. Putting consecutively 2, 3, 4, 5 in n=\ 9
the general
term xni we obtain
sin (jt/2) «
#
l
1
~~
3 3
sin (4it/2) ~
: "" U;
4
sin(5jt/2) _
~~
1
*
5 5
1.6.2. Knowing the first several terms of the sequence, write one
of the possible expressions for the general term:
(3 '
I' 8 »
13 »
18 '
23
;
(b) 1, -g-, 2, -g- , 3, , 4, -g-.
Afofe. A knowledge of the first several terms of a sequence is not

sufficient to define this sequence. That is why this problem should
be understood as one of finding a certain simple inductive regula-
rity compatible with the given terms.
Solution, (a) Note that the numerator of each of the given terms
of the sequence equals the square of the number of this term plus
unity, i.e. n 2 +\, while the denominators form the arithmetic prog-
ression 3, 8, 13, 18, ... with the first term and the com- ^=3
mon difference d = 5. Hence,
a„ = a + d(n— = 3 H- 5
1 1 ) (az — = 5n —
l) 2,
thus we have
_ n* + l
*»~~5/i —2 '
(b) Here the general term of the sequence can be written with
the aid of two formulas: one for the terms standing in odd places,
the other for those in even places:
j k at n = 2k—l 9
xn =\ 1/(6+1) at n = 2k.
It is also possible to express the general term by one formula,

which will be more complicated, for instance,
^=^[i-(-i)i+^- [i+(-i)"]- 2
1.6.3. Find the first several terms of the sequence if the general
term is given by one of the following formulas:
(a) jc„ = sin (nn/3)\
(b) x n = 2~ n cos nn\
(c) x n =(l +
l/n)\
1.6.4. Using the definition of the limit of a sequence, prove that
(a) lim x n =l x n = (2n— l)/(2n+ 1),

if
(b) limx n = 3/5 if x n = (3n

2
l)/(5n 2 + — 1). Beginning with which
n is the inequality |
— 3/5 < 0.01 |
fulfilled?
Solution, (a) For any e > let us try to find a natural number
Af(e) such that for any natural number n>N(z) the inequality
\x n — l|<e is fulfilled.
For this purpose let us find the absolute value of the difference
2/1—1
-1
2/1+1 2/i+ 1
I
2n+
Thus, the inequality |
jc„ — <e 1 1
is satisfied if
2
^qrf <8 » whence
n > 1/e— V 2 - Hence the integral part of the number 1/e — V, may
be taken as jV(e), i.e. W = £(1/8 — 7,).
So, for each e > we can find a number A/ such that from the
inequality n >N it will follow that \x n 1 e, which means that — < 1
lim f^ = l.
(b) Let us find the absolute value of the difference — 3/5|:

3/z 2 +l 3 I 8
5/z 2 — 1 5 |
5(5/i 2 — 1)'
Let e > be given. Choose n so that the inequality
5(5/i 2 -l)
is fulfilled.
Solving this inequality, we find
_ 8 . 1 .1 t/8 + 5b
Putting
N = £ (i/?±5),
we conclude that at n> N
K-3/5|<e,
which completes the proof.
If e = 0.01, then
*-*(•{- -5.
and all terras of the sequence, beginning with the 6th, are contai-
ned in the interval (3/5—0.01; 3/5 0.01). +
1.6.5. Given a sequence with the general term x n = 9n _^ 4 . It is
known that lim x„= 1/3. Find the number of points xn lying out-
side the open interval
Solution. The distance from the point x n to the point 1/3 is

equal to
X»
1 19 _ 19
3 3(9/z + 4) 3(9/z + 4)
Outside the interval L there will appear those terms of the se-
quence for which this distance exceeds 0.001, i.e.
19
3(9/z + 4) ^ 1 000
whence
18988 ?n 7
1 <; n <
Hence, 703 points (x lt x2 , .
,
x 103 ) are found outside the inter-
val L.
1.6.6. Prove that the number 1 = is not the limit of a sequen-

ce with the general term xn (n
2
= — 2)/(2n — 2
9).
Solution. Estimate from below the absolute value of the diffe-
rence
n2 —2 -0
1^-2 1
2n 2 —9 |2/2
2 —9 |
At n ^
3 the absolute value of the difference remains greater
than the constant number 1 /2 hence, there exists such e > 0, say,
;
e= 1
/ 29 that the inequality
2ril —9 — >
1
holds true for any n^3.

The obtained inequality proves that / = is not the limit of the
given sequence.
1.6.7. Prove that the sequence
1,
112
^
1
with the general term

\/n, if /2-2/e— 1,
n/(n + 2), if n = 2k,

has no limit.
Solution. It is easy to show that the points x n with odd num-
bers concentrate about the point 0, and the points x n with even
numbers, about the point 1. Hence, any neighbourhood of the point 0,
as well as any neighbourhood of the point 1, contains an infinite
set of points x n Let a be an arbitrary real number. We can always
.
choose such a small e >

that the e-neighbourhood of the point a will
not contain at least a certain neighbourhood of either point or

point 1. Then an infinite set of numbers x n will be found outside
this neighbourhood, and that is why one cannot assert that all the
numbers x m beginning with a certain one, will enter the e-neigh-
bourhood of the number a. This means, by definition, that the
number a is not the limit of the given sequence. But a is an arbit-
rary number, hence no number is the limit of this sequence.
1.6.8. Prove that lim* n =l if xn = (3 n + l)/3

n
.
1.6.9. Prove that limx„ = 2 if x n = (2n 3)/(n 1). Find the + +

number of the term beginning with which the inequality
|
(2n + 3)/(/2+ 1) — < 2 e,
|
where 8 0.1; 0.01; 0.001, is fulfilled. =
1
J_ _L i. _L L
2 '
2 '
4 '
4 '
8
with the general term
I
l
-^rm if n is odd *
*»=f — i
j77
n/2
if n
.
is even,
{ 2
has no limit.
1.6.11. Prove that at any arbitrarily large a>0 limA;„ = if
xn = a n /nl
Solution. Let a natural number k > 2a. Then at n >k
an a \ a a
a a a f a a ( a
^x
n\~ \ 2 n \ 1 2 k ) \k+l k +2 n
<a.(i)--W.(^"
Since lim(l/2)" = (prove it!), then at a sufficiently large n we
have: (4-) <—— and, hence, a n
/n\ < e, which means that
V 2 J (2a) k
Y\m(a n /n\) = 0.
1.6.12. Test the following sequences for limits:
(a) x„=1/(2jx);
i 1 for an even n,
(b) xn — <j
^
(C) Xn = ~ C0S nn
1
~2" I
(d) ^„ = ai [!-(-!)»].
1.6.13. Prove that the sequence with the general term
x n =l/n* (k>0)
is an infinitely small sequence.
Solution. To prove that the sequence xn is infinitely small is to
pro\e that = 0.
lim x n
n -* oo
Take an arbitrary e > 0. Since \x n \

= l/n k , we have to solve the
inequality
l/n* < e,
whence n> y^l/s. Hence N may be expressed as the integral part

of j/T/e, i. e. = E(^1/S).
1.6.14. Prove that the sequences with the general terms
(a) xn = ,
(b) xn = ± sin [(2/i - 1)
are infinitely small as n —oo.

1.6.15. Show that the sequence with the general term
xn = (
— l)
n
2/(5f/n+l) is infinitely small as n—>oo. Find a num-
ber N
beginning with which the points x n belong to the interval
(-1/10, 1/10).
Solution. Take an arbitrary e>0 and estimate \x n \:
I
JC l-
»l""
2
^
^
2
3/-
^ 3/-
2
- 1
3/-'
l
c 3/ .
f
-
That is why |xj<e a; soon as n > 1/e

3
. Hence limxn = 0, i.e.
n oo
the sequence is infinitely small.

We take now 8=1/10. Since |^n |< l/\/n x n will necessarily y
be smaller than 1/10 if \lV~n< 1/10 or n 1000. Hence N may >

be taken equal to 1 000. But we can obtain a more accurate result
by solving the inequality
\
X n\ = 3/- < TO •
It holds true at n > (19/5) = 3. = 54.872. 3 3

Hence N may be
taken equal to 54^1000.
1.6.16. It is known that if x n a = +
a n where a n is an infinite- ,
simal as n—^00, then lim x n a. =

Taking advantage of this rule,
n -»• oo
find the limits:

3"+! + sin (wi/4) 2* + (— 1)"
w+1 -f- sin (nn/4)

c <
Solution, (a) x n
. .
=3 "s^
— — 3+ ,
,
where a n = sin (nn/A)

„
v
3
.
is
an infinitesimal as — oo, hence lim x n = 3.

n -+ oo
1.6.17. Prove that lim y^n=l.

n -+ <x>
as the
Solution. Let us prove that the variable
sum l+a„, where a n is an infinitesimal as n oo.
Yn can be represented
—
Let us put \/n= 1 + a n Raising to the nth power we obtain .
rc= (1 +aj n = 1 +Aia„ + ^=â^+ . . . +a"ny

wherefrom we arrive at the conclusion that for any n > 1 the fol-
lowing inequality holds true:
n>l + njn^l) a%
(since all the terms on the right are non-negative). Transposing the
unity to the left and reducing the inequality by n 1 we obtain —
whence
r
it follows that 2/n >a n
2
or > a n > 0. Since
lim V 2/n = i
lim an also equals zero, i. e. an is an infinitesimal.
n -* oo n oo
Hence it follows that
lim yn=l.
n -* oo
is infinitely large as >oo. n—

Solution. Let us take an arbitrary positive number M and solve
the inequality
z V n
> AJ.
Taking the logarithm, we obtain
Vn> log, /W, n> (log 3

M) 3 .
If we now take /V = E (log 3

A4) 3 , then for all n> N the inequa-
lity |x n |>A4 will be fulfilled, which means that the sequence is
infinitely large.
1.6.19. Prove that

lim y/a = 1 (a > 0).
n -* oo
§ 1.7. Evaluation of Limits of Sequences

If the sequences {x n \ and {y n \ are convergent, then
(1) ±
Km (x n y n ) liin x n lim = ±
(2) lim lim xn lim -
If x„ ^ */„ then lim x n ^ lim y n.
1.7.1. Find lim x„ if

n qo
^
W xvn~
3n 2 +5n
2 +n
+4
2
,
'
W n ~
5b» + 2«»-3/i + 7
4n 3 -2n+ll
.
'
— + +
4n 2 -4n 3 l»+2»+ ...+««
+ 4'
2/i3 3/i
.
w n_ 5/i3 + n+l
.
(e) *n =
1+2
+r + " -
3+A+l
n
Solution, (a) ~ 2_
S+
n*
1
lim x = — lim (3 +
7-o
5//z + 4//z
x
2
— -3.)
(d) Recall that

(2/i+l)
l* +2 +3 2 2
+... +n 2
Hence
_ + ^ A-f-l
2
/i(/i+1)(2/i+1)_ 2n 3 + 3n 2 + Ai n^ n 2
6(5/i* /i+1) 6(5n3 + n+l) + ^6 6»

30 _L
n1 u3
1
lim #„ = 1/15.
1.7.2. Find lim x ny if

ft -+ 00
/ 3n + n— \32 2n 3 + 2n +l \« 2
W
, ,
— ^4„2 + 2n + 7j
.
' W /
U" + 7 " + 3n + 4j
3 2 •
(c) xn = ybn; (d) xn = pi*;

{t)xn =yn>; (l)xn =^6n + 3.
Solution, (a) Urn (^^r)' =
_ m 3n + n — \ 3n + n — \ 3n + n — __
,. /'
2
/
2
/ 2
\
«i « U« + 2n+7j U« + 2« + 7j U/t + + 7j 2 2 2
2/l
= lira 3+1/rt — \ 3_\ 2//i 2 3 ' 3

_27
4+2/rt „ + 7/rc 2
,4 J -64
§ 17. Evaluation of Limits of Sequences 49
In solving this example, and also the rest of the examples

(c)
of Problem 1.7.2, take advantage of the following equalities (see
Problems 1.6.17 and 1.6.19):
lim i/n=l and lim \/a=\. (1)
We have
lim xn = lim i/5n = lim lim \/n,
n co n -+ <x> n -* co n -* co
but from (1) it follows that limv^5=l and lim y^/i = 1 ; hence
lim xn = 1 • 1 = 1
1.7.3. Find
' 2ft 3 .
1— 5ft 2 \
Solution. Summing the fractions, we obtain
xn—
2ft
3 — 13ft 2 +3
i Oft 3 + 2ft 2 + +3
15ft
*
Whence
_ 2ft 3 — 13ft 2 +3 _ 1
lim x n - r
iirn^
10n3+ 2 n 2 + i5 n+3- 5
•
n -+ oo
Note. If we put
_ 3
_ 1— 5ft 2
^ — 5ft+
2ft
— 2ft
2
+3 ;
1
'
then the limit of their sum ) though each of the \im{y n + z n = 1/5,
summands an infinitely large quantity. Thus, from the conver-
is
gence of a sum of sequences it does not, generally speaking, follow

that the summands converge too.
1.7.4. Find lim xn if

n -* co
(a) x„ = V2n+3 — Vri—l' 9
n
(b) xn = Vn + n+ —]/ n — n+ 1;2
\
2
(c) xn = n (n—Vn + \y,2 2
(d) xn =¥n — n + n\ 2 3
(ej xn _
— ]Tn*+\ + Vn— ,
.
4 /
, 1—2 + 3 — 4 + 5 — 6+ ... —2ft

(
(h) xn = ± + ± + ±+...+.
1-2 1
2-3 1
3-4 1
' ' * 1
ft(ft+l)
Solution, (a) xn = Vn(V2 +

3/n—Vl l/n)-+ -f oo as n — — oo
since the second multiplier has a positive limit.
(c) Xn
_ n*(n-VlF+\ _ ) -n* _
1
n+K/iHl
=—n 1
— — oo as n — oo.
(d) xn = ~ . =
(n 2 —n 3 2/3
)
— n y n — n* + n* *
It means, x n — >•
1/3.
Factoring out the terms of the highest power in the numera-
(e)
tor and denominator, we have:
1/4 >
+ oo as n — oo.
1.7.5. Find \\mx n if

n-*<x>
(c) xn =Y — n l
3
+n; (d) *„ = -^ cos n3 — ;
\ — 2n n+l n n(— 1)" .

I
(e)
— C0S —
1 2« 1 l—2n n 2 +l *
(f) xn = j j j- .
1+T+T+-+*
§ 1.8. Testing Sequences for Convergence
Bolzano-Weierstrass* theorem. A monotonic bounded sequence has
a finite limit.
Theorem on passing to the limit in inequalities. If xn ^y ^z n n
and liin x^lim zn = c, then \imy n c too (c = is a number,
n -y cc n -> cc n -* cc
-{ 00 or — 00 but not oo).

1.8.1. Prove that sequence with the

the general term xn =
= (2n— l)/(3n+ 1) an increasing one.
is
Solution. We have to prove that x n+1 >x n for any n, i.e. to

prove that
2/z-j-l 2n— l
3/Z + 4 > 3rc+ 1

'
The latter inequality is equivalent to the obvious inequality
e>n
2
+ 5n+\ > 6m + 5n — 4. 2
Hence, x n+l > xn .
1.8.2. Given a sequence with the general term
Xn ~ n\
'
Prove that this sequence decreases at n^lO.

Solution,
Xn+1 ~~
_ 10"+ 1
_ 10" 10 __ 10
(n+ 1)! n\
'
/i+ 1 n+ 1
'
Since < 1 at ft ^10, then x n+1 <x n beginning with this
number, which means that the sequence decreases at n ^ 10.
1.8.3. Test the following sequences for boundedness:
( a ) Xn= ^j'>
#„ = (-l) ~sinft;
n
(b)
(c) zn = n cos nn.

Solution, (a) The sequence {x n \ is bounded, since it is obvious that
(b) The sequence {y n } is bounded:
l^l = l(-D"|-^T |sin«|<^T <2.

(c) The sequence {z n \ is not bounded, since
\z n \
= \ncosnn\ = n.
Xn Xi Xo Xn i
(a > 1, x > 0) converges.

Solution. Let us prove that this sequence is monotonic and
bounded. Firstly, x n x n _ l as <
Hence, the given sequence is a decreasing one. Secondly, all its

terms are positive (by condition a and x 0), which means > >
that the sequence is bounded below. Thus, the given sequence is
monotonic and bounded, hence it has a limit.
5+1
1
1
52
11
+l ^5 3 +l 1
' • • 1
i
5»H-1
î.e. x1 — b+{ ;
x2
1,1
~ 5+1 + 5 +1 a ;
x2 — 5+1+ 1,1,1 5H-1 ^ 5a +l '
converges.
Solution. The sequence \x n ) increases, since x n+1 xn 1/(5"
4-1
= + + 1)
and, hence, #„ +1 >
x n Besides, it is bounded above, since 1/(5"+
. 1) <
< 1/5" at any n and
+ _L_ + _^
^3 +l +
+-J—
1
" 5+1 I
1
5 +l 5*22 i 1 I
1
3 l
1 I1 '
'
'
' '
' I
1
Kn
5»+l ^ _L_ 1
iii 1/5-1/5" + * _ W,
^ îli^
/ 1 1\ 1
^ 5 52 53 ' ' ' ^ 5" 1 — 1/5 4 V 5" J ^ 4

'
Hence, the sequence converges.
1.8.6. Taking advantage of the theorem on the existence of a

limit of a monotonic bounded sequence, prove that the following
sequences are convergent:
(a) xn = —r- ;
(b) * B = 2 + l + l+...+I.
1.8.7. Prove that the following sequences converge and find their
limits:
(a) xx = V~% *2 = ^2+l/"2;
n radicals
2n
(b) *„
(/i + 2)! '
(c) x
(d) the sequence of successive decimal approximations 1; 1.4;
1.41; 1.414; ... of the irrational number j/~2;

(e) xn = n\/n n .
Solution, (a) It obvious

is that xx x2 < <x < 3
. . . < xn <
<*„ + < 1
• i.e. sequence
the is increasing. It now remains to
prove that it is bounded.
We have = V 2 + x n _ n = 2 3,
xn Sinc e x = V~2<2 l9 y
. . .
1
xt = V2 + x 1 <V2 + 2 = 2 x, = V~2+r x <V2 + 2 = 2, .... Let it

9 2
be proved that x n _ <2. Then x n =

1 ]/ 1 2 + x n _ < V2 + 2 = 2. Thus,
with the aid of mathematical induction we have proved that x n 2, <
i.e. the sequence is bounded. Hence, it has a finite limit. Let us
find it. Denote
lim x n = y.
n -> oo
Then, x a = V~2 Jr x n-v raising to the second power, we obtain
Passing to the limit, we can rewrite this equality as follows
lim x\ = lim (2 + ,^), or y 2 = 2 + y.

n -* oo n -* cc
The roots of the obtained quadratic equation are:
ft = 2; y2 =— 1-
The negative root does not suit here, since x„>0. Hence, lim xn =y l
= 2.
n -* oo
(c) We have ny — < E (ny) ^ ny

1 or y—l^tL^^y. Bui the
sequences j#
— -^-j and {y} converge, their limit being y, that is
why lim x n =
This sequence is non-decreasing, since each following term

(d)
x n+1 obtained from the preceding one x n by adding one more
is
significant digit to the decimal fraction. The sequence is bounded

above, say, by the number 1.5. Hence, the sequence converges, its
limit being V 2.
(e) The sequence decreases monotonically. Indeed,
(fl-f-l)! _ n\ _n\ nn __ nn
Xn+1 ~~
(n+\) n Xn
' ~~~ '
~~nP (n+l)»
Since
(n+\)*
< l
>
X» + 1 <X »*
Then, since xn > 0, the sequence is bounded below, hence

n
lim x n
-+ cc
exists. Let us denote it /. Obviously, /= lim xn ^0. Now let us

n -+ cd
show that / = 0. Indeed,
nn \ n J \ n J n
Hence,
^n
< -y and xn+l <-^x n .
Passing over to the limit,
we obtain
which, together with /^0, brings us to the conclusion:
/ = 0.
1.8.8. Find the limits of the sequences with the following gene-
ral terms:
n n
zn = -
1 1 1
r
-
*-...+
V n*+\ ]fn*+2 Vn* + n
Solution. Let us prove that lim#„=l. Indeed,
n -* oo
= n— Vn + n 2
l*»-l| 1
}Tn? +n
V n 2 + n(n+ V n 2 + n) 2n'
We can prove similarly that

lim z n = 1.
-»• 0»
Then,
+ = 2M .
On the other hand,
l
=x ri
VrP + n V~n* +n Vri* +n }^n 2 +n

Thus,
x n<y n < z n, Urn *„ = lim z„ = 1
n -* cc n -* oo
and according to the theorem on passing to the limit in inequalities
lim yn =\.
1.8.9. Using the theorem on passing to the limit in inequalities

prove
lim JX'a=l (a>0).
1.8.10. Prove the existence of the limit of the sequence y n = a 1/2n

(a > 1) and calculate it.
1.8.11. Taking advantage of the theorem on the limit of a

monotonic sequence, prove the existence of a finite limit of the
sequence
Xn "1"
22 32 /I
2 *
1.8.12. Taking advantage of the theorem on passing to the limit

in inequalities, prove that
lim x n =\ if xn = 2n{Vn +l—n). 2
n -* qo
x1 = Va\ xa =----j/ a + V a\
r
xs = V<* + V a + Va; xn --= ]f a + Va+ . . .
+ Va
n radicals
(a>0)
has the limit b = {VAa+\ + l)/2.
_ 1 1 l_
*n
~3+l + 3 +2 + 2 ' ' ' ^~?> n
+n
has a finite limit.
1.8.15. Prove that a sequence of lengths of perimeters of regular

2 w -gons inscribed in a circle tends to a limit (called the length of
circumference).
§ 1.9. The Limit of a Function

A
point a on the real axis is called the limit point of a set X
ifany neighbourhood of the point a contains points belonging to X
which are different from a (a may be either a proper or an impro-
per point).
Let the point a be the limit point of the domain of definition
X of the function f(x). The number A is called the limit of the
function f (x) as x—»a, A= lim /(.*:), if for any neighbourhood V
x -* a
of the number A there exists a neighbourhood u of the number a
such that for all x£X
lying in a, f(x)£V (the definition of the
limit of a function after Cauchy). The number A may be either
finite or infinite. In particular, if the numbers A and a are finite
we obtain the following definition.
A number A is called the limit of a function f(x) as x—-+a y
A= lim f (x), if for any e > there exists a number 8(e) > such
x -+ a
that for all x satisfying the inequality and belon- 0<\x — a|<8
ging to the domain of definition of the function f (x) the inequality
\f(x) — A\<s holds true (the "e-8 definition").
If a=+oo, the definition is as follows. A number A is called
the limit of a function f (x) as x «>, A lim f (x), if for any —+ =
X -* + 00
e >there exists a number M(e) such that for all x satisfying >
the inequality M(e) and belonging to the domain of definition
of the function f (x) the inequality \f(x) A\ holds true (the — <e
"e-M definition").
The notation lim / (x) = 00 means that lim | f (x) \
= + 00. The
x -+ a x -* a
rest of the cases are considered similarly.
The definition of the limit of a function after Heine. The nota-
tion limf(x) = A means that for any sequence of values of x con-
x -+ a
verging to the number a
-^2> • • •
•^l> • • * •
(belonging to the domain of definition of the function and differing

from a) the corresponding sequence of values of y
0i = /(*i); 0i = /(*i); 0n = /(*n). •••
has a limit, which is the number A.

1.9.1. Taking advantage of the definition of the limit after
Heine (i.e. in terms of sequences) and of the theorems on the limits
of sequences, prove that
,. 3*+l 1
Solution. Let us consider any sequence x l9 x 2 ... satisfying ,
the following two conditions: (1) the numbers x lt x2y ... belong
to the domain of definition of the function f (x) =(3x+ l)/(5* 4) +
(i.e. x n =£ —
4/5); (2) the sequence {x n } converges to the number 2,
i.e. lim x n 2. =
n -+ 00
To the sequence {x n } there corresponds the sequence of values

of the function
3^+1 .
3;c a +l .
5*1 +4 ;
5* 2 +4 » ' ' '
proceeding from the theorem on the limits (§ 1.7),
0^l) a j±i a l
l
„T.'
im fi xW )= Hm
3*„+l_ lim
ii.5^ + 4- lim (5*„ + 4) 10 + 4- 2
•
§ 1.9. The Limit of a Fund ion 57
Thus, independently of the choice of a sequence {xn \ which

converges to the number 2(x n ^= 4/5), the corresponding sequences —
of values of the function f (x n ) converge to the number 1/2, which,
according to the definition of the limit of a function, means that
3x+l 1
5^+4= 2
•
Note. The definition convenientlyof the limit after Heine is

applied when we have to prove that a function
no limit. f (x) has
For this it is sufficient to show that there exist two sequences {xn }
and {xn } such that lim xn lim x'n = a, but the corresponding =
n -* qo n -* oo
sequences {f(x'n )\ and {f (x„)\ do not have identical limits.
1.9.2. Prove that the following limits do not exist:
(a) lim sin 737; 3 ^m (c) lim sin x.

Y 1 C )X -+ K —* CO
Solution, (a) Choose two sequences
xn = 1 H1
nn
1
and xn = 1
—j-t—
+ (4n+l)ji
Ta
2
(n
v
= 1 ,
»
2,
»
. .
. ),
/»
for which
lim x n = lim a:^ = 1.
n -+ <x> n -* oo
The corresponding sequences of values of the function are:
Kx n = s\n + l/{ nn) _ =smnn =

l
)
l l
and
f «) - sin + g/[(4w ^
{ t) n] _ = sin f£+i = sin
{
ji
(
2/m +
f) = 1
Hence,
lim / (x n ) = and lim f (x'n )
= 1
1. e. the sequences {/(*„)} and {/«)} have different limits, whence

it follows that
x
lim sin
-* 1
— 1
does not exist.
(c) Choose two sequences, x n nn and x'n = = 2nn + n/2 (n = 1
2, ...), for which lim jc n = lim a^ = oo. Since

n 00 n -* cc
lim sinx„= lim sin jt/z = 0,

and
n
lim
-* cp
sin<= lim sin (2jw
n -* 00
+ ji/2) = 1,
lim sin * does not exist.

Note. The above examples show that one cannot draw the con-
clusion about the existence of the limit of a function proceeding
from the sequence of values of x of a particular form (for example,
proceeding from x n 1 +2/((4n+ 1) n) =
in the item (a) of this prob-
lem), but it is necessary to consider an arbitrary sequence x 19
x 2) ...,#„, ... having a given limit.
1.9.3. Proceeding from the definition of the limit of a function

u
after Cauchy (i.e. in the terms of e-5"; "e-M", etc.), prove that
(a) Km (3x— 8)- —5;
(e) lim arc tan x = ji/2;

(f) lim sin* =1/2.
x -* Jt/6
Solution, (a) According to the "e-8" definition we are to prove

that for any e > there exists 8 > such that from the inequality
I*— 1|<8 it follows that f (x) ( 5) f (x) 5 |
e. — — |
= |
+ |
<
In other words, it is necessary to solve the inequality
|3a: — 8 + 5| = 3|a: — < 1 |

e.
The latter inequality shows that the required inequality /(jt)-f5 |<e
is fulfilled as soon as \x —
1 < e/3 = 6. Hence, lim (3x 8) = 5.
1
|
— —
According to the "e-M" definition of
(b) the limit one has to
show that for any e > it is possible to find a number M>
such that for all x > the inequality M
will be fulfilled.
Transforming this inequality, we obtain
5* -j-l 5 14
3x +9 3 |3* + 9| <8.
Since x > 0, it remains to solve the inequality
whence
^ 14—9e
14 7" 98
hence M= 3e
.
Thus, for e > we have found M = —^—g 14 e

such that for all
values of x >M the inequality (*) is fulfilled, and this means that
Let, for example, 8 = 0.01; then M=^ ^ ^- = 463= 1

-|-.
(c) We have to prove that for any K> there exists 8 >
such that from the inequality
|x-l|<8
there always follows the inequality
1 1
I
(I-*) 2 (i-*) 2 >K.

Let us choose an arbitrary number K> and solve the inequality
1
K, (**)
(l-*) 2
whence
|l_x|<-L: (K>0).
Thus, if we put 8 = -J=-, then the inequality (**) holds true as

VK
soon as \x — < 1
|
6, which means that lirn
n __ .
2
= +oo.
(d) We have to prove that for any there exists K> M>
such that from the inequality x there always follows the ine- >M
quality \og a x>K.Let us choose an arbitrary number and K>
consider the inequality \og a K. If we put a K x>
M, then at =
x>M the inequality \oga holds true. Hence, x>K
lim log a *= +oo.
X -»» + 0O
1.9.4. Prove that lim cosjc does not exist.

X -+ 00
1.9.5. Using the sequences of the roots of the equations sin (\/x)=l
and sin (l/x)= —
1, show that the function / (x) = sin (l/x) has no
limit as x >0.—
1.9.6. Proceeding from Cauchy's definition of the limit of a

function prove that:
— 2) = x ~ {
(a) lim (3a: 1; (b) lim =2;

K - 1 X - 1 V X— 1
(c) lim sin.x: = 0; (d) lim cosx= 1;

x -* o x o
,
v ,. 2x — 1 2
(e)
.'^.s+s'tJ
(f) lim a x = +oo (a > 1);
*-* + 00
(g) lim —= 0.
§ 1.10. Calculation of Limits of Functions
I. If the limits \\mu{x) and limy (a;) exist, then the following
x -» a x-+a
theorems hold true:
(1) lim [u (x) ± v (x)} = lim u (x) ± lim v (x);

x -+ a x -* a x -+ a
(2) lim [u (x)-v(x)] = lim u (x) • lim v (x)\
^
x^ a a a
x -* x -*
lim U (x)
(3) = W^) (Hm»(x)^0).
II. For all main elementary functions at any point of their do-
main of definition the equality lim / (x) = / (lim x) = f (a) holds true.
x -+ a x-+a
III. If for all values of a; in a certain neighbourhood of a point a
(except for, perhaps, x a) the functions f(x) and q) (x) are equal =
and one of them has a limit as x approaches a, then the other one
has the same limit.
IV. The following limits are frequently used:
(1) hm
x-+0
— x
=1;
(2) lim(l
x -* <x>
+ l/jc)* =alim(l +a) '* = e = 2. 71828.
-*
l
.
.;
(3) lim
lo ^ il+x) = log a e (a>0;a^l);
x^Q x
/+ =
... ,. In (1 a:) «
(4) hm x
'
1;
xÔ
(5) lim ^=^ = lna (a > 0).

1.10.1. Find the limits:
i \ i; +
W I™m 3*.+-f*r+T
ixb 9x 7 /u\ r1 xs + 3x* — 9x—2
'
(b
>
I
™ *3_ x _ 6 ••
m P and q
(C)
^VeZU^' (d)
il
,^_J
integers);
(e) Hm ^+^- 3
;
(f)lim^^;
|/*Fp7 —3 J/" 2* —3 .
.
r f\ x —3
(g)
*-+22 r/ — 2o 3/0=
3/-Xft
j/Ff6-2j/3^5
x+6 r/ 3x— 5
5 '
(h)
*-3 L
*-*3|_
l0g «
VHT+6-3J'
> — x — a:+
2 1
r r
v + K8a:+ 1
Solution, (a) Since there exist limits of the numerator and deno-
minator and the limit of the denominator is different from zero,
we can use the theorem on the limit of a quotient:
+ 9* + 7 ^ î
lim V(4* 5 t 9x+7)
+ t ;
+9+7
lim 4*» J
4 ,
^-13^ + ^+1 lim (3jc

6
+x + 3
1) 3+1+1
(b) The above theorem cannot be directly used here, since the
limit of the denominator equals zero as x—-»2. Here the limit of
the numerator also equals zero as x —»2. Hence, we have the
indeterminate form -jj- . For ^^2 we have
A;3 + 3x2_9 ._2 _ (x—2) (x + 5x + = + 5a:+

A
2
1) jc
2 1
x — x— 6
3 ~~ (x—2) (x + 2x + 3) --jc + 2*+3" 2 a
Thus, in any domain which does not contain the point x =2 the
functions
r, x £3 + 3x — 9x— 2
2 , , . x2 + 5x+l
are equal; hence, their limits are also equal. The limit of the
function y(x) is found directly:
hm /vi- * + 5x+ = 15
cp(^)^ im
2
1
x -.2 ^2- + 2*+: n 2 ;
hence,
£/ x
x3 + 3x — 9x— 2 2 15
X-+2 X-+2 X X O 11
(c) Just as in (b), we remove the indeterminate form by

^
transforming
lim = lim (*+0(^^3-3,) =
<->-! 3 (1— *) ~ i>
1.10.2. Find the limits
— 4 — 3x + 2
X3 X2
(a) lim 2
( 3a:
(b)
*
lim
-* + 00
(j/9je 2 + — 3x); l
(c) lim
2^+3^/7+5^/7
v^ + oo V 3*— 2+ 3/2x—
(d) lim 0/2x 2 — 3— 5x);
* -* — 00
(e) lim x(Vx +l 2

— x)\
X -* + QO
/f , ,. V2x^f3 , ]/~2x 2 +3
(g) lim S^** 3 *.
X -* OO
So/««on. (a) lim

(^-^) .
Here we have the indeterminate form oo — oo; let us subtract the

fractions
lim
jcôo
f^_^)=lim
—
\3* 2+ 4 3x 2 y ^ 9x
3
+
2x3+4x2
6a: 2 — \2x— 8
-
2 + 4/x
= lim 9 —
2
'
^.Too + 6/a:— 12/a: 2 8/a:

3
9
Note. We see that in such examples the limit is equal to the
ratio of the coefficients at the superior power of x (provided the

polynomials are of the same degree).
/ux ,.
—
(}^9x 2
+i — 3x) = ,. 1
n
= 0.
(b) lim -J-
'
lim r
+ 1
x -+ + cc J^-j- 1+3*
(c) In handling such examples bear in mind that the function
f{x)= \/p n where p n (x) is a polynomial ofjdegree n, tending
(x),
n
to infinity in the same way as the function "j/ This allows us .
to single out the superior power of x and divide both the nume-
rator and denominator by this power of x. In the given example
§ 1. 10. Calculation of Limits of Functions 63
the divisor is [/ x\ then we obtain:
2J/T+3 j/T _ 2 + 3/*/T+5/ff^

+ j/3x— 2 + l/2x—3 x^ + oo |/"3 — 2/x+ £/4/x— 12/x + 9/x 2 3
(d) Since the sum of two positive infinitely large quantities is
also an infinitely large quantity, then
Urn
- cc
(1/2j?^3— 5*)= lim [V2x* — 3 + (—5*)] -
X -> X-*-oo
(f) At a; > we have |/#2 = .x;, therefore
. ^*«(2 + 3/*«) _ *V~2 + 37F _

"","™
+ 2/*) ~
,
'
+ ao *(4 4
At a: < we have \^x2 = —x and, hence,
, im
y* a (2 + 3/**) = Hm ^2 + 3/^ _ ]/"2~
*(4 + 2/x) *(4 + 2/*)

*x 3
/Vote. From this it follows, incidentally, that lim ^ \ ~t does
not exist.
lim 2x/(* + 3)
(g) lim5 2 ^* +3 ) = 5*- fl0

= 5 = 25.
2
* -*• oo
(a) Hm^T&l ; (b) H ra _*±I ;
(c) hm ,^5 X - <

; (d) lim-*-
x-+0
— (k positive in-
teger);
,
x ,. sin (x— ji/6)
^3 — 2cosx' U *-jt/2 sin*) 2 '
, v
y x—
2 sin 2 x-f sin 1
*i™/6 2sin
2
x— 3sinx+l '
(method of substitution), (a) Let

Solution us put 26 +x= z 3
.
Then x=^z*— 26 and 2—^3 as
lim 3/ ^Z_
* 1; hence
2
= lim ^ - lim ^-3)(^+3*+9) _
= lim2(z + 3z + 9) = 54.
2
2^3
(d) Let us put l+x=z k ; then x=z k — 1 and 2—^1 as x —+ 0.

Hence,
lim
^T^~ l
-i; m _£z± *
(S ee Problem 1.10.1 (d)).
X-y 2 -* 1
(e) Let us put x ji/6--=2; then — x= z + n/S and z — () as

x — On substituting we obtain
lim
t-jT/6 l/"3
yÛ
— lim—— 2 cos a: 2-0 /"3
sin2
2 cos (z -f ji/6)
sin z 2 sin (z/2) cos (2/2)
lim
—
= lim
>-o J^~3 "K" 3 cos 2+ sin 2 2 - o 2
/" 3 sin 2 (z/2) + 2 sin (z/2) cos (z/2)
CQS(2/2)
= lim _ =1.
z - o yA 3 sin (2/2) + cos (z/2)
,
(a)
x ,.
hm
xÔ
—— —
1 cos x
-5
x
;
/L v
(b) lim
x-0
i
• tan x
X3
— sin a:
cos^2)
(c) lim
a:- 1
f
Solution, (a) lim
/vi-
*-»o
1 — cos
x5
—= a:
lim
*-»o
2 sin 2 (x/2)
x r^'-
l Hm
2 J™
/
^
sin(^/2)
x/2
y
J
==1
2
,
'
/ux ,. tan a: —
- sinA; = lim sin a: (1 —cos a:)
—
(b) lim '
*-* C0S x x * '
= limcos x 1 sin* 1 — cos a; 1
x2
Let
(c) us put 1 x= z. Then x=l — z and z —O as x — 1.
Hence,
/ Jl TC \
lim
1— X
—
COS -y X
- — = lim
2-0
= lim
2-
sin -7- z
2 3T
Note. For a simpler method of solving similar problems see § 1.12.

(a) lim(l + l/*) 7
*; (b) lim(l +x) l
^\
x-»
(c
^ .fe)* ra
In (\+x)
;
(d) lim(l+£/A:) m *;
e4 *— 1 .
(e) lim t
(f) lim
x-+ o
3*— 1
1
*-* tan a:
*
In (a+x) — In a
(h) lim
t
(g)
x ->
... In x — 1
(0 llm -r=r •
Solution, (a) lim H— y x = Hni 1 +

1 \*
lim 1 + = e'\
In +x)
(e) lim
u" 1
= lim (1
3* — 1
1
ln3'
x -* x -v
(i) Put jt/e— l=z; then x--=e(z+l); z—*0 as x—*e. On substi-

tuting we obtain
Hm lni = J = lim ln(x/ e)

— ,,-ilin, e
1^1+^1 z e
_1)
z
1.10.6. Find
•im (1+^
x
Solution, lim (l+l) =lim |^1 +1^*] = 1.
(a) lim/l-MV 1 -^ 1 -^.
x 2Jr 2 X — 1 \(2* + 1 >/( at— 1)

(b) lim
2x 2 — 3a: —V
Solution, (a) Denote:
fW = (l+x)/(2 + x);
/ \ 1— ^*
Hm^(x) = Hml±^ = -|;

^~ x
lim
x-*\
(p (x) = lim
x->\
^~
l x
= ir 1
1
.
But at finite limits lim / (x) = A > 0, lim q) (.*:) =5 the following
relation holds true:

lim (p (x) In / (x)
lim [/
{x) = ex ~+ a
=e B\nA == j[B t
Hence,
°- r ""'"-"-(!)" !
lim =/!
{x)
/Vo/e. If in handling examples of the form lim [/
(x)]v it turns
x-+a
out that Vunf(x)=l and lim cp (jc) = oo, then the following
3— 31 43
transformation may be recommended:
lim [f(x)]*<*> = lim {1 + [/ (*)— 1 }}*

{x) -
x^-a x -a
l«»n(p(x)L/U)-l]
= lim {
[ 1 + (/ (x) — 1 )]
<*> 1
>}* <*> U <*>
- n = e**a . (*)
(a)
(c)
lim(
lim (1
^r
2
+sin :rix)
+3
cot
;
™;
(b) li.n(4±^)
,/,m
*;
/U Q)
(d) lim (a¥=kn> with & an integer).
(f|^~y
Solution, (a) Let us denote:
^^StT' <P(*) = 8* + 3; a
lim/(x) = lim§^=l;
lim q) (jc) = lim (8a:
2
-f 3) = oo.
Use the formula (*):
8a: 2 +3 ,im V (*> ^ (•*>-!]
lim (ttt—)
/W— I=|S±4 — 1
:
[/(*)-!] = - lim = - 8.
2
gf +
3)
lira q> (x)
Therefore
1;
/2xH-3\8x« + 3
1.10.9. The function /(#) is given with the aid of the limit
/? -> cc
x r 1
Investigate this function and graph it.

Solution. Consider three cases:
2n --=
(1) \x\> 1. Since in this case lim x 00, then
(2) |jc|<1. In this case lim Jt

a,,
= 0; therefore /(*) = — !.
(3) jt-=±l. In this case x 2n =\ at any n and therefore /t

(jc) = 0.
Thus, the function under consideration can be written in the
following way:
( 1 if
M>
\x\<l
1
f(x)=\-l if
{ if * = ±1
or, briefly, / (x) = sign (| x |
—
1) (see Problem 1.5.11 (n)).
The graph of this function is shown in Fig. 27.
1.10.10. The population of a cou-

ntry increases by 2% per year.
By how many times does it increase
in a century?
Solution. If we denote the ini-
tial number
inhabitants of a of
given country as A, then after a
year the total population will amo-
unt to
Fig. 27
After two years the population will amount to A[ 1 + p-q) • After

100
1 \
100 years it will reach the total of A (
I+j-q) » i- e - it will
\ 50
have increased times. Taking into account that
lim —l
=e> we can approximately consider that

1
e.
J
Hence, after 100 years the population of the country will have
increased e'1 7.39 times. «
Of course, this estimation is very approximate, but it gives an
idea as to the order of the increase in the population; ^the quan-

/ 1
\ioo \
tity ^1 =7.245 to within three decimal places).
50/
, x i- cos x-\~A tan x 1
lim-
(a)
2 — x — 2x*
(b)
z /__ x _2 i
(c) lirnÊ^;
2x- — 5x + 4 .
(d) lim
5a-
2 — 2x — 3 *
(e) lini(j/>+ 1— — 0;
X-> QO
a
(c) lim .
1
s
~~
'".,
a "' n , ;
(d) lim tan 2x tan (n/4 — x);
gc-»ji *-*ji/4
— 3 tan x
(e)
.
™
..
3
tan 3 x
cos (x+n/6)
*

*"* -1
(a) lim (1 + 4/x)* +3 ;
(b) lim ;
(c) lim^=^;
*
(d) lim (1
*->0
+ 3 tan 2
x) cot2 *;
x->0
22
(e) lim (sin2*)t- *; (f) lim lg-r\) ;
x-yn/4 x-y cd \
zx -r 1
/
(g) lim (tan ;c)

tan 2
*; (h) lim (sin x) ian x \
X-+JI/2 X-yJt/2
(k) hm
x +0
— x
.

,- arc cos (1—^) In tan a:
hm
. ,
(a) i nn ; (b) ;
(c) lim —— In (1 + a sin a:).

sin x
o
§ I. II. Infinitesimal and Infinite Functions.
Their Definition and Comparison
The function a(x) is called infinitesimal as or as .v-^ oo

if lim a (x) = or lima (*) = 0.
x-*a X-y CD
The function f (x) is called infinite as ^->a or as x-oo if
lim/(A;) = oo or lim / (a;) = oo.

§ 1.11. Infinitesimal and Infinite Functions 69
A quantity inverse to an infinite quantity is called an infinite-

simal.
Infinitesimal functions possess the following properties:
(1) The sum and the product of any definite number of infinite-
simal functions as x a are also infinitesimals as x -> a.
The product of an infinitesimal function by a bounded function
(2)
is an infinitesimal.
Comparison of Infinitesimals. Let the functions a(x) and P (x) be
infinitesimal as x a. If
a (x) , .
lim q-K- =
*-*a P to
where c is a certain finite number different from zero, then the

functions a (x) and P (x) are called infinitesimals of the same order.
If c=l, then the functions a(x) and P (x) are called equivalent;
notation: a(x)~$(x).
If £ = 0, then the function a(x) is called an infinitesimal of a
/ug/ier order relative to P (*), which is written thus: a (x) o (P (x)), =
and P(jc) is called an infinitesimal of a /oayer order with respect
to a (x).
a
If lim R /f where 0<|c|< + oo, then the function a (x)
is called an infinitesimal of the nth order as compared with the

function P(x). The concept of infinite functions of various orders
is introduced similarly.

X 4
(
a ) f( x ) = 2Trrr
5
as 2 >
f(x) = (x — l) sin
2 3
(b) as a: 1 are infinitesimals.
Solution, (a) It is sufficient to find the limit

2x ~\ = 0.
lim / (x) = lim
* 2 « -> 2 r o
(b) Firstly, the function cp(x) = (x — l) 2

is infinitesimal as x 1;
indeed, lim (a: — l)

2
— 0. Secondly, the function
X >
—
1
\J)(a:) - sin 3
j ; 1,
is bounded:
I
sin 3 — < 1.
Hence, the given function f (x) represents the product of the

bounded function by the infinitesimal
\p(x) (p(x), which means that
f(x) is an infinitesimal function as x -+ 1.
70 Ch. 1. Introduction to Mathematical Analysis
(a) /(*)= as x -+ 4
^-yi >
(b) f (x) = as x — oo
are infinitesimal.
1.11.3. Find
lim x sin (l/x).
x - o
Solution. Since x is an infinitesimal as x and the function —

sin (l/x) is bounded, the product xs\n(\/x) is an infinitesimal, which
means that lim x sin (l/x) = 0.
* -* o
1.11.4. Compare the following infinitesimal functions (asx— ())
with the infinitesimal (p(x) = A;:
(a) /,
= tan x (b)
(x) 3
; / 2 (x)
= ;
(c) U*) = K9 + x-3.

Solution, (a) We have
x ->
im ifE^ =
* x
li
-
m
tan a:
3
x,1 =
2
l
J X
,.
lim
li
-
——
tan*
^
3
liin a:
2
= 0.
Hence, tanx 3 is an infinitesimal of a higher order relative to x.
(b) We have
* - * x -* X2 3/^
J/
Hence, jj/sin 2 * is an infinitesimal of a lower order as compared

with A'.
(c) We have
lim ——— = li in r = ir.

x-> o
x x-+o + 6
Hence, the infinitesimals V~9-{-x —3 and x are of the same order.
1.11.5. Determine the order of smallness of the quantity p with

respect to the infinitesimal a.
(a) p=rcosa — cos2a; = tana — sina. (b) P
Solution, (a) p^=cosa — cos 2a = 2 sin -| a sin
Whence
2sin(3a/2)sin a/2)
|in] J^2 =lim
« «2
(
= 3_
2
a o a - o
Hence, P is an infinitesimal of the same order as a2 , i. e. of the

second one with respect to a.
§ 1.12. Equivalent Infinitesimals 71
1.11.6. Assuming * — oo, compare the following infinitely large

quantities:
(a) / (*)
= 3* + 2* + 5 2
and cp (x) = 2* 3
+ 2x — 1
f(x) = 2x + 3x (*) = (* + 2)
2 2
(b) and cp ;
(c) / (x) = Y x + a an d <p(x)=l/x.

The infinite function 3* 2 + 2* + 5
Solution, (a) is of a lower order
as compared with the infinite function 2* 3 + 2* — 1, since
*-> oo
2* 3 + 2x— 1 2 + 2/x — 1/x 2 3
1.11.7. Prove that the infinitesimalsa=* and P=*cos(l/*)(as*—^0)

are not comparable, i. e. their ratio has no limit.
X cos 1 x\
Solution. Indeed, lim -L-L=
(
y
1
lim cos (l/x) does not exist (prove

x -> o * x -> o
it!), which means that these infinitesimal functions are not com-
parable.
1.11.8. If * — 0, then which of the following infinitesimals is

(are) higher
of a order than x\ of a lower order than x; of the
same order as *?
(a) 100*; (b) *2 ;
(c) 6 sin*; (d) sin 3 *; (e) j/tan 3 *.
1.11.9. Let x —
O. Determine the orders of the following infini-
tesimal functions with respect to x:
(a) 2 sin 4 a;—* 5 ;

(b) [/sin 2 * +* 4
;
(c) Vl +x — 3
1; (d) sin 2* — 2 sin*;
(e) 1 — 2cos(x+£); (f) 2 |/sin*;
(g) TZTf; (
h ) tan* +* 2
;
(i) cos* l/cosx; (j) — cos*.

1.11.10. Assuming the side of a cube to be an infinitesimal, de-
termine the order of smallness of the diagonal of the cube (d), of
the area of its surface (S); of its volume (V).
§ 1.12. Equivalent Infinitesimals.

Application to Finding Limits
If the functions a (*) and P (*) are infinitesimal as *— and if
a (x) ~y (
x )> P (*) ~ 6 (*), then
lim -77t4-= Hm -^y- (replacing an infinitesimal by an equivalent one).

If
lim f(x) =k f < |*| < oo,

x -* a
then
/ (x) a (x) ~ ka (x).
If
a.(x)~y(x) 9
P(*)~Y(*),
then
a (x) ~ P (x).
For two infinitesimal functions to be equivalent it is necessary
and sufficient that their difference be an infinitesimal of a higher
order as compared with each of the two.
Listed below are infinitesimal functions:
(a(x) is an infinitesimal as x—+0)
(1) sin a (x) ~ a (x); (2) tan a (x) ~ a (x);
(3) 1— cosa(x) ~ [a(Jc)] 2 /2;
(4) a (x) ~ a (x)\

arc sin (5) arc tan a (x) ~ a (x);
(6) ln[l+a(x)] ~a(x); (7) a
a{x}
— ~ a (x) \na
1
(a > 0), in particular, e* U) — ~a 1 (x);
(8) [l+a(x)] p — - Pa(x), 1 in particular, (V l+aW-l- a^.

1.12.1. Prove that as x-^0
a x; (b) 1_
<
>
l
~YTTi~'* TT7~* ;
(c) sin
Solution, (a) By formula (8) at P=l/2 we have
1 1
X
o/T+^-Dî. X.
(c) By formula (1) we have
sin]/ xV~x '~y x)/"x ' ' =x 3

<\
y 2
+ Vx =X 3 3
'*V 1 +X 1/2 ~X 3
'\
whence sin y xV x ~y x*+yx*-

1.12.2. Replace each of the following infinitesimals with an equi-
valent one:
(a) 3sina — 5a 3
;
(b) (1 — cosa) + 2
16a 8 + 5a 4
-f 6a 5 .
§ 1.12. Equivalent Infinitesimals 73
Solution, (a) Note that the sum of two infinitesimals a and |3
of different orders equivalent to the summand of the lower order,

is
since the replacement of an infinitesimal with one equivalent to it

is tantamount to the rejection of an infinitesimal of a higher order.
In our example the quantity 3 sin a has the order of smallness 1,

—
(— 5a 3 ) the order of smallness 3, hence
3 sin a+ ( — 5a ~ 3 sin a ~ 3a.
3
)
(b) (1— cosa) 2

+ 16a 3 + 5a + 6a = 4siii
4 5 4
-| + 16a 3 + 5a + 6a4 5
.
The summand 16a 3 is of the lower order, therefore
(1— cosa) 2 + 16a 3 + 5a + 6a ~ 4 5

16a 3 .
1.12.3. With the aid of the principle of substitution of equiva-

lent quantities find the limits:
/ v I-
llm
sin 5* /i\ i- 1 — cos a;
a) (b) l"n
(
.0 mna-.,^
1" 0+4*)' * '
_ ,_ cos |
•
/
(c)
\ r
hm 4
In cos
—
a:
;
,,v
(d) hm
1; V^" 1 -f-
\
a: -f- a:
^
2 — 1
,
, x ,. sin 2*4- arc sin 2 *— arc tan 2 *

(e) hm ±- 3- ;
.pv |. 3 sin* * * — + 2 3
]" n tan* 2sin 2 * 5* 4

+ + ;
sm x ~ tan ^ ~^ 2
1
~~ cos 2 x )*-r xb
(g) Hm ^
7 tan 7 x-\- sin 6 x-j-2 sin 5 x

.
x ^ o
(h) lim — sin j/

X
* In (1 +3*)
3
, ,
^°(arctan |/* ) (*
5
/ - *
1
j.
1 — cos *+2sin x— sin x— * + 3* 3 2 4
'
'
x ^ tan * — 6 sin- x
3
* — 5* -}-
3
Solution, (a) We have sin5A:^5x; In (^1 4x) — 4x (see the + list
of equivalent infinitesimals on page 72). Therefore

sin 5* 5* 5
= ,.
J™lF(T+W , i o-4F = T- Il ,1
*2»JL- = lim ln " +

S*-' =
)1
(c) lim
= 4 hm „ cos* — = — 4 hm
I A ,.
—x — =
2
/2
—2.
7i
a: -> * x -> o
x
(d) From the list of equivalent infinitesimals we find:
V 1 +x + x — - 2
1 (x -\-
2
)/2 ~ x/2, sin \x ~ \x.
74 Ch. I. introduction to Mathematical Analysis
Therefore
vT+7+x — —
—=—
5
lllTl r—-j
l
r
lUTl —x/2 I
(e) Using the list of equivalent infinitesimal functions given on

page 72 we obtain
sin 2x + arc sin 2
x — arc tan x ~ sin 2x ~ 2x.
2
Hence,
,. sin 2x-\- arc sin 2 x — arc tan 2
x
= r 2x
= t2
hm Yx
Ilrn
35
•
<h)sin>/7~ Y~x\ In (1 + 3x) - 3x;
arc tan K7; e

5
^* — — 5 1 J^jc;
lim
sin^ln(l+3,_)
m ^.3,_3
'
(arc tan V~xY

1.12.4. Find the approximate values of the roots (/ 1.02 and
1^0.994. Estimate the absolute error.
Solution. Use the approximate formula
]/"l+x ~l+x/2 (*)
(for x sufficiently close to zero). In our case
+ 0.02- 1+^= 1.01;

J/1
V 1—0.006 ~ — = 0.997. 1
To estimate the error we note that
|_(Kr+7- 1) =! (*- 2 ^1+^+ 2) =

= j(*+i- 2 K^+i) = l(^^
Hence, the absolute error of the approximate formula (*) is esti-
X2
mated by the quantity y.
Using this estimate we find that the absolute error of the root
1^1762« 1.01 is = 0.00005, and the absolute error of
j/0994 ^ 0.997 to « -^^ « 0.000005.

(
amounts
o
1.12.5. Prove that, as x —- 0,
(a) f/T+x—l~jx;
§ J. 13. One-Sided Limits 75
(b) arc tan mx ~ mx\

(c) 1 — cos 3
.x;~ ~ sin 2
x.
1.12.6. For x-^0 determine the order of smallness, relative

to the infinitesimal $(x) =
x, of the following infinitesimals:
x '^ { x
(a) Vs\n 2 x + x*; (b) ^ 2* .
1 + /
1.12.7. For x—+2 determine the order of smallness, relative to
the infinitesimal $(x)=x — 2, of the following infinitesimals:
(a) 3(x — 2) + 2(jk —

2 2
4); (b) j/sin~^.
1.12.8. Making use of the method of replacing an infinitesimal

with an equivalent one, find the following limits:
, , ,. sin3x /U v
,. In (1 + sin 4a:)
e 'i"3*_l
< C
>
^ In (1 + tan 2x)
— cos 2x)
'
<
d>
J
1
arc tan 3x
™ aTclI^r =
/vi- In (2 /fx V\-\- sin 3x—

n
l» ln*(8in3x+l) ;
(
f
)
J™ 0+ tan 2.)
in
•'
In (1+2.-3^ + 4*3) .
/T+^-l
g
"i o
In (1— — 7jc») '
y }
"J" 1-cosx
*
3,
1.12.9. Find an approximate value of the root j/ 1042.
§ 1.13. One-Sided Limits

A number A is called the limit to the right of the function f (x)
as x —>x ()
(A= lim / (#) / (x 0)) if =
for any +
there exists 8>0
x -* * +
6 (e) > such that for all x satisfying the inequality < x x Q < 6 (e) —
and belonging to the domain of definition of the function f (x) the
inequality \f(x) 4 — <e holds The limit to the left of the
true.
— —
|
function / (* 0) as x— > x defined in a similar way. If *

is 0, =
then we write simply x—* + or -— and, respectively, /( 0) +
and f(— 0).
1.13.1. Find the one-sided limits of the functions:
|
— 2x + 3 if 1,
(a) /(*)={ asx->l;
3x — 5
q r .
>
.
{ if x I
(
b) / (*) = 77=777 as
r{ s2x
(c) f(x)= -r as x 0;
(d) f(*) = 3+ '
as x —
+ 7l /(] _„
1;
i
(e) f (x) =cos(n/x) as x — 0;
(f) f(x) =5/(x-2)* asx-+2.

Solution, (a) Letx<l. Then / (x) = — 2x + 3. Hence, f(l—0) =
= lim f(x) = \ is the limit to the left.
*- -0 1
If #>1, then f (x) = 3x — 5; hence,

f(l+0)= lim f(x) =—2 is the limit
x - 1+0
to the right (see Fig. 28).
1^1—-cos 2* Y'l sin 2 #

(c) /(*) =
V2 |
sin x|
but
= j
sin a:, if < < n/2 a:
f
|
sin x
-ji/2 < x < 0.
|
) —sin a:, if
Fig. 28 Hence,
0)= Hm /(*)= Urn — |/2

f( + 0)= Hm / (*) = lim (1/2-
AT- +0 *-> + V
(d) The expression 1/(1— a:) tends to +oo, when x tends to 1,
remaining less than 1, therefore
lim 7'/<i-*> =+ oo, lim

1
_ = 0, f (1-0)^3.
* -> 1 - X 1 - 1 + '
y)
Further, as x —^1+0 we have 1/(1 a:)— — oo. Therefore

lim 7i/< 1 -*>=0,
x-*• +01
" ,+0 = »
i
! t1.(
3+
H^')" 3+, = 4 -
(e) Let us choose two sequences, {xn \ and {a:;}, with the general
terms
and X (n=l, 2, ...)

Tn "-2^TT
respectively.
§ J. 14. Continuity of a Function. Points of Discontinuity 77
Then lim x n = lim x'n = and
lim / (#„) = lim cos 2nn = 1;
limf (*;) = limcos(2n+ l)f = 0.
Hence, the function /(*) has no limit to the right at the point 0;
taking into account that f (x) is an even function, we conclude that
it has no limit to the left either (see Fig. 29).
^^^^
1-/I
ii
Fig. 29
1.13.2. Prove that, as x —^1, the function
'
W" j
3x +2 at 1 <x<3
has a limit to the left equal to 2 and a limit to the right equal to 5.
1.13.3. Find the one-sided limits of the following functions as

X — 0:
(b) f{x) = ei>*;

sin x
^(^) = i- r
I
i \ £ / \
-L. 1
(c)
£ Continuity of a Function.
Points of Discontinuity and Their Classification
Let the function y =
f(x) be defined on the set X and let the
point x £X
be the limit point of this set. The function f (x) is said
to be continuous at the point x if lim f (x) f (x ). The latter con-
=
dition is equivalent to the condition lim
Ax - o
&y(x )=^ lim [/(x
a* -* o
+ Ax) —
-/(*.)] = <>.
The function f (x) is continuous at the point x ()

if and only if
f(x o -0) f{x o =

0) = f(x o ).
+
The function f (x) is continuous on the set X if it is continuous
at every point of this set.
Points of Discontinuity of the First Kind. Let the point x be the (]
limit point of the domain of definition X of the function f(x).

The point x is called a discontinuity of the first kind of the fun-
ction f (x) if there exist the limits to the right and to the left and
they are finite. If f (x u 0) = / (x -\-0) f (x ),
—
then x u is called
a removable discontinuity. Further, if f (x 0) f (# 0), then x u
{)
— +
is a non-removable discontinuity of the first kind, and the d-ifference
f(x 0)
{) + —
/(* —
0) is called a jump discontinuity of the function
f (x) at the point x, r
Points of Discontinuity of the Second Kind. If at least one of

the limits of f (x 0) and f(x —
0) is non-existent and infinite, +
then point x is called a discontinuity of the second kind of the fun-
ction f(x).
1.14.1. Using only the definition prove discontinuity of the fun-

ction f(x) - 3jc 4 + 5r + 2jc 2 + 3jc + 4 at any x.
J
Solution. Let x be an arbitrary point on the number scale. First

find lim f (x):
Urn f (x) = Urn (3x 4 + 5x + 2x + 3x + 4) = 3jcJ + 5*„ + 24 +

3 2
3.v + 4.
Then compute the value of the function at the point a' :
/ (*«)
=- 3a 4
+ 5x» + 2x1 + 3x + 4. {)
Comparing the results thus obtained, we see that
lin f(x )^=f(x () {

).
Hence, the function f (x) is continuous at the point x by definition.

Since ,v is an arbitrary point on the number scale, we have proved
continuity of the function for all values of x.
1.14.2. Given the functions:
for — oo < x ^ 1,
(a) f(x) for 1 < x < 3,

for 3 ^ x < oo;
for x<3,
(b) fix)
for x > 3;
(c) / (*) = 2x— 3

§ 1.14. Continuity of a Function. Points of Discontinuity 79
Find the points of discontinuity (if any). Determine the jump

discontinuities of the functions at the points of discontinuity of the
first kind.
Solution, (a) The domain of definition of the function is the entire

—
number scale ( oo, oo). In the open intervals ( 00, 1), (1, 3), —
(3, 00) the function is continuous. Therefore discontinuities are pos-
sible only at the points x=l, x 3, at which analytic representa- =
tion of the function is changed.
Let us find the one-sided limits of the function at the point 1:
/(1_0)= lim l(2* a + 3)--= 1;

x-+ -0 b1
/(l+0)= lim (6 — 5*)=1.

x +0 1
The value of the function at the point x= 1 is determined by the

first analytic representation, i. e. /(l) (2 = + 3)/5 = 1. Since
/<l-0) = /(l+0)=/(l),
the function is continuous at the point x=l.
Consider the point x 3\ —
/(3 — 0)= lim (6— 5x) = — 9;
x-> 3-0
/(3 + 0)== lim (x — 3) = 0.
*-* 3 +
We see that the right-hand and the left-hand limits, though finite,
are not equal each other, therefore the function has a disconti-
to
nuity of the first kind at the point x = 3.
The |ump of the function at the point of discontinuity is
f(3 + 0)-f(3-0)=0-(-9) = 9.
The function is
(c) defined and continuous throughout the entire
number scale., except at the point #=3/2. Since 2x — 3>0 for
x > 3/2 and 2x- 3 < for x < 3/2,
i 1 at x > 3/2,
\ — 1 at *<3/2.
Hence,
/ (3/2 + 0) = / (3/2-0) = -1. 1 ,
Therefore, at the point x = 3/2 the function has a finite discon-

tinuity of the first kind. The jump of the function at this point
/ (3/2 + 0) — / (3/2 — 0) is equal to 1 —— = ( 1) 2.
1.14.3. Test the following functions for continuity:

( sin x f , n
(a) l(x)=\ a
[ 1 furx=-0;
(b) /(*)= sin(l/x);

x sin (\/x) for x =^= 0,
(C) /(*) =
for 0;
= (
4-3* for x < 0,
(d) (x)
2a + x
/
for 0;
(e) /(x) = arctan(l/*); (f) / (x) = (** + l)/(* + 1).
Solution, (a) The function is continuous at all points ,v=^Q. At
the point x we have =
sin x
/(0)=1; liin :
= liin
x
1.
Hence, at this point the function is continuous as well, which

means that it is continuous for all values of x.
Fig. 30
(b) is defined and continuous for all

The function x^0. There
3re limits at the point x
no one-sided (cf. Problem 1.13.1 =
(e)). Therefore, at the point x =
the function suffers a disconti-
nuity of the second kind (see Fig. 30).
(d) / (—0) =
4, and / (+0) -
2a; the equality / (—0) f ( 0) / (0) = + -
will be fulfilled, i. e. the function f (x) will be continuous at the
point x = if we put 2a = 4, a = 2.
(f) f (_i_0) = /(— 1+0)= lirn
-
(x 2 — *+l) = 3, i.e. both one-
x-> 1
sided limits are finite and coincide. But at the point x =— 1 the
function is not defined and, therefore, is not continuous. The graph

of the function is the parabola y x2 x 1 with the point =(— 1, 3) — + M
removed. If we redefine the function putting / ( 1)=^3, then it —
will become continuous. Thus, at x== 1 the function has a remo- —
vable discontinuity.
1.14.4. Test the following functions for

continuity:
(a) / (x)
= E (x).
should be borne in mind that the function
It
E(x) is defined as the maximum integer n contained in the num-
ber x, i. e. as a number satisfying the inequality n^x.
(b)
|
1 if a; is rational,
\ if x is irrational.
X(x) is called the Dirichlet function. For instance, X (0) = 1; h(— 1/2) = 1;
X (|/""2) = 0; *(jx) = f
etc.
Solution, (a) The function E (x) is defined throughout the entire
number scale and takes on only integral values. This function is
discontinuous at every integral value n of the independent va-
riable, since E (n — 0) = n— 1;
E (
n + 0) = n (see Fig. 31). 9>
(b)Let us choose an arbitrary 3 r**>
point x on the x-axis; two cases
2
are possible: (1) the number x i
! '
is rational; (2) the number x is 1

irrational. -3 -2 -1
,
i
i
i
i i ——
i
In the case X(x )=l. In

first ! I ! o 12 3 4
any vicinity of a rational point ! ! H- -1
there are irrational points, where
X(x) =
0. Hence, in any vicinity
I
fJ
i
-z
of x there are points x for which i-*J -3
|A*/|HM*o)-M*)Hi- Fig. 31
the second case

In X (x ) 0. =
In any vicinity of an irrational point there are rational points
at which X(x)=l. Hence, it is possible to find the values of x for
which
\Ay\ = \X(x )-X(x)\ = 1.
Thus, both cases the difference A*/ does not tend to zero as
in
Ax—>0. Therefore, x is a discontinuity. Since x is an arbitrary-
point, the Dirichlet function X(x) is discontinuous at each point.
The graph of this function consists of a set of points with irratio-
nal abscissas on the x-axis and of a set of points with rational
abscissas on the straight line y 1, that is why it is impos- =
sible to sketch it.
1.14.5. Using the definition of continuity of a function in terms

of "e —
6", test the following functions for continuity:
(a) f(x) = ax + b (a^=0);
|
x2 if x is rational,
(b) f (*) — |
— x2 jj x j s i rra tional
Solution, (a) Choose an arbitrary point x According to the "e 8" .

—
definition it is necessary to show that for any preassigned, arbitra-
rily small number it is possible to find a number such e>0 S>0
that at a: x |
—
6 the inequality / (x) f (x ) <
e holds true.
| |
— |
<
Consider the absolute value of the difference
J
f(x) — f (x ) |
= |
(ax + b) — (ax + b)\ = \ax + b— ax —b\ = \a\\ x—x |
Let us require that \f(x) /(* )|< e This — - requirement will be

fulfilled for all x satisfying the inequality
\a \
\x —x < \
e or |
a: — x < e/|a| \
(a^=0).
Hence, if we take 8ê/|a|, then at \x — a: |<8 the inequality

\f( x ) —
/(*o)l< e is fulfilled. Continuity is thus proved for any
point x =x .
(b) Choose an arbitrary point x . If {x n \ is a sequence of rational

numbers tending to x , then lim f(x n ) = xl. If {a:^} is a sequence
of irrational numbers tending to x 0y then lim f(x'n ) = — x%. At x =£Q
the indicated limits are different and hence the function is discon-
tinuous at all points x=?^0.
On the other hand, let now x 0. Find the absolute = value of
the difference \f(x) f(0)\: —
\f(x)-f{0)\ = \±tf-0\=.x*.
It is obvious that x 2 <e at |x| <j/"e. If e>0 is given, then,
putting 6<j/e and \x 0| = |x]< 6, we obtain A/(0) = x < e.
2
— |
|
Hence, ?t the point x the function is continuous. And so, the =

point x = is the only point at which the function is continuous. Note
that the function under consideration can be expressed through the

Dirichlet function (see Problem 1.14.4 (b)): / (x) x 2 [2k (x)— 1] =
1.14.6. Determine which kind of discontinuity the following
functions have at the point x x= :
(a (x) = \
x+2 for x<2 f
—
<
f
1 x 2
1 for x>2; x = 2;
(b) / (x) = arc tan ^-5 x = 5; o (c) f (x) = j^H* x» = 0;
(d) /(A') = tanx; x = n/2' f
(e) f(x) = Vx — E{V

x)\ x -----n\ where n is a natural number.
Solution, (a) Find the one-sided limits at the point x Q -=2\
f(2 — 0)= lim (* + 2) = 4;

X-+ 2 -
f(2 + 0) = X*lim+ (x 2 — 1) = 3.
2
Here the limits to the right and to the left exist, are finite but
do not coincide, therefore the function has a discontinuity of the
first kind at the point x = 2.
(e) The function E {V x) has discontinuities of the first kind at

every point x =n 2
y
where n is a natural number (see Problem
1.14.4 (a)), whereas the function |/ x is__continuous at all xÔ.
Therefore the function f(x) = V x — E (V x) has discontinuities of
the first kind at the points 1, 4, 9, . . .
,
n2 , ...
1.14.7. Test the following functions for continuity

ex —\
(a) f(x) = '-
X
/ ex
\
(b) /(*) = <

*
I 3 for * = 0;
K/MorÔ,
| for x = 0;
(d) /(*)-- lim (smxr; (e) f(*) = ii|££l
(f) f(x) = E(x) + E(-x).
1.14.8. For each of the following functions find the points of
discontinuity and determine the jumps of the function at these
points:
(a) /W = F=^T- 1
;
(b) /W = * + r^!|;
—x for
(d) f(x)=\
(
(
_ 2 .
for X>L
1,
1.14.9. Redefine the following functions at the point x = so as

to make them continuous:
/
(a)
x r
/(*)
, v
= _;
tan x
(b) f(x) = 5x2 2x 3 *

(c) f(x) = v~n~x—
X
sin'2 x
id) f(x) = — cos x
1
'
§ Arithmetical Operations on Continuous

1.15.
Functions. Continuity of a Composite Function
If the functions f (x) and g(x) are continuous at the point x =x 0f

then the functions
(1) f{x)±g(x); (2) f(x).g(x)\ (3) l^l(g(x )^0)
are also continuous at this point.

If the function u =
(p(x) is continuous at the point x x and =
the function y =
f(u) is continuous at the point u Q tp(x ) then = f
the composite function y =

f [<p(x)] is continuous at the point x x = .
1.15.1. Test the following functions for continuity:

2x b — 8x 2 +ll
(a) f(x)-. '
x4 + 4x 3
+8a; 2 + +4 8a:
»
3 sin 3 #+ cos 2 x-\- 1

#
(b) f(x)-
4 cos a: —2 *
3 cos x^r x 2 sin *

(c) f(x) =x cos ( 1 /sin x)
Solution, (a) A function representing a ratio of two continuous

functions (polynomials in this case) is discontinuous only at points
for which the denominator becomes zero. But in our case
+ 4x + 8x + 8x + 4 = (x + 2x + 2)
x* 3 2 2 2
,
and since x 2 + 2x + 2 = (x + + >0 at any x, the

1)
2
1 denominator
never becomes zero. Hence, the function f (x) is continuous through-
out the entire number scale.
(b) The function f(x) suffers discontinuities only at points for
which the denominator equals zero, i.e. at points which are the
roots of the equation
4cos# —2 = or cosa:=1/2,
whence
x = x n = ±n/3 + 2nn (n = 0, ± 1 , +2, . . . ).
Thus, the function f (x) is continuous everywhere, except at the

point x n .
§ 1.15. Operations on Continuous Functions 85
(c) Just as in the preceding example, the numerator is continu-

ous throughout the entire number scale. As far as the denominator
is concerned, according to the theorem on continuity of a composite
function, it is continuous at points where the function u=l/s\nx

is continuous, since the function cos u is continuous everywhere.
Hence, the denominator is continuous everywhere, except at the
points x =
kn (k an integer). Besides, we must exclude the points
at which cos(l/sinx) 0, i.e. =
the points at which l/sinx =
= (2p+\)n/2 (pan integer), or sin x 2/[(2p 1) n]. Thus, the = +
function f (x) is continuous everywhere except at the points x = kn
and x = {
— 1)" arc sin
2
^n
+ nn (&» P» n = 0» ± U ± 2, ... ).
1.15.2. Test the following composite functions for continuity:
(a) y = cosx n y
where n is a natural number;
(b) y=^ cos log x\
(c) y =V 1/2 — cos 2

x.
Solution, have a composite function y = cosu, where

(a) We
u = xn . = cosu is continuous at any point u, and
The function y
the function u = x n is continuous at any value of x. Therefore, the
function y = cosx n is continuous throughout the entire number scale.
(c) Here y = 1/2
2
V —
where u = cosx. The function Vl/2 —u 2
y
is and continuous on the interval

defined J/ 2/2, ]/~2/2|, the [ —
function u = cosx is continuous throughout the entire number scale.
Therefore, the function y = Vl/2 — cos 2
x is continuous at all values
of x for which
(
n/4 + 2nn<^x<^3ji/4 + 2nn,
|cos*|<l/ 2/2, i.e.
^ 5n/4 + 2 nn^x<^7n/4 + 2nn.
1.15.3. For each of the following functions find the points of
discontinuity and determine their character:
(a) y = -TTT- -
— o » where u = ——7 {
( x— for1 0,
(b) y = u\ where u = x+l fof X<Q .
(c) y= j-j-^-^
l
where w = tan*.
Solution, (a) The function
, v 1
suffers a discontinuity at the point x=\. The function
suffers a discontinuity at points where u 2 + u — 2 = 0, i.e. ul =—2

and w 2 =l. Using these values of w, find the corresponding values
of x by solving the equations:
whence x =1/2 and x = 2.

Hence, the composite function is discontinuous at three points:
Xj--=l/2, x 2 =l, x 3 = 2. Let us find out the character of disconti-
nuities at these points.
Jim y = lim y = 0,
X -y 1 U -* cc
therefore x 2 =l is a removable discontinuity.
lim y = lim y = oo; lim */ = lim y ^ oo\

x -* \/2 u -+ — 2 x -> 2 u -* 1
hence, the points = 1/2, #, =2 are discontinuities of the second

kind.
1.15.4. Given the function f(x) l/(l—x). Find = the points of

discontinuity of the composite function
y = f{f[f(x)]\.
Solution. The point x = 1 is a discontinuity of the function
If x^=\, then
" =f [/(*)] = ,
-i/'d-v, "T 1 -
Hence, the point * = is a discontinuity of the function

« = /[/(*)]•
If jc=^0, 1, then
* = /</[/<*>]Htz^^ = *
iscontinuous everywhere.
Thus, the points of discontinuity of this composite function are
x =x=l, both of them being removable.
9
§ 1.16. Fund. Cont. on Closed Interval: Properties 87
§ 1.16. The Properties of a Function Continuous on a

Closed Interval. Continuity of an Inverse Function
I. The function f (x) y
continuous on the interval [a, 6], possess-
es the following properties:
(1) f (x) is bounded on [a, b]\
(2) f (x) has the minimum and maximum values on [a, b]\
(3) If m= min f (x), M=
max f(x), then for any A satisfy-
ing the inequalities A^.M there exists a point x £[a, b] for
which f(x ) A. =
In particular,
if f(a)-f (b) 0, then we can find a point <
c (a <c<fr) such that f(c) 0. =
II. Continuity of an Inverse Function. If the function y = f(x) is
defined, continuous and strictly monotonic on the interval X, then
there exists a single-valued inverse function x = y(y) defined, con-
tinuous and also strictly monotonic in the range of the function
y=f(x).
1.16.1. Does the equation sin*— x+ 1 =0 have a root?
Solution. The function
/ (x) — sin x — * + 1
is continuous over the entire number scale. Besides, this function

changes sign, since /(0)=1, and / (3ji/2) = 3n/2. Hence, by pro- —
perty (3) within the interval [0, 3jt/2] there is at least one root
of the given equation.
1.16.2. Has the equation x b — 18*-}- 2 = roots belonging to the

interval [ — 1, 1]?
1.16.3. Prove that any algebraic equation of an odd power with

real coefficients
a * 2w + + a,* + 2
" . . .
+ a 2n x + a 2n = , 1 (*)
has at Least one real root.

Solution. Consider the function
/ (*) =a x*"
* 1
+ a lX +...+ a *n
2n x -f a 2n+v
which is continuous throughout the number scale.

Let, for determinacy sake, a > 0. Then
11 rn f(x)=-\- oo, and Urn f(x) =— oo.
X + oo x -* - crj
Hence, we can find numbers a, b a0. By property (3), between a and b there exists a number

c such that f(c) Q, = which proves that the equation (*) has at
least one real root.
1.16.4. Let the function f (x) be continuous on [a, b\ and let

the equation f(x) =
have a finite number of roots on the inter-
val [a, b\. Arrange them in the ascending order:
a < x l < x2 < x, < . . .

< xn < b.
Prove that in each of the intervals

(a, xj, x 2 ), (x 2 ,
a:,), ... , (* M ,
b)
the function retains the same sign.

Solution. If the function changed its sign on a certain interval,
then we could find one more root of the function, which contradicts
the condition. To determine the sign of the function on any of the
indicated intervals it is sufficient to compute the value of the func-
tion at an arbitrary point of the appropriate interval.
1.16.5. Given a function on the interval [ — 2, +2]:

/ x2 + 2 if —2 < x < 0,
f(x) =\ — (x + 2) 2
if 0<x<C 2.
Is there a point on this closed interval at which f(x) 0? =

Solution. At the end-points of the interval [ 2, +2] the given —
function has different signs:
f(_2) = +6; f(+2) = -6.

But it is easy to notice that it does not become zero at any point
of the interval [—2, +2]. Indeed, jk
2
and (x
2
2) <0 + 2>0 — +
at any x; this is due to the fact that f (x) has a discontinuity at
the point x = 0.
1.16.6. Does the function
/ (x) = x /4 — sin nx + 3
3
take on the value 2y within the interval [—2, 2]?
Solution. The function f(x) = x 3 /4 sin:ru; 3 is continuous — +

within the interval [—2, 2]. Furthermore, at the end-points of this
interval it attains the values
/(-2)=1.; /(2)=5.
Since 1 < 2-^- < 5, then, by property (3), withi-n the interval
[—2, 2] there exists at least one point x such that f(x) = 2-y.
1.16.7. Show that the function
( 2x +\ for-l<x<0,
f(x)= |
2 X for x = 0,
[
2*— 1 for 0<*< 1,
§ 1.16. Funct. Cont. on Closed Interval: Properties 89
defined and bounded on the interval 1, 1], has neither maximum, [

—
nor minimum values.
Solution. In the interval [—1, 0) the function increases from 3/2
to 2 and in (0, 1] it increases from to 1, it does not attain either
the value 2 or 0. Therefore the function is bounded but never reaches
its upper and lower bounds. This is because there is a discontinuity
at the point x 0. =
1.16.8. Show that on any interval [a, b] of length greater than
unity the function f (x) =x — E (x) attains its minimum value but
never reaches its maximum.
Solution. In any interval \n, where n is an integer, the
given function f (x) increases from to 1, never attaining the maxi-
mum. Hence, 0^f(x)< for any 1
x. Since on the interval [a, b] we I
can find at least one internal in-

tegral point n then f(n) and y
= , *i , , ,
lim /(*) =
but f{x)=£\ for any
!,
/ / // / % x
x. It means that the function reaches -2-10 12 3
its minimum value but never p .
32
reaches its maximum. This is be-
cause there is a discontinuity at the point x =n (see Fig. 32),
2n +
Prove that the function y =
1.16.9. \/ x (n a natural number)
is continuous throughout the number scale, considering it as a function
inverse to y = x
2n+1
.
Solution. The function y = x 2n+i is continuous and increases from

— oo to oo over the entire number scale. Hence, the inverse function
+
jt = 2/ '
j/ y is defined for all y continuous and increasing. Denoting
y
the independent variable again as x we find that the function

y
y = 2n+ y/ x possesses the required properties.
1.16.10. Prove that for any function of the form

"" 1
y -a x 2 " +1
+ a^ 2
+a 2
x*«~ 3 + . . .
+ a x + a n+ n , , (*)
where a a iy a 2 ,
an are positive numbers, there exists
, ,
an inverse function increasing and continuous throughout the num-

ber scale.
x 2n+l increase
b
Solution. As is known, the functions x, r 1
,
x , . . .,
throughout the entire number scale. Then, since the coefficients

a (i = 0, 1,
i
.., A2+1) are positive, the function f (x) =a u x
.
2n+1
+
2n ~
-\-a x x
+ + a n x-\- a n+l also increases. Furthermore, it is con-
l
. . .
tinuous. Therefore, for a function of the form (*) there exists an

inverse function increasing and continuous over the entire number
scale.
Note. This example establishes only the existence of an inverse

function x =
g(y), but gives no analytic expression for it. It is not
always possible to express it in radicals. The problems of the exis-
tence of an inverse function and of expressing it analytically should
not be confused.
1.16.11. Prove that there exists only one continuous function
x = x(y) ( — oo < y <
which satisfies the Kepler equation:
oo)
x e$'mx = y (0<e<l). —
Solution. Let us show that y(x) is an increasing function. Let
x < x 2 be arbitrary points on the number scale. Then
x
y (* 2 ) —y (*i) = (*, — e sin — (*, — e sin x x ) =

= (x — xj — e (sin x — sin x
2 2 t ).
Estimate the absolute value of the difference slnx^ — sin x \ x |:
|
sin x 2 — sin x l
sin
2
cos—
-x
<2 sin
1
<2 •
\x 2 x t \
(x 2 x x ).
Since <e< 1,
— sin x < (x — x
e |
sin x2 x \
2 1) y
whence
(x 2 — x — e(s\nx — smx )=y(x — y(x >
l
) 2 l 2) l)
0.
Since y (x) is a continuous function in the interval (— oo, oo), the

inverse function x is a single-valued and
continuous function of y.
1.16.12. Show that the equation
has one root on the interval [1, 2]. Cal-

culate this root approximately to within
two decimal places.
1.16.13. The function f (x) is defined

on the interval [a, b\ and has values of
the same sign on its end-points. Can one
assert that there is no point on [a, b]
at which the function becomes zero?
1.16.14. Prove that the function
(
x+ 1 at — <x<0,
1
f{x) = \
_x at 0<x< 1
is discontinuous at the point x and still has the = maximum and

the minimum value on [— 1, 1] (see Fig. 33).
§ 1.17. Additional Problems
1.17.1. Prove the inequalities:

(a) n \
< {^Y^y for a natural n > 1;
2n ~
(U\
(D) — — — < y2n~TT
1 l
'
2 4
'
6 2n' ' ' '

(a) 20 2 303 > 303 202 ;
(b) 200! < 100 200 .
1.17.3. Solve the inequalities:
(a) ||*|-2|<1;
(b) ||
2 — 3^|— J_|_> 2;
(c) (*— 2)K* +1 2

>x + 2. 2
1.17.4. Can a sum, difference, product or quotient of irrational

numbers be a rational number?
1.17.5. Do the equations
(a) |
sin x\ =sin x + 3, (b) |
tanx| -=tanx-\-3
have any roots?
1.17.6. Prove the identity f^4^T + -^4^V = *

2
(
2
-
1.17.7. Prove the Bernoulli inequality

(\ +Xl )(\+x 2 ) ... (\ + x )^\+x n }
+x, + ...+x nt
where x lt x 2 >
... y
xn are numbers of like sign, and l+x,->0
(/ = 1, 2,

(a) f(x)=]/"^=7*;
(b) f(x)=}/ s'mVx;
(c) / (x) =- V — sin 2
nx\
(d) /(x)=- 7 =J== and g(x)=-=L=.\

1/ |
x |
—x V x— \x\
(e) (a:) = arc sin (| a: |
— 3);
(f) /(A')=arcco SinL.
1.17.9. Are the following functions identical?
(a) / = and cp(x)= 1;
^b) / (jc) = log x and (a) = 2 log a; 2

<p
(c) f(x)=x and q) (x) = (|/ x) 2 ;
(d) f(x)^l and cp (a:) == sin 2 x-\- cos 2 x\

(e) f (*) = lqg(*— l) + log(x
— 2) and cp(*) = log (*— 1) (*— 2).
1.17.10. In what interval are the following functions identical?
(a) f(x)=x and cp (x) =
10 lQ g*;
(b) f(x)=VxVx— 1 and <p(*) = K* (*— 1).
1.17.11. An isosceles triangle of a given perimeter 2/? ^=12 revol-
ves about its base. Write the function V (x), where V is the volume
of the solid of revolution thus obtained and x is the length of the
lateral side of the triangle.
1.17.12. Investigating the domain of definition of functions,

(a) solve the inequality
K* + 2 + Vx—5 > l/"5^;

(b) prove that the inequality
log 2 -,(*-3)>-5
has no solutions.
1.17.13. The function y = s\gnx was defined in Problem 1.5.11 (n).

Show that
(a) ^^signx;
(b) x = |
x sign
|
x\
(c) sign (sign x) = sign x.

1.17.14. Prove that if for a linear function
/ (x)
= ax + b
the values of the argument 1 2, x = x n (n =
) form an arithmetic ,
. .
.
progression, then the corresponding values of the function
yn = f(x») (/i = l,2, ...)
also form an arithmetic progression.
1.17.15. Prove that the product of two even or two odd functions
is an even function, whereas the product of an even and an odd
function is an odd function.
Prove that if the domain of definition of the function

1.17.16.
f(x) symmetrical with respect to x=0 then f(x)
is f(—x) is an +
even function and f (x) /( x) is an odd one. — y
1.17.17. Prove that any function f (x) defined in a symmetrical

interval ( — /, /) can be presented as a sum of an even and an odd
function. Rewrite the following functions in the form of a sum of

an even and an odd function:
(a) /(*)=-£±l; (b) y=a*.

1.17.18. Extend the function f(x)=^x 2 x defined on the inter- +
val [0,3] onto the interval [ —
3,3] in an even and an odd way,
1.17.19. The function {x\ = x — E (x) is a fractional part of a

number x. Prove that it is a periodic function with period 1.
1.17.20. Sketch the graph of a periodic function with period

T= 1 defined on the half-open interval (0, 1] by the formula y-—x 2 .
1.17.21. Let us have two periodic functions f (x) and y(x) defined
on a common set. Prove that if the periods of these functions are
commensurate, then their sum and product are also periodic functions.
1.17.22. Prove that the Dirichlet function X(x) (see Problem
1.14.4(b)) is a periodic one but has no period.
1.17.23. Prove that if the function
f (x) -=s\nx-{-coscix
is periodic, then a is a rational number.
1.17.24. Test the following functions for monotony:

(a) f{x) = \x\; (b) f(x)=\x\-x.
1.17.25. Prove that the sum of two functions increasing on a
certain open interval is a function monotonically increasing on this
interval. Will the difference of increasing functions be a monotonic
function?
1.17.26. Give an example of a non-monotonic function that has

an inverse.
1.17.27. Determine the inverse function and its domain of de-

finition if
(x if — oo < x < 1,
2
x' if 1<x<4,
2X if 4 < x < oo.
1.17.28. Show that the equation x 2 + 2jc+ 1 == — + Yx1 has no
real roots.
1.17.29. Construct the graph of the function
y = f(x-i) + f(x+0,
where
k(\—\x\ll) at
fW =
|
j
at \x\ > /.
1.17.30. Knowing the graph of the function y = f(x), sketch the

graphs of the following functions:
(z)y = F(x); (b) y = VW)\ (c) y = f[f{x)].
1.17.31. Prove that the graphs of the functions y = log fl
x and
y = \og a nx can be derived from each other by changing all ordinates
in the ratio 1 l/n.
1.17.32. Prove that if the graph of the function y-=f(x) defined y
throughout the number scale, is symmetrical about two vertical

axes x = a and x = b (a<fr), then this function is a periodic one.
1.17.33. Let the sequence x n converge and the sequence y n diverge.

What can be said about convergence of the sequences
(a) xn + yn \
(b) xny n ?
1.17.34. Let the sequences x n and y n diverge. Can one assert that
the sequences x n +
y tn x n y n diverge too?
1.17.35. Let an be an interior angle of a regular n-gon (ai=3,
4, ...). Write the first several terms of the sequence a n Prove .
that \ima n = n.
1.17.36. Prove that from \imx n =a it follows that lim |a:
w |
=|a|.
n -> oo n -* oo
Is the converse true?
1.17.37.sequence has an infinite limit, does it mean that this

If a
sequence is unbounded? And
if a sequence is unbounded, does it
mean that it has an infinite limit? Prove that x n = n - l)H is an {
unbounded but not an infinite function.

1.17.38. Prove that the sequence {a„}, where a n is the nth digit
of an arbitrarily chosen irrational number, cannot be monotonic.
1.17.39. Prove that if the sequence {ajb n } (b n >0) is monotonic,

then the sequence
fli + fl2+---+a« I
will also be monotonic.
1.17.40. Prove the existence of limits of the following sequences

and find them.
(a) V2 y j/ 2 [ 2 y V 2 V 2 |/T
(b) xn = c"lYn\ (c>0, k>0);

(c) xn = ajn, where a n is the nth digit of the number jt.
1.17.41. Prove that at an arbitrarily chosen x the sequence

j£û)j>
is bounded.
(
£(*) + £ (2*) 1- ...+E (nx) \
has the limit x/2.
1.17.43. Prove that

lima*=l (a>0).
h ->

1 +# for x 0,
for x = 0.
Prove that
li:n /(*) = 1.
x -
1.17.45. Let
Prove that
oo, if n > m,
lim P(x) = |
a /6 , if n =m y
0, if n < m.
1.17.46. Find the constants a and & from the condition:
(b)
X
lim
-> - oo
()/ > — x + — ax— b) = 1 Q.
1.17.47. Sketch the graphs of the following functions:
(a) f(x)= lim ,V 1 0);

/Z -v 00
(b) /(*)= lim sin 2 "*.

n -* go
1.17.48. Prove that
lim [(1 + x) (1 + x») (1 . . . ( 1 -!- a 2 ")] = (| x \ < 1).

« -> GO
1.17.49. Can one replace infinitesimal summands by equivalent

infinitesimals in computing a limit?
1.17.50. Determine the order of smallness of the chord of an

infinitely small circular arc relative to the sagitta of the same arc.
1.17.51. Determine the order ofsmallness of the difference of the

perimeters of an inscribed and circumscribed regular n-gons rela-
tive to an infinitely small side of the inscribed n-gon.
1.17.52. The volumetric expansion coefficient of a body is con-

sidered to be approximately equal to the triple coefficient of linear
expansion. On equivalence of what infinitesimals is it based?
1.17.53. Does the relation log(l+A;)~A hold true as x-^0?

1.17.54. Will the sum of two functions f(x) g(x) be necessarily +
discontinuous at a given point x if:
(a) the function f (x) is continuous and the function g(x) is dis-
continuous at x x = ,
(b) both functions are discontinuous at x = x ? Give some examples.
1.17.55. Is the product of two functions f(x)g(x) necessarily

discontinuous at a given point x if:
(a) the function f (x) is continuous and the function g(x) is dis-
continuous at this point;
(b) both functions f (x) and g(x) are discontinuous at x = x ? ()
Give some examples.
1.17.56. Can one assert that the square of a discontinuous func-

tion is also a discontinuous function? Give an example of a func-
tion discontinuous everywhere whose square is a continuous function.
1.17.57. Determine the points of discontinuity of the following

functions and investigate the character of these points if:
' x)
(b) / {x) = 2^ l/ii
\
(c) y(x) = x[l — 2X(x)] y

where X(x) is the Dirichlet function (see
Problem 1.14.4 (b)).
1.17.58. Test the following functions for continuity and sketch

their graphs:
(a) y = x— E{x)\
(b) y = x* + E(x>);
(C) y={-\)^).
(d) y= 1 n
+ (2sin^
-
fl ! L l
1.17.59. Investigate the functions f[g(x)] and g[f(x)] for conti-

nuity if / (x) = sign x and g(x) x(l x 2 ). = —
2x at — 1<*<0,
f(x) =
x + V, at 0<x<l
is discontinuous at the point x = and nonetheless has both maxi-
mum and minimum values on [ — 1, 1].
Ascertain that on the interval [ 2, 2] the function takes on —

all intermediate values from /(— 2) to f (2) although it is discon-
tinuous (at what point?).
1.17.62. Prove that if the function f (x): (1) is defined and mo-
notonic on the interval [a b]\ (2) traverses all intermediate values
y
between f (a) and f(b), then it is continuous on the interval [a, b],
1.17.63. Let the function

y f(x) be continuous on the interval =
[a, b], range being the same interval a^yîb. Prove that on
its
this closed interval there exists at least one point x such that
f(x)=x. Explain this geometrically.
1.17.64. Prove that if the function f (x) is continuous on the
interval (a, b) and x19 x 2 x n are any values from this open
,
interval, then we can find among them a number £ such that
f(l) = j[ U(x + f(x )+...+f(xn

1) 2 )].
1.17.65. Prove that the equation x 2*=1 has at least one posi-
tive root which is less than unity.
1.17.66. Prove that

if a polynomial of an even degree attains at
least one value the sign of which is opposite to that of the coeffi-
cient at the superior power of x of the polynomial, then the latter
has at least two real roots.
1.17.67. Prove that the inverse of the discontinuous function

y = (l-\-x 2
)
sign x is a continuous function.
4-3148
Chapter
DIFFERENTIATION
OF FUNCTIONS
§ 2.1. Definition of the Derivative
The derivative f (x) of the function y — f(x) at a given point x

is defined by the equality
r/ / \ i- &y t- f (x+Ax) — f (x)
A* - A* Ax - ax
If this limit then the function f (x) is called differen-

is finite,
tiable at the point x\ and it is infallibly continuous at this point.
Geometrically, the value of the derivative /' (x) represents the
slope of the line tangent to the graph of the function y = f(x) at
the point x.
The number
r / v \_ Hm f(x+Ax)—f(x)
is called the right -side derivative at the point x

The number
is called the left-side derivative at the point x.

The necessary and sufficient condition for the existence of the
derivative /' (x) is the existence of the finite right- and left-side
derivatives, and also of the equality fi (x) f+ (x). =
If f(x) =
oo, the function f(x) is said to have an infinite deri-
vative at the point x. In this case the line tangent to the graph
of the function y =
f(x) at the point x is perpendicular to the
x-axis.
2.1.1. Find the increment Ay and the ratio ^ for the following
functions:
(a) y = ]/"x at x= and A* = 0.0001;

(b) y =x l
at x=l and A* = 0.2.
*
+ x _Q
§2.1. Definition of the Derivative 99
Solution, (a) Ay = V x + Ax — |/"x =--V0. 0001 -=0.01;
^
Ax
^L =10o
= 0.0001
2.1.2. Using the definition of the derivative, find the derivatives
of the following functions:
(a) y = cos ax\ (b) y = 5x

2
— 2x.
Solution, (a) Ay = cosa (x + Ax) — cosax =
= — 2sin âx + Y Ax^ sin y x\
-2sin( ax-\--^- Ax j sin ~ Ax

Ax Ax
— CL
Ax
Aa: - lx
'
2 lim
A* -*
sm( ax + £.
V 2
^ Hm —
^VaxÔ / Ax -
sin
A*
= — a sin ax.
In particular, if a^=l, then */ = cosx and y' = — sinx.
2.1.3. Show that the following functions have no finite derivati-
ves at the indicated points:
(a) y= ]/ x at 3
the point x=0;
(b) y=\/x — 1 at the point x=l\
(c) y = 3\x\+l at the point x = 0.
Solution, (a) Ay - j/(x+ Ax) 3 — j/x 3
.
At x--=0we have A#=j/573

v
=
'Ax Ax
=
F
=; hence, ,'(0)^^ = 00,
i.e. there is no finite derivative.

(c) At Ax>0 the increment of the func-
Fig. 34
tion Ay at* = will be: Ay 3 (0+ Ax) = + — 1 1 =3Ax. Therefore
Ax - +
lim g
ax
= 3.
At Ax<0 the increment of the function Ay will be

Ay = — 3(0 + Ax) + — 1= — 3Ax, 1
hence,
A*
lim
-> -0
Ay
aa:
= — 3.
Since the one-sided limits are different, there is no derivative at
the point x =
(see Fig. 34).
4*
100 Ch. II. Differentiation of Functions
2.1.4. Investigate the function y = \\nx\ for differentiability at

the point x 1. =
Solution. At x 1 =
Ay = \\n(l + Ax)\ — |
In 1 1
= |
ln(l + Ax)|,
l. e.
Ar/=|ln(l+A^)H|^
ln(l
( + A*) at Ajk>0,
ln | + Ax) 1 at Ax<0.
Therefore
I n (1 + Ax) at A* > 0,
Ax
Ax ln(1
+ Ajc)
at Ax<0,
whence
lim ^=+1 and lim ^= 1.
Since the one-sided limits are different, there is no derivative.

Hence, the function y \\nx\ = is not differentiate at the point
x= 1 (see Fig. 35).
2.1.5. Find the average velocity of
motion specified by the formula
s= (f — 5^ + 2) m2
from tt = 5 sec to = 15 sec. /

2
2.1.6. Using the definition of the

derivative, find the derivatives of
the following functions:
Fig- 35 (a) y =x 3
\
(b) y-^\/x\
2.1.7. Investigate the function y = \cosx\ for differentiability at
the points x n/2 = + nn (n an integer).
§ 2.2. Differentiation of Explicit Functions
I. Basic Rules of Differentiation
(1) c'=0;
(2) (u±v)' = u' ±v'\
(3) (cu)'=cu'\
(4) = u'u + uv', the product rule;
(uv)'
(5)
^y = ii!E=ffHl(t,^0) the quotient rule. f
Here c = const, and u and v are functions of x which have deri-

vatives at a corresponding point.
(6) If the function u =
y(x) is differentiate at the point x 0t and
the function y =
f(u) is differentiate at the point a <p(*o)> ên =
§2.2. Differentiation of Explicit Functions 101
the composite function y = f(y(x)) is differentiate at the point *
and y' (x
x ) = y'u (u )
u'x (x ), the function of a function, or chain, rule.
II. Differentiation of Basic Elementary Functions
(I) (u = nu n ~ u'\ (2) (sin u)' = cos u-u';

n )' 1
(3) (cosu)' = — s\nu-u';
(4)/
V
(tanw)'=-^-;
V (5) (cot u)' = — -X-
/ V / \ / ;
»
cos w sin u 2 »
2
(6) (lna)' = -^;

(7) (fl
B
)' =a 8
lna-«'; 8) (<?")'=<?"«';
(9) (sinh */)' = cosh w';
(10) (cosh a)' = sinh h-w';
(II) (arcsin a)' =
"
= — (arccos^)'; =
(12) (arc tan^)'=

T -Î = — (arc cot w)'.
2.2.1. Find y\ if:
(a) = 5*
*/
2/3 — 3* 5/2 +2*" 3 ;
a
(b) y= 3
(a, b constants).
y xl x y x
Solution, (a) y' = 5.1

3
X *'*- 1
— 3 — x*'*- — 2.3*"
2
1 3 "1
=
3 ^ a:
lo
15 ,/-- b
'
2
2.2.2. Find */', if:
n — cos x
(a) v = 3cos* + 2sin x\ (b) v = -
/ o
\ o /u\ .
• si at }—
smx — cos x
!
/y */ ;
v '
1 v '
(c) y = (x + 1) arc tan*; (d) */=:* arcsin*.

2 3
Solution, (a) = 3 (cos*)' + 2 (sin *)' = — 3 sin * + 2 cos*;

(sin x-\- cos x)' (sin — cos x) — (sin x — cos x)' (sin x-\- cos a: a:)
(b) y'
(sin — cos x) a:
2
(cos a:— sin (sin x — cos x) — (cos x-{- sin (sin #+cos x) jc) a:)
(sin x — cos a:)

2
*
(sin a:— cos a:) 2
3
= (* *+ (arc sin *)' * = 3x
a:
(d) y' 3
)' arc sin 3 2
arcsin * + -^-== .
2.2.3. Find the derivative of the given function and then com-
pute the particular value of the derivative at the indicated value
of the argument:
(a) /(*) = !
— 1/7 + 16/* 2
at x = — 8;
(b) f(x) = (\—Vr x) /x at x = 0.01; 2
(c) f(t) = (cos 0/(1 — sin at f=n/6.
Solution, (a) f (*) = — {r" J - 16;r = a

^7=—^.
•*
x
3 1/ x
Putting * = — 8, we obtain
— sin / (1 — sin + cos/)

2
/ _ 1
(c) /'(/) ~~
(1 — sin t)
2
1 — sin /
*
Whence /'(ji/6) = 2.
2.2.4. Taking advantage of the differentiation formulas, find the
derivatives of the following functions:
(a) y = 2x + 3x— 5;
3
(b) y = V~7+ -^= + 0Ax 10
;
.
x 2x 2 + + a; 1 , x-{- V*
-cos (p
ex -j- sin x
(g) y = ex (cosx+smx)\ (h) #
2.2.5. Taking advantage of the rule for differentiation of a com-

posite function find the derivatives of the following functions:
(a) # = sin #; 3
(b) = lntanx; (c) y = 5 C0SX \
# = lnsin (x + = arcsin [/l—

3 2
(d) 1); (e) */ a; ;
(f) = In (tan 3*);

*/
5
(g) y = sin |/T/(T=7)
2
Solution, (a) Here the role of the external function is played by

the power function: sin x is raised to the third power. Differentiating
this power function with respect to the intermediate argument
(sin x), we obtain
(sin 3 a:) sin * = 3sin 2
x\
but the intermediate argument sin a: is a function of an independent

variable x\ therefore we have to multiply the obtained result by
the derivative of sin a: with respect to the independent variable x.
Thus, we obtain
yx = (sin 3 = 3 sin 2
A:) S in x (sin x)'x x cos x\
(b) y; = (lntanx) tan ,(tanx); =

y' = (5 x (cos x)'x = 5
C0S
= |2I
* In 5 (— sin x) = — 5 C0S * sin C0S
J^^ if
.;
(c) x )cos x a; In 5;
§ 2.2. Difjerentiadon of Explicit Functions 103
(d) yx = [In sin (x + 3

l)]'sin (JC , + 1)+ 1)], + x *+ l)'x =
[sin (x 3 3
1 [
= sin C0S
^ + 3x2 = 3 *' C0t + ^ 1 >
(J* + '
— 1 '
(e) = (arc sin Kl-x a

) K ([/"TT7) _
7 -x = 1 x2 ( 1
2
);
1 1
(-2*)=- * . (xÔ).
V\ — (\ —
,
x2 ) 2V~T^x* iîi^T^j?
2.2.6. Find the derivatives of the following functions:
y = (\ + 3x + 5x y = (3 — sin x)
2 3
(a) (b) )*; ;
(c) y= j/sin x + 1/cos 2 2

jc;
(d) + 2*+l + ln *;
j/2e* 5
(e) # = sin 3x + cos(x/5) + tan \f*>\

(f) # = sin(x — 5x + 1) + tan (a/x);
2
(g) y = arc cos ;
(h) */ = arctan(lnx) + ln(arctan.x:);

(i) In 2 arc tan (x/3);
(j) y= V + -\fx + V*-

x
Solution, (a) y' = 4 + 3x + 5x ( 1
2
)
3
(1 + 3x+ 5x = 2
)'
= 4(l+3x + 5x 2
)
3
(3+10x:);
l 1 1
2.2.7. Find the derivative of the function

2x
# = arc sin *
1 + X*
We
/e have
0'
l
1
2n-
2 (1 + s — 4x _2
)
a
2(1— a:
2
) 2(1— x 2 )
2x~y (i+* 2 2
y-(\-x*)*(\+x 2 |i-* 2 l(i+* 2 r
i. e.
v k
i
-i- x-
) )
frb at M<i.
f
,
= \ 2
"t|*|>i.
l-TT?
At |x;| = l the derivative is non-existent.

(a) r/ = sinh5xcosh(x/3);
(b) */= coth (tanx) — tanh (cot x)\

(c) y = arc cos (tanh x) + sinh (sin 6x);
(d) */= sinh # + cosh x 2 3 3 2
;
s\nh ax
e
(e) # ~ sinh 6a;— cosh 6a;
Solution.
(a) = (sinh 5x)' cosh ~ + sinh 5x( cosh )

=
= 5 cosh 5x cosh y + y sinh 5x sinh y
(c) y' = —-J= hx) +cosh(sin6x)(sin6x)
r
^
V — tanh 1
2
x
1/cosh 2 x
— sinh x) /cosh x
|^(cosh 2
x 2 2
+ 6 cos 6x cosh (sin 6x) = — c J h + 6 cos 6x cosh (sin 6*).

(a) y =
y 4/=-;
l—x
(b) » = U) (u(*)>0);
x 2
sin 3 * cos 2 x\
K~~
x+
(d) y^(VT^c) \
Solution, (a) Apply the method of logarithmic differentiation.
Consider, instead of y, the function
z = ln|y| = Inj/" 10^1 = ln\x\ +±\n(x*+ l)-^\n\5-x\.

Taking into account that (In u\ )' = u'ju we havey
1.2*.
|
= 1 — 24a; +125x — 3 2
+ 75
14a:
Z
x
1
3(a: 2 +1) 1
15(5 — a:) \bx(x 2 -\-

\) {b — x)
But z' = (In y\Y = y'ly>
|
whence
(a:
2
+ 1) — 24x + 125x —
3 2
14a: + 75
15x(x 2 +1)(5-a;)
Y^=~x
(b) (x) and v(x) have
Suppose the functions u derivatives in the
given domain of definition. Then the function
z=\ny = vlnu
also has a derivative in this domain, and
z' = (v In u)' = v' In u + v — .
§ 2.2. Differentiation of Explicit Functions 105
Hence, the function

y = e ln v==e z
also has a derivative in the indicated domain, and

y' = ez z' = yz'
Thus,
y' = u v {v' Inu + v^^vu"- ^ +u v \nu-v'. 1
2.2.10. Show that the function y = xe~ x ^ 2

satisfies the equation
xy'=(l— x 2 ) y.
Solution.
y> = e -x>/2 — X e -x*,2 = e -x>/2
2
(
1 _X 2).
Xy'= xe -x*/2(l_ x 2),

Hence,
xy' = y(l—x 2
).
2.2.11. Show that the function y = xe~ x satisfies the equation

xy'=-(\ — x)y.
2.2.12. Investigate the following functions for differentiability:
(a) y = arc sin (cosx); (b) y = ]/~ 1 Vl — x 2

.
o < / ,
Solution, (a) y
, . .
= — (cos
v
x)'
V 1— cos
—
2
*
— sin x
V sin 2
x
— sin*
sin x \
'
Hence, y' = — at points where sin^>0; y'=\ at points where

I
sinx<0. At points where sinxÔ, i.e. at the points x^=kn

(k = ±1, ±2, ...) the function, though continuous, is not dif-
y
ferent iable.
(b) The domain of definition of this function is the interval
— 1.
y' ,r
{
-
7=êe
•
l/^— 2
(-2x) at x=^0 and *=^± 1.
As X—+1—0 or x— — 1+0 we have y—>+oo. Let us find

out whether the derivative y' exists at the point x = 0, i.e. whe-
ther lim - - exists.

x
Ax->0 Ax
Since V 1— Ax 2 — 1 yAx 1
2
, then
X=- as Ax—> + 0,
Thus, yL (0) =/= y'+ (0), which means that the function under consi-
deration has no derivative at the point x = though it is conti- y
nuous at this point.

Note. There are cases of failure of existence of /' (x) and even of
f'+{x)and fL (x) at a given point, i.e. when the graph of the func-
tion has neither a right-, nor a left-side tangent at the given point.
For instance, the function
xs\n(\/x) at x^=0
nx >-\
f |
at x =
y
is continuous at the point x = 0> but does not have even one-sided
derivatives, since -ltl-
*
Ax
= sin-r—
Ax

(a) f(x) = s\nh(x/2) + cosh{x/2);
(b) /(*) = ln[cosh*]; (c) f (x) = 2 j/"cosh x— 1;
(d) f(x)= arc sin [tanh x]\
(e) f (*)
= Kl+sinh 4.r, 2
(f) f(x) = e
ax {zoshbx+smhbx).
2.2.14. Applying logarithmic differentiation find the derivatives

of the following functions:
(a) ^(cos^'-^Jb) y= Vx^h- x

<
(c)
f/(x+2)* V(x+ 3)"
2.2.15.
COS 2 X
'
4- sin 2 a:
show that
f(ji/4)-3/'(*/4) = 3.
* 2* 2
satisfies the differential equation

xy' + 2y = e-**.
2C
(a) # = IncosKarcsin 3~ (x > 0);
(b) y = ârc tan j/ cos In 3 x.

§ 2.3. Successive Differentiation. Leibniz Formula 107
§ 2.3. Successive Differentiation of Explicit Functions.

Leibniz Formula
If the derivative of the (n l)th order of a function y f(x) is — =
already found, then the derivative of the nth order is determined
by the equality
y
{n)
(x) = [y {n -» (x)]
f
.
In particular, y" (x) =

[y' (*)]', y"' (x)= \y" (x)]\ and so on.
If u and v are functions differentiate n times, then for their
linear combination c x u-{-c 2 v (c ly c 2 constants) we have the following
formula:
(c x u + c v) (m = c
2 x
u (n) +c v 2
(n
\
and for their product uv the Leibniz formula (or rule)

n(n X)
(uvy a = u (n) v nu '- l >v'
+ u {n '*>xf+ +
^
{
>
+ ... +uv = %C
=
nu
- (n) k (n k)
v {k \
k
n(n ~ " {n ~ k +
wneie u {0)
where =uu, u {0)
v =vu and
diiu C =
k
o„
-2-3.
= -— £)!
1
1) '
- .As fc!(/t
are
binomial coefficients. Here are the basic formulas:

m
(x (n) =m(m—\). ..(m
(1) )
~
\)x m n — n+ .
(2) (a*Y = a \n a(a>0). In

x n
n)
particular, (e
x
)
{n)
= ex .
(3) (In^r =(-1)"-^=^.

(4) (sinx) = sin(jt + njt/2).
(n)
(5) (cos*) = cos(* + wr/2).

(n>
2.3.1. Find the derivatives of the nth order of the following

functions:
(a) y=\nx\ (b) y= e kx \
(c) y = s'mx\ (d) = sin 5a: cos 2x;
*/
(e) # = sin xcosx; (f) sin 3x cos 2 a:; (g) # = In (x + x— 2). 2
(a) ^ = i = x-i; /=(-l)x-; iT'=l-2^- 3 ;
(-!)*-! (/I- 1)1

y
(4) --=— 1-2-3a:" 4 ; */
(/2)
=(— l)"" (n— 1) 1
I jt
(c) y' =cos a;= sin (# + n/2);

= cos (* + n/2) = sin (a: + 2ji/2).
In general, if we assume that for a given n=k
y
{k)
=sm +
then it will turn out that
0<* + i>=COS U + k^ :S1I1 (k+l)-
Whence, by virtue of mathematical induction we conclude that for

any natural n
(d) y=- sin 5x cos 2x =y [Vin 7x + sin 3.x:

j
Therefore
y
m = J. \jn sin (7x
+ n^+ 3" sin ^3x + n^) ]
2x+l
y'<
(g) +a:— 2
x2
'
To simplify the computations let us transform the obtained

function:
2x+ 1 (jc+2) + (jc-1) 1
:(^_1)-i
+ (x + 2)- 1
y '
x2 + x— 2 (x— l)(* + 2) -1
1
x +2 .
Whence
/=-l(*-l)- -l(* + 2)- 2 2
;
i/" = l.2(x— 1)~ + 1-2(a: + 2)-

3
^») = (_i)»-i(n— 1)! (*_l)-» + (* + 2)-»] =

= — l)"" — 1)
(
1
(AZ !
1 1
{x-\)» 1
(x + 2)»
2.3.2. #
v = cx
^4^;
«
find
-f
Solution. Transform the given expression in the following way:
ax -\-b be — ad be — ad
= ex-\- d c c(cx-\-d)
[CX + d)~K
Whence
— ad
y' = (-i) be c (cx + d)~
y" = (-l)(-2) b-^C*(CX + d)-\
y'" = (- 1) (-2) (-3) C3 (CJC + ^ 4>
/
<»)= = (_l)»n!*E_£^ C "( w + d)-«»+i) =
§ 2.3. Successive Differentiation. Leibniz Formula 109
2.3.3. y = x/(x 2 — 1); find y

(n)
.
Solution. Transform the given expression
_ x _ i r_j .
i_i
y x2 - 1 2 U+l" "^— lj 1 •
therefore (see Problem 2.3.2):

n n\
y
{n)
= (—\) +1 1
(x-
2.3.4. Using the Leibniz formula, find the derivatives of the

indicated orders for the following functions:
(a) y=x 2
smx\ find # (25) ;
(b) y = e x (x
—l); find 2
*/
(24)
;
(c) y = x
e° s'm$x\ find y
(n)
.
Solution, (a) y (2b) = (sin x-x 2

Y
2b) = (sin *)< 2B)
x2 + 25 (sin jk)
(24) 2
(a: )' +
+ ^^(sin #) (23) 2
(jk )", since the subsequent summands equal zero.
Therefore
*/
(25)
=x l
sin (x + 25 y) + 50a: sin (x + 24 ~ ) + 600 sin + 23 y ) =
= (a:
2 — 600) cos # + 50* sin x.
2.3.5. Compute the value of the nih derivative of the function
3 2
u = .,
* '
c at the point x = 0.
Solution. By hypothesis we have y(x)(x 2 — 2* + 5) = 3x + 2. Let
us differentiate this identity n times using the Leibniz formula;
then (for n^ we obtain
2)
-
y
n
(x) (x
2
— 2x + 5) + ny< ~ n l)
(x) (2x—2) + n(n {)
y<»-» (*). 2 = 0.
Putting # = 0, we have
W n)
(0) — 2ny {n ~ 1)
(0y+ n(n—l) y
(n ~ 2)
(0) - 0.
Whence
y<»> (0)=±ny<»-» (0)-^^^~ 2
>
(0 ).
We
have obtained a recurrence relation for determining the nth
derivative at the point x = 0(n^ 2). The values y (0) and y'(0)
are found immediately: (0) = 2/5;
110 Ch. II. Differential ion of Fuw turns
W= -3x
2
>i -4x-f
\
19 ,
n 19
(0)==
y (*-2*+5)' ;
^ 25'
Then, successively putting n = 2, 3, 4, find the values of

the derivatives of higher orders with the aid of the recurrence
relation.
For example,
Z
y 5
" '
25 5 5 125
•
y
"'(C)\ —L ^ 56 3»2
"
19
~
234
625'
5 125 5 25
2.3.6. Find the derivatives of the second order of the following

functions:
(a) y = xVT+7*; (b) y = ^M^; (c) y=e~*.
y = cfi 2x + c xe2x + ex 2
.
Show that this function satisfies the equation
y" — \y' + \y=ex .
2.3.8. Using the Leibniz formula give the derivatives of the in-
dicated orders for the following functions:
(20)
(a) y=x*s'mx; find */ ;
y = e~ smx;
x
(b) find y"'\
(c) = ex (3x — 4);
y 2
find y
(n)
;
(d) y= (1— x cos 2

) a:; find y
(2n)
.
2.3.9. Using the expansion into a linear combination of simpler

functions find the derivatives of the 100th order of the functions:
/ \ 1 /u\ l+#
y = x n [c t
cos (In x) +c 2
sin (In a-)]
(c l9 c2 , n constants) satisfies the equation
x 2 y" + - 2n) xy' + +

( 1 (
1 ai
2
) y = 0.
2.3.11. Prove that if /(a:) has a derivative of the nth order, then
[/ (ax + b)] (ni

= a'f n > (ax + b).
§ 2.4. Inverse, Implicit and Parametrically Represented Functions 111
§ 2.4. Differentiation of Inverse, Implicit and Parametri-

cally Represented Functions
1. The Derivative an
Inverse Function. If a differentiable
of
function y = f(x) t
a <x<b has a single- valued continuous inverse
function x = g(y) and y'x ¥=0 then there exists also
1
Xy , .
yx
For the derivative of the second order we have
Y" yxx
2. The Derivative of an Implicit Function. If a differentiable

function y = y(x) satisfies the equation F (x, y) = 0> then we have
to differentiate it with respect to x, considering y as a function
of x, and solve the obtained equation j^F(x y y) = with respect
to y'x . To find y"xx the equation should be twice differentiated with
respect to x y
and so on.
3. The Derivative of a Function Represented Parametrically. If
the system of equations
* = <P(0. = *(O. <*<'<P,

where <p(t) and if (t) are differentiable functions and q/(/)=^=0,
defines y single-valued
as a continuous function of then there
exists a derivative and
,
= q't (t)
=
The derivatives of higher orders are computed successively:
y'
xx = -^, y£ = ^>
{
and so on.
xt Xf
In particular, for the second derivative the following formula is
true:
- x't ytt—x'tt y't
(xt)*
2.4.1. For the function

y = 2x' + 3x + x;
i b
(a) find x'y ;
(b) y = 3x — (cos#)/2; find x"yy ;
(c) y= x + e x ; find xyy .
We x = Sx + +
2
Solution, (a) have y' 15x 4 1, hence,
l 1
yx
6x 2 +15x 4 +l
(c) #i=l + e*. y"xx = ex ,

hence,
1 „ ex
2.4.2.Using the rule for differentiation of an inverse function,

find the derivative yx for the following functions:
(a) y=i/x\ (b) # = arc sin/*; (c) y = In Vl + x
2
.
Solution, (a) The inverse function x = y s

has the derivative
xy = 3y 2 Hence, .
, 1 1 1
(c) At x > the inverse function a:=[/^— 1 has the derivative

xy =
f
e
2
y/Ve y—l. Hence,
2
yx = — 1
=^
Ve~iy~—i
- - x
2.4.3. For each of the following functions represented paramet-

rically find the derivative of the first order of y with respect to x:
(a) x = a(t sin/), — y = a(l — cos/);

(b) x = ks'mt sin&/, — y = kcost+coskt;
(c) Jt = 21ncot/, */= tan/ + cot /;
(d) x = ec\ y = e~ ct
.
Solution, (a) Find the derivatives of x and y with respect to the

parameter /:
x't = a (I — cos /); y\ = a sin /.
Whence
dy
~
dx
=—
a
a sin
— cos 7r = cot 172(1
i
/ , /
(/
v =7^
7 2fcri).
7
.
n , x
, . — 2 cosec 4
2
^
^ ~ cot! sin~27
*
dy
-f = sec
ctf
— cosec =
9
2 ,
/
9
2
,
/
4- cos 2t
.
sin 2 2/
20 .
•
d#
dx
4 cos 2/ sin 2/
4 sin 2 2t
^ V
, kn
2
2.4.4. The functions are defined parametrically:
(a) (x = acos 3
/, (b) jx=
fx=/ + 3/+l, /
3
;
y = bsm*t\ t* — 3t + U
\y = t*-3t l;
(c) ( x = a (cos + sin / / /), (d) fx = e* cos/,
y = a(s\n — /cos/); t
I
= ^sin
t/ /.
Find for them the second derivative of y with respect to x.

§ 2.4. Inverse, Implicit and Parametrically Represented Functions 113
Solution, (a) First find y'x .
= 3b sin
y'
t
2
/ cos t\ x\ — — 3a cos 2
1 sin t;
2
3b sin
^= - = -a
/ cos / b , . / . .
/rt « . . . ji \
3acos^sin/
ta "^ (' =*<2*+ 1) T ).
Then we shall find y"^ using the formula
u =^ xt
where
= ~acos 2
t'
Whence
„ b b
a cos 2 1 ( — 3a cos sin 2
/ /) 3a 2 cos 4 1 sin T
(d) = cos t
— e 1 sin / — e (cos — sin
x
/ t);
y
f
t
=e l
sin t + cos < = e' (cos + sin / /);
,
cos / + sin /
#
^* cos / — sin t
*
/ cos / + sin A'

«
= = \c°s / — sin ///
= 2
x\ e* (cos / — sin /) e t (cos / _ s in /) 3
'
2.4.5. Find y'

xxx :
(a) y = t
3
;
(b) x = sec/; #=tan/.
Solution, (a) First find
= y't
= 3t 2
y
whence
yx = — Wle- = — 3eH\ %
Then find the second derivative
6
xt
And finally, find the third derivative
= = 3g2 ' [2(/ 2 + 2Q + 2/ + 2] = 6g3f (/2

+ 3/ +
xt —e~ t
2.4.6. Find the derivative yx of the following implicit functions:

(a) x* + + tfy y* = 0; +
(b) ln* e-""=c;
(c) jk
2
+ # — 4x— 10*/ + 4-0;
2
(d) x
M
+ tf
'»
Solution, (a) Differentiate with respect to x 9

considering y as
a function of x; we get:
3x 2 + 2xy + x*y' +2yy = 0. f

Solving this equation with respect to y' find
,
3x 2
+ 2xy
2.4.7. Find yxx if:
(a) arc tan y y — +x = 0; (b) e

x — ey =y — x\
(c) x+y = e*~y.
Solution, (a) Differentiate with respect to x y
considering y as a
function of x and determine y'\
_X__^ + =0 l , whence y' = -±£ = y-> +

X
1.
Differentiate once again with respect to x:
Substituting the value of y' thus found, we finally get
2(l+y2 )
yXX yb
2.4.8. Find the value of y" at the point x=l if
a;
3 — 2x + + */— 5-0 2
*/
2
5.x: and y\ x=x = 1.
Solution. Differentiating with respect to x y

we find that
3* 2
— Axy — Ax yy' + 5 + y' =
2 2
0.
Putting x = 1 and = obtain the value of at x=\:

3 _4_4^ + 5 + ^_0; ^' = 4/3.
Differentiate once again with respect to x:
6x — 4y — 8xyy' — 8xyy' — \x
2 2 y' 2
— \x yy" + y" = 2
0.
Putting x=l; y^=l and t/' = 4/3, find the value y" at #=-. 1:
a
6 — 4— y64 — T
A 64
— 3^/
„ n
=0, y
„
- — 827.
Q 22
2.4.9. Find y' for the following implicit functions:

x
(a) x + V xy + y = a; (b) arc tan (y/x) = In V + y 2 2

\
(c) e*sin# — cos * = 0;

(d) ey + xyê; find at the point (0, 1).
2.4.10. Find y xx of the following implicit functions:

(a) y =x +
arc tan*/;
(b) x2 + 5xy + y 2
— 2x + y— 6 = 0; find at the point (1, 1).
2.4.11. For each of the following functions represented parame-

trically find the indicated derivatives:
sin c COS
(a) X
(i /
y= /
find y'x ;
1 +6 cos t
'
1
-f b cos I
(b) X= ln(l + f»), y= t — arc tan t; find y'x ;
X = + 2, y=
3
(c) /* t /3-t; find yxx ;
(d) X = e-*\ y=- arc tan (2/+ 1); find y'x \

= 4tan y = a sin t + b cos
2
(e) X (//2), t; find y'x ;
(0 X -= arc sin (t* — 1), y = arc cos 2t;

- find y'x ;
(g) X = arc sin t, y = find y"xx .
2.4.12. Show that the function y f(x), defined by the parametric =

equations x =e t
s'mt i
=
y e cost, satisfies the relation y"(x
t
yf + =
= 2{xy'-y).
§ 2.5. Applications of the Derivative

The equation of a line tangent to the curve of a differentiate
function y = y(x) at a point M(x 0i y ), where y =y(x )> has the
form
y—y Q
=y' W (*— *<>)•
A straight line passing through the point of contact perpendicu-

larly to the tangent line is called the normal to the curve. The
equation of the normal at the
point M will be ^ i
y—yo = —77
y' (*o)
(x # ),
y' (x )¥>o.
The segments AT, AN are
called subtangent and the
the
X
subnormal, respectively; and the o
lengths and MT are the MN Fig. 36
so-called segment of the tangent
and the segment of the normal,
respectively (see Fig. 36). The lengths of the four indicated segments
are expressed by the following formulas:
AT = _y_
y'
AN: MT =
MN = \y\V\ + {y') 2
.
2.5.1. Write the equations of the tangent line and the normal:
(a) to the curve y=^x s 3a; — +2 at the point (2, 4);
(b) to the parabola y —x + 5

= 2x 2
at x = — 0.5;
(c) curve
to the y=x* + 3x — 16 2
at the points of intersection
with the parabola y = 3x 2
.
Solution, (a) Find the derivative at the point x = 2:

y' = 3x -3 2
y
*/'(2) = 9.
The equation of the tangent line has the following form:
0—4= 9{x—2) or 9x— y— 14 = 0.

The equation of the normal is of the form:
y— 4 = — y {x— 2) or x + 9y— 38-0.

(c) Solving the system of equations
y = x* + 3jk 2 — 16,
y = 3x
2
y
we shall find the points of intersection of the curves
xt = — 2, x2 = 2, yx =y = 2
12.
Now we find the derivatives at the points x =—2 and x = 2:

y' = 4a: + 6x,
3
y' (—2) = — 44, y' (2) = 44.
Therefore, the equations of the tangent lines have the form
y— 12=— 44(* + 2), y—\2 -44(^ — 2).
The equations of the normals have the form
y-\2 = ±(x + 2), y -\2 = -±(x-2).

2.5.2. Find the points on the curve y = x' 3
— 3x + 5 at which the
tangent line:
(a) is to the straight line y
parallel 2x; =—
(b)perpendicular to the straight line y
is #/9; =—
(c) forms an angle of 45° with the positive direction of the x-axis.
Solution. To find the required points we take into consideration
that at the point of tangency the slope of the tangent is equal to
the derivative y' 3x 2 = —
3 computed at this point,
(a) By the condition of parallelism
3x 2 —3 = — 2,
whence x t = — l/j/3, x 2 =l/\/~3. The required points are:
^(-1/^3,5 + 8^3/9), M, (1/1/3, 5-8^3/9).

(b) By the condition of perpendicularity
3a:
2 — 3-9,
whence ^ = — 2, x2 = 2. The required points: M — 2,
x
( 3), M 2 (2, 7).
2.5.3. Find the angles at which the following lines intersect:
(a) the straight line y and the parabola y = 4—x = 4—x 2

/2\
(b) the sinusoid y =
s\nx and the cosine curve y = cosx.
Solution, (a) Recall that the angle between two curves at the point
of their intersection is defined as the angle formed by the lines tan-
gent to these curves and drawn at this point. Find the points of
intersection of the curves by solving the system of equations
y=4 — x,
y = 4—x /2. 2
Whence
AMO, 4); Af 2 (2, 2).
Determine then the slopes of the lines tangent to the parabola at

the points l
and M 2 M :
0'(O)=O f
y'{2)=-2.
The slope of a straight line is constant for all its points; in our
case it equals —
1. Finally, determine
the angle between the two straight

lines:
tanq^ — 1; ^ = 45°;
tan q) 2 = —1+2
+ 2 ~~
1 1
3
,
<P 2
arc tan « 18.5°.
2.5.4. Prove
that the segment of
the tangent the hyperbola y
to c/x =
which is contained between the coordinate axes is bisected at the
point of tangency.
Solution. We
have y' =—
c/x 2 hence, the value of the subtangent \
for the tangent at the point (x y ) will be M ,
y_
y'
i.e. Ox = x T (Fig. 37), which completes the proof.

Whence follows a simple method of constructing a tangent to the
hyperbola y = c/x: lay off the ^-intercept OT = 2x Then MT will .
be the desired tangent.

2.5.5. Prove that the ordinate of the catenary y acosh (x/a) is =

the geometric mean of the length of the normal and the quantity a.
Solution. Compute the length of the normal. Since
y' = sinh (x/a),
the length of the normal will be
MN = \y\ V l+(y') = y V 1+sinh % 2

(x/a) = y cosh (x/a) == y 2
/a t
whence y =a-MN, and y — Y a- MN

2 ',
2.5.6. Find the slope of the tangent to the curve
( x = + 3t — 8,
t
2
y = 2t — 2t — 5
2
\
at the point M
(2, —1).
Solution. First determine the value of t corresponding to the gi-
ven values of x and y. This value must simultaneously satisfy the
two equations
|
t* + 3t—8 = 2
\ 2t
2
— 2t — 5-— 1.
The roots of the first equation are t 1 2; t 2 5 the roots of the = =—

=—
y
second equation t t = 2; t 2 1. Hence, to the given point there

corresponds the value t 2. =
Now determine the value of the deri-
vative at the point M:
y =2
== == = T'
' '*
(il)/=2 (27 + 3)^2
And so, the slope of the tangent at the point M (2, —1) is equal
to 6/7.
2.5.7. Prove that the tangent to the lemniscate p aKcos26at =

the point corresponding to the value O ji/6 is parallel to the x-axis. =
Solution. Write in the parametric form the equation of the lem-
niscate:
= p cos = a j/"cos 26 cos 0,

x
y = p sin = a Yôs 20 sin 0.
Whence
a cos 6 sin 20 ,r 7^
xq = r
~ ay n
.
cos 20 sin 0,
^cos26
y {)
= a sin 6 sin 26 -./•
^ cos 0,n
+ a V cos 20
.
V cos 26
Xu(n/6) = -aY2, 0e(jx/6)-O.

y (Ji/b)
Thus, the slope k-=—Q = 0. Consequently, the line tangent to
*e (n/6)
the lemniscate at the point with o
= ji/6 and p = aV cos 20 =o
= a/V 2 is parallel to the x-axis.
2.5.8. Find the equations of the tangent and the normal to the
following curves:
(a) 4a;
3
— 3xy + Sx — 5xy — 8y + 9x+ 14=0 at
2 2 2
the point (—2, 3);
(b) x b
+ y — 2xy = at the point
b
(1, 1).
Solution, (a) Differentiate the implicit function:
1 2x 2 — 3y — 6xyy' +
2 — 5y — 5xy — 6yy' + 9 =
1 2x ' 1 0.
Substitute the coordinates the point M —

of ( 2, 3):
48 — 27 + 36#' — 24 — 5 + 0y' — +9= 1 1 48*/' 0;

whence
£' = -9/2.
Thus the equation of the tangent line is
»-3=-y(* + 2)
and the equation of the normal
y-3-=±-(x + 2).
2.5.9. Through the point (2, 0), which does not belong to the
curve # = Jt\ draw tangents to the latter.
Solution. Let (x Qy x* ) be the point of tangency; then the equation
of the tangent will be of the form:
y—4=y' (x {x—x ) )
y—xi = 4x (x—x 3
).
By hypothesis the desired tangent line passes through the point

(2, 0), hence, the coordinates of this point satisfy the equation of
the tangent line:
— %t — 4xq (2 x )\ 3x1 — = 0,
whence x x = 0; =
8/3. Thus, there are two points of tangency:
AM0, 7W 2 (8/3, 4096/81).
0),
Accordingly, the equations of the tangent lines will be
A 4096 2048 / 8 \
2.5.10. f (x) 3x b I5x 3 = — +

5x—7. Find out at which of the points x
the rate of change of the function is minimal.
Solution. The rate of change of a function at a certain point is
equal to the derivative of the function at this point
/'(*)= 15a:
4
— 45a: + 5 = 15 [(x — 1/2) + 1/12].
2 2 2
The minimum value of f (x) is attained at x ±l/j/*2. Hence the =

minimum rate of change of the function f (x) is at the point
r
x =± 1 /y 2 and equals 5/4.
2.5.11. A
point is in motion along a cubic parabola I2y=x 3 .
Which coordinates changes faster?

of its
Solution. Differentiating both members of the given equation with
respect to t we get the relation between the rates of change of the
coordinates:
\2y't = 3x 2
-x't
or
yt _x 2
(1) at —
2 < x < 2 the ratio y\\x\ is less than unity, i.e. the rate
of change of the ordinate is less than that of the abscissa;
(2) at x = ±2 the ratio y\\x\ is equal to unity, i.e. at these
points the rates of change of the coordinates are equal;
(3) at jc<-t-2 or x > 2 the ratio y\\x\ is greater than unity, i.e.
the rate of change of the ordinate exceeds that of the abscissa.
2.5.12. A body of mass 6g is in rectilinear motion according to

the law s 1 =— +
ln(/+ l) (t l)
3
+ +
(s is in centimetres and /, in
seconds). Find the kinetic energy (mv 2 /2) of the body one second
after it begins to move.
Solution. The velocity of motion is equal to the time derivative
of the distance:
»(0 = s;= r pr + 3 ('+ 1

)
a
-
Therefore
*<l) = 12l and ^ = |(l2i-) 2

-=4684(erg).
2.5.13. The velocity of rectilinear motion of a body is proporti-

onal to the square root of the distance covered (s), (as, for example,
in free fall of a body). Prove that the body moves under the action
of a constant force.
Solution. By hypothesis we have
v = s'i=ay s (a = const);
whence
s"ti =v' t
=a jy=- s = a t
1
12.
But according to Newton's law the force

F = ks tt (k = const).
Hence,
F^ka 2
/2 = const.
2.5.14. A pulled to the bank by means of a rope which
raft is
is wound on a drum,
at a rate of 3 m/min. Determine the speed
of the raft at the moment when it is 25 distant from the bank m
if the drum is situated on the bank 4 above water level. m
Solution. Let s denote the length of the rope between the drum
and the raft and x the distance from the raft to the bank. By
hypothesis
s
2 2
42 =x + .
Differentiating this relation with respect to /, find the relation-

ship between their speeds:
2ss/ = 2xX{,
whence
s ,
x t-T St '
Taking into consideration that

s; = 3; x = 25; s = /25 + 4 « 25.3, 2 2
we obtain
^25 a +4a
^;==
25
.3» 3.03 (m/min).
2.5.15. (a) Find the slope of the tangent to the cubic parabola
y=x* at the point x = ]/~3/3.
(b) Write the equations
of the tangents to the curve 1/(1 x2 ) y= +
at the points of its intersection with the hyperbola y=l/(x+l).
(c) Write the equation of the normal to the parabola y #2 -f-4x-f 1 =
perpendicular to the line joining the origin of coordinates with the
vertex of the parabola.
(d) At what angle does the curve y = ex intersect the y-ax\s?
2.5.16. The velocity of a body in rectilinear motion is determi-

ned by the formula v = 3t-\-t 2 . What acceleration will the body
have 4 seconds after the start?
2.5.17. The law of rectilinear motion of a body with a mass of

100 kg is s 2/ 2 = +
3/+ 1. Determine the kinetic energy (mv 2 [2) of
the body 5 seconds after the start.
2.5.18. Show that if the law of motion of a body is s ae be = 1

+
then its acceleration is numerically equal to the distance covered.
2.5.19. A body is thrown vertically with an initial velocity of

a m/sec. What altitude will it reach in t seconds? Find the velocity
of the body. In how many seconds and at what distance from the
ground will the body reach the highest point?
2.5.20. Artificial satellites move round

the Earth in elliptical
orbits. The distance r of a satellite from the centre of the Earth
as a function of time t can be approximately expressed by the fo-
llowing equation:
r =a ^
1 — e cos M — y (cos 2M — 1)
where M = ^-(t — t
n)
= time parameter
t
a = semi-major axis of the orbit

e = eccentricity of the orbit
P = period of orbiting
/„ = time of passing the perigee
1
ty the satellite.
Here a, 8, P and /„ are constants.

Find the rate of change in the distance r from the satellite to the
centre of the Earth (i.e. find the so-called radial velocity of the
satellite).
§ 2.6. The Differential of a Function.

Application to Approximate Computations
If the increment Ay of the function y = f(x) can be expressed as:
Ay =f (x + Ax) — f (x) = A(x) Ax + a (x, Ax) Ax,

where
lim a (x, Ax) = 0,
Ax -> o
then such a function is called differentiable at the point x. The

principal linear part of this increment A(x)Ax is called the diffe-
rential and is denoted df (x) or dy. By definition, dx Ax. =
For the differential of the function y f(x) to exist it is nece- =
ssary and sufficient that there exist a finite derivative y' A(x). =
The differential of a function can be written in the following way:
dy = y' dx = f {x)dx.
1
The perigee of the satellite orbit is the shortest distance from the sate-
llite to the centre of the Earth.
§ 2.6. Differential of a Function 123
For a composite function y = f(u) y

u = y(x) the differential is
retained in the form

dy = f (u) da
(the invariance of the form of the differential).
With an accuracy up to infinitesimals of a higher order than Ax
the approximate Ay^dy formula takes place. Only for a linear
function y = ax b do we have Ay=dy.
+
Differentials of higherorders of the function y = f(x) are succes-
sively determined in the following way:
d 2 y=d(dy); d 3 y=d(d 2 y) y
dny = d(d"~^y).
If y = f(x) and x is an independent variable, then
d 2
y = y" {dx)
2
\
d3y = y" (dx)\ . .
. ,
dny = y<» {dx)\
But if where u = cp (x), then d 2 y = f" (u) du 2 2
-\-f (u) d u, and
so on.
2.6.1. Find the differential of the function
y = In +e + arc tan e bx
10x
( 1 )
.
Calculate dy at x-= 0; d^^0.2.
_ r (i-fgiox)^ e *xy -I
_ 5g 5x (2g5-y— ^
— [
i+«,io*
(
l+gio*J — 1+gio*
1)
Substituting * = and d.x: = 0.2, we get

5
dtf|x=0; djc=0.2 = y0.2=0.5.
2.6.2. Find the increment and the differential of the function
y = 3x* + x—\
at the point x= \ at Ax = 0.1.
Find the absolute and relative errors allowed when replacing the
increment of the function with its differential.
Solution.
Ay = [3 (x + Ax) + (x + Ax) — —
3
1 ] (3a:
3
+ x— = 1 )
= 9xAx + 9x Ax + 3 Ax + Ax,
2 2 3
dy = (9x + 2
I) Ax.
Whence
Ay—dy = 9* A* + 3Ax 2 3
.
At a: = 1 and Ax = 0.1 we get
Ay—dy = 0.09 + 0.003 - 0.093,

dy=l; Ay = 1.093.
The absolute error |

Ay — dy = 0.093, the relative error
|
ky—dy
&y
°- 093
1.093'
0.085 or 8.5%.
2.6.3. Calculate approximately the increment of the function
y =x s
— 7x + 8
2
as x changes from 5 to 5.01.
2.6.4. Using the concept of the differential, find the approximate

value of the function
Solution. Notice that from = y(x+Ax) — y(x)

Ay we get
y(x + Ax) = y(x) + Ay,

or, putting Ay « dy,
y(x+Ax)^y(x) + dy.
In our problem let us put x = and A# = 0.15. Then
y 5 V \2 — x) (2 + x) 2 '
y'(0)= — i, d^ = — -g-.0.15 = — 0.03.

Hence,
y (0. 15) « y (0) + dy = — 0.03 - 0.97.1
The true value of y (0.15) = 0.9702 (accurate to 10" 4 ).
2.6.5. Find the approximate value of:
cos31°;
(a) (b) log 10.21; (c) j/33; (d) cot45°10'.
Solution, (a) In solving this problem we shall use the formula (*)
of the preceding problem. Putting x jt/6, rc/180, we compute: = =
^(^)
/x V3
= cos-g-n = -g-;
y (
x) = _ S in-g= — — ;
cos 3 1° = cos + ^) « 1 iL = 0.85 1
(c) Put a; = 32; A#=l. By formula (*) we get
^33«^3l + (t/ x); =3 2-l=2 + ^= = 2H-l = 2.0125.

§ 2.6. Differential of a Function 125
2.6.6. All faces of a copper cube with 5-cm sides were uniformly
ground down. As the weight of the cube was reduced by
a result
0.96 g. Knowing the specific weight of copper (8) find the reduction
in the cube size, i.e. the amount by which its side was reduced.
Solution. The volume of the cube u==x 3 where x is the length ,
of the side. The volume is equal to the weight divided by the den-
sity: v =
p/d\ the change in cube's volume Av 0.96/8 0.12 (cm 3 ). = =
Since Av approximately equals dv and taking into consideration that
tiv = 3x 2
dx we shall have 0.12 = 3x5 2
x Ax, whence
Ax = ^L= 0.0016 cm.
Thus, the side of the cube was reduced by 0.0016 cm.
2.6.7. Find the expressions for determining the absolute errors in

the following functions through the absolute errors in their argu-
ments:
(a) y = lnx; (b) y = logx;
(c) y = s'mx (0 < < ji/2); jk (d) = tan* (0 <x< ji/2);
(e) = log(sin#) (0 < x < ji/2);

(f) = log(tanx) (0<x<ji/2).
Solution. If the function f (x) is differentiate at a point x and
the absolute error of the argument A* is sufficiently small, then
the absolute error in the function y can be expressed by the number
= I y'x I
A*.
(a) Ay = \(\nx)' \ X AX = — , i.e. the absolute error of a natural
logarithm is equal to the relative error in its argument.
(b) Ay = {\ogx)' A X = ™A X where M = log e = 0.43429;

,
(e) A y = [log (sin x)]' A X = M \cotx\ A x

\ |
;
(f) A^lllogaan^ÂîÂ,.
From (e) and (f) it follows that the absolute error in log tan x
is always more than that in log sin x (for the same x and A x ).
2.6.8. Find the differentials dy and d 2 y of the function
y = 4x 7a: 3,
b
— 2
+
assuming that:
(1) a: is an independent variable;
(2) x is a function of another independent variable.
Solution. By virtue of the invariance of its form the differential
of the first order dy is written identically in both cases:
dy = y' dx = (20x 4
— 1 4x) dx.
But in the first case dx is understood as the increment of the

independent variable Ax (dx = Ax), and in the second, as the diffe-
rential of x as of a function (dx may not be equal to Ax).
Since differentials of higher orders do not possess the property
2
of invariance, to find d y we have to consider the following two
cases.
(1) Let x be an independent variable; then
2
d y = y" dx 2
= (8(k 3
— 14) dx 2
.
(2) Let x be a function of some other variable. In this case

d*y = (8(k 3
— 14) dx + (2(k — 14*) d
2 4 2
x.
2.6.9. Find differentials of higher orders (x an independent va-

riable):
(a) */= 4"* find d y\ 2

;
2
y = V~\n x — 4; d 2 y\
2
(b) find
(c) y = sm x\
2
find d y.
3
2.6.10. y = \nj^— I ; find d2 y if: (a) x is an independent variable,

(b) x is a function of another variable. Consider the particular
case when x =
tant.
2.6.11. The volume V of a sphere of radius r is equal to

4 3
yjir . Find the increment and differential of the volume and
explain their geometrical meaning.
2.6.12. The law of the free fall of a material point is s gt 2 /2. =

Find the increment and differential of the distance at a moment t
and elucidate their mechanical meaning.
2.7.1. Given the functions: (a) f(x) \x\ and (b) = cp (a;) = |
x3 [.
Do derivatives of these functions exist at the point ^ = 0?

2.7.2. that the curve y = e\*\ cannot have a tangent line
Show
at the x = 0. What is the angle between the one-sided tan-
point
gents to this curve at the indicated point?
f(x) = \x—a\y(x),
where q(x) is a continuous function and (p(a)=^=0, has no deriva-
tive at the point x =
a. Find the one-sided derivatives /1(a) and
/; (a).
j
x2 s\n(l/x) at x^=0,
f
\ at * = 0.
Use this example to show that the derivative of a continuous
function is not always a continuous function.
2.7.5. Let
^ ^ / -^
2
>
if a: ^x ,
\ ax +b y
if a: > a* .
Find the coefficients a and b at which the function is continuous

and has a derivative at the point x .
2.7.6. By differentiating the formula cos 3a: = cos 3

x — 3 cos x sin 2
x
deduce the formula sin 3a: = 3 cos 2
x sin x — sin 3
a:.
2.7.7. From the formula for the sum of the geometric progression
deduce the formulas for the following sums:
(a) \ + 2x + 3x + +nx n ~ 2
. . .
1 '
(b) \ + 2 x+3 x +
2 2
+n xn -\2 2
. . .
2
2.7.8. Prove the identity

s' n2nx
cosx + cos3*+ . . . +cos(2tt— 1)a:=
1 v 7 ,
x^-kn
^
2 sin x •
and deduce from it the formula for the sum

sin x + 3 sin 3x + . . .
+ (2n — 1) sin (2n — l)x.
2.7.9. Find y' if:
(a) #-/(sin 2 *) + /(cos 2

x); (b) y = f(e*)e'™;
(c) y = logcp {x) (x) (cp (x) > 0; (x) > 0).
2.7.10. Is it reasonable to assert that the product F (x) f (x) g(x)

=
has no derivative at the point a: a: if: =
(a) the function f (x) has a derivative at the point x 0l and the
function <p(x) has no derivative at this point?
(b) neither function has a derivative at the point x ?
Consider the examples: (1) f(x)=x, g(x) \x\; =
(2) /(x) = |*|. g(x)= \x\.
Is it reasonable to assert that the sum F (x) = f (x) + g (x) has no

derivative at the point x x if: =
(c) the function f (x) has a derivative at the point x , and the
function g(x) has no derivative at this point?
(d) neither function has a derivative at the point x ?
2.7.11. Prove that the derivative of a differentiate even function

is an odd function, and the derivative of an odd function is an
even function. Give a geometric explanation to these facts.
2.7.12. Prove that the derivative of a periodic function with
period T is a periodic function with period T.
2 J A3. Find F' (x) if
X X* xd
F(x) = 1 2x 3x2
2 6x
2.7.14. Find the derivative of the function y = x\x\. Sketch the
graphs of the given function and its derivative.
2.7.15. Suppose we have a composite function y f(u) where = 9
u^=(p(x). Among what points should we look for points at which

the composite function may have no derivative?
Does the composite function always have no derivative at these
points? Consider the function y u2 u \x\. = y
=
2.7.16. Find y" for the following functions:
x 2 sin (l/x), x =7^ 0,

(a) y = \x*\\ (b) y =
at * = 0.
Is there #"(0)?
2.7.17. (a) f(x)=x n show ; that

(2)
/'(l) d)
. /
n\
= 2n .
n~1 l/x
(b) f (x)=x e ; show that
(n=l, 2, ...).
2.7.18. y=x e~ x/a2

; show that
n
(—\) n(n- i)
~2
an
2.7.19. Show
that the function y atcs\nx satisfies the relation =
(\—x 2 )y" = xy Find y in) (0) (n
f
. 2) by applying the Leibniz for- >
mula to both members of this identity.
2.7.20. Prove that the Chebyshev polynomials

satisfy the equation
(1 -x 2
)
r; (x)-xT'n (x) +n 2
T n (x) = 0.
2
2.7.21. The derivative of the nth order of the function e'* has
the form
(e-x
2
y>»= e - x2 H n (*),
where H n (x) is a polynomial of degree n called the Chebyshev-

Hermite polynomial.
Prove that the recurrence relation
H n+l (x)-2xH n (x) + 2nH n _ (x)^0 1

2, ...)
is valid.
2.7.22. Show that there exists a single-valued function y=y(x)

defined by the equation y* 3y + =
x and find its derivative y'x
y
.
2.7.23. Single out the single-valued continuous branches of the

inverse function x = x(y) and find their derivatives if y 2x 2 4
= — .
2.7.24. u = ^r\n^
2
r^-\
\—v
check the relation ^^
dvdu
= 1.
2.7.25. Inverse trigonometric functions are continuous at all

points of the domain of definition. Do they have a finite deriva-
tive at all points of the domain? Indicate the points at which the
following functions have no finite derivative:
(a) ^ = arccos^y?-; (b) */ = arc sin
2.7.26. Show
that the function y y(x), defined parametrically: =
x = 2t — \t t
2
, y = + t\t\
is differentiate at f but its deriva-
9
=
tive cannot be found by the usual formula.
2.7.27. Determine the parameters a, b, c in the equation of the

parabola y ax2= bx + +
c so that it becomes tangent to the straight
line y =
x at the point 1 and passes through the point ( 1, 0). —
Prove that the
2.7.28. curves y 1 =
f(x) (/ (x) 0) and y2 > =
= f(x)sinax, where f (x) is a differentiate function, are tan-
gent to each other at the common points.
2.7.29. Show any point

that (x 09 y ) of the equilateral
for M
hyperbola x2 —y = a
2 2
the segment of the normal from the point
M to the point of intersection with the abscissa is equal to the
radius vector of the point M.
2.7.30. Show that any position of the generating circle the

for
tangent line and normal to the cycloid x=a(t
the sin/), —
y = a(l— cost) pass through the highest (at, 2a) and the lowest
[at, 0) points of the circle, respectively.
5—3148
2.7.31. Show that two cardioids p = a (1 -f-coscp) and p =

= a (I — coscp) intersect at right angles.
2.7.32. Let y = f(u), where u = y(x). Prove the validity of the

equality
d*y = \"' (u) du + 3/" (u) du d*u + f (u) d*u.
3
2.7.33. Let y = f(x), where x = <p(t); the functions f (x) and

cp(0 are twice differentiable and dx=^=0. Prove that
„
_
~
d 2 ydx—dyd 2 x
»
dx*
where the differentials forming the right member of the relation

are differentials with respect to the variable t.
2.7.34. How will the expression
<«-*> S- 4?+ y
be transformed (where y is a twice differentiable function of x) if
we introduce a new independent variable putting x cost? =

2.7.35. In determining an electric current by means of a tangent
galvanometer use is made of the formula
/ =k tan (p,
where / = current
k = factor of proportionality (depending on the instrument)
cp= angle of pointer deflection.
Determine the relative error of the result which depends on the
inaccuracy in reading the angle cp. At what position of the pointer
can one obtain the most reliable results?
Chapter
3
APPLICATION OF DIFFERENTIAL
CALCULUS TO INVESTIGATION
OF FUNCTIONS
§ 3.1. Basic Theorems on Differ entiable Functions
Fermat's Theorem. Let a function y =

f(x) be defined on a cert-
ain interval and have a maximum or a minimum value at an in-
terior point x Q of the interval.
=
If there exists a derivative f (x ) at the point x 0y then f'(x ) 0.
Rolled Theorem. If a function f (x) is continuous in the interval
[a, 6], has a finite derivative at all interior points of this interval,
and f(a) =
f(b), then inside [a, b] there exists a point b)
such that f'{l) 0. =
Lagrange's Theorem. If a function f(x) is continuous in the in-
terval [a, b] and has a finite derivative at all interior points of
the interval, then there exists a point £g(a, b) such that
f(b)-f(a) = (b-a)f®.
Test for the Constancy of a Function. If at all points of a cer-
tain interval /'(x) =
0, then the function f (x) preserves a constant
value within this interval.
Cauchy's Theorem. Let (p (x) and ip(x) be two functions continu-
ous in the interval [a, b] and have finite derivatives at all inte-
rior points of the interval. If these derivatives do not vanish si-
multaneously and (p (a) <p (b) then there exists £€(a, b) such that
9
t|)(fr)-t|)(a) ^tK (E) '

<p(&)'— <p(a) <p'(g)
Does the function f(x)

3.1.1. 3x2 1 = —
satisfy the condition of
the Fermat theorem in the interval [1, 2]?
Solution. The given function does not satisfy the condition of the
Fermat theorem, since it increases monotonically on the interval
[1, 2], and, consequently, takes on the minimum value at x = 1 and
the maximum one at x = 2, i. e. not at interior points of the in-
terval. Therefore, the Fermat theorem is not applicable; in other
words, we cannot assert that /'(l) = /'(2) = 0. Indeed, ^ (1) =
6,
/'(2)=12.
5*
132 Ch. III. Differential Calculus: Investigation of Fund's
3.1.2. Do the following functions satisfy the conditions of the

Rolle theorem?
(a) f(x) = \-Vtf in [-1, 1];
(b) f(x) = In sin x'm [ji/6, 5ji/6];
(c) /(*)= 1-|*| in [-1, 1].
If they do not, explain why.

Solution, (a) The function is continuous in the interval [ — 1, 1];
furthermore, / ( 1) /(1) — = =
0. Thus, two conditions of the Rolle
theorem are satisfied. The derivative f'(x) = — 2/(3 y/ x) exists at
all points except # = 0.
Since this point is an interior one, the third
condition of the theorem is not satisfied. Therefore, the Rolle the-
orem is not applicable to the given function. Indeed, /'(jk)^0 in
[-1. !]•
3.1.3. Prove that the equation

3x 6 + 15*— 8-0
has only one real root.
Solution. The existence of at least one real root follows from the
fact that the polynomial f (x) 3* 5 15* = +
8 is of an odd power. —
Let us prove the uniqueness of such a root by reductio ad absur-
dum. Suppose there exist two roots x 1 <
x 2 Then in the interval .
[*!, x 2 ] the function f (x) 3x b I5x = —

8 satisfies all conditions of
the Rolle theorem: it is continuous, vanishes at the end-points and
has a derivative at all points. Consequently, at some point \,x x \ x2 < < ,
f'(t) 0.=But /'(*)= 15 (x*+ 1) 0. This >

contradiction proves that
the equation in question has only one real root.
Does the function f(x) = 3x2

3.1.4. 5 satisfy the conditions of—
the Lagrange theorem in the interval [ 2, 0]? If it does, then —
find the point g which figures in the Lagrange formula f (b) f(a) = —
= f(E) (b-a).
Solution. The function satisfies the conditions of the Lagrange
theorem, since it is continuous in the interval [ 2, 0] and has a —
finite derivative at all interior points of the interval. The point £
is found from the Lagrange formula:
-5-7 _ __ R
f> /tx
/ \±) — At _ /(Q)-/(-2)
_ ufc
0— (— 2) 2 ~" '
whence £ =— 1.
3.1.5. Apply the Lagrange formula to the function f(x)=\nx in

the interval [1, e] and find the corresponding value of £.
3.1.6. Ascertain that the functions f(x) =x 2

— 2* + 3 and g(x) =
=x 3 —
7x 2 20x + —
5 satisfy the conditions of the Cauchy theorem
in the interval [1, 4] and find the corresponding value of £.
§ 3.1. Basic Theorems on Different iable Functions 133
Solution. The given functions f (x) and g(x) are continuous eve-
rywhere, and hence, in the interval [1, 4] as well; their derivatives
= —
f (x) 2x 2 and g' (x) 3x
2
\4x = — +
20 are finite everywhere; in
addition, g' (x) does not vanish at any real value of x.
Consequently, the Cauchy formula is applicable to the given
functions:
/(4)-/(l)
= nE)
S(4)-s(l) g'iQ*
i. e.
27 —9 3g 2 — 14g + 20 y ^6^ ;
Solving the latter equation, we find two values of £:E 1 =2 and

i 2 =4.
Of these two values only ^=2 is an interior point of the interval.
3.1.7. Do the functions f(x) = ex and g (x) = p-X ^ 2
satisfy the con-
ditions of the Cauchy theorem in the interval [—3, 3]?

3.1.8. On the curve y =
x* find the point at which the tangent
line is parallel to the chord through the points A(— 1, —1) and
fl(2, 8).
Solution. In the interval [
—
1, 2], whose end-points are the abs-
cissas of the points A andB, the function y x 3 is continuous and =

has a finite derivative; therefore the Lagrange theorem is applicable.
According to this theorem there will be, on the arc AB, at least
one point M, at which the tangent is parallel to the chord AB.
Let us write the Lagrange formula for the given function:
/(2)-/(-l) = /'(g)[2-(-l)] f
or
8 + = 36
1
2
• 3;
whence
Ex 1. I,= l.
The obtained values of £ are the abscissas of the desired points

(as wethere exist two such points). Substituting
see, and S- 2 in
the equation of the curve, we find the corresponding ordinates:
yi = l'i = l; y2 =U= — l.
Thus, the required points are: y (\,

M
1) and Af 2 (— 1, —1), of which
only the former is an interior point on the arc AB.
Note. This problem can be solved without using the Lagrange the-
orem; write the equation of the chord as a straight line passing
through two given points, and then find the point on the curve at
which the tangent is parallel to the chord.
3.1.9. Taking advantage of the test for the constancy of a func-

tion, deduce the following formulas known from elementary math-
ematics:
(a) arc sinx+arc cosx = ji/2;
2
(b) sin x = (l— cos2x)/2;
(c) arc cos -r-, — 2 arc tan a; at O^Zx <oo;
defined in the interval f— 1, 1]. The derivative of the indicated

function inside this interval equals zero:
According to the test for the constancy of a function / (x) = const,

i. arcsin A:
e. +
arccosA: C = (— 1<#<1).
To determine the constant C let us put, for instance, x=0; then
=
we have jx/2 C, whence
arc sin a: + arc cos = jt/2
a: ( — <x<
1 1).
The validity of this equality at the points a:=±1 is verified

directly.
(b) Let us take the function
/ (x) = sin 2
x + y cos 2a;
defined throughout the number scale: — oo < < a; oo. The derivative
of this function is everywhere equal to zero:
f (a;) = 2 sin a: cos a; — sin = 2a; 0.
According to the test for the constancy of a function
sin 2 x + y cos 2a: = C.

To determine C put, for instance, x = 0; then we get 1/2 = C.
W herefrom
sin 2 x+y cos2x = y
§ 3.1. Basic Theorems on Differ entiable Functions 135
or
.
9
s\n 2 x = — cos
^
1 2x
•
(c) Let us introduce the function
/ (x)
== arc cos l~* — 2 arc tan x,
—X 2
determined along the entire number scale, since
\
^2
< 1.
The derivative of the function f (x) is zero for all x > 0:
p / x
~ ~~~j/7
1 —4a;
~Y 2 4x 2
=^
1 ~ 3 ( 1 * a)a +7 * 2x (
i
~|- x~)
^ y
According to the test for the constancy of a function
arc cos t— — — 2 arc tan x = C

- at x > 0.
To determine C let us put, say, x = 1, which gives C = arccos0

— 2 arc tan 1 =0.
The validity of the proved formula at x = is verified directly.
Note, At x = the function arc cos r-j —

1 -\-x
r
2
has no derivative. At
x < its derivative is
l —x 2 \' 2
arc cos -
which enables us to derive the formula
arccos -j-p^2 = —2 arc tan # (x < 0).

The latter formula can be obtained on the strength of the fact
*
that arc cos I ,
is an even function, and 2 arc tan x is an odd one.
3.1.10. As is known, (e*)'=e* for all x. Are there anymore

functions that coincide with their derivatives everywhere?
Solution. Let the function f (x) be such that f'(x)^f(x) every-
where.
Let us introduce the function
V(x) = 1g- = f(x)e-*-

The derivative of this function equals zero everywhere:
cp' (x) =f (x) e-
x — e- x f {x)
= 0.
By the test for the constancy of a function f{x)/e

x = C, whence
f(x) = Ce x.
136 Ch. 111. Differential Calculus: Investigation of FuncVs
And so, we have proved that the group of functions for which
f'(x)^f(x) is covered by the formula f(x)^Ce x .
3.1.11. Prove the inequality
arc tanx 2 — arc tan x 1 < x2 —x lt
where x 2 > xr
Solution. To the function / (x)
= arc ianx on the interval [x l9 x2 ]
apply the Lagrange formula:
arc tan x 2 — arc tan x x

— —x i)>
where x x <£<x 2
.
Since
°<T^F < and ^2 — î>0, 1
then
arc tan x — arc tan x < x — x
2 l 2 1
.
In particular, putting ^= and x 2 = x, we get
arc tan x < x (x > 0).

3.1.12. Show that the square roots of two successive natural
numbers greater than N 2 differ by less than 1/(2 A/).
Solution. To the function f(x) = Vx on the interval [n, n+1]
apply the Lagrange formula:
/(n+l)-/(n) = /^-|/"n=^,
where n <| < n+ 1
If n>/V 2
, then 1>N\ hence 1/(2 < 1/(2jV), whence
y n+ \—y n< \/(2N).
3.1.13. Using the Rolle theorem prove that the derivative f (x)
of the function
( x sin —* at x > 0,
c
f(*)
, v
= i
( at a: =
vanishes on an infinite set of points of the interval (0, l).
Solution. The function f (x) vanishes at points where
sin (n/x) ^=0, n/x^kn, x—l/k,

k= 1, 2, 3, ...
Since the function f (x) has a derivative at any interior point

of the interval [0, 1], the Rolle theorem is applicable to anyone
§ 3.1. Basic Theorems on Differ entiable Functions 137
of the intervals [1/2, 1], [1/3, 1/2], [l/(k+\) y

l/k], ... .
Consequently, inside each the intervals of the sequence, there is of

a point l k 1 /(£+')< l k
, 1/6. at which the derivative f'(l k )=0. <
And so we have shown that the derivative vanishes on an infinite
set of points (see Fig. 38).
3.1.14. The Legendre polynomial is
a polynomial defined by the following
formula (Rodrigues' formula):
2nn ! dx»
(X 1)" (Al = 0,
1, 2, ...).
Using the Rolle

Koiie theorem, prove
tneorem,
that the Legendre polynomial P n (x)
has n different real roots, all of them
found between — 1 and -f 1.
f(x) = (x 2
— \)» = (x— l)
n
(*+l)".
This function and its n 1 successi- —
ve derivatives vanish at the points Fig. 38
x= ±\ (use the Leibniz formula for

higher derivatives of the product of two functions).
It follows from f(l) f(—l) =
that inside the interval 1, 1] [. —
a point ^ can be found at which f (l l ) i.e. x l 1 will be the
r
= t
=
root of the first derivative. Now apply the Rolle theorem once again
to the function f'(x) on the intervals [ 1, gj, [| x 1]. We find — ,
that besides +1 and —1

the function f" (x) has two more roots
on the interval [ 1, —
1]. Reasoning as before, we will show that,
apart from +1 and —
1, fhe (n~l)th derivative has (n— 1) more
roots on the interval [ 1, 1], i.e. the function f —
(n ~ v
(x) has all in
all n+l roots on the interval [ 1, 1], which divide this interval —
into n parts. Applying the Rolle theorem once again, we ascertain
that the function f

in)
(x), and hence, the function P n (x)= ^j—j f
in)
(*),
has n different roots on the interval [ — 1, 1].
3.1.15. Check whether the Lagrange. formula is applicable to the

following functions:
(a) f(x)-- on [3, 4];x2
(b) f(x)-- on [1, 3];
-\nx
(c) f(x)^4x*
4r* 5x 2 + x 2 on — — [0, 1];
5
(d) f(x)=l/x* (x-\) on [-1/2, 1/2].
If it is, find the values of £ appearing in this formula.
3.1.16. Using the Lagrange theorem estimate the value ln(l-j-t?)

[38 Ch. III. Differential Calculus: Investigation of Fund's
3.1.17. Using the Lagrange formula prove the inequality
Y^rx < \n(l+x)<x at jc > 0.
§ 3.2. Evaluation of Indeterminate Forms.

L'Hospital's Rule
Indeterminate forms of the type ~. the functions (x)

I.
^ , If f
and g(x) are differentiate in a certain neighbourhood of the point a,

f
except, may be, at the point a itself, and g (x)=^0, and if
lim / (x)
= lim g(x) = or lim / (x) = lim g(x) =- oo,
x -+ a x -* a x -> a x -y a
then
,. f
lim '-)-{= lim
(X) 1- f (X)'
a
g(x) x _> a g'W
f (x)
provided the limit lim exists (U Hospital's rule). The point a /
- a 8 x
,
w
may be either finite or improper -f oo or oo. —
II. Indeterminate forms of the type O oo or oo oo are reduced —
to forms of the type -jj- or ^ by algebraic transformations.
III. Indeterminate forms of the type l 30 oo° or 0° are reduced ,
to forms of the type 0-oo by taking logarithms or by the transfor-

mation [f(x)]
, pw = evwinfw
3.2.1. Applying the L'Hospital rule, find the limits of the follo-
wing functions:
e x_ e -x.__2x. , In (1 + a:
2
(c) lim lim
)
* _ o x—smx ; (d)
\ /
^ ^ Q C
Q
os3a: — e~ x y
sin 3a:
2 1 /* — 2
/ v „ e
1171 lim ~
Incns^-*! ' (f) 75 t
2
.
Solution, (a) Here both functions f(x) = e ax e~ 2ax and —

«=ln(l+A:) are infinitesimals in the neighbourhood of zero, since
lim / (x) = 1 — = 1 0; lim g (*) = In 1 = 0.

x -> x ->
Furthermore /' (x) and exist in any neighbourhood of the

point x^=0 that does not contain the point x- —1, and
§ 3.2. Evaluation of Indeterminate Forms 139
Finally, there exists a limit of the ratio of the derivatives:
,.
Inn ^ttt
/' (x)
= ,.
lim —
ae ax +2ae- 2ax
T~n~\ —— i x
—= Q
oa.
Therefore the L'Hospital rule is applicable:
lim
e ax_ e -2ax
— j
In
——
—(\+x) r: :
lim —
ae ax +2ae~ 2ax
r-Tf—
1/(1+*)
— — = 3a. -
(*)7
x> x
/Vote. When the limit of the ratio is computed according to the

L'Hospital rule the result is usually written directly as shown in (*).
Whether the desired derivatives and limits exist is ascertained in
the course of calculation. In case the ratio of the derivatives '

,
^
again represents an indeterminate form, the L'Hospital rule should
be applied for a second time, and so on until the indeterminacy is
removed or until it becomes clear that the required limits do not
exist. Therefore, henceforward we write only the necessary transfor-
mations, leaving to the reader the task of checking whether the
conditions of their applicability are fulfilled.
(b) lim
j/T
v '
l/2
+ 2^+1
+ x+x
= lim
1/(2
3/0+25)*)
2/(3 v
r
J^2 + x)+l
,
—-=ir 4
9
»'
. v
r sin 3x
2
,. —6a: cos 3a:
2
cos (2x 2 — x)
v/
K ^ In cos (2x 2 — x) x ^ (4a:— 1) sin (2a:
2
—x)
= —6n t
. cos 3a:
2
cos (2a:
2 — a:) i .
-
lim -
lim
x - o
4a:
A
— 1
x ^ o sin (2a:
2 — x) '
The limit of the first factor is computed directly, the limit of the
second one, which represents an indeterminate form of the type -jj- is
found with the aid of the L'Hospital rule:
—6 ~ «.
lim
cos 3a:
2
cos (2a:
lim -r
2 — x) ,.
x - o 4a:
-=
—\ x ^
sin
1 — (2a:
2
a:)
= —6 —
— xlim —(4a:— n cos • j-
1
;
1)
^
,
(2a:
n 9
2
—\—r =
— x) = 6 — 1-1 • — 6.
_^
3.2.2. It is the functions x k (k

known 0); that, as x—>+oo, >
loga Jc; a* (a >
1) are infinitely large quantities. Applying the L'Hos-
pital rule, compare these quantities.
jloga*
Solution. 1. lim
l 0g x
-^f- = lim ,
k _l = \ogae lim —=
i
kx*
0;
2.
,.
lim —= x rn , .
lim
mx
—. —=
rn ~ 1
. . . = ,
lim
.
-7-71 —^ =
m\ r,
0-
Hence, the power function x k (k>0) increases more rapidly than

the logarithmic function \og a x(a> 1), and the exponential function
a x with the base exceeding unity increases more rapidly than the
power function x rn .
<
a> (b) Hra (cotx-
o
1 1 1
lim
(c)
e* — \
Solution, (a) We have an indeterminate form of the type oo — oo.

Let us reduce it to an indeterminate form of the type — and then
apply the L'Hospital rule:
,.
lim
^^jVlnA:
/
i
l l
x— \J :
\
= r
lim
x ^ j
x— 1— In* =
(x—
r—. —
1) In x
lim
x ^ 1
,
In
1 — 1/* =
,
— [/x,
1
,
= m IJ
_^ !
* In *
*T
+
1
—= r
l ^ ^
lim
,
In a: -|-2 2'
(a) lim x n
\nx(n > 0);
x ->
<b) lim fin (1 + sin 2

x) cot In
2
(1+ x)].
x ->
Solution, (a) We have an indeterminate form of the type 0-co.
Let us transform it to ^ , and then apply the L'Hospital rule:
lim x n In
at ->
a: =. lim
x -+
—- =x
In x
x
lim
->
———- =——\\m1 fx
r
1
" * -+
x n = 0, since rc>0.
(b) We have an indeterminate form of the type 0-oo:
lim [in (1 + sin- x) cot ln> + *)] = (1 Jim =

1
sin 2x
2 {1 -{-tan 2 [In 2 (1 ln(l+x)

1+*
sin*
= hm ln(l+x) = ,.
lim -
^ K^ Q
1+a:
(a) lim (l/x)

sin
*; (b) lim x^^^-".
x -* + x -* i
§ 3.2. Evaluation of Indeterminate Forms 141
Solution, (a) We have an indeterminate form of the type oo°.

Let y -~=(\/xY'mx ; then
In y = s\n x In (1/x),
lim \ny= lim sinxln(l/x) (indeterminate form of the type 0-oo).

x-^-fO x -> +
Let us transform it to ~ and apply the L'Hospital rule:
— —= — \/x
i-
lim \nu
x^ + o
,
= t
lim
.
x^ + oh'smx
-;
In
:
a:
x
,.
lim
^
—
—
;
(cos x)/sin 2 x
= hm
x _^
sin 2 x
xcosx
= A0.
Hence, lim y = e" = l.
x -> -I
tan
(a) lim (sin x) ^; (b) lim**.
X-7C/2 x-»0
3.2.7. Compute
cot *.
lim (tanx)
x -> + Jt/2-0
Solution. Let us take advantage of the identity
(tan co *
* = £ co * * ' n *
an x
but
lim cot X In \anx = lim

}
" tan x
tan *
= ylm — ]Hl =z Q m
x-*+jt/2-0 JC-+jt/2-0 |/=tan * -> + oo #
Whence
lim (ianx) coix = e° = 1.
x-> +jt/2-0
3.2.8. Ascertain the existence of the following limits:

x 2 sin (1/x) m
(a) lim
* ->
sin x
(b)
v
v
lim
,.
—-—
*-*oo(2x+sin2;t)6>
24-2aH-
-
sin 2x
rj—
s,n *
tan a:
(c) lim
X -+ 31/
Can the L'Hospital rule be applied in computing them?

Does formal application lead to the correct answer?
its
Solution, (a) The limit exists and equals zero. Indeed,
li m f__smJJ_/£) _ m jj
x
ii m^s j n _L ^ i .
o 0.
x-*0 smx x-+0 smx x-> x
But the limit of the ratio of the derivatives does not exist. Indeed,
i .
lim
2x sin
—— —=
(
K
1 fx)
1
cos ( 1 fx)
-
A
— lim cos —
, .
1
but lim cos (1 /'x) does not exist, hence the L'Hospital rule is not
x - o
applicable here.
(b) The limit of the ratio of the functions does not exist:
Hm
*->
2
oc(2*+sin2*)<?
+ 2,4-sin2,
sin
*
^
,.x\
,
{
'
2.v-|-sin
2
2xy
v
^ ,
but lim e~ sinx does not exist, since the function e~ Anx traverses the
X -> GC
values from 1/e to e infinitely many times.

Now we will show that the limit of the ratio of derivatives exists:
r
lim
2 -f 2 cos 2a:
—
sm
x-> cc [2 + 2 cos 2a; + (2x -|- sin 2x) cos x| e
A
cos- x
= lim -
A
4 cos "
.,
/n
#+(2x+sin
,
4
. .
,
2a) cos
x ^ x a'
* C° S *
= lim .
g -s.nx = 0>
v - x + 4 cos A' -j- si n Q
2x *
since the function e~* inx is bounded, and ~

2.V+4 cos
,
4 cos A
x+
—_ —+
sin 2a-a:-ôo
o.
Here cosx, which vanishes for an infinite set of values of x has y
been cancelled out. It is the presence of this multiplier that makes

the L'Hospital rule inapplicable in this case, sine? it simultaneously
nullifies the derivatives ofthe functions being compared.
tan sec 2 x sec a tan a

7
(c)
v
lim
SeC x
a-
= ,.
lim
SCC X tan X
t = ,.
urn , = Inn = ...
* -* jt/2 *-*JT/2 x-+n/2 tan x SeC x
Here application of the L'Hospital rule gives no useful result, though

there exists a limit:
tan* sin* cos
hm sec x
= lim
cosx
a;
= t
.
lim sin x
.
= t
1.
x->jt/2 x->n/2 Jt->.-r/2
3.2.9. Using the L'Hospital rule find the limits of the following
functions:
r In 2 — 3) a
Xnx —x
^
7 v
I™ ,3
(a:
+ 3*- 10
;
^
/Ux
l
l
™ =
<?H1=£. i-4s^(^/6)
(c)
v '
lim . v cin v '
(d) Hm - 1 X
X -* '
(e) lim arc sin —^— cot (x — a)\ (f) lim (n — 2 arc tan x) In x;
tdn
(g) Hm f-M *; (h) limfa 1
/*— 1)* (a > 0);
* + V * / * ->• GO
tan (jiA-/(2a))
(i) lim (cos mx)"'*
2
; (j) Mm (2— —
§ 3.3. Taylor's Formula: Approximate Calculations 143
1 (1) lim^/ |n <^- l

>;
X * 1
In a: In x J x -
(m) lim cot 2 X ) ; (n) lim x— x 2

In (
1 +y
x ->
l/sin a
2
(o) lim a; cosh -
(P) lim
-.0 \2+ V~9-{-x
J/x* 1
lim (lncot^) tan lim

(q) *; (r)
2 arc tan x 2 —n '
§ 3.3. Taylor's Formula. Application to Approximate

Calculations
If the function f (x) is continuous and has continuous derivatives
through order n — 1 on the interval [a, 6], and has a finite derivative
of the nth order at every interior point of the interval then at
[a, b] the following formula holds true:
+ f
(x-a)+
f(x)=f
^ f" (a)
^
(a) f (a)
+ r (a) + . .
.
+ -^^ + r ©
(
where
l = a + Q(x— a) and <9< 1.
It is called Taylor's formula of the function f (x).

If in this formula we put a = 0, we obtain Maclaurin's formula:
fw=f (0) r (0) x + r (0) J + . . . + -

» (0)
jj^L
+ f
ln}
&)%, where g=6*, 0<6< 1.
The last term in the Taylor formula is called the remainder in

Lagrange's form and is denoted R n (x):
R.(x)= ^ la+ ' ls - tt)]

ix-ar;
accordingly, the remainder in the Maclaurin formula has the form
3.3.1. the polynomial P (x)

Expand b
2x* x*— 2 1 =x — + + 2x—
in powers the binomial x of
1 using the Taylor formula. —
Solution. To solve the problem it is necessary to find the value
of the polynomial and its derivatives at the point x=\. The
relevant calculations are given below.
P(l)--0, P'(1) = 0,
P"(l)-0, P"'(l) = 18,
P (l)-72,
(4)
P (5)
(l)=120,
P {n) (x) = (n > 6)
at any x.
Substituting the values thus found into the Taylor formula, we get
P(x)^~(x-ir+^(x-iy + ^(x~\r;
P(x)=3(x— iy + 3(x— 1) 4
+ (jc— l)
5
.
3.3.2. Applying the Maclaurin formula, expand in powers of x

(up to x 9 y
inclusive) the function
f(x) = \n(\ + x) y
defined on the interval [0, 1], Estimate the error due to deleting
the remainder.
Solution.
/(0) = lnl=0.
The derivatives of any order of the given function (see § 2.3):
rw=(-ir i
{fî.
^ (0 ) = (_i)"-i („_i)! (n= i f 2, 3, ...).
Substituting the derivatives into the Maclaurin formula, we get
In (1 + *)=*— y + y— • • •+ y + îo W,
where the remainder R 10 (x) in the Lagrange form will be written
as follows:
=— =— x U)
(10)
(*)
= / (£)
*
10
9!
<£< *)•
10! 10! (1 + g)
10
10 (1 H- £)
10 (°
Let us estimate the absolute value of the remainder R 1Q (x)\ keeping

in mind that 1 Oâ:^
and I 0, we have >
l*io(*)h 10(1
3.3.3. How many terms in the Maclaurin formula should be taken

for the function f(x) e =
x so as to get a polynomial
representing
this function on the interval [ 1, —
1], accurate to three decimal
place ?
Solution. The function f(x)=e x has a derivative of any order
f
(n)
(x) = ex .
§ 3.3. Taylor's Formula: Approximate Calculations 145
Therefore, the Maclaurin formula is applicable to this function. Let

us compute the values of the function e x and its first 1 deriva- n—
tives at the point x =
0, and the value of the Aith derivative at the
point £ =
0jc(0<6< 1). will have We
=/"'- l »(0)=--l;
f(0) = r (0) = r(0) = ...
fin)
^ =e -
==e Bx m
Whence
/(*) = +TT+ir+
1
• • •
+(f^. + R ^ x ^
where
Since, by hypothesis, and 0<9< 1, then
\Rn{x)\ = -^^<±e< _3_

3
nl
"
Hence, if the inequality
4- < 0.001 (*)
is fulfilled, then the inequality
I I
< o.ooi
will be fulfilled apriori. To this end it is sufficient to take n 7 >
(7! = 5040). Hence, 7 terms in the Maclaurin formula will suffice.
3.3.4. At what values of x will the approximate formula
cos x
than 0.00005?
« 1
— 2T + TT
have an error less
Solution. The right member of the approximate equation repre-
sents the first six terms in the Maclaurin formula for the function
cos a: (the second, fourth and sixth terms are equal to zero; check it!).
Let us estimate R Q (x). Since (cosa:)
(6)
cos x then =— 9
cos Qx
!*.(*)!
-
6!
•
R
x6 < 1*1
6!
For the error to be less than 0.00005, choose the values of x that
satisfy the inequality
-L£J1 < 0.00005.
Solving this inequality, we get |#|< 0.575.
3.3.5. Compute the approximate values of:
(a) cos 5°; (b) sin 20°,

accurate to five decimal places.

Solution, (a) Into the Maclaurin formula
CO sx= 1— -g. + iL — . . . +(— i)«-gL + R 2n+2

substitute x=n/36\ since
2
lf=2^ = Tf=l(-f) = 2 410

- 003808 ""' -
'
we confine ourselves to the following terms:

cos a: « 1 x 2 /2 y
the error being estimated at

cos 6x
<-L^<2.5.10- 6
.
4T~
And 50, within the required accuracy
cos 5° = cos i
00
= 1 —0.00381 = 0.99619.
3.3.6. Compute the approximate value of Y 83 accurate to six
decimal places.

(a) x—x 2
/2 < > 0;
In ( 1 + x)< x at x
(b) tanx> < x < jx/2;

x + x /33
at
l+±x — ^<]/ T+x<l+-jX

r
(c) at 0<x<oo.
Solution, (a) According to the Maclaurin formula with the rema-
inder R2 (x) we have
In (1 +*) = *— 2(1 + 6)"
where < |< x.
According to the same formula with the remainder R s
(x) we have
In (1 +*) = *— 4" 3(1 + ^)3 >

whe re
0<l x
<x.
YiTTW > and > >

Since at * °> jt follows that
an+EO 8
x—x 2
/2 < In (1 +*)< a:.
3.3.8. Show that sin (a + ft) differs from sina + /icosa by not
more than h 2 /2.
Solution. By Taylor's formula
sin (a + h) = sin a + h cos a — 4r- sin 6;

§ 3.4. Taylor's Formula: Evaluation of Limits 147
whence
+ h) — (sina + /icosa) = —
h 2
h2
'sin (a |
I
sinE |
£ 3.4. Application of Taylor's Formula to Evaluation

of Limits
The expression
f(x) = f (a) + Lw {x _ a)+ rvi (x _ ay + + {x _ a)n H _
-\-o(\x — a\ n
)
is the Taylor formula with the remainder in Peano's form where

q (x) = o [ty (x)] means that, as x —>a the function cp (x) has a high-
y
er order of smallness than the function i.e. lim ?4-t = 0.

In particular, at a = we have
fw=/(0)+m, + m,2+ ... + /^ xn+o(Uh .
Peano's form of the remainder for Taylor's formula shows that,

when substituting the Taylor polynomial of degree n for / (x) in the
neighbourhood of the point a, we introduce an error which is an
infinitesimal of a higher order than (x a)
n
as +a. — x—
The following five expansions are of greatest importance in sol-
ving practical problems:
e*=l+x + -g.+ ... + Jnr+o(x«);

sin x = x—~ + ...+(- 1
j^L. + o (**");
(1 + *)- = + ax + 1
yl
*+
v-3
- +° ~
{a l)
'
'
nl
^^
Yn
x»+o (*");
In (1+*) = *—-£- + -j- +...+ (-I)"" 1
^+ o (x n ).
3.4.1. Expand the function / (x) sin 3 x =

x l e~ x in positive inte- —
gral powers of x up to the terms of the fourth order of smallness
with respect to x.
Solution. We have
/(*)= [x-^+owy-x* \-x + ^- + o(x*)

= x — — + (* — + X — + (X = X — -| X* +
2 6
)
2 3 4
)
3
(X 4 ).
3.4.2. Expand the following functions:
(a) f(x) = xV — x — cos x In (1 + x)\ \

2
(b) f (*) = ln(l+sin*)
in positive integral powers of x up to the terms of the fifth order

of smallness with respect to x.
3.4.3. Applying the Taylor formula with the remainder in Peano's

.form, compute the limits:
— V\+x 2
cosx
r
_
, . l
(*> l'i" _tan**
<b) lira
X - X
lim
cosx —e - X,L
x 2 /2
(c)
i- e x sin x —x (1 -\-x)
(d) hm ^ '»
x -» *
(e) hm —^- 2
.
x -> *
Solution, (a) Retaining the terms up to the fourth order with

respect to x in the denominator and the numerator, we get
lim
- Y + 1 X 2 COS X 1
— (1 -f-#
2
)
1/2
COS X
tan4 * 4
x - *
* + o(* J
x2 x4 1
l+y* + 1/2 (—1/2)
. ,
2
1
)|
-
2 , 7 ; d
4 , ; 4 [,
1 --2" + ,
24
.
+0( *
. .
= 11 m [ J
x - *
V
4 8 24 (x4 )
, .
hm
'
;
' '
= hm , . [
1
1 .
3
*
x-+ L 3
3.4.4. Expand the following functions in positive integral powers

of the variable x up to the terms of the indicated order, inclusive:
(a) f (
x) = e2x ~ x2 up to the term containing x b \
(b) In cos a: up to the term containing x 6 \
x
(c)
eX _ up to the term containing x*.
x
§ 3.5. Testing a Function for Monotonicity

Let a continuous function f (x) be defined on the interval [a, b]
and have a finite derivative inside this segment. Then:
(1) For f (x) to be non- decreasing (non-increasing) on [a, b] it is
necessary and sufficient that /' (x) (/' (x)^ 0) for all x in (a, b). >
(2) For f(x) to be increasing (decreasing) on [a, b] it is suffi-

cient to fulfil the condition f (x) > (/' (x) < 0) for all x in (a, b).
3.5.1. Determine the intervals of monotonicity for the following

functions:
(a) f(x) = 2x — \nx; 2
(b) f(x) = 2x* — 9a;- — 24* + 7;

(c) /(*)
(d) f(x) = \n\x\-
(e) f(*) = 4x — 21 * + s 2
18*4 20;
(f) f(x) = e* + bx.
Solution. The solution of this problem is reduced to finding the
intervals in which the derivative preserves its sign. If the function
f (x) has a continuous derivative in the interval (a, b) and has in
it a finite number of stationary points x u x 2i xn (a <x < x
<x < 2
. . . < xn < b) y where /' (x k ) = (& = 1 , 2, . .
. ,
n), then /' (x)
preserves its sign in each of the intervals (a, x x ), (jc
3 ,
x a ),
(a)The function is defined at x > 0.

Let us find the derivative
f'(x) = lx—\lx.
The function increases if 4x — l/x>0, i.e. * > 1/2.
The function decreases if Ax l/x < 0, i.e. x < 1/2. —
And so, the function decreases in the interval <x < 1/2 and
increases in the interval 1/2<a:<4°°-
(b) Evaluate the derivative
f (
X) = 6x 2
— 18a: — 24 = 6 (x — 3x — 2
4).
vanishes at the points

It 1 x=— and x=4. Since f (x) is a
quadratic trinomial with a coefficient at its highest-power term 6 > 0,
then f (x) in the >
intervals
(_ oo, —1), (4, oo), and /' (x)<0
in the interval (—1, 4). Con-
sequently, f (x) increases in the
first two intervals, whereas in
(—1, 4) it decreases.
(c) In this case the derivative
f'(x) (2x = —
x 2 )e~ x vanishes at
the points x and x=2. =
In
the intervals (— oo, 0) and (2, oo) Fig. 39
the derivative /' (x) and the <
function decreases; in (0, 2) the derivative f'(x)>0 and the func-
tion increases (see Fig. 39).
3.5.2. Find the intervals of decrease and increase for the follo-
wing functions:
(a) / (x)
= cos (n/x);
(b) f(x) = smx + cosx on [0, 2jx].
Solution, (a) The function y = cos (n/x) is defined and differentiable

throughout the number scale, except at the point x = 0\
tj
,
= — sin — ji . ji
.
As is obvious, the sign of y' coincides with that of the multi-

plier sin (n/x).
(1) s\n(n/x) > if
< n/x <

2kn (2k + 1) n (k = 9 ±1, ±2, . ..);
(2) sin (n/x) < if
(2k -f 1) Jt < jt/* < 2(6 + 1) n.
Hence, the function increases in the intervals

1 1
,2Jfe+l '
2k
and decreases in the intervals
26 +2 9
2k-\-\
3.5.3. Investigate the behaviour of the function f(x) = 2s\nx +

-ftanx 3x in the interval ( ji/2, n/2).
Solution. The derivative
f/ . v
= 2 cos # H
ri i 1 o
3 = (1 — COS *) (1 —+— ]
COS X — 2 COS 2
=
/ (a:)
v 7 5 1
2
'
1
COS X COS X
_ 4 sin 3
(jc/2) sin(3A:/2)
cos 2 x
is positive in the intervals ( — n/2, 0) and (0, n/2) and vanishes

only at x = 0. Hence, in (— k/2, jx/2) the function / (a:) increases.
3.5.4. Prove that at 0<#<1 the inequalities
x — x /3 < arc tan < x — x*/6

3
a:
are fulfilled.
Solution. We will prove only the right inequality (the left one
is proved analogously).
The derivative of the function
= arc tan x — x + X
3
f (x) -g-
is equal to
= *2(l 2 2
-i)
/ W
/'(*>=_!
1+JC 2
i
1
i
gl
2
(*
+x 2
)
The function f(x) is continuous throughout the entire number

scale, in particular, it is continuous in the interval [0, 1], and in-
side this segment /' (x) < 0. Therefore, f (x) decreases on the interval
[0, 1] and, consequently, for any point x, 0<x^ 1, the inequality

/(x)</(0) = or
3
arc tanx —x + < ~q
a:
is fulfilled, whence
arc tanx < x— -g-.
3.5.5. Prove the inequalities

x — x /6 < sin* < x
3
at x > 0.
3.5.6. Prove that for O^p^l
and for any positive a and b
the inequality (a by^.a p bP is valid.
+ +
Solution. By dividing both sides of the inequality by bP we get
or
where x = b
.
Let us show that the inequality (*) holds true at any positive x.
Introduce the function
f(x)= 1+xP— (\+x)P; x>0.

The derivative of this function
/' {x) = pxP' 1

— p(l+x)P' = p l
-p (X+'xy-p
1
is positive everywhere, since, by hypothesis, 1 and x > 0. pÔ

Hence, the function increases in the half-open interval [0, oo), i.e.
f(x)=l+xP (l+x)P>f{0) — =
whence 1 + xp >(1 +x)p, which 9
completes the prcof. If we put p=\/n then we obtain 9
3.5.7. Prove that the function y = x b 2x* x increases everywhere, + +

and the function y=l —
x* decreases everywhere.
3.5.8. Determine the intervals of increase and decrease for the

(a) f(x) = x* + 2x — 5; (b) f (x)
= In (1 — x 2
);
(c) f(x) = cosx—x; (d) f(x) = -jx?~;

(e)/W = H5; (0 f{x) =

3.5.9. Prove the following inequalities:
(a) tan* > x + x*/3, if (0 <x< jt/2);
(b) e
x
^ l+x for all values of x\
(c) > at a: > 1.
3.5.10. At what values of the coefficient a does the function

f(x) =x 3
— ax increase along the entire number scale?
?.5.11. At what value of b does the function
f (x) = s'mx — bx + c
decrease along the entire number scale?
§ 3.6. Maxima and Minima of a Function

If a function defined on the interval
y=f(x) then an in-
is X y
terior point x is called

of the point of maximum of
this interval
the function f (x) [the point of minimum of the function f (x)] if
there exists a neighbourhood of the point x U£X
such that the ,
inequality f(x)^.f(x ) [f (x)^f(x )] holds true within it.

fi
The generic terms for points of maximum and minimum of a fun-

ction are the points of extremum.
A Necessary Condition for the Existence of an Extremum. At
points of extremum the derivative /' (x) is equal to zero or does
not exist.
The points at which the derivative f (x) or does not exist are =
called critical points.
Sufficient Conditions for the Existence of an Extremum.
I. Let the function f (x) be continuous in some neighbourhood of
the point x .
1. If /'(#)> and /'(#)< at x <x

at x > x (i.e. if in mo-
ving from through the point x the derivative changes
left to right
sign from plus to minus), then at the point x the function reaches
a maximum.
2. If /' (x) < at x <x and /' (x) > at x >x (i.e. if in mo-
ving through the point x from left to right the derivative changes
sign from minus to plus), then at the point x the function reaches
a minimum.
3. the derivative does not change sign in moving through the
If
point x then there is no extremum.

,
II. Let the function f (x) be twice differentiate (that is /' (x ) 0) =

at a critical point x If f"(x )<0, then at x . the function has a
maximum; if /" (x ) 0, then at x >
the function has a minimum; {)
but if f" (x ) =
0, then the question of the existence of an extremum
at this point remains open.
III. Let f'(x n ) = r(x )=...=P''-»(x = 0, but /<»>(x )^0. H

o )
n is even, then at f (x < there a maximum at # and at

(n)
{) ) is ,
f
{rn
(x ) > 0, a minimum.
If /2 is odd, then there is no extremum at the point x .
IV. Let a function y = f(x) be represented parametrically:
* = <p(0» y = ty(t),
where the functions <p(t) and ty(t) have derivatives both of the first
and second orders within a certain interval of change of the argu-
ment t, and cp' (0=7^0. Further, let, at t = t
y (/) = o.
Then:
(a) if (/ ) < 0, the function y = f(x) has a maximum at x =
= *o = <P('o);
(b) if i|/ (/ ) > 0, the function y = f(x) has a minimum at x =
= *o = <p
(c) if i|5
/r
(/ ) = 0, the question of the existence of an extremum

remains open.
The points at which q/ (/) vanishes require a special study.
3.6.1. Using the first derivative, find the extrema of the follo-
wing functions:
(a) f(x) — 9x +
= -jx'— 3 2
7;
(b) /(*)=**— 8x* + 22x — 24x+ 2

12;
(c) = 1)
3
(*-3) 2
;
Solution, (a) The function is defined and differentiate over the

entire number scale. Therefore, only the real roots of the derivative
/' (x) = 3x* — 3x 2

— I8x = 3x (x + 2) (x—3)
are critical expression to zero, we find the
points. Equating this
critical points: x x = — 2,
(they should always be ar- x2 = 0, x3 =3
ranged in an increasing order). Let us now investigate the sign of
the derivative in the neighbourhood of each of these points. Since
there are no critical points to the left of the point 2, the x= —
derivative at all the points x 2 has one and the same sign: it <—
is negative. Analogously, in the interval ( 2, 0) the derivative is —
positive, in the interval (0, 3) it is negative, at x 3 it is posi- >
tive. Hence, at the points x 2 and x 3 3 we have minima x
=— =
/(— 2)^— 9 and / (3) = — 40-^- , and at the point *2 = 0, maxi-
mum / (0) = 7.
(c) Just as in item (a), the critical points are the roots of the
derivative /'(#), since the function is defined and differentiate
throughout the number scale. Find f (x):
f' (x) = (x + l)
3
- 3) + 3x (x +
(x
2
l)
2
— 3) + 2x (x +
(x
2
1
)
3
x
X (x — 3) = 3 (x + (x — 3) (2x -3x —
2 2
1
) 1).
Equating this expression to zero, we find the critical points:
^ = -1, x2 ^(3-|/T7)/4, *, = (3 + |/T7)/4, x4 = 3.

Let us tabulate the signs of the derivative in the intervals be-
tween the critical points:
Intervals x < X l
< X < x2 x2 < X < x3 x :i
< x < x4 xA < X
Sign of /' (x)
As seen from the table, there is no extremum at the point

is
xt =— there is a minimum at the point x 2 a maximum at the

1, ,
point x 3 and a minimum at the point # 4

,
.
3.6.2. Using the first derivative, find the extrema of the follo-
wing functions:
(a) / (x)
=3
(b) / (x)='t/(x-iy+V(x+iy.
Solution, (a) The function is defined and continuous throughout
the number scale.
Let us find the derivative:
From the equation f'(x) = we find the roots of the derivative:

x= ± 1.
Furthermore, the derivative goes to infinity at the point x = 0.

Thus, the critical points are * 1= 1, x 2 0, x 3 =l. The results =— =
of investigating the sign
of the derivative in the
neighbourhood of these
points are given in Fig.
40. The investigation
shows that the function
has two maxima: /( — 1) = 2; /(1) =2 and a minimum / (0) = 0.
3.6.3. Using the second derivative, find out the character of the
extrema of the following functions:
(a) y =2 sin a;-|-cos2x;

(b) f(x)^2x i — 15x 2 — 84* + 8.
Solution, (a) Since the function is a periodic one we may confine

ourselves to the interval [0, 2n}. Find the first and second deri-
vatives:
y'= 2 cosx — 2 sin 2x== 2 cos x (1 — 2 sin x)\
y"= — 2sinx — 4 cos 2x.
From the equation 2 cos (1 2s\nx) = a: determine the critical
points on the interval [0, 2n}\
x x
= jt/6, x 2 =ji/2, x3 = 5jx/6, x4 = 3ji/2.
Now find the sign of the second derivative at each critical point:
y" (xc/6) = — 3 < 0; hence, we have a maximum y (n/6) = 3/2 at
the point x = Jt/6;
l
(jx/2) — 2 > 0; hence, we have a minimum # (ji/2) = at the 1
point x = ji/2;
2
(/"(5ji/6) = — 3 < 0; hence, we have a maximum y (5ji/6) = 3/2

at the point = 5jx/6;a'
y" (3jx/2) = 6 > 0; hence, we have a minimum #(3^/2) = — 3 at

:j
the point x = 3jt/2 (see Fig. 41).

4
Fig. 41
3.6.4. Investigate the following functions for extrema:
^<Hr +3 ô
Solution, (a) Though the derivative
)
)
/
_2(jr<0),
1 3(x>0)
exists at all points, except the point x= 0, and changes sign from
minus to plus when passing through the point x= 0, there is no
minimum here:
/ (0) = 5 > / (*) at — l<x<0.
155 Ch. III. Differential Calculus: Investigation of Functus
This is explained by the fact that the function is discontinuous

at the point x 0. =
(b) Here the derivative /' (x) = 4x(x 0) also exists at all points,
except at x =
0, and it changes sign from minus to plus when pass-
ing through the point x =
0. Nevertheless, we have here a maximum
but not a minimum, which can readily be checked.
It is explained by the fact that the function is discontinuous at
the point x 0. =
3.6.5. Find the extrema of the following functions:
(a) / (x)
= 3 ^4
+ 8 ^3_ 18a: 2_|_ 6 o»
(b) f(x) = Ve*—\.

Solution, (a) Here it is simpler to find the extrema of the func-
tion f 1 (x) 3x* =
8x* + — 18x + 60. Since 2
x) = 2x + 24x — 36jc= + 2x — 3)
3 2 2
f[ (
1 1 2x (x
fi(*) = 12(3* + 4* —
a
3),
the critical points are:
x1 = 3, x2 = 0, x3 = 1
and the character of the extrema is readily determined from the

sign of the second derivative fl( 3) 0; hence, at the point x — > x
= —3
the function f 1 (x) has a minimum, and the given function f (x)
obviously has a maximum / ( 3) 2/3, f" (0) 0; hence, at — =— x <
the point x 2 =
the function f 1 (x) has a maximum, and f(x) a
= '
minimum /(0) 5/6; / 1'(l)>0; hence, at the point x 3 =^\ the

r
function f 1 (x) has a minimum, and f (x) a maximum /(lj 50/53. =

(b) In this case it is easier to find the points of extremum of
the radicand
which coincide with the points of extremum of the function / (x).

Let us find the critical points of f 1 (x):
f' (x)
l
= 2xe x2 f[(x)=0 at the point x
;
= 0. Determine the sign of
the second derivative at the point x 0: =
f\{x) = 2e*'{\+2x% fi(0)=2>0.
Therefore the point x= is a minimum of the function f^{x)\ it
will also be a minimum of the given function f (x): f(0)=0.

3.6.6. Investigate the character of the extremum of the function
y =
cosh x +
cos x at the point x 0. =
Solution. The function y is an even one and apparently has an
extremum at the point x 0. To determine the character of the =
extremum let us evaluate the derivatives of this function at the

point x = 0:
y'^smhx — sin x, y'(0) = 0;
= cosh^ — cos x,
y" y"(0) = 0;
y"' = sinh x + sinx, y"'(0)=0;
# = cosh x + cosx;
(4)
y
u >=2 > (0) 0.
Since the first non-zero derivative at the point x = is a derivative

ofan even order, which takes on a positive value, we have a mi-
nimum y(0) = 2 at this point.
3.6.7. Investigate the following functions for an extremum at
the point x 0: =
(a) y= cosx— — (b) y = cos x— + -y 1 .
Solution, (a) = — sinx + x a:
^" j/'(0) = 0;
;
= — cosx+1 x; y" (0)=0;

«/
#//
= sinjc— y"'(0) = — 1=^0.
1;
And so, the first non-zero derivative at the point x = is a deri-

vative of the third order, i. e. of an odd order; this means that
there is no extremum at the point x 0. =
Investigate the following functions for extrema:
3.6.8.
(a) x*e~*
/
2
(x) =
(b) / (x) sin 3x 3 sin*.
;
= —
Solution, (a) The function f (x) x*e~ x * is continuously differen- =
tiate everywhere. Equating the derivative
/' (x) = 4x»e- x2 —2x e~ x2 = x e~ x2 (4 — 2/)

b 3
to zero, find the critical points:
= — V2\ x2 = 0; x., = V2.

Compute the values of the second derivative at the critical points:
/" (x) = Wx^-** — 8x*e~ x *

—lOx'e-** + 4x«e-* = 2
= 2x e- x2 {6 — 9x -\-2x y 2 2 A
>
r(0) = 0; r(-K2)<0; f" (V~2 < 0. )
Consequently, at the points x = — K2 and x =-f-j/2 the function t 3
reaches a maximum (±1^2) = 4e~ = \ As far as the critical

f
2
.
point x = concerned, nothing definite can be said as yet, we

2
is
have to find derivatives of f (x) of higher orders (up to the fourth

order!). But this process cumbersome, therefore we will turn to
is
the first sufficient condition of an extremum: let us find the signs

158 Ch. III. Differential Calculus: Investigation of fund's
of the first derivative in the neighbourhood of the critical point

*a =0:
/'(-1)<0; /'(1)>0.
Hence, at the point x = the function has a minimum /(0) = 0.
3.6.9. The function y = f(x) is represented parametrically:
|
= <p(t) = — 5P — 20/ + 7,
x t
b
\
y== ty(t) = 4P — 3t —\8t \-3 2
(—2 <t <2).
Find the extrema of this function.
Solution. have We
<p'(0 = 5f — 4
15/
2
— 20.
In the interval (—2, 2) y'(t)=£0.
Find \|/ (/) and equate it to zero:
i|/(0= 12/
2
— 6/ — 18 = 0.
Whence *, = 1 and f = 3/2.

?
—
These roots are interior points of the considered interval of va-
riation of the parameter t.
Furthermore:
f (t) = 24/ — 6; ^ — = _ 30 <( 1 ) 0, f (3/2) = 30 > 0.
Consequently, the function y=:f(x) has a maximum #=14 at

/= — 1 (i.e. at #=31) and a minimum */
=— 17.25 at t = 3/2
(i. e. at *= — 1033/32).
3.6.10. Find the maxima and minima of the following functions:
Ax
(a) f{x) = x*e-*\ (b) f(x) = '
xa -|- 4
(c) f(x) = - x- V(^W; (d) f (x)

=
a:
4 — 8x + 2'
14
a
(e) / (*)
= ^/2x + 3jc — 36jc;3 2
(f ) / (*)
= x In *; 2
(g) f(x) = x In 2
a:.
3.6.11. Investigate the following functions for an extremum at

the point x 0: =
(a) f (x)
= s'm x—x\ (b) f(x)==s'mx —x + x*j3;
(c) f(x) = s\nx—x + ^ — £;
e v*, if x=£0,
<d > '<*>=.(>. if x=
§ 3.7. Finding the Greatest and the Least Values 0/ a Function 1 59*
§ 3.7. Finding the Greatest and the Least

Values of a Function
The greatest value of a continuous function f (x) on an

(least)
interval [a,attained either at the critical points, or at the
b] is
end-points of the interval. To find the greatest (least) value of the

function we have to compute its values at all the critical points
on the interval [a, b], the values f(a), f (b) of the function at the
end-points of the interval and choose the greatest (least) one out
of the numbers obtained.
If a function is defined and continuous in some interval, and if
this interval is not a closed one, then it can have neither the
greatest nor the least value.
3.7.1. Find the greatest and the least values of the following
functions on the indicated intervals:
(a) f(x) = 2x — 3a: - \2x+ on

3 2
1 [—2, 5/2];
f(x) = x \nx on [1, e\\
2
(b)
(c) f (x)
= xe~ x on [0, +00];
(d) f(x) = V(\-x*)(l + 2x*) on [-1. 1].
Solution, (a) Find the derivative f (x):

I' (x) = 6x — 6x — 12.
2
It vanishes at two points: x = —

1
and x = 2. They both lie in-
1
2
side the indicated interval

'
-2
z - consequently both of them
>
2
must be taken consideration. To find the extreme values of
into
the function it is necessary to compute its values at the points
a-j and x 2 and also at the end-points of the segment:
,
f(-2) = -3, f(-l) = — 19, f(A) = _16 2

= 8; /(2)
1
'
/2 J
Hence, the greatest value /(— 1) = 8 and the least f(2) = — 19.
is
(b) Find the critical points: f (x) = x(l + 2\nx). The derivative
f (x) does not vanish inside the given interval [1, e]. Therefore
there are no critical points inside the indicated interval. It now
remains to compute the values of the function at the end-points of
the interval [1, e]
f(l) = 0; f(e) = e\
Thus, /(1) = is the least value of the function and f(e) =e 2
the
greatest.
(a) f/ = sin a; sin 2a; on (—00, 00);

(b) # = arc cosx on [— V 2

2/2, J/2/2];
(c) y = x + ]/ x on [0, 4].
Solution, (a) Represent the function y = sin sin 2x in the form
—
a:
= cos x — cos 3x
y 2 «
whence it is seen that the function an even one and has

is a pe-
riod 2jx. Hence, it is sufficient greatest and the
to seek the least
values among the extrema on the interval [0, n]. Find the deri-
vative y'\
y' = y(3sin 3x — sin x).

In [0, n] the derivative vanishes at the points
x, = 0, x2 = arc cos yL= ,

x3 = arccos^ — '
x^ rz=zjl -
Compute the values of the function at these points:
y(0) = y(n) = y y arc cos , ,

l
r_ = 4-
3 vi
Hence, the least value of the function in the interval (— oo, oo)
is equal to —4/(3/3), and the greatest to 4/(3 1/3).
3.7.3. The function

f(x)=ax + ^ (a, b, x>0)
consists of two summands: one summand is proportional to the
independent variable x, the other inversely proportional to it.
Prove that this function takes on the least value at x = ]/~b/a.

Solution. Find the roots of the derivative f (x) in the interval
(0, oo):
f'(x) = a-± =
at x = V'b/a (x>0). Since f" (x) = 2b/x> > for any x > 0, the func-
tion f (x) reaches a minimum at this critical point. This is the
only extremum (minimum) in the interval (0, oo). Hence, at
x = Vb/a the function f (x) attains the least value.
3.7.4. As a result of n measurements of an unknown quantity
x the numbers x lt x 2 , . . . ,
xn are obtained.
It is required to find at what value of x the sum of the squares
of the errors
f ( X) = (
X - Xl r + (x-x + 2 )*
. . .
+ (*-*B )
2
will be the least.

§ 3.7. Finding the Greatest and the Least Values of a Fund ion 161
Solution. Compute the derivative
1' (
X) = 2 (x—x + 2(x—x )+...+2
t ) t
(x —x n ).
The only root of the derivative is
Then, for all x we have f" (x)

==== 2n > 0. Therefore, the function
f(x) has its minimum at the point
Being the only minimum, it coincides with the least value of

the function (cf. Problem 1.3.8).
And so, the best (in the sense of "the principle of the minimum
squares") approximate value of an unknown quantity x is the arith-
metic mean of the values x ly x 2 , . . .
,
xn .
3.7.5. Find the largest term in the sequence
°n = *
/i
3
+200
Solution. Consider the function f (x) =^ 2QQ

in the interval
[1, oo). Since the derivative

x (400-^)
1 W~ (x 3 + 200) 2
is positive at <x< jJ/400 and negative at x> j/400, the
function < x < j/^400 and decreases at # > jJ/400.

f (x) increases at
/
From the inequality 7<j/ 400 <8 follows that the largest term it
in the sequence can be either a or a Since a =49/543 > a =8/89,

7 8
.
7 8
the largest term in the given sequence is
(a) f(x) = ]-x*-^x»-jX* + 2 on [-2, 4];
(b) f (*) = |/4 — x on [—2, 2];

2
(c) / (x) = arc tan x— y In on jt
(d) /(jc)=2sinjc + sin2A' on 0, |k ;
6 -3148
(e) f(x) = x — 2 In x on [1, e]\
(i)f{x)=
i
2^ 2 +|2
for -2<*<0; 0<*<2,
^ 1 for * = 0.
§ 3.8. Solving Problems in Geometry and Physics

3.8.1. The force of a circular electric current acting on a small
magnet with the axis perpendicular to the plane of the circle and
passing through its centre is expressed by the formula
Cx
F=
(a 2 +* 2
)
3/f
where a = radius of the circle

x = distance from the centre of the circle to the magnet
(0 < x < oo)
C--= constant.
At what x will the value of F be the greatest?
Solution. The derivative
a2 ~ 2x
F'(x) =C l
(a 2
+x v 2
)
«
has a single positive root x =

alV~2. This solves the problem.
Note. It often happens that reasons of purely physical or geo-
metric character make it unnecessary to resort to the differential
methods in investigating a function for the greatest or the least
value at the point under consideration.
Determine the most economical dimensions of an open-air

3.8.2.
swimming pool of volume 32 m 3 with a square bottom so that the
facing of its walls and bottom require the least quantity of ma-
terial.
Solution. Let us denote the side of the bottom by x and the
height by y. Then the volume V of the pool will be
V = x 2 y = 32, (*)
and the surface S to be faced

S — x2 + 4xy.
Expressing y through x from the relation (*), we get
S = x2 + \x —1
X }
=x + l
1
— X
.
Investigate the function thus obtained for a minimum in the

interval (0, oo):
S' = 2jc—^; 2*-^ = 0; x = 4.

§ 3.8. Solving Problems in Geometry and Physics 163
The single point thus found will obviously yield

the least value
of function S, since it has no greatest value (it increases un-
the
boundedly as x O and —
> oo). x—
And so, the required dimensions of the pool are: x 4 m, y ^2m. =
3.8.3. Inscribe into a given sphere a cylinder with the greatest
lateral surface.
3.8.4. 20 m of wire is available for fencing off a flower-bed

which should have the form of a circular sector. What must the
radius of the circle be if we wish to have
a flower-bed of the greatest possible surface y
area?
Solution. Let us denote the radius of the
circle by x and the
y
length of the arc by y
(see Fig. 42). Then
2O = 2jc + 0,
whence
Fig. 42
= 2(10-*).
The area of the circular sector S = ^xy = x (10 x) (0<#<10).
The derivative S'(#) 10 = — 2x has a root x--=5.
Since the least value S = reached at the end-points of the
is
interval [0, 10], the obtained value

x =
5 yields the greatest surface area S.
3.8.5. It is required
construct to
an open cylindrical reservoir of capa-
city V The thickness of the material
.
is d. What dimensions (the base radius
and height) should the reservoir have

so as to ensure the least possible
expenditure of the material?
Solution. Figure 43 represents a lon-
gitudinal section of the reservoir,
where the radius of the base of the
inner cylinder is denoted by x and
the height of the inner cylinder,
Fig. 43
by h. The volume of the bottom
and the wall of the reservoir
V= jx (x + df d + n [(x + df — x 2
] h = nd (x + df + nh (2xd + d a
). (
*)
On the other hand, by hypothesis we must have

V = nx 2
h
whence
164 Ch. III. Differential Calculus: Investigation of Functus
Substituting into (*), we get
V^nd (x + dY + ^(2xd + d')^nd(x+dr + 2 ^ + ^.

Now we have to investigate the obtained function V (x) for an
extremum at x > 0.
We have
V (x) = 2nd(x + d)- -^-

2
-^
2
= M<«+">y-r.>
The only positive root of the derivative is the point x= y'Vjn.
This solves the problem:
3.8.6. A factory D is to be connected by a highway with a

straight railway on which a The distance DB
town A is situated.
from
factory to the railway is
the
•0 equal to a the segment AB of the y
/ railway equals /. Freight charges on

1
/ the highway are

! times higher than m
a
/
/
on the railway (m
\
How should the highway

\
1).
be
>
DP
— ^
A x
~
p
/'
A
connected with the railway so as to
[
ensure the least freight charges from

factory to town?
b
AA Solution. First, let us make a draw-
r is 44
ing (see Fig. 44). It is absolutely
clear that the highway must also be
straight (a straight line is shorter than any curve connecting two
given points!). Furthermore, the point P cannot lie either to the
left of the point A or to the right of the point B. If we denote
the distance AP by x it will mean that
y
0^*^/.
Let the freight charges on the railway (per ton-kilometre) be k,
then the freight charges on the highway will be km. The total
freight charge N
for transporting loads from to A amounts to D
N = kx + km Va + (l — x)\ 2
Hence, we have to find the least value of the function
f X) = x + m y a + (x—i)\ o<*</.
*
Take the derivative

m (x — /)
§ 3.8. Solving Problems in Geometry and Physics 165
It vanishes only at one point:
x = l-
j/~m 2 - 1
If this point lies in the interval [0, /], i.e. if
or
V 1
tri
then it yields the least freight charge (which is easy to check).

If the indicated inequality is not observed, then f (x) increases on
[0, /] and therefore the least freight charge is obtained at x 0. =
In constructing an a-c transformer it is important to insert
3.8.7.
into the coil a cross-shaped iron core of greatest possible surface
area. Fig. 45 shows the cross-section
of the core with appropriate dimen-
sions. Find the most suitable x and
y if the radius of the coil is equal
to a.
3.8.8. the source of current is an

If
electric cell,then the effect P (watts)

obtained by cutting a resistance R
(ohms) in the circuit is expressed by
the formula
E2 R
P =
(/?+/?/)*' Fig. 45
where E is electromotive force in volts and R the internal resis- {
tance in ohms.
Find the greatest effect which can be obtained at given E and R { .
3.8.9. A tin of a given volume V has the form of a cylinder.

What must be the ratio of its height h to diameter 2R so as to
use the least amount of material for its manufacture?
3.8.10. In a given cone inscribe a cylinder having the greatest

lateral surface so that the planes and centres of the base circles
of the cylinder and cone coincide.
3.8.11. Given a point (1, 2) in the orthographic coordinates.

Through this point draw a straight line so that it forms, together
with the positive semi-axes, a triangle of the least area.
3.8.12. Given a point M

on the axis of the parabola y 2 2px =
at a distance a from its vertex. Find the abscissa of the point
on the curve nearest to the given point.
166 Ch. III. Differential Calculus: Investigation of FuncVs
3.8.13. The expenses sustained

in one hour's sailing of a ship
are expressed empirical formula of the form
in roubles by an
a +
bv\ where a and b are constants for a given ship, and v is the
ship's speed in knots (one knot is equal to 1.85 km/hr). In this
formula the constant part of the expenses a refers to depreciation
and crew's upkeep, and the second term (bv's ) to the fuel cost.
At what speed will the ship cover any required distance at the
lowest cost?
3.8.14. A troughthree boards of equal width. At

is built from
what slope should the boards be placed to ensure the largest
lateral
cross-sectional area of the trough?
3.8.15. A tank with a vertical wall of height h is installed on

a horizontal plane. Determine the position of an orifice, at which
the range of a liquid jet will be the greatest if the velocity of
flow (according to Torricelli's law) is equal to Y 2gx, where x is
the depth of the orifice.

3.8.16. Two are flying in a straight line and in the
aircraft
same plane an angle of 120° to each other and with an equal
at
speed of v km/hr. At a certain moment one aircraft reaches the
point of intersection of their routes, while ihe second is at a dis-
tance of a km from it. When will the distance between the
aircraft be the least and what is that distance?
£ 3.9. Convexity and Concavity of a Curve. Points of

Inflection
If f"(x)<0 (> 0) on an interval (a, b) y then the curve y = f(x)
on this interval is convex (concave), i.e. it is situated below (above)
any of its tangent lines.
If /"(a: ) =
or does not exist but f (x ) does exist and the second
derivative f" (x) changes sign when passing through the point x Qy
then the point (x 0l f (x )) is the point of inflection of the curve
3.9.1. Find the intervals in which the graphs of the following

functions are concave or convex and locate the points of inflection:
(a) y = x' + x* — 18a: + 24a:— 12; 2
(b) y---3x* — 8x + 6x*+l2; 3
(d) y=x+x l
'>;
(e) = 4l/>-l) + 20K(x-l)

(/
5 3
(x^ly,
y = ^r (*>°)
]J
(0 ;
§ 3.9. Convexity and Concavity. Points of Inflection 167
(g) = *sin(ln*) (x > 0);

(h) = 2-|**-l|.
Solution, (a) Find the derivatives:
y' = 4x + 3x —36x + 24,
3 2
12* 2 + 6* — 36 = 12(V+-|- — 3) ,
whence y" = at x1 = —
2, x 2 = 3/2.
Hence, y" > on the intervals (—00, —2) and (3/2, 00); <
—
on the interval ( 2, 3/2). The sign of the second derivative deter-
mines the convexity or concavity of the curve in a given interval*
This enables us to compile the following table:
X x < — 2 ~2 x
3
< < 1
Sign of y" + +
Conclusion Concavity Convexity Concavity
Since the second derivative changes its sign when passing through
the points x 1 = —2
and * 2 3/2, the points ( 2, =
124) and — —
» are points of inflection.
(d) Find the derivatives:

^ 10
y
/
=
1
l
i
+T x "<
2 /
f= 9-i77-
//
The second derivative is non-zero everywhere and loses its meaning

at the point x = 0. At x < we have y" < and the curve is con-
vex, at x > we have y" > and the curve is concave.
At the point x=0 the first derivative y'=l, the second deriva-
tive changes sign when passing through the point x=0. Therefore
the point (0, 0) is a point of inflection.
(g) Find the derivatives:
y' = sin (In x) + cos (In *),

y" = y [cos (In x) — sin (In*)] = -~- sin (-4- — In .
The second derivative vanishes at the points
Xk = î/A + kn
9
& = 0, ±1, zh 2, ... .
The function sin(jt/4 —

In*), and together with it y\ changes sign
when passing through each point x k Consequently, the points x h .
168 Ch. II L Differential Calculus: Investigation o) Fund's
are the abscissas of the points of inflection. In the intervals

^>2fcjl-3Jl/4 £2&Jl + JT/4^
the curve is concave, and in the intervals

^>2fcjl + JT/4 £2&Jl+ 5JT/4^
it is convex.
(h) The given function can be written in the following way:
2— (x — b
1), x> 1,
y= 2+ — (a:
5
x< 1.
1),
Therefore
r _ \
— 5x\ *> 1,
y ~\ 5x\ x<l.
At the point x= 1 there is no derivative. Further,
/
—20a: 3 ,
x> 1,
20x\ x<l;
y" = at the point x = 0. Hence, we have to investigate three in-
tervals: (— oo, 0), (0, 1), (1, oo).
Compile a table of signs of y"\
X x < < x < 1 x > 1
Sign of y" +
Conclusion Convexity Concavity Convexity
The point (0, 1) is a point of inflection, the point (1, 2) being

a corner point.
3.9.2. What conditions must the coefficients a, b c satisfy for 9
the curve y ax* —

bx 3 +
cx 2 -{-dx +
e to have points of inflection? +
Solution. Find the second derivative:
y"= \2ax* + Sbx + 2c.
The curve has points of inflection if and only if the equation
e>ax 2
+ 3bx + c =
has different real roots, i.e. when the discriminant 9b 2 — 24ac > 0, or
3b 2
— 8ac > 0.
§ 3.9. Convexity and Concavity. Points of Inflection 169
3.9.3. At what values of a will the curve
y = x* + ax* + ^-x + 2
1
be concave along the entire number scale?

Solution. Find y"\
y" ^ \2x* + 6ax + 3.
The curve will be concave along the entire number scale if y" ^0
for all values of x, i.e. when
Ax 2 + 2ax + >0 1 for all x.
For this it is necessary and sufficient that the inequality 4a 2 — 16 ^

be fulfilled; whence
|a|<2.
3.9.4. Show that the curve y --=
^ {
has three points of inflec-
}
tion lying in a straight line.

Solution. Find the derivatives:
, _ — x — 2x+ * 1
y ' '
l)
2 '
r
2x :i
-\ 6x 2 — 6x—
y " (x*-f-l) 3
The second derivative becomes zero at three points, which are the
roots of the equation
x [i
+ 3x — 3x — = 0,2
1
whence
x x
-— 2 — [/ 3~, x2 - —2 + [/T, x, = 1
Let us compile the table of signs of y"\
— 00 < X < -2- j/" 3 < v< -2-l ^3 <

X < x < oo
-2-
1
< J/T <-2r F"3 < x < 1
Sign of y" _L_

+
Conclusion Convexity Concavity Convexity Concavity
Hence, (- 2 - kT. —^) . (-2 + ^3. ^p-).d. D

are points of inflection. It is easy to ascertain that all of them
170 Ch. Ill . Differential Calculus: Investigation of Fund's
lie in a straight line. Indeed, the coordinates of these points satisfy

_2— }/T— 1 (1- }/T)/4— 1
the relation
-2+/" 3 -1 (\+V 3)/4+l
3.9.5. Investigate the curves represented by the following equa-
tions for convexity (concavity) and locate the points of inflection:
(a) y= x- ^(x-3) 2
;
(b) y=e
x
(— jx/2 < x < jt/2).
s[n
3.9.6. Show that the points of inflection of the curve y = xsmx

lie on the curve y 2 (4 + # = 4x 2
)
2
.
§ 3.10. Asymptotes
A straight line is called an asymptote to the curve y f(x) if =
the distance from the variable point of the curve to the straight M
line approaches zero as the point recedes to infinity along some M
branch of the curve.
We will distinguish three kinds of asymptotes: vertical, horizon-
tal and inclined.
Vertical asymptotes. If at least one of the limits of the function
f (x) (at the point a on the right or on the left) is equal to infi-
nity, then the straight line x a \s a vertical asymptote. =
Horizontal asymptotes. If lim f(x) A, then the straight line =
X -> ± 00
y=Â is a horizontal asymptote (the right one as x—> + and

the left one as x — — oo).
Inclined asymptotes. If the limits
X
lim
-* + cc
—=X
kv
*-» +
lim
oo
[f (x)
— k x]=b 1 A
exist, then the straight line y = k x+b 1 l

is an inclined (right)
asymptote.
If the limits
Urn
— X
=k 2
and lim
—
[f (x)
— k x]=b 2 2
X -+ CO X -* 00
exist, k 2 x-\-b 2 is an inclined (left) asymp-

then the straight line y —
tote. A
horizontal asymptote may be considered as a particular
case of an inclined asymptote at k 0. =
3.10.1. Find the asymptotes of the following curves:
(a) #= -^3 ;
(b) y= ^r+3x; (c) y = -^;
#= 7 + 4,v 2
(d) ; (e) y-=xe^; (f) In
</=--f
§ 3. 10. Asymptotes 171
(g) y = Vl + x* + 2x\ (h) y=}/\ +x 2

sin-L;
(i) y = 2Vtf + T.
Solution, (a) The curve has a vertical asymptote x = 3, since
5x
lim y = lim
,x — 3' oo
x^3T 0" x -+ 3 T- '
(the point x = 3 is a point of discontinuity of the second kind).

Find the horizontal asymptote:
lim y = lim = 5.
* -* ± GO X -+ ±00 X 6
And so, the curve has a vertical asymptote x =3 and a horizon-

tal one y = 5.
(b) The curve has a vertical asymptote x= 1, since
Â
a:
lim
-> I -
# =
a:
lim
-> 1 -
(-^13^==
\x ~~
— 00; 1
/
x-+
lim
1 +
# =
x-1+0
lim
\*~~
— -r
1
+ 3x = + /
I 00.
Find the inclined asymptotes:
k= lim — == lim .
——.
r
3 =3;
b = lim (y—kx) =
X -*± GO
= lim
3a:
3* — 3x = 3 )
a; -> ± co
Thus, the straight line y=3x+3 is an

Hg. 46
inclined asymptote Fig. 46). (see
(e) The curve has a vertical asymptote x= y since
lim */ = lim xe l/x = lim -^-=+00

X -* + X -> +
r = + oo
x
(see Problem 3.2.2.).
Find the inclined asymptotes:
& =
X
lim 4"=
- ± CO * Xv
lim
-+. ~
± CO
,
^= g°=l;
b = lim
±
(a^ 1 /* — x) = lim
±V3
— l x
e l
.
—
:
lim
'
* -* CO X ->
l/x -
Thus, the straight line y =

x+\ will be an inclined asymptote
of the curve (see Fig. 47). Note that
liin y = Urn xe l/x = 0.

-V -> — x -* -
(f) The function is defined and continuous at e— ^ > 0, i.e. at
x < and x > 3e

*
Since the function is continuous at

every point of the domain of definition,
vertical asymptotes can exist only on
finite boundaries of the domain of defini-
tion.
As x —> — we have
lim y= lim ^\n(e — ~) =

liin
In (e + z) =
(see Problem 3.2.2.), i. e. the straight

Fig 47
line a; = is not a vertical asymptote.
As x ^+ we have
lirn y = -
lim x\n[e
3x
K-*\/(3e) +
i.e. the line x=-l/(3e) is a vertical asymptote.

Now let us find the inclined asymptotes:
k = lira
X-*±cc X
b = lira [y — kx] = -
1r lirn x
X-+±cc
[n{l
-3x-e
lirn
x-*±<x> \ 3^ 2e
Hence, the straight line y— 3x2 2e

is an inclined asymptote
(see Fig. 48).
(g) The curve has no vertical asymptotes, since the function is
continuous everywhere. Let us look for inclined asymptotes. The
limits will be different as x—+ + oo and x— -oo, therefore we have
to consider two cases separately.
§ 3.10. Asymptotes 173
Find the right asymptote (as x-^ + oo):
*,= [i m
>" l
+ ', + 2« - lim
Yl + 1 +2
= 3;
X
*^ + CO X-* + CO
6, = lim (^1+a:2 + 2x — 3x) =

1+JC 2 !
— 3
= lim [|/T+72 --;e|: lim 0.

*-* + oo j/ 1 -f *
2
+*
Thus, as x—* + oo the curve has an asymptote y = 3x.
Fig. 48 Fig. 49
Find the left asymptote (as x-+ — oo):
* = Iim jClT^f^ lim iH^J

fe
2
= lim + x + 2a;— x\ =
2
lim =
0,
since both summands ([/l+r* and ( a;)) in the denominator are

positive at x 0. <
And so, the curve has an asymptote y = x as x —> oo. —
(h) The curve has no vertical asymptotes, since it is continuous
at x =^=0, and in the neighbourhood of the point x=^0 the function
is bounded.
Let us find the inclined asymptotes. We have
k = lim —= lim ± 1-0=0.

£-±00 X-+ ± 00
Then
f 1 as x
AT — + oo,
+ ^sin — =% _
I
! I
l 1
as v — oo.
Thus, the curve has two horizontal asymptotes: y 1 and y = 1 + —

(see Fig. 49). The same result can be obtained proceeding from
symmetry about the origin and keeping in mind that the function
y is odd.
3.10.2. Find the inclined asymptote of the graph of the function
y = j^-^ as x— oo and show that in the interval (100, oo) this
function may be replaced by the linear function y = x— 1 with an

error not exceeding 0.01.
Solution. Find the inclined asymptote:
k = lim —TT-t
X2
—= 1;
b= lim (
t4 x) =— 1.
x+cc V1+* J
And so, the equation of the asymptote is y = x — 1.

Form the difference:
-(*-!) =
l+JC
Hence, assuming
for all x> 100, we introduce an error of not more than 0.01.
3.10.3. Find the asymptotes of the following curves:
^==
*2 -6*+ 3 = xarctanx;
(a) ;
(b) y
y=x+(smx)/x; y = In (4— x
2
(c) (d) )\
(e) y = 2x — arc cos .
§ 3.11. General Plan for Investigating Functions and

Sketching Graphs
The analysis and graphing by elementary methods
of functions
were considered in Chapter Using the methods
I (§§ 1.3 and 1.5).
of differential calculus, we can now carry out a more profound and
comprehensive study of various properties of a function, and explain
the shape of its graph (rise, fall, convexity, concavity, etc.).
It is convenient to investigate a function and construct its graph
according to the following plan:

1. Find the domain of definition of the function.
2. Find out whether the function is even, odd or periodic.
§ 3.11. Investigating Functions and Sketching Graphs 175
3. Test the function for continuity, find out the discontinuities

and their character.
4. Find the asymptotes of the graph of the function.
5. Find the points of extremum of the function and compute the
values of the function at these points.
6. Find the points of inflection on the graph of the function,
compute the values of the function and of its derivative at these
points. Find the intervals of convexity of the graph of the function.
7. Graph the function using the results of this investigation. If
it is necessary to specify certain regions of the curve, calculate the
coordinates of several additional points (in particular, the x- and

^/-intercepts).
This is a very tentative plan, and various alternatives are pos-
sible.For instance, we recommend the student to begin sketching
the graph as soon as he finds the asymptotes (if any), but in any
case before the points of inflection are found. It should be remem-
bered that in sketching the graph of a function the principal refe-
rence points are the points of the curve corresponding to the extremal
values of the function, points of inflection, asymptotes.
3.11.1. Investigate and graph the following functions:
(a) y =x Q
— 3x* + 3x*— 5; (b) y--=\/~x— Vx+l;
(
c) y=-£zi*
-4 • (
d) y- -^ 1
(e) y = x+ln (x 2 — 1); (f) y = -j sin 2* + cos*;

(g) y = x*e l 'x
; (h) y = arc sin 1=^-
Solution, (a) The function is defined and continuous throughout
the number scale, therefore the curve has no vertical asymptote.
The function is even, since /( x) f(x). Consequently, its graph =
is symmetrical about the (/-axis, and therefore it is sufficient to
investigate the function only on the interval [0, oo).
There are no inclined asymptotes, since as x-ôo the quantity
y turns out to be an infinitely large quantity of the sixth order
with respect to x.
Investigate the first derivative:
0' = 6* 6
— I2x + = e>x(x* — 2x +
s
e>x
2
1) =§x{x 2 — l) 2 ;
the critical points are:
*i =— 1, *2 = 0> x» = 1 •
Since in the interval [0, oo) the derivative y'^0 y

the function
increases.
Investigate the second derivative:
y
" = 3(k 4
— 36* + 6 = 6 (5x — 6x +
2 4 2
1 ).
The positive roots of the second derivative:
*, = 1/K5, *2 =1.
For convenience and pictorialness let us compile the following
table, where all the points of interest are arranged in an ascending
order:
X (I. <*>) 2
(f?0
1
(°-
7?)
y' + +
25 ^5
y" (5
-f -t-
y 23
On the right one more additional value of the function is com-

puted to improve the graph after the point of inflection.
Using the results of the investigation and the above table and
taking into consideration the symmetry principle, we construct the
graph of the function (see Fig. 50). As is seen from the graph, the
function has roots where 1.6. + a&
(b) The function is defined and continuous over the entire number
scaleand is negative everywhere, since \/ x < \/ x + 1.

The graph has neither vertical, nor inclined asymptotes, since the
order of magnitude of y is less than unity as x—* oo. Determine
the horizontal asymptote:
lim y = lim (j/x— + l) =
Hence, the straight line y=^0 is the horizontal asymptote of the

graph.
§ 3.1 J. Investigating Functions and Sketching Graphs 177
The first derivative

l
3^/'(FH7 2
3
becomes zero at the point x2 = —y and infinity at the points
*i =— 1 ,
xb = 0.
-1
Fig. 50 Fig. 51
The second derivative
y
„_ J_/
~
___2 \ _J 1_ A) 1
3 V 3 J y# 3
does not vanish and is infinite at the same points x 1

=— 1, x3 = 0.
Compile a table:
1
X -l (--{) 2 K-°) (0, oo) 1
y' — 00 + 00 +
16
y" 00 00
1
9yj
1 ^ I
—0.26
With the aid of this table, and of the asymptote y = construct

the graph of the function (see Fig. 51).
(c) The function

is defined and continuous over the entire axis
except points
at the x=±2.
The function is odd, its graph is
symmetrical about the origin, therefore it is sufficient to investigate
the function on the interval [0, oo).
=
The straight line x 2 \s a vertical asymptote:
lim
2x s _ lim
2* 3
-\- OO
* -+ 2 - x2 —4 x-2 +
Determine the inclined asymptote:
2x 2
lim lim = 2,
v_w _1_ m X X -* + CD
b = lim (y — 2x) = lim

8x
X -* +CC X + CD
The curve has an inclined asymptote # = 2x, and
y—2x = Sx j > > 2,

at x
I < <2
at x
6x 2 (x 2 — 4) — 4x 4
2x (x — 12)
2 2
y ~ (x 2— 4) 2 ~ (x — 2
4)
2
in the interval [0, oo) vanishes at the points
= 0, *=2]/~3«3.46
*
and becomes infinite at the point x = 2.
„_ 16a: (A:
3
-h 12)
y ~ (at
2 — 4) 3
becomes zero at the point x= and infinite at x = 2.

Compile a table:
r
(0, 2) 2 (2, 2/" 3 ) 2 |/"3 (2V 3", oo)
y' — oo +
3
y" +o 00 + 2
+
Using the results of the investigation, sketch the graph of the

function (see Fig. 52).
(e) The function is defined and continuous at all values of x for
which x 2 — 1>0
or |x|>l, i.e. on two intervals: ( oo, 1) — —
and (1, + oo).
-A
V
4* l I
-1\
Fig. 52 Fig. 53
We seek the vertical asymptotes:
lim
- -
y = lim
- 1 -
[x + In (x — 2
= — oo;
1)]
1
lim y= lim [x+ In (x

2
— 1)] = — oo.
Thus, the curve has two vertical asymptotes:
x= — 1 and x = f 1
Find inclined asymptotes:

u
= —y = — In (a:
2 — = 1 ,. In (a:
2 — 1)
£
i-
lim
i-
lim
)
urn 1
]-•
b = lim [y — x] = lim In (x 2
— l)=+oo.
Jt -*± 00 X-+ ± oo
Hence, the curve has neither inclined, nor horizontal asymptotes.

Since the derivative
2*
^ 1+
y
i .
^r
exists and is finite at all points of the domain of definition of the
function, only the zeros of the derivative
x,= - \—V% x2=-l + \f2

can be critical points. At the point x = — +
2
1 the function is
not defined; hence, there is one critical point x = — l~l 2 belon-

ging to the interval ( —
00, —1). In the interval
A
(1, 00) both the

derivative y' >
and the function increase.
hence, the curve is convex everywhere, and at the point x x

=
= — — ^2^ — 2.41 the function has a maximum
1
y _ — J/2) « — — J/2 +
( 1 (2 + 2^2) « — 0.84. 1 In
To plot the graph in the interval (l,oo), where there are no characte-
ristic points, we choose the following additional points:
x=2\ */ = 2 + ln3^3.10 and x= 1.2; y = 1.2 + ln0.44 ^0.38.

The graph of the function is shown in Fig. 53.
(f) The function is defined and continuous throughout the num-
ber scale and has a period 2n. Therefore in investigating we may
confine ourselves to the interval [0, 2n\. The graph of the function
has no asymptote by virtue of continuity and periodicity.
Find the first derivative:
y' = cos 2x — sin x.
On the interval [0, 2k] it has three roots:
*i
= ji
"g" » ^2 = ~Q~
5ji 3ji
*8 == ~2~
Evaluate the second derivative:

y" == — 2 sin 2x — cos x.
On the interval [0, 2n] it has four roots:
x, = y, x2 = k + arc sin (1/4), x3 = ^, x4 = 2n — arcsin(l/4).

Let us draw up a table of the results of investigation of all cri-
tical points of the first and second derivatives (the table also in-
cludes the end-points of the interval [0, 2ji]).
Since in the interval ^0, —-^ the roots of the first and second
derivatives alternate, the signs of the second derivative in the in-
tervals between its critical points are indicated only for the last
three intervals.
The
results of the investigation enable us to construct the graph
of the function (see Fig. 54).
(g) The function is defined, positive and continuous on each of
the intervals ( —
00, 0) and (0, 00). The point x is a disconti- =
ji Jl 5ji 3ji
X 6~
T ~6~ AJ
2 (*• t) 2 (¥-) x4 (* 4 .2ji) 2ji
y' —2
y"
3 ]T3 3 ^ 3~
+
z 2
y l
nuity. Since (see Problem 3.2.2.)
lim y = lim jt
2
e /*
[
= Iirn ^- = oo ( t = —) ,
the straight line x=0 is a vertical asymptote. But

limy = lim x e Ux 2
= 0.
x-+ - *^ -
There are no inclined asymptotes, eince the function y=x e Ux 2
has the second order of smallness with respect to x as x—+±.oo.
n
\t/2 B jc
2%
i*
-1
Fig. 54
Let us find the extrema of the function, for which purpose we

evaluate the derivative:
y' = 2xe [/x
— e = 2e { ' x 1
'* {x— 1/2),
whence we find the only critical point *=y.

Since for x=^=0
y" (x) = 2*«/* - £ e*'* + 1 = 1 e u* (2x* - 2x + 1) > 0,
182 Ch.Iff. Differ eniial Calculus: Investigation of Fund's
on each of the intervals of the domain of definition the graph of

the function is concave, and at the point x 1
/ 2 the function has
=
a minimum
— — p*
* 1.87.
4
From the information obtained we can sketch the graph as in Fig. 55.
To specify the graph in the intervals ( oo, 0) and (V 2 oo) the — ,
following additional points are used:

x=— 1, = e~ ^ 0.37;
y l
x=l y
y = e^2J2.
(h) The function is defined and continuous
throughout the number scale, since at any x
1-f-x
2
< 1.
Since the function is even, we may confine

Fig. 55
ourselves to the investigation of the function
at *>0.
As the function continuous, the graph has no vertical asymp-
is
totes, but it has a horizontal asymptote:
x
limy
-* + oo
= =
arc sin (-
D=Hf.
— 2x (\+x — 2x(\—x 2
)
2
)
4*
X (l+x 2 2
2 |*|
X (1+x 2
~ 2
(1-**)* ) )
(1+x 2 2
)
is negative for x>0, therefore the function decreases

The derivative is non-existent
at the point x=0. By virtue of
the symmetry of the graph about
the y-axis there will be a maxi-
mum at the point #(0) = ^-. No-
tice that
at the point x the =
right derivative is equal to 1,
—
and the left one to 1. +
The second derivative is posi-
tive:
2(l+x 2 )2x 8x
y x) —2 > for a
(
(l-fx 2 ) 4 (i +* 2 )3
Hence, in the interval (0, oo) the graph of the function is concave.
Also note that the curve intersects with the x-axls at the points
x =± 1.
§ 3.12. Algebr. and Transcendent. Equations
Taking into consideration the results of the investigation, con-

struct the graph of the function (see Fig. 56).
3.11.2. Investigate and graph the following functions:
(a) y=\+x — y; (b) y = j~^\

2
(c) y = ~ + ^- (d) y = -^r x

\
(e) y =y^-¥^=A-
(f) y = x*\n(x+2); (g) y =
\x arc tan— atxÔ,
10 at x = 0.
§ 3./2. Approximate Solution of Algebraic

and Transcendental Equations
Approximate determination of isolated real roots of the equation
f(x) = isusually carried out in two stages:
1. Separating roots, i.e. determining the intervals [a, P] which
contain one and only one root of the equation.
2. Specifying the roots, i.e. computing them with the required
degree of accuracy.
The process of separation of roots begins with determining the
signs of the function f (x) at a number of points x a lt a 2 =
whose choice takes into account the peculiarities of the function / (x).
If it turns out that / (a k ) f (a k+1 ) 0, <
then, by virtue of the
property of a continuous function, there is a root of the equation
f(x) =6 in the interval (a ky a k+1 ).
Real roots of an equation can also be determined graphically as
^-intercepts of the graph of the function y =
f(x). If the equation
has no roots close to each other, then its roots are easily separated
by this method. In practice, it is often advantageous to replace
a given equation by an equivalent one
*i (*) = *« (*)•
where the functions ^
(x) and i|; 2 (x) are simpler than the function
f(x). Sketch the graph of the functions y = ty 1 {x) and y = 2 (x) y\)
and find the desired roots as the abscissas of the points of inter-
section of these graphs.
The Methods of Approximating a Root.1. Method of Chords. If
the interval [a, b] contains the only real root £ of the equation
f(x)= and f (x) is continuous on the interval, then the first ap-
proximation x 1 is found by the formula

To obtain the second approximation x 2 a similar formula is ap-

plied to that of the intervals [a, or [x 19 &], at the end-points
of which the function f (x) attains values having opposite signs. The
process is continued until the required accuracy is obtained, which
is judged of by the length of the last obtained segment.
Method of Tangents (Newton's method). If f(a)f(b)<0, and
2.
f (x) and F ( x ) are non-zero and retain definite signs for

then, proceeding from the initial approximation x (x £[a, b]) for
which f (x ) r(*o)>0, we obtain all successive approximations of
the root g by the formulas:
10 rw 2 1
/'w "' •
X"- Xn -*
r(*«-i)'
To estimate the absolute error in the nth approximation we can
apply the general formula
where
m = 1
min \f(x)\.
Under the above conditions the method of chords and the method
of tangents approximate the sought-for root from different sides.
Therefore, it is usual practice to take advantage of their combination,
i. e. to apply both methods simultaneously. In this case one can obtain
the most precise approximation of a root more rapidly and the cal-
culations can be checked. Generally speaking, the calculation of the
approximations x l9 x 2 , x n should be continued until the decimal
. . .
,
digits to be retained in the answer cease to change (in accordance

with the predetermined degree of accuracy!). For intermediate trans-
formations we have to take one or two spare digits.
3. Iteration Method. The equation f(x) is first reduced to the =
form x = cp (x) where q/ (x)
|
q < 1 (q = const) for a^x^b. Star-
| ^
ting from any initial value x £[a b] successive approximations of t y
the root I are computed by the formulas x 1 = q) (x ), x 2 = q> (Xj), xn = . . . ,
^y( xn-i)' The absolute error in the nth approximation can be

estimated by the following formulas:
\t— x n\<Tz^\xn -i—Xnl
if the approximations x n _ 1 and x n lie on the same side of the root,

and
I \ Xn I I
X n-1 Xn |»
J
if the approximations x n _ 1 and x n lie on different sides of the root.

§ 3.12. Algebr. and Transcendent. Equations 185
3.12.1. Locate the roots of the equation
Solution. Compile a table of signs of f (x) at some chosen points
X X
— 00 1
~3 3 +
— 1
+ +- oo 4-
+
From this table we draw the conclusion that the equation has
three real roots lying in the intervals (—3, —1), (0, 1) and (1, 3).
3.12.2. Determine the number of real roots of the equation
f(x)==x + e* = 0.
Solution. Since /' (x) = +e x >

1 0; / ( — oo) = — oo; f (+ oo) = + oo,
the given equation has only one real root.
3.12.3. An approximate value of the root of the equation f{x) =

= x*—x—l=0 is x=l.22. Estimate the absolute error in this
root.
Solution. We have /(*) = 2.2153 — 1.22 — = — 0.0047. 1 Since at
jc= 1.23
/ (x)
= 2.2888— .23— - 0.0588, 1 1
the root £ lies in the interval (1.22, 1.23). The derivative f (x) =
= 4r* — 1 increases monotonically, therefore its least value in the
given interval is
m = 4xl.22
1
3 — 1 = 4x 1.816— 1 -6.264,
wherefrom we get an estimate of the error
I
x — < LZiZU =^^l « 0.00075 < 0.001.
1 1
3.12.4. Solve graphically the equation
x log X — = 1 0.
Solution. Let us rewrite the equation in the form

186 Ch. Ill'.
Differential Calculus: Investigation of Fund's
Here \f 1 (#)
= log a:, i|)
2 W=j- There are tables for the values of
these functions, and this simplifies the construction of their graphs.
Constructing the graphs y = \ogx and y = -j (see Fig. 57), we find
the approximate value of the only

root 1 « 2.5.
3.12.5. Find the real root of the
equation
/ (x) = x — 2x + 3x—5 =
s 2
with an accuracy up to 10~ 4 :
(a) by applying the method of

chords,
(b) by applying the method of
tangents.
Pig 57 Solution. Let us first make sure
that the given equation has only
one real root. This follows from the fact that the derivative
/' (*)=3* a — 4* + 3>0.

Then, from /=(1) = —
3<0, (2) 1 it follows = >
that the given
polynomial has a single positive root, which lies in the interval (1, 2).
(a) Using the method of chords, we obtain the first approximation:
^=1—-=^.
1
4
1=1.75.
Since
/(1.75) = — 0.5156 <0,
and / (2) = > 0,
1 then 1.75 <£< 2.
The second approximation:
x2 = 1.75 + • 0.25 = 1 .75 + 0.0850 = 1 .8350.
Since / (1.835) =—
0.05059 0, then 1.835 g 2. < < <
The sequence of the approximations converges very slowly. Let
us try to narrow down the interval, taking into account that the
value of the function f (x) at the point x 2 1.835 is considerably =
less in absolute value than f (2). We have
f(1.9) = 0.339 >0.

Hence, 1.835 <£< 1.9.
§ 3.12. Algebr. and Transcendent. Equations 187
Applying the method of chords to the interval (1.835, 1.9), we

will get a new approximation:
*3= l-835- .3-;

+ S 9
059 •
0065 = L8434 -
Further calculations by the method of chords yield

x4 = 1.8437, xb = 1.8438,
and since /(1.8437)<0, and / (1.8438) > 0, then 1.8438 with

-4
the required accuracy of 10 .
(b) For the method of tangents we choose x = 2 as the initial

approximation, since /(2)=1 and /" (x) = 6x— 4 >
>0 in the in-
terval (1, 2). The first derivative f'(x) = 3x
2
4x 3 also retains its — +
sign in the interval (1, 2), therefore the method of tangents is ap-
plicable.
The first approximation:
^==2 — 1/7= 1.857.
The second approximation:

(1.857) 0.0779
-^27g= 1.8439.
-
^= 1.857-^-^=
QC7, / 7
1.857 , QC -
Q/1Qft
The third approximation:
^= 1.8439-^^= 1.8438,
already gives the required accuracy. Here the sequence of the ap-
proximations converges much more rapidly than in the method of
chords, and in the third approximation we could obtain an accuracy
up to 10" 6 .
3.12.6. Find the least positive root of the equation tan x = x with
an accuracy up to 0.0001 applying Newton's method.
3.12.7. Find the real root of the equation 2 x logx=0 by — —
combining the method of chords with the method of tangents.
Solution. Rewrite the left member of the equation in the follo-
wing way:
f(x) = (2-x) + (- log*),
whence it is seen that the function f (x) is a sum of two monoto-
nically decreasing functions, and therefore it decreases itself. Con-
sequently, the given equation has a single root £.
Direct verification shows that this root lies in the interval (1, 2).
This interval can be narrowed still further:
1.6<£< 1.8,
188 Ch. 111. Differential Calculus: Investigation of Fund's
since
^ (1.6) = 0. 1959 > 0; / ( 1 .8) = — 0.0553 < 0.
Then
f W = _l_l lege; /*(*) = -! log*
and
/'(*)< 0; rW>0interval
over the whole interval
both the method
[1.6; 1 8].
Applying to this of chords and the
method of tangents with the initial point ,v =1.6 we obtain the
first approximations:
= 1 '
6 - (
V(i78
1 6)
l /(i
(1 )
= 1
6H 0. 1559= 1 .7559;
) 6)
x[=l.6—pffi = )
1.6 + 0.1540= 1.7540.
Applying the same methods to the interval [1.7540, 1.7559], we

get the second approximations:
(1.7540-1.7559) / (1.7559)
jc,= 1.7559- = 1.75558,
/ (1.7540)—/ (1.7559)
/(1 7540)
x'-l
X 2 ~
7540-
l./DW
-
(1.7540)
1.75557.
Since x 2 — = 0.00001,
x'2 the root £ is computed with an accuracy
up to 0.00001.
3.12.8. Using the combined method find all roots of the equation
f(x)==x 3 — ox + = 1 accurate to three decimal places.
3.12.9. Applying the iteration method find the real roots of the
equation x — sin x= 0.25accurate to three decimal places.
Solution. Represent the given equation in the form x 0.25 sin a:. — =
Using the graphical method, we find that the equation has one
real root which is approximately equal
to x = 1.2 (see Fig. 58).
Since
sin 1.1 >
=0.8912 1.1 — 0.25,
sin 1.3 = 0.9636 < 1.3-0.25,
the root £ lies in the interval (1.1, 13).
Let us rewrite the equation in the
form
x = cp (x) = sin x + 0.25.
Since the derivative q/ (x) =
cos x in the interval (1.1, 1.3) does
not exceed cos 1.1 0.46 < <
1 in absolute value, the iteration method
is applicable. Let us write successive approximations
xn = sin x n _, + 0.25 (n=l, 2, ...),

§ 3.12. Algebr. and Transcendent . Equations 189
taking x = 1.2 for the initial approximation:
= sin 1.2 +0.25 = 0.932 +0.25=1.182;

jc, = sin 1.182 +0.25 = 0.925 +0.25= 1.175;
x3 = sin 1.175 +0.25 = 0.923 +0.25= 1.173;
x4 = sin 1.173 +0.25 = 0.9219 + 0.25 = 1.1719;
jc
6
= sin 1.1719 + 0.25 = 0.9215 + 0.25= 1.1715;
x6 = sin 1.1715 + 0.25 = 0.9211+0.25= 1.1711.
Since ^=0.46 and hence, yzrj< 1» we have g = 1.171 within the

required accuracy.
3.12.10. Applying the iteration method, find the greatest positive

root of the equation
x* + x= 1000
accurate to four decimal places.

Solution. Rough estimation gives us the approximate value of the
root = 10.
We can rewrite the given equation in the lorm
x= 1000 — a:
3
,
or in the form
1000 l_
X2 X 9
or in the form
x=jj/l000 — x and so on.
The most advantageous of the indicated methods is the preceding

one, since taking [9, 10] for the main interval and putting
(p (
X ) = Y 1000 — x,
we find that the derivative
<P'(*)= 3/
3 / (1000 -x) 2
does not exceed 1/300 in absolute value:
|q>' (X) |
<C ~ 300
6W = 4-
3 y 990*
Compute successive approximations of x n with one spare digit by

the formula
x n+1 =y\000-x n (n = 0, 1, 2, ...),
*o= 10,
x x
=V 1000— 10 = 9.96655,
x2 =V 1000— 9.96655 = 9.96666,
x3 = J/ 1000 — 9.96666 - 9.96667.
We may put £ =
9.9667 with an accuracy of 10~ 4 .
Note. Here, the relatively rapid convergence of the process of ite-

ration is due to the smallness of the quantity q. In general, the
smaller the q, the faster the process of iteration converges.
3.12.11. Applying the method of chords, find the positive root of

the equation
f(x) = x + l.U + 0.9*— 1.4 =
3 a
with an accuracy of 0.0005.
3.12.12. Using the method of chords, find approximate values of

the real roots of the following equations with an accuracy up to 0.01:
(a) l)2_2sin* = 0; (b) e
x — 2(1— = x)
2
0.
3.12.13. Applying Newton's method, find with an accuracy up

to 0.01 the positive roots of the following equations:
(a) *3 + 50*— 60 = 0; (b) x* + x — 32 = 0.
3.12.14. Using the combined method find the values of the root
of the equation
x3 — x— 1 -0
on the interval [1, 2] with an accuracy up to 0.005.
3.12.15. Applying the iteration method, find all roots of the equa-
tion \x —5 In x =
5 accurate to four decimal places.

3.13.1. Does the function
j x if x< 1
f{x)=
\\/x if *>1
satisfy the conditions of the Lagrange theorem on the interval [0, 2]?
3.13.2. Prove that for the function y ax $x y the number £ = 2

+ +
in the Lagrange formula, used on an arbitrary interval [a, b] is the y
arithmetic mean of the numbers a and b: £ = (a-f fc)/2.

3.13.3. Prove that if the equation

n~
a x n
+a x l
1
+ . . . +a n _ x = 1
has a positive root x ot then the equation

"
na x n L
+ (n—l)a x n ' + 1
2
. . . +a n _ = x
has a positive root less than x .
3.13.4. Prove that the equation x 4 — 4x — 1=0 has two different

real roots.
3.13.5. Prove that the function / (x) = x n + px-\- q cannot have

more than two real roots for n even and more than three for n odd.
3.13.6. Prove that all roots of the derivative of the given poly-
nomial f(x) =
(x+ l)(x— \)(x— 2){x— 3) are real.
3.13.7. Find a mistake in the following reasoning.
The function
x 2 sin (l/x) for x =^=0,
f{x) = |
\ forx =
is differentiate for any x. By Lagrange's theorem
x 2 sin —x = x( 2| sin
V
—
I
cos \-
I
whence
cos 4- = 2£ sin \ — x sin— (0 < £ < x).
As x tends to zero E will also tend to zero. Passing to the limit,
we obtain limcos (1/g) = 0, whereas it is known that Tim cos (l/x) is
l -* x-+
non-existent.
3.13.8. Find a mistake in the following deduction of Cauchy's

formula. Let the functions f (x) and y(x) satisfy all the conditions
of the Cauchy theorem on the interval [a, b]. Then each of them
will satisfy the conditions of Lagrange's theorem as well. Consequ-
ently, for each function we can write the Lagrange formula:
f(b)-f{a) = f®{b-a), a<t<b,

cp(ft) —v (a) = (I) (b—a),cp' a<l<b.
Dividing the first expression by the second, we obtain:
f(b)-f(a) f'($)(b-a)
= /'(E)
'
(p(ty-cp(a) y'(t)(b-a) (p'(g)
3.13.9. Prove the following inequalities:
(a) — —< In
a
J < —r-
-r- if 1.
3.13.10. Prove that all roots of the Chebyshev-Laguerre polynomial
are positive.
3.13.11. Prove that if the function f (x) satisfies the following

conditions:
(1) it is defined and has a continuous derivative of the (n— l)th
~
order f (n 1} (x) on the interval [x 01 x n ]\
(n)
(2) it has a derivative of the nth order f (x) in the interval
(*0» '0>
°(3) / (*o) = / (*l) =.-•=/ {*n) (X <X <...< *„),
1
then inside the interval [x ,

xn ] there is at least one point I such
that f
i,l)
(l) = 0.
3.13.12. The limit of the ratio of the functions
lim — -p-:
e-*(cosx + sinx)
—— =
e-'2x (cos x -f 2 sin x)
x
,.
hm
_ ^
+
^~ x,.1- L 2 tan*
r -Manx
I
n
is non-existent, since the expression ^ 1 -j-
8"
tan x
is discontinuous at
the points x n = + =
nn n/2 (rc 0, 1, ...), but at the same time the
limit of the ratio of the derivatives does exist:
lim -— (cos^
——
\e~ 2x (cos x~\-2 sin x)\'
-
nr-=
+ sinx)]'
l;
lim — ——.— =
—5e~
-2e-*sinxs
2x s'mx 5
2 x
lim e -*
^ x
_„
= n0.
xôo
Explain this seeming contradiction.
3.13.13. Prove that the number in the remainder of the Taylor

formula of the first order
fia + h) = f(a) + hf'(a) + ^ f" {a + Qh)

tends to 1/3 as h —* /"' (x)
if continuous at x — a and
is /"' (a)=^=0.
3.13.14. Prove that the number e is an irrational number.
3.13.15. Prove that for 0<x^n/2

the function f (x) ---(s'mx)/x
decreases. From obtain the inequality 2x/n
this < sin x <x for
< x < jt/2 and give its geometric meaning.
3.13.16. Show that the function / (x) = # + cos x —a increases;
whence deduce that the equation x + cos x = a has no positive roots
for a < 1 and has one positive root for a > 1.
3.13.17. Show that the equation xe x =2 has only one positive

root found in the interval (0, 1).
f {x ) = {i x + x2sln T for x¥=0 >
{ for x=
is not monotonic in any interval containing the origin. Sketch the

graph f(x).
3.13.19. Prove the theorem if: (1) f (x) and (p (x) are continuous
in the interval [a, b] and differentiate inside it; (2) f(a) = y(a)\
and (3) f'(x)>y'(x) (a<x<b), then f(x)>y(x) (a<x<b).
3.13.20. Show that the function / (x) = ^+ has neither maxima,
nor minima at ad — bc=^0.

3.13.21. In the trinomial x
2
+ px + q choose the coefficients p and q
so that the trinomial has a minimum at x =
3 and that the mini-
mum equals 5.
3.13.22. Test the function / (x) = (x— )

n
q) (x) for extremum at
the point x--x 09 where n is a natural number; the function (p (x) is
continuous at x =x and (p(# )^=0.
3.13.23. Given a continuous function
— sin 7)l*l *^=°>

f w = j( 2 at
at x= 0.
\
Show that /(#) has a minimum at the point x 0, but = is not

monotonic either on the left or on the right of x 0. =
(a) y = \x\ for —

(b) y = E(x) for — 2<x< 1.
3.13.25. Do the following functions have the greatest and the

least values on the indicated intervals?
(a) f(x) = cosx for — ji/2 ^ < a; ji,
(b) / (x) = arc sin x for — <x< 1 1
3.13.26. Prove that between two maxima (minima) of a continuous

function there is a minimum (maximum) of this function.
I a:
2
sin 2 (l/x) for x=£0,
f W
=
\0 for x =
has a minimum at the point # = (not a strict minimum).

3.13.28. Prove that if at the point of a minimum there exists a
right-side derivative, then it is non-negative, and if there exists a
left-side derivative, then it is non-positive.
f
I/* 2 (x>0),
\3* 2 (x<0)
has a minimum at the point #=^0, though its first derivative does
not change sign when passing through this point.
3.13.30. Let x be the abscissa of the point of flection on the i
curve y = f(x). Will the point x be a point of extremum for the

function y =
f (x)?
3.13.31. Sketch the graph of the function y = f(x) in the neigh-
x= —
bourhood of the point
f(-i>=2,
3.13.32. For
1 if
r<— i)
what choice
— of the
i. /"(-i)=o, r'(*) >o.
parameter h does the "curve of
probabilities"
= -4r «-*'** (ft>0)

V Tt
have points of inflection x = zko?
3.13.33. Show that any twice continuously different iable function

has at least one abscissa of the point of inflection on the graph of
the function between two points of extremum.
3.13.34. Taking the function y x* =

8x*+ 18a: 2 8 as an example, + +
ascertain that there may be no points of extremum between the
abscissas of the points of inflection on the graph of a function.
3.13.35. Prove that any polynomial with positive coefficients,

which is an even function, is concave everywhere and has only one
point of minimum.
3.13.36. Prove that any polynomial of an odd degree n^3 has
at least one point of inflection.
3.13.37. Proceeding directly from the definition, ascertain that

the straight line y = 2x+l is an asymptote of the curve y =
2x 4 -f-;c 3 +l
Chapter
INDEFINITE INTEGRALS.
BASIC METHODS OF INTEGRATION
§ 4.1. Direct Integration and the Method of Expansion

Direct integration consists in using the following table of integrals:
(1)
^
U »du = ^ +C
(2) ]^- = \n\u\ + C-
(3) aa dw-=â* + C; J
e = e + C;
tt
du n
^
(4) ^ cos a da = sin u + C; sin^dw = — cosw + C;
J
(5) ^ cosh u du = sinh u-\-C\ ^ sinh udu

= cosh w + C;
du
(6) f;
f t
2
dl
= — cot w + C;
1
J sin «
*
(7) f-^==l a rctan- + C=— iarccot-^ + C, {a > 0);

(8) [ r
du
= arc sin — + C= — arccos— 4-C. (a > 0);
-
r
(9) f
V u1 ±a 2
=ln(tt + J/ i^±tf) + C;
J
(10) \
- — =— - n —— + C.
In all these formulas the variable u is either an independent
variable or a differentiable function of some variable. If
\f(u)du = F(u) +C 9
then
J/
(ax + b)dx = jF (ax + b) + C.
The method of expansion consists in expanding the integrand into
a linear combination of simpler functions and using the linearity
196 Ch. IV. Indefinite Integrals
property of the integral:
J
\ 2=
1 1
a i fi (x)dx=^
i =
l
i
a i lf
J i
(x)dx f
\{=
S i
|fl/|>0).
/
Find the integral = f

*2 +5*- 1
dx.
4.1.1. /
J V x
Solution.
x2+
j = î~ 1
=j 3
(jc / 2 + 5*"', — x - v.) dx =
j
= }x z
i>dx + b^x i>dx—^x- l l
i*dx==
= fi_ + C + -3- x + C - 2*'/. + c = 2

3
/2
2 3
= 2^(^ + |-l) + C.
Note. There is no need to introduce an arbitrary constant after
calculating each integral (as is done in the above example). By com-
bining all arbitrary constants we get a single arbitrary constant,
denoted by letter C, which is added to the final answer.
a * o 1 P6*3 + * 2 — 2*+l ,
4.1.3. /= f
2 ,
2
.
J sin * cos *
Solution. Transform the integrand in the following way:
1 sin 2 *-)- cos 2 * 1 1
sin 2 * cos 2 a; sin 2 * cos 2 * cos 2 * sin 2 *'
Hence,
/ = f-^r-
2 + J[ -£t- = tan x — cot x + C.
J cos * sin x
1
2,
4.1.4. I = ^tan 2
xdx.
Solution. Since tan 2 x = sec 2
x — 1, then
/ = f tan 2 xdx= [— ^|
cos *
I dx = tanx — + C. jc
t/ J J
4.1.5. / =J (x 2 + 5) 3
dx.
Solution. Expanding the integrand by the binomial formula,
we find
/= j(x 6 + 15jc
4
+75a; 2 + 125)dx = y- + -^ + ^+ 125x + C.
4.1.6. / = J (3x + 5) 17
dx.
§ 4.1. Direct Integration and the Method of Expansion 197
Solution. Here it is not expedient to raise the binomial to the

17th power, since u =
3x-\-5 is a linear function.
Proceeding from the tabular integral
we get
_ l (3*+5)» .
r
i
is
rc-
4.1.7. /= r ^
4.1.8. / = J
cos(jtx +
Solution. Proceeding from the tabular integral (4)
J
cos w du = sin w + C,
we obtain
/ = J-sin (nx + 1) + C.
4.1.9. / = J
cos 4a; cos 7a: d*.
Solution. When calculating such integrals it is advisable to use

the trigonometric product formulas. Here
cos 4* cos 7* = (cos 3a: + cos 1 1a;)
and therefore
/ = —IP cos 3a: dx + y

\
1 P
\ cos
1
lU^^-g-sinSAr + ^sin 11a:+
1
C.
/Vote. When solving such problems it is expedient to use the

following trigonometric identities:
sin mx cos nx = -^ — n) x + sin (m-\-n) x];

[sin (m
sinmA:sin nx =^ [cos(m n) x — cos (m + n) x]\

2
cosmx cosnx =
Y [cos (m — n)x + cos (m + n) x].
4.1.10. / = J
cos a; cos 2x cos 5a: dA;.
Solution. We have
(cos x cos 2a:) cos 5x = y (cos + cos 3a:) cos 5x =

a:
= [cos 4a: + cos 6a:] + -j- (cos 2# + cos 8#)»

Thus,
= cos, 2 x dx + J cos d# ^ cos6xdx+ ^cos8xdx^ =

4a; -f-
= sin 2x + sin 4x + ~ sin 6a: + ^ sin 8a:

-g- -f- C.
4.1.11. / = $sin 2
3A:dA:.
Solution. Since sin 2 3A; = — £2!?!?

f
then
/ =Y l c
\ (1
11
— cos6A:)dA; =-ja: — ^sin6^: + C
4.1.12. / = ^cosh 2
(8x+5)dx.
Solution. Since cosh 2 w = cosh , then
/=-§ [1+ cosh (16*+ 10)] dx = lx + l s inh(16x+10) + C.

J
4.1.13. / = ^2 + ^ +5 •
dx dx
Solution. I = ]C x + 4x+5 = ^C
*
= arc tan (* + 2 + C )
d*
4.1.16. / =j
V4— 9x 2 '
Solution. /= 1-7=^=^ = 4- 1-7====- = 4- arc sin ^? + C.

4.1.17. /= r r
dx
=.
5o/^to«. 7=f^J|==r-
VS — x-t-Ax
J
7=iL
j/9-(x+2) J
== 2
arcsin^
3
+
^ C.
4-1.19. / =
|4
_^_.
Solution.
1 — { dx C ch
<ta
=TT7^1n
1
t
2 j/l + AT+2
J 4 — — 4x
2
J 8— t+2) 2
(a;
4^2 2 /*2 — (* + 2) + C.
4.1.21. Evaluate the following integrals:
, , P3 — 2 cot 2 * , C 2 + 3x 2 .
(
C
M cos** <
d>
JiM?^^
4.1.22. Integrate:
f — at^.h-
f cos 2*
^
(5)
^ ' \
y \ —x*
9
J cos*— sin* *
(c) \
^ dx; (d) \ (sin 5*— sin 5a) dx.
§ 4.2. Integration by Substitution
The method of substitution (or change of variable) consists in sub-

stituting cp(/) for a; where q) (t) is a continuously differentiate
function. On substituting we have:
lf(x)dx=lf[<p(t))<p'(t)dt,
and after integration we return to the old variable by inverse sub-

stitution / = q)
-1
(x).
The indicated formula is also used in the reverse direction:
5 f [

where x = y(t).
4.2.1. / = J xVx^dx.
Solution. Make the substitution
Whence
a:— 5=t\ x= t
2
+ 5, dx=2tdt.
Substituting into the integral we get
/=J(/»
+ 5)*.2* dt=2 j(/ + 5* 4 2
) df =2-£ + ^+ C.
Now return to the initial variable x:
2(*-5) 6/ »
3/
10(*-5) »
^
' = '
5
,
3
,
4.2.2. / = j- +d*e*
1
'
Solution. Let us make the substitution l+e*=t. Whence

e
x = — t \, x = ln — 1), dx=dt/(t — \).
Substituting into the integral, we get
dt
J \+e x Jn(/-I)'
<
But
1 1 l_
t(t — \)~t—\ t
y
^ -^
therefore
I = r = ln\t-l\-\n\t\ + C.
Coming back to the variable x, we obtain
I^ln^ + C^x-lnil+e^ + C.
Note. This integral can be calculated in a simpler way by mul-
tiplying both the numerator and denominator by e~ x :
§j^dx = -§ 7^- dx = -\n(e-*+l) + C = l
= -ln^-tl = x-ln(^+l) + C.
4.2.3. /= \
J
/2+3
V&x--5)3
dx.
(x2 - x)dx
4.2.4. /=r
4
+ 3x 2 +l)arctan^-t-J
Solution. Transform the integrand
(\ — \/x*)dx
+ l/x) 2
+l]arctan(.*:+l/;tr
Make the substitution x + -j=t; differentiating, we get
(l-±)dx = dt.
Whence
d
/-f ± '
J (/
2
+l)arctarW
Make one more substitution: arc tan t = u. Then
dt
2 +\' du
t
and
/ = j^=ln|«| + C.
Returning first to t t
and then to x, we have
/ = In |
arc tan t\ + C = In arc tan (* + 7) + c# |
4.2.5. / = I
V a '~ x2 rlr
x*
Solution. Make the substitution:
x= T 1
\ dx =
,
— -^ dt
Hence,
Now make one more substitution: \faH 2 — 1=2. Then 2a 2 / dt

= 2zdz and
Returning to t and then to x, we obtain
^ 2 (3 / ~~ r "
2
J a sin
2
x+b 2 cos 2 a:
Solution.
dx
-
\ a 2 sin 2 A; +& 2
cos 2 a: b2 I a?2 ,
tan 2
+ .
, cos 2 a:
"
J J b 2
a: 1
dx
Make the substitution ^-tanx
&
= t; dt = -%-
6 cos 2
9
a:
. Then
Returning to a:, we obtain
/ = ~ arc tan tan + C.

4.2.7. / = S
£/T + 3 sin xcosjcd*.
Solution. Make the substitution l+3sinA: = / T 3cosjc<2a; = <#. Then
/ = ^7d^^pv. a=i.4<v.+c= (l+3 ; n ^ <

/,
+c.
4.2.8. / = j"
cos X
202 Ch. /V. Indefinite Integrals
4.2.9. f ^-yr .
b
J (arc cos x) y \ — x' 1
dx
Solution. Make the substitution: arccosx^/; r = dt. Then
1^ l - X*
/ =— [
J/ 5 J
Cr 5
d/-4-^"
4
4
+ ^
—
~ c = 4 arc cos rx + c.
.10. f ^^±L=dx.
sin 2x
4.2.11. / = j_5!^L<fr.
+ sin x^1
2
Solution. Make the substitution:
1 -f- sin 2 x=t; 2 sin x cos x dx = sm 2x dx ~^dt.

Then
= ln/ + C = ln(l + sin 2
*) -|- C.
/=Jy-
««
Solution. Substitute
3 + #lnA; = /, (1 + \nx)dx = dt
and get
/ =§j = \n\t\ + C = \n\3 + x\nx\ + C.

4.2.13. Evaluate the following integrals:
In x) x
4.2.14. Find the following integrals:

x dx
(a) ^yi-xdx; (b) f-i^
In
-
In x
(c)
J
5
cos x V^sin .xrck; (d)
^
y* Ax -
§ 4.3. Integration by Parts

The formula
^ udv = uv — J
ydw
is known as the formula for integration by parts, where w and v
are differentiate functions of x.
§ 4.3. Integration by Parts 203
To use this formula the integrand should be reduced to the pro-

duct of two factors: one function and the differential of another
function. If the integrand is the product of a logarithmic or an
inverse trigonometric function and a polynomial, then u is usually
taken to be either the logarithmic or the inverse trigonometric func-
tion. But if the integrand is the product of a trigonometric or an
exponential function and an algebraic one, then u usually denotes
the algebraic function.
4.3.1. / = J
arc tan x dx.
Solution. Let us put here
w = arc tan a;, dv = dx>

whence
dx
^ = r+^
*
;
V==X)
I = jarc tan x dx = x arc tan x — ^ iq~^2 = x arc * an *

—y ^ ~^ x ^
4.3.2. /= ^ arc sin xdx.
4.3.3. / = J
xcosxdx.
Solution. Let us put
u = x\ dv = cos a; <£*;,
whence
du = dx\ v = s'mx i
1= ^ A;cosA:dA:
= Arsin a:
— J sinA:dA:== Arsin x-fcosx + C.
We will show now what would result from an unsuitable choice

of the multipliers u and dv.
In the integral ^xcosxdx let us put
// ^ cos x] dv = x dx,
whence
du — — sin a: dx; a =y a;
2
.
In this case
/ ~ -1 cos x +yJX 2
sin a: <2a;.
As is obvious, the integral has become more complicated.
4.3.4. /= J
a:
3
lnA:dA:.
u = In x\ dv =x 3
dx,
whence
dx
du
i
=—X
;
'
v = — .,\ 1
/ = jc
4
In x — 4" J ^ ^ T — -j^x^dx = ^-x*\nx — -^x + C. nA: 4
4.3.5. 7 = 5
— 2* + 5)<T*dx.
(a;
2
w =x 2 — 2x + 5; dv = e~ x dx y
whence
= (2x — 2)d;c; u = — e~ x \
1 =^ (x 2 —2x+5)e- x dx = — e- x {x — 2x + 5) + 2 2
J
(x—\)e~ x dx.
We again integrate the last integral by parts. Put
x— = l u\ dv = e~ x dx>
whence
du = dx; v = — e~ x .
71 =2 J
(jc— \)e~ x dx=— 2e- x (x— l) + 2 J
e-*dx= — 2*g-* + C.
Finally we get
7 = — e-*(* 2
— + 5) — 2xe-* + C = — e~
2a;
x (x 2
+ 5) + C.
Note. a result of calculation of integrals of the form ^ P (x) e ax dx
As
we obtain a function of the form Q(x)e ax where Q(x) is a poly- ,
nomial of the same degree as the polynomial P(x).

This circumstance allows us to calculate the integrals of the in-
dicated type using the method of indefinite coefficients, the essence
of which is explained by fhe following example.
4.3.6. Applying the method of indefinite coefficients, evaluate
1= J
(3x 3 —\7)e 2x dx.
Solution.
J
(3a:
3
— 17) e 2x dx= (Ax 3 + Bx2 + Dx + E) e 2x + C.
Differentiating the right and the left sides, we obtain
(3a;
3 — 1 7) e
2x = 2 (Ax + Bx + Dx + E) e2x + e2x (3 Ax + 2Bx + D).
3 2 2
Cancelling e 2x , we have
3x — 17 = 2 Ax + (2B + 3 A) x + (2D + 2B) x+(2E + D).
3 3 2
Equating the coefficients at the equal powers of x in the left and

right sides of this identity, we get
3= 2A\ = 2B + 3A;
0-2D + 2B; — 17=2£ + D.
Solving the system, we obtain
A— — -
R— ~- D—— '
F — —
Hence,
(3x 3 — 1 7)
a
*dx - (-| a; 3 — ~ + -J x — a;
2 2
) e * + C.
j £?
^
4.3.7. Integrate:
/ =J (*
3
+ \)cosxdx.
a = a;
3
+ 1 ; do = cos x dx,
whence
da = 3jc 2
(iA;; o = sinA:.
/ = (^3 +1) sin a: —3 ^ a;
2
sin x dx= (x 3
+ 1) sin x — 3/ lf
where lx = ^
2
x sin x dx.
Integrating by parts again, we get
/x =— a;
2
cos x + 2/ 2,
where /2 = ^ a: cos atâ:.
Integrating by parts again, we obtain

I2 = xsmx + cos x + C.
Finally, we have:
/ = ^ (x*+ 1) cosA:(iA:=(A: 3 + l)sinA;+3A: cosA: 6a: sin x — 6 cosa: + C =

2
= — 6a: + sin x + (3a: — 6) cosa:+C.

(a;
3
1)
2
Note. The method of indefinite coefficients may also be applied

to integrals of the form
P (x) sin ax dx, ^ P (x) cos ax dx.

J
4.3.8. /= J
(a:
2
+ 3a: + 5) cos 2a: g!a:.
jj
(a:
2
+ 3a:+ 5) cos 2a:g!a: =
= (A x 2
+ A x+ A x 2)
cos 2a: + (B a:
2
+B x+B x 2)
sin 2x + C.
Differentiate both sides of the identity:
+ 3x + 5) cos 2x = —2 (A x + A x + A sin 2x +
(x2 2
x 2)
+ (2A x+A cos 2a; + 2 (B x + B^x+B^) cos 2x+ (2B x+B

l)
2
1
)s'm2x=
= [2B x + (2B, + 2A x+(A,+ 2B cos 2x +
2
) 2 )}
+ [— 2A x + (2B — 2A,) x -f (B —2A 2

l 2 )]
sin 2x.
Equating the coefficients at equal powers of x in the multipliers

cos 2x and sin 2x> we get a system of equations:
2B = 1; 2(B 1 + A ) = 3; A + 2B = 5; 2
—2A = 0; 2 (B — A,) = 0; B
1
x
— 2A = 2
0.
Solving the system, we find

A
/i
_A. d
— u, /Jo — J_2 .
'
/I
i
__J_.
— 2 '
r
i
_
— A2 •
'
/i
2
_
— A4 •
'
R
2
—A
— 4
•
Thus,
J
(*
2
+ 3x + 5)cos2xdx= ( y + t) cos2jc+ (y * + jx + ~) sin 2 2
x+C.
4.3.9. /= J
(3a:
2
+ 6a: + 5) arc tan a: dA:.
w = arc tan a:; do = (3a: + 6a: + 5) dx

2
y
whence
^w ^ I
"»
o = + 3a: + 5
a:
3 2
a:.
Hence,
/ = (*» + 3a: + 5x) arc tan x— ^"^ 3 5* dx.
2
j ff
Single out the integral part under the last integral by dividing
the numerator by the denominator:
=J + 3x + 2j^-3.J I ^ 1 =f + 3*+21n(x
2
+l)-3arctanA'+C.
Substituting the value of / lf we finally get

I = (x* + 3x + 5a: + 3) arc tan x—
2 2
/2 — 3x— 2 In (x
2
+ + C.
1)
4.3.10. Find the integral
/ = ê bx cosAxdx.
e bx = ^; cos 4.v =. do,
§ 4.3. integration by Parts 207
whence
5e bx dx = du\ v = sin 4a:.
Hence,
/ — —4 e 1
5*
sin \x — 5 P
\ e 5 *sin4A:dA:.
Integrating by parts again, we obtain
I 1 z=
r*
j
sin Axdx^ — j
5
-^-e *cos4.*; + ^-
5 P
\
5
e *cos4A:dA:.
Thus,
/ = -j e bx sin 4a:
1
— —5/1
—j ( e 5 * cos \x + 5 P
\ e bx cos 4xdx)
\
9
l. e.
/ = -L e bx ^sin 4x + -|-cos 4a: ^

Whence
/ = ê bx ^sin 4a: + -|- cos 4a:
j
+ C.
4.3.11. / =$ cos(lnx)dx.
£/ = cos(lnA:); dv = dx y
whence
= — sin (In x) ^ ; ^ = a:.
Hence,
/ = J
cos (In A:)rfx = xcos(lnA:) + ^ sin (In a:) dx.
Integrate by parts once again
// -=sin (In x)\ dv=dx,

whence
dx
da = cos (In a:) — \ v=x.
Hence,
/i =Js * n (lnA:)dx = xsin(ln a:) — J

cos (ln^)dA:.
Thus
/ = cos (In a:) dv = a: cos (In a;) + x sin (In x) — /.
Hence
/ =y [cos (In a:) -f- sin (In a:)] + C.
4.3.12. / = jxln(l +-j)dx.

Solution. Let us transform the integrand
In
(
1 + 1 ) = In i±l = in (x + - In x. 1)
Hence
/ = J
x In (x+ \)dx — \^x\nxdx = — Ii /2
Let us integrate /x and /2 by parts. Put

a = ln(x+ 1); dv = xdx,
whence
Hence
/1 = Jxln(^+l)dx = -i(x«- l)ln(x+l)-l - \)dx (
j
X —
2
In X +l)-±§(x-\)dx = £ (x+l)-i*« + ±x + C.
1
2 (
T±ln
Analogously,
I2 =^ xln xdx=Y^ nx — -j" *a
Finally we have
/ = J*ln(l +^)^ = l(x 2

-l)ln(x+l)-^lnv+|+C.
4.3.13. / = j ^^R[ln(^+l)-2 1n,]^

Solution. First apply the substitution
Then
2dx dx
<« = -V or
? ?
I j,
<ft.
Hence,
The obtained integral is easily evaluated by parts. Let us put
u = \nt; dv = \/"tdt.
Then
du = Y> v =-~t\ft
Whence
-1 Jj/Tln*<tf=— -g-
\jtVT\nt— =
Returning to x, we obtain
'—
(^+i)
9x 3
1
[2—3 In
(1+33)] +C
4.3.14. / = sinx In tanxdx.
4.3.15. / = J \n{VT^x+VT+lc)dx.
u = In + /l+x) ; dv = dx,
whence
~~
2 '
j/t^*+ i^tt* '
]/T=i2~~ 2 '
* ^r=^
= a;.
Hence,
/ = x in {VT=Tx + VTTi) - \
j
x^^^dx =
= ^1^/1 — x + + x) — yjc + y arc sin x-fC.

Note.In calculating a number of integrals we had to use the
method integration by parts several times in succession. The
of
result could be obtained more rapidly and in a more concise form
by using the so-called generalized formula for integration by parts
(or the formula for multiple integration by parts):
û (x) v (x) dx = u (x) v (x) — u' (x) v (x) + u" (x) v (x) — ...
l 2 9
. .
+ (— l)"- ^"- (x) v n (x) — (— l)"" J u (x) vn (x) dx,
.
1 1
*
1 {n)
where
v1 (x)=^v (x) dx; v 2 (x) =^v x
(x) dx; . .
. ;
v n (x) = J vn _ x
(x) dx.
Here, of course, we assume that all derivatives and integrals appea-

ring in this formula exist.
The use of the generalized formula for integration by parts is
especially advantageous when calculating the integral ^ P n (x) cp (x) dx,

where P n (x) is a polynomial of degree n and the factor y(x) is y
such that it can be integrated successively 1 times. For example, n+
J
P n (x) dx = P n (xf£ - P'n (x) ~+
e
. . .
+
+ (-!)» PJ» {x)£t + C =
4.3.16. Applying the generalized formula for integration by parts,

find the following integrals:
(a) ^(x 3 — 2x + 3x—l)cos2xdx,2
(b) (2x + 3x — 8x +
3
V2x + 6 dx.
2
1 )
J
Solution.
(a )
^( x
3
—2x + 3x—l) cos 2x dx = (x
2 3
— 2x + 3x — 2^ 2
1 )
— {3x — 4x + 3)
/o 9
2 /« i o\ f
^
cos
^
2a: \ /c
j+(€>x,
— 4)l^
/ sin 2x\
—J — c6—cosj^— + Cn =
2a: .
= 5!£?f (2^3 _ 4x + 3*) + 2

( 6* 2 — 8* + 3) + C
(b) l(2x
3
+ 3x 2
— Sx + l)V~2x + 6dx =
= (2x + 3x 3 2
— 8x+ \)
{2x + 6)3/2 --(e>x + 6jc— 8) (2 2
*y +
= (2* + 6) (7 0x a — 45x - 396* + 897) + C.
2
Evaluate the following integrals:
4.3.17.
J
In (x + Vl+x*)dx.
4.3.18.
J
Yx{\nxfdx.
arc s ' n ****
4 3 19
J
x cos # dx
4.3.20.
sin 3 a:
4.3.21. 3* cos x dx.

J
4.3.22. ^(x 3 — 2x + 5)e

2 3x dx.
4.3.23. ^(l+x2 )
2
cosxdx.
4.3.24. ^(x 2 + 2x—l)sm3xdx.

4.3.25.
J
(a:
2
— 2x + 3)\nxdx.
3
4.3.26. jc arc tan jc dx.
J
4.3.27. arc cos jc dx.

J
4.3.28. Applying the formula for multiple integration by parts,

calculate the following integrals:
x * ~* x+
+ x — 2) sin
*
(a) ^(3x 2 2
(3* + \)dx\ (b) dx.
^
§ 4.4. Reduction Formulas
Reduction formulas make it possible to reduce an integral depend-

ing on the index n 0, called the order of the integral, to an >
integral of the same type with a smaller index.
4.4.1. Integrating by parts, derive reduction formulas for calcu-

lating the following integrals:
(a) ! = (b) ~m = dx;

»
I (** + a*)« ;
I
(c) /„= J
(a
2 —x 2
)
n
dx.
Solution, (a) We integrate by parts. Let us put
dv = dx,
(*
a
+ a 2\na
)
whence
du
,
= — r(x^-
2n x dx
—
-â n + 1 i
2 ir
,
9
v = x.
)
Hence,
/
(x 1 + a'
1
)
n
whence
n+ ~~ lna h
l
a1 n
1 '
{x 1 -f- a 1 )' 1
2n '
The obtained formula reduces the calculation

of the integral I n+1
to the calculation and, consequently, allows us
of the integral In
to calculate completely an integral with a natural index, since
/, = f 2 — 2
—
+ a = a arc tan a + C.
J x
'
For instance, putting n=l, we obtain
+ a + 2a? ai C ¥+C
*
~~ 2a 2 ~~ 2a 2 x* T an
2
J (** +a 2 2
)
'
a:
2
+a 2
2a 2 1 2
putting m = 2, we get
/ —
~~
C
dx —
~~
1 a: 3 ,
7 2 ~~~
_
3
J (*
2
+a 2 3
)
4a 2 "
(x 2 +a 2 2
)
4a 2
x x 3 ^
~~
1
4a 2" (x 2 +a 2 2
)
+ 8a* '
x^fa~2 + 8^.
alX 130 "7 + C i \ r*

"
(b) Let us apply the method of integration by parts, putting
w=sin'«-i1 x;
•
z
ay
j
= cos —a
sin *
—
* OT
,
a:,'
whence
du
da = (n — 1) sin" -2 x cos
v 7
a: d^:; v = (m— ,
1X
1)
1
m -1 #
cos™
. (m
v ^ 1)
Hence,
/ _ sin" -1 * n — 1 C sin"
si
-2 x dx
~~
(m — 1 ) cos"2
-1
x ~ m— 1
J ~ccos 7™
- #
_ sin"- 1 * — 1
m ^=1
(m— 1) cos* 1
-1
* m— 1 "" 2 a-*' ( /-
(c) Integrate by parts, putting

u = (a 2 — a;
2
)"; dv = dx,
whence
= — 2nx (a 2 — a:
2
)"
-1
dx; v = x.
Hence
j n=zX (a* — x + 2n $ x
2
)
— x )"- dx =
n 2
(tf
2 1
= x(a —x + 2n —a + a
2 2
)
n
J
(a:
2 2 2
) (a 2 — a;
2 )"" 1
dx =
= a: (a 2
—a; 2
)" — 2n/ + 2na I w
2
n_v
Wherefrom, reducing the similar terms, we obtain

(\+2n) In =x(a 2
—x 2
)
n
+ 2na 2
In _ 1 .
Hence,
j __ x(a
2 —
x2 ) n 2na 2 ,
7/1 ~ 2/z+l ~»2/i+l

7 »-i'
For instance, noting that

.
I- 1/2
_
= P
\ ^
ax
dx
r -n » " arC sin ~x + Cn .
>
we can find successively
= JV^"^ ^ = |(a -x 2 2 )i/2

+ ^/_ 1/2 ^
= -i[/5î + -ârcsin| + C,
= J(a 2
~A; 2 ) 3 / 2 ^= -J-(a
2
—* 2
)
3/ 2
+ -|-a 2
/ 1/2 , and so on.
4.4.2. Applying integration by parts, derive the following reduc-

tion formulas:
(a) /„= (\nx) n dx = x(\nx) n — nl n _

^
x \
J
*>"
(b) /„ = \ x« (In x)» dx = - /„ x
(a^-1);
(c) /„ = J
xn e x dx = x? ex —nl n -
l
1 ;
(d) /„= J^'sin"**^

= -n—oSin" ^(asinx — ncos^H —^
^- /„_,.
4.4.3. Derive the reduction formula for the integration of /„ =

= (-^-
sin"
and use it for calculating
to
the integral / 3a =f-^r-.
sin* x
J a: J
4.4.4. Derive the reduction formulas for the following integrals:
(a) In = J tan n xdx; (b) /„ = J

cot" xdx\
J V x*-\-a
C h a p t e r
BASIC CLASSES
OF INTEGRABLE FUNCTIONS
§ 5.1. Integration of Rational Functions

P (x)
If the denominator Q (x) of the proper rational fraction pry-( can
be represented in the following way:
Q {x) = {x—a) k {x — b) 1
. . . (x
2
+ ax+$) r
(x
2
+ yx + \i)
s
.
where the binomials and trinomials are different and, furthermore,

the trinomials have no real roots, then
P(x) Ax A2 Ak ~
~ ~ (x—a)*
. . .
Q (x) x—a 1
(x—a)* ' ' '
x— b 1
(x—b)* 1
* " *
M*— b) 1
1
* '
#
x2 + ax + p (*
2
+ cu;+p) 2 (jca + ouc + P)' 1
*a + V^+M- (*
2
+ Y*+M0 2
(*
a
+ Y*+H) 5
where
i4 lf i4„ . .
. ,
B 19 B„ . .
. ,
M N lf 19 7W 2 ,
N 2y . . . ,
/? lf L lf /? 2 , L2 ,
. . .
are some are determined by

real constants to be determined. They
reducing both sides of the above identity to integral form and
then equating the coefficients at equal powers of x, which gives
a system of linear equations with respect to the coefficients. (This
method is called the method of comparison of coefficients.) A system
of equations for the coefficients can also be obtained by substitu-

ting suitably chosen numerical values of x into both sides of the
identity. (This method is called the method of particular values.)
A successful combination of the indicated methods, prompted by
experience, often allows us to simplify the process of finding the
coefficients.
If the rational fraction is improper, the integral part should
first be singled out.

§5.1. Integration of Rational Functions 215
5.1.1.
, f* \5x 2* — 4x — oi
81
<ix
+ 4) (*-l)
c
Solution. The integrand is a proper rational fraction. Since all

roots of the denominator are real and simple, the integral will
appear in the form of the sum of three simple fractions of the form
15a:— Ax — 81
2 A . B . D
(x — 3) (x+4) (x— 1) X— 3 1
a+4 1
x— 1»
where A, B, D are the coefficients to be determined. Reducing the
fractions to a common denominator and then rejecting it, we obtain
the identity
15x 2 - 4x— 81 = A(x + 4) (x— + B {x—3) (x— + l) 1)
+ D(x-3)(x + 4). (*)
Comparing the coefficients at equal powers of x in both sides of

the identity, we get a system of equations for determining the
coefficients
A+B+D- 15; 3A — 4B + D = —4; — 4 A + 3B — 1 2D = — 8 1
Solving the system of equations we find A = 3, B = 5, D-— 7.

Hence,
= 31n|x — 3 + 5 In |
|
x + 4 + 7 In |*— + C =
|
1
|
= In |
(x—3) 3
(x +• 4)
fi
(x— I)
7
1 -f C.
AWe. Let us use the same example to demonstrate the applica-

tion of the method of particular values.
The identity (*) is true for any value of x. Therefore, setting
three arbitrary particular values, we obtain three equations for
determining the three undetermined coefficients. It is most conve-
nient to choose the roots of the denominator as the values of x t
since they nullify some factors. Putting x 3 in the identity (*), =

we get A =3; putting x = — 4, we obtain B^=5; and putting x-=l,
we get D = 7.
5-1.2. / =j (2 + x)(S-ir
Xi 2
5.1.3. I =$ -<Z7X dx -
Solution. Since of the numerator is higher than that

the power
of the the fraction is improper, we have to single
denominator, i.e.
out the integral part. Dividing the numerator by the denominator,

216 Ch. V. Basic Classes of Integrable Functions
we obtain
x* — 3x — 3*— 2
2
,
t
x +2
x — x — 2x
3 2 1
x(a: 2 — x — 2)"
Hence,
, rx* — 3x*—3x—2, p, .
1W p (x + 2)d*
Expand the remaining proper fraction into simple ones:

x+2 _ A B D
*(*— 2)(x+l)~~ x — 2 + a:+1*
Hence
+ 2 = A(x—2) (x + l) + Bx (x+ + Dx (x—2).
x l)
Substituting in turn the values ^ = 0, x = 2, x = — (the 2 3

1 roots
of the denominator) into both sides of the equality, we obtain
/4= _l; fi= |; D= |.
And so
= ^ + * + ln|*| — -|ln|x— 2|— yln|x+l| + C.

5,4 I =) ^-2 fl+x 3
dx -
Solution. Here the integrand is a proper rational fraction, whose

denominator roots are real but some of them are multiple:
x3 — 2x 2
+ x=x(x—l) 2
.
Hence, the expansion into partial fractions has the form

2x* — 3*-f3 _ A B D
x* — 2x'>+x~~~2
x
~"
(x— l) 2 + x— 1
'
whence we get the identity:
2x 2
— 3* + 3 == A (x— l)
2
+ Bx+Dx(x—l) =
^(A + D)x + (—2 A — D + B)x + A.
2
(*)
Equating the coefficients at equal powers of x we get a system

of equations for determining the coefficients A 6, D: y
A+ D = 2; —2A — D+B=—3\ A=3.

Whence A=3; B = 2; D = — 1.
Thus,
/=3j? + 2j 7 ^-J^ = 31n|,|-^T-ln|,-H + T

C.
§ 5.1. Integration of Rational Functions 217
Note. The coefficients can be determined in a somewhat simpler

way if in the identity (*) we put ^=0; x2 = 1 (the denominator
roots), and x 3 equal to any arbitrary value.
At x=0 we get 3=A\ at #=1 we will have 2-=B; at x =2
we obtain 5-4 + 2B + 2D; 5 = 3 + 4 + 2D; whence D = — 1.
5.1.6.
Solution. Since x 3 1 (x 1) (a:

2
+ = +
1) (the second factor is — x+
not expanded into real multipliers of the first power), the expan-
sion of the given fraction will have the form
x _ A Bx-{-D
X3 + 1 X+\ X2 —# + 1
'
Hence,
x = A(x 2
— x + + (Bx + D) (x+l) =
l)
= (A + B)x + (—A + B + D)x + (A + D).

2
Equating the coefficients at equal powers of x, we get
A=
-T> fi = T' D = T-
Thus,
To calculate the integral
let us take the perfect square out of the denominator:

3
and make the substitution x —\ = t. Then
j t * ^ x tdt ,
3 V d/
lln(/*+4) + K3 arctan^L + C.
/1 = jln(x 2
— x+l) + K3" arctanî + C.
Thus,
= — y ln|* + 1
| + -g- ln(x 2 — *+l) + X^-arctan -y~ 2
+C.
Solution. The denominator has two pairs of different conjugate

complex roots, therefore
l _ Ax + B + Dx + E
(jc*+1)(jc 2 + 4)~~ *a +4 '
hence
l=(Ax + B) {x 2
+ 4) + (Dx+E) (x + 2
1).
Here it is convenient to apply the method of particular values for

determining the coefficients, since the complex roots of the deno-
minator (# = ±i and x = ±2i) are sufficiently simple.
Putting x = i, we obtain
3fl+3>U = l,
whence ,4=0, B =^ - Putting x = 2i 9

we obtain —3E — 6Di = 1,
whence D = 0, E=— . Thus,

r dx l r i_ c dx _
""
2
J (x +1)(a:
2
+ 4)~~ 3 J x2 +l 3 J x2 +4
=
y arc tan x — -g- arc tan + C.
(x + 1 ) dx
5.1.8. / ~ i (* + x + 2) (** + 4* + 5)'
2
x + 4x +llx +12x-f-8
' 4 3 2
(* + 2x + 3) (x+l)
2 2
J
Solution, Here we already have multiple complex roots. Expand
the fraction into partial fractions:
^4 + 4a: 3_|_ 11 ^2 + 1 2^ +8_ Ax+B Dx + E f_
(x 2 + 2x + 3) 2
(x + 1 )
*~ (a;
2
+ 2x + 3) 2 '
x2 + 2* + 3 + * + 1
'
Find the coefficients:
-4=1; B = —U D = 0; £ = 0; f=L
Hence,
^4 + 4a: +1U +12a; + 8-dx-
3 2
(jt
2
+ 2x + 3) (x+\) 2
§ 5.2. Integration of Certain Irrational Expressions 219
=
Calculate /,
j (
,2 / ^
= (x+
2 3)2
dx.
Since x2 + 2x + 3 l)
2
+ 2, let us make the substitution
x-\- 1 Then we obtain
I i=\ dt = dt ~2 I V = ~~ 2/ 2'
(/2 +2 )
2
^ 2
+ 2) 2
2 (/a
+ 2)
The integral
2 2 2
J(/ +2)
is calculated by the reduction formula (see Problem 4.4.1):
.
_ l / l r dt _ l t l , t
r
Thus
^r
*
/ } L_ ! o rn for,
i-
i
7
2
2(/ +2) 2(/>+2) 2}A2 f2
/ £±J !_arctan^tl-J C
l
i~ + 2a;+3)
!
+ 2a;+3)
2(x 2 2(a: 2 2 V2 ^2
We finally obtain
J (*2 + 2*+3) (a:+1) 2
2
= ln|x + 11— 0/ tan^y + C.
YU^2
o*,"^ arc
2 + + 3) 1 1
(a:
2
2a:
i ox
2 V^2
Find the following integrals:
5#1 10 . .
j
^"^^ — 22a; — 8 ^>
5x 3 -j-9x 2
3 r;;
-4a
—
dx
5.1.11.
(jc+1)(a:+2) 2 (jc + 3)»'
5.1.12.
4a:+4) (a:
2 — 4a: +5)
dx
5.1.13.
(1+*)(1+* > )(1+* 8 )'
x +3 3
5 ' U4 IcT+W+T)^'
'
£ 5.2. Integration of Certain Irrational Expressions
Certain types of integrals of algebraic irrational expressions can

be reduced to integrals of rational functions by an appropriate change
of the variable. Such transformation of an integral is called its ra-
tionalization.
I. If the integrand is a rational function of fractional powers of

( p-± 2l\
an independent variable x, i.e. the function R\x x«*, x^j 9 ,
then the integral can be rationalized by the substitution x=t m ,
where m is the least common multiple of the numbers q lf q 2 ., q k , . . .
II. If the integrand is a rational function of x and fractional
powers of a linear fractional function of the form then ra-
tionalization of the integral is effected by the substitution

ax b
~t
cx-\-d
,
= t
m
,
'
where m has the same sense as above.
5.2.1. l=\ x+ Y*+¥~x dx.
Solution. The least common multiple of the numbers 3 and 6 is 6,
therefore we make the substitution:
x = t*, dx^Wdt,
whence
= 6p 3
d/ + 6 J^^7 = y/ + 6arctan/ + C. 4
2
# 3
+6 arc tan 1/ x+C.
5.2.2. 7 = 1 f*+K* dx.
5.2.3. / = t j
(2x —
3
3) +1
Solution. The integrand is a rational function of £/2x — 3, the-
refore we put 3 t
Q
2x— = y
whence
—
dx = 3t b dt; (2x — 3) _L
T= t
3
\
(2x— 3) 3 =/ 2 .
Hence,
1+/ 2
«3-y~ 3-J- + 3-^ — 3^ + 3 arc tan + C. /
§ 5.2. Integration of Certain Irrational Expressions 221
Returning to x, we get
-6 — - -
y (2x— 3) —
11
/ =3 I
-J-
(2x— 3) -i-
l
(2x— 3) 6
l
2
— (2x— 3) 6"+ arc tan (2x— 3) T + C ]

*
5.2.4. /= f- 3
/•"
Solution. The integrand is a rational function of x and the ex-
2qr^» therefore let us introduce the substitution
1
K 2 +x 2 +x '
whence
Hence
j
7 ~~
P 2(l+/ 3 )
2 />12/ 2 ,
ai ___3_
~~ " r
2j^""4/ 2 t U#
A " i
J 16/
6
(1 + '
3 2
)
Returning to x 9
we get
'=4K(W+ C -
5.2.6. /= (*__=£_.
Solution. Since
/(x-lH* + 2) = (x-l)(x + 2) ]/ £±2, 5
jZTj
-
; therefore let
us introduce the substitution:
x-\
~ l
K %-l 9
whence
_ tj—
a:—
/
4
+2 .
a: — —_ 1
1
, 4
3
,*
.
#-f-z
i _
— 3/4 .
, _
—
-12/ a ^
(/4 1)2
Hence,
~ _ Jp
(/4-l)(/4— 1) \2t*dt 4 C dt__±_i_ r
l
3.3^(^ 4 -D 2 "~ 3 J /2-3/ + Lt
Returning to x> we obtain
dx
5.2.7. f
J (1- x) V 1— * 2
d*
5.2.8
J 7^=
t+1) 2
(x — l)
4
5.2.9.
j (,-2) /}±£xx dx.
§ 5.3. Euler's Substitutions
Integrals of the form \ R(x y

[/ ax 2 -\- bx + c) dx are calculated with
the aid of one of the three Euler substitutions:
(1) V ax* + bx + c = t±xVa if a> 0;

(2) V~ax 2
+ bx + c=tx± ]/"c if c> 0;
(3) V~ax 2 + bx-\-c = (x— a)t if
ax 2
+ bx + c = a(x — a) (x— f$),
2
i.e. if a is a real root of the trinomial ax -\-bx-\-c.
5.3.1. /=f / x
Solution. Here a=l > 0, therefore we make the substitution
Vx + 2x-\-2 = — x. 2
t
Squaring both sides of this equality and reducing the similar terms,
we get
2x + 2tx = — 2, t
2
whence
_ t
2 —2 ,
_ + 2/ + 2
/
2
*~~2(1 +/) ;
2(l + 2 '
1
+ ^+ 2*
+ 2 = + <-^=%T7F 1 *
Substituting into the integral, we obtain

. P 2(l+<)(< 2 + 2/ + 2) _ f
(<
2
+ 2<+2)rf< '
J «
2
+ 4i + 4)2(l+0 2
J (l+0(<+2) 3
§ 5.3. Euler's Substitutions 223
Now let us expand the obtained proper rational fraction into par-
tial fractions:
/
2
+ 2/ + 2 _ A B D
(t + 1 ) + 2) ~~ + + + 2 + + 2)
(/
2
t I / (t
2 '
Applying the method of undetermined coefficients we find: i4=l,

B = 0, D = — 2.
Hence,
p r2 + 2/ -f- Ai _ p d/ Q p d/ _ 1 i y i iii 2 , ^>
2
/ = In (x + + j/> + 2x + 2) H x
1
r + C. _
+ 2+ y x* + 2x + 2
5.3.2. /= f *
c+ V * 2 — x-\- 1
Solution. Since here c=l>0, we can apply the second Euler

substitution
Vx 2
— x+ 1 = tx— 1,
whence
(2/-l)x = (/
2
-l)x 2
; * = ^rr;
^=-2^^^; x + Kx'-x+^y^y.
Substituting into /, we obtain an integral of a rational fraction:
r — 2; + 2/— 2
J /(/ — !)(/ + I)
2 '
-2/ 2 + 2/ — 2 i4 . B . D . E
t(t—\)(t+\) 2 t
1
/ — 1
1
(t + \)
2 1
* +
By the method of undetermined coefficients we find
A = 2; B = -y; Z> = -3; £ = - |.

Hence
Cdt_ C dt dl 3 P
7
,
- Z9 1_ o C
J / 2 J/-1 ^J(/+l) 2 2 Ji/ +r
= 21n|/|—[ln|/-l + I
In + l
\ +C 9
where / =
)/"x a — x+l + l
x
d*
5.3.3. /=f (1+x) ^l + x— a;
2
224 Ch. V. Basic Clastes of Integrable Functions
x dx
5.3.4. / = f-
(]ffx—\0 — x*) 3 '
Solution. In this case a and c therefore neither the first, < <
nor the second Euler substitution is applicable. But the quadratic
trinomial 7x 10 —
x 2 has real roots a 2, P 5, therefore we = =
use the third Euler substitution:
y'7x—\0—x = V(x— 2) (5 — *) = (x—2)t. 2
Whence
5— 2) t
%
\
_ +
X—
5 2/ 2
ax _
6/ dt
™
,
+ .
1
i v2
/
2 »
'
(1 + /
2 2
)
•
Hence
= + 2)*--»(-f + a
,
.«^--»fl| ) +C .
where /-
. — — 1/"7jc
^ 10 jc
2
.
Calculate the following integrals with the aid of one of the Euler
substitutions:
dx
5.3.5. f j±2
J x— Vx + 2jc + 4
dx
5.3.6.
J Y — — 1 AT
2
1
dx
5.3.7.
j ]/"(2x— 2 3
)
5.3.8.
f
^+.GI?"
yT+x
^ 2
§ O^ter Methods of Integrating Irrational Expressions

The Euler substitutions often lead to rather cumbersome calcu-
lations, should be applied only when it is difficult
therefore they
to find another method for calculating a given integral. For calcu-
lating many integrals of the form
J
R(x, Vax 1
+ bx + c) dx,
simpler methods are used.
I. Integrals of the form
I=f Vax 2 + bx+c

dx
J
§ 5.4. Methods of Integrating Irrational Expressions 225
are reduced by the substitution x + -^ = t to the form
*J Vat'+K *J Vat* + K
where M N K
l9 ly
are new coefficients.
The first reduced to the integral of a power function,
integral is
while the second, being a tabular one, is reduced to a logarithm

(for a >
0) or to an arc sine (for a 0, 0). < K>
II. Integrals of the form
P- W dx,
J Vax 2 + bx + c
where P m (*) is a polynomial of degree m, are calculated by the
reduction formula:
where
L
J
p
KaA: 2
P m _ 1 (x) is
:?Tl-
+ ^+c =v
a
1
polynomial
(x)^Hw/([-7
of degree m
J
—
^ V ax
1,
z
and
+r
K
=-,
bx + c
is
(i)
some
constant number.
The coefficients of the polynomial P m (x) and the constant ^.
1
number K are determined by the method of undetermined coeffici-

ents.
III. Integrals of the form
dx
(x—ai)" 1
Vax 1 -\-bx-\~c
are reduced to the preceding type by the substitution
x-a = T
1
.
x
IV. For trigonometric and hyperbolic substitutions see § 5.7.
5.4.1. 7 =f .
J V 4a:
2
+ 4*—
Solution. Make the substitution 2#+l = /, whence
/ — 1
— -^dt.
, dx
2
Hence,
1
4
CiLh5)dt_ 1 — + 14ln|/ + ^
^-4
./-tj 2 i 2
-4| + C.
J
I =ly^ + 4x—3 + ^\n\2x+l + [

r 4xi
+ 4x—3\ + C.
226 Ch. V. Basic Classes of Integrable F uncttons
5X
5.4.2. /= f +- H*.
J ^x + 2x+5
2
5.4.3. /== f f-x-zjL^dx.

+2x+2 2
Solution. Here P m (x) = x 3 — x — 1. Hence,

P m - (x) = Ax + Bx + D.
1
2
We seek the integral in the form
/ = (
A x> + Bx + D) Vx 2 + 2x + 2 + K f -= dx
j y * 2 + 2* +
Differentiating this equality, we obtain
/'
— x— l
r
V x2 + +2 2x
= (2 Ax + B) Vr x* + 2x + 2 + (Ax + Bx + D) 2
4
y x' ~\~ 2x ~ j 2
*
+ Vx' #
1
+ 2x+2
Reduce to a common denominator and equate the numerators
x* — x— = (2Ax + B) (x + 2*+ 2) + (/l* +

1
2 2
+ D) (* -f 1) + K.
Equating the coefficients at equal powers of x, we get the following
system of equations:
2A + A=1, B + 4A + B + A=0;
2B + 4A + D + B = —l\ 2B + D + K = — 1.
Solving the system, we obtain
Thus,
dx
*
+ 2x + 2
where
1
J Vx 2 + 2x+2 J 1)2+1
v
^ -r /-r
5.4.4. /= JjAx 2
— 4x + 3dA:.
Solution. Transform the integral to the form
/=[ Jf- *±^ dx = (Ax + B)V4x -4x + 3 + K

4
2
f
^
A
§ 5.4. Methods of Integrating Irrational Expressions 227
Applying the method of undetermined coefficients, we get
V2 4f J V(2x-\\> + 2
= )
]/4x 2 — 4x+3 + y In (2x — 1 -|- j/4j^ — 4x+3) + C.
9' 3 -3' 2 +2 -
5.4.5. f rf
J K3x 2 — 2x+l
5.4.6. 1 dx.
(*+4)dx
5.4.7. =
/- r
1 (f±
J (a:— 1) (a:+2) 2 +
Solution. Represent the given integral as follows:
(x-{-4)dx C x+4 dx
(a:+2) 2
(x-l)(x + 2) 2 j7> + jc+1 J
(at— 1)
yx + x+{
2
Expand the fraction ?

\X
. T\X -\- 9 2
Z)
into partial fractions
1 )
x+4 A . B . D
(x— l)(* + 2) 2
x—\ 1
(x + 2) 2 1
x +2
Find the coefficients
Hence,
5 5 dx
1 = 2 1
9(x-\) 3(x+2) 2 9 (a:+2)J

y + x+
xt \
dx 2 f dx
(x— 1) + 1
3
J (x+2) 2 |A^-f-x-fl
dx
_ "
5_ f
9
J + 2) /x + x-H 2
The first integral is calculated by the substitution x— 1 ==— , the
second and the third by the substitution x +2= -j-'
We leave the solution to the reader.
f*
3 - 6**+H*- 6 ^
5.4.8.
J W+4x+3
f 3^ + 5^-7,+ 9 Wt ,
5.4.9.
J |^2a; 2 + 5a: +7
1
xdx
5.4.11
I (x 2 — 3x + 2) Vx 2 — 4* + 3
r dx
5.4.12. "
J (^+l) 3
/^ + 37+2
{ l)dx
x Vf-
5.4.13. r .
J l+3x + 2
;c
4
§ 5.5. Integration of a Binomial Differential
The integral ^ xm (a + bx n y dx, where m, n, p are rational num-

bers, is expressed through elementary functions only in the follow-
ing three cases:
Case I. p is an integer. Then, if p > 0, the integrand is expanded
by the formula of the Newton binomial; but if p < 0, then we
put x=t k where k is the common denominator of the fractions m
y
and n.
Case II. is an integer. We put a + bx n = t
a
, where a is
the denominator of the fraction p.
Case III. ^-^- + p is an integer. We put a + bx n = t«x n , where a

is the denominator of the fraction p.
5.5.1. / = $
j/x(2+j/"x) 2 dx.
\2+x-V
= r -(
3 2
Here = 2, an
Solution. I
J
x y dx. /? i.e. integer; hence,
we have Case I.
+ 4xT +4) dx = J (x T + 4x T +4xT ) dx =

*3"
/ = (jc
J
3 - 24 ii -
2 \ -1
5.5.2. / = J* 3
G+* 3
) dx.
C-/T+V*
5.5.3. / = \
*
, J_ dx.
J
Solution. I =
J
( 1 + xT ) T d.*:.
m+l = (-4+0 =
Here m = — j2 .
; n = j\ 1
p = -^ 1
; 1,
{
i.e. an
3"
integer.
§ 5.5. Integration of a Binomial Differential 229
We have Case II. Let us make the substitution
l+x* =f« ;
±x 3
dx = 2tdt.
Hence,
/ =6§ t
2
dt = 2t + C = 2(1 +xT ) T + C.
3
= T ) Tdx.
J*~(2 +
5.5.4. / JK
5.5.5. [ = i
\
x b {\ +x 2 )~ dx.
j
5.5.6. / = jV 11
(\ + x*)~dx.
= — -j a fraction, = + = — also
Solution. Here p y is
a fraction, but +p = y= ---3 an integer, i.e. we have is
Case III. We put + x* = xH Hence 1

2
.
tdt
x— dx — 1 ,
— 2(f — 2 4 2 4
(t l) I)
Substituting these expressions into the integral, we obtain
= _^ ^_i)M/=-^+|-4+c.
(
3 r~
5.5.7. \ jj- dx.
dx.
5.5.8. j-
2
5.5.9. (!+**) dx.
J*»
230 Ch. V . Basic Classes of Inlegrable Functions
5.5.11. \Vx Y^ + V x * dx -
§ 5.6. Integration of Trigonometric and Hyperbolic

Functions
1. Integrals of the form
I = ^ s'm m xcos n xdx,

where m
and n are rational numbers, are reduced to the integral
of the binomial differential
n- 1
/=$/*(l— /*) 2
dt 9
t = smx
and are, therefore, integrated in elementary functions only in the
following three cases:
(1) n is odd {j-^- an integer^,
(2) m is odd (^rp an integer^,
(3) m+n is even ^^y-^ + ^p an integer^.
If an odd number, the substitution sin#=/ is applied.

is
If m
an odd number, the substitution cosa;=/ is applied.
is
If the sum m + n is an even number, use the substitution
tan#=/ (or cot^=/).
In particular, this kind of substitution is convenient for integrals
of the form
^tan n xdx (or ^cot n xdx^>
where n is a positive integer. But the last substitution is inconve-
nient both m and n are positive numbers. If m and n are non-

if
negative even numbers, then it appears more convenient to use the

method of reducing the power with the aid of trigonometric trans-
formations:
cos 2 x=y(l + cos 2x) t
sin 2 A; = y (1 cos2jc)
or sin x cos x =y sin 2x.

s
5.6.1. 1= f , iilli- dx.
J j/cos**
§ 5.6. Integration of Trigonom. and Hyperbolic Functus 231
Solution. Here m = 3 is an odd number. We put cosa: = /,

sin xdx = — dt, which gives
=3 j/cos a:
(
y cos 2
jc— 1 ) +C .
c
5.6.2. / -^d*.
=fJ sin x 6
5.6.3. / — ^ sin 4 x cos 6 a: c(a:.
Solution. Here both m

and n are positive even numbers. Let us
use the method of reducing the power:
/ =~ (2 sin a: cos a:)

4
cos 2 xdx = — sin 4 2jc(1 + cos 2x) dx = lx +/ 2 .
J J
The second of the obtained integrals is calculated by the substitu-
tion:
sin 2x =t, cos 2a: dx = -^dt,
/, = gg Jsin 4
2a: cos 2x dx t* dt ==^ + C ==^sm* 2x + Q.
We again apply to the first integral the method of reducing the
power:
Ii = 52§ sin4 2x dx = y§§ J —cos 4a:) dx= (

1
2
J_ x — y sin 4#j
128 +2§6 J + cos 8a:)c(a: =
:
A*-i sin4 *+^ sin8 *+ c -
And so, finally,
/__L *^ —
256
^
256
sin
0111
™+
^ 2048 sin
4a: ^
^" + ^
" 320 sin 8a: 1
5
2a: + 0.
5.6.4. /
r
= \ —
2
r sin x
g-dA:.
6
J cos *
«
Solution. Here both m and n are even numbers, but one of them
is negative. Therefore, we put
tan a:= /; —^— = +
cos 2 a;
1
1
1
2
; T~ti==dt.
cos 2 a;
Hence,
Q
5.6.5. /= [ ^dx.
2
J sin x
Solution. Here we can put cotx=t, but it is simpler to integrate
by expansion:
sin 2 * 2 1
J J Vsin x J
= — cot* 2a; -f y j" (1 — cos2x)dx =

cotx +
. .
—r + -j\
3x\
sin 2x ~
+ C.
-
.
,
d*
5.6.6. / = cos 4
J a:'
5.6.7. /
-J jj/sin 11 x cos a:
Solution. Here both exponents (~ y and — y) are negative
numbers and their sum — ~—y = — 4 is an even number, there-
fore we put
dx
tan x=t; ?
= dt.
4 11 11
J cos * j/tan * J j//
3 (1 + 4 tan 2
x)
•C.
2
8 tan a: ^/tan^Sc
5.6.8. Find the integrals of tanx and cot x.
Solution.
\ tanxdx= \ dx = — In |
cos x -)- C;
J COS
|
J AT
\ cot x dx — \
sin
dx = In |
sin x |
+ C.
J <J
at
5.6.9. /=$ ian xdx. 1
Solution. We put tanx=tf, x = arctan?; dx = We get
/ /5
= -' , , =
=I TT7-. K' + '-TTT.)*
= -jr tan 6 a:— tan 4 x + y tan 2

x+ In |
cos x \
+ C.
§ 5.6. Integration of Trigonom. and Hyperbolic Fund's 233
5.6.10. (a) /= ^cot«xdx; (b) / = $ tan 3 xdx.
5.6.11. / = C£2^fdjc.
sin * 3
J
Solution. Here sin a; is raised to an odd power. Let us put

cosa: = t y
— sin xdx = dt.
We obtain an integral of a rational function.
/
4
-dt.
Here, it is simpler to integrate by parts than to use the general

methods of integration of rational functions (cf. Problem 4.4.1 (b)).
Let us put
Then
du = 3t 2 dt\ v — 2(1— 1
2
)
Hence,
/ = t* r t
2
dt _
2(1— 2
)
t* i + i
dt:
2(1— P) 1
2 J 1— t*
t
3
3,3.
+ ln I _±i l
"2(1 — 2
) 2
1
4 |
\—t
cos x 3
3 + cos * 1
-cosx + T ln — I
+c.
5.6.12. /= f
J COS*
—
dx.
2 COS X 1
II. Integrals of the form i?(sin.x:, cos x) dx where /? is a rational

J
function of sin x and cos x are transformed into integrals of a rational
function by the substitution:
tan l^-jj =t (— ji <x< n).
This is so-called universal substitution. In this case
sin x
2t
cos x-
l — t
2
IT/*' 1 + /
2 '
2dt
x = 2arc tanf; dx =
i+t 2
'
Sometimes instead of the substitution tany = / it is more advan-
tageous to make the substitution cot-|- = / (0 <x< 2n).
Universal substitution often leads to very cumbersome calculations.

Indicated below are the cases when the aim can be achieved with
the aid of simpler substitutions:
(a) if the equality
R ( — sin a;, cosjc) = — R(s'mx t

cos x)
or
R(s\nx, — cos x)== — R(s\nx y
cosx)
is satisfied, then it is more advantageous to apply the substitution

cosx = t to the former equality, and s'mx = t to the latter;
(b) if the equality
R( — sin — cos x) = R (sin

x, x, cosx)
is fulfilled, then a better effect is gained by substituting tanx = t
or cot x= t.
The latter case is encountered, for example, in integrals of the

form J
R (tanx)dx.
5.6.13. /= —
/0
J sin x (2+ cos
.
dx
x—2sj^— r
sin x)
.
Solution. Let us put tan~ = t; then we have
2dt
Expand into simple fractions

l + /
2
A . B D
t(t — 3)(t—\) —3 t—\
1 1
t /
A
=b B =ih D =- 1 -
Hence
Jt_ d£ f dt C dt
J ~ 3 J
f
t +
,
_5_
3 J t-3 J
= -iln|/| + |ln|/ — 3 1— ln| — + C = / 1

\
T + |-ln|tan|— 3| — In
X
T ln |
tan tan-
2
+ C.
§ 5.6. Integration of Trigonom. and Hyperbolic Funcfs 235
dx
5.6.14. /:
5+ sin x + 3 cos x
dx
5.6.15. /:
sin a; (2 cos 2 £— 1) *
1
Solution. If in the expression
cos 2 x — we substitute
sin x (2 1)
—sin x for sinx, then the fraction will change its sign. Hence, we
take advantage of the substitution / cosa;; dt sinxdx. This = =—
gives
dl
J (
2/2_ 1}
'
Since
1 (2 — 2/ — (1— 2/
2
)
2
) 2 1
(1— 2 )(\— 2t 2 )'

1 — )(1— 2<
(1 <
2 2
) 1 — 2/ 2 1— 2
then
— +c=
dt dt
In In
1— 2/ 2 1-/1^2 2 l t
i+ VT COS JC ,
1 , II — COS X
1-^2 "•"T + COS X I
1
= -£=ln 1+^2
1
— cos #
+ ln tan
I
1 Y^2
5.6.16. /
_ C sin 2 # cos #
dx.
J sin #:+ cos #
Solution. Since the integrand does not change sign when sin*
and cos x do change their signs, we take advantage of the substi-
tution
t = tan*; dt = COSdx X
"
2
'
Hence,
7r
_ C tan 2 X cos 4 x m
dx _ C t
2
dt
J (tanx+l) cos 2 * (/o_iw/2_

J (*+l)(/ +l)
2 2
Expand into partial fractions
/* A Bt +D Et +F
(/+l)(/a + 1)
2-
t+\ ^ t
2
+\ M/ 2 +l) 2 *
B= — 1
D-- =
1-
2 '
F=-± 2
Hence,
<— 1
=11 '+1 dt+

4 lt 2 +\ 2
236 Ch. V. Basic Classes of fntegrable Functions
/ »
4
In -J±L-4
^ + 1 ^2 4
•
^+
1 + /
2
C=
= In I
sin a; 4- cos x |
— -|- cos x(smx + cos + C.
2 tan ^ + 3
5.6.17. /= f sin
t
+2 cos si 2 2
^
J a: a:
Solution. Dividing the numerator and denominator by cos 2 jc and

substituting tanx=/;
dx
2
dt. we obtain —— =
cos AT
(2 tan x + 3)
^_
.
r 2tan*+3 C cos 2 x
'
J sin
2
x+ 2 cos 2
a:
j tan 2 x+2
j|J|^ = ln(^ + 2) + -ârctan T + C

=
i= =
= In (tan * + 2) + -2= arc tan tan *
2
5.6.18. /= [j^-dx.
J + sin x 1
Solution. This integral, of course, can be evaluated with the aid
of the universal substitution tany = /, but it is easier to get the
desired result by resorting to the following transformation of the

integrand:
sin x sin#(l — sin x) sin — sin x)
1-J-sin* (1 -j- sin a:) — sin x)

(1 cos 2 x
sin x sin'2 x sin x
-tan 2 x.
COS 2 X COS 2 X COS 2 X
Whence
/ = C
S1 " *
9
2
J COS X
dx — J[ sec 2
x dx + Jf dx = —
1
COS
?
AT
tan a: + # + C.
1 1
5.6.19. /= f 4 2
1
. dx.
J cos x sin a:
the substitution tanx^/ can be applied, but it

Solution. Here
is simpler to transform the integrand. Replacing, in the numerator,
unity by the trigonometric identity raised to the second power, we
get
2 4
C (sin cos 2 x) 2 2 2 4
C sin #-f- 2 sin x cos x-f cos x
r
— 4 ~~ cl.s 4
^ ~~
J cos x sin'2
x J * sin 2 x
= [
Sm
4
J cos x
/
dx-\-2 2
J cos *
1
C— — h C
1 f
2
J sin *
-
d
=
J
ftan 2 x ^ + 2 tan — cot x =
cos 2 a:
1
a:
= y tan 3
a: + 2 tanx — cot x-\-C.
§ 5.7. Integrating Irrational Functions by Substitutions 237
III. Integration of hyperbolic functions. Functions rationally de-

pending on hyperbolic functions are integrated in the same way as
trigonometric functions.
Keep in mind the following basic formulas:
cosh 2
x — sinh x =2
1 ; sinh 2 x = y (cosh 2x — 1 );
cosh 2 x = y (cosh 2x + 1); sinh x cosh x = y sinh 2x.

X 2t ,^2
If tanh ~2=ty then sinh x = _ ] (2
; cosh x = ^_ 1
1
t2
\
x = 2 Artanh/ = ln({-±j) (-1< < * 1); dx =

5.6.20. /= J
cosh 2 xck.
Solution.
I = j y (cosh2*+ 1) d*=|-sinh2* + y.x: + C.
5.6.21. /= J
cosh 3 ^d^.
Solution. Since cosh x is raised to an odd power, we put sinhjc = /;

co$hxdx = dt. We obtain
/ = j cosh 2
* cosh xdx - j(l +/ 2
) d/ = +y + C =
*
= sinh x + y sinh 3
a: + C.
5.6.22. Find the integrals:
(a)
v '
f sinh
J
2
# cosh 2 x dx; »
(b)
v /
C-r-r
J sinh
—x+2^— — cosh
r
a;
£ 5.7. Integration of Certain Irrational Functions with the

Aid of Trigonometric or Hyperbolic Substitutions
Integration of functions rationally depending on x and \âx 2 bx c + +

can be reduced to finding integrals of one of the following forms:
I. $/? (ty P H* + q*)dt-

II. \R(t, ]fpH* — q*) dt;
III. J'/?(f, dt 9
238 Ch. V. Basic Classes of fntegrable Fund ions
where t = x + ^\ ax 2 + bx + c = ± p ± q 2 2
t
2
(singling out a perfect
square).
Integralsof the forms I to III can be reduced to integrals of
expressions rational with respect to sine or cosine (ordinary or hy-
perbolic) by means of the following substitutions:
I. / = — tan 2 or t-= — sinh2.

P P
II. t = — sec2 or t = — cosh 2.

p p
III. t = — sin2 or t = — tanhz.

p P
dx
5.7.1. I=[-r .
Solution. 5 + 2x + x = 4 + (x+ 1)
2 2
. Let us put x+l=t. Then
/_ - j" (4+/
1 7(1+2^+^)3 2)3
We have obtained an integral of the form I. Let us introduce

the substitution:
* = 2tan2; d*=-^-; /4+7~ = 2/1 + tan 2 2

z
cos z
We get
cos zdz =
t
1 • ^ , ^ 1 tan z .
^ 1 2 .
^
+ x 1
-C.
4 y5 + 2* + jc
a
dx
5.7.2. /=f
J (*+l)» V*a+2*+2
Solution. x 2
+ 2x + 2 = (x+l) + L 2
Let us put #+1=/; then

dt
Again we have an integral of the form I. Make the substitution

tf = sinhz. Then
dt = cosh z dz\ Vt +\ = Kl+sinh 2 2

2 = cosh z.
§ 5.7 . Integrating Irrational Functions by Substitutions 239
Hence,
,
/ = C
\
cosh z dz
r —= C
\ U92
.
dz
= — C0th Z + C^ = ,
J sinh- z cosh z J sinh z
V l+sinh2 2 VT+Zi/- — ^+ 2
2jc +2 , ^>
5.7.3. I = ^x Vx^-ldx. 2
5.7.4. /-j^d*.
5.7.5. / = $K(* 2
— l)
3
d*.
Solution. Perform the substitution:
x = cosh?; dx = sinh/d^.
Hence
/ = $j/"(cQsh*/ — l)
3
sinh t dt=\ sinh 4 / =
= -i j cosh 2
2/ df —i j cosh 2/ df + -J- J
d/ =
= y j(cosh«+ l)df — -J-
sinh 2/ + -J-
^ =
= ± sinh At — \ sinh 2t + 1- + C. 1
Let us return to x:
= Arcosh x = In (x + Vx* —
t 1 );
sinh 2t = 2 sinh cosh = 2x V'x — l; t /

2
sinh 4/ = 2 sinh 2* cosh 2t = 4* Vx —\ (2x

2 2
— 1).
Hence
5.7.6. /=f * _
J (1+ V x)Vx-x*
Solution. We make the substitution:
x = sin 2
/; dx = 2sin*cos/df
and get
j _ r
f 2 sin / cos t dt f 2 dt
J (1 + sinO Vsin 2 ^ — sin 4

^ J 1 + sin /
=2 [tz™li
2
dt =2 tan
cos
t
COS
^- + C '
=
J ^ t
7 i =+ c-
2(l
CLi 1)
+c.
5.7.7. / = [VZ — 2x— 2

dx.
r ~> o / r
5 ,.s.
,=j 3
(x 2 — 2*+5) 2
£ 5.5. Integration of Other Transcendental Functions
5.8.1. / = Ji2£djc.
Solution. We integrate by parts, putting
w = In x\ *
dv = ^;
x 2 *
, dx 1
•
a: a:
,
7
_
-
In *
T + J"^-
,
C dx In
a:
x
—T + ^ 1
5.8.2. /=fi^.
Solution. Let us put: = e*dx = dt. We get:
j (i +<*)*
Apply the reduction formula (see Problem 4.4.1):
' = YiJ^T) + T arctan + C = 2(1$^ + T arctan + c t -
5.8.4. / = $e-*ln(e*+l)dx:.
§ 5.8. Integration of Other Transcendental Functions 241
Solution. We integrate by parts:
u = \n(ex + 1); dv = e~ x dx;

du= +f *e xY dx;
t
'
v = e~ x \
*
1
/ = - e -ln(l+^) +
JT
^=- e -Mn(l+^) +
j
-^±i^^ =
e
= —e - * In (1 +ex + x — ln(l + e*) + C.

)
~aarc tan *
5.8.5. / = l- — dx.
J (1 +* 2
)
2
Pa: arc tan x dx

g g g ^
J jn~+T2
Solution. Integrating by parts, we get
xdx
u = arc tan x; dv =
V i +**
'
= T+x|/
2
arc tan x — In (x + |/* + l) + C.
a
§ 5.9. Methods of Integration

(List of Basic Forms of Integrals)
No. Integral Method of integration
1
[ ^[(pMlfW dx Substitution (p (x) = t
2
J
/ W M dx <p' Integration by parts
J
/ W 9' W d*= / (x) (p (x)— 9 (x) /' (x) dx.
This method is applied, for example, to
integrals of the form p (x) f (x) dx> where

^
p (x) is a polynomial, and f (x) is one of the

e^'y cos ax; sina#; In x\
arc tan x\ arc sin x, etc.
and also to integrals of products of an expo-

nential function by cosine or sine.
3
^ / W <f
(n)
W d* Reduced
{n)
to integration of the product
(x)q*(x) by the formula for multiple in-
f
tegration by parts
/ (
x) (p
(w)
(x) dx = f (x) q><»
- 1) (x) —
J
— f'(x) q>
( "- 2 > (x) + /" (x) (p
( "- 3 (*)—...
>
... + (-_l)i.-i/(i.-D X )q,

( w+
+ (-l)» J
fi»)(x)q>(x)dx
4 pn (x) dx, Applying the formula for multiple integra-

J tion by parts (see above), we get
where p n (x) is a polyno-
mial of degree n.
^
e**p n (x) dx =
~ L a a2 1
a3 *
* '
~ ~
t
P"
+ ( »'< j? +c
C Mx + N
5
J x
2
+ px+q ^ X> Substitution
p
2 — 4q <
§ 5.9. Methods of Integration (List of Basic Integrals) 243
a.
r dx
o Reduction formula is used
x 2n —3
n
~(2n — 2)(x 2 + l)"- 1 1
2/1 —2
7 fPW
\ttH
, .
P(x)
where 7r^-J - Integrand is expressed in the form of a
sum of partial fractions
is a proper rational frac- P(x) A± A2 At
tion
Q(x) (x—x x )
1
(x— x t 2 1 ••• 1
X — Xl )l~T~
= (x — x
) (
Q (x) 1)
1
(x-
-x )" ... 2 (x 2 + px + 1
(x-x 2 )
1
(*-* 2 )
2 1 1
(*-* 2 )'« 1
+ ...
*a+p* + <7
1
(x* + px + q)* 1
{x * + px + q) k
8 Reduced to the integral of a rational frac-

1")^,
a:
by the substitution x t k where k is a
tion = ,
where is a rational func- common denominator of the fractions

tion of its arguments.
m r
~~n
'
T
9 Reduced to the integral of a rational frac-

H X dx by the substitution
) [
'
[cx+d) \
>
tion
where R is a rational
function of its arguments. cx-\~d
10 f —j Mx+W dx By the substitution x-\--^ = t the integral

+ ^+c
I
2
J |/a^
is reduced to a sum of two integrals:
,
Mx+N *-mS m +
f t
The first integral is reduced to the integral

J N 2
-f/n
of a power function and the second one is a

tabular integral.
244 Qh. V. Basic Classes of Integrable Functions
No. Integral Method ot integration
R (x, \âx 2 -\-bx-\-c) dx y

Reduced to an integral of rational fraction
by the Euler substitutions:
where is a rat ion a
function of x and Vax* + bx + c = t ± x V~a {a > 0),
lâx'1 -|- bx +c Vax 2

+ bx + c=tx ± V^~c (c > 0),
V ax 2
-\~bx+c=t (x—x x )
(Aac —b 2
< 0).
where x ± is the root of the trinomial ax 2 -\-

+bx + c.
The indicated integral can also be evalua-
ted by the trigonometric substitutions:
V 2 - 4ac
sin /
2a
X
+2a=< V b2 — 4ac cos t (a < 0,
2a
\ac —b 2
< 0)
V b2 - 4ac
sec /
2a
X + Ta =
y 2 — Aac
2a
cosec / (a > 0,
Aac —b 2
< 0)
y 4ac--b 2
tan t
2a
V 4ac--b 2
2a
cot / (a > 0,
4ac- -6 2 > 0)
12
Pn W dx t
Write the equality
}/"ax 2 + +c toc
P„ (x) dx
where P„ (*) is a polyno-
j/~a# 2 + +c
w _
mial of degree n. 1 6#
dx
|/~ax 2 -j-kx;-j-c
where Q n -i(x) is a polynomial of degree

n —
1. Differentiating both parts of this equa-
lity and multiplying by Vax -\-bx-\-c we

2
t
get the identity
Pn (x) s i (*) (<u 2 + bx + + c)
+ |q b . 1 W(2(w + 6) + A i
which gives a system of n-\-\ linear equa-

tions for determining the coefficients of the
polynomial Q n - 1 (x) and factor k.
§ 5.9. Methods of Integration (List of Basic Integrals) 245
And the integral
dx
V ax' + bx + c
2
is taken by the method considered in No. 10

(M = 0; N=l).
dx
13 This integral is reduced to the above-con-
(x — Xi) m Vax 2 + bx + c sidered integral by the substitution
14 xm (a + bx n )Pdx, This integral is expressed through elemen-

^
tary functions only if one of the following
where m, n, p are rational conditions is fulfilled:
numbers (an integral of a (1) if p is an nteger,
binomial differential).
.m+ 1
(2) if is an integer,
n
m+\ -p
(3) if is an integer.
1st case
(a) if p is a positive integer, remove the

brackets (a +
bx n )P according to the Newton
binomial and calculate the integrals of powers;
(b) if p integer, then the
is a negative
substitution is x =
the common t
k, where k
denominator of the fractions and n leads m y
to the integral of a rational fraction;
2nd case
*
if is an integer, then the substitu-
n
tion a-\-bx n applied, where— t
k is k is the
denominator of the fraction p;
3rd case
if ~~~ JrP ls an integer, then the substi-
tution a-\-bx n x n t k is applied, where k— is

the denominator of the fraction p.
15
^ R (sin x t
cos x) dx Universal substitution tan -?r=t.
If R( — s'mx y cosx)= R(slnx t

cos x),
then the substitution cosa: = / is applied.
If R (sin x y — cosx)= R (sin x y
cos a:),
then the substitution smx=^t is applied.

If R( s\nx, —
cos x) R — = (sin x }
cos*),
then the substitution tanx / = is applied.
246 Ch. V. Basic Classes of Iniegrable Functions
No. Integral Method of Integration
16
^ R (sinh x, cosh x) dx The substitution tanh-^- = t is used. In
this case
sinhx = —-; cosh*^^; d* =1

__ /2
.
17
^ sin ax sin bxdx Transform the product of trigonometric
functions into a sum or difference, using one
bxdx of the following formulas:
^ sin ax cos
sin ax sin bx —
^ cos ax cos bx dx —
= -^- [cos (a b) a;— cos (a-\-b) x]
cos ax cos 6* =
— -i- [cos (a — fr) x-j-cos (a-\-b) x]
sin ax cos bx =
=-^- [sin (a b) x-j-sin (a+b) x)
m mis an odd positive number, then apply

18
^ sin x cos" x dx, If
the substitution cosa: /. =
where m and n are inte- If n is an odd positive number, apply the
gers. substitution sin# /. =
If m-j-n is an even negative number, apply
the substitution tan# /. =
If mand n are even non-negative numbers,
use the formulas
sin 2
n
x—
1 — cos 2x ; cos 2 x =— —
1+cos2x
Reduce the the binomial

^ sin^ x cos^
19 to integral of
differential by the substitution sin# = /
(0 < x < ji/2),
and # — rational num-

sinP xcos? xdx=^ tP(\— t*)**- 1
dt
p J
bers. (see No. 14).
20 Transform into an integral of a rational

function by the substitution e ax = t
Chapter
6
THE DEFINITE INTEGRAL
§ 6.1. Statement of the Problem.

The Lower and Upper Integral Sums
Let a function f (x) be defined in the closed interval [a, b]. The
following is called the integral sum:
i-
where a =x < x A <x < 2

.
< x n _ < xn = b,
. .
x
hXi = x i+1 — t \ li£[x h x i+l = 0, ]

(J 1, . . . ,
/2 — 1).
n- i
The sum S„ = 2= M,. Aa:,- is called the upper (integral) sum, and
/
A2 — 1
sn = ^ niikXi is called the lower (integral) sum, where M = (

i =
= supf(x) [m^mlfix)] iovx£ [x h x i+1 ].
The definite integral of the function f {x) on the interval [a, fr] is
the limit of the integral sums
\ / (*) = lim 2 f (£/) Aa;,- when max |

Aa:,- |
—> 0.
If this limit exists, the function is called integrable on the inter-

val [a, b]. Any continuous function is integrable.
6.1.1. For the integral

JX
J
sin xdx
find the upper and lower integral sums corresponding to the division
of the closed interval [0, n] into 3 and 6 equal subintervals.
248 Ch. VI. The Definite Integral
Solution. Divide the closed interval [0, n] into 3 equal parts by

the points:
%o
_
—n = n
-^2
_
— ~3~ 2ji
xs _
— *^ •
^ » »
The function sin* increases monotonically on the interval 0,

yj t
and therefore for this interval we have ra o = sin0 = 0, M =

= sin^- = -?y-. The least value of the function on the interval
[iT' T"j
* s m =s
i
i n »
an d the greatest value is M =
x
= sin-^-=l. On the interval Jtj the function sin* decreases
monotonically and therefore
m = sin n = 0,
n M
AA 2ji
= sin -y V3
= -^y- .
2 2
Since all Ax k are equal to -y
s3 =Z Ax, = iL (o + + o) = ^p- ~0.907,
k=0
When subdividing the closed interval [0, k] into 6 equal intervals

ill JX
= ^-> Jl
= -y Jl 2jl
=
i
=
.
rv
by the x x2 x3
points y *t=-g-, >
^^T* Xb
== —
5ji
f
x6 = jt, we find by analogy:
m = 0, 7W = sin-^- = 2
l
m = sm TT ~~
x
. ji 1
2
'
M 1
== sin = 2
m = sin T "~
2
Jl
2
»
M = sin -y =
2 1,
m = sm T3
2
9
7W 3 = sin-y = 1,
5ji
= sin -y = 2ji /3
m = sin-^-
4
6
"~
. 1
2
'
7W 4
.
m = sin 5
ji == o, M = sin — =
5
1
2
'
§ 6.1. The Lower and Upper Integral Sums 249
For this division we obtain
s,=}K + «i+... +m )=^-(l+K3)~ 1.43, 5
S«= £ (Af + M + + M = £ (3 + /3 « 2.48.

1
. . .
5) )
As would be expected, the inequalities

ji
s3 ^ ^ s6 ^ sin x dx ^S ^S 6 3
hold true (the exact value of the integral is equal to 2).
6.1.2. At what 8>0 does the relation
n-i
J
sin xdx — 2 sm %k & x < 0.001 k
follow from the inequality max hXi < 8.
Solution. Since s n /„ < <

S„, then for the required inequality to
hold true it is sufficient that the upper and the lower integral sums
differ by less than 0.001:
0<S„-s„< 0.001.
But
n-i n-i
S n -s n = 2 (Mi—m,) A*, < 5 2 (M,— /n f ),
i =o i = o
where M
and m are the greatest and the least values of the func-
f {
tion sinx on the interval [x h x i+1 ] (i = 0, 1, — 1). Assuming

for simplicity that the point ^- is chosen as one of the points of
division and taking advantage of monotonicity of the function s'mx
on the intervals 0, -^-j and , nj , we obtain
X(M — m = 2 (sin-|— sin0) =

f f)
2.
Consequently, the required inequality is satisfied if 26 < 0.001,

i.e. 5 < 0.0005.
6.1.3. Show that the Dirichlet function [see Problem 1.14.4 (b)]
is not integrable in the interval [0, 1].
In dividing the closed interval [0, 1] into a fixed num-

Solution.
ber of parts we must take into consideration, in particular, two
possible cases: (1) all points are rational; (2) all points are
irrational. In the first case the integral sum is equal to unity, in

the second to zero. Hence, no matter how we reduce the maximum
length of subintervals, we always get integral sums equal to unity
and integral sums equal to zero. Therefore, the limit of integral
sums is non-existent, which means that the Dirichlet function is
not integrable on the interval [0, 1].
6.1.4. Find the distance covered by a body in a free fall within

the time interval from t a sec to t b sec. = =
Solution. A body moves
in a free fall with constant acceleration g
and initial velocity = 0.
Consequently, the velocity at the instant t
v
is equal to the velocity increment within the time interval from
to /, i.e. v(t) =
kv. For a short time period A^ the velocity incre-
ment is approximately equal to the acceleration at the instant t
multiplied by A/. But in our case acceleration is constant, there-
fore Au = g"A/, and hence, v(t) gt since At t /. = y
= — =
Let us subdivide the time interval from t=a to t = b into n
equal parts; then the duration A/ of each subinterval will be equal
to A/^^1n^. We assume that during
to
each subinterval of time the
body moves uniformly with a velocity equal to its velocity at the
beginning of this interval, i.e.
v =ga,
( . , b —a
v n -i = g[<* + (n— J
)
Whence we find the distance covered by the body during the ith
Vi
subinterval: . The entire distance covered by the body is
approximately equal to
^^s n = ^(v + v + 1
. . . +v n _ = 1)
b—a
-7T8 na+ «
1
n.
\-2
n
\-
1
. . .
1 v
1)
'
n
— anin —
+—
b 1)1
= (b—a)g
z i \ T
[a
,
2
2~~"J
With n increasing the distance covered can be evaluated more accu-

rately. The exact value of s is found as the limit sn as n—> oo:
s= Urn s n = lim^g(b—a) [a+y (b — a) (l — =

-^-)J
= g(b-a) fl+ l(ft_fl)] = !( & 2__ a2) .
Since s„ is an integral sum

n- l
t =o
the distance s is an integral:
b b
s =^ vdt = ^gtdt = j- (b — a 2 2
)
.
a a
6.1.5. Proceeding from the definition, compute the integral

i
Solution. By definition,
\ xdx = \\m 2 S/Ax; as max Ax, — 0,

t =
where
= * <*i<--. <*„=!.
A#; = a^^+j x^-.
UK
1. Subdivide the closed interval [0, 1] into n equal parts by the
points x. = -L(i = 0, 1, 2, . ...n).
The length of each subinterval is equal to Ax t

=^ , and -i- —*
as n —> oo.
Let us take the right-hand end-points of the subintervals as the
points lt \ g. = *. +1 ==^(/ = 0, 1, — 1).
Form an integral sum:
',=s,="f '±1.1=1 (1+2+ ... +(l) = iii±i>.

i =o
As az— oo the limit of this sum is equal to
1
lhn
«... 2n
~2-
Hence,
i
x dx = y .
J
o
2. Using this example, we will show that for any other choice
of points l f the limit of the integral sum will be the same.
Take, for instance, the mid-points of the subintervals as £t
- =
f
+T '
= -}T (t ==0 *
1 n ~V-
Form an integral sum

^'2t'+l 2n 2
,
/
»
o
= S-sr •T=2^L 1+3+5
c
+ 1 1 |~
i i i i
---
i/o ixl
+(2n_1) ==
4^
= ^-l
J
i =
Hence
lim / n = i-*
n -+ cc
6.1.6. Proceeding from the definition, compute the integral:
\x m dx (m=^— 1, 0<a<b).
a
Solution. In this example the following points can be conveniently

chosen as points of division:
_L JL A
x = a\ Xi=a(±-\
a
,
xâiÂ
a J
,...,*„ =a a
. J \ \ .
They form a geometric progression with the common ratio
The length of the ith subinterval is equal to
Sx— aq i+1 — aqtâq 1

(q—l).
Therefore the maximum length of the subintervals equals max Ax t

=
n-l J_
= aq n " 1
(q— l) =a — 1 and tends to zero with
increasing n, since lim^ = l.
Now let us choose the right-hand end-points of the subintervals

as I,-: £,. = x f+1 = a<7''
+l (i=0, 1, 2, .... n— 1).
§6.1. The Lower and Upper Integral Sums 253
n- 1 n- 1
In = 2 8^*,= 2flY + ^(?-l) = 1!/B
i=0 = t
= i«+l( 9 _l) 9 *[l+ 9 *+l+. ..

fl
0» + = l>]
n (m + l)n 1 ,o 1
= a^{q-\)q- qm+l _ ^(b^-a^)q-

q
{
1
Let us calculate the limit of the integral sum as max Ax {

— 0,
t. e. as q—+ 1:
Hm /„H^ +1 -a" +1 )H^

Thus,
d* =
7h~T\
(
bm+1 — am+1 )-
6.1.7. Proceeding from the definition, compute the integral:
it-
i
Solution. Subdivide the interval [1, 2] into n parts so that the

points of division x t (i'=0, 1, 2, n) form the geometric pro-
gression:
* =1; x 1 =q; x 2 =q 2
\
x3 =q 3
\ . ..; xn = q n = 2,
whence q= j/2.
The length of the ith subinterval is equal to
Ax, = <7'
+1
— f'^'fa-l),
and so t q (q— 1) >0 as
max Ax = n~
- oo, i. e. as g x
1. — n— —
Now us choose the right-hand end-points of the sub intervals
let
as the points \ h i.e., li x i+ ^=q i+1 = .
i=0 fe
' i=0 9
2
T
"( 2 "-0 =ln2>
lim /„= lim
n -* cc n -* <x>
—
2"
—
1
since 2 — ~ — In 2 1
J
as n—oo.
And SO,
dx
= ln2.
*
6.1.8. Evaluate the integral

5
/=$ J/25 —x 2
dx 9
proceeding from its geometric meaning.
The curve y=V2h—x 2 is the upper half of the circle

Solution.
x* + y = 25. The portion of the curve corresponding to the variation
2
of x from to 5 lies in the first quadrant. Hence, we conclude

that the curvilinear trapezoid bounded by the lines x = 0\ x = 5;
y = y
and y = V25—x 2
is a quarter of the circle x 2 -\-y 2 = 25\ and
.,
its area
.
is equal to
i j.
—25ji
Hence,
I = ^V25-x 2
dx = ^.
6.1.9. Evaluate the integral, proceeding from its geometric

meaning:
/ =$ (4x— l)dx.
6.1.10. Prove that

X
^-arcsin-J (0<x<a).
Solution. The integral

3 l
4
2
J1 a: dx
H u
x
y \
1
la.
,27
expresses the area S 0AMx of the por-
tion of a circle of radius a lying in
the first quadrant (see Fig. 59).
V This area equals the sum of the
areas of the triangle and theOMx
sector OAM.
>OMx
Fig. 59
The area of the sector
where sin/ = -
Hence,
a2 x
ÔAM — ~Y arC Sin "J
c .
'
and consequently,
I = ~ya
2
2 — # + 4r arc sin—
2
2 a
.
6.1.11. Proceeding from the geometric meaning of the integral,

show that
2ji 1 1
(a)
J
s\n 3 xdx = 0; (b) [ e~*
2
dx^2 \e~*
2
dx.
Solution, (a) The graph of the function y s\n 3 x is shown in =

Fig. 60. Let us show that the area situated above the x-axis is
equal to that lying below this axis. Indeed, let n^x^.2n, then
x=n + x 1 where O^x^rc
and sin 3 x sin 3 (n xj sin 3 ^. = + =—
Therefore, the second half of the
graph is obtained from the first one
by shifting it to the right by n
and using the symmetry about the
x-axis. Hence,
2jx
^ s'm
3
xdx= 0.
o Fig. 60
6.1.12. Given the function f(x) =x 3

on the interval [—2, 3],
find the lower (s n ) and the upper (S n ) integral sums for the given
interval by subdividing it into n equal parts.
6.1.13. Proceeding from the geometric meaning of the definite

integral, prove that:
2jx
(a) $ s\n2xdx = 0\ (b) \cos 3 xdx = 0;

o o
2 3
§(2x+\)dx = 6; jK9=]?dx =
9jt
(c) (d)
2
6.1.14. Passing to the limit from the integral sums, compute the
integral
by subdividing the interval [1, 4]:

(a) into equal parts;
(b) by points forming a geometric progression. In both cases
choose as:
(1) left-hand end-points of the subintervals;
(2) right-hand end-points of the subintervals;
(3) mid-points of the subintervals [x h x i+1 ].
§ 6.2. Evaluating Definite Integrals by the

Newton-Leibniz Formula
The following is known as the Newton-Leibniz formula:
b
\f{x)dx=F(x) \
b
= F(b)-F(a),
where F (x) is one of the antideri vatives of the function f(x), i.e.
F'(x) = f(x) (a
Solution. Since the function F (x) = arc tan x is one of the anti-
derivatives of the function f (x)— _^V2 ,

using the Newton-Leibniz
{
formula we get
/— f
dx —
~~
J l+* 2
6.2.2. Compute the integrals:
ji
2 2
6
§ 6.2. Newton- Leibniz Formula 257
I
x2 for <x< 1
f (
x) =\ for 1 <x<2.
2
Evaluate J f
(x)dx.
6
2
Solution.
\f(x)dx=^f
12
By
(x)
the additivity property of the integral
dx+ ^f(x)dx=^ x dx+ \ V~xdx =

122
1 o I
I 3 2
_
~
X*_ I
^
, 2^
X
~2
3 I
3
1

2
/ = Sl l—x\dx.
Solution. Since
,
| 1 —x for 0<*< 1,
1
^ 1
1 x— 1 for 1 <*<2,
we obtain, taking advantage of the additivity property of the integral,
2 1 2
1— x\dx= 5 (1— J
(x— l)d;c =
J |
-*) 2 (*-
(1
2
,
l)
2
=1+1=1
2^2 l#
a
where a < b.
Solution. If 0â<6, then / (*)
= 1~- == 1 , therefore
a
b
= b — a. If a<b<0, then /(*) =— 1 and $ f(x)dx = — b — (— a) =
= a— 6. Finally, if a < < ft, then divide the integral ^ /

into two integrals:
b b
\f{x)dx = \f{x)dx+\f (x) dx=b— {—a).

a a
The above three cases may be represented by a single formula:

b
§l±\-dx = \b\-\a\.
a
Note. When evaluating integrals with the aid of the Newton-Leib-

niz formula attention should be paid to the conditions of its legi-
timate use. This formula may be applied to compute the definite
integral of a function continuous on the interval [a, b] only when
the equality F' (x) f (x) is =
fulfilled in the whole interval [a, b]
[F (x) is an antiderivative of the function f(x)]. In particular, the
antiderivative must be a function continuous on the whole interval
[a, b], A discontinuous function used as an antiderivative will lead
to the wrong result.
6.2.6. Find a mistake in the following evaluation:
V3
dx
T-i t
— — arc
= 2
1
rr tan
-77tanct
, 2x
\—x
Vs
o 2
[arctan( V 3) — arc tan 0] = ^ ,
where (± arc tan y^r) = yq^(*=?M).
Solution, The result is a priori wrong: the integral of a function

positive everywhere turns out to be negative. The mistake is due
l 2x
to the fact that the function -y arc tan
{
_ x2 has a discontinuity
of the first kind at the point x=l:

n
lim
1
-rr arc tan

, 2x
t = -t\ urn — arc tan
1 , 2x
- = TC
r .
The correct value of the integral under consideration is equal to
V3
dx ,
arc tan x
V3
= arc tan ]/"3 — arc tan = -?-
o
o
Here the Newton-Leibniz formula is applicable, since the function
F (x) = arc tan x is continuous on the interval 0, and the equality

F'(x) = f(x) is fulfilled on the whole interval.
§ 6.2. Newton- Leibniz Formula 259
6.2.7. Find a mistake in the following evaluation of the integral:
jt
dx
J 1 + 2 sin H 2 2
J cos' x-j- 3 sin x
71 dx
+3 tan 2
1
arc tan (V 3tanx) = 0.
-f I * j/"3~
(The integral of a function positive everywhere turns out to be

zero!)
Solution. The Newton-Leibniz formula is not applicable here,
since the antiderivative Z (x) 7

=-pLr arc tan (K 3 tan#) has a discon-
tinuity at the point x = -S-. Indeed,
lim F(x)= lim — arc tan ([/" 3 tan x) =
—F=-arc tan(+oo) = —— ,
lim F (x) = lim — arc tan (K 3 tan

X
x) =
n
1
arc tan ( — oo)=-

V 3 '
2V~3
The correct result can be obtained in the following way:
C dx __ C 1 dx
2
J cos # + 3 sin 2 # J cot 2 # +3 sin 2 *
= 7rr=- arc tan (V 3 cot x)

1/3 1^3
It can also be found with the aid of the function F (x) =

arc tan (K3 tan a:). For this purpose divide the interval of
/3
integration [0, ji] into two subintervals, 0, yj and j~y ,
jij , and
take into consideration the above-indicated limit values of the func-
tion F (x) as x —y=F0. Then the antiderivative becomes a conti-
nuous function on each of the subintervals, and the Newton-Leibniz

2-0 Ch. VI. The Definite Integral
formula becomes applicable:
o _JT_
arc tan ([/ 3 tan a:) arc tan (|/~3 tan a;)
^"3
6.2.8. Compute the integral
dx.
, , . »/ I + cos 2a: , / 2 cos 2 x

=
Solution, cos a:
y
:
I I
cos x, O^Zx^-j ,
— cos X, y < x < Jl.
Therefore
cos 2x
1 -j-
dx = j
cosxdx + ( cosx)c(a: =
J
= sin x 7 + (—sin*) ^ =(l_0) + (0— (— 1)) = 2.
Note. If we ignore the fact that cos* is negative in y , ji

j
and put
f 1 4-co s 2x
— cos a;,
V 2
we get the wrong result:
^ cos x dx = sin x^ = 0.
§ 6.2. Newton-Leibniz Formula 261

IOOji
/ = J
Kl— cos 2xdx.
Solution. We have
Vl— cos2a:= [/~2 |

sin jc |.
Since |
sin a: |
has a period ji, then
IOOji IOOji
jj \
r
\—cos2xdx=\r 2 J |
sin a: |
rfA: =
o o
ji
= 100 J/T J
sinxdA;=200|/2 .
6.2.10. Evaluate the integrals:
= }' 7ITW (b)/ =

<»>'
-2
;
Ix—
-3
2 1
'
2
/= sin a ydx; (d) / =
J ^qry^;
(c)
j
-ji o
_2_
JT
-
I
sin
£ f
i_
n
1
(g) Htt^' (h)

i+* 8 »
3
xdx
JI
T
(j) 7=5 Kcos* — cos 3 Jed*;
jt
2
~
"3
§ 6.3. Estimating an Integral.

The Definite Integral as a Function of Its Limits
1. If f(x)^q>(x) for a^x^b, then

b b
(x)dx<^ q>(x)dx.
J f J
a a
In particular,
b b
\f(x)dx < \\f{x)\dx.

a a
b
2. m (b—a) ^lf{x)dx<^M (b — a) 9
where m
is the least value, and the greatest value of the func- M
tion f(x) on the interval [a, b] (estimation of an integral).
3. If the function f (x) is continuous on the interval [a b], then y
\f(x)dx=f(&)(b—a), a<l<b
a
(mean-value theorem).
4. If the functions f(x) and y(x) are continuous on [a, b] t
and
<p(x), in addition, retains its sign on this interval, then
b b
\f{x)y{x)dx = f{l)\y(x)dx, a<l<b

a a
(generalized mean-value theorem).

x a
5. / (/) dt = / (*); -gj-j /(*)# = — /(*) at each point x of

J
continuity of the function f(x).
6.3.1. Estimate the following integrals:

JT
~3~
(a) / = j K3 + dx; (b) / =

j
2
+2
Solution, (a) Since the function / (x) = )/3 + x increases mono-
2
tonically on the interval [1, 3], then m = 2, M = /30, b—a = 2.

§ 6.3. Estimating an Integral. Integral as a Fund, of Its Limits 263
Hence, the estimation of the integral has the form

3
2-2 < Jl/T+l?dA:<|/"30-2,

l. e.
4 < ^3 + *
$
3
dx:<2|/"30^ 10.95.
(b) The integrand f(x) = —j— decreases on the interval

r ji
'
jt
3"
L"4
since its derivative
^ x cos x — sin x (x — tan x) cos x ^

q
j,
Hence, the least value of the function:
m-
3 V3
2ji
its greatest value being

2 V'2
M=
Therefore
3 /*3 / ji Jt
2ji
i. e.
6.3.2. Estimate the absolute value of the integral

19
C sin;
J
~> dx.
10
Solution. Since |sin*|<l, for x>10 the inequality
[-\-x 9
< 10" 8 is fulfilled.
Therefore
19
C sin
TT X 8
<(19— 10)10" < 8 10"?
J
10
(the true value of the integral ^ — 10 8

).
6.3.3. Which of the two integrals

1 i
x 3 dx
is the greater?
Solution. As is known, Y* > x 3 for 0<x< 1. Therefore

i i
^ Y xdx > ^ x 3 dx.

i i
2
(a) 0< \ 3
/l^_ < j\ (b) 1 < [e* dx<e.
b ^ 1+ * 8
b
Solution, (a) Since < ~ < for 0<x<^l,

V +% l
8
then
i i
o < (-
s
*M=< [*dx=± 1
(b) Since for <x< 1 there exists the inequality 1 < ex < e,
*
then
i ) i
2
^dx < ê* dx< ^ edx.
Hence the inequality under consideration holds true.
6.3.5. Prove the inequality

ji
T
ê-R*tox dx< JL(i_e-«) (R>0).
o
Solution. Since the function f(x) = —n—x si

decreases on
^
^0,
jx \
[see
J
Problem 6.3.1 (b)], then for 0<x<j

/(*)=—x >^t] = «-
c .
x sin .
£ ( ji \ 2
Hence, on this interval sin#>—2 x, therefore
sin *
<^ g n
and
2 2
~2R
ji
r -—
2^
^
e~ R sin
*dx < ^
e n *
dx -
6.3.6. Prove that for any functions f (x) and g(x), integrable on
the interval (a, b), the Schwarz-Bunyakovsky inequality takes place:
^f(x)g(x)dx <:

F(x) = [f (x)-kg(x)}>,
where k is any real number. Since F (x) 0, then
or
& /; b
X2
J
g
2
(x) dx—Tk
J
/ (x) g (x)dx + J
/* (x) dx > 0.
a a a
The expression in the left side of the latter inequality is a quad-

ratic trinomial with respect to X. It follows from the inequality that
at any X this trinomial is non-negative. Hence, its discriminant is
non-positive, i. e.
J
/ (x) g (x) dx -Jf 2
(x) dx
J
g
2
(x) d* < 0.
j
Hence
r?
/~ b 1
y(x)g(x)dx 2
<j/ ^f*(x)dx ^g (x)dx,
6.3.7. Estimate the integral from above

i
, (' sin x i
Solution. By the generalized mean-value theorem we have
sin x
I
dx = sin l\
b
—~—t2 = sin \ arc tan x = -J-sin6(0<6< 1).
J l+x' J l+*
Since the function sin a: increases on the interval [0, 1] then

sin £< sin 1. Whence we get an upper estimate of the integral:
C sin x
< —r sin « 0.64.
A aA
, . «
-7-
\ T dx 1
It is possible to get a better estimation if we apply the same

theorem in the form
P sin x
dx sin x dx = y^|2 (1
— cos < 1 cos 0.46.
i
+ Sj J
!
') 1
Proceeding from geometric reasoning, prove that:

6.3.8.
(a) if f (x) increases and has a concave graph in the
the function
interval [a, b] then y
(b—a)f(a) <^f(x)dx<(b—a)-
f(a) + Hb) .
(b) if the function f (x) increases and has a convex graph in the
interval [a, b] 9
then
b
f(a)+f(b)
(b-a) < §f(x)dx<(b—a)f(b).
Solution, (a) Without limitation of generality we may assume

f(x)>0. Concavity of the graph of a function means, in particu-
lar, that the curve lies below the chord
through the points A(a,f(a)) and
B(b, f(b)) (see Fig. 61). Therefore the
area of trapezoid aABb is greater than
that of the curvilinear trapezoid boun-
ded above by the graph of the func-
tion, i. e.
b
^f ix)dx< SaABb = (b-a).L^+m,
(b-a)f(a)< \f(x)dx
a
is obvious.
i
6.3.9. Estimate the integral ^\fl-\-x*dx using

(a) the mean-value theorem for a definite integral,

(b) the result of the preceding problem,
(e) the inequality ]f 1 + x* < 1 +

(d) the Schwarz-Bunyakovsky inequality (see Problem 6.3.6).
Solution, (a) By the mean-value theorem

i
/ = J \fl+x i
dx = V'T+V, where 0<£< 1.
But
i < VT+T* < V%
whence
1< / < K2 » 1.414.
(b) The function f(x) = VY+ is concave on the interval [0, 1],
since
rw=f^±|>o, o<,o.
On the basis of the preceding problem we get
i ,
1 < ^\f\T7*dx< -±f-l& 1.207.
1 1
(c) l<I = frl+T dx< j(l 1 !+-!.= 1.1.
r
(d) \ +x*,g(x)
Put f(x) = \/ = 1 and take advantage of the Schwarz-
Bunyakovsky inequality
i i
r~\ i
\V\+#dx =^Vl+x*dx = I<y 5(i+^)dx.J l*dx =

= l/T2T« 1.095.
6.3.10. Find the derivative with respect to x of the following

functions:
(a) F(x)=\\ntdt (x > 0),

X2
Vx
(b) F(x)= ) cos (t
2
)dt (x>0).
Solution, (a) Write the given integral in the following way:

C XZ *3 X'
1
F (x) = \ In MM- \ In tdt = \ \nt dt— $ In tdt,

K2 C C C
where >
is an arbitrary constant.
c
Now us find the derivative F' (x) using the rule for differen-
let
tiating a composite function and the theorem on the derivative of

an integral with respect to the upper limit:
/ ~ X2
F'x (x) = $ \ntdt 5 \ntdt (x% = In x 3 3x> — In a;

2
2x =
_C Xs
= (9x* — 4x)lnx.
C Vic
<b) F(x) = \ cos (t*)dt-\- $cos(* 3 )d* =
-J
cos (t*)dt + J
cos (t*)dt;
F'(x) =- cos (t
2
) di
i
( x )*
+ cos (P)dt
J J
= — COS-o1/ 1
5 + COS*
, 1
Trrr == -.cos
1
-1,1^cosa;.
2
6.3.11. Find the derivative with respect to x of the following

functions:
23
(a) f(x) = ji!lii«; (b) F (x)=^T+T*dt.

X
6.3.12. Find the points of extremum of the function F (x) = ^^-dt

^
in the domain x > 0.
Solution. Find the derivative
sin x
x
The critical points are:
x=nn 2, ...), where sinjc = 0.

Find the second derivative at these points:

— sin X
cv,
t M=
,
x x cos x
;
F" (nn)
v 7
=—
nn
cos (nn) = — —
nn v '
(
v 1
)
7
n
¥= 0.
Since second
derivative is non-zero at the points x
the nn =
(n = these points are points of extremum of the function,
1, 2, . .
.),
namely: maxima if n is odd, and minima if n is even.
6.3.13. Find the derivative of y, with respect to x y

of the func-
tion represented parametrically:
x= ^ i/ z\n zdz\ y = J
z
2
\nzdz.
yt
Solution. As is known, =
xt
Find x and t y\\
x't = [^Y~z\n zdz) (t

3
y
t
= t\nt».3t = 9t 2 3
\nt' y
. (
yt
\VT 'YT 2Yt
whence
= - 36/
31n
y'
x = f l
2
J/7 (f > 0).
-\Vt \nt
(arc tan x) 2 dx
^ sin Iâ: ^
(a) lim 2 -5 (b) lim

^ ;
'
x-+ + a> Vx2 + 1
2
N
"°
(c) lim
+ oo
o
x2
Solution, (a) At x^-=0 the integral sin equals zero; it is
^
o
easy to check the fulfilment of the remaining conditions that ensure
the legitimacy of using the L'Hospital rule. Therefore
2
^ sin V".x dx sin V^xdx (* h
J
lim = lim 3x 2
= lim 2x 3xsin x2
2
3
'
x-+0
(c) We have an indeterminate form of the type—. Use the

L'Hospital rule:
* \ 2
/
2 2 2
ê* dx 2 ^ <?* dx-?*
lim
o
= lim
X—*- + CD *-* + 00
2
2 e*
J
: lim
o
2
= lim 2-^— = 0.
X-> + CO e*
6.3.15. Find the derivative ^ of the following implicit functions:
(a) ê~ t2 dt+ ^

sin 2 tdt=0\
o o
y *
(b) ê 1
dt+ sin/d* = 0;
J
o o
(c)
^
|/"3 — 2sin 2
ede+ ^cos/d/=0.
Solution, (a) Differentiate the left side of the equation with

respect to x, putting y=y(x):
y -x 2
r 1 dy ,
sin"
J
.0 _o
e
-' 2
^ + sin a
* 2 2x- = 0.
Hence, solving the equation with respect to j- , we get
2
^^_2A;/? + ^ sin 2 A:
2
.
dx
(c) Differentiate the left side of the equation with respect to x,

putting y = y(x):
' y
dy
2 sin 2 z dz + ^ cos / dt
dx
= 0.
-0
X
Whence
/3 — 2 s ^ V3 — 2 sin 2
x
i
2
x + cos y ^ = 0; dx cos y
6.3.16. Find: (a) the points of extremum and the points of

inflection on the graph of the function
X
/ = l)(f — a
2) df;
o
(b) curvature of the line defined by the parametric equations:
x = aVn cos
( %- dt,
J
o
. — r»
t
y = ay sin
sir -77- dt
(the Cornu spiral).
Solution, (a) The function is defined and continuously differen-

tiate throughout the entire number scale. Its derivative
/; = (*-l)(*-2) 2
equals zero at the points 1 l, 2

x = x = 2 and
when passing through y
the point x 1 it changes sign from minus to plus, whereas in the

neighbourhood of the point x 2 the sign remains unchanged. Conse-
quently, there is a minimum at the point 1, and there is no ^=
extremum at the point x 2 2. =
rx = 3x 2 — lOx + 8
4
vanishes at the points x 1 -~y* x 2 =2 and changes sign when pas-
sing through these points. Hence, these points are the abscissas of
the points of inflection,
(b) We have
x't =ay
—n cos
. 1
Ttf^
= a yr n sin
— Tit
2
, y\ ,
hence,
*7 */ ^ cos 3 -g-
whence the curvature

Vnt
K_ \y"\
3
__
a
'
U + (y')
2
l
6.3.17. Prove that the function L(*), defined in the interval

(0, oo) by the integral
is an inverse of the function ex .
Solution. Let us take the derivative
=7 (*>0).
Since the derivative is positive, the function y^--L(x) increases

and, hence, has an inverse function
x=L~ 1
(y).
The derivative of this inverse function is equal to

dx 1
__
Ty~L r{x)~ X '
whence it follows (see Problem 3.1.10) that
x = Cey.
To find C, substitute x = l. Since

L(1) = 0, i.e. y\ Xssl = 9
then
1 =C^°-C,
which proves our assertion:
x=L' l
(y)=ey.
6.3.18. Given the graph of the function y=f(x) (Fig. 62), find
X
the shape of the graph of the antiderivative I='^f(t)dt.

o
Solution. Onthe interval [0, a], the given function is posi-

tive; consequently, the antiderivative increases. On the interval
0, the derivative of the given function is positive; hence, the

yj
curve / = / (x) is concave. On the interval y , a the derivative
of the given function is negative; consequently, the curve / = / (x)
is convex, the point *^y being a point of inflection. The inter-
val [a, 2a] is considered in a similar way. The point x 1 is a =

point of minimum, since the derivative I'(x)=f(x) changes its
sign from minus to plus; the point x 2 a is a point of maximum, =
since the sign of the derivative changes from plus to minus.
A 3a 2a.
2 2
Fig. 62 Fig. 63
The antiderivative I (x) is a periodic function with period 2a,

since the areas lying above and below the x-axis are mutually
cancelled over intervals of length 2a. Taking all this into account,
we can sketch the graph of the antiderivative (see Fig. 63).
6.3.19. Find the polynomial P (x) of the least degree that has
a maximum equal to 6 at x=l, and a minimum equal to 2 at
x = 3.
Solution. The polynomial is an everywhere-differentiable function.
Therefore, the points of extremum can only be roots of the deriva-
tive. Furthermore, the derivative of a polynomial is a polynomial.
The polynomial of the least degree with roots = 1 and x 2 = 3
has the form a(x l)(x 3). Hence, — —
P\x) = a (x — 1) (x — 3) = a (x — Ax + 3).
2
Since at the point x= 1 there must be P(l) = 6, we have

X x
P(x)=l P' {x)dx-\-& = a^(x* — 4x + 3)dx + 6 =

i i
= a y — 2* + llx — ) 2 1 -!- 6
( 1
The coefficient a is determined from the condition P(3)=2, whence
a = 3. Hence,
P(x) = x — 6x + 9jc-|
3 2
2.
6.3.20. Find the polynomial P (x) of the least degree whose graph
has three points of inflection: (—1, 1), (1, 1) and a point with —
abscissa at which the curve is inclined to the axis of abscissas
at an angle of 60°.
Solution. Since the required function is a polynomial, the abscis-
sas of the points of inflection can only be among the roots of the
second derivative. The polynomial of the least degree with roots
— 1, 0, 1 has the form ax(x 2 — 1). Consequently,
P"(x)=a{x 3 — x).
Since at the point x = the derivative P' (0) = tan60° = j/3, we

have
X
P' (x)=§P"(x)dx + V3 = a^-^) + VZ.

Then, since P(l) = l, we get
P(x)=^P'(x)dx+\ = a(£-£ + l)+V3(x-l)+l.

1
The coefficient a is determined Jrom the last remaining condition
=— = ^(T/a-i)
/>(_!) 1, whence a . H ence,
"
]/ 3 ~"
P(x)- 7
1
(3a:
5 — 1 Ojc
3
) + x J/3.
6.3.21. Taking advantage of the mean-value theorem for the
definite integral, prove that
i
(a) 3< \ Vq + x*dx< 10,

o
JT
2 ,
(b) ]/ 1 + ysin 2 *d*<^ y |,

2Jt
?n P dx 2jt
/
(C)
\
13 ^ J 10 — — 3 cos x
|
7
*
6.3.22. Using the Schwarz-Bunyakovsky inequality, prove that
^ V +x 1
3
dx < -Xp- . Make sure that the application of the mean-
o
value theorem yields a rougher estimate.
§ 6.4. Changing the Variable in a Definite Integral 275

X x2
(a) F(x) = ^lntdt(x>0); (b) F(x) = §y.

1
_2_
x
6.3.24. Find the derivative ^ of functions represented paramet-
rically:
In /
(a) x = ^dz,
j
= j
e z dz\
2 5
sin / VT
(b) x= C
\ arc sin zdz, # = \
—
C sin- z2
dz.
6.3.25. Find the points of extremum of the following functions:
(a) F(x) = )e 2
(l — t
2
)dt;
i
X2
(b) /7
W =f!!=g±id/.
+ J 2 e<
£ Changing the Variable In a Definite Integral
If a function ^=
the following conditions:
cp(/) satisfies
(1) q)(/) is a continuous single-valued function
defined in [a, p]
and has in this interval a continuous derivative q/(/);
(2) with / varying on [a, P] the values of the function x y(t) =
do not leave the limits of [a, b]\
(3) q)(a)=a and y($) = b,
then formula for changing the variable (or substitution) in the
the
definite integral is valid for any function f (x) which is continuous
on the interval [a f
b]:
lf(x)dx=\f[<p(t)] y'(t)dt.
a a
Instead the substitution x

of y(t) the inverse substitution =
t = yp(x) frequently used. In this case the limits of integration
is
a and p are determined directly from the equalities a ^(a) and =

p =
x|;(fr). In practice, the substitution is usually performed with
the aid of monotonic, continuously differentiate functions. The
change in the limits of integration is conveniently expressed in
the tabular form:
X t
a a
b p
r
J \ \—x*dx.
Solution. Make the substitution = 2sin/, je assuming that

— ^T-^^C-y.
3~ The function # = (/) = 2 sin
(p t on the interval
[""IT' "t]
sa ^ s fi es a " ê conditions of the theorem on changing
the variable in a definite integral, since it is continuously differen-
tiable, monotonic and
<p(-t) = -^. <p(t) = ^ 3 -
And so,
a: = 2 sin = 2 cos / dt\ V ^ — x* = 2 |

cos / 1
= 2 cos /,
Tl Jl 1
since cos/>0 on the interval 3"
*
~3~J
Thus,
J
s
M— a
dx = 4 $
JT
cos 2 tdt=2 J
Jl
(l + cos20d/=»
" 3 "a
i
=2 [/ +
lsin2<]y = ^ + )/3.
iL
3
4 .

p 1/^2
\ — 4
= 2 sec/; *l /
d* = 2^df; 2
2
cos /
ji
4
T
On the interval 0, the function 2 sec t is monotonic, there-
~J
fore the substitution is valid.
Hence,
f ~4
_/ f J^sec 2 / —4 ^ sin/
16 sec 4 / ^cos 2 /^
H 3
Vz
=4i sin si
2
/ cos tdt =j2 s ' n3 ^
o
~ 32
*
a VI
dx
(a) §x*Vtf=^dx; (b)
j ^=
1
f
"8 dx
(a)7
v
J 6— 5 *f .
sin x-\- sin ,2 *

; (b)
v 7
f
J 2 + cos x '
Solution, (a) Apply the substitution
sinje = /; X /
cos x dx = dt\
ji
~2~ 1
The inverse function x = arc sin t[Q < -y for < < t 1) satisfies
•all conditions of the theorem on changing the variable. Hence,
cos xdx —3 i
_4_
/
6— 5 sin x-\- sin 2 x J 6- 5/-j-/ 2
" :ln-
—2 3
'
o o
(b) Make the substitution / = tan-^X2

2dt X
jc = 2arctanf, dx = t
l + t*'
ji
~2 1
which is valid due to monotonicity of the function tan 4" on the

interval —
1 2dt _ 9 f
df
J 2 + cos.v J
2+ \+t*
2
2
= -7=- arc tan -= . /
''arc tan —U — arc tan 0^ = 71
'
j/3 j/3 J/3 ^3 3 1^3
dx
2
(o>0, b>0).
cos 2 x-\-b 2 sin 2 a:
tan# = /,
dx
dt,
Hence,
<4
dx _— dt C dt _
~
+ bH ~ b*J
\
f
2
cos 2 x-\-b 2 sin x J cP a2 .
2
• -r arc tan —
= b —a arc tan—a = ab a \o
If #,= 6=1, '

then
ab
arc tan—
a
= arc tan — 1
ji
4
, , which exactly coin-
cides with the result of the substitution a = 6=1 into the initial
integral
n jt
4
dx [* ji
dx -
"
I a 1 cos 2 x-j-fr 2 sin 2 # 4
dx
dx; (b) f-
J X l^l + ln *
'
(0 dx.
j
.1
C x sin x
/=\
^ cos i x
dx -
Solution. Reduce this integral to the sum of two integrals:
ji
2 3T
~2~
To the integral
x sin x
'
2
dx
J 1 -j-cos 2 x
apply the substitution
= n—
x /, X t
dx = — d/, ji n
T T
ji
Then
JI
2
— Q sin — (ji t) r (n — /) sin /
' 2 d/
J l+cos 2 (Jt-0
1 J 1-j-cos 2
1 /
C smt C t sin I
'
n df.
) 1-f-cos 2 /"*
J 1 + cos 2
/
Hence
. _ . . _ C x sin x . P sin / rf/ f* / sin /
/ =/i + / s
- J + cos x^
1
2
71
) 1 + cos 2
/ J 1 + cos 2
/
Since the first and the third integrals differ only in the notation
of the variable of integration, we have
n
) 1 + cos 2
*
To this integral apply the substitution

a = cos t y
t a
du = — sin t dt, u i
l
ji
T
2
du du
/ = 1+M a
Jl'
/Vote. The indefinite integral

d
r*
\ t-
1
x sin x
-j- COS X
— is not expressed in
elementary functions. But the given definite integral, as we have

shown, can be computed with the aid of an artificial method.
\n(\+x)
2
dx.
1-f-x

x = tan/, X /
dt
dx -=
cos 2 /
Jl
1
T
Hence,
Jl
4
In (1+ tan/) sec 2
sec 2 /
/
<tf = Jln(l +tan/)^.
Transform the sum 1 + tan/:
/2~sin (
/ -|
1 + tan = tan / -f- tan / =

cos /
Substituting into the integral, we obtain

jt n_ n_
~T 4 4
/ - j lln2d/ + Jlnsin(/ + -J-)df— Jlncos/<//-- =
(I
T 4
= 1/1112 T + jlnsin(/ + -5-)d/ — Jlncos/d/ =

°
JT Jt
T T
= -y ln2 + jlnsin^/ + dt— j In cos t dt - ln2 + /2
Now let us show that I^=I 2

. To this end apply the substitution
= — dz
jt
dt y
JT
T
T
to the integral /2 = In cos t dt.

J
Then
/2 = — ^ In cos ^— = ^ In sin
i
ji
2
= j" lnsin ^- + 2^ dz =I t.
Therefore
/=T ln2 -
Note that in this problem, as well as in the preceding one, the

indefinite integral ^~~jz^r ^ x ls no * expressed in elementary fun-
ctions.
6.4.9. Prove that for any given integral with finite limits a and b
one can always choose the linear substitution x = pt q (p, q con- +
stants) so as to transform this integral into a new one with limits
and 1.
Solution. We
notice that the substitution x pt q satisfies = +
explicitly the conditions of the theorem on changing the variable.
Since t must equal zero at x = a and / must equal unity at x =b
we have for p and q the following system of equations
= p-0 + q,
a
b = p.\+q,
whence p = b — a, q = a. Hence,
b l
lf(x)dx = (b — a)$f [(b — a) + a] dt. t

6.4.10. Compute the sum of two integrals

_2_
-5 3
$e<* +5 >
2
dx+3 .•(-*)' dx.
Solution. Let each of the given integrals into an

us transform
integral with limits and
preceding problem). 1 (see the
To this end apply the substitution x = t 4 to the first inte- — —
gral. Then dx = — dt and
-5 1 1
l i==
^
e ( x+ v 2
dx=-le<- t
+ ^ dt2
=- J
e«
- 1
>
2
dt
-4
Apply the substitution x = -^-\-~- to the second integral. Then
dx = ^ and
_2_
o
3 / 2 \ 2 1
/2 =3$ e*\
x
~) dx = \ev-^dt.
Hence
+ =- je (
'- 1)2
dt+ J
= 0.
9
Note that neither of the integrals ^ e u+5)2 and $£ ^* 3
^ rfx is
evaluated separately in elementary functions.

6.4.11. Prove that the integral
p sin 2kx .
\
i
:
sin y
dX
equals zero if k is an integer.

x t
= —
1
x tc /,
n
dx = — dt,
Jt
Then at k an integral number we get:

Ji n
P sin 2kx
^ _
~"
P sin 2fe (ji—
—
/)
^_ P sin 2&/
^
J sin a:
J sin(Jt /) J sin/
Since the definite integral does not depend on notation of the vari-
able of integration, we have
/ = — /, whence 7 = 0.
2
dx
VT-
Solution. Apply the substitution x s\nt (the given function is =

not monotonic), dx =
cost dt. The new limits of integration t 1 and
t 2 are found from the equa-
y~3
y = sin = sin
i . . .
,
tions /;
2
We may put =y and t2 =

=y , but other values may
also be chosen, for instance,
2jt
'
3
In both cases the variable Fig. 64
x= s\nt runs throughout the
entire interval'
1
2
^3 (see Fig. 64), the function sin/ being
»
2 J
2jt 5ji
monotonic both on and ~6~
|_<r> tJ 3
Let us show that the results of the two integrations will coincide.
Indeed,
V 3
2
r
dx — r
c° s * di
_ r _ n tan —
J a: 1^1 —x 2
J sin / cos / ,1 sin /
j
= lntan- -In tan —= In

1
/-
12 V^3
On the other hand, taking into consideration that cost is negative

on the interval j^-y; -yj , we obtain
V 3 2n 5jt
2 3 6
i
*
dx _ r cos t dt _ r dt
__
J xVl—x 2
J sin cos t) J sin/
5jt 2jx
6 3
, 5
tan ^"3
2-f
= In tan -In ln
^"3
tan^
Note. Do not take t

1
=^ 9
/a =y f
since, with / varying oo
the interval
|y , -yj , the values of the function x = sin t lie beyond
the limits of the interval

2 '
2
6.4.13. Prove that the function L(x) defined on the interval

X
(0, oo) by the integral L (x) = — possesses the following properties:

J
i
L(x l x2 ) = L(x + L(x l) 2 ),
L^j=L(Xl )-L(x 2 ).
Solution. By the additivity property
1 1 X,
Let us change the variable in the second integral
t z
t=x z l y
1
dt = x dz x y
*•
Then
L(x,x 2 ) = §j +
i
^=
i
L(x l ) + L(x 2 ).
Putting here x 1 x 2 =x 3]
x2 = x—
x
3
, we obtain
1
L(x 3 )^L(Xl ) + L^, i.e. L(^)=L(x )-L( Xl 3 ).

m
It is also easy to obtain the other corollary L\x n
)= —
n
L(x) for
any integral m and n.

Indeed, for positive m and n this follows from the relations
L{x n
J =^mL\x n \ L(x)=nL(x n ) ,
and for a negative exponent, from
L(1) = 0, L(x- l ) = L(±)=L(l)-L(x) = -L(x).

Now, taking advantage of the continuity of the integral as a fun-
ction of the upper limit, we get the general property L(x a ) aL(x). =
Note. As is known, L(x) = \nx. Here we have obtained the prin-
cipal properties of the logarithm proceeding only from its determi-
nation with the aid of the integral.
3
6.4.14. Transform the integral ^(x — 2fdx by the substitution
{x — 2f--^t.
Solution. A formal application of the substitution throughout the
interval [0, 3] would lead to the wrong result, since the inverse
function x = y(t) is double-valued: x = 2±Y~i> the function x
i.e.
has two branches: x = 2 — 1

[/" t\ x 2 -=2 + Y The former branch can-
t.
not attain values x > 2, the latter values x < 2. To obtain a cor-
rect result we have to break up the given integral in the following
way:
3 2 3
J
(x — 2) 2
dx = J (x — 2) 2
dx + J (x — 2) 2
dx,
2
and to put x = 2—Y t in the first integral, and x =2+V t in the

second. Then we get
2 4
3 1 1
/2 =j(^-2)Mx=j/ -^ =i-Ji/rd/=i.
2 r
2
Hence, / = —8 -(- — = 3,
1
which is a correct result. It can be easily

verified by directly computing the initial integral:
(x-2f
§{x-2)* dx -
= i+T= 3 -
Jt
Jt
4
, , _ C sin # + cos x i
dx
(e ) 7 =j 3 + sin2* >
(f) /=jV ]/ j^-d*. a>0;

b
2a 1
/ = ^2ax-x*dx; (h)/ =
(g)
j fr ^ J
6.4.16, Applying a suitable change of the variable, find the fol-

lowing definite integrals:
dx dx
.1 ^+1+^+1)3 ;
J; + / —x 2 2 '
V(a* + b*)/2
(c (d)
).fmf^) ;
.f yV —a )(&
2 2 2 —* 2
)
V'(3a* + fc
2
)/2
6.4.17. Consider the integral It is easy to conclude

j 4 _^ 2
.
that it is equal to -j . Indeed,
4 +
— =
d#
a;
2-
2
1
-7T
2
arc tan
, #
2
2
-2~~ 2 [t-(-t)
Jl
T
-2
On the other hand, making the substitution #--=y, we have
X *
d*
,
= — dt
—2 1
2
1
2
2
[ dx r # r 1
arc tan 2t
2
_ _JX
2
+l
'
This result is obviously wrong, since the integrand

4 +* > s
0, and,
consequently, the definite integral of this function cannot be equal
to a negative number — ^- . Find the mistake.
2n
6.4.18. Consider the integral

to
/= \
P
J 5
——
=
dx
2 cos x
. Making& the substi-
tut ion tan -^ = t we have

2ji ,
dx 2<ft
0.
j
5-2 cos* 5-2 1 — 2
=
(1+<1))
j
The result is obviously wrong, since the integrand is positive,
and, consequently, the integral of this function cannot be equal to
zero. Find the mistake.
_2
6.4.19. Make sure that a formal change of the variable t —x h
leads to the wrong result in the integral

J
-2
Y 2
dx. Find the
mistake and explain it.
6.4.20. Is it possible to make the substitution x=sec/ in the

i
integral / = J
J/V+l dx?
o
i
6.4.21. Given the integral —x 2

dx. Make the substitution
o
n
jc = sin*. Is it possible to take the numbers n and as the limits
for /?
6.4.22. Prove the equality
a a
J
f(x)dx=l[f(x) + f(-x)]dx
-a
for any continuous function f (x).

2ji
6.4.23. Transform the definite integral §f(x)cosxdx by the sub-

o
stitution sin^ =
§ 6.5. Simplification of Integrals Based on the

Properties of Symmetry of Integrands
1. If the function f (x) is even on [
— a, a], then
a a
J
f(x)dx = 2^f(x)dx.
-a
2. If the function f (x) is odd on [
— a, a], then
a
J
f(x)dx--=0.
-a
3. If the function f (x) is periodic with period 7\ then

b + nT
\
j
f{x)dx^ J / (*)
a a+ nT
where /i is an integer.
i

J
|*|d#.
—l
Solution. Since the integrand /(a:) = |a:| is an even function, we
have
i i i
i
^\x\dx = \x\dx = 2^ xdx = x

i
2
2 \
)

7
4
f a: sin # ,
-7
Solution. Since the integrand is odd, we conclude at once that

the integral equals zero.
§ 6.5. Simplification of Integrals 289
6.5.3. Evaluate the integrals
ji
(a) cosnxdx;
J /(*)
— jr
jt
(b) ^ / (x) sin nxdx,

-71
if: (1) f (x) is an even function; (2) f(x) is an odd function.

5
6.5.4. Calculate the integral dx.

J ^4 + 2* + 2
t
*
sin 2a:
J cos 4 x-\- sin 4 a:
Solution. The integrand is a periodic function with period jt, since
sin 2 (x-\- sin 2a

/(* + jt) = cos ji)
"
:/(*).
4
(#4- n) + sin 4
(x + ^) cos 4 x-f sin 4 a:
Therefore it is possible to subtract the number n from the upper

and lower limits:
5
T 37
sin 2a: C sin 2a:

_2 f
tan xdx
J cos 4 a: -4- sin 4 x J cos 4 a: + sin 4
x 2
J cos x (1 +tan 4 a:)
JT
Make the substitution

X /
t = tan a:,
dt 9
ji
'
cos 2 x
T 1
JT
T 1
r tan x dx C 2t dt
2
JT
arc tan
J cos 2 #0 4" tan 4 x) J \+t*''
t
—
J
cos xf (x 2 )dx =2 ^ cos xf (x 2 dx. )
Solution. It is sufficient to show that the integrand is even:

cos( — x)f [( x) 2 \ = cosxf (x 2
).
C 2s 7
~~ ,
3* — >^ —
+— 6
— 10a: 5 7a;
3
1 + +
2a: 2---r^-ri
x 1 .
J +2 a;
2
-K~2
Solution.
V~2
C 2x7 + 3a;
6 — 1 0x b — 7x* — 1 2a;
2
+x+ 1 ,
dX ~~
J a; +22
-V~2
r 2xi—\0xb —7x*+X j ,
C 3a; 2 - 4) +
(a;
4
1
A
aX_t"
J *2 +2 J ^+2
-K~2 -K~2
V~2
= +2 [3(^-2x )+ 1?iT ]^ = 2
J _ 4^3 + 2
arc t an
|/"
x I v-2
= _ 16 "
K~2 2 |o 5 2 l/"2
Tn calculating we expanded the given integral into the sum of

two integrals so as to obtain an odd integrand in the first integral
and an even integrand in the second.
cos* In
1 —x dx.
Solution. The function f(x) = cosx is even. Let us prove that the
function
+X
q>(x) = lny^
1
is odd:
cp <_*) = in {=£ _ In ({±f )- = -In {±f - _„ W .
Thus, the integrand is the product of an even function by an odd

one, i.e. an odd function, therefore
cos* — X = 0.
ln\^dx 1
" 2
6.5.9. Prove the validity of the following equalities:

(a)
J
x s s'm d xdx -=0; (b) $ e cosx dx = 2 ê cosx dx\
ji 1 o
~T "T
ji
(c)
J
sinrnxcostt^djc —O (m and n natural numbers);
— JT
Q
(d) [ s'mxf (cosx)dx = 0.

-a

h b
^f(x)dx^^f(a + b — x) dx.
a a
Solution. In the right-hand integral make the substitution
X t
x^=a + b — /, dx=^ — dt y a b
b a
Then we obtain
b
j f
a
(a
baa
a
+ b—x) dx=—lf(t)dt^lf(t)dt = \f (x) dx.

Note. The relation established between the integrals can be explai-
ned geometrically.
The graph of the function f(x), considered on the interval [a, b]
+ —
y
is symmetrical to that of the function f(a b x) considered on y
the same interval, about the straight line x = ^-^-. Indeed, if the
point A lieson the x-axis and has the abscissa x, then the point A\
which is symmetrical to it about the indicated straight line, has
the abscissa x' =a + b — x. Therefore, (a + b— x') = [a + b — (a+
+b — x)] = f(x).
f f
But symmetrical figures have equal areas which are
expressed by definite integrals. And so, the proved equality is an
equality of areas of two symmetrical curvilinear trapezoids.
\f{x)g{t — x)dx=\g(x)f(t — x)dx.

Solution. Apply the substitution t — x=z in the right-hand integral;

then we have
o /
-lg(t-z)f(z)dz~lf(z)g(t- z) dz.
Jl Jl
2 T
m xdx=^ cos m xdx and apply
6.5.12. Prove the equality sin the
J
o o
obtained result in computing the following integrals:

Jl JX
T 2
cos a #dx and ^

sin 2 **!*:,
J
o o
Solution. On the basis of Problem 6.5.10 we have

Jl 71
2 , v 2
(y —
sin"
2
xdx = sin
m = cosmxdx.
J J J
Hence, in particular,
71 71
2 2
/^=^ sin 2 xdx=^ cos 2 xdx;

o o
add these integrals:

71 7l_
2 2
2/ = (sin 2 jc + cos 2
x) dx = dx = ~\
^ J
hence, / = -j
ji 2
f (sin jc) dx = 2 ^ f (sin x) dx.

o
Solution. Since
^ /
(sin x) = jj
/ (sin x) dx +^ / (sinx) dx,
it is sufficient to prove that

ji
ji T
^
f(s'mx)dx= / (sin x) dx.
J
_jx_
In the left integral make the substitution
X t
x=n — /,
71 JT
dx= —dt, T
n
Then
JT
J
f(s\nx)dx = — ^ f [sin (n
— t)] dt
= J
/ (
sin *=J / ( sin x) dx -
o o

ji n
^ xf (s\n x) dx = ^- / (sin a:) dx.

J
Solution. In the left integral make the substitution
X t
x=n — t,
dx= — dt, JT
ji
Then we obtain
J
a:/ (sin x)dx = — {[
(ji — [sin (ji — /)]d/ =
Jt
JT Jt
= 5"/ (sin/)*— S'/ (smt)dt.

Whence
%\xf (s\x\x)dx= n J
f(sin x)dx,
which is equivalent to the given equality.

6.5.15. Using the equality
s\n[n + j]x
= + cos^-|-cos2a;+
j
-x- . . . -j-cosnx,
2 im-
prove that
ji sin ^H- "^") x

^
. x
dx — n.
sin
T
6.5.16. Prove that if q>(x) =ya +a i
ôsx + b 1 sin x+a cos2x+
2
-\-b 2 sin 2.x: + . . . +an cos nx + bn sin fix, then

2ji 2ji
(a)
J
cp (jc) = Jta ;
(b)
J <P (x) cos kx dx = na k \
o o
2ji
(c)
J (p (x) sin kxdx = nb k (k=l, 2, tt).
§ Integration by Parts. Reduction Formulas
If u and t; are functions of x and have continuous derivatives,

then
b b
J
u (#) u' (a:) dx = u (x) v (x) j
&
— (x) u' (x) dx
a a
or, more briefly,

b b
ûdv = uv^ — ^vdu. a

a a
I
6.6.1. Compute the integral ^xe x dx.

o

x= u, ex dx — dv\
du = dx\ v = ex >
which quite legitimate, since the functions u

is x and v = = ex are
continuous and have continuous derivatives on the interval [0, 1].
Using the formula for integration by parts, we obtain
i i
\ xe x dx = xe x \
— \ ex dx = e — ex L = 1
§ 6.6. Integration by Parts. Reduction Formulas 295
ax sin
6.6.2. Compute the integral I= \ e bxdx.
o
= s'mbx,
u dv = e° x dx;
du = b cos bx dx, v = — e° x
Since the functions u = smbx, v = — eax together with their deri-

vatives are continuous on the interval [0, jt], the formula for in-
tegration by parts is applicable:
/= —
a
e
ax smbx
a J
e
ax cos bx
dx -
= — —a [ e ax cos
J
bx dx = —
a
- L.
1
Now let us integrate by parts the integral /t . Put
u = cos bx, dv = eax dx,

du = — b sin bx dx, v = — e ax .
a
Then
/ = —a
-
\
—e
a
ax cos bx — ,
b r
e
ax s\n bxdx =
b f e~ \\ b* b G T +l) b*_
a \ a a J a2 ~ a2 a2
f
Hence
a? + b* b{eT +\) , b{e T +\)
In particular, at a = b= \ we get
ê x s'mxdx = -j(e n +1).

6.6.3. Compute the integral ^\n 3 xdx.
2
jt
6.6.4. Compute the integral ^s'mV x dx.
Solution. First make the substitution
Vx=t, X
x = t\
dx = 2t dt y ji
2
ji
Whence
.n
4 T
J
sin J/* dx: =2 J
/ sin / dt.
Integrate by parts the latter integral.

Put
t = u; sin tdt = dw,
du = dt\ v = — cost.
Then
^
2
I
t sin tdt =2 171
~z
+) cos / dt =
2 sinH
lo
2 =2.
6.6.5. Compute the integral /= ^-^y==^- dx.

J
x 2 s\nxdx.
6.6.7. Compute the integral In =^(a 2 —x 2

)
n
dx> where n is a na-
o
tural number.
Solution.The integral can be computed by expanding the integrand
(a 2 —x 2
)
n
according to the formula of the Newton binomial, but it
involves cumbersome calculations. It is simpler to deduce a formula
for reducing the integral /„ to the integral I n _ l To this end let .
us expand the integral I n in the following way:

a a
/„ = I
(a 2 — a;
2 )"" 1
(a
2 —x 2
) dx = a 2 l n _ x — \ x(a 2 — a:
2 )"" 1 xdx
and integrate the latter integral by parts:
= x;
u (a 2 —x 2
)
n~x
xdx = dv,
du = dx\ = —^-(a —x 1
2 2 n
v ' ) (n=£0).
2n
We obtain
= *i n -i + -*r -~k I ^ ~ x2 dx = a2/« -

in
ix («
2
| o
Y'
~i 7 «-
Whence
2n
/
l
»
=aU 2
I
2/1+1 '"-I*
This formula is valid at any real n other than and — ~.

In particular, at natural n, taking into account that
Io = l dx = a>
o
we get
*~~ U
2n (2/1-2) (2/1-4) . . .
6-4-2 _ aw + 1 (2/z)H
/
(2/i+ 1) (2/i— 1) (2/1 — 3) . . . 5-3 (2/1+1)!!'
where
(2/i)!! =2-4-6 ... (2/i),
(2/i+l)!! = 1-3-5 ... (2/1+1).
6.6.8. Using the result of the preceding problem obtain the fol-
lowing formula:
1
3 ^ 5 7 "I- • • • -r V U 2/1 + l (2/2+ 1)!!
'
where are binomial coefficients.
Solution. Consider the integral

Expanding the integrand by the formula of the Newton binomial

and integrating within the limits from to 1, we get:
In = \(\—x*) n dx =
o
i
= J(1— C} x + C t
2 2
*4 - C*x« + ...+(— 1 )
n
C nn x 2n ) dx =
x ~T
o ~T~ c 7 i • • •
2/7 + 1
^1 ^2 ^3 (-1)"
3^5 7
" "
t
2n+l »
m
// w ==5 sin^jcrfA;^^ cos xdx
(m a natural number).
Solution. The substitution

=
sin x f, t
cosjcd.x: = d/,
31
"2 1
reduces the second integral to the integral
Hm = \ (1— sin 2 *) 2
cosxdx=\{\—t*) 2
dt t
considered in Problem 6.6.7 with a=^\ and n = Therefore,
the reduction formula

m— 1
(m=^0, m^= 1)
is valid here, since

If m is an odd number, the obtained reduction formula reduces

ti m to
= J
cos#djc= 1,
therefore
_ (m-l)ll
* ml!
•
If m is an even number then the reduction formula transforms

Hm into
therefore
„ (m— 1) !! ji
ji
I = ^xsm m xdx
o
(ra a natural number).
Solution. Taking advantage of the results of Problems 6.5.14 and

6.5.13, we get
oo
ji
ji ji 2
/ = §xsm
m xdx =
j sm xdx = n
m s\n m xdx,
J J
o
which, taking into consideration the result of Problem 6.6.9, gives

(ji 2
(m— 1)!!
L L
-2 '
-7^fT - lf
™ 15 even '
ji rr
1 — u m is odd.
mil
i
6.6.11. Compute the integral /n = ^ x m (\n x) n dx\ m> 0, n is a

o
natural number.
Solution. First of all note that, though the integrand f (x)

^= x m (\n x) n
has no meaning at # = it can be made continuous on the interval
[0, 1] for any m> and n > 0, by putting f (0) ^=0. Indeed,
lim x m (\nx) n = lim \x n

\nx) =
X-+ +Q X-+ +
by virtue of Problem 3.2.4.

Hence, in particular, it follows that the integral l n exists at
m> 0, n > 0. To compute it we integrate by parts, putting
u =(\nx) n y dv = x m dx,
,
du =— n(\nx) n - 1
#
— ,
ax, a = ——
xm +
m+
l
Hence,
i
The formula obtained reduces /„ to I n _ v In particular, with a na-

tural n, taking into account that
i
we get
"
= (-!)" (m+l)" +
!
/„ 1 '
6.6.12. Compute the integral / Wt „= $ xw ( 1 — x) n

dx,
o
where m and n are non-negative integers.

( l
—x )
n = a; dx = dv;
du= — n(l —x) n~l
dx; v= .
Then
is
The obtained formula
a positive integer, then,
is valid for all n
applying this
> 0,
formula
and — 1.
successively
If n
n
times, we get
n l)
/
— m+ i
/
1 m + i.»-i""( /0 +l)(m
+ 2)
7
+
==
• •
•
_ n(n— I). .[n — (n— \)] . ,
~~ m+n °'
* * '
(m+l)(m + 2)...(m + /i) *
But
l
m+n dx: xm + n + l
u0 =\x
f» 1
'
m+n+ 1 o m+ /i +
o
Hence,
. /i (/i — l) (/i — 2) 3-2-
J
«. n
(
m+ J) ( m + 2). ..(m + /i) (m +n+ I)
'
The obtained result, with m a non-negative integer, can be written

in the form
. _ m\n\
m *
n ~ (m+n+ 1)!
"

i i
(a)
J
arc tan V x dx\ (b) ^(x — \)e~ x dx\
o o
~3~
i
. . C Xdx r>
\c) (d) arc tan

j iiH^' j a: a: g(a:;
(e) J*ln(l +x 2
)dx\ (f)
J
ln(l + tanx)dx;
o o
ji
T 16
(g) ^ sin 2a: arc tan (sin x)dx\ (h) arc tan j/" V^x—ldx.
J
o 1
6.6.14. Prove that

i i
^(arccosArJ^dArÂi^y
1
— (/t — 1)
J
(arc cos a;)"
~2
dA; (n> 1).
o o
6.6.15. Prove that if f" (x) is continuous on [a, &], then the fol-
lowing formula is valid
b
S *r (x) = [bf (b)-f(b)] - [af (a)-f (a)].

a
§ 6.7. Approximating Definite Integrals

1. Trapezoidal formula. Divide the interval [a, b] into n equal
parts by points x k = a-\~kh, where h = -^

b
y
k = 0, 1, n, and
apply the formula
b
j / (jc)dx * ^ [I / (* ) +/ + . . .
+ /(*„_,) + y f (*„)] .
The error R in this formula is estimated as follows:
|R|<^, where M=a

sup \f (x)\
(assuming that the second derivative is bounded).

2. Simpson's formula. Divide the interval [a, b] into 2n equal
parts by points xk = a + kh 9
where ^ = ~^» an d apply the formula
b
^f(x)dx^ -^.{f(x <) ) + f(x + 4[f(x + f(x
in ) l ) 3 )+...
iv
+f (*,„ - 1)] +2[f(x t ) + f (* +...+f (x
4) 2n _,)]}.
Assuming that f {x) exists and is bounded, the error in this formula
is estimated in the following way:
C dx
6.7.1. Approximate the integral I=\y^r£ using the trapezoidal
o
formula at n = 10.
Solution. Let us tabulate the values of the integrand, the ordi-

nates y { =
/ (x ) (i 0, 1, t
=
10) being calculated within four de-
cimal places.
1 +*i yi=
TTT„
Xi 1 +Xi M= TT7 t
0.0000 1.0000 1.0000 0.6000 1.6000 0.6250

0.1000 1.1000 0.9091 0.7000 1.7000 0.5882
0.2000 1.2000 0.8333 0.8000 1.8000 0.5556
0.3000 1.3000 0.7692 0.9000 1.9000 0.5263
0.4000 1.4000 0.7143 1.0000 2.0000 0.5000
0.5000 1.5000 0.6667
Using the trapezoidal formula, we obtain
r dx / 1.0000+0.5000 n qoqq
+ Aftnn1
, 1 .
?
.
/=3T+7~Tol 0.9091+0.8333 + ,
+ 0.7692 + 0.7143 + 0.6667 + 0.6250 + 0.5882 + 0.5556 +

+ 0.5263 ) = 1 6.9377 = 0.6937 « 0.6938. • 1
Estimate the error in the result obtained. We have f" (x) = ^_^ 3
.
Since 0<jc<1, then |f (x)|<2. Consequently, we may take the

number 2 as M 2
and estimate the error:
l«l<72^ = 6^0< - 0017 -
We calculated the ordinates accurate to four decimal places, and

Q Q5
the round-off error does not exceed °
^ (l +9xl) = 0.00005 ^more
Q,Q Q5
precisely,
^ . 9 = 0.000045, since the ordinates y and */ 10 are
exact numbers^ . Thus, the total error due to using the trapezoidal
formula and rounding the ordinates does not exceed 0.0018.
off
Note that when computing the given integral by the Newton-
Leibniz formula we obtain
dx
l+JC
ln(l+x) '
o
= In 2^0.69315.
Thus, the error in the result obtained does not exceed 0.0007, i. e.
we have obtained a result accurate to three decimal places.
1.5
C e°'
lx
6.7.2. Evaluate by Simpson's formula the integral \ —^-dx
0.5
accurate to four decimal places.
Solution. To give a value of 2n which ensures the required accu-
racy, we find /
iv
(a;). Successively differentiating f(x) — e-^-, we get
fiv (A;
)
= ^ (0.0001a;* — 0.004a; + 0.1 3
2a;
2 — 2.4x + 24) = ^g°- 1 *,
where the polynomial in parentheses. On the interval

P (x) is
[0.5, 1.5] function cp (a;) =e°- lx increases and therefore reaches

the
its greatest value at x =1.5: (p (1.5) e°-
lb
1.2. The upper estimate = <
of the absolute value of the polynomial P (x) divided by x can be
b
obtained as the sum of moduli of its separate terms. The greatest

value of each summand is attained at a:=0.5, therefore
\P(X) 0.0001 0.004 0.12 2.4 24

^ x
+ x2 *3 + x* xb ^
< 0.0002 + 0.0 6 + 0.96 + 38.4 + 768 < 808.
1
And so, |/
iv
(a;)| < 1.2x808 < 1000. Hence, the number 1000 may
be taken as Mv
We have to compute the integral accurate to four decimal places.

To ensure such accuracy it is necessary that the sum of errors of
the method, operations and final rounding off should not exceed
0.0001. For this purpose we choose a value of 2n (which will de-
termine the step of integration h) so that the inequality
|i?|<{ -0.0001 = 5- 10" 6
is satisfied.
Solving the inequality
5
l xl 000
180 (2/2)
4
<5xl0"
we obtain
2n > 19.
Let us take 2n = 20; then the step of integration h will be equal to
A more accurate calculation shows that at 2n = 20

|Z?|<3.5xl0- 5
.
If we
calculate y t within five decimal places, i. e. with an error
not exceeding 10"
5
then the error of the final rounding off will
,
also be not greater than 10~

6
Thus, the total error will be less
.
than 4.5xl0" 5 < 0.0001.

Now compile a table of values of the function y=—^~ f° r ê va '
lues of x from 0.5 to 1.5 with the step h = 0.05. The calculations
are carried out within five decimal places.
IX,
i xi OAxi e o. yi
0.50 0.050 1.05127 2. 10254

1 0.55 0.055 1.05654 1.92098
2 0.60 0.060 1.06184 1.76973
3 0.65 0.065 1.06716 1.64178
4 0.70 0.070 1.07251 1.53216
5 0.75 0.075 1.07788 1.43717
6 0.80 0.080 1.08329 1.35411
7 0.85 0.085 1.08872 1.28085
8 0.90 0.090 1.09417 1.21574
9 0.95 0.095 1.09966 1.15754
10 1.00 0.100 1.10517 1.10517
11 1.05 0.105 1.11071 1.05782
i %i 0. l*j e 0.lXi y\
12 1.10 0.110 1.11628 1.01480

13 1.15 0.115 1.12187 0.97554
14 1.20 0.120 1.12750 0.93958
15 1.25 0.125 1.13315 0.90652
16 1.30 0.130 1 . 13883 0.87602
f\OA *70 1
17 1 .00 0. loo 1 . 14454 U.o47ol
18 1.40 0.140 1.15027 0.82162
19 1.45 0.145 1.15604 0.79727
20 1.50 0.150 1.16183 0.77455
For pictorialness sake we use the tabular data to compile the

following calculation chart:
i x\ at 1 = and
i = 20 at an odd t at an even i
0.50 2.10254
1 0.55 1.92098
2 0.60 1.76973
3 0.65 1.64178
4 0.70 1.53216
5 0.75 1.43717
6 0.80 1.35411
7 0.85 1.28085
8 0.90 1.21574
9 0.95 1.15754
10 1.00 1.10517
11 1.05 1.05782
12 1.10 1.01480
13 1.15 0.97554
14 1.20 0.93958
15 1.25 0.90652
16 1.30 0.87602
17 1.35 0.84781
18 1.40 0.82162
19 1.45 0.79727
20 1.50 0.77455
Sums 2.87709 12.02328 10.62893
Using Simpson's formula, we get

I .5
j
^^^^(2.87709 + 4x 12.02328 +
0.5
+ 2x 10.62893)==^. 72.22807- 1.2038.

6.7.3. The river is 26 m wide. The table below shows the succes-
sive depths of the river measured across its section at steps of 2 m:
X 2 4 6 8 10 12 14 16 18 20 22 24 26
y 0.3 0.9 1.7 2.1 2.8 3.4 3.3 3.0 3.5 2.9 1.7 1.2 0.8 0.6
Here x denotes the distance from one bank and y the correspond-
y
ing metres). Knowing that the

depth (in mean rate of flow is
1.3 m/sec, determine the flowrate per second Q of the water in the
river.
Solution. By the trapezoidal formula the area 5 of the cross-sec-

tion
26
S = J#dx«2 y (0.3 + 0.6) + 0.9+ 1.7 + 2.1+2.8 + 3.4 +

o
+ 3.3 + 3.0 + 3.5 + 2.9 + 1.7+1.2 + 0.8 = 55.5 (m 2 ).
Hence,
Q = 55.5 x 1.3 « 72 (m 3 /sec).
impossible to estimate the error accurately in this case. Some
It is
indirect methods of estimation enable us to indicate approximately

the order of the error. The error in 5 is about 3
2
hence, the m ,
error in Q is about 4 3
/sec. m
6.7.4. Compute the following integrals:
(a) ^y?-dx accurate to three decimal places, using Simpson's

j
n
4
formula;
i
2
(b) ê-* dx accurate to three decimal places, by the trapezoidal
o
formula.
6.7.5. By Simpson's formula, approximate the integral
1 .36
/ = ^ / (*) Ax*
1.05
if the integrand is defined by the following table:
X 1.05 1.10 1.15 1.20 1.25 1.30 1.35
fix) 2.36 2.50 2.74 3.04 3.46 3.98 4.6

( l—x at 0<*<1,
f(x) = \
at l<x<2,
(
(2— xf at 2<x<£3.
Check directly that the function
X
F(x) = \f(t)dt
is continuous on the interval [0, 3] and that its derivative at each
interior point of this interval exists and is equal to f(x).
/ x In x , r\ . . ,
f(x) =\ at x =
^ — 1 at x=\
is integrable on the interval [0, 1].
6.8.3. Can one assert that if a function is absolutely integrable

on the interval [a, b] y
then it is integrable on this interval?
6.8.4. A line tangent to the graph of the function = / (x) at the
point x=a forms an angle -j with the axis of abscissas and an
angle ~ at the point x = b.

b
Evaluate (x) dx if f" {x) is a continuous function.

J f" y
6.8.5. Prove that
[ E (x) dx = E(x) {E M ~ + E (x) [x — E (x)]. {)

308 Ch VI. The Definite Integral
ji
f* dx
6.8.6. Given the integral \
=- . Make sure that the fun-
te
J 1+ COS 2 X
o
ct ions
fc .
1
i
(xy
\
'
= —7r=-arc cos
V
1
2
}^2cosx
Vr + cos
l
= 2
x
z?
i
and r 29 m = -7=^
/
V
\
7
>^2
1
arc
j.tan ^
tan—^
1^2
are antiderivatives for the integrand. Is it possible to use both an-

tiderivatives for computing the definite integral by the Newton-
Leibniz formula? If not, which of the antiderivatives can be used?
6.8.7. For f (x) find such an antiderivative which attains the given
magnitude y = y at x =x (Cauchy's problem).
6.8.8. At what value of £ is the equality ê 2x dx=e 2 ^(b — a) ful-
filled? Show that
6.8.9. Investigate the function defined by the definite integral

X
6.8.10. Show that the inequalities

1
0.692 < 1
are valid.
6.8.1L With the aid af the inequality x^s'm x^ ^x (q^Lx^L
<£) show that \<^dx<$

6.8.12. Using the inequality smx^x— ^-(xÔ) and theSchwarz-
Bunyakovsky inequality, show that
r
1.096 < \ V xs\nxdx< 1.111.
6.8.13. Assume that integrable functions p x (x), p 2 (x) p 3 (x), p A (x) y
are given on the interval [a, b], the function p ± (x) is non-negative,
and the functions p 2 (x), p3 (x), p4 {x) satisfy the inequality
Ps (x)<Pt (x)<P 4 (*)•
Prove that
b b b
$ p3 (x) p x
(x) dx^\p 2
(x) px (x) dx^ I p,
(x) pi (x) dx.
a a a
6.8.14. Let the function f (x) be positive on the interval [a, b].
Prove that the expression
b h
a a
reaches the least value only if f (x) is constant on this interval.
6.8.15. Prove that

ji
farctanx^l /
^
J x 2 Jf sin /
o o
6.8.16. Prove that one of the antiderivatives of an even function

is an odd function, and any antiderivative of an odd function is
an even function.
6.8.17. Prove that if f (x) is a continuous periodic function with

a+T
period 7\ then the integral /= J
f(x)dx does not depend on a.
a
6.8.18. Prove that if u = u(x), v v(x) and their derivatives =

through order n are continuous on the interval [a, b], then
$
UV {n) dx=[uV Kn - 1)
—U V , {n -2
>+. . . +(—l) n - 1 U in - l)
v] \a +
a
b
+ (— 1)" I
u {n >vdx.
Chapter
7
APPLICATIONS
OF THE DEFINITE INTEGRAL
§ 7.1. Computing the Limits of Sums with

the Aid of Definite Integrals
It is often to compute the limit of a sum when the

necessary
number of summands
increases unlimitedly. In some cases such li-
mits can be found with the aid of the definite integral if it is pos-
sible to transform the given sum into an integral sum.
For instance, considering the points — —2

1
, , . . . ,
—n as points of
division of the interval [0, 1] into n equal parts of length A*=—

for each continuous function f(x) y
we have
7.1.1. Compute
— —
n
im
UI1J
-* Co
—n
jt I
\
sin
.
—
ji
n
.
h sin
. 2ji
n
,
h . . . + sin
,
1
(n
v
n
1)} Jil
.
Solution. The numbers in brackets represent the values of the

function f{x) =
s\nx at the points
ji 2ji t #
{n — 1)ji
subdividing the interval [0, ji] intorc equal parts of length \x = ^.

Therefore, if we add the summand sin^ = to our sum, the lat-
ter will be the integral sum for the function f(x) = s\nx on the
interval [0, ji].
By definition, the limit of such an integral sum as n —> oo is the

definite integral of the function / (x) sin x from to n: =
n f
lim n
\
sin—ji
n
.
— .
\-
1
.
sin — 2ji
n
4-
1
. . .
'
sin
(n—l)n.
n — — hsin
1
. nn\
n )
=
7\
=j sin xdx = — cos a:

o
= 2.
o
§ 7.1. Computing the Limits of Sums 311
7.1.2. Compute the limit
lim ( r
1
+ *
+ . . . + -
1
V
Solution. Transform the sum in parentheses in the following way:
1
+ !
+ .
.+-=1
r
V~4n 2 -l '
V~4n 2 -2 2 \ 4n i -
1
+ ..-+•
'W-± V -(4)' 4
The obtained sum is the integral sum for the function f (x)
=
= on the interval [0, 1] subdivided into n equal parts.
Y==^
The limit of this sum as n— oo is equal to the definite integral
of this function from to 1:
lim ( r
1
+ r — .-
+ . . . + r
1
) =
1
=1 , = arc sin -77

1
- il
7.1.3. Compute
lim JL
AZ -* CD
1
+/^+/HT5+/5T5+--- + /»-+3fcr,]
Solution. Transform the given expression in the following way:
3 1
The obtained sum is the integral sum for the function / (x)
=
= Y 011 ê interval therefore, by definition,
^^
[0, 3];
A™ H l+ 3
+
3
+ + /--+s5j=d) -
,
Y dx = 2VT+x = 4—2 = 2.
=
J V T^ c
dx = ^(l + x)~
312 Ch. VII. Applications of the Definite Integral
7.1.4. Using the definite integral, compute the following limits:
(C) lim
^V*+V}±:.+V'n.
(d)K ra ^(l + cosg + cosg + ...+coseLzi)»
7.1.5. Compute the limit 4= lim—i/n!

n-.cc
Solution. Let us take logarithms
In A = lim In J^JiL ^ li m — In —n + ln —n +1 1
. . . +1
In
The expression in brackets is the integral sum for the integral
= (A;ln x — x)
i
In #dA;
J {
Consequently, In A =— 1 and lim

n
= g" 1
.
n->-cc
§ 7.2. Finding Average Values of a Function

The average value of / (x) over the interval [a, b] is the number
u
li=~§f(x)dx.
i
i
% \
2
2
The square root
b __ a j
[/ (x)] of the average value of the
j
a
square ol the function is called the root mean square (rms) of the
function f (x) over [a, b].
7.2.1. Find the average value jli of the function f(x)=i/ x over
the interval [0, 1].
Solution. In this case
j/ xax = -
1
—A
o~~ 4
c
§7.2. Finding Average Values of a Function 313
7.2.2. Find the average values of the functions:
(a) / (x) = sin 2

x over [0, 2n]\
(b) over 2 ]-
/(*)=?qrr [0.
7.2.3. Determine the average length of all vertical chords of the
hyperbola ^
a
— j^b =
2 2
l over the interval a^x^2a.
Solution. The problem consists in finding the average value of
the function f (x)

= 2y = 2 — V~x 2
— 2
over the interval [a, 2a]:
2a
1 C b
'
2b
a2 K^^-f ln(* + Vxâ^ =b[2\^3-\n(2 + V3)].
7.2 .4. Find the average ordinate of the sinusoid y = sinx over
the interval [0, ji].
Solution:
—
ji jt
u
^
=— jt
[ sin xdx = - cos x
ji
- « 0.637.
ji
J
o
Rewrite the obtained result in the following way:
C
\i- n = —2 • ji = \
.
sin x dx.
Using the geometric meaning of the definite integral, we can say

2
that the area of the rectangle
& with the altitude ii
^
=— ji
and the base
ji equals the area of a figure bounded by a half-wave of the sinu-
soid y=smx, Oîx^n,
and by the #-axis.
7.2.5. Find the average length of all positive ordinates of the
circle x2 +y = 2
1.
7.2.6. Show
that the average value of the function f (x)> conti-
nuous on the interval [a, b], is the limit of the arithmetic mean
of the values of this function taken over equal intervals of the
argument x.
Solution. Subdivide the interval [a, b] into n equal parts by the
points x ( â-\-i —
b
- (/=0, 1, 2, n).
Form the arithmetic mean of the values of the function f(x) at

n points of division x ,
x ly . .., xn _ ± :
„ /(*o) + /(*i)+-.. 1
Vf/r)
/ =
This mean may be represented in the following form:
n~ l
Vn=— ^f(X l
a i)
Ax i>
t =
where
b—
AX:-
n
The latter sum is the integral sum for the function f(x), the-
refore
n ~ l b
X
l
V>n = ~a Hm =
l
lim f (*d Ax ~b~â i \ f (

x )dx \l,
which completes the solution.

7.2.7. Find the average value of pressure (p m ) varying from 2
to 10 atm if the pressure p and the volume v are related as follows:
_3_
pv 2 = 160.
Solution. As p varies from 2 to 10 atm, v traverses the interval

[4j/4, 4 j/T00]; hence
4 j/7oo
3
pm =
'
3 160u" 2
do =
4(j/l00-j/4) Jf
3
4 /4
3
320
— U
v
= —3—- « 4.32 atm.
4(^/100-^4) 4
J/ 4 ^20(^/10+^/2)
7.2.8. In hydraulics there is Bazin's formula expressing the velo-
city v of water flowing in a wide rectangular channel as a function
of the depth h at which the point under consideration is situated
below the open surface,
2
v = v -20V HL (A) ,
where v is the velocity on the open surface, is the depth of H

the channel, L its slope.
Find the average velocity v m of flow in the cross-section of the
channel.
§ 7.2. Finding Average Values of a Function 315
Solution. We have
H
7.2.9. Determine the average value of the electromotive force

Em over one period, i.e. over the time from £=0 to t T if = 9
electromotive force is computed by the formula
£=£ sin-^-
where T is the duration of the period in seconds, £ the amplitude

(the maximum value) of the electromotive force corresponding to
the value / =0.257. The fraction —
2nt
is called the phase.
Solution.
e *=t) s,n — *=ra - cos "f-Jo =0 -
Thus, the average value of the electromotive force over one pe-
riod equals zero.
7.2.10. Each of the two vertical poles OA and CD is equipped

with an electric lamp of luminous intensity i fixed at a height h.
The distance between the poles is d. Find the average illumination
of the straight line OC connecting the bases of the poles.
7.2.11. Find the average value of the square of the electromotive
force (E 2 ) m over the interval from / = to t =T ~2 (

see Prob-
lem 7.2.9).
Solution. Since
E =E sin ~y~
we have
2nt
™dt=±rEl
2 C
{
- Q0S —
j ô J sin' \ Y^-dt^
T . 4nt ~ _ £o
t- -T- Sin
4jt
-TfT-
T o"
~ 2 '
7.2.12. If a function f (x) is defined on an infinite interval

[0, oo), then its average value will be
b
\i= lim ~ \
f(x)dx 9
if this limit exists. Find the average power consumption of an

alternating-current circuit if the current intensity / and voltage u
are expressed by the following formulas, respectively:
/=/ O
cos (cot + a);
u=u cos (cot +a+ cp),
where cp is the constant phase shift of the voltage as compared

with the current intensity (the parameters co and a will not enter
into the expression for the average power).
Solution. The average power consumption

T
wm = lim ~ C/ cos((o/-f a) u cos (cot +a -\- w) dt
Taking into consideration that
cos a cos P = y [cos (a + P) + cos (a — P)]

we will get
T
wm = lim ^ f
[cos (2a>/ + 2a + cp) + cos cp] dt =
= Iim
|ô. sin(2cor + 2a + cp)-sin (
2a + + cp)
^ cos
^|
= ^
Hence, it is clear why so much importance is attached to the
quantity coscp in electrical engineering.
7.2.13. Find the average value jli of the function f (x) over the
indicated intervals:
(a) f(x)=2x 2 + 1 over [0, 1];
(b) /(*)=- 7 over f

1
'
23 ;
(c) f(x) = — 2x + 3
3* over [0, 2].
7.2.14. A body fa lling to the ground from a state of rest acqui-

res a velocity v l = \^2gs^ on covering a vertical path s=s v Show
that the average velocity v m over this path is equal to -y-
7.2.15. The cross-section of the trough has the form of a para-

bolic segment with a base a and depth h. Find the average depth
of the trough.
7.2.16. Find the average value l m of alternating current intensity
over time interval from to — (see Problem 7.2.12).

§7.3. Computing Areas in Rectangular Coordinates 317
7.2.17. Prove that the average value of the focal radius of an

ellipse p:
1 — 8 COS (p
, where p=—\ a, b are the semi-axes and e is
eccentricity, is equal to b.
7.2.18. the segment AB of length a a point P is taken at a

On
distance x from the end-point A. Show that the average value of
the areas of the rectangles constructed on the segments AP and
a2
PB is equal to -g-
7.2.19. Find the average value of the function
cos2 *
f(1 X ')=
'
sin 2 # + 4 cos 2
x
over the interval ^0, yj. Check directly that this average, equal
to ~, is the value of the function f (x) for a certain x= l lying
within the indicated interval.
§ 7.3. Computing Areas in Rectangular Coordinates
If a plane figure is bounded by the straight lines x=a, x =

= b(a<b) and the curves y=y 1 {x) y=y 2 {x), provided y 1 {x)^ i z
^y 2 (
x) (a^x^b), then its area is computed by the formula
b
s= 5 [ftW- yi (x)] dx.
In certain cases the left boundary x=a (or the right boundary
x = b) can degenerate into a point of intersection of the curves
(a)
Fig. 65
y =y and y
l
(x) =
y 2 (x). Then a and b are found as the abscissas
of the points of intersection of the indicated curves (Fig. 65, a, b).
7.3.1. Compute the area of the figure bounded by the straight
lines x 0, x =
2 and the curves y= 2x , y 2x—x2 (Fig. 66). = =
Solution. Since the maximum of the function r/ = 2x— x 2

is at-
tained at the point x=l and is equal to 1, and the function
y=2 ^l
x on the interval [0, 2], we have
s-ii 2 "-< 2 *-*°>i«-£i:-("-?)i:-^-f

o
7.3.2. Compute the area of the figure bounded by the parabolas

x= —2y\ x=l — 3y 2 (Fig. 67).
Fig. 66 Fig. 67
Solution. Solving the system of equations
I x = 2y 2 ;
\ x =l—3y\
find the ordinates of the points of intersection of the curves y x = — 1,

y2 = 1 Since 1
. 3y 2 ^—
2y 2 for —
then we have
S= f [(l-m-(-W)]dy=2(y-^)\l = l
Find the area of the
7.3.3.
figure contained between the 2
parabola x 2 =\y and the witch x ^y
of Agnesi y = -p-^ g
(see Fig. \ 2
68). U = J-
Solution. Find the abscis- <Z£f i
sas of the points A and C of

-z 2
intersection of the curves. For
this purpose eliminate y from Fig. 68
§ 7.3. Computing Areas in Rectangular Coordinates 319
the system of equations
y= —
x2
.
whence
a:
2
8
+4
- = 4
or x* + 4x — 32=0. 2
The real roots of this equation are the points xt = —2 and

8
#2 = 2. As is seen from the figure,
x2 +4
A.
on the interval
[— 2, 2]. (It is also possible to ascertain this by directly computing
the values of these functions at any point inside the interval, for
instance, at jc = 0.)
Consequently,
2
+4 dx = ^4 arc tan-|- — -V
\2J
7.3.4. Find the area of the figure bounded by the parabola

y = x +l
2
and the straight line x + y = 3.
7.3.5. Compute the area of the figure which lies in the first qua-
drant inside the circle x2 y
2
+ =
3a2 and is bounded by the parabo-
las x2 = 2ay and y 2 = 2ax(a > 0) (Fig.
69).
Solution. Find the abscissa of the po-

int A of intersection of the parabola
y =2ax and the
2
circle x2 y -3a
2 2
+ .
Eliminating y from the system of equa-

tions
\
= 2ax, y
2
we obtain x2 + 2ax — 3a = 0, 2
whence we Fig. 69
get the only positive root: x A a. Analo- =
gously, we find the abscissa of the point D of intersection of the
circle x2 y
2
+ =
3a2 and the parabola x2 = 2ay\ xD = a\r
2^.
Thus, the sought-for area is equal to
aVl
Sr= [y*(x)—yi(x)]dx 9
J
where y x (x) =^ = \V2,ax for 0<x<a,

\
r
V 3a 2
— 2
for a<xâV2.
By the additivity property of the integral

a aV2
V 2a • it
3
x 2
— 6a s- +1
2
V 3a 2 — x + — arc sin —-f——7t
2
26a Jo
J o L a |^3
^~2
2 ^~2
3
a2 6^2
3a 2
( arcsin —f=
drLMn }A 3
— arcsm —|A 1
\^ 3
1
arcsin -tt a\
3 2
Here we make use of the trigonometric formula:
arc sin a — arc sin p = arc sin (a K — — p Vl—a 1

2 2
) (aP > 0)
for transforming
arcsin |/^y— arcsin = arcsin ^ 3 V3 V3

= arcsin y .
7.3.6. Compute the area of the figure lying in the first quadrant
and bounded by the curves y 2 4x = y
x2 \y and x2 y
2
5. = + =
7.3.7. Compute the area of the
figure bounded by the lines y =
x+l, y = cosx and the A:-axis
(Fig. 70).
-7 1 Tt/2
Solution. The function
Fig. 70
(
x+ 1 if —1 <x<0,
y=f(x)-. cos a: if 0<x<-|
I
is continuous on the interval — y 1, . The area of the curvili-
near trapezoid is equal to
5= j
f(x)dx= ^ {x+ l)dx + ^cosxdx = + sinx 2 = -
7.3.8. Find the area of the segment of the curve y l = x ~x

6 l
if
the line x =
2 is the chord determining the segment.
From the
Solution. equality y
2
x 2 (x- -1) it follows that =
2
* (;e— 1)>0, therefore either x or In other words, the =
domain of definition of the implicit function y
2
x3 2
consists = —
of the point x =
and the interval [1, oo). In computing the area
the isolated point (0, 0) does not play any role, therefore, the
interval of integration is [1, 2] (see Fig. 71).
Passing over to explicit representation
y = ±.xVx— 1, we see
that the segment is bounded abov e by the curve y = xVx— and 1
below by the curve y = —xVx—l. Hence,

2 2
S_ J [x\fjT^l — (—x\fl^l)]dx = 2lx)/'x—\dx.

x— l-/ 2
,
dx=2t dt, 1
2 1
Then
1
S = 4j(/'+l)/«<tf = 4[£ + f);-M5.

7.3.9. Determine the area of the figure
bounded by two branches of the curve
(y— x) 2 = x 3 and the straight line x=l.
Solution. Note first of all that y, as an Fig. 71
implicit function of x, is defined only for
xÔ; the left side of the equation is always non-negative. Now
we the equations of two branches of the curve y = x — xVx,
find
y =x+x V Since xÔ, we have

x. x + xVx^x— xV x, and
therefore
S=^(x + xV~x—x +xV~x)dx = 2 §*V~x dx-
7.3.10. Compute the area enclosed by the loop of the curve

y*=x(x-\) 2 .
Solution. The domain of definition of the implicit function y is

the interval 0^jc< +
oo. Since the equation of the curve conta-
ins y to the second power, the curve is symmetrical about the
322 Ch VII. Applications of the Definite Integral
x-axis. The positive branch y x

(x) is given by the equation
y X) _V x \x- 1 - y-{x _
( 1
<j
l)f x > L
The common points of the symmetrical branches y l
(x) and y 2 {x) =
= — y (x) must lie on the x-axis. But y (x) = *
x A V \
x— 1 1
= only
at x l = and at x 2 = 1. _
Consequently, the loop is formed by the curves y = \f x(\— x)
and y = — Y x(l—x), 0<*<1 (see Fig. 72), the area enclosed
being
S=2jV*(l — x)dx = 2 ^x —x 2 2
)dx = ^.
7.3.11. Find the area enclosed by the loop of the curve

y* = {x-\)(x-2)K
2
y =x(x-1)
Fig. 72 Fig.73
7.3.12. Find the area of the figure bounded by the parabola

y= — x 2 — 2x + 3, the line tangent to it at the point /W(2, —5)
and the y-axls.
Solution. The equation of the tangent at the point M(2 —5) 9
has the form y 5 + =—

6(x— 2) or y = 7 €>x. Since the branches —
of the parabola are directed downward, the parabola lies below the
tangent, i. e. 7— — +
6a:>— x2 2x 3 on the interval [0,2] (Fig. 73).
Hence,
2 2
S= j*[7 — 6x— (— x — 2x + l
3)] dx = J(x2 — 4x + 4)dx = -|
7.3.13. Find the area bounded by the parabola 2

2x+2, y=x —
the line tangent to it at the point 7W(3, 5) and the axis of ordi-
nates.
7.3.14. We take on the ellipse
£ + £-1 (a>b)
a point M(x,
y) lying in the first quadrant.
Show that the sector of the ellipse bounded by its semi-major
axis and the focal radius drawn to the point M has an area
So =
ab
-77-
2
arc cos —xa .
With the aid of this result deduce a formula for computing the
area of the entire ellipse.
y
1 M(x,y)
Fig. 74 Fig. 75
Solution. We have (Fig. 74):
Sqmao — âom#+ S MABM ;

S^ 0MB — y
a a
Smabm = §ydx = §± V~¥=T* dt = ~[t |/^=7~2 + «2 arc sin

|
X X
= £ [-xVtf^Z+cP (y- arc sin

j
Since
2
— arc sin —a = arc cos—a ,
1
we obtain
>MABM — 2a -x Va — x + a
2 2 2
arc cos ~
j
Hence
Sqmao — $ao mb + Smabm + ~y arc cos •

At x = 0, the sector becomes a quarter of the ellipse, i.e.
—4 êllipse
9 "
ab
2
A
--arccos0 =—ab n
y=T
ab
•
and consequently, S ellipse = nab. At a = b we get the area of a circle

^circle
= •
7.3.15. Find the area bounded by the parabolas y = 4x #=g-

2
,
and the straight line y = 2.

Solution. In this case it is advisable to integrate with respect
to y and take advantage of the symmetry of the figure (see Fig. 75).
Therefore, solving the equations of the parabolas for x, we have:
x=±
By symmetry of the figure about the (/-axis the area sought is
equal to the doubled area S 0AB0 :
2 ?
S = 2S 0AB0 = 2
j (3
V~y-\ V~y ) dy= 5 J Vy dy = ^p-
7.3.16. From an arbitrary point M(x, y) of the curve y x m =
(m > 0)
perpendiculars MN
and ML(x>0) are dropped onto the
coordinate axes. What part of the area of the rectangle ON does ML
the area ONMO
(Fig. 76) constitute?
Fig. 77
7.3.17. Prove that the areas S 5X , ,

S2 53, , bounded by the
x-axis and half-waves of the curve y = e~« x s\n$x, x^0 form a
y
geometric progression with the common ratio q = e P.

Solution. The curve of Fig. 77 intersects the positive semi-axis
Ox at the points where sinpx=0, whence
Xn — nil
^ i n — u, /-v «
l,
q
z, ... .
The function y = e" ax sin fix is positive in the intervals (x 2ki x 2k+1 )
and negative in (x 2k+l x2k+2 ) i.e. the sign of the function in the
> y
interval (xn xn+1 ) coincides with that of the number (

,
1)". Therefore —
(n+ 1) jt (n+ l)jt
Sn = ^ \ y \dx={—\) n \ e-« x s\nfixdx.

nn
3
But the indefinite integral is equal to
e~ ax sin fix dx =—a a

+fl
a ^
a s * n + P cos P*) + ^-
Consequently,
(n+ 1) ji
= [^^(asinpx + pcosPx)
13
a
[g-a(«+l)«/pp (_l)»+i_ga«JV3p (—1)"] =
tP
*« 2
+P
Hence
-a(/z + I ) jt/3
7.3.18. Find the areas enclosed between the circle x 2 +y 2 - 2x+

y= — x + 2x+l—2\/ ~3.
r
+4(/— 11=0 and the parabola *
Solution. Rewriting the equations of the curves, we have:

(x-l) 2
+ (*/+2) 2
=_t6,
y = -{x-\f-2 V 3 + 2.
Consequently, the centre of the
circle the point C(l,
lies at 2) —
and the radius of the circle equals
4. The axis of the parabola coin-
cides with the straight linex=l
and its vertex lies at the point
fl(l, 2, -2J/T) (Fig. 78).
The area S ABDFA of the smaller
figure is found by the formula
u
SaBDFA = S (*/par — i/circle) dXf
Fig. 78
where xA and x D are determined from the system of equations
j
(*-l)« + (» + 2)« = 16,_
\ y+ =
2 -(x-\) -2V3 +4, 2
whence x A =— 1, xD = 3.
Hence,
3
Sabdfa =
-
J
1
t(-^
2
+ 2x + - 2,/T + 2 + VT6-(*-l)*)]
1
> (
=
= [-j + x* + (3-2V3) x + x 16— (at— 1)* + ~V

+ y arc sin ~1J = — 8 |/T + 2 J/"T2 + 16 arc sin 1 =
*
32 . ,
8
The area of the second figure is easy to determine.

Note. The computation of the integral can be simplified by using
the shift x 1 — =
2 and taking advantage of the evenness of the
integrand.
7.3.19. Compute the area bounded by the curves y = (x — 4)
2
,
r/ = 1
6 x2 and the #-axis.
7.3.20. Compute the area enclosed between the parabolas
x = y2 \
x=^y 2
+l.
7.3.21. Compute the area of the portions cut off by the hyper-
bola x 1 3y
2
—
from the ellipse x 2=\
4y
2
+ = 8.
7.3.22. Compute the area enclosed by the curve y 2 =(\ —x 2
)
3
.

4 (y 2 ~~x 2 ) +x 3
== 0.
7.3.24. Compute the area of the bounded by the curve

figure
r
Yx +]/' y = \ and the straight line x-\-y=l.
7.3.25. Compute the area of the figure enclosed by the curve
y
2
= x (l— x
2 2
).

x3 + x — y = 0.
2 2
7.3.27. Compute the area bounded by the axis of ordinates and

the curve x 2
y (l y). = —
7.3.28. Compute the
area bounded by the curve y x* 2x 3 = — +
+ # + 3, 2
the axis of abscissas and two ordinates corresponding to
the points of minimum of the function y(x).
§7.4. Areas with Parametrically Represented Boundaries 327
§ 7.4. Computing Areas with Parametrically Represented

Boundaries
If the boundary of a figure is represented by parametric equations
x = x(t), y=y(t) 9
mulas:
S
Art
then the area
= -^y(t)x'(t)dt;
u
of the figure
S=§x(t)y'(t)dt;
a
is evaluated by one
S=
of the
| J (xy' -yx') dt,

a
three for
where a and P are the values of the parameter / corresponding

respectively to the beginning and the end of the traversal of the
contour in the positive direction (the figure remains on the left).
7.4.1. Compute the area enclosed by the ellipse
x = acos/, y=bs'mt (0^t^2n).

Solution. Here it is convenient first to compute
xy' ~—yx'=a cos txb cos / + b sin / x a sin t = ab.
Hence
2,1 2ji
S =yj i
x y' — yx')dt = -j^ abdt = nab.
o o
2 _2_
7.4.2. Find the area enclosed by the astroid

T 3
=1-
(^) +("f)
Solution. Let us write the equation of the astroid in parametric
form: a; =
a cos 3 /, */ asin 3 /, =
Here it is also conve- 0^/^2k.
nient to evaluate first
xy' —yx' =a 2
(cos 3 1 • 3 sin 2 / cos / + sin 3
/ • 3 cos 2 / sin t)=
= 3a 2
cos 2 /sin 2 /.
Hence,
2jt 2ji
S = -i (xy'—yx')dt=ja*§ sin 2 2tdt = \a 2

n.
j
o o
7.4.3. Find the area of the region bounded by an arc of the

cycloid x = a(t — sin/), y = a(l — cos/) and the #-axis.
Solution. Here the contour consists of an arc of the cycloid
(0</<2ji) and a segment of the #-axis (0^x^2na). Let us
apply the formula S = — ^yx'dt.

Since on the segment of the x-axis we have r/ = 0, it only re-

mains to compute the integral (taking into account the direction
of a boundary traversal):
2ji
^ a (I— cos*)a(l— cost)dt=a 2

$ (1— cos t) 2 dt =
2n
=a 2
j 2cos/ + 4"0+ cos2 ')] dt = 3na 2
.
7.4.4. Compute the area of the region enclosed by the curve

jc = asin/, y = bs\n2t.
Solution. When constructing the curve one should bear in mind
that it is symmetrical about the axes of coordinates. Indeed, if we
substitute ji t —
for t the variable x remains unchanged, while y
y
only changes its sign; consequently, the curve is symmetrical about

the ;t-axis. When substituting n t +
for / the variable y remains
unchanged, and x only changes its sign, which means that the
curve is symmetrical about the y-axis.
x=asint
Fig. 79
Furthermore, since the functions x =

a sin/; y =
bsm2t have a
common period 2jt, it is sufficient to confine ourselves to the fol-
lowing interval of variation of the parameter: 0^.t^l2n.
From the equations of the curve it readily follows that the va-
riables x and y simultaneously retain non-negative values only
when the parameter t varies on the interval u '

-2 therefore at
0</<y we obtain the portion of the curve situated in the first
quadrant. The curve is shown in Fig. 79.

As is seen from the figure, it is sufficient to evaluate the area
enclosed by one loop of the curve corresponding to the variation
§ 7.4. Areas with Parametrically Represented Boundaries 329
of the parameter t from to n and then to double the result
5=2
oo
ji
J
yx' dt =2
ji
^ b sin 2t x a cos / dt
ji
= \ab ^ cos
o
2
t smt dt =
7.4.5. Find the area of the region enclosed by the loop of the
curve
%=4(6-0; 0=4(6-0-
Solution. Locate the points of self-intersection of the curve. Both
functions x(t) and y(t) are defined throughout the entire number
scale oo — <
t <oo.
At the point of self-intersection the values of the abscissa (and

ordinate) coincide at different values of the parameter. Since x=
= 3
— y(/ — = 3±
2
3) , the abscissas coincide at t 1 'k. For the func-
tion y (t) to take on one and the same value at the same values
of the parameter /, the equality iid^(3^X) = -^^(3 + X) (
must
be fulfilled for X=^=0, whence X = ±3.
Fig. 80 Fig. 81
Thus, at ^= and at t2 = €> we have x(t x ) = x(t^) = 0, and

y(t l )= y(t 2 ) = 0, i.e. the point (0, 0) is the only point of self-
intersection. When t changes from to 6, the points of the curve
are found in the first quadrant. As / varies from to 3, the
point M(x, y) describes the lower part of the loop, since in the
3t x
indicated interval x(t) and y(t) = -^- increase, and then the func-
tion x(t) begins to decrease, while y (t) keeps increasing. Fi-
still
gure 80 shows the traversal of the curve corresponding to increas-
ing t (the figure remains on the left).
In computing the area enclosed by the loop sought it is conve-

nient to use the formula
(6-0Ûl 27
'
24 5
7.4.6. Find the area enclosed by the loop of the curve: x = t

2
\
y = t—J-
7.4.7. Compute the area enclosed by the cardioid: x =
= a cos t (1 +cos /); y = as\n t (1 +cos /).
Solution. Since a;(/) and */(/) are periodic functions, it is suffi-

cient to consider the interval [ ji, ji]. The curve is symmetrical
about the x-axis, since on substituting t for / the value of the —

variable x remains unchanged, while y only changes its sign, and
i/>0 as / varies from to ji.
As / changes from to jt the function u = cost decreases from
1 to — 1, and the abscissa x = au (1 + u) =a

^
— j + (û -f
y)
2
first decreases from x = 2a to x\ =— i and then increases
to a:| u= _ 1 = 0. We can show that the ordinate y increases on the

interval and decreases on the interval ^y</^Ji^.
The curve is shown in Fig. 81, the arrow indicating the direc-
tion of its traversal as t increases.
Consequently,
ji ji
S=
y §(xy' — yx') dt = a 2
J
( 1 + cos t)
2
dt = j na\
-JT

x — cos/, y = bs\n 3
1.
7.4.9. Compute the areas enclosed by the loops of the curves:

(a) x=t —\, 2
y = — t
3
t;
(b) x = 2t — t
2
\ y = 2t —P\ 2
(c) x=t 2
;
y=L (3 - n .

x = a cos/; y = b sin /cos 2 /.
7.4.11. Compute the area enclosed by the evolute of the ellipse
x = — cos 3
/; y~ — ^-sin a
/; c
2
=a*—b 2
.
§ 7.5. The Area in Polar Coordinates 331
§ 7.5. The Area of a Curvilinear Sector in Polar

Coordinates
In polar coordinates the area of a sector bounded by the curve
p = p(cp) and the rays q) x and (p 2 =a
P is expressed by the integral =
S = -j jp 2 (<P)d<P-
7.5.1. Find the area of the region situated in the first quadrant
and bounded by the parabola y
2
= 4ax and the straight lines
y=x —
a and x = a.
Solution. Let us introduce a polar system of coordinates by
placing the pole at the focus F of the parabola and directing the
polar axis in the positive direction
along the #-axis. Then the equation of
the parabola will be p^
1
—— —-
coscp
, whe-
re p is the parameter of the parabola.
In this case p =
2a, and the focus F has
the coordinates (a, 0). Hence, the equa-
tion of the parabola will acquire the
form p =i — zzz^z ,
'
and those of the
l cos cp
straight lines will become cp = and
cp = (Fig. 82). Therefore,

Jl
4a 2
SpARP
FABF
"
2 (f— COS(p) 2
dcp = 2d 2 r dcp
\
.
4sin^"
JT
Changing the variable:

Z
cot tt = z, 2 sin 2 (cp/2)

dz, ji/4 cot (k/8)
k/2 1
we obtain
cot (Jt/8)
S FABF = a* (i +z .) dz=aBfl .(cot-J + lcot»i-l-l)

|
4*
or, taking into account that cot ~= 1
"t^fj"/
sin (jx/4)
= +K 2
1 ,
>FABF = 2fl"(l+y^2
7.5.2. Compute the area of the region enclosed by
(a) the cardioid p = l+cos(p;

(b) the curve p = acosq).
7.5.3. Find the area of the regions bounded by the curve
p =
2acos3(p and the arcs of the circle p a and situated outside =
the circle.
r = -y»
2ji
Solution. Since the function p = 2acos3q) has a period
the radius vector describes three equal loops of the curve as q) va-
ries between —
n and jx. Permissible values for q) are those at
which cos3(p^0, whence
_£ + ^ <(p< *
+ {k = 0i ±lf ±2 , ...)•
Consequently, one of the loops is described as q) varies between

— and and the other two loops as cp varies between y and
5ji i v a 7ji j 3ji ,
-g- , and between -g- and respect-
ively (Fig. 83). Cutting out the parts,

belonging to the circle p=a, we get
the figure whose area is sought. Cle-
arly, it is equal to the triple area
Smlnm*
Let us find the polar coordinates of
the points of intersection and N. M
For this purpose solve the equation
2a cos 3cp = a, i . e cos 3(p
. = . Between
—~ and y only the roots — and

Fig. 83
~q" (k=0) are found. Thus, the point N

is specified by the polar angle q) x =—y , and the point M by q) 2 =y .
As is seen from the figure,
&MLNM ~ SoMLNO ÔMNO

~
JT/9 JT/9
=| j 4a 2 cos 2 3(pdq) L
J
a2 dcp = a2 ^+ iLij .
-!t/9 -Jt/9
7.5.4. Compute the area of the figure bounded by the circle

p=3|/"2 acoscp and p = 3asincp.
Solution. The first circle lies in the right half-plane and passes
through the pole p = 0, touching the vertical line. The second circle
§ 7.5. The Area in Polar Coordinates 333
is in the upper half-plane and passes through the pole

situated
p = 0,
touching the horizontal line. Consequently, the pole is a point
of intersection of the circles. The other point of intersection of the
circles B is found from the equation 3|/2 acoscp = 3a sin q), whence
B (arc tan 1^2 ,
a|/6). As is seen from Fig. 84, the sought-for area
V3'
A
f
Fig. 84 Fig. 85
S is equal to the sum cf the areas of the circular segments_CMBO

and OCBO adjoining each other along the ray (p=arctan^2 The .
arc BAO is described by the end-point of the polar radius p of

the first circle for arctan}/2 <(p^y, and the arc OCB by the
end-point of the polar radius p of the second circle for O^cp^
ârctan^2 . Therefore
J
cos 2 cp dq> =4 a2 (t — arctan ^ 2 >
arc tan V 2
arc tan V 2
S 0CB0 -=Y a2
J
sin 2 q)dq) = -|-a 2
ârc tan|/2 —^p) •
o
Hence,
Soaho + Socbo = 2.25a 2
(ji— arc tan ]/Y—VT).
7.5.5. Find the area of the figure cut out by the circle p = ]f3 sin cp
from the cardioid p = l-f-cos(p (Fig. 85).

Solution. Let us first find the points of intersection of these cur-
ves. To this end solve the system
I p = |/~3 sin cp, O^cp^rc,

\ p == + cos 1 q),
whence <Pi =y > ^2 =n -

The sought-for area is the sum of two areas: one is a circular

segment, the other a segment of the cardioid; the segments adjoin
each other along the ray <P
=y . The arc BAO is described by the
end-point of the polar radius p of the cardioid as the polar angle cp
changes from y to ji, and the arc OGB by the end-point of the
polar radius p of the circle for O^cp^y.

Therefore
S = -I 3 sin 2 cp dcp +-I (1 + cos cp)

2
dcp =
J J
= 4(<P-^)|f + |(cp + 2sincp + f sin 2<f
= T (lt ]/T).
7.5.6. Find the area of the figure bounded by the cardioid

p = a(l— coscp) and the circle p = a.
7.5.7. Find the area of the region enclosed by the loop of the
folium of Dtscartes x 9 y* 3axy. + =
Solution. Let us pass over tj polar coordinates using the usual
. « . formulas x p cosq), # psincp. Then = =
^ the equation of the curve is:
3
p (cos
3
cp + sin 3
cp) = 3ap 2
sin cp cos cp,
or
3a sin (p cos cp
cos 3 cp + sin 3
cp
3a sin 2cp
(sincp + cos(p) (2 — sin 2cp)
It follows from this equation that,
firstly, p = at cp = and at cp = ,
Fig. 86
and secondly, p — oo as cp
3ji
r and
.
cp ——y . The latter means that
the folium of Descartes has an asymptote, whose equation y =

= x —
a can be found in the usual way in rectangular coordinates.
Consequently, the loop of the folium of Descartes is described
as cp changes from to y and is situated in the first quadrant
(see Fig, 86).

$ 7.5. The Area in Polar Coordinates 335
Thus, the sought-for area is equal to
1 C 9a 2 cos 2 cp sin 2 cp ,
>OAO
J (cos 3 (\ + sin 3
cp) 2 ^
Taking advantage of the curve's symmetry about the bisector
y = x, i.e. about the ray cp = ~, we can compute the area of half
of the loop ^from cp = to (

P=-j) an(J then double it. This enab-
les us to apply the substitution
9 z
tan cp = 2,
dcp
cos 2 (p
T 1
which gives
4 1
_ qya
'
2
C cos 2 cp sin 2 cp ,
2 f 2 tfz
~* y " 2a(p ~""
ôao J (cos 3 cp+sin 3 cp) Jd + z 3 2
)
Still new substitution

z V
1+ = v, 2
3
3z dz = da,
2 1
1 2
leads to the integral
dv 3
~ 2
a2 .
7.5.8. Compute the area of the region enclosed by one loop of

the curves:
(a) p=acos2cp; (b) p=<2sin2(p.
7.5.9. Compute the area enclosed by the portion of the cardioid

p =a (1 —cos cp) lying inside the circle p=acoscp.

p =osin(pcos 2 (p, a > 0.
p =a cos 3
(| (a > 0).
7.5.12. Compute the area of the portion flying inside the circle
p = yL=^ of the figure bounded by the Bernoulli's lemniscate p =

=aVcos 2cp.
7.5.13. Passing over to polar coordinates, compute the area of
the region enclosed by the curve (x 2 2 3
y ) Aa 2 x 2 y 2 + = .
7.5.14. Passing over to polar coordinates, evaluate the area of

the region enclosed by the curve x* y* a 2 (x 2 2
y ). + = +
§ 7.6. Computing the Volume of a Solid
The volume of a solid is expressed by the integral
V = \S{x)dx
a
where S(x) is the area of the section of the solid by a plane per-
pendicular to the x-axis at the point with abscissa x\ a and b are
the left and right boundaries of variation of x. The function S(x)
is supposed to be known and continuously changing as x varies
between a and b.
The volume V x of a solid generated by revolution about the
x-axis of the curvilinear trapezoid bounded by the curve y f(x) =
(f(x)^0) the x-axis and the straight lines x a and x b (a<b)
y
= =
is expressed by the integral
b
Vx = ji
\y
2
dx.
a
The volume Vx of a solid obtained by revolving about the x-axis

the figure bounded by the curves y = y 1 (x) and y = y 2 (x) [0 y x (x) < <
^^W] an d the straight lines x=a y
x =b is expressed by the
integral
V x = n\{y\—y\)dx.
a
If the curveis represented parametrically or in polar coordinates,

the appropriate change of the variable should be made in the above
formulas.
7.6.1. Find the volume of the ellipsoid

Solution. The section of the ellipsoid by the plane # = const is
an ellipse (Fig. 87)
+ = i
with semi-axes b y 1
— ;
cy 1 — . Hence the area of the
section (see Problem 7.4.1)
r i—^
S(x)=TLbJ/ r xc ^ = nbc[\— (—a:
Therefore the volume V of the ellipsoid is
— ^ dx = nbc x — x
3 a 4
V = ^ nbc ^
1
3 a2 -a
= o
n u
abc.
In the particular case a =b=c the ellipsoid turns into a sphere,

and we have Inhere =• 4 na*
Fig. 87 Fig. 88
7.6.2. Compute the volume of the solid spherical segment of two

bases cut out by the planes x 2 and x =
3 from the sphere =
x2 +y +z =
2 2
16.
7.6.3. The axes of two identical cylinders with bases of radius a

intersect at right angles. Find the volume of the solid constituting
the common portion of the two cylinders.
Solution. Take the axes of the cylinders to be the y- and e-axis
(Fig. 88). The solid OABCD constitutes one-eighth of the sought-
for solid.
Let us cut this solid by a plane perpendicular to the #-axis at
a distance x from 0. In the section we get a square EFKL with
side EF = Va — 2 2
, therefore 5 (x) =a —x 2 2
and V =8 (a
2
— **) dx =
7.6.4. all chords (parallel to one and the same direction)

On of
a circle radius R symmetrical parabolic segments of the same

of
altitude h are constructed. The planes of the segments are perpen-
dicular to the plane of the circle.
Find the volume of the solid thus obtained (Fig. 89).
Fig. 89 Fig.
Solution. First compute the area of the parabolic segment with

base a and altitude h. If we arrange the axes of coordinates as
indicated in Fig. 90, then the equation of the parabola will be
y = ax 2 h. +
Determine the parameter a. Substituting the coordinates of the
point b(Jj, 0^, we get = aj-M, whence a= — ^; hence the
equation of the parabola isy=—-a^x' + h> and 1

the desired area
a_ o_
2 2
S = 2 §ydx = 2^—^x +h}dx=jah. 2
Now find the volume of the solid. If the axes of coordinates are
arranged as indicated in Fig. 89, then in the section of the solid
by a plane perpendicular to the x-axis at the point with abscissa x
we obtain a parabolic segment of area S^~ah y
where a = 2y =
= 2/ R — x2 2
. Hence,
R R
S(x) = j\^W^ handV=[S(x)dx^jh^W^ dx = ^ 2 2
nhR\
-R
7.6.5. The plane of a moving triangle remains perpendicular lo

the fixed diameter of a circle of radius a: the base of the triangle
is a of the circle, and its
chord vertex lies on a straight line pa-
rallel the fixed diameter at a distance h from the plane of the
to
circle. Find the volume of the solid generated by the movement
of this triangle from one end of the diameter to the other.
7.6.6. Compute the volume of the solid generated by revolving
about the x-axis the area bounded by the axes of coordinates and
_L _L _L
the parabola x 2 -\-y 2 2
=a '
Solution. Let us find the points of intersection of the curve and

the axes of coordinates: at x y a, at y x a. Thus, we= = = =
have the interval of integration [0, a],
( - — -Y
From the equation of the parabola we get y= \a 2
x 2
J ; there-
fore
a a _1_
4 a _3_ _1_
_1
V =n 2
y dx =n \a
2
—x 2
dx =n â2 —4a 2
x
2
+ 6ax-
J J J j
— Aa 2
x 2
+x jdx= ^ na*.
2
7.6.7. The figure bounded by an arc of the sinusoid y s\nx = y
the axis of ordinates and the straight line \ revolves about y^

the y-ax\s (Fig. 91).
1 1
I
1
1
1
>
1/2 2
Fie. 91 Fig. 92
Compute the volume V of the solid of revolution thus generated.

Solution. The inverse function x = arcsinj/ is considered on the
interval [0, 1]. Therefore
V= 2
j\^ x dy = n^ (arc sin y) 2
dy.
Apply the substitution arc smy = t. Hence

y
*/= sin t t
dy = cost dt,
1 n/2
V = jt n2 —&)
And so, \ t
2
cost dt. Integrating by parts, we get y=n (
4
i
7.6.8. Compute the volume

of the solid generated by revolving
about the the figure bounded by the parabola */
.xr-axis 0.25x 2 2 = +
and the straight line 5x 8^/ 14 0. — + =
Solution.The solid is obtained by revolving the area ABCA
(Fig.92) about the #-axis. To find the abscissas of the points A
and B solve the system of equations:
y = ±-x + 2, 2
5x—8y+H== 0.
Whence x i4 =y; x B =2. In our case y x (x) =^ +2 2
and y 2 (x) =
= (5/8)jc + 7/4. Hence,
, 891
lra(T*+ 7 ) -(T + 2 )']^' JC,
1 280
Jl.
1/2

about the (/-axis the figure bounded by the parabolas y x 2 and =
8x =
y*.
Solution It is obvious that x 2 (y) = \^y^x l (y) = ^- on the in-
terval from the origin the coordinates to the point of intersec-

of
tion of the parabolas (Fig. 93). Let us find the ordinates of the
points of intersection of the parabolas by excluding x from the sy-
stem of equations
y = x\
{ y
2
= 8x.
We obtain y 1 = 0, #2 = 4. Hence, V=n

J [y
— fj) dy=* 24 -
jt.
7.6.10. Compute the volume of the solid torus. The torus is a

solid generated by revolving a circle of radius a about an axis
lying in its plane at a distance b from the centre (bâ). (A tire,
for example, has the form of the torus.)
7.6.11. Compute the volume of the solid obtained by revolving

about the x-axis the figure bounded by two branches of the curve
(y—x) 2 = x 3 and the straight line x=l.
7.6.12. Find the volume of the solid generated by revolving
about the line y = — 2a the figure bounded by the parabola y 2 = 4ax
and the straight line x = a (Fig. 94).
-x
a
-2a C\ B -x
0'
Fig. 93 Fig. 94
Solution. we transfer the origin of coordinates into the point

If
O'(0, — 2a)
retaining the direction of the axes, then in the new
system of coordinates the equation
of the parabola will be
(y' — 2a) 2
= 4ax. B
Hence y 2 = 2a + \Z~4ax (for the cur-

ve OAB), and y 1 = 2a ]^4ax (for —
the curve OCD). The sought-for vo-
lume is equal to X
V = n§(yl — yl)dx = n§[(2a +

o o
+2^) ~ (2a— 2 V^f] dx = f na\

2
Fig. 95
7.6.13. Find the volume of the solid generated by revolving about

the x-axis the figure enclosed by the astroid: x a cos 3 1\ y as\n 3 t. = =
Solution. The sought-for volume V is equal to double the volume
obtained by revolving the figure OAB (Fig. 95). Therefore,
'
= 2n ^ y 2
dx.
Change the variable

x = a cos 3 /, t
dx = —
3a cos 2 / sin t dt,
jx/2
y = a sin 3
1 9
a
Hence,
V = 2n^ a 2 sin 6 / ( — 3a cos 2

ts'mt) dt =
it
IT
= 6jxa 3
sin 7 /d/ - j sin 9 / dt
J
- o o
Using the formula from Problem 6.6.9 for computing the above
integrals, we get
t, c o / 6 4 2 8 6 4 2 \ 32 ,

one arc of the cycloid x a(t sin/), y a(l =
cos/) about the — = —
Ar-axis
7.6.15. Compute the volume of the solid obtained by revolving

about the polar axis the cardioid p = a (1 +cos cp) shown in Fig. 81.
Solution. The sought-for volume represents the difference between

the volumes generated by revolving the figures and OKLO MNKLO
about the x-axis (which is the polar axis at the same time).
As in the preceding problem, let us pass over to the parametric
representation of the curve with the polar angle cp as the parameter:
x — p cos = a cos cp (p ( 1 + cos cp),
y = p sin =a sin (p cp (1 + coscp).

It obvious that the abscissa of the point
is equals 2a (the value M
of cp=0), the abscissa of the point K being the minimum of
x at
the function x =
a(\ +cos cp)cosq).
Let us find this minimum:
dx
^ = — a sin cp (1 +2 coscp) =0,
<p, =0; (p 2 =— ji.
At ^=0 we obtain xM = 2a t
at (p 2 = |n, xK =
Hence, the sought-for volume is equal to

2a
V z=n ^ y\dx — tc y\ dx.

J
Changing the variable x = acosq)(l + coscp), we get
X 9 X *
1
= a + cos
2 2
sin 2 2
r/
dx =
( 1
— a sin (1 +2 cos cp
(p) q),
cp) dcp,
— a/4 2jx/3 > —a/4 2n/3
2a ji
Thus,
o
V= ji
^ a2 (1 +cosqp) 2 sin 2 qp [ — a sin cp (1 + 2 coscp)] dcp —
2
— jt a2 (l +coscp) 2 sin 2 (p [
— a sin (p (1 + 2 cos (p)] d(p
==
J
2
= jia 3
^ sin 3 q) (1 +coscp) 2 (l + 2cosq))d(p =
o
i
= jia 3
+w) 2 (l +2u)du = ^ na 3
(w = cosqp).
-
7.6.16. Compute the volume of the solid bounded by:
(a) the hyperboloid of one sheet + ^r — ^ = 1 an d the pla-
nes 2 =— 1 and z=l;

(b) the parabolic cylinder 2=4 — */
2
, the planes of coordinates
and the plane x a\ =
(c) the elliptic paraboloid z = ^ +^ and the plane z =k (k > 0).
7.6.17. A wedge is cut off from a right circular cylinder of radius

a by a plane passing through the diameter of the cylinder base and
inclined at an angle a to the base. Find the volume of the wedge.

the figure bounded by the following lines:
(a) xy = 4, x= \, x = 4, */ =
about the x-axis;
(b) y = 2x — x'\ y = about the #-axis;
y = x y = 0, x=2 about the (/-axis;

3
(c) ,
(d) y = s\nx (one wave), y = about the #-axis;

(e) x
2
—
# = 4,
2
y=±2
about the (/-axis;
(0 (y
—
a) 2 = ax x = 0, y = 2a about the x-axis.
y
7.6.19. Find the volume of the solid obtained by revolving the

ax ~7*
curve yu 2 1
= about the x-axis.
a

about the #-axis the figure bounded by the lines y^s'wiX and
y = —2 x.
about the x-axis the curvilinear trapezoid bounded by the catenary
y = ^-(ê a
-\-e
a
) = acosh-^ and the straight lines xt =—c y
x2 =
=c (c> 0).

about the #-axis the figure bounded by the cosine line (/ cosxand =
the parabola y^^* 2
-

about the #-axis the figure bounded by the circle x 2 y
2
and + =l
the parabola y
2
= ^x.
7.6.24. On the curve y x 3 take two points A and B, whose =
abscissas are 1 and b a=
2 respectively. = )
Find the volume of the solid generated by revolving the curvi-

linear trapezoid aABb about the #-axis.
7.6.25. An arc of the evolute of the ellipse x acost\ y = = bs\nt

situated in the first quadrant revolves about the x-axis
Find the volume of the solid thus generated.
the region enclosed by the loop of the curve x = at 2

y = a[t —
>
^j
about the #-axis.
7.6.27. Compute the volumes of the solids generated by revolv-

ing the region enclosed by the lemniscate (x 2 2 2
y ) a 2 (x 2 2
) + = —
about the x- and (/-axes.

the region enclosed by the curve p acos 2 cp about the polar axis. =
§ 7.7. Arc Length of a Plane Curve in Rectang. Coord's 345
§ 7.7. The Arc Length of a Plane Curve in Rectangular

Coordinates
If a plane curve is given by the equation y y(x) and the =
derivative y' (x) is continuous, then the length of an arc of this
curve is expressed by the integral
l=lVl+y' 2
dx
a
where a and b are the abscissas of the end-points of the given arc.
7.7.1. Compute the length of the arc of the

semicubical parabola y 2 =x 3
between the points
(0, 0) and (4, 8) (Fig. 96).
Solution. The function y(x) is defined for

xÔ. Since the given points lie in the first qu-
_3_
adrant, y =x 2
. Hence,
3
y
f
= ~Vx and l/l+*/' 2 = j/ 1+-
Consequently,
Fig. 96
3 4
2
=A (10 KTo-i).
7.7.2. Compute the length of the arc cut off from the curve
4
y
2
=x 3
by the straight line * = -g.
7.7.3. Compute the arc length of the curve y=\ncos>x between
the points with the abscissas x=0, x=^.
Solution. Since y' = — tan x, then |/l + y' = \f\ -f tan x = sec x.
2 2
Hence,
n
4
3n
' sec xdx= In tan +\ = lntan-g
. .
7.7.4. Compute the arc length of the curve y = In Jx

e
|
from
x1 =a to x2 =b (b > a).
7.7.5. Find the arc length of the curve x=-^y 2 — y \ny between
the points with the ordinates y=l and y = 2.

Solution. Here it is convenient to adopt y as the independent
variable; then
'
*'-U-w and ^ T+ 7 - 1
/(T^+i7=^+i-
Hence,
2 2
/ = J ^jqpî dy = I (i y + 4) dy =T + T ln 2 -
1 1
_2_ _2_ _2_
7.7.6. Find the length of the astroid x 3

+y =a
3 3
.
Solution.As is known, the astroid is symmetrical about the axes

of coordinates and the bisectors of the coordinate angles. Therefore,
it is sufficient to compute the arc length of the astroid between
the bisector y = x and the x-axis and multiply the result by 8.
3
~3 — ~\ T
In the first quadrant y =\a ( 3
x J and y = at x = a, y=x
at * =
Further,
2 '
y
'
=± ( a ^-jc T )
( -|) x~^~ = — x~~* [aJ-x^)
and
Consequently,
a
c
.L _J_
/ =8 ) a 3
x 3
Jx = 6a.
a
3/
/f
2
jVo/e. If we compute the arc length of an astroid situated in

the first quadrant, we get the integral
a
j i
3 3
^ a x dx y
whose integrand increases infinitely as x — ().

§ 7.7. Arc Length of a Plane Curve in Rectang. Coord's 347
7.7.7. Compute the length of the path consisting of por- OABCO

tions of the curves y 2 2x 3 and x 2 y
2
=
20 (Fig. 97). + =
Solution. It is sufficient to compute the arc lengths and
since by symmetry of the figure about the
x-axis
Solving the system of equations

x2 + y = 20,
2
y
2
= 2x\
we find the point A (2, 4).
Find l fa. Here
2
J/2*, V\+y'*= )/ i+ J*.
Hence,
2
Since on the circle of radius ^20 /;

*ab xs *' ie l en §th of an arc cor-
responding to the central angle arc ta
tan 2, i
l
XB = ]/20arc tan 2.
Finally we have
I = A(ioi/TO— l)+4/5 arc tan 2.
7.7.8. Compute the arc length of the curve:
(a) = y— 1 cut off by the x-axis;
(b) */ = ln(2cosx) between the adjacent points of intersection with

the x-axis.
(c) 3y =x(x
2
l)
2
—
between the adjacent points of intersection
with the j-axis (half the loop length).
7.7.9. Compute the arc length of the curve
y = Ux V~#-\ — In (x + KF=T)]
between
x = 1 and x = a+l.
7.7.10. Find the arc length of the path consisting of portions of

the curves x2 =
(y+\y and y 4. =
§ 7.8. The Arc Length of a Curve Represented

Parametrically
If is given by the equations in the parametric form x = x(t),

a curve
y = y(t)and the derivatives x' (t), y' (t) are continuous on the in-
terval [t lt t 2 ], then the arc length of the curve is expressed by the
integral
l = [Vx ,2
{t) + y' 2
{t)dt,
where t x and t 2 are the values of the parameter t corresponding to

the end-points of the arc (t 1 t 2 ). <
7.8.1. Compute the arc length of the involute of a circle x =
= a(cos/ + /sinO. y = a(sini — tcost) from / = to t=2n.
Solution. Differentiating with respect to /, we obtain
x\ = at cos t, y't = at sin t,
whence Vx' t
2
+ y' = at.
t
2
Hence,
2jt
2n
/ = at dt = =2an 2
.
J o
o
7.8.2. Find the length of one arc of the cycloid:
x = a(t — sin t), = a(l — cos/).

y
7.8.3. Compute the length of the astroid: x = acos 3

t y y = asm*t.
Solution. Differentiating with respect to t, we obtain
x't= —3a cos ts'mt; 2
y' = 3a sin cost. 2

1
t
Hence
J/"*;* + ^;« = \f9a 2

sin 2 / cos 2 / = 3a \
sin / cos / 1 = ^2 |
sin 2t |.
Since the function |sin2/| has a period
/ = 4xy jsin2/df =6a.

o
Note. If we forget that we have to take the arithmetic value of

the root and put Yx 2
-\- y' 2 = 3asmt cos t , we shall obtain the wrong
§ 7.8. Arc Length of a Curve Rep res' d Parametric ally 349
result, since
2ji
2ji
sin/cos/d/ = ySin 2
/ -0.
7.8.4. Compute the length of the loop of the curve x=V3t 2

,
y = t-t*.
Solution. Let us find the limits of integration. Both functions
x(t) and y(t) are defined for all values of Since the function
x= V 3/ ^0, 2
the curve lies in the right half-plane. Since with a
change in sign of the parameter x(t) remains unchanged, while
y(t) changes sign, the curve is symmetrical about the #-axis. Furth-
ermore, the function x(t) takes
on one and the same value not
more than twice. Hence, it follows
that the points of self-intersection
of the curve lie on the x-axis, i.e.,
at y= (Fig. 98).
The direction in which the mo-
ving point M
(x, y) runs along the Fig. 98
curve as / changes from oo to oo —
is indicated by the arrows.
But y = at /
t
= 0, t
2 , 3
= ±1. Since x (t 2 ) =x (t 3 ) = J^3, the point
(^3, 0) is point of self-intersection of the curve. Conse-
the only
quently, we must integrate within the limits t 2 -1 and / =1.
8
=—
Differentiating the parametric equations of the curve with respect
to f, we get x't = 2V~3t, y't=\—3t\ whence
Consequently,
i(l+3t*)dt = 4.
= 2 — -j
Q
t t*
7.8.5. Compute the arc length of the curve #=-g- f y
between the points of intersection with the axes of coordinates.
7.8.6. Compute the arc length of the ellipse ^ + ^=1.

Solution. Let us pass over to the parametric representation of the
ellipse
x=acost, y = bs'mt i
O^t ^ 2n.
Differentiating with respect to t y
we obtain
xt = — a sin t\ y't = b cos t.

350 Ch. VII. Applications o) the Definite Integral
whence
Vx + y' = Va
t
2
t
2 2
sin 2 / +b 2
cos 2 / = a V\—e 2
cos 2 /
where e is the eccentrici-ty of the ellipse,
6 = —a = -V a*a— b*
c
.
Thus
2ji T
l = a \V 1— 2
zos 2 tdt = Aa[\f\ — e 2
cos 2 / df.
The integral — 2
cos 2 tdt is not taken in elementary func-
o
tions; it is called the elliptic integral of the second kind. Putting
/ = Y — x, we reduce the integral to the standard form:

ji ji
T 2
\ V\ — 2
e cos 2
/d/ = $ Kl— 2
sin 2 xdT=£(e),
where E (e) is the notation for the so-called complete elliptic inte-
gral of the second kind.
Consequently, for the arc length ot an ellipse the formula
/ =4a£(e) holds good.
It is usual practice to put e sina and to use the tables of va- =
lues for the function
E l
(a) =E 1
(arc sin e) =E (e).
For instance, if a = 10 and 6 = 6, then
V l0^— 6 = A0.8Q = sin 53°.

2 2
8 =-
— r Q0 .
Using the table of values of elliptic integrals of the second kind,

we find / = 40£ l
(53°) = 40xl.2776 «51.1.
* = </=y(' 2 -3)
between the points of intersection with the x-axis.
7.8.8. Find the arc length of the cardioid:
x=a (2 cos t — cos 2t),

y = a(2s\n — sin 2t). t
§ 7.9. Arc Length of a Curve in Polar Coordinates 351
7.8.9. Find the length of the closed curve
x = 4 \^2as\nt; y = asm2t.
7.8.10. Find the arc length of the evolute of the ellipse
x = — cos 3
/, — ^-sin 3
/, c
2
=a 2 —b 2
.
x=(t 2 — 2)sin/ + 2/ cos/,

y = (2 — + 2/ sin/
2
/ )cos/
between /, = and /
2
= ji.
7.8.12. On the cycloid x = a (/ — sin /); y = a(\— cos/) find the

point which divides the length of the first arc of the cycloid in
the ratio 1 :3.
§ 7.9. The Arc Length of a Curve in Polar Coordinates
If a smooth curve is given by the equation p p((p) in polar =

coordinates, then the arc length of the curve is expressed by the
integral:
*pi
where cp
t
and cp 2 are the values of the polar angle rp at the end-
points of the arc (cp, < cp 2 ).
7.9.1. Find the length of the first turn of the spiral ot Archi-
medes p = aq).
Solution. The first turn of the spiral is formed as the polai angle
cp changes from to 2n. Therefore
2ji 2ji
/ =\ \fa y 2 2
+a 2
dy = a\ Vy + 2
1 dcp
o o
=a [n |/"4ji
2
+ 1 -f 1 In (2n + + 1 ) .
Find the length of the logarithmic

7.9.2. spiral p=ae m v between
a certain point (p (p ) and a moving point , (p, cp).
Solution. In this case (no matter which of the magnitudes, p or

p is greater!)
,
<p
<Po
= a|/l + m 2
^
m
e '*dq) =a
IP — Po
V +m
1
2
.
Ap|,
i.e. the length of the logarithmic spiral is proportional to the

increment of the polar radius of the arc.
7.9.3. Find the arc length of the cardioid p = a(l+coscp)

(a>0, 0<q)<2ji).
Solution. Here p^= — asincp,
VV^+V^- K2a (l+cos 2
(p) - K4a 2
cos 2 (cp/2) -
= 2a cos W^J _ 2a
I
> I
^ C os (cp/2), jx < < 2k.cp
Hence, by virtue of symmetry

2jt
/ = 2a cos y d(p = 4a ^ cos dcp = 8a.

J |
o o
Find the length of the lemniscate

7.9.4. p
2
= 2a 2 cos2cp between
the right-hand vertex corresponding to cp = and any point with a
polar angle (p <~ .
Solution. If 0<(p<~, then cos2q)>0. Therefore
p = a j/*2 cos 2cp; pjp

= a V2 sin 2(p #
j^cos 2cp
s in
2
2cp \ _ a /"2
Y cos2(p 7
t V ]^ C os2(p
Hence,
o o
V — 2sin
1
2
cp
The latter integral is called the elliptic integral of the first kind.
It can be reduced to a form convenient for computing with the
did of special tables.
§ 7.9. Arc Length of a Curve in Polar Coordinates 353
7.9.5. Find the arc length of the curve p = asin y. 3
7.9.6. Compute the length of the segment of the straight line
p = asec ^q)
— y^ between (p = and cp =y
Solution. p^ = asec^q) — y^ tan ^(p
—y j ;
K*P
2
+ p; = «sec((p-y) ]/ l+tan ((p-f )=asec ((p-f
2 2
(The sign of the modulus in the function sec(q) — -^-j is omitted,
since on the interval [o.f this function

4
p=asln -£
is positive.)
l =a j* sec 2 ^ q) —y dq)
4 V3 a.
^
7.9.7. Find the length of the closed
p = asin y.
4
curve
Solution. Since the function p

r = asin 4
4
is even, the given curve is symmetrical about
Fig. 99
the polar axis. Since the function sin 4 y has
a period 4jx, during half the period from to 2n the polar radius
increases from to a, and will describe half the curve by virtue
of its symmetry (Fig. 99).
Further, p^--=asin 3 (q)/4)cos(q)/4) and
Kp + P^ = Va? sin
2 2 8
(cp/4) +a 2
sin 6 (cp/4) cos 2 (cp/4) - a sin 3
(cp/4),
if 0<(p<2jT.
Hence,
'271 JT/2
/ = 2a ^
sin 3
(cp/4) d(p = 8a 3
j sin /d/ = ya ((p = 4f).
b o
7.9.8. Find the length of the curve (p = y(p+l/p) between p=2

and p = 4.
Solution. The differential of the arc dl is equal to
2
d/ = Kp + p; dq) = Kp
a f
2
d(p» + dp = a
|/ P
2
(-^) ^ 1*.
From the equation of the curve we find ^=y — ^2") • Hence,

4 4
/ =
I
2
V P
2
4( -^) + ldP=j
1
2
2
V|(p -2+^ + 2
4)dp =
= 3 + ^.
4
;-.(/Kf*=i(f+'"' 2 ^
7.9.9. Find the length of the hyperbolic spiral p(p=l between

3 4
=t ^=y
1
<Pi and •
7.9.10. Compute the length of the closed curve p = 2a(sinq) + cos(p).

7.9.11. Compute the arc length of the curve p = 1+ c 0S(p
* rom
<Pi = — 2 t0 ^^y-
§ 7.10. Area of Surface of Revolution

The area of the surface generated by revolving about the x-axis
the arc L of the curve y = y(x) (a^x^b) is expressed by the
integral
b
P = 2n\y VT+V dx. 2
It is more convenient to write this integral in the form P = 2n ^

ydl 9
L
where the differential of the arc length.
dl is
If is represented parametrically or in polar

a curve coordinates,
then it is sufficient to change the variable in the above formula,
expressing appropriately the differential of the arc length (see §§ 7.8
and 7.9).
7.10.1. Find the area of the surface formed by revolving the

_2_ 2 _2_
astroid x 3 y
3
+
a 3 about the #-axis. =
Solution. Differentiating the equation of the astroid we get
fx 3
+4*/ 3 y' = 0,
whence
y'=- y-
Then, Vl+y'* = Since the astroid is sym-

3 3
X" \x\
metrical about the (/-axis, in computing the area of the surface we
may first assume x^0
and then double the result. In other words,
y
the desired area P is equal to

u u 2
P = 2x2n §yV\ + y' 2 dx = 4n

J
a3 —X 3
X
a3 —x =t\
3
1
a3
—4" x 3
dx = 2tdt,
a
fl
l/3
y_
12
Then P=\2na 3
t*dt=^n<&
J
o
7.10.2. Find the area of the surface generated by revolving about

the x-axis a closed contour formed OABCO
by the curves # A;
2 =
and x y 2 (Fig. 100). =
Solution. It is easy to check that the 1
given parabolas intersect at the points
(0, 0) and B (1, 1). The sought-for area
= +
P P 1 P 2 where the area P x is formed
,
by revolving the arc OCB and P 2 by revol- y
ving the arc OAB.

Compute the area P v From the equation
x = y2 we get y = V* and #' = ^y=f.
Hence, Fig. 100
i i
+ ±dx = 2n [ ^ *+ ^=
4
P = 2n§V~x }/~
!
1
1
4(4* + !)"
|(5K5-1).
Now compute the area P 2 We

. have y~x 2
,
y' = 2x and
P 2 = 2n I
x2 Vl + Ax 2
dx.
The substitution # = ysinh£, dx = ~ cosh t dt gives

Arsinh 2
Arsinh
= ^^sinh4^ —
2
P^JI sinh 2 /cosh 2 /d/ t
J o
= 9-TL -3> ln ( 2 + ^).

Thus, _
p=p 1+ p = 1^5-0^
2
(5
+ _qg£L _^ Jtln(2+|/-5 ^ )
7.10.3. Compute the area of the surface generated by revolving:

2
X
(a) the portion of the curve y = -^, cut off by the straight line
y = —3 , about the (/-axis;
(b) the portion of the curve y

2
= 4 + x, cut off by the straight
line x=2 about the x-axis.
y
7.10.4. Find the surface area of the ellipsoid formed by revolving

the ellipse -^--j--p-=l about the #-axis (a>b).
Solution. Solving the equation of the ellipse with respect to y
for y^0, we get
b , r , b x
Va* — x 2
'
Hence
-a
~b
where the quantity e= j/ a =~ is the eccentricity of the
ellipse.
When b —+a the eccentricity e tends to zero and
arc sin e
t
lim
.
8
= ,
I,
e - o
since the ellipse turns into a circle, in the limit we get the surface
area of the sphere:
P = 4jia a
.
§ 7 JO. Area of Surface of Revolution 357
7.10.5. Compute the area of the surface obtained by revolving

the ellipse 4x2Jry 2 =
4 about the (/-axis.
7.10.6. An arc of the catenary
y = ~{e a +e" a ) = a cosh

whose end-points have abscissas and x, respectively, revolves about
the x-axis.
Show that the surface area P and the volume V of the solid thus
generated are related by the formula P= —
2V
m
Solution. Since y' = sinh — , we have V\ + y' = cosh — 2

. Therefore
hence, P = —a
.
7.10.7. Find the area of the surface obtained by revolving a loop

of the curve 9ax 2 = y(3a — y)
about the (/-axis. 2
Solution. The loop

described by a moving point as y changes
is
from to 3a. Differentiate with respect to y both sides of the

equation of the curve:
1 8axx' = (3a — y) 2
— 2y(3a—y) = 3(3a — y)(a — y) y
whence = ^ ~y a~y
a ^
Using the formula computing the
xx'
^ . for
area of the surface of a solid of revolution about the (/-axis, we have
P = 2n j xV~l+x' 2 dy=2n j Vx + (xx')

2 2
dy =
3a 3a
b o
— y )dy = 3na
2 2
.
7.10.8. Compute the area of the surface generated by revolving

the curve 8y 2 =x 2 — x* about the x-axis.
7.10.9. Compute the area of a surface generated by revolving

about the x-axis an arc of the curve x = t
2
\ y=-^(t 2 — 3) between
the points of intersection of the curve and the x-axis.
Solution. Putting y = 0,
and, hence,
find ^ = and /
2t3
= ±K3,
^= and jc
2>3 = 3. Whence
the curve intersects it follows that
with the x-axis at two points: (0, 0) and (3, 0). When the para-
meter t changes sign, the sign of the function (x)t remains unchan-
ged, and the function y (t) changes
y, its sign, which means that the curve
1 m is symmetrical about the #-axis
(Fig. 101).
X To find the area of the surface it
is sufficient to confine
ourselves to
the lower portion of the curve OnB
-1 n that corresponds to the variation
of the parameter between and
Fig. 101
+ 1^3. Differentiating with respect
to we find
ty
-.2t; y = t*-l
t
and the linear element

dl = Vx? + y? dt (l + t
2
)dt.
Hence,
P = 2n\\y(t)\Vx' 2 + y' t t
2
dt =
VT vT
= 2n$ 3
(t*—3)(l
\
L "
°j + )dt = — ^-n
i" I 1 t
2
j (t* — 2t* — 3t)dt=3n.
7.10.10. Compute the surface area of the torus generated by re-
volving the circle x 2 +
(y— b) 2 r 2 (0 r b) about the *-axis. = < <
Solution. Let us represent the equation of the circle in parametric
form: x =
rzost\ y = b rs'mt. +
Hence
x't = — r sin t\ y't=r cos t.
The desired area is

2JT
P-2jiJ (b + r sin t) )/*(— r sin tf + (r cos tf dt =

= 2nr\ (b + r sin t) dt = 4n 2
br.
7.10.11. Compute t he are a of the surface formed by revolving

the lemniscate p =
a}/cos2q) about the polar axis.
Solution. Real values for p are obtained for cos2(p>0, i. e. for
— X^^^T^^ ri ght-hand branch of the lemniscate), or for

3 5
-j-ft^ ty^-j n (the left-hand branch of the lemniscate).
The linear element of the lemniscate is equal to
d/ = Kp» + p^dcp= if a cos2y+( a Y dip =-

,"|Acos2(p/
cos '
2
|^cos2(p \
^l
V 2(p
Besides, # =
psincp=asin(p}/cos2q).
The sought-for surface area P is equal to double the area of the
surface generated by revolving the right-hand branch. Therefore
P = 2 x 2n f y dl
= 4«o« f
^^^P^ = 2«a« (2 - j/2 ).
J J K cos 2cp
7.10.12. Compute the area of the surface formed by revolving

y = a the quarter of the circle x
+ y =a +
2 2 2
about the straight line x
between A (a, 0) and B (0, a).
Solution. Find the distance from the moving point MN
(x, y), M
lying on the circle x 2 y
2
a 2
to the + =
straight line x y a: , + =
\x+ ya 2 —x 2 — a\ :+ Va? — x* — c
MN =
V2 V2
since for the points of the circle that lie in the first quadrant
x-\-yâ. Further,
dl = yi + y'2 dx= if + ( i
x
Ydx=-^=
Hence,
'
x+ Va* — x 2 -
a dx
P= 2n
V~2 V^ — x l
V2na [- V'c x2 + x — aarcsin- = (*-*)•
7.10.13. Compute the

area of
the surface formed by
revolving
one branch of the = lemniscate p
r cos2q> about the straight line
=a\ _
it
Solution. From the triangle OMN Fig. 102

(Fig. 102) we find the distance MN
of an arbitrary point M of the right-hand branch from the
axis of revolution q) =x :
MN = p sin — (p^
=a|/cos2cpsin — qp^ ;
then
a dw
j1
dl = r - .
V cos 2cp
Jl/4
Therefore P = 2n j aKcos2q5 sin q>) y=?= = 2na 2

.
-Jt/4

3
about the x-ax\s the arc of the curve # = xy between x =—2 and
x = 2.
one half-wave of the curve y =
s\nx about the x-axis.
7.10.16. Compute
the area of the surface generated by revolving
about the the arc of the parabola x 2
(/-axis 4ay between the points =
of intersection of the curve and the straight line y 3a. =
7.10.17. Find the area cf the surface formed by revolving about
the ;t-axis the arc of the curve x e'sin/; y^^cost between =
*=0 and *=y
7.10.18. Compute the area of the surface obtained by revolving
about the #-axis the arc of the curve x = -^-\ t

3
y =4 —^
t
1
between
the points of its intersection with the axes of coordinates.

the curve p =
2asinq) about the polar axis.
about the x-axis the cardioid
x = a (2 cos — cos 2t),
t
y = a(2s'mt — sin 2t).
§ 7.11. Geometrical Applications of the Definite Integral
7.11.1. Given: the cycloid (Fig. 103)

x = a{t — sin t)\ yâ(l — cos/); 0< ^ 2n.
/
Compute:
(a) the areas of the surfaces formed by revolving the arc OBA
about the x- and (/-axes;
(b) the volumes of the solids generated by revolving the figure OB AO
about the */-axis and the axis BC\
(c) the area of the surface generated by revolving the arc BA
about the axis BC\
§ 7.11. Geometrical Applic's of the Definite Integral 361
(d) the volume of the solid generated by revolving the figure

ODBEABO about the tangent line DE
touching the figure at the
vertex B\
(e) the area of the surface formed by jj
revolving the arc of the cycloid [see
item (d)J.
Solution, (a) When revolving about —
c
the jc-axis the arc OBA generates a sur-
face of area Fig. 103
231
P x = 2n y dl = 2n j a ( 1
— cos t) 2a sin y dt =
J
64jta 3
8a sin 3
'ji
J y dt
When revolving about the (/-axis the arc OB A generates a sur-
face of area
TT
Py = 2n xdl = 4na 2
J
(t — sin t) sin y dt +
+ 4to 2 — sin0sinyd/ = 4jia 2

| (/ — sin siny dt = 16jt 2 2
.
Jl
(b) When revolving about the y-axis the figure OBAO generates
a solid of volume
oo
2a 2a 2a
Vy = n^ (x\ — x\)dy = n J
jc| rf^ — Jt
J
x2 dj/,
o
where x = x 1(y) is the equation of the curve BA, and x x2 (y) is =

the equation of the curve OB.
Making the substitution y a(l cos/), take into consideration = —
that for the first integral t varies between 2k and ji, and for the
second integral between and n. Consequently,
TT Jt
Vy = n J
a2 (t — sin t)
2
asm tdt — n J
a2 (t — sin t)
2
asm tdt =
2jt
= na 3
J
(/ — sin t)
2
sin tdt =
2n
- -
= jia 3
J
/
2
sin/d/ — J
t{\— cos2t)dt+ 3
$ sin /d/ =6ji 3 a 3 .
_2t 2jt 2ji
For computing the volume of the solid obtained by revolving

the figure OBAO about the axis BC it is convenient first to trans-
fer the origin into the point C, which yields the following equa-
tions in the new system of coordinates
x' = a(t — n — sin t)\ y' = a(l — cos/).

Taking into account only the arc BA, we get
2a jt,
V=n ^ x'
2
dy' = na 3
^ (t
— — sin ji t)
2
sin t dt.
2ji
Putting t —n = z> we obtain

JT
V = — na 3
\(z + sin z) 2
sin zdz — na 3
\ (z + sin z) 2
sin z dz =
n
=^ (9jx*— 16).
(c) Making the above-indicated shift of the origin, we get
dl = 2a sin y |
dt \
= — 2a sin y dt.
Therefore
2a
P= ^2nxdl = — 4na 2 j*
(/ — — sin/) sin y dt =
jt
2n
ji
= 4na (z+
y dz = 4 (^2n — ^
2
sin z) cos
J*
b
(d) Transferring the origin into the point B and changing the
direction of the r/-axis, we get
x' = a(t — ji — sin /), y
f
=a (1 +cos /).
Putting / —n = z, we have
x' = a (z + sin 2), y' =a ( 1 — cos 2),
z changing from — ji to ji for the arc OB A. Hence

ji
V=n ^ a3 {l— cos z) 2 (\ + cos z) dz = ri a l 3

.
-ji
ji ji
(e) P = 2n ydl = 4na 2

j (1— cosz)cosyd2 =y jta
2
.
J
-jt -ji
the volume of the solid bounded by the surfaces

7.11.2. Find
Z2 = 8(2— x) and x2 + y 2 = 2x.
Solution. The first surface is a parabolic cylinder with generat-
rices parallel to the r/-axis and the directrix z 2 = 8(2 a:) in the —
plane xOz, and the second is a circular cylinder with generatrices
§ 7.11. Geometrical Applies of the Definite Integral 363
parallel to the z-axis and the directrix x2 + y = 2x 2
2
in the plane xOy.
The volume V is computed by the formula V =^ S(x)dx. S(x) re-

o
presents the area of a triangle whose base is equal to 2y and alti-
tude to 2z\
S(x) = 2yx2z = 4 V2x—x 2
1/8(2—*).
Hence,
2 2
V= 5 4 Vx (2 x) 8 (
2 x) dx = 4 \/"8
$ (2
— *) Vx dx =
256
= 4/8 (|2 J/* 3
--?-!/**) :
15
'
7.11.3. Prove that if the figure S is bounded by a simple con-

vex contour and is situated between the ordinates y x and y 2 (Fig. 104),
then the volume of the solid ge- .y
nerated by revolving this figure
about the #-axis can be expressed
by the formula
y*
t b/
&
V = 2n yh dy y 1 \
^ y
yt
where
h=x 2
{y)—x 1 (y) 9
x =x 1 (y)
being the equation of the Si
left portion of the contour and

x = x2(y) that of the right portion.
Solution. Let the generating fig- Fig. 104
ure S be bounded by a simple
convex contour and contained between the ordinates y t and y 2 .
Subdivide the interval [y ly y 2 ] into parts and pass through the

points of division straight lines parallel to the axis of revolution,
thus cutting the figure 5 into horizontal strips. Single out one
strip and replace it by the rectangle ABCD, whose lower base is
equal to the chord AD =
h specified by the ordinate y its altitude y
AB being equal to Ay. The solid generated by revolving the rectangle

ABCD about the #-axis is a hollow cylinder whose volume may
be approximately taken for the element of volume
AV « n (y + Ay) h — ny h = 2nyAyh + nh (Ay)

2 2 2
.
Rejecting the infinitesimal of the second order with respect to Ay,

we get the principal part or the differential of volume
dV = 2n yh dy.
Knowing the differential of the volume, we get the volume proper
through integration:
V = 2n J
yhdy.
Thus, we obtain one more formula for computing the volume of

the solid of revolution.
7.11.4. The planar region bounded by the parabola y 2x 2 -\-3, =
the *-axis and the verticals and x=0
revolves about the x=\
(/-axis. Compute the volume of the solid of revolution thus generated.
Solution. Divide the area of the figure into elementary strips by
straight lines parallel to the (/-axis. The volume AV of the elemen-
tary cylinder generated by revolving one strip is
AV = n (x+ Ax) 2
y — nx y = 2n xy Ax + ny (Ax)
2 2
y
where Ax is the width of the strip.

Neglecting the infinitesimal of the second order with respect to Ax,
we get the differential of the desired volume
dV = 2n xydx.
Hence
i i
V= I
2n xydx = 2n \ x (2x2 + 3) dx = 4n.
o
7.11.5. Compute the area of the portion of the cylinder surface

2
x2 -f- y =t ax situated inside the
sphere
x2 + y + z =a 2 2 2
.
Solution. The generatrices of the

cylinder are parallel to the 2-axis,
x 2
=^
the
serving
circle
as
^
directrix
—jY+y (Fig. 105
^ shows a quarter of the sought-
a for surface).
Subdivide the portion of the
circle shown in Fig. 105 into
small arcs AL The generatrices
y passing through the points of di-
Fig. 105 vision cut the cylinder surface
into strips. If infinitesimals of
higher order are neglected, the area of the strip ABCD is equal to
CD- AL
If p and cp are the polar coordinates of the point D, then
p = acoscp and CD = ]/ a2 — p = asinq),
2
and A/ = a.A(p, whence we
§ 7.11. Geometrical Applies of the Definite Integral 365
find the element of area:
dP =a 2
sin cpdcp.
Hence,
P=4 J
a 2 sin cp dcp == 4a 2 .
7.11.6. Find the area of the surface cut off from a right circular
cylinder by a plane passing through the diameter of the base and
inclined at an angle of 45° to
the base.
Solution. Let the cylinder axis
be the e-axis, and the given dia-
meter the x-axis. Then the equa- C/
tion of the cylindrical surface q
will be x
2
+
y =a\ and that of
2
the plane forming an angle of 45°

with the coordinate plane xOy n y^^^?^
will be y = z.
The area of the infinitely nar- x
row strip ABCD (see Fig. 106) will Fig. 106
be dP =
zdl (accurate to infinite-
simals of a higher order), where dl is the length of the elemen-
tary arc of the circumference of the base.
Introducing polar coordinates, we get
z = y = a sin (p; dl = a dqp.
Hence dP =a 2
s\nq)dq) and
jt
P = a2 J
sin qp dy =a 2
[
— cos cp]* = 2a 2
.
7.11.7. The axes of two circular cylinders with equal bases inter-
sect at right angles. Compute the surface area of the solid constitu-
ting the part common to both cylinders.

about the y-axis the figure bounded by the parabola x 2 1, the = y—
axis of abscissas and the straight lines and x=l. x~0
7.11.9. Find the area 5 of the ellipse given by the equation
Ax 2 + 2Bxy+Cy 2 = l(8 = AC— 2
> 0; C> 0).
Solution. Solving the equation with respect to y, we get
_ — Bx — VC—dx _ — Bx+ VC — bx*

1
.
where the values of x must satisfy the inequality

C — 6jk 2 >0.
Solving this inequality, we obtain the limits of integration:
Consequently, the sought-for area is equal to
/? Vt
C_
6
7.11.10. Find the areas of the figures bounded by the curves rep-
resented parametrically:
(a) x=2t — t 2
; y = 2t 2
— t
3
;
7.11.11. Find the areas of the figures bounded by the curves given
in polar coordinates:
(a) p = asin3(p (a three-leaved rose);
(b) p =
1 — cos cp
(c) p = 3sinq) and p = K3cosq).

IP riirvp
7.11.12. Find the arc length of the curve /y 2
y
= —4^ (2 — x) 3
cut off
by the straight line x = — 1.
7.11.13. Find the length of the arc OA of the curve

i
^
4
where (0, 0); Al±,
2
,
.In-i
a In
7.11.14. Compute the arc length of the curve y

2
= —2 (x — l) 3
con-
tained inside the parabola y

2
= -^.
7.11.15. Prove that the length of the ellipse
x = V^2sin/; y = cost
is equal to the wavelength of the sinusoid y= s\nx.
7.11.16. Prove that the arc of the parabola y = ^ x2 correspon-
ding to the interval O^xâ
has the same length as the arc of
the spiral p =
Pcp corresponding to the interval O^pâ.
7.11.17. Find the ratio of the area enclosed by the loop of the
curve y=± -o- — * Wx to the area of a circle the circumference
of which is equal to the length of the contour of this curve.

§ 7.12. Computing Physical Quantities 367
7.11.18. Find the volume of the segment cut off from the ellipti-
cal paraboloid
^+^ =x by the plane x = a.
7.11.19. Compute the volume of the solid bounded by the hyper-

boloid ^t + ^t — 4-= — 1 and the planes
r z =c and 2 = />c.
a o c
7.11.20. Find the volume of the right elliptical cone whose base
is an ellipse with semi-axes a and fr, its altitude being equal to h.

about the x-axis the figure bounded by the straight lines y x+l; =
y =
2x+l and x 2. =
about the #-axis the figure bounded by the hyperbola — = 1,
the straight line 2ay — bx = and the axis of abscissas.
7.11.23. Find the volume of the solid generated by revolving the

curve p = acos 2
q) about the polar axis.
7.11.24. Find the areas of the surfaces generated by revolving the

following curves:
(a) y = tan x ^0<jc<^-^ about the #-axis;
(b) y = x \f -^-(O^xO) about the #-axis;
(c) x + y — 2rx =
2 2
about the x-axis between and h.
§ 7.12. Computing Pressure, Work and Other Physical

Quantities by the Definite Integrals
I. To compute the force of liquid pressure we use Pascal's

law, which states that the force of pressure of a liquid P on
an area 5 at a depth of immersion h is P = yhS y
where y is the
specific weight of the liquid.
II. If a variable force X = f(x) acts in the direction of the *-axis,
then the work of this force over an interval [x lt x 2 ] is expressed
by the integral
x2
A= ^f(x) dx.
III. The kinetic energy of a material point of mass m and velo-

city v is defined as
IV. Electric charges repulse each other with a force F = ,
where e x and e 2 are the values of the charges, and r is the distance
between them.
Note. When solving practical problems we assume that all the
data are expressed in one and the same system of units and omit
the dimensions of the corresponding quantities.
7.12.1. Compute
the force of pressure experienced by a vertical
triangle with base b and altitude h submerged base downwards in
water so that its vertex touches the surface of the water.
Solution. Introduce a system of coordinates as indicated in Fig. 107
and consider a horizontal strip of thickness dx located at an arbi-
trary depth x.
Assuming this strip to be a rectangle, find the differential of area
dS = MNdx. From the similarity of the triangles and ABC BMN
we have -^
b
= 4-,
h
whence MN = ^-h and dS = ^dx>
h
Fig. 107 Fig. 108
The force of pressure experienced by this strip is equal to dP —xdS

accurate to infinitesimals of higher order (taking into consideration
that the specific weight of water is unity). Consequently, the entire
force of water pressure experienced by the triangle is equal to
h h
P = xdS = -^ |
x*dx = -jbh*.
J
n
7.12.2. Find the force of pressure experienced by a semicircle of

radius R submerged vertically in a liquid so that its diameter is
flush with the liquid surface (the specific weight of the liquid is y).
7.12.3. A dam
has the form of a trapezoid whose upper
vertical
base is 70 m lower one 50 m, and the altitude 20 m.
long, the
Find the force of water pressure experienced by the dam (Fig. 108).
Solution. The differential (dS) of area of the hatched figure is
approximately equal to dS =
MNdx. Taking into consideration the
similarity of the triangles OML and OAE, we find ~= 2°

2Q
x
;
whence ML = 20 — MN = 20 — x + 50 = 70 — x. Thus, dS = MNx

xdx=(70 — x)dx
y
and the differential of the force of water pressure is
equal to
dP = xdS = x (70 — x) dx.
Integrating with respect to x from to 20, we get
P= \
(70x — x )dx=
2
11 333-
7.12.4. Calculate the work performed in pumping the water out

of a semispherical boiler of radius R.
A rectangular vessel is filled with equal volumes of water
7.12.5.
and water is twice as heavy as oil. Show that the force of pres-
oil;
sure of the mixture on the wall will
I
reduce by one fifth if the water is
replaced by oil.
n
Solution. Let h be the depth of
the vessel and / the length of the h_
wall. Let us introduce a system of 2

coordinates as shown in Fig. 109. x
Since the oil is situated above the
h
water and occupies the upper half of
the vessel, the force of the oil pres-
sure experienced by the upper half a
of the wall is equal to Fig. 1(
jl
2
The pressure at a depth x >y is made up of the pressure of the
oil column of height A

—2 and that of the water column of height
h
, and therefore
dP.=
h
2
Xy+U-I
f
h
ldx=[ x ~)idx.
Consequently, the force of pressure of the mixture on the lower half
of the wall is
The entire pressure of the mixture on the wall is equal to
r-^
P—P 4-
+ fi - — 4-
P — ———
+i6 - m 4 i6
lh 2 '
If the vessel were filled only with oil, the force of pressure P on
the same wall would be
h
IK1
=4.f xl dx = -r .
Hence,
P — P = l:lh*
16
= i-P.
5
7.12.6. The electric charge E concentrated at the origin of coor-

dinates repulses the charge e from the point (a, 0) to the point (ft, 0).
Find the work A of the repulsive force F.
Solution. The differential of the work of the force over displace-
eE
ment dx is dA =F dx = -^-dx.
Hence
eE
As b—* oo the work A tends to —a
7.12.7. Calculate the work performed in launching a rocket of
weight P from the ground vertically upwards to a height h.
Solution. Let us denote the force of attraction of the rocket by
the Earth by F, the mass of the rocket by Ri and the mass of m
the Earth by mE .
According to Newton's law
F =k X'
where x is the distance between the rocket and the centre of the
Earth. Putting km Rm E =K>

= we S et F x ( )
= -^> +
R being the radius of the Earth. At x=R the force F (R) will be
the weight of the rocket P, i.e. F(R) = P = -^2y whence K = PR*

and F (x) = .
Thus, the differential of the work is
dA = F(x) dx = ^-dx.
Integrating, we obtain
R+h R+fi
PRh
The limit lim A (h) = lim PR is equal to the work performed

h —* h —> <n ^ l~
by the rocket engine to achieve complete escape of the rocket from

the Earth's gravity field (the Earth's motion is neglected).
7.12.8. Calculate the work that has to be done to stop an iron
sphere of radius R rotating about its diameter with an angular ve-
locity (0.
The amount of required work is equal to the kinetic

Solution.
energy the sphere. To calculate this energy divide the sphere
of
into concentric hollow cylinders of thickness dx; the velocity of the
points of such a cylinder of radius x is cox.
The element volume of such a cylinder is dV = 4nx]^R 2
of x2 dx — y
the element of mass dM = ydV, where y is the density of iron, and

the differential of kinetic energy dK = 2nya) 2 x 3 2
x2 dx. VR —
Hence,
R
2
VR — x
2
AnyR* Mco 2 /? 2
K = 2nya dx =
co /?
2 3 2 2
jx 5
o
7.12.9. Calculate the kinetic energy of a disk of mass and ra- M

dius R rotating with an angular velocity co about an axis passing,
through its centre perpendicular to its plane.
7.12.10. Find the amount of heat released by an alternating si-
nusoidal current
2ji
I = /„ sin
,
during a cycle T in a conductor with resistance R.

Solution. For direct current the amount of heat released during
a unit time is determined by the Joule-Lenz law
Q = 0.24 I 2
R.
For alternating current the differential of amount of heat is
dQ = 0.24 a (t)Rdt, whence
Q=0.24/? [l*dt.
In this case
Q -0.24 RIl j sin 2 {^t — ^dt^

2jc
v
T_
-0.12 RI 2 '
0.12/? 77*.
2ji
7.12.11. Find the pressure of a liquid of specific weight d on a

vertical ellipse with axes 2a and 2b whose centre is submerged in
the liquid to a level h(h^b).
7.12.12. Find the pressure of a liquid of specific weight d on the
wall of a circular cylinder of base radius r and altitude h if the
cylinder is full of liquid.
7.12.13. Calculate the work performed to overcome the force of
gravity in pumping the water out of a conical vessel with the vertex
downwards; the radius of the cone base is R and its altitude is H.
7.12.14. Compute the work required to stretch a spring by 6 cm,

if a force of one kilogram is required to stretch it by 1 cm.
§ 7 J 3. Computing Static Moments and Moments of

Inertia. Determining Coordinates of the
Centre of Gravity
In all problems of this paragraph we will assume that the mass
isdistributed uniformly in a body (linear, two- and three-dimensional)
and that its density is equal to unity.
1. For a plane curve L the static moments x and y
about M M
the x- and (/-axis are expressed by the formulas
M x = J y dl, M y = J x dl.
L L
The moment of inertia about the origin of coordinates
/o=S(^ + »
f
)d/.
L
If the curve L is given by the explicit equation y = y(x) (a^x^b) y
then dl has to be replaced by Vl+y' 2 dx in the above formulas.

If the curve L is given by the parametric equations x x(t), =
y = y(t)(ti<: U)> then dl should be replaced by Vx' 2 + y' 2 dt in
these formulas.
2. For the plane figure bounded by the curves y y1 = {x), y =y 2
(x),
y 1 (x)^y 2 (x) and the straight lines x a, x b = = (a^x^b) the

static moments are expressed by the formulas
b b
M *=\ \ (yl — yl)dx; M y = §x(y — y )dx. t 1
a a
3. The centre of gravity of a plane curve has the following coor-

Mv M
dinates: xc =~ t yc = —j-x i
where / is the length of the curve L.
§ 7.13. Computing Static Moments and Moments of Inertia 373
The centre of gravity of a plane figure has the coordinates: x c = -^-,

yc where S is the area of the figure.
=^Y>
7.13.1. Find the static moment of the upper portion of the ellipse
a* ^ b*
about the #-axis.
Solution. For the ellipse
ydi dx = Vy + (yy'ydx;
2
since y 2 =b 2
— ^-x 2
and yy
f
= — ~i x * we have
r
ydl = ]/ b2 —^x + ^x 2 2
dx =^Va — e x 2 2 2
dx y
where e is the eccentricity of the ellipse, e = .
Integrating from —a to a, we find

a a
Mx = — [ Vet 1 —£ * 2 2
dx = —[ V — e 2 2
x 2 dx =
-a
Z==:
~a{K
a ^r *
— ~T
ar ° s n 8 ^
)
=^ ^+ arc sln p )«
In the case of a circle, i. e. at a = b, we shall have M x = 2a 2

,
~ 11. arc sin e

since 8 = and urn
8
= .
e -
7.13.2. Find the moment of inertia of a rectangle with base b and

altitude h about its base.
Solution. Let us consider width dy cut an elementary strip of
out from the rectangle and base and situated at parallel to the
a distance y from it. The mass of the strip is equal to its area
=
dS bdy, the distances from all its points to the base being equal
to y accurate to dy. Therefore, dl x by 2 dy and =
h
Ix = i
\ by 2 dy = -^-.
j
7.13.3. Find the moment of inertia of an arc of the circle x* +y =R

2 2
lying in the first quadrant about the #-axis.

7.13.4. Calculate the moment of inertia about the #-axis of the
figure bounded by the parabola y 2 =--Aax and the straight line x- a.
Solution. We
have dI x 2
=
x' dS, where dS is the area of a vertical
strip situated at a distance x from the y-ax\s (Fig. 110):
dS = 2\y\dx = 2\/4axdx.
374 Ch. VII . Applications of the Definite Integral
Hence,
a a 5
Ix =J 4x2 Vaxdx = 4V~a ^x T dx = -f a *-

o b
7.13.5. In designing wooden girder bridges we often have to deal

with logs flattened on two opposite sides. Figure 111 shows the
cross-section of such a log. Determine the
moment of inertia of this cross-section
about the horizontal centre line.
- — -z-^x
*~x
Fig. 110 Fig. Ill
Solution. Arrange the system of coordinates as is shown in the

accompanying drawing. Then
dI x =y 2
dS, where dS = MN dy = 2xdy = 2 VR — y 2 2
dy.
Whence
h h
lx = l\
-h
y
2
VW^ 2
dy = \\y 2
VW^ 2
dy.
Substituting y = Rs\nt; dy =R cos / dt; ^=0; t

2
= arc sin (h/R),
we <;et
h arc sin (h/R)
lx =4
Jy
2
VR — y 2 2
dy =4 [ R 2 sin 2
/ . R cos tR cos t dt =
arc sin (h/R) arc sin (h/R)
= 4R* sin 2 / cos 2 /d/ =^ f

(1— cos4/)df =
J
= ^ arc sin A + A (2h — R VR — h 2 2

)
2 2
.
When = we obtain the moment of inertia of the circle
about one of its diameters: L 4

'
7.13.6. Find the moment of inertia about the *-axis of the figure
bounded by two parabolas with dimensions indicated in Fig. 112.
Solution. Arrange the system of coordinates as shown in Fig. 112
and write the equations of the para-
bolas.
The equation *
of the left parabola is:
2
b2 ( ,
a
y ^Ya[ X +2 the equation of the
right parabola, # 2 = ^(y —

For the hatched strip the moment
of inertia is
dI x = y*dS = y*\MN\dy,
where
\MN = x \
2
—x 1
.
2 f
— —y = f-
2
Hence,
6/2 b/ 2
ab*
y*(a-%tf)dy = 2
.
/,= j> u
j i,2
-b/2
7.13.7. Find the static moments about the x- and (/-axis of the
arc of the parabola y 2 2x between x and x 2 (y>0).= = =
7.13.8. Find the static moments about the axes of coordinates
of the line segment -^--|--|-=1 whose end-points lie on the coordi-

nate axes.
7.13.9. Find the static moment about the #-axis of the arc of
the curve y = cosx between x l
=— and x,=-2 •
7.13.10. Find the static moment about the x-axis of the figure
r
bounded by the lines y =x 2
\ y = \/ x.
7.13.11. Find the moments of inertia about the x- and (/-axis of
the triangle bounded by the lines x = 0, y = andy + -|-= 1 (a> 0,
b>0).
7.13.12. Find the moment of inertia of the trapezoid ABCD about
its base AD
if AD = a, BC=b and the altitude of the trapezoid
is equal to h.
7.13.13. Find the centre of gravity of the semicircle x 2 -\-y 2 --=a 2
situated above the x-ax\s.
Solution. Since the arc of the semicircle is symmetrical about the
(/-axis, the centre of gravity of the arc lies on the */-axis, i. e. x c 0. =
376 Ch. VI I. Applications of the Definite Integral
To find the ordinate y c , take advantage of the result of Problem
7.13.1: M x —2a 2
; therefore y c = —= Thus, xc = 0, yc = -.
7.13.14. Find the coordinates of the centre of gravity of the ca-
tenary y =y + = cosh x between /I (0, 1) and B(a, cosh a).

Solution. We have
dl=\/ 1 + */'
2
dx= V\ + sinh 2
xrfx = cosh xdx
whence we find
/ = ^ d/
= ^ cosh xdx = sinh a.
l o
Then
a a
M y =^xdl=^ x cosh x dx = x sinh x\ — J sinh

a
Q
=
L
= a sinh a — cosh a + 1.
Hence,
a sinh a — (cosh a — = a- 1) cosh a-
- = a — tanh-£-
sinh a sinh a
Analogously,
a a
M x = ^ydl= ^
cosh 2
x dx =
y J
( 1 + cosh 2*) djc =
I / . sinh 2x sinh 2a
=t[ x +-t-
a sinh 2a
TH 4 cosh a
sinh a 2 sinh a
7.13.15. Find the centre of gravity of the first arc of the cycloid:
x =a(t —
sin /), y a(1 =
cos /) (0 t —
2n). < <
Solution. The first arc of the cycloid is symmetrical about the
straight line x =
na, therefore the centre of gravity of the arc of
the cycloid lies on this straight line and x c na. Since the length =
of the first arc of the cycloid l Sa we have = 9
2ji 2ji
ye =-j §ydl=--±2a J 2
(1— cos t)sm±dt=-j§ sin 3 ~2 dt =Ya '
7.13.16. Determine the coordinates of the centre of gravity of the

2 2 2
portion of the arc of the astroid x 6

'
+y =a* 3
situated in the first
quadrant.
7.13.17. Find the Cartesian coordinates of the centre of gravity of

the arc of the cardioid p=a(l cosq)) between q) and y n. + = =
Solution. Let us represent the equation of the cardioid in para-
metric form:
x = p cos = a + cos
q) ( 1 (p) cos (p;
y = p sin = a (1 +cos
q) cp) sin (p.
As the parameter qp varies between and ji the running point describes

the upper portion of the curve. Since the length of the entire car-
dioid equals 8a and
dl =V + (y'^f
9/
d(p - 2a cos ^ d(p (see Problem 7.9.3), we have
JT
xc = \ \y ^ ^Va [
asin W + cos <P) ^ a cos-|- dcp =
:
2a ^ cos 4 y sin -|- dcp = — 4- a cos y 5
i 4
5
Analogously,
JT
yc=z
Ta \
xdl =^«cosq)(l + cos(p) 2acos^d(p =
L
ji n
=aj cos cp cos 3

yd(p = a
J
^2cos 6 -|- — cos 3
dep.
Putting y= / we get (see Problem 6.6.9)
_JT_
T
yc = 2a \ (2 cos b t — cos* t)dt =4ftg^ — 2ay = -^-a.
And so, xc = y =^.

c
It interesting to note that the centre of gravity of the above-

is
considered half of the arc of the cardioid lies on the bisector of

the first coordinate angle, though the arc itself is not symmetrical
about this bisector.
7.13.18. Find the centre of gravity of the figure bounded by the

ellipse 4a: 2 9y
2
+ =
36 and the circle x 2 y
2
9 and situated in the + =
first quadrant (Fig. 113).
Solution. Let us first calculate the static moments

3 3
2
dx =
3
=y x V~9 —x 2
dx = 3\
I
b
3 3
5 — x2 ) dx =5.
The area of a quarter of a circle of
radius 3 is equal to ~ , and the area of a
quarter of an ellipse with semi-axes a =3

b=2
3ji
x
and equals y , therefore the area
of the figure under consideration is

9jx 3jt 3ji
Fig. 113 q
Thus,
My 4 Mx _ 20
~S"~ 3n'
7.13.19. Find the centre of gravity of the figure bounded by the

_L _L _L
parabola x 2
+y =a 2 2
and the axes of coordinates.
7.13.20. Find the Cartesian coordinates of the centre of gravity

of the figure enclosed by the curve p acos 3 q) (a 0). = >
Solution. Since pÔ
in all cases, the given curve is traced
when q) changes from — y to y . By virtue of evenness of the fun-
ction coscp it is symmetrical about the polar axis and passes through
the origin of coordinates at (p = ±y.
Compute the area S of the figure obtained:
iX3x5
S = 2xy j =a ji
2 2
p dy 6 2
Jcos (pd(p= a 2x4x6 X '32
2
Now arrange the axes of coordinates as shown in Fig. 114. Then

the parametric equations of the curve are
x — p cos (p = a cos 4
cp;
y = p sin (p = a sin cp cos 3 (p.
The centre of gravity of the figure lies on the x-axis, i.e. y c =

by virtue of symmetry about the #-axis. Finally, determine x c \
_
2
r*
\
,
xy ax —jt
-y
ji
: - JL_ — = ~- J cos 10 (p sin 2 (p dtp = ~-

J
(cos 10 cp — cos 12
(p) d(p =
8fl 3 / 1X3X5X7X9 1x3x5x7x9x11 \ ji 21

:
(5/32)jia 2 ^2x4x6x8x10 2x4x6x8x10x12; 2

~~40 fl "
Fig. 114 Fig. 115
7.13.21. Find the coordinates of the centre of gravity of the

2
figure bounded by the straight line y=—x and the sinusoid
*/ = sinx (*>0) (Fig. 115).
Solution. The straight line y = —2 x and the sine line y = s\nx inter-
sect at the points (0, 0) and , 1^ . The area of the figure
bounded by these lines is
5=
J
^sin jc ——x dx-
Hence,
yj( sin2 *-^*

2
d*
)
4 — jt
sin 2x 4
J_ Xr
4 — ji 2 4 3jx2- "6(4 — ji)
•
^x(^sinx — ^-x^dx —
12 — ji
2
dx — —
Jt (4- 3(4 ji) 12 3ji"
7.13.22. Prove the following theorems (Guldin's theorems).

Theorem 1. The area of a surface obtained by revolving an arc
of a plane curve about some axis lying in the plane of the curve and
not intersecting it is equal to the product of the length of the curve
by the circumference of the circle described by the centre of gravity
of the arc of the curve.
Theorem 2. The volume of a solid obtained by revolving a plane
figure about some axis lying in the plane of the figure and not
intersecting it is equal to the product of the area of this -figure by
the circumference of the circle described by the centre of gravity of
the figure.
Proof. (1) Compare the formula for the area of the surface of
revolution of the curve L about the x-axis (see § 7.10)
P=2n\ ydl }
L
with that for the ordinate of the centre of gravity of this curve
MX
y c = -f=T)ydi.
f Ai I
L
Hence we conclude that
P = 2n lyc = l-2ny ct
where / the length of the revolving arc, and 2ny c is the length
is
of a circle ofradius y c9 i.e. the length of the circle described by

the centre of gravity when revolving about the x-axis.
(2) Compare the formula for the volume of a solid generated

by revolving a plane figure about the x-axis (see § 7.6)
-yl) dx
a
with that for the ordinate of the centre of gravity of this figure
b
M x
yc = -j
Hence we conclude that
V= ji • 2S y c =S • 2 ny c
where S is the area of the revolving figure, and 2ny c is the length
of the circumference described by the centre of gravity when revol-

ving about the x-axis.
7.13.23. Using the first Guldin theorem, find the centre of gra-
vity of a semicircle of radius a.
Solution. Arrange the coordinate axes as shown in Fig. 116. By
virtue of symmetry ^=
0. Now it remains to find y c If the semi- .
circle revolves about the *-axis, then

the surface P of the solid of revo-
lution is equal to Ana 1 and the ,
arc length l =
na. Therefore, accor-
ding to the first Guldin theorem,
Ana 2 = na • 2ny c
7.13.24. Using the second Gul-

din theorem, find the coordinates
\ yc — 2—.
Fig- 116
-a
Y a
X
of the centre of gravity of the

figure bounded by the x-axis and one arc of the cycloid: r =
—
=a (t sin/); y a ( 1 cos t). = —
Solution. By virtue of the symmetry of the figure about the
straight line x~na its centre of gravity lies on this straight line;
hence, x c na. =
The volume V obtained by revolving this figure about the x-axis
is equal to 5n 2 a 3 (see Problem 7.6.14), the area S of the figure
being equal to 3na 2 (see Problem 7.4.3). Using the second Guldin
theorem, we get
V 5ji 2 a 3 _
_5a
~
y° ~~ 2nS 2jt-3jia 2
~~ ~6 '
7.13.25. An equilateral triangle with side a revolves about an

axis parallel to the base and situated at a distance b a from the base. >
Find the volume of the solid of revolution.
Solution. There are two possible ways of arranging the triangle

with respect to the axis of revolution which are shown in Fig. 117,
a and b.
The altitude of the equilateral triangle is h--

aV 3 the area
a 2 |/"3
The centre of gravity 0' is situated at the point of
a }A3
intersection of the medians and at a distance of b- from the
Q ^"3
axis of revolution in the first cas. j
, and b+ in the second.
Fig. 117
By the second Guldin theorem

2na 2 V3 aY3_
n
a*bV3
6 2
2na*V~3 a]/"3 a*bV3

v 9 = 6 4 "*\ 2 ^ 4
= ji 1
V J
7.13.26. Find the centre of gravity of the arc of a circle of radius
R subtending a central angle 2a.
7.13.27. Find the centre of gravity of the figure bounded by
the arc of the cosine line y = cosx between x= — y and x = ^ and
the straight line # = y.

7.13.28. Find the coordinates of the centre of gravity of the
figure enclosed by line y 2 = ax 3 x4 — .
7.13.29. Find the Cartesian coordinates of the centre of gravity

of the arc of the logarithmic spiral p = aev from (p, = y to (p 2 = ji.
7.13.30. A regular hexagon with side a revolves about one of

its sides. Find the volume of the solid of revolution thus generated.
7.13.31. Using Guldin's theorem, find the centre of gravity of
a semicircle of radius R.
7.14.1. Find the area of the portion of the figure bounded by

the curves y m =
xn and y n x m (m and n positive integers) situated
=
in the first quadrant. Consider the area of the entire figure depen-
ding on whether the numbers m
and n are even or odd.
7.14.2. (a) Prove that the area of the curvilinear trapezoid
bounded by the x-axis, = = straight lines x a, x b and parabola
y= Ax* Bx 2 + Cx + D can be computed using Chebyshev's formula
+
'
(a-\-b I b — a\ ,
fa + b\ ,
fa + b . 1 b — a\~]
(b) Prove that an analogous area for a parabola of the fifth order
y = f(x) = Ax + Bx" + Cx + Dx + Ex + F
b 3 2
can be computed using the Gauss formula
S=
^[ ,(£±»-/F?) + 8f(^) +
5
7.14.3. Show that the area of a figure bounded by any two ra-
dius vectors of the logarithmic spiral p ae m and its arc = '+
is pro-
portional to the difference of the squares of these radii.
7.14.4. Prove that if two solids contained between parallel pla-

nes P and Q possess the property that on being cut by any plane
R parallel to these planes equivalent figures are obtained in
their section, then the volumes of these solids are equal (Cava-
lien's principle).
7.14.5. Prove that if the function 5 (x) expressing (O^x^h)

the area of the section of a solid by a plane perpendicular to the
x-axis is a polynomial of a degree not higher than three, then the
volume
Using
of this solid is equal toV = ^ S (0) + 4S
deduce formulas for computing the volume of
this formula,
^ +S (h) .
a sphere, spherical segments of two and one bases, cone, frustrum

of a cone, ellipsoid, and paraboloid of revolution.
7.14.6. Prove that the volume of a solid generated by revolving

about the (/-axis the figure a^x^b
O^y^y(x), where y(x) is y
a single-valued continuous function, is equal to

b
V= 2ji
J
xy (x) dx.
7.14.7. Prove that the volume of the solid formed by revolving,

about the polar axis, a figure Oâ^cp^fJ^rc, O^p^p(q)),
is equal to
P
V == —
2ji r
l
p
3
((p) sin q)d(p.
a
7.14.8. Prove that the arc length of the curve given by the pa-
rametric equations
jc = nOcos/ + nOsin*,
y = — f"(t)sint + f'(t)cost
is equal to [f (t) + f (t)]l\.
7.14.9. Find the arc length of the curve represented parametri-
cally
x
cos
= C—
\
z
^— dz,
,
y = \
—
C sin z ,
dz
between the origin and the nearest point from the vertical tan-
gent line.
7.14.10. Deduce the formula for the arc length in polar coor-
dinates proceeding from the definition without passing over from
Cartesian coordinates to polar ones.
7.14.11. Prove that the arc length l(x) of the catenary y —
= cosh x measured from the point (0, 1) is expressed by the for-
mula / (#) =sinh a; and find parametric equations of this line,
using the arc length as the parameter.
7.14.12. A flexible thread is suspended at the points A and B
located at one and the same height. The distance between the
points is AB =2fr, the deflection of the thread is /. Assuming the sus-
pended thread to be a parabola, show that the length of the thread
at a sufficiently small y.
7.14.13. Find the ratio of the area enclosed by the loop of the
curve y =±
.
3
— x Vx to the area of the circle, whose circum-
ference is equal in length to the contour of the curve.
7.14.14. Compute the length of the arc formed by the intersection
of the parabolic cylinder
and the elliptic cone
between the origin and the point M (x, y y

z).
7.14.15. Prove that the area of the ellipse
Ax 1
+ 2Bxy + Cy* + 2Dx + 2Ey + F = (AC — B > 0)
2
is equal to
A B D
S = where A -= B C E
(AC—B- D E F
7.14.16. Find: (a) the area S of the figure bounded by the hy-
perbola x 1 y'1 — U the positive part of the x-axis and the radius
vector connecting the origin of coordinates and the point M(x, y)
lying on this hyperbola.
(b) The area of the circular sector Q bounded by the x-axls and
the radius drawn from the centre to the point (x, y) lying on N
the circle x 2 +
y*— 1. Prove that the coordinates of the points M
and N are expressed respectively through the areas S and Q by
the formulas
a;^ =cosh 2S, yM = s'mh2S, xN = cos2Q, y N = s')n2Q.

7.14.17. Using Guldin's theorem, prove that the centre of gra-
vity of a triangle is one third of the altitude distant from its base.
7.14.18. Let £ be the abscissa of the centre of gravity of a cur-

vilinear trapezoid bounded by the continuous curve y f(x), the =
x-axis and the straight lines x--^-a and x — b. Prove the validity
of the following equality:
b b
\
(ax + b) f (x) dx - (a% + b) \ f
(x)dx
a a
(Vereshchagin's rule).
7.14.19. Let a curvilinear sector be bounded by two radius vec-

tors and a continuous curve p^/((p). Prove that the coordinates
of the centre of gravity of this sector are expressed by the follow-
ing formulas:
<p 2 <p 2
y 3
p cos (p dcp p sin cp dcp
^
X = 2
3"
cp
N-' 2 H 2
2
^ P *P
386 Ch. VI /. Applications of the Definite Integral
7.14.20. Prove that the Cartesian coordinates of the centre of

gravity of an arc of the curve p =
/((p) are expressed by the fol-
lowing formulas:
<p 2
^
pcoscp |/"p 2 -|- p"2 rfqp
^ p
sin cp ^~p 2 + p' 2
cfcp
= 2i ; « =JBi-
Chapter
8
IMPROPER INTEGRALS
§ 8.1. Improper Integrals with Infinite Limits

Let the function f (x) be defined for all xâ and integrable on
A
any interval [a, A], Then lim \ f (x) dx is called the improper
integral of the function f (x) in the interval [a, +00] and is de-
+ 00
noted by the symbol ^ f(x)dx. We similarly define the integrals

a
B +00
^ f
(x)dx and ^ f (x)dx.
— GO — 00
Thus,
+ GO A
f / (x) = lim ^ f
(x) dx;
a A-> + v a
B B
^f(x)dx= lim ^f(x)dx\

-oo A-*-a> A
+ 00 C B
\ f{x)dx= lim \f(x)dx+ lim [f(x)dx.

-oo A -~™ A fl-* + co£
If the above limits and are finite, the appropriate integ-

exist
rals are called convergent;otherwise, they are called divergent.
Comparison test. Let f (x) and g(x) be defined for all and xâ
integrable on each interval [a, A], If (x) (x) for Aâ. 0^/ ^g
00
f»
all xâ, then from convergence of the integral \

j
g(x)dx it fol-
a
CO GO
lows that the integral ^f(x)dx is also convergent, and ^((xjdx^.

388 Ch. VIII. Improper Integrals
< ^ g (x) dx; from divergence of the integral ^f(x)dx it follows that
a a
OO
the integral ^g(x)dx is also divergent.

a
Special comparison test. If as x —oo the function f(x)^0 is an
infinitesimal of order X > as compared with -~ » then the integral
+ X
J
f(x)dx converges for X > 1 and diverges for X^l.
a
Absolute and conditional convergence. Let the function f(x) be
00
defined for all xâ. If the integral

J | / (x) \
dx converges, then the
a
00
integral [f(x)dx also converges and is called absolutely convergent.

a
In this case
00 CO
\f{x)dx <^\\f(x)\dx.
a a
If the integral \
j
f(x)dx converges, and [
)
\f(x)\dx diverges, then
a a
x
the integral ^f(x)dx is called conditionally convergent.

a
The change of the variable in an improper integral is based on
the following theorem.
Theorem. Let the function f (x) be defined and continuous for
xâ. If the function x y(t) defined on the interval = y
a</<|3
(a and P may also be improper numbers oo and oo), is monoto- —
nic, has a continuous derivative q/(/)^0 and limq) (t) = a,
t-> a+O
lim cp (t) = + °°» then
t > p - o
J
oo (
,
[f {x)dx= \f(y(t))v' (t)dt.

a a
Integration by parts involves no difficulties.
Evaluate the following improper integrals with infinite
8.1.1.
limits or prove their divergence taking advantage of their defini-
tion.
w]lW-x> W L+t +-co

5
; (c)J*sin*d*.
e2
Solution, (a) By definition,
jxln 3 *
e2
^ Jx
e2
In 3 x aI™ x { 2 In 2 *
= lim
(b) By definition,
oo o A
2
+ 2x + 5
(instead of the point x any other finite point of the #-axis= may
be taken as an intermediate limit of integration).
Compute each of the limits standing in the right side of the
above equality:
o
s
= y arctan T + T
1 , 1 .
Jl
,
= * + 1 ^ ji l , l
1™ f o ,
1*
i c 'i m 4- arc tan
2 =T— arctan -2'
o
Hence,
—X
+ 2* + 5 2
"
(c) By definition,
qo A
^xs\nxdx= lim [xs'mxdx.
o ^ + 00
o
Putting w = cfo = sin xdx and integrating by parts, we get
A / A x
14
lim ^xsinxdx^ lim f a;cosa;|^ + J cosArdA; j==
o
= lim ( A cos A + sin ^4).

j4 -+ + CO
But the last limit does not exist. Consequently, the integral
QO
a: sin atâ; diverges.

J
8.1.2. Evaluate the following improper integrals with infinite

limits on the basis of their definition:
dx
x+x* (O
j V" (4x
2
+l) 3 '
QO
dx dx
(! + *)'
(e)
] x2 — 6x+10 ;
(f)
J
Solution, (a) By definition

OD A
C x dx j. C xdx
lim _l_ (x 2 — 3)~ 1/2
/4 —>• + GO 2 -\I2
=— lim
\
-3
— = 1 1.
y4 -> + GO V <4
2
8.1.3. Prove that the integrals of the form
e" px dx and g^dx

J J
converge for any constant p > and diverge for /? < 0.

8.1.4. Test the integral
00
dx
f
J l+2x 2 + 3x 4
for convergence.
Solution. The integrand
l + 2x + 3x
2 4
is positive and is an infinitesimal of order A, =4 as compared with

•j as a; —oo. Since 4 > 1, the integral converges according to the
special comparison test.
C dx
J x+sin 2
for convergence.
Solution. The integrand f(x)- is continuous and posi-

x+sin 2 x
tive for x ^ 1.
As jc—* oo the function f(x) is an infinitesimal of order X=l

as compared with ;
according to the special comparison test the
integral diverges.
8.1.6. Test the following integrals for convergence:

cc co
i
1 r
1 1
3+arc sin —x f arc tan x

+ cosa:
.
, x i 2 , ,n , .
x V ,
J J i j/ x
i
(*+ + 1) dx
for convergence.
Solution. The integrand is continuous and positive for x^\.
Determine its order of smallness X with respect to as x — oo;
since
x+ Vx+\ 1
1
+ 1^ * +
*2 +2 ^+l *
1+2 l/_L.±
the order of smallness According to the special comparison
00
test the integral (

*+ — Hy diverges.

GO
r dx
J Vx (x-\)(x-2)
3
for convergence.
Solution. Since the function
f(x)= r _r_-±—- /
—=~ x
is an infinitesimal of order ^ = y3 with respect to —
1
as # — +oo,
according to the special comparison test the integral converges.
for convergence.
Solution. The integrand is continuous and positive for x^2.

Determine its order of smallness with respect to — as x — + oo:
= -TT X
V
v-35
Since the second multiplier has the limit \/2 as x —+ <*), we

have 1 = -^- < 1. Consequently, the given integral diverges.
1 — cos—x dx
for convergence.
Solution. The integrand
f(x)=
'
v
'
1 — cos- = 2sin -
X
2
y ~2 — = -p-,
1
is positive and continuous for x^l. Since 2 sin 2 f
J
the given integral converges (by the special comparison test).

i
"
e
In dx. n>0
for convergence.
Solution. Transform the integrand:
1 + -
Since the function is an infinitesimal as x—«>, then
/(*) l
f(x)< . In other words, lim According to
the special comparison test the given integral diverges.

1 — 4 sin 2x ,
dx
for convergence.
§ 8.1. Improper Integrals with Infinite Limiis 393
Solution. The function f(x) = j ^ sin 2x

TT^" c h an 2es its sign together
x 3 -\- y x
with the change in sign of the numerator. Test the integral
1 — 4 sin 2x|
|
dx
'[/ x
— 4 sin
<^ —
1 2a: |
5
for convergence. Since 3
, and the integral
COnV/erges, the integral f — ^-^^-dx converges as well (ac-

j J x*+y x
cording to the comparison test). Thus, the given integral converges
absolutely.
8.1.13. Prove that the Dirichlet integral
J x
converges conditionally.
Solution. Let us represent the given integral as the sum of two
integrals:
= \—
C slnx
dx=
. C sinx
\
— dx+ \— . P sin* ,
dx -
The first is a proper integral (since lim

x
= 1
) . Applying the
\ J x-> o
method of integration by parts to the second integral, we have
oo A
smx
dx = m\\ dx'-
J
A C cos x dx P cos*
= lim
x*
A -> co IL J
2 Jl
T J
But the improper integral

j
-^^dx converges absolutely, since
71
2
co
^ ^ ^
, and the integral [ ^ converges.

Therefore, the integral ^-^~dx converges.

o
Reasoning in a similar way it is easy to prove that the integral
CO
c ° sx
[
X
dx also converges. Now let us prove that the integral
n
2
J_5!2_LL dx diverges. Indeed,

J
sin x |
^ sin 2 x 1 — cos 2x
x x 2x
but the integral
PI— cos2x , 1 fdx 1 pcos2x ,
/I -»• CO
31 Jl
T 2 T 2
=4 lim In 4 - 1 In £-1 j
It ji 1
00
^
C cos 2x
X
<te
f* cos 2x
diverges, since lim In A =oo, and the integral \ converges.
JX
T
8.1.14. Prove that the following integrals converge
Ot 00 00
2 2 4
(a) ^ s\n (x ) dx; cos (a; ) dx; (b)
J
2x cos (jc ) dx.
J
Solution, (a) Putting % = K*, we find

00 00
jVin (x 2 )dx = -^ -y^dt.

^
b o
Let us represent the integral on the right side as the sum of two
integrals:
V' J V i J V i
The first summand is a proper integral, since lim -^£i- = 0. Let

/- + o V t
§8.1. Improper Integrals with Infinite Limits 395
us apply to the second summand the method of integration by

parts, putting
u= jV U \ sin t dt = dv,
C sin / , , cos I 00
_L f
CQS t dt _ 1 C cos I dt.
Jt/2 Jt/2 jt/2
COS / I
The
I
last integral converges absolutely, since

^"^yr » an ^
f3/2
QO
the integral ^"^72 converges. We can prove analogously that the

Jl
QO
T
integral cos(x 2 )dx is convergent. The integrals considered are cal-
J
led FresneVs integrals. They are used in explaining the phenome-

non of light diffraction,
(b) By the substitution x2 = t this integral is reduced to the
QO
2
integral Jcos(/ )d/. The latter integral converges as has just been
proved.
Note. Fresnel's integrals show that an improper integral can con-
verge even when the integrand does not vanish as x oo. The last —
convergent integral considered in item (b) shows that an improper
integral can converge even if the integrand is not bounded. Indeed,
at x = l/nn (n = 9 1, 2, ...) the integrand attains the values
=h \/ tin, i.e. it is unbounded.
8.1.15. Evaluate the improper integral
QO
dx
? n natural number.
J (l+* 2 )"'
Solution. Make the substitution x = tan t, where 0^£<y. Then
x=0 at /=0, x — + 00 as t —>— — jt

and x't^—^^Q.
1
Conse-
quently, by the theorem on changing a variable in an improper
integral
jt ji
~2~
oo 2
"
I(TW«I« xsec2M/= I cos2,, 2/d/ -
On changing the variable we obtain the proper integral which

was computed in Problem 6.6.9.
Therefore,
n/2 n = l >
? dx
aX
= f
>
1.3.5. ..(2/i — 3)
) ji
^
\
J<
l +* a
>"
\ 2.4.6. ..(2.-2) 'T- *> L
,
X
8.1.16. Compute the integral / = j-j-p^d*.
Solution. Apply the substitution
*=l/f; dx = — (l/t 2
)dt\ ^ = 00, f2 = 0;
If another integral
/
/
?
J
is
*2
l+* 4
added
//y _? (^
00
J
to
l
the
+
4
)^
l// 4
right
_?
J
and
/
4
+ l
•
left sides then we
2
dt.
r2 -f \/t
= t— l/t, = dz.
— + 00, z — ++ oo.
Make the substitution z (I l/t
2
)dt Then, as
* —H-O, — — 00 z and as t »- Hence,
00 r a
C dz dz dZ
lb
..
. 1 1
f ,
f
-00
1
lim arc tan—^L-H !—=- lim arc tan—

2 )/" 2 B -* - CD V2 2^2 ,4 -+ ^2
I / JI Jl \ Jl
+
8.1.17. Evaluate the following improper integrals:
X 00
2
(a) l~Y^dx\ (b) §e-* x 2m+1 dx.
accurate to two decimal places.

Solution. Represent the given integral in the form of a sum of
two integrals
x5 + * +l 2
1 7v
§ 8.2. Improper Integrals of Unbounded Functions 397
Compute the former with the required accuracy, using Simpson's

formula, and estimate the latter. Since for 1 we have x^
then
0< /2 = ^x-T*dx = ^N-*'*.

N
At N= 7 we get the estimate L<^ X —-= <1
0.0031.
Computation of the integral
by Simpson's formula for a step = l gives
5^0.2155,
and for a step -^- = 0.5
S . 5
=0.2079.
Since the difference between the values is 0.0076, the integral I t
gives a more accurate value S 5
= 0.2079 with an error of the order
°^L6 - 0.0005.
Consequently, the sought-for integral is approximately equal to
/ & 0.208
with an error not exceeding 0.004, or / = 0.21 with all true deci-
mal places.
§ 8.2. Improper Integrals of Unbounded Functions

If the function f (x) is defined for a^x integrable on any
interval [a, b — e], 0<e<b— a and unbounded to the left of the
point b, then, by definition, we put
b b-v
\ f (x)dx= lim \
f (x)dx.
If this limit is existent and finite, then the improper integral is

said to be convergent. Otherwise it is called divergent.
Analogously, if the function f (x) is unbounded to the right from

the point a, then
b b
\ f (x) dx= lim \ f(x)dx.

Finally, if the function is unbounded in the neighbourhood of an

interior point c of the interval fa, 6], then, by definition,
b c b
J f {x)dx = $ f (x)dx + J f
(x) dx.
a a r
Let the function be continuous on the interval [a, b] except

f (x)
at a finite number If there exists a function F (x) conti-
of points.
nuous on [a, b] for which F' (x) f (x) except at a finite number =
of points, then the Newton-Leibniz formula
]f(x)dx=F(b) — F(a)
a
holds good.
Sometimes the function F (x) is called a generalized ant [derivative
for the function on the interval [a, b}.
f (x)
For the functions defined and positive on the interval a^x<b
convergence tests (comparison tests) analogous to the comparison
tests for improper integrals with infinite limits are valid.
Comparison test. Let the functions f (x) and g(x) be defined on
the interval a^x<band integrable on each interval [a, b e], —
< e <b —a. If ^f(x)^g(x), then from the convergence of the
b b
integral \^g(x)dx follows the convergence of the integral ^f(x)dx,

a a
h b
and ^ f (x)dx^^ g(x)dx; from the divergence of the integral

a a
b b
^f(x)dx follows the divergence of the integral ^g(x)dx.

a
Special comparison test. If the function f(x)^0 is defined and
continuous on the interval and is an a^xb — 0, then

b
the integral ^f(x)dx converges for X< 1 and diverges for X^\.
a
In particular, the integral
b
C dx
(b — xf
converges for X 1 <
and diverges for % 1. ^
Absolute and conditional convergence. Let the function f (x) be
defined on the interval a?Cx<b and integrable on each interval
[a, b— e]; then from the convergence of the integral ^\f(x)\dx

a
b
follows the convergence of the integral ^f(x)dx.

a
b
In this case the integral ^f(x)dx is called absolutely convergent.

a
b b
But if the integral ^f(x)dx converges, and the integral ^\f(x)\dx

a a
b
diverges, then the integral ^ f(x)dx is called conditionally convergent.

a
b
Analogous tests are also valid for improper integrals ^f(x)dx,

a
where f(x) is unbounded to the right from the point a.
8.2.1. from the definition, evaluate

Proceeding the following
improper integrals (or prove their divergence):
2
dx
(a)
l4^T : (b)
I cos,-
3 2
(c )
J
1
f4 x-x>-3
;
^ J FTT^f
P* 3 + ?/*— 2 f dx
V
Solution, (a) The integrand / (x)
=— .. \ is unbounded in the
x y \wx
neighbourhood of the point jt = l. It is integrable on any interval
[1+e, e] 9
since it is a continuous function.
Therefore
f^==
J
\
xr/\nx
V
lim
S-++Q J
1+8
\-J*
xW\nx
V
= lim
e^+o [i
3
/lH^|
e
|l
1
+-ej
=
8^+0
lim !* /ln»(l+e)
2 2
3 3_
2
'
(b) The integrand !(k) = —— is unbounded in the neighbourhood
of the point * =y and integrable on any interval

y as
a continuous function. Therefore
-t
dx
= lim
J cos* e ^+ Jf cos x
= lim
e-> + o
In tan f
W4 + ^1 4 /
T
o
e
= Hm
b-> +o
In tan (£
\ 1
— 1
Hence, the given integral diverges.

(c) The integrand is unbounded in the neighbourhood of the points
x= 1 and x = 3. Therefore, by definition,
c dx r dx r dx
J V~4x— x'
2 — 3 ~~ J V~4x— x2 —3 J VT? -a;
2 —3
{instead of the point x =

2 we can take any other interior point of
the interval [1, 3]). Let us now compute each summand separately:
f ^ ^ = lim ( r
dx
== lim arc sin — 2) 2
1 + tr
1 1 + <?
= lim [0 — arc sin (e— 1)] ;
3 3-t
f
J
-==
^-^-3
d*
6-.
lim
+
(
J
-7=^==
\
r \-(x—2)*
lim arc sin (x
8-. +
— 2)
2
' =
2 2
= ^[3^ sin (l-e)

-0]=f
Hence,
(d) The integrand f (x) = r is unbounded in the neigh-
bourhood of the point x=l, which is an interior point of the

interval of integration. Therefore, by definition,
2 I l
r
dx ~ r
dx i r — 2
JC
|
§8.2. Improper Integrals of Unbounded Functions 401
Evaluate each summand separately. If 0^*< 1, then

1 -e
r dx _ r dx _ Hm P dx _
lim arc sin x = lim [arc sin (1 — e) —-0] =~1

8 -* + 8- +
If 1 < x<2, then
dx
J V^l 1 —x 2
\ J ^2_! p _+ J
j/- x 2 —\
1 1 1 +e
lim lnU + K* 2
— 1)
8-> +
8
lim [ln(2
-> +
+ K3) — ln(l + e + +e) 2 - l)] = In '2 + j/"3 ).
Hence,
= | + ln(2 + |/3).
(e) Represent the given integral as a sum of three items, divid-

3
ing each term of the numerator by £/ x ,
dx
"
„3/5
The first summand is a proper integral evaluated by the Newton-

Leibniz formula:
i
17' lo
17'
The second and third summands are unbounded to the right of the
point x = 0. Therefore,
i i
dx
lim ^ x 11 / 15
1
_ !^
* 4/15 8-> +
analogously,
f-^L = ii tn f JL = m A ^2/6 __5_

~2 ii
1
Hence,
i
625
dx = T7 + T- 2 'Y = "187*
(f) Represent the integrand f(x) = -

in the form of a sum
of partial fractions;
1 1 x+2
/(*) =
1
— a:
3
(1— x)(l +x+x 2
)
TU
Then
P
j
_ _
djc
= - jC 1 dx
+ - Jf
,
1
o
x +2 ,
c .
Since
1-8
dx l-e
: lim lim In (1 x)
J I-* E _
8-> +
the given integral diverges. There is no need to compute the second

summand representing a proper integral.
Note. Evaluation of the improper integrals from Problem 8.2.1
(a to f) can be considerably simplified by using a generalized anti-
derivative and applying the Newton-Leibniz formula. For instance,
3 3 s
in Problem 8.2.1 (a) the function F (x) -^ In 2 * is continuous on = /
the interval [1, e] and differentiate at each point of the interval
1 <xîe, and F' (x) = f (x) on this interval. Therefore
dx
= \\/\tfx _3
2
*
8.2.2. Proceeding from the definition, compute the following

improper integrals (or prove their divergence):
2/JT
2x dx
(a)
('
(b) fsinl.§;
8.2.3. Evaluate the following improper integrals:

3
-3 b
Solution, (a) Find the indefinite integral
i-, ^9=^) + C.
TCT =
J"
(9 arc sin
J
The function r
(x)= y^9 arc siny — xK9 x 2
^j is a generalized anti-
derivative for f(x) = y==== on the interval [ — 3, 3], since it is
continuous on this interval and F'(x)=f(x) at each point of the

interval (—3, 3). Therefore, applying the Newton-Leibniz formula,
we get
-7= = Tr 9 arc sin —x y 9 x- =-r n.
-3
(b) Transform the integrand
\/~*±x== 2
+* = 2 X
f (x)=s +
The indefinite integral is equal to
]/ |^dx^2arcsin|-l/4^7 + C.
2
j
The function f (x) =2 arc sin -|- — ]/~^ —x 2
is a generalized antideri-
vative for on the interval [0, 2], since it is continuous on

f (x)
this intervaland F'(x)=--f(x) on the interval [0, 2).
Therefore, applying the Newton-Leibniz formula, we get
2
2
^/2^ dx== ^ 2arcsin ^_ |

/4ZI72 = + 2.
Jl
f j| o
b

1
dx
-1
x /x
3
for convergence.
Solution. At the point x = the integrand goes to infinity. Both
i^
integrals f 3* and ( 3* diverge, since ^ = -g- > 1. Gorisequent-

J X v
*/ X J X y /X
-\
ly, the given integral diverges. If this were ignored, and the New-
ton-Leibniz formula formally applied to this integral, we would
obtain the wrong result:
1
dx 3
f
= —6.
1: i
- i
And this is because the integrand is positive.
8.2.5. Test the following improper integrals for convergence:

i i
/
(a)
\
\—F==dx\
P j e* u\
b)
Psin#-{-cos#
^
w -
,
dx.
v
Solution, (a) The integrand is infinitely large as x —+ + 0. Since
V 1 — cosx = J/2 sin ~ -J^p x as + (),
the integrand has the order h=l as compared with ~. According

to the special comparison test the given integral diverges.
(b) Rewrite the integrand in the following way:
r .
/(*)=
, sin #-|-cos x
5
~ 1
l/\ +x +x* y\~x

This function is infinitely large as x—+ 1, its order is equal
to h = -^r as compared with y~^, since the first multiplier tends
to 1 as x —>0. Therefore, by the special comparison test, the given

integral converges.
8.2.6. Test the following improper integrals for convergence:
(a) e5in ,_, dx; (b)

\
b T
K
i
,
(c)
x i
J v/ x
cos
—
—
a: d#
sin
n (l 4-
In i/x?)
Solution, (a) The integrand f (x)= J,; ^
is positive in the
interval (0, 2) and is not defined at x = 0. Let us show that

lim f(x) oo. Indeed, since =
e smx_i _ s'mx~ x, In (1 + l/'x*) as x-*0,
we have
lim
*-+ o
— .
v
1
= hni
x ->
- —=
*
lim .
*- o ^/* 2
= oo.
At the same time we have shown that /(x)~y-L= as x > 0, i.e.
that /(a;) is an infinitely large quantity of order k = —2 < 1 as com-
pared with — . Consequently, by the special comparison test, the

given integral converges.
(b) Determine the order of the infinitely large function f (x)
=
— in the neighbourhood of the point x =2 with respect
j/16-x4
l
to ^4^- T° this end transform the expression for f (x):

2 — x'
r '
j/l6-A: 4~ y4+x*l/ 2 + x 1/2 — x

Hence it is obvious that the function f (x) is an infinitely large
quantity of order A, = y<l as x —*2. According to the special
comparison test the given integral converges.
(c) The integrand / (x) =

y — sin x
.
/
CQS x
x
— is unbounded in the neigh-
bourhood of the point x = 0. Since

tt x
cos a: cos* 1
A
V x — smx y x i/*
as +
jc—> the function / (*) is an infinitely large quantity of order
k=-j < 1 as compared with and, by the special comparison test,
the integral converges.

8.2.7. Investigate the following improper integrals for conver-
gence:
ex dx ,i x f x l dx
(b)
(a)
.1 Tct ;
j K y u=ji? *
b o
1 2 _
^ /n r in +
fe)
V
'
J
r
x-sinx ' W J
(
gtanx_ 1
1
f sin•
7 dx
1 Vx
converges.
Solution. For <x< 1
sin —
x
Vx
C dx
But the integral y= converges, therefore, by the comparison
j
b
i
sin(i/*)
test, the integral
[
^x a j sQ conver g eS) an(j consequently
the given integral converges absolutely.

8.2.9. Prove the convergence of the integral
/ =^ In sin xdx
and evaluate it.
Solution. Integrate by parts, putting u = \n(s\nx), dx = dv:
lnsin xdx=x\ns\n x —C \
J
COS X j
x - —
dx=
sin x
—
J
C
\ t
X
tan x
dx.
,
Since lim
tan x
1, lim
i^^ — Q' ê ' as * integral is a proper
one. Consequently, the initial integral converges.

Now make the substitution x 2t in integral = /. Then dx = 2dt;
x = at / 1= =0; x =j at t2 = ^. On substituting we get:
n/2 JT/4 Jl/4
5 In sin xdA: = 2
J
lnsin2^=2 J
(In 2 + In sin + In cos f)df =
t
JT/4 JT/4
4
= 2Hn2f + 2
J
lnsin/<« + 2
J
lncos/df =
o o
JT/4 jt/4
= yln2 + 2 lnsin /df + 2 j In cos /d/.

J
In the last integral make the substitution £ = jt/2 — z. Then

dt = — dz\ = t at 2,=ji/2; t-=n/4 at 22 = jx/4. Hence,
JT/4 JT./4 77/2
2^ In cos tdt — — 2^ In cos — z^dz = 2 ^ \n sin zdz.

JT/ 2 jt/4
Thus,
JT./2 JT/4 JT./2
In sin xdx = j In 2 + 2 In sin tdt +2 In sin 2 =

J J
Jt/4
Ji/2
=J ln2 +2 lnsinfd* = -jln2 + 2/.

J
Whence
^=
J
\ns\nxdx=- — In 2.

1
dx n
fx
yr==z(n a natural number).
Solution. The integrand is an infinitely large quantity of order

A, =y with respect to -j-^ as x—>1 — 0. Therefore, the integral con-
verges.
Make the substitution x = sin / in the integral. Then dx = costdt,
x = at ^ = 0, x=l at t = n/2. On substituting we get
00
1 Jl/ 2 JT/ 2
—rxn dx = C s\n n /-cos tdt

= f .
s\n n tdt.
„ ,
J 1^ l
-* 2
J
cos/
J
The last integral is evaluated in Problem 6.6.9:

n — n —3 jx
1
yy, 1
ft even,
— ——
n
-
n
^
2
. . .
~3 »
1
n odd.
Consequently, the given integral is also computed by the same

formula.
8.2.11. Evaluate the following improper integrals (or prove their
divergence):
f
dx r dx f 3a:
2
+2
(a)
J
Tun; (b) (c) y^rdx.
J k
1
8.2.12. Compute the improper integral

i
In = § x m \n n xdx (n natural, m> — 1).
Solution. At n = the integral is evaluated directly:

i
l =^xm dx= m+ :
-
1 o m+ 1
"
For n > integrate I n by parts, putting
u = In" x\ do = xm dx\
X m+l
J
du=n\n n-i
n 1
x
x —
<lX
u = -
x
;
m +
We get
^
1
" m+ 1 o
XT Jf În^^dx-
m+ 1 m-f- 1
/„-,.
n 1
This gives a formula by means of which one can reduce /„ to / for

any natural n:
=— n
= n ^n — x
)
= __ (— 1 )" /z!
/
I I i
And finally,
"
w (m+l)» + 1

2.0
j _ C e~ x dx
~ J ^/ 2-\-x — x 2
0.3
accurate to 0.03.
Solution. The integral has a singularity at the point x=2, since
2 + x — = (22
x)(l+x). Let us represent it as the sum of two in-
tegrals:
2 -e 2
2 -e
Now compute the first integral to the required accuracy, and estimate
the second one. For eÔ.l we have
0</ < n5x ^'= (u53s

-
^lp==r -
§ 8.3. Geometric and Physical Applies of Improp. Integr's 409
Putting e = 0.1, we get the estimate / 2 < 0.028. Evaluation of the

integral
1.9
e~ x dx
V*-
by Simpson's formula with a step h = 0.8 gives
S . 8
= 0.519,
and with a step ft/2 = 0.4,
S . 4
= 0.513.
And so, integral /j gives the more accurate value, 0.513, with an
error not exceeding 0.001. Taking into consideration that integral 2
is positive, we round off the obtained value to
/ ^0.52
with an error not exceeding 0.03.
Note. By putting e = 0.01, we get the estimate / 2 < 0.005, but the
computation of the integral
1.99
f e-*dx
4
0.3
would involve much more cumbersome calculations.

8.2.14. Investigate the following integrals for convergence:
i i
dx /i x f dx
(a) (b)
f-£=;
JV sin* Jf
ex -
F=;
Ccos 2 xdx f tanxdx sin#dx
.
< c)
x
Jo^gr: (d)
j
.
(e)
,
j
f
_1_
2
^
$ 8.3. Geometric and Physical Applications of Improper
Integrals
l
8.3.1. Find the area of the figure bounded by the curve y--
l +x*
(the witch of Agnesi) and its asymptote.
Solution. The function y = is continuous throughout the en-
} _^ 2
tire number scale, and lim y = 0. Consequently, the x-axis is the asymp-
X-+ CO
tote of the given curve which is shown in Fig. 118. It is required

to find the area 5 of the figure that extends without bound along the
x-axis. In other words, it is required to evaluate the improper integ-

CO
dx
ral S= JC \ -t-t—
\ -\-
r
X"
. By virtue of the symmetry of the figure about the
— oo
y-ax\s we have
^=
CO GO
A
S=f -4^- = 2[ T 2 lim arctanx = 9Z'-tt^ = „
31
Jl.
8.3.2. Find the surface area generated by revolving about the

x-axis the arc of the curve y
Solution. The area of the sur-
y\ face is equal to the improper in-
1 tegral
i i
-7 / X
Fig. 118 Making the substitution e~ x = /,
—
= e~* dx, we get x^=0 at
t = 1 , x = oo at £ = 0; hence
i
S = 2ji [\fT^dt = 2n-±[tVT^ + \n(t + VT+T*)]\ =
= ji[J/'l+ln(l+j/~2)].
8.3.3. Compute the area enclosed by the loop of the folium of Des-
cartes
x 3 -f if — 3axy = 0.
Solution. The folium of Descartes is shown in Fig. 86. Let us re-

present the curve in polar coordinates:
x^pcoscp; y = psincp.
Then p 3 cos 3 cp -f-
3
sin 3 cp — 3a p 2
cos cp sin = 0, whence,
cp cancelling p 2 ^
we get
3a cos cp sin cp
^ ~~ cos 3 (p+ sin 3 cp

*
Since the loop of the curve corresponds to the variation of cp between

and y the sought-for area is equal to
jt
i r* o j ^
9a 2
n
r
2
sin 2
sin'cpcos'cp
cp cos
2
cp ,
§ 8.3. Geometric and Physical Applic's of Improp. Integr's 411
To evaluate the obtained proper integral make the substitution
tan cp = t\
CO gf
= dt\ cp = at t = 0, cp =y at t = oo. Thus we get
c 9a 2 f t
2
dt 9a 2 ,. C i
2
df 3a 2 ,. 1 1
<4
"
2
8.3.4. Find the volume of the solid generated by revolving the cis-
soid y 1
= 2a— a k° u * ^s as Y m ptote x=2a.
Solution. The cissoid is shown
in Fig. 119. Transfer the origin of
coordinates to the point 0' (2a, 0) without changing the direction of
the axes. In the new system of coordinates
X= — x 2a, Y =
y the equation of the cissoid
has the following form:
Y2
(X + 2a)*
-X
The volume of the solid of revolution about

the axis X = 0, i.e. about the asymptote, is
expressed by the integral
V = Ji$ X 2
dF = 2jcJX 2
dY.
Let us pass over to the variable X. For this

purpose we find dY= Y' dX. Differentiating the
equation of the cissoid in the new coordinates
Fig. 119
as an identity with respect to X, we get
3(X + 2a) 2 X — (X + 2a)3 2 (X~^~2a) 2 (X — a)
2VT' = -
X2 X 1
whence for Y > we have

(X + 2a)*(X-a) (X + 2a) (X-a) '
X2Y X }f-(X + 2a)/X

2
Hence,
V = —2n C(X± 2a) (X — a) :dX.

J V- (X + 2a)/X
Make the substitution (X + 2a)/X = — 2
; X = — 2a at t = 0, X=0at
/ = oo. Then:
2a \at 2at*
X 1+/ 2
dX:
+
:
(1 /
2 2
)
dt\ X + 2a- + 1 /
2
X-a = 3o + fl/
2
(
1 + /
2 5
whence
v
V ~ 9
f + at
2at 2 (3a 2
) 4at dt _
~
) t(\ + t*)(\ + t'*)(\ + t*)*
= 48 ™ 3
J (trw dt + 16 ™ 3
j (ttf¥ dt -
n
Putting t tanz, dt = = sec 2

zdz, we get ^ = at z = 0, ^=oo at
z = n/2. Hence,
JT/2 JT/2
V= A8na's [ sin 2 z cos 4 zdz + 16jta 3 ^ cos 2 z sin 4 z dz =

JT,/ 2 JT/2
= 48jia 3
^ cos 4 zdz — 48na 3
J
cos 6 zdz +
JT/2 JT/2
+ 16jta 3 ^ sin 4 zdz— 16jia 3 ^ sin 6 zdz.
JT/2
Using the known formulas for the integrals s\n n xdx,

J
JT/2
^cos^xdx (see Problem 6.6.9), we get
V = 64 jca 3 — • —— 64 ra 3 — .
1x3x5 -
2
8.3.5. Prove that the area of the region bounded by the curve
t/=-7=L=, the axis of abscissas, the axis of ordinates and the
V 1 —* 2
asymptote x — 1 is finite and equals -y
8.3.6. Prove that the area of the region bounded by the curve
y = y^=t » ê axis of abscissas and the straight lines x=±:l is
finite and equals 6, and the area of the region contained between
the curve # = ~2> the axis of abscissas and the straight lines x=±:l
is infinite.
8.3.7. Find the volumes of the solids enclosed by the surfaces

generated by revolving the lines y = e~ x , x = 0, y = (0 < -(- oo):
(a) about the x-axis,

(b) about the y-axis.
§ 8.3. Geometric and Physical Applic's of Improp. Integr's 413
8.3.8. Compute the area contained between the cissoid y 2 = 2a _ x

and its asymptote.
8.3.9. Compute the area bounded by the curve y = e~ 2x (at x > 0)

and the axes of coordinates.
8.3.10. Find the volume of the solid generated by revolving, about

the x-axis, the infinite branch of the curve y = 2^ — ^j for x> 1.
8.3.11. Let a mass m

be located at the origin and attract a
material point M
found on the x-axis at a distance x from and
having a mass of 1, with a force F= -£r (according to Newton's
law). Find the work performed by the force F as the point moves M
along the x-axls from x=r to infinity.
Solution. The work will be negative, since the direction of the
force is opposite to the direction of motion, hence
oo N
A=\ — ^xT dx=
2
lim \ ^rdx
2
= —— r
J n- oo J *
r t
During the reverse displacement of the point from infinity to M

the point x =
r the force of Newtonian attraction will perform posi-
tive work ^- . This quantity is called the potential of the force
under consideration at the point x=r and serves as the measure of

potential energy accumulated at a point.
8.3.12. In studying a decaying current resulting from a discharge

"ballistic" instruments are sometimes used whose readings are pro-
00
portional to the "integral current intensity" g= J

/ dt or the "inte-
00
gral square of current intensity" S=Î 2

dt and not to the instan-
2
taneous value of the current intensity / or to its square / . Here t
is time measured from the beginning of the discharge; / alterna- is
ting-current intensity depending on time. Theoretically, the process

continues indefinitely, though, practically, the current intensity be-
comes imperceptible already after a finite time interval. To simplify
the formulas we usually assume the time interval to be infinite in
all calculations involved.
Compute g and S for the following processes:
(a) I =/ e~ kt (a simple aperiodic process); k is a constant coeffi-
cient, which is greater than zero.
(b) / =I
e~ kt sin cot (simple oscillating process); coefficients k and a>
are constant.
Solution.
(a) g= \
l e~ k1 dt= lim f V"*' dt = / lin
oo
s J /;*-•*«<«= -'J-
= ;
oo A
(b) g= \ I ()
e~ kt s\n cot dt = lim ^ V"*' sin cot dt =
o
A - 00
n
= -^f7T2
+o)
2 2
!™ [(cocosorf
LV + ftsinwOe-"]^-^^;
or-f-&
7 J o 2 *
/2
^ _ n '
oo ^4
~ co * 2<ot
5- f sin 2
(o/d/= lim I
li e
-* kt l
^
dt =
J A - co J
=—A lim
co
2
I
+6 ;
2
(A cos 2cot + (o& sin 2a)t) 9 -2kt
46 /l - 00
4k (k* + 2
co )
8.3.13. Let an infinitely extended (in both directions) beam lying

on an by a concentrated force P. If the
elastic foundation be bent
x-axis is brought to coincidence with the initial position of the axis
of the beam (before the latter is bent) and the y-axis is drawn
through the point (at which the force is applied) and directed
downwards, then, on bending, the beam axis will have the follo-
wing equation
y = ^ e~ a 1x1
(cos ax + sin a x \
|),
where a and k are certain constants. Compute the potential energy

of elastic deformation by the formula
oc
W = Ee J (y"f dx (E, e const).
Solution. Find y"\
4
y"
Pol
= -£-e~ ax [(cos ax
+ sin ax) — 2 (— sin ax+cosax) +
-j-( — sin ax — cos ax)] = —-e~ ax (sin ax— cos ax).
Hence,
P 2 a 6 Ee
w k2 J
f
e~ 2ax (1 — 2 sin ax cos ocx) dx =
o
P 2 a Q Ee f 1 2a P 2 a*Ee
2a 4a + 4a 2
2
4fc
2
8.3.14. What work has to be performed to move a body of mass m

from the Earth's surface to infinity?
8.3.15. Determine the work which has to be done to bring an
electric charge e 2 =l from infinity to a unit distance from a
charge e1 .
dx
1 xP In* x
i
converges for p > 1 and q < 1.

00
converges absolutely for — l<(/?+l)/g<0 and converges condi-

tionally for 0<(p+ l)/</ < 1.
8.4.3. Prove that the Euler integral of the first kind (beta func-
tion)
i
B(p, q)=\ xP- j

1
(l — x)"- 1
dx
converges for p > and q > 0.

8.4.4. Prove that
lim
7
\ sinax-sinP^dxÔ,
7 -> a> J
if |a|=HP|.
8.4.5. Prove that
oo
I = ê- xi
-x* n + 1
dx=Y (n natural).
416 Ch. VI II. Improper Integrals
8.4.6. Prove that if the integral J-^— converges for any posi-
a
tive a and if f(x) tends to A as x—> 0, then the integral
00
r /(ax)-/(N
*
<fa .
(a>0( p>0)
J
converges and equals /4ln(P/a).

8.4.7. Prove that
00 cc
g" ax — cosfk
__ f cos
X—
f ^
J *
^x
J x
dx ~ \nn "cT
'
JT/2
8.4.8. At what values of m does the integral

I
——
J cos X
dx con-
verge?
ji
P dx
J
j^^k converges if k < 1, and
diverges if k^\.
00

sin *(*^~ CQS *)
<j x converges if
J
<s< 4, and converges absolutely if 1 < s < 4.
8.4.11. Suppose the integral
+ 00
lf(x)dx (1)
a
converges and the function y(x) is bounded.

Does the integral
lf(x)<p(x)dx (2)
necessarily converge?
What can be said about the convergence of integral (2), if integ-
ral (1) converges absolutely?
8.4.12. Prove the validity of the relation
/ (x) = 2/ (ji/4 + x/2) 2/ (n/4 jc/2) —x In 2
where f(x) =— ^ In cos y dy.

Compute with the aid of the relation obtained

JT
/ (-^) = —j In cos ydy.
8.4.13. Deduce the reduction formula for the integral
JT
In =^\ncosx-cos2nxdx (n natural)
and evaluate this integral.

ANSWERS AND HINTS
Chapter I
=—
2
p
1.1.5. (b) Hint. Prove by the rule of contraries, putting 2 , where/? and q
are positive integers without common multipliers.
1.1.8. Hint. You may take k = -
2s
*
1.1.9. (b) x^?4, *<0; (c) — 4<x<2.

1.1.11. (a) x < — 1 or x^sl. ////i/. The equality is valid for those values
x—
of x for which
1
^
^ 0; (b) 2^x<:3. //m/. The equality holds true for those
values of x for which x 2 — 5x-\-6^0.

2
1.1.13. (a) x < -g- or a: > 8; (b) x < or < x < 5. //m/. The inequality
\a — b\ > — 16| holds good when a and 6 are opposite in sign or when
\a\<\b\.
'• 2 - 3
' °=
[aV + ta +
a
W 3
<«•+«>«"-»>
^
1.2.4. + a&+a
&'- + 2
;
(
/t
)
_L 1-2 .6. 4 V'l+l;
4" '
; 2 ^10-5.
g
1.2.11. /(*) = 10 + 5x2*.
,,,,
1 = 5; / (0) = 4; / (0.5) = 4; / (-0.5) = -I— / (3) = 8.

.2. 14. / (2) ;
+ = follows that n + = a Xn+i = a Xn

+d
1.2.15. tfm/. From + ! = a Xn a a it i/
1 .
1.2.16. x=±2; ±3. 1.2.17. / (x) = x — 5x+6. 1.2.18. (*) = 23; (x) = 527. 1
/ 2. 11.2.20. P = 26 + 2

^
- S = &^1--^x. 1 a:;
1.2.21. (b) (2, 3); (c) (— oo, —1) and (2, oo); (d) x = ^ + 2kn(k = 0, ±1,
±2, ...). Hint. Since sin#^l, the function is defined only when sinx=l;
(g) (-^, 2) and (3, oo); (h) [1, 4); (i) (-2, 0) and (0, 1); (j)
-iL + 2tot<*<
< j+2kn(k = f ±1, ±2, ...).

Answers and Hints to Ch. I 419
1.2.22. (d) The function is defined over the entire number scale, except the
points x= ±2.
1.2.24. (a) (~oo, oo); (b) (3 — 2jt, 3 — ji) and (3, 4); (c) [—1, 3];
(d) (—1, 0) and (0, oo). 1.2.25. (b) 5<x<6.
1.2.26. (a) 2£ji<a:<(2£+1) ji (* = 0, ±1, ±2, ...); (b) — y, —1 .
*2
1.3.3. (b) Hint. Consider the difference B
^— -
\+x\ \+x\
1.3.4. (b) It increases for — —.-ffcri < x < ^ + kn (k = 9 ±1, ±2, ...)
and decreases on the other intervals.

JI
1.3.7. The function decreases on the interval < x^—- from 4-oo to 2 and
4
ji ji
increases on the interval < y f rorn 2 to +°°-
1.3.9. (c) The function
neither even, nor odd, (d) even. is
14
1.3.10. (a) Even; (b) odd; (c) odd; (d) neither even, nor odd; (e) even.
1.3.12. (a) \A |==5, co = 4, cp = 0, ^=y; (b) |

A |
= 4, (0 = 3, q> = -5-,
T= —
2ji
; (c) |/1 1 = 5, = —
= arc tan y r = //m/. 3 sin y + 4cos y =
(o ,
cp , 4ji.
a: a:
= 5sin ^y+cp^, where cos q^^, sincp = y. 1.3.13. (b) T = 2n\ (c) T= 1.
1.3.16. The greatest value /(1) = 2. The function reaches the greatest ////if.
value at the point where the quadratic trinomial 2x 2 — 4#+3 reaches the least
value.
1.3.17. (a) Even; (b) even; (c) odd; (d) even.
1.3.18. (a) T = n; (b) r = 6ji.
1.3.19. Hint, (a) Assume the contrary. Then
#+r+sin (x-\-T) = x-\-s'm x,

—T— \ = T
whence cos
(
f x-\
J
= , which is impossible for any constant 7\
1
^ 2sin "^-
since the left side is not constant; (b) suppose the contrary. Then
cos \^x+T = cos V x, whence either Vx+ T4- V~lc=2nk, or— 7= = = >Y
= 2nk (k = ±\ t ±2, ...), which is impossible, since the left-hand

t
members
of these equalities are functions of a continuous argument x.
1+arcslnj/
1.4.6. (a) x= ; (b)x = 3sin*/; (c) x = y lo 8 5
(y > 0); (d)* =
3
log 2 ^ lQ gj/
1 ° g2 ^ 1
==
.og|
(0 < y < 2 or 2 < y < 00).
1.6.3. (a) — — , ,
0, — ,
(b) —y ,
T>
- y, ,
(c) 2;2.25;2g; 2±{g
1.6.9. Hint. The inequality |^±^~2 < e is satisfied lor n > N =

420 Answers and Hints
= Ey- — lj. At 8 = 0.1 the inequality is fulfilled beginning with /z=10, at
e= 0.01 beginning with /z=100, at 8 0.001 beginning with /z= 1000. =

1.6.10. Hint. Verify that the sequence {x 2n -\} tends to as n oo, 1 —
and the sequence {x 2n \ tends to as n oo. —
1.6.12. (a) It has; (b) it does not have; (c) it has; (d) it does not have.
1.6.14. Hint. <a)|*J<-|; (b) \x n \<±.

1.6.19. Hint. For a > put y/~a= + a n
1 1 (a n > 0) and, with the aid of the
inequality a = (\-\-a n )
n
> na n prove that a n
,
is an infinitesimal. For a < 1 put
n /— =-— 1
—a =
1
n
V/ a (a n > 0) and make use of the inequality ([ -\- an ) > na n .
I -)- <x n
1.7.1. (b) A; (c) 0; (e) 1. 1.7.2. (b) 1; (e) 1; (f) 1.
1.7.4. (b) 1; (f) 0. Hint. Multiply and divide by imperfect of a sum, square
± —— 1
and then divide by n 3

; (g) ; (h) 1. Hint. Represent each summand
of xn in the form of the difference
1X2
1
J_
2' 2x3
111 2 3 '
'
1
/i(/i+l)~
11
n n -f
1'
which will bring xn to the form x n = —^ 1 .
1.7.5. (a)
y ;
(b) 1; (c) 0, (d) --i. Hint. The quantity — is an in-
4
finitesimal, and cos n 3 is a bounded quantity; (e) 0; (f)
y .
1.8.6. (b) Hint. The sequence is bounded due to the fact that n\=^\x2x
w-
X3x...X/?^2 and1
therefore
^H4)'+-.+(^n=3-(±)-'<,
1.8.7. (b) 0. Hint. Take advantage of the fact that 2«±!
xn
= n— o
< i.
-f-
1.8.9. //t/i/. For all /z, beginning with a certain value, the inequalities
—
i
< a < n are fulfilled; therefore —n J=- < ,"/ a < {!/ n,
*
and lim iV /i =
iy
i
= lim -
+
1.8.10. Hint. The sequence {y n \ decreases, since y n ±i =a 2
'l 1
= a 2flX2 =
= J^G/« > 0-
The boundedness of the sequence from below follows from a > 1. Denote
lim y n by & and from the relation y n + i= Vy n b=\.
fl -> oo
1.8.11. //m/. Ascertain that the sequence increases. Establish the bounded-
ness from the inequalities
n n(n—\) n — i n v 7
*<'+('-±)+(T-T)+-+(.-±r-7)-»-f
A n skits and /lints to Ch. 1 421
2n
1.8.12. Hint. Transform x n into the form xn = and take advan-
yn + 2
1 -j- n
tage of the inequalities
2n 2n
< < 1.
2i + 1
Y& + + n I
////i/. See Problem 1.8.7 (a).

1.8.13.
//m/. Establish the boundedness of
1.8.14. the sequence by comparing xn
with the sum of some geometric progression.
1.9.2. (b) Hint. Choose the sequences
and xn = - </i=l, 2, ...)
and ascertain that the sequences of appropriate values of the function have
different limits:
i
—=
Xn
x'
n
lim 2 + oo, lim 2 =0.
1.9.3. (e) Hint. Take advantage of the inequality
71
— arc tan x < tan ( — arc tan = (x > 0).
2 V 2
(f) Hint. Transform the difference
sin x —— = 1
sin x — sin . jx
-g-
into a product and apply the inequality |sina|^|a|.

1.10.1. (d) £-\ (e) ~; (f)—~. Hint. Multiply the numerator and de-
q o \Z
nominator by imperfect trinomial square (jj/ 10 — x-f-2); (g) ~~ ; (h) log fl 6.

'
— = log a (*-3)(^* + 6 + 3)
Hint, lim log a lim = log a 6;(i)
+ 6 — 3J x—3 *
at-v3L V x L*-3 3
<j)^.
1.10.2. (e) — . ////i/. On removing the irrationality to the denominator divide
the numerator and denominator by x.
1.10.3. (b) 32. (c) -5-. Hint. Put x=z 15 ; (f) oo. Hint. Put x =z;
as x ' (g) —3. //m/. Put s\nx = y.

2
i
1
1.10.5. (b) e
J
(c) *" (d) (f) 4; (g) (h) 2.
1.10.7. (b) 1 1.10.8. (b) 1; (d)

cot a
2
1.10.11. (a)
_1_
'
(c)
1
(cl) ' (e) 0; (f) - 1.

2 5
n
1.10.12. (a) j_ (b) -2; (c) y> (d) e) ~ 24
20
;
y ; (
-
2
1.10.13. (a) e 4 ; (b) -1; (c) 2 In a; (d) e 3 ; (e) e ; (f)e-J; (g) 1;
(h) 1; (i) 9; (j) 1; (k) a — p. Hint.
$X
lim
X-*
oiX
X
—= X
lim e?*i
~* Q
rj e (at -fax
-
X
1
L=a— 0.
1.10.14. (a) j/lT. Hint. Replace arc cos (1— x) by arc sin }/~2a;— x 2 (b) l;(c)a. \
1.11.5. (b) It is of the third order of smallness. Hint.
|.
m tan a— sin a 1
a- n a3 ~~T'
1.11.6. (b) They are of the same order; (c) they are equivalent.
1.11.8. (a) 100* is an infinitesimal of the same order as x\ (b) x 2 is an
infinitesimal ot an order higher than x\ (c) 6 sin a: is an infinitesimal of the
same order as x\ (d) sin 3 x is an infinitesimal of an order higher than x\
(e) \/ tan 2 x is an infinitesimal of an order of smallness lower than x.

1.11.9. (a) It is of the fourth order of smallness; (b) of the first order of
smallness; (c) of the third order of smallness; (d) of the third order of
smallness; (e) of the first order of smallness; (f) of the order of smallness
A ; (g) of the first order of smallness; (h) of the first order of smallness;
(i) of the second order oi smallness. Hint. Multiply and divide the difference
cos x—
jj/cos x by imperfect trinomial square; (j) of the first order of smallness.
The diagonal d is of the first order of smallness; the area S is of
1.11.10.
the second order of smallness; the volume V is of the third order of smallness.
1.12.3. (b) 4; (f) 3; (g) ~; (i) 2. 1.12.6. (a) 1; (b) 2.
1.12.7. (a) 1; (b) A . 1.12.8. (a) A ; (b) A ;

(c) A ;
(d) A ; (e) A ;
(f) A ; (g) _2; (h) 1. 1.12.9. 10.14. Hint. 1042= 10 3 x (1+0.042).

1.13.1. (b) Ml— 0) = — 2, /(l+0) = 2; (f) /(2-0) = -oo;
/(2 + 0)=+co.
1.13.3. (a) /(-0) = A f /(
_ .
f ) = 0; (b) /(-0) = 0, /( + 0)=+oo;
(c) /(— 0) 1. /( + 0)=1.
1.14.2. (b) The function has a discontinuity of the first kind at the point
x = 3. The jump is equal to 27.
1.14.3. (c) The function is continuous everywhere; (e) the function has
a discontinuity of the first kind at the point # = 0; the jump equals ji. Hint.
arc tan (— oo)= — ^- , arc tan (+ oo) = -]--^-
1.14.6. (b) At =5 the point # there is a discontinuity of the first kind:
/(5_0)= —y JI
+ =,
— /(5
(c) 0)
JX
;
at the point a: = 0, a discontinuity of the
JT
first kind: /(— 0)=1, / ~H 0) = 0; ( (d) at the point ^ =y , an infinite dis-
continuity of the second kind:
1.14.7. (a) At the point x~0 there is a removable discontinuity. To remove

the discontinuity it is sufficient to redefine the function, putting / (0)= 1; (b) at
Answers and Hints to Ch. 1 423
the point x = removable discontinuity. To remove the discontinuity

there is a
it is sufficient the function putting /(0)=1;
to extend (c) at the point # =
there is a discontinuity of the second kind: /(— 0) 0, / ( -|-0) -j- oo; (d) at = =
the points x — (2k hO-j (^ = 0« ±1» ±2, • ••). removable discontinuities, since
f
if I sin x | < 1,
flx)=UmJs\nx)*»=\ { jf | s in^ |
= 1;
(e) at the points x = kn (k~ 0. ±1, ±2, ...), discontinuities of the 1st kind,
since
- . .
_ |
sin x |
j 1 if sin x > 0,
sin a: ) —1 if sin a: < 0;
(f) at the points x = n — 0, ±1, ±2, ... removable discontinuities, since
(
. .
I
— 1 if x=n t
1.14.8. (a) At the point x=

is an infinite discontinuity of the second 1 there
kind; (b) at discontinuity of the first kind (the jump
the point x —— 2, a
being equal to 2); (c) at the point # 0, an infinite discontinuity of the second =
kind, at the point x=\ a discontinuity of the first kind (the jump being equal
t
to —4); (d) at the point x—\, an infinite discontinuity of the second kind.
1.14.9. (a) /(0)=1; (b) /(0) = --|; (c) /(0)=^; (d)/(0) = 2.

1.15.2. (b) The function continuous on the interval (0, +oo).
is
1.15.3. (b) The function continuous everywhere. At the only possible point
is
of discontinuity x = we have
lim y—\\mu =\\ 1
lim y = lim u 2 =l\ y \ x_ =y | B= _ t = 1
(c) at the points x = + (n = 0, ±1, ±2, ...) there are removable dis-
continuities, since \imy= lim y —— 1.

Jl U-*±cc
X ~>
2
Yes.
1.16.2. 1.16.12. 1.53. 1.16.13. No. For instance, the function y =x 2
on the interval [
— 1, 1].
1.17.1. (a) Hint. Multiply the obvious inequalities:
n-
n < '
2
n -4- 1
V2(n-\) <
/(S=T^<=±1;
2n —
rr*
T34 X T56 X...X —>2n
1 1
.f. t i a
b) Hint.
j
Let ^-o-X
2
.
D 2 4 6 2n
'
3 5 7 '
' 2n f- 1
Then A < B since — —< —

2/7—12/t
2n n
2n-\-\
—. and A'1 < AB^-- —
r
2n -f
i
-
r .
1.17.2. (a) Hint. Extract the 101st root from both sides of the inequality
2
and reduce both sides by 1 1 .
(b) Multiply the obvious inequalities:
99 X 101 < 100 2 ,
98 X 102 < 100 2 ,
2 x 198 < 100 2 ,
1 X 100 X 199 X 200 < 100 4 .
1.17.3. (a) —3 < a < — 1 or 1 < x < 3; (b) x < —— 1
or x >
5
y ; (c) the
inequality has no solutions, since it is equivalent to the contradictory system

x— 2 > 0, a:(4a: 2 *+4) < 0. —
1.17.4. Yes. 1.17.5. (a) No; (b) Yes.
1.17.7. Hint. Apply the method of mathematical induction. At /i=l the
relation is obvious. Supposing that the inequality
(l+*i)(l+* 2 ) ... (l+*„-i)^l+*iH-*a+---+*«-i

holds true, multiply both its sides by \-\-x n and take into consideration the
conditions \-\-x n >0, Xi>x n > (i=l, 2, n— 1).
1.17.8. (a) [1, +oo); (b) (2/zjt) 2 < x< (2/i+ 1)
2
jx
2
(/i = 0, 1,2, ...);
(c) a:=0, ±1, ±2, (d) (— oo, 0) for f (x); g (x) is nowhere defined;
(e) [-4, -2] or [2, 4]; (f) *=(2/i+l)y (/i = 0, ±1, ±2, ...).
1.17.9. (a) No: cp (0) = and / (0) is not defined;

1 , (b) No: f (x) is defined
for all a: 7= 0, and cp (x) only for x > 0; (c) No: f (x) is defined for all x, and
cp (a:) only for x^0; (d) Yes; (e) No: f (x) is defined only for x > 2, and cp (*)
for x 2 and for x < 1.
>
1.17.10. (a) (0, oo); (b) [1, oo). 1.17.11. V 8xt (x 3) (6 x), 3 < x < 6. = — i
—
1.17.12. (a) x =
5. ////i/. The domain of definition is specified by the inequa-
lities x-f-2^0, x —
5^0, 5 xÔ, which are fulfilled only at the point —
# = 5. Verify that the number x 5 satisfies the given inequality, (b) Hint. =
The domain of definition is specified by the contradictory inequalities x 3 > 0; —
2 — x > 0.
nX _J_ n — X r, — X
1.17.17. (a) f (x )
9
7l
y
= (b) a*=
T
± fiX
"
+ T±1?
(see Pro- ; + /
blem 1.17.16).
1.17.18, An even extension defines the function
, , / f(x) =x 2
-{-x for 0< a:<3,
x for —3 < x < 0.
An odd extension defines the function
I f(x)=x 2 + x for 0<*<3,
1.17.21. If the function

////if. / (x) has a period 7\, and the function cp (a;)
has a period and T l n 1 d, T 2
7\>, =
n 2 d (n lt n 2 positive integers), then the pe- ~
riod of the sum and the product of these functions will be T nd, where n is =
the least common multiple ot the numbers n x and n 2 .
1.17.22. Hint. For any rational number r
i / * * , ^ , i for rational x,
>=m*)={
.
'
n
; for irrational x.
But there is no least number in the set of positive rational numbers.

1.17.23. Hint. II wc denote the period of the function f (x) by T, then from
f(T) = f{0) = f(-T) we get
sin T-\~ cos aT = 1 = sin — 7)-]- cos — aT),
( (
whence sin 7 = 0, cos^r^l, and hence T = kn y

aT — 2nn i
a ——
2n
is rational.
k
1.17.25. The difference of two increasing functions is not necessarily a mono-
tonic function. For example, the functions f(x) x and g(x) x 1 increase for = =
xÔ, but their difference / (*)— g (x) ==x x 2 is not monotonic for xÔ: it —
increases on 0,
^
and decreases on
1.17.26. Example:
a: if a: is rational,
-{- x if x is irrational.
x = -j\n 1+0 (-1 < <

1.17.27. (a) y 1);
J -y
(b)
^
y for — oo < y < 1,
x=< V~y lor 16,
I log 2 y for 16 < y < oo.
1.17.28. Hint. The functions y=x 2 -\-_

+2x+l and y
(x 1) = — X+V*
(xÔ) mutually inverse, but the
are
equation y x, i. e. x
2
=
2x-f- x has + = 1
no real roots (see Problem 1.4.4).

1.17.30. (c) Hint If £ is the domain .
of definition of the function / (x),

then the function y = f[f(x)\ is defi-
ned only those for
for which x£E
f(x)£E. How
the points of the desired
graph are plotted is shown in Fig. 120.
1.17.32. Hint. The quantity T 2 (b = a) a period: from the conditions is of
symmetry f = f (a x) and f(b + x) = f(b — x) follows that
(a -\-x) it
/ [x + 2 (b-a)] = f[b + (b + x—2a)]=f(2a-x) = f[a + {a--x)] = f(x).

1.17.33. (a) It diverges; (b) it may either converge or diverge. Examples:
yn
--
L
+ (-I)»]. lim (x n y n ) = 0,
£ n -* oo
xn = —n yn =n z
lim (x n y n ) = oo.
n -+• oo
1.17.34. (a) No. Example: xn yn = —n-{-\; (b) No.

n{n ~ 2)
1.17.35. a„ = (n
x
= 3, 4, ...). 1.17.36. Hint. Take into account
that
1
1
xn |
— |
a 1
1 ^ |
xn —a |. The converse is incorrect. Example: = 1)" + 1
.
1.17.38. Hint. The sequence a„ only the following values: may attain
0,1, 9. If this sequence turnedout to be monotonic, then the irrational
number would be represented by a periodic decimal fraction.
1.17.39. Hint. If the sequence ~
On
increases, then
i. e 6„ +l a (
-
< an + l
bj (/= 1, 2, «)>
whence it follows that

t>
n + i(<*i + a +...+a n
2 ) < a n + l (b 1 + b +...+bn
2 ) t
and hence
"1 + ^2+ + + •• 0/1 1 ^1+ 02+ "-+fl/l ==

&1 + &2 + ... + ^ w + l &1+ &2+...+&*
== W + ft +
g ^ + + 1 8 + + + + fl2
>
(bi + b +...+b n + )(b l + b +...+b n '
i 1 i )
1.17.40. (a) 2; (b) 0; (c) 0. 1.17.41. Hint. From the inequalities

nX ^
nx — 1 < E (nx)
v
'
^ nx it follows that x — 1 < x
n
< ^
^
n
^ x.
1.17.42. Hint. From the inequalities

n n n
2 ix 2 £ ^2
^=1 /z=i fe=i
it follows that
/i+l 1
2/i n nl 2n
1.17.43. Hint. Take advantage of the fact that lim a

n
= lim yfa = 1 (see
« -* go n cc
1
Problem 1.6.19), lim a " = =1, and for a > 1, |

/i |
< —n the ine-
rt * 00
lim V a
n -> co
1_ 1
qualities a
rt
— 1 < ah — 1 < a
n — 1 take place.
1.17.45. Hint. Divide the numerator and denominator by x m .
1.17.46. (a) a=l; b = —l; (b) a=l; b=-^ . Hint. To find the coefficient
a divide the expression by x and pass over to the limit.

1.17.47. (a)
for 0<x<
f
/ W = x
M-{ <
1
for x > 1.
1,
(b)
for x ?= — +/2JI,
JX
/(*)H ^ (/i = 0, ± 1, ±2, ...).

1 for x = — -{-nji
1.17.48. Hint. Take advantage of the identity
(1 — x)(\+x) (1+x 2
)...(1+a: 2 ") = 1— x* n .
1.17.49. Generally speaking, one can't. For example,
lim
ln(l + *) + ln(l-*) = ]ijn
ln_0- f 2)
=_ 1
X -> X2 X -* X2
— x) —x
~ X =Q. +
and if we replace In ( 1 a:) by x and In (1 by we will get the wrong
X
result: lim
1.17.50. — . Hint. If a is a central angle subtended by the arc under con-
sideration, then the chord is equal to 2R sin ~ Ra, and the sagitta to
a2
R(\— cosa) ~ R-.
1.17.51. 2. Hint. The difference of the perimeters of a circumscribed and
inscribed regular /z-gons is equal to
on
2tf/z
/.
tan
jt
n
sin
.
—n = 2nR..tana —
jr \
a
sin a n
~ tiRcc .,
2
,
\ J
where a — —,
n
and the side of an inscribed rc-gon is
2Rs\n— = 2R sina~ 2Ra.

n
1.17.52. On the equivalence of (1-j-a) 3 — 1 and 3a as a —>0.

1.17.53. No, logd+^l!^-^ as*-+0.
1.17.54. (a) Yes. Hint.
If the function cp (x) / (x) -\-g (x) is continuous at =
the point x =x
then the function g (x) cp (x), f (x) is also continuous at this = —
point; (b) No. Example: / (x) —g(x) sign x (see Problem 1.5.11 (p)); both = =
functions are discontinuous at the point x 0, and their sum is identically —
equal to zero, and is, hence, continuous.
1.17.55. (a) No. Example: / (x) =x is continuous everywhere, and g(x)~ sin-^-
for Xf:0, g(0) = being discontinuous at the point # = 0. The product of these
functions is a function continuous at #= since lim x sin — = 0; (b) No. Exam-

x— o x
°r
pie:
r /(#) = — g (x) = <
—1 , r
X''
^ *
both functions are discontinuous at the
for x < 0;
point x = 0, their product
J
f (x) =— 1 being continuous everywhere.

m ^ is rational,
t 1-r ec
1.17.56. No. i-
Example:
K
i f /
f(x)
v
\
'
=< — I
irrational.
1 if
if
.
is
. We may 3
.,
write
\ 1 a:
f (x) 2\ (x) = 1, where X (x) is — 1.14.4 (b)). the Dirichlet function (see Problem
1.17.57. (a) # is a discontinuity =
of the second kind, \ is a disconti- x—
nuity of the first kind; (b) is a discontinuity of the first kind: f(\ 0) 0, x=\ — =
/ ( 1 H— 0) = 1 ;
(c) cp (x) is discontinuous at all points except x = 0.
1.17.58. (a) x = n = 0, ±1, ±2, ... are discontinuities of the first kind:
lim y=\, lim y=y \
x==n = 0. The function has a period of 1; (b) x= ± n
x-+ n-0 x-> n +
(n—±\ t ±2, ...) are points of discontinuity of the first kind:
lim y = 2n — 1; lim y = y\ X ~_ y— = 2n
x -* Vn-0 x Vn-rO
The function is even; (c) (/i=±1, ±2, ...) are the points of x=±V'n
discontinuity of the first kind; at these points the function passes over from the
value to 1 —
and returns to 1. The function is even;
1
(d)
x if |
sin x |
< y , i.e. —^-\-nn < x < -^-f-ji/z,
y=] y
-75- if |
1
sin
s a:
I
1
= ~2 ,
•
i.e. x =± 111 ^--f-jx/z,
6~
if
I
sin x I
> — , i.e. —+ J1/1 < x < — |-jw.
x=± — -\-nn JI
are discontinuities of the first kind.
1.17.59. The function f[g{x)\ has discontinuities of the first kind at the
points x =—
1; 0; +1. The function g[f(x)] is continuous everywhere. Hint.
The function f (u) is discontinuous at u 0, and the function g (x) changes sign =
at the points # =
0, ±1. The function g[/(#)|==0, since f (x) attains only the
values 0, ±1.
1.17.61. Hint. Write the function in the form
x+ 1 for — 2<x < 0,
f(x) = { ° for
2
* = '
x
(*+l)2 tor < x<2.
Make sure that the function increases from — 1 to 1 on the interval f — 2, 0)
3
and from to — on the interval [0, 2]. Apply the intermediate value theorem
to the intervals — — [ 2, 1] and [0, 2]. The function is discontinuous at the point
* = 0:
/(-0)=1, /(+0) = 0.
1.17.62. Hint. Suppose e > is given and the point x £[a, b\ is chosen. We
may consider that
e<min[/(x )-/(a), /(&)-/ (*„)].
Choose the points x x and x 2 x 1 < x < ,

x 2 so that
/(*i) = /(*o) — c, / (* 2 ) = /(*o) + e.
and put 6 = min(^ — *i> *2 *o)- —

1.17.63. Hint. Apply the intermediate value theorem to the function
g(x) = f(x) — x.
1.17.64. Hint. Apply the intermediate value theorem to the function f (x) on
the interval [x lt x n ], noting that
min [/ to), . . . , / (x n )] <1 [/ to) + / to) +...+/ to)] < max [/ to), • . , / (x„)].
1.17.65. Hint. Apply the intermediate value theorem to the function g (x) =
= 2X —~ on the interval lj .
At sufficiently large values of the independent variable the

1.17.66. Hint.
values of the polynomial of an even degree have the same sign as the coefficient
at the superior power of x; therefore the polynomial changes sign at least twice.
1.17.67. Hint. The inverse function
Y—y—\ for y <-l,

for y = 0,
Vy—\ for y > 1
is continuous in the intervals (—00, —1) and (1, 00) and has one isolated point
y = 0.
Answers and Hints to Ch. II 429
Chapter II
2.1.1. (b) -—
20
. 2.1.2. (b) 10a; — 2. 2.1.5. v av = 25 m/sec.
2
2.1.6. (a) y' = 3x 2
\ (b) y' =— . 2.1.7. The function is non-differentiable at
2
= — -j
—3
-—
3 =
the indicated points. 2.2.1. (b) y' ax • 2.2.2. (c) y'
= 2*arctan*+l. 2.2.3. (b) —9000. 2.2.4. (a) i/' = 6a:

2
-f- 3; (b) */' ="2^ -
1
«._ , ^ 7y+^+ 2 + 8^+ 2^./x

w ,__ 3^6(*-2
~2x
+r : (0
(
' D2
;
(d)
*
;
l^x '
~
e'»'
a;
= T5r
(cos a:— sin a;)— sin
;
a:—?*
(0 *' = 2e*+y; (g) *' = 2e*cos*; (h) j,'
x2e x
2.2.5. (f) 301nMtan3*)^; (g)s\ny^= —^

l
2
2(1— a:)
„ ~ „
2.2.6. (b) */'
/
= -3(3-sinA:) o /o x9
2
cosA:; (c)
/
— 3^==+—5—; 2 cos a: . 2 sin
cos
a:
y
3 sin x sin x 2 *
u'———2e* + 5 In 4
2* In 2 *
(d)
K)
. a;
;
.
(e)
.
,
*/' = 3 cos 3a;
1
=- sin
.
+
5^ ,
3 3/(2e*-2* + l)
a * 5
-j —
2M
l
sec'2 VI; (f) ^' = (2a: — 5)cos(a: — 5A:-H)-Asec — 2
a:
2
a:
; ( h) =
= 1
H 7
h-ri 5;
tanx^\+x* — (i) (/' = 21narc tan-^-3 •
•
9+^
tti
—
x/l + ln** arc
arc tan
3
2.2.8. (b) »' =- sinh , \ taax)

^ 2
*+ cosh 2 cosec* x;
{^ (d) </' = 3*x
X(Jtsinh2x» + coshjt*.sinh2x*); (e) y' = e sinh« e ta<acosh«+6>_
/
I X 2 1 2a; \
2.2.9. (c) j/^y-p^sinSArcos 2 ^^ ;
(x+\)
(d) y' = (tan*) 2

— lntanA:-[ sin^ 2a:
2.2.13. (a) /' (*) =y (coshy + sinh-0 ; (b) /' (x) = tanhx; (c) /' (x) =
j/cosh *+l; = -Jj— (d) /'(x) ;
(e) /' (a;) = 4 sinh 4a;; (i) f
f
(x) = (a+b)e ax X
X (cosh fo + sinh ^) = (a + e + b)x fr)
(a .
2.2.14. (a) j/' = (cos

cos a:)
s,y
51
"* (cos x In cos a;— tan x sin *);
cos 3a:
(b) y'=-
— sin
j/ sin 2
—
4
3a: (1 3a:)
5a:
2 -}-
x — 24
(c) r-
j_ jl
2 3 2
3(x-l) U+2) (*-j-3)
2.2.17. (a) y
/
=—r.
^81*—
In 3
1
tan ârc sin3" 2 *
Jârcsin3- 2 *
— ;
...
(b) y
,
=
5x j/cos 4
c /
In 3
— x
. r ,
sin In 3
cos 2 In 3 *)
#-ln x
v
2
o r
— 5/ cos
•
j/ arc tan In3 ^2

2.3.1. (b) k n e**\ (e) 2"" 1 sin f 2x + n ^) ;
(f) -1- sin (x +ny + )
+ y sin ^3* + n-y)+^sin ^5x + /iy

ax 71
2.3.4. (b) e* (a:
2
+ 48a: + 551); (c) e j sin $x jâ"- ^~ 1)
a«- 2 p 2 + ... | +
T
--"- "ix'.'gr --^-]}-
. n (n — IH/i-^8 1 \
+ COS p*
2* 2
+ 3' 1 +^» are sinx
2.3.6. (a)
(1+a: 2 ) )^1 +a: 2
(b)
'
A
3 +7r \,;
'(1— 2
)
2
(c)2r-"x
2
(1— a:
2
)
X (2x 2 -l).
2.3.8. (a) * 3 sin x— 60 x 2
cos x— 1140 X sin x-j-8640 cos x\ (b) 2e~* x
X (sin x + cos x)\ (c) [3* 2 + 6/™: + 3rt (n— 1) — 4]; (d) (— 1)" [(4/z
2
+ 2/z +
+ 1— a:
2
) cos a: — 4n^sinA:].
lx3x5x...X l97x(399 — x)
2.3.9. (a) .00! [_^__^_] :
b
201
2
2ioo (l—^)
1_ _1_
Hint. y = 2 (1— x) 2
— — x) (1
2
.
~ s -
2.4.1. (b) Xyy
/, "
= —7^-+—sin 4 cos
i
:
a:
rr-
3
(6 a:)
2.4.3. (b) y'

x = -cot^f; (d) *,; = - 2£>- 2 ''.
2.4.4. (b) ^=_*L_ ; ( C)

1^=57^7 •
2.4.5. (b) y'£x = — 3 sin/ sec 2

/.
=JL+ e "T- £=- jA.

2.4.6.
2 4 7
(b)
(b)
y'
x
tr
(e x -ey )V-* x+y )
(c) ifi=^=|;
.
(c)
„~
(d)
_ ^ _
«.» « JtjJ, W £J.

2.4,,. ,.,
^2_|_i ^2 (a cos/ — 6 sin /) cos 3 y

{c)
~W l (d) ~~
2t (2/ 2 + 2/+l)
1
; (e)
4 sin —
;
(
f
)
- Yt=F •
~ ^l^- fe)
2.5.1. (b) 6a: + — 9 = 0;

2*/— 6y+ 37 = 2a: 0.
Answers and Hints to Ch. II 431
2.5.3. (b) cp = arctan2|^2. 2.5.8. (b) * + — 2 = 0; y = x.

2.5.15. (a)
-J;
(b) y=l, * + 2</-2 = 0; (c)
0+^= —3 (*+|- J
(d) 2.5.16. 11.
-J.
2.5.17. 26,450. 2.5.19. s = at-^; v = a-gt; s max = s[ _ q

~ T
2.5.20. y = r;=^^sinM(l+2ecosM). 2.6.3. Ay « dy = 0.05.
2.6.5. (b) log 10.21 » 1.009; (d) cot45°10' « 0.9942.
2.6.7. (c) = cos* A x (d) A y = + tan ( 1
2
a:) A*.
I I
;
~ 4 ~ ln *
2.6.9. (a) d*i/ = 4-* 21n4(2A: ln4— l)dA:
2 2 2
;
(b) d2 ^ 4 ln *
a:
2
y (In — 4)
2
a:
*
(c) d 3y = —4 s'm2xdx :i
.
oaift (a)d.2
2.6.10. y
M = 4(1+3a: 4 ),
dx 2*'>
.
.
2
—(TZ^r ^ - f="p ^
4(1+3a: 4 ), 2
4a: ,
2
^^4)2 1
4
in particular at # = tan /, d2 y= ~ CQS 2 2/
4
2.6.11. AV = 4nr 2
Ar-\-4nr Ar 2
+—o jxAr 3
is the volume contained between
two spheres of radii r and r Ar; dV 4nr 2 Ar is the volume of a thin layer+ =
with a base area equal to the sphere's surface area 4:xr 2 and a height Ar.
2.6.12. = g/A/-|—i- g At
As the distance covered by a body within the 2
is
time At; ds=gt At = vdt the distance covered by a body which would move at
is
a velocity v = gt during the entire interval of time.
2.7.1. (a) It does not exist; (b) it exists and equals zero.
2.7.2. 90°. Hint. Since
(e*, x^0
y ~ \ e-* x < % 0,
/1(0) = -I, /;(0)=1.
2.7.3. r_(A)=-<p(fl); /;
2.7.4. Hint. For x ^ the derivative
n*) = -cos(l)+2xsin(l
At a: = the derivative equals zero:
a
Aa:- sin
9
— 1
/'(0)= lim
XT-^^ '
Thus, the derivative /' (a:) exists for all x, but has a discontinuity of the
second kind at the point x = 0.
2.7.5. a 2x b = x\. ,
=—
2.7.7. Hint. The formula for the sum of a geo-
metric progression represents an identity with respect to x. Equating the deri-
vatives of both sides of the identity, we get
"'" +1
1 2r 3* + + 1
1
+
; . . . +^- = ^+;>'" +
multiplying both sides of this equality by x and differentiating again, we get
\+x-{ n+ \)* X° + Un^+*i-\) X»+>-nx>> +

l» + 2»*+...+i»V-= ^
2.7.8. sin x + 3 sin 3x + . . .
+ (2/i — 1) sin (2«— 1) * —
_ (2/i-|- 1) sin (2/i — 1 ) jc— (2/i — 1 ) sin (2/1+1)*
4 sin 2 x
//m/. To prove the identity multiply its left side by 2 sin* and apply the
formula 2 sin a sin P cos (a P)= cos(a p). — —
To deduce the desired formula +
differentiate both sides of the identity and equate the derivatives.
2
2.7.9. (a) sin 2x [/'(sin *) 2
/' (cos *)]; (b) <*> [e x —
f (e x ) /'(*)/ (e x )\; +
ln
(C ) ^11?} .
1
<P' (*) . ^ (*)
In 2
"
* ' cp (*)
\|) (#) ln cp (a:) (p (x)
2.7.10. (a) No; (b) No; (c) Yes; (d) No.

2.7.11. Hint. Differentiate the identity /(—*) /(*) or f(— x ) —f(x). = =
This fact is easily illustrated geometrically if we take into consideration that
the graph of the even function is symmetrical about the y-axis, and the graph
of the odd function about the origin.
2.7.12. Hint. Differentiate the identity / (*+ T) f (x). =
2.7.13. F'(x) 6x 2 = 2.7.14. y' 2\x\.
. 2.7.15. The composite function =
/ [cp (x)] may be non-differentiable only at points where cp' (x) does not exist
and where cp (#) attains such values of q>(x) u at which f (u) does not exist. =
But the function y u2 = =
\x\ 2 has a derivative y' Q at the point #=0, =
though at this point the function u \x\ has no derivative. =
2.7.16. (a) y" = 6\x\; (b) 2 sin ——
XX
- cos-!
X
j sin i- at
x
x £ 0,
y" (0) does not exist, since y' (x) is discontinuous at x = 0.

2.7.17. Hint, (a) Verify that = C£ (k = 0, 1, .... /i) and take advan-
tage of the property of the binomial coefficients. (b) Designate: f(x) un =
show that u'
n
= (n— 1) u n -^ —
w„_ 2 and use the method of mathematical
\
induction.
2.7.18. Hint. Apply the Leibniz formula for the /ith derivative of the pro-
X
duct of the functions u =e a

and v =x 2
.
I at n = 2k
2.7.19. yM(0)={ [1x3x...X(2/j— l)]
2
at n = 2k+\
I (k=\, 2, ...).
Hint. Differentiate the identity n —2 times and, putting # = 0, obtain
y(n)
(0) = — 2) (/i
2
\fn -2 >
(0) (/i ^ 2)
2.7.21. Hint. Take advantage of the definition
e-*
2
Hn + 1
(x) = (e-x 2 )<n
+ l
) = (— 2xe-**Y n >
and the Leibniz formula for the nth derivative of the product u~e~ x2 and
V =-2 X yx
. 2.7.22.
=j^jy
2.7.23. x 12 = ±V + Y\- y \ (_oo <{/<l),
*3,4=±Kl-l/7=^ (0<y<l),
(t=l, 2, 3, 4) for Xi ^ 0, ±1.
4x/(l-jtf)
Answers and Hints to Ch. Ill 433
Hint. Solve the biquadratic equation x* — 2x 2

-{-y — and find the domains
of definition of the obtained functions x ( (y).
2.7.25. (a) x x 3; * 2 =1; =—
(b) ±\. x=
2.7.26. Hint. Note that the function x=-2t — \
t |
==
^ /^q' haS n ° deri "
<J
vative at / = 0. But / = <0 therefore we can ex P ress y ~t 2 + * V 1

=
^ x
= f 2/ 2 / > [ 2x
y=< „
2
< ^
< ^ '
through x: y ' '
This function is differentiate eve-
|
0, t * j 0, x < 0.
rywhere. 2.7.27. a =c= -!- ; 2.7.28. Hint. The curves intersect at the
4 2
points where sinaA:=l. Since at these points cosa#=0,
y'
2
= f' (x) sin ax-\-f (x)acosax = f (x) = y'v
i.e. the curves are tangent.
2.7.30. Hint. For / y= nn the equations of the tangent and the normal are
reduced to the form:
y = co\^-{x — at)-\-2a\ y — — tan y (x — at),

respectively. For l =
n(2k— 1) (k=\, 2, ...) the tangent line (y 2a) touches =
the circle at the highest point, and the normal (x at) passes through the high- —
est and lowest points; for t 2kn (k 0, 1, .) the tangent line (x =at) pas- = . .
—
ses through both points, and the normal (y 0) touches the circle at the lowest =
point.
v 2.7.34. ^-
dt
+ y.y
l 1
2.7.35. The relative error 6 = ^^ I sin 2cp
. The most
reliable result, i.e. the result with the least relative error, corresponds to the
value cp 45°. =
Chapter III
3.1.2. (b) Yes; (c) No, since the derivative is non-existent at the point 0.
3.1.5. l = e—\. 3.1.7. No, since g(— 3) =g(3). 3.1.9. (d) Hint. Consider
Ihe functions
2x
f(x) = arc sin ^
-j-2 arc tan x for |
a: |
> 1,
2x
g(x) = arc sin y-j— — 2 arc tan x for |
x |
< 1
3.1.15. (a) £ = y; (b) a-^; (c) £ = 10 ; (d) it is not appli-
cable, since the function has no derivative at the point # = 0.

3.1.16. 1.26 < ln(t+£?) < 1.37. Hint. Write the Lagrange formula for the
function f(x)~\nx on the interval [e, e-\-\] and estimate the right-hand side
in the obtained relation: In (1 -\-c) = 1 (e < £ < e-\- 1).
3.1.17. Hint. Apply the Lagrange formula to the function f (x) = \n x on the
interval [1, l+*]> x > 0, and estimate the right-hand side in the obtained re-
lation ln(l+x) =y (1 < £ < 3.2.1. (c) 2; (d) 0; (f) —1
3.2.3. (b) 0. Hint. Represent cot x -—= — JI!iLf
; ( c) JL 3.2.5. (b) gi = e.
3.2.6. (a) 1; (b) 1. 3.2.9. (a) -i I (b) lna-1; (c) 2; (d) ; (e) - ;
/ b a
rr.2
n
(f) 0; (g) 1; (h) In a; (i) e

2
; (j) — ;
(k) -1; (1) e; (m) 4":
(n)i-; (o)
^; (p) e~™; (q) 1; (r) -1 . 3.3.5. (b) 0.34201.
3.3.6. V 83 ^ 3.018350. tfm/. ^83 = V 81+2 =3 (l+gy) *

- Apply the bi-
'
nomial formula and retain four terms.

3.3.7. Hints, (b) Write the Maclaurin formula for the function /(A:) tan* =
with the remainder R 4 (x); (c) write the Maclaurin formula for the function
j_
/ (#) =(1 +*)
2
with remainders R 2 (x) and R 3 (x).
3.4.2. (a) / (x) = 1 ^ _i. x s
_^ x + , (jcB); (b) /w = x _^ + f!_fl +
+ g+o(* 5
).
3.4.3. (b) —i; (c) -1; (d) 1; (e) 1.
3.4.4. (a) 1+2,+, 2 -4,3_|, 4 __L, 5; (b)-^-g+g; (c)

1-f +
+ 12 720*
3.5.1. (d) The function decreases on the interval (— oo, 0) and increases on
(0, oo); (e) the function increases on the intervals ^— oo, y^j
and (3,-j-oo)
and decreases on (^-^ , 3 (f) the function increases over the entire number
scale.
5t
3.5.2. (b) The function increases on the intervals
(
( 0, — JT \
and
(
(
^- , 2jt
J
and decreases on ,
^
3.5.8. (a) increases throughout the number scale;
The function (b) the func-
tion increases on the interval (—1, 0) and decreases on (0, 1); (c) the function
decreases throughout the number scale; (d) the function increases on both
intervals ( —
oo, 0) and (0, oo) where it is defined; (e) the function decreases
on the intervals (0, 1) and (1, e) and increases on (e, -j-oc); (f) the function
decreases on the intervals ( oo, 1) and '(I, oo), increases— on ( 1, 1). —

3.5.10. a<0. 3.5.11. b^{. 3.6.1. (b) The minimum is / (1)-=/ (3) 3, the =
7 \
j=— —
'
maximum f(2) 4; (d) the=minimum / 3.6.2. (b) The minima

1
are f (±\)=V~3; the maximum /(0) = 2.

3.6.3. (b) The maximum is / (— 2)= 160; the minimum /(0) = 2.
3.6.7. (b) The minimum is /(0) = 0.
3.6.8. (b) On the interval [0, 2jt]: the minimum is /
^-^-j=r— 4; the maxi-
mum f 4 - 3 6 -10- (a)

- The minimum is / (0) = 0, the maximum / (2) =^4e~ 2 ;
(b) the minimum is /(— 2) = — 1, the maximum / (2)= 1; (c) the maximum is
/ 25
—5 =— —
\ f s
= 0, y — —
1
/(0) the minimum f i \

; (d) the maximum is / (±2) == 1,
the minimum /(0) = 7; (e) the maximum is / ( 3) = 3, the minimum

/(2)=-j/44.
3.6.11. (a) There is no extremum; (b) there is no extremum; (c) the ma-
ximum is /(0) = 0; (d) the minimum is /(0) = 0.
3.7.1. (c) The greatest value is the least value /(0) = 0; (d) the
greatest value is /
^
± = ^L_ - , the least value f(±\) — 0.
3.7.2. (b) The greatest value is j/(0) = y, the least value

î^y^^lh
(c) the greatest value is y(4) — 6, the least value y(0) = 0.
3.7.6. (a) The greatest value is /(— 2) = — the least value /(3) = — ^4
o
(b) the greatest value is /(0) = 2, the least value /(±2) = 0; c the greatest
( )
value is f(— —
) = -£-(-0.25
l
In 3, the least value /( V 3) = ^-0.25 In 3; (d) the

Yl) 6 1
*
'
3
^
greatest value
j
=^ is / if" » the least value / ^-^j=— 2; (e) the great-
est value is /(l)=l, the least value / (2) = 2 — In 2); there

(1 (f) is no great-
est value, the least value is /(0)=1.
3.8.3. H=R V =where H =the height the cylinder,
2, R the radius of
is of is
the sphere. 3.8.7. x a sin a, acos a, where a = 0.5 arc tan i/ 2.
Hint. The problem reduced finding the greatest value
is the function to of
S = 4xy-\-4x — = 4a 2a — sin a) (y x) 2 (sin 2
in the interval < a < ~ P max = ^r W = W 3.8.9. A = 2/?=

. 3.8.8. at £
1
.
4 4Wi
3u~
- — R
r=—
2tl
. 3.8.10. The radius of the cylinder
j base is
2
, where R
is the radius of the cone base. 3.8.11. The equation of the desired straight
. . x y .
T+T —
,
e 15
3.8.12. x=a p for a > p and x = for a^p.

3
3.8.13. v= 1/
V
-ttt
2b
• //in/. It will take —v 1
hours to cover one knot. The
appropriate expenses are expressed by the formula T = ^~^- = ~-\-bv 2 .
3.8.14. cp —— TL
. Hint. At the board width a the cross-sectional area of the
u
trough is equal to a 1
(1 +cos cp) sin cp. where cp is the angle of inclination of the
walls to the bottom.
3.8.15. . Hint. The point of fall of the jet is at a distance of

V
^ ^ from
2 g
the tank base, where h x is the height at H— — which the orifice should be
located, v is the rate of flow; therefore the length of the jet is determined by
the expression
V2gx y ^—^=2 Vx(h-x).

3.8.16. After hours the least distance will be equal to ^- km.

2v 2
3.9.1. (b) The intervals of concavity are

^
— oo, and (1, oo), of conve-
xity , 1^; the points of inflection are , , (1, 13); (c) the in-
tervals of concavity are (— )/~~3,

0) and (|^~3, oo), of convexity (— oo, — >^3)
and (0, V"s); the points of inflection are ^
— V~3, _ "7(^)> < ' °)> ( ^>
— — J
;
(e) the curve is concave everywhere; (f) the intervals of concavity are
3 -V~5 3 + VI
(0, x x ) and (x 2) oo), of convexity (x lt x2) t
where x x =e 2
, x2 =e 2
; the
points of inflection are (x lt y x ), (x 2 , y2) t
where
3.9.5. (a) The point of inflection is (3, 3); the curve is convex for x < 3 and
l/~~5 — i
concave for x > 3; (b) the abscissa of the point of inflection #=arc sin - ;
the curve is concave

•
in f
I
—— n
, arc sin
• V"5-A
- I ,
,
and convex
•
in
{
( arc sin -
• V^—\
—- ,
2~
3.10.1. (c) y = 0; (d) x = 0; y = 2x (i) as x—>+oo and y=—2x as

x —— > oo . 3.10.3. (a) * = 3, y = x — 3; (b) y= ± —
JTX
1; (c)y = x;
(d) x=±2; (e)y = 2x-^.

3.11.2. (a) The function is defined everywhere, it is even. The graph is sym-
metrical about the (/-axis and has no asymptotes. The minimum is y(0)-=l t
3 / V"3 23 \
maxima y (1) ==y (— =y 1) . The points of inflection are ( ± —g— ,
(b) the
J;
function is defined in ( —
and ( oo,
1, —
+oo). The graph has a vertical
1) —
asymptote x= — 1 and an inclined asymptote y=x 3. The minimum is y (0)^0, —
maximum y ( — 4) = — 256
The points of inflection are
.
/
6,
3296
and f — ——J
2, —6 \
1
J;
(c) the function is defined in (— oo, 0) and (0, +oo). The graph
has a vertical asymptote x = 0. The minimum is y (^-^-^=3. The point of inflec-
tion is
( -- V~* \
(d) the function is defined in the intervals (— oo, —
( ,
)
; 1),
g
(—1, 1) and (1, oo); it is odd. The graph is symmetrical about the origin, has
two vertical asymptotes x— ±1 and an inclined asymptote y x. The minimum =
is y ( V 3) = +3 — — , the maximum y{—V 3)--= 2~ ' point of
inflection is (0, 0); (e) the function is defined everywhere, it is even. The
graph is symmetrical about the (/-axis and has a horizontal asymptote y = 0. The
minimum is y (0) = j[/ 4, the maxima y (± }^ 2) =
2 jj/ 2. The points of inflec-
tion are (±2, \/ A)\ (f) the function is defined in ( — 2, +oo). The vertical
asymptote is x = — 2. The minimum is i/ (0) = 0, the maximum y ( — 0.73)^0.12.
The point of inflection is ( — 0.37; 0.075); (g) the function is defined everywhere.
The horizontal asymptote is y = as x —+ oo. The maximum is y

/ 3 \
f
/ 3 \ 3
Ae
The points of inflection are (0, 0),
-v~$ t
(tzVjy vi-sy (*±_o, (ttfiy e e
-z-vi
(h) the function is defined and continuous everywhere. The horizontal asymptote
is y—\. The minimum is i/(0) =
0, the point (0, 0) being a corner point on the
graph: y_ (0) = -i ,
f
y +{ 0) = +~.
3.12.6. 4.4934. x2 3.12.8.
0.202; jc 3 ^=—2.330;
2.128. 3.12.11. 0.6705. = =
3.12.12. (a) 0.27; 2.25; (b) 0.21. 3.12.13. (a) 1.17; (b) 3.07. 3.12.14. 1.325.
3.12.15. 0.5896 and 2.2805. Hint. To approximate the smaller root more precisely
write the equation in the form x e°- 8x =
- 1 to find a more accurate value of the
i
larger root represent it in the form 1 .25

(1 In x). x= +
3.13.1. No. Hint. Show that at the point the derivative is non-exis- x~\
tent: /'-(1)=1; Ml)=-1.
3.13.2. Hint. Check the equality / (b)
— f (a) = (b — a) f 1
a +b
2
3.13.3. Hint. Apply the Rolle theorem to the function / (x) a^ n -f- =
,..-\-a n - x (x) on the interval [0, x ].
3.13.4. Hint. Make sure that the derivative f (x)
f
A(x 5 1) has only one = —
real root, #=1, and apply the Rolle theorem.
3.13.5. Hint. The derivative /' (x) nx n ~ l p has only one real root at an — +
even n and not more than two real roots at an odd n.
3.13.6. Hint. The derivative is a polynomial of the third degree and has
three roots. Take advantage of the fact that between the roots of the polyno-
mial lies the root of its derivative.
3.13.7. Hint. From the correct equality limcos-r =0 (0 < £ < x), where g is
x ~> o s
determined from the mean value theorem, it does not follow that lim cos— = 0,
x ~> x
since it cannot be asserted that the variable £ attains all intermediate values
in the neighbourhood of zero as x 0. Moreover, g takes on only such a sequ- —
ence of values E for which lim cos — =0 (£££).
3.13.8. Hint. The mistake is that in the Lagrange formula one and the same
point £ taken for / (x) and cp (x).
is
3.13.9. Hint. Apply the Lagrange formula to the function \nx on the interval
[by a]; (b) apply the Lagrange formula to the function zP on the interval [y, x].
3.13.10. Hint. With the aid of the Leibniz formula ascertain that the
coefficients of the Chebyshev-Laguerrc polynomial alternate in sign, the odd
powers of x having negative coefficients. Whence deduce that L n (x) > for
x < 0.
3.13.11. Hint. Using the Rolle theorem, show that inside the interval \x xn
there are at least n roots of the first derivative, n 1 roots of the second deri- — , \
vative, and so on.

3.13.12. Hint. The L'Hospital rule is not applicable here, since the deriva-
tives, of both the numerator and denominator vanish at all points where the
factor sin* (which we cancelled in computing the limit ol the ratio of deriva-
tives) vanishes.
3.13.13. Hint. Write the Taylor formula with the remainder R2 :
6
h2 h'
f (a + h) = f(a) + hf' (a) + f" (a)+-^ /"' (a +6^).
Comparing it with the expansion given in the problem, get the equality
f" (a + Qh) — f" (a)
= }_f,„ a+^ n ( ) and pass over to the limit as h — > 0.
h 6
3.13.14. Hint. Prove by using the rule of contraries. Suppose that e== "~»
where p and q are natural numbers, p > q > 1, and, using the Taylor formula,
get for n> p
i =1+ _L + Jr+ ... + Jr+ _^r (£)«,„ <9< ,,
Multiply both sides of this equality by n\, and noting, that ~ n\ and
1
1
1!
1
. ..4 1
— n\ J
\- \ n\ are positive
r integers
b and —+r-r( —
/i
< —77 — <
n-\-\
•
q
1,
1
\ q J
obtain a contradictory result.
3.13.15. Hint. Verify that the function
sin* n
=}
i
x '
q x< 2 '
is continuous on the interval
f(x)
\ x=--0 1,
Ascertain that the derivative f (x) < is inside the interval.

3.13.16. Hint. Show that /' (*) ^ 0. Ascertain that
( /m 1
> for a <
= l-a<[ <Qfora>1> / 1,
/(0)
and take advantage of the fact that the function increases.

3.13.17. Hint. Show that the function f(x)=xe x 2 increases — and has oppo-
site signs at the end-points of the interval (0, 1).
3.13.18. Hint. Show that the derivative
/' W=-^+ 2xsin T — C0S T {x

* 0)
is equal to —3 at the points *=

9
— (n = ,
1
0, ±1, ±2, ...), and to

1
at
the points a: = 2~~» i- e - the derivative changes sign in any vicinity of the
origin.
3.13.19. Hint. Ascertain that the auxiliary function \p (x) / (*) cp (x) in- = —
creases.
3.13.20. Hint. Make sure that at all points of the domain of definition of the
function the derivative retains its sign if ad bcÔ. But if ad bc 0, i.e. — — =
—— d
c
then the function is constant. 3.13.21. p — — 6, #=14.
3.13.22. A minimum /(* ) = if cp (x ) >0 and n is even; a maximum
f(x ) if =
(p (* ) < and n is even; the point x is not an extremum if n
is odd. Hint. At an even n, in a certain neighbourhood of the point x the func-
tion retains its sign and is either rigorously greater than zero or rigorously less
than zero, depending on the sign of cp (x ). At an odd n t lie function changes
sign in a certain neighbourhood of the point x .
Answers and Hints to Ch. IV 439
3.13.23. Hint. For x ^ f (x) > 0, hence / (0) is a minimum. For x >
= 2— sin — + — cos — = -^—
the derivative /' (#)
xx x
is positive at the points x
ZiTin
and negative at the points x = ——/ . The case x < is investigated ana-
(2/i+1)ji
logously. 3.13.24. (a) and 0; (b) 1 and —2.
1
3.13.25. (a) The least value is non-existent, the greatest value equals 1; (b)
the function has neither the greatest, nor the least value.
3.13.30. Yes. Hint. Since /" (x) changes sign when passing through the point x ,
the latter is a point of extremum for the function /' (x).

3.13.31. The graph passes through the point ( 1, 2) and has a tangent M —
line y — =—
2 M
is a point of inflection, the curve being concave down-
ward to the left of the point M, and upward to the right of it. Hint. The func-
tion f" (x) increases and changes sign when passing through x 1. =—
3.13.32. h = —L=.
3.13.33. Hint. According to the Rolle theorem, between the roots of the first
derivative there is at least one root of the second derivative. When passing
through one of these roots the second derivative must change sign.
2n 2
3.13.35. Hint. The polynomial has the form a x' -\-aiX^"- -\- -{-a n - x x 2 -{-a n . . . .
Polynomials of this form with positive coefficients have no real roots.

3.13.36. Hint. Take advantage of the fact that a polynomial of an odd degree
(and, hence, also its second derivative) has at least one real root and changes
sign at least once.
3.13.37. Hint. Find

2
lim ( *L
**
!
+ +
Chapter IV
4.1.2. /=A: 3 + A: 2 + 0.51n| 2x— |+C. 1
3
2 -2- 2 - -
4.1.7. / = — (*+l) + ^-* 2 2
+C. Hint. Eliminate the irrationality from the
denominator.
4.1.14. / = ~arctan~ + C.
4.1.15. ^
/=-!=arctan-^±J+C.
^3
4.1.18. / = ln|* + Vx* + iyx+_\ + C.
3-
|
V\Qx-V7
4.1.20. I —
= U= In
2 ^70 V 10*+ V~7
+c
x—3
4.1.21. (a) —arc tan C; (b)
T (*-4) y X +C; (c) 3tanx-f
+2cotA: + C; (d) l-arc tan x-\- C.

x
4.1.22. (a) In (x-\- V 1
-f-*
2
)-farc sin x-f-C; (b) sin x — cos x-\-C\
2
(c)
In 5 5-l^
2 "*
+ C; (d) —0.2 cos 5* — xsin5a + C.
37
=
— 5 + C.
4.2.3. /
4 Y~2x
4.2.8. t= — 2 Vcosx+C. 4.2.10. / =— (a:

3
+ + 3a: 1)
3 +C.
4.2.13. (a) 0.75 + In a:)

4
+ C; (b) ln|ln*| + C; (c) arc sin -^-{-C;
(d) ârctan^ + C; (e) — 2 cos yV+C; (f) i- In 2 x+ In |

In a:
|
+C.
4.2.14. (a) __(35-40x+ 14* 2

)(1-a:) 3
+ C;
(b) |-(ln*-5)^l + lnx+C;
/ 2 2 2 \
(c) f
y — y sin 2
a:-|-— sin 4 a:
J
j/sin 3 *+C;
(d) — .1(8
2 + 4a:
2
+ 3a: 4 ) ^T^7 +C. 2
4.3.2. a: arc sinA: + + C.a:

2
4.3.14. — cos a: In tan x + In tan |
-J
+ C.
4.3.17. xln(x+ + rO — |/"i+x + C.
J^l 4
4.3.18. ^xf/x [(ln*)*-|ln* + | + C.

4.3.19. 2 |Ar+Iarcsin* + 4 J/T^t+C,
3* (sin * + cos sin 3)

4 3 21
- "
-
l+(ln 3)* + ^C
,
4.3.22. (±x»-x* + lx+!fje** + C.

4.3.23. (x* — 10x + 21) sinA: +2
A;(4A:
2 — 20) cos + C. a:
4.3.24. — cos 3*-) ^ — sin3x + C.

4.3.25. ^j-xt+ix^J \nx — £^+^_3*+C.
4.3.26. iârctanx-^+-J + C.
4.3.27.
y arc cos * — -H±^ j/T^F+C.
4.3.28. (a) _ 18^+g-13 s
.
n(6x+2) _6^tl cos(fa+2)+ ^,3 +

5
2
_|_i_ x 2_^ +C; (b) .|_ (jc

2__ 7jc+1)
( 2x-f-l)3
-A (
2a:-7)(2a:+1) 3
+
27 -i-
+ 320 (
2 *+!) 3
+C '
4.4.2. (d) Hint. Apply the generalized formula for integration by parts and
express In from the relation thus obtained
i
n -- —™ sin""
e •
„ i
1 ,
x (a sin x
•
— n cos x)-\ x
n (n—\)
3 — ,
/„-•>
n1
r ' n-
Answers and Hints to Ch. V 441
4.4.3. /„ = -
(nS* 2);
(n—\) sin"- 1
x n—l
cos x I cos a; 1
,
+ C.
, .
,
tan
2 sin 2 x 2 sin 2 x
1
'
2 2
4.4.4. (a) /„ = _L_tan"-iA:-/„_ 2 /i = — ln |

cos x |+C; / = a;-|-C;
(b) '„ = -' /i = In| sinx|+C; / = + C;a: (c) /„ =
1
v"- 1 Vx a/„_ 2 ; j/> + a + C; / = In |

x + Vx*+~a |
+ C.
Chapter V
5.1.2. *-2*+| In (Jt+1)»

1
+y ln|*+2| + C.
5.1.5. 2 ln |
x— 1 — In |* + C.
(x-
|
2
I)
9x4-1
5.1.8.
2
t^t- arc tan —>^7
^ I
arc tan (x + 2) + C.
3 V7 3
5.1.10. 5x + ln x + a
|
^ —2 +C |
3
5.1.11.
9* a + 50x + 68 1
ln
(x+\) (* + 2) 16
+ C.
4 (x + 2) (x + 3) 2
(* + 3)H
5.1.12. ~ — arc tan — + (x 2) C,
5.1.13. — 1
Jn
(1+,).
—x +x
1
arc tanx-
6(1+*)^ 6 1
l 1
arc tan — + C.
3 V3 1^3
5.1.14.
x+2
2 arc tan x + In ^+1
2(jc . + l) •
_"V?+T _
5.2.2. 4 £/*+6 y x+24 1
$/' x+24\n \
^ x—\\ + C.
5.2.4. Lârc tan ?4±+ln +

V"3
+ C, where
5.2.7. j/^1x + C. 5.2.8. -

2
5.2.9. f
\--x V )
1 — arc sin * + C.
5.3.3. -2 arc tan + 1
) +C -
5.3.5. 2 ln j/V-f- 2* -[-4 — *|

+ 2*- -4 — x—
|
2(/* 2
r l)
1+l ^-**+2arc
5.3.6. tan
|/ j±^+C.
5.3.7. |- 5.3.8.
(*+/T+?)" +e .
^ ~
2xX
'
L
15
5 Yx* + 2x 1-5— ln(x+l + y X + 2x+5) + C.
'
i
5.4.2.
3* + *--l
a
5.4.5.
o
y 3x 2_ 2x+ + c. \
5.4.6. Hiii j/x + a:+1 + |- In + 2 J/V-j-x+l + C.

2
|
2a: 1
-f- |
5.4.8. -1- — 14x + 111) Vx*-\-4x-r3- 66 In * + 2 +

(jc
2
+ 3 |+C. | -f- 4jc
o
5.4.9. —
64
(32a- 2 — 20*— 373) Vw + bx+7 -\
128
329 '
/2
In 1
4x+5+-
+ 2 V\x*+ 10a:+14| + C.
5.4.10. +
5 .4. 1 ..-^- 4X
,
+ 3 -2arcsin-L + C.
2
1
5.4.12. — y j-p- +C -
5.4.13. nh + + X + 3^ +
2 1 ^ 4 1
+ C. ////i/. First make the substitution
I
x2 = t.
5
5.5.2. 3arctan V x-{-C. 5.5.4. —\2 + x* J

4
-y \2 + x 3
J
4
+C.
1 8 5
5.5.5. A(l+xV -|-(I+*V + jg («+**) +C. s
^ |/ (l + ^/^) _3 ^(i + C. 7
5.5.7.
V
a/ - + 77= + c
-
5.5.8. 3 In ,
-
5.5... C+^f-^+C.
5.5.10.
^+^^- -11 +C .
A 8
+ / T4 + C
3
5.5.11.
|V (' )
-
5.6.2. —L
3 sin^ #
_L—
5 sin 5 x
f_c. 5.6.6. tan* + 4-tan**-|-C.
3
5.6.10. (a) —cot x+y cot x — 3 cot 5 x — x+C\

(b) JL tan 2 a:—-1 In ( 1 + tan 2
x) + C \ tan x + In
----
2
|
cos a-
|
+ C.
Answers and Hints to Ch. VI 443
5.6.12. — sinx — ]r- sin 3 x-\--^ In

1+
1 — sin x\
sin a: I
5.6.14. —ârc tan V 7= 7 +C
5.6.22. (a) -£+5!|^ + C;

x sinh4*
(b) — 2
arctan^
^tanh-^+r
j^— J + C.
5.7.3. }^?^) + â:(2*
l n (x+— 2
1) j/'^T+C.
o o
5.7.4.
5.7.7.
In (s + V^TT)--
/ = arc sin ^în-C.

+ +C-
^ 1
5.7.8. /==
*~ 1
-\-C.
4 ]/> — 2a: +5
5.8.2. / = 4 ^ — x+ 2 In (2 — x— 2 V —a:) — 2 (l +
1 1
V \ —x) In x + C.
cos + sin cc„
.- o , *
\-C where = arc tan x.
/ / , ,
a r^—
5.8.5. I'-'-e^ t t
/
2
-f-l
Chapter VI
6.1.9. / = 4» ^~^ ^ = 44 as the area of a trapezoid whose height is 5 — 1=4

and bases 4x1 — 1=3 and 4x5—1 = 19.
« i io s
6.1.12. n
"
=16— it:
1
4
175
2/i
— i
125
r"T~92
1
4/i
?
*
= 16-4—,175
—- ift
1
2/z
,125
1
—
\-—
4/z
t
2
.
6.2.2. (a) 1; (b) A ;

(c) . 6.2.10. (a) 1; (b) i- In ~ ; (c) jx;
-J
(d ) -J- arc tan -J; (e) In 2; (f) 1; (g) arc tan e~; (h) i; (i)
j|
(J)4- ;
(
k) J^lziO. 6.3.1. (c) 3 < / < 5. Hint. M = /(0)=A m =
= /(2)=A. 6.3.11. (a)^HL^; (b) - V 1-f-x 4
. 6.3.14. (b) ^ . 6 3. 15. (b) ^=
= -e-»smx. 6.3.23. (a) In x; (b)~. 6.3.24. (a) ^ = jjL ; (b) yi=
6.3.25. (a) The maximum is at *=1, the minimum at # =— 1; (b) the
minima are at x = — 2; 0; 2, the maxima at x= ± 1.
6.4.3. (a) -— (substitution x = a sin 0; (b) o (substitution *= tan 0-

lb I
Û+ln 2+]^ r
3
6.4.6. (a) ^2 ;
(b) 2(y 3-l); (C ) 8 + ?-^ji.
]A 3 2 1
+ K 2
. n
Sm
24
6.4.J5. (a) 2 — 2 1n2; (b) 0.2 In 112; (c) ; (d)V3— 0.5 In (2 +
sin | sin ^
+ x— cos x = t)\ —
V^3); (e)0.251n3 (substitution
—
sin (i) a3
^
(substitution x = acos/); (g) (substitution x = 2a sin 2
/); (h) -^-^-j •
6.4.16. (a) -g-; (b) -5-; (c) -j- In (substitution a:

4 = t)\ (d) -^(substi-
tution x 2 =a 2
cos 2 / + 6 2 sin 2 /).
6.4.17. The substitution x = -j~ w iH n °t do, since this function is disconti-
nuous at / = 0.
6.4.18. The substitution / = tan —x will not do, since this function is discon-
tinuous at x = 7i.
6.4.19. Hint. The inverse function b

x=±
V~ t is double-valued. To obtain
the correct result it is necessary to divide the initial interval of integration into
two parts:
2 _ o _ 2
_
^ \/ x 2 dx = ^ ]/ x 2 dx +^ \/ x 2 dx
-2 -2 _
and apply the substitutions x—— Y~t b in —2 < x < and x=-f V~t b in <
< x < 2.
6.4.20. It is impossible, since sec 1 and the interval of integration is
[0, 1].
6.4.21. It is possible; see Problem 6.4.12.
a a
G.4.22. Hint. On writing ^ / (x) dx ==

^ / (*) d* +J / (*) dx, make the sub-
-a -a
stitution x —— / in the first integral.
l —l o
6.4.23.
J
/ (arc sin 0^+ J
/ (n — arc sin /) dt+
J
/ (2ji + arc sin /)
o l -l
Hint. Represent the given integral as the sum of three integrals for the in-
tervals: 2jT and substitute tne variable: x =

y) (t' "IT") '
) '
= arc sin t, x = n — arc sin/, x = 2ji + arc sin respectively. /
6.5.3. (1) If f (x) is an even function, then

31 31 71
[ f (x) cos nxdx — 2^ f (x) cos nx dx, and

^ / (x) sin X dx — 0.
-JT — JT
Jl JT
(2) If / (a:) is an odd function, then / (x) cos nx dx == 0, and

^ / (x) sin nxdx—
J
-JT — JT
JT
= 2
^ / (x) sin /i dx.
6.5.4.0. 6.6.3. 6 — 2e. 6.6.5. jt }/" 2 — 4. 6.6.6. jt — 2. 6.6.13. (a) -5- — 1
(b) -|; (c) (d) 1; (e) In 2- ± ;

(f) In
|
<g)f-l; (h)i|l-2^3.
Answers and Hints to Ch. VI 445
6.6.14. Hint. Integrate by parts twice, putting u = (arc cos x) n the first
time and u = ~
(arc cos x) n l the second time.
6.6.15. Hint. Integrate by parts, putting u x. =
6.7.4. 0.601. Hint. Estimate 1V (x) on the interval and put
(a) | / \
yj
2/z = 6; (b) 0.7462. 6.7.5. 0.96
6.8.1.
for 0<x<l,
F(x)--
-i for 1 < *<2,
Continuity is checked directly. The assertion concerning the derivative re-

quires checking only at the points *=1, x 2. —
6.8.2. Hint. Make sure that the function / (x) is continuous both inside the
interval (0, 1) and at the end-points ( lim f(x) f(0) and lim f(x) f{\)). = =
X-> + *->l-0
6.8.3. No. Hint. Consider the function
| 1 if x is rational,
cp (x) == 1
if x is irrational on the interval [0,
1 1].
b
6.8.4. 1 — ]/"3. Hint.

J
/" (x) dx = f* (b)
—f (a).
a
6.8.5. Hint. Putting for definiteness x > and
E (x) = n^ x < /z+1,
take advantage of the additivity of the integral
x \ 2 n x
E (x) dx= ^ E (x) dx+ ^ E (x) dx+ ... + [ E (x) dx-\-

^
E (x) dx.
1 n-l n
6.8.6. The antiderivative F 1 (x) will lead to the correct result and F 2 (x) to
the wrong one, since this function is discontinuous in the interval [0, jc].
X
6.8.7. F {x) =y + J / (0 dt. Hint. Any antiderivative F (x) can be represen-
ted in the form F (x)

^ f (t) dt + C. Putting * =x 0t find C yo-
=
6.8.8. l=-^\n
2 2b — 2a '
The function is defined on the interval

6.8.9. 1, 1), it is odd, and [
— in-
creasing; convex on the interval [—1, 0] and concave on the interval [0, 1];
the point [0, 0] is a point of inflection.
6.8.10. Hint. The function
x x at i < a:^ 1
'<H I at, =
1_
is continuous on the interval, it reaches the least value m=e e

^0.692 at
=— and the greatest value M— 1 at x = and at x—l.

2 sin x
6.8.11. Hint. Integrate the inequality
6.8.12. Hint. Integrate the inequality
and write Schwarz-Bunyakovsky inequality
J
V x sin x dx ^ J
x dx
J
sin # dx — 8"
-=
"J/ o
2 VI
6.8.14. //i/i*. Apply the Schwarz-Bunyakovsky inequality in the form
b ~|2 b b
b
a -I a a
6.8.15. Hint. Make the substitution arc tan * = -y-

X
6.8.16. Hint. If f (x) is an even function, then F(x) = ^f(t)dt is an odd

o
function, since
—X X
F(-x) = ^
f(t)dt = —^f(—z)dz = —F(x) (/ =— z).
And if f (x) is an odd function, then F(x)—^f(t)dt is an even function,

n
since
—X X
F(—x)= ^f(i)dt = -^f{-z)dz = F(x) (t = -z);

n o
all the remaining antiderivatives have the form F {x)-\~C and, therefore, are
also even functions.
6.8.17. Hint. The derivative of the integral / with respect to a equals zero:
~ = /(a + 7W(fl) = 0.
Chapter VII
7.1.4. (a) In 2; (b) -| (2 l/"~2- l); (c) A; (d) 1; (e) ±.
7.2.2. (a) ~2 i
v
(
b) 4-4-4-
2^2 !n -t-t^ °- 283 -
7 2 5 * '
-
X- 7 2 - 10
- '
^
'
' e-j-1 4 '
'
'
h V& + h*
7.2.13.(a) (i=y; (b) (1=10 2; (c) (1=^ + 2. 7.2.15. y. 7.2.16.
^
7.3.4. ~. 7.3.6. ~+4- arc sin A- 7.3.11. A. 7.3.13.9. 7.3.16. —i-r
6 3 2 1
5 15 m-\-l
Answers and Hints to Ch. VII 447
r n
7.3.19. y. 7.3.20. y. 7.3.21. 2ji— (2 V 3)ln(2+V 3). 7.3.22. 0.75jt.
98 4 R Q
7.3.23. -i-
1
15
. 7.3.24. —o
1
. 7.3.25.
6
. 7.3.26. —
15
. 7.3.27. -i-
lz
1
. 7.3.28. —
oU
.
7.4.6. —5 . 7.4.8. 0.75jtafr. Hint. The curve is symmetrical about the coordinate
axes and intersects them at the points x= ±a, y = ±b.
7.4.9. (a) — . Hint. The curve is symmetrical about the #-axis, intersecting
it twice at the origin at / =± 1. The loop is situated in the second and third
quadrants; (b) — . Hint. The points of self-intersection ot the curve are found
in the following way: y — tx (/), therefore y = x — x

(t x ) t
x {t x ) t 2 (t 2 ) at t1 ^zt 2 and
x(t 1 ) = x(t
7.4.10. 0.25nab.
2 ), only if
D
x (t^ = x
Hint. The curve is symmetrical with respect to both axes of

(t 2 ) = 0, i.e. ^ = 0; = 2\ (c) t
2
8
^ .
coordinates and passes twice through the origin forming two loops. Therefore, it
is sufficient to compute a quarter of the desired area corresponding to the variation
of / from to y and multiply the obtained result by 4.
7.4.11. ^-J-. Hint. The curve resembles an astroid extended in the vertical
Sab
direction.
The curve
7.5.2. (a) 5^.; (b) . Hint. is a circle of radius
y passing
through the pole and symmetrical about the polar axis, —^"^(p^y.
XI JX
7.5.6.2^(^1-.). 7.5.8.(a)^; (b) f£ . 7.5.9.^(^-^3").

'
Tia
7.5.10. -575- . Hint. The curve passes through the pole forming two loops located
oZ
symmetrically about the (/-axis in the first and fourth quadrants. It is sufficient
to calculate the area enclosed by one loop corresponding to variation of cp from
ji
to y and double the result thus obtained.
7.5.11. —5 jt<2
2
. Hint. The curve passes through the pole, it is symmetrical
about the polar axis and situated in the first and fourth quadrants. It is sufficient
to calculate the area of the upper portion of the figure which corresponds to
variation of cp from to — ji
and double the result thus obtained.
(,+_£).
7.5.13.
zia^
— 1
. Hint. The curve is symmetrical about the coordinate axes and

intersects them only
the origin, forming four loops one in each quadrant
at —
(a four-leaved rose). Therefore, it is sufficient to find the area of one loop corres-
ponding to the variation of cp from to y and multiply the result by 4.
2
7.5.14. Y~2 na Hint. The curve is symmetrical about the axes of coordi-
.
nates and the bisectors of the coordinate angles; it cuts off equal intercepts on
the axes. The origin is an isolated point. It is sufficient to compute the area of
one-eighth of the figure corresponding to variation of cp from to and mul-

tiply the result by 8.
2
7.6.2. 9— ji. Hint. A plane perpendicular to the *-axis at the point x will
cut the sphere along a circle of radius r= j/"l6 —x 2

therefore the cross-sectio-
—
,
nal area S(#) ji(16 x 2 ). =

7.6.5. 0.5na 2 h. Hint. The area of a triangle situated at a distance x from
the centre of the circle is equal to h Yâ 2 —x 2 .
7.6.10. 2n 2 a 2 b. 7.6.11.
y (see Problem 7.3.9). 7.6.14. 5jx
2
a3 .
7.6.16. (a) 2nab(^\+^j\ (b) |a; (c)

^ abk 2
7i. 7.6.17. |-a 3 tana.
16 64 64 4
7.6.18. (a) 12ji; (b) —
10
jx; (c) -=-ji;
O
(d) ji
2
; (e) ~ji;
O
(f) -r-
o
no*.
3 2 / 11 _ 2£A
^ 3 9
1
7.6.19. 5jL. 7.6.20. 2L. 7.6.21. \na*\ e « -e a

) +na c = 2
sinh
^+
2
-f jra c. 7.6.22. ^(6ji-|-5 }/" 3). Hint. The abscissas of the points of intersex
tion are: x x1 = —~
3
tt
;
*2 = tt
. 7.6.23.
1Q
—ji.
48
7.6.24. ^
127
7
ji. 7.6.25.-^^2
Ifinr 6
3 105a/?
——
2
c
//t/i/. Represent the evolute of the ellipse parametrically as follows: x cos 3 /
y = — ^-s\n 3
t, where c= y a 2 — b 2 . 7.6.26. ^-na A . 7.6.27. -^L
3
jra
2 In ( 1
-f- |^ 2)
— |-j . Hint. Pass over to polar coordinates,
7.6.28. 1 Jta 3 . 7.7.2.^. 7.7.4. In

J^—
^ . 7.7.8. (a) f 6+ ^2+ V 3)
In (
(b)2ln(2-^3). //W. ^=-i;x,

2
= '
2
-2-;
3 '
(c)
v 7
3
. 7.7.9.
+ ^ 2
7.7.10. 10^+^5^. 7.8.2. 8a. 7.8.5. y. Hint. The curve intersects the
axes at t
x
= and =iV 8. 7.8.7. 4 j/~3.
*
2
7.8.8. 16a. 7.8.9. 8na. Hint.
See Fig. 79. 7.8.10. ii^Z^l 7.8.11. y . . 7.8.12. At * = ~L the point
2jt j/"3\ 3a
al
3
7.9.5. 1.5jxa. 7.9.9. In -|- . 7.9,10. 2 ^2 jia.
2
//m/. The curve p = 2 ^ 2 a cos cp — is a circle.

^ ~J
7.9.11. p[|^2+ln(l + r2)]. 7.10.3. (a) Jii ;
(b)^.
7.10.5. 2jt ( l
+ 3-i^V 7.10.8. -. 7.10.14. (34 V~\j-2)?-.
V V 3/ 2 9
7.10.15. 2jx [ ^24-^(1+ 1^2)]. 7.10.16. y^ 2

- 7.10.17. --y-^ jx (<?* — 2).
7.10.18. 29.6ji. 7.10.19. 4n 2 a 2 . 7.10.20. jta 2 . 7.11.7. 16a 2 where a is

o
Answers and Hints to Ch. VII 449
the radius of the cylinders base. 7.11.8. 1.5jx. 7.11.10. (a) -;

^~T arctan T- 71,n ' (a) (b) £(3 + 4 Y~2)\
3— 1). V2
(0 l(53i
o
+ 6 V 3). 7.11.13.
£
(2 In 7.11.14. (5 J/5-2 V 2).
^3 7.11.18. na 2 /"p?. 7.11.19.

7.11.17. 2jx
15
7.11.20.
nabh
7.11.21. I2ji. 7.U.22. l
/
4 ^- 6
')^fl. 7.11.23. — jta 3 .
^5—1
7.11.24. (a) jx (^5-0) + (^2+l)
+ 21n ^±fL3) (c)2n ,fc
=f\ 7.12.9.^.
7.12.2. 7.12.4.
;
-f^.
7.12.11. nabhd. 7.12.12. nrdh'K 7.12.13. nR * H ' 7.13.3. 0.25:rc/? 3 .
j2
M (5^ 5-1); M y = ^V 5 + lln(2+^ 5). Mx =

1
7.13.7. 7.13.8.
3
r
ft
V fl
2
+^ 2
;
M y =-j V'a 2Jt-b^ 7.13.9. /"2 + ln(l+ /"2). 7.13.10.0.15.
„ « rt « ,
7.13.11.
.
/* = f2 aft
3
. a3b
7.13.12.
(a + 3b) /i
3
7.13.16. x c = f/ = 0.4a.
; /jf r
12 12
7.13.19. Xc = y c = -j. 7.13.26. = ; yc = 0. 7.13.28.^ = ^; yr = 0.

0.2 (2e 2
«— 71
)
0.2a (e
2 *— 2e*)
4.5na y
.
7.13.29. *r 0r = 7.13.30. .
4/? m—n m— n
7.13.31. x,
n 7.14.1. ; 4
m-\-n
if both m and n are
m-\-n
m—n
even; 2 if both m and /z are odd; if m and n are of different
m-\-n m-\-n
evenness. Hint. The curves y m =x n and y = n x m have two common points (0, 0)
and (1, 1) in the first quadrant. The area of the figure situated in the first
quadrant is equal to dx . Depending on evenness and oddness
of m and n this figure is mapped symmetrically either about the coordinate

axes (m, n even) or about the origin (m, n odd). If and n are of different m
evenness, then the curves enclose only the area lying in the first quadrant.
7.14.3. Hint. Take advantage of the formula for computing the area in
polar coordinates.
7.14.4. Hint. Since the figures are of equal area, the function S(x) appear-
b
ing in the formula for the volume V= ^ S (x) dx is the same and, consequently,
a
the values of the integrals are also equal.
7.14.5. Hint. The formula follows directly from Simpson's formula
h
/ (x) dx = ±[f(0) + *i(j^+f(h)].

J
2
x2
for a sphere S(x) = n(r — x 2 2
)\ for a cone 5 (x) — nr 2
; for a paraboloid of
^
revolution S (#) = 2jip# and so on.
Hint. Divide the curvilinear trapezoid into strips A# wide and write
7.14.6.
an expression for the element of volume AV 2n xy Ax. =
7.14.8. Hint. Use the formula for calculating the length of a curve represen-
ted parametrically.
7.14.9. In — . Hint. The point nearest to the origin with a vertical
tangent corresponds to / — y.
7.14.13. 7.14.14. V2-z. 7.14.16. (a) 0.5 In (x+y);
(b)
JX
- — 0.5 arc sin x.
Chapter VIII
8.1.2. (b)
y ln2; (c) 1; (d) 1 — In 2; (e) ji; (f) 1.
* n
8.1.6. (a) It diverges. Hint. ^* ^ > y for X > Y'e—l; (b) converges;
M^
2 *
(c) diverges. Hint. > _J__ .
converges; (e) diverges.
Vx
8.1.17. (a) 0. Hint. Represent the integral as the sum of two items:
CO 1 cc
n X nX nX
f
^
dx—[ } x> dx-\- f } dx Make the substitution * — 4" in the se-

+* + iy .
t
2
J 1 J l
J 1
-r ^ '
o o i
CO 1
" X " X
cond summand and show that \ . dx=- — \ .
, dx; (b) ^
1 o
_2_ _
3
8.2.2. (a) 9a ; (b) it diverges; (c) diverges; (d) 6^2; (e) y;
(f) converges for p < 1 and diverges for 1.
8.2.7. (a) It converges; (b) diverges; (c) converges; (d) converges;
(e) diverges; (f) converges. 8.2.11. (a) It diverges; (b) 2 V^ln 2; (c)
y.
8.2.14. (a) It converges; (b) diverges; (c) diverges; (d) converges;
-1
(e) converges. 8.3.7. (a) y ;
(a) 2ji. 8.3.8. 3jw 2 . 8.3.9. . 8.3.10.
^p.
8.3.14. mgR. Hint. The law of attraction of a body by the Earth is deter-
= niP R 2
mined by the formula f -^y— 2
, where m is the mass of the body, r is the
distance between the body and the centre of the Earth, R is the radius of the
Earth.
8.3.15. e v Hint. Electric charges interact with a force -jf- , where ex and e2
are the magnitudes of the charges and r is the distance between them.
8.4.1. Hint. Represent the integral in the form of the sum
+ oo a +co
dx _C dx C dx
f
J
1
xp \n9 x~ j
1
xp \n^x J
a
x? M x
{a > l)
Answers and Hints to Ch. VIII 451
and apply special tests for convergence, taking into consideration that in the
integral In a:= ln[l
first + —
and in the second integral 1)] — a: — 1 as x — 1,
the logarithmic function increases slower for q < than any power function.
8.4.2. Hint. Making the substitution xQ=t, reduce the given integral to the
form ± —y
I
\ t
£+l
q
~ x
smtdt. Represent the integral
y\ t
p±i
q
~ l
sin / dt as the sum
o o
1 + GO
f —— dt + C — dt, where a = 1 — , and show that the integral conver-

J t* J t* q
o i
ges absolutely for 1 <a<2 and conditionally for 0<a^l. Note that at
îi_L — the integral is reduced to the conditionally converging integral
-f CC +X
^jl-dt, and at —— 1 to the diverging integral ~^-dt.
J J
o o
1/2
8.4.3. Hint. Represent the given integral as the sum xp

- l (\—x) (i- dx- v 3
r
J
o
l
-}- xP~ l
(1 — x)*!- 1
dx and apply the special comparison test.
1/2
T
8.4.4. Hint. If |
a |
7= | p |, then sin ax- sin fix dx is bounded.
8.4.5. Hint. By substituting t =x 2

the integral is reduced to the Euler
gamma-function.
8.4.6. Hint. hM-fMdxJ fJ&dx-[ L&dxJU!!^

J % J x J x X
J
a aa a& aa
= -4 In h \ dx. Applying the generalized mean value theorem, show

OC J X
aa
that the last integral tends to zero as a 0. —
8.4.7. Hint. Take the function f(x) e~ x for the = first integral, the function
f(x) =
cosx for the second and take advantage of the results of Problem 8.4.6.
8.4.8. It converges for < 3 and diverges for m m ^ 3. Hint. Take advantage
of the equivalence of 1 — cos a: and — as x— 0.
ji 2
P dx C dx
8.4.9. Hint. Represent \ —as the sum of two integrals
5 \
k -f-
J (sin x)* J (sin x)
n
P dx
+ \ ; reduce the second integral to the first one by making the substi-
J (sin*)*
ji
tution x = — tl t and take advantage of the equivalence of sin* and x as x — 0.

sinx{[ - cosx) ™*(i-™x)

8 4 .,o.
. mni . dx= dx +
] §
CC
X C0S
4 ^ ^'^s
^ dx '
The inte & rand of tne first summand on the right side is
y
an infinitely large quantity of order s 3 as x >0. By the special comparison — —
test the first integral converges absolutely for s 3 < 1, i.e. s < 4, and diverges —
for s^4. The second integral in the right side converges absolutely for s > 1,
since the function sin#(l cos a:) is bounded. But if < s —
the second in- < 1 ,
tegral converges conditionally as the difference of two conditionally converging

00 cc
^s\nx-josx
integrals *-^dx and dx (§ee Pro5 em i 8>I j 3 ).
^
TT JT
2 T
8.4.11. Hint. Integral (2) can diverge. For example, let
( 2/ux<A:<(2/z4-i)ji,
YV(x) —
1,
cp <
1—1, (2/i H- 1)ji < x < (2/z4-2) jt.
CD CO
^^-^^(x)dx =
S
The integral -^^- dx converges (see Problem 8.1.13). But
J
o o
00 CO
S *" X
= ^ I
<2a: diverges (see the same problem). But if the integral

^ / (x) dx con-
o a
cc
verges absolutely, then the integral f (x) cp (#) dx also converges absolutely: if
J
a
|
cp (#) | < C, then | / (a:) cp (a:) |
< C \ f (x) |, and it remains to use the comparison
theorem.
ji
T" x
8.4.12 //m*. Transform the integral f (x) into f (x)= ^ In sin z dz by the
JT
substitution
Jl
— z. Taking into account that sin z — 2 sin —
Z
• cos —Z
,
reduce
the above to the sum ol three integrals.

8.4.13. Hint. Putting w=ln cos x\ cos 2nx dx = dv, integrate by parts and get
T
p sin
=— x
1
the equality
J /,, \ sin 2/?a
cos*
dx, n ^ 0. Since
' 2/2 J
sin 2/2* = sin (2n — 2) x-cos 2a, 4- sin 2a -cos (2/z — 2) a:,
Answers and Hints to Ch. VIII 453
r n_
2
ln
~2n
-
f /0
\ sin (2n
.
— nx2) ^—
sin a: , ,
cos #
L
n
2 2
+ jsin(2n-2>*.sin 2 ^ + U^ 2 cos (2/z — 2) a:
J
Check by direct calculation that for n ^2 the second and the third summands
equal zero. Therefore, for n^2
\ sin (2/z
v
— 2) 7
A'
cos x
dx =
n
/„_,.
n 1
2/i J
n n
Since /,---—
2
\ sin 2jt
cos
4* = ^ we have
4
/2 == — —2 4 *
2
/ 3 =.-.-.
1
2
'
T~^3^4
J a: <j
and by induction, /„ = (— 1)"- ^ 1

.

Thomas Calculus

Uploaded by

Copyright:

Available Formats

Thomas Calculus

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thomas Calculus

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What topics are covered in Chapter I?

What topics are covered in Chapter I?

I.A.

(With elements of theory)

FROM THE RUSSIAN

First published 1973

Revised from the 1970 Russian edition

From the Author 9

Chapter I. Introduction to Mathematical Analysis 11

§ 1.1. Real Numbers. The Absolute Value of a Real Number 11

§ 1.2. Function. Domain of Definition 15

Chapter II. Differentiation of Functions 93

§ 2.1. Definition of the Derivative 98

§ 2.4. Differentiation of Inverse, Implicit and Parametri-

§ 2.5. Applications of the Derivative 115

Chapter III. Application of Differential Calculus to Investigation of

§ 3.1. Basic Theorems on Differentiable Functions .... 131

§ 3.2. Evaluation of Indeterminate Forms. L'Hospital's Rule 138

§ 3.5. Testing a Function for Monotonicity 148

§ 3.6. Maxima and Minima of a Function 152

Chapter IV. Indefinite Integrals. Basic Methods of Integration .... 190

§ 4.1. Direct Integration and the Method of Expansion 195

Chapter V. Basic Classes of Integrable Functions 214

§ 5.1. Integration of Rational Functions 214

§ 5.8. Integration of Other Transcendental Functions . . . 240

Chapter VI. The Definite Integral 247

§ 6.1. Statement of the Problem. The Lower and Upper

Chapter VII. Applications of the Definite Integral 310

Chapter VIII. Improper Integrals 387

From the Author

This textbook on mathematical analysis is based on many years'

vely, that is to say, if he studies the relevant theoretical material

§ 1.1. Real Numbers.

Nonperiodic decimal fractions are called irrational numbers.

The properties of absolute values are:

1.1.1. Prove that the number

By a direct check it is easy to ascertain that q^ 1

Prove the reverse, namely, y = ]/ x + x + 32

g4-2?» + 7g*-6? + 9 (^-^ + 3)2 _ ^_^ + 3 ,

(b) Prove that \/~2 is an irrational number.

is also rational, since it is the quotient of two rational numbers.

1.1.7. Prove that for every positive rational number r satisfying

<r + 2r/z + That

1.1.9. Solve the following inequalities:

(b) (x-2) 2 >4;

\x — 7x+ 12 >x — 7x+

The inequality |2# — 3|<1 is eqivalent to the

(d) The given inequality is valid for those values of x at which

1.1.10. Find out whether the following equations have any

(a) \x\=x b] (b) |x|=x +5? —

1.1.11. Determine the values of x satisfying the following equa-

1.1.12. Determine the values of x satisfying the following equ-

(b) |(x — 4) — (x + 2)| = |x — 4 — |x + 2|.

Solution, (a) The equality |a + 6| = |a| + |fe| is valid if and

at any values of x the equality is satisfied at those values 9

1.1.14. Find the roots of the following equations.

— sinjc = sinA;+ 1, or sin a: =— 1

hence 0<<2l/T. If<K"2,