^^J
Digitized by the Internet Archive
in 2007 with funding from
IVIicrosoft Corporation
http://www.archive.org/details/easymathematicscOOIodgrich
EASY MATHEMATICS OF ALL KINDS
VOL. I. CHIEFLY ARITHMETIC
EASY MATHEMATICS
CHIEFLY ARITHMETIC
BEING A COLLECTION OF HINTS TO TEACHERS, PARENTS,
SELF-TAUGHT STUDENTS, AND ADULTS
AND
CONTAINING A SUMMARY OR INDICATION OF MOST
THINGS IN ELEMENTARY MATHEMATICS
USEFUL TO BE KNOWN
BY
Sm OLIVEE LODGE, F.RS.
D.SC. LOND. ET OXON.
LL.D. ST. ANDREWS, GLASGOW, AND A'ICTORIA
PRINCIPAL OF THE UNIVERSITY OF BIRMINGHAM
MACMILLAN AND CO., Limited
NEW YORK : THE MACMILLAN COMPANY
1906
All rights reserved
First Edition 1905.
Reprinted 1906.
GLASGOW : PRINTED AT THE UNIVERSITY PRES8
BY ROBERT MACLKHOSB AND CO. LTD.
PA37
"The parent inherits a primal tendency to revert
to the fixed and rooted form, while the child is ' free-
swimming ' ; it is the natural explorer. And for ages
we the parents through the teachers have been more
and more successfully trying to train and educate our
' free-swimmers ' into fixed and rooted prisoners ; thus
atrophising or mutilating their discovering and inter-
pretative powers just as our own were injured at the
same age. Lady "Welby.
" There are several chapters in most arithmetic books
that are wholly unnecessary .... but a writer of a
school-book for elementary schools is not his own
master ; he must comply with the often unwise demand
of teachers and examiners." A. SONNKNSCHEIN.
L5t
PKEFACE.
This book is written without the least regard to any demand
but those of children and of life and mental activity generally.
In places where the author is mistaken he cannot plead that
he has been hampered by artificial considerations. His object
in writing it has been solely the earnest hope that the teaching
of this subject may improve and may become lively and
interesting. Dulness and bad teaching are synonymous terms.
A few children are born mentally deficient, but a number are
gradually made so by the efibrts made to train their growing
faculties. A subject may easily be over-taught, or taught too
exclusively and too laboriously. Teaching which is not fresh
and lively is harmful, and in this book it is intended that the
instruction shall be interesting. Nevertheless a great deal is
purposely left to the enterprise of the student and the living
voice of the teacher, and the examples given for practice are
insufficient. The author has usually found that examples
and illustrations are likely to be most serviceable, and least
dull, when invented from time to time in illustration of the
principles which are then being expounded; but a supple-
mentary collection of exercises for practice is necessary also,
in order to consolidate the knowledge and establish the
principles as an ingrained habit. Wearisome over-practice
and iteration and needlessly long sums should be avoided;
because long sums, other than mechanical money addition,
seldom occur in practice, and especially because many kinds
vm PREFACE.
of future study, especially the great group of sciences called
Natural Philosophy, will be found to afford plenty of real
arithmetical practice ; and even ordinary life affords some,
if an open mind is kept. The cumbrous system of weights
and measures still surviving in this country should not bo
made use of to furnish cheap arithmetical exercises of
preposterous intricacy and uselessness. There is too much
of real interest in the world for any such waste of time
and energy.
The mathematical ignorance of the average educated person
has always been complete and shameless, and recently I have
become so impressed with the unedifying character of much of
the arithmetical teaching to which ordinary children are liable
to be exposed that I have ceased to wonder at the widespread
ignorance, and have felt impelled to try and take some step
towards supplying a remedy. I know that many teachers
are earnestly aiming at improvement, but they are hampered
by considerations of orthodoxy and by the requirements of
external examinations. If asked to formulate a criticism I
should say that the sums set are often too long and tedious,
the methods too remote from those actually employed by
mathematicians, the treatment altogether too abstract, didac-
tic, and un-experimental, and the subject-matter needlessly
dull and useless and wearisome.
Accordingly, in spite of much else that pressed to be done,
a book on arithmetic forced itself to the front. It is not
exactly a book for children, though I hope that elder children
will take a lively interest in it, but perhaps it may be con-
sidered most conveniently as one continuous hint to teachers,
given in the form of instruction to youth ; and it is hoped
that teachers will not disdain to use and profit by it, even
though most of them feel that all the facts were quite well
known to them before. It is not intended to instruct them
PREFACE. IX
in subject-matter, but to assist them in method of presenta-
tion; and in this a good deal of amplification is left to be
done by the teacher. But it is of the first importance that
the teacher's own ideas should be translucently clear, and
that his or her feeling for the subject should be enthusiastic :
there is no better recipe for effective teaching than these two
ingredients.
For supplementary hints in connexion with the teaching
of very small children, a subject which occupies the first four
chapters, a couple of little books by Mrs. Boole recently pub-
lished by the Clarendon Press may be mentioned ; and as a
convenient collection of suitable examples for practice I suggest
a set by Mr;_^;^ Tucke^j)ublished by Bell and Sons. For
supplementary information on the higher parts of the work
such a book of reference as Chrystal's Algebra is probably
useful.
The author has to thank Mr. T. J. Garstang, of Bedales
School, Petersfield, Hampshire, and also Mr. Alfred Lodge,
of Charterhouse, late Professor of Pure Mathematics at
Coopers Hill, for reading the proofs and detecting errors
and making suggestions.
CONTENTS.
CHAPTER I.
The very beginnings. Counting. Extension or application of the idea
of number to measuring continuous quantity. Introduction of the
idea of fractions. Practical hints for teaching the simple rules.
Addition. Subtraction. Multiplication. Multiplication of money.
Division. Division of money. Origin of the symbols - pp. 1-27
CHAPTER II.
Further considerations concerning the Arabic system of notation, and
extension of it to express fractions decimally and duodecimally.
Numeration pp. 28-36
CHAPTER III.
Further consideration of division and introduction of vulgar fractions.
Extension of the term multiplication to fractions. Practical
remarks on the treatment of fractions - - " PP- 37-44
CHAPTER IV.
Further consideration and extension of the idea of subtraction. Addi-
tion and subtraction of negative quantities - - - pp. 45-50
CHAPTER V.
Generalisation and extension of the ideas of multiplication and division
to concrete quantity. First idea of involution - - pp. 51-57
CHAPTER VI.
Factors of simple numbers pp. 58-61
xii CONTENTS.
CHAPTER VII.
Dealings with money and with weights and measures. Modern treat-
ment of the rule called ''practice." The practical advantages of
decimalisation. Decimalisation of money. T3rpical exercises.
Binary scale. Decimal system of weights and measures. Decimal
measures. Angles and time. Further exercises - - pp. 62-85
CHAPTER VIII.
Simple proportion. Breakdown of simple proportion or " rule of
three " pp. 86-94
CHAPTER IX.
Simplification of fractions pp. 95-101
CHAPTER X.
Greatest common measure and least common multiple. Rule for
finding G.C.M. Algebraical statement of the process for finding
G.C.M. pp. 102-108
CHAPTER XI.
Easy mode of treating problems which require a little thought
pp. 109-115
CHAPTER XII.
Involution and evolution and beginning of indices - pp. 116-121
CHAPTER XIII.
Equations treated by the method of very elementary experiment.
Further consideration of what can be done to equations
pp. 122-137
CHAPTER XIV.
Another treatment of equations. Introduction to quadratics
pp. 138-148
CHAPTER XV.
Extraction of simple roots. Surds . . . . pp. 149-163
CHAPTER XVI.
Further consideration of indices. Fractional indices.
Negative indices pp. 154-^161
I
CONTENTS. xiii
CHAPTER XVII.
Introduction to logarithms pp. 162-164
CHAPTER XVIII.
Logarithms. Common practical base. Examples. Examples for
practice. Fundamental relations ... - pp. 165-176
CHAPTER XIX.
Further details about logarithms pp. 177-183
CHAPTER XX.
On incommensurables and on discontinuity - - - pp. 184-192
CHAPTER XXI.
Concrete arithmetic. The meaning of significant figures and practical
accuracy pp. 193-197
CHAPTER XXII.
Practical manipulation of fractions when decimally expressed.
"Order" of numbers pp. 198-209
CHAPTER XXIII.
Dealings with very large or very small numbers - - pp. 210-218
CHAPTER XXIV.
Dealings with vulgar fractions. Numerical verifications pp. 219-224
CHAPTER XXV.
Simplification of fractional expressions - - - . pp. 225-229
CHAPTER XXVI.
Cancelling among units pp. 230-234
CHAPTER XXVn.
Cancelling in equations. Caution - - - - pp. 235-238
CHAPTER XXVIII.
Further cautions. Cautious of a slightly more advanced character
pp. 239-245
XIV CONTENTS.
CHAPTER XXIX.
Illustrations of the Practical Use of Logarithms.
(i). How to look out a logarithm, (ii). How to look out the number
which has a given logarithm. Examples. Logarithms of fractions,
(iii). How to do multiplication and division with logs. pp. 246-257
CHAPTER XXX.
How to find powers and roots by logarithms. Exercises.
Roots of negative numbers pp. 258-264
CHAPTER XXXI.
Geometrical illustration of powers and roots. Further geometrical
methods of finding square roots .... pp. 265-276
CHAPTER XXXII.
Arithmetical method of finding square roots - - pp. 277-282
CHAPTER XXXIIL
Simple algebraic aids to arithmetic, etc. Illustrations. Problems.
Proof of square root rule. Cubes and cube root. Approximations
pp. 283-310
CHAPTER XXXIV.
To find any power of a binomial. Exercise. Examples pp. 311-321
CHAPTER XXXV.
Progressions. Examples. Algebraic digression. General expression
for any odd number. Arithmetical progression. Other series.
Geometrical illustrations pp. 322-338
CHAPTER XXXVL
Means. Examples. Mean or average of a number of terms. Weighted
mean. General average. Geometric mean. Example pp. 339-351
CHAPTER XXXVIL
Examples of the practical occurrence of progressions in nature or art.
Interest pp. 352-358
CONTENTS. XV
PART IL
MISCELLANEOUS APPLICATIONS AND INTRODUCTIONS.
CHAPTER XXXVIII.
Illustrations of important principles by means of expansion by heat.
Examples. Cubical expansion .... pp. 361-371
CHAPTER XXXIX.
Further illustrations of proportionality or variation.
Inverse variation. Summary pp. 372-400
CHAPTER XL.
Pumps and leaks. Leaks. Cooling. Electric leakage. Continuously
decreasing G.P. Summary pp. 401-411
CHAPTER XLI.
Differentiation. Examples pp. 412-418
CHAPTER XLII.
A peculiar series. Natural base of logarithms - - pp. 419-426
APPENDIX.
I. Note on the Pythagorean numbers (Euc. I. 47) - pp. 427-430
II. Note on Implicit dimensions pp. 430-434
III. Note on factorisation pp. 434-435
IV. Note on the growth of population .... pp. 435-436
CHAPTER I.
The very beginnings.
Concerning the early treatment of number for very small
children the author is not competent to dogmatise, but he
offers a few suggestions, the more willingly inasmuch as he
is informed by teachers that a great deal of harm can be and
often is done by bad teaching at the earliest stages, so that
subsequently a good deal has to be unlearnt. The principle
of evolution should be recollected in dealing with young
children, and the mental attitude of the savage may often be
thought of as elucidating both the strength and the weakness
of their minds.
Counting is clearly the first thing to learn; it can be
learnt in play and at meals, and it should be learned on
separate objects, not on divided scales or any other con-
tinuous quantity. The objects to be counted should be
such as involve some childish interest, such as fruit or sweets
or counters or nuts or coins. Beans or pebbles will also do,
but they should not be dull in appearance, unattractive as
objects of property, and so not worth counting. The pips
on ordinary playing cards will also serve, and they suggest
a geometrical or regular arrangement as an easy way of
grasping a number at a glance.
Counting should begin with quite small numbers and
should not proceed beyond a dozen for some time, but there
2 EASY MATHEMATICS. [chap.
is no object in stopping or making any break at ten. Several
important facts (the facts only, not their symbolic expression)
can now be realised : such as that 3 + 4 = 7, that 7 - 4 = 3,
that two threes are 6, and that three twos are the same,
without any formal teaching beyond a judicious question or
two. The lessons, if they can be called lessons, should go
on at home before school age ; but, whether this initial train-
ing is done at school or elsewhere, formal teaching at this
stage should be eschewed, since it necessarily consists largely
in coercing the children to arrive at some fixed notion which
the teacher has preconceived in his mind — a matter usually
of small importance. The children should form their own
notions, and be led to make small discoveries and inventions,
if they can, from the first. Mathematics is one of the finest
materials for cheap and easy experimenting that exists. It
is partly ignorance, and partly stupidity, and partly false
tradition which has beclouded this fact, so that even influen-
tial persons occasionally speak of mathematics as " that study
which knows nothing of observation, nothing of induction,
nothing of experiment," — a ghastly but prevalent error which
has ruined more teaching than perhaps any other misconcep-
tion of the kind.
As soon as small groups can be quickly counted, and
dimple addition and subtraction performed with a few readily
grasped and interesting objects — and the more instinctively
such operations can be done the better, — the time is getting
ripe for the introduction of symbols — for that arbitrary and
conventional but convenient symbolism whereby !•! is de-
noted by a crooked line, 5, and so on : a symbolism which the
adult is only relieved from the necessity of elaborating and feel-
ing difficult because of the extreme docility and acquisitiveness
of childhood. It has already learned 26 symbols, it will
patiently absorb nine or ten more, especially as they are soon
I.]
COUNTING.
3
found to be real conveniences; though if an adult wishes
to realise the genuine difficulty of the process — always a
most desirable thing to do — he should set to work to learn
the Morse telegraphic alphabet, especially in the forms used
for cable telegraphy.
I see no reason now why '.'. .'. should not be written
4 + 5 = 9, or soon afterwards why ; •(:) might not be written
5-2 = 3; but let no one suppose that these steps in nomen-
clature are easy. The nomenclature introduced is just as
hard as that of trigonometry or the calculus, only adult
persons are accustomed to the one and are often unacquainted
with the other. A set of little blocks, or some simple cheap
squared paper lends itself to statements like the following :
5=
cn
H
n
B
6 = 2+2+2 = 3+3
Fig. 1.
4+2
5+1
I see no reason for troubling about the names " addition "
and "subtraction," nor yet for artificially withholding them.
Jf they come naturally and helpfully, let them come.
Nothing is gained by artificial repression at any stage.
Premature forcing of names is worse than artificial with-
holding of them, but both are bad. If a gas, bubbling out
of soda water and extinguishing a flame, is familiarly known
as " carbonic acid," let it be called so : it is a help to have
4 EASY MATHEMATICS. [chap.
a label with which to associate observed properties, just as
it is convenient to call a certain flower " daisy," or a certain
star "Sirius." But to supply the label and withhold the
object, to lecture about daisies or stars or numbers before
they have been seen, is, let us politely say, unwise.
It seems to me that card games with counters may now
be introduced, to enable the children to realise that their
property may mount up beyond the smaller numbers that
would be wholesome with sweets ; and they can learn how to
group their counters into packets of six, or even into dozens,
and then they will have simply to count their packets and the
odd ones over. A child with four packets of six and three
over would have a real idea of his wealth, though " twenty-
seven " might still be a meaningless expression.
Differently coloured counters are now serviceable to replace
the packets, and thus the idea, but not the word, of different
" denominations " will be imperceptibly arrived at : and it
will be clinched by the at first unexpected discovery that
even strangers will accept one white coin as equivalent to
six much larger brown ones.
After this, some approach toward the admirable Arabic
notation, whereby value is symbolised by place or position
as well as by shape of digit, may be unobtrusively entered
on. The idea of boxes or cases, or spaces of different
value, in one of which odd counters or pennies are to be
stored, another one in which packets, or silver coins, are to
be kept ; and ultimately, but not too soon, a third one which
is to be occupied by packets of packets, or gold coins ; if ever
such wealth were attained.
While there is every advantage in thus emphasising atten-
tion to the value or place of the digit, and so to a system of
numeration, there are many reasons against concentrating
attention on the particular number " ten " prematurely : it is
I.] COUNTINa. 5
not a specially natural number, for one thing ; for another
thing it is so large that ten packets of ten are unlikely to occur,
whereas four packets of four, or six of six are quite possible.
Another reason is that it is undesirable to suggest, what habit
will subsequently only too erroneously enforce, that there is
something special and divine about the number ten, so that
the arrangement of digits 12 cannot help meaning a dozen.
This false idea, due merely to habit, will not occur to a child,
nor will he know intuitively that twelve pence make a
shilling, or twenty shillings a sovereign; indeed, strange to
say, he is usually somewhat callous as to the importance of
this pivot of human existence; and, though he soon gets to
like coins, he attends chiefly to their number without much
regard to their denomination, unless some are specially new
and bright.
Having got so far, the conventional symbolism, in which
practice has been quietly going on in the background during
the few more formal school quarter-hours, may be extended,
and the digit-symbols written in spaces drawn to represent
the boxes, or on paper ruled into quarter-inch squares, which
is cheaply and plentifully accessible, so that a 4 put in one
box shall signify 4 counters, while a 4 put in another box
shall signify 4 packets of say ten counters each, so that at the
end of a game | 0 | 3 | shall mean that the loser has no packets
and only three counters altogether, while another child may
have 3 I 0 ; that is, three complete packets and none over.
A third may have two packets and five over ; that is to say
and another, the winner at the game, may possess
or in words, 1 packet of packets, 5 simple packets,
2 5
15 2
and 2 odd ones.
The packets may be represented by otherwise coloured
counters, or the well known Tillich bricks or other Kinder-
6 EASY MATHEMATICS [chap.
garten devices can be employed for convenience ; the
important thing is not prematurely (i.e. not until the under-
lying reality has been essentially grasped) to proceed to the
only partially expressive symbolism 25 or 152, which to
us by mere habit looks so living and significant. Let the
elementary teacher reflect that to a mathematician the symbol
i
e~* dx looks equally living and significant, and be not hasty
with the children.
At the same time there is no need for artificial delay. A
child brought along the right lines will jump forward without
difficulty, will recognise the places without the boxes, will
get accustomed to the savage's mode of reckoning by tens
without being encouraged to go through the savage process
of counting on his fingers, and before long will be able
to interpret such a complicated symbolism as 50327, or
.£175. 16s. lid. The last, indeed, is properly spoken of as
" compound " instead of simple, for in it " scales of notation "
are badly mixed up. The reckoning proceeds by tens, by
dozens, and by scores, sometimes one and sometimes another,
occasionally by quarters also.
The poor child who finds himself able to master this and
the operations which arise out of it, need not be deterred by
any legitimate obstacles in mathematics until he comes to its
really higher walks, beyond simple differential equations : a
step which he will not be called upon to take at all unless he
is born to be a mathematician, in which case difficulties of any
ordinary kind will barely be felt.
The operations of addition and subtraction may now be
extended. 7 + 5 may be done into a packet of one dozen, or
into a packet of ten and two over, and denoted by 1/- or 12
according to which plan of grouping is adopted.
So also 8 + 7 may be called either 1/3 or 15, the former
1.] COUNTING. 7
being the custom if they are pennies, the latter if they are
nuts.
It is necessary to apologise to children for this needless
complication; but they inherit some things that are good,
to make up for several things that are stupid, and therewith
they will have to be content : —
8 + 7 + 9, if shillings, will be grouped differently again, and
be denoted by £1. 4s.; if pennies, they will be denoted thus,
2/- ; if ounces, they will be written 1 lb. 8 oz. ; if feet, they will
be called 8 yards ; if farthings, they will be written 6d. ; if
oranges, they will be called 2 dozen; but if boys, they will
be written 24.
I do not recommend anyone to confuse the minds of children
by pointing out these anomalies, or by quoting a sample of
them simultaneously as above. Children will not detect their
true character, but will docilely receive them as if all this
rubbish were part of the laws of nature. This may account
for their disinclination later on to make acquaintance with any
more of those laws than they can help, but at this stage they
are docile and assimilative enough : they can at this stage be
taken advantage of with impunity. But I should very much
like to confuse the minds of some teachers, and of some school
inspectors — especially some varieties of school inspector and
university examiner — and get them into a more apologetic
and humble mood at having to insist on filling the mind of a
child with any more of these artificial insular conventions
than is absolutely necessary in the present stage of British
political and commercial wisdom.
It is undesirable to hasten forward to numbers involving
3 digits too quickly ; they can be mentioned and illustrated
when convenient, but real work should for some time be limited
to 2 figure numbers, because in these the real principles can
be recognised and grown accustomed to in the simplest way.
6
EASY MATHEMATICS.
[chap.
The early operations in which practice can be given are
such as the following : Suppose counters are employed and
that little cases have been made which just hold six or ten
or any convenient number, suppose ten :
Then 13 will stand for one packet of ten and three counters
over;
17 added to it will amount to two packets and ten
counters over; which the child, if encouraged by
the sight of an unused case available, may wish
to make up into 3 whole packets, and so recognise
the propriety of denoting the number by
30
Similarly 15 + 17 will make up into three packets and
2 over, which may be shown thus :
make
which equals
while 25 + 37 will equal five packets and twelve over, or six
packets and 2 over; 29 + 37 = 66, but it is equally per-
missible to keep it as 5 packets and 16 counters over, if it
should happen to be convenient — as it sometimes is.
To take 4 from 17 is easy,
but to take 9 from 1 7 will involve emptying a case ; and only
8 counters will be left.
To take 13 from 25 can be done by removing 1 case and
2 counters ;
to take 15 from 25 is also easy ;
but to take 16 from 25 involves the breaking up of a packet.
tens
1
1
ones
5
7
2
12
3
2
i.j COUNTING. d
After a time these operations can be followed when nothing
concrete is present; but abstractions are not natural co
children, and before calling upon them to follow a difficult
conventional subtraction sum like
82
37
45
the operation of breaking up packets should be introduced
into the symbolism which is employed to faintly shadow the
concrete reality.
It is perfectly right to speak of 3 packets and 13 loose
counters, although they may be more compactly grouped as
4 packets and 3 counters. So if we have to subtract say 7
from 43 we shall first break up one of the four packets, so
as to turn 43 into 3 packets and 13, and then subtract the
7 without difficulty, leaving what is abbreviated into 36.
Hence before doing the above conventional little sum,
8 packets and 2 should be expressed as 7 packets and 12, or
From this 3 packets and 7 have to be removed,
7 12
leaving obviously 4 packets and 5. Wherefore 82-37 = 45
without any argument.
The abbreviated form of the above breaking-up operation,
called borrowing, will now gradually almost suggest itself, if
many sums of the kind are given to be done. But the best
and easiest method of subtraction is the complementary
method, and if this is taught from the first, the complexity
of borrowing becomes unnecessary.
The adult cannot too clearly realise that many of the
operations to which he has grown accustomed are labour-
saving shorthand devices with the vitality and principle
abbreviated out of them; quite rightly so for practical pur-
poses but not for educational purposes. The race invented
10 EASY MATHEMATICS. [chap.
them at first in more elaborate shape, and gradually abbre-
viated them into their present-day form. The child will
likewise get accustomed to this form in due time, but he
should not be over-hurried into it.
After adding two numbers for some time we may proceed
to add more than two,
and find that 7 + 9 -f 6 = 22, etc. ;
also that 7 -I- 7 +.7 = 21,
and it is natural to speak of this as three sevens.
So also the fact that 5 + 5 + 5 + 5 = 20 will naturally be
quoted as four fives make twenty ; and thus the essential idea
of multiplication will arrive, as a shorthand and memorised
summary of the addition of a number of similar things,
without any use of the name multiplication or any feeling
of a new departure. To find the value of three seventeens,
that is, to group them into tens and ones, is a problem for
an afternoon, and if it be done with counters in the first
instance, and ultimately with symbols, the meaning of the
operations having been realised beforehand with the counters,
so much the better.
The operation of adding or multiplying means grouping the
whole number into tens and ones, or into hundreds, tens, ones,
etc., instead of in the given groups.
A child must not be expected to be able to formulate his
conception of the operations, or to express them accurately in
words, at this stage. It is a capital exercise later, but it is
enough at first for him to realise the meaning of what he is
doing in the back of his mind. From time to time he can be
encouraged to interpret processes into words, but they must
have become familiar first. To be able to apply a rule, from a
precise statement in words of what has to be done, is an adult
accomplishment, often not reached by adults. To dissect out
and state a rule in words, from a knowledge of what the
I.] EDUCATION. 11
operation really is, is perhaps easier, and is a desirable gift, but
it is a training in the use of language rather than in the subject
matter of the craft. It is most appropriate and valuable prac-
tice for children at the proper stage, a stage reached much
earlier with some children than with others. Children who
reach the word-expression stage late are usually called "stupid."
If this adjective implies a stigma it is usually undeserved.
There is a performance appropriate to each stage of develop-
ment, and opprobrious epithets are generally employed by
those who seek to force things several stages too soon. A
highly trained and clever dog would soon prove himself
" stupid " if tested by a formula, or by words even of only
3 letters. An adult who can hum or whistle an air may be
told that he ought to be able to sit down and write it in the
recognised musical notation. Similarly he ought to be able to
read off a piece of music handed to him. He might resent
being called stupid if he found it difficult to do these, to some,
so simple things.
" Badness " of many kinds may exist in spoiled children (and
there are several ways of spoiling them), but badness in un-
spoiled children is rare, and stupidity is almost non-existent
unless they are physiologically out of order and therefore
mentally deficient. Stupidity is however a product easily
cultivated by improper feeding, especially improper mental
feeding. The " badness " of children is largely the effort which
nature makes at self-preservation ; for inattention and laziness
are the weapons whereby an attack of mental indigestion can
be warded off.
The only fault with very young children is that they are
too good, and therefore too easily damaged. Later on, a spirit
of rebellion acts as a preservative, but it would be better to
dispense both w^ith the rebellious spirit and with the causes
which necessitate it.
U EASY MATHEMATICS. [chap.
Returning from this digression, which is either false or else
of very extensive application, to our immediate subject, viz.,
the introduction of the fundamental operations to be performed
on number, — and remember that what are called the first four
simple rules are tremendously fundamental and important,
more important than anything which follows, until involution,
evolution, and logarithms are arrived at, — we must exercise
children in Multiplication and teach them something of the
multiplication table, at first experimentally, but afterwards by
straightforward memory work, for it is one of the things with
which the memory may be rightly loaded. We can next
recognise that Division too can be unceremoniously introduced
by trying to split up numbers into equal parts. The endeavour
to share sweets or fruit or cards or counters is an obvious
beginning. Then, since children are docile, they can be asked
to split up 2 packets and 7 into three equal groups, or they
can be asked to split up 2 packets and 4 into eight equal groups,
and so on ; for no reason assigned. But it must be recognised
that the operation of division in general is rather hard, and
involves a good deal of tentative procedure or guess work.
In other words it involves the rudiments of experiment and
verification. Gradually, when the multiplication-table is fairly
known over some little range, children can be encouraged to
apply theory before practice and actually to think out the
result before trying it ; but this is a lesson in deductive
reasoning, and represents the nascent beginnings of a loftier
mode of procedure than ordinary adults are accustomed to
apply to their affairs. When asked to split 28 into four equal
heaps, it is an application of pure theory to remember that 4
sevens are 28 and then to count out seven counters into each
heap at once. The empirical mode would be a method of
dealing out singly into four groups and then counting the
result. It is easily done with ordinary playing cards, but
I.]
EDUCATION. 13
its value as training is much enhanced if theory is applied
first.
If for instance 30 cards were given, to be dealt to four
players, the residue that will not go round to be put in the
middle or pool, a decided effort is required for a child to
perceive that there will be two for the pool and seven for
each player : but if he could have time allowed him so to
think it out, and then to make the experiment, he would be
conscious that his powers were developing, and he would in
reality be introduced to the first beginnings of a mode of
comprehending nature such as is in the higher stages reserved
for men of science, — using the term science in its most com-
prehensive signification.
It is very often a mistake for teachers to suppose that some
things are easy and other things are hard ; it all depends on
the way they are presented and on the stage at which they
are introduced. To ascend to the first floor of a house is
difficult if no staircase is provided, but with a proper staircase
it only needs a little patience to ascend to the roof. The
same sort of steps are met with all the way, only there are
more of them. To people who live habitually on the third
floor it is indeed sometimes easier to go on to the roof than
to descend into the basement. Educators should see that they
do not forcibly drive children in shoals up an unfinished or ill-
made stairway, which only the athletic ones can climb. It is
extremely difficult in familiar subjects not to go too fast. The
effort sometimes results in a process of going too slowly, which
is wearisome and depressing and the worse fault of the two.
Extension or Application of the idea of number to
measuring continuous quantity.
So far we have been employing number to count discrete
objects, and to perform simple operations of addition, and
14 EASY MATHEMATICS. [chap.
the like, among them. It is now appropriate to introduce
the idea of multiples of a unit, so that one thing can be twice
as long or twice as heavy as another, without being in another
sense "two" at all. The lines on ruled paper enable one
easily to draw across them a line twice or three times or six
times as long as another. So also letter-scales can be used to
show that a penny is twice as heavy as a half-penny, that a
half-crown weighs how many sixpences, and the like.
Given a foot rule they can measure the size of furniture,
or of books. Given a few ounce weights they can make very
rough estimates of the weights of things that have or might
have to go by post.
It is desirable not to dwell on these things at this stage,
but simply to accustom a child to recognise a rod 6 inches
long, and such like, and to see instinctively and without
formula or expression that number may he applied to con-
tinuous magnitude by the device of a unit of measurement.
Adults may realise that there is a real step here, by remem-
bering that if they were set to express the strength of an
electric current, or the electric pressure on a main, or the
strength of a magnet, numerically, they would be nonplussed,
unless they knew something about the units which within a
generation or two have been introduced for the purpose, —
the ampere, the volt, and the line of force ; so that nowadays
the British workman is able to speak familiarly of an electric
current of so many amperes — (sometimes pronounced
" hampers "). There is nothing really numerical about the
length of a table or the height of a door or the weight of
a sack or the brightness of a lamp or the warmth of a
room or the length of a day ; and its numerical expression
will depend entirely upon what conventional unit is em-
ployed, and may vary in different countries accordingly.
Do not assume therefore that a child is stupid to whom
I.] EARLY OPERATIONS. 15
an application of arithmetic to weighing and measuring is
not obvious.
Introduction of the idea of fractions
In the same way the idea of fractions can naturally occur ;
a halfpenny and a half ounce and a half inch being fairly easy
examples : but not the easiest. There can be no doubt that
just as numbering ought not to begin with continuous quantity
but with discrete objects, so fractions should be first displayed
as actually cut and broken things.
The proper fractions to begin with are halves and quarters
and eighths; and apples do admirably for that. Oranges
suggest further modes of subdivision, except that the removal
of the peel may constitute an unexpressed but felt complication.
Folding of a ribbon or paper easily leads to thirds and a^ny
other fractions wanted. Any child can be sent to cut off a
quarter of a yard, or a yard and a half, or even a foot and
three quarters, of tape. But again do not be surprised if this
last mode of specification is found occasionally puzzling : it is
of the nature of a problem, and requires time. The form of
difficulty which may properly occur to some children is " a half
of what "or " three quarters of what " : and if they bring the
foot and the 3 quarters all separate, i.e. if they cut the tape
into four pieces altogether, that is very well for a beginning.
They should not be supervised or fidgeted during the solution
of a problem. They cannot think if they are. These expres-
sions, 6 miles and a half, etc., have a conventional ring, to
which we have grown thoroughly accustomed, but they are
shorthand terms not really fully expressive : it might possibly
ambiguously suggest 9 miles.*
The measure of time in half and quarter hours may also be
*Cf. George Meredith's "Rhoda Fleming," Chap. 3.
16 EASY MATHEMATICS. [chap.
appealed to as illustrative of fractions ; but in this form they
are somewhat abstract. The divisions on a foot rule or metre
scale are easier, and for further progress are indeed the easiest
illustration to be borne in mind. Afterwards, the halfpenny,
the half crown, the half sovereign, etc., and the other fractions
of money may be brought in, whenever they appear to be
natural.
Practical hints for teaching the simple rules.
Simultaneously with all this introduction of fresh concep-
tions, mechanical practice in operations with symbolised num-
bers can be proceeded with : —
Addition.
About addition there is little to be said : the idea of packets
must have made everything concerning the carrying-figure
easy.
The principle being understood, it is now only a question of
practice in attaining quick and sure execution, as quick and
sure as it is worth while to aim at at this stage.
Addition of money is a useful accomplishment, and since the
packets into which it is to be made up are varied, it affords
good practice, involving a certain amount of constant thought
and care. It is wrong to try to force a child to acquire the
facility of a bank clerk in adding up long columns : that will
come in due time and is quite a useful faculty : it is clearly a
thing to acquire in commercial schools, but not while still
young and receptive.
It is well to begin thus :
£ s. d.
6 . 15 . 3
5.4.9
12 . — . — .
I.] EARLY OPERATIONS. 17
where the packets to be carried forward are complete. Then
change the 3 into a 4 or 5 and get 1 or 2 pence over; then
change the 15 into 16 or 17 and get some shillings over, and
so on, gradually. Always begin with what illustrates the procedure
in the simplest form and gradually complicate it.
There is one remark about addition worth making. In
adding say 43 + 8, some beginners are told to bethink them-
selves that 3 + 8 = 11, and so arrive at the digit 1 of the
result ; while others are told to think of the sum as 43 + 7 + 1,
stepping on to the intermediate stage of the complete packet
en route to 51 ;
e.g. 77 + 9 = 77 + 3 + 6 = 80 + 6 = 86.
Perhaps it is permissible to introduce this aid as a temporary
measure, but ultimately addition ought to proceed by instinct
and without thought. It is a mechanical process, and a bank
clerk who stopped to think, while adding, would be liable to
make a mistake.
Subtraction.
There appears to be no doubt now but that the "shop
method " of subtraction is the handiest and quickest : it may
as well, therefore, be acquired almost from the first.
37 Three and four make seven.
1^ One and two make 3.
24 Put down the figures in black type.
Verify by adding 13 to 24. Take another example :
174
Qo Eight and six make fourteen.
—^ Nine and one and seven make seventeen.
I do not think that children need find this method hard or
L.E.M. B
18 EASY MATHEMATICS. [chap.
unnatural, but practice will be needed before going on to
money sums, such as :
£ s d
, - ' " Four and seven make 11.
II ' g * ^ Eight and eighteen make 26.
~T 7^ H 12 and 5 make 17.
Verify by addition of the two lower lines. Get the children
never to pass and hand in a result as finished unless they have
taken pains to assure themselves that it is right. This does
not mean that they are not to hand in a confessedly unfinished
sum if they find they cannot do it without help.
Multiplication.
At good Kindergarten schools, a step beyond the first in
multiplication is often introduced by some such questions as
this :
How many stamps will three children have if each has 14 1
They first add 14 three times, and they are allowed to do
that till they find it quicker to use the phrase " three times,"
which, if they know the multiplication table, they can hardly
help doing in the process of adding ; and so they get to be
able to give the answer "3 times 14" instantly, without
necessarily having had time to realise what the operation
would result in when executed. This kind of intermediate
answer is to be encouraged.
In entering upon multiplication, employ a single digit as
one factor, and do it first as an addition sum, e.g. :
142
142
142
142
142
710
I-]
EARLY OPERATIONS. 19
then proceed
142
5
173 125
5 8
865 1000
£.
12 .
8.
7 .
d.
6
4
710
49 .
10 .
—
doing this latter also by addition first :
£. s. d.
12 . 7 . 6
12 . 7 . 6
12 . 7 . 6
12 . 7 . 6
49 . 10 . 0
but it is well to lead up to the last type of sum by simple
cases first, e.g. 4 x 2/6 = 10/-; 4 x 5/- = £1 ; 4 x 7/6 = 30/-
= £1 10/- ; 4 X 3d. = 1/- ; 4 x 1/3 = 5/- ; 4 x 10/- = £2 ;
4x11/3 = £2. 5s.
Do not hurry. If the child can be allowed time to see
a connexion between the three last statements, or the like,
so much the better. The value of these trifles is when they
are discovered; there is hardly any virtue in them if they
are pointed out, and none at all if they are laboriously
emphasised. If they are not glimpsed let them pass. We
all of us doubtless miss discoveries, most days, for lack of
attention and insight.
Next comes multiplication with two digits : first by numbers
like 10, 20, 70, etc.
Multiplying by ten means making every unit into a packet,
every packet into a set of packets, and so on.
Wherefore
0 I 1 I 3 I 4 when multiplied by ten becomes
13 4 0
the 1 being shifted into the empty compartment, and every
other digit likewise moved; the unit box, or box for single
counters, being left empty.
20 EASY MATHEMATICS. [chap.
If we multiply by 20, the shift takes place similarly, and
also every digit is doubled, yielding 2680.
So now start multiplying a number like 53 by 20, getting 1060.
Then a number with a carrying figure from the units place,
^^^® 47x20 = 940;
then one involving two carryings, like
or,^ c^ 57x20 = 1140,
and so on. '
Next take multiplication by a number like 23. Let it be
realised once more that 23 is short for 20 + 3, so that it may
be felt to be natural to multiply by 20 and by 3 successively
and add the results, which is what we do. At first let it be
worked in this way ; for instance, to find
824x23
= 20 X 824 or 16480
and 3 x 824 or 2472
added together make 18952
but gradually get it abbreviated into the usual form
824
23
1648
2472
18952
without necessarily putting in the cipher after the digit 8.
There appears to be no doubt now that it is best in mul-
tiplication to begin with the most important figure, so that
sums look thus :
173 173 768
56 156 107
865 173 768
1038 865 5376
9688 1038 82176
t.]
EARLY OPERATIONS.
21
a trivial matter to all appearance, but helpful in later stages,
and therefore better practised from the first.
[In my opinion it is thoroughly unwise to reverse the digits
of any factor before multiplying with them, though some
teachers of immense experience think otherwise.]
Multiplication of money, at least of English money, is
more difficult of course, because, in the specification of money,
scales of notation are so mingled; thus, depicting the com-
partments and labelling them when necessary :
£
s.
d.
4
3
5
1
7
1
1
at the double line the scale is changed from ten to a dozen,
and at the treble line it is changed again from ten to a score.
So if we have to double this sum, even doubling it is com-
plicated, and results in
£871 . 15 . 10
Let no one suppose that this is an easy process, for a child or
anyone.
It could in this case be performed more easily by simple
addition :
£ s. d.
435 . 17 . 11
435 . 17 . 11
871 . 15 . 10
but that is hardly applicable to larger factors. Not only
is doubling hard, but multiplying even by 10 is hard too.
Take the amount £5. 17s. Ud. and multiply it by 10; it
becomes the totally different-looking amount
£58. 19s. 2d.
Multiplying by 12 will of course turn all the pence into
shillings, and multiplying by 20 will turn shillings into
22 EASY MATHEMATICS. [chap.
pounds, but multiplying by any other factor is hard, and
is probably best deferred for the present.
If multiplication of money by a number like 23 is wanted,
not only must the 23 be divided into two parts 20 + 3, and
the multiplication done separately as usual, but it is generally
needful to resolve the 20 into two parts also, say 10 + 10,
and then add the three results together.
If however multiplication by 24 were desired, it would
be possible to split it into two factors 8x3, and to multiply
first by one and then the result by the other, without any
addition of results; but there is great danger of confusion
here, and there are plenty of what are considered and are
really "higher" parts of arithmetic which are much easier
than this. Low class or unskilled labour is not necessarily
easy : it may in some cases be terribly laborious, like un-
loading a ship. Another way of multiplying by 20 is to
split up 20 into the two factors 2x10 or 4x5 and employ
them successively. In that case the result of multiplying
by 23 is ultimately obtained by multiplying the original
sum by 2, the result by 10, the original sum by 3, and then
adding the last two results.
The fact is that with money specified in the customary
English way, the only operations that can comfortably be
performed on it are addition and subtraction, and these are
the only really frequent operations in practice.
To apply multiplication and division it is best to express
the money differently, in fact to decimalise it before commenc-
ing operations. This will be explained later (Chap. VII.),
though of course to most teachers it is a process already well
known. It ingeniously evades the difficulties caused by our
currency, and converts its treatment into almost a worthy
intellectual exercise.
1.] EARLY OPERATIONS. S3
Division.
First take simple sums to introduce the notatioUj such as
y = 3, or 21-r7 ^ 3.
21
Let it be realised also that -5- = 7, and that 3x7 = 21.
o
There are a multitude of interesting things to be learnt
before long about factors, and criteria for division, etc., but
not yet; let the child learn how to perform the process on
numbers of which he knows no factors. But at first do not
trouble him with remainders : let him at first be given simple
sums that divide out completely.
Thus we can tackle such sums as
71491036 , . , , ... , .,, 491036 ^^,,^
' — , which should be also written - — = — = 70148.
70148 7
The treatment of remainders is for subsequent consideration.
It is well to give the complementary sum 7x70148,
especially since the teacher will thus have but little trouble in
checking results — at least until the child finds out the dodge
—a discovery which is to be encouraged like all other
discoveries.
At good Kindergarten schools, a step beyond the first in
division is often introduced by some such plan as the
following :
To prove that 96 ^ 4 = 24.
Take nine bundles and six sticks over, deal out into four
places, two bundles in each place; and then deal sixteen
sticks, four into each place, giving the result 24. And so on
with other numbers.
As soon as short division is thoroughly understood, long
division may introduce itself as an assistance when more
difficult divisors are involved; for instance 988 -7-19. This
24 EASY MATHEMATICS. [chap.
being difficult to do by short division, where the multiplica-
tion and subtraction have to be done in one's head, it is
permitted to write the operations down, at first both of them,
thus :
19)988(5
95
3
Afterwards, perhaps, only the result of them, 3, which in
short division would likewise not appear, nothing but the
quotient being written in short division. Long division is
therefore not harder than short division, but easier : it is the
identical process, only written out more fully, so as to be
applicable to harder sums. It is the largeness of the figures
dealt with that makes it hard.
For long division it appears to be felt that by aid of the
shop system of subtraction there is no undue strain on the
brain by the use of the abbreviated method.
I would have it understood however that long division
sums are among the moderately hard things of life, and that
mathematicians seldom trouble themselves to do them. They
can be deferred until many other things have been done and
some familiarity with figures acquired. It is a gymnastic
exercise to perform even so simple a long division sum as the
following, and if attempted too early will involve strain.
72)5286456(73423
246
304
165
216
This is the process :
Sevens in 52 ? guess 7 times and write 7 as the first digit in
the quotient, then 7 x 2 = 14, to which add 4 to make 18.
Seven sevens = 49, say 50, to which add 2 to make 52 ;
record only the figures here printed in black type; bring
1.] EARLY OPERATIONS. 25
down the rest of the dividend 6456 or as much of it as is
wanted; only 6 is wanted so far, and we guess 3 for the
next digit in the quotient. Three times 2 and 0 make 6,
three times 7 and 3 make 24. Bring down more of the
dividend, say 456, or at least 4, and guess 4 for the next
digit.
4x2 = 8 and six are 14.
4 X 7 = 28, say 29, and 1 are 30.
Bring down the 5, and guess 2 for the next digit of the
quotient ; twice 2 = 4 and 1 = 5, etc., and then finally bring
down 6, and it goes 3 times exactly.
If the sum is neatly done the corresponding places are
vertically under each other, a detail of appearance emphasised
by the presence of a decimal point.
Let the result be written
52^56 ^ ^3^23.
Do not forget to set also the complementary sum
72 X 73423.
It will be well also to set the exercise whose result is
5286456 ^^
^^ . „ = 72, as a separate sum.
If the connexion is automatically noticed, it is well ; it will
prepare the mind for the later -on extremely important and
constantly occurring connected relations,
if T = c, then - = b, and he = a,
be
but refrain from using this abstract language at present.
Watch for the time when it can without strain be naturally
introduced. It is a great help when that step is reached,
and it represents a vital stage of real mental progress. The
mind should be soaked with particular instances however
before generalisations can be usefully and permanently grasped.
26 EASY MATHEMATICS. [chaf.
Division of money is of course difficult, even when the
divisor is a small number, because of our complex system
of notation, unless the money is first expressed in decimal
form.
To divide by 23 moreover it is not correct to divide by
20 and then by 3 and add the results, as it was with mul-
tiplication. A long-division sum is necessary, and that is no
joke with money as usually specified. Division by 24 can
indeed be done in two stages, by help of its factors 3 and 8
consecutively applied, but that only masks the essential
difficulty by a device applicable only to special cases.
My object in introducing these remarks about complex
money-sums here (and the same thing applies to weights and
measures sums) is to urge that they really belong to a later
stage, and to beg teachers to defer them beyond the early
years at which they are too often introduced. For their
premature employment has often resulted in giving children
an effectual and lifelong disgust with what they have docilely
conceived to be arithmetic ; whereas much of what they had
to do was really a mechanical and overstraining grind, having
as much relation to mathematics as carrying heavy hods of
bricks all day up a ladder has to architecture.
Origin of the symbols.
It is amusing to speculate on the probable origin of the
symbols for the digits. It appears likely that if a single
horizontal stroke meant 1, a double horizontal stroke hastily
drawn would give ~Z. or something like a 2.
It is less easy to make a sort of 3 out of three such strokes,
but it is possible.
The symbol for four would seem to be representative of a
four-sided figure or badly drawn square, '-|- , and the figure
8 was probably originally a pair of such squares Q •
I.] COUNTING. 27
But at this stage it appears likely that some skilled person
took pains to design digit symbols of distinctive form by
combination of a stroke and a semi-circle, making 9, set like
this :
I 2. 3 y b 1 S 10 II etc.,
and that the notion of the value of "place" was a develop-
ment from the further stages of this mode of representation.*
So also it is believed that the Roman symbol X for ten was
the result of counting by strokes and crossing off every tenth
stroke, thus :
1 1 1 1 1 1 1 1 ni 1 1
a practice not unknown among workmen to this day.
Two such crosses would naturally mean 20, etc., while half
a cross or V could conveniently be used to denote 5.
It has been suggested that the rounded M for 1000, W,
sometimes inscribed CIO, if halved, would give the D for 500;
and that a square C for 100, if halved, would furnish an L for
50; but this may be fanciful. The symbol 00 1 00 was used,
it is said, for 100,000, and 0001000 for a million.
* The above however is not history. The real history of the symbols
is complex, and stages of it are given in Dr. Isaac Taylor's learned work
on the Alphabet, especially Vol. II. pp. 263 et seq.
It appears that our digit symbols originated in India, and that several
of them, especially 7, represent a corruption of the initial letters of the
words previously employed to denote the numbers.
" They were introduced by the Arabs into Spain, from whence during
the 12th and 13th centuries they spread over Europe, not, however,
without considerable opposition. The bankers of Florence, for example,
were forbidden, in 1299, to use them in their transactions, and the
Statutes of the University of Padua ordain that the stationer should
keep a list of the books for sale with the prices marked ' not by ciphers
but in plain letters '. . . . Their use was at first confined to mathema-
tical works, they were then employed for the paging of books, and it
was not till the 15th century that their use became general."
CHAPTEE II.
Further considerations concerning the Arabic system
of notation, and extension of it to express
fractions.
Having become acquainted with the fundamental plan of
the system of notation in use, and the mode of expressing
any whole number of things by a combination of ten digits
arranged in places of different value, not all places necessarily
occupied — that is, by means of nine significant digits and a
cipher to express emptiness in whatever place emptiness may
occur, — it is permissible to elaborate it further, with a little
repetition occasionally.
At the beginning of each chapter there is liable to be a little
repetition of something that has already been explained, but
in a slightly different form. This amount of repetition is
purposely introduced and is useful : it is intended to link the
new knowledge on with the old. A new subject should not
be introduced as if it belonged to a perfectly distinct region of
thought; its connexion with what is known should be indicated,
and sufficient of the old should be reproduced to make the
connexion secure. Repetition of a judicious kind is by no
means a thing to be avoided, though it is easy to overdo it ;
and in every way the best kind of repetition is that which
repeats the old idea in a different form of words, or which
looks at something already known from a new aspect.
The beginnings of each new chapter should be easy, and the
CHAP. II.] NUMERATION. 29
steps to higher flights should be regular and moderate, like a
staircase.
Now we know that the symbol 304 means usually that there
are 3 packets of a hundred things each
no packets of tens and
4 single things,
but the '* ten-system," though customary, is not an essential
part of this plan of notation.
40 and 4/- are both constructed essentially on this plan, both
are understood to signify 4 packets and no odd units, though
the number in the packets is not the same in the two cases.
£4 . — • . — signifies again 4 of another variety of packet.
Three dozen and six pennies may be written either 3/6 or
42 pence. It would have been far more convenient if the
human race had agreed to reckon everything in dozens, and so
to express this number by the digits 3 6 instead of by the
digits 4 2 ; but as they have in early semi-savage times arranged
otherwise, we must now make the best of it. The general
idea is the same, only that whereas in ordinary life things are
commonly and conveniently reckoned by dozens, it is customary
in arithmetic to reckon by packets of ten, the symbols being
called digits because they used to be reckoned by actual
fingers : which by some simple persons are so employed still.
Thus whereas 7/6 is understood to mean seven dozen and six
pence, it is customary to mean by 76, seven packets of ten and
six units over; that is to say, if the units were pennies, the
same as 6/4. So also, instead of grouping dozens into a gross,
as in ordinary life, in arithmetic we group tens into a large
packet of ten tens, which we denote by 100. The symbol 346,
therefore signifies six single units, 4 packets of ten each, and 3
packets of a hundred each. If there are as many as ten sets
of 100, they are to be specified by 1000, and so on, as
ordinarily learnt.
80 EASY MATHEMATICS. [chap.
This system of notation extends as far as we like to the left
of the units place, and if six empty boxes follow the digit 1,
it means a million. But we might suppose boxes added to
the right of the units place ; can we find any use for them 1
Let us mark the unit box by a double line nearly round it, so
that in a long row
1 1 1
1
there need be no hesitation about which is the unit box;
then put the digit unity into each box. In the unit box it
means one of some thing, in the next box on the left it means
one packet of ten of those same things, and so on ; each digit
to the left having ten times the value of the one immediately
on the right. If this convention were extended to the box
on the right hand of the units place, the 1 there would
signify the tenth part of a unit, and a 1 in the next box on
the right would signify the tenth of a tenth, and so on.
For we know that not only can we group things together
into an aggregate, it is possible also to cut them up, or split
them into fractions.
Thus the things counted may be bags of money, and each
bag may be known to contain or to be worth 100 sovereigns.
In that case the figure 6 might signify six bags, and so stand
for 600 sovereigns. And each sovereign might be called a
fraction of the contents of a bag, viz., a hundredth part. But
in some of the bags the value might be made up with ten
pound notes, and each of them would likewise be fractions of
the contents of a bag, viz., the tenth part.
Three such notes may therefore be specified either as
3* ten pound notes
or 30- pounds value
or three tenths or -3 of the value of a bag ;
II.] NUMERATION. 31
the stop or mark or point being introduced whenever it is
necessary for clearness. Any mark will do. In foreign
countries a comma is commonly used, whereas we use a dot
placed about the middle of the figure. In early days a |
mark of this kind was used. Thus 346 | 57 used to be written
where we should now write 346*57, or a Frenchman 346,57, the
digits 5 7 being partitioned off to signify that they represent
fractional parts of objects or units ; the digit 6 refers to whole
objects or units, the digit 4 to packets of ten, the digit 5
to fractions of one-tenth, and the digit 7 to one-tenth part
of tenths, that is to say, it signifies seven hundredths of
a unit.
Suppose, for instance, the unit was a bag of sovereigns, as
above specified, then the number written 346J57 or 346 1 57 or
346-57 would mean 346 complete bags of a hundred pounds
each, with 5 ten pound notes and 7 sovereigns loose. The
money specified would be equal in value to
3465-7 ten pound notes
or to 34657* sovereigns
or to 34*657 thousand pound notes
or to -034657 million pound notes,
the position of the figures being changed according to the
unit intended, and the dot or other mark being used to
signify where whole numbers end and fractions begin.
The position of the above numbers relative to each other
is constant, viz. the order 3, 4, 6, 5, 7 ; but their absolute
position, or position relative to the unit place, is different in
the different cases, and is specified by the dot, which is always
and invariably placed after the units digit whenever it is
inserted at all. It is not always necessary to insert it. For
instance the number 3 might be written more completely and
equally well 3* or 3*0 or 3-000, in which case it definitely
I
32 EASY MATHEMATICS. [chap.
signifies 3 units of something, and the 0 would indicate the
fact that there was no fraction to be attended to. If the
dot is placed thus 30*, it would mean 3 packets of ten
units; if placed thus 300-, it means three groups of ten
packets each ; and any digit placed after the dot thus -3 means
a fraction, viz. three-tenths of a unit. Whereas if a digit
occurs 2 places to the right of the dot, as -03, it means three
himdredths of a unit ; as for instance 3 sovereigns would be
3 hundredths or "03 of a bag in the above example, or -3 of
a ten-pound note. Similarly a florin is one-tenth of a pound
or £0-1. Again it is the hundredth, or '01, of a ten-pound
note.
This use of the dot is only a matter of nomenclature, and
its importance lies in its simplicity and convenience. It is
always possible to write -03 as ^fo- i^ we please, just as it
is possible to denote 1864 by MDCCCLXiv if we like; but
it is not so simple.
It may be as well to observe that although there is no
numerical difference between 6 feet and 6*00 feet, there is a
practical and convenient difference of signification. In
practice 6 feet would mean something approximately the
height of a man, whereas 6 00 feet would be understood to
signify either that you had measured a length accurately to
the hundredth of a foot or something like the tenth of an
inch, and found no fraction; or else that you wished some-
thing to be made to that amount of accuracy.
Another way of reading the symbol -03 is three per cent.,
or three divided by one hundred. So also five per cent, is
•05; twenty per cent, is *20; seventy-four per cent, is -74,
and so on.
In the case of twenty per cent, it may obviously be written
•2 or Y% or ^ So also -5 being 5-tenths or 50 per cent, is the
same as | ; and one-half is often the neatest way of speaking
II.] NUMERATION. 33
of it and writing it. Again twenty-five per cent., or -25, is the
same as J, being 25-hundredths ; and '125 or 125-thousandths
is the same thing as ^. Sometimes one specification is
handiest, sometimes the other.
Unfortunately it is not very easy to denote either J or J or
§ in any other very convenient way on our decimal system of
notation, as it would have been if we had arranged to reckon
in dozens.
One-third of 1/6 is easy enough, being sixpence, while
two-thirds is 1/0: but one-third of 16 is an inconvenient
number to write in the ordinary notation. It is ^^-, that is
16 divided by 3, that is 5*333333... without end, as you find
by simple division.
So also I of 16 is 10-6666....
These are called repeating or circulating decimals, and
their frequent occurrence in ordinary transactions is caused
by our unfortunate custom of reckoning in tens instead of
in dozens. A simple circulating decimal may always be
interpreted as so many ninths : thus whereas '3 means 3 tenths,
•333... means 3 ninths, which is the equivalent of one-third;
•6666... means 6 ninths, and so on.
A third of ten is 3^333...
A sixth of ten is 1*666 ...
Two-thirds of ten is 6*666 ...
and even other fractions are not very convenient.
Thus a quarter of ten is 2*5
an eighth of ten is 1 '25
a sixteenth of ten is -625
three-quarters of ten is 7*5
and the only simple things to specify are -5
not often wanted, viz. 2, and a half of ten, which is 5.
L.E.M. C
34 EASY MATHEMATICS. [chap.
This may be contrasted with the convenience of reckoning
in dozens :
a third of a dozen is 4
a sixth of a dozen is 2
two-thirds of a dozen is 8
a quarter of a dozen is 3
half a dozen is 6
three-quarters of a dozen is 9
an eighth of a dozen is IJ
a sixteenth of a dozen is }.
Circulating decimals would not be avoided by the duo-
decimal notation, but they would be rarer, for they would
then in the simplest possible cases signify fifths or sevenths
or elevenths, which are not the commonest fractions to come
across in practice.
It should be remarked that in actual practice circulating
decimals only occur in the translation of numerical fractions ;
and then the decimals always either terminate or recur : but in
real concrete measurement, or subdivision of continuous magni-
tude, circulating decimals never occur, because such a specifica-
tion would signify an infinite accuracy, which is impossible.
In all practical cases measurements can only be accurate
to a certain number of significant figures, and though it may
once in a lifetime happen that these figures are all the same
by accident— as for instance 4*4444 — it cannot matter in the
end whether the last figure is 3 or 5 or even some other
digit. When the figures have expressed the actually attained
accuracy, all subsequent ones are superfluous and even mis-
leading, because they pretend to an amount of accuracy not
really attained.
For this reason the doctrine of circulating decimals belongs
rather to pure than to applied mathematics.
1
¥
1
"ST
1
ten
1
eleven
1
II.] NUMERATION. 35
In the duodecimal system the ordinary fractions would be
denoted as follows : , ^
^=? .
^ = '2497 or approximately '25
^ = •2
^ = -186^35
16
14
12497 or approximately '125
illlll
twelve
Once we have realised the advantages of what is known as
the duodecimal system, it is painful to have to return and
use the decimal notation.
Nevertheless a change from one to the other would necessi-
tate the uprooting of too deep-seated traditions. Among
other things it would alter the multiplication table, that
necessary but laborious thing to learn. In teaching children
it should be realised by the teacher that the multiplication
table is hard and tedious, and too much should not be ex-
pected of them ; but for convenience of life it is one of those
things that it is best to know thoroughly, and it is useful as
a matter of discipline. Its rational basis should be understood,
and experiment should be encouraged in the first instance to
find out what, say, four sixes or seven nines are. It is fairly
easy to see that four sixes will make two dozen, it is not so
easy to see that they will make two packets of ten and four
over, but, the fact having been ascertained, it should be learnt
that four sixes are 24, or four times six are 24 — either way,
whichever happens to be asked, but not both ways at the
same time so as to spoil the rhythm.
36 EASY MATHEMATICS. [chap ii.
Similarly it can be ascertained that five sixpences amount
to half-a-crown or 2/6 ; but that five sixes are 30, that is
they just make three packets of ten.
It is a serious addition to the work of childhood in this
country that they have to learn virtually two distinct multi-
plication tables, viz. the duodecimal pence table and the
decimal or ordinary numerical table. There is plenty of
scope for discipline in these things, and so if it is possible to
relieve the tedium in other places it is permissible.
The extent of multiplication table to be learnt is merely
a matter of convenience, and it is handy to learn beyond
12 times 12. Especially is it convenient to remember that
13x13 = 169 17x17 = 289
14x14 = 196 18x18 = 324
15x15 = 225 19x19 = 361
16x16 = 256 20x20 = 400
Also that 9 X 16 = 12 X 12 = 144 = 1 gross.
[The square numbers may with advantage be specially
emphasised ; 1, 4, 9, 16, 25, 36, and so on ; and it is easy also
as an exercise to ascertain and remember the powers of 2,
especially that 32 is the fifth power of 2 : they are
2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, etc.,
the last written being the 10th power.
A few of the powers of 3 are also handy.
3, 9, 27, 81, 243, 729.
The cubes or third powers of the simple numbers are useful.
1x1x1= 1 Cube of 7 = 343
2x2x2= 8 „ 8 = 512
3 X 3 X 3 = 27 „ 9 = 729
Cube of 4 = 64 „ 10 = 1000
„ 5 = 125 „ 11 = 1331
„ 6 = 216 „ 12 = 1728
All this is to be arrived at merely by simple multiplication,
and the phrase cube number need not yet be used.]
CHAPTEE III.
Further consideration of Division, and introduction of
Vulgar Fractions.
Just as Multiplication is cumulative addition, so Division
may be regarded as cumulative subtraction. Thus, for instance,
when we say that 7 will go in 56 eight times, we mean that it
can be subtracted from 56 eight times. From 59 it can like-
wise be subtracted eight times, but there will be 3 over.
This is the meaning of remainders.
To divide £748. 6s. lid. by £320. 2s. 4d. we can proceed if
we like by subtraction — it happens indeed to be the easiest
way, — and having subtracted it twice, we find that that is all
we can do, and that there is .£108. 2s. 3d. over. So we say
that the smaller sum goes twice in the bigger one, and
leaves a certain remainder.
In general however it is more customary to regard division
as the inverse of multiplication; and, so regarded, it leads
straight to fractions and to factors. Thus the fact that 3
multiplied by 4 equals 12, (3x4 = 12), may be equally well
expressed by saying that 12 divided by 3 equals 4, (-5- = 4),
/12 \ ^ /
or that 12 divided by 4 equals 3, (-^- = 3), or that 3 and 4
are corresponding factors of 12. Similarly 2 and 6 are other
corresponding factors, since 12 -r 6 = 2 and 12-7-2 = 6.
A number like 144, or one gross, has a large number of
factors. It is a good easy problem-exercise to suggest to a
38 EASY MATHEMATICS. [chap.
child to find them all. They are 2, 3, 4, 6, 8, 9, 12, 16, 18,
24, 36, 48, 72. The factors of 1728 are of course still more
numerous. And even the number 60 has a fair number of
factors, viz. 2, 3, 4, 5, 6, 10, 12, 15, 20, 30. These may be
contrasted with the poor show of factors exhibited by 100,
viz. 2, 4, 5, 10, 20, 25, 50.
Children can readily be set to find the factors of numbers,
and will thus incidentally be doing many simple division sums.
Their attention must not however be too exclusively, i.e.
for too long together, directed to integer or whole number
factors; they must be prepared to write down the result of
division when it is not a whole number, but a fraction, or a
whole number plus a fraction. Thus -^ for instance will be
found to be 28 and four over, the meaning of which should
be carefully explained, being first thoroughly understood and
led up to by the teacher.
To lead up to it, it may be pointed out that just as
28 oranges = 20 oranges + 8 oranges
so 28 half oranges = 20 half oranges + 8 half oranges
and 28 halves = 20 halves + 8 halves
and 28 quarters = 20 quarters + 8 quarters ; just as much
as 28 farthings = 20 farthings + 8 farthings.
Now ^ = 14, while — - + ^ = 10 + 4 = 14,
28 ^ ,., 20 8 ^ „ ^
— = 7, while -r + y = 5 + 2= 7,
4 4 4
28 _ 20 8 _2,.8_2-8
28 20 8 ,8
but it is neater to write it
25 3_3„6 ^_
= T+ 5 = ^-^5 = ^ + 10 = ''^-
III.} FRACTIONS. 39
So now the child should realise that, since 144 = 140 + 4, so
i|^ = -|-^ + i; which indicates a division that can be done
and a division that cannot be done. The division that can be
done has the result 28 ; the division that cannot be done is
4 -r 5, and it must be left, either in the form of ^, or in the
form —^ or -8. So the whole result is expressible as 28*8.
A.ccordingly a better way of saying that i|^ is 28 and four
over, is to say that it equals 28 + -|> ^^ 28*8.
To get it in the latter form directly and easily, the original
144 should be written 144*0, and then the sum will run
51 144-0 •, ^ 11
I quite naturally.
28*8
Take another example, because the mind of a child is often
sadly fogged about this elementary and important matter.
3^ =104 = 51:«f:.= 10-333,..,
a result found by simple division, a process which in this case
shows not the slightest sign of terminating but goes on for
ever.
Again -^' = 7 and 1 over, = 7 + 1, or as it is usually written
7J. But in thus writing it the question should occur, How
then would one write 7x11 and why does not 7 J mean seven
halves, or seven multiplied by a half, or 3 J ? It is a mere
convention, and not a consistent one, that 7 J shall signify 7 + 1
and not 7 x J, and some confusion is thereby caused. By no
means need the practice be altered : children must learn to
accommodate themselves to existing practice, and must begin
reform later in life if ever ; but the teacher should realise that
the simplicity of 7 J to him is only because he has got accus-
tomed to it, that it is a confusing thing in reality, and that a
child who is confused by it is likely to be the bright child and
not the dull one.
40 EASY MATHEMATICS. [chap.
Expression of vulgar fractions as decimals.
There is nothing new to be learnt about expressing a
vulgar fraction in the decimal notation, it is only a question
of practice. It is probable that beginners will find no diffi-
culty, but will simply divide out. If any difficulty is felt it
can be met by some such initial treatment as the following :
1 . ,, 2 .3 4 5
- is the same as - or - or - or — ,
and each one of these may therefore be written '5, which means
5 things in the tenths place or compartment devoted to tenths.
A florin for instance is the tenth part of the value of a
sovereign, so 5 florins = J a sovereign. £7'b means 7 pounds
+ 5 florins or £7. 10s. or £7 J.
So also 1 = ? = A = ?i, etc.,
4 8 12 10'
so to express J in decimals we shall have to put 2 J in the
tenths place ; but it is not customary to place fractions there,
the J is best set down as 5 in the next place to the right, as
•25. In that place 5 will mean yj^ths, and that is the same
thing as I a tenth, viz. ^V*h.
So I of a ten pound note = £2. 10s. = £2 J
= £2-5 = -25 ten pound note,
and generally ^ = -25.
So also f = -75, I = -125, etc.
The expression of any fraction as a decimal involves nothing
more than simple division ; thus ^ can be written ready for
operating 7 1 3-00000, and the quotient, written below, will be
•42857 etc.
In this particular instance however there happens to be no
simplification, so the operation is hardly worth performing in
that case.
III.] FRACTIONS. 41
To prove that ?i|^ = 57-57
work thus :
6 1 345-42
57-57'
To find ^y-^ we do the simple division sum
8|3475-000
434-375'
Hence 34^75 ^ ^.3^3^^ . 3;475 ^ .^3^3^^^ ^^
8 0
It is not really necessary to write it out in the division form :
simple division can be performed on the fraction as it stands.
In every case of writing decimal numbers one under the
other, the rule is to keep the column of decimal points
vertical ; in other words, adhere to your system as to which
is the units place, which the tens place, and which the tenths,
etc., throughout.
Extension of the term multiplication to fractions.
i The ordinary idea of multiplication involves the repetition of
the same thing several times, as three times four, or seven nines.
The adding of seven nines together is what is called multi-
plying nine by seven.
The payment of four £5 notes is not called multiplying
£5 by 4 : but if a conjuror extracted ten apples out of a hat
into which one had been put, he might be said to have
multiplied it.
So also seed corn is multiplied into an ear; and thus the
notion of increase is associated with the notion of multiplying,
i But it is best to dissociate the notion of increase from the
notion of technical multiplication, and to be prepared to
multiply by 1 if need be, leaving it the same as before, or even
by 1 leaving it smaller than before. This phrase " multiply
42 EASY MATHEMATICS. [chap.
by a half " is not a simple and natural one : it is a permissible
extension, such as we constantly make in mathematics, when
any operation that has been found practically useful is applied
over the whole range within which it is possible, and some-
times a long way beyond where it appears possible at first sight.
Multiplication by J has some points in common with the
addition of a negative quantity ; it results in diminution, and
it is a process that would not have occurred to us to do except
as an extension of a straightforward process. To multiply by
J and to divide by 2 is precisely the same thing. Why not
call it then dividing by 2 ^ Well, we do very often, but not
always, and a beginner must be content to be told that it is
useful to extend the nomenclature of operations in this way.
We shall speak of multiplication by J if we choose, when we
mean division by 3. We shall occasionally speak of adding
- 4 to a number when we really mean taking 4 from it. We
shall do any of these things when we have good reason for
doing so, and not otherwise.
Suppose we say that 2h sovereigns are equivalent to 50s.,
we arrive at the result by multiplying 20 by 2J, that is first
by 2 and then by J, and adding the separate results. It
would be a nuisance to be obliged to say that we multiply
20 by 2 and divide 20 by 2 and add the results, though it
would be quite true.
The fact that the half of 20 is 10 may be written if we like,
thus : I X 20 = 10 ; or, of course, %»_ or 20 -r 2 = 10 ;
^ of 24 = 8 may be written i x 24 = ^*- = 8 ;
of J may be written J x ^ =
1 .
and that the half of the third of an apple or ribbon is a sixth
of the apple or ribbon is easily verified by experiment. An
experiment need not always be performed; after a time it
can be vividly imagined, with advantages on the side of clear-
ness of apprehension.
III.] FRACTIONS. 4:3
The natural word to use for taking the fraction of a thing
is the word " of," like the half of an orange or a quarter of
a pound or one-sixth of the revenue ; and we shall gradually
find that in all arithmetical cases the word "of" has to be
interpreted as an instruction to perform the operation denoted
Jby X ^ that is to say, the operation we have been accustomejj
to call multiplication.
Practical remarks on the treatment of fractions.
It so happens that the multiplication of vulgar fractions
is easier than addition and subtraction, and so it may take
precedence. One half of one quarter is one eighth : as can
be found by concrete experiment, for instance on an apple,
or by looking at the divisions on a 2-foot rule.
1 nf 1 — 1
1 nf 3 — 3
Y 01 ^ - j-^
Ifvfl— 1.1 of 7^— 7.3p.f7_21
Such a statement as the last must be, and is, led up to;
and gradually the empirical rule can be perceived, that in
multiplication of fractions the numerators must be multiplied
for the new numerator, and the denominators must be multi-
plied for the new denominator.
[But initial difficulties and confusion must be expected
between this and the addition of fractions. Thus, for
instance:
4 "^8" 32 ~ 32'
This is set down here as a warning.
The greatest difficulty in dealing with fractions is felt as
long as they are abstract, "f of what?" is constantly or
should constantly be asked by a child. In the above two
sums the answer to this question would be different : —
44 EASY MATHEMATICS. [chap. lil.
In one it is | of a fraction, viz. | of | of a unit, such as a
foot, that has to be found. In the other it is f of one foot
which has to be added to | of another.]
It is convenient to ascertain and remember that ^ + ^ = -I
[whereas ^ of ^ = yV] ; also that i + xV = h or J-yV = t-
Exercise.
Find the third plus half the third of eight. The answer is
4, but the decimal notation confuses the matter :
J of 8 is 2-6666... and half this is 1-333...
so the sum is 3-9999..., that is 3| or 4.
So also a third + half a third of ten would seem troublesome,
though it results simply in five. But a third + half a third of
a dozen is simple enough, being 4 + 2 = 6. And always
1 . 1 _ 1
Division of fractions may be exhibited thus :
Suppose we have to find what |- -f f amounts to,
write it thus, seven eighths
three tilths
seven x five fortieths 7x5 35 7 5
= 7:X^; .
three x eight fortieths 3x8 24 83'
wherefore instead of dividing by |, we find we may multiply
The idea underlying the above process is that things called
eighths have to be divided by things called fifths, and that
to make it possible they must be expressed in the same
denomination, which in this case is fortieths. Thus we get
the rule, invert the divisor and multiply. Or otherwise
expressed : to divide by a number multiply by its reciprocal.
Division by ^ is the same thing as multiplication by 2. The
symbol -f J is equivalent to the symbol x 4.
CHAPTEE IV.
Further consideration and extension of the idea of
subtraction.
If a man gains £21. 6s. 5d. and loses £15. 4s. 4d., his nett
gain is found by subtraction, and is called the "difference," viz.
£6. 2s. Id. ; the total money which has changed hands being
the " sum," viz. £36. 10s. 9d. A loss may be called a negative
gain ; thus a gain of £10 minus £6, would mean a gain of
£10 accompanied by a loss of £6, or a nett positive gain of
£4. This leads us to discriminate between positive and
negative quantities, and to regard subtraction as negative
addition. Subtracting a positive quantity is the same as
adding an equal negative one.
Geometrically it is sometimes convenient to discriminate
between the journey A to B, or AB, and the journey B to Aj
or BAj just as a French-English dictionary is not the same as
an English-French dictionary. When expressed numerically
a length AB may be denoted by its value, say 3 inches, or
3 miles ; and the reverse journey may be denoted by - 3
inches or - 3 miles, because this when added to, or performed
subsequently to, the direct journey, will neutralise it and
leave the traveller where he started. The two opposite signs
cancel each other in this sense, and the two quantities added
together are said to amount to zero algebraically — that is
when their signs are attended to, and as regards the end
46 EASY MATHEMATICS. [chap.
result only ; but the traveller will himself be conscious that
although he is where he started from, he has really walked
6 miles ; so that for some purposes such quantities may be
added, and they are then said to be "arithmetically" or
better " numerically " added ; for other purposes they are to
be "numerically subtracted," or, as it is called, "algebraically
added," that is with their signs attended to, and with "minus"
neutralising an equal " plus."
If a height above sea level is reckoned positive, a depth
below may be reckoned negative ; so that a well may be
spoken of either as 60 feet below or as - 60 feet above the
sea level.
The latter mode of specification sounds absurd, but one
should gradually accustom one's self to it, for practical pur^
poses later on.
If children feel a difficulty with these negative quantities,
as they have every right to, they can be accustomed to them
gently, as a horse to a motor car. Mathematicians found
some difficulty with them once upon a time, so the difficulty
is real, though like so many others it rapidly disappears by
custom. Debts, return journeys, fall of thrown-up stones,
losings, apparent weights of balloons or of corks under water,
dates of reckoning B,C., and many other things will serve as
illustrations; not, however, to be taken all at once.
Time is the one thing that never goes backwards; but
nevertheless intervals of time may be considered negative if
they date back to a period antecedent to the era of reckoning.
In a race, for instance, it would be an ordinary handicap
if one of the competitors was set 12 yards behind scratch, or
if he was made to start from scratch 3 seconds late. In either
case he could be said to have a negative start.
In golf handicaps it is customary to denote these positions
behind scratch as positive, because they are added to the
IV.] NEGATIVE QUANTITIES. 47
score. This is because the object in golf is to get as low a
score as possible, not a high one as at cricket.
Addition and subtraction of negative quantities.
Suppose a man inherited a lot of debts, his property
would be diminished by their acquisition. The addition to
it would be negative, and would be indistinguishable from
subtraction.
A debt of £300 added to a possession of £500 would result
in nett property of £200 ; which we might express by saying
that -3 + 5 = + 2.
Or of course the debt might exceed the possession and
leave a balance of debt. For instance -8 + 5 = - 3 ; where
the unit intended by these digits might be a hundred or a
thousand pounds. This may be taken as an illustration of the
gain of a negative quantity. Take another.
An axe-head at the bottom of a river weighs 3 lbs. Some
corks, which, when submerged, pull upwards with a force
equal to the weight of 49 ounces, are attached to the mass
of iron. Its weight is thus more than counteracted, and it is
floated upwards with a force equal to the weight of 1 ounce,
, because 48 - 49 = - 1.
I . A raisin at the bottom of a champagne glass, or a speck of
grit in a soda-water bottle, can often be seen to accumulate
bubbles on itself till it floats to the surface and gets rid of
some, when it sinks again, and so on alternately.
The negative or upward weight of the corks, or of the
bubbles, counteracts and overbalances the positive weight of
the iron or of the fruit. It may be said that we have sub-
tracted more weight from it than it itself possessed, and so
left it with a negative weight — like a balloon. The weight
of a balloon is not really negative, but it superficially appears
to be ; because the surrounding air buoys it up with a force
48 EASY MATHEMATICS. [chap.
equal to the weight of the air it displaces, which represents a
greater weight than its own.
When we have to subtract a bigger number from a smaller,
we must not always merely say we cannot do it. It is con-
venient in subtraction sums to say so, and to " borrow " from
the digit in the next higher place (i.e. to undo one of the
available packets and bring the contents one step down), so
long as there is something there to be " borrowed," but if we
perceive that at the end of the sum there will be a manifest
deficiency we must proceed differently.
Suppose we were told to collect £S from a man who had
only £Sj we could not really do it; but we might report to
our chief, "if we do we shall leave him £5 in debt to
somebody," which could be expressed arithmetically thus :
3 - 8 = - 5.
Suppose we were told to pull 5 feet of a gate-post out of
the ground, and when we came to try we found that it had
only 2 feet buried; we might at first say that it could not
be done; but on second thoughts we could say that it was
hard to do, and that the only plan we could see would be to
pull it minus 3 feet out first, that is to get a mallet and drive
it 3 extra feet in, before pulling at it at all.
Suppose a stone were 30 feet above the ground, and we
were told to drop it 36 feet, that is to subtract 36 feet from
its height of 30 feet. It would not be easy to do, but it
could be done, for we might dig a hole 6 feet deep; or it
might even be sufficient if we dropped it over a pond of that
depth. In either case it would afterwards be 6 feet below
the surface of the ground, for 30 - 36 = - 6 ; it would then
be at an elevation of - 6 feet, which means the same as a
depression of 6 feet.
To speak of a depth of 6 feet as a negative height, in
ordinary conversation, would be absurd ; but to interpret an
IV.] NEGATIVE QUANTITIES. 49
arithmetical answer, which gives a height as - 6 feet, to mean
that a thing is not elevated at all but is depressed 6 feet,
would be quite right and in accordance with commonsense.
Hence the following examples are correct :
4- 9 = - 5
17- 39 = - 22
546-827 = -281
But now here is a necessary caution. Take the last case.
We see that it is right, for if we add 281 to 546 we get 827 ;
but suppose we had put it down like an ordinary subtraction
sum and noticed nothing wrong with it, it would have
looked like this
546
827 (example of the way not to do it).
-319
We should have said in the old-fashioned way 7 from 6 we
cannot, so borrow 10 from the next place; 7 from 16 is 9, put
it down. Now we have either 2 from 3, or what is more
commonly said, and comes to the same thing, 3 from 4,
leaving 1, which we put down; and then we have to take
8 from 5. There is nothing more to borrow, so we must set
it down as - 3. Well that is not wrong, but it requires
interpreting, and it is not convenient. The minus sign only
applies to the 3, which, being in the third place, means 300 ;
the other figures, the 19, were positive. Hence the meaning
is -300 + 19, or in other words -281. It might be written
319, with the minus sign above and understood to apply only
to the digit 3, but it could not properly be written - 319.
The above is therefore a very troublesome way of arriving
at the result. The convenient way is not to begin performing
the impossible subtraction, but to perceive the threatening
dilemma, and invert it at once ; then subtract the smaller
I4.E.M, D
50 EASY MATHEMATICS. [chap. iv.
number from the bigger in the ordinary way, labelling the
result however as negative. This is of course what we really
do when we say 5-8 = - 3. We do not begin saying " 8 from
5 we cannot, so borrow" from nowhere, for there is nowhere
to borrow from. We stop, invert the operation, and record
the result as negative ; because a-h = - (b-a).
One more case we must take however, viz. where the
quantity to be subtracted is itself negative : and its subtrac-
\. tion therefore represents a gain. The loss of an undesirable
• burden was esteemed by Bunyan's Pilgrim to be a clear gain.
A negative subtraction is a positive addition.
6-(-3) = 9; 7-(-9) = 16.
This is sometimes expressed by saying that two minuses
make a plus. The effect of a minus is always to reverse the
sign of any quantity to which it is prefixed, so if applied to a
negative quantity it turns it into a positive quantity. It is
equivalent to more than the removal, or subtraction, of a debt,
which would be effected by an equal sum added. A loss is
more than neutralised by a negative sign, it is reversed.
Add - 31 to 114, the result is 83 ; but subtract - 31 from
114, and the result is 145.
No more words are necessary. Familiarity and practice
will come in due course as we proceed. A surviving puzzle
may occasionally be felt, and can from time to time be
removed. It is a mistake to hammer at a simple thing like
that till it becomes wearisome ; for trifling puzzles or foggi-
nesses evaporate during sleep, and in a few years have
automatically disappeared, from children properly taught.
They continue to trouble too many adults at present.
CHAPTEE V.
Generalisation and extension of the ideas of multi-
plication and division to concrete quantity.
The idea of multiplication arose as a convenient summary
of a special kind of addition, viz. the addition of several
things of the same magnitude to each other. Thus four sixes
added together, if counted, make 24, and so it is summarised
and remembered as 4 sixes are 24, or 4 times 6 = 24 ; and 4
and 6 are called ' factors ' of 24.
Originally therefore the two factors in multiplication
signified, one of them the size of the quantity of which several
are to be added together, and the other the number of times
it was to be so added.
Thus 3x6, read 3 times 6, meant a summarised addition
sum, 6 + 6 + 6. But if read 6 times 3 it meant the addition
sum 3 + 3 + 3 + 3 + 3 + 3. That the result is the same may be
treated as a matter of experience, and may be demonstrated
by grouping, but it is not to be regarded as
self-evident. Nevertheless the diagram (fig. 2) ••....
demonstrates that 3 rows of 6 each is the ••••••
same as 6 columns of 3 each. And the ^^' ^*
counting of window panes and postage
stamps are illustrations of practically the
same thing.
^ — ' — ' — ' Thus we get led to the area of a
Fio. 3. rectangle of length a and breadth h as
a X J, or briefly written ah (fig. 3).
52
EASY MATHEMATICS.
[chap.
But the idea of multiplication soon generalises itself, and
the expression ah gets applied to a number of things to which
a simple numerical idea like 3 times 6, or a times h, would
hardly apply.
It may be worth showing however that the numerical
notion will apply further than might have been anticipated,
for instance the rectangle (fig. 4) is built up of 5 equal staves
each of them say 3 inches long and an inch wide. The area
of each stave is thus 3 inches x 1 inch, or 3 square inches.
And by adding 5 of the staves together (or multiplying one
of them by 5) we get the total area.
Fig. 4.
Fio. 5.
And the same area could be equally well obtained by
putting together 3 staves each of 5 square inches area (fig. 5).
The number 12 can be resolved into two factors 3 and 4, as
is shown by the annexed group which consists of 3 rows of
four dots each, or of 4 columns of 3 dots each, proving that
3 times 4 -= 4 times 3.
Fig. 6.
Fig. 7.
A dozen can equally well be grouped as in fig. 7 : its large
number of factors confers distinction on the number 12.
The number 10 has only two factors, viz. 2 and 5, since
the name "factor" is usually limited to whole numbers. It is
possible to say that 3| is a factor of 10, because if it be
v.] CONCRETE QUANTITY. 53
repeated 3 times the number ten results ; as is shown by the
following set of 3^ disks repeated 3 times, where the central
sectors have each of them an angle 1 20° or q
1^ of a revolution, and so make up a disk q
when put together. But the name " factor " q
is not usually applied to fractions. O O O ®
Again, a slab of any given area and unit Q
thickness will have a bulk which, measured O
in cubic inches, is numerically equal to its Q
area in square inches. If such a slab is mul- ^^°" ^'
tiplied or repeated, each slab being piled up on similar ones,
say 7 times, then 7 times its bulk will give the volume of a
rectangular block ; or the volume of a block may be said to
be obtained by multiplying its length, breadth, and height.
There is no reason to take one of these factors as numerical
more than another, and the truth is that none of them need
be numerical.
When we say volume = Ibh, or length x breadth x height,
we may and should mean by I the actual length,
hy b „ „ breadth,
and by h „ „ height,
— not the number of inches or centimetres in each — and the
resulting product is then the actual volume, and not any
numerical estimate of it. [If anyone disagrees with this
they are asked to withhold their disagreement for the present.
This is one of the few things on which presently I wish to
dogmatise. See Chap. XXVI. and Appendix II.]
From this point of view the symbols of algebra are concrete
or real physical quantities, not symbols for numbers alone, and
algebra becomes more than generalised arithmetic.
In such cases however the old original definition of multipli-
cation requires generalisation, and a good deal can be written
on it ; but no difficulty arises, and the question, being inter-
54 EASY MATHEMATICS. [chap.
esting chiefly from the philosophic point of view, does not in
this book concern us.
"We may proceed without compunction to multiply together
all sorts of incongruous things if we find any convenience in so
doing. Thus, a linear foot multiplied by a linear foot gives a
square foot,
6 feet X 3 feet gives 18 square feet,
4 feet X 3 feet x 2 feet gives 24 cubic feet.
In all these cases something real and intelligible results;
but if we multiply square feet by square feet, nothing intelli-
gible results ; consequently such a process will never appear
in a correct end-result, though we shall find that it often
appears as a step in a process without any detriment.
Again we may multiply a weight by a length, say 3 lbs. by
7 feet, and get what is called 21 foot-lbs., where the unit has
a meaning which can be interpreted, viz. the work done in
raising a 3 lb. weight 7 feet high against gravity, or else the
moment of a force round an axis. But if we try to multiply
3 lbs. by 7 lbs., we should get 21 square lbs., which has no
intelligible meaning and is nonsense. There is nothing in the
symbols to tell us whether it is sense or not : operations can
be consistently performed even on meaningless symbols.
To discriminate sense from nonsense, appeal must be made
to reality and to actual life and instructed experience.
Division is merely the inverse of multiplication, and similar
considerations apply to it.
If we divide 1 by any quantity we get what is called the
reciprocal of that quantity.
Thus J is the reciprocal of 2. xir is the reciprocal of 10.
— P — is the reciprocal of a length, and could be read
o leet
1 per yard.
v.] CONCRETE QUANTITY. 55
-— — -^ might represent the number of telegraph posts
per mile.
— -. is the reciprocal of a time, and might be read
^^ 'once every tenth of a second'; or it could be
simplified into a repetition of something ten
times a second, or 10 per second. It is what
is called a ' frequency,' and is in constant use for
vibrations.
-^ is a slow frequency, the frequency with which
a cycle of astronomical eclipses approximately
recurs.
6000 revolutions
5 minutes
wheel of a small engine, and may be read as
1 200 revolutions per minute, or 20 revolutions
per second.
If we divide a length by a time, as for instance
. ^^ ^^, we get a velocity; e.g. the speed of an express
l^"""" train.
is exactly the same velocity.
is a frequency of rotation, as of the fly-
1 second
^^ ^^, or approximately ^^^ ^ , is a walking pace.
No hesitation must be felt at thus introducing the units
into the numerator or denominator of fractions. If they are
left out, the residue becomes a mere numerical fraction, the
ratio of two pure numbers ; whereas with the units inserted
they are real physical quantities with a concrete meaning, and
are capable of vaTied numerical specification.
56 EASY MATHEMATICS. [cha1».
Thus the velocity of sound in air at the freezing point is
1090 feet 33000 centimetres 1 mile
or 5 =, or
1 second 1 second 5 seconds
•33 kilometres 1 kilometre • . i
or — . — or -^; -^ — approximately
1 second 3 seconds
10 minutes' walk 240000 miles
3 seconds a fortnight
First idea of involution.
When a number of the same things were added together
many times, the process was specially treated and called
multiplication. When a number of the things are multiplied
together several times, the process is likewise worthy of special
treatment, and is called " involution " or the raising of a thing
to a certain "power."
The raising to a power is compressed or summarised multi-
plication. The expression 4x3 meant four added to itself
3 times (or 12), whereas 4^ is understood to mean 4 multiplied
by itself 3 times (or 64).
So 25 = 32, 63 = 216,
103 = 1000, 106 = a million,
122 = 144^ and can be read 12 square, for short ; though
really a square number is an absurdity. It is called " twelve
square" because if the 12 represented inches, 12^ would mean
a square foot.
If ft is a length, a^ is truly a square whose side is of length
a, and a^ is truly a cube whose side is of length a. So 4^ is
read "4 square," and 6^ is often read "six cube," by analogy.
It is also true that 2* = 16, but here there is no geometrical
analogy, and it is read " 2 to the fourth power " simply, the
word "power" being often omitted in practice. Similarly a
million is " ten to the sixth " or 10^.
v.] CONCRETE QUANTITY. 57
A length divided by a time is a velocity (v)y and a velocity
divided by a time is an acceleration (a).
V
So in mechanics we find such an expression as
where f^ is often read as the square of the time, although
strictly speaking such an expression is nonsense. We can
have a square mile, but not a square fortnight; there is no
meaning to be attached to the term ; time cannot be multi-
plied by time with any intelligible result. Whenever such an
expression occurs, it is to be understood as an abbreviation
for something : in the above case for this
s = ■^{at)t,
where the at is v, and is a real and simple physical quantity.
s is a velocity multiplied by a time, and the double
reference to time is caused by the introduction of the specially
defined quantity "acceleration," which is often expressed
correctly as so many feet per second per second; the two
units of time in the denominator being conveniently spoken of
as the square of the time — by analogy with geometry again —
without thought and without practical detriment, though
confusing to anyone who seeks a real philosophic meaning in
the expression.
CHAPTEE VI.
Factors of simple numbers.
A CHILD should be encouraged who notices that no factor is
ever greater than half the number; for though there is nothing
in that but what is obvious, yet that is the type of noticing
which frequently leads to observations of interest. An even
number always has this largest factor, but an odd number
can never have a factor greater than a third its value ; and
frequently its largest factor is less than this. Some numbers
have no factors at all; like 7 and 11 and 13 and 29 and 131.
These are called prime numbers, and a child should make a
small list of them as an exercise. But do not attempt to
make it learn them or anything of this kind by heart. Ease
and quickness of obtaining when wanted is all that is
practically needed.
A child should be encouraged to discover criteria for the
existence of simple factors; but is hardly likely to be able
to notice the facts without aid.
Any number (written in the decimal notation) which is
divisible by 3 (i.e. which has 3 as a factor) has the sum of its
digits also divisible by 3. But this, though convenient as a
rule, is in no sense fundamental : it depends merely on our
habit of grouping in tens. In the duodecimal system every
number ending in 0 would necessarily be divisible by 3 as well
as by 4 and by 6 ; and extremely convenient the fact would be.
CHAP. VI.] FACTORS. 59
For instance, 1/- and 2/- and 4/- and 5/-, or any number of
shillings, can be divided by 3, 4, or 6 ; that is, can be parcelled
out exactly into a whole number of pennies.
By reason of the system of reckoning 12 pence to a
shilling, any sum of money can be subdivided into three or
six equal parts without halfpence or farthings ; thus J of a
pound is 6s. 8d., two-thirds is 13s. 4d., one-sixth is 3/4, one
8th is 2/6, and one-twelfth is 1/8.
In the decimal notation a number has to end in 00 in order
to be certainly divisible by 4; and in 000 in order to be
certainly divisible by 8. And the division is seldom worth
doing even then, because it hardly results in simplification.
The number 5 in the decimal system has an artificial sim-
plicity conferred upon it, but it is not often that we should
naturally group things in 5, except for the accident of our
5 fingers : and one of them is a thumb.
The advantage of working in at least two diff'erent scales
of notation is that it becomes thereby easy to discriminate
what is essential and fundamental from what is accidental
and dependent on the scale of notation employed. Thus the
curious properties of the number nine or eleven are artificial,
and in the duodecimal scale are transferred to eleven and
thirteen respectively.
The well-known criterion for divisibility by 3 or 9, viz.
whether the sum of the digits is so divisible, is accidental
again, and disappears in another scale of notation — for instance
when units are grouped in dozens instead of tens, — to give
place however to a much simpler rule.
The rule about divisibility of the sum of the digits applies
to eleven in the duodecimal scale, and indeed would always
apply to the number which is one less than the group number
artificially selected.
But the existence and identity of prime numbers is not
60 EASY MATHEMATICS. [chap.
accidental at all, but fundamental, and so also is the existence
of any given numbers of factors to a number — however it be
specified.
Thus one gross can be parcelled out into factors or equal
groups in a given number of ways, whether it be denoted by
1/0/0 or by 144 or by any other system of notation.
So also the number one-hundred has only six factors
whether it be denoted by 8/4 or by 100 (one nought nought),
and its factors are (in the duodecimal scale) :
4/2 2/1 1/8 t 5 2,
that is these actual numbers, however they are denoted. In
the duodecimal scale it is needful to have single symbols for
ten and eleven ; and the initial letters serve the purpose.
An actual number is easily exhibited by means of counters
or coins or marbles : its expression in digits is an artificial
arrangement and is adopted simply for convenience : it is
analogous to sorting the marbles into bags of which each
must contain an equal number — whatever number may be
chosen as suitable and fixed upon for the purpose.
It may be interesting to write down the numbers in the
duodecimal scale which would be divisible by 5.
5, t, 1/3, 1/8, 2/1, 2/6, 2je, 3/4, 3/9, 4/2,^4/7, 5/0, ...,
and the even numbers in the above are divisible also by ten.
The above numbers should be read five, ten, one and three,
one and eight, two and one, two and six, two and eleven, etc.,
meaning one dozen and three, one dozen and eight, two dozen
and six, two dozen and eleven, etc.
Numbers which have the factor 7 are
7, 1/2, 1/9, 2/4, 2/^, 3/6, 4/1, 4/8, 5/3, 5/^, 6/5, 7/0, ...,
and the even ones are divisible also by fourteen.
Numbers which have the factor eleven {e) are
e, Ijt, 2/9, 3/8, 4/7, 5/6, 6/5, 7/4, 8/3, 9/2, tjl, e/0, ...,
VI.] FACTORS. 61
namely eleven, one and ten, two and nine, and so on : the
last one written being read eleven dozen.
Numbers divisible by thirteen (1/1) are
1/1, 2/2, 3/3, 4/4, 5/5, 6/6, 7/7, 8/8, 9/9, % e/e, 1/1/0, ....
In the last two cases a law or order among the digits is
manifest, but in all four cases it may be noticed that every
digit makes its appearance in the units place, though only in
the last two cases do they appear in a simple order.
Numbers divisible by 3 are
3, 6, 9, 1/0, 1/3, 1/6, 1/9, 2/0, 2/3, 2/6, 2/9, 3/0, ....
and the even ones are divisible by 6. Every third one of the
above series, viz. those in thick type, are divisible by 9.
Numbers divisible by 4 are
4, 8, 1/0, 1/4, 1/8, 2/0, 2/4, 2/8, 3/0, 3/4, 3/8, 4/0, ....
Alternate ones are divisible by 8, and those in thick type
are divisible by sixteen.
Numbers divisible by twelve, that is arrangeable in dozens,
are of course,
1/0, 2/0, 3/0, 4/0, etc., 1/0/0, ...,
the last written being the symbol for a dozen dozen or one
CHAPTEE VII.
Dealings with money and with weights and measures.
In the British Isles it is customary to count pennies by the
dozen, the value of which when coined in silver is called a
" shilling " ; and shillings are counted by the score, the value
of which is called a "pound sterling," or when coined in gold
a " sovereign." Five dozen pence, or a quarter of a pound
sterling, when in a single silver piece used also to be called a
"crown." And these, together with the half-sovereign, half-
crown, half-penny, etc., are the chief names in vogue ; except
the "guinea" and the "farthing," neither of which need much
concern us. The " florin " is an attempt at a decimal coinage,
being the tenth of a pound; and the double-florin is an
attempt at an international currency or equivalence with the
dollar and the five-franc piece.
The addition of money is a practical operation in constant
use, and plenty of practice in addition is obtainable by its
means. No other addition sums are worth attention for their
own sake : but in addition of money it is worth while taking
pains to acquire a fairly quick and accurate style. At the
same time it is to be remembered that it is a purely mechanical
process — one that in large offices is better, more rapidly and
more accurately, performed by a machine, into which the
figures are introduced by pressing studs, and then the addition
performed instantaneously by turning a handle.
CHAP. VII.] MONEY. 63
Nothing that can be performed by turning a handle can be
considered an element in a liberal education : it can only be a
practical and useful art. That however it is ; partly because
a machine is seldom available, partly because it is ignominious
to be helpless without a tool of this kind, chiefly because
addition of money is an operation which is called for by
commonplace daily life more often than any other.
Nothing much need here be said about it. The columns
of an actual account hook are the best addition sums to set for
practice. Also, in writing figures down, it is well to take
care to place the unit digits under each other, leaving a place
for a left-hand digit whenever such occurs in the pence and
shillings columns, and to be equally careful to write the
pounds with the corresponding places vertical. Also to write
all figures very plainly. This last always, and for all purposes :
A good clear style of figure- writing should be cultivated.
Subtraction of money is greatly facilitated by the use of
the " shop " method : the old-fashioned process of " borrowing"
was troublesome, and moreover only enabled one row of
figures to be subtracted from one row, whereas with the shop
or complementary method any number of rows may be sub-
tracted from another row, and the process is practically only
addition. For instance suppose it is wished to subtract all
the smaller amounts from the larger in the annexed statement:
£
s.
d.
341
8
7
less 19
5
9
and 14
0
3
and 36
17
5
271
5
2
The process is, to say, 5 and 3 and 9 make 1/5 and 2 make
1/7, put down the 2d. and carry 1/-; then 18 and 0 and 5
make 23 and 5 make 28, put down 5/- and carry £1 ; then
64 EASY MATHEMATICS. [chap.
7 and 4 and 9 make 20 and 1 make 21, put down 1 and carry
2; then 5 + 1 + 1 = 7 and 7 make 14; and finally 1 and 2
make 3. In reading, emphasise all the black figures.
Verify by adding the four lower lines.
As to multiplication and division of money or of weights
and measures we will deal briefly with them : the old-
fashioned practice in such matters was tedious and was
pushed in childhood into needless intricacies. Dulness is
apt to line all this region, unless skill is expended on it and
due care taken, and no more practice should be enforced in
it than is required for ordinary life. Discipline and punish-
ment lessons might possibly with advantage be confined to this
region. Even for punishment it is however hardly necessary
to inflict sums dealing with acres, furlongs, poles or perches ;
or with bushels, pecks, scruples, quarters, pennyweights, and
drams. Hogsheads, kilderkins, and firkins may perhaps at
length be considered extinct, except for purposes connected
with the study of folk-lore. There are plenty of real and
living units to be learnt in Physics : we need not ransack
old libraries and antique country customs for them. And,
though the humanity involved in and represented by old
names has been a relief to some children, during their dismal
lessons, far too much has been made of the trivial and dull
operations suggested by tables of British weights and measures.
The sooner most of them are consigned to oblivion the
better.
Real living arithmetic is the same in any country ; and
the most important of all is that which must necessarily
be the same on any planet.
The units that are at present worthy of terrestrial attention
are the following :
Units of length — inch, foot, yard, mile, millimetre,
centimetre, metre, kilometre.
VII.] UNITS. 66
units of time — second, minute, hour, day, week
year,
units of area — square mile, square foot, square
centimetre, etc.
units of volume — cubic foot, cubic inch, cubic yard,
cubic centimetre, cubic metre ;
occasionally also litres, gallons
and pints,
units of mass — pound, ton (ounce, grain, hundred-
weight occasionally), gramme,
kilogramme, milligramme,
units of money — pence, shillings, pounds, francs,
marks, dollars.
But conversion from one to the other of the last-mentioned
denominations should in every case be only approximate.
Accurate work when wanted is done by tables, and the rate
of exchange is constantly varying.
For division of money, and of weights and measures, the
orthodox school rule is called "practice"; and it sometimes
happens that by excessive practice children are able to do
this kind of sum much better than adults — better even
than mathematicians; but since school time is limited, such
extravagant facility in one direction is necessarily balanced
by extreme deficiency in many others, and is therefore to be
deprecated. The world is too full of interest to make it
legitimate to exhaust the faculties of children over quite
needless arithmetical gymnastics, which confer no mathe-
matical facility, but engender dislike of the whole subject.
Modern treatment of the rule called " Practice." The
practical advantages of decimalisation.
In old days some very long sums used to be invented for
British children whereby our insular system of coinage and of
66 EASY MATHEMATICS. [chap.
weights and measures was pressed into the service to make
difficult exercises. The form was usually something like this :
Find the cost of 131 tons 5 cwt. 3 qrs. 24 lbs. 5 oz. at
£4. 13s. 9id. a ton.
Mr. Sonnenschein led the way, I think, towards taking all
the sting out of these outrageous problems, and reducing them
to useful though unimpressive and essentially insular exercises,
by introducing the chief advantage of the decimal system into
the working, before it had been embodied by Parliament in
a legal system of weights and measures and coinage itself.
If such sums have to be done, and a moderate amount of
" practice " in that direction is quite legitimate, decimalisation
of at least one of the quantities specified, that is, expressing it
in terms of one denomination, is undoubtedly the proper
initial step to take ; and then if we are asked the cost of so
much goods at a given price, the matter becomes a mere
straightforward multiplication ; while if we are asked to find
the price of a given amount of goods which have cost so much
money, or the amount of goods which can be obtained for a
given sum of money at a given price, we have only a straight-
forward division sum to do ; — once the complication of many
denominations, that is to say the "compound" nature of the
specification, with scales of notation mixed up, is by an initial
process got rid of. It is always possible, and sometimes
advocated, to reduce everything to the lowest denomination,
e.g., in the sum above to halfpennies and ounces ; but that is
terribly long and tedious. Expression in terms of the highest
denomination is much neater. The initial process is as
follows :
Decimalisation of money.
To express any sum of money in terms of a single unit, say
£1, which is the best unit for the purpose, it is sufficient to
viT.] PRACTICE. 67
notice and remember a few simple convenient facts. They
are all painfully insular, and are not an essential part of real
arithmetic at all, but if properly and lightly treated they
afford to British children an amount of easy practice which
foreign children are destitute of. It is only when trivial facts
and insignificant sums are laboured at, till they kill all interest
of the British child in real arithmetic, that they become deadly
and deserving of the harshest epithets.
The decimalisation of money in terms of a pound is easy,
since a florin is the tenth of a sovereign ; so any number of
shillings is easily expressed in decimals of a pound.
2/- = <£•!, 1/- = £-05,
3/- = £-15, 6d. = £-025,
4/- = £% 1/6 = £-075,
5/- = £-25, 2/6 = £-125,
6/- = £-3, 3/6 = £-175,
7/- = £-35, 4/6 = £-225,
and so on. etc.
A penny is 2To*^ ^^ ^ pound, but that is not specially
convenient when expressed as a decimal ; a farthing is -g^^^th
of a pound, and that is approximately iwuo ^^ £'001.
Since money is never needed closer than to the nearest
farthing, except in the price of cotton per lb. and a few rare
cases, the approximation of £-001, sometimes called a mil, for
1 farthing, or the writing of a farthing instead of £'001, often
suffices ; especially in interpreting results.
The following expressions are all equivalent in value :
^ a sovereign = 10/- = 5 florins = £-5.
So also are the following, each row among themselves :
£7. 10s. = £7i = £7-5 = £7 + 5 florins.
J of a ten-pound note = '25 ten-pound note = £2*5 = £2. 10s.
15/- = £| = £-75.
68 EASY MATHEMATICS. [chap.
150/- = 75 florins = £7'5 = £7. 10s.
18/- = 9 florins = <£-9.
12/- = 6 florins = £'6.
£1. 12s. = <£l-6.
£4:. 18s. = £4-9.
£7. 19s. = £7-95.
All these expressions should be read backwards as well as
forwards.
So also
£b. 2s. 6d. = £5-125.
£3. Is. 6d. = £3-075.
£3. lis. 6d. = £3-575.
£3. lis. ^d. = £3-577, almost exactly.
Take a few examples of the interpretation of decimals of a
pound into ordinary coinage :
£1-2 = £1. 4s.
£4-25 = £4. 5s.
£7-904 = £7. 18s. Id., the four mils being practically a
penny.
£13-127 = £13. 2s. 6Jd., the -125 being 2/6, and the 2 extra
mils ^d.
£1-178 --= £1. 3s. 6|d., the -15 being 3/-, -025 = 6d., and
there being 3 mils more.
£-025 = 6d.
£-026 = 6|d. almost exactly.
£-027 = 6Ld.
£-028 = 6|d. „ „ .
£-029 = 7d.
£•030 = 7 |d. „ „ or exactly 7ld.
We are now ready to do any number of sums like the
following :
Find the cost of 324 horses at £17. 9s. 6d. a horse.
Now 9/6 = 8/-+ 1/6 = £-4 + £-075 = £-475,
VII.] PROBLEMS. 69
so the answer is merely
324x<£l7-475 = £5561-9 = £5561. 18s.
Find the cost of 900 things at £9. 7s. 4Jd. (Sonnenschein.)
Answer is £9-36875x900 = £8431-875 by simple multiplication
= £8431. 17s. 6d.
How much a year is £31. 9s. 9d. per day?
Answer 365 x £31-4875 = £11492-9375 = £11492. 18s. 9d.
How much interest must be paid for 43 days' loan of a sum
of £543. 17s. 6d. at the rate of 3 J per cent, per annum"?
(Sonnenschein.)
Here £3| must be paid for each hundred pounds lent for a
year, so for 43 days only ^^g^^ths of that sum has to be paid.
Now 17/6 = 8 florins + 1/6 (or, otherwise, seven-eighths of a
pound) = £-875 ; so the amount to be paid is :
that is to say, x £5-43875.
This yields £2-243 = £2. 4s. lOJd., the answer.
Typical exercises.
There are certain time-honoured exercises of a type such as
the following, in which a fair amount of practice is desirable.
[Type only here given.]
If 3 peaches cost a shilling, what will 20 cost *?
If I have to pay 15 workmen at lOd. an hour for 8 hours,
how much money do I need 1
If the butcher supplies 7 J lbs. of meat for 5s., what has he
charged per pound 1
70 EASY MATHEMATICS. [cHAf.
and so on. The last being a troublesome kind of sum fre-
quently occurring to housekeepers, but usually and most easily
done by tables.*
Examples like these are quite harmless and give needful
practice, but when they become complicated a little of them is
sufficient, except for discipline, and the more concrete and
amusing they can be kept for ordinary purposes the better.
A slight further development, not quite so harmless, is of
the following type :
Find the cost of 6 lbs. 11 oz. 9 dwt. at 1 7s. 8 Jd. per ounce.
In British schools there is far too great a tendency to limit
all exercises to pseudo-commercial matters. In real business
this kind of sum hardly occurs ; and besides, greater interest
can be obtained by opening up fresh ground for the sub-
ject matter of examples.
A few specimens may be here suggested. A great deal of
what has to be laboriously taught later as physics is nothing
but simple arithmetic, and could easily be assimilated uncon-
sciously while doing sums.
1. If the sound of thunder takes 10 seconds to reach our
ears, how far has it come % (See p. 56 for velocity of sound :
it travels approximately a mile in five seconds. For more
accurate specification the temperature would have to be known.)
* Answers to these sums are as follows : Each peach costs the third
of Is., so twenty peaches will cost the third of £1, or 6s. 8d.
The 15 workmen's wages will amount to £5, since 80d. is the third
of a pound.
The price is 8d. a lb., since 7id. doubled three times makes 60d.
or 5s.
But it will be observed that in each case some accidentally convenient
relation is seized and utilised. That is the essence of mental arithmetic:
it is a training in quickness and ingenuity, not in mathematics ; and its
merits can be appraised accordingly.
VII.] PROBLEMS. 71
2. If a pistol shot is heard across an estuary 15 seconds
after the pistol was fired (which can be told by observing the
flash), how wide is the estuary ?
3. If light reaches the earth from the sun in 8 minutes,
what is its velocity? (The distance of the sun being 93
million miles.)
4. How long does it take to come from the moon 1
5. How long would it take to travel a distance equal to
seven times the circumference of the earth 1
6. If it takes 5 years to arrive from a star, how far off is
that star 1
7. If a locomotive could be run 60 miles an hour day and
night, how long would it take to go round the earth ?
8. How long to reach the sun? etc.*
Answers should be given in weeks or years or whatever
unit is appropriate and most suggestive. This is a good rule
always, and is the real use of units to which people are accus-
tomed. Conversion of miles into inches is tedious and use-
less : but stating a big result in miles, a small result in
inches, and a moderate result in feet or yards, is right and
illuminating.
■'^Answers :
1. 2 miles.
2. About 3 miles.
3. 93 million miles -f 8 minutes
93000
— ^— thousand miles per second
8 X bO
-j^ = nearly 194 thousand miles per second.
4. About a second and a half.
5. About 1 second.
6. 5 X 365^ X 86400 x 194,000 miles.
7. Nearly 17 days.
8. About 180 years.
72 EASY MATHEMATICS. [chap.
9. If a pistol shot fired in a valley, at a spot which is
distant from the summit of a mountain by an amount which is
represented by a length of 4 inches on an ordnance map of
scale 1 inch to the mile, is heard on that mountain top 25
seconds after the flash, how high is the mountain above the
valley 1 (Ans. : 3 miles.)
This is perhaps hard : it can be done by drawing and
measuring, after it has been perceived that the sound has
travelled 5 miles in a straight line.
10. If a motor car is travelling 21 miles an hour, how long
will it take to go 100 yards 1
Ans. : 9*74 or 9 J seconds.
11. If the estimate of time were f second out, what error
would be made in reckoning the speed from the measured
distance 1
Ans. : I sec. is j\th of -^^ sec, so the error in estimate of
speed would be about
f 1^ miles an hour, or about 7^ per cent.
12. If a volunteer corps of 84 members shoots 160 rounds a
day each for 5 weeks, and if each bullet weighs | of an ounce,
what weight of lead will they have expended 1
13. If each bullet needed one halfpennyworth of powder to
propel it, and if lead cost 17/- a cwt., what would be the cost,
in powder and shot, for a regiment of 12 such corps, in the
course of 5 weeks ?
14. If an iron rod expands J per cent, of its length when
warmed 200 degrees, what allowance must be made for the
expansion of a bridge girder, ^ mile long, between a winter
temperature of - 40° and a summer temperature of 110°?
15. With the above data how much will an iron rod a foot
long expand if warmed one degree 'i
16. If a snail crawl half an inch each minute, how far will
it go in 3 hours 'i
vir.] PROBLEMS. 73
17. If sound goes a mile in 5 seconds, how long would it
take to go a foot ^
18. If sound reverberated between two walls 10 feet apart,
how many excursions to and fro will it make per second 1
19. If light takes 8 minutes to travel 93 million miles, how
long would it take to go one yard"? How many kilometres
would it travel per second? How many centimetres per
second 1 *
* Answers to the above :
12. 84 X 160 X 35 X f ounces = about 10 tons.
13. In shot, about £170 ; since a shilling per cwt. is a pound per ton.
In powder, 84 x 160 x 35 x ^ pence = 7 x 80 x 35 shillings = 28 x 35
pounds per corps.
14. The range of temperature is 150°; for this range iron expands |
of J per cent, of its length ; that is,
»"ile = t^-^ feet = ^^ = 1 '2375 feet, or nearly 15 inches.
Id X oUU ovv
—-^th of i per cent, of its length ; which is so.ooo^^ ®^ ^ ^°°*'
or "0000125 expansion per unit length per degree ; which is about the
riglit value for iron.
17. 1 second -7- 1056, or about the thousandth of a second.
18. In each excursion to and fro it will have to travel 20 feet ; but it
can travel 1056 feet in a second, therefore it has time to make 52 "8
excursions per second. If the walls were only 2 feet apart instead of
ten, the rate of reverberation would be 5 times as rapid, and would
correspond to the note /k i — This therefore is the musical note
heard if a short sharp noise, like a blow or clap, be made between two
walls two feet apart.
19. To travel 1 mile, light would take 8 minutes -r 93 million ; there-
fore to travel 1 yard it would take l/1760th part of this.
^"'^ '• ^^ = sTxTTo = Ml "'""°"*'^^ "' " ^^'=°"'*-
Light travels 300,000 kilometres per second or 3 x 10^" centimetres
per second ; as nearly as experiment at present enables us to eay.
74 EASY MATHEMATICS. [chaf.
20. If you buy a large number of oranges at three a penny,
and an equal number at two a penny, and then sell them all at
five for twopence, how much have j^ou lost on the transaction 1
(Ans. : a penny for every 5 dozen sold.
The buying price per couple is Jd. + ^d. ;
2
the selling price per couple is -yd.
So the loss per couple is^ + l^-^ = |-| = wu^-)
There are many ways of doing this problem, and it should
not be left till it is fully realised. Other problems depend
on the same principle, which is an important one. For
instance :
21. An oarsman rows a boat a certain distance up a river and
back, and then across the river, or on a lake, the same distance
and back. Which will be the quickest to and fro journey 1
22. If a steamer travels down a river at a rate of 19 miles
per hour, and up the same river with the same engine-exertion
at 7 miles an hour, what is the speed of the river 1 How long
would the steamer take to go a journey of 65 miles and back 1
(Ans. : The speed of the boat in stagnant water is the
half-sum, viz. 13, the speed of the river is the the half-
difference, viz. 6 miles per hour. The journey of 130 miles
would take ten hours in stagnant water, but up and down
the river it will take nearly thirteen hours.)
The general principle is that whereas {l+x) + (\ -x) = 2,
+ - does not equal 2
1 +a: 1 -«
9
but does equal — ^, which is greater than 2 ; though not
much greater when x is small. This applies to (20) (21) and (22).
23. If a couple of travellers sharing expenses are found to
be out of pocket in the course of the day. A, £2, 4s. 6d., and
VII.] PROBLEMS. 75
B, £1. 3s. 4d., what sum must be transferred from one to the
other to equalise matters %
(Ans. : Half the difference, viz. 10s. 7d. ; and the cost to
each has been half the sum, viz. .£1. 13s. lid.)
24. If three travellers on a tour have expended when
they return
^ £17 . 4 . 6
B £4.3.2
C £7 . 5 . 4
how can they best arrange to share expenses equally 1
{Ans. Find the mean expenditure by adding the items
together and dividing by 3 ; and then take the difference
between this mean and the expenditure of each. B and C
will then have to pay their respective differences to A. Their
two deficiencies from the mean, added together, should equal
^'s excess expenditure over the mean ; if this is not the case
a mistake has been made.)
The same rule would apply to any number of travellers.
Observe how it works for the couple of last question.
These exercises do not contain examples of so many quarts,
pecks, pennyweights, and drams. Such sums have no business
to occur. If artificial complexities of that sort are set, any
way of dealing with them will do : the simplest way is the
best way.
If a pupil is constrained to bethink himself of how the
teacher intends him to do a sum, it destroys originality. His
effort should always be devoted to find the best and simplest
way. This a teacher can help him to find, but a self-found
way is more wholesome in many respects than a coerced way,
even though the latter is neater. Originality should always
be respected : it is rather rare. Perhaps docility is made too
much of, and budding shoots of originality are frozen.
76 EASY MATHEMATICS. [chap.
Binary scale.
Although the natural method of dealing with multiples of
a unit is to employ the same system of notation as is in vogue
in arithmetic, and although therefore it is natural to specify
large numbers of things by powers of ten, there is a natural
tendency also to deal with fractions on a different basis, viz.
to proceed by powers of J. We see this on a foot rule, where
the inches are first halved, then quartered, then divided into
eighths, then into sixteenths, and sometimes even into thirty-
second parts of an inch.
The same method of dealing with fractions is found in
prices, as for instance of cotton, or any commodity which
requires a penny to be subdivided. Below the halfpenny and
the farthing we find the eighth, sixteenth, thirty-second, and
sixty-fourth of a penny in use for quotations; and these
ungainly figures are, or used to be, even telegraphed and
automatically printed on tape. So also a carpenter will
understand a specification in sixteenths of an inch, while a
decimal subdivision would puzzle him.
A thousandth of an inch is sometimes used however in
fine metal fitting work, and the thickness of a rod wanted
may be specified to a fitter as the thousandth of an inch
greater than 2j^^ inch.
These peculiarities are insular and not to be encouraged,
having originated in laziness and ignorance ; but they are
not nearly so bad as the weights and measures which people
who ought to know better still require that children shall
be taught.
It is quite possible to word arithmetic itself on the binary
scale, counting in pairs only; thus 10 (read one nought) may
be understood to mean 1 pair; 100 may mean 1 pair of pairs
(or 4), 1000 on the same plan will mean 2x2x2 or eight,
and so on. And on this scale *1 would mean J, -01 a quarter,
VII.] COINAGE. 77
•001 an eighth; so that one and a quarter plus an eighth
would be written 1*011.
The natural tendency to this kind of subdivision is apparent
in coins, even in countries with a decimal currency. For
instance in America you find the half and the quarter dollar,
beside the dime and the cent. In France you find the double
franc, franc, half-franc, and quarter-franc. So in Germany we
have as a drink-measure the halh-Uter and viertel-liter. And
in England we have half-sovereigns, half-crowns, also three-
penny bits, sixpences, shillings, florins, and double florins,
each double the preceding; the double florin being roughly
equivalent to a dollar or to a five-franc piece.
So also the commonest gold piece in France is the Napoleon
or 20-franc piece; not the ten-franc, or the hundred-franc
piece, though they both exist.
This natural tendency is the chief difficulty in introducing
a purely decimal coinage ; another is the convenience of the
penny and the shilling. If a decimal system is to be intro-
duced, one or other of these coins must give way. If the
Bhilling gives way, we can have an approximation to the franc,
and much inconvenience or grumbling in connexion with cab
fares, etc. If the penny gives way, and is made the tenth
of a shilling, we approach closely to the German system ;
and many commodities used by poor people will automatically
rise in price.
In Austria an attempt is being made to replace the gulden
and kreutzer by their respective halves, called krone and heller,
which correspond approximately with the franc and centime ;
but the older denominations persist, and it is quite likely that
the two will co-exist and be convenient.
It may be asked "why mention these things in a book of
this kind " 1 And the answer is because children can take an
intelligent interest in them, and because it is instructive for
78 EASY MATHEMATICS. [chap.
them to realise that our present coinage is not a heaven-sent
institution, but is susceptible of change, — change too in which,
when adult, some of them can take their part, either in pro-
moting or opposing. There is therefore a reality about these
things, and arithmetical ideas can inculcate themselves in
connexion with them without labour.
Decimal system of weights and measures.
Although the present division of money is so deep-rooted
that decimal coinage is difficult of introduction, and although
the decimal system in arithmetic is not the best that could
have been devised ; yet its advantages over most other systems
are so enormous that in connexion with weights and measures
it undoubtedly ought speedily to be introduced.
The first and easiest place to introduce it is in connexion
with weights. No one really wants to reckon in ounces and
pennyweights and grains and scruples and drachms. Ounces
used to be perpetuated and popularised by the Post Office
regulations; but now that a quarter of a pound will go
for a penny, and, under certain restrictions, an eighth of a
pound for a halfpenny, the necessity for ounces has really
disappeared. It would be quite easy to make the halfpenny
postal regulations refer to a tenth of a pound instead of
an eighth, and to construct ten-pound weights, hundred-
pound weights, and their convenient doubles and halves and
quarters.
There is however this fundamental question to be considered :
shall the British pound be adhered to, or shall we adopt the
unit of our neighbours and employ the kilo (short for kilo-
gramme) or the demi-kilo 1
The kilo is too big for many ordinary purposes. In France
small marketing is still done by the demi-kilo, because it
represents a reasonable and commonly-needed amount of stufE
VII,]
MEASURES. 79
It is altogether handier than the kilo. A demi-kilo might be
introduced, and with us might still be called a pound, or, for
a time, an " imperial pound," though its value would have to
be increased by ten per cent, above our present pound. The
kilo is approximate!}^ 2-2 lbs., so the new pound or demi-kilo
would be one and a tenth old pounds. The gramme would be
•002 new pound.
The disadvantages of any change are obvious. The advan-
tage would be that we should then be using practically the
same unit as our neighbours.
All other denominations could be swept away ; except, for
occasional rough use, the ounce and the ton, which continue
useful ; for the ton would be 2000 of the new pounds, and
would correspond exactly with the French tonne ; and the
ounce, slightly changed, would be yY of the new pound, or it
might be changed so as to be one-tenth of it. The grain or
TWO" P^^ ^^ ^^® ^^^ pound might easily give place to a new
grain yxjj^jtr P^^* ^^ *^® ^^^ <^^®-
These handy names are useful for common purposes and for
speech. All accurate specifications should be made in terms
of the pound, and of that alone. Thus 1*4903 lbs. would be
a specification accurate to the nearest grain of a weighing of
something like a pound and a half.
3 '01 4 tons would be a statement, intended to be accurate to
the nearest pound, of the weighing of a 3-ton mass.
Let me emphasise what may be regarded as one of the
special advantages of this simple and easily introduced change.
Children could then be practised in weighing at once : to the
vast advantage of their education. At present an apothecary's
scales are an abomination, and no child can weigh satisfactorily
with the weights of a letter balance, which are all in the binary
scale ; though, as aforesaid, these serve as an introduction to
ideas of weighing, etc, in quite early stages. Letter weights go
80 EASY MATHEMATICS. [chap.
down too rapidly ; there are not enough subdivisions ; and the
result cannot easily and quickly be specified, except as an
awkward series of vulgar fractions, or else in the binary scale
of arithmetical notation.
The only way in which school weighings can be satisfactorily
done now is by the use of grammes and kilogrammes : and
there is a foreign feel about these things; which those who
learn chemistry indeed get over, but which gives it a flavour
distinct from ordinary life.
What we want is that children shall weigh and measure all
sorts of things, and do a large part of their arithmetic in
terms of their own weighings and measurings : thus making
it real and concrete and if possible interesting.
Weighings of plants and of growing seeds, of rusting iron
and of burning candles, of dissolving salts and of evaporating
liquids, can all be made interesting and instructive.
Weighings in air and water, and finding thereby the specific
gravity or the volume of irregular solids, can easily be over-
done and made tedious, but, short of this, such operations are
quite instructive.
G-auging and measuring of regular solids is an equally in-
structive way of arriving at their specific gravity, or, as it may
be more scientifically called, " density." The approximate
relative densities of such things as stone, lead, iron, gold,
copper, platinum, cork, air, referred to water, are worth
remembering : stone say 2*5, lead 11, iron 7, gold 19, copper 8,
platinum 21, cork ^, air -g^.
Decimal measures. — Continued.
The introduction into commerce of "the decimal system"
is a more difficult matter however. The admirable duo-
decimal division of the foot into inches (like that of the
shilling into pence) stands in the way. The foot and the
VII.] MEASURES. 81
inch and the yard seem ingrained in the British character,
and will give place to the metre and the centimetre only
with difficulty.
The fact is that the introducers of the " metre" made a great
mistake by not adopting the yard or the foot or some other
*existing unit as its value : they would also have been wise
if they had adopted the pound as their kilogramme, and left
the dimensions of the earth alone. It is the magnitude of
the human body which really and scientifically specifies and
confers any meaning on absolute size : our bodily dimensions
and time relations must be the basis of all our measures and
ideas of absolute magnitude. To abandon the human body
and to attend to the dimensions of the earth was essentially
unscientific or un philosophical : it has all the marks of faddism
and self-opinionatedness. However these unwisdoms of sections
of the human race we have to put up with, and at any rate
the French evolved a better system on the whole than that
which had come down to us by inheritance and tradition from
uncivilised times.
If we were at liberty to adopt the foot as our standard, and
to call its decimal subdivisions inches, or if a new foot were
made ten inches long, the change would not be so very
difficult. If it had been extensively customary to divide the
inch too into twelfths (called lines) the change would be
harder ; but divisions of the inch in the binary scale have been
customary, and these are not really convenient : a decimal
system is better than that ; and foot rules decimally divided
and subdivided could easily be supplied and used.
But then, as in the case of our present pound, we should be
using an insular measure different from all the rest of Europe,
and amid the stress of industrial and engineering competition
this is a serious handicap.
A metre scale is a rather unwieldy thing : a half-metre
82 EASY MATHEMATICS. [chap.
scale is handier for many purposes, and might be made like
a folding two-foot rule.
There is no help for it : we must get used to metres and
centimetres, and the sooner we begin the better.
Angles and Time. •
There are two things which have not yet been subdivided
decimally with any considerable consensus of agreement : they
are Angles and Time.
The division of the right angle into 90 equal parts is
convenient. The subdivision of the degree into sixtieths
and again into sixtieths (called respectively partes minutas and
paries minutae secundae, now abbreviated into " minutes " and
"seconds") is peculiar and sometimes troublesome but not
exactly inconvenient, though a decimal subdivision of the
degree would be simpler.
As to time, the fundamental unit is the day or period of
the earth's rotation (this being the most uniformly moving
thing we know). Its subdivisions (into 24 parts, and then
into sixtieths, etc.) are curious, but too deep-rooted for
anyone to attempt to alter; and fortunately they are the
same in all countries.* Legitimate practice in dealing with
different denominations can therefore be afforded to children
by our large admixture of universally understood measures of
time; including weeks, months of different kinds, years of
different kinds, and centuries. All other weight and measure
complications, especially those of a merely insular and boorish
character., should forthwith cease to be instilled into children.
Further exercises.
It is worth noticing and remembering that a kilometre = 10^
centimetres.
■^A third subdivision, the sixtieth part of a secoud, is sometimes
known as a " trice."
yii.] MEASURES. 83
It is also ten minutes' walk, or very roughly two-thirds of
a mile.
A cubic metre is a million cubic centimetres.
A cubic kilometre is a trillion cubic millimetres ; meaning
by "trillion" a million million million, after the English
custom. (But the French use the term " billion " to signify a
thousand million ; and a million million they accordingly call
a trillion ; while the above number would by them be desig-
nated a quintillion : in any case it is 1 followed by eighteen
ciphers).
A cubic centimetre is 1000 cubic millimetres, and is xoVo ^^
a litre.
A gallon of cold water weighs 10 lbs., by definition of a
gallon ; a pint therefore weighs a pound and a quarter.
A cubic metre of water is a tonne, and very approximately,
though accidentally, equals an English ton also.
A cubic centimetre of water, at its temperature of maximum
density, weighs a gramme exactly, from the definition of a
gramme.
The speed of an express train, 60 miles an hour, is only 15
times a walking pace.
The speed of a bullet, say 1800 feet a second, is twenty
times that of a train.
The speed of sound is comparable with that of bullets.
The speed of light is a million times the speed of sound in
air.
Four miles an hour is 2 yards a second, approximately,
or accurately 60 miles an hour is 88 feet a second.
It is an instructive exercise to let a boy find out the
sizes and distances of the planets of the solar system, and cal-
culate a numerical model illustrating them on any convenient
scale.
84 EASY MATHEMATICS. [chap.
I have myself found a local topographical scale the most
convenient : one on which the earth was about the size of a
football, and the sun the size of some public building a mile
or two distant. The other planets distribute themselves
about the town and county; some of them extending into
more distant counties.
It is instructive to try to place the nearest fixed star in
such a scale, and to find that it will not come on to the earth
at all.
The price of a railway ticket to the nearest fixed star, at
Id. per hundred miles, can also be calculated; and found to
approach or exceed the National debt.
The earth takes a year to go round the sun in a circle of
93 million miles radius : how fast does it go ?
(Ans. : About 19 miles a second.)
Light goes 10000 times as fast as this.
How fast would a train have to run on the equator if it
were to keep up with the apparent motion of the sun, so that
it should continue the same time of day 1
(Ans. : About 1000 miles an hour.)
How far from the North Pole could the same thing be
accomplished by a man walking 4 miles an hour ?
(Ans. : About 30 miles away.)
If a man walked 30 miles South from the North Pole, and
then walked 40 miles due West, how far, and in what direc-
tion, would he have to go to get back to the Pole 1
(Ans. : 30 miles due North.)
What is the density of a rectangular block whose height is
5 inches, length 11 inches, breadth 8 inches, and weight
82 J lbs. ^
(Ans. : 3 ounces per cubic inch.)
VII.] EXERCISES. 85
Directly the elements of mechanics and of heat and of
chemistry have been begun, any number of useful and fairly
interesting examples can be constructed. They afford
practice in arithmetic of the best and most useful kind;
quick and ingenious computation being what is wanted,
not laborious dwelling upon long artificial sums. Long sums
are never done in adult practice : there are always grown-up
methods of avoiding them.
It is cruel to subject children to any such disciplinary
process, as part of what might be their happy and stimulating
education. Before they have been subjected to it they are
often eager to have lessons; but experience of the average
lesson, as often administered, soon kills off any enthusiasm,
and instils the fatal habits of listlessness and inattention
which check the sap of intellectual growth for a long time.
If the teacher of arithmetic knows arithmetic and nothing
else, he is not fit to teach it. His mind should be alive with
concrete and living examples ; and it is well to utilise actual
measurings, weighings, surveyings, laboratory-experiments,
and the like, to furnish other opportunities for arithmetical
exercises.
Arithmetical exercise can be obtained unconsciously, as
bodily exercise is obtained by playing an outdoor game.
The mechanical drill or constitutional-walk form of exercise
has its place doubtless, but its place among children is limited.
There used to be too much of it, and too little spontaneity
of bodily exercise, in girls' schools. Now the spontaneity and
freshness is permitted to the body, but too often denied to
the mind.
The same kind of reform is called for in both cases. The
object of this book is to assist in hastening this vital reform.
CHAPTEE VIII.
Simple proportion.
Any number of sums are of the following character :
If 3 sheep cost £20, what will 100 cost ?
Now the so-called "rule of three" method of dealing with
sums of this kind, though permissible, is not really a good
method, because it leads to nothing beyond, and employs an
antiquated system of notation.
The answer is one hundred thirds of twenty pounds
= ioo x£20 = ^%^-^ = X666-6 = £666f = .£666. 13s. 4d.
If the answer is not obvious, it can be arrived at by the
intermediate step of considering one sheep, which will cost
the third of .£20, namely, £6. 13s. 4d.*
And so a hundred sheep will cost 600 pounds, 1300
shillings, and 400 pence.
The 1300 shillings reduce to 65 pounds, since 100 shillings
is five pounds; and the 400 pence make £1. 13s. 4d., since
240 pence is a pound, and so 400 pence is thirty shillings
and 40 pence (or 3s. 4d.) over.
This is not an orthodox way of doing the sum, but it is
just as good as any other, and it is one that a boy might
* [It would not come out even so well as this but for the fortunate
duodecimal division of the shilling into pence, so that one-third of a
pound, viz. 6s. 8d., and two-thirds, viz. 13s. 4d., can be exactly
specified without fractions. These amounts are worth remembering as
one-third and two-thirds of a sovereign.]
CHAP, vin.] SIMPLE PROPORTION. 87
scheme for himself. There would be no need to snub him
for it. Everything which is troublesome about such a sum
results from the miserable property of the number ten, that
it is not divisible by 3.
If we had set the following very similar question :
If 3 sheep cost £24, what would 100 cost ?
An infant could answer £800, doing it in its head. But it
would clearly do it by the same process, viz. the process of
considering the price per single sheep, and that is therefore
the natural and simplest method.
To summarise : The childish method is the i^iethod by
units, and may be written out at length ; the adult method is
the method by ratio ; what place is there for the rule of
three 1 The rule of three with its symbols : : : :
is reserved for antiquated school instruction.
Observe, there is no harm in writing a ratio as 2 : 3 or a : 6,
and occasionally it may be convenient to do so, though 2 -r 3,
9
or a -r 6 is precisely the same thing, and usually the form ^ or
-, or a/b, is in every way better. So the symbol : : is needless,
because replaced by =. The fact is that : connotes the
theoretical idea of ratio, while -f indicates the practical
operation of division, which is the actual means of working
a ratio out. The vulgar-fraction form may be used instead
of either of these signs and is usually best. The division
may or may not be actually performed, as we please.
I feel inclined to illustrate good and bad methods at this
stage a little further, by taking a few more very simple
examples. For instance :
If twenty dogs pulling equally at a sledge exert a hori-
zontal force of 1 cwt., what force do any three of them
exert?
88 EASY MATHEMATICS. [chap.
Adult method :
J.ths of 1 cwt. = ^
!x 112 lbs.
20
20
Good cMldish method :
20 dogs pull
112 lbs.
10 dogs pull
56 „
1 dog pulls
5-6 „
3 dogs pull
3x5-6 = 1
= 16-
16-8 lbs.
If it be asked why not stop at g^^ths of a cwt. and give the
answer as '15 cwt., I reply, no reason against it at all; but
children should be accustomed to realise forces and other
things, in p^jtual homely units that they can feel and appre-
ciate ; and a cwt. is too big for them.
Mechanical method :
20 : 3 :: 112 : the answer.
Rule. Multiply the means and divide by one extreme and you get
the other extreme.
.'. the answer is, etc.
British Method:
There is indeed a barbarous way of complicating the sum, which
is typical of much that goes on in these islands at inferior schools :
lbs. oz. drachms
20 1 336. 0. 0
16 . 12 . 12|
which is done thus :
Twenty into 336 goes 16 and 16 over, that is 16 lbs. over, which
equals 256 ounces. Twenty into this goes 12 times and 16 over, that
is 16 ounces or 256 drachms ; into which twenty again goes 12 times
and Jf ths over, which last equals f ths, that is fths of a drachm.
So the answer is 16 lbs. 12 oz. 12f drachms.
On this one has to remark that since the unfortunate ^ has
to appear (as it happens) sooner or later, why should it not
appear at firsts Why is |ths of a drachm easier to understand
than Iths of a pound? The fact is that it is not easier to
understand, and by children is not understood : the " 4 over "
which remains at the end is a continual puzzle to them.
VIII.] SIMPLE PROPORTION. 89
They have been so accustomed to getting rid of fractions by
reducing to a lower denomination, that at the end, when
lower denominations unaccountably fail them, they are non-
plussed. Quite rightly so ; the fault is not with the children.
Whenever an attentive child finds a persistent difficulty,
teachers should be sure that there is something wrong with
their mode of presenting it, probably with their own compre-
hension of it. Nothing is difficult when properly put. The
whole art of teaching should be so to lead on that everything
arrives naturally and easily and happily, like fruit and flowers
out of seeds.
Another British method. Usually however the sum is not recorded
80 briefly as this, but is written out in what is known as the long-
division plan ; and it is perhaps the safest mode of getting the right
answer if the answer is required to be thus barbarously specified, for
it certainly shirks nothing. This is the way of it :
To divide 336 lbs. av. into 20 equal parts
lbs. oz. dr.
2,0)33,6(16 . 12 . 12|
20
136
120
16
16*
96
256 oz.
240
16
_16*
96
16
256 dr.
240
16 remainder, and ^^ = -| dr.
* If any mathematician glances through this book, as I hope he may,
he will require at these stages to be reminded if British, to be informed
90 EASY MATHEMATICS. [chap.
This may look like a parody, but it is soberly the way in
which innumerable children have been taught in the past to
do such a sum. And the fact that they have been so taught
can easily be tested by setting it to people who were children
a few years ago.
Another method. If the factor plan of division is adopted
there is great danger of confusion and error about the carrying
figure. For instance, in dividing 336 lbs. into 20 equal parts,
a child as sometimes now" taught will proceed thus :
2 [336 lbs.
1 0 1 16,8
16 and 8 over.
8 what over ? They are apt to take it as 8 lbs. over, and so
interpret it as 128 ounces, and proceed to divide these again
by 20 by the same process
2 1 128
10 I 64
6 and 4 over
apt to be called 4 ounces over, which are interpreted as 64
drachms, and so on.
if Foreign, that in these islands a drachm is defined to be the sixteenth
of an ounce, and that an ounce avoirdupois is one sixteenth of an
avoirdupois pound ; moreover that a drachm is the lowest recognised
denomination of avoirdupois weight : after that fractions are permitted.
Pennyweights and grains belong to a system of measures to which the
name of " Troy " is (for some to me unknown reason, perhaps from
Troyes in France) prefixed. There is a "Troy pound" and a "Troy
ounce," for "metallurgical" use, but they differ from their "grocery"
cousins which are explicitly asserted "to have some weight." Then
between grains and Troy ounces there are other denominations used by
"apothecaries," called scruples and drams. This dram is not the same
as the grocery drachm. There appears however to be only one kind of
"grain," and 7000 of these make 1 lb. avoirdupois, while 5760 of them
make 1 lb. Troy.
VIII.] SIMPLE PROPORTION. 91
This is quite wrong. The 8 over in the first little sum was
really 8 double-pounds, and so the second little sum is all
wrong. If it had been right, the 4 over could not have been
4 ounces, but 4 double-ounces ; but what needless trouble and
risk of error is introduced by having to perceive this !
Again let many children be asked to divide £336 by 25,
they will few of them have been taught to proceed thus :
%Y- = 3-36x4 = £13-44
= £13. 8-8s.
= £13. 8s. 9-6d. or about 9Jd.
but they will proceed, either by long division on much the
same lines as in the last example, which is long to write, or
else by short division, dividing by 5 twice over, which is not
too long to write,
£ s. d.
5 I 336 .0.0
5 1 67 . 4 . 0
13 . 8 . 9|
short to write, but rather hard to do. Such trivial sums
should not call for so much brain power as is involved in
various and complicated carryings.
Money sums however are the best examples of the kind.
If it was 336 tons that had to be divided into 25 equal parts,
grown people would be satisfied to say that each part must be
13-44 tons; but at some schools it would have to be done
thus, — if not by a still longer process equally liable to acci-
dental error :
tons. cwt. qrs. lbs. ozs. dr.
5 I 336 . 0 . 0 . 0 . 0 . 0
^ 167.4.0.0.0.0
13 . 8 . 3 . 5 . 9 . 9j Ans.
92 EASY MATHEMATICS. [chap.
Breakdown of simple proportion or "rule of three."
Simple proportion, or the rule of three, is by some teachers
regarded as a kind of fetish ; moreover its extreme simplicity
makes it 9- rather favourite rule with children and they will
naturally do many exercises in it. Not always, it is to be
hoped, by the same mechanical method.
But there is all the more necessity for bringing home to
them the fact (strange if it is unknown to any teacher), that
it does not always work. For instance :
A stone dropped down an empty well 16 feet deep reaches
the bottom in one second. How deep will a well be if a
stone takes two seconds to reach the bottom 1
The answer expected is of course 32 feet; but it is not
right. The correct answer is 64 feet.
If a stone drops 16 feet in one second, how far will it drop
in J second ] (Ans. : 1 2 inches.)
Again, if a stone dropped over a cliff descends 64 feet in
2 seconds, how far will it drop in the next second 1
(Ans. : 80 feet.)
A steamer is propelled at the rate of 8 knots by its engines
exerting themselves at the rate of 1000 horse power. What
power would drive it at 12 knots?
Probably no one would expect the answer 1500 to this ; for
on that principle 10000 horse power would propel it at
80 knots.
An initial velocity of 1600 feet a second will carry a rifle
bullet 3 miles. What velocity would carry it 6 miles 1
An ounce weight drops 4 feet in half a second. How far
will a pound weight drop in the same time ?
(Ans. : By experiment, 4 feet likewise. A most important
fact, discovered by Galileo, and illustrated from the tower of
Pisa.)
VIII.] NON-SIMPLE PROPORTION. 93
Let it not be dogmatised on, but illustrated by dropping
things together ; and if it appears puzzling, so much the better.
Ignoring or eliminating the resistance of the air everything
falls at the same pace. The air has very slight influence
on the drop of smooth spheres through a moderate height.
Cotton wool and feathers and bits of paper will drop more
slowly, but the reason is obvious : a bullet will drop more
slowly in treacle than in air. That is because the air resistance
is small : it is not zero, and if a bullet and a pea were dropped
from too great a height, air friction would begin perceptibly
to retard the lighter body. So it is that big rain-drops fall
quicker than little ones ; and these small drops quicker than
mist and cloud globules. So also does heavy fine powder,
even gold powder, fall slowly in water, not because it is
buoyed up, but because it is resisted. Eemove the air, and in
a vacuum a coin and a feather will fall at the same rate. The
statement does not explain the fact. The full explanation of
the fact is not even yet known. But a very great deal more
is known about the whole subject than is or can be here
expressed. That is characteristic of elementary books through-
out, and the object of the learners should be to get through
all this easy stuff, and get on into more exciting matters
beyond : matters which the majority of the human race never
have the least knowledge of, because their early education
has been neglected.
A balloon 18 feet in diameter can carry a load equal to one
man. What load can a similar balloon carry which is 36 feet
in diameter. (Simplest rough answer, 8 men.)
A rope stretches half an inch when loaded with an extra
hundredweight.
How much would it stretch if loaded with an extra ton 1
94 EASY MATHEMATICS. [chap. viii.
A half crown is ten times the value of a threepenny bit.
How many threepenny bits can lie flat on a half-crown
without overlapping the edge*? (Ans. : By experiment, one.)
A boy slides 20 yards with an initial run of 10 feet. What
initial run would enable him to slide half a mile ?
If 2 peacocks can waken one man, how many can waken
six?
If a diamond is worth ten thousand pounds, what would
950 similar diamonds be worth ?
If a camel can stand a load of 5 cwt, for 6 hours, for how
long could he stand a load of ten tons ?
These things cannot be done by simple proportion. They
require something more to be known before they can be done
at all ; and accordingly it would appear as if generations of
teachers had discreetly shied at them all, indiscriminately,
and had excluded them from arithmetical consideration
altogether. It is just as if in geometry, finding straight lines
simpler than curves, they had agreed to found all their
examples upon straight lines.
CHAPTER IX.
Simplification of fractions.
Vulgar fractions are much harder to deal with than
decimals ; but as sometimes several have to be added together
it is desirable to know how to do it. Besides, the exercise so
afforded is of a right and wholesome kind.
Consider the following addition : J + J. Small children can
see (by experiment on an apple) that the result is |, and they
can also be taught to regard it as f + ^ = f^, which should be
read in words — two quarters added to one quarter make
three quarters.
Thus, it can be realised that when the denominators are all
the same, addition of fractions becomes simple addition of the
numerators.
For just as 5 oranges + 6 oranges =11 oranges, so
5 I 6 _ 11
TT'^TT ~ TT'
reading " seventeenths " instead of " oranges."
When denominations differ, therefore, the first thing to do
is to make them the same.
Thus, for instance, 3 apples + 4 oranges, is an addition which
can only be performed by finding some denomination which
includes both, say "pieces of fruit."
So also 7 horses + 3 pigs =10 quadrupeds. 5 copies
of Bobinson Crusoe + 3 copies of Ivanhoe = 8 prize-books,
perhaps.
96 EASY MATHEMATICS. [chap.
Eeduction to the same denomination cannot always be
done, when denominations are anything whatever, except by
using the vague term "objects" or "things"; but with
numerical denominators it can always be done, and the method
of doing it has to be learnt, y + tt = xi' ^^^ ^^^h like, are
easy examples. |- + y 2 = ^ is a slightly harder one.
It is done by saying -^^ + iV = A = J-
So also ^ + i = I, being equal to ^.
A harder example is | 4- 1, which can be written
21,10 _ 3 1 _ Kl
In the decimal notation this would appear thus :
3-5 + 1-666... = 5-1666....
A still harder example can be worked out thus :
9i5 _ 6. -J, 40 _ 103 _ 147
8"+'T — S'F"'" "56 "5 6" -^ -^TF'
though the final step is one that need not always be made.
Now it is evident, or at least it will gradually be found true,
that in a mechanical process of this kind there is always some
simple rule by which the result can be obtained without thought.
What is that rule ? If the child can find it out for himself, by
experimenting on lots of pairs of fractions, so much the better.
A week is none too much to give him to try, for if he finds it
out himself he will never forget it.
The rule is : cross-multiply for the numerators, and multiply
the denominators.
1 1 ^6 + 2 _ 8 2
2'^6 ~ 12 " 12~3*
1 !_& + «
a b ~ ah '
3 4 _ 27 + 28 _ 55
T'^g" 63 ~"63'
a c _ad+hc
h^d~'~hdr'
but it would be a pity to spoil this by premature telling.
IX.] FRACTIONS. 97
The fact that the sum of two reciprocals is the sum of
the numbers divided by their product, is worth illustrating
fully and remembering : remembering, that is, by .growing
thoroughly accustomed to it, not exactly learning by heart.
There is hardly any need to learn easy things like that by
heart: nevertheless it is a very permissible operation, whenever
the fact to be learnt is really worth knowing.
11 7 ., , . ., sum
0 + T = Ts. that IS the 3 — -•
3 4 12 product
1 1 _a + h
a h~ ah '
}_l_ 28
23 5 115*
2 + 49 "98 — 100- ^''•
the symbol aCh meaning " approximately equals."
[The approximation is seen to be true because adding 1 to
50 makes the same proportional difference as adding 2 to 100.
If this is too hard, it can be postponed. It is unimportant,
but represents a kind of thing which it is often handy to
do in practice.]
But this rule of cross-multiplication hardly serves for the
addition of three or more fractions, at least not without
modification. Take an example,
1^2 7_14 21_26i
6'''3"*'2~6"''6"^ 6 ~ 6 - ^*
Take another, 1 ^ 1 + 1 = i±|±l = 7,
where the three fractions |, f, and \^ all having the same
denominator, are written all together, with the addition of the
numerators indicated, and subsequently performed,
r. 111 3 + 4^ + 1 8j 17
One more, + + = — — = -^ = r^.
3 2 9 9 9 18 ^
This might hardly be considered a legitimate procedure, but
L.E.M. G
98 EASY MATHEMATICS. [chap.
there is nothing the matter with it. You might, instead,
proceed thus :
1 1 1 _ 18^27 A^^l^lZ
;V^2"^9 ~ 54 "^54 "^54 ~ 54 ~ 18'
and that is equally a correct method.
But neither of these plans is quite the grown up plan. Let
a better plan be found; but first let the above plans be
formulated and expressed. Is it not plain that the numerator
of each particular fraction is found by multiplying two of
the denominators together, while the common denominator of
all the fractions is found by multiplying all the denominators
together 1 Apply this rule :
1 1 1 _ 20 + 24 + 30 _ 7£_37
6"^5"^4~ 120 "120 "60*
For instance, a sixth of an hour + a fifth of an hour + a
quarter of an hour = 37 minutes, a minute being the sixtieth
of an hour. Now a sixth of an hour is ten minutes, a fifth is
12 minutes, and a quarter of an hour is 15 minutes: conse-
quently the neatest way of doing the sum would be
1 1 1 10 + 12 + 15 37
6"^5^4~ 60 60*
. ^, , 111 180 + 36 + 720
Another example, ^ + ^^ + 3 = -^f^^.-J-^
but here every term in numerator and denominator can be
divided by 3 and by 1 2, so that the above may be written
12^60 3 60 60 30
And it would have been neater to write it so at first — neater
but not essential, and sometimes not even the most rapid
plan.
To illustrate the above example :
i\th of a day is 2 hours.
^^0 th of a day is 24 miuutes.
^rd of a day is 8 hours.
IX.] FRACTIONS. 99
Consequently the sum of these fractions of a day is 10 hours and
24 minutes,
which is lOf^ of an hour [ = lOyV = 10'4 hours] or ^j + ^V of a day,
which again may be written iriJ + FT7 = FIT =^ ^iyths = •043, as before.
The form of the general rule, then, is given by
I 1 1 _ bc + ca + ab ^
a b G ahc '
but in practice it is possible to abbreviate this in some cases,
when one of the denominators contains the others as factors,
or when some simple relation of the kind exists between them.
This is what was made use of in the early simple cases,
such as y*2 + 2^ ; we did not proceed to write "" ^ and
then simplify it, but we wrote at once ^t + 2T ^ tt y ^^^^ i^
to say we perceived that 24 would do for the new denomin-
ator, and we adjusted the numerators accordingly.
Perhaps we had better display this algebraically. Let each
denominator contain a common factor, say r?, so that the
reciprocals to be added are h ^ H — , then if we applied the
^ na nb nc ^^
, , 111 -^ '^^^<^ + 1^'^<^<^ + i^^o^b ,
mere general rule we should write g-y , but the
repetition of the powers of n is manifestly needless, since they
cancel out ; and it is much neater to write for the new
denominator an expression which contains the common factor
, , , bc + ca-\-ab
n only once, thus : ^
The denominator so obtained is called the least common
multiple of the three denominators ; and it is frequently, in
examination papers, denoted by the letters L.C.M. It is not
an important idea at all. Sums can be done quite well
without it, but its introduction affords some scope for neat-
ness and ingenuity. Easy processes can be given for finding
100 EASY MATHEMATICS. [chap.
it, but they are hardly worth giving, as in real practice they
are seldom used : they are of most educational service if
employed as an exercise for the student's invention. They
will be dealt with sufficiently in the next chapter.
Now take a numerical example :
Add together | + J + f + t\ + ^2-
Here 32 is evidently the l.c.m. of the denominators, since it
contains all the others as factors. So that will serve as the simplest
common or combined denominator. The first numerator accordingly
will be 16, the second 8, the third 4 but taken 5 times and therefore
20, the next 2 taken 3 times, and the last 1 taken 7 times.
Consequently the sum is written as follows :
1 1 5 ^ J _ 16 + 8 + 20 + 6 + 7 _57
2'^4"^8"^16'*"32" 32 32*
Take another example of addition :
1 1 1 J^ _ 72 + 9 + 56 + 8 _ 145
7"^56'^9'''63~ 504 "504
Here 7 is plainly a factor of both the larger denominators, and 8 and
9 are the other factors, so the least common denominator will only
contain 7 ard 9 once, and will equal 7x8x9 = 504, and this being the
smallest common multiple possible, no further simplification can be
eflfected ; beyond of course expressing the result as a decimal if we so
choose. To express it as a decimal we must eflfect the division
indicated ; the result happens to equal '2877 almost exactly.
It is worth noticing that the series of powers of -J, viz. :
add up very nearly to 1 ; and the more nearly the more
terms of the series are taken.
It can be shown, not by trial indeed, but by simple reason-
ing, that if an infinite sequence of this series are added
together the result is exactly 1. Thus the first term con-
stitutes half of the whole quantity, say a loaf, the second
term added to it gives us three quarters, the third term gives
us |-th more, and we only need another eighth to get the
IX.] FRACTIONS. 101
whole. The next term gives us half of the deficiency, and
now we need the other sixteenth to make the whole. We
do not get it however: we get half of it in the next term,
and thus still fall short, but this time only by ^^ ; and so at
the end of the above series, as far as written, our deficiency
is 6T*^- Each term therefore itself indicates the outstanding
deficiency, and as the terms get rapidly smaller and smaller,
so does the deficiency below 1 get rapidly diminished till
it is imperceptible. (Compare p. 325.)
It is convenient to plot these fractions as lengths (setting
them up at equal distances along a horizontal line), say half a
foot, then a quarter, then an eighth, and so on. Then joining
their tops we get a curve which has the remarkable property
of always approaching a straight line, but never actually
meeting or coinciding with it, or at least not meeting it till
infinity ; when at length it has become quite straight.
There are many curves with such a property, but fig. 9 may
be the first a child has met. He can of course continue the
curve in the other direction — the direction of whole numbers,
or powers of two, and observe how rapidly it tilts upwards ;
but there is no straight line in this direction to which it tends
to approach ; this end proceeds to infinity both upwards and
sideways, not only upwards, though it proceeds far more
rapidly in the vertical direction than in the horizontal ; and
this end of it never becomes straight.
CHAPTER X.
Greatest Common Measure and Least Common
Multiple.
Another name of slight importance, which is usually paired
off with Least Common Multiple (page 99), is Greatest
Common Measure or Highest Common Factor : often denoted
by G.C.M. or by H.C.F.
The two numbers 24 and 16 have several factors common
to both of them, for instance 8 ; and this as it happens is the
greatest common factor, the others which they possess in
common being 4 and 2.
The numbers 20 and 35 have 5 as the largest factor common
to both of them. The numbers 72 and 84 have 12 ; while 72
and 96 have 24 as their g.C.m.
The numbers 23 and 38 have no factor, above unity,
common to both. In fact 23 has no factor at all.
The word "common" so used does not mean "ordinary,"
as children sometimes think, nor does it mean vulgar, but it
has the signification which it possesses in "common friend,"
or in vulgar phrase " mutual friend," or when people are said
to own property "in common."
To find common factors of two numbers, one way is to
arrange all the factors of each in two rows one under the
other and see how many correspond. Inspection will then
readily show which pair is the biggest.
CHAP. X.] GREATEST COMMON MEASURE. 103
Suppose the two numbers given were 40 and 60 ; the
following are the factors of 60,
2, 3, 4, 5, 6, 10, 12, 15, 20, 30,
and the following are the factors of 40,
2, 4, 5, 8, 10, 20.
Of these, the 2, 4, 5, 10, 20 are common to both, and 20 is
the largest of them.
In old-fashioned language, factors were called "measures,"
and the largest common factor was called the "greatest
common measure," and abbreviated into g.c.m.
What is the use of it 1 Very little ; but the meaning is
perfectly simple and should be understood. It can be utilised
for finding the Least Common Multiple of a set of numbers,
that is to say the smallest number which contains them all
as factors ; for the G.C.M. represents what may be struck out,
once at least and sometimes more than once, from the product
of a set of numbers, in order to leave behind the smallest
number which they are able to divide without a remainder.
Thus take the numbers 40 and 60, their product is 2400,
and of course they will both divide that ; but their G.C.M., 20,
may be cancelled out of it, leaving 1 20 ; and both 40 and 60
will divide that too. It is the least number which they can
both divide exactly, i.e. it is the least number of which they
are both factors, it is in fact their least common multiple.
Example.— Of the numbers 12, 20, 24 what is the g.c.m.
and L.C.M. ? Of these, 12 need not be attended to in finding
the largest common factor, because it is itself a factor of 24.
Of the numbers 20 and 24, 4 is a common factor ; so divide
all by that, and we get left with 3, 5, 6.
No factor will divide all these, so 4 was the g.c.m. of the
original numbers.
Their least common multiple is not 3 x 5 x 6 = 90, because
that would have omitted the factor 4 which they possess in
104 EASY MATHEMATICS. [chap.
common. The common factor need not be repeated more than
once, (for if it is, though you get a common multiple, you do
not get the least common multiple), but it must not be omitted
altogether, or you will not get a common multiple at all.
The L.C.M. accordingly is 3 x 5 x 6 x 4 = 360, and of that it
will be found that the given numbers 12, 20, 24, are factors.
The product of those numbers is 5760, and out of that the
G.C.M. 4 can be struck twice before arriving at the L.C.M.
Anyone therefore can invent a rule for finding the L.C.M.
of a set of numbers; it is, find their g.c.m. and divide or
cancel it out of all the numbers but one, then multiply the
quotients together.
But a rule for finding the g.c.m. is by no means so easy to
invent : it is an ingenious process, and the whole subject is
essentially a little bit of rudimentary pure mathematics ; it
has no practical importance or application except when dealing
with the properties of numbers.
The proof of the rule is an interesting and easy exercise in
the application of reason and commonsense to arithmetic, but
perhaps it is better deferred.
Rule for finding G.C.M.
The rule depends on the demonstration that any factor of
two numbers is likewise necessarily a factor of the remainder
left when one is divided by the other.
Thus consider the two numbers 40 and 24. Divide one
by the other, we get 1 and 16 over. The above sentence in
black type assumes or asserts that every factor of 40 and 24
must also be a factor of 16. In this case, as a matter of fact,
40 = 24 + 16
and it is manifest that a number which divides 24 and does
not divide 16 cannot divide 40.
Well that is the whole idea.
X.] GREATEST COMMON MEASURE. 105
If we were told to find the G.c.M. of 40 and 24, we could
by this means reduce the problem to finding the G.c.M. of 24
and 16. And then, repeating the division process, we should
observe that 24=16 + 8,
so that the problem becomes reduced still further into finding
the G.C.M. of 16 and 8. There is no question but that this is
8 ; — as indeed we might have guessed at first if our object
had been attainment of a result, instead of explication of a
process — and the way to clinch that is to perform the division
again and to find that there is now no remainder at all.
The matter can be stated algebraically, but beginners can
skip the algebra and come to the " illustration " which follows.
Algebraic proof of the process for finding G.C.M.
To find a common factor of two numbers P and Q, of
which P is the bigger,
let X be one common factor,
P Q
then — and - will be the complementary factors.
An extreme case is when P is divisible by Q without a
remainder, in that case x = Q. Suppose however that when
P is divided by Q the remainder is i?,
Q)P{n
nQ
R
so that P = iiQ + R; then if i? is a factor of Q it must be one
of P also (because P equals a multiple of Q plus E), so try
if i? is a factor of Q.
If it is, it is the common factor required ; but if not, work
out a division again, and let the remainder be >S^,
R)Q{m
mR
S
so that Q = mR + S.
106 EASY MATHEMATICS. [chap.
Then if aS' is a factor of R it must be one of Q too, and so also
of P, and in that case S will be the common factor required.
But if not, we must repeat the process and see what the
remainder is when R is divided by S, Call it 2\
S)R{1
IS
so that i? = IS-hT. ^'
Now once more if 2^ is a factor of aS^ it is necessarily a factor
of R, and therefore of Q, and therefore also of F, and so 2'
is the common factor required.
If not, the process must go on until there is no further
remainder; and then the last remainder (or divisor) is a
common factor of the two original numbers P and Q. Let us
assume that T divides S without a remainder, then 2' is the
common factor of all the numbers P, Q, R, S, T.
It is likewise the largest common factor which exists. Why?
because it has to be a factor not only of P and Q but also of
R, of aS', and of T ; and certainly T is the largest factor of Ty
therefore it is likewise the largest common factor of the others.
Statement in another form.
The whole process can be written thus :
To find the g.c.m. of P and Q^ work successive division
sums thus : ■ P R
Q = "■'?
Q . S
R^'^'-R^
S ~ ^^aS'
p 1 1 1
or ^ = rt 4- 7Tn\ — n-\ ;— = 7i +
g ■ Q/i? ' ^ 1 ' ^ 1 '
^^aS/T
the process terminating only when S\T is an integer.
X.] GREATEST COMMON MEASURE. 107
The r is a factor of all the numbers P, Q, R, S, T; and
since it must satisfy this condition if it is to be a factor of
P and Q at all, it is necessarily the greatest common factor
of F and Q, and indeed of the others too.
Or the whole process may be written (as usually performed)
in one sura thus :
Q)P(n
nQ
B)Q(m
mil
'~S)R{1
IS_
T)S{k
TcT
Then the last remainder (or divisor) T is the G.C.M. of P
and Q.
Illustration (modified from Kirkman and Field).
Let the two numbers be 492 and 228. Go through a
process of successive divisions.
228)492(2
456
36)228(6
216
^)36(3
36
Hence 12 is the G.C.M. of the two original numbers, and it
likewise is a factor of the intermediate divisor, viz. 36.
The argument runs as follows :
The common factor of 492 and 228 must also be a factor
of the remainder when 492 is divided by 228, for in fact
492 = (2x228)4-36,
108 EASY MATHEMATICS. [chap. X.
so that anything which divides 228 and fails to divide 36
cannot possibly divide 492.
Hence the problem reduces itself to finding the common
factor of 228 and 36.
But now 228 = (6 x 36) + 12,
hence the factor required must likewise divide 12, as well as
36. The numbers 2, 3, 4, 6, 12 all satisfy that condition, and
hence all these are factors of both the original numbers, but
of them 12 is the biggest.
Therefore 12 is the G.C.M. of the two given numbers 492
and 228. (Verify this by actual division. The quotients
are 41 and 19, and neither of these have any factors at all.)
CHAPTER XL
Easy mode of treating problems which require a
little thought.
Many of the problems set for purposes of arithmetic are
best done in the first instance by rudimentary algebra, that is
by the introduction of a symbol for the unknown quantity,
so that it can be tangibly dealt with. This introduction and
manipulation of a symbol for an unknown quantity need not
be discouraged, even from the first. It confers both power
and clearness. Many arithmetical sums are needlessly hard
because x is forbidden. There is a certain amount of sense in
the artificial restriction, but in complicated sums and in physics
the symbolic treatment of unknown quantities is essential,
and the sooner children are accustomed to it the better.
The introduction of a symbol for an unknown quantity is a
device to enable a sum to be clearly and formally stated.
After the sum has been solved by this aid, it is well to try
and express it so that it can be grasped and understood
without such assistance. The fear of those who object to x
in arithmetic is that this final step may be omitted. The grasp
is clearer when an auxiliary symbol can be dispensed with; but
that is not possible always at first. The x is to be thought of
as a kind of crutch : but sometimes it is like a leaping-pole
and enables heights to be surmounted which without it would
be impossible.
110 EASY MATHEMATICS. [chap.
Example. — How soon after twelve o'clock will the hour and
minute hand of a clock again be superposed ?
It is plain that it is soon after 1 o'clock, and that it is an
amount which has been traversed by the hour hand while the
minute hand, travelling twelve times as quickly, has gone
that same distance and 5 minutes more ; but it is not easy to
think out the required fraction in one's head, though ex-
ceptional children can do it.
But let it be postulated as n. minutes after 1 ; the hour hand
travels, starting from mark I, a distance n, while the minute
hand, starting from mark XII five minutes further back, has
to travel 5 + w in order to catch it up ; so the relative speeds
of the two hands are as (w. + 5) : w, and are also as 12 : 1 ;
wherefore 7j, + 5 12
or 12w = 71 + 5, •
or Wn = 5,
5
" = n'
and so the time required (or the answer) is five minutes and
five elevenths of a minute {i.e. yy hour) past one o'clock.
Take another question. — Start with a clock face indicating
9 o'clock, and ask when the hands will for the first time be
superposed.
The slow-moving hand has forty-five minutes' start; so, how-
ever many minutes it goes, the quick one has to go 45 minutes
more, at twelve times the pace. Wherefore .t -f 45 = 1 2aj, or
the meeting point is yy = 4y\ minutes after the mark IX;
or ^fths of 45 minutes, i.e. j^ths of an hour, since 9 o'clock.
The start in this case is nine times as great as was allowed
after one o'clock, in the previous question, and accordingly the
distance before overtaking occurs is likewise nine times as
great : in accordance with common sense.
XI.] PROBLEMS. Ill
The constant occurrence of 11 in such sums shows that 11
must have a decipherable meaning : it means the excess pace,
or relative velocity, of the quick hand over the slow. And
when this has been perceived, the easiest way to do such
sums in the head is self -suggested, viz. to treat it as a case of
relative velocities, with the hour hand stationary, and simply
ask how soon the minute hand will move to where the hour
hand was^ if it (the minute hand) went at -54ths of its real
speed.
The interval between successive overlaps is therefore always
yfths of an hour, or 65^^ minutes.
Exercise. — The hands make a straight line at 6 o'clock,
when will they be at right angles 1 Ans. : One has to gain
relatively 15 minutes on the other, and since its relative
speed is jjths of an hour per hour, the time required is
15x|f minutes, that is to say lj\- minutes more than a
quarter of an hour.
Pains should always be taken to express an answer com-
pletely and intelligibly. If any joy is taken in work, it
should be decorated and embroidered, so to speak, not left
with a minimum of bare necessity.
Moreover, never let it be taught (as Todhunter taught)
that the x or other symbol so employed is always necessarily
only a pure number. When we say " let a; be the velocity of
the train," or "the weight of the balloon," etc., we should
mean that x is to stand for the actual velocity, the actual
weight: however they be numerically specified. (Appendix 11.)
Some teachers of importance will demur to this. I assert
with absolute conviction that it is the right plan, and will
justify it hereafter. But it is a matter for adults to consider,
and is only incidentally mentioned here.
The dislike felt by teachers of arithmetic to the intro-
duction of z prematurely, is because there is a tendency
112 EASY MATHEMATICS. [chap.
thereafter to do arithmetical problems so easily that their
features are not grasped, and so some useful perceptions are
missed. If this were a necessary consequence it would be
a valid argument against the introduction of an algebraic
symbol, but it is not a necessary consequence.
For instance, in examples about the supply of a cistern by
pipes, or the work of men per day, it is admittedly desirable
to realise that we are here often dealing with the reciprocals
of the specified quantities ; and this may be masked by the
use of algebra, possibly, but it need not. I suggest that
algebra is the right way of discovering the fact, but that after
its discovery the fact itself may be properly dwelt on, and
thereafter directly applied. There is indeed too much ten-
dency to hurry away from an example when its mere
"answer" has been obtained, without staying to extract its
nutriment and learn all that it can teach : sometimes without
even trying whether the answer found will really fit or
satisfy the data in question. That is altogether bad. The
full discussion of a sum, in all its bearings, after the answer
is known, is often the most interesting and instructive part
of the process.
Children should always be encouraged to do this, and to
invent fresh ways of putting things, or detect or devise a
generalisation of their own for any suitable special case. Here
is afforded a first scope for easy kinds of originality of a
valuable kind.
Girls especially would find the benefit of being encouraged
to seek the general under the mask of the special. It seems
to fail to come to them naturally.
Illustrative Examples, showing the advantage of intro-
ducing symbols for unknown quantities.
Three pipes supply a cistern which can hold 144 gallons.
XI.] PROBLEMS. 113
One supplies a gallon a minute, another 2 gallons, and the
third 3 gallons per minute. How soon will the cistern be full 1
Let t be the number of minutes before the cistern is full
after the pipes are all turned on simultaneously ; then in t
minutes the first pipe will have supplied t gallons, the second
2^ gallons, and so on,
hence t + 'it + U = 144.
So t = U.
This is easy enough, but I think even this is made easier
by the introduction of a symbol for the unknown quantity.
Take however the following variation of the same problem :
A cistern is to be filled by three pipes labelled A^ B, and G;
Pipe A alone would fill the cistern in 2 hours 24 minutes.
Pipe B alone in 1 hour 12 minutes.
Pipe C alone in 48 minutes.
How soon would they all three fill it 1
This form of statement evidently makes the problem harder,
and it is clearly desirable to simplify it by ascertaining the
rate of supply of each pipe. This can be done at once if we
say, let n be the number of gallons corresponding to the
contents of the cistern, for then the data give us that
Pipe A supplies at the rate of n gallons in 144 minutes
or — — gallons per minute,
B supplies at the rate -- gallons per minute,
72
and C supplies at the rate -- gallons per minute.
48
So the set of pipes together supply, at the combined rate,
n n n _ n
l44"^72'^48 ^ T
that is to say, n gallons in the unknown time t, which time is
the thing to be found.
L.E.M. H
114 EASY MATHEMATICS. [chap.
We now see that the contents of the cistern is immaterial,
when the data are thus specified, for n cancels out of the
equation, and leaves us with the relation
11 1 1
t ~ 48 "^72 "^144"
We have thus discovered the mode of dealing with problems
of this kind, viz. to take the reciprocals of the times given.
In other words, to say that the rate of supply is inversely as
the time taken, or that it is proportional to the reciprocal of
that time ; and hence, writing the combined rate as the sum
of the rates, we get the equation directly as last written.
Now it is true that a mathematician would have seen this
at once, and written the equation as above without appearing
to think about it; but a child cannot be expected to think
out such a relation, at least not for a long time, unless he is
encouraged to consider, either tacitly or explicitly, the
contents of the cistern ; when it at once becomes, not exactly
easy but, possible.
The above equation may be called "the solution" of the
problem, so far as it involves reasoning or thought; the
subsequent arithmetical working necessary to obtain a
numerical result is comparatively mechanical, but it should
not be omitted.
J_ J^ 1 ^ 3 + 2 + 1 ^6^1
48 ^72 "^144" 144 ~ 144 ~ 24'
This is the reciprocal of the time ; and thus the time required
for the conjoint filling is 24 minutes, as we found in the first
or easy mode of statement, where the rates were explicitly
specified among the data.
Another question of the same kind : If A can build a wall
in 30 days, B in 40 days, and C in 50 days, how soon can
they all build it, if they can all work together without
interfering with each other |
XL] PROBLEMS. 115
Answer in x days, where
-1 i- i_ = 1-
30'^40"^50 x'
because, during each day, A does ^^^th of the wall, B does
^th, and G does ^^^th ; so the three together do, each day,
what is represented by these fractions added together. Hence
the number of days will be the reciprocal of the sum of these
fractions.
It is probably undesirable to assist a beginner to so easy a
solution of this class of problem prematurely, or until he has
been afforded an opportunity of expending some thought
upon it ; for it is difficult to get a good grip of a thing which
is too smooth and slippery.
CHAPTER XII.
Involution and evolution and beginning of indices.
Because 6x6= 36, which may be called 6^ (six square),
and 6x6x6= 216, and may be called 6^ (six cube),
6x6x6x6 = 1296 = 64, (six to the fourth power),
and so on,
it is customary to call 6 the square root of 36 ;
it is also the cube root of 216,
the fourth root of 1296,
and so on ;
and the process of finding, or extracting, the root of any
number is called evolution, — though the name is of small
importance.
The idea of roots and powers however is of great importance
and it is necessary to know how to find them.
The square root of 49 is 7 ; as we know from the multipli-
cation table. So also we know in the same way, that is by
direct experiment, that the square root of 64 is 8 ; because
this is only another way of saying that 8 square, 8^, or 8 x 8,
equals 64.
The statement that 9^ = 81 is identical, in everything except
in form, with the statement that the square root of 81 is 9.
The square root of 100 is 10,
that of 144 is 12,
and of 400 is 20.
CHAP. XII.] POWERS AND ROOTS. 117
A notation or mode of writing is necessary for roots, to
avoid having constantly to write words, and for compactness ;
just as 3 is handier to deal with than "three," though it
means the same thing.
The notation employed in involution or raising to powers
we have already stated (p. 56), viz. little figures or indices
placed after the main figure, as for instance 4^ = 1 6, the index
denoting how many fours are to be multiplied together.
So 6^ means that three sixes are to he multiplied together ;
and that is all that the index shows.
9^ means that five nines are to be multiplied together ; and
the result is a big number, which a child may at once be set to
calculate. He might also calculate such numbers as 2^, 3^, 4^,
55, ..., 99, 1010.
Moreover, he should at once write down the values of the
following: 102, 10M0M0M0«, ....
and perceive that in each case the number of ciphers following
the one is indicated by the index. So he can write down in
full 10^^ without consideration, and can be told that the short
form is a compact and handy and universally adopted method
of expressing large numbers.
From all this, if a sharp child were asked to invent a
notation for roots, he might perhaps, though it is much to
expect if really ignorant of the convention, but he might
suggest that since 4^ = 16, perhaps 16^ = 4; or perhaps
he might suggest 16 "^ as a suitable notation. In either
case he should be much encouraged.
Of the two notations, thus suggested, the first is correct
and is employed. The second is employed for something
also, but for something totally different from a root, viz. a
reciprocal.
Let us get used to the notation for roots by fractional
118 EASY MATHEMATICS. [chap.
indices, and at the same time justify it as a consistent and
convenient method.
First of all it must be admitted as not easy to put into
words. The index 2 signifies that the number is to be raised
to the second power, or multiplied by itself, so that
42 = 4x4 = 16;
hence we might say that 16^ means that the number is to be
raised to the half power, or multiplied, — how *? It is hardly
an interpretable phrase ; so we must proceed more gradually.
First of all, it is simple to suppose that if the index is unity
it should be understood to leave the figure unaltered, so that
41 == 4 and I61 = 16 ;
therefore we may write indices on both sides, thus, 16^ = 4^;
let us next suppose that we may halve the index on each side
getting 16^ = 4^, and read this, root 16 equals 4. We might
halve the indices again, and get 16* = 4*; which equals the
square root of four, or 2 ; so that we may surmise that the
fourth root of 16 is 2 ; and verify it thus,
2x2x2x2 = 16.
Similarly, 27 = 33,
27^ = 31,
which agrees with the fact that the cube root of 27 is 3,
(27 = 3 X 3 X 3).
Again, 81^ = 9,
8ii = 9* = 3.
All that we have here assumed (and it is a large assumption)
is that in an equation involving terms with indices, if we
perform an operation on the indices — provided we perform
the same operation on both sides, — the equality remains
undisturbed.
This is an assumption, a guess, an expeetation, to be justified
or contradicted experimentally by results. We shall find that
Xii.l POWERS AND ROOTS. 119
its truth depends entirely on the kind of operation so per-
formed. We happen to have hit first upon trying multipli-
cation and division as applied to indices, and that seems to
work correctly. But we shall try other operations shortly
and will find them fail.
Those who imagine or assert that experiment has no place
in mathematics do not know anything about mathematics.
Sometimes results are arrived at by theory, sometimes by
experiment, sometimes by a mixture of the two ; — either
theory first and confirmation by experiment, or experiment
first and justification by theory : just as in Physics or any
other developed science.
Let us now press our assumption to extremes and experi-
ment on it in various ways so as to see whither it will lead us.
Start with any equation, such as
42 = 161.
Double each index, and we get
44 ^ 162 ^ 256.
So the fourth root of 256 is given as 4, and the eighth root
will accordingly be 2, or 256 is asserted to be the eighth
power of two ; which is the fact : eight twos multiplied
together do yield 256.
Treble the index, and it becomes
46 = 163 = 256 X 16 = 4096,
or conversely, 4096^ = 4.
Hence the sixth root of 4096 is given as 4, and
4096tV = 4* = 2,
that is, its twelfth root is 2. Again fact agrees with theory.
2 multiplied by itself 12 times does equal 4096.
Hence it appears that the operation of multiplying indices
by any the same factor on each side of an equation may be
trusted to give true results.
120 EASY MATHEMATICS. [chap.
So also division of indices by any the same number may be
trusted too ; thus starting as before with
161 = 42^
quarter each index 16^ = 4^ = 2,
halve each again 16^ = 4^ = 2^.
How are we to calculate 2^ 1 That is not an easy matter :
we will leave it unvalued for a time and merely call it the
square root of 2. It is often denoted by a sort of badly-
written long-tailed -^ in front of the digit, thus, 4/'! or ^2.
Try again, 27 = 33,
27* = 3,
27* = 3*,
there is the same difficulty about interpreting 3^ ; no whole
number will serve. We can call it the square root of 3, or
briefly " root 3," and can denote it by writing ^3 as before.
^16 means the same as 16^, namely 4 ; and ^4 = 4^ = 2 ;
but whereas the fractional index contains an important and
valuable idea, which remains to be developed, the symbol J
is nothing but shorthand for the word "root," and is itself
trivial and inexpressive, though quite harmless and of constant
service.
What we have learnt from the above examples resulting in
^2 and ^3 is that when employing fractional indices we can
arrive at something, easy of interpretation indeed, but not
easy of numerical evaluation ; there is no need to mistrust the
result but only to wait till more light can fall upon it.
Now try some other operations applied to indices, we shall
find that wariness is necessary, and that mere guesses and
surmises as to what it is permissible to do to equations are
not worth much. Everything must be tested. Suppose we
try squaring them on both sides as thus : Starting with
42 = 161,
xii.] POWERS AND ROOTS. 121
squaring indices would give us
44 ^ 161,
since the square, or any other power, of 1 is 1,
1x1x1x1 = 1.
The result, that 16 is both the square and the fourth power
of 4, is false and absurd : and hence the sham equation is
erased.
So we learn that whereas multiplication of indices by any
factor is an operation that can be trusted to give true results,
and division of indices by a factor can probably be trusted
too, since one operation is the inverse of the other, yet that
involution is not an operation that can legitimately be per-
formed upon indices, but only upon the numbers themselves.
Suppose we try addition, equal additions to the indices on
each side; add 1 for instance, we get 4^ ^ 16^, which is a
falsehood if the equality sign is left unerased.
It is time we began to consider what operations are really
legitimate and what are not ; and gradually in both cases we
must proceed to ask. Why ?
CHAPTER XIII.
Equations (treated by the method of very elementary experiment).
It is therefore convenient at this stage to introduce the idea
of an expressed equality, which is called an equation, and to
consider what are the operations to which an equation can be
subjected without destroying the equality.
It is customary to postpone this subject to Algebra, but we do
not wish to perpetuate any sharp distinction between algebra
and arithmetic, and it is useful to begin experimenting with
equations while still they are expressed in terms familiar to
beginners.
Typical equations are of many kinds, of which we may
now consider the following :
The addition kind, 3 + 2 = 5.
The subtraction kind, 3-2 = 1.
The multiplication kind, 3x2 = 6.
The division kind, 3 -r 2 = 1-5.
The involution kind, 3^ = 9.
The evolution kind, 9* = 3.
There are plenty of others, but these will do to begin with.
Every equation has two sides, called respectively the left-hand
side and the right-hand side; the symbol = is the barrier.
It is not an impassable barrier, but terms get reversed when
they are taken across it ; positive becomes negative, and vice
versa. In order to find out what may be done to equations
we can experiment.
CHAP. XIII.] EXPERIMENTS ON EQUATIONS. 123
Take any of these equations and try experiments on it.
For instance, add something to or subtract something from
each side. So long as we add the same thing to each side no
harm is done : the equality persists. For instance, start with
the first two of the above equations and modify them by
addition or subtraction in various ways : —
3 + 2 + 7 = 5 + 7 = 12;
3-2-1 = 1-1=0;
3 + 2-6 = 5-6= -I,
3-2 + ft = l+a,
x-h3- 2-« = x-a + lj
3-2 + 2 = 1+2 = 3,
3 + 2 + 1 = 51
So far everything is very simple and safe.
Not only may we add the same thing to each side, but we
may add equal things to each side (which may be regarded as
an illustration of the axiom, that if equals be added to equals
the wholes are equal).
Thus 3 + 2 = 5,|
and also 7 + 6 = 13. J
So 3 + 2 + 7 + 6 = 5 + 13 = 18.
Or again, 3-2
and 32
So 3 -2 + 32 =1+9 = 10.
Also take the following :
3-2 = 1, 1
and 5-6 = -l;j
.-. 3-2 + 5-6 = 1-1=0.
But take an equation of the multiplication kind,
3x2 = 6,
a little caution is necessary in adding anything to the left-
hand side.
:l-}
124 EASY MATHEMATICS. [chap.
We might have (3 x 2) + 1 = 6 + 1 = 7,
or we might have 3 x (2 + 1) = 3 x 3 = 9.
If we only write 3x2 + 1, without brackets, it is ambiguous ;
for the value depends on whether the addition or the multi-
plication is performed first: that is, on whether the 2 is
grouped along with the 3 or with the 1, but the brackets
enable us to indicate the grouping clearly.
Take another example,
7 X 8 = 56,
(7 X 8) - 4 = 52,
although 7 X (8 - 4) = 28 ;
but the last is quite a different equation, and is not deduced
by simple subtraction of 4 from both sides.
About the other forms of equations there is no difficulty ;
we will just write them, with something either added to or
subtracted from each side :
1-1 = 1-5-1 = -5 = I,
32 - 1 = 9 - 1 = 8,
9^+2 = 3 + 2 = 5.
Incidentally we here observe the advantage of the fractional
notation over the + notation. If we had written 3 + 2-1 we
should have had to avoid ambiguity by the use of brackets, as
was necessary in multiplication ; but f - 1 is unambiguous.
Unity is subtracted from the whole fraction, not from either
numerator or denominator. If unity were subtracted from
the numerator it would not be right,
^^1-5-1 = -5;
nor will it do to subtract from the denominator, nor from both.
So much at present for addition and subtraction ; now try
multiplication and division : start with
3 + 2 = 5;
XIII.] EXPERIMENTS ON EQUATIONS. 125
double each terra, 6 + 4 = 10 ;
treble each term, 9 + 6 = 15;
halve each of these terms,
41 + 3 = 7i.
So here we are safe.
Proceed now to the factor or multiplication form of equation :
3x2 = 6;
double each digit, 6x4 ^ 12,
and we get wrong.
We learn that we must not double each factor in a product,
though we must double each term of a sum ; hence the expres-
sion 3 + 2 is commonly spoken of as containing two terms, but
3 X 2 is spoken of as a single term.
To double the single term it is sufficient to double one of its
factors ; so if we write
6x2 = 12
we get right again.
Similarly we must halve one of the factors only,
11 X 2 = 3,
or else 3x1 = 3.
Now attend to the quotient form,
double every digit, f ^ 3,
and we get wrong.
Double the denominator only,
it is still wrong.
Double the numerator only, f = 3,
and we get right.
^ So in multiplication and division of a quotient by a whole
* number, the factor has to be applied to the numerator
only.
126 EASY MATHEMATICS. [chap.
Take another example, ^ = 4,
6 _ 4 _ 9
3 _ 4 _ 1
^-T - ^»
-V- = 4x2 = 8.
Finally, take the involution form,
32 = 9.
What are we now to double if we want to double both
sides 1 62 = 18 is wrong,
3^ == 18 is also wrong.
We cannot do it quite so simply ; so we must write merely
2 X 32 = 18,
which leaves the step really undone and only indicated.
But take another example,
3 X 32 = 27.
This could be written 3^
Again, 33 x 32 = 27 x 9 = 243 = 3^,
and a rule of extraordinary interest and usefulness is suggested.
Think it over, we shall return to it in Chapter XVI.
Further consideration of what can be done to equa-
tions.
A sentence like the following :
" If both sides of an equation be treated alike, the equality
will persist," might easily be considered axiomatic ; but so
much caution is required before we can be sure that both
sides have been really treated alike, that it is highly dangerous
to employ such an axiom. We have already come across
some cases of the danger, but the subject is very important
and will bear fuller treatment.
The general doctrine may be laid down that before we
understand properly what can be done, or what it is per-
missible to do, in any subject whatever, we should take pains
XIII.] EXPERIMENTS ON EQUATIONS. 127
to ascertain also what cannot be done under the same circum-
stances, i.e. what it is not possible to do without error.
This latter part should not be too long dwelt upon, because
error is most simply excluded by attention to and familiarity
with the correct processes, so that presently all others instinc-
tively feel wrong ; but once at least we should examine the
whole matter, and learn, if we can, why one set of things are
wrong and another set right. This remark applies also to
other things than arithmetic.
An equation consists of two sides, and each side consists of
terms. Frequently the right hand side is zero, especially in
algebra and in higher mathematics. Sometimes, instead of
being zero, it is some constant or other independent quantity,
and is called " the absolute term," because it is undetermined
by anything on the left hand side : to which however it is
equated.
An equation is the most serious and important thing in
mathematics. The assertion that two quantities or two sets
of quantities are equal to each other, whether it is meant
that they are always equal, or only that they are equal
under certain circumstances which have to be specified, is a
very definite assertion and may carry with it extraordinary
and at first unsuspected consequences.
The equations we are now using as illustrations are by no
means of this high character ; they are usually mere identities,
and depend on the truism that a combination of things grouped
or expressed in one way are unchanged in number when
grouped or expressed in another way."*^ But although it may
* We call this a truism ; but it is a dangerous term to employ, and
when we come to Chemistry we must be on our guard against assuming
that the volume of Hg + O is equal to that of HgO, under the same
external circumstances. It is true of weights (as nearly as we can tell)
but it is not even approximately true of volumes,
128 EASY MATHEMATICS. [chap.
be some time before they realise the vast importance attaching
to equations, children will take it on trust that they are now
entering the central arcana of the subject, and will be willing
to give the needful attention to the processes which have
constantly to be employed. An initial account of them is
given in the following chapter, parts of which may be read
before the whole of the following introductory matter.
When a number of quantities are multiplied together, they
are held to constitute one term. Whenever the sign + or -
intervenes, it interrupts the term, and each such sign has a
term on either side of it.
Thus a + bj 70-6, are each two terms ; but ab, and 70 x 6,
and abc, and 10^, and 5aJ'2, and are all single terms.
X
What about such an expression as
5ab-x
abx
where there is one (or more than one) addition or sub-
traction sign in the numerator ?
Answer : So long as it is kept all together it can be called
one term, but it can easily be split into two, viz.
X ab^
and for some purposes its terms can be considered plural
without re-writing.
The long line of division in the original expression however
may be held to weld the whole into one term ; and brackets
have the same effect. Thus,
(a + b), (70-6), 5{a + b)x, J (ax -by)
are all single terms once more ; until the brackets are
removed. And removal of brackets is an operation to be
performed cautiously. Rubbing them out is not a legitimate
way of removing them.
XIII.] EXPERIMENTS ON EQUATIONS. 129
For instance, 3(7-4)= 9;
but 37-4, and 3x7-4, and 3-7-4
are all different.
Again, ^(16 + 9) = 5,
but V16+V9 = 7,
and ^^6+ 9 = 13,
while 16+^9 = 19;
the three are entirely distinct statements from the first, and
are not deducible from it.
So we learn that the right removal of brackets is a matter
to be studied.
When we assert that the same operation can properly be
applied to each side of an equation then, we must be careful to
interpret it always as an operation applied to the whole side,
and not to any part of it. We may not tamper with one
term and leave the others alone, nor must we tamper with a
part of a term only. Nor must we repeat the operation for
each of the factor components of a single term.
This must be illustrated :
Given that a + b = c,
it is correct to say that
2a + 2b = 2c,
or that 2(« + &)= 2c;
but given that axb = c,
it is not correct to say that
2a X 2b = 2c.
For here the term ab is one, and it only needs doubling once.
Given that a^ = J^,
it is true that 5^2 = 553^ (1)
but it is not true that (5a)2 = (55)3, (2)
for that would mean 26a^ = 125b^,
which, subject to the given data, is absurd, unless a and h are
both zero.
L.E.M. I
130 EASY MATHEMATICS. [chap.
In reading the two lines labelled (1) and (2) it is customary
to read them carefully in order to discriminate what otherwise
would sound quite similar. The former of the two lines is
read t( five a-square = five b-cnhe " ;
the latter of the two lines is read
" five a, squared = five b, cubed " ;
and these are quite different. They cannot under any circum-
stances be both true (unless indeed a and b are both zero).
They are therefore called " inconsistent " equations
(like X = y, and x = 2y, which cannot both be true).
To illustrate the inconsistency, take an example :
82 = 43, both being 64,
and so also 5 x 8^ = 5 x 4^, both being 320,
but (5 X 8)2 ^ (5 X 4)8, the one being 1600
and the other 8000.
Read the sign ^ as " does not equal. "
Given again that a^ = 6^,
it does not follow that a^ = b\
although we have done the same thing, that is added 1, to the
index on each side.
Nor would it be true to say that
although we have now squared the index on each side.
But it does turn out true that if we double or treble the
index, the equality persists : given that a^ = b^
it is true that a^ = ¥
and that a^ = b^,
so that it appears as if it were permissible to multiply the
index on each side by any the same factor. We must examine
this later, but at present we will merely verify the truth of
these last assertions by an arithmetical example :
XIII.] EXPERIMENTS ON EQUATIONS. 131
For instance 8^ = 4^,
whereas 8^ ^ 4^
one being 512, and the other 256 ;
but 84 = 46,
both being (64)2 ^r 4096.
Likewise 8^ = 4^,
both being (64)3 or 262,144.
Now take a slightly more general type of equation.
ax = hy^
it follows that (aa;)^ = (5y)^,
but it is by no means necessary that aa:^ si^^ll equal li^.
For instance, 7x4 = 14x2,
and (7x4)3 = (14x2)3;
but 7x43^14x23,
for one equals 448, and the other equals 112.
Take, as given, the equation db — xy, and let us multiply,
add, and divide on both sides, so as to illustrate legitimate
and illegitimate operations; the pupil being left to devise
numerical illustrations and tests for himself.
First multiply or divide by any quantity whatever, say c.
abc = cxy ;
or, assuming another quantity z = c, we may write it
abc = xyz.
So also ^ = %
c z
axb-^c = xxy~z = yxx-i-z,
-(ao) = -xo = ax- = xx- = yx- = ~(xy).
c^ ^ c c z ^ z z^ ^'
Next add or subtract something to or from each side.
ab-\-c = xy + c,
ab-b = xy-b ^ xy-y,
ab-x = xy-x,
ab-xy = xy-xy = 0.
132 EASY MATHEMATICS. [chap.
This last is worth attention. The result has been to
transfer a term from one side of the equation to the other,
its sign being changed in the process. This is important and
demands further illustration.
Let JB = 6.
We can equally well write it, by subtracting 6 from both
sides, a; - 6 = 0,
and the 6 has been transferred, with change of sign.
Or let X = y.
Subtract y from both sides,
then x-y — 0.
Again let a; = - y,
then add y to both sides and we get
x + y = 0.
Or let ax + by = -Cj
add c to both sides, ax + by + c = 0.
This kind of simple operation has constantly to be per-
formed.
One more illustration therefore:
Let ax-{-by — cx + dy.
We can subtract the right hand side from both sides ; in
other words, transfer it to the left, with change of sign;
getting ax + by-cx- dy = 0,
which is more neatly written
{a-c)x + {b-d)y — 0;
or again, (a-c)x = {d- b)y.
Here the last mode of expression is deserving of attention.
We will arrive at it more directly.
To this end start again with
ax + by = cx + dy;
transfer by to the right, and ex to the left ; thus we get
ax -ex = dy-by,
XIII. ] EXPERIMENTS ON EQUATIONS. 1^3
or what is the same thing,
(a-c)x = (d-b)y.
Divide each aide/^y the product {a-c)(d-b) and the
equation becomes / \
/(a-c)x _ (d- l)y
(a-c){d-h) ~ (a-c)(d'-by
In each of these terms there is a common factor in
numerator and denominator, so we can cancel them, and are
left with -^ = -y..
d-o a-c
Or we might have divided otherwise, and arrived at any of
the following :
X _ d-b
y ~ a-c
x:y = (d-b):{a-c),
y:x = (a-c):{d-b\
d-b
X = y,
a-c
(i (a-c)x _ -
(a-c)x _
{d-b)y '-"'
^1^-2^ = 0
d-b X
All these are entirely equivalent forms ; and as an exercise
they should all be deduced, any one from any other.
And all can be numerically illustrated, by attributing to
the symbols some particular values ; for instance by taking
a; = 24, y = 2, a = 5, 6 = - 19, c = 1, d = 29.
Let this be done, as an exercise.
134 EASY MATHEMATICS. [chap.
Among things that can legitimately be done to equations
are certain operations which are by no means obvious, and
demand attention.
Suppose we are told that
? = ^ (1)
We are not allowed to say — ^ = —I— ; but we are entitled
to say that ^
x-y _a-h ..
~V~~V' ^^^
because this is equivalent to subtracting unity from both
sides, Le. is equivalent to
?-l=^l.
y 0
So also we might have truly written
^±i' = ^ (3)
But from the truth of these two operations it follows that
we might also have written
x-y_^a-h .^.
x + y a + b ^ ^
For this would be obtained if we had divided each side of
equation (2) by the corresponding side of equation (3) ; for if
equals be divided by equals the quotients are equal.
Let us illustrate this important result arithmetically.
Start with -^ = —-, which can be easily proved true, and
may be taken as corresponding to (1).
Then it follows that
14 _ 6 49-21 . ,, ,8 28 , . ,
— - — = — — — ; in other words - = r^; which
6 21 ' 6 2r
corresponds to the form numbered (2) ;
XIII.] EXPERIMENTS ON EQUATIONS. 135
1 1-1 /Qx ..i> ^ 14 + 6 49 + 21 . .. . 20 70.
also, like (3), that = — ^r — , *•«• that -g- = gj '
A VI fA\ ^v, ^ 14 + 6 49 + 21 ^^ ... 20 70
and, like (4), that ^^^^-^ = ^^-^, or that _ = -.
Or each member of any of them may be inverted : for
instance the last : A - ??
20 " 70'
Starting once more with
X _ a
we might equally well write it
1 = 1 (5)
a 0
for this is the result of multiplying both sides by ^ ; so there-
fit
fore it is true to say that
a^-^ ^ IzA (6)
and — — = ^-T-, (7)
a 0
also that ^:i^ = ^, (8)
x + a y + b
and that is really, though by no means obviously, the same
thing as equation (4).
Illustrate this too, numerically, with the same numbers as
^^^^^« '- M = 1. corresponds to (5)
35 15
- -— = - ^— corresponds to (6)
49 21
where the minus signs may equally well be omitted or can-
celled by multiplying each side by - 1 ;
35 15
and _ = -^ corresponds to (8)
136 EASY MATHEMATICS. [chap.
or again any of them may be inverted, e.g.:
49 21 ^
Now let us apply the so-called involutional operations to
both sides of an equation, and ascertain what we may do and
what we may not do.
Begin with ah = xy.
Square both sides, {ahY = {xyy.
Square each factor, a^h^ = z^y^.
Square one of them only, and we get wrong,
a&2 ^ xy'^.
Take the square root of both sides,
J{ab) = J{xy\
also of each factor, Jajh = Jxjy,
or what is the same thing,
ar¥ x^y^.
So far we are all right except in the one marked instance :
as can be tested by giving suitable numerical values to the
four symbols.
But now take an equation with more than one term on a
side, say x + y == c.
Square both sides (x + yY = c^.
Square each term, x^ + y^ ^ c\
and we get wrong.
This is a mistake constantly being made by beginners, and
it must be further emphasised. As an example,
4 + 5 = 9,
(4 + 5)2 = 92 =, 81,
but 42 + 52 = 16 + 25 = 41 ^ 92.
The following fallacy may serve as an illustration :
^25 = J(U + 9), .'. 5 = JU+J9 = 7.
XIII.] EXPERIMENTS ON EQUATIONS. 137
Observe that these numerical instances, if they lead to
error, show quite decidedly that the operation tested is wrong.
They do not prove with equal validity that it is right,
if they turn out correctly : certainly a single instance of correct-
ness is insufficient. They render its rightness probable, but
the rationale of it has to be further investigated. A single
instance of real error however is sufficient to invalidate any
operation under test.
Exercises. — Test the correctness of the following horizontally
juxtaposed statements :
2x3 =
6.
2 + 3 = 5.
22 X 32 =
62.
22 + 32 ^ 52.
5x6 =
30.
5 + 6 = 11.
52 X 62 =
302.
52 + 62 i? 112.
23 X 63 =
123.
23 + 63 ^ 123.
23 + 63 ^ 83.
4x9 =
36.
4 + 9 = 13.
4^ X 9^ =
36* = 6.
4^ + 9* = 5.
9x144 =
: 1296.
9 + 144 = 153.
^9x^144 = V(1296). ^9 + ^144 = 15.
27 X 216 = 5832. 27 + 216 = 243.
27^ X 216^ = 5832i 27^ + 216^ = 9.
or 3 X 6 = 18.
But now it must be admitted that this experimental mode
of treatment may not be considered the best mode of beginning
the experience of equations : and it is certainly not the most
conducive to rapid progress; it may be better therefore to
apply treatment like that of the present chapter at a rather
later stage and to use it as a cautionary and salutary exercise.
The importance of the subject is so great that it can hardly
be over- emphasised, nor is one mode of approach sufficient.
In the next chapter a somewhat more orthodox and quite
effective mode of procedure is adopted.
CHAPTER XIV.
Another treatment of Equations.
Equations may be classified in various ways: there are such
things as differential equations, there are quadratic equations
and equations of the fifth degree, etc., but for the present we
will classify them under three simple heads : —
1st. Statements of specific or particular fact, such as :
3 + 4 = 7,
or 9,7(144) = 108;
these involve only definition and re-grouping.
2nd. Statements of general or universal truth, such as :
n^-l = (n + l){n-l)
log a' = xloga;
these are called identities, and are frequently denoted by a
triple sign of equality = , for instance a + 6 = 6 + a, whenever it
is desired to emphasise their distinction from the third class.
3rd. Equations proper, or statements of condition or in-
formation such as :
17a; = 34,
or 6x^ = 40 ;
statements which are not by any means generally true, but
are only satisfied by some implicit datum, such as, in the
above instances, x = 2.
4th. There is also, from this point of view, a fourth class of
CHAP. XIV.] EQUATIONS. 13d
equations, expressive of a relation between two quantities,
such as 3x + iy = 12,
or x^ + y^ = 25;
which are satisfied not by all possible values of x and ?/, as
an identity is satisfied, but by an exclusive and definite
though infinite series of values.
The first is satisfied, on a certain geometrical convention, by
all the points which lie on a specific straight line, the second
by all the points which lie on a definite circle.
If the equations are given simultaneously, they are satisfied
together by two and only two points, viz. the points where
the straight line cuts the circle.
With this fourth class we have nothing to do just yet : it
opens up a large and exhilarating subject.
With the first kind of equation we have constantly had to
do already : all purely arithmetical equations are necessarily of
this kind.
The second kind is constantly encountered throughout
algebra and trigonometry ; identities represent the skeleton
or framework of mathematical science, all its universal and
undeniable truths can be thus expressed.
The third kind of equation, or equation proper, — equations
which have a definite solution, equations which convey specific
information about an unknown quantity, and express it in
terms of numbers or known quantities of some kind — those
are the equations with which we deal in this chapter, and
that is the kind which gives immediate practical assistance
towards the solving of problems.
The process of " solving " an equation is simply the act of
reducing it to its simplest possible form. Written in any
form the equation conveys the same information, but in some
forms it is not easy to read ; the solving of it is analogous to
140 EASY MATHEMATICS. [chap.
the interpretation of a hieroglyph or the translation of an
unknown phrase.
For instance the following equations
13a; = 65
x= b.
all express the same fact and convey the same information
concerning x, but the last obviously conveys it in simplest
form, and it is called the " solution " of the first, the second
being an intermediate step.
The two sides of an equation may be likened to the two
pans of a balance, containing equal weights of different
materials or of the same material differently grouped. It
is permissible to take from or to add equal quantities to
both pans, the balance or equality being still preserved ; but
a weight must not be taken out of one pan and added to the
other, unless its force be reversed in direction and made to
act upwards instead of downwards ; which can be actually
managed by hanging it to a string over a fixed pulley, the
other end of the string being attached to the pan.
This fact is most simply expressed by saying that if any
term or quantity is transferred from one side of an equation
to the other, it must be reversed in sign, if the equality is still
to persist, i.e. if the equation is to remain true.
This is a simple but important matter of constant practical
use, and it requires illustration :
Let the equation be given
a; -2 = 3.
We can transfer - 2 to the other side of the equation, where
it will become + 2, giving us a; = 3 + 2 or in other words
X = b. (We may consider that we have added 2 to each side.)
The value 5 obviously satisfies the equation in its original
form, because it is true that 5-2 = 3 ; and the substitution
XIV.] EQUATIONS. 141
of the found value in the original equation and then seeing if
it fits or holds good, is called ' verifying ' the solution.
Take another case :
3x2+17 = 4a;2- 8.
Getting the unknown quantities on one side, and the known
on the other, it becomes
3x^-ix^ = -8-17;
if we like we may now reverse the sign of every term, which
will give us
4a;2-3a;2 = 17 + 8
or fl;2 = 25
or ic = 5.
Thus all the equations we have written recently happen to
be expressive of the same fact : namely that the particular x
denoted by them is merely the number 5. Substituting this
number in the above equation, it becomes
3x25 + 17 = 4x25-8
or 75 + 17 = 100-8
which we perceive to be an arithmetical identity, since both
sides = 92,
thus the solution is verified, or the value a; = 5 is proved to
satisfy the given equation.
We do not know for certain that it is the only value that
will satisfy it, but at any rate it is one solution. It so
happens that the equation last written will also be satisfied
by the solution x = -6 ; and this is characteristic of square
or quadratic equations in general, that there are two answers
instead of only one.
An equation of the third degree, that is an equation
involving x^, will in general have three answers ; and so on.
Take one more quite simple example, for practice :
given 7a;- 12 = 5a; + 6, to find a;.
142 EASY MATHEMATICS. [chap.
Subtract 5x from both sides, or, what is equivalent, transfer
5x over to the other side with change of sign,
we get 2a;- 12 = 6.
Now add 12 to both sides, or, what is the same thing, transfer
- 12 over to the other side, and it becomes
2a; = 6 + 12 = 18;
wherefore a; = 9 is the solution.
Try it in the original equation in order to verify it and we
get 63-12 = 45 + 6,
which is an arithmetical identity.
As to algebraic identities, it is probably needful to remind
young beginners occasionally even of such simple facts as
these : at the same time making mysterious hints that there
are possible interpretations, to be met with hereafter, wherein
even these simple statements lack generality and are open to
reconsideration, a + b = b + a^
and ab = ba;
and they should be frequently reminded of such useful
identities as (a + bf = a^ + 2ab + b\
\a-hf = a^-2ab + b^,
{a + b){a-b) = a^-b\
Oral questions should be asked at odd times concerning
equivalent expressions for such things as
{V + q)\ {^-W, {x + y){x-y\
{X+\){X-1\ (71+1)2, (^_l)2,
(l + a)2, {\-a)\ (a; + 3)2,
{x - 5)2, (3 + 72)2, (7 4. ia,)2^ etc., etc.,
since a pupil's knowledge of such fundamental things should
be ready for immediate application — like a well constructed
machine.
We have not yet taken an example of an equation involving
a;2 as well as a;, because they are not quite so easy to solve ;
XTV.] EQUATIONS. 143
but a parenthetical remark may be introduced even at this
stage. We know that quantities of different kinds do not
occur in one expression; in other words, that all the terms
of an expression must refer to the same sort of thing, if they
are to be dealt with together or equated to any one value.
Nevertheless an expression like a^ + 6x^ + 2x -h 6 is common,
and X may be a length; which looks as if we could add to-
gether a volume, an area, a length, and a pure number.
Not so, really, however : see Appendix.
The equation (x - 3) (re - 4) = 0,
written out, becomes ic^ _ 7^; + 12 = 0,
or x^ = 7a;- 12.
We may guess at numbers which will satisfy this equation,
and we have been told there must be two, because it is a
quadratic : it contains x^. By trial and error it will be found
that the number 3 and the number 4 will both satisfy it ; for
insertion of the first gives the identity 9 = 21-12, and inser-
tion of the second gives the identity 16 = 28-12; but no
other number whatever, when substituted, will result in an
identity, that is to say no other number will satisfy the
equation ; the equation has two, and only two, solutions, or,
as they are often called, "roots."
Looking at the factor form of the equation with which we
started, (x - 3) (a; - 4) = 0,
it is obvious that either 3 or 4 will satisfy it; because the value
3 makes the first factor zero, and the value 4 makes the second
factor zero. It is not necessary that both factors shall be zero
— either will do — hence the useful answer is not necessarily
both 3 and 4, but either 3 or 4, or possibly both.
The factor form of writing the equation, therefore, contains
the solution in so obvious a manner, that it is sometimes
spoken of as " the solution " ; and if an equation like
3a;2 + 7a;-31 = 11 -8a;
144 EASY MATHEMATICS. [chap.
were, by any process of manipulation, reduced to the form
(a;-2)(a; + 7) = 0
it would be considered solved ; because it is then obvious that
the values + 2 and - 7, that is to say either a; = 2 or a; = - 7,
or both, satisfy the equation. Inserting them successively,
for the purpose of verification, we get for the value x = 2
12 + 14-31 = 11-16
which is an identity ;
and for the value a; = - 7
147-49-31 = 11+56
which is another identity.
In collecting the terms of the given equation the two x
terms can be put together, making 15a;, and the two absolute
terms can be put together, making 42, but neither of these
pairs can be merged in the other, nor in the term 3^^ . there
are essentially three distinct kinds of terms in, the equation,
and they must be kept distinct.
Introduction to Quadratics.
When beginning quadratic equations, it is a good plan
to give them first of a kind that can easily be resolved into
simple factors, so as to remove the appearance of difficulty,
and yet to suggest a real method of solution.
For instance, a;^ - 7a; + 1 2 = 0
has roots 3 and 4, for these numbers add together to 7 and
multiply together to 12. So the expression on the left hand
side can be resolved into factors as
(x-S){x-4)
and the equation can be re-written
(a;-3)(a;-4) = 0.
Again a;2-5a;+6 = 0
is plainly satisfied by the values a; = 3 and a; = 2.
XIV.] QUADRATIC EQUATIONS. 145
Once more, a;^- 11a; + 30 = 0
has the roots 5 and 6, and is equivalent to
(x-5){x-6) - 0.
If we had chosen the equation
a;2 + lla; + 30 = 0
the roots would have been - 5 and - 6, and the equation
written in the factor form would have been
{x + 5){x + 6) = 0.
And so on, according to innumerable examples given in every
text book of algebra.
When a quadratic expression equated to 0 is solved, it is
always really resolved into two factors, for it is always
virtually expressed in the form
(x -a){x-h) = 0,
where a and h are the two numbers which satisfy the equa-
tion, its two " roots " as they are called ; a term which is thus
used in a new sense, having no reference to square or cube
root.
Multiplying out the above expression, it takes the form
x^-{a-{-h)x + db = 0
so that the coefficient of the middle term is the sum of the
roots, and the absolute term is their product; provided that
the coefficient of the quadratic term is unity, and the sign of
the middle term is negative.
The process of solving the equation is the same as that of
resolving the above expression into factors, and one way of
achieving it is to think of two numbers which add together
into the middle term and multiply together into the absolute
term, provided the coefficient of the quadratic term is unity.
Suppose, for instance, the equation given were a general
quadratic in a;,
Ax^ + Bx + C =(i.
146 EASY MATHEMATICS. [chap.
Divide everything by A in order to reduce the coefficient of
the quadratic term to unity, getting
A A
Here we know that the sum of the roots of the equation must
be equal to the ratio - BjA and that the product of the roots
must be equal to the ratio GjA. (See also Appendix III.)
In the above cases it was easy to guess the roots, but it is
by no means always easy. A process must be used for find-
ing them, and as far as possible the pupil should be left to
find it out: with guidance, but no more actual telling than
may be found necessary. But time and perseverance will be
required. If the child has no head for it the attempt may be
useless, and should not be persisted in unduly; nor should any
disgrace attach to failure; success is a triumph rather than
otherwise.
If the equation x^-\-\^x = 24 is given, it happens to be
rather obvious that 12 and 2 must be the numbers concerned,
if the signs are properly attended to ; but the rule for finding
them in general will have to be evolved from a consideration
of the chief quadratic identity,
{x + of = ic2 + lax + a2.
Suggest a trial of this to the pupil, and if necessary suggest
trying the value 5 for the auxiliary and gratuitously intro-
duced symbol a, because that will give
3;2+l0a: + 25,
which imitates the only hard part, the left-hand side, of the
given equation; for a little complication of the easy, the
numerical, part on the right-hand side, does not matter. So
we might write the given equation now if we like,
.iB2 + 10a;-t-25 = 24 + 25 = 49;
XIV.] QUADRATIC EQUATIONS. 147
but directly we have done that, the equation is practically
solved, for it is plainly equivalent to
(a; + 5)2= 72,
and therefore to a: + 5 = ± 7 ;
that is to say x = either 2 or - 12, as the case may be; for
either will satisfy or solve the equation.
Wherefore the given equation, with the roots wrapped up
fl;2 + l0a;-24 = 0,
may likewise be written (a; - 2) (a; + 12) = 0,
with the roots visible.
Another example or two to clinch the matter :
let it be given that a;2 + 14a; = 15,
here if we try to throw the left-hand side into the form
{x + a)2, the auxiliary number a is given by
2a = 14, so a^ = id;
and the equation becomes
a;2+14a; + 49 = 15 + 49 = 64,
or (a; + 7)2 = 82;
whence x = - 7 ± 8,
wherefore x = either 1 or - 15.
One more plainly numerical example :
a;2-6a; = 20,
here a is manifestly 3, and the equation becomes
a;2-6a;+9 = 29,
or (a;-3)2 = 29;
wherefore a; = 3 ± J(29)
and it can only be carried further by extracting the numerical
and incommensurable root.
Now a slightly more general touch :
given a;2-12a; = n.
To reduce this to the form (x - 6)2, we must add 36 to each
side, getting a;2 - 1 2a; + 36 = ?i + 36
whence the solution is {x-%) = ± J{n + 36)
148 EASY MATHEMATICS. [chap. xiv.
Finally let the given equation be
ic^ _ 2ax = 71^.
Complete the square on the left hand side,
x^-2ax + a^ = n^ + a^
it becomes {x - of == n'^ + a?
or a; = a ± J{n^ + a?) ;
which is essentially a general result.
The form of this result is easy to remember, and it is
really general ; for if the quadratic equation had been given
in the manifestly general form
Ax^ + Bx + C = 0,
where the coefficients A, B, C, stand for any known quantities
of any kind whatever, it can be reduced to the above form by
first dividing by A^ and then instituting comparisons between
it and the above ; for we then see that correspondence requires
the following identities :
71^ = _-- and 2a = --J,
B^
so that
a^ =
' 4.A^'
wherefore the solution of the
general quadratic is
B
^= -2A
-\iA^ a)
" 2A^
-B±J{B^-4AC)}.
But this should not be given to pupils for a long time yet,
and perhaps we have already been attracted a little further
than in the present book is legitimate. The pupil should by
no means be thus hurried. A month's practice at the numerical
and factor forms of expression may be desirable before passing
to even slightly more general forms.
CHAPTER XV.
Extraction of Simple Roots.
The last arithmetical lines of Chapter XIII. practically
asserted that 18 = 5832^; and it can easily be verified by
multiplication that
18x18x18 = 5832
or that 18 is the cube root of 5832.
Here then is a method, automatically suggested, for finding
cube and other roots : — Analyse the number into factors
whose roots are known, as 5832 was analysed into 27 and
216, at the end of the chapter referred to. It cannot always
or often be done, but whenever it can it is quite the best way.
But to be able to apply this method we must cultivate an
eye for factors, and we must also recognise or know by heart
a certain collection of cube and square numbers.
Thus 1728 = 12x12x12
or the cube root of 1728 is 12.
This is easy to remember because it represents the number
of cubic inches in a cubic foot. In a country with a purely
decimal system of measures, this fact would not be known
with the same ease. They would know well however that
1000 = 10x10x10
or that the cube root of a thousand is ten ; and so do we.
We may also know that 729 has 9 as its cube root, since it
evidently equals 81 x 9, that is 9^ x 9 or 9^. The cube root of
343 is 7.
150 EASY MATHEMATICS. [chap.
The square root of 10,000 is 100, and the square root of
this is 10 ; but what its cube root is is not so easy to say.
The cube root of 1000 is 10 ; but what its square root is
is not so easy to say.
The fifth root of 32 is 2, but neither its square root nor
cube root is simple.
It is valuable to remember thoroughly that 2 is the cube
root of 8.
The square root of a million is 1000,
the cube root of a million is 100 ;
the sixth root of a million is 10.
But there is no need to trouble about remembering any
more than a few ordinarily occurring square roots and cube
roots ; for the sixth and higher roots are seldom wanted, and
they can usually be derived from square and cube roots.
A number like 64 can resolve itself into 8x8
or into 4 x 16.
Its cube root is therefore easily stated as 2x2, viz. 4, and
its square root as 2 x 4, viz. 8.
144 again = 12x12 and also = 9x16,
and either pair of factors gives its square root but not its cube
root.
Surds.
Now let us proceed to ask what is the root of a number
like 12 [where the word "root" is used alone, square root is
understood]. We can resolve it into factors and find the
root of each
12 = 4x3,
so V12 = 74x^3 = 2x^3.
So the result may be stated that
VI 2 = 2^3.
Similarly JS = 2J2.
XV.] ROOTS. 151
Again, let us find the cube root of say 24.
24 = 8x3 and 4/8 = 2;
so 24^, which is often written 4/24, = 24/3.
Note the following :
732 = V16 J2 = 4^2,
•^'-V©=f=l-
8
Thus it would appear that 4^2 must equal -j-. If we
\/2
multiply or divide each of these numbers by J2 we can easily
verify this asserted quality. For multiplication by J 2 makes
them both 8 ; division by J2 makes them both 4.
Verify the following statements :
4/56 = 2^7.
v/20 = 2^5.
718 = 372.
V27 = 3^3.
772 = 3^8 = 6J2.
v/50 = 5^2.
V(200) = 1072.
^216 = 6J6.
^(360) = 6710.
V(810) = VIO.
V(490) = 7^10.
V(125) = 5^5.
V(iooo) = 10710.
V(1728) = 12^12.
7343 = 777.
V512 = 8^8 = 16^2.
The last five are all cube numbers, and their value suggests a
rule for expressing the square roots of any cube number ; e.g.
V27 = 3^3, or >3 = ^Jn.
152 EASY MATHEMATICS. [chai*.
N.B. — The best way of interpreting the word "verify" at
the head of the above set of examples, or in any similar place,
is for the pupil to take the left-hand expression, and try as an
exercise, independently, to simplify or otherwise express it,
and see if he can reduce it to the form given on the right-
hand side. He will thus perceive that numbers which have
two factors can have the expression for their roots put into
another form, which is often a more simple form ; and that
a large number of roots could be found numerically if the
roots of a few prime numbers were known.
The number ten, as usual, has an unfortunate disability, in
that neither of its factors is a perfect square, as one of the
factors of 12 is. All we can do with ^10 therefore is to
say that it equals J2 Jb, which is of extremely little use.
It is better kept as ^^10 and considered to be one of the
things to be found. Since the sq. root of 9 is 3, while 4 is
the sq. root of 16, the pupil may make a guess and try
whether 3"1 approximates to the sq. root of 10 or not. He
can easily do this by multiplying 3-1 by itself. By this
means he can gradually correct its value. He can in a
similar way make guesses also at ^20 and ^30 and ^^50. Let
him try.
No simplification, by resolution into factors, can be made
with any such numbers as
^/2, JS, A v/7, jn, J13, J17, 719, V23,
and so on; that is, no simplification of this kind can be
applied to any root of any prime number, naturally.
The roots of even numbers may always have a ^2 exhibited:
^6 may be written ^3 ^2,
JU „ „ J7J2,
but it is seldom useful to express them in this way.
XV.] ROOTS. 153
Let us see if we must draw a distinction between (^9)^ and'
^(9^), that is between the cube of root nine and the root of
nine cubed. Now
(^9)3 = 3^ = 27,
while J{9^) = V(729) = 9J9 = 27 likewise..
So they turn out to be the same.
The cube of a root appears to be equal to the root of a cube-.
That is curious, and may well be unexpected. It is not the
sort of thing at all safe to assume. Plausible assumptions
are always to be mistrusted and critically examined ; occasion-
ally, as in this instance, they turn out true.
Let us consider the fact more generally, and see whether it
is always true that
J{n') = Unf.
The other and more expressive notation for roots will here
come to our aid.
J(n^) may be written (n^)^,
and (Jn)^ may be written (71^)^,
so it looks as if both could be written as n^ or n^^, or, more
properly, ?i^ + ^.
This last is a thing we have not yet learnt how to interpret.
We may assume however, as an experimental fact, that
J(n^) = Jn^sjn = njn,
hence the interpretation nJn, that is ?i x ?i^, suggests itself
for n}^ or ?i^ + ^; and it is the right interpretation.
Here again (as on page 126) we have arrived at a striking
circumstance about indices, which is now well worthy of
examination.
CHAPTER XVI.
Further consideration of indices.
There are two things to which we might now appropriately
turn our attention : one is the numerical calculation of all
manner of roots, for instance, ^2, ^3, ^10, 4/2, ^3, etc., IJ2^
yXOOy and so on; evidently a large subject, since we may
require to find any root of any number; the other is the
discussion of that curious property of indices, which has been
dimly suggested by certain of the examples chosen, viz. the
suggestion that
and that (a;**)'" = («*")" = a;^».
Of these two directions along which we could now continue
the discussion, the latter is undoubtedly the easier, and so we
will proceed this way first; and incidentally we shall find
ourselves led to a very practical and grown-up way of dealing
with the former more difficult line of advance.
What we found experimentally (on page 1 26) was that
3x32 = 27 = 33;
also that 32 x 3^ = 9 x 27 = 243 = 3^.
And so we might have taken other instances :
22x24 = 4x16 = 64 = 2^,
2 X 25 = 2 X 32 = 26,
23 X 23 = 8 X 8 = 26,
2 X 22 X 23 = 2 X 4 X 8 = 26.
CHAP, XVI.] HORSE-SHOE-NAILS PROBLEM. 155
What does all this look like 1
Manifestly it looks as if to effect a product among the
powers of a given number, we must add the indices of the
several powers. It looks like
23 X 24 = 27,
23 X 28 = 211,
68 X 62 = 65,
2i X 22 = 2^ = (^2)5 = ^(26) = ^32 = 4^2,
2* X 2^ = 2*^ = 2^ = ^(29) = (72)9,
= V(51 2) = 7(2x256) = 16^2.
Now when the indices are whole numbers it is very easy to
see the reason of this simple rule. What does 2^ mean 1 It
means that four factors each of them 2 are to be multiplied
together. The index is only an indication of how many times
the similar multiplication is to be performed.
2^ means simply 2x2x2x2; the number of multiplication
signs being one less than the index, i.e. one less than the
number of factors of course. Similarly 2^ is merely an
abbreviation for 2x2x2. Hence
2^x23 = 2x2x2x2x2x2x2,
that is seven 2's are to be multiplied together; and so it is
naturally indicated by 2''.
The index counts the number of similar factors; hence
when the factors are increased in number the index shows
the simple increase; but the effect of the continued multi-
plication on the resulting number may be prodigious.
The anecdote about the nails in the horse's shoes here
appropriately comes in :
A man, who objected to the price asked for a horse, was
offered the horse as a free gift, thrown into the bargain, if he
would buy merely the nails in its shoes, of which there were
6 in each foot; at the price of a farthing for the first nail,
156 EASY MATHEMATICS. [ghap.
2 farthings for the second, 4 farthings for the third, 8 for the
fourth, and so on. The offer being accepted, he had to pay
.£17,476 5s. 3|d. for the nails; and he did not consider the
horse cheap.
The number of farthings in this sum is very great, but it is
simply one less than 22^.
If a beginner wishes to verify the above by multiplying 2
by itself 23 times, he can easily do it, though it will take a
little time; and he can then reduce the result to pounds
shillings and pence, as he has been no doubt so well taught
how to do. It is not a grown-up way of ascertaining 22-*, but
it serves. (Reference to pp. 166 and 259 may be convenient.)
If he is properly sceptical about the magnitude and correct-
ness of the above sum, he should do it. It is good practice in
easy multiplication ; and sums which are set by the pupil to
himself are likely to secure greater attention from him than
those enforced from outside. It is probably desirable that
children should often set sums as well as work at them. I
would even sometimes encourage them to set examination
papers. It is a good way of getting behind the scenes.
As regards the verification of a"* x a" = a"*"^" therefore, the
idea is very simple, so long as m and n are whole numbers ;
because it is a mere matter of counting the number of similar
factors.
"When we say that five sixes multiplied together equal 7776,
we are employing the number five in this very way. The
expression 6^ does not mean five sixes added together, or 30 ;
but it means five sixes multiplied together, yielding a much
larger result.
So also six tens multiplied together make a million, whereas
added they only make 60. In fact, as we said before, page 56
and Chap. XII., while multiplication is abbreviated addition,
involution is abbreviated multiplication.
XVI.] ^^,>><^ INDICES. 157
Fractional indices.
When the indices m and n are fractions, the idea they express
is not so simple, and the above relation a"* x a" = ft*"^" is not
so easily justified; but we may be willing to accept it by
analogy and see how it works.
If asked wherein the proof consists for fractional indices,
we must answer in " consistency," constant coherence and
agreement with results so obtained, and in corresponding
convenience of manipulation.
At one time 2^ and such like were called irrational quan-
tities because it was difficult to attach a commonsense signifi-
cance to "2 multiplied by itself half a time"; and it is
certainly not to be interpreted as half 2 multiplied by itself,
for that would be unity.
There is nothing irrational about this quantity however :
it has a value approximately 1*4142 ... though it will hereafter
be found that it will not express itself exactly by a finite
series of digits in any system of notation whatever.
It may rightly be styled "incommensurable" therefore,
but it is in no sense irrational.
"Irrational" however was a term at one time applied to
any power of a number whose index was not a positive integer.
The thing has to be mentioned, for historical reasons, but
the term "irrational" should now cease to be used. The term
" surd," being meaningless, may be employed if we like, but it
is never really wanted : it only serves as a heading to a
chapter to indicate its contents.
Negative indices.
But let us go on and ask how shall we interpret the ex-
pression if one of the indices be not fractional but negative ?
For instance, how shall we interpret
3-2 or 2-6.
158 EASY MATHEMATICS. [chap.
Suppose for instance we had
or X a-",
we should naturally say that the result must be a*""**.
Very well, let m = 2 and n = 3,
then 0? X a~^ = a~^.
What does fl~^ mean ?
How can we multiply a number by itself a negative number
of times ? At first the term " irrational " was applied to such
quantities as these : but a consistent interpretation was soon
found for them. If addition of indices means multiplication,
it is natural that subtraction of indices shall mean division.
Make the hypothesis therefore that a*""** can be interpreted
as a"" -f a", and let us see how that works.
Suppose we had 2^ ~ 2^, it could be written out in full ,
2x2x2x2x2
2x2x2 '
and the result after cancelling would be 2 x 2,
that is 22 = 25-3.
The whole thing is therefore quite simple.
Take other examples :
3(6-5)
=
3«
3^^ ~
3,
a'-*
=
a'
a' ~
a\
^«-.
=
x^ x" '
7 X T''
74-2+1 = 1^ = 73 = 343^
XVI.] INDICES. 159
7* 1
na-2 _ _ _ _!_7«
' - 72 - 49 ' '
7*
This last is a most interesting and useful result.
If the index is zero, the quantity, whatever it may be, is
reduced to unity ; for
a— = ^ = 1 ;
it equals 1 whatever a may be.
a<^ = 1 is the brief summary of this important consequence
of our notation. The index 0 would have been hard to
interpret, just as fractional and negative indices were hard to
interpret, but fortunately it thus interprets itself.
A negative sign applied to an index turns out therefore to
have the effect of giving the reciprocal of the quantity;
for since „»
we have only to take the case where m is zero, in order to get
-n _t - L
Hence 2~'^ =
2-2 =
3-2 =
3-3 =
2-6 =
1
2'
1
?
1
9'
27'
J_
32'
160 EASY MATHEMATICS. [chap.
Hence while 2^ means ^2, we thus find that 2"^ means ^ ;
or, in general, i . , 1
X = — and x~^ = —
«" X
Take the last simple and useful mode of expression. To
verify it, simply multiply both sides by x, thus
a;-i X a;+i = a;-i+i = a;0 = - = 1
X
Similarly x-^ = i ^ g)' = {x-%
And this suggests powers of powers; like ( 10^)2, that is the
square of a thousand, which is a million, 1 followed by 6 ciphers,
or 106.
So also (10^)2 = 10^2 = a billion; the indices being in this
case multiplied to give the result.
So now we leave addition and subtraction among indices,
which merely meant multiplication and division among the
quantities themselves, and begin to study multiplication among
indices.
Consider for instance what the meaning should be of 4^^^ ;
it equals 4^ = (43)2 = (42)3 = 4096,
So multiplication among indices means involution among
the quantities themselves.
So also division among indices will signify evolution among
quantities, thus
7t = (73)^ = (7^)3,
the order of the factors (which in this case are m and - j
being indifferent. ^
XVI.] INDICES. 161
If it were worth while we might proceed further, and con-
sider what would be the meaning of the process " involution "
applied to indices; how would that affect the quantities
themselves ? What for instance is the meaning of 2^ 1 but it
is a mere curiosity and is hardly worth while. Suffice it to
say that the numbers so reached become rapidly prodigious.
10^0 is a number with ten ciphers after the 1, or ten thousand
million ; but 10^^ possesses a hundred cyphers, and represents
a number far greater than that of all tTie atoms of matter in
the whole solar system — earth, sun, and all the planets, — not-
withstanding the fact that a speck containing a million-million
atoms is only visible in a high power microscope.
CHAPTER XVII.
Introduction to Logarithms.
The equation y = a;", that is, the ti*^ power of «, may be
equally expressed as a; = yn^ that is, the v>'^ root of y ; this is
not an inverse expression, but the same in inverse form.
So also the equation xy = \, which represents a multiplication
sum, can also be written y=l/Xj which represents a division
sum; and x^y^ ■■= c^ can appear as c = ±xy, the double sign
representing an ambiguity or double solution, because either
+ c or -c would when squared give the right result.
If y is the n}^ power of x, it is easy to say that x is the rt*^
root oi y ; we can also say that n is the index or exponent of
X which yields the value y ; but how are we to express the
relation that n bears to y1
It is a thing we have not yet come across.
It is called a logarithm; it involves a reference to both
X and y ; it is called the logarithm of y to the base x.
Let us understand this matter.
Write down 100 = IO2,
2 is called the logarithm of a hundred to the base ten.
Conversely 10 = 100* so J might be called the logarithm
of ten to the base a hundred.
Write down 25 = 52
2 is the logarithm of 25 to the base 5.
CHAP. XVII.] LOGARITHMS. 163
The logaritlim of a number is defined as the index of the
power to which the base must be raised in order to equal
the given number.
Thus if we are told that 3 is the logarithm of a thousand to
the base ten, it is another mode of stating that 10^ = 1000.
So 3 is the logarithm of 8 to the base 2,
2 is the logarithm of 49 to the base 7,
5 is the logarithm of 32 to the base 2,
3 is the logarithm of 216 to the base 6,
4 is the logarithm of 81 to the base 3,
and so on.
It looks a cumbrous and roundabout mode of expressing
what is more neatly expressed by the index notation, but it is
an exceedingly practical and convenient mode of statement all
the same, and is a great help in practical computation.
What is the logarithm of 343 to the base 7 ? Answer, 3.
What is the logarithm of a million to the base 10 ? Answer, 6.
What is the logarithm of 64 *? It is 6 to the base 2, and 2 to
the base 8.
What is it to the base 10? Answer, something less than 2
and more than 1.
What is the logarithm of 10 to the base 10, or of any
number to its own base ? Answer, unity, for a = a^.
What is the logarithm of unity itself ?
The answer is 0,to any base, because 1 ^a".
What is the logarithm of a fraction, say J, to the base 2 ?
Answer, a negative quantity, in this instance - 2, because
J = 2-2.
So also - 2 is the logarithm of y^^j- to the base 10, because
yJ^ = 10-2. And the fact that ^^ = 10-^ can be expressed
by saying that - 6 is the log of a millionth to the base 10.
It appears therefore that the logarithms of reciprocals or of
numbers less than 1 are negative, the log of 1 itself being 0.
164 EASY MATHEMATICS. [chap. xvii.
This is satisfactory. Everything greater than 1 has a positive
logarithm, everything less than 1 has a negative logarithm;
provided always that the base itself is greater than one.
The further a number is removed from 1 both ways, whether
in the direction of greatness or of smallness, the larger
numerically is the logarithm ; but it is positive bigness in the
one case, negative bigness in the other. It is natural there-
fore that the logarithm of 1, to any base, should be zero.
Mathematicians know how to calculate the log of any
number, no matter how complicated, and they have recorded
the results in a book called a table of logarithms; just as
grammarians and scholars know how to translate any foreign
word, and have recorded the results in books called dic-
tionaries. A Table of Logarithms is to be used like a
dictionary. It can be readily used, and is used every day,
by those who would find it difficult to construct it. It should
puzzle children sometimes how the meaning of words in dead
foreign languages were ascertained ; they mostly take it for
granted and do not think about it. So also, for a time, and
until they make some approach to becoming budding mathe-
maticians, they need not learn how to compute a table of
logarithms; but they must imbibe a clear idea as to their
meaning. They must also, and that is an easier matter still,
learn their practical use, and be able to use a table as they
have learnt how to use a dictionary.
CHAPTER XVIII.
Logarithms.
When we express a number thus :
64 = 82,
1000 = 108
32 = 25,
or, in general, n = a*,
we are said to express it " exponentially," that is, by means
of the index or " exponent " of the power to which a certain
other number called a base is to be raised in order to be equal
to the given number.
In the above equation n stands for the number, a for the
base, and x for the index or exponent of that base.
The question naturally arises, what relation does x bear to
n, for it manifestly depends upon both n and a 1 If the base
has been specified and kept constant, then x will vary only
as n varies. It is plain that x will increase as n increases, but
not nearly so fast.
Take a few examples, and first take the number 2 as base :
2 = 21,
4 = 22,
8 = 23,
16 = 24,
32 = 25
64 = 26,
166 EASY MATHEMATICS. [chap.
1024 = 210,
16,777,216 = 224.
Here the index runs up slowly, 1, 2, 3, 4, etc., according to
what are called the " natural numbers " ; whereas the number
on the left-hand side runs up very quickly. The index is said
to progress "arithmetically," that is, by equal additions; the
number on the other hand is said to progress " geometrically "
(a curious use of the word), that is, by equal multiplications.
There is evidently some law connecting the index and the
number, when a base is given; and the following nomenclature
is adopted :
5 is called the logarithm of 32 to the base 2 ;
3 is the logarithm of 8 to the base 2 ;
4 is the log of 16 to base 2 ;
6 = log 64 (base 2),
which is usually abbreviated still further :
6 = log2 64 ;
10 = logg 1024,
the base being indicated as a small suffix to the word log.
Make now a more complete table ; first of powers :
1 = 2°,
2 = 21, - - - 1
4 = 22.
8 = 23
16 = 24,
32 = 25
i = 2-2,
i = 2-3,
1 _ 0-4
T¥ - ^ '
1 ' _ 0-5
^2" - '^ '
1024 = 210, . . _^_ = 2-10,
and then of the corresponding logarithms :
XVIII.] LOGARITHMS. 167
From the above table it follows that (with the base 2)
logl = 0,
log 2 = 1, logi= -1,
log4 = 2, logi= -2,
logS = 3, logl = -3,
log 16 = 4, logTV= -4,
log 32 = 5, log^V= -5.
It would be a good thing to plot both these tables on
squared paper, representing for the first the indices 1, 2, 3, 4
as horizontal distances, and the numbers 2, 4, 8, 16 as vertical
distances; and for the second measuring distances to repre-
sent the 2, 4, 8, 16 numbers horizontally, and the logarithm
numbers 1, 2, 3, 4 vertically.
The first is called an exponential curve, or curve of exponents
or indices ; the second is called a logarithmic curve, or curve of
logarithms. The two curves turn out to be identically the same,
only differently regarded, — to make their identity apparent,
the paper can be turned round and looked through at the light.
If drawn on the same sort of squared paper the curves
will fit. They may either of them be said to represent the
relation between Geometrical and Arithmetical progression :
in one direction distances proceed arithmetically, or by equal
differences ; in the other geometrically, or by equal factors.
These curves will do for any base, if their scale is suitably
interpreted. The divisions we have labelled 2, 4, 8, etc., may
equally well be considered to represent 3, 9, 27, etc., or
a, a% a^, etc., or 10, 100, 1000, etc.
That is the advantage of a curve. Once drawn, it represents
to the eye a general kind of relationship ; and nothing but an
interpretation of its scale is necessary to make it fit any
required instance of that relationship. The shape of the above
suggested curve is drawn on pages 101 and 179.
168 EASY MATHEMATICS. [CHAf.
Verify the following statements :
3 = log 27 to the base 3,
4 = log3 81,
6 = logs 729,
.2 = log, 16,
4 = log, 256,
5 = log, 1024,
2 = logs 36,
3 = loge216,
3 = logy 343,
3 = log, 729,
2 = log, 81,
but 9 ^ log2 81,
2 = log,2 144,
3 = logi2 1728.
This last seems a curious and roundabout way of expressing
the fact that 12x12x12 = 1728, and if it did not turn out
practically very convenient there would be no justification for
introducing such a complication as the logarithmic notation
instead of the index notation; but it is constantly to be
noticed, when a new notation has been introduced into mathe-
matics, that it confers on us an extraordinary power of
progress, and enables difficulties further on to be dealt with
which were before intractable.
Any complication which is of no use — or let us say of no
obvious and well-known use — anything which should not be
familiar to every educated person — is not treated of in this
book ; the justification of any notation is that though for the
expression of simple and already well-known facts it may look
cumbrous and inexpressive, yet when we want to express
harder and at present unknown ideas it becomes helpful and
luminous.
I
xviii.] LOGARITHMS. l6&
Common practical base.
The case when the logarithmic base is the same as the
base adopted for our system of numerical notation, is worth
special attention, because it is the one most frequently used
in practice. What the base for numerical notation may be, is,
as we know, a pure convention ; and, as we have explained, it
is perhaps an unfortunate but now irremediable convention
that the base of notation is ten. It does not follow that the
logarithmic base must also be ten : it is perhaps possible to
find a natural base, involving no convention. If so, such a
base would of course be important and interesting; but
meanwhile we will take ten as the base also of a practical
system of logarithms.
Let us first make a table of powers of ten.
1 = 100,
10 = 101, _i^ = 10-1,
100 = 102, ^ ^ 10-2,
1000 = 103, ^V^ = 10-3,
1,000,000 = 10^ ..ooi.ooo = 10-^
Whence it follows that (with base 10)
to ^^l logl = 0,
log 10 = 1
log 100 = 2
log 1000 = 3
log 1,000,000 = 6
log-rV = log-1 = -1,
logTk = log-01 = -2,
logToVo =log-001 = -3,
log,.ooi.ooQ = log -000001 = -6,
Hence (with base 10) the logarithms of numbers between
10 and 100 lie between 1 and 2, that is to say consist of 1
and a fraction : the log of 1 1 will be 1 and a small fraction,
170 EASY MATHEMATICS. [chap.
the log of 99 will be 1 and a large fraction — very near to 2 in
fact. Consequently with all double-digit numbers the char-
acteristic property of the logarithm is that it begins with 1.
All numbers which consist of three figures lie between 100
and 1000, and these have the characteristic 2 ; that is to say
they all consist of 2 + a fraction. This is true even of such a
number as 999-99, provided the 9's are not repeated for ever;
because although the log of such a number is very nearly 3,
it is not quite 3 until 1000 is reached.
A number consisting of five digits will have a log whose
characteristic is 4, and so on; the characteristic is always
equal to the number of digits on the left of the unit digit,
which is taken as a zero of reckoning. Thus the characteristic
of the log of any of the following numbers (1200, 1728, 5760,
9898, 1431-8, 1696-25) is 3.
The logarithm of every fraction between 0 and '1 will be a
negative fraction : it will not be quite equal to - 1, but it may
be put equal to - 1 plus a positive fraction.
The logarithm of every number between *! and '01 will lie
between - 1 and - 2, and therefore may be expressed either
as — 1 minus a fraction, or as - 2 plus a fraction ; and the
latter is the usual plan.
The rule for the characteristic therefore is to count always
to the first, i.e. the most important significant figure, starting
from the units place as zero. On this method of expression
it is easy to write down the characteristic of the logarithm of
any number at sight.
The best plan is to employ the term " order," connoting by
the (yrder of the number the index of the power to which the
base say ten must be raised in order to give a number with
that number of digits. E.g. the order of 100 is 2, because it
equals 10^, and all the numbers 121, 256, 780, 900 may be
technically designated as of the same "order"; because.
XVIII.] LOGARITHMS. 171
though greater than 10^, they are less than 10-^ ; and the
amount by which they exceed 10- is shown by the fractional
part of the logarithm, not by its integer part or characteristic.
But 1000 is of the order 3, and so likewise is 1728, etc.
17 is of the order 1, and so is 14-58,
4 is of the order 0, and so is 4*6,
•3 is of the order - 1, and so are -35 and -78,
•02 and -035 and -016 are of the order - 2.
Accepting this nomenclature, which is useful in quite rudi-
mentary arithmetic, e.g. in long division and the like, we are
able to say simply that the characteristic of a logarithm is the
" order " of its number.
Let there be no confusion between the table on page 169
and the one on page 167 to base 2. They involve different
bases ; and though the base is not expressed every time, but
only in the heading, that is merely because of the needless
trouble of frequently printing or writing suffixes, like this :
3 = log2 8 = logs 216 =logio 1000.
Examples.
The characteristic of the logarithm of the following numbers
(also called the " order " of the number itself) is as here given:
The logarithm, to base 10, of each of the numbers
5, 8-7, 1-23, 9-99, Mil has the characteristic 0 ;
and this is the '* order " of each number.
Of each of the numbers
11 1 the order is 1
of 300, 981-4, 101-01, it is 2
it is - 1
it is -2
it is - 3
17,
94,
17-65,
300,
981-4,
101-01,
•17,
-94,
-11101,
-08,
•01,
•0999,
•002,
-0056,
-009846,
108-001,
17-9909,
5-0652,
8 X 106,
1-0056,
10-001,
256000,
•0000256.
ly^ Easy mathematics. [chap.
Examples for Practice.
Write down the characteristic of the logarithm of each of
the following numbers, (in other words express the " order " of
each number) :
56, 108, 56-75,
8-3, 8300, 5065-2,
-56, -0056, -008309,
99-9, 9-9, -099,
Fundamental relations.
There are a few fundamental properties appropriate to
logarithms belonging to any base whatever.
One of them is that log 1=0,
and another is that log (base) = 1,
but there are others which we have already several times
hinted at.
Let us recollect once more what an index or exponent
signifies. It signifies the number of similar factors which
have to be multiplied together. For instance,
25 = 2x2x2x2x2,
64 = 6x6x6x6,
or, in general, a^ = axaxa,
a** = axaxax... to n factors.
So now if we write a number N equal to any of these, as for
instance N = a^,
the index, or exponent, which we shall now call the logarithm
of i\r to the base a, simply counts the number of times the
base occurs as a factor in the number N.
N = axax.a.
Suppose now we took some number not quite so easy to
deal with as those in the examples we have hitherto considered.
XVIII.] LOGARITHMS. 173
a number which cannot be represented as any simple power
of any integer, say for instance the number 30 ; and ask what
will the logarithm of 30 be to the base 10.
First of all we see that it must be between 1 and 2, because
1 is the logarithm of 10, while 2 is the logarithm of 100; and
the logarithm increases with the number, but arithmetically
instead of geometrically. So as 30 lies roughly half way
geometrically between 10 and 100, it may be expected that
its logarithm will be somewhere about halfway arithmetically
between 1 and 2. It will be 1 + a fraction ; and what that
fraction is can be approximated to more or less closely by
examining and measuring the logarithmic curve which we
ought to have carefully drawn, as indicated on p. 179, and
specially labelled so as to suit the base 10. Measuring that
curve for the logarithm of 30, it suggests a value something
like IJ or 1'5. This would be the result to "two significant
figures," but if the curve has been carefully drawn, it might
give us 3-figure accuracy, that is, would enable us to express
the result correctly to 3 significant figures ; in that case we
might estimate log 30 to be about 1 48.
The number which really lies geometrically half-way
between 10 and 100 would be V(IOOO), since 10:n/1000 =
a/IOOO : 100; and ^(1000) is accordingly called the geometrical
mean of ten and 100. Hence the logarithm of VlOOO is
exactly 1*5 or 1^. Similarly the logarithm of mJIO is '5 or J.
It is a curious thing that though we do not yet know how to
calculate the root of 10, we know its logarithm; and this
suggests — what frequently happens — that the logarithm of
the result of an arithmetical operation is easier to perceive
than the result itself.
We can examine the curve again, to see if it will show
us what number has a logarithm exactly 0*5; we shall see
that it indicates something like 3-1 or 3*2, and if it were
iT4 EASY MATHEMATICS, [chap.
drawn carefully it might indicate 3*16. This is one of the
values that we ought already to have arrived at by trial and
error, as recommended on p. 152 ; taking different numbers
between 3'1 and 3 "2 and squaring them, to see how nearly
the square would approach 10.
No number that we can select will, when squared, exactly
equal ten. It has no square root that can be expressed
numerically with exactness. Nor has any number except the
square numbers, 1, 4, 9, 16, 25, etc. These numbers, i.e. the
group ordinarily denoted by these symbols, are square
numbers in any system of notation, their square roots can be
numerically expressed precisely ;* and for no other numbers
can the same be done, however many fractions or combina-
tions of fractions, or however many decimal places, are
employed. Nor can it be done in any other system of
notation. In other words,
x/2, V3, VS, J7, n/1000, etc., etc.,
are all incommensurable.
But their logarithms are easily expressed, to any base
whatever, in terms of the logarithm of the number itself to
the same base. Thus
log V2 = J log 2,
log V3 = J logs,
log x/5 = J log 5,
logVlO = JloglO.
* Caution. — It is not intended, and it is not true, that the above digits
express square numbers when interpreted in accordance with any scale
of notation ; for instance, the amount of money represented by 2/5 is
not a square number of pennies, but the number we are accustomed
to designate by 25 is a square number, and 25 coins can be easily
arranged to form a square.
XVIII.] LOGARITHMS. 175
Hence to the base 10,
logViO = J log 10 = 0-5,
logVlOO = J log 100 = 1-0,
logv/1000 = J log 1000 = 1-5,
logv/10000 = i"iog 10000 = 2-0,
logVlOOOOO = J log 100000 = 2-5.
Similarly we may guess that
log^lO = -J log 10 = -3333...,
and so we may refer to the curve and see what number has
the logarithm J, for that will be the cube root of ten. We
find that it is 2-1544347, or approximately 2*14; and if we
multiply this by itself three times, 2'14x2-14x2-14, we shall
get a number not far off ten — a trifle greater than ten.
Similarly 21*4 will be approximately the cube root of ten
thousand, and 214 of ten million.
No exact numerical specification of the cube roots of any
number can be given, except of the cube numbers, that is,
those numbers which, given in the form say of marbles, can
be built up to represent cubes ; namely such numbers as
1, 8, 27, 64, 125, and so on.
For the cube root of any other number, if it could be
expressed, would be a fraction ; and a fraction multiplied by
itself necessarily remains a fraction; it can never yield an
integer. You cannot fractionate a fraction into a whole.
This remark is further developed in Chapter XX.
We now know how to find the logarithm, to the base ten,
of any power of ten, whether integral, negative, or fractional.
Examples :
log 103 = 3^ log 10-3 = - 3, log 10* = J ;
log 104 = 4^ log 10-4 = - 4, log 10* = J ;
so generally log^^lO* = a;,
and this may be easily generalised so as to apply to any base.
176 EASY MATHEMATICS. [chap, xviii.
For log ft* = X log ft,
but we know that log (base) = 1,
so we see that the logarithm of any power of the base is equal
to the index or exponent of the power, or
log (base)* = x.
We have thus arrived at the original definition of a
logarithm from which we started, — having reasoned *'in a
circle."
The advantage of reasoning in a circle is that we thereby
check and verify to some extent the intermediate steps, for if
any of them had been inconsistent we could not have worked
round to our starting point ; unless indeed we had happened
to make a pair of errors which cancelled each other : a thing
which is sometimes done — especially when the conclusion is
consciously in our minds. Working round a circle of reasoning
is in that case no adequate check. It is not possible to get
round by any odd number of errors, but with an even number
of errors it is possible though not very probable; unless
indeed we know our destination too well beforehand.
The real test of truth is that it shall turn out to be
consistent with everything else which we know to be true^
No one chain of reasoning, however apparently cogent, is to
be absolutely trusted, — for there is always the danger of
oversight due to defective knowledge. Complete consistency
i-s the ultimate test of truth ; and convergence of a number of
definite lines of reasoning is an admirable practical test.
CHAPTER XIX.
Further details about logarithms.
Involution and evolution become easy directly we employ
logarithms :
To obtain any root, say the r*^ root, of any number :
Find the logarithm of the given number to any base,
calculate -th of this logarithm, then find the number which
r
has this value for its logarithm to the same base; that
number is the ?'*^ root of the given number.
Or put it thus : utilising the logarithmic curve, page 179.
Take a length on the horizontal line as representing the
given number; find its logarithm, as the vertical distance to
the curve at this point ; calculate -th of this length, and find
on the curve a point whose vertical height is equal to it;
then the foot of the perpendicular from this point marks out
on the horizontal line a length which represents the Hh root
of the given number, on the same scale as the number itself
was represented. Thus, for instance, \y6. of the height of the
curve at division 8, projected back horizontally, should meet
the curve above the division 2 ; because 2 is the cube root of 8.
You see it is worth while to draw the curve neatly and
carefully, so that fairly correct measurements may be made
upon it. Besides, accurate drawing is a useful art, and it
takes a little time to employ drawing instruments accurately
so as to make no blots or smudges, and to get all lines
uniformly thick and accurately passing through the points
L.E.M. M
178 EASY MATHEMATICS. [chap.
intended. It is an art worthy of cultivation for future use.
Much information can be gained from such curves, not only
in science but even in business and in politics.
It may be said that by this process of drawing and
measuring, a logarithm or a root can after all only be attained
approximately. Yes, but the same is true of any process,
so far as accurate expression is concerned. A logarithm or
a root in general requires an infinite series of digits to express
it ; all finite expression is approximate.
I do not however say that a mathematician would calculate
logarithms or roots by such a curve: he would know plenty of
other contrivances for such things, and perhaps we may know
some of them later on ; but he would not despise the curve
method, at least in more really difficult investigations. He
would use it frequently. But for mere logarithms he would
use a table, somewhat as indicated in Chapters XXIX. and XXX.
Now let us see if we can calculate a few other logarithms.
We can obtain any we want from the curve, but if we could
obtain a few once for all, and label them, and then be able
to express the logarithms of other numbers in terms of these,
it might save us time and trouble ; and besides it is a de-
sirable and useful thing to be able to do.
We have managed to find the logarithm of any power of
ten (p. 175), let us see if we can manage the logarithm of any
product containing ten as one of its factors.
We have indeed already tried one of them, viz. 30 (see
p. 173); let us try another, say 50, likewise to the base 10.
We want to find to what power 10 must be raised in order
to equal 50.
Let X be the power, then
50 = 10*
is an equation from which we have to find x.
This is only another mode of stating that x = log^Q 50.
X
■ - - J_
it J"
2 J:
Ui
^
h- ^ '
t x^
T °' 9 a
t >< T3^
4 M' -i-T
1 op^-C
" " "I^X
gi_|:X
1 L X n'
t 3 5::t:
V r ir-s
I f- 4a XX
4 :s T3^
1 ^ i:r
ii _g ^^
+ CO _o j:_|_
\ 0 T '52
4 B -fcX
J S ^ 5
JL « _x It -^
L C ^3^
t C XX
4 2 ill:
4 S S— ii
JL * _j5 js-4:
C ^ ^X
V IS 4fl^i
A 19 g^T
4 -r r
" ^ S-
^ is"ir
4 i| x
V- 1 -X-
4t l-i:
1 <N 2r cj-
r
5
V
^
;s^
5
s
^
^^=^^
y "6 in| g~ in o in o
oj- ^ 6 ^ - pj
XIX.] . LOGARITHMS. 181
But now resolve 50 into two factors, and write
5 X 10 = 10%
10*
then 5=i^=W-^;
hence x-1 = log 5,
or a; = 1 +log5 = 1-7 about (by the curve).
Thus log 50 = log 5 + log 10,
which is a special case of a general assertion that
log nm = log n + log m.
Examine this :
Let n = a% so that x = log w,
and let m = a^, so that y = log m ;
then nm = a*o? = (f^^-,
.'. log nm = %-\-y = log n + log m.
Hence by using logarithms, multiplication is turned back
into addition, just as involution was turned back into multi-
plication. So also division is turned into subtraction, just as
evolution was turned into division.
The fundamental relations are as follows ; and although we
have stated them several times before, they are supremely
important and will bear repetition.
Let a* = n and a" = m,
so that X = logn and y = logm,
then nm = a*ft* = a"'^" ;
.*. lognm = x + y = logn + logm.
Furthermore, — = — = a*-*;
.*. log£ = o^-y = logn-logm.
Moreover, n"^ = {ay = a^ -,
.'. logTi"* = xy = mlogw.
182 EASY MATHEMATICS. , [chap.
Likewise log %"'" = - m log w,
„ log 71^ = log'^n = -log 71.
Apply these ideas. We can write at once that to the base 10
log 5000 = log 5 + log 1000 = 3 + log 5,
log 500 = log 5 + log 100= 2 + log 5,
log 50 = log 5 + log 10= l+log5,
log 5 = log 5 + log 1 = 0 + log5,
log ^jj = log -5 = log 5 - log 10 = - 1 + log 5,
log -05 = log 5 - log 100 = - 2 + log 5,
log -005 = log 5 -log 1000 = -3 + log5.
If we know log 5 therefore we should know the logarithm
of five times any power of ten, or even of five times any root
of ten ; for
log 5^10 = log 5 + 1 log 10 = -5 + log 5,
log 571000 = logs 4-i log 1000 =-- 1-5 + log 5,
log vfo = Iog5-ilogl0 =--5 + log5,
log 5^10 = Iog5 + Jlogl0 = •3 + log5;
but this is perhaps hardly worth stating.
How are we to find log 5? We can, if we choose, express
it by means of log 2, thus :
log5 = log-V- = Iogl0-log2 = 1 -log2,
or log 2 + log 5 = 1.
Similarly log 20 + log 50 = 3,
log 20 + log 5 = 2.
So also log 2 + log 6 = log 12,
log 3 + log 4 = log 12,
log 7 + log 9 = log 63,
log 8 + log 8 = log 64,
log 9 + log 9 = log 81,
log 17 + log 13 = log 221,
XIX.] LOGARITHMS. 183
log 6 - log 2 = log 3,
log 9 - log 3 = log 3,
log 4 -log 3 = log l'^,
log 5 -log 2 = log 1-5,
log 5 - log 3 = log 1 '6,
log 7 - log 5 = log 1 -4,
log 9+ log 16 = log 144,
2 log 12 = log 144,
3 log 12 = log 1728,
ilogl2 = logV12 = log 2^3 = log2 + logV3
= log2 + Jlog3,
ilogl6 = log4,
Jlog 8 = log2,
-1- log 49 = log 7,
ilog25 = log5,
Jlog72 = -Uog(36x2) = log6H-Jlog2
= Jlog( 9x8) = log3 + log2V2
= log 3 + log 2 + J log 2.
This might have been set as an exercise. Prove that
I log 72 = log 3 + 1-5 log 2.
One way to prove it would be to double both sides,
log 72 = 2log3 4-31og2
= log 32 + log 23
= log 9 + log 8
= log(9x8)
= log 72. Q.E.D.
Exercises. — Verify, by means of the curve in this chapter,
the following approximate statements,
log 2 = -3, log 4 = -6, log 8 = -9,
710 = 3-16..., 710 = 2-15.,.,
711-6 = 3-4..., 7115 = 4-86...,
77 = 2-6..., 7841 = 29-0.
CHAPTER XX.
On incommensurables and on discontinuity.
By this time it should have struck pupils with any budding
aptitude for science, and for such alone is this particular
chapter written, that it is strange and rather uncanny, un-
expected and perhaps rather disappointing, that magnitudes
should exist which cannot be expressed exactly by any finite
configuration of numbers : not only that they should exist,
but that they should be common. Draw two lines at right
angles from a common point, each an inch long; then join
their free ends, and measure the length of the joining line
(which is often called the hypotenuse of the right-angled
isosceles triangle that has been constructed) : that is one of
the quantities that cannot be expressed numerically in frac-
tions of an inch, i.e. in terms of the sides. Its value can be
approximated to and expressed, say in decimal fractions of an
inch, to any degree of accuracy we please; but the more
carefully it is measured the more figures after the decimal
point will make their appearance : the decimal is one that
never stops and never recurs. An infinite number of digits
are necessary for theoretical precision, though practically six
of them would represent more accuracy than is attainable by
the most careful and grown-up measurement. It is therefore
incommensurable, and can only be expressed exactly by another
incommensurable quantity, viz. in this case the square root
CHAP. XX.] INCOMMENSURABLES. 185
of 2. The length is ^2 times an inch, or about 1-4142 ...
inches. Draw a square upon it and it will be found to be
two square inches in area. That is just the fact which (when
proved) enables us to assert that each of its sides is of length
J^ ; since that is the meaning of the phrase " square root."
It may be proved by the annexed figure :
Fig. 11.
Where the shaded area ABC is an isosceles right-angled
triangle, the area of which is repeated several times in the
figure; four times inside a square drawn on the hypotenuse
AB, and twice inside a square drawn on one of the sides AC.
Wherefore the square on ^^ is twice the square on AC.
Observe however that there is nothing necessarily incom-
mensurable about a hypotenuse itself : it is only incommen-
surable when the sides are given. It is easy to draw a
hypotenuse of any specified length, say IJ inches long, and to
complete an isosceles right-angled triangle ; but now it is the
sides that will be incommensurable. The real incommensura-
bility is not a length, but a ratio, that is a number from
which dimensions have cancelled out. No length is incom-
mensurable, but it may be inexpressible in terms of an
arbitrarily chosen unit, i.e. it may be incommensurable with
the unit selected, and the chances are infinity to one that any
length pitched upon at random will be in this predicament.
It will not be precisely expressible in feet or metres, nor even
in fractions of them, though it can be expressed with any
degree of accuracy required.
186 EASY MATHEMATICS. [chap.
The hypotenuse of most right-angled triangles will be
incommensurable with both the sides, but there are a few
remarkable exceptions ; one in especial, known to the ancients,
viz. the one where the sides are in the ratio of three to four.
If such a triangle be drawn, with the sides respectively three
inches and four inches long, the hypotenuse will be found to
be five inches long ; the more accurately it is measured the
nearer it approaches to 5. It can indeed be shown theoreti-
cally, it is shown in Euclid I. 47, that it equals 5 exactly : a
surprising and interesting fact.
With an isosceles right-angled triangle however, no such
simple relation holds : the hypotenuse is ^2 one of the sides,
and ^2 is incommensurable; for, as we have previously
suspected and may now see, every root, whether square or
cube or fourth or any other root, of every whole number, is
incommensurable, unless the number be one of the few and
special series of squares or cubes or higher powers. Cf. p. 1 75.
To prove this we have only to observe that :
The square or any higher power of a fraction can never be
other than a fraction ; for you cannot fractionate a fraction
into a whole.
The square of a fraction cannot be an integer. Hence
no integer can have a fraction as its square root.* Yet every
integer must have a square root of some kind, that is a
quantity which, squared or multiplied by itself, will equal the
given number; but this quantity, though it may be readily
exhibited geometrically and otherwise, can never be exhibited
as a fraction, i.e. it cannot be expressed numerically by any
means, either in vulgar fractions or in decimals or in duo-
decimals or in any system of numerical notation; in other
words, every root of every integer except unity is incom-
mensurable (incommensurable, that is, with unity or any
* Attend here. It is easy to miss the meaning.
XX.] INCOMMENSURABLES. 187
other integer), except of those few integers which are built
up by repeating some one and the same integer as a factor ;
for instance the following set :
4 = 2x2
8=2x2x2
9 = 3x3
16 = 4x4
25 = 5 X 5
27 = 3 X 3 X 3
32 = 2x2x2x2x2
36 = 6 X 6
49 = 7 X 7
and so on ;
which class of numbers are therefore conspicuous among the
others and are called square and cube numbers, etc. Every
root of every other number is incommensurable, and most
roots of these are too.
Not roots alone but many other kinds of natural number are
incommensurable : circumference of circle to diameter, natural
base of logarithms, etc., etc. ; everything in fact not already
based upon or compounded of number, like multiples, etc.
Incommensurable quantities are therefore by far the com-
monest, infinitely more common in fact, as we shall find, than
the others : " the others " being the whole numbers and
terminable fractions to which attention in arithmetic is
specially directed, which stand out therefore like islands in
the midst of an incommensurable sea; or, more accurately,
like lines in the midst of a continuous spectrum.
What is the meaning of this ? The meaning of it involves
the difference between continuity and discontinuity. There
is something essentially jerky and discontinuous about number.
Numerical expression is more like a staircase than a slope : it V
necessarily proceeds by steps : it is discontinuous.
188 EASY MATHEMATICS. [chap.
A row of palings is discontinuous : they can be counted,
and might be labelled each with its appropriate number.
Milestones are also discontinuous, but the road is continuous.
The divisions on a clock face are discontinuous and are
numbered, and, oddly enough, the motion of the hands is
discontinuous too (though it need not theoretically have been
so, and is not so in clocks arranged to drive telescopes). The
hands of an ordinary clock proceed by jerks caused by the
alternate release of a pair of pallets by a tooth wheel — an
ingenious device called the escapement, because the teeth are
only allowed to escape one at a time ; and so the wheels
revolve and the hands move discontinuously, a little bit for
every beat of the pendulum, which is the real timekeeper.
The properties of a pendulum as a timekeeper were discovered
by Galileo ; an escapement of a primitive kind, and a driving
weight, were added to it by Huyghens, so that it became a
clock.
Telegraph posts are discontinuous, but telegraph wires are
continuous. They are discontinuous laterally so as to keep
the electricity from escaping, but they are continuous longi-
tudinally so that it may flow along to a destination.
But, now, are we so sure about even their longitudinal
continuity 1 The pebbles of a beach are discontinuous, plainly
enough ; the sand looks a continuous stretch ; but examine it
more closely, it consists of grains ; examine it under the
microscope, and there are all sorts of interesting fragments to
be found in it : it is not continuous at all. The sea looks
continuous, and if you examine that under the microscope it
will look continuous still. Is it really continuous 1 or would
it, too, appear granular if high enough magnifying power were
available 1 The magnifying power necessary would, indeed,
be impossibly high, but Natural Philosophers have shown
good reason for believing that it, too, is really discontinuous.
XX.] DISCONTINUITY. 189
that it consists of detached atoms, though they are terribly
small, and the interspaces between them perhaps equally
small, or even smaller. But even so, are they really dis-
continuous? Is there nothing in the spaces between them,
or is there some really continuous medium connecting
theml
The questions are now becoming hard. Quite rightly so ; a
subject is not exhausted till the questions have become too
hard for present answer.
There are several curious kinds of subterranean or masked
continuity possible, which may be noted for future reference.
Look at a map of the world ; the land, or at least its islands,
are after a fashion discontinuous, the ocean is continuous ; but
the land is continuous too, underneath, in a dimension not
represented on the map, but recognisable if we attend to
thickness and not only to length and breadth.
Human beings are discontinuous : each appears complete
and isolated in our three-dimensional world. If we could
perceive a fourth dimension, should we detect any kind of
continuity among them %
The questions have now become too hard altogether; we
have left science and involved ourselves in speculation. It is
time to return. A momentary jump into the air is invigorat-
ing, but it is unsupporting, and we speedily fall back to earth.
But how, it may be asked, does this discontinuity apply to
number? The natural numbers, 1, 2, 3, etc., are discontinuous
enough, but there are fractions to fill up the interstices ; how
do we know that they are not really connected by these frac-
tions, and so made continuous again ? Well, that is just the
point that deserves explanation.
Look at the divisions on a foot rule ; they represent lengths
expressed numerically in terms of an arbitrary length taken as
190 EASY MATHEMATICS. [chap.
a unit: they represent, that is to say, fractions of an inch;
they are the terminals of lengths which are numerically
expressed ; and between them lie the unmarked terminals
of lengths which cannot be so expressed. But surely the
subdivision can be carried further; why stop at sixteenths
or thirty seconds 1 Why proceed by constant halving at all ?
Why not divide originally into tenths and then into hun-
dredths, and those into thousandths, and so onl Why not
indeed ? Let it be done. It may be thought that if we go on
dividing like this we shall use up all the interspaces and have
nothing left but numerically expressible magnitudes. Not so,
that is just a mistake ; the interspaces will always be infinitely
greater than the divisions. For the interspaces have all the
time had evident breadth, indeed they together make up the
whole rule ; the divisions do not make it up, do not make any
of it, however numerous they are. For how wide are the
divisions'? Those we make, look, when examined under the
microscope, like broad black grooves. But we do not wish to
make them look thus. We should be better pleased with our
handiwork if they looked like very fine lines of unmagnifiable
breadth. They ought to be really lines — length without
breadth ; the breadth is an accident, a clumsiness, an unavoid-
able mechanical defect. They are intended to be mere
divisions, subdividing the length but not consuming any of
it. All the length lies between them ; no matter how close
they are they have consumed none of it ; the interspaces are
infinitely more extensive than the barriers which partition
them off from one another ; they are like a row of compart-
ments with infinitely thin walls.
Now all the incommensurables lie in the interspaces; the
compartments are full of them, and they are thus infinitely
more numerous than the numerically expressible magnitudes.
Take any point of the scale at random : that point will cer-
XX.] INCOMMENSURABLES. 191
tainly lie in an interspace : it will not lie on a division, for the
chances are infinity to 1 against it.
Let a stone — a meteor — drop from the sky on to the earth.
What are the chances that it will hit a ship or a man ] Very
small indeed, for all the ships are but a small fraction of the
area of the whole earth ; still they are a finite portion of it.
They have some size, and so the chances are not infinitesimal ;
one of them might get struck, though it is unlikely. But the
divisions of the scale, considered as mathematically narrow,
simply could not get hit accidentally by a mathematical point
descending on to the scale. Of course if a needle point is
used it may hit one, just as if a finger-tip is used it will hit
several ;. but that is mere mechanical clumsiness again.
If the position is not yet quite clear and credible, consider
a region of the scale quite close to one of the divisions already
there, and ask how soon, if we go on subdividing, another
division will come close up against the first, and so encroach
upon and obliterate the space between them. The answer is
never. Let the division be decimal, for instance, and consider
any one division, say 5. As the dividing operation proceeds,
what is the division nearest to it 1
At first 4 of course,
then 4-9,
then 4-99,
then 4-999,
and so on.
But not till the subdivision has been carried to infinity, and
an infinite number of 9's supplied after the decimal point, will
the space between be obliterated and the division 5 be touched.
Up to that infinite limit it will have remained isolated, stand-
ing like an island of number in the midst of a . blank of
incommensurableness. And the same will be true of every
other division.
192 EASY MATHEMATICS. [chap. xx.
Whenever, then, a commensurable number is really associ-
ated with any natural phenomenon, there is necessarily a
noteworthy circumstance involved in the fact, and it means
something quite definite and ultimately ascertainable.
For instance :
The ratio between the velocity of light and the inverted
square root of the electric and magnetic constants was found
by Clerk Maxwell to be 1 ; and a new volume of physics was
by that discovery opened.
Dal ton found that chemical combination occurred between
quantities of different substances specified by certain whole or
fractional numbers ; and the atomic theory of matter sprang
into substantial though at first infantile existence.
The atomic weights are turning out to be all expressible
numerically in terms of some one fundamental unit; and
strong light is thrown upon the constitution of matter thereby.
Numerical relations have been sought and found among the
lines in the spectrum of a substance ; and a theory of atomic
vibration is shadowed forth.
Electricity was found by Faraday to be numerically con-
nected with quantity of matter ; and the atom of electricity
began its hesitating but now brilliant career.
On the surface of nature at first we see discontinuity, objects
detached and countable. Then we realise the air and other
media, and so emphasise continuity and flowing quantities.
Then we detect atoms and numerical properties, and discon-
tinuity once more makes its appearance. Then we invent the
ether and are impressed with continuity again. But this is
not likely to be the end ; and what the ultimate end will be,
or whether there is an ultimate end, are questions, once more,
which are getting too hard.
CHAPTER XXL
Concrete Arithmetic.
It is highly desirable that arithmetical practice should be
gained in connexion with laboratory work, for then the
sums acquire a reality, and interest is preserved. It is
absolutely essential that all concrete subject-matter be based
upon first-hand experience, for unless that can be appealed to,
abstractions have no basis, but are floating unsupported in
air. It far too frequently happens that a child, constrained
to do sums expressed in terms of weights, has never weighed
a thing in its life. It is the same mistake as is made when
a child is drilled in the formal grammar of a language about
which it knows absolutely nothing. In every case concrete
experience should be the first thing provided, and abstractions
may follow. The teacher is apt not to realise this, because
grown persons have necessarily acquired some first-hand
experience in the ordinary course of life ; but a teacher who
is really educated all round and has a living acquaintance
with a great number of subjects should be able to enliven a
lesson into something quite exciting, if only he or she can
cultivate the patience necessary to allow time for the indi-
viduals of a class to attain some first-hand experience for
themselves.
L,.E.M. N
194 EASY MATHEMATICS. [chap.
This is the real object of school laboratory work, and the
mathematical teacher should seek to keep in touch with, and
to be aware of, what the pupils are doing under other
teachers, so as to illuminate his abstractions with concrete
instances and examples. By far the best kind of examples
are not those contained in books, but those which arise
naturally or are invented by a stimulating teacher in the
course of his exposition, or as a result of actual manipulation
on the part of the taught.
The result of a laboratory measurement is always an incom-
mensurable number; for the mere counting of a number of
distinct objects is not to be called a laboratory measurement.
No measurement of length, for instance, could ever be
expressed as a whole number of inches, nor yet as a whole
number plus a definite fraction of an inch. No measurement
that ever was made could be expressed by either a termi-
nating or a recurring decimal, nor by a vulgar fraction;
for any of these modes of specification would imply infinite
accuracy.
Suppose that an astronomical measurement is expressible
by the number 17 "4673, it is absolutely certain that 3 cannot
be the last digit of the series if it is to be expressive of
absolute fact. It may be that the next is 0, and perhaps the
next also, but unless you can guarantee that all the digits
to infinity are 0, the only reason for stopping at 3 (and it
is a good reason) is that we can measure no more.
So a decimal expressing the result of measurement cannot
terminate, neither can it recur. For, suppose the result,
as nearly as we could get it, were 4*6666, how do wc know
that the next digit is going to be 6, and the next and the
next also 1 We cannot know it.
If it did recur it would be the vulgar fraction 4|; hence,
this also is strictly an impossibly accurate result of measure-
XXI.] SIGNIFICANT FIGURES. 195
mcnt. The same with every vulgar fraction: it may be an
approximate result, but no more.
The phrase a quarter, or a half, or seven-eighths, is
appropriate therefore to rough specifications of approximate
magnitude, but is inappropriate to precise specification of
anything beyond counting of objects and fractions of an object.
Measurements should be expressed in decimal notation, and
the number of significant figures given should be characteristic
of the order of accuracy of the work.
The meaning of significant figures and practical
accuracy.
Rough workshop measurements are accurate, let us say, to
3 significant figures. Students' measurements in Physics,
which are naturally more difficult than those of the workshop,
if of the schoolboy kind, do well if they are accurate to two
significant figures. For instance, if the latent heat of melting
ice came out 79 or 80, it is quite as good as can be expected.
A great deal of trouble is necessary to get a third figure right,
for of course it means just ten times the accuracy. A good
student would however try to get the third figure right, and
might succeed, if it were not too complicated a measurement.
The Demonstrator, and senior students who give some months
to the work, would aim at 4 figure accuracy, and, if they
attained it, would do well. A few exceptionally skilled
experimenters with a genius for the work, devoting a year to
a research, might attain 5 figure accuracy, but such accuracy
as this is generally limited to the astronomical observatory,
where the measurements are fairly simple and the theory of
the errors to which instruments are necessarily liable has been
studied for centuries. In taking the mean of a number of
astronomical observations, even 6 figure accuracy is attainable,
but beyond this it is extremely difficult to go.
196 EASY MATHEMATICS. [chap.
The fundamental measurements that have to be made are
the following :
length
time
angle
mass
and of these, oddly enough, length is by far the hardest to do
accurately, though the easiest to do approximately.
Time is measured with considerable accuracy, even by a pocket
watch. Suppose the watch were uncertain by 3 seconds a day,
it would not be bad. If it lost or gained regularly it would
be a perfect time keeper, for a regular loss can be estimated
and allowed for ; but that is not feasible except in elaborate
chronometers carefully preserved. What is meant by the
above is that having allowed for any known regular loss, it
may lose or gain 3 seconds a day irregularly, so that to be
quite safe we might consider it uncertain to the amount of
plus or minus 3 seconds, or 6 seconds altogether. There are
86,400 seconds in a day, so the outstanding possible error
would be 6 parts in 86,400, or 1 part in 14,400, or 7 parts in
a hundred thousand, or '007 per cent., and it would therefore
be liable to cause a bad error in the fifth significant figure — an
error which even slightly affects the fourth. Still for a cheap
watch that is good performance, and means long hereditary
skill on the part of makers of watches.
You could not hope to measure a mile with the same
accuracy as you can measure the length of a day.
Angles are not very difficult to measure, because a number
of disturbing causes have no effect on the divisions of a circle.
If the weather gets warmer or colder, your yard and other
measures change, and clock-pendulums and watch hair-springs
change too; but though a circle expands and its divisions
grow wider with heating, their number is not affected ; the.
XXI.] SIGNIFICANT FIGURES. 197
expanded circle is still divided into 360 equal parts or
degrees. There is something essentially numerical about the
divisions of a circle ; and measurement of angle is subject to
fewer disturbing causes than measurement of length.
But the really easy thing to determine accurately is mass or
weight. For it never changes whatever you do to it. The
weight of a piece of matter is constant, whether it be hot or
cold, or whether it be evaporated to a gas, or dissolved in a
liquid, or whether it be molten, or boiled, or vaporised, or
chemically decomposed, or burnt up, or subjected to any other
operation. So far as is known, its weight continues absolutely
unchanged; although in combustion it appears to increase in
weight, because it combines with other things. Moreover the
balance is an easy and an accurate instrument. Even a
beginner can weigh on a reasonably delicate balance to 4
significant figures, that is, he could weigh ten grammes to the
nearest milligramme. He could hardly do better than that.
It is possible, however, with elaborate care to weigh to 6
significant figures, i.e. to weigh 10 grammes to the hundredth
of a milligramme ; but it needs a good balance and precaution
against currents of air, dust, warmth of observer's body, acci-
dental electrification, in some cases, and other disturbing
causes. These things do not really disturb either the weights
or the thing weighed, but they disturb the balance.
However, this is a digression, so as to make clear what
is meant by a reasonable number of significant figures. We
see that the number that is properly to be recorded will de-
pend upon circumstances, that every additional figure expressed
is a claim to greater accuracy, and that it is always better to
aim at too many than too few ; but we should cultivate an
instinct for knowing when we have recorded as many as the
experiment, or the observation, or the circumstances will
justify.
CHAPTER XXII.
Practical manipulation of fractions when decimally
expressed.
Since the results of all actual measurements yield incom-
mensurable numbers, it is desirable to be able to deal with
them freely. The present chapter will be considered very
elementary, but it is inserted thus apparently out of place in
order to emphasise the desirability of reintroducing familiar
matter with variations, and also, more particularly, to uphold
the doctrine that the other things treated are equally easy :
ease is only a matter of use and custom.
In the easy manipulation of fractions there is much to be
learnt, and considerable practice is necessary to attain facility.
It is not worth while to exaggerate this practice, because the
resulting art is not an accomplishment capable of giving plea-
sure to other people, like some other arts which can be
attained by practice ; nevertheless, some practice in arith-
metic is essential, and on this part of the subject some of the
time which has been saved from hogsheads and drachms
can be usefully and interestingly expended.
First of all, we may notice that the manipulation of frac-
tions is much simplified when they are stated in the ordinary
arithmetical notation, utilising the same system as is em-
ployed for whole numbers. The ease conferred is similar to
that gained by abolishing strange denominations of every kind.
Thus it is simpler to deal with 17 '34 cwts. than it is to deal
CHAP. XXII.] DECIMALS. 199
with it when expressed as 17 cwts. 1 quarter, 10 pounds, 1
ounce, 4^f drachms, which is the way that helpless children
are constrained to deal with it.
So also it is simpler to deal with a fraction when expressed
as 4-4 inches, than when expressed as 25 "mils" more than
4 inches and 3 eighths, or 4 + f + y|^o inches, which is, how-
ever, the way the British workman seems to prefer to have
it expressed — to the detriment of international engineering
operations.
In other words, it is always simpler to express a thing
numerically in a single denomination than to employ a multi-
tude of denominations or denominators.
Even such a thing as 1 + 1^ + J + ^ + yV is more simply
expressed as 1*9375, though still better as 2 - y\, or f J. Sim-
plicity is attained by use of a single denominator, whether
sixteenths or tenths, or whatever it may be. It is the admix-
ture of denominations or denominators that is troublesome.
So also the manipulation of fractions when expressed
decimally is as easy as the manipulation of whole numbers.
Care has to be taken about the position of the digits in either
case, and the explicit writing of the decimal point almost makes
the matter easier. The essential rule is, keep the decimal
points under one another, and they will then keep the
places of the digits right.
Thus, add together
4-375 + -025 + 53-1.
The sum is written 4*375
•025
53-1
57-500
and the result is verbally expressed as 57 J.
For just as various denominations, inches, weeks,
months, ounces, tons, gallons, are handy in speech and for
200 EASY MATHEMATICS. [chap.
realising and speaking of magnitudes after they have been
calculated, so vulgar fractions are often handy enough to
express a result at the end. When they are complicated,
however, they should only be used to quickly express approxi-
mate results. For instance, 5*12 inches might be spoken of
as about 5Jth inches. So also 35*9 inches might be spoken of
as about a yard. And the number 14 '34 might be spoken
of as about 14 J. For instance, if it expressed a length in
feet, the length should be called 14 feet 4 inches if we were
speaking to a carpenter. And similarly 5*67 feet would
be approximately 5f feet, or 5 feet 8 inches.
In subtraction just the same rule holds : keep the decimal
points vertical. E.g. to subtract 15*43 from 304,
write it 304-00
15-43
288-57
and there is nothing more to be said.
In countries with decimal coinage, this is all the arith-
metic that book-keeping clerks have to employ. Although
they may use the terms dollars, and quarters, and dimes, and
cents in ordinary speech, they do not express a sum of money
after our fashion, as
Dollars. Quarters. Cents.
17 3 18
but they express it simply as 17*93 dollars.
So, also, if another amount of 3 dollars, 2 quarters, and
17 cents has to be added, it is never expressed in that way,
but as 3-67 dollars; 17'93
3-67
the addition is then quite easy, viz. 21*60 dollars. All addi-
tion becomes simple addition ; and compound addition no
more exists.
To express the resulting amount in the form given by com-
XXII.] DECIMALS. 201
pound addition (which try), viz. as 21 dollars, 2 quarters, and
10 cents would be unnatural ; but it might, of course, be
spoken of as 21 dollars and 60 cents, for that etymologically
means precisely the same as 21 and 60 hundredths, i.e. 21*60.
The use of variegated and picturesque units, like weeks, and
fortnights, and centuries, and acres, and hundredweights, and
quarts, is to relieve the monotony of conversation ; they should
not be introduced into the workings of arithmetic. The end
result can be interpreted into them, for vivid realisation, as occa-
sion arises, and the instructed person should always be able to
speak to the uninstructed person in his own language. For
an instructed youth to expect workmen and others, who have
not had his advantages, to appreciate his scholastic mode of
expression, is barbarous, and shows a pitiful lack of sense on
his part.
So long as popular units exist they should be employed in the
proper place : they are part of folk-lore, and are often inter-
esting enough ; it is only when they are allowed to get out of
their proper place and spoil the lives of children that they are
to be condemned. In arithmetic proper they are out of place.
Now take multiplication. It is a little more troublesome,
of course, but not much.
Keep the points vertical, as before ; in other words, keep
the digits expressive of the same denomination under each
other, i.e. the units under the units, the tenths under the
tenths, etc. ; then the denomination of the answer looks after
itself without any trouble.
For instance, multiply 30*57 by 4*3. Write it thus:
30*57
4*3
122*28
9*171
131*451
202 EASY MATHEMATICS. [chap.
I need not have written the last figure of the result, for
most purposes ; for since the data are only given to two places
of decimals, an appearance of three decimal places in the result
may give a notion of spurious and deceptive accuracy, and so is
often better eschewed.
But this idea of approximate accuracy does not apply to
results in pure mathematics, such as the properties of numbers,
and things like that : it is the results of practical measurement
that are not wanted to impossible accuracy, just as the price of
a ship, or a railway, or a war, is not wanted closer than the
nearest penny, if indeed so close.
I may say, however, that when we are dealing with the
results of practical measurement, it is the number of significant
figures in the whole specification, rather than the number of
decimal places, which is the thing to be attended to. In the
above sum the data involved four significant figures, and so a
sixth significant figure in the result would be without meaning,
and ought not to be written.
Now take a further example in multiplication : suppose
we had 5-4306 grammes to multiply by 70*2 : the Avhole sum
would stand thus :
5 "4306 grammes
70-2
380-142
1-08612
381-22812 grammes
The weighing was only given to 5 figure accuracy, so any-
thing more is delusive in the result. Six figures may perhaps
be permitted, that is as far as 381*228, but the last two figures
after this, the 12, which are really -00012, have no useful
meaning, and need never have been written. And even the
8 is quite uncertain, so that the way to state the result with
the same accuracy as the data is 381*23 grammes. Three being
XXII.] dp:cimals. • 203
put as the last digit instead of two, because the next digit,
viz. 8, carries it more than half way to the higher figure.
We observe then that when we multiply by a figure in the
units place, we place the digits of the product under the cor-
responding digits of the multiplicand. When we multiply by
a figure in the ten's place, we shift each digit one place to the
left. If we multiplied by a figure in the hundred's place we
should shift them two places to the left. Whereas when we
multiply by a digit in the tenth's place, that is one place to the
right of the decimal point, we shift the resulting figures of the
product one place to the right, instead of writing them im-
mediately under the corresponding digits of the multiplicand.
The rule about division is similar. Let us divide 470-82 by
5*7. Write it in its first stage :
5-7) 470-82 (8
456-
Now, here 8x5-7 = 45*6, whereas in order to perform the
subtraction we really require 456, else the decimal points
would not be in the right position : hence the 8 is not
really 8, but 80 ; that is it is not in the units' place, but in the
ten's place, and so the decimal point is to be placed after the
next digit.
Performing the subtraction indicated above, we see that the
next digit of the quotient is a 2, and so the sum goes on
without any further trouble or attention :
5-7) 470-82 (82-6
456-
14-8
11-4
3-42
3-42
And there happens to be no remainder. But, if there were, it
204 EASY MATHEMATICS. [chap.
would give no trouble ; we should not take it up and express
it as a vulgar fraction, but should continue the sum in the
same way as before, bringing down ciphers as long as we chose,
that is until we had got the quotient to the required degree of
accuracy.
Dealing with fractions then in the decimal notation is just
as easy as dealing with whole numbers in the same notation.
The process is just the same, only we must be careful to put
the decimal point in the right place. So, however, we must
with whole numbers, only we do not have to actually write a
decimal point in their case (except in the quotient perhaps) ;
but we always have to be careful to interpret the quotient as
meaning hundreds or thousands, or whatever it is, correctly,
and that is essentialy the same thing as attending to the
position of the decimal point.
For instance, divide 729 by 14.
14
) 729- (52-07143
70
29-
28-
1-00
•98
•020
014
60
.56
40
The last figure in the quotient is not exactly 3, but that is the
nearest, and it is quite time to stop, as we have already reached
the extravagant accuracy of seven significant figures. If we
wanted to go on, however, there is not the slightest difficulty.
We simply go on till the remainder is negligible, not because
it is itself numerically small, but because it occurs so many
xxii.] DECIMALS. 205
decimal places away from the left hand significant figure that
only an utterly insignificant fraction is left. For instance, in
the above sum, the last remainder which is indicated, as giving
the quotient 3 in the fifth decimal place of the quotient, is
really '00040, and the multiplication of the divisor by 3 would
give "00042, which leaves a remainder of - 2 in the fifth
decimal place, to be divided by 14; with a result wholly
trifling.
In the above sum the decimal points and a few preceding
ciphers are indicated to show where they really occur, and to
show how they might be indicated all the way along, if we
chose; but there is no real need to indicate them anywhere
except in the quotient. At the same time it sometimes helps
to keep us right and clear to put all the points into the
process, where they ought by rights to be, and always to see
that they keep strictly vertical.
" Order " of Numbers.
As has been said before, in another connexion, p. 171, an
extremely useful idea is the "order" of a number, that
is to say the index of its order of magnitude : in other words,
the power of 10 which it represents. This can be definitely
specified by the distance of its highest significant figure to the
left or the right of the unit's place : distances to the left being
called positive, to the right negative ; the unit's place itself
being characterised by the order 0, and everything being
reckoned from that as the zero position.
For instance, any single digit, like 6, would be of the order
0; 26 would be of the order 1 ; 526 of the order 2; 8526 of the
order 3, and so on ; the order being given by the position of
the highest significant figure, and by nothing else. Thus
8526-79 would still be of the order 3; so would 8000, or 7000,
or 1000.
206 EASY MATHEMATICS. [chap.
26*79 is of the order 1.
6*79 is of the order 0.
•79 is of the order - 1.
•09 is of the order - 2.
What, then, is the order of -00058 ? Here the highest sig-
nificant figure is 5, and its position is 4 places to the right
of the unit's place ; hence the order of this number is - 4. So
also the numbers '0001 and "0009578 are of the minus-fourth
order; but 1-0009 is of the order 0 again, and 27-0009 is of
the order 1.
In the example -00058 it is right to say that the digit 5
is of the order - 4, the digit 8 of the order - 5 ; and it is
right to say that the number 58, which it contains, is also of
the order - 5. Again, in the number 525, we may say that
the 52 which it contains is of the order 1, that is to say,
that it occurs one place to the left of the unit's place.
It is often in practice convenient thus to attend to the order
of particular digits, or pair of digits.
The rule for multiplication and division can now be given
thus :
For multiplication of two numbers, take the highest signifi-
cant figure of each, multiply them together, and give the
resulting product a position representing the sum of the
orders of the two digits taken. For instance, multiply 36 by
745. You take the two highest digits, 3 and 7, the sum of
whose orders is 1 + 2 = 3. The product, which is 21, has to be
placed so that it shall have the order 3, that is to say, the
unit's figure of the 21 is to be 3 places to the left of the unit's
place.
Or take this example : — Multiply 081 by -742. We say 8
times 7 is 56, and this is to have the order compounded
of - 2 and - 1, that is to say - 3. Hence the 56 is to be placed
so that its unit's digit is 3 places to the right of the unit's
XXII.] DECIMALS. 207
place ; or in other words, there will be one 0 between the 5
and the decimal point.
The rule for division is to be stated similarly : —
Take the first significant figure of the divisor, and the first
one or two of the dividend : enough, that is, to be able to
effect a division. Then the resulting quotient will have the
order of this part of the dividend minus the order of the
figure taken in the divisor.
For instance, if we had to divide 81 by 742, there would be
no difficulty. We should take the 7 from the divisor, which
is of the order 2, and 8 from the dividend of the order 1 ; and
the quotient, has an order equal to the difference of the two
orders, viz. - 1.
But if, on the other hand, we had to divide 742 by 81, we
should take 8 from the divisor, where it is of order 1 ; but it
would be useless to take 7 from the dividend : we must take
74, its place being also of the order 1 ; so that the resulting
quotient will be of the order 0.
These matters are not particularly easy, they can be much
simplified by employing powers of 10, as we will soon show ;
but meanwhile we will do sums of this kind on commonsense
principles, as follows : Divide 742 by -081. A simple and
favourite way of doing such sums, is to get rid of the
decimals as much as we please by shifting the decimal point
in both equally, that is, multiplying them both by the same
power of ten, so that it would be transformed into 742 -f 81
simply. The answer comes out about 9'1605.
One more example. Take the inverse of this sum.
■742 ) -081 ( -1
i . — :
Here the first product 742 is required shifted one place to the
right in order to come under the proper digits of the dividend,
i
208 EASY MATHEMATICS. [chap.
so the quotient must be not unity, but one tenth, or -1. That
once determined, the rest is quite ordinary.
•742) -0810 (-109164
•0742
6800
6678
1220
742
4780
4452
3280
Now, here it must be admitted that people clever at arith-
metic do not write long division sums in so full and lengthy a
manner. They do both the multiplication and the subtraction
in their head, and write down the remainder only ; so that
the sum just done would look like this when people have done
it by aid of the " shop " method of subtraction :
•742) -0810 (-109164
6800
1220
4780
328
I can do it this way if I am put to it, but it seems to me a
needless tax upon the brain, at least when grown up ; and I
am more likely to make mistakes and am less able to check
them when made. Consequently for myself, I prefer the
longer method, for it is the same sum in reality, the only
difference is in the amount of it recorded on paper. I suppose
that very clever people indeed would record nothing of it
except the quotient : all the rest they would do in their head,
as if it were a short division sum, or would even perceive, in-
tuitively as it were, that --;=p = -109164, Boys have been
XXII.] DECIMALS. 209
known to be able to do things like this, and they are called
calculating boys. They are, however, rather rare. Never-
theless, people when young are much cleverer at learning
things than old folk, so perhaps they will get used to the
abbreviated method of recording, if they begin young enough,
and may like it better than the other. It is, I believe,
found so.
One other point, however, I must not forget to mention
here, and that is that if I had a sum like '742 -^ '081 to do,
742
I should first write it thus : -^y, and then proceed to look for
factors. If they do not occur easily, it is not worth while to
spend time in hunting for them ; still less is it worth while
to go through the farce of finding G.C.M. or H.C.F., or what-
ever it is called : one might as well be doing the long
division sum as that. And then I should proceed to look
out logarithms, and so turn it into simple subtraction. In the
particular instance I have chosen, however, it is hardly worth
while taking even this trouble, for directly you write 81, you
see that you can divide by 9 in two stages ; and although this
might be found a little unsafe in old-fashioned times, when
one had remainders to express as vulgar fractions, now that
we know how never to be troubled with remainders, we
proceed to divide numerator and denominator by 9 twice
over, as follows :
742 82-444
-^ = ^^^4^*"' =9-16049382716,
that is, for all practical purposes, 9*1605, as we found before
(p. 207) by long division.
CHAPTER XXIII.
Dealings with very large or very small numbers.
But there is a mode of dealing with all these sums which is
of great simplicity and service, and is more particularly useful
when the figures to be dealt with are nowhere near the region
of unity. In ordinary life we usually have to deal with a
moderate number of things, or a few simple fractions of things ;
we seldom have to deal with billions or trillions, or with
billionths or trillionths ; but in science there is no restriction
of this kind : we may have quantities of every order of magni-
tude to deal with. The human body is our natural standard
of size, and on it our measuring units are or ought to be
based. Everything much bigger than our body requires a
large number to express it; so also anything incomparably
smaller requires a very minute fraction to express it. We
must be prepared to deal easily and familiarly with very large
and very small numbers, and we need never suppose that a
large number requires a great number of significant figures to
express it ; for by that means it would not be of any different
size, it would only be expressed with preposterous accuracy.
A number like 17,199,658 is for most purposes quite suflS-
ciently expressed as 17*2 millions or 17,200,000.
So also our lifetime constitutes a natural human standard of
time, and our walking and other movements are standards of
velocity ; but, to express the facts of nature in general, these
CHAP. xxiiT.] ORDER OF MAGNITUDE. 211
magnitudes may have to be multiplied or subdivided to almost
any extent. The distance of the fixed stars, and the velocity
of light, and the age of the earth, are examples of one kind of
magnitude. The size of atoms and the duration of their
collisions lie towards the other end of the scale.
In many cases the precise numerical specification is of less
importance than is the order of magnitude ; sometimes because
it is not accurately known, sometimes because it may be
variable within certain limits. The "order of magnitude"
may roughly be said to be given by the number of digits
involved in its specification ; in other words, by the power of
ten concerned, without much regard to the particular figures
that precede that power. Thus, for instance, in 3 x 10^^ it is
the index ten which gives the order of magnitude; the numbers
4 X 10^0 and 5 x lO^o and even 8 x 10^^ or i x iQio ^re of the
same " order," viz. ' ten.'
So also the numbers 30 and 70 are of the same order of
magnitude, viz. ' one,' though one of the two numbers is more
than double the other.
The closeness of specification required depends upon the
subject matter and the object for which it is wanted. Occasion-
ally, though not often, it would be possible to consider ten and
a thousand as practically, though not technically, of the same
order of magnitude : they would be roughly alike as compared
with either a billion or a billionth.
Now let us take some examples of the index method of
dealing with figures. Take first mere numbers of different
orders of magnitude. For instance, divide, multiply, add, and
subtract the following pair of numbers in every way :
a =17,400,000, J =-0015;
which may be called 17-4 millions, and 1*5 thousandths, or
17 '4 X 10^, and 1-5 x 10"^ respectively.
212 EASY MATHEMATICS. [chap.
First we notice that when numbers differ greatly in magni-
tude, addition and subtraction are operations that are useless;
a + b and a~b are to all intents and purposes the same as a in
the above case ; the larger magnitude dominates the smaller, so
far as addition or subtraction is concerned. A million plus or
minus three is practically the same as a million. So no finite
quantity added to infinity makes the smallest difference to it.
This is a frequently useful fact : small quantities can be neg-
lected when added to or subtracted from large ones.
1+0=1, when z is small enough.
a^ -x^ = a^, when x is small compared with a,
which may happen either when a is very big or when x is
very small, or even when both are big or both small so
long as a is much bigger than x ; in other words, so long as
X
the ratio - is small. The term "small," so used, signifies small
a
compared with the other quantities concerned in the expression;
or sometimes, as in this case of the ratio, small compared with
unity.
But when we proceed to multiplication or to division, we find
a very different state of things ; there is then no domination
of a big quantity over a small one ; the bigness may be
exaggerated, or it may be partially destroyed, by the influence
of the small one.
Take the example suggested above :
a& = 17-4 X 106x1-5x10-3 = 26-1x103 = 26100 (i)
1 = 17-4 X 106 ^ (1-5 X 10-8) = 1 X 17-4 x 109= 11-6 ^ 10^ (2)
&_ 1-5 xl0-3_ 10-^ _ 10-^0 _
a~ 17-4x106" 11-6" i-ie - ^^-""^^ ^'^)
These results are numbered (1) (2) (3) for reference. The
three results are of different orders of magnitude. The
middle result is about half a million times the first ; it is so
much greater because a number has been divided by the small
XXIII.] ORDER OF MAGNITUDE. . 213
quantity b instead of being multiplied by it. The ratio between
results (1) and (2) is therefore exactly equal to //^, that is
(1-5)2 X 10-6 or 2-25 x IQ-^ or -00000225.
The first result is more than a hundred billion times the
third ; the ratio between them being a^, that is
(17-4)2 X 1012 = 302-76 x 10^2 = 3 x lO^^ approximately,
an enormous number, but not bigger than what we have
frequently to deal with in physics. The particles in a candle
flame are quivering with about this number of vibrations per
second, otherwise we should not be able to see the light.
Everything self-luminous must be quivering at this or at a
somewhat greater rate, consequently such rates of vibration
are quite common.
The result (2) compared with result (3) shows a still greater
difference in order of magnitude. To express the ratio,
11-2
which is tq^t^ x lO^'-*, a number of 21 digits is required, viz.
the number l-3xl02o, more accurately 1-2993x1020, a
number which is of the same order of magnitude as the
number of atoms in a drop of water.
Now take another example. If light travels a distance
equal to seven and a half times round the world in a second,
how long does it take to come from the sun, a distance of 93
million miles 1 How long does it take to travel 1 foot, or
say 30 centimetres. And how long to travel from molecule
to molecule in glass, supposing that they are the ten millionth
of a millimetre apart 1
• The circumference of the earth is just 40 million metres,
by the definition of a metre. It therefore equals 4 x 10^
centimetres. 7 J times this equals 3 x lO^^^ centimetres ; and
this distance traversed per second gives the velocity of light.
A mile is about 1 -6 kilometres, so the distance of the sun is
i-io X 1-6 = 149 million kilometres, as nearly as it is at present
214 . EASY MATHEMATICS. [chap.
known; in other words it is 149x10^x10^ centimetres ==
1-49 X 10^^ centimetres. Hence the time required by light for
its journey from the sun is
1 -49 X 1013 centimetres 1 490
„ ^^,„ cenUmetres
3 X 10^^
= 497 seconds.
second,
or about 8 minutes and a quarter.
This begins to illustrate the right method of dealing with
units. We shall have occasion to illustrate and emphasise it
later at much greater length ; but it will be seen already
that the "centimetres" in numerator and denominator cancel
out, and that the " seconds " in a denominator of the denomi-
nator come up to the top, and gives Us the units of the answer.
If this is not clear, never mind, we shall return to it and to
much more like it. We might have written the whole working
thus :
93 million miles _ 93 miles
7*5 X 40 million metres per second ""300 metres per second
= •31 X 1600 seconds = 496 seconds
= SJ minutes approximately,
and that is really the best and safest way to do it. We have
here put the actual data into the fraction, and then cancelled
out the " millions " ; next expressed numerically the ratio of
miles to metres, which is 1600, since 1*6 kilometre is a mile ;
and then we bring the " per second " out of the denominator,
and call it " seconds " in the numerator.
In so far as the two answers are not identical to the nearest
second, that is simply because of the approximate working,
which is justified by reason of the uncertainty of the data.
If the result were expressed as 496-666 seconds it would be
merely dishonest. The velocity of light and the distance of
the sun are both quantities which have had to be experi-
mentally determined, and neither is known with more tha?:-'
xxm.] LARGE AND SMALL NUMBERS. 215
three figure accuracy. In fact, the latter is not known quite
so closely as this. Moreover, it is at best only an average value :
the sun is not always at the same distance from the earth,
since the earth's orbit is not circular ; the distance we have
chosen is an approximation to the mean or average distance.
Now take the latter parts of the question, viz. the time
required by light to travel a foot, or say 30 centimetres ; and
the time required to travel a molecular distance of the ten
millionth of a millimetre, or 10"^ centimetre.
These are quite easily ascertained, since the velocity is
given as 3 X 10^^ centimetres per second. To travel 30 centi-
30
metres liffht takes ^ ■,^,^ second, that is 10"^ second, which
° ox 10^"
means the thousand millionth part of a second.
10-8
To travel from molecule to molecule, it takes ^ — i7uo~ J 10"^^,
or say the third part of the trillionth of a second. Here
the digit 3 is quite unimportant. The order of magni-
tude is all that is of any use, and that is the trillionth or
10 "1^ of a second. Molecular magnitudes are not known more
accurately than that. It may be considered remarkable that
they have been measured at all. The way they are obtained,
a way necessarily indirect, can only be understood later.
With attention to these early stages, this and much else can
presently be understood by everybody. At present, grown
people are ignorant of all these things, because they have not
prepared their minds.
Now take a more childish example, akin to the horseshoe
nails, page 155, and perhaps equally surprising.
A country the size of England was being besieged by a
hostile fleet, and its inhabitants were in danger of starvation
because they did not grow their own corn. Under these
circumstances the captain of a merchant steamer craved
216 EASY MATHEMATICS. [chap.
permission from the enemy to run the blockade with a chess
board full of wheat for his starving wife and family, the board
to contain a single grain of wheat on the first square, two
grains on the next, four on the next, and so on.
But when the enemy's admiral had had the necessary
calculation made, by a Japanese sailor who happened to be on
board, and was informed that the corn thus to be passed
through his lines was sufficient not only to feed but to
smother every living soul in the country, in fact to cover
the whole land with a layer of grain more than a dozen yards
thick, he declined to grant the request unless the whole
supply were delivered at one operation.
To do the sum, proceed as follows : — The number of grains
is 2^\ or, strictly speaking, one grain less than this number.
A mode of arriving at this, if it is not obvious, will be given
below, but it could be reasoned out by an intelligent beginner.
Call the number 71.
then log 71 = 64 log 2,
and log 2, either from the curve (p. 179) or from a table of
logs, is approximately -3, more accurately -30103 ;
hence logTi = 19*266.
n is therefore a number with twenty digits of " order " 19 ;
in other words, it is approximately eighteen trillion; more
accurately it is 1*845 x 10^^.
This number is not so great as the number of atoms in a
drop of water, but it is a large number. To see what it
means : buy half a pound of wheat as imported, without its
husk, etc. — it costs only a penny — and devise a plan of practi-
cally counting the grains in say a cubic inch, without actually
counting so many individually. This should not be beyond a
youth's ingenuity.
I find that on the average a grain is J inch long and } of an
inch broad. So if they were regularly arranged, in what is
XXIII.] GRAINS OF WHEAT PROBLEM. 217
called square order, there would be 28 of them lying in a
square inch ; and if piled up an inch high, also in regular order,
there would be 7x28 = 196 in the cubic inch so constructed.
An allowance for irregularity should doubtless be made,
but it is uncertain ; it is not even quite clear whether more or
fewer could be got into a given space by a higgledy piggledy
arrangement than regular packing in artificially square order.
It will be near enough if we take it as about the same, and so
estimate 200 grains of wheat to the cubic inch.
We are now prepared to go on with the sum set. The area
of a country the size of England is given in the geography
books, or the Penny CyclopaBdia, as 50,000 square miles. A
mile is 1760 x 36 inches = 6-336 x lO'* inches, so a square mile
is the square of this, viz. 40-145 x 10^ square inches, of which
the first two digits are sufficient for our purpose.
Hence the area of a country as big as England is 5 x 10* x
40 X 10^ = 2 X 10^* square inches. Now the number of grains
which are to be distributed over this area is given by our
previous working as 1*845x10^^, and we have ascertained
that, roughly speaking, 200 of the grains will occupy a cubic
inch. Hence the number of cubic inches which have to be
provided to hold all the corn, is the 200th part of 1-845 x lO^^,
that is to say, -922 x lO^'' ; or just less than the tenth of a
trillion cubic inches. To provide this capacity on the surface
of the country, the grain would have to be spread all over it
in a uniform layer, of thickness
•922 X lOi'' cubic inches ._, ^^o t . i
—^ — Tmi = — c — = '4ol X 10"^ linear inches.
2x 10^^ square inches
In other words, the corn would flood the whole country
to a depth of 461 inches, or 38-4 feet, which is as high
as an ordinary house. All cottages would therefore be
completely submerged by the chess board full of grain
distributed uniformly over the face of the country.
218 EASY MATHEMATICS. [chap.
Perhaps, however, our initial step, that the number of
grains is precisely 2^^ - 1, was not obvious. It can easily be
seen thus. The number on any square will be one grain more
than all those on the preceding squares added together.
Thus, for instance, the number on the third square is 4, and
the two previous squares contain one grain and two grains
respectively, or 3 grains together ; which is one less than the
number on the next. The number on the next following
square is 8, and the three previous squares together hold 7, or
again one grain less, and so on. Hence the number on the
tenth square would be the number on the nine previous
squares added together, plus one. The number on the tenth
square is 2^, so the number on the 9 previous squares added
together will be 2^ - 1. The number on the sixth square is
2^ = 32, hence the number on the five previous squares added
together will be 31 ; and so it is, viz. 1 + 2 + 4 + 8 + 16, which
equals 31. Compare page 323.
Now the total number of squares in a chess board is 8^, or
64 ; the number of grains on a 65th square, if there were one,
would be 2^^, hence the number on the 64 previous squares
added together (which is just what we want) is 2^* - 1.
This peculiar result of continued doubling, that the product
each time just exceeds the sum of all the preceding products,
has suggested a plan of what is called "breaking the bank," at
a place where you stake on one of two events, either of which
is equally probable, say red or black, and . win back double
your stake if you win, that is receive your stake and another
added to it by the " bank." The simplest rule for " breaking
the bank " is simply this : Begin small, and double your stake
every time you lose ; whenever you win, begin again.
If it were feasible to continue this process you could never
really lose, because your stake would always just exceed the
sum of your previous losses, so that whenever you won you
XXIII.] LARGE NUMBERS. 219
would get them all back plus 1 counter more. Winning would
be slow but sure. To work the process you must be prepared,
however, with a considerable number of counters to stake
with, if you happen to lose many times in succession. And as
a matter of fact every ' bank ' protects itself against so simple
an arithmetical device by declining to receive more than a
certain maximum stake. If therefore you have staked the
maximum and lost, you have no way of getting your losses
recouped ; and so it is universally conceded, even by gamblers,
that there are more profitable, as well as more useful, ways
of earning a living.
The operation of constant doubling is a particular case of
what is generally called geometrical progression, and it
is remarkable how rapidly we can thus reach enormous
magnitudes.
Of course, if instead of doubling we treble or quadruple
each time, the large result will be reached still sooner, but, as
a matter of fact, any constant factor greater than 1, repeated
often enough, will grow to any magnitude ; whereas any
constant factor less than 1 repeated often enough will obliterate
or reduce to insignificance any initial magnitude.
Take an example. Let the factor be 1*1, that is one and a
tenth. Multiply it by itself 20 times, so that the result is
(1-1)20, whose value can be found by logarithms easily enough,
thus :
Call it X, then log x = 20 log 1 -1 = 20 x -0414 = -828,
so:c = 6*6; showing that the initial rate of increase is slow.
Look, for instance, at the geometrical progression or compound-
interest curve on page 357, which is the same as the
exponential curve on page 101 taken backwards, and note
that it begins slowly. But continue the process until we have
reached (l-l)^^^, whose logarithm will be 100 x -0414, or 4-14,
and already the result is 13,800. To get really large numbers
220 EASY MATHEMA1 ICS. [chap, xxiir.
with such a factor as 1-1, we should therefore have to repeat
the operation very often indeed.
So also to reduce to insignificance by means of such a factor
as '9, we should have to repeat the multiplication very often :
for let (-9)100 = x
then \ogx = 100 log -9 = 100 log ^^
= 100 (log 9 - log 10) = 100 log 9 - 100
= 95-42-100 = -4-58 = -5 + -42 =5-42,
wherefore a; = -0000263,
a number which is of the " order " - 5.
Illustrations of excessively rapid multiplication by geometri-
cal progression occur in Natural History, where certain
organisms are known to increase at a prodigious rate, this rate
of increase being the cause of plagues, like a plague of
locusts, or blight, or like certain kinds of disease. For sup-
pose a parent insect laid and hatched a thousand eggs
(which is indeed a very moderate number), and suppose each
of these also hatched a thousand, and suppose each genera-
tion only required a month to come to maturity, and lived for
a year ; the number of descendants in the course of twelve
months would be a thousand raised to the twelfth power,
that is to say a number of the " order " 36, or 1 followed
by 36 ciphers, or a trillion-trillion.
Some diseases are caused by the fission or splitting up
of cells into two or more, which rapidly grow and split up
again. In such case the rapidity of increase can be still more
prodigious, because the time which need elapse between
the splitting and the re-splitting of cells may be short.
It does not follow that these geometrical-progression rates
of increase apply without qualification to every kind of
population, nor to one whose needs are in excess of available
supply. For some actual tacts, however, see Appendix IV.
CHAPTER XXIV.
Dealingfs with Vulgar Fractions.
Having now exhibited the easy mode of dealing with frac-
tions, we must proceed to the more difficult method where the
division operation has not been performed, but is only indi-
cated : the same sort of indication as has been used in algebra.
For instance, to divide a hy b you cannot really do it in
algebra, you can only indicate it, as ct -r & or t. So also in
arithmetic if one has to divide 3 by 4, we can, if we choose,
do it, and write '75 simply, but for many homely purposes it
is sufficient to indicate it only, and leave it in the form of
3 -r 4 or f . Fractions left like this are not so easy to deal
with, but they usually apply to such simple magnitudes that
they are simple enough. For anything complicated, however,
they are unsuitable, and they must be simplified ; moreover,
as we have seen in Chapter XXL on concrete arithmetic, they
seldom occur in practical measurements; nevertheless we
must learn to perform the fundamental operations upon
vulgar fractions without having necessarily to reduce them
first to a simpler form. One way of dealing with mixed
units, such as cwts., quarters, and lbs., or pounds, shillings,
and pence, is to reduce them all to some one denomina-
tion; but it would be rather stupid if we did not know
how to treat them in any other way. So also one way
222 EASY MATHEMATICS. [chap. xxiv.
of treating collections of vulgar fractions is to reduce them
all to some one denomination, or to decimals ; but we ought
to learn how to manage them without this preliminary-
operation.
So we will proceed to illustrate by example some simplifying
processes, first reminding ourselves of the fundamental opera-
tions of addition, subtraction, multiplication, division, involu-
tion, and evolution, applied to vulgar fractions.
No further explanation is needed beyond what has gone
before, Chaps. III., IX., etc. For addition, make a common
denominator and cross multiply.
X y hx + ay
or another example
a 0
X
;7 + y =
For subtraction, the same.
ah ■
>
x + ay
a
hx-ay
ah '
uals —
ay
a b
or if given --y, it equals
For multiplication, multiply numerators together and
denominators together.
x^y^xy
a b ab^
or, in the common case when one denominator only appears,
X nx
-xn=—.
a a
For division, invert the divisor, and multiply
X y X h _hx
a ' b a y ay'
XX
or -^n=--.
a na
XXIV.] VULGAR FRACTIONS. 223
For involution, operate similarly on both numerator and
denominator. /a^\« r,^
41J a"
For evolution, just the same ; only we may write ?i as — if
it is a fraction, if we like,
Numerical Verifications.
The simplest fractions of all, to deal with, are those which
are not really fractions, but integers in disguise, like ^^- or -%*-,
and these serve for testing any operation easily and quickly.
If these two are added, for instance, the result according to
the above rule is 12 24
96 + 72
24
168
~ 24 " ^•
Subtracted, the result is
96-72
24
= ?-^ = l
24 ^•
Multiplied, the result is
288
24
Divided, they become
= 12.
12 8
3 ""24
= 1=1-333....
Squared, the results are
g* = 16, and -
576 _
64 ^'
respectively.
Square-rooted, we get
n/(12) 2^3
v/3 V3
2 for one.
^"^ ^ = l72 = 72 = N/3f-'''«°*''-
224 EASY MATHEMATICS. [chap.
Of course, in practice we should get rid of such pretence or
imitation fractions, by expressing them as whole numbers at
once, before beginning operations on them. But we may often
have fractions not unlike them ; composed, that is to say, of a
whole number and a proper fraction in addition. Those
fractions, larger than unity, are sometimes called "improper
fractions," and when expressed as an integer + a fraction, they
are called "mixed numbers"; but these terms are hardly ever
used out of the schoolroom.
A pretended fraction like y- cannot be concretely exhibited
to small children unless there are several things to be cut up.
With several apples we can do it ; for if we cut them all up
into thirds, and then pick out 12 of the thirds, we shall find
that we can build up with them 4 apples.
So also -^/- would give us 4 apples and J of another ; and we
could not exhibit it properly unless we had 5 apples to start
with.
These fractions are called improper fractions because they
are not fractions of one thing, but fractions of a lot of things.
To exhibit a proper fraction like | is easy, for we have only
to cut an apple into half quarters and then remove five of
these, leaving f ths of the apple behind.
.It is from this point of view that vulgar fractions are simpler
than decimals. There is always some good reason why a
popularly employed nomenclature has been hit upon. For
these simple things it is excellent : it is only when we come to
complicated things that we find it rather difficult. In practice,
however, difficult sums never occur in this form, and there is
no reason for wasting the time and brains of children in
simplifying unwieldy artificial complications ; for these things
may give them much trouble when young, whereas later, if
they ever learn mathematics, they will experience none at all,
even if they come across the most complicated of them.
CHAPTER XXV.
Simplification of fractional expressions.
We Mall attend now to certain simple operations which
constantly occur in practice ; it is easy to get accustomed to
them and to take an interest in them, as in any natural
exercise of intelligence. First of all we will take "cancelling,"
that is, striking out of common factors, a process in which
useful ingenuity can be trained. Examples are better than
precept, so try the following : Simplify
36 X 108 X -91
17-28 X 65 x8-r
Here we see at a glance that a lot of factors can be struck
out, because
36 = 3x12, 108 = 9x12, 91 = 7x13,
though that last is not so likely to be known, unless an
extended multiplication-table has been learnt — a very useful
accomplishment ; moreover
1728 = 123, 65 = 5x13, and 81 = 92.
As to the position of the decimal point, that is a matter
that gives no trouble at all. The decimal point must always
occur somewhere ; it is understood and not written at the end
of integers, but it is there all the time ; and its influence can
be attended to after the cancelling has been done. Of course
we might shift it equally to the right in both numerator and
JL.E.M, P
226 EASY MATHEMATICS. [chap.
denominator, and so get rid of its explicit appearance, and
we shall do this ultimately, but we will not do that at present
since we want to use it as an example ; we will cancel factors
as they stand, and leave the decimal points unchanged in
position till the end ; and the result, written down in practice
in a few seconds without all this talk, is
3x9x-Q7 _3x-07_-07 7
•12x5x8-l~-6x-9 "OS" 18'
This will not simplify further, because 7 is a prime number
and does not go into 18.
It would be very seldom useful to write the result as
1 1
or
2| 2-5714...'
but it would often be useful to write it as
3*5
-Q- = -3888 ... or approximately '4.
Take another example : one less likely to occur however,
one of a double fraction.
8 w 2 1 •
TT¥ ^ T(J(T
Here we may cancel out factors among the numerator frac-
tions, and likewise among the denominator fractions; but
we must not cancel a factor in the upper numerator against a
factor in the lower denominator. 119 contains the factor
17, and also the factor 7, being equal to 7 x 17 ; and obviously
543 contains the factor 3, since its digits divide by 3. 100,
as usual, is useless for factor purposes.
So we re-write the fraction (with needless elaboration)
1 4 181 4x181
1 "" 9 "" ~r ~9~
or
1 "" 100 100
XXV.] FRACTIONS. 227
the middle line being drawn rather longer and stronger than
the other two, so as to show that the upper of the two
fractions is to be divided by the lower.
We have now to multiply the extremes and divide by the
means, a convenient rule to remember, giving us
400x181
63 •
But a rule of this kind should never be given, it should be
ascertained and, if possible, invented by the pupil. To invent
a handy rule involves a little bit of original thought, and the
opportunity for exercising that vital power should never be
lost.
Hence we should not at first make the above convenient
short cut, but proceed thus :
rpi. 4x181 7
This means —9—^100.
and is an example of division of fractions, so invert the
second number and multiply
724 100 72400
9 "" 7 ~ 63 •
To express this in decimals we might proceed thus, for
although 9 is not a factor of the numerator it will be approxi-
mately one, and can at any rate be divided out, leaving
^^^ = 1 1 49-206349206349 ....
Digression on recurring decimals. — No importance attaches
to the notation of the superposed dot or dots for circulating or
recurring decimals. Children may write as many of the recur-
ring places as it amuses them to write. In practice, the result
would not usually be wanted beyond the first '2, which is nearly
equivalent to six figure accuracy, since the next figure is a 0.
The interpretation of recurring decimals as vulgar fractions,
228 EASY MATHEMATICS. [chap.
with a certain number of nines and noughts in the denomi-
nator, is of no practical moment. It should be reserved for
the more intelligent and irrepressible children, and by them it
might be found out, with great advantage. Children who
delight in finding out such things are on the way to acquire
some of the powers and tastes of the pure mathematician.
The simplest case may, however, be known, and perhaps
this amount of hint would be necessary even to sharp boys.
A recurring decimal is a geometrical progression, with fractional
common ratio, and extending to infinity.
Thus the commonest of all
•3 or -3333
roVim^ •
the common ratio being Yuth.
Hence its sum, by the rule for G.P. [see Chap. XXXV.] is
a _ '3 3 1
l-r~l-TV~9"3-
The precise value of the answer on p. 227 expressed as a
vulgar fraction, though never really wanted, is well known by
everybody (needlessly well known for so trivial and useless a
thing) to be 11 49f f ||||. (End of digressim.)
The fact that ^ equals %
y
as is proved by writing it
a h a y
— — — = — X —
x' y X V
is worth remembering : most easily remembered as worded in
the rule, multiply the extremes for the new numerator and
the means for the new denominator.
XXV.] CANCELLING. 229
The y may be called a double denominator, and we observe
that it comes up into the ultimate numerator.
The rule for cancelling may also be similarly illustrated :
na
a
In
X
y
or in
~mx
my
e Ti's
and the
m's cancel out.
But
in such a
\. fraction as
this
ra
X
T
ry
the r's, so far from cancelling, appear in the result twice over,
that is, squared ; for it equals
r^ay
bx '
The rule for cancelling in the case of double fractions there-
fore is : cancel common factors from alternate members in the
double fraction, then deal with the extremes and means to
attain the simplified result.
It may be preferred, and it may be safer, to perform the
latter operation first, and so keep all the cancelling for an
expression in simple fractional form ; but either is a correct
procedure.
CHAPTER XXVI.
Cancelling among units.
It is not only numbers that can so be cancelled : we may
and often do, have fractions composed of concrete or physical
quantities — quantities with length, and breadth, and thickness,
and weight, and velocity, and other things. It will be found
that cancelling can conveniently go on among these also.
Suppose we had the following ratio to interpret:
330 yards x 16 square yards x 77 lbs.
4 inches x J mile x '14 ton x 5 minutes'
an experienced eye would see at once that the result was a
velocity, i.e. that it could be expressed as so many miles an
hour, or feet per second. And the working is on the following
lines, though again the actual operation is much speedier than
is the explanation of it : —
First we have the ratio of yards to inches, which is 36, and
this is most conveniently and safely recorded by erasing the
word **yard " and replacing it by " 36 inches."
Next in the numerator we have square yards, and in the
denominator we have a linear mile, which is 1760 linear
yards, and that value is therefore conveniently substituted for
"mile."
Then we have the ratio of tons to pounds, which is 2240 ;
and we get left with one of the " yard " factors of the square
yards uncancelled in the numerator, and with "minutes"
CHAP. XXVI.] CONCRETE QUANTITIES. 231
uncancelled in the denominator. The result, before any cancel-
ling is done, will be the following :
330 X 36 inches x 16 yards x yards x 77 lbs.
4 inches x ^ x 1760 yards x -14 x 2240 lbs. x 5 minutes'
Now we can strike out a number of units common to both
numerator and denominator, and can at the same time do
some numerical cancelling, of which we will indicate the
steps sufficiently, noting that 11 is a factor of 1760, because
the sums of its alternate digits are equal.
30x9x32x11 yards
160 X -02 X 2240 x 5 minutes*
Now, we see that 16 will cancel out with 32, and of
course the ciphers can go from the 30 and the 160.
So we get it thus
27x2x11 yards
•02 X 2240 X 5 minutes*
So many yards by so many minutes ; in other words, a
velocity of so many yards per minute. How many 1
27x11 297
•1x1120 112
yards per minute.
Here, perhaps, it is simplest to resort to long division, since
no more factors are obvious : so we might leave the answer as
2-6518 yards per minute, which is a sort of racing snail's
pace ; or we might reduce it to other units. This last is a
thing which often has to be done, and so no opportunity for
showing the right way to do it must be lost at this early
stage.
This is the easiest and only safe way :
^97 y^^ds ^ 297 X 3 feet ^ 297 ^^^^^^
112 minutes 112 x 60 seconds 2240 ^
232 EASY MATHEMATICS. [chap.
or again
miles
297 yards ^ 297 neo 297 x 60 miles
112 minutes" 112 ^«~ 112x1760 hours
27x3 81 .,
= 56716 = 896 ^^^"^P"^^^^"-
It may be said that simple reductions like that can easily be
done without writing them down fully. So they can, but
they can easily be done wrong. Change of units is a subject
extraordinarily easy to make a slip in, especially by multiplying
where one ought to divide. It is at best a mechanical process,
and it should be done mechanically ; that is by a straight-
forward method which involves no delicate thought, and affords
no loopholes for mistakes to creep in.
To check the above result, we can recollect that 4 miles an
hour is about 2 paces or 6 feet per second ; so that the ratio of
the above two specifications for the same thing should be
roughly as 3 to 2. And so it is ; for the first is very roughly
4- foot per second, while the second is roughly yV mile per
hour; and the ratio of 7 to 11 is not very different from that
of 6-7 to 10, which is frds.
This rough-and-ready checking, in terms of anything that
comes handy, and with quite rough approximation to the
figures, is very useful, and, in real practice, wise ; else we exhibit
the ridiculous result of academic correctness in minutiae, and
commercially hopeless error in the order of magnitude ; so
that, for instance, a quantity pretending to be accurate to four
or five significant figures may be all the time a thousand
times too great or too small.
This is the kind of thing that always moves the practical
man to legitimate and sarcastic mirth, because he could get
nearer than that by his own untutored instinct and common-
sense. People who have been elaborately tutored, but have
XXVI.] DEALma WITH UNITS. 233
not taken care for themselves of their own birthright of
coramonsense, are denominated "prigs," and their existence
tends to bring education into contempt.
We must take another example of this cancelling of units,
and we will take an instance of the occurrence of a double
denominator in them. Suppose the following given :
15 cwt. X (32 ieety
1080 grammes x 400 centimetres per second x | yard'
Here the experienced eye will see that the result must be a
time, for every kind of unit will cancel out except the "per
second " in the denominator. This is what I call a double
denominator, for the per alone would put it in a denominator ;
so the result is that it comes up into the ultimate numerator.
To work the sum, proceed thus (with full elaboration
shown, because it is an illustration) :
15 X 112 lbs. X 32x32 x (feet)^
- ^oA ^ metres , ^ r /
1080 fframmes x t- x 1*5 feet
° second
1120 lbs. x8x 32 feet
1AOA 1 metre'
1080 grammes x r
° second
18 454 erammes x 256 feet ,
■=-= X ^ -„ - — - seconds
17 grammes x 3*28 feet
18x454x256 ,
- 17x3-28 ^^^^^^-
To work this out, either a slide-rule or logarithm-table would
be advantageous. Suppose we take this as an opportunity for
utilising a table of four-figure logarithms, and see what we get.
log 18 = 1-2553 log 17 = 1-2304
log 454 = 2-6571 log 3-28= -5159
log 256 = 2-4082
6-3206 1-7463
234 EASY MATHEMATICS. [chap. xxvi.
Subtracting, we get 4*5743,
which is the logarithm of 37530 to four significant figures ;
and 37,530 seconds is therefore the answer.
This is equal to 10 hours 25 J minutes; and so the com-
plicated expression, involving many kinds of units, with
which we started, represents really nothing more elaborate
than the length of a working day.
These examples are rather dull and artificial ; but to take
a real example, which would lead to this kind of concrete
result, would assume some knowledge of mechanics or physics.
Suffice it to say that plenty of quite similar examples will
occur when we get to real subjects like those, and, meanwhile,
all that we can show is that they involve no difficulty
of dealing with and interpreting. No admixture of units
involves anything the least difficult : it only wants disen-
tangling; and, in order to disentangle it securely and
easily, the best plan is not to be afraid of writing out the
thing at length, with all the factors present — both the
numerical and the concrete units, or standards — and so
gradually boil it down by a mechanical process involving
no troublesome thought.
Whenever thought is necessary, it is to be exercised vigor-
ously, but it should not be wasted over simple mechanical
operations. Take thought once for all, learn good methods,
and so economise thought in future. This is, indeed, the
principle of any mathematical machine. Machines can be
constructed, and are used, for performing really intricate
mathematical operations; for analysing out the harmonic |
constituents of a tidal or other irregularly periodic curve, for
instance. To devise such a machine required thought, and
indeed genius, of the highest order: to work it, requires
nothing beyond what an intelligent office boy can learn.
CHAPTER XXVII.
Cancelling in Equations.
One more detail concerning cancelling may here be
mentioned. It relates to cancelling on either side of an
equation. One side of an equation may be considered as
divided by the other, and the result equated to unity, so that
the rules for cancelling are easily deduced.
r or instance m — = -^
a 0
the n's may be cancelled, for it is equivalent to
nx
ny
t
nxb ^ bx ^ ,
or • — = l,or — = 1, or ox = ay :
nya ay
as might have been seen at once by cross multiplying.
Suppose, for instance, that it had been written thus :
nx f^y _r.
it would have been the same thing ; and the left hand of this
equation might be reduced to a common denominator, with
the subtraction performed as far as possible, by writing
hnx — nay_^
ab
236 EASY MATHEMATICS. [chap.
How comes it that this is the same thing as bx = ay1
Because it may be written
^(bx-ay) = 0,
and, in order that this may be true, one of the two factors
must vanish, that is, must itself equal 0. For you cannot get
zero by multiplying two finite quantities together. Hence
either -r mubt equal 0, which is in some cases possible,
but is clearly not here intended ; or else bx-ay = 0. And the
latter cannot happen unless ay and bx are equal.
So we get this simple rule, that when an expression is
equated to 0 any factor can be struck out, without having
to be accounted for, provided always that that factor is not itself
zero.
This last is a most tremendously important proviso, and its
neglect may land you in the utmost absurdity. If we strike
out a factor zero, from an expression equated to zero, we may
be striking out the very and only factor which made it zero ;
the factor left behind may have any value whatever: the
equation declines to tell us for certain what that value is,
and we must not proceed to work on the assumption that it
does. Similarly, if a zero factor is cancelled on either side
of an equation, we can make no deduction concerning the
equality or otherwise of the residual factors.
Caution.
This inequality of zeros is a matter of great importance, and
I must proceed to illustrate it even at this stage, though we
shall find plenty of instances later on.
Suppose an expression like this were given, from which to
find X. iej(n^-4)(x^ + ¥)
XXVII.] CANCELLING IN EQUATIONS. 237
we should be quite safe in striking out 16 and likewise ^ (viz.
the 3 in the denominator), for these numerical factors are
certainly not zero, so we should get left with
VK- 4)^ = 0.
Now, if we strike out the factor J(n^ - 4) and the factor
-, we shall be left with the impossible result x'^ + b'^ = 0.
Why impossible ? Because it means that
x^= -h\
and the square of a real number, whether that number be
positive or negative, cannot possibly be negative ; for two
similar signs multiplied together give a positive sign always ;
-3x -3=+9 just as much as 3 x 3 does.
What the equation suggests is that, under the circumstances,
n must equal 2. It is the {n - 2) component of the (w^ - 4)
factor, and not the factor containing x, which is responsible
for the zero value of the whole; and the equation tells us,
therefore, nothing at all about the value of x.
I do not say that that is all that can be deduced from the
equation, but that is all that lies on the surface.
To clinch the danger of striking out a factor, without at the
same time recollecting the possibility that it may be itself the
essentially zero factor, the following absurdity may be given.
To " prove " algebraically that 2 = 1.
Let ic = 1, so that a; - 1 = 0,
then a;2 = 1^ and ic^ — 1 = 0.
So «2 - 1 =a; - 1, since both equal zero, *
wherefore (a; + 1) (a; - 1) = (a; - 1).
Cancel out the factor (a; - 1) from both sides, and we get left
a;+l = l;
but we knew all the time that a;=l, therefore the left hand
side is 2, and so 2 = 1.
258 EASY MATHEMATICS. [chap, xxvii.
Instead of going through the above farce, it would be
briefer to say 2x0 = 0;
divide both sides by 0, hence
2=1;
or instead of 2 you may put any quantity you please.
It is a point that may possibly require emphasis, so we will
put it still more evidently :
It is undeniable that 7 x 0 is 0,
and also that 6 x 0 is 0,
if then it be argued that .*. 7x0 = 6x0,
and that the 0 factor may be cancelled out, it seems to follow
that 7 = 6.
It is unsafe then to press the axiom that things which are
equal to the same thing are equal to one another, to cover the
case when " the same thing " is zero.
It is a question whether we have a right to say that 7x0 =
6 X 0 at all, although they are both zero. It rather depends
on what we mean by 0. It is certainly untrue to say that ^ = J
always, because clearly any numbers might be substituted
for the 7 and the 6. Do not, however, assume that J is
gibberish. A meaning can be found for everything if you
are patient and persevering. At any rate, we have no right
to cancel out the zero factor which alone is responsible for the
pretended equality 6x0 = 7x0. Of course the expres-
sion does not in practice occur in this crude form, but it
occurs in some masked form, such as
18 («2-4) = 39 (a;-2),
whence, cancelling out the common factor 3 {x-2), we get
6 (a; + 2) = 13, ori)j = J;
which may, however, be quite false, and is not at all a necessary
consequence of the equation from which it is supposed to
be deduced; it is a possible consequence, or "solution," but
« = 2 is another, and may be the only real one.
CHAPTER XXVIII.
Further Cautions.
Before leaving the subject of "cancelling," it may be well
to append a caution concerning a small point which does
sometimes give trouble to a beginner. The fractions so far
chosen for simplification had both numerator and denominator
composed of factors ; in other words, numerator and denomi-
nator was each really a single " term " : they were not
composed of a number of terms united by the sign + or - .
Compound fractions of this latter kind are more troublesome.
In arithmetic they do not often present themselves in this
form for simplification, because when they occur, the addition
or subtraction can be so easily performed that naturally it is
done before any process of simplification is thought of. But,
in algebra, addition and subtraction are operations that cannot
be done, they are only indicated. Indeed that is one of the
chief advantages of algebra, that the operations to be per-
formed are preserved intact and evident, and are not masked
by the poor achievement of performance.
Suppose then we had ,
^^ 2inx
the whole thing is full of factors, but we may not cancel any.
If only the -\- were replaced by x we could cancel everything,
and leave nothing but unity; but as it is, the fraction is
already in its simplest form, unless indeed we choose to split
it up into two fractions.
240 EASY MATHEMATICS. [chap.
Why may we not cancel anything 1 Because a factor, in
order to be cancelled, must apply to the whole denominator
and to the whole numerator. In the above, there is no factor
which applies to the whole of the numerator. So there we are
stopped. Let us however resolve it into two fractions
371 8a;
2inx 24:nx
and from each of these cancelling is easy, yielding
Sx^Sn
This form may be preferable to the first given form, or it may
not. It depends on what we want to do with it.
Suppose however we take another example, very like the
first, but upside down,
36my
4m + 9y
Still no factors can be cancelled, for there is no factor common
to the two terms of the denominator ; but now we cannot even
separate it into two fractions. The attempt is often made by
beginners ; they try to write it
a splendid simplification certainly, but bearing no resemblance
whatever to the originally given fraction of which it is
supposed by the mistaken beginner to be a counterpart.
The mistake is so often made that it is worth numerical
illustration.
Example (i). ^^. Example (ii). "^^^
144 ^ ' ■' 24 + 7
The first can be split into two fractions
24 _7^_1 _^
144 "^144 6 "^144'
XXVIII.] CANCELLING IN FRACTIONS. 241
The second can not be split up at all. It could be written,
if it were worth while,
6 6
1 + 2V ~ 1-2916
Of course both, being arithmetical, can be written
l44 ^^^ "sT ^^^P®^^^^®^^''
and that is just why these forms do not occur in arithmetic as
they do constantly in algebra.
Is no cancelling ever to be done when a numerator or
denominator contains more than one term^ Certainly there
is, if each term has a common factor. For instance,
Tia + nb 21 + 51
-MT °' '^^ -TOTT-
If the + were replaced by x the n^ would cancel out alto-
gether ; but as it is, only one n cancels out, and the result is
a + b 7 + 17
1 or ■ .
nab 357
I have found beginners who thought that if they used the
factor in the denominator to cancel one of the terms in the
numerator, they could not use it likewise to cancel the other
term; who would wish therefore to divide the 1071 by 9
instead of by 3, and to write it 119 in the result, because a 3
has been cancelled out of each term in the numerator, and
therefore it looks as if a 9 should be cancelled from the
denominator. But there is every difference between striking
out a factor from each of two terms, and striking it out from
each of two factors. The mistake arises in fact from a
momentary confusion between + and x .
When the expression is — — - — ^,
the result is ^ :
mz
L.E.M. Q
242 EASY MATHEMATICS. [chap.
but if the expression had been
rnxx^my
the result would be ^ :
z
but just as the former expression with the + sign would
hardly occur as such in arithmetic, so the latter with the
X sign would hardly occur as such in algebra ; it would be
written — -^
m^z
and no shadow of doubt could arise. The doubt seems to
occur only when there are several terms.
Take the case of more than one term in both numerator
and denominator, like
a + b
x + y'
Can we split this up into two fractions 1 Certainly, but not
ah
into - + - j
X y
the two fractions into which it splits up are
a b
x + y x + y^
the whole denominator occurring in both.
Cautions of a slightly more advanced character.
There is another mistake often made by beginners later on ;
and we may as well mention it here, along with the other
cautions. When we have a simple factor applied to two terms,
like n(a + b),
we may take away the brackets and apply it to each term,
getting the equivalent form
vu + nb,
XXVIII.] OPERATIONS. 243
But although this is legitimate with a factor, it is not legiti-
mate with everything that can occur outside a bracket, — not
legitimate with a symbol of operation for instance ; neither
is it legitimate with a square root or a logarithm.
Thus: n{a + b) = na-\-nby
but J{a + b)\Ja + Jb,
and log (a + b)^ log a + log b.
The sign ^ is to be read " is not equal to " or " does not
equal."
The two root expressions are quite different, and each is
already in its simplest form. To illustrate numerically :
7(4 + 9)^2 + 3,
for J\3, so far from being 5, is something between 3 and 4 ;
because 3^ = 9 and 4^= 16, so ^13 lies between them, and as a
matter of fact is 3-6055513... .
(Never imagine from the accidental repetition of some
figure, like the 5 in this number, that it is going to " recur."
A root cannot possibly be a recurring decimal, for, if it could,
it would be a fraction, and therefore commensurable ; and a
root is always incommensurable, except when it is an integer.
See Chapter XX.).
So again of course (a + b)^ ^a^ + b%
and (x - yf ^^ - y^.
As to log a + log &, so far from equalling log {a + b\ we know
that it equals log (axb), that is log ab.
So also ft^+^Vja^ + a",
but, instead, a'^^ = a* x a".
Ja + Jb is by no means equal to J{ab\ although log a + log h
does equal log ab ; nor does cos x + cos y equal either cos xy or
cos (x + y) ; they are all different. So we learn to be cautious
with symbols of operation and not to treat them as factors
nor to treat them all alike. We have to be very cautious
about the removal of brackets in their case, and must always
244 EASY MATHEMATICS. [chap.
be sure that we understand the meaning and value of the
symbol outside them. Some operations can be treated in this
way, and some cannot ; and we must learn to discriminate.
Before long we shall find that a highly important operator,
denoted by -y-, can be treated in this way ; so that
d , . d d
(uJ^'V)=-j-U + -r-V.
dx^ ' dx dx
And another operation denoted by Idx can likewise be so
treated, so that
l{u + v)dx= ludx+ jvdx;
but these things have to be proved, they must never be
assumed ; and the time for discussing them is not yet.
We may notice however that the familiar symbol of opera-
tion X is one that can be treated in this way
7x(4 + 6) = (7x4) + (7x6),
whereas the symbol -r c&nnot be so treated
7-f(4 + 6)H(7-^4) + (7-6).
Nor can the symbol + be so treated. Anything which can be
so treated is said to be subject to " the distributive law," that
is it may, and indeed must, be distributed among all the terms.
There is another law, spoken of as the " commutative law,"
which is sometimes applicable and sometimes not ; that is to
say it applies to some things and not to others. It applies
when the order can be inverted ; for instance,
axb has the same value as 6 x a.
3 times 4 gives the same number, though it does not
suggest the same grouping, as 4 times 3.
Similarly a-{-b is the same as b + a,
but a - Z> is not the same as b-a;
it is numerically equal but is opposite in sign : an important
distinction.
XXVIII.] OPERATIONS. 245
Nor is a^h the same as 6 ^ a, not even numerically equal.
It is not opposite, but "reciprocal."
Again there is a permutative law :
cxab is the same as a x c& or bx ac,
so also x + y-hz = y + x + z etc. ;
and under certain circumstances, though not invariably,
d d . .
but ajh is not the same as J{ab), nor the same as bja ; the
three things are in fact equal to J((i^b), J{ab), and J{ab^)
respectively.
The expression nlogx is not the same as \ognx, it equals
log (a;").
CHAPTER XXIX.
ILLUSTRATION OF THE PRACTICAL USE OF
LOGARITHMS.
(i). How to look out a logarithm.
Below is given the simplest table of logarithms that can be
used. You can buy four-figure logarithms conveniently printed
on a card, and perhaps you may prefer to use them at once,
because four-figure logarithms are accurate enough for many
practical purposes, and are handy in actual work. But to ex-
plain the method of using a table and the principle of it, without
niceties and details, the annexed table will serve quite well.
You will find this table repeated at the end of the book,
folded in such a way that it is easy to refer to.
Table of 3-figiire Logarithms.
0
1
2
3
4
5
6
7
8
9
10
000
004
009
013
017
021
025
029
033
037
11
041
045
049
053
057
061
065
068
072
075
1-2
079
083
086
090
093
097
100
104
107
111
1-3
114
117
121
124
127
130
134
137
140
143
1-4
146
149
152
155
158
161
164
167
170
173
1-5
176
179
182
185
188
190
193
196
199
201
1-6
204
207
210
212
215
218
220
223
225
228
1-7
230
233
236
238
241
243
246
248
250
253
1-8
255
258
260
263
265
267
270
272
274
277
1-9
279
281
283
286
288
290
292
295
297
299
2
301
322
342
362
380
398
415
431
447
462
3
477
491
505
519
532
544
556
568
580
591
4
602
613
623
634
644
653
663
672
681
690
5
699
708
716
724
732
740
748
756
763
771
6
778
785
792
799
806
813
820
826
833
839
7
845
851
857
863
869
875
881
887
892
898
8
903
909
914
919
924
929
935
940
945
949
9
954
959
964
969
973
978
982
987
991
996
6HAP. XXIX.] USE OF LOGARITHM TABLES. 247
The triple digits which occur throughout this table are the
decimal parts of the logarithms of the numbers on the left and
above. The decimal point is not printed, but it is always to
be understood, and on taking out the triple figures a decimal
point must always be written in front of them.
Now let us use the table to find a few logarithms. The
most obvious record in the table is that
logl =0, log M = -041, log 1-2 = -079,
logl-3 = -114, etc. logl-9 = -279,
log 2 =-301, log 3 =-477, log 4 ='602,
and so on.
Next we have, by the use of the top row of figures combined
with the left-hand column,
log 2-1 = -322, log 2-2 = -342, log 2-3 = '362, etc.
log 3-1 = -491, log 3-2 = -505, etc.
log 4-1 = -613, etc.
and so on.
For all these figures there is nothing more to do than just
extract the logarithms from the table as they stand.
But now suppose we wanted the logarithm of 20 or 30.
We know that log 30 = log 10 + log 3 = 1+ log 3, hence look
out log 3, and write log 30 = 1*477.
Similarly log 20 = 1-301,
log 70 = 1 -845, and so on.
So all we have to do in that case is to prefix a 1 to the
decimal point.
If we wanted the logarithm of 11 or 12 or 13 it would be
just the same, we must prefix a 1 to the decimal point, so that
log 11 = 1-041, log 12 = 1-079, log 13 = 1114.
Similarly
log 21 = 1-322, log 22 = 1-342, log 31 = 1-491, etc.
1 is called the " characteristic " of any number of " order " 1.
248 EASY MATHEMATICS. [chap.
9
Further if we want the logs of 100 or 200 or 300, we must
prefix a 2 to the decimal point because
log 300 = log 100 + log 3 = 2 + log 3 = 2-477.
Similarly log 110 = 2*041, log 120 = 2'079,
log 210 = 2-322, log 220 = 2-342,
log 310 = 2-491, log 320 = 2-505.
So the logarithm of any number consisting of two signifi-
cant figures can be readily obtained from the table, and the
" characteristic " or integer part of the logarithm is given by
the "order" of the number. "Characteristic" and "order"
are in fact two names for the same thing, except that the first
is appropriate to a logarithm, and the second is applied to
the original number.
The logarithm of every number with only two significant
figures is therefore directly contained in the little table printed
above, no matter how big the number may be. For instance,
log 98,000,000 = 7-991.
But suppose the number had 3 significant figures. What is
the logarithm of 215 for instance"? Well, it will lie approxi-
mately half-way between log 210 and log 220. Not exactly
half-way because the number grows in G.P. while the logarithm
grows in A.P., but half-way is near enough for most practical
purposes. So we can see that approximately log 215 = 2-332,
because that is half-way between 2-322 and 2*342.
But suppose the number whose log was wanted did not lie
half-way between others, but only one-tenth of the way ;
suppose for instance log 211 was wanted, we should have to
take one tenth of the difference between 322 and 342, which
difference, being 20, the tenth of it is 2 ; and this would have
to be added on, as representing one-tenth of the interval.
So log 211 would equal 2-324.
xxix.] USE OF LOGARITHM TABLES. 249
We can in fact make an extension of the table for any third
significant figure in the number whose log is required, thus,
log 210 = 2-322, log 211 = 2-324,
log 212 = 2-326, log 213 = 2-328,
log 214 = 2-330, and so on, up to
log 220 = 2-342.
Take a few more illustrations of this.
Wanted log 2-35.
From the table, log 2-30 = -362 and log 2-40 = -380,
so log 2-35 = -371.
Wanted log 3-41.
log 3-40 = -532, log 3-50 = 544,
so log 3-41 = -533 approximately.
Similarly log 3*42 = -534 approximately.
Wanted log 5-63.
log 5-6 = -748, log 5-7 = -756,
.*. log 5-63 = -750,
being three tenths of the interval added on to the smaller one.
Wanted log 5-67. We might add on seven tenths of the
interval to the smaller one, or, rather better, subtract three
tenths of the interval from the bigger one, getting
log 5-67 = -754.
But the table contains more than I have at present described
and used. The first half of the table gives the logarithms of
numbers near to unity, so we can get out logarithms of 1 -01
or r02 etc. up to 1*99, the numbers being expressed to 3
significant figures and all the logarithms recorded. It is a
help to have this given, as a sort of extra, because for these
small numbers the logarithms change so rapidly that the jump
is too great for easy and safe treatment by attending to the
diff'erences, and when we come to look out anti-logs (see next
p^e) they will fall in gaps of too large size.
250 EASY MATHEMATICS. [cha^.
Using this part of the table we see that
log 1 -01 = -004, log 1 -02 = -009,
log Ml = -045, log 1-25 = -097,
log 1-53= -185, log 1-99 = -299.
And consequently
log 10-2 =1-009, log 102- =2009,
login- =2-045, log 125- =2-097,
log 12-5 = 1 -097, log 19-9 = 1 -299,
log 199- =2-299, log 1990- =3-299.
The characteristic of the logarithm is always the " order "
of the number.
(ii). How to look out the number which has a given
logarithm.
To look out the number which possesses a given log we
have only to use the table backwards. It is quite simple and
obvious in idea, the only trouble is that we shall not usually
find the given logarithm actually in the table. If it is an
extensive table we are more likely to find it, and that saves
thought, but involves the turning over of many pages ; with a
little compressed table like the one given, we are not likely to
find a number exactly entered, and a trifle of thought is
necessary. That is no defect however for our present purpose,
which is not immediately to facilitate practice, but to furnish
instruction which shall facilitate practice by and bye.
The phrase " number which possesses the logarithm " so and
so, is rather long and unwieldy, and it is commonly shortened
into anti-log.
Thus log 2 = -301, or 2 is the anti-log of '301.
Given then the following logarithm, -380, what is its anti-log ?
Referring to the table, we see that it is 2-4.
Given -663, the anti-log is 4-6, and so on.
But suppose the given logarithm had been 1 -380, what then ?
XXIX.]
USE OF LOGARITHM TABLES.
251
We should look in the table for the decimal part only, for it
is only the decimal part which is ever there recorded. The
prefix 1 before the decimal point tells us the power of ten in
the result, shows in fact that the result lies between ten and a
hundred. The antilog is therefore not now 2 '4 but 24*
So also the antilog of 1-663 is 46- not 4*6.
The antilog of 2-663 is 460',
of 3-663 is 4600-,
and so on. It is safer to actually write the decimal points at
the end of whole numbers in this sort of case.
The integer part of the logarithm, often called its " charac-
teristic," has simply the efiect of determining the order of
magnitude of the result (p. 171). Surely however a most
important effect, and one not to be slurred over.
Examples.
What is the antilog of 1*672 'i Answer 47*.
What is the antilog of I'SOl 1 Answer 20*.
What is the antilog of 2-041 ? Answer 110*.
What is the antilog of 3-699 1 Answer 5000'.
Employing the upper part of the table we see that
log 1-34= -127,
log 1-01= -004
log 101 =2-004,
log 1-09= -037,
log 109 =2-037.
Likewise the antilog of -111
1-111
2-111
log 1-18 = -072,
log 10-1 =1-004,
log 1010 =3-004,
log 10-9 =1037,
IS 1-29
is 12-9
is 129-
•097 is 1-25.
1-097 is 12-5.
•196 is 1-57.
2-196 is 157-.
252 EASY MATHEMATICS. [chap.
So far the logarithms have been found in the table, because
we chose numbers of only two significant figures. The case
when a logarithm does not occur exactly in the table causes
no difficulty, it only gives a little more trouble.
What is the antilog of '389? Answer 2*45; because it lies
about half-way between '380 and '398, so the answer lies
half-way between 2-4 and 2-5.
What is the antilog of 2-389 ? Answer 245.
What is the antilog of 1-675 1 Answer 47*3.
Why? because the given log lies one-third of the way be-
tween 672 and 681 ; so its antilog will lie about one-third of
the way between 47 and 48. As to the position of its decimal
point, that is determined by the "characteristic" or integer
part of the given logarithm, which was unity.
Logarithms of fractions.
So far all the antilogs have turned out greater than 1,
because all the logarithms chosen have been positive. The
characteristics have all been either 1 or 2 or 3 or 0 ; for the
logarithm -672 has the characteristic 0. It might be, and
often is, written 0-672.
But now suppose it had a negative characteristic ; for
instance 1*672, where the minus sign is placed above the 1
instead of in front of it, in order that it may not be applied to
the whole number, but only to the 1 ; which is a conventional
but convenient mode of representing an important distinction.
The meaning, written out fully, is - 1 -i- -672.
Naturally this might be written, if we liked, - -328, but if
we did that we should want another table full of negative
numbers wherein to look out the logarithms of fractions. By
the above device we can use the same table all the time, and
only adjust the position of the decimal point in accordance
with the characteristic, so that it fixes the " order " as usual.
XXIX.] USE OF LOGARITHMS. 253
For instance,
antilog of '672 from the table is 4*7,
so antilog of 1-672 will be '47,
and antilog of 2-672 will be '047,
antilog of 3-672 will be '0047,
the negative characteristic indicating the position of the
highest significant figure counting from the units' place.
The antilog of 1*672 is of course 47-
of 2-672 470-
and of 3-672 4700-
and so on ; the positive characteristic counting the number of
places to the left of the units' place.
(iii). How to do Multiplication and Division with
Logs.
We know that log ah = log a + log b,
and that log - = log b - log c.
c
So we know that log — = log a + log b - log c,
c
and likewise log —r = log a + log b - log c - log d - log e
= (log a + log b) - (log c + logd + log e).
In other words, whenever we have a fraction consisting of a
number of factors in numerator and denominator, we must
look out the logarithm of each factor. All those in the
numerator, arrange in one column, and add ; all those in the
denominator, arrange in another column, and add ; then
subtract one addition from the other so as to get the logarithm
of the quotient. We have then only to refer to the table to
find the number which possesses this logarithm, and the
quotient is found.
254 EASY MATHEMATICS. [chap.
For instance, take this fraction
6-7 X 43 X 170
74x3-2xl-3"
Look out the logs of the factors as stated :
log 6-7= -826 log 74- =l-869
log 43- =1-634 log 3-2= -505
log 170- = 2-230 log 1-3= -114
add 4-690 add 2-488
subtract 2-488
2-202
2-202 is therefore the log of the resulting quotient. Eeferring
to the table, we find that the number possessing this log is
159-2. Hence that is the answer.
It may be asked, why do it this way when we could easily
do it by simple multiplication and division 1
Reply : Very little, if any, advantage in such a simple case
as that. No advantage at all if you can easily see factors
which may be struck out.
But people who often have to do such sums get rather tired
of frequent multiplication and division, and they usually prefer
logarithms as a quicker and surer way. It becomes quicker
and surer with practice. Engineers usually employ what is
really the same process, but they have their table of logarithms
constructed in wood; and instead of looking out the logarithms,
they slide a slip along this rule, till a mark on it points
to the number printed where its logarithm ought to be,
and so attain the result in an ingenious manner, without
actually recording or thinking of any logarithm at all. They
shift the pointers, of which there are a pair, alternately to one
factor after another, taking numerator and denominator factors
alternately, and then at the end they read off the result as
indicated by one of the pointers.
XXIX.] USE OF LOGARITHMS. 255
The instrument is called a "slide-rule"; it is in fact a
mechanical table of logarithms arranged ingeniously for quick
and practical use, and it gives you about 3-figure accuracy if
it is of a simple and well made pocket kind. More elaborate
and larger instruments can give 6-figure accuracy. The in-
genuity belongs to the devising and making of the rule : the
use of it is quite simple, but it has to be learnt. It should not
be learnt as a substitute for other methods, but as a supplement.
Pupils are not recommended to learn the slide-rule till they
can use a numerical table of logarithms. Nor are they recom-
mended to use logarithms till they can multiply and divide
with facility. In other words, these aids to rapidity should
be kept in their proper place, — not to make people helpless
without them, but to assist people who can work quite well
without them to obtain results more quickly and with less
labour.
Another example. Find the value of
27-1 X -16 X -089
•00055x3430 *
Look out logs and record them as below :
log 27-1 =2 '^^^ log '^0^^^ = 5'*^^^
log -16 =1-204 log 3430- =3-535
log -089 = 2-949 0»275
T-586
subtract 0-275
1-311
Antilog of this is -205, which is therefore the result, and may
be recorded as equal to the above fraction to something like
3-figure accuracy. This should be checked by actual multi-
plication. Indeed for some time, and especially when there
are negative characteristics, it is safest to check over the
result by other means than the inere logarithms. It is the
256 EASY MATHEMATICS. [chap.
order of magnitude that runs great risk of going wrong : the
actual digits can easily be got right.
A few more examples for practice :
What is the log of -05 1
And what is the antilog of I'89 ?
The log of 5 in the table is '699,
so the log of -05 is 2-699.
The antilog of '89 estimated from the table is about 7*77,
so the antilog of T'89 is -777.
What is the antilog of -049 ?
We find this number in the upper part of the table, and
the antilog required is 1-12.
What is the antilog of 2-049 1 Answer 112.
T-0491 Answer -112.
2-0491 Answer -0112.
What is the antilog of -023 1 It does not occur even in the
upper part of the table, but it lies half-way between two
numbers which we find there; so we estimate the antilog
as 1-055.
The antilog of 3-023 = 1055'
ofT023= -1055.
This is the use of the upper part of the table, as previously
half explained, that it gives us the logarithms of numbers only
slightly greater than 1 in greater profusion ; and it is just
here that profusion is necessary, for in other parts of the
table logarithms lie much closer together in value than they
do here. Consequently what would naturally be the first row
of the table is spread out into nine rows, the first row itself
becoming thereby a column, reaching from Ito 2, and giving
all the tenths of this interval.
If a beginner likes to think out the reason and meaning of
the different closeness of distribution in various parts of a
XXIX.] USE OF LOGARITHMS. 257
logarithm table, he should by all means do so, but he need
not be made to do it. The reasonableness of it can be put
thus:
All the 900 integers between 100 and 1000 have logarithms
lying between 2 and 3 : this unit difference of logarithms is
therefore spread over all that range ; while the same loga-
rithmic interval, viz. that between 0 and 1, has to be squeezed
between the numbers 1 and 10, covering only nine consecutive
integers.
The logarithmic interval between 1 and 2 has to serve for
the 90 whole numbers between 10 and 100, while the same
logarithmic interval, viz. between 3 and 4, is all that can be
allowed to cover the 9000 numbers between 1000 and 10000.
Hence manifestly the logarithms of integers between 1 and
10 must be few, and the intervals between must be great,
though they may be conveniently filled up with the logarithms
of intervening fractions; but the logarithms of integers
between 1000 and 10000 are close together, their value
increasing only slightly for each addition of unity to the
number. In other words, the logs of integers take 9000 steps
to go from 3 to 4 ; they only take 9 steps to go from 0 to 1.
The one is a trip, the other is a straddle.
L.E.M. R
CHAPTER XXX.
How to find powers and roots by logarithms.
The finding of any power, or any root, is now an extremely
simple operation.
We know that log x" = ?i log x, and this holds whether n be
an integer or any fraction.
In other words, as said before,
log x^ = 2 log X,
log x^ = 3 log X,
logJx = logx^=^^\ogx,
log ^a; = log a;^ = J log a;,
and so on. Hence the method suggests itself, and we need
only proceed to examples.
To find the value of ^2, the logarithm of it will be half the
logarithm of 2, and that we look for in the table, and find to
be '301, so half of it is -1505. This we do not find exactly in
the table, but we see that it is the logarithm of a number
lying between 1-41 and 1-42, and Ave estimate the number as
being 1'414. This is of course only an approximation,
because no arithmetical specification of it can be anything
but approximate. If calculated more elaborately, it comes
out 1*4 142136..., but it can neither stop nor circulate.
Similarly ^3 = 1*732 approximately,
or more nearly 1 -7320508 . . . ,
again without either recurrence or termination.
CHAP. XXX.] USE OF LOGARITHMS. 259
Now take a case of a power. Suppose we want to calculate
the 2^4 involved in an example on page 156. Its logarithm
will be 24 log 2 = 24 x -301 = 7*224 and we have only to look
out the number which has this logarithm, that is look for the
antilog of 7-224. We shall find 225 in the table (p. 246), and
that is really better than 224, because when we multiply a
number by so big a factor as 24 there must probably be some
carried forward figure to be attended to. Anyway we find
that '225 is the logarithm of 1 •68. This is not the result, of
course, since we have not yet attended to the characteristic,
which is 7. The characteristic is indeed, in these big numbers,
usually the most important thing to notice. The charac-
teristic 7 shows that the number is of the order seven, i.e. that
it lies between 10'^ and 10^; in other words, that it requires
eight digits to express it, and so it is approximately
16,800,000,
that is sixteen million eight hundred thousand, so far as we
can express it with 3-figure accuracy. There are eight digits
in this result, but only three of them are "significant," the
others are mere ciphers to indicate the order of magnitude.
The neatest way of recording such a result is therefore
1-68x107,
and the characteristic of the logarithm will always give us the
index of the power of ten when the number is so written.
For instance, antilog 19 '330 = 2 -1 4 x 1 O^^,
antilog 6 -552 = 3 -56 X 1 0-6,
antilog 2-950 = 8-92 x 10-2= 0892.
The operation of finding a root will look thus :
To find the fifth root of 1930.
log 1930 = 3-286,
ilogl930 = 0-657 = logof 4-54...
wherefore 4-54... = ^(1930).
260 EASY MATHEMATICS. [chap.
Observe that when dividing a logarithm the characteristic
is to be included in it and divided with the rest of it. It is
only in dealings with the " table " that the characteristic does
not appear. It should however always be supplied and should
not be forgotten or ignored.
Thus if we had wanted the fifth root of 193 or of 19-3 or
of 1 93 we should have obtained a totally different number :
not the same number with the decimal point shifted, but a
different number altogether. For instance,
log 193 = 2-286,
1 log 193= -457 = log of 2-865.
log 19-3 = 1-286,
I log 19-3= •257 = logof 1-81.
log 1-93 = 0-286,
^ log 1-93= -057 = log of 1-U.
But now suppose we required the root of a fraction, i.e. of
something whose logarithm was negative. We must think
how to proceed in that case. Suppose for instance we want
the fifth root of -193,
log -193 = 1-286,
that is to say a negative part and a positive part ; it means
- 1 + -286.
In order to divide this by 5 conveniently, it is best to increase
both the negative and the positive parts by any convenient
equal amounts : in this case the convenient amount is 4.
Add - 4 to the negative part, and add + 4 to the positive part:
the value will thereby be unaltered, but it will be written as
-5 + 4-286
and now it is quite easy to divide by 5, yielding
-1 + -857 or T-857,
which is the log of -72. Wherefore
•72=^-193.
i
XXX.] USE OF LOGARITHMS. 261
The root is bigger than the number. That is universal with
roots of proper fractions. When we square a fraction we
diminish it ; when we square-root a fraction, consequently, we
increase it. Think it out; it is all in accordance with common-
sense.
But we must take another example.
Let us find the square root of •0054.
log -0054 = 3-732
= -4 + 1-732,
Jlog-0054= -2+ -866
= 2-866 = logof -0735,
wherefore -0735 = ^-0054.
We might have made an approximate guess at this, because
^•0049 could have been written down as '07 by inspection,
and so ^'0054 will be a little bigger ; how much bigger it is
not so easy to guess.
But suppose we had wanted J'Obiy we should have found
nothing like a 7 in the root. Let us do it :
log -054 = 2-732,
1 log -054 =1-366 = log of ;232.
So -232 =^054, whereas -0735 = ^-0054,
-0232 = 7-00054, -735 =^-54,
2-32 =75-4, 7-35 =^54,
23-2 =V540, 73-5 =^5400.
A little easy repetition on this point may be useful so as to
emphasise it.
^49 =7, and ^100= 10;
so V4900 = 70,
and ^490000 = 700 ;
V-49 =-7,
V0049 = -07,
V-000049 = -007 ;
262 EASY MATHEMATICS. [chap.
but if instead of two ciphers we had suffixed or prefixed only
one cipher, we should have had quite different results, and not
so easy to ascertain, viz. the following :
74-9 = 2-214,
V'490 = 22-14,
V'49000 =221-4,
^•049 = -2214,
^•00049= -02214. .
Exercises.
^407 = 7-41
because log 40 7 = 2 -6 1 0,
and one-third of it is -870 = log of 7*41.
Similarly work out the following :
^•407 = -741,
^407000- =74-1,
^-000407= -0741,
so that whereas for square roots the noughts can be added in
pairs to leave the digits unaltered, for cube roots the ciphers
must be added in triplets if they are to make no change in the
digits. This is an immediate consequence of the fact that
^1000 = 10.
The last case, for instance, works out thus :
log -000407 = 4-61 = - 6 + 2-61,
of which one-third is 2-870 = log of -0741.
The simplest way of dealing with these things however is
to express them in powers of ten.
Thus -000407 = 407 xlO-«,
so its cube root is
4/407 X 10-2 = 7-41 ^ 100 = -0741.
But now suppose the digits had not been added in triplets.
Find cube root of 40-7.
log 40-7 = 1-61,
XXX.] ROOTS OF NEGATIVE NUMBERS. 263
a third of that is '537, which is the logarithm of 3-44... which
is therefore the root required.
Again, to find ^4-07.
log 4-07 = -61,
one-third is 2033 and the number corresponding to this log is
nearly TG.
So (4070)^=16 nearly; more accurately 15*966...
Also (-00407)^ = -16.
Find the cube root of '0078.
We may ^vrite it as 7*8x10"^,
and so express its cube root as
1-98x10-1 = -198.
Find the cube root of -000000081.
Express it as 81 x 10"^.
Its cube root is 4-33 x lO'^ = 00433.
Roots of negative numbers.
Perhaps it is not likely to occur often in elementary practice,
but it is worth noticing that the cube root of a negative
number is by no means impossible. What, for instance, is the
cube root of - 8 ; that is, what number multiplied twice by
itself will make - 8 ] The answer is - 2, for
-2x-2=+4 and +4x-2=-8.
So y-27=-9, y- 1728= -12, and so on.
Also y- 407 =-7-41, see above.
The square root of a negative number has no simple meaning.
If we tried to find the square root of - 9 or - 25 we could not
do it, for - 3 X - 3 = + 9 and - 5 x - 5 = + 25. Hence nega-
tive numbers have no square roots, but they have cube roots.
Having no square roots of course they cannot have fourth
roots, for a fourth root is simply the square root of a square
root. But they have fifth roots and seventh roots and any
odd numbered roots, because an odd number of minus signs
264 EASY MATHEMATICS. [chap. xxx.
multiplied together make minus. Negative numbers have no
even roots.
This is not all that can be said concerning the roots of
negative numbers, by any means : Pure mathematicians know
a great deal more than that about them ; and later, children
who like the subject may learn some of it, but not yet. In
order however to prepare them for a convenient way of
dealing with the matter, I will point out that any negative
number can be said to have - 1 as a factor ; for instance,
- 8= 8x -1,
-16 = 16x -1,
- 27 = 27 X - 1, and so on.
Hence any root of any negative number is equal to the same
root of the same positive number multiplied by the appropriate
root of - 1. For instance,
y-8 = ^8x4/-l = 2^-l,
^-27=y27x4/-l = 34/-l,
^-32 = 2^-1, and so on.
[Rememher thai,tj(xi/) = jxjy or that (a5)" = a"6^]
But the same method may be extended to even roots, thus
V-16 = V16xV-l = V-l,
V-9 = 3V-1,
4/- 81 = 3^-1,
4/-64 = 2«/-l, and so on.
It is true that we do not yet know what to make of J- 1
or 4/- 1 or y- 1 ; it is an impossible or imaginary quantity ;
but though we think that we do know what to make of I/- 1
or ^ -I, viz. although we know that they = - 1, do not let
us be too sure that we know all about even these. It is at
any rate true that -Ix -Ix -1= -1, and that is all that
need now concern us ; but it is not, strange to say, the whole
truth concerning even the odd roots of minus one.
CHAPTER XXXI.
Geometrical Illustration of Powers and Roots.
Geometrical illustration, or illustration of number by-
simple diagrams, cannot be pressed very far with advantage for
elementary purposes. But for simple things the illustrations
are so vivid and useful and interesting that they should
often be employed, and especially be set as exercises so as
to infuse life and interest into what might otherwise be dull.
The simplest illustration of all relates to the squares or
cubes of integers. That the square of 3 is 9 is illustrated in
the most conspicuous manner by the diagram.
Fia. 12.
So also that the square of 4 is 16, and the square of 5 is 25.
That the cube of 2 is 8 is illustrated thus,
but the best plan of dealing with solids is
to use cubical wooden blocks and build
them up.
8 blocks will build a cube whose side is 2
27 „ „ „ 3
64 » » » 4
and so on.
/ /
/
/ / /
/
/
/
y
Fio. 13.
266 EASY MATHEMATICS. [chap.
The same blocks laid flat on the table will serve con-
veniently for squares and rectangles and commensurable areas
generally. They will also serve to outline commensurable
triangles : with conspicuous advantage in some cases.
By this kind of practice a reality about square and cube
numbers is attained which can be got in no other way.
Naturally also the area of any rectangle can be thus
illustrated as the product of length and breadth ; and the
volume of rectangular solids as the product of length, breadth,
and height.
If we try to illustrate fourth or higher powers in this way
we shall find ourselves helpless. Space is only of 3 dimensions.
There are length and breadth and thickness, and no more.
Some have tried to imagine what a fourth dimension would
be like, but for the present we will be content with an actually
experienced and familiar three dimensions.
So much for powers ; now what about roots ?
The few commensurable roots that exist must all be whole
numbers, and they will be represented, so far as square and
cube roots are concerned, by the length of the sides or edges
of the squares or cubes which have so far been drawn or built
up. Thus, for instance, the square root of 16 is 4, and the
cube root of 27 is 3. But this fact, which is experimentally
obvious in the commensurable case, where the square or the
cube can be built of blocks, is true also in the general case.
The length of a side of a square is the square root of its area
always, and the length of the edge of a cube is the cube root of
its volume always. This represents the geometrical
notion of a root so far as geometry can illustrate it.
We will now proceed a little further.
Suppose we take a square and draw a diagonal
Fio. 14. across it, what is the length of that diagonal 1 It
is evidently greater than a side, and not so great as two sides.
XXXI.]
GEOMETRICAL ILLUSTRATIONS.
267
Fig. 15.
I
If we measure it carefully we shall find it rather less than
a side and a half. It will be about one and two-fifths or 1*4
times a side.
Now construct a square on the diagonal, i.e. a fresh square
with the old diagonal for one of its sides. We may not know
how to do it accurately on blank paper, but it is quite easy
to do if we use paper ruled faintly in squares, such as can
easily be obtained in copy-books. Or the
figure may be constructed by folding over
the tongue of a sort of square envelope.
In any case it is quite easy to see that
the square on the hypothenuse is twice
the area of the square on either side of
the isosceles rt.-angled triangle. For
produce the sides along the dotted lines.
The larger square is thereby cut up into four parts each of
which is half of the smaller square : see fig. 15. Therefore
the areas of the squares are as 2 to 1.
But the side of a square is the square root of its area, hence
a side of the new square is ^2 times a side of
the old one. In other words, the diagonal of a
square is ^2 times the length
of one of the sides.
Or, expressing it otherwise,
the hypothenuse of an isosceles right-angled
triangle is J'2 times either of the sides.
If we were to draw a square on one of
the sides and a square on the hypothenuse,
the two squares would be as 2 to 1.
(The area of the triangle itself is evi-
dently i on the same scale.)
Drawn thus, we might not see how to
prove it, but drawn as in the previous figure the proof is
268 EASY MATHEMATICS. [qbA^.
obvious. To be sure that there is no mistiness about it, a
beginner should write the proof out for himself, expressing it
as well as he possibly can. The inventing and writing out of
proofs is good exercise, and to do it really well demands some
thought and a little skill. The skill so cultivated is of a
useful kind in life.
An example is necessary ; but the danger of an example
is that it is apt to become stereotyped. It may "be varied
in innumerable ways, and a way invented by the pupil is
better than one which he has to learn. If there are actual
errors in his proof they can be pointed out, but defect of
taste and style, though much to be deprecated in adult
persons, must be eliminated gradually from a beginner. He
cannot be expected to concoct a proof in finished style from
the first.
Something like the following would be good enough : —
To prove that the square on the hypothenuse of an isosceles
right angled triangle is double of the square on either of the
sides.
Construction. — Draw the triangle ABC with right-angle at C\
so that AC^ BC are the equal sides, and AB the hypothenuse.
P g Now draw a square on AC, and draw
it so that the equal sides of the triangle
shall serve as two of the equal sides of
the square. That is draw the square
ACBR
Next draw a square on AB, and draw
it so that C lies in the middle of it,
which is best done by producing AC bh
equal length to E, and producing BC
an equal length to F, and then joining up so as to make the
square ABEF, which, being a quadrilateral figure with equal
diagonals at right angles to each other, must be a square.
f ""'7
^
\
\
1/
^_
\\^
C /E
^x j ^'
F
Fig
18.
XXXI.] GEOMETRICAL ILLUSTRATIONS. 269
Proof. — The square so constructed contains the area of the
original triangle four times, while the former square contains
it only twice. Therefore the square on the hypothenuse is
double the square on one of the equal sides of an isosceles
right-angled triangle, q.e.d.
Beginners can and should realise the fact, immediately and
without words, by having given to them small triangles in
wood, and by then piecing them together so as to make the
above figure. In a short time, left to themselves, the realisa-
tion becomes vivid.
Now proceed to a right-angled triangle with unequal sides.
Suppose as a special case the hypothenuse is double one of
the sides. It is not difficult to devise a way
of drawing this figure if we use a pair of
compasses.
For let AB be one of the sides. Double
it and you get AC. With centre A and
radius AC mark off a circle. This gives the
length of the hypothenuse.
At B draw a line perpendicular to AB
till it meets the circle at D; then join A and D. The triangle
ABD is the triangle required, viz. a right-angled triangle with
its hypothenuse double one of the sides.
If we were to draw a square on AD and another on AB^
the area of the one square would be quadruple that of the
other; because the sides are as 2 to 1, therefore the squares of
the sides will be as 4 to 1.
What about a square drawn on BD% If drawn and
measured it will be found to be about f of the big square.
It can be shown by geometry that its area is exactly three
quarters of the big one. In other words, that the middle
sized square and the small one added together exactly equal
the big square in area. This is a most curious and important
870 EASY MATHEMATICS. [chap.
fact : about as important as anything we have come across, if
it applies, as it does, to all plane right-angled triangles without
exception. But they must be plane triangles ; i.e. the sides
must be straight.
We are not yet supposed to know how to prove it. We
can verify it approximately by drawing the squares carefully
and cutting them out in wood or cardboard or sheet lead,
and then weighing them. The two smaller squares will be
found just to balance the big one, if they are cut out neatly
and if the sheet was uniform in thickness and material. This
is not a proof that they are mathematically equal, but it is a
verification that they are approximately equal, equal "within
the limits of error of experiment."
That is a kind of equality by no means to be despised. In
some difficult cases it is all the equality that can be ascertained.
In the present case it is by no means all ; but no proof of
exact equality can be obtained by empirical or experimental
processes, no matter how carefully they are carrried out.
Exactness is a prerogative of mathematical reasoning, that is
reasoning on pure abstractions from which all flaws and imper-
fections and approximations are by hypothesis eliminated.
The fact that the squares on the sides of any right-angled
triangle are together equal to the square on the hypothenuse,
was known to the ancients. It was called the theorem of
Pythagoras ; and a classical proof, a fine example of ingenious
reasoning, is given as the 47th proposition of the collection of
geometrical propositions made by the Greek Geometer Euclid
in his first book. Translations of that ancient treatise are
sometimes still learnt by schoolboys in this country, and may
be considered a part of classical education. It is an antiquated
and slow way of learning geometry however, and in fact can
hardly be intended seriously for that purpose. Nevertheless
it is a delightful literary work and pleasant for reference.
XXXI.]
GEOMETRICAL ILLUSTRATIONS.
271
People who are not acquainted with it are hardly educated in
the usual English sense.
Many proofs can be given of any proposition, and the fact
itself is of more importance than any one proof of it.
That does not for a moment mean that proofs can be dis-
pensed with, for without a proof we should not really know
the fact. We could know it approximately but not rigorously
and exactly; and it should be always a joy to feel that a
theorem or a statement can be made without limitation or
approximation. Such statements are the only ones that can
be pressed into extreme cases, with perfect confidence that
whenever applicable, that is whenever the postulated data are
satisfied, they will be always precisely true.
What we are doing at present however does not necessarily
demand extreme accuracy. We have been finding roots, which
we can only do approximately, and we now want to illustrate
them. It will suffice for our present purpose if we assume
Pythagoras's theorem as experimentally verifiable with suffi
cient accuracy for our present purpose, and proceed to use it.
The most remarkable of all right-angled triangles is the
one whose sides are all commensurable, namely 3, 4 and 5.
The square of 5 is equal to the sum of the squares of 4 and 3.
25 = 16 + 9. See fig. 20.
Of course the sides might equally
well be 6, 8, and 10 ;
also 9, 12, and 15;
12, 16, and 20;
and so on.
Also they could be 1-5, 2, 2*5 ;
•75, 1, 1-25;
and so on.
So long as the proportion holds, the absolute length of the
sides is only a matter of "scale."
Kyv
)C^4
/ 3
Fio. 20.
272
EASY MATHEMATICS.
[OHAP.
There is no other commensurably-sided right-angled triangle
until we come to the one with sides 5, 12, 13 ; and the next
one has sides 8, 15, 17. [See Appendix.]
Triangles with commensurable sides can be outlined by
children by surrounding them with square blocks or slabs;
and it is especially instructive to outline right-angled triangles
in this way, because then the squares on the three sides can,
after suggestion, be completed, and the number of blocks in
each counted : When it will be perceived that
9 + 16 = 25, 144 + 25 = 169, 225 + 64 = 289;
a fact which ought to arouse some curiosity, since it
represents the first inkling of one of the most simple
fundamental and universal truths in existence.
What we have learnt by assuming Pythagoras's proposition,
so far, enables us to say that in a right-angled triangle with
the hypothenuse double the base the vertical side is ^3 times
the base. For the squares on each are as 1:3:4; therefore
the sides are as ^1 : ^^3 : ^4, that is as 1 : ^3 : 2.
If the hypothenuse is treble the base, the squares will be
as 9 to 1, and so the square on the vertical
side will be represented by 8 on the same
scale, and the vertical side itself will be ^8,
which equals 2J2.
This should be examined and verified.
It will be easy for a beginner to devise a
verification of it. For instance, thus :
Draw any vertical AB. Draw half a
square on it, as shown, ADB,
Take half one of these sides, and lay it off horizontally, BC.
Then, this being called 1, BD ov AD will be 2, and AB
will be 2^2 or ^8 ; and so therefore AC should equal 3 on
the same scale, because 8 + 1 = 3^. See if ^C does measure 3.
Fig. 21.
XXXT.3 GEOMETRICAL ILLUSTRATIONS. 273
Further geometrical methods of finding square roots.
Let us now attend a little more carefully to the important
statement that the square root of the area of a square is the
length of one of the sides. We have seen that it is true
numerically, now see if it is true and sensible physically. The
point to attend to is that the square of a length is an area,
and the square root of an area is a length, not proportional to
a length or numerically represented by a length, but actually
and physically a length.
Ja'^ = a.
a being a length, a^ is an area, and Ja^ is therefore a length
again. But there is no reason why the area need be square.
Suppose it were oblong, and given as axh; if J{ah) = a-,
X would still be a length. What length 1
Answer. The length whose square is equal to the product
a6, the geometric mean of the two lengths a and h ; for if
Jab = c,
ah = c^,
and so - = - ot a:c = c:b,
c b
or the three quantities a, c, h are in geometrical progression,
for they dififer by a constant factor, viz.
r = - = -
a c
They might be written -, c, cr.
r
The term c is the mean of the other two terms in the G.P.,
so it is called their geometric mean.
Can it be found geometrically 1 It can, and this is another
most interesting proposition known to the ancients and
recorded by Euclid, It is called the 14th proposition of his
second book. Though perhaps not easy to prove, it is
pxtremely easy to state. We will state it now and prove it
274
EASY MATHEMATICS.
[chap.
Fig. 22.
later. The statement without a proof is a poor thing, but the
statement as a prelude to the proof — a
statement which shall provide a niche
for the proof in the mind of a beginner
and cause him to welcome it when it
comes — is an excellent thing.
Construction for finding the geo-
metric mean of two lengths.
Lay off the lengths end to end as
JB + BC.
Draw a circle on the combined lengths as diameter, and
erect a perpendicular at the junction-point B till it meets the
circle in D ; then BD is the geometric mean of JB and BC.
The figure shall be repeated below, with the lengths labelled,
and the rectangle ab shown. (Fig. 23).
"Geometric mean" is an arithmetical
or algebraical sort of term. What will
it mean geometrically^ It will mean
that the square on BD has to equal
in area the so-called rectangle AB.
BOf which means the real rectangle
AB . BE. That is to say the square
on the length Jab has to equal the area ab.
That is precisely what remains to be proved.
D If c2 = ab
then c is the geometric mean of a and J,
a c
or - = V.
c b
The only practical difficulty is how to find
the length c, and that is overcome in a very
simple manner by the circle in the above
construction.
If any one has got as far as the 35th proposition of Euclid's
FiQ. 23.
XXXI.]
GEOMETKICAL ILLUSTRATIONS.
275
Fio. 25.
third book, they can devise a proof of this curious and very
important property of the circle for themselves ; in fact the
figure annexed suggests it at once, as soon
as we know that the rectangles contained
by the segments of two chords are equal.
Given this simple and beautiful con-
struction, we can at once find a length
numerically representing the square root
of any given number n ; for we can take
the two initially given lengths as n and
1 respectively, so that their product is n, and the geometric
mean will then represent the square root of w, because it
will be equal to J(n x 1). (Fig. 26).
For instance to construct ^4.
Take a line 5 inches long as the dia-
meter of the circle, mark off 4 inches
and draw a perpendicular to meet the
circle; this will be ^4, and if measured
will be found to equal 2 inches.
To find *y5 geometrically.
Draw a circle of radius 3 inches, so that its diameter is
6 inches. At the first inch draw a perpendicular and measure
its length. That will be the root of 5.
It should equal 2*236 inches if carefully
drawn and measured.
For the root of 7 the same construc-
tion exactly is to be carried out, only
the circle will be 4 inches in radius.
For the root of 2 the circle will be
1*5 inches radius, or 3 inches diameter.
And for the root of any number whatever, ??, the radius of
Fig. 26.
Fio. 27.
the circle will be
n+l
276
EASY MATHEMATICS.
[OHAP. XXXI.]
For roots of large numbers, this method will not be con-
venient, but for roots of fractions not too far removed from
unity it serves well.
For instance, to find the root of 3-6. Take a circle of
2*3 inches radius, and make the construction, erecting a
perpendicular to the diameter at the end of the first inch.
Its length gives the root, and should equal 1 -9 inches.
Fia. 29.
To find the root of -75.
Take a circle J x 1-75 = -875 inch radius, and at the first
inch of it erect the perpendicular.
Its length will be greater than "75, as necessary for the
root of a proper fraction, and it should equal -866.
This particular result could however have been still more
easily calculated, or at least expressed in terms of ^^3; for
•75 = }.
so
7-75 = ^= J73 = 1x1-732... = -866....
It must be understood then that a geometrical construction
in these cases, though it may be regarded as a simple method
of arriving at the result, is more particularly an illustration of
a result otherwise arrived at. This is however not always
the case, and sometimes by construction results can be found
which it would be extremely difficult to get in any other way.
Engineers and building constructors know this well : and
graphical methods are in constant practical use.
CHAPTER XXXII.
Arithmetical method of finding Square Roots.
We now know three methods of finding a square root.
1. The factor method, when it is applicable, which it seldom
is ; whenever it is easily applicable it should be used. Often
it becomes a matter of guessing and trial and error, with
the error gradually corrected or diminished.
2. The logarithm method, which is the real practical plan,
and is frequently done with a slide rule.
3. The graphical method.
4. There is however another, an arithmetical method, which
is usually learnt, though seldom really employed. It is an
ingenious plan and is not at all bad for finding square
roots. For cube roots it gets complicated, and for higher
roots like fifth and seventh it would be altogether too
difficult for anyone but a mathematician, and he would
never think of employing it.
To find a really high root, for instance a 9th root, the
logarithm method is the only reasonable one; though we
might take the cube root twice over. A sixth root is the
square root of a cube root. An eighth root is the result of a
square root operation three times repeated. An eleventh root
I could only do by logarithms, and with them it is so easy
278 EASY MATHEMATICS. [chap.
that nothing better is needed. Let us see, for instance, what
is the eleventh root of 2,
^^••' = -0273664 ... = log of 1-06503 ... ,
which is therefore the root required.
[If any part of such an answer as the above pretended to
" circulate," we should know that the recurrence was spurious,
and only due to the fact that not enough digits in log 2 had
been taken into account. Roughly speaking we may say
that all numbers are incommensurable, except those specially
selected to be otherwise.]
Why then learn any arithmetical method for finding square
roots, other than the logarithm method 1
Answer. Because we might not have a table of logarithms
handy, and because it is ignominious to be dependent on
material tools except in operations which are complicated.
To find a cube root by direct process is rather complicated,
and I do not recommend its being learnt except by enthusiasts :
and they will forget it again. But the rule for square root
is fairly easy and often useful. It will however be the
hardest thing we have attempted yet, and the proof will be
deferred to the next chapter. It is not usually considered
hard, but all the things before this have been easier in
reality, though people often shy at them. I hope they will
do so no longer.
To find the square root of 256 by direct arithmetic. Set
it down like a long division sum, but with the digits marked
out in pairs, by dots or commas or other marks, as shown,
beginning with the units place, then work as follows :
1)256(16
1
26 ) 156
166
xxxu.] RULE FOR SQUARE ROOT. 2*79
First guess the square root of 2, or the integer smaller.
It is 1, so put it in two places, and multiply and subtract as
in long division. Then double the 1, and place it on the left
as 2, and see how many times it will go into something less
than 15 ; guess 6.
Set down 6 in two places as shown, multiply and subtract,
and there is no remainder. The sum comes to an end : the
root is 16.
If we had guessed 7 instead of 6, as might seem natural,
the product treated as above would have been 189, and been
too big.
If we had been given the number 2560, it would have been
dotted off in pairs as follows :
2566,
and the result would have been quite different. We should
now have to guess the root, not of 2, but of 25, which is very
easily done. The process would then have looked like
this :
5 ) 2S66 ( 50-6
25
1006 ) 6000
6036
-36
so that 50-6 is approximately the square root of 2560.
The small remainder shows that the result is not quite
accurate, and its negative value shows that the result is
slightly in excess.
(Observe that ciphers, like the other figures, are always
brought down in pairs. If it were a cube root we were finding
they would be brought down in triplets.)
It is natural to put 6 in the second stage, after the 0, as we
have done above, because it is very nearly right. It is a little
too big however, and if we wanted to work the root out
%80 EASY MATHEMATICS. [chap.
further, we should put 5 and be sure that the next figure
would be 9.
A more exact result is 50-5964426....
To find the square root of 6241.
Set it down, and again partition off the figures in pairs, be-
ginning with the units place, by dots or other marks, as shown
7 ) 6241 ( 79
49
149 ) 1341
1341
Guess the root of 62 or the next lower integer; guess 7.
Set it down in 2 places, multiply, and subtract. Double 7,
and see how many times it will go in 134 ; guess 9 times.
Set down 9 in 2 places, multiply as shown, and subtract.
There is no remainder : the root required is 79 exactly.
We might have guessed this. Looking at the number we
see that the root will be less than 80, for 80^ = 6400. But
it will not be much less than 80, because a moderate
difference in a square is but a small difference in the root.
So we might try 78. Multiplying out, we should find
78 X 78 = 6084, which is about as much too small as the other
was too big. Hence we know that it is either 79 or something
very near to 79.
Take another instance of guessing : choosing a number quite
at random, say ^(596). We know that
242 = 4x122 = 4x144 = 576,
while 252 = 625. So here again the number lies about half
way between 24 and 25, but a little nearer the smaller of the
two ; and we might see how 24*4 would answer.
Multiplying 24-4 x 24-4 we should get 595*36, which is very
close. As a matter of fact,
V(596) = 24-4131112...
XXXII.]
RULE FOR SQUARE ROOT.
281
which you can proceed to ascertain by the arithmetical process
worked out at length, as thus :
44
596 I 24-41311123
4
196
176
484
2000
1936
0
1
900
469
48^
4885
!3
640
488
151
146
48826
48826
)1 5
4
43100
88261
121
5483900
4882621
48826221
488262222
60127900
48826221
1130167900
976524444
153643456
We see that the next digit will be 3, and have placed it
in position, but we consider that as we have now obtained
ten significant figures, we have gone far enough, especially
as we know that there can be no end.
If we have to find the square root of a decimal, we can mark
it oif into pairs, as before, always beginning with the units
place. Thus 17*8534 is marked off properly for the purpose
of extracting its square root, which is plainly 4 decimal
something.
So also 6-060576 is properly marked off, and its root is '024.
The marking off in pairs is manifestly connected with the
fact that 7100 = 10. It is to get the power of ten in the
282 EASY MATHEMATICS. [chap, xxxit.
answer right. The number of dots gives the number of figures
in the answer, if the units place is included in it. To find a
cube root, the dots would be placed on every third digit,
but always beginning with the units place, because any root
of 1 is 1.
There is not much more than this to be learnt about this
ingenious and practical process, until we are able to prove it
and see the reason of the successive steps : this will be fully
attended to in the next chapter, pages 296 to 299. There are
however a great number of far more important things, and I
only place this brief record of the process here, because
I by no means wish to extrude it; moreover it is an in-
teresting thing to prove. It is essentially a limited process,
however, since, for any useful purpose, it only applies to
square roots ; though a complication of it, on the same principle,
will apply to cube, and even to higher, roots. At the same
time it is undeniable that square roots and cube roots occur
much more frequently than do others, just as second and
third powers do ; partly because they cover the actual
dimensions of our space.
CHAPTER XXXIIT.
Simple Algebraic Aids to Arithmetic, etc.
A VERY little knowledge of algebra enables us to make
better estimates, and to approximate as closely as we please,
l)oth to powers and to roots ; and it is worth while to show
this now : this chapter being chiefly one for exercise and
practice. It may be regarded as a chapter of miscellaneous
worked out examples, rather than as a progressive chapter;
though it contains the proof or explanation of the ordinary
square root rule.
First of all consider the multiplication of two binomials, that
is two factors each consisting of two terms, say {a + b){c-\-d).
Every term will have to be multiplied by every other, for it
means a(c-{-d)-\-b(c + d), that is ac + ad-hbc + bd.
So for instance (3 + ^2) (4 + ^3)
will equal (3 x 4) + 3 ^3 + 4 ^2 + ^2 ^3
= 12 + 3^3 + 4^2+^6.
Or take this example,
(J2 + 2){J3-S)
multiplied out it becomes ^6 + 2^3-3^2-6.
But take a more easily verifiable example, say
(17-5)(13-10)
= 221-170-65 + 50
= 271-235 = 36;
rather an absurd way to do such simple arithmetic as 12x3.
284 fiASY MATHEMATICS. [chap.
Very well, now take the case of {a + hy.
It means (a + h){a-\- h).
And this multiplied out equals a^ + ab + ba + b^ ;
but ab + ba = 2ab,
so (a+b)2 = a2+2ab+b2.
Similarly (a-b)2 = a2-2ab+b'».
Now let us use these results to obtain powers and to ap-
proximate to roots. Suppose we want lOS'^, work it out thus:
(103)2 = (100 + 3)2
= 1002 + 6x100 + 32
= 10,000 + 600 + 9
= 10609.
Again, to find (998)2 ; write it as (1000 - 2)2
= (1000)2-4000 + 4
= one million less 3996
= 996004.
Similarly: (125)2 = (120 + 5)2 = 14400 + 1200 + 25
= 15625
or (125)2 = (130-5)2 = 16900-1300 + 25
= 15625.
(79-2)2 = (80 --8)2 = 6400 -1-6x80 + -64
= 6400-64-128
= 6272-64.
(5-11)2 = (5 + -11)2 = 25 + 1-1 + -0121
= 26-1121.
(39)2 = (40-1)2 = 1600-80 + 1
= 1521.
A further algebraical aid is often of great use, especially in
preparing for logarithmic calculations.
The value of {a + b)(a - b) when multiplied out is
a^-ab + ba-b'^:
XXXIII.]
ALGEBRAIC AIDS.
285
the two middle terms destroy each other, and so only «2 _ 52
is left.
This is a most useful fact to remember,
a2-ba = (a+b)(a-b).
For instance 92 - 42 = (9 + 4) (9 - 4) = 13 x 5 = 65,
(17-31)2 -(2-69)2 = 20x14-62 = 292-4,
(•019)2 _ (.^08)2 = -027 X -Oil = 2-97 x 10"*,
(1-05)2 -(-95)2 =, 2x-l = -2.
The fact is so important that it is worth learning in words.
The difference of two squares is equal to the product of
sum and difference.
Expressed thus it suggests a geometrical way of putting it :
K
C!
M
G^.
Fia. 30.
Let AB and ^C be any two given
lengths.
Erect a square on each, viz. the square
AD and the square AE, drawing them
so that they are superposed.
The difference of the two squares is
shown by the irregular six-sided rect-
angular figure with what is called a
" re-entrant " angle at E.
We have to show that this area is
equal to that of a rectangle bounded by
lengths representing the sum and the difference respectively
of the two given lengths.
To construct such a rectangle in a convenient position,
produce CE both ways to F and G, making CG = CA. Then
FG is equal to the sum of the two given lengths, viz. AB + AC;
and GH, which is the same as CB, is equal to their difference.
Therefore the area of the rectangle GHDF exhibits the
product of the sum and difference. Hence we have to show
that this rectangle is equal to the area of the irregular figure
CELKDBC, the difference of the two squares.
286 EASY MATHEMATICS. [chaf.
Now the two areas have a great part common, viz. the
rectangle BF; so we have only to show that the residues LF
and GB are equal.
By producing LE to M, another rectangle EB is constructed
equal to GB ; and this rectangle is plainly equal to LF, because
the height and base of the one correspond to the base and
height of the other.
The proof is therefore completely indicated. It has been
rather long and not particularly neat, but it is such a proof as
could be invented by an industrious beginner for himself.
The proposition is really an ancient one, and is established
with due ceremony in Euclid Book II., Propositions 5 and 6.
We observe from this example that a geometrical proof is
or may be hard, while an algebraic proof of the same thing is
absurdly easy : so it often is, though not always. As usual
there are plenty of ways of proving a proposition ; the pro-
position itself is more important than any one proof of it.
The geometrical illustration has been introduced here to em-
phasise the extreme importance and usefulness of the fact that
{x + y)(x-y) = x^-y^.
Now let us proceed to show how it is employed for adapting
things to logarithmic calculation.
Suppose we had to find the value of the following :
(8-131)2 -(4-026)2.
We might look out the logarithm of each, double it, find the
antilog of each, and then subtract them.
But on the other hand we might first throw it into the form
12-157 X 4-105,
look out the logarithms of these two numbers, add them,
and find the antilog of the sum. And this is a shorter process
than the preceding.
In general, sums and differences are awkward for logarithmic
calculation, while products and quotients are convenient.
xxiiii.] ALGEBRAIC AIDS. 287
Take another example of finding the value of a difference of
two squares :
(15) ~ V35 j ^ \T5'^ 35 j Vr5 ~ 35;
~ 105 ^ 105
_40_
(105)2*
And it is easy to look out the necessary logarithms :
log 40 = 1-6021 ; log 105 = 2-0212.
2 log 105 = 4-0424
difference 3-5597 = log of -003625.
•003625 is therefore the result.
We might indeed have done the above differently, because
we happen to see a common factor in the given expressions,
and can take it outside brackets, thus,
V15; \35/ " W \V3/ \7/ J 25 • 21 • 21
40 160
~ 25 X (21)2 44100'
log 16 = 1-2041
log 4410 = 3-6444
difference 3-5597 = log of -003625 as before.
This therefore serves as a check, and is itself instructive.
Sums of this kind, given as exercises, will call out nascent
ingenuity and will furnish much better and more real
arithmetical practice than a quantity of routine examples
without much variety.
In so far as the actual arithmetical operations to be
performed are usually simple and short, that is a peculiarity
pharacteristic pf nearly all the real sums that have to be
288 EASY MATHEMATICS. [chap.
done in practice; always excepting the long and intricate
operations occasionally undertaken for special purposes by
pure mathematicians — a matter with which children have
nothing whatever to do.
Sometimes the converse use of the proposition
^2-52 = (a + b)(a-h)
is convenient. For instance, suppose we had to find the value of
(N/3 + y2)(V3-V2).
It would be very clumsy to interpret it arithmetically thus :
(1-732 + l-4U2)(l-732- 1-4142)
= 3-1462x0-3178,
whose logarithm is -4977
plus 1-5022
equals 1-9999
which is the log of something extremely near to unity, and
perhaps unity itself if we had taken more places in the
logarithms.
I say this would be an extremely clumsy way.
The neat and direct way is to write the product as the
diflference of the two squares, thus :
(73 + J2) U3 - V2) = USy - (72)2 = 3 - 2 = 1,
which shows that it is unity exactly.
Take other examples of {a -\-b){a-b) = o? -l)^:
(Vl4-V8)(V14 + ^8) = 14-8 = 6.
(x/7-x/3)(77 + V3)= 7-3 = 4.
(n/5 + 1)(x/5-1) = 5-1 = 4.
(v/57-l)(V57 + l) = 57-1 = 56.
(1+V17)(1-^17) = 1-17 =-16.
(V3T4T59 + l)(x/3T4T59-l) = 2-14159.
cxiii.] ALGEBRAIC AIDS. 289
{1 + (-0012)^} {1- (-0012)^"} = 1--0012 = -9988.
This last might have been done thus :
3 + 1 3-1 ^ 8
JS ' J'S ~ 3'
(6 +^20) (6 -^20) = 36-20 = 16.
(v/5-2)(v/5 + 2) = 5-4 = 1.
(m + Jn)(m-Jn) = rn^-n.
2 a%^-n^
62 dP-'
2 x^ - a^
(M)(M)
/ m n\ /m n\ _m'^ n^
( X a \ f X a\ _x'^ a^ _
\sjci Jx) \Ja ~ Jx) ~ a~~x ~
(30 + iV^)(30-i» = 900-^.
\x^ 6;W'*'6;~ir4 3e
^ 576_-^o
36a;4 *
{ab-Jab)(ab + Jah) = a^^ - ab.
(^a-hi/b){^a-l'b)^J-b^.
{a'^b){a'-b) = cv^-b\
(a" + ft-")((t"-a-«) = a^-a-^
L.E.M.
290
EASY MATHEMATICS.
[chap.
(a*" + «")(«"' -a") = a"
{ajx + hjy){ajx-hjy) = a^x-ll^y.
(l+a")(l-«") = l-d"^.
\ _x-\
X X '
3\/. , 3\ _ 9 16a;2
(.v.+|)(v-j.) -
16.T
a; X
_ (4a; + 3) (42; -3)
(1 + sjlog n) (1 - x/log n) = 1 - log 71.
{jT^ + sfl3y)(JPf3i-sF73y) = l-73j/--73y = y.
(J(l +m) . u - Jm . u){J(l +m) . u + Jm . u) = ul
{JJ+l + 76) {Ja + h - Jh) = «.
If we have now driven home the important fact that
{a + h){a-h) = a'--b^ (1)
sufficiently, we will proceed to illustrate geometrically those
other equally important truths, viz. that
(fl + 6)2 = a^ + 2ab + b^, (2)
{a-bf = a^-2ab + b\... (3)
or, expressed in words, the square of a binomial is the sum of the
squares of its terms plus twice their product.
Or expressed geometrically. (2) The
square on a line made up of two parts is the
sum of the squares on the parts plus twice
the rectangle contained by the parts.
The annexed figure makes this obvious.
For the base of the big square is made up
of two parts labelled a and b.
And we see that it is built up of the square on a, plus a
ab
b'
a=
ba
ci b
Fio. 31.
D a
ab
(a-bf
b""
A a-b C
XXXIII.] ALGEBRAIC AIDS. 291
square virtually on h, plus two rectangles each equal to the
product ah.
It is a quite ancient proposition known as the fourth
proposition of Euclid's second book.
(3) The statement for the squared difference {a - by expres-
sion may be worded geometrically thus :
If a straight line is divided into two parts, the sum of
the squares on the whole line and one of
the parts is equal to twice the rectangle
contained by the whole line and that part
together with the square on the other
part.
The same figure serves, differently
labelled ; but a separate figure may make
it clearer. The square of AB, which is a'^,
together with the square on BC, which is F <^
b'^j exceeds the square on AC, which is
(a - by, by twice the rectangle ab, that is by the two rectangles
DE and EF.
This proposition, which asserts that
a^ + b'^ = (a-by + 2ab,
is known as the 7th proposition of the second book of Euclid.
It may be illustrated, like the preceding, by the folding of paper.
The process of putting these propositions into proved
geometric form is, we see, liable to be rather troublesome and
long. Algebraically they are quite easy. Geometry illustrates
the algebra, but it does not in this latter instance illustrate it
strikingly; and it is quite possible to spend too much time
over such geometrical illustrations, unless they are made out
by pupils for themselves, which is an admirable exercise. A
great deal, though not all, of Euclid's second book is of this
character, and represents an antique method of expressing
algebraic results without employing algebra. For good reason
292 EASY MATHEMATICS. [chap.
in those days, — because algebra was not then invented.
Children need not be dosed with too much of this rather con-
fusing and nearly useless kind of geometry at the present time.
Illustrations.
Let us write down some illustrations of the use of these
results in simplifying algebraic expressions, and in finding
roots. Write the results compactly thus,
(a±by = a^±2ab + h\
and then illustrate them :
{Ja + Jhf = a + 2J{ab) + h
{x^-y^Y = x + y-2jxi/.
{6+xy = 36 + l2x + x^
(a;-l)2 = a;2-2x+l.
("-;)
2 1
= a;24-2 +
X
2
+ a;2_2.
(5-^2)2 = 25-10^2 + 2 = 27-14-142...
= 12-858....
(1-73)2 = 4-2^3 = 2(2 -V3) = '536....
Notice that although ^3 is greater than 1 the squared
difference cannot help being positive.
/ 1 \2
W3
-TV^y
— -j-r^ -rt
J — W tJ,
(..
= 5-2 + i
I = 3-2.
(121)2
= (120)2 +
240 + 1 =
14641.
(119)2
= (120)2-
240 + 1 =
14161.
(1-5)2
= (1+^)^
= 1+1+1
- = 2-25.
(1-3)2
= i-6 + -i =
= 1-7.
XXXIII.] M.GEBRAIC AIDS. 293
Problems.
1. If any diagram has all its linear dimensions increased
by one-sixth, by how much is the whole area of the figure
increased ?
The answer liable to be given is one thirty-sixth, but it is
not right. The right answer is Jfths, or a little more than
one-third of the original area. The first answer attends only
to the little corner squares and neglects the two strips, for
{a + hf-a^ = &2 + 2a&;
the 2ah being much bigger than J^.
The simplest solution is to say that in the linear dimensions
throughout, 6 has become 7, hence, in the area, 36 has become
49 ; wherefore the superficial increase is 13 of the same parts,
that is 13/36ths of the original.
2. If a block is reduced in the ratio of 3 : 2 linear, that is if
its length, breadth, and thickness are all made two-thirds of
what they were, the shape being preserved, what change has
been made in the surface or superficial area and in the volume
or cubical contents ?
Answer. The linear dimensions being reduced by one-third,
or from 3 to 2, the superficial are reduced by five-ninths, or
from 9 to 4 ; and the cubical are reduced by nineteen twenty-
sevenths, or from 27 to 8. In other words the surface is less
than half what it was, and the volume is less than a third
what it was.
3. If every linear foot becomes 13 inches, every square foot
becomes 169 square inches, and every cubic foot becomes 2197
cubic inches. So, while the linear increase is y^^th of the
original, the superficial increase is y^^^ths, or a little more
than Jth of the original area ; the volume increase is yYA*^^'
or distinctly more than ^th of the original volume.
4. If one per cent, is docked ofi" linear dimensions, about
294 EASY MATHEMATICS. [chap.
two per cent, are thereby taken from area, and about three
per cent, from bulk.
Now use the same equation to find soLuare roots.
Suppose we want the square root of 50. We see instantly
that it is a little more than 7, let us call it 7 + x, then write
50 = {7 + xf = 4:9 + Ux-hx\
or, subtracting 49 from both sides (i.e. transferring 49 over to
the left with change of sign),
1 = Ux + x\
wherefore x = j^ is a first approximation, for the x^ is a very
small number, almost negligible, x is really a trifle less
than y'j, though not so much less as ^^ would be, for its defect
is yj of x^, which is approximately ^wois ^^^7- ^^ i^ impossible
to express the root accurately, and the result obtained by
neglecting x^ is usually a quite sufficiently close approximation.
So the root is l + ^t = 7-0714. The error is in the last
place ; the 4 is too big, it ought to be a 1.
So also to guess the root of 143
Write it 12-«.
143 = (12-a;)2 = 144-24a; + a;2,
neglect x% and ^ = aV = *^^ ^ 7»
so approximately JUS = 12 --0417 = 11-9583.
Its real value is 11-9582607 ....
What is the square root of 99 ?
99 = (10-x)2 = 100-20x + a;2^
so X = 217 = '^^
and ^99 = 9-95....
What is the root of 395 1
Let it be 20 - x.
395 = 400 - 40a; + a;2
or a: = /^ = i = -125,
so 7395 = 19 875....
xxxiii.] ALGEBRAIC AIDS. 295
What is the square root of 1,000,015'?
that is of 10<5 + 15.
Call it 103 + a;;
then 10« + 15 = (W + x)^ = 10^ + 2000a; + x^ ;
15
whence x = ^^^ = -0075,
and so the required root is 1000*0075.
In such a case the extra quantity is extremely small, and
we see that in the root it is just half the value of the corre-
sponding quantity in the given square.
This is a handy approximation which may be generalised
and recollected. It is an immediate consequence of neglecting
a;2 and writing (1 +xY = l + 2x approximately when x is small.
So \/l + 2x = 1+x approximately.
For instance ^^(1*008) will equal 1-004 ;
and ^(100-084) = 10^1*00084 =0. 10 x 1-00042
= 10-0042 ;
or the equation may be written
J{l+x) = 1 + |a; approximately.
The following relation,
^(10" + a;)^10^" + 10~^\|a:,
when X is moderately small, is a general result; but for
memory it is best to make the first term unity ; and so in the
numerical example just above, the factor 100 was first taken
outside the root, where of course its value is 10. If the factor
had been 1000 instead of 100, that is, if there had been an
odd number of ciphers in it, this could not have been done so
easily : we should then have had a ^10 to deal with, and that
would destroy the advantage of the process.
The process applies most obviously to numbers which can be
separated into two very unequal portions, one of which has a
known square root. If they are not very unequal, the neglect
296 EASY MATHEMATICS. [chap.
of x^ becomes of more consequence, and the same sort of
process must be continued further, before the square of an
outstanding error is neglected.
Suppose for instance we wanted the square root of 72 ; we
could write it as (8 + «)2 or we could write it as (9 - yf^
so that 72 = 64 + 16a:4-a;2,
or 72 = 81 -18«/ + ?/2,
whence approximately x = ^^ = i, and we should be
neglecting \\
or 2/ = tJ^ ^^^ w® should be neglecting a trifle less.
So the answer would be roughly 8*5,* but this would be a
little too big, and the process must be continued, by successive
approximations, beyond 8*4, in such a case; the process
develops, in fact, into the ordinary arithmetical method of
finding a square root, as described but not explained in the
last chapter. We can now explain it, for it all depends on
what we have just been doing ; it involves an ultimate ignoring
of an Qi?, but it carries the process of surmising the root to any
desired degree of approximation, before the inevitable out-
standing error is considered so minute that its square may
safely be neglected.
To illustrate the process arithmetically, and at the same
time display its rationale algebraically, take any simple
number at random, say for instance 33, call it iV, and proceed
to approximate to its square root.
(1) First guess the nearest lower integer root, namely 5,
call it a in general, and write x for the unknown necessary
complement to be found, so that
*In the particular example chosen it happens to be very easy to
calculate the square root, because the 'factor method' would apply.
Beginners may be reminded always to keep an eye open for the simple
and satisfactory factor-method, such as this :
72 = 9 X 8,
80 ^72 = 3x^8 = 3x2^2 = 6^/2=8-48528....
XXXIII.] PROOF OF SQUARE ROOT RULE. 297
i\^ = {a + xy = a^ + 2ax + x'^,
or 33 = (5 + x)2 = 25 + lOx + xK
From this we deduce that the deficiency N-a^ = x{1a-\rx),
or that 8 = x{\0^-x).
This gives us our first approximation to the required com-
plement, or error in our rough estimate of the root, namely
X = -TK , or say "7 as the first di^it of it.
(2) Thus we can now make a closer guess at the root,
namely 5*7, and start afresh for a second approximation, x\
writing 33 = (5*7 + xj = 32-49 + 11 -^x' + x'%
so the second deficiency is '51 = x'{\\'4: + x'),
which gives, as the second outstanding error,
*51
x' = ■— — , = -04 as the first diijjit of that.
ll-4 + « ^
(3) Our approximation to the root has now become 5*74,
and we start ofi" a third time to write
33 = (5-74 + a;72 = 32-9476 + ll-48a;" + a;"^
whence the third deficiency -0524 = ic"(ll-48 + ic"), which gives
us x" = -004 as the next digit of the rapidly diminishing
outstanding error.
(4) The approximation is now getting closer, being 5-744,
and so we start again, saying
33 = (5-744 + a;"')2,
whence the fourth deficiency comes out
•006464 = a;'"(ll-488 + a;'"),
yielding x'" = -0005 ... as the error still remaining.
(5) We have now arrived at ^33 = 5*7445 . . . , and we can
continue the process as long as we like ; but, at this (or at any
other) stage, we can take refuge in simple division, to get at
once a still closer approximation. For hitherto we have not
neglected the square of any small quantity : everything so far
298 EASY MATHEMATICS. [chap.
has been exact ; but sooner or later exactness will have to be
abandoned, because we know that a number really has no
exact numerical root. It was considered too inaccurate to
neglect the gquare of x, but we might perhaps have neglected
the square of x\ or at least of x'\ We did not neglect even
this however, but we are now going to neglect the square of
x'" ; so after reckoning the present deficiency, -00071975,
instead of saying ^^^^ _ -00071975
^ ~ 11-4890 + x'""
which would be continuing the process, we will say simply
"" _ '00071975
"" " 11-4890"'
very nearly, and divide straight out, getting -00006265 as the
result. Wherefore finally the approximation at which we
have arrived is ^/33 = 5-74456265 ....
If the process thus elaborated be compared with the
operation as ordinarily performed, a little thought will make
everything clear without more words.
The only thing that can require explanation is the actual
mode of reckoning the successive outstanding deficiencies, viz. :
N-a^', N~{a + xf; N - {a ■\- x ^^ x'f ; a,ndiN-{a + x + x' -{■x"f.
The original number N is not in practice thus manifestly
reverted to for the purpose of getting these values — which in
the above numerical example are successively
8; -51; -0524; and -006464,—
but exactly the same result is obtained by the successive
subtractions as ordinarily performed : the value of an ex-
pression like (N - a^) - x{2a -\- x) being practically employed,
each time, instead of the equivalent N -{a + xY, because
(having already found N- a^) it is quicker to reckon.
The well known ordinary process is here exhibited for the
same number, in order that it may be compared with the
XXXIII.]
PROOF OF SQUARE ROOT RULE.
299
above fully explained treatment. To find the square root of
33, write
r
33- 1 5-7445
■
25-
10-7
8-00
i-
7-49
11-44
•5100
-4576
11-484
•052400
•045936
t
11-4885
•00646400
f
•00574425
11-4890
•00071975
and the outstanding error in the root is very closely indeed
equal to the residual deficiency divided by twice the root
so far found, that is to say, -00071975 -^ ir4890, or 00006265.
The advantage of the approximation we noted on p. 295,
is so great that even when the first number is conspicuously
not unity, it is often convenient to make it so by division.
For instance to find ^85, it equals J{^l-\- 4)
= 9V(1+A)=^*9a+A) = 9 + 1 = 9-2.
And so with some of the other examples, they too may be
done this way.. We will therefore repeat them.
750 = ^(49 + 1) = 1JU^^^1(1+^)
= 7x1-0102 = 7-0714.
In this case the approximate value -0102 is obtained thus.
98 is two per cent, less than 100, so — is two per cent, greater
than -01.
7143 = 7(144-1) = 127(1 -tJ^)- 12(1 -^1^) = 12-^V
*At this stage the second term is halved and the root sign dropped.
300 EASY MATHEMATICS. [chap.
^99 = 10V(1-tJo)=^10(1-^1^) = 10 --05 = 9-95.
V375 = V(400 - 5) = 20J{1 - ^\) - 20(1 - ^i^)
= 20-i = 20 --125 = 19-875.
Perhaps decimals might be preferred throughout. Some-
times they would be handier, sometimes not.
^396 = 7(400 - 4) = 207(1 - -01) ii= 20(1 - -005)
= 20--1 = 19-900.
The result of this convenient approximation is always to give
slightly too big a value for the root, and this whether terms
under the root are separated by a negative or a positive sign.
Thus for instance the approximation to JlOl namely 10-050,
and to 799 namely 9-950 are both of them a trifle too big.
The error itself can be estimated by a further stage of
approximation, and so gradually we can get as nearly accurate
as we please, but we leave it there for the present.
The error in either case is about -000125, so the digits as
they stand above are fairly near the truth.
Cubes and Cube Root.
Now let us see what we can get of the same kind to help us
in other cases. Suppose we cube a binomial, what shall we get ?
First notice that
{a + b){c + d){e+f) = (a + h){ce + cf+de-]-df)
= ace + acf + ade + adf
+ bee + bcf+ bde + bdf,
eight terms altogether.
So take the three factors all alike.
{a + by = (a + b){a + b)(a + b)
= acut + aab + aba + abb
+ baa + bab + bba + bbb
= a^+a% + a% + ab^
+ a^b + ab^ + ab^ + b^
= a^ + 3a% + Sab'^ + b^
XXXIII.] ALGEBRAIC AIDS. 301
Not a very simple expression at first sight, but quite simple
when you get accustomed to it, and very easy to remember
and write down.
Notice first that every term is of the same "dimension,"
that is to say it involves three letters multiplied together,
no more and no less. There is no term involving only a^j
nor only b^, nor a alone, nor is there anything like a\ The
expression is a cube, and every term is of the nature of a cube.
If a and b were lengths, the cube is a volume, and every term
is necessarily a volume. You cannot with any sense add an
area like a^ to a volume like a^, but you can add a volume like
a^b or like ab^ to another volume like a^, and you can add each
more than once, in fact 3 times if you choose.
Notice next that the power of a decreases by one each term,
and the power of b increases. We might, if we liked, introduce
the index 0, because we know that
ao = 1 = 60.
So the more fully written expression
would represent, with needless explicitness, the truth that the
sum of the indices of each term is 3.
As to the big 3's prefixed to the two middle terms, they are
styled coefficients, or numerical factors ; we have seen exactly
how they arise, simply because we had to add three equal
terms. They take the place of the 2 in the middle term when
we were squaring a binomial.
We illustrated the square of a binomial by fig. 31, — where
the a^ and the b^ and the two rectangles each equal to ab
are obvious, and plainly make up the (a + by.
So also we can geometrically illustrate the cube of a
binomial ; taking a cube whose every edge is divided into any
two parts, respectively a and b, we get a figure like 33, which
is more easily realised when built up or sawn out of wood.
302
EASY MATHEMATICS.
[chap.
Such of the portions as are visible are labelled with their
respective volumes. There is first a big cube a^^ then there
are three slabs each of area a^ and thickness h, but one of them
in the figure is invisible at the back; there are 3 rods or
'^b^-...
prisms each of the length a and sectional area h"^ ; and lastly
there is a little cube b^ diagonally opposite the big one ; and
these make up the 8 pieces, out of which the whole cube has
been built up, (a + hf.
This then is a solid figure illustrating the cube of a bi-
nomial in the same sort of way that Euclid 11. 4 illustrates
the square of the same quantity.
Suppose we wished to illustrate the fourth power of a
binomial by geometry. We could not possibly do it in any
natural fashion, for we have already exhausted all the dimen-
sions of space. Hence geometrical propositions on involution
are not only complicated and wordy, but are feeble and limited.
Algebra is not limited at all; we can raise a binomial
to the fourth, fifth, fifteenth, or any other power that we
please, and presently we will do it. But first we will take
a few examples and applications of what we have learnt
about the cube or third power.
XXXIII.] ALGEBRAIC AIDS. 303
First a mere numerical illustration or verification :
(5 + 2)3 _ 53 + (3x25x2) + (3x5x4) + 23
= 125+ 150 + 60 +8
= 343.
Then take a case where the first terra is unity :
{l+xf = l + 3x + 3x^ + x^,
and then one with the second term negative :
(l-xf = l-3x + 3x^-xP.
Notice in this case that the signs in the expansion are
alternate, because the powers of {-x) are alternately odd
and even : the odd will all be negative, and the even will be
positive. The general case, with the negative sign to the
second member of the binomial, ought also to be recorded :
(a - bf = a3 _ Sa% + Sab'^ - b%
(x-lf = .'c3-3a;2 + 3a:-l,
but this is just the same as (1 -xy with the sign of every
term reversed.
It is worth obtaining the general result for the third power
of a±b in another way, by help of what we know about its
second power.
{a±by = (a±b)(a^±2ab + b^)
= a^±3a% + 3ab^±b^
Observe that the alternative sign affects only alternate
terms, viz. those which involve the odd powers of the possibly
negative quantity b. Among its even powers there is never
any variety.
Another special case is
x + ~) = x^ + 3x + - + -^.
x) XX?
This is rather a curious case, considered from the point
of view of the ' dimensions ' of each term, x^ looks like a
volume, and would be a volume if x were a length, and
304 EASY MATHEMATICS. [chap.
3a; would be merely treble that length ; then come reciprocals.
How can this be possible 1 Answer : — It is never possible to
have different dimensions in different terms of an expression.
It is quite easy and common to have factors of different
dimensions, as components of a single term, united by the
sign X , but different terms united by the sign + or - are
always of the same dimensions.
Apply that to the case of lx + -j and we see that x^
cannot possibly be a volume, nor can a; be a length, if 1 is
a pure number. It can in that case only be of the same
dimension as its reciprocal.
Length and volume are all very well as illustrations, but
it would be a great mistake to suppose that algebraic sym-
bols can express nothing else. The terms "square" and
"cube" suggest geometrical signification, and that doubtless
was their original meaning, but now they have been so
generalised that the original geometrical signification is
almost forgotten. Cube is still used merely as short for
"third power," and square is short for "second power," but
the things that we raise to powers may be anything whatever
that we find convenient. Often they are mere numbers,
like a number of oranges. If we speak of 3^ oranges meaning
27 oranges, it may be a pedantic mode of statement, but it
is not incorrect. Even if we spoke of a cube of 3 oranges,
or 3 oranges cubed, we might possibly be understood, as
meaning a cubical box full of oranges with 3 in each edge,
9 in each face, and 27 in the box.
But this expression would not bear close examination,
unless we put it in brackets, thus, (cube of 3) oranges, and
then it does express more than merely 27. For 27 might
be lying about anyhow, but (cube of 3) signifies that they
are packed in a certain compact arrangement..
XXXIII.] ALGEBRAIC AIDS. 305
Why is cube of (3 oranges) wrong? Because that would
mean 3^ x oranges^ ; and the latter factor has no meaning.
Cube of (3 feet) is perfectly right, for that means
33 X feet3 = 27 cubic feet.
You can have a cubic foot, but you cannot have a cubic
orange ; or rather perhaps you cannot have anything linear or
superficial in oranges, as you can with feet or metres or inches.
Eeturning to the expression x + - then, what can x mean 1
Only a thing whose dimensions are the same as its reciprocal,
that is to say, a thing which has no " dimension," not a con-
crete thing at all, but an abstract number, a number of things
abstracted from " things " altogether and contemplated alone.
That is what we mean by an "abstract number" or "pure
number." It is the simplest kind of "abstraction" there is,
and the first we arrive at ; later we shall employ plenty more.
If It is a pure number, like 2,
- is likewise a pure number like n-
n^ is also a pure number, and n^, and any power.
Jn or any root is also a pure number.
So is log n.
We cannot assert that a*" is a pure number for certain,
because it depends entirely on what a is.
a might be a length, in which case a^ would be an area, and
a^ be a volume, a^ would in that case have no assignable
physical meaning, but it would certainly not be a pure
number.
There is therefore no difficulty about an expression like
3?~ ic3 '
(■
-S'
= x'
-Zx + ~-
X
every
term must
be a
pure
number.
L
E.M.
u
306 EASY MATHEMATICS. [chap.
This is not necessary with the next example, because there
all the terms in the expansion have the same dimensions;
provided always that a and b are quantities of similar kind.
(a2 - &2)3 = ^6 _ 3ft4J2 + 3^254 _ J6,
In the next case, however, a must be a pure number,
because of the term unity. If 1 means 1 something, the
something cubed can go outside the brackets : it must apply
equally to both a^ and 1.
(^2-1)3 = a^-3a^ + 3a^-l.
(1-^2)3= l-3J2 + (Sx2)-{j2f
=1+6-372-272
= 7-572 = 7-7-071 = --071.
(7-1)3 = 73 + -3x72 + 21 X -01 +001
= 343 + 3x4-9 + -211
= 357-911.
(57)3 = (50 + 7)3 = 125000 + (21x2500) + (150x49)
+ 343 = 185193;
but in this case it would be easier to do it by simple
multiplication, 57 x 57 x 57, or perhaps by logarithms. The
worst of logarithms for finding a positive integer power is
that they only give it approximately, unless you take a
considerable number of places ; and an integer power never is
approximate, it can always be numerically expressed, because
we start with a number and only multiply it by itself.
By "integer power" or "integral power" I do not mean a
power of an integer, I mean any number raised to a power
whose index is a whole number and not a fraction. If the
index is fractional it represents a root. The case is entirely
different with a root, for then we are endeavouring to find
something which multiplied by itself will produce a given
number : and the result is usually incommensurable.
XXXIII.] FINDING CUBE ROOTS. 307
But for integer indices, whether positive or negative, we
can always get an exact result by straightforward multiplication;
for instance 2^^, or 2~% or (1'2)^.
(1-2)3 = 1 + (3 X -2) X (3 X -04) + '008
= 1 + -6 + -12 + -008
= 1-728,
which is a familiar number — expressing the thousandth part
of a cubic foot, if 1"2 means the tenth of a foot in inches.
Now find a cube root or two by the approximation method,
choosing numbers which are not very different from a perfect
cube.
Say we want the cube root of 65, call it 4 + a;.
65 = (i + xf = 64 + 48x+12a;2 + ic3;
So the first approximation to x is ■^^.
This however is a trifle too big, because 12a;2 has been
neglected. So we might call it ^g or even -^q, at a shot,
and say that the answer is 4-02 ... . As to neglecting x^ it
is of slight consequence. This process, elaborated, is the basis
of the arithmetical cube-root rule.
Take only one more example of finding cube roots, because
they are usually done most easily by logarithms.
To find ^341.
341 = (7 - xf = 343 - 3 X i9x + 21a:2 - a;^;
2
.*. approximately x = ^ — j^, or, as this is a trifle too small,
2 1
say - — — = — = -0139. So approximately
v^341 = 7 --0139 = 6 9861,
which is still a trifle too small in the last place. The digit
1 ought to be a 3 or a 4.
308 EASY MATHEMATICS. [chap.
As an exercise it would be desirable to establish a method
akin to the square root approximation, like this,
(341)^ = (343 -2f = 7(1-^1^)^
= 7(1 - Y^\y-) approximately
= 7 X y^l^ ; which equals -2 per cent, less than 7,
= 6-986 roughly ;
or generally, when a; is a small quantity,
^{l±x) = l±lx approximately ;
which is equivalent to neglecting squares and cubes and all
higher powers of x.
Approximations.
The fact that the square of a small quantity is very small,
and the cube of it extremely small, is easy enough to under-
stand ; and since it is extremely useful in application, it should
be thoroughly understood and remembered. Let the small
quantity be 1 per cent., for instance, or "01 or Yho- ^^^ square
is X(tJo oj 0"6 ten-thousandth ; and its cube is a millionth ;
If then we have to find (I'Ol)^, it will
= 1 + -03 + -0009 + -000001
= 1 030901,
of which the first significant digit of the decimal is decidedly
the most important, the second is sometimes worth attention,
denoting a value about ^^rd of the previous one, and the last
is utterly trivial, except for exact mathematical purposes.
A cube of a foot and one inch (or 13 inches cubed),
(13 inches)^, is decidedly bigger than a cubic foot; but never-
theless a cubic inch is almost negligible in comparison with a
cubic foot : it is only the ytVf^^ P^^^ ^^ i*-
Let us examine this, because beginners often make mistakes
here.
XXXIII.] INCREASES IN BULK AND AREA. 309
( 1 foot + 1 inch)^ they incline to write down as a cubic foot
plus a cubic inch : which is just the mistake of thinking that
(a + 6)3 equals a^ + b^; in other words it is the mistake of
altogether ignoring Sci^b + Sab'^, three slabs and three rods, and
attending only to the little insignificant corner cube of the
small quantity b (supposing 6 to be a small quantity) in fig. 33.
The true value is
(1 foot+1 inch)3 = (1 foot)3 + 3(feet)2xl inch
+ 3feetx(inch)2 + (linch)3
= 1 cubic foot
+ 3 slabs a foot square and an inch thick
+ 3 rods a foot long and a square inch section
+ a cubic inch.
The last term is the most trivial of the eight terms, and the
3 slabs are the most important after the cubic foot itself.
Translating to inches, we see that
(13inches)3 = 1728 + (3 x 144) + (3 x 12) + 1
= 2197 cubic inches,
which is otherwise very easily arrived at.
If instead of a foot and an inch we had taken a yard and an
inch, the smallness of everything except the slabs would have
been accentuated ; and if we take a metre and a millimetre we
shall see it still more forcibly :
(1 metre + 1 millimetre)^ = 1 cubic metre + 3 slabs a metre
square and a millimetre thick
+ 3 lines a metre long and a
square millimetre cross section
+ a millimetre cube ;
or expressing it all in cubic centimetres
= 1 million c.c. + three hundred thousand c.c. + three c.c.
+ a thousandth of a c.c.
= 1,300,003-001 c.c.
310 EASY MATHEMATICS. [chap, xxxiii.
When things expand by heat, the expansion is usually very
small ; the increase of bulk is not so small as the increase of
length however. If the edge of a cube expands 1 per cent,
the volume of it expands just about 3 per cent., and the area
of one of its faces about 2 per cent. This follows from what
we have been saying. Compare page 293, No. 4.
It is sometimes expressed by saying that the proportional
superficial expansion is twice the linear, while the cubical
expansion is three times the linear. We will employ the
subject of expansion to furnish us with a few interesting
arithmetical examples of an easy and uncommercial kind in a
future chapter, but first we will do some algebraic expansions.
CHAPTER XXXIV.
To find any power of a Binomial.
Suppose we have to find {a + &)^ we have only to multiply
a + ?> by itself four times, and write down the result. We
might write it thus
(a + hf{a + hf
= (a2 + 2ab + 62) (^2^ ^ah + h'^)
= a^ + 2a%+ a%-^
+ 2a% + 4a%^ + 2ab^
+ a%^ + 2ab^ + b^
= a^ + ^a% + 6fl2/>2 + 4a j3 _^ j4
Now here we see the same sort of law as was observed in
the expansion of (a + b)^ ; the indices of a decrease regularly,
and those of b increase regularly, so that every term is of the
fourth degree. The numerical coefficients follow a less
obvious law. Let us write them down for the cases that we
know.
for (a + b) 1 1
„ (a + 6)2 1 2 1
„ (a + bf 1 3 3 1
„ {a + by 1 4 6 4 1
The law is fairly plain, and we might guess the coefficients
for the next sets :
(a + bf 1 5 10 10 5 1
(a + bf 1 6 15 20 15 6 1
312 EASY MATHEMATICS. [chap.
and then we can verify them, by direct multiplication, thus
(a + 5)5 = {a + bf(a + by
= a^ + ba^b + 10^^362 + lOa^^ + 5a¥ + b\
(a + bf = (a + bfia + bf
= ft6 + 6^56 + 15a452 + 20a3^>3 + 15^2^4 + Qab^ + jo,
A guessing process like the above, which is subsequently
verified and obviously extensible to the case of any positive
integer as index, is a method of frequent and considerable use
in order to first ascertain a rule or law or method of pro-
cedure ; but one should not rest satisfied without perceiving
the rationale of it, and so to say " proving " it or reasoning it
out; otherwise it remains what is called an "empirical" law,
meaning a law ascertained by experiment and observation
without a full knowledge of the reason. Some laws have to
remain of this character, when the subject matter is difficult
or obscure; but that is not the case with little calculations
like the present : the reasonableness of the result can always
be made out, and it is a most wholesome exercise. In the
present instance the method of expanding any binomial as an
empirical process seems to have occurred to Isaac Newton
while still quite young ; and the reasoned proof of this process
is what we now know as " the binomial theorem."
We will not go into this fully just at present, nor at all
more fully than is needed for practical purposes, but for a
positive integer the empirical process itself is easy and worth
while for anybody to know.
First write down what we have observed, for any positive
integer index n, concerning (a + by : —
We know that the powers of a will begin with a", and
decrease by one each time down to a^ or unity.
The powers of b will begin with b% or unity, and climb
by one each time up to b" ; so that as regards the algebraic
part of the expansion, the terms will be
XXXIV.] BINOMIAL EXPANSION. 313
or, or-% ar-w, ar-n)\ a%''-\ a}f-\ b'\
the sura of the indices of a and b always adding up to n,
which may be called the " order " or " degree " of the whole.
Now what about the numerical coefficients 1 We can obtain
them as follows. Take the coefficient of any term, multiply it
by the index of a in that term, and divide by the number
of terms preceding the next term, the result will give the
coefficient for that next term. This is what we have ascer-
tained empirically, though we did not word or express it
before, but it is what we did or might have done ; because,
take the case of {a + 6)^,
the first term is a',
so the coefficient of the next term is 5, giving
a^ + ba^b.
Now take the 5 and the 4, multiply them together, and
divide by 2 ; we get 10, which is the coefficient of the next
term, carrying us as far as three terms,
a^ + ba^b + lOa^^.
Then take the 10 and the 3, multiply them, and divide by
the number of terms ; thus we get the next coefficient, viz. 10,
a^ + 5a'^b+10a%^ + l0a^b^
Now take the 10 and the 2, multiply them, and divide by 4,
and we get the coefficient of the next term, viz. 5.
Then take 5 and 1, multiply, and divide by 5, and we get
the coefficient of the last term, viz. 1, giving the whole ex-
pansion, with six terms in all,
a^ + ba^b + lOa^b^ + I0a%^ + bab^ + b^
In the last term the index of a is zero, hence a does not
appear, because a^ = 1 ; and if we apply the rule further it
will give us zero as a factor of the next and of every succeeding
term ; which therefore all vanish, so the series terminates.
314 EASY MATHEMATICS. [chap.
Try this rule also for {a + b)^ and (a + hf, getting the result
in the latter case,
a^ 4- 7a% + 2la^b^ + 35a^b^ + 3ba%^ + 2U%^ + lab^ + b\
and then apply it to (a + 6)",
a** + na" ^b + — ^r — ' a" ^^ + -^ ^r-^-^ ^ a** ^b^
+"(»-^)("-^)<"-3V-'y + etc.
Now this is a most interesting example of a very important
algebraic thing called a ' series.' It appears to go on for ever,
but, as we have seen, it does not go on very long when n is a
positive integer, for sooner or later there will come the index
n-n, whose value is 0 ; and as this quantity n-n will enter
as a factor into every subsequent coefficient, they all vanish,
and the term with index 0 applied to a is the last term.
Thus for (ft + by* there were six terms in the series, and no
more. For {a + b)^ there were seven terms, and no more ; and
for (a + by- there will be n + 1 terms, and no more, provided n
is a positive integer. All subsequent terms are zero, because
they all contain the factor n-n. But if the index n \% z,
fraction, or if it has a negative value, even a negative integer
value, the cause of stoppage will no longer occur ; for,
naturally, a numb&r-of-terms can never be a fraction or negative.
There will therefore never be an index n-n-, there will be
71-7, 71-8. 71-9, etc., but none of these can possibly be zero
unless n itself is a positive integer.
Consequently in these cases the series does not stop, but
goes on for ever, extending to infinity. It may happen
however that its later terms become insignificantly small, and
that all after a certain number can be neglected for practical
purposes. This a point to which attention must be specially
directed, because it is exceedingly useful in practice.
XXXIV.] FACTORIALS. 315
Notice that we have not yet established or proved the above
series for the expansion or power of a binomial, even for the
case of n a positive integer. We will defer the proof for the
present : so far we have only arrived at it by experiment.
The proof is not difficult for n a positive integer, but it will
come better later. Mathematicians know how to prove it for
a fractional and a negative index, that is for the case of an
infinite series, which however is exactly of the same algebraic
form as the one we have written.
The method of experiment and observation is quite a good
practical method, only it might in some cases lead us wrong
unless it can be checked over and reasoned out by some more
intellectual process.
For the present we will accept the series and study it.
Notice first the denominators of the several terms. They consist
of a series of consecutive natural numbers 1 .2.3.4.5, etc.,
multiplied together. This sort of product often occurs, and it
is convenient to have a symbol for it. [_5 is the way it is
written, 5 ! is the Way it is printed, and it is called " factorial
5." They all mean the same thing, viz. 1x2x3x4x5, that
is to say 120.
So 4 ! has the value 24, since it means 1x2x3x4
M
>> 5>
6,
\1
)) J J
2,
\1
" "
1,
11
» J)
720,
7! or [7
equals
5040,
8!
=
40,320,
and so on.
So that factorial 20 is an enormous number.
Now look at the numerators. They too are factorials of a
kind, but they do not begin at 1, they begin at the other end ;
they represent a part of factorial w, with the early part cut off.
316 EASY MATHEMATICS. [chap.
" )
\n '
They might be denoted by — ^ ; for
-V, »
71 . 71 - 1 . 71 - 2 . 71 - 3 . . . w - r + 1 =
(71 - r) !
So the successive numerators are as follows :
[n \ii^ \n I
etc.,
\n 1 71-1 1 71-2 1 71 -3'
being 1, ti, n{n-l\ n{n-l){n-2), etc., respectively.
Hence any of the coefficients may be written in this form
\L'
while as to the ab part corresponding to this general coefficient
it will be a'-'"^''.
Hence the whole series may be neatly written as the sum
of a number of terms all of this kind, for every value of r from
0 to 7i; and such a summation is usually expressed by the
capital letter sigma ; hence
^ ^ ^Q 1(71-7')! r! J
which means that you write down all the terms of this form
in regular order from r = 0 up to r = tj, and then add them
together. Try to do this, for different values of n, for instance
3 or 4 or 5 or 6, and see that you get the series already obtained.
The only thing that requires explanation, until we come to
fractional and negative indices, is how to interpret " factorial
nought." To common sense such an expression sounds meaning-
less ; and to understand it fully, together with the factorials of
negative and fractional numbers, a good deal of mathematics
must be conquered. It is easy however to show that 1 0 must
be interpreted as unity, that is to say that [J_and [O^are alike
equal to 1.
XXXIV.] BINOMIAL EXPANSION. 317
Proof. \n = n\'n -\ ; but in the special case when n = 1
n and n are the same thing ; hence in that case \n-\ is unity,
but it is also factorial nought.
Exercise. — Make a table of binomial coefficients up to say
the index 12 as the finish. For answer, see p. 334.
A special case of frequent occurrence is when one of the
terms of the binomial is unity, as for instance (1 +xY.
Consider this case. Any binomial can be thrown into this
form by an obvious process, as follows :
{a + hy = «"(l+^y = a'\\+x)\
where a" is a factor taken outside brackets, and the ratio hja
is treated as a single quantity, a pure number, and called x.
Observe that a and h might be anything, so long as they
are the same thing, but that x must be a number, in order
that it may be added to 1 ; and being a ratio of similar
quantities, it is a number. The most important case, with
fractional and negative indices, is when a; is a small number,
for then the series, or expansion in powers of x, will rapidly
diminish, and all beyond a few terms can be neglected. The
meaning of this will become clearer soon.
First apply the ordinary rule for the expansion, observing
that 1""^ 1""'^, etc., need not be written, because they are all
mere unity factors. We have nothing therefore to write but the
successive binomial coefficients and the ascending powers of x.
/, v_ , 71.71-1 o n .n-\ .n-2 ^
{l+xf = 1 +nx + — 2l — ^ + 13 ^ + ••• »
a very useful expansion ; and if x is really small, so that x^ may
be neglected, it gives us this extremely handy approximation,
(1 + j)" sQb 1 + wa; when x is very small.
As a matter of fact we have used this already for extracting
approximate roots (p. 308), arriving at it by a different process.
318 EASY MATHEMATICS. [chap.
Thus to find J{1 i-x) when x is small we have only to put
» = i
^(1 +a:) = (1 -\-xy ^ 1 +^a; approximately.
E.g. v/(l'01) = 1-005 approximately.
^(1-008) = 1-004
VlOO-6 = 10^(1 + 006) ^ 10 X 1-003 = 10-03.
So also for cube or other roots.
^(l+x) = (l+x)^^l+^x,
i/1003 = 10yi003 = 10(1 003)^ =£3^10x1-001 = 10-01.
^33 = 2V(UA) = 2(1+^V)'^2(1+^)
= 2 + Jlg. = 2-0125.
Or take an example of a negative index.
-^ = (i+.r^-^i-i..
But this case of a negative index will bear examining more
fully.
Let us write n = -m, and then interpret the general
expression for the special case of a negative index. Observe
that it is no new expansion, only the old one re-written with
the sign of the index changed, but it looks different :
(a + b)-"^ = a-^ + {-m)a-^-'b+ ~^^[~'^^~\-'^-'b^+ ...
i . A
ffl(l+m)(2 + m)^_(^^3)^3
= J_ ^^ m.m + 1.62 m(m + l)(m + 2)73 ,
a^ ~ a"*+^ "^ |_2 a"*-^" |_3 qT'^^
the terms having alternate signs.
(a — 6)""* would be similar but have all the terms positive.
XXXIV. ] BINOMIAL EXPANSION. 319
Hence also •
/I ^ \-m 1- m.m + 1 „_m. 771 + 1. m + 2 „ , "
(l±x)-"' = 1+mx-^ — r-^ ^ + ■ I Q ■ ^ + + etc.,
where it will be observed that with the -H sign on the left,
the terms on the right are alternately + and - ; but with
the — sign on the left they are all + on the right.
The series is infinite, but if x is small a few terms practically
sufiice.
Examples.
Take some examples or special cases :
(l+ic)-i = I-x + yttX^ r^—x^+...
= 1 -X-{-X'^-X^+X*-X^+ ... y
of which only a few terms are important if x is small, e
p^ = (l-Ol)-i = 1-01 + -0001- -000001+ .
= -990099... ib-9901,
7YJ^ = (l+a;)-2 = l-2x + Sx'^-4:X^ + 5x^-...,
J(l-hx) = (l+a:)^= i + ix + i^^x''
+ [^ ^+-
v/r^ = 1 - la: - 1^2 - ^^x^ - ^^x^ - ... ,
-i(-i-2)(-i-2)^
il
320 EASY MATHEMATICS. [chap.
A curious case is afforded when both terms of the binomial
are unity, like (1 + 1)". When the index n is not a positive
integer the series is divergent and useless; but when ti is a
positive integer it is simple enough, for the sum is finite. It
is a mere curiosity, but we may as well find a power in
this way.
For instance to find 2^,
,, ,,, , . 5.4 5.4.3 5.4.3.2 |5
(1 + 1)^ = 1+54-^- + -^^+—^^— + !^
= 1+5 + 10 + 10 + 5 + 1 = 32.
Similarly 2^ = 1+6 + 15 + 20 + 15 + 6 + 1 = 64.
This set of numbers, as tabulated in their early stages on
pages 311 and 334, are called the binomial coefficients; and
you observe that each set of them adds up to a power of 2.
We had not noticed this before.
Now what is the good of an expansion generally 1 Is it of
any practical use? Well it is, but it is the first few terms
which are the most useful. The expansion of some power of
(1 +a;) is specially useful when x is very small, for then
^(l+a:)^l + |a;.
This approximation is said to be correct to the first order of
small quantities, or to be an approximation of the first order.
To be correct to the second order of small quantities we
must introduce the terms involving «2, and so on.
When X is only moderately small, third and even fourth
terms may have to be employed, and the more terms intro-
duced the more accurate will the result be.
XXXIV.] INFINITE SERIES. 321
If X is greater than 1, the series becomes hopeless, but if x
is only slightly less than 1, it can always be approximated to
sufficiently, by taking enough terms, though it is not then
really useful.
The series is said to be convergent or converging when x is
less than 1. A converging series is one whose terms con-
tinually decrease in such a way that the sum of an infinite
number of them is finite.
For instance, l+J + |- + J^ + yV+'-is ^ converging series,
and its value, to an infinite number of terms, is 2 ;
but 1 + 1 + T. + J+1+ ...
happens not to be convergent, for though the terms keep on
diminishing, they do not diminish with sufficient rapidity to
be able to stop at any point and say 'we will neglect the rest.'
Those which we neglected would in fact amount to more than
those we took into account, for the sum of an infinite number
of terms of such a series is infinite. It is not a convergent
series at all, although each term is smaller than the preceding
one. A curious case.
The first is called a geometrical progression, the second is
called a harmonic progression, because it gives the series of
the harmonics or simplest overtones in music. The time of
vibration of the fundamental note being called 1, a trained ear
can hear, when a string is struck or plucked or bowed, or
when an open organ pipe is blown, other superposed notes,
with their times of vibration \^ ^, J, etc., of the first ; and
these superposed or secondary tones are called harmonics. So
the series is called a harmonic series.
An arithmetical series is one whose terms proceed by simple
addition. In a harmonic series it is the denominators or
reciprocals of the terms which proceed in this way. For-
tunately we seldom or never want to sum a harmonic series.
L.E.M.
CHAPTER XXXV.
Progressions.
We have now, in the last chapter, arrived at an example of
a series or progression. The subject of 'series' is immense
and endless, but there are a few simple ones which are excep-
tionally easy to deal with.
Of these, three are commonly treated quite early, viz. the
three called Arithmetical, Geometric, Harmonic, respectively.
In an arithmetical series the terms proceed by a common
difference.
In a geometric series the terms proceed by a common factor.
In a harmonic series the reciprocals proceed by a common
difference.
Thus 1, 2, 3, 4, 5, ... is the simplest example of an a.p.
1, 2, 4, 8, 16, ... „ „ „ G.P.
11111 TT p
But the common difference may be negative, or the common
factor less than 1, so that
7, 6, 5, 4, 3, 2, 1, 0, - 1, - 2, ... is an example of A.P.
1111 n T>
1111—1-1 TTT>
T> T> ^> ^J ^J T» ••• " >» "•^•
Also 1, 1-25, 1-5, 1-75, 2 ... is an a.p.
1000, 100, 10, 1, -1, -01, ... is a G.P.
^ + A + F+ 1 + 6 + 36 + 216 ... is also a G.P.
-1, -1-6, -5, 5, 1-6, 1, -714285, -5, -45 is an H.P.
CHAP. XXXV.] PROGRESSIONS. 323
The latter is perhaps too much disguised for a beginner, but
if the terms be written as vulgar fractions it is plain enough :
the denominators are in A.p. with common difference 2, for it
is the same as
5
5
5 5 5 5 5 5
5
-5'
-3'
-1' 1' 3' 5' r 9'
11'
I
The thing that generally has to be done with a series is to
evaluate the sum of its terms ; and the most important are
those whose terms decrease, so that an infinite number of
them have a finite sum, which can be ascertained. Otherwise
we must know how many terms we are intended to add up.
Another thing that may be necessary to do with a series,
especially those which do not converge, or which actually
increase as they go on, is to find the value of the nth term.
Thus in the horse-shoe nail question, page 155, we had
really to add 24 terms of a G.P., beginning with 1 and proceed-
ing by a common factor 2 ; but the finding of the 25th term
of that series was sufficient, because we could see experi-
mentally that each term was almost precisely equal to the
sum of all that had preceded it, being, as a matter of fact,
just one in excess. (Compare page 218.)
1 -h 2 with 1 added made the next term 4.
1+2 + 4 „ „ „ „ 8.
1 + 24-4-J-8 „ „ „ „ 16.
1 + 2 + 4 + 8 + 16 „ „ „ „ 32.
etc.
So all the first 24 terms, with 1 added, were equal to the
25th term of the series.
But the 25th term was 2-% therefore the whole 24 terms of
the series, added up, was one less than 2^* ; that is to say the
series equalled 2^4-1.
324 EASY MATHEMATICS. [chap.
Of the three progressions we have mentioned, G.P. is
certainly the most commonly occurring and the most useful.
Let us take it first.
Let the common factor be called r, and the first term a, so
that the terms run thus
a-\-ar + ar^ + ar^ + . . . ,
r being any number whole or fractional.
If ?•, when interpreted arithmetically, is a negative quantity
the terms will have alternately opposite signs, and the result
will be a combination of alternate addition and subtraction ;
which however can conveniently be called the algebraic sum,
meaning the sum when written algebraically with sign implied
but unexpressed, but of course subtracting from the series
those terms with negative signs when arithmetical interpre-
tation is entered upon.
One sees at once that since the second term is
ar
and the third a?-2,
the fourth ar^^
the nth. term must be a?'""^
The sum of the first n terms will therefore be
-: - a(l +r + r^-l- ... +r""^).
Now this is a thing we have already come across ; it was
when r was small, that is to say
= 1 + r + r2 + r^ . . . ,
but r must be less than 1, or the series will not converge;
every term will get bigger than the preceding one if r is
greater than 1, and there would be no meaning except infinity
in an infinite number of such terms. But the expansion is
XXXV.] PROGHESSIONS. 325
only true for an infinite number of terms; consequently it
is only serviceable when r is less than 1.
However, that is a very important case : the most important
case. Let us apply that to a few examples before we go further.
Find the sum 1 + i + i + 1 • ■ • •
Here a = 1 and r = i,
the sum then will be 1 x ( j = - — - = 2, which we already
knew. (pp. 321 and 100.)^^ "^'^ ^~2
This series can be well illustrated by cutting up an apple or
a loaf of bread ; for if such an object be taken and first a half
cut off", then a quarter, then an eighth, then a sixteenth, and
so on, all the cutting can be performed on a single object, and
however long the cutting be continued the single unit will
not be exhausted : and yet if the cutting be continued ad
infinitum the apple will be all exactly used up. In other
words, although the sum of any finite number of terms of
the series J, J, \^ etc., is less than unity, the sum of the
infinite series of these fractions is exactly one whole, no more
and no less, that is to say
i + T + i + TV+---«^*V- = 1-
As another example take
1 + -1 + -01 + -001 + ...;
-.1 1 1 10 , .
It equals
1-1 -9
T=^^'
as is otherwise obvious by simp]
le addition of the terms.
Again
12 + 444 +
4 4
27 + 8T+--
The sum
A-^ ^
= 12x|=18.
Another
way of putting it is
to say that
1 1 1
3'^9'^27"^""
= r = 1.
326 EASY Mathematics. [chap.
The series 10 + 9 + 8*1 +7*29 + ..., to infinity, is a G.P. that
does not decrease very fast, but it converges nevertheless, and
the value towards which it converges, constantly approaching
though never actually reaching, is — — as usual, that is 100 ;
the sum to infinity = :j ^ = j^ — ^ = 100 exactly.
In general, so long as r is less than 1, it matters not how
little less, the series will converge, and we can find the sum of
an infinite number of terms. Suppose the common ratio were
•999 for instance, and the first term were 1, the sum to
infinity would be ^-^^, that is to say 1000. If the first
1 ~ To IT (7
term of this series had been anything else than 1, say 56 for
instance, the sum would have been merely 56000. Or if the
first term had been 4*35782, or any number you please,
the sum would have been 4357*82, if the common factor were
*999 as supposed.
The first term therefore causes no difficulty, it is the com-
mon ratio or factor that requires attention when a finite
number of terms is wanted; and a finite number of terms
always is required whenever the common factor is greater
than 1, and often is when it is less.
How are we to find the sum of n terms then ?
It can be done by a contrivance : .
Write down the series, and then write' it down again with
every term multiplied by r, and then subtract the two series,
thus :
Call *S^ the sum of n consecutive terms of the series.
,S^ =
1 + r + 7*2 + r^ + .
..7-1;
.*. •
r^ =
r + r2 + ?-3 + . ,
..r"-l + r^
Now subtract
^-
-rS= 1
_,.n
because all the other terms
go
out;
XXXV.] PROGRESSIONS. 327
I _ 7-" if
therefore S = -^ , or
1-7' r-V
which is the same thing.
If the first term is (X, then the above expression has to be
multiplied by a ; so that in general, whatever r may be, the
sum of n terms of a geometrical progression is
5 = a r-.
r- 1
If n should be 00 there is no finite meaning in the series
unless r is less than 1 ; in that case r" = 0, because higher
powers of a proper fraction keep on diminishing, so an infinite
power must disappear altogether ; we then get the case which
we already know, viz.
Examples.
Apply this to the sum of 24 horse-shoe nails with one
farthing for the first, and with common factor 2. (p. 156.)
224 _ \
Ans. : The price is a -^ — 3- = 2^4-1 farthings.
Find the sum of six terms of the series
100 + 200 + 400 + etc.
06 _ 1
Jns.: It equals lOOx-^r — - = 6300.
Z — L
Find the sum of 1 + 3 + 9 + 27 + etc. to six terms.
. ^, 36-1 728 ^^,
Ans.: The sum = -^ — ^ = -^ = 364.
Find the value of 64 + 16 + 4 + 1 +i + y^ + ^V
Ans. : This is a G.P. of seven terms with common ratio J
and first term 64.
s. . = ..-:iffi.J.e..(-^.)
256
~ 3
328 EASY MATHEMATICS. [chap.
1 — r**
The numerator of the fraction , in a case of many
terms with a fractional ratio, is of small significance : it is
nearly unity.
Algebraic Digression.
The result we have arrived at, as the sum of a G.P., may be
regarded as an expansion for an algebraical division.
1 -?•"
T = l+r-\-r^ + r^+ ... .
1 -r
This might be generalised hypothetically thus,
x~y
which could be verified by direct division, or more easily
by multiplication, and could be led up to experimentally thus :
x^ -y^
x-y
— = x'^ + xij + y^j
X — y
— ^ x^-\- x^y + xy"^ + «/3,
and so on.
If we try the positive sign between the terms on the left,
the matter is a little more troublesome.
Try it first in the denominator only :
— = X-y,
x + y ^
— ^1^ will not go without a remainder,
x + y ^
t^ = x^-x'^y + xf--y\
0^ -y^
X + y
and so on. •
will not go again.
XXXV.] ALGEBRAIC DIVISIONS. 329
Now try the positive sign in numerator only:
^ -^ will not fifo, i.e. will have a remainder ;
x-y
- — ^ will not ffo either.
x-y ^
Now try positive signs in both numerator and denominator:
x^ + y-
x + y
x^ + y^
x^ + y^
x + y
x^ + y^
will not go,
xy + y'^;
will not go,
c* - x^y + a^y - xy^ + y*.
So it makes all the difference whether the indices are even
or odd. All the above can easily be verified by direct
operation ; and the reason of the failure to divide out, when
they do fail, will also be manifest on trial. The reason is that
the last term, the y^ or yS etc., would have the wrong sign.
To sum up what we have observed :
a;" - ?/" is divisible hy x-y whatever n is,
and likewise hy x + y when n is even.
«" + y'" is divisible hy x + y when n is odd,
but is not divisible by x-y, whatever n is; understanding
by " divisible," divisible without a remainder, that is, that the
denominator is a fadm- of the numerator ; and understanding
by n always a positive integer.
Another way of putting it is as follows :
x-y and x + y are both factors of «" - y" if n is even,
x-y only is a factor of x" -y" ii n is odd,
x + y only is a factor of xT" + y'^ ii n is odd,
neither is a factor of x" + y^ if n is even.
330 EASY MATHEMATICS. [chap.
Or thus, which forms an easy way of remembering the facts :
^.3 _|. yZ ig divisible by aj + y,
a;3 _ yZ ig divisible by « - y,
x^ - y2 is divisible by both,
01? 4- ip' is divisible by neither.
General expression for any odd number.
It is handy to be able to discriminate between an odd and
an even number algebraically.
It is done thus :
2/1 is always even, if n is an integer.
In ± 1 is always odd, again if n is an integer.
The wth odd number is ^n-\ (hence this is commonly the
expression used for an odd number) ;
e.g. 5 is the third odd number and is equal to (2 x 3) - 1 ;
11 is the sixth odd number and is equal to (2 x 6) - 1 ;
and so on.
The hundredth odd number is therefore 199 and the 365th
is 729.
Arithmetical Progression.
Now take some examples of A. P.
An interesting case is to find the sum of the first n
consecutive odd numbers added together, that is to find the
value of
1+34-5 + . .. + (271-1).
This sum might be found by experiment, thus :
1 + 3 = 4 = 22,
1+3 + 5 = 9 = 32,
1+3 + 5 + 7 = 16 = 42.
So the sum of the first four odd numbers is 42 and of the
first five will be found to be 52 = 25.
XXXV.]
PROGRESSIONS.
331
9
7
5
3
1
Fig. 84.
Trying a few more we come to the experimental conclusion
that the sum of the first n odd numbers will be n^.
1 + 3 + 5 + 7 + .. . + 27Z-1 = n\
The annexed diagram illustrates to the eye the facts that
1 + 3 = the square of 2,
1 + 3 + 5 = the square of 3,
1+3 + 5 + 7 = the square of 4,
and so on.
The sum of the first n natural num-
bers is not so simple, but is a good
problem to solve experimentally, thus :
1 + 2 = 3,
1+2 + 3 = 6,
1 + 2 + 3 + 4 = 10.
So we have a series, rather a notable series, with differences
increasing by 1 each time.
For the sum is
one term two terms three terms four terms five terms six terms
13 6 10 15 21, etc.
2x5 21x6 3x7
What would be the number thus to be placed under n terms %
The answer is \n{n-{-\)', and it is possible but not quite
easy to guess that.
It is worth remembering however : the sum of the first n
natural numbers is \n{n-\- 1).
As for the sum of the even numbers, that is a very simple
matter, for it is merely double the preceding ; or it may be
regarded as the sum of the odd numbers plus n ;
2 + 4 + 6 + 8 + 10... + 2n = 7i2 + ^
= w(7i+l).
This method of guessing and verifying will not carry us
far : a reasoned process of arriving at a result is far more
332 EASY MATHEMATICS. [chat.
powerful and effective. Algebra enables us to reason things
out; and the customary method for the sum of an A. P. is
as follows :
Let the general arithmetical progression be the following,
to n terms,
a, a + b, a-\-2b, a + Sb ... a + {n-l)b;
write it again, but backwards,
a + {n-l)bj a + (n-2)b, a + {n-3)b, ....... a.
Now add the two series together, term by term, as they
stand one under the other ; and the result will be 2a+{n-l )b
every time.
Hence, since there are n terms, the result of the double
series added together, if *S' is the sum of a single series, will be
2,8' = n{2a + {n-l)b};
.'. S = na + ln(n-l)b.
This is the general result for an A. P.
For example, to test it by speci-al cases :
In the case of the first n natural numbers a = b = 1, and so
*S' = ni-^n{n-l)
= in(n + l),
as we have already found by experiment.
In the case of the first n odd numbers, a is 1 and b = 2 ;
S = n + n(n- 1) = n^,
as we also found experimentally.
It is very instructive and pleasing to see how a general
formula thus gives special cases, and it is one of the verifications
by which a general formula should always be tested.
The following is interesting for practice :
1 = P
3 + 5 =23
7 + 9 + 11 =33
13 + 15+17 + 19 = 43
etc. . .
XXXV.] SERIES. 333
Other Series.
The number of series or progressions that can be dealt with
is enormous, is indeed infinite; and is too large a subject for us
to enter upon in this book. Suffice it to say that many others
occur in practice besides the simple ones which are best known.
This series, for instance,
124-22 + 32 + 42
is neither a geometric nor an arithmetic nor a harmonic
progression. Something like it occurs in che overtone fre-
quencies of vibration of plates and bars.
Manifestly we might have
12, 32, 52, .. ,
or P, 23, 43, 83,
and so on ; any number of such series could be invented.
There is one simple series that we came across recently
on p^ge 331, the difference of whose terms was constantly
and steadily increasing: the series 1, 3, 6, 10, 15, etc.
If we started with this series and took the differences we
should get an A. p. series, and this is a process we might
continue ; thus :
Start with this,
0 1 3 6 10 15 21 28....
Take differences,
12 3 4 5 6 7 ...anA.P.
Take second differences,
111111 a series of constants.
Take third differences,
0 0 0 0 0 a series of zeros.
Suppose we start with a different series, say the natural
series of square numbers,
0, 1, 4, 9, 16, 25,
the differences of these will give the series of odd numbers ;
while the second differences would be constant.
334 EASY MATHEMATICS. [chap.
If we took any geometrical series the differences would be
the same series again, multiplied by a factor, the factor being
one less than the common ratio.
Hence the differences of the powers of 2, viz. 1, 2, 4, 8, 16, 32,
would be the same series over again.
The binomial coefficients can be obtained by interjecting a
single 1 into the middle of a row of noughts and then adding
adjacent terms to make a term of the next series, as thus,
00000 10 0000
0000 1 10000
0 0 0 1 2 1 0 0 0
0 0 13 3 10 0
0 14 6 4 10
1 5 10 10 5 1
1 6 15 20 15 6 1
1 7 21 .35 35 21 7 1
18 28 56 70 56 28 8 1,
19 36 84 126 126 84 36 9 1
1 10 45 120 210 252 210 120 45 10 1
1 11 55 165 330 462 462 330 165 55 11 1
1 12 66 220 495 792 924 792 495 220 66 12 1
The simplest illustration of an arithmetical progression is
the natural series of numbers — the ordinary counting of a
child. The most important instance of an arithmetical
progression that occurs in nature is afforded by time. It is
true that it progresses continuously and not by jerks, but the
motion of a clock hand is a jerky motion, and the succession
of days, weeks, and years divide the continuum into units for
measuring purposes, and represent a perfectly uniform and
inexorable constant rate of progress.
A set of numbers are said to be in geometrical progression
when their logarithms are in arithmetical progression. The
notes of a piano are in this predicament, when estimated by
their vibration frequencies. The chromatic scale, on a tempered
XXXV.]
PROGRESSIONS.
335
instrument like a piano, proceeds by equal musical intervals,
but these intervals are characterised by equal ratios of
vibration frequency, every octave having double the vibration
rate of its predecessor ; in other words the factor 2 carries us
over the interval of an octave, the factor f gives successive
fifths, and so on ; so that the same musical interval, in
different parts of the register, is characterised by a constant
difference of logarithm.
A set of numbers are said to be in harmonic progression
when their reciprocals are in arithmetical progression.
The series of square numbers have their roots proceeding
in A. P.; another series we have encountered has consecutive
differences in A. P. ; another series might have successive ratios
in A. P. ; and so on.
Geometrical Illustrations.
The heights of a row of palings may be used to illustrate the
three best known modes of progression, if their tops all reach
a sloping straight line. If they are spaced simply at equal
intervals, they of course form an A. P. ; if they are spaced so
that lines drawn from the foot of each to the top of the next
are all parallel, they will form a g.p. ; but if spaced so that a line
from the foot of each to the top of the next-but-one bisects the
intermediate one, they form an H.p. The three figures annexed
illustrate this.
Arithmetical
Fig. 35.
Concerning fig. 35 there is nothing to be said but what is
obvious.
336
EASY MATHEMATICS.
[chap.
Concerning fig. 36 it can be pointed out that the triangles
formed are all similar, that the lengths of the slant lines are
in G.P. as well as the vertical lines ; and so are the areas of the
Geometrical
Fig. 36.
triangles. They may be said to illustrate the successive heights
attained by a bouncing ball : which heights are also in G.P.
Fig. 37 is the most notable; it may be regarded as the
perspective view of a series of equal rectangles or parallelo-
grams— the perspective view in fact of a uniform fence. Hence
it is useful in drawing metrical perspective figures.
Proofs. — The proof that fig. 36 represents a geometrical pro-
gression is almost obvious, since by construction the triangles
are similar, their sides being parallel ; hence
h^b=:yi = h==h etc
^1 h y^ h yi
The fact that fig. 37 gives a harmonic progression can be
established thus :
Let a, Z>, c be three verticals erected so that a line from the
foot of a to the top of c, or from the foot of c to the top of a,
bisects the intermediate height &, which therefore divides the
XXXV,]
PROGRESSIONS.
337
base in some ratio m : n, then it can be shown that
harmonic mean of a and c ; for by similar triangles
is the
ib_ n
a m + n
and ^* _ ™ ;
c m + n
therefore
a c
or
1 1 2
FiQ. 38.
wherefore the reciprocals are in arithmetical progression.
In fig. 37 it is convenient to call the slant lines transversals,
and to say that the transversal from the foot of each passes
through the mid-point of the next to the top of the next-but-one.
Another geometrical illustration of a g.p. is the following —
a sort of straight line spiral; the inclinations being any constant
angle other than 45°, the vertex angles being 90°.
L.E.M. Y
338
EASY MATHEMATICS.
[chap. XXXV.
Successive sides of the spiral are in g.p., and so are the
distances of successive vertices from the centre.
To convert fig. 35 or
fig. 41 into the repre-
sentation of a G.P. as
it stands, the roof must
be made of a logarith-
mic or exponential
curve instead of a
straight line.
Thus fig. 9 and
fig. 47 already repre-
sent a G.P. ; each verti-
cal height is the Geo-
metric Mean of any
pair of heights equi-
distant on either side
of it.
The 'amplitudes' of
the swings of a dying-
out pendulum consti
tuteaG.R: the 'periods'
of successive swings
constitute an A. P.
See (fig. 40.)
The temperatures of
a cooling body, read
every minute, consti-
tute an approximate
G.P., and if plotted would give a logarithmic curve : looking
like fig. 9 or fig. 78, or the dotted line in fig. 40, or part of
the figure on page 179-
CHAPTER XXXVI.
Means.
A THING of some interest and use is the mean or average of
a set or a pair of terms in a progression. In an A. P. the mean
can be found by adding and halving the two extreme terms.
Thus for instance in the progression
7 9 11 13 15
1 1 is the mean term, and it can be found as the half sum of 9
and 13, or the half sum of 7 and 15.
The arithmetic mean of a and c
is \{a + c); for calling this h, it
makes h -a = c-h, that is, it gives
a common difference in the pro-
gression a, b, c; and it is illustrated
by the figure, where b is the mean
height of the trapezium shown, whose area is therefore h
times the base.
Fig. 41.
The arith
metic mean of 1 and 7 is 4
of 0 and 100 is 50
of 0 and 9 is 4J
of 6 and 16 is 11
of - 1 and + 1 is 0
of - 6 and + 8 is 1
of - 3 and + 9 is +3
of - 9 and + 3 is - 3
of 12 and 90 is 51
because 51 -
- 12 = '39, and 90 - 51 = 39 ; or because
51 = 6 + 45.
340 EASY MATHEMATICS. [chap.
In general then the arithmetic mean is the half sum, the
sum being understood as the algebraic sum, paying attention
to sign.
The geometric mean of two terms is the square root of their
product, because this would give a common ratio; thus if
three terms a^ b, c are in a. p., b must equal J{ac)f because
^(ac) :a = G : sj(ac) = \l[~), the common ratio ;
* be If
in other words the common ratio - or r must be equal to ^/-.
In the progression a, ar^ ar^ the middle term is plainly the
square root of the product of the end terms.
The Geometric Mean is also called a " mean proportional."
To illustrate a geometric mean it is customary to use either
a right-angled triangle or a circle. Thus if the two lengths
whose geometric mean is required are CA and CB, any circle
drawn through A and B has the property that its tangent
drawn from C is equal to the geometric mean required;
for by Euclid III. 36, CF^ = CA.CB; hence incidentally
we arrive at the proposition that
all the circles that can be drawn
through the two points A and B
can be cut at right angles by a
certain circle drawn round C as
centre; because the length CP is
constant. If only the points A and
B had been initially given, then a
number of such points C could be found, each with its ap-
propriate length of radius, by drawing a tangent, or a series
of tangents, to any one of the circles.
If a right-angled triangle ABC be drawn, and a per-
pendicular be let fall from the right angle C on to the opposite
XXXVI.]
MEANS.
341
side, the length of this perpendicular is a mean proportional
between the segments of the base :
CD = J AD . DB, since -tttt = ^Diy
Similarly ^C is a mean proportional between AB and AD,
and BC „ ,,
The same thing is true for
a semicircle, since the angle
in a semicircle is a right angle.
Euclid III. 31.
Hence an easy construction
for finding the Geometric
Mean of two lengths is to
place them end to end, as
AD, DB, construct a semicircle on the whole length AB thus
compounded, and erect a perpendicular DC at the junction
point of the two lengths. This is the g.m. required. (Cf. p. 274.)
Or if the two given lengths had been A B sand AD, then the
distance AC would be their G.M.
The harmonic mean of two terms
is su(5h that it would be the arithmetic
mean of the two terms inverted.
For instance, J, J, J are in H.P.,
Fig. 43.
Fig. 44.
and
is the harmonic mean be-
tween ^ and J.
arithmetical progression, and - = - ( -
0 A \0j
Let a, b, c be in harmonic progression, then -^ j-^
are in
wherefore
2ac
a + c
The harmonic mean can therefore be described as twice the
product divided by the sum.
342
EASY MATHEMATICS.
[chap.
Geometrically it could be represented by setting up the two
given numbers as parallel measured lengths, like a and c, and
joining their ends both direct
and crosswise.
Then the parallel drawn
through the crossing point is
the harmonic mean of a and c.
It is represented by a dotted
line in the figure.
The proof of this construction is given above in connexion
with figs. 37 and 38.
If the two outer lines of the figure are continued till they
meet, the fourth position thus determined forms a H.P. with
the three positions determined by the crossings depicted in the
figure ; and they are familiar in elementary geometrical
optics.
Examples.
The student should cover up
the
right-hand
column and
reckon the entries in it. They are all intended to be done in
the head.
The geometric mean of
1
and
9
is
3.
of
9
and
9
is
9.
of
1
and
100
is
10.
of
0
and
100
is
0.
of
0
and
n
is
0.
of
9
and
36
is
18.
of
4
and
25
is
10,
or generally, of
W2
and
^2
is
mn.
of
40-5 and
24-5
is
31-5,
that is to say of
|-m2
and
|7l2
is
\mn.
of
1
and
81
is
4-5.
of
7
and
I
is
1.
xxxvi.] MEANS. 34S
is 1-3.
is 6.
is 2-3,
is 2-3.
the G.M. of
1
and
16
of
1
T
and
252
of
147
and
1
2T
of
49
and
1
9
or generally, of
am'^
and
1
— , IS
m
of am and — is a,
m
e.g. of 72 and 8 is 24.
of - 1 6 and - 1 is - 4.
of -16 and -100 is -40.
of -i and -72 is -6.
of -14 and -7 is -7^2.
of ab and a is ajb.
To find the harmonic mean b of two terms a and c, we can
write down that — t = i j
a 0 b c
^, ^ b-a c-b
or that — J— = -^ — ,
ab be
, , b-a a
or that
c-b c
The two most usual expressions for a three-term harmonic
1 1 2
progression are _ + - = -,
J 7 2ac
and b = — ■ — ;
a + c
these are simple and useful, and we come across them con-
stantly in elementary optics. In reflexions from a small
curved mirror the object and image and the centre of curvature
of the mirror are situated at points whose distances from the
mirror itself run in a harmonic progression.
So also in lenses, if a is the distance of object from the lens,
and c the distance of the image, ^b is its focal length.
This is set down here merely as a reminder.
344
EASY MATHEMATICS.
[chap.
Illustrations. — The harmonic mean of J and |^ must be i ;
and twice the product by the sum of the two numbers is
if 120 5*
2x1x1.
1 4- 1
T + F
The harmonic mean of 4 and 6 is
2x4x6 48 , ^
-TT6- = io==^'^-
2 X 99
The harmonic mean of 1 and 99 = ..^^ = 1-98.
100
The harmonic mean of 0 and 1 is 0.
The harmonic mean of 17 and 13 is
34x13 442
30 ~ ~ 30
The harmonic mean of
of
14-7i
-2and+6 = ^
2 and +2 = -oo.
70
24
+ 4
= -6.
of - 5 and - 7 =
of - 1 and - 9 =
-12
18
10
— f^s
= -1-8.
of 1 and 00 = 2.
The geometric mean of 6 and 20 is Jb x 20
ratio being 2.
The geometric mean of
of
of
of
of
of
10, the common
4.
1 and 16 is
2 and 32 is 8.
4 and 9 is 6.
8 and 2 is 4.
8 and - 2 is - 4.
8 and + 2 is imaginary.
of - 1 and - 9 is - 3.
of a and b is J(ab).
Comparing the three means of the same two quantities a and by
XXXVI.] • MEANS. 345
the arithmetic mean is J {a + h) and is the biggest of the three,
the geometric mean is sfabj
the harmonic mean is f and is the smallest of the three.
a + h
The H.M. may be considered as ^^-^ — —. that is the ratio of
^ A.M. '
the square of the G.M. to the A.M. ; or it is equal to the G.M.
multiplied by the proper fraction g.m./a.m.
Examples.
Take any two numbers, say 4 and 9 :
the arithmetic mean is 6*5,
the geometric mean is 6*0,
the harmonic mean is ^f = 5'53846... .
Let the two numbers be 1 and 25 :
the A.M. is 13,
the G.M. is 5,
the H.M. is fl = 1-923... .
Let the two numbers be 49 and 36 :
the A.M. is 42-5,
the G.M. is 42-0,
/42)2 49
theH.M.is^ = ~x42,
which is necessarily less than the G.M.
Let the two numbers be 0 and 4 :
the A.M. is 2,
the G.M. is 0,
the H.M. is 0,
but not the same 0, it is half the G.M squared.
Let the two numbers be 6 and - 6 ;
the A.M. is 0 :
the G.M. is imaginary,
the H.M. is - 00 .
346 EASY MATHEMATrcS. [chap.
Let the two numbers be - 3 and ~ 6 :
the A.M. is - 4-5,
the G.M. is - 372 = - 4-2426,
the H.M. is - 4-0.
In some cases the order of numerical magnitude is inverted ;
but, when compared with positive, the smallest negative
quantity is represented by the largest number. If heights of
mountains were reckoned from sky instead of from earth, as
by dropping a plummet from a balloon, the lowest mountain
would need the longest plumb-line to reach it. The lowest
parts of the solid earth are beneath the sea and require a long
sounding line to reach them.
The A.M. is of course always half way between the two
numbers.
The G.M. is nearer to the smaller one, and ultimately
coincident with the smaller one when the other is infinitely
bigger.-
The H.M. is less than the G.M. in the ratio of — — ' \ or
A.M.
H.M. X A.M. = G.M.2, or the G.M. is a mean proportional between
the other two.
Mean or Average of a Number of Terms.
In taking a mean of terms there is no need whatever that
those terms should form any sort of progression or ordered
series. Hitherto we have only taken the mean of two terms,
and two terms cannot possibly determine any kind of pro-
gression, any more than two points can determine a curve.
But we can reckon the arithmetical or the geometrical mean of
any number of terms as follows :
Suppose we want the mean of a set of observations of
temperature, taken at every hour of the day, so as to determine
the mean temperature during the day of 12 hours, say from
XXXVI.]
MEANS.
347
8 a.m. to 8 p.m. Let the thermometer readings be the
following — there will be 13 readings, because of the beginning
and end points of time between which the twelve hours lie :
Add them up and divide by the number of
them, that is by 13. This is the mean or
average of the readings, and is found to be
67 "58. It is apparently a summer day with a
warm and probably cloudy morning giving place
to a clearer sky and cooler evening.
If the temperature readings were plotted and
joined, the result would be a curve (fig. 46) ;
and the average height of this curve would be
the mean temperature.
The average height must be approximately
67*58; but when the curve is drawn by a
recording thermometer, so as to give the
temperature not only at every hour, but at every instant, a
more exact determination of the average can be made.
60-5
65-0
67-2
67-3
71-25
75-0
79-0
77-0
74-6
70-3
62-4
553
53-7
13 I 878-5
67-58
■---
zzz
.^^
-^^
^
- —
"^
X
^
Mean
Temp.
8
a.m. £
) 1
0 1
1 1
2
p.nui
Fig
2 C
46.
} <
I I
3 e
i
T i
i
The average or mean height of such a curve is the height of
a rectangle with the same base which shall equal the curve in
area, as shown in the figure by the dotted line.
One way of finding out the area of such a curve, and in
that way of obtaining its mean height, is to cut the above figure
out in cardboard or tinfoil or sheet lead, and then weigh it ;
348 EASY MATHEMATICS. [chap.
weighing at the same time a rectangle or square of known
dimensions cut out of the same sheet.
Thus suppose the curve carefully drawn on such a uniform
sheet, on a scale which gave 1 horizontal centimetre to each
hour and 1 vertical millimetre to each degree, and that the
figure bounded as above was carefully cut out and found to
weigh 4'98 grammes; while a rectangle of the same base 12
centimetres, and height 7 centimetres, was found to weigh
5*16 grammes.
We should know that the area of the curve-bounded figure
was ^7j-^x 7 X 12 sq. centimetres, and that its average height
was — -— X 70 millimetres.
516
This process would give the average height, and therefore
the number of degrees in the average temperature, as 67 '56 ;
and, if carefully carried out, it should be more correct than
merely averaging numerical readings taken each hour, for it
averages the temperature recorded from instant to instant.
To follow a process of this kind profitably, the best plan is
actually to do it, and then the method of working will
naturally occur to you with a little thought; and a good
result can be obtained with some handicraft skill. It is a
practical method of experimentally performing the operation
known as integrating; it is integration between definite
limits, or the finding of a definite integral of a function
represented by a curve.
Weighted Mean.
Generally the word " mean " implies the simple arithmetic
mean, and the niean of several numbers, say Wj, n^ and n^, is
yii + yi2 + % ^^.^j^ .g ^i^g^ written n and read "ri bar."
XXXVI.] MEANS. 349
The average of a^, a^.-.a^ is a =- -(ttj + a2+ ••• +<^n) ^ but
sometimes we have to do with a weighted mean. One case is
when a series of observations of the same thing are taken
under different circumstances, and some of the circumstances
are more favourable than others; for instance if the height
of a flagstaff were being measured by the length of its shadow,
at noon on successive days at the time of the summer solstice,
and suppose the record of the shadow measurements was
entered thus :
15-46 feet day fair observer W. Smith.
15*30 „ cloudy „ „
15'47 „ day bright observer E. Jones.
15 '50 „ weather hazy ,, „
25-6 „ day bright observer J. Williams,
and the most probable result were required.
First of all the last observation would have to be thrown
aside altogether, because Williams has evidently made a
mistake in the very first significant figure, and although his
observation may be correct in the last figure and almost
certainly means 15*6, it is hardly safe to begin doctoring
results ; it is safer to reject any that thus show obvious signs
of carelessness.
The other observations may have different weights attached
to them, and to know how to attach weights satisfactorily
needs considerable experience, and experience too with these
same observers, because it may happen that Jones is known to
be a more trustworthy and exact observer than Smith ; each
person has what is called a " personal equation," some always
tend to read slightly too large, others slightly too small, while
others cannot be trusted to more than say three significant
figures.
Let us suppose that an experienced person decides to attach
350 EASY MATHEMATICS. [chap.
the weight 3 to the first observation, the weight 1 to the
second because of the clouds, the weight 6 to the next because
Jones is a good observer, and the weight 2 to the next
because of the haze, the weighted mean of the set of obser-
vations would be obtained as follows, attending only to the
decimal places because the 15 is common to all :
(3 X -46) + (1 + -30) + (6 X -47) + (2 x -50)
3+1+6+2
1-38 + -30 + 2-82 + 1-00 5-50 ,^„^
= ^ .= -^ = -4583;
wherefore the result as thus determined would be given as
15'458, or say 15*46 feet to four significant figures: some
probable error affecting the last place.
General Average.
An average in general may be better expressed thus :
Let Tij observations give a result x^
let Tin ,, ,, ., Xa
"^G " " " "^C'
then the total number of observations is Wj + ?i2+ •••^6' ^^^
the appropriate weighted mean or average of the whole
number of observations,
W^ + ^2 + • • • ^6
1 vx - ^(nx)
commonly written as ic = ^\ /,
2(71)
the 2 being read " sum."
This is a most important and commonly occurring form of
average, or arithmetic mean, of any number of like and unliko
quantities.
XXXVI.] MEANS. 351
Geometric Mean of several Numbers.
To find the g.m. of two quantities we multiply them
together and extract the square root. Similarly to find the
G.M. of three quantities we should multiply them together and
extract the cube root; and to find the g.m. of say six
quantities, multiply them all and extract the sixth root. But
this would have to be done by logarithms ; so the process is
better put into logarithmic form from the first.
To find the geometric mean of n quantities
Find the arithmetic mean of their logarithms :
the resulting a as thus calculated from its logarithm will be
the G.M. required.
Example. — To find the g.m. of
92, 100, 121, and 89;
look out their logarithms, add, and divide by four,
1-9638
2-0000
2-0828
1-9494
4 I 7-9960
1-9990 = log of 99-8
Therefore the g.m. required is 99-8.
The A.m. would have been 100*5.
As to the H.M. of more than two quantities I do not re-
member that it is often required : it can be got by taking the
arithmetic mean of the reciprocals of the given set of numbers
and then the reciprocal of that.
CHAPTER XXXVIL
Examples of the practical occurrence of Progressions
in Nature or Art.
In illustrating a subject by examples there is a great
advantage in selecting natural examples in place of artificial
ones. Natural examples may not be quite so easy as
artificial ones, but they are vastly better worth studying.
Artificial examples are often easy and handy for practice,
and many of them can be done in a short time, but a real
or naturally occurring example will take you into the essence
of the subject and is worth dwelling on long and steadily.
Every such example is more than an example : it opens a
chapter, and sometimes needs a treatise.
The chief instance of the natural occurrence of a geometrical
progression is in the theory of " leaks " — a leak of steam out
of a boiler, or of compressed air or water out of a reservoir,
of heat out of a cooling body, of electricity out of a charged
conductor; these and many other instances are all subject
to the same mathematical law — the law of a decreasing
geometric progression. See Chapters XL. to XLIL
A commercial example of the occurrence of G.P. is the
institution known as compound interest, when money in-
vested in some business undertaking is allowed to supply the
necessities and the supplementary accessories of effective
CHAP. XXXVII.] INTEREST. 353
human labour ; on the strength of which, under good manage-
ment, the sum invested increases in value at an ascertainable
or arbitrarily specified rate, and may accumulate until it
becomes a very large fortune.
Let us take this case first, for it may perhaps seem simpler
than an example from Physics.
Interest.
Suppose ,£1000, invested in machinery and wages, enables
a workman to produce fifty pounds worth of goods every
year more than need be expended on advertising the goods,
carrying them to their destination, feeding and clothing the
workman, patching up his shed and repairing the machinery ;
it is called capital, and is said to increase at the rate of
5 per cent, per annum, the increase being called interest.
If the fifty pounds is taken away and otherwise utilised,
so that the original capital of £1000 remains what it was,
without increase or diminution as the years go by, it is called
simple interest, and is an example of arithmetical progression.
But if the fifty pounds is invested in improved machinery
and in extra assistance, in such a way that it too brings in a
profit at the same rate ; and if this is steadily done each year
as interest accumulates, so that it is always added to the
capital, which thus goes on increasing ; it is called compound
interest, and is governed by the law of geometrical pro-
gression.
Every year the capital is increased in the ratio of yg^j ^^
every pound at the beginning of the year becomes a guinea
at the end of it. If this is supposed to go steadily on, what
will be the result after say 1 5 years 1
The result will be that the original capital has been
increased f^ths in one year, and this new capital has been
increased f Jths in the second year, or the original capital
354 EASY MATHEMATICS [chap.
f J X f J- in two years. In three years the original capital
will have increased in the ratio ( ^ ) > and in 1 5 years ( — j .
Now M + — J could be found by the binomial theorem if
we liked : and we know that as a first approximation it will
bel+i|.
Introducing the second order approximation
- n .n-1 „
l+nx + — j-^ — x%
we shall find its value to this order of approximation as
1.5 15xU 1 7x15
^20 2 400 "^ 400
- 1-3. 105 _
which equals approximately 2 ; that is to say the capital in
15 years will be by this operation a little more than doubled,
and will have become rather more than .£2000.
Now a quantity which is frequently doubled becomes, as
we know, extraordinarily large after the operation has been
repeated several times. If it doubles every 15 years, in 60
years it will have become £16000, and in a century and a half
it will have to be multiplied by 2^^, viz. 1024; that is to say
the capital will have swelled to more than a million.
A " penny " put out to 5 per cent, compound interest in the
time of Caesar would now theoretically far exceed all the
material wealth of the world in value.
But though an approximate calculation of compound interest
is instructive, there is no need to make it approximate, we can
calculate it exactly if we choose. We have only to raise the
ratio of the G.P., viz. 1*05 or 1^^^, to the 15th power in
order to find the value after 1 5 years.
xxxvii.] INTEREST. 355
This we should naturally do by logarithms,
log 1-05 = -0211893,
15 X log 1-05 = -3178365 = log of 2-078914.
So the original capital of .£1000 becomes increased to
£2078. 18s. 3d. in fifteen years, at five per cent, per annum
compound interest.
Suppose we wished to find at what rate of compound
interest a sum of money would exactly double itself in any
given time, say for instance in 12 years, we should have to
proceed thus:
Let X be the rate of interest,
then (l+icp has to equal 2,
wherefore 1 + « = ^^2.
We must use logarithms to find the twelfth root of 2.
log 2 = -301030,
TVlog2 = -025086 = log of 1-05945,
which equals l+x, wherefore x = -05945,
or the rate of interest must be 5-945, or nearly six, per cent.,
in order that doubling may occur every dozen years.
Let us see at what rate doubling will occur in thirty years.
^i_log2 = -0100343
= log of 1-0241,
or a little more than 2-4 per cent., about £2. 8s. 2d. added to
every hundred pounds per year.
Any rate of interest will double property if sufficient time
be allowed to it; but if we wanted to double capital every
six years, we should need a high rate of interest :
(l+xf = 2,
ilog2 = -050172 = log of 1-12247,
indicating about 12 J per cent.
Approximately therefore the doubling time and the rate
of interest vary inversely. If one is increased, the other can
356 EASY MATHEMATICS. [chap.
be decreased in roughly something like the same proportion,
especially when the rate is small.
. That is apparent at once if we expand by the binomial and
put it to double itself in n years :
{1+xy =* l+nx approximately = 2, say,
wherefore nx = 1 approximately,
or n and a; vary inversely as one another, to a first approxi-
mation. (See Chap. XXXIX.)
But then this first approximation is exactly simple interest,
it ignores the x^ and x^ and higher terms ; and it is just in
the presence of those terms that the virtue of compound
interest consists.
If X is added to each pound every year, but the interest is
not allowed to become part of the capital, so as to increase, the
amount becomes at the end of n years simply (l-{-nx) times its
original value. Thus at 5 per cent, simple interest, so that
X = -05, .£1000 in 1 year will become 1-05 times £1000,
or £1050. In three years it will be £1150, and in 20 years it
will be multiplied by 1 + (20 x -05) = 2 ; that is to say it will be
just exactly doubled in twenty years.
So whereas compound interest at 5 per cent, doubles an
amount in about 15 years, simple interest does the same
thing in 20 years ; and the interest could have been drawn and
otherwise utilised all the time.
At 10 per cent, simple interest a sum would be doubled in
ten years, and at 1 per cent, simple interest a sum would
be doubled in a century; whereas at 1 per cent, compound
interest, in a century, it would have increased in the pro-
portion (l-01)i<>o = 2-705, that is would have considerably
more than doubled, though it would not have trebled.
The advantage of compound interest over simple interest tells
more at high rates, for then the higher powers in the expansion
become important; and therein lies the difference It is in-
XXXVII.
INTEREST.
357
structive to plot the two things. Simple interest increase would
be represented by a straight line law, compound interest by an
exponential curve, the two starting off together, but ultimately
separating with greater and greater rapidity as time passes.
Corr
our
Sim
Irtereit
7 8 9
Time
Fig. 47.
10
12
13 14 15
One is an A. P., the other a G.P. ; one is a straight line law,
the other a compound interest law. One proceeds by constant
increment, the other by constant factor. (See pages 404, etc.)
As an example consider interest at the rate of ten per cent,
per annum, and tabulate the value of <£1000 after successive
years on the two systems.
Value at 10 % Interest.
Time.
Factor.
Simple.
Compound.
at start
£1000
£1000
1
1 year
1100
£1100
11
2 years
1200
£1210
(1-1)2=1 + -2+ •01 = 1-21
3 years
1300
£1331
(1-1)»=1-331
4 years
1400
£1464. 2s.
(1-1)4=1-4641
5 years
1500
£1610. 10s. 2id.
(l-l)5=l + -5+-10+-01+0005
6 years
1600
(1-1)6 [+-00001 = 1-61051
10 years
2000
£2593. 16s.
(1-1)10
15 years
2500
20 years
3000
50 years
6000
100 years
11,000
£13,781,000
(1.1)100
200 years
21,000
£189,910,000,000
(1-1)200
358 EASY MATHEMATICS. [chap, xxxvii.
The rate of interest plotted in fig. 47 is arbitrary, and
depends upon the original sum or 'principal'; this would
appear below the base line, and the diagram represents only
its growth. The curve on page 179 is really the same curve,
but to get the right aspect it must be looked at through the
paper ; and the scale of plotting is unsuitable.
The right hand column of the above tabulated numbers
contains the factors by which the original sum must be
multiplied to give the amount at compound interest. It is
hoped that the binomial coefficients will be recognised.
It may be noticed that during the length of an active life-
time the difference between simple and compound interest is
not extravagant, even at so high a rate as ten per cent ; but
that if a sum is locked up during a long minority, or if
otherwise the interest be left to accumulate for a long time
without being contemporaneously expended, the growth by
compound interest becomes enormous.
The operation of a " sinking fund " for the annihilation of
great debts can thus be illustrated. But it is not to be
supposed that interest accumulates, without effort, auto-
matically. It is the result of human skill, brains, labour,
and management.
PART 11.
MISCELLANEOUS APPLICATIONS AND
INTEODUCTIONS.
CHAPTER XXXVIII.
Illustrations of important principles by means of
expansion by heat.
Everyone has seen a telegraph wire by the side of a
railway and observed the peculiar effect of its sag, as the
train passes along, when such a wire is near the window; it
seems to be moving up and down. A wire or rope or chain
stretched between two posts cannot be perfectly straight, but
sags, something like the top of a lawn tennis net.
In hot weather the sag of a given span of wire is greater,
in cold weather it is less; because the material expands by
heat and contracts by cold. Suppose the length of a wire on
a certain span is I during a night of light frost, then by noon,
when the sun has been up some time, it will have increased to
V : the increase of length being l' - 1.
This notation for the same kind of quantity under different
circumstances, by means of the same letter with a dash affixed
to it, is in constant use, and must be grown accustomed to ;
the new length should be read " Z-dash " ; the increase should
be treated as a single quantity and should be read "/-dash
minus I." And here it is desirable to remark that all mathe-
matics is intended to be read, and that it is good and necessary
practice to read it.
Algebra is a language — a very expressive language ; and
although it appeals primarily to the eye, it should be made to
362 EASY MATHEMATICS. [chap.
appeal to both eye and ear, that is it should always be " read,"
if only to oneself. It is a great mistake to treat it as a silent
language and only to look at it, beginners must learn to read
it ; and it would be well now to turn back and read aloud all
the equations and other algebraical expressions we have
employed so far. It cannot be done properly without a good
deal of practice. Everything written on a blackboard by a
teacher should be spoken also.
Now consider what the increase in length depends on. In
the first place it depends on the original length of the wire.
No one would expect a wire a few inches long to elongate as
■much as one a few hundred yards long. It is only common
sense to realise that every yard of the wire will elongate an
equal amount under the same conditions, and that therefore
20 yards will lengthen 20 times as much as one yard ; so I' -I
is proportional to /. Next it will depend on the change of
temperature, which we may treat as a single quantity and call
T' - Tf where T was the original temperature and T' the new
temperature.
We have no guarantee that the lengthening is proportional
to the rise of temperature, but it is a natural assumption to
begin with, and will have to be corrected if necessary later.
We can assume that it expands twice as much for two degrees'
rise as it does for one degree, and ten times as much for ten
degrees' rise. There are only a few substances for which this
is really and precisely true, but, for all, it is a rough approxi-
mation to the truth, and will do for the present.
Lastly, the lengthening will depend on the material of which
the wire is composed. If it were a copper wire it would be
found to expand more than if it were of iron Every material
has its own "expansibility," by which is meant the rate of
expansion, or increase per unit length per degree rise of
temperature.
xxxviii.] EXPANSION BY HEAT. 363
If we denoted the " expansibility " of the material by k, we
should be able to express in one line all we have so far said,
thus: I' -I = U{r-T) (1)
for this asserts that the lengthening is dependent on and
proportional to three factors, viz. (i) a constant representing
the properties of the material, (ii) the length of the piece
of that material which is under consideration, and (iii) the
rise of temperature or the warming to which it has been
subjected.
Of the three factors, k may be styled a " constant " to be
determined by experiment in the laboratory, a thing depending
on the properties of a material, and beyond our control, except
in so far as we can select the material ; Z is an arbitrary con-
stant entirely in our control, depending only on what we
choose to attend to. We might observe the lengthening of
the whole of a span of wire, or of any portion of it, or we might
select spans of different length, or we might cut off a bit and
attend only to that. T' -T represents the change or variation
of a variable quantity, in this case temperature; it is some-
times called the independent variable, for its changes go on
independently of any of the other things we have considered,
and the change of length is dependent on it. One could hardly
make the converse statement and say that change of tempera-
ture was caused by change of length, even though the length
was that of a thermometer column (though it might be rash
to stigmatise even this statement as absurd under all circum-
stances : there are things which get warm when and by reason
of being stretched), but it is extremely natural to say that
the change of length is caused by change of temperature;
so the cause is called the independent, and the effect the
dependent, variable. By some people the names " principal "
and " subordinate " variables are preferred to " independent "
and "dependent."
364 EASY MATHEMATICS. [chap.
A change of temperature might be caused out of doors by
the appearance and disappearance of the sun, or by a change
of wind ; in a laboratory it could be caused by the application
and removal of flame, or of an electric current, or of some
other means of heating. Anyway it is to be observed by
a thermometer of some kind, and T -T may be considered
as being measured by the rise of the thermometer.
/' — Zis the change of the dependent or subordinate variable ;
and its dependence might be conveniently indicated by putting
the two variables on one side of the above equation, and the
constants on the other, as for instance :
ji' _ ji — "'^ (2)
To emphasise the fact that the two terms T' -T represent a
thing which is really one quantity viz. a warming, a difference
of temperature, it is convenient to have a single symbol for it;
and the symbol usually chosen is an abbreviation for difference
of temperature, namely 87' or dT ; meaning
diff. of temp. = T'-T,
or diff. T = T- T, or simply dT = T - T.
This mode of expression is very handy and extraordinarily
convenient. It can be applied of course to all kinds of
quantities, so V -I may be written " diff. length," or dl ;
wherefore our two above numbered forms of a proportionality
statement become abbreviated into
dl = kldT (1)
and ^ = kl (2)
respectively.
These equations are labelled like the previous ones because
they say precisely the same thing as the others did, and in
the same way. It is generally understood that the symbol d
XXXVIII.] EXPANSION BY HEAT. 365
is used for a difference only when it is an infinitesimal
difference. For finite differences delta I is used, or simply V - I.
Form (1) gives explicitly the change of length in terms of the
original length and the change of temperature ; form (2) gives
the ratio of the changes of the two variables in terms of the
constants — viz. the expansion-property of the material, and
the length selected for observation.
Another way of writing the equation is often useful, in
which the expansion per unit length is explicitly attended to :
that is the lengthening by heat of any one yard or foot or
metre of the wire, without regarding the whole wire. To get
this we have only to divide the length out, and so get
j = kdT, (3)
a quantity which is often technically referred to as " the ex-
pansion " ; it is defined as the ratio dl/l, and it is equal to the
expansibility multiplied by the rise of temperature.
Let no beginner suppose that these various forms of the
equation are different. They are all essentially the same, but
they emphasise features differently; just as in any language
a sentence may be recast so as to say the same thing with
various emphases. Never forget to regard algebra as a
language, in which statements of singular definiteness and
precision can be compactly made.
Now suppose the temperature fell instead of rose : the ex-
pression T' -T would be negative, and we might sometimes
choose to denote the fall of the temperature hjT- T. At
the same time most substances contract with cold, and so
V -I would also be negative, and the contraction could be
written l-V or -dl; but usually dl and dT would not have
the negative sign actually prefixed to them, it would be
sufficient to say that dT and usually dl are both negative, for
the case of a fall of temperature.
366 EASY MATHEMATICS. [chap.
Now let us begin again, and look at the matter afresh and
in a still simpler manner. Take a rod of length 1, that is to
say 1 foot or one metre or one inch in length, at a temperature
7*, and warm it one degree. Its length will now be increased
by an amount we will call ^, meaning Tc feet or metres or
inches, according to our choice of unit length.
Warm it 2 degrees and its increase of length will be 2^,
and so on, as shown thus :
its length being 1 at temperature T,
its length is 1+^ „ „ T+1
i> >j 1 + 2a; „ „ T+2f
l + Sk „ „ r+3,
l+nk „ „ T + n.
This temperature T + n we will call T' so that n = T'-T.
If the original rod had been originally of length I instead of
length 1 and had been all of it treated alike, every unit would
have expanded by the same amount, so the final length would
in that case be /(I +nk), which we may. call l.
Hence I' = /(1+n^)
or V -I = Ink
= kl{r-T\
thus arriving at the same result as before, which we will now
write in any one of four equivalent ways, e.g.
dl = kUT, (1)
S=*' (^)
f = WT, (3)
i ^-ifc (4)
XXXVIII.] EXPANSION BY HEAT. 367
The last may be taken as a definition of the expansibility h,
and shows the principle of what we must do in the laboratory
in order to measure it.
We must take a rod or wire or something convenient of
the given material, and measure either its whole length /,
or a length I between two marks or scratches on it ; then we
must subject it to a measured rise of temperature, and observe
the increase in length of the chosen portion carefully, with
a microscope or micrometer by preference.
Then it is best to cool it down again and see that the rod
recovers its original length ; and then the warming can be
repeated, and the increase in length observed again, and so
on several times, to avoid accidental errors and to get the
true reading as nearly as we can.
Thus the three required quantities dl, dTy and /, are all
measured; and, dividing dl by / and by dT^ we get the
expansibility of the material as a result.
These are not laboratory instructions, and accordingly little
or nothing shall here be said about the practical mode of over-
coming difficulties. Suffice it to say that the readiest mode of
securing measurable differences of temperature, is by making
use of properties of substances designated by such phrases as
boiling oil, molten lead, melting ice, boiling water, condensing
steam, and the like ; and that the chief precautions needed, in
order to measure with precision the expanded length, are
those which shall guard the measuring scale, or standard of
length, from being likewise affected by the high temperature
of the rod to which it has in some sort to be applied.
With this hint the somewhat elaborate arrangements de-
picted in text books of Physics can be appreciated.
This matter has been gone into at some length, because
it is typical : it is always worth while to master a type, and
nothing is gained by haste.
368 EASY MATHEMATICS. [chap.
Examples.
A bar of iron 10 yards long expands '444 inch when taken
out of ice and put into steam or boiling water ; what is its
expansibility^ i.e. what is the increase per unit length per
degree for iron; meaning by a degree Centigrade the hun-
dredth part of the interval between freezing and boiling
water, and by a degree Fahrenheit the 180th part of that
same interval.
•444
Answer. -— — — - = -0000123 per degree Centigrade,
or -00000683 per degree Fahrenheit.
The numerical result is worth remembering as specified in
the Centigrade scale of thermometry, which is the most used
for scientific purposes. Observe that there are 4 ciphers
before the significant figures, which happen to be the first
three natural numbers, and so are quite easy to remember ; the
amount is about IJ in a hundred thousand, or about 12 parts
in a million : meaning that iron expands this fraction of its
length for each Centigrade degree rise.
Brass would give a number about 18 instead of 12; and
zinc, which is one of the most expansible metals, would expand
nearly 25 parts, or just double as much as iron. Platinum how-
ever, and glass, would have been found to expand only about
8 or 9 parts in a million, per unit length per degree Centigrade.
If a material expanded 1 per cent, of its length for a rise
of 100°, its expansibility would be -0001. If it expanded J
per cent, for a rise of 250°, its expansibility would be -00002,
which lies between that of brass and that of zinc.
Cubical Expansion.
It would be a mistake to suppose that a rod increases in
only length when heated : it swells in every direction, just
XXXVIII.] EXPANSION BY HEAT. 369
as if it were slightly magnified. Its increase in length is
most noticeable, because that was originally its greatest
dimension, but its increase in thickness is proportionately as
much. Thus if a bar were a yard long and an inch thick it
would expand in length 36 times its increase in thickness,
but its proportional expansion, its y, would be the same in
every direction.
Consider an iron plate 10 metres long, 1 metre broad, and
1 millimetre thick, and let it be warmed 406 degrees. Its
linear expansion is
406 X -0000123 = -00500,
or five parts in a thousand (or the half of 1 per cent.).
So its increase in length is 5 centimetres ;
its increase in breadth is 5 millimetres ;
its increase in thickness is -005 millimetre,
or 5 millionths of a metre,
or 5 mikrons,
a mikron (sometimes spelt micron) being a convenient unit
for microscopic work, and being sometimes mconveniently
denoted in biological books by the symbol /x. Units or
standards should be expressed in words; symbols are never
used for them by mathematicians.
What is the increase in area and in bulk of such a plate
when so heated 1
The first thing to learn is that we must not take the
increases and multiply them together. (Cf. p. 293.)
The increase in area is not 5 centimetres x 5 millimetres.
The increase in volume is not 5 centimetres x 5 millimetres
X 5 mikrons.
But the new area is 1005 x 100-5 sq. centimetres, whence the
increase in area is 1002-5 sq. centimetres.
^.B.M. 2 A
370 EASY MATHEMATICS. [chap.
The new volume is 1005 x 100*5 x -1005 cubic centimetres
= 103 X 102 X 10-1 X (1-005)3
= 104x1-015075
= 10150-75 c.c.
The old volume was lO^xlO^xlO"! = 10* c.c, so the
increase in volume is 150'75 c.c.
But, as usual, there is a quicker and better mode of making
the numerical calculation, by first treating it algebraically.
Let V = l{\ +kt) be the new length,
b' = b(l+kt) the new breadth,
and z' = z{l+ U) the new thickness,
where t stands for the rise of temperature T' - T.
Then the new volume is
I'h'z' = lhz{l+ktf;
that is, calling the old volume V and the new volume V\
V = V{\-\-Uf
= r(i + 3^^+3F^2+^/3).
but, since kt is a small quantity, this is, to a first approxi-
mation, V' = r(l + Zkt\
or F'-F= 3kFt = 3kF(T' - T),
or -j^ = 3kFy or -pr = 3kt;
a result which can be expressed by saying that the cubical
expansibility, viz. 3^, is three times the linear expansibility.
Similarly the coefficient of superficial expansion is twice
the linear.
This is equivalent to neglecting squares and cubes of small
quantities ; and for most purposes that can safely be done
in cases of solid expansion. Hence to do the above sum,
very approximately, all that is necessary, after observing that
xxxviiT.] EXPANSION BY HEAT. 371
the linear proportional increase is '005, is to say : — The
original area = 10 square metres, so the increase of area is
2 X "005 X 10 square metres
= 0-1 sq. metre ^ 1000 sq. centimetres approximately.
The original volume is
10 metres x 1 metre x '001 metre = -01 cubic metre
= 10,000 c.c.
so the increase of volume is 3 x -005 x 10,000 c.c.
= 150 c.c. approximately
CHAPTER XXXIX.
Further Illustrations of Proportionality or Variation.
One of the most important things to understand, in order
to be able to apply elementary mathematics to simple
engineering facts, is the law of simple proportion. Two
quantities are said to be proportional, or to vary as each other,
if they are both doubled when one is doubled and if they
vanish together.
Thus for instance the stretch of an elastic and the force of its
pull are proportional. For first of all they vanish together : if
the elastic is not pulled it is not stretched. Secondly, if it has
been stretched with a certain pull, and you double the pull,
the stretch also will be found to be doubled. You can try this
by hanging up an elastic or a spiral spring and loading it with
different weights. As the weight increases, the stretch
increases, and in a simply-proportional manner. This is
the principle of a spring balance.
Take all the load off, and the pointer returns to zero,
indicating no stretch. Observe, it is not the length of the
elastic that is proportional, for that does not become zero, nor
is it doubled when the load is doubled, but it is the stretch or
increase of length that is proportional to the load. Suppose
however there had been some irremovable load on all the time,
as indeed there is, for the spring or elastic itself has a trifle of
weight of its own, how do we allow for that % Answer : by
CHAP. XXXIX.] VARIATION. 373
always attending to the variable part of the load only, just as
we attend to the variable part of the length only. The " load "
must really signify the load added or subtracted ; and the
stretch corresponding thereto signifies the increase, or it may
be the decrease, of length which accompanies that variation of
load ; so that instead of saying the length I varies as w the
load, which is not true, we ought to say, change of length
varies as change of load, or dl varies as dw^ which is quite
true ; unless indeed the spring is overloaded and permanently
strained or injured so that it cannot recover; or, in other
words, unless it is not perfectly elastic. So long as it is
perfectly elastic, the law of simple proportion holds ; and the
test of whether it is perfectly elastic or not is to see if it can
completely recover when the load is removed.
Some substances stand a great deal of loading, such as steel ;
some stand only a little without giving way, like glass or
copper ; and some stand hardly any, or none, like lead or
straw or dough.
There are two methods of giving way, one by breaking, like
glass, the other by permanently bending, like lead. There is
plenty to be learnt about all these things, but the time for
learning them will come later. All that we have to note at
present is that the law of proportion is not to be expected to be
verified when the substance experiences a permanent set, or
deformation of any kind, from which it cannot recover ; and of
course not when it is broken.
The law holds "within the limits of elasticity," and it is
known as "Hooke's law," because that great and ingenious
man Robert Hooke experimented on it and emphasised its
simplicity and convenience more than two centuries ago.
You may think that it is so simple as not to be much of a law
of nature; but you will find that all the most fundamental laws
are simple. Simplicity and importance may quite well go
374
EASY MATHEMATICS.
[chap.
together, though there is perhaps no necessary connexion
between them.
Now take another example of simple proportion. Let a
balloon ascend at a perfectly steady pace from the ground.
Its height is proportional to the time which has elapsed since
it started. In one second it went up let us say a yard. In a
minute it will go up 60 yards, and in an hour 3600 yards,
provided the same upward speed was all the time maintained.
If it went sometimes faster and sometimes slower, the simple
proportionality would not hold. The height and the time
vanish together, for we began to reckon time at the instant it
was let go, and we were careful to measure height always to
that same point of the balloon which touched the earth at the
moment of letting go.
But now take an example where simple proportion does not
hold between two connected variable quantities.
20 Years
Fia. 48.
The height of a child depends upon his age, and increases
with his age, but it is not proportional to it : in other words,
his height does not vary with his age, in the technical sense.
XXXIX.]
STRAIGHT LINE LAW.
375
For first of all they do not vanish together : the child had some
length when he was born ; and next, they are not doubled
together. A child has not twice the height at 6 that he
had at 3.
If we were to plot age and height together, it would be
instructive, and the result might be something like fig. 48,
where age is plotted horizontally and height vertically.
The point 0 is called the origin, and represents the epoch
of birth. If the curve passed through this point 0, zero
initial length would be signified ; so the curve does not pass
through it, but starts above it at the infant's length at birth :
say fourteen inches
Such a curve is not simple proportion at all. It is easy
enough to understand, but the law represented by the curve is
not a simple one.
Simple proportion would be represented, on the same plan,
by a straight line through the origin ; as for instance if the
stretch and the load of an elastic thread or spiral spring were
plotted : they vanish together. (Fig. 49.)
But suppose the two quantities did not vanish together, we
might still have them plotted as a straight line. For instance.
376
EASY MATHEMATICS.
[chap.
the length of a rod at different temperatures (or the total
length of a piece of elastic under different loads) :
Temperature
Fig. 60.
and this, though it is not exactly simple proportion, is the
next thing to it, and is sometimes called "a straight line
law." (Fig. 50.)
It can be made simple proportion by deducting a constant,
by deducting the original length for instance, I' -I = kit,
whereas if the length I had not been deducted it would have
been expressed by I' = 1 + Ikt,
and this is characteristic of a
straight line law.
In general a straight line law is
represented by
y = a + hx,
whereas simple proportion would be
y = bx
without the constant a.
So by subtracting the constant a
a straight line law can always be
expressed as simple proportion. So it can if we attend to
changes only. For let y become y', and x become x, while a
and h remain constant all the time, we should have
Fio. 61.
XXXIX.] PARABOLIC LAW. 377
y = a + hx,
y'= a + hx\
y'-y = b{x'-x),
or dy = hdXf
and the a has disappeared. The two differerices, dy and dxy are
simply proportional; for they increase together in a constant
ratio, and they vanish together : one cannot become zero
without the other.
The weight of a boy is not represented by a straight line,
even if his weight at birth is deducted. His law of growth is
not a straight line law but a more complicated law : it could
be plotted as a curve, from successive observations.
Any law can be expressed by a curve ; thus we might have
a parabolic law, meaning that the curve of plotting is a para-
bola as nearly as can be told.
A parabolic law is expressible in algebra thus,
y = a-hbx + cx'^j
and it would be instructive for children to plot the curve
represented by this equation, and see what it looks like. To
carry out the plotting we must be told the values which are
to be attributed to a, b, and c, and can arrange the scale to suit.
Thus let a = 4, b = 1, and c = J.
Make a table of corresponding successive values of x and y.
When X = 0,
y = a, or in
this case 4 ;
a:= 1,
y = a + b + c, or in
this case 5 J ;
x = 2,
y = a + 2b + 4c,
or 7;
x = S,
y = a + 3b + %
or 91;
a: = 4,
y = a + 46+16r.
or 12;
x = 5,
y = a + 5b + 25r,
or 15J;
x = 6,
y = a + 6b + 3Qc,
or 19;
etc.
378 EASY MATHEMATICS.
Plot this, and it looks somewhat thus :
[chap.
15-
10-
3 4
Fig. 52.
But we need not necessarily limit ourselves to the positive
side of the vertical axis ; we might ascertain and plot values of
y when x is negative, otherwise the curve is incomplete. You
could hardly tell it was a parabola from its appearance so far.
When X =
-1,
y = a-b + c =
3^;
X =
_ 9
-'5
y = a-2b + 4:C =
3;
X =
-3,
y == a-Sb + 9c =
3i;
X =
-4,
y = a - 4b + 16c =
4;
X =
-5,
y = a-bb + 25c =
5i;
X =
-6,
2/ = a- 66 + 36c =
7;
X =
-7,
y =
9i;
X =
-8,
y =
12;
X =
-9,
y =
15i;
X =
-10,
y =
19; etc.,
XXXIX.]
PLOTTING A CURVE.
379
every value on this side being equal to a certain value for
another x on the other side. The whole curve is symmetrical
about a vertical axis through x = -2.
The dotted line is called the axis of the parabola. (Fig. 53.)
Another example of what may be a straight-line-law is the
slope or gradient of a railway. At a certain place it may be
said to rise 1 in 30, for instance ; meaning that if you go
30 feet along the railway you have ascended 1 vertical foot j if
you go 30 yards you have ascended 1 vertical yard, and so on.
A uniform gradient is naturally plotted by a straight line,
and if the vertical height is called y while the horizontal
distance is called a:, the gradient is approximately denoted by
— . Not exactly, because the denominator is usually measured
along the sloping railway, and not horizontally.
380
EASY MATHEMATICS.
[chap.
slope are -^ and
dx
Either way of measuring the gradient is a good method;
and sometimes one is used, sometimes the other. If the
slant distance is called ds^ the two chief ways of measuring
-, respectively.
Except when the slope is steep, the difference between these
two methods of measuring it
is not marked : — the connexion
between the two methods is
easily shown by a diagram.
A considerable but feasible
slope for an ordinary railway
would be a gradient of 1 in 30, that is to say -^ = — ,
Often it is not more than 1 in 100. The actual gradient
is frequently written up on low posts by the side of the
line.
But take the case, impossible for a practical
railway, where the slope is 45°, which would
be a steep mountain side, dy and dx are in
that case equal, and ds = ^2 either of them,
dy _ , dy 1
so j^ = 1, whereas -5- = —p..
dx ds J'2
Take the case of a precipice or a steeple, almost
vertical, so that dx is extremely small.
dy and ds are now nearly equal and dx is nearly
0 ; in the limit quite zero.
So approximately, for an angle nearly 90°,
dy _ ,
and zr- = 7^
dx 0
or at any rate a very great number; in the limit quite
infinite.
XXXIX.]
GRADIENT.
381
Take the case of that famous right-angled triangle with
commensurable sides, and express the slope of its hypothenuse :
^ _ 3 dy _Z
dx ~ 4' ^s ~ 5'
5X
3 .
4 dx
Fio. 57. Fia. 58.
irse always d&^ = d^ + dy'^^
dy_
dy dy dx
ds J(dx^ + dy^) ji+f^l)'
so that
It is often convenient to measure slope in yet another way,
viz. by the angle of slope, which we denote by a or ^ ; then the
ratio of the height to the slant length, -j-^ is called the sine of
the angle, and is written sine 0 or sin ^, while the ratio of the
height to the base, -^, is called the tangent of 0, and is written
tan Q, So we see from the above that an angle whose sine
is I, or -6, has a tangent whose value is f or -75 ;
also that the sine of 45' is -p,
while the tangent of 45" is 1,
and that sin 90° = 1, while tan 90° = oo. (Cf. fig. 56.)
On the other hand, when the slope is very slight, the
sine and tangent are about equal, and it does not much
matter which measure of the slope we employ. (Cf. fig. 54.)
382
EASY MATHEMATICS.
[chap.
Ultimately when the slope vanishes they also vanish, and
vanish on terms of equality, so that sin 0 = tan 0 = 0.
Here is a case where two things vanish together but are by
no means proportional ; they are approximately proportional,
or indeed equal, for small angles, but the tangent increases
faster than the sine ; and as the angle grows, it increases very
much faster ; so that, by the time the sine has reached unity,
the tangent has gone, with a rush, to infinity.
Plotting them they would look thus :
where one curve represents the sines,
and the other the tangents, of angles
from 0° to 90°
Fia. 59.
It may be worth showing, even at this stage, what suggested
these curious names. The name " tangent " especially sounds
XXXIX.]
MEASURES OF ANGLE.
383
curious when applied to a ratio. The idea arose from drawing
a circle round an angle and seeing all the different ways in
which it might be measured.
In this figure the angle is at C, and
AD is a, bit of a circle with centre C.
This figure long ago suggested a bow
and arrow, hence EB was called the
sagitta, and ABD was called the arc : the
string A ED is called the chord, and half
of it under certain restrictions is called
the sine, presumably because the point E
of the string is pulled to the breast before
releasing the bow. The tangent can then
be measured as part of a line drawn through B, touching the
circle, when the circle is drawn of unit radius.
Let CA or CB equal 1 on any arbitrary scale; then C or
ACB is the angle, a,
AE is its sine,
AB is its arc,
EB is its sagitta,
and BF is its tangent,
always provided CA or OB are equal to
1, that is taking the radius of the circle
as unity.
The size of the circle is quite arbitrary :
any length whatever may be chosen as a
unit of measurement.
But it is desirable to bear in mind
that angles are not measured by length
but by ratio; and accordingly the state-
ment that
AB
Fio. 61.
angh
arc
radius AC'
384 EASY MATHEMATICS. [chap.
or intercepted part of arc -f radius, is a better statement than
to say that i , i- i
'' angle = arc when radius = 1.
And to say that
, intercepted perpendicular AE
sinanffle = ~ ^^—- = —r-y
° radius A C
is better than saying that it equals AE when AC equals 1.
So also
f i p - intercepted portion of tangent _ FB
° ~ radius ~ BC'
It may be worth while also to state here that the length of
that boundary line of the angle which cuts the circle and is
produced to meet the tangent, viz. CF, is called the " secant "
when the circle is of unit radius ; or in general, dividing by
the radius, ^r»
secant of angle (7, or sec C, = y^.
These different fractions or ratios are all measures of the
angle : they are quite independent of the size of the circle
or of any linear dimension whatever. They indicate angular
magnitude alone.
In any right-angled triangle, if the length of the hypothenuse
is called r, and the angle at the base be called d, then the
length of the base is, by definition of cosine, equal to r cos 6 ;
that is to say, it is the length r multiplied by a proper fraction,
which fraction is called the cosine of the angle 0.
The base may be thought of as the projection of r on to a
direction inclined at the angle ^ to it ; the shadow as it were
XXXIX.]
MEASURES OF ANGLE.
385
of a slant rod of length r thrown by vertical rays of light on a
horizontal ground (Hg. 64).
If the projection were made in a direction at right angles to
the first, then the projection is r sin 6^ provided 6 still means
the same angle as before. So that the sides in any right angled
^
Fio. 65.
triangle are related as in the diagram (fig. 62); and it is
obvious, by definition of the tangent ratio, that
sin <9 ^
7, = tan u.
cos t)
It also follows, by Euc. I. 47, that
(r sin (9)2 + (r cos (9)2 = r\
or that (sin Bf + (cos df= 1,
a fundamental identity.
One peculiarity of angular magnitude is that it is un-
magnifiable. Look at an angle through a magnifying glass,
and though its sides lengthen, the angle continues constant.
A right-angle, for instance the corner of a book, continues a
right-angle, and 45° continues 45°. A degree is always the
360th part of a circle, however big the circle ; a quarter of
a revolution, or a right angle, is always 90° ; and so on. The
size of the divisions of a circle change when it is magnified,
but their number remains constant.
Number is another thing which is unmagnifiable. Magnify
3 oranges, the oranges look bigger, but they still look 3.
So when a plate expands with heat, if it is uniform, any
L.E.M. 2 b
386 EASY MATHEMATICS. [chap.
angles it may have remain constant. If there were a hole
in the plate, the hole would expand just as if it had been
filled with solid ; its boundary line might have been drawn
with ink on a similar solid plate. Everything expands
with heat as if it were looked at through a weak mag-
nifying glass. So a hollow space is not encroached upon
by expanding walls, but is enlarged as if it were full of
substance. A hollow bulb, for instance, has a greater capacity
when heated than it had before. It may not necessarily hold
more fluid, not more weight or mass of fluid, for the fluid
might expand still faster than the solid, but it holds a greater
volume. A thermometer bulb containing mercury is in this
predicament ; and what we observe, when the thermometer
rises, is the apparent expansion of the whole bulb-full of
mercury swelling in the only direction open to it, viz. in the
narrow stem. It is called "apparent" expansion because it
represents what is visible, viz. the excess of the expansion
of the mercury over that of the vessel which contains it.
If the vessel expanded more than the liquid, the rise in the
stem, indicating the apparent expansion, would be negative ;
but this is hardly a possible case in practice, for all liquids
expand more than any solid Nevertheless the true or
absolute expansion of a liquid is always greater than it
appears to be, unless we could observe it in a vessel which
did not expand with heat.
Let V be the volume of the vessel, and a the expansibility
of its material, that is to say its increase in bulk per unit
original bulk per degree rise of temperature, then the total
swelling for any rise of temperature I" - T is
v' -V == av{T' -T\
or dv = avdT,
or — = adT;
xxxTX.] VARIATION. 387
or, in words, the proportional expansion equals the change
of temperature multiplied by the expansibility. If we assume
simple proportionality between change of volume and change
of temperature, it is the same thing as assuming that a is a
constant ; and in that case the expansion is said to vary with
the temperature.
The term " varies with " or " varies as " is a technical term,
and is understood to mean "is proportional to." The latter
is really the better expression, for in common language two
things can vary or change together without being pro-
portional, like the age of a child and its weight, or the
amount of sunshine and the cheapness of corn, or the height
of a look-out man on board ship and the distance of his
horizon, or the amount of oil consumed in a lamp and the
brightness of the consequent light.
But the term "varies as" or "varies with" is understood
to indicate more than merely changing together: it means
that they vary in a simply proportional manner, so that
if one is doubled or trebled or increased 1 per cent.,
the other is doubled or trebled or increased 1 per cent,
too.
When y varies as x, in the technical sense, it can be written
y ^ x; where oc is a mere symbol, not much used, to denote
" varies as."
Or it may be written y :'.x^ and read y is proportional to x ;
or it may be written, the ratio «/ : a; is constant ; or y -=- « or
^ is constant, equal to h say ; or y = hx.
This last is the simplest and most satisfactory mode of
stating simple proportionality, h being understood to be a
constant, that is something not at all dependent on the value
of X and y, which are the variables.
A straight line law is slightly more general than this :
388 EASY MATHEMATICS. [chap.
it includes the main idea of proportionality, but it exempts
from the necessity of vanishing together ;
y = a + hx
is the typical straight line law.
Either can be expressed as
dy = bdx,
for when we attend to variations, the constant term a has
no influence, and so disappears. The constant factor b, which
is only part of a term, by no means disappears ; b represents
the rate of increase of y with respect to x; for instance it
represents the slope of the line, being the change of elevation
per unit step along the base; it is in fact the 'tangent' of
the slope.
The next more general law is the parabolic law.
y = a + bx + cx^j (68 i.)
or it might be y = a-bx + cx^, (66.)
or y ^ a + bx- cx\ (68 ii.)
or y = a-bx- cx^, (68 iii.)
all of them parabolic, with different appearances.
The slope of such a curve is of course not constant.
To find an expression for the slope, we must take two
points near together and compare the vertical with the
horizontal step, that is find the dy corresponding to a given
small dx. To do this we let x change to x' and y to y', and
then write the relation once again for the changed values
y' = a + bx' + cx'^,
and now subtract the old value from the new, so as to get the
difference, y -y = {a-a) + b{x' -x) + c{x'^ - x^)
= {b + c(x' + x)}(x' -x);
. diff. y 1 / , .
• . -..rn = b + C(X +X).
diff. X ^ '
XXXIX.]
SLOPE.
389
Now if the step is made quite small, the x and ic' are
extremely nearly equal, so that it matters little whether we
Fio. 66.
write y^ ■\-x or 2ic or 2a;' ; and in the limit when the step is
infinitesimal they become actually equal, and then
dy
dx
= J + '2cx,
It is not
and this is the gradient of the curve at any point,
constant, but it follows a
straight line law.
The rate of change of
the angle of slope may
be called the curvature.
When the slope is constant,
as in a straight line, the
curvature is nothing, but
when the slope changes, as in the last case, the curvature can
be measured as the rate of change of the angle of slope per
unit step along the curve. Suppose the gradient or angle of
Fig. 67.
390 EASY MATHEMATICS. [chap.
slope at any point is denoted in angular measure by <^, then the
curvature could be defined as -^, and that is its usual measure.
as
Another method of measuring it, which seems simpler but
is not so satisfactory, is to indicate the slope by the tangent of
the angle <^, that is by the gradient -j-, which we may denote
by a single letter g, and then to denote the curvature by the
rate of change of gradient per horizontal step, that is by -t-;
On this plan the value of the gradient for the above
parabola is g = b + 2cx,
and the rate of change of gradient is
dx - ^'^
because if x changes to x'j g changes to ^' = J + lex', and when
you come to reckon the difference ratio, the 6's go out :
g' -g ^ 2ca;^ - 2ct ^ ^^
X — X X X
It is not customary or necessary thus to introduce a new
d^-y
symbol g -, it is neater to express -^ as — — , that is to say as
70 O/X OjX
y|; and this it is which is equal to 2 c.
CLX
So, in the original parabolic expression
y = a + hx + cx'^,
the meaning of the constant c is half the rate of change of
gradient, which is a principal term in the curvature; the
meaning of the constant h is the slope, or gradient itself,
especially the slope at the place where x is 0, that is to say
where it cuts the axis of y ; and the meaning of a is the height
at which the curve cuts the axis of y, that is to say the inter-
cept on that axis. Compare the equations on page 388 with
the curves drawn.
XXXIX.]
ALGEBRA AND GEOMETRY.
391
392
EASY MATHEMATICS.
[chap.
Observe, with regard to slope, that when the curve slopes
upward where it cuts the vertical axis, as in fig. 68 i. or ii., the
coefficient b, which measures the slope there, is positive; but
when it slopes downward as in fig. 68 iii. the term involving b is
negative. If the curve cut the vertical axis without slope, or
horizontally, the term involving b would disappear, and in
such a case the parabola would be represented by y = ai- cx^
(fig. 69).
Pio. 69,
If the curve cut off no length on the axis of y, that would
be indicated by the non-existence of the constant a, so that
that parabola would be written y = cx^ (fig. 70),
Fio. 70.
and this is the simplest expression for a parabola possible.
XXXIX.] ALGEBRA AND GEOMETRY. 393
Observe that it is the same curve all the time ; it is only
shifted with respect to the axes by the different values which
can be attributed to the constants.
c is the curvature term, and when c is positive it curves
upwards, like 68 i. ; when c is negative it curves downwards,
like 68 ii. or iii.
y = -cx^ would be like this (fig. 71) :
Fio. 71.
If wo want the parabola to look like this (fig. 72) :
FiQ. 72.
we have only to turn it through a right-angle, that is to say,
interchange the axes of x and y, and write
X = cy\
394 EASY MATHEMATICS.
or if we wrote x = -cy'^ it would look like this
[chap.
Fig. 73.
If we wrote x = a -{-cy^ it would become like this
Fio. 74.
where the x intercept is a.
ALGEBRA AND GEOMETRY.
395
If we wrote ic = a + &y + c/ it would slope up where it cuts
the axis.
Fig. 76.
If we had x = a-hy + cff^ it would slope down at the foot,
like this :
Fig. 76.
396 EASY MATHEMATICS. [chap.
li X = -a + cy^ the curve would cut the x axis on the left :
Pig. 77.
If we introduce a term containing x^ as well as a term
containing y^, and if necessary a term xy, we can tilt the
parabola in any desired direction and place it anywhere in the
plane : though in these cases there is a risk that it may cease
to be a parabola; and if we introduced a term x^ or y^ its
parabolic character is bound to be spoilt ; just as the
introduction of either x^ or y^ destroyed the straightness of the
line y = a + bx.
The next step towards algebraic complexity is what is called
the cubic parabola y = a + bx + cx'^ + dx^,
where again the coefficients or constants, a, J, c, d, may be any
of them positive or any of them negative.
The beginner can plot this and see what it looks like, —
proceeding after the same fashion as before ; that is attributing
any arbitrary value, positive or negative, to the four constants
ahcdj and then reckoning the value of y for different values of x ;
afterwards plotting them to any convenient scale — remem-
XXXIX.] INVERSE VARIATION. 397
bering that the horizontal, scale and the vertical scale need not
necessarily be the same, but may be chosen independently, to
suit the convenience of the draughtsman.
Inverse Variation.
It frequently happens that two quantities are connected in
such a way that one increases when the other decreases. For
instance, the plentifulness of a commodity, say corn, and its
price. In a year of good harvest the price of wheat drops.
During a famine the price rises. It might happen that the
total money to be obtained, for the produce of a certain farm
acreage, was constant, whether the crop was plentiful or sparse.
Such a simple relation as that is not likely to hold exactly ; but
if it did, the two quantities — the price and the supply — would
be said to vary inversely as one another, that is to vary in
inverse simple proportion, so that their product remained
constant ; whereas if they varied in direct simple proportion
it would be their ratio which remained constant.
It does not follow that this law of simple inverse variation
holds because one quantity decreases and the other increases ;
all manner of complicated relations may hold between such
quantities ; the law of inverse proportion is the simplest
possible, and there are a great many cases where it holds,
or holds very approximately.
Take a piece of india-rubber cord or tubing, and pull it out
longitudinally ; as the length increases, the sectional area
diminishes, and it is a matter of measurement to ascertain
what relation holds between these two things.
If the tube were filled with water and were then pulled out,
the behaviour of the water would furnish a test of how the
sectional area varied with the length. If the water continued
to fill, or to stand at the same level in, the tube (which
might terminate at one end in a piece of glass tubing,
398 EASY MATHEMATICS. [chap.
for convenience of observation, and be closed with a solid
plug at the other), that would mean that the sectional area
and the length varied inversely as one another; in other
words that their product was constant; for the product of
length and sectional area is the volume, and it is the volume
which the water measures. Try the experiment. As a matter
of fact the water will be found to sink a little as the tube is
stretched, showing that the volume increases slightly : the
law of simple inverse proportion does not hold in this case.
Consider another case then. Take a vessel of variable
capacity, for instance a cylinder and piston ; or a tube open at
one end, which can be plunged mouth downward under a
liquid, like a long diving bell, and can be lowered or raised so
that the air in it shall be compressed or expanded at pleasure.
If a pressure gauge is attached, it will be possible to read
how the pressure increases as the volume diminishes ; and it
will be found that the two vary inversely as one another,
provided one is careful to take the whole dry-air pressure and
the whole volume into account. The product of pressure and
volume will be found experimentally constant ; because if one
is halved, the other will be doubled ; if one is trebled, the
other becomes one-third of its original value; and so on, a law
which is written :
1
pec-,
V
or vac -, or pv = constant.
P
These are all statements of the same fact ; p standing for
the pressure, and v for the volume of a given quantity of dry air
at constant temperature. If the vessel leaks, so that the
actual amount of air under observation changes, the law will
not apply. Nor will it hold if the temperature is allowed to
change. For if air is warmed it expands,, that is, it increases
XXXIX.] VARIATION DIRECT AND INVERSE. 399
in volume or in pressure or in both together ; and there is no
necessity at all for the pressure to decrease as the volume
increases unless the temperature is maintained constant.
Hence a complete statement is that pv = constant if T is
constant; or, if we choose so to express it, jpv = const, pro-
vided dT = 0, i.e. provided the difference of T is zero, which
is only another way of saying " if T is constant."
Suppose, subject to dT = 0, the pressure was increased by
a small amount dp, and the volume thereby decreased a small
amount - dv, we should have the new product of pressure and
volume expressed thus :
p'v = (p + dp)(v + dv\
and this product must equal the old product, pv, because of
the law that the product of pressure and volume is constant ;
so, multiplying out, we get
pv +p dv + vdp + dpdv = pv.
Wherefore, ignoring the second-order small quantity dpdv,
we have vdp+pdv = 0
dp dv
or -^ =
p V
as another statement of the law of inverse variation.
Summary.
Direct Variation
y = kx,
or di/ = kdx,
dy _ 1 _1
dx~ ~ X
or xdy-ydx = 0,
dy da
dy dx ^
or -^ = 0,
400 EASY MATHEMATICS. [chap, xxxrx.
Inverse Variation.
xy =^k,
or xdy + ydx = 0,
dy y k
dx X a;2'
di/ dx
or — + — = 0.
y X
^o that if J = ±|, that is if the ratio of differences is
numerically the same as the ratio of the quantities themselves,
it is a case of simple proportion ; but two distinct cases are
given by the alternative sign :
if the sign is + it is direct proportion ;
if the sign is - it is inverse proportion.
CHAPTER XL.
Pumps and Leaks.
When you pump water out of a reservoir, taking a barrel
full of water out at each stroke, the quantity of water
remaining decreases in an arithmetical progression, of which
the first term was the contents of the well, and the com-
mon difference is the contents of the pump barrel. If one
were called F, the other v (read big F and little v), the
level in the well would fall after successive strokes in the
following series :
Aq, /ij, W-g, ... /«-„,
where h^ is the height of the water before pumping began, and
h„ is the height after n strokes of the pump,
such that ^ = A- = -A_ = _^A__ = ...,
F V-v F-2v F-3v
a mode of writing which is called a continued proportion.
The quantity of water remaining in the well descends in a
decreasing Arithmetical Progression,
F, F- V, F- 2v, etc. ... V- nv,
and the well is empty when nv = F; or the number of strokes
required to empty it is the ratio of the capacities concerned, F/v.
The height or level of the water in the well goes in the
same sort of progression, and h„ is zero after F/v strokes.
L.E.M. 2c
402 EASY MATHEMATICS [cuap.
But now consider an air pump instead of a water pump.
The peculiarity of air or any other gas is that it always fills
the vessel which contains it, and does not accumulate in one
part as a liquid does. A bottle may be said to be " full " of air,
whether it contains much or little, in the sense that all parts
are equally full. It is always full in this sense, and it can
never be full in any other sense ; because however much air
is in, some more can always be pumped in : the only limit is
the bursting and destruction of the bottle. Or, if it were
made of porous material, it could be said to be as full as it
would hold when the rate of leak was equal to the rate at
which air was being pumped in; but even that could be
exceeded by beginning to pump a little faster.
With a liquid, on the other hand, a bottle may be properly
said to be " half-full " ; it can also be completely full, for you
cannot pump more than a certain quantity into a closed vessel.
If it is an open vessel the rate of leak at a certain definite
point becomes suddenly equal to the rate of supply, and the
vessel overflows ; which is a good practical method for main-
taining a constant level.
There is no such easy method for providing a constant air
or gas or steam pressure, though something of the kind is
attempted by means of a leak so adjusted as to suddenly
change from near zero to something considerable, at a certain
critical pressure, — such an arrangement being called a " safety-
valve." Locomotive boilers are usually filled with steam to
this pressure before a train starts on a long journey, and any
excess steam which the furnace generates blows off noisily in
a visible cloud.
If you were to pump air into a closed chamber, a barrel full
of atmospheric air would be injected at every stroke, and the
pressure would rise in an increasing arithmetical progression.
XL.] PUMPS. 403
Pq being the initial pressure before pumping, and j?„ the
pressure after n strokes,
such that % = P-l- = -=.&- = ... = -^^,
the pressure being proportional to, and a measure of, the extra
quantity of air injected. But if a pump is used to eject the
air, that is to say, to draw out from a closed chamber a barrel
full of air at every stroke, the law of decreasing pressure is
different : it then forms a geometrical progression.
For the same quantity of air is not removed each time.
The same volume is removed, but it is removed from air of
gradually diminishing density. The air keeps on getting
rarefied, and this rarefied air it is which has to supply the
pump barrel ; so that during every direct stroke the air which
at first occupied F" expands to occupy V+v, and then the
excess is ejected into the atmosphere at the return stroke of
the pump, ready for the expanding operation to begin again.
Thus, assuming the temperature to remain constant, we
have the pressure diminished at every stroke in the constant
ratio -j= — ; and the series p^, p^, p^, ... p^ is a decreasing
geometric progression with the common ratio F/(F+v).
F
So that p^ = F+^^«'
_ F _ f V \
\FVv)
2
Pi)'
Ps - yj^^P^ ~\f+v) ^«'
etc.,
the ratio of the pressure at beginning and end of any stroke
being constant, viz.
Pn V ^^ .
Hence the operation of an ordinary exhausting air pump is
404 EASY MATHEMATICS. [chap.
governed by the law of a decreasing geometrical progression ;
and an infinite number of strokes would be necessary completely
to empty the vessel, that is to reduce the pressure to zero.
Leaks and Compound Interest.
Now suppose instead of being pumped out the vessel were
full of compressed air and were to leak ; or suppose there
were a cistern full of water with a crack in the bottom of
it; the pressure in the one case and the level in the other
would fall according to a certain law. If the leakage rate
were constant, that is to say if the same amount of material
escaped every second, the law would be a decreasing A.P. ;
but that is never the case in fact. The size and circumstance
of the leak-orifice being constant, the amount of matter which
escapes through it depends on the force with which it is
urged, that is to say on the pressure behind it. A high
pressure reservoir, or a tall full cistern, would leak fast, the
air or water rushing out of the leak with violence ; whereas
towards the end, when the vessel was nearly empty, the rush
would have degenerated into a mere dribble or ooze, unless of
course it had worn the hole larger — which we will not suppose
to be the case. With a constant sized orifice the rate of leak
is therefore proportional to that which causes the leak, viz.
the pressure ; and so the pressure keeps on falling, at a rate
depending on itself : a curious and important, because, in one
form or another, a frequent case.
When you come to think, it is just the compound-interest
case, but inverted. Capital increases at a rate proportional to
itself : when small it grows slowly, that is by small additions,
when large it grows quickly. If we call the capital at any
dx
moment aj, its rate of growth will be — , since dx means an
CLl
increase of capital, and dt the time during which this increase
XL.] LEAKS. 405
occurs. In the case of capital the increase is somewhat
discontinuous : the interest is added every year, or it may be
every month, or perhaps every day, but not every instant.
liCt us assume that it is continuous however, so that it
increases from moment to moment at the rate -=- ; this rate of
Cit
increase will be proportional to z itself, and of course to the
percentage which is granted.
Suppose for instance it was 5 per cent., or -05, the law of
increase would be dx ^-
the interest, dx, is proportional to the rate, *05, to the capital
on which it is paid, x, and to the time during which it has
accumulated, dt; or dx = 'Obxdt
If it were 4 per cent., or 3 per cent , or 2 J per cent., of
course we should substitute '04, or -03, or -025, for the '05
numerical coefficient.
So with the leak, we have similarly to express that - dx^ the
loss of pressure, is proportional to the pressure, and to the
time, and to a leak-aperture constant which we will call k ; so
dx = -kxdt;
for to express that it is a loss and not a gain, a decrease not an
increase, we must apply to it a negative sign. The x might be
pressure, or it might be level or "head," — the two are
proportional in the case of water ; but level has no meaning in
the case of gas, so we will take "pressure" as the more
general term, and, denoting it by p instead of x, write the
law of leak, in the simplest possible case of a constant orifice, as
dt~ "P'
the rate of fall of pressure is proportional to the pressure from
instant to instant, diminishes as it diminishes, and does not
reach zero till the pressure reaches zero. The pressure in fact
406
EASY MATHEMATICS.
[chap.
decreases as a geometric progression. But it is a geometric
progression with one curious feature about it, it is continuous,
not discontinuous like numbers ; it does not go in steps or jumps,
like compound interest added every year or every day, but it
is like compound interest added or rather subtracted every
instant, with complete continuity, according to a smooth curve,
the logarithmic or g.p. curve, see page 357 or 101 or 179.
Cooling.
The cooling of a hot body under simplest conditions follows
just the same law ; the rate of fall of temperature is propor-
tional to the actual excess of temperature above surrounding
objects. If we denote this temperature by 6 and time by t^
dd
dt
= -kd
expresses the simplest possible law of cooling as the heat
escapes or leaks from the body into surrounding air or space.
It is instructive to put a thermometer into a flask or pan of
very hot water, and read the thermometer from time to time ;
at first every half-minute, or oftcner, then every minute, and
then as it cools more and more slowly, it will suffice to read it
every five minutes ; finally plotting the result thus :
Atmospheric
Temperature
Fio. 78.
XL.] LEAKS. 407
By choosing different vessels, say one black and one bright,
or by choosing similar vessels and filling them with different
liquids, one water and another turpentine say, many instructive
observations can be made; but a discussion of these would
carry us too far at present.
The curve of cooling is identical with the curve of leaking ;
and the curve of leaking might be plotted by reading a
pressure gauge, or by reading the level of a leaking water-
reservoir, from time to time. And both are curves of de-
creasing G.P. or are logarithmic curves.
Electric Leakage.
Experiments in electricity are more difficult, but if it were
possible to read satisfactorily by means of an electrometer the
potential of an electrified body or Ley den jar or condenser
which was steadily leaking, it would be found to obey the
same law.
Continuously decreasing G.P.
Now see how to express a quantity which decreases geo-
metrically with perfect continuity, and not by steps, as time
goes on. Notice that time is a continuous progression ; there
is no means of hurrying it ; one day is like another, and they
follow on with absolute regularity. Time is an inexorable
arithmetical progression, an increasing one if you think of
your birth, a decreasing one if you attend to the other end of
your life. Whatever can be varied, time cannot : at least not
by us.
Now when a vessel is leaking, the pressure is to be multi-
plied by a constant factor, some proper fraction, at each
successive equal interval of time.
SoPq at the start, when time is 0, or say at 12 o'clock noon,
becomes let us suppose ^p^ = Pj after the lapse of 1 hour, or
408 EASY MATHEMATICS. [chap.
at 1 p.m. If so, then in another hour it will have fallen to J^j,
or what is the same thing J^q ; in yet another hour, that is at
3 o'clock, it will be ^p^, and in n hours it will be —Pq ; hence
it will have fallen continuously down the decreasing geo-
metrical-progression-curve depicted on page 101.
But why should we suppose it halved in each unit of time 1
We can be more general than that, and say that it is reduced
to -Pq (read, ' one r-th of jo-nought ') after the lapse of one hour,
where r is some number greater than unity ; then in another
hour the pressure will have become -p-^, or what is the same
thing ijPo-
So the pressure at 2 p.m. is p^^ = r'-p^,
at 3 p.m. jOg = r->o,
and at n hours after noon p^ = '^~"Po'
Or we might say that, at any time t after the start, the
pressure is p = t~*Pq.
This then is the law
p = p^r
P -t
or — = r \
Po
or log^ - -t\ogr,
or log/?o-log^ = nogr,
log. = l^il^Y^i?,
all expressive of the very same fact.
Now r is a constant depending on the size of the opening the
viscosity of the escaping fluid (or on the covering and contents
XL.] LEAKS. 409
of the cooling body), and any other circumstance which can
affect the rate of leak, other than pressure (or temperature) and
time. And the log of r is the diminution of the log of the
pressure, during any lapse of time, divided by the time which
has elapsed. It is the ratio of the logarithmic diminution, or
decrement, to the time ; it is the decrement of the logarithm
of the pressure per unit time, and is technically known as
the " logarithmic decrement " of the pressure (or of the
temperature in the case of a cooling body, or of the potential
in the case of an electric leak, or of the level in the case of a
leaking cistern).
To measure log?-, all we have to do is to read the
pressure (or temperature, etc.) at any one instant, and then
read it again some time later, observing the interval of
time.
Let the two readings be denoted by p^ and ^„, and let the
intervening time be n seconds, then
log;?o - log j^n
n
is the logarithmic decrement per second, and is a measure of
the constant we have called log r.
Thus the law which was at first expressed in diiferential
form as ^ == _ K, or -^ = - kdt,
at ^ p
can also be expressed in integral form as
P "= Po^~' OT \ogpQ -\ogp = tlogr;
and it now becomes necessary to ascertain and express the
relation between the two constants k and r, which evidently
refer to the same sort of thing, viz. the fixed circumstances
of the leak.
Now remembering what we know of exponentials, let us see
if we can puzzle out the connexion between these constants.
410 EASY MATHEMATICS. [chap.
The law that we have written expresses the fact that
pressure decreases geometrically as time increases arithmeti-
cally : a constant factor is characteristic of one progression,
while a constant difference characterises the other.
We know that if p^ is the pressure at the era of reckoning,
that is at the instant from which time is to be reckoned, then
at any time t the pressure is ^ = i'o^ "^ ^^^ ^^ ^^7 ^^^^^
time f the pressure is^ = PqV'^', therefore
p = dp and
P
Now
let
the change
be small.
so that jt/
f
-t =
dt;
then the last
equation is
p + dp _
P
-f='
-dt
or
dp^
= r-'^'-l =
-dt. logr.
p
The last step we are not supposed to know enough yet to
justify ; but, assuming it and deferring its justification to
page 425, we see that
J- =, - log r . dt,
p ^ '
|=-^log
and this we can compare with the equation at which we
started (p. 405), ^jp
Thus the relation between the constants k and r is simply
k = logr,
and accordingly k is itself the logarithmic decrement of the
pressure per second.
XL.]
LEAKS.
411
The physical meaning of ^" is — -4^,
the physical meaning of log r is
log;?- log jp'
Summary.
These two
things turn
out to be
mathemati-
cally the
same thing.
But when logr is thus written, what base is intended for
the logarithm? There is nothing to say that the base is 10,
and indeed no explicit assertion has been made about any
base whatever : all that has been asserted is that p is to be a
quantity whose rate of change is proportional to itself, or
equal to itself when multiplied by the constant h or log r.
There is evidently something here worth investigation from
the purely mathematical point of view. It is a definite
mathematical question to put "What is that quantity whose
rate of change shall be proportional to itself? how is such a
quantity to be expressed in generaU" To investigate this
question, we can study the rates of variation of various
algebraic expressions.
^^
-^ ^,„ CHAPTER XLL
Differentiation.
Take the area of a square, and ask how it varies with the
side which contains it, when the square slightly expands. We
already know, but we will go through the process, especially
for a very small or infinitesimal increment of the side. Let
the side be x, and the area of the square be called «/, so that
y = a;2, then when the square is warmed a little, x increases by
the amount dx^ and y increases by the amount dy, such that
y + dy = (x + dxy
= x^ + 2xdx-h{dxy.
Now let dx be so small that the square of dx may be utterly
neglected in comparison with dx itself. In the limit suppose
dx actually infinitesimal, so that {dxY, being dx x dx, is again
or still further infinitesimal, even in comparison with dx ; then
y + dy = x^ + 2xdx;
but y = x^, therefore, subtracting, there remains dy = 2xdx,
1 = ^-
whence the rate of change of area of an expanding square, per
unit expansion of edge, is twice the length of one of the sides :
a very elementary statement, but not obvious. It is of course
a general analytic or algebraic result, and in no way depends
upon any geometrical meaning attached to y'^. The geometrical
CHAP. XLi.] RATES OF CHANGE. 413
square is only a special case, and it is convenient as an
illustration ; but it would be equally true for any other
variation of one quantity as the square of another; for
instance, the relation between the velocity of a falling body
and the height it has fallen, so well known in mechanics, is
written v^ = 2^/i, and this we can re- write in differential form
d{v^) = Ivdv = "Igdh,
dv a
or — = -
dh v'
which gives us the extra speed gained for each additional foot
or centimetre or other small unit of height.
Suppose for instance the height already fallen were 100 feet:
a dropped stone would have acquired a speed of 80 feet a
second. By the time it has dropped a foot more, the above
equation asserts that its speed will have increased by the
amount f f = f = '4 feet per sec.
We might also get the above relation thus :
v^ = 2ghj
v'^ = 2g{h+l);
:. v'^-v^ = 2g,
2(7 2q a
V +v 2v V
but in this case there is an approximation, because 1 foot
added to 100 is by no means infinitesimal though it is
moderately small. Consequently a sort of average or mean has
to be taken between v and v\ which in the limit would be
ultimately equal.
The expression dy = 2xdx we long ago illustrated by the
two strips, each equal to xdx in area, which went to form the
increase of surface in a square plate x'^ expanded by heat
(page 369) ; the little corner bit {dxf being ignored, because
when the strips themselves are infinitesimal, the infinitesimal
414 EASY MATHEMATICS. [chap.
bit of each at the ends is nought in comparison, or is said to be
an infinitesimal of the second order.
Similarly we may deal with the expansion or variation
of a cubical block of side x.
Denote its volume by y = a;^,
then when it expands infinitesimally
y + dy =^ {x + dxf
= x^ + Sx^dx + infinitely smaller quantities ;
.*. dy = 3x^dx,
dx
or the rate of expansion of a cubical volume, per unit increase
of a side, is three times the area of one of its faces.
Observe that the rate of increase of an area is a length,
while that of a volume is an area; but that is because the
rate of increase is taken per unit of length. If it were taken
per unit of time or of temperature, and if, as before, we write
y = x^, we could say that
dt ~ '^'^ dt'
or the rate of variation of the volume with respect to any
outside or independent variable, such as temperature, (-^\
(i.e. the cubical expansion per degree), is greater than the
rate at which each edge expands for the same variable, (j-j
(the linear expansion per degree), in the ratio of three times
the area of one of its faces.
Stated thus it is perhaps hardly geometrically evident,
nor need it be made so. What is geometrically capable of
illustration is the fact that
dy or d(x^) = Sx^dx.
xLi.] DIFFERENTIATION. 415
Other expressions of the same kind of fact are best
treated as mere analytic or algebraic statements, without any
necessary geometrical signification.
So we learn that to get the small change of any quantity
we have only to attend to the early terms of a binomial
expansion : two only, if the change is infinitesimal.
For instance, to find d{x% that is to express it in terms
of dx, we let x increase by dx, and then expand and neglect
all beyond the first power of dx \ thus
{x + dxf = x'^ + ioi^dx + higher powers ;
but x + dx = x\ and d(x^) means x'^-x\
therefore d{x!^) = A:X^dx.
Similarly d{:i^) = bx^dxy
d{x^) = Qx^dXy and so on, until
f/(«i2) = I2x^idx,
and dx"" = nx'^'^dx.
So also with fractional indices :
For instance, to find d Jx. Expand, and ignore all higher
powers of the infinitesimal quantity dx.
(x -^ dxy = x^ + Ix ^^ic + higher powers,
but (x + dxy = Jx',
so djx = Jx' -Jx = ^x'^dx = j^.
3 3 1
Again {x-\-dxy= x^ + ^x^dx,
3 1
therefore d Jx^ = dx^ = ^x^dx = ^Jx.dx.
Or take negative indices :
To find d(-\ or dx~'^; expand again,
{x + dx)~'^ = x~'^ - x'^dx,
416 EASY MATHEMATICS. [chap
but {x + dxy^ = '-,,
X
\X/ X X
di^] =±,-t-^-x-'^dx = -~.
And in general, whatever n may be,
dx«» = nx^-^dx,
a perfectly general result, worth thoroughly learning and
applying to special cases.
Even in the case when n = -0 it holds good ; for then it says
that dx^ = 0,
which we know is true, because x^ = 1 = constant, and so its
differences or variations must be zero.
If 71 = 1 it gives dx = 1 . dx, which is a mere identity.
If w = 2, it gives dx^ = 2xdx;
if w = 3, „ da^ = 3x^dx,
which we separately verified ; and so on.
Examples.
Check the following statements : —
dxy = 7afidx; ^ - 7x^; ~ = Sx^ ;
dax = adx; d{ax^) =2axdx; daa^ = Sax^dx.
d7x^
—1 — = 35.^* ; d(x + y) = dx + dy; d{ax + by) = adx + hdy ;
d{ax + hx'^) = adx + 2bxdx = {a + 2bx)dx;
d
^{cty + by^) = a + 2by; d(a + bx) = bdx;
XLi.] DIFFERENTIATION. 417
-^(a + bx) = b; d(a + bx + cx^) = {b + 2cx)dx;
^{a -{■bx + cx^ + x^) = b->r 2cx + 3x^ ; ~(A + Bx"") = nBx''-^ ;
dx dx^ '
^(A+Bx+Cx^-hD:^ + Eo^+ ..+Zx^')
dx
B + 2CX + Wx^ + iEx^ + . . . + nZx''-^ ;
dx, d (l\ _ I
y? ' dx\x) ~~ x^
dx\xy ~ x^' dx\x^) ~ x^
A.(—\ - — ' —(4- E\ - -A
dx\x ) ~ x^' dx\x ) ~ x'
d / a\ a d fA r, rt \ ^ n
dx\Fx) =- -w^' dxK^^^^^V = -^2+^;
4
dxy.
^^ + i + C+Dx + Ex^ = -'^-^, + D + 2Ex;
X ) xy x^
dx'-^x^dx; dx^"" '-^ 2jx'^
dx^ = ^x^dx; ^(»' = W^;
-j^a^x + b + -j-^j - j-j-^-^^-j^ = 2jxV~x)'-
~(a + bxf = ^(a+2abx + bV} = 2a6 + 2i% = 2b{a + bx);
L.E.M. 2d
418 EASY MATHEMATICS. [chap. xli.
dii^) = 2vdv; d(2gh) = 2gdh', j^{v^) = 2vj^ ;
j-(au^ + bu + c) = 2au + b', j-av^ = 2av~r- ;
du^ ' ' du du
d f o t ^ c^ du ,du ,» ,.du
j^iau^ + bn + c) = 2a«^ + 6^ = <2«« + *)^< '
CHAPTER XLII.
A Peculiar Series.
We are now able to write down a set of algebraic terms,
each of which shall be the differential-coefficient of the one
following it :
0 + l+x + ix^.
Of this we might make a regular series, for just as ^x^
differentiated gives x, so ^x^ differentiated would give x^, and
-1— would give ^x^. So also - — - — - differentiated would give
X?
- — -, and so on ; hence the series
a;2 a;3 x^ x^
l+^ + 2l + 3! + 4! + 5! + -
is a series which, when differentiated, gives as result
/g2 /g3 ^4
0 + l+»' + 2! + 3! + ¥! + -'
the very same series, — provided both extend to infinity :
a very curious case, the rate of variation of the series is
equal to itself. (Cf. p. 411.)
Such a series must therefore be appropriate for use in the
theory of leaks, that is for dealing with a quantity whose
rate of change is proportional to or equal to itself. We can
guess therefore that such a series must, when plotted, give
a curve of the nature of the exponential or logarithmic or
I..E,M, 2 D 2
420 EASY MATHEMATICS. [chap.
geometrical-progression or compound-interest curve. If we
call its value ?/, it satisfies the equation -f-, = y (cf. page 405).
It is a notable series. It is plainly convergent if x is less
than 1 ; but it is convergent even when x is equal to 1 or
greater than 1, because the denominators increase so fast;
they increase so fast indeed that a moderate number of terms
are generally sufficient to evaluate it fairly. The powers of x
grow fast in size when x is greater than 1, but the factorials
of the corresponding index-number grow still faster, and so
must ultimately get bigger ; for x stays as a constant factor
while being raised to any power, while in 'factorials' the
factor keeps on increasing. See page 315.
Let us try what the value of this series is when x — 1:
I + I+-I- + J+2V+ xhs + tI^ + Wtct + —
Greater than 2 and apparently less than 3, because
1 + 1 + J + J + 1 + xV + • • • would equal 3.
With patience its value can be reckoned to any desired
degree of accuracy, and it comes out
2-71828...,
a remarkable number, usually called e.
So now we can reckon what the series is when x has any
other value than unity. If we try it arithmetically f or a; = 2
we shall get
where we observe that though at first the numerators are
bigger than the denominators, afterwards, in spite of the well-
known rapid increase of the powers of 2, the factorials in the
denominators soon overpower them; for 2^2 = 4096, whereas
1 12 = 479,001,600, and is thus a hundred thousand times as
great.
To get a good value for this last series we must take a fair
XLii.] EXPONENTIAL SERIES. 421
number of terms, ten or a dozen, into account ; and if we do
we find the result
7-389...,
which is e^.
Similarly if we put a; = 3 we shall get 20-09... ,
which is e^ ;
whereas if we put a; = i we get 1-6467... ,
which is ^e.
Thus we suspect that the series
is in fact e* ; which is true, and it is called the exponential
series accordingly.
It has the singular and very useful property that its rate of
change is equal to itself, that is to say that
as we have already proved by differentiating each term of its
series separately and observing that the series is unchanged by
the process, being simply repeated over again.
Natural base of logarithms.
We can now apply this to logarithms ;
Let y = (^,
so that log 1/ = a; log ^, or log^ y = a; ;
we have just learnt that in this case
dx ^'
wherefore d log^ y == dx = —.
That is to say, the rate of change of the logarithm of a
yariable number js ec^ual to the rate of change of the number
422 EASY MATHEMATICS. [chap.
itself divided by that number ; provided the base of logarithms
is e.
If we take any other base than e we shall not get quite so
neat a result.
For let u = ?-*, where r is any number whatever,
then log M = a; log r (or log^ u = x\
and so — = c?a; . log r ;
wherefore d loff, u ^ dx = —. ,
o*- li log r
which only reduces to the above simple form when r = e;
otherwise it requires the natural logarithm of r to appear.
For instance, suppose we put u = lO^'^isosos^ ^g g, re-
presentation of r*, and change the index by a small amount,
say to 2*9180408; then, by referring to an ordinary seven-
figure table of logarithms, we shall see that the corresponding
change in the number u is -02
since u = 828-00 and v! = 828-02.
Now our assertion is that the change in the logarithm
(x'-x or dx, viz. -0000105) would have been equal to the
change in the number (u' -u or du) divided by the number
'02
{u or u'), that is to say would be practically equal to — — , if the
828
base had been e; but since the base is 10 this result has to be
divided by the further fixed quantity — the natural logarithm
of our artificial base 10 (which is a number approximately
equal to 2-3), in order to give the right result.
And it will be found accordingly that
-02
ooQ — o^ = -0000105, almost exactly ;
which illustrates the last algebraic line above.
XLiL] BASE OF LOGARITHMS. 423
Let us illustrate the occurrence of this natural logarithm of
10 by another numerical example, and at the same time make
an estimate of its value.
Suppose we put 10"2 to represent r*, and then allow the index
X to increase somewhat, say to 2-01 ; what will be the corre-
sponding change in r* %
We might write u = lO^oo, v! = lO^oi ;
so that <fo^H:^«^10^°;-10-^IO.o._l (1)
But in general when w = r*, _ = ?•* log ?', wherefore
nil
— = dxAogr = -OlloglO; (2)
and from these two expressions for the same thing we can
approximately evaluate the natural log 10. For equating (1)
and (2), we get
lO'Oi - 1
log 10 = ,Q^ = 100 X (1001 - 1)
= 100(^710-1) = 100(1-0232-1) = 2-32,
the last digit being affected by an error caused by the increase
in X not being infinitesimal.
This 2*3... is approximately the logarithm of 10 to a certain
base which has not been artificially specified, and which there-
fore must have entered automatically and "naturally" without
convention or artifice. What is that base %
It is a number such that if raised to the 2-3... power it will
equal 10. Call it w, then
2-3...1ogio7i = logiolO = 1,
or logioTi = 2^37^7= -434...,
wherefore n = 2-717... ;
which plainly points to e, with a deficiency of one part in two
thousand, or a twentieth of one per cent., due to approximations.
424 EASY MATHEMATICS. [chap.
Clearly therefore there is something peculiar about e as the
base of an exponential system : it is simpler than any other,
and it occurs automatically or naturally, unless we force some
other base in ; for when one finds that
whereas -y- = e*,
it becomes apparent that the base here automatically indicated
is such as to make loge = 1.
The fact is that logr, wherever it naturally occurs, means
log^r, and not a logarithm to any base at random. There
appears therefore to be a natural base for logarithms ; in this
respect differing entirely from the base or radix of the scales of
notation in ordinary counting. Ten, or twelve, or any number,
might be used for that — it was a pure convention ; but though,
as soon as we have adopted ten as the numeration base, ten
becomes specially convenient for practical calculations by aid
of logarithms also, yet ten is not the natural base of logarithms ;
nor is it the simplest base for an exponent. That property
specifically belongs to the incommensurable number called e.
The expansion of any exponential, such as r*, is now easily
managed in terras of e; for r may be expressed as e*, whence
r* = e** ; and we already know that
But since r = e^ it follows that k = log/, hence the above
expansion may also be written as
7^= l+a;logr + ^ ,| ^ +...,
where the logarithms are all to the base e.
XLii.] LEAKS. 425
For the special case when x is infinitesimal, say dt,
that gives us r*" = 1 + ^^ . log r,
wherefore j*^' - 1 = dt . log r,
which justifies a step assumed above (page 410) ; where,
however, it happened that the dt had a negative sign.
The whole theory of leaks or cooling is now quite easy, after
this incursion into the elements of pure mathematics : for
given that any quantity 'p (say pressure or temperature or
potential) changes at a rate proportional to itself, we can
write down instantly the following equivalent expressions
{t meaning time) :
dj)
It
kp,
P
d\ogp = -Tcdt,
log/ -log;? = k{t-t'\
\ogp + U = constant = logj?Q,
p =p^e'^.
All these are different modes of expressing the same
physical fact : the law of a cooling body, or a leaking reser-
voir, or any other of the many cases where rate of change
of a quantity is proportional to the quantity itself; and the
last gives explicitly the value of that quantity at any
instant, in terms of the initial value, the logarithmic decre-
ment, and the time.
And this must be regarded as typical of the way in which
general facts in Physics are simplified, summarised, and com-
pactly treated, by aid of more or less easy mathematics.
426 EASY MATHEMATICS. [chap. xlii.
So ends the present book, but in a subject like this there
can be no termination; every avenue leads out into infinity
and must be left with its end open. In no science are there
any real boundaries. In an advanced book a subjective
boundary may be reached, viz. the boundary of our present
knowledge ; but in an elementary book like the present that
is immensely far away, and the only terminus that can here
be reached is a terminus of print and paper.
APPENDIX.
I. Note on the Pythagorean Numbers (Euc. I. 47).
See Chapter XXXI.
By the Pythagorean numbers I mean simply those triplets of
integers which serve to express the relative lengths of the sides of
commensurate right-angled triangles : numbers therefore which
satisfy the conditions of Euclid I. 47, that any two of them are
greater than the third, and that the sum of the squares of two of
them equals the square of the third.
The only numbers mentioned in the text, page 272 are :
3, 4, 5 ; 5, 12, 13 ; and 8, 15, 17 ;
but there are innumerable others.
The subject is of no practical importance, and is only mentioned
here as an example of an easy kind of investigation in pure
mathematics which an enthusiastic and advanced pupil might be
encouraged to undertake, and which might lead him to take some
interest in less simple parts of the theory of numbers. The result
of the investigation, in this case, might be worded thus :
In general the sides and hypothenuse of a right-angled triangle
are incommensurable, but an infinite number of such triangles
exist in which the three sides may be represented by integers.
These are of some interest, and the simplest of them, when the
sides are in the ratio of the numbers 3, 4, 5, is frequently used by
surveyors.
A formula from which all such sides may be calculated is the
identity .«,. + (^J.(2!|*^)V
428 EASY MATHEMATICS.
meaning that ah and \{a'^-h'^) represent the sides containing the
right angle, and that \{a^ + h^) represents the hypothenuse.
To get a number of these triangles rapidly, without repetition
of shape, i.e. without obtaining mere multiples of other sets,
it is sufficient to choose as the auxiliary integers a and 6 any odd
numbers which are prime to each other. The reason for choosing
the auxiliary integers, a, 6, as odd numbers prime to each other, is
simply that if they contained a common factor the triplets obtained
from them would be merely a multiple set representing the same
shape as a simpler set ; whereas if one was even while the other
was odd, then a^-W would be odd, and \{a^-¥) would not be an
integer ; or if everything were doubled it would be merely
repeating the sides of some previous shape in another order.
Excluding multiple sets, one of the sides and the hypothenuse
are always represented by an odd number, and the other side by
an even number.
It is easy to prove that one of the sides containing the right angle
must always be a multiple of 4, that one of them (it may be the
same) must be a multiple of 3, and that one of the three sides (again
it may be the same) must be a multiple of 5 :
One of the two sides must be a multiple of 3.
Of course a, or ?>, may itself be a multiple of 3, thus satisfying
the condition for the side ah. If neither of them is, then a 6 is not
a multiple of 3, but in that case their squares must be of the form
3m + l, 3w+l [or rather of the form 6m + 1, 6w + l, since they are
odd numbers], and so the other side, viz. \{a^-b^) is then of the
form 3(m-w).
One of the two sides must be a multiple of 4.
Taking a = 2m + 1,
h = 2^1 + 1,
^(a2_62) ^ 2{m-n){m + n + \\
and either m- ,^or m + n-\-\ must be an even number, since their
difference is an odd number.
One of the three sides must be a multiple of 5.
If a or & is a multiple of 5, one of the sides, viz, the odd side,
is the required multiple. If not, its square must be of the form
40m + 9 or 40m + 1. If the squares of both have the same remainder,
APPENDIX.
429
the even side is a multiple of 5. If one has remainder 9 and the
other 1, the hypothenuse is a multiple of 5, of the form 20m + 6.
So if neither a or 6 is a multiple of either 3 or 5 it follows that
the number representing the even side has all three of the factors,
3, 4, 5 ; i.e. that it is a multiple of 60.
Moreover it can be shown that the hypothenuse is always itself
the sum of two square numbers, one odd and one even, and that the
odd side is the difference of those same squares. Thus, writing the
odd side as
(2m + l)(2ri+l) = {m + n-\-\f-{m-n)\
the even side is '2,{m-n){m-\-n + \ ),
and the hypothenuse is {m — 7if + {m-\-n-\- Vf.
The following is a table of the Pythagorean triplets, with the
mode of obtaining them displayed.
Auxiliary pair
Odd
Even
of numbers.
side.
side.
Hypothenuse.
3,1
3 = 4-1
4
5 =
4 + 1
5,1
5 = 9-4
12
13 =
9 + 4
7,1
7
24
25 =
16 + 9
9,1
9
40
41 =
25 + 16
11,1
11
60
61 =
36 + 25
13,1
13
84
85 =
49 + 36
15,1
15
112
113 =
64 + 49
17,1
17
144
145 =
81+64
19, 1
19 = 100-
81 180
181 =
100 + 81
5,3
15
8
17 =
16 + 1
7,3
21
20
29 =
25 + 4
9,3
27
36
45
etc.
11,3
33
56
65
13,3
39
80
89
15,3
45
108
117
17,3
51
140
149
19,3
57
176
185
7,5
35
12
37
9,5
45
28
53
11,5
55
48
73
13,5
65
72
97
15,5
75
100
125
17,5
85
132
157
19,5
95
168
193
430 EASY MATHEMATICS.
Auxiliary pair
Odd
Even
of numbers.
Bide.
side.
Hypothenuse.
9,7
63
16
65
11,7
77
36
85
13,7
91
60
109
15,7
105
88
137
17,7
119
120
169
19,7
133
156
205
11,9
99
28
101
13,9
117
44
125
15,9
135
72
153
17,9
153= 169-
-16 104
185 = 169-hl6
19,9
171
140
221
13, 11
143
24
145
15, 11
165
52
173
17, 11
187
84
205
19,11
209
120
241
The left-hand column is simply a series of pairs of odd numbers,
mainly prime to each other (but a few are included, for the sake of
systematic completeness, which are not prime, and therefore involve
repetition) ; the second column is their product ; the third column
half the difference of their squares ; and the fourth column half the
sum of their squares ; the incipient columns merely illustrate the fact
that the hypothenuse is the sum of two square numbers, one odd,
one even, whose difference is equal to the odd side.
The identity (m^-ny + {%n7if = {m^+'nFf
represents the facts most simply, where m and n are any integers.
One of these integers must be even and the other odd, with no
common factor, if mere multiples or repetitions of shape are to
be avoided.
II. Note on Implicit Dimensions {see pp. 53, 1 1 1, 143, 230).
The treatment of algebraical symbols as representing concrete
quantities, with all the simplification and increased interest which
this treatment involves, was first effectively called attention to by my
brother, Alfred Lodge, at that time Professor of Pure Mathematics
at Coopers Hill, and now a Mathematical Master at Charterhouse.
See Nature for July, 1888, vol. 38, p. 281, which was the sequel
APPENDIX. 431
to a pioneer paper read by him before the Association for the
Improvement of Geometrical Teaching, in January, 1888.
The subject was subsequently and independently developed by
Mr. W. Williams of South Kensington, now of Swansea ; and,
whether it has received full recognition or not, it has undoubtedly
justified itself in the eyes of all who have put it to the test of
practical experience. The whole subject is too large for this place,
but a few elementary remarks are appropriate :
Quantities of different kinds do not occur in one expression ;
in other words, the terms of an expression must all refer to the
same sort of things, if they are to be dealt with together or
equated to any one thing. Nevertheless an expression like
is common, and ^ may be a length ; which looks as if we could add
together a volume, an area, a length, and a pure number. Not so,
really, however ; suppressed or implicit or unexpressed or masked
dimensions must in that case exist in the numerical coefficients;
the coefficient 5 must implicitly or tacitly refer to a length, 2 to an
area, and 6 to a volume, if ^ is a length ; and thus all the terms are
really of the same kind. So they always will be in every real
problem.
When an equation contains terms of essentially different kind, it
must really consist of two or more equations packed together into
the apparent form of one. Thus ^/( - 1) is a quantity of essentially
different kind from 1 or (v' + l) ; the former being imaginary, the
latter real. Hence if ever they occur together in an equation, as
for instance in such an equation between complex quantities as
«v/( + l) + 6V(-l) = c^{ + l) + d^(-l\
or what is the same thing (writing /^( - 1) as i, for short, and s/{ + l)
as an unexpressed unity factor)
a + bi = c + di,
it must be regarded as a double equation, unless some of the
quantities a, 6, c, d are themselves complex ; for it can only be
interpreted and satisfied by the two separate equations
a = c and h = d.
In other words it is really two equations packed together for
brevity into a single statement,
432 EASY MATHEMATICS.
For if either of these conditions is not satisfied, if for instance b
is less than d, it is impossible to fill up the deficit by any increase
in the value of a, since that refers to a quantity of totally different
kind. A deficiency of oxygen in the atmosphere cannot be com-
pensated by a surplus of gold in a bank ; nor can deficiency of
beauty be effectively counterbalanced by excess of size.
The group met with in a German philosophical treatise (according
to a writer in the Hibhert Journal)^ as representing the class which
does not " count " for moral and intellectual purposes,
" cows, women, sheep, Christians, dogs,
Englishmen, and other democrats,"
cannot be regarded as classified according to a satisfactory system,
any more than can the somewhat similar group of tax-payers which
is at present disfranchised by Act of Parliament.
So that any conclusions, inferences, or results due to the
aggregation of such individuals in a community must be separable
into a series of independent conclusions, inferences, or results, except
in so far as some of these things are themselves complex, partaking
more or less of each other's characteiistics.
Sometimes we have equations among integers or other commen-
surable numbers, with incommensurables likewise involved, such as
If now m, ?i, ^, y are all to be considered integers or any vulgar
fractions or terminating or recurring decimals, i.e. unless some of
the quantities m, ?i, .r, y are in whole or in part themselves surds,
it must follow from the above statement that
X = m and y = n,
otherwise the equation cannot be satisfied.
Again suppose x means a distance measured horizontally, and y
a distance measured vertically, and the equation given is
ax-hby = cx + dy ;
it consists of two distinct and independent equations, unless a, b,
c, d are themselves directional quantities and not mere numbers ;
in that case, however, i.e. in case a, c are vertical lengths and 6, d
are horizontal lengths, the equation is quite homogeneous and
satisfactory, and denotes certain relations among rectangular areas.
Or a, c may be reciprocals of horizontal lengths, and b, d re-
ciprocals of vertical lengths ; and so on. But if a, 6, c, d are mere
APPENDIX. 433
numbers, we are bound as before to equate the coefficients, that is
to say to admit that a = c and h = d ; for no amount of horizontal
travel is equivalent to a rise, nor can horizontal dimension make
up for a deficiency in height.
In any single equation therefore, like v^ = 2gh for instance, where
one side is plainly the square of a velocity, the other side must
also be really, though not obviously, the square of a velocity.
And since g is an acceleration and A is a height, those who
know any mechanics will realise that the necessary condition is
thoroughly satisfied.
But when g is interpreted as 32 or 981, the fact is masked, as
facts often are masked by the incomplete method of arithmetical
or numerical specification. If 32 or 981 is regarded as a pure
number, which is all of g that it is customary actually to express in
writing, then the equation becomes an absurdity, since it appears
to assert that a velocity multiplied by itself results in a certain
multiple of an elevation.
But when it is remembered that the 32 means really 32 feet per
second per second, everything is perfectly right ; for, the height
being expressed in feet, the right-hand side of the above equation
is so many square feet per second per second, or square feet
divided by square seconds, which is the square of a velocity, in
perfect agreement with the left-hand side.
So also in the equation to a parabola, y = a-\-hx-\-cx^^ the
convention is that y is a vertical height, and x^ the square of a
horizontal length ; but, since all the terms must really be alike in
kind, it follows that a must be a vertical height (and it is : viz. the
intercept on the vertical axis), that h must be a ratio of vertical to
horizontal (and it is the tangent of an angle accordingly, namely
the value of -^ at the place where the curve cuts the axis of y) :
and further that c must be a sort of curvature, a quantity in-
volving a vertical direction once in the numerator and a horizontal
dimension twice in the denominator. It is in fact half . . •;„ ; it
{dxy
represents the rate at which the tangent to the curve swings round
as the ordinate travels uniformly along the axis of x ; and this
rate, when measured by changes in the tangent of the angle of
slope, is constant. Compare Chapter XXXIX,
434 EASY MATHEMATICS.
But 2^ = ;r is also a possible equation, and looks as if a vertical
height could be equivalent to a horizontal length. But it is only
an appearance, due to suppressed quantities. The coefficient 1,
not written, is really the tangent of an angle of 45°, and involves
the ratio of vertical to horizontal required to restore the balance
and common sense.
So also when y — x^ there is an unwritten unity coefficient which
is not a pure number, but an actual quantity, the ratio y : x^^ which
the equation asserts has in this case the magnitude 1.
Or when ^ = 6, if x is a length, it follows that the 6 is a numeri-
cal abbreviation for 6 feet or 6 centimetres or 6 miles, measured in
the same direction as x. See Article in " School World " for July,
1904.
It is frequently best to express these units fully, and not to get
too exclusively into the habit of writing a length as 50 without
saying whether inches or centiinetres is intended, or an age as 15
without saying years or months, or a price as 42 without saying
shillings or pounds, or whether it is per hundredweight or per ton
(compare pages 232, 4). For though these and other less customary
abbreviations are permissible among experts, beginners who get too
used to them are apt to degenerate into slovenly incompleteness
and inaccuracy, and to suffer by finding difficulties hereafter where
none exist.
III. Note on Factorisation (see Chapter XIV.).
A quadratic expression ax'^ -^-hx-^-c can be resolved into factors if
the middle term hx can be separated into two parts such that when
multiplied together the product is acx'^.
Thus take 3:pH10^+7,
and write it 3jp2 + ^x + 3.r + 7 ;
it becomes at once (3.r + 7) (.r + 1 ).
Again take bx"^ + 27.r + 28,
and write it hx"^ + 7a; + 20a; + 28 ;
it becomes (5a? +7) (a? + 4).
When a quadratic expression is thus written in four terms, such
that the product of the means is equal to the product of the
extremes, the four terms are necessarily proportional ; and if such
proportionality does not hold, you cannot factorise.
APPENDIX.
435
When they are proportional, as in the above case, and their sum
equated to zero, {5x^-h7a:) and (20.r + 28) must have a common
factor ; so also must {5x'^ + 20.r) and (7^ + 28) have another common
factor.
If we write such four proportional terms with the common factors
displayed, they must have the form
ac + ad+bc + bd ;
which terms geometrically represent themselves thus :
a h
ac
Ic
ad
bd
bj:^-
7.
20^
28
In the above example either aisx and b is 4, which are equivalent ;
or a is 5.r and b is 7, which are also equivalent. For instance the
diagram explicitly applied to the above case would look thus : no
scale being implied in the drawing, but simply a framework.
(5x) (7)
(X)
(4)
IV. Note on the Growth of Population (page 220).
In spite of what has been said in the text as to the danger of
applying the geometrical law of increase, or indeed any fixed law
of increase, to a given country or to any assemblage without taking
into account all the circumstances, nevertheless the growth of the
population of England and Wales during the last century illustrates
with remarkable closeness the geometrical-progression law.
The following is a table of the common logarithms (to base 10) of
the population of England and Wales for each decade from 1801 to
1901 ; together with their differences. If the geometrical law held
precisely, these differences would all be constant ; as it is, they
hover about a mean value, except in some of the early years of
the century, when they are abnormally big, — apparently as a
reaction from the Napoleonic wars, but doubtless also on account
of some applications of science, and other economic conditions.
436 EASY MATHEMATICS.
Population table of England and Wales for last century.
Logs.
Diffs.
1801
6-949026
-058050
1811
7-007076
-072114
1821
7-079190
-063724
1831
7-142914
-058869
1841
7-201783
-051739
1851
7-253522
-048944
1861
7-302466
-053794
1871
7-356260
•058286
1881
7-414546
-047890
1891
7-462436
•049796
1901
7-512232
From these data the curve of population might be plotted, and
it will be seen that from 1841 onwards it would be fairly steady,
the mean or average diflFerenee for 10 years during this period
being -051741. So the difference of logarithms for one year is -00517,
or to base e, -0119, or say -012. But we know that dlogep = -^
(see Chap. XLII.) ; hence -012, or 12 per thousand per annum, is
the average rate of increase of the population since 1841. The
curve is mainly a geometrical-progression or exponential curve,
with this value as the common ratio.
All the fluctuations noticed in such a curve could doubtless be
explained instructively, though to some extent hypothetieally, by
a Historian. For instance there is an excessive rate of growth in
the decade 1871 to 1881, which probably includes a period of good
trade ; but even that is not equal to the rates of increase nearer the
beginning of the century, when presumably the population was
emerging out of extreme poverty.
GLASGOW : PRINTED AT THE UNIVERSITY PRESS BY ROBERT MACl.BHOSE AND CO. LTD.
OF 3-FIGURE LOGARITHMS
2
3
4
5
6
7
8
9
^
'009
013
017
021
025
029
033
037
5
049
053
057
061
065
068
072
075
3
086
090
093
097
100
104
107
111
7
121
124
127
130
134
137
140
143
9
152
155
158
161
164
167
170
173
9
182
185
188
190
193
196
199
201
)7
210
212
215
218
220
223
225
228
13
236
238
241
243
246
248
250
253
»8
260
263
265
267
270
272
274
277
n
283
286
288
290
292
295
297
299
!2
342
362
380
398
415
431
447
462
)1
505
519
532
544
556
568
580
591
3
623
634
644
653
663
672
681
690
)8
716
724
732
740
748
756
763
771
^5
792
799
806
813
820
826
833
839
)1
857
863
869
875
881
887
892
898
)9
914
919
924
929
935
940
945
949
39
964
969
973
978
982
987
991
996
14 DAY USE
RETURN TO DESK FROM WHICH BORROWED
LOAN DEPT.
RENEWALS ONlr-TEl. NO. M2..340S
This book « due on the last date stamped below,
on the date to which renewed
^ewed books are subject to immediate recall
or
^
m