LIBRARY OF

WELLESLEY COLLEGE

PRESENTED BY TVof Horafo-rd

HIGHER ALGEBRA
A SEQUEL TO

ELEMENTARY ALGEBRA EOR SCHOOLS.

s.

HIGHER ALGEBRA
A SEQUEL TO

ELEMENTARY ALGEBRA FOR SCHOOLS

BY

H.

S.

HALL,

M.A.,

FORMERLY SCHOLAR OF CHRIST'S COLLEGE, CAMBRIDGE, MASTER OF THE MILITARY AND ENGINEERING SIDE, CLIFTON COLLEGE
AND
S.

;

K

KNIGHT,

B.A.,

FORMERLY SCHOLAR OF TRINITY COLLEGE, CAMBRIDGE, LATE ASSISTANT-MASTER AT MARLBOROUGH COLLEGE.

FOURTH EDITION.

Honfcon:

MACMILLAN AND
AND NEW YOKE.
1891
[The Right of Translation
is

CO.

reserved.}

2-

TO

422

12/

First Printed 1887.

Second Edition with corrections 1888.

Third Edition revised and enlarged 1889. Reprinted 1890. Fourth Edition 1891.

}

PREFACE.
The
present work
is

intended as a sequel to our Elementary

Algebra for Schools. The first few chapters are devoted to a fuller discussion of Ratio, Proportion, Variation, and the Progressions, which in the former work were treated in an elementary manner and we have here introduced theorems
;

and examples which are unsuitable
reading.

for

a

first

course of

From part new

this point the

to
:

work covers ground for the most the student, and enters upon subjects of special
these

importance

we have endeavoured

to treat

minutely

and thoroughly, discussing both bookwork and examples witli that fulness which we have always found necessary in
our experience as teachers.
It has

been our aim to discuss
within

all

the essential parts
limits
it

as

completely as possible

the

of

a

single

volume, but in a few of the later chapters
possible to find

has been im;

room

in all such cases
first

more than an introductory sketch our object has been to map out a suitable
for

course of reading, referring the student to special treatises

for fuller information.

In the chapter on Permutations and Combinations we
are

much indebted to the Rev. W. A. Whitworth for permission to make use of some of the proofs given in his Choice and Chance. For many years we have used these proofs in our own teaching, and we are convinced that this

;

vi

PREFACE.

part of Algebra

sense reasoning from first principles by a system of than by the proofs usually found in algebraical text-books. The discussion of Convergency and Divergency of Series

made common
is

far

more

intelligible to the

beginner

always presents great difficulty to the student on his first The inherent difficulties of the subject are no reading.

doubt considerable, and these are increased by the place it has ordinarily occupied, and by the somewhat inadequate treatment it has hitherto received. Accordingly we have
placed this section somewhat later than is usual; much thought has been bestowed on its general arrangement, and

on the selection of suitable examples to illustrate the text and we have endeavoured to make it more interesting and intelligible by previously introducing a short chapter on Limiting Values and Vanishing Fractions. In the chapter on Summation of Series we have laid

much

stress

on the " Method of Differences" and

its

wide and
is

important applications.

The

basis of this

method

a well-

known formula

in the Calculus of Finite Differences,

which in

the absence of a purely algebraical proof can hardly be considered admissible in a treatise on Algebra. The proof of the
Finite Difference formula which 396,

we have given

in Arts. 395,

we

believe to be

new and

original,

and the development

of the Difference

Method from

this

formula has enabled us to

many interesting types of series which have hitherto been relegated to a much later stage in the student's reading.
introduce

We

have received able and material assistance in the

chapter on Probability from the Rev. T. C.
Christ's College, Brecon,

Simmons

of

and our warmest thanks are due to him, both for his aid in criticising and improving the text, and for placing at our disposal several interesting and
original problems.
It
is

hardly possible to read any modern treatise on

Analytical

Conies or Solid Geometry without some know-

PKEFACE.
ledge of Determinants and
their
applications.

yii

We
may

have

therefore given a brief elementary discussion

of Determi-

nants in Chapter

xxxm.,

in the

hope that

it

provide

the student with a useful introductory course, and prepare him for a more complete study of the subject.

The
in the

last

chapter contains

all

the most useful propositions

Theory of Equations suitable for a first reading. The Theory of Equations follows so naturally on the study of Algebra that no apology is needed for here introducing prowhich usually find place in a separate treatise. In fact, a considerable part of Chapter xxxv. may be read with advantage at a much earlier stage, and may conveniently be studied before some of the harder sections of previous
positions

chapters.
It will

be found that each chapter
itself,

is

as nearly as possible

complete in

so that the order of their succession can
;

be varied at the discretion of the teacher
*

but

it is

recom-

mended

that

all

sections

marked with an

asterisk should be

reserved for a second reading.

In enumerating the sources from which we have derived
assistance in the preparation of this work, there
to
is

one book
indebted.

which

it is

difficult

to

say

how

far

we

are

Todhunter's Algebra for Schools and Colleges has been the recognised English text-book for so long that it is hardly

any one writing a text-book on Algebra at the present day should not be largely influenced by it. At the same time, though for many years Todhunter's Algebra has been in constant use among our pupils, we have rarely adopted the order and arrangement there laid down; in many chapters we have found it expedient to make frequent use of alternative proofs; and we have always largely supThese notes, plemented the text by manuscript notes. which now appear scattered throughout the present work, have been collected at different times during the last twenty
possible that
H. H. A. b

;

Viii

PREFACE.
it is

years, so that

impossible to

make

definite acknowledge-

ment

in every case

other writers.

where assistance has been obtained from But speaking generally, our acknowledge-

ments are chiefly due to the treatises of Schlomilch, Serret, and Laurent; and among English writers, besides Todhunter's Algebra, we have occasionally consulted the works of De Morgan, Colenso, Gross, and Chrystal. To the Rev. J. Wolsienholme, D.Sc, Professor of Mathematics at the Royal Indian Engineering College, our thanks
are

due for his kindness in allowing us to select questions from his unique collection of problems and the consequent
;

gain to our later chapters
It

we

gratefully acknowledge.

remains

for

us to express our thanks to our colleagues
so largely assisted us in
;

and friends who have
the Rev. H.

reading and

correcting the proof sheets

in particular

we

are indebted to

C

Watson
it.

of Clifton College for his kindness in

revising the whole work,
in every part of

and

for

many

valuable suggestions;

**'

1887

"

H.

S.

S. R.

HALL, KNIGHT.

j
<

PREFACE TO THE THIRD EDITION.
and examples are substantially the same as in previous editions, but a few articles havej been recast, and all the examples have been verified again. We have also added a collection of three hundred Miscellaneous Examples which will be found useful for advanced These examples have been selected mainly but students. not exclusively from Scholarship or Senate House papers much care has been taken to illustrate every part of the subject, and to fairly represent the principal University and
In
this edition the text

Civil Service Examinations.
March, 1889.

CONTENTS.
CHAPTER
I.

ratio.
PAGE
.

Commensurable and incommensurable quantities Ratio of greater and less inequality
i

2 3

a
a1

_c _e

_

/pan + qc n +re n +

...\

n
'

b~d~f~"'~\pb n + qd n + rf n +...J
+ a 2 + a 3 +... + an b l + b 2 + b.i + ... + b n
lies

between greatest and least of fractions

flh

V

an

Cross multiplication

8
9

Eliminant

of three linear equations
I

Examples

10

CHAPTER
Definitions

II.

proportion.
13
definitions

and Propositions Comparison between algebraical and geometrical Case of incommensurable quantities Examples II.

1G
17 19

CHAPTER
If

III.

VARIATION.
21

Ace B, then

A = mB

Inverse variation
Joint variation
li

22
23
is

Ace

B when G

constant,

and A

& C when B
.

is

constant, then

A=mBG
Illustrations.

23

Examples on

joint variation

.

.

.

.

.21
20 b-1

Examples

III

;

.

CONTENTS.

CHAPTER
Sum
of

IV.

ARITHMETICAL PROGRESSION.
PAGE
series

n terms of an arithmetical Fundamental formulae Insertion of arithmetic means Examples IV. a
Discussion of roots of dn~ + (2a

28

29
31

31

-d)n-2s =

33
35

Examples IV. b

f'

CHAPTER

V.

GEOMETRICAL PROGRESSION.

Insertion of geometric

means

....
.

38
39

n terms of a geometrical series of an infinite geometrical series Examples V. a.
of

Sum Sum

40
41

Proof of rule for the reduction of a recurring decimal

43

n terms Examples V. b
of

Sum

of

an arithmetico-geometric

series

44 45

CHAPTER VI.

HARMONICAL PROGRESSION. THEOREMS CONNECTED WITH THE PROGRESSIONS.

Reciprocals of quantities in H. P. are in A. P.

Harmonic mean

......
.

Formulae connecting A. M., G.M., H.M. Hints for solution of questions in Progressions

Sum of squares of the natural numbers Sum of cubes of the natural numbers 2 notation
Examples
VI.
a.

.....
pyramid on a square base

Number

of shot in

Pyramid on a triangular base Pyramid on a rectangular base Incomplete pyramid Examples VI. b

.....
VII.
scales of notation.

5J

CHAPTER
Examples VII.
a.

Explanation of systems of notation
Expression of an integral number in a proposed scale Expression of a radix fraction in a proposed scale

57 59
.

59
01

CONTENTS.

XI
PAGE

The

difference between a

number and

tho

sum

of its digits

is

divisible

by r - 1
Proof of rule for " casting out the nines "
Test of divisibility by r

62

C3
64

+1

Examples VII. b

65

CHAPTER

VIII.

SURDS AND IMAGINARY QUANTITIES.
a
sjb

nationalising the denominator of

+ jc + s/d

07

Rationalising factor of fJa±Z/b Square root of a + Jb + *Jc + Jd

68

69

Cube root of a + *Jb Examples VIII. a. Imaginary quantities

70
72 74
.

J -ax J -b= - sjab
If
If

75
.

a + ib = 0, then a = Q, b =
a

75

+ ib = c + id, then a = c,
of product
is

b

=d

75
77

Modulus

equal to product of moduli

Square root of a + Powers of i
.

ib

77

79
;

Cube roots of unity Powers of u Examples VIII. b.
.

1 + w -f or =

79

80
81

CHAPTER
A
Conditions for

IX.

THE THEORY OF QUADRATIC EQUATIONS.

quadratic equation cannot have more than two roots
real, equal,

...
equal in magni-

83

imaginary roots
c

84

Sum

of roots

=


b

a

,

product of roots = a

Formation of equations when the roots are given
tude and opposite in sign,
(2)

.....
(1)

85 86

Conditions that the roots of a quadratic should be
reciprocals

88 88
in general the

Examples IX. a For real values of x the expression ax 2 + bx + c has
sign as a
;

same
90 92
93

exceptions

Examples IX. b
Definitions of function, variable, rativnnl integral function

...
a

Condition that ax 2 + 2hxy+ by 2 + 2gx
linear factors

+ 2fy + c may

be resolved into two
9
i

Condition that ax
root

2

+ bx + c = and

a'x- + b'x

+ c' = may have

common
96
96

Examples IX.

c.

Xll

CONTENTS.

CHAPTER
Reciprocal equations

X.

MISCELLANEOUS EQUATIONS.
page
.

Equations involving one unknown quantity

97 100
101

Examples X. a
Equations involving two unknown quantities

Homogeneous equations Examples X. b
Examples X.
c.
;

....

103

104
106 107 109 111

Equations involving several unknown quantities
Indeterminate equations
easy numerical examples

Examples X. d

113

CHAPTER XL
Preliminary proposition

permutations and combinations.

.115

115 Number of permutations of n things r at a time 117 Number of combinations of n things r at a time The number of combinations of n things r at a time is equal to the .119 number of combinations of n things ?i-rata time Number of ways in which m + n +p + ... things can be divided into
. .

classes containing m, n, p,

...

things severally

....
when p

120
122

Examples XI. a
Signification of the terms 'like'

and 'unlike'
all at

.....
a time,

124

Number
Number

of arrangements of

n things taken

things
.

are alike of one kind, q things are alike of a second kind, &c.
of

125

permutations of n things r at a time, when each

may

be

repeated

126
127
greatest
.

The total number of combinations of n things To find for what value of r the expression n Gr is Ab initio proof of the formula for the number
things r at a time

.

.

127 128

of combinations of n

Total number of selections of

p + q+r+

...

things, whereof

p

are alike

of one kind, q alike of a second kind, &c

129
131

Examples XI. b

CHAPTER
Illustrations of the

XII.

mathematical induction.
133
134

method of proof

Product of n binomial factors of the form x + a

Examples XII

135

CONTENTS.

Xlii

CHAPTER XIII.
11
,

BINOMIAL THEOREM.

POSITIVE INTEGRAL INDEX.

Expansion of (x + a) when n is a positive integer General term of the expansion The expansion may be made to depend upon the case in which the first term is unity Second proof of the binomial theorem Examples XLII. a The coefficients of terms equidistant from the beginning and end
are equal

....

PAGE
137
139

140
141

142 143

Determination of the greatest term

143

Sum Sum

of the coefficients
of coefficients of

146

odd terms

is

equal to

sum

of coefficients of even

terms

146

Expansion of multinomials Examples XIII. b.

146
147

CHAPTER

XIV.

BINOMIAL THEOREM.

ANY INDEX.
150
153 155
155

Euler's proof of the binomial theorem for any index

General term of the expansion of

(1

+ x)'

1

.....

Examples XIV. a Expansion of (l-rx) n is only arithmetically intelligible when x<l The expression (.r-f?/)'1 can always be expanded by the binomial
theorem General term of the expansion of
(1

-

_n .r)

....
(l

157
157
158 159 161

Particular cases of the expansions of

(1

- x)~ n

Approximations obtained by the binomial theorem Examples XIV. b.
Numerically greatest term in the expansion of

+ x)

n
.

162
164 105
166

Number

of

homogeneous products of

r

dimensions formed out of n

letters

Number of terms in the expansion of a multinomial Number of combinations of n things r at a time, repetitions being allowed
Examples XIV.
c

107

CHAPTER XV.
General term in the expansion of
positive integer
(a

MULTINOMIAL THEOREM.
+ bx + ex 2 + dx 3 +
...) p ,

when ^
n
,

is a

170
<lv :i

General term in the expansion of (a + bx + cx- +
is

+

...)

when

//

a rational quantity

171

Examples

XV

173

XIV

CONTENTS.

CHAPTER
Definition.

XVI.

LOGARITHMS.
PAGE
175 176

N=a)og a N

Elementary propositions Examples XVI. a Common Logarithms
Determination of the characteristic by inspection

178

.....
• •

.179
180
181

Advantages of logarithms to base 10 Advantages of always keeping the mantissa positive Given the logarithms of all numbers to base a, to find the logarithms
to base b
.

....
.

182
183 183
185

.

log a &xlog 6 a = l

Examples XVI. b

CHAPTER
Expansion
e is

XVII.

EXPONENTIAL AND LOGARITHMIC SERIES.
e

of ax .

the limit of

(
(

Series for l\ n
1

187
is infinite

+-

V

V
)

,

when n

188
191

Expansion of

log,, (1

+ x)
.

Construction of Tables of Logarithms

192
e

Rapidly converging series for

log,,

(n

+ 1) - log n

194
195 195

The quantity e is incommensurable Examples XVII

CHAPTER
Interest

XVIII.

INTEREST AND ANNUITIES.
. .
.

and Amount of a given sum at simple interest Present Value and Discount of a given sum at simple interest Interest and Amount of a given sum at compound interest Nominal and true annual rates of interest
.

.198
.

.

198 199

.

.

Case of compound interest payable every moment Present Value and Discount of a given sum at compound interest Examples XVIII. a
Annuities.
Definitions

....
.

200 200
201 202

.

202
203 203 204
204
.

Amount Amount

of unpaid annuity, simple interest
of unpaid annuity,

compound

interest

Present value of an annuity, compound interest Number of years' purchase Present value of a deferred annuity, compound interest

.

.

.205
206
206

Fine for the renewal of a lease Examples XVIII. b

CONTENTS.

XV

CHAPTER
Elementary Propositions
Arithmetic

XIX.

INEQUALITIES.
PAGE 208
is

mean of two positive quantities

greater than the geometric

mean The sum of two
equal

209
quantities being given, their product
:

is

greatest

when
are

they are equal

product being given, the

sum

is least

when they

210
211

The arithmetic mean of a number of positive quantities is greater than the geometric mean Given sum of a, &, c, ...; to find the greatest value of am b n c p Easy cases of maxima and minima Examples XIX. a The arithmetic mean of the ?/i th powers of a number of positive
quantities
is

212 212

213

greater than

m th

except
If a

when

m lies between

power of and 1
(

their arithmetic

mean,
214

b

and

b are positive integers,

and a>b,

1

+-

)

>

(

1

+^

)

216

? 1> * > » >0'Vrr| > vrrf y
'a

217 217 218

+ b\ a+b

Examples XIX. b

CHAPTER XX.
Definition of Limit

LIMITING VALUES AND VANISHING FRACTIONS.

a when x is zero By taking x small enough, any term of the series a + a rr + a^x- + ... may be made as large as we please compared with the sum of all that follow it; and by taking x large enough, any term may be made as large as we please compared with the sum of all that
Limit of a
x3
is

+ a x x + a 2 x" + a 3

+ ...

....
.
.

220
222

precede

it

.

222
.

Method

of determining the limits of vanishing fractions

221
226

Discussion of
equations

some

peculiarities

in the

solution

of

simultaneous

Peculiarities in the solution of quadratic equations

....

227 228

Examples

XX
XXI.
CONVERGENCE AND DIVERGENCY OF

CHAPTER

Case of terms alternately positive and negative
Series is convergent
if

.....

SERIES.
230
232

u Lim ~ n
«u
-i

is less

than

1

.

XVI

CONTENTS.
PAGE

Comparison

of 2rtn with

an auxiliary

series

2v n

.

.

234 235
237

The

auxiliary series

^p

+ 2P + 3~p +

.

Application to Binomial, Exponential, Logarithmic Series
log
71

.

Limits of

Product of

n an

and nx n when n

is infinite

.

238
238
241

infinite
a.
.

number

of factors

.

Examples XXI.

.

w-series is convergent

when

u
v- series is

v
.

convergent,

if

243 244 245 247 248
248

u ji-i

v n-l
.

Series is convergent

if

Lim Lim

]n
<

I

\un+i

—— - 1
)

)

J
]
.

Series
Series

is

convergent
(n)

if

(

n log
series

—— >
2a
,l

20

compared with

{n)

.

The

auxiliary series

2

.

n

(log n) p

Series is convergent if

Lim
series

\n

(

—~

1

J

-

l| log

n

two infinite Examples XXI. b.
Product
of

249 252

CHAPTER
If the equation

XXII.

UNDETERMINED COEFFICIENTS.
.
.

f(x)-=0 has more than n roots, it is an identity Proof of principle of undetermined coefficients for finite series

254

.

.

254

Examples XXII. a
Proof of principle of undetermined coefficients for infinite series
.

256
.

Examples XX1T. b

257 2C0

CHAPTER
Use of partial fractions Examples XXIII

XXIII.

PARTIAL FRACTIONS.
261

Decomposition into partial fractions
in expansions

265 265

CHAPTER XXIV.
Scale of relation

recurring series.
267

Sum

of a recurring series

269
269

Generating function

Examples XXIV

272

.

CONTENTS.

XV ii

CHAPTER XXV.

continued fractions.

Conversion of a fraction into a continued fraction Convergents are alternately less and greater than the continued fraction Law of formation of the successive convergents

....
.

PAGE
273
275

275 27G
277

Pn&w-l-Pn-l4n=(~ Examples XXV. a. The convergents gradually approximate to the continued fraction Lhnits of the error in taking any convergent for the continued fraction Each convergent is nearer to the continued fraction than a fraction
.

1)n

278
279

with smaller denominator
Pp'
,:>

.

280
p'
.

or <x~, according

P as->
q

or

<—
q

281

qq

Examples XXV. b

281

CHAPTER XXVI.
Solution of ax-bi/ = c

indeterminate equations of the first
DEGREE.
284

Given one solution, to find the general solution Solution of ax + by = c

286 286

Given one solution,

to find the general solution

287

Number

of solutions of

ax + by = c

.

287
c'z
.

Solution of ax + by + cz = d, a'x + b'y +

Examples XXVI.

...

= d'

289

290

CHAPTER XXVII.
Numerical example

recurring continued fractions.
292
is

A

periodic continued fraction

equal to a quadratic surd

.

.

.

293
21)4

Examples XXVII. a
Conversion of a quadratic surd into a continued fraction
.
.

.

295

The quotients recur The period ends with a partial quotient 2a x The partial quotients equidistant from first and The penultimate convergents of the periods Examples XXVII. b.
. .

296 297
last are equal
.
.

298
299 301

CHAPTER

XXVIII.

indeterminate equations of the second
DEGREE.

Solution of ax 2 + 2hxy

+ by* + 2gx + 2fy + c =
be solved

303

The equation # 2 - Ny2 =l can always

304

xviii

CONTENTS.

CONTENTS.

XIX
PAGE

If

a

prime to b, then different remainders
is

a, 2a, 3a,

......
...

(6-

1)

a when divided by

6 leave

350
352 352

(p(abcd...)=<p(a)(p(b)<p(c) <p(d)

PO-»'(i-i)(i-J)(i-l)
Wilson's Theorem
:

1

+

\p

-

1

=M

(p)

where p

A

property peculiar to prime numbers

Wilson's Theorem (second proof)
Proofs by induction

.......
the
general

....
is

a prime

354 354
355

35G
357

Examples XXX.

b.

CHAPTER XXXI.
Law
a,+

theory

of

continued

FRACTIONS.
of formation of successive convergents
-

.

359

—— —a.2

+

...

has a definite value
to

if

Lim

'"
y

n4' 1

>0

.

362

n+l

The convergents

\

h
. .
.

a l ~ a 2~
if

are positive proper fractions in ascend-

ing order of magnitude,

a n <kl

+ bn
.

363

General value of convergent when a n and b n are constant Cases where general value of convergent can be found
a x + a2 +
is

364
365

incommensurable,

if

— <1
cl
.

366

Examples XXXI. a
Series expressed as continued fractions

367
369
371

'Conversion of one continued fraction into another

Examples XXXI. b

372

CHAPTER XXXII.
and illustrations. (Examples XXXII. a /Compound Events
Definitions
I
[

probability.
.

Simple Events

373

376
377
is

Probability that two independent events will both happen

pp'

.

378 379
381

The formula holds

also for dependent events

.

Chance of an event which can haj^pen in mutually exclusive ways Examples XXXII. b Chance of an event happening exactly r times in n trials Expectation and probable value "Problem of points"
.

383
385

....... .......

386

388

XX

CONTENTS.
PAOE

Examples XXXII.
Inverse probability

c.

389
391
.

Statement of Bernoulli's Theorem
Proof of formula

P P Q r = ^rj-jn

Concurrent testimony Traditionary testimony

.... ....
dktkrminants.

392
396
899

Examples XXXII. d
Local Probability. Geometrical methods Miscellaneous examples

401

402
405

Examples XXXII.

e

CHAPTER XXXIII.

Eliminant of two homogeneous linear equations Eliminant of three homogeneous linear equations Determinant is not altered by interchanging rows and columns Development of determinant of third order
.
.
.

.....
.

409

.410
.

410
411

Sign of a determinant
.

is
.

altered by interchanging
. . .

fcw<

adjacent rows or

columns If two rows or columns are identical, the determinant vanishes A factor common to any row or column may be placed outside Cases where constituents are made up of a number of terms
.
.

412
. .

112

.

.

412

.

.

413

Keduction of determinants by simplification of rows or columns Product of two determinants

.

.111
417 419 422
423
42jl

Examples XXXIII. a
Application to solution of simultaneous equations

Determinant of fourth order Determinant of any order
Notation

Sia^-^
b.

...
. .

.

.

.

... ...
.

425
\-r,

Examples XXXIII.

.

CHAPTER XXXIV.

miscellaneous theorems and examples.
429

Keview of the fundamental laws of Algebra f(x) when divided by x - a leaves remainder Quotient of / (x) when divided by x - a Method of Detached Coefficients
.

432 433

434
434
435

Horner's Method of Synthetic Division Symmetrical and Alternating Functions

.

Examples

of identities

worked out
.

437

List of useful formula?

438

CONTENTS.

XXI
PAGE

Examples XXXIV. a
Identities proved

438
440
441

by properties of cube roots of unity

Linear factors of a 3 + 6 3 + c 3 - Sabc

Value of an + bn + cn when a + b + c = Q

442
442

Examples XXXIV.
Elimination

b.

.

444
444 445 446
of elimination

Elimination by symmetrical functions
Euler's

method

Sylvester's Dialytic

Bezout's method

....
.

Method

446
447 449

Miscellaneous examples of elimination

Examples XXXIV.

c.

.

.

.

CHAPTER XXXV.

theory of equations.
.
.

Every equation of the n th degree has n roots and no more Kelations between the roots and the coefficients These relations are not sufficient for the solution Cases of solution under given conditions Easy cases of symmetrical functions of the roots

.

452 452
454 454
455

Examples XXXV. a. Imaginary and surd roots occur in pairs Formation and solution of equations with surd
Descartes' Kule of Signs

456
roots

....

457

458
459

Examples XXXV. b
Value of /(.r +
Calculation of
//).

Derived Functions
process

460 462
463 464

f(x+h) by Horner's
value gradually

/

(

x)

changes

its

If f(a)

and/ (b)
6

are of contrary signs, f(x)

=

has a root between
464

a and

An An
If

equation of an odd degree has one real root
equation of an even degree with
roots
a,
its last

465
465

term negative has two real
466 467 468

has r roots equal to Determination of equal roots

/ (x) =

f (x) =

has r - 1 roots equal to a

.

/'(*)_
J
(X)

I

1
,

1
,

x-a
of

x-b

x-c
.

an assigned power of the roots Examples XXXV. c Transformation of equations Equation with roots of sign opposite to those of f(x) = Equation with roots multiples of those of f{x) =0

Sum

468

470
471
.
.

.

471

.

472

XX11

CONTENTS.

Equation with roots reciprocals of those of /
Discussion of reciprocal equations

....
(x)
.

=

.

.

PAGE 472 473
475

Equation with roots squares of those of f(x) = Equation with roots exceeding by h those of f (x) = Bemoval of an assigned term Equation with roots given functions of those of f{x)-. Examples XXXV. d Cubic equations. Cardan's Solution
Discussion of the solution
Solution by Trigonometry in the irreducible case
.

.

.

475

.

476
477

.

.

.

.

.

Biquadratic Equations.
Descartes' Solution

Ferrari's Solution

.

.

.

Undetermined multipliers
Discriminating cubic
;

.

roots all real

.

478 480 481 482 483 484 486 486
487
488 490
525

Solution of three simultaneous equations

y =1, a+\ + b + \ + c + \

x

&c.

.

Examples XXXV. e. Miscellaneous Examples Answers

HIGHER ALGEBRA.
CHAPTER
RATIO.
I.

Ratio is the relation which one quantity Definition. bears to another of the same kind, the comparison being made by considering what multiple, part, or parts, one quantity is of the
1.

other.

A

The and

ratio of

B
To

is usually written A to The quantities B. are called the terms of the ratio. The first term is
:

A

B

called the antecedent, the second
2.

term the consequent.
is

find

what multiple or part A

of B,

we

divide

by
-^
,

B

;

hence the ratio

A

:

B may

A

be measured by the fraction

and we

shall usually find it convenient to

adopt this notation.

In order to compare two quantities they must be expressed in terms of the same unit. Thus the ratio of £2 to 15s. is measured

2x20 .... by the traction —^—
Note.

or

8
.

ratio expresses the number of times that one quantity contains another, and therefore every ratio is an abstract quantity.
3.

A

Since by the laws of fractions,

a
b
it
:

=

ma
mJ'
:
;

follows that the ratio a b is equal to the ratio ma mb that is, the value of a ratio remains unaltered if the antecedent and the consequent are multiplied or divided by the same quantity.

H. H. A.

1

2
4.

HIGHER ALGEBRA.

or more ratios may be compared by reducing their Thus suppose equivalent fractions to a common denominator.

Two
_

bx x and — = =— hence h by by y the ratio a b is greater than, equal to, or less than the ratio x y according as ay is greater than, equal to, or less than bx.
_

a

:

o

and x

:

y J are two

ratios.

x-

a aV JNow - = ~
xt

i

,

3

:

:

:

5.

The

ratio of

two fractions can be expressed as a
Ch

ratio

of

two

integers. °

Thus the

ratio


b

:

— a

C

is

measured by the

a
fraction


c

,

or =—
be

:

and

is

therefore equivalent

to

the

ratio

d
ad
:

be.

or both, of the terms of a ratio be a surd quantity, then no two integers can be found which will exactly measure their ratio. Thus the ratio J'2 1 cannot be exactly expressed by any two integers.
6.

If either,

:

If the ratio of any two quantities can be Definition. expressed exactly by the ratio of two integers, the quantities otherwise, they are said to be are said to be commensurable incommensurable.
7.
;

Although we cannot find two integers which will exactly measure the ratio of two incommensurable quantities, we can always find two integers whose ratio differs from that required by as small a quantity as we please.

Thus
and therefore
,
,

.

J5 V= J5 — > mm(>
-.

2-236068...
4

=
,

A1I , -559017...

4

559017

and <

559018
-jooOOOO
;

so that the difference

between the

ratios

559017

:

1000000 and

than -000001. By carrying the decimals further, a closer approximation may be arrived at.

J5

:

4

is less

Ratios are compounded by multiplying toDefinition. gether the fractions which denote them ; or by multiplying together the antecedents for a new antecedent, and the consequents for a new consequent.
8.

Example.

Find the

ratio

compounded
:

of the three ratios
,
:

2a

Sb, Q>ab

:

5c 2 c

a

KATIO.

m The
.

required ratio =
.

,

,.

2a

-x Gab x a 6b be
c --„
1
'

_4a
~ DC
9.

Definition. When the ratio a b is compounded with 2 2 and is called the duplicate ratio itself the resulting ratio is a b 3 3 Similarly a b is called the triplicate ratio of a of a b. b.
: :

,

:

:

:

Also a

2
:

b

" 2

is

called the subduplicate ratio of
(1) (2) (3)
:

a

:

b.

Examples.

The duplicate ratio of 2a 3b is 4a 2 96-. The subduplicate ratio of 49 25 is 7 5. The triplicate ratio of 2x 1 is 8a; 3 1.
: : :

:

:

said to he a ratio of greater inequality, of less inequality, or of equality, according as the antecedent is greater than, less than, or equal to the consequent.
10.

Definition.

A

ratio

is

11.
less
its

of greater inequality is diminished, and a ratio of inequality is increased, by adding the same quantity to both
ratio
..

A

terms.

a , ,, Let T be the
6

ratio,
its

and

,

,

,

let

a+x = be the new ratio formed by J b + x

adding x to both

terms. erms.

Now

a
b

a+x
b

ax — bx

+x

b(b+x) x(a — b)
~b(b + x)
}

and a - b is positive or negative according as a less than b. a +x a y ence it a > b, T > ^ o b + x

is

greater or

H.
«

/,

;

it d.

j

a <b,

a
7-

<

a
b

~t~

x
;

b

+ x

which proves the proposition.
can be proved that a ratio of greater inequality increased, and a ratio of less inequality is diminished, by taking
Similarly
it

is

the

same quantity from both
12.

its terms.

two or more ratios are equal many useful propositions may be proved by introducing a single symbol to

When

denote each of the equal ratios.

1—2

;

4

HIGHER ALGEBRA.
will illustrate

The proof of the following important theorem the method of procedure.

//
,
.

a b
.
.

c

e
f n
11
'

d

1

each of these ratios J

= —
(

+ qc + re + js \pb n + qd n + rt n +

/pan r-

.

.

.\ n
)
,

.

. .

/

where

p, q,

r,

n are any
7

quantities ivhatever.

Ijet

ace = —
—,

b

d

~>—

• •

—&
j

')

J

then

a—

bk,

c

= dk,
n

e

=fk, ...;
re
n

whence

pan =pbn
'''

^

qc"
n

= qd"k
...
'

,

= rf"k",...
n
n

pa" + gc" + re +
H

_pb"k" + qd"k + rf"k +
n fb + qcl
n

pb + qd"+rf+...

+r/i +...

= k";
i

'pa"

+

e
By

2)b"

a c + re" + .\ n _ = k = ^ = -,= d b + qd" + ?'/" + .../
qc"
. .

,

giving different values to p, q, r, n many particular cases of this general proposition may be deduced ; or they may be proved independently by using the same method. For instance,

a

_c

e

b-~d'f-'"
each of these ratios

b+d +f+
:

a result of such frequent utility that the following verbal equivalent should be noticed When a series of fractions are equal, each of them is equal to the sum of all the numerators divided by the sum of all the denominators.
Example
1.

If b

(I

= C =€>

d

,

shew that

J

+ 2c 2 e - Sae 2/ _ ace ~b 4 + 2^/-36/3 ~bdf
az b

Let

_ _^._A,, «-£-£-X;.
6
rf

then

a = bk,

c

= dk, e =fk

;

RATIO.
aa6+2c»g-3qgy _ "" k 4 + 2tl-f - Bbf3

5
+2d?fk* - 3bf 3 k 3 4 + 2r/-/ - 36/8 a
c
e

*

'*

W
fc

...

=
Example
2.

ace

bdf'
If

a

=f=
b

c

,

prove that
3

*2 + a2

#+a
x Let a
then

y2 + &2 y+b
A;
,

+ c2 _ z+c
?/

2

(

.c

+ y +2 2 + (a + & + c) 2 a; + ?/ + 2 + a + &+c
)
2;
;

=r=-=
y
c

it

z

so that x = «£,

= 6/c, = ch
(k*

s a + a3
a:

+a
z

= aW+a* =
-

— +a ah
(ife

L__ ' Jc+1
(&
2

+ —l)a

;

x*+a*
ar

y

a

+y
+
&

ga +e»_

a

+a

?/

+

c

'

/c

+l)o +1

+ l)&

(fc

2

£+1

&

+ l)c +1

Jfc 2 + l)(a + 6 + c) fc+1
Jfc
8

(a+6+c) 8 +(a+ 6+c) a
6

&(a + & + c) + a +
_ (lea

+c

_

+ he) % + (a + b + c) 2 (ka + kb + kc)+a + b + c (x+y+z)*+(a+ b + cf
+
kb

x+y+z+a+b+c
13.

If

an equation

is

homogeneous with respect

to certain

quantities,

we may

for these quantities substitute in the equation

any others proportional to them.
lx
is
3

For

instance, the equation

2 2 y + mxifz + ny z —

homogeneous in
x, y, %
11

x, y, z.

Let

a, j3,

y be three quantities pro-

portional to

respectively.

Put h = — =
a

x

£

75

= y
3

z

,

so that
4

x-

ak,

y=

/3k, z

= yk

;

then
that
is,

Ia f3k

+ ma(3 2 yk* + n^y'k4 =
3
2 + ma/3 2 y + nj3 y = 2

0,
;

7a /?

an equation of the same form as the original a, /?, y in the places of x, y, z respectively.

one, but with

.

.


HIGHER ALGEBRA.
is

6
14.
.

The following theorem
. . i

important.

Q

If

y*
1

,

y~
2

,

.— ,....
3

n

r-

be

unequal fractions, of which the dethe fraction

n

nominators are

all

of the same sign, then

+ a8 + a 3 + b +b 2 +b 3 +
a,
l

...

+ an

•'•

+bn
least

lies

in magnitude between the greatest

and

of them. Let
=*

Suppose that
least fraction,

all

the denominators are positive.
it

be the

and denote
a
b

by k

;

then

— /c
1

'

i

.'.a r
.-.

ko r

>

«

a
y-

>k

:

b
l

a> kb

i

'

a2
ba

> k;
so on;

.-.

a 2 > kb 2

;

and
.*.

by

addition,

a ,+«2

+ «3 +
l

+ an >

(

bl

+b + K +
,
.
'

+K) k
ar
br

'>

a + a 2 + a3 +

+ au

b.+b 2 + b.+ 9 3
1

+b n
that

Similarly

we may prove
l

a + a2 + a 3 +
6
.

+ an

+*.+*

+

+K

<

at

V
all

where

^

is tlie

greatest of the given fractions.

In like manner the theorem may be proved when denominators are negative.
15.
in

the

The ready application

of the general principle involved

Art. 12 is of such great value in all branches of mathematics, that the student should be able to use it with some freedom in any particular case that may arise, without necessarily introducing an auxiliary symbol.

Example
prove that

X
1.

If

-

b

+ c-a
+
c

—=
c

V -

+ a-b

=
a

z
,

+ b-c

x+y+z
a+b

= *&+*)+? (*+*)+ * (*+V)
2(ax + by + cz)

=

;

;

RATIO.

t Each
i

e ,i iof the given e actions = fi

sum °f numerators — __ sum of denominators
x+y+z
a

_
'

+b+c

"
(

''

Again, if we multiply both numerator and denominator of the three given fractions by y + z, z + x, x + y respectively,

each fractions

+ z) \ j + c-a) = (y + z)(b
{l
-

?(« + *>
(z

__.

-

'(* + *
(x + y)

)

+ x)

(c

+ a-b)

(a+b-e)

sum of numerators sum of denominators
=

x

(y

+z)

+ y (z + x ) +z {x + y) 2ax + 2by + 2cz

(2).

.'.

from

(1)

and

(2),

x+y+z
a

_x

(y

+ z)+y
2

+ b + c~

+ x)+z (x + y) (ax + by + cz)
(z

Example

2.

If

l(mb + nc-la)
I

m(nc + la-mb)

n
n

(la

+ mb - nc)

'

prove that

m
y
(cz

x(by + cz-ax)

+ ax-by)
y

z(ax + by -cz)
z

x

We have

I

m
la

mb + nc —

nc

+ la-mb
n

n la + mb —

nc

=m

v z -+'"2/a"

= two

similar expressions

ny + mz _lz + nx b a
Multiply the
y,

_ mx + ly
c

first

of these fractions above
;

and below by

.r,

the second by

and the third by

z

then

nxy + mxz ax

_

Jyz

+ nxy _ mxz + lyz
by
cz

= _2lyz
by

+ cz- ax
similar expressions

= two
I

m
y
(cz

x

(by

+ cz -ax)

+ ax-by)

z

n (ax + by-cz)'

:

8
16.

HIGHER ALGEBRA.
If

equations containing three quantities in the first degree, such as

we have two

unknown
(1),
(2),

a x+b y+
l
l

c
l

z=Q

a2 x +

b 2y

+ c2 z =
;

we cannot
form

solve these completely

but by writing them in the

we

can,

II X by regarding - and - as the unknowns, solve in the

z

z

ordinary

way and
x
%

obtain
b c2
l

- b2c " afi 2 - a 2 b
x - b2c

i '

y
z

__

cxa

"

2

- c2 a

l

>

l

afi 2

-a

b 2

'
l

or,

more symmetrically,
y
x

b c2
x

cla

2

-

c a,
2

afi 2

-

a_p x

'

,(3).

It thus appears that

represented by (1) write down the ratios x y z in terms of the coefficients of the equations by the following rule
:
:

when we have two equations of the type and (2) we may always by the above formula

Write down the coefficients those of y; and repeat these as

of x, y, z in order, beginning with in the diagram.

Multiply the coefficients across in the way indicated by the arrows, remembering that in forming the products any one obtained by descending is positive, and any one obtained by ascending is negative. The three results
h i cz- h fv
c
x

a2-

c

2

an a

A- a b
2

>

are proportional to

x, y, z respectively.

This

is

called the

Rule of Cross Multiplication,

RATIO.
Example
1.

9

Find the

ratios of

x
t

:

y

7x=4y + Qz

from the equations 3z = 12x + Uy.
:

z

By

transposition

we have 7x - Ay - 8-2 = 0, 12x + lly-Sz = 0.

"Write

down

the coeilicients, thus

-4
11

-8 -3

7

-4
11,

12

whence we obtain the products

(-4)x(-3)-llx(-8),
or
•'*

(-8)xl2-(-3)x7,
100,

7 x 11 - 12 x (-4),

-75,
y

125;
z

x x
-

100 ~ ^75~"125'

x,

that

,.
is,

=

-*-

y

4

-3

=z ?
5

.

Example

2.

Eliminate

x, y, z

from the equations
?/

+ ^ 1 + c 12 = a^ + ^y + c^^O
a 1 a;

(1),

(2),

Ogaj+fegy+c^^O

(3).

From

(2)

and

(3),

by cross multiplication,
k> C 3

*__ " Vs

_
C 2«i

y ~ C 3«2


«2 6 3

j*

. '

~ ll ih

denoting each of these ratios by and dividing out by A-, we obtain
Oj

k,

by multiplying up, substituting in

(1),

(Va is

63ca)

+ &i

(^'"3

- c 3 a a) + ('i

A-

" A-)

= °-

This relation
Example
3.

called the eliminant of the given equations.

Solve the equations

ax + by + cz =

(1),
(2),

= hex + cay + abz = (b - c) x+ y+
z

(c-a) (a-b)

(3).

From

(1)

and

(2),

by cross multiplication, x z —y

b-c
.-.

= ^— =
c-a
c),

x = k (b-

T — k, suppose a-b y — k (c - a), z — k(a- b).
:

Substituting in

(3),

k {bc(b-c) + ca

(c

-

a)

+ ab (a -

b)}
\

={b- c)

(c

-

a) {a
{a

-

b),
b)

k{-{b-c)(c^ln'nce

a) {a - &)
.-.

= (b-e) [e - a)
z

-

;

fcss-lj

x = c -b, y — a-r,

= b - a.

10
17.

HIGHER ALGEBRA.
If in Art. 16

we put
ax+
x

z

=
+ +

1,

equations (1) and (2) become

bxy

c
t

= 0,

>

v+
and
(3)

h 2y

c2

becomes
b x c2

x -

y
b 2c l
cx a
a

- c2a
*

i

aj> %

- a2 b

'
]

a
l

b2

-a b

l

afi 2

-a b
2

l

Hence any two simultaneous equations involving two unknowns in the first degree may be solved by the rule of cross
multiplication.

Example.

Solve

5x-3y -1 = 0, x + 2y = 12.
5x - 3y - 1

By

transposition,

= 0,
1

x + 2y -12 = 0;
*'•

x 36 + 2

=

- 1 + 60 ~
38

y

10 + 3

;

whence

x=

= is' y lS'

59

EXAMPLES.
1.

I.

Find the ratio compounded of
(1) (2)

the ratio 2a

:

36,

and the duplicate
:

ratio of 9b 2

:

ab.
:

the subduplicate ratio of 64

9,

and the
,

ratio 27

56.

(3)

the duplicate ratio of

2a
-j:

/6a?
-

M

--

and the

ratio

Sax

:

2by.

2.

If

#+7

:

2 (# + 14) in the duplicate ratio of 5
:

:

8,

find x.

3.

Find two numbers in the ratio of 7

12 so that the greater

exceeds the less by 275.
4.

What number must
it

be added to each term of the ratio 5

:

37

to

make
5.
6.

equal to
:

1

:

3

\

If x

y=3

:

4,

find the ratio of

7x-4y
x
:

:

3x+y.
y.

If 15 (2a-2

- y 2 ) = *7xy,

find the ratio of


RATIO.
7
If
2rt 4 &
2

"

11

?=£ = «
+ 3a -5eV _^__^__ = _
2 2

e

prove that

"

l

8.

If v

= = 6ca

7

,

prove that

a

-j

is

equal

t<»

y
a
9.

If

_

y

q + r-p
(q

r+p-q

p + q-r

shew that
10.

- r) x + (r - p) y + (p - q) z = 0.

If

——

x-z

==-


z

-

=-

y

,

find the ratios of

x

:

y '

:

z.

ii

#

if

y+ z = z+ v == r+ ^ pb + qc pc + qa pa + qb'
'

-

Khew

tliat

2 (*+?+*)

a+o + c

_ (6+o).r+(C +«)y+(« ± i)i
6c + <?a-|-a6

12.

If
.tfS

i'=^ = 2 -,
a
o
c
3
?/

shew

tliat

—„-

+a

3

.r-fa
1<J.

„ t 2

y

3

t- ;,, y* + b 2
-

+6

3

z3 + c 3 _(.y + + *) 3 + (q + &+c) + 2 + c- — (>c+y + \« /_ + o + c)_\s 5) + (a
.,

2

.,

,

i

.

j.

,

II

2y + 2g-.v _ 2g + 2.-c-y —
i

_ 2A-+2y-g — =
2a + 26-c

J

BheW that
14.

26 +

2c-a

=
2c
2

+ 2a")

6

If

shew that
15.

If

prove 1
16.

+y 2 + ^2 = («.v+^ + ^) 2 x a=y b = z c. I (my + rut - Ix) = m (nz + Ix - my) = n (Ix + my - nz\ y+z-x = z+x-y x+y-z = n
(a 2 +6 2 + c2 )
(.i,
: : :

-j

I

m

——
-

Shew

that the eliminant of
bz

ax + cy +
is

= Q,
a3

= 0, bx + </y + c; = 0, + & 3 + c3 -3«6c = 0.
cx + by + az

17.

Eliminate
ctx

x, y, z

from the equations

+ hy + (/z = 0, hx + by-\-fz = 0, gjc+fy+C2=0.

:

12
18.

HIGHER ALGEBRA.
If

x = cy + bz, y = az+cx z=bx + ay,
}

shew that
19.

X z = j—* = y 1 - a \-b \—c
II

2

1

i •> l

n

9 L

Given that a(y + z)=x, b(z + x)=y, c(x+y)=z,
bc

prove that

+ ca + ab + 2abc = l.
:

Solve the following equations
20.

3x-4y + 7z =

0,

21.

x+y=
3x-2y+17z =
x* + 3f + 2z
s

z,

2x-y-2z = 0, to?-f+£=l8.
22.

0,

= l67.

tyz + 3sa?=4an/, 2tys - Sac = 4ry,

23.

3x 2 - 2y 2 + oz 2 = 0,
7a*

- 3y 2 -I5z 2 = 0,
5.0-4^ + 73 = 6.

ff+2y+32=19.
24.

If

.--*

+-^L^
+

*„<>,
n </c+V«
?==-

I

m
Jb+Jo
(b

Ja+Jb
shew that
Solve the equations
25.

+

'

—— =
(a-b)(c-\/ab)

- c)

=
(c

=- a)
(b

(a -

V be)

-

\J ac)

ax + by + cz = 0, bcx + cay + abz = 0, xyz + abc (a3x + b 3y + &z) = 0.
a.-£+&y

26.

+ C2=a 2# + & 2y +

6' 2

2==0,

x + y + z + (b-c)(c-a) (a-b) = 0.
27.

If

a(y+x)=x,
2
-

b(z + x)=y, c(x+y)=z,
?/

prove that 1
28.

— X-be) = b(l-ca) = c(l-ab) —7— — a
2
,

*

s2

{I

If

ax + ky+gz = 0, kx + by + fz^0, gx+fy + cz = 0,

prove that

^
(2)

x2 bc-f 2
(be

y
2
)

2

z2

ca-g 2
{ea -

ab-h 2
)

-f

g

2

(ab

- h 2 ) = (fg - eh)

(gk - af) (A/- bg).

;

.

CHAPTER

II.

PROPORTION,
Definition.
c

18.

When
them

quantities composing
ft

are equal, the four are said to be proportionals. Thus
ratios

two

if

-

= - then
,

a, b, c,

d are
is

proportionals.

This

is

expressed by
is

saying that a

is

to b as c

to d,
:

a
or

b

:

:

and the proportion c d
: :

written

a

:

b

—c

d.

The terms a and d are
19.

called the extremes, b

and

c

the means.

If four quantities are in proportion,
is

the

product of

the

extremes

equal

to the

product of the means.
proportionals.

Let

a, b, c,

d be the

Then by J
whence

definition


b

=.


d

ad =
if

be.

Hence
fourth

any three terms
found.

of

a proportion are given,

the

may be

Thus

if a, c,

d are

given, then b

=

Conversely, if there are any four quantities, a, b, c, d, such that ad = be, then a, b, c, d are proportionals ; a and d being the extremes, b and c the means ; or vice versa.

Quantities are said to be in continued Definition. proportion when the first is to the second, as the second is and so on. Thus to the third, as the third to the fourth are in continued proportion when a, b, c, d,
20.
;

a

bed

b

c

14

HIGHER ALGEBRA.
If three quantities a,
b,

c

are in continued proportion, then
:

a
.-.

b

b

:

c

\

ac

= b2

.

[Art. 18.]

In this case
c
\

and
21.

c is

be a mean proportional between a and said to be a third proportional to a and b.
b is said to
to

three quantities are proportionals the first is third in the duplicate ratio of the first to tJie second.

If

the

Let the three quantities be 1

a. b, c:

then T = -.

be

Now

a

cue
=r =
a
X

a

x -

b

a
6
c

=
:

a

2

b

F,;
b
2
.

that

is,

a

:

= a2

It will be seen that this proposition is the
of duplicate ratio given in Euclid,

same as the

definition

Book
:

v.

22.

If

a

:

b - c

:

d and

e
,

'f=g
e

h,

then will ae

:

bf= eg dh.
:


.b

a
or
F b

=

c -

d

and >=!-:

g

j
ae
•'*

h

=

eg

bf
or ae
:

dh
eg
c

y

bf=
:

:

dh.

Cor.

If

a
b

b

=

:

d,
:

and
then
This
is

:

x=d

v/,

a
the theorem

:

x=

c

:

y-

known

as ex cequali in Geometry.
b,
c,

d form a proportion, many other proportions may be deduced by the properties of fractions. The results of these operations are very useful, and some of them are often quoted by the annexed names borrowed from
23.

If four quantities a,

Geometry.

:; ;

PROPORTION.
(1)

15
c.

If a

:

b

=c

:

d,

then b
1
-f-

:

a=d
-r-

:

[Invertendo.]

For - = b

;

therefore

=-

=1
=
c

-_

d'

b

d'
:

that
or

is

-

a
b
:

a=d
:

:

c.

(2)

If
acZ

.

a
-

:

b=c
;

:

d,

then a

c

=
;

b

:

d.

[Alternando.]

For
that or
(3)
•!
,

be

therefore

—j = —
c

.

a
c

is,

= =

b
-,
:

a
b
:

a
If «
7:

:

d.

6

=

c

:

d, tlien

a+

b

:

b

=

c

+d

:

d.

[Componeudo.']

lor
that
or
(4)
is

=

-,

:

o

d

therefore

s-

+1 =
7

-,

o

d
c

+

1

a+b
o

—=
6

+d —d=—
c

:

a+ b
If

:

=

+d
:

:

d.

a

:

6

=

c

:

d,

then

a-b
7

b

= c- d

:

d.

[Divideudo.]

For
.

=-

= -,
d

:

therefore b

b
,
"

.

that
or

is,

—1 =- - 1 d a— —7— = — -=—
:

bc-d
3

a-b
(5)

b

-

c

-d
:

:

d.

If

«
(3)

:

6

-

c

:

df,

then a

+6 a — b=c+d:c— d.

For by

r = -ja—

and by
.

1

1

/

*

\

(4)
. . .

-j

bc—d
=
c

^-j

.'.

by J

,

« +&
=

+d

division,

or

a+
This proposition
is

b

:

a-b c-d' a-b = c + d c-d.
:

usually quoted as

Componeiuh

a) id JJivi-

dendo.

Several other proportions

may

be proved in a similar way.

:

'

16
24.

HIGHER ALGEBRA.
The

results of the preceding article are the algebraical equivalents of some of the propositions in the fifth book of Euclid, and the student is advised to make himself familiar with them

in their verbal form. follows

For example, dividendo may be quoted as
the excess

When
fourth
25.

there are
to

four proportionals,

of

the first above

the second is

the

second, as the excess of the third above the

is to the fourth.

We

shall

now compare
is

the algebraical definition of pro-

portion with that given in Euclid.
Euclid's definition
as follows
:

Four quantities are

said to be proportionals

when

if

any

equi-

multiples whatever be taken of the first and third, and also any equimultiples wJiatever of the second and fourth, the multiple of the third is greater than, equal to, or less than the multiple of the fourth, according as the multiple of the first is greater than, equal to, or less than the multiple of the second.

In algebraical symbols the definition may be thus stated

:

Four quantities
according as p>a
I.

a,

b,

c,

d

are in proportion

when

p>c

= qd

=

qb,

p and

q being

any positive

integers tcJudever.

To deduce the geometrical
a
-z

definition of proportion

from

the algebraical definition.
Since

- -

c
,

by multiplying both

sides

u by -

,

we

obtain

pa
qb

2)C

qd

hence, from the properties of fractions,

pc = qd according as pa =

qb,

which proves the proposition.
II.

To deduce

the algebraical definition of proportion from

the geometrical definition.

Given that pc = qd according as pa =
a
b

qb, to

prove

=

c
~d'

PROPORTION.
If
is

17

-j-

not equal to -

,

one of them must be the greater.

Suppose

g

>

-^

;

then

it

will be possible to find
<p

some fraction 2

which

lies

between them, q and
b

being positive integers.
-

P
0).

Hence
and

>

p < ?

(1)

(2>

From
from

pa>qb;
2)c<qd\

(2)

and these contradict the hypothesis.
Therefore y and - are not unequal; that
the proposition.
It should be noticed that the geometrical definition of proportion deals with concrete magnitudes, such as lines or areas,
26.
is

- =

-•

which proves

represented geometrically but not referred to any common unit of measurement. So that Euclid's definition is applicable to incommensurable as well as to commensurable quantities ; whereas the algebraical definition, strictly speaking, applies only to commensurable quantities, since it tacitly assumes that a is the same determinate multiple, part, or parts, of b that c is of d. But the proofs which have been given for commensurable quantities will still be true for incommensurables, since the ratio of two incommensurables can always be made to differ from the ratio of two integers by less than any assignable quantity. This lias been shewn in Art. 7 ; it may also be proved more generally as in the

next

article.

Suppose that a and b are incommensurable; divide b into m equal parts each equal to /?, so that b = m/3, where m is a positive integer. Also suppose f3 is contained in a more than n times and less than n+ 1 times;
27.

then
that

i,

nB a (n+1) B - > -^ and < * /^
.

o

mp

imp

,

is,

-=

lies

between

o

m

and

m

;

so that

-j-

differs

from

— by

a quantity less than

.

And

since

we

H. H. A.

2

,

1

'

18

HIGHER ALGEBRA.

can choose B (our unit of measurement) as small as we please,
be made as great as we please.
as

m can

Hence

1 —

can be made as small

we

will

and two integers n and m can be found whose ratio express that of a and b to any required degree of accuracy.
please,

28.

The propositions proved

In solving problems. tions is greatly facilitated by a skilful use of the operations componendo and dividendo.
Example
If
1.

in Art. 23 are often useful in particular, the solution of certain equa-

(2ma + 6mb

+ Snc + 9wtZ) (2ma - Gmb - Snc + 9nd) = (2ma - 6mb + Snc - 9/uZ) (2mm + Gmi - Snc - dnd),

prove that a,b,

.*.

d are proportionals. 2ma + Gmb + Snc + 9nd _ 2ma + Qmb - Snc - 9nd 2ma - bmb + Snc - \)nd 2 ma - 6mb - Snc + 9nd componendo and dividendo, 2 (2ma + Snc) _ 2 {2ma - Sue) 2~{Gmb + 9nd) ~ 2 (Smb - 9m/)
c,

'

Alternando,

2ma + Snc -

=— = -— -— ,. n 2ma-Snc bmb-vna
:

.)nd Gmb + (


;

Again, componendo and dividendo,

Ama _ \2mb
One
a

~

lQnd
b
-,

whence
or

c

=

a
:

a

:

b

—c

d.

Example

2.

Solve the equation

Jx+l + Jx^l _ 4a; -1 2 Jx + l- Jx-1
We have,
componendo and dividendo,

Jx+l _ 4a; +
.r

*'•

+ l_ x-1 "

16a; 2
16a;
2

+

8a;+l
*

-24a; + 9

Again, componendo and dividendo,

2x
2

_ 32a; 2 - 16a; + 10 ~ ~ 32a; - 8

"
whence

X

~
2

16a; 2

-8a; + 5
'

16a;- 4

16a;

-

4a;

= 16a; - 8a; + 5
2

;

5 x = -.

:

c

PROPORTION.

19
II.

EXAMPLES.
1.

Find the fourth proportional to

3, 5, 27.

2.

Find the mean proportional between
(1)

6

and

24,

(2)

36'0a 4
II

and 250a 26 2

.

3.

Find the third proportional to a
:

X
'

-

y
If
4. 5.

-f

x

- and

x
-.

y

b

=c

:

d,
:

prove that

a2 c + ac2

b 2d
:

+ bd2 = (a + c) 3

:

(b
:

+ df.
pc 2 — qd2
.

pa 2 + <?6 2

£>a 2

— qb 2 =pc 2 + qd2
:

.

6.

a-c

:

b-d=*Ja 2 + c 2
:

*Jb

2

+ d2
<

7.

\/a 2 ~+"c2

\/^+d^=jS /ac + ^

:

^Jbd+j.

If a, 6,
8.
9.

c, o?

are in continued proportion, prove that
:

10. 11.

+ ^=03 <?d+d\ 3a-4d=2a 3 + 3b 3 3a3 -46 3 2a + 3(i (a 2 + b 2 + c (b 2 + c2 + d2 = (aft + &c 4- c^) 2
a
:

6

:

:

.

2

)

)

.

If b is a

mean

proportional between a and

c,

prove that

a 2_fr2 + c2

a- 2 -6- 2 + c- 2
12.

If

a

:

6=c

:

d,

and
:

e

:

/=#

:

h t prove that
:

ae + bf

ae-bf=cg + dh

cg-dh.

Solve the equations
13.

2afi-3afi+a;+l 2073-3072-07-1

3^-^ + 507-13
307 3

-072

-507+13*

14.

Zx*+x2 - 2o7 - 3 _ 5o74 + 2o72 -7o7 + 3 ~~ 3ot* - x 2 + 2o; + 3 5o? 4 - 2o7 2 + 7o- - 3
(m-\-n)x — (a- b)

'

15.

(m-n)x — (a + 6)
If a,
&, c, o?

(m + n)x + a + c (wi — n)x + a-c'

16.

are proportionals, prove that
a-

+ d=b + c +

7

.

(a K

— b)(a — c)

^
a

-.

17.

If a, b,

c,

d, e are in

continued proportion, prove that
2

(ab + be + cd + e&) 2 = (a 2 + 6 2 + c 2 + rf 2 ) (6 2 +

+ d + e2
2

).

2—2

20
18.

HIGHER ALGEBRA.
If the work done by x — 1 men in x 2 men in x - 1 days in the ratio of 9

+ 1 days is
:

to the

work done

by x +
19.

10, find x.

Find four proportionals such that the sum of the extremes is 21, the sum of the means 19, and the sum of the squares of all four numbers is 442.

Two casks A and were filled with two kinds of sherry, mixed 20. in the ratio of in the cask A in the ratio of 2 7, and in the cask What quantity must be taken from each to form a mixture 5. 1 shall consist of 2 gallons of one kind and 9 gallons of the other \ which
:

B

B

:

wine; it is then filled with water, then nine gallons of the mixture are drawn, and the cask is again filled with water. If the quantity of wine now in the cask be to the quantity of water in it as 16 to 9, how much does the cask hold?
21.
full of

Nine gallons are drawn from a cask

If four positive quantities are in continued proportion, shew that the difference between the first and last is at least three times as great as the difference between the other two.
22.

In England the population increased 15*9 per cent, between 1871 and 1881; if the town population increased 18 per cent, and the country population 4 per cent., compare the town and country popula23.

tions in 1871.

In a certain country the consumption of tea is five times the consumption of coffee. If a per cent, more tea and b per cent, more coffee were consumed, the aggregate amount consumed would be 1c per cent, more but if b per cent, more tea and a per cent, more coffee were consumed, the aggregate amount consumed would be 3c per cent,
24.
;

more

:

compare a and

b.

Brass is an alloy of copper and zinc bronze is an alloy 25. containing 80 per cent, of copper, 4 of zinc, and 16 of tin. fused mass of brass and bronze is found to contain 74 per cent, of copper, 16 of zinc, and 10 of tin find the ratio of copper to zinc in the composition
;

A

:

of brass.

crew can row a certain course up stream in 84 minutes; 26. they can row the same course down stream in 9 minutes less than they could row it in still water how long would they take to row down with the stream ?
:

A

CHAPTER

III.

VARIATION.
Definition. One quantity A is said to vary directly as another B, when the two quantities depend upon each other in such a manner that if B is changed, A is changed in the same
29.
ratio.

Note.
as B.

The word

directly is often omitted,

and

A

is

said to vary

a train moving at a uniform rate travels 40 miles in 60 minutes, it will travel 20 miles in 30 minutes, 80 miles in 120 minutes, and so on; the distance in each case being increased or diminished in the same ratio as the time. This is expressed by saying that when the velocity is uniform

For instance

:

if

the distance is ptroportional to the time, or the distance varies as the time. 30.

The symbol
is

A

on

B

read

"A

used to denote variation varies as B."
oc

is

\

so

that

31.

If A.

varies as B, tlien

A is equal to B multiplied by some
b,

constant quantity.

For suppose that values of A and B.
mi
i i

a,

a

lt

a„,

a3 ...,

b
x

,

b 2 , b3

...

are corresponding

Inen, by deimition,

•,•

a b — = =«,

V
:

a b — =— ^ K

;

a —
»3

=

b
-r
;

and

,

so on,

K
.

/. s-i
6,

= =* = y^=
62
.

,.j-

63

each being equal to T
b
.

TT

Hence
that
is,

-=

the corresponding value ot

any value r

of

A
_

=—=

B
is

is

always the same

:

— —

7u,

where
.'.

m

constant.

A=mB.

;

22
If

HIGHER ALGEBRA.
any pair
of corresponding values of

A and B
if

the constant

m can be

determined.
3

For instance,

are known, A = 3 when

^=12, we have
and
32.

=m x

12;

A = \B.

One quantity A is said to vary inversely Definition. as another Z?, when A varies directly as the reciprocal of B.
Thus
if

A

varies inversely as B,

A = -^

,

where

m is
:

constant.
If 6

the same work in and so on. Thus it appears that 4 hours, 2 men in 24 hours when the number of men is increased, the time is proportionately decreased; and vice-versa.
hours, 12
;

The following is an do a certain work in 8

illustration of inverse variation

men

men would do

Example 1. The cube root of x x=8 when y = 3, find x when y = l^.

varies inversely as the square of y

;

if

By

supposition £/x=


if

,

where

m is constant.
Tit

Putting x = 8, y = 3, we have
.*.

2
?;t

=n>
= 18,

and

v *jx
a;

=—

18

r
,

hence, by putting y = ^, we obtain

= 512.

Example 2. The square of the time of a planet's revolution varies as the cube of its distance from the Sun; find the time of Venus' revolution, assuming the distances of the Earth and Venus from the Sun to be 91-J and 66 millions of miles respectively.
Let
or

P
;

be the periodic time measured in days,

D

the distance in millions

of miles

we have
some constant.

P aD
2

3
,

P*=kD 3
is

,

where k

For the Earth,

365 x 365 = k x 91± x 91| x 91£,
k=

whence

4x4x4
365

4 x4x4 p2 _ " r ~ 365 "
.

'

.

VARIATION.
For Venus, whence

23
;

pa^i^ili x 66 x 66 x 6G
3 b.)

P = 4x66

::

V

/264
365
approximately,

= 264 x a/*7233, = 264 x -85

= 224-4.
Hence the time
33.
of revolution is nearly 224£ days.

One quantity is said to vary jointly as a number of others, when it varies directly as their product. Thus A varies jointly as B and C, when A = mBC. For instance, the interest on a sum of money varies jointly as the
Definition.
principal, the time,

and the rate per

cent.

3 -i.

Definition.

A

is

said to vary directly as
.

B

and

in-

versely as C,
35.

when A

varies as -^

7/*A varies as
is constant,

B

when C

is constant,

when
vary.

B

then

tvill

A

and

A varies as C
B and C

vary as

BC

ivhen both

B and partly on Suppose these latter variations to take place sepathat of C. rately, each in its turn producing its own effect on A also let a, b, c be certain simultaneous values of A, B, C.
The variation
of

A

depends partly on that of

;

Let C be constant while B changes to b ; then A must undergo a partial change and will assume some intermediate value a\ where
1

"=2.

(1)

Let

B

be constant, that

is,

let it retain its

changes to c ; then intermediate value

A must
a' to its

complete its final value a, where

value b, while C change and pass from its

From
that
or
is,

(1)

and

(2) x
'


a

x

- =
a


b

x

c

:

A = =be

.

BC,

A

varies as

BC.

;

;

;

24
36.

HIGHER ALGEBRA.
The following are
article.

illustrations of the

theorem proved in

the last

work done by a given number of men varies directly as the number of days they work, and the amount of work done in a given time varies directly as the number of men therefore when the number of days and the number of men are both variable, the amount of work will vary as the product of the number of men and the number of days.
The amount
of

Again, in Geometry the area of a triangle varies directly as its base when the height is constant, and directly as the height when the base is constant and when both the height and base are variable, the area varies as the product of the numbers
;

representing the height and the base.

The volume of a right circular cone varies as the square of the the base when the height is constant, and as the height when the radius of base is constant. If the radius of the base is 7 feet and the height 15 feet, the volume is 770 cubic feet ; find the height of a cone whose volume is 132 cubic feet and which stands on a base whose radius is 3 feet.
Example.
Let h and r denote respectively the height ani radius of the base measured in feet also let V be the volume in cubic feet.
;

Then

V=mr

2

h,

where

m is

constant.

By

supposition,

770 — m x

2 7 x 15

;

whence

m=—

22
;

.*.

by substituting

V=

132, r = S,

we

get

132=- xOxft;
—X

22

whence and therefore the height
37.

7i= 14
is

14

feet.

case in

The proposition of Art. 35 can easily be extended to the which the variation of A depends upon that of more than

Further, the variations may be either direct or The principle is interesting because of its frequent ocinverse. For example, in the theory of currence in Physical Science. gases it is found by experiment that the pressure (p) of a gas varies as the "absolute temperature" (t) when its volume (v) is constant, and that the pressure varies inversely as the volume when the temperature is constant ; that is

two

variables.

2? oc

t,

when v

is

constant

, ,

;

VARIATION.

25

and

p
these results

cc

-

,

when

t

is

constant.
that,

v

From

we should expect
the formula

when both

t

and v are

variable,

we should have

p

cc

-

,

or

pv = kt, where k
this is

is

constant
the- case.

and by actual experiment

found to be

Example. The duration of a railway journey varies directly as the distance and inversely as the velocity; the velocity varies directly as the square root of the quantity of coal used per mile, and inversely as the number of carriages in the train. In a journey of 25 miles in half an hour with 18 carriages 10 cwt. of coal is required; how much coal will be consumed in a journey of 21 miles in 28 minutes with 16 carriages?
Let
t be the time expressed in hours, d the distance in miles,

v the velocity in miles per hour, q the quantity of coal in cwt., c the number of carriages.

We have
and

t oc

v
*!l
c

v

oc

whence
or
t

t oc


k is constant.

— —7- where
,

Substituting the values given,
1

we have
;

_ k x 18 x 25 ~ 2 jm
that
is,

k

=
^

25x36"
. .

Hence

t

=

cd —^—T 2o 36 Jq
v/lO
x
t,

Substituting now the values of question, we have

c,

d given in the second part

of the

710x16x21 60" 25x36^2
28

.

'

a

that

••

is,

s/q=—15x28
/

n/10x 16x21

=5^10,

^

,-

whence

q
of coal
is

— -=- = 6|.

Hence the quantity

6|cwt.

26

HIGHER ALGEBRA.

EXAMPLES.
1.

III.

If

x

varies as y,

and

#=8 when y = 15,
and

find

x when y = 10.
find

2.

If

P

varies inversely as Q,

P=7

when # = 3,
and

P when

3.

If the square of

x

varies as the cube of y,
.

x—3 when y = 4,
3
10 C=—

find the value of y

when #=-y-

4.

A

varies as

B

and
i?

C jointly;
i? varies

if

A=2

when # = - and
±Z? and

,

find
5.

when A = 54 and

= 3.
and
as C, then
J.

If .4 varies as C, each vary as C. If J. varies as

\/

AB will

6.

BC, then

Z>

varies inversely as -7

C
.

7.

P

varies directly as

Q and

inversely as

R\

also

P=o ~

2

when

^ = - and
8.
9.

R =—
a'

:

find

(^

when P=a/48 and jR=\/<5.
x2 +y 2
;

If

varies as y, prove that

varies as

x2 -y\

y varies as the sum of two quantities, of which one varies x and the other inversely as x and if y = 6 when x=4, and y = 31 when x = 3 find the equation between x and y.
If directly as
;

If 3/ is equal to the sum of two quantities one of which varies 10. as x directly, and the other as x2 inversely; and if y 19 when x=2, or

=

3

;

find
11.

y

in terms of x.

If

A

the cube of C, and

varies directly as the square root of and inversely as when 24 if 3 when .£=256 and C=2, find

4=

B

B

A=

and

C=g

.

12.

Given that x + y varies as
between x and
z,

z

+z

,

and that x — y varies as
z

z


z

,

find the relation

provided that

=2 when x =3
varies as Z> 2 ,

and

y = \.
and C jointly, while If J. varies as varies as D. varies inversely as A, shew that
13.

B

B

and

C

A

varies as the sum of three quantities of which the first is constant, the second varies as .r, and the third as x 2 and if y = when x=l, y l when x=2, and y 4 when x = 3; find y when x=7.
14.

If

y

;

=

=

When a body falls from rest its distance from the starting 15. point varies as the square of the time it has been falling if a body falls through 402^ feet in 5 seconds, how far does it fall in 10 seconds ? Also how far does it fall in the 10th second?
:

VARIATION.

27

Given that the volume of a sphere varies as the cul>c of its 16. radius, and that when the radius is 3-&- feet the volume is 179rj cubic feet, find the volume when the radius is 1 foot 9 inches.
of a circular disc varies as the square of the radius when the thickness remains the same; it also varies as the thickness when the radius remains the same. Two discs have their thicknesses in the ratio of 9 8 find the ratio of their radii if the weight of the first is twice that of the second.
17.
:

The weight

;

a certain regatta the number of races on each day varied jointly as the number of days from the beginning and end of the regatta up to and including the day in question. On three successive days there were respectively 6, 5 and 3 races. Which days were these, and how long did the regatta last?
18.

At

19.

The

price of a

diamond

varies as the square of its weight.

Three rings of equal weight, each composed of a diamond set in gold, have values «£«., £b, £c> the diamonds in them weighing 3, 4, 5 carats respectively. Shew that the value of a diamond of one carat is

the cost of workmanship being the same for each ring.

persons are awarded pensions in proportion to the square One has served 9 years root of the number of years they have served. If the receives a pensio?i greater by ,£50. longer than the other and length of service of the first had exceeded that of the second by 4| years How long 8. their pensions would have been in the proportion of 9 had they served and what were their respective pensions ?
20.
:

Two

attraction of a planet on its satellites varies directly as the mass (M)of the planet, and inversely as the square of the distance (D) also the square of a satellite's time of revolution varies directly as the distance and inversely as the force of attraction. If m v d v t v d2 £ 2 are simultaneous values of J/, D, T respectively, prove and 2
21.
;

The

m

,

,

,

that

of revolution of that moon of Jupiter whose distance is to the distance of our Moon as 35 31, having given the Earth, and that the that the mass of Jupiter is 343 times that of Moon's period is 27*32 days.

Hence

find the time

:

The consumption of coal by a locomotive varies as the square of the velocity; when the speed is 10 miles an hour the consumption of
22.

coal per hour is 2 tons if the price of coal be 10s. per ton, and the other expenses of the engine be lis. 3c/. an hour, find the least cost of a journey of 100 miles.
:

.

CHAPTER

IV.

ARITHMETICAL PROGRESSION.
Quantities are said to be in Arithmetical Definition. Progression when they increase or decrease by a common dif38.

ference.

Thus each
Progression
:

of

the following
11, 15,

series

forms an Arithmetical

3,

7, 2,

8,
a,

-4, -10,
d,

a+

a + 2d, a + 3d,

The common
the series

by subtracting any term of from that which follows it. In the first of the above
difference is found
is

examples the common difference
the third
39.
it is d.

4

;

in the second it

is

—6

;

in

If

we examine
a,

the series
d,

a+

a+

2d,

a + 3d,

. .

we

notice that in

any term
3 rd 6 th

the coefficient
seiies.

of d

is

always

less

by one

than the number of the term hi the

Thus the

term

is

a + 2d; a + 5d ;
a+ a+
I9d',
(

20 th
and, generally, the
If
?i

term term

is
is is

th p term

p—
d.

\)d.
if
I

n be the number

of terms,

and

denote the

last,

or

th

term,
40.

we have
To find
the

I

=a+

(n — 1)

sum of a number of
first

terms in Arithmetical

Progression.

Let a denote the
the

term,

d the common

difference,

number

of terms.

Also

let I

denote the last

and n term, and s

ARITHMETICAL PROGRESSION.
the required
8

29

sum

;

then
d)

= a+(a +

+

(a

+ 2d) +

...

+

(I

-

2d)

+ (l-d) +

l;

and, by writing the series in the reverse order,
s

= I+

(I

-

d)

+

(I

- 2d) +

...

+ (a + 2d)+

(a + d)

+

a.

Adding together these two
2s

series,

= (a +

l)

+
I),

(a

+

l)

+ (a + l)+

...

to

n terms

= n (a +
•'•

s

= ^(a +
a

l)

(1);

and
.-.

l~a + (n-l)d
s

(2),

= -^{2a + (n-l)d\
last

(3).

three useful formula; (1), in each of these any one of the letters may denote (2), (3) the unknown quantity when the three others are known. For instance, in (1) if we substitute given values for s, n, I, we obtain an equation for finding a ; and similarly in the other formulae. But it is necessary to guard against a too mechanical use of these general formulae, and it will often be found better to solve simple questions by a mental rather than by an actual reference to the requisite formula.
41.
tlie

In

article

we have

;

Example 1. Find the sura of the series 5^, Here the common difference is 1^; hence from
the

GJ, 8,
(3),

to 17 terms.

sum

=

\

2 x

— + 16 *li|

= y (11+20)
17x31 ~2

= 263£.
Example
400
:

2.

find the

The first term of a series is 5, the last 45, and number of terms, and the common difference.
(1)

the

sum

If

n be the number of terms, then from
400 = "
(5

+ 4r>);

whence

n = 10.

;

.

;

30
If

HIGHER ALGEBRA.
d be the

common

difference

45= the
whence
42.
If

16

th

term = 5 + 15d;

d = 2f

any two terms

of

an Arithmetical Progression be

given, the series can be completely determined; for the data furnish two simultaneous equations, the solution of which will give the first term and the common difference.

Example.
23
rd

The 54 th and 4 th terms
first

of

an A. P. are - 61 and 64
difference,
;

;

find the

term.
If

a be the

term, and d the

common
4
th

- 61 = the 54 th term = a + 53d

and
whence we obtain

64 = the

term = a + 3d
;

;

d=

5 -jr, a = Hh.

and the 23 rd term = a + 22d = 16£.

Definition. When three quantities are in Arithmetical Progression the middle one is said to be the arithmetic mean of the other two.
43.

Thus a
44.

is

the arithmetic

mean between a — d and a + d.
mean
betiveen
;

To find
b

the arithmetic

two given quantities.
the arithmetic mean.

Let a and

be the two quantities

A

Then

since a, A, b are in

A. P. we must have b - A = A — a,
difference

each being equal to the

common
a A—

whence
45.

+
2

b

Between two given quantities it is always possible to insert any number of terms such that the whole series thus formed shall be in A. P. and by an extension of the definition in
;

Art. 43, the terms thus inserted are called the arithmetic means.
Example.
Insert 20 arithmetic

means between 4 and

67.
;

to find a series of 22 terms in A.P., of

Including the extremes, the number of terms will be 22 so that we have which 4 is the first and 67 the last.

Let d be the
then

common

difference
;

67 = the 22 nd term = 4 + 21d
is 4, 7, 10, 7, 10, 13,

whence d = S, and the series and the required means are

61, 64, 67 58, 71, 64.

;

;

;

;

ARITHMETICAL PROGRESSION.
46.

31
betiveen

To

insert

a given number of arithmetic means

two

given quantities.

Let a and

b

be the given quantities, n the number of means.

Including the extremes the number of terms will be u + 2 so that we have to find a series of n + 2 terms in A. P., of which a is the first, and b is the last.

Let d be the common difference
then
b

= the

(n

+

2)

th

term

whence

d=

r

:

71+

1

'

and the required means are
a+
Example
their squares
1. is

b

—a
,

n+l
293
;

a

2 (b
H
*

— — n+l

=-'

a)
,

a+

nib — — — a)
*

n+l

_-

'

.

The sum

of three find them.

numbers

in A.P. is 27,

and the sum of
;

numbers are a Hence

Let a be the middle number, d the d, a, a + d.

common
;

difference

then the three

a-d + a + a + d = 27
.-.

whence a = 9, and the three numbers are 9 (9-rf) 2 + 81

d, 9,

$

+ d.

+ (9 + d) 2 = 293;

whence
and the numbers are
term
4, 9, 14.

d=±5;
of the first

w"'

Example 2. Find the sum is 3n - 1.

p terms

of the series whose

By

putting

n=l, and n=p
first

respectively,
last

we obtain

term = 2,

term =3p — 1

.-.

sum=|(2 + 3i>-l)=|(3p + l).

EXAMPLES.
1.

IV.

a.

Sum

2, 3|, 4J,...

to 20 terms.

2.

Sum

49, 44, 39,... to 17 terms.

3.

Sum-,
4

3

2

-, o


7
I

,...

to 19 terms.

.

32
4.

HIGHER ALGEBRA.
Sum
3,

-, If,... to n terms. o

7

5.

6.
7.

Sum Sum Sum

3'75,

35,

3-25,... to

16 terms.

-Tl, -7, -6J,... to 24 terms.
1-3,

-3-1, -7-5,... to 10 terms.

8.

Sum Sum
Sum Sum
n

-.-

6

,

12 3 x/3, -75 »... to 50 terms.
4

3
-j=
,

9.

-tt

,

V5

;-..

to 25 terms.
to 40 terms.

10. 11. 12.

a - 36, 2a - 56, 3a - 76, 2a tt

. . .

6,

4a a,

36,

6a ,...

56,.
,

. .

to
,

n terms.

+6 Sum -£-

,

3a-6 ^

——

to 21 terms.

,,

13.

Insert 19 arithmetic
Insert 17 arithmetic

means between - and — 9|.
means between 3^ and — 41§.
means between 36.17

14. 15.

Insert 18 arithmetic
Insert
as

and
1.

S.v.

16.
17.
18.

arithmetic

means between x2 and
first

Find the sum of the
In an A. P. the
first

n odd numbers.
is 2,

term

the last term 29, the

sum

155;

find the difference.
19.

The sum
5
;

of 15 terms of

an A. P.

is

600,

and the common

differ-

ence

is

find the first term.

20. find the 21.

The

sum

third term of an A. P. of 17 terms.

is 18,

and the seventh term
is 27,

is

30

;

The sum

of three

numbers

in A. P.

and their product

is

504

;

find them.

22.

The sum
is

of three

numbers

in A. P. is 12,

and the sum of

their

cubes
23.
24.

408

;

find them.
series

Find the sum of 15 terms of the

whose nth term

is 4?i4- 1.

Find the sum of 35 terms of the
Find the sum of p terms of the

series

whose p ih term

is

^ + 2.
+ b.

25.
26.

series

whose n th term

is

-

Find the sum of n terms of the series 2 a2 - 1 6a2 -5 3 4a a a a
,

,

,

.

.

;

;

ARITHMETICAL PROGRESSION.
47.

33

In an Arithmetical Progression when s, a, d are given, to determine the values of n we have the quadratic equation
s

= ^ <2a + (n- l)d\

;

when both

roots are positive

and integral there

is

in interpreting the result corresponding to each.

no difficulty In some cases
n.

a suitable interpretation can be given for a negative value of
Example.
taken that the

How many
sum may be

terms of the series G6 ?

-9,

-6, -3,...

must be

Here
that
or
is,

? {-18 + (»-l) 3}=66;

n--ln-U = Q,
(n-ll)(n+4)=0;
.'.

?i=ll or -

4.

If

we take 11 terms

of the series,
6,

we have
21

- 9, the

-

3, 0, 3, 6, 9, 12, 15, 18,

sum
If

of which

is 66.

at the last of these terms and count backwards four terms, the and thus, although the negative solution does not directly answer the question proposed, we are enabled to give it an intelligible meaning, and we see that it answers a question closely connected with that to which the positive solution applies.

we begin

sum

is

also 66;

can justify this interpretation in the general case in the following way.
48.

We

The equation

to determine

n

is

dn 2 + (2a-d)n-2s =

(1).

Since in the case under discussion the roots of this equation have The last opposite signs, let us denote them by n and - n term of the series corresponding to n is
.

l

a + (n l

1

)

d
yi.,

term and count backwards, the common terms is difference must be denoted by - d, and the sum of
if

we

beirin

at

this

|{2
and we
shall

+ »,-!</) +

(»,

-!)(-</)}
6-.

shew that

this is equal to

H. H. A.

3


34
For the expression

'

HIGHER ALGEBRA.

==

?

\

2a + (2n

{

— n — l)dl
2

^ 1
1

2an2 + 2n n 2 d - n 2 (n2 +
x

1)

d

= ^ I 2n n 2 d - (da* - 2a - d .n2 )\
x

= l(4s-2s) = s,
since

— n2

satisfies

dn 2 + (2a — d)

n—

2s

=

0,

and — n n 2
}

is

the

product of the roots of this equation.
the value of n is fractional there is no exact number of terms which corresponds to such a solution.
49.

When

Example.

How many
?

terms of the series 26,

21, 16, ...must

be taken to

amount
Here
that
or
is,

to 71

n ~ {52 +
5)i2

(n-l)(-5)} = 74;
- 57u + 148 = 0,

(n-4)(5n-37) = 0;
.*.

?i

=4

or 1%.

Thus the number
is greater,

of

terms

is 4.

It will

be found that the

sum

of 7

terms

while the

sum

of 8 terms is less than 74.

50.

We
1.
:

add some Miscellaneous Examples.
The sums of n terms of two arithmetic 4« + 27; rind the ratio of their 11 th terms.
series are in the

Example
ratio of
7?t

+l

a„,
"

Let the first term and d 2 respectively.

common

difference of the

two

series be

a v d x and

We have
Now we
obtain

|M^* +
2a 2

£+1
4?i

{n-l)d 2

+ 27
by putting n—21, we x

have

to find the value of

a 2 + l0d 2

tttt', hence, '

2^ + 20^ _ 148 _ 4 "~ 3 2a 2 + 20d 2 ~ 111
thus the required ratio
is

_

4
,

:

3.

Example series whose
1, 3, 5, 7,...
;

2.

If

Su S2

S&...S,, are the

first

terms are 1, 2, 3, 4,... find the value of
#L + <Sf2

sums of n terms of arithmetic and whose common differences are
..

+£3 +. + £„.

ARITHMETICAL PROGRESSION.

We have

S^

{2

+

(n - 1)}

AH+D =n

,

.S>^{2i>+(»-l)(2i ;-l)}= - {(2p-l)n+l};
•.

?i

the required

sum=
=

w

- {(n m
(n

+

l)

+

(3n

+ l) +
~
.
.

(2/>- 1

.

n+

1)}

?l

-

{

+ 3n + 5n +

.2p

-1

.

;/)

+ p)

= ±{n(l + 3 + 5+...21>-l)+p} =r 2
2

1

(»l>

+P)

EXAMPLES.
1.

IV.

b.

Given
?

a=

-2, c?=4 and terms of the

.5

= 100,

find n.

2.

How many

series 12, 16, 20,...

must be taken
first

to

make 208
3.

In an A. P. the third term sixth term is 1 7 find the series.
;

is

four times the

term, and the

4.

The
;

2n

*1

,

31 st

,

and

last

terms of an A. P. are 7j, 5 and

-6j

respectively
5.

find the first
,

term and the number of terms.
terms of an A. P. are 0, - 95 and - 1 25 term and the number of terms.

The 4 th 42nd and
,
;

last

respectively
6.

find the first

arranges to pay off a debt of £3600 by 40 annual instalments which form an arithmetic series. When 30 of the instalments are paid he dies leaving a third of the debt unpaid: find the value of the first instalment.
7.

A man

Between two numbers whose sum
:

means is inserted; the sum number by unity how many means are there The sum of n terms of the series 2, 5, 8.
arithmetic

2£ an even number of of these means exceeds their
is
2

8,... is !>">0

:

find

n.

3—2

36
9.

HIGHER ALGEBRA.

Sum

the series -

r

,

-_

,

,

.— ,

...

to

n

terms.
is

If the sum of 7 find the sum of n terms.
10.

terms

is 49,

and the sum of 17 terms

289,

11.

If the

p

th
,

that
12.

th r ih terms of an A. P. are a, b, c respectively, shew q (q-i')a + (r-p)b+(p-q)c = 0.
,

p

;

find
13.

The sum of p terms of an A. the sum ofp + q terms.

P. is q,

and the sum of q terms and

is

The sum

of four integers in A. P.

is 24,

their product is

945

;

find them.

Divide 20 into four parts which are in A. P., and such that the product of the first and fourth is to the product of the second and third in the ratio of 2 to 3.
14.
15.

m

The p th term

of an A. P.

is q,

and the q th term

is

p

;

find the

tb

term.
16.

How many
If the

terms of the
of

series 9, 12, 15,...

must be taken
find the
tth
?

to

make 306?
17.

sum

n terms
th

of an A. P.

is

2n + 3n 2
is to

,

term.

'in

2

If the sum of 18. 2 ?i , shew that the to
19.

m terms of an A. P. m term is to the n

the

sum

of

n terms as
2n —
1.

th

term as 2m —

1 is to

Prove that the sum of an odd number of terms in A. P. to the middle term multiplied by the number of terms.
20.

is

equal

If 5 = n (pn

-

3) for all values of

n

t

find the

p th term.
;

of terms in an A. P. is even the sum of the odd the even terms 30, and the last term exceeds the first by terms is 24, of 10 1 find the number of terms.
21.

The number

:

22.

There are two sets of numbers each consisting of 3 terms in A.

P.

and the sum of each set is 15. The common difference of the first set is greater by 1 than the common difference of the second set, and the
product of the the numbers.
23.
first set is

to the product of the second set as 7 to 8

:

find

Find the relation between x and y in order that the ,th mean between x and 2y may be the same as the th mean between 2x and y,
?
?-

n means being
24.

inserted in each case.

If the

that

its

sum

for

sum of an A. P. is the same p + q terms is zero.

for

p

as for q terms, shew

CHAPTER

V.

GEOMETRICAL PROGRESSION.
Quantities are said to be in Geometrical Definition. Progression when they increase or decrease by a constant factor.
51.

Thus each
gression
:

of the following series forms a Geometrical Pro12,
1

3,

G,

24,

1

-

I
9'
3
,

3'
,

-I 27'

a, ar,

ar 2 ar

The constant factor is also called the common ratio, and it is found by dividing any term by that which immediately iwecedes
it.

In the

first of
it is

the above examples the
;

common

ratio is 2

;

in

the second

—o

in the third it is

r.

52.

If

we examine

the series
3 ar2 ar
,
,

a, ar,

ai A , r is

we

notice that in

any term

the

index of
is is

always

less

by one

tlian the

number of the term in
the 6
th

the series.

Thus

the 3 rd term

ar

2
;

term

ar

s
;

the 20 th term
and, generally,
If

is

ar

19
;

the

ih p term

is

a?^

-1
.

term,

n be the number we have

of terms,
l

and

if I

denote the

last,

or

n ,h

= ar"~\

When three quantities are in Geometrical Definition. Progression the middle one is called the geometric mean between the other two.
53.

;

;

38 To find

HIGHER ALGEBRA.
the geometric
b

mean

between two given quantities.
;

Let a and

be the two quantities

G

the geometric mean.

Then

since a, G, b are in G. P.,

_G G~ a'
b

each being equal to the

common
.-.

ratio

whence
54.

G 2 = ab; G = Jab.

To

insert

a given number of geometric means between

two given quantities.

Let a and

b

be the given quantities, n the number of means.

In

all

series of

n

there will be n + 2 terms ; so that + 2 terms in G. P., of which a is the
;

we have
first

to find a
b

and

the

last.

Let r be the common ratio
then
b

= the
«r"

(n
+1
;

+

2)

th

term

"

~a'
i

••"©"
Hence the required means
value found in
Example.
(1).


of,
a?- ,...
2

are

ar

n
,

where r has the

Insert 4 geometric

means between 100 and
is

5.
first,

We
sixth.

have to find 6 terms in G. P. of which 160

the

and 5 the

Let r be tbe

common

ratio

tben 5 = the sixth term

= 160?'5
.
*

;

1

'

~32'

whence

r= o'
80, 40, 20, 10.

and the means are

GEOMETRICAL PROGRESSION.
55.

o!)

To find

the

sum of a number

of terms in Geometrical

Progression.

Let a be the first term, r the common Then terms, and s the sum required.
8

ratio,

n the number
l

of

= a + car + ar +
2

n + ar

~2

+ ar"~

;

multiplying every term by
rs

r,

we have
+ ar"~ 2 + ar"" +
1

= ar + ar 2 +
rs
.-.

ar*,

Hence by

subtraction,

— s = ar n — a

;

(r-l)s = a(r"-l);

,..-5fe^a
r

-

(i).

1

Changing the signs in numerator and denominator,

.-?S=*3 1 -r
Note.
for
s.

(2).

It will
(2)

using

be found convenient to remember both forms given above in all cases except when? isj^ositive and greater than 1.
1

Since ar'^ 1

^

1,

the formula

(1)

may
S=

be written

rl-a 7--T

:

a form which

is

sometimes useful.
the series 3 = -2

Example.

Sum

,

-1, -,

3

to 7 terms.

The common

ratio

;

hence by formula

(2)

the

sum

=—

(-23
2187]

II
2

128

I

~3

X

2315

2

128 * 5

403

40
56.

HIGHER ALGEBRA.
Consider the series
n

1,

r,

111
^2
,

~3

,

2
The sum
to

n terms

=

>-

-H
.2From
this

1

27
2

2
result
it

«-i

appears that however many terms be taken the sum of the above series is always less than 2. Also we see that, by making n sufficiently large, we can make the fraction
njr-^i

as

sma U a s we

please.

Thus by taking a
differ

sufficient
little as

number
please

terms the sum can be made to from 2.
of

by as
is

we

In the next
57.

article a

more general case
s

discussed.

From

Art. 55

we have

=

\ -r
1

a
1

ar"
1

—t

—r

'

Suppose r

is is

a proper fraction; then the greater the value of the value of
?•",

n the

smaller

and consequently

ar'

1

of

;

and
of

therefore by

making n

sufficiently large,

we can make

the

sum

n terms

of the series differ

from

^

by as small a quantity as

we

please.

This result

is

usually stated thus

:

the

sum of an

infinite
^
:

number of terms of a decreasing Geometrical Progression
or

is

1

r

more

briefly, the
1.

sum

a
to infinity is

1-r'
is

Example
product
is

Find three numbers in G. P. whose sum
ar; then - x a x ar = 216
r

19,

and whose

216.
a,
;

Denote the numbers by -,
r

hence a = 6, and

the

numbers are r

,

6, 6r.

'

.

.

GEOMETRICAL PROGRESSION.
6
-

41

r
.-.

+ 6 + 6r=19;

6-13r + 6r2 = 0;
r

whence

= - or

3

-

2

Thus the cumbers
Example
the
2.

are 4,

6, 9.

The sum
first

of an infinite

number

of terms in G. P. is 15,

and

sum

of their squares

is

45

;

find the series.
r

Let a denote the
(l

term,

the

common
'
-z

ratio
ci^

;

then the sum of the

terms

is

1

-r

;

and the sum of

their squares is

1

-r*
(1),

Hence

,—=15 1 - r
1

Dividing

(2)

by

(1)

~=
+r

a2 45 _ 72 =
9

(

2)-

(3),

and from

(1)

and
,

l
(3)
z

= 5;

whence r=x and therefore a = 5.
an, Thus +1 series is o, the

2

... — — 20 10 o y
,

,

EXAMPLES.
1.

V.

a.

Sum -,-,-,...
A

112
O
l£,

to 7 terms.

9

2.

Sum

-2, 2^, -3i,... to 6 terms.
3,...

3.

Sum ^t, Sum Sum Sum Sum Sum
Sum
2,

to 8 terms.

4. 5.

-4,

8,...

to 10 terms.

16'2, 5-4, 1-8,... to 7 terms.
1,

6.

5,

25,...

to

p

terms.

7.

3,

-4,
N/3,


o

,...

to 2n terms.

8.

1,
1

3,...

to 12 terms. to 7 terms.

9.

-j/2 v

,

-2,

8
-jr
s '2
,...

:

42
10.

HIGHER ALGEBRA.

Sum -~,

11

3 3, -j,.- to ^ terms.

11.

Insert 3 geometric Insert 5 geometric Insert 6 geometric

means between 2^ and means between 3f and means between 14 and

4

.

12.

40|.
-

13.

64

I —

.

Sum
14.

the following series to infinity |, -1, ?,...
1-665, -1-11, -74,...
3,

15.

-45,

-015, -0005,...

16.
18.

17. 19.

3" 1 3~ 2
,

,

3-',...

v/3,

1,...

7,

N/42,

6,...

20.

The sum
3 terms
;

the

first

of the first 6 terms of a G. P. find the common ratio.

is

9 times the

sum

of

21.

The

fifth

term of a G. P.

is 81,

and the second term

is

24; find

the series.

The sum of a G. P. whose common ratio is 3 is 728, 22. last term is 486 ; find the first term.
In a G. P. the first term is 7, the last 23. 889 find the common ratio.
;

and the

term 448, and the sum

24.

The sum

of three

numbers

in G. P.

is 38,

and their product

is

1728; find them.

The continued product of three numbers in G. P. is 216, and the sum of the product of them in pairs is 156 find the numbers.
25.
;

26.
s

If

p the

sum

denote the sum of the series l+rp + r2p +... of the series 1 — rp + r2p - ... ad inf., prove that

Sp

ad

inf.,

and

/Op

+ Sp == ^*ij'2p'
b, c

27.

If the

p

th
,

th

q

,

that
28.

r th terms of a G. P. be a, a«- r 6 r -*c*-«=l.

respectively, prove

sum

The sum of an infinite number of terms of a G. P. of their cubes is 192 ; find the series.

is 4,

and the
infinite

Recurring decimals furnish a good illustration of Geometrical Progressions.
58.

Example.

Find the value of
•423 =-4232323

"423.

~ io +

4

23 iooo

+

23

iooooo

+
;

~io + IP +

10

5+

'

GEOMETRICAL PROGRESSION.
that,*,

43

-«3_

4

10
_4

+
,

23 /
3 103

1

io V

+
1

io

2

+

1

\

+
io-*

)

10

+

23
10
:: '

_ _1_
10"

4_
1

23
"

100
99

"io" "!^3
~~

+ 10 990
990
rule.

4

_23

_419

"

which agrees with the value found by the usual arithmetical
59.

any recurring decimal to a vulgar fraction may be proved by the method employed in the but it is easier to proceed as follows. last example
rule for reducing
;

The general

Tojind
Let

the value

of a recurring decimal.

denote the figures which do not recur, and suppose them j> in number; let Q denote the recurring period consisting of q figures let D denote the value of the recurring decimal ; then
;

P

D = 'PQQQ
.-.

;

10>xD = P'QQQ

and
therefore,

10T+'

*D = PQ-QQQ
1)

>

by subtraction, (10 p+y - lC)
10" (10'

D = PQ-P;
;

that

is,

D = PQ - P

. ' '

PQ-P D _ (10''- 1)10'''

10"- 1 is a number consisting of q nines; therefore the denominator consists of q nines followed by p ciphers. Hence we have the following rule for reducing a recurring decimal to a
vulgar fraction
:

Now

For
the
the

numerator subtract nonrecurring fgures from
the

the integral the integral

number number

consisting of consisting of

recurring figures ; for the denominator take a number consisting of as many nines as there are recurring jig n n 8 followed by as many ciphers as there are non-recurring figures.

non-recurring

and

r

r ;

44
60.

HIGHER ALGEBRA.
To find
a,

the

sum ofn
r,

terms of the series
r
2
,

(a

+

d)

(a

+ 2d)

(a

+ 3d)

r

3
,

in which each term is the product of corresponding terms in an arithmetic and geometric series.

Denote the sum by

S

;

then
2

S=a+(a + d)r+(a + 2d)r
.-.

+
. .

...

+

(a

+ n~^ld)r"-'
~
l

rS=

ar + (a + d)r2 +
subtraction,
.
. .

.

+(a+ n-2d)r n + (a + n -ld)r n .

By

2 S(l - r) = a + (dr + dr +

+ dr""
,

1

)

-

(a

+ n-lct)
;

n

dr(\-r n ^ =a+
1

~

l

)
=

-

(a v
1

N + n - Id) r „
'

'

•'•

dr(l-r"+ l-r (1-r) 2
a
Write
a

)

_ ( a + n~^\d )

r"
'

T^r
(a

Cor.

S in
+

the form

dr

dr"

+

l-r
then
if

(l-ry~
In this

(l-r)*
small as

n~ld)r\ T^r
we
all the

;

r<l, we can make
can
be
7^

r" as

please by taking
be neglected,

n

sufficiently great.

case,

assuming that
they

terms which

involve r n

made
-r)

so small that

may

we

obtain

-z

1-r

+

(1

r„ for the sum to

infinity. J

We

shall refer

to this point again in Chap.

XXI.

In summing to

infinity series of this class it is usually best to

proceed as in the following example.
Example
1.

If

x <1, sum the
l

series

+ 2ar + 3x 2 + 4x 3 +

to infinity.
;

Let
.-.

S = l + 2a; + 3a: s + 4a s +

xS=

x + 2x* + 3x*+

;

S(l-x) = l + x + x 2 + x*+
1

~l-x

]

a-

I

J

GE0METK1CAL PROGRESSION.
Example
T
2.

45

Sum
»

the series 1

+

-

+

-,

o

o-

+ —+ 0"*
.

. .

.

to n terms.

Let
1

i S== i

^

+i+
1

7 _ _ + 10 +

+

Sn-2 _..
3n - 5

•"•

5*4
,

4

7

5

+
/3

52

+ 53+
3

+ -57^-+
3 \

3n-2
5

,—

3

3n-2
\
"

1

3

A
"

1
f

1 5

1
"

-" 7 1

3n_2

+5«-»j

"~5«~

=

,

1

+

3

*

DN

'>-i
'

3
1

/

=

+

" I (1 5
_

1\ —
J

- -5.-

3«-2

7

~4
•'*

12w + 7 ~ 4 5* .

;

6

~

35 16

12/t+7
16
.

5"- 1

'

EXAMPLES.
1.

V.

b.

Sum

1 4-

2a 4- 3a2 + 4a 3 43
7
77.

. . .

to

n terms.
to infinity.

2.

Sum 1 + - 44

lb

+ ^, + zoo + c^rr. 64 +
4
-3

15

31

. . •

3.

Sum 1 + 3.r + 5d'2 + 7o? + 9.z4 + ... Sum 1 + - 4-,

to infinity.

2

4.

3 2

+

. . .

to

n terms.

5.

Sum 1 + + + Q + 2 4 o
-7

3

5

7

...

to infinity.

6. 7.

Sum l + 3^ + 6 f
l

2

4-10ji>3 4-... to infinity.

Prove that the (n + l) th term of a G. P., of which the first term th term of a G. P. of is a and the third term b, is equal to the (2»+l) which the first term is a and the fifth term 6.

whose first term is a and common ratio r is equal to the sum of n of a G. P. whose first term is b and common ratio r 1 Prove that b is equal to the sum of the first two terms of the first series.
8.

The sum

of

2n terms of a G.

P.

.

46
9.

;

HIGHER ALGEBRA.
Find the sum of the
l

infinite series
2

+ (l + b)r + (l + b + b
of three
4,

)r 2 +

{l

+ b + b 2 + b 3 )r3 +...,
;

r

and
10.

b being proper fractions.

The sum

be multiplied each by find the numbers.
11.

numbers in G. P. is 70 if the two extremes and the mean by 5, the products are in A. P.
infinite G. P. are together equal to 5, the sum of all the terms that follow it; find

The

first
is

two terms of an
3 times

and every term
the
series.

Sum
12. 13.

the following series

:

.r+a,

,v

2

x

(x

+ if)
j

+ 2<x, .r3 + 3a. to u terms. + x2 (x2 + y 2 + a? (a*3 + if) +
..

)

. . .

to

n terms.

14.

«+o
2 3

3« - 2
33

,

ha +
2

— +... to 2p terms.
+ p + -76

15.

16.

454545
7

+ ^2 +
~

3

+ 34 +
-

3

3

35

to mfinity-

72 +

73

74 + 75

+

.

.,

to llifinity-

17.

If a,

b, c,

d be in G. P., prove that (b - cf + (c - a) 2 + (d - b) 2 = {a- d) 2
:
:

.

If the arithmetic mean between a and b is twice as great as the 18. geometric mean, shew that a 6 = 2 + ^/3 2-^3.
19.

Find the sum of n terms of the

series the r th

term of which

is

(2r-f-l)2'\

Find the sum of 2n terms of a series of which every even term is a times the term before it, and every odd term c times the term before it, the first term being unity.
20.

21.
a,

If

Sn Sv

denote the
ratio
JS2 ,
r,

and common
22.

of n terms of a G. P. whose first term find the sum of Slf S3 /8'5 ,.../8 2B _ 1
r

sum

is

,

.

If
first

S3 ,...SP

are the

sums

whose

terms are

1, 2, 3,..,j2,

and

of infinite geometric series, whose common ratios are
respectively,

2' 3' 4

*

'

'

'

^Ti
. . .

prove that
23.

&\

+ S2 + S3 +

+ Sp =f (p + 3).
integer,

If r

<

1

and

positive,

and
wl

m is a positive
2wi

shew that

(2»i

+

l)r

(l-r)<l-r
small

+1
.

Hence shew that nr n

is indefinitely

when n

is indefinitely great.

CHAPTER
HARMON ICAL PROGRESSION.

VI.

THEOREMS CONNECTED WITH

THE PROGRESSIONS.
61.

DEFINITION.

Three quantities
c

a, b, c

are said to be in

Harmonical Progression when - =

o

7

— —

.

c

Any number

of

Progression when monical Progression.
62.

quantities are said to be in Harmonical every three consecutive terms are in Har-

The reciprocals of quantities in Harmonical Progression

are in

A rithmetical

Progression.
b, c

By

definition, if «,

are in Harmonical Progression,

a

a—

b
}

~c^~b^~c'
.'.

a(b —

c)

=

c (a

b),

dividing every term by abc,

1111
c

b

b

a'

which proves the proposition.

Harmonical properties are chiefly interesting because of their importance in Geometry and in the Theory of Sound in Algebra the proposition just proved is the only one of any importance. There is no general formula for the sum of any number of quantities in Harmonical Progression. Questions in H. P. are generally solved by inverting the terms, and making use
63.
:

of the properties of the corresponding A. P.

;

.

48
64.

HIGHER ALGEBRA.
To find
a,

the

harmonic

7tiean

between two given quantities.

Let
then -

b
,

be the two quantities,
T are in b

H their

harmonic mean;

a

,

-~
11

A. P.
1

''11

I
2

a~b

I

I

IV
&'

H~ a
,,

11 +
2ab

a+b
Example.
Insert 40 harmonic

means between

7

and ^
term
is ;

Here 6

is

na term of an A. P. whose the 42

first

let

d be the

common

difference

;

then
6

=^ +

41d ; whence d
are -

= -.
41
--;

Thus the arithmetic means
monic means are
65.
7
3£, 2\,...~.

2
,

3

-

,

and therefore the har-

*

If A, G> II be the arithmetic, geometric,
b,

and harmonic

means between a and

we have proved

A=

a+

b
(!)•

~Y~

G = Jab

(2).

H=^ a+b
_,

(3). v '

Therefore
that
is,

a+ Tr All = —.

b — 2

2ab
.

a+b

T

= ab = G 2
7

~

:

G

is

the geometric

mean between A and
see that
,-z-

//.

From

these results

we

A-

.

b- 2 Jab ~ a+b G=-~-Jab = a + g-^_

:

;

HARMONICA! PROGRESSION.
which
is

49

a and b are positive; therefore the arithmetic mean of any two positive quantities is greater than their geometric mean.
positive
if

Also from the equation G*--A1I, we see that G is intermediate in value between A and 11; and it lias been proved that

A>

G, therefore

harmonic means order of magnitude.
66.

II ; that is, the arithmetic, geometric, and between any tioo positive quantities are in descending

G>

Miscellaneous questions in the Progressions afford scope for skill and ingenuity, the solution being often neatly effected by some special artifice. The student will find the following
hints useful.
1.

If the

the terms of the same common difference as before.
2.

same quantity be added to, or subtracted from, all an A P., the resulting terms will form an A. P. with
[Art. 38.]

terms of an A.P. be multiplied or divided by the same quantity, the resulting terms will form an A. P., but with a new common difference. [Art. 38.]
If all the
3.

same same common
4.

If all the terms of a G.P. be multiplied or divided by the quantity, the resulting terms will form a G.P. with the

ratio as before.
d...

[Art. 51.]

If a,

b, c,

are in G.P., they are also in continued pro-

portion^ since,

by

definition,

a

bed
,

b

c

1
'

r

Conversely, a series of quantities in continued proportion be represented by x, aw, xr'2
in

may
b

Example H. P.

1.

If a

2
,

b

2
,

c

2

are in A. P.,

shew that

b

+

c,

c

+ a, a +

are

By adding
a*

ab

+ ac + bc + c),

to each term,

we

see that
2

+ ab + ac + bc,
{a

b2

+ ba + bc + ac,

c'

+ ca + cb + ab
(c

are in A.P.

that
.-.,

is

+ b)

(a

{b

dividing each term by (a

+ c)(b + a), (c + a) + b)(b + c) (c + a),
.

+ b)

are in A. P.

-.

,

b

+c
b

c

+a

a
a

+b

are

m A. P.
P
4

that

is,

+ c,

c

+ a,

+b

are in H.

H. H. A.

'

50
Example 2. If n terms of an A.
I

HIGHER ALGEBRA.
the last term, d tlie common difference, and s the sum 2 P. be connected by the equation Sds={d + 2l) prove that
,

of

d = 2a.
Since the given relation
is

true for any

number of terms, put n= 1

;

then

a=

l

= s.
,

Hence by
or

substitution,

8ad = {d + 2a) 2
(d-2ay- = 0;
.-.

d — 2a.
s
th

Example

3.

If the

p

th
,

th

q

,

r

th
,

terms of an A. P. are in G.

P.,

shew that

p-

q, q

-

r,

r - s are in G. P.

"With the usual notation

we have
LAlt bb
-

a+

(p-l)d_ a + (g-l )d_a+ (r-l)d

^V(q^lJd-^T¥^l)~d-aT(^l)d
.*.

-

(4)J

'

each of these ratios

+ (p-l)d}-{a+(q-l)d\ _ {a+ (q- 1) ~ {a + (q - 1) d\ - {a+ (r- 1) d) " {a+(r - 1)
{a

- \a + (r- 1) d} d] - {a + {s - 1) d\
d]

=
Hence p 67.

p-q^q-r
q-r
r,

r
r

—8
s

q, q

-

are in G.P.

The numbers
;

1,

natural numbers
the
first

the

n

th

are often referred to as the term of the series is n, and the sum of
2, 3,

n terms

is

- (n +1).

68.

To find

the

sum of

the squares

of

the first

n natural

numbers.

Let the sum be denoted by

S

;

then

£=l

2

+

2

2

+
a

3'

+
2

+n
1

2
.

We have
(

n3 - (n n _ \y _ 3 (w - 2) _ (71 (n 3 2
1
3

l)

= 3n - 3n+

;

and by changing n into n—l,
similarly

= 3(™ - l) 2 - 3(w 2 3 3) = 3(w - 2) - 3(n 3

2)

1)

+ +

1
1

;

2)

;

3 3

-2 =3.3 -3.3+l; -l 3 =3.2 -3.2 + l; -0 =3.1 -3.1 + 1.
3
2 2

3

2

; ;

THE NATURAL NUMBERS.
Hence, by addition,

51

^3 =

3(l

2

+

— 6b

2 + 3 + ...+»')-3(l + 2 + 3 + ...+n) + * 3n(n + l) + it.

2

3

a
.

•.

3aS

=n -n+

-±a


l)

= n(n +
.

1) (n

1 4- ;;);

••

„ S=

n(n+ l)(2n +
6

C9.

To fiml

the

sum of

the

cubes

of the frst

n natural

numbers.

Let the sum be denoted by

S
3

',

then

£=l +2
3

3

+ 33 +
4?i

+n\
;

We

have

4 (n - 1 ) - (?* w _ 2) 4 - ( w (
n*

-

(n

- Gn 2 + 4n - 1 4 3 8 2) = 4 (n - 1 ) - 6 (n - 1) + 4 (n - 1) 3 3)« = 4(ra- 2) - 6 (n- 2) + 4 (n - 2) l)
4

=

1
1

2

3
2
1

4

4

4

-2 = 4.3 -6.3 + 4.3-l; -l = 4.2 -G.2 + 4.2-l; -0 = 4.1 -6.1 + 4.1-1.
4

3

2

4

3

2

4

3

2

Hence, by addition,

.\

w4 = 4#-6(l 2 + 2 2 +...+?i 2) + 4(l +2 + .. + 7i)-n; 4S = n 4 + n + 6(l 2 + 2 2 +...+7t 2 )-±(\+2 + ...+n) = n* + 7i + 7i (n+ 1) (2n+ 1) -2n(7i + 1) = 7i (71 + 1) (?r -7i+\ + 2n+\-2) = 7i (n + 1 ) (?r + n)
;
,

a g _ w'(n + l) *

_f W (n+l) )'
~
(

'

"

*

4

2

J

cz^es of the f7'st n natural 7iumhers is equal to the squa7'e of the siwi of these 7iumbers.
^Ae s?<m o/*

Tims

Me

The formulae

applied to find of the terms of the series
«,

and the two preceding articles may be the sum of the squares, and the sum of the cubes
of this

a+

d,

a+

2d,

4—2

;;

}

:

52
70.

HIGHER ALGEBRA.

In referring to the results we have just proved it will be convenient to introduce a notation which the student will freWe shall denote the quently meet with in Higher Mathematics.
series
1

+

2

+

3

+

.

. .

+ n by

2,n

1* l
3

+ 2* + 3* +

...

+ 2 3 + 3 3 +...

+na by %n* +«8 by 2n

3
;

where 2 placed before a term signifies the sum which that term is the general type.
Example
1.

of all terms of

Sum

the series
1
.

2

+2

.

3

+3

.

4+

.

.

.to

n terms.

The w th term=ra(n+l)=n2 +«; and by writing down each term in a similar form we shall have two columns, one consisting of the first n natural
numbers, and the other of their squares.
.•.

the

sum = 2m2 + 2?i
_w(m+1)
6

(2m + 1)

T
'

n(n +
2
)

l)

n(n+l) j2n+l
n(n + l)(n+2)
3

Example

2.

Sum

to

n terms the
;

series

whose M th term

1_1 is 2'

+ 8m 3 - 6m 2

.

Let the sum be denoted by S then S = 2 2»- 1 + 82>i3 -62n 2
1 _ 2" " 2- 1

+

8m 2

(m
~~

+ 2 4~ 1) _

6m ( m+1)(2m +
6

1)

= 2» - 1 +« (m + 1) {2m (m + 1) - (2m + 1)
= 2' -l + n(n + l)(2n 2 -l).
l

EXAMPLES.
1.

VI.

a.

Find the fourth term
(1)
2, 2,

in each of the following series

(2)
(3)
2.

2J, 21,
2f,

3i,...
3,...

2,

3i,...

Insert two harmonic
Insert four harmonic

means between means between

5

and

11.

3.

2 - and o

2 —
1«3

.

EXAMPLES OX THE
4.

riiOGltESSlOXS.

53

ively,

and 9 l are the geometric and harmonic means, respectbetween two numbers, find them.
If 12
:

If the harmonic mean between two quantities is to their geo5. metric means as 12 to 13, prove that the quantities are in the ratio of 4 to 9.
6.

If a,

b, c

be in H.

P.,
:

a
7.

shew that a—b=a+ c

:

a — c.
and the u lh term be

If the

lh
iii

term of a H.

P. be equal to n,
is

equal to m, prove that the (m + n) th term
8.

equal to

mn

m+n

If the

p

th
,

th
<7
,

that
9.

(<j

-

rth terms of a H. P. be a, b, c respectively, prove r) be + (r — p) ca + (p-q) ub 0.

=

If b

is

the harmonic
1
j-

mean between a and

+ — =-+o —a b a c c
,

111
series

c,

prove that

.

Find the sum of n terms of the
10.

whose n th term
12.

is

3n*-n.

11.

ns +^n.
3" -2".
,

»(»+2).
3
(4' l

13. 16.
(

»2 (2»+3).
If the

14.

15.

+ 2; 2 )-4/i
i

:J
.

i.

P.,

and

on,

difference to the first

(?^+l) th and (r+ l) th terms of an A. P. are in n r are in H. P., shew that the ratio of the common 2

(m+iy\
y

term

in the A. P. is

— n

.

If I, m, n are three numbers in G. P., prove that the first term 17. th th of an A. P. whose £th terms are in H. P. is to the common , and ?i difference as m-\-\ to 1.
,

m

18.

If the

term and
19.

of n terms of a series be the nature of the series.

sum

a + bu + cri2

,

find the

nth

Find the sum of n terms of the series whose n th term 2 4?i(?i 2 +l)-(6>i -fl).
If

is

20.

between any two quantities there be inserted two arithmetic
; ;
: :

two geometric means G ly G2 and two harmonic means means A u A H1 7/2 shew that 6^0',, II1 H.2 = A l + A 2 I^ + IL,.
,

;

21.

If

p

be the
of

and q the

first

n arithmetic means between two numbers, n harmonic means between the same two numbers,
first

of

prove that the value of q cannot
22.

lie

between p and

f

p.
J

that

it

Find the sum of the cubes of the terms of an A. is exactly divisible by the sum of the term-.

P.,

and shew

;

.

;

54)

HIGHER ALGEBRA.

Piles of Shot and Shells.
71.

To find

the

number of
base.

shot

arranged in a complete

pyramid on a square

Suppose that each side of the base contains n shot then the 2 2 number of shot in the lowest layer is n in the next it is (n—l) 2 and so on, up to a single shot at the in the next (n-2)
; \ ; ;

top.
.-.

S^n
=

2

2 + (n-l) 2 + (n-2) +... +

l

n(n+l)(2n + l)
6

72.

To find
the base

pyramid

number of shot arranged in a complete of which is an equilateral triangle.
the
;

Suppose that each side of the base contains n shot

then the

number

of shot in the lowest layer is

n+
xi

(n

-

1)

+ (n 1)

2)
1

,

+ + n)

+

1

j

that

is,

n(n + —V«

2 or - [n

In
the

this result write

n — 1, » — 2,

for n,

and we thus obtain

number

of shot in the 2nd, 3rd,
.-.

layers.

S=i($n* + 2,n) >(n + l)(» + 2) M
number of
is

[Art7a]
arranged in a complete

73.

To find
the base

the

shot

pyramid
Let

of which

a

rectangle.

m and
shot

n be the number
consists of

of shot in the long

and short

side

respectively of the base.

The top layer

a single row of

m—

m — (n — l),

or

n+1

in the next layer the in the next layer the

number

is is

2

(in

— n + 2);

number

3 (in

—n+

3)

and

so

on

;

in the lowest layer the

number

is

n (m — n + n).

;

;

)

PILES OF SHOT
.-.

AND

SHELLS.
...

55

S= (m -01 + 1) + *2(m-n + 2) + 3(w-w + 3) +
= (m - n) (1 + 2 + 3 +
(wi
...

+n(rn-n + n)
+ n2)

+

n) + (l + 2
1)

2

2

+

3

s

+

...

- n) n (n +
2

1 )

+

w (n +

(2n +1)

6
l}

=

n (n + 1 ){3(m-n) + 2n +
l)

_n(n + =
74.

(3m- n+
6

1)
'

To find
the base

the

number of
is

shot arranged in

an incomplete
sides of the

2>yramid

of which

a

rectangle.

Let a and
top layer,

b denote the

number

of shot in the

two

n

the

number

of layers.
of shot is ab
is
is
;

In the top layer the number
in the next layer the

number
number

(a
(a

+ +

1) (6

+

1)

in the next layer the

2) (b

+

2)

\

and so on
in the lowest layer the

number

is

+ n --

1) (b
2
.

+n—

1)

or
.-.

ab + (a + b)(n-l) + ()i-l)

S = abn + (a + 6) % (n- 1) + % (n- Vf ("-l)w(a + 6) (n-l)n(2 .n-l + + =
abn

+

1

2

6
b) (n

=

| {6ab +

3 (a

+

-

1)

+

(w

-

1) (2m

-

1)}.

In numerical examples following method.
75.

it is

generally easier to use the

Example. Find the number of shot in an incomplete square courses, having 12 shot in each side of the top.
If

pile of 16

we

the base,

place on the given pile a square pile having 11 shot in each side of we obtain a complete square pile of 27 courses;
in the complete pile

and number of shot
also
.*.

=

^

= <)'.)30 = 506;

;

[Art. 7

1

.]

number

11 x 12 x 23

of shot in the added pile=
in the incomplete pile =6424.

number of shot

56

HIGHER ALGEBRA.

EXAMPLES.
Find the number of shot in
1.

VI.

b.

2. 3.

A square pile, having 15 shot in each side of the base. A triangular pile, having 18 shot in each side of the base. A rectangular pile, the length and the breadth of the base con-

taining 50 and 28 shot respectively.

and

An incomplete triangular pile, a side of the base having 25 shot, side of the top 14. a
4.
5.

An

incomplete square pile of 27 courses, having 40 shot in each
if

side of the base.

The number of shot in a complete rectangular pile is 24395 6. there are 34 shot in the breadth of the base, how many are there in length ?
7.

;

its

and

of shot in the top layer of a square pile is 169, in the lowest layer is 1089 ; how many shot does the pile contain ?

The number

Find the number of shot in a complete rectangular pile of 8. 15 courses, having 20 shot in the longer side of its base.
Find the number of shot in an incomplete rectangular pile, 9. number of shot in the sides of its upper course being 11 and 18, and the number in the shorter side of its lowest course being 30. What is the number of shot required to complete a rectangular 10.
the
pile
its

having 15 and 6 shot in the longer and shorter upper course?
11.

side, respectively, of

half the

The number of shot in a triangular pile is greater by 150 than number of shot in a square pile, the number of layers in each being the same; find the number of shot in the lowest layer of the tripile.

angular

Find the number of shot in an incomplete square pile of 16 12. courses when the number of shot in the upper course is 1005 less than in the lowest course.
13.

number
14.

Shew that the number of shot in a square pile is one-fourth the of shot in a triangular pile of double the number of courses.
If the

of shot in a triangular pile is to the number of shot in a square pile of double the number of courses as 13 to 175 ; find the number of shot in each pile.

number

The value of a triangular pile of 16 lb. shot is ,£51 if the 15. value of iron be 10s. 6d. per cwt., find the number of shot in the lowest layer.
;

If from a complete square pile of n courses a triangular pile of 16. the same number of courses be formed ; shew that the remaining shot will be just sufficient to form another- triangular pile, and find the number of shot in its side.

.

\

CHAPTER

VII.

SCALES OF NOTATION.

The ordinary numbers with which we are acquainted in Arithmetic are expressed by means of multiples of powers of 10;
76.

for instance

25-2
4705 = 4

x 10
x

+ 5; 3 10 + 7 x 10 2 +

x

10 +

5.

This method of representing numbers is called the common or denary scale of notation, and ten is said to be the radix of the scale. The symbols employed in this system of notation are the nine digits and zero.

manner any number other than ten may be taken as the radix of a scale of notation thus if 7 is the radix, a number 3 expressed by 2453 represents 2x7 + 4x7" + 5x7 + 3; and in
In
like
;

this scale no. digit higher

than 6 can occur.

Again in a scale whose radix is denoted by r the above number 2453 stands for 2r 3 + 4? ,2 + hr + 3. More generally, if in the scale whose radix is r we denote the digits, beginning with that in the units' place, by a a,, a2 ,...aj then the number so formed will be represented by
tt

,

a r +a

n

n

~

1

,r

+a
,

y~~ +

.

.

.

+

a/

2

+

a,r

+ a,
less

where the coefficients a a ,,...«,. are integers, all which any one or more after the first may be zero.

than

r,

of

in this scale the digits are r in number, their values ranging from to r — 1

Hence

The names Binary, Ternary, Quaternary, Quinary, Senary, Septenary, Octenary, Nonary, Denary, Undenarv, and Duodenary
77.

are used to denote the scales corresponding to the values three,... twelve of the radix.

fae»,

58

HIGHER ALGEBRA.
. .

In the undenary, duodenary, scales we shall require symbols to represent the digits which are greater than nine. It is unusual to consider any scale higher than that with radix twelve ; when necessary we shall employ the symbols t, e, T as digits to denote ten eleven and twelve
.

'

'

',

'

'

'.

It

is

especially
'

worthy
',

of notice that in every scale 10 is the
itself.

symbol not for
78.

ten

but for the radix

The ordinary operations of Arithmetic may be performed in any scale but, bearing in mind that the successive powers of the radix are no longer powers of ten, in determining the carrying figures we must not divide by ten, but by the radix of the scale
;

in question.
In the scale of eight subtract 371532 from 530225, and multiply the difference by 27.
1.

Example

530225 371532
136473

136473
27

1226235 275166

4200115
After the first figure of the subtraction, since we cannot Explanation. take 3 from 2 we add 8 thus we have to take 3 from ten, which leaves 7 then 6 from ten, which leaves 4 then 2 from eight which leaves 6 and so on.
;

;

;

;

Again, in multiplying by

7,

we have
2.

3x7 = twenty one = 2x8 + 5;
we
therefore put

down

5

and carry

Next
put down 3 and carry 6

7x7 + 2 = fifty
;

one = 6x8 + 3;
is

and so on, until the multiplication
3

completed.

In the addition,

+ 6 = nine = lx8 + l;
1.

we

therefore put

down

1

and carry
2

Similarly

and and so on.

+ 6 + l = nine=l x 8 + 1; 6 + l + l = eight = lx8 + 0;
2.

Example

Divide 15et20 by 9 in the scale of twelve.
9)15<?£20

lee96...G.

Explanation.

Since 15 = 1 x
8.

T + 5 = seventeen = 1 x9 + 8,
seven = e x 9 + 8
;

we put down
Also 8 x

1

and carry

T + e = one hundred and
down
e

we

therefore put

and carry

8;

and so on.

;

SCALES OF NOTATION.
Example
3.

59

Find the

.square root of

442641 in the scale of seven.

442641(646
34

134 1026 G02

1416112441 12441

EXAMPLES. Vila.
1.

Add

together 23241, 4032, 300421 in the scale of

five.

2. 3.
4. 5.

Find the sum of the nonary numbers 303478, 150732, 2G4305.
Subtract 1732765 from 3673124 in the scale of eight.

From 3^756

take 2e46t2 in the duodenary scale.

Divide the difference between 1131315 and 235143 by 4 in the

scale of six.
6. 7.
8.

Multiply 6431 by 35 in the scale of seven.

Find the product of the nonary numbers 4685, 3483.
Divide 102432 by 36 in the scale of seven.

9.

In the ternary scale subtract 121012 from 11022201, and divide

the result by 1201.
10. 11.

Find the square root of 300114 in the quinary
Find the square of
tttt

scale.

in the scale of eleven.
scale of seven.

12. 13. 14.

Find the G. C. M. of 2541 and 3102 in the

Divide 14332216 by 6541 in the septenary scale.

Subtract 20404020 from 103050301 and find the square root of the result in the octenary scale.
15.
16.

Find the square root of ce^OOl in the scale of twelve.

The
(1)
(2)

following

numbers

are in the scale of six, find
:

by the

ordi-

nary

rules,

without transforming to the denary scale
the G. C. M. of 31141 and 3102

the L. C. M. of 23, 24, 30, 32, 40, 41, 43, 50.

79.

To express a given integral number in any proposed

scale.

Let iV be the given number, and r the radix of the proposed
scale.

Let a u an a 2 ,...a be the required digits by which iV expressed, beginning with that in the units' place; then _1 a n r" + a n — ,r" + ... + ar~ + a.r + a 2
,
t

is

to be

N=

1

10
lt

.

We

have now to find the values of «

,

a,, "_...",,

60
Divide

HIGHER ALGEBRA.

N by

r,

then the remainder
l

is

a nr"
if

+ a71—1 r

n

2

+

...

+a

and the quotient r+a
a
,

is

l

.

If this quotient is divided

by

r,

the remainder

is

the next quotient
so on, until there
is

a a2
i

;

;

and

no further quotient.
, x

the required digits a a ag1 ...an are determined by successive divisions by the radix of the proposed scale.

Thus

all

,

Example

1.

Express the denary number 5213 in the scale of seven.
7)5213
7)7447
7)106.
7 )15.

5

2
1 1

2

5213 = 2x7 4 +lx7 3 + lx7- + 2x7 + 5; Thus and the number required is 21125.

Example

2.

Transform 21125 from
e)21125
e)1244T
~e)Gl3.

scale seven to scale eleven.

t

t

.-.

the required

number

is 3t0t.
first line

Explanation.

In the

of

work

21 = 2x7+l=fifteen = lx<?
therefore on dividing

+ 4;
4.

by

e

we put down
nine = 2 x
7
e
;

1

and carry
;

Next
therefore

4x7 + 1 = twenty

+7

we put down 2 and carry

and

so on.
to scale ten by scale twelve.

scale ten,

Example 3. Reduce 7215 from scale twelve and verify the result by working in the
r

working in

7215

f)7215
«)874.
t)t^.
t)10.
1.

1
1

JL2
In scale
of ten

80
-

In scale
of twelve
.4
.2

12

1033 12
1 12401

J

Thus the
scale ten.

result is 12401 in each case.

1 x 12 + 5 in calculation is most readily effected by writing this expression in the form [{(7 x 12 + 2) } x 12 + 1] x 12 + 5 thus we multiply 7 by 12, and add 2 to the product; then we multiply 86 by 12 and add 1 to the product; then 1033 by 12 and add 5 to the product.

Explanation.

7215 in scale twelve means 7 x 123 + 2 x 12 2 +
;

The

;

SCALES OF NOTATION.
80.

f>l

Hitherto we have only discussed whole numbers; but fractions may also be expressed in any scale of notation thus
;

•25 in scale ten denotes

5 2 —+— 10*' 10
,

2

•25 in scale six denotes

-=

+

^ G
.

5

;

•25 in scale r denotes


2

h

r

5 — r

Fractions thus expressed in a form analogous to that of ordinary decimal fractions are called radix-fractions, and the point The general type of such fractions in is called the radix-point. scale r is

~
r

r

3

~

r

3

~

>

where 6 b 2 6 a ... are or more may be zero.
, , ,

integers, all less

than

r,

of

which any one

81.

To

express

a given radix fraction in any proposed

scale.

Let
scale.

F

be the given fraction, and r the radix of the proposed
b b 3 ,...

Let b then left
;

,

,

be the required digits beginning from the
b b ^ F J-X + A+ r3+ r r

We

have now to find the values of

6 p b2

,

63 ,

Multiply both sides of the equation by r

;

then

rF=b+Hence
l

2

+ -l+

h

;

b is equal to the integral part of

rF

;

and,

if

we denote

the fractional part by

F

x

,

we have

H-i. + J +
the integral part of rF ; and similarly by successive multiplications by r, each of the digits may be found, and the fraction expressed in the pro posed scale.

Multiply again by r\ then, as before, b
x

is

4

62
If

HIGHER ALGEBRA.
in

the successive multiplications by r any one of the products is an integer the process terminates at this stage, and the given fraction can be expressed by a finite number of digits. & But if none of the products is an integer the process will never terminate, and in this case the digits recur, forming a radixfraction analogous to a recurring decimal.
13

Example

1.

Express

-^ as

a radix fraction in scale

six.

13

16
7

7. x6= -8- = 4 + 8'
ft

13x3

a

7x3
lx3

Kj.

1

1

a

I..

1

^x6 = 3.
.-.

the required fraction

=g + ^+ p + = •4513.
and the

4

5

13
Qi

Example

2.

Transform 16064-24 from scale eight

to scale five.

We must

treat the integral

fractional parts separately,
'24

5 )16064 5 )2644 ...
5)440...

5
1*44

5)71. ..3
5)13... 2

J>_
2-64
£_

2...1

4-04
5_ 0-24

After this the digits in the fractional part recur; hence the required

number
82.

is

212340-1240.

In any scale of notation of which the radix is r, the sum of the digits of any whole number divided by r - 1 will leave the same remainder as the whole number divided by r — 1.
a n the digits begindenote the number, a a lt a 2 Let ning with that in the units' place, and S the sum of the digits; then ~ = a Q + a r + a2r 2 + + a„_/ 1 + arn
,
,

N

N
1

,

x

;

S=a
r .tf-S=a

+ a n _ + an + a +a 2 + (r-l) + a 2 (r°--l)+ + •„_, (i*- - 1) + «, (f - 1).
x

l

SCALES OF NOTATION.

03
divisible

Now

every term on the right hand side

is

by r —

1

iV-S
* . .

rr

=1

— an integer y 6
;

that

is,

-

-=/ +
1

r

-

'

1

when; /

is

some integer

j

which proves

tlie

proposition.

Hence a number in scale r will be sum of its digits is divisible by r — 1.
83.

divisible

by

?•

1

when

the

taking ?-=10 we learn from the above proposition that a number divided by 9 will leave the same remainder as the sum of its digits divided by 9. The rule known as " casting out the nines " for testing the accuracy of multiplication is founded

By

on

this property.

The

rule

may

be thus explained

:

Let two numbers be represented by da + their product by P; then

b

and 9c

-f

d,

and

P^Slac + %c + 9ad + bd.
Hence — has the same remainder
s?nn
as -^
;

and therefore the

of the digits of /*, when divided by 9, gives the same remainder as the sum of the digits of bd, when divided by 9. If on trial this should not be the case, the multiplication must have been incorrectly performed. In practice b and d are readily found from the sums of the digits of the two numbers to be
multiplied together.
Example.

Can the product

of 31256

and 8127 be 263395312

?

The sums of the digits of the multiplicand, multiplier, and product are 17, 21, and 31 respectively; again, the sums of the digits of these three numbers = 8x3 = 24, which has 6 for the sum of the are 8, 3, and 7, whence digits; thus we have two different remainders, 6 and 7, and the multiplication

M

is incorrect.

denote the scale of'r, and the difference, supposed positive, between the sums of the digit* in the odd and the even places; then — or + is a multiple, of
84.

If

N

denote

any number in

D

N D

N D

r+

1.

64
Let a
,

HIGHER ALGEBRA.
«!,

a

,

a n denote the
a,r

digits beginning
-1
,r"
2

with that

in the units' place; then

JV= a +
.-.
A

+ a r 2 + ar* +
...=<*,

+ a
(r

+a

r".

+ a 3 (r3 + 1) + ...; n and the last term on the right will be aw (r"+l) or a n (r — 1) Thus every term on the right is according as n is odd or even. hence divisible by r + I
2

Ar -a + a -a + a3 -

(r+1) + « 2

-

1)

;

^
NOW
.'.

!

r

+
}

%
1

i

'-

= an

mteo-er. °

a0~ a +fl 2~
=— is

CC

3

+

—^D)

an integer;

which proves the proposition.
Cor.

the

sum
1.

of the digits in the even places is equal to = 0, and is divisible of the digits in the odd places, If the

sum

D

N

by r +

Example whose radix

1. is

Prove that 4 •41 is a square greater than 4.
;

number

in

any

scale of notation

Let r be the radix

then
4-41

=4+- +
r

rz

i=(2 + -Y;
\

rj

thus the given number

is

the square of 2*1.
is

Example

2.

In what scale
then
2+ -

the denary

number 2-4375 represented by

213?
Let r be the scale
;

13 + 2=2-4375=2-^;
7 r
r-

16
;

whence
that
is,

"

7r 2 -

16r-48 =

(7r+12)(/--4) = 0.
is 4.

Hence the radix

Sometimes
Example by 101215 ?
;

it is

best to use the following method.

3.

In what scale will the nonary number 25607 be expressed

The required scale must be less than 9, since the new number appears the greater also it must be greater than 5 therefore the required scale must be 6, 7, or 8; and by trial we find that it is 7.
;

SCALES OF NOTATION.

65

Example 4. By working in the duodenary scale, find the height of a rectangular solid whose volume is 364 cub. ft. 1048 cub. in., and the area of whose base is 46 sq. ft. 8 sq. in.

The volume
264-734 cub.
ft.

is

364-^i? cub.
sq.
ft.,

ft.,

which expressed in the

scale of twelve is

The

area

is

46^ 4

We

have therefore to

which expressed in the scale of twelve is divide 264*734 by St-OS in the scale of twelve.
3*08)26473-4(7-e

3<-08.

22*48 36274 36274

Thus the height

is 7ft. lliu.

EXAMPLES.
1.

VII.

b.

Express 4954 in the scale of seven.
Express 624 in the scale of
five.

2.
3. 4.
5. 6.

Express 206 in the binary

scale.

Express 1458 in the scale of three.

Express 5381 in powers of nine.

Transform 212231 from soale four to scale Transform 6£12 from

five.

7.
8. 9.

Express the duodenary number 398e in powers of
scale twelve to scale eleven.

10.

Transform 213014 from the senary to the nonary
Transform 23861 from
scale nine to scale eight.

scale.

10.

11. 12. 13. 14. 15.
16.

Transform 400803 from the nonary to the quinary

scale.
12.

Express the septenary number 20665152 in powers of

Transform
Express

ttteee

from scale twelve to the common

scale.

3 — as a radix fraction in the septenary scale.

Transform 17 "15625 from scale ten to scale twelve. Transform 200 "2 11 from the ternary to the nonary

scale.
scale.

17. 18.

Transform 71*03 from the duodenary to the octenary
Express the septenary fraction

1552 —— as a denary vulgar fraction

in its lowest terms.
19.
20.
21.

Find the value of
In what scale
In what scale
is
is

and of '42 in the scale of seven. the denary number 182 denoted by 222?
*4

the denary fraction

25 -—

denoted by -0302?
5

H. H. A.

66
22.

HIGHER ALGEBRA.
Find the radix of the scale in which 554 represents the square

of 24.

In what scale is 511197 denoted by 1746335 ? Find the radix of the scale in which the numbers denoted by 24. 479, 698, 907 are in arithmetical progression. In what scale are the radix-fractions *16, "20, '28 in geometric 25.
23.

progression ?
26.

The number 212542
Shew that
148'84
is

is

in the scale of six; in

what

scale will

it

be denoted by 17486?
27.

a perfect square in every scale in which the
is

radix
is

is

greater than eight.

28.

Shew that 1234321
;

greater than 4
digits.

a perfect square in any scale whose radix and that the square root is always expressed by the

same four
29.

Prove that 1-331 is a perfect cube in any scale whose radix is greater than three. Find which of the weights 1, 2, 4, 8, 16,... lbs. must be used to 30. weigh one ton. Find which of the weights 1, 3, 9, 27, 81,... lbs. must be used 31. to weigh ten thousand lbs., not more than one of each kind being used but in either scale that is necessary.
32.

radix

is

Shew that 1367631 greater than seven.

is

a perfect cube in every scale in which the

Prove that in the ordinary scale a number will be divisible by 8 if the number formed by its last three digits is divisible by eight. Prove that the square of rrrr in the scale of s is rm^OOOl, where 34.
33.
q, r, s

any three consecutive integers. and a new number N' be taken in the scale If any number 35. formed by altering the order of its digits in any way, shew that the be and N' is divisible by r — 1. difference between
are

N

?*,

N

If a number has an even number of digits, shew that it is 36. divisible by r+1 if the digits equidistant from each end are the same.
37.
JV,

If in the ordinary scale
>S'

St

be the

sum

of the digits of a

number

and 3#2 be the sum of the digits of the number 3iV, prove that the is a multiple of 3. difference between aS^ and 2 Shew that in the ordinary scale any number formed by 38. writing down three digits and then repeating them in the same order is a multiple of 7, 11, and 13.
In a scale whose radix is odd, 39. digits of any number will be odd if the
the

shew that the sum of the number be odd, and even if

number be even. If n be odd, and a number in the denary scale be formed 40. by writing down n digits and then repeating them in the same order, shew that it will be divisible by the number formed by the n digits, and also by 9090... 9091 containing n- \ digits.

+

CHAPTER

VIII.

SURDS AND IMAGINARY QUANT1T1KS.
85.

In the Elementary Algebra, Art. 272,

it is -rr

proved that
r-

the denominator of any expression of the form
rationalised

Jb + Jc
by Jb — Jc, the surd conjugate
to the denominator.

can be

by multiplying the numerator and the denominator

Similarly, in the case of a fraction of the form

where the denominator involves three quadratic two operations render that denominator rational.
For,
first

Jb + Jc + Jd surds, we may by
'

multiply both numerator and denominator by Jb + Jc — Jd; the denominator becomes (Jb + Jc) 2 — (Jd) 2 or Then multiply both numerator and denominator b + c- d + 2 be.

J

by (b + c - d) — 2 J be; the denominator becomes which is a rational quantity.
Example.
Simplify
12

(b

+

c

- d) 2 — Abe,

3+^/5-2^/2

The expression me

- 12 _

(

3

+ y/5 + V*)

(3+ ^ 5)S _
6

(V2 )»
)

^ 12(3 + ^/5 + 2^/2
+ 6^5
2 (3

v/5+ 2^/2)^5-1)

U/5+l)U/5-l)

2+V5+V10-V2
= 1+^5+^/10-^2.

5—2

68
86.

HIGHER ALGEBRA.
To find
I.

the factor

which will rationalise any given bino-

mial surd.

Case
xn and y
n

Suppose the given surd
x,

is

£]a

- $b. p and
q
;

Let ZJa =

?Jb = y, and let are both rational.

n be the

l.c.m. of

then

Now

n xn — y is divisible by x - y for all values of n, and ~ n ~' + y"" ). +x n 2y + xn Sf + ar- _ y" = (x-y) {x
1

Thus the

rationalising factor
1

is

n y + x ~y + n and the rational product is x — y'\
.X'""

+ xn

~2

+ y"~

l ;

Case
Let
(1)

II.

Suppose the given surd

is

p

Ja +

fjb.

x, y,

n have the same meanings

as before; then

n is even, xn — y" is divisible by x + y, and ~ n~ n n x - y = (x + y) (xn - x 2y + + xf~* - y""
If
l

1

).

Thus the

rationalising factor

is

ur -ary +
l

+ ay"-'- 3

r

1

;

and the rational product
(2)

is

x"
is
x

—y

n
.

If

n

is

odd, x"
(x

+

y"
n~

divisible

by x + y, and

n n x +y =

+

y) (x

- xn Sj +
is

- xy n ~ 2 + y n ~').
-2

Thus the

rationalising factor

x

n~1

-x y+
is

n~2

-xy n
+y
n
.

+ y"-

1

;

and the rational product
Example
i
,

x"

1.

Find the factor which
i
;

will rationalise ^/S

+ ^/5.

Let x = 3 2 y = 5 5

then x b and y 6 are both rational, and
;

x e - y 6 — (x + y) (x 5 - x*y + xhj 2 - xhj 3 + xy i - y 5 )
thus, substituting for x
32

541 3223 53+
32
.

and

y,

the required factor
32
53

is
5

14
.

.

32

.

53+ 32

53

5

53,

5

13
55

2

14

or

32-9.
is

+
6

32~.
6

53-15 + 32. 53-5 5

;

and the rational product

32

- 5 s = 3 3 - 5 2 = 2.
"

i

V
69

SURDS AND IMAGINARY QUANTITIES.
Example
2.

Express

(&+&) *

\5 5 -9s )

as an equivalent fraction with a rational denominator.
i
i

i

To
34

rationalise the denominator,
;

which

is

equal to 5" -3*, put 5 2 = x,

=y

then since

x4
is

-y A = [x 3 2

y) (x*
1
.

+ xhj + xy 2 + if')

12
+ 5]
4
.

3

the required factor

5-

+ 52
4

3-*

3*

+ 3~4

;

and the

rational

denominator
/
i

is

52
/ 3

- 3-* = 5 2 - 3 = 22.
2
.

i\
'-

1

12
3~ 4
.
'

3\

.•.

+Vl

the expression

=

5 V5

2 + 3 V—\& + 5 — 3^ + 5-

+3

'

4
_

3
.

1

2

2

13
+3

4
t

52 + 2

5 2 .3 j

+ 2.

5 .3 4"+2.5 5 .3 j
,J

3

1
.

113
5.3 2 +5 2 .3 j
11

22

_l l + o 2

3j +

87.

We

have shewn in the Elementary Algebra, Art. 277,

how to find the square root of a binomial quadratic surd. We may sometimes extract the square root of an expression containing more than two quadratic surds, such as a + Jb + Jc 4- Jd.
Assume
.'.

Ja + Jb + Jc + Jd = Jx + Jy + Jz
2

j

a+

Jb + Jc + Jd = x + y + z +
2

Jxy +

2

Jxz +

2

Jyz.

If then

Jxy =

Jb, 2

Jxz =

Jc, 2

Jyz = Jd,
satisfy

and if, at the same time, the values of x, y, z thus found x + y + z = a, we shall have obtained the required root.
Example.

Find the square root

of 21

- 4^/5 + 8^/3 - 4^/15.
\

Assume
.'.

V 21 - V 5 + V 3 - V 15 = slx + Jy - slz

21 -

4^5 + 8^/3 - 4J15 = x + y + z + 2 Jxy- 2 Jxz - 2 Jyz.

Put
by multiplication,

2jxy = 8JS, 2jxl = 4J15, 2jyz = ±Jo\
xyz = 240
;

that

whence

it

follows that

,Jx = 2j3,

Jxyz=4 s/lo Jy = 2, „Jz = s/5.
is

;

And since these values root is 2^3 + 2-^/5.

satisfy the equation

x+y+z

= 21,

the required

;;

,

70
88.

HIGHER ALGEBRA.
If J a, +

Jh = x + Jy,
+

then

ivill

J

a,

— Jh = x — Jy,

For,

by cubing, we obtain
ci

Jb=x

3

+ 3x 2 ^/y + 3xy + y Jy.

Equating rational and irrational parts, we have

a = x3 + 3xy, Jb
.'.

= 3x 2 Jy + y Jy a- Jb = x3 - 3x2 Jy + 3xy -y Jy;
\

that

is,

J a - Jb = x— Jy.
by the help
if

Similarly,
it

of the

Binomial Theorem, Chap. XIII.

may

be proved that

Ja + Jb = x + Jy,
where n
89.
is

then

Ja - Jb = x- Jy,

any

positive integer.

the following method the cube root of an expression of the form a ± Jb may sometimes be found.

By

Suppose
then
.-.

Ja + Jb = x + Jy

Ja- Jb = x- Jy.
Jtf^b=x -y
2

(1).

Again, as in the last

article,
(2).

3 a = x + 3xy

The values
In
(2)

of

x and y have to be determined from

(1)

and

(2).

(1)

we
is,

suppose that obtain

Ja

2

— b=c; then by
c)

substituting for y in

a = x3 + 3x (x 2 —
that

kx 3 — 3cx —

a.

If
trial,

from

this equation the value of
is

the value of y

x can be determined by obtained from y = x 2 — c.

do not here assume sjx + sly for the cube root, as in the extraction of the square root; for with this assumption, on cubing we should have

Note.

We

a + Jb = xjx

+ Sxjy + Syjx + yjy
hand

y

and since every term on the right rational and irrational parts.

side is irrational

we cannot equate

,

1

;

SUKDS AND IMAGINARY QUANTITIES.
Example.

71

Find the cube root of 72 - 32 x/5.
sf 72

Assume
then

-'62^5 = x - ^/y

;

^72 + 32 s/5 = x + s/y.
multiplication
,

By
that
is,

^5184 - 1024

x 5 = a; 2 - y

;

± = x'--y
72 - 32^/5 = .c 3 - fkcPJy

(1).

Again

+ Sxy - y^'y

;

whence

72 = x
(1)

3

+ 3.t//

(2).

From
that
is,

and

(2)

72 = x
ar
}

:i

+ Sx (x- - 4)

-3x = 18.
and the cube root
is

By

trial,

we

find that x = S; hence y = o,

3-^/5.

the binomial whose cube root we are seekhi" consists of two quadratic surds, we proceed as follows.
90.

When

Example.

Find the cube root of 9 N/3 + ll v/2.

By

proceeding as in the last

article,

we

find that

.-.

the required cube root

=*J3

(

1

+

*

/J

= v/3+v/2.
91.

We
1.

add a few harder examples in

surds.
4

'Example

Express with rational denominator

V9-^3 N

+

r

The expression

=—
fj

4
^

3 - 33

+

(J + l)
l3 3 + l) (3 3

-3 +
ri

l)

iM±i]-3 ~d
3

3

+1

+1 +l

"

:

72
Example
2.

HIGHER ALGEBRA.
Find the square root of

l(x-l) + j2x -7x-i.
<i

f

The expression = \

{3x - 3

+ 2 J(2x + l)(x-4)

}

a

= ±{(2x + l) + {x-±) + 2jc2x + l){x-£)};
hence, by inspection, the square root
is

Example

3.

Given

^5 = 2-23607,

find the value of
(5

J2 + J7- 3J5'
Multiplying numerator and denominator by >J2,
the expression
2

^6-2^/5

+ ^14-6^/5
n/5-1

2

+ 3-^/5

EXAMPLES. Villa.
Express as equivalent fractions with rational denominator
1
i L-

1

+ V2-V3'
1

2 A

^
J2+J3-J5'

3.

-

sfa + s/b + s/a + b'

=

.

4
fi

^^
i i

*Ja-l-\/2a + *Ja + l
(j3 + x/5)(j5 + ^/2)

^10 + ^5-^/3
Find a factor which
7.

will rationalise
^/5

#3 -a/2.
N/3-l.
3

8.

+ ^/2.

9.

06+6*.

10.

11.

2

+ 4/7.

12.

4/5-^3.

=

,

SUKDS AND IMAGINARY QUANTITIES.
Express with rational denominator:

73

16

*/3

17

v 8 + ^4

W

Find the square root
19.

of
20. 22.

16-2 N /20-2
G

v/28 + 2 N /l3.->.

21. 23.
24.

+ ,/12-^24-,/8.

24+4^15-4^21-2^35. 5- x /10- N /15 + N /G.

a+36+4+4^/a-4^6-2V3oS
21+3 N /8 - 6 N /3 - Jl - v'24
(5

- N /56 + 2 N /21.

Find the eube root of
25.

28.

10+6 JZ. 38^14-100^2.

26.

38 + 17^5.

27.
30.

99-70^/2.

29.

54^3 + 41^5.

135^3-87^6.

Find the square root of
31.

a + x + \J%ax + x2
l

.

32.
i

2a - \/3a 2 - 2ab - b'2
i

.

33.
35.

+ « 2 + (l+a 2 + a 4 ) 2
a

.

34.

l+(l-« 2 )" 2

.

If

= ——j-

,

6

= —-i—
S

,

find the value of

7«2 +

1 1

ab -

lb'

2
.

36.

If

x

jl'jl

,

y = /|z7^

find the value of 3<t 2
"

»

~ 5j y + 3^-

Find the value of

V26-15J3 5V2-V38T573'
1

V33-19 N
"
1

/

6 + 2V3 /32
_>

39.

(28

- 10 N/3) - (7 + 4 v/.3)
2

2
.

40.

(26 + 15 s fzf

- (26 +

15 N/3)

3

41.

Given s/b = 2-23607, find the value of lOx/2 N /10 + N/18
'

N/18
42.
43.

- a/3+V5

N /8

+ V3 - V5
*/2.
.

Divide

x*+ 1 + 3# #2 by * - 1 +
(b 2

Find the cube root of 9a6 2 +
Evaluate
V^'
2

+ 24a 2 ) ^6* -3a8

44.

-1
,

when

2.r;

= Ja+ -i\'"

x-s/x*-\

74

HIGHER ALGEBRA.

Imaginary Quantities.
Although from the rule of signs it is evident that a negative quantity cannot have a real square rootlet imaginary quantities represented by symbols of the form J- a, J- 1 are of frequent occurrence in mathematical investigations, and their
92.

therefore proceed to explain use leads to valuable results. in what sense such roots are to be regarded.

We

When the quantity under the radical sign is negative, we can no
as indicating a possible arithmetical longer consider the symbol a may be defined as a symbol which obeys operation ; but just as a x Ja = a, so we shall define J— a to be such that the relation

J

J

J

J— a x J— a = - a, and we shall accept the meaning to which this assumption leads us.
It will be found that this definition will enable us to bring imaginary quantities under the dominion of ordinary algebraical rules, and that through their use results may be obtained which can be relied on with as much certainty as others which depend
solely

on the use

of real quantities.

93.

By

definition,
.-.

J- I
x

x

J-I = 2

1.

Ja.J-l
(

Ja.
.

J-l^a^l);
be regarded as equivalent to

that

is,

J a J- 1) = - a.

Thus the product

J a J— 1 may
.

the imaginary quantity
94.

J— a.

It will generally be found convenient to indicate the imaginary character of an expression by the presence of the

symbol

J- 1

;

thus

JZjtf = Jja
95.

2

x

(

-

1)

= a J7 J-T.

always consider that, in the absence of any statement to the contrary, of the signs which may be prefixed before a radical the positive sign is to be taken. But in the use of imaginary quantities there is one point of importance which
shall

We

deserves notice.

SURDS AND IMAUiNAltY QUANTITIES.
Since
(- a) x (root,
b)

75

ab,

by taking the square

we have

J- a x J- b = ± Jab.
Thus
in funning the product of

J

a and

J— b

it

would appear

that either of the signs + or This is not the case, for

— might be placed before Jab.
I

J -a x J- b = J a J.

x \/b

.

J-l

= - Jab.
96.

It

is

usual to apply the term

c

imaginary

'

to all expres

sions which are not wholly real. Thus as the general type of all imaginary expressions. are real quantities, but not necessarily rational.
97.

a+bj—l may

be taken Here a and b

In dealing with imaginary quantities we apply the laws of combination which have been proved in the case of other surd
quantities.

Example Example

1. 2.

a

+b

J - 1±

(c

+d

J-

1)

= a ± c + (b ± d)
c

J-

1.

The product

of a 4 b

J - f and
*/

+ dj-l

= (a + bj^l)(e + dj^l)
= ac - bd 4- (be + ad)
98.

-

1.

If a + b
if

J — 1 = 0,

2/tew

a=

0,

em(£ b

=

0.

For,

a+

then

J^\ = 0, bJ—\=-a;
b

.'.
2

—6" = a";
b
2

.'.

a +

= 0.
;

are both positive, therefore their sum cannot be zero unless each of them is separately zero that is, a — 0,

Now
6

« 2 and

b

2

and

= 0.
7f& + bJ^T = c + d J~l

99.

1,

then a

= c, andh
;

<I

For, by transposition, a
tlierefore,

—c+
and

by the

last article,

— d) J- 1 = a — c = 0, and 6 — ^ = 0;
(b

that

is

a=

c,

6

</.

76

HIGHER ALGEBRA.
Thus in order that two imaginary expressions may necessary and sufficient that the real parts should be
imaginary parts should
be equal. be equal it

is

equal,

and

the

When two imaginary expressions differ Definition. 100. only in the sign of the imaginary part they are said to be conjugate.
Thus a — b J —
Similarly
1 is

conjugate to a + b
is

J — 1.

^2 + 3^-1

conjugate to

J - 3 J- 1.
'2

The sum and the product of two conjugate imayinary 101. expressions are both real.

For
(

Again

J -\ +a-b J-\ = 2a. (a + b J- 1) (a - b J- 1) = a - (- b
a +
b
2

2

)

= a + b2
2

.

102.

a +

2

b

2

is

Definition. The positive value of the square root of called the modulus of each of the conjugate expressions

a +b
103.

J— 1
the

and a —

b

J—

1.

The modulus of

sions is equal to

product of two imaginary expresthe product of their moduli.

Let the two expressions be denoted by

a+bj— 1

and c+dJ—\.
which
is

Then

their

product = ac

— bd +
2

(ad

+ bc)
be)
2

J — 1,

an

imaginary expression whose modulus

= J(ac -

bd)

+ (ad +

= Jasc* + b*<f + a*d* + bs(?
=
=

J (a

2

+

b
2

2

)

(c

2

+

dr)

Ja

2

+

b

x

2 2 Jc + d

;

which proves the proposition.
form a + bj— 1, may be rationalised by multiplying the numerator and the
104.
If the

denominator of a fraction

is

of the

it

denominator by the conjugate expression a —

b

J—

1.

)

;

SURDS AND IMAGINARY QUANTITIES.
For instance

77

dj - 1 a + b J=l
c

+

(c

+

dJ-\)(a-bJ -\
— be) J —
b
72
1

~(a + b J~i)(a-bJ-T)
ac + bd + (ad

a +
ac

2

+ bd a + b~

-i

ad — be v a' + b

_•

Thus by reference to Art. 97, we see that the sum, difference, product, and quotient of two imaginary expressions is in each case an imaginary expression of the same form.
105.

To find

the square root

of a + h

J—

1.

Assume

J a + b V— 1 =x + y s/ — 1,
a+

where x and y are real quantities.

By

squaring,

bj—\=x -y
2

2

+ 2xy J —
parts,

1

therefore,

by equating

real

and imaginary
2

x2
2

-y =
^!/ =
2 2
)

a
b

(1),

(2);
2

.-.

(x

+y

=

(x
2

- ff + (2xyY
,

=a +
.•
.

b

2.2
;

x s + if =
obtain

Ja

2

+

6"

(3).

From

(1)

and
.

(3),

we

x

Ja -*

2

+2b 2 + a

.

.»"=*—

Ja + b -a i
2

2

;

Thus the required root

is

obtained.
(3)

Since x and y are real quantities, x 2 + y- is positive, and therefore in the positive sign must be prefixed before the quantity *Ja 2 + b-.

Also from (2) we see that the product xy must have the same sign as hence x and y must have like signs if b is positive, and unlike signs if b
negative.

b

;

is

;

;;

78
Example
1.

HIGHER ALGEBRA.
Find the square root
of

-

7

- 24

JN/

1.

Assume
then

J- 7- 21*/- l = x + y J- 1
-7-24 N/T l =
•••

-^

2

-2/

2

+ 2^

T l;
(1),

* 2 -?/ 2
2a;?/

=-7
=
-24.

and

= 49 + 576 = 625; .\ar + 2 = 25
2/

(2).

From

(1)

and

(2),

x- = 9

and
.-.

= 16 x= ±3,
2
?y

?/==

±4.

Since the product xy

is

negative,

we must take
3,
?/

x-

3,

y = - 4 j or # = -

= 4.
1

Thus the
that
is,

roots are

3 - 4 „/ - 1

and -3 + 4

*J

-

7 - 7 - 247"TT= ± (3 - 4 J ~i).
2.

Example

To

find the value of ^/

- 64a 4

.

It

remains to find the value of

\/ ±*J -

1.

Assume

;

SURDS AND IMAGINARY QUANTITIES.

70

— 1 is often represented by the letter i; but 106. The symbol until the student has had a little practice in the use of imaginary

J

quantities he will find

it

easier to retain the

symbol
1

J—
;

1.

It is

useful to notice the successive powers of

J—

or

i

thus

(7-1)^1,
?',

i*=ij
it

and since each power is obtained by multiplying the one before by J — 1, or we see that the results must now recur.
107.

investigate the properties of certain imaginary quantities which are of very frequent occurrence.
shall

We

now

Suppose
that
.

x - ^1
(x

;

then x 3 =
1
)

1,

or

x3 —

1

=

;

is,

-

(x

2

+

x+

1)^0.
1

\

either

x—

1

=

0,

or

x2 + x +
x=

=

;

whence
It

35=1, or

-W-3

.

be shewn by actual involution that each of these Thus unity has three cube values when cubed is equal to unity.
roots,

may

-l+JZTs
2
"'

-1-733~'
2

two

of

which are imaginary expressions.
ft
;

Let us denote these by a and
of the equation

then since they are the roots

x2 + x +
their product is equal to unity

l

=0.

that

is,
.

aft=
\

1

;

that

is,

.

2 aft = a = a2 ft

;

,

since a
2

8

= 1.

Similarly
108.
other, it

we may shew

that a =

ft

.

Since each of the imaginary roots is thr, square of thr is usual to denote the three cube roots of unity by 1, <d, «>~.

,

r

;

1

80
Also
a)

HIGHER ALGEBRA.
satisfies

the equation x + x +
.

2

1
;

=

;

•.

1

+

to

+ w2 =

that

is,

the

sum of the

three cube roots of unity is zero.
2

Again,

to

.

o>

—w =

1

therefore (1) the product of the two imaginary roots is unity 3 (2) every integral power of w is unity.
It is useful to 109. integral powers of a> are 1,
it

;

notice
to,
;

that the
2
;

must be
If
-rt>

of the

form

3m

and w for, and to" = w3m —
3, it

if
1.

successive positive n he a multiple of 3,

n be not a

multiple of
1

must be
3m +
o>

of the
3m

form

3m +

1

or

3m + 2.
It

KO
110.
"

n n = dm + 1 O w = om +-,
.

Wn —
"

=to

.to
2

=

to.

w=w 3m +2 =w3m .w=to.
2

see that every quantity has three cube roots, two of which are imaginary. For the cube roots of a 3 are those 2 7 Similarly the cube roots of a x 1, and therefore are a, ao>, aw 2 7 of 9 are ^9, o> ^9, a> ^9, where ^ 9 is the cube root found by the In future, unless otherwise stated, ordinary arithmetical rule.
.

We now

the symbol %ja will always be taken to denote the arithmetical cube root of a.
(9
4.

Example

1.

Reduce

*

3

/
,

_

1 \2
.

to the

form A +

2 + N/-l

Bj - 1.

The expression

^

4-9 + 12^-1
2

+ v/ :=~l

(-5 + 12 N/Jl)(2-V^l)
(2+

J-l)(2-J-T)

-10 + 12 + 29 4+1

J~l

=2+ 5
which
is

W(x

29

/—

1;

of the required form.
2.

Example
Since

Resolve x 3 + y 3 into three factors of the first degree.

x 3 + if =
.'.

+ y)

x*

for

+ y 9 = (x + y) w + w2 = -1, and w 3 = l.

-xy + y 2) (x + toy) (x + ury)
(x 2

;

;

: ;

;

SURDS AND IMAGINARY QUANTITIES.
Example
3.

81

Shew

that
(a

(a

+ wb + arc)

+ w2 6 + toe) - a 2 + b2 + c 2 a

60 - ca

- ab.

In the product of
the coefficients of
Zr

a

+ wb + arc and
are or, or 1

+ orb + wc,

and

c

2

the coefficient of be the

= w2 + o>4 = or + o> = - 1 coefficients of ca and a& = 2 + o> = - 1 (a + cob + arc) {a + urb + wc) = a 2 + b 2 + c 2 - be - ea o>
.*.

ab.

Example

4.

Shew

that
(l

+ «-«>)'-(l-M+U?)*=0.
1

Since
(1

+ w + o> 2 = 0, we have
(1

+ u- w 2

3
)

-

- w + wa)3=( - 2m2) 3 -

(
{

- 2o>) :J

= -8o> 6 + 8a/ = -8 + 8
= 0.

EXAMPLES.
1.

VIII.

b.

Multiply 2 \/~~3 + 3
Multiply 3

V 3^

by 4 *J^3- 5 a/^2.
by 3

2.

V ^7 - 5 V^ l+V 3^

V^+ 5 V^.
.

3.

Multiply
AT
IX' Multiply
I

e^ -1 +e'^~ 1 by e^ _1 -e -V-*.
a;

4.

A

5

by #

.

l-V^ =

Express with rational denominator
_
0.

— 3-V-2
1
,

3 a/~2~ + 2 *J~-h

o.

,

,

.

3V-2-2V-5
a+rV^l
a-WisT-i
(a +

3+2

V~l

2-5\/
(.f

r

3-2 V^l
2

a-x»f-i
a+a?V^-l"

l

+

5V :r l'
ia
+
3
,

g

+ V-l)a .v-V-1

(W- l)^
.r

+ V-1
\/

V-lja-Cft-V-l)^ (a + \/-l) 2 -(«- V-l) a#
when w
is

11.

Find the value of ( Find the square of

-

l) 4n

a positive integer.

12.

Jd + 40 V"-T+ V9 - 40 V -?.
(j

H. H. A.

a

82

HIGHER ALGEBRA.
Find the square root of
13.
16.

-S + ISV^L

14.

-ll-COV 17

!".

15.

-47 + 8V-3.

-SV^l.

17.

a 2 -l+2a^^l.

18.

±ab-2(a 2 -b 2 )*/^T.

Express in the form
iy *

A + iB
ZU>

2-3r
a + O8
3

2V3-i\/2'
(^ + ^)
2

zu

1-T

09
If
24.
26.

9<*

(«-^) 2
a + io

—i

a — ib

2 1, co, g>

are the three cube roots of unity, prove
25.
(1

(l+co 2 ) 4 = co.
(1

(l-co + co 2 )(l+co-or)

= 4.

- co)

(1

-

co-

)

- CO 4 )

(1

- co 5 ) - 9.

27. 28.
29.

(2

+ 5co + 2co 2 6 = (2 + 2co + 5a> 2)6 = 729.
)

(l-co

+ co 2 )(l-co 2 +

co

4

)(l-co 4 + co 8 )... to

2>i

factors = 2 2 ».

Prove that A3 +yZ + £> _ 2#gZ = (x+y+z) (# + i/a> + Za> 2 ) (x +y<o 2 + Za).

30.

If

x=a+b
xyz=a3 +b3
^- 2
?/
.

t

y — aw + Z>co 2

,

s=«co 2 +

6co,

shew that
(1)

(2) (3)

31.

If
(

shew that

+ 2 + 5 2 = 6a6. a3 +y3 +s3 =3(a3 +&3) ax + cy + bz = X, ex + by + az = I", Zu- + ay + gs = if, 2 + 6 2 + c2 - be - ca - ab) (x2 +y2 + z2 -yz- zx - xy) = X 2 +Y2 + Z - YZ- XZ- XY.
k

2

2

CHAPTER

IX.

THE THEORY OF QUADRATIC EQUATIONS.
111.

After

suitable reduction every quadratic equation m;iy
(Lif Jrbx

be written in the form

+

C

(1),

and the solution

of the equation is

x=

-

b

± Jtf^iac

\ 2a

(2). v
'

We

shall

now prove some important
all

propositions connected
is

with the roots and coefficients of
the type.
112.

equations of which (1)

A
if

quadratic equation cannot have more than
possible, let the equation

tiro roots.

have three Then since each of these values must different roots a, f3, y. satisfy the equation, we have
For,
c

ax 2 + bx +

=

aa 2
2

-t-

ba +

c~0
c

(1), (2),

afi

+ bp +

=

ay 2 + by + c =

(3).

From

(1)

and

(2),

by subtraction,

a(a 2
divide out

-p

2 )

+ b(a-P) = 0;
is

by a — fi which, by hypothesis,

not zero; then

a (a + ft) + b = 0.
Similarly from (2) and (3)

a
.•.

{fi

+

y)

+

b

=

;

by subtraction
is

a (a - y) —

;

which

impossible, since,
y.

not equal to

by hypothesis, a is not zero, and u Hence there cannot be three different roots.

is

G—

.

84
113.
/3,

HIGHER ALGEBRA.
In Art. Ill -b +
let

the two roots in (2) be denoted by a and

so that

a

=

2 Jb - Aac

2a

, '

P
'
:

—b- Jb — Aac =
2

_

2a

'

then we have the following results
(1)

a and

(3

(2)

- Aac (the quantity under the radical) are real and unequal. 2 If b — Aac is zero, a and ft are real and
If b
2

is

positive,

equal, each

reducing in this case to
(3)
(4)

— 77-

2a

If b

2

- Aac

is

negative, a

and

ft

are imaginary and unequal.
/3

If b

2

Aac

is

a perfect square, a and

are rational and

unequal.

applying these tests the nature of the roots of any quadratic may be determined without solving the equation.
Example
Here
1.

By

Shew that the equation 2x 2 -6a; + 7 =
x.

cannot be

satisfied

by any real values of

a=

2, b

= - 6, c — 7

;

so that

&2_ 4ac= (_6)2-4.2.7=-20.
Therefore the roots are imaginary.

Example

2.

If the equation

a?

2

+2

(k

+ 2) x + 91c = has equal roots, find

l\

The condition

for equal roots gives
(fc

+ 2) 2 = 9£,

fc2_5ft + 4=0,

(fc-4)(fc-l)=0j
.-.

k = A, or

1.

Example
are rational.

3.

Shew

that the roots of the equation

x 2 - 2p3 +p 2

-q 2 + 2qr- r2 =

2 2 2 roots will be rational provided (-- 2p) - 4 (p - q + 2qr-r2) is a 2 But this expression reduces to 4 (q -2qr + r2), or 4:(q-r) 2 perfect square. Hence the roots are rational.

The

.

- - ,

114.

D Since
.

a=

-

b

+ Jb 2 - Aac

^

'

?=

-b- Jb

2

- Aac
'

2a
2

we have by

addition
-

*

+

^ __M_b
2a

b

+ Jb 2 - Aac

-b- Jb - Aac
0);

2a

a

THE THEORY OF QUADRATIC EQUATIONS.
and by multiplication we have
n = (- b

85

+

J¥~r^

)

(_.

5

_ j,/r^c)

_

4ac
2

~4a

=

c

«
in the
c

:

(2).

By writing the equation
2

form
'

°

a
these results

a

may

also be expressed as follows.
e(
l

* qUadratiC
unity,

uation

«**" **
is

coefficient

of

the first

term

is

its

4**5d?
(ii)

the roots

equal t0 the coefficient of

with

the product of the roots

is

equal to the third term.
not contain the 8nta0

*£%* £»"& ££&£**"
115.

Since

__ =a+ £
- a; + - =

and £
be written

*

the equation ar+

may

x 2 -(a + f])x + ap =

(1).

Hence any quadratic may also be expressed in the form 2 x - (sum of roots) x + product of roots =
Again, from
(1)

(2).

we have

(x-a)(x-p) = Q

We

(3).

may now
1.

easily

form an equation with given
and 2.

roots.

Example

Form
is

the equation whose roots are 3
(* - 3)

The equation
or

(*+ 2)=0,

««-*- 6=0.

When
metnou.

the roots are irrational

it is

easier to use the following ~>

;

.

86
Example
2.

HIGHER ALGEBRA.
Form
the equation whose roots are 2

+ ^3 and

2 -

^3.

We have
. •

sum
is

of roots

= 4,

product of roots = 1
.

the equation

x(2) of

- Ax + 1 = 0,

by using formula

the present article.

a method analogous to that used in Example 1 of the last article we can form an equation with three or more given
116.
roots.
7 o

By

Example
positions

1.

Form

the equation whose roots are
satisfied

2,

-

3,

and ^

The required equation must be
:

by each of the following sup-

#-2=0,
therefore the equation

# + 3 = 0,

7 £-- = ();

must be

(*-2)(*+3)(*-|)=0j
that
is,

[x

- 2) (x +3) (5a;- 7) =0,
3

or

5a;

-2a; 2 -37a; + 42 = 0.
0,

Example

2.

Form

the equation whose roots are

±«, j

.

The equation has

to be satisfied

by

x = 0, x = a.
therefore
it is

x=-a, x=}

c b
;

x
that
or
is,

(x -J- a) (x

- a)

(

x--

j

=

;

- a 2) (bx - c) = 0, bx 4 - ex 3 - a~bx- + a-cx = 0.
x
(x 2

117.

The

results of Art.

generally sufficient In such questions the roots should never be roots of quadratics. considered singly, but use should be made of the relations obtained by writing down the sum of the roots, and their product, in terms of the coefficients of the equation.
are

114 are most important, and they to solve problems connected with the

Example
(l)a 2
,

1.

If

a and
.

/3

are the roots of

x--px + q = 0,

find the value of

+ /3 2 (2)a s + /3 3
have

We

a + (2=p,
a(3
.-.

a

= q.

a2

+ /3 2 =(a + /3) 2 -2a/3
=p*-2q.

.

m

;

:

THE THEORY OF QUADRATIC EQUATIONS.
Again,
a
;{

87

+ ft = (a + /3)

(a 2

+ p" 2 -

a/3)

a =i>{(a + /3) -3a/3]

=*(?»- 89).
Example
2.

If a,

p"

are the roots of the equation

/.r'-'-|

mx + 7i = 0,

find the

equation whose roots are


p

,

-

a
of roots

"We have

sum

=
p
-

+
.

-

=
ap

^
,

a.

product of roots =
.-.

-=1

p a
by Art. 115 the required equation
is

or

apx2 -

(a

2

+ p~2 £ + 0/3 = 0.
)

O

As
.,

in the last

example o2 +j8B =

=—
L

O

7

,

and
?i

a/3

=

--

.

t

.*.

the equation

,.

n
is
V

?ji

2

-2wZ
-=
v

x-

x+

y
it

= 0,

_

or

n /x- 2 -

(

2

- 2nZ) x + nl = 0,

Example

3.

When
it

.r

=

^
if

,

find the value of

2x 3 + 2x 2 -7x+l'2
x.

;

and shew that

will

be unaltered

£-

be substituted for

a

Form
the

the quadratic equation whose roots are
of the roots

^

;

sum

=3
17 =—
2.r 2
;

the product of the roots

hence the equation
.*.

is

-

6.r

+ 17 =

;

2x 2 -6x + 17

is

a quadratic expression which vanishes for either of the

, values

3*5^/-"!

^
m

Now

2a* +

2.t

2

- Ix + 72 = x

(2.r 2

-

C>.c

+

17)

+4

(2.r

2

- Cx + 17) + 4

=xx0+4x0+4 = 4;
which
is

the numerical value of the expression in each of the supposed cases.

88
118.

HIGHER ALGEBRA.
To find
c

Vie condition

that

the

roots

of the equation

ax + bx +
in sign,

2

=

should be (1) equal in magnitude

and

opposite

(2) reciprocals.

The roots will be equal in magnitude and opposite hence the required condition is their sum is zero
;

in sign

if

_-=
a

0,

or b

= 0.
is

Again, the roots will he reciprocals when their product unity ; hence we must have
c
i — = 1. or

a

c

= a.

these results is of frequent occurrence in Analytical Geometry, and the second is a particular case of a more general condition applicable to equations of any degree.
first of

The

Example. Find the condition that the roots of ax2 + bx + c = may be both positive, (2) opposite in sign, but the greater of them negative.

(1)

We
(1)

have
If the roots are

a

+ B=


b

,

a

a8=-

c
.

a

both positive,

a/J is positive,

and therefore

c

and a

have

like signs.

Also, since a
signs.

+ fi

is positive,

is

negative; therefore b and a have unlike

Hence the required condition and opposite to the sign of b.
(2)

is

that the signs of

ft

and

c

should be

like,

a have unlike

If the roots are of opposite signs, a/3 is negative, signs.
it is

and therefore
negative,

c

and

Also since a +/3 has the sign of the greater root
fore - is positive; therefore b
(X

and

there-

and a have
is

like signs.

Hence the required condition and opposite to the sign of c.

that the signs of a and b should be

like,

EXAMPLES.
Form
.

IX.

a.

the equations whose roots are 4
o

3
/

m
n
5.

n

m
6.

p-q p+q

p+q

p—y

4.

7±2

N /5.

±2«/3~5.

-p±2s/Tq.

)

=

.

:

THE THEORY OF QUADRATIC EQUATIONS.
7.

89

-3±5l
-3, |, i.
Prove that
(

8.

-a±ib.
|,
0,

9.

±i(a
2±«/3,

-b).

10.

11.

-|.

12.

4.

13.

tlio

roots of the following equations are real
c2

1

x2 - 2ax +a a - 6a (a

- 0,
6)
.v

(2)

- b + c) r- +4 (a -

+ (a - b - c) = 0.
lias

14.

If the equation

x2 - 15 -m(2x-8) =

equal roots, find the

values of m.
15.

For what values of

m will the

equation
(3

x2 - 2x (1 + 3//0 + 7
have equal roots
16.
?

+ 2m) =

For what value of

m

will the equation

x*

— bx

ax-c

m- 1 m+ 1
?

have roots equal in magnitude but opposite in sign
17.

Prove that the roots of the following equations are rational:
(1) (2)

+ c-b)x 2 + 2cx + (b + c-a) = 0, abc 2 x2 + 3a 2 cx + b 2 ex - 6a 2 -ab + 2b' = 0.
(a
1

If

a, /3

are the roots of the equation

ax2 + bx + c = 0,
20.

find the values of
2

18.

»,+£.

19.

aW + aV.
= 1 + 2/.
3+ ?

(|-f)

.

Find the value of
21. 22.

a3 + s2 - X + 22 when

.r

x3 - Zx2 - 8x +15 when x
.t-

23.

3

- «.r2 + 2a 2.r + 4« 3 when

-= 1 - J - 3.
a

If o and /3 are the roots of 24. whose roots are (a- ft) 2 and (a + /3) 2
.

x*+px+q=O

t

form the equation

25.

Prove that the roots of (x — a) (.»; -b) = h 2 are always red.
I f
.'-, ,

26.

x%

are the roots of ctx*+bx

+

c

= 0,

find tho value

i

>f

(1)
(2)

(ax l

+ b)- 2 + («x + b)- 2
i

,

(ax^byt+iaxt+b)-*.

90
27.

HIGHER ALGEBRA.
Find the condition that one root of ax2 + bx-\-c =
shall

be

n times the other.
28.

If

a, (3
2

are the roots of
2

roots are a
29.

+ /3 and o~

2

+/3

-2

ax2 + bx + c = 0, form the equation whose

,

the equation whose roots are the squares of the of the difference of the roots of
2x* + 2
30.

Form

sum and

(m + n) x + m2 + n 2 =0.

Discuss the signs of the roots of the equation

px2 + qx + r = 0.
The following example illustrates a useful application 119. of the results proved in Art. 113.
Example.
can have
all

If

x

is

a real quantity, prove that the expression
lie

x- + 2x

— 11
-.

numerical values except such as

between 2 and

6.

Let the given expression be represented by

y, so that

a 2 + 2:r- ll_

2(s-3)
2

~ y;

then multiplying up and transposing, we have
rr

+2.r(l-?/)

+ 6f/-ll = 0.

a quadratic equation, and in order that x may have real values 4(1 -i/) -4(Gy — 11) must be positive; or dividing by 4 and simplifying, 2 Hence 8*/ + 12 must be positive ; that is, (y - 6) (y - 2) must be positive. ?/ the factors of this product must be both positive, or both negative. In the former case y is greater than 6; in the latter y is less than 2. Therefore y cannot lie between 2 and 6, but may have any other value.

This

is

2

In this example it will be noticed that the quadratic expression y — 8y + 12 is positive so long as y does not lie between the roots 2 of the corresponding quadratic equation y — Sy + 12 = 0.
2

This is a particular case of the general proposition investigated in the next article.
120.
the

For

all real values
a,

of x

tlie

expression ax 2

same sign as

are real

when the roots of the and unequal, and x has a value lying between
except
I.

+ bx+c has 2 equation ax +bx + c =0
them.

Case

Suppose that the roots of the equation

ax 2 + bx +
are real
;

c

=
let a

denote them by a and

ft,

and

be the greater.

THK THEORY OF QUADRATIC EQUATIONS.
Then
2

91

((.r

4

bx +

c

—a

(

x* + -

V

a

X+
+

)

aj

= a {x2 = a (x -

(tt

ft)

X+

aft

a) (x
a,

- ft).

the factors x — a, x — ft are both if x is less than ft, the factors x — a, x — ft are both positive negative; therefore in each case the expression (x — a)(x — ft) is 2 But if x has a positive, and ax + bx + c lias the same sign as a. value lying between a and ft, the expression (./• - a) (x - ft) is negative, and the sign of ax" + bx + c is opposite to that of a.
if
;

Now

x and

is

greater than

Case

II.

If a

and

ft

are equal, then
c

ax 2 + bx +

= a(x—

a)

2
,

and

(x - a)

2

is

positive fur all real values of
a.

x

;

hence ax 2 + bx +

c

has the same sign as

Suppose that the equation ax 2 + bx + imaginary roots ; then

Case

III.

c

=Q

lias

ax 2 + bx +

c

— alx* {-•
(/

a

x+

a\

b\-

iae-b')

But iac-b*
4a
2

b
.

2

4«c
.

is

negative since the roots are imaginary

;

hence

is positive,

and the expression
2

x + ^(
is

Aac

b

2

&)

+
x
;

\a 2
2

positive for all real values of
a.

same sign as
121.

therefore ax + bx + This establishes the proposition.

c

has the

the preceding article it follows that the expression ax + bx + c will always have the same sign whatever real value x may have, provided that b 2 - Aac is negative or zero; and if this condition is satisfied the expression is positive or negative accord2

From

ing as a

is

positive or negative.
2

Conversely, in order that the expression ax + bx + c may be always positive, b 2 — Aac must be negative or zero, and a must be 2 positive; and in order that ax + bx + c may be always negative 1 I - Aac must be negative or zero, and a must be negative.


92
Example.

— —
HIGHER ALGEBRA.
lie in

;

;

Find the limits between which a must
ax 2 - Ix + 5 5x 2 - Ix + a

order that

may

be capable of

all values,

x being any
1
i-

real quantity.

_ Put
then
expression

ax -lx + 5
o
rr

=v;

(a-5?/)a:2 -7.r(l-?/)

+ (5-a?/):=0.

In order that the values of x found from this quadratic
49
that
is,

may

be real, the

(1

- y)'2 - 4 (a - 5y)
2
2

(5

- ay) must be

positive,

- 20a) y + 2 (2a + l)y + (49 - 20a) must be positive hence (2a 2 + 1) 2 - (49 - 20a) 2 must be negative or zero, and 49 - 20a must be
(49
positive.

Now
that
is,

(2a2 +

2

1)

-

(49

- 20a) 2

is

negative or zero, according as

2 (a 2

- 10a + 25) x 2
4 (a -

(a 2

+ 10a - 24)
(a

is

negative or zero negative or zero.

according as

o) 2 (a

+ 12)

- 2)

is

This expression is negative as long as a lies between 2 and - 12, and for such values 49 - 20a is positive; the expression is zero when a = 5, - 12, or 2, but 49 -20a is negative when a = 5. Hence the limiting values are 2 and - 12, and a may have any intermediate value.

EXAMPLES.
1.

IX.

b.
lie

Determine the limits between which n must

in order that

the equation

2ax (ax + nc) + (n 2 - 2) c2 =

may have
2.

real roots.

If

x be real, prove that x
^•2
.77
'-

x x — bx + 9
l

-5

'-

must

.

lie

between

1

and - r^
11

1

.

4-

3.

Shew
If
9.

that -=

1

x- + x+\

lies

between 3 and - for
3
3
^ 1

all real

values of x.

4.

x be

sb

real,

prove that x

34a?— = =—++ 2# — 71 can have no value between x 7

5

and
5.

Find the equation whose roots are
s]a ± sja

-b
find the value of

6.

If

a, /3

are roots of the equation

x2 — px+q=0,

(1) (2)

atitfp-i-fl + ptfPa-i-a),
(

a

-p)-* + (P-p)-\

THE THEORY OF QUADRATIC EQUATIONS.
7.

!>:{

If the roots of ht?+

nx+n=0 be in

the ratio of p

:

q prove that
t

8.

If

x be

real,

the expression
2)i

r n 2 (x - n)
.

admits of

all

values

except such as
9.

lie

between
y
-

and 2m.

Tf the roots of the equation

ax2 + 2hx + c=() be
be
a-ffi

a

and

(3,

and

those of the equation Ax'

+ 2Ux+C=0
~~a 2

and

fi

+ d,

prove that

b*-ae_ B*-AC

~

A2

'

10.

Shew that the expression —
when x
is real,

values

provided

p + 3x - 4x* that p has any value
.#4-2
*

—5

will be capable of all L

between

1

and

7.

11.

Find the & greatest value of n 2 n 2x + 3x + 6

for real values of x.

12.

Shew

that

if

x

is real,

the expression
i

(x 2

-bc)(2x-b-c)~
c.

has no real values between b and
If the roots of the roots of
13.

ax2 + 2bx + c =

be possible and different, then
(ac

(a

+ c) (ax 2 + 2bx + c) = 2
vice versa.

-

b2 )

(.r

2

+ 1)

will

be impossible, and

14.

Shew that

the expression

values

when x

is real, if

a2 - b 2

,! — a) (ex — a) and c 2 — d 2 have the same

- fl

{-)-

will be capable of all
sign.

(ox

*122. shall conclude this chapter with some miscellaneous It will be convenient here to introduce theorems and examples. a phraseology and notation which the student will frequently meet with in his mathematical reading.

We

Definition. Any expression which involves x, and whose value is dependent on that of x, is called a function of X. Functions of x are usually denoted by symbols of the form f(x),
F(x),<f>(x).

Thus the equation y =f(x) may be considered as equivalent to a statement that any change made in the value of as will produce a consequent change in and vice versd. The quantities x and y are called variables, and are further distinguished as the independent variable and the dependent variable.
;//,

94

HIGHER ALGEBRA.

independent variable is a quantity which may have any value we choose to assign to it, and the corresponding dependent variable has its value determined as soon as the value of the inde-

An

pendent variable
%

is

known.

123.

An

expression of the form
n +p x
x
]

pjs"

+ pjf

2

+

.

.

.

+ pn _ x + p n
t
lt

where n is a positive integer, and the coefficients p p lt p a ,...p n do not involve x, is called a rational and integral algebraical function of x. In the present chapter we shall confine our attention to
,

functions of this kind.

*124. function is said to be linear when it contains no higher power of the variable than the first ; thus ax + b is a linear function is said to be quadratic when it function of x. contains no higher power of the variable than the second ; thus ax2 + bx + c is a quadratic function of x. Functions of the third, fourth,... degrees are those in which the highest power of the Thus in the last variable is respectively the third, fourth, th degree. article the expression is a function of x of the n

A

A

The symbol fix, y) is used to denote a function of two 2 2 variables x and y thus ax + by + c, and ax + bxy + cy + dx + ey +f are respectively linear and quadratic functions of x, y.
*125.
;

The equations fix) =
ratic,
...

are said to be linear, quadfix, y) — according as the functions f(x), f(x, y) are linear, quad0,

ratic,....

have proved in Art. 120 that the expression ax + bx + c admits of being put in the form a (x — a) (x — fi), where a and j3 are the roots of the equation ax2 + bx + c — 0.
*126.
2

We

Thus a quadratic expression ax 2 + bx-\- c is capable of being resolved into two rational factors of the first degree, whenever 2 has rational roots ; that is, when the equation ax + bx + c = 2 b - iac is a perfect square.
*127.

To find

the condition that

a quadratic function ofx,y

may

be resolved into two linear factors. y)

Denote the function hy f(x,

where
c.

2 z f{x, y) = ax +'2hxy + by + 2gx+ 2fy +

THE THEORY OF QUADRATIC EQUATIONS.
Write
thus
ax*
this in descending

95
it

powers of
y)

x,

and equate
c

to zero;

+ 2x (hy +

+ by 2 + 2fy +

-

.

Solving this quadratic in x

we have
2fy

2 _ - (h + (j) ± J{hy + y)* - a (by + x—

+
+

c)
,

a

<

>v

ax + hy + g = ±

Jy

2

(h*

-

ab)

+ 2y

(hy

a/)

2

(g

-

ac).

in order that J\.r, y) may be the product of two linear factors of the form px + qy + r, the quantity under the radical

Now

must be a perfect square
(kg

;

hence
(h-

- a/) 2 =

a,

ab) {<f

- ac).

Transposing and dividing by
abc + 2fyh
2

we obtain
2

— af —

by

ch

2

=

;

which

is

the condition required.
is

This proposition
*128.

of great

importance in Analytical Geometry.

To

find the condition that the equations

ax 2 + bx +

c

— 0, ax 2

f b'x

+

c

-

may have

a

common

root.

Suppose these equations are both
aa.
2

satisfied
0,

by x

a

;

then

+ ba +

c

=

a'a +b'a + c'
.*.

2

= 0;
1

by cross multiplication
a"

a
b'c

be

ca — c'a

ab'

— ab

'

To eliminate
equate
it

square the second of these equal ratios and to the product of the other two ; thus
a,

a
(ca
.'.

a
2

1

(ca

— c'a) — ca) 2 =

(be

b'c)

'

(ab'

— ab)

'

(be

b'c) (ab'

— ab),

which
It

is
is

the condition required.

easy to prove that this is the condition that the two 2 quadratic functions ax 2 + bxy + cy 2 and a'x + b'xy + c'y' may have

a common linear

factor.

1

06

HIGHER ALGEBRA.

^EXAMPLES.
1.

IX.

c.

For what values of

m will

the expression

2 y + 2xy + 2x + my - 3

be capable of resolution into two rational factors

?

which will make equivalent to the product of two linear factors.
2.

Find the values of

m

2.v

2

+ mxy + 3y 2 - 5y - 2

3.

Shew that the expression

always admits of two real linear
4.

factors.

If the equations

x2 + px + q = 0,
have a

x2 + p'x + q' =
must be
either

common

root,

shew that

it

p'l'-p'q nr

9-q

q-q
5.

p-p
+ m'xy
-f-

Find the condition that the expressions
Lv2 + mxy + ny 2 ,
l'x2
n'y'
1

may have
6.

a

common

linear factor.

If the expression

%a? +

2Pxy + 2y 2 + 2ax - 4y +

can be resolved into linear factors, prove that roots of the equation P'2 + 4aP + 2d1 + 6 = 0.
7.

P

must be one of the

Find the condition that the expressions

ax2 + 2hxy + by 2

,

a'x2

+ 2k'xy + b'y 2

may

be respectively divisible by factors of the form y
8.

-mx, my + x.

Shew

that in the equation

x 2 - Zxy + 2y 2 - 2x - 3y - 35 = 0,
for every real value of x there is a real value of y, value of y there is a real value of x.
9.

and

for every real

If

x and y are two

real quantities connected
92.r

by the equation

9x2 + 2xy +y 2 then
will

- 20y + 244 = 0,
1

x

lie

between 3 and

6,

and y between

and

10.

10.

may

If (ax2 + bx + c)y-\-a'x 2 + b'x + e' = 0, find the condition that be a rational function of y.

x

CHAPTER

X.

MISCELLANEOUS EQUATIONS.
In this chapter we propose to 129. consider some miscellaneous equations ; it will be seen that many of solved by the ordinary rules for quadratic equJtions, but others require some special artifice for their solution

the^l

Z

Example

_

3_

3

I.

Solve
.r

8x 2n

-8x~^=63.

Multiply by

2n

and transpose; thus
i. 8x n - 63x 2 '*-8 = 0;


-

L
«2n = 8,

(a?"-8)(8x^+l) = 0;

or--; 8'
2n

1

=(*)* «(-p)*;
.-.*=«», or
Example
2.

A.
^

Solve

2

V

/-+
«

3

/- =

6a 6
I

«

.•.%+! = * +
</

«

2<% 2 -6a 2?/-& 2 + 3a& = 0;
(2ay~&)(ty-3a)=0;
6

3a
9a 2

*

&2

4a

"l
'

&a

*

H. H. A.

98
Examples.
Solve

HIGHER ALGEBRA.
(*-5)(a:- 7)(«
(x
2

+ 6)

(*

+ 4) = 504.

We have

- x - 20) (x 2 - x - 42)

= 504

;

which, being arranged as a quadratic in x 2 (a2
.-.

x, gives

- x) 2 - 62
(x2

(x2

- x) + 336 =

-a:-6)(x2 -x-56) =
f

.-.

X *- X -Q =

or a2 -a;
8,

-56 =

whence
130.

x = S, -2,

-7.

Any

equation which can be thrown into the form
2 ax 2 + bx + c + p J ax + bx +

c
2

—q
+ bx +
c,

may

be solved as follows.

Putting y =

J ax

we

obtain

Let a and

ft

be the roots of this equation, so that
2

J ax
"When no sign

+ bx +

c

= a,

Jax

2

+bx + c = ft

;

from these equations we
is

shall obtain four values of x.

prefixed to a radical it is usually understood that it is to be taken as positive; hence, if a and ft are both positive, all the four values of x satisfy the original equation. If however a or ft is negative, the roots found from the resulting

quadratic will satisfy the equation

ax2 + bx +

c

— p J ax +
2

bx + c =

q,

but not the original equation.
Example.
Solve x2 - ox + 2
;

Jx 2 - 5z + 3 = 12.
+ 2 N/^-5a; + 3 = 15.
;

Add

3 to each side

then
2

rc

-5a; + 3

Putting

Jx2 -5x+3 = y, we obtain y 2 + 2y - 15 = Thus *Jx 2 - 5x + 3 = + 3, or Jx 2 -6x + S = -5.
= x^
:

whence y = 3 or -

5.

ing Squaring, and solving the resulting quadratics, we obtain from the

ic=6 or -1; and from the second
satisfies the given equation,

—^

first

.

The

first

pair of values

but the second pair

satisfies

the equation

x2 - 5x -2

Jx2 -5x + 3 = 12.

.

;

;

MISCELLANEOUS EQUATIONS.
131.

99

Before clearing an equation of radicals it is advisable to examine whether any common factor can be removed by
division.

Example.

Solve *J x'2 - lax

+ 10a'2 - Jx- + ax- 6a- — x- 2a.

We

have

*J(x-2a)(x-5a) - J{x-2a) (x+Sa) = x-2a.

The

factor *J x - 2a can
.'.

now

be removed from every term

sjx

-5a- Jx + 3a = „Jx - 2a
(x

x - 5a + x + 3a - 2 *J(x - 5a)

+ 3a) = x - 2a
;

;

x = 2 Jx'2 - 2ax - 15a'2

3ar-8aa;-60a 2 = 0;
{x

-6a) (3a; + 10a) =0;
ba, or x—c

——
10a
obtain x = 2a.
:

Alsoxby equating to zero the factor

Jx - 2a, we

On

trial it will be

the roots are

—— and
D

found that x = 6a does not
2a.

satisfy the equation

thus

The student may compare a similar question discussed in the Elementary Algebra, Art. 281.
132.

The following
Solve J'3x- -

artifice is

sometimes useful.
11

Example.

4.x

+ 34 + JSx'2 - 4x -

=9

(1).

We

have identically
(3x--4a; + 34)-(3a; 2

-4x-ll) = 45
(1);

(2).

Divide each

member

of (2)

by the corresponding member of

thus
(3).

J'dx- -

4.r

+ 34 - JSx2 - 4x -

11

=5

Now (2) is an identical equation true for all values of x, whereas (1) is an equation which is true only for certain values of x hence also equation (3) is only true for these values of x.
;

From

(1)

and

(3)

by addition
2 v/3x -4a;

+ 34 = 7;
or

whence

as

= 3,

--.

7—7

.

.

100
133.

HIGHER ALGEBRA.
The
solution of an equation of the form

ax4 ± bx 3 ± ex 2 ± bx + a =
in

0,

which the coefficients of terms equidistant from the beginning and end are equal, can be made to depend on the solution of a
quadratic.
tions,

Equations of this type are known as reciprocal equaand are so named because they are not altered when x is
its

changed into

reciprocal -

For a more complete discussion of reciprocal equations the student is referred to Arts. 568 570.

Example.

Solve

12a; 4

- 56x 3 +

89a; 2

-

56.x

+ 12 = 0.

Dividing by x2 and rearranging,

12/W- ) -56^+^+89 =
2

0.

Put

x + -=z: then x
.-.

a;

2

+ — = z 2 -2;
x-

12

(z 2

-

2)

-56^ + 89 = 0;
-=.

whence we obtain

z

=2
1
a;

,

or
5

u

13
6

2

By

solving these equations

we

find that x = 2, -

13 ,

,

-

2

.

The following equation 134. solved in a similar manner.
Example.
Solve
6
6a; 4

though not reciprocal may be

-

25a; 3

+ 12a; 2 + 25a; + 6 = 0.

We have
whence

2 (^ +^i) - 25 fx - -\ + 12 = 0;

6(a;


a;

j

-25

(a;

— 1+24 =
- -

0;

.-.

2

(^--^-3 = 0,
= 2, - ,

or 3

fx- -]-8-0;

whence we obtain

3,

When one root of a quadratic equation is obvious by 135. inspection, the other root may often be readily obtained by making use of the properties of the roots of quadratic equations
proved in Art. 114.

MISCELLANEOUS EQUATIONS.
Example.
This
is

101

Solve

(

1

-

a-) {x

+ a) - 2a
is

(

1

-

ar)

= 0.

a quadratic, one of whose roots

clearly a.

Also, since the equation

may

be written
(1

2ax- + (1 - a 2 ) x - a

+ a 2 ) = 0,
is

the product of the roots

is

-

-

;

and therefore the other root

.

EXAMPLES.
Solve the following equations
1.
:

X.

a.

a- 2 -2x~ 1 = 8.
1

2.

9

+ a- 4 = 10a--'.
3
1

_J
*.

3.

2jx + 2x
2

§ =51

4.

6a?*~7**-8a7
JL

1

5.

3"+6=5#».

6.

3.f 2n

-.rri -2=0.

7
-

»>/;+Vj-"*
i

a
10. 12.

\/S + \/?'
1+8.^ +
5 (5*

2*.

9.

6 x/a=5a

2

-13.
3*.

9^ =

0.

11. 13.

32*+ 9« 10.
2 2* + 8 + 1

+ 5-*) = 26.

= 32.2'.

14.

2 2* + 3 -57 = 65(2*-l).

15.

,/*+£-*
(x

16.

^.-#=5A-

17.
18.
19.

+ 1 = 1 680. (x + 9) (x - 3) (x - 7) {x + 5) = 385. x (2x + 1 (.v - 2) (2a - 3) = 63.
- 7)
(a-

- 3) {x + 5)

(.v

)

)

20.

(2a-7)(a 2 -9)(2a + 5) = 91.
A'

21.
22.

2

+ 2 >/a + 6a = 24 - 6x.
2

3a2 - 4a + s/'3x i -4x-6 = l8. 3a2

23.
24.

-7 + 3
2

(N

/3sa -16a? + 21
(x

= 16a;.
7a.

8+9

J(&v -1)

-2)

= 3.c 2 -

25

.

^- +s

/,,,_- 5 , +3=

^:.


102

.

HIGHER ALGEBRA.

26.

7.-^hs±i_c»x x
\*j

W

.Y.
j

27.

J4x2 -7x-lb >/2^2 -9^ + 4 + 3
2

>Jx2 -3x=

Jx2 - 9.
J7x2 -6x-l = 0.
.

28.
29.

J%v-l = J2x2 + 2\x-l\.
2

V 2^ + 5^-7 + V3(a;
s/a
2

-7a; + 6) -

30.
31.

+ 2ax-3x2 - Ja 2 + ax-6x 2 = ^/Sa 2 + 3ao; - 9a;2

J2x2 + bx - 2 - V2#2 + 5a; - 9 = 1.
2 2 x/3^ - 2x + 9 + x/3.r -

32.
33. 34.

2o;

-4 = 13.
1.

V2^ 2 -7a;+l >/ 3^
2

./2a;2

-9a; + 4 =

- 7^ - 30 -

*/2o; 2

- 7x -

5

=

a?

-5

35.

o^ + a?-4o;2

+ a;+l = 0.
37.

36.

x*

+ ^x2 + l = 3x3 +3x.
9

8

3*+l-3(«s +#)=2tf*.

38.

10(o7t +l)-63a?(a;2

-l) + 52a;2 = 0.
a + 2a; +

x+J\2a-x _*Ja+\
x - <Jl2a-x
.

J'a2 - 4x2 _ bx
a
'

sJct-V
a;

a+2x- J a 2 - 4a;2 ~
1
1

41.

1

a;

+ sjx 2 - 1
,

-

^/.r 2

-

x - sjx 2 -

1

x + a/^ 2 |. 2

\

= 8x jx 2 - 3x + 2.
./». £".+ V # -l
2 a;

,—

-

42.

>/^+#I = Jtf-x
2

43.

44.

2*

:

2 2* = 8

:

1.

45.

a2 *(a 2 +

I)

= (a?* + a*)a.

46

^/a?-5
3a;-7
2
"

= V3a?-7
x-b
'

18 (7a; - 3)
'

_ 250 V^+T
"
3n/7^3

2a;+l

2

1

48.
49.

(a + a;) 3 + 4 (a
>/a;2

- *) = 5

(a

2

3 - a;2 )"

.

+ aa;-l - Jx2 + bx-l = J a - K/b.
2

50.

^B+-»-^El. ##+
V^'
1

8a

\/a'2

1

51.

.v

4

- 2.v3 + a; = 380.

52.

27^ + 2U- + 8 =

.

MISCELLANEOUS EQUATIONS.
136.

108
of

We
1.

shall

now

discuss

some simultaneous equations

two unknown
Example

quantities.
Solve
(z

x+2+y+S+ J(x + 2) (y + 3) = 39.
+ 2) 2 + (y + 3) 2 + (:r + 2)( / + 3)=741.
2
;

Put x + 2 = m, and y + 3 = v

then
(1),

u+v + Juv = Sd
w + v + wv = 741
2 2

(2),

hence, from (1) and

(2),

we obtain by

division,
(3)

u + v - Juv = 19

From
and
or

(1)

and

(3),

u+t?=29;

Juv = 10,
wv = 100;

whence
thus

w = 25,

or 4; v

= 4,

or 25

;

x = 23, or 2; y=l, or 22.
2.

Example

Solve

.r

4

+ y*= 82

(1),

ar-y=2
Put
then from
(2)

(2).

# = w + t>, and y = u- v;

we obtain
(1),

v
(w
.-.

= l.

Substituting in

whence
and

+ l) 4 + (u- 1) 4 = 82; 2(m4 + 6m2 + 1) = 82; u 4 + 6u 2 -40 = 0; w2 = 4, or — 10
;

u= ±2,
ysal, -3,

or

± >/~ 10*

Thus

x=s, -l, i± V^iO;

-li^-10.
10
(1),

JEa;ampZe3. e

Solve

f^ - —^ = 2A Sx-yx +
y 7x + 5y = 29
y*)
;

(2).

From

(1),

15

(2a;

2

+ Sxy + y* - 3z 2 + Axy - y-) = 38 (3.x 2 + 2xy
.-.

129o;

2

-29xy-38?/

2

= 0;

.-.

{Sx-2y)(iBx + 19y)=0.
Sx = 2y
(3),

Hence
or

43# = -19y

(1).

;

104
From

HIGHER ALGEBRA.
x
(3),

= y __7x + 5y =
3

2

29

= 1,
.-.

by equation

(2).

x = 2, y = 3.
y 7x + 5y - 82
29
equation
(2),

Again, from

x
(4),

19 ~ ^43 ~

= -gg, by
•*'

X~

551

~ 82 82 ,V

_ 1247
"

Hence
Example

x = 2, y = 3; or

x-

-

551

,

11

= -^-

1247

4.

Solve

4# 3 + 3a; 2 f/ + ?/ 3 =8,
2z 3 -2a; 2?/ + £?/ 2 = l.

Put y = mx, and substitute in both equations.
z 3 (4 + 3m + m 3 ) = 8 z 3 (2-2m + m2) = l
•*'

Thus
(1).

(2).

4 + 3m + m z _

2-2m + m2 ~

'

m3 -8»i9 +19m
that
is,

-12 = 0;

(/;i-l)(?/i-3)
.*.

(m-4) = 0;
3,

m=l,

or

or

4.
(2).

(i)

Take
(2),

m = l,

and substitute
# 3 = 1;

in either (1) or
.*.

From
and
(ii)

x = l;

y=mx=x=l.
Take

m = 3,

and substitute
5:r
3

in (2)
3
.*.

thus

= l;

x=

/l
k'->

\/
*/
3

and
(iii)

y = vix = 3x = 3
Take7?& = 4;

3/1

-.

we obtain
10.r 3 =l;
.-.

/ x=^-;
3

1

and

y = mx = 4x=4. /r^.

/I

MISCELLANEOUS EQUATIONS.
Hence the complete solution
is

L05

* =1,

V5'
s

v
4

To*

•"

=1
'

\/l>

\/^*
always be used when the

Note.

The ahove method

of solution

may

equations are of the same degree and homogeneous.

Example

5.

Solve

3 lx 2 y 2

-7y 4 - 112^ + 64 =

(1),
('2).

x 2 -7xy + 4y 2 + 8 =

From

(2)

we have -8 = x2 - Ixy + 4//'-

;

and, substituting in
)

(1),

3\x2 y 2 - 7# 4 +
.-.

Uxy

{x 2

- Ixy + Ay 2 +

{x 2

- Ixy + 4 j/")- =
)

;

31x-y 2 - 7 j/ 4
.-.

+ (x 2 - Ixy f Ay 2 ) (Uxy + x 2 - Ixy + Ay 2 =

;

Slx 2 y 2 -7y* + (x2 + 4y 2 ) 2 -(7xy) 2 = 0;

that

is,

s*-10sy+9y4 =0
.'.

(3).

(x2

-y 2 )(x 2 -9y 2 ) = 0;
ov

hence

x=±y,

x=

±3y.
(2),

Taking these cases in succession and substituting in x = y=±2;

we obtain

x=-y=± ^J
x=±3, y=±l\
3

-

>/-17'^ =T \/

-

yj

Note. It should be observed that equation (3) is homogeneous. The method here employed by which one equation is made homogeneous by a suitable combination with the other is a valuable artifice. It is especially useful in Analytical Geometry.

Example

6.

Solve

(x+yft+2

{x

- ?/)* = 3

{x 2

-

y*fi

(1).

3x-2y=13
i

(2).
i

i

Divide each term of

(1)

by (x 2 - y 2
i

)

,

or {x

+y)*
i

(x

-

r

-

y)

;

.

(

x +y\s

\x-yj

+2 (-~ \.v

y

+ gj

Y=3

106

HIGHER ALGEBRA.
i

This equation

is

a quadratic in
i

[x
(

+ v\ a
1 ,

from which we easily

find,

( ±y-Y = 2oTl; \x-yj
.'.

x

whence

^=8 x-y

or 1

;

7x = 9y, or y = 0.
(2),

Combining these equations with

we obtain
13

x=9, y = 7;

or x = -^,y=0.

EXAMPLES.
Solve the following equations
1.
:

X.

b.

3x-2y = 7,
xy = %).

2.

bx -y = 3,
2 2 y - 6# = 25.

3.

4^-3^ = 1,

12^ + 13y 2 = 25.
ocy

4.

a,

4

+ #y+2/ 4 = 931,

5.

x2 +

+3/

2

= 84,

x2 — xy +y2 = 19.
6.

x - *Jxy+y =6.
7.

x + Jxy +y = 65, #2 + #y

+y 2 =2275.
9.

x +y = 7 + \A?y, x2 +y 2 = l33-xy.
10.
3.r

8.

3#2 -5y 2 = 7,
Zxy - 4y 2 = 2.

5y 2 -7^ = l7,

bxy - 6x2 = 6.
12.

+ 165 = 16.ry, 7^y + 3y 2 = 1 32.
2

11.

3x2 +xy+y 2 = l5,
Zlxy - 3x2 -bf = 45.

#2 + y 2 -3 = 3.zy,
2x2 - 6 + y2 = 0.

13.

.r

4

+y 4 =706, x+y = 8.

14.

xA +y* = 272,

15.

^-y = 992,
5

x-y = 2.
17.

x-y = 2.
18.

16.

,r+i = l, y
e ?/+-=25.

£+£-£, x 2 y
3

|+t = 5. 2 5
2 5 5 - + - = 7;

4

= 1.
21.

19.

11
^3

x +y = 1072,

20.

11 xy^+yx^=20,
33
+y
^- 2 2

11
6(.i?

+y = ie.
3

=65.

11 =
2

# +y2 = 5,
2
2
)

+y

5.

.

.

MISCELLANEOUS EQUATIONS.
22.

107

Jx+y+J7-y = 4,

23.

y+

Jx 2 - 1 = 2,
+

24

.

JZ+JZ^,
tf*

25.

f~f
.

v^

7 = i_

,

26.

27. 28. 29. 30.

+ 4y 2 - 15* = 10 (3y - 8), xy = 6. 2 2 2 2 .r y + 400 = 41ay, y = 5.ry - 4.r
4i- 2
9. c 2

+ 5y=6+2Qay-25ya + 2.v,

7#-lly = 17.
,

+ 33.r -12 = 1 2xy - 4,y 2 + 22y
- y2 )
(.r

at?

- ovy = 18.

(.-v

2

- y) = 1 6a^,

(a4

- y4

)

(a3

- y 2 ) = 64Ga?y.

31
32.

2.v 2

- xy +y 2 =2y,
2
,

2x 2 + 4xy = 5^.
-o8
5.v

y—-% + +
(.r

y)

(#-y) 2

—% =

,

- 7y = 4.
-4xy + 2x 2 ) + 8 = 0.

33. 34.
35. 36. 37. 38.

2 y(y -3.r#-.r2 ) + 24 = 0,

x(j/ 2

3a-3

- 8ay2 + if + 2 1 = 0,

a2

(y

- x) = 1
108.

2 2 y (4v - 108) = x (x3 - 9y 3 ),

2x 2 + 9xy + y 2 =

6xi + x 2y 2 + l6 = 2x(\2x+y 3 ),

x2 + xy-y 2 = 4.
.

x (a + x)=y(b+y),
xy + «Z> = 2ax,
fir_

ax + by = (x + y) 2
.

xhf + a2 b 2 = 26 2y2
y—a

a

39.
40.

a2
6.v3

+ .lzi = _J b2 x—b
>

L

a—b

L

.

=0

.

= 10a
\y

2

6.r

+ 3a3y,

ay 3 = 10ab 2y + 3b 3x.
i

41.

2a(--'A+4a2 = 4:X2 +^-t
xj
2a

a*

Equations involving three or more unknown quantities can only be solved in special cases. We shall here consider some of the most useful methods of solution.
137.

Example

1.

Solve

x

+ y +z =13 .7^ + 2/2 + 22 = 65 xy = 10
(x

(1),
(2),

(3).

From
Put u

(2)

and

(3),
;

+ yf + * 2 = 85.
u*+z*=85.

for

x+y

then this equation becomes

108
Also from
(1),

HIGHER ALGEBRA.
u +z =13;
z

whence we obtain u = l or 6;

=6

or
7,1
\

7.

Thus we have
Hence the
solutions are

x+y=

#?/ = 10

and

£ + ?/ =

6,

acy = 10

x=5,
*

or

2,'|

.r

= 3db\/-l.,

y = 2, or

5,1

or

=6

y^W^T,
z=l.

5

J

Example

2.

Solve

(a;

+ y) {x + z) = 30, + z)(y + x) = 15,

{y

[z+x)(z+y) =18.
Write
m,
1;,

w

for

?/

+ 2,
viv

a

+

as,

a;

+y

respectively

;

thus
(1).

— 30,

tvu = 15,

mv = 18

Multiplying these equations together, we have

wVu>2 _ 30 x 15 x 18 = 15* x 6 2 .*. uvw = ±90.
Combining

;

this result with each of the equations in (1),

we have

u = 3, v = 6, w = 5\ or w = -3, v = -6,
.-.

w=-5;

y+z=-S,\ y + z=3,\ z + x = $, > or z+x = -d>,\ x + y = -5,i x + y = 5))

whence

ce=4, y = l, 2
3.

= 2;

or

x=-i, y=-l, «=-2.
(1),
(2),
:

Example

Solve

y

2 2

+ ys + 2 2 = 49
2

+ z:r + a; = 19 x* + xy + y 2 =39
2

(3).

Subtracting

(2)

from

(1)

y
that
is,

2

-x2 + z{y -«)=30;
(4).

(y-x){x + y + z) = 30
(1)

Similarly from

and

(3)

[z-x){x+y+z)*=10
Hence from
(4)

(5).

and

(5),

by division

y-*- 3 «-«
whence
y

.

= 3z-2x.

;

;

MISCELLANEOUS EQUATIONS.
Substituting in equation
(3),

10f)

we obtain

z*-8xa+8zs =13.

From

(2),

x 2 + xz

+

z~

= 19.
4, Art. 130,

Solving tbese homogeneous equations as in Example

we obtain

a;=±2,
or
jc= ±-ts,

z

= ± 3 and
;

therefore y

= ±5

2= ± -t^
.t

;

and therefore
y^
a;

y= T —-,
— xy = c 2
; .

Example

4.

Solve

2

-yz = a 2

,

-zx = 6 2

,

z2

Multiply the equations by
c
2
.r

y, 2,

respectively

and add

then
(1).

+ « 2 + & 2z =
//

Multiply the equations by

z, x,

y respectively and add

;

then
(2).

b2x

+ c-y + a*z =

From

(1)

and

(2),

by cross multiplication,
2

~^¥c = V^W =
k 2 (a 6

^W =
2
Z>

k su PP° se

'

Substitute in any one of the given equations

;

then

+ b 6 + c 6 - 3a2
z
2

2

c 2)

=1
1

x

11

a 4_^2c 2

^4_ c

ft2

C 4_

a 7/j

i

*Ja*+b*+c*-3a?tP<?

EXAMPLES.
Joh

X.

c.

110
9.

HIGHER ALGEBRA.
x*y*zhi=\%
aPy*z=12
t

#VW=8,
3

x*yz2 u2 =

l,

3xy 2z2 u 2 = 4.

10.
11.

^-

=54, .*% 32 2 = 72.
12.

ay+#+y=23, xz+x + z = 4l, yz + ij + z = 27.

2^-4?+2/ = 17, 3yz+y-6z = 52, §xz + 3s + 2#= 29.

13. 14. 15.
16.

xz+y^lz, yz + x=8z, x + y + z = l2.
.r

+y 3 + ^3 =a3 ^2 +y 2 + 22 = a2 # + # + s = a. ^2 +y 2 +22 =3/^ + 2^ + .«y = « 2 3.r-# + 2 = a*/3. #2 +y2 -M2 = 21a2 ys + ^-.ry = 6a 2 3x+y-2z = 3a.
3
,
,

,

,

,

Indeterminate Equations.
138.

Suppose the following problem were proposed for

solu-

tion

:

person spends .£461 in buying horses and cows; if each horse costs £23 and each cow £16, how many of each does he buy
1

A

?

Let

x,

y be the number

of horses

and cows respectively
461.

;

then

23a;

+ 16^ =

Here we have one equation involving two unknown quantities, and it is clear that by ascribing any value we please to x, we can obtain a corresponding value for y thus it would appear at first sight that the problem admits of an infinite number of solutions. But it is clear from the nature of the question that x and y must and with this restriction, as we shall see be positive integers later, the number of solutions is limited.
;

;

quantities is greater than the number of independent equations, there will be an unlimited number of solutions, and the equations are said to be indeterminate. In the present section we shall only discuss the simplest kinds of indeterminate equations, confining our attention to positive integral values of the unknown quantities ; it will be seen that this restriction enables us to express the solutions in a very
If the

number

of

unknown

simple form.

The general theory
in Chap. xxvi.

of indeterminate equations will be found

;

;

:

;

INDETERMINATE EQUATIONS.
Example
1.

Ill

Solve 7# +
7,

12j/

= 220

in positive integers.
;

Divide throughout by

the smaller coefficient

thus

x+

y+^ =31 + -;
x + y+^-~

.-.

=31

...

(1)

Since x anil y are to be integers,

we must have

5y-S
7

= integer
integer

;

and therefore
that

l%-9 =

is,

*-=-= %-l+ w-2 integer;
1/-2
:

and therefore
.-.

integer

=p

suppose.

or

y-2 = 7p, y = lp + 2
(1),

(2).

Substituting this value of y in
.r

that

is,

+ 7p + 2 + 5> + l = 31; x = 2§-l2p

.(3).

If in these results we give to p any integral value, we obtain corresponding integral values of x and y; but if p > 2, we see from (3) that x is negative and if p is a negative integer, y is negative. Thus the only positive integral values of x and y are obtained by putting p = 0, 1, 2.

The complete

solution

may

be exhibited as follows

p=
a:

0,

1,

2,

= 28, y= 2,
Note.
to

16,
9,

4,

16.

When we

obtained

5y-S
integer,

we multiplied by

3 in order

make

artifice

the coefficient of y differ by unity from a multiple of 7. A similar should always be employed before introducing a symbol to denote

the integer.

Example

2.

Solve in positive integers,

14x -

11//

= 29.

(1).

Divide by 11, the smaller coefficient; thus

x+

Sx
11

i/-2 + ir

;

3x-7
11

= 2 - x + y = integer

112
hence
that

HIGHER ALGEBRA.
12£ - 28

—^
g*

= m teg er
.

*>

is,

x - 2+
'— ^
Qfc

——^ = integer

;

.-.

—=
y

integer =_p suppose;

.*.

X = \\p + §

and, from

(1),

— 14p + 5 !

This

any and y

called the general solution of the equation, and by giving to p positive integral value or zero, we obtain positive integral values of x
is
;

thus we have

p = 0,
.t

1,

2,

3,

= 6, y = 5,

17, 28, 39, 19,

33, 47,

the

number
Example

of solutions being infinite.
3.

In

how many ways can £5
5^ + 4y = 200;
•••

be paid in half-crowns and florins?
;

Let x be the number of half-crowns, y the number of florins

then

x +y+\=
.

5 °;

.'

.

x 2 — integer = 2^ suppose

;

.*.

x=4p,
y

and

= 50-5p.

Solutions are obtained by ascribing to p the values 1, 2, 3, ...9; and If, however, the sum may be paid either therefore the number of ways is 9. in half-crowns or florins, p may also have the values and 10. If ^ = 0, then x = 0, and the sum is paid entirely in florins if p = 10, then y = 0, and the sum is paid entirely in half-crowns. Thus if zero values of x and y are admissible the number of ways is 11.
;

Example
each

4.

The expenses
5s.,

paid there of each?

man

each

woman

of a party numbering 43 were £5. 14s. Qd. if 2s. 6d., and each child Is., how many were
;

Let x, y, z denote the number of men, women, and children, respectively; then we have x + y + z= 43 (1),
10.r

Eliminating

z,

we obtain

+ 5?/ + 2z = 229. 8x + By = 143.
is

The general

solution of this equation

x=Sp + l, y = 45-8p;

:

:

INDETERMINATE EQUATIONS.
Hence by substituting
in
(1),

113

we obtain z = 5p-3.

Here p cannot be negative or from 1 to 5. Thus

zero, but

may have
4,

positive integral values

p= x-

1,

2,

3,

5;

4,

7, 10, 13,

16;
5;

*

y = 37, 29, 21, 13,

2=2,

7, 12, 17, 22.

EXAMPLES.
Solve in positive integers
1.
3.i

X.

d.

+ 8y = 103.

2.

5#+2y=53.
23a?+25y=915.

3.
6.

7.>;+

12y=152.

4.

l&P+lly=414

5.

4L>;

+ 47y = 2191.
least values

Find the general solution in positive integers, and the
of

x and y which
7.

satisfy the equations
8.

5.v-7y = 3.

6a?-13y=l.
19y-23a?=7.

9.

8#-2ty=33.
77y-3Qa?=295.
;

10.

I7y-13#=0.

11.

12.

buying horses and cows if each horse costs £37 and each cow £23, how many of each does he buy ?
13.
14.

A farmer spends £752 in
In

how many ways can £5 be paid
?

in shillings

and sixpences,

including zero solutions
15.

Divide 81 into two parts so that one
5.

may

be a multiple of 8

and the other of
16.

What

is

the simplest

way

for a person

who has only guineas
?

to

pay

105. 6d. to

another

who has

only half-crowns

Find a number which being divided by 39 gives a remainder 17. and by 56 a remainder 27. How many such numbers are there ?
18.

16,

the smallest number of florins that must be given to discharge a debt of £1. (5s. 6d., if the change is to be paid in half-crowns
is

What

only?
Divide 136 into two parts one of which when divided by 5 leaves remainder 2, and the other divided by 8 leaves remainder 3.
19.

20.

I
:

buy 40 animals consisting
spend £301, how
pocket
I

at

£17

if I

many

of rams at £4, pigs at £2, of each do I buy ?

and oxen

have 27 coins, which are sovereigns, half-crowns and the amount I have is £5. 05. 6d. how many coins of each sort have I ?
In 21. or shillings,
;

my

H. H. A.

8

;

;

CHAPTER XL
Permutations and Combinations.
Each of the arrangements which can be made by taking 139. some or all of a number of things is called a permutation.
Each of the groups or selections which can be made by taking some or all of a number of things is called a combination.
Thus the •permutations which can be made by taking the letters a, b, c, d two at a time are twelve in number, namely,
ab,
ac,

ad,
da,

be, cb,

bd,

cd,

ba,

ca,

db,

dc
letters.

each of these presenting a different arrangement of two

a,

The combinations which can be made by taking the namely, b, c, d two at a time are six in number
:

letters

ab,

ac,

ad,

be,

bd,

cd;
letters.

each of these presenting a different selection of two

forming combinations we are only concerned with the number of things each selection contains whereas in forming permutations we have also to consider the order of the things which make up each arrangement; for instance, if from four letters a, b, c, d we make a selection of three, such as abc, this single combination admits of being arranged in the

From

this it appears that in

following

ways

:

abc,

acb,

bca,

bac,

cab,

cba,

and

so gives rise to six different permutations.

PERMUTATIONS AND COMBINATIONS.
140.

115
of
this

Before

discussing

the

general

propositions

chapter there is an important principle which we explain and illustrate by a few numerical examples.

proceed to

If one operation can be performed in m ivays, and (when it has been performed in any one of these ways) a second operation can then be performed in n tvays ; the number of ways of performing the two operations ivill be m x n.
performed in any one way, we can associate with this any of the n ways of performing the second operation and thus we shall have n ways of performing the two operations without considering more than one way of performing
If the first operation be
:

the

first;

and

so,

corresponding to each of the

m

ways

of per-

forming the first operation, we shall have n ways of performing the two; hence altogether the number of ways in which the two operations can be performed is represented by the product

m x n.
There are 10 steamers plying between Liverpool and Dublin; in how many ways can a man go from Liverpool to Dublin and return by a different steamer?

Example

1.

There are ten ways of making the first passage and with each of these there is a choice of nine ways of returning (since the man is not to come back by the same steamer) hence the number of ways of making the two journeys is 10 x 9, or 90.
;

;

extended to the case in which there are more than two operations each of which can be performed in a given number of ways.
This principle
easily be

may

Example
hotels;

2.

in how different hotel?

Three travellers arrive at a town where there are four many ways can they take up their quarters, each at a

The first traveller has choice of four hotels, and when he has made his selection in any one way, the second traveller has a choice of three ; therefore the first two can make their choice in 4 x 3 ways and with any one such choice the third traveller can select his hotel in 2 ways hence the required number of ways is 4 x 3 x 2, or 24.
; ;

141.

To find

the

number of permutations of

\\

dissimilar things

taken r at a time.

we

This is the same thing as finding the number of ways in which can fill up r places when we have n different things at our

disposal.

The
things

first

place

may

be

may be tilled up in n ways, for any one of the n taken when it has been filled up in any one of
;

8—2

116

HIGHER ALGEBRA.

these ways, the second place can then be filled up in n - 1 ways ; and since each way of filling up the first place can be associated with each way of filling up the second, the number of ways in which the first two places can be filled up is given by the product n (n - 1). And when the first two places have been filled up in any way, the third place can be filled up in h — 2 ways. And reasoning as before, the number of ways in which three places can be filled up is n (n - 1) (n - 2).

Proceeding thus, and noticing that a with each new place filled up, and that of factors is the same as the number of have the number of ways in which r equal to

factor is introduced at any stage the number places filled up, we shall

new

places can be filled

up

n (n- l)(n—
and the r th factor
is

2)

to r factors

;

n — (r—

1),

or

n — r+1.
n
1).

Therefore the number of permutations of a time is

things taken r at

n{nCor. a time
is

1)

(n-

2)

(n-r +

The number

of permutations of

n things taken

all

at

or
It
is

n (n - 1) (?i - 2) n(n — Y)(n—2)
Also n\
is

to

n

factors,

3.2.1.
\n,

usual to denote this product by the symbol

which

is

read "factorial n."
142.
of

sometimes used for

\n.

We

shall in future denote the

number
n

of permutations

n things taken

r at a time

by the symbol

P

r

,

so that

"Pr = w(w-l)(w-2)
also

(n-r + 1);

"P =

\n.

In working numerical examples it is useful to notice that the n suffix in the symbol Pr always denotes the number of factors in the formula we are using.
143.

The number
also be

of permutations of

n things taken

r at

a time

may

found in the following manner.

Let "Pr represent the number of permutations of n things taken r at a time.

PERMUTATIONS AND COMBINATIONS.

117
1

Suppose we form all the permutations of n things t;iken r — at a time ' the number of these will be "Pr—l
;
.

each of these put one of the remaining n — r + 1 tilings. Each time we do this we shall get one permutation of u things r at a time; and therefore the whole number of the permutations n of n things r at a time is x (n - r + 1) that is, r_

With

P

;

]

By

writing

r—l

for r in this formula,
1

we obtain

"P_ = '^
similarly,

r

_2

x(n-r-f2),
x (n

'P

= 'Pr _a

-r+

3),

"P^P.x (71 -I),
"P
x

=7l.

Multiply together the vertical columns and cancel like factors from each side, and we obtain
n

P

r

= n(n-l)(n-2)

(n-r+l).
in

Example
seats
;

in

Four persons enter a railway carriage how many ways can they take their places ?
1.

which there are

six

and then the second person seat himself in 6 ways in 5 ; the third in 4 and the fourth in 3 and since each of these ways may be associated with each of the others, the required answer is 6x5x4x3, or 360.
first

The

person

may
;

;

;

Example

2.

How many different numbers
1, 2, 3, ...9?

can be formed by using six out

of the nine digits

Here we have 9 different things and we have to find the number of permutations of them taken 6 at a time
;

.

*

.

the required result = 9 P6

=9x8x7x6x5x4
= 60480.
144.
tilings

To find taken r at a

the

number of combinations of n

dissimilar

time.

Let "Cr denote the required number of combinations.

Then each
|r

of these combinations

consists of

dissimilar things which

can be arranged among

a group of r themselves in

ways.

[Art. 142.]

118

HIGHER ALGEBRA.
\r

Hence "C r x
things taken

is

equal to the
;

number

of arrangements of

n

rata
r
|

time

that

is,

*C x\r = "Pr

=n
_
Cor.

(n — 1) (n

2)

.

.

.

(n - r + 1)

;

tt(w-l)(w-2)...(w-r+l)
|r
V
'"

n This formula for C r may also be written in a different form ; for if we multiply the numerator and the denominator by \n — r we obtain

n (n -

1) (n

-

2)
\r

...

{n

-r+

1) x

n—r
\

n—r

The numerator now numbers from n to 1
;

consists of the product of all the natural

\n
.'.

"C r =

.

~

(2).

It will be convenient to remember both these expressions for C r using (1) in all cases where a numerical result is required, and (2) when it is sufficient to leave it in an algebraical shaj)e.
n
,

Note.

If in

formula

(2)

we put r = n, we have
\n n
i

~jn|_0" |0'
is to

but n Cn =l, so that if the formula be considered as equivalent to 1.

be true for r = n, the symbol 10 must

Example. From 12 books in how many ways can a selection of 5 be made, (1) when one specified book is always included, (2) when one specified book is always excluded ?
Since the specified book is to be included in every (1) have only to choose 4 out of the remaining 11.
selection,

we

Hence the number

of

ways = n C4
1 1x10 x9x8 ~ 1x2x3x4

= 330.

;

PERMUTATIONS AND COMBINATIONS.
(2)

119

Since the specified book is always to be excluded, we have to select the 5 books out of the remaining 11.

Hence the number

of

ways = n C6
_ 11x10x9x8x7

1x2x3x4x5
= 462.
145.

equal

to

The number of combinations of n things r at a time the number of combinations of\\ things n — r at a time.

is

In making all the possible combinations of n things, to each group of r things we select, there is left a corresponding group of

n-r

things

;

r at a time is n — r at a time

that is, the number of combinations of n things the same as the number of combinations of n things
.-.

"C r =

n

C

n—

.

r

The proposition may
"0

also be

proved as follows
\n =—

:

_r

=

n—r n—

(n

- r)

[Art. 144.1

n
n—
r

r

Such combinations are
Note.

called complementary.

Put r=w, then
result

tt

C = n Cn =l.
just proved
is

The

we have
of 14

useful in enabling us to

abridge arithmetical work.
Example.

Out

men

in

how many ways can an

eleven be chosen?

The required number = 14 C U

14 x 13 x 12

1x2x3
= 364.
If

we had made

use of the formula

uC

n we should have had
,

expression whose numerator and denominator

to reduce each contained 11 factors.

au


120
146.

.

HIGHER ALGEBRA.
Tojind
the

number of ways in which m + n things can be divided into two groups containing in and n things respectively. This is clearly equivalent to finding the number of combinations of ra + n things ra at a time, for every time we select one group of ra things we leave a group of n things behind. Ira + n Thus the required number = h= 1
ra
\7b

Note.
different

If

n = m, the groups are equal, and in
of subdivision is -

this case the
it

number
is

of

ways

~ —\2m —

[9

;

for in

any one way

possible

to interchange the two groups without obtaining a

new

distribution.

147.

To jind the number of ways in which
groups containing m,
n,

m + n + p things can
severally.

be divided into three

p things

First divide ra + n + p things into two groups containing and n + p things respectively the number of ways in which this
:

m

\m +

can be done

is

-r=

\m

n+p n+p

Then the number of ways in which the group of n+p things can be divided into two groups containing n and p things respec-

\n+p
tively
is

Hence the number
groups containing m,

of

n p ways in which the subdivision into three
things can be

n,

p

m + n+p
in

n+p
x
, ,

made is \m + n + ]>
Ira

,
5

or
\n \p

n+p

\n \p
J3wi

Note.

If

we put ?i=p = m. we obtain

:

—r=-|—

;

hut this formula regards

as different all the possible orders in which ~th.e three groups can occur in any one mode of subdivision. And since there are 13 such orders cor-

responding to each mode of subdivision, the number of different ways in
|3ot

which subdivision into three equal groups can be made
Example.

is

771

r^fm
771

r^ |3

The number
,

of
---

ways in which 15
;

recruits can be divided into

115

three equal groups is

-

and the number

of

ways in which they
15
I

can be drafted into three different regiments,

five into each, is -_—Hr [6 5 [6
J

PERMUTATIONS AND COMBINATIONS.
148.

121

In the examples which follow it is important to notice that the formula for 'permutations should not be used until the suitable selections required by the question have been made.
Example
1.

be formed; in
(1)

From 7 Englishmen and 4 Americans a committee of is to how many ways can this be done, (1) when the committee conAmericans ? choose 2 Americans and 4 Englishmen.
;

tains exactly 2 Americans, (2) at least 2

"We have to

The number of ways in which the Americans can be chosen is 4 C, and the number of ways in which the Englishmen can be chosen is 7 C4 Each of
.

the

first

groups can be associated with each of the second

;

hence

the required

number

of

ways = 4 C 2 x 7 C4

=

li
|~2"[2

X

\1

TTJ3

17
'J^

= 210.
2, 3,

|2|2|3
(2)

The committee may contain

or 4 Americans.

"We shall exhaust all the suitable combinations by forming all the groups containing 2 Americans and 4 Englishmen then 3 Americans and 3 Englishmen; and lastly 4 Americans and 2 Englishmen.
;

The sum of the three results will = *C2 x 7 C4 + 4 C3 x number of ways
17
[2
1

give the answer. 4 7 (7 C4 x 7 C, 3 +
|4 17

Hence the required
17

2

X TTT^
[4
j_3

+

TK X rl ^ + 1 X j_3 |3|4 [2)5

= 210 + 140 + 21 = 371.
In this Example we have only to make use of the suitable formulae for combinations, for we are not concerned with the possible arrangements of the members of the committee among themselves.

Example

made each

2. Out of 7 consonants and 4 vowels, containing 3 consonants and 2 vowels?

how many words can
,

be

of ways of choosing the three consonants is 7 C3 and the of choosing the 2 vowels is *Ca and since each of the first groups can be associated with each of the second, the number of combined groups, each containing 3 consonants and 2 vowels, is 7 C 3 x 4 C2
; .

The number number of ways

Further, each of these groups contains 5 letters, which among themselves in [5 ways. Hence
the required

may

be arranged

number

of words = 7 C 3 x

4

C2 x

Jo

~|3|4

X [2]2

X

"

= 5x|7
r

= 25200.

122
Example
3.

HIGHER ALGEBRA.
How many

words can be formed out of the letters article, so the even places? that the vowels occupy

Here we have

to put the 3 vowels in 3 specified places,
;

nants in the 4 remaining places

the

first

and the 4 consooperation can be done in 3 ways,
1

and the second in
the required

1

4

.

Hence

number

of words

=|3x[4

= 144.
In this Example the formula for permutations is immediately applicable, because by the statement of the question there is but one way of choosing the vowels, and one way of choosing the consonants.

EXAMPLES

XI.

a.

In how many ways can a consonant and a vowel be chosen out of 1. the letters of the word courage?

There are 8 candidates for a Classical, 7 for a Mathematical, and In how many ways can the 4 for a Natural Science Scholarship. Scholarships be awarded?
2.
3.

Find the value of

8

P

25
7,

P

24
5
,

<74

19
,

CU

.

4.

How many

different

arrangements can be made by taking 5

of the letters of the
5.

word equation ?

If four times the number of permutations of n things 3 together is equal to five times the number of permutations of n — 1 things 3 together, find n.

permutations can be made out of the letters of the word triangle? How many of these will begin with t and end with e ?
6.

How many

different selections can be made by taking four of many different numbers can be formed the digits 3, 4, 7, 5, 8, 1 ? with four of these digits ?
7.

How many
If 2n C3

How
3,

8.

:

n Oj

= 44

:

find n.

9.

10.

How many changes can be rung with a peal of 5 bells ? How many changes can be rung with a peal of 7 bells, the tenor
?

always being last

On how many nights may a watch of 4 men be drafted from a 11. crew of 24, so that no two watches are identical ? On how many of these would any one man be taken?
arrangements can be made out of the w ord draught, the vowels never being separated ?
12.
r

How many

letters of the

PERMUTATIONS AND COMBINATIONS.
13.

1

23
;

how many committees can
and 3 aldermen
14.
?

In a town council there are 25 councillors and 10 aldermen be formed each consisting of 5 councillors
of the letters A, B, C, p, q, r how many arrangements can (1) beginning with a capital, (2) beginning and ending with a

Out
]

be made
capital
15.

Find the number of combinations of 50 things 4G at a time.
If n

16.
17.

C12 = n Cs

,

find n C17 ,

22

<7n .

In
if

arranged,
18.

the letters of the word vowels be the letters oe can only occupy odd places
]

how many ways can

From
(1)

chosen
officer?

4 officers and 8 privates, in how many ways can 6 be to include exactly one officer, (2) to include at least one

In how 10 persons
19.
?

many ways can
,

a party of 4 or more be selected from

20.

If

™Cr = ls Cr + 2

find'<75

.

Out of 25 consonants and 5 vowels how many words can be 21. formed each consisting of 2 consonants and 3 vowels ?

many ways
23.

In a library there are 20 Latin and 6 Greek books; in how can a group of 5 consisting of 3 Latin and 2 Greek books be placed on a shelf ?
22.

In how
?

many ways

can 12 things be divided equally among 4

persons

From 3 capitals, 5 consonants, and 4 vowels, how many words 24. can be made, each containing 3 consonants and 2 vowels, and beginning with a capital ?
election three districts are to be canvassed by 10, 15, and 20 men respectively. If 45 men volunteer, in how many ways can they be allotted to the different districts ?
25.

At an

4 Latin and 1 English book be placed on a shelf so that the English book is always in the middle, the selection being made from 7 Latin and 3 English books?
26. 27.

In how

many ways can

A

boat

is

to be

manned by

on bow side and 1 can only row the crew be arranged ?
28.

eight men, of whom 2 can only rowon stroke side; in how many ways can

There are two works each of 3 volumes, and two works each of 2 volumes in how many ways can the 10 books be placed on a shelf so that volumes of the same work are not separated ?
;

29.

In

how many w ays can 10 examination papers
r

be arranged so

that the

befit

and worst papers never come together?

124
30.

HIGHER ALGEBRA.

men,
not

An eight-oared boat is to be manned by a crew chosen from 11 of whom 3 can steer but cannot row, and the rest can row but canif

steer.

men

In how many ways can the crew be arranged, can only row on bow side?

two of the

Prove that the number of ways in which p positive and n negative signs may be placed in a row so that no two negative signs shall be together is p + 1 Cn
31.
.

32. 33.

If

56

Pr +

54
6
:

Pr + = 30800
3

:

1,

find

r.

different signals can be made by hoisting 6 differently coloured flags one above the other, when any number of them may be hoisted at once ?

How many

34.

U^C

24
:

2r

C2r _ 4 = 225

:

11, find

r.

Hitherto, in the formulae we have proved, the things have been regarded as unlike. Before considering cases in which some one or more sets of things may be like, it is necessary to point out exactly in what sense the words like and unlike are When we speak of things being dissimilar, different, unused. like, we imply that the things are visibly unlike, so as to be On the other hand we easily distinguishable from each other. shall always use the term like things to denote such as are alike For to the eye and cannot be distinguished from each other. consonants and the vowels may instance, in Ex. 2, Art. 1-48, the be said each to consist of a group of things united by a common characteristic, and thus in a certain sense to be of the same kind; but they cannot be regarded as like things, because there is an individuality existing among the things of each group which makes them easily distinguishable from each other. Hence, in the final stage of the example we considered each group to consist of five dissimilar things and therefore capable of [5
149.

arrangements among themselves.
150.

[Art. 141 Cor.]

Suppose we have to find all the possible ways of arranging 12 books on a shelf, 5 of them being Latin, 4 English, and the remainder in different languages.
be regarded as belonging to one class, united by a common characteristic ; but if they were distinguishable from each other, the number of permutations would be )12, since for the purpose of arrangement among themin each language
selves they are essentially different.

The books

may

'

PERMUTATIONS AND COMBINATIONS.
If,

125

however, the books in the same language are not distinguishable from each other, we should have to find the number of ways in which 12 things can be arranged among themselves, when 5 of them are exactly alike of one kind, and 4 exactly alike, of a second kind a problem which is not directly included in any of the cases we have previously considered.
:

To find the number of ways in which n arranged among themselves, taking them all at a
151.

things
time,

may

be

when p

of

of one kind, q of them exactly alike of another kind, r of them exactly alike of a third kind, and
the things are exactly

alike

the rest all different.

Let there be n letters suppose p of them to be to be b, r of them to be c, and the rest to be unlike.
;

a,

q of them

Let x be the required number of permutations ; then if in any one of these permutations the_p letters a were replaced by p unlike letters different from any of the rest, from this single permutation, without altering the position of any of the remaining letters, we could form p new permutations. Hence if this change were made in each of the x permutations we should obtain x x \p
I

permutations.
Similarly,
if

the q letters b were replaced by q unlike letters,

the

number

of permutations

would be
x \p x
|<7.

x
In
like

manner, by replacing the r
\q x

letters c
\r

by

r unlike letters,

we

should finally obtain x x \p x

permutations.
of \n

But the things are now all different, and therefore admit permutations among themselves. Hence
x x
\p

x

\q

x

r-

\n;

that

is,

x—
the required

~.

—r
p

\p \g

which

is

number

of permutations.
all

Any

case in which the things are not

different

may

be

treated similarly.

126
Example
letters
1.

HIGHER ALGEBRA.
permutations can be made out of the of the word assassination taken all together ?

How many

different

We

have here 13

letters of

which 4 are

s,

3 are a, 2 are

i,

and 2 are

n.

Hence the number of permutations

~|^[3|2j£

= 13.11.10.9.8.7.3.5 = 1001 x 10800 = 10810800.
Example
2.

How many
so

1, 2, 3, 4, 3, 2, 1,

numbers can be formed with the that the odd digits always occupy the odd places?
can be arranged in their four places in

digits

The odd

digits 1, 3, 3, 1

way s

(1) -

l^2
The even
digits 2, 4, 2

can be arranged in their three places in
13

y^

ways

(2).

Each

of the

ways

in

(1)

can be associated with each of the ways in
14 13
-j^

(2).

Hence the required number = y^=x- x

= 6 x 3 = 18.
things r at a

To find the number of permutations of n time, when each thing may be repeated once, twice, times in any arrangement.
152.

up

to

r

Here we have

to consider the

number

of

ways in which r

places can be filled disposal, each of the

up when we have n different things at our n things being used as often as we please in

any arrangement. The first place may be filled up in n ways, and, when it has been filled up in any one way, the second place may also be filled up in n ways, since we are not precluded from using the same thing again. Therefore the number of ways in which the first The third place can two places can be filled up iswxn or n2 also be filled up in n ways, and therefore the first three places in
.

n3 ways.
Proceeding in this manner, and noticing that at any stage the index of n is always the same as the number of places filled up, we shall have the number of ways in which the r places can be
filled

up equal to n r

.

PERMUTATIONS AND COMBINATIONS.

127

Example. In how many ways can 5 prizes be given away to 4 boys, when each boy is eligible for all the prizes?

Any one of the prizes can be given in 4 ways; and then any one of the; remaining prizes can also be given in 4 ways, since it may be obtained by the boy who has already received a prize. Thus two prizes can be given away in 4a ways, three prizes in 4 ways, and so on. Hence the 5 prizes can be given away in 4 5 or 1024 ways.
:! ,

153.
to

To find
tiling

the total

number of ways in which
some or
all

it is

possible

make a
Each

selection by taking

of

\\

things.

be dealt with in two ways, for it may either be taken or left; and since either way of dealing with any one thing may be associated with either way of dealing with eacli one of the others, the number of selections is

may

2x2x2x2
But
This
of
this includes the case in

to

n

factors.
all

which

the things are
of

left,

therefore, rejecting this case, the total
is

number

ways

is

2"-l.

often spoken of as "the total

number

of combinations"

n

things.

has 6 friends more of them to dinner?
Example.

A man

;

in

how many ways may he
;

invite one or

He
ways

has to select some or
2s

all of his 6 friends

and therefore the number of

is

-

1,

or 63.

This result can be verified in the following manner.

The guests may be invited singly, in twos, threes, number of selections = 6 C1 + 6 C2 + 6 C3 + 6 C4 + 6 C5 + <>C6

;

therefore the

= 6 + 15 + 20 + 15 + 6 + 1 = 63.
154.

To find for what value of
r at

r the

number of combinations

of

n things
Since

a time
=

is greatest.
1

"C

^( ?l -

)( n

- 2) 1.2.3

(w-r + 2)(n-r + (r-l)r

l)

,

_

n(n-l)(n-2)
1.

(w-r +
(r-1)
1

2)

2.3
r+

"C

= nC

.

x

n—

r
factor

The multiplying
which shews that
it

may be

written

—-

-

1,

decreases as r increases.

Hence

as r receives


128
the values
71 4- 1

;

HIGHER ALGEBRA.
1, 2,

3

in succession,
1

n

G

r

is

continually increased
1.

until

1

becomes equal to

or less than

r

Now
r
so long as
i

1^1,
71+1
r

>z >

^
;

that

is,

——

-

r.

We
(1)

have to choose the greatest value of r consistent with

this inequality.

Let n be even, and equal to 2m; then

n+

-2
?•

1

2m +1

2—

+

1

s

;

and

for all values of

up to
,

?n inclusive this is greater

than

r.

Hence by putting
combinations
(2)
is

r
.

m= —

we

find that the greatest

number

of

"Cn
2

Let n be odd, and equal to

— =-5—
n+
1

2m +

1
-

;

then

2m +

2

» + li
is

and

for all values of r

but when r -

m+ 1 *C.= n Cm
mi+I

greater than r the multiplying factor becomes equal to 1, and
inclusive this
: '

up to

m

that

is, '

"C n+ +
1

n

C7i—l
2

J

2

and therefore the number
things are taken

of combinations is greatest

when the

,

or

—^—

at a time; the result being the

same

in the

two

cases.

The formula for the number of combinations of n things r at a time may be found without assuming the formula for the numbes of permutations.
155.

Let "C r denote the number of combinations of n things taken r at a time; and let the n things be denoted by the letters
a, b, c, d,

PERMUTATIONS AND COMBINATIONS.
Take away a; then with
"~
X

120

C
n
is

eaeli

of

remaining letters we cm form combinations of n— 1 letters taken r - 1 at a time. With of these write a; thus we see that of the combinations tilings r at a time, the number of those which contain
tin
1

a

w~ l

b is

n~

C
,

x
:

x

C

the number of those and so for each of the n letters.
\

similarly

which

contain

Tlierefore

n

x "~*Cr _ l is equal to the

number

of combinations
b,

r at a time which contain a, together with those that contain those that contain c, and so on.

in this manner, each parFor instance, if r=3, the ticular one will be repeated r times. combination abc will be found anions; those containing a, amonir those containing b, and among those containing c. Hence

But by forming the combinations

*c= n - cr— x-.
x
r

r

i 1

.,

By

writing

u—

1

and r — 1 instead

of
ni

n and
1

r respectively,

r-l°
Similarly,

-V^ = ^G _ x »-2
r

z

n— r + 2/~1
2

_n-r + \ri ^i

U—
>.

T
2

+

o
;

and

finally,

n - r+1

C = »-r + 1.
1

Multiply together the vertical columns and cancel like factors from each side thus
;

"C

.

n (rc-l)(n-2)

(n-r+
1

l)

r(r-l)(r-2)
the total

156.
to

To find
selection

make a

number of ways in which by taking some or all out qfip +

it is

-possible

c

x

+r +

are alike of one kind, q alike of a second kind, r alike of a third kind; and so on.
tilings, ivJierenf-p

The p things may be disposed of in p + 1 ways for wo may take 0, 1, 2, 3, p °f thorn. Similarly the q things may be disposed of in q + \ ways; the r things in r+1 ways; and
;

so on.
H.
II.

A.

9

:

130

HIGHER ALGEBRA.
Avays in which all the tilings
1) (r

Hence the number of disposed of is (^ + 1) (q +
But
taken
;

may be

+

1)

this includes the case in therefore, rejecting this

which none
case,

of the things are

the

total

number

of

ways

is

(jp

+ l)fe+l)(r +

.l)

-1.

general formula expressing the number of permuta157. tions, or combinations, of n things taken r at a time, when the

A

things are not all different, may be somewhat complicated particular case may be solved in the following manner.

;

but a

Example. Find the number of ways in which (1) a selection, (2) an arrangement, of four letters can be made from the letters of the word
proportion.

There are 10

letters of six different sorts,

namely o, o,o; p,p;

r,

r; t; i; n.

In finding groups of four these
(1)
(2)
(3)

may

be classified as follows

Three

alike,

one

different.

Two Two

alike,
alike,

two others alike. the other two different.

(4) (1)

All four different.
selection can be made in 5 ways for each of the five letters, can be taken with the single group of the three like letters o.
;

The
n,

p,

r, t, is

selection can be of the three pairs o, o; p, p;
(2) (3)

The

made
r, r.

in 3 C2 ways for we have to choose This" gives 3 selections.
; ;

two out

This selection can be made in 3 x 10 ways 3 pairs, and then two from the remaining 5 letters.
(1)

for

we

select

one of the

This gives 30 selections.

letters

This selection can be made in 6 C4 ways, as we have to take 4 different This gives 15 selections. to choose from the six o, p, r, t, i, n.
total

Thus the

number

of selections

is

5

+ 3 + 30 + 15

;

that

is,

53.

all

In finding the different arrangements of 4 letters possible ways each of the foregoing groups.
gives rise to 5 x

we have

to

permute in

(1)

=

,

or 20 arrangements.

(2)

gives rise to 3 x -^=^

,

or 18 arrangements.

(3)

gives rise to 30 x

-=,

or 360 arrangements.
or 3G0 arrangements.

(4)

gives rise to 15 x j4

,

Thus

the total

number

of arrangements is 20

+ 18 + 360 + 360;

that

is,

758.

PERMUTATIONS AND COMBINATIONS.

131

EXAMPLES.
1.

XI.

b.

letters

Find the number of arrangements that can he made out of the of the words
(1)

independence,
(3)

(2)

superstitious,

institutions.
if

can 17 billiard balls be arranged, them are black, 6 red, and 4 white %
2.

In how

many ways

7 of

room is to be decorated with fourteen flags ; if 2 of them are 3. blue, 3 red, 2 white, 3 green, 2 yellow, and 2 purple, in how many ways
can they be hung?
4.

A

How many numbers
2, 3, 0, 3, 4, 2,

greater than a million can be formed with

the digits
5.

3?

Find the number of arrangements which can be made out of the letters of the word algebra, without altering the relative positions of vowels and consonants.

On three different days a man has to drive to a railway station, 6. and he can choose from 5 conveyances in how many ways can he make
;

the three journeys
7.

?

have counters of n different colours, red, white, blue, in I make an arrangement consisting of r counters, supposing that there are at least r of each different colour ?
I
;

how many ways can

In a steamer there are stalls for 12 animals, and there are 8. cows, horses, and calves (not less than 12 of each) ready to be shipped; in how many ways can the shipload be made?
9.

In
is

there

how many ways can n things be given to p persons, when no restriction as to the number of things each may receive ?
In how

10.

many ways

can

five

things be divided between two

persons
11.

?

How many different arrangements can be made out of tl ie letters
a z b 2 c* when written at
full

in the expression

length?

letter lock consists of three rings each marked with fifteen 12. different letters find in how many ways it is possible to make an unsuccessful attempt to open the lock.
;

A

13.

Find the number of triangles which can be formed by joining

three angular points of a quindecagon.
library has a copies of one book, b copies of each of two 14. books, c copies of each of three books, and single copies of d books. In how many ways can these books be distributed, if all are out at once I
15.

A

How many

numbers
7

less
?

than 10000 can be made with the

eight digits
16.

1, 2, 3, 0, 4, 5, 6,

In how many ways can the following prizes be given away to a class of 20 boys: first and second Classical, first and second Mathematical, first Science, and first French ?

9—2

132
17.

HIGHER ALGEBRA.

A telegraph has 5

positions, including the position of rest signals that can be made ?
18.

arms and each arm is capable of 4 distinct what is the total number of
;

In how many ways can 7 persons form a ring? In how many ways can 7 Englishmen and 7 Americans sit down at a round table, no two Americans being together?
possible to draw a sum of money from a bag containing a sovereign, a half-sovereign, a crown, a florin, a shilling, a penny, and a farthing?
19.

In how

many ways

is it

3 cocoa nuts, 4 apples, and 2 oranges, how tions of fruit can be made, taking at least one of each kind
20.

From

many
?

selec-

21.

Find the number of different ways of dividing
groups.

mn

things into

n equal

How many signals can be made by hoisting 4 flags of different 22. colours one above the other, when any number of them may be hoisted at once ? How many with 5 flags ?
Find the number of permutations which 23. the letters of the word series taken three together ?
can be formed out of

There are p points in a plane, no three of which are in the same 24. straight line with the exception of q, which are all in the same straight line; find the number (1) of straight lines, (2) of triangles which result from joining them.
points in space, no four of which are in the same plane with the exception of q, which are all in the same plane; find how many planes there are each containing three of the points.
25.

There are

p

There are n different books, and p copies of each; find the number of ways in which a selection can be made from them.
26.

Find the number of selections and of arrangements that can be made by taking 4 letters from the word expression.
27.

28. letters of the 29.

How many

permutations of 4 letters can be made out of the
all

word examination ?
1, 3, 5, 7, 9,

Find the sum of
Find the sum of

using the digits
30.

numbers greater than 10000 formed by no digit being repeated in any number.

numbers greater than 10000 formed by using the digits 0, 2, 4, 6, 8, no digit being repeated in any number. If of p + q + r things p be alike, and q be alike, and the rest 31. different, shew that the total number of combinations is r (p + l)(q+l)2 -l. Shew that the number of permutations which can be formed 32. from 2n letters which are either a's or 6's is greatest when the number of a's is equal to the number of Z>'s. If the n -f 1 numbers a, b, c, d, 33. be all different, and each of them a prime number, prove that the number of different factors of the expression ambcd is (m + 1) 2 W — 1.
all

CHAPTER

XIT.

Mathematical Induction.
important mathematical formula? are not easily demonstrated by a direct mode of proof; in such cases we frequently find it convenient to employ a method of proof known as mathematical induction, which we shall now illustrate.
158.

Many

Example
of the first

1.

Suppose

it

is

required to prove that the
is

n natural numbers

equal to

<

—^——

sum

of the cubes

.

'J-

We can easily see by trial that the statement is true in simple cases, such as when re=l, or 2, or 3 and from this we might be led to conjecture that the formula was true in all cases. Assume that it is true when n terms are taken that is, suppose
; ;

13

+ 2 3 + 33 +
term, that
is,

to

itteims=|
1)
3

H ( ;t+1
)j

3
.

Add
13

the («+

l) th

(n+
n

to each side
^
|

;

then

+ 23 + 33 +

to

n + 1 terms =j

^ +1
2

\+(n+iy

= {n + iy-('j+n + l\
(n+l) 8 (na +4n+4)
4
\

(n

+ l)(K + 2)
2

)\
'

-\

!

which is of the same form as the result we assumed to be true for n terms, n + 1 taking the place of n in other words, if the result is true when we take a certain number of terms, whatever that number may be, it is true when we increase that number by one; but we see that it is true when 3 terms are taken therefore it is true when 4 terms are taken it is therefore true when Thus the result is true universally. 5 terms are taken; and so on.
;
;

;

x

:

134
Example
x + a.
2.

HIGHER ALGEBRA.
To determine
the product of n binomial factors of the form

By

actual multiplication
(x

we have
;

+ a)
(x

(x

+ b)
(x

(x

+ c) = x 3 + (a + b + c) x 2 + (ab + bc + ca) x + abc
(x + d)

(x+a)

+ b)

+ c)

= x*+(a + b + c + d)x 3
+ (ab + ac+ ad + bc+ bd + cd) x~ + (abc + abd + acd + bed) x + abed.

In these results we observe that the following laws hold
1.

The number

of terms
left.

on the right

is

one more than the number of

binomial factors on the

2. The index of x in the first term is the same as the number of binomial factors and in each of the other terms the index is one less than that of the preceding term.
;

the coefficient of the second the coefficient of the third term is the sum of the letters a, b, c, ; term is the sum of the products of these letters taken two at a time; the coefficient of the fourth term is the sum of their products taken three at a time and so on the last term is the product of all the letters.
3.

The

coefficient of the first

term

is

unity

;

;

;

Assume that

these laws hold in the case of n - 1 factors that is, suppose ' ~ ~ (x + a) (x+b)... (x + h) = x71 1 +p 1 x n 2 +p.2 x n 3 +p.i x n i + ... +p> n
;

^

,

where

p 1 = a + b + c+
p.2

...h;
;

= ab + ac + + ah + bc + bd+ p 3 = abc + abd+
...
;

p n_x = abc...h.
Multiply both sides by another factor x + k
(x
;

thus

+ a)

(x + b)

...

(x + h) (x + k)
n (p 3 + pJc) x

= x n + (p + k)
x

x

n~l

+ (p.2 +p x k)
.

n~* +

~3

+... +l^ n- x k.

Now

^i

+ A;:=(a + & + c + ..+/*) + & = sum of all the n letters
1

a, b, c,...k;

p.2

+p k=p.2 + k (a + b + ... + h)

= sum
n
p.A

of the products taken two at a time of all the
...

letters a, b, c,

k;
.
.

+ ah + bc + .) +p.2 k =p 3 + k (ab + ac + = sum of the products taken three at a time of
. . .

all

the n letters
2? n _ 1 A*

a, b, c,

...

k;
b,
c,
...

= product

of all the n letters a,

k.

MATHEMATICAL INDUCTION.

135

If therefore the laws hold when ?t-l factors are multiplied together they hold in the case of n factors. But we havo seen that they hold in the case of 4 factors; therefore they hold for 5 factors; therefore also for 6 factors and so on thus they hold universally. Therefore
;
;

[x + a) (x + b) {x

+c)

...

(x

+ k) = x + ,V U_1 + Stfp-* + S.A x n ~* +
11

.

.

.

+ 8n

where

S^the sum <So = the sum

of all the n letters a,

b, c

...

Js;

of the products taken two at a time of these n letters.

S n =the product
159.

of all the n letters.

Theorems relating
Shew
u
.-c

to divisibility

may

often be esta-

blished by induction.

Example.
values of
n.

that

-l

is

divisible
l-1

by

x-1
; *

for all positive integral

By J
if

x
division

n

-l

x-1
1 is divisible
;

= xn ~

z'
l

-l

^

x-1
3

therefore x n
1

~l -

But x' fore x 4,

1 is

divisible by x -

1 is divisible by^r

-

also divisible by x - 1. therefore x - 1 is divisible by x - 1 ; there1 1, and so on ; hence the proposition is established.
1,

by x -

then x* - 1

is

Other examples of the same kind Theory of Numbers.
1G0.

will be

found in the chapter on the

the foregoing examples it will be seen that the only theorems to which induction can be applied are those which admit of successive cases corresponding to the order of the natural numbers 1, 2, 3, n.

From

EXAMPLES.
Prove by Induction
1.
:

XII.

1+3 + 5+
l2

+ (2n-l) = n 2

.

2.

+ 2 2 + 32 +

+ n 2 =i?i(n+l)(2tt+l). + 2» = 2(2'
l

3.

2

+ 22 + 23 +

-l).

4.

T~o + o~q + q-~T + 1.22.33.4

ton terms =
.r

-^ n+1

.

5.

Prove by Induction that

n

— yn

is

divisible

by

x+y when

n

is

even.

CHAPTER
Binomial Theorem.
161.
(x

XIII.

Positive Integral Index.

It

may be shewn by

actual multiplication that

+ a) (x + b) (x + c) {x + d) = x4 + (a + b + c + d) x3 + (ab + ac + ad + + (abc + abd + acd + bed) x + abed

bc

+ bd + cd) x*
(1).

We

may, however, write down this result by inspection ; for the complete product consists of the sum of a number "of partial products each of which is formed by multiplying together four If we letters, one being taken from each of the four factors. examine the way in which the various partial products are
formed, we see that
(1)

the term x

4

is

formed by taking the

letter

x out

of each

of the factors.
(2) the terms involving of any three factors, in
6, c,

out

letters a,
(3)

x3 are formed by taking the letter x every way possible, and one of the d out of the remaining factor.

the terms involving x 2 are formed by taking the letter x out of any two factors, in every way possible, and two of the letters a, b, c, d out of the remaining factors. the terms involving x are formed by taking the letter x out of any one factor, and three of the letters a, b, c, d out of the remaining factors.
(4)

(5)
«, b,
c,

the term independent of x
d.
.

is

the product of

all

the letters

Example

+ 3) (x - 5) (x + 9) = x 4 + (- 2 + 3 - 5 + 9) z 3 + (- 6 + 10 -18 -15 + 27 -45) a 2 + (30 - 54 + 90 - 135) x + 270 = x4 + 5a; 3 - 47.<c 2 - 69z + 270.
1.

(x

-

2) (x

BINOMIAL THEOREM.
Example
2.

POSITIVE INTEGRAL INDEX.

137

Find the coefficient of x* in the product (x - 3) (* + 5) [x - 1) (x + 2) (x - 8).

The terms involving x* are formed by multiplying together the x in any three of the factors, and two of the numerical quantities out of the two remaining factors
;

of the quantities

hence the coefficient is equal to the sum of the products - 3, 5, -1,2, - 8 taken two at a time.
coefficient

Thus the required

= -15 + 3- G + 2-1- 5 + 10-40- 2 + 8- 10 = -39.
1G2.
If in equation (1) of the preceding article

we suppose

b=c=d=a, we

obtain

(x

+

a)

4

= x4 + iax* + 6a

V + 4a

3

as

+ a4

.

The method here exemplified of deducing a particular case from a more general result is one of frequent occurrence in Mathematics for it often happens that it is more easy to prove
;

a general proposition than

it is

to prove a particular case of

it.

next article employ the same method to prove a formula known as the Binomial Theorem, by which any binomial of the form x + a can be raised to any assigned positive integral power.
shall in the

We

163.
integer.

To find

the

n expansion of (x + a) ivhen n

is

a positive

Consider the expression
(x

+ a)

(x

+ b)

(x

+

c)

(x

+

k),

the

number

of factors being n.

The expansion of this expression is the continued product of the n factors, x + a, x + b, x + c, x + k, and every term in the expansion is, of n dimensions, being a product formed by multiplying together n letters, one taken from each of these n factors. The highest power of x is x n and is formed by taking the letter x from each of the n factors. ~ The terms involving xn are formed by taking the letter x from any n—\ of the factors, and one of the letters a, b, c, ... k
, l

n in the from the remaining factor thus the coefficient of x final product *is the sum of the letters a, b, c, k; denote it
1

~

;

by^.
The terms involving x n 2 are formed by taking the letter x from any n — 2 of. the factors, and two of the letters a, b, c, ... k n from the two remaining factors thus the coefficient of x ~ in the final product is the sum of the products of the letters a, b, c, ... k taken two at a time; denote it by S2
;
.

~

\

1

;

138

HIGHER ALGEBRA.
n~r

are formed by taking generally, the terms involving x — r of the factors, and r of the letters the letter x from any n a, b, c, ... k from the r remaining factors ; thus the coefficient of

And,

r x"~ in the final product is the sum of the products of the letters a, b, c, ...k taken r at a time; denote it by Sr
.

The

last

term in the product

is

abc

...

k; denote it

by

Sn
.

.

Hence

= x n + Sx

12
(x
n~l

(x + k) + a)(x + b)(x + c) + SjxT* + ••• + £rx"~ + ...+Sn— ,x + Sn
r

In $j the number of terms

is

n

;

in

S2

the

number of terms

is

the same as the number of combinations of n things 2 at a time n n that is, C2 ; in S3 the number of terms is C3 ; and so on.

Now
"Ca:
(x

suppose
a)
n

b,

c,
2
:

...

k,

S, becomes "C\a

S
1

each equal to a; then S becomes becomes "Cjf: and so on: thus
l

+

= xn + n C ax n
l

+

n

C a xn
2
2

~2

+ "C^aV" 3 +

.

.

.

+ "Ca"

;

substituting for *Clt
n n (x+a) = x"+nax

n

C

2,

...

we

obtain
„_„ J

»

«-il

+

n(n—l)

- —r-^oV
n+
1

+

n(n — \)(n—2) /v v l
1

a 3x n

n

3

+...

+ an

,

the series containing

terms.

is

the Binomial Theorem, and the expression on the right said to be the expansion of (x + a)*.

This

is

164.

The Binomial Theorem may
induction

also be proved as follows

:

the product of the n factors x + a, x + b, x + c, ...x + k as explained in Art. 158, Ex. 2; we n can then deduce the expansion of (x + a) as in Art. 163.
find

By

we can

165.

The

coefficients

in the expansion of (x

+
,

a)" are
n
,

very
n
.

We

C3 ... C n conveniently expressed by the symbols "C,, "C 2 shall, however, sometimes further abbreviate them by omitting With this notation we have n, and writing (7,, C 2 C 3 ... C n
,
, .

(x
If

+ a) n =
write

x"

+ C ax
x

n~

l

+ C 2 a2 x n

~2

+ C3 a 3x n

~3

+

...

+ Ca\

we

—a

in the place of a,
-l

we
n-2

obtain
3
3 (-a) 3 xn - +... + Cn (-a) n

(x

n -a) n = x" + C\(- a) x

+

C

2

(-a) 2x

+C

= xn - C,axn
1

~

l

+ C„a 2xn 2

-2

- C,a3xn ~ 3 + 3
1

...

+ (- IYG na\
\ /

n Thus the terms in the expansion of (x + a) and (x — a) are numerically the same, but in (x - a)' they are alternately positive and negative, and the last term is positive or negative according as n is even or odd.
n

BINOMIAL THEOREM.
Example
1.

POSITIVE INTEGRAL INDEX.
{x

1

39

Find the expansion of

+ y) 6

.

By

the formula,
{x

+ yf = x« + 8tfi xhj + 6 C^xY + C3 afy8 + B C 4 .r 2 4 + 8 <7e x,f + »Ce ,/' = z? + 6.1-V + loo; 4 2 + 20a; 3 3 + loxhf + Gxi/> + if',
fi

*/

?/

?/

on calculating the values of 6 C1

G
,

C2

6
,

C3

,

Example
[a

2.

Find the expansion of
X

(a

-

2.r) 7 .

-

2x) 7

= a7 - 7C

a c (2x)
'

+ 7 C2 a 5 (2a;) 2 - 7 C3 a 4 (2a;) 3 +
,

to 8 terms.

7

Now remembering that n Cr = n Cn _ r after calculating C 3 the rest may be written down at once; for 7 GX = 7 C^
,

7

the coefficients up to Cr — 7 C< \ and so on. x

Hence
(a

- 2x) 7 = a7 - 7a B

{2x)

+ jp| a 5 (2xf -

\^-\ «4 (2a;)
- 21a 2
(2a;)

3

+

= a7 - la 6 (2x) + 21a 5 (2a;) 2 - 35a 4 {2xf + 35a 3 (2a-) 4
5

+ la

(2s) 6 - (2.r)~

= a 7 - Ua 6x + 84a5
Example
3.

a;

2

- 280a 4 .r 3 + 560a 3a; 4 - 672aV + USaafi -

128a; 7 .

Find the value of
(a

+ Jtf^ly + (a- Ja- -

1)".

We have here the sum of two expansions whose terms are numerically the same ; but in the second expansion the second, fourth, sixth, and eighth terms are negative, and therefore destroy the corresponding terms of the first expansion. Hence the value
= 2 {a7 + 21a 5 (a2 - 1) + 35a 3 (a 2 - l) 2 + la (a 2 = 2a (64a 6 - 112a 4 + 56a 2 - 7
).

l) 3 }

the coefficient of the second term is C of the third term is G 2 ; of the fourth term is "C3 ; and so on ; the suffix in each term being one less than the number of the term to which it applies hence "C r is the coth This is called the general term, terin. efficient of the (r + l) because by giving to r different numerical values any of the n coefficients may be found from C r ; and by giving to x and a their appropriate indices any assigned term may be obtained. Thus the (r + l) th term may be written
166.
of (x
a)
n
,

In the expansion
;
l

+

n

n

;

Cjrw,

or

»(»-l)("-2)-(»
t

+»,,-,„,.

In applying this formula to any particular case, it should 1><> observed that the index of a is the same as the svffix of C, and that the sum of the indices ofx and a is n.


140
Example
1.

;

.

.

HIGHER ALGEBRA.
Find the
fifth

term of

(a

+ 2a; 3 ) 17
4
)

.

The required term

= 17 C4 a13 (2a; 3

17.16.15.14

1.2.3.4

xl6ft13 .T 12

= 38080a 13 x 12
Example
2.

.

Find the fourteenth term of
(

(3

-

a) 15 .

The required term

= 15 C13 (3) 2 - a) 13 = 15 C2 x(-9a13 = - 945a 13
)
.
.

[Art. 145.]

The simplest form of the binomial theorem is the expansion of (l+x) n This is obtained from the general formula of Art. 163, by writing 1 in the place of x, and x in the place of a. Thus
167.
(1

+

x)

n

= l+ H C x + "C2 x 2 + n(n-\) 1 ^
i

.

.

.

+ "C x + ..+ "Cx n
r

r

1

+ nx +

— 1.2

zr—i ar

2

~

+

4- r" +

'

}

the general term being

n(n—l)(n—2)

(n-r+

1)
tb

,

.

The expansion of a binomial may always be made upon the case in which the first term is unity thus
;

to

depend

{x

+y

yJ(X (i + l)J
= xn (l + z) n where
,

z

=V
x
(as

Example

1.

Find the

coefficient of
2

a;

16

in the expansion of
20

2

-

2a;)

10
.

We have
and, since
a;

(a;

-

2a;)

10

=

a;

(1 - -

V
/
(

20

multiplies every term in the expansion of

1 -

2\ 10 we have in
,

J

this expansion to seek the coefficient of the

term which contains
4

Hence the required

coefficient

= 10 C4
10
.

(

.

2)

9

8

.

7

1.2.3.4

xl6

= 3360.
In some cases the following method
is

simpler.

BINOMIAL THEOEEM.
Example
2.

POSITIVE INTEGRAL INDEX.
x r in the expansion of
term.
./•-

141
-

Find the

coefficient of
tlio

(

i

j

.

Suppose that x r occurs in

(p

+ l) ,h

The

(p

+

1)°'

term

= *CP (x-) n -i> (iY
= n Cp x"' -*".
1

But

this

term contains x r and therefore
,

2n-5p = r,
5

or

p=-

2*1

r

5

Thus

the required eoellicient = n CiJ = n Co, l _,.
n

g(2n-r)
2n —— —
i'

= (3n

+

?•)

Unless

is

a positive integer there will be no term containing x r in

the expansion.

In Art. 163 we deduced the expansion of (x + «)" from the product of n factors (x + a) (x + b) ... (x + k), and the method of proof there given is valuable in consequence of the wide gene1G8.
rality of the results obtained.

But the following shorter proof
method

of

the Binomial Theorem should be noticed.
It will be seen in Chap. xv. that a similar
to obtain the general
is

used

term
(a

of the expansion of

+

b

+ c+

)".

161).

To prove

the

Binomial Theorem.
1

The expansion of (x + a)' is the product of n factors, each equal to x + a, and every term in the expansion is of n dimensions, being a product formed by multiplying together n letters, Thus each term involving one taken from each of the n factors. r x"~ a is obtained by taking a out of any r of the factors, and x Therefore the number of out of the remaining n — r factors. terms which involve x"~ a must be equal to the number of ways that is, the coellicient in which r things can be selected out of n n~ "6' and by giving to r the values 0, 1, 2, 3, ... n in cC is of x Hence succession we obtain the coefficients of all the terms.
r r r
;

r

r

,

(x

+

a)

n

=x

l

4-

m

C X*- a +
1 J

n

C,,x -°-a

n

2

+

.

.

.

+

n

C x"-a +
r r

...+

a",

since

*C and "C n are each equal

to unity.

:

142

HIGHER ALGEBRA.

EXAMPLES.
Expand the
1.

XIII.

a.

following binomials
.

(#-3) 5

2. 5.

(3^ + 2y) 4

.

3.

{Zx-yf.
(1-^j/) 7
.

4.

(l-3a 2 ) 6

.

{a?+x)\

6.

">•

g*-js)'The 4th term

n
of
(a?

-

(H':

12
-

NT
of (1 - 2x) 12
.

Write clown and simplify
13.

-

5)

13
.

14.
.

The 10 th term

15.

The 12 th term

of (2#

- 1) 13

16.

The 28 th term

of (5x + 8y) 30

.

17.

The
The The

4 th term of

(a U + 96\ J /

10
.

18.

5

th

term of (2a -

b\ 8
.

-J

19.

V th

term of (^' -

^-Y
5.

.

8

20.

The

5 th term of

(

— -V x

%

Find the value of
21.

(x + s/2y +
(

(x-j2)\
.

22. 24.

(V^^+^-CV^3^-^)
(2-Vr^) 6 + (2 + v I^^) 6
/
.

5
-

23.

v/2

+ l) 6 -( N/2-l) 6

25.

Find the middle term of
Find the middle term of Find the
Find the

f

-+\x a
1

a

cV\ 10

26.

(

- "—


j
.

27.

coefficient of
coefficient of

a.

18 in L'V2

+—
-

]

.

28.

x 18

in (axA

bx) 9
in

.

29.

Find the

coefficients of

x 32 and #~ 17
/

(

1\ 15 ( A x - -g
J

30.

Find the two middle terms of 3a (

a3\ 9

)

.

BINOMIAL THEOREM.
31.

POSITIVE INTEGRAL INDEX.

1

b3

Find the term independent of x in

(

-x2 —

)

.

32.

Find the 13th term of

Ux -

\-\

.

33.

If x* occurs in the expansion of

lx+-\
/

,

find its coefficient.
3»i
.

34.

Find the term independent uf

;/;

in

f

x— -.,
I

1 \
j

35.
....
,

If
.

1\'-' / xp occurs in the expansion of ( xr+-

1

,

prove that

its co-

eihcient

is
1

\2n .

(4
j3

"-^ \@n+p)

170. In the expansion of (1 4- x) u the coefficients of terms equidistant from the beginning and end are equal.

The
"C..

coefficient of the

(r

+

l)

th

term from the beginning

is

Tlie

(r+l) th term from the end has n + 1— (r+1), or
it;
th
,

n-r

therefore counting from the beginning it is the (n — r + l) term, and its coefficient is "Cn _ r which has been Hence the proposition shewn to be equal to "Cr [Art. 145.]

terms before

.

follows.

171. (l + x)»

To find

the

greatest

coefficient

in

the

expansion of
m

The

coefficient of the general

term of

(1

+x)"

is

C

r

j

and we
is

have only to find for what value of r

this is greatest.

By

Art. 154,
is

when n
it is

is

even, the greatest coefficient
,,

"C n
i

;

and when n
being equal.
172.

odd,

"C
2

or "C

,

,

;

these

two

coefficients

2

To find
have

the greatest

term in the expansion of (x +

a)".

We

(x

+

a)"

= x" (l +

-Y

;

therefore, since

n x multiplies every term

in

( 1

+ -j

,

it

will

be

sufficient to find the greatest

term in this latter expansion.

;

144
Let
the r th

HIGHER ALGEBRA.
and (r+l) th be any two consecutive terms. term is obtained by multiplying the r th term by
is,

The (r+l) th
.


x

:

that

by

(

1

)

-

.

r

\
Vh

r

J

x
as

[Art. 166.1 L J
_

The

+
r

]

factor
is
1

1

decreases

r

increases

;

hence the

(r+l) th term
until
(

not always greater than the rth term, but only
)

\

r

J

- becomes equal x
/n +
\

to

1,

or less than

1.

Now
,

—r

1
1 1

\

-

a

J x
1
.,

>

1•j

so long as

n+

1

>-

x
a
;

that

..

n+
is,

1

>-+

r

x a

1,

or


a
If

>r

( 1 ).

——
-+ a
1

be an integer, denote

it

by

j>

j

then

if

r

])

the

multiplying factor becomes 1, and the (p + l) th term th and these are greater than any other term. /> ;

is

equal to the

71+1
If

——
1

be not an integer, denote

its

integral part

by q

- +
a

then the greatest value of r consistent with (1) th term is the greatest. (q + 1 )

is

q\ hence the

Since we are only concerned with the numerically greatest term, the investigation will be the same for (x-a)"; therefore in any numerical example it is unnecessary to consider the sign Also it will be found best of the second term of the binomial. to work each example independently of the general formula.

-

;

BINOMIAL THEOREM.
Example
1.

POSITIVE INTEGRAL INDEX.
of

1

to

If

x = -, find the greatest term in the expansion

(1+

I

Denote the

?•"'

and

(/-

+ l) tu terms by Tr and Tr± 1

respectively; then

9-r
hence
.

4

T7^. > Tr
l

,

so long as °

9-r
r

4

x

->1;
3

that
or

is

36 - 4r > 3r,

3G>7r.

The
is

greatest value of r consistent with this
its

is

5

;

hence the greatest term

the sixth, and

value

3i4 243~

Example

2.

Find the greatest term in the expansion of (3- 2a:)9 when

(3-

2^ = 3^1

2

|J;

(2rV -—
1
J

.

„ Here


-*r+i

= 9-r+l
10 - r
X

2*
-~o~

x

Tr
;

,

numerically,

...


r

2 3

^r,.
,

iience

Tr+1 > Tr
10 ~ r
2
6

long as so i

x ->

i

1

that

is,

20>5r.
for all values of r

Hence

up to

3,

these are the greatest terms. Thus the 4"' and 5 th terms are numerically equal and greater than any other term, and their value

T r+x = T r> and

we have Tr+l >Tr

;

but

if

r=4, then

=3"x»C,x f|
H. H. A.

J

=36 x 84x8 =489888.
10

146
173. To find of (I +x)".
the

HIGHER ALGEBRA.
sum of
+ a?)n =
1

the coefficients

in the

expansion

In the identity
put x =
1
;

(1

+ C x + G 2 x 2 + C3 x 3 +
x

. . .

+C

af,
t

thus

2*=l + Cx + C2 + C3 +... + Cn = sum of the coefficients.
Cor.
that is "the total [Art. 153.]
C\

+C 12 + C +
g q 3

...

+

Cn =T-l;
'

number

of combinations of

n things"

is 2"

1.

174.

To prove

that in the expansion of (1
is

+

x)

n
,

the

sum of

the coefficients

of the odd terms

equal

to the

sum of the
...

coefficients

of the even terms.

In the identity put x = - 1 thus
;

(

1

+

x)

n

=

1

+ C x + C2x2 + C 3x3 +
x

+ C x\

= l-C + a8 -(7a + (74 -C6 +
1

;

...

i

+ 1+ c4 +

;.....

-0
=
1

1

+'Ca + C.+
all

(sum of
n-l

the coefficients)

=

2

175. The Binomial Theorem may also be applied to expressions which contain more than two terms.
:

expand

Example.

Find the expansion of
3 2 2
)

(x z

+ 2x- l) 3

.

Regarding 2x -

1 as a single term, the
2

= (x-) + 3 (a (2x - 1) + 3a; = x 6 + 6a; 3 + 9a; 4 - 4c 3 - 9a; 2 + 6x — l,
176.

expansion (2x - l) 2 + (2x -

3

l)

on reduction.

The following example
If
(1

is

instructive.
,

Example.

find the value of

and

+ x) n = c + c x x + c#? + +c n xn c + 2c 2 + 3c 2 + 4c 3 + + n +l)cn 2 2 c 1 + 2c 2 + 3c 32 + +nc n2
(

(1),
(?).

The

series (l)

= {c + c + c 2 +
1

+ c n + (c x + 2c 2 + 3c3 +
)

+nc n

)

=2 w +

?i

Jl

+ (/t-l) +
1

v

J_L

/

+

+ il

= 2 n + n(l + l) n ~ -2n +w.2»-1
.


BINOMIAL THEOREM.
To
find the value of the series c x x + 2c2 x 2

4

;

:

POSITIVE INTEGRAL INDEX.
(2),

1

IT

we proceed thus

+ 3c 3 x* +
8>

+ ncn x n
*» +

=»«{l-Hn-l),+ <"-^= nx (1 + x)"1
;

+ *»-}

hence, hv chauging x into -

x

,

we have

&+£+!*+ X X' X
s

+ ^=!(i + i)*n
X

1

X \

xj

(s) .

W
;

Also
If

cQ

+ c 1 x + c.2 x 2 +

+ en z% =(l+z) n

(4).

we

series on the left-hand sides of (3) and (4), see that in the product the term independent of x is the series (2) hence n l\ u_1 f the series (2) = term independent of x in - (1 + x) n ( 1 + • I

we multiply together the two

term independent of x in


)l

(l

+ x)-'

1

'1

= coefficient

of x n in

n

(1

+ .r) 2>l-l

= ?ix 2n -

1
<

i2n-l

n-1

In-

1

EXAMPLES.

XIII.

b.

In the following expansions find which
1.

is

the greatest term

(x

2. 3.

when #=11, y = - 3y) 28 when x = 9, y = 4. ( 2x (2a + b) u when a =4, 6 = 5.
y)
30

4.

(3

+ 2x) lb when x — ss

5

.

In the following expansions find the value of the greatest term
5.

:

(1

+ x) n when x = -

2

o

,

n = 6.

6.

+ #)* when a= s

,

.r

= -,

??=9.

10—2

148
7.

HIGHER ALGEBRA.
Shew that the

equal to the "- 1 (1+tf) 2
.

sum

middle term of (l + x) 2n is of the coefficients of the two middle terms of
coefficient of the

be the sum of the odd terms and n even terms in the expansion of (x + a) prove that
8.

If

A

B

the

sum

of the
n
)
.

,

A 2 -B2 = (x2 -a2
(x+y) n are

th 2 nd, 3 rd , 4 terms in the expansion of 1080 respectively ; find x, y, n.

9.

The

240, 720,

10.
11. 12.

Find the expansion of Find the expansion of
Find the
r th

+ 2x - x )\ (Zx 2 -2ax + 3a
(1
2

2 3
)
.

term from the end in (x + a) n

.

(]\ x--j
14.

2n +

l

In the expansion of (1 + #) 43 the coefficients of the (2r + th the (r + 2) terms are equal; find r.

l)

th

and

Find the relation between r and n in order that the 2n may th th be equal. of the 3r and (r + 2) terms of (l+x)
15.
16.

coefficients

Shew that the middle term
1

in the expansion of (1

+x) 2n

is

.3.5...(2n-l)
hi

sn

^
(1

If c

,

Cj, c 2 , ...

<?

n

denote the coefficients in the expansion of

+x) n

,

prove that
17.

^ + 2^ + 303 +
c

+ncn =n.2n - 1
.

.

18.

- + +i 42
i

c,

c, 2 - 4-

3
3c8
1

o

.

H +

-l — = 2 n+\ n+l
cn
1

n+

.

io iy.

c.2c


c

2

r

+~
,

ncn
x

cx
v
,

C2
Cl + c 2
N )

cn _

n{n+\
O A
.

)

20.

(co+ej
a 2c

(

(c n

. 1 + cn )

= -2n +

c,c, 1 2

cn

^

(n+l) H '1

M 21.
22.

+ —i+ 2
2 2c,
1

2 3c2
3

+

—+
2 4c,
?

+
|2w
i

4

—+-p^ n
1

cn l

3' 1 +

-1

n+l

-=-..

c

(

f+c +c2 +
j

+ c« =

7i

]^

23.

c cr

+ CjC r + + c 2cr + 2 +

+cn _ r cn = --

\2n =~~

.

CHAPTER
Binomial Theorem.
177.

XIV.

Any

Index.

chapter we investigated the Binomial Theorem when the index was any positive integer; we shall now consider whether the formula? there obtained hold in the case of negative and fractional values of the index.

In the

last

Since,

by Art. 167, every binomial may be reduced to one
type,
it

common

will be sufficient to confine our
(1

attention

to

binomials of the form

+x)

n
.

By

actual evolution,

we have
1

(1

+ xf = V 1 + X =
division,
-

+ ^ X - - X2 +

yr.

x3 -

;

and by actual
(1

- x)~ 2 =

7^

-

xa

=

1

+ 2x + 3x* + ix 3 +
[Compare Ex.

:

1,

Art. CO.]

and

in each of these series the

number

of terms

is

unlimited.

In these cases we have by independent processes obtained an
i

expansion for each of the expressions (1 + x) and (1 + x)~~. We shall presently prove that they are only particular cases of the n general formula for the expansion of (1 + x) where it is any
2
,

rational quantity.

This formula was discovered by Newton.

Suppose we have two expressions arranged powers of x, such as
178.
,

in ascending

I

+

mx+

m (m v
-

1 )

'x-+

(m mx -

1 )

(m /x

-

2)

, 'a?+

(I

).

and

l+n.v +

g
l

>x-+-±
]

K J -

a?

+

(2).


150

xx

;

HIGHER ALGEBRA.
of these
will be a series in as-

two expressions cending powers of x\ denote it by 1+ Ax +Bx2 + Cx3 + Dx4 +
then
it
is

The product

;

are functions of and n, in any particular and therefore the actual values of A, B, C, and n in that case. But case will depend upon the values of the way in which the coefficients of the powers of a; in (1) and (2) combine to give A, B, C, is quite independent of and n ; in other words, whatever values in and n may have, A, B, C, If therefore we can determine preserve the same invariable form. the form of A, B, C, for any value of and n, we conclude that A, B, C, will have the same form for all values of
clear that A, B, C,

m

m

m

m

m

and

n.

often referred to as an example ; of "the permanence of equivalent forms " in the present case we have only to recognise the fact that in any algebraical product the form of the result will be the same whether the quantities involved are whole numbers, or fractions ; positive, or negative.
principle here explained
is

The

use of this principle in the general proof of The proof which Ave the Binomial Theorem for any index. give is due to Euler.
shall

We

make

179.

To prove

the

Binomial Theorem ivhen

the

index

is

a

positive fraction.

Wliatever be the value of m, positive or negative, integral or fractional, let the symbol f(m) stand for the series
,

1

+ mx +

m (m ^ 1) m —(m-Y) (m — 2) -x-

—y—
\

+

v

' v

s 3

'

+

...

then.y(n) will stand for the series
-

1

+ nx +

n(n — l)„ 2
'

x +

n(n — l)(n — ' v
v

2)
'-

„ 3

+

....

multiply these two series together the product will be another series in ascending powers of x, whose coefficients loill be and n may be. unaltered inform whatever
If

we

m

To determine this invariable form of the product we may give to m and n any values that are most convenient for this purpose suppose that m and n are positive integers. In this casey(m) m andy*(?i) is the expanded form of is the expanded form of (1 + x) n and therefore (1 +x)
;
,

;

BINOMIAL THEOREM.
f(m) xf(n) but when
(1

ANY INDEX.
+ a?)" =
(1

1.51

+

x)

m

x (1

+ x) m+ \
+ x)"
,

m

and n are
-,

positive integers the expansion of (1
v

+

"

/

(m +

n)

(m + n I
.

1)

.

the form of the product of f(m) x/(><) in o# cases, whatever tlie values of and n may be; and in agreement with our previous notation it may be denoted hyf(m + n) ; therefore for all values ofm and n
is

This then

m

/(m) xf(n)=f(m + n).
Also

/(w) x/(n) x/(^) =/(w +

»)

x/( p)
similarly.

=f(m + n +p),
Proceeding in
tliis

way we may shew

that
+...to k terms).

f(m) xf(n)

x/(j;)...to k factors

=/(»» + n +p
?i,

Let each of these quantities m,

j),

be equal to =
rC

,

where h and k are positive integers

;

but since h

is

a positive integer,

f (h) = (1 + x)

h
;

but

y*
(

y

)

stands for the series

,

h

k\k

J

2

,,
.*.

(

1

+

vi
a;)

=

,

1

+

h
T

x+

k \k x

J /

,

x-

+

,

«

1.2
for

which proves the Binomial Theorem
index.

any positive fractional

;

152
180.

HIGHER ALGEBRA.
To prove
the

Binomial Theorem when

the

index

is

any

negative quantity.
It has

been proved that

f(m) x/(w) =/(w* + n)
for all values of
positive),

m

and

n.

Replacing

in

by — n

(wliere

n

is

we have
f(-n) •xf(n)=f(-n + 7i)
=/(0)

=%
since all terms of the series except the first vanish

•'•

/hr

/( - n)

'

but/(w) = (l +

x)'\ for

any positive value

of n;

or

(1

+ *)"" =/(-*)•

But f(—n) stands
1

for the series

+ (- n) x +

^
1
.

'-f—,

= L

ar

+
*>

;

...

(1

+

«.)-

=

1

+ (_ W )

a.

<" " " + (rg )

+

;

which proves the Binomial Theorem for any negative index. Hence the theorem is completely established.
181.

The proof contained

in the

two preceding

articles

may
dif-

not appear wholly satisfactory, and will probably present some ficulties to the student. There is only one point to which
shall

we

now

refer.
is finite

In the expression iov f(in) the number of terms
vi is

when

a positive integer, and unlimited in all other cases. See Art. 182. It is therefore necessary to enquire in what sense we

BINOMIAL T11EOUEM.

ANY INDEX.

153

are to regard the statement thaty(m) x/(n) =f(m + n). It a\ ill be seen in Chapter xxi., that when x< 1, each of the series/^/), /( n )i/( + n ) * s convergent, and/(m + «) is the true arithmetical equivalent of f(m) *f(n). But when sol, all these series are divergent, and we can only assert that if we multiply the series denoted by/(m) by the series denoted by f(u), the first r terms of the product will agree with the first r terms of f(m + n), whatever finite value r may have. [8ee Art. 308.]

m

3

Example

1.

Expand

(1

2 - xf

to four terms.

3

Id- 1 ),

,.J(H(S-)

Example

2.

Expand

(2

+ 3a;) -4

to four terms.
4

(2

+ 3z)- 4 = 2-<(l + ^)~

182.

In finding the general term we must now. use the

formula

m(w-1)(w-2)

(n-r +

l)

xr

written in full

;

for the

symbol "C can no longer be employed
r

when n

is

fractional or negative.
coefficient of the general
its

Also the

term can never vanish unless
is

zero; the series will thereterm, when n — r + 1 is zero ; that is, when fore stop at the r r=oi+ l but since r is a positive integer this equality can never Thus the hold except when the index n is positive and integral. expansion by the Binomial Theorem extends to w+1 terms when n is a positive integer, and to an infinite number of terms in all other cases.
th
;

one of the factors of

numerator

.

154
Example
1.

HIGHER ALGEBRA.
1

Find the general term in the expansion

of (1 +x)'\

The (r+l) th term-

—-


r

-5)
2 r lr

(-2r + 3)
af.

The number
tive
;

of factors in the numerator is r, and r - 1 of these are negatherefore, by taking -- 1 out of each of these negative factors, we may

write the above expression

(-i)~
Example

1

-

8

-

6

-<»-V
i

2.

Find the general term in the expansion of

(l-nx) n .

The

(r

+ 1)'- term = »

V"

A

"

/

E
w
I

M
(1-F^Un)

£

(

- «»)r

/
_

= !(!-«) (l-ar.)r

^
^

r

l(l-n)(l-2n)

(1-r-l.n)

=

_
(

i)r
(

_

i)r-i

(n-l)(2n-l)

(r-l.n-1)

^

(n

-

1) (2«

-

1)

.

....(^l.n-l)

since

(_1)»-

(_

l)r-i

= (_ i)2r-i = _ 1#
- x)~ 3

Example

3.

Find the general term in the expansion of
3'(- 4 >'-_5

(1

.

The(r + irterm=<-

)^-(- 3 -'-+ 1
r
(r

)

-,)r
(

= (1)r 3.4.5
~
1.2.3

+ 2)

(1)ffa,

[

}

X
r

~
by removing

_

(r+l)(r+2) . *» 1.2
from the numerator and denominator.

like factors

BINOMIAL THEOREM.

ANY INDEX.
a.

155

EXAMPLES. XIV.
Expand
1.
1

to 4 terms the following expressions:

(l+xf.

s

156
and,
if

HIGHER ALGEBRA.
not,

under what conditions the expansion of
true equivalent.

(1

+

x)

n

may

be used as

its

Suppose, for instance, that
(1

n——

l; then

we have
x*

-x)~ =
r

1

+ x + x2 + x 3 +
obtain
2
3
2 4

+

(1);

in this equation put x

=
1

2

;

we then
+
2

(-l)~

=l+2

+

+

2

+

This contradictory result take
,

is sufficient

to

shew that we cannot

n(n-l) l+nx+— — ~—
\
'

x2 +

as the true arithmetical equivalent of (1

+ x) n
of

in all cases.

Now
gression,
series
(1) v '

from the formula for the sum

we know

that the

sum
- xr

a geometrical proof the first r terms of the

=

1
-z

I

—x
1

1

-X

x 1 - x
1,

r

y

and,
large

when x

is

numerically less than

by taking r
;

sufficiently

we can make
number

x
^

r

as small as

we

please

that

is,

by taking

a

sufficient

of terms the
^
r
.

sum can be made
But when x
is

to differ as

little

as

we

please from

numerically

greater than

1,

the value of

x ^ r increases with 1 - x
of
JL

r.

and therefore

no such approximation to the value

is
vC

obtained by taking

any number

of terms of the series
1

+ X + Xs + X3

4-

It will be seen in the chapter on Convergency and Divergency of Series that the expansion by the Binomial Theorem of

(1+x)" in ascending powers of telligible when x is less than 1.

a?

is

always arithmetically

in-

But

if

x

is

greater than

1,

then since the general term of

the series
,

1

+ nx

n(n-\)
H
.j

x"
.

+

I

-

BINOMIAL THEOREM.
r
,

ANY INDEX.

157

contains x it can be made greater than any Unite quantity by taking r sufficiently large in which case there is no limit to the value of the above series; and therefore the expansion of (1 + x) n as an infinite series in ascending powers of x has no meaning arithmetically intelligible when x is greater than 1.
;

184.

We

may remark
;

that
for

by the Binomial Theorem
either of the

we can always expand (x + y)" we may write the expression in
:

two following forms
x"

('*!)'

'(•ff.
first

and we obtain the expansion from the according as x is greater or less than y. To find in its expansion of (1 — x) -u
185.
.

or second of these

simplest

form

the

general term in the

The

(r

+

l)

th

term

(-

n)(-n-

I)

(-71-

2)...

(-n-r+1)
(-*y
1)

+1 Hw + 2) -(** + »•= (- iy »(*
= (_ I)*
ttv*+l)(tt+2)...ytt +

r-l )

xr

n

(n

+

1)

(n + 2)

...

(n + r

-

1)

r

From
(1

this it appears
is positive.

that every term in the expansion of

-x)~*

Although the general term in the expansion of any binomial may always be found as explained in Art. 182, it will be found more expeditious in practice to use the above form of the general term in all cases where the index is negative, retaining the form n(n- l)(n-2) ... ( n - r + 1 ) , x
i

t
only in the case of positive indices.


158
Example.

;

HIGHER ALGEBRA.
Find the general term in the expansion
of
-

_

.

.

—The
(r

1

= (l-3x)

-1 3
.

+ l) th term

1.4.7

(3r-2) 3rrr

1.4.7
^H
:

(Sr-2)
r

w

^

_i

If the given

formula for

we should have used the same the general term, replacing Sx by - 3x.
expression had been
(1
3

+ Sx)

186.
(1

The following expansions should be remembered
1

:

(1 (I

- x)' = 1 + x + x 2 + x 3 + - x)~ 2 = 1 + 2x + 3x2 + ±x 3 +
- x)~ 3 =1 + 3x + 6x* + 10x 3 +

+ xr +

+

(r

+
(

1)

xr +
l

+

r+

J%—K +
r

n when n is unrestricted in value, will be expansion of (1 + x) found in Art. 189 ; but the student will have no difficulty in applying to any numerical example the method explained in
,

Art. 172.
Example.
2 x=3
,

Find the greatest term

in the expansion of (l+a;)~ n

when

and n — 20.
fi j_
<t'

We have

^

^V+i—
-

,xxTr
?
-'r»

,

numerically,

19+r
'r+l >

r

•"•

so long as

2 (19 + r) £

—>

1

that

is,

38 >r.
for all values of r

Hence
,

up

to 37,

we have

these are the greatest terms. Thus the 38 th and 39 th terms are equal numerically and greater than any other term.

I^k = T,. and

jrr+1

>Tr

;

but

if

r=38, then

BINOMIAL THEOREM.
188.

ANY INDEX.

159

useful applications of the Binomial explained in the following examples.

Some
1.

Theorem are

Example

Find the

first

three terms in the expansion of
i

_i
3.

(l

+ 3*) -(l-2x)

r

Expanding the two binomials

as far as the term containing x'\

we have

,

/3

2\

/8

3

2

=1 +
If in this

1

13
-Q

X+

55 72

.

X "'

so that ar = -000004, we see that the third term is a decimal fraction beginning with 5 ciphers. If therefore we were required to find the numerical value of the given expression correct to 5 places

Example ^='002,

of decimals

it

would be

sufficient to substitute *002 for
.

x in

1

+

o

x, neglect-

ing the term involving x 2

When x is so small that its square find the value of be neglected,
Example
2.

and higher powers may

J(± + xJ*
Since x- and the higher powers may be neglected, it will be sufficient to retain the first two terms in the expansion of each binomial. Therefore
i

the expression

_tl±±±l b(i+|.)

-K-S-).
the term involving x- being neglected.

:

160
Example
3.

HIGHER ALGEBRA.
Find the value of -rj=
x/47
_i 1 --

to four places of decimals.

--

1 /

2

\

2

-^ = (47)
1/

*=(7*-2)*=-(l-n)
1

3

-7^ + 72 + 3 -74 + 2 -7G+--7 + 73 + 2
To obtain
*

^

5

:L_

75

+

7+ "" 2 *7
we proceed
as follows

the values of the several terms

1)1
7
7 7 7
)
)

-142857 -020408 -002915 -000416 •000059
5

=t,
'

!

)

= 7-3,

)

= ^;
.

and we can
5 ciphers.

see that the

term -

1 =

is

a decimal fraction beginning with

.-.

-i\/47

= -142857 + -002915 + -000088

= •14586,
and
this result is correct to at least four places of decimals.

Example

4.

Find the cube root
!

of 126 to 5 places of decimals.

(126)3

= (5 3 + l) a
1

5

~5
~

/

t

1

1
:J

V
1
3
'

3"5
1_

M
J.
'

9'5« + 81*5 9
_1

5

1

\ '")

1
9

1
•"

~
_

+

1^1^
3*"l0
-04
2

~ 5
2

55

+
5

81 *5 7
1

9'10 + 81
h

W_ "•• *10 7

= 5-1 ^

-00032
9
...

-0000128 —
81
...+...

3

=5-f -013333

- -000035

= 5 '01329,

to five places of decimals.

~

.

.

BINOMIAL THEOREM.

ANY INDEX.
b.

101

EXAMPLES. XIV.
Find the (r+1)" term
1

in each of the following expansions
(l-.t-)

:

!•

(l+#) (l+#)

-I 2
.

2.

-5
.

i

3.

(l+3.e)

:]
.

4.

J 3
.

3

5.

(l+.r2 )-3.

6

.

(i-2.v)~*.

7.

(a+fo?)" 1

.

8.

(2-.r)~ 2
*

.

9.

tt{rf-x*)\
*

10

-

7=A=.
</T+2*

11.

f/ 3/

N

(l-3.^

12.

V&Z^
,
:

Find the greatest term
13.
(

in

each of the following expansions
.

1

+ .v)

7

when

4 x=— lo
2

14.

(

1

+ a?) 2 when
— 74?)
11

a?= 5
1

4

15.

(1

wheu# = -.
o
a?

16.

(2a?

+ 5J/) 12 when
4.v)

=8
25

and y = 3.

17.

(5

-

~7

when v=t

18.

(3-r2

+ 4/) - n when x = 9, y = 2,

«=

1 5.

Find to
19.

five places of

decimals the value of
4/998.
1

v98.

20.

21.

\ 1003.
3

3/

4/

22.

\ 2400.

23.

^=.

1

24.

(1^)3.

25.

(630)

*.

26.

tfilla

If x be so small that its square find the value of
1

and higher powers may be neglected,

3

27.

(l-7tf) s (l

+ 2a?)"*.

162

HIGHER ALGEBRA.

31.

V^+C+jj
Prove that the
'

'

32

.

^T^-^1^
(1+5*)*+

(4+|Y
*

33.
is

coefficient of sf in the

expansion of (l-4r)

v2

31
35.

Prove that

(1

+*)*=2-

|l

-—
1

^

+

-^ (f^

)

f

Find the

first

three terms in the expansion of

'

(1

+ x) 2 Vl + 4x

36.

Find the

first

three terms in the expansion of
3

(! + #)* +

*Jl

+ bx
expansion of
(1

37.

Shew that

the
th
.

nth

coefficient in the

- x)~ n

is

double of the (n-l)
189.

of (1 + x)

n

,

To find the numerically greatest term in for any rational value of\\.

the expansion

Since we are only concerned with the numerical value of the greatest term, we shall consider x throughout as positive.

Case

I.

Let n be a positive integer.
is
is,

The (r+l) th term
by
.

obtained by multiplying the r th term

x

;

that

by

f

1
J

x

;

and therefore the

terms continue to increase so long as

Or
that
.,

1 'n+ 1

-1
,

)- 1

'

(n+ l)x
is.

— > 1 + x,
)

r

or

(n

*— + 1 1+02

x — >r.

BINOMIAL THEOREM.
(ll 4- 1

ANY INDEX.
it

103

If

^

X —

be an integer, denote

by p; then
is

if

r=p,

the

multiplying factor is 1, and the (;>+l) th term th and these are greater than any other term. ^>
,

equal to the

If

( 71 4,

1

)

X —

be not an integer, denote
is
7,

its

integral part by q

;

then the greatest value of r
greatest.

and the

(q

+

l) th

term

is

the

Case

II.

Let n be a positive

fraction.
is

As
rm
_.

before, the
,

(r+
t

th

term by
(1)

(n +
(

l)

term

obtained by multiplying the

1

\

I

)x.

x be greater than unity, by increasing r the above multiplier can be made as near as we please to - x so that after a certain term each term is nearly x times the preceding term numerically, and thus the terms increase continually, and there is no greatest term.
If
;

than unity we see that the multiplying factor continues positive, and decreases until r > n + 1 and from this point it becomes negative but always remains less than 1 numerically therefore there will be a greatest term.
(2)

If

x be

less

,

;

As

before,

the

multiplying factor

will

be greater than

1

so Ions: as
( Jl 4-

-^
1
\ X -— be an

(n + l)x
1

+x

—>

r.

If

^

integer, denote it

by p

;

then, as in Case

I.,

the (p + l) th term any other term.
If

is

equal to the

th

£>

,

and these are greater than

^p
4-

( 7t 4"

1

)

X

be not an integer,
is

let q

be

its integral

part; then

the (q

l)

th

term

the greatest.

Case

III.

Let n be negative.
in,

Let n - —

so

that

m

is

positive
nil _L

;

then the numerical

value of the multiplying factor

f

is

J

.

x

;

that

is

m-l
(

+

\
1
)

x.

r

J

11—2

164
(1)

HIGHER ALGEBRA.
If
II.,

x be greater than unity we may shew, as in Case that there is no greatest term.
(2)

If

x be
1,

less

than unity, the multiplying factor will be

greater than

so long as

(m that
is,

1

)

x —>

1

- x,

r
or
lyn.

(m—\)x — — > r. I -x
-.

If

^—

x

— be — x
1
)

CC

a positive integer, denote

it

by p

\

tlien the

th (p + l) term is equal to the any other term.
(fjr

p

th

term, and these are greater than

If

*

1

— be -x
)

1

£C

'

positive but not an integer, let q be its inte(q

gral part

;

then the
'-

+

l)

th

term

is

the greatest.

If

i

be negative,

tlien

m

is less

than unity

;

and by
x,
J

writing the multiplying factor in the form (1
:

we

hence each term is less than see that it is always less than 1 the preceding, and consequently the first term is the greatest.
190.
sions that

To find the number of homogeneous products of v dimenand their can be formed out of the n letters a, b, c,
by the Binomial Theorem, we have

powers.

By

division, or

= 1

— ax
1

=

1

2 2 3 3 + ax + a x + a x +

,

1

— bx
1

=

1

+ bx +

b

2

x2 +

b

3

x3 +

,

1

=
ex

1

+ ex +

c

2

x2 +

c

3

x3 +

,

;

BINOMIAL THEOREM.
Hence, by multiplication,
1
1

ANY INDEX.

KJ5

1
1

1

_ ax
...)

— bx
b*x*

1

ex

=

(1
1 1

+ ax + + x (a +
t

aV +
b

(1

+ bx +
2

+

...) (1
b'

+ ex +
2

c°x

2

+
.
.

...) ...
)
.

=

+

c

+

...)

+x

(a

2

+ ab + ac +
suppose

+ bc±c 2

4-

+

...

=

+ S x + Sjfx? + Saxa +

where
a,
b,
c,

Slt

>S'.,,

SaJ
three,

are the

duets of one, two,

homogeneous prodimensions that can be formed of

sums

of the

and their powers.
the

equal to 1 values of

each ; 1, and the l9 S2 S so obtained give the number of the Sl9 homogeneous products of one, two, three, dimensions.
b, c,

To obtain

number
:i

of these products, put a,

each term in
,
,

JS

S2 S
,

:i

,

now becomes

1

1

1 1

Also
1

— ax -

1

— bx
or (1


a;)

ex
".

becomes
(1


x) of

Hence

S = coefficient
r

r x in the expansion

of (1

— x)~

~

n(n+

l)(n + 2)
jr

(n+r-

1)

n + r—1
\r

\n—

1

To find the number of terms in the expansion of any multinomial when the index is a positive integer.
191.

In the expansion of
(at

+ aB + aB +

+a )",
r

every term is of n dimensions; therefore the number of terms is the same as the number of homogeneous products of n dimensions and their ... a that can be formed out of the r quantities a,, a powers ; and therefore by the preceding article is equal to
,
r
,

I?'

+ n—
r

1

n

—1

166
192.

HIGHER ALGEBRA.
From

the result of Art. 190 we relating to the number of combinations of

may deduce
n
things.

a theorem

then if we were to write Consider n letters a, b, c, d, ; down all the homogeneous products of r dimensions which can be formed of these letters and their powers, every such product would represent one of the combinations, r at a time, of the n letters, when any one of the letters might occur once, twice,
thrice,
...

up

to r times.

Therefore the number of combinations of n things r at a time when repetitions are allowed is equal to the number of homogeneous products of r dimensions which can be formed out of n
\n
letters,

+r—
,

1
-

and therefore equal to
\r

n—\

,

or

n+r

*CT

.

things r at a time when repetitions are allowed is equal to the number of combinations of n + r— 1 things r at a time when repetitions are excluded.
is,

That

the

number

of combinations of

it

193.

We

shall

conclude this chapter with a few miscel-

laneous examples.
(1

Example

1.

Find the

coefficient of

x r in the expansion of
...

- 2a;) 2

~

.

The expression = (1 - Ax + 4.x 2)

(1

+PyC +p^xr +

+p rxr + ...)
,

suppose.
1,

The

coefficient of

-4,4

respectively,

xr will be obtained by multiplying p r and adding the results hence
;

p r - x p r -» by
,

the required coefficient =p r - 4p r _ x

+ 4p r _ 2

.

But

p r =(- iy fe±afc±9
coefficient
1)r . 1

.

[Ex 3 Art. 182.]
. ,

Hence the required

=

.
(

1)

r

(r+lHr + 2) _ 4( _

rJ^ + 4( _ ira I ^r
(

= ^-[(r + l)(r + 2) + 4r(r + l)+4r(r-l)]
f-l) r

3

;

BINOMIAL TIIEOltEM.
Example
2.

ANY INDEX.

Hi?

Find the value
„ 42
,

of the scries

5
.

|_2.

3 5
.

T

5.7
|3.3
2

^

5.7.!)
1

-4-

4

.

3

:J

+

•••

mU The

.

expression v

— =2+3 —
.

[2

1 1 — + 3.5.7 — + 3.5.7.9 _ + —- 1 3.

:

.

1

3

;!

14

34

3

5 2
~
1

3

5
'

7
'

3

5
'

7
'

1)

-Q

2

Ll?
|2

2

2

2

'3?*"

~|3~
5

2 2J '3 :i+
3
5

2

2

'

2

24

]i~
7

*3 5+

••'

3 z 2
1

3 z 2

*3

+

2' 2 2
"~J2~

/2\/2\

a

2*2*2 /2V 2 2 2

1

o-.r-ffl"
= 35 =V3
Example 3. If ?t is any n (3 + Jl) is an odd number.
Suppose I
-

positive integer,

shew that the integral part
(3

of

to denote the integral

and/
.

the fractional part of

Then

I+f=3 +C
n
is

1

S"- s/7

i

+

a2 S
less

n~2

+ a/7)'
(1).

1
.

7+(78
1,

3*-«^7) 8

+
1

Now 3- N/7
.•./' = 3 n

positive

and

than

therefore

(S-^)'

is

a proper

fraction; denote it by/';

-C
(1)

,

3' l

1

-V7 + C'
(2)
;

3' l
2

-2

.7+C3 3' -3( v/7) 3 +
l
.
.

(2).

Add

together

and

the irrational terms disappear, " I+f+f = 2 (3» + C2 3' 1 2 7 + )
.
.

and we have

= an
But since/ and/' are proper
:\

even integer.

fractions their

sum must

be

1

I=an odd

integer.

EXAMPLES. XIV.
Find the
1.

c.

coeflicient of

xm
an

in the expansion of

2.

in the expansion of

- xf 4 + 2a - a 2
(1

'

3.

«* in the expansion of

X "T" X

168
4.

HIGHER ALGEBRA.
Find the
coefficient of

xn

in the expansion of

2 4- x + X2
(

1

+ ^J

5.

Prove that
1

1
*

2
6.

2

1.3 + 271

1
'

1.3.5

2_

2*

2.4.6'2 3

+

1-3.5.7 1^ 2.4.6.8'2*

V 3'

Prove that
N/8

3.5.7 - + t^~; + = 1 + 3 H 3^5 H \ \ n + 4 4.8.12 4. 8
'
~ ,

7.

Prove that

2n 2n(2n + 2) + ~3 + ~ 3.6 " +

2n(2n + 2) (2n + 4)

3.6.9
3.6.9

+

~2
8.

V
?

+ 3 + _ 3T6~ +

+
J

Prove that
7'

(^-1) h + i + ^ 7.14 + n{n-\)(n-2) + + 7.14.21 7
.

1

J

J

±n Ji

(2
.

%

j.

»(*+!) , n(» + l)(n + 2) + + 2.4 2.4.6
"

\
J

9.

Prove that approximately, when x

is

very small,

"7!
2

1+
(

9

\2

256'

'

r 'V 6
+2
>JQ)

10.

Shew that the

integral part of (5

n is odd, if

n be a n be a

positive integer.
11.

Shew

that the integral part of (8 + 3 V/7) H

is

odd,

if

positive integer.
12.

Find the

coefficient of

xn

in the expansion of
)-*.

(l-2.v + 3.v2 -4.v 3 +
13.

Shew

that the middle term of

/
(

x+.

1\ 4 1
'

1

is

equal to the coefficient

of

xn

in the expansion of (1

-Ax)

^"

2

14.

Prove that the expansion of
(1

- xf n + 3nx (1

— x^) n may be put into ) - xf n ~ 2 + 3n @ n 3 x i (i _ xyn - 4 +
(1

the form

\

3

_

BINOMIAL THEOREM.
15.

ANY INDEX.
ill

L69

Prove that the coefficient of
1

at*

the expansion
1,


1.

,

is

1,0,

-

according as n

is

of the form 3m,

3m (1)

or 3//<.+

(2)

In the expansion of (a + b + c) s find the sum of the coefficients of the terms.
16. 17.

the

number

of terms,

Prove that

l\

n-l
,

111
if
\'S

n be an even
\b\

integer,
1

2"" 1
,1
|rc
'

n-

n-5

u-\
\

18.
(1

If c
u
,

(',,

+.f)

when n
a)
(2)

is

the coefficients in the expansion of a positive integer, prove that
C2 , fn are

I//-1
c

-c l+ c 2 -c3+

+(-mv-(-i)'1/

,
|

;^r_ 1

.

^-2^ + 3^-4^+
c*- c *+c£-c*+
is

+ (_i)n (/i4

.

1)t M
.

= 0>
(-1)^,

(3)

+ (-l)»cn2 =0,

or

according as n
19.

odd or even.

If *„
(1)

denote the

sum

of the first n natural numbers, prove that

(l-;r)- 3 = ^ + %^ + ^.^+
2 (*! *, B + 82*2, _j +

+V»~
l)

1

+...

(2)

+ 8n8n +

= -—

j2^ + 4

—^

.

20.

T.

1.3.5.7
fr(!)

(2)i-l)
,

If

2 .4. 6
?2n +
l

8

2n
1

.

P*>™

that

+ <Mj» + Man - +
- ?1 y,
(l

+ 2n- \<ln + 2 + ?«?• +1 = 5+
+ - 1)" (
1

(2)

2 { ?2n

_

j

+ g^a.

2

tj

n_

#„ + J

in

Find the sum of the products, two at a time, of the the expansion of (1 +x)n when n is a positive integer.
21.
,

coefficients

a

n If (7 /3, where n v/3) proper fraction, shew that (1 -f3)(p

22.

+4
,

=p +

+ p) = l.

and p are positive

integers,

and

|9

23.
(1

If c

<?!,

c^,

+#)*, where

?i

is

are the coefficients in the expansion of a positive integer, shew that
rn

c2

c
.

\

(-I) n_1 fn
n

,11
2

1

2

3

3

n

,

CHAPTER

XV.

Multinomial Theorem.
have already seen in Art. 175, how we may apply the Binomial Theorem to obtain the expansion of a multiIn the present chapter our object is not nomial expression. so much to obtain the complete expansion of a multinomial as to find the coefficient of any assigned term.
194.

We

Example.

Find the

coefficient of
(a

a4 b"c 3 d 5 in the expansion of
.

+ b + c + d) u

the product of 14 factors each equal to a+b + c + d, and every term in the expansion is of 14 dimensions, being a product formed by 5 4 2 taking one letter out of each of these factors. Thus to form the term a b c*d , we take a out of any four of the fourteen factors, b out of any hco of the remaining ten, c out of any three of the remaining eight. But the number of ways in which this can be done is clearly equal to the number of ways of arranging 14 letters when four of them must be a, two 6, three c, and five d that is, equal to

The expansion

is

;

114

412
This
is

A

TTralg 3 5

.

[Art. 151.] J L

therefore the number of times in which the term a 4 b 2c*d 5 appears in the final product, and consequently the coefficient required is 2522520.

To find pansion of (a + b +
195.

the coefficient of any assigned term in the exp where is a positive integer. c + cl +
.

..)

,

p

The expansion is the product of p factors each equal to a + + c + cZ + ..., and every term in the expansion is formed by taking one letter out of each of these p factors ; and therefore the number of ways in which any term a a b^cyd8 ... will appear in the final product is equal to the number of ways of arranging p letters when a of them must be a, (3 must be b, y must be c; and so on. That is,
ft

the coefficient of

aabPcyds
a+
j3

...

is

=

— o ~f~^ p \y 6
\p
...

where

+ y+

S

+

...

=p.

MULTINOMIAL THEOREM.
Cok.
Jn the expansion of
(a

171

+ bx +

cx~
...

+ da? +
is

...

)'',

the term involving a"b&cyd 6

^L-.a-^^v^).
or

i—T5-Ht^t/3

aWcyd5

...

xfi+2y+M +

..

where a +
This

+y+

$

+

...

= p.
general term of the expansion.
expansion of
(a

may be

eallecl tlte

Example.

Find the

coefficient of «* in the

+ i.c + ex 2

)-'.

The general term

of the expansion is

-ii-«a
where a + p + y = \).

^V

+2?

(i),

We
which

have to obtain by trial all the positive integral values of /3 and 7 satisfy the equation fi + 2y = 5; the values of a can then be found from
a + /3 + 7 = 9.
/3

the equation

Putting 7 = 2, we have

= 1, and a = G; putting 7 = 1, we have /3 = 3, and a = 5; putting 7 = 0, we have /3 = 5, and a = 4.
sum
of the corresponding values of the

The required
expression
(1).

coefficient will be the

Therefore the coefficient required
|9
19 19

= 252a 6 6c 2 + 5Q4a*&c + 12Ga 4 b\
19G.

To find

the general
(a

term in the expansion of
clx
3

+ bx + ex 2 +

+

n
. .

.)

,

vjhere

n

is

any

rational quantity.
is

By

the Binomial Theorem, the general term
,,_

n(n-l)(n-2)...(n-p + l)
(

v

+ rf +

^+

where jp

is

a positive integer

172

HIGHER ALGEBRA.
195, the general term of the expansion of
(6a;

And, by Art.

+

ex'

+ dx 3 +

...)''

\P

\pjy_\o_—

where

ft,

y, 8

.

.

.

are positive integers whose

sum

is p.

Hence the general term
pression
is

in the expansion of the given ex-

where
197.

/?

+y+

S

+

...

= /?.

Since (a + bx

+ ex 2 + dx 3 +
+
c
2

..)"

may
...

be written in the

form

„A 6 ail +-x
\
it will

a

-x+-ar+
3

a*

y
,

a

a

J

be sufficient to consider the case in which the hrst term of the multinomial is unity.

Thus the general term
(1

of
. .

3 + bx + ex 2 + dx +

.)"

is

n (n-l)(n-2). \p \v

(n-p +
8
\

l)

bpcyd8

^ +9f+u+

where
Example.

fi

+ y + &-\-...=p.
x 3 in the expansion of

Find the

coefficient of

(l-3z-2.r 2 + 6x' 3 )3.

The

general term

is

S(S-0(t-»)...(|-* + o V
,.,
,,

iO-sA-^e) /-^^
8

(i).

have to obtain by trial all the positive integral values of /3, 7, 5 which satisfy the equation j3 + Zy + 35 3 ; and then p is found from the equation 5. The required coefficient will be the sum of the corresponding 2>=/3 + 7 +

We

=

values of the expression

(1).

MULTINOMIAL THEOREM.

173

In finding /3, 7, 5, ... it will be best to commence by giving to 5 successive integral values beginning with the greatest admissible. In the present case the values are found to be

8=1,

= 0, 5 = 0,
5

= 0, p=l; 7 =1, 0=1, p=2; 7 = 0, 0=3, p = 3.
7 = 0,
18

Substituting these values in

(1)

the required coefficient

^)<^)(->)<- 3"- 2>+^#^ (-3)
4_4_4
3

s

3~3
it is

198.

Sometimes

more expeditious

to use the Binomial

Theorem.
Example.

Find the

coefficient of

x 4 in the expansion

of (1

- 2x + 3.r 2 ) -3

.

found by picking out the coefficient of x x from the first few terms of the expansion of (1 - 2x - Sx 2 ) * by the Binomial Theorem that is, from
coefficient is
;

The required

1

+ 3 {2x - Sx 2 + 6
)

(2.r

- 3x2 ) 2 + 10 {2x -

3.r-) :J

+ 15

(2.r

-

3.r

2 4
)
;

we stop
than
x*.

at this

term

for all the other

terms involve powers of x higher
2

The

required coefficient = 6

.

9

+ 10

.

3

(2)

(

-

3)

+ 15

(2)

4

=

-66.

EXAMPLES. XV.
Find the
1.

coefficient of

a 2 Pc4 d in the expansion of
ry

(a-b — c+d) w
.

.

2.

a 2 b d in the expansion of (a + b — c — d) s
a?b s c in the expansion of (2a + 6-f 3c) r
.

3. 4.
5.

x~yhA in the expansion of

{cub

- by + cz)9
.

.

x3

in the expansion of

(l+3# — 2a2 )3
(l
)

6.
7.

xA in the expansion of
.'•"

in tlie

expansion of
(if

8.

A"

8

in the expansion

+ 2.r + 3.r 2 10 2 (1 + 2.v - x 2 (1 - 2.r + 3# - 4.r'
.

)'\

!

4

)

.

.

174
Find the
9.
.r

HIGHER ALGEBRA.
coefficient of

23

in the expansion of (1 -

2x + 3x2 - x4 - .i/') 5
i

.

10.

x

5

in the expansion of (1
in the expansion of (1

-2x + 3x

2
)

2
.

i

11.

x

3

- 2x + 3x 2 - 4a3 ) 2 - X + X*\) —
'-»

.

12.

x8

in the expansion of

(
(

2

~2
.

1

13.
14. 15.

x* in the expansion of (2 -

4x + 3x2 ) ~ 2
2

.

3

X in the expansion of x 12

s

(

1

+ Ax + 1 Ox + 20^G ) " *
4

in the expansion of (3
i

- 15x* + 18^') - l x2
.

.

16.

Expand

(1

- 2x - 2x
2

2

)*

as far as
2

17.

Expand Expand
If

(1

+ 3x - 6x*)

3

as far as

x5

.

4

18. 19.

(8

- 9^ + 1 8a4 ) 3
"

as far as

x8
l

.

(l+x + x2 +

+xP) n = a

+ a x + a.^v +
2

a xn
llf>

r>,

prove that
(1)

a

+a1 +aa +

+a^=(p+l)n

.

(2)

a1 +2a2 +3a8 +
a
2
)
,

+«p.a«p=5»i>(p+l)*.

20.

If

of

(1

+x+x
a
2

a 15 a2

>

ft 3 •••

are the coefficients in order of the expansion

n
,

prove that

-a 2 + a 2 -a 2 +
If the

+ (-l)n - 1 aU 1 =^an {l-(-l)^an}.
(1
)

21.

expansion of

be

a
«o + a 3 + a6

+x + x2 n + a x+a 2x2 + ... +araf +
r l

...

+a 2n x2n
8

,

shew that

+

...

=a + a + a-+
l

4

...

=« + a6 +a +
2

...

= 3 n_1

.

;

CHAPTER

XVI.

Logarithms.
The logarithm of any number to a given Definition. base is the index of the power to which the base must be raised Thus if a x = JV, x is called in order to equal the given number. the logarithm of to the base a.
199.

N

Examples.

(1)

Since 3 4 = 81, the logarithm of 81 to base 3

is 4.

(2)

Since

lO^lO,
3,...

10 2 = 100, 103 = 1000,

the natural numbers to base 10. 1000,

1, 2,

are respectively the logarithms of 10, 100,

The logarithm of iV to base a is usually written log a jy, 200. so that the same meaning is expressed by the two equations
x a =

N; x = \og

a

N.

From

these equations

we deduce

an identity which
Example.

is

sometimes

useful.
to base 2 N/2.

Find the logarithm of 32 £/!

Let x be the required logarithm; then,

by

definition,

(2

x/2)« = 32 4/4
1

2

.-.

(2. 2*)*
3
.-.

=2

s
.

2*
2

2^ = 2 5 ^;
3 27 - x = -r
;

hence, by equating the indices,

.'.

x=

—= o

3-6.

.

)

176
201.

HIGHER ALGEBRA.

understood that a particular system of logarithms is in use, the suffix denoting the base is omitted. Thus in arithmetical calculations in which 10 is the base, we instead of log 10 2, log l0 3, usually write log 2, log 3,
it

When

is

might be taken as the base of logarithms, and corresponding to any such base a system of logarithms of all numbers could be found. But before discussing the logarithmic systems commonly used, we shall prove some general propositions which are true for all logarithms independently of any particular
base.

Any number

202.

TJie

logarithm of
all

1 is 0.

For
the base
203.

= 1 for may be.

values of a

;

therefore log

1-0, whatever

The logarithm of the base
1

itself is 1.

For a = a
201.

;

therefore loga a

=

1

To find

the logarithm
let

of a product.

Let
suppose

MN be the product;
a:

a be the base of the system, and

=

log. J/,

y = \oga J\T;
a*

so that

a*

= M,

= N.
ay

Thus the product

MN==ax x

= ax+y ;
whence, by definition, log a

MN = x + y
= 100^1/"+
low

N.

Similarly, \og a 3INP

= \oga M+

loga iV+ loga P;

and so on

for

any number

of factors.

Example.

log 42 = log (2 x 3 x 7

= log2 + log3 + log7.
205.

To find

the logarithm

of a fraction.

Let

M be the
-zz

fraction,

and suppose

so that

x= ax =

\oga

M
i

2/

= loga iT;

M

t

ay = N.

LOGARITHMS.
Thus the
fraction

whence',

by

definition,

loga

*—-=x — y

'

178
Example
log
c.

HIGHER ALGEBRA.
1.

Express the logarithm of

a —^ m

r

3

terms of log

a, log b

and

.

Ja*

a2

3

= log a 2 - log (c^ 2
5

)

=3 -log«-(logc + logfc
3

2
)

= = log a - 5 log c - 2 log &.
Example
2.

Find

a;

from the equation a x c~2*=&3a!+1
.

.

Taking logarithms of both

sides,

we have
;

x log a - 2x
.-.

x

= (Sx + 1) log 6 (log a - 2 log c - 3 log b) = log b
log c

;

_

lo 2 6

log a - 2 log

c

- 3 log

b

EXAMPLES. XVI.
Find the logarithms of
1.

a.

16 to base J2, and 1728 to base 2 v '3.

2.

125 to base 5 v/5, and *25 to base
stt. to

4.

3.

256

base 2 x/2, and
2,

'3

to base

9.

4. 5.

'0625 to base

and 1000
and i
2

to base -01.
to base 9^/3.

-0001 to base '001,
4

6.

kI gp

/~*r
,

i

3
,

j

*/

r~-^
a
to base a.

a?
7.

Find the value of
l0g8 128, l0g6

^,

log-frgj,

log3 43 49

'

Express the following seven logarithms in terms of log a, logb, and
logo.
8.

log(N^)

fi

.

9.

log{Va 2

xyb s

).

10.

logflcFW).

"

=
179

LOGARITHMS.
11.

log^o^x^oJR).
log

12.

log(^a

V63 -j-VP7a).
f

13.

14.

logj^J +
2

ffi

3

15.

Shew

that log |f

'

;

£^ = 1

VW7J2
72!)

4

logo - | log2
5

3

,

log 3.

16.

Simplify

logV

V

9" 1 27 3
.

.

17.

Prove that log

75 — - 2 log 5 + -

1<

>g

—=
39

h .g

2.

Solve the following equations:
18.

o«=c&*

19.

a2».63*=cs
'

.

90 U
*

°^ - "& &*-*
c

21

-

a 2* m6 3*.62 a »=m10J
• l

^

'

22.

If \og(x'2y 3 )

=a

}

and log- = 6,
b 3x ,

find log*-

and

log//.

23.

If

a3 " *

.

V> x

= ax + \

shew that x log (-") = log a.

24.

Solve the equation
(a*

- 2a*b* + b*) x - 1 = (a -

ft)

2*

(a + 6) -

-'.

Common Logarithms.
Logarithms to the base 10 are called Common Logarithms; this system was first introduced, in 1615, by Briggs, a contemporary of Napier the inventor of logarithms.
208.

the equation 10 - JV, it is evident that common logarithms "will not in general be integral, and that they will not

From

x

always be positive.

For instance
.*.

3154 > 10 and < 10 4
'

;

log

3151«=3 + a

fraction. ion.

12—2

.

180
Again,
.*.

HIGHER ALGEBRA.
-06

> 10~ 2 and < 10 _l

;

log *06

=-

2

+a

fraction.

The integral part of a logarithm is called Definition. 209. the characteristic, and the decimal part is called the mantissa.
characteristic of the logarithm of any number to the base 10 can be written down by inspection, as we shall now shew.

The

210.

To determine

the characteristic

of the logarithm of any

number

greater than unity.

Since

10 - 10,
1

10 =100, 10 -1000,
3

2

number with two digits in its integral part lies 2 between 10' and 10 a number with three digits in its integral 3 2 Hence a number part lies between 10 and 10 ; and so on. 10" _I and 10". with n digits in its integral part lies between
it

follows that a

;

Let
then

N be

a

number whose

integral part contains

n

digits;

J\T— in(tt-l)+ a fraction

.*.

log

iV=

(n — 1) + a fraction.

characteristic is n — 1 \ that is, the characteristic of the logarithm of a number greater than unity is less by one than the number of digits in its integral part, and is positive.

Hence the

To determine the characteristic of the logarithm of 211. decimal fraction.
Since

a

10°=

1,

1(rs

=iJcr

01 '

10-8

=i=' 001

>

LOGARITHMS.
it

181

follows that a decimal with one cipher immediately after the decimal point, such as -0324, being greater than -01 and less -1 than -1, lies between 10~ 2 and 10 ; a number with two ciphers after the decimal point lies between 10 _:i and 10""; and so on. Hence a decimal fraction with n ciphers immediately after the decimal point lies between 10~ " + 1) and 10~".
(

Let

D

be a decimal beginning with n ciphers
/)
.-.

;

thou

_

1

f)~(w + l) +

fraction.

log J)

= — (n +
is

l)

+n

fraction.
1)
;

Hence the
of

characteristic

- (n+

that

is,

the characteristic

of a decimal fraction is greater by unity titan the number of ciphers immediately after the decimal point, and is
the logarithm

negative.

212.

The logarithms

to base 10 of all integers from
j

1

to

200000 have been found and tabulated
given to seven places of decimals. This use, and it has two great advantages
:

in
is

most Tables they are the system in practical

the results already proved it is evident that the characteristics can be written down by inspection, so that only the mantissse have to be registered in the Tables.
(1)

From

The mantissse are the same for the logarithms of all numbers which have the same significant digits; so that it is
(2)

sufficient to tabulate the mantissse of the logarithms of integers.

This proposition we proceed to prove.

Let be any number, then since multiplying or dividing by a power of 10 merely alters the position of the decimal point without changing the sequence of figures, it follows -~ 10 7 that x 10''. and where p and q are any integers, are numbers whose significant digits are the same as those of N.
213.

N

N

N

,

Now
Again,

log

(N x

10 p )

= log N+p log 10

= log J\r +p
log (AT 1
9

(1
1

).

)

- log N - q log

= logiV-7
an integer subtracted from log
(1)

(2).

In

is
;

added to
that
is,

logiV^,

and

N

in (2)

an integer

is

the mantissa or decimal portion

of the logarithm remains unaltered.

;

;

182

HIGHER ALGEBRA.

In this and the three preceding articles the mantissse have been supposed positive. In order to secure the advantages of

we arrange our work so as always to keep the mantissa positive, so that when the mantissa of any logarithm
Briggs' system,

has been taken from the Tables the characteristic is prefixed with its appropriate sign according to the rules already given.

In the case of a negative logarithm the minus sign is written over the characteristic, and not before it, to indicate that the characteristic alone is negative, and not the whole expression.
214.

Thus 4-30103, the logarithm of -0002, is equivalent to -4 + -30103, and must be distinguished from — 4*30103, an expression in which both the integer and the decimal are negative. In working with negative logarithms an arithmetical artifice will sometimes be For instance, necessary in order to make the mantissa positive. a result such as - 3*69897, in which the whole expression is negative, may be transformed by subtracting 1 from the Thus characteristic and adding 1 to the mantissa.

- 3-69897 - - 4 +

(1

- -69897) = 4-30103.

Other cases will be noticed in the Examples.
Example
1.

Required the logarithm of -0002432.

In the Tables we find that 3859636 is the mantissa of log 2432 (the decimal point as well as the characteristic being omitted) and, by Art. 211, the characteristic of the logarithm of the given number is - 4
;

.-.

log -0002432 = 4-3859636.

Example

2.

Find the value of ^-00000165, given
log 165 = 2-2174839, log 697424=5-8434968.

Let x denote the value required
log
a-

;

then
I

= l©g (-00000165) 5 =

l

= log (-00000165)
o

= i (6-2174839)
the mantissa of log -00000165 being the characteristic being prefixed by the rule.

same as that of log

165,

and the

Now

- (6-2174839)

=-

(10

+ 4-2174839)

= 2-8434968

;

'

J

LOGARITHMS.

183

and -8434908 is the mantissa of log 007424; hence x is a number consisting of these same digits but with one cipher after the decimal point. [Art.
211.

Thus
215.

a:

= -0097424.

The method of calculating logarithms will be explained in the next chapter, and it will there be seen that they are first found to another base, and then transformed into common logarithms to base 10.
It will therefore be necessary to investigate a method for transforming a system of logarithms having a given base to a new system with a different base.

216.

Suppose that the logarithms of
it is

are known and tabulated, to base b.

all numbers to base a required to find the logarithms

Let
quired.

N

be any number whose logarithm

to

base

b

is

re-

Let y = log6 iV, so that
••

b

y

=N
= logJT;

log. (&")

that

is,

ylog£ = log,JV;
J — log r X log N. 0u a6
1/
'

or

k&^wK*
1U Oa°

10

^
log b are
all

C

1 )-

and b are given, los: N and from the Tables, and thus log^V may be found.

Now

since

N

known

Hence
to base b

it

appears that to transform logarithms from base a
to multiply r J
is

we have only J

them

by J

;

this

is

a

log b

constant quantity and modulus.
217.

given by the Tables;

it is

known

as the

Tn equation
.

(1) of the preceding article

put a for

N\

thus
i
On
,

i

Oa

log/t x log 8 /j

=

1

.

:

184
This result

HIGHER ALGEBRA.

may

also be

proved directly as follows
a*

Let

x = log/?, so that
b,

=b

;

then by taking logarithms to base

we have

x \og b a = \ogb b

.-.

loga 6xlog 4 a

=

l.

218.

The following examples

will illustrate

the utility of

logarithms in facilitating arithmetical calculation ; but for information as to the use of Logarithmic Tables the reader is referred to works on Trigonometry.
4

5

Example

1.

Given

log 3 = -4771213,

find log {(2-7) 3 x (-81)»-H90)*}.

The

required value

= 3 log j= +

27

4 81 5 - log - log 90 1Q
-=

= 3(l<^3»-l)+|oog3*-2)-|(log3*+l)

KM)"*a-K +
= ^log3-5H

t)

= 4-6280766-5-85
= 2-7780766.
notice that the logarithm of 5 powers can always be obtained from log 2 ; thus
log 5

The student should

and

its

= log

— = log 10 - log 2 =
digits in

1

-

log

2.

Example

2.

Find the number of
log 2

875 1C given
,

= -3010300,

log 7

= -8450980.

log (875 16 )

= 16 log (7x125) = 16 (log 7+ 3 log 5) = 16(log7 + 3-31og2) = 16x2-9420080
=47-072128;
is 48.

hence the number of digits

[Art. 210.]

.

;

;

;

LOGARITHMS.
Example 3. Given log 2 and log value of x from the equation
3,

185
two places of decimals the

find to

Taking logarithms of both
(3
.-.

sides,

we have
(x

- 4a) log G +

+ 5)

log 4 = log 8

(3

-

4.r)

(log 2

+ log 3) +

(x
2)

+ 5)

2 log 2

= 3 log 2
- 10 log 2

.-.

.r (

- 4 log 2 - 4 log 3 + 2 log
.r

= 3 log 2 - 3 log 2 - 3 log 3
+ 3 log 3 2 + 4 log 3

= 10

log 2

2 log

_ 4-4416639 ~2-al054a2

= 1-77...

EXAMPLES. XVI.
1.

b.

Find, by inspection, the characteristics of the logarithms of

21735, 23-8, 350, '035,

%

-87, -875.

The mantissa of log 7623 is '8821259 2. of 7-623, 762-3, -U07623, 762300, '000007623.
3.

;

write

down

the logarithms

How many

digits are there in the integral part of the

numbers

whose logarithms are respectively
4-30103,
4.

1-4771213,

3-69897,

-56515

1

(Jive the position of the first significant figure in the

numbers

whose logarithms are
2-7781513,
-6910815,

5-4871384.
7

Given log 2 = -3010300, log 3 ='4771213, log
value of
5.

= -8450980,
log -128.

find

the

log 64.
log -0125.

6.

log 84. log 14-4.

7.

8.

9.

10.

log 4^.
4

11.

log^l2.

12.

logW —
\
-i

13.

logN/ 0l05.
:

14.

Find the seventh root of -00324, having given that
log 44092388 = 7-6443036.

15.

Given log 194*8445 = 2'2896883,

find the eleventh root of (39-2) 2

.

186
16.

HIGHER ALGEBRA.
Find the product of 37-203, 3-7203, -0037203, 372030, having
log 37-203

given that

= 1-5705780,
3,

and log!915631 = 6:28231 20.
3

17.

Given log 2 and log

find log

//3 2 5**\ /( —y )

.

X

18.

Given log 2 and log 3, find log (#48 x 108

1

-f ^6).

19.

Calculate to six decimal places the value of

given log
20.

2,

log

3,

V 42 x 32 J log 7; also log 9076-226 = 3-9570053.

V

/294 x 125 \ 2
'

Calculate to six places of decimals the value of

(330^-49) 4 ^\/ 22x70;

given log

2,

log 3, log 7

;

also

log 11
21.

= 1-0413927,

and logl7814-1516 = 4-2507651.
digits in 3 12

Find the number of

x 28

.

22.

Shew

that

/21\ 100
(

is
J

greater than 100.

23.

Determine how many ciphers there are between the decimal
/1\ 1000
first significant digit

point and the

in

(

j

Solve the following equations, having given log
24.

2,

log 3,
55

and

log
.

7.

3*~ 2 = 5.

25.
1

5*

= 10l
28. 30.

26.

~ 3 *=2* + 2
1

27.
29.

2F = 2 2 * +
2x + y

.5 3

-.

2*. 6*- 2

=5 2 *. 7
\

"*.

= 6»
'

"I

3l

~ x - y =4-y ~l

3*
31.
32.

=3

22,

+ ij-

2 2x

=3 3j/_a;J

Given log 10 2 = -30103, find log25 200.

Given log 10 2 =

-.30103, log 10 7

= -84509,

find log 7N/2

and logV27.

CHAPTER

XVII.
SERIES.

EXPONENTIAL AND LOGARITHMIC
219.

In Chap. XVI. it was stated that the logarithms in common use were not found directly, but that logarithms .are first found to another base, and then transformed to base 10.

In the present chapter we shall prove certain formulae known as the Exponential and Logarithmic Series, and give a brief explanation of the way in which they are used in constructing a
table of logarithms.

220.

To expand a 1 in ascending powers of x.
if

By

the Binomial Theorem,

n>l,

K)"
=
1

+ nx - + n
.

1

nx(nx—\) v
r
.

——
)

1

2

-» n"

+

nx (nx — *

1) (nx- 2) J ~±
-

1
.

3

n

-s 6

+

x (x
I?

x (x

)

(x — -\
(i).

3
I

.

By

putting sb=1,

we

obtain

(-')•

188
hence the seizes
(1) is

HIGHER ALGEBRA.
the xih power of the series (2); that
is,

1

+ x+

,„

+

rz

3

and

this is true

indefinitely

however great n may increased we have

be.

If therefore

n be

1+ * +
The
is

x2
|2

+

x3
]3

+

xA
|_4

+

=

/-'ill 1+
1

+

(

U+ ^ ^
+

+

y

series

1

+
e

1

+

1

+

— +—+
1

1

usually denoted by

;

hence
5

,=
Write ex

l

+«+

x-

+

X3

+

X4

+

for x, then
6** =

-

CX 4

cV
tjj-

+

cV +
-ry
;

Now
obtain

let

e

e

=

«,

so that c

= log/*

by substituting

for c

we

a'

= l+x\og a +
e

—Vo
If

+

+

lr

This
Cor.

is

the Exponential Theorem.
the limit of

When n

is infinite,

(

1

+-

)

= e.

[See Art. 266.]

Also as in the preceding investigation,

it

may

be shewn that

when n

is

indefinitely increased,
(, 1+

x\ n

n)

=1+X+

-

X2
Y2

+

x3
]3

+

x4
\i

+

.

EXPONENTIAL AND LOGARITHM!*! SERIES.
tli.tt is,

180

when n

is infinite,

the limit of

(

+- — nj V
1
)

c

T
.

x
l>y putting

—=

1
,

we have

Now m

is

H)--K)~={K)7 when n (x\ — —
infinite
is infinite; n
*.

1

J

e

Hence the
221.

limit of

(1 (-.')' = e~\
)

In the preceding

article
is less

no restriction

is

placed upon

the value of x; also since -

than unity, the expansions we
intelligible.

have used give results arithmetically

[Art. 183.]

But there
deserves

is

another point in the foregoing proof which

notice.

We
nj

have assumed that when n
/ :)•••(• nj

is

infinite

/
the limit or

1\7
\

2\

r-V
n
J
.

\

xr
\r

is -r-

,

\r

for

all values

of r.

Let us denote the value of
iB a!

(

"3( ~3 {
a,

x-

r

-^r) by u
as

.

r

™,

Then
Since n

u

u —

H(*^)
z,

1 / = -lx

r— 1\ =
n
)

1

r \

J

r

n

+

1 —

nr

is infinite,

we have
. .

U —— = X u
,

r

;

that

is,

u ~ — ur
r
a

x

.

l

It is clear that the limit of
ft>

u

is r^-;

hence the limit of u 3

is

,-x-;

that of

uA

x
is .—r;

and generally that

of

ut

x
is .—
.

190
222.

HIGHER ALGEBRA.
The
series

ii

111
+

E

H

+

+

'

~@

which we have denoted by e, is very important as it is the base Logarithms to this to which logarithms are first calculated. base are known as the Napierian system, so named after Napier their inventor. They are also called natural logarithms from the fact that they are the first logarithms which naturally come into
consideration in algebraical investigations.

logarithms are used in theoretical work it is to be remembered that the base e is always understood, just as in arithmetical work the base 10 is invariably employed.
the series the approximate value of e can be determined to any required degree of accuracy ; to 1 places of decimals it is

When

From

found to be 2-7182818284.
Example
1.

Find the sum of the
-,

infinite series

1

1

1

We

have

e

= l + l+

--

+—+
ex ,

-rg

+

;

and by putting x = -

1 in

the series for

e

" 1=1 -

1+

i2-i3

+

n-

hence the

sum

of the series is

- (e

a

+ e~

x ).

Example
1

2.

Find the

coefficient of

x r in the expansion of

ex

- ax — x o

= (1 - ax - x 2

)

e~x

n n fi = (l-a,-^)|l-, + « «* + -^--

...

+ L_L_ +

(-l) rx r

1

j.

-

-

EXPONENTIAL AND LOGARITHMIC
The
coefficient required

SERIES.

101

(-!)>•
r

(-l)'-ia

(-1)'--

r-1

r-2

-l) r

{l + ar-r(r-l)},

223.

7V>

expand

log, (1

+

tt) ira

ascending powers of
8 3

\.

From

Art. 220,
a"
=-•

1

+ y 1< >g

e

r6

+

/r (loge a)

+

^ y ^

(log. c v

L
+x

2

|

4+
'

'

lii

this series write 1

for a; thus

(1 +x)'J

=

1

+ y log,

(1

+ *) +

f {log, (1 2

+ *)} +

£ {log
1

e

(1

+

a;)}

3

+

...

(1).

Also by the Binomial Theorem, when x <

we have
rf+
(2).

(i+«y-i+» +
Now

g^*+ y fry
is

1 >fr- 8)

in (2) the coefficient of

,+ 1.3* + 1.2J
.

+
rp&
\K/

1.2.3.4
•>••*

+

'

m

that

is,

£ — -- + —
%K/

r**^

«C»

-

+

2i

o

±

Equate

this to the coefficient of

y in

(1)

;

thus

we have

l0g t,(l

+Ct')

= 7J--+
t

-- -+
in ascending powers of
(2),
.t.

This

is

known
If

as the Logarithmic Series.
{log,, (1

Example.

x < 1, expand

+ x)} 9

equating the coefficients of y 2 in the series (1) and required expansion is double the coefficient of y'2 in
?/(!/-!) ____.r2 + y (y x
.

By

we

see that the

1) (y

2)

.

1.2.8

+
,

y(y-i)(y-2)(y-3)

1.2.3.4
1) (y - 2) (y

^+

'

that

is,

double the coefficient of y in

(y-l)(y-2) y-1 + 1.2* 1.2.3

* +

(y

-

3)

"1.2.3.4

+
}.

Thn8

{log.(l +

*)P=2{^-i(l+l)^(l + l + l)«*-

.

192

HIGHER ALGEBRA.

Except when x is very small the series for log e (l + x) 224. can, however, is of little use for numerical calculations. deduce from it other series by the aid of which Tables of Logar-

We

ithms

may

be constructed.

By

writing

for x

we

obtain log.
1 1

:

hence
1

lo S.(

n+1 )- lo& w =
for

S"2? 3^"
n 1
;

+

(1)

-

By

1

writing

x we obtain log e

hence, by changing

signs on both sides of the equation,
log 8 n

-

log e (n

-

1)

= - + s—a +

#7

—3 +

(2).

From

(1)

and

(2)

by addition,

log.("

+ l)-log,(n-l) =

2(-

+_+__+
...J

(3).

obtain log e 4 — log e 2, that is log e 2 ; and by effecting the calculation we find that the value of log 6 2- -69314718...; whence log e S is known.

From

this formula

by putting n =

3

we

Again by putting n =

9

we

obtain log e 10

— log e 8; whence we

findloge 10 = 2-30258509....

To convert Napierian logarithms

into logarithms to base 10

we

multiply by

.

=-j=

,

which
value

is

the modulus [Art. 216] of the
1

common
we

system, and

its

is
/x.

^' oOJjOoDk) J .


.

,

or '43429448...;

shall denote this

modulus by

In the Proceedings of the Royal Society of London, Vol. xxvn. page 88, Professor J. 0. Adams has given the values of e, /x, log e 2, log e 3, loge 5 to more than 260 places of decimals.
225. If we multiply the above series throughout by /x, we obtain formulae adapted to the calculation of common logarithms.

Thus from

(1), /x

loge (ra

+

1)

- /* log

e

?i

=£ _

^ + JL

-. ...

;

EXPONENTIAL AND LOGARITHMIC SERIES.
that
is,

193

log I0 ( M + 1)

-

log,

n=£-

-t-

+

Jt.

_

m
(2)
.

Similarly from
l

(2),

^-lo >-l)^ + ^ + + ^
g]

From
ot one of

either of the above results

we

see that

if

two consecutive numbers be known, the logarithm of the other may be found, and thus a table of logarithms can be ° constructed.
only needed to J2 ; +^ calculate the logarithms of prime numbers, for the logarithm ot a compose number may be
Sl

the logarithm

" ld

e

«*"«*«*

that the above formula are

logarithms of

its

component

obtained by adding together the

factors.

either 7^+ then find

In order to calculate the logarithm of any one of the smaller prime numbers, we do not usually substitute the number in either of the formula or (2), but we endeavour to find (1) some value ot n by which division may be easily performed, and such that
1

or

n-l

log(n+l)

the given number.
Example.

contains the given number as a factor. or log(w -l) and deduce the logarithm of °

We

Calculate log 2 and log 3, given ^=-43429448.
(2),

By

putting n = 10 in
1

we have the value

of log

10-

log 9; thus

-

2 log 3

= -043429448 + -002171472 + -000144765 + -000010857
+ -000000868 + -000000072 + -000000006
1-2
log 3 =-045757488,
;

log 3 = -477121256.

Putting M = 80 in
4 log 3 - 3 log 2 - 1

(1),

we obtain

log 81 -log 80; thus

= -005428681 - -000033929 + -000000283 - -000000003 = -908485024 - -005395032, log 2 = -301029997.
article

3 log 2

In
iog9
{7i

the

next

we

shall

give

another

reader is referred to Mr Glaisher's article on Logarithms in the hncyclopcvdia Britannica.
H. H. A.
I

+ l)-\ ge n which is often useful in the construction of Logarithmic Tables. For further information on the subject the

series

for

•>

x

x

194
226.

HIGHER ALGEBRA.
In Art. 223 we have proved that
2

3

log e (l

+

x)

=

x~2
2

+

3"-~-"'

changing x into -

cc,

we have
3

log.(l -«)=-*—2 ~

J"
x5


\

By

subtraction,
.

1

+x

{

a /

x3

Put

=

l-x
+
\) ;

-

= —

,

n
log. &e

so that

x—

-x

Zn +

= 1

;

we thus
+

obtain

los* (n oeV

n = 2< 5

(2w +

7

+

777^

l

3(2?i

+ l) 3

=

va

5(2w +

j-t=

^-r. 5

+

...}-.
J

l)
is

Note.

This series converges very rapidly, but in practice

not always

so convenient as the series in Art. 224.

227. chapter.

The following examples

illustrate the

subject of the

Example
that

1.

If a,

8 are the roots of the equation ax 2 + bx + c = 0, shew
a 2 ,02
-

\og(a-bx + cx'1 ) = loga+(a. + p)x

-J—x2 + a —^- x3 -...
.

3

ffi

o

Since a

+ 8=


a

,

a/3

= - we
,

a

have

a - bx + cx 2 =a {l

+

{a

+ B)x + aBx 2 )
(1

=a
.-.

(1

+ cur)

log (a

- bx + ex 2 = log a + log
)

+ px). (1 + ax) + log (1 + Bx)
B2x 2

a _ _-... + Bx-'^-+^--... = loga + ax- a-x' + a x
.

2

3

3

B3x 3

= \oga + {a + B)xExample
log (1
2.

a

^^

2

+ *+^
of
is

a

3

-

...

Prove that the coefficient

xn in the expansion
3.

of

+ x + x2

)

is


Q

2

n

or - according as

1

n

n

is

or

not a multiple of

log (1

+ x + x 2 = log-)

1-x 3
X6

—— =log
X9

(1

- x'3) - log
X 3r

(1

(

x)

X2

X3

xr

\

'

'

3

'

EXPONENTIAL AND LOGARITHMIC SERIES.
If

1

93
_I

n

is

a multiple of
first series,

3,

denote

it

by

3,-;

then the eoeffieient of *»
the second 3eries;

j,

from the

together with
or _ ?

I
g

fr

om

^ J
therefore

coefficient is

- n

+

n

n
3,

If

.

is

not a multiple „f
is .

*» does not ocour in the

first series,

the required coefficient

n

228.

To prove
if

that e is incommensurable.

For

not, let e

= ™ where m
1

and n are positive integers;
1

then

^ui..
7i

1
1

.

1

If

(w

|n+l

"

multiply both sides by \n;
'•

m ^irJ
But

= integer + -i- +

w+1

___J___
(n+l)(n+2)

*

1

(n+])(jw.2)(n+S)
1

+ -"

n+

—L + _____J_____
1

(n

+

l)(n+ 3)
it is

(n

+ 1 )~(^T2)^T3) +
*

*

'

is

a proper fraction, for

greater than

and

less

than the

geometrical progression
_

n+\
that
is,

+

_i
(n+1)
2
'

i__
\3 (n+iy +

•••

'j

less

than I; hence an integer
is

is

equal to an integer plus

a fraction, which

absurd; therefore

e is

incommensurable.

EXAMPLES. XVIL
1.

Find the value of

-0 + 5-7 2^3
2.

4

+ 5~6 ? -.-+...

Find the value of
2 . 22

2

^3

.

23

i72 4

+

5

.

25

"

* '

15

9

+

o
.

;

1

19G
3.

HIGHER ALGEBRA.
Shew
that
a?
\

a a? hge (n+a)-logt (n-a)=2l- + ^.+_+
*\*&
»</

/>»o
%A/

4.

if

y=*'-2 +-3 2
3

*t/

O"**

4 +•••»
. . .

shew that
5.

x =y +^- +
that

^+
' )

Shew
a

+~ 2

(-

:

V~^" /

+ )+^P 3 V

a

+...

= log a-log
e

e

Z>.

6.

Find the Napierian logarithm of

——

correct to sixteen places

of decimals.
7.

Prove that
Prove that

e" 1

1 = 2 (/ .- +

2
.-=

+ nr +....)

3

\

.

8.

iog,d + xr« ( i -»•)'-'=«
9.

(£ +
4
)

+

o
+
• •

+••••)

Find the value of
''- 2

H
10.

-f

j2

(-'

4

-#

+

|i

^-

//,m
-

Find the numerical values of the common logarithms of and 13; given ^ = -43429448, log 2 = '30103000.
11.

7, 1

Shew that

if

ax2 and


2

are each less unity

12.

Prove that
logc (l

+ 3a-+2^2 = 3.r)

—+—
S5X3

—+
65.iT4

...

and

find the general

term of the

series.

13.

Prove that
,

1

+ 3.?

5x2

and

find the general

term of the

series.

14.

Expand

—^— in a series of ascending powers of

x.

;

EXPONENTIAL AND LOGARITHMIC SERIES.
15.

197

Express 25

(e ix

+ e~ ix

)

in ascending

powers of

.r,

-where i'=

</

— 1.

16.

Shew that

17.

If a

and

/3

be the roots of

x 2 -jtxr + tf = 0, shew that
a 2 4- ft 2
n 3 4ft 3

18.

If .r<l, find the
1

Bum
2

of the series
,

3

4

.

19.
i

Shew that
i\ A + -) n =1 (1
-,

i

i

log,

nj

2(» + l)

2.3(>4-l) 2

3.4(>t

+ l) 3

""

20.

If log,

^,^^,2^ l+.r + .^ +
., x

,

3 3

De expanded in
coefficient of
n is

a series of ascending

.t-

powers of
the form

#,

shew that the
3 if

o:

1
if

n be odd,

or of

4m + 2. and
Shew
that

n be

of the

form 4m.

21.

1

+

23
]2

+
,

33
J3

+

43
(4

+ - =5e

-

22.

Prove that
2 log,

n - log,

+

1)

-

log, (»

_ 1)=-

+

—+ _+
"

23.

Shew

that


ft

+1

1


1

2(?i+l) 2
1_
2?i
2

—+
'

3(7i+I) 3
1

n
<)

3n 3
81
1()
'

24.
log,

If log,

= Yq

-

24
«,

loge

=~ 25
=1
1

?;

go

= C sheW
'

that

2 = 7a - 26 +

3c,

log, 3

a - 36 + 5c,

log, 5

= 1 6a - Ah + Vc

and calculate

log, 2, log, 3, k>g e 5 to

8 places of decimals.

CHAPTER

XVIII.

INTEREST AND ANNUITIES.

In this chapter we shall explain how the solution of 229. questions connected with Interest and Discount may be simplified by the use of algebraical formulae.
Value in their ordinary arithmetical sense ; but instead of taking as the rate of interest the interest on ,£100 for one year, we shall find it more convenient to take the interest on £1 for one year.
shall use the
Interest, Discount, Present

"We

terms

230. To find the interest given time at simple interest.

and amount of a given sum in a

Let
year,

P be

n
;

the principal in pounds, r the interest of £1 for one the amount. the number of years, I the interest, and

M

is

The interest Pnr that is,
Also

of

P

for one year is Pr,

and therefore

for

n years
(1).

/

=Pnr

M = P + I;
M=P(l+nr)
(1)
r,

that

is,

(2).
r, 7,

From
or P,
?i,

and (2) we see that if of the quantities P, n, M, any three be given the fourth may be found.
the present value

231.

To find

and

discount of a given

sum

due in a given time, allowing simple

interest.

Let

P be

the given sum,

r the interest of

£1

the discount, the present value, for one year, n the number of years.

V

D

I

'

INTEREST AND ANNUITIES.

199

Since V is the sum which put out to interest at the present time will in u years amount to P, we have

P=

V(\+nr);
+ nr

1

And

D=P1

P
+ nr
'

Pnr 1 + nr
The value of D given by this equation is called the true discount. But in practice when a sum of money is paid before it is due, it is customary to deduct the interest on the debt instead of the true discount, and the money so deducted is called the banker's discount; so that
Note.
Banker's Discount = Pnr.

True Discount =

Pnr
1

+ nr'

Example. The difference between the true discount and the banker's discount on £1900 paid 4 months before it is due is 6s. 8d.; find the rate per cent., simple interest being allowed.
Let r denote the interest on £1 for one year; then the banker's discount 1900r

1900
-

is

——

~~
,

^~

''

* and «. * the true adiscount
i

is

i*
1900r

""•

~3~

1900r

~3~

1
;

7~T~3 l r
+i
1900r 2 =3

whence
*'•

+ >-;
3800
'

r

_ ~ 1 ±Jl + 22800 ~
3800
,

1±151
152

Rejecting the negative value,
.-.

t,

.

1 we nave f—aSui = o?

«

rate per cent.

= 100r = 4.

To find the interest and amount of a given sum in a given time at compound interest.
232.

Let

P denote

n

the

number

the principal, 7? the amount of £1 in one year, the amount. of years, I the interest, and

M

.

200 The amount
of

HIGHER ALGEBRA.

the end of the first year is PR ; and, since this is the principal for the second year, the amount at the end of 2 x Similarly the amount at the or PR the second year is 3 end of the third year is PR and so on ; hence the amount in n years is PR" \ that is,

P at

PR R

.

,

.'.

M=PR"; I=P(R -l).
n

Note.

If r denote the interest

on £1

for

one year, we have

R = l+r.
In business transactions when the time contains a 233. fraction of a year it is usual to allow simjyle interest for the Thus the amount of ,£1 in ^ year is fraction of the year. v reckoned 1 + - ; and the amount of in 4f years at compound

P

interest

is

PR*

(1 + ^ r

.

Similarly

the

amount

of

J

P

in

n+

m years is PR" (In \

— m/

)

more than once a year there is a distinction between the nominal annual rate of interest and that actually received, which may be called the true annual rate ; thus if the interest is payable twice a year, and if r is the nominal
If the interest is payable

annual rate of

interest, the

amount

of

£1

in half a year

is 1

r +-^

,

and therefore in the whole year the amount
or
1

of <£1 is (1

+ -J,
is

+

r

+

r

2
;


4

so

that

the

true

annual

rate

of

interest

r

2

If the interest is payable q times a year, and if r is the nominal annual rate, the interest on .£1 for each interval is

234.

r
,

and therefore the amount

of

P

in

n

years, or

qn

intervals, is

In
(f

this case the interest is said to be "converted into principal"

times a year.

INTEREST AND ANNUITIES.
If the interest
is

201

convertible into principal every moment, then q becomes infinitely great. To find the value of the amount, r 1 put - = — so that q - rx thus x q
, :

the

amount =

P

(l

+-Y =P(l + -Y" = P {(l + i)T'
= Pe nr
,

[Art. 220, Cor.,]

since

x

is infinite

when q

is infinite.

235.

To find

the present value

due in a given time,

and discount of a given allowing comjwund interest.

stun

R

be the given sum, V the present value, the discount, the amount of £1 for one year, n the number of years.

Let

P

D

Since

V

is

time, will in

n

the sum which, put out to interest at the present years amount to P, we have

P=VR
it

n
',

and
Example.

D = P(l-R-).
The present value of £672 due in a certain time is £126; interest at 4£ per cent, be allowed, find the time; having given
log 2
if

compound

= -30103,

log 3 = -47712.

Here

Hol) = iI' *»**=!'
;

Let n be the number of years

then

672=126
.

y
;

;

25

•'•

?ll0g

=

.

672
°g
i26-'

1

24
.

or
.-.

100
96

?ilog

=logy

.

16

n (log 100 - log 96)

= log 16 - log 3,
4 log 2 - log 3
2 - 5 log 2

n=

- log 3
Veiy nea y
'

"
thus the time
is

=

•72700
-01773

=

'

very nearly 41 years.

202

HIGHER ALGEBRA.

EXAMPLES.
When
log 2 = -3010300, log 7
1.

XVIII.

a.

required the following logarithms

may

be used,

log 3 = '4771 213, log 11

= -8450980,

= 1-0413927.
cent,

Find the amount of £100 in 50 years, at 5 per

compound

interest; given log 114-674 = 2-0594650.

interest the interest on a certain sum of money is ,£90, and the discount on the same sum for the same time and at the same rate is £80 ; find the sum.
2.

At simple

3.

In how

many

years will a
?

sum

of

money double

itself at 5

per

cent,

compound

interest

Find, correct to a farthing, the present value of £10000 due 4. years hence at 5 per cent, compound interest given 8 log 67683-94 = 4-8304856.
;

5.

In

how many
interest
?

years will £1000 become £2500 at 10 per cent,

compound
6.

mean
7.

at simple interest the discount between the sum due and the interest on it.

Shew that

is

half the harmonic

Shew that money will increase more than a hundredfold a century at 5 per cent, compound interest.
8.

in

What sum
to

amount

£1000

of money at 6 per cent, Given in 12 years ?

compound

interest will

log 106
9.

= 2-0253059,

log 49697

= 4-6963292.
:

borrows £600 from a money-lender, and the bill is renewed every half-year at an increase of 1 8 per cent. what time will elapse before it reaches £6000 1 Given log 1 18 = 2-071882.

A man
"What

amount of a farthing in 200 years at 6 per cent, Given log 106 = 2-0253059, log 11 5-0270 = 2-0611800. compound interest?
10.
is

the

Annuities.
*

An annuity is a fixed sum paid periodically under 236. certain stated conditions ; the payment may be made either once Unless it is otherwise a year or at more frequent intervals.
stated

we

shall suppose the

payments annual.
;

annuity certain is an annuity payable for a fixed term of a life annuity is an years independent of any contingency annuity which is payable during the lifetime of a person, or of the survivor of a number of persons.

An

1

;

;

INTEREST AND ANNUITIES.

203

deferred annuity, or reversion, is an annuity which does not begin until after the lapse of a certain number of years ; and when the annuity is deferred for n years, it is said to commence after n years, and the first payment is made at the end of n + 1
years.
If the annuity is to continue for ever it is called a perpetuity
if it
;

A

does not commence at once

it is

called a deferred perpetuity.
is

annuity left unpaid for a certain number of years to be forborne for that number of years.
237.

An

said

To find

the

amount of an annuity
interest.

left

unpaid for a given

number of years, allowing simple
Let

be the annuity, r the interest of £1 for one year, n the number of years, the amount.

A

M

the end of the first year A is due, and the amount of this sum in the remaining n - 1 years is A + (n — 1) rA ; at the end of the second year another A is due, and the amount of this sum in the remaining (a— 2) years is A + (ii — 2) rA and so on. Now is the sum of all these amounts

At

M
.-.

M={A + (n-l)rA} + {A + (n-2)rA} +
n terms
2
.-.

+ (A + rA) + A,

the series consisting of

J/=Wil+(l +
= nA +
K

+ 3+
'-

+n-l)rA

—— —

rA.

238. To find the amount of an annuity left given number of years, allowing compound interest.

unpaid for a
one year, n

Let
the

A

be the annuity,
of years,

number

M the amount.
AR

R

the

amount

of <£1 for

the end of the first year A is due, and the amount of this n~ sum in the remaining n— 1 years is at the end of the ; second year another A is due, and the amount of this sum in the n~ 2 remaining n - 2 years is and so on.

At

AR

x

;

.-.

M = AR
.

n

~

x

+ AR"- +
2

2

+AR
to

2

+

AR + A

= A(l

+R + R

+

n terms)

Rm =A R^l-

;

204
239.

HIGHER ALGEBRA.
In finding the present value
of annuities it is always

customary to reckon compound interest; the results obtained when simple interest is reckoned being contradictory and untrustworthy. On this point and for further information on the subject of annuities the reader may consult Jones on the Value of Annuities and Reversionary Payments, and the article Annuities
in the Encyclopaedia Britannica. 240.

a

(jiven

To find t/ie present value of an annuity to continue for number of years, allowing compound interest.

be the annuity, R the amount of £>\ in one year, n the number of years, V the required present value.

Let

A

The present value
the present value of
the present value of

of

A due

in

1

year

is
is

AR~
AR~

l

;

A due

in 2 years

AR~'J
3

;

A

due in 3 years

is

;

and so

on.

[Art. 235.]
is

Now V
payments
.-.

the

sum
1

of the present values of the different

V=AR- +AR= AR~
l

3

+ AK- 3 +

tow terms

1

- R~"
1

l-R-

=A
Note.
This result
given in Art. 238, by

1-R-"

R-l
may
.

Rn

also be obtained [Art. 232.]

by dividing the value of M,

Cor. If we make n infinite of a perpetuity

we

obtain for the present value

R-V
mA

'

r

is the present value of 241. If years' purchase. is said to be worth

an annuity A, the annuity

m

In the case

of a perpetual annuity
1

mA — —

;

hence

m=r

=

100
rate per cent.


INTEREST AND ANNUITIES.
205
is

that is, the number of years' purchase of a perpetual annuity obtained by dividing 100 by the rate per cent.

instances of perpetual annuities we may mention the income arising from investments in irredeemable Stocks such as

As

many Government

Debentures. nished by the number of years' purchase of its Stocks ; thus the 2| p. c. Consols at 96} are worth 35 years' purchase Egyptian 4 p. c. Stock at 96 is worth 24 years' purchase while Austrian 5 p. c. Stock at 80 is only worth 16 years' purchase.
;

A

Securities, Corporation Stocks, and Railway good test of the credit of a Government is fur-

;

242.

To find
at the

the

present

value
to

commence
Let

end of p years and

of a deferred annuity to continue for n years, allow-

ing compound interest.

A

be the annuity,

R

the

amount
at

of

£1
end

in one year,

V the

present value.

The

first

payment

is

made

the

of

(;>

+

l) years.

[Art. 236.]

Hence the present values
ments are respectively
{

of the

first,

second, third... pay-

AR- * +l \ AR- (p+2 \ AR- (p+3 \
.'.

...

V=AR- (p+l) + AR- {p+ »+AR1

(1,+3

>+

ton terms

= AR~ (p+1)

-

7?~"

AR~ V

A R- p -"

Cor. The present value of a deferred perpetuity to commence after p years is given by the formula
V

~R-V
is

243.

A freehold

estate

an estate which yields a perpetual
is

annuity called the rent ; and thus the value of the estate to the present value of a perpetuity equal to the rent.

equal

It follows from Art. 241 that if we know the number of years' purchase that a tenant pays in order to buy his farm, we obtain the rate per cent, at which interest is reckoned by dividing 100 by the number of years' purchase.

~

206

HIGHER ALGEBRA.

Example. The reversion after 6 years of a freehold estate is bought for £20000; what rent ought the purchaser to receive, reckoning compound Given log 1-05 = -0211893, log 1-340096 = -1271358. interest at 5 per cent. ?

The
years,

rent is equal to the annual value of the perpetuity, deferred which may be purchased for £20000.
'

for 6

Let

£A

be the value of the annuity; then since

.R

= l-05, we

have

20000^* ^° 5
•0o

'- 6
;

A x (1-05) 6 = 1000; log A -6 log 1-05 = 3,
.-.

log
.-.

A = 3-1271358 = log

1340-09G.
is

A = 1310-096, and

the rent

£1340.

Is. lid.

Suppose that a tenant by paying down a certain sum lias obtained a lease of an estate for p + q years, and that when q years have elapsed he wishes to renew the lease for a term p + n years the sum that he must pay is called the fine for renewing n years of the lease.
244.
;

Let A be the annual value of the estate then since the tenant has paid for p of the p + n years, the fine must be equal to the present value of a deferred annuity A, to commence after
;

p years and

to continue for
.

n years

;

that

is,

the fine

n

AR-* = -= A—1

AR- p n A—1

~

.

[Art. 242.1

r

4

EXAMPLES.
The
1.

XVIII.

b.

interest is supposed

compound

unless the contrary

is stated.

person borrows ,£672 to be repaid in 5 years by annual instalments of ,£120; find the rate of interest, reckoning simple interest.

A

Find the amount of an annuity of ,£100 in 20 years, allowing compound interest at 4| per cent. Given
2.

log 1-045 = -0191163,

log24-117 = 1-3823260.
at what rent should it on the purchase money ?
;

be

freehold estate is bought for £2750 3. let so that the owner receive 4 per cent,

A
A

may

freehold estate worth 4. rate of interest.

£120 a year

is

sold for £4000; find the

INTEREST AND ANNUITIES.

207

How many years' purchase should be interest being calculated at 3i per cent.? estate,
5.

given for a freehold

6.

amount
7.

If a perpetual annuity is worth 25 years' purchase, find the of an annuity of £625 to continue for 2 years.
If a

perpetual annuity is worth 20 years' purchase, find the annuity to continue for 3 years which can be purchased for £2522.

the rate of interest is 4 per cent., find what sum must be paid now to receive a freehold estate of £400 a year 10 years hence; having given log 104 = 2-0170333, log 675565- -8296670.
8.

When

9.

Find what

sum

will

interest being payable every
10.

amount to £500 in 50 years moment; given e _1 = '3678.

at 2 per cent.,

n

If 25 years' purchase must be paid for an annuity to continue and 30 years' purchase for an annuity to continue 2?i years, find the rate per cent.
years,

man borrows £5000 at 4 per cent, compound interest if the 11. principal and interest are to be repaid by 10 equal annual instalments, find the amount of each instalment having given
; ;

A

log 1-04 =-01 70333
12.

and log 675565 = 5-829667.

has a capital of £20000 for which he receives interest at 5 per cent. if he spends £1800 every year, shew that he will be ruined before the end of the 17 th year; having given
;

A man
log 2

= '3010300,

log 3

= '4771213,

log 7 = '8450980.
;

of an estate is £500 if it is let on a lease of 20 years, calculate the fine to be paid to renew the lease when 7 years have elapsed allowing interest at 6 per cent. having given
13.
;

The annual rent

logl06 = 2-0253059,
14.

log4-688385 = '6710233,

log3'118042 = '4938820.

If a,

b, c

tinue n,

2/i, 3?i

years' purchase must be paid for an annuity to conyears respectively; shew that

a2 — ab + b 2 = ac.

What is the present worth of a perpetual annuity of £10 15. payable at the end of the first year, £20 at the end of the second, £30 at the end of the third, and so on, increasing £10 each year; interest being taken at 5 per cent, per annum ?

c

; ;

;

CHAPTER

XIX.

INEQUALITIES.

245.

Any

quantity a

quantity b when a -b is because 2 - (- 3), or 5 is positive. Also b is said to be less than a when b- a is negative; thus -5 is less than -2, because — 5— (— 2), or - 3 is negative.

said to be greater than another positive; thus 2 is greater than -3,
is

In accordance with this definition, zero must be regarded as greater than any negative quantity.
shall suppose (unless the contrary is directly stated) that the letters always denote real and positive quantities.

In the present chapter we

246.

If

a >

b,

then

it is

evident that

a+
a—
c

> > >

b b

+c —c

;

ac

be
b

a

will still hold after each side has been increased, diminished, multiplied, or divided by the same positive quantity.

that

is,

an inequality

247.

If
c to

a-ob,
each
side,

by adding

a>b+ c;
which shews that in an inequality any term

may

from one
If

be transposed

side to the other if its sign be changed.
b,

a >

then evidently

b

< a

that is, if the sides of an inequality be transposed, the sign of inequality must be reversed.

;

INEQUALITIES.
If a
is,

209

>

b,

then a - b

is positive,

and

b-a

is

negative; that

—a—

(—

b) is

negative,

and therefore

—a < —b

;

hence, if the signs of all the terms of the sign of inequality must be reversed.

an inequality

be changed,

Again,

if

a >

b,

then

—a

< — b, and
be

therefore

— ac < —

that is, if the sides of an inequality be multiplied by the same negative quantity, the sign of inequality must be reversed.
248.
If

a.>b,,

a.

>

boi

a^>b.,,

a

>b
...

,

it

is

clear

that

a + a 2 + a 3 +...+ am >
l

6,

+ b^+ba +
b r
,>

+ bm

;

and
249.
1

a a 2 a,--' a ,n
:

>h AK-'V

11

If

a>b, and

if

p, q are positive integers, then ^/a>^Jb,
V
a'
1

or a > b 9 ; and therefore positive quantity.

>

b'

;

that

1

is, a'

>

b'\

where n

is

any

Further,

—<

=-

;

that

is

n a~ <

b~".

The square of every real 250. Thus (a therefore greater than zero.
.

quantity

is

positive,
;

and

2

b)

2

is

positive

.

a
• .

2

2ab +
b
2

b

>

;

.

a2 +

>

2ab.

Similarly

-—^
mean of
mean.

> Jxy

;

that

is,

the arithmetic

tivo positive quantities is greater

than

their geometric

The
equal.

inequality becomes an equality

when

the quantities are

251. The results of the preceding article will be found very useful, especially in the case of inequalities in which the letters

are involved symmetrically.
H. H. A.

li

>
210
Example
1.

x

HIGHER ALGEBRA.
If a,
b, c

denote positive quantities, prove that

a 2 + b 2 +c 2 >bc + ca + ab;

and
For

2 (a 3 + b 3 + c 3 )>bc (b

+ c) + ca(c + a) + ab (a + b).
(1);

& 2 + c 2 >2bc

whence by addition
It

+ a 2 >2c«; a 2 + b 2 >2al); a 2 + b 2 + c 2 > be + ca + a&.
c2
is

may

be noticed that this result
(1)

true for any real values of a,

b, c.

Again, from

b2

-bc + c 2 >bc
b3

(2);
(3).

.-.

+ c 3 >bc(b+c)
we obtain

By

writing

down
2 (a 3

the two similar inequalities and adding,

+ b 3 + c3
if

)

be (b

+ c) + ca

[c

+ a) + ab{a+b).
(2)

It

should be observed that

(3) is

obtained from

by introducing the

factor b + c, longer hold.

and that

this factor be negative the inequality (3) will

no

Example
.r

2.

If

x

may have any
x3

real value find

which

is

the greater,

3

+l

or x 2 + x.

+l- (x 2 + x) =x 3 - 2 - (x - 1) = (x 2 -l)(x-l) = (.r-l) 2 (* + l).
hence

Now

[x

-

l) 2 is positive,

x3 + 1
according as x
If

>

or

<

x2 + x
is,

+1

is

positive or negative; that

according as x

>

or

<

-

1.

x—

-

1,

the inequality becomes an equality.

252. Let a and b be two positive quantities, and their product ; then from the identity

$

their

sum

P

4a6 = (a + bf - (a - b)\

we have

iP = S - (a - b)
2

2
,

and S 2 =
is

±P+(a- b)

2
.

Hence,
given,

if

S

is

given,

P

greatest

when a — b\ and

if

P

is

S is

least

when

a= b;
if the sum of two positive quantities is given, their product is greatest when they are equal ; and if the product of two positive quantities is given, their sum is least when they are equal.

that

is,

:

INEQUALITIES.
253.

211

To find

the greatest value

of a product the

sum of whose

factors is constant.

Let there be n factors a, b, sum is constant and equal to s.
Consider the product abc any two unequal factors. If
ii>
...

c,

...

k,

and suppose that their

a, b

Tii two equal factors —— by the

and suppose that a and b are we replace the two unequal factors a+b a+b ..
k,
,

-—

the product

,

,

,

is

increased

while the sum remains unaltered ; hence so long as the product contains two unequal factors it can be increased tvithout altering the sum of the factors ; therefore the product is greatest when all In this case the value of each of the n the factors are equal. s /s\" factors is - and the greatest value of the product is ( - ) or

n

,

\nj \n/

, '

/a + b + a
\

c+ c + n

...

+k\ +k\"
)

Cor.

If «,

b, c,

...

k are uneqiud,
c

/a + b +
(

\

+ n

...

+k\ n
)

> abc

7

...

k

7

;

J

that

is,

a + b + c+

...

+k

n

>

(abc

...

k)".

By an

and geometric mean
the arithmetic

extension of the meaning of the terms arithmetic this result is usually quoted as follows

mean

mean of any number of positive

quantities is greater

than the geometric mean.
Example.
where
r is

Shew

that (l r + 2 r + S r +

.

.

.

+ nr n > nn ( \nY
)

;

any

real quantity.
lr

c Since
.

+ 2 r+Zr +
n

+nr

1

>(l r .2 r .3 r

« r )'

1
;

.'.

(

)

>l r .2 r .3 r

n r that
,

is,

>(|»)

r
;

whence wo obtain the

result required.

14—2

212

HIGHER ALGEBRA.
m
.

254. To find the greatest value q/'a b"cp is constant; m, n, p, ... being positive integers.

. .

when a + b +

c

+

...

Since m, n, p,... are constants, the expression am b n cp ... will

— (-) be & greatest when \mj \nj expression is the product of
(
)

(

m + n +p +
or a

\pj
)

...

is

greatest. &
...

But

this last
is

factors
. .

whose sum

m — )+ n (—) + *p \nj
(

\mj

(

—) +

.. .,

+

b

+

c

+

.,

and therefore con-

stant.

Hence am b n c p ...

\pj
will

be greatest when the factors
o
c

a
ni

n

p

are

all equal,

that

is,

when
b
c

a

m
m
of

n

p
is

+ m+n+p +
b
c

a+

+

Thus the greatest value
m

np

n

p
. . .

/a +
(

b

+ c+

...\
)

M4*4*+"

Example. Find the greatest value of x numerically less than a.

(a

+ x) s

(a -a:) 4 for

any

real value

The given expression
the

is

greatest

when

(

——
is

(
J

—j—

j

is

greatest

;

but

sum

of the factors of this expression
(a - x)* is greatest

3

(

—^- J +
,

4

I

—^—
.

)

,

or 2a;

hence {a + x) 3

when —^— =
. *

—^—

or

x=

--

Thus the

63

84

greatest value is

a7

.

The determination of maximum and minimum values 255. may often be more simply effected by the solution of a quadratic equation than by the foregoing methods. Instances of this have already occurred in Chap. ix. ; we add a further
illustration.

Example.
is

Divide an odd integer into two integral parts whose product

a

maximum.
;

Denote the integer by 2/i + 1 ; the two parts by x and 2n + 1 - x the product by y then (2n + 1) x - x* = y ; whence
;

and

2x = (2n +

1)

±

V^h + I) -^
2

;

but the quantity under the radical must be positive, and therefore y cannot
be greater than
-

11
(2/t

;

.

INEQUALITIES.

213

+ l) 2
in

,

or

n'

2

+n+-

;

and since y
1,

is

integral its greatest

value must be n-

+ n\

which case x = n+

or n

;

thus the two parts are n

and n+1.
256.

Sometimes we may use the following method.
Find the minimum value
;

Example.
Put
..

of

'


'

c

+x

'

c+x=y
.

then

the expression

=

(a-c + y){b-c + y) ^-^ y
c) (b

_ (a -

- c)

+y+a-c+b-c
~ C)

a~

C

b

(

Jy

-^yy + a-c + b-c + 2j (a-c)(b-c)
is

.

is

Hence the expression is a minimum when the square term when y=J(a -c)(b- c).
Thus the minimum value
is

zero

;

that

a-c + b-c + 2
and the corresponding value of x

*J(a -

c) (b

-

c)

is */(«

-

c) {b

-

c)

-c.

EXAMPLES. XIX.
1.

a.

Prove that (ab + xy) (ax + by) > 4abxy.

2.
3.

Prove that

(b

+ c) (c + a) (a + b) > 8abc.
sum
2.

Shew
is

that the

of

any

real

positive quantity

and

its

reciprocal
4. 5.

never less than
l,

If

a 2 + b2 =

and x2 +y 2 = l, shew that ax + by<\. and x2 +y 2 + z2 = l, shew that

If « 2 + 6 2 + c 2 =l,

ax + by + cz <
6.

1.

If

a>

b,

shew that a a b b > a b b a and loe - <
,

losr =

.

7.

8.
9.

^ + y 2 z + z2x) (xy 2 +yz 2 + zx2 ) > Find which is the greater 3«6 2 or aP+263

Shew that

(.r

2

D.'-'-V - 2.

.

Prove that a3 6 + ab3 < « 4 +

64

.

10.

Prove that 6abc < be (b + c) + ca(c + a) + ab (a + b).

11.

Shew that

b 2 c°-

+ c2 a2 + a 2 b 2 > abc (a + b + c).

214
12.
13.
14.

HIGHER ALGEBRA.
Which
Shew
is

the greater x3 or x2 +x + 2 for positive values of x%

that x3 + lSa 2x

>

hax* + 9a3

,

if

x>

a.

Find the greatest value of x greater than x^ + Hx.
15.

in order that

7x2 + 11

may

be

Find the minimum value of x2 - 12#+40, and the value of 24? - 8 - 9x2
.

maximum

16. 17.

Shew that ( \nf > »• and
Shew that
(x +y + ,s) 3
1
.

2

.

4

.

6.

. .

2?i< (w +

l) n .

18.
19.

Shew that n* >
If
?i

3

.

> 27^^. 5 .(2n - 1
. .

).

be a positive integer greater than
2 ft

2,

shew that

>l+?iV2 ,7_1

.

21.

Shew
(1) (2)

that

(x+y +z) 3 > 27 (y + z- x) (z + x - y) (x+y - z). xyz>(y+z-x)(z + x-y)(x+y-z).

22.
7

Find the

maximum value of
• •

(7

- x) A

(2

+ #) 5 when # lies

between

and - 2.
no
23.

Find xu minimum value the
T7v
1

f (5 of

x)(2 + =-*-* + x) 1+*

.

*257.

am +bm

>

/a
(

To prove m —+ b\

that if a "

and b are

positive

and

unequal,

)

,

except

when

m

.

xs

a positive proper jr action.

We
since
L

have
is

m a" + 6 =
1

f

-y- +

+

(

-^-J

-^

g-J
of

;

and
these

—~—

less

than


-

,

we may expand each

expressions in ascending powers of

——
2

.

[Art. 184.]

•'•"~2~ =

a" + 6

m

m /a + b\

+

m (m - 1) (a + b\"—- fa - b\*
1.2

v~2-J

\~r)
"A

\~r)
4

m (m - \)(m - 2)(m-3) fa + bV"+
1.2.3.4
2

J

4 fa - b\ + '" j

12


INEQUALITIES.
(1)

215

all

a positive integer, or any negative quantity, the terms on the right are positive, and therefore
If
is

m

a" +
=

n

b'

>

b fa + 6 V"
s

\2~)
less

'

(2)

If

m

is

positive
first

and
m

than
m

1,

all

the

terms on

the right after the

are negative, and therefore

a + --

——
b
b \

m

<

b\ fa + fr

(3)

If on

>

1

and

positive,
i

put
i

m=76
i
7l

where n <

1

;

then

fa +

m

m m

{-2
'a

)
1

=
1

(-2-)
1

H fd" fa + b n \ n
;

n

m

— o—

+ bm \ m

(a*)

m

+
o

(b»)

H
,
i

>
i
>

]j

/ox

y

(

2)

+

b'"\"

1

>

a+b

—~

.*.

— — > fi-7
=
is

Hence the

proposition

established.

If

m = 0,

or

1,

the

inequality becomes an equality.

*25&.

If there are n 'positive quantities a, b, c, ...k, then a m +b m + c m + ... + k m /a + b + c+...+k > n n \

)

unless

m&

rt

positive proper fraction.

Suppose

on to

have any value not lying between
1

and

1.

m m Consider the expression a + b + c" + ... + k"\ and suppose that a and b are unequal ; if we replace a and b by the two equal
....

quantities
altered,

a+
_
i

——
b

,

a+ —
-j

b
,

the value or a +
1

..

,

,

+ c+...+fc remains un7

m but the value of a" + b m + c +

...

+

k'" is

diminished, since

216

HIGHER ALGEBRA.
so long as

any two of the quantities a, b, c,...&are unequal m m m the expression am + b + c + ... + k can be diminished without altering the value of a + b + c + ...+k; and therefore the value m n of a" + b' + c + + km will be least when all the quantities a, b, c,...k are equal. In this case each of the quantities is equal
1

Hence

.

. .

to

a+

b

+

c

+

...

+k
;

n
of

and the value

m a + b m + cm +

...

+km
c

then becomes
.

n
Hence when

fa +
{

b

+

+

.

.

+ k\ m

»
are unequal,

)

a, b, c,...k

am + b m + cn +...+le m

c + ... + k\ /a + b + c+ ...+fc m > fa

n
If ?n lies

between

and

1

we may

in a similar

manner prove

that the sign of inequality in the above result must be reversed.

The proposition may be stated verbally

as follows

:

is

The arithmetic mean of the m th powers of n positive quantities greater than the m th power of their arithmetic mean in all cases

except

when

m

lies

between

and

1.

*259.

If a and h

'

;

INEQUALITIES.
X

217

*2G0.
if'x

To prove

that

/I + 'I±^-

i-x'

>

v

/ lh V

i-y

5

and y
For

are proper fractions
x

and ])ositive, and x >

y.

/l+x .71±f >flr<
1
.

//g*

|

,.

1

according as

+a

-lo<?-n

l-x

1 1 +y > or < - log y °l-2/'

But

S^lzS" 2 ^ +? + ?+•••)»

t

AAM ^

and

Ilog^2(l + !\ & 3 1-2/ 5 \
2/

7
,

'

- log x *

1

.

1 4- a;

l-x

>y

1

.

log

-

\+y
-

°

1

y

and thus the proposition

is

proved.

•261.

To prove

that (1
,

+

x)

,+x

(1

-x) 1_x >l,

if

x<l, and

to

d

777
i'
<

luce that

/a + b\ a+b k aa bD > ( _
J

Denote

(1 +jb)

1+ *

(1

-a)

1- *

by P; then
(l

logP = (l+a)log(l + x) +
= x {log (1 +x)=

-aj)log(l-a;)

log (1

- x)) +
\

log

(1

+

x)

+ log (l—x)
\

2x^
r\
I
--.

**s

+
3

JO

+
JC

_ / *C

i//

^

+

...)-2^
Jit

+

1

+

SI/

-

+

6

...)

/

•*/

\

Hence
that
is,

log

P is

positive,
(1

and therefore
1

P> 1

+*)

+

r

(l

-*)'-'>!.

1

218
In this result

HIGHER ALGEBRA.
f9

put x = — where u > z u
,

;

then

1

+

sY + w/,
u)

Z

Z

{
'u

^-i) %1
J
-z

z\ ~u
l

;

+ z\

\
.'.

u

J
(u

y-^Y- >r>rl u
\

.

+

z)

u+

*(u-z) u

>u 2u

.

Now

put u + z =

a,

u — z = b,

so that

w = ———

;

(TJ
* EXAMPLES.
1.

XIX.

b.

Shew Shew

that 27 (a 4 + & 4 + c4 ) that

>
3

(a + b + c)\

2.

n (n + l) 3 < 8

(l

+ 2 3 + 3 3 + ... + n 3 ).
of the first

3.

bers

is
4.

Shew that the sum of the m ih powers greater than n (n + l) m if m > 1.
,

n even numthat

If a

and

/3

are positive quantities, and a

> /3, shew

(

-SM-jrthe value of
(

Hence shew that
2-718...
5.

if

n>

1

1

+-

)

lies

between 2 and

If a,

b, c

are in descending order of magnitude, shew that

/a + c\ a

^.

fb + c\

b

\a-c)
6.

\b-cj
c

Shew

that (

'a a

v v + b + c+...+k\ a + b +
~
~

+

--

+ ,i

'

'

)

< a a b b<*.
+ a n ),
if

.

.>&*.

7.

Prove that
If
ii

- log (1 + am <
)
lib

- log
Ih

(1

m > n.

8.

is

a positive integer and 1 _ #n + J
1

?&

+

<

x < 1, shew _ A.n
n

that

)

:

INEQUALITIES.
9.

219
.

If a,

b,

c are in

H.

P.

and

»> 1,

shew that n n -\-c n > 2b n
if
i
i.s

10.

Find the

maximum

value of

x3 (4a - .r) 5
i

x

is

positive

and

less

than 4a; and the
fraction.

maximum

value of

x*(\—xf when x
< x and

a proper

11.

If

x

is positive,

shew that

log (1 +.r)

"

1+.?;

12.

If
(1

x + y + z=l, shew that the

least value of -

-\

x

y

hz

is

i)

and that
13. 14.

- x) (1 - y)
that

(1

- z) > 8xyz.
(a 2 + 6* + c2 + cl2 f.

Shew
Shew

(a+b+c+d) (a 3 + 6 3 + c 3 + a73 ) >

that the expressions
(c

a(a-b)(a-c) + b (b- c) (b-a) + c (c-a)
and
cfi{a-b)(a-c)
arc both positive.
15.

- b)

+ b 2 (b-c){b-a) + c2 (c-a){c-b)
<
n

Shew

that (xm + y m ) n

(.t'

+y n
b
.

MI
)
,

if

m > n.

16.

Shew that abfr <
If
(1
at,

fa 4-h\

a

+

(^p)

17.

6, c

denote the sides of a triangle, shew that
(q

a 2 (p-q)(p- r) + b 2 (q - r)
q, r

-p) + c2

(r

-p) (r - q)

cannot be negative; p,
(2)

being any real quantities;

ah/z + b 2zx + c2 xy cannot be positive, if
[1

x + y + z = 0.

18.

Shew that

j3 15

\2n-l >

(\n)n

are positive integers, If a,b,c, d, to n, shew that the least value of
19.

p

whose sum
r

is

equal

\a\bJ±\ c

L

where q

is

the quotient and r

(|g+ 1 ) > the remainder when n is divided by £>.

»

P "r (|g)

;

CHAPTER XX.
LIMITING VALUES

AND VANISHING FRACTIONS.
a

262.

If a be a constant

finite quantity,

the fraction

--

can
that

x

be made as small as
is,

we

please

by

sufficiently increasing

x

;

we can make

a - approximate to zero as nearly as
;

we

please

by taking x large enough
"

this is usually abbreviated

by saying,

when x

is infinite

the limit of

x

is

zero."

Again, the fraction - increases as x decreases, and by making

x

x as small as we please we can make
thus

as large as

x

we

please

when x

is

zero - has no finite limit;
JO

this is usually ex-

pressed by saying, "
263.

when x

is

zero the limit of -

is infinite."

say that a quantity increases without limit or is infinite, we mean that we can suppose the quantity to become greater than any quantity we can name.

When we

when we say that a quantity decreases without limit, we mean that we can suppose the quantity to become smaller than any quantity we can name.
Similarly

used to denote the value of any quantity is used to which is indefinitely increased, and the symbol denote the value of any quantity which is indefinitely dimi-

The symbol

go

is

nished.

LIMITING VALUES.
204.

221

The two statements
:

of Art. 2G2

may now

be written

symbolically as follows
if

x

is co

,

then -

is

;

x
is

if

x

is

,

then

co

.

x

But in making use of such concise modes of expression, it must be remembered that they are only convenient abbreviations
of fuller verbal statements.

The student will have had no difficulty in understanding the use of the word limit, wherever we have already employed it; but as a clear conception of the ideas conveyed by the words limit and limiting value is necessary in the higher branches of Mathematics we proceed to explain more precisely their use and
26~>.

meaning.
If y =f(x), and if when x approaches a Definition. value a, the function f(x) can be made to differ by as little as we please from a fixed quantity b, then b is called the limit of

266.

y when x — a.
For instance,
if

S
;

denote the sum of n terms of the series
'

1+2 +

2

2

+

2~ J

+

'"

then

S'

=

2

~2^^—
is,
,

t

Art 56 *1

Here S
as

is

a function of n, and

can be made as small

we

please

by increasing n

;

that

the limit of

S

is

2

when

n

is infinite.

shall often have occasion to deal with expressions 267. consisting of a series of terms arranged according to powers of some common letter, such as

We

a + a x + a ax" + a 3 x3 +
x

where the
unlimited.

coefficients
x,

o

,

a,,

a2

,

a3

,

...

are

finite

independent of

and the number

of terms

may

quantities be limited or

It will therefore be convenient to discuss some propositions connected with the limiting values of such expressions under

certain conditions.

,

222
268.

HIGHER ALGEBRA.
The
limit of the series

a + a x + a 2 x 2 + a 3x 3 +
x

when x

is indefinitely

diminished

is

a

.

Suppose that the

series consists of

an

infinite

number

of terms.
...
;

let

be the greatest of the coefficients alf asi o3 , us denote the given series by a + S ; then

Let

b

and

S<bx + bx + bx
2

3

+

...

;

and

if

x<

1

,

we have
is
;

S< — 1 x
-=

bx

.

Thus when x
small as

we

please

can be made as hence the limit of the given series is a
indefinitely diminished,
.

S

If the series consists of a finite

number

of terms,

S

is less

than in the case we have considered, hence a fortiori the proposition
269.
is true.

In

the series

a

+ a,x + a2 x + a 3 x,3 +

.

.

.

by taking x small enough ive may make any term as large as we please compared with the sum of all that follow it ; and by taking x large enough we may make any term as large as we please compared with the sum of all that precede it.

The
it is

ratio

of

the term an x

n

to the

sum
a

of all that follow

ax an+1 x
n+l

n

+a n+2x"
;

+2
'

+

or
'

...

a n+1 x + a u+2 x*+..

When

x

is

indefinitely small the

as small as we please as we please.

that

is,

denominator can be made the fraction can be made as large
n

Again, the ratio of the term a nx precede it is

to the

sum

of all that

a xn
a n—l ,cc
n
l

+an — »x
2

n

2

+...'

or

a a n — ,y + a n—2<J 2 +...' ay
lts

where u = y x

.


LIMITING VALUES.

.

223
;

indefinitely largo, y is indefinitely small hence, fraction can be made as large as as in the previous case, the

When

x

is

we

please.

270. position

The following particular form
is

of the

foregoing pro-

very useful.

In the expression
a x + a H — ,x
ii

H-i

+

+ a.x + a
1

1

, »

consisting of a finite number of terms in descending powers of x, by taking x small enough the last term a can be made as large as we please compared with the sum of all the terms that precede it, and by taking x large enough the first term ax* can be made as large as we please compared with the sum of all that follow it.

Example 1. By taking n large enough we can make the first term of n 4 - 5/i3 -7/i + 9 as large as we please compared with the sum of all the other terms that is, we may take the first term ?i 4 as the equivalent of the whole expression, with an error as small as we please provided n be taken large
;

enough.
3.t
3

Example
zero.
(1)

2.

Find the limit of =-=

2x'2
:

—4

— when
may

(1)

x

is infinite

:

(2)

x

is

In the numerator and denominator we
3a; 3
;

disregard

all

terms but

the

first

hence the limit

is

^-s
OXr

,

3 or ^
O

.

(2)

When

x

is

indefinitely small the limit

is —--

-4
,

8

or - 2

1

Example
Let have

3.

Find the limit of

^

/1 +x - when x V 1—x
*

/

is

zero.

P

denote the value of the given expression

;

by taking logarithms we

log

P=i

X

{log (1+ x)

-log (1-x)}
[Art. 226.]

^(l + ^' + '^+.-.V
Hence the
required
is
2 e' .

limit of log

P

is 2,

and therefore the value of the

limit

.

224

HIGHER ALGEBRA.

VANISHING FRACTIONS.
271.

Suppose

it is

required to find the limit of
2 x 2 + ax — 2a

x 2 - a2

when x = a.
If

we put x = a + h, then h
a. x,
2

will

approach the value zero as x

approaches the value

Substituting a + h for

x +

2

ax— 2a
2

x

-a

2

=

3ah + h 2
2ah + h
2

=

3a + h

2a~+h'
this

and when h
.

is

indefinitely small the limit of

expression

is

-.
a

3

There

is

however another way
x 2 + ax - 2a2 x 2 — a2
(x

of regarding the question; for

- a) (x + 2a) (x — a)(x + a)
the

x + 2a x+a
of

'

and
^r
-j
,

if

we now put x = a

value

the

expression

is

as before.
2
.

fi

2

If in the given expression

x^ — a
it

~— we

pat x = a before
,

simplification

it

will be

found that

assumes the form -

the

value of which is indeterminate ; also Ave see that it has this form in consequence of the factor x -a appearing in both numerator and denominator. Now we cannot divide by a zero factor, but as long as x is not absolutely equal to a the factor x - a may be removed, and we then find that the nearer x approaches to the value «, the nearer does the value of the
3 fraction approximate to ^
,

or in accordance with the definition of

Art. 266,

when x = a,
i

the limit of

,

!

i

.

.

,

n

x t ax —
^

.

x~ -^a"


Jia

.

is

2

o

^
VANISHING FRACTIONS.
272.
If

.

.

225

f(x) and

<f>

(x) are

becomes equal to zero for
fraction

two functions of x, each of which some particular value a of x, the
and
is

^~
<f)

(a)

takes the form Ki v

called

a

Vanishing

Fraction.

Example

1.

If

x = S, find the limit of T 3 -5:r 2 + 73-3 3 x- — ox— '6 .T

When x = 3,

the expression reduces to the indeterminate form ^; but by

removing the factor ~ 2x+1 becomes 2 x + 2x + 1
,

x-3 from numerator and When x = S this reduces to

denominator, the fraction ,

which

is

therefore the

4

required limit.

Example

2.

The

fraction

J'^-a-Jx + a
x-a

becomeg

when % _ a

To

find its limit, multiply

jugate to

numerator and denominator by the surd conJ'dx-a- Jx + a; the fraction then becomes

(Sx-a)-(x + a)
^

2
or
,

[x- a )(Jdx-a + Jx + a)' whence by putting x = a we

J'6x-a+

>Jx +

a

find that the limit is

—j=

1

Example

3.

The

fraction

becomes ^ when x=l. 1 _% x
the Binomial Theorem.

-

21 x

To find its limit, put x = l + h and expand by Thus the fraction
1

-

(1

+

fe)*

_

V

3

9

l-(l + /0i

l-(l+J*-^»F+-.)
1

3
1

+

1, -;<

9 2
7

-5 +

/l

5

25
is

Now h = when
273.

*ael; hence the required limit

-

Sometimes the roots of an equation assume an indeterminate form in consequence of some relation subsisting between the coefficients of the equation. lo H. H. A.

226
For example,
if

HIGHER ALGEBRA.
ax + b = ex +
(a
d,

— c)x = d — b,
x=

d-b
a—
j-

c

But

if

c

= a, then x becomes
is

——

,

or go
if

;

that

is,

the root of
is

a simple equation
indefinitely small.

indefinitely great

the coefficient of x

274.

The

solution of the equations

ax + by + c =
be'

0,

a'x + b'y
ca'
1

+c =
'

0,

ab'

- b'e — ab

— c'a

ab'-a'b
infinite.
a',

If ab'

— a'b =

0,

then x and y are both
;

In this case
the

— =—= m

suppose

by substituting
c — = 0.

for

b\

second

equation becomes ax + by +
c—

If

m
ii

is
c'

not equal to

c,

the two equations ax + by +

c

=

and

ax + b J

H

=

differ

Ml

only in their absolute terms, and being J

inconsistent cannot be satisfied

by any

finite values of

x and

y.

If

m is

equal to ^

c.

we have

-=-=-,
a
b
c

and the two equations

are

now

identical.

Here, since be —

b'e

=
,

and

ca'

— c'a —

the values of x and y
is

Q each assume the form fact,

and the solution

indeterminate.

In

the present case we have really only one equation involving two unknowns, and such an equation may be satisfied by an unlimited number of values. [Art. 138.]
in

acquainted with Analytical Geometry will have no difficulty in interpreting these results in connection with the geometry of the straight line.
is

The reader who

:

;

VANISHING FHACTIONS.
shall now discuss some peculiarities 275. arise in the solution of a quadratic equation.

227
which may

We

Let the equation be

ax 2 + bx +
If c

c

-

0.

= 0, then
2 ax + bx = 0;

whence
that
is,

x=
one of the roots
is

0,

or

:

a
is finite.

zero and the other

If 6
sign.

=

0,

the roots are equal in magnitude and opposite in

[Art. 118.]

If

a=
in

that

the equation reduces to bx + c = this case the quadratic furnishes
0,

;

and
only

it

appears one root,

namely —

=- .

But every quadratic equation has two

roots,

and in

order to discuss the value of the other root

we

proceed as follows.
of fractions

Write —
*J

for

x in the original equation and clear
cy + by + a
2

thus

=

0.

Now

put a =

0,

and we have
cy
2

+ by = 0;
or

the solution of which

is

y J

0,

—c>
b
;

that

is,
'

x=

oo,

or

c

T
b

.

Hence, in any quadratic equation one root will become
if the coefficient

infinite

ofx

2

becomes zero.

This is the form in which the result will be most frequently met with in other branches of higher Mathematics, but the student should notice that it is merely a convenient abbreviation of the following fuller statement

In the equation ax 2 + bx + c = 0, if a is very small one root is very large, and as a is indefinitely diminished this root becomes indefinitely great. In this case the finite root approximates
to

-y
o

as its limit.

The

may

cases in which more than one of the coefficients vanish be discussed in a similar manner.

15—2

228

HIGHER ALGEBRA.

EXAMPLES. XX.
Find the limits of the following expressions,
(1)
-

when #=oo,
-3) 7#2
-

(2)

when x = 0.
(3r?I)*

(2s

(3

-5*)

)

CHAPTER

XXI.
SERIES.

NVERGENCY AND DIVERGENCY OF
076

terms are formed expression in which the successive series if the series terminate at w ^ ! lr by some reguJ« law is called a number q{ lb ca'iea a lt:

Ax

some

assigiMgl
is

t^

^ m
x

^e

;

terms

unfimitfd, it

is

called

an

B infinite series.

^J.

^

In the preset chapter we an expression c- ° ue form u + n2 + 3 +
i/-

shall usually denote a series

by

+u +

/

consisting of w terms. Suppose that we have a series function of n; if n increases The sum of the series will be a equal to a certain

tends to become indetinitely, the sum either becomes infinitely great. finite fcmi*, or else it

of

numerically the first n terms cannot quantity however great n may be.

An

infinite series

is

said to be convergent

when

the

sum
finite

exceed some

be divergent when the sum of numerically greater than any finite the first n terms can be made great. quantity by taking n sufficiently

An

infinite series is said to

978
series

W examining whether the made
finite,

of a given we can find the sum of the first n terms divergent or we may ascertain whether it is convergent becomes inIf series

remains

finite,

or

when n

is

indefinitely great.
of the first
...

For example, the sum
1

n terms of 1-*"
.

the series

2 + x + x + x* +

is

.

_a

.

T
230
If

HIGHER ALGEBRA.
x
is

numerically

less

than

1,

the

sum

appro.

;

-,

the

finite limit

j—-

,

and the

series is therefore

converge,^!
sum
of
t„ first

If

x

is is

n terms

—y
1,

numerically greater than
,

1,

the

and by taking n

sufficiently great,

d e greater cUve?ge n t. d

""*

™Y

tk can
*>•»
is

&n * e

l""*^

thus

serie's

r i^entUm * ^ **
x= the series becomes

"

^

«»

*

*" d

***»

^

If

1-1+1-1+1 -1+
The sum of an even number of terms is of an odd number of terms is 1 and thiw +\L between the values and 1. Thi t,fe Ac which may be called o-iBo^ or
I

X A^c^
«*•

^™ ^
5

,;L the sum •„

S:
280.

are many 0ases in which we'haic , sum of the first n terms of a series. therefore to investigate rules by which we
f of

/^' ?ihere finding the.

T
,

^
,„.

1

°f a

^

We

p"
h od
its

can test the cTt
effecting

-*Hout

4n

tra/mfe series

w* taAicA *Ae ferw are alternately

Let the

series

be denoted by
u,

M -%+%where
6 giVe

+ u -M +
5

w1 >^>^a >w,> M o * 4
"
SerfeS

....

may

for,™

be Written in each of the following

K-«,)+(«,-«0 +(».-«,) +

^
(

».-K-«J-(«4 -«,)-K-«r)From

2 ).

(1) we see that the sum of any number of terms is a positive quantity; and from (2) that the sum of any nnmber of terms is less than «, ; hence the series is convergent.

CONVEltGENCY AND DIVERGENCE OF SERIES.
281.

231

For example, the
,

1

11111
f1

series

h

2
is

3

4
-•

5
1

6
in Art. 223,

convergent.
is

By

putting x

we

see that its

sum

log e

2.

Again, in the series

T~2
each term
series is
is

23 +
i

4

_5 3"4 +

6

5

_7 + ~6

'

numerically less than the preceding term, and the therefore convergent. But the given series is the sum of
1

11111 -2 3-4 5"6
+
+

+
+

m
'
,

(1) '
(2).

and

1-1+1-1
Now
(1) is

+

1-1

equal to log e
of terms is

2,

and

(2) is equal to

or 1 according'
series

as the
is

number

even or odd.

Hence the given

convergent, and its sum continually approximates towards log,, 2 if an even number of terms is taken, and towards 1 + log 8 2 if an odd number is taken.
282. An infinite seizes in which all the terms are of the same sign is divergent \f each term is greater than some finite quantity

however small.
each term is greater than some finite quantity a, the sum of the first n terms is greater than na and this, by taking n sufficiently great, can be made to exceed any finite

For

if

;

quantity.

Before proceeding to investigate further tests of convergency and divergency, we shall lay down two important principles, which may almost be regarded as axioms.
283.

remain convergent, and if divergent it will remain divergent, when we add or remove any finite number of its terms for the sum of these terms is
I.

If a series is convergent it will

;

a finite quantity. the terms are positive is convergent, then the series is convergent when some or all of the terms are negative ; for the sum is clearly greatest when all the terms have the same sign.
II.

If a series in

which

all

We

shall suppose that all the
is

terms are positive, unless the

contrary

stated.

;

232
284.

HIGHER ALGEBRA.
infinite series is convergent if from and after some the ratio of each term to the preceding term is numerically

An

fixed term less than some quantity zuhich

is itself

numerically

less

than unity.

Let the

series

beginning from the fixed term be denoted by
u,

+u 12 +
c>

u^ 3

+ u^ + 4
r

and

U
let

—<
U
:

r,

U1 -* — < r, UA < Us U2
+

where

r

<

1.

Then

u,

+u 12 + u+u
r,

3

A 4

/_.
1

u9

u.
<f a

V
(1

^
+
1.

ua u
x

n u u u 3 u 2 Ul
)

1

<
tliat
is,

it,

r

+ r2 + r 3 +

<
1

u. ~1

-r

,

since r

<

Hence the given
285.

series is convergent.

In the enunciation

of the preceding article the student

should notice the significance of the words " from and after a fixed term."

Consider the
1


CONVERGENCY AND DIVERGENCY OF
SERIES.
'2Xi

infinite series in which all the terms are of the same 286. sign is diverge) it if from and after some fixed term the ratio of ea<-li

An

term

to the 'preceding

term

is

greater than unify, or equal to unify.
.

Let the fixed term be denoted by t* If the ratio is equal to of the succeeding terms is equal to u unity, each and the sum is equal to nu of n terms hence the series is divergent.
, l

;

than unity, each of the terms after the and the sum of n terms is greater fixed term is greater than u than nu hence the series is divergent.
If the ratio is greater
x

,

;

}

In the practical application of these tests, to avoid having to ascertain the particular term after which each term is greater or less than the preceding term, it is convenient to find
287.

the limit of

-

when n

is

indefinitely increased; let this limit

be denoted by
If If If

n—\ A.
is

X< \>

1,
1,

the series

convergent.
divergent.

[Art. 284.]
[Art. 286.]

the series

is

the series may be either convergent or divergent, and a further test will be required ; for it may happen that

X=l,

—— <
n—
1

1

but continually approaching to
m

1

as

its

limit ivhen

n

is

indefinitely increased.

quantity r which

is

In this case we cannot name any finite itself less than 1 and yet greater than X.

Hence the

test of Art.

284
1

fails.

If,

however,

uH —
is

u —— >
I

1

but con-

tinually approaching to Art. 286.

as its limit, the series

divergent by

We

shall

use " Liin

un —

—— "
1
,

as

an abbreviation

of the

words

"the limit
Example 1

of

U — when n un —
.

is infinite."

1

1.

Find whether the

series

whose n lh term

is

-

.,

is

con-

di-

vergent or divergent.

?/

n

(n

+ l)a: n
n
2

ru^1
{h

(n
2

+ l)(n-l)n*

('„_!

-

1)

him


"n
I

—x\

;

; ;

234
hence
if
if

HIGHER ALGEBRA.
x < 1 the x > 1 the
series is convergent series is divergent.

If

x = l, then Lim

u — =1, and

a further test

is

required.

Example

2.

Is the series
l2

+ 2 2x + 3 2 x 2 + 4?xs +
u —n
7jL

convergent or divergent?

_ Here
Hence
x = 1 the
3.
if
if

T Lim
.

un-i

=Lim-.

~ n2 x n l —. n (n-l)-x „
is

9 2

=x.

x < 1 the series

convergent
divergent.
. . .

x> 1

the series
l2

is

If

series

becomes

+ 2 2 + 32 + 4 2 +

,

and

is

obviously divergent.

Example

In the series
.

~ a+(a + d)r+{a + 2d)r 2 +... + (a + n-1 d)r n 1 +

...,

, Lim
.

»*-i

a d w —n —=Lim — (n -^--.r = a (n-2)d
T
.

-t-

1)

—.

+

r;

thus

if

r< 1

the series

is

convergent, and the

sum

is finite.

[See Art. 60, Cor.]

two infinite series in each of which all the terms are ])Ositive, and if the ratio of the corresponding terms in the two series is always finite, the two series are both convergent,
288.

If

there are

or both divergent.

Let the two

infinite series
x

be denoted by
,

u + ua + ua + w4 +
and
v,

+ 12 +
v,

v, 3

+

v.

4

+

The value

of the fraction

u + u , + ua^
i

+n

n

lies

between the greatest and

least of the fractions

\
and
is

-*,

-»,

[Art. 14.1

therefore

a,

finite quantity,

L

say

Hence

if

one

series is finite in value, so is the other; if

one
the

series is infinite

in value, so

is

the other;

which proves

proposition.

CONVERGENCY AND DIVERGENCY OF
289.

SERIES.

235

The application

by means of it we series whose convergency or divergency has been already established. The series discussed in the next article will frequently be found useful as an auxiliary series.
290.

of this principle is very important, for can compare a given series with an auxiliary

The

infinite series

1111 y
T
2P

3P

4.1'

is

always divergent except when p

is positive

and

greater than

1.

Case I. Let;? > 1. The first term is 1 the next two terms together are
;

less

than
fol-

2 j—; Z

4 the following four terms together are less than-—;
.

the

4
less

lowing eight terms together are
the series
that
2
is,

than
8


o

;

and so

on.

Hence

2
is less

than

I

+ t^+th+ttt, P 4'
2 o
1

4

+•••;

less

than a geometrical progression whose
1,

common

ratio

~j

is less

than
II.

since

p>

1

;

hence the series

is

convergent.

Case
The

Let_p=l.

series

now becomes

1

+^ +
2

-^

+ - + =;+
4
5

...

3

The

third and fourth terms together are greater than - or
t

2

1

^ —

;

4 1 the following four terms together are greater thau ^ or the o 2 8 1 following eight terms together are greater than and so or ;

;

on.

Hence the

series is greater

1111
+ + +
2

than
2

+

'"'

2

2

and

is

therefore divergent.

[Art. 2^6.]

Case III. Let p<\, or negative. Each term is now greater than the corresponding term
Case
II.,

in

therefore the series
series is

is

divergent.

Hence the

p

is

always divergent except in the case when positive and greater than unity.


236
Example.

HIGHER ALGEBRA.
Prove that the series
2 3

4

n+1

is

divergent.

Compare the given

series

with 1 + «

+ «+v

"^

•"••••

Thus

if

«*

th terms of the given series and the n and v n denote the n

auxiliary series respectively,

we have
n
2

un
i'

_n + l

.

1
re

_ w+ 1
?i

n

'

7/

hence

Zi-m,

— =1,

and therefore the two

series are

both convergent or both

divergent. But the auxiliary series is divergent, therefore also the given series is divergent.

This completes the solution of Example

1.

Art. 287.

291.
limit of

In the application

of Art.

288

it is

necessary that the

should be finite

;

this will be the case if

we

find our

auxiliary series in the following

way

:

Take u
u
- is finite

,

the

nth term

of the given series

and retain only the
)

highest powers of n.

Denote the

result

by v n

then the limit of
7i

by Art. 270, and v may be taken as the

th

term of

the auxiliary series.
3/2n2
is
,,

Example r
divergent.

1.

Shew that

the series

whose n th term

-

Z/S?vi +

2n+5

=
1

is

As n

increases,

un approximates

to the value

l/w
Hence,
if

or
'

4/3

*

i

n 12

v„=-r ,we ~ n1

1

have

Lim

— = ^r, 3 v
u
/2

3

which

is

a finite quantity;

n

v/

therefore the series

whose nth term

is


1

may
;

be taken as the auxiliary
is

series.

But

this series is divergent [Art. 290]

therefore the given series

divergent.

CONVERGENCY AND DIVERGENCY OF
Example
2.

SERIES.

237

Find whether the
vn =

series in
:}

which

^/;<

+l -n

is

convergent or divergent.

Here

"»=«

\\/ * + tf ~
1

*J
1

//

+

(

»-» + -;"
5

)

~3n 2
If

+

9><

we take v n =

=

,

we have

vM N
Luti

3

9n'J

— =x. v„
n

3

But the auxiliary

series

JL

Jl JL P + 22 + 3 2+

'"
is

n

l " 1+

-

is

convergent, therefore the given series

convergent.

To shew that the expansion of (1 292. T/teorem is convergent when x < 1.
Let u r u r+l represent the pansion ; then
,

+ x) n by

the

Binomial

?* th

and (?-+l) th terms

of the ex-

u wr
.

,

n-r+1
r
is

negative; that is, from this point the terms are alternately positive and negative when x is positive, and always of the same sign when x is negative.
this

When r>?6+l,

ratio

Now when
since

r

is

infinite,

Lim —

7/

— =x

numerically

;

therefore

the series is convergent if all the terms are of the same sign; and therefore a fortiori it is convergent when some of [Art. 283.] the terms are positive and some negative.
1

x <

To shew that the expansio?i of a x in ascending 293. of x is convergent for every value of x.

p owers
1

n-1 «„_, the value of x; hence tlie series

Here

W —

*-

=

#

^ ——
1°#«
-

;

and therefore Lim
1
1

7-

U —

=-

<

11 whatever
1

be

«*__,
is

convergent.

.

238
294.

HIGHER ALGEBRA.

To shew that the expansion of log (1 + x) in ascending powers of x is convergent when x is numerically less than 1.

Here the numerical value of
is

—— = u
ni
,

-i

n't

I

n

x.

which in the limit
is less

equal to x
If
a5

\

hence the series is convergent when x
the series becomes

than
is

1.

= l,

1— k + 77-t+--> 4 3 2
1

an ^

con "

vergent.
If

[Art. 280.]

q a 4: O This shews that the logarithm of zero is divergent. [Art. 290.] infinite and negative, as is otherwise evident from the equation

x~ — 1,

the

series

becomes —

— -—

_

t"

•••>

an(^

*s

e-°°=0.
295.

The

results of the

two following examples are important,
\q<j

and

will be required in the course of the present chapter.
1.

Example

Find the limit

of

x —2-- when x

is infinite.

Put x = ev; then
logs
y
~ eV y*

y
yi

X

i

-

y
\2

y'

2

y
also

^

3

+"
is

when x

is infinite

y

is infinite

;

hence the value of the fraction
is infinite

zero.

Example
Let

2.

Shew

that

when n

the limit of nx n = 0,

when x<l.

x=if

,

so that

y>l;
n\ogy = logz; then
y

also let y n =z, so that

fu^=— = i ^^ = logy' n z'logy
Now when
therefore


and

logz
.

z

n

is infinite z is infinite,

—s_ = 0;
z

also

logy

is finite;

Lim nx n = 0.
It
is

296.

sometimes necessary to determine whether the
is finite

product of an infinite number of factors

or not.

Suppose the product to consist of n factors and to be denoted by io uMAia then if as n increases indefinitely u <<1, the product will ultimately be zero, and if u n > 1 the product will be infinite ; hence in order that the product may be finite, u- must tend to the limit 1
;

2

.

CONVERGENCY AND DIVERGENCY OF
Writing
1

SERIES.

239

+ vn

for

u n the product becomes
,

(l+* )(l+*,)(l+*8)
1

(l+Oj

Denote the product by
1

P and

take logarithms

then
(1),

logP = log(l+v ) + log(l+v 8 ) +...+ log(l + vJ
and
in order that tlie product

may

be

finite this series

must be

convergent.

Choose as an auxiliary

series

v,+v 2 + v3 +
/

+v n

(2).

_1

Now

r Lim-2-l
.

log(l +
vn
is

t;)

^ = Lim\

_.

r\

2** +
vn

.

"

/==1,
I

since the limit of v n

when

the limit of
is

un

is 1

Hence
product

if

(2) is

convergent, (1)

convergent, and the given

finite.

Example.

Shew

that the limit,

when n
7

is infinite,

of

13
2'
is finite.

3

5

5

'4*1*6' 6

2n-l 2n + l ~JT~'~2ir
successive pairs by

Uj,

The product consists of 2n factors; denoting the m 2 Ug,... and the product by P, we have
,

P=u
where
but
**

x

v 2 u3

u n>
,

s n = — 2m

2n-l 2«+l

-5

2n

—=

1

- t-?; 4«(1),

1

logP = logM 1 + log« 2 + logM 3 + ...+logM n
to

and we have

shew that

this series is finite.

Now

log« n = log (l
2,

-^)=-

~^

'•'

32/i-1

therefore as in Ex.
is finite.

Art. 291 the series is convergent,

and the given product

In mathematical investigations infinite series occur so 297. frequently that the necessity of determining their convergency or divergency is very important ; and unless we take care that the series we use are convergent, we may be led to absurd conclusions. [See Art. 183.]

240
For example,
if

HIGHER ALGEBRA.
we expand (l—x)~ 2 by
1

the Binomial Theorem,

we

find
(

1

- a;)- 2 =

3 + 2x + 3ar + 4a +

obtain the sum of n terms of this series as explained in Art. 60, it appears that

But

if

we

in +
1

2.*;

+

3ar

O

9

+

...

+ nx

n-l

=

*
t=

^

(1

- x)

ra

—=

^^ 1 - x

:

whence

(l-x) 2

I

+'2x+ 3x~ -'—'— +

...

+ nx -

4-

7z

(1-a?)9
i

^+

1-*

By making n J °

infinite,

we

see that -z

(l-x) 2

- can only be rea

garded as the true equivalent of the
1

infinite series

+ 2x + 3x2 + ix3 +

when
If

-rz

(1

nx x vanishes. ri + =— l-x -x)~
is

n

infinite,

this quantity

becomes

infinite

when x=l,
295], so

or aj>l,

that

it is

and diminishes indefinitely when a,*<l, [Art. only when x < 1 that we can assert that

\

Ta =* 1

+ 2x + 3x2 + 4#3 +

to

inf.

j

if we were to use - x)~ 2 by the Binomial Theorem as if it were the expansion of (1 In other words, we can introduce the true for all values of x. 2 infinite series 1 + 2x + 3x + ... into our reasoning without error

and we should be

led to erroneous conclusions

if

the series
divergent.

is

convergent, but

we cannot do

so

when the

series

is

divergent series have compelled a distinction For to be made between a series and its algebraical equivalent. 2 example, if we divide 1 by (1 - x) we can always obtain as many terms as we please of the series
difficulties of
,

The

l

2 + 2a;+3£ +4a;3 +

whatever x

may

be,

and

so in a certain sense
;

-p.

^

niay be
equi-

called its algebraical equivalent

yet, as

valence does not really exist except

we have seen, the when the series is

con-

1


SERIES.

CONVERGENCY AND DIVERGENCY OF
vergent.
It
is

241

therefore

more appropriate

to speak of


(l

X)

as the generating function of the series
1

+2a,-

+ 3a 2 +
alge-

being that function which Avhen developed by ordinary braical rules will give the series in question.

term generating function will be more fully explained in the chapter on Recurring Series.
of the

The use

EXAMPLES. XXI.
Find whether the following
.

a.

1

#

Ill
1

series are
1

convergent or divergent.:

x

x+a

x-^-2a

.v

+ 3a
1

_ 4.

x and a

being positive quantities.

+ + + + 1.2 273 371 475
6
-

1

1

1

_1_

1

1

1

xy

(*+i)(y+i) >+a)(y+*)
x2
1

4

(*+3)(y+3)

+

'

x and y
4

being positive quantities.

x

x3
1 1

x*
h
.

1.2^2.3^3.4^4.5
tf
/>»

/)»2
\Mj

+ + + + T72 3T4 576 778
6
n
,
-

\Mj

o»o

%A/

o***

1+

I+ I + I+
2

22

32

42

7

-

\/l + \/i + \/f + \/1 +
1

8.

+ toe + bx + la? + 9af* +
__

2

y
*

"i^

1 i. + 2/' + 3p + A + 4p
+
2
.,

ia

1

+

+
5

Ib

+ - + ,^TT +

11.

x+

3 2 8 , 15 A - x + -x* + x + 5 10 17

.

.

.

+

n 2 -\
-.,—-..

nl +

xn + 16

H. H. A.

242
-.n
-,

HIGHER ALGEBRA.
l

12.

+ g * +5 *» +I ^P + „. +
,

2

6

14

,


2' l

-2*p-i+

1
14.

— — —
.

2.r+— + -- +
8 27

..+

-,—
?t

3

+
4\_4\ 3
=" 3^

1C 15
'

/2 2

2\-!

(p-l)
2

+

/3 3
3 (2-

3\- 2

-2)
4

+«= \3
4

/4 /4_

4

16.

1

+

--

4 2 3 +- + - +-+

3

17.

Test the series whose general terms are
(1)

Jn*+l-n.

(2)

jtF+l- Jnt-^i.

18.

Test the series
/1N
1
.r

1

1

1

/on (2)

A-

-+ # -1 A+l ^ -2 a T5+ r-+——-+ 5+— 2
+

11

A+l

a+2
1

x+3
1
1

x

being a positive fraction.
19.

Shew that the

series

1+
is

2"

I I E

+

3^

+

4"

+

convergent for
20.

all

values of p.
infinite series

Shew that the

ux + 2i 2 + u 3 + u± +
is

convergent or divergent according as Lim^fu n
21.

iti

<1, or >1.

Shew

that the product
2
1
*

2

4

4
'

3' 3' 5

6 5

2ti-2 2tt-2 2n 271-3' 2»-l"S^Ti

is finite

when n

is infinite.

22.
infinite,

Shew

that when x=\, no term in the expansion of (1 +#)" is except when n is negative and numerically greater than unity.

CONVERGENCY AND DIVERGENCY OF
*298.
sriven

SERIES.

243

tests of convergency and divenrencV & cxltv Art* 9x7 ogi 11 «» . proved in the next article enables of

m

The

wp nave we i.™
b

n/^T ^ ST^
+
1

.1 I+ +

1

3?

+

•••

,7

+

---

6

venter

"

ddit£0U,a

**

wUch

^

S01 » eti »' es

fa-d

con-

— ^
<
Case
I.

tergent

when
;

the v-servs is convergent if after

some particular term
«,/t(!re the v-sertes is

onrf

Me „*»*»
> _-n V.,
'

«,<« J,

AWjori

divergent if

5-

U n-i

Let us suppose that Let

Wj

and

»,

are the particular terms.

* < Ei &<!
2

.

then

w,

+

w + u3 +
fl

=

2t

V

w

i

**

2

w,

y

that

is,

<
if

(v. -f

-y

+v +

)

Hence,
vergent.

the ^-series is^convergent the w-series

is

also con-

Case

II.

Let

- > 3»
2

^^
tt,
a

**,

v,
i

«

;

then

2

M + ^9 +
i

U,

+

V

V,

Va

27,

J

16—2

:

244
that
is,

HIGHER ALGEBRA.
>
if

(v

l

+ v2 + v 3 +
is

...).

Hence,
vergent.

the ^-series

divergent the

it-series

is

also di-

*300. have seen in Art. 287 that a series is convergent th or divergent according as the limit of the ratio of the ?i term In the to the 'preceding term is less than 1, or greater than 1. remainder of the chapter we shall find it more convenient to use this test in the equivalent form
:

We

convergent or divergent according as the limit of th the ratio of the n term to the succeeding term is greater than 1,
is

A series

or less than Similarly

1

;

that

is,

according as
of

Lim

—— >

1,

or

<

1.

the

theorem

the

preceding

article

may be

enunciated

The

w-series will be convergent

when the
;

v-series is

convergent

provided that

Lim

u v —— > Lim ——
is

and the

it-series will

be

di-

vergent when the v-series

divergent provided that

Lim
*301.

^^ Lim ^.
is
1,
(

The

series

whose general term

vergent according as

Lim \ n

——
-

un

is

convergent or di-

1 \

>>

or

< 1.

Let us compare the given

series

with the auxiliary series

whose general term v

is

.

the auxiliary series case the given series is convergent if
1

"When p >

is

convergent, and in this

un Un +
thatis,if

(n+iy
,

or (l

+ iy.

n?
l

JS

B

.

u n+i

> l + g + -J CP-lV+ „
>

n

2n~

n
that
is, if

/

u

,\
l

KCr
I

r
(

p+
-

p (p-l)

^
)

+

Lim \n

Wh

1

1

>)>.

J)

.

'

'

CONVERGENCY AND DIVERGENCY OF
But the auxiliary
series is

SERIES.

245

by a Unite quantity however small proposition is established.

convergent if y; is greater than 1 ; hence the first part of the

the auxiliary series is divergent, and by proceeding as before we may prove the second part of the proposition.
1

When p<

Example.

Find whether the
a;

series

l
is

+

1 2* 3

^ + L3
if

x=
5

2.4*

+

2~i.d'T + '"

1.3 .5

x[

convergent or divergent.

Here Lim
the series
If
is

u n+l

1 —— =-;
it

x"

hence

x<l

the series

is

convergent, and

if

x>l

divergent.

x= 1, Lim

u —— = u
n+l

1.

In this case
-

1 M" _

3 4

-

5 6
......

(2w(2n -

3)

1

~2

.

.

2) ' 2~/T=T

and

wn un+1
'•

2n(2n+l)
(2n - 1) (2n

-

1)

"Urn

J"

(2n-l)2

'

hence when

a;

=1

the series

is

convergent.
is
1,

*302.

T/ie series

whose general term

un
or

is

convergent or di-

vergent, according as

Lim ( n

log

—-

j

>

<

1.

Let us compare the given series with the series whose general

term

is


n

-

l

When

p>

1

the auxiliary series
is

is

convergent, and in this

case the given series

convergent

if

u-sn+
I

>

/„
1

lv + ij

;

[Art. 300.]

that

is, if

log

—— > p log (1

h

)

:

or

!

if

log

**
" 71+

'u

.,
1

>P «
'

^-5 + 2

7J

2n

"3

.

:

246
that
is, if

HIGHER ALGEBRA.

Lim In log

——

)

>p.

Hence the first part of the proposition is established. When p < 1 we proceed in a similar manner in this
;

case the

auxiliary series
Example.

is

divergent.
series

Find whether the
2 2z 2

3 s x3

4*r 4

5 5 x5

is

convergent or divergent.

„ Here

wn+l

u —n = n x — *-

n n
|n_
'

'- v

(n + l) n+l /pW+1
.

7^

[n

+l

(n+l)**

H)'
[Art.

A lyV
220 Cor.l.

.-.

Lim

wn+1

3l = -1
<?*

Hence

if

a?<- the

series is convergent, if

#>-

the series

is

divergent.

If:r=-,then
e

^St—
u n+l
-n— = loge-wlog( 1

•.log

+-

)

_1

~2n
.

3n8+ "
1

J_

'

un

1

.

Lim [ n log ——
is

1

=-

hence when x = - the series

divergent.

\ u n+ ) tests given in Arts. 300, 301 are not applicable.

*303.

If

Lim -^- = 1, and wn+1

also

Liminf-^- - l)) =
i

1,

the

J)

To
series

discover a further test

we

shall

make
.

use of the auxiliary

whose general term

is

—n

(log n) p

r-

In order to establish

the convergency or divergency of this series proved in the next article.

we need

the theorem

,

CONVERGENCY AND DIVERGENCY OF

SERIES.

247

*304. If $ (n) is positive for all positive integral values of n and continually diminishes as n increases, and if a be any posit ive integer, then the two infinite series
</>(l)

+

<£(2)

+

</>(3)
</>

+
)

...

+

</>(n)

+

...,
.
.

and

a<£ (a)

3 2 + a <£(a 2 ) + a

(a

3

+

.

. .

+ a n <£ (a n ) +

.

are both convergent, or both divergent.

In the

first series let
k

us consider the terms

</>(«*+ 1), <f>(a

+

k

2),

<f>(a

+

S),
</>(«*).

i+1

<M«

)

0)

beginning with the term which follows

of

The number of these terms is ak+l - ak or ak (a- 1), and each them is greater than <£(a* +1 ); hence their sum is greater than
,

1

ak (a-

k+1

1) <f>(a

);

that

is,

greater than

x

ak+l

cf>

(a

k+1
).

By

giving to k in succession the values
4>(2)

0, 1, 2, 3,...

we have

+ 4>(3)-f<M4) + +
2)

++W>^x«*W;
Co

<]>(a

+

1)

+

<£(«

+

<f>(a+

3)+

+ <£(«*)>

x

a 2<f>(a 2 )

;

therefore,

by addition,

$!

<£(1)

>
ct

S

2

,

denote the sums of the first and second series respectively; therefore if the second series is divergent so also is the
£,
,

where
first.

S

2

Again, each term of (1) is less than <£(«*), and therefore the sum of the series is less than (a— 1) x ak <j>(ak ).

By

giving to k in succession the values

0, 1, 2, 3...

we have

<j>{2)
<f>(a

+

<£(3)
<f>(a

+ +

4>(4)

+
<f>(a

+

<£(«)

< (a-

1) x <£(1);
2

+

I)

+

2)

+

+

3)

+

+<f>(a

)<(a-

1) x a<f>(a);

therefore,

by addition

4-+(l)<(«-l){4 + *(l)};
hence
if

the second series

is

convergent so also »'

is

the

first.

Note. To obtain the general term of the second series we take </>(») the general term of the first series, write a n instead of n and multiply by a n
.

248
*305.
if p

HIGHER ALGEBRA.
The
series

whose general term

is

—^ r— n(logn)
p

is

convergent u

>

1,

and

divergent if p

=

1,

or p

<

1.

the preceding article the series will be convergent or divergent for the same values of p as the series whose general

By

term

is

1
ft" \l

1
c\y

(\V

11
X

'

a"(loga") p

'

(n\oga) p

'

(log a)'

np
;

The constant

factor

7=

r_

is

common

to every term

there-

fore the given series will be convergent or divergent for the

same

values of p as the series whose general term

is

-

.

Hence the

required result follows.
*306.

[Art. 290.]

The

series

whose general term

is

un
1
>

is

convergent or di-

vergent according as

Lim

\\\

(

-

1

]

log

n

>

1,

or

<

1.

Let us compare the given

series

with the series whose general

term

is


j)

-.

n (log ny
>
1

When

the auxiliary series
is

case the given series

convergent, and in this convergent by Art. 299, if
is

un M,+i

(w + l){log(n+l)}'

n

{log n) v

(!)•

Now when n
log (n

is

very large,
1

+

l)

= log n + log
(1)

(

+

J

=

log

n+-

,

nearly;

Hence the condition

becomes

un +
.

,

l

VN

nj
'

V N

n

log O

n

, '

thatis,

u H+l

l)(l+P ^>(l + nj \ nlog n) n log \
i

'

that

is,

u —- > u
n+l
.,

1

1 +- +
?i

P

wlogw o

CONVERGENCY AND DIVERGENCY OF
1
)

SERIES.

240

>

1

+

P
.

;

l0g?4

or

<n

CSr

1

)- 1 } 10*"^

Hence the
second part
Example.

first

may

part of the proposition is established. The be proved in the manner indicated in Art. 301.

Is the series

22

2 2 .4 2

2 2 .4 2 .6 2

^3 2 ^3 2 .5 2 ^3 2
convergent or divergent?

.5 2 .7 2

Here
ti

A. = *£* . 1 + I +

*
(1).

.-.

Lt«i

—*- =1, and we proceed to the next

test.

Fromfl),

»fe-l)=l+5
-1 - -1)1=1, and we
pass to the next
test.

<

2>-

.-.

Lim In

(

*-»
since Lt/u

ffe[Art. 295];

1

)- 1 } 108
1

^'^
1

•••^"[ffe—^— =
n

)-

108

}

"] 30
is

'

hence the given series

divergent.

Art. 183 that the use of divergent series in mathematical reasoning may lead to erroneous results. But even when the infinite series are convergent it is necessary to exercise caution in using them.

*307.

We have shewn in

For instance, the
-

series
JC
%)C

+

Ou

Jb

+

4/2~J/3
is

474~^5

+

'"

convergent

when x=l.
1

series
1

by

itself,
1

[Art. 280.] But if we multiply the 2n the coefficient of x in the product is

+

1

+

1

»

'

.

250
Denote
this

HIGHER ALGEBRA.
by a2n
;

then since
1

1

J^
when n
is infinite.

a„ 2

>
"

—— Jn
;

, '

and

is

therefore infinite

If

x=l, the product becomes %-a +a3 - aB f ... + a - a 2n+1 + a2>I+a x
gJ1
ffl

...,

._ and since the terms ol, a2/i+ 1' a2;i+2 .,. 2h' no arithmetical meaning.

...

are infinite, the series has
'

This leads us to enquire under what conditions the product of two infinite convergent series is also convergent.
*308.

Let us denote the two
x

infinite series
. . .

a + a x 4- a 2 x + a 3x +
bQ

+ a2 x
+
b 2n x
2n

+

.

.

.,

+

b^x

+

b 2x

2

+

b3 x

3

+

.

.

.

+

.

.

by

A
If

and

B

respectively.

we multiply
a <A + ( a

these series together

we

obtain a result of

the form

A + a(A) x +

(

a J>o + a fii +

afiz)

x2 +

...

Suppose this series to be continued to infinity and let us denote it by G ; then we have to examine under what conditions C may be regarded as the true arithmetical equivalent of the product AB.
First suppose that all the terms in
2/1
, ,

A and B

are positive.

Let A„ B„ C„ denote the series formed by taking the 2« 2« 2w + 1 terms of A, B, C respectively.
If

JO

first

multiply together the two series A 2ai B2ni the coefficient of each power of x in their product is equal to the coefficient of the like power of x in C as far as the term x 2 " ; but in A 2n B, n 2n there are terms containing powers of x higher than x whilst 2n x is the highest power of x in C0n ; hence

we

,

^o B* >
2/i

2/1

C

2 2/1

.

If

we form

C 2n includes all besides ; hence

the product A B the last term is a b x 2n but the terms in the product and some other terms
;

C. %n

>A B
ii

.

ii

CONVERGENCY AND DIVERGENCY OF
intermediate in value between 1 whatever be the value or n.
T

SERIES.

251
,

Thus
Let

C

2"

is

/.

A B B B and A B
2/»

2/i'

-4

and

B

be convergent series

;

put
Y.

A = A-X, B

=B-

where and Y are the remainders after n terms of the series have been taken; then when n is infinite and Y are both

X
.-.

X

indefinitely small.

therefore

A n B H = (A-X)(B-Y) = AB-BX-AY+XY' the limit of A B is AB. since A and B are botli finite.
}

Similarly, the limit of

A 2n B„ a

is

AB.

Therefore
since
it lies

which is the limit of C 2n must be equal to between the limits of A B B B and A 2« B„ n
.

C

AB

2;«

Next suppose the terms
sign.

in

A and B

are not

all of

the same

the inequalities A 2n B n > C„ > A b B b are n I 2b 2b necessarily true, and we cannot reason as in the former case.

In

this

case

not

Let us denote the aggregates of the positive terms in the two series by P P' respectively, and the aggregates of the negative terms by iV, N'; so that
t

A = P-N, B^F-N'.
each of the expressions P, vergent series, the equation
if

Then

P\

JV,

N' represents a con-

AB = PF- NF- PN' + NN\
each of the expressions PP\ NF, PN\ NN' is a convergent series, by the former part and thus the product of the two series A and of the proposition B is a convergent series.
intelligible, for
;

has a meaning perfectly

Hence
that the

the

sum
if

product of two series will be convergent provided of all the terms of the same sign in each is a con-

vergent series.

each of the expressions P, P', N' represents a divergent series (as in the preceding article, where also = = N), then all the expressions PF, NF, PN\ NN' are and N' divergent series. When this is the case, a careful investigation is necessary in each particular example in order to ascertain whether the product is convergent or not.
y

But

N

F P

252

HIGHER ALGEBRA.

^EXAMPLES. XXI.
Find whether the following
1
.r

b.

series are convergent or divergent

2

1.

1+ 2*4 + 2.4.6'8 + 2.4.6.8.10'
3

1.3.5

#*

1.3.5.7.9

£«

12

+

3.6

2.

1

+ ^+7.
22
2

3.6.9 ^ 3.6.9.12 2 ^ +7.10.13* + 7.10.13.16 A + 10
2 2 .4 2
G

3.

o

^+374^+3.4.5.6^+3.4.5.6.7.8*°+

a

2 2 42 6 2
.
.

o

4 *'

n 2# ^ 3
1

2

.?

2

43

^
1

54

|2

|3

^

1

^h

|4

,5

1

12

13

14

*

+ 2 2 .4 2 r + 2 2 .4 2 .6 2 ^ 2 + 22
'

l2

12

.3 2

12

.3 2 .5 2

*

7

i

g(l-a)
,

'•

X "T

"*"

,

(

l

+ a)«(l-g)(2- g)
l2
.

12

~

22
I2

(2

+ q)(l+a)q(l-a)(2-,a)(3-q)
.

22 32
.

a being a proper
8

fraction.

a+x
*

IT*—
1+ 1
+
.

(a + 2#) 2 12~

(a + 3ai) 3 + "13"" +

9

.

^ ^MM,
y
1
.

Z

.

y (y+1) a(a + l)(a + 2)/30+l)(/3 + 2)

1.2.3.y(y+l)(y + 2)
10.

"*"
'

x1 (log 2)* + a?3 (log 3)i + a?4
i

(log 4)*

+ +
where *
is is

11.

+a+

__^+—_-^
=
1

12

If
-

^;

integer,

^_a_i

shew that the series w j s positive, and divergent if

'^r^w^S^' + + +
?^

a positive
\

A -a- 1

2

«3

is

convergent negative or zero

if

CHAPTER

XXII.

Undetermined Coefficients.
309. In Art. 230 of the Elementary Algebra, it Avas proved that if any rational integral function of x vanishes when x = a, it is divisible by x — a. Cor.] [See also Art. 514.

Let

n n p x +p x
x

"
'

+ pjf

"2

+

+pn

be a rational integral function of x of n dimensions, which vanishes when x is equal to each of the unequal quantities
«!>

«*,

%i

«„•

Denote the function hy f(x); by x - a we have
,

tlien

since

f(x)

is

divisible

l

the

f(x)=:(x-a )(p x"- + quotient being of n — 1 dimensions.
l

i

),

Similarly, since f(x) is divisible

by x -a,

7 ,

we have

2W

n~X

+

= (x-aj-(pjf- +
— 2 dimensions; and

the quotient being of n

Proceeding in this way,
visions

we

shall finally obtain after

n

di-

f(x)
310.

=p

(x

- a) (x-a}(x-aa)

(x- aH).

If a rational integral function of\\ dimensions vanishes for more than n values of the variable, the coefficient of each power of the variable must be zero.
Let the function be denoted hyf(x), where
f(x)
!>x"

+p x"~
)

x

+p,c'-' +

+p

n

;

~
254

:

HIGHER ALGEBRA.
is

and suppose tha,tf(x) vanishes when x a n ; then unequal values a lt a2i a3

equal to each of the

f( x) =Po (x - a i) ( x ~ a 2 ) ( x ~ °0
Let
c

(

x ~ a ,);

be another value of x which makes f(x) vanish since f(c) = 0, we have
Po
(

then

c

~ a i)

a*)

(

G

~ a s)

(c-«J = 0;
of

and therefore p = 0,
factors
is

since,

by hypothesis, none

the other

equal to zero.

2\x

n-

x

Hence f (x) reduces to "~ n-2 +p2x + 2) 3X 3+ +Pn

-

By
of x,

hypothesis this expression vanishes for more than and therefore p = 0.
x

n values

2>o,

In a similar manner we may shew that each of the P3 Vn mus t be equal to zero.
,

coefficients

This result

may

also

be enunciated as follows

If a rational integral function of n dimensions vanishes for more than n values of the variable, it must vanish for every value
of the variable.
If the function f(x) vanishes for more than Cor. has more than n roots. (x) — of x, the equation

n

values

f

Hence
roots
it is

also,

if

an

equation of n dimensions has more than

n

an

identity.

Example.

Prove that
(x
(a

-

b) (x

-

c) c)

(x
{b

- c)
-c)

(x
(6

-b) (a-

- a) - a)

(x

-

a) (x

— b) _

1

(c-a) (c-b)~

This equation is of tivo dimensions, and it is evidently satisfied by each of the three values a, 6, c ; hence it is an identity.

If two rational integral functions of n dimensions are equal for more than n values of the variable, they are equal for
311. every value of the variable.

Suppose that the two functions
2)

xn

+p x
1

n-1

+2> 2 x"- +

2

+pH

,

n 2 qox + q^"- + q 2 x- +
1

+ qmt
2

are equal for
U>»

more than n values
'1

of

x ; then the expression

- %) x +

(Pi

-

?i)

x

"~ l

+

(p»

"~ - ad x +

+ (p* -

?.)

UNDETERMINED COEFFICIENTS.
vanishes for more than preceding article,

255

n values

of x;

and therefore, by the

that

is,

2\

= %>

Pi=9li>

Pi^Vv

l> n

=

*
<l n
>

Hence the two expressions are
equal for every value of the variable.

identical,

and therefore are

Thus

if two rational integral functions are identically equal, we equate the coefficients of the like powers of the variable.

may

This is the principle Art. 227.

we assumed

in the

Elementary Algebra,

This proposition still holds if one of the functions Cor. For instance, if of lower dimensions than the other.

is

p
we have
q =
0,

x"

+ pff~ + pjf~ 2 + pjf~* + = q 2x n ~ 2 + q 3 x n ~ 3 + +q n
l

+pn
,

only to suppose that in the above investigation q o = and then Ave obtain

0,

^o=°> Pi=°> P2=vs Ps=q 3
>

>

p,,

=

q»-

312.

The theorem

of the preceding article is usually referred

The application to as the Principle of Undetermined Coefficients. of this principle is illustrated in the following examples.
Example,
1.

Find the sum of the series 1.2 + 2.3 + 3.4+

+n(n+l).

Assume

that

1.2 + 2. 3 + 3. 4 + ... + n(n + l)=A + Bn+Cn 2 + Dn3 + Eni +..., where A, B, C, D, E,... are quantities independent of n, whose values have
to be determined.

Change n
1. 2

into

n+

1

;

then
(?t

+ 2.3+...+?i(;i + l) +

+ l) (n + 2) = A+B(n + l) + C(n+l)* + D(n + l)3 + E(n + iy+....
{2n + l)

By subtraction, (n + 1) [n+2) = B+C

+D

(3}v>

+ 3}i + l) + E
;

{±-n*

+ 6ri- + ±n + l)+

..

.

This equation being true for all integral values of n, the coefficients of the respective powers of n on each side must be equal thus E and all succeeding
coefficients

must be equal

to zero,

and

3D = 1;
whence
1)

3D + 2C = 3;

D + C + B = 2;
B=
2
-

=-

1
,

o

(7=1,

.

o

256
Hence the sum

HIGHER ALGEBRA.

=A + — + n2 + - n3 o
o

.

To

find A, put

n = l; the

series

then reduces to

its first

term, and

2=A

+ 2,

or

A = 0.
(n

Hence
Note.

1 .2

+2

.

3

+ 3.

4

+ ... + n(;i + l) = - n

+ 1)

(n + 2).

be seen from this example that when the n lh term is a rational integral function of n, it is sufficient to assume for the sum a function of n which is of one dimension higher than the w th term of the
It will

series.

Example

2.

Find the conditions that x3 +px2 + qx + r may be x 2 + ax + b.
x3 +px2 + qx + r=(x + k)
coefficients of the like
(x 2

divisible

by

Assume
Equating the

+ ax + 6).
x,

powers of

we have

k + a=p, ak + b = q, kb = r.

From

the last equation k =
r

-

;

b

hence by substitution we obtain
ar — +b = q;
,

r
that
is,

n + a=p, and

r

=b

(p-a), and ar = b (q-b);

which are the conditions required.

EXAMPLES.
l 2 +3*

XXII.

a.

Find by the method of Undetermined
1.

Coefficients the ,sum of

+ 5*+7*+...to n
2 2

terms.
..

2. 3.

1.2. 3 + 2. 3. 4 + 3. 4. 5 +
1. 2 2
2

.ton terms.
terms.

4.
5.

+ 2.3 + 3.4 + 4.5 +... to n I 3 + 33 + 5 3 + V 3 + .to n terms. l 4 + 2* + 3 4 + 4 4 + ...to?i terms.
. .

Find the condition that x3 -3px + 2q may be factor of the form a?+%ax + a2
6.
.

divisible

by a

7.
8.

Find the conditions that ax3 + hv2 -\-cx + d may be a perfect cube.
Find the conditions that a2 AA + bx3 +cx2 + dx+f 2 may be a
Prove that ax2 + 2bxy + cif- + 2tlv + 2ey +/
d- = a/, e2
is

perfect square.
9.
if b'
1

a perfect square,

= ac,

= cf.

UNDETERMINED COEFFICIENTS.
10. 11.
12.

257
<id

If

a.<

:i

+ bx2 + cx + d is
is

divisible

by x2 + h 2 prove that
,

= bc.

If

3tP

— f>qx+4r

divisible
:

by

(x

— c) 2 shew
,

that g*=r*,

Trove the identities

(

}

a2 (x-b)(x— c) b 2 (x-c)(x — a) + ~(b-c){b-a)~ (a-6)(«-c)
(y-^>)(^-c)<.y-cQ
t"

+ ~Jc^aJ(c-b)

c2 (x

- a) (x - b) _ "

2

w
/

n

(ff-c)(#- eg) (.?-«)

(rt-6)(a-c)(a-J)" (b-c)(b-d)(b-a)

+
13.

(x - d) (x - a) (x

+ \d-a){d-b)\d-c)** {c-d){c-a)(c-b)

~b)

(x

- a)

(x

- b) (x - c)

'

Find the condition that

ax2 + 2/ixy + by 2 -f 2gx + 2fy + c

may

be the product of two factors of the form
jfctf+gy+r,
jt/.t'

+ ^'y + r'.
l~,
77,

14.

If £

= lx + my + nz,
t

r)

= nx + ly + mz, £=mx + n// +
y, z

and

if

the

same equations are true for all values of x, changed with x y, 2 respectively, shew that
l

when

£,

£ are inter-

2

+2mn = l,
3
,

m 2 + 2ln = 0,

n 2 + 2lm=0.
-//

15.

Shew that the sum
2
,

of the products

-

/•

together of the n

quantities a, a

a
y

,..a n is

(

«

+

1

-l)(tt*- + a

-l)...(a»-l)
1)

(a

-1)

(a 2

-

1).. .(a*-'-

a

i(„-r)(»-r+l).

313. is equal If the infinite series a + a x + a.,x + a 3 x + to zero for every finite value of x for which the series is convergent, tit en each, coefficient must be equal to zero identically.
2

2

3

Let the series be denoted by S, and let S\ stand for the ex2 pression a + a 2 x + a x + then S = a + xS and therefore, = for all finite values of x. But since S l>y hypothesis, a + xS is convergent, #, cannot exceed some finite limit; tlierefore by taking x small enough xS may be made as small as we please. In this case the limit of & is a but S is always zero, therefore a Q must be equal to zero identically.
;
l

:i

t

,

t

x

;

x;

Removing the term a we have xS = for all finite values 2 that is, a + a 2x + ajc + vanishes for all finite values of
, x x

of
x.

Similarly,
coefficients a n

we may prove
a.,,

a

is

succession that each equal to zero identically.
in

of

the

H. ir.A.

17


258

;

. .

HIGHER ALGEBRA.

314. If tivo infinite series are equal to one another for every finite value of the variable for which both series are convergent, the coefficients of like powers of the variable in the two series are equal.

Suppose that the two

series are

denoted by

3 2 a + ax x + a x + a 3 x +

and

A +A
" Ao +
"A

x

x+

A x + Aj? +
2 2

;

then the expression
«o
(«i
i

)

x + (a2 -

A

2 2)

°°

+ (« - A a)

°f

+
;

vanishes for all values of x within the assigned limits by the last article

therefore

a
that
is,

-A = O a.-A^Q,
t

aB -A a = 0, a3
a2 =

-A a = 0,
,

« =

^

>

°i-^n

A

2,

a.

d

= A3

;

which proves the proposition.
Example
1.

Expand
.

2
-=

+ x2

in a series of ascending powers of x as far

5 as the term involving x

Let

r—
,

2

+ X2
'-

— =a
2

+ a x x + a 2 x'2 + a.jx? +
whose values

...,

where a

fl

a

,

a. 2

,

a

:i

,...

are constants
)

are to be determined; then
re

2

+ x 2 — (1 + x - x 2

(a Q

+ Oj a; + a 2 ar + o a

3

+

. .
.

)

In this equation we may equate the coefficients of like powers of x on n each side. On the right-hand side the coefficient of x is a n + a u _ 1 - a n _ 2 2 and therefore, since x is the highest power of x on the left, for all values of ?t>2 we have
,

this will suffice to find the successive coefficients after the first three

have

been obtained.

To determine
a

these

we have the equations
a. 2

= 2,
a

a1

+ a = 0,

+ a1 -a = l;
a 2 =5.
a 3 = -7;
;

whence
Also
a3

= 2, ^=-2,

+ a 2 -a 1 = 0, whence
~~

a4 + a 3

a 2 = 0, whence a 4 = 12

and
2

a 5 + a±-a 3 = 0, whence a 5 = - 19
,

thus

l + ic-a;

+ X~ = 2 2x + 5x 2 „ 2

7.t 3

+ 12x 4 - 19a 5 +

. .


Example
2.

.

.

UNDETERMINED COEFFICIENTS.
Prove that
if

250

n and

r are positive integers

*-.fr-y+«fez3 .-y- -fr-?)fr-«> fr
(

£

3

~+

.„

I

is

equal to

if r

be less than n, and to

|w if r

=n

We

have

= x n + terms containing
Again, by the Binomial Theorem,

higher powers of x.

.

.(1).

(g*-l)n =c»w -ne(»-l)ai + ^_(±Ll) e(n-2)*_

j

(

2 ).

By expanding each
of

of the terms

e

nx
,

e (n

~ l)X
y

...

we

find that the coefficient

xr in

(2) is

nr
|r

(»-l) r
[r

n(n-l) (n-2) r
j2
|r

w(m-1)(w-2) (n-3)*
|3

1

r

and by equating the

coefficients of

x r in

(1)

and

(2)

the result follows.

y = ax + bx 2 + ex3 + express x in ascending powers of y as far as the term involving y 3

Example

3.

If

,

.

Assume
and substitute
y = a{py

x=py + qy 2 + ry 3 +
in the given series
;

,

thus
t

+ qy* + ry 3 +...) + b(py + qy 2 +...y2 + c{2>y + qif+...y
coefficients of like

+....

Equating

powers of
;

y,

we have
.

an = 1

whence p = whence q =

a

aq

+ bp- =

;


a

a6

5

.

ar

+ 2bpq + cp 3 —
V &'V" # = •'---4-

;

whence

r

= = —5

,

a

1

m,

Thus
This
Cor.
is

+

(2&
-

-

-

ac) y

?—

an example of Reversion of
If

Series.

the series for

?/

be given in the form

y= k

+ ax + bx 2 + ex? + ...

put
then
z

y-k = z;
— ax + bx- + ex3 +
.

.

;

from which x

may

be expanded in ascending powers of

z,

that

is

of y - k.

17—2

I


HIGHER ALGEBRA.

« '

260

EXAMPLES.
Expand the

XXII.

b.

following expressions in ascending powers of

x as

far

1+2^; 1_^_^.24
3+*
'

l-&g
"'

i_#_6#21

*
.

l+x 2+.r + .r2'

2

-x—x

2

5

+ ax — ax2 — .r3

l -

6

th Find « and b so that the n term in the expansion of

a
7-

-4-

bv
--,

(l-.r)-

may

be (Sn-2)xn
7.

~

1
.

Find
2
3

a, b, c
,

so that the coefficient of
,

xn

in the expansion of

a + bx + cx

—(l-.r)^— may be J
8.

n- +
(3/

,

l.

If y 2 +

% =#

+ 1),
|.r

shew that one value

of

y
i

is

+ s .r-js S A +
e?/

9.

If cxz

+ ax -y = 0,
y a

shew that one value of x
3

3c2;/5

12c3;/ 7
«'°

a4

a7

Hence shew that
equation
3

x= -00999999 is an approximate x + 100.? -1 = 0. To how many places of
?

solution of the

decimals

is

the

result correct
10.

the In the expansion of ( 1 + x) ( 1 + ax) ( 1 + a\c) ( 1 + a\v) factors being infinite, and a < 1, shew that the coefficient of number of
,

Xr

.

1S

I
2 3

\

11.

(l-a)(l-a )(l-« ) When a < 1, find the coefficient
(1

(l-O
of

n

hr(r-l)

xn

in the expansion of

- ax)

(1

—a 2x) (1 — dAx)
n ^~ 1 '
K

to inf.

12.

If

n

is

a positive integer, shew that

(1)

nn+1 -n(n-l)n+1 +
n n -(n+l)(n-l) n +

(n-2)* +1 -

=jn\ n+.l
=1;

;

(2)

——

^-

(n-2)«-

the series in each case being extended to
(3)

n terms
\n;

;

and

l"-»2»+

7t

^~ 1

<

3a -

=(-l)w

(4)

(n+p) n -n(n+p-l) n

+ —^—

'

(n+p-2) n -

= '— \n;

the series in the last two cases being extended to

n + 1 terms.

CHAPTER

XXIII.

Partial Fractions.
In elementary Algebra, a group of fractions connected by the signs of addition and subtraction is reduced to a more simple form by being collected into one single fraction whose denominator is the lowest common denominator of the given
315.

process of separating a fraction into a group of simpler, or jwtial, fractions is often required. For — 5a; 3 ^--„ in a series of ascendexample, if we wish to expand 1 — iX -r OXT ing powers of x, we might use the method of Art. 314, Ex. 1, and so obtain as many terms as we please. But if we wish to find the general term of the series this method is inapplicable, and it is simpler to express the given fraction in the equivalent form 1 2 -1 and (1 — 3aj) -1 Each of the expressions (1 —a;)
fractions.
1-

But the converse

— x l — ox can now be expanded by
I

.

the Binomial Theorem, and the general

term obtained.
316. In the present chapter we shall give some examples illustrating the decomposition of a rational fraction into partial fractions. For a fuller discussion of the subject the reader is referred to Serret's Cours d'Algebre Superieure, or to treatises on In these works it is proved that any the Integral Calculus. rational fraction may be resolved into a series of partial fractions;

and that to any

linear factor

x—a

in the

denominator there

cor-

responds a partial fraction of the form
factor

x-

b occurring twice in

to any linear X cc the denominator there correspond
;

——
If

two partial fractions,
times, there is

— x—
7?

l

-j

and

-.

b

(x

— —
7?

*__

.

by
.

x — b occurs
so

three

an additional fraction

hnl au d (x-b)"

on

-

To

;

s

262
any
quadratic
factor

HIGHER ALGEBRA.

fraction of the form
twice, there
'

+px + q —Px + Q if
x2
:

there

corresponds

a

partial

x' +])x

+q

2 the factor x + vx + q occurs

is

a second partial fraction y—~ r (x-+2)x
quantities
x.

P x + QL
+

;

and so

on.

q)

Here the

Av B

lt

P B
2
,

3,

P, Q,

Pv Q

are
x

all

independent of

We
follow.

shall

make

use of these results in

the examples that

Example

1.

5x — 11 Separate =-^ ^ into partial fractions.

Since the denominator

+ x - 6 = (x + 2) (2x - 3), we assume 5.r-ll B A + x+2 2.c-3' 2x 2 + x-§
2.r
2

where A and
determined.

B

are quantities independent of

x whose values have

to be

Clearing of fractions,

5x-ll = A (2x-S) + B(x + 2).
Since this equation powers of x thus
;

is identically true,

we may equate

coefficients of like

2A+B = 5,
whence
'"'

~SA + 2B=-U;

A = 3,
5.r-ll
2x
2

B= -1.
3
1

+ x-6~ x + 2
r~.

2x-B'

Example

2.

Resolve
.

;

(x

- a)
z-.

(x

+ b)

r,

into partial fractions.

Assume

A B mx + n =-r = h x-a x + b' (x-a)(x + b) mx + n = A {x + b) +B (x-a)
-.

.'

.

(1).

We
is

might now equate

simpler to
Since

and find the values proceed in the following manner.
coefficients

of

A and

B, but

it

A and B
put

are independent of x, ov x = a; then

we may give to x any value we please.

In

(1)

x-a = 0,

ma + n A= ra+b
putting x + b
t

n = 0,

or

x—

-

,

b,

— B— nib-n
-r,

.

CI "T*

mx + n (x - a) (x + b)

1

/ma + n

mb-ii\
x+
b

~ a b + \

x-a

J


23
v

;

r

.

PARTIAL FRACTIONS.
Example
3.

203

Resolve

7^7—-.-rrx-— —^ mto y£x — I) (J x
j

- 11

2 r'

Assume
. • .

=

(2.c-l)
(3

^-^

23x-ll.r 2

rm r= n r+ 5 + x 3 2.c-l (3 + x)(3-.r)
(3

ABC
H 5

partial fractions.

3-x
1) (3

w
i

1)

5

23x - lLc2 = .1

+ x)

-x)+B

(2x - 1) (3 - x)

+ G (2x -

+ x).

By

putting in succession

2^-1 = 0,

3

+ x— 0,
-1.
4

3 -#

= 0, we

find that

4 = 1, B = i, C=
•'•

- lis 2 1 + 2 )~2x-l (2.c-l)(9-x
23.c

1_
3 -a:'

3+x

Example

4.

Resolve

Assume
.•
.

3s 2

z-^-n (x - 2)- (1

—^— — ^ + s-2 — —— 3.t
'

-.

[x —

2

+x—2
'

,

:

into partial fractions.

&)"

(J.

&x\

s-7
2x)
(x

"

~k

1

k~ H 2x
(1

B « + x-2
2x) {x

G
(x
2)

-

2)

2

'

%x-

+ x-2 = A

-

2)

2

-

-

+ C (1 - 2x).

Let
let
a;

1

- 2x = 0, then
then

A= -o

-2 = 0,

C=-4.
;

To

find B, equate the coefficients of x 2

thus

3

= A - 2B whence B = - ^
;

o

3.r' '

+xa

2

(x-

2)

(1

- 2x)

3(1- 2x)
19a;
-r-,

3 {x

- 2)

(x

-

2)

2

'

Example 1

5.

Resolve

[x

42 —— into partial fractions. +l)(x-4]
2

-r

L

Assume
.-.

42-19.C
-7-3

(.^

+ 1)^-4)
19.r

tt,

r;

=

Ax + B —3 + x-+l

C
.i--4'

-

42 -

= (Ax +B)

(x-

4)

+ C (x* + l).

Let x = 4, then
equating coefficients of
x'
2
,

C=-2;
= A + C, and .4=2; 42 = - 4Z? + C, and B = - 11,
2s -11
2

equating the absolute terms,
•'"

42 - 19a

p+l)(x-4)"^TT *-4*
employed in the following example
will

317.

The

artifice

sometimes be found useful.

,

;

2G4
Example. r
Resolve
-;

HIGHER ALGEBRA.
9x* — 24rc2 + 48# (a:-2) 4 (a:+l)
7^t~.
=rr-

into partial fractions.

Assume
where A
is

9z 3 -24;r 2 +48;c —. ^tx~? rrv (x-2)*(x + l)

— —n + x+1

A

f(x)

{x-2) 4
of x

'

some constant, and / (x) a function
.-.

whose value remains

to

he determined.

9x3 -24£ 2 + 48x = ,l (x-2)*+(x + l)f{x).

Let

x= - 1,
(x

then
^4

A=

-

1.

Substituting for

and transposing,
9a3 - 24s3 + 48x = x4 + x* + 16* + 16

+ 1) / (a) = [x - 2) 4 +

.'./(*)

= £ 3 + 16.
•r

To determine
.r

the partial fractions corresponding to
3

-

\x—2)

+ ——
3

16

,

put

x-2 = 2;

+16

(2

+ 2) 3 + 16
Z*

23

+ 6^ 2 +122 + 24
z*

then

(x-2) 4
1

~z

+
1

6
z^

+

12
~z^

+

24
~z

~

4

6
'

12

24

+ /. nva /_ n\n + ~ x-2^(x-2)'2 + (x-2f^ (x-2f
9x3

"
318.

-24j; 2 + 48*
(x

(x-2) 4

+ l)

= "

x+1
-i

1,1,+
-\

x-2

«

+ (x-2)"
/

6

,

12
TZ

Z

Svi

(x-2f

^-J

+

24

(x-2) 4

*

the preceding examples the numerator has been of lower dimensions than the denominator ; if this is not the case, we divide the numerator by the denominator until a remainder is obtained which is of lower dimensions than the denominator.

In

all

Example. *

Resolve

-zr-=

6r 3 + 5# 2 -7 =- into partial fractions. ox- - 2x - 1

By

division,

3a:

2

8a;

2.x

-

v-

1

= 2x + 3 +
5
3.c

Sx 2 - 2x 1

1

and

— -4 = 3x -2x-l
^-=
2
pr

^

s

+

= l

+

x-1'

'*

1 5 6^ + 5^-7 = 2.r + 3 + - - + 2 *-l' 3.T+1 3x -2x-l

319. fractions

We

shall

now

explain

how

resolution

into

partial

used to facilitate the expansion of a rational fraction in ascending powers of x.

may be

PARTIAL FRACTIONS.
Example
1.

265
when expanded l
in a

Find the general term of
powers of
x.

—-——— _ 2—
;;

,'-

-|-

;•

.

(*- 2)-(l - 2x)

series of ascending

By Ex.

4,

Art. 316,

we have

3.r 2

+ .r-2
2

(*-2) (l-2*)

3(1-2*)

3(1-2*)

15 15
+
1

4

3(*-2)
3(2-*)

(*-2) 2
4
(2-a?) a

Hence the general term
/
V

of the expansion

is

r+6
3

r+l\
sr

6

2r

y
and
find

Example

2.

Expand
(1

7

+*
.

+

r-^r— *) (I + *~)

in ascending powers of *

the general term.

Assume
.\

.

(1

H

+ *) (1+* ) 7 + * = J(l + * 2 + (E*+C)(l + *).
)

+ -—*
7

-=J

—+ 1+x
.4

JB*

+C 1 + *2

;

Lctl + *=:0,
equating the absolute terms,
equating the coefficients of *
2
,

then
7

A = 3;
whence C = i
whence
;

= A + C, = A + B,
4-3*
1

B— - 3.

+* _ 2 (1 + *)(1+* )
7

+ 1 + *^ 1+* 2

3

= 3(1 + .r)- + (4 = 3{l-* + * 2 + (4-3*)
To
(1)

3*) (1

+ x 2 )~l
...j

+ (_l)P;C P +

{l-.r 2 + *-»-

+ (-1)p*'^+...}.
r

find the coefficient of xr
If
/•

:

is

even, the coefficient of * r in the second series is
r

4(-l) 2
r-l

;

therefore in the expansion the coefficient of x r is 3
(2)

+4 (-

1)

2
.

If r is odd, the coefficient of *r in the
r+l

second series

is

-3(-

1)

'-

and the required

coefficient is 3

(

-

1)

2

-

3.

EXAMPLES.
Resolve into partial fractions
,
:

XXIII.

lx-\ l-bj; + 6jf-

46+13.r
'

l+3. r + 2 .r 2
(1 -2.r) (1
-.//-')'

12.t 2

-lU--15'

}

266
.y

HIGHER ALGEBRA.
+13 2 (x-l)(x -5x+6)'
2

-

10a;

'

2x*+x2 -x-3 x(x-l)(2x + 3)'
„ 7*

9 6
*

x*- 3x* -3a;2 + 10
(#+l) 2 (#-3)
2^2 -lLr + 5
(^--3) (^ 2

(a;-l)(^ + 2) 2

'

26^2 + 208o;

Q

+ l)(^ + 5)' 3^-8x2 +10
(a;

2

+ 2^'

5)

,,

5^ + 6.r 2 + 5.r
(^
2

(07-1)*

*

-l)(^+l) 3

Find the general term of the following expressions when expanded in ascending powers of x.
12#

+ 3# l + llo; + 28^* 2#-4
l
t^

5a;
'

+6

(2+a?)(l-#)'

16.

u
(

#2 + 7;f + 3
tf

2

+ 7a- + uy

15.

(1

2 - x5tt^ ) (1

—^ -

4 + 3^+2a'2
(1

* •

2.r)

- x)

1

+ x - 2x 2
'

)

17.

7,

w ttx* (l+a?)(l-4a?) 2
.

3

+ 2x-x2
,

• *

no 18.

4 + 7x
(2

19

.

*"*
(^-1)(^2 +1)"

+ 3a;)(l+.r) 2 1 -* +i *
(1-tf) 3

20.

21

.

„cw?) (1

1
(1

- te)

(1

- co;)

*

22

.
'

»-«•
(2

- 3.r + a2) 2

'

23.
(l)
[

Find the sum of n terms of the
I

series

(i+^)(i+^2 )
(1

+

— + + (i+^2 )(n-^) (i+^)(i+^4 )
+
(1

.

.

x (1 - ax)
+x)
a?

ax

(1

- a2x)
(1

^

'

(l

+ ax) (1 + a%)
find the

+ ax)

(1

+a%)
xA

+ a 3.r)

+

24.

When

< 1,

sum
x2

of the infinite series

1

(l-x)
25.

+ (l-x3
)

(1

-a?)

(1

-.r5 )

+

(1-tf5 )

(1

-^) +

Sum

to

n terms the

series

whose p th term

is

xp(1+xp + 1 )

(l-^)(l-.^ + 1 )(l-^ + 2 )'
Prove that the sum of the homogeneous products of n dimen26. sions which can be formed of the letters a, b, c and their powers is

an + 2 (b -c) + bn + 2 (c- a) + cn + 2 (a-b) a2 (b-c) + b 2 (c-a) + c2 (a-b)

*

CHAPTER XXIV.
Recurring Series.
320.
in

A series

u + u + u 2 + u3 +
l

which from and after a certain term each term is equal to the sum of a fixed number of the preceding terms multiplied respectively by certain constants is called a recurring series.
321.

In the

series
1

+ 2x +

3ar

+ 4a? +

5a;

4

+

,

each term after the second is equal to the sum of the two preceding terms multiplied respectively by the constants 2x, and - x 2 j these quantities being called constants because they are Thus the same for all values of n.

5x4
that
is,

= 2x

.

4a;

3

+ (- x2 )
;

.

3a;

2
;

u4 = 2xn3 — x2u 2
and generally when n is greater than with the two that immediately precede

1,

it
.

each term is connected by the equation
,

uh —
or

2xiin— 1
,
,

— x 2 u n—2*
ii

u H — 2xu n — + x 2 u
1

—„ 2

=

0.
,

In this equation the coefficients of u n «*,_,, and l*,_ a taken with their proper signs, form what is called the scale of relation.
,

Thus the
is

series
1

a recurring series

+ 2x + 3a; 2 + 4a; 3 + 5x 4 + in which the scale of relation 2 1 - 2x + x
.

is

322.

If the scale of relation of a recurring series is given,
sufficient

any term can be found when a

number

of the preceding

268

HIGHER ALGEBRA.

As the method of procedure is the same terms are known. however many terms the scale of relation may consist of, the following illustration will be sufficient.
If
is
1

- px - qx2 - rx 3

the scale of relation of the series

a + a<x + a a x 2 + ajc3 +

we have
3 2 + rx3 a n - 3x + x " a n - 2x am =pan_, + &».- + «- 3 or thus any coefficient can be found when the coefficients

anx*=px a n -i x
-

"~ 1

'l

~

'l

~

<l

-

i

m

5

of the

three preceding terms are known.

Conversely, if a sufficient number of the terms of a series be given, the scale of relation may be found.
323.

Find the scale of relation of the recurring series 2 + 5x + 13x2 + 35x 3 + Let the scale of relation be 1 -px - qx*-, then to obtain p and q we have 13 - 5p - 2q = 0, and 35 - 13p - 5q = the equations
Example.
;

whence p = 5, and q= -

6,

thus the scale of relation
1

is

- 5x + 6a;2

.

If the scale of relation consists of 3 terms it involves 324. 2 constants, p and q ; and we must have 2 equations to deTo obtain the first of these we must know termine p and q. at least 3 terms of the series, and to obtain the second we Thus to obtain a scale of must have one more term given. relation involving two constants we must have at least 4 terms 'O

given.
2 3 to find the the scale of relation be 1 — px — qx - rx To obtain the first of 3 constants we must have 3 equations. these we must know at least 4 terms of the series, and to obtain the other two we must have two more terms given hence to find a scale of relation involving 3 constants, at least G terms of the

If

,

;

series

must be

given.

we

Generally, to find a scale of relation involving must know at least 2m consecutive terms.

m

constants,

Conversely,

if

2m consecutive

terms are given, we

may assume

for the scale of relation
1

~ l\ x ~ l\ x * ~ lhx* ~

-PJ**


RECURRING
325.

SERIES.
series.

2G9

To find

the

sum ofn

terms of a recurring

of finding the sum is the same whatever be the scale of relation ; for simplicity we shall suppose it to contain only two constants.

The method

Let the

series

be

a u + axx + a 2 x 2 + aj£ +
the sum be S ; let the scale of relation be so that for every value of n greater than 1, we have

(1)
1

and

let

— px — qx*

;

Now S—a + a.x + a,x + ...+ a — px S= — pa x — pa x* — ... — - qa? S= - qajt? - ... -qa
2
it

1

2

/»—

,x"~\
1 '

2^>ci

x

H_2

x

n~ l

— pa
i

x*t
u

H

_ 3x*x

1

-qa H _ x n -qa H _ x
l

+

\
1

...

(i

- px _

qtf)

S-

a + {a

x

-pa ) x - {pan _ + qan _ a ) x n - qa^x**
power
of

,

for the coefficient of every other of the relation

x

is

zero in consequence

a n-P a n-l-<2 a «-2=
.

'

s_
is

% + («, -P<-Q x
1

(P a

,t

-,

-px— qx

2

H+l + qan - 3 ) x" + qa n _ x 2 1 - px — qx
}

Thus the sum
nominator

of a recurring series is a fraction

whose de-

the scale of relation.

32G. If the second fraction in the result of the last article decreases indefinitely as n increases indefinitely, the sum of an
infinite

number

of terms reduces to
this fraction in

—\—
1

— px — — qx"
!

-

^—

.

ascending powers of x as many terms of the series as we please; for this reason the expression original
If

we develop

explained in Art. 314,

we

shall obtain as

1

—px — qx

2

is

called the generating function of the series.

327.
°,

From
,

the result of Art. 325,
.)

we
...

obtain

an + ( a v 1

—P a X ———%— = a -px
'

£

qx'

lt

°

+ a.x + ax- + 2
'

o

+a

" -1

n+ x.xi

l

1

- px— qx

270

HIGHER ALGEBRA.

from which we see that although the generating function
1

— px —

qx 2

may
it

be used to obtain as many terms of the series as we please, can be regarded as the true equivalent of the infinite series

a + a x + a 2x 2 +
l

,

only

if

the remainder
( I**,.-,

+qan1

xn
2)

+ (2 a n-^" +l
;

— poj — qx 2
in other

vanishes when when the series

n
is

is

indefinitely increased
v

words only

convergent. o

When the generating function can be expressed as a 328. group of partial fractions the general term of a recurring series may be easily found. Thus, suppose the generating function can be decomposed into the partial fractions
1— ax
Then the & general term
r

ABC
h

1

+ bx
r

h

(I— ex) 2
r

'

is

{Aa + (the method of Art. 325.

l)

r

M+

(r

+

1)

Cc } x\
using

In this case the sum of n terms

may be found without

Example. Find the generating function, the general term, and the sum to n terms of the recurring series
1

- Ix - x 2 1

43.-C

3

-

Let the scale of relation be

-px - </.r 2 then -l + 7j>-<z = 0, -43 + 2> + 7</ = 0;
;

whence p = l,

5 = 6;

and the

scale of relation is 1

-x;

6.r 2 .

Let S denote the sum of the

series

then

S = l-lx- x 2 -4Sx s -

-xS= -Qx S=
2
.-.

- x + 7x 2 +

x*+

-6x 2 + 42.r 3 +

(l-x-6x 2)S = l-8x, J" 8 * s-

which

is

the generating function.

'

RECURRING SERIES.
If

271
we obtain
-

we separate
l)

1

-

8.r

1-x-U.rtU

^-„ into partial fractions, ±
is

1

— +
2

2a;

1

1-305'

;

whence the (r+

or general term

{(-lyw^-v ].<>.
Putting
the
r
?i

= 0,

1, 2,...n

-1,

sum

to

terms

=

{

2

- 2 x + 2%2 -...
2
1"1

+ (1

I)"" 1 2" a;'
'

1

"1
}

-

(1

+ 3a + 3%* +

.

.

.

+ 3"- xn~
1

l

)

_ ~ 2 + - I)'
(

2 n+1

xn

_ 1_- 3* xn - 3x~

l+lte

term and sum of n terms of the we have only to find the recurring series a + a + a_,+ general term and sum of the series a + a l x + a2x2 + and put x — 1 in the results.
329.

To

find the general
i

,

,

Example.

Find the general term and sum of n terms of the
1

series

+ 6 + 24 + 84+

The

and the generating function
This expression
is

scale of relation of the series 1 + 6.r 1 +x *—
is


4

+ 24x 2 + 84x3 +

.

.

.

is 1

- ox + Ooj2

,

-

1

— OX + OX"


3

.

equivalent to the partial fractions

1 If these expressions

- Sx
3r

1

- 2a;
2 r ) xr.
.

be expanded in ascending powers of x the general
(4
.

term
of

is

-

3

.

Hence the general term n terms is

of the given series is 4 2 (3' 1 - 1) - 3 (2' 1 - 1).

3r

-3.

2r ;

and the sum

may remind the student that in the preceding 330. article the generating function cannot be taken as the sum of
the series
1

We

+6x + 24:x +8±x 3 +
2

except

when x has such a value as to make the series convergent. Hence when x = 1 (in which case the series is obviously divergent)
is

the generating function But the general term of
1

not a true equivalent of the

series.

+

6

+ 24 + 84 +

is

independent qfx, and whatever value x be the coefficient of x" in
1

may

have

it will

always

+ Gx + 24* 2 + 84a 3 +
its

therefore treat this as a convergent series and find general term in the usual way, and then put x = 1.

We

:

a

272

HIGHER ALGEBRA.

EXAMPLES. XXIV.
Find the generating function and the general term of the following
series
1.
l

+ 5.r + 9.r 2 +13.r 3 + + 3x + 5x2 + 9x3 +
6a?

2.

2-.v + 5.r2 -7.r3 +
7

3.
5.

2

4.

-6x + 9x2 + 27x4 +

3+

+ Ux2 + 36.r* + 98.^ + 276.1-5 +
series
:

Find the n th term and the sum to n terms of the following
6.
8. 9.

2 2
1

+ 5 + 13 + 35+ + 7^ + 25^ + 91^+

7.

-l+6.v2 + 30.v 3 +

+ 2.v + 6x2 + 20# 3 + 66x* + 212^ +

10.

-^ + 2 + + 8+
Shew that the
series
1
2

11.

+ 2 2 + 3 2 + 42 +

+ n2

,

13

+ 23 + 3 + 4 +
3

3

+n

3
,

are recurring series, and find their scales of relation.
12.

Shew how

to

deduce the

sum

of the first

n terms

of the re-

curring series

a from the sum to
13.

+ a x + a2 x2 + a^v3 +
x

infinity.
1

Find the sum of 2n +

terms of the series

3-1 + 13-9 + 41-53+
14.

The

scales of the recurring series

are general term

+ vv+ a^x2 + a3.r3 + b + b 1x+b^c2 -{-b 3 3 + 2 l + rx + sx2 respectively; shew 1 +px+qx
a
.v
, ,

,

,

that the series whose
is

is
l

(<x n

+6 n )^"

is

a recurring series whose scale

+

(p

+ r)x + (q + s +pr) x2 + (qr +ps) x3 + qsx*.
formed having
for its

15.

If a series be

nih term the sum

of

n terms

of a given recurring series, shew that it will also form a recurring series whose scale of relation will consist of one more term than that of the given series.

CHAPTER XXV.
CONTINUED
FllACTIONS.

331.

All expression of

the form a +
c

is

+ e

a

called

a

+

...

continued fraction here the letters a, b, quantities whatever, but for the present
;

c,

may denote any
shall only consider

we

the simpler form a +
x

,

where a n a 2i « 3 ,... are positive

2

a3 +
1

...

integers.

This will be usually written in the more compact form
1

a,

+

a2 + a3 +
of quotients a

332.

When

the

number

a

,

« 3 ,...

is finite

the

continued fraction is said to be terminating ; if the number of quotients is unlimited the fraction is called an infinite contirmed
fraction.
It is possible to reduce every terminating continued fraction to an ordinary fraction by simplifying the fractions in succession

beginning from the lowest.
333.
tn

To convert a given fraction into a continued fraction.

Let

be the tnven fraction
j>

;

divide in by n, let a

be the

quotient and

the remainder

;

thus

m — a. +p —
n
si.

=a, + — n n

1

:

P
H.

A

18

274
divide

HIGHER ALGEBRA.
n by ^»,

^

«„

be the quotient and q the remainder

;

thus

n V
divide
on.

1 - = a . + q = as + - ;
'

V

'

P
9.

p by

q, let a.6

be the quotient and r the remainder
1
1

;

and

.so

Tims

— = a.
rn,

n

+
«o

=

a.

+

1

1

i

+

a2 + a3 + and we put

a3 +.
first
7)1

If

m is less

than

?t,

the

quotient
1
ti

is zero,

n
and proceed as
before.

m
It will be observed that the above process is the same as that and n ; hence if of finding the greatest common measure of and n are commensurable we shall at length arrive at a stage

m

m

where the division is exact and the process terminates. Thus every fraction whose numerator and denominator are positive
integers can be converted into a terminating continued fraction.
251
to a continued fraction.

Example.

Reduce

^^

process,

Finding the greatest we have

common measure
5

of 251

and 802 by the usual

CONTINUED FRACTIONS.
335. To shew that the convergents ewe alternately greater than the continued fraction.
less

275

and

Let the continued fraction be a +
l

1

1

a 2 + a3 +
too small because the part
is

The
aa +a3 +

first

convergent
omitted.

is «,,

and

is

is

The second convergent °
is

a
l

-i

— a
-\

,

and

is

k

too great, because the denominator aa

too small.

The

third
is

convergent
great
;

is a,

-\

,

a 2+
so on.

and

is

too small because a

CC

3

and

— %

too

the given fraction is a proper fraction a = ; if in this case we agree to consider zero as the first convergent, we may enunciate the above results as follows
t

When

:

The convergents of an odd order are
of an even order are
336.
vergents.
all greater,

all less,

and

the convergents

than the continued fraction.

To

establish the

law of formation of

the successive con-

Let the continued fraction be denoted by
1
1 1

a +
x

a 2 + a3 + a 4 +

then the

first

three convergents are
a.
1

a x a3 + a2

1

o, (a, a,

+

!)

+ «,
1

a3

.

a2 +

and we see that the numerator of the third convergent may be formed by multiplying the numerator of the second convergent by the third quotient, and adding the numerator of the first convergent also that the denominator may be formed in a similar manner.
;

Suppose that the successive convergents are formed, in a similar way; let the numerators be denoted by^,^.,, p 3 ,..., and the denominators by q lt q q 3 ,...
,

Assume
that
is,

that the law of formation holds for the

»tt
-

convergent;

suppose
1\

= »J».-i +P»-i

In

=

<*

n

?.-,

+

Q„- 2

18—2

27g

HIGHER ALGEBRA.
l)
th

The (*+

the quotient a n +

±

convergent

differs

from the

ft*

only in having

in the place of

aj hence

the (» + 1)-

«*

vergent

^n-n ^»

« n+1

+ ^»-i by + ?„_!' ?„
?

supposition.

If therefore

we put
th

denominator of the (» + l) conco. th«t the numerator and we case of which was supposed to hold in the P f

^^ ^ tttVs
^ ^*
337.

vergent, hence it holds universally.

conhold in the case of the third so on; therefore * holds for the fourth, and

to call It will be convenient

aH the n* partial quotient;

at this stage being a n + a„ 4 the complete quotient +1 + «« +2 quotient at'any stage by usually denote the complete

ft.

We shall We have seen that

&

by m ; then x differs from continued fraction be denoted let the quotient ft instead of the partial only in taking the complete
quotient a„
;

thus

X_

ft

j^i-l
n_x

~kq

+ ff»-2 + qn -2
to

'

338

// Eb
Q

6e

tfl6

n th convergent

a continued fraction, then

denoted by Let the continued fraction be
a,1

+

111

a Q + a 3 + a4 +

CONTINUED FRACTIONS.
then

277

= ("

1

2 )

(P.-

2

9«-a

-iV,

^-2)1 similarly,

But
hence

p 2 q -]\
x

q,
/>„

=

(<h

% + 1) - «x
g,

a,

= 1 = (-

l)

2
J

g^, -#,_,

=

(-

1)".

When the
still

hold

if

continued fraction we suppose that a x =

is less

0,

than unity, this result will and that the first convergent

is zero.

are calculating the numerical value of the successive convergents, the above theorem furnishes an easy test of the accuracy of the

Note.

When we

work.

Cor. q n had a

Each convergent is in its lowest terms for iipn and common divisor it would divide pn q nl — pn_ l q ni or unity
1.
;

;

which

is

impossible.
2.

Cor.

The

difference

between two successive convergents
is

is

a fraction whose numerator

unity

;

for

?«_i

q n qn ^

q,,q n -i'

EXAMPLES. XXV.
Calculate the successive convergents to
1.

a.

2

2.

2+ 2+ 3+ 1+
3+

3.

1111111 111111
2

+

l *

l '

l

6+ 1+ 1+ 11+

44-

2+

6

3+ 1+ 2+ 2+ 1+

9"

Express the following quantities as continued fractions and find the fourth convergent to each.

729
4.

278
12.

HIGHER ALGEBRA.

39*37079 inches, shew by the theory of continued fractions that 32 metres is nearly equal to 35 yards.
is

A metre
Find a

13.

days

series of fractions converging to "24226, the excess in of the true tropical year over 365 days.

14.
A,
.

A kilometre
..

is

5

18

the fractions -,

^

,

very nearly equal to "62138 miles; shew that 23 64 the == , ^z are successive approximations to «.
.

•*•*

ratio of a kilometre to a mile.

scales of equal length are divided into 162 and 209 equal parts respectively; if their zero points be coincident shew that the th 31 st division of one nearly coincides with the 40 division of the other.
15.

Two

16.

If

n3 + n u + n + l

s

is

converted into a continued fraction, shew
alternately,

that the quotients are cessive convergents.
17.

n — 1 and n+l

and

find the suc-

Shew that
Pn + \~Pn - 1
(!)

_ Pn
9.n

2n + 1
(2)

9.n

-1

(^-O^-fH-vrPn Pn U
\
/ \
+

2

-1
,

X

g«-l

\ c Jn

18.

If

is

the n th convergent to a continued fraction, and a n the

corresponding quotient, shew that

339.

Each convergent

is

nearer

to the

continued fraction than

any of the

'preceding convergents.
fraction,

Let x denote the continued

and

*—

"-

,

^-*±J

—"-±2

9*

?« + !

^+2

three consecutive convergents; then x differs from *-a±l only in

taking the complete (n + 2) th quotient in the place of a
this

;

denote

by

k:

thus

x=

? n+l +Pn

;

and

^^ ~
Pn +
1

a;

= Pn^l^n-Pn^l
?„ +
,

1
"

&+

1

(%» +l +

7.)

y. + ,

(%„ + +
,

?„)


Now
botli

«

CONTINUED FRACTIONS.
k
is

270
than q
;

greater than unity, and qm

is less
--"
'

lience

on

accounts the difference between

'

and x

is less

than the
is

difference

between

—" and

x: that

is,

every convergent

nearer

to tlie continued fraction than the next preceding convergent, and therefore a fortiori than any preceding convergent.

Combining the
follows that
tli^

result of this article with that of Art.

.°>3.>,

it

convergent of an odd order continually increase, hat are always less than the continued fraction ;
covrergents of an even order continually decrease, hut are always greater than the continued fraction.
tin'

340.

To find
p

limits to the error

made in taking any convergent

for

the

continued fraction.

Let

Y
,

p -^

1

p r_n±2
]

)0

three consecutive convorgents, and let

k denote the complete (n + 2) th quotient;

then

x=
p
<ln

^^
>+ n J
..

t

k
<ln( k <ln +

1

'.(*•«

+9 i)
i

Now
p.. — is

k
.

is

greater than

1,

therefore the difference between x and
-

less

than

i
,

and greater than

Again, since
less

<7, 1

+

l

><7„,

the error in taking -" instead of x
1

p

is

than
341.


1

5

and greater than 77-0—
last

.

?.

v. +

,

From the
p — qm

article it

appears that the error in
1

taking
or
(I

instead of the continued fraction

is

less

than

,

?.?.+,

a (a / V

;

II

., +1 ill

7

+q 2

;

;

that

is,

less

than

,) Ft— 1/

« 11+ .7 ili
.

3 "

:

hence the larger

I

a i+l

is,

the nearer does £2 approximate to the continued fraction;
/

'

.

280
therefore,

HIGHER ALGEBRA.
any convergent which immediately precedes a a near approximation to the continued fraction.
is less

large

quotient is

Again, since the error

than

g

,

it

follows that in order

to find a convergent which will differ from the continued fraction

by

less

than a given quantity a

,

we have only
is

to calculate the

successive convergents

up to

,

where q n 2

greater than

a.

properties of continued fractions enable us to find two small integers whose ratio closely approximates to that of two incommensurable quantities, or to that of two quantities whose exact ratio can only be expressed by large integers.
342.

The

Example. Find a series of fractions approximating to 3* 14159. In the process of finding the greatest common measure of 14159 and
100000, the successive quotients are
7, 15, 1, 25, 1, 7, 4.

Thus
4

3-14159 =

3+1
22
'

1

1

1

111

7+ 15+ 1+ 25+ 1+ 7+
333
'

The

successive convergents are
3
1

7

106

'

355 113
is

this last convergent

which precedes the large quotient 25

a very near
less

approximation, the error being less than

^-

,

and therefore

than

25TP5)-

»

•° 00004

-

343. Any convergent is nearer to the continued fraction than any other fraction whose denominator is less than that of the
convergent.

Let x be the continued fraction,
r

V

,

P '-*=* two consecutive
?.-,
s is less

°n
8

convergents, - a fraction whose denominator °
If possible,
let
7)

than q
r

.
"

v r - be nearer to x than —
«

,

then - must be
s


;

nearer to x than
£-5=?
it

-Ji

^1 [Art. 339]

and since x

.

lies

between
'

P -I"

and

In - J

follows that S

must

lie

between

— and —
%

9.-X

?»-!

.

.

+

.

;

CONTINUED FRACTIONS.
Hence
r

281

P»-*P.

Pn -i fWi<5 ^

l

.'.

rqn _ x ~ spn _ x

;

that

is,

Therefore

an integer p —

&

must

impossible. r be nearer to the continued fraction than less
;

than a fraction

which

is

*

344.

P If -

,

P' —

be

two consecutive conver gents
greater or less than

to

a continued
- is

fraction x, then

,

is

x 2 according as
,

greater or less than


q

,

Let k be the complete quotient corresponding to the convergent immediately succeeding ° J
°

q"
,

;

then x — -f—.
lcq '
'

— +

,

q

'

''

5
,

"*=

WW^YY w
(tfp'q
,

{hq

+ qY " "'

w
q'

pY]

=
The
lience

-pq)(pq'-2>'q)

qq'(kq'

+ q) 2
>p,
>q, and

factor ky'q'

- pq

is

positive, since p'

k>

I

—>
pp'

or

< x 2 according
,

as

]iq'

—p'q

is

positive or negative

;

that

is,

according as - > or

<

,

from the above investigation that the ex2 - q 2 M 2 q' 2 x 2 —p' 2 have the same pressions ]iq'—2 q VP ~ c c L L^-> p
Cor.
It follows
)/
i

,

sign.

EXAMPLES. XXV.
1.

b.

222

Find limits to the error
is

in taking

yards as equivalent to

a metre, given that a metre

equal to 1-0936 yards.

)

282
2.

HIGHER ALGEBRA.
Find an approximation to

JL J_ J- JL JL + 3+ 5+ 7+ 9+ 11+
which
3.

differs

from the true value by

less

than -0001.
99

Shew by the theory

of continued fractions that =- differs from

1*41421 by a quantity less than

.

4.

„ Express
1

,. ,. , n a 3 + 6a 2 + 13a+10 \ + ,14a--+15a + 7 as a continued fraction, and A « 1K rs 4 +6a 3 a
,
,
.

,

find the third convergent.
5.
is

Shew

that the difference between the

first

and n th convergent

numerically ecpial to
1
1

Mi Ms Wh
6.

+

1

(-l)n
...+

9n-l2n
s-5
,

Shew

that

if

an

is

the quotient corresponding to

p

a
^
'

Pn-1~

"

«n-l+ «u- 2 + 0»-3+

'"

«3+

a2+

'

«1

(2 )

_i- =an+ _L_ ^2
qn-1

L_
tt

...

J_
«3+

1.
«2
,

,

an-l+ «»-2+

7.

In the continued fraction
(

1111 — a+ + + «+
« «
>

n-3+

shew that

1

Pn

(

2)
is

+P\ + 1 =Pn - lPn + 1 + £>„£>„ + 2 Pn = q n -lthe
?i
th

8.

If

convergent to the continued fraction

111111 a+ a+ «+
b+
b+
6

+
• -

shew that
9.

q 2n =p 2n + u

#».. q2n - 1 = r r - n b

a

In the continued fraction

1111 a+ 6+ «+ 6+
9n + 2

'

shew that
Pn + 2~ ( ah + 2 ) P n +Pn-2 = °i

~ ( ab + 2 )

?u + •?•*- 2 = °-

''

CONTINUED FRACTIONS.
10.

283

Shew that
/ a[a\x + \

111
ar

a.v.j,+

3

+
+

oa;4

+

to 2/i quotients

x.

L

\
/

= «.v,H
:r.,+

to 2/i quotients.

oa?3

.v,+

11.

If

-r;

,

-

,

- are the n tU (n —
,

iV

,

l)

th
.

(?i-2) th convergent* to the

(^

A3

continued fractions

111
((

111
'

111
'

a i+

2+ a S+

tt

*+ a B +
T

W 4+

(

':5+

"4+ a 5+

'

respectively,

shew that

J/= OjP + 5,
12.

iV

= (a^ + 1) P + aJL

If


pn

is

the n th convergent to
j.
i

i_

a+ «+ a+ shew that
expansions of

"

'

and q n are respectively the

coefficients of

xn

in the

#
1

.

— ax — x 2

and
1

«.#

+ x2
'

— «-r — x 2
>

Hence shew that
equation
t

a pn —<In-i =

n

_ Qn
i

where

a, /3

are the roots of the

2

-

at

-

1

= 0.

13.

If


9n

is

the n th convergent to
_l
1
1

1_
b-t-

a+

b-\-

a+

"

'

shew that

pn

and q n are respectively the
.

coefficients of

xn

in the

expansions of

1

x + bx2 — ^ ax+(ab + l)x2 — xA and - (ab + 2) x2 + x* 1 - (aft + 2) x 2 + xA

'

Hence shew that
op, n = bq 2n _ x = ab

a

,

where

a, /3

are the values of

x2 found from the equation

l-(ab + 2)x2 + xA = 0.

CHAPTER XXVI.
INDETERMINATE EQUATIONS OF THE FIRST DEGREE.
345. In Chap. X. we have shewn how to obtain the positive integral solutions of indeterminate equations with numerical coefficients; we shall now apply the properties of continued fractions

to obtain the general solution of the first degree.
346.

any indeterminate equation

of

equation of the first degree involving two unknowns x and y can be reduced to the form ax±by = ± c, where a, 6, c are positive integers. This equation admits of an unlimited number of solutions but if the conditions of the problem require x and y to be positive integers, the number of solutions may be
;

Any

limited.

It is clear that the equation ax + by = — c has no positive integral solution ; and that the equation ax — by = — c is equivalent

— ax — ax ±by — c.
to by
If a

c) hence it will be sufficient to consider the equations

have a factor m which does not divide c, neither of the equations ax±by = c can be satisfied by integral values of x and y for ax ± by is divisible by m, whereas c is not.

and

b

;

If a,

b, c

have a common factor
shall suppose a,
b,

so that

we

c

can be removed by division; to have no common factor, and
it

that a and
347. equation

b

are prime to each other.
the general solution

To find

in positive integers of the

ax — by —

c.

Let - be converted into a continued fraction, and
6

let

— denote
q

the convergent just preceding j

;

then

aq—bp = ±l.

[Art. 338.]

;

.

INDETERMINATE EQUATIONS OF THE FIRST DEGREE.
I.

285

If aq

— bj) —

1,

the given equation

may
;

l»e

written

ax — by — c (aq —
.-.

b]j)

a(x —
b

cq)

-b

(y

c/>).

Now
divisible

since a

by

b

;

have no common factor, x — cq must be hence x — cq = bt, where t is an integer,

and

x- cq
b

y — cP.
a

that

is,

x=

bt

+

cq,

y—

at

+

cj)

\

from which positive integral solutions
to
t

be obtained by giving any positive integral value, or any negative integral value
en
-j,

may

numerically smaller than the less of the two quantities
also
t

cd

\

may
If

be zero; thus the number of solutions

is

unlimited.

II.

aq — bp — —

1,

we have
c

ax — by — —
.'.

(aq


+

bji)
',

;

a(x +
x + cq =—

cq)

=

b (y

cj))

• . .

-

+ cp — = t, an = y
a

integer

o

lience

x=

bt

— cq, y — at —

cp;

from which positive integral solutions may be obtained by giving to t any positive integral value which exceeds the greater of the

two quantities -=-,—; thus the number
o

CO

CD

of solutions

is

unlimited.

a

III.

If either a or b is unity, the fraction

j-

cannot be con-

verted into a continued fraction with unit numerators, and the investigation fails. In these cases, however, the solutions may be written down by inspection; thus if 6 = 1, the equation becomes ax — y = c; whence y = ax—c, and the solutions may be found by
ascribing to x any positive integral value greater than Note. It should be observed that the two arithmetical progressions in which the
respectively.
series of values for x

a
and y form are b and a

common

differences

;

286
Example.

HIGHER ALGEBRA.
Find the general solution in positive integers of
29.r

-

42*/

= 5.

In converting
13
is -jr
;

— into a continued fraction

the convergent just before

we have

therefore

.-.

29xl3-42x9 = -l; 29x65-42x45 = - 5;
29 (* + 65)

combining this with the given equation, we obtain

=42(# + 45);
u

x + 65
•*•

-£j- =

+ 45
29"

= *»

an mte 8 er

5

hence the general solution

is
a:

= 42«-65,

ij

= 20t~4o.
the equation

Given one solution in positive integers of ax — by = c, to jind the general solution.
348.

Let

h,

k be a solution of
.'.

ax-by = c;

then ah — bk =
;

c.

ax — by = ah - bk

.'.

a (x — h) — b(y — k);
.'.

x—h — — = y—k = a
z

t.

an integer

;

b

.'.

x=h+

bt,

y—k+

at

which

is

the general solution.

349.

To Jind

the general solution in positive integers

of

the

equation

ax + by =

c.

a Let t be converted into a continued
b

fraction,

and

let

— be the
q

D

convergent just preceding j
I.

;

then aq — bp = ±

1.

If aq

— bp=l, we have
ax +by =
.'.

c

(aq

— bp);
c2));

a(cq — x) = b(y +
cq

—x
b

.'

.

-=-=

——— — = y +acp = L an integer °
'

;

'

. ' .

x = cq — bt, y — at - cp

;

; ;

;;

INDETERMINATE EQUATIONS OF THE FIRST DEGREE.
from which positive integral solutions
to
t

287

may be
CI)

obtained by giving

positive integral values greater than
of solutions is limited,

and

less
is

than j-

CO

.

Thus the number

and

if

there

no integer

fulfilling these conditions there is

no solution.

II.

If aq

— bp = -

1,

we have

ax + by =
.-.

c

(aq

bp)

;

a(x + cq) = b(cp-y);
x + co

——
-

o

JL

en =— « y=

a

t. '

an integer 7
;

.

.

x=bt — cq,

y

= cj) —

at

from which positive integral solutions
to
t

may be

obtained by giving

co positive integral values greater than -~
before, the

and

less

than

cP

As
no

number
a or

of solutions

is

limited,

and there may be

solution.

III.

If either

b is equal to unity, the solution

may

be

found by inspection as in Art. 317.
Given one solution in positive integers of Ike equation ax + by = c, to find the general solution.
350.

Let

A,

k be a solution of ax
'

-f

by —

c

;

then ah + bk =

c.

.

.

ax + by — ah + bk
a (x — h) — b (k y)

.'.

x
.'.

7— =- —-— ——hk—y o a
bt,

t,

an integer

;

.'.

x=h +

y-k—

at

which

is

the general solution.

351.

To find

the

equation

number of solutions in positive ax + by = c.

integers

of the

Let T be converted into a continued fraction, and
b

let

- be the
q

convergent just preceding j

;

then aq — bp

= at 1.

.

288
I.

HIGHER ALGEBRA.
Let aq -bp

=

l

;

then the general solution
ep.

is

x = cq-bt, y = at-

[Art. 349.]
t

Positive integral solutions will be obtained by giving to
positive integral

values

not

greater

than °f
o

,

and

not

less

than

— a
Suppose that - and T are not integers. °
c c

(i)

a

b

Let

— = m+f. J
a


b

=n+

a. y

'

where m, n are positive integers and J\ g proper fractions then the least value t can have is m+ 1, and the greatest value is n;
;

therefore the

number

of solutions is

n-vi = -±b

cq

cp

— +f- J c J J g=-—+f-g. J
.
j.

a

ab

Now
—r- a

this is

an

integer,

and may be written

— + a fraction,
ab

or

fraction, according as /is greater or less

than

g.

Thus the

number

of solutions

is

the integer nearest to

,

greater or less

according
(ii)

as/ or g

is

the greater.
z-

Suppose that
this case

is

an

integer.
of

In
this,

g-

0,

and one value
c

x

is zero.

If

we

include
in-

the

number
C

of solutions is

ao
teger.
C

-r+f, which must be an
is

Hence the number
or -j
,

of solutions

the greatest integer in

^7+1
(iii)

according as
c

we
an

include or exclude the zero solution.

Suppose that cc

is

integer.

In
this,

case/=0, and one value of y is zero. If we include the least value of t is m and the greatest is n; hence
this of solutions is
71

the

number

—m+

l.

or —r

-

ab

q+

1.

Thus the

INDETERMINATE EQUATIONS OF
number

TJIE FIRST
c

DEGREE.

289
c

of solutions is the greatest integer in -7

cording as

we

ab the zero solution. include or exclude
c -

+1

or —=. ae-

ab

(iv)

Suppose that

and

c

7 are

both integers.

In this case
value.

and y = 0, and both x and y have a zero If we include these, the least value t can have is m, and

f—
;

the greatest

is

n

hence the number of solutions
the zero values the

is

11-111+

1,

or

-y + ab

1.

If

we exclude

number

of solutions is

ab

4-i.
II.

If

aq

-bp= -

1,

the general solution
bt

is

x=

cq

}

y— cp — at,

and similar

results will be obtained.

352. To find the solutions in positive integers of the equation ax + by + cz — d, we may proceed as follows.

transposition ax + by = d — cz ; from which by giving to z we obtain equations of in succession the values 0, 1, 2, 3, the form ax + by = c, which may be solved as already explained.

By

we have two simultaneous equations ax + by + cz=d, ax + b'y + cz = d\ by eliminating one of the unknowns, z say, we obtain an equation of the form Ax + By = C. Suppose that x —f, y — g is a solution,
353.
If

then the general solution can be written

x=f+Bs, y = g-As,
where
s is

an

integer.

Substituting these values of x and y in either of the given equations, we obtain an equation of the form Fs + Gz = II, of which the general solution is
8

=h +
we

Gt,

z

= k - Ft

say.

Substituting for

s,

obtain

x=f+Bh + BGt,
and the values
integral values.
H. H. A.

y = g-Ah-AGt; by giving to
t

of x, y, z are obtained

suitable

19

:

290
354.

HIGHER ALGEBRA.
If one solution in positive integers of the equations

ax + by +

cz

=

d,

ax + b'y + c'z = d',
may
be obtained as follows.
;

can be found, the general solution
Lety,
g,

h be the particular solution

then

af+ bg +

ch

=

d,

a'f+

b'g

+ ch = d'.

By

subtraction,

a(x-f) + b(y-g)+c(z- h) = 0, a'(x-/) + b'(y-g) + c'(z-h) = 0;
whence

1

j

x-f
be

— b'c

=

y-g _ z-h _
ca

t

— c'a

ab'

— a'b

k

'

where
be

t


x

b'c,

an integer and k is the H.C.F. of the denominators Thus the general solution is ca — c'a, ab' — a'b.
is (be'

=f+

b'c)

j
/c

,

y—g+

(ca'

c'a)

=• ,

z

=h+

(ab'

- a'b) ?.
fc

fc

EXAMPLES. XXVI.
j

Find the general solution and the
1.

least positive integral solution, of
3.

775.r-711y =
In how
?

l.

2.

455#-519y=l.
,£1. 19s. 6d.

436#-393y = 5.

4.

many ways

can

be paid in florins and half-

crowns
5.

Find the number of solutions in positive integers of

lLe+15y=1031.
6.

Find two fractions having 7 and 9

for their denominators,

and

such that their
7.

sum

is 1

£-§-.

Find two proper fractions in their lowest terms having 12
for their

and 8
8.

denominators and such that their difference

is

— 24

.

A

of

y pounds x

certain sum consists of x shillings ; find the sum.

pounds y

shillings,

and

it is

half

Solve in positive integers
9.

21 6#+ty + 4s=122\ lhr + 8y- 6^=145 5J

10.
'

1 2x 12.r-lly 1 \y

-

4z = 221 + 4^=22

-4.v+ 5y+ z=ll)

INDETERMINATE EQUATIONS OF THE FIRST DEGREE.
11.

291

20^-21^=381 3y+ 4s =34/
7.r

12.
'

13^ + 1 Is =103)
7z

- by=

"

4J

13.
15.

+ 4y + 19^ = 84.
3, 2,

14.

23.r+17.y +

lU = 130.

Find the general form of
8 leave remainders

all positive

integers which divided

by

5, 7,

5 respectively.
3, 7, 11

16.

Find the two smallest integers which divided by
1,

leave

remainders
17.

6, 5 respectively.

of three digits in the septenary scale is represented in the nonary scale by the same three digits in reverse order if the middle digit in each case is zero, find the value of the number in the
;

A

number

denary

scale.

If the integers 6, «, b are in 18. possible values of a and b.
19.

harmonic progression, find

all

the

rods of equal length are divided into 250 and 243 equal parts respectively if their ends be coincident, find the divisions which are the nearest together.
;

Two

to toll at the same time, and tolled at intervals of 23, 29, 34 seconds respectively. The second and third bells tolled 39 and 40 seconds respectively longer than the first ; how many times did each bell toll if they all ceased in less than 20 minutes?
20.

Three

bells

commenced

21.
7.r

+ 9y = c may
22.

Find the greatest value of c in order that the equation have exactly six solutions in positive integers.

Find the greatest value of c in order that the equation 14r + lly=c may have exactly five solutions in positive integers. Find the limits within which c must lie in order that the equation 19x + 14y = c may have six solutions, zero solutions being
23.

excluded.
greatest value of c in order that the equation ax + by = c may have exactly n solutions in positive integers is (n + l)ab-a-b, and that the least value of c is (n-l)ab + a + b } zero solutions being excluded.
24.

Shew that the

in

o

CHAPTER

XXVII.

RECURRING CONTINUED FRACTIONS.
have seen in Chap. XXV. that a terminating continued fraction with rational quotients can be reduced to an ordinary fraction with integral numerator and denominator, and but we shall prove that a therefore cannot be equal to a surd quadratic surd can be expressed as an infinite continued fraction whose quotients recur. We shall first consider a numerical
355.
;

We

example.
Example. Express ^19 as a continued fractions approximating to its value.
x/19 = 4 v/19
3
N /19
fraction,

and find a

series of

+ ( v/19-4) = 4+ Tl9 - -; v +
,x/19 z _2_ +
3

3

+ 4_ 2

5

,
'

V19 + 2
.

5

+ 2_ 1j ^19-3 = , 1 +

=1+^—
_

2

5

\/19 + 3'
5
.

,/19 + S 2

^£9-8
2

\/19 + 3'

v/19

+ 3_ 1

1

+V
1

1

9-2_ 1
1

5

~ + N/
,

3

/L9 + 2

n iN/19-4
2

+

3
(

~

0i

1

\/19 + 4'

N /19

+4=8+

N/19-4)

=8+
hence
1

after this the quotients 2, 1, 3, 1, 2, 8 recur;
1

V 19 - 4 + 2+ 1+ 3+ 1+ 2+8+
It will

1

Jl_£

2. •••

be noticed that the quotients recur as soon as we come to a quotient which is double of the first. In Art. 361 we shall prove that this is always the case.

»

,

RECURRING CONTINUED FRACTIONS,

293

[Explanation. In each of the lines above we perform the same series of operations. For example, consider the second line we first find the
:

greatest integer in is

—-—
o

;

this is 2,

and the remainder

is

-


6

2,

that

^— ^

o

.

We

then multiply numerator and denominator by the surd
so that after inverting the result
5
.

conjugate to

^19-2,

,

we begin a

new

line with a rational denominator.]
first

The

seven convergents formed as explained in Art. 336 are
4
1
'

9 2
'

13
3
'

48
11
'

61

170
'

1421
'
'

14

39

326
,
'

The
less

eiTor in taking the last of these

is less

than

_

,

and

is

therefore

than

.

-

,

or

,

and

a fortiori

less

than -00001.

Thus the

seventh convergent gives the value to at least four places of decimals.

356.
roots

Every periodic continued fraction is equal to one of the of a quadratic equation of which the coefficients are rational.
fraction,

Let x denote the continued and suppose that

and y the periodic

part,

x=

a+

1
z

1
,

b

+

c

+

:

:

294 The equation
s'y
2

HIGHER ALGEBRA.

+ {s — r) y — r = 0, which gives the value of if the positive value of y, has its roots real and of opposite signs v'y + p — on rationalising the denominator be substituted in x = -,
;
ii

qy+q

,

the value of x

is of

the form

~G

,

where A, B,

C are integers,

B

being positive since the value of y
Example.
Express
l

is real.

,1111 — — ~—
+s
5

^-—

...

as a surd.

Let x be the value of the continued fraction

;

then x - 1 = =

£>

— O— + +
1

1

\X


1)

;

whence 2x + 2x 2

The
is

7 = 0. continued fraction is equal to the positive root of this equation,

and

therefore equal to

^—-

.

EXAMPLES. XXVII.

a.

Express the following surds as continued fractions, and hnd the sixth convergent to each
1. 5. 9.

v/3.
v/11.

2.
6.

^5.
x/13.

3. 7.

y/6.

4.
8.

s/8.

x/14.

V22.
4 N/10.

2^3.

10.

4 v/2.

11.

3^5.

12.

13

-

j&-

14
-

15

V33-

-

\/s-

16 -

\/n-

17.

Find limits of the error when
T

268
65

—— is
is

taken for N /17.

18.
19.

916 Find limits of the error w hen '—-

taken for
is

v/23.

Find the

first

convergent to N /101 that convergent to

correct to five places

of decimals.
20.

Find the

first

VI 5

that

is

correct to five places

of decimals.

Express as a continued fraction the positive root of each of the following equations
21.
24. 25.

x* +

2x-l = 0.

22.

a8 -4*?- 3=0.

23.

la?- 8x- 3=0.

Express each root of x2 - 5^ + 3 =

Find the value of 3 + 5— x— x--

Ill
-

as a continued fraction.

6+ 6+

6+

26.

Find the value of

,

1+ 3+ 1+

3+

RECURRING CONTINUED FRACTIONS.
27.

295

Find the value of

3+

28.
29.

Find the value of 5 +

111111 1+ 3+ 1+ + 1111 1+ 1+
2

+

2

3

+

,

1+

10 +

Shew that
*+

i+6+ i+ e+""~*\ 1+ a+ 2+ 3+ 2+
infinite

;

30.

1+
*357.

111111 3+ 3+ 5+
5+ 1+

Find the difference between the
•"'

111111 3+ 3+ 5+
1+ 5+

continued fractions

1+

""

To convert a quadratic surd into a continued fraction.

Let

and

let

a positive integer which is not an exact square, a be the greatest integer contained in then
x

N be

JN
=

j

N/iV = «,

+ (Jff- a,) =

«,

+

-j£—

,

if r,

W- »,\
'

Let

b

be the greatest integer contained in —
r
i

;

then

JM+a
where

l

=

b

JN-b r
x
|

x

+a
x

^

h

JN-a
,

2

^h +
.

«2 = b r
i

1

— a and
x

r r2
x

= N — a„ 2

Similarly
2

r2

»*

-

2

JN + a./

where

«3 = bf2 — as and r 2 r 3 — on
;

N— a
.
,

2

3

;

and

so

and generally

JN+as=i = -

,

.

b M-l
,

+

JN-a v —
'

"

= b "-' +
=

r

H— 1

(1-1

jy + a
,

"

'
it

>

where

an =

&„_,/•„_,

-

a„_

Hence
and thus
"We

*JN= a, +

1111
1

and

?•„_,*•„

N - a/.

JN can be expressed as an infinite continued fraction.
this fraction consists of
re-

shall presently prove that

curring periods ; it is evident that the period will begin whenever any complete quotient is first repeated.

;

296

HIGHER ALGEBRA.
shall call the series of quotients

We

JA
the
first,

T
.

JAr + a,
r
x

JN + a
r2

2

JJST+ a 3
r3
)

second, third, fourth

complete quotients.

the preceding article it appears that the quantities a v rv b v b b are positive integers; we shall now prove 3 that the quantities a2 a 3 a4 are also positive inr8 r 3 r 4
, , , ,
,

*358.

From

,

,

, .

.

.

tegers.

p p Let — , —..
q

p
—r.

q

q

be three consecutive convergents to JN. and ° x
corresponding to the partial quotient b n
is
.

P" let

— be the convergent

The complete quotient

at this stage

;

hence

v^=
we have

r

p +p
,t

=P

JW+a„P+r p
v

Clearing of fractions and
parts,

equating rational and irrational
ck<l

«y + rnP = ^Y>

+ rn q =p

;

whence a n ( pq - pq) =pp* ~ <Z<7 '^j rn {ptf —p<i) = ATq' 2 —p' 2 But pq' —p'qssdslf and pq —pq-, pp' —qq'N, Nq 2 — p 2 have the same sign [Art. 344] hence a n and rn are positive integers.
.

;

Since two convergents precede the complete quotient
this investigation holds for all values of

-

r*

n greater than

1.

*359.

To prove

that the complete

and partial
l

quotients recur.
.

In Art. 357 we have proved that rn r n _ = N—a 2 Also r n and r n _ are positive integers ; hence a n must be less than ^/JV, thus an cannot be greater than a v and therefore it cannot have any
l

values except 1, 2, an cannot exceed a
x

3, ...a x }

'

that

is,

the

number of different values of

.

Again, a n+1 =rv b u -a that is rn b n = a n + a n+v and therefore r n b n cannot be greater than 2a also b n is a positive integer hence rn cannot be greater than 2a v Thus rn cannot have any values except 1, 2, 3,...2a ; that is, the number of different values ofi\ cannot exceed 2a r
h
, l

;

1

1

1

1

.

' ;

RECURRING CONTINUED FRACTIONS.
Thus the complete quotient
2a*
different values
;

297


rn

cannot have more than

that

therefore all subsequent ones,

some one complete quotient, and must recur.
is,

Also
quotients

b n is the greatest integer in


rn

;

hence

the partial

must

also recur,

and

the

number of partial
2
']

quotients in

each cycle cannot be greater than 2a

*3G0.

To prove

that a,
«,_,

< au +

rn

.

We

have

+ an = bH _ rn _
1

l

;

«»_i

+ a «= or >?, ,-i
t

5

since 6„_ l

is

a posit ive integer

;

But

N"-a;=rnrn _
a ~ a n < rn
i

-

l i

,

which proves the proposition.
*361. To shew that the period begins loith the second partial quotient and terminates ivith a partial quotient double of the first.

seen in Art. 359, a recurrence must take place, let us suppose that the (n+ l) th complete quotient recurs at the (*+ l) th ; then
Since, as
a.

we have

=a

,

r,

=r

,

and

b.

=

b

;

we

shall

prove that
a, — 1 4
.

=a n —

,

1'

.

rm »—

,

=r

ii

, — 1'
,

b, *—

=

b —
ii

,

We

have
r. — *
,
i

r* =

N -a,

2
'

»

iV

— a = r H — ,r —r
2
it
i

ii

H

. l

r, *

v

=r

Again,

a
.-.
"

,
'

_a
n-1

*

~

1

-b

7

,

— om = zero,
7 .

or an integer.

1

298

HIGHER ALGEBRA.
l

But, by Art. 360, a
a.

-a

n_

l

<.r _
ii

,
x

and a
.

l

-a _ <r _
s
l

s

i

;

that

is

-a.

,

<r

' ,

:

therefore a

,

- aa

,

<r
zero.
fl

;

hence

-^
n-

is less

than unity, and therefore must be

Thus

«,_!
if

= «„_!, and

also 6#_1
th

=6

_i.

th the (n + l) complete quotient recurs, the ?^ comth plete quotient must also recur; therefore the (n- l) complete quotient must also recur; and so on.

Hence

This proof holds as long as n is not less than 2 [Art. 358], hence the complete quotients recur, beginning with the second
It follows therefore that the recurrence rx begins with the second partial quotient b ; we shall now shew that it terminates with a partial quotient 2a

quotient 1

-

.

x

x

.

Let

-

"be the complete quotient which

just precedes the

second complete quotient 1

when

it

recurs

;

then

——
H

a

ancj

v
»",

l

ar e

two consecutive complete quotients

;

therefore

but

N- a* =
Again, a
y

r,

;

hence rn =
?'„,

1.

— aH <

that

is

<

1

;

hence a - an x

0,

that

is

««

= «,•
bn

Also an + a
proposition.

= rn bn — bn

;

hence

= 2a

i

;

which establishes the

*362.
distant

To shew
the

from

any period the partial quotients equibeginning and end are equal, the last partial
that in

quotient being excluded.

Let the

last

complete quotient be denoted by * rn =l,
ci n


r.

;

then

=a
x

,

bn

=2a r

We shall prove that
^- 2 =^ 2)
«h-2=«3
>

^.-2= & 2^

+

.

;

RECURRING CONTINUED FRACTIONS.

299

We

have
r«-x

=

r*

rH_t

=-N- a,;
+ a H = rm_,
?•,&,;

N - a; &„_,

r,

Also
»„_,

+a =
x

«„_,

=

r, &„_,

and

«,+«., =

.

.

-=i

-

= 0, —
is

o M-1

= zero,
1 ,

or an integer,

But
unity
;

"-' I

^

<

CT
i

~ a "~

l ,

that

<

a '~ a ',

which

is less

than

thus a 2 - an_x =
rn _ 2

hence
,

a,,.!

= « 2 and
,

o,,^

=

b
l

.

Similarly

= r2

«„_ 2

= «3

6„_ 2

-

b2

;

and

so on.

*363. From the results of Arts. 3G1, 362, it appears that when a quadratic surd v/iVr is converted into a continued fraction, it must take the following form

1

J_ J_ J_ + & +6 + &i
3

J_ J_ J
°3

1_

3

+

&2

+

°i

+2a,+
of
the recurring

*364.
periods.

To obtain

the penultimate convergents

Let n be the number of partial quotients in the recurring period then the penultimate convergents of the recurring periods are the ?iih 2n th 3n th convergents ; let these be denoted by
;
, ,
,

V -\ ^=,
xt

^,
b3

respectively.

Now

/v JiV = a + v
l

b

i

111 ——— + + +
j
i

b2

b _
7l

11 —+ — +
-

2a
7)

l

l

so that the partial quotient corresponding to
Pn+X

+1

is

2a
t

;

hence

SWl

= ^Pn+Pn-l " 2 «1 9n + ?n-l

'

Tlie complete quotient at the

same stage
:

consists of the period

2«,+T
bi

-r

,

+

K

+

6 «-i

300
and
is

HIGHER ALGEBRA,
therefore equal to a +
x

J'N

;

hence

Clearing of fractions and equating rational and irrational parts, we obtain

*iP.+JV-i = -jfy«i a i9 H + 9 n-i=^«
,

r

(!)•

Again
quotient

— can be obtained from — and
2* 1

-^

by taking

for the

+
.

1

1

V?V^
Thus

C
1

which

is

equal to

rtj

+


in

&=1
**

2^

=

U+%)qn + q^
.

& P« + %.qn

,

from

(1);

l(A + *&)
if

(2)>

?2 „

In

like

manner we may prove that
c th

-

is

the penultimate

icn

convergent in the

recurring period,

«i ^c«

+Fcn-i

= Nqmi

a, q cn

+ qm_l =#*,
,

and by using these equations, we may obtain £—
cessively.

,

suc-

It should be noticed that equation (2) holds for all multiples
of

n

;

thus

Ol

the proof being similar to that already given.

In Art. 356, we have seen that a periodic continued fraction can be expressed as the root of a quadratic equation with rational coefficients.
*365.

y

RECURRING CONTINUED FRACTIONS.
Conversely,

301

we might prove by

the method of Art. 357 that
>

an expression of the form
integers,

tt~~

where

-<4>

B,

C

are positive

not a perfect square, can be converted into a recurring continued fraction. In this case the periodic part will not usually begin with the second partial quotient, nor will the last partial quotient be double the first.

and

B

For further information on the subject
fractions

of recurring continued

Cours cVAlgebre Superieure, and to a pamphlet on The Expression of a Quadratic Surd as a Continued Fraction, by Thomas Muir, M.A., F.R.S.E.

we

refer the student to Serret's

^EXAMPLES. XXVII.
:

b.

Express the following surds as continued fractious, and find the fourth convergent to each
1.

N/a

2

+ l.

2.

Ja* -a.

3.

N /«--l.

4.

V/T7T.
Prove that

5

.

y«"^f
—— -i

.

6

.

^l
r-

7.

J9a* + 3 = 3a+
and
find the fifth convergent.
8.

2a + 6a + 2a + 6a +

Shew

that
2

p
9.

i+

p+

1111 i+
p
\ /
Ct

i—»

V
10.

/111
Shew
that

a \-\

\
If

P9

Cl

2

+
1

tt

3 + P9 4 +

=P a + aCL
\

111—

l

+

PaZ +

OCl \

+

Ja'2 +

be expressed as a continued fraction, shew that

2(a*+l)q n =p n _ 1 +p n + 1
11 11.

Tf If

.%'=

a x + « 2 + a i+ a 2 +

1111
1

,

2p n = q n _ l + q n + l
...,

.

?/

~2a +
~3tf 1

1111 1111
+~3a 2 +
)

2« 2 + 2a x + 2a 2 +

""'

shew that

3^ + 3« 2 + x {f- - z-) + 2y (z 2 - .r2 + 3z {a? -

'"'
1
)

= 0.

302
12.

HIGHER ALGEBRA.
Prove that

(
13.

b

JL J_ Jl J_ ^ =+ a+ b + a+ '")\b+ a+ b+ a+ '")"b
X ~ a+

V————
••'
•••'
2

U

J_ J_ J_ b+ b+ a+ a+

1

shew that
14.

(ab 2

J_ J_ J_ J_ y- b + a+ a+ i+ b+ + a + b)x- (a b + a + b)y = a -b2
2 th

.

If

— be the n
P2
2

convergent to

Ja 2 + l.

shew that

+P32 +'"+P 2n + l = Pn + lPn + 2-PlP2

15.

Shew that

(— — \a+ + a+
v

b

c+

'

c

+

1

1

1

\ /

l

+ bc

"/ \
ih

^+

a+

c

+
2

l+ctb'

16.

If

— denote the r
qr

convergent to ^—— &


i

,

shew that
1

Pi+Pi>+ >~+P°.n-l=P2n-p<L,
17.

? 3 + ?5 + infinite
i

+ &» - = ?8» ~ ft.
continued fractions
'••'

Prove that the difference of the

_i_j_2_
a+ b+ c+
is
•'

i

b+ a+ c+

equal to
1

+ ao

=

.

18.

number

If s/JV is converted into a continued fraction, of quotients in the period, shew that

and

if

n

is

the

De converted into a continued fraction, and if the penultimate convergents in the first, second, ...kth recurring periods be denoted by n lt n2i ...nk respectively, shew that
19.

If

\/^

*CHAPTER

XXVIII.

INDETERMINATE EQUATIONS OF THE SECOND DEGREE.
*366. The solution in positive integers of indeterminate equations of a degree higher than the first, though not of much practical importance, is interesting because of its connection with In the present chapter we shall confine the Theory of Numbers. our attention to equations of the second degree involving two
variables.

obtain the positive integral values of x arid y which satisfy the equation
to

*367.

To shew Iww

ax 2 + 2hxy + by 2 + 2gx + 2fy +
a, b, c, f, g,

c

=

0,

h being

integers.
x,

Solving this equation as a quadratic in

as in Art. 127,

we

have

ax +

hy+g = ±J(h

2

-ab) y 2 +

2 (hg

- af)y+(g 2 -ac)...(l).

x and y may be positive integers, the expression under the radical, which we may denote kv py 2 + 2gy + r, must be a perfect square that is
in order that the values of
;

Now

2 py + 2qy + r =

z

2
,

suppose.

Solving this equation as a quadratic in y,
2 2 py + q = ± Jq -jjr+pz
;

we have

and, as before, the expression under the radical square ; suppose that it is equal to t 2 ; then
2

must be a perfect

t

- pz 2 = q 2 - pr^
and
j\ q, r are constants.

where

t

and

z are variables,

304

HIGHER ALGEBRA.

Unless this equation can be solved in positive integers, the original equation does not admit of a positive integral solution.

We

shall return to this point in Art. 374.
If
a,
b,

h are

all

positive, it is clear that the

number

of

limited, because for large values of x and y the sign 2 2 of the expression on the left depends upon that of ax + 2hxy + by [Art. 2G9], and thus cannot be zero for large positive integral

solutions

is

values of x and

y.

Again,
negative, solutions
is

if

h*

and

negative, the coefficient of y in (1) is by similar reasoning we see that the number of

ab

2

is

limited.
Solve in positive integers the equation
a;

Example.

2

- 4xy + &y* - 2x -

20*/

= 29.

Solving as a quadratic in x, we have

x = 2y + 1 ± ^30 + 24// - 2y\

But 30 + 24?/ - 2j/ 2 = 102 - 2 (y - G) 2 hence (y - 6) 2 cannot be greater than By trial we find that the expression under the radical becomes a 51. 8 perfect square when (y-6) =l or 49; thus the positive integral values of y
;

are

5, 7, 13.

When

?/

= 5,

x = 21

or

1;

when y = 7, x = 25

or

5;

when y = 13,

x = 29 or 25.

have seen that the solution in positive integers *3G8. of the equation
2 ax 2 + 2hxy + by + 2gx + 2fy +

We

c

=

can be made to depend upon the solution of an equation of the form

x2 ±
where
iV*

Ny = ± a,
2

and a are positive

integers.

The equation x 2 + Ny* = — a has no real roots, whilst the 2 2 equation x + Ny = a has a limited number of solutions, which

we shall therefore confine our attention be found by trial 2 2 to equations of the form x - Ny = ± a.

may

;

*369. To sJiew that the equation x solved in positive integers.

2

-Ny =l
2

can always

be

Let
2_

JN
9

be converted hito a continued fraction, and

let

l-

'— be any three consecutive convergents; suppose that

q

q

y

U

1

INDETERMINATE EQUATIONS OF THE SECOND DEGREE. 305
t

17"

a,/
is

,n
the complete quotient corresponding to
*. (/"/

~

V'q)

W
1 ]} (

;

then
[Art. 358].

-P"

2

But
/

r„

=

1

at the end of
..
]>

any period [Art.
"

.'3(51]

— JVq

—]

)(

2

)

,

being the penultimate convergent of any recurring period.
If the

number
P
;

of quotients in the period is even, -,

i.s

an even

convergent, and
greater than

is

therefore

greater

than

v/iV,

and therefore

thus

pq —pq = 1.
is

Jn

this case p'*

— N"q' a = J,
=
1.

and therefore x=]>\ y = q
Since

a solution of the equation xr — Ny*

p —

is

the

penultimate convergent of any recurring
is

period, the
If the

number

of solutions

unlimited.
is

number

of quotients in the period

odd, the penultima

1

convergent in the tirst period is an odd convergent, but the penultimate convergent in the second period is an even convergent. Thus integral solutions will be obtained by putting x=p', y — q\

where
sixth,


q

is

the penultimate convergent in the second, fourth,

recurring periods. of solutions is unlimited.

Hence

also in this case the

number

*370.

To obtain a

solution, in positive inte<iers

of the equation

As

in the preceding article,

~KT

we have
'
I

v -Jq =pq-pq>
'2

If

the

number

of quotients in the period is odd,

and
>

if
<1

is

an odd penultimate convergent in any recurring period, -,-<-,

and therefore

In this equation x 2 —

pq —pq — — 1. case p' — Nq '

2

2

X

1

=—

1

will

-\, and integral solutions of the be obtained by putting x =p\ y — q
',

is the penultimate convergent in the first, third, q recurring periods. nappeu i.H.10 we can discover ,, „
11.

where

fifth...

,

n. A.

irv

2fc

1:0

306
Example.

HIGHER ALGEBRA.
Solve in positive integers x 2 - 13y 2 = ±1.

We

can shew that

^13-3 +
Here the number
vergent in the
first

1+

11111 1+ITl+
hence
a;

6+
odd
is
;

of quotients in the period is

the penultimate con-

18
period
is -=;

= 18,

y=o

a solution of

x 2 -13y 2 =-l.

By

Art. 364, the penultimate convergent in the second recurring period is
1

2U
hence # = 649, y = 180
is

/ 18

'180' J' a solution of x 2 - 13y 2 =l.
18

+

5

io\ Xl3

649 ,u * thatlS

By forming
periods

we can obtain any number
x2 13?/
2

the successive penultimate convei'gents of the recurring of solutions of the equations

=

-

1,

and x 2 - lSy 2 = + 1.

one solution in positive integers of x 2 — Nif = 1 lias been found, we may obtain as many as w e please by th following method.
*371.

When

r

Suppose that x = h, y = k is a solution, h and k being positive n 2 integers; then (A — Nk 2 ) = 1 where n is any positive integer.
,

Thus
.

x2
•.

-Ni/= (h* - m?y
- yJJST) =
1

.

(x

+ yjN)

(x

(h

+

kJN)
=

n

(h (h

- k s !X)\
- kJN)"
;

Put x +

yJN = (h + kJX)' x-y JN
,

kJN) + (h-kJJYy; 2 Us in = (h + kjiry - (h - kjNy.
.-.

2x = (h +

n

Tlie values of

x and y so found are positive integers, and by
1, 2, 3,...,

ascribing to n the values can be obtained.

as

many

solutions as

we

please

Similarly

if

x —
2

Xy = —
2

1,

x=h and if n

i

y = k is a solution of the is any odd positive integer,

equation

x*

- Ntf =
1, 3,

(h

2

- Nk 2 )\

Thus the values

of

x and y are the same as already found, but
5,
2 2

n

is

restricted to the values

By putting x = ax', y = ay the equations x — Ny — ± a 3 become of 2 — IFyf* = d= 1, which we have already shewn how to
*372.
solve.

,

INDETERMINATE EQUATIONS OF THE SECOND DEGREE
*373.

307

We
if

have seen in Art. 3G9 that n P ~ N(f~ = ~ rn ( V<1 ~ V<l) = * r..
is

Hence

a

a denominator of any complete quotient which

occurs in converting

JX
i

into a continued fraction,

and

if


q

is

convergent obtained by stopping short of this complete 2 quotient, one of the equations x~ — Xy — ±a is satisfied by the
the
values

x =p\ y = q'all less

Again, the odd convergents are

than
if

JN, and

<2
,

the

even convergents are
convergent,
is

all

greater than
is

JX

;

hence

is

an even
t

x=p, y = q

a solution of x*
,

-Xif — a: and
'

if —.

an odd convergent, x
*374.
Tlie

=p

y—q

is

a solution of Xs

— Xy = — a.
2

q

method explained

in the preceding article enables

us to find a solution of one of the equations x 2 — Xy 2 — only when a is one of the denominators which occurs in the process of into a continued fraction. converting For example, if Ave convert J7 into a continued fraction, we shall find that

±a

JX

and that the denominators

of the complete quotients are 3, 2,

3, 1.

The

successive convergents are
2

3

5 2'

8

37

45
17'

1'

1'

3'

IT
2
CI

82 31'

127

18

'

;

and

if

we take the
I"

cycle of equations
ar
2

x 2 - ty 2 = —

O
3,

— Iff = 2,
*"

o

ar

f o — 7^ = — 3,
o

o

cc

- y2 =
h'
/

1

1

we
for

shall find that they are satisfied

by taking

x the values

2, 3, 5, 8,
1,
1,

37, 45, 82, 127,
2, 3,

and iovy the values
*375.

14, 17, 31, 48,

It thus appears that the

tions in integers of with certainty is very limited.

number of cases in which solu2 2 the equations x — Xy = ± a can be obtained
it

In a numerical example however, sometimes happen that we can discover by

may,
a

trial

20—2

308

HIGHER ALGEBRA.
2 2

positive integral solution of the equations x — Ny = =*= a, when a is not one of the above mentioned denominators ; thus we easily
find that the equation

When

by y=2, # = 9. one solution in integers has been found, any number of
is

# 2 -7?/ 2 = 53

satisfied

solutions

may be

obtained as explained in the next

article.

Suppose that x =f, y = g is a solution of the equation x _ Ny = a and let x = h, y - k be any solution of the equation x2 - JSfy 2 = 1 then
*376.
2
2
;

;

x*

~ Ny 2 = (f 2 - Kg ) (h 2 - Nk ) = (fh±Ngk) 2 -N{fk±gh)\
2 2

By

putting

x

-fh ± Kgk, y -fk ± gh,

and ascribing

we may

to h, k their values found as explained in Art. 371, obtain any number of solutions.

*377. Hitherto square ; if, however,

form x - n y =
2 2

2

a,

is not a perfect has been supposed that is a perfect square the equation takes the which may be readily solved as follows. it

N

N

Suppose that a = be, where b and then of which b is the greater
;

c

are

two

positive integers,

(x

+ ny) (x — ny) =
c
;

be.

the values of x and y found from these equations are integers we have obtained one solution of the equation ; the remaining solutions may be obtained by ascribing to b and c all their possible values.
b,

Put x + ny =

x - ny =

if

Example.
equal to 60.

Find two positive integers the difference

of

whose squares

is

Let

x,

y be the two integers
is

;

then

,xr

- y 2 = 60

;

that

is,

(as

+ y)

(x

-

y)

= 60.

Now

60

the product of any of the pair of factors

1,60;

2,30; 3,20;

4,15; 5,12; 6,10;

and the values required are obtained from the equations
ic

+ y = 30,
2;

# + y = 10,

x-y=
Thus the numbers

x-y=

6;

the other equations giving fractional values of x and y.
are 16, 14; or
8, 2.

;.

INDETERMINATE EQUATIONS OF THE SECOND DEGREE.
Cor.

309

In like manner we
ax'

may

obtain the solution in positive
*2/'y

integers of

+ 2hxy + by 1 + %jx +

+

c

= k,
two
rational linear

if

the left-hand

member can be

resolved into

factors.

*378. If in the general equation a, or b, or both, are zero, instead of employing the method explained in Art. 3G7 it is simpler to proceed as in the following example.
Example.
Solve in positive integers
2.ry

- 4a- 2 + V2x - 5y = 11

Expressing y in terms of
4a; 2

x,

we have
5
;

+ V= —^r-r—
In order that y

- 12* 11 „ 6 , =2tf-l+; 2x - 5 2x n

may be an
or

integer
2iX

—O

=

must be an integer

hence

2.r

-5

must be equal

± 2, or ± 3, or ± G. The cases ±2, ±6 may clearly be rejected; hence of x are obtained from 2x - 5 = ± 1, 2x - 5 = ± 3
to

± 1,

the admissible values

whence the values

of

.x-

are

3, 2, 4, 1.

Taking these values

in succession

we obtain the
?/

solutions

x = S, y = ll; s=2, y = -3; # = 4,

= 9;

ar=l,

y=

-1;

and therefore the admissible solutions are a; = 3, y = 11; x = 4, y = 9.
principles already explained enable us to discover for what values of the variables given linear or quadratic Problems of this functions of x and y become perfect squares. kind are sometimes called DiopJiantine Problems because they

*379.

The

were first investigated by the Greek mathematician Diophantus about the middle of the fourth century.
Example 1. Find the general expressions for two positive integers which are such that if their product is taken from the sum of their squares the difference is a perfect square.
Denote the integers by x and y
x.-.
;

then
;

-xy + y'2 = z'2 suppose
x(x-y) = z
2

-y-.

This equation

is satisfied

by the suppositions
n (x -y)

mx= n {z + y),
where

= m (z

-

y),

m

and n are

positive integers.

,

310
Hence

HIGHER ALGEBRA.
mx - ny - nz = 0,
these equations

nx + (m -n)y- mz - 0.

From

we
x

obtain by cross multiplication

_
2

2mn - ri
solution

m
is

2

V - n2

2
.

m

2

- mn + ri1

'

and since the given equation
x = 2mn-n 2
,

homogeneous we may take
2

for the general

y=m

-n 2

,

z

= m 2 -vin + n 2

.

Here m and n are any two positive integers, »i being the greater; thus ?n = 7, n = 4, we have x = ±0, y = SS, 3 = 37.

if

Find the general expression for three positive integers in arithmetic progression, and such that the sum of every two is a perfect

Example

2.

square.

Denote the integers by x-y,

x,
,

x + y; and
,

let
;

2x-y=p 2
then
or

2x = q 2 2x + y = r 2
,

2 2 2 p + r = 2q

i*-q*=q*-p>.
This equation
is satisfied

by the suppositions,
n
(r

m (r - q) = n (q - j>),
where

+ q) = m (q +p),

m

and n are

positive integers.

From

these equations

we obtain by
2

cross multiplication

V
w2 + 2mn-m
Hence we may take

=

<l

m +n
2

_
2

r

m + 2mn- n
2

'

2

for the general solution

p=n* + 2mn-m*,
whence

q

= m2 + n 2
,

,

r

= )u 2 + 2mn-u 2

;

x = = {m 2 + w 2 ) 2 y = inin (m 2 - w2)

and the three integers can be found.

From
odd
;

the value of also their values

x it is clear that m and n are either both even or both must be such that x is greater than y, that is,
(m 2 + n 2 ) 2 >8mn{m 2

-n 2 ),
;

or

mz (m - Sn) + 2inn
is satisfied if

2

+ 8m n* + n 4 >

which condition

m>Sn.

= 9, w=l, then a- = 3362, y =2880, and the numbers are 482, 33G2, If 6242. The sums of these taken in pairs are 3844, 6724, 9604, which are the squares of 62, 82, 98 respectively.

m

:

INDETERMINATE EQUATIONS OF THE SECOND DEGREE. 311

*EXAMPLES.
Solve in positive integers
1. 3.

XXVIII.

5.

= 77. y 2 -4.ry + 5.r2 -10.i- = 4. 3.y + 3.ry-4j/ = 14.
5a- 2 -10.iv/

+

7?/

2

2.

7^-2^+3y2 =27.
xy - 2.v - y = S.
4^ 2 -y 2 =315.

4.

6.

Find the smallest solution in positive integers of
7.

10.

-14y 2 =l. x2 - 61/ + 5 = 0.
.r

2

8.

^-19^=1.
11.

9.

2
.t-

= 4iy

2

-l.

x 2 -7y 2 -9 = 0.

Find the general solution
12.
.r

in positive integers of

2

-3/=l.

13.

x 2 -5y 2 =l.

14.

.v

2

- 17y 2 = -

1.

Find the general values of x and y which make each of the following
expressions a perfect square
15.
18.
:

x2 -3xy + 3y 2

.

16.

afi+2xy + 2f.

17.

5^+y2

.

Find two positive integers such that the square of one exceeds the square of the other by 105.
19.

Find a general formula

for three integers

which may be taken

to represent the lengths of the sides of a right-angled triangle.

Find a general formula to express two positive integers which are such that the result obtained by adding their product to the sum
20.

of their squares

is

a perfect square.

" There came three Dutchmen of acquaintance to see me, 21. being lately married they brought their wives with them. The men's
;

my

Claas, and Cornelius; the women's Geertruij, Catriin, and Anna but I forgot the name of each man's wife. They told me they had been at market to buy hogs each person bought as many hogs as they gave shillings for one hog; Hendriek bought 23 hogs more than Catriin; and Claas bought 11 more than Geertruij likewise, each man laid out 3 guineas more than his wife. I desire to know the name of each man's wife." (Miscellany of Mathematical Problems, 1743.)

names were Hendriek,
:

;

;

22.

Shew

square, if n is and k' the numerator of an even convergent to N ^2.

that the sum of the first n natural numbers is a perfect equal to k 2 or k' 2 - 1, where k is the numerator of an odd,

CHAPTER XXIX.
SUMMATION OF
Examples
summation
SERIES.

380.
in

of

of certain series

have occurred

previous chapters ; it will be convenient here to give a synopsis of the methods of summation which have already been explained.
(i)
(ii)

Arithmetical Progression, Chap. IV. Geometrical Progression, Chap. Y.
Series

(iii)

which are partly arithmetical and partly geoof the

metrical, Art. 60.
(iv)

Sums

powers of the Natural Numbers and

allied

Series, Arts.
(v)

68 to 75.

Summation
Recurring

by

means
Chap.

of

Undetermined

Coefficients,

Art. 312.
(vi)

Series,

XXIY.

proceed to discuss methods of greater generality ; but in the course of the present chapter it will be seen that some of the foregoing methods may still be usefully employed.

We

now

th If the r term of a series can be expressed as the dif381. ference of two quantities one of which is the same function of r that the other is of r - 1 the sum of the series may be readily
,

found.

For

let

the series be denoted by

and

its

sum by S
1 ;

,

and suppose that any term u r can be put in

the form v r -v r _

then
!

^.=(«i-0+(w.-«i)+(*.-«f) + »- + (w.-i-0 + (*.- ,,--i)
=v -v
n
.

SUMMATION OF
Example.

SERIES.

313

Sum
1

to n

terms the series
1
.
.

+ (l »„v„ n T + „ 3j-)(l + U 2u-)(l (l + s)(l+2s) + + 3.r) (l + + 4ar)
.

1

,

.

If

we denote the

series

by

*-±(—-—\ + + 3*/'
-

a;\l

2x

1

Ws

_!/
a;\l

1

1

\
'

+ 3#

1

+ 4*,/

x\l + nx
b}'

i+ n+ i.x/
l

addition,

SL=-

ar\l

I

^

+ a;
n

+w + l.a?/
..r)

I

(1

+ x)

(1

+n+ l

Sometimes a suitable transformation may be obtained by separating u into partial fractions by the methods explained
382.
in Chap.

XXIII.
Find the sum
of
1

Example.

(l

+ x)(l + ax)
th term= n ft*

+

(l

n,

+

x~n z~r + ax)(l+a*x)

a

a2
t-,

(1

+ a-x)

.,

>

,,

(1

+ a 3 x)

5—

v

+

.

.

.

to n terms.

nu, The


{l

« n_1

+
.'.

n, a n 1 x)(l
.

an

~x

—. — w + a — = l + a"- t- + l + anx n x) ^ -A (1 + a**) + B (1 + a '- *).
1

A

B

suppose:

**

7

1

By

putting 1

+

«" _1 .r, 1

+ a u x equal

to zero in succession,

we obtain

A=Hence
similarly,
u, 1

a n-l
1

-a'

,

B= —1 -a'
a
1

nn

=—
1

1/1
1

- a \1 + a:
/
,

+ ax J
2

....

a

t**=;

a — -5—

\
5
.

1

/

a"- 1
1

aw

\

Wn

~l-a Vl+a*"

*

l

+

an

j/

'"•

*~l-o\l + *

l

+ a n x)'

;

314
383.

HIGHER ALGEBRA.
To jind
the

of\\ terms of a series each term of which is composed of r factors in arithmetical progression, the first factors of the several terms being in the same arithmetical progression.

sum

Let the where
u„

series

be denoted by u x + u2 + us +

+ un
(a

,

-

(a

+ nb)

(a

+n+

1

.

b) (a

+n+2

.

b)

...

+n+r—

1

.

b).

Replacing n by
k «„ »-i
!

n—
.

1,

we have
+ nb) (a + n +
(a
1
.

=
.'.

(a

+n—
+
?i

1

&) (a
.

6)
.

.

..

(a

+n+r— 2

.

b)

1

b)

un =
1

+n+

/•

1

b)

?.«.„_!

=

vn

,

say.

Replacing n by n +

we have
+ r.
6)tf» =

(a + w
therefore,

«Il+1 j

by subtraction,

(r+l)b un = vn+ i-vn
.

.

Similarly,

(7+1)6. wB_, =

r

/(

-

/<„_,,


SUMMATION OF
Example.

SERIES.
series

315

Find the sum of n terms of the

1.3.5+3.5.7+5.7.9+
The n th term
is

{2n - 1) (2«

+

1) (2n

+ 3)

;

hence by the rule
,

- _ (2w-l)(2n + l)(2n + 3)(2n+5) n >\+C7.

^
C
;

To determine
we have

C, put n

=1
-j

;

then the series reduces to
:

its first

term, and

15

= ——
-

8

f-

whence C

=—
8

;

•'•

S *~

(2n-l)(2n+l)(2n+3)(2n+5)
8

+

15
8"

= n (2n3 + 8«2 + In - 2),
384.
also be

after reduction.

The sum

of the series in the preceding article

may

found either by the method of Undetermined Coefficients [Art. 312] or in the following manner.

We

have u n = (2w .

1)

(2w +
3

1)

(2w +
2

3)

= $n 3 +
3m,

12>i

2

- 2m -

3;

\

Sm = 82?i + 122^ - 22m ;

using the notation of Art. 70
.

\S

U

= 2m 2

(m + l) + 2m (n
2

2

+

1)

(2n + 1) - n (n +

1)

- 3m

= w(2m 3 + 8m + 7?i-2).
It should be noticed that the rule given in Art. 383 is 385. only applicable to cases in which the factors of each term form an arithmetical progression, and the first factors of the several terms are in the same arithmetical progression.

Thus the sum 1.3.
5

of the series

+ 2.4.

6

+ 3.5.

7

+

to

n terms,

be found by either of the methods suggested in the preceding article, but not directly by the rule of Art. 383. Here

may

un = n

(m

+

2) (m

+

4)
2)

-n

(m

+1 +
\)

1)

(m

+2+

2)

= n{n+ l)(™ + = n (m +
The
rule can
1)

+ 2n(n+

+ u(a + + 3m.
;

'2)

+

2/4

(m

+

2)

+ 3m (n +

1)

now be

applied to each term

thus
2)

Sn = \n (m+
\
)>

1)

(m + 2)(m + 3)

+n (n+

1) (m

+

+ |« (»+

1)

+C

(r

h

l)(/r»-l) (» +

5),

the constant being zero.

;

316
386.

HIGHER ALGEBRA.
To find
the

of\\ terms of a series each term of which is composed of the reciprocal of the product of r factors in arithmetical progression, the first factors of the several terms being in tlie same arithmetical progression.

sum

Let the where
4i

series

be denoted by u +
x

it,

+u +
:i

+ un

,

Un

— = (a + nb) (a + n + 1
Replacing n by n -

.

b) (a

+n+2

.

b)

(a

+ n + r-l

.b).

1,

—=
M„_i n-l
l

(a
(a

+n+n+

1

.

b) (a

+ nb)

(a

+n+

1

.

b) ...(a

+ n + r—2
say.

.

b)

.'.

r-l

.

b)
1,

un =

(a

+n-

1

.

b)

un _ x = vni

Replacing n by n +

we have
;

(a+nb)un = vn+1
therefore,

by subtraction,

(r-l)b. un = vn -vn+1
Similarly
(r

,

(r

1) b

.

un _ x =
.

v _l
ll

-

vn

,

— l)b
1) b

u.2

=

v.2

vSt
v.2
.

(r—

.

Wj

=v —
x

By
tnatis
, ,
,

addition,

(r

— 1) b Sn = v — v n+1
.

x

;

.

Q

^i

-

*"-(r-l)6~

?W _ r U

_

(

a + nb) un

(r-l)6
n,

'

where

C

is

a quantity independent of

which may be found by

ascribing to n some particular value.

Thus

Sn = C- ,

*
.

(r-l)6

(a

+ n+L.b)... (a + n +

=

.

r-l.
rule
:

,

b)

Hence the sum may be found by the following
th Write doivn the n term, strike off a factor divide by the number offactors so diminished difference, change the sign and add a constant.

from the beginning, and by the common

The value

of

C= (r — Vv~7 o
-,

=

t

1)

(r

tti

u

i

'>

but

**

is

advisable in

1) 6

each case to determine

C by

ascribing to

n some

particular value.

:

.

SUMMATION OF
Example
1.

SERIES.

317

Find the sum of

re

terms of the series

1.2.3.4
The
1

+ 2.3.4.5 + 3.4.5.6 +
1

re'

'

term

is

-

»(n + l)(n + 2)(n + 3)'

hence, hy the rule,

we have

3(n+l)(w + 2)(» + 3)
rut »=1, then

^=0-3^;
X

whence (7=1;

5

*
3(re

"

18

+ l)(re + 2)(re + 3)
fi^

By making n
Example
2.

indefinitely great,

we obtain

=—

Find the sum
3

to n terms of the series

+ a—T7—- r r—5 + 1.2.42.3.5 + 3.4.6
t

4

5

Here the rule is not directly applicable, because although 1,2,3, , the first factors of the several denominators, are in arithmetical progression, In this example we may the factors of any one denominator are not. proceed as follows

n+2
"
re(re+l)(re
re (re
:

(n+2) 2

+ 3)

n{n+l)

(n

+ 2)

(re

+ 3)

4-1)

+ 3re + 4
3
(re

re(re

+ l)(re + 2)(re + 3)"
1

(re

+ 2)(w + 3)

+

l)(re

+ 2)(re + 3)

w(re+l)(n + 2)(re + 3)'
?i

Each
to

which

of these expressions may the rule is applicable.

now
3

be taken as the

th

term of a

series

S

-c

l

4

n+S

2(re

+

2)(re

+ 3)

3 (re+ 1)

(re

+ 2)

(re

+ 3)

'

put re=l, then

17271=
_29
n

3-13
C
1
re

"4 " 27174 " 372.3 .4' Whence C = 36
3 2
(re

4

29
;

4 3
(re

36

+3

+ 2) (re + 3)

+ 1)

(re

+ 2)

(re

+ 3)'

;

318
387.

HIGHER ALGEBRA.
In cases where the methods
of Arts. 383,

386 are directly applicable, instead of quoting the rules we may always effect the summation in the following way, which is sometimes called the
'

Method

of Subtraction.'

Example.

Find the sum of n terms

of the series

2.5 + 5.8 + 8.11 + 11.14+
The
arithmetical progression in this case
2,
is

5,8, 11, 14,

In each term of the given series introduce as a new factor the next term of the arithmetical progression denote this series by &", and the given series by S; then
;

S'
.-.

= 2.

5. 8

+ 5. 8.

11

+ 8.

11.

14+

+(3w-l)(3n+2)(3»+5);
to

£'-2.5.8 = 5.8.11 +

8. 11.14

+ 11. 14.17+...

(u-1) terms.

By

subtraction,

_2.5.8=9[5.8 + 8.11 + 11.14+...to(»-l)terms]-(3n-l)(3n+2)(3n+5),
-2
.

5

.

8 = 9 [S - 2
(3/i

.

5]

-

(3/i

-

1)

(3n+2) (3n+5),
8

9S =

-

1) (3/i

+ 2)

(3/i

+ 5) -2. 5.

+ 2, 5.0,

fif=n(3n3 +6n+l).

When the nth term of a series is a rational integral 388. function of n it can be expressed in a form which will enable us readily to apply the method given in Art. 383.
(n) is a rational integral function For suppose dimensions, and assume
<j>

of

n

of

p

cf)(n)

=A
JB,

+Bn+
C,

Cti(n +

1)

+B)i(u+ l)(n + 2)+

,

where A, number.

D,

are undetermined constants

p+

l

in

values of n, we may equate the coefficients of like powers of n; we thus obtain ^> + 1 simple equations to determine the p + 1 constants.

This identity being true for

all

Example.

Find the sum of n terms

of the series
.

whose general term

is

n*+6n3 + 5w2
Assume
7i

4

+ 6/i3 + 5/t 2 = A + Bn+ Gn

[n + 1)

+ Dn [n +

1) (n
;

+ 2) + En (n + 1) (n + 2) (w + 3)
2,

it is

at once obvious that ,4=0, 2? = 0, 0. successively, we obtain C = - 6, J)

E=1
Thus

and by putting n = -

n=- 3

=

« 4 + 6» 3 + 5/< 2 =//(n +

l)

(n+2)

(?i

+ 3)-6/t(/i + l).

)

SUMMATION OF
Hence

SERIES.

310

Sn = s n
o

(/t

+

l)(»

+ 2)(n + 3)(?i + 4) - 2n(n + l)(n + 2)
i

= \n(n+l)(n+2){n' + 7n + 2).
o

Polygonal and Figurate Numbers.

n+ \n(n— l)b, which is the sum arithmetical progression whose first term is 1 of n terms of an and common difference b, we give to b the values 0, 1, 2, 3, •> we get u, \n (n + 1), n* \n (Bn — 1) n
389.
If in the expression
s
j

which are the u ih terms
third, fourth, fifth,

of the Polygonal

Numbers

of the second,

each term

is

fourth, fifth,

orders; the first order being that in which unity. The polygonal numbers of the second, third, orders are sometimes called linear, triangular
)

square, pentagonal

390. To find the j>olygonal numbers.

sum of the

first

n terms of

th the r order

of

The n ih term
.-.
i

of the rtb order is

n + \n (n -

1) (r

2);

$ =$n + l(r-2)%(n-l)u = \n (n + 1) + 1 (r -2)(n-l) n (n +
= in(n + l){(r-2)(n-l) +
$}.

1) [Art.

383]

391.

If the

sum
th

of

n terms
1,

of the series
,

1,

1,

1,1,
series,

be taken as the

?*

term of a new

we

obtain

1,2,3,4,5,
If again

we take
?t

n —in +
-j

1
,

which

is

the

sum

of

n terms

of the

last series, as the

th

term of a new
1, 3, 6,

series,

we

obtain

10, 15,

proceeding in this way, we obtain a succession of series such that in any one, the nih term is the sum of n terms of the preceding series. The successive series thus formed are known as Figurate Numbers of the first, second, third, ... orders.

By

320

HIGHER ALGEBRA.
and
the

To find the n th term 392. order offigurate numbers.

sum of n

terms of

ilie

rth

The nih term
second order
is

th term of the of the first order is 1; the n th n; the n term of the third order is Hn, that is
tYl

\n
-—
xl

(n

+

1); the

n

term
tIl

of the fourth order is
,
.

2 "V
1
.
.

>

tnat

is

2

n(n+l)(n+2) L± 1.2.3
.

..

'-

the

u th term
,

of the fifth order is

.

~

M

,

^ n(n+l) (n+2) —*-=
2,
1

—^5
.

2

.

o

,

that

is

w(w+l)(n + 2)(M + — —
2

^-n 4

3)
;

and

so on.
of the r th order is

Tims

it is

easy to see that the

n th term
.

w(?*+l )(w + 2)...(n + r-2)

|rc+r-2
01
i

r-1
A«rain, the

n—

1

I

r

1

sum of n terms n (n + 1) (n +
of tlie (r

of the r th order is
2)
. . .

(w

+

r

-

1)

which

is

the

w th term

+

l) th order.

In applying the rule of Art. 383 to find the sum of n terms of order of figurate numbers, it will be found that the constant is always any

Note.

zero.

The properties of figurate numbers are historically 393. interesting on account of the use made of them by Pascal in his Traite du triangle arithmetique, published in 1665.
The following
simplest form
1 1
1
...

table exhibits the Arithmetical Triangle in its

;

SUMMATION OF

SERIES.

321

Pascal constructed the numbers in the triangle by the following rule
:

Each number
immediately
thus
15

is the

sum of

that immediately above

it

and

that

to the left

=

of it; 5 + 10, 28 = 7 + 21, 126 = 56 + 70.

the mode of construction, it follows that the numbers in the successive horizontal rows, or vertical columns, are the hgurate numbers of the first, second, third, orders.
. . .

From

drawn so as to cut off an equal number of units from the top row and the left-hand column is called a base, and the bases are numbered beginning from the top left-hand corner. Thus the 6th base is a line drawn through the numbers 1, 5, 10, and it will be observed that there are six of these num10, 5, 1 bers, and that they are the coefficients of the terms in the expansion of (1 + x) 5
line
;
.

A

The properties
:

of these

numbers were discussed by Pascal

with great skill in particular he used his Arithmetical Trianyle to develop the theory of Combinations, and to establish some
interesting propositions in Probability. The subject treated in Todhunter's History of Probability, Chapter n.
is

fully

"Where no ambiguity exists as to the number of terms in a series, we have used the symbol % to indicate summation but in some cases the following modified notation, which indicates the limits between which the summation is to be effected, will be found more convenient.
304.

Let

cf>

(x)

be any function of

x,

then 2
x=l

<f>

(x)

denotes the

sum

of the series of
1

terms obtained from <f> (x) by giving to x inclusive. tive integral values from I to 'a

all posi-

m

required to find the terms of the series obtained from the expression
it is

For instance, suppose

sum

of all the

(p-l)(p-2)...(p-r)

by giving to p
H.H.
A.

all

integral values from r

+

1

to

j>

inclusive.

21

7

1

322

HIGHER ALGEBRA.
*=* - r)—- - r + (p — sum = 2 (p

Writing the factors of the numerator in ascending order,
.

the required

1)

...

(p —

-

1)

= i{1.2.3.. ..r+2.SA....(r+l)+...+(p-r)(p-r+l)...(p-l)}

=l
\r

(p-r)(p-r + l) „(p-l)p
r

+

1

[Art. 383.] L J

=

y~(-l)(y-2)...(^-r) jr+1
i

Since the given expression is zero for all values of r inclusive, we may write the result in the form

p from

1

to

%

p

(p-l) ( p-2)
\r

•(p- r) _ p(p-l)

(

ff-2) ...(p-r)
r
|

v-\

+

EXAMPLES. XXIX.
Sum
1.

a.

the following series to n terms

:

1.2.3 + 2.3.4 + 3.4.5 +

2.
3.

1.2.3.4 + 2.3.4.5 + 3.4.5.6 +
1.4.7 + 4.7.10 + 7.10.13 +
1.4.7 + 2.5.8 + 3.6.9 +

4. 5.

1.5.9 + 2.6.10 + 3.7.11
the following series to
I
1

+
to infinity
:

Sum

n terms and

1

1.2^2.3
7
'

174
1

II +
4.
1

3.4^

+ 77l0 +
1

1

1_ + 3.5.7 + 5.7.9 + 1.3.5

1.4.7
4

1_ 1 + 4.7.10 + 7.10.13 +
5
1

10

6
1-

1.2.3^2.3.4
*

4-

.3.4.5^
*""' 7
i

11

3.4.5
1 ——

J_ + _1_ + _JL + 4.5.6 5.6.7
3
-l

io \9,

— 1.2.3

5

u

2.3.4

3.4.5

4.5.6

SUMMATION OF
Find the sum of n terms of the
13.
1

SERIES.
:

323

series

,3.22 +2.4.3a +3.5.4*+
2

14.

(?i

-l 2 + 2<> 2 -2 2 ) + 3(> 2 -3 2 ) +
)

Find the
15.

sum
2

of

n terms of the
16.

series
(n*
?*

whose na term
2

is

»*(»*?i

1).

2

(?i'

-l)

4

+ DR + 4)(n 2 + 5n + H). + 2/> 3 + h -1
u-

17.

A i 2 4w 9 -l

*&•

v,

+

/i

1Q iy.
21.

n*+3n?+2n+2
n* +
.,

7i*+n 2 + l
.

_ 2)i

zu.

,

iv
th

+n
order of figurate numbers
is

Shew
If the

that the
th

equal to the r
22. the (n

term of the term of the n order.
?i
tXx

r

th

+ 2)

th

n th term of the r th order of figurate numbers term of the (>-2) th order, shew that r=n+%

is

equal to

23.

Shew that the sum

numbers from the

of the first n of all the sets of polygonal linear to that of the ? ,th order inclusive is

{r-\)n(n + \),

oN

Summation by the Method of Differences.
395.
let Mj, u.2
1
,
,

Let un denote some rational integral function of », and w 3 tt4 ,... denote the values of u n when for n the values
,
. .

2, 3, 4,

.

are written successively.

proceed to investigate a method of finding un certain number of the terms u u.2 w3 u4 ,... are given.
x ,
, ,

We

when

a

From
by
follows

the series u x u 2 u 3 u A u 5 ,... obtain a second series subtracting each term from the term which immediately
,

,

,

,

it.

The

series
u.2

— w,, u s -u.

2

,

u4 — u3

,

u 5 -u 4 ,...

thus found is called the series of the first order of differences, and may be conveniently denoted by

Aw,,

&u~

&u A

,

At*4 ,...

subtracting each term of this series from the term that immediately follows it, wr e have

By

Am.,- A?*,,

Attg — Awg,
series

Aw4 — Awj,...
of
differences,

which may be called the and denoted by

of

the second order

A../',,

A.,?'.,,

AjWg,...

)

.

324

HIGHER ALGEBRA.
this series

From

we may proceed
of
,

to form the series

third, fourth, fifth,... orders these series being Aur 3ur ,

differences,
?t r) ...

of the the general terms of

A

A

A

5

respectively.

From

the law of formation of the series
Uj
t

u.2 ,

u3
Au.2
,
,

,

u±,

u5

,

u6

,

Attj,

Au3
2,

,

Aw4
,

,

Au
2
i
,

5

,

A.y^
.

A2w

A.m3
3

Au
A u3
3
,

A

3

Wj,

A u,,

appears that any term in any series is equal to the term immediately preceding it added to the term below it on the left.
it

Thus

u.2

= «j

4-

Au ly and

Ait.2

= Au +
i

A.m^

.

By

addition, since u.2

+ Au.2 = u 3 we have
ii.j

= t^ + 2Au +
±

A.2 u x

.

In an exactly similar manner by using the second, third, and fourth series in place of the first, second, and third, we obtain

Au3 = Au +
x

2A.2u 1

+

A^.

By

addition, since
?f 4

u3 + Au3 = u4i we have
x

- u + ZAu +
x

SA^ + A^

.

So far as we have proceeded, the numerical coefficients follow the same law as those of the Binomial theorem. We shall now prove by induction that this will always be the case. For suppose that

un+i = «i + mAmj +

v

X

-J

9

A,u +
l

...

+ "CVA^j +

+

AnWj

i

then by using the second to the (n + 2) th series in the place of the th first to the (n + l) series, we have

Au n+1 = A%! + nA. u + -A.
2
}

it (11

——
.

1

jr-f-

A 3Wj +

.

.

.

+ B C r_ 1 A^w1 +
f

.

.

.

+

A n ^u

Y

.

By
M»+a

addition, since

u n+l + Au n+1 = un+2i we obtain
. .

=

Mj

+ fa +

1)

Awj +

+

n
(

Cr + *Gr _j) A ru

x

+ ...+ A„ +1«,

)

)

.

;

SUMMATION OF
But
*Cr + HJr -i -

SERIES.

32.",

(—
(n +

^

+ l)

x

»C r _ x = ?i±i

x "C,,.,

l)w(w-l) ...(w+l-r+1) _ 1.2. 3... (r-l)r
it

Hence
f£ n+8 ,

if tlie

law of formation holds for u n+l

also holds for

true in the case of therefore universally. Hence
it is

hut

w4

,

therefore

it

holds for

u rn and

"„

=

Ui

, + (n -

1X 1)

.

A?^ +

(w-l)(w-2) £*_ *

A

.

2

?^

+

...

+ An.iWi.

39G.

To

find the

sum

of

w terms

of the series

in terms of the differences of

ul

.

Suppose the
of the series

series u^, u.2
Vl,

,

u 3 ,...
v3
,

is

the

first

order of differences

v. 2)

v4 ,...,
...

then v n+1 = (vn+1

- vn) +
^»+l

(v n

- vn _ t) +

+

(v2

-v) +
x

v x identically

'•

= u a + u n-l +
va1
1

••

+ u 2 + u\ + v
v5
4
J

l

Hence

in the series
0,

v3i
2 3

v4
J

,

Aw1? Aw 2
the law of formation
•'

,

Aw

3

is

the same as in the preceding article;

«»+i

=

+ «Wi + -4
...

—s— Aw

x

+

.

.

.

+ A„?^

;

that

is,

Wj

+ w3 + u z +
y

+ un
>

- nu +
x

n —(n—\) AWj + n(n—l)(n-2) A M! + -— -^ —r —
t

2

.

.

.

+

A„?f

,

The formula)

of

this

and the preceding
,

article
:

may be

ex-

pressed in a slightly different form, as follows

term of a given series, (I x d2 d3 ,... cessive orders of differences, the n th obtained from the formula
,

a is the first the first terms of the sucterm of the given series is
if

'

326
and the sum of n

HIGHER ALGEBRA,
terras is

^i"^" 2
Example.

1

)^

»(»-l )("- 2
,

)

n(»-l)(tt-a)(n-3)
f/

j3

4

Find the general term and the sum
12,

of

?*

terms of the series

40,90, 168, 280, 432,

The

successive orders of difference are
28, 50, 78, 112, 152,

22, 28, 34, 40,
6,
0, 6, 6,
0,...

Hence the n th term

= ,c + ™,(re - ,x + 12 28 1)
=
?i

22(re-l)(re-2) K
-

~P

'

+

6 (re- 1) (re- 2) v

-

M

M

(re

-3)
-

II

3

+ 5re2 + 6>t.

2re

3

The sum of n terms may now be found by writing down the value of + 52re2 + 62re. Or we may use the formula of the present article and

obtain

S^ia^ 28"'"- 1

'

+

22 "'"- 1)( "- 2 »

+

«M»-D (»-2) (-8)

= ^(3re2 + 26re + 69re + 46),

= in(re+l)(3n 2 + 23re + 46).
397.
It will be seen that this

method

of

summation will only

when the series is such that in forming the orders of differences we eventually come to a series in which all the terms This will always be the case if the n th term of the are equal.
succeed
series is a rational integral function of n. ^»
j

For simplicity we will consider a function of three dimensions; the method of proof, however, is perfectly general.
Let the
u.1

series

be

+ ua + u a + 2 3

+ u +u n + .,+u n+2 + u u + „ + „ 3
.
,

ii

\

where

u " = An 3 + Bn 2 + Cn + D.
,

and

let v n'

w

, ii*

%
ii

denote the

?i

th

term of the

first, *

second, third
*

orders of differences;

SUMMATION OF
then
that
is,

SERIES.
l)

327
1)
;

vh

-

m m+1 — un = A(3n* + 3n+
vn =

+ 2?(2» +

+C:

Similarly

w

3Au 2 + (3A + 2B) n + A + 11 + C = v — v = 3A (2n + I) + 3A + 211
.

,

and

zH

.— =w »ti iv

=6-4.
ii

Thus the terms
and
generally,
if

in the third order of differences are equal; the n ih term of the given series is of p dimensions,
th

the terms in the

p

order of differences will be equal.

Conversely, if the terms in the ]j th order of differences are equal, the u tu term of the series is a rational integral function of ii of p dimensions.
Example.

Find the « th term of the

series

-1, -

3, 3,

23, G3, 129,

The

successive orders of differences are

-2,

6, 20, 40, GO,
8, 14,

20,20,
6,
0,
;

6,

Thus the terms assume
where A, B, G,
Putting
1,

in the third order of differeLces are equal
it

hence we

may

H

= A+Bn+Cn + Dn
2

3
,

D have
2,

to be determined.

4 for 7i in succession, we have equations, from which we obtain A =3, B = -3, C = - 2,
3,

four simultaneous

D—\

;

hence the general term of the

series is 3

- 3n - 2n 2

+ n9

.

398.

If a

is
ri

a rational integral function of p dimensions
a,

in

n, the series

+

ax + ajx

12
+
{a,
...

2

+

...

+ a xn
»i

is

a recurring
Let

series, ivhose scale

of relation
;

is (I

x) p+1

.

S

denote the sum of the series
--

then
.

S (1

-

x)

a o + (a - a )x
x

- ajx* +
-

..

+

(a,

- a ,_>" - ax" +
n.

l

=a +
here b n

b
t
,

x + bjc2 +
"•

+ bx" - ax" + \ say;
1

=a —a
h

it

1

, '

so that 6 n is of x p
1

dimensions in

Multiplying this last series by

x,

we have
+l

S(i-xy

=s+(^-a„)*+(^-^K+..-+(6n-6„-iK-(« ,+6>"
J

+«X

+a

= c^+{b-a )x+c^ + c i x^...+cX-{a +b
2
i

i

y^+
,

a
i

,+
:c

% say;

here cn -b n

-b u u

so that cn

is

of

p-

2 dimensions in n.

328

HIGHER ALGEBRA.
it

follows that after the successive multiplications by n products 1 — x, the coefficients of x in the first, second, third, orders of differences are general terms in the first, second, third,

Hence

.

.

.

.

.

.

of the coefficients.

hypothesis a n is a rational integral function of n of p dimensions ; therefore after p multiplications by 1 - x we shall arrive at a series the terms of which, with the exception of p terms at the beginning, and p terms at the end of the series, form a geometrical progression, each of whose coefficients is the same.

By

[Art. 397.]

Thus
where k
is

S (1 - xf = k(xp + x>' +1 +

...+ x") +/(a?),

(x) stands for the a constant, and the beginning and p terms at the end of the product.

f

p terms

at

r.Sil-xy
that
is,

J^l^K/ix);
^

a=
is

kx»(l-x"-^) + (l-x)f(x)
(1

-x)

p+l

'

thus the series

a recurring series whose scale of relation

is

(l-x) p+1

.

[Art. 325.]

not given, the dimensions of an are readily found by the method explained in Art. 397.
If the general

term

is

Example.

Find the generating function of the
3

series

+ 5a; + 9a; 2 +15a; 3 + 23a; 4 + 33a; 5 +
of differences of the coefficients,
G, 8,
2,

Forming the successive orders
the series
2,

we have

4,

10,
;

2,

2,

2,

thus the terms in the second order of differences are equal hence a n is a rational integral function of n of two dimensions ; and therefore the scale We have of relation is (1 - a;) 3
;

.

+ 15.r 3 + 23a; 4 + 33a; 5 + - SxS = - 9.r - 15a; 2 - 27.x- 3 - 45a; 4 - 69^ 5 4 2 3 9a; + 15a; + 27.r + 45a; + Sx 2 S = - 3^- 5a;4 - 9a; 5 -xs S=
9a;
2

S = 3 + 5x +

By

addition,

(

1

- a;) 3 S = 3 b

4a;

+ 3a; 2
2
*

;

3-4.r +

3a;

•*•

~

(1-a;) 3

SUMMATION OF
399.

SERIES.

329

have seen in Chap, xxiv. that the generating function of a recurring series is a rational fraction whose denomiSuppose that this denominator can nator is the scale of relation. then the be resolved into the factors (1 — ax) (1 — bx) (1 — ex) generating function can be separated into partial fractions of the
to rm
,

We

1

- ax

ABC
1-

;

1

bx

1

- ex

can be expanded by the Binomial Theorem in the form of a geometrical series; hence in this case the recurring series can be expressed as the sum of a number of geometrical series.

Each

of these fractions

however the scale of relation contains any factor more than once, corresponding to this repeated factor there
If

1

- ax
be

will

which r=, - ax) when expanded by the Binomial Theorem do not form geometrical series; hence in this case the recurring series cannot be expressed
partial

fractions

of

the form

-^

(1

-axy

A—

A

-7,

...

:

(1

as the

sum

of a

number

of geometrical series.

The successive orders of differences of the geometrical 400. progression
a,

ar,

ar

2
,

ar

3
,

ar\ ar
,

n
,

are

«(r-l), a(r—l)r, a(r-l)r 2 a(r—\)r?
'

a(r-l) 2 a(r-l) 2r, a(r-\fr 2
,

,

which are themselves geometrical progressions having the same

common
401.

ratio r as the original series.

Let us consider the

series in

which

where </>(rc) is a rational integral function of n of p dimensions, and from this series let us form the successive orders of differences. Each term in any of these orders is the sum of two parts, one arising from terms of the form ar n ~\ and the other from terms of the form <£(?i) in the original series. Now since <f>(n) is of ;; dimensions, the part arising from <f>(n) will be zero in the (p + l) th and succeeding orders of differences, and therefore these series will be geometrical progressions whose common ratio is r.
[Art. 400.]

330
Hence
if

HIGHER ALGEBRA.
the
first

few terms

of a series are given,

and

if

the

p

order of differences of these terms form a geometrical progression whose common ratio is r, then we may assume that the general term of the given series is ar"" +f(n), where f(n) is a rational integral function of n of p - 1 dimensions.
th
1

Example.

Find the n th term
10,

of the series
60, 169, 494,

23,

The

successive orders of differences are
13, 37, 109, 335,

24, 72,

216,

Thus the second order
the

of differences

common

ratio is 3

;

hence we

may
.

a geometrical progression in which assume for the general term
is

un —a

S n -^
b, c,

+ bn + c.
make n equal
to
1, 2,

To determine
then

the constants a,

3 successively;

a + b + c=10, 3a + 2b+c = 23, 9a + 3b + c = 60;
a = 6, 6=1,
un = 6
.

whence

c

= S.
.

Thus

3' 1

"1

+n+3=2

3» + n +

3.

In each of the examples on recurring series that we 402. have just given, on forming the successive orders of differences we have obtained a series the law of which is obvious on inspection,

for

and we have thus been enabled to the ?4 th term of the original series.
If,

find a general expression

however, the recurring series is equal to the sum of a number of geometrical progressions whose common ratios are n~ + Cc n ~\ «, b, c, ..., its general term is of the form Aa"' + Bb and therefore the general term in the successive orders of differences is of the same form ; that is, all the orders of differIn this case to ences follow the same law as the original series. find the general term of the series we must have recourse to the more general method explained in Chap. xxiv. But when the coefficients are large the scale of relation is not found without considerable arithmetical labour ; hence it is generally worth while to write down a few of the orders of differences to see whether we shall arrive at a series the law of whose terms is
1
l

evident.

403.

We

add some examples in further

illustration of the

preceding principles.

SUMMATION OF
Example
1.

SERIES.

33]

Find the sum

of n terras of the series

1.2'3 + 2.3'33+ 3.4'35 + 4.5

*3^

+


"

2« + 3
1l(ll

1

+ l)
.4

3"
7?

Assuming

=-+ n(u+l) n
=
.

2n — +3

n+
-1.
-

=

,

1

we

find
tt Hence

A =3,
t/,.

B=
.

"

=

/3 \n
(

1

\
)

n + 1)
#,, n

1 — =1 3"

n

3"- 1


1
.

,

11 —
.

n +

1

3"'

.

and therefore

=1

n+1

3' 1

.

Example

2.

Find the sum of n terms of the
1

series

3

+ _3_ +
3. 7

_5
3. 7. 11

+

7

3. 7. 11. 15

+

The
.

ri h

term

is

.,„,., 3.7 11
.

r, kt: (An- 5) (4/i-l)
•,

ssume
3 7
.-.

2n-l
(4n _
5) (4 n

_i)

~3
>

A
.

(n

+ 1) + B
"

An + B
3.7
(4„ - 5)
'

7 ......4»-l

2rc-l = ,4n + (J+I> )-(.-t» + .B)(4?i-l).

On equating coefficients we have three equations involving the two unknowns A and B, and our assumption will be correct if values of A and B
can be found to satisfy
all three.
,

Equating

coefficients of n 2

we obtain ^1=0.
it

Equating the absolute terms, -1 = 2B; that is B = -%; and found that these values of A and B satisfy the third equation.
1
""'

will

be

1

V,l

~2 *3.7
S„
"

1

1

(4»-5)

2'3.7

(4»-5)(4»-l)
(4»-l)

;

hence

=
2 2

.

3.7.11
series

Example

3.

Sum
G. 9

to

n terms the
21

+ 12.

+ 20.

37

+ 30.

57 + 42. 81

+
find that the
;t

By

the

method

of Art. 396, or that of Art. 397,
6,

we

th

terra

of the series

12, 20, 30, 42,

is

?r

+ 3» + 2,

and the

th

term of the
9,

series

21, 37, 57, 81,

is2n*+6n+l.

;

3.32

HIGHER ALGEBRA.
»„=(« + 1)

Hence

+ 2) {2m (m+3) + 1} = 2m [n + 1) (?i + 2) (»+ 3) + (n + 1) (m + 2)
(a

•'•

S«=ln(»+l)(»+2)(n+3)(n+4)+|(n+i)(n+2)(n+8)-2.
4.

Example

Find the sum

of

??.

terms of the series

2.2 + 6.4 + 12.8 + 20.16 + 30.32+
In the series
2, 6, 12, 20, 30,
2

the

?i

th

term

is

n2 +

n

;

hence

un = {n +
(rc

n) 2 n .

Assume

2

+ m) 2' = (An2 + Bn+ C)2 n 1

{A (n-l) 2 + B

(n

-

1)

+ C\
n,

2"- 1
;

dividing out by

2' 1_1

and equating
2=A
t

coefficients of like

powers of

we have

2 = 2A+B,

0=C-A + B;
0=4.
l) 2

whence
.-.

A=2,
wn = (2?i2 - 2n + 4)
2n

B=
{

-2,

and

S n = (2m2 - 2m + 4)

+ 4 2"2 n - 4 = (na - n + 2) 2*« - 4.
- 2 (n - 2 (n 1)
}

1
j

EXAMPLES.
4,
8,

XXIX.

b.

Find the n th term and the sum of n terms of the
1.

series

14, 30, 52, 80,

114,

2. 3.

26, 54, 92, 140,
12, 36, 80, 16, 0,

198,

2,
8,

150, 252,

4.
5.

-64, -200, -432,

30,

144, 420, 960, 1890, 3360,
series

Find the generating functions of the
6.
7.
1

:

8.

9.

10.

+ 3x + 7x2 +13.^ + 21a4 + 31a6 + 2 3 1 + 2a + 9a + 20a + 35a4 + 54a 3 + 2 + 5a + 10a2 + 1 7a-3 + 26a4 + 37a-5 + - 3a + 5a2 - 7 Xs + 9a4 - 11a6 + I 4 + 2% + 3 4 a 2 + 4 4^ + 5 4 a 4 +
1

Find the sum of the
11.

infinite series

:

3

+
2

32

+

33

+
2

g4

+
"

12 1Z>

i2
*

_?5

+ ??_iV _ + + 5« 5* 53 52

62

SUMMATION OF

SERIES.
:

333

Find the general term and the sum of n terms of the series
13.
9,

16, 29, 54,

103,

14. 15. 16.
17.

-3, -1,
2,
1,

11, 39, 89,

167,

5, 0,

12,
1,

31,
8,

8(i,

29, 80,

193,
Tr»5

4,

13, 35, 94, 262,

Find the sum of n terms of the
18. 19.
1

series

:

+ 8* + 3.>/- +

4./,-' ;

+
5

5.t-

1

+

1+ 3.i- + 6x2 + lO.f' + 1 5.r* +
4
1

onJLi + 1.2*2
21
'

2.3

"2

'
:2

f

6 3.4'2 + 4.5'2 4 +
5
1 1
!

2T3-

4+

i£i11

4S+

4^

5- 4 ' +

0- 44+
+ +
4
4-

22.

3.4 + 8.

+ 15.20 + 24.31+35.44+

23.

1.3 + 4.7 + 9.13 + 16.21+25.31
1.5 + 2.15 + 3.31+4.53 + 5.81
1

24.

oC 25

2
1

3
A

'

1.3^1.3.5
1.2
2.2-'

1.3.5.7
4.2 4

k

— 1.3.5.7.9

nn 26

'

^- + 14- + -T5- + -T6- +
1

3.23

27.

2.2 + 4.4 + 7.8 + 11.16+16.32 +
.

28.
rtr
.

3

+3

.

32

+ 5 3 3 + 7 34 +9.
. .

3>

+

...

1
'

1.3

1.3.5

1.3.5.7

2.

42. 4. 62.4. 6. 82. 4. 6. 8. 10

^ 30
32

-± + 2+i5L 92, l l 3+ 2.3' 2+ 3.4' 2 + 4.5" 2 + 1.2
_4_
1

+—

1.2.3*3
|4
|5

1 1 _5_ + 2 5" 33 + 2.3.4' 3 3.4.
(J

±+A + H + ^ + (3^
|6

33
'

19
1
.

I
.

2

3

'

+2 4

28
.

1
.

3

4

"

8

+

_39_
3
.

J_
*

4

.

5

16

+

52
4
.

1
.

5

6

'

32

+


334
404.

HIGHER ALGEBRA.

There are many series the summation of which can be brought under no general rule. In some cases a skilful modification of the foregoing methods may be necessary ; in others it properties of Avill be found that the summation depends on the
certain

known expansions, such

as those obtained

by the Binomial,

Logarithmic, and Exponential Theorems.
Example
1.

Find the sum of the
2
[I

infinite series

+

12
|2

+

28
|3_

+

50
|I

+

78
|5

+
..

term of the

series 2, 12, 28, 50, 78.

v

.

is 3n-

+ n -2; hence

3h 2 +j«-2
a »"
]n

3h(h-1)+4»-2
|n

2

;i-2

+

n-1

In"

Put n equal

to 1, 2, 3, 4,... in succession

;

then we have
3

",

= 4--;

2

„2 = 3

4 + ri i

2
;

r2

«3 =j

+ ^ - -gj i

4

2

and so

on.
,Sf„

Whence
Example

= Se + 4e - 2 {e If (1

1)

= 5e + 2.
.

2.

+ a;) n = c + c rr + c 2 .r 2 +
l-c 1

.

.

+ c nx n

,

find the value of

+ 2 2 c 2 + 3 2 c 3 +... + n\v
shew that
. .
.

As

in Art. 398
l2

we may

easily

+ 2 2 .r + 3 2 .r 2 + &x 3 +...+ n-x n ~ l +
. .

=
.

Also cn + c n _ x x +

.c. 2

£n

-2

+ c^

11

'1

+c

xn =

(1

+ .r) n

Multiply together these two results;
(l

the coefficient of x 11

x

in

+

n+1 .r)
.„
,

.

,.,

(1

- x) A
jc

that

then the given series (2 - 1 - x) n+1
.

is

equal to

is,

in

-

-

7

J=

(1

-

x) 3

.

The only terms containing
2"+! (1 .r)- 3

n-1 in this expansion arise from
l - .t)" 2 + \!l±Jl %*-i

-

(n

+

l 1) 2> (1

(i

_ ^-l.

.-.

the given series =

f

L^+3 2»+i _ „ („ + 1) gn + ?ii"±:
.

l
)

2 h-i

-n(w+l)2«- !


Example
IP _ (n _
3.


335
is

MISCELLANEOUS METHODS OF SUMMATION.
If b

1) „,,.

-+

= a + l, and »-»>(»-«>
)i

a positive integer, find the value of
.

rfj.

_ C-3)(»-4)(»-5)
\6

|2

^

+

By

the Binomial Theorem,

we

see that

(n-8) (n-2)
are the coefficients of x n
1-2
,

(n-5)(n-4)(»-3)
1
,

.r'

,

.r

n_4

-*

5

.r'

,

in the expansions of (1

x)

',

(1-.t)- 2
e<pial

Hence the respectively. (1-x)-*, (l-.r) -4 to the coefficient of x* in the expansion of the scries
,
,

sum

required

is

1

ax*

1-bx
this series

{1-bx)

+ 3
'

a-x 4

a*x 6
(1

(1-fcc) 8

- bx)*
finite

'

and although the given expression consists only of a

number

of terms,

may

be considered to extend to infinity.
series

But the sum of the

=

,
;

1-bx
1

•-

( 1

\
i

+

,

1-bx J
,

)

=

z

1-bx + ax"

z

-

(a

+ l)x + ax-

since b

— a+1.

Hence the given

series

= coefficient = coefficient
a H+l

of x n in

(l-x)(l-ax)
a=
(

of

xn in

1

\1

-ax

-

~ 1—x)
-

}

_

1

a-1
Example
,

4.

If the series

1

+

x3
J3

+

xe
JG

+

r

+

X*
]5

'

+

X7
|7

+
'

x'

2

+

X5
|5

+

X8
|8_

+

|2_

are denoted by a,
If

b, c

respectively,

shew that a8 + 63 + c8 -3o6c=l.
of unity,
(a

w

is

an imaginary cube root
c3

a3 + b 3 +

- Sabc = {a + b + c)

+ wb + w'-c
xz
\3

)

(a

+

w-b

+ ojc)

.

Now

'

lA ~ h

+ c = 1+x +

.t

2

~\9

+

+

.t

4

Tl

+

Xs
~\5

+
w'.r'

w-.r-

and

>/

+

lob

+ OJ-C-1+ C0X+

—+
\

ur\r-

-r^\

+ —T- +
\

w 4 .c 4

-r=~
\

=e
similarly

I

a

+
•.
,

io'-b
,
.,
,

+ wc = c
,

0)=X

o

,

bc

=

X
1,

uX
since l

co

2

X

(l+u> + w ! )x

+ w + ur = 0.

)

.

'

336
405.

HIGHER ALGEBRA.
To find
the

sum of

the r th

powers of

the first

n natural

numbers.

Let the sum be denoted by
r

Sn

;

then
r

SH =V+2 +
Assume that
Sn =A nr+i +An + A n where A A^ A A 3 ...
r

3

+

...

+n

r
.

r

1

n

1

2

+An ~
r

2

3

+

...

+An + A
r

r

+1
1

(1), / \

,

2

,

,

are quantities whose values have to be

determined.

Write n +
(n + l)
r

1

in the place of

n and
{{n
x

subtract; thus

=A

{(n

+

l)
1

r+1

- n r+1 ] + A
1

+

1)'

- nr ]
-2
}

+ A 2 {(n+ l)- -n'-

}

+ +

A
n.

3
r

{(n

r + iy~ 2 -n

+

...

+A

r ...(2).

Expand (?i+l) r+ \
efficients of like

(n

l)

,

(n+l) r_1

,

...

and equate the

co,

powers of

By

equating the coefficients of n r

we have

l=A.

(r

+

1),

so that

A =
a
!

1

T

.

By

equating the coefficients of
r

n
;

r

,

we have
1
x

=

A (r+
°

l)r —— + A

r
x

whence A = ^
substitute for

Equate the

coefficients of

n

r

p
,

A and A Jf and

multiply both sides of the equation by
\P

r(r-l)(r-2)
we thus obtain
i

...

{r- 2)+

;

1)

~p
In
r

+

+
l

+A
2

'r

+A

'r(r-l)

+

^ r(r n and
r

1)

(r-2)

+

"^
1

(1) write

w—

1

in the place of
r

subtract; thus

n =A

{n

r+l

-(n-iy +i }+A

{?i
l

-(n-l) } + A 2
r

{n

r

'

1

-(n-iy- } +
for

...

Equate the

coefficients of

n

~p
,

and substitute
4

A A
,

1

;

thus

o

p+

'4
1

2

+

^-^gzi) +i /^;);^) 2

-....

r

3

r

(?•

1

v

(?•

1) (r

2)

w

MISCELLANEOUS METHODS OF SUMMATION.
From
2
(3)

337

and

(4),

by .addition

.and subtraction,

p+

1

"r

*

r(r-l)(r-2)

o^/_^)^/ 0'-i)(p-g(^-3) + 3

r (r

1)

r

(?•

1) (r

2) (r

3)

w
. . .

(6).

ascribing to }> i n succession the values 2, 4, 6, from (G) that each of the coefficients A.^ A 5> A.,... to zero; and from (5) we obtain
1

By

,

we

see

is

equal

r
J

___1_
'

r

(r-l)(r-2)
,

6 " 1^

30
.

li

_J_ r(r-l)(r-2)(r-3)(r-4) 8 ~42"
|6
(2),

;

By

equating the absolute terms in

we

obtain

\=A^A
and by putting
l

X

+ A% + AZ +

+A
r

-

r

n= 1 in equation (1), we have +A + A r+l 1 = A + A + A + A9 + a
A r+1 =
0.

;

thus
406.

The result of the preceding expressed by the formula,

"

article is

most conveniently
_3

n r+x

1

,

„ r
l

r

_

x

_ r(r-l)(r-2)
3

r

r+1

2

2

4

r(r-l)(r-2)(r-3)(r-4)
'6

^

+

w lprp
}

7?

-

i

7?-

1

7?

i

7?

i

7?

5

quantities 2? ... are known as Bernoulli's Numbers; , 3 5 for examples of their application to the summation of other series the advanced student may consult Boole's Finite Differences.

The

B B
x

,

,

Example.
ttt

Find the value of
S„ = ^r.

l5

+ 25 +
.

35

-f

+ n5
.
.

.

We

,

have

n6

+ ^ + ^ -^ n* - Ba
n5
"2

n5

^,5
5?i 4

_5 4 3 n* + C, —j— _

~6 +
the constant being zero.
II. ii.

_?t6

+

n2

l2~r2'
22

A.

1

HIGHER ALGEBRA.

EXAMPLES. XXIX.
the following Find the sum of
series:

c.

JL + ^

^

A ^

5.

l

+ ^+-\T'T
r

|1

2 -3

ii
r/3

3

p
6.

p

r-i

q

£± £ + f^l.2- +
12

to r

+

1

terms.

*rz

(1
7-

TX^" 1+tm;

+ .r) _ »^"2) "'

+2 ^ "•(1+^)"
X
_

?i(?i-l)(™- 2 )

1+3a?

-

to n terms.

2n + l

,

2/i + K /

x

Y+
12

...

to

n

terms.

o,2

W

2

( 7l

2

_

12)

7i

2 (ft2

2

)^
1

2

9.

i-j[i+-ii7?

.2*.3

2

^

2
)

+

to

w + 1 terms.

+ 23

1L

+ 3^T5 + 5T677 + r2T3
2

12.

ji

+

3^6 +
]|
[3

+

11
|4

+

18
|5

+
121s 6 _
16

2a 8
is.
1

^+W

23s 5

"[7-'|r

+-J2-|3
14

Without

fuming

the formula, find the
+»«•

sum

of the series:

W

!«+*+*+

«

17

+ 2; + 3? +

+ "-

SUMMATION OF
15.

SERIES.
53
I*

339

Find the sum of

l3

+ 23 + - +

33

43

— + _+
I*

B
16.

Shew
"'- 1
/-

that the coefficient of x n in the expansion of
1

(l-X) 2 -r.r
)

.,

is

Y

fl

+
1

("

2

-l)(« 2 -4)

if"
If

n
)(' t

c+ (n*-lKn*-4)(n*-9) * +
,
1

[7

}'

17.

n

is

a positive integer, find the value of

8 .-

(

»-i)^ + e»- g 2
\

- 8)

2

^- (»- 8 )(»- 4 )»- 6
\o

)

g

^+

and

if

11

is

a multiple of

3,

shew that

1

-

(

»- 1)+ (»-»H»-3) _(»-8)(»-4)(»-6) +
?i

=( _ 1)n

18.

If

is

a positive integer greater than

3,

shew that
(
...

rf+ «flga (.-

y+ "(»-i)<«-«)(— *) ,l 4y +
??.

=»»*(» + 3)
19.

SP*" 4.

Find the sum of

terms of the series
2
l

:

W
(2)

1
i

+ i2 +

+
i"4

+2

2

+ 2 4+ l+3 2 + 3 4 +
17

3

_5__J_+JL__L+i3 _JLL+
2.3 3.4
4.5

5.6

6.7
is

7.8
(-l) n + xn ?i(n+l)(n + 2)
1

20.

Sum

to infinity the series

whose n th term

If (1 21. x)n Cq c vv integer, find the value of

+

— +

+ c^v 2 + CyV3 +

+ cn#n n
,

being a positive

(n
22.

- \)\ + (n - 3) 2 c3 + (?i - 5) 2 c5 +
series
:

Find the sum of n terms of the

„N

2

4

8

16

32

1.5

5.7

7.17
17

17.31
31

31.65
49

^
23.

7

1.2.3

+ 2.3.4 3.4.5
if

4.5.6
1

_71_ 5.6.7

Prove that,

a < 1,
'

(

1

+ or)

(

+ A) ( 1 + a?x) ....
aPx3
' !

= 1 + 5- -,+T: =K 1-a 2 (l-« )(l-a 4 + (l-« 2 )(l-a 4 )(l-a'
)

ax

a*x2 =5wi 2

"

)

22—2

340
24.

HIGHER ALGEBRA.
If

A

r

is

the coefficient of xr in the expansion of
(i

+ ,f(i +

2/ (i
2

2 *A2

/

^.\2

|)

+
,

|)
v

(i

+ 2~ J
;

3

'

prove that

^l r

=

2^
-

2s

,,
(^4 r -i

+ ^r-

2)

>

j and ^4 =
,

1072
"3^5

25.

If

n

is

a multiple of

n

-^~\i
[3

6,

shew that each

of the series

3+
[5
1
,

-

3

"
•-

w(w-l)(w-2)
11

*3 +

n(n-l)(n-2)(n-S)(n-4)
|5

1 ""'32

is

equal to zero.
26.

If

n

is

a positive integer, shew that

pti

+1

_ qn + 1
.

is

equal to

27.

If

P =(w-r)(»-r+l)(n-r+2)
r

(n-r+^-1),

&=r(r+l)(r+2)
shew that

(r+^-1),

P&
28.
,

+ P2Q2 + P3Q3+
?i

+ P »-i^-i =
3,

ho

k \n-l+p + q
+ g +l|n-2

|>

If

is

a multiple of

shew that

1

"^"

»-3

+ "~

(m-4)(w-5)
|3

(w-5)(w-6)(w-7)

+ (-!)
is

^

H
(n-r-l)(w-r-2)...(tt-2r + l)
u.
"'"•••'
,

3 equal to - or u

n x


1

n

,

according as n

is

odd or even.

29.

If

is

a proper fraction, shew that

x 1_^2

xz T l_#6

x5
l_a?io
1

x

x3
'

Xs
l+.r10

+.v2^1+^

;

CHAPTER XXX.
Theory of Numbers.
In this chapter we shall use the word number as equivalent in meaning to positive integer.
407.

not exactly divisible by any number except itself and unity is called a prime number, or a prime; a number which is divisible by other numbers besides itself and unity is called a composite number \ thus 53 is a prime number, and 35 is a composite number. Two numbers which have no common factor except unity are said to be prime to each other thus 24 is prime to 77.
is

A

number which

408. propositions,

We shall
some

make frequent use

of the following elementary

which arise so naturally out of the definition of a prime that they may be regarded as axioms.
of
(i)

If a
b, it

factor

number a divides a product must divide the other factor c.

be

and

is

prime to one

For since a divides be, every factor of a is found in be; but since a is prime to b, no factor of a is found in b; therefore all the factors of a are found in c that is, a divides c.
;

prime number a divides a product bed..., it must divide one of the factors of that product ; and therefore if a prime number a divides b", where n is any positive integer, it must divide b.
(ii)

If a

a is prime to each of the numbers b and c, it is prime to the product be. For no factor of a can divide b or c therefore the product be is not divisible by any factor of a, that is, a is prime to be. Conversely if a is prime to be, it is prime to eacli
(iii)

If

;

of the

and c. Also if a is prime to each of the numbers b, c, d, ..., it is prime to the product bed... and conversely if a is prime to any number, it is prime to every factor of that number.
b
;

numbers

342
a and integral power of a
(iv)

HIGHER ALGEBRA.
If
b are

prime to each other, every positive is prime to every positive integral power of b. This follows at once from (iii).
(v)

If

a

is

prime to

b,

the fractions

bo
=-

and j- are in their
ft

lowest terms,

n and

m being any positive integers.
fractions,

Also

if

j and

- are any two equal
c

and
of a

j

is

in its lowest terms, then
b respectively.

and d must be equimultiples
409.

and

The number of primes
if

is infinite.

p be the greatest prime number; then the product 2 3 5 7 11 .p, in which each factor is a prime numand therefore ber, is divisible by each of the factors 2, 3, 5, .p the number formed by adding unity to their product is not hence it is either a prime divisible by any of these factors number itself or is divisible by some prime number greater than p in either case p is not the greatest prime number, and thereFor
not,
.

let

.

.

.

. .

.

.

;

;

:

fore the

number

of primes is not limited.

410.

No
only.

rational

algebraical

formula can
2

represent

prime

numbers

the formula a + bx + ex + dx + ... represent the value of prime numbers only, and suppose that when x = the expression is ]), so that
If possible,
let

3

m

3 2 p — a + bm + cm + dm +

;

when x = m + np
a +
that
or
is,

the expression becomes

b (m

+ np) +

c

3 {m + np) 2 + d (m + np) +

...,

a+

bm + cm 2 + dm3 +

.

. .

+ a multiple

of p,

p+
is

a multiple of p,

thus the expression

divisible

by

£>,

and

is

therefore not a prime

number.
411.

A number

can
the

be resolved into

prime factors in only one
,

way.

Let

N
/3,

denote

number; suppose

N = abed...,

where

a, b, c, d, ...

where

a,

Suppose also that are prime numbers. Then y, 8, ... are other prime numbers.
abed...

JV = a/3yS...,

=

a/3yS...

;

.

THEORY OF NUMHEHS.
;

343

hence a must divide; abed... but eacli of the factors of this product is a prime, therefore a must divide one of them, a suppose. But a and a are both prime therefore a must be equal to a. Hence bed. =/3yS. and as before, /? must be equal to one of the
;
. .
.

;

.

factors of bed... J and so on. Hence the factors in a/3y<$... are equal to those in abed..., and therefore iV can only be resolved into prime factors in one way.

412.

Let
a, b,

N

c, ...

integers.

number of divisors of a composite number. denote the number, and suppose N"=apbg<f..., where are different prime numbers and p, q, r, ... are positive Then it is clear that each term of the product
To find
the

(l+a + a' +
is
;

...+a'')(l+b +

b

2

+

...

+ V)

(I

+

c

+

c

2

+ ...+c r )...

a divisor of the given number, and that no other number is a divisor hence the number of divisors is the number of terms in the product, that is,

This includes as
413.

(f>+l)fe+l)(r + l) divisors, both unity and the number
the

itself.

To find

number of ways in which a composite number

can be resolved into two factors.

Let N" denote the number, and suppose a, b, c... are different prime numbers and ]), integers. Then each term of the product
(I
is

N=
c

a'tyc'

.

. .

,

where

q, r...

are positive

+ a + a2 +

...

+

of) (1

+

b

+

b

2

+

. . .

+ b' )
1

(1

+c+

2

+

.

.

.

+

r

c

)

.

.

a divisor of iV; but there are two divisors corresponding to each way in which iV can be resolved into two factors hence the required number is
;

}(!>+l)& + l)(r + l)
This supposes quantities^, q,
If

N not
r, ...

a perfect square, so that one at least of the
is

an odd number.

a perfect square, one way of resolution into factors is x/iVx JNj and to this way there corresponds only one divisor JX. If we exclude this, the number of ways of resolution is
is

N

!{(p+l)(? + l)(r + l)...-l},
and
to this
for the

we must add required number

the one

way

JN x

N

/iV; thus

we obtain

\{(P + !)(</+ !)(<•+

l)- +

lj


344
414.
other.

-

;

HIGHER ALGEBRA.
To find
the

number of ways in which a composite
two factors which are prime
to

number can

be resolved into

each

= av b qc .... Of the two factors before, let the number one must contain ap for otherwise there would be some power of a in one factor and some power of a in the other factor, and thus

As

N

r

,

Similarly b q the two factors would not be prime to each other. must occur in one of the factors only ; and so on. Hence the required number is equal to the number of ways in which the product abc... can be resolved into two factors; that is, the

number
the

of

ways

is

-(1 + 1)(1 + 1)(1 +

1)...

or 2""

1 ,

where n

is

number
415.

of different
the

prime factors in N.
divisors of a number.
r

To find

sum of the

Let the number be denoted by ap b q c term of the product
(1
is

...,

as before.

Then each
r

+a +

2 a + ...+ar )(l+b + b +

2

...

+

b'

1

)

(1

+

c

+

c

2

+ ...+c

)...

a divisor, and therefore the product ) that is,
the

sum

of the divisors is equal to this

sum

required

=

a

.

_
a

i
*

&»+'_!

c
"

r+1

-1

b-l
33

-l c-1
52 = 2 3
(2

Example
Since

1.

Consider the number 21600.

21600 = 6 3

.

102 = 2 3
(5

.

.

22
(3

.

.

33

.

5 2,

the number of divisors =
..

+ 1)
.

+ 1)
1

+ 1) = 72
1

the

sum

of the divisors = — ? 2 — 1
.

...

,.

26-1

3*-l
5

o

-l —— — — —
53
.

5

= 63x40x31 = 78120.
Also 21600 can be resolved into two factors prime to each other in 2 3_1 or 4 ways.
,

Example

2.

If

n

is

odd shew that n

(n2

-

1) is divisible

by

24.

We have
Since n
is

n(n2 odd, n - 1
is divisible

l)

= {n7i

1)

(n+1).
;

one of them

and n+1 are two consecutive even numbers by 2 and the other by 4.
;

hence

is divisible

3,

Again n - 1, n, n + 1 are three consecutive numbers hence one of them by 3. Thus the given expression is divisible by the product of 2, and 4, that is, by 24.

THEORY OF NUMBERS.
Example
3.

34".

Find the highest power of 3 which

is

contained in
J

100.

Of the first 100 integers, as many are divisible by 3 as the number of times that 3 is contained in 100, that is, 33 and the integers are 3, G, 9,... 99.
;

Of these, some contain the factor 3 again, namely 9, 18, 27,... 99, and their number is the quotient of 100 divided by 9. Some again of these last integers contain the factor 3 a third time, namely 27, 54, 81, the number of

them being the quotient
factor 3 four times.

of 100 by 27.

One number

only, 81, contains the

Hence the highest power required = 33 + 11 + 3 + 1 = 48.
This example
article.
is

a particular case of the theorem investigated in the next

416.

To find
In.

the highest 'power

of a prime number a which

is

contained in

Let the greatest integer contained in -,
Cv

n — 2
iii

,

n —
tJ

...

respectively

Ct

CL

be denoted by /
1,2, 3,
...

(

-]

,

/(-,], /(-§),...

Then among the numbers

n. there are

/ ( - which
j

contain a at least once, namely

the numbers

a,

2a, 3a, 4a,

...

Similarly there are

I[-A

which

contain a 2 at least once, and I

(


g

)

which contain « 3 at

least once;
is

and so

on.

Hence the highest power

of a contained in \n

'©'©)*'6) + ~
In the remainder of this chapter we shall find venient to express a multiple of n by the symbol Jl(n).
417.
it

con-

418.
divisible

To prove
by
|r.

that the prodicct

of

r consecutive

integers is

Let
least of

P

n

which

stand for the product of r consecutive integers, the is n ; then

Pn = n(n+l)(n + 2)
and

...

(u + r-l),

\

Pn+l = (n+l)(n + 2)(n+3) nPm+i = (n + r) P = nP + rP
n

...(n+r);
;

H

.-.

1>

p -P =lsxr
= r times the product
of r

1

consecul ive integer-.

.

)

:

;

;

346

HIGHER ALGEBRA.
if

Hence
\r

the product of r

—1
m

consecutive integers

is

divisible

by

1,

we have

Pm+1 -P
Now
also P.
,

= rM(\r-l)

= M(\r).
P, =
|?',

and therefore

P
(r.

2

is

a multiple of

\r

\

therefore
if

P

,

.

.

.

are multiples of

We

have thus proved that
is

the product of

r— 1

consecutive integers

divisible

by

\r
;

1,

the

product of r consecutive integers is divisible by \r but the product of every two consecutive integers is divisible by 2
1

therefore the product of every three consecutive integers by 3 ; and so on generally.
1

is

divisible

This proposition

may

also be proved thus

is

that every prime factor contained in \n + r as often at least as it is contained in \n \r.
of Art. 416,

By means
This
419.

we can shew

we

leave as an exercise to the student.
is

If p

a prime number,

the coefficient

the expansion q/*(a

+

b) p

,

except the first
first

and

last,

of every term in is divisible by p.

"With the exception of the efficient of the form

and

last,

every term has a co-

p(p-l)(p-2)...(p-r +
'-

l)

where r may have any integral value not exceeding p — 1. Now this expression is an integer; also since p is prime no factor of r is a divisor of it, and since p is greater than r it cannot divide any factor of \r that -is, (p — 1) (p — 2)... (p - r + 1) must be Hence every coefficient except the first and divisible by |r. the last is divisible by p.
\

;

420.
(a

If p

is

a prime number,
...)p
;

to
1'

prove that
cp

+b+c+ d+

= a5 +

b +

+ dp +

.

.

.

+ M(p).

Write

ft

for b

+c+
p

. .

then by the preceding article

Again

J3

py = a* + p' + M(p). p = (b + c + d+ = (b + y) p suppose
(a +
. . .

= bp + y + M{p).

By

proceeding in this

way we may

establish the required result.

=

;

THEORY OF NUMBERS.
421 prime to
[Fermat'a Theorem.]
p, then
1

347

If p is a prime number and N"" -lis a multiple of p. +

N

is

We
let

have proved that
(a

+

b

+

c

d+

...y^a' + V+c* + d"+
...

...

+

M

each of the quantities «, 6, Cj 4 pose they are in number ; then

(p);
'

N

be equal to unity, and sun J P

But

,V

is

prime top, and therefore iV'- -

1 is

a multiple of p.

°°\ Si ^ is P ™°> P-li* an even number except when r />=J. lherefore
'

}

T

l

Hence
that
is

either

2^ +
= 7^ ±

1

or

S^ K

1

«

a multiple of ft
integer.

.V

-•

1,

where

is

some positive

422.

It should be noticed that in the course of Art. 421

it

tins result is

sometimes more useful than Fermat's theorem.
Shew
that n 7 - n
is

Example
Since 7 a 1S °
T Now

1.

divisible

by

42.

is

a prime,
n? -

n7 - n

M
;

(7)

n:=n

G

-l) =
|3

>i(n

+ l)(n-l)(n* + nS + l).
is divisible

(n

-

1)

n (n +

1) is divisible

by

hence n? - n

by 6 x 7, or

42.

Dowfra
Let
thatls

mXpleof^
.r,

iS pHme DU ber of^J; tSL* mimbeiS GXCeedS thGshew that the difference of the p" ^ 7 dlfference of the numbers

\

'

b/a

>

whence we obtain the required

y be the numbers then * p -x=M(p) and y»-y=M (p), *p -yp -(*-y)=^(p);
;

result.

Example
Tf

3.

Prove that every square number

is

of the form Sn or on

± 1.

v

V
-

iS
-

UOt

~l S£r^PTi?5n 1 is a multiple of 5 i n»- 1 or N*+l
eitner
l

m t0 ? L lG
r

5

x'J

e

have
iS
i*

AT2

= 5;i
;

where »
is,

w some

n ultipIe 0f 5
that

^

positive integer
;

Fermat '« theorem tf»=5w ± 1.

thus

348

HIGHER ALGEBRA.

EXAMPLES.
1.

XXX.

a.

i

Find the least multipliers of the numbers
3675,
4374,

18375,

74088
perfect squares.

respectively,
2.

which

will

make the products

Find the

least multipliers of the

numbers

7623,
respectively,
3.

109350,

539539

which

will

make

the products perfect cubes.

a?—y 2
4. is

is

If x and y are positive integers, divisible by 4.

aud

if

x—y

is

even,

shew that
square

Shew that the
If

difference

between any number and

its

even.
5.

-ix-y

is

a multiple of

3,

shew that 4x 2 + 7xy - 2y 2

is divisible

by

9. 6.

Find the number of divisors of 8064.
In how
?
'1

7.

many ways can

the

number 7056 be
by
15.
6.

resolved into

two factors
8. 9.

Prove that 2 4 Prove that n
(?i

1 is divisible is

+ 1) (n + 5)

a multiple of
its

Shew that every 10. the same remainder.
11. 12.
13.

number and

cube when divided by 6 leave

If

n

is

even,

shew that n (;i 2 + 20)
(Sn + 2)
2,

is divisible

by

48.

Shew that n (?i 2 - 1)
If

is divisible

by

24.
is

n

is

greater than

shew that n 5 — 5n 3 -f 4?i

divisible

by

120.
14. 15.

Prove that 3 2n + 7
If

is

a multiple of

8.

n

is

a prime

number

greater than

3,

shew that

?i

2

-

1

is

a multiple of
16.

24.
is divisible

Shew that n5 — n
n
is

by 30

for all values of n,

and by

240

if

odd.
difference of the squares of
is divisible

17.

Shew that the
6

any two prime

numbers greater than
18.

by

24.
is

Shew that no square number

of the form

3?i

— 1.
or

19.

Shew that every cube number

is

of the form

9?i

9n±L

THEORY OF NUMBERS.
20.
is 0, 1

349
7,

Shew that
or
6.

if

a cube

number

is

divided by

the remainder

21.

form 7n
22.
23.

If a number is both square or 7?t+l.

and cube, shew that

it is

of the

Shew
If 2»

that no triangular
4- 1

number can be

of the form 3u
l
2
,

-

1.

a prime number, shew that divided by 2>i+l leave different remainders.
is

2 2 , 3 2 ,...n 2

when

24. may be. 25.

Shew

that ax

+ a and

a* - a are always even, whatever a and

x

Prove that every even power of every odd number

is

of the

form 8r + l.
26.

Prove that the 12 th power of any number

is

of the form I3)i

or 13ft+l.
27.

Prove that the 8 th power of any number

is

of the form I7n

or

I7n±l.
prime number greater than
5,

If n is a 28. divisible by 240
29.

shew that n 4 —

1

is

n —
G

If n is any prime 1 is divisible by 168.

number

greater than

3,

except

7,

shew that

30.

Show

that

?i

36

-

1 is divisible

by 33744

if

n

is

prime to

2, 3,

19

and

37.

When p + l and 2p + l are both 31. «** — 1 is divisible by 8(p l)(2/) + l), if

+

prime numbers, shew that x is prime to 2, £> + l,and

2p+h
If 32. divisible by
33.

p

p

is r .

pV p1 a prime, and X prime to p, shew that x ~

-

1

is

If

m

is

a prime number, and

a, b

two numbers

less

than m,

prove that

am
is

-2

+ am ~ 3 b + am - 4 b'i +

+ bm ~ 2

a multiple of m.

any number, then any other number may be expressed in the form = aq + r, where q is the integral quotient when is divided by a, and r is a remainder less than a. The number a, to which the other is referred, is sometimes called the modulus ; and to any given modulus a there are a different
423.
If

a

is

N

N

N

:

;

;

350
forms
of a

HIGHER ALGEBRA.

number iV, each form corresponding to a different value of r. Thus to modulus 3, we have numbers of the form 3q + 2; or, more simply, 3q, 3q±l, since 3q + 2 is 3<7, 3q + l, In like manner to modulus 5 any numbe^ equal to 3 (q+ 1) - 1.
will be

one of the

five

forms

5q,

5q

±

1,

5q

±

2.

424. If 6, c are two integers, which when divided by a leave the same remainder, they are said to be congruent with respect to the modulus a. In this case b — c is a multiple of a, and following the notation of Gauss we shall sometimes express this as follows b = c (mod. a), or b - c=0 (mod. a).

Either of these formulae
425.

is

called a congruence.

pb and

are congruent with respect pc are congruent, p being any integer.

If

b, c

to

modulus

a,

then

For, by supposition, b - c = ?za, where w is some therefore ])b — pc — pna ; which proves the proposition.
426.

integer

If a

is

prime
a,

to

b,

and

the quantities

2a, 3a,

(b

1

)

a

are divided by b, the remainders are all different.

For

if

possible, suppose that

ma when
then

divided by b

two of the quantities ma and leave the same remainder r, so that
m'a = q'b + r
;

ma = qb + r,

(m - 7/1') a = (q-q')b

— m', since therefore b divides (m — m') a ; hence it must divide it is prime to a ; but this is impossible since and m' are each

m

m

less

than

b.

Thus the remainders are
quantities
order.
is

all different,
b,

exactly divisible by
1, 2, 3,

and since none of the the remainders must be the
but not necessarily in this

terms of the series

b

1,

Cor. of the a.

If
p.

a

is

prime to
c

b,

and

c is

any number, the
+
(b

b

terms

c,

+

a,

c

+

2a,

c

1) a,

.

THEORY OP NUMBERS.
when

351

divided by b will leave the same remainders as the terms
c,

of the series

c+

1,

c+

2,

c

+ (b;

1),

though not necessarily in mainders will be 0, 1, 2,
427.
-(/"b., b.j

this

order
1.

and therefore the

re-

b-

are respectively congruent to c n c„, c wn7A regard to modulus a, £/te?i //te products b,baba ..., ^c.c.^ (o-tf also congruent.
,

b3

...

,

...
...

}

For by supposition,
b
1

-c = n
l

x

a,

b2

-c = u
2

2

a,

b

3

-c

:i

u./i,

...

where n lt n 2
.

,

n3

...

are integers;
=(<?!

•.

bx baba

...

+
2

»,») (ca
...

+

=

c,c c

3

+

M

rc «) 2

(ca

+w

a)
:j

.

.

(a),

which proves the proposition.
428.

We
be

can

now

give another proof of Fermat's Theorem.

If p

a prime number and

N

prime

to p,

then

N

1

'

-1

1

is

a multiple of p.
Since JV and

p

are prime to each other, the numbers
2tf,
3.V,

if,

(p-l)iV
(p-1)

(1),

when divided by p

leave the remainders
1,

2,

3,

(2),

though not necessarily in this order. Therefore the product of all the terms in (1) is congruent to the product of all the terms
in (2),

p being the modulus.
i

That is, |^—1 N''~ and divided by p ; hence

\p

-

1

leave the

same remainder when

but i^—l

is

prime to p

;

therefore
1

it

follows that

JP" 429.

1

=

M

(j>).

denote the number of integers less than a number a and prime to it by the symbol <f> (a) ; thus <f>(2) = 1 ; <£(13) = 12; </>(18) = G; the integers less than 18 and prime to It will be seen that we here it being 1, 5, 7, 11, 13, 17. consider unity as prime to all numbers.
shall

We

,;

352

HIGHER ALGEBRA.
a,

To shew that if the numbers 430. each other,
(f>

b,

c,

d,

...

are prime to

(abccl

.

.

.)

= <£ (a)
;

.

</>

(b)

.

<£ (c)

.

.

.

.

Consider the product ab then the first ab numbers can be written in b lines, each line containing a numbers ; thus
1,

2,

k,

a,

a+l,
2a +1,

a+

2, 2,

a+

k, k,

a+
2a +
(b

a, a,

2a +

2a +

(&_

\)

a+

1,

(6-

1)

a+

2,

...

(b-l)a +

k,

...

-

1)

a+

a.

Let us consider the vertical column which begins with h ; if k is prime to a all the terms of this column will be prime to a but if k and a have a common divisor, no number in the column Now the first row contains <£ (a) numbers will be prime to a. prime to a \ therefore there are <£ (a) vertical columns in each let us suppose that the of which every term is prime to a vertical column which begins with k is one of these. This column is an A. p., the terms of which when divided by b leave remainders 0, 1, 2, 3, ... 6 — 1 [Art. 426 Cor.]; hence the column contains to b. <£ (b) integers prime
;

Similarly, each of the term is prime to a contain
.

</>

table there are <f> (a) cj> (b) also to by and therefore to ab

columns in which every hence in the (b) integers prime to b integers which are prime to a and
(a) vertical
; ;

that
(a) (a)
(a)
.

is

(ab)
...)

=
=

<£ (6).

Therefore

cf>

(abed

<f>

.

<j>

(bed
.

. .

.)

cj>

.

(f)(b)

<j>

(cd

...)

=

<f>(a).<t>(b).<t>(c).<}>(d)....

431. To find the number given number, and prime to it.

of positive integers

less

than a
r

Let JV denote the number, and suppose that JV = ap bq c ... where a, b, c, ... are different prime numbers, and p, q, r ... Consider the factor a ; of the natural numpositive integers. p p — the only ones not prime to a are bers 1, 2, 3, ... a 1, a
1
'

,

a,

2a,

3a,

...

(a*-

1

-

I) a,

(a1

"

1

)

a,

.

;

THEORY OF NUMBERS.
and the number
of these
(av )
i

353

is a''~

;

hence

4>

= a" ,

a'-' c\

=a?(l-

-^

.

Now
.

all
</>

the factors ap
(a)Vc
r
.

b'\

...

are prime to each other
1

\

.

.)

=

<j>

(a

1

')

.

</>

(b

)

.

(c

r

)

.

.

-H)-H)-H)^••K)H)H)that
is,

^
to
it is

W

= iir(i-i)(i-J)(i-I)....
of all the integers

Example.

and prime

Shew that the sum ^N<p (N).

which are
then

less

than

N
an

If x is any integer less than integer less than and prime to it.

N

N

and prime and

to

it,

N-x

is also

Denote the integers by

1,

p, q,

r, ...

,

their

sum by

S; then

S = l+p + q + r+... + (N-r) + {N-q) + (N-p) + {N-l),
the series consisting of

(N) terms.

Writing the series in the reverse order,

S = {N-l) + (N-p) + (N-q) + (N-r)+...+r + q+p + l;
.-.

by addition,

2S = N + N + N+
.-.

...

to

<p

(N) terms;

S = $N<p(N).

From the last article it follows that the 432. integers which are less than J¥ and not prime to it is
:

number

of

'-'(.4>(>-i)(» 3("i)-'
tli at is,

N N N N N _++_+..._ ao ac N a be b
c

. . .

+

N+
abc

....

Here the term

N —

gives the

number
...

of the integers

a,

la, da,

— .a
(t

which contain
H.
II.

a as a factor; the term —= gives the number of
ao

N

A.

23

354

HIGHER ALGEBRA.

the integers ab, 2ab, Sab,

-j ab, which contain ab as a factor, ao and so on. Further, every integer is reckoned once, and once only ; thus, each multiple of ab will appear once among the multiples of a, once among the multiples of b, and once negatively among the multiples of ab, and is thus reckoned once only. iV Again, each multiple of abc will appear among the j- ,
...
,

N

.

.

N — N — a o c
;

terms which are multiples of a, b, c respectively; among the iV JV =- terms which are multiples of ab, ac, be respectively r r J

& — —
,

,

ab'

ac'

be

'

'

and among the -j- multiples

of abc; that

is,

since

3-3+1

=

1,

each multiple of abc occurs once, and once only. cases may be discussed.

Similarly, other

433.
»

[Wilson's Theorem.]

If -p

be

a prime number,

1

+ \p -

1

is divisible

by
2,

p.

By
Ijp-1

Ex.

Art. 314
i)""
1

we have
i)

= (P~

- (P -

(P -

2r

'

+ _

^z^ipD (p - $y->
irl+
tlie
1

Jp-l)(p

2)(p-3)
{p

top _ lterms
(j)

.

and by Fermat's Theorem each of (p-2) p ~\ (p-2>y~\ ... is of the form

expressions

-

l) p

~

l

,

+M(p)-, thus
2)

p-l = M(p) +fl-(p-l) + (P~
= M(p) —
Therefore
1

l

)(P~

-...top-

I

terms!

=M(p) + {(i-iy->-(-iy->}
1,

since
1

p—
=

1 is

even.

+ pI

M

(p).

This theorem is only true when p is prime. For suppose p has a factor q; then q is less than p and must divide \p — 1 hence
;

1

+

\p

1 is

not a multiple of

q,

and therefore not a multiple

of p.

Wilson's Theorem may also be proved without using the result quoted from Art. 314, as in the following article.

.

.

1

THEORY OF NUMBERS.
434.

,'}:>5

[Wilson's Theorem.]

If p

be

a prime number,

1

+ lp—

is divisible

by

p.

Let a denote any one of the numbers
1,

2,

3,
if

4

5

...

(p-1)
(^;

(1),

then a
are

is

prime to p, and

the products
3. a,

\.a, 2. a,

1 )

a
leaves

divided by p, one and only one of them mainder 1. [Art. 426.]

the re-

ma; then we can shew that the numbers m and a are different unless a=j)~ 1 or 1- For if a were to give remainder 1 on division by^>, we should have (mod. p), a~ - 1 =
Let
this be the product
2

and since p is prime, this can only be the case when a + or a — 1 — 0; that is, when a=p— 1 or 1.
;

1

- p,

Hence one and only one of the products 2a, 3a, ... (p — 2) a gives remainder 1 when divided by p that is, for any one of the series of numbers in (1), excluding the first and last, it is
possible to find one other, such that the product of the pair

the form

M

is

of

(p)

+

1

Therefore the integers 2, 3, 4, ... (p-2), the number of which is even, can be associated in pairs such that the product of each pair is of the form (j?) + 1

M

Therefore by multiplying

all

these pairs together,

we have

2.3.4
thatis,

...

(p-2) = M(p) + l;

1.2.3.4
\p

...

whence
or
1

-

1

(p-l) = (p-l){M(p) + = (p) +p - 1

l} ;

M

j

+

1^

1

is

a multiple of p.

Cor.

If 2p
l.

+

l

is

a prime

number
is

/jp\

2

+

(-

iy

is

divisible

by 2p +

For by Wilson's Theorem 1 + \2p n = 2p + 1, so that p+ 1 = n —p then
;

divisible

by 2p +

1

.

Put

\2p

= 1.2.3.4 p(p+l)(p + 2) (n-1) = 1 (w-1) 2(n-2) 3(»-3) ... p (n-p) = a multiple of n + (- iy (\p) 2
.

Therefore

1
2

+

(

-

l) p (\p)
is

2

is

divisible

by n or 2p +

1,

and

therefore (|^)

+ (— l/

divisible

by 2;;+l.

23—2

;

:

356
435.

HIGHER ALGEBRA.

Many
1
.

theorems relating to the properties of numbers

can be proved by induction.
Example
If

p

is

a prime number, x p - x
;

is divisible

by p.

Let x p - x be denoted by f(x)

then

/ (x + 1) -/ (x) = {x + 1)p - (x + 1) - {xP - x) 1) " =px p ~ l + P< f~ * p 2 + +px
. .

J.

a

.

=a
.-.

multiple of p,

if

p

is

prime

[Art. 419.]

f(x + 1) =f(x) + a multiple of p.
by^i, so also is/(.r
2*5

If

therefore/^)

is divisible

+ l);

but

/(2) =

~ 2 = (1 + 1)^-2,
[Art. 419]
;

and this

is

a multiple of p
is

when p is prime

therefore / (3)

is divisible is

by p, therefore /(4)
universally.

divisible

by^, and so on; thus the proposition
if

true

it

~ follows that x p J

This furnishes another proof of Fermat's theorem, for

x

is

prime to

p,

-

1 is a multiple of p.
24/i

Example

2.

Prove that 5 2,l+ 2 24?i

- 25

is

divisible

by 576.

Let 5 2n+2 -

- 25 be denoted by f(n)
/(?i+l) = 5 2n+4 -24(w + l)-25

then

= 5 2 .5 2w+2 -24n-49;
.-.

f(n+l) - 25/ (n) =25

(24n + 25) - 24u - 49

= 576 (n + 1).
Therefore if f(n) is divisible by 576, so also is /(u + 1); but by trial we see that the theorem is true when n = l, therefore it is true when n=2, therefore it is true when ?i = 3, and so on; thus it is true universally.

The above

result

may

also be proved as follows

52n+2

_ 2in - 25 = 25 M+ 1 -

- 25 = 25 (1 + 24)" -24rc-25
24;i

= 25 + 25 n 24 + M (24 2 = 576n + iW(576)
. .

)

- 24n - 25

= i)/(576).

EXAMPLES.
1.

XXX.
divisible

b.

Shew that 10 n + 3

.

4" +
.

2

+5

is is

by

9.

2. 3.

Shew
Shew

that 2
.

.

7n

+3

5H

-5

a multiple of 24.
9.

that 4 6 n 8
.

+ 5 n + x when divided by 20 leaves remainder

4.

Shew that

7n

+ 4" + 2

is

of the form 24 (2r -

1).

b

THEORY OF NUMBERS.
5.

357

If

p

is

prime, shew that 2
l

\p-3 + l

is

a multiple of p.

6.

Shew Shew

that a v, +

-a

is

divisible 1>y 30.

7.

that the highest power of

2

contained in

2r

1

is

2''-;--l.
8.

Shew

that 3 4 + - + 5 2 +* is a multiple of 14.
'1
'1

9.

Shew that

3** +6 +160»a

- 56n- 243

is divisible l»y

512.

Prove that the sum of the coefficients of the odd powers of x 10. 2 in the expansion of (l+ # 3 + .r 4 ) n "" 1 , when n is a prime number other than 5, is divisible by n.
<

r+# +

If n is a 11. divisible by 504.
12.

prime number greater than

7,

shew that n°-l

is

If

n

is

an odd number, prove that

5

?i*

+ 3>i 4 + 7>i 2 - 11

is

a

multiple of 128.
a prime number, shew that the coefficients of the terms of (H-a?)*-* are alternately greater and less by unity than some multiple of p.
13.

If

p

is

a prime, shew that the sum of the (p-l) th powers of any p numbers in arithmetical progression, wherein the common difference is not divisible by p, is less by 1 than a multiple of p.
14.

If

p

is

15.

Shew

that a 12 -

12

is divisible

by

91, if

a and

b are

both prime

to 91.
16.

If

p

is

a prime, shew that \p

-2r
1

2r

-

1

-1

is divisible

by p.

If 17. 1, n + 1 are 2 that n(?i -4) is divisible

n—

both prime numbers greater than 5, shew by 120, and ?i 2 (>i 2 + 16) by 720. Also shew

that n must be of the form 30£ or 30^ + 12.
18.
,

Shew
.

that the highest power of n which

is

contained in \nr ~

1

is

equal to

n r — nr + r-

1
.

n- 1
p

19.

If
c2
1 is

number
l(p-D

is a prime number, and a prime to ]), and if a square can be found such that c 2 — a is divisible by pt shew that

a2

20.

divisible

by p.

Find the general solution of the congruence
98a;-

1=0

(mod. 139).


358
21.

HIGHER ALGEBRA.

Shew that the sum of the squares and prime to it is a given number

N

of all the

numbers

less

than

?(i--30-J)0-9- + ?ci-.»a-.)a-*-.
and the sum of the cubes
is

?(i-3(x-])(t-9- + ?a-^ci-«a-*.,
a,b,c... being the different prime factors of
22.
iV.

If

divisible

and q are any two positive by (|£>)«. |# and by (\q) p \p.
jt?
.

integers,

shew that \pq

is

23.

Shew that the square numbers which

given by

are also triangular are the squares of the coefficients of the powers of x in the exL

pansion of

an<^ ^hat the square numbers which are also r— \)X -f" X2> pentagonal by the coefficients of the powers of x in the expansion of

1_

Shew that the sum of the fourth 24. and prime to it is less than

N

powers of

all

the numbers

5

\

a

-gg(l-<*)(l-i»)(l -«*)...,
a, 6, c,...

being the different prime factors of
If
(iV) is
if

A".

25.

(f>

the
is

number
prime to

of integers
JV,
1

which are

less

than JV and

prime to

it,

and

x

shew that
(mod.
JV).

^^26.

=

If

dv d2 ds
,

,

...

denote the divisors of a number

JV,

then

(f>

(dj

+

(c? ) 2

+ <£ (d3 + ... =iV.
)

Shew
<t>

also that

(!)

r-

;

—9~0( 3 r — + 0( 5 )T^
1
)
;

fi

™-

---

odinf. = -?

~—kJ.
l

.;

*

CHAPTER XXXI.

The General Theory of Continued Fractions.
*-436.

In Chap. xxv. we have investigated the properties of
a,

Continued Fractions of the form

+

-

«,+

%+

-

.

.

.

,

where a2 a 3
,
'

, . . '

are positive integers, and a^ is either a positive integer or zero. shall now consider continued fractions of a more general

We

type.
*-i37.

The most general form
'

of

a continued fraction

is

~' a
1

=*=

„ * rZl « 2 "i a3

=*=

where a i> a2> a3>

••> *,»

KK

•••

represent

any quantities whatever.

The

fractions


a,

,

shall confine our attention to two cases; continued fraction. (i) that in which the sign before each component is positive (ii) that in which the sign is negative.

We

a2

,

a3

,

.

.

.

are called components of the

*438.

To

investigate the

law of formation of

the successive

convergents to the continued fraction

b
i

b2

b3

The

first
6,

three convergents are ergents

a2 b

x

a 3 .a 2 b

i

+b 3 .b

l

AVe see that the numerator of the third convergent may be formed by multiplying the numerator of the second convergent by a 3 and the numerator of the first by b 3 and adding the results also that the denominator may be formed in like together manner.
,

;

n

;

360

HIGHER ALGEBRA.

Suppose that the successive convergents are formed in a similar way; let the numerators be denoted by p it p 2 p 3 ..., and the denominators by q lt q 2 q3 ...
,
, ,

Assume
vergent
;

that the law that is, suppose

of

formation holds for the nth con-

p In

-anlp

n—\
,

+

b n-l n — 2> p „,
differs
;

q
J-

=a q

n-Ln —

,

1

+

bq _. h-Ih—2
only in having

The (n+l) th convergent
a
h

from the

wth

—— a
n+
\

in the place of a

hence

the (n+ l) tb convergent

If therefore
?>

we put

^,=a

nP +b

,,p

,,

q

+

,

=a

+,q

+b

,,q

,,

numerator and denominator of the (u + l) th convergent follow the law which was supposed to hold in case of the th But the law does hold in the case of the third convergent ?t hence it holds for the fourth and so on therefore it holds

we

see that the

.

;

;

universally.
*-439.

In the case of the continued fraction
b,

b2

b3
CC

«1

-

a2 ~

~
3

we may prove
Vn

that

= anPn-l ~ kPn-*
.

,

Qn

=

»«?«-!

~

k<ln-2

',

a result which may be deduced from that of the preceding article by changing the sign of b n
*440.

In the continued fraction

h
a +
1

K
a2 +

K
«3 +

we have

seen that
J-

p n =a nl p

ii

—\

,+bp
n-L

„, n — 2'

q J-n

^an-ln — ,+bq q n-J-n—2
1

a.

GENERAL THEOHY OF CONTINUED FRACTIONS.

Ml

?„ + ,

\9n

Qn-J'

but

and
less

is

therefore a proper fraction: hence

"

+

1


ft

is

numerically

&.+1

than

— — ——
In

,

and

is

of opposite sign.

In-}

reasoning as in Art. 335, we may shew that every convergent of an odd order is greater than the continued fraction, and every convergent of an even order is less than the continued fraction ; hence every convergent of an odd order is greater than every convergent of an even order.

By

Thus 2--'^ 32/1+]

^
9an
1

is

positive

and
1
2/1

less

than £ss=l 2ft*-J

^

2

" ;

hence

2*2/1

2 2

2/1

+

1

1

2/1

+

1

2 2/1-1

Also ?*=i 22/1 —
1

2 --- is
22/1

positive

and
2
2/1

less

than

^=* - &=s
22/1 — 1

hence

22/1-2

2 2

2/1

—2
—2
all

2/1

2

2/1

greater than the continued fraction but continually decrease, and the convergents of an even order are all less than the continued fraction but continually increase.
of

Hence the convergents

an odd order are

Suppose now that the number of components is infinite, then the convergents of an odd order must tend to some finite limit, and the convergents of an even order must also tend to some
these limits are equal the continued fraction tends to one definite limit ; if they are not equal, the odd convergents tend to one limit, and the even convergents tend to a different limit, and the continued fraction may be said to be oscillating; in this case the continued fraction is the symbolical representation of two quantities, one of which is the limit of the odd, and the other that of the even convergents.
finite limit
;

if

a

.

362
*441.

HIGHER ALGEBRA.
To shew
that the continued fraction
n+I

a +
x

a 2 + a3

+
is

has a definite value if the limit of
greater than zero.

rf""

when n

is infinite

The continued

fraction will have a definite value

when n

is

infinite if the difference of the limits of

-^
9n+l

and

— is equal to zero.

Now
whence we obtain
Pjt+X
<2n+l

n+lffn-l

fPn

_ Pn-1
£«-!>

?«+!

_ £» =

(_

1

)-l 6 "+^»-l & ng»-3

KV* KVl (P* _ Pl\

But

k.-i?

k.-M? n+lin-1
li+1 !7u

+

^B+lS'n-l

an+l Qn

k^g
an(

+

J

an+l q n
"n+lSn-l

a.

i

K?-i + &«£«-*) _ «*»«+!
^n+lSn-l
^n+1

+

^n+A^-2 b^ q
x

K

also neither of these
"

terms can be negative; hence
is

if

the limit of
;

n+1

is

greater than zero so also

the limit of
1

"+

in

which
£a

case the limit of

-J**^
Qn+i

is

less
m

than
t

:

and therefore ^±i_ -^
q„+i

qn

the limit of the product of an infinite

number
that
is,

of proper fractions,

and must therefore be equal to zero
the same limit
.For
;

:

-^
?«+:

and

-- tend to

qn

which proves the proposition.

example, in the continued fraction

V
a.
1

T-

3-

n~

3+5
Lim -f-^- = Lim*«+i
.

2n+l +
\, (n+iy
7 2

»

-'

=4

:

and therefore the continued fraction tends to a

definite limit.

;

GENERAL THEORY OF CONTINUED FRACTIONS.
*442.

3G3

In

the

continued fraction
c\i,
1

cl

2

H;„ A

if the denominator of every component exceeds the numerator by unity at least, the convergents are positive fractions in ascending order of magnitude.

in

a2 a ? each of which the denominator exceeds the numerator by

By

supposition


a

'

,

-*

,

3
,
.
. .

are positive proper fractions

i

unity at

least.

The second convergent

is

1

,

and since a

i

a
exceeds
tliat «,

--

2

ft
t

by unity at
2

least,

and
ft,;

-* is

a proper fraction,
is,

it

follows

is

greater than

that

the second convergent
it

is

a positive proper fraction.
that
2
,

In like manner
;

may be shewn
it

is

a positive proper fraction

denote

by f, then

the third convergent
fraction.

is


ft,

!

-.

,

and

is

therefore a positive proper

Similarly
;

we may shew

that

—— —— —
K
a,

is

a positive

proper fraction

hence also the fourth convergent

K
CV
2

K

«1
is

~

%~
2
ii

a positive proper fraction

;

and

so on.

Again, O

'

i

p

ii

— anl p
^
2

ii

,

ft

1

2' »/ n — .,

p

q—aa

iuii-

,

ft

I

q 2 niii- „

'

;

hence ^s±J -

^
<?i

and

;

- - ^=*
n
1

have
h

tl ie

same

sign.

But P °~

^=
,

g

«\
2
,

- - =
r 'i

-^
2

,

and

is

therefore positive

92

<V* 2 -ft 2
(

<M
?3
;

hence ^ 2
9jt

>^ ^ > ^7, l,

^>^
<74

and so on; which proves the

7,

%

proposition.

;

.

;

,

364

HIGHER ALGEBRA.

If the number of the components is infinite, the conCor. vergents form an infinite series of proper fractions in ascending order of magnitude ; and in this case the continued fraction must tend to a definite limit which cannot exceed unity.

*443.

From
Pn =

the formula
<*nPn-i

+ hPn-t*

9n

= a&n-, +

&.&_«>

we may always determine in succession as many of the convergent^ as we please. In certain cases, however, a general
expression can be found for the
Example.
11

th

convergent.
c c

c

To

find the

w th convergent
2
;

to -

5- 5- 5series

We have p n = Sp,^ - 6p n _
any three consecutive terms

of

hence the numerators form a recurring which are connected by the relation

Pn " 5p«-! + §Pn-2'
Let
then, as in Art. 325,

S =p1 +p.2x +Prf? +
we have

• • •

~ +p nx n 1 +
5Pl)
2

.

.

S= Pl1+-(P2~ 5x
,

X

+ 6a;

But the

first

two convergents are 6
1

=-^

18

12
1

- hx + Qx.

l-3x
.

- 2x

'

whence
Similarly
if

p n = 18
S'
x

3"- 1 - 12

2' 1

"1 =

6 (3" - 2").
. .
.

= q + q& + q 3x* +...+ q nxn ~ x +

we

find

^=
gn

___ = ___ _
i

r

;J

whence

=9

.

3*- 1 - 4

.

2*- 1

= S'^ 1 - 2' +
l
'

]
.

w y w _ 6(3"-2 )
"""

ffn

~ 3n+1 - 2?l+1

This method will only succeed when a„ and b n are constant Thus in the case of the continued fraction for all values of n.

a+

a+ a+

...

,

we may shew

that the

numerators of the

successive convergents are the coefficients of the powers of

x in

the expansion of

^

7-2

,

and the denominators are the
ft
-I-

ti /Y*

coefficients of the powers of x in the expansion of

^

1 ~~ OjX

7-2 — ox

.

1

GENERAL THEORY OF CONTINUED FRACTIONS.

305

*444. For the investigation of the general values of pn and qn the student is referred to works on Finite Differences ; it is only The in special cases that these values can be found by Algebra.

following method will sometimes be found useful.
Example.

Find the value

of

3 -— =— 5— o 1 z

12
+

+

+

The same law of formation holds for p n and u n = nun _ x + nu n _ 2 either of them then
; ,

qn

;

let

us

tal<e

i*

n to denote

or
Similarly,

un -

(n

+ 1) «„_! = tt

(uB_, -

?w n _ 2 ).

i^j - RU _s = - (« u _ 2 -ra- 1 « n_ 3 ).

whence by multiplication, we obtain
u n -(n + l)u n _ 1 =

(-iy^(u 2 -3u
;

i

).

The

first

two convergents are 1

1

2
,

T -A

hence
qn

n p n -(n + l) Pn ^=(-l) -\

-(n + l)q n -i = (- I)""2
0n
gn-1
to

-

Tims

^n
7i

Pn-1

(" !)

?l-l

+l

m
|n-l

_

2 (~ I)""

lra+1
l)"- 8
In

iw+l
ffn-1

+l

'

At! _ Art = (~
In

In

9»-2 n -

In

Ps_Pi
13
|2'

?3

_

<h

1
[3'

3

2

1
2

3i
'2

2'

2

|2'

whence, by addition
|n

+l
7a

+
12
|3

+
|4

Lit n+1
j

1
. '

n+1
By making n
infinite,

= 1-7^+1i- "77 +
|2
|3

,

1

1

1

,

(- l) n

~a

|1

n+1

we obtain
e)

2n

e

e-V

which

is

therefore the value of the given expression

*


HIGHER ALGEBRA.
*445.

;

366

If every component of

—— —— a — — + a + +
3j
j

...

is

a proper

o

3

fraction with integral numerator fraction is incommensurable.

and denominator,

the continued

For

if

possible, suppose that the continued fraction is
-^ XL
,

com-

mensurable and equal to
then
-7

where A and

B

are positive integers

=
...
;

*—m

,

where

f,

denotes the infinite continued fraction
*—-

^-

hence

f=
,

-

=^

suppose.

Now
infinite

A, B, «x 6 X
,

are integers and Similarly -= J

B

C=

a2


b

f

is positive,

therefore

C

is

a positive integer.
continued

+f
. .

*-=

where fa denotes the
hence

fraction

—*- —*a3 + a +
.

.

;

/

=

—S^
,

*

=
;

7*

suppose
on.

;

and as

before, it follows that
.

D
,

is

a positive integer
are

and so
,
;

Again,

B
-7
,

C =

D
,

jy

...

proper tractions

..

tor -j is less

5

.

,

than
less

ax
than

,

which

is

a proper fraction
so on.

:

B

=

is

less

than -* as

;

-^

O

is

;

and

form an infinite series of positive integers which is absurd. Hence the in descending order of magnitude given fraction cannot be commensurable.

Thus A, B, C, D,

...

;

holds if some of the components are not proper fractions, provided that from and after a fixed component all the others are proper fractions.

The above

result

still

For suppose that
proper fractions
;


n

and

all

the succeeding components are
just proved, the infinite conis

thus, as

we have
an

tinued fraction beginning with -s
it

incommensurable

;

denote

7)

by

k,

then the complete quotient corresponding to
of the continued fraction
is

—n
in

k
is

^
.

and therefore the value

^-^
9n-l

/ —hn-2 +

"~ 2

GENERAL THEORY OF CONTINUED FRACTIONS.
This cannot be commensurable unless

3G7
this

V n_1 —
tfn-l

P -^^ =l
?«-«

and

condition cannot hold unless ?2=a = *-=2 , Pn= B ?n-a ?n-3 Qn-z
finally

= P^*
9n-4
;

t

;ultl

P — H= P

;

that

is

£>6

=

0,

which

is

impossible

hence the

given fraction must be incommensurable.
*446.

%

.

?,

//* eirary

component of

K —- — — a a a
1
1

«-

*""

...

?'s

i

""

a

~

a proper

3

fraction with integral numerator and denominator, and if the value of the infinite continued fraction beginning with any component is less than unity, the fraction is incommensurable.

The demonstration

is

similar to that of the preceding article.

*
1.

EXAMPLES. XXXI.

a.

Shew that

in the continued fraction

\
ax

h_ - a2 - a 3 J

_h

'

Pn = a nPn - 1 ~ ^nPn - 2
2.

Convert

rators.

"

m
'^|

Qn = a n9.n -\~

"?i?n

-2

)

into a continued fraction with unit

nume-

3.

Shew

that

« V*+6=«+^
(2 )

^
±r.
.......

V^=«-

2 |_

4.

In the continued fraction

«1~ a%~ a3~

—— —— ——

.... if

the denominator
least,

of every

component exceed the numerator by unity at
increase with n.

shew that

p n and qn
5.

If a lf a,, rtg,..^,, are in

harmonical progression, shew that
1 1 1

^"2^

an

1

«2

2^

2~="

2^

SJ'

.

1

368
6.

HIGHER ALGEBRA.
Shew that
cc+
V
-

2a+ 2a +
:

—- -—)

...

+
J

[a V
1

2a- 2a1

...

= 2a
J

2
,

and
7.

(

a

— ... \/ a +s 2«+ 2a+"7\
(

-

7T2a-

o^--)= a 2a- "7
b

\

,
~

2a2 -

2a'2

In the continued fraction
b
b

a+ a+ a+
shew that
8.

pn + = 6an
x

,

bq n + 1 - apn +
ax

=
1

%

n_

x

.

Shew that

— a+
b

b
-

b

a+ a+

= 6.-^7 — ^n x+ x+
a
1

—p

— 8P

i

.v

being the

number

of components,

and

a, /3

the roots of the equation

k2 — ak — b = 0.
9.

Prove that the product of the continued fractions

J_ L _L A_ 6+ c+ d+ a+
is

'"'

,7

_1_ -c +

-6+ -a+

J_

J_

-0?+

_x _

'"'

equal to

1

Shew that
1

4

9

64

(?i

2

-l) 2

(n

+ l)(w + 2)(2/i + 3)
6

10.

1- 5- 13- 25-

»2 +(»+l) 2

11.

-L JL _§_ 1- 5- 7-

^2 ~1 _ ?*fo+ 3 ) 2» + l~ 2

'

12.

£- 3- ji24-

13

.

i !> «* 1- 3- 4- 54
6

?i

£±L 5±|- 1+ i+i« + |. + ...+|«. I— L_ + l— + 2
?i
!

-2=1 w+l2/i

=.-1.
2(e2 -l)
e2

8

14.

1+ 2+ 3+
3.3

+2 n+
3(n + 2)

+l
_'6(2 e3+l)
"

3.4 3.5

15.

1+
If u, 1

2+

3+
Ucy
a

n+
a+
f
,

5e 3

-2

a + 26 bein» formed by taking the denominator and the sum of the numerator and denominator of the preceding fraction for its numerator and denomi,

16.

=

v,

=

Uo

=

—f

,

each successive fraction

e>

•*

nator respectively, shew that u „=**—=

.

CONVERSION OF SERIES INTO CONTINUED FRACTIONS. 309
17.

Prove that the n ih convergent to the continued fraction
J'

f

)'

y'H

+1

y.
'

r+l- r+l- r+l18.

is

»* +1 -l

Find the value of
positive

cij

+ l-

\

a.,+

%l-

a 3 +l-

%

'

a19 a2 a 3v being
>

and greater than unity.

19.

Shew that the n lh convergent
1) ;
lh

to

1

--

-

is

equal to

the (2m — v

convergent to 6

— -— — 2 + 1+2+1+ -—
n

,

20.

Shew that the 3n th convergent

1111111
1

to

n
is
3>i

5- 2- 1- 5- 2- 1- 521.

+l

Shew that

2

3

2+ 3+ 4+

3-d e-2
.

'

hence shew that

e lies

between 2§ and 2 r8T

Conversion of Series into Continued Fractions.
*417.
It will be convenient here to write the series in the
1 1 — +—
^6
1

form

u2

+
1

u3

1 — +..

1

Put

370
Similarly,

HIGHER ALGEBRA.

11111 — — — —=—
u
+
x

u2

+

u3

u

us + x2 u u UQ2 u - u + u2 — u2 + u3
x x

12
+
x

=

1

u —+— + x u u
2
y

x

2

a

'

x

x

and

so

on
1

;

hence generally
1 1

CONVERSION OF SERIES INTO CONTINUED FRACTIONS.
By
,
.

.',71

1

putting

x

—=
V,l=

1
,

we obtain

"n x
a n+l - a nx
;

hence we have

X
<l

.T
rt

3

.T
(l

3

XA
(J

X
(

(IfX

l

2

3

i

'l

+

a 2~ a l X +

x a3~ U,2 X +
<i

2

2

<i.."x

a ,~ a i x

+

.-.

log (!

+ *) =

x

Vx

2°-x

3 2 .r

1+ 2-x+ 3-2x+ ~i-Sx +

In certain cases we may simplify the components of the continued fraction by the help of the following proposition
*44S.
:

The continued

fraction
&.

K
a2 +

h
a3 +

K
a4 +

ax +
is

equal to the continued fraction
C

A
+
,

5 CA
c 2a
2

^3
c

C 3C

cxa x

+

a + 3 3

A
+

...

c/c4

where

cn

c2 ,

c

3,

c

4

are any quantities whatever.
7

Let

/"

denote

—— ——
=
«i

7

;

tlien

the continued fraction

+/i
...

ci«i

+

c i/i

Let

/„

denote

3

-

——
«2

;

then

+/2
;

C 2<\.

+

C X2

'

Similarly, c
is

f

=

£ Q Q
-

'^-~
G

C 3 rt 3

+

J

and

so on;

whence the proposition

3

established.

24—2

-a

372

HIGHER ALGEBBA..

^EXAMPLES.
Shew that
1.

XXXI.

b.

1111
+
Wj

UQ

u2

m3

+
1

— + (-l) u n
,

,.

1

Jl

ic

2

«
2.

+
+
J.

Mj

- UQ +

U. 2

u t2 -

u2n _ n-1

?<<!

+

Un ~ u n-l
lv

1 -+-

&n

ClfiCCt

—+
X
r

X2
Cl^Qz-tCto

+
Ctry.0

a?
CIqCC -iCt-v
• • •

&JI

iX-xJu

^n

1

*^

r—\
3.
?•

4.

1111 .2^^—^— — — — — 1-4-1—4n+l
-.

-2

~~

r+1 »~ r + 1,

+2 r + 2?'
,

-.

5.

r

,

l

+s + 3 + « 2
'

11
22

+
+

?i+l

6.

l2
x

11 i»+oa+
x

(n + 1) 2

111
1l
2

1114
4

to

n quotients.
9

...

n2

1- 3- 5- 7-

2n+l
?i

4

+ 22 -

n*+(n+Yf'

7.

e

8.

1111
a
,

= l + 1x + 2- x + S- x+4ab
1
r

x

2x

3x

+

abc
1

-i

— + ...= a+ b-l+ c— abed 1+ d—l +
5—,
1

la
1

9.

l+- + -i + -a + -iB+ ...=i+ — r lb r r
?**
?
,J

1

r3

rn1— + —
r

7-3
,5 r5 +
,

r5

,

1

yr-r — r+1—

•••

a x + a2 +

a3 +

'

an

1+
'

«x +
c q

« 2 -f

«3 +

'
'

<(„_ H-l

u.

if

p=4 a+
9i

,4- 4b+ c+

b+ c+ d+

'

shew that
12.

P (a+ 1 + Q) = + Q.
o
1

Shew that
9*

\-

tinued fraction

——

M2 Ms Mt
/>» />»

...

is

equal to the
q„,
a*.
...

con-

....

where

#.,

are

the

denominators of the successive convergents.

'

,

CHAPTER

XXXII.

PROBABILITY.
Definition. If an event can happen in a ways and fail in b ways, and each of these ways is equally likely, the probability,
449.

or the chance, of

its

happening rr
°

is

z

a+

,

and that

b

of its failing G

is

a+

b

For instance,

if

in a lottery there are 7 prizes
.

and 25 blanks,
prize
is

the chance that a person holding

1

ticket will

win a

7 —

and

his chance of not

winning
for the

25
is

.

Oa
450. bability
If

The reason may be made

mathematical definition of pro:

clear

by the following considerations

an event can happen in a ways and fail to happen in b ways, and all these ways are equally likely, we can assert that the chance of its happening is to the chance of its failing as a to b. Thus if the chance of its happening is represented by ka, where k is an undetermined constant, then the chance of its failing will be represented by kb. .-. chance of happening + chance of failing = k (a + b) Now the event is certain to happen or to fail therefore the sum of the chances of happening and failing must represent certainty. If therefore we agree to take certainty as our unit, we have
;

1

= k (a +
v

b),
'

or

k


a+

T b

:

.-.

the chance that the event will happen

is

a+
is

b b

and the chance that the event
Cor.
If

will not

happen

a+b

the probability of the happening of an event, the probability of its not happening is 1 — p.

p

is

;

374
451.

HIGHER ALGEBRA.
Instead of saying that the chance of the happening of
is
,

it is sometimes stated that the odds are a T a+o in favour of the event or b to a against the event.

an event

to

b

,

The definition of probability in Art. 449 may be given 452. in a slightly different form which is sometimes Useful. If c is the total number of cases, each being equally likely to occur, and of these a are favourable to the event, then the probability that the
event will happen
is

c

,

and the probability that

it

will not

happen

is

1

.

Example 1. What is the chance of throwing a number greater than 4 with an ordinary die whose faces are numbered from 1 to 6? There are 6 possible ways in which the die can fall, and of these two
are favourable to the event required

therefore the required chance

=-= -

.

Example 2. From a bag containing 4 white and 5 black balls a man draws 3 at random what are the odds against these being all black ?
;

The total number of ways in which 3 balls can be drawn is 9 <73 and the number of ways of drawing 3 black balls is 5 C 3 therefore the chance
, ;

of drawing 3 black balls

~*C% ~ 9.8.7
Thus the odds against the event
Example
3.

=
42
5.

'

are 37 to

Find the chance of throwing
dice.

at least

one ace in a single

throw with two

The

possible

number

of cases is 6 x 6, or 36.

on one die may be associated with any of the 6 numbers on the other die, and the remaining 5 numbers on the first die may each be associated with the ace on the second die thus the number of favourable cases

An

ace

;

is 11.

Therefore the required chance

is

— 3b
:

.

Or we may reason

as follows

There are 5 ways in which each die can be thrown so as not to give an That is, the chance ace ; hence 25 throws of the two dice will exclude aces. 25 of not throwing one or more aces is so that the chance of throwing one ; 36

^

ace at least

is 1

-

^ do

,

or

^, oo

;

5

;

;

PROBABILITY.
Example
3 dice.
4.

375

Find the chance

of

throwing more than 15 in one throw with

throw amounting to 18 must be made up of 6, G, G, and this can occur in 1 way; 17 can be made up of G, G, 5 which can occur in 3 ways; 16 may be made up of G, G, 4 and 6, 5, 5, each of which arrangements can occur in
3 ways.

A

Thereforo the number of favourable cases
1

is

+ 3 + 3 + 3,
,

or

10.

And

the total

number

of cases is 6 3 or 21G;

therefore the required chance

=^ = 108 21G
:

has 3 shares in a lottery in which there are 3 prizes and share in a lottery in which there is 1 prize and 2 blanks 6 blanks B has 1 shew that A's chance of success is to ZJ's as 1G to 7.

Example
;

5.

A

A may draw
he

3 prizes in 1 2 prizes

way
1

may draw may draw

and

blank in —^— x 6 ways • m
JL
.

3

2

:

he
the

1 prize

and

6 2 blanks in 3 x -r-^r JL m

ways

sum

of these

numbers

is 64,

which

is

the

win a

prize.

Also he can draw 3 tickets in
therefore

number 9.8.7
'

of

ways

in

which A can

'

,

or 84 ways

-4's

chance of success = chance of success
1

84

—r = —21

.

Z»"s

is

clearly o

;

C
:

1

therefore

A 's

chance

:

B's

chancer—

L

O
7.

= 16

:

Or we might have reasoned thus: A
20 ways J
; '

6.5.4
will get all blanks in
* '

,

or

the chance of which

20
is —-.
,

84

or

5 — 21

;

therefore A's chance of success = 1

-— = -—.
—1

ZL

Suppose that there are a number of events A, B, C,..., of which one must, and only one can, occur ; also suppose that a, b, c, ... are the numbers of ways respectively in which these events can happen, and that each of these ways is equally likely to occur it is required to find the chance of eacli event.
453.
;

The
and

total

of these

number of equally possible ways is a + b + c+ ..., the number favourable to A is a; hence the chance

;

376
that
will

HIGHER ALGEBRA.

A

will

happen rr
is
-.

is

=

a+
c

b

+ c+
:

.

Similarly the chance that J

B

...

happen ri

a+

b

+

+

and

so on.

...

From the examples we have given it will be seen that 454. the solution of the easier kinds of questions in Probability requires nothing more than a knowledge of the definition of Probability,
and the application
tions.

of the laws of

Permutations and Combina-

EXAMPLES.
1.

XXXII.

a.

In a single throw with two dice find the chances of throwing

(1) five, (2) six.

a pack of 52 cards two are drawn at random chance that one is a knave and the other a queen.
2.

Prom

;

find the

bag contains 5 white, 7 black, and 4 red chance that three balls drawn at random are all white.
3. 4.

A

balls:

find the

If four coins are tossed, find the chance that there should be
tails.

two heads and two
5.

one

is

of two events must happen given that the chance of the two-thirds that of the other, find the odds in favour of the other.
:

One
If

will

from a pack four cards are drawn, find the chance that they be the four honours of the same suit.
6.

7.
it is

Thirteen persons take their places at a round table, shew that five to one against two particular persons sitting together.

There are three events A, B, C, one of which must, and only 8. one can, happen; the odds are 8 to 3 against A, 5 to 2 against B: find the odds against C.
9.

dice,

Compare the chances and 12 with three dice.

of throwing 4 with one die, 8 with two

In shuffling a pack of cards, four are accidentally dropped the chance that the missing cards should be one from each suit.
10. 11.

;

find

has 3 shares in a lottery containing 3 prizes and 9 blanks B has 2 shares in a lottery containing 2 prizes and 6 blanks compare their chances of success.
:

A

12. Shew that the chances of throwing six with 4, respectively are as 1 6 18,
;

3,

or 2 dice

;

'

PROBABILITY.
13.

377

There are three works, one consisting of 3 volumes, one of 4, and the other of 1 volume. They are placed on a shelf at random prove that the chance that volumes of the same works are all together
;

3
18

140
14.
-1

and

B

throw with two dice

;

if

A

throws

9,

find

i>'s

chance

of throwing a higher number.
15.

a row

:

The what

letters
is

forming the word Clifton are placed at random in the chance that the two vowels come together ?
is

In a hand at whist what 16. held by a specified player ]
17. line

the chance that the 4 kings are

a

:

There are 4 shillings and 3 half-crowns placed at random in shew that the chance of the extreme coins being both halfis .

crowns

Generalize this result in the case of

m

shillings

and

n half-crowns.

455. have hitherto considered only those occurrences which in the language of Probability are called Single events. When two or more of these occur in connection with each other, the joint occurrence is called a Confound event.

We

For example, suppose we have a bag containing 5 white and 8 black balls, and two drawings, each of three balls, are made from it successively. If we wish to estimate the chance of chawing first 3 white and then 3 black balls, w^e should be
dealing with a

compound

event.

In such a case the result of the second drawing might or might not be dependent on the result of the first. If the balls are not replaced after being drawn, then if the first drawing gives 3 white balls, the ratio of the black to the white balls remaining is greater than if the first drawing had not given three white; thus the chance of drawing 3 black balls at the second trial
the balls are replaced after being drawn, it is clear that the result of the second drawing is not in any way affected by the result of the first.
is

affected

by the

result of

the

first.

But

if

We are thus

led to the following definition

:

Events are said to be dependent or independent according as the occurrence of one does or does not affect the occurrence of the others. Dependent events are sometimes said to be contingent.

;

;

;

378
456.
babilities

HIGHER ALGEBRA.
If there are two independent
of which are known,
first

to

find

events the respective prothe probability that both will

happen.

event may happen in a ways and fail in b ways, all these cases being equally likely ; and suppose that the second event may happen in a' ways and fail in b' ways, Each of the a + b cases may all these ways being equally likely. be associated with each of the a + b' cases, to form (a + b) (a! + b')

Suppose that the

compound

cases all equally likely to occur.

In aa' of these both events happen, in bb' of them both fail, in ab' of them the first happens and the second fails, and in a'b Thus of them the first fails and the second happens.
aa
is

(a

+ b){a'+b')
bb'
is

the chance that both events happen
the chance that both events

fail

(a

+ b)(a+b')
ab'
is

the chance that the
fails

first

happens and the second

(a

+ b)(a'+b')
a'b
is

(a

+ b)(a'+b')

the chance that the
happens.

first

fails

and the second

Thus if the respective chances p and p\ the chance that both

reasoning will apply in Hence it is easy to see that if p lf p 2 p 3 ... are the events. respective chances that a number of independent events will separately happen, the chance that they will all happen is p p 2 p 3 ... ; the chance that the two first will happen and the rest fail is 2\Po (1 — P ) (1 —pj'- > an d similarly for any other par3
,
,

two independent events are will happen is pp'. Similar the case of any number of independent
of

x

ticular case.

the chance that an event will happen in one trial, the chance that it will happen in any assigned succession of r trials is p ; this follows from the preceding article by supposing
457.
If

p

is

r

'

P =P 2 =P 3 =
1

=Pleast of the events will

To

find the chance that
:

some one at

the chance that all the events fail is (1 -p (1 -]).,) (1 -p 3 ) -.-j and except in this case some one x) of the events must happen ; hence the required chance is

happen we proceed thus

=
PROBABILITY.

'

379

Example 1. Two drawings, each of 3 balls, arc made from a bag containing 5 wbitc and 8 black balls, the balls being replaced before tbe second find the chance that the first drawing will give 3 white, and the second trial 3 black balls.
:

The number

of

ways

in

which 3

balls

may

be drawn

is 13

C3

;

3white
3black
Therefore the chance of 3 white at the
first trial

5C.,;

*C Z

.

= •—^-f1.2" 1.2.3
13. 12. 11 _
:

143
=

and the chance of

3 black at the second trial

8.7.6 1.2.3
x

1.2.3

28 143*
•j

>

therefore the chance of the

compound event =--

"143

14o

=

<

.

20449

Example 2. In tossing a coin, find the chance of throwing alternately in 3 successive trials.
Here the
first

head and

tail

throw must give either head or
first is

tail

;

the chance that the

second gives the opposite to the
is

-

,

and the chance that the third throw

the same as the

first is

^

.

Therefore the chance of the

compound

event =- x = 2 2

=

-j

.

4

Example 3. Supposing that it is 9 to 7 against a person A who is now 35 years of age living till he is 65, and 3 to 2 against a person B now 45 find the chance that one at least of these persons will be living till he is 75 alive 30 years hence.
;

The chance
the chance that

that

A

will die within

30 years
3

is

9 —

;

B

will die within 30 years is ^

;

therefore the chance that both will die is

9 ^x3 lb o

',

or

27 —

-

8U
is

;

therefore the chance that both will not be dead, that
.

that one at least will

..

.

be alive,

,

27

>sl-8o>

53 or-.

By a slight modification of the meaning of the symbols 458. in Art. 45G, we are enabled to estimate the probability of the concurrence of two dependent events. For suppose that when the first event has happened^ a denotes the number of ways in which the second event can follow, and b' the number of ways in which it will not follow then the number of ways in which the two
;

:

.

'

;

380

HIGHER ALGEBRA.
is

events can happen together

aa\ and the probability of their

aa
concurrence
is
-.

+

o) (a

7

,

,

,

—tr +
o
)

the probability of the first event, and p' the contingent probability that the second will follow, the probability of the concurrence of the two events is pp

Thus

if

p

is

Example 1. In a hand at whist find the chance that a specified player holds both the king and queen of trumps.
Denote the player by
13

A

;

then the chance that

A

has the king

is clearly

^;

for this particular card

can be dealt in 52 different ways, 13 of which

fall

to A.

The chance
12 — ol
:

that,

when he has the

king, he can also hold the queen is
fall

then

for the

queen can be dealt in 51 ways, 12 of which
chance required = u
,
n
.

to A.

m * « Therefore the

— x 12 = 17 ^ ol 52
1
--:

13

.

Or we might reason

as follows

The number of ways in which the king and the queen can be dealt to A is equal to the number of permutations of 13 things 2 at a time, or 13 12. And similarly the total number of ways in which the king and queen can be
.

dealt is 52

.

51.

Therefore the chance =

13

.

12
„„

*

52.51

=—
1

,

as before.

17

Example 2. Two drawings, each of 3 balls, are made from a bag containing 5 white and 8 black balls, the balls not being replaced before the second trial: find the chance that the first drawing will give 3 white and the second 3 black balls.
At the first trial, 3 balls may be drawn in and 3 white balls may be drawn in 5 C3 ways;
therefore the chance of 3 white at first trial
13

C3 ways
13
.

5.4 1.2"

12

.

11

5

1.2.3

143

When
2 white

3 white balls have been
balls
;

drawn and removed, the bag contains
in
10

and 8 black
balls

therefore at the second trial 3 balls

and 3 black

may

be drawn in

may be drawn 8 C ways 3
;
.

C 3 ways

;

therefore the chance of 3 black at the second trial

8.7.6

"1.2.3
therefore the chance of the
5

'

10.9.8 1.2.3
7
=-=
.

~

_

1_

m

15

compound event
143

x

15

=

7

429 Ex.
1,

The student should compare

this solution with that of

Art. 457.

;

. . .

PROBABILITY.

381

459. If an event can happen in ttvo or more different ways which are mutually exclusive, the chance that it wilt happen is the sum of the chances of its happening in these different ways.

This is sometimes regarded as a self-evident proposition arising It may, howimmediately out of the definition of probability. ever, be proved as follows
:

Suppose the event can happen in two Avays which cannot
concur
;

and

let

^

,

=*

be the chances of the happening of the

event in these two ways respectively. Then out of bfi 2 cases there are a b 2 in which the event may happen in the first way, and a b ways in which the event may happen in the second; and tliese ivays cannot concur. Therefore in all, out of b l b 2 cases there are a,b„ + a ,b, cases favourable to the event: hence the chance that the event will happen in one or other of the two
x
J
k

ways

is

a b 2 + a2 b
x

a_
x
x

a,

bh 12

6,

bf
number
of ex-

Similar reasoning will apply whatever be the clusive ways in which the event can happen.

an event can happen in n ways which are mutually exclusive, and if plt p p^ ---Pn are the probabilities that the a event will happen in these different ways respectively, the probability that it will happen in some one of these ways is

Hence

if

,

Pi+Pl+Pa*
Example
with two
9 can be
1.

+Pnat least in a single

Find the chance of throwing 9

throw
4
,

dice.

made up

in 4 ways,

and thus the chance

of throwing 9 is

.

10 can be

made up

in 3 ways,

and thus the chance

of throwing 10 is

^
2

3

11 can be

made up made up

in 2 ways,

and thus the chance of throwing 11 is-and thus the chance of throwing 12
less
is

12 can be

in 1 way,

-^

chance of throwing a number not separate chances
.*.

Now the

than 9

is

the

sum

of these

the required chance =

ou

<V/ .

=

lb

n

.

;

:

382
Example
2.

HIGHER ALGEBRA.
One purse contains
1 sovereign

and 3 shillings, a second purse contains 2 sovereigns and 4 shillings, and a third contains 3 sovereigns and 1 shilling. If a coin is taken out of one of the purses selected at random, find the chance that it is a sovereign.
Since each purse
the
first is

is

equally likely to be taken, the chance of selecting
is

-

;

and the chance of then drawing a sovereign
it

-

;

hence the
purse
is
it

chance of drawing a sovereign so far as
- x j , or =^ 12 3 4
.

depends upon the

first

Similarly the chance of drawing a sovereign so far as
is

depends on the second purse
chance of drawing a sovereign

is

12 13 - x -

- x 3 6

,

or 9
1

1

;

and from the third purse the

,

or -

.-.

the required chance =

—+-+ y
x*5

-.

=o

.

tc

In the preceding article we have seen that the pro460. bability of an event may sometimes be considered as the sum of the probabilities of two or more separate events ; but it is very important to notice that the probability of one or other of a series of events is the sum of the probabilities of the separate events only when the events are mutually exclusive, that is, when the occurrence of one is incompatible with the occurrence of any
of the others.

Example.

From
:

drawn

at

random

20 tickets marked with the first 20 numerals, one find the chance that it is a multiple of 3 or of 7.
is

is

The chance that the number
it is

a multiple of 3

is

,

and the chance that
hence the

a multiple of 7
.

is

2 —

;

and

these events are mutually exclusive,

required chance

,

,

is

6 2 — + -^

,

2 or -

.

But if the question had been: find the chance that the number multiple of 3 or of 5, it would have been incorrect to reason as follows
Because the chance that the number
chance that the number
it is

is

a

is

a multiple of 3
is

is

,

and the

is

a multiple of 5
is

4 —

,

therefore the chance that
ticket

a multiple of 3 or 5

^+^

6

4

,

or 5,

1

.

For the number on the

might be a multiple both of 3 and of are not mutually exclusive.
461.
It

so that the

two events considered

should be observed that the distinction between simple and compound events is in many cases a purely artificial

:

PROBABILITY.
;

383

one in fact it often amounts to nothing more than a distinction between two different modes of viewing the same occurrence.
Example. A bag contains 5 white and 7 black balls; if two drawn what is the chance that one is white and the other black?
(i)

balls arc

Regarding the occurrence as a simple event, the chance

= (5*7H.=C 2 = (ii)

.

66

of the two following
(1)

The occurrence may be regarded compound events

as the

happening of one or other
is

drawing a white and then a black

ball,

the chance of which

°r 132 12 * 11
(2)

;

drawing a black and then a white
7

ball,

the chance of which

is

X

5

i2

ir

0r

35 132'

And

since these events are mutually exclusive, the required chance

- j$5 + ^5_ _35 132~66' 132
be noticed that we have here assumed that the chance of drawing two specified balls successively is the same as if they were drawn simultaneously. A little consideration will shew that this must be the case.
It will

EXAMPLES. XXXII.
1.

b.

the chance of throwing an ace in the successive throws with an ordinary die ?
is

What

first

only of two

Three cards are drawn at random from an ordinary pack 2. the chance that they will consist of a knave, a queen, and a king.

:

find

The odds against a certain event are 5 to 2, and the odds in 3. favour of another event independent of the former are 6 to 5 find the chance that one at least of the events will happen.
;

The odds against A solving a certain problem are 4 to 3, and 4. the odds in favour of B solving the same problem are 7 to 5 what is the chance that the problem will be solved if they both try 1
:

the chance of drawing a sovereign from a purse one compartment of which contains 3 shillings and 2 sovereigns, and the other 2 sovereigns and 1 shilling ?
5.
is

What

A

bag contains 17 counters marked with the numbers 1 to 17. 6. counter is drawn and replaced; a second drawing is then made: what is the chance that the first number drawn is even and the second

A

odd?

384
7.

HIGHER ALGEBRA.

Four persons draw each a card from an ordinary pack: find the chance (1) that a card is of each suit, (2) that no two cards are of
equal value.
8.

Find the chance of throwing

six

with a single die at least once

in five trials.

be favourably reviewed by three independent critics are 5 to 2, 4 to 3, and 3 to 4 respectively what is the probability that of the three reviews a majority will be favourable ?
9.

The odds that a book

will

;

bag contains 5 white and 3 black balls, and 4 are successively drawn out and not replaced what is the chance that they are alternately
10.
;

A

of different colours
11.

%

In three throws with a pair of dice, find the chance of throwing

doublets at least once.
12.

If 4 whole

numbers taken at random are multiplied together
last digit in the

shew that the chance that the
.

product

is 1, 3, 7,

or 9

16

1S

625'

In a purse are 10 coins, all shillings except one which is a 13. Nine coins are taken sovereign in another are ten coins all shillings. from the former purse and put into the latter, and then nine coins are taken from the latter and put into the former find the chance that the sovereign is still in the first purse.
;
:

14.

will

two coins are tossed 5 times, what be 5 heads and 5 tails
If
\

is

the chance that there

If 8 coins are tossed, 15. one will turn up head?
16.

what

is

the chance that one and only

order cut a pack of cards, replacing them after each cut, on condition that the first who cuts a spade shall win a prize find their respective chances.

A, B,

C in

:

A and B draw from a purse containing 3 sovereigns and 17. 4 shillings find their respective chances of first drawing a sovereign, the coins when drawn not being replaced.
:

of n ^persons sit at a round table, find the odds against two specified individuals sitting next to each other.
18.

A party
A
is

one of 6 horses entered for a race, and is to be ridden by one of two jockeys B and C. It is 2 to 1 that B rides A, in which case all the horses are equally likely to win if C rides A, his chance what are the odds against his winning ? is trebled
19.
;
:

If on an average 1 vessel in every 10 is wrecked, find the chance that out of 5 vessels expected 4 at least will arrive safely.
20.

.

PROBABILITY.
462.
trial

385

The probability of the happening of an event in one being known, required the probability of its happening once,
...

twice, three times,

exactly in

n

trials.

Let p be the probability of the happening of the event in a single trial, and let q = 1 -p\ then the probability that the event will happen exactly r times in n trials is the (r + l) th term in the expansion of (q + p)*.
set of r trials out of the total number n, the chance that the event will happen in every one of r q"~ [Art. 456], and as these r trials and fail in all the rest is
if

For

we

select

any particular

a set of r

ways, all of which are equally applicable to the case in point, the required chance is
trials

can be selected in

n

p

Cr

C rp q
If

.

we expand
2f

(/;

+

q)"

by the Binomial Theorem, we have
...

+ "C 2) n
1

-

1

q

+ n C jS'- 2 q 2 +

+"C n _ p
r

r

q"- r

+

...

+ qn

;

thus the terms of this series will represent respectively the probabilities of the happening of the event exactly n times, n — 1
times,

n—

2 times,

...

inn

trials.

463.
twice,
...

If

the event happens

n

times,

or

fails

(n — r) times, it happens r times or more ; chance that it happens at least r times in n trials is
n n P + "Cy-*q + "C aP -Y+
...

only once, therefore the

tv^r.
what
the chanco of

or

the

sum

of

the

first

n—r+

1

terms of the expansion of

Example 1. In four throws with a pair of throwing doublets twice at least ?
In a single throw the chance of doublets
failing to
is

dice,

is

c

i
,

-^ do

or ^ o

;

and

the chance of
if

throw doublets

is

5 ^

.

Now

the required event follows
;

doublets

are thrown four times, three times, or twice
is

therefore the required chanco

the

sum

of the first three terms of the expansion of

/l 5\ h+d

4
.

Thus the chance =
H. H. A.

19 1 — (1 + 4.5 + 6.5-)= -^

25

;

:

386
Example
:

HIGHER ALGEBRA.
2.

bag contains a certain number of balls, some of which are replaced, another is then drawn and replaced; on if p is the chance of drawing a white ball in a single trial, find and so the number of white balls that is most likely to have been drawn in n trials.
white; a ball
is

A

drawn and

The chance
find for

what

of drawing exactly r white balls is n C rp r q n value of r this expression is greatest. nC
r

-r
,

and we have

to

Now
so long as

p

r

q

n- r

> n Cr - l p r - l q n - (r ~
l)p>(p + q)r.

l

\

(n-r + l)p>rq,
(n +

or

But p + 5 = 1; hence the required value of p[n + l).
If

r is

the greatest integer in

n

is

such that pn
failures.

is

an

integer,

the most likely case

is

that of

pn

successes

and qn

Suppose that there are n tickets in a lottery for a prize of £x; then since each ticket is equally likely to win the prize, and a person who possessed all the tickets must win, the money value of
464.

each ticket

is

£-

x
:

n

in other

words

this

would be a

fair

sum

to

pay

for each ticket; hence a person

who

possessed r tickets might

reasonably expect

£—
TX

as the price to be paid for his tickets
is,

by

any one who wished to buy them; that

he would estimate
in-

£- x
n

as the worth of his chance.
:

It

is

convenient then to

troduce the following definition
If

represents a person's chance of success in any venture and the sum of money which he will receive in case of success, the sum of money denoted by is called his expectation.

M

p

pM

In the same way that expectation is used in reference 465. to a person, we may conveniently use the phrase probable value applied to things.
Example 1. One purse contains 5 shillings and 1 sovereign a second purse contains 6 shillings. Two coins are taken from the first and placed in the second then 2 are taken from the second and placed in the first find the probable value of the contents of each purse.
: ;

The chance that the sovereign is in the first purse is equal to the the chances that it has moved twice and that it has not moved at all

sum

of

8
PROBABILITY.
that
is,

,387
8

the chance

=

6
.*.

112 .1=-t. -+5
.

4

3

4

.

the chance that the sovereign

is

in the second purse =-r.

Hence the probahle value

of the first purse

=T
4
.*.

3

of 25*.

+

1
.

4

of 6*.=£1.

O.s-.

3r/.

the probable value of the second purse

=31*.-2Q£«.=10*.

<></.

Or the problem may be
The probable value

solved as follows

:

of the coins

removed

=s

of 25s.

= 8^s.;

the probable value of the coins brought back

=^of

(Gs.+S V>\)=3 r
:

W,

.\ the probable value of the first purse

= (25-81 + 3^)

shillings

= £1.

Ck

3d., as before.

Example 2. A and B throw with one be won by the player who first throws 6.
their respective expectations?
1

die for a stake of
If

A has

the

first

£11 which is to throw, what are
5
1

In his

first

throw A' a chance

is

b

;

in his second

it is

-^

x - x o 6

5

,

because
;

each player must have failed once before A can have a second throw in his 1 /5\ 4 third throw his chance is x ^ because each player must have failed
(

J

twice; and so on.

Thus A's chance

is

the

sum

of the infinite series

5MKi©
Similarly #'s chance
is

4+
}•

the

sum

of the infinite series

WM»
.-.

,+

4

G)

*
J-

A' a chance

is

to 7>"s as G is to 5; their respective chances are therefore
their expectations are

--

and =y, and

£6 and £5

respectively.

26—

388

HIGHER ALGEBRA.

shall now give 466. and interesting results.

We

two problems which lead

to useful

Example 1. Two players A and B want respectively m and n points of winning a set of games their chances of winning a single game are p and q respectively, where the sum of p and q is unity the stake is to belong to the player who first makes up his set determine the probabilities in favour
; ; :

of each player.

Suppose that A wins in exactly m + r games; to do this he must win the The chance of last game and m-1 out of the preceding m + r-1 games. ™+^- 1 p™- 1 q r 2h or m^~1 Cm-1 pm qr . this is m_ 1
the set will necessarily be decided in m + n - 1 games, and A may 1 games, ... or + n - 1 games; win his m games in exactly m games, or therefore we shall obtain the chance that A wins the set by giving to r the Thus A s chance is values 0, 1, 2, ... n - 1 in the expression m+r 1 Cm _ 1 p m q r

Now

m+

,

m

J

.

similarly B's chance is

n(nA-l\

1.2

*

+ n-2 jm-1
\m
ii

)

This question is known as the " Problem of Points," and has engaged the attention of many of the most eminent mathematicians since the time of Pascal. It was originally proposed to Pascal by the Chevalier de Mere in 1654, and was discussed by Pascal and Fermat, but they confined themselves to the case in which the players were supposed to be of equal skill their results were also The formulae we have given are exhibited in a different form. assigned to Montmort, as they appear for the first time in a work The same result was afterwards obof his published in 1714. tained in different ways by Lagrange and Laplace, and by the latter the problem was treated very fully under various modi:

fications.

Example 2. There are n dice with / faces marked from 1 to /; if these are thrown at random, what is the chance that the sum of the numbers exhibited shall be equal to p?
Since any one of the
the

number

be exposed on any one of the n dice, of ways in which the dice may fall is / n

/

faces

may

.

their

Also the number of ways in which the numbers thrown will have sum is equal to the coefficient of x p in the expansion of
{x l

p

for

+ x* + x 3 + ... + xf

n
)
\

for this coefficient arises out of the different ways in 1, 2, 3, .../can be taken so as to form p by addition.

which n of the indices

PROBABILITY.

3-S!)

Now

the above expression

=x

11

(l

+ x + x2 +

...

+ xf

')"

-(£?)"
We
have therefore to find the
(I
.

coefficient of

x p ~ n in the expansion of
.

-

x') n (I

- x)~ n

t

n(n-l) »(n+l)
*

.,.

n(n-l)(n-2)

.,,

and

<1

-

v

.r)

« -" = , 1

+ nx +


2

' a;

+

?t(»+l)(w + 2) _ x3 +... 1 ^ 3
.

~ Multiply these series together and pick out the coefficient of x p n in the product we thus obtain
;

n(n+l)...{p-l)
\

P -n

it

n(n+l)...(p-f-l) p - n -f
\

+

n

(n

-

1)

M(;t

+ l)...(j>-2/-l)
\

1.2

•"

p-n-2f
The
.

where the series is to continue so long as no negative factors appear. required probability is obtained by dividing this series by/ n

This problem
in 1730
j

is

due to

De Moivre and was

published by him

it illustrates

a method of frequent

utility.

Laplace afterwards obtained the same formula, but in a much more laborious manner he applied it in an attempt to demonstrate the existence of a primitive cause which has made the planets to move in orbits close to the ecliptic, and in the same direction as the earth round the sun. On this point the reader may consult Todhunter's History of Probability, Art. 987.
;

EXAMPLES. XXXII.
1.

c.

of

.1

In a certain game A'a skill is to winning 3 games at least out of 5.

2>'s

as 3 to 2

:

find the chance

is

coin whose faces are marked 2. the chance of obtaining a total of 12 ?

A

2,

3

is

thrown 5 times

:

what

In each of a set of games it is 2 to 1 in favour of the winner 3. the previous game what is the chance that the player who wins of the first game shall win three at least of the next four ?
:

There are 9 coins in a bag, 5 of which are sovereigns and 4. the rest are unknown coins of equal value find what they must be if the probable value of a draw is 12 shillings.
;

390
5.

HIGHER ALGEBRA.

A coin

present itself

n times, what an odd number of times ?
is

tossed

is

the chance that the head will

is

From a bag containing 2 sovereigns and 3 shillings a person 6. allowed to draw 2 coins indiscriminately; find the value of his ex7.

pectation.

Six persons throw for a stake, which is to be won by the one who first throws head with a penny ; if they throw in succession, find the chance of the fourth person.
3 are placed in a bag, and one is withdrawn and replaced. The operation being repeated three times, what is the chance of obtaining a total of 6 ?
8.
1, 2,

Counters marked

9.

A coin whose faces are marked 3 and
sum
of the

5

is

tossed 4 times
less

:

what

are the odds against the
10.

numbers thrown being

than 15?

Find the chance of throwing 10 exactly in one throw with

3 dice.
players of equal skill, A and B, are playing a set of games they leave off playing when A wants 3 points and wants 2. If the stake is £16, what share ought each to take \
11.

Two

;

B

12.

A

and

B throw with
number

3 dice
?

:

if

A

throws

8,

what

is Z?'s

chance

of throwing a higher

A had in his pocket a sovereign and four shillings taking out 13. two coins at random he promises to give them to B and C. What is the worth of (7's expectation ?
;

14.
(1)

In

five

throws with a single die what

is

the chance of throwing

three aces exactly, (2) three aces at least.
15.

B of 5s. to 2s. that in a single throw with two dice he .will throw seven before B throws four. Each has a pair of dice and they throw simultaneously until one of them wins find B's
makes a bet with
:

A

expectation.

one the common cube, and the other a regular tetrahedron, the number on the lowest face being taken in the case of the tetrahedron; what is the chance that the sum of the numbers thrown is not less than 5 ?
16.

A person

throws two

dice,

bag contains a coin of value J/, and a number of other coins whose aggregate value is m. A person draws one at a time till he draws the coin 31 find the value of his expectation.
17.
:

A

6n tickets numbered 0, 1, 2, 6n- 1 are placed in a bag, and three are drawn out, shew that the chance that the sum of the numbers on them is equal to 6?i is
18.

If

3?&

(6n-l)(6n-2)'

;

PROBABILITY.

3Dl

*Inverse Probability.
all the cases we have hitherto considered it lias been supposed that our knowledge of the causes which may produce a certain event is sucli as to enable us to determine the chance of the happening of the event. have now to consider problems of a different character. For example, if it is known that an event has happened in consequence of some one of a certain number of causes, it may be required to estimate the probability of each cause being the true one, and thence to deduce the probability of future events occurring under the operation of the

*467.

In

We

same

causes.

*468.

Before discussing the general case

we

shall

give a

numerical illustration.

Suppose there are two purses, one containing 5 sovereigns and 3 shillings, the other containing 3 sovereigns and 1 shilling, and suppose that a sovereign lias been drawn it is required to find the chance that it came from the first or second purse.
:

Consider a very large number iV of trials ; then, since before the event eacli of the purses is equally likely to be taken, we may

assume that the

first

purse would be chosen in ^ iV of the

trials,

and

5 in - of these a sovereign

would be drawn

;

8

thus a sovereign °
first

would be drawn

5
-

1

o

x ~iV, or
1

5 —N times from the lb
in -

purse.

The second purse would be chosen
3

N

of the trials,

and

in

j

of these a sovereign

would be drawn

;

thus a sovereign would

be drawn -JV times from the second purse.
very large but is otherwise an arbitrary number let us put iV-16n; thus a sovereign would be drawn 5 n times from the first purse, and Qn times from the second purse; that is, out of the lln times in which a sovereign is drawn it comes from the first purse bn times, and from the second purse 0?i
JV
is

3

Now

:

392
times.
first

HIGHER ALGEBRA.
Hence the
is

probability that the sovereign
it

came from the

purse


5

,

and the probability that

came from the

A' second is rr.
*469. It is important that the student's attention should be directed to the nature of the assumption that has been made in Thus, to take a particular instance, the preceding article. although in 60 throws with a perfectly symmetrical die it may not happen that ace is thrown exactly 10 times, yet it will doubtless be at once admitted that if the number of throws is continually increased the ratio of the number of aces to the number of throws will tend more and more nearly to the limit

6


6

.

There
;

is

no reason

why one

face should appear oftener than

hence in the long run the number of times that each of the six faces will have appeared will be approximately equal.
another

The above instance is a particular case which is due to James Bernoulli, and was

of a general
first

theorem given in the Ars

Conjectandi, published in 1713, eight years after the author's Bernoulli's theorem may be enunciated as follows death.
event happens in a single trial, then if the number of trials is indefinitely increased, it becomes a certainty that the limit of the ratio of the number of successes to the number of trials is equal to p ; in other words, if the number of

If p

is the

probability that

an

trials is

N,

the

number of successes may

be taken to be

pN.

proof See Todhunter's History of Probability, Chapter vn. of Bernoulli's theorem is given in the article Probability in the Encyclopaedia Britannica.

A

observed event has Jiappened through, some one of a number of mutually exclusive causes : required to find the probability of any assigned cause being the true one.

*470.

An

.

suppose that the probability of the existence of these causes was estimated Let p r denote the probability that when the at P P2 P3 ... Pn r* cause exists the event will follow after the event has occurred th cause was the it is required to find the probability that the r true one.
the event took place
'

Let there be n causes, and before
x ,
, ,
.

1*

:

PROBABILITY.
;

393

Consider a \ cry great number JV of trials then the first cause of these, and out of this number the event follows exists in P similarly there are p^^N trials in which the event in p Px j follows from the second cause; and so on for each of the other causes. Hence the number of trials in which the event follows is
x

N

X

N

;md the number in which the event was due to the r th cause
'P,.I\N
',

is

lience after the event the probability that the r
is

th

cause

was the true one
tli at

pJPjr+NUpP);
is,

the probability that the event was produced by the

r"'

cause

is

Mvn
It
is

PrK

necessary to distinguish clearly between the probability of the existence of the several causes estimated before the event, and the probability after the event has happened of any The former are usually called assigned cause being the true one. P a priori probabilities and are represented by ... n \ probabilities, and if we denote the latter are called a posteriori them by Q t1 Q„, Q 3 ... Q Hf we have proved that

*471.

P

,

,

x

P

.

P

,

Qr

2 ( P P)

'

where p r denotes the probability of the event on the hypothesis
th of the existence of the r cause.

which is otherwise evident as the event has happened from one and only one
this result it appears that

From

S

(Q)

=

lj

of the causes.

give another proof of the theorem of the preceding article which does not depend on the principle enunciated in Art. 469.
shall

We

now

observed event has happened through some one of a member of mutually exclusive causes : required to find the probability of any assigned cause being the true one.

*472.

An

suppose that the probability of the existence of these causes was estimated at P P2 Pz , ... P n Let p r denote the probability that when the ?- th cause exists the event will follow ; then the antecedent probability that the event would follow from the r th cause is p r P r
the event took place
t

Let there be n causes, and before
,
.

,

.

;

394
Let

HIGHER ALGEBRA.

be the a posteriori probability that the r cause was the th true one; then the probability that the r cause was the true one

Q

th

r

is

proportional to the probability that, would produce the event
.

if

in existence, this cause

"

pA pA

A._JL =

'"

_

<?»

.

s(<?)

_
)

i
.

p. p ,
Pr pr

Hp p

s(P P)'

<?,=

2 (pP)

appears that in the present class of problems the product Prp r will have to be correctly estimated as a first step; in many cases, however, it will be found that P lt P 2 P3 ... are all equal, and the work is thereby much simplified.

Hence

it

,

,

,

Example. There are 3 bags each containing 5 white balls and 2 black and 2 bags each containing 1 white ball and 4 black balls a black ball having been drawn, find the chance that it came from the first group.
balls,
:

Of the

five bags, 3

belong to the

first

group and 2 to the second

;

hence

If

a bag
2
;

is

selected

from the

first

group the chance of drawing a black
is

ball is 7

if

from the second group the chance

4 o

;

thus

px = /

2
,

p.2

=

4
o
;

"
Hence the chance

lh

p 1_

6

3o'

lh

2

~~25'
of the first

^

that the black ball

came from one
43*

group

is

JL^/A h 8\15 \35"
35
'

25/

an event has been observed, we are able by the method of Art. 472 to estimate the probability of any particular cause being the true one ; we may then estimate the probability of the event happening in a second trial, or we may find the probability of the occurrence of some other
*473.
event.

When

For example, pr is the chance that the event will happen from the rth cause if in existence, and the chance that the r th cause is the true one is Qr hence on a second trial the chance Therefore that the event will happen from the rth cause is p r Qr the chance that the event will happen from some one of the causes on a second trial is 2 (2} Q)'
;
.

PROBABILITY.
Example.

395

shillings; 2 coins are

what

is

purse contains 4 coins which arc either sovereigns or drawn and found to be shillings: if these are replaced the chance that another drawing will give a sovereign?

A

This question
separately.

may

be interpreted in two ways, which we shall discuss

If we consider that all numbers of shillings are a priori equally likely, shall have three hypotheses; for (i) all the coins may be shillings, (ii) three of them may be shillings, (iii) only two of them may be shillings.
I.

we

Here
also

P^P.^P.^;

^=
iirst

1,

J>a

=g, P»=q-

Hence probability of

hypothesis

= 1-5- (1 + o
o

+ r) — +
f )

tTv^ Qi>

probability of second hypothesis =

^ (*+2
-f-

=

To

~ ^'

J'

probability of third hypothesis — -

(

1

+~ +

.-

)

=T7\=Qy

Therefore the probability that another drawing will give a sovereign

1

3_

2

~4
II.

*10

+

1_

5^

4

'To~40

-1 8*

If

each coin

is

equally likely to be a shilling or a sovereign, by taking
I

the terms in the expansion of
1

/l .

+

-

IV
,

we

see that

the

chance of four
6 —

J

4
,

shillings is

r-^

,

lb

of three shillings is 77:

of

two shillings

is

-.

10

10

;

thus

P_l il_ 16'
also, as before,
iJ i

P

2


~1G'

P _A. ^»-16'
Ps—r'

= l>

Pi— a*
Qi +

Hence

Qi_Q-2_Q*_
6
'

Q2+Q*
24

_

12

'

6

1 24'

Therefore the probability that another drawing will give a sovereign

= (<2ix0)+(q,x^ +
~

((?

:j

x|)

8

+

16

4

396

HIGHER ALGEBRA.

shall now shew how the theory of probability may *474. be applied to estimate the truth of statements attested by witshall nesses whose credibility is assumed to be known. be the truth, suppose that each witness states what he believes to whether his statement is the result of observation, or deduction, or experiment; so that any mistake or falsehood must be attributed to errors of judgment and not to wilful deceit.

We

We

problems we shall discuss furnishes a useful intellectual exercise, and although the results cannot be regarded as of any practical importance, it will be found that they confirm

The

class

of

the verdict of

common
it is is

sense.

*475. When speaks the truth

ments made by him of those which are true to the whole number.
*476.

asserted that the probability that a person p, it is meant that a large number of statehas been examined, and that p is the ratio

independent witnesses, A and B, whose probabilities of speaking the truth are p and p' respectively, agree in making a certain statement what is the probability that the statement is true %

Two

:

Here the observed event is the fact that A and B make the same statement. Before the event there are four hypotheses for A and B may both speak truly or A may speak truly, B falsely; or A may speak falsely, B truly or A and B may both speak The probabilities of these four hypotheses are falsely.
; ; ;

PP\

p( l ~P\

P'Q-P)*

(

1

-P)( 1 ~P')
which

respectively.

Hence

after the observed event, in

A and B make
;

the

same statement, the probability that the statement is true is to that the probability that it is false as pp to (1 - p) (1 -p')
is,

the probability that the joint statement
pp'

is

true

is

pp' + (l-p)(l-p')'
Similarly if a third person, whose probability of speaking the truth is p", makes the same statement, the probability that the statement is true is

ppp
and so on
for

./„//
}

ppY + {1-p){i-p')(i-p")
any number
of persons.

PROBABILITY.
*477.
;

397

In the preceding article it lias been supposed that we have no knowledge of the event except the statement made by A and B if we have information from other sources as to the probability of the truth or falsity of the statement, this must be taken into account in estimating the probability of the various
hypotheses.

For instance, if A and B agree in stating a fact, of which the a priori probability is P, then we should estimate the probability of the truth and falsity of the statement by Ppp* and (1 - P) (1 — p>) (1 — p') respectively.
Example. There is a raffle with 12 tickets and two prizes of £9 and £3. A, B, C, whose probabilities of speaking the truth are ^, §, f respectively, A and B assert that he has report the result to D, who holds one ticket. won the £9 prize, and C asserts that he has won the £3 prize; what is D's expectation?
Three cases are possible; D may have won C may all have spoken falsely. Now with the notation of Art. 472, we have
«

£9, £3, or nothing, for A, B,
the a priori probabilities
3

P-i P-A P-™. *i- 12 *a-l2' ^ ~12'
also

Pi~2

12 X 24 X
3

5-30>

**~~2 X 3 X 5

1133 ~
3

30

'

A_ 2 * 3

1

1

X

2_ 2
5~3()
;

"
hence D's expectation

4

20

27'

4 3 =— of £9 + — of £3

=£1.

13s. id.

*478. With respect to the results proved in Art. 47G, it should be noticed that it was assumed that the statement can be made in two ways only, so that if all the witnesses tell falsehoods they agree in telling the same falsehood.

not the case, let us suppose that c is the chance that the two witnesses A and B will agree in telling the same falsehood then the probability that the statement is true is to the probability that it is false as pp' to c (1 —p) (1 — p').
If this
is
;

general rule, it is extremely improbable that two independent witnesses will tell the same falsehood, so that c is usually very small; also it is obvious that the quantity c becomes These consmaller as the number of witnesses becomes greater. siderations increase the probability that a statement asserted by two or more independent witnesses is true, even though the credibility of each witness is small.

As

a

.

398

HIGHER ALGEBRA.

Example. A speaks truth 3 times out of 4, and B 7 times out of 10; they both assert that a white ball has been drawn from a bag containing 6 balls
all of different

colours

:

find the probability of the truth of the assertion.
;

There are two hypotheses
false.

(i)

their coincident testimony is true,

(ii) it

is

Here

Px =•1 ~ 6'
,

P2 ~6' P- 5

^ 1- 4 X

10'

P2 ~25 X

4

X

10

;

for in estimating p.2 we must take into account the chance that A and both select the white ball when it has not been drawn ; this chance is

B

will

11 X
5

1

°r

*

5

25

Now

the probabilities of the two hypotheses are as

P^

to

P po,
2

and

therefore as 35 to 1; thus the probability that the statement is true

is

35

*479. The cases we have considered relate to the probability of the truth of concurrent testimony; the following is a case of traditionary testimony.
states that a certain event took place, having received an account of its occurrence or non-occurrence from B, what is the
If

A

probability that the event did take place

1

The event happened

(1) if

they both spoke falsely ; one of them spoke the truth.

they both spoke the truth, (2) if and the event did not happen if only

Let p, p denote the probabilities that A and B speak the truth ; then the probability that the event did take place is
pp' +

(l-p)(l-p)

}

and the probability that

it

did not take place

is

p(l- 2))+p'(l-p).
*480. The solution of the preceding article is that which has usually been given in text-books; but it is" open to serious objections, for the assertion that the given event happened if both A and B spoke falsely is not correct except on the supposition that Moreover, the statement can be made only in two ways. although it is expressly stated that A receives his account from B, this cannot generally be taken for granted as it rests on
A'& testimony.

:;

PROBABILITY.

399

the different ways of interpreting the question, and of the different solutions to which they lead, will be found in the Educational Times Reprint, Yols. XXVII. and XXXII.
full discussion of

A

^EXAMPLES. XXXII.

d.

There are four balls in a bag, but it is not known of what 1. colours they are one ball is drawn and found to be white find the chance that all the balls are white.
;
:

In a bag there are six balls of unknown colours; three balls are drawn and found to be black; find the chance that no black ball
2. is left in

the bag.

A letter is known to have come either from London or Clifton 3. on the postmark only the two consecutive letters ON are legible what is the chance that it came from London ?
;

Before a race the chances of three runners, A, B, C, were estimated to be proportional to 5, 3, 2 but during the race A meets with an accident which reduces his chance to one-third. What are now the respective chances of B and C ?
4.
;

purse contains n coins of unknown value a coin drawn at random is found to be a sovereign; what is the chance that it is the only sovereign in the bag ?
5.
;

A

has 10 shillings and one of them is fcnown to have two heads. He takes one at random and tosses it 5 times and it always what is the chance that it is the shilling with two heads ? falls head
6.
:

A man

bag contains 5 balls of unknown colour; a ball is drawn and replaced twice, and in each case is found to be red if two balls are now drawn simultaneously find the chance that both are red.
7.
:

A
A

purse contains five coins, each of which may be a shilling 8. or a sixpence two are drawn and found to be shillings find the probable value of the remaining coins.
;
:

9.

A die
is

is

thrown
10.

15

:

thrown three times, and the sum of the three numbers find the chance that the first throw was a four.

A

what is the probability that they the same fact ?

speaks the truth 3 out of 4 times, and B 5 out of 6 times will contradict each other in .stating

; :

400
11.

HIGHER ALGEBRA.

speaks the truth 2 out of 3 times, and B 4 times out of 5 they agree in the assertion that from a bag containing 6 balls of different colours a red ball has been drawn find the probability that the state-

A

:

ment

is true.

One of a pack of 52 cards has been lost ; from the remainder 12. of the pack two cards are drawn and are found to be spades ; find the chance that the missing card is a spade.
with 10 tickets and two prizes of value £5 holds one ticket and is informed by that he has won the £b prize, while C asserts that he has won the ,£1 prize is denoted by §, and what is A's expectation, if the credibility of that of C by f ?
13.

There

is

a

raffle

and £1

respectively.

A

B

B

purse contains four coins two coins having been drawn are found to be sovereigns find the chance (1) that all the coins are sovereigns, (2) that if the coins are replaced another drawing will give a sovereign.
14.
;
:

A

makes a bet with Q of ,£8 to £120 that three races will be the three horses A, B, C, against which the betting is 3 to 2, 4 to 1, and 2 to 1 respectively. The first race having been won by A, and it being known that the second race was won either by B, or by a horse against which the betting was 2 to 1, find the value of P's
15.

P

won by

D

expectation.

a bag containing n balls, all either white or black, all numbers of each being equally likely, a ball is drawn which turns out to be white; this is replaced, and another ball is drawn, which also turns out to be white. If this ball is replaced, prove that the chance
16.

From

of the next

draw giving a black

ball is - (n

— 1)

(2n + l)~ l

.

purses, n into each, coins have been distributed into coins will be found in the same find (1) the chance that two specified
17.

If

mn

m

(2) what the chance becomes when r purses have been and found not to contain either of the specified coins. examined

purse; and

A, B are two inaccurate arithmeticians whose chance of solving a given question correctly are -£ and y1^ respectively if they obtain the same result, and if it is 1000 to 1 against their making the same mistake, find the chance that the result is correct.
18.
;

witnesses, each of whom makes but one false statement in six, agree in asserting that a certain event took place ; shew that the odds are five to one in favour of the truth of their statement, even
19.

Ten

although the a

'priori probability of the

event

is

as small as ^9

—r

.

.

1 :

PRoUAl'.ILITY.

41

1

Local Probability.
*481.

Geometrical Methods.
of

questions of Probability requires, in general, the aid of the Integral Calculus; there are, however, many easy questions which can be solved by

The application

Geometry

to

Elementary Geometry.
Example 1. From each of two equal lines of length I a portion is cut random, and removed what is the chance that the sum of the
:

off at

remainders

is less

than

I?

Place the lines parallel to one another, and suppose that after cutting, the right-hand portions are removed. Then the question is equivalent to asking what is the chance that the sum of the right-hand portions is greater than the sum of the left-hand portions. It is clear that the first sum is equally likely to be greater or less than the second; thus the required
probability
is

a

Cor.

Each

of two lines
is

is

known

to be of length not exceeding
Z

I:

the

chance that their sum

not greater than

is

a

Example 2. If three lines are chosen at random, prove that they are just as likely as not to denote the sides of a possible triangle.
one must be equal to or greater than each of the other two denote its length by I. Then all we know of the other two lines is that the length of each lies between and /. But if each of two lines is known to be of random length between and 1, it is an even chance that their sum is greater than /. [Ex. 1, Cor.]
lines
;

Of three

Thus the required

result follows.

Example 3. Three tangents are drawn at random to a given circle shew that the odds are 3 to 1 against the circle being inscribed in the triangle formed by them.

P
O

Draw
draw

three random lines P, (), 11, in the same plane as the circle, to the circle the six tangents parallel to these lines.

and

H. H.A.

2G

.

402
Then

HIGHER ALGEBRA.

of the 8 triangles so formed it is evident that the circle will be escribed to 6 and inscribed in 2 and as this is true whatever be the original directions of P, Q, R, the required result follows.
;

Questions in Probability may sometimes veniently solved by the aid of co-ordinate Geometry.
*4:82.

be con-

Example.

random:
coincide.

a rod of length a + b+c, lengths a, b are measured at find the probability that no point of the measured lines will

On

Let AB be the line, and suppose AP = x and PQ = a; also let a be measured from P towards B, so that x must be less than b + c. Again let AP' = y, P'Q' = b, and suppose P'Q' measured from P' towards B, then y must be less than a + c. Now in favourable cases we must have AP'>AQ, or else AP>AQ\

hence Again for

y>a + x,
all

or

x>b + y

(1).

the cases possible,

we must have

x>0, and
2/>0, and

<& + c) <a + c)
make OX
equal to
b

Take a pair of rectangular axes and equal to a + c.

+ c, and
and the

OY
line

Draw

x = b + y represented by KB.

the line y

= a + x,

represented by

TML

in the figure;

Q
•A

n

P'

W

B

f
031,

-0.—Q

b

-

K
a.

X

Then YM,

EX are each equal to c,

OT are

each equal to

conditions (1) are only satisfied by points in the triangles and ItXR, while the conditions (2) are satisfied by any points within the rectangle OX, OY;
.*.

The

MYL

the required chance

=

c2

{a + c)(b

+ c)
with some Miscellaneous

*483.

We
A

shall close this chapter

Examples.
balls are

Example 1. thrown

box

is

at

random

partments each containing a balls, q compartments each containing r compartments each containing c balls, and so on, where

divided into equal compartments into which n find the probability that there will be p com;

m

b balls,

Z>a+qb + rc +

=n.

PROBABILITY.

403

Since each of the n halls can fall into any one of the m compartments n and these are the total number of cases which can occur is all equally , likely. To determine the number of favourable cases we must find the number of ways in which the n balls can be divided into p, <1, r, ... parcels containing a, b, c, ... balls respectively.

m

First choose

any

g

of the compartments,
this can be

where
is
-.

s

the

number

of

ways in which

done

\s

—\m-s ——

stands for \m

p + q + r + ...
-v

;

(1).
'

Next subdivide the s compartments into groups containing p q, severally; by Art. 147, the number of ways in which this can be done is
t

r, ...

\»\1

r

(2).
..

Lastly, distribute the n balls into the compartments, putting a into each of the group of p, then b into each of the group of q, c into each of the group of r, and so on. The number of ways in which this can be done is
In

(\a)*(\b)«(\c_)

-

(3).

of ways in which the balls can be arranged to satisfy the required conditions is given by the product of the expressions (1), (2), (3). Therefore the required probability is

Hence the number

\m

t
\

m"

(\a)>>

(\b)i ([£)-•

|p|j|r.

m-p-q-rmade in succession, find the chance that the the balls are replaced after
:

Example 2. A bag contains n balls and the ball on each occasion is found
;

;

k drawings are

to be white

next drawing will give a white ball'; (i) when each drawing (ii) when they are not replaced.
(i)

Before the observed event there are n + 1 hypotheses, equally likely; bag may contain 0, 1, 2, 3, ... n white balls. Hence following the notation of Art. 471,
for the
1

= -Pj — P-2 = P 3 =

.

.

.

= Pn

;

Hence

after the observed event,

Qr = l»+2*+ 3*+. ..+n*

7-*

Now

the chance that the next drawing will giv< a white ball xt give

=2 - Qr

\

thus the required chance

=
n

p + 2* + 3* + ...+n*
Art. 405.

and the value of numerator and denominator may be found by

26—2


404

1

'

HIGHER ALGEBRA.

In the particular case when k = 2,
the required chance

=- <—*-=

S

-4-

^

3 (n + 1)

~2(2n+l)'
If

n

is

indefinitely large, the chance
1

is

equal to the limit,

when n

is in-

V^
'

2
'

«*+!
fc

finite, of

n
is

k

+2
fe

+ 1.'

and thus the chance

+1

&+2*
not replaced,
r r
"

(ii)

If the halls are

r

-

1
'

r- 2
"
'

r

-

A;

+

n n— 1 n- 2

»— k+1

and

Q r =i it

p,
r

(r-k + l)(r-k + 2)
r=K
r=0

(r-l)r
(r-l)r

y (>--£+l)(r-ifc + 2)

(u-ifc

+ l)(n-Jfe + 2)

(n-l)n (n+1)

The chance

that the next drawing will give a white ball=

2
r =0

.

U—

Qr

Ii

(;i

-

A) (u

-

/c

+ 1)
fc

?i

(n + 1) r=0
*

s"(r-fc)(r-fc+l)
(/i-A-)(n-/v

(r-l)r
n(»i

_

+1
n(n +
l)

(n-k)(n-k + l)

+ l) 2~ k~+

+ l)

Jfc+1

~k + 2'
which
is

independent of the number of balls in the bag at

first.

Example 3. person writes n letters and addresses n envelopes ; if the letters are placed in the envelopes at random, what is the probability that every letter goes wrong ?
Let un denote the number of ways in which all the letters go wrong, and represent that arrangement in which all the letters are in their let abed own envelopes. Now if a in any other arrangement occupies the place of an assigned letter b, this letter must either occupy a's place or some other.
.

A

. .

Suppose b occupies a's place. Then the number of ways in which all the remaining n - 2 letters can be displaced is u n _ 2 and therefore the numbers of ways in which a may be displaced by interchange with some one of the other n- 1 letters, and the rest be all displaced is (n - 1) «„_ 2
(i)
,
.

PROBABILITY.

405

(ii) Suppose a occupies i>'s place, and b does not occupy a's. Then in arrangements satisfying the required conditions, since a is fixed in &'s place, the letters b, c, d, ... must be all displaced, which can be done in h__j ways; therefore the number of ways in which a occupies the place of another letter but not by interchange with that letter is (n - 1) u n - l ;
.-.

v n = (n-l)
4-44,

(M n _!

+ «„_„);

from which, by the method of Art.
Also n 1

we

find u n - nu n _ 1

=

(

-

l) n (ttj

-

Uj).

= 0,

tig

=1

;

thus we finally obtain
,

f

1

i

i

(- 1 )'

1

!

the total number of ways in which the n things can be put in n therefore the required chance is places is In
;

Now

11 +
[2

1
|4

_

'••

+

(- 1)"
in
'

and in lias maintained a permanent place some of its many It was first discussed in works on the Theory of Probability. by Montmort, and it was generalised by De Moivre, Euler, and
Tlie

problem

liere

involved modifications

is

of considerable interest,

Laplace.

*484. The subject of Probability is so extensive that it is impossible here to give more than a sketch of the principal algebraical methods. An admirable collection of problems, illustrating every algebraical process, will be found in "NVliitworth's Choice and Chance; and the reader who is acquainted with the Integral Calculus may consult Professor Crofton's article Probability in the Encyclopcedia JJritannica. complete account of the origin and development of the subject is given in Todhunter's History of the Theory of Probability from the time of Pascal to that of Laplace.

A

the theory of Probability to commercial transactions are beyond the scope of an elementary treatise ; for these we may refer to the articles Annuities and Insurance in the JEncyclopcedia Britannica.

The

practical applications of

^EXAMPLES. XXXII.
L
What

e.

are the odds in favour of throwing at throw with two dice ?
2.

lea.st 7 in

a single

In a purse there are 5 sovereigns and 4 shillings. If they are drawn out one by one, what is the chance that they come out sovereigns und shillings alternately, beginning with ;t sovereign?

:

406
3.

HIGHER ALGEBRA.
what

is

If on an average 9 ships out of 10 return safe to port, the chance that out of 5 ships expected at least 3 will arrive 1
4.

the tickets are blanks but one; each person draws a ticket, and retains it shew that each person has an equal chance of drawing the prize.

In a lottery

all

:

One bag contains 5 white and 3 red balls, and a second bag 5. contains 4 white and 5 red balls. From one of them, chosen at random, two balls are drawn find the chance that they are of different colours.
:

throw a die in the order named Five persons A, B, C, B, until one of them throws an ace find their relative chances of winning, supposing the throws to continue till an ace appears.
6.
:

E

is

Three squares of a chess board being chosen at random, what the chance that two are of one colour and one of another 1
7.
8.

one the common cube, and the other a regular tetrahedron, the number on the lowest face being taken in the case of the tetrahedron find the average value of the throw, and compare the chances of throwing 5, 6, 7.

A person throws

two
;

dice,

A's skill is to 2?'s as 1 3 9. 3 ; to Cs as 3 2 ; and to Z)'s as 4 find the chance that in three trials, one with each person, will succeed twice at least.
: : :

A

10.

A

certain stake is to be
:

won by
if

an ace with an octahedral die chance of the last ?
11.

the first person there are 4 persons

who throws
what
is

the

Two

players A,

wants 2 games

of equal skill are playing a set of games ; A to complete the set, and wants 3 games: compare

B

B

their chances of winning.

purse contains 3 sovereigns and two shillings a person draws one coin in each hand and looks at one of them, which proves to be a sovereign shew that the other is equally likely to be a sovereign or a shilling.
12.
:

A

;

A and B play for a prize A is to throw a die first, and is to he throws 6. If he fails B is to throw, and to win if he throws If he fails, A is to throw again and to win with 6 or 5 or 4, 6 or 5. and so on find the chance of each player.
13.
if
;

win

:

Seven persons draw lots for the occupancy of the six seats in 14. a first class railway compartment find the chance (1) that two specified persons obtain opposite seats, (2) that they obtain adjacent seats on
:

the

same
15.

side.

A

number

consists of 7 digits
is

chance of
16.

its

being divisible by 11

.4 —

whose sum
.

is

59

;

prove that the

Find the chance of throwing 12 in a single throw with 3

dice.

PROBABILITY.
17.

407
0, 1, 2, ...G

A

bag contains

7 tickets

marked with the numbers
;

ticket is drawn and replaced find the chance that respectively. after 4 drawings the sum of the numbers drawn is 8.

A

There are 10 tickets, 5 of w hich are blanks, and the others are marked with the numbers 1, 2, 3, 4, 5 what is the probability of drawing 10 in three trials, (1) when the tickets are replaced at every
18.
T
:

trial, (2) if

the tickets are not replaced

?

19.

If

n

integers taken at

random

are multiplied together, shew
is 1, 3, V,

that the chance that the last digit of the product
An _

or 9

is


o
.pi

;

the chance of

its

being
is

2, 4, 6,

or 8 is

—=——

9>i
;

Kn

of its being 5 is

and of
20.

its beinc:

10 H -8'l -5 n + 4 n 10*

purse contains two sovereigns, two shillings and a metal dummy of the same form and size ; a person is allowed to draw out one at a time till he draws the dummy find the value of his expectation.
:

A

sum of money persons A, B, C who first throws to throw in the order named until
21.

A

certain

be given to the one of three 10 with three dice; supposing them the event happens, prove that their
is to
.

chances are respectively

/8\ 2
(ja)'
22.

56

W>

and

/7\ 2 [&)'

Two
5

persons, whose probabilities of speaking the truth are

2 -

and -

respectively, assert that a specified ticket has been

drawn out

of a bag containing 15 tickets: the assertion ?

what

is

the probability of the truth of

23.

A

bag contains

-

counters, of which one
;

is

marked

1,

two are marked 4, three are marked 9, and so on a person puts in his hand and draws out a counter at random, and is to receive as many
shillings as the

number marked upon

it

:

find the value of his ex-

pectation.
24.

If 10 things are distributed

among

3 persons, the chance of
is
_

a particular person having more than 5 of them

....

.

If a rod is marked at random in n points and divided at those points, the chance that none of the parts shall be greater than
25.

— th n

of the rod

is

an

.

.

; ;

408
26. shilling,

HIGHER ALGEBRA.

There are two purses, one containing three sovereigns and a and the other containing three shillings and a sovereign. A coin is taken from one (it is not known which) and dropped into the other and then on drawing a coin from each purse, they are found to be two shillings. What are the odds against this happening again if two more are drawn, one from each purse 1
If a triangle is formed by joining three points taken at random in the circumference of a circle, prove that the odds are 3 to 1 against its being acute-angled.
27.

Three points are taken at random on the circumference of a what is the chance that the sum of any two of the arcs so determined is greater than the third ?
28. circle:

into three parts, that they form the sides of a possible triangle ?
29.

A

line is divided at

random

what

is

the chance

Of two purses one originally contained 25 sovereigns, and the other 10 sovereigns and 15 shillings. One purse is taken by chance and 4 coins drawn out, which prove to be all sovereigns what is the chance that this purse contains only sovereigns, and what is the probable value of the next draw from it?
30.
:

31. find the 32.

a straight line of length a two points are taken at random chance that the distance between them is greater than b.

On

A straight
If

line of length
;

points taken at
33.

random

a is divided into three parts by two find the chance that no part is greater than b.

on a straight line of length a + b two lengths a, b are measured at random, the chance that the common part of these lengths
shall not exceed c is
c —r
2
,

ab

where

c is less

than a or b

;

also the chance

that the smaller length b

lies entirely

within the larger a

is


(Jj

.

on a straight line of length a + b + c two lengths a, b are measured at random, the chance of their having a common part which
34.
If

shall not exceed

d

is

T
(c

+ a)(c+6)'

.

.

7

.

,

where d

is less

than either a or

b.

Four passengers, A, B, C, D, entire strangers to each other, are 35. travelling in a railway train which contains I first-class, second-class, and n third-class compartments. and are gentlemen whose respective a priori chances of travelling first, second, or third class are represented in each instance by X, fi, v, C and are ladies whose similar a priori chances are each represented by I, m, n. Prove that, for all values of X, fi, v (except in the particular case when

m

A

B

D

X

:

p

:

v=l.

:

m

:

oi),

A

and

B

are

company

of the

same lady than

likely to be found both in the each with a different one.

more

;

CHAPTER

XXXIII.

Determinants.

devoted to a brief discussion of determinants and their more elementary properties. The slight introductory sketch here given will enable a student to avail himself of the advantages of determinant notation in Analytical Geometry, and in some other parts of Higher Mathematics fuller information on this branch of Analysis may be obtained from Dr Salmon's Lessons Introductory to the Modern Higher Algebra, and Muir's Theory of Determinants.
485.

The

present chapter

is

48G.

Consider the two homogeneous linear equations

a x+
]

b
l

y=

0,

a 2 x + b 2 y = 0;
multiplying the first equation by bsi the second tracting and dividing by x, we obtain

by 6

,

sub-

This result

is

sometimes written
a
x

b
x

-0,

a„

b„

and the expression on the left is called a determinant. It consists of two rows and two columns, and in its expanded form each term is the product of two quantities; it is therefore said to be
of the second order.

« determinant, and
letters

The

,

b

tile

aa b 2 are called the constituents of the terms «,/>,,, ab. are called the elements,
,

410
487.

HIGHER ALGEBRA.
Since
a.

a

A - «A =

a.

a„

a.

follows that the value of the determinant is not altered by changing the rows into columns, and the columns into rows.
it

488.

Again,

it is

easily seen that

«

DETERMINANTS.
hence
*i
(I.

411

«.,

ft..

*,

K
C is
2

\
C3

a
that
is,

^3

C3

<\

the value

of

the

rows into column*, and
491.

not altered by changing the the columns into rov)S.

determinant

From

the preceding article,

=«,
c2
a.,

6.

412
492.

HIGHER ALGEBRA.
The determinant
a
x

bl
h2 b3

c

x

<**

C2 C3

%
= «, (K C3 ~ K C 2) +
="
hence
«,

h l ( C 2«3

b l(a2CB

~

«« C »)

~ Cl

l

- C /h) + C (« " «3 6 2 ) C C C b (K a 3 ~ 6A) ( 2 3 - J>>) ,

A

l

>

K
b2
c.

bx

ax
a
(i

a2

3

C

3

appears that if two adjacent columns, or rows, of the determinant are interchanged, the sign of the determinant is changed, but its value remains unaltered.

Thus

it

If for the sake of brevity

we denote
b
t

the determinant

a
(l

x

cx
C
2

2

«3

h K

°3

by

(a b 2 c 3 ), then the result
x

we have

just obtained
s)-

may be

written

(VsO = - (»Ac
Similarly

we may shew
(

that

c

i«A) = - ( a

Ab

3)

= + («A C 3 )of
tlie

493. If two rows or two columns identical the determinant vanishes.

determinant are

For let D be the value of the determinant, then by interchanging two rows or two columns we obtain a determinant whose value is — D; but the determinant is unaltered; hence J) = — D, that is D = 0. Thus we have the following equations,
a A, —
1 1

aA
2

2

n

+ aJL m = D. 3 3

494. If each constituent in any row, or in any column, is multiplied by the same factor, then the determinant is multiplied

by that factor.

DETERMINANTS.
For
tna
x

413

by

ma ma

2

a

414
These results

HIGHER ALGEBRA.

be generalised; thus if the constituents of the three columns consist of m, n, p terms respectively, the determinant can be expressed as the sum of mnp
easily

may

determinants.

Example

1.

Shew

that

b
c

+c

+a
+b

a- b b-c
c

a
b
c

— Babe -a 3 - b 3 — c 3

.

a

—a

The given determinant
b

:

DETERMINANTS.

415

and the last two of these determinants vanish [Art. 494 Cor.]. Tims we see that the given determinant is equal to a new one whose constituents of first column is obtained by subtracting from the the first column of the original determinant equimultiples of the
corresponding constituents of the other columns, while the second

and third columns remain unaltered.
Conversely,

a + j)b + qc
{ x

b
}
{

C2

a + PK +
.,

<7 C 2

^2
c„

a

c.

and what has been here proved with reference to the first column hence it appears is equally true for any of the columns or rows that in reducing a determinant we may replace any one of the rows or columns by a new row or column formed in the following
;

way
of the row or column to be replaced, and increase or diminish them by any equimidtij)les of the corresponding constituents of one or more of the other rows or columns.

Take

the constituents

After a little practice it will be found that determinants may often be quickly simplified by replacing two or more rows or columns simultaneously for example, it is easy to see
:

that

a +2
i

}

b

}

b

}

- qc

x

c,

€l

l

% + Ph

2

1 K ~ CC

2

°2

Ct

2

b2
b„
c.,

a3 +2 jb 3

K-Q c

s

c

a..
3

but in any modification of the rule as above enunciated, care must be taken to leave one row or column unaltered.

on the left-hand side of the last identity the constituents of the third column were replaced by c l +rali c 2 + ra^ respectively, we should have the former value inc, + ra creased by
Thus,
if

a + 2>b
x

t

b

x

— qc

ra
x
x

ra„

« a + i'K

K - vc

ra..
i

416
and

HIGHER ALGEBRA.
of the four determinants into
is

there

which this may be resolved one which does not vanish, namely
ra,

pbs

-

qc 2

ra 9 „
.

Example

1,

Find the value of

I

29 25

26
31

22

27 46

!

63

54

The given determinant
3

26-4l = -3x4x
31

1

-6
9

54

-4 -8

DETERMINANTS.

417

[Explanation. In the first new determinant the first row is the sum of the constituents of the three rows of the original determinant, the second and third rows being unaltered. In the third of the new determinants the first column remains unaltered, while the second and third columns are obtained by subtracting the constituents of the first column from those of the second and third respectively. The remaining transformations are sufficiently obvious.]

Before shewing how to express the product of two determinants as a determinant, we shall investigate the value of
497.

»A + h A + c i?2 Vi + KPt + c*y, %% + h A + c*y a a + h A + Wi %% + b A + ^ y
«i«i

+

&

A + ^7,

a a3 + b
i

A+

r
i7-s

2

« 2 <* 3 +
«3 a 3

h

A+

c

a
3

ya
3

3

i

3

2

+

&A + c y

Art. 495, we know that the above determinant can be expressed as the sum of 27 determinants, of which it will be sufficient to give the following specimens
:

From

*1<S

a a
x
{

« 2 ft 3

a3a 2

«3 a 3

418
where

HIGHER ALGEBRA.
l

X = a^ + a x \ ^=A*i+/W we have Substituting for X, and X in
2
2

.(2).

2

(1),

(a^ + 6^)
(at a

x,

+
x

b2 P

}

)

+ (a^+bfij x 2 = 0\ x + (a 2 a, + bfi 2 ) x2 = Oj
x

(3).

In order that equations values of x and x2 other than
x

a^
But equations

may simultaneously zero, we must have = + bfr a a + bfi
(3)
x

hold for

2

2

(4).

(3) will

hold

if

equations (1) hold, and this

will be the case either if
(5),

aa
or
if

b2

X

=
l

and

X

2

= 0;
.(6).

which

last condition requires that
a.

ft

=o

a.

ft

Hence if equations (5) and (6) hold, equation (4) must also hold and therefore the determinant in (4) must contain as
;

factors the determinants in (5) and (6) ; and a consideration of the dimensions of the determinants shews that the remaining

factor of (4)

must be numerical

;

hence
«i

al

a
i

+

&

A

»i a,

+

h

A

the numerical factor, by comparing the coefficients of afyafl, on the two sides of the equations, being seen to be unity.

DETERMINANTS.

419

4

420
12.

HIGHER ALGEBRA.
Without expanding the determinants, prove that x y z b a b c 1

x

y

z

x
z

a
c

P
r

P
a
b

r
c

P
13.
1)

r

Solve the equations

14.

15.

1

a
b
c
1

a" b2
i

= (b-c)(c- a)

(a

-

b).

1 1

16.

1

1 c c3

=

(b

- c)

(c

- a)

(a

- b)

(a

+ b + c).

a
a?
17.

b

x
x2
yz

y

6
*2

= (y-z)(z-x) (x -y)(yz + zx + xy).

y

2

zx

xy
a+b

18.

-la b+a
c

a+c
b

4(b + c)(c + a)(« + b).

+a

-26 c+b

+c

-2c
a-

19.

(b+cY
62
^2 {c

2dbc{a+b+cf.

+ af

62

(a+bf
f
c
c

20.

Express as a determinant

b

b

DETERMINANTS.
22.

421

Fi

422

HIGHER ALGEBRA.
usefully em-

The properties of determinants may be 499. ployed in solving simultaneous linear equations.
Let the equations be

+ c z + d = 0, a 2X + b 2 y + c 2 z + d2 = 0, aja + bg + cji + d^O;
ax+
x

h xy

x

x

multiply them by A lt -A s A.A respectively and add the results, A , A A 3 being minors of alt a2i aa in the determinant
, j

D=

a,

*,

a„

«3

^3

^

The

coefficients of

y and z vanish in virtue of the relations proved

in Art. 493,

and we obtain

(Mi
Similarly

-

M* + M

3)

*+

(M

i

-

M

2

+

Mb) =

°-

we may shew

that
2/

(6,5,

-6

A+
t

6

A)
)

+ (<*A - <*A + <*A) = +
(dfi,

0,

and
fcff,

- e,C, + c 3 C3

- dfia + d3 C,) = 0.

Now

«A - a A, + «A - - (6,5, - 6,5, + 6A)
may be
written
z

hence the solution

x
<*,

—y

DETERMINANTS.
a x + b y + c,s + d u — a2x + b 2 y + c 2 z + d2u = a 3 x + b 3 y + c 3 z + d3u = ax + b W + c 4z + du = y 4 4
x
x x

423
0,
0,

0, 0.

A

A

From the
x

last three of these,

we have

as in the preceding article
z

—y

—u

K

424

HIGHER ALGEBRA.
is

The left-hand member of this equation consists of n rows and n columns, and is
the
th II
-

a determinant which called a determinant of

order.

discussion of this more general form of determinant is beyond the scope of the present work ; it will be sufficient here to remark that the properties which have been established in the case of determinants of the second and third orders are quite general, and are capable of being extended to determinants of

The

any

order.

For example, the above determinant
equal to

of

the
1

n th

order

is

a
or
ct

1

l

A -b B + c C -d D 1+ A -a A + a A -a4A +
1

...
...

1

l

1

1

1

+ (-l)"- k
1

1

K

l,

1

2

2

B

3

4

+ (-1)"- a n A n

,

according as we develop it from the first row or the first column. Here the capital letters stand for the minors of the constituents denoted by the corresponding small letters, and are themselves Each of these may be exdeterminants of the (n-l) th order. th pressed as the sum of a number of determinants of the (n — 2) order ; and so on ; and thus the expanded form of the determinant may be obtained.

Although we may always develop a determinant by means of
the process described above, especially when our object the whole determinant, as elements.
502.

not always the simplest method, is not so much to find the value of to find the signs of its several
it is

The expanded form
a
l

of the

determinant

DETERMINANTS.
cording as
it

425

can be deduced from the leading element by an even or odd number of permutations of two suffixes ; for instance, the element a 3 b 2 c is obtained by interchanging the suffixes 1 and 3, therefore its sign is negative ; the element ajb c 2 is obtained by first interchanging the suffixes 1 and 3, and then the suffixes 1 and 2, hence its sign is positive.
1 1

The determinant whose leading element may thus be expressed by the notation
503.

is

a
x

b 2 c 3 dA

...

%^aJ>aeBdA
the
of all the elements

,

2 * placed before the leading element indicating the aggregate
which can be obtained from suffixes and adjustment of signs.
it

by suitable

interchanges of

Sometimes the determinant is still more simply expressed by enclosing the leading element within brackets; thus (a^crf ...) is used as an abbreviation of 5 ± a,b„c„d ....
A

Example.
the element

In the determinant

(a^c^e^ what

sign is to be prefixed to

a^c^e.,1

the leading element by permuting the suffixes of a and d we get from this by permuting the suffixes of b and c we have a 4 b 3 c 2 d 1 e 5 ; by permuting the suffixes of c and d we have a i b.i c 1 d2 e 5 finally by permuting the suffixes of d and e we obtain the required element rt 4 & 3 c 1 rf 5 ? 2 and since we have made four permutations the sign of the element is positive.
a 4 b2c 3 d x e 5
; ; ;

From

504.

If in Art. 501, each of the constituents 6
;

equal to zero the determinant reduces to a A it is equal to the product of a and a determinant of the (n — l) th order, and we easily infer the following general theorem.
y

k is in other words
,

c

,

...

the first row or column of a determinant is zero except the first, and if this constituent is equal to m, the determinant is equal to times that determinant of lower

If each of

the constituents

of

m

order ivhich is obtained by omitting the first column row,

and

first

Also since by suitable interchange of rows and columns any constituent can be brought into the first place, it follows that if any row or column has all its constituents except one equal to zero, the determinant can immediately be expressed as a determinant of lower order.
This is sometimes useful in the reduction and simplification of determinants.

426
Example.

HIGHER ALGEBRA.
Find the value
of

DETERMINANTS.

427

EXAMPLES.

XXXIII.

b.

Calculate the values of the determinants
1.

1
I 1
1

1

1

2

3 4

3.

a
1 1

:

b

k

428
Solve the equations
12.

HIGHER ALGEBRA.

x+ y+ 0=1, ax + by + cz=k, a 2x + b y + c z = l
2

13.

2

2
.

ax + by + cz=k, a2x + 2y + c2z = 2 a?x + b 3y + c?z= P.

,

14.

x + y+ z+ u=l, ax + by + cz+ du=k, a2x + b 2y + c2z + d2 u = k 2 3 s 3 cfix + b y + c z + d u = P.
,

15.

Prove that

b+c—a—d c+a—b—d a+b-c-d

be

— ad ca — bd ab — cd

+ d)-ad(b-t-c) ca(b-\-d) — bd(e + a) ab (c + d) -cd{a + b)
be (a
\

= -2 (b- e) (c-a) (a-b) (a-d) (b - d) (c-d).
16.

Prove that

a2
b2
c
2

a 2 -(b — c) 2
b2
c2

be

-(c-a) 2
- (a b) 2

ca

ab
).

= (b-c)(c-a){a-b)(a + b + c)(a 2 + b 2 + c2
17.

Shew

that

a

b

c

d
c

e

f
e

f
e

«

b

d
c

ABC CAB
B C A

f
e

a

b
cc

d
c

d
c

f
e

b

d
c

f
e

a

b

b

where

f a A=a 2 -d + 2ce
2

d

-2bf,

B=e -b
2 2

2

+2ac-2df,

C=e -f + 2ae-2bd.
2

If a determinant is of the ?i th order, and if the constituents of its first, second, third, ...nth rows are the first n figurate numbers of the first, second, third, ...nth orders, shew that its value is unity,.
18.

CHAPTER XXXIV.
MISCELLANEOUS THEOREMS AND EXAMPLES.
with some remarks on the permanence of algebraical form, briefly reviewing the fundamental laws which have been established in the course of the work.
506.
shall begin this chapter

We

In the exposition of algebraical principles 507. at the outset we do not lay down new analytically
:

we
of

proceed

names and

new

abstract Arithmetic ; we prove certain laws of operation which are capable of verification in every particular case, and the general theory of these operations constitutes the science of Algebra.
ideas,

but

we begin from our knowledge

usual to speak of Arithmetical Algebra and SymIn the bolical Algebra., and to make a distinction between them. former we define our symbols in a sense arithmetically intelligible, and thence deduce fundamental laws of operation ; in the latter we assume the laws of Arithmetical Algebra to be true in all cases, whatever the nature of the symbols may be, and so find out what meaning must be attached to the symbols in order that they may obey these laws. Thus gradually, as we transcend the limits of ordinary Arithmetic, new results spring up, new language has to be employed, and interpretations given to symbols which were not contemplated in the original definitions. At the same time, from the way in which the general laws of Algebra are established, we are assured of their permanence and universality, even when they are applied to quantities not arithmetiit is

Hence

cally intelligible.

Confining our attention to positive integral values of the symbols, the following laws are easily established from a priori
508.

arithmetical definitions.

:

430
I.
(i)

HIGHER ALGEBRA.
The Law of Commutation, which we enunciate as follows Additions and subtractions may be made in any order.
a+

Thus
(ii)

b-c = a-c + b = b-c + a.
and
divisions

Multiplications

may

be

made in any

order.

Thus

axb=bxa; axbxc = bxcxa = axcxb'
ab-±- c

}

and so

on.

=a

x b

-=-

c

=

(a

-f-

c)

x b

= (b + c)

xa.
as follows :

II.

The Law of Distribution, which we enunciate and
divisions

|

Multiplications

may

be distributed over additions

and

subtractions.

Thus

{a
(a

-b+
b)(c

c)

m = am —bm + cm,
ac

— d) =

— ad —

bc

+

bd.

[See Elementary Algebra, Arts. 33, 35.]
since division is the reverse of multiplication, the distributive law for division requires no separate discussion.
III.
(i)

And

The Laws of Indices.
n m+n am xa = a
3

a +a =a
(n)

m

H

m -".

[a

)

=a

.

[See Elementary Algebra, Art. 233 to 235.]

These laws are laid down as fundamental to our subject, having been proved on the supposition that the symbols employed are positive and integral, and that they are restricted in such a way that the operations above indicated are arithmetically intelligible. If these conditions do not hold, by the principles of Symbolical Algebra we assume the laws of Arithmetical Algebra to be true in every case and accept the interpretation to which this assumpBy this course we are assured that the laws of tion leads us. Algebraical operation are self-consistent, and that they include in their generality the particular cases of ordinary Arithmetic.

From the law of commutation we deduce the rules 509. for the removal and insertion of brackets [Elementary Algebra, Arts. 21, 22] ; and by the aid of these rules we establish the law

; ;

MISCELLANEOUS THEOREMS AND EXAMPLES.
of distribution as in Art. 35.

431

For example,

it is

proved that

(a

-b)(c — d)~ac — ad—bc +

bd,

with the restriction that a, b, c, d are positive integers, and a Now it is the province of greater than b, and c greater than d. Symbolical Algebra to interpret results like this when all restricHence by putting a = and c = 0, we obtain tions are removed. (— b) x (— d) = bd, or the product of two negative quantities is and c= 0, we obtain a x (—d) =—a<I, positive. Again by putting 6 = or the product of two quantities of opposite signs is negative.

"We are thus led to the Rule of Signs as a direct consequence of the law of distribution, and henceforth the rule of signs is included in our fundamental laws of operation.
510. For the way in which the fundamental laws are applied to establish the properties of algebraical fractions, the reader is referred to Chapters xix., xxi., and xxn. of the Elementary Algebra

there be seen that symbols and operations to which we cannot give any a priori definition are always interpreted so as to make them conform to the laws of Arithmetical Algebra.
it will

Chapter xxx. Elementary Algebra. When m and n are positive integers and m > n, we prove directly from the definition of an index that
511.
of indices are fully discussed in
of the

The laws

a

m

xa n = am+n

;

am

1

-r a'

= am

~n
j

(a

m
)

n

= am ".

We

then assume the first of these to be true when the indices are free from all restriction, and in this way we determine meanings for symbols to which our original definition does not apply.
p

The interpretations

a~" thus derived from the first law are found to be in strict conformity with the other two laws and henceforth the laws of indices can be applied consistently and with perfect generality.
for a\ a
,

512.

In Chapter

vill.
2

obeying the relation i From this definition, and by making i subject to the general laws of Algebra we are enabled to discuss the properties of expressions of the form a + ib, in which real and imaginary quantities are combined. Such forms are sometimes called complex numbers, and it will be seen by reference to Articles 92 to 105 that if we perform on a complex
.

we =— 1

defined the symbol

i

or

J— 1

as

number the operations of and division, the result is

addition, subtraction, multiplication, in general itself a complex number,

432

HIGHER ALGEBRA.

Also since every rational function involves no operations but those above mentioned, it follows that a rational function of a complex number is in general a complex number.
x+ly \og(x±iy) cannot be fullyExpressions of the form a treated without Trigonometry; but by the aid of De Moivre's theorem, it is easy to shew that such functions can be reduced to complex numbers of the form A + iB.
,

x+iy is of course included in the more general The expression e x+i form a \ but another mode of treating it is worthy of attention.

We

have seen in Art. 220 that

(x\ — nj
1 H
;

n
,

)

when n

is infinite,
x+i!/

x being any real quantity the quantity defined by means of the equation
*+iy

e

may be

similarly

e

= Lini

(1 \

H

n

)

,

when n

is infinite,

J

x and y being any
found fully

real quantities.

The development

complex numbers will be discussed in Chapters x. and XI. of Schlomilch's
of the theory of

Handbuch

der algebraischen Analysis.

513. shall now give some theorems and examples illustrating methods which will often be found useful in proving identities, and in the Theory of Equations.
514.

We

of x

is

To find the remainder ivhen any rational integralfunction divided by x - a.

Let fix) denote any rational integral function of x ; divide f(x) hyx-a until a remainder is obtained which does not involve x ; let Q be the quotient, and R the remainder then
;

f(x)
Since value

= Q(x-a) + R.

R
we

does not involve x it will remain unaltered whatever give to x ; put x = a, then

f(a)

= QxO + R;

now Q

is finite

for finite values of x, hence

:

.MISCELLANEOUS THEOREMS AND EXAMPLES.
Cor.

433

f(a) =

;

== 0, that is If f{x) is exactly divisible by x - a, then hence if a rational integral function of x vanishes when

R

x—

a,

it

is divisible

by x -

a.

515.

The proposition contained

in the preceding article
it

is

so

useful that

give another proof of of exhibiting the form of the quotient.

we

which has the advantage
be

Suppose that the function is of n dimensions, and denoted by "~ }+ 2 n 3+ -+P»> P^"~ +P-^"" p x +2\ x
then the quotient will be of n q{f

let it

1

dimensions
n -3
2

;

denote
H_
l

it

by

c"-

x

+ qi xn

-2

+q x

+

...

+q

;

let

R

be the remainder not containing x
1
1

;

then

pjf +p x-

+2> 2 x'-

2

+p.ax"~ +
(q

3

•••

= (x-a)

X^

+P„ + qi x"~ 2 + q 2x"~ 3 +

...

+ qa _

x

)

+

R
x,

Multiplying out and equating the coefficients of like powers of we have

- a Qi=P 2 q3 2 = ihi
9. 2

>

or qs

= a4i+Pa
2

'>

n

or

& = «<i

+ ih
_l

;

R - oqn - =Pn
%

,

or

R = aq

n

+p n

;

thus each successive coefficient in the quotient is formed multiplying by a the coefficient last formed, and adding The process of finding next coefficient in the dividend. successive terms of the quotient and the remainder may arranged thus
Po
Pi
«?0

by
the the

be

P2
Ct(

P3
Cl(

Pa-X
Cl(

Pa
(l(

lx

l2

In-2

2n-l

%
Thus

q

x

%
1

v,

?.-,

x

R = aq^ +p - « («<?„-- +#.-i) + P* =
n
n

~2

=P<P* +P^"~ +P/<<
If tlie divisor is

+

•••

+P,r
in

x + a the same method can be used, only

this case the multiplier is - a.

H. H. A.

28

434

HIGHER ALGEBRA.
3a;7

is

Example. Find the quotient and remainder when divided by x + 2.

- a; 6 +

31a: 4

+ 21a; + 5

Here the multiplier

is

-

2,

and we have
31
3
7a;
5

3-10
3

-6 -7
is

14 -28 -6
14

00
12 12

21

-24
- 3

5 6

-6

11

Thus the quotient
mainder
516.
is 11.

3.r 6

-

+ 14a; 4 + 3a; 3 - 6a; 2 +12a;-3, and

the

re-

In the preceding example the work has been abridged
coefficients of the several terms, zero

by writing down only the
coefficients

being used to represent terms corresponding to powers This method of Detached Coefficients may of x which are absent. frequently be used to save labour in elementary algebraical processes, particularly when the functions we are dealing with The following is another illustration. are rational and integral.
Example.
Divide
1
3a; 5

-

8a; 4

-

5a; 3

+ 26a;2 - 33a; + 26 by

a;

3

-

2a;

2

- 4a; + 8.

+ 2 + 4-8)3-8- 5 + 26-33 + 26(3-2 + 3 3 + 6 + 12-24 -2 +

-2-

MISCELLANEOUS THEOREMS AND EXAMPLES.
we form the next horizontal
this gives 3,

435

line, and add the terms in the third column; which is the coefficient of the third term of the quotient. By adding up the other columns we get the coefficients of the terms in

the remainder. ]

Example.

Divide 6a 5 + ba*b - 8a?b 2 - 6a 2 b 3 - 6a ¥ by 2a 3 + 3a 2 6 -

bz

to four terms in the quotient.

2

436
z = 0, then 3 pansion of (x + y)

HIGHER ALGEBRA.

Put

A=
l,

3,

2 being the coefficient of x y in the ex-

.

Put x = y = z =

and we get 27 = 3 +

(3 x 6)

+B

;

whence

B = 6.
Thus (x + y +
z)
3

= x3 + y 3 + z 3 + 3x 2y + 3xy 2 + 3y 2z + 3yz 2 +
520.

3z x

3

2 + 3zx + 6xyz.

function is said to be alternating with respect to its variables, when its sign but not its value is altered by the interchange of any pair of them. Thus x — y and
2 a 2 (b-c) + b (c-a) +

A

c (a

2

-

b)

are alternating functions.

no linear alternating function involving more than two variables, and also that the product of a symmetrical function and an alternating function must be an
It is evident that there can be

alternating function.

Symmetrical and alternating functions may be concisely denoted by writing down one of the terms and prefixing the symbol % ; thus %a stands for the sum of all the terms of which a is the type, %ab stands for the sum of all the terms of which ab is the type; and so on. For instance, if the function involves
521.

four letters a,

b, c,

d

}

^a-a + b + c + d;
%ab = ab + ac + ad +bc + bd+ cd;

and so

on.
if
2

Similarly

the function involves three letters
(b

a, b,

c,

$a
and
so on.

-c) = a (b-c)±
b ca
2

2

b (c

2

- a) +

2

c

(a

- b)

%a 2 bc = a 2 bc +

+

2

c

ab;

%a

2

It should be noticed that when there are three letters involved b does not consist of three terms, but of six thus
:

2<a b

2

= a 2 b + a c + b 2c + b2 a +
2

2

c

a+

c b.

2

The symbol 2 may also be used to imply summation with regard to two or more sets of letters; thus
%yz (b-c) = yz (b~c) + zx (c-a) + xy (a - b).

MISCELLANEOUS THEOREMS AND EXAMPLES.
522.

437

The above notation enables us to express in an abridged form the products and powers of symmetrical expressions thus
:

(a+b +
(a

c)

3

= %a 3 + 32a 2 b +

Gabc

j

+

b

+ +

c

+ df = 2« 3 + 3$a 2 b + Gtabc; +
4

(a

b

c)

= %a A + i%cfb + 6Sa"6 + 1 2%a 2 bc;
fl

%a x 2« 2 = 2a 3 + %a 2 b
Example
1.

.

Prove that
(a + b) 5

- a5

-

b*

= 5ab (a + b) (a 2 + ab + b 2
;

).

vanishes

Denote the expression on the left by E then E is a function of a which hence a is a factor of E similarly 6 is a factor of E. when a = Again E vanishes when a— - b, that is a + b is a factor of E; and therefore E contains ab(a + b) as a factor. The remaining factor must be of two dimensions, and, since it is symmetrical with respect to a and b, it must be of the form Act? + Bab + Ab'z thus
;
;
;

(a

+ b) 5 - a 5 -

b5

= ab (a + 6)
and
b.

(Aa*

+ Bab + A b~),

where

A and B

are independent of a
1, b

= 1, we have 15 = 2A + B a = 2, b = - 1, we have 15 = 5A - 2B putting whence A = o, J5 = 5; and thus the required result
Putting a =
;

;

at once follows.

Example

2.

Find the
(&3

factors of

+ c 3)

(b-c)

+ (c 3 + a 3 (c-a) + (a 3 + b 3
)

)

(a

- b).

then is a function of a which vanishes therefore contains a - b as a factor [Art. 514]. Similarly it - c) (c - a) (a - b) as a contains the factors b-c and c-a; thus contains (b
;

Denote the expression by

E

E

when a = b, and
factor.

E

first

E is of the fourth degree the remaining factor must be of the degree; and since it is a sj^mmetrical function of a, b, c, it must be of the form M{a + b + c). [Art. 518];
Also since
.-.

E = M (b-c)

(c-a) (a-b)(a + b + c).

we may give to a, b, c any values that we find most conTo obtain venient; thus by putting a = 0, 6 = 1, c = 2, we find M=l, and we have the required result.
Example
(x
3.

M

Shew

that
(y

+ y + zjt-x5 -y b - z? = 5

+ z)

(z

+ x)

(x

+ y)

(x2

+ y 2 + z~ + yz + zx+ xy).

Denote the expression on the left by E then E vanishes when and therefore y + z is a factor of E; similarly z + x and x + y are
;

y=-z,
factors;

therefore

E

contains

(y

+ z)

(z

+ x)

[x

+ (/)asa

factor.

Also since

E

is

of the

438
fifth

HIGHER ALGEBRA.
it is

degree the remaining factor is of the second degree, and, since symmetrical in x, y, z, it must be of the form

A
Put»=2/=z=l;
put
thus
2

(x

2

+ y2 + z 2 + B
)

(yz

+ zx + xy)

.

10=^1+5;
thus 35

x=2, y=l,

= 0;

whence

= 5A + IB A=B = 5,

;

and we have the required

result.

collect here for reference a list of identities which 523. are useful in the transformation of algebraical expressions; many of these have occurred in Chap. xxix. of the Elementary Algebra.

We

^bc (b

— c) = —(b-c)(c- a)

(a

-

b).

$a2 (b-c) = -(b-c)(c-a)(a-b).
$a(b 2 -c 2 ) = (b-c)(c-a)(a-b).

2a 3 (b-c) = -(b-c) (c-a) (a-b) (a + b + c). s 3 3 a + b + c - 3abc = (a + b + c)(a2 +b 2 + c 2 - bc-ca- ab).
This identity

may be

given in another form,
b

a +

3

b

3

+

c

3

-3abc = l(a +

+ c){(b-c) 2 + (c-a) 2 + (a-b) 2 }.

(b-c) 3 + (c-a) 3 + (a-b) 3 = 3(b-c)(c-a)(a-b).
(a

+

b

+

c)

3

-a 3 -b -c 3 = 3(b + c)(c + a)(a + b).
3

Hbc

(b

+

c)

+ 2abc =

(b

+ c)(c + a)(a +

b).

%a 2 {b + c) + 2abc =(b + c)(c + a) (a +
(a
2

b).
b).

+
2

b

+

c) (be

+ ca + ab) - abc =(b +
2 2

c)(c

+ a) (a +

2b c

+

2c

V + 2a

b

-tf-fr-c*
b

= (a +

+

c)(b

+ c-a)(c + a-b)(a+b-c).

EXAMPLES. XXXIV.
1.

a.

Find the remainder when

by x + 5.
2.

3^ + 1 1^ + 90#2 - 19# + 53

is

divided

Find the equation connecting a and b in order that

2xi -7x3 +ax + b

may

be divisible by x - 3.

MISCELLANEOUS THEOREMS AND EXAMPLES.
3.

439

Find the quotient and remainder when
jfi

_ 5#4 + 9 iV3 _ q xi _ iq v + 13
tf _ 2xA

j >s

divided by

x2 - 3v + 2.

4.

Find a in order that x3 -7x + 5

may

be a factor of

- 4^ + 19.V2 - Six + 12 + a.
descending powers of x to four

5.

Expand

^.^^.g ^

terms, and find the remainder.

Find the factors of
6. 7.

a(6-c) 3 + 6(c-a) 3 + c(a-6)3.
a4 (6 2 - c2) + 6 4 (c2 - a 2 ) + c4 (a2 (a
6 2 ).
.

8.
9.

+ 6 + c) 3 -(6 + c-a) 3 -(c+a-6) 3 -(a + 6-c) 3
(c

a (6 - cf + & a
(6
4

- af + c(a- 6) 2 + 8a6c.

10. 11.
12.
13.

- c4 ) + b

(c 4

- a4 + c(a i - 6 4
)

).

(6c

+ ca + a6) 3 - J 3

3
**

- c%3 - a 3 6 3

.

(a + 6 + c) 4
(a + 6
(tf

-(6 + c) 4 -(c + a) 4 -(a + 6) 4 + a 4 + 6 4 + c4
.

.

+ c) 5 -(6 + c-a) 5 -(c + a-6) 5 -(a + 6-c) 5
.

14.

- a) 3 (6 - cf + (x - b) s (c - af + (x - c) 3 (a - 6) 3
:

Prove the following identities
15.

2 (6 + c - 2a) 3 = 3(6 + c- 2a) (c + a- 26) (a + 6- 2c).
c{a-bf _ a(b-cf He-*)* fl|M (c-a)(a-6r (a-6)(6-cr (6-c)(c-a)
.

i0,

g

17 '"

J^6
a+

_?L
6

2c
c

(6-c)(c-a)(a-6)_
(6

+c

+a

+ c)(c + a)(a+6)

3

18.

2 a2(& + c)-2a3 -2a&c=^& + c-a)(c + a-6)(a + &-c)-

iy

*

(a-6)(a-c)^(6-c)(6-a)^(c-a)(c-6)
42(6-c)(6 + c-2a) 2 = 92(6-c)(6 + c-a) 2
.

20.

21.
22.

ty+z)*(e+x)*(x+y)*=tx*(y+zY+2(^z)3 -2^

2z*'

^ («6 - c2 (ac- 6 2 = (26c) (26c - 2a2 ).
) )

23.
24.

«6c (2a) 3

- (26c) 3 = abc 2a 3 - 26 3 c3 = (a 2 - 6c) (6 2 - ca) (c2 - a6).
;

5(6- c) 3 (6 + c - 2a) =

hence deduce

2

- y) (£ + 7 - 2a ) 3 = °-

"

440
25.
(b

HIGHER ALGEBRA.
+ cf+(c + af+(a + bf-Z(b + c)(c + a)(a + b)

= 2(a3 -\-b 3 + c3 -3abc).
26.

If

x=b+c-a, y = c + a-b, z = a + b-c, shew that #3 + ^3 + # _ g^g = 4 (a 3 + 6 3 + c 3 - 3a6c).
if

Prove that the value of a 3 + b 3 + c3 - 3a6c is unaltered substitute s-«. s- b, s-c for a, 6, c respectively, where
27.

we

3s

= 2(a + 6 + c).

Find the value of
28
.

« w ,w i (a-b)(a-c)(x-a)
,

»_ w _ » + + /1 (b-c)(b-a)(x-b)
c2
(c

(c-a) (c-b) (x-c)'
'

29.

— cz — a 2 ;+77 7 + ft7 wt (b -c)(b — a) (a -b)(a- c)
b2
x

a 2 — b 2 — c2

— a2 -

Z>

2

-a) (c-b)

30.

(

a +P)( a + <l)

,

(a-b)(a-c)(a+x)

(b+p)(b + q) (b-c)(b-a) (b + x)
.

+

(

c

+p)(c + (J )
'

(c-a) (c-b)(c + x)

31.

3 (a -b) (a- w — d)' w c) (a ^
If

__

^

32.

s (a-

b) (a

— c) (a-

'

d)

33.
'jp

x + y + z = s, and #yz =£< 2 shew that
,

_y\(p__z\
p)\zs

,y«

A /£. _ /£ A /> y\ + fp__ pj\xs #\ + \xs _ pj\ys _ p) \zs p) p)
Miscellaneous Identities.

4
'

8

524. Many identities can be readily established by making use of the properties of the cube roots of unity; as usual these
will be denoted

by

1, w, o>

2
.

Example.

Shew that
(x

+ yf -x7 -y7 = Ixy
the
left

(x

+ y)

(x 2

+ xy + y 2

2
)
.

The
hence
it

expression, E, on

vanishes

when x = 0, y = 0, x + y = 0;

must contain xy

(x

+ y)
ta)7

as a factor.

Putting x = coy, we have

E = {(1 +
hence

- W7 -

1}

y7=

{(_

w 2)7 _ w 7 _

!} y

i

= (_ w 2 - w -l)y7 = 0;
contains x - wy as a factor and similarly tains x - ury as a factor; that is, is divisible by
;

E

we may shew that
.

it

con-

E

(x-

ury)

(x

-

2 to ?/),

or x^

+ xy + y 2

=

;

;

MISCELLANEOUS IDENTITIES.
Further,

441

seven, and xy(x + y) (x2 + xy + y2) of five dimensions, the remaining factor must be of the form A (x 2 + y'2 ) + Bxy thus
;

E being of
+ y)
7
7

(x

- x - y7 = xy {x + y) (x

2

+ xy + y (Ax + Bxy + Ay2 ).
2 2
)

Putting
putting

a;

= l, y = l, we
2,

have 21 = 2^+5;

x

y=-

1,

we have 21 = 5^1 -2B;
2
)
.

whence
.-.

A = 7, B = 7 2 2 (x + y) 7 - x 7 - y7 = Ixy (x + y)(x + xy + y

525.

We know from
b
3

elementary Algebra that
b

aa +
also

+

c

3

- 3abc = (a+
3,

+

c) (a

2

+

b

2

+

c

2

-be- ca —

ab)

;

we have
a*
3

seen in Ex.

Art. 110, that
(a

+

b

2

+

c

2

— be — ca — ab =
— 3abc can be
(a

+ ub +

ore) (a

+

2

<a

b

+

wc)

hence a + b 3 + thus

c

3

resolved into three linear factors;

a3 +

b

3

+ c 3 - 3abc =
Shew

+

b

+c) (a + mb +

arc) (a

+

2

<D

b

+

wc).

Example.

that the product of
a;

a3 + b 3 + c 3 - dabc and

3

+ y 3 + z3 - Sxyz

can be put into the form

A 3 + B 3 + C3 - SABC.
(a

The product

= [a + b + c)
x
(x

+ y + z)

+ wb + ore) (a + w 2 & + wc) (x + uy + urz) (x + w 2 y + uz).
+ b + c)
(x

By

taking these six factors in the pairs (a
(a

+ y + z);

+ u>b + w2 c)

(x

+ cry + uz)

;

and

(a

+ urb + uc)
2

(x

+ wy + urz),

we obtain the three

partial products

A + B + C, A + wB + u-C,
where

A+u B + u)C,
(A

A = ax + by + cz, B — bx + cy + az, C = cx + ay + bz.

Thus the product = (A + B +

+ wB + u 2 C) = A 3 + B 3 +C3 -SABC.
C) [A

+ orB + «C)

526.
a,
b,

a+

b

In order to find the values of expressions involving c when these quantities are connected by the equation + c = 0, we might employ the substitution a—h+
k,

b

=

ioh

+

2
(x>

k,

c

=

ufh

+

u>k.

however the expressions involve a, b, c symmetrically the method exhibited in the following example is preferable.
If

1

1

1

.

442
Example.
If a

HIGHER ALGEBRA.
+ b + c = 0, shew that 3 3 3 5 6 (a 5 + b 5 + c ) = 5 (a + 6 + c )

(a 2

+ 6 2 + c 2 ).

We have
where

identically
(1

+ ax)

(1

+ bx)

(1

+ cz) = 1 +px + qx 2 +rx3

,

p — a + b + c,

q

= bc + ca + ab,

r — abc.

Hence, using the condition given,
(1

+ ax)

(1 -t

bx)(l + cx)

= l + qx 2 + rx3

.

Taking
(~
'

logarithms
)

and equating the
of

coefficients

of

xn

,

we have

n

(a n

+ b n + c n = coefficient
of

xn in the expansion of log(l + qx 2 + rx3)
{qx2

= coefficient
By

xn in
rc

[qx2

+ rx3 - ^
)

+ rx 3 ) 2 + ^ (qx2 + rx3) 3 -

.

.

putting

= 2,

3,

5

we obtain
a3 + b 3 + c 3

a2 + b2 + c 2

j— =*•
»-

3— =r
~

a 5 + b 5 +c 5
'

T- = '

<?r;

whence
and the required
If

—=
c

«

result at once follows.

a=fi-y,

6

= 7 -a,

= a-/3,
5

the given condition
/3,

is

satisfied;

hence

we have

identically for all values of a,
}
)

y

5 6{(iS-7) 5 + (7-«) +("-/3)

= 5{(/3- 7 3 + (Y-a) 3 + (a-/3) 3 }
that
is,
5

{{§-

y?+ (y- a) 2 + (a-/S) 2

}

(/3-7)

+ (7-a) 5 + (a-^) 5 =5( 8-7)( 7 -a)(aJ

i

3)(a 2 +

^+7

2

-/37-7a-a^;

compare Ex.

3, Art. 522.

EXAMPLES. XXXIV.
If (a integer (a b
1.
,

b.

+ b + cf = a3 + bz + c3 shew that when n + + cf n + = a2n + + b 2n + + c2n + K
that
2

is

a positive

2.

Shew

(a + <ob + a)
3.

c) 3

+ (a +

Shew that

+ a>c) 3 = (2a - b - c) (2b - c - a) (2c - a - b). (x+y) n -x -y n is divisible by xy(x2 + xy+y2
a>

2

b

1l

),

if

n

is

an odd positive integer not a multiple of
4.

3.

Shew that
cy) 3

a3

(bz

+ b 3 (ex - azf + c3 (ay - bx) = Sabc (bz - cy) (ex - az) (ay - bx).
3

6

c

MISCELLANEOUS IDENTITIES.
5.

443

Find the value of
a>c) (c

(6

-c)(c — a) (a-b) + (b6.

- a>a)
be

(a

- cob) + (6 - eo 2 c)
(x2
-f

(c

- arc<)

(a

-

2 a> 6).

Shew that

may
2

be put into
7.

2 2 y + z — yz - zx - xy) the form A 2 + B 2 + C 2 -BC-CA- AB.

(a2

+ b 2 + c2 -

— ca - ab)

Shew that
2
,

A + AB + B and
Shew that
8. 9.

(a2 + ab + b 2 ) (x2 + xy + y 2 ) can be put into the form find the values of A and B.

2 («a + 26c) 3 - 3 (a 2 + 26c) (6 2 + 2ca) (c2 + 2ab) = (a3 + 6 3 + c3 - 3a6c) 2
2 (a 2 - fc) 3 - 3 (a2 - be) (b 2 - ca) (c 2 - ab) = (a 3 + 6 3 + c3 - 3a6c) 2

2
)
.

.

+ 6 2 + c 2 3 +2(6c + ca + a6) 3 -3(a 2 + 6 2 + c2 )(6c + ca + a6) 2 = (a?+b3 + c3 -3abc) 2 If a + 6 + c = 0, prove the identities in questions 11 — 17.
10.

.

11.
12.

2(a 4 + 6 4 + c 4 ) = (a2 + 6 2 + c2 ) 2

.

a5 + 6 5 + c5 =

-

13. 14.

a6 + 6 6 + c6 = 3a

W
c

5a6c (6c + ca + ab).

- 2 (6c + ca + a6) 3

.

3(a 2 + 6 2 + c2 )(a 5 + 6 6 + c5 ) = 5(a 3 + 6 3 + c3 )(« 4 + 6 4 + c4 ).

,_
15.

a7 + b 7 + c7
m

5 5 5 = a +6 +c !

a0

+ 62 + c2
-R

/6-c
16.

c--a
b

a-b\

( a

b
c—a

a
2 (6 c

J\b — c

c +-—

\

a

-J

=9-

17.

+ c2 a + a2 6 - 3a6c) (6c 2 + ca2 + ab 2 - 3abc) = (be + ca + ab) 3 + 27a2 62 c2

.

18.

25 {Q, - zf + (z - x) 7 + (x - y) 7 } {{y - zf + (z- xf + (x- yf)

= 21
19.

{(y

- zf + (z - xf + (x- yf] 2

.

{(y-z) 2 + (z-xf + (x-y) 2 } 3 -54:(i/-z) 2 (z-x) 2
2

(x-yf
.

= 2(y + z-2x) (z+x- 2y) 2 (x +y- 2z) 2
20.
(6

- cf + (e - a) 6 + (a - 6) 6 - 3
7 7 7

(6

- c) 2 (c - a) 2 (a - bf

= 2 (a2 + b2 + c2 -be- ca- ab) 3
21.

.

(6-c)

+ (c-a) + (a-6) = 7(6-c)(c-a)(a-6)(a2 + 6 2 + c2 -6c-ca-a6) 2
shew that
2
)

.

x+y + z = 0, 4 (ax + by + czf -3(ax + by + cz) (a2 + 2 +
22.

If

a + 6 + c = 0, and

(x2 +y 2 + z 2 )

-2(b-c)(c-a)(a-b)(y-z)(z- x) (x-y) = 54abcxyz.

+
444
If

HIGHER ALGEBRA.
a+
b

+ c + d=0, shew
g

that

a5 + &5 +c5 + c#>
23.

g? + b 3 + (? + d3 -•

3—

a2 + b 2 + c2 + d2
2

24.

(a3

+

3
Z>

+ c3 + d3 2 = 9 (M
)

cda + da& + abc) 2 = 9 (be- ad) (ca - bd) (ab ,

cd).

25.

If 2s

=a+b+c

= a2 + 6 2 + c2 prove that 2 2 2 (4s2 + 5 (s - b) (s - c) (a - a + 5a6cs = {s and
2o2 2
)

cr )

2 a- ).

26.

Shew that

(a?

+ 6.2% + 3.vy 2 - y3 3 + (v 3 + 6xy 2 + 3x2y ~ x 3 = Zlxy (x+y) (x 2 + xy +y 2
3
) )

3
)
.

27.

Shew

a
that
2

5

2 (a — b)(a — c) (a - d)

= a + b + c + d2 + ab + ac + ad+bc + bd+cd.
2 2

28.

Resolve into factors

2« 2 6 2 c2 + (a3 + b 3 + c3 ) abc + Z> 3 c3 + c-%3 + a 3 Z> 3

.

Elimination.

In Chapter xxxiii. we have seen that the eliminant of a system of linear equations may at once be written down in the form of a determinant. General methods of elimination applicable to equations of any degree will be found discussed in treatises on the Theory of Equations in particular we may refer the student