Little Mathematics Library

o o

V. G. BOLTYANSKY
I )I FFE RENTIATION
EXPLAINED

Mir Publishers. Moscow

nonyjiflPHbiE jiekuhh no matemathke

B . r . E o jI T f lH C K H H

MTO TAKOE
^HO mDEPEHUMPOBAHHE?

«<J>H3 MaTTH3 » MoCKBa

 LITTLE MATHEMATICS LIBRARY

V. G. Boltyansky

D IF F E R E N T IA T IO N
E X P L A IN E D
Translated from the Russian
by
M. Samokhvalov, Cand. Sc. (Tech.)

MIR PUBLISHERS • MOSCOW

 First published 1977

Ha aHZAuucKOM tnbiKe

© English translation, Mir Publishers, 1977

 CONTENTS

Author’s P r e fa c e ................................................................... 6

The Problem of a Free Falling B o d y ............................... 7

Statement of the Problem 7
Qualitative Solution of the Problem 9
Formula for the Velocity of a Falling Body The Number e . 13

D ifferentiation ..................................................................................... 25
The Concept of the D erivative......................... 25
The Differential Equation . . . . . . . . . 27
Two Problems Leading to Differential Equations . . . . 28
Napierian Logarithms . . . 33

Harmonic O scillation s ................................................................... 34

The Problem of Small Oscillations of a Pendulum 34
The Differential Equation of Harmonic Oscillations . . 42
The Oscillatory C ircuit................................................................ 45
Oscillations Produced by the Action of the Elastic Force of
a Spring. . 47

Some Other Applications of the Concept of the

D e r iv a tiv e ..................................................................................... 52
Maximum and Minimum Values ................................................. 52
The Problem of Drawing a Tangent . 58
Modelling . . . . 60

Afterword.................................................................................. 62
 AUTHOR’S PREFACE

High school students, especially those interested in mathematics,

physics and engineering, often ask, ‘What is “higher” mathematics?’
Sometimes they discuss this and similar questions at mathematics
clubs at schools.
In this book I have tried to explain, in a way a high school
pupil would understand, certain concepts of higher mathematics**
such as the derivative, differential equation, the number e, and
natural logarithm (pupils are more apt to be aware of and interested
in the latter two concepts). Wherever possible, I have tried
to illustrate the concepts with problems taken from physics.
In addition, I have tried to show that the concepts of “higher
mathematics” are mathematical reflections of actual processes, that
mathematics and life are connected, not separated, and that
mathematics is a growing, not an unchanging, completed science.
Not all proofs and arguments are presented with complete mathe­
matical rigour. Some arguments are presented for illustration.
This method seems to me more appropriate for a general book.
The book can be used by mathematics and physics clubs at
school. Part of the material is taken from lectures the author
gave at the request of the advisers of school mathematics clubs
at the Moscow State University.
I would like to take this opportunity to express my sincere
gratitude to A. I. Markushevich and A. Z. Ryvkin for their valuable
advice and comments on the manuscript.

** The reader may become acquainted with certain concepts of higher

mathematics by reading other books of this series: A I. Markushevich
Ploshchadi i logariphmy (Squares and Logarithms) (No. 9) and I. P. Natanson
Summirovanie heskonechno malykh velichin (The Summation of Infinitesimals)
(No 12)

6
 The Problem o f a Free Falling Body

Statement of the Problem

We shall first estimate the velocity of a body that falls to the
ground from a certain height.
Elementary physics tells us that a body falling freely in a vacuum
obtains after t seconds from the start of the process the velocity
v = v0 + gt (1)
where v0 is the initial velocity and g the gravitational acceleration.
In certain cases formula (1) remains approximately valid for
a body falling in the atmosphere (as distinct from a vacuum),
in other cases it can lead to gross errors. For example, it is
applicable to a body falling a short distance. When a body falls
from a great height, the expression of velocity greatly differs from
formula (1). In 1945 V. G. Romanyuk in a free-fall jump fell
12 000 m before he opened his parachute. In a vacuum a body
which fell from such a height (without initial velocity) would at
ground level have attained a speed of around 500 m/s. In fact,
qt2
it follows from the formula s = y that falling time equals (in
a vacuum)

'2s 2 x 12 000 m
t= ; 49.5 s
9 9.8 m/s2
and from (1) we obtain the speed
v = gt x 9.8 m/s2 x 49.5 s 485 m/s
(We could have used the formula v2 = 2gs right away.) At the same
time it has been established that the speed of the parachutist
in the course of a free-fall jump reaches 50-60 m/s and does not
exceed this value. Thus, in this case formula (1) leads us to an
erroneous conclusion.
Here is another example: the parachute is designed so that a man
descending with it approaches the ground with a speed of

7
approximately 6.5 m/s no matter from what altitude the descent
is being made.
Clearly, formula (1) is in this case of no value.
This points to the conclusion that the speed of a body falling
in the atmosphere reaches a constant value in the course of time.
In other words, some time after the start of the fall the motion
of the body becomes uniform and its acceleration becomes zero.
This means that the resultant force (the sum of all forces) acting
on the body is zero.
It is easily seen why formula (1) is of no use for calculating
the velocity of a body falling in the atmosphere. The formula is
deduced on the assumption that the body moves under the action
of only one force, namely, the force of gravity
P = mg (2)
On the other hand, we have established that for a body falling
in the atmosphere the resultant force (some time after the motion
has begun) becomes zero, i. e. the force of gravity P is compensated
by some other force which was not taken into account when formula
(I) was deduced. This compensating force is the air resistance,
or drag. It is the drag which prevents the parachutist from attaining
excessive speed —it “supports” him, so to speak.
How can drag be taken into account? Suppose there is no wind.
For a stationary body the drag is zero. The faster the body moves
the more difficult it becomes for it to penetrate the air, i. e. the
drag increases. This fact may be easily demonstrated on a still day
if you increase your speed: start walking, then running, cycling, etc.
We shall suppose that the magnitude of this force is proportional
to the velocity, i. e. that it equals bv, where b is the proportionality
factor and v the speed of motion. This assumption is well supported
by experiment for low speeds*’ not exceeding 1-2 m/s. The factor b
depends on the dimensions and on the shape of the body.
For instance, the speeds being equal, the drag of a ball is
approximately 20 times that of a “cigar-shaped” body of equal
cross section (Fig. 1).
Confining ourselves to these brief remarks, in future we shall
assume that the drag (for which we shall adopt the designation S) is
S = -b v (3)
the minus sign shows that the drag has a direction opposite to that
of velocity.

*’ Note that for speeds above 1-2 m/s the drag exceeds the value bv.
It is sometimes assumed that the drag is proportional to the square
of velocity

8
 All said, we shall assume that a body thrown downwards with
an initial velocity is acted upon by only two forces: the force
of gravity P and the drag S. Making use of Newton’s Second Law
we may write:
ma = P + S (4)
where m is the mass of the body and a its acceleration. It is more
convenient to point the positive direction vertically downwards than
upwards since the velocity of a falling body is directed downwards,
and, if this convention is adopted, it will be positive. The force
of gravity which is also directed downwards will also be positive.
On the other hand, the drag will have the direction opposite
to the velocity, i. e. upwards, and will for this reason be negative.
Hence, substituting the values (2) and (3) of P and S into formula
(4) we obtain
ma = mg — bv
or
a= (5)
m V b
It is natural to assume the acceleration to be positive if it is
directed downwards and negative if it is directed upwards.
Equation (5) relates the acceleration to the speed of motion which
we do not know yet. It is this equation which we shall have
to use to determine how the velocity of the moving body will
change in time.
Qualitative Solution of the Problem
As a result of the above discussion we have obtained equation
(5) for the velocity of a falling body. Now it is up to us to solve
the equation. For this reason the following arguments are of
a purely mathematical nature, although to help the reader see
the point we shall continue to treat the matter of the velocity
of a falling body.

9
 Equation (5) connects two unknown quantities: velocity and
acceleration. Assigning an arbitrary value to the acceleration we
shall, using equation (5), obtain a definite value for the velocity.
Therefore, our first impression is that one equation (5) is not enough
to find two values v and a.
However, this impression is erroneous. The acceleration of a body
is completely determined by the manner in which its velocity
changes with time. Therefore equation (5) contains not completely
arbitrary values, but related values v and a. This makes it possible
to solve equation (5). Studying the relation between the velocity
and the acceleration, we shall later on arrive at the concept of
the derivative.
We shall prove two properties of velocity which follow from
equation (5); these properties will make the process perfectly clear
(in the assumptions made above). In due course we shall also
obtain an accurate formula for the velocity.
property l I f initially the velocity v0 of a falling body is less than
— then in the course of motion the relation v < always
b b
holds. If; on the other hand, v0 > then v > will always be
b b
true.
Suppose this is not so. For instance, let v0 < and at a cer-
b
tain instant f, (i. e. f( seconds after the body started falling) let
the velocity exceed It follows that at some intermediate
b
moment (perhaps, on numerous occasions) the velocity was equal to
ma
—r—. Let t0 be the last instant (in the first tj seconds) when the
b
velocity was equal to so that the inequality v > — - held
b b
for the interval f0 —t t. It follows from formula (5) that during
all this time the acceleration a was negative. However, this
contradicts the statement that within the corresponding interval
of time the velocity has exceeded The resulting contradiction
b
proves that the velocity cannot exceed the value of
b
The case v0 > should be considered in the same way.
b
mg
property 2 I f v0 < ——,the velocity of a falling body increases

with time, all the time approaching the value if, on the other
b
10
hand, v0 > the velocity decreases steadily also approaching the
value -!*
b
Evidently, if, for instance, i>0 > the relation v > - will,
b b
as we know from Property 1, hold for the motion as a whole.
It follows from formula (5) that the acceleration will be negative
and therefore the velocity will steadily decrease.
Let us prove that in the course of time the difference i> — ~
b
will become less than any predetermined small quantity h (this
may be, for instance, made equal to 0.001 m/s). Therefore, consider
the instant
m
t*
hb
During the time from the start of the motion up to the instant
t * the velocity of the falling body has decreased from the value i>0

to a value not less than i. e. it has decreased no less than by

b
an amount
mg hb
Vo T~ ~ — t*
b m
It follows that at some intermediate instant the acceleration
was not greater than — , for should the acceleration have been
m
greater during the whole interval of time the velocity would have
decreased more than by the amount — f*.
m hb
So let us assume that at the instant t' I a I <
In accordance with (5) it follows that
mg m ■ , m hb
a < - r ------ = h
~b b m
i. e. at the instant t' the difference between the velocity and the
value is less than h. This will hold for all future instants as
b
well, for the velocity continues to decrease always remaining greater
mg
than

11
 It should be noted that we have thus proved a somewhat more
precise statement than Property 2, namely, that not later than
after the time

t* 6
( )

after the fall has started the different e between the velocity and
the value will be less than h.
b
Properties 1 and 2 provide us in a certain sense with a solution
of the problem we have set ourselves. Although we have not yet
obtained a precise formula for the velocity, we have, nevertheless,
found out qualitative laws for the variation of the velocity, i. e. we
know now how it is going to change with time.
For instance, let us observe the motion of the parachutist.
If he opens his parachute the moment he jumps, his velocity will
fflQ
increase from zero, but shall never exceed the value —r—. The
b
value mg (the weight of the jumper with the parachute) is known,
and b depends on the parachute’s diameter. This makes it possible
to calculate the dimensions of the parachute so that the maximum
possible speed of descent equal to guarantees a safe landing
b
for the parachutist. On the other hand, in case of a free-fall jump
(with the parachute tucked away) the coefficient for the drag will
have a new value, which we shall denote by b' and which is
smaller than in case of a descent with an open parachute For
this reason the maximum speed - is greater in the latter case.
b
It follows that before the parachute is opened in the course
of a free-fall jump the parachutist’s speed will be greater than
and, according to Property 1, after the parachute is opened
b
mg
it will decrease approaching at the same time remaining
b
greater than Thus in this case, too, some time after the
b
parachute has been opened the jumper is able to make a safe
landing.
Let us take a numerical illustration.
Example I. Let a parachute be designed so that a jumper’s speed
with the parachute open approaches the limiting value of 6 m/s.

12
i. e. = 6 m/s. The jumper opens his parachute while falling
b
at a speed of 50 m/s. How much time will lapse before his speed
becomes less than 10 m/s, i. e. differs from the limiting value
- = 6 m/s by an amount less than h = 4 m/s?
b
mg
Solution. From the equation —;—= 6 m/s we obtain:
b
m mg 1 6 m/s
= 0.6 s
T b g 10 m/s2
Next using formula (6) we conclude that the speed of descent
wui
will differ from the limiting value —— = 6 m/s by h = 4 m/s not
b

later than after a time ^i>0 — ~ -jj- seconds, i. e. with the

assumptions made, after

(50 m/s —6 m/s) x 0.6 s x ----- —= 6.6 s

4 m/s

Formula for the Velocity

of a Falling Body.
The Number e

Properties 1 and 2 demonstrate how the velocity of a falling

body changes with time. In this section we intend to obtain
an exact formula for the velocity of a falling body. A certain
number enters into the expression for the velocity whose value
within the fifth decimal place is 2.71828.... Frequently used in
“higher” mathematics, this number is denoted by the letter e
(in the same way as another frequently used number 3.14159...,
which expresses the ratio of the circumference to the diameter,
is denoted by the letter it). Later we shall see why this number
e = 2.71828... enters into the formula for the velocity and how it
should be accurately calculated. For the present, we shall write
the formula for the velocity of a falling body without deduction
and shall consider some examples illustrating its application.
l£t v0 be the initial velocity of a falling body and v, the velocity
of the body at the instant t (i. e. t seconds after the body has

13
started falling). Then we have
mg
”‘ = - r + { ‘" - n r ) ' w

This is the exact solution of equation (5). The proof of formula

(7) will be given below. Let us consider some examples.
Example 2. Let us demonstrate that formula (7) immediately
leads to the qualitative laws for the variation of the velocity
(Properties 1 and 2) obtained above.
_±
The number e m, which is obtained by raising the number e
to a negative power, is, in effect, positive and less than unity,
_A _A, ( -AY
i. e. 0 < e m< 1. As t increases, e m = \e m) decreases (and
may be made for sufficiently large r’s as small as desired). Therefore
mq
it follows from formula (7) that, for instance, for v0 > —— the
b

velocity v, always exceeds the value ^for v0 — — > 0^ and

mg
that it decreases with time approaching — .
b
Example 3. Let us calculate, using formula (7), the velocity
of the parachutist 6.6 seconds after he opens his parachute in
a free-fall jump. We shall use the same numerical values as
mg
in Example 1 (p. 12), i. e. —— = 6 m/s, v0 = 50 m/s. (We have seen
b
that this velocity must be less than 10 m/s.)
Solution. We have
b mg , , . 5 1
— = g : —— % 10 m/s : 6 m/s = —----
m b 3 s
Next, using the tables of logarithms (the common logarithm of
the number e is approximately 0.4343), we easily obtain

logi , f\e- 1«>

') )= -■ k t log10 e % — - x 6.6 x 0.4343 =
m ~~ 3
= -4.7773 = - 5 + 0.2227 = 5.2227
whence
-A,
e ™ *0.0000167
14
Substituting this value into formula (7), we obtain
v6 6 s ~ 6 m/s + (50 m/s —6 m/s) x 0.0000167 a 6.000735 m/s
Similarly, with the aid of formula (7), it may be easily calculated
that the parachutist speed will be 10 m/s (the conditions being
31og10 11
the same) t = ——----------s a 1.44 s after the parachute opens.*'
5 logioc
Thus, when the parachute in a free-fall jump is opened, the speed
of descent decreases in 1-2 s from 50-60 m/s almost to the normal
speed of 6-7 m/s under an open parachute. Here the parachutist
slows down considerably, i. e. undergoes the great force of the
upward jerk of the parachute, caused chiefly by air resistance. One
who has watched free-falls (at aviation holidays, for example) may
have seen how a person rapidly plunging to the ground, suddenly,
the moment the parachute opens, slows down sharply; it may even
seem that for an instance he dangles in midair.
Example 4. Let a parachutist’s rate of descent in a free-fall
jump approach the limiting value of = 50 m/s. We shall set
b
the speed at the start at zero. What will the error be if instead
of formula (7) we use formula (1), applicable for a body falling
in a vacuum?
Solution We have
_ A = _ * _ 10 m/s2 = _ o.2 -
m mg ~ 50 m/s s
~T
Thus, in accordance with formula (7), the parachutist’s rate of
descent will be
v, = 50(1 - e - 02‘)
We obtain the following value of the speed of a body falling
in a vacuum from formula (1).
v, = gt % lOf
Thus the ratio of the velocities becomes
50(1 —e~0 2') _ 5 _ - 0 2 l\

gt * t
*' The speed of descent will in fact approach the limiting value of
6 m/s still more rapidly since expression (3) for the drag is really valid
only for small velocities. For greater speeds the drag’s dependence
on speed is greater than bv

15
 Putting t = 1 s, we obtain (after some easy calculations with the
aid of the logarithmic tables) for the above ratio the value x 0.91,
and for t = 2 s —the value * 0.82. We see that already the first
few seconds after the body has started falling, its speed (because
of the drag) differs noticeably from the value gt.
Now let us prove formula (7). Therefore, let us first elucidate
the relation between the velocity and the acceleration. If v, is the
velocity of the body’s motion at the instant f, and vt+h its velocity
h seconds after this instant (i. e. at the instant t + h), the ratio
V, +h ~ v,
----------- is termed the average acceleration of the body in the
interval h and denoted by aav:

If h is very small (say, 0.01 s, or smaller, depending on the

nature of the body’s motion) the acceleration changes little during
such a small interval, so that aav differs little from the value a,
at the instant t. The difference between a, and uav is the smaller
the less is h. In other words, should we put h progressively
smaller (say, 0.1 s, 0.01 s, 0.001 s, etc.), leaving t constant, aav
would change approaching a, closer with every step. In mathematics
this fact is written down as follows:

a, = lim aav lim Vt +h - v,

h->0 /l—o h

The sign lim means the limit of the expression which follows
it (i. e. the expression «av); the designation h->0 below indicates
that we have in mind the limit uav when h tends to zero.
Thus we have obtained the relation expressing the dependence
of the acceleration on the velocity. Let us now prove three
additional properties of the velocity of motion we have been
discussing. These properties will help us to prove formula (7).
PROPERTY 3 I f the velocity and the acceleration of a moving body
satisfy relation (5), this means that further variation of the velocity
is uniquely determined by the initial velocity v0.
Suppose the contrary is true. Let two bodies T and T* with
equal values of m and b move in such a manner that their
velocities and accelerations satisfy relation (5). Suppose at the
instant t = 0 these bodies had equal initial velocities v0, but fj
seconds later their velocities became different, say the velocity
of the first body t’x became greater than the velocity of the second

16
 nui
body t>f. To be definite, suppose that t>0 > —r— (should the op-
b
posite hold, the proof would be quite similar). Let t0 be the last
instant (during the first seconds) when the velocities of both
bodies were equal. In that case in the period tj —t0 the velocity
of the first body v was all the time greater than the velocity
of the second body v*, i. e. v > v*. Hence it follows that
mg mg
v— > v"
b b
ttm mg
both quantities v ---- and v* — being positive since
mo
d0 > — (see Property 1). In accordance with formula (S) we
b
conclude from the inequalities
mg * mg
v ---- 2- > v * ---- > 0
b b
that the accelerations of both bodies a and a* are negative, the
magnitude of a being greater than that of a*. This, however, means
that the velocities of the bodies T and T* have decreased during
the period rt —f0, the velocity of the body T having decreased
more than that of the body T *, i. e. at the instant it the velocity
v should be less than p* (because at the instant t0 these velocities
coincided). We, however, made the opposite assumption. The result­
ing contradiction proves that Property 3 is correct.
p r o p e r t y 4. I f two identical*' bodies T and T* start falling
simultaneously with initial velocities v0 and vg, their velocities v, and
vf at any instant t will satisfy the relation
mg mg
b b
( 8)
mg mg
»« b v0 ~

To prove this consider an imaginary body T which moves so

that at the instant t its velocity is equal to

vt = qv, + (l -<?)-—

In the sense that their respective m's and b's are equal.
where

<1= mg
v° - - v

We shall demonstrate that the velocity and the acceleration of

this imaginary body will satisfy relation (5). Find the average
acceleration <3av of this imaginary body inside the interval t, t + h.
We have
mg mg
(ivl+h + (1 - g) qv, + (1 - q)
v,+h ~ v, b b
a»v =
h
Vt+h - V,
= g
h
= <7«av

where «av is the average acceleration of the body T inside the same
interval. Should we take progressively smaller values of h in the
relation <5av = qaay, <5av would approach the acceleration ci, of the
imaginary body at the instant t. and «av the acceleration a, of the
body T at the same instant. In this way we shall obtain a, = qa,
(for an arbitrary instant t) and with the aid of equation (5)
b mg
a = qa = —q —
m b
b_ mg mg
qv + (1 - q)
m b b
i. e. relation (5) is satisfied for the imaginary motion we have been
discussing above.
Next the initial velocity of the imaginary body T is equal to
, mg / mg
v0 = gv0 + (1 - q)— = q I i)0 - ~y ~

m g\ mg
b ) b

18
Thus both bodies T and T* have identical initial velocities and
both move so that their velocities and accelerations satisfy equa­
tion (5). In accordance with Property 3 it follows that their
velocities D, and vf coincide at every instant t, i. e.

v? =qv, + (I - q ) —

Hence we obtain
mg , my my
qv, + (1 - q ) -^ -
b_ b
= g =
my my
v,
b T
my
vo*
_b_
my
»o b
thereby proving Property 4.
PROPERTY 5 For any instants t and t the relation holds
my my my
vl + T v, -
b
(9)
my my my
Vo b
where v0, vx, v„ vl+x are the velocities of the falliny body T at the
instants 0, t, f, t + t.
Indeed, start observing how the body T is falling from the instant
x onwards. In t seconds (i. e. t + x seconds after the motion has
started) the velocity will be i>(+t. This means that if at the instant
t = 0 we in addition to the body T drop another body T* whose
initial velocity v* equals vx, the velocity of the latter body r(* at
the instant t will be equal to u(+t, i. e. vf = u, +t. Hence we obtain
from (8)
mg mg
Vt+z vx -
b
my mg
v,
b Vo -

2* 19
or otherwise
mg
b
Dividing both parts of the equation thus obtained by
, we obtain the required relation (9).
After deriving formula (9) let us start to calculate the exact
value of the velocity v,. In order to avoid lengthy formulas let us,
as a temporary measure, introduce the notation
mg
v, -
b
u, =
mg
v0 -
b
thus simplifying formula (9):
u,+t = u,ut ( 10)

For x = t formula (10) yields

“ 2 . = “.2
Similarly, putting x = 2t. we obtain from (10)
u3, = u,u2, = u,uf = uf
and for x = 3f the result will be
= u,u3, = u,u? = uf
etc. Continuing in the same way, we shall find that for any
positive integer n the relation holds
uM= uJ (11)

Setting in this equation t = ~ s and extracting the root of the

n-th power, we obtain
i

n
Next setting in equation (11) t = 1 s and substituting p for n, we
obtain
Up = uf
It follows from the last two equations that

20
 Thus, for any positive rational number t (i. e. the number of
the form -jj-, where p and n are positive integers) it holds

u, = u\
or returning to the previous notation

mg
b
( 12)
mg
”o -

Here vt is the velocity of the falling body at the instant t = 1 s.

Inasmuch as relation (12) is valid for rational t's it is valid for
all t's.
Let us take, for example, the instant t = J/2 s = 1.414... s.
The numbers 1.4, 1.41, 1.414, etc. are rational and for this reason
relation (12) holds for these values of t:

Should we set t equal to rational values each of which is

progressively a more exact approximation of the number j/2
the left-hand side of equation (13) would approach the limit

mg mg V*
72- V, -
b *>
and the right-hand side the limit
mg
v° - ~ r v0 ~

** See, for example, the concept of an irrational power in the course

of algebra by A. P. Kiselev, part II (p. 97 in the 1960 Russian edition).
21
Thus we obtain in the limit

Similar treatment is, of course, applicable not only to j/2, but

to any irrational value of t. Hence, relation (12) is valid for any t.
Introducing the notation
mg
t>. —

mg
v0 ~
b
we obtain from (12)
mg
v, -
mg
v0 b
whence

Formula (14) obtained above for the velocity of a falling body

is not the final one since we do not know the value of the number
c which enters this formula. To calculate the number c, let us
find from formula (14) the acceleration of the body at the initial
instant of its motion. The average acceleration for the first h seconds
has according to (14) the following value

vh - v 0 Tb + \ b J (/ m g \ c 1' - 1
= --------------- h--------------- = r “ T j ~
As h tends to zero, this expression gives us the acceleration a0
at the initial instant:

a0 = lim a3t = lim (v0 - ) C 1 (15)

)i-.o t-oy b J h
Denoting the expression ch — 1 by x, we obtain
ch — 1 = x, ch = 1 + x, h = log,. (1 + x)
22
Thus instead of the expression fv 0 — - r - \ ------- which enters
V b J h
(15) under the limit sign, we obtain the expression

x
log, (i + x) _L
— logc(l + x) logc (1 + X)*
x
Note that as h tends to zero the limit of the number ch is
unity and that the number x = ch — 1 approaches zero. Hence we
may write:

a0 = lim (16)
*-o i
logc(l + x)*

The limit of the expression (1 + x)' as x tends to zero is termed

the number e. We shall leave without proof that this limit exists,
_ i_

i. e. that the expression (1 + x)x does in fact approach some value

as x -»0. The proof (by the way, quite elementary) of the existence
of this limit may be found in the first chapters of any textbook
on higher mathematics. *’ t
We shall confine ourselves to calculating the values (1 + x)* for
x= 0.1, x = 0.01, x = 0.001, and x = 0.0001. The results of such
calculations are presented below (these calculations may be
performed with the aid of logarithmic tables, preferably seven-
place ones; one may also use Newton’s binomial formula):

(1+ 0.1)01 = l .l 10 k 2.59374

i
(1 +0.01)001 = l.Ol100 « 2.70481
i
(1 + 0.001)0001 = l.OOl1000 « 2.71692
_i_
(1 + 0.0001)00001 = 1.000110000 w 2.71814

*’ See, for instance, V I. Smirnov, Kurs vysshei matematiki (A Course

in Higher Mathematics), V. 1

23
These calculations clearly demonstrate that as x -»0 the limit
X
of the expression (1 + x)x is e = 2.71... . We now obtain from (16)

1/0 b
«o —;--------
lo&e
On the other hand, it follows from (5) that

Equating the expressions obtained for a0, we obtain

m
v0 ~ b mg

whence
lofcc b )
m b
i m e as b C
lofce--T , C

__6
Finally substituting e m for c into formula (14) we obtain formula
(7) thus concluding the proof.
 Differentiation

The Concept o f the Derivative

Equation (S) as we have just seen may be solved quite accurately.

This equation relates the quantity v (the velocity of a falling body)
with the quantity a which shows how rapidly the quantity v changes
(acceleration is “the rate of change of speed”).
When we talk of the rate of change of some quantity, we assume
that the quantity we are dealing with is not a constant characterized
by a number, but a variable quantity, i. e. a quantity whose value
changes with time. Here are some examples of quantities that
change with time (or depend on time): the velocity and the accelera­
tion of non-uniform motion, the magnitude of alternating current,
etc.
Let y be a quantity whose value changes with time. For the sake
of convenience we shall denote by y, the value this quantity
assumes t seconds after the process under observation has started.
The difference y,+» - y, shows how much the variable y has changed
in h seconds (in the interval t, t + h seconds after the process has
started). The ratio

for its part shows how much y changes on the average every
second (within this interval of time), i.e. this ratio represents
the average rate of change of the variable y. Putting h progressively
smaller, we shall obtain the values of the average rate over
progressively shorter intervals starting from the instant t. As a limit­
ing case (when h tends to zero) relation (17) yields the rate of
change of the variable y at the instant t. We know already that
this rate of change is designated in mathematics by
tt+» - y,
lim ( 18)
fc-0 h
25
 Expression (18) is termed the derivative of the quantity y with
respect to time t, and as we have seen it represents the rate
of change of the variable y. We may deal with a variable which
changes not with time, but depends on some other quantity.
For instance, the area of a circle depends on its radius. Denoting
the area of a circle of radius R by SR, we obtain
SR = kR2 (19)
Should we investigate the dependence of the circle’s area on its

radius we would arrive at the ratio , which expresses

the mean rate of change of the area with the change of radius.
The limit of this ratio (as h -> 0) is the derivative of S with
respect to R.
The concept of the derivative is one of the basic concepts of
higher mathematics. In cases when the variable y changes in
accordance with variations of a quantity x (or when, as may also
be said, y is a function of x) the derivative of y with respect
to x is denoted by y', or more often, by one of the symbols
dy d
dx ’ dx

£ = lim yx+h - yx
dx h->o h

here the letter d should not be cancelled out since it is not

a multiplicand but denotes the process of finding the derivative,
or, as is also said, the differentiation.
dS
Let us by way of example calculate the derivative of func­
tion (19):
dS .. SR+h — SR n{R + h)2 —nR2
-= lim ------ = lim
dR h->o h h- o h
- lim (2nR + nh) = 2ttR
/l —o
i. ,e. the derivative of the circle’s area with respect to its radius
equals its circumference.
ds
For the second example let us find the derivative —f- of the
dt
distance with respect to time. Denote the distance covered by

26
a body up to the instant t (i. e. t seconds after it has started moving)
by st. Then the ratio b' will represent the average speed
during the interval t, t + h, the limit of this ratio for h -> 0 being
the value of the speed at the instant t:
ds
v, = lim
/1-.0 ~di
dv
In the same way one may calculate the derivative — . The ra­
tio
v, +h ~ v,

is the average acceleration during the interval t, t + h, the limit

of this ratio beihg the value of the acceleration at the instant (
(compare with what was said on p. 16):
dv
= lim v, +h - Vt
(i-*0 h ~dt
The relations that have just been proved

( 20)

and

( 21)

play an extremely important part in mechanics.

The Differential Equation
Let us again turn to equation (5). According to (21) this equation
may be rewritten in the form
dv b( my \
-dr= - - ^ { v- - r ) (22)
Now it is evident that this is an equation with one unknown
quantity v. However, this is not an algebraic equation, but an
equation relating the quantity v to its derivative. Such an equation
is termed a differential equation. Comparing differential equation
(22) with its solution (7) and denoting — by k and by c, we
tn b
may make the following statement.
27
 THEOREM. The solution of the differential equation
dv
= —k(v — c) (23)
~dt
is represented by the expression
v = c + (v0 - c) e~ta (24)
where v0 is the initial value of v (i. e. its value for t = 0).
In future we shall be able, using this theorem, to calculate some
other physical phenomena, as well.

Two Problems
Leading to Differential Equations

(a) Switching on an Electric Current Consider an electric

circuit consisting of a coil and a battery (Fig. 2). The electrical

Fig. 2

properties of a coil are rather complex, but for some cases they
may be described with great accuracy with the aid of two quantities:
its resistance and its inductance. The coil is in fact usually represented
by two sections connected in series: by the resistance and the
inductance (Fig. 3). The voltage drop across the resistance is
proportional to the current i passing through the coil (Ohm’s Law)
V = Ri
2X
 Coil

r - n n r — ^w o it h

1 11 Switch J
Bottory T

I--- #|s—I Fig 3

The proportionality factor R is termed the resistance of the coil.

The voltage drop across the inductance is proportional to the rate
of current variation. Denoting the current variation rate by w
(it is measured, for example, in amperes per second) and the propor­
tionality factor by L, we obtain for the voltage drop the expression
V = Lw
The quantity L is termed the inductance of the coil. The voltage
drop across the coil is the sum of voltage drops across the
resistance and the inductance, i. e. it is expressed by the formula
V = L w + Ri (25)
Formula (25) is well supported by experiment (if the frequency
of the current passing through the coil is not too high). This is
the formula we are going to use. Denote the electromotive force
(emf) of the battery by E. Equating the emf of the battery to the
voltage drop across the coil, we obtain on the basis of KirchhofFs
Second Law the equation (we neglect the internal resistance of the
battery and the resistance of the conductors):
E — Lw + Ri
or

The solution of this equation may be easily obtained from

the theorem formulated on p. 28. Indeed, denoting the current at
the instant t by i„ we may contend that the quantity

is the average rate of change in the current during the interval

f, t + h. When h -»0 we obtain the rate of the current change

29
at the instant t
di
w = lim i,+h - i,
fc-0 h It
Hence the quantity w is the derivative of the current i, and
equation (26) may be rewritten in the form
dj__ _ _ R / . _ _ E
d t~ L 11 R
The only difference between this equation and equation (23) is
that the variable we are seeking is denoted by i and not by v
which is, of course, inessential. The constants k and c contained
in equation (23) in this case assume the following values

Hence the solution of this differential equation will be of the form

(see (24)):
E (■ E\ - f '
'< = - 5 T + « o - -5 T *

Should the current i0 the instant the battery was switched on

(t = 0) be zero the formula would assume a simpler form

It follows from this formula that the current, zero at the

instant of switching on, grows continuously approaching the value
£
— , i. e. the value of the current which would pass through the
R
coil if it had the same resistance R and no inductance at all.
(b) Radioactive Decay. Suppose we have a piece of rock
which contains some radioactive material. The atoms of the radio­
active material decay and are transformed into another chemical
substance —the product of decay. Hence, in the course of time
the amount of the radioactive material contained in the piece
of rock decreases. Let us introduce the concept of the decay rate.
Suppose at some instant t the amount of the radioactive material
contained in the rock had been m, grammes and h years later
it decreased (because of the decay) to mt+h g. The expression
m, +h - m,
h

30
shows by what amount, in grammes per year, the mass of the
radioactive material has on the average decreased (during the
respective period of time). It would be natural to term this
expression the average (mean) decay rate for the period. The limit
to which this expression tends as h -*■ 0 is the decay rate at the
instant t. Denote it by u
m,+h - m, dm
u = lim
h ~dt
Note that the decay rate is negative since the mass of radioactive
material decreases with time.
What does the decay rate depend on? When the amount of
radioactive material contained in the rock is small, it may be
accepted that the decay rate is proportional to the amount of radio­
active material contained in the piece at a certain instant, i. e.
u — —km
where m is the mass of radioactive material and k a positive
constant (proportionality factor).
The arguments in favour of approximate validity of this law are
simple if one assumes that the decay of some atoms of radioactive
material does not affect the state of the others. In these conditions
it may be assumed that the amount of radioactive material decaying
per unit time out of every gramme is always approximately the
same, say k grammes, no matter how much radioactive material
still remains in the rock. In this case the amount of radioactive
material decaying per unit time out of m g will be km g.
Is the assumption that the decay does not affect the state of
the remaining radioactive atoms correct? One should bear in mind
that the decay products may strike another atom, thereby causing
it to decay; new particles will thus be generated —this may lead
to the decay of further atoms, etc. Such a chain reaction (this,
for example, is the basic process in the functioning of the atom
bomb) would not fit with the assumption of the independent decay
of the atoms. To make this chain impossible, the particles emitted
in the course of the decay should be lost and should not (in most
cases) strike other radioactive atoms. This will be the case if
radioactive material makes up only a small percentage of the mass,
the bulk of the rock being non-radioactive. Then the overwhelming
majority of the particles emitted in the course of the decay will
be lost in collisions with non-radioactive atoms of the rock, and
a chain reaction will not be possible. Therefore, for small amounts
of radioactive material contained in the rock the decay of the
31
individual atoms may, to a good approximation, be assumed to be
independent.
We obtain accordingly the following differential equation for
evaluating the mass of the remaining radioactive material
dm

which differs from equation (23) in that the unknown quantity

is designated by the letter m instead of v, and that the constant
c is in this particular case zero. Hence, according to (24) the
solution will take the form
m, = moe"b (27)
where mo is the mass of the radioactive material at the initial
instant (when we begin to take interest in the process of decay).
Example 5. How many years will it take for the amount of
the radioactive material to decrease by half?
Solution. To obtain an answer to this problem, we should
determine t from the equation
m^e ” = — mo
Cancelling out mo and taking the logarithm, we find
0.69
t
¥ log*2 k
This period is termed half-life of the corresponding radioactive
material. Note that this period does not depend on the initial
amount of the radioactive material, but only on k, i. e. on the
nature of the material. For instance, half-life of radium is
1590 years, that of uranium 238 - about 4.5 thousand million years.
Example 6 . Formula (27) permits us to draw certain conclusions
concerning the age of the Earth.
Suppose a piece of rock extracted from the depths of the Earth
contains together with some inclusions m g of the radioactive
material and p g of its decay product. Suppose in addition that
every gramme of this radioactive material yields (upon complete
decay) r g of the decay product. This means that p g of the decay
product was produced from y g of the radioactive material.
Hence if we assume that the radioactive decay process started at
some instant (i. e. up to that instant the material had not contained
a single atom of the decay product), we must put the initial mass
of the radioactive material equal to m + —. To calculate the time
r
that has passed from that imaginary instant (since the beginning
32
of the decay) until today, we must in accordance with (27) solve
the equation

with respect to t, whence we obtain

' “ T k»-(1 + ™)
Such calculations performed for some minerals contained in the
Earth yield approximately the same value for t — around
2 x 109 years. Thus the conditions on the Earth that permitted
a normal decay process persisted for several thousand million years.
It may be conjectured that some thousand million years ago
the matter which now makes up the Earth existed under quite
different conditions which facilitated the production of radioactive
atoms from the simpler atoms and from other particles.
Napierian Logarithms
The formulas for solving the problems presented above contain
the exponential function with the base e. When making calcula­
tions based on these formulas with the aid of logarithmic tables,
one can exclude some operations if one uses logarithms with the
base e. For instance, taking the logarithm in formula (27) with
the base e and with the base 1 0 , we obtain
log* m, = - kt + log, mo, logi0 m, = - k t log10 e + log10 mo
In the second instance an extra logarithm has to be taken and
an extra multiplication performed. Moreover, problems result
in forihulas which contain logarithms with the base e as we have
seen in Examples 5 and 6 . The number e comes up in other
mathematical problems as well, and it is very convenient to use
logarithms with the base e, especially in theoretical questions.
logarithms with the base e are termed Napierian, or natural, loga­
rithms and denoted by the symbol In: the meaning of the expression
In x is the same as that of log, x. Common and natural logarithms
are related by the expression
logio x = M ■In x
where
M = logio e « 0.4343
This relation is easily obtained if one takes the logarithm with
the base 1 0 of the identity etnx = x

33
 Harmonic Oscillations

The Problem of Small Oscillations

of a Pendulum
Suppose a thread of length / is fixed at point C with a body
M suspended from its other end (a pendulum). The problem is
to find out the nature of the motion of the body M. To solve
this problem mathematically, we shall simplify it somewhat. First
of all we shall assume that the thread from which the pendulum
M is suspended is inextensible and weightless.
We shall examine the motion of the pendulum which takes place
in a single vertical plane passing through the suspension point.
Because the thread is inextensible we may be sure that the body
M will move along a circumference of radius / with the centre
at point C. The assumption about the negligible mass of the thread
means that its weight is negligible as compared with that of the
body M. This makes it possible for us to assume that external
forces act only on the body M. We shall assume the pendulum
M to be a heavy point (i. e. we shall attribute to it a mass m
and neglect its dimensions). From all the forces that act on the
body M we shall in addition to the tension of the thread take
into account only the force of gravity. When considering this
problem, we may also neglect the air resistance (for instance, we
may presume the body M to be confined to a hermetic evacuated
vessel). The problem of how much the motion of the pendulum
in the atmosphere differs from its motion in a vacuum is treated
in the note on p. 49.
Let the body M at a certain instant be at point A on the
circumference along which it is moving. Denote the lowermost
point of the circumference by Q, the length of arc QA by s, and
the central angle QCA corresponding to it (Fig. 4) by a
(in radians). Then
s = /a (28)
We shall assume the arc s and the angle a to be positive when
point A is to the right of Q and negative in the opposite case.

34
 c

Q K ig 4

Now let us deduce the equation from which we are to learn

the law of the pendulum’s motion. The difference in altitude of
points A and Q is

h = QB = CQ — CB — I — I cos a = / (1 — cos a) = I ■2 sin2

Assuming the potential energy of the pendulum to be zero at

point Q, we find for the potential energy of the pendulum at point
A the value

Wip) = mgh = mg2l sin2

The kinetic energy of the pendulum is

w <u =

where v is the velocity of the body Af.

Hence the total energy E of the pendulum (at point A) is
expressed by the formula

(29)

The pendulum does not perform any work in its motion (we
neglected friction and drag!), and for this reason its energy E
remains all the time constant.
We shall somewhat simplify equation (29) by treating the
problem of small oscillations of the pendulum, i. e. of such motion
of the pendulum in the course of which the deflection angle from
its equilibrium position Q remains small. Let us clarify what we
mean by “small angles”. The point is that an exact solution of

35
equation (29) cannot be written with the aid of any known opera­
tions. Therefore the question arises as to whether equation (29)
may be replaced by a simpler one. Naturally this simplification
should be such that the solution of the simplified equation would
be a very accurate approximation to the solution of equation (29).
Note that hereby we do not in principle introduce a new inaccuracy
since relation (29) is in itself approximate*’, so the applicability
of an approximation of a definite sort depends only on the degree
of accuracy we want to obtain.
The usual simplification to which equation (29) is subjected is
the substitution of an angle (p for sin (p. That such a substitution
is permissible for small angles (p will be seen from Fig. 5, which

shows an arc A'Q'B' of the circumference of radius C'Q' = 1;

the angle (p is plotted on both sides of radius C'Q'. The length
of section A'B' is 2 sin (p (for A'S' is the sine line), the length
of arc A'B' being 2 9 (the angles are, of course, measured in radians!).
It is clear from the figure that small angles 9 differ little, this
difference being the less the smaller is 9 . For instance, it may
readily be checked with the aid of trigonometric tables that
for angles not exceeding 0.24S radian (i. e. « 14°) the ratio

*’ In deducing equation (29) we have made some simplifications:

neglected the drag, the weight of the thread, the dimensions of the
body M, etc.
Note that any physical law or any mathematical relation linking physical
quantities (for example, relations (1), (2), (3), (4), (5), (25), and (29)) is
an approximate one, for in every case there are “forces” which were not
taken into account when the physical law or mathematical relation was
formulated. This, however, does in no way belittle the enormous importance
of physical laws. For instance, the accuracy of Newton’s Second Law
and of Ohm’s Law is under ordinary conditions extremely great.

36
sin (p
^ differs from unity by less than 0 .0 1 ; for angles less than
1° (0.017 radian) this difference drops to about 0.0005.
Accordingly, assuming the deflection angle of the pendulum
a a
to be small, we substitute — for sin —, i. e. we substitute a new
equation for equation (29) which differs little from the former:
mv2 „ . / a V

Taking into account (28), we may write this equation in the form
mv mgs“
= E
21
or
/ , , 2IE
— tr + s2 (30)
9 mg
This equation contains two unknown variables: s and v (we
assume that constants g, I, m, and E are known). Yet this equation
(same as equation (5)) may be solved, for quantities s and v are
not arbitrary but related by expression (20). It follows from (20)
that equation (30) may be written in the form

so it is, in effect, an equation of one unknown. Let us solve this

equation.
Let us choose a coordinate system lying in a plane with the

value of s measured along the x- -and the value of ^ ~ v al°ng

the y-axes. At any instant t there is a definite value of the distance
s and the velocity t> of a body M, i. e. a definite point N in the
plane (Fig. 6 ). Conversely, if we know where point N lies, we can

37
find its coordinates s and / — v, i. e. we can learn the position
' 9
of the pendulum and its velocity. Thus at any instant t the
pendulum M is symbolized by some point N. The length of section
ON is easily calculated with the aid of the Pythagorean theorem:

ON = ]/P N 2 + OP2 + s2
i. e. (in accordance with relation (30))
'Hm
ON =
mg
As the pendulum swings, the values s and v will change, i. e.
point N will move across the plane which contains the coordinate
system. But the distance from point N to the origin of coordinates
,f 2lE
always remains the same, i. e. equal to Hence point N
mg
moves along a circumference of the radius
2IE
R= (32)
mg
This circumference is termed phase circumference.
Let us find the velocity with which point N moves along
the circumference. This velocity is directed along a tangent to the
circumference. Let it be represented, for example, by a vector NA
(Fig. 7). Let us decompose the vector NA into a horizontal and
a vertical components. Then a horizontal component NB will
represent the velocity of motion of point P along the x-axis.
The distance from point P to 0 being s, the velocity of motion
. • „ . ds
of point P is ——= v, l. e. NB = v.
v dt

38
 Now since triangles ONP and NAB are similar, we obtain

PN : ON = NB : NA or ~ v : R = v:N A
9
From the latter proportion we obtain

NA = R

Such is the velocity with which point N moves along the circum­
ference.
Denote the initial deflection and the initial velocity of a pendulum
by s0 and v0, respectively, and the corresponding point on the
phase circumference by N0. Then the radius of the phase circum­
ference will be

R= — Vo + So (33)
9
(see (30) and (32)). An angle <p0 = /i XON0 will be determined
from the relation

so
(Fig. 8). Next, t seconds after the motion of the pendulum has

N0N = R -j-1 along the phase circumference, and an angle

39
^ N0ON will accordingly be | / f t . As a result (Fig. 9)

Fig. 9

<p = ^ XON /i XON0 - ^ N0ON =

■ < P o -]/ft
Hence we obtain

OP = R cos (p = R cos ^(p0 - = ^ c°s 1 ~ <Po^

PN = R sin (p = R sin ^(p0 — ^ = —R sin 1 ~ <Po^

Remembering finally that OP = s and PN = / — v, we obtain

If 9

s = R c o s ( j / ^ f - ( p 0) , v = - | / ^ R s i n ^ | / ^ t - ( p 0^ (35)
These formulas express the deflection and the velocity of the
pendulum t seconds after it has started to swing, i. e. they represent
a complete solution of problem of pendulum's motion (with the
simplifications we have made). Let us consider some examples.
Example 7. The pendulum is initially deflected to the right
to a distance s0 and dropped with a zero initial velocity. Find
its deflection and velocity at the instant t.
Solution. In this case R = s0, <p0 = 0, and from formulas (35)
we obtain .— _ ,—
s = s0 cos l/- y t, v= - / — s0 sin l / — f

Example 8 . The pendulum was initially in the equilibrium

position Q and was thrown out of this position by a thrust which
imparted to it the initial velocity i>0 directed to the right (i. e.

40
a positive initial velocity). Find the deflection and the velocity
at the instant t.
Solution. From formulas (33) and (34) we obtain in this case

that R = j / 1 , 0 , <Po = -y. and from formulas (35):

Example 9. Let us find the derivatives of functions sin tot and

cos cut.
Since u is the time derivative of s, comparing the values of s
and v in Example 8 , we conclude that

Similarly, from Example 7 we obtain

^ ( so c o s | / f f ) = —| / T So « i n |/ r

Setting, for instance, 1 and denoting the quantity

’ v° = s° =

by (o, we obtain from these formulas:

sin a» = to cos (at, ——cos (Dt = —(Dsin (Dt (36)

dt dt
Example 10. Since cosine and sine are periodic functions
after some time T, termed the period of oscillations, the pendulum
again returns to the initial position and again repeats the same
motion. Let us find the period of oscillations of the pendulum.
Solution. The values of sine and cosine remain unchanged when
their arguments change by 2n. Therefore the period of oscillations
of the pendulum is the period T during which the expression
in parentheses following the sine and cosine symbols in equations
(35) increases by 2n. In other words, the difference between the

values of the expression t — (p0 at the instants t and t + T

41
must be 2n

~r (f + T) —(p0 — - t - i p0 + 2rc

From this relation T is easily obtained

7 = In (37)

In this way the motion is periodically repeated every T seconds.

The pendulum oscillates periodically. During each period (i. e. every
T seconds) the pendulum, as may be seen from (35), once assu'mes
the extreme right-hand position (cosine becomes equal to +1)
and once the extreme left-hand position (cosine becomes equal
to —1). At these instants of maximum deflection the velocity
of the pendulum turns zero (see (35)). When cosine turns ± 1,
the sine of the same argument turns zero. The maximum velocity
of the pendulum is attained (sine turns +1) when it passes through
point Q (cosine turns zero).
The Differential Equation
of Harmonic Oscillations
We have deduced formulas describing the motion of the pendulum
from equation (30) or, which is the same, differential equation (31).
There is another differential equation which, too, describes the
motion of the pendulum we have been discussing above. Its deduc­
tion is very simple.
Suppose a body JVf at some instant is at point A of the
circumference along which it moves. Let us decompose the force
of gravity (which we assume to be equal to mg and to be directed
vertically downwards), using the parallelogram rule, into two
components: one tangent to the circumference at point A and
the other perpendicular to this tangent. The component which is
perpendicular to the tangent tends to stretch the thread and is
compensated by the tension of the thread (for the thread is presumed
to be inextensible). A force F acts along the tangent in the direc­
tion of point Q and is equal in value, as may readily be seen,
to mg sin a (Fig. 10), i. e. the force F is negative for positive
a’s, and vice versa. Thus
F = —mg sin a
If both mutually compensating forces —the tension of the thread
and the component of the force of gravity perpendicular to the
tangent —are ignored, there remains only the force F acting on the

42
 c

F ig 10

body M (we neglected the air resistance), and we may therefore

write on the basis of Newton’s Second Law

a = —g sin a
Recalling that we are now dealing only with small oscillations
of the pendulum, and that this makes it possible to approximate
sin a by a, we may write this equation in the form
a = -get
or according to (28)

a + 4 -s = 0 (38)

This is just the required equation. Let us show how it may be written
in the form of a differential equation. It follows from relations
a = —r and v = —— that if we take a derivative of a distance s
dt dt
once and then take the second derivative of the first one obtained
(i. e. of the velocity), we shall obtain the acceleration. In other
words, the acceleration is the second derivative of the distance s
(with respect to time t). This is written down as follows

or
ad2s
s
a = —r^r (39)
dt2
43
 u S
The symbol (the second derivative of s with respect to t) is
not treated as an algebraic expression, but as a single symbol;
we are not allowed to perform any operations with its components
(for instance, to cancel anything out of the “fraction”). It follows
from (39) that equation (38) may be written in the form of
a differential equation as follows
d2s
s=0 (40)
dF
Note that we are already able to solve this equation. Indeed,
it describes the law of variation of the quantity s, i. e. the law
of the oscillations of the pendulum, and this we have already
investigated. Therefore we may state at once that the solution
of equation (40) is contained in the first formula of (35). We may
express this in more detail as follows.
The solution for differential equation (40) is

s = R cos

where R and (p0 are given by formulas (33) and (34). Note that
to find the quantities R and <p0 we must know the initial deflection
s0 and the initial velocity v0, i. e. the values of s and
ds
— taken at the initial instant.
dt

Denote a quantity |/ ^ - by to. Then we may express the state­

ment made above in the following form.

THEOREM. The differential equation

=0 (41)

has the solution

s = R cos (tot - <p0) (42)
ds
where R and (p0 depend on the initial values of s and — .
dt
Equation (41) is known as the equation of harmonic oscillations.
Any quantity described by this equation is said to be oscillating
harmonically; this means that the respective quantity changes with
time in accordance with the law (42). The quantity co which enters

44
differential equation (41) and its solution (42) is termed angular
2n
oscillation frequency, and T = oscillation period. When the
quantity s oscillates harmonically, any value it assumes is repeated
every T seconds (see Example 10).
Let us compare differential equations (23) and (41). Equation (23)
contains only the first derivative, therefore it is termed a first-order
differential equation. Equation (41) is a second-order differential
equation because it contains the second derivative. Note that only
the initial value of the quantity v itself had to be known to solve
the first-order equation (23). To solve the second-order equation
(41), on the other hand, we must know not only the initial value
ds
of the quantity s itself, but of its derivative — as well. In short,
to solve a first-order equation we must know the initial value of
only one quantity; to solve a second-order equation, of two.
Note that we have obtained the solution of equation (40) from
physical considerations': both equations (31) and (40) describe the
same physical phenomenon, and for this reason they should have
identical solutions expressing the law of pendulum’s oscillations.
Of course, such argumentation is guesswork and not a rigorous
mathematical proof. It may be proved by purely mathematical
means that equations (31) and (40) are equivalent, i. e. that then-
solutions are the same: after differentiating both sides of equation
(31) we obtain equation (40). Conversely, one may obtain equation
(31) from equation (40), but then one will have to use an operation
reverse to that of differentiation. Such an operation (termed
integration) forms together with differentiation the basis of higher
mathematics as a whole. A more detailed discussion is beyond
the scope of this booklet.
Still, by applying formulas (36), the reader will easily find the
second derivative of function (42) and make sure for himself that
the function satisfies equation (41).
Let us consider two examples from the field of physics which
lead to the equation of harmonic oscillations.

The Oscillatory Circuit

Let us consider an oscillatory circuit, i. e. a closed electric circuit
consisting of a coil and a capacitor. The coil possesses an inductance
and a resistance (see p. 29). The complete circuit may be represented
in the form of a circuit diagram (Fig. 11). Let us denote by q
the electric charge which goes over from one plate of the capacitor

45
 Coil

R L
Switch /
Capacitor

F ig 11

to another and by i the current in the circuit. (We presume that

initially the capacitor had a charge q0, and that a current i0 was
in the coil, and are interested in subsequent changes in these
quantities.) In that case the voltage across the capacitor will be
where C is its capacitance; the voltage drop across the coil
will be equal to Lw + Ri, where R is the resistance and L the
inductance (see (25)). According to KirchhofiPs Second Law the sum
of the voltage drops in the circuit is zero, i. e.

Lw + /?/ + -— = 0 (43)

A quantity i is the derivative of q with respect to t. Indeed,

if the values of the charge q were q, and q, +h at the instants
t and t + h, respectively, this means that during that interval
a charge passed through the cross section of the circuit’s wire
(anywhere in the circuit) equal to q, +h — q,. Therefore the average
current over the interval t, t + h was
<?<+). - <7,

Taking the limit, we obtain

<l<+h - q , = dq
/ = lim
h-*0 h dt

It follows from relations i = and w = that w is the de-

dt dt
rivative of / = i. e. that w is the second derivative of q:
dt ’
d2q
W~~di2
46
Hence equation (43) may be rewritten in the form:
i + r d(i + « _ n (44)
Li F + R ^ r + ~ c-°
The differential equation we have just obtained is more complex
than equation (41), since in addition to the unknown function q
and its second derivative it contains also the first derivative
dt2
dq_
We shall not, however, try to solve equation (44) (see the
d t'
note on p. 49), but shall only discuss the case when the resistance
of the coil R is negligible (compared to the quantities L and C),
neglecting on this account the term R ~ in equation (44). Then
dt
the equation will assume the form:

i *dt2 + iC- o
or
d2q 1 A
(45)
1 F + lc«= °
Equation (45) is, evidently, an equation of harmonic oscillations
(see (41)) with the oscillation frequency of the circuit
1
oo =
\/lc
and the period of oscillations expressed by the formula
T = 2tc \/LC
The solution of equation (45) is of the form (see (42)):

q = R cosI ' ~ <Po

\]/L C
where R and <p0 depend on the initial values, i. e. on q0 and i0.
Oscillations Produced by
the Action of the Elastic Force
of a Spring

Suppose a weight of mass m is suspended from a spring.

The spring will be somewhat elongated under the action of the
force of gravity (the elongation will stop as soon as the spring

47
tension compensates the force of gravity); in this position the weight
and the spring can remain stationary (be in a state of equilibrium).
Should we shift the weight from its equilibrium position pulling
it downwards, the spring tension would exceed the force of gravity
and the resultant force would be directed upwards. Should we,
on the contrary, shift the weight to a point above the equilibrium
position, the resultant force would be directed downwards. Thus,
this resultant force “tries” to return the weight to its equilibrium
position.
For the sake of simplicity we shall content ourselves with only
the vertical motion of the weight: up and down. Denote the posi­
tion of equilibrium by 0, the position of the weight at some
instant by A, and a distance OA by s, assuming the positive
direction along the vertical to be downward from point 0, i. e.
we will take s to be positive if the weight (point A) is below point
0, and negative if it is above point 0. Denote the resultant force
of the spring’s gravity and tension by F, and the air resistance
by S. We shall assume that no other forces except F and S act
on the weight. In accordance with Newton’s Second Law we may
write:
ma = F + S

where a is the weight’s acceleration. The force F which tends

to return the weight to its equilibrium position is directly propor­
tional to a deviation s, i. e. it is numerically equal to the quantity
ks, where k is a proportionality factor. This assumption is well
supported by experiment (for not very great deviations from the
equilibrium position). The quantity k is termed the stiffness of the
spring. For positive s (point A is below 0) the force F is directed
upwards, i. e. it is negative, for negative s the force F is positive.
In other words, the sign of the force F is opposite to that of the
deviation s, i. e.
F = —ks

We shall assume for the force S the same value as above (see
(3)), i. e.
S = -b v
In this way we obtain the following equation of motion of the
weight
ma = —ks — bv
48
or
ma + bv + ks = 0 (46)

Since v = and a = — this equation may be written in the

dt dr
form
d2s ds
m -jT- + b — + fcs = 0 (47)
df2 dt
Differential equation (47) is analogous to differential equation
(44), which we obtained while solving the problem of the oscillatory
circuit. We shall not solve equation (47) (see the note below),
but consider only the case when the air resistance may be neglected
(i. e. when a quantity b is very small compared to quantities m
and k). In that case equation (47) takes the form
d2s (48)
dt2
Relation (48) is an equation of harmonic oscillations with the
frequency __

and the period

m
T = 2k
T
The solution of equation (48) is according to (42) of the form:

s = R cos t - Vo (49)

where R and <p0 depend on the initial conditions, i. e. on s0 and v0.

Note. In order to obtain the equation of harmonic oscillations
we neglected the forces of friction and air resistance when studying
the oscillations of a pendulum and a weight suspended from the
spring, and the resistance of the coil when studying the oscillatory
circuit. Physically it means that no energy is spent in the process
of the oscillations, mathematically this leads to the omission of
the term containing the first derivative. As a result we have obtained
harmonic oscillations, i. e. oscillations that remain constant from
cycle to cycle, or sustained oscillations.
What difference would it make if while solving the above
problems we took into account the force of air resistance and

49
the voltage drop across the resistance? For instance, what is the
difference between the solutions of equation (44) and those of equa­
tion (45)? Mathematical calculation (which we shall not present
here) shows that when R is not too great, equation (44), too,
describes an oscillatory process. The oscillations described by equa­
tion (44), however, become weaker with time; for this reason they
are termed damped oscillations. The physical explanation of this
difference is that the energy of oscillations constantly decreases
being transformed into heat because current passing through the
resistance R generates heat. In the same way the oscillations
of the pendulum, too, become gradually weaker because due to
friction and air resistance the pendulum’s energy is gradually spent
on heating the pendulum itself and the surrounding atmosphere.
However, when the resistance is small, the difference between the
damped and the sustained (harmonic) oscillations during a short
interval of time (for example, for several periods) is not great.
The damping makes itself felt only after a substantial time. Should
we, for example, suspend a heavy weight from a string and deflect
it slightly from its equilibrium position, the decrease in the amplitude
of the oscillations after 10-15 periods would be quite negligible
and unnoticeable to the eye. We would be able to observe it only
several minutes after the oscillations have started.
Here for the purpose of comparison is the exact solution of
equation (47) (without deduction). We shall presume that the value
of the factor b in the expression for the force of air resistance
is not too great (namely, b < 2 ymk). Then the solution of equation
(47) takes the form
h_ _b_
k_
s = Re 2m COS (50)
m 2m I f-<Po
where R and <p0 are determined by the initial conditions. It may
be seen from this formula that s decreases infinitely with time

(multiplicand e 2m becomes less and less as time f increases).

Figure 12a, b shows the graphs of functions (50) for various values
b b
of a factor • The smaller is the slower is the damping of
2m 2m
the oscillations. Compare these graphs with those of harmonic
oscillations (49) shown in Fig. 12c I formula (50) coincides with
b \
formula (49) for —— = 0 .

50
 Note in addition that for large values of the factor b (for
b > 2 \/mk) formula (50) is replaced by another formula. In this
case the weight will pass through the equilibrium position not more
than once and after that will slowly approach this position all
the time remaining on the same side of it (below or above).

51
 Some Other Applications
of the Concept of the Derivative

Maximum and Minimum Values

Let us consider some variable quantity y whose value depends

on some other quantity x. When you say that y depends on x,
or that y is a function of x, you mean that there is a quite
definite value of y to correspond to every value x assumes..For
instance, the area of a circle is a function of its radius, i. e. the
area of the circle depends on its radius. Sine, cosine, tangent, etc.
depend on the value of an angle, i. e. they are functions of the
angle. These functions are termed trigonometric functions.
So let us suppose that y is a function of x. Let us pose
the following problem: find such a value of x that y would be
at its maximum. Before proceeding with the solution, let us introduce
the concept of the domain of definition of a function. We shall
discuss this concept by using some examples.
For the first example let us take the following function. Suppose
V is the volume of a kilogramme of water at normal atmospheric
pressure and at a temperature t° (Celsius). Here V depends on t,
i. e. V is a function of the quantity t. Evidently, this function is
defined only for f’s lying inside the interval 0-100c. Indeed, at normal
atmospheric pressure water cannot have a temperature t < 0°
(water will turn into ice) or t > 100° (water will turn into steam).
Hence, the function V is defined only for such values of t that
satisfy the inequalities t ^ 0 and t < 100. Usually the two in­
equalities are written together: 0 < t < 100. Thus the function V
is defined only for
0 < t < 100
In other words, the domain of definition of the function V is made
up of numbers which satisfy the condition 0 < t < 100. Such
a domain of definition is termed a number segment because, when
numbers are represented on the number axis, all the numbers
which satisfy the condition 0 < t < 100 fill a complete segment
of the number axis. The numbers 0 and 100 are termed the

52
terminals or end points of the number section 0 < f < 100. All the
other points of the section are termed its internal values or internal
points. It is peculiar to every internal value t0 that its number
segment contains both points greater and smaller than t0. The terminals
of the segment do not possess that property.
For the second example let us consider the current i passing
through the electric circuit (shown on the diagram in Fig. 3)
t seconds after the circuit has been closed. In this case i is
a function of time t. The dependence of / on t is shown' by the
formula presented on p. 30. For what values of t is the function
/ defined? Evidently, up to the instant the circuit was closed,
i. e. for t < 0, there had been no current passing through it, therefore
it is reasonable to consider the current i only for t > 0. Hence,
the domain of definition of the function i will be made up of all
the numbers that satisfy the condition t > 0. Such a domain of
definition (it may be termed a number semi-axis) has only one end
point t = 0, all the other points are internal.
Finally, for the third example let us consider a function y = sin x.
It is defined for any x, i. e. the domain of definition of this func­
tion is the entire number axis. This region has no end points.
There are functions whose domains of definition are very
intricate. We, however, shall consider only such functions which
have for their domain of definition the number segment, number
axis or number semi-axis.
Let us now turn again to the problem of finding the maximum
value of a function. Can a function attain its maximum value
at an end point of its domain of definition? Of course, it can.
Consider, by way of example, the function V discussed above,
which defines the volume of a kilogramme of water at normal
atmospheric pressure and at a temperature t°C. Since the volume
of water increases upon heating (from 4°C upwards), it is clear
that the function V will have its maximum value at t = 100°, i. e.
at the end point of the domain of definition.
In many cases differentiation enables the position of the function’s
maximum to be determined quickly. Namely, the following
proposition holds.
Let y be a function of the variable x. I f this function attains its
maximum value at an internal point x = a of its domain of definition,
dy
the derivative —- at this point becomes zero*',
dx
*’ Subject to the condition that the derivative exists. There are such
functions in mathematics that do not have a derivative

53
 Let us prove this proposition. We shall denote the value of y
that corresponds to the value of x (chosen from the domain
of definition of the function) by yx. We have presumed that the
value ya the function y assumes at x = a is its maximum
value, i. e.
y .> y x (5i)
for any x (chosen from the domain of definition of the function).
The derivative for x = a is determined by the relation
ax
dy ya+h - ya
lim (52)
dx for x ~ a li-O h
Let us prove that this derivative is zero.
We shall start making h approach zero, attributing various
positive values to it. As long as the numerator ya+h —ya of the
fraction under the limit sign satisfies the inequality ya+h - ya < 0
(see (51)), and h > 0, the fraction under the limit sign will itself
be non-positive (i. e. it may be zero, or negative). But in that case
the limit of this fraction, too, cannot be positive, i. e. derivative
(52) cannot be a positive number.
Next let us make h approach zero, attributing various negative
values to it. In this case, as before, ya+h —ya < 0 (see (51)), but
h < 0, and for this reason the fraction under the limit sign will
itself be non-negative. But in that case the limit of this fraction
(i. e. the derivative we are interested in) cannot be negative.
dy
Thus, the value of the derivative - - for x = a can be neither
dx
positive nor negative, therefore, it must be zero, and this proves
our proposition.**
In this proof the essential fact was that a is an internal point
of the domain of definition of the function. Indeed, we attributed
positive and negative values to h so that a + h assumed values
greater and less than a.
Suppose a is an end point. Then the domain of definition
includes either values greater than a, or less than a, i. e. the above
proof is inapplicable.
When dealing with the problem of the minimum (and not maximum)
value of a function, the analysis is quite similar. As a result we

*’ It would certainly not be out of place to point out that if the limit
exists, its value must be the same for h > 0 and h < 0, i. e.
lim = lim —translator's note.
h>0 h< 0

54
shall prove that if a function assumes its minimum value at an
internal point of its domain of definition, the function’s derivative
at this point will become zero. Combining both these cases —
of maximum and minimum values —we obtain the Fermat’s theorem,
named after a French mathematician of the XVIIth century.
th e o r e m I f a function assumes its maximum ( or minimum) value
at an internal point of its domain of definition, the function's
derivative at this point becomes zero.
This theorem serves as a basis for obtaining maximum and
minimum values with the aid of differentiation. We must find the
derivative of the function in question and the corresponding
internal points of its domain of definition where the derivative
becomes zero. One should look for the point where the function
attains its maximum (or minimum) value among these points
(where the derivative becomes zero) or among the end points of
the domain of definition.
Example 11. A battery having an electromotive force E and
an internal resistance r is connected to the terminals of a conductor
(for instance, a heater). What should the conductor’s resistance be
so that it draws the maximum power from the battery?
Solution. Denote the conductor’s resistance by R. In this case
the total resistance of the circuit will be R + r, and, accordingly,
£
the current passing through the circuit will be i = -=-----
K+r
The power supplied by the battery to the conductor is expressed
by the formula W = i2R, i. e. W = —~- R _ (53)
(.R + r)2
Hence, problem may be formulated as follows: for what R does
function VFas expressed by formula (53) attain its maximum value?
The domain of definition of the function W is the semi-axis
R > 0 (we consider only the case of positive ohmic resistance).
dW
Find the derivative -----•
dR ‘

E2 (R + h) _ E2R
dW (R + h + r)2 (R + r)1
r—^ = llm --------------------;----------------
dR h->o h
E2 [(R + h)(R + r)2 - R {R + h + r)2]
= lim
h-0 h(R + h + r)2 (R + r)2
r2 - R 2 - R h r2 - R2 r-R
I1™ (R + h + r)2 (R + r)2 ~ (R + r)4 “ (K + r)3

55
I or the derivative —— i. e. the fraction

r-R
(R + r)3
to become zero, its numerator r — R should become zero, i. e.
it should be R = r.
Hence the power W can attain its maximum (or minimum) value
either for R = r or at the end point of its domain of definition
R = 0. For R = 0 the power W is zero, too (this is the minimum
value). Therefore the power can attain its maximum value only
for R = r, i. e. when the conductor's resistance is equal to the
internal resistance of the battery.
Shall the power really be at its maximum for R = r? Indeed,
we have proved only that the power may attain its maximum
value for R = r, but this in itself does not mean that it really
will be so.
It is not difficult to make sure that when R = r, the power W
is really at its maximum. Indeed, when R = 0, the power W is
zero, too; when R is very great, the current i will be very small
and hence the power will be small (for the voltage drop across
the conductor’s terminals does not exceed E). It is therefore clear
that the power must attain its maximum value for some (noi
very great) value of R. But since the power must attain its maximum
value (and it may do so only for R = r), therefore it stands
to reason that we do, indeed, obtain the maximum power when
R = r.
Example 12. The problem is how to build a steam boiler
in the shape of a cylinder so that it will have the required
volume V It is desirable to keep the total surface of the boiler
down to a minimum (in that case the minimum amount of metal
will be - required to build the boiler; besides the smaller is the
surface of the boiler the less are the heat losses due to contact
with the atmosphere). Find the best (optimum) dimensions of the
boiler.
Solution. Denote the radius of the cylinder’s base by R and its
height by h. Then
V = nR2h
l. e.
V
h=
kR2

56
 The cylindrical surface is S = 2nR2 + 2nRh, i. e.
, 2V
S = 2nR2 + — (54)
R
We must find out for what values of R the quantity S (which
depends on R, i. e. is a function of radius R) assumes its minimum
dS
value. Find the derivative — :
dR
r 2 n ( R + h)2
,, + 2v i r 2 n R 2, + —2 v i
+ n,
—
dS .. R
L R \
—— = lim
dR h^o

2Vh
2M2Rh + h2) -
= lim ----------
h-0 h
2V 2V_
= lim 4kR + 2rt/i — 4xcR -
h-0 R(R + h) ]" R2

dS
Equating the derivative to zero, we find: R = and
dR 1
therefore
>4V
=2

In other words, the height of the cylinder should be equal to its

diameter.
Do we really obtain the minimum value of the cylindrical
surface for that value of R2 This we can easily check. Indeed,
for very great values of R the surface S is also very great (since
the value of the first term in the expression for S is great —see
(54)). For very small values of R the value of the surface S is
also very great (this time it is the second term that accounts for
this). Therefore for some intermediate value (not too great and
not too small) of R the quantity S must assume its minimum value.
dS
But since the derivative —— becomes zero only for one value of
dR
R, this is the value of R to which the minimum surface of the
cylinder corresponds.
We shall confine ourselves to these two examples. If instead,
the reader will find many such examples in textbooks and problem

57
books. The reader may be advised to solve some of them provided
he does not ignore the final stage, i. e. the proof that there is
indeed a maximum or a minimum value at the point obtained.
More perfect methods are described in higher mathematics courses
which enable the decision to be made whether the function does
indeed assume its minimum or maximum value at the point
obtained. Besides, there are rules for calculating derivatives.
The author worked on the assumption that the reader is not
familiar with these rules, and for this reason the derivatives
in the examples above have been obtained with the aid of direct
calculations.

The Problem of Drawing a Tangent

Let L be a curve and Af0 a point on it. Let us discuss the

problem of drawing a tangent to the curve L at point M0. First
let us discuss how the tangent is defined in mathematics. Choose
point M also lying on the curve Land draw a straight line M0M
which we shall term a secant, for it intersects the curve L at least
at two points M 0 and M. Should point M move along the curve
L approaching point M 0 (Fig. 13 shows the sequence of positions

M, M', M", ... of point M), the secant M0M would turn around
point M0. If the secant M0M in turning tends to coincide with
some straight line M 0K, this limiting line M 0K is termed the
tangent to the curve L at point M0.
Suppose now that the curve L is drawn in a plane containing
a coordinate system so that each point M of the curve L has its
corresponding x- and y-coordinates. Denote the x-coordinate of
point M 0 by a (Fig. 14), and the length of section N0N by h.
Then the x-coordinate of point M will be a + h. We denote
the y-coordinate of point M 0 by y„, and that of point M by

58
ya+h- The length of section MP is
MP = MN - PN = MN - M 0N0 = ya+h - ya
and we therefore have
MP MP Va + h - y a
tan z. PM0M = (55)
M0P N0N h
Denote an angle PM0K, i. e. the angle between the x-axis and
the tangent, by a. Then as M approaches M 0, i. e. as a section
N0N = h tends to zero, the angle PM0M will approach a, and
the tangent of the angle PM0M tana. Hence, we obtain from
relation (55) in the limit (for h -»0)
dy
tan a = lim
h-'O dx for x = a
It follows that the slope of the tangent line is equal to the value
of the derivative of the y-coordinate with respect to the x-coordinate
for x = a, where a is the point of tangency.
Example 13. Consider a sine curve (Fig. 15), i. e. a curve

59
whose x- and ^-coordinates are related by the equation
y = sin x
How should a tangent line be drawn to this curve at some point
M 0 whose x-coordinate is equal to a?
We know already how to find the slope of this tangent line
dy_ d .
tan a = = —— sin x = cos a
dx for x = a dx for x —a

(see (36)). Hence, to draw a tangent line we must find cos a

(which is quite easy since we know the length of section
M 0N 0 = y = sin a) and draw a straight line M0K so that
tan a = cos a. For instance, when a = 0, we obtain: tan a = cos 0 = 1,
i. e. the tangent line to a sine curve drawn through the origin of
coordinates makes an angle ~ with the x-axis. For a = we

obtain: tan a = cos — = whence, using the trigonometric tables,

we obtain that the angle a, which the tangent line makes with
the x-axis, is equal approximately to 26°34'.

Modelling

We have seen above that various physical phenomena which

at first glance have nothing in common may be described by
the same differential equations. That was the case with a body
falling in a resistant medium, with the current being switched
on in a circuit and with radioactive decay. This was the case with
the three problems leading to harmonic oscillations. But when two
phenomena are described by one equation, then the solution of
this equation describes both phenomena. In other words, both
phenomena take the same course (i. e. the values of the corres­
ponding physical quantities change identically with time). Therefore
we may, for example, study the electromagnetic oscillations
in a circuit by observing the oscillations of a pendulum, and vice
versa. This very simple remark proves to be very significant.
Suppose we have a design of an intricate machine before us
and want to check whether the design was calculated correctly.
It takes much time and money to build the machine, and still
we must make sure beforehand that the design is the right one.
Let us write out the equations of motion for the machine (these
are usually differential equations). Should we solve these equations
we would know how the machine is going to work and whether

60
it has been correctly designed. But it often proves easier to act
as follows. Let us build another device (for instance, assemble
an electric circuit) which would be described by the same differen­
tial equations as the machine to be designed. If we are able to do
it, it will be sufficient to study the behaviour of the device we
have built to learn how the machine we are designing is going
to work. Thus, this device is' a model of machine.
Such “modelling”, too, uses the concept of the derivative as its
basis, for to be able to build the right model one must know
the differential equations which describe the operation of the
machine.
 A fterw ord

The concepts of the derivative and the differential equation are

used extremely wide in mathematics, physics, astronomy, and
engineering. The discipline that studies the properties and the
applications of these concepts is “higher mathematics”. Unfortu­
nately, within the space of this booklet it would have been
difficult to explain ideas underlying the definition of integration *)
which, in a sense, is an operation reciprocal to differentiation and
together with the latter forms the cornerstone of higher mathe­
matics.
The author would like to draw the reader’s attention to the
point that the concepts of higher mathematics, including the
concept of the derivative, are not abstract, but constitute
a mathematical reflection of processes taking place in nature (for
instance, of the speed of mechanical motion). These concepts
developed in close contact with problems posed by everyday
life —primarily in the fields of mechanics (the problem of calculating
velocities) and geometry (the problem of drawing a tangent).
F. Engels wrote: “Mathematics together with other sciences arose
out of man’s requirements.” This statement fully applies to higher
mathematics. The concept of the derivative created in conjunction
with the study of the motion of bodies and the variation of
quantities itself reflects this motion, for it treats a variable quantity.
‘The turning point in mathematics was Descartes’ variable quantity.
Thanks to it mathematics was invaded by motion and dialectics
and it, too, created an immediate need for differential and integral
calculus the rudiments of which were soon to be established and
which was completed as an entity but not discovered by Newton
and Leibnitz” (Engels).

*’ See I. P. Natanson Summirovanie beskonechno malykh velichin (The

Summation of Infinitesimals) published in this series (No. 12).

62
 Thus the concepts of higher mathematics arose out of man’s
requirements, primarily in connection with the study of mechanical
motion. But motion in nature is not only limited to its mechanical
form. Time and again new phenomena reveal themselves to
scientists, and mathematics has to describe newly discovered
phenomena and forms of motion. Physical theories created in recent
times (the theory of relativity, quantum physics, the theory of the
nucleus) require new mathematical apparatus. Some mathematical
disciplines were born during the last decades. Mathematics is not
an abstract science but a science closely connected with life which
develops together with the growth of our knowledge of the physical
world.
 TO THE READER
Mir Publishers would be grateful for your comments on
the content, translation, and design of this book. We would
also be pleased to receive any other suggestions you may
wish to make.
Our address is:
USSR, 129820, Moscow 1-110, GSP
Pervy Rizhsky Pereulok, 2
Mir Publishers

I’rinted in the Union of Soviet Socialist Republics