Treatiseonalgebr 00 Smituoft
Treatiseonalgebr 00 Smituoft
Treatiseonalgebr 00 Smituoft
to the
LIBRARY
of
the
UNIVERSITY
OF TORONTO
from
the estate
of
7"?J!TJ ?.. DA7IZ3
BY THE SAME AUTHOR.
ELEMENTARY ALGEBRA. Sixth Issue. Revised
and
Enlarged.
Globe Svo.
4-r.
6d.
KEY. i or. 6d.
For the Second Edition the whole book was
thoroughly
revised,
and
the
early chapters
remodelled and
simplified.
The number of
examples
was also
very greatly
increased.
A TREATISE ON ALGEBRA. Seventh Issue. Crown
8vo.
-JS.
6d.
KEY. i or. 6d.
This book has been made more
complete by
the addition of a
Chapter
on the
Theory
of
Equations.
AN ELEMENTARY TREATISE ON SOLID
GEOMETRY. Sixth Issue. Crown 8vo.
y.
6it.
GEOMETRICAL CONICS. Crown Svo. 6s.
AN ELEMENTARY
TREATISE ON CONIC
SECTIONS. Thirteenth Issue. Crown Svo.
js.
6d.
KEY. 1 0$-. 6d.
By
C.
SMITH, M.A.,
and S.
BRYANT,
D.Sc.
EUCLID'S ELEMENTS OF
GEOMETRY. Globe
8vo.
BOOK I. is.
BOOKS I. and II. is. 6et.
BOOKS III. and IV. is. BOOKS I. IV.
35.
MACMILLAN AND CO.
LTD.,
LONDON.
A TREATISE
ON
ALGEBRA
A TREATISE
ON
ALGEBEA
OHAELES
SMITH,
M.A.
MASTER OF SIDNEY SUSSEX
COLLEGE,
CAMBRIDGE.
Hontion
MACMILLAN
AND
CO.,
LIMITED
NEW YORK: THE MACMILLAN COMPANY
1900
First
Edition,
1888.
Second
Edition,
1890.
Third
Edition,
with
additions,
1892.
Fourth
Edition,
with
additions,
1893.
Fifth Edition,
1896. Reprinted 1898,
1900.
PREFACE TO THE FIRST EDITION.
THE
following
work is
designed
for the use of the
higher
classes of Schools and the
junior
students in the
Universities.
Although
the book is
complete
in
itself,
in
the sense that it
begins
at the
beginning,
it is
expected
that students who use it will have
previously
read some
more
elementary
work on
Algebra
: the
simpler parts
of
the
subject
are therefore treated somewhat
briefly.
I have ventured to make one
important change
from
the usual order
adopted
in
English
text-books on
Algebra,
namely by considering
some of the tests of the conver-
gency
of infinite series before
making any
use of such
series : this
change
will,
I feel
sure,
be
generally approved.
The order in which the different
chapters
of the book
may
be read
is, however,
to a
great
extent
optional.
A
knowledge
of the
elementary properties
of Deter-
minants is of
great
and
increasing practical utility
;
and
I have therefore introduced a short discussion of their
fundamental
properties,
founded on the Treatises of Dostor
and Muir.
No
pains
have been
spared
to ensure
variety
and inte-
rest in the
examples.
With this end in
view,
hundreds
of examination
papers
have been consulted
;
including,
with
VI PREFACE.
very
few
exceptions, every paper
which has been set in
Cambridge
for
many years past. Amongst
the
examples
will also be found
many interesting
theorems which have
been taken from the different Mathematical Journals.
I am indebted to
many
friends for their kindness in
looking
over the
proof-sheets,
for
help
in the verification
of the
examples,
and for valuable
suggestions. My especial
thanks are due to the
following
members of
Sidney
Sussex
College
: Mr S. R.
Wilson, M.A.,
Mr J.
Edwards, M.A.,
Mr S. L.
Loney,
M.A.,
and Mr J.
Owen,
B.A.
CHARLES SMITH.
CAMLBIDOE,
December
I2llt,
1887.
PREFACE TO THE THIRD EDITION.
A
Chapter
on
Theory
of
Equations
has been
added,
which it is
hoped
will increase the value ol the book.
CONTENTS.
Definitions
CHAPTER I.
PACE
1
CHAPTER II.
FUNDAMENTAL LAWS.
Negative quantities
Addition of terms
Subtraction of terms
Multiplication
of monomial
expressions
....
Law of
Signs
The factors of a
product may
be taken in
any
order
Fundamental Index Law
Division of monomial
expressions
Multinomial
expressions
Commutative
Law,
Distributive Law and Associative Law
9
10
11
14
15
16
18
19
22
23
CHAPTER III.
ADDITION. SUBTRACTION. BKACKETS.
Addition of
any
multinomial
expressions
Subtraction of multinomial
expressions
Brackets
Examples
I
26
27
28
29
Vlll
CONTENTS.
CHAPTER IV.
MULTIPLICATION. PAGE
Product of two multinomial
expressions
... , . 32
Detached coefficients
36
Square
of a
multinomial
expression
3'j
Continued
products 40
Examples
IL
. . 42
CHAPTER
V.
DIVISION.
Division
by
a multinomial
expression
...... .46
Extended definition of division
50
Examples
III.
51
CHAPTER VI.
FACTORS.
Monomial factors .
^
53
Factors found
by comparing
with known identities ... 54
Factors of
quadratic expressions
found
by inspection
55
Examples
IY 56
Factors of
general quadratic expression
57
Factors found
by
rearrangement
and
grouping
of terms 60
Examples
V - .64
Divisibility
of x
n
a"
by
x a 65
Remainder Theorem 66
An
expression
of the nth
degree
cannot vanish for more than n
values of
a:,
unless it vanishes for all values of a; . . . <',;i
Cyclical
order 71
Symmetrical expressions
71
Factors found
by
use of Remainder Theorem 72
Examples
VI 73
CHAPTER VII.
HIGHEST COMMON FACTOR. LOWEST COMMON MULTIPLE.
Monomial common factors
76
Multinomial common factors
77
Examples
VII
84
Lowest
common
multiple
...
84
Examples
VIIL .
86
CONTENTS. IX
CHAPTER VIII.
FRACTIONS.
PAGE
A fraction is not altered
by multiplying
its numerator and denomi-
nator
by
the same
quantity
....... 88
Reduction of fractions to a common denominator .... 90
Addition and subtraction of fractions
91
Multiplication
and division of fractions
93
Important
theorems
concerning
fractions formed from
given
fractions 96
Examples
IX. .
99
CHAPTER IX.
EQUATIONS.
ONE UNKNOWN
QUANTITY.
General
principles applicable
to all
equations
106
Simple equations
107
Special
forms of
simple equations
108
The
problem
of
solving
an
equation
the same as the
problem
of
finding
the factors of an
expression
110
Quadratic equations
110
Discussion of roots of a
quadratic equation
113
Zero and infinite roots
114
Equations
not
integral
116
Irrational
equations
119
A
quadratic equation
can
only
have two roots 121
Relations between the roots and the coefficients of a
quadratic
equation
1-3
Relations between the roots and the coefficients of
any equation
.
1
Equations
with
given
roots ........
125
Discussion of
possible
values of a trinomial
expression
. . .
127
Examples
X
130
Equations
of
higher degree
than the second
135
Equations
of the same form as
quadratic equations
. . .
135
Reciprocal equations
137
Roots found
by inspection
138
Binomial
equations
139
Cube roots of
unity
. . . . - . . . . . 140
Examples
XI
141
X
CONTENTS.
CHAPTER X.
SIMULTANEOUS
EQUATIONS.
PAGE
Equations
of the first
degree
with two unknown
quantities
. .
1^5
Discussion of solution
1*9
Equations
of the first
degree
with three unknown
quantities
. .
150
Method of undetermined
multipliers
......
151
Equations
with more than three unknown
quantities
. . .
154
Examples
XII. . .
155
Simultaneous
equations
of the second
degree
. . . .
157
Examples
XIII . .
163
Equations
with more than two unknown
quantities
. . .
165
Examples
XIV. .170
CHAPTER XT.
PROBLEMS.
Problems not
always
satisfied
by
the solutions of the
corresponding
equations
. . .
173
Examples
XV ,
, 177
CHAPTER XII.
MISCELLANEOUS THEOREMS AND EXAMPLES.
Examples
of elimination 1S2
Equations
with restrictions on the values of the letters . . . 186
Identities deduced from the factors of
"
+
6
s
+
c
3
-
Salic . . . 187
Examples
XVI 190
CHAPTER XIII.
POWERS AND ROOTS. FRACTIONAL AND NEGATIVE INDICES.
Index Laws
200
Boots
of arithmetical numbers 202
Sards
obey
the Fundamental Laws of
Algebra
.... 203
Fractional
and
negative
indices 205
Rationalizing
factors 210
Examples
XVII 212
CONTENTS.
XI
CHAPTER XIV.
SURDS. IMAGINARY AND COMPLEX
QUANTITIES.
PAGE
Properties
of Surds
214-216
If
a+,Jb=c
+
^/d,
where a and c are rational and
^/b
and
*Jd
are
irrational,
then a=c and b
=
d
217
If either of two
conjugate quadratic
surds is a factor of a rational
expression,
so also is the other ......
217
Square
root of a
+
^/b
218
Examples
XVIII 220
Imaginary
and
complex quantities
221
Complex quantities obey
the Fundamental Laws of
Algebra
. - 223
Definition and
properties
of the modulus of a
complex quantity
224-226
If either of two
conjugate complex quantities
is a factor of a real
expression,
so also is the other . .
226
CHAPTER XV.
SQUARE
AND CUBE ROOTS.
Square
roots found
by inspection
229
Square
root of
any algebraical expression
230
Square
root found
by equating
coefficients
231
Extended definition of
Square
root
232
When
any
number of terms of a
square
root have been
found,
as
many
more terms can be found
by ordinary
division . . 233
When n
figures
of a
square
root have been found
by
the
ordinary
method,
n- 2 more
figures
can
always
be found
by
division . 234
Cube root 235
Method of
finding
the nth root of
any algebraical expression
. . 238
Examples
XIX
238
CHAPTER XVI.
RATIO AND PROPORTION.
Eatio.
Compound
ratio.
Duplicate
ratio
241
A ratio is made more
nearly equal
to
unity by adding
the same
positive quantity
to each of its terms .....
242
Incommensurable numbers 243
Proportion
244
Continued
proportion.
Mean
proportional
245
Geometrical
and
algebraical
definitions
compared
.... 246
Xll CONTENTS.
PAGE
Variation 248
Indeterminate Forms 251
Examples
XX 253
CHAPTER XVII.
THE PROGRESSIONS.
Arithmetical
progression
255
Geometrical
progression
260
Harmonical
progression
265
Examples
XXI. . . 2G7
CHAPTER XVIII.
SYSTEMS OP NUMERATION.
Expression
of
any integer
in
any
scale of notation .
. . 271
Radix fractions . 273
The difference between
any
number
(expressed
in the scale
r)
and
the sum of its
digits
is divisible
by
r
-
1 275
Rule for
casting
out the nines . . 276
Examples
XXII. 277
CHAPTER XIX.
PERMUTATIONS AND COMBINATIONS.
Permutations of different
things
280
Permutations all
together
of
things
which are not all different . 281
Combinations 283
<7r=C.-r
-
'
^ 285
Greatest value of
n
C
r
286
(3H-V)^n
=
x^n
+
x^n-l -V^l
+ +
v^n
'
' 287
Vandennonde's Theorem
[see
also
p.
310j
288
Homogeneous products [see
also
p. 352]
289
Examples
XXIH
294
CHAPTER XX.
THK BINOMIAL THEOREM.
Proof of the binomial theorem for a
positive integral exponent
. 297
Proof
by
induction 298
Greatest term 301
Examples
XXIV. . . . .
303
Properties
of the coefficients of a binomial
expansion
. . . 305
CONTENTS. Xlll
PAGB
Continued
product
of n binomial factors of the form x
+
a,
x
+ b,
&c.
308
Vandermonde's Theorem
[see
also
p. 288]
310
The multinomial theorem
311
Examples
XXV
314
CHAPTER XXI.
CONVERGENCE AND DIVERGENCY OF SERIES.
Convergency
and
divergency
of
series,
all of whose terms have the
same
sign
318-32ft
Series whose terms are
alternately positive
and
negative
. . . 326
Application
to the
Binomial,
the
Exponential
and the
Logarithmic
Series 327-320
Product of
convergent
series
329
If So
r
x
r
=S6
r
o;
r
1
for all values of x for which the series are con-
vergent,
then a
r
=b
r
331
Examples
XXVI . 332
CHAPTER XXII.
THE BINOMIAL THEOREM. ANY INDEX.
Proof of the theorem 335
Euler's
proof
336
Greatest term
340
Examples
XXVII 341
Sum of the first r
+
1 coefficients of a
+
a^x
+
a
2
a;
2
+ 343
Binomial Series 344
Expansion
of multinomials ........
348
Combinations and Permutations with
repetitions
....
350
Homogeneous products [see
also
p. 289]
352
Examples
XXVHI
355
CHAPTER XXIII.
PARTIAL FRACTIONS. INDETERMINATE COEFFICIENTS.
Decomposition
into
partial
fractions ...... 362
Case of
imaginary
factors . 365
Case of
equal
factors 366
Indeterminate coefficients 368
Examples
XXIX. .
369
XIV
CONTENTS.
CHAPTER XXIV.
EXPONENTIAL THEOREM. LOGARITHMS. LOGARITHMIC SERIES.
PAG
The
Exponential
Theorem .37
Examples
XXX
37
Properties
of
logarithms
38
The
logarithmic
series
.38
Cauchy's
theorem
38
Series for
calculating logarithms
38
Examples
XXXI 38
Common
logarithms
38
Compound
interest and annuities 31
Examples
XXXIL 3J
CHAPTER XXV.
SUMMATION OF SERIES.
Sum of series found
by expressing
w
n
in the form v
n
-
H_,
. . 395
Sum of series whose
general
term is
2.b . . 397
Sum of series \vhose
general
term is
)...(a+n+r-2.b)}
. 400
Sum of
squares
and sum of cubes of the first n numbers
.
403
Sum of l
r
+ 2r+... +
nr . . - .....
404
Piles of shot .
'
. .
-
;
.
"
:
^
.
Figurate
Numbers ....'.'...
^6
Polygonal
Numbers . . -
.
. . .....
407
Examples
XXXIIL ........
*W
.
Sum of series whose
general
term is
...(b
+
n-l .
x)
,
where a
p
=A
rpr
+
A^p
r
-
l
+
... + A,
t
Series whose law is not
given
. *
Method of Differences
Recurring
Series
. . .
.,..;..;
417
Convergency
of Infinite
products
.....
m(m-l) in(m-l)(in-'2)
._.
Conditions for
convergency
of 1 m
+
-- '-
v
1 ^ g
-
+
'
2u
m
and Sa
n
u
,,
both
convergent
or both
divergent
.
426
CONTENTS.
XV
PAGE
n convergent
or
divergent according
as the limit of
1
.
428
Examples
XXXIV......
CHAPTER XXVI.
INEQUALITIES.
Elementary Principles
......... 435
Product of
any given
number of
positive quantities,
whose sum is
given,
is
greatest
when the
quantities
are
equal
. . .
43G
The arithmetic mean of
any
number of
positive quantities
is
greater
than their
geometric
mean .......
437
The sum of
any given
number of
positive quantities,
whose
product
is
given,
is least when the
quantities
are all
equal
. . .
-438
If
TO, o,
|8,...
are
positive
and m
=
a
+
/3
+
...,
then
Sa
z
(Sal
*
> < V
,
unless x is a
positive
fraction ....
442
n
(n )
Examples
XXXV...... .....
443
CHAPTER
XXVII.
CONTINUED
FRACTIONS.
Convergents
of a
+- ... are
alternately
less and
greater than
C
T ~T~
the continued fraction . . . . . . . .
447
Law of formation of successive
convergents
.....
447
Eeduction of
any
rational fraction to a continued fraction . . 449
Properties
of
convergents
........ 450
Examples
XXXVI
........... 454
General
convergent
..........
456
Periodic continued fractions ........ 458
Convergency
of continued fractions ....... 460
Eeduction of
quadratic
surds to continued fractions . . . 465
Series
expressed
as continued fractions ......
470
Examples
XXXVII.....
.473
XVI
CONTENTS.
CHAPTER
XXVIII.
THEORY OF NUMBERS
I'AOE
The Sieve of Eratosthenes
479
Properties
of
primes
480-482
Highest power
of a
prime
contained in
[ji
483
The
product
of
any
n consecutive
integers
is divisible
by
jn
. . 483
Format's Theorem
[see
also
p. 493]
485
Number of divisors of a
given
number . . . . . .
^86
Number of
positive integers
less than a
given
number and
prime
to
it
[see
also
p. 495]
487
Forms of
square
numbers 489
Examples
XXXVIII 490
Congruences
492
Wilson's Theorem 494
Extension of Format's Theorem 497
Lagrange's
Theorem .........
498
Reduction of Fractions to
Circulating
Decimals .... 499
Examples
XXXIX 501
CHAPTER XXIX.
INDETERMINATE
EQUATIONS.
Integral
solutions of
axby
=
c can
always
be found if a and b are
prime
to one another
504
General solution of ax
-
by=c, having given
one solution . . 504
General solution of ax +
by
=
c,
having given
one solution . . 505
Number of
positive integral
solutions of ax
+
by
=
c . . . . 506
Integral
solutions of ax
+
by+cz d,
a'x
+
b'y
+
c'z
=
d' . . . 508
Examples
XL
510
CHAPTER XXX.
PROBABILITY.
Definition of
probability
;"12
Exclusive Events 514
Independent
Events 515
Dependent
Events 516
CONTENTS.
XV11
PAGE
Probability
of an event
happening
r times in 71 trials . .
. 518
Expectation
520
Inverse
Probability
,
. 521
Probability
of
Testimony
523
Examples
XLI 526
CHAPTER XXXI.
DETERMINANTS.
Definition and
properties
of determinants
.... 530-542
Multiplication
of determinants 543
Simultaneous
Equations
of the First
Degree
545
Elimination
547
Sylvester's
method of Elimination
548
Examples
XLII. .1
549
CHAPTER XXXII.
THEORY OF
EQUATIONS.
Every equation
of the 7?th
degree
has n roots
553
Eolations between the roots and the coefficients of an
equation
.
554
Sum of the with
powers
o f the roots
556
Symmetrical
functions of roots 557
Transformation of
Equations
559
Beciprocal Equations
561
Examples
XLIH
564
Imaginary
and
quadratic
surd roots occur in
pairs
. . .
565
Eoots common to two
equations
567
Eoots connected
by any given
relation 567
Commensurable Boots
568
Examples
XLIV 569
Derived Functions 571
Equal
Roots 572
Continuity
of a rational
integral
function 574
If
/(a) and/(/3)
are of
contrary signs,
a real root
of/(.t)-0
lies
between a and
/3
575
The
Discriminating
Cubic 576
Eolle's Theorem . 577
XV111 CONTENTS.
PAGE
Descartes' Rules of
Signs
.
578
Examples
XLV
....
580
Cubic
Equations
582
Biquadratic Equations 583
Sturm's Theorem .
535
Synthetic
Division . .
,
590
Horner's Method of
approximating
to the real roots of
any equation 593
Examples
XLVI
596
Miscellaneous
Examples 599
Answers
to the
Examples
. . .
5^3
CHAPTER I.
DEFINITIONS.
1.
ALGEBRA,
like
Arithmetic,
is a science which treats
of numbers.
In Arithmetic numbers are
represented by figures
which have determinate values. In
Algebra
the letters of
the
alphabet
are used to
represent
numbers,
and each
letter can stand for
any
number
whatever,
except
that in
any
connected series of
operations
each letter must
through-
out be
supposed
to
represent
the same number.
Since the letters
employed
in
Algebra represent any
numbers
whatever,
the results arrived at must be
equally
true of all numbers.
2. The numbers treated of
may
be either whole
numbers or fractions.
All concrete
quantities
such as
values,
lengths,
areas,
periods
of
time, &c.,
with which we have to do in
Algebra,
must be measured
by
the number
of
times each contains
some unit of its own. kind. Thus we have
lengths
of
4,
f
,
oi,
the unit
being
an
inch,
a
yard,
a
mile,
or
any
other
fixed
length.
It is
only
these numbers with which we are
concerned,
and our
symbols
of
quantity,
whether
figures
or
letters,
always represent
numbers. On this account the
word
quantity
is often used instead of number.
3. The
sign
+
,
which is read
'
plus,'
is
placed
before a
number to indicate that it is to be added to what has
gone
s. A. 1
2 DEFINITIONS.
before. Thus 6 + 3 means that 3 is to be added to 6
;
6
-f
3 -f
2 means that 3 is to be added to 6 and then 2
added to the result. So also a
+
b means that the number
which is
represented by
b is to be added to the number
which is
represented by
a
; or,
expressed
more
briefly,
it
means that b is to be added to a
;
again
a
+
b + c means
that b -LS to be added to a and then c added to the
result
4. The
sign
,
which is read
'minus/
is
placed
before
a number to indicate that it is to be subtracted from what
has
gone
before. Thus a b means that b is to be subtracted
from
a;
a b c means that b is to be subtracted from
a,
and then c subtracted from the
result;
and a b
+ c means
that b is to be subtracted from
a,
and then c added to the
result.
Thus in additions and subtractions the order of the
operations
is from left to
right
5. The
sign
x,
which is read
'into/
is
placed
between
two numbers to indicate that the first number is to be
multiplied by
the second. Thus a x b means that a is to
be
multiplied by
b;
also a x 6 x c means that a is to be
multiplied by
b,
and the result
multiplied by
c.
The
sign
x is however
generally
omitted between two
letters,
or between a
figure
and a
letter,
and the letters
are
placed consecutively.
Thus ab means the same as
a x
b,
and 5ab the same as 5 x a x b.
The
sign
of
multiplication
cannot be omitted between
figures
: 63 for
example
does not stand for 6x3 but for
sixty-three,
as in Arithmetic.
Sometimes the x is
replaced by
a
point,
which is
placed
on the
line,
to
distinguish
it from the decimal
point
which is
placed
above the line. Thus a x b x
c,
a.b.c and abc all mean the
same,
namely
that a is to
be
multiplied by
b and the result
multiplied by
c.
6. The
sign
-T-,
which is read 'divided
by'
or
'by/
is
placed
between two numbers to indicate that the first
DEFINITIONS. 6
number,
called the
dividend,
is to be divided
by
the
second
number,
called the divisor. Thus a-r-b means
that a is to be divided
by
b
;
also a
-=-
b -j- c means that
a is to be divided
by
b,
and the result divided
by
c;
and a -f- 6 x c means that a is to be divided
by
b and the
result
multiplied by
c.
Thus in
multiplications
and divisions the order of the
operations
is from left to
right.
7. When two or more numbers are
multiplied together
the result is called the continued
product,
or
simply
the
product;
and each number is called a
factor
of the
product.
When the factors are considered as divided into two
sets,
each is called the
co-efficient,
that is the
co-factor
of the other. Thus in
3abx,
3 is the coefficient of
abx,
3a is the coefficient of
bx,
and Sab is the coefficient of x.
When one of the factors of a
product
is a number
expressed
in
figures,
it is called the numerical
coefficient
of the
product
of the other factors.
8. When a
product
consists of the same factor
repeated any
number of times it is called a
power
of that
factor. Thus aa is called the second
power
of
a,
aaa is
called the third
power
of
a,
aaaa is called the fourth
power
of
a,
and so on. Sometimes a is called the first
power
of a.
Special
names are also
given
to aa and to aaa
;
they
are called
respectively
the
square
and the cube of a,
9. Instead of
writing
aa, aaa, &c.,
a more convenient
notation is
adopted
as follows : a
2
is used instead of
aa,
a? is used instead of
aaa,
and a* is used instead of
aaaa
,
the factor a
being
taken n
times;
the small
figure placed
above and to the
right
of a
shewing
the
number of times the factor a is to be taken. So also
a
3
6
2
is written instead of
aaabb,
and
similarly
in other
cases.
The small
figure,
or
letter,
placed
above a
symbol
to
12
4 DEFINITIONS.
indicate the number of times that
symbol
is to be taken
as a factor is called the index or the
exponent.
Thus a"
means that the factor a is to be taken n
times,
or that
the nth
power
of a is to be
taken,
and n is called the
index.
When the factor a is
only
to be taken
once,
we do
not write it a
1
,
but
simply
a.
10. A number which when
squared
is
equal
to
any
number a is called a
square
root of
a,
and is
represented
by
the
symbol f/a,
or more often
by
*Ja:
thus 2 is
*J4s,
since 2
s
=
4.
A number which when cubed is
equal
to
any
number
a is called a cube root of
a,
and is
represented by
the
symbol f/a
: thus 3 is
^27,
since 3
s
=
27.
In
general,
a number which when raised to the nth
power,
where n is
any
whole
number,
is
equal
to
a,
is
called an nth root of
a,
and is
represented by
the
symbol ^/a.
The
sign
V
was
originally
the initial letter of the
word radix. It is often called the radical
sign.
11. A root which cannot be obtained
exactly
is
called a
surd,
or an irrational
quantity:
thus
*/7
and
^4
are surds.
The
approximate
value of a
surd,
for
example
of
V7,
can be
found,
to
any degree
of
accuracy
which
may
be
desired,
by
the
ordinary
arithmetical
process
;
but we
are not
required
to find these
approximate
values in
Algebra:
for us
*/7
is
simply
that
quantity
which when
squared
will become 7.
12. A collection of
algebraical symbols,
that is of
letters,
figures,
and
signs,
is called an
algebraical expression.
The
parts
of an
algebraical expression
which are con-
nected
by
the
signs
+
or are called the terms.
Thus 2a 3bx
+
ocy*
is an
algebraical expression
con-
taining
the three terms
2a, 3bx,
and
DEFINITIONS.
5
13. When two terms
only
differ in their numerical
coefficients
they
are called like terms. Thus a and 3a are
like terms
;
also 5a
3
6
s
c and 3a
8
6
8
c are like terms.
14. An
expression
which contains
only
one term is
called a monomial
expression,
and
expressions
which
contain two or more terms are called multinomial
expres-
sions;
expressions
which contain two
terms,
and those
which contain three terms
are, however,
generally
called
binomial and trinomial
expressions respectively.
Thus
3a6
s
c is a
monomial,
a
2
+
36
2
is a
binomial,
and ax*
+
bx
+
c
is a trinomial
expression.
15. The
sign
=,
which is read
'equals,'
or 'is
equal
to,'
is
placed
between two
algebraical expressions
to denote
that
they
are
equal
to one another.
The
sign
> indicates that the number which
precedes
the
sign
is
greater
than that which follows it. Thus a > b
means that a is
greater
than b.
The
sign
< indicates that the number which
precedes
the
sign
is less than that which follows it. Thus a < b
means that a is less than 6.
The
signs
=, :J-
and are used
respectively
for is not
equal
to,
is not
greater
iiian,
and is not less than.
The
sign
v is written for the word because or since.
The
sign
.'.is written for the word
therefore
or hence.
16. To denote that an
algebraical expression
is to be
treated as a
whole,
it is
put
between brackets. Thus
(a
+
b)
c means that b is to be added to a and that the
result is to be
multiplied by
c
; again (a b)(c
+
d)
means
that b is to be subtracted from
a,
and that d is to be added
to
c,
and that then the first result is to be
multiplied by
the second
;
so also
(a
+
b)
3
(c
+
df
means that the cube of
the sum of a and b is to be
multiplied by
the
square
of the
sum of c and d.
Brackets are of various
shapes
:
thus,
( ), { }, [ ].
Instead of a
pair
of brackets a
line,
called a
vinculum,
is
often drawn over the
expression
which is to be treated as
C DEFINITIONS.
a
whole : thus a
b c is
equivalent
to a
(b c),
and
Ja
-I- 6 is
equivalent
to
*J(a
+
b).
It should be noticed
that where no vinculum or bracket is
used,
a radical
sign
refers
only
to the number or letter which
immediately
follows it: thus
V2a
means that the
square
root of 2 is to
be
multiplied by
a,
whereas
J2a
means the
square
root of
2a
;
also
\/
4
x means that a; is to be added to the
square
root of
a,
whereas
Ja
+x means that x is to be added to a
and that the
square
root of the whole is to be taken.
The line between the numerator and denominator of
a fraction acts as a
vinculum,
for is the same as
Note. It is
important
for the student to notice that
every
term of an
algebraical expression
must be added or
subtracted as a
whole,
as if it were enclosed in brackets.
Thus,
in the
expression
a
+ be d
-=-
e
+f,
b must be
multiplied by
c before
addition,
and d must be divided
by
e before
subtraction,
just
as if the
expression
were written
EXAMPLES.
1. Find the numerical values of the
following expressions
in each of
which
a=l,
b
=
2,
c
=
3,
and d=4.
(i)
5a+3c-3b-2d,
(ii)
26
(iii)
afe
+
36c-5d,
(iv)
bc-ca-ab,
(v)
a
+ bc
+
d&nd
(vi)
bcd + cda +
dab + abc.
Ant.
0, 12, 0, 1, 11,
50.
2. If a
=
3,
6=1 and c
=
2,
find the numerical values of
(i)
2a-36
s
-4r,
(ii) 2a*6-3&V,
(iii)
^c-^,
(iv)
a
s
+3ac*-8ac-c,
and
(v)
2aWc
-
Wc'a
-
2ea
a
fc.
AM.
19,
C, 0, 1,
0.
DEFINITIONS.
3. Find the values of the
following expressions
in each of which a
=
6
=
2,
c=-.\ an.lrf
=
0.
(i) (3a
+
4d)(26-3e),
(ii)
2
2
-(&
2
-3c
2
)d,
(iii)
a
3
-6
3
-
2
(a-
&
+
<:),
(iv)
a
(6
2
-
c
2
)
+
6
(c
2
-
d
z
)
-f
d(a
s
-
c
5
),
(v)
3
(
+
6)
2
(c
+
d)
-
2
(b
+
c)
2
(a
+
d),
2a
2
2b
2
2c
2 3
ana
(vi)
----
,
--
,
H
v/
----
,
--
,
---
.
b
+
c c
+
a b
+
d a
+
b
Am.
9, 13,
3, 11,21,
3.
4. Find the values of
and
4/
when a
=
5,
6
=
4,
c
=
3.
vins.
3, 13, 60,
5.
5. Shew that a
3
-6'
3
and
(a
+
b) (a -6)
are
equal
to one another
(i)
when a
=
2,
6
=
1;
(ii)
when a
=
5,
6
=
3;
and
(iii)
when a
=
12,
b =~j.
6. Shew that the
expressions
a
3
-b
3
,
(a-b)(a*
+
ab+b-), (a-b)
3
+
3ab(a-b),
and a
+
6
are all
equal
to one another
(i)
when a
=
3,
6
=
2;
(ii)
when a
=5,
6
=
1;
and
(ii)
when a
=
6,
6
=
3.
CHAPTER II.
FUNDAMENTAL
LAWS.
17. WE have said that all concrete
quantities
must
be measured
by
the number of times each contains some
unit of its own kind. Now a sum of
money may
be either
a
receipt
or a
payment,
it
may
be either a
gain
or a loss
;
motion
along
a
given straight
line
may
be in either of two
opposite
directions
;
time
may
be either
before
or
after
some
particular epoch
;
and so in
very many
other cases. Thus
many
concrete
magnitudes
are
capable
of
existing
in two
diametrically opposite
states : the
question
then arises
whether these
magnitudes
can be
conveniently
distin-
guished
from one another
by special signs.
18. Now whatever kind of
quantity
we are consider-
ing
+
4 will stand for what increases that
quantity by
4
units,
and 4 will stand for whatever decreases the
quantity by
4 units.
If we are
calculating
the amount of a man's
property
(estimated
in
pounds),
+
4 will stand for whatever increases
his
property by
4,
that is
+
4 stands for 4 that he
possesses,
or that is
owing
to him
;
so also 4 will stand
for whatever decreases his
property by
4,
that
is,
4 will
stand for 4 that he owes.
If,
on the other
hand,
we are
calculating
the amount
of a
man's
debts,
+ 4 will stand for whatever increases his
FUNDAMENTAL
LAWS. 9
debts,
that
is,
+
4 will -now stand for a debt of 4
;
so also
4 will now stand for whatever decreases his
debts,
that
is,
4 will stand for 4 that he
has,
or that is
owing
to
him.
If we are
considering
the amount of a man's
gains,
-f
4
will stand for what increases his total
gain,
that
is,
+
4
will stand for a
gain
of 4
;
so also 4 will stand for what
decreases his total
gain,
that
is,
4 will stand for a loss of
4. If however we are
calculating
the amount of a man's
losses,
+
4 will stand for a loss of
4,
and 4 will stand for
a
gain
of 4.
Again,
if the
magnitude
to be increased or diminished
is the distance from
any particular place,
measured in
any particular
direction, +
4 will stand for a distance of
4 units in that
direction,
and 4 will stand for a distance
of 4 units in the
opposite
direction.
19. From the last article it will be seen that it is not
necessary
to invent
any
new
signs
to
distinguish
between
quantities
of
directly opposite
kinds,
for this can be done
by
means of the old
signs
+
and .
The
signs
+
and are therefore used in
Algebra
with
two
entirely
different
meanings.
In addition to their
original meaning
as
signs
of the
operations
of addition and
subtraction
respectively, they
are also used as marks
of
distinction between
magnitudes
of
diametrically opposite
kinds.
The
signs
+
and are sometimes called
signs of
affection
when
they
are thus used to indicate a
quality
of
the
quantities
before whose
symbols they
are
placed.
The
sign
+,
as a
sign
of
affection,
is
frequently
omitted
;
and when neither the
+
nor the
sign
is
prefixed
to a
term the
+
sign
is to be understood.
20. A
quantity
to which the
sign
+
is
prefixed
is
called a
positive quantity,
and a
quantity
to which the
sign
is
prefixed
is called a
negative quantity.
The
signs
+
and are called
respectively
the
positive
and
negative signs.
10 FUNDAMENTAL LAWS.
Note.
Although
there are
many signs
used in
algebra,
the name
sign
is often used to denote the two
signs
+
and
exclusively.
Thus,
when the
sign
of a
quantity
is
spoken
of,
it means the
.+
or
sign
which is
prefixed
to
it;
and when we are directed to
change
the
signs
of an
expression,
it means that we are to
change
the
+
or
-
before
every
term into or
+
respectively.
21. The
magnitude
of a
quantity
considered inde-
pendently
of its
quality,
or of its
sign,
is called its absolute
magnitude.
Thus a rise of 4 feet and a fall of 4 feet are
equal
in absolute
magnitude
;
so also
+
4 and 4 are
equal
in absolute
magnitude,
whatever the unit
may
be.
Addition.
22. The
process
of
finding
the result when two or more
quantities
are taken
together
is called
addition,
and the
result is called the sum.
Since a
positive quantity produces
an
increase,
and a
negative quantity produces
a
decrease,
to add a
positive
quantity
we must add its absolute
value,
and to add a
negative quantity
we must subtract its absolute value.
Thus,
when we add
+
4 to
+
6,
we
get
+ 6
+
4
;
and when
we add 4 to
+ 10,
we
get
+
10 4.
Hence + 6
+
(+ 4)
=
+
6 +
4,
and + 10
+
(- 4)
=
+
10
-
4.
So
also,
when we add + b to
+
a,
we
get
+
a
+
6
;
and
when we add 6 to
+
a,
we
get
+
a
-
b. Hence
+
a
+
(+
6)
=
+
a
+
b,
and +
a +
( b)
=
+ a b.
We therefore have the
following
rule for the addition
of
any
term : to add
any
term
affix
it to the
expression
to
which it is to be
added,
with its
sign unchanged.
When numerical values are
given
to a and to
b,
the
numerical values of a
+
6 and a b can be found
;
but
ADDITION OF TERMS. 11
until it is known what numbers a and 6 stand
for,
no
further
step
can be
taken,
and the
process
is considered
to be
algebraically complete.
23. When b is
greater
than
a,
the arithmetical
operation
denoted
by
a b is
impossible.
For
example,
if
a
=
3 and 6
=
5,
a b will be 3
5,
and we cannot take
5 from 3. But to subtract 5 is the same as to subtract 3
and 2 in
succession,
so that
3-5
=
3-3-2
=
0-2
=
-
2.
We then consider that 2 is 2 which is to be sub-
tracted from some other
algebraical expression,
or that
2 is two units of the kind
opposite
to that
represented
by
2
;
and if 2 is a final
result,
the latter is the
only
view that can be taken.
In some
particular
cases the
quantities
under con-
sideration
may
be such that a
negative
result is without
meaning
;
for
instance,
if we have to find the
population
of a town from certain
given
conditions
;
in this case the
occurrence of a
negative
result would shew that the
given
conditions could not be
satisfied,
and so also in this case
would the occurrence of a fractional result.
Subtraction.
24. Since subtraction is the inverse
operation
to that
of
addition,
to subtract a
positive quantity produces
a
decrease,
and to subtract a
negative quantity produces
an
increase. Hence to subtract a
positive quantity
we must
subtract its absolute
value,
and to subtract a
negative
quantity
we must add its absolute value.
Thus,
to
subtract
+
4 from
4-10,
we must decrease the amount
by
4
;
we then
get
+10
4.
Also
to subtract 4 from
+
6,
we must increase the
amount
by
4
;
we then
get
+6 +
4.
Hence
+10-(+4)
=
+
10-4
=
+
6,
and +
6-(-4)
=
+
6
+
4
=
+
10.
12 FUNDAMENTAL LAWS.
So
also,
in all cases
a
(+ 6)
=
a
b,
and a
(b)
=
a
+
b.
We therefore have the
following
rule for the subtraction
of
any
term : to subtract
any
term
affix
it to the
expression
from
which it is to be subtracted but with its
sign changed.
25. We have hitherto
supposed
that the letters used
to
represent quantities
were restricted to
positive
values
;
it would however be
very
inconvenient to retain this
restriction. In what follows therefore it must
always
be
understood,
unless the
contrary
is
expressly
stated,
that
each letter
may
have
any positive
or
negative
value.
Since
any
letter
may
stand for either a
positive
or for
a
negative quantity,
a term
preceded by
the
sign
+
is not
necessarily
a
positive quantity
in
reality
;
such terms are
however still called
positive
terms,
because
they
are so in
appearance
;
and the terms
preceded by
the
sign
are
similarly
called
negative
terms.
26. On the
supposition
that b was a
positive quantity,
it was
proved
in Articles 22 and
24,
that
a
+
b .........
(i)
\
a-b.........
(ii)
I
a-b.........
(iii)
and a
( b)
=
a
+
b.....
...(iv)j
We have now to
prove
that the above laws
being
true
for all
positive
values of b must be true also for
negative
values.
Let b be
negative
and
equal
to
c,
where c is
any
positive quantity ;
then
+
b
=
+
( c)
=
c from
(ii),
and &
=
( c)
=
+
c from
(iv).
Hence,
putting
c for
+
b,
and
+
c for b in
(i), (ii),
SUBTRACTION OF TERMS. 13
(iii), (iv),
it follows that these relations are true for all
negative
values of
b,
provided
a +
( c)
=
a
c,
a +
(+ c)
=
a
+
(,
a
(c)=a
+
c,
and a
(+ c)
=
a
c,
are true for all
positive
values of c
;
and this we know to
be the case.
Kence the laws
expressed
in
(A)
are true for ail values
of 6.
27. Def. The
difference
between
any
two
quantities
a and b is the result obtained
by subtracting
the second
from the
first.
The
algebraical
difference
may
therefore not be the
same as the arithmetical
difference,
which is the result
obtained
by subtracting
the less from the
greater.
The
symbol
a
~
b is sometimes used to denote the arithmetical
difference
of a and 6.
Def. One
quantity
a is said to be
greater
than another
quantity
b when the
algebraical difference
a b is
positive.
From the definition it is
easy
to see that in the series
1, 2, 3, 4, &c.,
each number is
greater
than the one before
it
;
and
that,
in the series
1, 2, 3, 4, &c.,
each
number is less than the one before it.
Thus
7, 5, 0, 5,
7 are in
descending
order of
magnitude.
EXAMPLES.
Ex. 1. Find the sum of
(i)
5 and
-
4,
(ii)
-
5 and
4,
(iii)
5,
-
3 and
-6 and
(iv)
-3, 4,
-6 and 5. Ans.
1, -1, -4,
0.
Ex. 2. Subtract
(i)
3 from
-
4,
(ii)
-
4 from
3,
and
(iii)
-
a from
-b.
Ans.
-7,7,
-b
+
a.
Ex. 3. A barometer fell -01 inches one
day,
it rose -015 inches on
the next
day,
and fell
again
-01 inches on the third
day.
How
much
higher
was it at tie end than at the
beginning?
Ans. "005 inches.
Ex. 4. A thermometer which stood at 10
degrees centigrade,
fell
20
degrees
when it was
put
into a
freezing
mixture: what was
the final
reading?
Aiis.
-
10.
14
FUNDAMENTAL LAWS.
Ex. 5. Find the value of a-6
+
c and of -a +
6-c,
when
a=l,
6
=
-2andc
=
3. .4ns.
6, -6,
Ex. 6. Find the value of
-
a
+
b
-
c when
a
=
l,
b=
-2,
c=
-1;
also when
a=-2, 6=-l,
e
=
-3. Ans.
-2,4.
Ex. 7. Find the value of a
-
(
-
6)
+
(
-
c)
when
a=
-3,
6=
-2,
c=-l.
Ans. -4.
Ex. 8. Find the value of
-
a+
(
-
b)
-
(
-
c)
when
a=
-2,
6=
-3,
c= -5. Ant. 0.
Ex. 9. Find the value of
-
(
-
a)
+
b
-
(
-
c)
when
a=-l,
b=
-2,
c=-3. Ans. -6.
Multiplication.
28. In
Arithmetic,
multiplication
is first defined to be
the
taking
one number as
many
times as there are units in
another.
Thus,
to
multiply
5
by
4 is to take as
many
fives as there are units in four. As
soon, however,
as
fractional numbers are
considered,
it is found
necessary
to
modify
somewhat the
meaning
of
multiplication,
for
by
the
original
definition we can
only multiply by
whole numbers.
The
following
is therefore taken as the definition of
multiplication
:
"
To
multiply
one number
by
a second is to
do to the
first
what is done to
unity
to obtain the second."
Thus 4 is 1
+
1
+
1
+
1;
.-. 5 x 4 is 5
+
5
+
5
+
5.
Again,
to
multiply fy by ^,
we must do to what is
done to
unity
to obtain
f
;
that
is,
we must divide
^
into
four
equal parts
and take three of those
parts.
Each of
K
the
parts
into which &
is to be divided will be ._
7 ,
and
7x4
by taking
three of these
parts
we
get
7.
Thus = x '.
_
5x :\
"7x4'
MULTIPLICATION OF MONOMIAL EXPRESSIONS. 15
So
also,
(
-
5)
x 4
=
(
-
5)
+
(
-
5)
+
(- 5)
+
(
-
5)
=5555
=
-20.
With the above
definition,
multiplication by
a
negative
quantity presents
no
difficulty.
For
example,
to
multiply
4
by
5. Since to subtract
5
by
one subtraction is the same as to subtract 5 units
successively,
-5
=
-1-1-1-1-1;
.-. 4 x
(
-
5)
=
-4-4-4-4-4
=
-20.
Again,
to
multiply
5
by
4. Since
-4
=
-l-l-l-l
5
+
5
+
5
[Art. 26]
=-+20.
We can
proceed
in a similar manner for
any
other
numbers,
whether
integral
or
fractional,
positive
or
nega-
tive.
Hence we have the
following
rule :
To
find
the
product of any
two
quantities, multiply
their
absolute
values,
and
prefix
the
sign
+
if
both
factors
be
positive
or both
negative,
and the
sign if
one
factor
be
positive
and the other
negative.
Thus we have
.........
(i)
(ii)
(
+
o)x(-6)
=
-ofc .........
(iii)
(
-a)
x
(-&)=
+
at.........
(iv)
The rule
by
which the
sign
of the
product
is determined
is called the Law of
Signs.
This law is sometimes
enunciated
briefly
as follows : Like
signs give
+,
and unlike
signs give
.
16 FUNDAMENTAL LAWS.
29. The factors of a
product may
be taken in
any
order. It is
proved
in Arithmetic that when one
number,
whether
integral
or
fractional,
is
multiplied by
a
second,
the result is the same as when the second is multi-
plied by
the first.
The
proof
is as follows: when the numbers are
integers,
a and b
suppose,
write down a series of rows of
dots,
putting
a dots in each
row;
and take b
rows,
writing
the
dots under one another as in the
following
scheme :
******
******
******
******
a in a row
b rows.
Then the whole number of the dots is a
repeated
b
times,
that is a x b. Now consider the columns instead
of the rows: there are
clearly
b dots in each
column,
and
there are a columns
;
thus the whole number of dots is b
repeated
a
times,
that is b x a.
Hence,
when a and b are
integers,
ab
=
ba.
When the numbers are
fractions,
for
example
and
f
,
5 3 5x3
we
prove
as in Art. 28 that = x
=
^
-. .
And, by
the
7 4 7x4
Cf
.
,
5x33x5,
5335
above
proof
for
integers,
- -
=
7
=
;
hence = x
7
=
7
x =.
i X 4 T*X< i 4 4 i
Hence we have ab
=
ba,
for all
positive
values of a and
6;
and the
proposition being
true for
any positive
values of
a and
6,
it must be true for all
values,
whether
positive
or
negative;
for from the
preceding
Article the absolute value
of the
product
is
independent
of the
signs,
and the
sign
of the
product
is
independent
of the order of the factors.
Hence for all values of a and b we have
ab
=
I>a
(i).
If in the above scheme we
put
c
iu
place
of each of the
MULTIPLICATION OF MONOMIAL EXPRESSIONS. 17
dots;
the whole number of the c's will be
ab;
also the
number of c's
in the first row will be
a,
and this is
repeated
b times.
Hence,
when a and b are
integers,
c
repeated
ab
times
gives
the same result as c
repeated
a times and this
repeated
b times. So that to
multiply by any
two whole
numbers in succession
gives
the same result as to
multiply
at once
by
their
product;
and the
proposition
can,
as
before,
be then
proved
to be true without restriction to
whole numbers or to
positive
values.
Thus,
for all values
of
a,
b and
c,
we have
a xbx c= a x
(be) (ii).
By
continued
application
of
(i)
and
(ii)
it is
easy
to
shew that the factors of a
product may
be taken in
any
order, howe.ver
many
factors there
may
be. Thus
abc cab
=
cba,
&c
(C).
30. Since the factors of a
product may
be taken in
any
order,
we are able to
simplify many products.
For
example:
3ax4a
=
3x4xaxa
=
12a
8
,
(- 3a)
x
(- 46)
=
+
3a x 46
=
3 x 4 x a x b
=
IZdb,
(a&)
2
=
abx.ab
=
axaxbx.b
=
a*b*,
(V2a)
2
=
V2a
x
*J2a
=
\/2
x
V2
x aa
=
2a
8
.
Although
the order of the factors in a
product
is
indifferent,
a factor
expressed
in
figures
is
always put
first,
and the letters are
usually arranged
in
alphabetical
order.
31. Since a
2
=
aa,
and a
3
=
aaa;
we have
a
8
x a
8
=
aa x aaa
=
a
5
=
a
2+s
.
So also
a
3
x a*
=
aaa x aaaa
=
a
1
=
a
3
"
1
"
4
,
and a
4
x a
=
aaaa x a
=
a
5
=
a
4*1
.
In the above
examples
we see that the index
of
the
product of
two
powers of
the same letter is
equal
to the sum
of
the indices
of
the
factors.
We can
prove
in the
following
s. A. 2
18 FUNDAMENTAL LAWS.
manner that this is true whenever the indices are
positive
integers
:
since
by
definition
a
1"
=
aaaa ... to m
factors,
and a*
=
aaaa ... to n factors
;
.'. a
1"
x a*
=
(aaa
... torn
factors)
x
(aaa
... to n
factors)
=
aaa ... to
(m
+
ri)
factors,
=
a
m+M
,
by
definition;
hence a x a"
=
a"
rt1*
.......................................
(D).
The law
expressed
in
(D)
is called the Index
Law.
32. Since
(- a)
x
(- a)
=
+
a
8
=
(+ a) (+ a) [Art. 28],
it follows
conversely
that the
square
root of a* is either
+ a or a: this is written
\fa
2
=
a,
the double
sign being
read
'
plus
or minus.'
Thus there are two
square
roots
of any algebraical
quantity,
which are
equal
in absolute
magnitude
but
opposite
in
sign.
EXAMPLES.
1.
Multiply
2o
by
-
4&,
o
8
by
-
a
3
and
-
2a
s
6
by
-
Sab
3
.
Ans.
-Safe,
-a
5
,
6aW.
2.
Multiply
-
2xy* by
-
3j/*z, 3ax*y by
-
5a
2
X(/
2
,
and Sa?bc?x
by
12afe
2
cx*.
Ans.
6xy*z,
3-
Multiply
7a6Vby
-3a
s
iV,
and -2ai
s
zy by
-
Ans. -21a
7
W>,
4. Find the values of
(
-
a)
1
, (
-
a)
8
,
(
-
a)
4
and
(
-
a)
5
.
Ans. a
2
,
-a
s
,
a
4
,
-a*.
5. Find the values of
(
-
a6)
a
, (a-b)*
and
(
-
3aV*c
a
)*.
Ant.
6. Shew that the successive
powers
of a
negative quantity
are
alternately positive
and
negative.
7. Find the cubes of
2u*b,
-
3ab-c
a
,
and
-
2o?bx
s
y'
2
.
Am. 8a6
3
,
-
27aic
9
and
-
8a
6
b
3
x
9
y
8
.
DIVISION OF MONOMIAL EXPRESSIONS. 19
8.
Find the value of
(
-
a)
2
x
(
-
fc)
3
,
of
(
-
2a&
2
)
3
x
{
-
3a-b)
3
,
and
of
(-36c)
3
x(2a
3
&)
3
.
Ans. -a
2
*)
3
,
216a
9
6
9
,
T2a
s
b
s
c~.
9.
Find the value of Zabc
-
2a
2
6c
3
+
4c
4
,
when a =
2,
6=-l,
and
c=-2.
^Iras. 12.
10.
Find the value of 2o
J
&c
-
3b-cd
+ 4c*da
-
5d?ab,
when
a=-l,
6=
-2,
c= -3 and d= -4.
.4n*. -148.
Division.
S3. Division is the inverse
operation
to that of multi-
plication;
so that to divide a
by
6 is to find a
quantity
c
such that c
x b
=
a.
Since division is the inverse of
multiplication
and
multiplications
can be
performed
in
any
order
[Art. 29],
it
follows that successive divisions can be
performed
in
any
order. Thus a-4-6-i-c
=
a4-c4-6.
It also follows from Art. 29 that to divide
by
two
quantities
in succession
gives
the same result as to divide
at once
by
their
product.
Thus a + b + c
=
a +
(be),
which
is
usually
written a -4- be.
Not
only may
a succession of divisions be
performed
in
any
order,
but divisions and
multiplications together
may
be
performed
in
any
order. For
example
For a
=
a
-=-
c x
c;
.*. ax6
=
a-=-cxcx6
=
a-=-cx6xc;
[by
Art.
29]
therefore,
dividing
each
by
c,
we have
ax6-=-c
=
a^-cx6.
Hence we
get
the same result whether we divide the
product
of a and 6
by
c,
or divide a
by
c and then
multiply
by
6,
or divide b
by
c and then
multiply by
a.
34. The
operation
of division is often indicated
by
placing
the dividend over the divisor with a line between
2 2
20 FUNDAMENTAL LAWS.
them: thus
r
means an- 6. Sometimes
a/6
is written for
o
r
. When a -r- 6 is written in the fractional form
r ,
a is
o
b
called the
numerator,
and 6 the denominator.
Since
-
=
1 -r-
c,
c
-xc=l-r-cxc
=
l.
c
Also ax-xc
=
ax(- xc]=axl=a.
1
A \
ax-xc
=
ax(-xc=axl='
c
Vc /
Therefore,
dividing by
c,
1
a x
-
=
a -7-
c,
c
so that to divide
by any quantity
c is the same as to
multiply by
the
quantity
-.
c
Hence axb-^c
=
a-rcxb,
can be
written,
i
*
T
i
axox-=ox-xo,
c c
in which form it is seen to be included in Art. 29
(C).
35. Since a
3
x a
8
=
a
8
,
and a
7
x a
3
=
a
10
;
we have con-
versely
a
8
-T- a*
=
a
2
,
and a
10
-e- a
3
=
a
7
.
And,
in
general,
when m and n are
any positive integers
and m >
n,
we have
a"
1
-r a"
=
a'"~*,
for
by
Art. 31
a"
1""
x a"
=
a
m
.
Hence if one
power
of
any quantity
be divided
by
a lower
power
of the same
quantity,
the index of the
quotient
is
equal
to the
difference
of the indices of the
dividend and the divisor.
DIVISION OF MONOMIAL EXPRESSIONS.
21
and aW+aW
=
a
5
&
3
.
36. We have
proved
in Art. 28 that
ax
( b)
=
ab
;
/.
( ab)
H-
( 6)
=
a,
and
( ab)
-=-
a
=
b
;
we have also
proved
that
/.
(+ ab)
-T-
(- a)
=
-
6,
and
(+ ab)
-s-
(+
a)
=
4-
b.
Hence if the
signs
of the dividend and divisor are
alike,
the
sign
of the
quotient
is
+
;
and if the
signs
of
the dividend and divisor are
unlike,
the
sign
of the
quo-
tient is
;
we therefore have the same Law of
Signs
in
division as in
multiplication.
Thus -a
3
6
6
-o6
2
=
-a^
4
,
and
-
2a
5
6c
7
-r-
-
3a
4
for
=
ac
5
.
EXAMPLES.
1. Divide 10a
by
-
2a,
Sa
2
**
3
by
-
2afe
8
,
and
-
7a
5
6
3
c
4
by
-
3a
2
6
2
c
2
.
3 7
Ana.
-5, --a,-a
s
bc*.
a O
2. Divide
-
2a
5
6
7
c
8
by
4a
3
6c
7
,
-
6z
5
i/< by
3x
3
y,
and
-
5a
2
6
4
x
7
y
8
by
-2a&
4
zV.
^n.
-Hfl
2
&
8
c,
-2V, lazy.
2 J
3.
Multiply
-
2a
3
6c
5
by
-
3ai*
7
c
2
and divide the result
by
8a
3
6
6
c
6
.
3
Ans. -rab-c.
4
37. The fundamental laws of
Algebra,
so far as
monomial
expressions
are
concerned,
are those which were
22 FUNDAMENTAL LAWS.
marked
A, B, C,
D in the
preceding
articles,
and which are
collected below :
+
(+
a)
=
+
a
+
(-)
=
-a
( a)
=
+
a
(B),
abc
=
cba
=
cab
=
&c
(C),
It should be remarked that the laws
expressed
in
(A),
(B), (C)
have been
proved
to be true for all values of a and
6;
but both m and n are
supposed
in
(D)
to be
positive
integers.
Multinomial
Expressions.
38. We now
proceed
to the consideration of multi-
nomial
expressions.
We first observe that
any
multinomial
expression
can
be
put
in the form
a
+
b + c +
&C.,
where
a, b, c,
&c.
may
be
any quantities, positive
or
nega-
tive.
For
example,
the
expression 3x*y xy* *7xyz,
which
by (A)
is the same as
%x*y
+
( ^xy*)
+
( 7xyz),
takes
the
required
form if we
put
a for
3#*y,
6 for
fxy*
t
and c
for
(
7xyz).
It therefore follows that in order to
prove any
theorem
to be true for
any algebraical expression,
it is
only necessary
MULTINOMIAL EXPRESSIONS. 23
to
prove
it for the
expression
a
+
b
+
c
+
&c.,
where
a,b,c,
&c.
are
supposed
to have
any
values,
positive
or
negative.
39. It follows at once from the
meaning
of addition
that the sum of two or more
algebraical quantities
is the
same in whatever order
they
are added. For
example,
to
find how much a man is
worth,
we can take the different
items of
property, considering
debts as
negative,
in
any
order.
Thus a + b
+
c
=
c
+
a
+
b
=
b
+
c
+
a
=
&c
(E).
The laws
[C]
and
[E]
are
together
called the Com-
mutative
Law,
which
may
be enunciated in the
following
form : Additions or
Multiplications may
be made in
any
order.
40. Since additions
may
be made in
any
order,
we have
+...)
+
a
(from E)
...
-f
a
=
a + b+c +
d+ ...
(from E).
Hence,
to add
any algebraical expression
as a whole is
the same as to add its terms in succession.
Since the
expression
+
a
b+
c d
may
be written in
the form
+
a
+
( 6)
+
c
+
( d),
we have
+
{+
a
-
b
+
c
-
d\
=
+
{+
a
+
(-
6)
+ c
+
(- d)}
=
+
a
+
(- 6)
+
c
+
(- d).
When we
say
that we can add the terms of an
expres-
sion in
succession,
it must be borne in mind that the
terms include the
prefaced signs.
41. Since subtraction and addition are inverse
opera-
tions,
it follows from the
preceding
that to subtract an
expression
as a whole is the same as to subtract the terms
in succession. Thus
a
(b
+
c
+ d+
...)
=
a b c d ...
24 FUNDAMENTAL LAWS.
42. If c be
any positive integer,
a ?nd b
having any
values
whatever,
then
(a
+
6)
c
=
(a
+
6)
+
(a
+
6)
+
(a
-f-
b}
+
...
repeated
c times
=
a
+
6
+
a
+ +
a
+ +...
[Art. 40]
=
a
+
a
+
a
+
. . .
repeated
c times
+
6
+
b
+
b
+
. . .
repeated
c times
=
ac
+
be.
Hence,
when c is a
positive integer,
we have
(a
+
b)c
=
ac +
bc..................
(F).
Since division is the inverse of
multiplication,
it follows
that when d is
any positive integer
And hence
(a-f 6)
x c -T- d
=
{(a
+
b)
x
c}
-r- d
=
(ac
+
6c)
-T- d
=
ac + d
+ be -r-
d,
C C C
that is
j ^ j.
add
Thus the law
expressed
in
(F)
is true for all
positive
values of
c;
and
being
true for
any positive
value of
c,
it
must also be true for
any negative
value.
For,
if
(a
+
6)
c
=
ac
+
be,
then
(a
+
6) ( c)
=
-
(a
+
6)
c
=
ac
-
be
=
a(-c)+b(-c).
Hence for all values of
a,
6 and c we have
(a
+
6)c
=
ac
+
6c..................
(F).
Thus the
product
of the sum of
any
two
algebraical
quantities by
a third is
thq
sum of the
products
obtained
by multiplying
the
quantities separately by
the third.
The above is
generally
called the Distributive Law.
MULTINOMIAL EXPRESSIONS. 25
43. Since
(a
+
b)
-r- c
=
(a
+
6)
x
-
1 1
=
a x
-
+
b x
-
=
a-7-c
+
6-^c.
c c
we see that the
quotient
obtained
by dividing
the sum of
any
two
algebraical quantities by
a third is the sum of the
quotients
obtained
by dividing
the
quantities
separately by
the third.
44. From Art. 40 it follows that
c +
(d
+
e)+
...
+
d)
+ e+...=
&c.,
so that the terms of an
expression may
be
grouped
in
any
manner.
Again,
from Art.
29,
it follows that
abcde ... =a
(be) (de)
... =a
(bed)
e ...
=
&c.,
so that the factors of a
product may
be
grouped
in
any
manner.
These two results are called the Associative Law.
45. We have now considered all the fundamental laws
of
Algebra,
and in the
succeeding chapters
we have
only
to
dev
elope
the
consequences
of these laws.
CHAPTER III.
ADDITION. SUBTRACTION:. BRACKETS.
Addition.
46. WE have
already
seen that
any
term is added
by
writing
it
down,
with its
sign unchanged,
after the
expres-
sion to which it is to be added
;
and we have also seen
that to add
any expression
as a whole
gives
the same
result as to add all its terms in succession. We therefore
have the
following
rule : to add two or more
algebraical
expressions,
write down all the terms in succession with
their
signs unchanged.
Thus the sum of a 26
+
3c and 4>d oe +
Gf
is
a
-
26
+
3c
-
4>d
-
5e
+
6/.
47. If some of the terms which are to be added are
'like'
terms,
the result
can,
and
must,
be
simplified
before
the
process
is considered to be
complete.
Now two 'like' terms which have the same
sign
are
added
by taking
the arithmetical sum of their numerical
coefficients with the common
sign,
and
affixing
the com-
mon letters.
For
example,
to add 2<i and 5a in succession
gives
the same
result,
whatever o
may be,
as to add
la;
that
is,
+2a +
5a= +7a. Also,
to
subtract 2a and 5a in succession
gives
the same result as to subtract
la;
that
id,
-
'2a
-
fxt=
-
la.
ADDITION. SUBTRACTION.
27
Also two 'like' terms whose
signs
are different are
added
by taking
the arithmetical difference of their
numerical coefficients with the
sign
of the
greater,
and
affixing
the common letters.
For
example,
+5a-3a=+2a +
3a-3o=+2a,
also +3o-5a=+3a-3a-2o=
-2a.
Thus,
when there are several 'like' terms some of
which are
positive
and some
negative, they
can all be
reduced to one term.
Ex.1. Add
2a+56toa-6fc.
The sum is o-66+ 2o+5&
=Sa-b.
Ex.2. Add 3a
2
-5aft
+ 76
2
,
-
4a
2
-
2a&
+
36
!
,
and 2a
2
+
5a&-86
8
.
The sum is 3o*
-
Sab
+ 7V*
-
4a
a
-
2ab
+
3&
2
+
2a
a
+
Sab
-
8b.
The terms 3a
2
,
4a
s
,
and
+
2a
2
can be combined
mentally;
and we
have a
2
.
Similarly
we have
-
2ab and +
26
2
.
Thus the
required
sum is a
8
-
2a6
+
26
2
.
The
beginner
will find it desirable to
put
like terms
under one another.
Subtraction.
48. We have
already
seen that
any
term
may
be sub-
tracted
by writing
it
down,
with its
sign changed,
after the
expression
from which it is to be subtracted
;
and we have
also seen that to subtract
any expression
as a whole
gives
the same result as to subtract its terms in succession. We
therefore have the
following
rule : To subtract
any alge-
braical
expression,
write down its terms in succession with
all the
signs changed.
Thus,
if a 26
4- 3c be subtracted from 2a 36
4c,
the result will be 2a
-
36
-
4c a
+
26 3c
=
a b 7c.
49. The
expression
which is to be subtracted is some-
times
placed
under that from which it is to be
taken,
'like'
terms
being
for convenience
placed
under one another
;
28 BRACKETS.
and the
signs
of the lower line are
changed mentally
before
combining
the 'like' terms.
Thus the
previous example
would be written down as under:
2a-36-4c
a
-
26
+
3c
a- fe-7c
As another
example,
if we have to subtract 3oi-5oc
+
c
s
from
o
2
-
Sab +
2ac
-
2fc'
2
,
the
process
is written
3ab-5ac
Brackets.
50. To indicate that an
expression
is to be added as a
whole,
it is
put
in a bracket with the
+
sign prefixed
But,
as we have seen in Art.
46,
to add
any algebraical
expression
we have
only
to write down the terms in suc-
cession with their
signs unchanged.
Hence,
when a bracket is
preceded by
a
+
sign,
the
bracket
may
be omitted.
Thus +
(2a
-
5&
+
7c)
=
+
2a
-
56
+
7c.
Hence
also,
any
number of terms of an
expression may
be enclosed in brackets with the
sign
+
placed
before each
bracket. Thus
3a
-
26
+
4c
-
d
+ e
-/=
3a
-
26
+
(4c
-
d
+
e
-/)
=
3a
+
(-26
+
4c)
-
d
+
(e -/).
When the
sign
of the
first
term in a bracket is
+
it
is
generally
omitted for
shortness,
as in the
preceding
example.
51. To indicate that an
expression
is to be subtracted
as a
whole,
it is
put
in a bracket with the
sign prefixed.
But,
as we have seen in Art.
48,
to subtract
any alge-
braical
expression
we have
only
to write down the terms
in succession with all their
signs changed.
BRACKETS.
29
Hence,
when a bracket is
preceded by
a
sign,
the
bracket
may
be
omitted,
provided
that the
signs
of all the
terms within the bracket are
changed.
Thus
a
-
(26
-
c
+
d)
=
a
-
26 + c
-
d.
Hence
also,
any
number of terms of an
expression may
be enclosed
in a bracket with the
sign prefixed, provided
that the
signs
of all the terms which are
placed
in the
bracket are
changed.
Thus
a
-
26 +
3c
-
d
=
a
-
(26
-
3c
+
d)
=
a
-
2b
-
(-
3c
+
d).
52. Sometimes brackets are
put
within brackets: in
this case the different brackets must be of different
shapes
to
prevent
confusion.
Thus a
[2b {3c
(2d e)}]
;
which means that we
are to subtract from 26 the whole
quantity
within the
bracket marked
{ },
and then subtract the result from a
;
and,
to find the
quantity
within the bracket marked
{ },
we must subtract e from
2d,
and then subtract the result
from 3c.
When there are several
pairs
of brackets
they may
be
removed one at a time
by
the rules of Arts. 50 and 51.
Thus a
-[b
+
{c-(d-e')}]
=
a
-
[b
+
{c
-
d
+
e}]
=
a
[b
+ c d
+
e]
=
a b c
+
de.
EXAMPLES
I.
1. Add 3x
-
5y,
5x
-
2y
and
7y
-
4cc.
2. Add So;
-
5y
+
2s, 5x-7y-
5z and
6y
-
z
-
IQx.
3. Add
\a \b
+
\c, \b -^c
+
^a
and
^c-^a
+
^
4. Add a
3
-
a
2
+
a,
a
3
-
a + 1 and a* a
a
-
1.
5. Add X
s
-
5xy 7y*
and
3y
2
+
4xy
-
x*.
30 EXAMPLES.
6. Add m*
-
3mn +
2n',
3rc
2
-
m
a
and 5mn
-
3n* + 1m*.
7. Add 3a
2
-2oc-
2ab,
26
s
+ 36c + 3o6 and c*
-
2oc
-
26c.
8. Add a
2
6
-
5ab* + 76
s
,
2a
3
-
a*6 + 5a6
8
and 36
3
-
2a
:i
.
9. Subtract 3a 46 + 2c from a + 6
-
2c.
10. Subtract
^
+ -6
-
-5
c from c
^
a
- ~
6.
Sf
4
d A
}
11. Subtract 3x* -4x + 2 from 4a;
a
-
5x
-
1.
12. Subtract 5a
4
-
3a
3
6 + 4o*6
2
from 56
4
-
3a6
3
+ 4aV.
13. What is the difference between
-
3jc*
5xy
+
ktf
and
-5a? +
2xy-3y*l
14. What must be added to 26c 3ca 4a6 in order that
the sum
may
be be +
ca 1
15. What must be added to 3a*-26
2
+ 3c
2
in otder that
the sum
may
be be + ca
16.
Simplify
3x
-
{2y
+
(5x
-3x +
y)}.
17.
SimpHfy
x
-
[3y
+
{3z
-
x-2y]
+
2x].
18.
Simplify y
2x
{z
x
y
x +
z}.
19.
Simplify
a
[a-b {a
b + c a-b + c-
d}~\.
20.
Simplify
2x
-
[3x
-
9y
-
{2x -Sy-(x+ 5y)}].
21.
Simplify
a
-
[3a
+ c
-
{4a
-
(36
-
c)
+
36}
-
2a}
22.
Subtract x
(3y
z)
from
y {2x zy}.
23. Subtract 2m
-
(3m
-
2n-m)
from 2n
-
(3n
-
2m
-
n).
24.
Find the value of
{a
-
(b
-
c)}
1
+
{b-(c- a)}'
+
{c
-
(a
-
6)}*
when
a
=
-l,
6
=
-2,
c=-3.
25. Find the value of
{a*
-
(6
-
c)
8
}
-
{6
s
-
(c
-
a)
8
}
-
{c*
-
(a
-
6)
2
}
when
a
=
1,
6
=
2,
c
=
-
3.
CHAPTER
IV.
MULTIPLICATION.
53. Product of monomial
expressions.
The
multiplication
of monomial
expressions
was considered in
Chapter
II.
,
and the results arrived at were :
(i)
The factors of a
product may
be taken in
any
order.
(ii)
The
sign
of the
product
of two
quantities
is
4-
when both the factors are
positive
or both
negative
;
and
the
sign
of the
product
is when one factor is
positive
and the other
negative.
(iii)
The index of the
product
of
any
two
powers
of
the same
quantity
is the sum of the indices of the factors.
From
(i), (ii)
and
(iii)
we can find the continued
product
of
any
number of monomial
expressions.
Thus
(
-
2a
s
6c
3
)
x
(
-
3a
3
6
2
c)
=
+
2a
2
6c
3
x
So^c,
from
(ii),
=
2 x 3 x a- . a? . b . b" . c
3
.
c,
from
(i),
=
6a
5
&
3
c
4
,
from
(iii).
Again, (
-
3a
2
b) (-
Bab
3
) (-
7a
4
6
2
)
=
{
+
3a
2
6 . 5ab
3
} (
-
7a
4
6
2
)
= -3 . 5 . 7 . a
2
. a . a
4
. b . b
3
. V*=-W5a?b
e
.
54. Product of a multinomial
expression
and a
monomial. It was
proved
in Art. 42 that the
product
of the sum of
any
two
algebraical quantities by
a third is
equal
to the sum of the
products
obtained
by multiplying
the two
quantities separately by
the third.
32 MULTIPLICATION.
Thus
(
+
y)
z
=
xz
+
yz
..................
(i).
Since
(i)
is true for all values of
x,
y
and
z,
it will be
true when we
put (a
-f
6)
in
place
of x
;
hence
{(a
+
b)
+
y\
z
-
(a
+
b)
z
+
yz
=
az
+
bz
+
yz.
:.
(a
+
b +
y)
z
=
az
+
bz
+
yz.
And
similarly
(a
+ 6 + c +
d + . .
.)
t
z
=
az
-f
bz
+
cz + dz 4 . .
.,
however
many
terms there
may
be in the
expression
a
+ b + c+ d+
...
Thus the
product of any
multinomial
expression by
a
monomial is the sum
of
the
products
obtained
by multiplying
the
separate
terms
of
the multinomial
expression by
the
monomial.
55. Product of two multinomial
expressions.
We now consider the most
general
case of
multiplication,
namely
the
multiplication
of
auy
two multinomial ex-
pressions.
We have to find
and,
from Art.
38,
this includes all
possible
cases.
Put M for a?
+
y
+
z
+ . . .
; then,
by
the last
article,
we have
,
=
Ma
+
Mb
+
Mc
+
...
=
ax +
ay
+ az+
...
+bx +
by
+ bz+
...
+
cx
+
cy
+ cz+
...
Hence
(a
+
b
+ c+
...) (x
+
y
+
z
+
...)
-
ax
+
ay
+
az
+
. . .
+
bx
+
by
+
bz
+
. . .
+
ex
+
cy
+
cz
-f
. . .
MULTIPLICATION.
33
Thus,
the
product of any
two
algebraical expressions
is
equal
to the sum
of
the
products
obtained
by multiplying
every
term
of
the one
by every
term
of
the other.
For
example
(a
+
6) (c
+
d)
=
ac 4- ad + be 4- bd
;
also
(3a
+
56) (2a
4
36)
=
(3a) (2a)
+
(3a) (36)
+
(56) (2a)
+
(56) (36)
=
6a
+
9a6
+
10a6
+
156
2
=-6a' + 19a6
+
156
2
.
Again,
to find
(a 6) (c d),
we first write this in the
form
{a +(- 6)} {c +( d)\,
and we then have for the
product
ac
+
a
(- d)
+
(- 6)
c
+
(- 6) (- d)
=
ac ad be + bd.
In the rule
given
above for the
multiplication
of two
algebraical expressions
it must be borne in mind that the
terms include the
prefixed signs.
56. The
following
are
important examples
:
L
(a
+
6)*
=
(a
+
6) (a
+
6)
=
aa
+ ab
+
ba
+
66
;
Thus,
the
square of
the sum
of any
two
quantities
is
equal
to the sum
of
their
squares plus
twice their
product.
II.
(a-6)
2
=(a-6)(a-6)=aa
+
a(-6)
+
(-6)a
Thus,
the
square of
the
difference of any
two
quantities
is
equal
to the sum
of
their
squares
minus twice their
product.
III.
(a
+
6) (a
-
6)
=
aa
+a
(- 6)
+
ba
+
b
(- 6)
-
b~
;
Thus,
the
product
of
the sum and
difference of any
two
quantities
is
equal
to the
difference of
their
squares.
s. A. 3
34 MULTIPLICATION.
57. It is usual to exhibit the
process
of
multiplication
in the
following
convenient form :
a
o'-2ol)
a
4
+
2a
3
6
-
a*6*
-
2a'&
-
4a*&
8
+
2a&
8
_
0*6'
+
2o6
3
-
6
4
a
4
-
4a
s
&*
+
4a&
3
-
6*.
The
multiplier
is
placed
under the
multiplicand
and a
line is drawn. The successive terms of the
multiplicand,
namely
a
2
,
+
2a&,
and
6*,
are
multiplied by
a*,
the first
term on the left of the
multiplier,
and the
products
a
4
,
+
'2a*b and a*6*
which are thus obtained are
put
in a
horizontal row. The
terms
of the
multiplicand
are then
multiplied by
2o6,
the second term of the
multiplier,
and the
products
thus obtained are
put
in another hori-
zontal
row,
the terms
being
so
placed
that 'like' terms
are under one another. And
similarly
for all the other
terms of the
multiplier.
The final result is then obtained
by adding
the rows of
partial products ;
and this final
sum can be
readily
written
down,
since the different sets
of
'
like' terms are in vertical columns.
The
following
are
examples
of
multiplication
arranged
as above described :
a
+
b a
+
b cP
+
db
+ b*
a + b a-b a-b
a'
2
+
ab a
2
+
ab a
3
+
a
2
6
+
ab*
-ab-V -a?b-ab*-b*
a 6* a
a
+
b
+c
3x*-xy
+
2y*
a +b +c 3z
3
+gy-2y
a
d'+ab +ac 9z*
-
3xy
+ 6x
2
z/
a
+
ab
+b*+ be
+3x*y
_
+ac
_
+ bc +c*
-
6
j-y
+
2xy
3
-
4y
4
58. If in an
expression
consisting
of several terms
which contain different
powers
of the same
letter,
the
MULTIPLICATION. 35
term which contains the
highest power
of that letter be
put
first on the
left,
the term which contains the next
highest
power
be
put
next,
and so on
;
the
terms,
if
any,
which
do not contain the letter
being put
last
;
then the whole
expression
is said to be
arranged according
to
descending
powers
of that letter. Thus the
expression
is
arranged according
to
descending powers
of a. In like
manner we
say
that the
expression
is
arranged according
to
ascending powers
of 6.
59.
Although
it is not
necessary
to
arrange
the terms
either of the
multiplicand
or of the
multiplier
in
any
particular
order,
it will be found convenient to
arrange
both
expressions according
to
descending
or both
according
to
ascending powers
of the same letter: some trouble in
the
arrangement
of the different sets of
'
like' terms in
vertical columns will thus be avoided.
60. Definitions. A term which is the
product
of n
letters is said to be of n
dimensions,
or of the nth
degree.
Thus Sabc is of three
dimensions,
or of the third
degree
;
and 5a
s
b
2
c,
that is
5aaabbc,
is of six
dimensions,
or of the
sixth
degree.
Thus the
degree
of a term is found
by taking
the sum of the indices of its factors.
The
degree of
an
expression
is the
degree
of that term
of it which is of
highest
dimensions.
In
estimating
the
degree
of a
term,
or of an
expression,
we sometimes take into account
only
a
particular
letter,
or
particular
letters : thus ax*
+
bx
+
c is of the second
degree
in
a,
and is often called a
quadratic expression
in a;
;
also
ax*y
+
bay
+
ex* is of the third
degree
in x and
y,
and is
often called a cubic
expression
in x and
y.
An
expression,
or a
term,
which does not contain x is said to be of no
degree
in
x,
or to be
independent
of x.
When all the terms of an
expression
are of the same
dimensions,
the
expression
is said to be
homogeneous.
Thus a
3
+
3a
2
6 5&
3
is a
homogeneous expression, every
32
36 MULTIPLICATION.
term
being
of the third
degree
;
also ax*
+
bxy
+
cy*
is a
homogeneous expression
of the second
degree
in a; and
y.
61. Product of
homogeneous expressions.
The
product
of
any
two
homogeneous expressions
must be
homogeneous
;
for the different terms of the
product
are
obtained
by multiplying any
term of the
multiplicand by
any
term of the
multiplier,
and the number of dimensions
in the
product
of
any
two monomials is
clearly
the sum
of the number of dimensions in the
separate quantities;
hence if all the terms of the
multiplicand
are of the
same
degree,
as also all the terms of the
multiplier,
it
follows that all the terms of the
product
are of the same
degree
;
and it also follows that the
degree
of the
product
is the sum of the
degrees
of the factors.
The fact that two
expressions
which are to be multi-
plied
are
homogeneous
should in all cases be noticed
;
and
if the
product
obtained is not
homogeneous,
it is clear
that there is an error.
62. It is of
importance
to notice
that,
in the
product
of two
algebraical expressions,
the term which is of
highest
degree
in a
particular
letter is the
product
of the terms
in the factors which are of
highest degree
in that
letter,
and the term of lowest
degree
is the
product
of the terms
which are of lowest
degree
in the factors : thus there is
only
one term of
highest degree
and one term of lowest
degree.
63. Detached Coefficients. When two
expressions
are both
arranged according
to
descending,
or to
ascending,
powers
of some
letter,
much of the labour of
multiplication
can be saved
by writing
down the coefficients
only.
Thus,
to
multiply
3x
s
x
+
2
by
&c*
+
2x
-
2,
we write
3-1
+
2
3
+
2-J
9-3
+
6
6-2+4
-6+2-4
9
+ 3_2+ 6-4
MULTIPLICATION. 37
The
highest power
of x in the
product
is
clearly
as
4
,
and the rest follow in order. Hence the
required product
is
9#
4
+
3a;
8
-
2x*
+
Qx
-
4.
When some of the
powers
are absent their
places
must
be
supplied by
O's.
Thus,
to
multiply
x* 2x*
+
x 3
by
as*
+
#
3
x
3,
we write
1+0-2+1-3
1+1+0-1-3
1+0-2+1-3
1+0-2+1-3
-1-0+2-1+3
_
-3-0+6-3+9
1+1-2-2-5-1+5+0+9
Hence the
product
is
x
s
+
5a? + 9.
This is
generally
called the method of detached
coefficients.
64. We now return to the three
important
cases of
multiplication
considered in Art. 56
;
namely,
&
2
..................
(i),
6
2
...............
(ii),
6
2
...............
(iii).
A
general
result
expressed by
means of
symbols
is
called &
formula.
Since the laws from which the above formulae were
deduced were
proved
to be true for all
algebraical
quantities
whatever,
we
may
substitute for a and for b
any
other
algebraical quantities,
or
algebraical expressions,
and the results will still hold
good.
38
MULTIPLICATION.
We
give
some
examples
of results obtained
by
substi-
tution.
Put b in the
place
of b in
(i);
we then have
{a
+
(- b)\
2
=
a
8
+
2a
(- b)
+
(- by,
that is
(a
-
6)
2
=
a
2
-
2a& + 6
2
.
Thus
(ii)
is seen to be
really
included in
(i).
Put
V2
in the
place
of b in
(iii)
;
we then have
(a
+
^2) (a
-
V2)
=
a
8
-
(V2)
8
=
a
8
-
2.
[We
here, howevei,
assume that all the fundamental
laws are true for surds: this will be considered in a
subsequent chapter.]
Put b
+
c in the
place
of b in
(i);
we then have
/.
(a
+
b
+
c)
s
=
a
8
+
2a6
+
2ac
+
b*
+
26c
+
c
8
......
(iv).
Now
put
c for c in
(iv),
and we have
/.
(a
+
b
-c)
8
=
a*
+
2a6
-
2ac
+
6
2
-
26c+
c
8
.
Put 6
+
c in the
place
of b in
(iii);
we then have
.-.
(a
+
b
+
c) (a
-
b
-
c)
=
a
2
-
6
8
-
26c
-
c
2
.
The
following
are additional
examples
of
products
which can be written down at once.
(a*
+
26
s
)
(a
8
-
26
2
)
=
(a
s
)
2
-
(26
2
)
8
=
a<
-
46*.
(a
s
-
^36*)
=
(a
2
)
8 -
(^ft
8
)
2
=
a*
-
36.
=
(a
8
+
6
s
)
2
-
)
=
{(^
+
x)
+
(x
3
+
z)
8
-
(*
s
+
1)
8
=
x
+
2x
4
4- .T
8
-
(x*
4-
-2.r
2
i-
1)
=
x
8
+
x*
-
z
-
1.
MULTIPLICATION. 39
65.
Square
of a multinomial
expression.
We
have found in the
preceding
Article,
and also
by
direct
multiplication
in Art.
57,
the
square
of the sum of three
algebraical quantities;
and the
square
of the sum of
more than three
quantities
can be obtained
by
the same
methods. The
square
of
any
multinomial
expression
can
however best be found in the
following
manner.
We have to find
Now we know that the
product
of
any
two
algebraical
expressions
is
equal
to the sum of the
partial products
obtained
by multiplying every
term of one
expression by
every
term of the other. If we
multiply
the term a of
the
multiplicand by
the term a of the
multiplier,
we
obtain the term a
2
of the
product:
we
similarly
obtain
the terms 6
2
,
c
2
,
&c. We can
multiply any
term,
say
b,
of the
multiplicand by any
different
term,
say
d,
of the
multiplier;
and we thus obtain the term bd of the
product.
But we also obtain the term bd
by multiplying
the term d of the
multiplicand by
the term b of the
multiplier,
and the term bd can be obtained in no other
way,
so that
every
such term as
bd,
in which the letters
are
different,
occurs twice in the
product.
The
required
product
is therefore the sum of the
squares
of all the
quantities
a, b, c,
&c.
together
with twice the
product
of
every pair.
Thus,
the
square of
the sum
of any
number
of algebrai-
cal
quantities
is
equal
to the sum
of
their
squares together
with twice the
product of every pair.
For
example,
to find
(a
+
6
+
c)
2
.
The
squares
of the
separate
terms are a
2
,
6
2
,
c
3
.
The
products
of the different
pairs
of terms are
ab,
ac and be.
Hence
(a
+
&
+
c)
2
=a
2
+
6
2
+ c
a
+2a& +
2ac + 26c.
Similarly,
(a
+
26
-
3c)
2
=
a
2
+
(26)
2
+
(
-
3c)
2
+
2a
(26)
+
2a(
-
3c)
+
2
(26) (
-
3c)
=
a
2
+
46
2
+ 9c* + 4a6
-
6ac
-
126c.
40 MULTIPLICATION.
And
-
2ad
-
2bc
+
2bd
-
2cd.
After a little
practice
the intermediate
steps
should be omitted
and the final result written down at oiice. To ensure
taking
twice
the
product
of
every pair
it is best to take twice the
product
of each
term and of
every
term which
follows
it.
66. Continued Products. The continued
product
of several
algebraical expressions
is obtained
by finding
the
product
of
any
two of the
expressions,
and then
multiplying
this
product by
a third
expression,
and so on.
For
example,
to find
(x
+
a) (x
+
b) (x
+
c),
we have
x +
a
x
+
c
x*
+
(a
+ b+
c)x'-
+
(ab
+ ac
bc)x
+ nbc
In the above all the terms which contain the same
powers
of x are
collected
together:
it is
frequently necessary
to
arrange expressions
in this
way.
Again,
to find
(a?
+
a
5
)
2
(x
+
a)
2
(x
-
a)
2
.
The factors can be taken in
any
order
;
hence the
required product
=
[(x
-
a) (x
+
a) (o?
+
a
a
)]
J
=
[(x
3
-
a
2
) (x
2
+
a
8
)]
2
=
(x*
-
a
4
)
2
=
x
6
-
2a*x*
+
a
8
.
67. We have
proved
in Art. 55 that the
product
of
any
two multinomial
expressions
is the sum of all the
partial products
obtained
by multiplying any
term of one
expression by any
term of the other.
To find the continued
product
of three
expressions
we
must therefore
multiply
each of the terms in the
product
of the first two
expressions by
each of the terms in the
third;
hence the continued
product
is the sum of all the
partial products
which can be obtained
by multiplying
together any
term of the
first,
any
term of the
second,
and
any
term of the third.
MULTIPLICATION.
41
And
similarly,
the continued
product
of
any
number of
expressions
is the sum of all the
partial products
which
can be obtained
by multiplying together any
term of the
first,
any
term of the
second,
any
term of the
third,
&c.
For
example,
if we take a letter from each of the three
factors of
(a
+
b) (a
+
b} (a
+
b),
and
multiply
the three
together,
we shall obtain a term
of the continued
product;
and if we do this in
every
possible way
we shall obtain all the terms of the continued
product.
Now we can take a
every
time,
and we can do this in
only
one
way;
hence a
8
is a term of the continued
product.
We can take a twice and b once,
and this can be done
in three
ways,
for the b can be taken from either of the
three binomial
factors;
hence we have 3a
2
b.
We can take a once and b
twice,
and we can do this
also in three
ways;
hence we have 3a&
2
.
Finally,
we can take 6
every
time,
and this can be done
in
only
one
way;
hence we have b
3
.
Thus the continued
product
is
a
8
+
3a
2
6
+
3a6
2
+
b*,
that is
(a
+
b)
3
=
a
3
+
3a*6
+
Sab
2
+
b
3
.
The continued
product (x
+
a) (x
+
6) (x
+
c)
can simi-
larly
be written down at once.
For we can take x
every
time : we thus
get
a.
3
.
We can take two ass and either a or b or c : we thus
have of
a,
x*b and #
2
c.
We can take one x and
any
two of
a, b,
c: we thus
have
xab, xac,
and xbc.
Finally,
if we take no
#'s,
we have the term abc.
42
MULTIPLICATION.
Thus,
arranging
the result
according
to
powers
of
x,
we
have
(x
+
a) (x
+
b) (x
+
c)
=
a?
+
a?
(a
+
b
+
c)
+
x
(ab
+
ac
+
be)
+ abc.
68. Powers of a binomial. We have
already
found
the
square
and the cube of a binomial
expression;
and
higher powers
can be obtained in succession
by
actual
multiplication.
The method of detached coefficients should
be used to shorten the work.
The
following
should be remembered:
(a
+
6)
2
=
a
2
+ 2ab
+
b*,
(a
+
b)'
=
a
3
+
3a
2
6
+
3a6
2
+
b
3
,
and
(a
+
6)
4
=
a*
+
4a
s
6
+
6a
2
6
2
+
4a6
3
+
b
4
.
To find
any power, higher
than the
fourth,
of a binomial
expression
a formula called the Binomial Theorem should
be
employed:
this theorem will be considered in a subse-
quent chapter.
EXAMPLES IL
L
Multiply
2x
-
a
by
x
-
2a.
2.
Multiply
3x
-
% by ^x
-
3.
3.
Multiply
x"
+ x + I
by
x 1.
4.
Multiply
of
xy
+
y* by
x +
y.
5.
Multiply
l+x+at? + a?
by
x- 1.
6.
Multiply
x* +
x*y
+
x^y
1
+
xy
3
+
y* by y
x.
7.
Multiply
x*
-
x + 2
by
x* + x
-
2.
8.
Multiply
1 + ax + a*x*
by
1
-
ax +
a*x*.
9.
Multiply
x
4
+ x" + 1
by
x*
-
x" + 1.
EXAMPLES. 43
10.
Multiply
3a:
8
-
xy
+
2y* by 3y*
-
xy
+ 2x*.
11.
Multiply
y?
-
5x? + I
by
2x? + 5x + 1.
12.
Multiply
2x*
-
5x*y
+
y> by /
+
Zxy*
+ U.
13.
Multiply
3a
8
-2a
8
6 + 3a&
2
-36
8
by
2a
a
+5a'b-4:ab'+b
9
.
14.
Multiply
2aV
-
3aVy*
+
5y* by
aV+
4a*y
-
2/.
15.
Multiply
2a
-
3a
8
+ 5a
8
-
7a
s
by
1
-
2a
8
+ 6a
4
.
16.
Multiply
a*
-
ab
-
ac + b*
-
be + c
8
by
a + b + c.
17.
Multiply
a?
+y*
+ z
>
-
yz
zx
xy by
x +
y
+z.
18.
Multiply
4a*+ 96
a
+ c
a
+ 36c + 2ca
-
Gab
by
2a + 36
-
c.
19.
Multiply together
a;
4
+
1,
a^+1 and x*-l.
20.
Multiply together
a;
4
+
16y
4
,
x*+
ty*, x+2y
and a;
-
2y.
21.
Multiply together (a:
-
y)', (x
+
y)*
and
(a:
8
+
y
3
)'.
22.
Multiply together (*
+
I)
8
,
(a;
+
I)
3
and
(a;
-
I)
3
.
23.
Multiply together
x* a; +
1,
a^+a;+ 1 and x*
-
x* + 1 .
24.
Multiply together
a*
-
2ab + 46
s
,
a
2
+2o6 + 46* and
a<_4a'& + 16&*.
25. Find the
squares
of
(i)
a +
26-3c,
(ii)
a'-ab +
b*,
(iii)
bc + ca +
ab,
(iv)
1
-
2a; +
Sa;",
and
(v)
a? + x* + x + 1.
26. Find the cubes of
(i)
a + b +
c,
(2)
2a
-
36
-
2c and
(iii)
1 + x + x
9
.
27.
Simplify
(a;
+
y
+
z)'
-
(-
x +
y
+
z)*
+
(x
-
y
+
zf- (x
+
y
-
z)*.
28. Shew that
29. Shew that
(y
+
z)
3
+
(z
+
xf
+
(x
+
y}*
-
x
a
-y*-z
3
=
(x
+
44 EXAMPLES.
30.
Simplify {x (x
+
a)
-
a
(x a)} {x (x a)
-
a
(x
+
a)}.
31. Shew that
32. Shew that a
3
+ b
3
=
(a
+
b)
3
-
Sab
(a
+
b),
and that
a
4
+ b
4
=
(a
+
b)
4
-ab(a
+
b)'
+ Za'b
3
.
33. Shew that
(x*
+
xy
+
y*)*
-
4xy (x*
+
y
!
)
=
(x*-xy
+
y*)
a
.
34. Shew that
(y
~
*)*
+
(
z +
*)*
+
(x
+
y)*+2(x
+
y)(x
+
z)+2(y
+
z) (y
+
x)
+ 2
(z
+
x) (z
+
y)
=
4
(x
+
y
+
z)*.
35. Shew that
(a
8
+
b*) (c*
+
d*)
=
(oc
+
bd)
a
+(od- be)*.
36. Shew
that,
if x
=
a +
d,
y
=
b +
d,
and z
=
c + d
;
then
will y?
+
y*
+ z*
-
yz
-
zx
xy
-
a' + b
9
+ c*
-
be
-
ca ab.
37. Shew
that,
ifa;
=
6 +
c, y
=
c +
a,
and z
=
a +
b;
then
will x
3
+
y
3
+ z
3
yz
zx
xy
=
a* + b
3
+ c* be ca ab.
38. Shew that
2(a-6)(a-c)
+
2(6-c)(6-a)
+
2(c-a)(c-6)
=
(6 -c)*
+
(c-a)'+ (a -6)".
39. Shew that
(x
3
+
y*
+ z
2
) (a*
+ b' + c
8
)
-
(ax
+
by
+
cz)'
=
(bz cy)'
+
(ex azf
+
(ay bx)*.
40. Shew
that,
if x
=
a
3
be, y
=
b*
ca,
z
=
c* ab
;
then
will ox +
by
+ cz
=
(x
+
y
+
z) (a
+ b +
c),
and be
(x* yz)
=
ca
(y* zx)
=
ab
(if xy).
41. Find the value of
(x-a)
a
+
(x-b)
3
+
(x-c)
8
-3(x-a)(x-b)(x-c)
when 3x
=
a +
b + c.
42. Shew that
(a
8
+ b
9
+ c
8
)
8
=
(&
+ c
)
+
(ab
+
ac)*
+
(ab
-
ac)>
+ a*
=
(be
+ ca +
ab)'
+
(a*
-
be)*
+
(&'
-
ca)
8
+
(c
8
-
ab)\
43. Shew that
(x
9
+
xy
+
y*) (a
8
+
ab +
b')
=
(ax
-
by)'
+
(ax
-
by) (ay
+ bx +
by)
+
(ay
+ bx +
by)~.
EXAMPLES. 45
44 Shew that 1
+ a" + b* + c
3
+ b*c* + cV + a*b*
-(l-bc-ca- ab)
s
+
(a
+ b + c-
abc)*.
45 Shew that
(a*
+ b* + c
8
+
d*)
3
=
(a
3
+ b*-c
3
-
d
3
)
3
+
4(ac
+
bd)*+
4
(ad
-
be)
3
.
46. Shew that
(i) (a
4-
2)*- 4(o
+
1)
2
+ 6a
2
-
4(a- 1)
2
+
(a-2)
2
=
0.
(ii) (a+2)(&
+
2)-4(a
+
l)(6+l)
+ Ga6
-
4
(o
-
1) (6
-
1)
+
(a
-
2) (b
-
2)
-
0.
47. Shew that
(i)
(a+2)
8
-4(a
+
l)
3
+ 6
3
-
4
(a-
1)
3
+
(a
-
2)
3
-
0.
(ii) (a
+
2) (6
+
2)(c
+
2)
-
4
(a
+
1)(6
+
l)(c
+
1)
+ Qabc
48. Shew
that
(a
+ b +
c)
3
+
(b
+ c
-
a) (c
+
a
-
b) (a
+ b
c)
=
la
3
(b
+
c)
+ 46
2
(c
+
a)
+ 4c
2
(a
+
b)
+ 4a6c.
49. Shew
that
50.
Multiply
a + b
3
+ c
3
+ d
3
be
-
ca ah ad bd cd
by
a+b
51.
Shew that
(x
s
+x+l)(x
t
-x+l)(x
4
-x+l)(x-x*+l)...(x*'
t
-x
2
*~ l
+l)
CHAPTER V.
DIVISION.
69. Division
by
a monomial
expression.
We
have
already
considered the division of one monomial
expression by
another. We have also seen
(Art. 43)
that
the
quotient
obtained
by dividing
the sum of two
alge-
braical
quantities by
a third is the sum of the
quotients
obtained
by dividing
the
quantities separately by
the
third;
and we can shew
by
the method of Art. 54 that
when
any
multinomial
expression
is divided
by
a monomial
the
quotient
is the sum of the
quotients
obtained
by
dividing
the
separate
terms of the multinomial
expression
by
that monomial.
Thus
(cfx 3cw?)
-:- ax
=
d?x -ax Sax -f- ax
=
a 3.
And
(12#
8
-
5cw?
-
2a'a?)
+ 3x
=
12x* -r- 3x
-
Sax* -r &c
-r- 3#
=
4#
2
ew; a*.
70. Division
by
a multinomial
expression.
We
have now to consider the most
general
case of
division,
namely
the division of one multinomial
expression by
another.
Since division is the inverse of
multiplication,
what
we have to do is to find the
algebraical expression
which,
when
multiplied by
the
divisor,
will
produce
the' dividend.
Both dividend and divisor are first
arranged according
DIVISION.
47
to
descending powers
of some common
letter,
a
suppose
;
and the
quotient
also is considered to be so
arranged.
Then
(Art. 62)
the first term of the dividend will be the
product
of the first term of the divisor and the first term
of the
quotient
;
and therefore the
first
term
of
the
quotient
will be
found by dividing
the
first
term
of
the
dividend
by
the
first
term
of
the divisor. If we now
multiply
the whole divisor
by
the first term of the
quotient
so
obtained,
and subtract the
product
from the
dividend,
the remainder must be the
product
of the
divisor
by
the sum of all the other terms of the
quotient;
and,
this remainder
being
also
arranged according
to
descending powers
of
a,
the second term
of
the
quotient
will be
found
as
before by dividing
the
first
term
of
the
remainder
by
the
first
term
of
the divisor. If we now
multiply
the whole divisor
by
the second term of the
quotient
and subtract the
product
from the
remainder,
it
is clear that the third and other terms
of
the
quotient
can
be
found
in succession in a similar manner.
For
example,
to divide 8a
8
+
8a
2
6
-
4a&'
2
+
V
s
by
2a
+
b.
The
arrangement
is the same as in Arithmetic.
2a
+
b
)
8a
s
+
8a*b
+
4a&
2
+
b
s
(
4a
2
+
2a&
+
6
s
8a
8
+
4a
2
6
_
4a6
+
4a6
2
+
6
s
2a6
s
The
first
term of the
quotient
is 8a
9
4- 2a
=
4a
s
.
Multiply
the divisor
by
4a
2
and subtract the
product
from
the dividend : we then have the remainder 4id
2
b
+
4<aV
+
b
a
.
The second term of the
quotient
is 4a
2
6 -4- 2a
=
2a6.
Multiply
the divisor
by
2a6,
and subtract the
product
from the remainder: we thus
get
the second remainder
2a6
2
+ b*. The third term of the
quotient
is 2a6
2
+ 2a
=
6*.
Multiply
the divisor
by
6
2
,
and subtract the
product
from
48 DIVISION.
2o6
8
+ 6
s
,
and there is no remainder. Since there is no
remainder after the last
subtraction,
the dividend must be
equal
to the sum of the different
quantities
which have
been subtracted from
it;
but we have subtracted in suc-
cession the divisor
multiplied by
4a*,
by
+
2a&,
and
by
-f
&*;
we have therefore subtracted
altogether
the divisor
multiplied by
4a*
+
2a6
+
6*.
And,
since the divisor mul-
tiplied by
4a
s
-I-
2a6 4- 6
s
is
equal
to the
dividend,
the
required quotient
is 4a
8
+
2ab
+
6*.
The dividend and divisor
may
be
arranged according
to
ascending
instead of
according
to
descending powers
of
the common
letter,
as in the last
example
considered with
reference to the letter b
;
but the dividend and the divisor
must both be
arranged
in the same
way.
71. The
following
are additional
examples:
Ex. L Divide a
4
-
a6
+
2a*&
s
-
o& +
b*
by
o
+
fc
2
.
a
2
-f
&
2
}a*
-
a
s
b
+
2a
2
6
2
-
ab
3
+
b*
(a*
-
ab + b*
'*
-a
3
b -ab
3
+
6*
Ex. 2. Divide a
4
+
a
s
6
2
+
b
4
by
o
s
-
ab
+
&.
+
b*
+
a*b
:l
-ab
3
+ b*
In this
example
the terms of the dividend were
placed apart,
in
order that 'like' terms
might
be
placed
under one another without
altering
the order of the terms in
descending powers
of a. The
subtractions can be
easily performed
without
placing
'like* terms
under one
another;
but the
arrangement
of the terms
according
to
descending (or ascending) powers
of the chosen letter should never
be
departed
from.
Ex.3.
Divide o
+
&
+
-
3o6c
by
a
+
fe + c.
DIVISION. 49
a^b
-
a
2
c
-
Babc+ b
3
+
c
3
a?b
-
ab
z
-
abc
-
a
2
c
+
a&
s
-
2abc +
b
3
+ c
3
-a'-c
-
abc
-
ac
l
+
ab*
-
abc + ac*
+
b
3
+
c
3
-abc
+
ac^-Wc
+
-abc -&
2
c-
Where,
as in the above
example,
more than two letters are
involved,
it is not sufficient to
arrange
the terms
according
to
descending powers
of a
;
but b also is
given
the
precedence
over c.
By using brackets,
the above
process may
be shortened. Thus
72. The method of detached coefficients
may
often be
empl oyed
in Division with
great advantage.
For
example,
to divide
2#
6
-7o:
5
+5a;
4
+
3a;
8
-3#
8
+
4a;-4
by
2a?
-
So?
+
a;
-
2,
we write
2-3
+
1-2)2-7
+ 5-1-3-3 +
4-4(1-2-1
+
2
2-3
+
1-2
_
-4+4+5-3+4-4
-4
+ 6-2
+
4
__
_2+7-7+4-4
-2+3-1+2
4-6+2-4
4-6+2-4
The first term of the
quotient
is a? and the other
powers
follow in order : thus the
quotient
is
s. A.
50 DIVISION.
73. Extended definition of Division. In the
process
of division as described in Art.
70,
it is clear that
the remainder after the first subtraction must be of lower
degree
in a than the
dividend;
and also that
every
re-
mainder must be of lower
degree
than the
preceding
remainder. Hence
by proceeding
far
enough
we must
come to a
stage
where there is no
remainder,
or else
where there is a remainder such that the
highest power
of a in it is less than the
highest power
of a in the
divisor,
and in this latter case the division cannot be
exactly per-
formed.
It is convenient to extend the definition of division to
the
following
: To divide A
by
B is to
find
an
algebraical
expression
G such that B x C is either
equal
to
A,
or
differs
from
A
by
an
expression
which is
of
lower
degree,
in some
particular
letter,
than the divisor B.
For
example,
if we divide a*
+
3a6
+
46*
by
a
+
6,
we
have
a
+
b
)
a
2
-f
Sab + 46
2
(
a
+
26
a*+
ab
2a6
+
46
s
2a6
+
26'
+ 26'
Thus
(a
2
+
3a6
+
46
2
)
4-
(a
+
6)
=
a 4
26,
with remainder
26
s
;
that is a* 4- 3a6
+
46*
=
(a
+
6) (a
+
26)
+
26
s
. We have
also,
by arranging
the dividend and divisor
differently,
I
+
a
)
46
s
+
3a6 +
a
s
(
46
-
a
46
2
+
4a6
-
ab
+
a*
ab a*
Hence a
change
in the order of the dividend and
divisor leads to a result of a different form. This
is,
how-
ever,
what
might
be
expected considering
that in the first
DIVISION. 51
case -we find what the divisor must be
multiplied by
in
order to
agree
with the dividend so far as certain terms
which contain a are
concerned,
and in the second we find
what the divisor must be
multiplied by
in order to
agree
with the dividend so far as certain terms which contain b
are concerned.
When therefore we have to divide one
expression by
another,
both
expressions being arranged
in the same
way,
it must be understood that this
arrangement
is to
be adhered to.
74. Def. A relation of
equality
which is true for all
values of the letters it
contains,
is called an
identity.
The
following
identities can
easily
be
verified,
and
should be remembered
(a?
+
2ox
+
a
2
)
(a?
Zax
+
a
2
) (a; a)
=
cc a.
(a?
a
2
) (x a)
=
x
+
a.
(a?
+
a
8
) (as
+
a)
=
a? cix
+
a*.
(x*
-
a
4
)
-
(x
+
a)
=
a? ax*
+
a?x a
3
.
(x
4
+
aV
+
a*)
-f-
(a?
+
ax
+
a
2
)
=
a? ax
+
a
2
.
H-
(x
+
y
+
z)
=
a?+
y*+
#yz
zx
xy.
EXAMPLES IIL
1. Divide of
9y* by
x +
3y.
2. Divide x*
-
IQy* by
x*
-
4y*.
3. Divide 27a?
+
64/ by ly
+ 3a.
4. Divide 3a?
-
4xy ly* by 2y
x.
5. Divide 1
-
5ce* + 4x
5
by
1
-
x.
6. Divide X
s
-
5xy*
+
4y
5
by
x
-
y.
42
62 EXAMPLES.
7.
CHAPTER VL
FACTORS.
75. Definitions.
An
algebraical expression
which
does not contain
any
letter in the denominator of
any
term is said to be an
integral expression:
thus
|a
s
& b
3
is an
integral expression.
An
expression
is said to be
integral
with
respect
to
any
particular
letter,
when that letter does not occur in the
denominator of
any
term : thus
I
--
,
is
integral
with
a a
+
o
respect
to x.
An
expression
is said to be rational when none of its
terms contain
square
or other roots.
76. In the
present chapter
we shall shew how factors
of
algebraical expressions
can be found in certain
simple
cases.
We shall
only
consider rational and
integral expres-
sions
;
and
by
the
factors
of an
expression
will be meant
the rational and
integral expressions,
or the
expressions
which are rational and
integral
in some
particular
letter,
which
exactly
divide it.
77.
Monomial
Factors. When some letter is
common to all the terms of an
expression,
each
term,
and
therefore the whole
expression,
is divisible
by
that letter.
Thus 2ax
+
x
2
=x(2a+x),
ax
+
a
2
a:
2
=
ax
(1
+
ax),
Such monomial
factors,
if there be
any,
are obvious on
inspection.
54 FACTORS.
78. Factors found
by comparing
with known
identities. Sometimes an
algebraical expression
is of
the same form as some known result of
multiplication:
in this case factors can be written down at once.
Thus,
from the known
identity
a?- 6*=
(a
+
6) (a -6),
vie have
a
2
-46
s
=a
2
-(26)
2
=(a
+
26)(a-2&),
a
-
2=a
2
-
U/2)*= (a
+
x/2) (a
-
^2),
a
-
166
4
=
(a)
-
(4ft
2
)
2
=
(a
3
+
462) (a
2
-
46
2
)
=
(a
2
+
4ft
2
) (a
+
26) (a -26),
and a
-
9ai
2
=
a
(a
2
-
96
2
)
=
a
(a
+
36) (a
-
36).
Again,
from the
identity
we have
o*+86
3
=a
s
+
(26)
s
=
(a
+
26) {a
2
-a(
=(a+26)(a
2
-2a6
+
46
8
),
8a
3
+ 276"
=
(2a)
+
(36
a
)
=
(2a
+
36
2
) { (2o)
a
-
(2a) (36
s
)
+
(36
s
)
1
}
=
(2o
+
36
s
) (4o
2
-
6a6
8
+
96*)
,
and a
9
+
*
=
(a
3
)
3
+
(z*)*
=
(a
3
+
x
3
) (a
6
-
aV+
a*)
=
(a+ x)
(a*
-
ax
+
ar) (a
6
-
oV
+
a:
6
).
And,
from the
identity
o
-
6=
(o
-
6) (a
2
+
ab
+
6
s
),
we have
-a6xy
+
-xV
J
The
following
are additional
examples
of the same
principle
:
(i)
=
(a
+
6 +c
+
d)(a+6-c-d).
(ii)
4a6-
(o
s
+6
-
C
2
J
2
=
{2a6
+
(o
2
+6-c
s
)}
{2o6
-
(rf+P -c*)};
and,
since
and 2a6-a
f
-6
2
+c
s
=c*-(o-6)
a
=(c+a-6)(c-o
we have
finally
FACTORS. 65
(iii) (a
+
25)
3
-(2o+&)
8
=
{(a+2&)-(2a
+
6)}{(a+2&)3
+
(a
+
2&)(2a
+
6)
+
(2a
+
&)
a
}
B
(6
-
a) (7a
+
13a&
+
76
2
).
79. Factors of
af+pas
+
q
found
by inspection.
From the
identity
(x
+
a) (x
+
b)
=
of
+
(a
+
6)
x
4-
ab,
it follows
conversely
that
expressions
of the form
x*
+px
+
q
can
sometimes,
if not
always,
be
expressed
as the
product
of two factors of the form as
+
a,
x
+
b.
We shall
presently give
a method
by
which two factors
of
af+px
+
q
of the form x
+
a and x
+
b can
always
be
found
;
but whenever a and b are
rational,
the factors can
be more
easily
found
by inspection.
For,
if
(x
+
a) (x
+
b),
that is x
2
+
(a
+
b)
x +
ab,
is the same as a?
+
px
+
q,
we
must have a
+
b
=p
and ab
=
q.
Hence a and b are such
that their sum is
p,
and their
product
is
q.
For
example,
to find the factors of a;
2
+ 1x +
12. The factors will
be x
+
a and x
+
b,
where a
+
b 7 and ab= 12. Hence we must find
two numbers whose
product
is 12 and whose sum is 7:
pairs
of
numbers whose
product
is 12 are 12 and
1,
6 and
2,
and 4 and
3;
and the sum of the last
pair
is 7. Hence x
2
+
Ix +
12
=
(z+ 4) (x
+
3).
Again,
to find the factors of x
2
-
Ix + 10. We have to find two
numbers whose
product
is
10,
and whose sum is -7. Since the
product
is
+10,
the two numbers are both
positive
or both
negative;
and since the sum is
-
7,
they
must both be
negative.
The
pairs
of
negative
numbers whose
product
is 10 are -10 and
-1,
and -5
and
-2;
and the sum of the last
pair
is -7. Hence x
2
-7a; +
10=
(-5)(*-2).
Again,
to find the factors of x
2
+
3x
-
18. We have to find two
numbers whose
product
is
-
18 and whose sum is 3. The
pairs
of
numbers whose
product
is
-
18 are
-
18 and
1,
-
9 and
2,
-
6 and
3,
-
3 and
6,
-
2 and 9 and
-
1 and 18
;
and the sum of 6 and
-
3 is 3.
Hence
a?+
3x
-
18
=
(x
+
6) (x
-
3).
It should be noticed that if the factors of
x*+px+q
be
x
+
a and x
+
b,
the factors of x
3
+pxy
+
qy*
will be x
+
ay
and x
+
by]
also the factors of
(x
+
y)*+p (x
+
y)
z
+
qz?
will bex
+
y
+
az and x
+
y
+
bz.
56 FACTORS.
Hence from the above we have
x
2
+
Ixy
+
12j/
8
=
(x
+
4y) (a:
+
3j/)
,
x
s
+
3*y
8
-
18y*=(x
+
6y
a
)
(x
-
3j/
2
),
(o
+
6)
2
-
7(fl+6)s
+
10x=(a+6-6x) (0+6- 2x),
and x-5^
+
4=(a~
1
)
2
-5a:
3
+4
=
(x
it
-4)(z
2
-
=
(*
+
2) (*- 2) (*
+
!)(*-!).
EXAMPLES IV.
Find the factors of the
following expressions
:
1.
a
4
-166
4
. 2. 16a;
4
-81a
4
6
4
.
3.
16-(3a-26)*.
4.
4y
t
-(2z-x)\
5. 20aV
-
45axy*.
6.
36aV-4aVy
4
.
7.
(3a
8
-&')
2
-(a
2
-3&
2
)'.
8.
(5a
8
-
36*)*
-
(3a
2
9.
(5^
+
2a;-3)'-(x
2
-2a:-3)
8
.
10.
(3x
8
-
4aj
-
2)
2
-
(3^
+ 4x
-
2)
f
.
11. 32a
8
6
3
-46
9
. 12.
(a
f
-
26c)
8
-
86V.
13. a*-2o-8. 14, aj+12-x
8
.
15. l-18x-63x*. 16. 8a-4a
2
-4.
17. a*b
-
4a'6
8
+ 3ab
3
. 18.
19.
(6
+
c)
8
-6a(6
+
c)
+
5a'.
20.
9(a
+
6)
8
-6(a
+
6)(c
+
rf)
+
21. a;
4
-29x
a
+100. 22.
23. x
4
-8a;
2
7/V+16yV.
24. 9a-
25. x
f
-
2ax
-
6
8
+ 2ab. 26. x*
+
2xy
-
a*
-
2ay.
27. 4
(06
+
cd)
f
-
(a
8
+ &'
-
c*
-
d')>.
28.
4x/-a6
2
-x* +
2
-a*-6
28
.
FACTORS. 57
80. Factors of
general quadratic expression.
We
proceed
to shew how to find the factors of
any
ex-
pression
of the second
degree
in a
particular
letter,
x
suppose.
The most
general quadratic expression [Art. 60]
in x
is ax*
+
bx
+
c,
where
a,
b and c do not contain x.
The
problem
before us is to find two factors which are
rational and
integral
with
respect
to
x,
and are therefore
each of the first
degree
in
x,
but which are not
necessarily,
and not
generally,
rational and
integral
with
respect
to
arithmetical numbers or to
any
other letters which
may
be involved in the
expression.
The method of
finding
the factors of ax*
+
bx
+
c con-
sists in
changing
it into an
equivalent expression
which
is the
difference of
two
squares.
We first note that since a?
+
Zax
+
a
2
is a
perfect
square,
in order to
complete
the trinomial
square
of which
x* and 2ax are the first two
terms,
we must add the
square
of
a,
that
is,
we must add the
square of half
the
coefficient
of
x.
by
the addition of
(
-
I
;
also
x*-px
is made a
perfect square, namely
For
example,
x
2
+
5x is made a
perfect square, namely
[
x
+
^
\
,
the addition of
(
-
I
;
also
3?-px
is ma
\*/
(x
-
|Y,
by
the addition of
(-
1
Y=
.
81. To find the factors of aa?
+
ax
2
+
bx
+
c
=
a
(
a?
+
-
x
+
-
}
.
\
a
aj
Now
x*+
-
x is made a
perfect square, namely (x
+
)
,
a
\ 2a/
by
the addition of f
=-)
==
^
.
And,
by adding
and sub-
6
8
tracting
r-^
to the
expression
within
brackets,
we have
58
FACTORS.
/
,
,
b b* 6 c\
a I a.
;
+
-
a?
+
7-2-7-1
j
+
-
V
a 4a 4a
8
a/
&Y
/6
2
c
^-
-
-7-5
2a/
V4a
a/
Hence as the difference of
any
two
squares
is
equal
to
the
product
of their sum and
difference,
we have
oa?
s
+
fa? + c
Thus the
required
factors have been found*.
Ex. 1. To find the factors of a? +4a; + 3.
Ex. 2. To find the factors of a;
8
-
5x + 3.
Ex. 3. To find the factors of 3s
3
-
ix
+
1.
Ex. 4. To find the factors of 3? +
2ox
-
6
s
-
2a&.
*
It will be
proved
later on
[see
Art.
91]
that an
expression containing
x can be resolved into
only
one set of factors of the first
degree
in x.
FACTORS.
59
82. Instead of
working
out
every
example
from the
beginning
we
may
use the formula
cw;
2
+
bx
+
c
b
/b*-4>ac} f
b
/&'-4oc)
I5W~"l^H*
+
^
V"^
1
"
"I*
and we should then
only
have to substitute for
a,
b and c
their values in the
particular
case under consideration.
Thus to find the factors of 3z
2
-
4x
+
1 . Here a
=
3,
6
=
-
4,
c
=
1.
/6
2
-4ac 716-12
/I
1
., .
HenCe
V
-*?-
=
V
~36-
=
V
9
=
3
5
* ex
P
resMOn
therefore
equivalent
to
3(2!-_--f-Va;----)=:3(a;-s]
(*-!)
83. We have from Art. 81
a#
2
4- bx
+
c
b &'-4acW b
___
Now,
for
particular
values of
a, b, c, 7-5
may
be
positive,
zero,
or
negative.
if
-
^fftf*
L Let
j-j
be
positive.
Then the two factors of
o#*
+
bx
+ c will be rational or irrational
according
as
V 4sac . .
c
8
is or is not a
pertect square.
TT T 4.
~
U fPK
II. Let
7-*
e zero.
I hen
4a
a^
+
6a;
+ c
=
a(a;+ =-] [+ -^ )
.
V 2aJ \ 2a/
Hence oaf
+
bx
+
c is a
perfect square
in
x,
if 6
2
4oc
=
0.
O
___
tLftif*
III. Let
j-j
be
negative.
Then no
positive
or
negative quantity
can be found whose
square
will be
equal
to
a ;
for all
squares,
whether of
positive
or
nega-
4fl
tive
quantities,
are
positive.
60 FACTORS.
Expressions
of the form V
a,
where a is
positive,
are
called
imaginary,
and
positive
or
negative quantities
are
distinguished
from them
by being
called real.
We shall consider
imaginary quantities
at
length
in a
subsequent chapter
: for our
present purpose
it is sufficient
to observe that
they obey
all the fundamental laws of
Algebra
;
and this
being
the
case,
the formula of Art. 81
will hold
good
when b
z
4>ac is
negative.
Note. For some
purposes
for which the factors of
expressions
are
required,
the
only
useful factors are those
which are
altogether
rational : on this account irrational
and
imaginary
factors are often not shewn.
Thus,
for
example,
the factorisation of of 8 is for
many purposes
complete
in the form
(x 2) (#
2
+
2#
+
4)
*,
the
imaginary
factors of a?
+
2x
+
4,
namely
x
+
1
+
V^~3 and x
+
1
-
V^3,
not
being
shewn.
84. We have in Art. 81 shewn how to resolve
any
expression
of the second
degree
in a
particular
letter into
two factors
(real
or
imaginary)
of the first
degree
in that
letter.
It should be noted that the factors of the most
general
expression
of the third
degree,
or of the fourth
degree,
can be
found,
although
the methods are
beyond
the
range
of this book
;
expressions
of
higher degree
than the fourth
cannot
however,
except
in a few
special
cases,
be resolved
into factors.
85. Factors found
by re-arrangement
and
grouping
of terms. The factors of
many expressions
can be found
by
a suitable
re-arrangement
and
grouping
of the terms.
For
example
ax-2?-ax*
*
The reason of this will
appear
from Art. 179 and Art. 193.
FACTORS. 61
or we
may
write the
expression
in the form
1 -a? +
ax- aa?=l
-a?+ax(l
-a
8
),
and the factors 1 -a;
8
,
1
+
ax are now obvious.
For the best
arrangement
or
grouping
no
general
rule
can be
given
: the
following
cases are however of
frequent
occurrence and of
great importance.
I. When one of the letters occurs
only
in the
first
power,
the factors often become obvious when the
expres-
sion is
arranged according
to
powers
of that letter.
Ex. 1. To find the factors of ab + be
+
cd +
da.
Arranged according
to
powers
of a we have a
(b
+
d)
+
be
+
cd,
which is at once seen to be
a(b
+
d)
+
c
(b
+
d) (a
+
c) (b
+
d).
Ex. 2. To find the factors of a?
+
(a
+
b
+
c)
x
+
ab
+ ac.
The
expression
=
a
(x
+
b
+
c)
+
x*
+
bx
+ ex
=
(a
+
x) (x
+
b
+
c)
.
Ex.3. To find the factors of ax
3
+x +
a
+ 1.
Ex. 4. To find the factors of a
2
+
2a6
-
2ac
-
3&
2
+ 2&c.
The
given expression
is of the first
degree
in c
;
we therefore write
it in the form a
s
+
2ab
-
36
2
-
2c
(a
-
b)
II. When the
expression
is of the second
degree
with
respect
to
any
one of the
letters; factors,
which are rational
and
integral
in that
letter,
can be found as in Art. 81.
Ex. 1. Find the factors of a?
+
36
8
-
c
2
+
2bc
-
4a6.
Arranging according
to
powers
of
a,
we have
a*
-
4a& +
36*
-
c
a
+2bc
=
o
2
-
4aZ>
+
4ft
2
-
46
+
3&
2
-
c
2
+
26c
=
(o
-
b
-
c) (a
-
36
+
c)
.
Ex. 2. Find the factors of a
2
-
ft
2
-
c
j
+
d
2
-
2
(ad
-
be).
The
expression
=a
2
-
2ad
-
i
2
-
c+eP+
26c
=a-2ad+d
2
-&
2
-c
2
+
26c=(a-d)
a
-(6-
:=
(a
--
d
+ b
-
c) (a
d
-
b
+
c).
62 FACTORS.
Ex. 8. Find the factors of a? +
2ab
-
ac
-
3&*
+
5bc
-
2c*.
The
expression
=a+
a
(26
-
e)
-
36* +
5&c
-
2c*
=
(a
+
?^
V
- ^
{46
s
-
4bc +c
+
126
2
-
20bc
+ Sc
3
}
Ex. 4. Find the factors of x*
+
x*
-
2ax
+
1
-
a
2
.
Arranging according
to
powers
of
a,
we have
III. When the
expression
contains
only
two
powers
of a
particular
letter and one of those
powers
is the
square
of the
other,
the method of Art. 81 is
applicable.
Ex. 1. To find the factors of
*
-
or thus:
Ex. 2. To find the factors of x*
+
x*
+ 1.
Two real
quadratic
factors can be found as follows :
Ex. 3. To find the factors of x
8
-
28x +27.
**
-
28x
+ 27
=
af
-
28x
+
14
-
14
+ 27
=
(a?
-
14)*
-
13
In this
ease,
and also in Ex.
1,
two factors can be seen
by
inspection,
as in Art. 79.
FACTORS. 63
Ex.4. To find the factors of a
4
+
fc
4
+
c
4
-
26
2
c
2
-
2c
2
a
-
2a
2
6
2
.
Arranging according
to
powers
of
a,
we have
a
4
-
2a
2
(fc
2
+ c
2
)
+
fc
4
+
c
4
-
26
2
c
2
=
a
4
-
2a
2
(fc
2
+
c
2
)
+
(fc
2
+
c
2
)
2
-
(fc
2
+
c
2
)
2
+
fc
+
c
4
-
2fc
2
c
2
=
{a
2
-
(b*
+c
2
)}
2
-
4fc
2
c
2
=
(a
2
-
fc
2
-
c
2
-
2bc) (a
2
-
6
2
-
c
3
+
2fcc)
=
{a
2
-(6
+
c)
2
}{a
2
-(6-c)
2
}
=
(a
+ fc
+
c)
(a
-
6
-
c) (a
-
fc
+
c) (a
+
6
-
c)
.
IV. Two factors of aP*
+
bP
+
c,
where P is
any
expression
which contains
x,
can
always
be found
by
the
method of Art. 81
;
for we have
/ I
/6
2
-4ac\ / 6
/&*
a
(
p+
25+ V -SP-
) (
p+
5- V
Ex. 1. To find the factors of
(a;
2
+
x)*
+
4
(,r
2
+
*)
-
12.
Since P
2
+
4P-12
=
(P-2) (P
+
6),
the
given expression
=
(y?
+ x
-
2) (a;
2
+
x
+
6)
=
(a;+2)(a;-l)(x
2
+x+
6),
the factors of a;
2
+
x
+
6
being imaginary [see
Art.
83,
Note].
Ex. 2. To find the factors of
(a;
2
+
x
+
4)
2
+
8x
(y?
+
x +
4)
-f
15z
2
.
The
given expression
=
{ (x
2
+
x
+
4)
+
3x} { (a?
+
x
+
4)
+
5x]
4)
Ex. 3. To find the factors of
Since
2P"+
5PQ
+
2Q
2
=
the
given expression
=
(So;
2
+ 6x +
3)
(Sx
2
+
12a;
+
3)
Ex. 4. To find the factors of
(a?
+
x +
1) (a?
+
*
+
2)
-
12.
The
given expression
=
(a;
2
+
z)
2
+
3
(3?
+
x)
-
10
64 EXAMPLES.
EXAMPLES V.
Find the factors of the
following expressions
1. x*
+
ax*
-
x
-
a.
2. ac bd ad+bc.
3. ac' + bd'-ad'-bc*.
4. acx
9
+
(be
+
ad) xy
+
bdy'.
5. acx
3
+ bcx* + adx + bd.
6.
(a
+
&)'
+
(a
+
c)-(c
+
rf)
f
7. a* + a
3
b
-
ab
3
-
b\
8. a
*
-
a*b
-
ab* + b\
9. a'b'-a*-b*+l.
10.
xy-*V-yV
+ 3
4
.
11. a;
9
yV
x*z
y*z
4- 1.
12. x
4
+
x*y
+ xz
s
+
yz
3
.
13. x
(x
+
z)
-
y (y
+
z).
14. x
4
-7x* -18.
15. x
4
-
23x* + 1.
18. x
4
-
UxY
+
y\
17. x' + x
f
+l.
18.
x'-2(a'
+
b*)x
t
+
(a'-b)\
19. a;
4
-4x
4
y
2
2
3
+
4yV.
20.
x*-2(a
+
b)x-ab(a-2)(b
21. x
3
+ bx* + ax + ab.
22.
(l+yy-2x
t
(l+y>)
+ x
4
(l
23. x
9
-
y
9
-
82*
-
2xz +
iyz.
FACTORS. 65
24.
2y*
~
oxy
+ 2x*
-ay
ax -a".
25. a
3
-
36*
-
3c
8
+ 106c
-
2ca
-
2ab.
26. 2a
2
-7a&-22&'-5a4-356-3.
27. l +
(6-a
2
)ar-a&c
3
.
28. 1
-
2ax
-
(c
-
a
8
)
x* + aca;
3
.
29. a
s
(6-c)-(-6-'(c-a)+c
8
(a-6).
30. 6*0 + be* + c
2
a + ca* + a*b + ab* + 2abc.
31. a
2
6
-
aft* -f a'c
-
ac
s
~
'lobe + b'c + be
1
.
32.
33.
34. 2a;
3
-
4x
2
y
-
3^2
+
2x2^
+
2xyz
-
y*z.
35.
36.
37.
38.
(a; +1) (a: +2) (a; +3) (a:
+
4)
-24.
39.
(a;+l)(x
+
3)(a;
+
5)(a;
+
7)
+ 15.
40.
4(
+
5) (a:
+
6)
(x
+
10) (SB
+ 1
2)
-
3x.
86. Theorem. The
expression
x
n
a" is divisible
by
a;
a,
for
all
positive integral
values
of
n.
It is known that a
a,
x
2
a* and x
s
a' are all
divisible
by
x a.
We have x
n
-
a"
=
x
n
-
ax"'
1
+
ax
n
~
l
-
a"
=
ar""
1
(a:
-
a)
+
a
(x
n
~
l
-
a"'
1
).
Now
if
x a divides x
n
~
l
a""
1
it will also divide
a;""
1
(x a)
+ a
(x
n
~
l
a""
1
),
that
is,
it will divide x
n
a".
Hence,
if
x a divides x
n
~
l
a
n
~
l
it will also divide
x*
-
a".
S. A.
5
66
FACTORS.
But we know that x a, divides x*
a*;
it will therefore
also divide x* a*.
And,
since x a divides x* a
4
it will
also divide a? a
6
. And so on
indefinitely.
Hence x* a" is divisible
by
x
a,
when n is
any
positive integer.
87. Since x*
+
a*
=
x* a"
+
2a" it follows from the
last Article that when x*
+
a* is divided
by
x a the
remainder is
2a",
so that x*
+
a" is never divisible
by
x a.
If we
change
a into
a,
x a becomes x
( a)
=
x
+
a
;
also x* a* becomes x*
( a)",
and x*
( a)"
is x*
+
a*
or x" a*
according
as n is odd or even.
Hence,
when n is odd
x*
+
a" is divisible
by
x
+
a,
and when n is even
x
n
a" is divisible
by
x
+ a.
Thus,
n
being any positive integer,
x a divides x* a"
always,
x a x*
+
a"
never,
x
+
a x* a" when n is
even,
and
x
+
a x*
+ a* when n is
odd.
The above results
may
be written so as to shew the
quotients:
thus
x a
a*- a"
the
upper
or lower
signs being
taken on each side of the
second formula
according
as n is odd or even.
88. Theorem.
If any
rational and
integral expres-
sion which contains x vanish when a. is
put for
x,
then will
tc-o.be a
factor of
the
expression.
FACTORS. 67
Let the
expression, arranged according
to
powers
of
x,
be
ax*
+
bx*~
l
+ cx
n~*
+ . .
Then,
by supposition,
aa"
+
ba
n
~
l
+
cat""
2
4 =0.
Hence ax*
+
bx"'
1
+
cx'~~ +
=
ax* + bx*'
1
+
caT*
+
-
(aa
n
+
fca"'
1
+ COL"'-
+ . .
.)
=
a
0"
-
a")
+
b
(x*-
1
-
a"'
1
)
+
c
(x*~*
-
a"'
2
)
+
But,
by
the last
Article,
x
n
-
a",
x"~
l
-
a"'
1
,
x
n
~
2
-
a""
2
,
&c. are all divisible
by
x a.
Hence also ax*
+
bx*'
1
+
cx*~* + is divisible
by
x a.
The
proposition may
also be
proved
in the
following
manner.
Divide the
expression
ax
n
+bx
n
~
1
+
cx
n~*
+
by
x
a,
continuing
the
process
until the
remainder,
if there be
any
remainder,
does not contain
x;
and let
Q
be the
quotient
and R the remainder.
Then,
by
the nature of
division,
aaT
+
bx*-
1
+
cx*~*
+
=
Q(x- o)+R,
and this relation is true for all values of x.
Now since R does not contain
a;,
no
change
will be
made in R
by changing
the value of x :
put
then x
=
a,
and
we have
aof
+
ba*-
1
+
ca"-
2
+
=
Q (a
-
a)
+R
=
R.
Hence,
if any expression
rational and
integral
in x
be divided
by
x
a,
the remainder is
equal
to the result
obtained
by putting
a in the
place of
x in the
expression.
It therefore follows that the
necessary
and sufficient
condition that an
expression
rational and
integral
in a?
may
be
exactly
divisible
by
x a is that the
expression
should vanish when a is substituted for x.
52
68
FACTORS.
Ex. 1. Find the remainder when x
3
-
4x
s
+
2 is divided
by
x
-
2.
The remainder
=2*
-
4 . 2
a
+
2
=
-
6.
Ex. 2. Find the remainder when X
s
-
2a
s
a;
+
a* is divided
by
x
-
a.
The remainder is a
3
-
2a
s
+
a
3
=
0,
BO that x
5 -
2a
s
x
+
a
3
is divisible
by
x
-
a.
Ex. 3. Shew
by
substitution that
x-l, x-5,
x
+
2 and x
+
4 are
factors of x*
-
23s
2
-
IS*
+
40.
Ex. 4. Shew
by
substitution that a
-
b is a factor of
Put a=b and the
expression
becomes o
s
(a-c)
+
a
3
(c-a),
which
is
clearly
zero : this
proves
that a
-
6 is a factor.
Ex. 5. Shew that a is a factor of
89. We have
proved
that x a is a factor of the
expression
ax*
+
bx*~
l
+
cx*~*
+
...... ,
provided
that the
expression
vanishes when a is
put
for x.
If the division were
actually performed
it is clear that
the first term of the
quotient,
which is the term of the
highest degree
in
x,
would be ax
n
~
l
. Hence the
given
ex-
pression
is
equivalent
to
(x-
a) (ax"'
1
+ &c.......
).
Now
suppose
that the
given expression
also vanishes
when
#=/3;
then the
product
of a; a. and ax"'
1
+ ......
will vanish when x
=
$\
and since x a. does not vanish
when x
=
yS,
it follows that ax
n
~
l
+
...... must vanish
when x
=
/8.
Hence x
/S
is a factor of ax*'
1
+
&c.
; and,
if the division were
performed,
it is clear that the first
term of the
quotient
would be ax
n
~*.
Hence the
original expression
is
equivalent
to
(x-a) (x
-
/3) (ax*~*
+ &c.......
).
Similarly,
if the
original expression
vanishes also for
the values
7,
8,
&c. of
x,
it must be
equivalent
to
(a?- a) (x
-
) (x -y)(x- 8)
......
(ax
n
~
r
+
&c.......
),
FACTORS. 69
where r is
equal
to the number of the factors x
a.,
x
$,
&c.
If therefore the
given expression
vanishes for n values
a,
$, 7,
&c. there will be n factors such as
xa,
and the
remaining
factor,
ax
n~*
+
&c. will reduce to
a;
and hence
the
given expression
is
equivalent
to
a(x a)
(x
/3) (x 7)
COR. If
any
of the factors x
a,
x
/9,
. . . occur more
than once in ax
n
+
bx
n
~
l
+
. . .
,
it can
similarly
be
proved
that the
expression
is
equivalent
to a
(x a)
p
(x
$)*
. . .
,
the factors x
a,
x
{$,
...
occurring respectively^?, q,
...
times,
and
p
+
q
+
. . .
=
n.
90. Theorem. An
expression of
the nth
degree
in x
cannot vanish
for
more than n values
of
x.
For if the
expression
ft fjf* ^\n
/i
/y*
.1.
f*
/Y* I
vanishes for the n values
a, /3, 7
,
it must be
equivalent
to
a
(x
OL) (x 0) (x
7)
If now we substitute
any
value,
k
suppose,
different
from each of the values
a, /3,
7,
&c.; then,
since no one of
the factors k
a.,
k
/3,
&c. is
zero,
their continued
product
cannot be
zero,
and therefore the
given expression
cannot
vanish for the value x
=
k,
except
a
itself
is zero.
But,
if a is
zero,
the
original expression
reduces to
bx
n
~
l
+
cx
n~*
+
,
and is of the
(n-l)*
degree;
and
hence as before it can
only
vanish for n 1 values of
x,
except
b is zero. And so on.
Thus an
expression
of the nth
degree
in x cannot
vanish
for
more than n values
of
x,
except
the
coefficients of
all the
powers of
x are
zero;
and when all these coefficients
are
zero,
the
expression
will
clearly
vanish for all values
of #.
70 FACTORS.
91. Theorem.
If
two
expressions of
the nth
degree
in x be
equal
to one another
for
more than n values
of
x,
they
will be
equal for
all values
of
x.
If the two
expressions
of the nth
degree
in x
ax*
+
bo;"'
1
+
c#"~*
+
......
,
and
px*
+
qx*~
l
+ rx*~*+
......
,
be
equal
to one another for more than n values of
x,
it
follows that their
difference,
namely
the
expression
will vanish for more than n values of x.
Hence,
by
Art.
90,
the coefficients of all the different
powers
of x must be zero.
Thus
a-p=Q, 6-2
=
0,
c-r
=
0,
&a
that
is,
a
=
p,
b
=
q,
c
=
r,
&c.
Hence,
if
two
expressions of
the nth
degree
in x are
equal
to one another
for
more than n values
of
x,
the
coefficient of any power of
x in one
expression
is
equal
to the
coefficient of
the same
power of
x in the other
expression.
When
any
two
expressions,
which have a limited
number of
terms,
are
equal
to one another for all values
of the letters
involved,
the above condition is
clearly
satisfied,
for the number of values must be
greater
than the
index of the
highest power
of
any
contained letter.
Hence when
any
two
expressions,
which have a
limited number of
terms,
are
equal
to one another for all
values of the letters involved in
them,
we
may equate
the
coefficients of
the
different powers of any
letter.
92. Theorem. A rational
integral expression
con-
taining
x can be resolved into
only
one set
of factors of
the
first degree
in x.
For,
if it be
possible,
let the
expression
ax*
+
bx"~
l
+ ...
be
equivalent
to
a
(x
-
a)" (x /S)
7
. . .
,
and also to a
(x
-
f)
1
(x- n)
m
...
FACTORS. 71
Put x
=
a in both
expressions
;
then a
(a
-
f)' (a ij)
m
. . .
must
vanish,
and therefore one at least of the
quantities
,
?;,
... must be
equal
to a. Let
=
a
;
remove one
factor a; a from both
expressions,
and
proceed
as before.
We thus
prove
that
every
factor of one
expression
occurs
to as
high
a
power
in the other
expression;
the two ex-
pressions
must therefore be identical.
93.
Cyclical
order. It is of
importance
for the
student to attend to the
way
in which
expressions
are
usually arranged.
Consider,
for
example,
the
arrange-
ment of the
expression
be
+
ca
+
ab. The term which does
not contain the letter a is
put
first,
and the other terms
can be obtained in succession
by
a
cyclical change of
the
letters,
that is
by changing
a into
b,
b into c and c into a.
In the
expression
a
2
(6 c)
-f b'*
(c a)
+
c
2
(a 6)
the same
arrangement
is
observed;
for
by making
a
cyclical change
in the letters of a
2
(b c)
we obtain 6
2
(c a),
and another
cyclical change
will
give
c
a
(a 6).
So also the second and
third factors of
(b
c)
(c a) (a b)
are obtained from the
first
by cyclical changes.
94.
Symmetrical expressions.
An
expression
which
is unaltered
by
interchanging any pair
of the letters which
it contains is said to be a
symmetrical expression.
Thus
a
+
b
+
c,
be
+
ca
+
ab,
a
8
+
6
8
+
c
8
8160 are
symmetrical
expressions.
Expressions
which are unaltered
by
a
cyclical change
of the letters involved in them are called
cyclically sym-
metrical
expressions.
For
example,
the
expression
(6 c) (c a) (a 6)
is a
cyclically symmetrical expression
since it is unaltered
by changing
a into
b,
b into
c,
and c into a.
It is clear that the
product,
or the
quotient,
of two
symmetrical expressions
is
symmetrical,
for if neither of
two
expressions
is altered
by
an
interchange
of two letters
their
product,
or their
quotient,
cannot be altered
by
such
interchange.
72 FACTORS.
It is also clear that the
product,
or the
quotient,
of
two
cyclically symmetrical expressions
is
cyclically sym-
metrical.
Ex. 1. Find the factors of
a?(b-c)
+ b
2
(c-a)
+
c-(a- b).
If we
put
b
=
c in the
expression
a
a
(6-c)
+
6
s
(c-a)
+
c(a-6)
(i)
it is
easy
to see that the result is zero.
Hence b- c is a factor of
(i),
and we can
prove
in a similar
manner that c
-
a and a b are factors.
Now
(i)
is an
expression
of the third
degree;
it can therefore
only
have three factors.
Hence
(i)
is
equal
to
L(b-c)(c~a)(a-b) (ii),
where L is some
number,
which is
always
the same for all values of
a, b,
c.
By comparing
the coefficients
[See
Art.
91]
of a
s
in
(i)
and
(ii)
we
see that L
=
-
1.
We can also find L
by giving particular
values to
a,
b and c.
Thus,
let o
=
0,
6
=
1,
c
=
2;
then
(i)
is
equal
to
-2,
and
(ii)
is
equal
to
2L,
and hence as before L=
-
1.
Ex. 2. Find the factors of
a?(b
-
c)
+
6
s
(c
-
a)
+
c(a
-
6).
As in the
preceding example, (6
-
c), (c
-
a)
and
(a
-
b)
are all
factors of
a
s
(&-c)
+
&
$
(c-a)
+
c
s
(a-&) (i).
Now the
given expression
is of the
fourth degree
; hence,
besides
the three factors
already found,
there must be one other factor of
the
Jirst degree,
and this factor must be
symmetrical
in
a, 6, c,
it
must therefore be a-i-
b
+
c.
Hence the
given expression
must be
equal
to
L(b-c)(c-a)(a-b)(a
+
b
+
c) (ii),
where L is a number.
By comparing
the coefficients of a
3
in
(i)
and in
(ii)
we see that
L=
-1;
hence
a(&-c)
+
6
s
(c-a)
+
c*(a-6)=-(6-c)(c-a)(a-6)(a
+
6
+
c).
We can also find L
by giving particular
values to
a, b,
and c.
Thus,
let
a=0,
6
=
1, c=2;
then
(i)
is
equal
to -6 and
(ii)
is
equal
to
6L,
so that L
=
-
1.
\Ve
may
also
proceed
as follows:
Arrange
the
expression according
to
powers
of a
;
thus
a
s
(b
-
c)
-
o
(b
3
-
c
3
)
+
be
(b-
-
c
2
).
FACTORS. 73
It is now obvious that 6
-
c is a
factor,
and we have
(b-c) {a
3
-a
=
(b
-
c) (c
-
a) {V
2
+
bc-a--ac}=-(b- c)
(c
-
a) (a
-
b) (a
+
b
+
c).
Ex. 3. Find the factors of 6
2
c
2
(b
-
c)
+
cW
(c-a)
+
a
2
6
2
(a
-
6).
By putting
b
=
c in the
.expression
fe
2
c
2
(&-c)
+
c
2
a
2
(c-a)
+
a
2
6
2
(a-&)
.....................
(i),
it is
easy
to see that the result is
zero;
hence b-c is a factor of
(i).
So also c
-
a and a
-
b are factors.
The
given expression being
of the
fifth degree,
there must
be,
besides the three factors
b-c, c-a, a-b,
another factor of the
second
degree
; also,
since this factor must be
symmetrical
in
a, b, c,
it must be of the form L
(a
2
+
fc
2
+
c
2
)
+
M
(be
+
ca
+
ab).
Thus
(i)
is
equal
to
Equating
coefficients of a
4
in
(i)
and in
(ii)
we see that
L=0;
and then
equating
coefficients of 6*c* we see that M 1. Hence
(i)
is
equal
to
-(b-c) (c-a) (a-b)
(bc
+
ca +
ab).
We
may
also
proceed
as follows.
Arranging according
to
powers
of
a,
the factor b c which does
not contain a becomes obvious
; then, arranging
according
to
powers
of
b,
the factor c-a which does not contain b becomes
obvious;
and
so on. Thus
&
2
c
2
(b-c)-
a?
(b
9
'-
c
3
)
+
a?
(b-
-
c
2
)
=
(b
-
c) {6
2
c
2
-
a-(b'
2
+
bc
+
c
2
)
+ a
3
(b
+
c)}
=
(&
-
c) {6
2
(c
2
-
a
2
)
+
a-b
(a
-
c)
+
a
2
c
(a
-
c)}
=
(b
-
c) (c
-
a) {(6
2
-
a
2
)
c
+
6
2
a
-
a
2
6}
=
-
(b
-
c) (c
-
a) (a
-
b) (be
+
ca
+
ab).
EXAMPLES VI.
Find the factors of the
following expressions:
1.
(y-z)*
+
(
z
-
x
)*
+
(
x
-yY.
2.
(y-
z
y
+
(
z
-
x
y
+
(
x
-yy.
3. a
4
(b
2
-
c
2
)
+ 6
4
(c
2
-
a
3
)
+ c
4
(a
2
-
&
2
).
4. ab-c
3
+ bc-a
3
+
ca-b}
3
.
74 EXAMPLES.
5. a
(b
-
cf
+ b
(c
-
a)
8
+ c
(a
-
&)*.
6. be
(b c)
+ ca
(c a)
+ ab
(a b).
7. 6V
(b
-
c)
+ cV
(c
-
a)
+ a
B
b
3
(a
-
b).
8. a
4
(&-c)
+ &
4
(c-a)
+ c
4
(a-6).
9. a
s
(6-c)
+ 6
5
(
c
-)
+ c
'(
a
-
&
)-
10.
(a
+ b +
c)
8
-
(6
+ c
-
a)
8
-
(c
+ a
-
6)
8
-
(a
+ b
-
c)
3
.
11.
(
a +
&+c)
5
-(&
+
c-a)
6
-(c
+
a-&)
6
-(a
+
&-c)
5
.
12. a
(6
+ c
-
a)
s
+ b
(c
+ a
-
b)*
+ c
(a
+ b
-
c)*
+
(b
+ c
-
a) (c
+ a
-
b) (a
+ 6
-
c).
13. a*
(b
+ c
-
a)
+ b'
(c
+ a
-
b)
+ c'
(a
+ b
-
c)
(6
+ c
a) (c
+ a
-
b) (a
+ 6
c).
14.
(b
+ c
a) (c
+ a
b) (a
+ b
-
c)
+ a
(a
b +
c) (a
+ b
c)
+
b(a
+ b
c) (
a + 6 +
c)
+ c
(-
a + 6 +
c) (a
-
6 -f
c).
15.
(6
-
c) (a
-
b +
c) (a
+ b
-
c)
+
(c
-
a) (a
+ b
-
c) (-a
+ b +
c)
+
(a
-
6)
(-
a + 6 +
c) (a
-
b +
c).
16.
(x
+
y
+
zl'-x'-if-z
3
.
17.
(x
+
y
+
zf
-
X
s
-
y*
-
z".
18.
(&-c)(&
+
c)'
+
(c-a)(c
+
a)*+(a-&)(a
+
&)
f
.
19.
(b
-
c) (b
+
c)
8
+
(c
-
a) (c
+
a)
3
+
(a
-
b) (a
+
b)
3
.
20.
(6-c)(6
+
c)
4
+(c-a)(c
+
a)
4
+(a-6)(a
+
6)
4
.
21. a
3
+b
3
+ c
a
+5abc-a(a-b)(a-c)-b(b-c)(b-a)
c
(c a)
(c 6).
22. a
3
(a
+
b)(a
+
c)(b-c)+b'(b
+
c)(b
+
a)(c-a)
+ c*
(c
+
a) (c
+
6) (a
-
6).
23.
(y
+
z) (z
+
x) (x
+
y)
+
xyz.
24. a
2
(&
+
c)'
+
&'(c
+
a)
f
+
c*(a
+
6)*
+
a5c(a
+ & +
c)
+
(a*
+ 6
f
+
c*)
(be
+ ca +
ab),
25.
(x
+
y
+
zy-(y
+
z)*-(z
+
x)'- (x
+
y)
4
+ x
4
+
y*
+ z\
26. a"
(b
+ c
-
2a)
+ b*
(c
+ a
-
2ft)
+ c'
(a
+ b
-
2c)
+ 2
(c
s
-a
s
)(c- 6)
+
2(a*-6*)(a-c)
+ 2
(&'-c')(6
-
a).
EXAMPLES. T5
27. b + c-a-db-ca-d
+
(
a + b- c
-d)
4
(a
-
b) (c-d).
28. Shew that
12
{(x
+
y
+
z)
3"
-
(y
+
z)
s"
-
(z
+
x)
3n
-
(x
+
y)
3'
+ of +
y
3'
+
**"}
is divisible
by
(a;
+
y
+
z)* (y
+
z)
4
(z
+
x)* (x
+
y)
4
+ x
4
+
y
4
+ z
4
.
29. Shew that
+
(a*
+ b* + c* be ca
a6) (b
+ c
a) (c
+ a
b) (a
+ b
c)
=
2abc
(be
+ ca +
ab).
30. Shew that
(b
-
c)
9
+
(c
-
a)
9
+
(a
-
b)
9
-
9
(6
-
c) (c
-
o)
f
(a- b)
3
=
2(a-b)
3
(a-c)
3
+
2(b-c)
3
(b-a)
3
+2(c-a)
a
(c-b)
3
.
31. Shew that
(6
+
c)
3
+
(c
+
a)
3
+
(a
+
b)
3
+
(a
+
df
+
(b
+
d)
3
+(c
+
d)*
32. Reduce to its
simplest
form
4
(a'
+ ab + 6
8
)
8
-
(a
-
6)* (a
+
2b)' (2a
+
b)'.
33. Shew that
a
(j
+ c
*
_
)3
+ ft*
(
c
*
+ o
_
&)
+ c
-
(
a
t
+ 6*
.
c
=)'
is divisible
by
a* + b
4
+ c
4
-
26V
-
2cV
-
2a*b*.
34. Resolve into
quadratic
factors
4
{cd (a
2
-
6
s
)
+ ab
(c'
-
d*)}'
+
{(a'
-
b
3
) (c'
-
d')
-
4a6/}
f
.
35. Shew that
(y
s
-
*)(!
+
y) (1
+
xz)
+
(a?-
x
3
)
(1
+
yg) (1
+
yx)
+(x*-y*) (1 +2x)(l
+y)
=
(y-z) (z-x) (x-y) (xyz +x+y+z).
36. Find the factors of
a
3
(b-c)(c-d)(d-b)-b
>
(c-d)(d-a)(a-c)
+
c*(d-a)(a-b)(b-d)-d
3
(a-b)(b-c)(c-a).
37. Find the factors of
F
(6
-
c) (c
-
d)
(d-b)-
<?d
3
a?
(c -d)(d-
a)
(a
-
c)
+ d*a~b'
(da)(ab)(b-d)-
a"b
2
c"
(a
-
b) (b c) (c a).
CHAPTER VII.
HIGHEST COMMON FACTORS. LOWEST COMMON
MULTIPLES.
95. A Common Factor of two or more
integral alge-
braical
expressions
is an
integral expression
which will
exactly
divide each of them.
The
Highest
Common Factor of two or more
integral
expressions
is the
integral expression
of
highest
dimensions
which will
exactly
divide each of them.
It is usual to write H.C.F. instead of
Highest
Common
Factor.
96. The
highest
common factor of monomial
expressions.
The
highest
common factor of two or more
monomial
expressions
can be found
by inspection.
Thus,
to find the
highest
common factor of a*6
2
c and a
4
^
8
.
The
highest power
of a which will divide both
expressions
is
a
8
;
the
highest power
of 6 is fc
2
;
and the
highest power
of c is
c;
and no other letters are common. Hence the H.C.I', is a
3
6
2
c.
Again,
to find the
highest
common factor of a
3
&
4
c
4
,
a-b
3
and a'fcc*.
The
highest power
of a which will divide all three
expressions
is a
2
;
the
highest power
of b which will divide them all is
b;
and c
will not divide all the
expressions.
Hence the U.C.F. is o
2
&.
From the above
examples
it will be seen that the
H.C.F.
of two or more monomial
expressions
is the
product
of
each letter which is common to all the
expressions
taken
to the lowest
power
in which it occurs.
HIGHEST COMMON FACTORS. 77
97.
Highest
common factor of multinomial
expressions
whose factors are known. When the
factors of two or more multinomial
expressions
are
known,
their H.C.F. can be at once written
down,
as in the
pre-
ceding
case. The H.C.F. will be the
product of
each
factor
which is common to all the
expressions
taken to the lowest
power
in which it occurs.
Thus,
to find the H.C.F. of
(x-2)
s
(z-l)-(x-3)
and
(x
-
2)
!
(x
-
1) (x
-
3)*.
It is clear that both
expressions
are divisible
by (x
-
2)
2
,
but
by
no
higher power
of x-2. Also both
expressions
are divisible
by
x
-
1,
but
by
no
higher power
of x 1
;
and both
expressions
are divisible
by x-3,
but
by
no
higher power
of a; -3. Hence the H.C.F. is
(*- 2)2
(3 -!)(*- 3).
Again,
the H.C.F. of a
-2
6
3
(a
-
6)
2
(a
+
6)
3
and a
8
6
s
(a-6) (a
+
6)
2
is
o6
2
(a-6)(a+6)
2
.
In the
following examples
the factors can be seen
by
inspection,
and hence the H.C.F. can be written down.
Ex. 1. Find the H.C.F. of a<6
2
-a
2
6* and a
Ans. a
2
Ex. 2. Find the H.C.F. of aP-cfb* and a
6
6
2
-16a
2
6
6
.
Ans. a
2
6
2
Ex. 3. Find the H.C.F. of a
3
+
3a?b
+
2aV
t
and a
4
+
6a
3
fe
+ 8a
2
6
2
.
Ans. a
(a
+
26).
98.
Although
we
cannot,
in
general,
find the factors
of a multinomial
expression
of
higher degree
than the
second
[Art. 84],
there is no
difficulty
in
finding
the
highest
com mon
factor
of
any
two multinomial
expressions.
The
process
is
analogous
to that used in arithmetic to find
the
greatest
common measure of two numbers.
If the
expressions
have monomial
factors,
they
can be
seen
by inspection
;
and the
highest
common factors of
these monomial factors can also be seen
by inspection:
we have therefore
only
to find the multinomial
expression
of
highest
dimensions which is common to the two
given
expressions.
78 HIGHEST COMMON FACTORS.
Let A and B stand for the two
expressions,
which are
supposed
to be
arranged according
to
descending
powers
of some common
letter,
and let A be of not
higher degree
than B in that letter. Divide B
by
A,
and let the
quotient
be
Q
and the remainder R
;
then
.: R
=
B-AQ.
Now an
expression
is
exactly
divisible
by any
other if
each of its terms is so
divisible;
and therefore B is divisible
by every
common factor of A and
R,
and R is divisible
by every
common factor of A and B. Hence the common
factors of A and B are
exactly
the same as the common
factors of A and
R;
and therefore the H.c. F. of A and R
is the H.c.F.
required.
Now divide A
by
R,
and let the remainder be S
;
then
the H.c.F. of R and S will
similarly
be the same as the
H.C.F. of A and
R,
and will therefore be the H.C.F. re-
quired.
And,
if this
process
be continued to
any
extent,
the
H.C.F.
of a/ny
divisor and the
corresponding
dividend will
always
be the H.C.F.
required.
If at
any stage
there is no
remainder,
the divisor must
be a factor of the
corresponding
dividend,
and that divisor
is
clearly
the H.C.F. of itself and the
corresponding
divi-
dend. It must therefore be the H.C.F.
required.
It should be remarked that
by
the nature of division
the remainders are
successively
of lower and lower dimen-
sions
;
and
hence,
unless the division leaves no remainder
at some
stage,
we must at last come to a remainder which
does not contain the common
letter,
in which case the
given expressions
have no H.C.F.
containing
that
letter.
Since the
process
we are
considering
is
only
to be used
to find the
highest
common multinomial
factor,
it is clear
that
any
of the
expressions
which occur
may
be divided or
multiplied by any
monomial
expression
without
destroying
the
validity
of the
process
;
for the multinomial factors
will not be affected
by
such
multiplication
or division.
HIGHEST COMMON FACTORS. 79
Thus the H.C.F. of two
expressions
can be found
by
the
following
Rule :
Arrange
the two
expressions according
to
descending powers of
some common
letter,
and divide the
expression
which is
of
the
highest degree
in the common
letter
by
the other
(if
both
expressions
are
of
the same
degree
it is immaterial which is used as the
divisor}.
Take
the
remainder,
if any, after
the
first
division
for
a new
divisor,
and the
former
divisor as dividend
;
and continue
the
process
until there is no remainder. The last divisor
will be the H.C.F.
required.
The
process
is not used
for
finding
common monomial
factors,
these can be seen
by
inspection
;
and
any
divisor, dividend,
or remainder which
occurs
may
be
multiplied
or divided
by any
monomial ex-
pression.
Ex. 1. Find the H.C.F. of a:
8
+x*
-
2 and x
3
+
2x
2
-
3.
/x
s
-
x-l
x-1
Thus the H.C.F. is x-1.
The work would be shortened
by noticing
that the factors of
the first
remainder, namely
x'
2
-l,
are x-l and x
+
1. And
by
means of Art. 88 it is at once seen that x
-
1
is,
and that x
+
1 is
not,
a factor of
x*+x*~
2.
Ex. 2. Find the H.C.F. of
x
s
+
4x
2
i/-8x7/
+
24y
3
and x
6
-
x*y
+
8x*y
3 -
8xy*.
The second
expression
is divisible
by x,
which is
clearly
not a
common factor : we have therefore to find the H.C.F.
of the first
expression
and x*
-
x*y
+
8xy*
-
8y*.
-
8y*
(x-5y
8x
8 ^
-
5x
3
y
+
8x
2
?/
2
-
l&xy
3 -
By*
-
5x
3
y
-
20x
V
+
40xy
3
-
80 HIGHEST COMMON FACTORS.
The ismainder
=
28j/
a
(z
2
-2xy
+
4</-j:
the factor
2%*
is
rejected
and x
2
-
2xy
+
4i/
2
is used as the new divisor.
x
2
-
2
xy
+
4t/
2
\ x
3
+ 4Z
2
?/
-
8
xy*
+
24y
3
/ x
+
6y
'x
3
-2x-
+
4xi
il
S
Hence a;
2
-
2xy
+
4y
5
is the H.C.P.
required.
Ex. 3. Find the H.C.F. of
and 3z<+ 15x
3
+
5x
2
+10x+2.
To avoid the inconvenience of
fractions,
the second
expression
is
multiplied by
2: this cannot introduce
any
additional common
factors. The
process
is
generally
written down in the
following
form :
23*4- 9ar+
Ux
+
3 \ 3x*
+
ISx*
+
Sar
2
+ Wx
+
2
/2
6x
4
+
SOx
3
+
10z
2
+
20x
+
4 / 3
)J
2x
21*
8
+
132X
2
+ 156x
+ 27 / 7
21x8+
x*-f 5x+
1
lx
(
-
5x-25x-5
Thus x
s
+ 5x +
l is the H.C.F.
required.
Detached coefficients should
generally
be used
[Art. 63].
99. The labour of
finding
the H.C.F. of two
expres-
sions is
frequently
lessened
by
a modification of the
pro-
cess,
the
principle
of which
depends
on the
following
Theorem : The common
factor of highest degree
in
a
particular
letter,
x
suppose, of any
two
expressions
A
and B is the same as the H.C.F.
of pA +qB
and rA
+sB,
HIGHEST COMMON FACTORS. 81
where
p, q,
r,
s are
any quantities positive
or
negative
which
do not contain x.
To
prove
this,
it is in the first
place
clear that
any
common factor of A and B is also a factor of
pA
+
qB
and
of rA
+
sB.
So
also,
any
common factor of
pA
4-
qB
and rA
+
sB
is also a factor of
s(pA
+
qB) q (rA
+
sB),
that
is,
of
(sp qr~)
A.
Hence,
as
(sp qr)
does not contain
x,
any
common factor of
p
A
+
qB
and rA + sB must be a factor
of
A,
provided only
that
p, q,
r,
s are not so related that
sp qr
=
0.
Similarly any
common factor of
pA
+
qB
and rA
+
sB is also a factor of r
(pA
+
qB) p (rA
+
sB),
that is of
(rq ps)
B,
and therefore of
B.
Since
every
common factor of A and B is a factor of
pA
+
qB
and of rA
+
sB,
and
every
common factor of
pA
+
qB
and rA
4-
sB is a factor of A. and of
B,
it follows
that the H.C.F. of A and B is the same as the H.C.F. of
p
A 4-
qB
and rA
+
sB.
Ex. To find the H.C.F. of
2x*+o?-Qx
z
-2x
+
3 and 2x
4
-
3z
3
+2x
-
3.
We
have,
by subtraction,
4ar-6x
s
-4ar + 6 ..............................
(I);
and,
by addition,
-as-8)
...................
(II).
The
required
H.C.F. is the H.C.F. of
(I)
and
(II),
and therefore
of
(I)
and
2*
2
-a;-3 ................................
(III).
Multiply (III) by
2 and add
(I),
and we have another
expression,
namely
4ui?-2x'*-Gx
=
2x(2x*-x-3)
.....................
(IV),
such that the H.C.F. of
(III)
and
(IV)
is the H.C.F.
required.
But the
H.C.F. of
(III)
and
(IV)
is
obviously
2x
2
-
x-3.
100. If
R,
S...... be the successive remainders in the
process
of
finding
the H.C.F. of the two
expressions
A and
B
by
the method of Art. 98
; then,
as we have
seen,
every
common factor of A and B is a factor of
R,
and therefore a
common factor of A and R.
Similarly every
common factor
of A and R is a common factor of R and /S. And so on
;
so
s. A
6
82 HIGHEST COMMON FACTORS.
that
every
common factor of A and B is a factor of
every
remainder,
and therefore must be a factor of the H.c.F.
Hence
every
common multinomial factor of two ex-
pressions
is a factor of their
highest
common multinomial
factor
;
and this is
obviously
true also of monomial factors.
Therefore
every
common factor of two
expressions
is a
factor of their H.C.F.
It can be shewn that
every
remainder,
in the
process
of
finding
the H.c.F. of two
expressions
A and B
by
the
method of Art.
98,
is
expressible
in the form FA
+ GB
where F and are rational and
integral
in x.
For,
if
Qj, Q
2 ,
Q3 ,
...
Qn
be the successive
quotients,
and R
l ,
R
2 ,
R
3 ,
...R
n
be the successive
remainders,
the
last of
which,
R
n ,
is the H.C.F. of A and B if A and B
have
any
common
factor,
but is
independent
of ar if A and
B have no common factor
containing
x\
then we have
Now
R!
is
clearly
of the
required
form
(F
being
Qj
and
being
1),
and
substituting
for
R^
in the second
equation
it will be at once seen that
R?
is of the
required
form.
Also R
k
is of the
required
form
provided
that both R
k-i
and
Rt-t
are of that form
; for,
if
Rk-i
=
LA
+
MB and
R
k_t
=
L'A
+
M'B
;
then
R
k
=
R
k-,_-Rk
-
l
Q
t
=
(L'-Q
k
.L)A
+
(M'-Qk
.M)B,
and the
expressions
L
Qk
. L and M'
Qk
. Mare both
integral,
since
by supposition
L, M, L',
M' and
Qk
are all
integral expressions.
Hence R
n
=
FA
+
GB,
where F and G are
integral
expressions.
If now A and B have no common factor in
x,
then R
n
docs not contain x.
And,
dividing throughout by
R
n>
since
FjR
n
arid
G/R
n
are
integral
in
x,
we have
l=P.A+Q.B,
where P auu
Q
are
integral
in x.
HIGHEST COMMON FACTORS. 83
Thus,
if
A and B be
any
two
integral expressions
con-
taining
x,
but which have no common
factor containing
x,
two other
expressions,
P and
Q,
can be
found
both
integral
in
x,
and such that PA
+
QB
1.
Ex. FindPand
Q
when A =x*-Sx? +
1 and B
=
x*
Am.
P=-i(8x
+
5), Q=j^(
101.
The H.C.F. of three or more multinomial
expres-
sions can be found as follows.
Let the
expressions
be
A, B, G, D,....
Find G the H.C.F. of A and B.
Then,
since the
required
H.C.F. will be a common
factor of A and
B,
it will be a factor of G : we have there-
fore to find the H.C.F. of
G, G,
D....
Hence we first find the H.C.F. of two of the
given
ex-
pressions,
and then find the H.C.F. of this result and of
the third
expression
;
and so on.
NOTE. The
highest
common
factor
of
algebraical
ex-
pressions
is
sometimes,
but
very inappropriately,
called
their
greatest
common measure
(G.C.M.).
If one
expression
is of
higher
dimensions than
another,
in a
particular
letter,
we have no reason to
suppose
that it
is
numerically greater:
for
example,
a
8
is not
necessarily
greater
than
a;
in
fact,
if a is
positive
and less than
unity,
a
3
is less than a.
It should also be noticed that if we
give particular
numerical values to the letters involved in
any
two ex-
pressions
and in their
H.C.F.,
the numerical value of the
H.C.F. is
by
no means
necessarily
the G.C.M. of the values
of the
expressions.
This is not the case even when the
given expressions
are
integral
for the
particular
values
chosen. For
example,
the H.C.F. of
\^X*+\DX+\
and
22#
2
+ 23o;
+
1 will be found to be x
+
1
;
but if we
suppose
x to be
J,
the numerical values of the
expressions
will be 12 and
18,
which have 6 for
G.C.M.,
whereas the
numerical value of the H.C.F. will be
f.
62
4 LOWEST COMMON
MULTIPLE.
EXAMPLES VII.
Find the n. c. F. of
1. a*
-
5ab
+ 46' and a
3
-
5a6 + 46
s
.
2. 2**- 5* + 2 and 12x*- So?- 3x+ 2.
3. 2x
4
-3a,y
+
y
4
and 2x
9
-
3x
4
y"
+
y*.
4. 2s
8
+ 3x
3
y
-
y*
and 4x*
+
xy*-y
3
.
5. x*
-
4y*
+
l'2yz
-
9^ and a* + 2a
-
iy
5
+
8ys
-
6. 20a
4
-
3a
8
6 + 6
4
and 64a
4
-
3o6
8
+ 5b\
7. a'
-
a'b + ab* + Ub
3
and 4a* + 3a'b
-
9a6
f
+ 26*.
8. 2a:
4
+ a;'-9x
8
+ 8x-2 and 2a:
4
-
7a^+ Use*
9. 1 lx
4
-
9<Ke"
-
aV
-
a
4
and 13a;
4
-
lOox
8
-
2aV
-
a
4
.
10. x
4
+ x
s
-9x'-3x+18 and x" + 6x*
-
49z + 42.
11. x
4
-2ic
8
+ 5ic*-4x+3 and 2
4
-"--
6x* + 2x + 3.
12. a;
4
+ 3
2
+ 6 + 35 and x
4
+ 2x
3
-
5x* + 26x + 21.
LOWEST COMMON MULTIPLE.
102. Definitions. A Common
Multiple
of two or
more
integral expressions,
is an
expression
which is
exactly
divisible
by
each of them.
The Lowest Common
Multiple
of two or more
integral
expressions,
is the
expression
uf lowest dimensions which
is
exactly
divisible
by
each of them.
Instead of Lowest Common
Multiple
it is usual to
write L.C.M.
LOWEST COMMON MULTIPLE. 85
103. When the factors of
expressions
are
known,
their L.c.M. can be at once written down.
Consider,
for
example,
the
expressions
a
8
&'(a;-a)
2
(tf-&)
8
and ah"
(x
-
a)
4
(x
-
6).
It is clear that
any
common
multiple
must contain a
8
as
a factor
;
it must also contain 6
4
,
(x a)
4
and
(x 6)
8
.
Any
common
multiple
must therefore have a
s
b*
(x
a)
4
(x 6)
8
as a factor
;
and the common
multiple
which has no un-
necessary
factors,
that is to
say
the lowest common
multiple,
must therefore be aV
(x a)
4
(x 6)
8
.
From the above
example
it will be seen that the L. c. M.
of two or more
expressions
which are
expressed
as the
product
of factors of the first
degree,
is obtained
by taking
every different factor
which occurs in the
expressions
to the
highest power
which it has in
any
one
of
them.
Ex. 1. Find the L.C.M. of
Bx*yz, 21a?yW
and
Ans.
Ex. 2. Find the L.C.M. of 6a&
2
(a
+
b)
2
and 4a
2
6(a
2
-6-).
Ans. 12a
2
6
2
(a
+
ft)
2
(a
-
b).
Ex. 3. Find the L.C.M. of
2axy (x
-
?/)
2
,
Bax*
(x
2
-
j/
2
)
and
4y
2
(x
+
y)
2
.
Ans.
12axV(*
2
-!/
2
)*.
Ex. 4. Find the L.C.M. of x
a
-
3x
+
2,
x
2
-
5x
+
6 and x
2
-
4a?
+ 3.
Ans.
(x-l)(x-2)(x-3).
104. When the factors of the
expressions
whose
L.C.M.
is
required
cannot be seen
by inspection,
their
H.C.F. must
be found
by
the method of Art. 98.
Thus,
to find the L.C.M. of x
3
+
x--2 and ar
3
+
2x
2
-3.
The H.C.F. will be found to be x
-
1
;
and ar
+
a;
2
-2
=
(x-l)(a;
8
+
2z
+
2),
x
3
+
2x*
-
3
=
(x
-
1)
(ar
2
+
Sx
+
3).
Then,
since x
2
+
2x
+
2 and x'
+
Bx
+
B have no common
factor,
the
required
L.C.M. is
(x
-
1) (x-
+
2x
+
2) (x-
+
3x
+
3).
86 LOWEST COMMON MULTIPLE.
105. Let A and B stand for
any
two
integral
ex-
pressions,
and let H stand for their
H.C.F.,
and L for their
L.C.M.
Let a and b be the
quotients
when A and B
respec-
tively
are divided
by
H;
so that
A
=
H. a and B
=
H. b.
Since H is the
highest
common factor of A and
B,
a and b can have no common factors. Hence the L.C.M.
of A and B must be Hx a x b. Thus
L
=
H.a.b.
Hence L
=
Ha x
-jj
=
A x -^
(i);
also LxH
=
HaxHb
=
A xB
(ii).
From
(i)
we see that the L.C.M
of any
two
expressions
is
found by dividing
one
of
the
expressions by
their
H.C.F.,
and
multiplying
the
quotient by
the other
expression.
From
(ii)
we see that the
product of any
two
expressions
is
equal
to the
product of
their H.C.F. and L.C.M.
EXAMPLES VIIL
Find the L.C.M.
of
1. 6x*
-
5ax
-
6a* and 4x
8
-
2ax
s
-
9a
s
.
2. 4a
f
-
5ab + b
1
and 3a
8
-
3a*6 + ab*
-
b*.
3. 3x*
-
I3x* -f 23x
-
21 and 63* + x'
-
44a: + 21.
4. x
4
-
lla? + 49 and 7x*
-
40x
3
+ 75^
-
40* +
7.
5. a^ + 6x* +
lla; + 6 and a;
4
+
*
-
4*
-
4x.
6. a:
4
-*
i
+8a:-8anda;
4
+ 4a;
3
-8x* + 24a;.
7. 8a'
-
1
Sab',
8a" + 8a'6
-
6ab* and 4a
8
-
806 4- 36'.
:
EXAMPLES. 87
8. x* -7x+l
2,
3
2
-
C>x
-
9 and '2x
3
-
Qx
2
-
Bx.
9. 8x
a
+
27,
1 Gx
4
+
36or + 81 and 6x* -5x-G.
10. x
a
-
Qxy
+
Qy\
x*
-
xy
-
Qy
3
and 3x
s
-
1
2y*.
11. x
3
-
7xy
+
12?/
2
,
x
s
-
Qxy
+
8y
3
and x"
-
5xy
+
6y*
12.
Shew
that,
if ax
s
+ bx + c and a'x* + b'x + c' have a com-
mon factor of the form x
+f,
then will
(ac'
-
a'c)
a
=
(be'
-
b'c) (ob
1
-
a'b).
13. Shew
that,
if ax
3
+
bx* + cx + d and a'x
3
+ b'x* + c'x + d'
have a common
quadratic
factor in
x,
then will
ba' b'a ca! c'a da d'a
ad' a'd bd' b'd cd' c'd
'
14. Find the condition that ax
3
+ bx + c and a'x
3
+ b'x + c'
may
have a common factor of the form x
+f.
15. If
<7j,
g
a
, <7
3
are the
highest
common
factors,
and l
lt
l
a ,
1
3
the lowest common
multiples
of the three
quantities
a, b,
c taken
in
pairs; prove
that
ff^^ffj^i^,
(abc)*.
16. If
A, ,
C be
any
three
algebraical expressions,
and
(BC), (CA), (AB)
and
(ABC)
be
respectively
the
highest
common factors of B and
C,
C and
A,
A and
B,
and
A,
B and
C;
then the L.C.M. of
A,
B and C will be
A. B.C.
(ABC)
-f
{(BC)
.
(CA)
.
(AB)}.
CHAPTER
VIIL
FRACTIONS.
106. WHEN the
operation
of division is indicated
by
placing
the dividend over the divisor with a horizontal
line between
them,
the
quotient
is called an
algebraical
fraction,
the dividend and the divisor
being
called
respec-
tively
the numerator and the denominator of the fraction.
Thus
r
means a + b.
b
Since,
by definition, ,
=
a -4-
b,
it follows that
r
x b
=
a.
o o
107. Theorem. The value
of
a
fraction
is not altered
by multiplying
its numerator and denominator
by
the same
quantity.
We have to
prove
that
a
_am
b bin'
for all values of
a,
6 and TO.
'
=
6
;
then
x x b
=
j
x b
=
a,
by
definition.
6
'
FRACTIONS. 89
Hence o?x6xm
=
axm;
.-. x x
(6m)
=
am.
[Art.
29,
(ii).]
Divide
by
bm,
and we have
n
, am
x
=
am -r-
(ym)
=
: .
bni
108. Since the value of a fraction is not altered
by
multiplying
both the numerator and the denominator
by
the same
quantity,
it follows
conversely
that the value of
a fraction is not altered
by dividing
both the numerator
and the denominator
by
the same
quantity.
Hence a fraction
may
be
simplified by
the
rejection
of
any
factor which is common to its numerator and
9
CuSO
denominator. For
example,
the fraction
TJ-
takes the
tXs
2
simpler
form
^
,
when the factor
x,
which is common to
its numerator and
denominator,
is
rejected.
When the numerator and denominator of a fraction
have no common
factors,
the fraction is said to be in its
lowest terms.
To reduce a fraction to its lowest terms we must
divide its numerator and denominator
by
their H.C.F.
;
for
we thus obtain an
equivalent
fraction whose numerator
and denominator have no common factors.
Ex. 1. Eeduce
~
^
to its lowest terms.
bcfcty
The H.C.F. of the numerator and denominator is
3axy
;
and
Sox
2
!/
_
'la
Es.2.
Simplify *
-
Ixy
+
10y
2
_
(x
-
2y) (x-oy)
_x-5y ~
90
FRACTIONS.
,
ar* ax
Ex. 3.
Simplify -; ?
.
or
-
x*
x*-ax
x(x-a)
a
2
-X
s
(a-x)(a-i-.r)
*
Since x
-
a=
-
(a
-
x),
if we divide the numerator and denominator
by
a
-
x,
we have the
equivalent
fraction
- - '-
;
and if we divide the
a
+x
x
numerator and denominator
by x-a,
we have .
-
r
.
By
the
-
(a+x)
Law of
Signs
in Division
-
= .
,
=
--
,
and the last form
a
+
x
-(a
+
x)
a+x
is the one in which the result is
usually
left.
Ex.4.
The H.C.F. will be found to be x*-3a; +
7; and, dividing
the
numerator and denominator
by
a;
a
-3x+7,
we have the
equivalent
fraction
109. Reduction of fractions to a common de-
nominator. Since the value of a fraction is unaltered
by multiplying
its numerator and denominator
by
the
same
quantity, any
number of fractions can be reduced
to
equivalent
fractions all of which have the same de-
nominator.
The
process
is as follows. First find the L.C.M. of all
the
denominators;
then divide the L.C.M.
by
the denomi-
nator of one of the
fractions,
and
multiply
the numerator
and denominator of that fraction
by
the
quotient;
and
deal in a similar manner with all the other fractions : we
thus obtain new fractions
equal
to the
given
fractions re-
spectively
and all of which have the same denominator.
For
example,
to reduce
a b
and
x*y(x
+
y)' xy*(x-y)
to a common denominator.
The L.O.K. of the denominators is x
s
i/
3
(x
j
-y
2
).
Dividing
this
L.C.M.
by
x
3
i/(x+y)l xy
l
(x-y)
and
x^y-(x--if),
we have the
FRACTIONS. 91
quotients y"(x-y),
x
z
(x
+
y)
and
xy respectively.
Hence the
required
fractions are
_ ay'*(x-y)
(x
+
y)
x
3
y (x
+
y)x y*(x
-
y)
~
x
s
y
3
(x
2
-
j/
2
)
'
b
bxo?(x+y)
_
_ bx*(x
+
y)
xy
3
(xy)
~
xy
3
(x-y)
x x
z
(x
+
y)
~
X
3
y
3
(x
2
-y'^)
t
c
_
cxxy cxy
~ ~
It is not
necessary
to take the lowest common
multiple
of the
denominators,
for
any
common
multiple
would answer the
purpose ;
but
by using
the L.C.M. there is some
saving
of labour.
110. Addition of fractions. The sum
(or
differ-
ence)
of two fractions which have the same denominator
is a fraction whose numerator is the sum
(or difference)
of
their
numerators,
and which has the common denominator.
This follows from Art. 43.
When two fractions have not the same
denominator,
they
must first be reduced to
equivalent
fractions which
have the same denominator : their
sum,
or
difference,
will
then be found
by taking
the
sum,
or
difference,
of their
numerators,
retaining
the common denominator.
When more than two fractions are to be
added,
or when
there are several fractions some of which are to be added
and the others
subtracted,
the
process
is
precisely
the
same. The fractions must first be reduced to a common
denominator,
and then the numerators of the reduced
fractions are added or subtracted as
may
be
required.
Ex. 1. Find the value of
--
H
--
-
.
a+b
a-b
The L.C.M. of the denominators is
(a
+
b) (a
-6);
and
1
,
1 a-b a+b
a+b
a-b
(a
+
b) (a-b) (a-b) (a
+
b)
(a+b) _
2o
~
a ab
Ex. 2. Find the value of
-
r
+
,-.
a-b b
2
-a
|)2
FRACTIONS.
The L.C.M. of the denominators is a
2
-
It";
and we have
a(a+b)
-ab
_a(a
+
b)-ab_
a
2 l~2 ^ "j.!(
^~
2 7
"
^
'2 Ijl
*
Ex. 3.
Simplify
-^-
+
-^
+
-5 -5
+
-_/
a x a
+x
-2 i -a .! , ~*
In this case it is not desirable to reduce all the fractions to a
common denominator at once : the work is
simplified by proceeding
as under :
a a
_
a(a
+
x)
+ a
(a-x)
_
2a
2
a-x a
+
x~ a
2
-x
2
~a
s
-s*'
2a 2a
2
+
,
. 4a* 4a*
and
finally
.
-
-.
*-
4
[The
above would be shortened
by observing that, except
for the
factor
2,
the second addition
only
differs from the first
by having
a
2
and x
2
instead of a and x
respectively
;
and hence the result of the
addition can be written down at once. So also the result of the
third addition can be written down from the first or
second.]
Ex. 4.
Simplify
+ .
J
x -3 x-1
x+1 x+3
Here
again
it is best not to reduce all the fractions to a common
denominator at once : much labour is often saved
by
a
judicious
arrangement
and
grouping
of the terms.
1 1
(a
+
8)-(s-8)
6
x-3 x
+
3~ x
a
-9
~x
2
-9
f
and
__3_
+
_j_
=
-3(x
+
D
+
3(x-l)
=
-6
x-1 x
+
1 x'
!
-l x
2
-!'
48 -6
_6(x*-l)-6(x
2
-9)
+
- ~
(x-9)(x
2
-!)*
a
2
6
2
c
2
Ex. 6.
Simplify
.
-
-r.
-
r-+
TT-
-
r71
-
r + -.
-
^.
-
rr
.
(a
-
6) (a
-
c) (6
-
c) (b
-
a) (c
-
a) (c
-
6)
The L.C.M. of the denominators is
(b
-
c)(c-a)(a-b) [See
Art.
93].
Hence we have
a?(c-b)
+
b*(a-c)+c*(b-a)
(b-c)(c-a)(a-b)
Now we
naturally test, by
the method of Art.
88,
whether either
of the factors of the denominator is a factor of the numerator: we
are thus led to find that the numerator is the same as the denomi-
nator,
so that the
given expression
is
equal
to
unity.
FRACTIONS. 93
Ex. (5.
Simplify
_
a
2
_ _
tf
_ _
c
a
(a-
b) (a
-
c) (x
+
a)
+
(b-c) (b -a)(x
+
b)
+
(c- a) (c
-
b) (x
+
c)
'
The L.C.M. of the denominators is
(b-c)
(c -a)(a-b) (x
+
a) (x
+
b) (x
+
c).
The
expression
is therefore
equal
to the fraction whose denominator
is this
L..C.M.,
and whose numerator is
a-(c-b)(x
+
b)(x
+
c)
+
b-(a-c)(x
+
c) (x
+
a)+c*(b-a) (x
+
a)
(x+b).
Arranging
the numerator
according
to
powers
of
x,
the coefficient
of x
a
is a
2
(c
-
b) +b*(a
-
c)
+
c*(b
-
a)
=
(b
-
c) (c
-
a) (a
-
b).
The coefficient
of * is a
2
(c
3
-
J>
2
)
+
6
s
(a
2
-
c
2
)
+
c
2
(b
2
-
a
2
)
=
0.
The term which does not contain x is
abc
{a (c
-
b)
+
b
(a
-
c)
+
c
(b
-
a)}
=
0.
Hence the numerator is a;
2
(5-c)(c-a) (a-b),
and therefore the
given expression
_
_
x
2
(6-c)(c-ffi)(a-6)
_
_
_
x
2
_
~(b-c) (c-a) (a-b) (x
+
a) (x
+
b) (x
+
c)
~
(x
+
a) (x
+
b) (x
+
c)
'
111.
Multiplication
of fractions. We have now
to shew how to
multiply algebraical
fractions.
fl (*
Let the fractions be
T
and
-
7
.
o a
r
a c
Let
x
=
-f
x
-
7 ;
b d
then xxbxd=
r x-jXbxd
b d
a
7
c
7
=
T
x b x
-
7
x
,
b a
by
the Commutative
Law.
fl (*
But,
by
definition,
. x b
=
a,
and
->
x d
c;
CL
ac
!) I- FRACTIONS.
Thus the
product of any
two
fractions
is another
fraction
whose numerator is the
product of
their
numerators,
and
whose denominator is the
product of
their denominators.
The continued
product
of
any
number of fractions is
found
by
the same rule. For
a c e
_ac
e
_
ace
b
X
d
X
f
=
bd
X
f~bdf'
and
similarly,
however
many
fractions there
may
be.
Hence
/a\* a a aa a* , /a\" a"
(b)
=
6
X
6
=
66
=
6'
;
and
'
m
general
(b)
=
^
'
CL G
112. Division
of fractions. Let
r
and
j
be
any
6 a
two tractions
;
and let a;
=
r
-5-
^
o a
m,
caeca
Then ajx-T
=
r-J-jX 7=7;
d b d d b
c d a d
.'.
a;x-jX-=
r
x-.
d c b G
,-r
ad
Hence x
=
T
x
,
c
c d cd
since -. x
-
=-=-
=
1.
d c ac
>
Thus o divide
by any fraction
-,
is the same as to
multiply by
the
reciprocal fraction
-
.
c
As
particular
cases of
multiplication
and
division,
we
have
a a c ac
_
x f
=
X =
b b }~
b
'
FRACTIONS. 95
a, a c a I a
and
r
-T- c
=
T-=-T
=
T
X
~
=
I~'
o o 1 o c be
Note. It should be noticed that the rules for the
multiplication
and division of
algebraical
fractions are
simply
rules
concerning
the order in which certain
opera-
tions of
multiplication
and division
may
be
performed,
and
have
really
been
proved
in Art 33.
Thus
r
x -,
=
(a
-r
6)
x (c -5-
d)
b d
=
a-i-b x c -r-d
ac
=
axc-=-6-=-d
=
(ac)
Ex. 1.
Simplify *
*
X
s
+
a
3
x-a
_
(x
3
+
a
8
)
(x
-
a)
_
(x
2
-
ax
+ a
2
) (x
+
a) (x
-
a)
_
x
s -
ax
+
a
8
Ex. 2.
Simplify
-
_
x
y xy
_
y
-
x
y*-x
s
_y-x
xV
8
X7/
1 1
y'-x
2
zy
'
x
2
^
2
x,y y^
yr x
+
y
1
"
2
^~
a+x a-x
Ex.3.
Simplify
a
"
X
a
+
x a-x
a x a
+x
a + x a-x
_
(a
+
x) (a+x) -(a-j) (a-x)
4ax
a-x a
+
x~ a
2
-x
2
~
a
8
-
x
,
a + x a-x
_
(a
+
a
) (a
+
x)
+
(a
-
x) (a
-
x)
_
2a
2
+
2z
a
a
-
x a
+ x a
2
-
x
a
"
a
2
-
x
3
*
96 FRACTIONS.
Hence the
given
fraction is
equal
to
4ax
a"
-
x
1
a*
-
as
8
a*
-
a;* 2a
a
+
2.r
2
a
2
-f x
3
"
113. The
following
theorems
(the
second of which
includes the
first)
are of
importance:
Theorem I.
If
the
fractions
^
,
^
,
^
8
,
<fcc. be all
equal
to one
another,
then will each
fraction
be
equal
to
pa,
+
qaa
+
ra. -f
. .
Let each of the
equal
fractions be
equal
to x.
Then,
since
r
1
=
a?,
a
l
=b
l
x\
r
so also
qay
=
qbt
x,
ra
s
=
rb
t
x,
Hence,
by
addition,
pa.
+
qa.
+ ra. +
a,
*___ *_
* 3
__
/p
.
i *
(VC
Theorem II.
//"
the
fractions
r
1
,
p
,
r*
,
.
to one
another,
then will each
fraction
be
equal
to
;f-jTi
where A is
any homogeneous expression
of
the nth
degree
in
a,, a,, a,,
<c. and B is the same
homogeneous
expression
with b
1
in the
place of a,, 6,
in the
place of
a
t
,
tfec.
Let each of the
equal
fractions be
equal
to
x,
so that
ct
=
6,#,
=
bx,
a
=
i>a,',
ic.
FRACTIONS. 97
Let
Xa^
a/
av... be
any
term of A
;
then
\bfbf
6
8
y
...
will be the
corresponding
term of B
;
and since the ex-
pressions
A and B are
homogeneous
and of the nth
degree,
Now
Xa,
a
a/ a/
. . .
=
X
(fc^
=
X
(&,
since a
+
yS
+
7
+
... =n*
Hence
any
term of A
=
of x
corresponding
term of B
;
/. sum of all the terms of A
=
of x sum of all the terms of
B,
that is A
=
x" . B
;
which
proves
the theorem.
Theorem III.
If
the denominators
of
the
fractions
r-
1
,
j*
,
j-*
,
...... be all
positive,
then will the
fraction
- 1~11!
the
greatest of
the
fractions
^
,
j*
,
&c.
Let
^
be the
greatest
of the
fractions,
and let
~
\
\ v\
then
r^
<
x, -^
<
x,
&c.
6
8 6,
Hence, 6,,
6
2
,...
being
all
positive,
we have
=
x . b
a; .
*
We have in the above assumed certain lesults which will be
proved
in
Chapter
XIII
T
S. A.
7
98 FRACTIONS.
Hence
by
addition
a
l
4-a
8
+
a
8
+ < x
(b,
+
6,
+
b,
+
);
Hence
r
1
r
8
r
1
is less than the
greatest
of the
b
1
+ b
t
+ b
3
+...
fractions;
and it can be
similarly proved
to be
greater
than the least of the fractions.
Ex. 1. Shew
that,
if
?
_
3
,
then will
-
=
-
.
b d a-b e-d
_
,
a
, c
Let
r=;
then
-
=x.
o d
a
+
b bx +
b x
+
l dx +
d c
+
d
Hence
-
=
-
=
-
= =
-
.
a-b bx-b x-l dx-d c-d
Or thus:
a c
Smce
6=3'
,
.
b d b d
a . c . a-b c-d
Also
T
-
1
=
-.
-
1,
that is
r
=
3-
.
b a b d
a+b a-b
c+d
c-d
Hence
_,.__
=
_.,-_;
a+b c+d
a-b~ e-d'
7
=
3,
o a
Ei. 2. Shew
that,
if
7
=
3,
then will each fraction be
equal
to
[This
is a
simple
case of Theorem
II.,
Art.
113.]
t j[
cy
+ te
=
<" +
c
*_te+ay
th
I TO n
6cx
cay
abz
Ex. 3. Shew
that,
if
=
-
=
,
then will
I m n
-al
+
bm
+
cn al-bm + cn al+bm-cn'
EXAMPLES.
99
Each of the
given equal
fractions
a 2bcx
-al+bm+ cn -al
+
bm
+
cn
2cay
2abz
and
similarly
=-
;
r
-
=
.
al-bm+cn
al + bm-cn
EXAMPLES IX.
Simplify
the
following
fractions :
SOaWafyV
SaWVys*
O R rt"
^
"
A "2. A~"~
3.
2
. 4.
(a;
3
y
3
)
(x
+
y)
R
&
"/ a a\~~7 \
**
O/3
.
Q/>* /*
_ /vfl /*
.1.
1
^*C -r */C
""
Jt *C^*C CTl
* '
i O,3 , O,2 i 0-v. i 1
'
m
2x*+2x+l' x
t
-2x*-x*-2x+l
Q _**- *:_
54a;
s
-27a;
4
-3a
8
-4
10.
11.
x*(y-z)
+
y*(z-x)
+ z*
(x-y)
(y
+
zf
+
(z+ x)*
+
(x
+
y)*
_ . a
(6 c) (c
-
c)
c
(o? a) (a 6)
b(c-d)(d-a)-d(a-b)(b-c)'
100
EXAMPLES.
x*(y-z) +y* (z-x)
+**
(x-y)
2a 26
^ **
16.
r
+
5
+
17.
a + 6 a
-
6 6* a*
'
3- 3 + a; 1
-
16s
1
-
3x 1 + 3x
18.
y
x+
2y 2y
x x* 4z
t7 *7 J
8aa
19
_
,
x+2a 2a x of
ML
-^*--
f4
X + 4 05 +
1 3 3 1
*" ~^""
"^~^T7a'
22.
_V_J-
+
-
6
-JL
+
x 2a x a x
23.
5xy
+
6y*
x*
4xy
+
3y*
a b
24
7
(a-0)(a-c) (0-c)(D-a) (c-a)(c-0)
as
*
.
y
,
c
*
(a
-
b) (a
-
c) (b
-
c) (b-a) (c- a) (c
-
6)
*
26.
^
6) (a c) (6 c) (6 a) (c a) (c
be
(a
+
d) ca(b
+
d) ab(c
+
d)
(a-b)(a-c)
+
(b-c)(b-a)
+
(c-a)(c-b)
'
(y+z)(y
EXAMPLES.
101
29
(y-x)(z-x)
(-y)(-y)
(x
-
2y
+
z) (x
+
y
-
2z) (x
+
y- 2z) (-
2x +
y
+
z)
+
(*-x)(z-y)
(-
2x +
y
+
z) (x
-
2y
+
z)
'
x + a x 4- b x + e
_
(x
+
a)(x
+
b) (x
+
c)
x-a x b x-c
(x a) (x
-
b) (x c)
x x x
n
x*+
(be
+ ca +
ab)
x
x a x b x c
(xa)(x-b)(x c)
a" 6" c"
(a-b)(a-c)
+
(b-c)(b-a)
+
(c-a)(c-b)'
32
__
(a-b)(a-c) (b-c)(b-a)
+
(c-d)(c-b)'
, (a
+
b)(a
+
c)
,
(b
+
c)(b
+
a)
(a-6)(a-c) (b-
c
)(b-a) (c-a)(c-6)'
/I 1\ /I 1\ /I 1\-
CM
T-
1 + o I
)~*"
c
l r)
\o c/ \c
a/ \o o/
31.
35.
1
,
_J
(o
-
b +
c) (a
+ b
>
-
c) (a
+ 6
-
c) (-
a + 6 +
c)
(-
a + 6 +
c) (a
-
b +
c)
'
36'
a
~(^T
+
ft^^'
+
c'
-""(a
-
6)"
37. Shew that
16+
j^L?
+
a
i^ -
2
5^}
*=
1 6
{x-a
x + a x +a
)
38. Shew that
a + b a-b a'x
+ b
s
y
_
aV
-
b
*y
3
ax +
by
ax
by
a~x* +
b'y*
a*x*
b*y
4
'
102
EXAMPLES.
39. Shew that
/
(i)
(a
-
6) (a
-
c) (1
+
as) (b
-
c) (b-a) (1
+
bx)
c 1
+cx)
a b
(a-b)(a-c)(l+ax) (b
-
c) (6
-
a) (1
+
bx)
C X
(iii)
(c
-
a) (c
-
6) (1
+
ex) (1+ ax) (1
+
bx) (1
+
ex)
'
1 1
(a
-
J) (a
-
c) (
1
+
ax)
(b
-
c) (b
-
a) (
1 +
bx)
I x
2
(c-a)(c-6)(l+ca;) (1 +oo;)(l +bx)(l
+
ex)
'
40.
Simplify
(a
-
6) (a
-
c) (x
+
a) (b
-
c) (b -a)(x
+
b) (c-a)(c-b) (x+c)'
41.
Simplify
a
(6
+ c
a)
6
(c
-f a
6)
c
(a
+ 6
c)
(a-6)(a-c) (6-c)(6-a) (c-a)(c-6)'
42.
Simplify
(a
-
6 +
c) (a
+ 6
-
c) (a
+ b
-
c) (-
a + 6 +
c)
(o
-
6) (a
c) (6
-
c) (b a)
(-
a + 6 +
c) (a
-
6 +
c)
43.
Simplify
a
(6
+
c)
6
(c
+
a)
c
(a
+
b)
6+c a c+a-6 a+6 c'
EXAMPLES.
103
/
IV / 1\
2
/ 1 \
2
44. Shew that I m +
)
+
(n
+ -\ +
(
mn +
)
\
mj \ nj
\
mn/
/ 1\ / 1\ / 1 \
{
m +
\ [n
+ I I mn H I
=
4.
\
m/ \
nj
\
mw/
45. Shew that
(2&c
.1 fl
t I
. fl5
'
)
'
W 1
1
&r+ i-
+
i s-r
+
1i
c
r
*
IT
+
oir=
6c
-
\6
+ c
J lc 6-2cJ (.6
+ c
J (6 c-26J
46. Shew that
b c c a a
b_
(b
-
c) (c
-
a) (a
-
b)
l+bc l+ca
1 + ab~
(1 +bc)(l +ca)(l
+
u.b)'
47.
Simplify
I 1 1\ /I 1 1
48. Shew
that,
if
y
+ z
_z
+ x
_
b c c a a b*
then will each fraction be
equal
to
Jx*
49. Shew that, if
-
=
T ,
then will
y
b'
x + a
y
+ b x +
y
+ a + b
50. Shew
that,
if
x
=
y
=
z
6-hC-a c + a-b a + b c*
then will
(b
-
c)
x +
(c
-
a) y
+
(a
-
b)
z
=
0.
1 04 EXAMPLES.
KI ou *i *
-
bz-cy
cx-az
51. Shew
that,
if -r
* =
,
6-c c a
.
ay
bx
and c be not
zero,
then will each
equal
=- .
a b
and a
(y z)
+ b
(z
-
x)
+ c
(x y)
=
0.
52. Shew that
a
4
b
4
c
4
+
(a-b)(a-c)(a-d) (6-c) (b-d) (b-a) (c-d) (c-a)(c-b)
d
4
(d-a)(d-b)(d-c)
53. Shew that
." .'
=
a + b + c
(al -a,)(al
-a
a
)
......
(a,
-
a.) (a2
-
aj (a,
-
a,)
......
(,-,)
o
r
+ ^
--s-
(a.
-
a,) (a.
-
a,)
......
(a.
-
_,)
is
equal
to zero if r be less than TO
1,
to 1 if r
=
n
l,
and to
a,
+ a
s
+ ...
+_,
if r
=
n.
54. Shew that
a,
CR a; cr. ..
1 +
-
-f-
7
-
-^
-
+
;
-
r-
* -_
+ ...
-a, (a:
-a,)
(a:
-a,)
(x
-a,)
(as-of)
(-,)
a a;""
1
a;"
55. Shew that
6 c
o(a+
6-fc-t- ... + k +
l) a(a
+
b) (a
+
b)(
I
... +
,
(a
+
b + ... +
fy (a
+ 6 + ... +
k +
J)
'
EQUATIONS.
ONE UNKNOWN
QUANTITY.
114. A STATEMENT of the
equality
of two
algebraical
expressions
is called an
equation;
and the two
equal
expressions
are called the
members,
or
sides,
of the
equation.
When the
equality
is true for all values of the letters
involved the
equation
is,
as we have
already
said,
called
an
identity,
the name
equation being
reserved for those
cases in which the
equality
is
only
true for certain
particular
values of the letters involved.
For the sake of
distinction,
a
quantity
which is
sup-
posed
to be
known,
but which is not
expressed by any
particular
arithmetical
number,
is
usually represented by
one of the first letters of the
alphabet,
a, 6, c, &c.,
and
a
quantity
which is
unknown,
and which is to be
found,
is
usually represented by
one of the last letters of the
alphabet
x,
y,
z,
&c.
115. We shall in the
present chapter only
consider
equations
which contain one unknown
quantity.
To solve an
equation
is to find the value or values of
the unknown
quantity
for which the
equation
is true
;
and
these values of the unknown
quantity
are said to
satisfy
the
equation,
and are called the roots of the
equation.
Two
equations
are said to be
equivalent
when
they
have the same roots.
106
EQUATIONS.
ONE UNKNOWN
QUANTITY.
An
equation
which contains
only
one unknown
quantity,
x
suppose,
and which is rational and
integral
in
x,
is said
to be of the
first degree
when x occurs
only
in the first
power ;
it is said to be of the second
degree
when a? is the
highest power
of x which occurs
;
and so on.
Equations
of the
first,
second and third
degrees
are
however
generally
called
simple, quadratic
and cubic
equa-
tions
respectively.
116. In the solution of
equations frequent
use is
made of the
following principles.
I. An
equation
is
equivalent
to that formed
by adding
the same
quantity
to both its members.
For it is clear that A
+ m
=
B
+
m
when,
and
only
when,
A
=
B.
II.
Any
term
may
be transformed
from one side of
an
equation
to the
other,
provided
its
sign
be
changed.
Let the
equation
be
a + b
c=p q
+ r.
Add
p
+
q
r to both sides
;
then a+ b c
p+q r=p q
+
r
p
+
q
r,
that
is,
a
+
b c
p
+
q
r
=
0.
We thus have an
equation equivalent
to the
given
equation,
but with the terms
p, q,+r changed
in
sign
and
transposed.
By
means of
transposition
all the terms of
any equa-
tion
may
be written on one side of the
sign
of
equality
and zero on the other side.
III. An
equation
is
equivalent
to that formed
by
multiplying (or dividing)
each of its members
by
the same
quantity
which is not
equal
to zero.
For,
if A
=
B,
it is clear that mA
=
mB.
Conversely,
if
mA
=
mB,
that is m
(
A
B)
=
0,
it follows that A B
=
0,
since m is not zero. Hence mA
=
mB
when,
and
only
when,
A=B.
EQUATIONS.
ONE UNKNOWN
QUANTITY.
107
The case of division
requires
no
separate
examination,
for to divide
by
m is the same as to
multiply by
.
117.
Simple Equations.
The method of
solving
simple equations
will be seen from the
following examples.
Ex. 1. Solve the
equation
13z
-
7
=
5x
+ 9.
Transpose
the terms 5x and
-
7;
then I3x
-
5a;= 7 + 9.
Thatia 8*
=
16.
Divide both sides
by
8,
the coefficient of
x\
then x
=
2.
Ex.2. Solve the
equation
-
2
=
+
5.
4 o
We
may get
rid of fractions
by multiplying
both members
by 20,
the least common
multiple
of the
denominators;
we then have
15* -40
=
8*
+
100,
or
transposing
15x
-
8x
=
100 +
40
;
A 7*=140.
Divide
by
7,
the coefficient of
x;
then a;
=
20.
Ex. 3. Solve the
equation
a
(x
-
a)=2ab -b(x-b).
Bemoving
the
brackets,
we have
ax
-
a?
=
2ab
-
bx
+
b",
or
transposing
ax
+
bx
=
2ab
+
b
2
+
a
2
,
that is
x(a
+
b)
=
(a
+
b)'*.
Divide
by
+
6,
the coefficient of
x;
then
a
+
6
From the above it will be seen that the different
steps
in the
process
of
solving
a
simple equation
are as follows.
First clear the
equation
of
fractions,
and
perform
the
algebraical operations
which are indicated. Then trans-
pose
all the terms which contain the unknown
quantity
to one side of the
equation,
and all the other terms to the
other side. Next combine all the terms which contain
the unknown
quantity
into one
term,
and divide
by
the
108
SIMPLE
EQUATIONS.
coefficient of the unknown
quantity
: this
gives
the re-
quired
root.
118.
Special
Case?.
Every simple equation
is re-
ducible to the form
ax+b
=
0,
the solution of which is
x . The
following
are
special
cases.
a
I.
If 6
=
0,
the
equation
reduces to ax
=
;
whence
<c
=
0.
II. If 6
=
and also a
=
0,
the
equation
is
clearly
satisfied for all values of x.
III. Ifa
=
and b
=t=
0.
Suppose
that while b remains
constant,
a takes in
succession the values
^
,
^
,
^
,.
. .
;
then will x take
in succession the values
106,
10
2
6,
10
8
6.... Thus as
a becomes
continually
smaller and
smaller,
x will become
continually greater
and
greater
in absolute
magnitude
;
moreover,
by making
a
sufficiently
small,
x will become
greater
than
any assignable quantity;
for
example,
in
order that the absolute value of x
may
be
greater
than
10* it is
only necessary
to
give
to a an absolute value
less than
y^
.
This is
expressed
by saying
that,
in the
limit,
when
a becomes
zero,
the root of the
equation
ax
+
b
=
is
infinite.
The
symbol
for
infinity
is oo .
EXAMPLES.
Solve the
equations
Ans. x=12.
2.
(*-3)-(x-8)
+
(x-5)=0.
Ant. x
=
0.
EQUATIONS.
ONE UNKNOWN
QUANTITY. 109
3.
a(x-a)
=
b(x- b).
Ana. x=
4.
(x
+
a)(x+b)-(x-a)(x-b)
=
(a
+
b)*.
Ans.
x=(
5.
a(2x-a)
+
6(3se-6)
=
2a&. -ins. z=
6.
(a?+x)(b*
+
x)
=
(ab+x)*.
Ans. s=0.
7.
3(aj
+
3)
a
+5(x
+
5)
2
=8(a;+8).
Ans. x=-6.
8.
(x
+
a)
4
-(aj-o)
4
-8a
8
+8a
4
=0. Ans. x=-a.
9.
(x-l)*+x
3
+
(x+l)
a
=3x(x
t
-l).
Ant. x
=
0.
10.
Ans.
x=--^
B
119.
Equations expressed
in Factors. It is clear
that a
product
is zero when one of its factors is zero
;
and
it is also clear that a
product
cannot be zero unless one of
its factors is zero.
Thus
(as 2) (x 3)
is zero when x 2 is
zero,
or when
x 3 is
zero,
and in no other case.
Hence the
equation
0-2)0-3)=0,
is satisfied if # 2
=
0,
or if # 3
=
0;
that
is,
if x
=
2,
or
if x
=
3,
and in no other case. The roots of the
equation
are therefore 2 and 3.
Again,
the continued
product
(x
a) (x b}(x c)...
is zero when x a is
zero,
or when x b is
zero,
or when
x c is
zero,
&c.
;
and the continued
product
is not zero
except
one of the factors x
a,x b,x
c
t
&c. is zero.
Hence the
equation
(xa)(x b)(xc}
. . .
=
is
equivalent
to the
system
of
equations
x a
=
0,
ab
=
0,
x c
=
0,
&c.
From the above it will be
apparent
that the solution
110 QUADRATIC EQUATIONS.
of an
equation
of
any degree
can be written down at
once,
provided
the
equation
is
given
in the form of a
product
of
factors
of the first
degree equated
to zero.
Now all the terms of
any equation
can be
transposed
to one
side,
so that
any equation
can be written with all
its terms on one side of the
sign
of
equality
and zero on
the other side.
It follows therefore that the
problem of solving
an
equation of nny degree
is the same as the
problem offinding
the
factors
of
an
expression of
the same
degree.
Ex.1. Solve the
equation
x* -5x=6.
Transposing,
we have x*
-
6x
-
6
=
0,
that is
(x-6)(x
+
1)=0;
.-.
x-6=0,
or x + l
=
0.
Hence
x6,
or x=-l.
Ex. 2. Solve the
equation
a*-a^=6x.
Transposing,
we have x
8
-
x
2
-
6x
=
0,
that is
x(z-3)(z+2)=0;
/.
x=0,
or
x=3,
or x
=
-
2.
120.
Quadratic Equations.
When all the terms
of a
quadratic equation
are
transposed
to one side it must
be of the form
aa?
+
bx
+
c
=
0,
where
a,
b,
c are
supposed
to
represent
known
quantities.
We have
already [Art. 80]
shewn
how to resolve a
quadratic expression
into factors : the same method will
therefore enable us to find the roots of a
quadratic equation.
Hence to solve the
quadratic equation
ant?
+
bx
+ c
=
0,
we
proceed
as follows.
Divide
by
a,
the coefficient of
a^;
the
equation
then
becomes
EQUATIONS.
ONE UNKNOWN
QUANTITY.
Ill
Now add and subtract the
square of half
the
coefficient
of
x,
that is the
square
of
\
-
. Then we have
CL
a,
that is
6
that is (.+
*}'-! /()}'
-0,
\
2aJ
(V
\4a
8
a/j
//&'
c\}(
b
//&" c\\_
V
Ua
2
~ayjr
4
"2a~V
\i?
"
Jj
"
6
/6*
-
4ac
Hence a;
+
-
H- .
/
-
A
-
,
-
=0,
2a
v
4-a*
6
/6
a
-
4ac
_ A
> >v / i r
"
a V
4a
or
'> >v i
2a V
4a
Thus there are two roots of the
quadratic equation,
namely
b_
/fij-Jiac
2a
f
V
4a
*
Ex. 1. Solve
We have X*-13
X+
(Y_
that is
or
\
x- -=0 or x-
.:
x=6,
or x
=
l.
112 QUADRATIC EQUATIONS.
Ex. 2. Solve Bar
8
-
lOz
+
6
=
0.
Dividing by
3,
we have
Hence *=-
(5-^7),
or ar=-
Ex. 3. Solve a
(s
8
+
1)
=
*
(a"
4-
1).
Divide
by
a and
transpose ;
then
,
flence *
3
-x
-
+
( ~H )- ( -ir- I
+1=0.
a
\
2a
/ \
2a
/
s
that is
a
+
l a
2
-! a +
l
+
"*
-
or *--
that is *--
=
0,
or z-a
=
0.
a
Thus the roots are a and
-
.
a
Note. In most cases the factors can be written down
at
once,
as in Art.
79,
without
completing
the
square;
and much labour is
thereby
saved.
EQUATIONS.
ONE UNKNOWN
QUANTITY.
EXAMPLES.
Find the roots of the
following equations
:
1. 9** -24* + 16=0.
2. 5
(x*
+
4)
=4
(a?* 9).
8.
4.
5. x
3
+(a-x)=(a-2*)
3
.
6.
a?+
(a
-
2*)
2
=
(o
-
3x)
3
.
7. *+a;=a
J
+
a.
8. x
2
+ 2aa;
=
6
3
+ 2a&.
9.
(*-a)
a
+
(x-6)
2
=(a-&)
9
.
10.
(a-x)
3
+(a;-6)
3
=(a-7>)
3
.
11.
12.
(x-a+2b?-(x-2a+b)
3
=(a
Aiu. -.
o
Am.
4.
Ans.
3,
-\.
O
^;.
-i
-5.
4 4
^n*.
0,
o.
Ans.
0,
.
I
Ans.
a,
-a-1.
Ant.
b,
-2-6.
Ant.
a,
b.
Ant.
a,
b.
Ans.
1,
^--
o
~~
C
Ans.
a-2b,
2a-b.
121. Discussion of roots of a
quadratic equation.
In the
preceding
article we found that the
quadratic equa-
tion ax*
+
bx
+
c
=
Q had two
roots,
namely
_!+
/:
2a
V
6
2
-4ac .
4oc
,
_
6 Ib
1
4oc
~
o_ A / 3~Tz
and
V
Since . /
j
real or
imaginary according
as
6
a
4ac is
positive
or
negative,
it follows that the roots of
aa?
+
bx
+
c
=
are real or
imaginary according
as 6
2
4ac
is
positive
or
negative.
The roots are
clearly
rational or irrational
according
as 6
2
4oc is or is not a
perfect square.
It should be
remarked also that both roots are rational or both
irrational,
and that both roots are real or both
imaginary.
S. A.
8
114-
SPECIAL FORMS OF
QUADRATIC EQUATIONS.
If 6* 4ac
=
0,
both roots reduce to
5-
,
and are thus
Zfll
equal
to one another. In this case we do not
say
that
the
equation
has
only
one
root,
but that it has two
equal
roots.
It is clear that the roots will be
unequal
unless
6* 4ac
=
0. Hence in order that the two roots of the
equation
o#*
+
bx
+
c
=
may
be
equal,
it is
necessary
and
sufficient that 6
s
=
4ac.
When 6
8
=
4ac,
the
expression
ax*
+
bx
+
c is a
perfect
square
in
x,
as we have
already
seen.
[Art. 83.]
122.
Special
Forms. We will now consider some
special
forms of
quadratic equations,
in which one or more
of the coefficients vanish.
L If c
=
0,
the
equation
reduces to
aa?
+
bx
=
0,
or
(ax
+
6)
=
0,
the roots of which are and .
a
II. If c
=
and also 6
=
0,
the
equation
reduces to
aa?
=
0,
both roots of which are zero.
III. If 6
=
0,
the
equation
reduces to aa?
+
c
=
0,
the
roots of which are +
./
. The roots are therefore
'
V
a
equal
and
opposite
when 6
=
0,
that is when the coefficient
of a is zero.
IV. If
a,
b and c are all
zero,
the
equation
is
clearly
satisfied for all values of x.
V. If a and 6 be zero but c not
zero,
put
x
=
-
in the
equation
cue
8
+
l)x
+
c
=
;
y
I
EQUATIONS.
ONE UNKNOWN
QUANTITY. 115
then we
have,
after
multiplying by y*,
cy*
+
by
+
a
=
0.
Now from I. and II. one root of this
quadratic
in
y
is zero if a
=
0,
and both roots are zero if a
=
and also
6
=
0.
But since #
=
-,
x is
infinity [Art. 118]
when
y
is zero.
\j
Thus one root of a#
2
+
bx
+
a= is infinite if a
=
;
also
both roots are infinite if a
=
and also 6
=
0.
Thus the
quadratic equation
has one root
infinite,
if
a=a';
it has two roots
infinite,
if a=a' and
also b
=
b';
and the
equation
is satisfied for all values of
x,
if
a=a',
b=b' and c=c'.
Again,
the
equation
a(x
+
b) (x
+
c)
+
b(x
+
c)(x+a)=c(x
+
a)(x+b),
is a
quadratic equation
for all values of c
except only
when c=a
+
b,
in which case the coefficient of x
2
in the
quadratic equation
is zero.
When c=a
+
b we
may
still however consider that the
equation
is a
quadratic equation,
but with one of its roots infinite.
Note. It is however to be remarked that since in-
finite roots are not often of
practical importance
in
Algebra,
they
are
generally neglected
unless
specially required.
123. Zero and infinite roots of
any equation.
The most
general
form of the
equation
of the nth
degree
is
ax*+
bx
n
~
l
+...+
kx
+
l
=
Q..............
(i).
If I
=
0,
the
equation may
be written
#
(a*"-
1
+
far
2
-K ..+
&)
=
0,
one root of which is
clearly
zero.
Similarly
two roots will be zero if I and also k
=
;
and so
on,
if more of the coefficients from the end vanish.
82
116
EQUATIONS
NOT INTEGRAL.
Put x
-
;
then we
have,
after
multiplying by y",
by+
......
+
ky
n
~
l
+
ly*
=
0.
From the
above,
one root of the
equation
in
y
will be
zero when a
=
0;
and two roots will be zero if a
=
and
also 6
=
0. But when
y
=
0,
x
=
-
<x>.
*/
Thus one root of
(i)
is infinite when a
=
0,
and two
roots are infinite when a and b are both zero
;
and so
on,
if more of the coefficients from the
beginning
vanish.
124.
Equations
not
integral.
When an
equation
is not
integral,
the first
step
to be taken is to reduce it to
an
equivalent integral equation.
An
equation
will be reduced to an
integral
form
by
multiplying by any
common
multiple
of the denominators
of the fractions which it
contains,
but the
legitimacy
of
this
multiplication requires
examination. For if we
multiply
both sides of an
integral equation by
an ex-
pression
which contains the unknown
quantity,
the new
equation
will not
only
be satisfied
by
all the values of the
unknown
quantity
which
satisfy
the
original equation,
but
also
by
those values which make the
expression by
which
we have
multiplied
vanish. Thus if each member of the
equation
A=B,
be
multiplied by
P,
the
resulting equa-
tion PA
=
PB,
or P
(A B)
=
0,
will have the same roots
as the
equation
A
=
together
with the roots of the
equation
P
=
0.
When however an
equation
contains fractions in whose
denominators the unknown
quantity
occurs,
the
equation
may
be
multiplied by
the lowest common
multiple
of the
denominators without
introducing any
additional
roots,
for
we cannot divide both sides of the
resulting equation by
any
one of the factors of the L.C.M. without
reintroducing
fractions,
which shews that there are no roots of the re-
sulting equation
which
correspond
to the factors of the
expression by
which we
multiply.
EQUATIONS.
ONE UNKNOWN
QUANTITY. 117
3 2x
Ex. 1. Solve the
equation
= H
5=6.
x-o x-3
Multiply by (x-5) (x-3),
the L. c. M. of the
denominators;
then
we have
3
(x
-
3)
+
2x(x- 5)
=
5
(x- 5) (x- 3);
.-. Sz
2
-
33x
+ 84=0.
Whence x=4 or
x=7.
a? Zx 1
Ex.2. Solve the
equation -5 T
+
2
+ -=0.
x^ l x 1
Multiply by
x
2
-
1,
the L.C.M. of the
denominators;
then we have
x*~Sx
+
2(a?-l)+x+l=Q,
which reduces to
3a;
2
-2x-l
=
0,
that is
(3z
+
l)(x-l)=0.
Thus the roots
appear
to be
-
and 1
;
the latter root is however
due to the
multiplication by
a?- 1.
ar'-Sx 1 x
the
equation
is
equivalent
to
x-1
+
2=0,
x
+
1
which has
only
one
root, namely
x=
-
J.
From the above
example
it will be seen that when an
equation
has been made
integral by multiplication,
some of the roots of
the
resulting equation may
have to be
rejected.
Ex. 3. Solve the
equation
:
x x-9
_x+l
x-8
x^2
+
x^7
~
x^l
+
x^~6
'
In this case it is best not to
multiply
at once
by
the L.C.M. of the
denominators of the
fractions;
much labour is often saved
by
a
judicious arrangement
and
grouping
of the terms.
By transposition
we have
x x+l x-9 s-8-
The first two terms
=
118 EQUATIONS
NOT INTEGRAL.
-2
tod the other terms
=
;
-
;
-
^-
.
(x-7)(x-6)
Hence the
equation
is
equivalent
to
2 2
Now
multiply by
the L. o. M. of the
denominators;
then
2
(x-7) (x- 6)
-2
(x-2) (x- 1)=0,
which reduces to
20x-80=0;
/. x=4.
Or thus:
The
equation
is
equivalent
to
x-2 x-7 ~x-l
222
that is
x-2 x-1 x-l x-6'
-
10
-
10
"(*-2)(x-7) (x-l)(x-6)'
from which we find as before that .r
-
4.
Ex. 4. Solve the
equation
:
x +
a x
+
b x
+
c
We have -^--1
+
^-1
+-^--1=0;
x+a x+b x+c
..
_
*_+_^_
=0.
x
+
a x
+
b
x+c
Hence- *=0
(i),
111
or else
H
jH
=0.
x+a x+b x+c
Multiply by
the
L.C.M.;
then
that is
EQUATIONS.
ONE UNKNOWN
QUANTITY.
119
the roots of which are
3
Thus there are three roots
given by (i)
and
(ii).
Ex. 5. Solve the
equation:
b+c
e
+
a a
+
b
_a
+
b +
e
bc-x ca-x ab-x~ x
The
equation
is
equivalent
to
b
+
c a c
+
a b a
+
b c
__
bc-x x
ca-x.
x ab-x x~
Taking
the terms in
pairs
we have
(a
+
b
f
c)x-abc (a
+
b
+
c)x-abc (a
+ b
+
c)
x- abc
_
n
x(bc-x) x(ca-x) x(ab-x)
Hence
(a
+
b
+
c)x-abc=0
1,
1 l
II. w*
7i \
x(bc- r) x(ca x) x(ab-x)
_, T
,
abc
From I. we have x= r .
a+b +
c
From II. we have on
multiplication by
the L. C.M.
(ca
-
x)
(ab
-x)
+
(ab
-
x) (be -x)
+
(be
-
x) (ca
-
x)
0,
that is 3x
2
-2x
(bc
+
ca+ab)
+
abc
(a
+
6
+
c)
=
0,
whence x=
J {be
+
ca
+
a&=fc
N/&-c
2
+ c
2
a
2
+
a
2
&
2
-
abc
(a+
b
+
c) }.
125. Irrational
Equations.
An irrational
equa-
tion is one in which
square
or other roots of
expressions
containing
the unknown
quantity
occur.
In order to rationalize an
equation
it is first written
with one of the irrational terms
standing by
itself on one
side of the
sign
of
equality:
both sides are then raised
to the lowest
power necessary
to rationalize the isolated
term
;
and the
process
is
repeated
as often as
may
be
necessary.
120
IRRATIONAL
EQUATIONS.
Ex. 1. Solve the
equation Jx
+ i
+
Jx
+
20
-
2
Jx
+
11
=
0,
We have
^x
+
4
+ x
+
20
=
Square
both members : then
which is
equivalent
to
Square
both members : then
whence z
=
5.
Ex. 2. Solve the
equation
Square
both members: then
2x
+
8
+
4(x+5)-4
.'. 3x
+
12
=
2
J2X
+
8
Jx+5.
Square
both members : then
9x
+ 72x -H44
=
4
(2x
+
8)
(x
+
5) ;
A *'
=
16,
whence a;
=
4 or x=-4.
Ex. 3. Solve the
equation *Jax
+
a.
+
Jbx
+
/3
+
v
/car +
7=0.
We have
Square
both members : then we have after
transposition
(a
+
b-c)x
+
a+p-y=
-2
Jax+a
Squaring
again,
we have
that is
+
2x
(aa
+
bfi
+
cy
-
by
-
c/S
-
co
-
ay
-
a/3
-
6a)
Thus the
given equation
is
equivalent
to a
quadratic equation.
It should be observed that it is
quite
immaterial what
sign
is
put
before a radical in the above
examples;
for there are two
square
roots of
every algebraical expression
and we have no
symbol
EQUATIONS.
ONE UNKNOWN
QUANTITY. 121
which
represents
one
only
to the exclusion of the
other;
so that
+
ijx
+
l and
-
Jx
+ 1 are alike
equivalent
to
,Jx
+
1
;
also
x+
Jx
+
1 has the same two values as
xijx
+
l.
126.
By squaring
both members of the rational
equa-
tion A
=
B,
we obtain the
equation
A
2
=
B*
;
and the
equation
A*
=
B',
or A* B*
=
0,
is not
only
satisfied when
A B
=
Q,
but also when A
+
B
=
0. Hence an
equation
is not in
general equivalent
to that obtained
by squaring
both its
members;
for the latter
equation
has the same
roots as the
original equation together
with other roots
which are not roots of the
original equation.
Additional
roots are not however
always
introduced
by squaring
both
sides of an irrational
equation.
For
example,
the
equation
as +
1
=
jx+
13 is
really
two
equations
since the radical
may
have either of two values
;
and
by squaring
both
members
we obtain the
equation
(x+
!)*
=
#+
13,
which
is
equivalent
to the two.
[See
Art.
152.]
127. A
quadratic equation
can
only
have two
roots.
We have
already proved
that an
expression
of the
nth
degree
in x cannot vanish for more than n values of
x,
unless it vanishes for all values of x. This shews that an
equation
of the nth
degree
cannot have more than n
roots,
and in
particular
that a
quadratic equation
cannot have
more than two roots.
The
following
is another
proof
that a
quadratic
equa-
tion can
only
have two roots.
We have to
prove
that aaf
+
bx
+
c cannot vanish for
a, ft, 7
three
unequal
values of x. That is we have to
prove
that
00*
+62 +c
=
....................
(i),
a/3
2
+
6/3
+
c
=
..................
(ii),
and
Off
+
by
+ c
=
...................
(lii),
cannot be
simultaneously
true,
unless
a,
b,
c are all zero.
122
A
QUADRATIC EQUATION
HAS ONLY TWO ROOTS.
From
(i)
and
(ii)
we have
by
subtraction
that is
(a-){a
But
a-=|=0;
hence
a(a
+
)
+
6
=
........... .- ......
(iv).
Similarly,
since
ft
7
4= 0,
we have from
(ii)
and
(iii)
a(/3+7)
+
6
=
..................
(v).
From
(iv)
and
(v)
we have
by
subtraction
a
(a 7)
=
.....................
(vi).
Now
(vi)
cannot be true unless a
=
0,
for a
7
0.
Also when a
=
0,
it follows from
(iv)
that 6
=
0,
and then
from
(i)
that c 0.
Thus the
quadratic equation
ao?
+
bx
+
c
=
cannot
have more than two
different
roots,
unless a
=
b
=
c
=
;
and when
a, 6,
c are all zero it is clear that the
equation
aa?
+
bx
+
c
=
will be satisfied for all values of x
t
that is
to
say
the
equation
is an
identity.
Ex. 1. Solve the
equation
a>
-*
-">
(a
-
6) (a
-
c) (6
-
c) (6
-
a)
The
equation
is
clearly
satisfied
by
x=a,
and also
by x=b;
hence
a,
6 are roots of the
equation,
and these are the
only
roots of the
quadratic equation. [The equation
is not an
identity,
for it is not
satisfied
by x=c.]
Ex. 2. Solve the
equation
(x-b)(x-c) (x-c)(x-a) (j-a)(j-6)
(a-b)(a-cy (ft-c)(6-o)
T
(e-a)(c-b)
The
equation
is satisfied
by x=a, by
x
=
b,
or
by
x=c.
Hence,
as it is
only
of the second
degree
in
x,
it must be an
identity.
Ex. 3. Solve the
equation
(x-b)(x-c)
,
(x-c)(x-a) (x-a) (x-b)
_
(a-b)(a-cy (6-c)(&-a)
r
(c-a)(c-b)~
EQUATIONS.
ONE UNKNOWN
QUANTITY.
123
The
equation
is satisfied
by x=a,
by z=6,
and
by
x=c
;
and
the
equation
is not an
identity,
since the coefficient of x
3
is not
zero.
Hence the roots of the cubic are
a, b,
c.
Ex. 4. Shew
that,
if
(a
-
a)
2
*
+
(a
-
)
2
y
+
(a
-
7)*z
=
(a
-
5)2,
(b
-
a)
2
x
+
(b
-
/S)
2
y
+
(b
-
y)*z
=
(b
-
6)
2
,
then will
where d has
any
value whatever.
The
equation
(X
-
a)
2
*
+
(X
-
19)
2
y
+
(X
-
y)*z
=
(X
-
5)
2
is a
quadratic equation
in
X,
and it has the three roots
a,
b,
c. It is
therefore
satisfied when
any
other
quantity
d is
put
for X.
128. Relations between the roots and the coeffi-
cients of a
quadratic equation.
If we
put
a and
/3
for the roots of the
equation
aa?
+
bx
+
c
=
0,
we have
y
=_A
2a
p
_
^
/^
2
""
^ac
By
addition
we have
(i).
u>
By multiplication
we have
i
a
6
2
-4ac c ....
(ll). 2
4a 4a
2
a
The formulae
(i)
and
(ii)
giving
the sum and the
product
of the roots of a
quadratic equation
in terms of the
coefficients are
very important.
124 RELATIONS BETWEEN ROOTS AND COEFFICIENTS.
129. Relations between the roots and the co-
efficients of
any equation.
By
the
following
method
relations between the roots and the coefficients of an
equation
of
any degree may
be obtained.
We have seen that if the
expression
of the nth
degree
in x
ax*
+
baT
1
+ c#
n~*
+
daT*
+
...
,
vanish for the n values x
=
a,
x
=
/3,
a;
=
7,
&c.,
then will
ax"
+
bx*'
1
+
caT*
+
dx"~*
+ ...
=
a
(a
-
a) (x
-
@) (x
-
7)...
We have therefore
only
to find* the continued
product
(x
o) (x /3) (a; 7)
...... and
equate
the coefficients of the
corresponding powers
of x on the two sides of the last
equation.
For
example,
if
a, /8,
7
be the roots of the cubic
equa-
tion ox
3
+
bo?
+
ex
+
d
=
0,
we have
cue
8
+
bx*
+
cx + d
=
a(x o)(x ft)(a;y)
=
a
{a?
-
(a
+
/3
+
7)
*
+
(Py
+
i*+a&)x
Hence,
equating
coefficients,
we have
a
C
7
a
+
a=
-,
N
d
It should be remarked that the sum of the roots of
any
equation
will be zero
provided
that the term one
degree
lower than the
highest
is absent*.
We
may
make use of the above to
prove
certain identical rela-
tions between three
quantities
whose sum is zero. For
a,
b,
c will
be the roots of the cubic x
3
+pr.
+
q
=
0,
provided
that a
+
b
+
c
=
0,
and
that
p
and
q satisfy
the relations
*
See Art. 437
EQUATIONS.
ONE UNKNOWN
QUANTITY.
125
bc
+
ca +
ab=p
............................................
(i),
abc
=
-
q
........................................
(ii).
Then,
since a + b +c
=
..........................................
(iii),
we have a?
+
V*+c?=(a
+ b +
c)*-2(be
+ca
+
ab)
=
-2p
.....................................
(iv).
Also,
since
a, b,
c are roots of
(v).
3
+pc
+
q
=
From
(v) by
addition
-3?
...........................
(vi).
Multiply
the
equations (v)
in order
by
a"~
3
,
b
n
~
3
,
c*~
3
,
and add
;
then
a
n
+ ft +
c
+
p (
a
-
2
+
fe-
+
c
1*-2
)
+
q (a
n
~
3
+
b
n~*
+
c"-
3
)
=
0.
Hence we have in succession
Hence also
=
2.
3 4
[See
also Art.
308,
Ex.
2.]
130.
Equations
with
given
roots.
Although
we
cannot in all cases find the roots of a
given equation,
it is
very easy
to solve the converse
problem, namely
the
problem
of
finding
an
equation
which has
given
roots.
For
example,
to find the
equation
whose roots are 4 and 5.
We want to find an
equation
which is satisfied when x
=4,
or
when x
=
5
;
that is when x
-
-i
0,
or when x
-
5
=
;
and in no other
cases. The
equation required
must be
(*-4)(*-5)
=
0,
that
is,
x*-9x +
-2U=0,
126
EQUATIONS
WITH GIVEN ROOTS.
for this is an
equation
which is a true statement when x-
4=0,
or
when x
-
5=0,
and in no other case*.
Again,
to find the
equation
whose roots are
2, 3,
and -4.
We have to find an
equation
which is satisfied when x
-
2=0,
or
when x
-
3
=
0,
or when x
+4
=
0,
and in no other case. The
equation
must therefore be
(x
-
2)
(x
-
3) (x
+
4)
=
0,
that is of -a?- 14a;
+ 24=0.
Ex. 1. If
o,
/3
are the roots of the
equation
ax*
+
bx
+
c=Q,
find the
equation
whose roots are and
-
.
/3
a
The
required equation
is
B)HH
thatis a?-*
^-41
=
0.
a
P
Now,
by
Art.
128,
we have
o+0=--, /3=-;
\a
8
a/
a ac
a/3
Hence the
required equation
is
J9
fr-aa*
Ex. 2. If
a,
/3, y
be the roots of the
equation
find the
equation
whose roots are
(iy, ya, a/3.
The
required equation
is
that is
*
The
equation
x*
-
9x +
20
=
is
certainly
an
equation
with the
proposed
and with no other roots
;
but to
prove
that it is the
only equa-
tion with the
proposed
and with no other
roots,
it must be assumed that
every equation
hat a root.
If,
for
example,
the
equation
x* + 7x* -2
= had no
roots,
then
(x
-
4
)
(x
-
5) (x
6
+ lx-
-
2)
=
would also be an
equation
with the
proposed
roots and with no others.
The
proposition
that
every equation
has a root is
by
no means
easy
to
prove;
the
proof
is
given
in works on the
Theory
of
Equations.
EQUATIONS.
ONE UNKNOWN
QUANTITY.
127
Now,
by
Art.
129,
we have
and
a/37=--.
Hence the
required equation
is
db d*
--
-^
a
2
a
2
131.
Changes
in value of a trinomial
expression.
The
expression
ax* + bx
+
c will alter in value as the value
of x is
changed
; but,
by giving
to x
any
real value
between oo and
+
oo
,
we cannot make the
expression
ax
9
+
bx
+
c assume
any
value we
please.
We can find the
possible
values of ax*
+ bx
+
c,
for
real values of
x,
as follows.
In order that the
expression
ax*
+
bx
+
c
may
be
equal
to X for some real value of
x,
it is
necessary
and sufficient
that the roots of the
equation
ax*
+
bx
+
c=\
be
real,
the condition for which is
6
2
_
4a
(
c
_
x)
>
o,
that is b* 4ac +
4aX>0 ..................
(i).
I. If 6* 4ac be
positive,
the condition
(i)
is satisfied
for all
positive
values of
4a\,
and also for all
negative
values of 4>a\ which are not
greater
than 6
2
4ac.
Thus,
when 6
2
4ac is
positive,
ax*
+
bx
+
c
can,
by
giving
a suitable value to
x,
be made
equal
to
any quantity
of the same
sign
as
a,
or to
any quantity
not
absolutely
u
^~
4*CLC
greater
than
-j
-
and whose
sign
is
opposite
to that of a.
II. If b* 4>ac be
negative,
the condition
(i)
can
only
be satisfied when 4aX is
positive
and not less than
4ac 6*.
128
CHANGES IN VALUE OF A TRINOMIAL EXPRESSION.
Thus,
when 6
8
4oc is
negative,
ax
9
+
bx
+ c must al-
ways
have the same
sign
as
a,
and its absolute
magnitude
.
,
4oc
-
6
s
can never be less than
--
.
4a
III. If 6* 4ac be
zero,
the condition
(i)
is satisfied
for all
positive
values of a\,
It follows from the above that the
expression
o.c*+ bx+c
will
keep
its
sign unchanged,
whatever real value be
given
to
as,
provided
that b* 4ac be
negative
or
zero,
that is
provided
that the roots of the
equation
ax*
+
bx
+ c
=
be
imaginary
or
equal,
and also that the
expression
can be
made to
change
its
sign
when the roots of ax*
+
bx
+
c
=
are real and
unequal.
We
give
another
proof
of this
proposition.
If the
equation
ox*
+
bx + c
=
have real
roots, a, /8
suppose,
then ox*
+
bx
+
c
=
a
(x
a) (x
-
$).
Now
(x ct)(x ft)
is
positive
when x has
any
real
value
greater
than both a and
/?,
or less than both a and
8;
but
(x CL) (x
ft)
is
negative
when x has
any
real
value intermediate to a and
ft.
Thus for real values of x the
expression
ox*
+ bx+c
has
always
the same
sign
as a
except
for values of x which
lie between the roots of the
corresponding equation
ax*
+
bx
4-
c
=
0.
132. We can also
prove
that the
expression
ax*+bx +c
will or will not
change sign
for different values of x accord-
ing
as 6* 4oc is
positive
or
negative,
as
follows.
I. Let 6
J
4ac be
positive.
The whole
expression
within
square
brackets will
clearly
be
negative
when x x-
; also,
when x is
very
EQUATIONS.
ONE UNKNOWN
QUANTITY.
129
. / b\* ... , &
2
-4oc
great,
I x
+
^-
1 will be
greater
than
-7-5
an(i there-
\
_''
4ft
fore the whole
expression
within
square
brackets will be
positive.
Thus when 6
s
4oc is
positive
the
expression
aa?+bx+c
can be made to
change
its
sign by giving
suitable real
values to x.
II. Let b* 4ac be
negative (or zero).
Since f x
+
=-
I is
positive
for all real values of
x,
and
j-j
is also
positive (or zero),
the whole
expression
within
square
brackets must be
always positive.
Thus when 6* 4oc is
negative
or
zero,
the
expression
ax*
+
bx
+
c will
always
have the same
sign
as a.
133. It follows from Article 131 or 132 that if an
expression
of the second
degree
in x can be made to
change
its
sign by giving
real values to
x,
then must the
roots of the
corresponding equation
be real.
Consider,
for
example,
the
expression
where the
quantities
are all
real,
and
a, @,
7
are
supposed
to be in order of
magnitude.
The
expression
is
clearly
positive
if x
=
a,
and is
negative
if x
{$.
Hence the
expression
can be made to
change
its
sign,
and therefore
the roots of the
equation
a
2
(x -) (x
-
7)
+
V(x -7) (a
-
a)
+
c
2
<>
-
a) (x
-
ft)
=
are real for all real values of
a, b, c, a, /3,
7.
Ex. 1. Shew that
(x
-
1) (x
-
3) (x
-
4) (x
-
6)
+
10 is
positive
for all
real values of x.
Taking
the first and last factors
together,
and also the other
two,
the
given expression
becomes
(x
n
-
-
7* +
6) (x
2
-
7x +
12)
+ 10
!
-7z)
+
82
which is
clearly always positive
for real values of x.
S.A.
130 EXAMPLES.
Ex. 2. Shew
that, by giving
an
appropriate
real value to
x,
4x2+36x
+
9
,
. .
r^j-j
---
can be made to assume
any
real value.
then
a:*(4-12X)
+
(36-8X)a;
+ 9-X=0.
Now in order that x
may
be real it is
necessary
and sufficient
that
(36
-
8X)
2
-
4
(4
-
12X) (9
-
X)
>
0,
or that X*-8X +
72>0,
or
(X-4)+$6>0,
which is
clearly
true for all real values of X. Thus we can find
real values of x
corresponding
to
any
real value whatever of X.
Ex. 3. Shew that
-5 s
. can never be
greater
than 7 nor less
than
-
for real values of x.
a
then *
2
(l-X)-3x(l
+
X)
+
4(l-X)
=
0.
In order that x
may
be real it is
necessary
and sufficient that
thatis -7X
8
+
50X-7>0,
or
-(7X-1)(X-7)>0.
Hence 7X
-
1 and X
-
7 must be of
different signs,
and therefore
X must lie between
-
and
7,
which
proves
the
proposition.
EXAMPLES
X.
Solve the
following equations:
2.
(c
+ a
-
26)
a? +
(a
+ b
-
2c)x
+
(b
+ c
-
2a)
=
0.
q
<' y
(x-a)'~ (x
+
b)*'
. a+x
b+x_
*
.
-
T
~
V>
6 + cc a-t- a;
5.
6.
7.
EQUATIONS.
ONE UNKNOWN
QUANTITY.
131
ax + b cx+ d
a + bx c +
dx'
a x
_l-bx
1 ax b x
'
3x-4 7
i^ ** T^ *Jt*/
Y
x+
1 x + 1
x
s
x
9
5s- 4
+
x
T
^T~x+l
+
"x
T
^T'
1111
9. + ~ +
f>+ 5
=
0.
x-8 x- 6 x + b x+ o
2534
10 H
=
+ .
x+ 8 x+ 9
x+ 15 x + 6
n
2 1 6
2x-3 x-2
._ x-a x-6 x-c
12.
r
+ + -=3.
x- b x c xa
13.
^
+
^
+
^f
=
3.
ax o-x c x
-. . **^-i~vw */--*^ CT~C
n
14.
-
+
-
+ =3.
x a x o x c
2x-l 3x-l x-7
15. +
^=4+
-.
x+l x+2 x-l
.. x 2 x 3
ID.
-^
+
-
="x
+
-
.
1r,
x + a x a x+b x b
17. + +
T+-T
x a x + a x b x + b
, x 1 x 4 x 2 x-3
I
H
j
i
AW.
x+l x+4 x+2 x+3
1 1
19.
1 1
x+an
5
x a +
x+b x
-
b
92
132
EXAMPLES.
20- +
*T
+-^-=0.
oa
-
x 36
-
x 3c x
x+b x+c x+b+c
a + c 6-i-c a + 6 + 2c
ff
~
^
-
^
x+'2b x + 2a x + a+ b
x-a
2(a-b)
-- ---
;
--
.
x a x o x a b
(x
+
a) (x
+
6)
=
(g
+
c) (a;
+
c?)
a+a+6
x+c+d
EQUATIONS.
ONE UNKNOWN
QUANTITY.
133
37.
Ja
+ x +
Jb
+ x
=
Ja
+ b + 2x.
38.
Ja
x +
Jb
+ x
=
J'2a
+ 26.
39.
J(a
+
x)(x
+
b)
+
J(a-x)(x~-
b)
=
2
Jax.
40.
Ja (a
+ b +
x) Ja (a
+ b
-
x)
=
x.
41.
Jx
2
+ax + b
s
-Jx*-ax
+'b*
=
2a.
42.
Jx
1
+ ax +
a
a
+
Jg^ "V r^f
43.
Jax
-b +
Jcx
+ b
=
Jax+'b
+
Jcx-b.
44.
Jx (a
+ b
-
x)
+
Ja (b
+ x
-
a)
+
Jb (a
+ x
-
b)
=
0.
45.
Jx
+ a +
Jx
+ b +
Jx
+ c
=
0.
46.
Jab (a
+ b +
x)
=
Ja (a
+
b)(b- x)
+
Jb (a
+
b) (a,
-
x).
47.
Jx
3
48. a
49.
^or
what values of x is
^14
-
(3x- 2) (x- 1)
real.
50. Shew that
5
--
=-
-
can have no real value between
ar+ 2x 7
5 and 9.
x
2
6x + 5
51. Shew thaK if x be real
-;
--
r- can never be less
x
8
+ 2x + 1
than
/*
_ O*
I 1
52. What values are
possible
for
-5
--
. x beinsr
real.
x
a
+ x + 1
'
53. Find the
greatest
and least real values of x and
y
which
satisfy
the
equation
x
3
+
y
a
=
6x
8y.
54. Find the
greatest
and least real values of x and
y
when
55. When x and
y
are taken so as to
satisfy
the
equation
(x
2
+
2/*)"
=
2a
2
(x'
-
y
2
),
find the
greatest possible
value of
y.
134 EQUATIONS
OF HIGHER DEGREE THAN THE SECOND.
56. Shew that if the roots of the
equation
x
1
(b'
+
b")
+
2x(ab
+
a'b')
+ a' + a"
=
be
real, they
will be
equal.
57. If the roots of the
equation
ay? + bx + c
=
be in the
ratio m :
n,
then will mnb*
=
(m
+
n)*
ac.
58. If oaf + 2bx + c
=
and a'x* + 2b'x + c'
=
have one and
only
one root in
common, prove
that b* ac and b'* a'c' must
both be
perfect squares.
59. If
a;,,
x be the roots of the
equation
ax* + bx + c
=
0,
x
9
x'
find the
equation
whose roots are
(i)
x^
and
xf, (ii)
and
-'
x
t
x
\
(iii)
b + ax
l
and b + ax
a
.
60. If x
l
,
x
a
be the roots of ax* + bx + c
=
0,
find in terms
of
a, b,
c the values of
x*
(bx
a
+
c)+
x*
(bxl
+
c),
and x'
(bxa
+
c)'
+ x*
(bxt
+
c)
9
.
61. Shew
that,
if x
lt
x
t
be the roots of X
s
+ mx + m* + a
=
0,
then will x* + x + x* + a
=
0.
O
62. If
x,
,
JT,
be the roots of
(x*
+
1) (a*
+
1)
=
max
(ax
-
1),
then will
(x*
+
1) (a;,*
+
1)
=
rnx^ (xt
x
t
-
1).
63. If
,,
x
t
be the roots of the
equation
A
(x*
+
m*)
+ Amx + JBm'x
2
=
0,
then will A
(x'
+
a;/)
+
^laj.a;,
+ Bx'x*
=
0.
64. Prove
that,
if * be
real,
2
(a x) (x
+
Jx
3
+ 6
2
)
cannot
exceed a* + 6*.
r
2x
4
-4^ + 9x'
65. Find the least
possible
value of
-
^-5
(
x +1
/
for real values of ax
EQUATIONS.
ONE UNKNOWN
QUANTITY.
135
Equations
of
higher degree
than the second.
134. We now consider some
special
forms of
equations
of
higher degree
than the
second,
the solution of the most
general
forms of such
equations being beyond
our
range.
135.
Equations
of the same form as
quadratic
equations.
The
equation
ax*
+
bx*
+
c
=
can be solved in
exactly
the same
way
as the
quadratic
equation
ax*
+
bx
+
c
=
;
we therefore have
b Jtf 4ac
x
=
+
2a
~
2a
/(
b
, N/6
2
-4acl
Hence #
=
+
/
J
+
^ -
>
.
-
V
I
2a
-
2a
J
Thus there are
four
real or
imaginary
roots.
Similarly,
whenever an
equation only
contains the
unknown
quantity
in two terms one of which is the
square
of the
other,
the
equation
can be reduced to two
alternative
equations
:
for,
whatever P
may
be,
is
equivalent
to P
=
- -
+
~
2a
~
2a
Ex.1. To solve x
4
-
lOx
2
+ 9
=
0.
We have
(x
8
-
9) (x
2
-
1)
=
;
.. x
2
=9,
giving
x=
3;
or else x
2
=l, giving
x= =fc 1.
Thus there are four
roots, namely +1,
-
1, +3,
-
3.
136 EQUATIONS
OF HIGHER DEGREE THAN THE SECOND.
Ex. 2. To solve
(x
a
+
x)
2
+
4
(x
8
+
x)
-
12
=
0.
The
equation may
be written
(x-
+
x
+
6) (x
5
+ x
-
2)
=
0.
Hence x*
+
z+6=0,
or x
2
+ x-2=0.
1
1
_
Therootsof x
3
+
x + 6
=
are
-
~--23.
Therootsof x
+
ar-2=0 are 1 and -2.
Thus the roots are
1, -2,
--
5N
/-23.
1.
Ex.3.
(x
+
The
equation
is
equivalent
to
Therootsof x
3
+
3x + 2=0 are
-
1 and -2.
Therootsof x
2
+
5x
+
2
=
are
-
.
' '
Thus the
equation
has the four roots
-
1,
-
2,
- -
o
Ex.4. To solve o
Put
then
whence we obtain two values of
y,
a. and
/3 suppose.
We then have ax
5
+
bx -f c
=
a
1
,
or
and the four roots of the last two
quadratic equations
are the roots
required.
Ex.5. To solve 2'-4a;+3
>/x
1
-2x
+ 6
=
15.
The
equation may
be written
Put
y=,J(3?
-
2x
+
6)
;
then we have
2y
J
+
3y
-
27
=0,
a
whence
y=8,
or
y=--.
Hence z
s
-2x+6
=
9, giving
x=3 or
-1;
ftl 1
orelse
x*-2z+6=-j-,
giving
x=l,
Thus the roots are 8
;
-
1
;
1
5
I
EQUATIONS.
ONE UNKNOWN
QUANTITY 137
g
Ex. 6. To solve
(x
+
o) (x
+
2a) (x
+
3a) (x
+
4a)
=
^
a
4
.
Taking together
the first and last of the factors on the
left,
and
also the second and
third,
the
equation
becomes of the form we are
now
considering.
We have
(x
2
+ 5ax
+
4a
2
) (x
s
+
Sax
+ 6a
2
)
=
^
a
4
.
Hence
(*
+
Sax)
8
+
10a
a
(x
2
+
Sax)
+24a
4
=
^
a
4
,
25
,
15
.. x
2
+
5ax=
-
a
2
,
or else x*+5ax=
--j-a
2
.
4 4
Hence x
+
s
a=0,
or
x+
a=
Sx
/10.
fi & i
mi it.
S 5 a
., rt
Thus the roots are
~o
~9
a
9
*l
*
136.
Reciprocal Equations.
A
reciprocal equa-
tion is one in which the coefficients are the same whether
read in order backwards or forwards
;
or in which all the
coefficients when read in order backwards differ in
sign
from the coefficients read in order forwards.
Thus
ax
3
-f bx* + bx
+
a
=
0,
ax
4
+
bx
3
+ ex* + bx + a
=
0,
and ax
s
+
bx
4
+
ex* ex* bx a
=
are
reciprocal equations.
[See
also Art.
4-42.]
Ex.1. To solve
We have a
(x
3
+
l)
+
fcx(x
+
l)
=
0,
that is
(x
+
l){a(x
2
-.-c
Hence x=
-
1,
or else
Ex.2. To solve
Divide
by
x
2
;
then we have
Now
put
x
+
-
=
y;
X
then x
2
+-5=t/
2
-2.
x-
'
Hence a
138
ROOTS FOUND BY INSPECTION.
Let the two roots of the
quadratic
in
y
be a and
ft;
then the
roots of the
original equation
will be the four roots of the two
equations
x+-=a and x
+
-=B.
x x
Ex.3. To solve ox
5
+&x
4
+
cx
3
-cx
>
-6x-a=0.
We have
a(x* -l)
+
bx(x*-l)
+cx*
(x-l)
=
0,
that is
(x-l){a(x*
+
x*
+
x*
+
x
+
l)
+
bx(x*
Hence
x=l,
or else
The last
equation
is a
reciprocal equation
of the fourth
degree
and is solved as in Ex. 2.
137. Roots found
by inspection.
When one root
of an
equation
can be found
by inspection,
the
degree
of
the
equation
can be lowered
by
means of the theorem of
Art. 88.
Ex. 1. Solve the
equation
x(x-l)(x-2)
=
a(o-l)(a-2).
One root of the
equation
is
clearly
a. Hence x
-
a is a factor of
x(x-l)(x-2)-a(a-l)(o-2),
and it will be found that
x(x-l)(x-2)-a(a-l)(o-2)
=
(ar-a){x
a
-(3-a)x
+
(a-l)(o-2)}.
Hence one root of the
equation
is
a,
and the others are
given by
x
2
-
(3
-
a)
x
+
(a
-
1) (a
-
2)
=0.
Ex. 2. Solve the
equation
Here we have to
try
to
guess
a root of the
equation,
and in order
to do this we take
advantage
of the
following principle
:
If x
=
3
be a root of the
equation
ox"
+
bx"~
l
+ . . . +
k
=
0,
where
P
a, b,...k
are
integers
and
-
is in its lowest
terms,
then a will be a
P
factor of k and
/3
a factor of a. As a
particular case,
if there are
any
rational roots of x*+... +
=
0,
they
will be of the form
x=a,
where a is a factor of k.
In the
example
before us the
only possible
rational roots are
1,
2, 3,
and 6. It will be found that x=2 satisfies the
equation,
and we have
Hence the other roots of the
equation
are
given by
x* + 4x-3=0,
and are therefore
-
2
^7.
EQUATIONS.
ONE UNKNOWN
QUANTITY.
139
Ex. 3. Solve
Since x=a and x
=
b both
satisfy
the
equation, (a: -a) (x-b)
will
divide
(a
-
x)*
+
(x
-
b)*
-
(a
-
6)
4
,
and as the
quotient
will be of the
second
degree,
the
equation
formed
by equating
it to zero can be
solved.
We
may
however
proceed
as follows. The
equation may
be written
Thus the
required
roots are
a,
b and the roots of the
quadratic
ar
-
x
(a
+
b)
+
2a?
-
3a6
+
2b*
=
0.
Ex. 4. Solve the
equation
(x-b)(x-c) (x-c)(x-a) t (x-a)(x-b)_
(a-b)(a-cy (b-c)(b-a)
+C
(c-a)(c-b)~
The
equation
is
clearly
satisfied
by
x
=
a,
by
x
=b,
and
by
x =c.
Also,
since the coefficient of a? is
zero,
the sum of the roots is zero.
[Art. 129.]
Hence the
remaining
root must be
-
a
-
b
-
c.
Thus the roots are
a, b, c,
-
(a
+
6
+
c).
138.
Binomial
Equations.
The
general
form of
a binomial
equation
is x
m
k
=
0.
The
following
are some of the cases of binomial
equations
which can be solved
by
methods
already given
for the
general
case De Moivre's theorem in
Trigonometry
must be
employed.
Ex.1. To solve a? -1
=
0.
Since x
3 -
l
=
(x-l)(x'-
+
x
+
l),
we have x-
1=0;
or else x
2
+ 05
+ 1=0,
the roots of which are
Hence there are three roots of the
equation
x
3
=l
;
that is there
aro three cube roots of
unity,
and these roots are
140
CUBE BOOTS OF UNITY.
Ex.2. To solve x*-l=Q.
Since x*
-
1
=
(a;
-
1) (x
+
1) (x
+
J^l) (x-J~l),
the
four fourth
roots of
unity
are
1,
-1,
,/Tl
and
-J^l.
Ex. 3. To solve X
s
-1=0.
x
Hence x
=
1
;
or else x*+x
3
+x*+x +
l
=
Q.
The latter
equation
is a
reciprocal equation.
Divide
by
x-
t
and
we have
Put
then
_
Hence
that is
Hence
or
or
Ex.4. Tb solve
Hence
139. Cube roots of
unity.
In the
preceding
article
we found that the three cube roots of
unity
are
i,
M-i+y-3), H-W-3).
An
imaginary
cube root of
unity
is
generally repre-
sented
by
o>; or,
when it is
necessary
to
distinguish
EQUATIONS.
ONE UNKNOWN
QUANTITY.
141
between the two
imaginary
roots,
one is called
&>,,
and the
other
&>,,
so that
1,
G>,
and &>
2
are the three roots of the
equation
a? 1
=
0.
Taking
the above
values,
we have
also a>
l0
>,
=
i(-
1
+
s)(-
1
-
3)
=
1.
These relations follow at once from Art. 129
;
for the
sum of the three roots of .-c
8
1
=
is
zero,
and the
product
is 1.
Again ,
=
i(-
1
+
J=S?
=
K~
1
~
v/
1
^)
=
and
,
=
i(-
1
-
N/^)
2
=
H-
1
+
7^3)
=
i,
so that
6),
2
=
ti)
2
and co*
=
(0
1
. These relations follow at
once from
WjtOjj
=
1 and m*
=
to
s
8
=
1.
Thus
ty
we
square
either
of
the
imaginary
cube roots
of
unity
we obtain the other.
Hence if &> be either of the
imaginary
cube roots of
unity,
the three roots are
1,
&) and o>
8
.
We know that
Hence a
+ b + c is a factor of a
3
+ 5
3
+
c
8
-
3abc,
and this is the case
for all values of
a, &,
c.
Hence a +
(w6)
+
(w
2
c)
is a factor of a
3
+
(w6)
8
+
(o^c)
3
-
3a
(ub) (a^c),
that is of a
s
+
&
3
+
c
3
-3a&c;
and a
+
0^6
+
we can
similarly
be shewn
to be a factor.
Hence
EXAMPLES XI.
Solve the
following examples
:
1.
x
4
-
2^-8
=
0.
2.
142
EXAMPLES.
3. x'
-
7aV
-
8a"
=
0.
x x
a
+ I
_
5
i?Tl
+
~^~~2'
x
3
+ 2 a;* + 4s + 1
_
5
5.
3 -A r
+
. o 77
6.
(a?
+ a;
+!)(;'
+ * +
2)
=
12.
7.
(cc
8
+ 7a; +
5)
2
-3x*-21a;=1
8.
N--=-
T
.
4
9. 6
10.
(a
11.
(a: +!)(
+
2) (x
+
3) (a;
+
4)
=
24.
12.
(a?
+
a)(x
+
3a) (ar
+
5a) (x
+
la)
=
384a
4
.
13.
(*
-
3a) ( -)(+
2a) (as
+
4o)
=
2376
4
.
14.
(x +2)(x
+
3) (x
+
8) (x
+
12)
=
4x
2
.
15. 2a*
-
3a;
-
21
=
2x
Jx*
-
3x + 4.
16.
x
4
-
2
(a
+
b)
x
2
+ a
2
+ 2ab + b*
=
0.
17. x
4
-
2*V
-
2z
2
&* + a
4
+ 6
4
-
2a
s
b
2
=
0.
18. 4a;
4
-4a;
3
-7^-4a; + 4
=
0.
19. 9a
4
-
24a^
-
2x*
-
2lx + 9
=
0.
20. a
8
* 1=0. 21. 0^-1
=
0.
22. 3x
8
-14a^ + 20a;-8
=
0.
23. x
4
-15a:
!
+10a; + 24
=
0.
24. x
4
+ 7x
3
-
1x
-
1
=
0.
25.
(a; -a)
8
(b
-
c)
8
+
(x
-
b)' (c-a)
8
+
(a:- c)
8
(a-6)'
=
0.
26.
x(x -l)(x- 2)
=
9. 8. 7.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
42.
43.
44.
45.
EQUATIONS.
ONE UNKNOWN
QUANTITY. 143
x
(x
-
1) (
-
2) (x
-
3)
=
9 . 8 . 7 . 6.
a
-
x 4- 6 a
=
a + b 2x.
(a -)
+
(* -6)*
=
(a -ft)
8
.
i/cT-^x
+
yx
-b
=
!ja-b.
$a
-
x +
*Jx
b
-
*Ja
b.
(a;
+ a +
by
-
(ax
+ bx +
ab)
a
=
0.
obcx
(x
+ a + 6 +
c)* (a$c
+ seca + xab +
abc)
9
=
0.
b-2x)* (a
+
bf -oc^i.-v,
x
4
+ b (a +
b)
x
3
+
(ab
-
2)
6
2
af
-
(a
+
b)
b'x + b
4
=
0.\^ y ''- -
(cc
+ 6 +
c)
(x + c +
a)
(x + a +
b)
+ abc
=
0.
=
6 + c x c + a as a + o a;
(g-a)
a
(x
-
b)* (x
-
c)
g
(x-a)
2
-(6-c)
2
+
(aj-6)'-(c-a)
f
+
(x-c)*-(a-b)*
~
(x
+
d) (x
+
b) (x
a)(x-V)_(x
+
c) (x +.
d)
(x -a) (x b) (x
+
a) (x
+
b) (x -c) (x d)
+
c) (
CHAPTER X.
SIMULTANEOUS
EQUATIONS.
140. A SINGLE
equation
which contains two or more
unknown
quantities
can be satisfied
by
an indefinite
number of values of the unknown
quantities.
For we can
give any
values whatever to all but one of the unknown
quantities,
and we shall then have an
equation
to deter-
mine the
remaining
unknown
quantity.
If there are two
-equations containing
two unknown
quantities (or
as
many equations
as there are unknown
quantities),
each
equation
taken
by
itself can be satisfied
in an indefinite number of
ways,
but this is not the case
when both
(or all)
the
equations
are to be satisfied
by
the
same values of the unknown
quantities.
Two or more
equations
which are to be satisfied
by
the
same values of the unknown
quantities
contained in them
are called a
system
of simultaneous
equations.
The
degree
of an
equation
which contains the unknown
quantities
x,
y,
z... is the
degree
of that term which is of
the
highest
dimensions in
x,
y,
z
Thus the
equations
ax
+
a*y
+
a*z
=
a*,
are of the
first,
second and third
degrees respectively.
SIMULTANEOUS
EQUATIONS
OP THE FIRST DEGREE.
145
141.
Equations
of the First
Degree.
We
proceed
to
consider
equations
of the first
degree, beginning
with
those which contain
only
two unknown
quantities,
x and
y.
Every equation
of the first
degree
in
x,
y,
z,...
can
by
transformation be reduced to the form
ax
+
by
4- cz
+
. . .
=
k,
where
a,b,c,...k
are
supposed
to
represent
known
quan-
tities.
NOTE. When there are several
equations
of the same
type
it is convenient and usual to
employ
the same letters
in
all,
but with marks of distinction for the different
equations.
Thus we use
a, b,
c... for one
equation;
a',
b'
,
c ... for
a second
;
a", 6",
c"... for a third
;
and so on. Or we use
a
lt 6j, c,
for one
equation
;
2
,
6
a
,
c
a
for a second
;
and so
on.
Hence two
equations containing
x and
y
are in their
most
general
forms
ax+by=
c,
and a'x
+
b'y
=
c,
and
similarly
in other cases.
142.
Equations
with two unknown
quantities.
Suppose
that we have the two
equations
ax+by
=
c,
and a'x
+
b'y
=
c'.
Multiply
both members of the first
equation by
&',
the
coefficient of
y
in the second
;
and
multiply
both members
of the second
equation by
b,
the coefficient of
y
in the
first. We thus obtain the
equivalent system
aVx
+
bb'y
=
cb',
a'bx +
bb'y
=
c'b.
S.A. 10
146
SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGliEE.
Hence,
by
subtraction,
we have
(aV
-
a'b)
x
=
cV- c'b
;
cb' c'b
whence
ab' a'b
*
Substitute this value of x in the first of the
given
equations;
then
cb'
-
c'b ,
a
r, 77
+
by
=
c,
ab ab
,
_
c
(ab' a'b) a(cb c'b)
^
~
ab'
-
afb~
~
' *
, ac' a'c
whence v
=
n 77
.
y
ab- ab
The value of
y may
be found
independently
of x
by
multiplying
the first
equation by
a and the second
by
a
;
we thus obtain the
equivalent system
a'ax
+
a'by
=
a'c,
a ax
+
ab'y
=
ac'.
Hence,
by
subtraction,
we have
(a'b ab') y
=
a'c ac'
;
_
a'c ac'
''
y
~a'b^ab"
which is
equal
to the value of
y
obtained
by
substitution.
NOTE. It is
important
to notice that when the value
either of a; or of
y
is
obtained,
the value of the other can
be written down.
For a and a' have the same relation to a; that b and V
have to
y\
we
may
therefore
change
x into
y provided
that we at the same time
change
a into
b,
b into
a,
a into
b',
and b' into a'. Thus from
cb' c'b ca c'a
as
=
n
TL
we have
y= j, p-
.
ab ab
y
ba
-
b a
SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGREE.
147
It will be seen from the above that in order to
solve two simultaneous
equations
of the first
degree,
we
first deduce from the
given equations
a third
equation
which contains
only
one of the unknown
quantities
;
and
the unknown
quantity
which is absent is said to have
been eliminated.
143. From the last article it will be seen that the
values of x and
y
which
satisfy
the
equations
ax
+
by
=
c,
and
4
a'x +
b'y
=
c',
can be
expressed
in the form
be b'c ca,' c'a ab' a'b
'
So
also,
from the
equations
ax 4-
by
+ c
=
0,
and a'x -f
b'y
+ c
=
0,
x
y
1
U7O M O \TC* ^ __ ___ _
we juctvt?
-77 =7- , ,
-.
, ,,
be be ca ca abab
It is
important
that the student should be able to
quote
these formulae.
Ex. 1. Solve the
equations
and
We have
X
^
-
2
.27-3 . 13
"
1377-27 . 3~ 3. 3-7 .2'
x
y
1
that13
I5
=
T5
=
5
;
15
and
y
= =
2.
102
148
SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGREE.
Ex. 2. Solve the
equations
*+=.
x
y
. 2 5
and
---
=
-
7.
x
y
These
may
be considered as two simultaneous
equations
of the
first degree
with
-
and
-
as unknown
quantities.
We therefore have
1 1
_
x
___
y
-1
8
(
-
7)
-
(
-
5)
a
*
2.2-(-7)4
=
~
1 11 2f>
x=-26'
or
*=-fi-
1 32 13
Ex. 3. Solve the
equations
x
-
y
=
a
-
b,
ax
-
by
=2
(a
2
-6
2
).
We have
x
y
-1
-
2
(a*-
6
2
)
+
6
(a- b)
~
a
(a
-
b)
-
2
(a
2
-
ir)
~
-
b
+
a'
that is
*
!T _
1
.
.'. x-
b-a
OM _ , _
and
i/
b-a
Instead of
referring
to the
general
formulae of Art.
143,
as we have
done in the above
examples,
the unknown
quantities may
be elimi-
nated in
turn,
as in Art. 142
;
and this latter method is
frequently
the
simpler
of the two. Thus in this last
example
we have at
once,
by
multiplying
the first
equation by
a and then
subtracting
the
second,
SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGREE. 149
Then x
=
(a
+
2b)+a
-
b\
144. Discussion of solution of two simultaneous
equations
of the first
degree.
We have seen that the
values of a; and
y
which
satisfy
the
equations
ax
+
by
=
c........................
(i),
and a'x +
b'y
=
c .....................
(ii),
are
given by
(db
r
a'b}
x
=
cb' c'b ...............
(iii),
(ba b'a) y
=
ca' c'a ...............
(iv).
Thus there is a
single
finite value of
x,
and a
single
finite value of
y, provided
that ab' a'b
4=
0.
If ab' ab
=
Q,x
will be
infinite [see
Art.
118]
unless
cb' c'&
=
; and,
if ab' a'b and cb' c'b are both
zero,
any
value of x will
satisfy equation (iii).
So
also,
y
will be
infinite
if ab' a'b
=
0,
unless ca'c'a
is also
zero,
in which case
any
value of
y
will
satisfy
equation (iv).
If ab' a'b
=
0,
then
=
r, ;
and if ab' a'o
=
and
a b
,, ,, ,,
a b c
also cb c b
=
0,
then
=
77
=
~/
a b c
When
equations
cannot be satisfied
by finite
values of
the unknown
quantities, they
are often said to be incon-
sistent. Thus the
equations
ax
+
by=c
and a'x
+
b'y
=
c'
are inconsistent if
-,
=
77 ,
unless each fraction is
equal
to
a b
f\
f ,
in which case the
equations
are indeterminate. In fact
when
=
77
=
it is clear that
by multiplying
the terms
of
equation (i) by
we shall obtain
equation (ii),
so that
CL
the two
given equations
are
equivalent
to one
only.
150 SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGREE.
We have hitherto
supposed
that
a, a', b,
b' were none
of them zero. It will not be
necessary
to discuss
every
possible
case :
consider,
for
example,
the case in which a
and a' are both zero.
4\
When a and a are both
zero,
we have from
(i) y
=
r ,
/
and from
(ii)
y-j7>
These results are inconsistent with
one another unless
r
=
r
.
.
b b
c c'
Hence,
if a
=
a'
=
0,
and
r
=
T/ ,
the
equations (i)
and
s*
(ii)
are satisfied
by making y
=
r
,
and
by giving
to #
any
finite value whatever.
c c'
If however
r =t= r/>
the
equations by
=
c and
b'y
=
c'
cannot both be
satisfied,
unless
they
are looked
upon
as the
limiting
forms of the
equations ax+by=c
and
a'x+b'y=c',
in which a and a' are
indefinitely
small and
ultimately
zero. But from
(iii)
we see that when a and a' diminish
without
limit,
as must increase without
limit,
cb' c'b not
being
zero.
Thus,
in the
equations (i)
and
(ii),
when a
and a' diminish without
limit,
and cb'
c'b,
the value of x
must be infinite.
Equations
with three unknown
quantities.
145. To solve the three
equations
:
d ...............
(i),
a'x +
b'y
+ c'z
=
df ...............
(ii),
a"x
+
b"y
+
c"z
=
d"...............
(iii).
Method of successive elimination.
Multiply
the
first
equation by
c',
and the second
by
c
;
then we have
ac'as
+
bey
+
cc'z
=
dc',
SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGREE.
151
and a ex +
b'cy
+ c'cz
=
d c
;
therefore,
by
subtraction,
(ac a'c)
x
+
(be' b'c) y
=
dc' d'c......
(iv).
Again, by multiplying
the first
equation by
c" and
the third
by
c and
subtracting,
we have
(ac"
-
a"c)
a;
+
(be"
-
b"c) y
=
dc"
-
d"c......
(v).
We now have the two
equations (iv)
and
(v)
from
which to determine the unknown
quantities
x and
y.
Using
the
general
formulae of Art.
143,
we have
=
-
(be'
-
b'c} (dc"
-
d"c)
+
(dc
-
d'c) (be"
-
b"c)
(ac' a'c) (be" b"c) (be
b'c)
(ac" a"c)
Method of undetermined
multipliers.
Multiply
the
equations (i)
and
(ii) by
X and
/A,
and add to
(iii);
then we have the
equation
x
(\a
+
IJM
+
a")
+
y(\b
+
pb'
+
b")
+
z
(Xc
+
fie'
+
c")
=
(\d
+
/j,d'
+
d"),
which is true for all values of X and
/JL.
Now let X and
p,
be so chosen that the co-efficients
of
y
and z
may
both be
zero,
\d +
ud'
+
d"
then as
=
,-> ,
\a
+
/j,a
+a
where X and
p,
are found from
and Xc
+
pc
+
c"
=
;
X
n
1
'*
b'c"
-
b"c
=
b"c
-
bo"
=
b7^T7c
'
Hence
_
d
(b'c"
-
b"c)
+
d'
(b"c
-
be"}
+
d"
(be'
-
b'c
)
"
a
(b'c"
-
b"c)
+
a'
(b"c
-
be")
+
a"
(be
1
-
b'c)
'
[The
numerator and the denominator of the first value
of
x,
which was obtained
by eliminating
z and
y
in succes-
152 SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGREE.
sion,
can both be divided
by
c
;
and the two values of # will
then be seen to
agree.]
Having
found the value of x
by
either of the above
methods,
the values of
y
and z can be written down.
For the value of
y
will be obtained from that of x
by
interchanging
a and
b,
a' and
b',
and a" and b". The
value of
y
can also be obtained from that of a;
by
a
cyclical change [see
Art.
93]
of the letters
a,
b,c;
a, b',
c
;
and
a", b",
c";
and a second
cyclical change
will
give
the
value of z.
It should be remarked that the denominators of the
values of
as,
y
and z are the
same,
and that there is a
single
finite value of each of the unknown
quantities
unless this denominator is zero.
Ex. 1. Solve the
equations
:
....................................
(i),
2=7....................................
(ii),
14
.................................
(iii).
Multiply (ii) by 3,
and subtract
(i)
;
then
5a; +
10y
=
15...................................
(iv).
Again multiply (i) by
3,
and subtract
(iii)
;
then
From
(v)
we have
y
=
1
; then, knowing y,
we have from
(iv)
x
=
1
;
and,
knowing
x and
j/j
we have from
(i)
2
=
1.
Thus
i=j/
=
2
=
l.
Ex. 2. Solve the
equations
:
l ..............................
(i),
d ........................ .....
(ii),
d*..............................
(iii).
Multiply (i) by
c and subtract
(ii)
;
then
(c-a)x+(c-b)y=c-d
........................
(iv).
Again multiply (i)
by
c
8
and subtract
(iii)
;
then
(c*-a*)x
+
(c*-V
i
)y=c*-d*
.....................
(v).
Now
multiply (iv) by
c
+
b and subtract
(v) ;
then
(c-</) (b-a)x
=
(c-d)(b-d);
^(b-d)(c-d)
(b-a)(c-a)'
SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGREE. 153
The values of
y
and z
may
now be written down :
they
are
._(c-d)(g-d). ,_(a-d)(b-d)
J
(c-b)(a-b)'
'
(a-c)(b-c)'
Instead of
going through
the
process
of
elimination,
we
may
at
once
quote
the
general
formulae. Thus
-
c
2
)
+
d?(c- b)
(be-
-
b*c)
+ a
(b*
-
c
2
)
+
a
2
(c
-
b)
~(b-c){-
=
,
~
(,
(b-a)(c
Ex. 3. Solve the
equations
:
,
~
(,~ !,
as above.
(b-a)(c-a)'
. ........
(i),
ax +
by
+cz
=
be
+ ca
+
ab
.....................
(ii)
,
bex
+
cay
+
abz
=
3abc
..............................
(in).
We have
_ (a
+
b
+
c) (ab*
-
cc
2
)
+
(be
+
ca
+
ab) (ca -ab)
+ Babe
(c
-
b)
afc
2
-
ac
2
+
a
(ca
-
db)
+
bc(c- b)
_
a
(b
-
c) { (b
+
c) (a
+
b
+
c)
-
be
-
ca
-
ab
-
3bc}
(b-c){ab
+
ac
-
a
2 -
be}
a(b- c)
2
~(o-6)(a-c)'
The values of
y
and z can now be written down:
they
are ,
6(c-a)
a
cCa-fc)
8
y
~
(b-c)(b-a)''
'
(c-a)(c-b)'
Ex. 4. Solve the
equations
:
................. .. ........
(i),
...........................
(ii),
c
s
=0
...........................
(iii).
The
equations may
he solved as in the
preceding examples,
or as
follows.
It is clear that
a, b,
c are the three roots of the
following
cubic
in X
Hence from Art.
129,
we have at once
and a;= -abc.
154
SIMULTANEOUS
EQUATIONS
OF THE FIRST DEGREE.
146.
Equations
with more than three unknown
quantities.
We shall return to the consideration of
simultaneous
equations
of the first
degree
in the
Chapter
on
Determinants,
and shall then shew how the solution of
any
number of such
equations
can be at once written
down.
The method of successive elimination or the method
of undetermined
multipliers
can however be extended to
the case when there are more than three unknown
quan-
tities. For
example,
to solve the
equations
ax
+by
+
cz +
dw =<?...............
(i),
a'x
+
b'y
+c'z +
d'w
=
e' .............
(ii),
a"x
+b"y
+
c"z
+d"w
=e"............
(iii),
a'"x +
V"y
+
c'"z
+
d'"w
=
e" ..........
(iv).
Multiply (i) by
X,
(ii) by p, (iii) by
v,
and add the
products
to
(iv).
Then we have
x
(aX
+
a'fji
+ a"v
+
a'")
+
y (b\
+
b'fi
+
b"v +
b'")
4
z
(cX
+
c>
+
c"v
+
c")
+
w(d\
+
d'n.
+
d"v +
d'")
e\ +
e'n
+ e"v +
d"....................................
(v).
Now choose
X,
jj,,
v so as to make the coefficients
of
y,
z and w in the last
equation
zero
;
then
e\
+
e'u, + e"v +
e"
,
.
N
where
X,
/A,
v are to be found from the
equations
6X
+
6>
+
b"v + b'"
=
cX +
c'(j,+c"v
+c'"
=
0> ............
(vii).
Hence we have to solve
(vii) by
Art. 145 and then
substitute the values of
X,
p,
and v in
(vi)
: this will
give
the value of
x\
and the values of the other unknown
quantities
can then be found
by cyclical
changes
of the
letters, a, b,
c, d,
&c.
SIMULTANEOUS
.EQUATIONS
OF THE FIRST DEGREE. 155
EXAMPLES XIL
Solve the
following equations.
r 11 1 94
l
x
y
_
t)
_
9
L
3~6~2'
*
x~y~^
_3y
=
l
!.,.= 10
5 10 2'
xy
37 4 3
K
6 3
-
=7^,
4 +5
=
-
+-
y
2 x
y
x
y
y
3
5. ax +
by
=
2ab,
6. a; +
ay
+
a
9
=
0,
bx
ay
=
b' a'.
x +
by
+ b
3
=
0.
7. a: +
y
=
2a,
8.
(6
+
c)
x
+
(b
-
c) y
=
lab,
(a b)x=(a
+
b)y.
(c
+
a)
x +
(c
a) y
=
2ac.
9. bx +
ay
=
2ab,
cfx +
b*y
=
a
3
+ b
3
.
10.
(a
+
6)
x
+
by
=
ax +
(b
+
a) y
=
a' b
3
.
11. x +
y
+
z=l,
12.
z
=
16.
2
13. x +
2y
+ 3z
=
3x +
y
+ 2z
=
2x+3y
+ z
=
G.
14.
y
+
z=2a,
15.
y
+ z
x=2a,
z +
x=2b,
z + x
y
=
2b,
x +
y=2c.
x +
y-z
=
2c.
1 -"6 EXAMPLES.
4
16.
y
+
z-3x=2a,
17. ax +
by
+
cz=l,
z +
x-3y
=
2b,
bx +
cy
+
az=l,
+ bz=l.
y+z-x
z+x-y
x +
y-z
iO. r
-- ---
=r
----
=
1.
b + c c
+ a a + b
19. x +
y
+ z
=
Q,
20.
ax +
by
+
cz=l,
bx +
cy
+ az
=
bc+ca+ab
t
a'x +
b*y
+ c*z
=
a + b + c. cx +
ay
+ bz
=
bc + ca+ab.
21 x +
y
+ z a + b +
c,
bx +
cy
+ az
=
a* + b
2
+ c-
8
,
ex +
ay
+ bz
=
a
3
+ b* + c
2
.
22. x +
y
+ *
=
0,
(b
+
c)
x +
(c
+
a) y
+
(a
+
b)
z
=
(b c) (c a) (a
-
b),
bcx +
cay
+ abz
=
0.
23. ax +
by
+ cz
=
a,
24. x
ay
+ a*z a*
0,
bx +
cy
+ az
=
b,
x
by
+ b*z
-
b*
=
0,
ex +
ay
+ bz
=
c. x
cy
+ c'z c
3
=
0.
^
25. ax +
by
+ cz
=m,
cfx + b
3
y
+ c*z
=
m
5
,
a*x +
b*y
+ c'z
=
m*.
26. ax +
cy
+ bz
=
a* + 2bc
t
cx +
by
+ az
=
b*+
2ca,
c* + 2ab.
27. x +
y
+ z
=
2a + 26 +
2c,
ax +
by
+ cz
=
2bc + 2ca +
2ab,
(b-c)x
+
(c-a)y
+
(a-b)z
28. ax +
by
+ cz
=
a + b +
c,
a*x -f
6*y
+ c
f
=
(a
+ b +
c)*,
bcx 4-
cay
+ abz
=
0.
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE.
157
29.
x+y
+ z
=
l + m +
n,
Ix +
my
+ nz
=
mn + nl +
Im,
30. lx +
ny
+ mz
=
nx +
my
+ lz
=
mx +
ly
+ nz
=
l
a
+ m
3
+ n
3
-
Zlmn.
31. l"x + m
a
y
+ n
s
z
=
Imx +
mny
+ nU
=
nix +
Imy
+ mnz
=
l + m + n.
32.
J-.+J-^U,!
6 +
a b +
fi
b +
,
c+a
c+p
c+y
33.
y
+ z + w
=
,
+ ? + cc
=
6,
M> + x +
y
=
c,
x +
y
+ z
=
d,
34. x +
ay
+ a?z + a*w + a
4
=
0,
a +
by
+ b*z + b
3
w + b*
=
0,
x +
cy
+ c*z + c
a
w + c*
=
0,
x +
dy
+ d?z + d?w + d
4
=
0.
Simultaneous
Equations
of the Second
Degree.
147. We now
proceed
to consider simultaneous
equa-
tions,
one at least of which is of the second or of
higher
degree.
We first take the case of two
equations containing
two
unknown
quantities,
one of the
equations being
of the first
degree
and the other of the second.
158 SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGHEE.
For
example,
to solve the
equations
:
From the first
equation
we have
Substitute this value of x in the second
equation ;
we then have
whence
y
2
+
14y
+
13
=
0,
that is
(z/
+
13)(y
+
l)
=
0;
..(/=-!,
or
y
=
-
13.
and if
y
=
-13,
x=ll.
Thus x
=
B, y=-l;
or 3=
11,
y=
-
13.
From the above
example
it will be seen that to solve
two
equations
of which one is of the first
degree,
and the
other of the second
degree,
we
proceed
as follows :
From the
equation
of the first
degree
find the value of
one of the unknown
quantities
in terms of the other un-
known
quantity
and the known
quantities,
and substitute
this value in the
equation
of the second
degree
;
one of
the unknown
quantities
is thus
eliminated,
and a
quadratic
equation
is obtained the roots of which are the values of
the unknown
quantity
which is retained.
The most
general
forms of two
equations
such as we
are now
considering
are
lac
+
my
+ n
=
0,
aa?
+
bxy
+
cy*
+
doc +
ey +f=
0.
From the first
equation
we have
my
+
n
ft*
-
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE. 150
Hence on substitution in the second
equation
we have
to determine
y
from the
quadratic equation
a
(my
+
ri)* Iby (my
+
ri)
+
cPy*
-
dl
(my
+
n)
+
ePy +fl*
=
0.
Having
found the two values of
y,
the
corresponding
values of x are found
by
substitution in the first
equation.
148. It should be remarked that we cannot solve
any
two
equations
which are both of the second
degree
;
for
the elimination of one of the unknown
quantities
will in
general
lead to an
equation
of the
fourth degree,
from
which the
remaining
unknown
quantity
would have to be
found
;
and we cannot solve an
equation
of
higher degree
than the
second,
except
in
very special
cases.
For
example,
to solve the
equations
oof
+
bx
+
c
=
y,
a;
2
+
y
2
=d.
Substitute ax
1
+
bx
+
c for
y
in the second
equation,
and
we have
a?
+
(ax*
+
bx
+
cf
=
d,
which is an
equation
of the fourth
degree
which cannot be
solved
by any
methods
given
in the
previous chapter.
149. There is one
important
class of
equations
with
two unknown
quantities
which can
always
be
solved,
namely, equations
in which all the terms which contain
the unknown
quantities
are of the second
degree.
The
most
general
forms of two such
equations
are
ox*
+
bxy
+
cy
1
=
d
and a'ar
+
b'xy
+
c'y*
=
d'.
Multiply
the first
equation by
d',
and the second
by
d
and subtract
;
we then have
(ad'
-
ad)
x* +
(Id
1
-
b'd) xy
+
(cd
1
-
c'd) f
=
0.
160
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE.
The factors of the above
equation
can be found either
by inspection,
or as in Art. 81
;
we therefore have two
equations
of the form Ix
+
my
=
either of which com-
bined with the first of the
given equations
will
give,
as in
Art.
147,
two
pairs
of values of x and
y.
Ex. 5. To solve the
equations
:
3/
2
-zt/
=
15
.................................
(i),
x
3
+
xy
=
14
.................................
(il).
We have
14(y*-xy)
=
15
(x*+xy);
that is
(5x
-
2y) (3x
+
7y)=C.
Hence 5x
-
2y
=
0,
or else Bx
+
ly
=
0.
If 5x
-
2y
=
0,
we have from
(i)
a
-g=15,
whence
y=5.
Hence also x=
2.
If 3x
+
lij
-
0,
we have from
(i)
whence
and then
X==F-
.
V
J
7 3
Thus
z=2, j/=5;
or x=-
75
,
y=
T
--.
V" v
j
150. The
following examples
will shew how to deal
with some other cases of simultaneous
equations
with two
unknown
quantities
;
but no
general
rules can be
given.
Ex.l. To solve
z-y=2,
xij
=
15.
Square
the members of the first
equation,
and add four times the
second
;
then
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE.
101
Hence x +
y
=
8,
which with
x-y
=
2,
gives
x=5 or
-3,
and
i/
=
3 or -5.
Thus x
=
5, y
=
3;
or z=
-3, y=
-
5.
Ex.2. To solve x
2
+
xy+y'
2
=a-
...........................
(i),
*
...........................
(ii).
Divide the members of the second
equation by
the
corresponding
members of the
first;
then
...........................
(in).
From
(i)
and
(iii) by
subtraction we have
6
4
2xy=a
1
-^-
............................
(iv).
From
(i)
and
(iv)
From
(Hi)
and
(iv)
we have
3&
4
-a
4
Finally,
from
(v)
and
(vi)
we have
1
f /3a
4
-6
4
/3&
4
-a
4
)
*
=
-
i * / 5
db
A / } ,
2] V
2a
2
V
2*
f'
1
[ /3a
4
-6
4
nnd v--
Ex. 3. To solve x
2
-
2t/
2
=
%
,
3x
2
+
xy-2y
2
=16y.
Multiply
the first
equation by 4,
and subtract the second
;
then
x
2
-xy-&y*
=
Q,
that is
(a;
+
2y) (x
-
3y)
=
;
or else
a;-3?/=0.
S.A.
11
162
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE.
If x +
'2y
=
Q,
the first
equation gives
4</--2j,*
=
4y;
.%
y=0
or
y
=
2,
whence z=0 or z=-4.
If x
-
3y
=
0,
the first
equation gives
whence
Thus x
=
0, y=0;
jr=-4,
12 4
*=y,
y-
t
.
Ex.4. To solve
By
subtraction we have
(x
+
y
+
i)
1
-
(*-
that is
(2z
+
3}(2y-l)
=
U.
Hence 2+3=0,
or
2y-l=0.
If
2x+3=0,
we have
9
whence
y
=
-
2.
If
2y
-
1=0,
we have
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE. 163
Hence
(b
+
1)*
+
(b
-
z)
4
=
20,*,
whence after reduction
Thus
and
EXAMPLES XIIL
Solve the
following equations
:
1. x +
y
=
a?
y*
=
23.
2. x*
-
4y*
+
x +
3y
=
2^
-
y
=
1.
3.
a:
2
+
a;2/-12,
4. a^+
2^=22,
B c a
- _!
-
y
~
x
~
84
'
x
+
7. a
(aj
+
y)
b
(x y)
=
xy.
a
l
+
l
-
l
9
a
+
b
'
10 O.
5
+
=
~2>
~5T
s
=
i">
a?
xy
a x
y
-
+
-
-
a6
-3
T
72
".
%f yx
o
xy
10. a +
y
=
2a,
11.
0^-^
+
^=109,
12. x
2
+
ay
+
^=133,
13. a; +
y=72,
112
164
EXAMPLES.
14.
1+1=2,
15.
*+y=l,
x
y
xyl
1--=8. o;' +
y
8
=
31,
x
y
16. x
s
+
y*
+
3xy
-
4
(x
+
y)
+ 3
=
0,
ajy
+ 2
(a;
+
y)
-
5
=
0.
17. x* +
xy
+
x=U,
18.
y*
+
xy
+
y
=
28.
*
-
xy
+
y*
=
3.
19.
x(y-b) =y(x-a)
=
2ab.
20. x + -=l. 21. ax +
by=2ab,
yy
1 a 6
-
=
4. -+-
=
2.
*
y
a:
22. -+
=
12,
23.
a?
a;
24.
xy--=a,
25.
y
1
X &
26. a; +
y
=
6,
27.
(*-.- y*) (a^
+
y*)
-
1440. a;
1
+
y
1
=
40x-y.
28. *'
-
xy
=
8x +
3,
29.
y-
y
=
3,
1
sry
xy-y*-8y-G.
a;-y
1
1 -I-
a;y
3
'
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE.
165
30.
x-y
=
a(x*-y
3
),
31.
+
=+
a o a o
-
+
-=?-
+
-=-
+ -.
ax o
y y
x
151.
Equations
with more than two unknown
quantities.
No
general
rules can be
given
for the solu-
tion of simultaneous
equations
of the second
degree
with
more than two unknown
quantities
: all that can be done
is to solve some
typical examples.
Ex. 1. Solve the
equations
:
(x
+
y)(x+z)=a* (i),
(y
+
z)(y
+
x)=V* (ii),
(z+x)(z+y)=c
a
(iii).
Multiply (ii)
and
(iii)
and divide
by (i) ;
22
then
Similarly
we have
.
ca
and x
+
y= (vi).
Also from the
original equations
it is clear that the
signs
mast all
be
positive
or all be
negative.
Add
(v)
and
(vi)
and subtract
(iv)
from the sum
;
then
2abc
j3h2-
So also
y=
and z
=
2a6c
1G6 SIMULTANEOUS
EQUATIONS
OF TUB SECOND DEGREE.
Ex.
2. Solve the
equations
:
x(y+z)=a
.................................
(i),
y(z+x)=b
.................................
(ii),
g(x
+
y)=c
.................................
liii).
We have
y(z
+
x)
+
z(x+y)-x(y
+
z)
=
b
+
c
-a,
that is
2yz
=
b
+
e
-
o.
Similarlj
2zx=c
+a-b,
and
2xy
=
a + b-c.
Hence
(2xy) (2zx)
=
(a
+
6-c)(c+a-6).
2yz
b+c-a
.
9r
,
c-a)
Hence
(c
+ a-
2(b+c-a)
,
. .. .
and
Bumlarly j,
=
and z
=
Ex. 3. Solve the
equations
:
a
(i),
a
(ii),
t*+2xy
=
b
(iii).
By
addition
(x+y
+
z)-=2a
+
b;
.:
x+y+z= j2a
+
b
(iv).
From
(i)
and
(ii) by
subtraction
Hence
x=y (v),
or else
x+y-2z=0 (vi).
L If
x=y,
we have from
(ii)
and
(iii) by
subtraction
/. z-x=
i/yi-a (vii).
Hence,
from
(iv), (v)
and
(vii),
=
1
3
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE.
167
IL When x
+
y
-
2a
0,
we have from
(iv)
and x +
y
-
5
Also,
from
(ii), y
s
+ a;
(x
+
y)
=
a,
which with the
previous equation gives
,
la-b
1
"
=
=F
V-T^S
Ex. 4. Solve the
equations
:
b*z
+
<Py
=
c
2
x -f- a
a
z
=
a
2
^
+
6
2
x
=
0^2.
We have l?z + c
i
y=xyz (i),
<?x +
a*z=xyz (ii),
and
a?y
+
&x=xyz (iii).
Multiply (i) by
-
a
8
,
(ii) by
6
s
,
and
(iii) by
c
2
,
and add
;
then 26
2
c
2
aj=(-a
3
Hence
*=0,
26
2
c
2
or else
yz=
If x
=
0, y
and z must also be zero.
Hence
x=y=z=Q;
26V
2c
2
a
2
and
sumlarly
Zx
=
c
3
+a
a_
b
*
2o
a
6
2
The solution then
proceeds
as in Ex. 2.
Ex. 5. Solve the
equations
:
x
a
-yz
=
a,
y*-zx
=
6,
z
2
-xy=c.
168 SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE.
We have
(x*
-
?/z)
2
-
(y-
-
zx)
(*
-
xy)
=
a
2
-
be,
that is x
(z
3
+
y
3
+
z*
-
Sxyz)
=
a
2
-
be.
Hence,
from the last
equation
and the two similar
ones,
a*
-
be b*-ca c*-ab'
Hence each fraction is
equal
to
/
*-v*
_
/
V (
2
-
^)
2
-
(fo
2
-
ca) (c*
-
ab)
~
V
Ex. 6. Solve the
equations
:
(i),
(ii),
+ +
=
3 ....................................
(in).
a b c
It is obvious that
x=a, y=b,
z=c will
satisfy
the
equations
:
put
then x=a
+ \,
y=b+fj,,
z=c
+
v,
and we have after reduction
................................
(iv),
<)
(vi)
From
(iv)
and
(v)
X
a]6
-
whence from
(vi)
X
Hence
ar=o, yb, z=c;
or else
x~a
y-6
g-c 2
(?>
-
c) (c
-
a)
fo
-
6)
a
(6-c)
~
b
(c
-
a)
~
c(a-b)
~
a'
2
(b-
Ex. 7. Solve the
equations
:
G,
xyz=
6.
This is an
example
of a
system
of three
symmetrical equations.
Such
equations
can
generally
be
easily
solved
by making
use of the
relations of Art. 129. Thus in the
present
instance it is clear that
x, y,
2 are the three roots of the cubic
equation
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE.
169
The roots of the cubic are
1, 2,
3.
Hence
*=1,
y
=
2,
*=3;
or
x=l, y
=
3, z=2;
or
=2,
j/
=
3, 2=1;
&c.
Ex. 8. Solve the
equations
:
x
+
y
+2
=
a ..............................
(i),
H+H
..............................
yz
+ zx +
xy
=
-e
2
...........................
(iii).
This
again
is a
system
of
symmetrical equations,
and two of the
relations of Art. 129 are
already given ;
we have therefore
only
to find
the third.
We have from
(ii),
yz
+ zx
+
xy
l
m
xyz
a'
/.
xyz=
-ac* .................................
(iv).
Then,
from
(i), (iii)
and
(iv),
we see that
x, y,
z are the roots of
the cubic X
s
-
aX
2
-
c
2
X
+
ac
2
=
0,
that is X
2
(X-o)-c
2
(X-a)
=
0;
/.
X=a,
or X= c.
Thus
x=a,
y=c, z=c;
&G.
Ex.
9. Solve the
equations
:
By
addition
that is
(y -z)(z- x) (x-y)
=
(b- c) (c
-
a) (a
-
b).
By multiplication
)(*-*)(z-y)=
"We*
(b
-
c)
(c
-
a) (a-b);
Hence
xyz
=
abc .................................
(i),
or
xyz=
-abc..............................
(ii).
Again a*(b-c)y
+
b*(c-a)x=x
3
y(y -z)+xy
2
(z-x)
=xyz(y-x)
........................
(iii).
Hence,
if
xyz
=
abc,
we have from
(iii)
{b
a
(c-a)+abc}x+{a?(b-c)-abc}y
=
Q,
that is
bx(bc+ca-al>)
-
ay(bc
+
ca-ab)
=
;
.*.
-
=
,- ,
and therefore each
=
-
.
a b c
170
EXAMPLES.
Thus,
when *wr=afec. we have
-
=
f
=
-
.
a o c
Hence each is
equal
to
* /
~j~
=
*fl-
Thus
-
=
!
=
-=!,
oi
ss^-
= -.
a o c aw ota co
or
=
-^-
=
1
CLdr btiF CuP
If
x//z
=
-abc,
we have from
(iii)
x
.,
\ V I
a b
Hence also each =
-
(ab
-be-
ca)
j/{
~
(be
-
co
-
ab) (ca
-ab-
be)
(ab-bc-
ca)}.
EXAMPLES XIV.
Solve the
following equations
:
1.
yz
=
cf,
2.
x(x
+
y
+
z).
zx
=
b
3
,
y(x
+
y
+
z)<
3.
-
5.
yz
=
by
+
cz,
6. x
1
+
2yz =12,
zx
=
cz +
ax,
y*
+
2zx
=
12,
7.
(y
+
z)(x
+
y
+
z)
=
a,
8.
(y
+ l
(z
+
x) (x
+
y
4-
z)
-
b,
(z
+
c) (x
+
a)=
b*
t
i-
y
+
z)
=
c.
(x
+
a) (y
+
b)
=
c*.
SIMULTANEOUS
EQUATIONS
OF THE SECOND DEGREE. 17 1
9.
x*-(y-z)*
=
a>,
10. x
(y
+ %
x)
=
a,
y(z+x-y)
=
b,
z
(x
+
y z)
=
c.
ii.
y^=*-?=
x
-y=2xyz.
a b c
y
+ z z + x x +
y
x*
zx
=
b + z +
x,
12.
13.
15.
yz-f*
=
cy
+
bz,
zx
g*
=
az
+
ex,
xy
ft
9
=
bx +
ay.
17. x +
y
+ z
=
6,
a* + b' + c
2
'
14.
yz
=
a
(y
+
z)
+
a,
zx
=
a
(z
+
x)
+
ft,
16.
18. x +
y
+ z
=
15,
xyz
=
6.
19. x +
y
+ z
=
9,
'=
41,
ays
=
105.
20.
x+y
+ z=
10,
yz
+ zx +
xy
=
33,
(y+z)(z+x)(x+y)
=
29.
yz
zx
_^
= =
_xy_ _
^x'
+
y'
+ z'
^
bz +
cy
ex + az
ay
+ bx a* + b' + c*
*
23. ax
=
-
no
,
y .
*
,
ZA h
JT-
+
-
=
i,
JC C
x b z
'
,
-+-
+-=1,
aye
x
y
c
_
1
a 6 2
a a;
6y
=
^
+
*'
=
*+*.
172
EXAMPLES
24.
y*
+ 2*
-
x
(y
+
z)
=
a*,
25. x' +
yz-a
t
=
y
t
+
zx-b
t
=!?+vy-<?=-
a
26. x
(x
+
y
+
z)
-
(y*
+ a*
+
yz)
=
a,
y (x
+
y
+
z)
(z*
-4-
X
s
+
zx)
=
b,
27.
x* +
y*
+ z'
=
a' + b* +
c,
(b c)
x +
(c a)
y
+
(a b)
z
=
0.
28.
Cx
+
y) (x
+
z)
=
ax,
29.
x*-yz
=
ax,
(y
+
)(y
+
*)
=
fy,
y
i
-zx
=
by)
(z
+
x)(z
+
y)=cz.
z*-xy
=
cz.
30. a;
2
+ a
(2s;
+
y
+
z)
=
y
2
+ 6
(2y
+ z +
x)
=
z
1
+ c
(2z
+ x +
y)
=
(x
+
y
+
z)
9
.
31.
y
%
+
yz
+ z
3
=
a
8
,
z
1
+ zx + x*
=
b*,
x* +
xy
+
y*
=
c*.
32.
ax
by
cz
I I 1111
-
+ +
=-
+
r
+-.
yz
zx
icy
a b c
CHAPTER XT.
PROBLEMS.
152. WE shall in the
present chapter
consider a class
of
questions
called
problems.
In a
problem
the
magni-
tudes of certain
quantities,
some of which are known and
others
unknown,
are connected
by given
relations;
and the
values of the unknown
quantities
have to be found
by
means of these relations.
In order to solve a
problem,
the relations between the
magnitudes
of the known and unknown
quantities
must
be
expressed by
means of
algebraical
symbols
: we thus
obtain
equations
the solution of which
gives
the
required
values of the unknown
quantities.
It often
happens
that
by solving
the
equations
which are the
algebraical
statements of the relations
between the
magnitudes
of the known and unknown
quantities,
we obtain results which do not all
satisfy
the
conditions of the
problem.
The reason of this is that in a
problem
there
may
be
restrictions,
expressed
or
implied,
on the numbers
concerned,
which restrictions cannot be
retained in the
equations.
For
example,
in a
problem
which refers to a number of
men,
it is clear that this
number must be
integral,
but this condition cannot be
expressed
in the
equations.
Thus there are three
steps
in the solution of a
problem.
We first find the
equations
which are the
algebraical
expressions
of the relations between the
magnitudes
of the
174 PROBLEMS.
known and unknown
quantities;
we then find the values
of the unknown
quantities
which
satisfy
these
equations
;
and
finally
we examine whether
any
or all of the values
we have found violate
any
conditions which are
expressed
or
implied
in the
problem,
but which are not contained in
the
equations.
The
necessity
of this final examination
will be seen from some of the
following examples
of
problems.
Ex. 1. A has 5 and B has ten
shillings.
How much mast A
give
to
B in order that he
may
have
just
four times as much as B ?
Let x be the number
of shillings
that A
gives
to B.
Then A will have 100
-
x
shillings,
and B will have 10 +x
shillings.
But, by
the
question,
A now has four times as much as B.
Hence we have the
equation
100-x=4(10+ar);
/. *=12.
Thus A must
give
12
shillings
to B.
It should be remembered that x must
always
stand for a number.
It is also of
importance
to notice that all concrete
quantities
of the
same kind must be
expressed
in terms of the same unit.
Ex. 2. One man and two
boys
can do in 12
days
a
piece
of work
which would be done in 6
days by
3 men and 1
boy.
How
long
would it take one
ma.n
to do it ?
Let x=the number of
days
in which one man would do the
whole,
and let
y=
the number of
days
in which one
boy
would do the whole.
Then a man does
-
th of the whole in a
day
;
and a
boy
does
-
th
x
y
of the whole in a
day.
By
the
question
one man and two
boys
do
^th
of the whole in a
day.
Hence we have
i#i-A
x
y
12
We have
also,
since 3 men and 1
boy
do
^th
of the whole in a
day,
5
+
i-i
x
+
y
6'
Whence x=20.
Thus one man would do the whole work in 20
days.
PROBLEMS. 175
Ex. 8. In a certain
family
eleven times the number of the children is
greater by
12 than twice the
square
of the number. How
many
children are there ?
Let x be the number of children
;
then we have the
equation
11*
=
2^
+
12,
or
2X
3
-11*+
12
=
0,
that is
(2x-S)(x-4)=0.
Hence
#
=
4,
or
x=%.
The value
x=f
satisfies the
equation,
but it must be
rejected,
since
it does not
satisfy
all the conditions of the
problem,
for the number
of children must be a whole number.
Thus there are 4 children.
Ex. 4. Eleven times the number of
yards
in the
length
of a rod is
greater by
12 than twice the
square
of the number. How
long
is the
rod?
This leads to the same
equation
as Ex.
3;
but in this case we
cannot
reject
the fractional result. Thus the
length
of the rod
may
be 4
yards,
or it
may
be a
yard
and a half.
Ex. 5. A number of two
digits
is
equal
to three times the
product
of
the
digits,
and the
digit
in the ten's
place
is less
by
2 than the
digit
in the unit's
place.
Find the number.
Let x be the
digit
in the ten's
place
;
then x +
2 will be the
digit
in the unit's
place.
The number is therefore
equal
to
1
Hence,
by
the
question,
Hence x
=
2,
or
x=-\.
Now the
digits
of a number must be
positive integers
not
greater
than
nine;
hence the value x=
-$
must be
rejected.
The
digit
in
the ten's
place
must therefore be
2,
and the
digit
in the unit's
place
must be 4. Hence 24 is the
required
number.
Ex. 6. A number of two
digits
is
equal
to three times the sum of the
digits.
Find the number.
Let x be the
digit
in the ten's
place,
and
y
the
digit
in the unit's
place;
then the number will be
equal
to
Wx+y.
Hence,
by
the
question,
/.
lx=2y.
17G
PROBLEMS.
Since x and
y
must both be
positive integers
not
greater
than
9,
it follows that x mast be 2 and
y
must be 7. Thus the
required
number is 27.
Ex. 7. The sum of a certain number and its
square
root is 90. What
is the number ?
Let x be the
number;
then we have the
equation
.-.
(x-90)
s
=,
or x
1
-181x
+
8100
=
0,
that is
(x-81)(x-100)=0.
Hence a;
=81,
or *=100.
If,
in the
question,
the
square
root means
only
the arithmetical
square root,
81 is the
only
number which satisfies the conditions.
If, however,
'
its
square
root
'
is taken to mean
'
one of its
square
roots,'
both 81 and 100 are admissible.
Ex. 8. The sum of the
ages
of a father and his son is 100
years ;
also
one-tenth of the
product
of their
ages,
in
years,
exceeds the father's
age by
180. How old are
they
?
Let the father be x
years old;
then the son will be 100
-
x
years
old.
Hence, by
the
question,
,V:t (100-*)=
a; +
180;
.-. x
2
-90j; +
1800=0,
that is
(z-60)(z-30)
=
0.
Hence
x=60,
or a-
=
30.
If the father is
60,
the son will be 100-60=10. If the father
is
30,
the son will be 100-30
=
70,
which is
impossible,
since the
son cannot be older than the father.
Hence the father must be 60 and the son 40
years
old.
Ex. 9. A man
buys pigs, geese
and ducks. If each of the
geese
had
cost a
shilling less,
one
pig
would have been worth as
many geese
as
each
goose
is
actually
worth
shillings.
A
goose
is worth as much as
two
ducks,
and fourteen ducks are worth seven
shillings
more than
a
pig.
Find the
price
of a
pig,
a
goose,
and s duck
respectively.
Let x=the
price
in
shillings
of a
pig,
?/= .. .1 ,, goose,
and 2=
,,
duck.
Then, by
the
question,
a
pig
is worth
y
times
(y
-
1) shillings;
.-. x
=
y(y-l)
................................
(i).
PROBLEMS. 177
Since a
goose
is worth 2
ducks,
.'.y
=
2*
..........................................
(ii).
And,
since 14 ducks are worth
7
shillings
more than a
pig,
Hz
=
7 +
x
....................................
(iii).
From
(i)
and
(ii)
we have the values of x and z in terms of
y
; and,
substituting
these values in
(iii),
we have
01
.-.
y
=
7,
or
y=l.
If
y=7,
ar
=
42 from
(i),
and z
\
from
(ii).
If
y
=
l,
x=0 from
(i),
and
z=\
from
(ii).
These values are how-
ever
inadmissible,
since
pigs
cannot be
bought
for
nothing.
Hence a
pig
cost
42s.,
a
goose 7s.,
and a duck 3*. M.
EXAMPLES XV.
1. Divide 50 into two
parts,
such that twice one
part
is
equal
to three times the other.
2. A has 5 less than
B,
G has as much as A and B
together,
and
A, B,
C have 50 between thorn. How much
has each?
3. One man is 70 and another is 45
years
of
age ;
when
was the first twice as old as the second ]
4. How much are
eggs
a
score,
if a rise of 25
per
cent, in
the
price
would make a difference of 40 in the number which
could be
bought
for a
sovereign
]
5. A
bag
contains 50 coins which are worth XI 4
altogether.
A certain number of the coins are
sovereigns,
there are three
times as
many half-sovereigns,
and the rest are
shillings.
Find
the number of each.
6. A can do a
piece
of work in 20
days,
which B can do
in 12
days.
A
begins
the
work,
but after a time B takes his
place,
and the whole work is finished in 14
days
from the
beginning.
How
long
did A work ?
S.
A.
12
178
EXAMPLES.
7. A man
buys
a certain number of
eggs
at two a
penny,
four times as
many
at 5d. a
dozen,
five times as
many
at Sd.
a
score,
and sells them at 3s. Sd. a
hundred,
gaining by
the
transaction 3s. 6c. How
many eggs
did he
buy?
8. A bill of .63. 5s. was
paid
in
sovereigns
and
half-crowns,
and the number of coins used was
100;
how
many sovereigns
were
paid
1
9. A man
walking
from a town A to another B at the
rate of 4 miles an
hour,
starts one hour before a coach which
goes
12 miles an
hour,
and is
picked up by
the coach. On
arriving
at B he observes that his coach
journey
lasted two
hours. Find the distance from A to B.
10. Two
passengers
have
altogether
600 Ibs. of
luggage
and are
charged
for the excess above the
weight
allowed 3s. 4d.
and lls. Sd.
respectively.
If all the
luggage
had
belonged
to
one
person
he would have been
charged
1. How much
luggage
is each
passenger
allowed free of
charge]
11. A
piece
of work can be done
by
A and B in 4
days,
by
A and C in 6
days,
and
by
B and G in 12
days
:
find in
what time it would be done
by
A,
B and C
working together.
12. A father's
age
is
equal
to those of his three children
together.
In 9
years
it will amount to those of the two
eldest,
in 3
years
after that to those of the eldest and
youngest,
and
in 3
years
after that to those of the two
youngest.
Find their
present ages.
13. A and B start
simultaneously
from two towns to meet
one another : A travels 2 miles
per
hour faster than B and
they
meet in 3 hours : if B had travelled one mile
per
hour
slower,
and A at two-thirds his
previous pace they
would have
met in 4 hours. Find the distance between the towns.
14. A traveller walks a certain distance : if he had
gone
half a mile an hour
faster,
he would have walked it in
-J-
of the
time : if he had
gone
half a mile an hour slower he would have
been 2 hours
longer
on the road. Find the distance.
EXAMPLES. 179
15. Divide 243 into three
parts
such that one-half of the
first,
one-third of the
second,
and one-fourth of the third
part,
shall all be
equal
to one another.
16. A sum of
money consisting
of
pounds
and
shillings
would be reduced to
one-eighteenth
of its
original
value if the
pounds
were
shillings,
and the
shillings pence.
Shew that its
value would be increased in the ratio of 15 to 2 if the
pounds
were
five-pound
notes,
and the
shillings pounds.
17. 1000 is divided between
A,
B,
C and D. B
gets
half as much as
A,
the excess of C"s share over D's share is
equal
to one-third of A's
share,
and if B'a share were increased
b'y
100 he would have as much as C and D have between
them
;
find how much each
gets.
18. Find two
numbers,
one of which is three-fifths of the
other,
so that the difference of their
squares may
be
equal
to
16.
19. Find two numbers
expressed by
the same two
digits
in different orders whose sum is
equal
to the
square
of the sum
of the two
digits,
and whose difference is
equal
to five times the
square
of the smaller
digit.
20. A man rode one-third of a
journey
at 10
miles
per
hour,
one-third more at 9 miles
per
hour,
and the rest at 8
miles
per
hour. If he had ridden half the
journey
at 10 miles
per
hour and the other half at 8 miles
per
hour,
he would have
been half a minute
longer
on the
journey.
What distance did
he ride ]
21. Two
bicyclists
start at 12
o'clock,
one from
Cambridge
to Stortford and
back,
and the other from Stortford to
Cambridge
and back.
They
meet at 3 o'clock for the second
time,
and
they
are then 9 miles from
Cambridge.
The distance from
Cambridge
to Stortford is 27 miles. When and where did
they
meet lor the
first time ?
22. Divide 1015
among A, B,
C so that B
may
receive
5 less than
A,
and G as
many
times B's share as there are
shillings
in A's share.
12 2
180
EXAMPLES.
23. On a certain road the
telegraph posts
are at
equal
distances,
and the number
per
mile is such that if there were
one less in each mile the interval l>etween the
posts
would be
increased
by 2y yards.
Find the number of
posts
in a mile.
24. The sum of two numbers
multiplied by
the
greater
is
144,
and their difference
multiplied by
the less is 14 : find
them.
25. A and B start
simultaneously
from two towns and
meet after five hours
;
if A had travelled one mile
per
hour
faster and B had started one hour
sooner,
or if B had travelled
one mile
per
hour slower and A had started one hour
later,
they
would in either case have met at the same
spot they
actually
met at. What was the distance between the towns ?
26. A battalion of
soldiers,
when formed into a solid
square, present
sixteen men fewer in the front than
they
do
when formed in a hollow
square
four
deep. Required
the
number of men.
27. A number of two
digits
is
equal
to seven times the
sum of the
digits
;
shew that if the
digits
be
reversed,
the
number thus formed will be
equal
to four times the sum of the
digits.
28. A sets out to walk to a town 7 miles
off,
and B starts
20 minutes afterwards to follow him. When B has overtaken
A he
immediately
turns
back,
and reaches the
place
from
which he started at the same instant that A reaches his
destination.
Supposing
B to have walked at the rate of 4
miles an hour : find A's rate.
29.
A starts to
bicycle
from
Cambridge
to
London,
and B
at the same time from London to
Cambridge,
and
they
travel
uniformly
: A reaches London 4
hours,
and B reaches
Cambridge
1
hour,
after
they
have met on the road. How
long
did B take
to
perform
the
journey
?
30. A number consists of 3
digits
whose sum is 10. The
middle
digit
is
equal
to the sum of the other two
;
and the
number will be increased
by
99 if its
digits
be reversed. Find
the number.
r
EXAMPLES.
181
31. Two vessels contain each a mixture of wine and water.
In the first vessel the
quantity
of wine is to the
quantity
of
water as 1
:
3,
and in the second as 3:5. What
quantity
must be taken from each in order to form a third
mixture,
which shall contain 5
gallons
of wine and 9
gallons
of water 1
32.
Supposing
that it is now between 10 and 11
o'clock,
and that 6 minutes hence the minute hand of a watch will be
exactly opposite
to the
place
where the hour hand was 3 minutes
ago
: find the time.
33.
A,
B and C start from
Cambridge,
at
3,
4 and 5
o'clock
respectively
to
walk,
drive and ride
respectively
to
London. C overtakes B at 7
o'clock!
and C overtakes A 4
miles further on at
half-past
seven. When and where will
overtake A 1
34. A train 60
yards long passed
another train 72
yards
long,
which was
travelling
in the same direction on a
parallel
line of
rails,
in 12 seconds. Had the slower train been
travelling
half as fast
again,
it would have been
passed
in 24
seconds. Find the rates at which the trains were
travelling.
35. A distributes 180 in
equal
sums
amongst
a certain
number of
people.
B distributes the same sum but
gives
to
each
person
6 more than
A,
and
gives
to 40
persons
less than
A does. How much does A
give
to each
person
1
36. Three vessels
ply
between the same two
ports.
The
first sails half a mile
per
hour faster than the
second,
and
makes the
passage
in an hour and a half less. The second
sails
three-quarters
of a mile
per
hour faster than the third
and makes the
passage
in 2 hours less. What is the distance
between
the
ports
?
37. Two
persons A,
B walk from P to
Q
and back. A
starts 1 hour after
B,
overtakes him 2 miles from
Q,
meets him
32 minutes
afterwards,
and arrives at P when B is 4 miles off.
Find the distance from P to
Q.
CHAPTER XIL
MISCELLANEOUS THEOREMS AND EXAMPLES.
153. Elimination. When more
equations
are
given
than are
necessary
to determine the values of the un-
known
quantities,
the constants in the
equations
must be
connected
by
one or more
relations,
and it is often of
importance
to determine these relations.
Since the relations
required
are not to contain
any
of
the unknown
quantities,
what we have to do is to eliminate
all the unknown
quantities
from the
given system.
The
following
are some
examples
of Elimination :
Ex. 1. Eliminate x from the
equations
ax
+
b
=
0,
a'x + b'=0.
From the first
equation
we have x=
,
and from the second
equation
we have x=
7
.
Hence we must have
-
=
-.
,
or
ba'-b'a=0;
which is
the
a a
required
result.
Ex. 2. Eliminate x and
y
from the
equations
ELIMINATION. 183
From the first two
equations
we have
[Art. 143]
y
be'
-
b'c
~
ca!
-
c'a
~
ab'
-
a'b
'
These values of x and
y
must
satisfy
the third
equation ;
hence
be' -b'c
,,ca'-c'a ,,
-
a
r,
-
rr+b ~TI
-
^
+
c
=>
ab ab ab ab
or a"
(be
1
-
b'c)
+
b"
(ca'
-
c'
a)
+ c"
(ab'
-
a'b)=0,
the
required
result.
The
general
case of the elimination of n
-
1 unknown
quantities
from n
equations
of the first
degree
will be considered in the
Chapter
on Determinants.
Ex. 3. Eliminate x from the
equations
As in Art.
143,
we have
X* X
be'
-
b'c ca'
-
c'a ab'
-
a'b
'
Hence
(be'
-
b'c) (ab'
-
a'b)
=
(ca'
-
c'a]
3
,
the
required
result.
It should be remarked that the above condition is also the
condition that the two
expressions
ax
2
+
bx
+
c and a'ac
2
+
b'x
j-
c'
may
have a common factor of the form x
-
a
;
for if the
expressions
have
a common factor of the form x
-
a
they
must both vanish for the
same value of x.
Ex. 4. Eliminate x from the
equations
ax*
+
bx
+
c
=
0,
a'x
3
+
b'x
-t-
c' 0.
As in Ex.
3,
we have
x
3
x 1
be' b'c ca' c'a ab'
-
a'b
'
be'
-
b'c
_
(ca! -c'aV
'*
ab'-a'b
~
\ab'-a'b)
'
.:
(be
1
-
b'c) (ab'
-
a'6)
2
=
(ca!
-
c'a)
3
,
the
required
relation.
184 ELIMINATION.
Ex. 5. Eliminate x from the
equations
..............................
(i),
a'x*+b'x*+c'x+d'=0 .............................
(ii).
Multiply (i) by
a'x,
(ii) by a,
and subtract
; tben,
(ab'-ba')x^
+
(ac'-ca')x
+
ad'=0
.....................
(iii).
We can now eliminate x from
(i)
and
(iii)
as in Ex. 3.
Ex 6. Eliminate
x, y,
z from the
equations
=a
..............................
(i),
..............................
(ii),
..............................
(iii),
xyz
=
d
3
..............................
(iv).
From
(i)
and
(ii)
we have
2yz
+
2zx
+
2xy
=
a
2
-
fe
3
.
From
(iii)
and
(iv)
we have
X
s
+
y
3
+
z>
-
3xyz
=
c*
-
3d
3
,
i.e.
(x+y+z)(x*
+
y*
+
z*-yz-zx-xy)=c*-3d*.
Hence
o{6
J
-i(a
2
-6
a
)}
=
c
3
-3d
s
;
/. a*
+
2c
s
-6d
3
-3a6
2
=0,
the
required
result.
Ex. 7. Eliminate
x,
y,
z from the
equations
x*(y+z)=a?
..............................
(i),
y*(z+x)
=
b
n
-
..............................
(ii),
*(x
+
y)=rc
2
..............................
(iii),
xyz=abc
...........................
(iv).
From
(i), (ii), (iii) by multiplication
*
V**(y +*)(*+*)
(x
+
y)
=
Hence,
from
(iv),
that
is,
the
required
result.
Ex. 8. Eliminate
I, m, n, I', in',
n' from the
equations
II'
'
=
a,
mm'
=
6,
nn'
=
c,
ELIMINATION.
185
By
continued
multiplication
of the last three
equations,
we have
8fgh
=
2lmnl'm'n'
+
IV
(m-n"
2
+
mPn-)
+mm'
(nT
+
n'
2
*
2
)
+
nn'
(Pm
1
-
+
I'W)
=IV
(mn
1
+
m'n)
2
+
mm'
(nl
+
n'l)
2
+nn'
(lm'
+
I'm)
2
-
iU'mm'nn'
=
40/
2
-f
460
2
+
4ch
2
-
4a6c.
Hence alc +
2fgh-af*-lg*-ch
2
=0.
154. To
find
the condition that the most
general quad-
ratic
expression
in X and
y may
be
expressed
as the
product
of
two
factors of
the
first degree
in x and
y.
The most
general quadratic expression
in x and
y may
be written in the form
ax*
+
Zhxy
+
by
2
+
2gx
+
2fy
+
c
(i).
What is
required
is the condition that the above ex-
pression may
be
identically equal
to
(Ix
+
my
+
n) (I'x
+
m'y
+
ri) (ii),
where
I, m, n, I', m',
n do not contain x or
y.
Now if
(i)
and
(ii)
are
identically equal
we
may
equate
the coefficients of the different
powers
of x and
also of
y [Art. 91].
Hence we have
II'
=
a,
mm
=
b,
nn
=
c,
mn +
m'n
2f,
nl'
+
n'l
=
2g,
lm
+
I'm
=
2&.
Eliminating
I, m, n,
I',
m',
n
[Art.
153,
Ex.
8],
we have
abc
+
2fgh
-
a/
2
-
bg*
-
ch*
=
0,
the condition
required.
Ex. 1. For what value of X is
12x*
-
lOxy
+
2y*
+ 11*
-
5y
+\
the
product
of two factors of the first
degree
in x and
y?
Ans. X
=
2.
Ex. 2. For what value of X is
12X
2
+
36z</
+
Xy"
+
6z +
6y
+
3
the
product
of two factors of the first
degree
in x and
y
1
Ans. X=28.
186
EQUATIONS
WITH
RESTRICTIONS.
155.
Equations
in which there is some re-
striction on the values of the letters. A
single
equation
which contains two or more unknown
quantities
can be satisfied
by
an indefinite number of values of the
unknown
quantities, provided
that these values are not
in
any way
restricted. If however the values of the un-
known
quantities
are
subject
to
any
restriction,
a
single
equation may
be sufficient to determine more than one
unknown
quantity.
For
example,
if we have the
single equation 2x+5y=7,
and restrict both x and
y
to
positive integral
values, the
equation
can
only
be satisfied
by
one set of
values,
namely
by
the values x
=
1,
y
=
1.
Again,
from the
single equation
3
(x
-
a)*
+
4
(y
-
&)'
=
0,
with the restriction that all the
quantities
must be
real,
we can conclude both that x a
=
0,
and that
y
6
=
0;
for the
squares
of real
quantities
must be
positive,
and
the sum of two or more
positive quantities
cannot be zero
unless
they
are all zero.
Ex. 1. If
(a
+
b
+
c)*=3(bc
+
ca +
ab),
then a=6=c.
We have a
2
+
6
a
+
c
2
-6c-ca-a6=0,
that is
i{(ft-c)
a
+(c-o)
a
+(a-6)
2
}=0.
Whence b
-
c,
c
-
a and a
-
b must all be zero.
Ex. 2.
If
x, of,
y, y'
be all
real,
and
2
(**
+
z*
-
xaf) (y
a
+
y*
-
yy
r
)
=
then will x=x and
y=y
/
.
We have
=
0,
that is
(x-x')*(y-y')*
+
(xy'-xry)*=0.
Hence
xy'-x'y
=
and
(x
-
x') (y
-
y
1
)
=
0.
From the second relation x
=
x' or
y
=
y';
and either of these
combined with the first relation shews that both x-=v! and
y=y\
IDENTITIES.
187
Ex.3. If a
1
s
+o
2
s
+rt
3
*+ ......
=p\
V+V+V
+ ......
=r.
and
Oj61
+
a
2&,+a3
6
3
+ ......
=P1,
the
quantities being
all
real;
then will
^i_23_^3_
&c
JP
7 ^-
ixc.
&!
fc
s
&
3
<z
Multiply
the
equations
in order
by q
2
, p
2
and
-
2pg respectively,
and add
;
we then have
(qal
-pb1)*
+
(qat -pbj*+(qas -pb3
)*
+ ......
=
0.
Heuce
qa1 -pbl
= =
qay -pb3 =qa:t -pb
a
=
...... &o.
Therefore
==%==
...... &c.
b! q
b
2
b
3
156. We have
already proved
that
a
3
+
6* + c
s
-
3a6c
=
(a
+
b
+
c) (a
2
+
6
8
+
c
2
-
be
-
ca
-
ab)
=
$(a
+ b
+
c){(b
-
c)
a
+
(c
-
a)
8
+
(a
-
6)
2
}
=
(a
+
b
-f-
c) (a
+
tab
+
a>
2
c)(a
+
co
3
b +
&>c),
where w is either of the cube roots of
unity. [See
Art.
139.]
From the above
many
other identities can be found.
Left side
=
{b
+
c
+c +
a
+
a
+
b} {(c
+ a-a
+
6)
2
+ two similar
terras}
Ex.2.
(6-fc-a)
s
+(
-3(6
+ c
Left side
=i(a
+
6
+
c){(26-2c)
2
+
two similar
terms}
=4(a
3
+6*+c
3
-3a6c).
Ex. 3.
(*'
-
y)
3
+
(f
-
2X)
+
(z
a
-
xy)
3
-
3
(*
-
yz) (t/
2
-
zx) (z*-xy)
Left side
=
^(x'^-^y^+z'-yz-zx-xy [(j/
2
-x-z
2
-a;i/)
2
+ two
similar
terms]
=
h(x
!i
+
y*
+
z
i
-yz-zx-xy)(z
+
y+z)V[(y-z)*
+
two
similar
terms]
=
(x
+
y
+
z)
2
(x-
+
y*
+
z-
-yz-zx- xy)-
1 88
EXAMPLES.
Ex.4. Shew that
(a?
+
y
3
+
s?
-
Sxyz) (a
s
+
fc
3
+
c
3
-
3a6c)
can be
expressed
in the form Z
3
+ Y
3
+ Z
3
-3XYZ.
We have
and
The continned
product
of the left members of the above
equations
and the continued
product
of the
expressions
on the
right
is
(ax
+
by
+
cz)
3
+
(bx
+
cy
+
az)
3
+
(ex
+
ay
+
bz)
s
3
(ax +by
+
cz) (bx
+
cy
+
az) (ex
+
ay
+
bz)
,
which is of the
required
form.
157. Definitions. The
symbol
=
is often used to
denote that the two
expressions
between which it is
placed
are
identically equal.
Thus a* 6
s
=
(a
+
b) (a 6).
The sum of
any
number of
quantities
of the same
type
is often
expressed by writing only
one of the terms
preceded by
the
symbol
2.
Thus S6c means the sum
of all such terms as be
;
so that if there are three letters
a, b, c,
S6c
=
be
+
ca
+
ab. So also the
identity
may
be written
(2a)
2
=
2a*
+
22ab.
The
product
of
any
number of
quantities
of the same
type
is often
expressed by writing only
one of the factors
preceded by
the
symbol
EL Thus II
(6
+
c)
means the
product
of all such factors as
(6
-f
c)
;
so that if there are
three letters
a, b,
c,U(b
+
c)=(b
+ c)
(c
+
a) (a
+
6).
158. The
following examples
illustrate cases of fre-
quent
occurrence.
EXAMPLES.
1 89
Ex. 1. If a
s
+ 6
3
+c
3
=
(a
+
b
+
c)
3
,
then will
a
2+l
+
&2+l
+
C
2n+l
=
(
a
+b +
c)
2"*1
,
where n is
any positive integer.
Since
(a
+
6
+
c)
3
=
a
s
+&
3
+ c
s
+
3
(b
+
c) (c
+
a) (a
+
b),
the
given
relation shews that
(b
+
c) (c
+
a)(a
+
b)=Q.
Hence either b
+ c=0,
or c
+
a=0 or a + &
=
0.
If b
+
c
=
;
then b
zn+l
=
(
-
c)**
+l
=
-
c
2tt+1
,
and therefore
Thus if i + c
=
0,
a
a>>-*-1
+ 6
2'
+1
+
c
2n+1
-(a
+
6
+
c)
2n+
1
becomes
a
2n-H
+
-.>n+l
+ C
2u+l _
a
'J+l
_
J,2-H
+
C
2n+l
_
Q.
Hence a
2"+1
+
fe
2n+1
+
c
2n+1
=(a
+
6
+
c)
2B+1
if &
+c=0;
and so also
if c +
a
=
0,
or if a
+
b
=
0. This
proves
the
proposition,
since
|
(a+6)=0.
Ex. 2.
If
x,
y,
z be
unequal,
and if
prove
that each
equals 2xyz,
and that x
+
y-r
Since
y
3
+
z
3
+
m
(y
2
+
z
2
)
=
z
3
+
y?
+
m
(z
2
+
a;
2
),
we have
that is
(y-x){
Therefore,
y
-
x not
being equal
to
zero,
we have
y*+xy+x
i
+m(x
+
y)=Q
.....................
(i).
So
also,
since
y^z,
z*
+
yz+y*+m (z
+
y)
=
.....................
(ii).
From
(i)
and
(ii)
we have
by
subtraction
Hence,
as x
4=
z,
we have
x
+
y
+
z
+ m=Q
..............................
(iii).
Substitute
-
(x+y +z)
for m in
(i)
;
and we have
x*
+
xy +y*- (x
+
y) (x +y
+
z)
=
Q;
.:
yz
+
zx
+
xy=0
...........................
(iv).
Then
y
3
+
z
3
+
m
(j/
2
+
z
2
) =y3
+z
3-^
+
z
^ (x
+
y
+
z)
from
(iii)
zx)
=
2xyz
froni.(iv).
190
EXAMPLES.
Ex. 3. Shew
that,
if a
+ b
+ c
+
d=Q,
then will
a* +
fc<
+ c
4
+
d*
=
2
(aft
-
cd)
2
+
2
(ac
-
6d)
a
+
2
(ad
-
fee)*
+
4fccd.
We have to
prove
that
2a=22a
2
&
2
-8a&cd.
Since
a+b+c+d=0;
we
have, by squaring
and
transposing,
o
2
+
fc
2
+
c
2
+ d
2
=
-
2
(fee
+
co
+
ab
+
ad
+
bd
+
cd).
Hence
by squaring
Now
(Sic)
2
=
2b
2
c
2
+ Gafcca
1
+
2fccd
(6
+
c
+
d)
+
2cda
(c
+
d
+
a)
+
2da6
(d
+
a
+
b)
+ 2a&c
(o
+ 6
+
c)
=
S6
2
c
3
+
6abcd
-
Sabcd.
Hence 2a<
+
22a
2
6
2
=
4Sa
2
6
2
-
8a6cd
;
.-. 2a
4
=
2Sa
2
&
2
-8a6cd.
Ex.4. Prove
that,
if ax +
by
+
cz
=
0,
and
-H
---
h
-
=
0,
then will
x
y
z
From the
given
relations we
have,
as in Art.
143,
o b c
y
z
~
z x~ x
y'
z
y
x z
y
x
Hence
[Art. 113]
each fraction is
equal
to
t
yj \x zj \y xj
z
y
x z
y
x
Hence
(y-z)(z-x)(x-y)
EXAMPLES XVL
1. Shew
that,
if
=
a,
=
b and
=
c : then
y
+ z z +x x +
y
... a b c
will . +
,
j
+ .
=
1.
1+a 1+6 1+c
EXAMPLES. 191
2. Shew
that,
if ax +
by
=
and ex* +
dxy
+
ey 0,
theu
will a*e + b*c
=
abd.
3. Eliminate
x, y,
z from the
equations
y
-
z
_
zx x
y
-
=
Or.
=
0,
--
=
C.
y+z
z+x
4. Eliminate
x,
yt
z from the
equations
y
x z
y
x z
-+-
=
a,
-+^
=
6,
-
+
-
=
c.
x z
y
x z
y
T O
v 1
ON/
11 1
5. If x +
-
1 and
y
+ -=l-
prove
that * +
y
z x
6. Eliminate x from the
equations
a 4- c
=
dx,
x
d
;
a
-
c
=
--
ox.
x
7. Eliminate
x, y,
z from the
equations
x*
yz=a, y* zx=b,
z
9
xy
=
c,
ax +
by
8. Prove that the
equations
a: +
y
+ a
=
a,
do not
give any
roots,
but
simply
a relation between
a,
b and c.
9. Shew
that,
if
bz+cy
=
cx + az
=
ay
+
bx,
and x* +
y*
+ 2*
2^3
-
2zx
-
2xy
=
;
then will a
=t
b c
=
0.
10. Shew
that,
if
-
+^
+-
=
1 and
-
+
-
+- =0: then
a b c x
y
z
11. Ifa: + -=w +
-
=
2 +
-;
then
y
z ae
192
EXAMPLES.
12. Shew
that,
if x
=
cy
+
bz, y
=
az + cx uuu s
=
then
13. Shew
that,
if x
3
=y
t
+ z* +
2ayz, y*
=
z
2
+ x* 4- 26sa; and
=
x* +
y
3
+
2cxif ;
then
1
-
a' 1-b
3
1
-
c-'
'
14. Shew
that,
if
x, y,
z be
unequal,
and
a + bz a + bx
y
=
T >
z
=
i~
anti 3
=
r^ >
c+efe c + oa; c +
dy
then will ad + 6c + 6* + c*
=
0.
15. Eliminate
x, y,
z from the
equations
a?
yz
. V* zx z*
xy
-+^=J, +-5
=
m,
+_
=
n.
yz
or zx
y* xy
z*
16. Eliminate
x, y,
z from the
equations
bx* + Ix + c
=
0,
cy*
+
my
+
a
=
0,
oz* + nz + b
=
0, xyz
=
1.
17. Eliminate
x, y,
z from the
equations
y
9
+ z*
=
ays,
z* + x*
=
bzx,
x* +
y*
=
cxy,
xyz
not
being
zero.
18. Eliminate
(i)
x, y,
z and
(ii)
a, b,
c from the
equations
,y
2 z 03,
,
a;
,
w
o-4-c-=a,
c-4-a-
=
o. and a
-
+ 6
-
=
c.
z
y
x z
y
x
19. Eliminate
x,
y,
z from the
equations
ax +
yz
=
be,
by
+
zx-ca, cz+xy
=
ab,
and
xyz
=
abc.
20. Eliminate
x, y,
z from the
equations
x*
xy
xz_
y* -yz-yx
_z*-zx-zy
abc
and ax +
by
+ cz
=
0.
EXAMPLES.
103
21. From the
equations a?yz
=
a*
(y
+
z)
3
,
b*&x
=
fi
2
(z
+
x}*,
c'xy
=
y* (x
+
y)
3
,
deduce the relation
abc
_
a
3
b* c
2
aySy
a
8
;8
2
y'
22. Prove
that,
if
?/
2
+ 2
2
-f
yz
=
a*,
z* + zx + x*
=
b
2
,
x* +
xy
+
y*
=
c
s
and
yz
+ zx +
xy
=
;
then will a 6 c
=
0.
23. Prove
that,
if
-
+
r
+
=
-. r r
,
then will
a b c
(a
+ b +
c)
111
1
a
sit-t 2+
c
zm
(a
+ 6 -(-
c)
2" +
where n is
any positive integer.
24. Shew
that,
if
b
s
+ c*
-
a
3
c* + a*- b
2
a
2
+ b
2
-
c
8
+
- -
+
S-T
=
1
2bc 2ca 2ab
then
(b
+ c
-
a) (c
+ a
-
b) (a
+ b
-
c)
=
0,
and
/6' + c'-aV*
+
/
r
c +a
'_yy
+ '
/q
2
+ 6
a
-
.
V
26c
V
*
\
2c
/
^
V
2a6
25. If oV+6 + cV
=
0,
and a
2
=
6
s
=
c
2
:
a:
y
prove
that V+ 6
4
y
3
+ cV
=
0,
and oV+
b'y
3
+ c
e
z
3
=
aV +
by
+ cV.
no T.T
a
2/
s asa; aa;
y 1
26. If x
-V
=
y
--
-j-
=2
--
-f
,
and
x, y,
z be
unequal;
ar
y^
z
then each member of the
equations
is
equal
to x +
y
+ z a.
(z
xY
27. If
x, y,
z be
unequal,
and if 2a
3y
=
-
and
2a
-
3z
=
(
X
V<L
,
then will 2a
-
3x
= fe f
,
and
x
x +
y
+ z
=
a.
13
194
EXAMPLES.
28. If x + -s
-
;
-
. be not altered in value
by
inter-
x* +
y"
+ z*
changing
x and
y,
it will not be altered
by interchanging
x and
z,
and it will vanish if x +
y
+
z=I,
the letters
being
all
unequal.
29. If
x,
y,
z be
unequal,
and
y*
+ z
8
+
m
(y
+
z)
=z* + x
3
+ m
(z
+
x)
=
x* +
y*
+ m
(x
+
y),
then each will
equal 2xyz.
30. If
x,
y,
z be
unequal,
and
y*
4- z
8
+
myz
=
z*
+
x* 4- mzx
=
x* +
y*
+
mxy,
then each will
equal ^(x*
+
y*
+
z*).
31. If
x, y
be
unequal,
and if
(**~y-*?
=
(*?
?
x
y
/22 x
y)*
then will each
equal
--ZL
.
z
32. Shew
that,
if
a, 6, c,
d be all real
quantities
not
zero,
and
(a
9
+ b
s
) (c
s
+
d*)
=
kabcd : then will a
=
b and c
=
d,
33. If
a, 6, c,
x be all real
quantities,
and
c b
then
r
=
-
=
x.
b a
34. Shew
that>
if
(a?
+
y
1
+ z*
)
(a
9
+ b' +
c*)
=
(ax
4-
by
+
cz}\
then
x/a
=
y/b
=
z/c.
35. Prove the
following
:
(i)
If 2
(a*
+
b') =(a
+
b)',
then a
=
b.
(ii)
If 3
(a
9
+ b' + c
1
)
=
(a
+ b
+
c)*,
then a= b
=
c.
(iii)
If 4
(a*
+ 6' + c
f
+ d*
)
=
(a
+ b + c +
d)',
then
a=b
=
c
=
d.
and
(iv)
If n
(a*
+ b' + c' + ......
)
=
(a
+
6 + c+ ......
)*,
then
a
=
6
=
c
=
......
,
7i
being
the number of the letters,
EXAMPLES. 195
36. Prove
that,
if
a, b, c,
d be all real and
positive,
and
a"
+ b* + c* + d*
=
4:abcd
;
then will a
=
b
=
c
=
d.
37. If
(n-
1)
x" +
2x
(al
-
a
m)
+
a,
2
+ 2a
8
8
+
2a
3
*
+ ... 4-
2a*_
1
+a*
=
2
(a^
+
a
a
a
3
+ ...... + a
n -iO
for real values of
x,
a,,
a
s
, ...,
a
m
;
then will
a
t a,
=
a
g
a
z
=
...
=
a
B _,
=
*.
"Verify
the
following
identities
:
38. a
8
(b
+
c)
+ b
3
(c
+
a)
+ c
8
(a
+
b)
+ abc
(a
+ b +
c)
=
(a
3
+ b* + c
9
)
(be
+ ca +
ab).
39.
(b
+
c-a-d)*(b-c)(a-d)
+
(c+a-b-dy(c-a)(b-d)
+
(a
+
b-c-d)
4
(a-b)(c-d)
=
16
(6 -c) (c
-
a) (a -b) (d-a) (d-b) (d
-
c).
40. 8
(
a + 6 +
c)
3
-(6
+
c)
8
-(c
+
a)
3
-(a
+
6)
8
=
3
(2a
+ 6 +
c) (a
+ 26 +
c) (a
+ & +
2c).
41.
(a+
& + c +
d)
5
-
(b
+c.+
d)'-(c
+ d +
a)
B
-(d+a
+
&)
6
-
(a
+ b +
c)
5
+
(6
+
c)
8
+
(c
+
a)
8
+
(a
+
b)>
+
(a
+
d)
s
+
(b
+
d)
6
+
(c
+
dy-a*
-
b>- c
5
-d
5
= GOabcd
(a+b
+ c +
d).
42.
(a
+ b +
c)
3
abc
-
(be
+ ca +
ab)
3
=
abc
(a
3
+
b
3
+ c
3
)
43.
(a
3
+ b* + c
2
)
3
+ 2
(be
+ ca +
ab)
3
-
3
(a
8
+ &
2
+ c
2
) (be
+ca +
ab)
3
=
(a
3
+ b
3
+ c
3
-
3a&c)
2
.
44.
(ca
-
b*)
(ab
-
c
2
)
+
(ab
-
c*) (be
-
a
8
)
+
(be
-
a
1
) (ca
-
b
2
)
=
(be
+ ca +
ab}
(be
+ ca + ab
-
a
2
-
6
2
c*).
132
196
EXAMPLES.
45. 2
(c'
+ ca +
o')
(a*
+ ab + I
s
)
-
(b'
+ be + c
3
)
9
+ 2
(b'
+ be +
c) (c*
+ ca +
a*) -(a'
=
3
(be
+ ca +
ab)*.
46. Shew that
(3a-&-c)
3
+(36-c-a)
8
+(3c-a-&)
3
c-a)(3c-a-6)
=
16
(a
a
+ b*+ c
3
-3obc).
47. Shew that
(na
-b
c)*
+
(nb
c
a)
3
+
(nc
-a
b)*
3
(na
-
b
c) (nb
c
a) (nc a-b)
=
(n+ I)
2
(n
-
2) (a
3
+ b
3
+ c
3
-
oabc).
48. Shew that
(x
s
+
2yz)
a
+
(y
+
2zx)
a
+
(*
+
2a#)
3
-
3
('
+
2y) (y
9
+
2ac) (z*
+
2xy)
=
(a;
3
+
T/
S
+ z
3
-
Sa^z)'.
49. Shew that
(5y
+
azf
+
(bz
+
oxf
+
(bx
+
ay)
8
-
3
(by
+
oz) (bz
+
ox)
(bx
+
ay)
=
(a
8
+ b
3
) (x*
+
y
3
+ z
3
-
3xyz).
50. Shew
that,
if 1 + o> + o>*
=
0,
then
[(b -c)(x-a)
+
<a(c-a)(x-b)
+
<a*(a -b)(x- c)]
1
+
[(b -c)(x-a)
+ to*
(c
-
a)
(a
-
b)
+ a>
(a
-
6) (x
-
c)]
8
=
27
(b- c) (c- a) (a
-
6) (x -a)(x- b) (x-c).
51. Shew that the
product
of
any
number of
factors,
each
of which is the sum of two
squares,
can be
expressed
as the
sum of two
squares.
52.
Yerify
the
identity
(a*
+ 6* + c
f
+
cF) (p*
+
(f
+ r*+
*)
=
(ap
+
fy
+ cr +
ds)'
+
(aq -bp
+ C8
drf
+
(ar
bs
cp+ dq)*
+
(as
+ br
cq
dp)*.
Hence she\\ that the
product
of
any
number of
factors,
each of which is the sum of four
squares,
can be
expressed
as
the sum of four
squares.
EXAMPLES. 197
53. Shew that
(x
s
+
xy
+
y
2
) (a"
+ ab + b
2
)
can be
expressed
intheform
Z' + -XT + Y*.
1
54. Shew that
(x*+pxy
+
qy*) (a
2
+pab
+
qb*)
can be ex-
pressed
in the form A'
2
+pXY
+
qY*.
55. Shew
that,
if 2*
=
a + b +
c,
(i)
a
(s
-
b) (s
-
c)
+ b
(s
-
c) (s
-
a)
+ c
(s
-
a) (
-
b)
+ 2
(s a) (s b) (s
-
c)
=
abc.
(ii) (s
-
a)
3
+
(s~ b)
3
+
(s
-
c)
3
+ 3abc
=
s
3
.
(iii) (b
+
c)
s
(*
-
a)
+ a
(s b) (s c)
-
2sbc
=
(c
+
a)
s
(s
-
b)
+ b
(s
-
c)
(s
a)
2sca
=
(a
+
b)
s
(s
-
c)
+ c
( a) (s 6)
-
2sab.
(iv) a(b-c)(s-a)*
+
b(c-a)(s-by
+
c(a-6)(s-c)
2
=0.
(v)
s
(s b) (s
-
c)
+ s
(s c) (s
-
a)
+ s
(s
-
a)
(s
-
b)
(s a) ( b) (s c)
=
abc.
(vi) (s
-
a)
2
(s
-
b)
s
(s
-
c)
2
+
s*(s- b)
3
(s
-
c)
8
+
s*(s- c)' (s
-
a)
2
+ s*
(s
-
a)' (s
-
b)
2
+
s(8-a)(s-b)(s- c) (a*
+ b* + c
2
)
=
a*b
s
c*.
56. Shew
that,
if 2s
=
a + b + c +
d,
Shew also that
a
(s
-
b) (s
-
c)(
-
d)
+ b
(s
-
c) (s
-d)(s-a)
+
c(s-d) (s-a) (s-b)
<-
+d(s-a)(8-b)(s-c)
+
2(s-a) (s-b) (s-c) (s- d)
s
(bed
+ cda
+ dab +
abc)
=
2abcd.
57. Shew
that,
ifa+& + c + c
=
0,
then
ad
(a
+
df
+
bc(a- d)
3
+ ab
(a
+
b)'
+
cd(a- b)*
+
ac(a
+
c)*
+ bd
(a
-
c)"
+ knbcd
=
0.
198 EXAMPLES.
58. Shew
that,
if
(a
+
b) (b
+
c) (c
+
d) (d
+
a)
=
(a
+ b + c
+
d) (bed
+ cda + dab +
abc)
;
then ac
=
bd.
59. Shew
that,
if a + 6 + c
=
and x +
y
+ z
=
0,
then M
4
(ax
+
by
+
cz)* -3(dx+by
+
cz) (a"
+ b' + c
8
) (a?
+
y
2
+ z
3
)
-
2
(b
-
c) (c
-
a) (a
-
b) (y -z)(z-
x)(x -y)
=
54abcxyz.
60. Shew
that,
ifa + 6 + c
=
0;
then
(iii)
a
9
+ b
9
+ c
6
=
3
2
6V+
(a
9
+ b* +
c*)
3
.
(iv)
25
(a
7
+ b
7
+ c
7
) (a
3
+ b
3
+ c
8
)
=
2 1
(a
s
+ 6
s
+
c
s
).
61. If a + b + c + d
=
0, prove
that
(a
8
+ b
a
+ c
3
+ d
3
)
2
=
9
(bed
+ cda + dab +
abc)'
=
9
(be- ad) (ca
-
bd) (ab
-
cd).
62. Shew
that,
if a + b + c
=
0,
then
/b-c c a a
b\/a
b c \
(
+
j
+
-
r
+ +
-
,
)
=
9.
\a
o c
/ \6
c ca a
o/
63. Prove
that,
if
J_
+
m
7
+"
m
~i,
e
<o)
-
1+t + m 1+m + mt 1 +
n +nm
AtA*. v\-t^w
i mi 1
.\/W-
and A
N
, =
j-
+ =
j
+ r
=
1,
Q} 1+i + m l+m + ml
* j -
and none of the denominators be
zero,
then will l
=
m
=
n.
61 Shew that
+ ...
=
1
-(!-)(!-
b)
(l-c)(l-d)
EXAMPLES. 199
65. Shew that
1=
1 + 2
(1
-
a)
+ 3
(1
-
a) (1
-
2a)
+ ...
+
{n(l-a)(l-2a)... (1-ro-la)}
+
l{(l_a)(l-2a)...(l-wa)}.
66. Shew that
a" + a'-
1
(1
-
a")
+ a'"
2
(1
-
a") (1
-
a"'
1
)
+ ...
+
{a (1
-
a') (1
-
a"
~
l
)
. ..
(1
-
a
8
)}
+
{(1
-
a") (1
-
a"'
1
)
...
(1
-
a)}
=
1.
67. Shew
that,
if n be
any positive integer,
l-a-
0-0(1-0
'
(1
-.0(1-0(1-Q
1-a l-a
8
l-a
3
(l-a-) (1-a"'
1
).. .(l-a)
4-
S ^
:
^
3
'
=
n.
l-a
68. Prove
that,
if
a + b + c + d=
0, 1/16,
-ty*w
*
+
y
+ Z +
u=0,
Y*
a
-^
and a* +
&y
+ cz + du
=
;
then 2
(a
4
a? +
b*y
+ c'z +
d*u)
<*
a>fae+*6$2&-&b
=
(a*x
+
b'y
+ c'z +
d*u) (a
3
+ b* + c' +
d*).
69. Prove
that,
if n be
any positive integer,
1-1 +
1- 1
+
.l._J L_
h
l
*
9*^^r
^
<-
"^
rt~' i
2n + 1 n + 2 2rt
70. Prove
that,
if
111 1 1 11
-
+
-
=
+
=
+
=
-
;
u v u+a vb u+a vb
f
then
/
2
(ab
1
-
a'bf
=
aa'bb'
(a
-
a') (b
-
b').
*w
^
+
oJ(
s
CXA<
-W
CHAPTER XIIL
POWERS AND ROOTS. FRACTIONAL AND NEGATIVE
INDICES.
159. THE
process by
which the
powers
of
quantities
are obtained is often called
involution;
and the inverse
process, namely
that
by
which the roots of
quantities
are
obtained,
is called evolution.
We
proceed
to consider some cases of involution and
of evolution.
160. Index Laws. We have
proved
in Art.
31,
that
when m and n are
any positive integers,
a
m
xa
n
=
a
mtn
(i).
This result is called the Index Law.
From the Index
Law,
we have
a
m
x a" x a?
=
a
m+*
x a*
=
a"*"**,
and so
on,
however
many
factors there
may
be.
Hence a"x a" x a" x ...
=
a"
rH"
T
*
(ii).
Thus the index
of
the
product of any
number
ofpowers
of
tiie same
quantity
is the sum
of
the indices
of
the
factors.
Also,
a
m
x a* x a" x . . . to n factors
=
(j
m
"
H""1"'"'1
"
to n terms
POWERS.
201
Hence
(a
m
)"
=
a
mn
........................
(iii).
Thus,
to raise
any power of
a
quantity
to
any
other
power,
its
original
index must be
multiplied by
the index
of
the
power
to which it is to be raised.
Again,
to find
(ab}
m
.
(ab)
m
=
ab x ab x ab x ...... to m
factors,
by
definition,
=
(a
x a x a ...... to m
factors)
x
(b
x b x b ...... to
in
factors), by
the Commutative
Law,
=
a
m
x b
m
,
by
definition.
Hence
(ab)
m
=
a
m
x b
m
.
Similarly (abc...)
m
=
a
m
x b
m
x c
m
x ......
(iv).
Thus,
the mth
power of
a
product
is the
product of
the
mth
powers of
its
factors.
The most
general
case of a monomial
expression
is
Now
(aW
......
)
m
=
)
m
(&)"
(c'F
...... from
(iv)
=
a
xm
b
vm
c*
m
............ from
(iii).
Hence
(aW
......
)
m
=
a*
m
b
vm
c*
m
. . . . .'.
(v).
Thus
any power of
an
expression
is obtained
by taking
each
of
its
factors
to a
power
whose index is the
product of
its
original
index and the index
of
the
power
to which the
whole
expression
is to be raised.
As a
particular
case
a
=
a X =-
161. It follows from the Law of
Signs
that all
powers
of a
positive quantity
are
positive,
but that successive
powers
of a
negative quantity
are
alternately positive
and
negative.
For we have
202
ROOTS OF ARITHMETICAL
NUMBERS.
(_a)*
=
(-a) (-)
=
+
a
2
,
(- a)
3
=
(- a)
8
(-
a)
=
(+
a
2
) (- a)
=
-
a
1
,
(_ )
=
(_
a
)
3
(- a)
=
(-
a
8
) (- a)
=
+
a\
and so on.
Thus
(- a}**
=
+
a
8
",
and
(- a)
2"
+1
=
-
a
2im
.
Hence aW even
powers,
whether
of positive
or
of
negative quantities,
are
positive;
and all odd
powers of
any quantity
have the same
sign
as the
original quantity.
162. Roots of Arithmetical numbers. The
approximate
value of the
square
or of
any
other root of
an arithmetical number can
always
be found : this we
proceed
to
prove.
It will be seen that the
process
described would be an
extremely
laborious
one;
we are
not however here concerned with the actual calculation
of surds.
Consider,
for
example,
$62.
First write down the
squares
of the numbers
1, 2, 3,
&c. until one is found
which is
greater
than
62 : it will then be seen that 7*
is
less and 8
s
is
greater
than 62. Now write down the
squares
of the numbers
7'1, 7'2,
7
'3, ...,
7'9 : it will then
be seen that
(7'8)
2
is
less,
and
(7'9)
a
greater
than 62.
Now write down the
squares
of 7
'81, 7*82, ...,
7 '89 : it will
then be seen that
(7'83)
2
is
less,
and
(7 '84)* greater
than 62.
By continuing
this
process,
we
get
at
every stage
two
numbers such that 62 is intemiediate between their
squares,
and such that their difference becomes smaller
and smaller at
every
successive
stage;
moreover,
this
difference
can,
by sufficiently continuing
the
process,
be
made less than
any assigned quantity
however small.
Thus,
although
we can never find
any
number whose
square
is
exactly equal
to
62,
we can find two numbers
whose
squares
are the one
greater
and the other less than
62,
and whose difference is less than
any assigned quantity
however small. The
limiting
value of these two
numbers,
SURDS OBEY FUNDAMENTAL LAWS. 203
when the
process
is continued
indefinitely,
is called the
square
root of 62.
The
process
above described for
finding
a
square
root
can
clearly
be
applied
to find
any
other root.
Thus an nth root of
any integral
or fractional number
can
always
be found.
163. Surds
obey
the Fundamental Laws of
Algebra.
The fundamental laws of
Algebra
were
proved
for
integral
or fractional values of the letters
;
and it can
be
proved
that
they
are also true for surds.
Consider,
for
example,
the Commutative Law.
We have to
prove
that
#ax 7&=7&x j/a.
We can find whole numbers or fractions
x,
y
and
p, q
such that
x>%a>y,
and
p>^/b> q;
and the difference between x and
y,
and also the difference
between
p
and
q,
can be made less than
any assigned
quantity
however small.
Hence x x
p
>
\]a
x
m
jb
>
y
x
q,
and
p
x x >
tyb
x
^Ja >qxy.
But,
since
x,
y, p, q
are
integral
or fractional
numbers,
we know that x x
p
=
p
x
x,
and
y
x
q
=
q
x
y
;
also the
difference between
px
and
qy
can be made less than
any
assigned quantity
however small.
It therefore follows that
/a
x
tyb
and
"Zjb
x
%/a,
which
are both
always
intermediate to
xp
and
yq,
must be
equal.
Thus the Commutative Law holds for
Surds,
and the
other laws can be
proved
in a similar manner.
164. We
already
know that there are two
square
roots,
and three cube roots of
every quantity
;
and we
may
remark that there are
always
n nth roots. Thus there is
204
PROPERTIES OF ROOTS.
an
important
difference between
powers
and roots
;
for
there is
only
one nth
power,
but there is more than one
nth root.
165. We have
proved
in Art. 160 that the mth
power
of a
product,
is the
product
of the mth
powers
of its
factors;
and,
since surds
obey
the fundamental laws of
Algebra,
the
proposition
holds
good
when all or
any
of the factors are
irrational. Hence
Also
(Jab...y
=
ab...,
by
definition.
Hence
\/a
x
*Jb...
must be
equal
to one of the
square
roots of ab... .
We can write this
Ja\/l>...
=
Jab...
,
meaning thereby
that the continued
product
of either of
the
square
roots of
a,
either of the
square
roots of
b,
&c. is
equal
to one or other of the
square
roots of ab ...
Similarly
we
have,
with a
corresponding
limitation,
...,
and
j/a
"/a
=
V &'
Also a
m
=
Xa"*,
for their
npth powers
are both
equal
to a
mp
.
Again,
since the nth
power
of a monomial
expression
is
obtained
by multiplying
the index of each of its factors
by
n,
it follows
conversely
that an nth root of a monomial
expression
is obtained
by dividing
the index of each of its
factors
by
n,
provided
the division can be
performed.
Thus one value of
\fa
4
is a
8
,
one value of
If
a
9
b
9
c
8
is
a
2
b
3
c,
and one value of
^/a"*
b** c"
Y
is a* W c*.
EXTENSION OF MEANING OF INDEX.
205
Fractional and
Negative
Indices.
166. We have hitherto
supposed
that an index was
always
a
positive integer
;
and this is
necessarily
the case
so
long
as we retain the definition of Art. 9
; for,
with that
definition,
such
expressions
as a and a~
2
have no
meaning
whatever.
We
might
extend the
meaning
of an index
by assign-
ing meanings
to a" when n is fractional and
negative.
It
is, however,
essential that
algebraical symbols
should
always obey
the same laws whatever their values
may
be;
we therefore do not
begin by assigning any meaning
to a"
when n is not a
positive integer,
but we first
impose
the
restriction that the
meaning of
a* must in all cases be such
that the
fundamental
index
law,
namely
a x a"
=
a
ro+n
,
shall
always
be true
;
and it will be found that the above
restriction is of itself sufficient to define the
meaning
of a"
in all
cases,
so that there is no further freedom of choice.
For
example,
to find the
meaning
of
a.
Since the
meaning
is to be consistent with the Index
Law,
we must have
Thus or must be such that its
square
is
a,
that is a
3
must be
\/a.
Again,
to find the
meaning
of a"
1
.
By
the index law
a
-j
a"
1
x a
3
=
a~
m
=
a
18
;
therefore a"
1
=
,
=
-.
a
3
a
Thus a"
1
must be
-
.
a
206 FRACTIONAL AND NEGATIVE INDICES.
167. We now
proceed
to consider the most
general
cases.
i
L To find the
meaning
of
a",
where n is
any positive
integer.
By
the index
law,
111
a" x a" x a" x to n factors
i+-+-+
to n terms
-
=
a*
=
a*
=
a
=
a.
Hence a" must be such that its nth
power
is
a,
that is
a^
=
^a.
m
II. To find the
meaning
of
a",
where m and n are
any
positive integers.
By
the index
law,
P
.
;-
+ +...., to terms
-
x
a* x a
n
x
to n factors
=
a
n n
=a* =a .
Hence a"
=
"Jo,"*.
We have also
11
i.i.
, r+~+
to term
a
n
x a* x to m factors
=
a*
=
a*.
Hence a"
=
(a")".
Thus we
may
consider that a" is an nth root of the
mth
power
of
a,
or that it is the mth
power
of an nth root
of a
;
which we
express by
With the above
meaning
of o" it follows from Art.
165 that a"
=
a
HI>
.
FRACTIONAL AND NEGATIVE INDICES.
207
NOTE. It should be remarked that it is not
strictly
true that
Z/(a
m
)
=
(^a)"' except
with a limitation corre-
sponding
to that of Art.
165,
or unless
by
the nth root of
a
quantity
is meant
only
the arithmetical root. For
example, fy(a*)
has two
values,
namely
a
2
,
whereas
( 4/a)
4
has
only
the value
+
a*.
III. To find the
meaning
of a.
By
the index law
v
_m
/,0+m f,m
.
f.0
.
f,i -|
a x a a a-
,
. . &
=
& -r a
=
i.
Thus a
=
1,
whatever a
may
be.
IV. To find the
meaning
of a"
1
",
where m has
any
positive
value.
By
the index
law,
a~
m
x a
m
=
a-'
n+m
=
a
;
and a
=
1,
by
III.
Hence a~
m
=
-=.
,
and a'"
=
-==.
a
m
of*
168. We have in the
preceding
Article found that in
order that the fundamental index
law,
a
m
x a"
=
a"'
+n
,
may
always
be
obeyed,
a
m
must have a definite
meaning
when
n has
any given positive
or
negative
value. We have now
to shew
that,
with the
meanings
thus
obtained,
a
1"
x a
n
=
a**",
(a
m
)
n
=
a
m
",
and
(ab)
n
=
a"6",
are true for all values of m and n. When these have been
proved,
the final result of Art. 160 is
easily
seen to be true
in all
cases.
208 INDEX LAWS.
I. To
prove
that a
m
x a"
=
a"
1
*",
for all values of m
and ??.
We
already
know that this is true when m and n are
positive integers.
Let m and n be
any positive
fractions
*-
T
and
-
respectively.
Then
? ?!
a
m
x a"
=
a* x a*
=
^
a* x
J/a
r
.
by
definition
x rf
=
Stf
[Art. 165]
rg
*=
a
'*
,
by
definition
+L
=
a
11
*
=
a .
Thus the
proposition
is true for all
positive
values of
m and n. To shew that it is true also for
negative
values,
it is
necessary
and sufficient to
prove
that
a"
1"
x cf
=
a"*"",
and a
m
x a""
=
a'"""
where m and n are
positive.
Now flf* x cf"
=
x
=
-^
=
a~"'~".
a
m
a" a
'
And,
if m w be
positive,
a
m~"
x
a*
=
a'",
and a"
1
x a~* x a"
=
a*"
;
therefore
a"'~
n
=
a
m
x a~".
Hence,
if m 71 be
negative.
-=-.
x
=
- -
,
a a a
that
is,
a
m
x a""
=
a
1
""".
Hence a
1"
x a"
=
a
m+
*,
for aW values of m and n.
COB. Since a"" x a"
=
a"' for all values of m and
n,
it follows that a"'
-=-
a"
=
a
w
"".
INDEX LAWS. 209
II. To
prove
that
(a"')"
=
a
mn
,
for all values of m and n.
First,
let n be a
positive integer,
m
having any
value
whatever.
Then
(a
w
)
n
=
a
m
x a
m
x a
m
x to n
factors,
__ in+ m+m+ to n terms
-
Next,
let n be a
positive
fraction
-
,
where
p
and
q
are
positive integers.
Then
(a
m
)
n
=
(a'7=#{(aT}>
=
U(a<
np
),
since
p
is an
integer,
Finally,
let w be
negative,
and
equal
to
p.
Then
(a-)"
=
(a
m
p
=
^
=
^
=
a-
=
a
m
".
Hence for all values of m- and n we have
(a
m
)
n
=
a
mn
.
III. To
prove
that
(ab)
n
=
a
n
b
n
,
for all values of n.
We have
proved
in Art. 160 that
(ab)
n
a
n
b
n
,
where
n is a
positive integer.
And,
whatever m
may
be,
provided
that
q
is a
positive
integer,
we have
(a
m
b
m
)
9
=
a
m
b
m
x a
m
b
m
x ... to
q
factors
_
Qin+m+...
to
q
terms
v
hn+m+ ... to
9
terms
Let n be a
positive
fraction
*-,
where
p
and
q
are
positive integers.
Then
s. A.
14
210
RATIONALIZING FACTORS.
(afy*
=
(a&)
=
*J(dbY
=
y(a
p
V),
since
p
is a
positive
integer.
Also
(a"6")*
=
a"
7
6"
?
,
since
q
is a
positive integer.
Hence a"6"
=
#(a"&
p
)
=
(06)".
Thus
(a&)"
=
a"6",
for all
positive
values of n.
Finally,
if n be
negative,
and
equal
to
m,
we have
(a&)"
=
(06)-
=
7-^
=
^
-
a-fc-
1"
=
a"b".
(ab)
m
a
m
b'
Ex.
(i) Simplify
a^ x a
~
i.
Ex.
(ii). Simplify
a*6 x a?b*.
a*b$ x a^&S
=
a?
+
s
6^
+
f=
a*b*s
=
a-l*.
Ex.
(iii). Simplify (
a
-
2
&*)~*.
(a-*6t)-!
=
a
(-)(-?)
6
i(-)
=a fc-
3=
!.
Ex.
(iv). Simplify -J(a
~
*
5c
~
*)
-f-
v^(a^
ft^-
1
).
169.
Rationalizing
Factors. It is sometimes re-
quired
to find an
expression
which when
multiplied by
a
given
irrational
expression
will
give
a rational
product.
The
following
are
examples
of
rationalizing
factors.
Since
(a
+
Jb) (a
-
^/fc)
=
a*
-
6,
it follows that
a^/&
is made
rational
by multiplying by
o=F>/6.
So also
ajbcjdis
made rational
by multiplying by
a
*Jb
T
c
*Jd.
Again
from the known
identity
2&
8
c +2c
j
a
a
+
2a
8
6
-
o*
-
fe
-
c*
=
(a
+
6
+
c) (
-
a +
6
+
c) (a
-
6
+
c)
(a
+
6
-
c),
RATIONALIZING FACTORS.
211
it follows that the
rationalizing
factor of
<JP
+
N/2
+
J
r is
(
~
The
rationalizing
factor of
/Jp
+
ijq+^r
may
also be found as
follows,
and
Thus the
required rationalizing
factor is
(JP
+
Jq
~
\/
r
) (P
+
q-r-
2
which is the same as before.
Again,
from the
identity
the
rationalizing
factor of a
+
b* is seen to be a*
-<$
170.
To
find
the
rationalizing factor of any
binomial.
p
r
.
Let the
expression
to be rationalized be ax
9
by'.
t
Put X=ax
9
,
and
Y=by',
and let n be the L.C.M. of
q
and s.
Then it is
easily
seen that X
n
and Y" are both
rational.
Hence,
from the identities
and
(X- 7) (Z"-
1
+
Z"-
2
F-f ......
+
F"'
1
)
=
X"
-
F",
the
rationalizing
factors of X
+
Y and X Y are seen to
be
respectively
X*~
l
-X"-*Y+......
+
(-
1
)"-'
F""
1
,
and X
n
-
1
+
X
n
~*Y+
......
+
F""
1
.
142
212
EXAMPLES.
Ex. To find a factor which will
rationalize
x*
-
ay*
.
Here
X=x$,
Y=ay%,
n=6.
The factor
required
is therefore
*
^
+
ax
t
y
I
+
cP&y
%
+
a?x%
y
$
+
a
4
x%
y
^
+
a
"y
EXAMPLES
XVIL
1.
Simplify
a$b x
a~^b~^.
2.
Simplify
a$ x a~^ x
(a*)~*
x
-
(a-^)
3.
Simplify
(oi'V)*
x
(a'd'c'^.
b+e 1 c+a 1 +b
J
4.
Simplify
(x
c
-
a
)'-
6
x
(-&)&
-
x
(a;
6
^*)*-
5.
Multiply r+J^g^
+
jr by x^-y^.
6.
Multiply
a
2
+ 1 + x~*
by
x*
-
1 + a;'
1
.
7.
Multiply
x +
8. Divide x-2 + x~
by
x-aj~.
9. Divide a?
-
x
by
a"
-
a
10. Divide a;*
-
xy^
+
x^y
-
y* by a*-?/*.
11. Shew that
x$
-
4a;* + 2x1 + 4a;
-
4o;& + x^
=
(a;*
-
2a;i +
12.
Multiply
4x
3
-5;c-4-7ar
l
+ 6ar
by
and divide the
product by
3x 10 + 10x~'
-
4a;~
s
.
EXAMPLES. 213
13. Divide
x
-
x~
'
-
2
(*
-
aT)
+ 2
(
-
x~
) by
x"
-
aT *.
a
lI
14.
Simplify
-f a
_ _
T^.
.
,
15. Divide
Q
+
^
by
-^
+
^
?/"*
a;'
i/
**
16. Shew that
x
7
-l a;
3
+l x'
3
1
17. Shew that
(2x
+
y~
l
) (2y
+ of
')
=
(2x*y^
+
x~*y~*y.
18. Shew that
19. Shew
that,
if
ar +
y*
+ z
3
=
0,
then
(x
+
y
+
z)
3
=
27xyz.
,
20. Find factors which will rationalize the
following
expressions
:
(i)
a^ +
b$,
(ii)
tft! +
yi,
(iii)
a -f- bx
3
+ ex
3
,
and
(iv)
x* +
if
+ z*.
21. Shew
that,
if
(1
-
3
)i (y
-
z)
+
(1
-
3,
3
)i (z- x)
+
(1
-
s
s
)i (x
-
y)
=
0,
and
x, y,
2 are all
unequal,
then
CHAPTER XIV.
SURDS. IMAGINARY AND COMPLEX
QUANTITIES.
171. Definitions. A surd is a root of an arithmetical
number which can
only
be found
approximately.
An
algebraical expression
such as
*Ja
is also often
called a
surd,
although
a
may
have such a value that
*Ja
is not in
reality
a surd.
Surds are said to be of the same order when the same
root is
required
to be taken. Thus
\/2
and
\/6
are called
surds of the second
order,
or
quadratic
surds
;
also
4/4
is a
surd of the third
order,
or a cubic surd
;
and
ya
is a surd
of the nth order.
Two surds are said to be similar when
they
can be
reduced so as to have the same irrational factors. Thus
V8
and
t/18
are similar
surds,
for
they
are
equivalent
to
2\/2
and
3\/2
respectively.
The rules for
operations
with surds follow at once from
the
principles
established in the
previous chapter.
Note. It should be remarked that when a root
symbol
is
placed
before an arithmetical number it denotes
only
the arithmetical
root,
but when the root
symbol
is
placed
before an
algebraical expression
it denotes one of the roots.
Thus
ya
has two values but
</2
is
only supposed
to denote
the arithmetical
root,
unless it is written
^2.
SURDS.
215
172.
Any
rational
quantity
can be written in the form
of a surd. For
example,
and a
=
4/a
2
=^a
s
=
#a
n
.
Also,
since
V
a x
v^
=
V
a&
[Art. 165],
we have
2V2
=
V*
x
^2
=
V(*
x
2)
=
V8,
^5
3
x
4/3
=
^(5
3
x
3)
=
#375,
and
a^oF
=
V
a
"
x
^
&
=
v^(a"
x
a&)
=
v'l
Conversely,
we have
^18
=
\/(9
x
2)
=
V9
x
*J2
=
and
+
4/40
=
^(3
3
x
5)
+
4/(2
3
x
5)
=
3^5
-f
2^/
173.
Any
two surds can be reduced to surds of the
same order. For if the surds be
f/a
and
m
jb,
we have
%a
=
"7a
m
,
and
76
=
n
^6" [Art.
165].
Ex. Which is the
greater,
4/14
or
4/6?
The surds must be reduced to
equivalent
surds of the same order.
Now
4/14
=
4/14
2
=
4/196,
and
4'6
=
4/6
3
=4/216.
Hence,
as
4/216
is
greater
than
4/19(>i 4^6
must be
greater
than
4^14.
Thus we can determine which is the
greater
of two surds without
finding
either of them.
174. The
product
of two surds of the same order can
be written down at
once,
for we have
^/a
x
tyb
=
%/al>.
Hence,
in order to find the
product
of
any
number of
surds,
the surds are first reduced to surds of the same
order : their
product
is then
given by
the formula
%/a
x
yb
x
%/c.
. .=
%/abc.
. .
Ex. 1.
Multiply ^5 by 4/2.
<s/6
x
4/2
=
4/53
x
4/2
2
=
4/(5
3
x
22)
=
4/500.
Ex. 2.
Multiply 3^/5 by
24/2.
3
v
/5 x
24/2
=,
3 x 2 x
V5
x
4/2
=
6 x
4/5
3
x
4/2
2
=
6
4/500.
216 MULTIPLICATION OF SURDS.
Ex. 3.
Multiply ^2 by
4/2.
-s/2
x
4/2=4/23
x
4/23=^2^x~2 =4/32.
Ex. 4.
Multiply
^2+^3 by
=/3
x
3+2 xS+3
x
/o
+
/2 x 5
Ex. 5. Divide
4/4 by
4/8.
175. The determination of the
approximate
value of
an
expression containing
surds is an arithmetical rather
than an
algebraical problem
;
but an
expression containing
surds must
always
be reduced to the form most suitable
for
arithmetical calculation. For this reason when surds
occur in the denominators of
fractions,
the denominators
must be
rationalized.
[See
Art.
169.]
The
following examples
will illustrate the
process:
2 2 X 5 2
3
3U/5
+
1)
_8
~
lV
"
176. The
product
and the
quotient
of two similar
quadratic
surds are both
rational.
This is obvious
;
for
any
two similar
quadratic
surds
can be reduced to the forms
a^b
and
c*Jb.
Conversely,
if the
product
of the
quadratic
surds
>Ja,
and
\/b
is rational and
equal
to
x,
we have x=
*Ja
x
V6;
therefore
x\/b
=
\/a
x
Jb
x
\Jb
=
b\/a,
which shews that the
surds are similar. So
also,
if
\/a
-r-
jb
is
rational,
the surds
must be similar.
SURDS.
217
177. The
following
theorem is
important.
Theorem.
If
a
+
<\/b
=
x
+
Vy,
where a and x are
rational,
and
*Jb
and
*Jy
are
irrational;
then will a
=
x,
and b
=
y.
For we have a
x+^b
=
>Jy.
Square
both sides
;
then,
after
transformation,
we have
2
(a x)
V&
=
y
b
(a xf.
Hence,
unless the
coefficient of
\/b
is
zero,
we must have
an irrational
quantity equal
to a rational
one,
which is
impossible.
The coefficient of
*/b
in the last
equation
must there-
fore be
zero,
so that a
=
x. And when a
=
x,
the
given
relation shews that
*Jb
=
\ly,
and therefore b
=
y.
As a
particular
case of the
above,
^a
4=
b
+
Vc,
unless 6
=
and a
=
c.
.
Hence
\/a
+
\/c
can
only
be rational when it is zero.
Ex. 1. Shew that
Ja+^fb
+
Jc^Q,
unless the surds are all similar.
For we should have
/a
+
Jb
=
-
^/c
;
and therefore a
+
b
+
2
Jajb
=
c .
Hence
Jajb
is
rational,
which shews
[Art. 176],
that
fja
and
Ijb
are
similar surds.
178.
The
expressions
a
+
*/b
and a
*Jb
are said to be
conjugate
quadratic
surd
expressions.
It is clear that the sum and the
product
of two
conju-
gate quadratic
surd
expressions
are both rational.
Conversely,
if the sum and the
product
of the
expres-
sions a
+
^/b
and c +
*Jd
are both
rational,
then a
=
c and
*Jb
+
*Jd
=
0,
so that the two
expressions
are
conjugate.
For a
+
c
+
*Jb
+
\jd
can
only
be rational when
>Jb
+
*/d
is zero.
[Art. 177.]
And,
when
*Jd
=
\/6,
the
product
(a
+
V&) (c
+
*Jd}
=
ac
+
(c a)
\/b b,
which cannot be rational unless c
=
a.
179. In the
expression
ax*
+
bx*-
1
+
cx
n~*
+
......
+
k,
where
a, b, c,
...... k are all
rational,
let a-f
yV3
be substi-
218
SURDS.
tuted for
x;
and let P be the sum of all the rational terms
in the result and
Q \//3
the sum of all the irrational terms.
Then the
given expression
becomes P
+
Q V$.
Since P and
Q
are
rational,
they
contain
only squares
and
higher
even
powers
of
\/&
and hence P and
Q
will not
be
changed by changing
the
sign
of
A/$.
Therefore when
a
V/3
is substituted for x in the
given expression
the
result will be P
-
Q V/3.
If now the
given expression
vanish when a +
V/3
is
substituted for
x,
we have
Hence,
as P and
Q
are rational and
\flS
is
irrational,
we must have both P
=
and
Q
=
;
aud therefore
P-QV=0.
Therefore if the
given expression
vanish when a
+
V/3
is substituted for x it will also vanish when a
\//S
is sub-
stituted for x.
Hence
[Art. 88],
if x a
Jft
be a factor of the
given
expression,
x a
+
V/3
will also be a factor.
Thus,
if
a rational and
integral expression
be divisible
by
either
of
two
conjugate quadratic
surd
expressions
it
will also be divisible
by
the other.
180. The
square
root of a binomial
expression
which is the sum of a rational
quantity
and a
quadratic
surd can sometimes be found in a
simple
form. The
pro-
cess is as follows.
To find
V(a
+
^b),
where
^b
is a surd.
Let
Square
both sides
;
then
Now,
since
*Jb
is a
surd,
we can
[Art. 177] equate
the
rational and irrational terms on the different sides of the
last
equation;
hence x
+
y
=
a,
and
4txy
=
b.
SURDS.
219
Hence x and
y
are the roots of the
equation
#
2
ax
+
-
=
0,
4
and these roots are
(a
+
V(a
2
-
6)}
and
J {a
-
V(a
2
-
6)}.
/a
+
V(a
s
-
6) /a
-
Thus
-
It is clear
that,
unless
\/(
a2
~
&)
is
rational,
the
right
side of the last
equation
is less suitable for calculation than
the left. Thus the above
process
fails
entirely
unless
a
8
6 is a
square number;
and as this condition will not
often be
satisfied,
the
process
has not much
practical
utility.
It should be remarked that if x and
y
are
really
rational,
they
can
generally
be written down
by inspection.
Ex.1.
Let
^(6
+
2J5)
=
Jx
+
Jy.
Then,
by squaring,
we have
=
x
+
y
+
2>Jxy-
Hence,
equating
the rational aud irrational
parts,
x +
y
=
6 and
xy
=
5. "Whence
obviously
x
=
1 and
y
=
5. Thus
Ex. 2. Find
^(28
-
5^/12).
Let
^(28
-5
t
jl'2)=
f
jx-^y.
Then,
as
before,
4xy=
25x12,
or
xy
=
75
and+y
=
28;
whence x
=
25 and
y
=
3. Thus
*J (28
-
5
^12)
=
5
-
iJ3. [If
we had taken x
=
3 and
y
=
25 we should have had the
negative
root,
namely
^/3
-
5.]
Ex. 3. Find
>/(18
+
12^/3).
In this case
^/(a
2
-
b)
is irrational and therefore the
required
root
cannot be
expressed
in the form
,Jx
+
>Jy
where x and
y
are rational.
The root can however be
expressed
in the form
j/x
+
j/y;
for
=4/243
+
4/27.
Ex.4. Find
^(10
+
2
^6
+
2^10
+
2^/15).
Assume
^/(
10
+
2^/6
+
2^/10
+
2^/15) ^Jx
+
Jy+Jz;
then 10
+
2^/6
"
"We have now to
find,
if possible,
rational values of
x, y,
z such that
xy
=
Q,
xz
=
10,
3/2=15
and x +
y
+
z
=
10. The first three
equations
are satisfied
by
the values
x=2, y
=
3,
z
=
5,
and these values
satisfy
x
+
y
+z=W.
Hence
220 IMAGINARY AND COMPLEX
QUANTITIES.
Ex. 5. Prove
that,
if
4/ (a +V
6
)
=
x
+
\/2/5
then wiu
j/(a-ijb)=x-jy.
We have a
+
^/fc
=
(a:
+
^y)
8
=
a?
+
3xy
+
Jy (Zx*
+
y).
Hence, equating
the rational and irrational
parts,
we have
a=x
s
+
3ary,
and
Jb=>
Hence
a-*J
EXAMPLES XVIII.
Simplify
the
following
:
""
Tni T
* ^* . _
5"
v
".
v
<O
l/ In . lr\^
/IT f
6.
-77;
7.
J2
+
J3
+
J5-
1
9.
10.
11
45
1L
W^l""^Tl'
U-
39=1
+
13'
1T^+T4-
14
72T^
IMAGINARY AND COMPLEX
QUANTITIES.
221
15.
17.
18.
19.
l/(97
-
56
J3).
20.
21.
22.
24.
V{6
+ 2
^2
+ 2
,/3
+
25.
26.
27. Shew that
1
28. Shew that
1 3
-
2
IMAGINARY AND COMPLEX
QUANTITIES.
181. We have
already
seen that in order that the
formula obtained in Art. 81 for the factors of a
quadratic
expression may
be
applicable
to all
cases,
it is
necessary
to consider
expressions
of the form
J a,
where a is
222 IMAGINARY AND COMPLEX
QUANTITIES.
positive,
and to assume that such
expressions obey
all the
fundamental laws of
algebra.
Since all
squares,
whether of
positive
or of
negative
quantities,
are
positive,
it follows that
<Ja
cannot
represent any positive
or
negative quantity;
it is on this
account called an
imaginary quantity.
Also
expressions
of the form a +
bj
1 where a and b are
real,
are called
complex quantities.
182. The
question
now arises whether the
meanings
of the
symbols
of
algebra
can be so extended as to
include
these
imaginary quantities.
It is clear that
nothing
would
be
gained,
and that
very
much would be
lost,
by extending
the
meanings
of the
symbols, except
it be
possible
to do
this
consistently
with all the fundamental laws
remaining
true.
Now we have not to determine all the
possible systems
of
meanings
which
might
be
assigned
to
algebraical
symbols,
both to the
symbols
which have hitherto been
regarded
as
symbols
of
quantity
and to the
symbols
of
operation, subject only
to the restriction that the funda-
mental laws should be satisfied in
appearance
whatever the
symbols may
mean: our
problem
is the much
simpler
and
more definite one of
finding
a
meaning
for the
imaginary
expression
J
a which is consistent with the truth of all
the fundamental laws.
183. We
already
know that 1 is an
operation
which
performed upon any quantity changes
it into a
magnitude
of a
diametrically opposite
kind.
And,
if we
suppose
that
J
1
obeys
the law
expressed by
1 x
J
1 x
J
1
=
1,
it follows that
<J
1 must be an
operation
which when
repeated
is
equivalent
to a reversal.
Now
any species
of
magnitude
whatever can be re-
presented
by lengths
set off
along
a
straight
line; and,
when a
magnitude
is so
represented,
we
may
consider the
COMPLEX
QUANTITIES.
223
operation
J
1 to be a revolution
through
a
right
angle,
for a
repetition
of the
process
will turn the line in the
same direction
through
a second
right angle,
and
the line
will then be
directly opposite
to its
original
direction.
Hence,
when
magnitudes
are
represented b}'
lengths
measured
along
a
straight
line,
we see that
J 1,
regarded
as a
symbol
of
operation,
has a
perfectly
definite
meaning.
The
symbol
/
1 is
generally
for
shortness denoted
by
i,
and the
operation
denoted
by
i is considered to be a
revolution
through
a
right angle
counter-clockwise,
i
denoting
revolution
through
a
right angle
in the
opposite
direction.
184. It is clear that to take a units of
length
and
then rotate
through
a
right angle
counter-clockwise
gives
the same result as to rotate the unit
through
a
right angle
counter-clockwise and then
multiply by
a. Thus ai
=
ia,
Again,
to
multiply
ai
by
bi is to do to ai what is done
to the unit to obtain
bi,
that is to
say
we must
multiply
by
b and then rotate
through
a
right angle;
we thus
obtain ab units rotated
through
two
right angles,
so that
ai x bi
=
ab
=
abii.
From the above we see that the
symbol
i is commuta-
tive with other
symbols
in a
product.
Since
(ai)
x
(ai)
=
aaii
=
a
2
( 1)
=
a
2
,
it follows
that
>/
a
8
=
ai
;
it is therefore
only necessary
to use one
imaginary expression, namely
J
1.
185. With the above definition of
J
1 or
i,
namely
that it
represents
the
operation
of
turning through
a
right
angle
counter-clockwise,
magnitudes being represented by
lengths
measured
along
a
straight
line,
the truth of the
fundamental laws of
algebra
for
imaginary
and
complex
expressions
can be
proved.
Some
simple
cases have been
considered in the
previous
Article: for a full discussion
see De
Morgan's
Double
Algebra;
see also Clifford's
Common Sense
of
the Exact
Sciences,
Chapter
iv. 12
and
13,
and Hobson's
Trigonometry, Chapter
xin.
224
CONJUGATE
COMPLEX EXPRESSIONS.
186. If a
+
bi
=
Q,
where a and b are
real,
we have
a bi. But a real
quantity
cannot be
equal
to an im-
aginary
one,
unless
they
are both zero.
Hence,
if a
4-
bi
=
0,
we have both a
=
and 6
=
0.
Note. In
future,
when an
expression
is written in
the form a
+
bi,
it will
always
be understood that a and 6
are both real.
187.
If a
+
bi
=
c
+
di,
we have a
-
c +
(b d)i=0;
and
hence,
from Art.
186,
a c
=
and b d
=
0.
Thus,
two
complex expressions
cannot be
equal
to one
another,
unless the real and
imaginary parts
are
separately
equal.
188. The
expressions
a
+
bi and a bi are said to be
conjugate
complex expressions.
The sum of the two
conjugate complex expressions
a
+
bi and a bi is
a+
a
+
(b 6)i
=
2a;
also their
pro-
duct is aa
+
obi abi 6V
=
a
2
+ 6
2
.
Hence the sum and the
product of
two
conjugate complex
expressions
are both real.
Conversely,
if the sum and the
product
of two
complex
expressions
are both
real,
the
expressions
must be con-
jugate.
For let the
expressions
be a
+
bi and c + di. The sum
is a
+
bi
+
c + di
=
a
+
c
+
(b
+
d)i,
which cannot be real
unless 6
+
d
=
0.
Again,
(a
4-
bi) (c
+
di)
=
ac+bci+adi +
bd?
=
acbd+(bc
+
ad)i,
which cannot be real unless be
+
ad 0.
Now,
if 6 + d
=
and also be +
ad
=
0,
we have 6
(c a)
=
;
whence
a
=
c or 6
=
0. If 6
=
0,
d is also
zero,
and both
expres-
sions are real
; and,
if 6
=t= 0,
we have a
=
c,
which with
b
=
d,
shews that the
expressions
are
conjugate.
189. Definition. The
positive
value of the
square
root of
a*+b*
is called the modulus of the
complex
MODULUS OF A COMPLEX EXPRESSION. 225
quantity
a +
bi,
and is written mod
(a
+
bi).
Thus
mod
(a
+
bi)
=-
+
Jo?
+
b*.
It is clear that two
conjugate complex expressions
have
the same modulus
; also,
since
(a
-f
bi) (a bi)
=
a
2
+
6
2
[Art. 188],
the modulus of either of two
conjugate complex
expressions
is
equal
to the
positive square
root of their
product.
Since a and b are both
real,
a
2
-f b* will be zero
if,
and
cannot be zero
unless,
a and 6 are both zero. Thus the
modulus of a
complex expression
vanishes if the
expression
vanishes,
and
conversely
the
expression
will vanish if the
modulus vanishes.
If in mod
(a
+
bi)
=
+
\/a
2
+ 6
2
we
put
6
=
0,
we have
mod a
=
-f
v/a
2
,
so that the modulus of a real
quantity
is
its absolute value.
190. The
product
of a
+ bi and c
+
di is
ac + bci
+
adi
+ bdi*
=
ac
bd+
(be
+
ad)
i.
Hence the modulus of the
product
of a
+
bi and
c +
di is
V{(c
-
bd)*
+
(be
+
ad)*}
=
V{(a
2
+
6
2
) (c
2
+
cT)}
Thus the modulus of the
product
of two
complex
expressions
is
equal
to the
product
of their moduli.'
The
proposition
can
easily
be extended to the case of
the
product
of more than two
complex expressions
; and,
since the modulus of a real
quantity
is its absolute
value,
we have the
following
Theorem. The modulus
of
the
product of any
number
of quantities
whether real or
complex,
is
equal
to the
product of
their moduli.
191. Since the modulus of the
product
of two com-
plex expressions
is
equal
to the
product
of their
moduli,
it
follows
conversely
that the modulus of the
quotient
of two
expressions
is the
quotient
of their moduli. This
may
also be
proved
directlv as follows :
S. A.
15
226
MODULUS OF A PRODUCT.
,
., a
+
bi c di
(a
+
bi)
-r-
(c
+
di)
=
-
-y.
x
-
T
.
c +
at c at
ac
+
bd
+
(6c ad)
i
, fa + Wl
V{(ac
+
Mf
+
(be -ad)*}
Hence mod 1
---
=
*-**-
i
~-
c*
+
cP
_
V(a
g
+
6'}
_
mod
(a
+
bi)
~ ~
mo
192.
It is obvious that in order that the
product
of
any
number of real factors
may
vanish,
it is
necessary
and sufficient that one of the factors should be
zero, and,
by
means of the theorem of Art.
190,
the
proposition
can
be
proved
to be true when all or
any
of the factors are
complex quantities.
For,
since the modulus of a
product
of
any
number of
factors is
equal
to the
product
of their
moduli,
and since
the moduli are all
real,
it follows that the modulus of
a
product
cannot vanish unless the modulus of one of its
factors vanishes.
Now if the
product
of
any
number of factors vanishes
its modulus must vanish
[Art. 189];
therefore the modu-
lus of one of the factors must
vanish,
and therefore that
factor must itself vanish.
Conversely,
if one of the
factors
vanishes,
its modulus will
vanish;
and therefore
the modulus of the
product
and hence the
product
itself
must vanish.
193. In the
expression
ax"
+
kr*"
1
+
ca
B
~
8
+...+
k,
where
a,b,c,...k
are all
real,
let a
+
/8
be substituted for x
t
and let P be the sum of all real terms in the
result,
and
Qi
the sum of all the
imaginary
terms. Then the
given
expression
becomes P
+
Qi.
Since P and
Q
are both
real,
they
can contain
only
MODULUS OF A PRODUCT. 227
squares
and
higher
even
powers
of
i,
and hence P and
Q
will not be
changed by changing
the
sign
of i Therefore
when a
fti
is substituted for x in the
given expression
the result will be P
Qi.
If now the
given expression
vanishes when a
+
j3i
is
substituted for
a,
we have P
+
Qi
=
0.
Hence,
as P and
Q
are
real,
we must have both P
=
and
Q
=
0,
and therefore
PQi=
0.
Hence if the
given
expression
vanishes when o
4-
/3i
is substituted for
x,
it will also vanish when a
/9i
is
substituted for x.
Therefore
[Art. 88]
if x a.
/3i
is a factor of the
given expression,
x a
+
(3i
will also be a factor.
Thus,
if any expression
rational and
integral
in
x,
and with all its
coefficients
real,
be divisible
by
either
of
two
conjugate complex expressions
it will also be divisible
by
the other.
152
CHAPTER XV
SQUARE
AND CUBE ROOTS.
194. WE have
already
shewn how to find the
square
of a
given algebraical expression;
and we have now to
shew how to
perform
the inverse
operation, namely
that
of
finding
an
expression
whose
square
will be
identically
equal
to a
given algebraical expression.
It will be seen
that our
knowledge
of the mode of formation of
squares
will enable us in
many
cases to write down
by inspection
the
square
root of a
given expression.
195. From the
identity
a* 2ab
+
tf
=
(a b)*,
we see that when a trinomial
expression
consists of the
sum of the
squares
of
any
two
quantities plus (or minus)
twice their
product,
it is
equal
to the
square
of their sum
(or difference).
Hence,
to write down the
square
root of a trinomial ex-
pression
which is a
perfect square, arrange
the
expression
according
to
descending powers
of some letter
;
the
square
root of the whole
expression
will then be found
by taking
the
square
roots of the extreme terms with the same or
with different
signs according
as the
sign
of the middle
term is
positive
or
negative.
Thus,
to find the
square
root of
SQUARE
ROOT. 229
The
square
roots of the extreme terms are + 2a
4
and
36
3
.
Hence,
the middle term
being negative,
the re-
quired square
root is +
(2a*
36
3
).
Note. In future
only
one of the two
square
roots of
an
expression
will be
given, namely
that one for which
the
sign
of the first term is
positive:
to find the other
root all the
signs
must be
changed.
196. When an
expression
which contains
only
two
different
powers
of a
particular
letter is
arranged
accord-
ing
to
ascending
or
descending powers
of that
letter,
it will
only
consist of three terms. For
example,
the
expression
a
2
+
6
s
+
c
2
+
2bc
+
2ca
-f
2ab when
arranged according
to
powers
of a is the trinomial
a-
+
2a
(6
+
c)
+
(6
2
+ c
2
+
26c).
It follows therefore from the
preceding
article that
however
many
terms there
may
be in an
expression
which
is a
perfect square,
the
square
root can be written down
by inspection, provided
that the
expression
contains
only
two
different powers of
some
particular
letter.
Ex. 1. To find the
square
root of
a
2
+
b- +
c-
+
2bc
+
2ca
+
2ab.
Arranged according
to
powers
of
a,
we have
a
2
+2a(6
+
c)
+
(6
+
c)
2
,
that is
Hence the
required square
root is a
+ b+c.
Ex. 2. To find the
square
root of
4s
4
+
9y
4
+
16z*
+
12s
2
?/
2
-
IQx-z-
-
The
given expression
is
4x*
+
4x
2
(3)/
2
-
4z
2
)
+
Qy*
-
24yW
+
1
that
is,
(2a;
2
)
2
+
2
(2a;
2
) (3?/
2
-
4*
2
)
+
(By
3
-
which is
{2x
a
+(3t/
2
-
4z
2
)}
2
.
Hence the
required square
root is 2x-
-t
3y-
-
4;
2
.
230 SQUARE
HOOT.
Ex. 3. To find the
square
root of
a
2
+
2afcc
+ (&
a
+
2ac)
x*+2bcx*+c*x*.
Arrange according
to
powers
of
a;
we then have
a
2
+
2a
(bx
+ex
3
)
+
h
2
*
2
+2bca?
+
Ac*,
that
is,
a
2
+
2a
(bx
+
cz
3
)
+
(bx
+c*
2
)
3
.
Hence the
required square
root is a
+
bx
+ ex
3
.
Ex. 4. To find the
square
root of
The
expression only
contains
y*
and
y
;
we therefore
arrange
it
according
to
powers
of
y,
and have
j/
2
+
2y
(x
3
-
x
3
+
x)
+
x*
-
2x*
+
Si
4
-
2x*
+x*.
Now,
if the
expression
is a
complete square
at
all,
the last of the
three terms must be the
square
of half the coefficient of
y
;
and it is
easy
to
verify
that
(*-*+a;)
2
=
x
8
-
2x*
+
3x*
-
2a? +*.
Hence the
required square
root is
y
+
x"
-
x*
+x.
197. To
find
the
square
root
of any algebraical
ex-
pression.
Suppose
that we have to find the
square
root of
(A
+
B)*,
where A stands for
any
number of terms of the
root,
and B
for the
rest;
the terms in A and B
being arranged
accord-
ing
to
descending (or ascending) powers of
some
letter,
so
that
every
term in A is
of higher
(or lower) degree
in thai
letter than
any
term
of
B.
Also
suppose
that the terms in A are
known,
and that
we have to find the terms in B.
Subtracting
A* from
(A +)*,
we have the remainder
Now from the mode of
arrangement
it follows that the
term of the
highest (or lowest)
degree
in the remainder is
twice the
product
of the
first
term in A and the
first
term
Hence,
to obtain the next term of the
required
root,
that
1
SQUARE
ROOT.
231
is,
to obtain the
highest (or lowest)
term of
B,
we subtract
from
the whole
expression
the
square of
that
part of
the root
which is
already found,
and divide the
highest
(or
lowest)
term
of
the remainder
by
twice the
first
term
of
the root.
The first term of the root is
clearly
the
square
root of
the first term of the
given expression;
and,
when we have
found the first term of the
root,
the second and other terms
of the root can be obtained in succession
by
the above
process.
For
example,
to find the
square
root of
x
6 -
4x* + 6x*
-
8x
3
4- 9x
2
-
4x
+4.
The
process
is written as follows :
x
6
(x
3
)
2
=x
(x
3
-
2x
2
)
8
=lc8
(x
3
-2x
3
+s-2)
a
=x
6
-4x
s
+6.r
4
-8x
:i
+ 9x
a
-
We first take the
square
root of the first term of the
given
expression,
which must be
arranged according
to
ascending
or de-
scending powers of
some letter : we thus obtain
a?,
the
first
term of
the
required
root.
Now subtract the
square
of x
3
from the
given expression,
and
divide the first term of the
remainder,
namely
-
4ar*,
by
2x
3
: we
thus obtain
-
2x
2
,
the second term of the root.
Now subtract the
square
of x
8
-2x
a
from the
given expression,
and divide the first term of the
remainder,
namely
2x
4
, by
2X
3
: we
thus obtain
x,
the third term of the root.
Now subtract the
square
of x
3
-
2x
2
+
x from the
given expression,
and divide the first term of the
remainder,
namely
-4x
3
,
by
2X
3
: we
thus obtain
-
2,
the
fourth
term of the root.
Subtract the
square
of x
3
-
2x*
+
x
-
2 from the
given expression
and there is no remainder.
Hence x
3
-
2X
2
+
x
-
2 is the
required square
root.
The
squares
of x
3
,
x
3
-
2x
2
,
<fec. are
placed
under the
given
expression,
like terms
being placed
in the same
column,
so tbat in
every
case the first term of the remainder is obvious.
198. The
square
root of an
algebraical expression may
also be obtained
by
means of the theorem of Art 91.
Take for
example
the case
just
considered.
232 SQUARE
ROOT.
The
required
root will be ax
3
+ bx*
+
cx
+
d,
provided
that the
given expression
is
equal
to
(ax
3
+
bx*
+
ex +
d)*,
that is
equal
to
aV
+
2abx*
+
(2ac
+
6
s
)
x*
+
2
(ad
+
be}
a?
-I-
(26d
+ c
2
)
x*
+
2cdx + <P.
Hence,
equating
the coefficients of
corresponding
powers
of x in the last
expression
and in the
expression
whose root is
required,
we have
a"=l;
The first four of these
equations
are sufficient to
determine the values of
a, b, c, d\
these values are
(taking
only
the
positive
value of
a),
a
=
1,
b
=
2,
c
=
1,
d
=
2.
The last three
equations
will be satisfied
by
the values
of
a, b, c,
d found from the first
four,
provided
the
given
expression
is a
perfect square,
which is
really
the case.
Thus the
required square
root is x
3
2x
2
+
x
-
2.
199. Extended Definition of
Square
Root. The
definition of the
Square
Root of an
algebraical expression
may
be extended so as to include the case of an
expression
which is not a
perfect square.
For,
although
an
expres-
sion
may
not be a
perfect square,
we can
find,
by
the
methods of Art. 197 or Art.
198,
a second
expression
whose
square
is
equal
to the
given expression
so far as
certain terms are concerned.
Thus the
square
root of x*
+
2x
may
be said to be
cc+l,
(x
+
I)* being equal
to x*+2as so far as the terms
which contain x are concerned.
Again,
the
square
root of 1
+
x
may
be said to be
2
l
+
s
orl+-
,
the
square
of the former
differing
from
2 i O
1
+
x
by
,
and the
square
of the latter
differing by
SQUARE
ROOT. 233
"J^
+
A**
Thus,
provided
x is
small,
1+s
*s an
2
J? l/
approximation
to the
square
root of 1
+
x,
and 1
+
= -~
c5
is a closer
approximation,
and
by continuing
the
process
we can
approximate
as
closely
as we
please
to the
square
root of 1
+
x
;
this however is
by
no means the case when
a; is not a small
quantity.
200. When
any
number
of
terms
of
a
square
root have
been obtained as
many
more can be
found by ordinary
division.
For
suppose
the
expression
whose
square
root is to be
found is the
square
of
The coefficients a
l}
a
t ,...a^.
can be found
by equating
the coefficients of the first 2r
powers
of x in the
square
of
the above to the coefficients of the
corresponding powers
of x in the
given expression.
The
square
of the above
expression
is
*'1
+
.
+
aX
1"""1
)
2
+
2
(a^
+ ...
+
2R
(a
Now,
since the
highest power
of x in R is x
n~*r
,
the
highest power
of a; in the
expression
within
square
brackets
is a;
8
"-*.
Hence the
expression
within
square
brackets will not
affect
any
of the terms from which
a,,
a
s
,
--.a^
are deter-
mined,
for the first 2r terms of the
given expression
ex-
tend from x* to
It therefore follows that if the
square
of the sum of
the first r terms of the root be subtracted from the
given
234 SQUABE
ROOT.
expression,
and the remainder be divided
by
twice the
sum of the first r
terms,
the
quotient
will
give
the next
r terms of the root.
201. When n
figures of
a
square
root
of
a number
have been
found by
the
ordinary
method,
n 1 more
figures
can be
found by
division,
provided
that the number is a
perfect square of
2n 1
figures
;
if
however this be not the
case,
there
may
be an error in the last
figure.
Let N be the
given
number,
which is the
perfect
square
of a number
containing
2n 1
figures,
and let
p
be the number formed
by
the first n
figures
followed
by
n 1
zeros,
and let
q
be the number formed
by
the
remaining
n I
figures.
Then
Now
2p
<
2 .10"" and
q
>
10*'
1
. Hence
q*/2p
must be a
fraction
;
whence it follows that if
p*
be subtracted from
N and the remainder be divided
by 2p,
the
integral part
of the
quotient
will be
q.
Next,
let
*/N
contain m
figures,
where m is
greater
than 2n 1.
Let
p
be the number formed
by
the first n
figures
of
the root followed
by
m n
zeros,
let
q
be the number
formed
by
the next n 1
figures
followed
by
m 2n
+
1
zeros,
and let r be the number formed
by
the m 2?i
+
1
remaining figures.
Then
.'.
(N
-
p*)/2jo -q
=
(q*
+
r*
+
2qr)/2p
+
r.
Now
io*>pH:l0
1 "'1
,
10"-" >
q
<
lO"
1'"-1
,
and 1(r
-**i
> r
^
1(r
-
CUBE ROOT. 235
whence it follows that
(<?
2
+
r
2
-i-2gr)/2p
is less than
I
/\>
2M-H
Hence
(q*
+
r*
+
2gr)/2;>
+
r is less than 2 x 10
m"*"+l
,
but
it is not
necessarily
less than 10
m
"
1 "+1
. Hence
(N-p*)jZp may
differ from
q by
more than lO"-"
1*1
;
it
must however differ
by
less than 2 x 10"
1
"
2""1
"
1
;
so that the
n 1 first
figures
of the
quotient
(N p*)/2p
are either
the n 1
figures
of
q
or differ
only
in the last
figure,
and in that case
by
1 in excess.
CUBE ROOT.
202. From the
identity
(a
+
6)
3
=
a
s
+
3a
2
6
+ 3a&
2
+ 6
s
,
we see that the cube of a binomial
expression
has
four
terms,
and that when the cube i&
arranged according
to
ascending
or
descending powers
of some
letter,
the cube
roots of its extreme terms are the terms of the
original
binomial.
Hence the cube root of
any perfect
cube which has
only
four terms can be written down
by inspection,
for we
have
only
to
arrange
the
expression according
to
powers
of some letter and then take the cube roots of its extreme
terms.
For
example,
if 27a
6
-
54a
5
6 + ZQatW
-
8a?b
s
is a
perfect
cube its
cube root must be 3a
2
-
2a6
;
and
by forming
the cube of 3o
a
-
2ab
it is seen that the
given expression
is
really
a
perfect
cube.
When an
expression
which contains
only
three different
powers
of a
particular
letter is
arranged according
to
powers
of that
letter,
there will be
only^owr
terms.
It therefore follows that however
many
terms there
may
be in an
expression
which is a
perfect
cube,
the cube
root can be written down
by
inspection, provided
that the
230
CUBE ROOT.
expression
contains
only
three different
powers
of some
particular
letter.
For
example,
to find the cube root of
a
3
+
b*
+
c
8
+
3a6
-f
3a
2
c + 3a5* -f
3ac
2
+ Gate
+
36^ + 36c
J
.
Arranged accovliug
to
powers
of
o,
we have
a
3
+
3a
3
(b
+
c)
+
3a
(&"
+c
-t-
2fcc)
+
6 +c +36c +
36c
2
,
that
is,
a
3
f
3a
(h
-f
c)
+
3a
(6
+
c)
2
+
(6
+
c)
s
.
Hence the
required
root is a
+
6
+
c.
203. To
fend
the cube root
ofany algebraical expression.
Suppose
we have to find the cube root of
(
A
+
B)
3
,
where A stands for
any
number of terms of the
root,
and
B for the
rest;
the terms in A and B
being arranged
according
to
descending (or ascending) powers
of some
letter,
so that
every
term of A is of
higher (or lower)
degree
in that letter than
any
term of B.
Also
suppose
the terms in A are
known,
and that we
have to find the terms in B.
Subtracting
A
3
from
(A
+
)
3
,
we have the remainder
Now from the mode of
arrangement
it follows that the
term of the
highest (or lowest)
degree
in the remainder is
3 x
square
of the first term of A x first term of B.
Hence to obtain the next term of the
required
root,
that
is,
to obtain the
highest (or lowest)
term of B we
subtract
from
the whole
expression
the cube
of
that
part
of
the root which is
already found
and divide the
highest
(or
lowest)
term
of
the remainder
by
three times the
square
of
the
first
term
of
the root.
This
gives
a method of
finding
the successive terms
of the root after the first
;
and the first term of the root
is
clearly
the cube root of the first term of the
given
expression.
CUBE ROOT. 237
For
example,
to find the cube root of
of
-
6ofy
+
21*V
-
44a;V
+
63zy
-
olxy*
+
27?/.
The
process
is written as follows :
-
54.r.v
3
+
27t/
-
8x
3
y
3
(
x
n
-
-2xy
+
3y
2
)
3
=
a?
-
Garty
+
21xV
-
44afy8
-f 63a;
V
-
Having arranged
the
given expression according
to
descending
powers
of
x,
we take the cube root of the first term : we thus obtain
x-,
the
first
term of the
required
root.
We then subtract the cube of 3? from the
given expression,
and
divide the first term of the
remainder, namely
-
6a;
1
fy, by
3 x
(a?)-
:
we thus obtain
-
2xy,
the second term of the root.
We then subtract the cube of a?
-
2xy
from the
given expression,
and divide the first term of the remainder
by
3 x
(x'-)
2
: this will
give
the third term of the root.
Note. The above rule for
finding
the cube root of an
algebraical expression
is
rarely,
if
ever,
necessary.
In actual
practice
cube roots are found as follows.
Take the case
just
considered
;
the first and last terms
of the root are #
2
and
3y*,
the cube roots of the first and
last terms of the
given expression ;
also the second term
of the root will be found
by dividing
the second term of
the
given expression by
3 x
(#*)
2
>
so tnat ^^ie second term
of the root is
2xy.
Hence,
if
the
given expression
is
really
a
perfect
cube,
it must be
(x* %xy
+
Sy")
3
,
and it is
easy
to
verify
that
(a?
Zxy
+
3?/
2
)
8
is
equal
to the
given expression.
Again,
to find the cube root of
a?
-
If the
given expression
is
really
a
perfect
cube the
first and last terms of the root must be
^/x
9
and
\/
27?/
8
respectively,
that is
a?
8
and
3t/
a
.
238 EXAMPLES.
The second term of the root must be
Qafy
-r- 3
(#
3
)*
2aty
;
and the term next to the last must be
Hence the
given expression, if
a cube at
all,
must be
(a? 2x*y
+
xy* Sy
8
)
3
;
and
by expanding
(of 2a?y
+
xy* 3^)
s
it will be found that the
given
expression
is
really
a
perfect
cube.
204. From the
identity [see
Art.
253]
(a
+
&)"
=
a"
+
na*~
l
b
+
terms of lower
degree
in
a,
it is
easy
to
shew,
as in Articles 197 and
203,
that the
n
th
root of
any algebraical expression
can be found
by
the
following
Rule.
Arrange
the
expression according
to
descending
or
ascending powers of
some
letter,
and take the n
th
root
of
the
first
term : this
gives
the
first
term
of
the root.
Also,
having found any
number
of
terms
of
the
root,
subtract
from
the
given expression
the w
th
power of
that
part of
the root which is
already found,
and divide the
first
term
of
the remainder
by
n times the
(n
1
)
th
power of
the
first
term
of
the root : this
gives
the next term
of
the root.
EXAMPLES XIX.
Write down the
square
roots of the
following expressions
:
2.
3.
4. 25a
4
+ 96
4
+ 4c
4
+ 1 26V
-
20cV
-
EXAMPLES.
239
Find the
square
roots of
5. a:
6
6. 4a
7. 49 + 1 12x* + 7Qx* + 64a:
4
+ SO*
5
+ 25a;.
8. x*
-
2x
8
+ 5x*
-
6x + 8
-
6aT' + 5aT
8
-
2aT
8
9
25a?
*
+ J^_20- +
-
y
+ 2.
y*
25*
'
y
5x
10.
*
-
11. x%
-
12. a:*
-
Za'tx** + 2a*x* + cT^a:^
-
2atx^ + a?.
Find the cube roots of
13. x*
14. x*
-
15. 1
-
9or + 33o;
4
-
63a:
8
+ 66x"
-
16. Find the
square
root of
2a
2
(b
+
cf
+ 2b*
(c
+
a)
2
+ 2c
2
(a
+
ft)
2
+
4aJc
(a
+ 6 +
c).
17. Find the
square
root of
a?
(x*
+
y*
+
s?)
-f
yV
+ 2a;
(y
+
) (y
-
a;
2
).
18. Find the
square
root of
(a
_
6)
4
-
2
(a
2
+ 6
2
) (a
-
6)'
+ 2
(a
4
+ 6
4
).
19. Shew that
(a;
+
a)
(x
+
2a) (x
+
3a) (a;
+
4a)
+ a
4
is a
perfect
square.
20. Prove that a;
4
+
px
3
+
qx*
+ rx + s is a
square,
if
p
3
s
=
r*
and
p
3
-
4pq
+ 8r
=
0.
240
EXAMPLES.
21. Find the values of
A,
B and C in order that
4x
-
24o;
5
+
Ax* + a? + Cx*
-
40* + 25
may
be a
perfect
square.
22. Shew
that,
if ax
3
+ bx* + cv + d be a
perfect
cube then
23.
*
Find the conditions that
ay? +
by
3
+ cz* +
2fyz
+
2gzx
+
Ihxy
may
be the
square
of an
expression
which is rational in
x,
y
and z.
24. Shew that if
(a
-
A)
x
2
+
(6
-
A.) y*
+
(c
-
X)
z* +
'Ifyz
+
Igzx
+
2hxy
be the
square
of an
expression
which is rational in
x, y
and
,
then will
a
-^
=
b -V=c-SfL
=
X
.
f 9
h
25. Shew that when the first r terms of the cube root of
an
algebraical expression
are
known,
r more terms can be
found
by ordinary
division.
26. When n +
2
figures
of the cube root of a number have
been obtained
by
the
ordinary
method,
n more can be obtained
by ordinary division,
provided
the number is a
perfect
cube of
2 + 2
figures.
27. Shew
that,
if n + 2
figures
whose numerical value
is a have been found of a
positive
root of the
equation
9? +
qx
r
=
0, q being supposed positive,
then the result of
dividing
r
qa
a*
by
3a
s
+
q
will
give
at least n 1 more
figures correctly.
CHAPTER XVL
RATIO. PECPORTION.
205. Definitions. The relative
magnitude
of two
quantities,
measured
by
the number of times the one
contains the
other,
is called their ratio.
Concrete
quantities
of different kinds can have no
ratio to one another: we
cannot,
for
example, compare
with
respect
to
magnitude
miles and
tons,
or
shillings
and weeks.
The ratio of a to b is
expressed by
the notation a : b
;
and a is called the first
term,
and b the second
term,
of the
ratio. Sometimes the first and second terms of a ratio are
called
respectively
the antecedent and the
consequent.
It is clear that a ratio is
greater, equal
or less than
unity according
as its first term is
greater, equal
or less
than the second. A ratio which is
greater
than
unity
is
sometimes called a ratio of
greater inequality,
and a ratio
which is less than
unity
is
similarly
called a ratio of less
inequality.
The ratio of the
product
of the first terms of
any
number of ratios to the
product
of their second
terms,
is
called the ratio
compounded
of the
given
ratios.
Thus ac : bd is the ratio
compounded
of the two ratios a : b and c : d.
The ratio a* : b* is sometimes called the
duplicate
ratio
of a : b
;
so also a
3
: b
8
,
and
\/a
:
V&
are called
respectively
the
triplicate,
and the sub-
duplicate
ratio of a : b.
S.A. 16
242
RATIO.
206.
Magnitudes
must
always
be
expressed by
means
of
numbers,
and tbe number of times which one number
contains another is found
by dividing
the one
by
the other.
Thus ratios can be
expressed
as
fractions,
The
principal properties
of fractions and therefore of
ratios have
already
been considered in
Chapter
viil.
Thus,
a ratio is unaltered in value
by multiplying
each
of its terms
by
the same number.
[Art. 107.]
Different ratios can be
compared by reducing
to a
common denominator the fractions which
express
their
values.
[Art. 109.]
The theorems of Art. 113 are also true for ratios.
The
following
theorem is of
importance
:
207. Theorem.
Any
ratio is made more
nearly
equal
to
unity by adding
the same
positive quantity
to each
of
its terms.
By adding
x to each term of the ratio a :
b,
the ratio
a
+
x : b
+
x is obtained.
,.
T
a a b , a
+
x
,
a b
Now
T
1
=
r~~
>
anc*
r~,
1
=
iT.
b b o+x
b
+
x
and it is clear that the absolute value of
7
is less than
b+as
that of
F
,
for the numerators are the same and the
o
denominator of the former is the
larger
: this
proves
the
proposition.
When x is
very great,
the fraction
r
is
very
small
;
and
, ,
which is the difference between and
1,
o
+
o?
b+as
can be made less than
any assignable difference by taking
as
sufficiently great.
RATIO.
243
This is
expressed by saying
that the
limiting
value of
a
+
x , . . , . .
^
,
when x is
infinite,
is
unity.
H-
x
Now two
quantities,
whether finite or
not,
are
equal
to one another when their ratio is
unity.
Thus a
+
x and
6
+
x are
equal
to one another when x is
infinite,
a
being
supposed
not
equal
to b.
[See
Art.
118.]
208. Since
any
ratio is made more
nearly equal
to
unity by
the addition of the same
quantity
to each of its
terms,
it follows that a ratio is diminished or increased
by
such addition
according
as it was
originally greater
or less
than
unity.
This
proposition
is sometimes enunciated :
A ratio
of greater inequality
is diminished and a ratio
of
less
inequality
is increased
by
the addition
of
the same
quantity
to each
of
its terms.
209. Incommensurable numbers. The ratio of
two
quantities
cannot
always
be
expressed by
the ratio of
two whole numbers
;
for
example,
the ratio of a
diagonal
to a side of a
square
cannot be so
expressed,
for this ratio
is
\/2
:
1,
and we cannot find
any
fraction which is
exactly
equal
to
\/2.
Magnitudes
whose ratio cannot be
exactly expressed
by
the ratio of two whole
numbers,
are said to be in-
commensurable.
Although
the ratio of two incommensurable numbers
cannot be found
exactly,
the ratio can be found to
any
degree
of
approximation
which
may
be desired
;
and the
different theorems which have been
proved
with
respect
to ratios
can,
by
the method of Art.
163,
be
proved
to be
true for the ratios of incommensurable numbers.
102
244"
PROPORTION.
PROPORTION.
210. Four
quantities
are said to be
proportional
when
the ratio of the first to the second is
equal
to the ratio of
the third to the fourth.
Thus
a, b, c,
d are
proportional,
if
a : b c : d.
This is sometimes
expressed by
the notation
a :b :: c :
d,
which is read
"
a is to b as c is to d."
The first and fourth of four
quantities
in
proportion,
are sometimes called the
extremes,
and the second and
third of the
quantities
are called the means.
211. If the four
quantities
a, b, c,
d are
proportional,
we have
by
definition,
a
_c
I'd'
Multiply
each of these
equals by
bd : then
ad
=
bc.
Thus the
product
of
the extremes is
equal
to the
product
of
the means.
Conversely, if
ad
=
be,
then
a, b, c,
d will be
propor-
tional.
For,
if ad
=
bc,
then
ad_bc^
bd~bd'
a
_c
"'
b~d'
that is
a : b c : d.
~
PROPORTION.
245
Hence
also,
the four relations
a : b
=
c :
d,
a : c =b :
d,
b : a
=
d :
c,
and b : d
=
a :
c,
are all
true,
provided
that ad
=
be. Hence the four
proportions
are all true when
any
one of them is true.
Ex. If a : b=c :
d,
then will a
+
b :
a-b=c+d : c-d.
This has
already
been
proved
in Art. 113 : it
may
also be
proved
as follows :
a
+
b : a-b=c
+
d :
c-d,
if (a
+
b)(c-d)
=
(a-b)(c
+
d),
that
is,
{/
ac-bd +
bc-ad=ac- bd-bc
+
ad}
or,
if
bc
=
ad.
But be is
equal
to
ad,
since a : b=c : d.
212.
Quantities
are said to be in continued
proportion
when the ratios of the first to the
second,
of the second
to the
third,
of the third to the
fourth, &c.,
are all
equal.
Thus
a, 6, c, d,
&c. are in continued
proportion
if
a: 6
=
6 : c
=
c : d
&c.,
,,,..,.
a b c
p
that
is,
it
T
= -
=
-7
=
&c.
bed
If a : b
=
6 :
c,
then b is called the mean
proportional
between a and
c;
also c is called the third
proportional
to a and 6.
If
a, b,
c be in continued
proportion,
we have
a b
b
=
c'
.'. V
=
ac,
or b
=
Jac.
246
PROPORTION.
Thus the mean
proportional
between two
given quantities
is the
square
root
of
their
product.
a b b b
Also T- x
-
=
-
x
-
,
b c c G
a 6* a*
tnat is
=
-,
=
TT
c c* 6
s
Thus,
if
three
quantities
are in continued
proportion,
the ratio
of
the
first
to the third is the
duplicate
ratio
of
the
first
to the second.
213. The definition of
proportion given
in Euclid is
as follows : Four
quantities
are
proportionals,
when if
any
equimultiples
whatever be taken of the first and the
third,
and also
any equimultiples
whatever of the second and
the
fourth,
the
multiple
of the third is
always 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.
If the four
quantities
a, b, c,
d
satisfy
the
algebraical
d C
test of
proportionality,
we have
T
=
-7
;
therefore for all
o a
, ma me
values of m and
n,
,
=
j.
no nd
>
>
Hence me
=
nd,
according
as ma
=
nb. Thus
a, 6, c,
d
<
<
satisfy
also Euclid's test of
proportionality.
Next,
suppose
that
a, 6, c,
d
satisfy
Euclid's definition
of
proportion.
If a and b are
commensurable,
so that a : b
=
m :
n,
where m and n are whole numbers
;
then
a m I*
' M
j
T
=
;
.'. na
=
mu.
b n
PROPORTION.
247
But
by
definition
nc
=
md
according
as na
=
mb.
Hence nc
md;
c
_m _
a
d n~b'
Thus
a, b, c,
d
satisfy
the
algebraical
definition.
If a and b are incommensurable we cannot find two
whole numbers m and n such that a : b
=
m : n.
But,
if
we take
any multiple
na of
a,
this must lie between two
consecutive
multiples, say
mb and
(m
+
1)
b of
b,
so that
na>mb and
na<(m
+
l)b.
Hence
by
the
definition,
nc > md and nc <
(m
+
1)
d.
Hence both
T
and -, lie between and .
b d n n
a c . 1
Thus the difference between
T
and
-
7
is less than
-
;
and
b d
n
as this is the case however
great
n
may
be,
-^
must be
equal
to
r ,
for their difference can be made less than
any
assignable
difference
by sufficiently increasing
n.
Ex. 1. For what value of x will the ratio 7 +
x : 12
+x be
equal
to
the ratio 5 : 6? Ans. 18.
Ex. 2. If &x
+
6y
3
=
13xy,
what is the ratio of x to
y?
Ans. 2 : 3 or 3 : 2.
Ex. 3. What is the least
integer
which when added to both terms of
the ratio 5 : 9 will make a ratio
greater
than 7 : 10? Ans. 5.
Ex. 4. Find x in order that x
+
l : x
+ Q
may
be the
duplicate
ratio of
3 : 5. Ans.
~. .
16
248 VARIATION.
Ex. 5. Shew
that,
if a : b : : c :
d,
then
(i)
cP+ab +
b* : c* + cd + d* :: a*-ab
+
b* : c'-c
(ii)
(iii)
:
(a
+
b
+c
+
d)*.
[See
Art.
113.]
Ex. 6. If a : 5 :: c :
d,
then will al
+
cd be a mean
proportional
between o*
+
c- and 6
s
+
d
1
.
VAKIATION.
214. One
magnitude
is said to
vary
as another when
the two are so related
that the ratio of
any
two values of
the one is
equal
to the
ratio of the
corresponding
values
of the other.
Thus,
if a
t
,
a
s
be
any
two measures of one of the
quantities,
and b
lt
b
t
be the
corresponding
measures of the
other,
we have
=
r
1
;
and therefore
-^
=
j
3
.
a, 6,' \ 6,
Hence the measures of
corresponding
values of the two
magnitudes
are in a constant ratio.
The
symbol
is used for the words varies as: thus
A
oc
B is read
'
A varies as B'.
If a
oc
6,
the ratio a : b is constant
;
and if we
put
m
for this constant
ratio,
we have
a ,
T-
=
77i: .'. a
=
7710.
To find the constant m in
any
case it is
only necessary
to know one set of
corresponding
values of a and 6.
For
example,
if a ac
b,
and o is 15 when b is
5,
we have
-
=
m
=
;
.\a
=
36.
:
VARIATION. 249
215. Definitions. One
quantity
is said to
vary
in-
versely
as another when the first varies as the
reciprocal
of the second.
Thus a varies
inversely
as b if the ratio a :
r
is
constant,
and therefore ab
=
m.
One
quantity
is said to
vary
as two others
jointly
when
the first varies as the
product
of the other two. Thus a
varies as b and c
jointly
if a
be,
that is if a
=
mbc,
where m is a constant.
One
quantity
is said to
vary directly
as a second and
inversely
as- a third when the ratio of the first to the
product
of the second and the
reciprocal
of the third is
constant.
Thus a is said to
vary directly
as b and
inversely
as
c,
if a : b x
-
is
constant,
that
is,
if a
=
m-
,
where m is a
c c
constant.
In all the different cases of variation defined
above,
the constant will be determined when
any
one set of
corresponding
values is
given.
For
example,
if a varies
jointly
as b and c
;
and if a is 6 when b
is 4 and c is
3,
we have
a=mbc,
and
6=mx4x3.
Hence m=-
,
and therefore
a=^bc.
t i
216. Theorem.
If
a
depends only
on b and
c,
and
if
a varies as b when c is
constant,
and varies as c when
b is
constant; then,
when both b and c
vary,
a will
vary
as be.
Let
a, b, c; a', b',
c and
a", b',
c' be three sets of
corresponding
values.
250 VARIATION.
Then,
since c is the same in the first and
second,
we
,
a b ,..
bave
a'
=
F
(l)
'
And,
since 6' is the same in the second and
third,
we
a c
have
=
(ii).
a c
Hence from
(i)
and
(ii),
=
TT-, ,
ct o c
which
proves
the
proposition.
The
following
are
examples
of the above
proposition.
The cost
[C]
of a
quantity
of meat varies as the
price [P] per
pound
if the
weight [W]
is
constant,
and the cost varies as the
weight
if the
price per pound
is constant.
Hence,
when both the
weight
and the
price per pound change,
the cost varies as the
product
of the
weight
and the
price.
Thus,
if C a
P,
when W is
constant,
and C <x.
W,
when P is constant
;
then C a.
PW,
when both P and W
change.
Again,
the area of a
triangle
varies as the base when the
height
is
constant;
the area also varies as the
height
when the base is
constant; hence,
when both the
height
and the base
change,
the
area will
vary
as the base and
height jointly.
Again,
the
pressure
of a
gas
varies as the
density
when the
temperature
is
constant;
the
pressure
also varies as the absolute
temperature
when the
density
is constant
;
hence when both
density
and
temperature change,
the
pressure
will
vary
as the
product
of the
density
and absolute
temperature.
Ex. 1. The area of a circle varies as the
square
of its
radius,
and the area of a circle whose radius is 10 feet is 314-159
square
feet. What is the area of a circle whose radius is 7 feet?
Am. 452-38896 feet.
Ex. 2. The volume of a
sphere
varies as the cube of its
radius,
and the volume of a
sphere
whose radius is 1 foot is 4*188 cubic feet.
What is the volume of a
sphere
of one
yard
radius? Am. 113-076 feet.
Ex. 3. The distance
through
which a
heavy body
falls from rest
varies as the
square
of the time it falls
;)
also a
body
falls 64 feet in
2 seconds. How far does a
body
fall in 6 seconds? Am. 576 feet.
Ex. 4. The volume of a
gas
varies as the absolute
temperature
and
inversely
as the
pressure;
also when the
pressure
is 15 and the
temperature
260 the volume is 200 cubic inches. What will the
volume be when the
pressure
becomes 18 and the
temperature
390?
Ant. 250 inches.
INDETERMINATE FORMS. 251
Ex. 5. The distance of the
offing
at sea varies as the
square
root
of the
height
of the
eye
above the sea
level,
and the distance is
3 miles when the
height
is 6 feet : find the distance when the
height
is 72
yards.
Ans. 18 miles.
INDETERMINATE FORMS.
217. A ratio or fraction sometimes assumes an in-
determinate
form for some value or values of a contained
letter.
Thus,
when x=Q both the numerator and the denominator of
a? x
the fraction -3
-
vanish,
and the fraction assumes for this value of
x*-x
x the indeterminate form
^
;
and this is also the case when x=l.
Again,
when x
=
00 both the numerator and the denominator of
the above fraction become
infinitely great,
and the fraction assumes
00
the indeterminate form .
oo
We
proceed
to shew how to find the
limiting
values
of fractions which assume these indeterminate forms.
x
2
1
Consider,
for
example,
the fraction
-5
=
,
which as-
"~ "~
-L
sumes the form when x
=
1.
and,
provided
ac 1 is not
really
zero,
we
may
divide the
numerator and denominator
by
as 1 without
altering
the
value of the
fraction,
and we can do this however small
xl
may
be.
TT i- 77
^
-
*
+
l
Hence,
when x 1 is
very
small,
-3
^
=
-5
-
=
,
and the
limiting
value of the latter
fraction,
as x
ap-
2
proaches indefinitely
near to
1,
is at once seen to be
^
.
252 INDETERMINATE
FORMS.
Hence,
as x
approaches indefinitely
near to
1,
the
a? 1
2
fraction
3
_
approaches indefinitely
near to the value =.
a? \ 2
This is
expressed by
the notation
.,
-. =
=
7:
.
1
or 1 3
Ex. 1. Find the
limiting
value of
-5
-
when
x=2.
x* 10x
+
lo
It follows from Art. 88 that x-2 is a common factor of the
numerator and denominator.
x
a
-5x
+
6
_ T (x-2)(x-3)_ r
*-3_l
~
*-
(x-2)(x-8)
*-*
x-8~6"
x x
Ex. 2. Find the
limiting
value of
^-5
:
-
when x
=
and when
2i%*
-f-
OX
_
2
*~
*
~
"
~
*~
2 3
since
-
and
-
are both zero when x is
infinite,
x x
ET. 3. Find the
limiting
value of the ratio l
+
2x : 2
+
3x when x
increases without limit.
/ 1\ 1
1 + 2*
*V x)
* 2
*"
/
2\~
*-
(
3+
i)
Ex. 4. Find the
limiting
value of -. when x becomes
Mr
-
40
indefinitely great.
2x
s
+ 100x + 500 .
x(6-^r
EXAMPLES.
253
EXAMPLES XX.
1. Shew
that,
if a +
b,
b 4-
c,
c +
a are in
continued
propor-
tion,
then b + c : c + a
=
c a . a b.
2. Shew
that,
if x : a
y
: b
=
z :
c,
then
x
a
y*
z
8
_(x
+
y
+
z)
3
tf
+
b
t
+
J~(a
+ b + C
y'
3. Shew
that,
if
(a
+ b + c +
d) (a
-
b
-
c +
d)= (a
-
b + c
-
d)
(a+
b c
d),
then
a, b, c,
d are
proportionals.
4. Shew
that,
if 6* + c
2
=
a
8
,
then
a + b + c : c + a b
=
a + b-c:b + c a.
5. What number must be subtracted from each of the
numbers
7, 10, 19,
31 in order that the remainders
may
be in
proportion
?
6. Find a : b :
c,
having given
b a+c-b a+b+c
7. If
a + b b+c a 2a + b+2c'
x
y
z
b+ca c+ab a+b c'
shew that
(a
+ b +
c) (yz
+ zx +
xy)
-
(x
+
y
+
z) (ax
+
by
+
cz).
8.
If a
(y
+
z)
=
b
(z
+
x)
=
c
(x
+
y), prove
that
y-z
zx x
y
a(b-c)
b
(c- a)
c(a-b)'
9.
Shew that the ratio
is intermediate to the
greatest
and least of the ratios a
l
: b
1}
a
t
: b
a
, &c.,
the
quantities being
all
positive.
10.
If a : b :: c :
d,
then
11.
Shew
that,
if
(a
+
b) (b
+
c) (c
+
d) (d
+
a)
=
(a
+ b + c +
d) (bed
+ cda + dab +
abc)
t
then
a
: b : : d
: c.
254 EXAMPLES.
12. If
(bed
+ nla + dab +
abc)*
-
abed
(a
+ b + c +
d)*
=
0,
then it will be
possible
to
arrange
a, 6, c,
d so as to be
propor-
tionals.
13. Shew
that,
if
then
a + 26 + c a c a 26 + c
'
abc
xz
x-2y
+ z'
14. Shew
that,
if ax
3
+
by
2
+ cz
3
+
2/y
4-
2^c
4
2hxy
=
and x +
y
+ z
=
are
only
satisfied
by
one set of ratios x :
y
:
z,
then be
f*
+
ca-g*
+ ab-h* + 2
(gh
-
of)
15. Shew
that,
if
a
then
p(px-qy-rz) q(qy-rz-px) r(rz-px-qyY
-- -
r
jy
- -
r
--
j
-
a
(ax by
cz)
b
(by
cz
ax) c(cz
ax
by)
16. Shew
that,
if ab
=
cd,
then either of them is
equal
to
(a
+
c) (a
+
d)
(b
+
c) (b
+
d)/(a
+ b + c +
d)*.
Also,
if.a+b
=
c +
d,
then either of them is
equal
to
, ,/l
1 1 1\
/,
,
abed
(
-
+
T
+
-
+
-j) (ab
+
cd).
\a
b c
dj
I
v
17. Find the
limiting
values of the
following
fractions
when x
=
2,
and when x
=
oo .
x*-7x+ 10 g
>
'
.r
3
-12a;+16'
18. Find the
limiting
values of the
following
when x
=
a.
CHAPTER XVII.
ARITHMETICAL, GEOMETRICAL,
AND HARMONICAL
PROGRESSION.
218. Series. A succession of
quantities
the members
of which are formed in order
according
to some definite
law is called a series.
Thus
1, 2, 3, 4, ,
in which each term exceeds the
preceding by unity,
is a series.
So also
3, 6, 12, 24, ,
in which each term is double
the
preceding,
is a series.
We shall in the
present Chapter
consider some
very
simple
cases of
series,
and shall return to the
subject
in a
subsequent
Chapter.
ARITHMETICAL PROGRESSION.
219. Definition. A series of
quantities
is said to be
in Arithmetical
Progression
when the difference between
any
term and the
preceding
one is the same
throughout
the series.
Thus, a, b, c, d,
&c. are in Arithmetical
Progression
[A. P.]
if b a
=
G b
=
d c
=
&c.
The difference between each term of an A. P. and the
preceding
term is called the common
difference.
256
ARITHMETICAL PROGRESSION.
The
following
are
examples
of Arithmetical
progressions
:
1,
3, 6, 7,
&c.
8, _i, -5, -9,
&c.
a,
a
+
26,
a
+46,
&o.
In the first series the common difference is
2,
in the second it is
-
4,
and in the last it is 2b.
220. If the first term of an arithmetical
progression
be
a,
and the common difference
d; then,
by
definition,
the 2nd term will be a
+
d,
3rd a
+
2d,
4th a
+
3d,
and so
on,
the coefficient of d
being always
less
by unity
than the number
giving
the
position
of the term in the
series.
Hence the nth term will be a
+
(n 1)
d.
We can therefore write down
any
term of an A. P.
when the first term and the common difference are
given.
For
example,
in the A. P. whose first term is
5,
and whose
common difference is
4,
the 10th term is 5 +
(10- 1)4
=
41,
and the
30th term is 5
+
29 x4
=
121.
221. An arithmetical
progression
is determined when
any
two of its terms are
given.
For,
suppose
we know that the mth term is
2,
and that
the nth term is
ft.
Let a be the first
term,
and d the common
difference;
then the mth term will be a
+
(m
-
1)
d,
and the nth term
will be a
+
(n 1)
d.
Hence a
+
(m l)d
=
at,
and a
+
Thus we have two
equations
of the first
degree
to
determine a and d in terms of the known
quantities
m, n,
a and
,3.
ARITHMETICAL PROGRESSION. 257
Ex. Find the 10th term of the A. p. whose 7th term is 15 and whose
21st term is 22.
If a be the first
term,
and d be the common
difference,
we have
o
+
6d=15,
and a
+ 20d=22.
1 9
Hence d
=
-
,
a
=
12. The 10th term is therefore 12
+
^
=
16.
-j 2i
222. When three
quantities
are in arithmetical
pro-
gression,
the middle one is called the Arithmetic Mean of
the other two.
If
a, b,
c are in A.
P.,
we
have,
by
definition,
b a
=
c
b;
and therefore 6
=
(a
+
c).
Thus the arithmetic mean
of
two
given quantities
is
half
their sum.
When
any
number of
quantities
are in arithmetical
progression
all the intermediate terms
may
be called
arithmetic means of the two extreme terms.
Between
any
two
given quantities any
number
of
arith-
metic means
may
be inserted.
Let a and b be the two
given quantities,
and let n be
the number of terms to be inserted.
Then b will be the n
+
2th term of the A. P. whose first
term is a.
Hence,
if d be the common
difference,
b
=
a
and therefore d
=
--
.
n
+
1
Then the series is
b a _ 6 a
the
required
arithmetic means
being
b a b a b a
a-\
--
,
a
+
2
-,
...... a
+
n
, ,
......
-
-,
n
+
l
n+1 n+l
na
+
b
(n-l)a
+ 2b
(n
-
2)
a
+
36 a
+
nb
n
+
l'
'
n
+
l
n+l
'
......
'
~n~+l
'
S.A.
258 ARITHMETICAL PROGRESSION.
223. To
find
the sum
of any
number
of
terms
of
an
arithmetical
progression.
Let a be the first term and d the common difference.
Let n be the number of the terms whose sum is
required,
and let I be the last of them.
Then,
since I is the nth
term,
we have
l
=
a
+
(n-l}d
.....................
(i).
Hence,
if S be the
required
sum,
Now write the series in the reverse order
;
then
S
=
I
+
(I
-
d)
+
(I
-
2d)
+
......
+
(a
+
2d)
+
(a
+
d)
+ a.
Hence,
by
addition of
corresponding
terms,
we have
...... to n terms
or,
from
(i),
(iii).
From the formulas
(i), (ii), (iii)
the value of all the
quantities
a, d, n, I,
S can be found when
any
three are
given.
Ex. 1. Find the sum of 20 terms of the arithmetical
progression
S
+
6
+ 9
+
&C.
Here
a=3, d=3, n=20;
.
20
Ex. 2. Shew that the sum of
any
number of consecutive odd
numbers,
beginning
with
unity,
is a
square
number.
The series of odd numbers is
1+3+5+
Here a
=
l, d=2;
hence the sum of n terms is
given by
ARITHMETICAL PROGRESSION. 259
Ex. 3. How
many
terms of the series 1
+
5
+ 9
+
...... must be taken
in order that the sum
may
be 190?
Wehave5'={2o
+
(n-l)d},
where
5=190, o=l,
d=4.
m
Hence n is to be found from the
quadratic equation
or 2n
2
-n- 190
=
0,
that is
(n
-
10) (2ra
+
19)
=
0.
19
Hence n=10. The value n=
-
is to be
rejected
for n must
I
necessarily
be &
positive integer*.
Ex. 4. How
many
terms of the series 5
+
7 + 9 + ...... must be taken
in order that the sum
may
be 480?
Here we have
-
480=0,
or
(n-20)(n+24)=:0.
Hence n must be
20,
for the value n=
-
24 must be
rejected
as a
negative
number of terms is
altogether meaningless*.
Ex. 5. What is the 14th term of the A. p. whose 5th term is 11 and
whose 9th term is 7? Ans. 2.
Ex. 6. What is the 2nd term of the A. p. whose 4th term is b and
whose 7th term is 3a
+
46 ?
Am. -2a-b.
Ex. 7. Which term of the series
5, 8, 11,
<fec. is 320?
Ans. The 106th.
Ex. 8. Shew
that,
if the same
quantity
be added to
every
term of an
A.
P.,
the sums will be in A. p.
Ex. 9. Shew
that,
if
every
term of an A. p. be
multiplied by
the same
quantity,
the
products
will be in A. p.
*
The inadmissible value is a root of the
equation
to which the
problem leads,
but it is not a solution of the
problem. [See Chapter xi.]
It should be remarked that a
negative
value of n cannot mean a number
of terms reckoned backwards.
172
260
GEOMETRICAL PROGRESSION.
Ex. 10. Shew
that,
if between
every
two consecutive terms of au
A.
P.,
a fixed number of arithmetic means be
inserted,
the whole will
form an arithmetical
progression.
lix. 11. Find the sum of the
following
series :
(i) 2i
+
4$
+
6|
+ ...... to 23 terms.
(ii)
2
+
~B~
...... to 12 terms.
(Hi) (a
+
9b)
+
(a
+
7&)
+
(o
+
5b)
+ ...... to 10 terms.
n-1 n-2 n-3
(iv)
h
--
-\
----
h ...... to n terms.
TI n n
Ans.
(i)
621,
(ii) -16,
(iii) 10a, (iv)
I
(n-1).
&
Ex. 12. The 7th term of an A. p. is
15,
and the 21st term is
8;
find
the sum of the first 13 terms. Ans. 195.
Ex. 13. Find the sum of 21 terms of an A. p. whose llth term is 20.
Ans. 420.
Ex. 14. Shew
that,
if
any
odd number of
quantities
are in A.
p.,
the
first,
the middle and the last are in A. p.
Ex. 15. Shew
that,
if
unity
be added to the sum of
any
number of
terms of the series
8, 16, 24, &c.,
the result will be the
square
of an
odd number.
Ex. 16. How
many
terms of the series 15
+
11
+ 7 + ...... must be
taken in order that the sum
may
be 35?
Ant. 5.
Ex. 17. The sum of 5 terms of an A. p. is
-
5,
and the 6th term is
-
13;
what is the common difference?
Ant. -4.
Ex. 18. Find the sum of all the numbers between 200 and 400 which
are divisible
by
7. Ans. 8729.
Ex. 19. If a series of terms in A. p. be collected into
groups
of n
terms,
and the terms in each
group
be added
together,
the results form an
A. P. whose common difference is to the
original
common difference as
*':!.
GEOMETRICAL PROGRESSION.
224. Definition. A series of
quantities
is said to be
in Geometrical
Progression
when the ratio of
any
term
to the
preceding
one is the same
throughout
the series.
GEOMETRICAL
PROGRESSION. 261
Thus
a, b, c, d,
&c. are in Geometrical
Progression
, . b c d
(G.P.)
if
-
=
r
=
-
=
&c.
a b c
The ratio of each terra of a
geometrical progression
to the
preceding
term is called the common ratio.
The
following
are
examples
of
geometrical progressions:
1, 3, 9, 27,
&c.
4, -2, 1, -,
&c.
a,
a
3
,
a
5
,
a
7
,
&c.
In the first series the common ratio is
3,
in the second series it is
-
i,
and in the third series it is a*.
225. If the first term of a G.P. be
a,
and the common
ratio r
; then,
by
definition,
the 2nd term will be
ar,
3rd ar
2
,
4th ar
3
,
and so
on,
the index of r
being always
less
by unity
than
the number
giving
the
position
of the term in the series.
Hence the nth term will be ar""
1
.
We can therefore write down
any
term of a G.P. when
the first term and the common ratio are
given.
For
example,
in the o. p. whose first term is
2,
and whose common
ratio is
3,
the 6th term is 2 x 3
5
,
and the 20th term is 2 x 3
19
.
226. A Geometrical
Progression
is determined when
any
two of its terms are
given.
For,
suppose
we know that the rath term is
a,
and
that the nth term is
J3.
Let a be the first
term,
and r the common ratio
;
then the mth term will be ar"
1
"
1
,
and the nth term will
be ar""
1
.
262 GEOMETRICAL PROGRESSION.
Hence ar
m l
=
a,
ar""
1
=
/3 ;
and .'. r
n "
=
-5
.
P
l-n 1-m
Hence r
=
a.
m
~
n
^
n
~
m
,
and therefore a
=
a
w
~p
>n
~
m
.
Ex. Find the first term of the a. p. whose 3rd term is 18 and whose
5th term IB
40$.
If a be the first
term,
and r the common
ratio,
we have
.81 ,9
0^=18,
tir*=~
;
.:r
3
=2*
4.
Hence a=18x
5
=8.
y
Thus the series is
8,
12, 18,
&c.
227. When three
quantities
are in
G.P.,
the middle
one is called the Geometric Mean of the other two.
If
a, b,
c are in
G.P.,
we have
by
definition
6
=
c. .
b
=
a b
'
~
Thus the
geometric
mean
of
two
given quantities
is
a
square
root
of
their
product.
When
any
number of
quantities
are in
geometrical
progression
all the intermediate terms
may
be called
geometric
means
of the two extreme terms.
Between two
given quantities any
number
of geometric
means
may
be inserted.
For let a and b be the two
given quantities,
and let n
be the number of means to be inserted.
Then 6 will be the
(n
+
2)th
term of a G.P. of which a
is the first term.
Hence,
if r be the common
ratio,
we
have
n+l
Hence the
required
means are
ar,
ar
!
, a?-",
jt_
1 n-l 2 1 n
that
is,
tt
n+1
6
H+1
,
tt*
+1
6*
+1
,
a*
+1
6
n>1
.
-
GEOMETRICAL PROGRESSION. 263
228. To
find
the sum
of any
number
of
terms in
geometrical
progression.
Let a be the first
term,
and r the common ratio. Let
n be the number of the terms whose sum is
required,
and
let I be the last of them.
Then,
since I is the nth
term,
we have I
=
ay""
1
.
Hence,
if S be the
required
sum,
Multiply by
r
;
then
Sr
=
ar
+ ar*+
ar
3
+
......
+ ar"'
1
+ a?-".
Hence,
by
subtraction,
S-Sr
=
a- ar
n
;
1 r
'
Ex. 1. Find the sum of 10 terms of the series
3, 6, 12,
&o.
Here
a
=
3,
r=2,
n
=
10.
1-2
10
Hence S=3
-=3(2
10
-1)
=
30G9.
1 i
229. From the
preceding
article we have
l-r*
_
a ar
n
\r \r 1 r'
Now when r is a
proper fraction,
whether
positive
or
nega-
tive,
the absolute value of r
n
will decrease as n increases
;
moreover the value of r
n
can be made as small as we
please by sufficiently increasing
the value of n.
Hence,
when r is
numerically
less than
unity,
the sum
of the series can be made to differ from
-
by
as small
a
quantity
as we
please by taking
a sufficient number of
terms.
264
GEOMETRICAL PROGRESSION.
Thus the sum of an
infinite
number of
terms of the
geometrical progression
a
+
ar
+
ar*
+ in which r is
numerically
less than
unity,
is -=
.
Ex. 1. Find the sum of an infinite number of terms of the series
9-6+4-
Here
Hence
Ex. 2. Find the
geometrical progression
whose sum to
infinity
is
4J,
and whose second term is
-
2.
Let a be the first
term,
and r be the common ratio.
Then we have ar=
-
2,
and :
-
=
-
.
1-r 2
Whence
1
5
.
Q
1 4
Hence r=-
5
,orr=
5
2
and the series is
6,
-
2,
-
,
<fcc.
4
The value r=
3
is
inadmissible,
for r must be
numerically
less than
unity.
Ex. 3. The 3rd term of a o.p. is
2,
and the 6th term is
-
J
;
what is
the 10th term? Ant.
-fa.
Ex. 4. Insert two
geometric
means between 8 and
-
1,
and three
means between 2 and 18. Am.
-
4, 2;
2^3,
6,
6/B.
Ex. 5. Shew that if all the terms of a Q.P. be
multiplied by
the same
quantity,
the
products
will be in o.p.
Ex. 6. Shew that the
reciprocals
of the terms of a o.p. are also in Q.P.
Ex. 7. Shew
that,
if between
every
two consecutive terms of a
o.p.,
a
fixed number of
geometric
means be
inserted,
the whole will form a
geometrical progression.
HARMONICAL
PROGRESSION. 265
Ex. 8. Find the sum of the
following
series :
(i)
12
+ 9 +
62+...
to 20 terms.
(ii) 1-5
+
^
+ ...
to 6 terms.
j y
(iii)
4
+
-8
+
-16 -)-...
to
infinity.
Am.
(i)
48
|l -(!)"},
(ii)|g,
(iii)
5.
Ex. 9. Shew that the continued
product
of
any
number of
quantities
n
in
geometrical progression
is
equal
to
(gl)*,
where n is the number of
the
quantities
and
g,
I are the
greatest
and least of them.
Ex. 10. Shew that the
product
of
any
odd number of terms of a G.P.
will be
equal
to the nth
power
of the middle
term,
n
being
the number
of the terms.
Ex. 11. The sum of the first 10 terms of a certain G.P. is
equal
to 244
times the sum of the first 5 terms. What is the common ratio ?
Ana. 3.
Ex. 12. If the common ratio of a G.P. be less than
J,
shew that each
term will be
greater
than the sum of all that follow it.
HARMONICAL PROGRESSION.
230. Definition. A series of
quantities
is said to be
in Harmonical
Progression
when the difference between
the first and the second of
any
three consecutive terms is
to the difference between the second and the third as the
first is to the third.
Thus
a, b, c,
d
&c.,
are in Harmonical
Progression
[H. P.],
if
a b : b c :: a :
c,
b c : c d:: b :
d,
and so on.
If
a, b,
c be in harmonical
progression,
we have
by
definition
a b : b c :: a :
c;
,\
c(a
b)
=
a(b- c).
266 HARMOXICAL PROGRESSION.
Hence,
dividiug by
abc,
we have
which shews that
-
,
j-
,
-
are in arithmetical
progression.
Thus,
if quantities
are in harmonical
progression,
their
reciprocals
are in arithmetical
progression.
231. Harmonic Mean. If
a, b,
c be in
harmonical
progression,
-
, r,
-
will be in arithmetical
progression.
a o c
211
Hence
r
=
-
+
-
;
b a c
2ac
.*. o
=
a
+ c
Thus
the harmonic mean
of
two
quantities
is twice
their
product
divided
by
their sum.
If we
put
A, G,
H for the
arithmetic,
the
geometric,
and the harmonic means
respectively
of
any
two
quantities
a and
b,
we have
Thus the
geometric
mean
of any
two
quantities
is also
the
geometric
mean
of
their arithmetic and harmonic
means.
232. Theorem. The arithmetic mean
of
two
unequal
positive quantities
is
greater
than their
geometric
mean.
If
a,
b be the two
positive quantities
we Lave to shew
that
(a
+
b)
>
or a
-
v/&
2
> 0.
EXAMPLES. 267
Now
(\/a V^)
2
is
always positive,
and therefore
greater
than
zero,
unless a
=
b.
Since the arithmetic mean of two
positive quantities
is
greater
than their
geometric
mean,
it follows from Art. 231
that the
geometric
mean is
greater
than the harmonic.
233. To insert n harmonic means between
any
two
quantities
a and b.
Insert n arithmetic means between
-
and
r ,
and the
a o
recipi'ocals
of these will be the
required
harmonic means.
The arithmetic means are
1 1
/I
_
IN 1 2 /I
_
IN
a
+
n
+
l(b a)'
a
+
n+l
(b
a)'
Hence,
by simplifying
these terms and
inverting
them,
the
required
harmonic means will be found to be
(n
+
1)
ab
(n
+
l)ab (n
+
l)ab
nb
+
a
'(n-I)b
+
2a'"
"
'
b
+
naT
234. It is of
importance
to notice that no formula can
be found which will
give
the sum of
any
number of terms
in harmonical
progression.
EXAMPLES XXI.
1. Shew
that,
if
a, b,
c be in A.
p.,
then will
a*(6
+
c),
b*
(c
+
a),
<?
(a
+
b)
be in A. p.
2. Find four numbers in A. p. such that the sum of their
squares
shall be
120,
and that the
product
of the first and last
shall be less than the
product
of the other two
by
8.
3. If
a,
6,
c be in A.
p.,
and
6, c,
d be in H.
p.,
then will
a : b
=
c : d.
4.. Find three numbers in G. p. such that their sum is
14,'
and the sum of their
squares
84.
268
EXAMPLES.
5. If
a, b,
c be in arithmetical
progression,
and x be the
geometric
mean of a and
b,
and
y
be the
geometric
mean of b
and
cj
then will x
a
, b*,
y
3
be in arithmetical
progression.
6. Shew
that,
if
a, b,
c be in harmonical
progression,
then
will
^ ,
-
7
and
-
,
be also in harmonical
b + c a c + a-u a+bc
progression.
7. Shew
that,
if
a, 6, c,
d be in harmonical
progression,
then will
8. Shew
that,
if
a, b,
c be in harmonical
progression,
2
=
1
_1_
6 b a b c'
9.
Shew
that,
if
a, b,
c be in
H.P.,
then will
b+a b+c
b a b c
=
2.
10. If
a, b,
c be in A.
p., 6, c,
d in G.
P.,
and
c, d,
e in H. p.
;
then will
a, c,
e be in
G. P.
11. If
a, b,
c be in H.
P.,
then will a
-,
^
,
c
^
be in G.P.
12. If
a, b,
c are in H.
p.,
then
a,
a
c,
a b are in H.
p.,
and also
c,
c
a,
c b are in H. P.
13. If
x,
a,,
a
t
, y
be in A.
p., x, g
lt
#
g
,
y
in
G.P.,
and
x,
A,,
/t
g
,
y
in
H.P.,
then
A,/tg
A
t
+
A,
'
14. The sum of the
first, second,
and third terms of a G. P.
is to the sum of the
third,
fourth and fifth terms as 1 :
4,
and
the seventh term is 384. Find the series.
EXAMPLES.
269
15. If
a,,
a
g
,
a
3
, ,
a_
be in harmonical
progression,
prove
that
a,a,
+
a
s
a
3
+
a
:i
a
t
+ + a
n
_
l
a
m
=
(n
-
1)
a
j
a
n
,
16. If
a, x,
y,
b be in arithmetical
progression,
and
a, w, v,
b be in harmonical
progression,
then xv
=
yu
=
ab.
17. Three numbers are in arithmetic
progression,
and the
product
of the extremes is 5 times the mean
;
also the sum of
the two
largest
is 8 times the least. Find the numbers.
io-r
a +
b,b+
C,. ,
1 ...
18. If
, r , 6,
-
=
be in A. P. : then a.
r ,
c will be
1 ab 1 be b
iu H. P.
19. If
a, b,
c be in A.
P.,
and a
3
,
b*
}
c* be in n.
p., prove
that
, b,
c are in G.
p.,
or else a
=
6
=
c.
2
20. If x be
any
term of the arithmetical
progression
and
y
be the
corresponding
term of the harmonical
progression
whose
first two terms are
a, b,
then will x a :
y
a :: b :
y.
21. Shew
that,
if a be the arithmetic mean between b and
c,
and 6 be the
geometric
mean between a and
c,
then will c be
the harmonic mean between a and b.
22. The series of natural numbers is divided into
groups
as
follows:
1; 2, 3;
4,
5, 6;
7,
8, 9, 10;
and so on. Prove that
the sum of the numbers in the &
th
group
is
^k (k* +1).
23. An A. P. and an H.P. have each the first term
a,
the
same last term
I,
and the same number of terms n
; prove
that
the
product
of the
(r
+
l)
th
term of the one series and the
(n r)
th
term of the other is
independent
of r.
24. Terms
equidistant
from a
given
term of an A. P. are
multiplied together
;
shew that the differences of the successive
terms of the series so formed are in A. p.
25. Shew
that,
if S
m
,
S
an
,
S
an
be the sum of n
terms,
of
2
terms,
and of 3n terms
respectively
of
any
O.P.,
then will
S.V^-SJ-^-Sf.
26. If
a, b,
c be all
positive
and either in A.
p.,
in
O.P.,
or in H.
P.,
and n be
any positive integer,
then a" + c*
> 26".
270 EXAMPLES
27. If
P, Q,
R be
respectively
the
j
th
, q
ih
,
and r"
1
terms
(i)
of au A.
P.,
(ii)
of a G.
P.,
and
(iii)
of an H.
p.,
then will
(i)
(ii)
(iii)
28. Shew
that,
if
a,,
a.,,
a
a ,
......
,
a
n
be in H.
P.,
then
a, a_
a
a
s
+ a
3
+ ... +a
n a, +, , ii
_
will be in H. P.
29. Shew
that,
if
a,,
a
s ,
3
,
a,,
be all
real,
and if
(a'
+ a' + +
<-,)(/
+
<+ +O
=
(at
a
t
+
a
a
a
a
+ +
a.^aj
2
,
then will
a,,
a
a
,
a
3
,
be in G.P.
30. Shew that
any
even
square, (2rz)
z
,
is
equal
to the sum
of n terms of one series of
integers
in A.
P.,
and that
any
odd
square, (2n
+
1)",
is
equal
to the sum of n terms of another A. p.
increased
by unity.
31. Prove that
any positive integral power (except
the
first)
of
any positive integer, p,
is the sum of
p
consecutive
terms of the series
1, 3, 5, 7,
&c.
;
and find the first of the
p
terms when the sum is
p
r
.
32. If an A. p. and a G. P. have the same first term and the
same second
term, every
other term of the A. p. will be less
than the
corresponding
term of the G.
p.,
the terms
being
all
positive.
CHAPTER XVIIL
SYSTEMS OF NUMERATION.
235. IN arithmetic
any
number whatever is
repre-
sented
by
one or more of the ten
symbols
0,1, 2, 3, 4, 5, 6,
7, 8, 9,
called
figures
or
digits, by
means of the convention
that
every figure placed
to the left of another
represents
ten times as much as if it were in the
place
of that other.
The
cipher,
0,
which stands for
nothing,
is
necessary
because one or more of the
denominations, units, tens,
hundreds, &c.,
may
be
wanting.
The above mode of
representing
numbers is called the
common scale
of
notation,
and 10 is said to be the radix or
base.
236. Instead of ten
any
other number
might
be used
as the base of a
System of
Numeration,
that is of a
system
by
which numbers are named
according
to some definite
plan,
and of the
corresponding
Scale
of
Notation,
that is
of a
system by
which numbers are
represented by
a few
signs according
to some definite
plan
;
and to
express
a
number, N,
in the scale whose radix is
r,
is to write the
number in the form .....
djl^d^d^,
where each of the
digits
d^d^d^d^
...... is less than
r,
and where d stands for d
units,
d
l
stands for
d^
x
r,
d
3
for d
a
x
r\
and so on.
Thus N=
Note.
Throughout
this
chapter
each letter stands for a
positive integer,
unless the
contrary
is stated.
272 SYSTEMS OF NUMERATION.
237.
Theorem.
Any positive integer
can be
expressed
in
any
scale
of
notation,
and this can be done in
only
one
way.
For divide N
by
r,
and let
Q
l
be the
quotient
and d
the remainder.
Then N
=
d
+
rxQ
l
.
Now divide
Q, by
r,
and let
Q
2
be the
quotient
and d
t
the remainder.
Then
Q
r
=
d
t
+
rQ
8
;
therefore N
=
d
+
rd
t
+
r*Q
s
.
By proceeding
in this
way
we must sooner or later
come to a
quotient,
Qn
= d
n
,
which is less than
r,
when the
process
is
completed,
and we have
N
=
d
9
+
rd
l
+
r*d
t
+
i*d
t
+ ...... r
n
d
n
,
so that the number would in the scale of r be written
d
Each of the
digits
d
,
d
v
d
v
...... is less than
r,
and
any
one or more of
them,
except
the
last,
d
n
,
may
be zero.
Since at
every stage
of the above
process
there is
only
one
quotient
and one remainder the transformation is
unique.
The
given
number N
may
itself be
expressed
either in
the common or in
any
other scale of notation.
Ex. 1.
Express
2157
in the scale of 6.
The
quotients
and remainders of the successive divisions
by
6
ore as under :
6
|
2157
6
|jJ59
remainder 3
=
rf
6[59
............
5
=
^
1
............
3=d,
Thus 2157 when
expressed
in the scale of 6 is 13553.
Ex. 2.
Change
13553 from the scale of 6 to the scale of 8.
We have the
following
successive divisions
hy 8, remembering
that since 13553 is in the scale of 6 each
figure
is six times what
it would be if it were moved one
place
to the
left,
so that to
begin
with we have to divide 1x6 +
3,
and so on.
SYSTEMS OF
NUMERATION.
273
8
|
.13653
8
[1125
remainder 5
8[53
5
4 1
Hence the number
required
is 4155.
Ex. 3.
Change
4155 from the scale of 8 to the scale of
10.
Proceeding
as
before,
we have
10
14155
10(327
remainder 7
10~[25
5
2 1
Thus 2157 is the number
required.
Or thus:
Since
4155=4 x8
3
+
l x8
2
+
5x8
+
5=
{(4x8
+
1)8
+
5}
8+5,
the
required
result
may
be obtained as follows :
Multiply
4
by
8 and add 1
;
multiply
this result
by
8 and add 5
then
multiply again by
8 and add 5.
Ex. 4.
Express
3166 in the scale of 12.
[Eepresent
ten
by
t,
and
eleven
by .]
Ana. IQet.
Ex. 6.
Express
^
in the scale of 4. Am.
^pj
.
_1 ill
Ex. 6. In what scale is 4950 written 20301? Ans. 7.
238. Radix Fractions. Radix fractions in
any
scale
correspond
to decimal fractions in the
ordinary
scale,
so that
a b c
'abc. . . stands for
-
+
-3
+
-
+
r r r
To shew that
any given fraction may
be
expressed by
a
series
of
radio;
fractions
in
any proposed
scale.
Let F be the
given
fraction;
and
suppose
that,
when
expressed by
radix fractions in the scale of
r,
we have
a b c
F='<*c
=
-
+
^
+
^+
where each of
a, 6,
c is a
positive integer (including
zero)
less than r.
S. A.
18
274
SYSTEMS OF NUMERATION.
Multiply by
r;
then
Fxr=a
-
+
~
r r*
Hence a must be
equal
to the
integral part,
and
-
+ -s +
must be
equal
to the fractional
part
of Fr.
r r
(If
Fr be less than
1,
a is
zero.)
Let F
t
be the fractional
part
of
Fr;
then
F
=-
-
i
r r
*
Multiply by
r;
then
Q
1
r
Hence b must be
equal
to the
integral part
of
F^.
Thus
a,b,c,
can be found in succession.
Ex. 1.
Express
5-
by
a series of radix fractions in the scale of 6.
Hence -012 is the
required
result.
Ex. 2.
Express
-
by
a series of radix fractions in the scale of 3.
= - =
-
?x8
=
2
+
l;
*x3
=
H-|; |
X 3
= 2
+ 1.
Hence -610212 is the
required
result.
Ex. 3.
Change
324-26 from the scale 8 to the scale
G.
The
integral
and fractional
parts
must be considered
separately.
SYSTEMS OF NUMERATION.
275
6|324
6
1
43 remainder 2
5 5
276
SYSTEMS OF NUMERATION.
Ex. 1. The difference of
any
two numbers
expressed by
the same
digits
is divisible
by
r- 1.
For the sum of the
digits
is the same for both
;
and since N,-S
and N
t
-S are both divisible
by
r-1,
it follows that
^-N^
is
divisible
by
r
-
L
Ex. 2. Shew that in the
ordinary
scale a number is divisible
by
9 if
the sum of its
digits
be divisible
by 9,
and
by
3 if the sum of its
digits
is divisible
by
3.
N- S is a
multiple
of 9
; hence,
if S be a
multiple
of
9,
so also is
N
; and,
if S be a
multiple
of
3,
so also is N.
Ex. 3. Shew that
any
number is divisible
by
r+1 if the difference
between the sum of the odd and the sum of the even
digits
is
divisible
by
r
+
1.
Let
and D
Then N-
Each of the terms on the
right
is divisible
by
r
+
1
[Art. 87J ;
.'. N-D is divisible
by
r
+
1. Hence if D is divisible
by
r
+
1 so also
is AT.
Ex. 4. If N, and N
a
be
any
two whole
numbers,
and if the remainders
left after
dividing
the sum of the
digits
in
.N\
,
N^
and in N
l xNy by
9 be
nj,
n
t
and
p respectively ;
then will
t^n,
be
equal
to
p,
or differ
from
p by
a
multiple
of 9.
For
#!
=
!+
a
multiple
of
9,
&udX.,=n
t
+
&
multiple
of
9;
therefore N,^N
t
=n
1
xn
i
+&
multiple
of 9." Hence
n^-t-
a
multiple
of 9 is
equal
to
p
+
a
multiple
of 9.
If the above is
applied
in
any
case of
multiplication,
and it is
found that
n^n2
does not
equal p,
or differ from it
by
a
multiple
of
9,
there must be some error in the
process
of
multiplication.
This
gives
a method of
testing
the
accuracy
of
multiplication;
the test is not however a
complete one,
for
although
it is certain that
there must be an error if n
:
x
n,
does not
equal p,
or differ from it
by
a
multiple
of
9,
there
may
be errors when the condition is
satisfied,
provided
that the errors neutralize one another so far as the sum of
the
digits
in the
product
is concerned.
This is called the "Rule for
casting
oat the nines."
Ex. 5. A number of three
digits
in the scale of 7 has the same
digits
in reversed order when it is
expressed
in the scale of 9 : find the
number.
Let
a, b,
c be the
digits ;
then we have
49a + 76 +c
=
81c
+ 96
+0,
where
a, b,
c are
positive integers
less than
7.
Hence
EXAMPLES. 277
Now 40c and 24 are both divisible
by
8;
therefore b must be
divisible
by
8. But b is less than 7 : it must therefore be zero. And
since b is
zero,
we have 5c
=
3o,
which can
only
be satisfied when
c=3 and a=5.
Thus the number
required
is 503.
Ex. 6. A number
consisting
of three
digits
is doubled
by reversing
the
digits ;
prove
that the same will hold for the number formed
by
the
first and last
digits,
and also that such a number can be found in
only
one scale of notation out of
every
three.
Let the number be abc in the scale of r.
Then we have
(a&c)
x 2
=
cba.
Since cba is
greater
than
abc,
e must be
greater
than a.
Hence we must have the
following equations
:
(i),
(ii),
(iii).
From
(i)
and
(iii)
we see that the number
represented by
ca is
double that
represented by
ac.
Also 4a
+
2
=
2c=a+r;
.-. r-2
=
3a.
Hence,
as a is an
integer,
r-2 must be a
multiple
of
3,
so that
the number must be in one of the scales
2, 5, 8, 11, &c.,
the numbers
corresponding
to these scales
being Oil, 143,
275, 3t7,
<fec.
EXAMPLES XXII.
1. Find the number which has the same two
digits
when
expressed
in the scales of 7 and 9.
2. In
any given
scale write down the
greatest
and the
least number which has a
given
number of
digits.
3. A number of six
digits
is formed
by writing
down
any
three
digits
and then
repeating
them in the same order
;
shew
that the number is divisible
by
1001.
4. Of the
weights 1, 2, 4, 8,
&c.
Ibs.,
which must be taken
to
weigh
1027 Ibs. ?
278
EXAMPLES.
5. Shew that the number
represented
in
any
scale
by
144
is a
square
number.
6. Shew that the numbers
represented
in
any
scale
by
121, 12321,
and 1234321 are
perfect squares.
7. Find a number of two
digits,
which are
transposed by
the addition of 18 to the
number,
or
by converting
it into the
septenary
scale.
8. A number is denoted
by
4 '440 in the
quinary scale,
and
by
4'54 in a certain other scale. What is the radix of
that other scale 1
9. If S be the sum of the
digits
of a number
N,
and
2Q
be the sum of the
digits
of
2^,
the number
being expressed
in
the
ordinary scale,
shew that
~
Q
is a
multiple
of 9.
10. If a whole number be
expressed
in a scale whose
radix is
odd,
the sum of the
digits
will be even if the number
be
even,
and odd if the number be odd.
11. Prove
that,
in
any
scale of
notation,
the difference of
the
square
of
any
number of three
digits
and the
square
of the
number formed
by reversing
the
digits
is divisible
by
r
3
1.
12. Prove
that,
in
any
scale of
notation,
the difference of
the
square
of
any
number and the
square
of the number formed
by reversing
the
digits
is divisible
by
r* 1.
13. A number of three
digits
in the scale of
7,
when
expressed
in the scale of 11 has the same
digits
in reversed
order : find the number.
14. Prove that all the numbers which are
expressed
in
the scales of 5 and 9
by using
the same
digits,
whether in the
same order or in a different
order,
will leave the same remainder
when divided
by
4.
15. There is a certain number which is
expressed by
6
digits
in the scale of
3,
and
by
the last three of those
digits
in
the scale of 12. Find the number.
EXAMPLES. 279
16. Find a number of four
digits
in the scale of 8 which
when doubled will have the same
digits
in reverse order.
17. The
digits
of a number of three
digits
are in A. p.
The number when divided
by
the sum of its
digits gives
a
quotient
15
;
and when 396 is added to the
number,
the sum
has the same
digits
in inverted order. Find the number.
18. Find the
digits a, b,
c in order that the number
13a645c
may
be divisible
by
792.
19. Prove that there is
only
one scale of notation in
which the number
represented by
1155 is divisible
by
that
represented by
12,
and find that scale.
20. Find a number of four
digits
in the
ordinary
scale
which will have its
digits
reversed in order
by multiplying
by
9.
21. In the scale of notation whose radix is
r,
shew that
the number
(r
8
l)(r" 1)
when divided
by
r 1 will
give
a
quotient
with the same
digits
in the reverse order.
22. Shew
that,
in
any
scale of
notation,
the
circulating period consisting
of all the
figures
in order
except
r 2 which is
passed
over. For
example,
in the
ordinary scale,
^
=
012345679.
23. There is a number of six
digits
such that when the
extreme left-hand
digit
is
transposed
to the extreme
right-hand,
the rest
being
unaltered,
the number is increased three-fold.
Prove that the left-hand
digit
must be either 1 or
2,
and find
the number in either case.
24. Find a number of three
digits,
the last two of which
are
alike,
such that when
multiplied by
a certain number it
still consists of three
digits,
the first two of which are alike
and the same as the former
repeated
ones,
and the third is the
same as the
multiplier.
CHAPTER XIX.
PERMUTATIONS AND
COMBINATIONS.
240. Definition. The different
ways
in which r
things
can be taken from n
things, regard being
had to
the order of selection or
arrangement,
are called the
per-
mutations of the n
things
r at a time.
Thus two
permutations
will be different unless
they
contain the same
objects arranged
in the same order.
For
example, suppose
we have four
objects, represented
by
the letters
a, b, c, d;
the
permutations
two at a time
are
ab, ba, ac, ca, ad, da, be, cb, bd, db, cd,
and dc.
The number of
permutations
of n different
things
taken r at a time is denoted
by
the
symbol n
P
r
.
241. To
find
the number
ofpermutations of
n
different
things
taken r at a time.
Let the different
things
be
represented by
the letters
a, b, c,
It is obvious that there are n
permutations
of the n
things
when taken one at a
time,
so that
n
P
l
=
n,
Now in the
permutations
of the n letters r
together,
the number of
permutations
in which a
particular
letter
occurs first in order is
equal
to the number of
permuta-
tions of the
remaining
n 1 letters r 1 at a time. This
is true for each one of the n
letters,
and therefore
PERMUTATIONS. 281
Since the above relation is true for all values of n
and
r,
we have in succession
^
=
(*
-
r
+
2)
x
B
_
r+1
P
r
_
r+1
.
But
n
-
r+1^
=
("-r+l).
Multiply
all the
corresponding
members of the above
equalities,
and cancel all the common
factors;
we then
have
.Pr
=
n
(n
-
l)(w
-
2) (n-r+I).
If all the w
things
are to be
taken,
r is
equal
to
n,
and
we have
n
P
n
=
n
(n
-
1) (n
-
2)
3.2.1.
Definitions. The
product
n
(n
-
1) (n
-
2)
. . . 2 . 1
is denoted
by
the
symbol
[n_
or
by
n! The
symbols
[w
and w ! are read
'
factorial n'
The continued
product
of the r
quantities
n,n l,
n
2,
(TO r+1),
n not
being necessarily
an
integer
in this
case,
is denoted
by
n
r
. Thus
n^
=
n(n 1) (n >2).
Hence we have
n
P
n
=
[n
,
and
n
P
r
=
w
r
.
%
242. To
find
the number
of permutations of
n
things
taken all
together,
when the
things
are not all
different.
Let the n
things
be
represented by
letters;
and
sup-
pose p
of them to be
a's,
q
of them to be
b's,
r of them to
be
c's,
and so on. Let P be the
required
number of
per-
mutations.
282
PERMUTATIONS.
If in
any
one of the actual
permutations
we
suppose
that the a's are all
changed
into
p
letters different from
each other and from all the
rest;
then,
by changing only
the
arrangement
of these
p
new
letters,
we
should,
instead
of a
single permutation,
have
\p
different
permutations.
Hence,
if the a's were all
changed
into
p
letters
different from each other and from all the
rest,
the
6's, c's,
&c.
being
unaltered,
there would be P x
|p
permutations.
Similarly,
if in
any
one of these new
permutations
we
suppose
that the 6's are all
changed
into
q
letters different
from each other and from all the
rest,
we should obtain
\q
permutations by changing
the order of these
q
new
letters. Hence the whole number of
permutations
would
now be P x
[p
x
jj.
By proceeding
in this
way
we see that if all the letters
were
changed
so that no two were
alike,
the total number
of
permutations
would be
Px\p
x
\q
x
|r...
But the number of
permutations
all
together
of n
different
things
is
\n.
Hence P x
\p
x
\qx
|r...= |n;
In
p
=
L=
\p\q\r...'
Ex. 1. Find
6
P
3
,
5
P
4
and
7
P
7
. Ant.
120, 120,
5040.
Ex. 2. Shew that
lo
P
4 =jP7
.
Ex. 3. If
m
P
5
=
12 x ? find n. Am. 7.
Ex. 4. If
^P,=
100 x
W
P
2
,
find n. Ans. 13.
Ex. 5. If
a,P,=2
x
W
P
4
,
find n. Ans. 8.
Ex. 6. Find the number of
permutations
of all the letters of each of
the words
acacia, Hannah,
success and
Mississippi.
Ans.
60, 90, 420,
34650.
Ex. 7. In how
many ways may
a
party
of 8 take their
places
at a
round
table;
and in how
many ways
can 8 different beads be
strung
on a necklace? Ans.
|7,
i
[7.
Ex. 8. In how
many ways may
a
party
of 4 ladies and 4
gentlemen
be
arranged
at a round
table,
the ladies and
gentlemen
being placed
alternately?
Ant. 144.
COMBINATIONS. 283
Ex. 9.
The number of
permutations
of n
things
all
together
in which
r
specified
things
are to be in an
assigned
order
though
not
necessarily
consecutive is
\nj\r.
Ex. 10. The number of
ways
in which n books can be
arranged
on a
shelf so that two
particular
books shall not be
together
is
(n
-
2) |n-l.
Ex. 11. Find the number of
permutations
of n
things
r
together,
when
each
thing
can be
repeated any
number of times.
Here
any
one of the n
things
can be
put
in the first
place; and,
however the first
place
is
filled,
any
one of the n
things
can be
put
in
the second
place;
and so on. Hence the number
required
=nx wxnx ...
=
n
r
.
COMBINATIONS.
243. Definition. The different
ways
in which a
selection of r
things
can be made from n
things,
without
regard
to the order of selection or
arrangement,
are called
the combinations of the n
things
r at a time.
Thus the different combinations of the letters
a, b, c,
d
three at a time are
abc, abd,
acd and bed.
The number of combinations of n different
things
r at
a time is denoted
by
the
symbol jOr
.
244. To
find
the number
of
combinations
of
n
different
things
taken r at a time.
Let the different
things
be
represented by
the letters
a, b, c,
...
Now in the combinations of the n letters r
together
the number in which a
particular
letter occurs is
equal
to
the number of combinations of the
remaining
n 1 letters
r 1 at a time. Hence in the whole number of combina-
tions r
together
every
letter occurs
H
.
l
C
r
.
l
times,
and
therefore the total number of letters is n x
n
_
l
G
r
_
l
;
but,
since there are r letters in each
combination,
the total
number of the letters must be r x C.
284 COMBINATIONS.
Hence r x G=nx
,0, ,
n r
it i '**
Since the above relation is true for all values of n and
of
r,
we have in succession
(r-2)x n
_
2
(7
r
_
2 =-(7
l
-2)x n
.
3
(7
r
.
9
,
Also
-
Hence,
by multiplying corresponding
members of the
above
equations
and
cancelling
the common
factors,
we
have
(r
x.O
r =n(7i-l)(n-2) (n-r
+
l),
th*tis
n
C
r
=
n
(
n
-
l
K
n
-V-(
n
-
r
+
V
=
^ (i).
vr Ir
v '
By multiplying
the numerator and denominator of the
fraction on the
right by
In
r,
we have
n
(n 1) (n 2).
.
.(n
r
+
1)
x
|n
r
" r
|r
In r
ttn
By comparing
the above result with that obtained in
Art.
242,
it will be seen that
JP
r
=
n
C
r
x
|
r. The relation
JP
r
=
J0
r
x
[r
can however be at once obtained from the
consideration that
every
combination of r different
things
would
give
rise to
[r
permutations,
if the order
of the
letters were altered in
every possible way.
Note. In order that the formula
(ii) may
be true
when r
=
n,
we must
suppose
that
1
=
1,
since
n
C
n
1.
We should also obtain the result
JO
=
1
by putting
n
=
1 in
the relation
|n
=
n
\n
1.
COMBINATIONS.
285
245. Theorem. The number
of
combinations
of
n
different things
r
together
is
equal
to the number
of
the
combinations n r
together.
The
proposition
follows at once from the fact that
whenever r
things
are taken out of n
things,
n r must
be
left,
and if
every
set of r
things
differs in some
par-
ticular from
every
other,
the
corresponding
set of n r
things
will also differ in some
particular
from
every
other
;
and therefore the number of different
ways
of
taking
r
things
must be
just
the same as the number of different
ways
of
leaving
or
taking
n r
things.
The result can also be obtained from the formula
(ii)
of the last Article.
For
n
C
r
\r\n-
It should be remarked that the first method of
proof
holds
good
when the n
things
are not all
different,
to which
case the formulae of Art. 244 are not
applicable.
Ex. 1. Find
10
<7
4
,
^fit
and
20^17-
^"*-
210, 220,
1140.
Ex.2.
If,,C,= nC'i2
,find
1,C',j.
Am. 153.
Ex.3. Find
n, having given
that
B
C
B
=
nC',.
Ans. 11.
Ex. 4. Find
n, having given
that 3 x
n
C7
4
=5 x
,-^j.
Ans. 10.
Ex.
5. Find
n, having given
that
n
(7
4
=210. Ans. 10.
Ex. 6. Find n and r
having given
that
n
P
r
=272 and
B
C
r
=136.
Ans.
n=17,
r=2.
Ex. 7. Find n and r
having given
n
C
r
.
l
:
B
C
r
:
B
(7
H
.
1
:: 2 : 3 : 4.
Ans.
n=34,
r=14.
Ex. 8. How
many
words each
containing
3 consonants and 2 vowels
can be formed from 6 consonants and 4 vowels ?
The consonants can be chosen in
6<7,=20 ways;
the vowels can be
chosen in
4
C
2
=6
ways;
hence 20 x 6 different sets of letters can be
chosen,
and each of these sets can be
arranged
in
B
P
5
=120
ways.
Hence the
required
number is 20 x 6 x 120.
286
COMBINATIONS.
Ex. 9. How.
many
different sums can be formed with a
sovereign,
a
half-sovereign,
a
crown,
a
half-crown,
a
shilling
and a
sixpence?
Number
required
=
t
C
1
+
6
(7
t
+
t
C
t
+
t
C
4
+
,(7,
+
t
C
e
=
63.
Ex. 10. Shew
that,
in the combinations of 2n different
things
n
together,
the number of combinations in which a
particular thing
occurs is
equal
to the number in which it does not occur.
Ex. 11. Shew
that,
in the combinations of 4n different
things
n
together,
the number of combinations in which a
particular thing
occurs is
equal
to one-third of the number in which it does not
occur.
Ex. 12. Out of a
party
of 4 ladies and 3
gentlemen
one
game
at lawn-
tennis is to be
arranged,
each side
consisting
of one
lady
and one
gentleman.
In how
many ways
can this be done ? Ans. 36.
Ex. 13. The
figures 1, 2, 3, 4,
5 are written down in
every possible
order: how
many
of the numbers so formed will be
greater
than
230007
Ans. 90.
Ex. 14. At an election there are four candidates and three members
to be
elected,
and an elector
may
vote for
any
number of candidates
not
greater
than the number to be elected. In how
many ways may
an elector vote?
An*. 14.
246. Greatest value of
n
C
r
. To
find
the
greatest
valtie
of n
C
r
for
a
given
value
of
n.
From the formulae of Art. 244 we have
Hence
n
C
r
=
n
C
r
_
1
,
according
as n r
+
1
=
r
;
that
is,
according
as r =
% (n
+
1).
Thus the number of combination of n
things
r
together
increases with r so
long
as r is less than
(n
+
1).
If then n be
even, m
O
r
is
greatest
when r
=
^
.
H n be
odd, .C
r >,<?,_,
as
r<(n
+
l),
and (?
=
(?,_,
when
r=(7i+l).
Thus,
when n is
odd,
these are the
greatest
values of
n
C
r
.
COMBINATIONS. 287
For
example,
if
n=10, n
C
r
is
greatest
when r=5. Also if
n=ll,
n
C
r
is the same for the values 5 and 6 of
r,
and
n
C
r
is
greater
for these
values than for
any
other value.
247. To
jn-ove
that
n
C
r
+
n
C
r_,
=
n+1
(7
r
.
If the total number of combinations of
(n
+
1) things
r
together
be divided into two
groups according
as
they
.do or
do not contain a certain
particular
letter,
it is clear that the
number of the combinations which do not contain the
letter is the number of combinations r
together
of the
remaining
n
things,
and the number of the combinations
which do contain the letter is the number of
ways
in which
r 1 of the
remaining
n
things
can be taken. Thus
The above result can also be
proved
as follows :
From Art. 244 we have
n(n-l)...(n-r+l) n(n- l)...(n
1.2 ...... r 1.2 ......
(r-1)
1.2. .....r
_
(n
+
I
)
n
(n
-
1)
. . .
(n
-
r +
2) _
1.2 ...... r
-+ir-
Ex. To
prove
that
nfl
P
r
=
n
P
r
+
r
^P^.
A
particular thing
is absent in
n
P
r
of the
permutations
of the
(n
+
1) things
and occurs in
n
P
r_j
;
also when it does occur it can be
in either of r
places.
Hence
n+l^r
=
n
P
r
+
r
n*V-l
248. Theorem. To
prove
that,
if
an and
y
be
any
two
positive integers
such that
x-\-y
=
m,
then will
Suppose
that ra letters
a, 6,...,
p, <?,...,
are divided
into two
groups
a,b,... t
and
p, q,...,
there
being
x and
y
letters
respectively
in these
groups.
Then the whole
number of sets of n out of the ra
things
will
clearly
be
288 COMBINATIONS.
equal
to the sum of the number of sets formed
by taking
n out of the first
group
and none out of the
second,
n I
out of the first
group
and one out of the
second,
n 2
out of the first
group
and 2 out of the
second,
and so on.
Now n can be chosen from the first
group
in
X
C
H
ways.
Also n 1 can be chosen from the first
group
in
X
G
W
_
1
ways,
and
any
one of these sets of n 1
things
can be
taken with
any
one of the
VG^
sets of 1 from the second
group,
so that the number of sets formed
by taking
n 1
from the first
group
and 1 from the second is
JG^
x
JG^
Similarly,
the number of sets formed
by taking
n 2
from the first
group
and 2 from the second is
,Cn_^
x
t
C
t
.
And,
in
general,
the number of sets formed
by taking
n r from the first
group
and r from the second is
,Gn_
r
x
M
C
r
.
Hence we have
If x or
y
be less than n some of the terms on the
right
will
vanish;
for
u
C
r
=
if r > n.
249. Vandermonde's Theorem. From the last
Article,
if
x,
y
and n be
any positive integers
such that
x
+
y
is
greater
than
n,
we
have,
since
J0r
=
.
,
...+,
\n-r
'
[r
[n'
Multiply
each side of the last
equation by
[n,
and we
have
\ /
m mf i
HOMOGENEOUS PRODUCTS.
289
The above has been
proved
on the
supposition
that x
and
y
are
positive integers
such that x +
y
is
greater
than
n;
and
by
means of the theorem of Art.
91,
the
proposi-
tion can be
proved
to be true for all values of x and
y.
For the two
expressions
which are to be
proved
identical are
only
of the nth
degree
in x and
y.
But,
if
y
has
any particular integral
value
greater
than
n,
the
equation
is known to be true for
any positive integral
value of
x,
and thus for more than n
values;
and
hence it must be true for that value of
y
and
any
value
whatever of x. Hence the
proposition
is true for
any
particular
value whatever of
x,
and for more than n values
of
y
;
it must therefore be true for all values of x and for
all values of
y.
This
proves
Vandermonde's
theorem,
namely
:
If
n be
any positive integer,
and
x,
y
have
any
values
whatever;
then will
n(n
l)
n
..
|r
Jn
r
HOMOGENEOUS PRODUCTS.
250. The number of different
products
each of r
letters which can be made from n
letters,
when each
letter
may
be
repeated any
number of
times,
is denoted
by
the
symbol n
H
r
.
For
example,
the
homogeneous products
of two dimensions formed
by
the three letters
a, b,
c are a
2
,
fc
2
,
c
3
, be, ca,
ab.
Tofind n
H
r
.
Since in each of the r- dimensional
products
of n
things
there are r
letters,
the total number of letters in all
the
products
will be
n
H
r
x r
; and,
as each of the n letters
occurs the same number of
times,
it follows that the
S..A. 19
290
HOMOGENEOUS PRODUCTS.
number of times
any particular
letter,
a
suppose,
occurs is
H
H
r
xr-i-n.
Now consider all the terms which contain a at least
once. If
any
one of these terms be divided
by
a the
quotient
will be of r 1 dimensions
;
and,
when all
the terms
which contain a are so
divided,
we shall obtain
without
repetition
all the
possible homogeneous products
of the n letters of r 1 dimensions. Now the
homogeneous
products
of r 1 dimensions are in number
n
H
r
_
l ; and,
by
the
above,
the number of a's
they
contain is
r 1
-
x
n
H .
Hence,
taking
into account the a which is
n
a factor of each of the
n
H
r
_
l
terms,
the total number of a's
which occur in all the r- dimensional
products
of the n letters
is
.ffr..
+
x
A-I.
that fe
^~^
A-1-
Hence
equating
the two
expressions
for the number
of
a's,
we have
r
n
_n
+
r 1 __
~"
X
n
l-L
r
X xZ
i
'.
n n
n+r-
1
H
r
=
----
x
#,_,.
Since the above relation is true for all values of n and
r,
we have in succession
n+r
-'2
""-1
=
x
"-*'
H ^xU J"t
r-l
_ 9
r i-"r-3>
x // X
n-"!' ! o
Also
n
//
t
is
obviously equal
to n.
EXAMPLES.
291
Hence,
by multiplying
and
cancelling
common
factors,
we have
r?
(??.
+
!)
......
(n
+
r
1)
ro ,
n
H
r
=
- - -
[See
also Art.
293].
Ex. 1. Find the number of combinations three at a time of the letters
a, b, c,
d when the letters
may
be
repeated.
Ans. 20.
Ex. 2. Find the number of different combinations six at a time which
can be formed from 6
a's,
6
6's,
6 c's and 6 d's. Ans. 84.
Ex.3. Shew that
n
H
r
=
_,.&,.+#,._!,
and deduce that
Ex. 4. Shew that
251.
Many
theorems
relating
to
permutations
and
combinations are best
proved by
means of the binomial
theorem :
examples
will be found in
subsequent chapters.
[See
Art.
292.]
Ex. 1. Find the number of
ways
in which mn different
things
can be
divided
among
n
persons
so that each
may
have m of them.
The number of
ways
in which the first set of m
things
can be
given
is
mn
C
m
; and,
whatever set is
given
to the
first,
the second
set can be
given
in
mn
-
m
C
m ways ;
so
also,
whatever sets are
given
to the first and
second,
the third set can be
given
in
OT,,-2m
C7
OT ways
;
and so on.
Hence the
required
number is
\m(n-l) jm(?t-2)
12m m
X X X
|m
\m(n-2)
\m
\m (TI
-
3) \m
L \m
\m
Ex. 2. Prove that
l-
n
C
1
.
n
H
1
+
n
C,
i
.
n
H
2
-
n
C
3
.
n
H
3
+ +
(-l) B
C
n
.
n
H
n
=0.
Since
BCW/r
=
n
-
(
^"
(
"" r +
1)
\L \L
_n
2
(n
2
-!
2
) (ra
a
-f^l
2
)
p. 2* r
~'
192
292
EXAMPLES.
we have to
prove
that
TI*
n(n-l
a
)
n
ii
(n'-l)(n'-2')
1A
+
1
s
. 2
2
1
s
. 2
2
. 3*
^(n-l)...(n'-;m*)
I
2
.2
2
...n
2
n
2
-!
2
Now the first two terms
=
p
;
(n
s
-l)(n
3
-2)
'. ti*
=
+
v
jf^
(n-l)(n
2
-2
2
)(n--3
J
)
" f<rar
=
-
l.2.3*
~'
and so on.
Hence the sum of all the terms on the left
_,
m
(n-l)(n-2) (n*-n)
~
(
'
'
1
2
.2
2
n
2
Ex. 3. Shew that n
straight
lines,
no two of which are
parallel
and
no three of which meet in a
point,
divide a
plane
into n
(n
+
1)
+
1
parts.
The nth
straight
line is cut
by
each of the other n
-
1 lines
;
and
hence it is divided into n
portions.
Now there are two
parts
of the
plane
on the two sides of each of these
portions
of the nth line which
would become one
part
if the nth line were
away.
Hence the
plane
is
divided
by
n lines into n more
parts
than it is divided
by
n
-
1 lines.
Hence,
if
F(x)
be
put
for the number of
parts
into which the
plane
is divided
by
x
straight lines,
we have
F(n)=F(n-l)
+
n.
Similarly
F(n- l)
=
F(n-2)
+
(n-l),
and
But
obviously
F
(1)
=
2.
Hence
F(n)=2
+
2
+
3
+
4
+ ...... -j-n
Ex. 4.
Suppose
n
things
to be
given
in a certain order of succession.
Shew thai the number of ways of
taking
a set of three
things
out of
these,
with the condition that no set shall contain
any
two
things
which were
originally contiguous
to each other is
^ (n
-
2) (n
-
3) (n
-
4).
Shew also that if the n
given things
are
arranged cyclically,
so that
the nth is taken to be
contiguous
to the
first,
the number of sets is
reduced to
Jn (n
-
4) (n
-
5).
EXAMPLES.
Consider
the second case first.
Let the different
things
be
represented by
the letters
a, b, c,
......
k,
I.
Suppose
that a is taken first.
Then,
if either of the two letters
next but one to a be taken
second, any
one of n
-
5 letters can be
taken for the third of the set.
If, however,
the second letter ia not
ne^t but one to
o,
but in either of the n
-
5 other
possible places,
there would be a choice of n
-
6
places
for the third letter of the set.
Hence the total number of
ways
of
taking
3 letters in order a
being
first is
2(n-5)
+
(7i-o)(w-6),
that is
(n-4)(n-5).
There is the
same number when
any
one of the other letters is taken first
; hence,
as the order in which the three letters in a set are taken is
indifferent,
the total number of sets is
%n (n
-
4) (n
-
5).
In order to obtain the first case from the
second,
we have
only
to
suppose
that a and I are no
longer contiguous.
Hence the number
in the first case is the same as that in the second with the addition
of those sets which contain a and
I,
and there are n-4 of these.
Hence the number in the first case is
Ex. 5. There are n letters and n directed
envelopes
: in how
many
ways
could all the letters be
put
into the
wrong envelopes
?
Let the letters be denoted
by
the letters
a, b,
c... and the corre-
sponding envelopes by
o', b',
c',
......
Let F
(n)
be the
required
number of
ways.
Then a can be
put
into
any
one of the n- 1
envelopes
6', d,
......
Suppose
a is
put
into V
;
then k
may
be
put
into
a',
in which case
there will be F
(n
-
2) ways
of
putting
all the others
wrong.
Also if
a is
put
into
&*,
the number of
ways
of
disposing
of the letters so that
k is not
put
in
a',
b not in
6',
<fec. is
F(n
-
1).
Hence the number of
ways
of
satisfying
the conditions when a is
put
into fc' is
F(n-l)
+
F(n -2).
The same is true when a is
put
into
any
other of the
envelopes b', c',...
Hence we have
F(n)-nF(n-l)=-{F(n-l)-(n-l)F (n-2)}.
Similarly F(n-l)
-
(n- 1) F(n- 2)=
-
{F(n- 2) -(n-2) .F(n-3)}
F
(3)
-
3F
(2)
=
-
{
F
(2)
-
2F
(1) }.
But
obviously
F
(2)
=
1 and F
(1)
=
;
F
(n)-F(n-l)
=
(-!).
Hence
*W-*>I
(-l>.
1.
jn
in-1
|n
294- EXAMPLES.
Hence, by addition,
the number
required.
EXAMPLES XXIIL
1. In how
many
different
ways may twenty
different
things
be divided
among
five
persons
so that each
may
have four ?
2. A crew of an
eight-oar
has to be chosen out of eleven
men,
five of whom can row on the stroke side
only,
four on the
bow-side
only,
and the
remaining
two on either side. How
many
different selections can be made 1
3. There are three candidates for a certain office and twelve
electors. In how
many
different
ways
is it
possible
for them
all to vote
;
and in how
many
of these
ways
will the votes be
equally
divided between the candidates ?
4. Shew that
^(7^
:
tm
C
n
is
equal
to
1.3.5......
(4n-l)
.
5. Find the number of
significant
numbers which can be
formed
by using any
number of the
digits
0, 1, 2,
3, 4,
but
using
each not more than once in each number.
6. Shew that in the
permutations
of n
things
r
together,
the number of
permutations
in which
p particular things
occur
i*..,^-,*^,-
7. There are n
points
in a
plane,
no three of which are in
the same
straight
line
;
find the number of
straight
lines formed
by joining
them.
8. There are
points
in a
plane,
of which no three are on
a
straight
line
except
m which are all on the same
straight
line.
Find the number of
straight
lines formed
by joining
the
points.
EXAMPLES. 295
9. There are n
points
in a
plane,
of which no three are on
a
straight
line
except
m which are all on a
straight
line. Find
the number of
triangles
formed
by joining
the
points.
10. Shew that the number of different n-sided
polygons
formed
by
n
straight
lines in a
plane,
no three of which meet
in a
point,
is
j
n 1 .
11. There are n
points
in a
plane
which are
joined
in all
possible ways by
indefinite
straight
lines,
and no two of these
joining
lines are
parallel
and no three of them meet in a
point.
Find the number of
points
of
intersection,
exclusive of the n
given points.
12.
Through
each of the
angular points
of a
triangle
m
straight
lines are
drawn,
and no two of the 3m lines are
parallel;
also no
three,
one from each
angular point,
meet in a
point.
Find the number of
points
of intersection.
13. The streets of a
city
are
arranged
like the lines of a
chess-board. There are m streets
running
north and
south,
and n east and west. Find the number of
ways
in which a
man can travel from the N.W. corner to the S.E.
corner,
going
the shortest
possible
distance.
14. How
many triangles
are there whose
angular points
are at the
angular points
of a
given polygon
of n sides but none
of whose sides are sides of the
polygon
?
15. Shew that 2n
persons may
be seated at two round
\2n
tables,
n
persons being
seated at
each,
in
-=j=
different
ways.
%
16. A
parallelogram
is cut
by
two sets of m lines
parallel
to its sides : shew that the number of
parallelograms
thus
formed is
^ (m
+
l)*(m
+
2)
2
.
17. Find the number of
ways
in which
p positive signs
and n
negative signs may
be
placed
in a row so that no two
negative signs
shall be
together.
18. Shew that the number of
ways
of
putting
m
things
in
n+1
places,
there
being
no restriction as to the number in each
place,
is
(m
4-
n) l/m
I n
(
296 EXAMPLES.
19. Shew that 2n
things
can be divided into
groups
of n
|2n
pairs
in
J==-
ways.
2
Jn
20. Find the number of
ways
in which mn
things
can be
divided into m sets each of n
things.
21. Shew that n
planes through
the centre of a
sphere,
no
three of which
pass through
the same
diameter,
will divide the
surface of the
sphere
into n*
-
n + 2
parts.
22. Shew that the number of
parts
into which an infinite
plane
is divided
by
m + n
straight
lines,
m of which
pass
through
one
point
and the
remaining
n
through another,
is
mn + 2m + 2n
1, provided
no two of the lines be
parallel
or
coincident.
23. Find the number of
parts
into which a
sphere
is
divided
by
m + n
planes through
its
centre,
m of which
pass
through
one diameter and the
remaining
n
through
another,
no
plane passing through
both these diametei-s.
24. Find the number of
parts
into which a
sphere
is
divided
by
a + 6 + c+...
planes through
the
centre,
a of the
planes passing through
one
given
diameter,
b
through
a
second,
c
through
a
third,
and so on
;
and no
plane passing through
more than one of these
given
diameters.
25. Shew that n
planes,
no four of which meet in a
point,
divide infinite
space
into
(n
3
+ 5n +
6)
different
regions.
26. Prove that if each of m
points
in one
straight
line be
joined
to each of n
points
in
another, by straight
lines termin-
ated
by
the
points
; then,
excluding
the
given points,
the lines
will intersect
fynn (m 1) (n 1)
times.
27. No four of n
points lying
in a
plane
are on the same
circle.
Through every
three of the
points
a circle is
drawn,
and no three of the circles have a common
point
other than one
of the
original
n
points.
Shew that the circles intersect in
^n
(n
-
1) (n
-
2) (n
-
3) (n
-
4) (2n
-
1
) points
besides the
original
n
points, assuming
that
every
circle intersects
every
other circle in two
points.
CHAPTER XX.
THE BINOMIAL THEOREM.
252. WE have
already [Art. 67] proved
that the con-
tinued
product
of
any
number of
algebraical expressions
is
the sum of all the
partial products
which can be obtained
by multiplying any
term of the
first,
any
term of the
second,
any
term of the
third,
&c.
253. Binomial Theorem.
Suppose
that we have n
factors each of which is a + b.
If we take a letter from each of the factors of
(a
+
6) (a
+
6) (a
+
6)
and
multiply
them all
together,
we shall obtain a term of
the continued
product;
and if we do this in
every possible
way
we shall obtain all the terms of the continued
pro-
duct.
[Art. 67.]
Now \ve can take the letter a
every
time,
and this can
be done in
only
one
way;
hence a
w
is a term of the
product.
The letter b can be taken
once,
and a the
remaining
(n 1)
times,
and the number of
ways
in which one b can
be taken is the number of
ways
of
taking
1 out of n
things,
so that the number is
,.(7,
: hence we have
JO.,
a*'
1
b.
298 BINOMIAL THEOREM.
Again,
the letter b can be taken
twice,
and a the
remaining
(n 2)
times,
and the number of
ways
in which
two 6's can be taken is the number of
ways
of
taking
2 out
of n
things,
so that the number is
n
(7
2
: hence we have
And,
in
general,
b can be taken r times
(where
r is
any positive integer
not
greater
than
ri)
and a the re-
maining
n r
times,
and the number of
ways
in which r
6's can be taken is the number of
ways
of
taking
r out of
n
things,
so that the number is
n
C
r
: hence we have
Thus
(a
-f
6) (a
+
b) (a
+
6)
...... to n factors
=
a"
+
(?,
. aT
l
b
+
,
. a*-*b*
+
......
+
J0
r
. a*~
r
b
r
the last term
being n
C
n
a
n
~*b*,
i.e. 6".
Hence,
when n is
any positive integer,
we have
The above formula is called the Binomial Theorem.
If we substitute the known values
[see
Art.
244]
of
jClt n
C
t
,
n
C
t>
... in the series on the
right,
we obtain the
form in which the theorem is
usually given, namely
L 2
-
a
"
\r \n
r
The series on the
right
is called the
expansion
of
(a
+
6)-.
254. Proof
by
Induction. The Binomial Theorem
may
also be
proved by
induction,
as follows.
BINOMIAL THEOREM.
We have to
prove
that,
when n is
any positive integer,
(a
+
b)"
=
a"
+ na-'fc +
n
^~
9
^
a"-*b*
+......
...
+-P
S
a"~
r
b
r
+
... +
6",
|r
n r
or that
(a
+
6)"
=
a"
+
,,CX~
1
& +
.OX""
6
'
+ +
nW'V
+
-"
+
b
n
.
Now if we assume that the theorem is true when the
index is
n,
and
multiply by
another factor a
+
b,
we
have,
when like terms of the
product
are
collected,
(a
+
6)"
=
a""
+
(1+
JCJ
a
n
b
+
(n
O
l
+
B
(7
2
)
a"'
1
6"
+
. . .
...
+
LCr
_
1
+
n
C
r
)a"-^b'+...+b
n+l
.
Now l
+
Cl
+
n-
and,
in
general,
Hence
Thus
if
the theorem be true for
any
value of
n,
it will
be true for the next
greater
value.
Now the theorem is
obviously
true when n
=
1. Hence
it must be true when n
=
2
;
and
being
true when n
=
2,
it must be true when n
=
3
;
and so on
indefinitely.
The
theorem is
therefore true for all
positive integral
values
of n.
Ex.1.
Expand (a
+
6)
4
.
We have
4.3. .4.3
(a
300 GENERAL TERM.
Ex.2.
Expand (2x-y)
3
.
Put 2a; for
a,
and
-
y
for b in the
general
formula : then
y*.
Ex. 3.
Expand (a
-
6)
n
.
Change
the
sign
of 6 in the
general formula;
then we have
1 . A
255. General term.
By
the
preceding
articles we
see that
any
term of the
expansion
of
(a
+
6)" by
the
Binomial Theorem
will be found
by
giving
a suitable
value to r in
In
On this account the above is called the
general
term
of the series. It should be noticed that the term is the
(r
+
l)th
term from the
beginning. [See
Note Art.
244.]
256. Coefficients of terms
equidistant respec-
tively
from the
beginning
and the end are
equal.
In the
expansion
of
(a
+
6)
n
by
the Binomial
Theorem,
the
(r+l)th
term from the
beginning
and the
(r
+
l)th
term from the end are
respectively
Jd
r
. a"-
r
b
r
and
n
C
n
.
r
. a
r
b
n
~
r
.
But
n
U
r
=
n
G
n_
r
.
[Art. 245.J
GREATEST TERM. 301
Hence,
in the
expansion
of
(a
+
&)",
the
coefficients of
any
two terms
equidistant
respectively from
the
beginning
and the end are
equal.
This result
follows, however,
at once from the fact that
(a
4-
&)",
and therefore also its
expansion,
would be unaltered
by
an
interchange
of the letters a and 6
;
and hence the co-
efficient of a
n
~
r
b
r
must be
equal
to the coefficient of 6*~
r
a
r
.
257.
If,
in the formula of Art.
253,
we
put
a
=
1 and
6
=
a?,
we have
(n I} \
n
V- H
&
+
-
+
r~T=
~
**
+...+*".
1.2 \r \n r
This is the most
simple
form of the Binomial
Theorem,
and the one which is
generally employed.
The above form includes all
possible
cases :
if,
for
example,
we want to find
(a
+
&)" by
means of
it,
we have
.
L b
=
aNl
+
n-
(
a
o
1.2
...
l . &
258. Greatest term of a binomial
expansion.
In the
expansion
of
(1
+
#)",
the
(r
+
l)th
term is formed
*i
7* -4- 1
from the rth
by multiplying by
-
x.
r
AT
n-r
+
1
/n+1 _ \
,
n
+
1 , .
Now
-
x
= --
1
} x,
and
-
clearly
r
\
r
J
r
diminishes as r increases
;
hence
-
as diminishes
r
as r is increased. If
- -
x be less than 1 for
any
802 GREATEST
COEFFICIENT.
value of
r,
the
(r
+
l)th
term will be less than the rth.
In order therefore that, the rth term of the
expansion may
be the
greatest
we must have
n-r
+
l n
_
r _i
+
i
-
x <
1,
and
----
x > 1.
r r 1
(n
+
l)v (n
+
1)
a;
Hence r
>
--
,
and r <
-
+
1.
x+
I
x+1
The absolute values of the terms in the
expansion
of
(1
+
)*
will not be altered
by changing
the
sign
of x
;
and
hence the rth term of
(1 #)"
will also be
greatest
in
absolute
magnitude
if
r >
^ '
,
and r <
--
f-
+
1.
X+l
.,
(n
+
1)
x
,
n r
+
1
,
If r
=
*-
^ ,
then
-
x
=
1
;
and hence there
#4-1
r
is no one term which is the
greatest,
but the rth and
r+
l|th
terms are
equal,
and these are
greater
than
any
of
the other terms.
/ \
n
Since
(a+x)"
=
a
n
( 1
+
-
j
,
the rth term of
(a-f z)
n
is the
greatest
when
+ L
'
5+i
a
Ex. 1. Find the
greatest
term in the
expansion
of
(1
+
z)
10
,
when
21 21
The rth term is the
greatest,
if r> -=- and r< 1
+-=-. Hence
o o
the fifth term is the
greatest.
Ex. 2. Find the
greatest
term in the
expansion
of
(1
+
z)
10
,
when
5
* =
&'
The rth term is the
greatest,
if r > 5 and < 6. Thus there is no
one term which is the
greatest,
bat the 5th and Gth terms of the
EXAMPLES. 303
expansion
are
equal
to one another and
greater
than
any
of the
other terms.
Ex. 3. Find the
greatest
term in the
expansion
of
(10
+
3x)
15
when
x=4. Am. The ninth term.
The
greatest
coefficient of a binomial
expansion
can
be found in a similar manner. For in the
expansion
of
(1
+
#)"
the coefficient of the
(r +
l)th
term is formed from
n T
-\-
1
that of the ?*th
by multiplying by
+ . Hence the
rth coefficient will be the
greatest
in absolute
magnitude,
. n
r+1
,
nrl
+1
if
-
< 1 and
;
> 1.
r rl
mi
.
. -
f
ri
+
1 ,
n+I
lhat is if r >
9
and < 1
H
^
.
Hence when n is
even,
the coefficient of the rth term
/n
is the
greatest
when ?-
=
-+!;
and when n is
odd,
the
JU
coefficients of the
^
th and
-
th terms are
equal
2 '2
to one another and are
greater
than
any
of the other terms.
For
example,
in
(1
+
a;)
20
the coefficient of the llth term is the
greatest;
and in
(1
+
a:)
11
the coefficients of the 6th and 7th terms
are
greater
than
any
of the others.
EXAMPLES XXIV.
Write out the
following expansions
:
1.
(z
+
a)
5
. 2.
(2a-x)
b
. 3.
(1-ar)
6
.
4.
(2a-3a
s
)
4
. 5.
(2a;
2
-3)
4
. 6.
(x
3
-
2y
3
)
5
.
7. Find the third term of
(x
-
By)
i0
.
8. Find the fifth term of
(3x
-
4)
20
.
9. Find the
twenty
first term of
(2 x)~
2
.
10 Find the fortieth term of
(x
-
y)
42
.
304
EXAMPLES.
11. Find the middle term of
(1
+
a;)
8
.
12. Find the middle terms of
(1
+
a;)".
13. Find the
general
term of
(x 3y)
m
.
14. Find the
general
term of
(x*
+
y
8
)".
15. Write down the first three terms and the last three
terms of
(3a; 2y)
15
.
16. Find the term of
(1+a;)"
which has the
greatest
coefficient.
17. Find the two terms of
(1
+
a;)'*
which have the
greatest
coefficients.
18. Shew that the coefficient of x" in the
expansion
of
(1
-I-
)*"
is double the coefficient of x" in the
expansion
of
19. Shew that the middle term of
(1
+x)
s*
is
l . o .
o
...
(4n
-
i)
2"x*^
20.
Employ
the binomial theorem to find 99
4
,
51* and
999
3
.
21. Shew that the coefficient of af in the
expansion
of
|n
-
I IS
rr-;
22. Find the middle term
/ l\
f
"
of (x
- -
J
.
23. The coefficients of the
5th,
6th and 7th terms of the
expansion
of
(1 +x)*
are in arithmetical
progression:
find n.
24. For what value of n are the coefficients of the
second,
third and fourth terms of the
expansion
of
(1 +)"
in arith-
metical
progression
1
25. If a be the sum of the odd terms and b the sura of the
even terms of the
expansion
of
(1
+
x)*,
shew that
(1
-*)'=
or -b'.
PROPERTIES OF THE COEFFICIENTS. 305
259.
Properties
of the coefficients of a
binomial
expansion.
It will he convenient to write the Binomial Theorem
in the form
where,
as we have
seen,
c
=
c
n
=
1
;
c
l
=
c
n
.
l
=
n
;
and,
in
general,
c
r
=
c,,_ r
=
**
.
I. Put x
=
1 in
(i)
;
then
Thus the sum
of
the
coefficients
in the
expansion of
(1
+
x)*
is 2".
II. Put x
=
-
1 in
(i)
;
then
Thus the sum
of
the
coefficients of
the odd terms
of
a
binomial
expansion
is
equal
to the sum
of
the
coefficients
of
the even terms.
III. Since c
r
=
c
n
_
r
,
we have
(1
+
x)'
:
=
c
+
CjX
+ c
2
x*
+
. . .
+
c
r
x
r
+
. . . +
c,X
l
,
and
(1
+
x)
n
=
c
n
+ c
n^x
+ c
n.^
+
. . . c
n
.
r
x
r
+
. . .
+ c a.-
n
.
The coefficient of x
n
in the
product
of the two series
on the
right
is
equal
to
C
+
C
l
+
C
2
2
+
...... + C
2
'
Hence
[Art. 91]
c
*
+
c
t
2
4-
. . .
+ c
r
2
+
......
+
c,,
2
is
equal
to the coefficient of x
n
in
(1
+
x)
n
x
(1
+
x)",
that
|2
is in
(1 +x)~"\
and this coefficient is
'
\n \n
s. A.
20
306 PROPERTIES OF THE
COEFFICIENTS.
Hence the sum
of
the
squares
of the coefficients in the
|2n
expansion
of
(1 +#)"
is
f^f-.
I
71
1'
1
IV. As in
III,
we have
(1
+
x)*
=
c + c
t
x
+
cj?
+ ...
4- c
n
x
n
,
and
(1 #)
n
=
c
n
c
n .,a;
+
c
n_#r +...+( I)"c
a;".
The coefficient of a;" in the
product
of the two series
on the
right
is
equal
to
The
coefficient of x" in
(1 +ar)
n
x
(1 a?)",
that is in
In
(1 flf)*.
is zero if n be
odd,
and is
equal
to
( I)
2
.
rr~ ir~
if n be even.
Hence c
2
c,
2
+
c
2
s
...+( l)"cB
s
is zero
or
( l)
2
n<!/(^?i!)
2
,
according
as n is odd or even.
Ex. 1. Shew that
Cj
+
2c
a
+ 3c
s
+ ... + rc
r
+ ... +
uc
n
=
7i2
n
-
J
.
We have c
1
+
2c
2
+3c
s
+... +
nc
B
Ex.2. Shew that
c,-sC1 +;C1
-
-K-l)"-^r7
=
^-i
2 o n
+
i n
+
l
11 1
In(n-l)
We have c
t
-~c
l
+
~c
a
-&c.=l--^n
+
^ j-jj
PROPERTIES OF THE COEFFICIENTS.
307
(n+l)n (n
+
l)7t(n-l)
H
+
l-
V
+
1.2 1.2.3
-
+
(-!)"}
n+l n
+
l 1.2 1.2.3
Ex. 3. Shew
that,
if n be
any positive integer,
c
c,
c
2
L_?__
I
(1)
x x
+
l
x+2
'
v
~
x+n
x(x
+
l)...(x+n)'
n
c
n
[2.
for oZi values of
x,
and for
any particular
value of n.
Change
x into x
+
1
;
then
=
fc
Hence, by
subtraction,
/7_ J- 1 /7_ -i_ /7_ /7 j_ /7
-...:.. +
(-l)
r
^
j.
f
_
iy+i
_
_zr=_
ar+n+1
x(x
+
l)
......
(x
+
n
+
l)'
But
B
C
r
+
n
C
r
_
1
=
B+1
(7
r
,
for all values of r
[Art. 247J.
Hence we have
_
,
(_
l\n+l
n+l
_ _ _ _
x
+
2
z+n+1 x
(x
+
I)
...
(x+n+l)'
Hence
t/
the theorem be true for
any particular
value of n it will
be true for the next
greater
value. But the theorem is
obviously
true
for all values of x when n=l: it is therefore true for all
positive
integral
values of n.
[See
also Art.
297,
Ex.
3.]
By giving particular
values to x we obtain relations between
c
,
Cj
,
<fec. For
example
:
Put x
=
1
;
then we have
202
308 PROPERTIES OF THE COEFFICIENTS.
I
=
2
;
then
T~
+
li~
=
1.3.5...(2+1)*
Ex. 4. Shew that
c
o-C! (a-l)
+ c
3 (a-2)-c,(a-3)
+ +
(
-
l)"cn (a-n)
=
0,
and that
c a
s
Cj (a l)
2
+ c
2 (a
-
2)
2
c
s
(a- 3)
3
+ +
( !)"< (a-n)*=0.
We have from
II.,
if n be
any positive integer,
!_
"(-!) (n-l)(n-2)
<-ii-o
Hence,
if n > 1
l-(n-l)+
- - -
+
(-l)
n
~
1
=0
.....(ii).
Multiply (i) by
a and
(ii) by
n and add
;
then
a-n(a-l)H (a -2)-
+
(- l)
n
(o-n)
=
(iii),
where n is > 1.
Change
a into a
-
1 and n into n
-
1 in
(iii)
; then,
n
being
>
2,
we
have
a
-
1
-
(n- 1) (o
-
2)
+ +
(
-
!)-> (o-n)
=0
(iv).
Now
multiply (iii) by
a and
(iv) by
n and add
;
then
-
?^-(-2)
s
-
+
(-l)(ci-)=0.
provided
n is
greater
tluin 2.
By proceeding
in this
way
we
may prove
that
provided
that
p
is
any positive integer
less than n.
[See
also Art.
305.J
260. Continued
product
of n binomial factors of
the form x
+
a,
a;
+
b,
x
+
c,
&c.
It will be convenient to use the
following
notation :
$j
is written for a+b
+
c+...,
the sum of all the
letters taken one at a time. S
t
is written for ab
+
ac
+
...,
the sum of all the
products
which can be obtained
by
PRODUCT OF BINOMIAL FACTORS.
309
taking
the letters two at a time.
And,
in
general,
S
r
is
written for the sum of all the
products
which can be
obtained
by taking
the letters r at a time.
Now,
if we take a letter from each of the binomial
factors of
(x
+
a) (x
+
b) (x
+
c) (x
+
d)
......
,
and
multiply
them all
together,
we shall obtain a term
of the continued
product
; and,
if we do this in
every
possible way,
we shall obtain all the terms of the con-
tinued
product.
We can take x
every
time,
and this can be done in
only
one
way;
hence x
n
is a term of the continued
product.
We can take
any
one of the letters
a, b, c...,
and x
from all the
remaining
n 1 binomial factors
;
we thus
have the terms cuc
n
~\
bo;"'
1
,
ac"'
1
, &c:,
and on the whole
8
t
.i-*.
Again,
we can take
any
two of the letters
a, 6,
c...,
and x from all the
remaining
n 2 binomial factors
;
we thus have the terms abx
n
~*,
acx
n
~*, &c.,
and on the
whole S
t
. a?~\
And,
in
general,
we can take
any
r of the letters
a, b, c...,
and x from all the
remaining
n r binomial
factors
;
and \ve thus have S
r
. x
n
'
r
.
Hence
(x
+
a) (x
+
b) (x
+
c)
......
the last term
being
abed......
,
the
product
of all the
letters
a, 6, c, d,
&c.
By changing
the
signs
of
a, 6, c, &c.,
the
signs
of
S
l}
S
3
,
S
6
,
&c. will be
changed,
but the
signs
of $
2
,
S
4
,
$,,
&c. will be unaltered.
Hence we have
(x
a) (x b) (x c)
......
=
x
n
-
8.af
310 VANDERMONDE'S THEOREM.
261. Vandermonde's Theorem. The
following
proof
of Vandermonde's Theorem is due to Professor
Cayley*. [See
also Art.
249.]
We have to
prove
that if n be
any positive
integer,
and a and 6 have
any
values whatever
;
then will
In
+
7-7^
.-A+
+!>
jr
J
r
Assume the theorem to be true for
any particular
value
ofn.
Multiply
the left side
by
a
+b
n;
it will then become
(a
+
b)^.
1
.
Multiply
the successive terms of the series on
the
right
also
by
a
+
b n but
arranged
as follows :
for the first term
{(a n)
+
b}
;
for the second
{(a
n
+
1)
+
(6
-
1)}
;
and for the rth
{(a
-
n
+
r
-
1)
+
(6
-
r
+
1)}.
We shall then have
(a
+
i)B+1
=
a.
{(a
-
n)
+
6}
+
.C,
.
0^6, {(a
-
n
+
1)
+
(b
-
1)}
-i
{(a
-
+
r
-
1)
+
(b
-
r
+
1)}
+ ...+&
{a
+
(6 -n)}.
Now
a,,{(a n)
+
b] =a^
+a
f,6
1
,
A b-A {(a
-
n
+
1)
+
(6
-
1)}
=
n
C,
(an
b,
+
a^
r
. a
n
.
r
b
r
{(a
-
n
+
r)
+
(6
-
r)|
=
n
C
r
(a..
*
Messenger of
Mathematics,
Vol. v.
r
MULTINOMIAL THEOREM. 311
Hence
(a
+
6)n+1
=
a
n+1
+
(
1
+
(?,)
a
n
6,
+
. . .
...
+
C.CU
+
,Cr)
a
n+l
.
r
b
r
+
. . .
+
b^
=
a+i
+
,,+iCX&l
+
+^C
r
a
n+l
_
r
b
r
+
...
+6
n+1
,
since
n
C
r_,
+
n
C
r
=
n+1
<7
r
.
Thus,
t/
the theorem be true for
any particular
value
of
n,
it will also be true for the next
greater
value. But
it is
obviously
true when n
=
1
;
it must therefore be true
when n
=
2
;
and so on
indefinitely.
Thus the theorem is
true for all
positive integral
values of n.
262. The Multinomial Theorem. The
expansion
of
the nth
power
of the multinomial
expression
a
+
b
+
c 4- ...
can be found
by
means of the Binomial Theorem.
For the
general
term in the
expansion
of
(a
+
b
+
c
+
d
+
...)",
that is of
{a +(b
+ c + d+
...)}",
by
the Binomial Theorem is
a
r
(b
+ c +
d
+
...)""
r
-
Similarly
the
general
term in the
expansion
of
by
the Binomial Theorem is
\n
s n r s
b'(c
+
d
The
general
term in the
expansion
of
(c
+
d
+
...)"
' *
by
the Binomial Theorem is
i s
'
(d
+
. .
.)"-'-'.
t
\n
r s t
Hence the
general
term in the
expansion
of
is
\n
r
\n
i s
\r\rir
\8\n-r
s \t\n-r st"
312 MULTINOMIAL THEOREM.
In
that is
where each of
r, s,
t ... is zero or a
positive integer,
?md
... =n.
The above result can however be at once obtained
by
the method of Art.
253,
as follows.
We know
[Art. 67]
that the continued
product
is the sum of all the different
partial products
which can
be obtained
by multiplying any
term from the first multi-
nomial
factor,
any
term from the
second,
any
term from
the
third,
&c.
The term a'b'c*... will therefore be obtained
by taking
a from
any
r of the n
factors,
which can be done in
n
G
r
different
ways;
then
taking
b from
any
s of the
remaining
n r
factors,
which can be done in
n_^Ct
different
ways
;
then
taking
c from
any
t of the
remaining
n r s
factors,
which can be done in
n
_
t_.C'f
different
ways;
and so on.
Hence the total number of
ways
in which the term
a
r
b' c'... will be
obtained,
which is the coefficient of the
term in the
required expansion,
must be
that is
|n
In r In r s
X
r-r=
-=
X
r-
1 -=
x ...
=
[r
In r
|s|n-r
s
[< "nrst
Irlslt..."
Hence the
general
term in the
expansion
of
(a
+
b
+
c
+
...)"
is
EXAMPLES.
313
Ex. 1. Find the coefficient of ale in the
expansion
of
(a
+
b
+
c)
3
.
13
The
required
coefficient
= =
6.
Ex. 2. Find the
coefficients of
a?b-,
bed
2
and abed in the
expansion
of
(a
+
b
+ c +
d)*.
We have the terms
|4
|4
|
a b
'
'-'-'
lcd
*
&nd
l^lljllj.
alcd-
Th'is the
required
coefficients are
6,
12 and 24
respectively.
263.
By
the
previous
Article,
the
general
term of the
expansion
of
(a
+
bx
+ ca? +
do?
+
)
B
is
141
a
r
(bx)' (GO?)' (dx?)
u
,
Cffife-
that i s
, -,
aWeZ"
rs
Hence to find the coefficient of
any particular power
of
#,
say
of of
1
,
in the
expansion,
we must find all the different
sets of
positive integral
values of
r, s, t,...
which
satisfy
the
equations
s
+
2t
+
3u +
......
=
a,
=n.
The
required
coefficient will then be the sum of the
coefficients
corresponding
to each set of values.
Ex. 1. Find the coefficient of x
6
in the
expansion
of
(1
+
2x
+
3x~)
4
.
ii
The
general
term is
-j
:
j-
2*3
t
x
l+yt
,
and the terms
required
are
those for which s
+
2t
=
5
anoTr+* +
t=4.
^ince each of the
quantities r,
s and t must be zero or a
positive
integer,
the
only possible
sets of values are t
=
2, 8=1,
r=l and
14
t
=
l,
s
=
3, -=0,
the
corresponding
coefficients
being
;
~
. 2
. 3
2
li 11 l
314 EXAMPLES.
14
and
TTT-^T
2
3
.
3,
that is 216 and 96
respectively.
Hence the
(0 |d
[1
required
coefficient is 312.
In
simple
cases the resnlt can be
readily
obtained
by
actual
expansion.
We have
(1
-I-
2*
+ 3x
2
)=
1 +
4
(2x
+ Sz
2
)
+ 6
(2x
+
3X
2
)
2
+
4
(2x
+3x
2
)
s
+
(2x
+
3x
2
)
4
.
Only
the last two terms will contain x
s
and the coefficients of x
5
in
these terms will be found to be 216 and 96
respectively,
so that the
required
coefficient is 312.
Ex. 2. Find the coefficient of x* in the
expansion
of
(1
+x + x
1
)
3
.
Arts. 6.
Ex. 3. Find the coefficient of X
s
in the
expansion
of
(l
+
x
+
x
2
)
4
.
Ant. 16.
Ex. 4. Find the coefficient of x
8
in the
expansion
of
(2
+
x
-
x
8
)
5
.
Ans.
Ex. 5. Find the coefficient of x
10
in the
expansion
of
(7
+
x
+x+x
s
+a
4
+ar
i
)
s
. Ant. 3S-
Ex. 6. Find the coefficient of the middle term of the
expansion
of
. An. 381.
EXAMPLES XXV.
1. Prove that
c -2
Cl
+ 3C
,-
......
+(-l)'(n+
2. Prove that
c
1
-2c
t
+ 3e
s
-
...... +
(- l)"-'
3. Prove that
4. Prove that
c,
+ 2c
3
+
3c
<
+ ......
+(-l)ct
=l
+(n-2)2
'.
5. Prove that
c + 3c
|
+5c
j
+ ......
+
(2n
+
l)c,
=
(7+ 1)2".
6. Prove that
3c,
+
7c,+ llc,+
......
4-
(4n
-
l)c.=
1 +
(2- 1)2".
I
EXAMPLES.
315
7. Prove that
c, c,
c,
c 2"
+l
-l
_"
I _l I
_S
I I
_ _
_____^
123 n+l n+l
8. Prove that
c c. c c
e
2"
U
i 8 t _ 4 i
8
i
~T
T
17
T - T
-f ......
~
T - ...... :
135/ W+ 1
9. Prove that
c
.
c
a
c
s
2"- 1
_I
i
__
3 i
__
*4.
_
246
~n+l*
10. Prove that
c
-
11
234 n+2
11. Prove that
l -*
4.
?
' -
c
-
! ! !
1 23
12. Prove that
,
c,
c_ ,
.....c
+ + { !>*= 4- 1- + .
T^O
T^ m
./ -|
T^ ,, T^ ...... l^
v/
ril2 w
c
f-4
+
7-
+(
:
1
13. Prove that
+ .. .. +C C
=;
,-r
14. Prove
that,
if
then n
"-1 e *~ l
............ +nc
i
x*~,
and
{1
+
(n+ l)a;}(l +x)
a~ t
=
c + 2c
t
x 4- ...... +
(w+ 1)0
Hence
prove
that,
\2n-l
2c
s
+ 3c-
4-
,
|n
1
\n-
1
-
I
316 EXAMPLES.
15.
Shew, by expanding {(1
+
x)"
1}"*,
where m and n
are
positive integers,
that
16. Prove
that,
if >
3,
(i) a-n(a-l)
+
?^f>(a-2)-
...... +
(- 1)" (-n)-
0.
(ii)
5- W
(a-l)(6-l)
+
r
^(a
(iii) a6c-n(a
...... +
(- I)'
(a -n) (b-n) (c -n)
=
0.
17. Shew
that,
if there be a middle term in a binomial
expansion,
its coefficient will be even.
18. Shew that the coefficient of x" in the nth
power
of
x* +
(a
+
b)
x + ab is
a" +
.0,
V
'
I +
.Ct
f
a-
f
6" + ...... +b\
19. If n be a
positive integer
and P
u
denote the
product
of
all the coefficients in the
expansion
of
(1
+
x)*,
shew that
P
u
[n
20. Shew that
21. Shew
that,
if n be a
positive integer,
l+x
w(n-l)
- --
1+Tia; 1.2
(L+nx)
_
n
(nj:_
1.2.3
22. Shew that
(a
+ b + c + d +
)
5
=
2a
s
+ 52a
4
6 + lOSa'ft' + 202a'6c
P1XAMPLES. 317
23. If
(1
+ x+
or)"
=
+
a,cc
+ a
a
x
s
+
,
prove
that na . + ^. a
,
+
,
:
==
(t
n
=
0,
|ji_
[? \n-r
unless r is a
multiple
of 3.
21. Shew
that,
in the
expansion
of
(1
+ x + x*+ +
x')
H
,
where n is a
positive integer,
the coefficients of terras
equi-
distant from the
beginning
and the end are
equal.
25. If
,
p
a
a
,
be the coefficients in the
expansion
of
(!+;
+
x*)"
in
ascending powers
of
or,
prove
that
c&
*
ct* + a
f
+ <z
gB a^,
and that
26. If
(
1 4- x +
x*)"
=
a +
ajc
+
a^x*
+ a
a
x
3
+
...,
prove
that
27. Shew
that,
in the
expansion
of
(a,
+ a
t
+ a
s
+ ... + a
r
)
a
,
where n is a whole number less than
r,
the coefficient of
any
term in which none of the
quantities
a,,
a
g
,
&c.
appears
more
than once is n\
28. Shew
that,
if the
quantities (1 +x), (l
+ x + x
2
),
.........
,
(1
+ x + x' + ...... +
x')
be
multiplied together,
the coefficients
of terms
equidistant
from the
beginning
and end will be
equal ;
and that the sum of all the odd coefficients will be
equal
to the
sum of all the
even,
each
being (n
+
1)
!
29. Shew that the coefficient of x" in the
expansion
of
(1
+ x H-ce
2
)"
is
n(n-\) n(n l)(n 2)(ii-o)
, 2
2
129202
......
30. Shew that 18 can be made
up
of 8 odd numbers in
792 different
ways,
where
repetitions
are allowed and the order
of addition is taken into account.
CHAPTER XXL
CONVERGENCY AND DIVERGENCY OF SERIES.
264. A series is a succession of
quantities
which are
formed in order
according
to some definite law. When a
series terminates after a certain number of terms it is said
to be a
finite
series,
and when there is an endless succession
of terms the series is said to be
infinite.
We have
already
found that when the common ratio of
a
geometrical progression
is
numerically
less than
unity
the sum of n terms will not increase
indefinitely,
but that
the sum will become more and more
nearly equal
to a
fixed finite
quantity
as n is increased without limit. Thus
the sum of an infinite series is not in all cases
infinitely
great.
When the sum of the first n terms of a series tends to
a finite limit
8,
so that the sum
can,
by sufficiently
increasing
n,
be made to differ from S
by
less than
any
assignable quantity,
however
small,
the series is said to be
convergent,
and S is called its sum. Thus !
+ + + +
...
is a
convergent
series whose sum is 2.
When the sum of the first n terms of a series increases
numerically
without limit as n is increased
indefinitely,
the
series is said to be
divergent.
Thus 1
+
2
+ 3
+ 4+...
is a
divergent
series.
CONVERGENOY
AND DIVERGENCY. 319
When the sum of the n first terms of a series does not
increase
indefinitely
as n is increased without
limit,
and
yet
does not
approach
to
any
determinate
limit,
the series is
neither
convergent
nor
divergent.
Such a series is some-
times called an indeterminate
or a neutral
series,
or the
series is said to oscillate.
For
example,
the series 1 1
+
1 1
+ ... is an oscilla-
tory
series,
for the sum of n terms is 1 or
according
as
n is odd or even.
It is clear that a series whose terms are all of the same
sign
cannot be
indeterminate,
but must either be conver-
gent
or
divergent.
For unless the sum of n terms increases
without limit as n is increased without
limit,
there must
be some finite limit which the sum can never
exceed,
but
to which it
approaches indefinitely
near.
265. If each term of a series be
finite,
and all the terms
have the same
sign,
the series must be
divergent.
For,
if
each term be not less than
a,
the sum of n terms will be
not less than
net,
and net can be made
greater
than
any
finite
quantity,
however
large, by sufficiently increasing
n.
266. The successive terms of a series will be denoted
by MJ,
u
t
,
w
s
,... ; and,
since it is
impossible
to write down all
-the terms of an infinite
series,
it is
necessary
to know how
to
express-
the
general
term,
u
nt
in terms of n.
The sum of the n first terms will be denoted
by
U
n ;
and the sum of the whole
series,
supposed convergent,
in
which case alone it has a
sum,
will be denoted
by
U.
Thus U=
M,
+
M
2
+
u
a
+
...
+
u
n
+
w
n+1
+
...,
and U
n
=
u
+
u
+ u+
...
+u.
267. In order that the series u
lt
u
t
,
w
8
,
u
4
,
......
,u
n
,
u
*\>
&c- mav be
convergent
it is
by
definition
necessary
and sufficient that the sum
320
CONVERGENCY AND DIVERGENCY.
should
converge indefinitely
to some finite limit U as n is
indefinitely
increased.
Hence U
Ht
U
n+i
,
U
H4Jt
,
&c. ... must differ from
U,
and
therefore from one
another,
by quantities
which diminish
indefinitely
as n is increased without limit.
Hence,
in order that a series
may
be
convergent,
the
(n
+
l)th
term must decrease
indefinitely
as n is increased
indefinitely,
and also the sum of
any
number
of
terms
sginning
at the
(n
+
l)th
must become less than
any
assignable quantity,
however
small,
when n is
indefinitely
increased.
ill i
For
example,
the series
T
+
S
+
O
+ -"'' 1"
cannot be con-
1 2 o n
vergent, although
the nth term diminishes
indefinitely
as n is increased
indefinitely ;
for the sum of n terms
beginning
at the
(n
+
l)th
is
f
-\ + ... +
jr-
,
which is
greater
than x
n,
that
is, greater
n
+
1 N -r
- 'An -n
than^.
268. We shall for the
present
consider series in which
all the terms have the same
sign
;
and as it is clear that
the
convergency
or
divergency
of such a series does not
depend
on whether the
signs
are all
positive
or all
negative,
we shall consider all the
signs
to be
positive.
The
convergency
or
divergency
of series can
generally
be determined
by
means of the
following
theorems.
269. Theorem I. A series is
convergent if
all its
terms are less than the
corresponding
terms
of
a second
series which is known to be
convergent.
CONVERGENCY AND DIVERGENCY. 321
Let the two series be
respectively
and V=v
l
+v
i
+v
3
+
......
Then,
since u
r
< v
r
for all values of
r,
it follows that U
is less than V.
Hence,
as F is
finite,
U must also be
finite : this
proves
the
theorem,
for a series must be
convergent
when its sum is finite and all the terms have
the same
sign.
It can be
proved
in a similar manner that a series
is
divergent
if all its terms are
greater
than the
corresponding
terms of a
divergent
series.
Ex.
(i).
To shew that the series
T
+
^
-
+ + . + ... is
J. JL J..A.O J..43.O.4
convergent.
The terms of the series are less than the terms of the series
111 1
l
+
l~2
+
l 2 2
+
l 2 2 2
+
""
SeneS 1S a
ge
"
metrical
progression
whose common ratio is
-
,
which is therefore
known to be a
convergent
series. The
given
series must therefore
also be
convergent.
Ex.
(ii).
Shew that the series
(a
+
x)
(b
+
x) (b
+
x)(2b
+
x) (b
+
x)(2b
+
x)(3b+x)
......
is
convergent
if
a,
b and x are all
positive,
and a < 6.
The terms of the
given
series are less than the
corresponding
a
+
x
(a
+
xf (a
+
x)
3
terms of the series
r
--
H
71
-
fa
+
h
-
: +
....
b
+
x
(b+x)* (b+x)
3
[since
-r
-
<
T
-
if r >
1, a,
6 and x
being positive
and 6 >
a].
The latter series is
convergent,
and therefore also the
given
series.
To ensure the
convergency
of the first series it is not
necessary
that all its terms should be less than the
corresponding
terms of the second
series,
it will be
sufficient if all the terms
except
a
finite
number of them
S.A.
21
322
CONVERGENCY AND DIVERGENCY.
be less than the
corresponding
terms of the
second,
for the
sum of a finite number of terms of
any
series must be
finite.
4 4
2
4* 4* 4
s
4
6
Ex. Shew that the series 1
+
^-+,-
+ + + -+ + ... is Con-
lz IS. c
'
& c (L
vergent.
From the sixth term
onwards,
each term is less than the corre-
4
s
4
8
spending
term of the series
=-r= +
-,
+ .... Hence the series
5
[
5*
beginning
at the sixth term is
convergent,
and therefore the whole
series is
convergent.
270. Theorem Hi
If
the ratio
of
the
corresponding
terms
of
two series be
always finite,
the series will both be
convergent
or both
divergent.
Let the series be
respectively
and
V
=
v
l
+ v
a
+
v
a
+
......
Then,
since the
quantities
are all
positive,
^.
must lie
f
between the
greatest
and least of the fractions
[Art.
1
13].
Hence U : V is finite. It therefore follows that if U is
finite so also is
V,
and if U is infinite so also is V.
8 16 8n
For
example,
the two series
_
+
^
+ ...... + + ...
and
r +
g
+ ...... H
---
!- &r^ both
convergent
or both
divergent.
X
w
For the ratio of the rth
terms,
namely
--
^-.
-
jrr-
:
-
is
equal
to
o^2
- --
,
which is > 1 and <
8 for all values of r. Now we have
already proved
that the second series is
divergent:
the first series is
therefore also
divergent.
271. Theorem III. A series is
convergent if, after
any particular
term,
the ratio
of
each term to the
preceding
is
always
less than some
fixed quantity
which is
itself
less
than
unity.
CONVERGENCY AND DIVERGENCY. 323
Let the ratio of each term after the r
th
to the
preceding
term be less than
k,
where k < 1.
Then,
since
^
<
k,
*^
2
<
/-, ,
u>r
u
rl
we have
nl
<
r
'
,
since k is less than 1.
J.
~~
K
Hence the sum of the series
beginning
at the r"
1
term
is
finite,
and the sum of
any
finite number of terms is
finite
;
therefore the whole series must be
convergent.
272. Theorem IV. A series is
divergent if, after
any particular
term,
the ratio
of
each term to the
preceding
is either
equal
to
unity
or
greater
than
unity.
First,
let all the terms after the r"
1
be
equal
to u
r \
then u
r+l
+
u
r
^
+
. . .
+
u
n+r
=
nu
r
,
and nu
r
can be made
greater
than
any
finite
quantity by sufficiently increasing
n. The series must therefore be
divergent.
Next,
let the ratio of each
term,
after the r
th
,
to the
preceding
term be
greater
than 1.
Then u
r+1
> u
r ,
u
r+a
> u
r+l
> u
r ,
&c.
Hence u
rl
+
u
r
^
+
...
+
u
n+r
> nu
r ;
the series must
therefore be
divergent.
1 2 2
2
2
3
2*-
1
Ex. 1. In the series
T
+
7>
+
"o+T
+ H H
foe ratio
U
A.
271
=
, ,
which is
greater
than 1
;
the series is therefore
n
+
1
divergent.
Ex. 2. In the series l
a
+
2
2
x
+
3
2
x
2
+
,
the test ratio is
^~^-x,
(IN
2
lH I x.
Now,
if x be less than
1,
and
any
fixed
quantity
k
be chosen between x and
1,
the test ratio will be less than k for all
terms after the first which makes
212
324 CONVERGENCY AND DIVERGENCY.
Hence the series is
convergent
if x < 1.
If x
=
1 the series is I
2
+
2
s
+
3
2
+ which is
obviously divergent,
and if x > 1 the series is
greater
than l
>
+
2
2
+
3
a
+
Thus the series I
j
+
2
2
x
+
3
8
x
2
+ is
divergent except
when x
is less than
unity.
273. When a series is such that after a finite number
at
of terms the ratio
-***
is
always
less than
unity
but
becomes
indefinitely nearly equal
to
unity
as n is in-
definitely
increased,
the test contained in Theorem III.
fails to
give any
result
;
and in this
case,
which is a
very
common
one,
it is often difficult to determine whether
a series is
convergent
or
divergent
For
example,
in the series
1111
the ratio
-5*1
=
r
w
Hence,
if k be
positive,
the test ratio is less than
unity,
but becomes more and more
nearly equal
to
unity
as n is increased.
We cannot therefore determine from Theorem III.
whether the series in
question
is
convergent
or
divergent.
274. To shew that the series
^
+
^
+
^
+
. . . is con-
1 .Z o
vergent
when k is
greater
than
unity,
and is
divergent
when
k is
equal
to
unity
or less than
unity.
First,
let k be
greater
than
unity.
Since each term of the series is less than the
pre-
ceding
term,
we have the
following
relations :
111
2*
*
3*
<
2*
'
CONVERGENCY AND DIVERGENCY. 325
! _L 1 JL _t
4*
+
5*
+
6* 7*
<
4*
'
and
T^rr
(2"
+
1)*
'
T
(2"
+1
-
1)*
2"*
Hence the whole series is less than
1 1 i. 1
2
"
1*
+
2*
+
4* S*
+ ...... +
2"
that
is,
less than
But this latter series is a
geometrical progression
whose common
ratio,
^-j
,
is less than
unity,
since k > 1.
Hence the
given
series is
convergent.
Next,
let k
=
1
;
then we can
group
the series as
follows :
X
therefore,
as each
group
of terms in brackets is
greater
than
^,
the
given
series taken to 2* terms is
greater
than
1+^
+
^
+
-^
+
...... taken to n
+
1
terms,
that
is,
greater
than 1
+
\n,
which increases
indefinitely
with n.
Hence
T
+
O
+
Q
+
...... is
divergent.
L Z o
Lastly,
let k be less than
unity ;
then each term of the
series
-^
+
-^
+
...... is
greater
than the
corresponding
term of the
divergent
series
-
+
-
+
......
;
the series is
L
'
therefore
divergent
when k
< 1.
326 CONVERGENCY AND DIVERGENCY.
275. The
convergency
or
divergency
of
many
series
can be determined
by
means of Theorems I. and
II.,
using
the series of the last Article as a standard series. The
method will be seen from the
following examples.
2n
Ex. 1. Is the aeries whose
general
term is
-^
^
convergent
or
divergent
?
Since
-?^->-,
if
n>l,
it follows that 2
-|^-,-
> 2
i
. But2-
n'-'
+
l n n-'
+
l n
is
divergent;
therefore 2
-^
=- is also
divergent.
Ex. 2. Is the series whose
general
term is
-
convergent
or
divergent?
n
+
2 n
+
2 3n 3 _ n +
2 ! . _ 1
Now
, ,
< =- <
-j-
<
-,
. Hence 2 -x
r
< 32
-=
. But 2
-&
n
s
+
l n
8
n
1
n* n
s
+l
n
a
n*
is
convergent [Art. 274];
therefore 2
-5
=- is also
convergent.
276.
We have hitherto
supposed
that the terms of the
series whose
convergency
or
divergency
was to be deter-
mined were all of the same
sign.
When, however,
some
terms are
positive
and others
negative,
we-first see whether
the series which would be obtained
by making
all the
signs positive
is
convergent;
and,
if this is the
case,
it
follows that the
given
series is also
convergent
;
for a con-
vergent
series,
all of whose terms are
positive,
would
clearly
remain
convergent
when the
signs
of some of its
terms were
changed.
If, however,
the series obtained
by
making
all the
signs positive
is a
divergent
series it does
not
necessarily
follow that the
given
series is
divergent.
For
example,
it will be
proved
in the next Article that
the series
|
+ +...
is
convergent, although
the
series
-j-
+ +
J+i
+ ... is
divergent.
A series which would be
convergent
if all the terms
had the same
sign
is called an
absolutely convergent
series.
277.
Many
series whose terms are
alternately positive
and
negative
are at once seen to be
convergent by
means of
CONVERGENCY AND DIVERGENCY.
327
Theorem V. A series is
convergent
when its terms
are
alternately positive
and
negative, provided
each term is
less than the
preceding,
and that the terms decrease without
limit in absolute
magnitude.
Let the series be
W
l
-
W
2
+
W
3
-
U
4+
U
n
+
U
+l
U
n+2
+
......
By writing
the series in the forms
and
we see
that,
since each term is less than the
preceding,
the
sum of the series must be intermediate to
u^
u
y
and u
l
;
and hence the sum of the series is
finite.
It is also
similarly
clear that the absolute value of U U
n
is intermediate to
the absolute values of u
n+1 u^
and w
n+1
,
and therefore
UU
n
becomes
indefinitely
small when n is increased
without limit. The series must therefore be
convergent.
For
example,
the series
T~H+Q-7
+
*s
convergent,
since
the terms are
alternately positive
and
negative
and decrease without
limit. The series
T-
+
O~T+
*B no* however a
convergent
series
although
its sum is a finite
quantity
between
-
and
2,
for the
a
nth
term,
namely
-
,
does not diminish
indefinitely
as n is
indefinitely
increased.
278. We will now
apply
the
preceding
tests of con-
vergency
to three series of
very great importance.
I. The Binomial Series. In the binomial
series,
namely
m (m
1)
,
1 + mac
-f
v
(
a?+
1 2
-\ :
328
CONVERGENCY AND DIVERGENCY.
the number of terms is finite when m is a
positive
integer
;
but when m is not a
positive integer
no one
of the factors
m,
ra
1,
m
2,
&c. can be
zero,
and
hence the series must be endless.
To determine the
convergency
of the series when m
is not a
positive integer
we must consider the ratio
XT
u
+i
w
+
1
ft
m
+
l\
u
n
..:u
n
. Now
!
=
- -
x
=
x 1
--
.
u
n
n
\
n
J
Hence,
for all values of n
greater
than m
+
1,
u
n+l
,
and
w
n
have different
signs
when a; is
positive,
and have the
same
sign
when x is
negative.
Moreover,
as n is in-
creased,
the absolute value of u
n+Jun
becomes more and
more
nearly equal
to x. If therefore x be
numerically
less than
unity,
the ratio u
9+Jun
will,
either from the
beginning,
or after a finite number of
terms,
be numeri-
cally
less than
unity.
Hence
by
Art. 271 the series
formed
by adding
the absolute values of the successive
terms will be
convergent,
and therefore also the series
itself must be
convergent,
whether its terms have all
the same
sign
or are
alternately positive
and
negative.
Thus the binomial series is
convergent,
if a; is numeri-
cally
less than
unity*.
II. The
Exponential
Series. In the
exponential
series,
namely
X
s
x*
the ratio u
n+Jun
is
x/n.
Hence the ratio u
n+Jun
is nu-
merically
less than
unity
for all terms after the first for
which n is
numerically greater
than x. The series is
therefore
convergent
for all values of x,
*
The series is also
convergent
when x
=
l,
provided
n >
-
1
;
and it is
convergent
when x
-
1,
provided
n > 0.
[See
Art. 338.
J
CONVERGENCY AND DIVERGENCY.
329
III. The
Logarithmic
Series. In the
logarithmic
series,
namely
a? a? .
n
x
n
*-2
+
3-
......
-<-!>
+
-'
the ratio
u^,Ju
is
--
^
=
a?
(
1
--
T ) '>
and hence
n+
1
\
n
+ lj
u
n+Jun
will be
numerically
less than
unity provided
x
is
numerically
less than
unity.
The
logarithmic
series
is therefore
convergent
when x has
any
value between
1 and
+
1.
If
#=1,
the series becomes
1|-
+...,
which is
convergent by
Theorem V.
If x
=
1,
the series becomes
(1
+ + +
. .
.),
which
is known to be
divergent. [Art. 274.]
279. The condition for the
convergency
of the
product
of an infinite number of
factors,
and also some other
theorems in
convergency,
will be
proved
in a
subsequent
chapter. [See
Art.
337.]
The two
important
theorems
which follow cannot however be deferred.
280. If the two series
and V
=
v + v
lx+vjc?
+
...... +v
n
a;
n
+
..
be both
convergent,
and the third series
x
+
(u
v
s
+
U&
+
W
2
v
)
a?
be
formed,
in which the coefficient of
any power
of x is
the same as in the
product
of the two first series
;
then
P will be a
convergent
series
equal
to Ux
V,
provided
(1)
that the series U and V have all their terms
positive,
or
(2)
that the series U and F would not lose their con-
vergency
if the
signs
were all made
positive
*.
*
This
Article,
and in fact the whole of this
Chapter,
is taken with
slight
modifications from
Canchy's Analyse Algibriqiie.
330 CONVERGENCY AND DIVERGENCY.
First,
suppose
that all the terms in U and V are
positive.
Then U
in
x F
XB
=
P^
+
terms
containing
#
2"
and
higher powers
of x. Hence
U^
x
V^
>
P^.
Also
P^
=
U
n
x V
n
+
other terms. Hence
P^
> U
n
x V
n
.
Hence
P^
is intermediate to U
n
x V
n
and
U^
x F
2n
.
Now,
the series ZJand V
being convergent,
/",
and
IT,
both
approach indefinitely
near to
U,
also F
2n
and F
n
both
approach indefinitely
near to
V,
when n is
indefinitely
increased. Hence
U^
x
F^,
and U
n
x V
n ,
and therefore
also P
an
which is intermediate to
them,
will in the limit be
equal
to U x V.
Hence,
when all the terms are
positive,
P=Ux V.
Next,
let the
signs
in the two series be not all
positive,
and let IF and V be the series obtained
by making
all the
signs positive
in U and
F;
and let P
1
be the series formed
from U' and V in the same
way
as P is formed from
ffand F.
Then
U^
x F
in P^
cannot be
numerically greater
than
U'^
x V
7
9n
P'.
tn ,
for the terms in the latter
expres-
sion are the same as those in the former but with all the
signs positive.
Now,
provided
the series U and F do not lose their
convergency
when the
signs
of all the terms are made
positive,
it follows from the first case that
If^
x
V^ P'^,
and therefore also U
in
x
F^ P^
which is not
numerically
greater,
must diminish
indefinitely
when n is increased
without limit. Hence the limit of
P^
is
equal
to the
limit of
17^
x
V^ ;
so that P must be a
convergent
series
equal
to the
product
of U and F.
If the series U and F are
convergent,
but are such
that
they
would lose their
convergency by making
the
signs
of all the terms
positive,
the series P
may
or
may
not be
convergent;
and,
when P is not
convergent,
the
relation UxV
=
P does not hold
good,
for P has no
definite value and cannot therefore be
equal
to U x
F,
CONVERGENCY
AND DIVERGENCY. 331
although
the coefficient of
any particular power
of x in
the series P is
always equal
to the coefficient of the same
power
of a; in the
product
of the series U and F*.
281. If the two series
and b
+
\x
+
bj?
+
b
a
x*
+
......
,
be
equal
to one another for all values of x for which
they
are
convergent
;
then will a
=
6
,
a
l
=
b
lt
a
2
=
6
2
,
&c.
For if the series are both
convergent,
their difference
will be
convergent.
Hence
a
-
b
+
(a
4
-
6,)
*
+
(a2
-
6
2)
a?
4-
......
=
.........
(i),
for all values of x for which the series is
convergent.
The last series is
clearly convergent
when x
;
and
putting
x
=
we have a b
=
0. Hence a
=
b .
We now have
as{al
-b
l
+(at -bjx+(at
-b
9
)as*
+
......
}
=
......
(ii).
Now for
any
value x
l
for which the series in
(i)
is
convergent,
a
2
6
a
+
(a
B b^)asl
+
...... is
equal
to a finite
limit,
L
l suppose.
Hence
(ii) may
be written
x^ [a,
b
1
+
x
l
L
l}
=
0; and,
since this is true for all values of a?
t
,
however
small,
it
follows that a
t
6
t
,
must be
numerically indefinitely
small
compared
with L
l
;
that
is,
a, 6,
must be zero. It can
now be
proved
in a similar manner that
2
6
2
=
0,
a
a
_
b
t
=
0,
&c.
Hence
if
two series which contain x be
equal
to one
another
for
all values
of
x
for
which the series are conver-
gent,
we
may equate
the
coefficients of
the same
powers of
x
in the two series.
The
particular
case of two series which have a finite
number of terms was
proved
in Art. 91.
*
It can be
proved
that P is
convergent
if either U or V is
absolutely
convergent.
See
ChrystaPs Algebra,
Part
n., p.
127.
332
EXAMPLES.
EXAMPLES XXVI.
Determine whether the
following
series are
convergent
or
divergent
:
1 1 1 1
+++ + + *"
1.23.45.6
_
a
(a
+
6)
+
(a+ 26) (a+3b)
+
(a
+
46) (a
+
56)
+
'"
3 3.4 3.4.5 3.
4...(n
+
2)
* +
4'
A I*
'
A f* O
' "' r
A / I ft iT>\
'
4 . O 4.O.O * .
O...(_7l
+ J)
3 3. 5 3.5.7 3.5.
7...(2n
+
l)
4
+
4~77
+
4.7. 10
+ '" +
4. 7.
10...(3+1)
+
1 1 . 3 1 .3.5
1.3.
5...(27
t
-l)
'
3
+
3 . 6
+
3 . 6 . 9
+ '" +
3.6. 9...3
+ "*
EXAMPLES.
333
1+2 1+3 l+
i +
i
+ +
2'+!' 3*+2'
,
_ _ m m* m
15. +
z
+
5
+
x + m x + 2m x + om
ifi
l m m
*
16.
?
+ 1-
5+
...
05+1
x + m x + m*
17-
'1.2.3
+
2.3.4
+ '" +
n(n+l)(n
+
2)
+
18
19
_N __ __5_
+ +
20. l_ l
3
21.
oq
2 4x 6x*
2
+
T
+
To
+ -'1
V+l
'
31
_!_
,
2n-5
_,
4
+
2
a:+
12
a:
*?-6**
25. Shew that the series
111 1
_M^_ J.
- - -
.^
-
J_ ^^^___ i r
l
is
convergent
for all values of
x, except only
when x is the
square
of an
integer.
-
3)}
*".
CHAPTER XXII.
THE BINOMIAL THEOREM. ANY INDEX.
282. IT was
proved
in
Chapter
xx.
that,
when n is
any positive integer,
...
r
......
"I
--
:
-
*
~r ......
If
We now
proceed
to
prove
that the above formula is
true for all values of
n,
provided
that the series on the
right
is
convergent.
When n is a
positive integer
the above series
stops,
as
we have
already
seen,
at the
(n
+
l)th term;
but when
?i is not a
positive integer
the series is
endless,
for no one
of the factors
n,
n
1,
n
2,
&c. can in this case be zero.
It should be noticed that the
general
term of the
,
. . , ,
w(n-l)(w-2)...(n-r
+
l)
,
binomial
series,
namely
s- -
p
a---'
of,
In
cannot be written in the shortened form
- *
X
T
unless
jr [n
r
n is a
positive integer;
we
may
however
employ
the
notation of Art.
241,
and write the series in the form
BINOMIAL THEOREM. ANY INDEX.
335
283. Proof of the Binomial Theorem.
Represent,
shortness,
any
series of t
'^+...
by/(m).
Thus
for
shortness,
any
series of the form 1
+
-^
a;
+
-^
a?
+
. . .
l [
and
-
(m
+
TO). (ra
+
TO) (m
4-
TO).
^ -
^ --
.......
a,
+
i^
+ ......
+ 5v4
.......
(m
4-
TO). r
j-
la
--
4
^
j
--
x
Now the coefficient of of in the
product f(m) x/(w)
is
Mr ,
m
r-lKl
W
r
.
g
n
a
^
r -.^
g
Wj
|r |r-l|_l lr-212"
1
"
[f^flf fc'
that is
I
r
And,
by
Vandermonde's Theorem
[Art.
249 or
261],
( 7W I 77 I
this coefficient is
equal
to
-
:
--
r
,
which is the coefficient
of of
in/(m
+
n).
Thus the coefficient of
any power
of as in
f(m
+
TO)
is
equal
to the coefficient of the same
power
of x in the
product /(w)x/(n);
also the series
/(m), /(TO)
and
f(m
-f
n)
are
convergent,
for all values of m and
TO,
when
x is
numerically
less than
unity [Art. 278].
It therefore
follows from Art. 280 that
for all values of m and
TO,
provided
that a; is
numerically
less than
unity.
336
BINOMIAL THEOREM. ANY
INDEX.
Now it is obvious
that/(0)
=
1,
and
that/(l)
=
(1
+
x)
;
we also know that if r be a
positive integer f(r)
=
(1
+
#)
r
.
Hence,
by
continued
application
of
(a),
we have
/(m) x/(w) x/0>)
x ...
=/(m
+n)
x/(p)
x ...
7*
Now let m
=
n=p
=...
=
-,
where r and 5 are
positive
s
integers;
then
taking
s
factors,
we have
K
r
)WH=^>-
-
,.,
But,
since r is a
positive integer, /(r)
is
(1
This
proves
the Binomial Theorem for a
positive
fractional
exponent:
the theorem is therefore true for
any positive
index.
And,
assuming
that the binomial theorem is true for
any positive
index,
it can be
proved
to be true also for
any
negative
index.
For,
from
(a),
/(-n)x/(n) -/(-+)=/(<>).
Hence,
as/(0)
=
1,
we have
n iS
P
sitive
'
.
Hence
(1
+
x)~~* =/( w),
which
proves
the theorem
for
any negative
index.
284. Euler's Proof. Euler's
proof
of the Binomial
Theorem is as follows.
KULER'S PROOF.
337
Represent,
for
shortness,
any
series of the form
m(m 1) 2 m(m l)...(m r+l) , '
- f
...
1.2 k
by/(m):
thus
/(m)=l
+
mo;
+
v
+ ...
+ a;'
+
......
(i),
+
......
(ii),
and,
Now,
if the series on the
right
of
(i)
and
(ii)
be multi-
plied,
and the
product
be
arranged according
to
ascending
powers
of
as,
the result must involve m and n in the same
wayjwhatever
their values
may
lae!
But,
when m and n
are
positive integers,
we know that
f(m)
is
(1
+
a?)
m
,
and
thatf(ri)
is
(1 +#)",
and the
product f(m) xf(n)
is there-
fore
(1
4-
x)
m
**,
which
again,
as m
+
n is a
positive integer,
is/(m
+
w).
Hence when m and n are
positive integers
the
product f(m) xf(n)
is
f(m
+
ri);
and,
as
the/brm
of
the
product
is the same for all values of m and n it follows
that
/(m)x/(n)=/(m
+
7i)
............
(a),
for all values of m and n
provided f(m)
and
f(n)
are
absolutely convergent. [Art. 280.]
From this
point
the
proof
is the same as in Art. 288.
Ex.1.
Expand (l
+
ar)-
1
.
Put
=
-
1 in the above
formula;
then we have
s. A.
22
338
BINOMIAL THEOREM. ANY INDEX.
This
example
illustrates the
necessity
of some limitation in the
value of x
;
for we know
[Art. 229]
that
1-x+x"-
is not
equal
to
;
unless x is between
-
1 and
+ 1.
1+*
Ex. 2.
Expand (1
-
ar)-
J
.
We have
Here
again
it is clear that the result cannot be true for all values
of x
;
if x
=
2,
for
example,
we should have
l
=
l
+
2.2
+
3.2
2
+
4.2
8
+
,
which is absurd.
Ex. 3.
Expand (1
+
xfi.
We have
(l
+
x)^=I
+
^x
the
general
term
being
t
Hence
(1+*)
2.4.6...2r
Ex.4.
Expand (l-)"i.
Wehave
(i-
x
)-^=l
+
(-^) (-x)
+
'
\
2/
(-j)2+.
X 91
\L
Hence
All the terms are
positive,
for in the
general
term there are 2r
negative
factors.
BINOMIAL THEOREM. ANY
INDEX.
339
Ex. 5.
Expand
(a
3
-
3a?x)^ according
to
ascendiug powers
of x.
-,
Sx
\\
*
=
52 5
2/_l\
2
3' 3V
3/
/
5 2
5.2.1.4.7...(3r-8)
/x /xV
~|
After the
second,
all the
signs
are
positive
;
for in the
general
term
there are r-
2-f r,
that is an even
number,
of
negative
factors.
285. The
(r
+
l)th
term of the
expansion
of
(1
/n w* .-I-
|
is obtained from the rth
by multiplying by
x,
T
that is
by (
1
H
-----
]
x. Now 1
H
--
is
always
V
r
)
r
negative
if n
+
1 is
negative
;
and,
whatever n
+
1
may
be,
n -4- 1
1
H
--
will be
negative
for all terms after the first for
which r>
n+
1.
Hence,
if x be
positive,
the ratio of the r
+
1th and rth
terms will be
always negative
when r > n
+
1. The terms
of the
expansion
of
(1
+
x}
n
will therefore be
alternately
positive
and
negative
after r
terms,
where r is the first
positive integer greater
than n
+
1.
If x be
negative,
the ratio of the
(r
+
l)th
and rth
terms will be
always positive
when r
>n+
1. The terms
of the
expansion
of
(1 x)*
will therefore be all
of
the
same
sign
as the rth
term,
where r is the first
positive
integer greater
than n
+
1
; and,
as a
particular
case,
all
the terms of the
expansion
of
(1 x)
n
are
positive
when n
is
negative.
222
340
GREATEST TERM.
For
example,
all the terms in the
expansion
of
(1
-
x)-
are of the
same
sign
as the
rth,
where r is the
integer
next
greater
than
}
+
1,
BO that r is 3.
Also,
after the
ninth,
the terms of the
expansion
of
(l+xp<
are
alternately positive
and
negative.
286. Greatest Term. In the
expansion
of
(1
+
x)*
by
the binomial
theorem,
we know that the ratio of the
nn __
/*+
|
"I
(r
+
l)th
term to the rth is +
-
#,
that is
r
+
x
[
1 I
;
we also know that x must be numeri-
\
r
/'
cally
less than
1,
unless n is a
positive integer.
First
suppose
that n
+
1 is
negative,
and
equal
to
m. Then the absolute value of the ratio of the
(r
+
l)th
term to the rth term is x ( 1
H
J
. Fonce the rth term
is
=
(r
+
l)th
term
according
as
#[lH )
=
1;
,. > mx
,.>
according
as r
=
,
that is
=
<
L x
<
1 x
Hence,
if
^
be an
integer,
r
suppose,
the rth
A
~~
C
term will be
equal
to the
(r
+
l)th
term,
and these will
(
J
_L
ft\ g>
be
greater
than
any
other terms.
But,
if ^r
JL
"
*C
be not an
integer,
the rth term will be the
greatest
when
.
.
,
.
,
(1
4-
n)
x
r is the
integer
next above ^=
*
.
X X
Next,
suppose
that n
+
1 is
positive,
and let k be the
integer
next
greater
than n
+
1.
Then,
if r be
equal
or
71 -4- 1
greater
than
k,
1 will be
negative
and less than
r
unity;
hence,
as x must be less than
unity,
each term
after the &th will be less than the one before
it,
and
therefore the
greatest
term must
precede
the kth. And
tl -4- 1
since,
for values of r less than n
+
1,
1 will be
r
GREATEST TERM. 341
positive
;
the rth term will be
=
(r
+
l)th according
as
[
1
)
a?
= 1
;
that
is, according
as r
=
*-= .
\
r
J > <
1+a?
Hence,
if *-=
be an
integer,
r
suppose,
the rth
1+a;
term
will be
equal
to the
(r
+
l)th,
and these will be
greater
than
any
other terms.
But,
if ^-= be not an
1+a;
integer,
the rth term will be the
greatest
when r is the
(n
+
1)
x
integer
next above
*
f
.
ac+I
Ex. 1. Find the
greatest
term in the
expansion
of
(l-a;)"*,
when
ft
. 1
=- . Here n
+
1 is
negative,
and
V'J"'"'
=
|-f
=4. Hence the
fourth and fifth terms are
equal
to one
another,
and are
greater
than
any
other terms.
Ex. 2. Find when the
expansion
of
(l-x)~
1
*~
begins
to
converge,
if
_3
(1 ~t~ w) x
is
x
Here n
+
1 is
negative,
and
^
'
=
^
,
*
=
22i. Hence the
1 x
%
convergence begins
after the 23rd term.
Ex. 3. Find the
greatest
term in the
expansion
of
(a
+
x)
1
&,
when
IA / x\W
Since
(a
+
x)
~f
=
of ( 1
+
-
J
,
the
greatest
term
required
is the
term
corresponding
to the
greatest
term in ( 1
+
-
J
. Now
(n+l)--T-(
1
+
-
)
=
.
7
-5-
7
=
-;
hence r must be the
integer
next
a
\ a/
^ 4 4 J
greater
than
-
,
so that the 5th term is the
greatest.
m
EXAMPLES XXVII.
1. Find the
general
term in the
expansion
of each of the
following expressions by
the binomial theorem.
(i) (l-cc)~
2
,
(ii) (1-a;)"
3
,
(iii)
'
(1 -x)~
a
.
342 EXAMPLES.
(iv) (1 +*)-, (v) (1+*)*, (vi) (!+),
(vii) (1-5*)-*, (viii) (1-5*)*, (ix)
(1 -*)-?,
(x) (2a+3x)-*, (xi) (o'-2a*)*
>
and
(xii) (4-7*)*.
2. Find the first
negative
term in the
expansion (i)
of
(1
+
*),
and
(ii)
of
(1
+
3.
Find the
greatest
term in the
expansion
of
(1
+
x)
18
when x
=
.
4. Find the
greatest
term in the
expansion
of
(1 1#)~
8
'
when x
=
.
5. After what term will the
expansion
of
(1 a;)
1
*
1
begin
to
converge,
when x
=
1
6. Shew that the coefficients of the first 19 terms in the
expansion
of -r=
ry
are all
positive,
and that the
greatest
(1 x)
of them is 100.
7. If
a,,
a
t
,
a
3
,
a
4
be
any
four coefficients of consecutive
terms of an
expanded
binomial,
prove
that
a,
o
s
2a
g
(t -\- (t Ct + O/ Ct -\- t/
8. Find the
general
term in the
expansion by
the binomial
theorem of each of the
following expressions according
to
ascending powers
of x:
(i>
..
a
+ *
(i
v) (a
+
a)i (a
-
x)~
i,
(v) (a
+
x)' (a
-
x)'
9
,
and
(vi) (a-xY
9. Shew that the coefficient of x" in the
expansion
of
(1 +*)'(! -)"
is 2n.
10. Shew that the coefficient of x" in the
expansion
of
(1
+
2*)
1
(1
-
xY*
is 27
(n
-
1),
n
4
3.
SUM OF COEFFICIENTS.
343
287. Sum of coefficients. The sum of the first
r
+
1 coefficients of the
expansion
of
(1
x)
n
can be ob-
tained as follows.
We have
also
(1 -x)~
1
=
l+x + x*+
...
+
x
r
+...
From
[Art. 281]
the coefficient of of in the
product
of the two series is
equal
to the coefficient of of in
(1 #)"
x
(1 x)~
l
,
that is in
(1 a)"'
1
',
hence we have
i-TT
+
S------K-i)
r
r
1
[1^
|2
\r
=
coefficient of of in
(1
-
x)*~
l
=
(-!)'
~
Similarly,
if <
(x)
=
+
a
t
a?
+
a
2
#
8
+
. . .
+
X
+
. .
.,
the
i
/_\
sum a
n
+
a,
+
. . .
+
a
r
will be the coefficient of of in
^
.
1 x
Thus,
to find the sum of the first r
+
1 coefficients in the
expansion
of
<J> (x),
we have
only
to find the coefficient of
., <f>
(x)
x
r
in the
expansion
of
^-
-
.
J. C
Ex. 1. Find the sum of the first r coefficients in the
expansion
of
(1-a:)-
8
. Am.
Jr
(r+l)(r+2).
The sum
required
is the coefficient of a:
1
""
1
in
(1
-
#)-
4
.
Ex. 2. Find the sum of n terms of the series
1.2.3 + 2.3.4 + 3.4.5+ ......
......
];
the
JV * V
sum
required
=
6 x sum of the first n coefficients in the
expansion
of
(1
-
x)~*=
6 x coefficient of a;""
1
in
(l-x)-
8
=-n(n+l) (n
+
2)(n
+
3).
Ex. 3. Find the sum of the first
n+r coefficients in the
expansion
of
344
BINOMIAL SERIES.
The sum
required
=
coefficient of x
n+r
~
l
in the
expansion
of
_
,
r)
j
+higher po
wers of
(i
_
_
JL 31
(l
+
x)
2 712
1*-1
n(n-l)2
3
Hence
(r^F
=
(T-^~{i^)
5+
~^^-"
+an
integnd
expression
of the
(n
-
3)th degree.
The coefficients of x
n+r
~
1
in
(l-x)-*, (1-x)-*
and
(l-ar)"
1
re-
spectively
are
-
(n+r)(n
+
r+l),
n
+
r,
and
1;
hence the coefficient of
2*-
1
(n-f r) (n+r+1) -^-^(n+rJ
+
^-'n
(n
-
1).
Ex. 4. Find the sum of n terms of the series
1.2
Ant.
(2n-l)I/n!(n-l)l.
288. Binomial Series. Series which are derived
from the
expansion
of
(1
+
#)" by giving particular
values
to x and n are of
frequent
occurrence: it is therefore of
importance
to be able to determine at once when a
given
series is a binomial series.
The case in which the index is a
positive integer
needs
no remark.
When the index is a
negative integer,
we have
n(n
...
-
|r
and it should be
carefully
noticed that this
expansion
can
be written in the form
I 71
~~
BINOMIAL SERIES. 345
When the index is
fractional,
p/q suppose,
we have
,
~
Here we notice that
(i)
there is an additional factor
both in the numerator and in the denominator for
every
successive
term,
(ii)
the successive factors of the numerator
are in an A. P. whose common
difference
is the denominator
of
the
index,
(iii)
the successive factors of the denominator
are
1, 2, 3, 4, &c.,
or
multiples
of these.
Bearing
in mind the above
laws,
there will be no
difficulty
in
determining
the
expression
which will
pro-
duce a
given
binomial series.
Ex. 1. Find the sum of the series
1 1.3 1.3.5
3
+
3T6
+
3TO
+
Writing
the series in the form
1
1
1^3 1^
1.3.5
1^
|T*3
+
[2/P
+
|3
*3*
+ ......
'
we see from
(A)
that it is obtained from the
expansion
of
(1
-
a;)~*
x 1
by giving
to x the value found from
-
=
-
.
2 3
1 3
"
...... =i+s
'
therefore
Ex. 2. Find the sum of the series
2 2.5 2.5.8
Writing
the series in the form
^
1
2_._5 ^
2.5.8 1
]T'6
+
~JF
6
s
+
|8
'
6
+
.......
,, BW ~.V~*
V ,
~ , ,.^^wv. iiV^ .^w _,..,*. x,^
^ -,,
346 BINOMIAL SERIES,
by giving
to x the value
-
=
^
. Hence the sum
required
is
Ex. 3. Find the sum of the series
+ + '-h to
infinity.
In this case the factors of the
denominator,
although multiples
of
1, 2, 3, 4, &c.,
do not
begin
at the
beginning.
Additional factors
must therefore be introduced in the
denominator,
and
corresponding
additional factors in the numerator. We then hare
|3
6
s
|4
6*
Now the terms of this latter series are terms of
(A),
if
<?=4, p= 5,
We can therefore find the
required sum,
as follows :
/
4\t_
5 1 5.1 1 5.1.3 1 5.1.3.7 1
\ ~6/
l'6
+
(2^6
+
jjJ
6*
+
|4
6*
_
5 5 5
[~3
3.7 3.7.11
~|
L
6
+
6. 12
+
6. 12
|_18
+
18. 24 18.24.30
""_]'
(i\*_l_
JL
+ JLs
\3y
o tff ta
Whence S
=
^{8^27-17}.
o o
K Q K *7
Ex. 4. Find the sum of the series
-
+
'
+
'
+
to
infinity.
4 *x o 4 . b . Xo
Ant.
Ex. 5. Find the sum to
infinity
of the series
1 1.3 1.3.5
23 2
[From (1
+
1)*].
^"*-
24
~
3
V
2'
Ex.6. Shew that
1-^
+
Jl*
+
1^1
+
4TT^
+
* ^
.
2 2.5 2.5.8 2.5.8.11
l
+ +
-
+ + +
*>
infim*.
THEOREMS OBTAINED BY
EQUATING
COEFFICIENTS. 347
289. We know from Art. 281 that if
any expression
containing
x be
expanded
in two different
convergent
series
arranged according
to
ascending powers
of
x,
the
coefficients of like
powers
of x in the two series will be
equal. By
means of this
very important principle many
theorems can be
proved.
Ex. 1. Shew
that,
if n be
any positive integer,
n n
2
(n
a
-l
2
)
n
a
(n
2
-
1
2
) (n
2
-
2
2
)
1
2+
1
2
.2
2
1
2
.2
2
.3
2
n
(n
-
1) (n
-
1) (n
-
2)
We have
(l-x)=l-nx+
\
'
x*
--
v
'\
L
a?+ ......
1 . &
JL.^.O
1.2...n
Also, provided
x >
1,
we have
/
1 _iy*_ 1
1
n(n
+
l)
1
n(n
+
l)(n
+
2)
1
\ /
i * /(/
n(n+l)...(n+n^T)
1
1.2...n x
, ^(n
2
-!
2
)
2
(n
2
-l
2
)...{n
2
-(n-l)
2
}
Hence l-n
2
+
*
a
-
+
(-l)
n
--
,3
*
5^
*+
(1\~*
1
- -
1
,
that is
equal
to the coefficient of x in
(
-
l)
n
x
n
,
which is zero.
[See
also Art.
251,
Ex.
3.]
Ex. 2. Find the sum of
i
-11 1
1>3 - 5 -
(2-l)
n
'2
+
^
n
'2.4
+ + '
2.4.6...2n
[Equate
coefficients of a;" in
(1
-
x)~^
x
(1
-
a;)-*
and in
(1
-
x)
"*.
]
5.7...(2n+3)
'
2.4...2n
*
Ex.3. Shew that
l-3n+
(3re
~^
)(
f"
2)
- =
(-!).
X i
1 + a; 1 1
We have
Hence
(l+x){l-a?
+
x
e
... +
(-l)
n
x**+
...}
=
1
+
x
(1
-
x)
+
a?
(1
-
x)
2
+
. . .
+a?"*
1
(1
-
3)
3
*+1
+ ...
The coefficient of a^
n+1
on the left is
(
-
1)
B
.
34-8
EXPANSION OF MULTINOMIALS.
The terms on the
right
which
give
<r
3n+1
are
a?"*
1
(1
-
x)**
+1
+
X
s*
(1
-
z)
+
o^-
a
(1
-
x)
3 1
+ . . .
;
and hence the coefficient of x
3""1
"
1
will be found to be
a_.(8it-l)(3n-2) (3n-2)(3n-3)(3n-4)
1.2 1.2.3
290.
Expansion
of Multinomials.
Any
multi-
nomial
expression
can be
expanded by
means of the
binomial theorem.
Since
(p
+
qx
+
ra?
+
...)" may
be written in the form
p
n
(l
+
-
x
+
-
x*
+
...
J
,
it is
only necessary
to consider
\
{* t?
'
expressions
in which the first term is
unity.
Now in the
expansion
of
{1
+
ox
+
bx* + ex*
+
...}",
that is of
(1
+(ox
+
bx*
+
cx*
+...)}", by
the binomial
theorem,
the
general
term is
n
(n 1) (n 2)
. . .
(n
r
+
1)
, ,
2 , xr
p
'(ox
+ bof
+
cof+
...)
r
;
also in the
expansion
of
{ox
+
bx*
+
ex* +
. .
.)
r
,
r
being
a
positive integer,
the
general
term is
by
Art. 262
where each of
a, /3,
7,...
is zero or a
positive integer,
and
Hence the
general
term of the
expansion
of the
multinomial is
|o |/8 |T_...
To find the coefficient of
any particular power
of
x,
say
of
x*,
we must therefore find all the different sets of
positive integral
values
(including zero)
of
a, /9, 7,...
which
satisfy
the
equation
o
+
2/8
+
87
+
...
=
k;
the cor-
responding
value of r is then
given byr=a+/3
+
7+...,
and the
corresponding
coefficient is found
by substituting
EXPANSION OF MULTINOMIALS.
349
in the formula for the
general
term. The
required
coeffi-
cient will then be the sum of the coefficients
corresponding
to each set of values of
a., $,
7
____
Ex. 1. Find the coefficient of a? in
(1
-
x
+
2x-
-
3x
3
)-^-
The values of
a,
/3, y
which
satisfy
+
2/3+87=6
will be found
to be
0,
1, 1; 2, 0, 1; 1,
2,
0; 3, 1, 0;
and
5, 0,
0. The cor-
responding
values of r will be
2,
3, 3,
4 and 5
respectively
;
and the
corresponding
coefficients will be
-!) (-1) 2/V
2/
_
' V ; ' V
' V
,,,
1 1
'
'
[ jj.
(JKJMliKJ)
....
9 45 15 35
63
thatis
~2' TU' 4' -16
and
256-
31
Hence the
required
coefficient is
.
*<Ot)
291. From the above
example
it will be seen that
the
process
of
finding
even the first six terms in the
expansion
of a multinomial is
very laborious;
in
many
cases, however,
the work can be much
shortened,
as in
the
following examples.
Ex. 2. Find the coefficient of x
13
in the
expansion
of
We
Hence the coefficient
required
is zero.
Ex. 3. Find the coefficient of x* in the
expansion
of
(1
+ x
+
x
2
+
x
3
}-
1
.
We have
-
1-x'
350 COMBINATIONS WITH REPETITIONS.
Hence the coefficient of X** is
1,
the coefficient of x*
r+l
is
-
1,
the coefficient of x*^
2
is
zero,
and the coefficient of x
4r+s
is zero.
Thus the coefficient of x" is 1 when n is of the form
4r,
it is
-
1
when n is of the form 4r
+ 1,
and it is zero when n is of either of the
forms 4r
+
2 or 4r
+
3.
Ex. 4. Find the coefficient of x
r
in the
expansion
of
(l
+
2ar
+
3x
2
+4i
s
+ ...... to
infinity)*.
Since 1
+
2^
+ 30?+
......
=
(l-x)~
2
,
the
required expansion
is
that of
(1
-
x)~
2n
;
the coefficient of x
r
is therefore
292. Combinations with
repetitions.
The number
of combinations of n
things
a
together
of which
p
are of
one
kind,
q
of a
second,
r of a
third,
and so
on,
can be
found in the
following
manner.
Let the different
things
be
represented by
the letters
a,b,c,...'y
and consider the continued
product
It is clear that all the terms in the continued
product
are of the same
degree
in the letters
a, b, c,...
as in
x\
it
is also clear that the coefficient of of- is the sum of all the
different
ways
of
taking
a of the letters
a, 6, c,...
with the
restriction that there are to be not more than
p
as,
not
more than
q
b's,
&c.
;
so that the coefficient of of- in the
continued
product gives
the actual combinations
required.
Hence the number of the combinations will be
given by
putting
a
=
6
=
c=...
=
l. Thus the number of the com-
binations of the n
things
a
together
is the coefficient of
of- in
(1
+
x
+
a?
+ ...+
x
f
) (1
+
sc+ ...+x
y
) (1
+
x
+
...
+&%.
Permutations. The number of
permutations
of the
n
things
a
together being represented by
P
a,
it is
easily
seen that
iar
2
-4- 4--^;
PERMUTATIONS.
351
For la x the coefficients of a? in
ax aV
a^
p
)
H
i !?
'
"
L bx 6V
x
l
1+
ll
+
^-
f
is the sum of all
possible
terms of the form
la
\i_\m
a
l
b
m
,...
for which I
+
m + ...
=
a,
and the number of
permutations
a
together
formed
by taking
I of the
a's,
m of the
b's,
&c. is
Ex. 1. Find the number of combinations 7
together
of 5
a's,
4 i's
and 2 c's.
The number
required
is the coefficient of x
7
in
...
x*)(l
+
x
+ x
2
),
that is in
(1-s
6
) (l-x
5
) (1-z
3
) (1-z)-
3
.
Bejecting
terms of
higher
than the seventh
degree
in the continued
product
of the first three
factors,
we have
(1
-
*
-
x
8
-
x
6
) (1
+
3x
+
6x
2
+
lOx
3
+
15x*
+
Six
3
+ 28x
+ 36x
7
+
...);
and the coefiScient of x
7
is 36
-
15
-
6
-
3= 12.
Ex. 2. Find the total number of
ways
in which a selection can be
made from n
things
of which
p
are alike of one
kind, q
alike of a
second
kind,
and so on.
The total number of the combinations is the sum of the coef-
ficients of a;
1
,
a;
2
,...,
x
n
in
(l
+
a;
+ ... +a;
p
) (l
+a;+ ...
+afl)... ;
and this
sum is obtained
by putting
x=l in the
product
and
subtracting
1
for the coefficient of x. Hence the
required
number is
The above result
can, however,
be obtained at once from the
consideration that there are
p
+
1
ways
of
selecting
from the
a's,
namely by taking 0,
or
1,
or
2,...
or
p
of
them; and,
when this is
done,
there are
g
+
1
ways
of
selecting
from the
b'a;
and so on.
Hence the total number of
ways, excluding
the case in which no
letter at all is
selected,
is
( p
+
1) (q
+
1)...
-
1.
[Whitworth's
Choice
and
Chance, Prop, xni.]
352 HOMOGENEOUS PRODUCTS.
Ex. 3. A candidate is examined in three
papers
to each of which m
marks are
assigned
as a maximum. His total in the three
papers
is
2m;
shew that there are
-(m
+
l)(m
+
2) ways
in which this
may
m
occur.
The number of
ways
is the coefficient of z
2"*
in
(l+x
+ z
2
+
...a*)
3
,
that is in
(l-
Hence the number
required
Ex. 4. Shew that the number of
permutations
four at a time which
can be made of n
groups
of
things
of which each consists of three
things
like one another but unlike all the rest is n
4
-
n.
The number
required
is
equal
to
|4
x the coefficient of x
4
in
293.
Homogeneous
Products. We have
already
[Art. 250]
found the number of
homogeneous products
of r
dimensions which can be formed with n
letters,
where each
letter
may
be
repeated any
number of times. We now
give
another method of
obtaining
the result
Suppose
the
letters to be
a, b, c,... ;
then if the continued
product
x
(1
+
ex +
cV + cV
+...)...
be
formed,
the coefficient of of will
clearly
be of r dimen-
sions in the letters
a, b, c,...,
and will be the sum of all the
possible ways
of
taking
r of the letters*. Hence the
number of the
products
each of r dimensions will be
given
by putting
a=6=c=...
=
l in the continued
product.
Thus the number
required
is the coefficient of of in
(1
+
x
+
of
+
. .
.)",
that is in
(1 #)~".
Hence
n(n
+
l)...(n
+
r
1)
\n
+
r 1
=
This result can be
expressed
in the form
JJr
=
n+r
_
l
C'
r
.
An
expression
for the sum of the
homogeneous products
will be
found in Art.
300,
Kx. -1.
EXAMPLES. 353
Con. The number of terms in the
expansion
of
n
+
r 1
(a,
+
a
2
+
a,+
... +
aj
is .
294. We shall conclude this
chapter by solving
the
following examples.
Ex. 1. Find
^14, by
the binomial
theorem,
to six
places
of decimals.
J l - *
-.--
2-^85-^476
=4
{1
-
-0625
-
-001953
-
-0001220
-
-0000095
-
-0000010}
=
3-741657.
Ex. 2. Shew
that,
when x is
small,
*)-
3
,
=
1
+
jr
x
approximately.
*)-i
Since x is
small,
its
square
and
higher powers may
be
rejected ;
and when all
powers
of x
except
the first are
neglected
the
given
expansion
becomes
equal
to
.
4
Ex. 3. Shew that the
integral part
of
Cs/3
+
1)**'
is
(^3
+
1)
2"^
-
U/3
-
l)-
Since
*JS
-
1 is a
proper fraction,
(^3
-
1)
2B+1
must also be a
proper
fraction. It therefore follows that if
(^3
+
l)
2n+1 -
(^/3
-
l)
2n+1
be an
integer,
it must be the
integral part
of
(*/3
+
1)
2""*"1
.
Now
=
{
3"
^3
+
(2n
+
1)3"
+
-{3 N/3-(2n+l)3+
all the irrational terms
disappearing,
S.A.
23
354
EXAMPLES.
Since the coefficients of all the different
powers
ot '6 in the last
expression
are
integers,
it follows that
(v/3
+
l)
;n
'
H
-
(^3
-
I)***
1
is
an
integer,
and is moreover an even
integer.
By
the
following
method it can be
proved
that
U/3
+
1)
2"*1
-
(^3
-
I)
2"*1
is an
integer
divisible
by
2+.
Represent (^3
+
lp+i
-
(^3
-
1)"'"" by 1^.
Then
I,
=
2;
and it will be found that I
8
=20,
and also that
Hence
U/3
-
It follows from the last relation that I
2n+3
w^^ ^e an
integer
if
^2n-H
and
^2n-i
are
integers.
Now we know that
Jj
and
/,
are
integers ;
hence
by
induction I
8n+1
is
always
an
integer.
The relation
(A)
also shews that
I^n+s
will De divisible
by
2*~
M
provided
lo^+j
is divisible
by
2
ft+1
and
/2j-i by
2". Now we know
that J
x
is divisible
by
2
l
and I
3
by
2-
;
hence Z
5
must be divisible
by
2
3
;
and it will then follow that I
7
must be divisible
by
2
4
;
and so
on,
so that I
2n-t-i
is
always
divisible
by
2
n+1
.
Ex. 4. To shew
that,
if n be
any positive integer,
a*
-
n
(a
+
&)
+
"
(
"
"
'^
(a
L . Z
Put
y
j
for x in the
identity proved
in Art.
259,
Ex. 3
; then,
after
reduction,
we have
fc
5
"
.
C
Q 'i
...(y
+
a+nb) y
+
a
- +
(
~
1)r
y
+
a
+
rb
+
Now
expand
the
expressions
on the two sides in
powers
of
-
.
|nfc (nfc
Left side
=
\ . <.
=
~,
T
+
higher negative
R+
powers
of
y.
hence the coefficient of
-J-T-,
on the
right
is
yk+l
(
-
l)*[Coa*-e,
(a
+
b)*+
...... +
(
-
l)'cr
Hence S
(
-
l)
r
c
r (a
+
rb)*
is zero if
k<n,
and is
equal
to
-l)fc
n
nif k=n.
EXAMPLES. 355
EXAMPLES XXVIII.
1.
Find the sum to
infinity
of each of the
following
series :
13 1.33
2
1.3.53'
+
jT2~
3
+
~J2r
2*
+
~~J3~2
i+
-
/\
i_il
1-3 1 1.3.5 1
22
+
2.42
3
2.4.62
3+
'"
41 4. 71 4.7.101
1+
]ll
+
^^
+
jT
3
+ -
3^5
3.5.7
_3 15^._7JL9_
^
376
+
3. 6.9
+
3. 6 . 9 . 12
+
3 3.4 3.4.5 3.4.5.6
__
'
24 2. 4. 6 2.4.6.8 2.4.6.8.10
2 2.5 2.5.8
+ + +
6712 628
3 3.5 3.5.7
-4
+
4-8-4T^T2
+
....
4 4.12 4.12.20
2 2.5 2.5.8
1++ +
(x)
9.18 9.18.27 9. 18. 27. 36
1 1 . 3 1.3.5
(X1)
2.4.6
+
2.4.6.8
+
2.4.6.8.10
+
(
"\ L
7 - 28 7.28.49
72
+
72. 96
+
72. 96. 120
+ *"
2. Shew
that
a
n(n+l)/
a
\
s
i + n
_
4-
v
_f I _
I 4-
a + b 1.2
\a
+
b)
'"
_?_
b
"
,
4-
,
-
n
---
,
a + b 1.2
\a
+ b
232
356
EXAMPLES.
3. Shew that
o.d
1-*
n(+l)/l-*\*
=2
n
\-n-
-
+
\
M-
-
)
1 + x 1.2
\1
+
x)
_
n(n
+
\)(n+ 2)
/l-g\
3
\
1.2.3
+ -'--
4. Shew
that,
if a; be
greater
than
-
,
a; 1 / x \* 1.3 / x \
3
5. Shew that
1.3.5/ a; \
4
+
2 . 4 . 6
VI
+
/
,
I <B
6. Shew that
a a; ?i (n +
1)
/a
\*
/a +
x\*
1
_l_ A) _
t
V
_
'
I V I - I
AT/t IT ...... I I .
a + x 1.2
\a
+
x/ \
2x
/
7. Shew that
1
+ *'"
=
1 +* + nx 1 + x-
1
+
8. Shew
that,
if a <
6,
& = f bV !---
- -
4 ' 5 ' 6
^
'
l&"
1 '6*
*
1.2 6* 1.2.36*"
1
"
('
9. Shew that
n + x
/
_
I \ / O\
~TT^
+
]2(l'+x)* |3(l+x)
10. Shew
that,
if the numerical value of
y
be less than one-
third of that of
x,
/
2y
\
n(n-4-l)(n+2)
/
2y
\
I
+
^ s : J
+
1.2
\x
+ 1.2.3
'x^*-l72-\x^J"
EXAMPLES. 357
11. Find the value of
to r terms.
12. Shew
that,
if n be a
positive integer,
n
a
(n
-
1)
n
a
(n'-r)(n-2)
111 '"
T
12
13
(1(2
M --n-
_
[s-g
13. Shew
that,
if n be a
positive integer,
(n'-l') M()i'-l')(M'-2')
TIT l! l
--
v /
14. Shew that if n be a
positive integer
< 4
^<.^a^_,*442fcfateia-;._,
1.2 1 8
1 . 2 . o 1 . J . o
15. Shew that
^
n
(n
+
1) (n
+
2) (n
+
3).
16. Prove that
.
n
(
+
!)(+ 2). ox/ o\ o
I
2ra +
o
-
Mn-3)(w-2)
+ ...
=
'
W
17. shew
that,
if
^
=
; 4 -62;,
p.
+
^.-IPI +A-IP.
+
+PIP.-I +p.=
358 EXAMPLES.
1.3.5...(2r- 1) 5.7...(2r-i-3)
2.4.6...2r
'
and
^
=
2.4...2r
>
that
pr
+pr
_
l
q
l
+pr
.
t
q
t
+ -. +
qr
=
|(
r + l
) (
r + 2
)-
19. Shew that
14.2/n n .
o.(n-8)(n-3) (n-3)(n -4) (n-5)
1.2 1.2.3
20. Shew that
- ~
=
(a
+
6)->
-
(n
-
2)
06
(a
+
6)-*
(i-3)(n-4)
.
+
S-
Jr%
--
cfb
(a
+
6)
*
-
...
21. From the
expansion
of
(1
+ 2x +
a^)
2"
prove
that
2n(2-l) 2n(2n-l)(2n-2)(2rc-3)
~{TJT~ ||2
[2
[4
+
[n
|n
=
|2n
[2
'
22. Shew that
n(n+
l)...(n
+ ?-
1) n(n+ l)...(n
+ m-
4)
nro
~|m-3
n
(n
-
1
) n(re+ l)...(n
+
?-7)
1.2 lTO-6
if in >
2rt,
and
=
1 if m
=
2n.
23. Find the coefficient of x" in
(1
+
x) (l+o^) (1+a;
4
) (1+a;
8
)
24. Shew
that,
if x be a
proper fraction,
25. In how
many ways
can 12
pennies
be distributed
among
6 children so that each
may
receive one at
least,
and
none more than three?
\
**
'
'
<
x<1
~
-^ *^
EXAMPLES.
359
26. There are n
things
of which
p
are alike and the rest
unlike
; pi'ove
that the total number of combinations that can
be formed of them is
(p
+
1)
2"~
p
-
1.
27. Shew that the number of
ways
in which n like
things
can be allotted to r different
persons,
blank lots
being
admis-
sible,
is
,+,._,
(7
p _,.
28. Shew that the number of combinations n
together
of
2n
things,
n of which are alike and the rest are all
different,
is 2".
29. The number of combinations n
together
of 3n
things,
of which n are alike and the rest all
different,
is
30. A man
goes
in for an examination in which there are
four
papers
with a maximum of m marks for each
paper
;
shew
that the number of
ways
of
getting
half marks on the whole is
31. Find the coefficient of x* in
(1
-
2r- 2
32. Find the coefficients of x
b
in the
expansions
of
x
4
)
6
and
(1
+ x + x* + x
3
+ x* + a;
5
)
8
.
33. In a
shooting competition
a man can score
5, 4, 3, 2,
1
or
points
for each shot. Find the number of different
ways
in which he can score 30 in 7 shots.
34. In how
many ways
can 20 be thrown with 4
dice,
each
of which has six faces marked
1, 2, 3, 4, 5,
6
respectively]
35. Find the coefficient of x" in the
expansion, according
to
ascending powers
of
x,
of
(4a
s
+ 6ax +
$x*)~
l
.
36. Shew that the coefficient of x
3
"1
in the
expansion
of
1 +
x .
-
--iT*
15 2 + 1.
(1
+ x + x
3
}
3
37. Shew that the coefficient of x
r
in the
expansion
of
(1
+ 2x + 3x*+
...)
2
is
(r
+
1) (r
+
2) (r
+
3).
360 EXAMPLES.
38. Find the coefficient of x* in the
expansion
of
{1
. 2 + 2 . 3a; + 3 . 4x* + ... to
infinity}
1
.
39. Find the coefficient of x' in the
expansion
of
(1
. 2 + 2 . 3 . 2* + 3 . 4 . 2V+ ...... +
(n
+
1) (n
+
2)
2V
+ ... to
infinity)'.
40. Shew that the coefficient of x' in the
expansion
of
41. Shew that if
p-q
be small
compared
with
p
or
q
t
then will
/p
=
(n+l)y
+
(n-l)g
near
-
42. If
(6^6
+
U)
8
"*^^,
and F be its fractional
part;
then will
43. If
(3
73
+
5)"*
1
=
I+F,
where / is an
integer
and F
a
proper
fraction,
then will
F(I+ F)
=
2*
r+1
.
44. Shew that the
integer
next
greater
than
(3
+
J7)
tm
is divisible
by
2*
1
.
45. If m be a
positive integer,
the
integer
next
greater
than
(3
+
^5)"
is divisible
by
2".
46. Shew that the
general
term in the
expansion
of
1 +
x +
y
+
xy
.-
*
-
z
\+>i4x+i}
--^
47. Shew that the coefficient of x" in the
expansion
of
x (. r"-l
f
(r
-
1') (r'
-
2*)
(T^F^
isr
{
l+
-i3-
c+
nf-
+
(^-l')(^^-3-)
<
,
+
..j.
EXAMPLES.
361
48. Shew that
+
(2n-3)(2n-2).
n +
(2n- 1)
2n.l
=
2V.
49. Shew that the coefficient of x"*
r
~
l
in the
expansion
of
50. Shew that the coefficient of a;"*'"
1
in the
expansion
of
>
is -!,
51. Shew that
n
n
-
n
(n 2)"
+
^
-
(n 4)"
-
. . . to n + I terms
=
2.4.6.
8...2n.
52. Shew that
53. If three consecutive coefficients in the
expansion
of
any power
of a binomial be in arithmetical
progression, prove
that the
index,
when
rational,
must be of the form
q*
2,
where
q
is an
integer.
54. Shew that the sum of the
squares
of the coefficients in
the
expansion
of
(1
+
x +
*)",
where n is a
positive integer,
is
\2n
'-
55. Shew
that,
if n is
any positive integer,
1 +
2(2r+l)
+
2.4(2r+l)(2r+3)
+ '"
_
r(r+l)(r
+
2)...
2r(2r+l)(2r+2)...(2
CHAPTER XXIII
PARTIAL FRACTIONS. INDETERMINATE COEFFICIENTS.
295. IN
Chapter
VIIL it was shewn how to
express
as
a
single
fraction the
algebraic
sum of
any
number of
given
fractions. It is often
necessary
to
perform
the converse
operation, namely
that of
finding
a number of
fractions,
called
partial fractions,
whose denominators are of lower
dimensions than the denominator of a
given
fraction and
whose
algebraic
sum is
equal
to the
given
fraction.
296. We
may always suppose
that the numerator of
any
fraction which is to be
expressed
in
partial
fractions
is of lower dimensions in some chosen letter than the
denominator.
For,
if this be not the case to
begin
with,
the numerator can be divided
by
the denominator until
the remainder is of lower dimensions : the
given
fraction
will then be
expressed
as the sum of an
integral expression
and a fraction whose numerator is of lower dimensions
than its denominator.
297.
Any
fraction whose denominator is
expressed
as the
product
of a number of different factors of the
first
degree
can be reduced to a series of
partial
fractions
whose denominators are those factors of the first
degree.
For let the denominator be the
product
of the n
factors x
a,
x
b,
x
c,...;
and let the numerator be
represented by F(x),
where
F(x)
is
any expression
which
is not
higher
than the
(n
1
)th
degree
in x.
PARTIAL FRACTIONS.
3(>3
We have to find values of
A, B, G,..,
which are
independent
of x and which will make
F(x)
ABC
T T
~"
TT
i
(x a) (x 6)
(x c).
. . x a x b x c
or,
multiplying by (x a)(x b)(x c)
,
F(x)
=
A(x-b}(x-c}
+
B
(x
-
a) (x
-
c)
+
C
(x -a) (x -b) (i).
In order that
(i) may
be an
identity
it is
necessary
and sufficient that the coefficients of like
powers
of x on
the two sides should be
equal.
Now F
(x)
is of the
(n
l)th degree
at
most,
and the terms on the
right
of
(i)
are all of the
(n l)th degree
;
hence,
by equating
the
coefficients of
x,
x
1
,...
x*~
l
on the two sides of
(i),
we have
n
equations
which are sufficient to determine the n
quan-
tities
A, B, C,
The values of
A, B, C,...
can however be obtained
separately
in the
following
manner. Since
(i)
is to be
true for all values of
x,
it must be true when x
=
a
; and,
putting
x
=
a,
we have F
(a)
=
A
(a 6) (a c)
;
and
therefore A
=
F(a)/(a b) (a c) Similarly
we have
B
=
F(b)l(b
-
a) (b
-
c).
. .
;
and so for
G, D,.
. ..
We have thus found values of
A, B, G,
... which make
the relation
(i)
true for the n values
a, b, c,
... of x
;
and
as the
expressions
on the two sides of
(i)
are of not
higher
degree
than the
(n l)th,
it follows
[Art. 91]
that the
relation
(i)
is true for all values of x.
Thus
F(x)
^^
F(a)
1
(x a)(x b)(x c)... (a 6)(a c)...
x a'
Sx 4- 7
Ex. 1. Eesolve -.
^- ^
into
partial
fractions.
(x
-
i)(x
-
Z)
3a;+7 A B
Assume
=
h
:
(z-l)(x-2)
x-l x-2'
then
364 PARTIAL FRACTIONS.
In this
identity put
x
=
l;
then 10 = -^. Now
put
x=
2;
then
18=JB.
3x
+ 7 13
JLO
(x-l)(*-2)~:c-2~*-l'
Ex. 2. Resolve .-^
;-,
-
rr?
-
r
into
partial
fractions.
(x-a)(x-b)(x-c)
Let
(fc-c)(c-a)(-t)
=
^ C
(x-a)(x-b)(x-c)
x-a x-b x-c'
then
(b-c)(c-a)(a-b)=A(x-b)(x-c)
+
B(x-c)(x-a)
IT C
(x-a) (x-b).
Putting
x
=
a,
we have
(6
-
c) (c
-
o) (a
-
6)
=
A
(a
-
b) (a
-
c)
;
there-
fore A
=
c
-
b;
and the values of B and C can be written down from
symmetry.
cc
-
a
(
a
-fr)
c
~
b ffl
~
c >-
Thus
=
(x
-
o) (x
-
b)(x
-
c)
x-a x-b x-c'
Ex. 3. Eesolve
;
-
=T-J
-
~
-
;
-
r into
partial
fractions.
2)...
(x+n)
Assume
x x+1
' '
x+r
Then,
we have
If we
put
z=0,
all the terms on the
right
will vanish
except
the
first,
and we shall have l
=
J x
In,
so that A
=ll\n.
To find the
general term,
put
x=-
r;
we then have
that is l
=
(-l)MP
[rJn-r;
hence
A,
=
(
-
l)
r
/[r_ |n
-
r.
Hence the
required
result is
1J1 \
n
1 1
\n\x~
- +
(
~
l
^
~
+ - +
L)
"
'
[See
Art. 259. Ex.
3.]
Ex. 4.
Express ;
. .
., ,
-
. in
partial
fractions.
(x-a)(x-b)(x-c)
*
Ant. S
pa
(a
-
b) (a
-
e)
x
-
a
'
PARTIAL FRACTIONS.
365
--
Ex. 5. Resolve -.
-
,. ,
--
=r- into
partial
fractions.
(x-l)(x
3
+
2x
+
5)
The factors of x
2
+ 2x +
5 are the
complex expressions
x +
1
+
2i
and x
+
1
-
2i,
where t is written for
>/
-
1.
A
Assume
,
.
,. .o . .. . K>
=
-
T +
Pat
=1;
then 16
=
84,
so that A
=
2.
Put
ar=-l-2i,
then
(l
+
2t)
2
+
15=5
(-2-2f) (-4i),
that is
12
+
4i =
B
(
-
8
+
8i)
;
therefore B
=
-
jr-^.
.
Change
the
sign
of i
2i
in the value of
B,
and we have C=
-
V -f
JLV *J O
+
t
Thus
2-2i
298. We have in the last
example
resolved the
given
fraction into three
partial
fractions whose denominators
are all of the first
degree,
two of the factors of the denomi-
nator
being imaginary. Although
it is for most
purposes
necessary
to do
this,
the reduction into
partial
fractions,
of
a fraction whose denominator has
imaginary
factors,
is often
left in a more
incomplete
state.
Take,
for
example,
the
fraction
just
considered,
and assume
A Bx
+
C
.,
r
[It
is to be noticed that we must now assume for the
numerator of the second fraction an
expression containing
x but of lower
degree
than the
denominator.]
Putting
a;
=
1,
we have 16
=
8.4,
so that A
=
2.
Put A
=
2 in the above
identity;
then after
transposi-
tion
-
a
2
-
4sx
+
5
=
(Bx
+
(J) (x
-
1)
;
or,
dividing by
x
I,
Bx
+
C
=
x 5.
ne* I
|
*"\
/ /> I
-
mi
**'
t^
Ac/ *. t*/
^^
^J
1 n i m
~
^
(a?-l)(^+2o;
+
5)
x-\
a?+%c+~5'
366
PARTIAL FRACTIONS.
299. We have hitherto
supposed
that the factors of
the denominator of the fraction which is to be
expressed
in
partial
fractions,
were all different from one another. The
method of
procedure
when this is not the case will be
seen from the
following examples.
O/r -L.
^
Ex. 1.
Express ,
-
-^-,
-
^-.
in
partial
fractions.
(x-l)*(x-3)
We
may
assume that
2x+5
A B C
_D_
(x^l)(x
-
3)
=
(x-l)
3+
(x^Tj
a
+
(x
-
1)
+
x^3
;
or,
clearing
from
fractions,
By equating
the coefficients of
x,
x
1
,
x
3
,
x
3
on the two sides of the
last
equation,
we shall have four
equations
to determine the four
quantities A, B, C, D,
so that the
assumption
made is a
legitimate
one. The actual values of
A, B, C,
D are not however
generally
best found from the
equations
obtained
by equating
the coefficients of
the different
powers
of x. In the
present case,
the
following
method
is more
expeditious.
Put x
-
l=y
;
then we have
Now
equate
coefficients of
y, y
1
,
y-, y
3
,
and we have 7= -%A
;
A-1B; 0=B-2C;
andO=D + (7.
Whence ^
=
-L
JB=
-
^,
C=
-^
and D
=
^
.
^ 4 o o
~-, T
, 11 7 11 11
Hence
, ,,,
)(x-3) 8(x-3) 2(x-l) 4(x-l)
s
8(x-
ll +x)*
Ex. 2.
Express
the fractional
part
of
-^ g~
in
partial
fractions.
Assume
(l+x)
ABC
mte ral
P*
S1 -
Then
x)*=:A+B (1
-
2x)
+ C
(1
-
2x)
s
+
(1
-
2x)
3
x
integral expression.
Now
put
1
-
2x
=
y
;
then
taining higher powers
of
y).
PARTIAL FRACTIONS. 367
Also
right
side
=
A +
By
+
Cy*
+
y
3
x
integral expression
in
y.
Hence, equating
coefficients of
y, y
l
,
y*,
we have
. 3 n3-'
n(n-l)3-'
A
=i'
B=
""' *+l
'
300. The
following examples
will illustrate the use
of
partial
fractions.
Ex. 1. Find the coefficient of x
n
in the
expansion
of
according
to
ascending powers
of x.
1 32
We have .
=
-
1 vx
+ Qx
2
i 3x 1
-
2x
=
3
{
1
+
Bx 4- (3x
Hence the
required
coefficient is 3.
n+1
-
2
n
'
H
.
Ex. 2. Find the coefficient of x
n+r
in the
expansion
of
rf-njpv.
(1 2x)
From Ex.
2,
Art.
299,
we have
(l
+
x)*_3
1
_"
3>t
~
1
1 n
(n
-
1)
3"-- 1
(1 2x)
3
~
2*
(1 2x)
5
2*
(1 2x)
2
2
n+1
1 2x
+
an
integral expiession
of the
(n
-
3)th degree.
Whence the re-
quired
result.
Ex. 3. Shew that the sum of all the
homogeneous products
cf n
dimensions of the three letters
a, b,
c is
equal
to
a**
8
(c
-
b)
+
fc+
2
(a
-
c)
+
c+
2
(b
-
a)
(b-c)(c-a)(a-b)
The sum of all the
homogeneous products
of n dimensions is the
coefficient of x
n
in the
product
) [See
Art.
293];
that is in
-^ r-^
r
;
,
which will be found to be
equal
to
(1
-
ox) (1
-
6x) (1
-
ex)
a
8
1 6
2
1 c
8
1
+
(a-6)(a-c)
1-ax
(6-c)(6-a)
l-6x
(c-a)(c-6)
i-cx'
and the coefficients of x" in the
expansions
of these
partial
fractions
is
easily
seen to be
a
n+2
jn+2 C
f2
(~a-6)(a-c)
+
(b^c) (b~^a)
+
(c
-
a) (c
-
b)
'
which
equals
(b-~c)(c-a)(a-b)
368
INDETERMINATE COEFFICIENTS.
Ex. 4. To find the sum of all the
homogeneous products
of n
dimensions which can be formed from the r letters a
lt a,, a,,
,
a
r
.
As in the
previous example,
the sum
required
will be the co-
efficient of ar" in
n r-^j ;
r
,
which will be found
(l-a-
to be
equivalent
to "S,
i-aa)(i-as
)---
l-<*i*
O.W+T-
i
Hence the
required
sum is 2
'(ai-ai
)(a1 -aj...(al
-a
r)'
301. Indeterminate coefficients. We shall con-
clude this
Chapter by
giving
two
examples
to illustrate a
method,
called the method
of
indeterminate
coefficients,
which
depends upon
the theorems established in
Articles
91 and 281.
Ex. 1. Find the coefficient ofx
r
in the
expansion, according
to
ascending
powers
of
x,
of
(1
+
cz) (l
+
c
2
*) (l+c
3
x)...(l
+
c*x).
The continued
product
is of the nth
degree
in x
;
we
may
therefore
assume that
where A
,
A
1
,
A
a
,...
do not contain x.
Now
change
x into ex
; then,
since A
,
A
v
A
v
&c. do not contain
x,
we have
Hence
Now
equate
the coefficients of x
r
on the two sides of the last
identity,
and we have
A
r
+
f^Af^m
Arc*
+
A^cf
;
t _
c
+l-
c
r
f _ f
c
-r+l
_
1
f '
A
r
--
cr_i
A
r-l
<
cr_i
A
r-l
......
W'
By
continued
application
of
(a)
we have
f
r
,r-l ,
3
,(C*- -1) (*-'+* -1)...(C*-* -l)(c"-l)
^1)...(C-1)(C-1)
1}
'
for J
-
obviousl7
L
INDETERMINATE COEFFICIENTS.
369
Ex. 2. To find the sum of the series I
2
+
2"
+
3
2
+ . . . +
n
2
.
Let l*
+ 2*+S*+...+n*
=
A
1
n
+
A
t
n*
+
A
s
n*......
(a)
for some
particular
value of
n,
where A
lt
A
z ,
A
3
do not contain n.
The relation
(a)
will be true for n
+
1 as well as for
n, provided
or, subtracting (a), provided
Now the last relation will be true for all values of n if we
give
to
A
lt
-4
S ,
A
3
the values which
satisfy
the
equations
found
by equating
the coefficients of n
2
,
n
1
and
n,
namely,
the
equations
3^3=1, 34,+2^2
=2,
and
Af
+
A^
+
A^l,
from which we obtain 6A
1
=2A,
t
=3A
s
=l.
Hence,
if the relation !
2
+
2
2
+ ... +n'
z
=^
n
+
s
n
a
+-
n
3
,
be true
O <& o
for
any
value of
n,
it will be true for the next
greater
value. But it
is
obviously
true when
n=l;
it will therefore be true when n
2;
and, being
true when
n=2,
it must be true when
n=3;
and so on
indefinitely.
The sum of the
cubes,
or of
any
other
integral powers,
of the
first n
integers
can be found in a similar manner.
[See
also Art.
321.]
EXAMPLES XXIX.
Resolve into
partial
fractions :
3.r
2
x+l
x* + 1x + 6
'
x
3
-
5x + 6
'
3
x
<_. ^
(2-ar)*(l
+
aj)*
x
3
-4af + x + 6*
a;
2
3
^
1 + 7a
-
x
2
9.
./-r*
10.
5
-
gx
(l-3x)*(l+x)'
S.A.
24
370 EXAMPLES.
11
6a + a;-l
x*+2
(x-2)
3
(x'+l)'
18.
* +
. , t
. n
*-* +l
(*-l)
i
(a;
a
+
4)' (x- !)*(*- 2) (x*
IS.
3-7
. . =-r . 16.
x-
!
(a;+2)
3
(a;+l)
17. Find the coefficient of x" in the
expansion
of
x 4- 4
x* + 5x + 6
'
18. Find the coefficient of x" in the
expansion
of
x-2
19. Shew that the coefficient of a;**"
1
in the
expansion
of
x + 5 I
(P- !)(*+ 2)"
2-'
20. Find the sum of the n first coefficients in the
expansion
39"K
Ot = =
-r-s.
21. Find the sum of the n first coefficients in the
expansion
. 2-5x
"^
"/V
"""""'
f * /\ t* \ y ^
o \
"
2x)
n
+ ,j.\
22. Find the coefficient of x" in the
expansion
of
\-
-... .
(l-x)<
*\ Find also the sum of the n first coefficients.
v
23. Shew that the coefficient of cc
n+r
in the
expansion
of
EXAMPLES. 371
24. Shew that r-. ^r -. r
(a5-a,)(o;-a8
) ...(x-aj
a,""
1"'
1
W T^ T^ W r
^4
.
^
(a, -a,) (a, -,)... -a,
25. Shew that the coefficient of z"~* in the
expansion
of
{(1 -)(!- cz) (1
-
c*z) (1
-
c
3
*)}-
1
is
(1
-
O (1
-
O(1
-
O/(l
-
c) (1
-
c
8
) (1
-
c
3
).
v
26. Prove that
a
(
b
-
c
) (be
-
oof) (cT
-
a"") b(c-
a)
(ca
-
bb')
(b
m
-
b'
m
)
a a' b b'
c(a-b) (ab
-
cc') (c
m
-
c'
m
)
c
-
c'
(b c)
(c
a) (a b)
(be
aa') (ca bb') (ab cc') ff^J
abc^
where aa'
=
bb'
=
cc',
and ff
m
_
a
is the sum of the
homogeneous
products
of
a, b, c, a', b',
c' of m 3 dimensions.
^
27. Shew that the
product
of
any
r consecutive terms of
the series 1
-
c,
1
-
c
8
,
1
-
c
8
,
. is divisible
by
the first r of
them.
4rv"-
ftyO,
.
br^Mj
U
2>b^
3^1
*CaQ
U-t
C. -xJ>
28. Shew
that,
if c be
numerically
less than
unity,
(1
+
ex)
(1
+
c*x) (1
+ c
a
x)
... to
infinity
_1
c ^
*
"
29. Shew
that,
if c be
numerically
less than
unity,
(1
+
ex) (1
+
c*x) (I
+
c*x)
... to
infinity
=
ll
c c
4
c^
^
30. Shew
that,
if c be less than
unity.
1
=
1+
x x*
242
372
EXAMPLES.
31. Shew
that,
if c be less than
unity,
(!+)(! +fa)(l+fa)...
1+c
(1
+ C
)(1+Q
(l-x)(l-c*)(l-c'a:)...
~
1-c
(l-c)(l-c)
[Gauss.]
^
32. Shew that the coefficient of x
r
in the
expansion
of
c
being
less than
unity.
33. Shew that
1
v v* v
3
4.
y
4.
&
i
y
i
1~
. T . . .
a; 1
-
ox 1
-
a x 1
-
a a;
1 x x* a?
.
_r _i. ____^^__ j_ j_ -*_
^i
o
'
t ^
'
l-y
I
-ay l-a'y
l-a
3
y
34. Shew that
x 2x* 3x* 4a;
4
n^
+
i^a?
+
r^a?
+
r^
4
+
a; a?
2
as*
35. Shew that Lambert's
series,
namely,
x X
s
x' x*
is
equivalent
to c w>
1-i-x .1+0^
a
l+a^
1
x r
+ x*
5
4- ar
^ -,
+ . . . Clausen.
I
l-x l-o
1
1-a:
CHAPTER XXIV.
EXPONENTIAL THEOREM. LOGARITHMS. LOGARITHMIC
SERIES.
302. The
Exponential
Theorem. If
l/n
be nu-
/ IV*
merically
less than
unity,
( 1
+
-
J
can be
expanded by
the Binomial Theorem
;
and we have
,
1V . . 1
,
nx
(me
-
1)
1
,
- -~
nj
n 1 . 2 w
2
nx(nx V)(nx 2)
1
nx(nx l)...(nx r+1)
1
1.2.3 ~^T
S+
~Y~
which
may
be written
IN-
i
/ IN / r-lN
x
(x
--
...fa;
--
nj
\
n
J
t
Putting
x
=
1,
we have
V.
374
EXPONENTIAL THEOREM.
1.1
(l -!)(!-?)
,
-,
,
_ a.
\
n>
V
^
,
l+ +.
-^-
+
The above relation is true for all values of n however
great,
and therefore when n is
infinite;
but when n is
infinite,
l/n
is
zero,
and the relation becomes*
Denoting
the series 1 +
1
+
T^
+
TK
+ + r- +
...
by
e,
I?
15
we have the
Exponential
Theorem,
namely
\*L \L
It should be remarked that the above series for e* is
convergent
for all values of x
[Art. 278].
303. The
quantity
e is of
very great importance
in mathematics.
It is obvious that it is
greater
than 2 and it is
clearly
less
than 1
+
1
+
2"
1
+
2~
s
+ 2~*+
.
..,
and therefore less than 3.
Its actual value can be found to be 271828....
*
This
requires
more careful examination not
only
to nud the limit of
each
term,
bnt also because the limit of a sum is not
necessarily equal
to the sum of the limits of its terms unless the number of the terms
is
finite.
This examination is however omitted here for the
investigation
in Art. 304 is
preferable.
EXPONENTIAL THEOREM. 375
To
prove
that e is an incommensurable number.
If
possible,
let e
=
m/n,
where m and n are
integers;
then we should have
l
+
n 12
\n \n+l |n+2 |n+3
Multiply
both sides
by
|n;
then all the terms will
become
integral except
1
1 1
i i \
'
/ . t>\ / i^\ / i ~i \
*"
* * *
Hence
1
__
"l
~i i ck\ / '. i \
'
/ . -\ / . r\ / .
-
\
i
must be
equal
to an
integer;
but this sum is less than
T~*~7 TY~*~? TV
"*"''
an^ therefore less than
/ [1 =] ,
that is less than -. But an
integer
n
+
I
I \
n
+ lj
n
cannot be less than
l/n
;
it therefore follows that e cannot
be
equal
to the commensurable number
m/n.
304. The
following proof
of the
Exponential
Theorem
is due to Prof. Hill*. It will be seen that it
only
assumes
the truth of the Binomial Theorem for a
positive integral
exponent.
2 r
Let/(m)
denote the series 1
+m +
^
If
Thus
f(m)
=
*
Proceedings of
the
Cambridye Philosophical Society,
Vol. v.
p.
415.
Substantially
the same
proof
is however
given
in
Cauchy's Analyse Alge-
brique.
376 EXPONENTIAL THEOREM.
ri* n'
r
|s
(m
+
n)
r+>
r
+s
'
[2 |p
"
Now the coefficient of raV
in/(m) x/(w)
is
and in
f(m
+
n)
the term m
r
n' can
only
occur in
\r
+ s 1
and its coefficient will therefore be . -,
,
that is
Hence,
as the series
/(m), f(n)
and
f(m
+
n)
are
convergent
for all values of m and
ft,
and.
the coeffi-
cient of
any
term m
r
n' is the same
iny(m) xf(n)
as in
/(m
-f
n),
it follows from Art 280 that
/(m)x/(n) =/(
+
)
........................
(i)
for all values of ra and n.
Now let a; be a
positive integer;
then from
(i)
we
have
/(I) x/(l) x/(l)
+
...... to a;
factors,
=/(!
+
1
+
1
+ ...... to x
terms),
..{/(!)}"=/(*)
...........................
(ii).
V
Next let a; be a
positive
fraction
-
,
where
p
and
q
are
positive
integers.
Then from
(i)
+ + + ...... to
?
terms
=
(7(1))',
from
(ii);
.-/(?)=(/(.)!.
Hence,
for all
positive
values of
x,
{/(!)}* =/(%)
EXPONENTIAL THEOREM. 377
Lastly,
let x be
negative,
and
equal
to
y,
so that
y
is
positive
;
then
/(- y) x/(y) =/(0)
from
(i) ;
but
/(O)
=
1,
therefore
/(- y)
=
l/f(y).
Hence
/(*) =/(- y)
=
7
=
'*
since
2/
is
Positive,
Heuce,
whatever a;
may
be,
But
therefore e*=a:
=
l
+ +
305. To shew that
We have from Art. 304
Also,
by
the binomial
theorem,
(f .!).=_
n . 4-ij.
+
n
J^
e
*-*>*
-....
(.T;-
x
3
\
n
+
775
+
7-5-+... )
is
zero,
if r is
L_ L_.
'
less than
n,
and is 1 if
r=n.
nln
1)
Also the coefficient of x
r
in e^-n*'
11
"
1
'*-!
^ ^
^e'
n
"
s
'*
... is
Hence, equating
the coefficients of x* in the
expansions
of the
two
expressions
for
(e*
-
l)
n
,
we have
378 EXPONENTIAL THEOREM.
The above theorem
may
be
generalised
as follows:
We have
12
and
(
e
ax _
e
bx)*
gnbx
(g(a-b)x
_
1)
=
e
t>x
j
(
o
_
-t. o+
J4)
x
_
_
Hence,
equating
coefficients of x* in the two
expressions
for
* -
e
te
)
n
,
we have
(
n^2 .a +
26) -...=(-&)".
If we
put
na
=
x and b-
a=y,
the last result becomes
. ..
=
(-!)". y-
n.
We have
also,
if i be
any positive integer
less than
71,
x*-n(x
+
y)*+
n
("~^
(ar
+
2y)*-
... to n
+
1 terms=0.
1 o
The
following particular
cases are of
importance,
k
being
less
than H.
a*_... to n
+
1
terms=0,
1 . 2
and m*-
(m
-
if+-JL^ (
m
_
2)*
-
... to n
+
1 terms=rO.
EXAMPLES XXX.
Ex. 1. Shew that the limit when n is infinite of
\
Ex. 2. Shew that the limit when n is infinite of
(
1
+
-
)
6
is e*
\ nJ
Ex. 3. Shew that
n
n+1 -
n
(n
-
l)
n+1
H :
-^
-
(n 2)*
+1
...
=
-
n
1 . a a
Ex. 4. Shew that
tt +-n
(n
-
I)"""
+
"
(
"
"
1}
(n
-
2)'^
-
...
=
(3/1
+
1)
ln
+
2)
1 & &*
-
LOGARITHMS.
379
Ex. 5. Shew that
O A
f*
Ex. 6. Shew that
e'^ja
+
r?
+
-p,
+ ...
C. j* II
Ex. 7. Shew that
3 1+2 1+2+3 1+2+3+4
' =1+
~"
+
~~
+
-- --
Ex. 8. Shew that
1 1 1
Ex. 9. Shew that
^1
_J _
4-
-
-
-t -i-
~1. 3 1.2.3.5^"'^
Ex. 10. Shew that
l 1 1
) (1
1 1
"
+ *"* *
Ex. 11. Shew that
1 1 1
Ex. 12. Shew that the coefficient of x
n
in the
expansion
of
2e
1 11 11
LOGARITHMS.
306. Definition. The index of the
power
to which
one number must be raised to
produce
a second number is
called the
logarithm
of the second number with
respect
to
the first as base.
Thus,
if a
x
=
y,
then x is called the
logarithm
of
y
to the base
a,
and this is
expressed by
the
notation x
=
log
a
y.
380
PROPERTIES OP
LOGARITHMS.
We
proceed
to
investigate
the fundamental
properties
of
logarithms,
and to shew how
logarithms
can be
found,
and how
they
can be
employed
to shorten certain
approxi-
mate calculations.
307.
Properties
of
Logarithms.
The
following
are
the fundamental
properties
of
logarithms.
I. Since
a
=
l,
for all values of
a,
it follows that
log
a
1=0.
Thus the
logarithm of
1 is
0,
whatever the base
may
be.
II. If
log
a
x
=
a,
Iog y
=
ft,
log
a
z
=
7,
. ..
then x
=
a
a
,
y
=
aP,
z
=
a
y
,...\
/.
xyz
...
=
a
a
. a? . a* . . .
=
a
a+p+lr+
-
'
lo
g. (
x
y
z-
..)
=
a
+ +
7+-
=
log.
x
+
log.y
4-
log
a
z
+
...
Thus the
logarithm of
a
product
is the sum
of
the
logaritiims of
its
factors.
IIL If
loga
x
=
a,
and
log
a
y
=
;
then x
=
a
a
,
y
=
aP,
and /.
#-t-y
=a
a
~P;
'
lo
ga
(#
-^
y)
=
a
-
=
loga
a;
-
log. y.
Thus fo
logarithm of
a
quotient
is the
algebraic differ-
ence
of
the
logarithms of
the dividend and the divisor.
IV. If x
=
a*
;
then at* a
ma
,
for all values of m.
Hence
loga
x
m
=
mi
=
m
log
a
x.
Thus the
logarithm of any power of
a number is the
product of
the
logarithm of
that number
by
the index
of
Hie
power.
V. Let
log
a
x
=
a,
and
log,
x
=
ft;
then x
=
a*
=
b
ft
;
f
and hence a
=
6
a
,
and a?
=
6.
LOGARITHMIC SERIES.
381
Therefore
=
log
ft a,
and
^
=
loga
&.
n fi
Hence
loga
b x
log
h
a
=
-3
x
-
=
1.
p
ot
Also
/3
=
a
Iog
6 a,
that is
Iog
a?
=
log
a
a; .
Iog
6
a.
Hence the
logarithm of any
number to the base b will be
fownd by multiplying
the
logarithm of
that number to the
base a
by
the constant
multiplier Iog
a.
308. The
logarithmic
series. Let a
=
e*,
so that
fc
=
log.
a;
then a*
=
<f*= e
xlo
' a
. Hence from Art.
304,
we have
Put a
=
1
+
y
;
then we have
Now,
provided y
be
numerically
less than
unity, (1
+
y)
x
can be
expanded by
the binomial theorem
;
we then have
x(x-V) #(#-!)(#- 2)...(ar-r+1)
l+xy+
\ 2
;
y
2
+...+-^
-^
-r/-
-
The series on the
right
is
convergent
for all values of a;
and
y,
and the series on the left is
convergent
for all values
of x
provided y
is
numerically
less than
unity.
Hence,
for
such values of
y,
we
may equate
the coefficients of x
on the two sides of the
equation.
We thus obtain
This is called the
logarithmic
series.
382
LOGARITHMIC SERIES.
Ex. 1. To
express
a*
+
b* in terms of
powers
of ab and a
+ b.
From the
identity (1
-
ax) (1
-
bx)
=
1
-
(a
+
b)
x
+
abx
2
=
l-sx+px
2
,
where s is
put
for a +
b and
j>
for
ab,
we have
log. (1
-
ax)
+
log. (1
-
bx)
=
log. (1
-
sx
+px*).
o
2
x
3
a
3
*
3
\
/,
i-x
2
_
Hence
o
2
x
3
a
3
*
3
\
-3-
+
^+
......
J
Equate
the coefficients of x
n
on the two sides of the last
equation.
[This
is allowable since the series can
clearly
be made
convergent by
taking
x
sufficiently small.]
Then the coefficient of x
n
on the left is
-(a*+l*).
On the
right
we have to
pick
out the coefficient of x
n
from the terms
(beginning
at the
highest
in which it can
appear)
the coefficient of x" is therefore
^^-i>."-'
?>
Hence we have
a
+
t
n
=
(a
+
&)
-
nod
(a
Ex. 2. Shew
that,
if a
+
6 +
c=0;
then will
10
(a'
+
V
+
S)
=
7
(a
+ 6
+c
2
) (a
8
+
&
+
<*).
Put
-jp
for bc
+
ca
+
ab,
and
g
for abc
;
then we have the
identity
(1
-
ox) (1
-
6x) (1 -cx)
=l
-px*- qx*.
Now take
logarithms,
and
equate
the coefficients of the different
powers
of x in the two
expansions.
This
gives
-
(a
r
+ b* + c
r
)
in
terms of
p
and
9,
and the
required
result follows at once.
[See
also
Art.
129.]
CAUCHY'S THEOREM. 383
Ex. 3. To
express
a
n
+
6"
+
c* in terms of dbc and bc
+
ca +
ab,
when
Put
-p
=
bc +
ca+ab,
and
q
=
abc;
then we have the
identity
(1
-
ax) (1
-
bx) (l-cx)
=
l-px*- qx
3
.
Hence,
by taking logarithms,
and
equating
the coefficients of like
powers
of
x,
we have
-
(a"
+
6"
+
c")
=
coefficient of z" in 2
-
x*
( p
+
qx)
r
,
n T
which
gives
the
required
result.
The terms in S
-
x*
(p
+
qx)
r
which contain x
6"1 *1
are
Now
by inspection
we see that the coefficient of x
m
~
1
in each
of the above terms in which it occurs contains
pq
as a factor
;
and
also that the coefficient of x
6"1"*"1
in each of the terms in which it
occurs contains
p
z
q
as a factor.
Hence,
when a
+
b
+c=0,
a
n
+b
n
+c
n
is
algebraically
divisible
by
dbc
(bc
+
ca
+
ab)
when n is of the form 6m
-
1,
and
a*+b*+c*
is
algebraically
divisible
by
abc
(be
+
ca +
at)
2
when n is of the form
6m+l.
If we
put c=-(a
+
6),
bc
+
ca
+
ab becomes
-(a?
+
ab
+
b
2
),
and
we have
Cauchy's
Theorem,
namely
that a* + 6"
-
(a
+
b)
n
is divisible
by
ab
(a
+
b) (a
2
+
ab
+
IP)
when n is of the form 6m
-1,
and
by
ab(a+b) (a
a
+a& +
6
2
)
2
when n is of the form
6m+l.
[See papers
on
Cauchy's
Theorem
by
Mr J. W. L. Glaisher and
Mr T.
Muir in the
Quarterly
Journal,
Vol.
xvi.,
and in the
Messenger
of Mathematics,
Vol.
vin.]
309. In order to diminish the labour of
finding
the
approximate
value of the
logarithm
of
any
number,
more
rapidly converging
series are obtained from the funda-
mental
logarithmic
series.
Changing
the
sign
of
y
in the
logarithmic
series
y-
+
-
+ ..............
(i),
we have
384
LOGARITHMIC
SERIES.
Hence
log.
j
-
=
log. (1
+
y)
-
log. (1
-
y)
+ +
......
)
......
(iii).
o
Put for
, ,
and therefore
-
for
y ;
then
n I
y
m
+
n
-
n 1 m
-
ri* 1 /m
-
We are now able to calculate
logarithms
to base e
without much labour. For
example
:
Put m
=
2,
n=
I,
in formula
(iv) ;
then
fl
1 1 11
loor
2
=
2-1 1 1 4- tv/
O
" **
1
Q
'
O Q3
'
C
'
OS
'
from which it is
easy
to obtain the value
log,
2
=
'693147...
Having
found
log.
2,
we have from
(iv)
log.8-log.2-2
ft +
J.i+i
.-U.
..1
=
-405465....
5* 5*^
Hence
log.
3
=
'693147 +
"405465
=
1-09861.
Proceeding
in this
way,
the
logarithm
to base e of
any
number can be found to
any requisite degree
of
approximation.
310.
Logarithms
to base e are called
Napierian
or
natural
logarithms.
The
logarithms
used in all theoretical
investigations
are
Napierian logarithms
;
but when
approximate
numeri-
cal calculations are made
by
means of
logarithms,
the
logarithms
used are
always
those to base
10,
for reasons
which will
shortly appear:
on this account
logarithms
to base 10 are called Common
logarithms.
We have shewn how
logarithms
to base e can be
found;
and
having
found
logarithms
to base
e,
the
logarithms
to
base 10 are obtained
by multiplying by
the constant
factor
Iog10
e,
or
by l/log.10. [Art.
307,
V.]
This constant
factor is called the Modulus: its value is '43429...
EXAMPLES. 385
EXAMPLES XXXL
1. Shew that
log (x
+
ri)
=
log
x +
log
(l +
-
J
+
+
-
]
-
V
w
-
1 +
a;/
2. Shew that
log. ^12-1
+
(|
+
i)
1
+
(i
+
i)i
4-
IN 1
3. Shew that
log. <s/
10
=
{i
+
39
+
5 <p
+
7 93
+ to
infinity
1+
lK+o7r8+Km
+
?F7v7+
*
infinity
I .
I 19 6 y y 7 y
/ \
"
*
4. Shew that
log.
2
-
=
^-+3--^+^-^+
......
to
infinity.
5. Shew that-
^1^+ g-L-g
+
jJJ-j
+ ...... to in-
finity
=
3
log,
2
-
1 .
6. Shew that
108,
=
x-l I x-l I a*-
7.
Shew that
log,
x
=
-
T H;
-
TTS O/
-
rrs
......
SB + 1 2
(x
+
I)
8
3
(a;
+
I)
3
8. Shew that
a +
a_
2aa; 1 / 2ax \
3
1 / 2aa \
5
g<
^^
=
^T^
5
+
3
V
+
aV
+
5
U
2
+ x
3
)
+ ......
S.A.
25
386 EXAMPLES.
9. Shew that
10. Shew that
fl ,
1
/I 1\
. 1 /I 1 1\
,
1
'-^*--^- +^
+
4(1
+
5+3)*'-...}.
11. Shew
that,
if
log,,
(1
+ x +
x*)
be
expanded
in
powers
1 2
of
x,
the coefficient of a" is either
-
or
--
,
and
distinguish
the
n n
cases.
12. If
loge (1
-
x +
tc*)
be
expanded
in
ascending powers
of
a; in the form
a,a;
+
a
t
x* +
aj?
+ ......
,
then will a
t
+ a
9
+ a
t
+ ...
X + oc + x?
13.
Expand loge
^
in
ascending powers
of ax
~ "*
14. Shew that
1 a; a?* a;*
"r
fl_
a; a;* of
}
'
(n~
jl(+l)
+
|2(n
+
2)
|3(n
+
3)
+
'"/
15. From the
identity
2
log (1 x)
=
log (1
2x +
a^), prove
16. If
log.
r
-
5
-
5
be
expanded
in a series of
positive
*
1-x-x +x?
1 3
integral powers
of
x,
the coefficient ot x" will be or
-
accord-
n n
ing
as n is odd or
eveit,
EXAMPLES. 387
17. Shew that the coefficient of x
r
in the
expansion
of
e"
1
I 1
=
-,
is
{T
+ 2
T
+ 3
r
4- +
n'\.
Hence find the sum of n
l-e
[r
l
terms of the series 1* + 2* + 3
2
+
...,
and also of I
8
+
2
8
+ 3
s
4- ...
18. Shew
that,
if a
r
be the coefficient of x
r
in the ex-
pansion
of ef*
t
then
1 l
r
2
r
3
r
Hence shew that
1 2 ;
[1
+
|2
+
|
and that
I
4
2
4
3
4
I 1 I O ; *i
I 1
19. Shew that
/
w
ro(n-l) n(n-l)(n-2) )
IP
1
2
.2
2
'
1
s
. 2
2
. 3'
"J
-1 .fa. 1 x,
w
,
|~'*
T
*
-
/ T
i
2
9
2
1
2
2* 3
s
20. Shew that the sum of w terms of the series = + -~ +
$
+ . .
.,
I
' '
2> 9
beginning
at the
(n+ l)th,
becomes
equal
to
log,
2 when n is
increased without limit.
21. Shew that
log. (1
+
n)
<
j
+
^
+ +
-
< ! + lo
ge (1
+
w).
22. Prove the
following
:
(i) (*
+
y)
7
-
x
1
-
y
1
=
Ixy (x
+
y)(x*
+
xy
+
y
2
)
2
,
(ii) (x
+
y)
11
-
a:
11
-
y"
=
llxy
(x
+
y} (ar
+
xy
+
y
3
)
(iii) (x
+
y)
13
-
x
13
-y
la
=l
Zxy
(x
+
y) (x*
)
+
2
252
388 COMMON LOGARITHMS.
23. Shew that x** +
y*
+
(x
+
j/)*"
n
(n 3) (n 4) (n
-
5)
_.
.
=
2p"
+ n
(n
-
2) p
"
V
+
^
' v ' v
p" *q*
3.4...2r
where
p
=
a? +
xy
+
y
3
and
q
=
xy
(x
+
y).
24. Shew
that,
(i)
if n be
any
uneven
integer, (6 c)"
+
(c
-
a)'
+
(a
-
b)"
will be divisible
by (b
-
c)
8
+
(c
-
a)
8
+
(a
-
&)';
(ii)
if n be of the form 6m
1,
it will be also divisible
by
(b
c)*
+
(c a)
2
+
(a
-
b)' ;
and
(iii)
if n be of the form
6wi + 1 it will be divisible
by (b c)*
+
(c
-
a)*
+
(a b)*.
COMMON LOGARITHMS.
311. In what follows the
logarithms
must
always
be
supposed
to be common
logarithms,
and the
base, 10,
need
not be written.
If two numbers have the same
figures,
and therefore
differ
only
in the
position
of the decimal
point,
the one
must be the
product
of the other and some
integral power
of
10,
and hence from Art.
307,
II. the
logarithms
of the
numbers will differ
by
an
integer.
Thus
log
421-5
=
log
4'215
+
log
100
-
2
+
log
4'215.
Again, knowing
that
log
2
=
"30103,
we have
log
'02
=
log (2
+
100)
=
log
2
-
log
100
=
-30103
-
2.
On account of the above
property,
common
logarithms
are
always
written with the decimal
part positive.
_
Thus
log
-02 is not written in the form
-
1 '69897 but
2'30103,
the minus
sign referring only
to the
integral portion
of
the
logarithm
and
being
written above the
figure
to which
it refers.
Definition. When a
logarithm
is so written that its
decimal
part
is
positive,
the decimal
part
of the
logarithm
is called the mantissa and the
integral part
the character-
istic.
CHARACTERISTICS FOUND BY INSPECTION.
389
312. The characteristic
of
the
logarithm of any
number
can be written down
by inspection.
For,
if the number be
greater
than
1,
and n be the number of
figures
in its
integral part,
the number is
clearly
less than 10" but not
less than 10""
1
.
Hence its
logarithm
is between n and n 1 : the
logarithm
is therefore
equal
to n I
+
a decimal.
Thus the characteristic
of
the
logarithm of any
number
greater
than
unity
is one less than the number
offigures
in
its
integral part.
Next,
let the number be less than
unity.
Express
the number as a
decimal,
and let n be the
number of
ciphers
before its first
significant figure.
Then the number is
greater
than 10"""
1
and less than
1(T".
Hence,
as the decimal
part
of the
logarithm
must be
positive,
the
logarithm
of the number will be
(n
+
1)
+
a decimal
fraction,
the characteristic
being (n
+
1).
Thus,
if
a number less than
unity
be
expressed
as a
decimal,
the characteristic
of
its
logarithm
is
negative
and
one more than the number
of ciphers before
the
first signifi-
cant
figure.
For
example,
the characteristic of the
logarithm
of 3571'4 is
3,
and that of -00035714 is 4.
Conversely,
if we know the characteristic of the
logarithm
of
any
number whose
digits
form a certain
sequence
of
figures
we know at once where to
place
the
decimal
point.
For
example, knowing
that the
logarithm
of a number whose
digits
form the
sequence
35714
is
3-55283,
we know that the number
must be 3571 '4.
313. Tables are
published
which
give
the
logarithms
of all numbers from 1 to 99999 calculated to seven
places
of decimals : these are called
'
seven-figure
'
logarithms.
For
many purposes
it is however sufficient to use five-
figure logarithms.
390 USE OF TABLES OF LOGARITHMS.
In all Tables of
logarithms
the mantissae
only
are
given,
for the characteristics can
always,
as we have
seen,
be written down
by inspection.
In
making
use of Tables of
logarithms
we
have,
I. to
find the
logarithm
of a
given
number,
and II. to find the
number which has a
given logarithm.
I. To
find
the
logarithm of
a
given
number.
If the number have no more than five
significant
figures,
its
logarithm
will be
given
in the tables.
But,
if
the number have more
significant figures
than are
given
in the
tables,
use must be made of the
principle
that
when the difference of two numbers is small
compared
with either of
them,
the difference of the numbers is
ap-
proximately proportional
to the difference of their
loga-
rithms. This follows at once from Art.
308,
for
Iog10 (N
+
x)
-
Iog10
N
=
lo
glo
l
+
j
=
p log.
l
+
=
A*
Hf"
^
Jp
+
'
)
=fj
*jj-
approximately,
when
~
is
small,
p being
the modulus
l/log
e
10.
An
example
will shew how the above
principle,
called
the
Principle of Proportional Differences,
is utilised.
Ex. To find the
logarithm
of 357-247.
We find from the tables that
log
3-5724=
-5529601,
and
log
3-5725
=
5529722;
and the difference of these
logarithms
is -0000121.
Now the difference between 3-57247 and 3-5724 is
Jyths.
of the
difference between 3*5724 and 3-5725
;
and hence if we add
-fyths.
of
0000121 to the
logarithm
of 3-5724 we shall obtain the
approximate
logarithm
of 3-57247. Now
&tht.
of -0000121 is
-00000847,
which
is nearer to -0000085 than to -0000084. Hence the nearest
approxi-
mation we can find to the
logarithm
of 3-57247
is
-6529601+ -0000085
=
5529686.
The characteristic of the
logarithm
of 357 '247 is
obviously 2,
and
therefore the
logarithm required
is 2*5529686.
II. To
find
the number which Jias a
given logarithm.
For
example,
let the
given logarithm
be 4*5529652.
We find from the tables that
log
3-5724
=
-5529601 and that
log
3-5725
=
-5529722,
the mantissa of the
given logarithm falling
COMPOUND INTEREST AND ANNUITIES. 391
between these two. Now the difference between -5529601 and the
51
given logarithm
is -? of the difference between the
logarithms
of
J-^J.
3-5724 and 3-5725
;
and
hence, by
the
principle
of
proportional
differences,
the number whose
logarithm
is -5529G52 is
3-5724 +
.^-
x -0001 =3-5724 +-00004=3-57244.
\.a J.
[The approximation
could
only
be relied
upon
for one
figure.]
Thus -5529652=
log 3-57244,
and therefore
4'5529652=
log
-000357244.
COMPOUND
INTEREST AND ANNUITIES.
314. The
approximate
calculation of
Compound
In-
terest for a
long period,
and also of the value of an
annuity,
can be
readily
made
by
means of
logarithms.
All
problems
of this kind
depend upon
the three fol-
lowing
:
[The
student is
supposed
to be
acquainted
with
the arithmetical treatment of these
subjects.]
I. To
find
the amount
of
a
given
sum at
compound
interest,
in a
given
number
of years
and at a
given
rate
per
cent,
per
annum.
Let P denote the
principal,
n the number of
years,
lOOr the rate
per
cent,
per
annum,
and A the
required
amount.
Then the interest of P for one
year
will be
Pr,
and
therefore the amount of
principal
and interest at the end
of the first
year
will be P
(1
+
r).
This last sum is the
capital
on which interest is to be
paid
for the second
year;
and therefore the amount at the end of the second
year
will be
{P (1
+
r)} (1
+
r)
=
P
(1
+
rf. Similarly
the
amount at the end of n
years
will be P
(1
-f
r)
n
.
Thus A
=
P
(1
+
r)
n
;
and hence
log
A
=
log
P
+ n
log (1
-I-
r).
If the interest is
paid,
and
capitalised,
half
yearly,
it
(T*
N
1
+
-
z
392 ANNUITIES.
Ex. Find the amount
of
350 in 25
years
at 5
per
cent,
per
annum.
Here
P=350,
r=^
and ;i
=
25;
f.
log ^=log
350+25
log
1
+
=
log
350 +
25
(log
105
-
log 100).
From the tables we find that
log
350
=
2-5440680 and
log
105
=
2-0211893;
hence
log
^=3-0738005. Whence it is found from the
tables that A
=
1185-22.
II. To
find
the
present
value
of
a sum
of money
which
is to be
paid
at the end
of
a
given
time.
Let A be the sum
payable
at the end of n
years,
and
let P be its
present
worth,
the interest on
money being
supposed
to be lOOr
per
cent,
per
annum. Then the
amount of P in n
years
at lOOr
per
cent
per
annum
must be
just equal
to A.
Hence from I. P
=
A
(1
+
r)-.
III. To
find
the
present
value
of
an
annuity of
A
payable
at the end
of
each
of
n successive
years.
If the interest on
money
be
supposed
to be lOOr
per
cent,
per
annum
;
then from II.
The
present
value of the first
payment
is A
(1
+
r)"
1
second A
(1
+
r)~*
nth
4(l+r)~".
Hence the
present
value of the whole is
Ex. l-'ind the
present
value of an
annuity
of 30 to be
paid
for 20
years, reckoning
interest at 4
per
cent
EXAMPLES.
393
(
/25N
20
]
Hence the
present
value
=
30 x 25
-jl
-
(
^r.
1 V.
Now
log
{log
25
-log 26}
=
20
{1-3979400- 1-4149733}
=
20(
-
-0170333)=
-
340666
=
1-659334
=log -456389,
from the Tables.
Hence the value
required=30
x 25 x
(1
-
-456389)
=
407 7...
EXAMPLES XXXIL
The
following logarithms
are
given
log
1-02
=
-0086002
log
1-6386
=
-2144730
log
1-025
=
-0107239
log
1-6387
=
-2144995
log
1-033
=
-0141003
log
1-7292
=
-2378452
log
1-04
=
-0170333
log
1-7349
=
-2392744
log
1-05 =-0211893
log
2 =-3010300
log
1-06
=
-0253059
log
2-0829
=
-3186684
log
1-1467
=
-0594498
log
3
=
-4771213
log
1-1468
=
-0594877
log
3-0832
=
-4890017
log
1-2258
=
-0884196
log
4-4230
=
-6457169
log
1-2620
=
-1010594
log
5-1
=
-7075702
log
1-4816
=
-1707310
log
5-577
=
-7464006
log
1-4817
=
-1707603
log
6-3862
=
-8052425
log
7-4297
=
-8709713
log
7-4298
=
-8709771
1. Find
jyi05.
2. Find
jy51.
3. Find the amount of 100 in 50
years
at 5
per
ceut
per
annum.
4. Shew that
money
will more than doable itself in 16
years
at 5
per
cent,
per
annum,
and in 18
years
at 4
per
cent.
per
annum.
394 EXAMPLES.
5. Find the amount of 500 in 10
years,
interest at 4
per
cent,
being paid
half
yearly.
6. The number of births in a certain
country every year
is 85
per
1000 and the number of deaths 52
per
1000 of the
population
at the
beginning
of
every year:
shew that the
popu-
lation will be more than doubled in 22
years.
7. A man invests 30 a
year
in a
Savings
Bank which
pays 2J
per
cent,
per
annum on all
deposits.
What will be
the total amount at the end of 20
years
?
8. What sum should be
paid
for an
annuity
of 100 a
year
to be
paid
for 40
years, money being supposed
to be worth
4
per
cent,
per
annum 1
9. A
corporation
borrows 30000 which is to be
repaid
by
30
equal yearly payments.
How much will have to be
paid
each
year, money being supposed
to be worth 4
per
cent,
per
annum]
10. A house which is
really
worth 70 a
year
is let on a
lease for 40
years
at a rent of 10 a
year,
the lease
being
re-
newable at the end of
every
14
years
on
payment
of a fine.
Calculate the amount of the
fine,
reckoning
interest at 6
per
cent.
CHAPTER XXV.
SUMMATION OF SERIES.
315. WE have
already
considered some
important
classes of
series,
namely
the
Progressions [Chapter xvn],
Binomial series
[Art. 288],
and
Exponential
and
Logarith-
mic series
[Chapter xxiv].
In the
present chapter
some
other
important types
of series will be considered.
316. The nth term of a series will be denoted
by
u
n
,
and the sum of n terms
by
S
n
. When the series is con-
vergent
its sum to
infinity
will be denoted
by
S^.
317. No
general
method can be
given by
which the
summation of series can be
effected;
but in a
great
number of cases the result can be obtained
by expressing
the
general
term of the
series,
u
n ,
as the difference of
two
expressions
one of which involves n 1 in the same
manner as the other involves n.
For
example,
in the series
"-
q a
j
+
(x
+
a) (x
+
2a)
+
(x
+
2o) (x
+
3a)
'
the nth
term,
namely
=
,
is
equal
to
(x +
n I.
a) (x+na)
,
=-r . Hence the series
may
be written
tc
+
(n l)a
x
+
na
396 SUMMATION OF SERIES.
_ -
x+a) \x+a x+2aj \a:
+
2a # +
+
!
-
7
-
TT
---
> : and it is now obvious that all
\x+(n I)
a a;
-f
no,)
the terms cancel
except
the first and last
;
hence S
n
=
x x
+
no,
x(x
+
no)
'
Ex. 1. Find the sum of n terms of the series
i
Ex. 2. Find the sum of n terms of the series
123 n
Am.
1--=.
n
+
1
11
1
Herew_=i
--
;
-
r-
Ans- i
~iTt*
n
n
n+1
n+1
Ex. 3. Find the sum to
infinity
of the series
J_ J_ J_
1
3|1
+
4|2
+
5 |3
11
1
Here
=
, ,
-
i s
Ans. =.
*
n+1 n+2
2
Ex. 4. Find the sum to
infinity
of the series
3 6 7 2n+l
Here _=--. -r . ^ls. 1.
1
n*
Ex. 5. Find the sum of n terms of the secies
_1
2 3 n
1. 3*1. 8. 6*1. 8. 5. 7*
+
1.3.6...(2n
+
l)*
1 1
~
1 . 3 .
6Z(2n
-
1)
~
1 . 3 .
5...(2n
-
1) (2nH
Ans- ^
SUMMATION OF SERIES. 397
Ex. 6. Sum to
infinity
the series
2 1 3
1
+
JL 1 4.
w+1
1.3 3^3. 5
'
3
2
5. 7
'
3
T ' T
(2- l)(2n+ 1)
3
2n-l
11
in
i
4"
Ex. 7. Find the sum to
infinity
of the series
22Ti
+
pTl
+
65TI
+
8^I
+
Am'
2'
Ex. 8. Find the sum of n terms of the series
*
2
*
3
(1-aou-^) (i-^)(i-*v (i-s'Mi-sV'
i i
.4ns.
318. To
find
the sum
ofn
terms
of
the series
{a (a
+
6)...(a
+ ^T.
b)}
+
{(a
+
b) (a
+
26)... (a
+
rb)}
+
...
+
{(a
+
n-
1.6)(a+n6)
...
(a+n
+
r-2
.6)}
+
...
In the above series
(i)
each term contains r
factors,
(ii)
the factors of
any
term are in arithmetical
progression,
and
(iii)
the first factors of the successive terms form the
same A.P. as the successive factors of the first term.
Consider the series which is formed
according
to the
same law but with one factor added at the end of
every
term,
and let v
n
be the nth term of this new
series,
so that
v
n
=
{(a
+
n-l.b)(a
+
nb)
...
(a
+
n
+
r 1 .
6)}.
Then
-
{(a
+
n
-
2 .
b) (a
+
n
-
I .
b)
. . .
(a
+
n
+
r
-
2 .
6)}
=
{(a
+
n-1 .
6)
. . .
(a
+ n~+r-2
.
b)}{(a+
n+r-1
.
6)
-
(a
+ n-2 .
b)}
J
=(r+l)6{(a+/7^T.6)...(a
+
n
+
r-2.6)}.
398
SUMMATION OF SERIES.
Hence v
m
v
n
_
l
=
(r+ 1)6
x u
n
.
Changing
n into w 1 we have in succession
Also
v,
1>
=
(r
+
1)
6 x w
t
,
where v is the term
preceding
v
t
which is formed accord-
ing
to the same
law,
that is v
c
=
{(a 6)
a
(a
+
&)...(a
+
r 1
&)},
so that f> is
obtained
by putting
n
=
in the
expression
for v
m
.
Hence
by
addition
Ex. 1. Sum the series 1.2
+
2.3 +
3.4+ ......
+n(n+l).
Here
K,,
=
n(n+l),
=
n(n+l)(n+2),
r
=
0. 1.
2,
r
=
2,
and
6=1.
Hence S
n
=
5 t(n+l) (n+2).
o
Or,
by using
the above method without
quoting
the
result,
which
is
preferable
in
very simple cases,
we have
n(n+l)=i{n(n+l)(n
+
2)-(n-
O
(n-l)n=i{(n-l)n(n
+
l)-(n-
Hence S=
3
n(n
+
l)(n+2).
O
Ex.
2. Sum the series 1.2. 3
+2.3.4 +
+n(n+l) (n
+
2).
Ant.
-n(n+l)(n
+
2)(n
+
3).
SUMMATION OF
SERIES. 399
Ex. 3. Sum the series
1.2.3.4 + 2.3.4.5 + ...... +
n(n
+
l)
(n
+
2) (n
+
3).
Ans.
n(n+l)(n
+
2)(n
+
3)(n+4).
Ex. 4. Find the sum of n terms of the series
3.5.7 + 5.7.9 + 7.9.11 + ......
Here
ft =(2n
+
l)(2n+3)(2
+
5),
v
n
=(2n
+
l) (2n
+
3) (2n
v
=1.3.5.7, r=3,
and &
=
2.
Many
series which are not of the
requisite
form can be
expressed
as the
algebraic
sum of a number of series
which are all of the
required
form;
and the sum of the
given
series can then be written down. The
following
are
examples.
Ex. 5. Find the sum of n terms of the series 1.3
+
2.4
+ 3.5 + ......
Here
=
(n
+
2)=n (n
+
l)+n.
The sum of the series 1.2
+ 2.3+... +
n(n
+
l)
is
and the sum of the series 1
+2+ ... +n
is
5 {n(n
+
l)
-0 .
1}.
m
Hence the
required
sum
is-(n
+
l) (n
+
2)
+
s
n
(n
+
1). "
o
Ex. 6. Find the sum of the series
2.3. 1
+3.4.4+4. 5.7+ ...... +
(n
+
l) (n+2) (3/1-2).
Here
B =(n+l) (+2) (3ra-2)
=
3n(n+l) (n
+
2)
-2
(n+1) (n
+
2).
400 SUMMATION OF SERIES.
319. To
find
the sum
of
n terms
of
the series whose
general
term is
I/{(a
+
n^l.
b) (a
+
nb) (a
+
n
+
1 .
6)... (a
+
n
+
r-2 .
&)}.
Consider the series which is formed
according
to the
same law but with one factor taken
away
from the
beginning
of each
term,
and let v
%
be the nth term of this
second
series,
so that v
n
=
1
/ {(a
+
nb)
. . .
(a
+
n
+
r 2 .
6)}.
Then
1
{(a
-i-
nb)
. . .
(a
-f
n
+
r
-
2 .
6)]
1
{(a
+
n
-
1 .
b) (a
+
nb).
.
.(a
+
n
+
r
-
3 .
b)}
{(a
+
n- 1 .
6).
.
.(a+
w+r-2
.
b)}
Changing
n into n 1 we have in succession
x w
v
2
fl
t
=
(r
-
1)
b x u
r
Also
Wj
v
=
(r 1)
6 x u
lt
where v is the term which
precedes
v
l
and which is formed
according
to the same
law,
that is
t>
=
1
/ {a (a
+
6)...(a
+
r^ 2 .
b)}.
Hence,
by
addition,
SUMMATION OF SERIES.
401
1 1
Mere u
m , . , , /.. . v
"
. n i *o n
1 l
1)1
1
Hence S
n
=
o
Ex. 2. Sum the series
+ _ + ...
+
-
1.2.3.4 2.3.4.0 i
to n terms and to
infinity.
1
+
3)
'
Hence
S.=
^j
{1^73
~
(
w
+
1) (n
+
2) (n
+
3)}
and
3.7.11 7.11.15
(4w-l)(4n
+
3)(4n+7)'
1 1
Ex. 3. Sum the series
,,-^n-
+
,, ,
+ ...+-
\47i
-
iji4
+
a; i*n4
-
j
! *
-<4nS. S
=
{ }
.
Many
series which are not of the above form can be
expressed
as the
algebraic
sum of a number of series
which are all of the
required
form;
and the sum of the
series can then be written down. The
following
are
examples.
Ex. 4. Sum the series z
^
+
^-7
+
5=
+ ...
Here
1
n+1
1 1
n(n+2) n(n+l)(n
+
2) (n+l)(n
+
2)
T
B
(n
+
l) (n
+
2)
*
The series whose
general
terms are -.
^-.
r-
and
;
=
s-
(n+l)(n
+
2) n(n
+
l)(?i
+
2)
are of the
required
form. Hence the sum of the
given
series is
given by
s
_
j
_
-
j
+
1 (_
i
\=~-
2n+i
2
\1
. 2
(n+l)(n
+
2)/
4 2
(n
+
1) (9
2/t+3
2
(n
+
1) (n
+
2)-
S.A.
26
402 SUMMATION OF SERIES.
1 1
Ex. 5. Sum the series
1.3.5 2.4.6
n(n+2)(n
+
4)'
1
(n
+
l)(n+3)
"n
(n
+
2) (n
+
4)
n
(n
+
1) (n
+
2) (n
+
3) (n
+
4)
n+3)(n-
+
n(n+l)(n
Hence
Q
-
1
f
* 1
I
,
3
\
l 1
I
"
2(2.3 (n
+
2)(n
+
3)J
4(1.2.3.4 (n
+
l)(n
+
2)(n
+
3)(n+
4)(
'
320. The sum of series of the kind
just
considered
may
be obtained
by
means of
partial
fractions.
The method will be seen from the
following example.
To find the sum of the series =
^
+
rr -.
+
-
-+ ... +
A. B
Let
^
=
+
s
; then,
as in
Chapter nun,
we find
that A
=
\
and B=
-5.
a u
_ 11
Hence 2u
n
=
.
1
n n
+
2
We have therefore the
following
series of
equations
:
_1_1
2
_1_1
2
_1
1
13 24 35
11 11 11
2u__.=
j. , 2u__,
=
=
T,
and 2u_
=
r.
1
n-2 n
l
n-1
n+1'
'
n
n+2
Hence, by addition,
the other terms all
cancelling.
3 2n
+
3
Hence S
n
=-
-
^
321. To
find
the sum
of
the rth
powers
of
the
first
n whole numbers.
We will first consider the two
simplest
cases.
SUM OF
SQUARES
AND
CUBES. 403
Case I. To find the sum of I
2
+
2*
+
3
2
+
...
+ n*.
Here u
u
= n*
=
n
(n+ 1)
n.
Hence,
by
Art.
318,
--
Case II. To find the sum of 1
s
+
2"
+
3'
+
. . .
+
n'.
Here u
n
= n*
=
n
(n
+
1) (n
+
2)
3n* 2n
=
n(n
+
l)(n
+
2)-3n(n
+
I)+
n.
Hence,
by
Art.
318,
m
+
-
-
Since 1
+
2
+
...
+n
=
5 n(n
the above result shews that
so that the sum
of
the cubes
of
the
first
n whole numbers is
equal
to the
square of
the sum
of
the numbers.
The sum of the cubes of the first n
integers
can also be
easily
found
by
means of the
identity
4n
s =
{n (n
+
1) }
3
-
{(n
-
1) n}
3
.
For we have in succession
4n
3
=
{n
(n
+
1)}
2
-
{(n
-
1) n}
2
,
4.2
3
=(2.3)-(1.2)
2
,
and 4 . 1
s
=
(1
.
2)
s
-
(0
.
1)
8
.
Hence, by
addition,
4/S
w
=n
2
(n+l)
2
.
262
404 SUM OF POWERS OF INTEGERS.
Case III.' To find the sum of 1"
+
2
r
+ 3
r
+
. . .
+
n
r
.
The sum for
any particular
value of r can be found
by
the same method as that
adopted
for the values 2 and 3.
For
example,
the sum of the fourth
powers
can be
written down as soon as n* is
expressed
in the form
+ 7n
(n
+
1)
-
n.
By
means of the Binomial Theorem a formula can be
found which
gives
the sum of the rth
powers
in terms of
the sum of
powers
lower than the rth
;
and this formula
can be used for
finding
the sum of the
2nd, 3rd, 4th,
&c.
powers
in succession. The formula has however the
great
disadvantage
that in order to find
by
means of it the sum
of the rth
powers,
it is
necessary
to know the sums of all
the
powers
lower than the rth.
By
the Binomial
Theorem,
we have in succession
(
r
+
1) (n
-
l)
r
+
(n
-
1)'"
1
JL X
Hence,
by
addition,
we have
(n
+
I)*"*
1
(+!)
where
S is written for the sum of n terms of the series
PILES OF SHOT. 405
We can in a similar manner find a formula for
summing
the
rth
powers
of
any
series of
quantities a,
a
+
b,
a +
2&,
...
in arith-
metical
progression.
The result is
where S
n
322. Piles of Shot. To
find
the number
of spherical
balls in a
pyramidal heap,
when the base is
(I)
an
equilateral
triangle, (II)
a
square,
and
(III)
a,
rectangle.
I. In a
pile
of this kind the balls which rest on the
ground
form an
equilateral triangle,
and
upon
this first
layer
a number of balls are
placed forming
another
equi-
lateral
triangle having
one ball fewer in each side than in
the side of the base
;
and so on
;
a
single
ball
being
at the
top.
If n be the number of balls in each side of the
base,
the total number in the base will be
that is
%n (n
+
1).
The whole number of the balls in the
pile
will therefore be
{n(n
+
l)
+
(w-l)n+...+
1.2},
that is TI
(n
+
1) (n
+
2).
II. In this case the balls in
any layer
form a
square
with one ball fewer in each side than in the
layer
next
below. Hence if n be the number of balls in each side of
the lowest
layer,
w
2
will be the number of balls in the
base,
and therefore the whole number of the balls will be
n
1
+
(n
-
1)*
+
(n
-
2)
2
+.
.
.+
1*,
that is
%n (n
+
1) (2
+
1).
III. In this case the balls in
any layer
form a
rectangle
with one ball fewer in each side than in the
layer
next below. Hence if n and m be the number of balls
in the sides of the lowest
layer,
nm will be the number of
balls in the base and therefore the whole number of the
balls will
be,
n
being greater
than
m,
nm
+
(n 1) (m
-
l)
+
(n 2) (m
-
2)
+
...
(n-m
+
1)
1
=(n
m
+
m)
m
+
(n
m
+
m
l)(m 1)+.
.
.(n
m
+
1)
1
406
FIGURATE NUMBERS.
=
(n
-
m)
[m
+
(m
-
1)
+...+
1}
+ m
2
+
(m
-
1)
8
+...+
I'
2
-
\ (n
-
m)
m
(m
+
1)
+
w
(m
+
1) (2m
+
1)
=
%m (m
+
1) (3n
-
m +
1).
Ex. 1. How
many
balls are contained in 8
layers
of an unfinished
triangular pile,
the number in one side of the base
being
12 ?
If the
pile
were
completed
it would contain
,.
12.13.14
balls;
and there are -.4.5.6
missing
from the
complete pile;
hence the
required
number is
^
(12
. 13 . 14
-
4 . 5 .
6).
Ex. 2. How
many
balls are contained in 10
layers
of an
incomplete
pile
of balls whose base is a
rectangle
with 20 and 25 balls in its
sides?
The number=2n
(n
+
5)
from n=ll to
=
20.
Ans. 3260.
323.
Figurate
numbers. Series of numbers which
are such that the nth term of
any
series is the sum of the
first n terms of the
preceding
series,
all the numbers of
the first series
being unity,
are called orders of
figurate
numbers.
Thus the different orders of
figurate
numbers are:
First
order, 1, 1, 1, 1, 1,
Second
order, 1, 2, 3, 4, 5,
Third
order, 1, 3, 6, 10, 15,
It follows from the definition that the nth term of the
second order of
figurate
numbers is
n;
the nth term of
the third order will therefore be
(1
-f
2
+
3
+
...
+
n),
that is
^n (n
+
1);
the nth term of the
fourth
order will therefore be
the nth term of the
fifth
order will therefore be
-
{n
(n+
1) (n
+
2)
+
(n-l)n(n
+
!)+...+
!. 2.
3},
that
- . >
POLYGONAL NUMBERS. 407
is
-- -
n
(n
+
1) (n
+
2) (n
+
3);
and so
on,
the nth term
2 . o . 4
of the
rth order
being
324.
Polygonal
numbers.
Consider the arithmetical
progressions
whose first two terms are
respectively
1,
1
;
1,
2
; 1,
3
; 1,
4
;
and so on. Then the series formed
by
taking
1, 2, 3,...,
n of the terms of these different
arith-
metical
progressions, namely
the series
1, 2, 3, , n,
1, 3, 6, ,
Jn(n
+
l),
1,4,
9, ,
w
2
,
1, 5, 12, ,
n
+
f(n-l),
1, r,
3r
3, ...,
n
+
^n(n 1) (r 2),
...
are called series of
linear,
triangular, square, pentagonal,.
.
r-gonal
numbers.
The sum of n terms of a series of
r-gonal
numbers
can be written down at
once,
for the sum of n terms of the
series whose
general
term is n
+
\n (n 1) (r 2)
is
\n (n
+
1)
+
(n
-
1)
n
(n
+
1) (r
-
2) [Art. 318].
EXAMPLES XXXIII.
Find the sum of n terms of each of the
following series,
and find also the sum to
infinity
when the series is
convergent.
1. 4.7.10 + 7.10.13 + 10. 13.16 + ...
1 1
_J
*
3.7.11
+
7. 11.15
+
11. 15. 19
+
**'
3. 1.3.
4 + 2. 4. 5 + 3. 5.6 + ...
4. 1.5 + 3.7 + 5.9 + 7.11 + ..
408
EXAMPLES.
5.
EXAMPLES. 409
24. Find the sum of the series n . n+
(n 1) (n
+
1)
+
(n- 2) (n
+
2)
+ ... +
2
(2n- 2)
+ 1 .
(2n
-
1).
25.
Find the sum of n terms of the series
ab +
(a
-
1) (b
-
1)
+
(a
-
2) (b
-
2)
+
...
26. Prove
that,
if S
u
'
=
l
r
+ 2" + ... n
r
\
then will
(i)
&s
n
*
=
6s
u
*
*
s;
-
s
n
*.
(ii)
8
u
'
+
S;-2(Sn*r.
27. Find the sum of the
following
series to n terms :
2 + 3
"
+ 2> + -
3 1 4 1 5 1
W
i.22
+
2.32*
+
3.42
3
+ "*
.....
4 /2\ 5 /2\ 6 /2\
<
1U
>
1T2(3)
+
2T3(3)
+
374(3)
+-
8 /5\ 9 /5\ 10 /5
8
9 /3\ 10
/3V
11
mw"
l
"O~4w
+
37TT5
H
15 /6\ 16 /6\' 17 /6 16 /6\'
2T374
(l)
28. Shew that the sum of all the
products
of the first n
natural numbers two
together
is
777 (w 1)
n
(n
+
1) (3w
+
2).
z*
29. Shew that the sum of all the
products
of the first n
natural numbers three
together
is
j^
(n
-
2) (n 1)
n*
(n
+
1)*.
30. Shew that the sum of the
products
of
every pair
of the
squares
of the first n whole numbers is
410
SUMMATION OF
SERIES.
325. To find the sum of n terms of the series
b
^
b(b+xyb(b+x)(b
+
2xy
a(a
+
.
6
(6
+
a?)... (6
+
n-
la;)
In the above series there is an additional factor both
in the numerator and in the denominator for
every
succes-
sive
term,
and the successive factors of the numerator and
denominator form two arithmetical
progressions
with the
same common
difference.
Consider the series formed
according
to the same law
but with an additional factor in the
numerator,
and let v
n
be the
general
term of this second
series,
so that
_
a
(a
+
x).
.
.(a
+
n 1
x) (a
4- nx)
Then
a
(a
+
x).
.
.(a
+
n 1
x} (a
+
nx)
'-
^
-
a(a
+
x)...(a
+
n
lx)
a(a
+
x)...(a+
n
lx)
,(b
+
n
-
1
x)
.'. tL
__.
=
M.
It- n
1
*
So also v
n
_
l v^j
=
u
n_
l
x
(a
-f
x
I
)
(a 4-x)
Also t>
t
=
a
r
-
=
(a
+
&
v
t
v
l
=
u
t
x
(a
+
x
6).
=
(a
+
&)
^
=
M,
x
(a
+
x
b)
+
&M,.
SUMMATION OF SERIES.
411
Hence S
n
x
(a
+ x
b)
=
v
n
a
;
a
f
(a
+
a")...(g
+
na;)
1
a
+
x-b\b(b
+
ai)...(b
+
:
^lx)
J*
The sum of n terms of the series
a
a(a x)
a,
(a x) (a
b
~
b
(b
+
x)
b
(b
+
x) (b
+
2*)
in which the successive factors of the numerator and
denominator form two arithmetical
progressions
whose
common differences are
equal
in
magnitude
but of
opposite
sign,
can be found
by changing
the
sign
of a in the
previous
result: the sum
can, however,
be obtained inde-
pendently by
the same method. Thus
a
_
1
|~a
a
(a x)~\
b
~
a
+
b-x
[1
4
~T>
J
_
a
(a x)
_
I Fa
(a
-
x) a(a x)(a
2
b(b
+
x)~
a
+
b
x\_
b b
(6
+
x)
...
(b
+
n
-lx}
Hence
S
_ ._
-
._.,
I \a
(a
x)...(a
n\
IK)
a
+
b-x
[b
(b
+
x).
.
.(b
-f
n-2x)
a(a x)...(a nx)
~|
b(b
+
x)...(b+n-l x)\
'
a+b-x
Ex. 1. To find the sum of n terms of the
series
-
-f-
2
'-
\
2 - 5 - 8
.
3
3.6
3.6.9
412 SUMMATION OF SERIES.
We have
?-V ?L\
2.5_l/2.5.8
2
-5\
3~2\
3
"I/
1
3.6~2~V
3.6
"
3
/
"
2.5.8...(8n-l)_l
(2.5.8...(3n+2)
_
2. 5.
8...(3n-l)l
3.6.9...3n
~2
t
3.6.9...3n 3 .6.
9...(3n-3)J
*
1
J2.5.8...(3n+2)
2
HeDCe g
-al
3.6.9...3n
[This particular
series is a binomial
series,
the successive terms
being
the coefficients of
x,
x
a
, &c.,
in the
expansion
of
(1-ar)
. Hence
[Art. 287]
l
+ S
n
=sum of the first
(n
+
1)
coefficients in the
expansion
of
(1
-
x)~$=
coefficient of x* in
(l-x)~^x (I- x)'
1
,
that is in
(1-z)"^].
2 26 26 10
Ex. 2. Find the sum of n terms of the series
-
+
^-
+ ''+...
o o . 7 O.7.
11
2 (6.10...
Ex. 3. Find the sum of n terms of the series
m
(m 1) m(m 1) (m 2)
1.2 1.2.3
Am.
(-1)*-
1
326. The sum of n
+
1 terms of the series
a
where
a,
is
any integral expression
of the rth
degree
in
n,
can be found in the
following
manner.
S
=
a
Hence S
m
x
(1
-
a;)'*
1
=
a
+
{at
-
(r
+
1)
a
}
x
+...
SUMMATION OF
SERIES.
413
f
Now a
p
is
by supposition
an
integral expression
of the
rth
degree
in
p
;
hence
where A
r
,
A
r
_
l
,...,
A do not contain
p.
Also,
by
Art.
305,
the sum of the series
p*
_
(
r
+
1) ( p
-
i)
+
L(p
_
2
)_...
to
(r+2)
terms,
is zero for all
integral
values of k less than r
+
1. Hence
a
p
-(r
+
l^a^
+
~^d^-
... to
(r
+
2)
terms
is zero for all values of
p.
All the terms of the
product
S
n
x
(1 a;)*"*
1
will there-
fore vanish
except
those near the
beginning,
or the
end,
for which the series a
p (r+l)ap
_
1
+...
is not continued
for
(r
+
2)
terms,
that is all the terms of the
product
will
vanish
except
the first r
+
1 terms and the last r
+
1 terms.
Hence
8
U
x
(1 -asf"
=
a
+
{at
-(r
+
1) aj
as
+...
whence the value of S
n
is found.
Ex. 1. Find the sum of the series
[all
the other terms
vanishing
on account of the
identity
fc-2
=1
-
(n
+
2)
x
n+*
. o
1
_
n
414 SERIES WHOSE LAW IS NOT
GIVEN.
Ex. 2. Find the sum of n
+
1 terms of the series
.2-4.1*)ar
s
[The
other terms all
vanishing,
since
fc
8 -
4
(
k
-
1)
8
+
6
(
k
-
2)
-
4
(*
-
3)
8
+
(k
-
4)
8
==
identically.]
Hence
S^
=
[1
+
4x
+
z
2
-
(n
8
+
6n* + 12n
+
8)
x
n+1
+
(3n
8
-f
15n
a
+
21n +
5)
*
-
(3n
8
4-
12n
2
+ 12n +
4)
*+
+
(n
+
1)
8
*+*] /(I-*)
4
.
When x is
numerically
less than
1,
the series is
convergent,
and the
urn of the series continued to
infinity
is
(1
+
4x
+
x
2
) / (1
-
x)
4
.
327. Series whose law is not
given.
We have
hitherto considered series in which the
general
term was
given,
or in which the law of the series was obvious on
inspection.
We
proceed
to consider cases in which the
law of the series is not
given.
With reference to series
in which the law is not
given,
but
only
a certain number of
the terms of the
series,
it is of
importance
to remark that
in no case can the actual law
of
the series be
really
deter-
mined: all that can be done is to find the
simplest
law the
few terms which are
given
will
obey.
There are for instance an indefinite number of series
whose first few terms are
given by
x
+
a?
+
a?
+
...,
the
simplest
of all the series
being
the
geometrical progression
whose nth term is #" : another series which has the
given
.
.,
,
f
.
c
terms is that formed
by
the
expansion
of
METHOD OF
DIFFERENCES. 415
which
agrees
with the
geometrical progression except
at
every
10th term.
Note. In what follows it must be understood that
by
the law of a series is meant the
simplest
law which
satisfies the
given
conditions.
METHOD OF
DIFFERENCES.
328. If in
any
arithmetical series
each term be taken from the
succeeding
term,
a new
series is
formed,
namely
the series
(a
s
-
a
t
)
+
(a,
-
cO
+.
.
.+
(a.
-
a
n_,)
+.
...
which is called the
first
order
of differences.
If the new series be
operated upon
in the same
way,
the series obtained is called the second order
of differ-
ences. And so forth.
Thus,
for the series
2, 7, 15, 26, 40,
...
,
the first order of differences is
5, 8, 11, 14, ...,
and the second order of differences is
3, 3, 3,
...
329. When the law of a series is hot
given,
it can often
be found
by forming
the series of successive orders of
differences
;
if the law of one of these orders of differences
can be seen
by inspection,
the law of the
preceding
order
of differences can often be
found,
and then the law of the
next
preceding
order of
differences,
and so on until the
law of the series itself is obtained. The method will be
seen from the
following examples.
Ex. 1. Find the nth term of the series
The first order of differences is 6
+ 17 +
35
+ 59 + 89 + ......
second
12
+
18 +
24
+ 30+ ......
,
third ,
6 + 6-1-6
+ .. .
416 METHOD OF DIFFERENCES.
The second order of differences is
clearly
an arithmetical
progres-
sion whose nth term is 6
(n
+
1).
Hence,
if v
n
be the nth term of the first order of
differences,
we
have in succession
Also
^=6.1-1.
Hence, by
addition,
t>
B =6(l
+
2
+ ......
+n)-l
=
3n(n
+
!)-!.
Then
again,
we have in succession u
n
-
u
w_j
=
v
n_j
=
3(n-l)n-l;
n
_
1
-u
B_,
=
3(n-2)(n-l)-l; ...;
t*
a
-Jt
1
=
3. 1.2-1. Also u
l
=
l.
Hence u
n
=3
{(n-l)n
+ ...... +
1
.2}
-n
+
2
=
(n-l) n(n
+
l)
-n
+2.
Ex. 2. Find the nth term and the sum of n terms of the series
6+9 + 14+28+40 + ......
The first order of differences is 3
+
5 + 9 + 17+ ......
second,, ,,2
+
4
+ 8+ ......
Hence the second order of differences is a
geometrical progression,
the
(n-l)th
term
being
2*"
1
.
Hence,
if r
m
be the nth term of the
first order of
differences,
we have in succession
Also
1
=3.
Hence,
by
addition,
r
m
=
(2
+
2
a
+ ...... +
2-
1
)
+3
=
Then
again,
we have in succession
-
-u
w
_
1
=r
w
_
1
=2
l*-1
+
u
-i-H-=
2
"~*
+
1
....... a
-u
1
=2
1
+
l. Also
1=
=6.
Hence
=
The sum of n terms of the series can now be written down : for
the sum of n terms of the series whose
general
term is 2*
+
n
+
3 is
Note.
By
the method
adopted
in the
preceding
examples
the nth term of a series can
always
be found
provided
the terms
of
one
of
its orders
of differences
are all
the
same,
or are in
geometrical progression.
330. It is of
importance
to notice that when the
nth term of a series is an
integral expression
of the rth
degree
in
n,
all the terms of the rth order of differences
will be the same.
RECURRING SERIES.
417
For,
if u
n
=
A
r
n
r
+
J.
r
_
1
n
r
"
1
+
. . .
-t-
A
,
where A
r
,
A
r
_
lt
...
do not contain
n,
the nth term of the first order of differ-
ences will be
which
only
contains n to the
(r l)th degree.
Similarly
the nth term of the second order of differ-
ences will be of the
(r 2)th degree
in n
;
and so
on,
the
nth term of the rth order of differences
being
of the
(r
r)th
degree
in
n,
so that the nth term of the rth order of
differences will not contain
n,
and therefore all the terms
of that order of differences will be the same.
When therefore it is found that all the terms of the
rth order of differences are the
same,
we
may
at once
assume that u
n
=
A
r
n
r
+
A
r
_
l
n
r
~
l
+
...
+
A
,
and find the
values of A
T ,
A
r_
l}
...,
A
by comparing
the actual terms of
the series with the values obtained
by putting
n
=
1,
n
=
2,
&c. in the assumed value of u
n
. This method will not
however
give
the value of u
n
in a convenient form for
finding
the sum of the series
; for,
if r be
greater
than
3,
the sum of n terms of the series whose
general
term is
A
r
n
r
+
A
r
_
l
n
r
~
1
+
... cannot be found
[see
Art.
321]
without
a troublesome transformation which will in fact reduce u
n
to the form in which it is obtained
by
the method of
the
preceding
Article. A much better method would be
to assume that u
n
=
A
r
(ri)r
+
A
r
_
l (n)r
_
l
+
. .
.,
and then to
find A
r
,
A^,..., AI
as above.
RECURRING SERIES.
331. Definitions. When
r-fl
successive terms of
the series a +
a^x
+
a
z
a?
+
......
+
a
n
x
n
+. . . are connected
by
a relation of the form a
n
x*
+
px (an
_
l
a;*"
1
)
+
qx* (aw
_
2
a;*"
8
)
+
...
=
0,
the series is called a
recurring
series of the rth
order,
and 1 +
px
+
qx*
+
... is called its scale
of
relation.
The relation does not hold
good
unless there are r terms
before the
nth,
so that the relation
only
holds
good
after
the first r terms of the series.
S.A.
27
418 RECURRING SERIES.
For
example,
the series 1
+
2x
+
4*3?
+
8x*
+
...... is a
recurring
series of the first
order,
the scale of relation
being
1 2x.
Again,
it will be found that the series
1
+ 3#
+
5x*
+
7x
3
+
9x* +...... is a
recurring
series of the
second
order,
the scale of relation
being
1 2x
+
of.
332. To
find
the sum
of
n terms
of
a
given recurring
series.
Let the series be a
+
a^x
4- ...... + a
n
x*
+.
.
.,
and let the
scale of relation be 1
+
px
+
qx*. [This
assumes that the
recurring
series is of the second
order,
but the method is
perfectly general].
Then
......
+
a
n
x
n
;
:. S
n
(1
+
px
+
qx*)
=
a
+
(at
+
pa )
x
+
(a^
+
p
. . .
+
(a.
+
pa^
+
qan^)
x*
+
( pan
+
qa^)
x"*
1
+
qa
n
x*"
=
a
-j-
(a,
-f
pa )
x
+
(pan
+
qan_J
x*+
l
+
qaH
x***,
since all the other terms vanish in virtue of the relation
a
t
x*
+
px (ak_i
x*~
l
)
+
qa? (at
_
t
ic*"
f
)
=
0,
which is
by sup-
position
true for all values of k
greater
than 1.
Hence
"
1
+
px
-f
qx*
If the
given
series be a
convergent
series,
the nth term
will be
indefinitely
small when n is increased without
limit;
and the sum of the series continued to
infinity
will in this case be
given by
*
I+px
+
qx*
The
expression , ,
is therefore such that if it
1
+px
+
qor
can be
expanded
in a
convergent
series
proceeding
accord-
ing
to
ascending powers
of
x,
the coefficient of x* in its
expansion
will be the same as in the
recurring
series.
RECURRING SERIES. 419
On this account the
expression
-:
. ,-
is
1
+px
+
qx
called the
generating function
of the series.
333. A
recurring
series
of
the rth order is determined
when the
first
2r terms are
given.
For let the series be
Then,
the series
being
a
recurring
series of the rth
order,
if we assume that the unknown scale of relation is
1
+pla;+p^+...+pr
af,
we have
by
definition the follow-
ing equations
a
r-H~^~Pi
a
'r+l "^Pflr
+ +
Pr
tt
l
=
^>
=0,
We have therefore r
equations
which are sufficient to
determine the r unknown
quantities p,, p2
, ...,
pr
in the
scale of relation
;
and when the scale of relation is deter-
mined the series can be continued term
by
term,
for a
iri
.
l
is
given by
the
equation
a,
tr+l
+
p^
+
. . .
+
p
r
a
r
=
;
and
when o
2r+1
is
found,
a^
9
can be found in a similar manner
;
and so on.
The series is
similarly
determined when
any
2r con-
secutive terms are
given.
334. From Art. 305 we know that if
p
<
r
+
1,
to r
+
2 terms
=
0,
for all values of k.
272
420
RECURRING SERIES.
This shews that the series
1-
+
x
+
3V
+.
.
.+
(n
+
1)V
-f.
. .
is a
recurring
series whose scale of relation is
(1 xf
1
'.
It also shews that the series
a,
+
a^x
+
aj?
+.
.
.-f
a
n
x*
+.
. .
is a
recurring
series whose scale of relation is
(1 #)
r+1
whenever
a.
is a rational and
integral expression
of the
rth
degree
in n.
335. In order to find the sum of
any
number of terms
of a
recurring
series
by
the method of Art
332,
it is neces-
sary
to know the
general
term of the
series;
we must
therefore shew how to obtain the
general
term of a
recurring
series when the first few terms are
given.
By
Art.
333 the scale of relation of a
recurring
series
of the rth order can be found when the 2r first terms are
given
; and,
having
found the scale of
relation,
the
genera-
ting
function is at once
given by
the formula of Art. 332.
Now,
provided
the scale of relation can be
expressed
in factors of the first
degree,
the
generating
function can
be
expressed
as a series of
partial
fractions of the form
A A
or of the form -r=
TV,
and the coefficient of
any
\-ix
(I ax)*
poAver
of x in the
expansion
of the
generating
function
can be at once written down
by
the binomial
theorem;
and thus the
general
term of the series is found.
When the value of x is such that the
given recurring
series is not
convergent,
the
generating
function will not
be
equal
to the
given
series continued to
infinity
nor can
it be
expanded
in a series of
ascending powers
of x
; but,
taking
as an
example
the
generating
function in Art.
332,
a.
+
(a,
+
pan) y
, ,
.
the
expression
-
l
* 1*
can
always
be
expanded
in
i+w+W
ascending powers
of
y,
if
y
be taken
sufficiently
small,
and
RECURRING SERIES. 421
the coefficients of
y
and
y
l
in this
expansion
will
clearly
be a and a
t respectively
and all
succeeding
terms will
obey
the law a>
t
+pa
t
_
1
+
qa
k
^.
i
=
}
and hence all the coeffi-
cients of the
expansion
will be the same as the corre-
sponding
coefficients in the
given
series. We
may
there-
fore in all
cases,
whether the series is
convergent
or
not,
find the
general
term of a
recurring
series
by writing
down the
expansion
of its
generating
function in
ascending
powers
of x on the
supposition
that x is
sufficiently
small.
Ex. 1. Find the nth term of the
recurring
series 3
+
4a;
+
Gx
2
+
IQx
3
+ . . .
In an
example
of this
kind,
in which the order of the
recurring
series is not
given,
it must
always
he understood that what is wanted
is the
recurring
series of the lowest
possible
order whose first few
terms
agree
with the
given
series. In the
present example
there is
a sufficient number of terms
given
to determine a
recurring
series of
the second
order,
but an indefinite number of
recurring
series of the
third,
or of
any higher
order than the
second,
could be found whose
first four terms were the same as those of the
given
series.
[See
Art.
327.]
Assuming
then that the scale of relation is
I+px
+
qx
3
,
we have
the
equations
6
+
4p
+
3q
=
0,
and 10
+
6p
+
4g
=
0,
whence
p
=
-
3 and
g
=
2. Hence the scale of relation is 1
-
3x+2x
2
.
The
generating
function is therefore
3
+
(4-9)x_
3-5x
2_
1
l-2x
Hence the
general
term of the series is
(2
+
2
n
~
1
)
ar
1
^
1
.
The sum of n terms can now be found
by
the method of
Art.
332;
the sum can however be written down at
once,
for the sum of n
terms of the series 2
(l
+
x
+x*+
...)
is 2
(1
-
x*) / (1
-
x)
and the sum
of n terms of the series l
+ 2o; +
4x
1!
+ ... is
(l-2
n
x
n
)/(l-2x).
We
may
remark that the
given
series is
convergent provided
x < .
Ex. 2. Find the nth term and the sum of n terms of the series
1
+
3
+ 7 +
13
+
21
+
31
+ ....
Consider the series 1
+
3x
+ 7x
2
+
13x
3
+
2Lr*
+
31x
5
+ . . .
Then, assuming
that the series is a
recurring series,
and also that
a sufficient number of terms are
given
to determine the
recurring
series
completely,
it follows that the series is of the third order.
Let then the scale of relation be
1+px
+
qx^+rx
3
;
we then have
the
following equations
to find
p, q,
r :
422 RECURRING SERIES.
13
+
7p+3g+r=0,
and 31 +
21p
+
13g
+
7r=0,
whence
p=-3, q=B
and
r=-l,
so that the scale of relation is 1
-
3z
+
3x
2
-
a-
3
.
The
generating
function is now found to be
JLfl=
2 2
1_
Hence the
general
term of the series
Thus the
general
term of the
given
series is n
2
-
n + 1.
Having
found the
general
term of the series the sum of the first n
terms can be written
down,
for the sum of n terms of the series
whose nth term
isn(n-l)
+
l is
3
(n-l)n(n+l)
+
n.
o
Ex. 3. Find the nth term of the series 2
+
2
+
8
+
20
+ ......
Considered as a
recurring
series of the lowest
possible order,
the
generating
function of 2
+
2z
+
8z
2
+
20x
s
+...
wiU be found to be
2-2*
1-2X-2X
3
'
Now the factors of 1
-
2x
-
2x
a
are
irrational,
and therefore the
nth term of the
series,
considered as a
recurring
series of the second
order,
will be a
complicated expression containing
radicals.
On the other
hand, by
the method of Art.
329,
we should be led
to conclude that the nth term of the series was
(3n
8
-
9n +
8)
x*"
1
,
which
by
Art. 334 is a
recurring
series of the third order.
As we have
already remarked,
the actual law of a series cannot be
determined from
any
finite number of its
terms,
and the above is a
case in which it would be difficult to decide as to what is the
simplest
law that the few terms
given obey,
for the
recurring
series
of the lowest order which has the
given
terms for its first four
terms is not the
recurring
series which
gives
the
simplest expression
for the nth term.
CONVERGENCT
AND DIVERGENCY.
336. We shall now
investigate
certain theorems in
convergency
which were not considered in
Chapter
XXI.
CONVERGENCY.
423
337.
Convergency
ofinfinite
products.
A
product
composed
of an infinite number of factors cannot be con-
vergent
unless the factors tend to
unity
as their limit
;
for
otherwise the addition of a factor would
always
make a
finite
chaoge
in the continued
product,
and there could be
no
definite,
quantity
to which the
product approached
without limit as the number of factors was
indefinitely
increased.
It is therefore
only necessary
to consider infinite
pro-
ducts of the form
where u
n
becomes
indefinitely
small as n is
indefinitely
increased;
and the
convergency
or
divergency
of such
products
is determined
by
the
following
theorem.
Theorem. The
infinite product
II
(1
+
u
r\
in which
all the
factors
are
greater
than
unity,
is
convergent
or
divergent according
as the
infinite
series Sw
r
is
convergent
or
divergent.
Since e*
> 1
+
x,
for all
positive
values of
x,
it follows
that
Hence,
if 2w
r
be
convergent,
II
(1
+
u
r
)
will also be
convergent.
Again,
(1
+
w,) (1
+
w
2
)
> 1
+
u
t
+
u
3
,
and so
on,
so that
n
(i
+
o
> i
+ 2w
r
.
Hence,
if Sw
r
be
divergent,
H
(1
+
u
r)
will also be
divergent.
m T. tu
. ...- .
Ex. 1. To shew that
T~-. ^frr ^ 77
-
T
'. is infinite or
zero,
when
b(b+l)(b
+
2)...(
n is
indefinitely
increased,
according
as a is
greater
or less than b.
424 CONVERGENCY.
For,
if a >
b,
the
expression may
be written in the form
which is
greater
than
1 1
|
__
+
_
+ ......
J.
But
Y
+
r
i
+
j
n
+ ... is a
divergent
series
[Art. 274]
: the
given
ex-
pression
is therefore infinite when n is
infinite,
a
being greater
than 6.
If 6>a: then as
before,
j
-
-44
-
;
......
is
infinite;
and
a(o+l)(a
+
2)
.....
..
, a(a
......
therefore
-=-75- 17-71 ?rr
-
must be zero.
Ex. 2. Determine whether the series
a
a(a
+
x)
a
(a
+
x) (a
+
2x)
b
+
b(b+x)
+
b(b+x)(b
+
2x)
+ ......
is
convergent
or
divergent.
From Art.
325,
we have
"
a
+
x-b
,
(a+
... . . .
Now bv Ex.
1,
- --
=
-
is infinite or zero
according
b(b
+
x)...(b
+
n-lx)
as a
+
x
^
6.
Hence the
given
series is
convergent,
and its sum is then
b-a-x'
if b >
a+x.
Also the series is
divergent
if ii< a
+ x.
Also if b=a
+x,
the series becomes
r + , + ; ^-
which is
I
b+x
b
+
2x
known to be
divergent [Art. 274].
338. The Binomial Series. We have
already proved
that the binomial
series,
namely
...
&
Zj
1 . Z . o
is
convergent
or
divergent,
for all values of
m,
according
as a; is
numerically
less or
greater
than
unity.
If x
=
1,
the series becomes
m(m-l) w(m-l)(m-2)
1.2
+
~
1.2.3 *,":'
CONVERGENCY. 425
Now we know that the terms of this series are
alternately positive
and
negative
after the rth
term,
where
r is the first
positive integer greater
than m
+
1. More-
over the ratio w
n+1 /wn
is
numerically
less or
greater
than
unity according
as m
+
1 is
positive
or
negative.
The
series will
therefore,
from theorem V.
Chapter
XXL be
convergent
when m
+
1 is
positive provided
the nth term
decreases without limit as n is increased without limit.
1 1.2
n
Now +
=
u
n ( w)(l m)...(n
1
m)
Now,
if m
+
1 be
positive
and less than
r,
the
product
of
the factors from the rth onwards is
greater
than
.
r m r
+
lm
and the
product
of the
preceding
factors is finite.
Hence,
when n is increased without
limit,
!/"
is in-
finitely great,
and therefore u
n
indefinitely
small,
provided
1
+
m be
positive.
Thus the binomial series is
convergent if
a?
=
1,
pro-
vided m
>
1.
If x
=
1,
the series becomes
The sum of n terms of the above series is
easily
found
to be
[see
Art. 287 or Art.
325]
(l-m)(2-w)(3-m)...(n-l -m)
1.2.3...(n-l)
The sum of n terms of the series is therefore
[Ex.
1,
Art.
337],
zero or
infinite,
when n is
infinite,
according
as m is
positive
or
negative.
Thus the binomial series is
convergent
when x
=
l,
provided
m is
positive.
426 CONVERGENCY.
339.
Cauchy's
Theorem.
If
the series w
t
-f
w
s
+
8
+ ... +
u
n
+
... have all its terms
positive,
and
if
each term
be less than the
preceding,
then the series will be
convergent
or
divergent according
as the series
w,
+
au
a
+
a*u
a
t
+
...
+
a*u
a
+
... is
convergent
or
divergent,
a
being any positive
integer.
For,
since each term is less than the
preceding,
we
have the
following
series of relations
i*
t
+
w,
+
...
+
u
a
<
CK^
<
(a 1) Wj-f !,
w
+
w
+
w * < a*
-
u<a-1
au,
u
a
,
+l
+
w
.+2
-t- . .
.+
WaH-i
<
(a"
+1
-
a")
w
a
, <
(a
-
1)
a
n
n
a
,.
Hence,
by
addition,
S <
(a 1)
2 +
w,
............
(I),
where S and
2 stand for the sum of the first and second
series
respectively.
Again,
we have since a is
<f 2,
o
(a
2
a)
w
a
o
(.-
1+1
+
M
a.-i+2
+. .+
a )
> a
(a*
a"~
l
)
u
a
. > o"w
a
..
Hence aS>'2-u
l
.'
(II).
From I and II it follows that if S is finite so also is
2,
and that if S is infinite so also is
S.
Ex. To shew that the series
-^ ^
is
convergent
if Jk be
greater
n(logn)*
than
unity,
and
divergent
if k be
equal
or less than
unity.
By Cauchy's
theorem the series will be
convergent
or
divergent
a*
according
as the series whose
general
term is
-
:
j
is
convergent
or
divergent.
CONVERGENCY. 427
V
_y
___
a
(log o)*
n*
(log a)*
~
(log a)*
n*
it therefore follows from Art. 274 that the
given
series is
convergent
if k > 1 and
divergent
if k
:f
1.
340. We shall conclude with the two
following
tests
of
convergency
which axe sometimes of
use,
referring
the
student to Boole's Finite Differences and Bertrand's Differ-
ential Calculus for further information on the
subject.
341. Theorem. A series is
convergent
when,
from
and
after any particular
term,
the ratio
of
each term to the
preceding
is less than the
corresponding
ratio in a known
convergent
series whose terms are all
positive.
For let the
series,
beginning
at the term in
question,
be
C
r
=
w
1
+
ti
a
4-
u
s
+... +u
n
+
...,
and the known
convergent
series,
beginning
at the same
term,
be
V
=
v
l
+
v
3
+
v
s
+ ...
+
v
n
+
. . ..
ni
Then,
since
-^
<
*
for all values of
r,
we have
u
r
v,
1
v
l
v
t
v
t
v
3
v
a
v
l
u u u u u u
>
-
1
(M,+W2
+
l*
3
+
W
4 +...)>
U.
U
\
U
i
Hence as V is
convergent,
U must also be
convergent.
The
given
series is therefore
convergent,
for the sum
of the finite number of terms
preceding
the first term of
U must be finite.
We can
prove similarly
that
if,
from and after
any
particular
term,
w
r+1
: u
r
> v
r+1
: v
r
,
and all the terms of Sw
r
have the same
sign ;
then 2w
r
will be
divergent
if 2v
r
be
divergent
428
CONVERGENCE
342. Theorem. A
series,
all
of
whose terms are
posi-
tive,
is
convergent
or
divergent according
as the limit
of
( u \
nil
--^
)
is
greater
or less than
unity.
For let the limit of n
(
1
- ^
* l
]
be
a.
V
uj
Consider the series 2
-5
=
2v_
: then
n?
/.,
v
..\ ((n
+
1)^
n
ft
] ftn^
+
lower
powers
of n
nil
**]
=nl -.
-^-^ }
=
c-
s j
----
7:
.
\
v
J
|
(n
+
l)
ts
)
w*
+
lower
powers
ot n
Hence the limit of n
(
1
--
'
1
,
when n is
infinitely
V
fl
n
/
great,
is
/9.
First
suppose
a >
1,
and let
ft
be chosen between
a
and 1.
Then since the limit of n ( 1
---
1
j
is
greater
than
the limit of nil
--
*
]
. there must be some finite value
V
V
n
'
of n from and after which the former is
constantly greater
than the latter.
But when n (1
-
%**\ >n(l-
-^A
,
we have
-5*-
1
>
-^-
l
.
V
n
U
*
Hence,
by
the
previous
theorem, Sw,
will be conver-
gent
if 2v
m
be
convergent;
but
Sv.
is
convergent
since
/8>1.
Similarly,
if a be
<
1,
and
ft
be taken between a and
1,
we can
prove
that 2w
m
is
divergent
if
Sv.
is
divergent,
and the latter series is known to be
divergent
when
ft
< 1.
If the limit of n
(
1
--
J
be
unity
the test fails.
EXAMPLES. 429
,
Ex. 1. Is the senes
^
+
-_
+
-__^
+ ...
convergent
or
divergent
?
Here
_?l
=
_-
x,
the limit of which is x.
Hence,
either from
u
w
b+n
the
beginning
or after a finite number of
terms,
--5*
1 >
1
according
as x
>
1.
< <
Hence the series is
divergent
if x >
1,
and
convergent
if x < 1.
If x
1,
the limit of is
unity.
But
".
b
+
nj'
the limit of which is b
-
a.
Thus,
if x
=
l,
the series is
convergent
when b-a>1 and
divergent
when b
-
a < 1. When
b=a+
1,
the series becomes
a a a
b
+
b
+
l
+
b
+
2
+
'
which is
divergent.
[These
are the results arrived at in
Ex.
2,
Art.
337.]
EXAMPLES XXXIV.
1. Find the sum of each of the
following
series to n
terms,
and when
possible
to
infinity
:
4 4.7 4.7.10
}y
5
+
5.8
+
5.8.11
+ ""
2 2.5 2.5.8
3.5 3.5.7
11 11.13 11.13.15
'
1V
'
U
+
U.16
+
1.IG.IS
+
""
2.
Find,
by
the method of
differences,
the wth term and
the sum of n terms of the
following
series :
(i)
2 + 2 + 8 + 20 + 38 + ....
(ii)
7 +
14 + 19 + 22 + 23 + 22+....
(iii)
1+4 + 11+26+57 + 120+....
430 EXAMPLES.
(iv)
1 + + 1 + 8 + 29 + 80+193+....
(v)
1 + 5 + 15 + 35 + 70 + 126+....
(vi)
1 + 2 + 29 + 130 + 377 + 866 + 1717 + ....
3. Find the
generating
function of each of the
following
series on the
supposition
that it is a determinate
recurring
series :
(i)
(ii)
(iii)
(iv)
1 + a; + 2aj*
+ 7z* + Ha;
4
+ 35a;
5
+ ....
(v)
I
1
+ 2
2
x + 3V + 4V+ 5V + 6V +
....
4. Find the nth
term,
and the sum of n terms of the
following recurring
series:
(i)
2 + 6 + 14 + 30+....
(ii)
2-5 + 29-89+....
(iii)
1+2
+ 7 + 20+....
5. Find the nth term of the series
1, 3, 4, 7,
<kc.
; where,
after the
second,
each term is formed
by adding
the two
preceding
terms.
6. Determine
a, b,
c,
d so that the coefficient of a" in
e
a + bx + cx' + dx?
the
expansion
of
-
^
-
^
-
may
be
(w+ 1)
.
7. Shew that the series 1'+ 2
r
x + 3V + 4V + ... is the
expansion
of an
expression
of the form
- -
^j
;
shew
also that a
r
=
0;
and that a
r_
l
=
a
l
_
1
.
8. Find the sum to
infinity
of the
recurring
series
2
+ 5x + 9x* + 15x + 25x
4
supposed convergent,
it
being given
that the scale of relation
is of the form 1 +
px
+
qx?
+ nf. Shew that the
(n
+
l)th
term
of the Hpj-ies is
(2"
+ 2n +
l)x*.
EXAMPLES.
431
9. Find the sum to
infinity
of the series
1 + 4* + llx* + 2Qx* + 57x
4
+ 120x* +
x
being
less than .
10. Find the sum of n terms of the series
x x(x +
a)
x
(x
+
a)
(x +
b)
I+- +
v
,
'
+
- J
'-+ .
a ab abc
11. Shew that
1 a
I
432
EXAMPLES.
16. Shew
that,
if m be a
positive integer,
2n+l
m(m-l)(2n
+
l)(2n
+
3)
1.2
(2n
+
2)(2r
+
4)
"'
(2w
+
2) (2w
+
4)
. . .
(2n
+
2m)
'
m n
+
1
m(m-l)
17. Shew
that,
if
TO,
n and m n
+
1 are
positive integers;
then
m-n+1 1.2
(m
n+
l)(m
n +
2)
n
(n
-
1) (ro
-
2)
_
m(m-l)(m-2)
_
1.2.3
(m
-
w +
1) (m
-
w +
2) (m
-
n +
3)
,
...
+ ...to
(w+ 1)
terms
=
-.
-
v
.
x /
v
-
'.
v
.
-'---
r.
(m
n +
1)
(m
-
n +
2)
. . .
(m
n +
m)
18. Shew
that,
if m + 1 >
0,
then
lm(m-l) lm(m-l)(m-2)
T72 5~
1.2.3
1
19. Shew
that,
if P
r
be the sum of the
products
r
together
of the first n even
numbers,
and
Qr
be the sum of the
products
r
together
of the first n odd
numbers;
then will
l +
P.+J^-f
+P
a
=
1.3.5...(2n+l),
and
1$.^
+
^+
+
#
-
=
2.4.6...2n.
20. Prove that
{a
+
(a+ !)
+
(
+
2)
+ ... +
(a
+
n)} {*
+
(+ l)
+
(a
+
2)
-t- ...
...
4-(a
+
w)}
=
a
3
+
(a-fl)
8
+...
+(a
+
n)
3
.
21. Shew that the series
1-q-
(l-q-)(l-q-') (1 -q") (1 -q-) (1
-q-
8
)
1-q
+
(l-q)(l-q
8
) (1
-
q) (1
-
a*) (1
-
q
8
)
is zero when n is an odd
integer,
and is
equal
to
(1 q) (1
-
q
3
)
...
(1
a""
1
)
when n is an even
integer [Gauss].
EXAMPLES. 433
22. Find the sum of the series
n n 1 n 2
1.2. 3^2. 3.4^3.4.5"
w(w
+
l)(n
+
2)'
23.
Sumtoinfinity
^-|^
+
i^-....
24.
Sum,
when
convergent,
the series
x a? x"
O
+
273
+ " +
n(n+l)
+ '"
'
25. Sum to
infinity
the series
1 .
2 . 3
+
3 . 4 . 5x
+ 5 . 6 . 7a* + 7 . 8 . 9x* +
...,
x
being
less than
unity.
26. Shew
that,
if n is a
positive integer
1 n ,8n(3n-3)
3rc(3tt-4)(3n-5)
_
1.2 1.2.3
27. Shew
that,
if a
lt
a
s
,
a
a
,...
be all
positive,
and if
aj
+
a
g
+
a
a
+ ... be
divergent,
then
is
convergent
and
equal
to
unity.
28. Shew that the series
2 T
is
convergent
if >
1,
and is
divergent
if x
^>
1.
29. Shew
that,
if the series u
l
+u
a
+
u
a
+ ... +u
n
+ ... be
divergent,
the series
will also be
divergent.
S. A. 28
434 EXAMPLES.
30. For what values of x has the infinite
product
(1
+
a) (1
+
ax) (1
+ aa:
2
) (1
+ ax
8
)
... a finite value?
31. Prove
that,
if v
n
is
always
finite and
greater
than
unity
but
approaches unity
without limit as n increases
indefinitely,
the two infinite
products
v^jojo^
. . .
,
vj)v*v*...
are either both finite or both infinite.
32. Test the
convergency
of the
following
series :
_i_ 2_
f
y n
n
. Ill
r*ptsp*
2.4 2.4.6
2.
4.6...
2*
+ ....
3.5.7...(2n+l)(27i
+
2)
J_
1.3 1.3.5
2.3
+
2.4.5
+
2.4.6.7
+ ""
+
2.4.6...2n(2n+l)
+
"' '
1.2.3.y(y
CHAPTER XXVT.
INEQUALITIES.
343. WE have
already proved [Art. 232]
the theorem
that the arithmetic mean of
any
two
positive quantities
is
greater
than their
geometric
mean. We now
proceed
to
consider other theorems of this
nature,
which are called
Inequalities.
Note.
Throughout
the
present chapter every
letter
is
supposed
to denote a real
positive quantity.
344. The
following elementary principles
of
inequal-
ities can be
easily
demonstrated :
I. If a > b
;
then a
+
x > b +
x,
and a x > b x.
IL If a > b
;
then a < b.
III. If a
>
b
;
then ma >
mb,
ana ma
<
nib.
IV. If a >
6,
a' >
&'',
a" >
b",
&c.
;
thena + a'+a"+...
> b
+ b'+
b"
+
...,
and oafa"... >bb'b"....
V. If a > b
;
then a" > b
m
,
and a"""' < If*.
Ex. 1. Prove that o
+
& > a?b
+
a&
3
.
We have to
prove
that
as-a*6-a6
2
+6
3
>0,
or that
(a
s
-
6)
(a
-
6)
>
0,
which must be true since both factors are
positive
or both
negative
according
as a is
greater
or less than b.
282
486
INEQUALITIES.
Ex. 2. Prove that a
m
+
a"** > a* +
or*,
if m > n.
We have to
prove
that
(a
m
-a*)(l-a~
m
~
n
)>Q,
which must be
the case since both factors are
positive
or both
negative according
as
a is
greater
or less than 1.
Ex. 3. Prove that
(P
+
m'
+
n
2
) (l
fi
+
m'*+n'*)>(ll'
+
mm'
+
nn')
J
.
It is
easily
seen that
(It
+
m
*
+
n
2) (i
<2
+
m
v
+
n
-s)
_
(M
.
+
mm
>
+nn
y
=
(mn
1
-
m'n)*+(nl'
-
n'l)
z
+
(lm'
-
I'm)*.
Now the last
expression
can never be
negative,
and can
only
be
zero when mn'
-
m'n,
nl'
-
n'l and lm'
-
I'm are all
separately zero,
the
conditions for which are
^
=
T
=
-7
.
I m n'
Hence
(P
+
m
2
+n
8
) (i'
2
+
m'
2
+
n'
a
)
>
(M'+mm'
+
nn')
2
,
except
when
Z/r=m/m'=n/n',
in which case the
inequality
becomes an
equality.
345. Theorem
I. The
product of
two
positive quanti-
ties,
whose sum is
given,
is
greatest
when the two
factors
are
equal
to one another.
For let 2a be the
given
sum,
and let a
+
x and a x
be the two factors. Then the
product
of the two
quanti-
ties is a*
a?,
which is
clearly greatest
when x is
zero,
in
which case each factor is half the
given
sum.
The above theorem is
really
the same as that of Art. 232
;
for
from Art. 232 we have
-*
> ab.
346. Theorem II. The
product of any
number
of
positive quantities,
whose sum is
given,
is
greatest
when the
quantities
are all
equal.
For,
suppose
that
any
two of the
factors,
a and
b,
are
unequal.
Then,
keeping
all the other factors
unchanged,
take
^(a
+
b)
and
^(a
+
b)
instead of a and b : we
thus,
without
altering
the sum of all the
factors,
increase their continued
product
since
^(a
+
6)
x
^(a
+
b)
>
ab,
except
when a
=
6.
Hence,
so
long
as
any
two of the factors are
unequal,
the continued
product
can be increased without
altering
the sum
;
and therefore all the factors must be
equal
to
one another when their continued
product
has its
greatest
possible
value.
INEQUALITIES.
437
Thus,
unless the n
quantities
a, b, c,
... are all
equal,
T j
/a+ b
+
c
+ d+
...\*
abed ...
<[
,
V
n
J
and therefore
"rC-j-Ct'T-...
-,/ 7 \
>
?/(abcd ...).
n
By extending
the
meaning
of the terms arithmetic
mean and
geometric
mean,
the last result
may
be enunci-
ated as follows :
Theorem III. The arithmetic mean
of any
number
of
positive quantities
is
greater
than their
geometric
mean.
Ex. 1. Shew that a
+
b
3
+ c
3
> Babe.
a
3
_|_
58
_j_ C
3
We have >
z](a?
. b
3
. c
3
)
> aoc.
o
Ex.2. Shew that
+ + + +
>n.
We have -(
+ + +
)
>
^/
(
.^..,
Ex. 3. Find the
greatest
value of
(a-x) (b-y) (cx
+
dy),
where
a, o,
c
are known
positive quantities
and
a-x, b-y
are also
positive.
The
expression
is
greatest
when
(ac
-
ex) (bd
-
dy) (ex
+
dy)
is
greatest,
and this is the
case,
since the sum of the factors is now con-
stant,
when ac
-
ex
=
bd
-
dy
=
ex
+
dy
. Whence the
greatest
value is
found to be
(a
Ex. 4. Find when x
a
y
z
1
has its
greatest
value,
for different values of
x, y
and z
subject
to the condition that x
+
y
+
z is constant.
Let
P=a;V;
then
_x
x x
y_
y y
z_ z_ z_
a' a' a
......
/3'j8
......
y
'
y' y'"
The sum of the factors in the last
product
is
constant,
since there
are a factors each
-
,
8 factors each
\ ,
and
y
factors each
-
,
and
ft 7
therefore the sum of all the factors is x
+
y+z.
438
INEQUALITIES.
Hence,
from Theorem
II,
(-jfl) (-)
nas its
greatest
value
x
y
z
when all the factors are
equal,
that is when
-
=
|
=
-
.
a
p y
It is clear that P is
greatest
when
P/a
a
/3\
y
is
greatest,
since
o, /3, 7
are constant
;
hence P is
greatest
when
a;/a=y//3=z/>.
In the above it was assumed that
a, /3, 7
were
integers
;
if this
be not the
case,
let n be the least common
multiple
of the denomina-
tors of
a, /3, 7.
Then x
a
y
z
y
will have its
greatest
value when
x
na
y
n
z
n
"
f
has its
greatest
value,
which
by
the
above,
since
na,
n/3
and
ny
are all
integers,
will be when
=
-^
=
that is when
na
np ny
Thus,
whether
o,
/5, 7
are
integral
or
not,
x
a
yz
y
is
greatest
for
values of
z,
y
and z such that x
+
y
+
z is
constant,
when
xja.
=
yjft
=
z/y.
347. Theorem IV. The sum
of any
number
of
positive quantities,
whose
product
is
given,
is least when
the
quantities
are all
equal.
First
suppose
that there are two
quantities
denoted
by
a and 6.
Then,
if a and b
are_unequal,
(/a
-
V&)
2
>
0,
and there-
fore a
4-
b >
Jab
+
Jab.
Hence the sum of
any
two
unequal quantities
a,
b is
greater
than the sum of the two
equal quantities
Jab, Jab
which have the same
product.
Next
suppose
that there are more than two
quantities.
Let
a, b,
any
two of the
quantities,
be
unequal.
Then,
keeping
all the others
unchanged,
take
Jab
and
Jab
instead of a and b : we
thus,
without
altering
the
product
of all the
quantities,
diminish their sum since
Jab
+
Job
<a+b.
Hence,
so
long
as
any
two of the
quantities
are
unequal,
their sum can be diminished without
altering
their
product;
and therefore all the
quantities
must be
equal
to one another when their sum has its least
possible
value.
INEQUALITIES.
439
348. Theorem V.
If
m and r be
positive,
and m > r
;
then,
unless a
1
=
a
z
=
0^
=
&c.,
a
n
m
.
, .,
=-
will be
greater
than
n
n n
We have to
prove
that
n
(a
+
a +
...)>
(a/
+
o,
r
+
. .
.) (a1
m
~
r
+
a
2
m
-
r
+...),
or that
(n
-
1) (a,-
+
a
+
...)>
2 a
2
m
-.+
<CX),
or that
2
(a,
1"
+
a,"
-
-'
-
a
1
-r
a
i
p
)
>
0,
every
letter
being
taken with each of the
(n 1)
other
letters.
Now
<
+
a,"
-
a,V"
-
<""<
=
-
O
"
~
<")
which is
positive
since
'a,
1
"
a/
and
aj
1"'1
"
a
2
m
~
r
are 6o^A
positive
or both
negative according
as a
t
is
greater
or less
than a
2
.
Hence 2
(a,"
+
a
2
m
-
a
1
r
a
8
m
~
r
-
a
1
l "-p
a
a
f
)
>
0,
which
proves
the
proposition.
By repeated application
of the above we have
^
n n w ?i
where
a, /3,
7,
... are
positive quantities
such that
Ex. 1. Shew that 3
(a
8
+
ft
3
+
c
3
)
>
(a
+
6
+
c) (a
2
+
6
2
-f-
c
2
).
Ex. 2. Shew that a
s
+
b
s
+c
s
> abc
(a
2
+6
3
+c
2
).
From Theorem
V,
3 3
.
abc,
from Theorem III.
440
INEQUALITIES.
349.
Theorem VI.* To
prove
that,
if
a,b,c,
... and
a,
/9, y,
. . . be all
positive,
then
g +
b/3
+
07
+
...y
a+b +
c
+
...
)
First,
let
a,b,c,
... be
integers.
Take a
things
each
a,
b
things
each
J3,
and so on.
Then,
by
Theorem
III,
... to a
terms)
+
($
+
/3
+
... to b
terms)
+
...
If
a, b, c,
... be not
integral,
let m be the least common
multiple
of the denominators of
a, b, c,
...
;
then
ma, mb,
me,
. . . are all
integers,
and we have
maa.
+
mbft
+
...
COR. I. Put a
=
-, fi
=
T , ...,
and let there be n of the
a o
letters
a, b,
...
;
then
ia
+
6
+
...j
'
a"b
b
...
COR. II. Substitute in
(A)
a
r
for
a,
b
r
for
6,
. . .
;
also
substitute a"*~
r
for
a,
6"*~
r
for
/S, ...,
where m> r.
*
See a
paper by
Mr L. J.
Rogers
in the
Messenger of Mathematics,
VoL xvii.
INEQUALITIES.
441
Again,
substitute a
r
,
b
r
,
..; for
a,
b,
...
respectively,
and
a*'", b*~*,
... for
a, /9
respectively,
where < r.
Then
Hence,
as m r and r t are both
positive,
we have
from
[B]
and
[C]
[q
m
+6
<n
+
...]^
:
|a
r
+6
r
+...j
>
Hence,
provided
m>r>t,
The
following
are
particular
cases of
[D].
Put t
=
; then,
since a
+ 6+
...
=
n,
we have
provided
m>r
Again, put
t
0,
ra
=
1
;
then since ra > r
>
t,
r must
be a
proper
fraction.
Hence,
if
r be a
properfraction,
iin,
put
t
=
0,
r
=
1
;
then m > 1.
Hence,
ifm>\we
have
n
Now
put
m
=
1,
r
=
0,
then t is
negative.
Hence,
pro-
vided t be
negative,
(a
+ b+
...)"*
x rT
1
x
(a'
+ 6'+
...)
> 1
;
442
INEQUALITIES.
From
[F], [G]
and
[H]
we see that
g*
+ 6'+...
>
["a
+ 6
+
...r
n
<\_
n
according
as a; is not or is a
proper
fraction.
350. We shall conclude this
chapter by solving
the
following examples. [See
also Art.
133.]
Ex. 1. Shew
that,
if
=
o
1
+
a
2
+. .. + ,
* n
2
-__
>
-7,
unless
a,
=
<!,=.
..
=
a
m
.
-a
a
-
Unless
04
=
0,=
. ..=a
w
,
we have
1 / S t 8 \ "/
_^_
nV-0!
+
-o,
+ +
s~aj
>
V (*-ai)(*-a2)...(*-aJ'
By multiplication,
since
(s
-
a
x
)
+
(*
-
a
a)
+ ... +
(s
-
oj
=
ns
-
s,
we have
n-1
/_
i t
,
\
n
2
\*-o
a
-
a,
-
a^y
Ex.
2. Shew
that,
ifa
+
6
+c +
d=3;
then will
For
So also
34/{(-c)(-d)(-a)}<6, 3^{(<-<Z) (-a) (-6)}
<c,
and
3^{(-a)(*-6) (*-c)}<d.
Hence 81
(*
-
o) (*
-
ft) (s
-
c) (
-
d)
< aftcd.
/
-2 i
,,2
i
,2\ Z+y+*
Ex. 3. Shew that
(
^
y
^
)
>
z*yVz*,
unless o;=w=z.
V *+y
+
/
First
suppose
that
x, y
and z are
integral;
then
by
Theorem
III.
... to x
terms)
+
(y
+
y
+ ... to
y tenns)
+
(z+^
+ ...
to z
terms)
x
+
y+z
>
If
x, ?/,
z be not
integral
let m be the least common
multiple
of their denominators
;
then
tnx,
my
and mz are
integral,
and we
have
by
the first case
/ TO
a
.T
3
+
my
+ mV\
mx+mv+m.
>
(Hia,-)"" (7nw)
V mx
+
my
+
mt
)
EXAMPLES.
443
/r
2
4- W
a
4- Z
z
\
m
<*+>+*)
that is
(
g
I X m(*+lH-*) >
(
x
*yV
z
i\m
x m
\ x+y
+
z
)
..
\
The Theorem can in a similar manner be
proved
to be true for
any
number of
quantities.
EXAMPLES
XXXV.
Prove the
following inequalities,
all the letters
being
supposed
to
represent positive quantities
:
1.
y*z*
+
z'x*
+
x*y* <fc xyz (x
+
y
+
z).
2.
(;
+
/+.'+
......
XV+V+V+
......
)
3.
4.
a b c
5-
c/
03
6.
7. a*cc? + 6
z
ac? + c*ab + d
2
bc
8.
(6c
+ ca +
a>)
2
<J;
3a6c
(a
+ b +
c).
9. a
4
+ b* + c*
<t
a6c
(a
+
6 +
c).
10. a
5
+ 6
s
+ c* +
c?
B
<
abed
(a
+ b + c +
d).
.. a x a* x* ..
11.
-
<
-5
-
^
,
if x < a.
a +
x a + or
1 1 1
1 1 1
12.
--
1
---
h <t ~~F
= "*"
"/
- "*
--
p^
7
a 6 c
Jbc J
ca,
Jab
13. Shew
that,
if x* +
x?
+ ...... +
x*
=
a,
then na>
(xl
+
x
2
+ ...... 4-
a;,,)
2
> a.
444
EXAMPLES.
14. Prove
that,
if
a;,,
x
t
,
,,
......
,
x
m
be each
greater
than
a,
and be such that
(a,
a) (xt a)
......
(xn a)
=
6",
the
least
value of
a^oyc,...^
will be
(a
+
6)*,
a and 6
being positive.
15. Shew that
ay
+ bx a + b
222 9
16.
T
+
-
+
-
r<t
--
E
-
6+c c+a a+6 a + 6 + c
16
__ ___ . _
c + a* + a d
+
a + b a + 6 + c a + 6+c+cT
18. Shew that if a
>
6
>
c;
then
19. If of
=
y*
+
2*,
then will a:"
>
y"
+ s"
according
as n
*"
2.
20. Shew that
(a6cc?)'
+rH4<
lies between 'the
greatest
and
1111
least of
a',
b
1
, <f,
d' .
21. Shew that
l+aj + x*+ ...... + x
tm
(2n+ l)x".
22. If n be a
positive integer,
and a
> 1
;
then
23.
24,
abc^, (6
+
c-a)(c
+
a-6)(o
+
6-c).
25. a6cd
<
(6
+ c + rf
-
2a) (c
+ d + a
-
26) (d
+ a + b
-
2c)
26.
where
(n !)*
=
,
+a
t
+
+a..
27. If
a, 6,
c be
unequal positive quantities
and such that
the sum of
any
two is
greater
than the
third,
then
1119
r
_ _
j_ _
_
j. . _ _, .
>
.^ .
6+c a
c+a 6 a+6 c a+6+c*
EXAMPLES. 445
28. Shew
that,
unless a
=
b
=
c,
29. Shew
that,
if
a, 6,
c be
unequal positive quantities,
then
30. Shew that
px
1
r
+
qx
r
~
f
+ rx
p
~
1
>p
+
q
+
r,
unless
*=1,
or
/>
=
y
=
r.
31. Shew that
^
-.
7
.
~
2.4. 6 . . . 2w
33. Find the
greatest
value of
ify**,
for different values
of
x, y,
and z
subject
to the condition that oaf +
by^
+ CZ
Y
=
d.
34. Prove
that,
if TO >
2,
([)*>
TO".
35. Shew
that,
if TO be
positive,
36. In a
geometrical progression
of an odd number of
terms,
the arithmetic mean of the odd terms is
greater
than
the arithmetic mean of the even terms.
37. Prove
that,
if an arithmetical and a
geometrical pro-
gression
have the same first
term,
the same last
term,
and
the same number of
terms;
then the sum of the series in
A. p.
will be
greater
than the sum of the series in G. p.
38. Shew
that,
if P
r
denote the arithmetic mean of all
those
quantities
each of which is the
geometric
mean of r out
of n
given positive quantities;
then P
lt
P
t
, ...,
P
n
are in
descending
order of
magnitude.
39. Shew
that,
if s
=
a+6 + c+
...,
'"-^-
-..<<-
n
being
the number of the
unequal positive quantities a, 6, c,
....
40.
Shew
that,
if TO be
any positive integer,
CHAPTER
XXVIL
CONTINUED FRACTIONS.
351. ANY
expression
of the form a b
cd
e &c.
is called a continued
fraction.
Continued fractions are
generally
written for con-
venience in the form
b d
f
a + ..
-
c e
g
352. The fraction obtained
by stopping
at
any stage
is
called a
convergent
of the continued fraction. Thus a and
a +
-
,
that is r- and
=
,
are
respectively
the first and
second
convergents
of the continued fraction a +
-
. . .
~c e
The rth
convergent
of
any
continued fraction will be
denoted
by
&
.
*ft
The fractions
a,
--
,
-,
&c. will be called the
first,
c e
second, third,
&c. elements of the continued fraction.
CONTINUED FRACTIONS. 447
353. In a continued
fraction of
the
form
a-\
-- -
...,
G
-f-
6 -f-
where
a, b, c,
&c. are all
positive,
the
convergents
are
alternately
less and
greater
than the
fraction itself.
For the first
convergent
is too small because the
part
... is omitted
;
the second
convergent,
a
+
-
,
is too
c+
c
great
because the denominator is
really greater
than
c;
then
again,
the third is too
small,
because c +
-
is
greater
&
than c H
--
. . .
;
and so on.
e
+
354. In order to find
any convergent
to a continued
fraction,
the most natural method is to
begin
at the
bottom,
as in Arithmetic : thus
A
'
If
only
one
convergent
has to be
found,
this method
answers the
purpose
;
but there would be a
great
waste of
labour in so
finding
a succession of
convergents,
for in
finding any
one
convergent
no use could be made of the
previous
results: the successive
convergents
to a continued
fraction
are, however,
connected
by
a
simple
law which we
proceed
to
prove.
355. To
prove
the law
of formation of
the successive
convergents
to the continued
fraction
The first three
convergents
will be found to bo
a
a^+
T
-
7
1'
b,
448 CONTINUED FRACTIONS.
Now the third
convergent
can be written in the form
from which it
appears
that its numerator is the sum
of
the
numerators
of
the two
preceding convergent^ multiplied
respectively by
the denominator and numerator
of
tfie last
element which is taken into
account;
and a similar law
holds for the denominator.
We will now shew
by
induction that all the
convergents
after the second are formed
according
to the above law
provided
there is no
cancelling
at
any stage.
For,
assume that the law holds
up
to the nth con-
vergent,
for which the last element is
a^Jb^,
and let
p
rfq
r
denote the rth
convergent
;
then
by supposition
Then the
(w
+
l)th convergent
will be obtained
by
a a a . a b
changing
-=-=*
into
T
*=*
r* ,
that is into
f
* i-5
.
&.-i
b
*-i+
b
n
b
*-A
+
a
.
Hence in
(i)
we must
put a,,.^,
for
a..,
and b
n
,
l
b
n
+
a
n
for
6^;
we then have
*-i [from I].
Similai-ly q^
=
b
nqn
+
0.7..,.
Thus the law will hold
good
for the
(n
+
l)th
con-
vergent
if it holds
good
for the nth
convergent.
But we
know that the law holds
good
for the third
convergent
;
it must therefore hold
good
for all
subsequent
ones.
COR. L
In the fraction
a,
-\
o
+ a+
CONTINUED FRACTIONS.
449
COR. II. In the fraction
y-
1
-^ r
'
?
Pn
=
P-l
~
a
^-8
tm
?n
=
Vn-l
~
?-,
Ex.
By
means of the law
connecting
successive
convergents
to a
continued
fraction,
find the fifth
convergent
of each of the
following
fractions :
1111 1 1 1 1 1
+
'
(U
12345 2222
2
+
3
+
4
+
5 + 6-
(1V)
3+
5
+
5
12345
/ -\ f ? ? i i
(V)
1+1 + 1+1+1'
tV1)
4
+ 3
+
2
+
1
+
2
+
"
22222 ..11111
(>
8-3-8-8-8-~
<
vm
)i_4-l-4-I-"
^,
a. #fc. iHfr, W, H. Hi, H, I-
356. The
convergents
to continued fractions of the
form a
+
r
-
-7 >
where
a, b, c, d,
... are all
positive
o
+ c
+d +
integers,
have certain
properties
on account of which
such fractions have
special utility:
these
properties
we
proceed
to consider. We first however shew that
any
rational fraction can be reduced to a continued fraction
of this
type
with a finite number of elements.
7?2
For let be the
given
fraction
; then,
if ra be
greater
7i
than
n,
divide ra
by
n and let a be the
quotient
and
p
the
remainder,
so that
=
a
+
-
. Now divide n
by
n n
and let b be the
quotient
and
q
the
remainder;
then
=-=
. Now divide
p by
o and let c be the
n
5
6
+
i
P
1 1
quotient
and r the remainder : then
-
=
- =
.
By
p p
r
J
1
-
c +
~
<1 q
s. A. 29
450 CONTINUED FRACTIONS.
ffl
proceeding
in this
way,
we find in the
required
form,
m
p
1 11
namely
=
a
+
-
=
a
+
=
a
+
T
-
....
J
n n
7 <7
o
+c +
o
+
i
P
Since the numbers
p, q,
r,
... become
necessarily
smaller at
every stage,
it is obvious that one of them
will sooner or later become
unity,
unless there is an
exact division at some earlier
stage,
so that the
process
must terminate after a finite number of divisions.
It should be noticed that the
process
above described
is
exactly
the same as that for
finding
the G.C.M. of m
and
n,
the numbers
a, b, c,
...
being
the successive
quo-
tients. On this account the numbers
a, b,
c &c. in the
continued fraction a
+
r
-
are often called the
b
+ c +
first, second, third,
&c.
partial quotients.
It is
easy
to see that the continued
fractions,
found as
above,
for and
-y ,
where k is
any integer,
will be the
n nfc
same.
491
Ex. Convert
-^r^
and 3*14159 into continued
fractions,
and find in
71 855
each case the fourth
convergent.
Ans. =
,
-
.
177 llo
\
357.
Properties
of Con
vergents.
Let the continued
fraction be a.
H
,
and let denote the nth
a
8 +a,+ qn
convergent
L From Art. 355 we have
CONTINUED FRACTIONS. 451
So also in succession
But
P& -p&
=
(at
a
z
+
1
)
-
ap,
=
1.
Hence
p.g^ -pn
_
t gB
=
(-!)"
........................
(i).
Hence also
fc_5eiLJ
...().
9. 9-1 J.^l
COR.
In the continued fraction
-
h...,
which is
,+ <*,
less than
unity,
we have
PnQm-P i?.
=
(- i)""
1
and
8f-fit==t-i-.
?. ?-i 9.9^.
II.
Every
common measure of
p
n
and
<?
must also be
a measure of
pn qn.^ p
n
-iq
n
,
that is. from
I.,
a measure of
+ 1. Hence
p
n
and
q
n
can have no common measure.
Thus all
convergents
are in their lowest terms.
III. If F= a
:
+
- -
. .
;
then F will be
a
2
+
a
s
+ +a
n
+
obtained from the wth
convergent by putting:
-
+*+!
+
in the
place
of .
+
+ ""^
^
Hence F
where \ is written instead of
-
.... so that \ is some
a
n-;.l
+
positive quantity
less than
unity.
292
452
CONTINUED
FRACTIONS.
Hence F
=
9.
Also F
..-*=-
Now \ is less than
1,
and
qn
is
greater
than
qn-l
;
hence
^
~^?
is less than F-
p
-*=*
.
9. ?-!
Thus
any convergent
is nearer to the continued
fraction
than the
immediately preceding convergent,
and
therefore
nearer than
any preceding convergent.
oc
IV. If
any
fraction,
-
suppose,
be nearer to a continued
\s
3C
fraction than the nth
convergent,
then
-
must from III. be
c/
also nearer than the
(n l)th
convergent;
and,
as the
continued fraction itself lies between the nth and the
fl9
(n
l)tlj
convergents [Art. 353],
it follows that
-
must
y
also lie between these
convergents.
Hence
fe -
-
must be <
1>
"--> - &
;
</.! y ?-, g.
e >
? p.., y
~
q^x.
Hence,
as all the
quantities
are
integral, y
must be
greater
than
qn
.
CONTINUED FRACTIONS. 453
Thus
every fraction
which is nearer to a continued
fraction
than
any particular convergent
must have a
greater
denominator than that
convergent.
V. We have seen in III. that
where X is a
positive quantity
less than
unity.
Hence F~
P
^>
-
^
-
x
;
ff-i
V**(S+9
r
J
also ^~h,'<_!_.
Thus
any convergent
to a continued
fraction differs
from
the
fraction itself by
a
quantity
which lies between
-j-j
and
-j7-j
-
j-v ,
where d. and cL are
respectively
the
d
A
d
i (i
+
2)
denominators
of
the
convergent
in
question
and the next
succeeding convergent.
Ex. 1. Shew
that,
if
pr /qr
be the rth
convergent
to the continued
fraction a,
+ .....
,
then will
03
+
03+
+o'
For we have
pn
=
a
npn
_
l
+
Pn
_
2 ,
2>n-l
=
"n-ll'n-a +Pn-3
p
p
1 1
Hence
-
=
a
_+^-
3
=
a_
+
-
=a_
+
-
Pn-l
*
Pn-l
'
_ . Pn-3
-
''ti i +
-
11 1
=
,+
- -
....
'
a
n-l
+
n-3
+
111
.
a
s
+
a,
+
<!
454 EXAMPLES.
It can be
proved
in a similar manner that
?
1 1
-5-
Ex. 2. To shew that
-^
=-
... to n
quotients,
where n is
i ^ a
~
a
positive integer.
We have
n 1 n-1 1
,21
^
-
...... to n
quotients.
2
Ex. 3. Shew
that,
if
pr jqr
be the rth
convergent
of
- -
;
EXAMPLES XXXVL
f
1. Shew
that,
if
*-*
, ,
be three successive
convergents
Vi
q,
ft
to
any
continued fraction with unit
numerators,
then will
P-Pi
:
?
a -Vi=P*
'
?,-
2. Shew
that,
if
be the nth
convergent
of
^
*
^
q,
b
t +6. +6.
+
...;
then will
----!'''^
3. Two
graduated
rulers have their zero
points
coincident,
and the 100th
graduation
of one coincides
exactly
with the
63rd of the other: shew that the 27th and the 17th more
nearly
coincide than
any
other two
graduations.
EXAMPLES. 455
4. Shew
that,
if a
t
,
a.
2 ,
......
,
a
B
be iii harmonical
progression
;
11
all 1
(t,
then will
a-
=
-
^
...
j,
-.
._,
2-2- -2-
a,
5. Shew that
111 111
...=w{a,
+
-5 -s ...
J. z
6. Shew
that,
if P=
-
\
-
...
.-
a +b +c
+ ++1
a 6 c A
and
V=T
j
"-
r'}
o + G +d + + k'
then will
P(a+
Q+ l)
=
a +
Q.
7. Find the value of
n n-l n-2 2
1^
1
n +n 1 +n 2-t-*" +2 + 1+2*
8. Shew
that,
whether n be even or
odd, v T T
'1 4 14
2n
to y*
quotients
= --
.
n+1
9. Prove that the
ascending
continued fraction
6
a
+
b,
+ b + . ,
,
b. b b
s
... is
equal
to
-J
+
'
+
-
+ ...
a a a a aa
10. If
JOB
be the numerator of the nth
convergent
to the
fraction
^-
=?
j
5
.... shew that a linear relation connects
6,
+ b
a
+ b
3
+
every
successive tour of the series
pf, pj
1
, p *,...;
and find
what the relation is.
11. If
pJq.
be the rth
convergent
of
-
r
-
T ...,
a +b +a + b +
shew that
p
2it+a
=p
an
+
bq
tu
,
and that
q
sn+i
=
ctp
2u
+
(ob
+
l)</2)i
.
456
EXAMPLES.
12. If
pjqr
be the rth
convergent
of the continued fraction
111111
L. -/k. 11
-
T
~ ~
r
~
>
shew that
=bp, +(oc+l)<7_.
a +b +c +a+b +c +
13. If
pjqr
be the rth
convergent
of
y ^
z-
-
...,
shew that
p
tn
q
tn
_,
-
qln p
vt
_
l
=
-
a"b\
ip
14. Shew
that,
if be the nth
convergent
to the continued
M
fraction
;L LLL
'
then
9 =p
^
and
bq
>^
=
a
^
+
abp
^-
15.
Shewthat,
if^rra +
i -
...i
\;
then will
5
ft-^c+ k + I
11 11
1^1
1 1 1
=
J_
I +k +
'"
+c +b +a I +k +
'"
+c +b
pq'
p
16. Shew
that,
if
-^
be converted into a continued
fraction,
H/
p
the first
quotient being
a,
and the
convergent preceding
-^
being
-
; then,
if be converted into a continued
fraction,
the last
convergent
will be
(P
-
aQ)/(p- aq).
fry
fft
17. Shew
that,
if
-
and
^
be
any
two consecutive couver-
gents
of a continued fraction
a;,
then will
,
^
x*
according
as
P>t
q<q"
3-)8. To
find
the nth
convergent of
a continued
fraction.
We have in Art. 355 found a law
connecting
three suc-
cessive
convergents
to a continued
fraction,
so that the
GENERAL CONVERGENT. 457
convergents
can
always
be determined in succession. In
some cases an
expression
can be found for
any convergent
which does not involve the
preceding convergents
: the
method of
procedure
will be seen from the
following
examples.
Ex. 1. To find the nth
convergent
of the continued fraction
1 1.3 3J5
5_/7
3+4+4 +
4
+"
LI. i A (2n
-
3) (2n
-
1)
Here the nth element is
'-^
'-
. and therefore
4
P*=*P-i
+
(2n
-
3) (2n
-
1) pn
_
.
The above relation
may
be written
Pn
-
(2n
+
!)*_,=
-
(2n
-
3) {p^
-
(2n
-
l)pn
_
2}
.
Changing
n into n
-
1 we have in succession
P-i
~
(2
-
l)l,-3=
-
(9n
-
5)
{i>n
_
2
-
(2n
-
3) pn
_
s},
ft~7ft=-3{ft-5ft}.
But,
by inspection, p=l,
P
2
=4;
..
ft-5ft=-l.
Hence
pn
-
(2n
+
1)
p
B_,
=
(
-
l)-i (2n
-
3) (2n
-
5)
... 3 . 1.
Then
again
ft. Pn-l
_A-^T
l
_
1.3...(2n
+
i)~1.3...(2n-l)~(2n+l)(27t-l)'
ft
1
and
1.3.5 1.3~~3.5~'
P,
1
1.8 1.8'
1 1
Since the denominators of
convergents
are formed
according
to
the
same law as the
numerators,
we have from the above
since
q
=
3 and
g
=
15.
458 GENERAL
CONVERGENT.
Hence
('2n
+
l)(2n-l)...3.1 (2n-l)...3.1
5.3.1 3.1
.-.
gn =1.3...(2n-l)(2tt
+
l).
Hence
pn/qn,
the nth
convergent required,
is
_1_
1
(
-
1)*"
1
1.8 3.5"
1
"
~*"(2;<
Ex. 2. To find the nth
convergent
of the continued fraction
1234
1
+
2
+
3
+ 4+
"
The
necessary
transformations are
given
in Ex.
5,
Art. 251.
1 1
l-l)*'
1
It will be found
thatj>w
=-,
- -
r
:
J
+ +
*-
-r
,
\&
\O
]tt
-f"
A
11
(
-
1)
and that
9=*--7n
+15
~
+
1
5~
I a
JO
\tl
~h
X
If n be
supposed infinitely great
359. Periodic continued fractions. When the
elements of a continued fraction
continually
recur in the
same
order,
the fraction is said to be a
periodic
continued
fraction
;
and a
periodic
continued fraction is said to be
simple
or mixed
according
as the recurrence
begins
at the
beginning
or not.
lillli
t+
1111
is a
simple,
and
a+b+b +b+
is a mixed
periodic
continued fraction.
GENERAL
CONVERGENT. 459
360. To
find
the nth
convergent of
a
periodic
continued
fraction
with one
recurring
element.
Let the fraction be a
+
- - -
...
Then,
for all
C+C+C+
convergent^
after the
second,
we have
p
n
=
(//>_,
+
bp
n_.
2
where b and c are
constants,
that
is,
are the same for all
values of n.
Now,
if
MJ
+ u
i
x
+
u
3
x*+...+u
i
x"~
1
+ ...'be the
recurring
series formed
by
the
expansion
of
-
7=,
the suc-
1
-
ex
-
bx*
cessive coefficients after the second are connected
by
the
law u
n
=
cu
n
.
1
+
bu
n
_
t
.
Hence,
if A and B are so chosen
that u
l =pl
and u
3 =pz
,
then will u
n pn
for all values of
n. The
necessary
values of A and B are
respectively pl
and
p^ cplt
that is a and 6.
Hence the numerator of the nth
convergent
to the
continued fraction a
+
-
... is the coefficient of x
n
~
l
in
c + c
+
f
a
+
bx
the
expansion
of
^
~~ij
J.
"
CSC
*~
0*X/
Similarly
the denominator of the nth
convergent
is
the coefficient of a;*"
1
in the
expansion
of
-
^2
j~
,
X
"*
CJ&
^~
u(JO
that is of
,
1 ex bx*
'
Ex. 1. Find the nth
convergent
of the continued fraction
333
+
The numerator of the nth
convergent
is the coefficient of x
re
~
J
1
+ 3* 3 1
in the
expansion
of
^-
_
a a
*"** 1S
*
oil S~l
~
o , o
Hence
i>=^3
+
(-!)}.
Also
gn
=
coefficient of x
n-1
in the
expansion
of
1 3
1
-
2x
-
3x
3
~
4(1
-
tec)
4
(1
+
x)
'
460
GENERAL
CONVERGENT.
Hence
?H
=i
{8- (-I)"}.
Thus the nth
convergent
is 2
la^i^ii*'
Ex. 2. Find the nth
convergent
of the continued fraction
a c a c
b+d+b+d+
""
We have
P2n
=
dp
P-tn-l
=
&P
Hence,
eliminating
j>2n-i
&n&
Fan-si
we have
=
0.
Since the last result is
symmetrical
in a and
c,
and also in i
and
d,
it follows that
Pfci-l
-
Hence the relation
pn
-
(a
+
c
+
bd) pn
_
2
-i-
ocpn_4
=
holds
good
for all values of n.
Hence
pn
will be the coefficient of a:*"
1
in the
expansion
of
provided
the values of A
, B, G,
D are so chosen thr.t the result holds
good
for the first four
convergents.
It will thus be found that
pn
is
the coefficient of x*'
1
in the
expansion
of
a
+
adx
-
acx
1
It will
similarly
be found that
gB
is the coefficient of x""
1
in the
expansion
of
)x-acx*
361.
Convergency
of continued
fractions.
When
a continued fraction has an infinite number of elements it
is of
importance
to determine whether it is
convergent
or
not. When an
expression
can be found for the nth con-
CONVERGENCY
OF CONTINUED FRACTIONS.
461
vergent,
the rules
already investigated
can be
employed;
the nth
convergent
cannot however be often found.
In the continued fraction
^ j-
s
j-
2
... it is
easy
to
\
+
b
t
+
6
8
+
shew,
as in Art.
357,
that
and hence that
,a,a ...a_
rn
5|
t
*l"'2
,
__
/ -I \i
r
...
-f
v. *)
Now,
if all the letters are
supposed
to denote
positive
quantities,
the terms of the series on the
right
are
alternately positive
and
negative;
also each term is less
than the
preceding,
for the ratio of the rth term to the
preceding
term is
f
^
r"
8
,
which is less than
unity
since
7
r
=
6
r ^
r
_
1
+ a
r
5
r
r_
2
. Hence the
series,
and therefore the
continued
fraction,
is
convergent provided
the nth term
diminishes
indefinitely
when n is
indefinitely
increased.
It can be shewn that the condition of
convergency
is
satisfied whenever the ratio
&>_!
: a
n
is finite for all
values of n*.
For let
&&_,
be
always greater
than k . a
w ,
where k is
some finite
quantity.
Thenw
=
b
But
Hence
Whence u<
2&
Todhunter's
Algebra,
Art. 783.
462
PERIODIC CONTINUED FRACTIONS.
But
(1
+
&)""*
increases
indefinitely
with
n,
since k is
finite;
hence u
n
decreases without limit as n is increased
without limit.
We have therefore the
following
Theorem. The
infinite
continued
fraction
j^ -^
,
?
.
..,
in which all the
letters
represent positive quantities,
is
convergent if
the ratio
b
n
b
n_!
: a
n
is
always greater
than some
fixed finite quantity.
It should be remarked that
any
infinite continued
fraction of the form a
+
T
-
.... in which
a, b, c,...
b
+ c
+
are
positive integers,
is
convergent
362. In the
following
five Articles the continued
fractions will all be
supposed
to be of the form
a
+
j-
...,
where
a, b, c,...
are
positive integers.
T
C
T
This form of continued fraction
possesses
two
great
advantages,
for we know that
every convergent
is in its
lowest
terms,
and we can also see
by inspection,
within
narrow
limits,
the difference between
any convergent
and
the true value.
363. Theorem.
Every simple periodic
continued
fraction
is a root
of
a
quadratic equation
with rational
coefficients
whose roots are
of contrary signs,
one root
being
greater
and the other less than
unity.
Also the
reciprocal
of
the
negative
root is
equal
in
magnitude
to the continued
fraction
which has the same
quotients
in inverse order.
Let the fraction be
1 1 1 1 1 1
b + c +'"
P P
Let
-jy
and
-^
be the last two
convergent^
of the first
*
Articles
363, 364,
and 3GS are taken from a
paper by
Gerono,
Nouvellet Annale* de
Mathtmatiquet,
1. 1.
PERIODIC CONTINUED FRACTIONS. 463
period.
Then
<rp+>
"
The roots of
(i)
are
obviously
of different
signs,
and the
positive
root is the value of the continued fraction.
Now,
from Art.
357,
Ex.
1,
P__
1 11
P'~
+
k+'"+b+a,'
01 1
and
7w=t + T.
...
.
T
+
o
TT 't 1
1 111
Hence,
if
y
=
I
+
T
... ....
k
+ +
a
+ l+k +
Py
+
Q
we have
y
=
*
,
n
,
5
(ii).
The roots of
(ii)
are
obviously
of different
signs,
and the
positive
root is the value of the continued fraction
1 11
+
k+
'"
+ a + J +
""
From
(i)
and
(ii)
we see that the
positive
root of
(ii)
is
equal
in
magnitude
to the
reciprocal
of the
negative
root
of
(i);
and therefore the
reciprocal
of the
negative
root of
,.;./,!
11
\
(l)
IS
Z+
y-
...
-j
....
\
k+ +a+
I + I
The
positive
roots of
(i)
and
(ii)
are both
greater
than
unity,
as
may
be seen
by inspection;
the
negative
root of
(i)
must therefore be less than
unity.
The fraction
-
. .
requires
no
special
4 6+ +
I
+a +
464
PERIODIC CONTINUED FRACTIONS.
examination,
for we have
only
to
change
x into
-
,
and
y
iuto
-
;
thus
-
r ,
...
,
... is
equal
to the
y'
a+b+ +k+ 1+
a 4-
positive
root of PV
(Q'P)
x
Q
=
Q,
and the
negative
f,
1
11)
root is
-U +
y ,
...
,
T
,
-4.
[
k+ +
b
+
a)
Hence,
as
before,
one root of the
quadratic equation
in
x is
greater
and the other is less than
unity.
364.
Theorem.
Every
mixed
periodic
continued
fraction,
which has more than one
non-periodic
element,
is a
root
of
a
quadratic equation
with rational
coefficients
whose
roots are both
of
the same
sign.
Let the fraction be
i
+
'"
and let
i
1
+
'
A' A
Let -~ and
-^
be the two last
convergent^
of the non-
Ij Jo
periodic part
;
then
_
yA
+ A'
([}
'yB
+
ff
-
P P
Let
jy
and
-^
be the last two
convcrgents
of the first
period
of
y
;
then
The elimination of
y
from
(i)
and
(ii)
will
clearly
lead
to a
quadratic equation
in x with rational coefficients.
Now,
if the
positive
root of
(ii)
be substituted in
(i)
we
REDUCTION OF
QUADRATIC
SURD. 465
shall
clearly
obtain a
positive
value of
x,
and this will be
the actual value of the
given
continued fraction.
Also,
from the
preceding
article,
the
negative
value of
-is
\v +
-V
; and,
if this value be substituted
y (
/*+ +J
in
(i),
we have
1 1 1
+
b+~' +
k-v
-^
+
- :
we have to shew that this is
positive.
If k > v the result
is
obvious; ifk<v,
;
-
... is
negative
but is less
k
-
V
-
fJL
+
than
1,
and
therefore x is
positive provided
one element at
least
precedes
Jc
;
also k cannot be
equal
to
v,
for in that case
the
periodic part
would
really begin
with k and not with
a. Hence both values of x are
positive
in all cases.
REDUCTION OF
QUADRATIC
SURDS TO CONTINUED
FRACTIONS.
365. It is clear that a
quadratic
surd cannot be
equal
to a continued fraction with a finite number of elements
;
for
every
such continued fraction can be reduced to an
ordinary
fraction whose numerator and denominator
are
commensurable. It will be shewn that a
quadratic
surd
can be reduced to a
periodic
continued fraction of the form
a +
r.
,
>
where
a, b, c,
... are
positive
integers.
T C +
The
process
will be seen from the
following example.
Ex. To reduce
/8
to a continued fraction.
The
integer
next below
^/8
is
2;
and we have
4 1
4
/H 4- 2
The
integer
next below
<v
j
is 1
;
and we have
S. A.
30
466 REDUCTION OF A
QUADRATIC
SURD
_ ,-_
__ _
"
+
~
L *
4 4
The
integer
next below
,^/S
+
2 is 4
;
and we have
=4
+_L_.
4
The
steps
now
recur,
so that
J8=2+J . \ . \ .
i
Thus
^/8
is
equal
to a
periodic
continued fraction with one non-
periodic element,
which is half the last
quotient
of the
recurring
portion;
and it will be
proved
later on that this law holds
good
for
every quadratic
surd.
366. We now
proceed
to shew how to convert
any
quadratic
surd into a continued fraction.
Let
^/N
be
any quadratic
surd,
and let a be the
integer
next below
*JN;
then
Na* 1
w
-
=
a
+
-
n
r
r
-
,
if
N+
a
jN+a
'
r
i
where
r,
=
Ncf.
Since
JN
a, is
positive
and less than
1,
it
follows
that
- -
is
greater
than 1. Let then b be the
integer
.
,
,
V^V"
+ . i
next below
;
then
T
N a*
where
a,
=
6r,
a and
r,
= *-
.
Then,
as
before,
*
is
greater
than
unity
;
and if
TO A CONTINUED FRACTION. 467
c be the
integer
next below
^
*,
we have
=
c f
r
=
cH
,
,
,
7 .
*
^r
=c
+
i a*
where
a.
=
cr
8
a. and
r,
=
-
3
.
t
n*
'
2
The
process
can be continued in this
way
to
any
extent
that
may
be desired. Thus
*/N=
a
+
r
-
o+
c
367. To shew that
any quadratic
surd is
equal
to a re-
curring
continued
fraction.
It is first
necessary
to
prove
that the
quantities
which,
in the
preceding
Article,
are called
a,
a
2
,
a
a
,...,
r
lt
r
t
,
r
a
,...
are all
positive integers.
It is known that N is a
positive integer,
and that
a, b,
c, d,
... are all
positive integers.
We have the
following
relations :
r^X
-
8
,
............
(i);
a,
=
6-
1
-a
,
r
J
r
2
=
iY-a
2
2
.........
(ii) ;
03
=
^-0,,
r
2
r
s
=
iV-a
3
2
.........
(iii);
a
4
=
dr,-o8
,
r
3
r
4
=
J\T-a
4
8
.........
(iv);
and so on.
Now it is obvious from
(i)
that r
l
is an
integer.
From
(ii)
we have
r,
=
-- !-
=
1
+ 2a6 6V .
^ * *
'
i
since Na*
=
r
l
.
Thus a
2
=br
l
a,
and r
2
=
1
+
2ab 6V
t
;
whence it
follows that a
a
and r
s
are
integers,
since
i\
is an
integer.
302
468 REDUCTION OF A
QUADRATIC
SURD
From
(iii)
we have
similarly
a
s
=
cr
2
a
s
and r
8
=
r,
+ 2a
2
c
-
cV
s
;
whence it follows that a
3
and
r,
are
integers,
since a
s
and
r
s
are
integers.
Then
again,
from
(iv)
we have
a
4
=
dr
3
a
a
and r
4
=
r
s
+
2a
8
d
-
cV
8
;
whence it follows that a
4
and r
4
are
integers,
since a-
3
and r
8
are
integers.
And so
on;
so that a
n
and r
n
are
integral
for all values
of n.
We have now to
prove
that a
n
and r
n
are
positive
for
all values of n.
We know that
a, 6, c,
&c. are the
positive integers
. ,
r V
'iV
'4-
a
JN+a
9
next below
>JN,
-
,
--
*,
&c. Hence
Jjya,
T r
i a
JN
a
s
, */Na
3
, &c.,
and therefore also N
a*,
N
a*,
Na*, &c.,
are all
positive.
That is r
1(
r
2
,
r
s
,
&c. are all
poit^'we.
Again,
since 6 is the
integer
weatf 6e/ow
-
,
it
**i
follows that
\/N+a<br
1
+
r
1
.
Now,
a cannot be
equal
or
greater
than br
v
for then
j2F<r
lt
and therefore a<r
1 ;
therefore a
< br
l
,
since r
l
is
positive
and b a
positive
integer.
Hence a<br
l
,
so that a
s
is
positive.
Again,
since c is the
integer
next below
---
*,
it
-
follows that
>JN+a
t
< cr
g
+
r . And we cannot have
a
f
>cr
s
,
for then
^N<r
t
,
and therefore a <r
a
<cr
s
,
since
r. is
positive
and c a
positive integer.
Thus
a,
< cr
t
,
so
that
8
is
positive.
And so
on;
so that a
n
is
positive
for all values of n.
Having
shewn that the
quantities
r
n r,,
r
s
,
&c. and
a,
a
a
,
a
8
, &c.,
are all
positive integers,
it follows from the
TO A CONTINUED FRACTION. 400
relation r
B
r
n.j
=
N a* that a
n
is less than
*/N,
so that
a,,:j>a;
hence the
only possible
values of a
n
are
1, 2,...,
a.
Then,
from the relation a
n
+
a
n+1
=
k. r
n>
where k is a
positive integer,
it follows that r
n
cannot be
greater
than 2a.
Hence the
expression
"
cannot have more than
'
n
2a x a different values
;
and therefore after 2a
2
quotients,
at
most,
there must be a recurrence.
368. Theorem.
Any quadratic
surd can be reduced
to a
periodic
continued
fraction
with one
non-recurring
element,
the last
recurring quotient being
twice the
quotient
which does not
recur;
also the
quotients of
the
recurring
period,
exclusive
of
the
last,
are the same when read back-
wards or
forwards.
Let
*JN
be the
quadratic
surd.
Then,
from the
preceding
Article,
we know that
*/
N
is
equal
to a
periodic
continued fraction.
We also know that
any periodic
continued fraction is
equal
to one of the roots of a
quadratic equation
with
rational coefficients
;
and the
only quadratic equation
in x
with rational coefficients of which
\J
N is one root is the
equation
a? N
=
0.
Now the roots of a? N=Q are both
greater
than
unity
in absolute
magnitude,
and the roots are of
different
signs
: it therefore follows from Articles 363 and 364 that
the continued fraction which is
equal
to
*JN
must be a
mixed
recurring
continued
fraction
with one
non-recurring
element.
Hence we have
-
4-11
*b+c+
'
11 1111
^-a=
-... r-
470 SERIES EXPRESSED AS CONTINUED FRACTIONS.
11 1111
Now
T
...
, r ,- , 7 r ,
1S tne
positive
6
+
c + +h+k+
I
+b+
root of a
quadratic equation
with rational
coefficients;
and as this
positive
root is
JN a,
the
negative
root
must be
*/Na. Hence,
Art.
363,
we have
I
+
k + h
+
'"
+c
b l
+
"
Ij.
1
-'
1
Ill
=
L
+
T r T T
k
+
h
+ + c+ b+
I
+
II 111
Hence I a
+
~. ...
-
-j
k+h+ + c+
b
+
I
+
11
111
6
+ c+
"'
+h + k+
1
+
'"'
whence it is
easy
to see that I a
=
a,
k
=
b,
h
=
c,
....
SERIES EXPRESSED AS CONTINUED FRACTIONS.
369. To shew that
any
series can be
expressed
as a
continued
fraction.
Let the series be
u
l
+
u
i
+
u
a
+
u^+
... +u
n
+ ...............
(i).
Then the sum of n terms of the series
(i)
is
equal
to
the nth
convergent
of the continued fraction
This can be
proved by
induction,
as follows.
Assume that the sum of the first n terms of
(i)
is
equal
to the nth
convergent
of
(ii).
Another term of the series
is taken into account
by changing
u
n
into u
n
+
u^
; and,
7/ U
by changing
u
n
into u
n
+
u^.
lt
"~* *
will become
SERIES EXPRESSED AS CONTINUED FRACTIONS.
471
U
n-3 (
U
n
+
W
+l) I' U
'
'I 1. '
,
which is
easily
seen to be
equal
to
U
n
_, +
U
n
-}-U
n
,
,
It tjL U 1L
**~1 "
+1
.
Thus the sum of
n+1
terms of
(i)
will be
equal
to the
(n+l)th convergent
of
(ii) provided
the sum of n terms of
(i)
is
equal
to the nth
convergent
of
(ii).
But it is
easily
seen that the theorem is true when
n is 1 or 2 or 3 : it is therefore true for all values of n.
Thus u
l
+
u
a
+
u
a
+
u
4
+
...to n terms
^
_u^_ u^ jy^
.. t otients ...
[A]
.
-
It can be
proved
in a
precisely
similar manner that
MJ
w
g
+
u
a
w
4
+
to n terms
* A>
U U 1b It/
*-5 s-i-
... to n
quotients
... FBI.
-
u
+
u u
+
The formula
[B]
can also be deduced from
[A] by changing
the
signs
of the alternate terms.
370. The
following
cases are of
special
interest :
a a a a aa
r-
1
+
-r^r*
+
'
%
3
. . . to n terms
. . to n
quotients.
. .
[C],
all the
upper signs,
or all the lower
signs, being
taken.
And +
|
+
l-
...to n terms
a a
4
*|
3
_ ... to w
quotients
. . .
[D],
d + CL + <Z + CL
1
'
2 1
'
3 i
-
all the
upper signs,
or all the lower
signs, being
taken.
These can be
proved by
induction as in the
preceding
Article.
4-72 SERIES EXPRESSED AS CONTINUED FRACTIONS.
Thus to
prove J^C].
It is obvious
by inspection
that the theorem
is true when n 2. Assume then that
[C]
is true for
auy particular
value of n
then,
to include another term of the series
-^
must be
"
changed
into
7^ .
" n+*
,
and therefore
-1 n
will become
which can
easily
be seen to be
equal
to
*-1 * * n+1
.
Thus,
o
n
a
n
=p
o^.,
a
n+1
if
[C]
be true for
any
value of
n,
it will be true for the next
greater
value;
hence as
[C]
is true when
n=2,
it is true for all values of n.
The
following
are
particular
cases of
[C].
a,
+
a,a,
-f-
c^a/i,
+
a,
a,,
-IU
-
f
...
ctj Ojd;, ttiCf2
of
3
a
l
a
i
ci
3
d
i
EXAMPLES.
473
Ex. 3. Find the value of
--
... to
infinity.
1+1 +2+3+4+
From
[F]
we see that
1123
"I -17!2
+
r7278-f7270
+ "
toinfinity=l-<ri.
133
Ex.4. Find the nth
convergent
of
-
....
o
+
2
+
2,
+
From
[F]
we have
!__!_
1
=
1 3 3
3 3.3
-
Hence the ?ith
convergent required=-
-jl
-
(
-
^ )
L .
r r 2r 3r
Ex.5. Shew that 1
+
T
-
-.
...
=
c
r
.
1
-
r
+
2
-
r+3
-
r
+
4
-
r_i
r r-r
+
172 1T23 r7
r r 2r Br
EXAMPLES XXXVII.
1. Find the continued fractions
equivalent
to the
following
quadratic
surds :
(1) Vl?,
(3) V1
40
>
(
3
) N/33, (4)
(5) ^(a'+l), (6) N/(a
2
2. Shew that
JN=
a +
=-
-^-
....where a has
any
value
2a + 2a +
whatever,
and b
=
N-a*.
3. Find the value of
1111
474 EXAMPLES.
ffll,
1 1 1 I i i
... to
fataitj.
^ '
2+3+4+5+4+5+
4. Shew that
1 1
+
__
..to
infinity
=
5
(1+jj
... to
infinity).
14 + 14+
\
A + A +
5. Shew that
11111
\ / . 1 1 1 1
----,-
-
)(
+
-
T j
a+ j +c+ d+a+
/\
c+6+a+a+
6 + d+ bed
~
a + c +
acft
*
1
1
6. Shew
that,
if x
=
y+<
r
... to
infinity,
then
&y
+
Zy
+
v
=
x
--
x
to
infinity.
2
-
2x
-
7. Shew
that,
if
1111
x
=
-
r
-
r
... to
infinity,
a + b +a +o +
then will
8. Shew
that,
if n be
any positive
integer,
n*-T n-2*
n'-y
~T~+~5~+
7 +'
9. Shew that
a 1 I ...
a +
a +
a a
to n
quotients.
EXAMPLES.
475
10. Shew
that,
'if
a c a c c a e a
X
=
T ~j T ~j
and
y
-i
r -5 r
b+d+b+d+ d+b+d+b+
then bx
dy
=
a c.
11. Shew that the ratio of
1111 1111
T
r
T
... too+
T
-
T T
is 1 + a :
12. Shew that the nth
convergent
of
1422
. 2'-l
3-3-3-3-'"
1S
13. Shew that the ?ith
convergent
of
i i i
"
2+2+2 +
"
(1
+
^2)-
-
(1
-
1222
14. Shew that the nth
convergent
of
y ^ -^ ^
15. Shew that the nth
convergent
of
1 ab ab a"
-
b'
a + b -a +
b -a + b
-"
a
n+l
-b
n+l
16. Find the nth
convergent
of the continued fraction
238 r
8
-!
1 -5-7-
"
-27T1-""
'!) 7}
17. In the series of fractions
,,
&c.. where the law
?i
<?
of formation is
pr
=
qr_i,
qr
=
(n
3
l)^r_i
+
%<Ir-i> pi'ove
that
jo
1
the limit of when r is
infinitely
great
is
^
.
qr
l+n
476 EXAMPLES.
18. Shew that in the continued fraction
a, Oj Og
6",
+
&,
+
&,+
'"
+b,
+
"
the nth
convergent
is to the
(n l)th
in the ratio
,
a
, ,
3
6.-t-
7
-a-
...
r
5
: o +
r-*
1-
r
1
-
6
.-,
+ + 6
*
6
.-i
+ + 6
i
19. Shew
that,
if
y
lt
yy
, <fec.,
be the
convergents
of a
simple
periodic
continued fraction found
by taking
1, 2, <fec.,
complete
P P'
periods,
and if
-^,
-^>
be the two
convergents immediately
Q Q
P'y ,+
p
preceding y
lf
then
yx
=
^J
+
^
.
20. If Z be
any integer
not a
perfect square,
and if
,JZ
be converted into a continued fraction
11 111
fm
J^
_
I
b + c +
' '
+ fc + 2a +b
and if the
convergents
obtained
by taking
one,
two, ...,
i
complete periods,
each
period terminating
with
k,
be denoted
byP,,
P
t>
...,
P
t
,
prove
that
21. Find the nth
convergent
of the continued fraction
1 1 1
a -r a~ a + a~
l
a + a"
-
and shew that the limit of the wth
convergent
when n is
indefinitely
increased is a or a"
1
according
as a is
numerically
less than or
greater
than
unity.
22. Shew that the nth
convergent
of
1233 . 3f
/lyi
2+2+2+24-
- 1S
8|/ V
-V
T
EXAMPLES. 477
23. Shew that
1 03 x 1x
(n-l)x
.
,. .,
5
=
T i ^
...
-
... to
infinity.
1 + 1 x + 2-oj + 3-a: + + n x +
24. Shew that
x x
3
x
3
_x
ax bx
a ab abc a+b x + c a? +
"
25. Shew that
11246
_i
... to
infinity
=
e
2
.
1+1+3+5+7+
26. Find the value of
1136
3(n-2)
-=
o
... ... to
infinity.
1+2+5+8+ + 3n
-
4 +
27. Shew that
1
2
.3
2
2
2
.4
2
3
2
.5
2
(n-l
3 +
~5" + 7 + ~~9~ +
"*
+ 2+l
1.3 2.4 '3.5 r
28. Shew that the nth
convergent
of
2222 2"-l
29. Shew that the nth
convergent
of
1414 .
6n-l +
(-!)"
3_3_3_3_
ls
e + 7 +
(-!)'
I
2
2
2
3
s
30. Shew that
-K-
^
i=-
... to
infinity
=
1.
o D /
"
31. Shew that
1 1.2 3.4 5.6
478
EXAMPLES.
32. Shew that
j__l
2.3 4.5 6.7
~72
"2+1 + 1 + 1 +
33. Shew that
1
3^4 5^6 TJi
9.10
34. Shew that
(1 +)"
nx
(n-l)x 2(n-2)x
3
(n
-
3)
x
35. Shew
that,
if n be a
positive integer,
1
o
/ o\
q
/ o\ /
1 N 1
1 n + 1- Ti+1
36. Shew that
(l
+
f)(l
+
f,)(l
+)...
a x + a+a x+a+a
37. Shew that
-
s
=
^
= ... to
infinity
n- 2n+l- 371+1-
1 n 2n
=
^
=- =
-v
... to infinity.
CHAPTER XXVIIL
THEORY OF NUMBERS.
371.
Throughout
the
present chapter
the word number
will
always
denote a
positive
whole
number;
also the word
divide will be used in its
primitive meaning
of division
without remainder. The
symbol M(p)
will often be used
instead of
'
a
multiple ofp.'
Definitions. A number which can
only
be divided
by
itself and
unity
is called a,
prime
number,
or a,
prime.
A number which admits of other divisors besides
itself and
unity
is called a
composite
number.
Two numbers which cannot both be divided
by any
number,
except unity,
are said to be
prime
to one
another,
and each is said to be
prime
to the other.
372.
The Sieve of
Eratosthenes. The different
prime
numbers can be found in order
by
the
following
method,
called the Sieve
of
Eratosthenes.
Write down in order the natural numbers from 1 to
any
extent that
may
be
required
: thus
1, 2, 3, 4, 5, 6, 7, 8, 9,
10
11, 12, 13, 14, 15,
16, 17, 18, 19,
20
21, 22, 23, 24, 25, 26, 27, 28, 29,
30
31, 32, 33, 34, 35, 36, 37, 38, 39,
40 &c.
Now take the first
prime
number, 2,
and over
every
second number from 2
place
a dot : we thus mark all
480 THEORY OF NUMBERS.
multiples
of 2.
Then,
leaving
3
unmarked,
place
a dot
over
every
third number from 3 : we thus mark all mul-
tiples
of 3. The number next to 3 which is unmarked is
5
;
and
leaving
5
unmarked,
place
a dot over
every
fifth
number from 5 : we thus mark all
multiples
of 5. And so
for
multiples
of
7,
&c.
,
Having
done
this,
all the numbers which are left
unmarked are
primes,
for no one of them is divisible
by
any
number smaller than
itself,
except unity.
It should be here remarked that if a
composite
number
be
expressed
as the
product
of two
factors,
one of these
must be less and the other
greater
than the
square
root of
the
number,
unless the number is a
perfect square,
in
which case each of the factors
may
be
equal
to the
square
root. Hence
every composite
number is divisible
by
a
prime
not
greater
than its
square
root. On this account
it
is,
for
example, only necessary
to
reject
as above mul-
tiples
of the
primes
2, 3,
5 and 7 in order to obtain the
primes
less than
121,
for
every composite
number less than
121 is divisible
by
a
prime
less than 11.
373. Theorem. The number
ofprim.es
is
infinite.
For,
if the number of
primes
be not
infinite,
there
must be one
particular prime
which is
greater
than all
others. Let then
p
be the
greatest
of all the
prime
num-
bers. Then
\p
will be divisible
by p
and
by every prime
less than
p.
Hence
\p
+
1 will not be divisible
by p
or
by any
smaller
prime ;
therefore
\p
+
1 is either divisible
by
a
prime greater
than
p,
or it is itself a
prime greater
than
p.
Thus there cannot be a
greatest prime
number
;
and therefore the number of
primes
must be infinite.
Ex. Find n consecutive numbers none of which are
primes.
The numbers are
given by
\n
+ 1
+
r,
where r is
any
one of the
numbers
2, 3, ..., (n+1).
374. Theorem.
No rational
integral algebraical
formula
can
represent prime
numbers
only.
THEORY OF NUMBERS. 481
For,
if
possible,
let the
expression
a bx ca? dx*...
represent
a
prime
number for
any integral
value of x
}
and
for some
particular
constant
integral
values of
a, b, c,
... .
Give to x
any
value,
ra
suppose,
such that the whole
expression
is
equal
to
p,
where
p
is neither zero nor
unity
;
then
p
=
a bm cm* ... . Now
give
to x
any
value m
+
np,
where n is
any positive integer;
then the
whole
expression
will be
a b
(m
+
np)
c
(m
+
np)*
...
=
abm cm* ...
+
M(p)=p+M(p).
Thus an indefinite number of values can be
given
to x
for each of which the
expression
a bx ex* ... is not
a
prime.
In connexion with the above
theorem,
the
following
formulae
are
noteworthy
:
(i)
a:
2
+
x
+
41,
which is
prime
if x < 40.
[Euler.]
,ii)
x*+x +
17,
which is
prime
if x < 16.
[Barlow.J
(iii)
2x
2
+29,
which is
prime
if x < 29.
[Barlow.]
375. The student is
already acquainted
from Arith-
metic with
many properties
of factors of numbers : these
all
depend upon
the
following
fundamental
Theorem :
If
a number divide a
product of
two
factors,
and be
prime
to one
of
the
factors,
it will divide
the other.
For,
let ab be divisible
by
x,
and let a be
prime
to x.
Reduce
-
to a continued
fraction,
and let
-
be the con-
x
q
vergent
which
immediately precedes
-
;
then
[Art. 357,
I.]
30
qapx=l;
.:
qabpxb
b. Now
qab
is,
by supposi-
tion,
divisible
by
x\
and therefore
qabpxb
must be
divisible
by
x,
that is b must be divisible
by
x.
From the above theorem the
following
can
easily
be
deduced :
S.A.
4-82
THEORY OF NUMBERS.
I. If a
prime
number divide the
product
of several
factors it must divide one at least of the factors.
II. If a
prime
number divide a" it will divide a.
III. If a be
prime
to each of
a, y3,
7,
... it will be
prime
to the
product
afiy
IV. If a be
prime
to
b,
a" will be
prime
to 6
m
.
V. If a number be divisible
by
several
primes
separately
it will be divisible
by
the
product
of them all.
376. Theorem.
Every composite
number can be re-
solved into
prime factors
;
and this can be done in
only
one
way.
For,
if N be not a
prime
number,
it can be divided
by
some
number,
a
suppose,
which is neither N nor 1
thus N=ab.
Again,
if a and b be not
primes,
we have
a
=
cxd,
6
=
ex/,
and therefore
N=cdef. Proceeding
in this
way,
since the factors diminish at
every stage,
we must at last come to numbers all of which are
primes.
Thus N can be
expressed
in the form
ax/3x7x8x...,
where
a, /S,
7, 8,
. . . are all
primes
but are not
necessarily
all
different,
so that N
may
be
expressed
in the form
a*$V...,
where
a, ft, %...
are the different
prime
factors
ofN.
Next,
to shew that there is
only
one
way
in which a
number can be resolved into
prime
factors.
Suppose
that N=abcd...
,
where
a, b, c, d,...
are all
primes
but are not
necessarily
all different
; suppose
also
that N
=
afiyS.
. .
,
where
a, y8,
7,
8. . . are also
primes.
Then
we have abed...
=
aj3y&
Hence a divides
a/ityS...;
and
therefore,
as all the letters
represent prime
numbers,
a
must be the same as one of the factors of
<*/3y8
Let
a
=
a;
then we have bed...
=$78...,
from which it follows
that b must be
equal
to one or other of
yS, 7,
B,
. . .
;
and so
on. Hence the
prime
factors
a, 6, c,...
must be the same
as the
prime
factors
a, /9,
7,
....
Ex.
Express 29645,
13689 and 90508 in terms of their
prime
factors.
Am. 5.T
2
.
II
2
,
3 . 13- and 2
2
. IP . 17.
THEORY OF NUMBERS. 483
377. To
find
the
highest power of
a
prime
number
contained in
|n.
Let
/(- )
denote the
integral part
of-: and let a be
\yj
y
any prime
number. Then the factors in In which will be
divisible
by
a are
a, 2a, 3a, ...,/(-).
a.
Thus /
(
-
)
factors
\aj
\a/
(n
\
-j)
factors will
a
I
be divisible
by
a*. And so on.
Hence the whole number of times the
prime
number a
is contained in In is
/(-) +/( )+/ (-5)
+
\aj \a'J \a
9
/
Ex. 1. Find the
highest powers
of 2 and 7 contained in
[50.
Here
,().. ,().* z(|)
=
6,
l(|)=3,
I (
^5 J
=
1. Hence U
47
is the
required highest power
of 2.
Again,
lf-=-J
=
7,
I (
^\
=
l. Hence 7
8
is the
required highest
power
of 7.
Ex. 2. Find the
highest powers
of 3 and 5 which will divide
[80.
Am.
3
s
",
5.
Ex. 3. Find the
highest power
of 7 which will divide
[1000.
Ans. 7
164
.
378. Theorem. The
product of any
r
consecutive
numbers is divisible
by
I r.
Let n be the first of the r consecutive numbers
;
then
we have to shew that
n
("
+
1
)(
m
+
2
> (
+
r-V
|r
\n
+ r 1
or
j
,
is an
integer.
\r \n J.
i/fi
_i_
f
i
The theorem follows at once from the fact that
v
r
v
~
[r [n
1
is
n+r^0r
,
and the number of combinations of n
+
rl
312
4-84
THEORY OF NUMBERS.
things
r
together
must be a whole number for all values
of n and of r.
The theorem can also be
proved
at once from first
principles by
means of Art. 377.
For it is obvious that I
f
71
^!} </()
+
l(-},
\
a J
\
a
)
\aj
I [-
^
J -/( i-)
+
-M-i)
and so on. Hence from
Art. 377 it follows that the number of times
any prime
number is contained in I n
+
r 1 can never be
less,
although
it
may
be
greater,
than the number of times
the same
prime
number is contained in
|w
1 x Ir. Thus
every prime
number which occurs in In 1 x
|r,
occurs to
at least as
high
a
power
in
\n
+
r
1,
which
proves
that
[
n
+
r 1 is divisible
by
[n
1 x
jr.
It can be
proved
in a similar manner that
is an
integer,
where
379.
If
n be a
prime
number the
coefficient of every
term in the
expansion of (a
+
&)" except
the
first
and last
terms is divisible
by
n.
For,
excluding
the first and last
terms,
any
coefficient
,
n(n l)...(w-r
+
1)
,
is
given by
-
p
--
,
where r is
any integer
between and n.
,
T
,
...
n(n l)...(n-r+l)
Now,
by
the
preceding
Article,
-
p
-
is an
integer;
and,
as n is a
prime
number
greater
than
r,
n
*v r (n-l)(n-2)...(n-r+l)
must be
prime
to
r;
and therefore
a- -
r
*
-~
LJ
|r
must be an
integer.
Hence
every
coefficient,
except
the
first and
last,
is divisible
by
n.
Similarly,
if n be a
prime
number,
the coefficient of
THEORY OF NUMBERS.
485
every
term in the
expansion
of
(a
+
6
+
c+...)
w
which
contains
more than one of the
letters,
is divisible
by
n.
For the coefficient
of
any
term which contains more
In
than one of the letters is of the form -.
jJr
-
,
where
a
+/3+y+...=
n. Now .
'
-
is an
integer;
and,
as
n is a
prime greater
than
any
of the letters
a, ft,
7,...,
n must be
prime
to
ja
\@
[7.
. .
;
and therefore the coefficient
of
every
term which contains more than one letter is
divisible
by
n.
Ex. 1. Shew that
n(n
+
l) (2n
+
1)
is a
multiple
of 6.
Ex. 2. Shew
that,
if n be
odd,
(n
2
+
3) (n
2
+
7)
=
M
(32).
Ex. 3. Shew
that,
if n be
odd,
n*
+
4n
2
+
11
=
M
(16).
Ex. 4. Shew that 1
+ T
2"-"
=
M
(8)
.
Ex 5. Shew that 19
2"
-
1 =
M
(360).
Ex. 6. Shew
that,
if n be a
prime
number
greater
than
3,
380. Fermat's Theorem.
If
n be a
prime
number,
and m
any
number
prime
to
n;
then m""
1
1 will be
divisible
by
n.
We know that when n is a
prime
number,
the coeffi-
cient of
every
term in the
expansion
of
(a
t
+
a
2
+...+
a
m)
n
,
which contains more than one of the
letters,
is divisible
by
n. Now there are m terms each of which contains
only
one letter and the coefficient of each of these terms is 1.
Hence,
putting
a
t
=
a
s
=...=
1,
we have
m
n
=
TO
+
M(ri)
;
/. m
(m"'
1
-
1)
=
M(ri).
Hence,
if m be
prime
to
n,
m""
1
1 will be a
multiple
of w.
Ex. 1. Shew
that,
if n be a
prime number,
l-i
+
2-i
+
3-i
+ ... +
(n-l)"-
1
+
l
=
M(n).
Ex. 2. Shew
that,
if a and b are both
prime
to the
prime
number :
then will a*"
1
-
fe"-
1
be a
multiple
of n.
Ex. 3. Shew that n
6
-
n= M
(30).
486 THEORY OF NUMBERS.
Ex. 4. Shew that n
7
-n=if{42).
Ex. 5. Shew that x
lj
-
j/
u
=
M
(1365),
if x and
y
are
prime
to 1365.
Ex. 6. Shew
that,
if m and n are
primes
;
then
m
n
~
l
+
n"
1
-
1
-
1
=
If
(mn)
.
Ex. 7. Shew
that,
if
m,
n and
p
are all
primes ;
then
(np)
m-1
+
(pm)"-
1
+
(mn)P-
1
-
1
=
M
(mnp).
Ex. 8. Shew that the 4th
power
of
any
number is of the form 5m
or 5m
+ L
Ex. 9. Shew that the 12th
power
of
any
number is of the form 13m
or 13m
+
1.
Ex. 10. Shew that the 8th
power
of
any
number is of the form 17m
or 17m 1.
381. To
find
the number
of
divisors
of
a
given
number.
Let the
given
number, N,
expressed
in
prime
factors,
be a
z
6V....... Then it is clear that N is divisible
by
every
term of the continued
product
Hence the number of divisors of
N,
including
N and
1,18
Ex. 1. The number of divisors of
600,
that is of 2
s
. 3 . 5
2
,
is
(3+l)(l
+
l)(2+l)=24.
Ex. 2. Find the sum of the divisors of a
given
number.
The
given
number
being
N=a*b*c?...
,
the sum
required
is
easily
seen to be
(1
-
a**
1
) (1
-
6H"
1
) (1
-
c**)...
Ex. 3. Find the number of divisors of
1000,
3600 and 14553.
Arts.
16,45,24.
Ex. 4. Shew that
6,
28 and 496 are
perfect
numbers.
[A perfect
number is one which is
equal
to the sum of all its
divisors,
not
considering
the number itself as a
divisor.]
Ex. 5. Find the least number which has 6 divisors. Ant. 12.
Ex. 6. Find the least number which has 15 divisors. Ant. 144.
Ex. 7. Find the least number which has 20 divisors. Ant. 240.
Ex. 8. Find the least numbers
by
which 4725 must be
multiplied
in
order that the
product may
be
(i)
a
square,
and
(ii)
a cube.
Ant.
21,
245.
THEORY OF NUMBERS. 487
382. To
find
the number
of pairs offactors, prime
to
each other
;
of
a
given
number.
Let the
given
number be N= a
x
6V...
; then,
if one of
two factors
prime
to
each other contains
a,
the other does
not
;
and so for all the other different
prime
factors.
Hence the factors in
question
are the different terms
in the
product (1
+
a*) (1
+
b
v
) (1
+
c*).
.
.,
the number of
them
being
2",
where n is the number of different
prime
factors of N. The number of different
pairs
of factors
prime
to each other is therefore 2""
1
,
in which result N
and 1 are considered as one
pair.
383. To
find
the number
ofpositive integers
which are
less than a
given
number and
prime
to it.
Let the
given
number be N= a*6V.
..,
where
a, b, c,...
are the different
prime
factors of N.
The terms of the series
1, 2, 3,.
.
.,
Nwhich are divisible
N N
by
a are
a, 2a, 3a,..., a;
and therefore there are
a a
N
numbers which are divisible
by
a. So also there are
-r-
numbers divisible
by
b,
r-
divisible
by
be,
-j-
divisible
by
abc,
and so on.
We will now shew that
every integer
which is less
than N and not
prime
to Nis counted once and once
only
in the series
-N
^
N _,
N
^
N , .
2 Z
-T + 2
-y--
2-r-
-j
+
(a).
a ab abc abed
Suppose
an
integer
is divisible
by only
one
prime
factor
of
N,
a
suppose;
then that
integer
is counted once in
jy
(a), namely
as one of the numbers which are divisible
CL
by
a.
Next
suppose
an
integer
is divisible
by
r of the
prime
factors
a, b, c,... ,
then that
integer
will be counted r
488 THEORY OF NUMBERS.
N r
(r 1)
N
times in
S
,
it will be counted -~=
s-
times in
S r
a 1.2 ab
it will be counted
- -
1^=
--
times in 2
r
and
1.2.3 abc
,
ab
so
on. Hence the whole number of times an
integer
divisible
by
r of the
prime
factors is
counted,
is
r(r-l) r(r-l)(r-2)
(
,r(r- !)...!
1.2 1.2.3
~]r~~
=
1-(1-1)'=1.
Thus
every
number not
prime
to N is counted once in
(a);
and therefore the number of
positive integers
less
than N and not
prime
to N is
given by (a)
;
provided
however that
unity
is considered to be
prime
to N.
Hence the number of
positive integers
less than Nand
prime
to N is
N N N
N-^- + ^-
h -^^-
+ ...
a ab abc
\
a ab abc
")
[Art
.
260]
.
Ex. 1. Find the number of
integers
less than 100 and
prime
to it.
Since 100=2* . 5
2
,
the number
required
is
Ex. 2. Find the number of
integers
less than 1575 and
prime
to it.
Ant. 719.
Ex. 3. Shew that the number of
integers, including unity,
which
are less than
N[N>2]
and
prime
to N is
even,
and that half
these numbers are less than
-^
.
|For
if a be
prime
to N so also is N
-
a
;
and if a >
,
then
A
T
-a<
-1
2
J
THEORY OF NUMBERS.
489
384 Forms of
square
numbers. Some of the
different
possible
and
impossible
forms of
square
numbers
will be seen from the
following examples
:
Ex. 1. Shew that
every square
is of the form 3m or 3m +
1.
For
every
number is of the form 3m or 3m 1. Hence
every
square
is of the form 9m or 8m
+
1.
Ex. 2. Shew that
every square
is of the form 5m or 5m
=t
l.
For
every
number is of the form
5m,
5m 1 or 5m
=t
2
;
and there-
fore
every square
is of the form
5m,
5m+
1 or 5m
+
4.
Ex. 3. Shew
that,
if a
2
+ &
8
=c
2
,
where
a, 6,
c are
integers;
then will abc
be a
multiple
of 60.
First,
every square
is of the form 3m or 3m +
1;
and therefore
the sum of two
squares
neither of which is a
multiple
of 3 is of the
form 3m +
2 which cannot be a
square.
Hence either a or b must be
a
multiple of
3.
Again, every square
is of the form 5m or 5m
=t
1. The sum of two
squares
neither of which is a
multiple
of 5 is therefore of one of the
forms
5m.,
or 5m =fc 2. Now no
square
can be of the form 5m 2
;
and
if a
square
be of the form
5m,
its root must be a
multiple
of 5.
Hence,
if ab is not a
multiple
of
5,
c will be a
multiple
of 5.
Thus,
in
any case,
abc is a
multiple of
5.
Lastly,
since
every
number is of the form
4m,
4m
+1,
4m
+2
or
4m
+
3, every square
is of the form
16m,
8m
+1,
16m
+
4. Now a
and b cannot both be
odd,
for the sum of their
squares
would then be
of the form 8m +
2 which cannot be a
square. Also,
if one is even
and the other
odd,
the even number must be divisible
by 4,
for the
sum of two
squares
of the forms 8m
+
1 and 16m
+
4
respectively
is of
the form 8m
+
5 which cannot be a
square.
It therefore follows that
ab must be a
multiple of
4.
Thus abc is divisible
by
3,
by
5 and
by 4; hence,
as
3,
4 and
5 are
prime
to one
another,
abc=M
(60).
Ex. 4. Shew that
every
cube is of the form 7m or
1m 1. Shew also
that
every
cube is of the form 9m or 9m
=t
1.
Ex. 5. Shew that
every
fourth
power
is of the form 5m or 5m
+ 1.
Ex. 6. Shew that no
square
number ends with
2, 3,
7 or 8.
Ex. 7. Shew
that,
if a
square
terminate with an odd
digit,
the last
figure
but one will be even.
Ex. 8. Shew that the last
digit
of
any
number is the same as the last
digit
of its
(4n
+
l)th power.
Ex. 9. Shew that the
product
of four consecutive numbers cannot be
a
square.
490
THEORY OF NUMBERS.
EXAMPLES XXXVIII.
1. Shew that the difference of the
squares
of
any
two
prime
numbers
greater
than 3 is divisible
by
24
2. Shew
that,
if n be a
prime greater
than
3,
n
('
-
l)(n
f
-
4)(w'
-
9)
=
M(T
. 3' . 5
.
7).
3. Shew
that,
if n be
any
odd
number,
(n
+
2m)" -(n
+
2m)
=
Jtf(24).
4. Shew that a
4
"*'
-
a
4
"*'
=
M(30).
5. Shew
that,
if N~a*
=
x and
(a+1)
1
N=y,
where x
and
y
are
positive;
then N
xy
is a
square.
6. How
many
numbers are there less than 1000 which are
not divisible
by 2,
3 or 51
7.
P, Q, R,
p, q,
r are
integers,
and
p, q,
r are
prime
P O ft
to one
another;
prove
that,
if + + be an
integer,
then
,
and will all be
integers.
P 1
r
8. Shew that 284 and 220 are two 'amicable'
numbers,
that is two numbers such that each is
equal
to the sum of the
divisors of the other.
9. Shew
that,
if 2"-l be a
prime
number,
then
2"'(2"-l)
will be a
'perfect' number,
that is a number which is
equal
to
the sum of its divisors.
10. Find all the
integral
values of x less than 20 which
make x
1 '
1 divisible
by
680.
11. Shew that no number the sum of whose
digits
is 15 can
be either a
perfect square
or a
perfect
cube.
12. Shew that
every square
can be
expressed
as the differ-
ence between two
squares.
THEORY OF NUMBERS.
491
13. Find a
general
formula for all the numbers which when
divided
by 7, 8,
9 will leave remainders
1, 2,
3
respectively;
and shew that 498 is the least of them.
14.
If n be a
prime
number,
and N
prime
to
n,
shew that
jyr*-n_i
_.
M
(n'),
and that tf*-*"
1
-
1
=
Jf
(').
15. Shew
that,
if n be a
prime
number and N be
prime
to
7i,
then will ^V
rl++-+<"-
1
'
=*=
1
=M(n*).
16. Shew
that,
if
p
be a
prime number,
and
(1 +xy~*=
1 +
ajc
+
ajc*
+ a
a
x* +
...;
then a
t
+
2,
a,-
3,
a
8
+
4,
<fec. will be
multiples
of
p.
17. Shew that if three
prime
numbers be in A. p. their
common difference will be a
multiple
of
6,
unless 3 be one of
the
primes.
|2a
126
18. Shew that .
L
:,
-= is an
integer.
a
jo
la + 6
|2n
19. Shew that -.
,,
is an
integer.
\n+
1
|_ro
\nr
20. Shew that -. i-r-
^
is an
integer.
21. Each of two numbers is the sum of n
squares
;
shew
that the
product
of the two numbers can be
expressed
as the
sum of
^n (n 1)
+ 1
squares.
22. Shew that a* + b* cannot be divisible
by
3,
unless both
a and 6 are divisible
by 3;
shew also that the same result holds
good
for the numbers 7 and 11.
23. Shew
that,
if a' + b*
=
c*,
then
ab(a*-b')
will be a
multiple
of 84.
24. Shew that no rational values of
a, b, c,
d can be found
which will
satisfy
either of the relations a* + b*
=
3
(c
1
+
cP),
25. Shew
that,
if a* + c
2
=
26
8
,
then a
2
-
b
s
=
M
(24).
492 THEORY OF NUMBERS.
CONGRUENCES.
385. Definition. If two numbers a and b leave the
same remainder when divided
by
a third number
c,
they
are said to be
congruent
with
respect
to the modulus
c;
and
this is
expressed by
the notation a
=
6
(mod. c),
which
is called a
congruence.
For
example,
21
=
1
(mod. 10),
and
(a
+
l)
2
= 1
(mod. a).
The
congruence
a
=
6
(mod. c)
shews that a b is a
multiple
of
c,
which can be
expressed by
a b
=
(mod. c).
386. Theorem.
Ifa1
=
b
t
(mod. x),
and a
g
=
b
2 (mod. x)
;
then will a
l
+ a
s
= b
1
+
b
3 (mod. x),
and
a^
=
bfa
(mod. x).
For let
a,
=
m^x
+
r
lt
and a
s
=
m^x
+
r
t
; then,
by sup-
position, 6j
=
n
v
x
+
r
x
and b
t
=
n
a
x
+
r
t
.
Hence a
l
+
a
s (6X
+
6J
=
(Wj
+
m
a n,
n
s
)
a;
;
.-.
(at
+
a
s)
-
(6X
+
6
S
)
=
(mod. a;),
or a
l
+
a,
=
6
t
+
6
2 (mod. a;).
Again,
it is
easily
seen that a
t
a
t bp^
=
a
multiple
of
x,
and therefore a
t
a
g
=
6j6s
(mod. x).
The
proposition
will
clearly
hold
good
for
any
number
of
congruences
to the same modulus.
387.
Congruences
have
many properties analogous
to
equations.
For
example,
if the
congruence
Ax*
+
Bx
+
C
=
(mod. p\
wherein
A, B,
C have constant
integral
values,
be satisfied
by
the three values
a, b,
c of
x,
which are such that a b
is
unity
or
prime
to
p,
and so for
every
other
pair,
then
the
congruence
will hold
good
for all
integral
values of
x,
and
A, B,
C will all be
multiples
of
p.
For we have
Aa*
+ Ba+G=0
(mod. p),
and Ab
2
THEORY OF NUMBERS.
493
.-.
by
subtraction
(a
-
b) {A (a
+
6)
+
B]
=
(mod. p).
Hence,
as a b is
unity
or
prime
to
p,
we have
A
(a
+
6)
+ B
=
(mod. p).
Similarly,
A
(a
+
c)
+
B
=
(mod. p).
Hence,
by
subtraction,
A
(b c)
=
(mod. p).
Therefore A
=
(mod. p),
since b c is
unity
or
prime
to
p.
Then,
since A =
(mod. p),
it follows that B = Q
(mod.
p),
and then that (7=0
(mod. p).
Then,
since
A, B,
C are all
multiples
of
p.
it follows
that Ax*
+
Ex
+
C is also a
multiple
of
p
for all
integral
values of x.
We can
prove
in a similar manner the
general
theorem,
namely
:
If
a
congruence of
the nth
degree
in x be
satisfied
by
more than n values
of
x,
which are such that the
difference
between
any
two is
unity
or is
prime
to the
modulus,
then
the
congruence
will be
satisfied for
all
integral
values
of
x,
and the
coefficients of
all the
different powers of
x will be
multiples of
the modulus.
388. Theorem.
If
a and b are
prime
to one
another,
the numbers
a, 2a, 3a,...,
(6 l)a
will all leave
different
remainders when divided
by
b.
For
suppose
that ra and sa leave the same remainder
when divided
by
6.
Then ra sa
=
M(b);
but if 6 divide
(r s)
a,
and be
prime
to
a,
it must divide r
s,
which is
impossible
if r
and s are both less than b.
Hence the remainders obtained
by dividing
a, 2a,
...
,
(6 1)
a
by
b are all
different;
and since there are 61 of
these
remainders,
they
must be the numbers
1, 2, ,
(b 1)
in some order or other.
494
THEORY OF NUMBERS.
'
If a be not
prime
to 6 the remainders obtained
by dividing a, 2a,
3a, ...,
(b
-
1)
a
by
b will not be all different. For let k be a common
factor of a and
b,
and let a
=
ka and b
=
kfi.
Then it is
easily
seen
that
(r
+
fi)a
and ra will leave the same remainder when divided
by
b,
and
(r
+
/3)
a and ra are both included in the series
a,
2a ('/
-
1)
a
provided
r
+0
< 6
-
J,
COR. If a be
prime
to
b,
and n be
any integer
what-
ever,
the remainders obtained
by dividing
n,
n
+
a,
n
+
2a,
. . .
,
n
+
(b 1)
a
by
6 will all be
different,
and will
therefore be the numbers
0, 1, 2, ...,
(6 1).
389. Fermat's Theorem. From the result of the
preceding
article,
Fermat's theorem can be
easily
deduced.
For,
if a and 6 are
prime
to each
other,
the numbers
a,
2a, ...,
(6 1)
a will
leave,
in some order or
other,
the re-
mainders
1, 2,
. .
.,
(6 1),
when divided
by
b. Hence we have
a. 2a . 3a...
(b^l
.
a)
=
1 . 2 . 3.
..(6
-
1) (mod. b).
that is
|6-1
(a*-
1
-
1)
=
(mod. b).
Now,
if b be a
prime
number,
|
b 1 will be
prime
to
6;
and we have a
6
"
1
1
=
(mod. 6),
which is Fermat's
theorem.
390. Wilson's Theorem.
If
n be a
prime
number,
1
+ 1
n 1 will be divisible
by
n.
If a be
any
number less than the
prime
number
n,
a will be
prime
to
n,
and
hence,
from Art.
388,
the remainders obtained
by dividing
a, 2a, ...,
(n 1)
a
by
n will be the numbers
1, 2, ...,
(n 1);
hence one and
only
one of the remainders will be
unity.
Let then ab be
the
multiple
of a which
gives
rise to the remainder
1; then,
if b were
equal
to
a,
we should have a*
=
1
+
M(n),
or
(a
+
1) (a 1)
=
M
(n),
and this can
only
be the
case,
since
n is a
prime,
if a
=
1 or a
=
n 1. Hence the numbers
2, 3,
...
(n
-
3), (n 2)
can be taken in
pairs
in such a
way
that the
product
of each
pair,
and therefore the
product
of
all the
pairs,
is of the form
M(ri)
+
1.
Thus 2 .
3
.
4...(n
-
2)
=
M(n)
4
1
;
/. In
-
1
=
M(n)
x
(n
-
1)
+
n
-
1.
THEORY OF NUMBERS.
495
Hence
|w-l
+
1
=
M
(n).
Wilson's theorem
may
also be
proved
as follows :
From Art.
305,
we have
Now
by
Fermat's theorem
(n I)""
1
=
1
4-
M
(n),
2y-
l
=
I+M(n),
&c.
Hence we have
that is
(1
-
I)"'
1
-
(-
I)*-
l
+M(n)
=
|ra-l;
hence,
as n-l
is
even,
I n 1 +
1
=
M
(n),
Wilson's theorem is
important
on account of its
express-
ing
a distinctive
property
of
prime numbers;
for 1
+
In 1
is not divisible
by
n unless n is a
prime.
For if
any
number less than n divide n it will divide In 1 and
therefore cannot divide In 1
+
1.
391. Theorem.
If
the number
of integers
less than
any
number n and
prime
to n be denoted
by <f> (n)
; then,
if
a, b, c,...
are
prime
to each
other,
<f> (abc.
.
.)
=
<f> (a)
x
<(&)
x
<(c)
,
provided
that
unity
is considered to be
prime
to
any greater
number.
First take the case of two numbers
a,
b and their
product
ab.
Arrange
the ab numbers as under :
1
,
2
,
3
, ,
& a
a + I
,
a
+
2,
a
+
3, ,
a
+
k,
2a
2a+l,
2a
+
2, 2a+3, ,
2a
+
&,
3a
496
THEORY OF NUMBERS.
Then it is clear that all the
integers
in the kih vertical
column will or will not be
prime
to a
according
as k is or
is not
prime
to a. Hence there are
<f> (a)
columns of
integers, including
the
first,
all of which are
prime
to a.
Then
again,
we know from Art. 388 that since a is
prime
to
b,
the remainders obtained
by dividing
the numbers
k,
a+
k,...,
(b I)a
+
k
by
6 are the numbers
0, 1, 2,...
(6 1);
and it is clear that a number is or is not
prime
to
b
according
as the remainder obtained
by dividing
the
number
by
b is or is not
prime
to b. Hence there are
as
many integers prime
to b in
any
one column as there
are in the series
0, 1, 2,...
(6 1),
that is to
say,
there are
in each column
(f>(b) integers prime
to b. Thus there are
(f> (a)
columns of
integers prime
to a and each column
contains
<f> (b) integers
which are also
prime
to b. But all
integers
which are
prime
to a and also to b are
prime
to
a x b. Hence the number of
integers
less than ab and
prime
to ab is
(a)
x
<f> (b),
so that
$ (ab)
=
(f> (a)
x
<f> (b).
The
proposition
can at once be
extended,
for we have
)
=
<f>(axbc...)=<f>
(a)
x
392. The number of
integers
less than a
given
number
and
prime
to it can be found
by
means of the theorem in
the
preceding
article.
For let the number be N
=
a
a
b^c^...,
where
a, b, c,...
are the different
prime
factors of N.
To find the number of
integers
less than a
a
and
prime
to
it,
(unity being
considered as one of these
numbers)
we
must subtract a
a
~
l
from
a;
for tlie numbers
a, 2a, 3a,...,
a""
1
. a are not
prime
to
a,
and these are the
only
numbers
which are not
prime
to a
;
thus
THEORY OF NUMBERS. 497
Similarly
<
(&*)
=
6*
(l
-
1}
,
<j> (c?)
=
c?
(
I
- -
},
&c.
\ O/ \ GJ
But, by
the
preceding
article,
Hence
f W
=
Jf(l
-
$ (l
-
i) (l
- -
1
)...,
where
a, b, c,...
are the different
prime
factors of
N,
and
unity
is considered to be
prime
to
a, b, c,
&c.
393. The
following
is an extension of Fermat's
Theorem :
If
a and m are two numbers
prime
to one
another,
and
<j>
(m)
the number
of integers, including unity,
which
are less than m and
prime
to
m;
then a*
<
m
>
1
=
(mod. m).
Let the
<f> (m) integers
less than m and
prime
to m be
denoted
by
1,
a, /3,
7,.
. .
,
(m 1).
Then the
products
a .
1,
aa, a/3,
ay,...,
a
(m 1)
must all leave different remainders
when divided
by
m,
for if
any
two,
ra and sa
suppose,
left
the same
remainder,
(r s)
a would be a
multiple
of
m,
which is
impossible
since a is
prime
to m and r s is less
than m. Moreover the remainders must all be
prime
to
m,
since the two factors of
any
one of the
products
are both
prime
to m
;
and therefore as the
<J> (m)
remainders are
all
different,
and are all
prime
to
m,
they
must
be,
in some
order or
other,
the
<j>(m)
numbers
1, a, /3, 7...
Hence
a. aa.
a/3
......
a(m 1)
=
1 . a .
/3
.7... (m l)(mod. m);
.-.
(
a
* () _
1}
l . a
. . .
.(
m
_
i)
=
o
(mod. m).
Hence as 1.
oc.j8...(ra 1)
is
prime
to
m,
we have
a
b(m)
_i
=
o
(mod. m).
If m be a
prime
number,
<
(m)
=
m
1,
and we have
Fermat's Theorem.
s.A.
32
498 THEORY OF NUMBERS.
394.
Lagrange's
Theorem.
Ifp
be a
prime number,
the sum
of
all the
products
r
together of
the numbers
1, 2,
3
...,
p
1,
is divisible
by p,
r
being any integer
not
greater
than
p
2.
Consider the
identity
(x
-
1) (x
-
2).
.
.0
-
p
+
1)
=
x*~
l
-
8
l
a**
+
S
2
+...+(
Change
x into x 1
;
theii
(a; -2)(#- 3).. .(*-/>)
=
(a-l)
p
-'
-,(*- i)'"
+(
Hence
1
-
s
l
x>~*
+
s
t
or
3
+
...+(-
i
I-
1
-
S *
-
1"-
+
. . .
-f
-
I)
Equate
the coefficients of the different
powers
of a: in
the above
identity
;
and we have
7
1.2
p(j)-
2
~
1.2.3
x>
1.2
f(p-l)(p-2)(p-3).
?
1.2.3.4 1.2.3
(p-2)(p-3)
1.2
,-...
s
'1.2...-
Since
p
is a
prime
the first term in each
right-hand
member is divisible
by p
;
whence it follows from the first
equation
that S
l
is a
multiple
of
p,
and then that S
9
is a
multiple
of
p,
and so on to
iS
M
.
THEORY OF NUMBERS. 499
Lagrange's
Theorem
may
also be deduced from the
Theorem of Art.
387,
assuming
that Fermat's Theorem is
known.
For the
congruence
(a?- 1) (x
-
Z)...(x-p
+
1)
-x^
1
+
1
=
(mod. p\
is of the
(p 2)th degree
in
x,
and
by
Fermat's Theorem
it is satisfied
by
the
p
I values
1, 2, ...,
p
1,
which are
such that the difference between
any pair
is either
unity
or is
prime
to
p.
Hence,
by
Art 387 it is true for all
integral
values of
x,
and the coefficients of all the different
powers
of x are
multiples
of
p.
It should be remarked that Wilson's Theorem follows
at once
by putting
x 0.
395. Reduction of fractions to
circulating
decimals.
It is obvious that a fraction whose denominator con-
tains
only
the factors 2 and 5 can be reduced to a ter-
minating
decimal,
for
a o.5
J>
.2
2*5~
If, however,
the denominator contains
any
factor
which
is
prime
to
10,
the fraction can
only
be reduced to a cir-
culating
decimal.
Let the fraction in its lowest terms be =
r ,
where
2
y
o* .
6 is
prime
to 10. Let this fraction be
equivalent
to a cir-
culating
decimal with a
recurring
and
/3
non-recurring
figures.
Then
a g . 5" . 2' N
2
P
. 5* . b
~
10
P+
". 6 T 10*
(10 -1)
'
Hence,
as b is
prime
to a and to
10,
10
a
1
=
M
(b),
and
a
is the lowest
power of
10
for
which this is
true,
for
500 THEORY OF NUMBERS.
otherwise the fraction could be
expressed
as a
circulating
decimal with fewer than a
recurring figures.
It should be noticed that the number of
recurring
figures
in the
circulating
decimal
depends only
on b and is
not affected
by
the
presence
of 2
P
5
?
in the
denominator,
for the number is
a,
where a is the lowest
power
of 10
which is
equal
to M
(b)
+
1.
We will now
prove
that a is either
equal
to
<f> (b)
or to
one of its
sub-multiples.
By
the extension of Format's Theorem
[Art. 393]
we
have
We have also 10"
-
1
=
M
(b).
Hence,
if a be not
<f> (b)
or one of its
sub-multiples,
let
(f> (b)
=-k'Ji
+
r,
where r
< a.
Then 10*<
6
> -
1
=
10* . 10
r
-
1
=
{
M
(b)
+
1
}MO
r
-
1
=
M
(b)
+
1 O
r
-
1
;
lQ
r
-I=M(b),
which is
impossible
since r
<
a,
and a is the lowest
power
of 10 which is
equal
to M
(b)
+
1.
Hence,
if
b be the
factor of
the denominator
of
a
fraction
which is
prime
to
10,
the number
of recurring figures
in the
equivalent
decimal is either
<f> (b)
or one
of
its
sub-multiples.
396. We shall conclude this
chapter by considering
the
following examples
:
Ex. 1. Shew that S
2**3
-
8n
-
9 is a
multiple
of 64.
We have
Ex. 2. Shew that S
2"
-
32?i*
+
24n
-
1
=
(mod. 512).
Let u
ll
=3
2n
-32n
s
+
24n-l;
then
u^+i
=
3
2 -w
-
32
(n
+
l)
a
+
24
(n
+
1)
-
1.
Hence u
n+l
-
9u
n
=
256n
s
-
256n
=
256n
(n
-
1)
=
M
(512),
since n
(n
-
1)
is divisible
by
2.
EXAMPLES. 501
And since
n
_
H
-&u
n
=
(mod. 612),
it follows that
+!
^
(mod. 512) provided
w
n
=
(mod. 612).
The theorem is therefore
true for all values of n
provided
it is true for n
=
l,
which is the case
since
1^
=
0.
Ex. 3. Shew that no
prime
factor of
n*+l
can be of the form 4m- 1.
Every prime
number,
except 2,
is of the form 2&+1. Let then
2fc
+
l be a
prime
factor of n
a
+
l. Then n is
prune
to 2fc
+
l,
and
therefore
by
Fermat's theorem n
2*
=
M
(2fc
+
1)
+
1.
But,
by supposition,
n
2
+
1
=
M
(2k
+
1) ;
Since n*
=
M
(2k
+
1)
+
1 and n
n
=M(2k
+
!)
+
(-!)*
it follows
that k must be
even,
and therefore
every prime
factor of n
2
+
1 is of
the form 4m
+ 1,
and therefore no
prime
factor can be of the form
4m- 1.
Since the
product
of
any
number of factors of the form 4m
+
1 is
of the same
form,
it follows that
every
odd divisor of n*
+
1 is of the
form 47/1+1.
Ex. 4. Shew that
every
whole number is a divisor of a series of nines
followed
by
zeros.
Divide the successive
powers
of 10
by
the
number,
n
suppose,
then
there can
only
be n different remainders
including
zero,
and hence
any particular
remainder must recur. Let then 10* and 10" leave
the same remainder when divided
by
n : then 10
Z
-
10" is divisible
by
n
and is of the
required
form.
EXAMPLES XXXIX.
1.
Prove the
following
:
(i)
2""
M
-9tt
2
+ 37i-2
=
Jf(54).
(ii)
5
2" +1
+ n"
-
5n
3
+ 4n
-
5
=
M
(1 20).
(iii)
4
2""
+ 3
n
^
=
(mod. 13).
(iv)
3
4" +s
+ 2 . 4
3- +1
=
(mod. 17).
2. Shew
that,
if a be a
prime
number,
and b be
prime
to
a;
then 1
2
6
2
, 2*6', .......
(^s^Y
& will
give
different re-
mainders when divided
by
a.
3. Shew
that,
if 4i
+ 1 be a
prime
number,
it will be a
factor of
{
1
2n\
a
+ 1
;
and
that,
if 4n 1 be a
prime,
it will be a
factor of
{ 1
2n
-!}'-!.
4. Shew
that,
if n be a
prime
number,
and r be less than
n;
then will lr-1
\n-r
+
(- I)
1
"
1
=
M(ri).
502 EXAMPLES.
5. Shew
that,
if m and n are
prime
to one
another,
every
odd divisor of m* + n* is of the form 4& + 1.
6. Shew that
y5+H5
+
5=
+
j;
+ .-. * i
where
2, 3, 5,...
are the
prime
numbers in order.
7. Shew that the arithmetic mean of all numbers less than
n and
prime
to it
(including
unity)
is
\n,
8. Shew
that,
if N be
any number,
and
a, b,
c,
... be its
different
prime
factors
;
then the sum of all the numbers less
than Jff and
prime
to N is
-^-
(1
)(^~r) (^ )
an^
the sum of the
squares
of all such
numbers is
N*
9. If <
(TO)
denote the number of
integers
less than m
and
prime
to
it;
and if d
lt
d
t
,
d
a
,...
be the different divisors
of
n;
then will 2
<f>
(d)
=
n.
10. Shew
that,
if a fraction
-j-
,
where 6 is
prime
and
prime
to
10,
be reduced to a
decimal,
and if the number of
figures
in the
recurring period
is even
;
then the sum of the first half of
the
figures
added to the last half will consist
wholly
of nines.
11. If be converted to a
circulating
decimal with
p
1
P
figures
in its
recurring period,
shew that
p
must be
prime
and
that the
recurring period being multiplied by 2, 3,
......
(p 1)
will
reproduce
its own
digits
in the same order.
12. Shew
that,
if =
has a
circulating period
of
p figures,
-*
of
q figures,
and
5
of r
figures,...,
and if
P, Q, R,...
are
prime,
then _ will have a
circulating period
of n
figures,
where
n is the L.C. M. of
p, q,
r....
CHAPTER XXIX.
INDETERMINATE
EQUATIONS.
397. WE have
already
seen that a
single equation
with more than one unknown
quantity,
or n
equations
with more than n unknown
quantities,
can be satisfied in
an indefinite number of
ways, provided
there is no restric-
tion on the values which the unknown
quantities may
have.
If, however,
the values of the unknown
quantities
are
subject
to
any
restriction,
n
equations may
suffice to
determine the values of more than n unknown
quantities.
We shall in the
present chapter
consider some cases of
equations
in which the unknown
quantities
are restricted
to
integral
values.
398. It is clear that
every equation
of the first
degree
with two unknown
quantities
x and
y
can be reduced
to one or other of the forms ax
+
by=c,
cuv
by=c,
where
a, 6,
c are
positive integers.
By changing
x into x and
y
into
y,
ax
+
by
=
c
will become ax +
by
=
c,
and ax
by
=
o will become
ax
+
by
=
c\
hence in order to shew how to find
integral
solutions of
any equation
of the first
degree
in x
and
y,
it is
only necessary
to consider the two
types
ax
+
by
=
o and
axby=c.
Now,
it is evident that the
equation
ax
by
=
c cannot
be satisfied
by integral
values of x and
y,
if a and b have
any
common factor which is not also a factor of c
; and,
if
a,
b and c have
any
common
factor,
the
equation
can be
504
INDETERMINATE
EQUATIONS.
divided
throughout by
that factor. In what follows it
will therefore be
supposed
that a and b are
prime
to one
another.
399. To shew that
integral
values can
always
be
found
which will
satisfy
the
equation axby
=
c,
provided
a and
b are
prime
to one another.
Let
T
be reduced to a continued
fraction,
and let
o
79 CL
-
be the
convergent immediately preceding
-=-.
Then,
from Art.
357,
aqpb
1;
/.
a(cq)-b(cp)
=
c
(i),
and a
( cq)
+
b(+ cp)
=
c
(ii).
Hence it follows from
(i)
that either x
=
cq, y cp
or
x
=
cq, y
=
cp
is a solution of the
equation
ax
by
=
c;
and from
(ii)
that either x
=
cq, y
=
cp
or x
=
cq,
y
=
cp
is
a solution of the
equation
ax
+
by
=
c.
Hence at least one set of
integral
values of x and
y
can
always
be found which will
satisfy
the
equation
ax
by=c.
The above
investigation
fails when a or b is
unity.
But the
equation
ax
y
=
c is
obviously
satisfied
by
the
values x
=
a,
+
y
=
c
aa,
where a is
any integer.
So also
x
by
=
c is satisfied
by
the values x
=
c
+
6/3,
y
=
ftt
where
(3
is
any integer.
Henoe the
equation
ax
by
=
c
always
admits of at
least one set of
integral
values.
400.
Having given
one set
of integral
values which
satisfy
the
equation
ax
by
=
c,
to
find
all other
possible
integral
solutions.
Let x
=
a,
y
=
/9
be one solution of the
equation
ax
by=c;
then aa
bft
=
c.
Hence,
by
subtraction,
a
(x
-
a)
-
b
(y
-
ft)
=
0.
INDETERMINATE
EQUATIONS.
505
Now since a divides
a(x a),
it must also divide
b
(y /3) ;
a must therefore be a factor of
y
ft,
since it is
prime
to b.
Let then
y
ft
=
ma,
where ra is
any integer
;
then
a
(as a)
=
mba,
and therefore as
=
a
+
mb.
Hence,
if x
a,
y
=
/8
be one solution in
integers
of
the
equation
ax
by
=
c,
all other solutions are
given by
x
=
a.
+
mb,
y
=
ft
+
ma,
where m is
any integer.
It is clear from the above that there are an indefinite
number of sets of
integral
values which
satisfy
the
equation
axby
=
c,
provided
there is one such set
;
and,
from the
preceding
article,
we know that there is one set of
integral
values.
It is also clear
that,
whether a and
ft
are
positive
or
not,
an indefinite number of values can be
given
to m
which will make a
+
mb and
ft
+
ma both
positive.
Hence there are an infinite number of
positive integral
solutions of the
equation
ax
by
=
c.
401.
Having given
one set
of integral
values which
satisfy
the
equation
aae
+
by
=
c,
to
find
all other
possible
integral
solutions.
Let
x
=
a,
y
=
ft
be one
integral
solution of the
equation
ax
+
by
=
c
;
then act
+
bft
=
c.
Hence,
by
sub-
traction,
a
(x a)
-f
b
(y ft)
=
0.
Now,
since a divides a
(x
a),
it must also divide
b
(y ft}
;
a must therefore be a factor of
y
ft,
since it is
prime
to b.
Let then
y
ft
=
ma,
where m is
any integer
;
then
a(x a)= b(y ft)
=
mob
;
and therefore x
=
a.
mb.
Hence,
if a;
=
a,
y
=
(31oe
one solution in
integers
of the
equation
ax
by
=
c,
all other
integral
solutions are
given
by
x
=
ct
mb,
y
=
ft
+
ma,
where m is
any integer.
506 INDETERMINATE
EQUATIONS.
From the
above,
together
with Art.
399,
it follows that
there are an indefinite number of sets of
integral
values
which
satisfy
the
equation
ax +
by
=
c. The number of
positive integral
solutions of the
equation
is, however,
limited in number.
402. To
find
the number
of positive integral
solutions
of
the
equation
ax
+
by
=
c.
We have
proved
in Art.
399,
that the
equation
ax +
by
=
c is satisfied
by
the values x
=
cq, y= cp,
or
by
the values x
=
cq,y
=
cp,
where
p/q
is the
penultimate
convergent
to
a/b.
First
suppose
that x
=
cq, y
=
cp satisfy
the
equa-
tion;
then all other
integral
values which
satisfy
the
equation
are
given by
x=cqmb, y
=
cp
+ ma
(i),
where m is
any integer.
From
(i)
it is clear that in order that x and
y may
both
be
positive,
and not
zero,
m must be a
positive integer,
and that the
greatest permissible
value of m is
/(
~
j
and its least value
/()
+
!,
so that the number of
\aj
different values of m is
/(-r
1
/(). Hence,
as one
\b/
\aj
set of values of x and
y corresponds
to each value of
m,
the number of solutions
is/f-^J
I
(
).
T
.
cq T
, ,
cp 7 ,
/ ,,
c
a(cq)-b(cp)
Let
-
=
L
+f.
and
=
Z. -f
/;
then
-?
=
v * x
,
v ^'
b a ab ab
Cq Cp T T * f rr T (
c
\ r r
m
-f
c
=
1
,
^
+-A /
-Hence 1
[
r is
/,
/ or
b a
\abj
7",
7
S
1
according
as
ft
is not or is less than
ft
.
Thus the number of solutions is
/(-r
INDETERMINATE
EQUATIONS.
507
according
as the fractional
part
of
-^
is or is not less than
C"D
the fractional
part
of .
a
It can be shewn in a similar manner that if
x
=
cq, y
=
cp satisfy
the
equation,
the number of solu-
/ c \ ( c \
tions in
positive integers
is / I
r]
+
1 or /
{
-^
1
according
as the fractional
part
of is or is not less than the frac-
a
tional
part
of
-~
.
Ex. 1. Find the
positive integral
values of x and
y
which
satisfy
the
equation
1x
-
ISy
=
26.
7111
We have
=
T
=-
5
,
the
penultimate convergent
is therefore
Id 1
~h
1
+
O
i. Then 7. 2- 13.
1=1;
/. 7
(2
x
26)
-
13
(26)
=
26.
2
Hence one solution is
z=52, y=26;
the
general
solution is there-
fore x=52
+
13m, t/
=
26 + 7m.
[In
this case the solution
x=Q,
y=
-
2 can be seen
by inspection;
and hence the
general
solution is x
=
I3m,
y=-2
+ 7m,
which is
easily
seen to
agree
with the
previous result.]
Ex. 2. Find the
positive integral
values of x and
y
which
satisfy
the
equation
7x +
ICty
=
280.
7111 2
Here
TTI
=
T
K ~5
*^e
penultimate convergent being
-
. Then
7.3-10.2
=
1;
.-.
7(3.280)
+
10(-2.280)
=
280.
Hence x
=
840,
y
=
-
560 is one solution in
integers.
The
general
solution in
integers
is therefore x=840
-10m, y=
-560
+
7m;
and,
in order that x and
y may
be
positive
m
:f
84 and m
4
80. Thus the
only
values are
a;=40, y
=
0; x=SO,
y
=
7;
a;
=
20,
j/
=
14; a;=10,
y
=
21; a:=0,
y=28.
Ex. 3. Find the number of solutions in
positive integers
of the
equation
3111
Here
^
=
T T
,
whence 3 . 2-5. 1
=
1;
.-. 3 .
(2
x
1306)
+
5
(
-
1306)
=
1306.
Hence the
general
solution is x
=
2612
-
5m,
y
=
3m
-
1306.
For
positive
values of x and
y
we must have m>435 and
Hence the number of solutions is 522
-
435=87.
508 INDETERMINATE
EQUATIONS.
403.
Integral
solutions of the two
equations
ax
+
by
+ cz
=
d,
a'x +
b'y
+
c'z
=
d'
can be obtained as follows.
Eliminate one of the
variables,
z
suppose
;
we then
have the
equation
(ac'
-
a'c)
a;
+
(be'
-
b'c) y
=
dc'
-
d'c..........
(i),
and this
equation
has
integral
solutions
provided
ac a'c
and be b'c are
prime
to one
another,
or will become
prime
to one another after division
by any
common factor
which is also a factor of dc' d'c.
Hence from
(i)
we
obtain,
as in the
preceding
articles,
the
general
solution
x
=
a.
+
(be' b'c)
n,
y
=
$
(ac a'c)
n,
where x
=
a,
y=$
is
any integral
solution,
and n is
any
integer.
Now substitute these values of x and
y
in either of
the
original equations
: we then obtain an
equation
of the
form Az
+
Bn
=
C,
from which we can obtain
integral
solutions of the form z
=
7
+
Bm,
n
=
8
Am,
provided
A
and B are
prime
to one
another,
or will become so after
division
by any
common factor which is also a factor of G.
Ex. Find
integral
solutions of the simultaneous
equations
Eliminating z,
we have 13a;
+
13f/
=
52,
or x
+
y=4.
Whence
x=2
+n,
y
=
2-n. Then 5
(2
+
n)
+
7
(2
-
n)
+
2*
=
24,
that isz-n=0.
Hence the
general
solution is x=2 +
n,
y=2-n,
z=n.
If
x, y
and z are to be
positive,
the
only
solutions are x
=
4, y
=
0,
z=2; x=3,y
=
l,
2
=
1;
and
x=2, j/
=
2, z=0; and,
if zero values are
excluded,
there is
only
one
solution,
namely x=B,
y
=
l,
z1.
404. The
following
are
examples
of some other forms
of indeterminate
equations.
Other cases will be found in
Barlow's
Theory
of Numbers.
Ex. 1. Find the
positive integral
solutions
(excluding
zero
values)
of
the
equation
3z
+
2y
+
8z=40.
It is clear that z cannot be
greater
than
4,
if zero and
negative
values of x and
y
are inadmissible.
INDETERMINATE
EQUATIONS.
509
Hence we have the
following equations
:
z
=
4,
3x +
2y=
8;
And it will be found that all the solutions
required
are
2, 1, 4;
4, 2, 3; 2, 5, 3; 6, 3, 2; 4, 6, 2; 2, 9, 2;
10, 1, 1; 8,
4, 1; 6, 7, 1;
4, 10,
1
;
and
2, 13,
1.
Ex. 2. Find the
positive integral
solutions of the
equation
We have
(3x-2y) (2a;-3y)
=
16; hence,
as x and
y
are
integers,
3x
-
2y
must be an
integer,
and must therefore be a factor of 16.
Thus one or other of the
following
simultaneous
equations
must
hold
good
3ar-2y=16,
510
EXAMPLES.
EXAMPLES XL.
1. Find all the
positive integral
solutions of the
equations
:
(1) 7x+l5y
=
53.
(2)
8a;+13y
=
138.
(3) 7x+9y=100.
(4)
I5x +
7ly
=
10653.
2. Find the number of
positive integral
solutions of
2a; +
3y=133
and of 7a: +
lly
=
2312.
3. Find the
general integral
solutions of the
equations
(1) 7x-l3y=l5.
(2) 9*-lly
=
4.
(3) 119x--105y
=
217.
(4)
49x-69y=
100.
4. Find the
positive integral
solutions
(excluding zero)
of the
equations
(1) 2aj+3y
+ 7=23.
(2)
7x +
4y
+ 18s
=
109.
(3)
5aj +
y
+
7s=39,
(4)
3x +
2y
+ Zz
=
250,
170.
5. Solve in
positive integers (excluding zero)
the
equa-
tions :
(i)
2
(ii)
x*
-
xy
+ 2x
3y
=
1
1.
(iii)
2x'
(iv)
2
6. Shew that
integral
values of
x, y
and z which
satisfy
the
equation
ax +
by
+ cz
=
d
t
form three arithmetical
progressions.
7. Divide 316 into two
parts
so that one
part
may
be
divisible
by
13 and the other
by
11.
8. In how
many ways
can 1. 6*. 6d. be
paid
with
half-crowns and florins 1
9. In how
many ways
can 100 be made
up
of
guineas
and
crowns ?
EXAMPLES. 511
10. In how
many ways
can a man who has
only
8 crown
pieces pay
11
shillings
to another who has
only
florins?
11. Find the
greatest
and least sums of
money
which can
be
paid
in
eight ways
and no more with half-crowns and
florins,
both sorts of coins
being
used.
12. Find all the different sums of
money
which can be
paid
in three
ways
and no more with
four-penny pieces
and three-
penny pieces,
both sorts of coins
being
used.
13. Find all the numbers of two
digits
which are
multiples
of the
product
of their
digits.
14. Two different numbers each of two
digits,
and which
end with the same
digit,
are such that when divided
by
9 the
quotient
of each is the remainder of the other. Find all the
sets of numbers which
satisfy
the conditions.
15. A man's
age
in 1887 was
equal
to the sum of the
digits
in the
year
of his birth
: how old was he ?
16. Shew
that,
if
(1 -x*i)(I
-)...
(1 -x**)
then the number of solutions in
positive integers (including
zero)
of the
equation
a
l
x
l
+ a
f
a;
g
+ ... -f a
u
x
u
=
m,
is A
m
,
a
p
a
g
, ...,
a
n being
all
integers.
The number of solutions of the
equations
x +
2y
=
is
At an entertainment the
prices
of admission were
Is.,
2s. and
5,
and the total
receipts
1000;
shew that there are 1005201
ways
in which the audience
might
have been made
up.
17. The
money paid
for admission to a concert was
300,
the
prices
of admission
being 5s.,
3s. and
Is.;
shew that the
number of
ways
in which the audience
may
have been made
up
is 1201801.
CHAPTER XXX.
PROBABILITY.
405. THE
following
is
generally given
as the defini-
tions of
probability
or chance:
Definition. If an event can
happen
in a
ways
and fail
in b
ways,
and all these
ways
are
equally likely
to
occur,
then the
probability
of its
happening
is r and the
pro-
Ot
-p
bability
of its
failing
is
^
.
To make the above definition
complete
it is
necessary
to
explain
what is meant
by
'
equally likely.'
Events are
said to be
equally likely
when we have no reason to
expect
any
one rather than
any
other. For
example,
if we have
to draw a ball from a
bag
which we know contains
unknown numbers of black and white
balls,
and none of
any
other
colour,
we have
just
as much reason to
expect
a black ball as a white
;
the
drawing
of a black ball and of
a white one are thus
equally likely.
Hence,
as either a
black ball or a white ball must be
chosen,
the
probability
of
drawing
either is
,
for there are two
equally likely
cases,
in one of which the event
happens
and in the other
it fails.
Again,
if we have to draw a ball from a
bag
which we know contains
only
black,
white and red
balls,
but in unknown
proportions,
we have
just
as much reason
to
expect
one colour as to
expect
either of the
others,
so
that the
drawing
of a
black,
of a white and of a red ball
PROBABILITY.
513
are all
equally likely
;
and hence the
probability
of draw-
ing any particular
colour is
,
for there are three
equally
likely
cases,
and
any particular
colour is drawn in one case
and is not drawn in the other two cases.
Another
meaning may
however be
given
to
'
equally
likely
;'
for events
may
be said to be
equally likely
when
they
occur
equally
often,
in the
long
run. For
example,
if
a coin be tossed
up,
we
may
know that in a
very great
number of
trials,
although
the number of
'
heads
'
is
by
no
means
necessarily
the same as the number of
'
tails/
yet
the ratio of these numbers becomes more and more
nearly
equal
to
unity
as the number of trials is
increased,
and that
the ratio of the number of heads to the number of tails will
differ from
unity by
a
very
small fraction when the number
of trials is
very great;
and this is what is meant
by saying
that heads and tails occur
equally
often in the
long
run.
Now,
if each of the a
ways
in which an event can
happen
and each of the b
ways
in which it can fail occur
equally
often,
in the
long
run,
it follows that the event
happens,
in the
long
run,
a times and fails b times out of
every
a
+
b cases. We
may
therefore
say, consistently
with
the former
definition,
that the
probability of
an event is the
ratio
of
the number
of
times in which the event
occurs,
in the
long
run,
to the sum
of
the number
of
times in which events
of
that
description
occur and in which
they fail
to occur.
Thus,
if it be known
that,
in the
long run,
out of
every
41
children
born,
there are 21
boys
and 20
girls,
the
probability
of
any
21
particular
birth
being
that of a
boy
is
-j-
Again,
if one of two
players
at
any game win,
in the
long run,
6
games
out of
every
8,
the
probability
of his
winning any particular
. 5
game
is-.
We
may
remark
that,
in the
great majority
of
cases,
including
all the cases of
practical utility,
such as the data
used
by
Assurance
Companies,
the
only way
in which
pro-
bability
can be estimated is
by
the last
method,
namely, by
finding
the ratio of the actual number of times the event
s- A.
33
514 PROBABILITY.
occurs,
in a
large
number of
cases,
to the whole number
of times in which it occurs and in which it fails.
406. If an event is certain it will occur without fail
in
every
case : its
probability
is therefore
unity.
It follows at once from the definition of
probability
that if
p
be the
probability
that
any
event should
occur,
1
p
will be the
probability
of its
failing
to occur.
When the
probability
of the
happening
of an event is
to the
probability
of its failure as a is to
b,
the odds are
said to be a to 6
for
the
event,
or b to a
against
it,
according
as a is
greater
or less than b.
407. Exclusive events. Events are said to be
mutually
exclusive when the
supposition
that
any
one
takes
place
is
incompatible
with the
supposition
that
any
other takes
place.
When
different
events are
mutually
exclusive the chance
that one or other
of
the
different
events occurs is the sum
of
the chances
of
the
separate
events.
It will be sufficient to consider three events.
Let the
respective probabilities
of the three
events,
expressed
as fractions with the same
denominator,
be
a a ,
a.
-/,
-
and
3.
d d d
Then,
out of d
equally likely ways,
the three events
can
happen
in
a,,
a
2
and
a, ways respectively.
Hence,
as the events never
concur,
one or other of
them will
happen
in
a,
+
a
a
+
a
8
out of d
equally likely
ways.
Hence the
probability
of one or other of the three
events
happening
is
a.
+
a
+
a
~
This
proves
the
proposition
for three
mutually
ex-
clusive events
;
and
any
other case can be
proved
in a
similar manner.
PROBABILITY.
515
Ex. 1.
Find the chance of
throwing
3 with an
ordinary
six-faced
die.
Since
any
one face is as
likely
to be
exposed
as
any
other
face,
there is one favourable and five unfavourable cases which are all
equally likely
;
hence the
required probability
is
^
Ex. 2. Find the chance of
throwing
an odd number with an
ordinary
die.
Ant. -.
Ex. 3. Find the chance of
drawing
a red ball from a
bag
which con-
tains 5 white and 7 red balls.
Here
any
one ball is as
likely
to be drawn as
any
other
;
thus there
are 7 favourable and 5 unfavourable cases which are all
equally
7
likely
;
the
required probability
is therefore
-^
.
Ex. 4. Two balls are to be drawn from a
bag containing
5 red and 7
white balls
;
find the chance that
they
will both be white.
Here
any
one
pair
of balls is as
likely
to be drawn as
any
other
pair.
The total number of
pairs
is
^Cg,
and the number of
pairs
which are both white is
7
(7
2
: the
required
chance is therefore
7.6 / 12 . 11
_
7
1.2
/
1.2
~22'
Ex. 5. Shew that the odds are 7 to 3
against drawing
2 red balls
from a
bag containing
3 red and 2 white balls.
Ex. 6. Three balls are to be drawn from a
bag containing
2
black,
2 white and 2 red balls
;
shew that the odds are 3 to 2
against drawing
a ball of each
colour,
and 4 to 1
against drawing
2 white balls.
Ex. 7. A
party
of n
persons
take their seats at random at a round
table : shew that it is n
-
3 to 2
against
two
specified persons
sitting together.
408.
Independent
Events. The
probability
that two
independent
events should both
happen
is the
product of
the
separate probabilities of
their
happening.
Suppose
that the first event can
happen
in
a,
and fail
in
&! equally likely ways;
and
suppose
that the second
event can
happen
in a
2
and fail in 5
S
equally likely ways.
Then each of the
o\
+
b^
cases
may
be associated with each
of the
Oj+fra
cases to make
(a1
+
b
1)(aa
+
b^ compound
cases which are all
equally likely;
and in
aft^
of these
compound
cases both events
happen.
Hence the
proba-
332
516
PROBABILITY.
bility
that both events
happen
is
;
-~-
=-. .
that
-
>
is
-
-^T-
x
j-
,
which
proves
the
proposition.
Thus the
probability
of the concurrence of two inde-
pendent
events whose
respective probabilities
are
>,
and
pt
Cor. If
P!
and
pt
be the
probabilities
of two inde-
pendent
evente,
the chance that
they
will both fail is
(1 P^)(!L ps),
the chance that the first
happens
and the
second fails is
pl (I p
t),
and the chance that the second
happens
and the first fails is
(1
It can be shewn in a similar manner
that,
if
plt
p
t
,p
t
,...
be the
probabilities
of
any
number of
independent
events,
then the
probability
that
they
all
happen
will be
Pi-Pi-Ps--,
and that
they
all fail
(1 p^O-
->
2)(1 p3)---,
&c.
409.
Dependent
Events. If two events are not
independent,
but the
probability
of the second is different
when the first
happens
from what it is when the first
fails,
the
reasoning
of the
previous
article will still hold
good
provided
that
p^
is the
probability
that the second event
happens
when the first is known to have
happened.
Thus
if
PJ
be the
probability
of
any
event,
and
pt
the
probability
of
any
other event on the
supposition
that the first has
happened
;
then the
probability
that both events will
happen
in the order
specified
will be
p1
'^p
t
^ And
similarly
for
any
number of
dependent
events.
Ex. 1. Find the
probability
of
throwing
two heads with two throws of
a coin.
The
probability
of
throwing
heads is
jj
for each throw
;
hence the
required probability
is,
by
Art.
408,
5
x
5
=
-
.
a & 4
Ex. 2. Find the
probability
of
throwing
one 6 at least in six throws
with a die.
PROBABILITY.
517
The
probability
of not
throwing
6 is
^
in each throw. Hence the
/5\
6
probability
of not
throwing
a 6 in six throws
is, by
Art.
408,
IF
and therefore the
probability
of
throwing
one six at least is
Ex. 3. Find the chance of
drawing
2 white balls in succession from a
bag containing
5 red and 7 white
balls,
the balls drawn not
being
re-
placed.
7
The chance of
drawing
a white ball the first time is
^
;
and,
13
having
drawn a while ball the first
time,
there will be 5 red and 6
white balls
left,
and therefore the chance of
drawing
a white ball
/>
the second time will be .
Hence,
from Art.
409,
the chance of
drawing
two white balls in succession will be
^
x
^
=
jr-
.
i& 11 &&
[Compare
Ex.
4,
Art.
407.]
Ex. 4. There are two
bags,
one of which contains 5 red and 7 white
balls and the other 3 red and 12 white
balls,
and a ball is to be
drawn from one or other of the two
bags ;
find the chance of
drawing
a red ball.
The chance of
choosing
the first
bag
is
-
,
and if the first
bag
be
chosen the chance of
drawing
a red ball from it is
;
hence the
i&
chance of
drawing
a red ball from the first
bag
is
H
x
TO
=
57
2 \.i 2-i
Similarly
the chance of
drawing
a red ball from the second
bag
is
131
oXTe
=
Tn' Hence,
as these events are
mutually exclusive,
the
a 10 10
,
.
,
. 5 1 37
chance
required
is
-
+
-
=
.
Ex. 5. In two
bags
there are to be
put altogether
2 red and 10 white
balls,
neither
bag being empty.
How must the balls be divided so as
to
give
to a
person
who draws one ball from either
bag,
(1)
the least
chance and
(2)
the
greatest
chance of
drawing
a red ball.
[The
least chance is when one
bag
contains
only
one white
ball,
and the
greatest
chance is vrhen one
bag
contains
only
one red
ball,
I /
the chances
being
-
and
respectively.]
518
PROBABILITY.
410. When the
probability
of the
happening
of an
event in one trial is
known,
the
probability
of its
happen-
ing exactly
once, twice,
three
times,
&c. in n trials can be
at once written down.
For,
if
p
be the
probability
of the
happening
of the
event,
the
probability
of its
failing
is 1
p
=
q.
Hence,
from Art.
408,
the
probability
of its
happening
r times
and
failing
n r times in
any specified
order is
p
r
q*~
r
-
But the whole number of
ways
in which the event
could
happen
r times
exactly
in n trials is
n
C
r
,
and these
ways
are all
equally probable
and are
mutually
exclusive.
Hence the
probability
of the event
happening
r times
exactly
in n trials is
n
C
r
p
r
q*~
r
.
Thus,
if
(p
+
q)
n
be
expanded by
the binomial
theorem,
the successive terms will be the
probability
of the
happen-
ing
of the event
exactly
n
times,
n 1
times,
n 2
times,
&c. in n trials.
Cor. I. To find the most
probable
number of successes
and failures in n trials it is
only necessary
to find the
greatest
term in the
expansion
of
(p
+
q)*.
Cor. II. The
probability
of the event
happening
at
least r times in n trials is
p
n
+
n.p"q
+
^Ill)^-y
+...+
t-X2".
1.2
jr \n
r
Ex. 1. Find the chance of
throwing
10 with 4 dice.
The whole number of different throws is 6
4
,
for
any
one of six
numbers can be
exposed
on each
die;
also the number of
ways
of
throwing
10 is the coefficient of x
10
in
(x+xt+.^
+
x
6
)
4
,
for this co-
efficient
gives
the number of
ways
in which 10 can be made
up by
the
addition of four of the numbers
1, 2, ..., 6,
repetitions being
allowed.
Now the coefficient of x
w
in
(x+x*+
...
+Z
6
)
4
,
that is in ar
1
(
l
^'Y,
is
easily
found to be 80. Hence the
required
chance is
80
_
5
6.6.6.6~1'
Ex. 2. Find the chance of
throwing
8 with two dice.
Am, .
36
PROBABILITY. 519
Ex. 3. Find the chance of
throwing
10 with two dice. Ans. -=.
13
Ex. 4. Find the chance of
throwing
15 with three dice. Ans.
rr^
.
J.uo
Ex. 5. A and B each throws a
die;
shew that it is 7 : 5 that A'B throw
is not
greater
than B's.
Ex. 6. A and B each throw with two dice : find the chance that their
throws are
equal.
73
Ans.
m
.
Ex. 7. A and B have
equal
chances of
winning
a
single game
at tennis :
find the chance of A
winning
the 'set'
(1)
when A has won 5
games
and B has won
4,
(2)
when A has won 5
games
and B has won
3,
and
(3)
when A has won 4
games
and B has won 2.
Ans.
(1)|, (2)Z,(3)g.
Ex. 8. A and B have
equal
chances of
winning
a
single game ;
and A
wants 2
games
and B wants 3
games
to win a match : shew that it is
11 to 5 that A wins the match.
Ex. 9. A and B have
equal
chances of
winning
a
single game;
and A
wants n
games
and B wants n +
1
games
to win a match : shew that
, t
1. 3. 5...
(2?i-l)
.
, 1.3.5...(2n-l)
the odds on^ are
1+
a . 4 . 6 )..a,
*> 1-
2 . 4
.6...2n
'
3
Ex. 10. A's chance of
winning
a
single game against
B is -: find the
B
chance of his
winning
at least 2
games
out of 3.
. 81
Ans.
TTp;.
125
2
Ex. 11. A's chance of
winning
a
single game against
B is
r
: find the
o
chance of his
winning
at least 3
games
out of 5. 192
Ex. 12. What is the chance of
throwing
at least 2 sixes in 6 throws
with a die?
12281
AnS'
46656'
Ex. 13. A coin is tossed five times in succession : shew that it is an
even chance that three consecutive throws will be the same.
Ex. 14. Three men toss in succession for a
prize
which is to be
given
to the first who
gets
'heads'. Find their
respective
chances.
421
An
*-r r r
520 PROBABILITY.
411. The value of a
given
chance of
obtaining
a
given
sum of
money
is called the
expectation.
If
j-
is the chance of
obtaining
a sum of
money
M,
then the
expectation
is Mx = .
For if E be the
expectation
in one
trial,
E(a
+
b)
will
be the
expectation
in a
+ b trials. But the chance
being
r
,
the sum M
will,
on the
average,
be won a times in
a
+
b
every
a
+
6 trials
;
and hence the
expectation
in a
+
b
trials is Ma. Hence
JE(a
+
b)
=
Ma;
therefore
E=Mx
a
+
b'
Thus the
expectation
is the sum which
may
be won
multiplied by
the chance of
winning
it.
Ex. 1. A
bag
contains 5 white balls and 7 black ones. Find the
expectation
of a man who is allowed to draw a ball from the
bag
and
who is to receive one
shilling
if he draws a black
ball,
and a crown if
he draws a white one.
7
The chance of
drawing
a black ball is
;
and therefore the
19
expectation
from
drawing
a black ball is Id. The chance of
drawing
a white ball is
;
and therefore the
expectation
from
drawing
a
Km
white ball is 2*. Id.
Hence,
as these events are
exclusive,
the whole
expectation
is 2s. Sd.
Ex. 2. A
purse
contains 2
sovereigns,
3 half-crowns and 7
shillings.
What should be
paid
for
permission
to draw
(1)
one coin and
(2)
two coins? Am.
(1)
4s.
6$d. (2)
9s. Id.
Ex. 3. Two
persons
toss a
shilling
alternately
on condition that the
first who
gets
'heads' wins the
shilling:
find their
expectations.
Ant.
8d.,
4J.
Ex. 4. Two
persons
throw a die
alternately,
and the first who throws
C is to receive 11
shillings
: find their
expectations.
Ant.
6s.,
5t.
PROBABILITY.
521
412. Inverse
Probability.
When it is known that
an event has
happened
and that it must have followed
from some one of a certain number of
causes,
the deter-
mination of the
probabilities
of the different
possible
causes is said to be a
problem
of inverse
probability.
For
example,
it
may
be known that a black ball was drawn from
one or other of two
bags,
one of which was known to contain 2
black and 7 white balls and the other 5 black and 4 white balls
;
and
it
may
be
required
to determine the
probability
that the ball was
drawn from the first
bag.
Now,
if we
suppose
a
great number, 2N,
of
drawings
to be
made,
there will in the
long
run be N from each
bag.
But in N
drawings
2
from the first
bag
there
are,
on the
average,
g
N which
give
a black
y
ball
;
and in N
drawings
from the second
bag
there are
^
N which
y
2
give
a black ball.
Hence,
in the
long run,
-
N out of a total of
y
-N+-N black balls are due to
drawings
from the first
bag ;
thus
y j
the
probability
that the ball was drawn from the first
bag
is
We now
proceed
to the
general proposition
:
Let P
t
,
P
2
,.
.
.,
P
n
be the
probabilities of
the existence
of
n
causes,
which are
mutually
exclusive and are such that a
certain event must have
followed from
one
of
them;
and let
p1( p2
,
...,pn
be the
respective probabilities
that when one
of
the causes P
lt
P
2
,
.
..,
P
n
exists it will be
followed by
the
event in
question;
then on
any
occasion when the event is
known to have occurred the
probability of
the rih cause is
Let a
great
number N of trials be made
;
then the
first cause will exist in N.P
1
cases,
and the event will
follow in N.
P, .j9,
cases. So also the second cause exists
and the event follows in N. P
2
.
p.2
cases
;
and so on.
Hence the event is due to the rth cause in N.P
r .p
r
522
PROBABILITY.
cases out of a total of .A
r
(P
l pl
+
P
2p2
4-
. . .
+
P
n p^)
',
the
P v
probability
of the rth cause is therefore
^ p
-
.
Having
found the
probability
of the existence of each
of the different
causes,
the
probability
that the event
would occur on a second trial can be at once found.
For let P
r
'
be the
probability
of the existence of the
rth cause
;
then
pr
is the
probability
that the event will
happen
when the rth cause exists
;
and therefore
P/
.
p
r
is
the
probability
that the event will
happen
from the rth
cause.
Hence,
as the causes are
mutually
exclusive,
the
probability
that the event would
happen
on a second
trial is
Ex. 1. There are 3
bags
which are known to contain 2 white and 3
black,
4
\j-hite
and 1
black,
and 3 white and 7 black balls
respectively.
A ball was drawn at random from one of the
bags
and found to be a
black ball. Find the chance that it was drawn from the
bag
con-
taining
the most black balls.
Here P
1=P,=P3
=g.
Also
ft|, ft=g
and
2>3
=^j-
Hence the
required probability
is
1
_7
3' 10 7
13 11 1
7_
15
3' 5
+
3* 5
+
3* 10
Ex. 2. From a
bag
which is known to contain 4 balls each of which is
just
as
likely
to be black as
white,
a ball is drawn at random and
found to be white. Find the chance that the
bag
contained 3 white
and 1 black balls.
The
bag may
have contained
(1)
4
white,
(2)
3 white and 1
black,
(3)
2 white and 2
black,
(4)
1 white and 3
black,
and
(5)
4 black
;
and
the chances of these are
respectively -5
,
^ ,
-s
,
and
^
.
lo lo 1(3 lo lu
Art. 410. Also the chances of
drawing
a white ball
in these
311
different cases will be
1,
-
,
-
,
j
and
respectively.
*. -
1
/ T o
Hence the
required probability
= . =
-
.
1841ol4o
PROBABILITY.
523
413.
Probability
of
testimony.
The method of
dealing
with
questions relating
to the
credibility
of wit-
nesses will be seen from the
following examples
:
Ex. 1. A ball has been drawn at random from a
bag containing
99
black balls and 1 white
ball;
and a man whose statements are
accurate 9 times out of 10 asserts that the white ball was drawn.
Find the chance that the white ball was
really
drawn.
The
probability
that the white ball will
really
be drawn in
any
case
is
r^
,
and therefore the
probability
that the man will
truly
assert
1 9
that the white ball is drawn is
=7^-
x
^
.
JLUU JLU
99
The
probability
that the white ball will not be drawn is
-75;
,
and
100
therefore the
probability
that the man will
falsely
assert that the
99 1
white ball is drawn is
^
x
.
100 10
Hence as in Art. 412 the
required probability
is
1 Q
ioo
x
io i
ioo
x
16
+
ioo
x
10
Ex. 2. From a
bag containing
100 tickets numbered
1, 2,
...,
100
respectively,
a ticket has been drawn at random
;
and a
witness,
whose statements are accurate 9 times out of
10,
asserts that a
particular
ticket has been drawn. Find the chance that this ticket
was
really
drawn.
In 1000.tf trials the ticket in
question
will be drawn ION
times;
and the witness will
correctly
assert that it has been drawn 9N times.
The ticket will not be drawn in
990iV
cases,
and the witness will
make a
wrong
assertion in 99N of these cases
;
but there are 99
ways
of
making
a
wrong
assertion and these
may
all be
supposed
to be
equally likely;
hence the witness will
wrongly
assert that the
particular
ticket has been drawn in N cases. Hence the
required
9
probability
is
^7:,
so that the
probability
is in this case
equal
to the
probability
of the witness
speaking
the truth.
Ex. 3.
A
speaks
the truth three times out of
four,
and B five times
out of six
;
and
they agree
in
stating
that a white ball has been drawn
from a
bag
which was known to contain 1 white and 9 black balls.
Find the chance that the white ball was
really
drawn.
The
probability
that the white ball will be drawn in
any
case is
524 PROBABILITY.
,
and therefore the
probability
that A and B will
agree
in
truly
asserting
that a white ball is drawn is
y^
x
-
x
-
.
The
probability
that a black ball will
really
be drawn in
any
Q
case is
==;
and therefore the
probability
that A and B will
agree
in
911
falsely asserting
that a white ball is drawn is : x
-g
x
g
Hence,
as in Art.
412,
the
required probability
is
J_
3 5
10
X
4
X
6 5
I
3 5 9 1 1 8'
Ex. 4. A
speaks
truth three times out of
four,
and B five times out of
six;
and
they agree
in
stating
that a white ball has been drawn from
a
bag
which was known to contain 10 balls all of different colours,
white
being
one. What is the chance that a white ball was
really
drawn?
The
probability
that the white ball will
really
be drawn in
any
case is
,
and therefore the
probability
that A and J> will
agree
in
i ^ *5 1
truly asserting
that the white ball is drawn is
Tfj
x
T
x
H
=
ff
The
probability
that the white ball will not be drawn in
any
case
9
1
is
Y
The
probability
that A will make a
wrong
statement is
j
'.
hence,
as there are nine
ways
of
making
a
wrong
statement which
may
all be
supposed
to be
equally likely,
the chance that A will
wrongly
assert that a white ball is drawn is
j
x
-
. Therefore the
chance that A and B will
agree
in
falsely asserting
that a white ball
is drawn is
9_ JL_
I 1
10
X
4~x"y
X
6x9~2160'
J_
1 fi
Hence the
required
probability
is
Ex. 5. It is 3 to 1 that A
speaks
truth,
4 to 1 tbat B does and G to 1
that C does : find the
probability
that an event
really
took
place
which A and B assert to have
happened
and which C denies
;
the
event
being,
independently
of this
evidence,
as
likely
to have
happened
as not. Am. *.
PROBABILITY. 525
414. We shall conclude this
chapter by considering
the
following examples, referring
the reader who wishes
for fuller information on the
subject
of Probabilities to
the article in the
Encyclopaedia
Britannica,
and to Tod-
hunter's
History
of the Mathematical
Theory
of Proba-
bility.
Ex. 1. A
bag
contains n
balls,
and all numbers of white balls from
to n are
equally likely
;
find the chance that r white balls in succes-
sion will be
drawn,
the balls not
being replaced.
The chance that the
bag
contains s white balls is
7
;
and the
chance that r balls in succession will be drawn from a
bag
contain-
. s
(*-l) ...(- r+1)
ing
n balls of which s are white is
--
-^
n(;t-l)..
Hence the chance
required
is
n(n-l)...(ii-r
+
l) (n- 1) (n-2)...(n-
_
+
In(n-l)...(n-r+l)
n
(n- l)...(n-
Now
{1.2...r}
+
{2.3...(r
+
l)}
+ ... +
{(n-r+l)...(n-l)n}
=
(n-r
+
l)(n-r
+
2)...(+l)
r+1
Hence the
required
chance is
-
,
which is
independent
of the
whole number of balls in the
bag.
If it be known that r white balls in succession have been
drawn,
the
probability
of the next
drawing giving
a white ball can be at
once found from the
preceding
result.
For in a
great
number
N,
of
cases,
there will be r white balls in
succession in =
cases,
and r
+
1 white balls in succession in
^
N N
r+1
cases. Hence the
required
chance is = -.
---
=
.
Ex. 2. Two men A and
B,
who have a and b counters
respectively
to
begin with, play
a match
consisting
of
separate games,
none of which
can be
drawn,
and the winner of a
game
receives a counter from the
loser. Find their
respective
chances of
winning
the
match,
which is
supposed
to be continued until one of the
players
has no more
counters,
the odds
being p
:
q
that A wins
any particular game.
526 PROBABILITY.
Let A'B chance of ultimate success when he has n counters be U
B
.
Then A's chance of
winning
the next
game
is
,
and his chance
p
+
q
of ultimate success will then he
n+1
;
also A's chance of
losing
the
next
game
is
,
and his chance of ultimate success will then be
Hence
=-^-+i
+
p
+
q
" '
p
+
q
"
-
(j>
+
9)
u
w
+
9-i=0,
from which it follows that w
will be the coefficient of it" in the
expansion
of
p
provided
A and B be
properly
chosen.
A+Bx C D
Now
-
j
-
r
-
5
can be
expressed
in the form
--
h
--
;
*
-* l-x
and hence the coefficient ofx*isD
+ {-)
.
P\pJ
Thus u_
=
D
+
-
(-) ,
where C and D have to be
determined.
P \PJ
But it is obvious that A's chance of
winning
is zero if he has no
counters and
unity
if he has a
+
b,
so that u.=0 and w
a+t
=
l
;
hence
C C /<A+*
=
JD +
,
and 1=D
+
- ?
) ,
whence the values of C and D
I
d,
a,
are
found,
and we have
M!)
Hence A's chance of
winning
the
game
is
Similarly
B's chance of
winning
the
game
is
EXAMPLES XLI.
1. A and B throw
alternately
with two
dice,
and a
prize
is to be won
by
the one who first throws 8. Find their
respective
chances of
winning
if A throws first.
EXAMPLES.
527
2.
A,
B and C throw
alternately
with three
dice,
and a
prize
is to be won
by
the one who first throws 6. Find their
respective
chances of
winning
if
they
throw in the order
A,
,C.
3. Three white balls and five black are
placed
in a
bag,
and three men draw a ball in succession
(the
balls drawn not
being replaced)
until a white ball is drawn : shew that their
respective
chances are as 27 : 18 : 11.
4. What is the most
likely
number of sixes in 50 throws
of a die 1
5. Shew that with two dice the chance of
throwing
more
than 7 is
equal
to the chance of
throwing
less than 7.
6. In a
bag
there are three tickets numbered
1, 2,
3.
A ticket is drawn at random and
put
back
;
and this is done
four times : shew that it is 41 to 40 that the sum of the
numbers drawn is even.
7. From a
bag containing
100 tickets numbered
1, 2,
3, ...100,
two tickets are drawn at
random;
shew that it is 50
to 49 that the sum of the numbers on the tickets will be odd.
8. There are n tickets in a
bag
numbered
1, 2, ...,.
A
man draws two tickets
together
at
random,
and is to receive a
number of
shillings equal
to the
product
of the numbers he
draws : find the value of his
expectation.
9. An event is known to have
happened
n times in
n
years
: shew that the chance that it did not
happen
in a
particular year
is
(1 )
.
\ /
10. If
p things
be distributed at random
among p persons ;
shew that the chance that one at least of the
persons
will be
j P
P
~
\P
void is
i^-
.
P
11. A writes a letter to B and does not
get
an
answer;
assuming
that one letter in in is lost in
passing through
the
post,
shew that the chance that B received the letter is
_ 1
^
=-
,
it
being
considered certain that B would have answered
2m- I'
the letter if he had received it.
528
EXAMPLES.
12. From a
bag containing
3
sovereigns
and 3
shillings,
four coins are drawn at random and
placed
in a
purse;
two
coins are then drawn out of the
purse
and found to be both
sovereigns.
Shew that the value of the
expectation
of the
remaining
coins in the
purse
is 11s. 6c7.
13. From a
bag containing
4
sovereigns
and 4
shillings,
four coins are drawn at random and
placed
in a
purse;
two
coins are then drawn out of the
purse
and found to be both
sovereigns.
Shew that the
probable
value of the coins left in
the
bag
is
29^ shillings.
14. If three
points
are taken at random on a circle the
chance of their
lying
on the same semi-circle is
f
.
15. A rod is broken at random into three
pieces
: find the
chance that no one of the
pieces
is
greater
than the sum of the
other two.
16. A rod is broken at random into four
pieces
: find the
chance that no one of the
pieces
is
greater
than the sum of the
other three.
17. Three of the sides of a
regular polygon
of 4 sides are
chosen at
random; prove
that the chance that
they being
produced
will form an
acute-angled triangle
which will contain
.
(n-l)(n-2)
the
polygon
is .
=(-;-.
-
<\
-
2)
18. Out of m
persons
who are
sitting
in a circle three are
selected at
random;
prove
that the chance that no two of
those selected are
sitting
next one another is
7
-
=44
-
HT
(m l)(m-2)
19. If m odd
integers
and n even
integers
be written down
at
random,
shew that the chance that no two odd numbers are
\n
\n+
1
adjacent
to one another is
;
-
, ^
. m
being
J> n + 1.
-
'
20. If m
things
are distributed
amongst
a men and b
women,
shew that the chance that the number of
things
,
.
,,
.
l(b
+
a)
m
-(b-a)
m
received
by
the
group
of men is
odd,
is
^
v
-~
^
-
.
EXAMPLES. 529
21. The sum of two whole numbers is
100;
find the chance
that their
product
is
greater
than 1000.
22. The sum of two
positive quantities
is
given; prove
that it is an even chance that their
product
will not be less
than three-fourths of their
greatest product; prove
also that
the chance of their
product being
less than one-half their
greatest product
is 1
-^
.
23. Two men A and B have a and b counters
respectively,
and
they play
a match
consisting
of
separate games,
none of
which can be
drawn,
and the winner of a
game
receives a
counter from the loser. The two
players
have an
equal
chance
of
winning any single game,
and the match is continued until
one of the
players
has no more counters. Shew that A's chance of
winning
the match is
-y
.
24. An urn contains a number of balls which are known
to be either white or
black,
and all numbers are
equally likely.
If the result of
p
+
q drawings (the
balls not
being
replaced)
is
to
give p
white and
q
black
balls,
shew that the chance that the
next
drawing
will
give
a black ball is
25. Two sides
play
at a
game
in which the total number
of
points
that can be scored is 2m + 1
;
and the chances of
any
point being
scored
by
one side or the other are as 2m + 1 x
to 2m + 1
y,
where x and
y
are the
points already
scored
by
the
respective
sides. Shew that the chance that the side
which scores the first
point
will
just
win the
game
is
(2m!
2m +
1!)
2
S. A.
3-4
CHAPTER XXXI.
DETERMINANTS.
415. IF there are nine
quantities arranged
in a
square
as under :
then all the
possible products
of the
quantities
three to-
gether, subject
to the condition that of the three
quantities
in each
product
one and
only
one is taken from each of
the rows and one and
only
one from each of the
columns,
will be
0,6,0,, o,68
c
2)
a
2
&
8c,
,
0,6,0,, a,6,c,,
and
o,68c,.
Let now these
products
be considered to be
positive
or
negative
according
as there is an even or an odd number
of inversions of the natural order in the suffixes
;
then the
algebraic
sum of all the
products
will be
for there are no inversions in
a,62c,,
there is one inversion
in
0,6,0,
since 3
precedes
2,
there are two inversions in
0,6,0,
since 2 and 3 both
precede
1,
there is one inversion
in
0,6,0,
since 2
precedes
1,
there are two inversions in
0,6,0,
since 3
precedes
both 1 and
2,
and there are three
inversions
in
o,68c,
since 3
precedes
both 1 and 2 and 2
precedes
1.
DETERMINANTS. 531
The
expression (A)
is called the determinant of the
nine
quantities a,, a,,
&c.,
which are called its
elements;
and the
products apz
c
s
,
ap^,
&c. are called the terms of
the determinant.
416. Definition. If there are w*
quantities arranged
in a
square
as under :
the members of the same row
being distinguished by
the
same
letter,
and the members of the same column
by
the
same suffix
;
and if all the
possible products
of the
quan-
tities n at a time are taken
subject
to the condition that
of the n
quantities
in each
product
one and
only
one is
taken from
every
row and one and
only
one from
every
column,
and if the
sign
of each
product
is considered to be
positive
or
negative according
as there is an even or an odd
number of inversions of the natural order in the suffixes
;
then the
algebraic
sum of all the
products
so formed is
called the determinant of the n
2
quantities
or elements.
To denote that the n
8
quantities
are to be
operated
upon
in the manner above
described,
they
are enclosed
by
two
lines,
as in the above scheme.
The
diagonal through
the left-hand
top
corner is called
the
principal diagonal;
and the
product
of the n elements
a
i \>
c
a>
'
m
n
which lie
along
it,
is called the
principal
term of the determinant.
All the other terms can be formed in order from the
principal
term
by taking
the letters in their
alphabetical
order and
permuting
the suffixes in
every possible way:
on this account a determinant is sometimes
represented
by enclosing
its
principal
term in brackets
;
thus the
above determinant would be written
|X&2
c
3 ...wJ,
the
342
532
DETERMINANTS.
determinant is also often
represented by
the notation
When
only
one determinant is considered it is
generally
denoted
by
the
symbol
A.
A determinant is said to be of the nth order when
there are n elements in each of its rows or
columns,
and
therefore also n elements in each of its terms.
417. Since there are as
many
terms in a determinant
of the nth order as there are
permutations
of the n
suffixes,
it follows that there are
[n
terms in a determinant of the
nth order. There
are,
for
example,
six terms in a deter-
minant of the third order.
418. The law
by
which the
sign
of
any
term of a
determinant is found is
equivalent
to the
following
:
Take the elements in order
from
the successive rows
beginning
at the
first
;
then the
sign of any
term, is
positive
or
negative according
as there is an even or an odd number
of
inversions in the order
of
the columns
from
which the
elements are taken.
We will now shew that the words row and column
may
be
interchanged
in the above law. To
prove
this,
consider
any product,
for
example, a^c^d^f^
and its
equivalent
c,/2
6
8
cJ
4
a
B
e
6
,
where in the first form the letters follow the
alphabetical
order and in the second form the numbers
follow the natural order.
We have to shew that the number
of inversions
in the suffixes in the first form is the same as the number
of inversions of the
alphabetical
order in the second form.
This follows
immediately
from the fact that
if,
in the first
form,
any
suffix follow r suffixes
greater
than
itself;
then,
in the second
form,
the letter
corresponding
to that suffix
must
precede
r letters earlier than itself in
alphabetical
order.
Thus,
in the
example,
2 follows four suffixes
greater
than itself in a
6b,
i
c
l
d
i
e
tlfyt
and
/precedes
four letters earlier
than itself in
c,/j6s
d
4
a
8
e
e
.
DETERMINANTS. 533
Since the words rows and columns are
interchangeable
in the law which determines the
sign
of
any
term,
we have
the
following
Theorem. A determinant is unaltered
by changing
its
rows into columns and its columns into rows.
For
example
Ex. 1. Count the number of inversions in
2314,
3142 and 4231.
Am.
2, 3,
6.
Ex. 2. Count the number of inversions in
4132,
35142 and 5312G4.
Ans.
4, 6,
7.
Ex. 3. What are the
signs
of the terras
Ifg,
cdh and
ceg
in the
determinant a b c
d e
f
9
h k
[The
order of the columns is
231,
312 and
321.]
Ans.
+, +,
-.
Ex. 4. What are the
signs
of the terms
byiq,
celn and
dfkm
in the
determinant la b c d
\
?
/
j
k
m n
p q
[The
order of the columns is
2314,
3142 and
4231.]
Ans.
+, -,
-.
410. Theorem I.
If
in
any
term
of
a determinant
any
two
suffixes
be
interchanged,
another term
of
the determinant
will be obtained whose
sign
is
opposite
to that
of
the
original
term.
Let P . h
a
.
kp
be
any
term of a
determinant,
P
being
the
product
of all the elements
except
h
a
and
kp
; then,
by
interchanging
a and
{3
we have P .
hp
. k
a
. Now since
P . h
a
.
kp
is a term of the
determinant,
P can contain no
element from the rows of tis and k's and no element from
534 DETERMINANTS.
the a. or
ft
columns
;
and this is a sufficient condition that
Phpka.
should also be a term of the determinant.
We have now to shew that the two terms have
different
signs.
First
suppose
that two consecutive suffixes are inter-
changed.
Consider the term Ah
a
kpB
where A denotes the
product
of all the elements which
precede
h
a
and B the
product
of
all the elements which follow
kp. By
interchanging
a and
ft
we have
Ahpk
a
B,
which we have
already
found is a term
of the determinant.
Now the number of inversions in the two terms must
be the same so far as the suffixes contained in
A,
or in
B,
are
concerned,
whether
compared
with one another or with
a and
ft;
but there must be an inversion in one or other of
aft
and
fta.
but not in both. Hence the numbers of
the inversions in the two terms differ
by unity,
and therefore
the
signs
of the terms must be different.
Now
suppose
that two non-consecutive suffixes are
interchanged
;
and let there be r elements between the two
whose
suffixes,
a and
ft
suppose,
are to be
interchanged.
Then a will be
brought
into the
place
of
ft
by
r+I
in-
terchanges
ofconsecutive
suffixes,
and
ft
can then be
brought
into the
original place occupied by
a
by
r
interchanges
of consecutive suffixes
;
and therefore the
interchange
of
a and
ft
can be made
by
means of 2r
+
1,
that is
by
an odd
number,
of
interchanges
of successive suffixes.
But,
by
the
first
case,
each such
interchange gives
rise to a loss or
gain
of one inversion
;
and hence there must on the whole be a
loss or
gain
of an odd number of inversions : the
sign
of the
new term will therefore be different from the
sign
of the
original
term.
420. Theorem II. A determinant is unaltered in
absolute
value,
but is
changed
in
sign, by
the
interchange
of
any
two columns or
any
two rows.
Suppose
that in
any
determinant the rows in which
the letters h and k occur are
interchanged.
Then,
if
DETERMINANTS.
535
A . h
a
. B .
kp
. G be
any
term of the
original
determinant,
the term of the new determinant formed
by
the elements
which occur in the same
places
as before will be Ak
a
BhpC;
and these two terms must have the same
sign
in the two de-
terminants. Now
by
Art. 419 we know that A . k
a
. B.
hp
. C
is a term of the
original
determinant and that its
sign
is
different
from that of A . h
a
. B.
kp
. G. Hence
any
term of the new determinant is also a term of the
original
determinant but the
sign
of the term is different : the two
determinants must therefore be
equal
in absolute
magni-
tude but different in
sign.
The
proposition being
true for rows
is,
from Art
418,
true also for columns.
For
example
two 421. Theorem III. A
determinant,
in which
rows or two columns are
identical,
is
equal
to zero.
When two rows
(or
two
columns)
are
identical,
the
determinant is unaltered either in
sign
or
magnitude by
the
interchange
of these two rows
(or columns).
But,
by
Theorem
II,
the
interchange
of
any
two rows
(or columns)
of a determinant
changes
its
sign.
Thus the determinant
is not altered in value
by changing
its
sign
: its value
must therefore be zero.
Ex. 1. Find the value of a a
j
b 6
2
It is obvious that two rows would become
identical,
and
therefore
the determinant would
vanish,
if a =b. Hence A must be
equal
to
an
expression
which has a
-
b as a factor.
Similarly
b
-
c and c
-
a
must be factors of A. But A is
by inspection
seen to be of the third
degree
in
a, b,
c
;
hence A=L
(b-c)
(c-a)(a-b),
where L is
numerical. The
principal
term of A is 6c
a
and this is the
only
term which
gives
be'
2
,
and the coefficient of 6c
2
in L
(b
-
c) (c
-
a) (a
-
b)
is L
;
therefore L
=
1. Thus A
=
(b
-
c) (c
-
a) (a
-
b).
536
DETERMINANTS.
Ex. 2. Find the value of
DETERMINANTS. 537
A minor determinant
is said to be of the
first
order,
or
to be a
first
minor,
when one column and one row are
suppressed
;
it is said to be of the second
order,
or to be a
second
minor,
when two columns and two rows are
sup-
pressed
;
and so on.
The determinant obtained
by suppressing
the line and
the column
through any particular
element is called the
minor
of
that
element,
and will be denoted
by
A
z
where x
is the element in
question.
and 6., c,
are first minors of
Thus
3
i
,
aud are A.
,
and A
a
respectively.
424.
Development
of determinants.
determinant of the fourth order
Consider the
A certain number of the terms of A will contain a
x
;
let the sum of all these terms be a
1
.A
l
.
Similarly
let
the sum of all the terms which contain a
a
,
a,
and a
4
,
be
respectively
cr
2
. A
z
,
a
a
. A
8
and a
t
. A
t
.
Then,
since no term
can contain more than one of the letters
Oj,
n
t
,
a
a
,
a
4
we
have
A
=
Oj
A
j
+
a^A2
+
a
a
A
3
+
a
t
A
4
............
(
i
).
Now,
since no term of A which contains
a^
can contain
any
element from the column or the row
through
a
lt
it
follows that
every
term of A which contains
a,
is the
product
of a
x
and some term of A
a
,; conversely
the
product
of
a,
and
any
term, T,
of A
ai
will be a term of
A,
and the
sign
of the term
Oj
. T of A will be the same as the
sign
of
the term T of A
0i
,
for there is no
change
in the number of
538 DETERMINANTS.
inversions. Hence the sum of all the terms of A which
contain
a,
is
a,.
A
0l
.
So
also,
every
term of A which contains
a,
is the
product
of
a.,
and some term of
A^,
and the
product
of a
a
and
any
term, T,
of
A^
will be a term of
A,
but there is
one more inversion in the term
Oj
. T of A than there is in
the term T of A
0i
,
since 2
precedes
1. Hence the sum of
all the terms in A which contain
c^
is
a,
.
A^.
Similarly
the sum of all the terms of A which contain
a
s
are
a,
. A
03
;
and the sum of all the terms which con-
tain a
4
are a
t
. A
fll
.
Hence
(ii).
By
means of Articles 419 and
420,
we can shew in a
similar manner that
A
=
-
b,
A
6l
+
b
2
A
6t
-
6,
A
6j
+ 6
4
A
6<
=
a,
A
ai
-
6,
A
6i
+
c,
A
Ci
-
d,
A
dl
=
&c.
Cor.
By comparing (i)
and
(ii)
we see that the co-
factors of
the elements
a,, a,,
&c.,
are
equal
in
absolute
magnitude
to the minors
of
the same elements.
425. We have in the
previous
article considered the
case of a determinant of the fourth
order;
the
reasoning
is
however
perfectly general,
so that if A be a
determinant
of the nth order
having a,, a,,...,
a
n
for the elements of its
first row or column then will
A
=
So also
A
=
(- I)-
1
(k,
. A
fli
-
a
2
A
a
,
+
. ..
+
(
-
I)"
fli>
.
A,,
-
k,
Where k
lt A;,,...,
k
For
example
i
fl
a s
. . .
+
(- 1)-
1
k
n
are the elements of the rth
row.
=
0,
=<h (Vs
-
fcj 63+0, fcj 6,
DETERMINANTS.
539
Prove the
following
;
1.
3.
5.
7.
9.
1 1
540
DETERMINANTS.
Now consider the determinant which differs from the
original
determinant
only
in
having
the sih column
identical with the
rth;
then
A,, B,,
&c. will be the same
in the new determinant as in the
original
one.
The value of the new determinant will
therefore,
by
Art.
425,
be
equal
to
since a
r
=
a
t
,
b
r
=
b
t
,
&c.
But,
from Art
421,
we know that the new determinant
is zero.
Hence a
f
.
A,
+ b
r
.
B,
+
. . .
=
0.
Thus in the
determinant
A^a^jCjdJ,
we have
A=OA
+
aA- +
a
0= b
1
b
3
A
s
+b
4
A,
also
427. Theorem VI.
If
each element
of any
row
(or
column) of
a determinant be the sum
of
two
quantities,
the
determinant can be
expressed
as the sum
of
two deter-
minants
of
the same order.
It will be sufficient to take as an
example
the deter-
minant
0,+a,
6
X
By
Art.
424,
we
have,
if A
lt
A
t
,
A
t
,
be the co-factors of
the elements of the first
column,
A
=
(a,
+
or,)
A
l
+
(at
+
2
2)
A
z
+
(a,
+
or,) A,
a, b, c,
+
a,
a
2
b
2
c
2
2.
s
l>
3
c
3
DETERMINANTS.
541
Similarly
it can be
proved
that
c
l
=
*
542 DETERMINANTS.
Add the first row to each of the others
;
then
=A abed
26 2c 2d
2c 2d
2d
Ex. 3. Shew that
2c 2d
2d
=
2o6j
2c 2d
!
2d
=
0.
a
+ lb
Take the second row from the
third,
and then the first from
the second.
Ex.4. Shew that b
+
c
DETERMINANTS.
543
It is in the first
place
clear that a term of the determinant of the
sixth order will be obtained
by taking any
term of
[a^^sl
wHo. an
'J
term of
[e^/S^].
Thus A
=
[Oj^Cjl
.
[a^y^
together
with terms
involving
I, m, n,
&c.
;
and we have to shew that all terms
involving
any
of the letters
I, m,
n,
<fec. will vanish.
Now,
in
every
term of the minor of
/,
three elements must be
chosen from the last three
rows,
and two
only
of these can be chosen
from the last two columns
;
hence one of the three elements must be
zero,
and therefore
every
term of A
{
is zero. Hence the minor of
I,
and so also the minor of each of the elements
m, n,
&c. is zero
;
this
proves
that there are no terms
involving any
of the letters
I,
m,
n,
&c.
It can be
proved
in a similar manner that
any
determinant of the
2nth order is the
product
of two determinants of the nth
order,
provided every
element of one of the nth minors of the
original
determinant is zero.
430.
Multiplication
of determinants. We shall
consider the case of two determinants of the third order:
the method is however
perfectly general.
To
easpress
as a determinant
of
the third
order,
the
product of
the two determinants.
and
A.
=
A
7i
A 7]
We know from Art. 429 that
A,
A
=
-1
544
DETERMINANTS.
-1, ,
, -1,
o
,
o
,
, ,
, ,
,
,
which is
by
Art. 429
equivalent
to the
product
of
-1
-
j
-
1
1-10
-1
,
that is
1,
and
O^
+
Ooj83
+
0373
,
Hence the
required product
is the determinant
last
written.
Ex. 1.
Multiply by
x
y
z
z x
y
y
z x
The
required product
is
Y=ay
+
bz+cx,
and Z=
a c b
b a c
c b a
X Y Z
Z X Y
Y Z X
where A'
=
ax
+
by
+
cz,
Since
=x*+y
3
+z*-3xyz,
and the other determinants x
y
z
z x
y
y
z x
\
are of the same
form,
we see that the
product
of
any
two
expressions
of the form x
3
+
y
3
+z
3
-
3xyz
can be
expressed
in the same form.
[See
Art.
156,
Ex.
4.]
Ex. 2. Shew that 2bc
-
o
8
,
c
2
,
Form the
product
of
Ex. 3. Shew that
a,
-b,
c
c, -a,
b
b,
-c,
a
and
A
-3o6c)
a
.
-a,
be
-c,
a b
-b,
c a
s
,
where A
lt lt
&Q.
I, B, C,
' *
c,
are the co-factors of a
}
,
6j
,
&c. in the
expansion
of the determinant
DETERMINANTS. 545
A
l
B
1 Cj
A
2 .Bj
(7
a
o,, a,
For
)+<V,=[
and
AJbl
+
Aj3%
+ A
s
b
3
=&o.
=
Q
[Art. 426].^
Hence
[A^C^
.
\
431. The notation
is
employed
to denote
the
system
of four determinants
obtained
by omitting any
one of the
columns.
432. We conclude with the
following important appli-
cations of determinants.
Simultaneous
Equations
of the First
degree.
The solution of
any
number of
simultaneous
equations
of
the first
degree
can be at once obtained
by
means of the
foregoing properties
of determinants.
First take the case of the three
equations
Multiply
the
equations
in order
by
A^
A
2
,
A
a
,
where
A^,
A
s
are the co-factors of
a,,
a
2
,
a
8 respectively
in the
determinant
Then we have
by
addition
z
=
35
546
DETERMINANTS.
that is
[at
b
t
c
a]
x
=
[k
l
b
z
c
s
],
for from Art. 426 the coefficients of
y
and z are zero.
Similarly
we obtain
k&t
c
Jy-fa*t<&
and
^
[a,
6
2 cj
*
=
[a,
6
8 &J.
Now consider n
equations
of the form
a^
+
&,#
+
c^,
+
d^xt
+
=
k
r
As
before,
multiply
the
equations
in order
by .4,',
A
t
,
A
8
,
&c. the co-factors
respectively
of
a,, a,,
a
,
&c. in the
determinant
\al \
c
t
. .
.] ;
then we have
by
addition
the coefficients of
y,
z,
&c.
being
all zero
by
Art. 426.
-dull CO w
~~~
f f
-i
So also
[a. A;, CL...]
Ex. 1. Solve the
equations
=7,
The values of
x,
y,
t are
respectively
623
741
14 2 9
DETERMINANTS. 547
We have
548 DETERMINANTS.
It can be shewn in a similar manner that the condition
that n
equations
of the form
CLJK
+
b
ty
+
. . .
+
k
l
=
0,
with
(n 1)
unknown
quantities, may
be
simultaneously
true is
[OAJ8,...*.]
=
0.
434.
Sylvester's
method of Elimination. This is
a method
by
which # can be eliminated from
any
two
rational and
integral equations
in x. The method will be
understood from the
following examples.
Ex. 1. Eliminate x from the
equations
From the
given equations
we have
axP
+ bx^+cx
=0,
px
3
+
qx*
+
rx
=0,
and
px-
+
qx
+ r=0.
Now we
may
consider the different
powers
of x as so
many
different
unknown
quantities ;
and the result of
eliminating
x
3
,
x* and x from
the four last
equations
is
by
Art. 433
a b c
|
=
0.
a b c
\pqrO
\0pqr
[This
result is
equivalent
to that obtained in Art.
153,
Ex.
3.]
Ex. 2. Eliminate x from the
equations
nx
3
+bx- +
cx
+
d Q and
From the
given equations
we have
ax*
+
bx
a
+
cx*
+
dx
-0,
ox
3
+
bx*
+
ex
+
d
=
0,
px
4
+
qx
3
+ rx*
=
0,
px
3
+
qx
a
+rx
=0,
Eliminating x*,
x
3
, x-,
x from the five last
equations
as if the
different
powers
of x were so
many
different unknown
quantities,
we
have the condition
a b c d =0.
a b c d
p q
r
p q
r
p q
r
EXAMPLES. 549
EXAMPLES XLH.
1. Shew that
550
EXAMPLES.
551
21. Shew that
a
2
a
2
-(6-c)
a
be
52
&_(c-a)
a
ca
c
a
c
a
-(a-b)
a
ab
22. Shew that
(6-c)
2
(0-&)
a
(a-c)
2
(b a)* (c a)
2
(b c)
2
(c-a)
a
(c-&)
a
(a-b)*
23. Shew
that,
if
any
determinant
vanishes,
the minors of
any
one
row will be
proportional
to the minors of
any
other row.
24.
25.
26. Shew that the determinants
a b c
z a b
y
a c
x b c
Shew that
CHAPTER XXXII.
THEORY OF
EQUATIONS.
435. ANY
algebraical expression
which contains x is
called a
function
of
x,
and is denoted for
brevity by f
(x),
F(x), <f>(x),
or some similar
symbol.
The most
general
rational and
integral expression
[Art. 75]
of the nth
degree
in a;
may
be written
1
+
. . .
+
a
n ,
where
Oo, Oj, a,,...
do not contain as.
Since all the terms of
any equation
can be
transposed
to one
side,
every equation
of the nth
degree
in x can be
written in the form
. . . + a
n
=
0,
where n is
any integer,
and the coefficients
Oo, a,, a,...
do not contain x.
Now
any equation
in a; is
equivalent
to that obtained
by dividing every
one of its terms
by any quantity
which
does not contain x
; and,
if we divide the left side of the
above
equation by
o<>,
the coefficient of x
n
,
we shall obtain
the
equation
of the wth
degree
in its
simplest
form,
namely
of
1
+p
1
x
n
-
1
+p
a
x
n-*
+
...
+
p
n
=
0,
where
p
lt
p
t>
^
3 ,...
do not contain
x,
but are otherwise
unrestricted.
-
THEORY OF
EQUATIONS.
553
436. If we assume the fundamental theorem* that
every equation
has a root real or
imaginary,
it is
easy
to
prove
that an
equation
of the nth
degree
has n roots.
For
suppose
the
equation
to be
f(x)
=
0,
where
f(x)
=
x
n
+p
1
x
n
~
1
+p&*-*
+
. . .
+p
n
.
Since
f(x)
=
has a
root,
a
l
suppose,
we have
f(ai)
=
0,
and therefore
[Art. 88] f (x)
must be divisible
by
x
a^,
so that
f(x)
=
(x
a^)
(#),
where
<(#)
is an
integral
function of x and of the
(n
l)th
degree. Similarly,
since the
equation <f> (x)
=
has a
root,
a
2
suppose,
we
have
(x)
=
(x
a^)-^r (x),
where
ty (x)
is an
integral
function of x of the
(n 2)th
degree.
Hence
f(x)
=
(x
-
Oi) (x
-
a
2
) ^r (x).
Proceeding
in this
way
we shall find n factors of
f(x)
of
the form x a
lt
and we have
finally
/()
=
(x
-
d) (x
-
a
2
).
.
.(x
-
a
n
).
It is now clear that
Oj,
a
z ,...,
a
n
are roots of the
equation f(x)
=
;
also no other value of x will make
f(x)
vanish,
so that the
equation
can
only
have these
n roots.
In the above the
quantities
a^,
a
z ,
a
3 ,...
need not be
all different from one
another;
but if the factors x-
a^,
x
a^x Ot,
&c. be
repeated p, q,
r,
&c. times
respectively
in
/(#),
we must have
f(x) =(x- ajr (x
-
a
2
)9 (x
-
a
3
)
r
. .
.,
where + +r +
...=n.
The
equation/(x)
=
has in this case
p
roots each
a,,
q
roots each a
a , &c.,
the whole number of roots
being
p+q
+ r+
... =n.
*
Proofs of this fundamental
proposition
have been
sriven
by Cauchy,
Clifford and others : the proofs are. however. Ions' and difficult.
554 THEORY OF
EQUATIONS.
437. Relations between the roots and the coeffi-
cients of an
equation.
We have seen that if
a^, a, , a,,...
be the roots of the
equation /(#)
=
;
then
f(x)
=
(x- a^ (x
-
a,).
.
.(as
-
a
n
).
Hence
[Art. 260]
-*-...+(- 1)
S
n,
where S
r
is the sum of all the
products
of
Oj, Oj, a,,...
taken r
together.
Equating
the coefficients of the different
powers
of x
on the two sides of the above
identity,
we have
438.
By
means of the relations obtained in Art.
437,
which
give
the
values of certain symmetrical functions of the roots of an
equation
in terms of its
coefficients,
the values of
many
other
symmetrical
functions of the roots can be
easily
obtained without
knowing
the
roots themselves.
The
following
are
simple examples
:
Ex. 1. If
a, b,
c be the roots of the
equation
x?+px*+qx+r=0,
find the value of
(i)
So
2
and
(ii)
Za
2
*
2
.
We have a
+
b+c=-pt
bc +
ca
+
ab
=
q
and abc
=
-r.
Hence
Also,
ZW:
2
=
(be
+
ca
+
ab)*
-
2abc
(a
+
b
+
c)
=
q*
-
2pr.
Ex. 2. If
a, b, c,...
be the roots of
z"+j>1
z-
1
+p2
3*-
8
+...+j7,,=0,
find the values of Za* and 2a>.
We know that 2a=
-pj, Sa&=jj,
and 2a6c=
-
p9
.
Now
(Za)
J
=(a
+
& + c +
...)
z
=Za
>
+2Sa6
[Art. 65];
Za
2
=
(Zo)
-
2Za6
=p^
-
2j>,.
Again
Za
a
. Zo
=
Za
3
+ Za
2
6,
and Za
2
fc
=
2a6 . Za
-
3Za6c.
THEORY OP
EQUATIONS.
555
[For
in 2oi . 2a there can
only
be terms of the
types
a?b and abc
;
of
these the term a?b will occur
once,
but the term abc will occur three
times,
for we can take either a or b or c from 2a and
multiply by be,
ca or ab
respectively
from 2a6. Thus 2a& . Sa= 2a
2
6
+
SZafcc.J
Hence
2a*=Sa
8
. 2a
-
2a& . 2a
+
32c&c=
(pf
-
2p2
) ( -pj
-
p, ( -pj
-
3p,.
439. Theorem. If there 'are
any
n
quantities
Oi,
a,, 03, &c.,
and m be
any positive integer
not
greater
than
n then will
The
following
relations hold
good
:
k..[A].
To
prove
the first relation it is
only necessary
to notice
that the
product
Sa
a
. 2a
1
tnr
-
1
can
only give
rise to terms
of the
types
of"
and
a^^a^;
also
every
term of either
type
will
occur,
and no term can occur more than once.
Thus
Set!
.
Soi"*-
1
=
Sd!
+
W*-
1
03.
The other
relations,
except
the
last,
will be seen to be
true in a similar manner.
Also,
the
product
SojO.,... 0^-1- 2^
can
only give
rise to terms of the
types
a
1
2
a
2
a
3
...a
m_i
and o
1
a
2
...a
ni
;
the first of these terms can
only
occur
once,
namely
as
aia2
a
8
...a
wl
_
1
x
Oj ;
the second term
will, however,
occur
m
times,
for we
get
the term
by taking. any
one of the
m factors it contains from
Sc^
and
multiplying
this
by
the
proper
term of o
1
o
9
...a
m
_
1
.
Hence
_i
+
m.
r^a^.
. .a
556 THEORY OF
EQUATIONS.
From the relations
[A],
we have at once
+
So^'S
...a
m
..................
[Bj.
If now
Oi, Oj, a,,
&c. be the n roots of the
equation
aP
+
frx"-
1
+
^a*;"-
2
+
...
+p
n
=
0,
we know that
2oi
=
Pn
^,a
1
a
t
=p
3 ,
So
1
a
2
a
8
=
j?
8 ,
&c.
Hence,
by substituting
in
[B]
and
transposing
we have
So,
+P!
.
Sc^"*-
1
+pa
1
T -*
+
. . .
+
P^-I^OI
m
=
Q ............
[C].
The formula
[C] gives
the sum of the mth
powers
of
the roots of an
equation
of the rcth
degree [m } n\
in
terms of the coefficients and the sums of lower
powers
of
the roots.
[See
also Art.
471.]
The sum of the mth
powers
of the roots of an
equation
can therefore be obtained from the formulae
+
jOjSo!
4-
Bp,
=
0,
4
=
0,
If we eliminate
equations
we have
and from the first three
Pi P 3ft
+
2
1
p
l
2p
t
1
ft
To find
2oi
m
we must eliminate
2oi
1
from the first w
equations,
and we have
=
So/*
j,
|=0.
1
ft
THEORY OF
EQUATIONS.
557
Pi j
2
1
P!
1
=
0.
0... 1
p
l
The coefficient of
Saj
m
is a determinant of which all
the elements on one side of its
principal diagonal
are
zeros,
the elements
along
the
principal diagonal being
all
equal
to
1;
the determinant is therefore
equal
to 1. Hence
Scij"*
is
equal
to an
integral
function
ofp
l}
p
z ,
&c.
If m be
greater
than n the relation
corresponding
to
[C]
can be
very easily
obtained.
For,
since
c^,
a
a>
... are
roots of
/(#)
=
0,
we have n
equations
of the
type
<h
n
+^ii
n
~
1
+
^2i
n
~
2
+ +
Pn
=
0.
Multiply by
aj
n
~
n
, a.,"
1
"
71
,...
respectively;
then we shall
have n
equations
of the
type
=
0.
Hence,
by
addition,
we have
2X
1"
+
pSaJ"-
1
+
p^aj
1^-*
+
...+p
n
By
means of the relations
[C]
and
[D],
which were
first
given by
Newton,
it is
easily
seen that the sum
of
the
rath
powers of
the roots
of any equation
can be
expressed
as
a rational and
integral function of
the
coefficients,
m
being
any positive integer.
440.
Any
rational and
integral symmetrical
function
of the roots of an
equation
can be
expressed
in terms of
the coefficients
by
means of the relations
,
=
-p
s ,
&c.
Consider the
symmetric
functions of the third
degree.
558 THEORY OF
EQUATIONS.
It is
easily
seen that
pf
=
Sc^
8
+
Thus we have three
equations
to determine
Sc^*, So^o,
and
SoiOjOs,
and these are the
only symmetrical
functions
of the third
degree.
Similarly
each of the
products pf, pSp
a ,
Pip
s ,
p*
and
p
4
can be
expressed
in terms of
symmetric
functions of the
fourth
degree,
and there will be as
many
such
equations
as
there are
symmetric functions
of the fourth
degree.
The same will hold
good
with
respect
to
symmetric
functions of
any
other
degree.
The sum of the suffixes of the
p's
will in all cases be
equal
to the
degree
of the
symmetrical
function.
441. A rational and
integral symmetrical
function of
the roots of an
equation
can also be
expressed
in terms of
sums of
powers
of the
roots,
and thence
by
Newton's
Theorem in terms of the coefficients of the
equation.
The method will be seen from the
following examples.
Ex. 1.
Express
Za/a.,'
in terms of
Za^, Za^
and 2a
1
'H
'.
Thus
2ojPaa
=
2aj
If, however,
p=q
;
then we have
Thus 2a
1
"a
2P=i(2o1P)
a
-i2rt1
2
P.
Ex. 2.
Express
Za
1
p
a.
! a,
p
in terms of the
sumv
of
powers
of the
separate
roots.
THEORY OF
EQUATIONS.
559
Hence S
p
.S
q
.S
r
=
2o
1
I
*
Hence,
from Ex.
1,
Zafafaf^
S
p
. 8
q
. S
r
-
S
M
. S
r
-
The above will
only
hold
good
when
p, q,
r are all different. If
=
q
=
r we shall have
V
=
S
3J>
+
ZVapaf
+
SZafafaf
;
J (Sp
3
-
3S
2P
TRANSFORMATION OF
EQUATIONS.
442. We now consider some cases in which an
equa-
tion is to be found such that its roots are connected with
the roots of a
given equation
in some
specified
manner.
I. To
find
an
equation
whose roots are those
of
a
given equation
with
contrary signs.
If the
given equation
be
f(x)
=
0,
the
required equa-
tion will be
f( y)
=
0.
For,
if a be
any
root of the
given equation
so
that/(a)
=
0,
then a will be a root of
/(-3/)
=
?.
Thus if the
given equation
be
Pox?
+
Pitc"-
1
+
p.zx
n~*
+
......
+p
n =Q,
the
required equation
will be
-
y)*-
8
+ ......
-f
p
n
=
o,
or
p y
n
-p
t
y
n
~
l
+p^y
n^
-
......
+
(- l)
n
p
n
=
0.
H. To
find
an
equation
whose roots are those
of
a
given equation
each
multiplied by
a
given quantity.
Let
/(#)
=
be the
given equation,
and let c be the
quantity by
which each of its roots is to be
multiplied.
f/ /7/\
Let
y
=
cx,
or
^
=
#;
then/
(-)
=
is the
equation
c \c
/
560 THEORY OF
EQUATIONS.
required.
For,
if a be
any
root of
/(x)
=
0,
so that
/(a)
=
0,
ac will be a root
off(-}
=
0.
\c/
Thus,
if the
given equation
be
the
required equation
will be
or
p y
n
-f
piCy
n
~
l
+
p^^*
+ ...... +
p
n
c
n
=
0.
The above transformation is useful for
getting
rid of fractional
coefficients.
Ex. Find the
equation
whose roots are the roots of
each
multiplied by
c.
The
required equation
is
We can now choose c so that all the coefficients
may
be
integers
;
the smallest
possible
value of c is
easily
seen to be 6.
III. To
find
an
equation
whose roots are those
of
a
given equation
each diminished
by
the same
given quantity.
Let
f
(x)
=
be the
given equation,
and let c be the
quantity by
which each of its roots is to be diminished.
Let
y
=
xc,
or x
=
y+c;
then
/ (y
+
c)
=
will be
the
equation required.
For,
if a be
any
root of
/(x)
=
0,
so
that/
(a)
=
0,
a c will be a root
off(y
+
c)
=
0.
An
expeditious
method of
finding f(y
+
c)
will be
given
later on.
[Art. 472.]
The chief use of above transformation is in
finding
approximate
solutions of numerical
equations
;
it can also
be used to obtain from
any given equation
another
equa-
tion in which a
particular
term is absent.
Ex. Find the
equation
whose roots are those of x
:l
-3x
2
-9x + 5=0
each diminished
by
e,
and find what c must be in order that in
THEORY OF
EQUATIONS.
561
the transformed
equation (i)
the sum of the
roots,
and
(ii)
the
sum of the
products
two
together
of the
roots, may
be zero.
The
equation required
is
f(y
+
c)=Q,
that is
or
The sum of the roots will be zero if the coefficient of
y*
be zero
;
that
is,
if c
=
l.
The sum of the
products
two
together
of the roots will be
zero if the coefficient of
y
be
zero;
that
is,
if c
2
-2c-3=0,
or
(c-3)(c
+
l)=0.
IV. To
find
an
equation
whose roots are the
reciprocals
of
the roots
of
a
given equation.
Let
/(a?)
=
be the
given equation.
Then the
equa-
tion
/(-)
=0 is satisfied
by
the
reciprocal
of
any
value of
\xj
x which satisfies the
original equation.
This transformation enables us to find the sum of
any
negative power
of the roots of the
equation f(x)
0,
for
we have
only
to find the sum of the
corresponding positive
power
of the roots of the
equation/
(
-
)
=
0.
443. A
reciprocal equation
is one in which the
reciprocal
of
any
root is also a root.
To
find
the conditions that an
equation may
be a
reciprocal equation.
Let the
equation
be
2
+
...... +
p
n
=
0.
Then the
equation
whose roots are the
reciprocals
of
the roots of the
given equation
is
or,
multiplying by
x
n
,
Po
+
PI&+PV&
+
......
+p
n
^
n
0.
s. A. 36
562 THEORY OF
EQUATIONS.
The
equation
last written must be the same as the
original equation,
the ratio of
corresponding
coefficients
must therefore be the same
throughout.
Thus
Pn Pn-i Pn-a PO
From the first and last we have
Pn=p*,
so that
Pn
=
+
PO>
whence it follows that the coefficients are the
same when read backwards as
forwards,
or else that all
the coefficients read in order backwards differ in
sign only
from the coefficients read in order forwards. These two
forms of
reciprocal equations
are often said to be of the
first and of the second class
respectively.
444. The
following important properties
of
reciprocal
equations
can
easily
be
proved.
I. A
reciprocal equation
of the first class and of odd
degree
has one
root
equal
to
-
1.
II. A
reciprocal equation
of the second class and of odd
degree
has
one root
equal
to
+
1.
III. A
reciprocal equation
of the second class and of even
degree
has
the two roots 1.
[These
follow at once from Art.
87.]
IV. After
rejecting
the factor
corresponding
to the roots
given
in
I,
II, III,
we are in all cases left with a
reciprocal equation
of the
first class and of even
degree.
Y. The
problem
of
solving
a
reciprocal equation
of the first class and
of even
degree can,
by
means of the substitution x+x~
1
=y,
be
reduced to that of
solving
an
equation
of half the dimensions. For
the
equation may
be written
(**
+
1)
+
Oj (X
4*-1
+
X)
+ . . .
=
0.
Divide
by
**
;
then
a
(x
+
ar)
+
Ol (x"-
1
+
ar*)
+
. . .
=
0.
Now,
if
x+x~
1
=y,
x
s
+x~
2
=y
a
-2;
and,
from the
general
relation
x
+ *-
=
(x-i
+
-"--i) (x
+
*-)
-
(a:"-
a
+
a-***)
,
it follows
by
induction that x
n
+x~
n
can be
expressed
as a rational
and
integral
expression
of the nth
degree
in
y.
THEORY OF
EQUATIONS.
563
Ex. Solve the
equation
6x
-
25s
8
+
Six*
-
3Lt
a
+
25x -6=0.
As in
III,
the
expression
on the left has the factor z
a
-
1 corre-
sponding
to the roots 1. Thus we have
6
(x
6
-
1)
-
25x
(**
-
1)
+
Six*
(x
2
-
1)
=
0.
Hence the
required
roots are 1 and the roots of
6x*
-
25x + 37s
2
-
25x
+
6
=
0.
Divide
by
x
3
;
then
Put
x+-=y;
.: a:
a
+-=j/
2
-2.
X X"
Hence
6y
2
-25y
+
25
=
0;
1 5 1
From x
+
-
=
-,
we nave ar=2 or
g
.
9m
&
From
+-
=
-,
we have x=-=
(5\/-ll).
x o
o
j j
_
Thus the
required
roots are
1, 2,
s
,
-
(5
v
-
11).
A O
445. The method of
dealing
with other cases of trans-
formation will be seen from the
following examples.
Ex. 1. If
a, 6,
c be the roots of the
equation
x
3
+px
2
+
qx
+
r=0,
find
the
equation
whose roots are
be, ca,
ab.
Since bc=
=
-,
if we
put y
=
-,
the three values of
y
cor-
a a x
responding
to the values
o, 6,
c of x will be
be, ca,
ab. Hence the
T
equation required
will be obtained
by substituting
-
for x in the
given equation,
so that the
required equation
ia
or
Ex. 2. Find the
equation
whose roots are the
squares
of the roots of
the
equation
x
3
+px
z
+
qx
+
r
=
Q.
We have x
(x?
+
q)
=
-
(px*
+
r);
Now
put y=x
2
,
and we have the
required equation, namely
.
362
564 THEORY OF
EQUATIONS.
Ex. 3. If
a, 6,
c be the roots of
x*+px*
+
qx+r=0,
find the
equation
whose roots are
a(b
+
c),
b
(c
+
a),
c
(a
+
6).
a
(b
+
c)=a (-p a)
;
&c.
Hence,
if we
put y
=
x(-p-x), y
will have the values
required
provided
x is restricted to the three values
a, 6,
c
;
that is
provided
x satisfies the
equation
Thus if we eliminate x between the
given equation
and the
equa-
tion
x*+px+y=0,
we shall
get
the
required equation
in
y.
Multiply
the second
equation by
x and
subtract;
then
(y-q)x=r.
Now substitute for x in the second
equation,
and we obtain the
equation required, namely
r*+pr(y-q)+y(y-q)*=0.
EXAMPLES
1. If
Oj,
a
s
,
a
3
be the roots of the
equation x*+px
+
q=Q,
find
the values of
(i) (ai
+a
t)(a,
+
a
1)(a1
+
a^. (ii)
(iii) Sa^.
(iv)
2^*. (v) 2V. () SoiS- ()
(viii)
2
(a/
-
a,^) (a,
-
OjOj). (ix) (a^
-
a-jO,) (a,
-
a^) (a,*
-
(
X
)
s .
(xi)
2
--
.
(xii)
a
s +a,
-
8. Find the sum
(i)
of the
squares, (ii)
of the cubes and
(iii)
of
the fourth
powers
of the roots of the
equation x*+px
+
q=0.
3. If
a, b,
c be the roots of the
equation x*+px*
+
qx+r=Q,
find
the values of
(i) (6+c-3o)
I 1\ /I 1\ /I 1
4. Find the sum of the
squares
and the sum of the cubes of the roots
of the
equations
(i)
i-14x +
8=0.
(ii)
x<
-
22s
2
+ 84x- 49
=
0.
THEORY OF
EQUATIONS.
565
0. If
a,
b, c,...
be the roots of the
equation
_,=0,
find the values of
(i)
2a
J
.
(ii)
Sa
s
.
(iii)
S
-^.
(iv)
Sy.
(v)
S^-.
(vi)
Sp.
6. Find the
equation
each of whose roots exceeds
by
2 a root of
the
equation
a;
8
-4a;
2
+3a;-l=0.
7. Find the
equation
whose roots are those of the
equation
each
multiplied by c,
and find the least value of c in order that the
resulting equation may
have
integral
coefficients with
unity
for the coef-
ficient of the
highest power.
8. If
a, b,
c be the roots of the
equation
x
3
+
px*
+
qx+r=0,
find
the
equation
whose roots are
(i) be, co,
oft.
(ii)
b
+
c,
c +
a,
a
+
b.
(Hi)
(iv)
a
(b
+
c), b(e
+
a), c(a+ b), (v)
6
2
+
c,
c
2
+
a
2
,
o
2
+
6
2
.
(vi)
be
-
a
2
,
ca
-
b*,
ab
-
c
2
.
9. If
a, b, c,
d be the roots of the
equation
find the
equation
whose roots are
(i)
b
+
c
+
d,
&G.
(ii)
b
+
c
+
d-2a,
<fec.
(iii)
6*
+
c
2
+
d
2
-o,
&c.
1O. Find the
equation
whose roots are the cubes of the roots of
the
equation
X
s
+px^
+
qx
+ r=0.
446. In
any equation
with real
coefficients
imaginary
roots occur in
pairs.
For,
if a
+
b\i ^1 be a root
of/ (x)
=
0,
x
-
a
-
b V~l
will be a factor of
f(x),
and therefore
[Art. 193]
x a
+
b ^T^ 1 will also be a
factor,
whence it follows
that a b "J 1 is also a root of
f
(x)
=
0.
Corresponding
to the two roots a b V 1 of
f(x)
=
0,
f(x)
will have the real
quadratic factor {(x
of
+
6
2
}.
566 THEORY OF
EQUATIONS.
447. In
any equation
with rational
coefficients
quad-
ratic surd roots occur in
pairs.
For,
if a
+
*Jb
be a root of
/(x)
=
0, Jb
being
irra-
tional,
x a
V&
will be a factor of
/(x),
and therefore
[Art. 179]
x a
+
\/b
will also be a factor
off(x),
whence
it follows that a
\/b
will also be a root of
f(x)
0.
Corresponding
to the roots a
\/b,
f(x)
will have the
rational
quadratic factor {(x a)
2
&}.
Ex. 1. Solve the
equation
x
4
-
2s
3
-
22s
2
+
62x
-
15=0,
having given
that one root is
2+^/3.
Since both 2
+
^/3
and 2
-
x
/3 are roots of the
equation,
that is
2
-
4x +
1,
mnst be a factor of the left-hand member of the
equation.
Thus we have
(x
2
-
4x
+
1) (a:
2
+
2x
-
15)
=
0.
Whence the roots
required
are
2*^3
and the roots of
Ex. 2. Solve the
equation
2x*
-
15x
a
+
46x
-
42
=
0,
having given
that
one root is 3
+
>J
-6.
Since 3
^/
-
5 are roots of the
equation
(x
-
3
-
V^S) (x
-
3
+
V^)
must be a factor of the left-hand member of the
equation,
which
may
be written
{(x-3)
2
+
5}(2x-3)
=
0.
o
Whence the roots
required
are 3
\l~-5> 5.
I
Ex. 3. Solve the
equation
x
8
-
4x
-
llx*
+
40*
3
+
llz
2
-
4x
-
1
=
0,
having given
that one root is
^/2 +^/3.
If
Ja+
Jb
be a root of
any equation
with rational
coefficients,
tja
and
*Jb
not
being
similar
surds,
then
^^6
will all four be
roots.
Hence in the
present
case
(x
-
^2
-
that is ;r
4
-10x
2
+ l will be a factor of
/(x).
The
equation may
therefore be written
(x*
-
10x
a
+
1) (x*
-
4x
-
1)
=
0,
whence the roots aie
,^2*^/3,
THEORY OF
EQUATIONS. 567
Ex.4. Solve x
4
-*
8
-
9^ -14*
+
8
=
0,
having given
that one root is
-
1
+
$3.
x
+
1
-
4/3
is a factor of
f(x)
;
and
therefore,
as
/ (or)
is
rational,
the rational
expression
of lowest
degree
of which x +
1
-
j/3
is a
factor,
namely
the
expression (ar
+
l)
3
-3,
must be a factor of
f(x).
Thus we have
Thus the roots are
4,
-
where u is an
imaginary
cube root of
unity.
448. Roots common to two
equations.
If the
two
equations f(x)
=
and
<f) (x)
=
have one or more
roots in
common,
f
(x)
and <
(x)
must have a common
factor,
which will be found
by
the
process
of Art. 98.
Ex. Find the common roots of the
equations
a? -3x
2
-
10*
+
24=0 and ar
8
-
6a?
-
40x
+ 192
=
0.
The H. c. F. of the left-hand members will be found to be x
-
4.
Hence x
=
4
gives
the common root.
449. When it is known that two roots of an
equation
are connected
by any given
relation,
these
roots can be
found.
Ex. 1. Solve the cubic x
8
-
3z
a
-
Wx
+
24
=
0,
having given
that one
root is double another.
Let a and b be the two roots and let a
=
26.
Then,
since a is a root of the
given equation
a
s
-3a
8
-10a
+
24
=
........................
(i).
Also,
since 6 is a
root,
or a
3
-6a
2
-40a
+ 192=0
(ii).
The factor common to the left-hand members of
(i)
and
(ii)
will
be found to be a -4. Thus a=4 and 6
=
2;
the
remaining
root of
the cubic is then
easily
found to be
-
3.
Ex. 2. Solve the cubic 2a;
3
-
15x
2
+
37o;-30=0, having given
that the
roots are in A. p.
The sum of the roots is
equal
to three times the mean
root,
568 THEORY OF
EQUATIONS.
a
suppose.
Thus 3a
=
-
ir
,
whence a
=
o-
Divide
/(x) by
the
i t
factor 2x
-
5,
and the
remaining
roots are
given by
a;
2
-
5x
+ 6.
Hence the roots are
2,
4J
,
3.
m
In the
general
case
suppose
that a and b are roots
of the
equation/(x)
=
connected
by
the relation
b=$(a).
Then
f(x)
=
and
/ {< (#)}
=
have a common
root,
namely
a;
and this common root can be found as in Art.
448. Thus a and
<f> (a)
can both be found.
Ex.
(i).
Find the condition that the roots of x
3
+px'
2
+
qx
+r=Q
may
be
(i)
in Arithmetic
Progression, (ii)
in Geometrical
Progression.
Let
o, I,
c be the roots in order of
magnitude.
(i)
o
+
6
+
c
=
36;
/. 6
=
-.
Hence,
as b is a
root,
we have
whence
2p
-
Qpq
+ 27r
=
0.
(ii)
abc
=
b*;
.:
6=4/17.
Hence,
as & is a
root,
we have
whence
p
3
r=q
3
.
450. Commensurable roots. When the coefficients
of an
equation
are all rational the commensurable roots
can
easily
be found.
It is at once seen that an
equation
with
integral
coefficients and with
unity
for the coefficient of the first
term cannot have a.
fractional
root.
For if
j-
be a root of
/(#)
=
0,
j-
being
a fraction in
its lowest
terms,
we have
n \
'*
I)
Multiply by
&*-
1
;
then all the terms will be
integral
except
the first which will be fractional
[for
a is
prime
to
THEORY OF
EQUATIONS.
569
b and therefore a
n
is also
prime
to
6],
and this is im-
possible.
Now,
from Art.
442, II.,
any equation
can be trans-
formed into another with
integral
coefficients and with
unity
for the coefficient of its first term
; hence,
from the
above,
we have
only
to find
integral
roots.
Now it is clear that if a be an
integral
root
off(x)
=
0,
so that # a is a factor
of/ (x),
a must be a factor of the
term which is
independent
of x. Thus if we
apply
the
test of Art. 88 to all the factors of
p
n
we shall discover all
the
integral
roots.
Ex. Find the commensurable roots of x*
-
27x
a
+
42x
+
8
=
0.
Here the commensurable
roots,
if
any,
are factors of 8. Hence
we have
only
to test whether
any
of the numbers
8, 4, 2,
1
are roots. It will be found that 4 and 2 are roots.
Having
found
two roots the
equation
can be
completely
solved
;
for we have
(x
-
2) (x
-
4) (*
a
+Qx
+
1)
=
0.
Hence the roots of the
equation
are
2,4,
-
3
2^/2.
EXAMPLES XLIV.
1. Solve
the
equation
x
4
+
'2x*
-
16x
8
-
22x
+ 7
=
0,
having given
that
2
+
,y/3
is one root.
2. Solve the
equation
3x
3
-
23x
!
+ 72#
-
70
=
0, haviug given
that
3
+
N/-
5 is one root.
3. One root of the
equation
3*
-
4*<
-
42
+
56*" + 27*
-
36
=
is
J2-J5,
find the
remaining
roots.
4. One root of the
equation
2a;
6
-3x
5
+
oa;
4
+
6x
3
-27a; +
81=0 is
^/2
+
\/-T.
Find the
remaining
roots.
5. Find the
biquadratic equation
with rational coefficients one root
of which is
^3
-
J5.
6. Find the
biquadratic equation
with rational coefficients one root
of which is
^2
+
N/-
3.
7. Shew that x
3
-
2z
2
-
2x
+
1
=
and x*
-
7x
2
+
1
=
have two roots
in common.
8. Solve the
equation
<r
-
4x
8
+
llx
2
-
14x
+
10=0 of which two
roots are of the form a
+
jS^/-
1 and a
+
2/3^/-
1.
570 THEORY OF
EQUATIONS*.
9. Find the condition that the roots of X
s
+px*
+
qx
+
r
=
may
be in
Harmonical
Progression.
10. Find the conditions that the roots of
x*+px
3
+
qx*+rx
+
s=Q
may
be in A. p.
11. Find the roots of the
equation
x
3
-
3x*
-
13x
+
15
=
0, having given
that the roots are in A. p.
13. Solve the
equation
x
4
+
2x*-21x*-22x
+
40=0, having given
that
the roots are in A. p.
13. Find the commensurable roots of
(i)
**- Tar* + 17x-
15=0,
(ii)
x*-x-13x
2
+
16x-48=0,
(iii)
3x
3
-26x
2
+34x-12=0.
1*. Solve the
equation
4x
s
-32j
2
-x
+
8=0,
having given
that the
sum of two roots is zero.
16. Solve the
equation
x
+
4x
3
-5x
s
-8x
+
6=0,
having given
that
the sum of two roots is zero.
16. Find the condition that the sum of two roots of the
equation
x
4
+j>x
3
+
gx
:1
+rx +
8
=
may
be
equal
to zero.
17. Solve the
equation
x
8 -
79x +
210=0, having given
that two of the
roots are connected
by
the relation a
=2/3
+
1.
18. Solve the
equation
3x*
-
32x* + 33x
+
108=0, having given
that
one root is the
square
of another.
19. Shew
that,
if the roots of the
equation
be in A.
p., they
will be obtained from
-
by giving
to r
the values
1, 3, 5,...
when n is
even,
and the values
2, 4, 6,...
when
n is odd.
ao. Find the condition that the four roots
a,
/3, 7,
5 of the
equation
z
4
+px
s
+
qx*+rx
+ $=Q
may
be connected
(i) by
the relation
0^=78,
and
(ii) by
the relation
0/9
+
75=0.
31. Shew
that,
if four of the roots of the
equation
be connected
by
the relation a +
/3
=
y
+
8,
then will 4<t//c
-
fc
1
-
8a*d
=
0.
33. If
a, ft, c,...
be the roots of the
equation
x'+fjx"-
1
+pyx*~*+
...
+pn =0,
prove
that
(1
-
a") (1
-
b*) (1
-
c)...
=
^
+
B
8
+
C-
3ABC,
where
^=Jm+Jm-s+-. B=P-i+P-4
+
-,
and C=
THEORY OF
EQUATIONS.
571
451. Derived functions. Let
f(x)
=
p<p
n
+
p&n-
1
+p.&
n
~
z
+ ......
+p
n ;
then,
if x
+
h be
put
for
x,
we have
f(x
+
h)
=
p (x
4
h)
n
+
Pl
(x
+
h)
n
~
l
+
p, (x
+
h)
n-*
+
. . .
+
p
n
.
If now
(x
+
K)
n
,
(x
+
h)
n
~
l
,
&c. be
expanded by
the Bi-
nomial
Theorem,
and the result
arranged according
to
powers
of
h,
we shall have
+
higher powers
of A.
This
expansion
is
usually
written in the form
/(*+*)=/(*)
[The
reader who is
acquainted
with the Differential
Calculus will see that the
expansion
of
f
(x
+
h)
in
powers
of h is an
example
of
Taylor's Theorem.]
It will be seen at once that
f (#)
is obtained
by
multi-
plying every
term of
f (x) by
the index of the
power
of x
it contains and then
diminishing
that index
by unity.
It will also be
easily
seen that
/" (x)
can be obtained
from
f (x)
in a similar
manner,
and so for
f" (x),
&c. in
succession. We shall however in what follows
only
be
concerned with
f(x).
Def. The function
f (x)
is called the first derived
function of
f(x),
the function
/"(#)
is called the second
derived function
of/ (x),
and so on.
Thus if
/ (x)
=
Po
x*
+p^+P&* +P&+p
4
,
f(x)
=
572 THEORY OF
EQUATIONS.
452. Theorem.
If(x)
be
any
rational and
integral
function of xandf
(x)
be its
first
derived
function,
then will
where
a^, a?, a,,
...... are the n
roots,
real or
imaginary, of
the
equation f(x)
=
0.
We know that
Hence
f(x
+
h)=p (x
a
1
The coefficient of h in the
expression
on the
right
is
by
Art. 260
equal
to
p
x
(sum
of all the
products
n I
together
of the n
quantities
xa^, xa,,
......
,
x
a^).
But
/(x
+
A)
=
/(x)
+
hf (x)
+
higher powers
of h.
Hence
f'(xi)=p
x
(sum
of all the
products
n 1
together
of the n
quantities
x
a^,
x
a,, .......
x a
n
).
Hence
x
Oj
x
a,
In the above the
quantities
c^, 03,
......
,
a
n
need not
be all different from one another
;
but if a
a
occur r
times,
and
a,
occur 8
times, &c.,
we shall have
x cr
a
x
a,
453.
Equal
Roots. We have seen in the
preceding
Article that if
c^, a,,
......
, a*
be the n roots of the
equa-
tion
/ (x)
=
0,
so that
f(x)
=
p (x aj) (x Oj)
......
(x
a
n
)
;
then will
f(x) =p
x
(sum
of all the
products
n1
together
of the n
quantities
x
Oi,
x
a, ....... ,x
a^.
Now,
if
any
root,
for
example
Oj,
is not
repeated,
so
that the factor x
a^
occurs
only
once in
f (x),
then the
factor x
ttj
will be left out of one of the terms of
f(x)
but will occur in all the
others;
whence it follows that
f'(x)
is not divisible
by
x
c^.
Thus a root
off(x)
=
which is not
repeated
is not a root
off(x)
=
0.
THEORY OF
EQUATIONS.
573
If, however,
r roots of the
equation /(x)
=
are
equal
to
cti,
the factor x
a^
will occur r times in
f
(x),
and
therefore x
Oj
will occur at least r 1 times in
every
term
of/'
(x),
for
every
term
of/' (x)
is formed from
/ (x)
by omitting
one of its factors. Hence a root
of f(x)
=
which is
repeated
r times is also a root
of /' (x)
=
re-
peated
r 1 times.
We can therefore find whether the
equation /(x}
=
has
any equal
roots,
by finding
the H. c. F. of
/
(x)
and
/' (a?)
;
and if
/
(x)
be divided
by
this H. c. F. the
quotient
when
equated
to zero will be an
equation
whose roots are
the different roots of
/(x)
=
0,
but with each root occur-
ring only
once.
Ex. 1. Find the
equal
roots of the
equation
x*
-
5x
3
-
9x
2
+
Six
-
108
=
0.
Here
/(x)=*
4
-5x
3
-9x
s
+81x-
108,
/' (x)
=
4x
-
15*
2
-
18x
+ 81.
The H.o.r. of
f(x) and/' (x)
will be found to be x
3
-
60;
+
9,
that
is(x-3)
a
.
Since
(x
-
3)
2
is a factor of
/' (x),
(x
-
3)
3
will be a factor of
/ (x),
and it will be found that
/ (x)
=
(x- 3)
3
(x
+
4).
Thus the roots of the
given equation
are
3, 3, 3,
-
4.
Ex. 2. Shew that in
any
cubic
equation
a
multiple
root must be
commensurable.
This follows from Art. 445 and
446,
and from the fact that a
cubic
equation
can
only
have three roots.
Ex. 3. Solve the
equation
x
s
-
ISx
3
+
10x
2
+
60x
-
72
=
by testing
for
equal
roots.
Here
/(x)=x
5
-15x
s
+10x
a
+
60x-72;
/' (x)
=
5x*
-
45z
2
+
20a;
+
60.
It will be found that the H.C.F. of
f(x) and/' (x)
is
If now we divide
f(x)
by
x
3
-x
a
-8x
+
12 the
quotient
will be
x
z
+ x
-
6,
and the roots of x
a
+x
-
6=0 are 2 and
-
3.
Thus the
given equation
has
only
two different
roots, namely
2
and
-
3
;
and it will be found that
/(x)
=
(x
-
2)
3
(x
+
3)
2
. Thus the
roots
of/(x)
=
arc 2,
2, 2, -3,
-3.
574 THEORY OF
EQUATIONS.
454.
Continuity
of
any
rational and
integral
function of
.
.
Let
p^c
n
+
p^x
n
~
l
+
p.M*-*
+
..... .
+
p
n
be
any
rational
and
integral
function of x
arranged according
to descend-
ing powers
of x.
Then each term will be finite
provided
x is finite
;
and
therefore,
as the number of the terms is
finite,
the sum of
them all will be finite for
any
finite value of x.
It can be
easily proved
that the first
(or any
other
term)
can be made to exceed the sum of all the terms
which follow it
by giving
to a; a value
sufficiently great
;
and also that the last
(or any
other
term)
can be made to
exceed the sum of all the terms which
precede
it
by
giving
to a; a value
sufficiently
small.
For let k be the
greatest
of the
coefficients;
then
Po*
n
- Po** Pox(x-l) PQ
Pl
x*-i
+ ...
+pn
k
(*-*
+ ...
1)
fcx" k
(X
7"
Now
g
(x-1)
can be made as
great
as we
please by sufficiently
K
increasing
x.
We can
prove
in a similar manner that
pJ(pn-jX
+ ...
+PO**)
can
be made as
great
as we
please by sufficiently diminishing
x.
Now
suppose
that x is
changed
into x
+
h
;
then we
shall have
where the coefficients
f'(x), f"(x),
&c. of the different
powers
of h are finite
quantities.
Then
by
the
above,
the first term on the
right (or
if this
term vanishes for
any particular
value of
x,
then the first
term on the
right
which does not vanish for that
value)
will exceed the sum of all the terms which follow
it,
provided
h be taken small
enough.
But the first term
will itself become
indefinitely
small when h is
indefinitely
small. Therefore
f(x
+
fi) f(x)
can be made as small
as we
please by taking
h
sufficiently
small. This shews
that as x
changes
from
any
value a to another value b
THEORY OF
EQUATIONS.
575
f{x)
will
change gradually
and without
any interruption
from
f
(a)
to
f
(b),
so that
f(x)
must
pass
once at least
through every
value intermediate to
f(a)
and
/(&).
It must be noticed that it is not
proved
that
f
(a?)
always
increases or
always
diminishes from
f(a)
to
f
(b),
it
may
be sometimes
increasing
and sometimes diminish-
ing
as a; is
changed
from a to
6;
what has been
proved
is
that there is no sudden
change
in the value of
f(x).
455. Theorem.
If /()
and
f(jB)
have
contrary
signs
one root at least
of
the
equation /(a)
=
must lie be-
tween a and
fi.
For since
/(a;) changes continuously
from
/(a) to/(/8),
it must
pass
once at least
through any
value intermediate
to
f(a)
and
f($)
;
it therefore follows that for at least one
value of x intermediate to a and
ft
it must
pass through
the value
zero,
which is intermediate to
f(a)
and
/(/8)
since
/(a)
and
/(/9)
are of
contrary sign.
Thus the
equation f(x)
=
is satisfied
by
at least one value of x
which lies between a and
@.
For
example,
if
f(x)=:f-4x
+
2,
then
/(!)
=
-
1 and
/(2)=2.
Hence one root of the
equation
x
3
-
4x
+
2
=
lies between 1 and 2.
456. Theorem. An
equation of
an odd
degree
his
at least one real root.
Let the
equation
be
/"(#)=
0,
where
Then
/(+<)
is
positive, /(0)
=
|)2n+1
,
and
f( oo)is
negative.
Thus there must in all cases be one real
root,
which
is
positive
or
negative according
as
p.>n+i
is
negative
or
positive.
457. Theorem. An
equation of
even
degree,
the
coefficient of
whose
first
term is
unity
and whose last term
is
negative,
has at least two real roots which are
of
con-
trary signs.
576 THEORY OF
EQUATIONS.
Let a?"
+
p^-
1
+ 4-
pan
=
be the
equation, p
m
being negative.
Then
/(+
oo
)
is
positive, /(O) =p
m,
and
/(
oo
)
is
positive.
Hence,
as
p
M
is
negative,
there must be one real root
at least between -f
oo and
0,
and also one at least between
and oo
.
458. The
following
is a
very important example.
To
prove
that
if
a, b, c, f, g,
h be all real the roots
of
the
equation
(x-a)(x-b)(x-c)-f*(x-a)-g*(x-b)-h?(x-c)-
Zfgh
=
0,
will
always
be real.
We
may suppose
without loss of
generality
that a > b >e.
Write the
equation
in the form
(x
-
a) {(x -b)(x- c) -J*}
-
{g* (x-b)
+
A
(x
-
c)
+
2fgh}
=0.
By substituting
+
oo
, b, c,
-
oo
respectively
for x in
(*-6)(*-c)-/,
we see that the roots of the
equation (x -V)(x- c) -/*=
are
always
real
;
and if a and
/3
be these
roots,
where
o>/9,
then
a>6>c>/3.
Now substitute
+
co
, a, /3,
-
oo for x in the left-hand member of
the cubic
equation,
and we shall have
respectively
the
following
results
+
00,
-{gja-b
+
hja-c}*,
+
{gjb~p+hjc- p}*,
-
oo.
Hence there is one root of the cubic between +
oo and
a,
one root
between a and
/3,
and one root between
/3
and
-
oo .
If, however,
a=p
the above
proof
fails
;
but if
a=P,
then
(x -b) (x-
c) -/
2
,
must be a
perfect square,
whence it follows that
l=c
and/=0.
The cubic
equation
in this case becomes
(x-a)(x-b)*-(g*+h*)(x-b)
=
0,
the roots of which are at once seen to be all real.
If a be a root of the cubic
equation itself,
there will be another
real root less than
p.
Hence all the roots of the cubic must be
real,
for the
equation
cannot have one
imaginary
root.
The cubic
equation
considered above is of
great importance
in
Solid
Geometry,
and is called the
Discriminating
Cubic.
459. Theorem.
Iff(a)
and
/(/3)
are
of contrary
signs,
then an
odd number
of
roots
of f (x)
=
lie between
a and
ft
;
also
iff (a)
and
f(ft)
are
of
the same
sign,
then
THEORY OF
EQUATIONS.
577
no roots or an even number
of
roots
of f(x)
=
lie between
a and
ft.
Let
a, b, c, ,
k be all the roots of the
equation
f(x)
=
which lie between a and
ft ;
then
f(x)
=
(x a) (x b) (x c) (x
k)
(j>(x),
where
(/> (x)
is the
product
of
quadratic
factors
(correspond-
ing
to
pairs
of
imaginary roots)
which can never
change
sign,
and of real factors which do not
change sign
while x
lies between a and
ft.
Then
/(o)=(o-o)(a-6)(a-c) (a-*)0(a),
and
f(ft)
=
(ft-a)(ft-b)(ft-c) (ft-k)^(ft).
Now,
supposing
a >
ft
all the factors a
a,a b, ,
a k are
positive;
and all the factors
ft
a,
ft
b, ,
ft
k are
negative
;
also
<f) (a)
and
<f> (ft)
have the same
sign.
Therefore if
/"(a)
and
f
(ft)
have
contrary signs
there must
be an odd number of the roots
a, b, c, ,
k.
Also,
if
/(a)
and
/(ft)
have the same
sign
there must be no such
roots or an even number of them.
460. Rollers Theorem. A real root
of
the
equation
f(x)
lies between
every adjacent
two
of
the real roots
of
the
equation f(x)
=
0.
Let the real roots of
f(x)
=
0,
arranged
in
descending
order of
magnitude,
be
a, b, c,
. .
.,
k. Then
f(x)
=
(x-a)(x- b).
.
.(x
-
k) <f> (x),
where <>
(x)
is the
product
of real
quadratic
factors corre-
sponding
to
pairs
of
imaginary
roots and these
quadratic
expressions keep
their
signs unchanged
for all real values
of x.
Then
f(x
+
X)
=
(x
-
a
+
\) (x
-
6
+
\)...(x-
k
+
X)
x
{< (x)
+
\(f)'
(x)
+
higher powers
of
X]
, (/(*) f(x) f(x) f(x)6'(x)
=f(x)
+
\\^^-
+
^f
+...+
jL
*-i
+
.
/V
'
(x
a x-b x k
<f>(&)
[See
Art.
452.]
s. A. 37
578 THEORY OF
EQUATIONS.
Hence/' (.).
+ +....+
as- a xb x k
e terms on the
right except
t
,
and that term is
(x b)(x c).
.
.(x k)
</> (x).
Now all the terms on the
right except
the first contain
the factor x
a,
and that term is
Hence
/(a)
=
(a
-
6) (a
-
c).
.
.(a
-
k) <f> (a).
Se
/' (6)
=
(b
-
a) (b
-
c). ..(&-&)
Now
0(a), </>(6), <(c),
&c. have all the same
sign.
Hence as a > b > c. .
.,
the
signs
of
/' (a), /' (b), f(c),
&c.
are
alternately positive
and
negative.
Hence there is at
least one root of
/' (x)
=
between a and
b,
one root
between b and
c,
&c.
461. Descartes' Rule of
Signs.
In
any equation
f(x)
=
the number
of
real
positive
roots cannot exceed the
number
of changes
in the
signs of
the
coefficients of
the
terms in
f(x),
and the number
of
real
negative
roots cannot
exceed the number
of changes
in the
signs of
the
coefficients
of f(-x).
We shall first shew that if
any polynomial
be multi-
plied by
a factor x
a,
where a is
positive,
there will be
at least one more
change
in the
product
than in the
original
polynomial
Suppose
that the
signs
of
any polynomial
succeed each
other in the order
+H
+ H
-----
1
,
in which there
are five
changes
of
sign.
Then
writing only
the
signs
which occur we shall have
Now we cannot write down the second
partial product
for we do not know that all the
possible
terms in the
THEORY OF
EQUATIONS.
579
polynomial
are
present
;
but whenever there is a
change
of
sign
in the first
partial product
it is clear that if there
is in the second row
any
term of the same
degree
in
x,
so that it would be
put
under this term which has the
changed sign,
it must arise from the
multiplication
of
the next
preceding
term so that the two terms would have
the same
sign.
Thus whenever there is a
change
of
sign
in the first
partial product
that
sign
will be retained in
the addition of the two lines of
partial products.
The
number of
changes
of
sign,
exclusive of the additional one
which must be added at the
end,
cannot therefore be
diminished.
Hence the
product
of
any polynomial by
the factor
x a will contain at least one more
change
of
sign
than
there are in the
original polynomial.
If then we
suppose
the
product
of all the factors
corresponding
to
negative
and
imaginary
roots to be first
formed,
one more
change
of
sign
at least is introduced
by
multiplying by
the factor
corresponding
to each
positive
root. Therefore the
equation f(x)
=
cannot have more
positive
roots than there are
changes
of
sign
in the
coefficients of the terms in
/(#).
The second
part
of the theorem follows at once from
the
first,
for the
positive
roots of
/( x)
=
are the
nega-
tive roots of
f(x)
=
0.
The above
proof may
be made clearer
by taking
as a definite
example
the
multiplication
of x
7
+
2x
e -
x*
+
4j;
3
+
3x-l
by
*-l.
The
signs
of the two lines of
partial products
will be
+ +
-
-5- +
-
In the third line the
only signs
written down are those under
the
changes
in the first
line,
which
changes
are all retained in the
final
product
Hence no matter what has occurred iu the intervals
the number of
changes
(exclusive
of the one at the
end)
cannot be
diminished.
462. Descartes' Rule of
Signs only gives
a
superior
limit to the number of real roots of an
equation,
but does
372
580 THEORY OF
EQUATIONS.
not determine the actual number of real roots. The
number of the real roots of
any equation
with numerical
coefficients can be found
by
means of Sturm's Theorem.
Before
considering
Sturm's Theorem we shall shew how
to find
algebraical
solutions of cubic and
biquadratic
(quartic) equations
in their most
general
forms. Abel
has
proved
that an
algebraical
solution,
that is a solution
by
radicals,
of a
general equation
of
higher degree
than
the fourth cannot be
found,
although particular
forms
of such
equations
can be
solved,
for
example any reciprocal
equation
of the fifth
degree
can
always
be
solved.
EXAMPLES XLV.
1. Solve the
following equations
each of which has
equal
roots :
(i)
x
a
-
-
12x
2
-
15x
-
4
=
0,
(ii)
!B
4
-6x
3
+
13
1
-24x +
36=0,
(iii)
16**-24x +
16x-3=0,
(iv)
2aT
2. Find the condition that the
equation
ax* +
3bx*
+ 3cz +
d
=
may
have two
equal
roots.
3. Shew
that,
if the
equation
ax*
+
3bx* +
3cx + d=Q have two
equal
roots, they
are each
equal
to
1 be -ad
4. Shew that the roots of the
equation
a b* c
8
fc
i i/
*
j
T >
i/
x-a *-o x-e x-k
are all real.
6. Shew that all the roots of the
equation
o* 6
s
c*
x-a
x-/3 x~7
are real.
THEORY OF
EQUATIONS.
581
6. If
a,, a.,,
a
s
, ..., a^
be in
descending
order of
magnitude,
and if
b be
positive, prove
that the roots of the
equation
a^ (x-aj... (x-oSB)
=
will all be
real,
and find their
positions.
7. Prove that if
a,
b, c,
d be
unequal positive quantities,
the roots of
the
equation
xxx
x-a x-b x-c
will all be real
;
and
that,
if the roots be
a, /3,
y,
S,
then will
a 62
_
(c-a)(c-p)(c-y)(c-
-
8. Form the
equation
whose roots are the values of
jpw
+
gw"
1
,
where
(o is a fifth root of
unity,
and shew that the
equation
is
X
s -
5pqx*
+
5p
2
g
s
-
P
5
~
9
8
=
-
9. If
a, /3, 7,
8 be the roots of the
equation
x
4
+
Ipx
3
+
6gx
a
+
4rx
+
s
=
form the
equations
whose roots are
(i) a/3
+
75, ay+pS,
ad
+
fiy.
(ii) (
a +
J8)(7
+
5), (a
+
y)(p
+
5), (a
10. If
a, /3, 7,
8 be the roots of the
equation
a;
4
+
4pz
8
+
6gz
3
+
4rx
+
=
form the
equation
whose roots are
(a
+
/3-7-5)
2
,
(a-0
+
7-5)
2
,
(a-^-
11. If
Oj,
a
2
,
a,
be the roots of
x
3
+p1
x* +
p*x +p3
=
0,
shew that 2a
3
a
582 THEORY OF
EQUATIONS.
CUBIC
EQUATIONS.
463. The most
general
form of a cubic
equation
is
We have however seen
[Art.
442,
in.]
that
by
in-
creasing
each root
by
=
,
the
equation
will take the
simpler
o
form of +
px
+
q
=
0.
We shall therefore
suppose
that the
equation
has
already
been reduced to this
simplified
form.
464. To solve the cubic
equation
a?
+
px
+
q
=
Q.
The solution is at once obtained
by comparing
the
equation
with
a?
-
Sabas
+
a
3
+.b
a
=
0,
i.e.
(x
+ a +
b) (a
+
ooa
+
&>
2
&) (as
+
o>
2
a +
toty
=
0,
where o> is an
imaginary
cube root of
unity [Art. 139].
Thus the roots
required
are
a
b,
wa (i>
2
b,
o>
2
a
a>b,
where a and b have to be determined from the
equations
p
=
Sab,
q
=
a
3
+
b
3
.
Whence a
:!
and b
s
are
given by
465. The
foregoing
solution is a
slight
modification of
that called Garden's solution. It is a
complete algebraical
solution of the
equation
and the values found for x would
satisfy
the
given equation identically.
If, however,
nu-
merical values be
given
to
p
and
q,
the numerical values
2
jjS
of a and 6 cannot be found when
^-
-f-
= is
negative,
for we
THEORY OF
EQUATIONS.
583
cannot reduce an
expression
of the form
(3
+
5\/
1)^,
for
example,
to the form a
+
/3V
1. Thus when
p
and
q
are
numerical
quantities
such that *r
+
~= is
negative,
Garden's
4) *
solution
altogether
fails to
give
a numerical result.
This case is called the 'irreducible
case/
and
we shall see further on
[Art.
467,
Ex.
3]
that when
7?
2
O*
*
+
-^
is
negative
all the roots of the cubic are real.
T? .Z i
It should also be noted that in
any
case the
approxi-
mate values of the real roots of a cubic can be obtained
much more
easily by
Homer's
general process [Art. 475]
than
by
Garden's solution.
Ex. Solve the cubic
equation
x
3
+
4x
-
5
=
0.
Comparing
with x
3 -
Sabx + a
3
+
b
3
=
0,
we have -3a&
=
4 and a
3
+
&
3
=-5,
whence a and 6 are
given by
The
approximate
values of a and 6 can therefore be
found,
and then
the roots are
a
b,
ua
uPb,
w
2
a ub.
In this
example
the solution can be obtained in a
very simple
manner.
For,
using
the test
given
in Art.
449 for commensurable
roots,
we are led to find that 1 is a commensurable
root,
and
writing
the
equation
in the form
(x
-
1) (x
2
+
x
+
5)
=
0,
the roots are at once
seen to be
i,
-(liV^)-
BIQUADRATIC EQUATIONS.
466. Several methods of solution of a
biquadratic
equation
have been
given.
In all of them the solution is
shewn to follow from the solution of a cubic
equation.
The
simplest
method of solution is that due to Ferrari.
584 THEORY OF
EQUATIONS.
To solve the
equation
of
1
+
pa?
+
qa?
+
rx
+
s
=
0.
Ferrari's Solution. Add
(ax+@y
to both sides of
the
equation
;
then
a*
-f
pa*
+
(q
+
a
2
)
a?
+ (r +
2a)
ae
+
s +
^
=
(ax
+
/9)
2
.
Now the left-hand member will be a
perfect square,
namely
(a?
+^x
+
\\
,
provided
\
^
/
Eliminating
a
and
yS,
we have a cubic
equation
to
determine
X,
namely
One root of this cubic
equation
is
always
real,
and if
this root be found the values of a and
yS
are determined.
We then have
whence
ac^+x +
\
(ax
+
/3)
=
0,
where
a, /3
and X are known. Thus the
biquadratic
equation
can be
completely
solved.
Ez. Solve the
equation
s*
+
6x
+
U*
+
22*
+
6
=
0.
Add
(ax
+
p)
2
to both
sides;
then
s* 4 6Z
3
+
(14
+
a
1
)
*
+
(22
+
2a/3)
a;
+
5 +
p
=
(ax
--
/S)
3
.
The left hand member is the
square
of x*
+
3z
+
X
provided
9 +
2A
=
14+a
8
,
6X=22-f2a^3
and X
s
=5+/3
2
.
Whence
(X
s
-
5) (2\
-
5)
-
(3\
-
II)
2
=
;
X*-7X
a
The real root of the cubic is 3.
THEORY OF
EQUATIONS.
585
Then, taking
X=3,
we have o
j
=l,
2o/3=
-4,
/3
a
=4.
Hence
(x*
+
3x
+
3)
=
(x
-
2),
whence we obtain the roots -2
N/3,
-l2
>/-i
>
STURM'S THEOREM.
467. Sturm's Theorem. Let
f(x)
=
be an
equa-
tion cleared
of equal
roots,
and let
/i (x)
be the first derived
function of
/ (x).
Let the
process
of
finding
the
highest
common factor of
/(x)
and
/j (x)
be
performed
with the
modification that the
sign of every
remainder is
changed
before
usin^
it as a
divisor,
and let the
operation
be con-
tinued until a remainder is arrived at which does not
contain x
(this
will
always happen
since
f(x)
=
has
no
equal
roots and therefore
f(x)
and
yj
(x)
have no
common measure in
x),
and
change
also the
sign
of this
last remainder.
Let
fi(x), /,(#),..., f
m
(x)
be the series of modified
remainders so
obtained,
of which the
last,
f
m
(x),
does not
contain x.
Then the number
of
real roots
of
the
equation f(x)
=
between a and
/3,
[#
>
a]
is
equal
to the excess
of
the num-
ber
of changes of sign
in the series
f(x), f
(x) ,
f
2
(x),
,..,
f
m
(x)
ivhen x
=
a.
over the number
of changes of sign
when
x
=
j$.
For,
let
(ft, q
a ,...,
q
m-i
be the successive
quotients;
then we have the series of identities
/m-a (#)
=
?m-i/i_i (#) ~fm (#)
Now
(i)
it is clear that no two consecutive functions
can vanish for the same value of
x,
for in that case all the
succeeding
functions,
including f
m
(x),
would vanish for
that value of
(x)
; and,
(ii)
it is also clear that when
any
586 THEORY OF
EQUATIONS.
one of the functions
except f
(x) vanishes,
the two
adjacent
functions will have
contrary signs.
It follows from
(i)
and
(ii)
that so
long
as the
increasing
value of x does not make
f(x)
itself
vanish,
that is unless
we
pass through
a real root
of
the
equation f(x)
=
0,
there
can be no alteration in the number
of changes of sign
in the
series
of
Sturm's
functions
;
for no function will
change
sign
unless it
passes through
a zero
value,
and when this
is the case for
any
function,
since the two
adjacent
func-
tions have
opposite signs,
there must be one and
only
one
change
in the
group
of three.
Next
suppose
that a is a real root of the
equation
/(tf)
=
0. Then
/(a
+
X) =/(a)
+
X/'(a)
+&c.
;
and as
/(a)
0,
the
sign
of the series on the
right
will,
if \ be
very
small,
be the same as the
sign
of
+
X/'(a).
Hence,
however small X
may
be,
the
sign
of
f(a X)
must be
opposite
to that of
f(a),
and the
sign
of
f
(a
+
X)
must
be the same as the
sign
of
f(a).
Thus 05 x increases
through
a real root
of
the
equation
/(#)
=
0,
the series
of
Sturm's
functions
will lose one
change
of sign.
Since we have
proved
that as x increases the series of
Sturm's functions never lose or
gain
a
change
of
sign
except
when x
passes through
a real root of the
equation
f(x)
=
0,
in which case one
change
of
sign
is
always
lost,
it
follows that the excess of the number of
changes
of
sign
when x
=
a. over the number of
changes
when 'x
=
(S
must
be
equal
to the number of real roots of the
equation
which
lie between a and
ft.
To find the total number of real roots of an
equation
we must substitute oo and
+
oo in Sturm's functions
;
then the excess of the number of
changes
of
sign
in the
series in the former case over that in the latter will
give
the whole number of real roots.
Ex. 1. Find the number of the real roots of the
equation
** +
4x-4a:-13=0.
Here
/(x)=x<
+
4x-4x-13,
THEORY OF
EQUATIONS.
587
N.B. We
may clearly multiply
or divide
by positive
numerical
quantities
as in the
ordinary process
for
finding
H. c. F. It will be
found that
/,(*)=
-19.
Substitute
-oo, 0,
+
in the above
functions,
and the series of
signs
will be
Thus there is one real root between
-
oo and
0,
and one real root
between and
+
oo .
Ex. 2. Find the number and the
position
of the real roots of the
equation
x
5
-
5x
+
1
=
0.
Here
/ (x)
=
x
5
-5x +
l,
It will be found that
/3 (*)=+255.
The
following
are the series of
signs corresponding
to the values
of x written in the same line
oo
, + -f-
-2,
-
+
-
+
-1, + +
0,
+
- -
+
1,
-
+ +
2,
+ + + +
Hence there is one real
negative
root between
-
2 and
-
1,
one
itive root between and 1 and another between 1 and
2,
the
remaining
two roots
being imaginary.
Ex. 3. Find the condition that all the roots of the
equation
may
be real.
The other functions will be found to be
588 THEORY OF
EQUATIONS.
The
signs
for
-
oo and + oo are
-, +,
+2p, -(2
and
+, +,
-2p, -(27
In order that the roots
may
be all
real,
it is
necessary
and
sufficient that there shall be three
changes
of
sign
in the first line
and none in the
second,
the conditions for which are that
p
and
27g*
+
4p
s
must both be
negative,
the second of which
implies
the
first.
468.
Although
Sturm's Theorem
completely
solves
the
problem
of
determining
the number and the
position
of the real roots of an
equation,
it is often a
very
laborious
process.
In some cases the
position
of the real roots can
be determined without
difficulty by
actual
substitution;
and sometimes the
necessity
for
using
Sturm's Theorem
can be obviated
by
some
special
device.
Ex. 1. Find the number and
position
of the real roots of the
equation
=
0.
Substitute in
f(x)
the values
1, 2, 3,
4, 5,
6 in
succession,
and
the
signs
will
be+, +, -, -, -,
+.
Hence there is one root
(at
least)
between 2 and
3,
and one
(at least)
between 5 and 6
;
but
by
Descartes' Rule of
Signs
there cannot be more than two
positive
roots.
Hence there are two
positive
roots which lie between 2 and 3 and
between 5 and 6
respectively.
We can find in a similar manner that there are two
negative
roots
which lie between
-
1 and
-
2 and between
-
6 and
-
7
respectively.
Ex. 2. Find the number and
position
of the real roots of the
equation
In this case we should
easily
find the two
negative
roots which
lie between and -1 and between -4 and -5
respectively.
The
positive
roots
would, however, probably escape
notice
(unless
Sturm's
Theorem were
used)
as
they
both lie between 2 and 3
;
it will in fact
be found
that/(2)
is
+,/(2i)
is
-, and/(3)
is
+.
Ex. 3. Find in
any
manner the number and
position
of the real roots
of the
equation
x
-
Sa
8
-
7;r> +
&r + 20
=
0.
By
Descartes' Rule of
Signs
we see
by inspection
that there
cannot be more than two
positive
roots and there cannot be more
than two
negative
roots.
Now
/(I)
is
-f
,
/(2)
is
-
;
thus one real root lies between 1 and
2. Since
/ (z>
)
is +
,
there must be another
positive
root which is
easily
found to lie between 5 and 6.
THEORY OF
EQUATIONS.
589
Change
* into
-
x,
then the
negative
roots of the
given equation
are
positive
roots of
of + 5af
-
7ar
2
-
8x
+
20
=
0.
Now
f(x)
must
clearly
be
positive
for all
positive
values between
and 1
;
and if x >
1,
which is
always positive
since
4x6x20-15
2
>0.
Hence there can be no real
negative
roots.
469.
Equation
of Differences. If
f(x)
=
be
any
equation
of which a and
(3
are
any
two
roots,
and if
y
=
a
fi,
we shall obtain an
equation
in
y
whose roots
are the differences of the roots of the
given equation by
eliminating
/3
from the
equations f(y
+
/3)
=
andy(/3)
=
0.
Since a
/3
and
/3
a will both be roots of this
equation,
it follows that the
equation
in
y
will
only
contain even
powers
of
y
;
and it is
easily
seen that the
equation
in
y
2
has as
many positive
roots as there are
pairs
of real roots
of the
equation f(x)
=
0,
and that the
equation
in
y
n
-
has
all its roots real and
positive
if the roots of the
equation
f(x)
=
are all real.
In the case of the cubic
a?+px
+
q
=
Q the
equation
in
y
2
can be found more
easily
as under.
To find the
equation
whose roots are the differences of the roots
of the cabic x
3
+
px
+
q
=
0.
Let
j/
2
=(a-/3)
2
,
where
a,
/3, y
are the roots of the cubic.
Then
2
=a+
/3)
2
-4aS=
!
+
^
;
Also
Hence
7=3?/(!/
a
+3>);
.-.
27?
s
+
3pq (y* +.p)
2
+
q (y +p)*
=
0,
which is the
equation required.
By
Descartes' Bule of
Signs
the above cubic in
y"
cannot have
three
positive
roots unless
p
and
4p
3
+
27g
2
are both
negative. Also,
590
THEORY OF
EQUATIONS.
if
4j)*
+
27g
2
be
negative
the
equation
in
y-
will have three
positive
roots which are
separated by +00, -2p, -p,
0.
Thus a
necessary
and sufficient condition that the cubic
may
have all its roots real is that
4p*
+
27<?
2
may
be
negative.
470. We shall conclude
by shewing
how to find the
approximate
values of the real roots of
any equation.
This can be done in various
ways;
we
shall, however,
only
give
Horner's method. We must first
give
the
explan-
ations of the
separate processes
which are
employed.
471.
Synthetic
Division.
Suppose
that when
f(x)
=
a^
+
(hx
+
ajc"-*
+
. . .
+
a^
is divided
by
X \ the
quotient
is
Q
=
fto*"-
1
+
hx*-*
+
fcoz"-"
+
. . . + &_
and that the remainder is
R,
where R does not contain x.
Then
/(a?)
=
Q
x
(as
-
X)
+ R.
But
Q
x
(x X)
+
R is at once seen to be
b<r
n
+
(6
a
-
X&
)
a;"-
1
+
(6,
-
X^)
x
n
~
a
+ ...
+
(&n-i
-
^n-a)
X
+
R- \b
n^.
Equating
coefficients of the different
powers
of x
mf(x)
and in the
expression
last
written,
we have
R
\6n_j
=
a
n
.
From the above relations it will be seen that the values
of b
, &!,
6
2,
&c. can be obtained at once
by
the
process
indicated below :
0* Oj O, O,
......
a_! On
6
&j 6,
b
s &_!
R
First &
=
Ooj
multiply
6
8
by
\ and add to
a,,
the sum is
THEORY OF
EQUATIONS.
591
l)i\
multiply
&!
by
X and add to a
2 ,
the sum is 6
a
;
proceed
in this
way
to the end.
Ex. Find the
quotient
and the remainder when
x
5
-6x*
+
2x
3
+
15x^ + 1
is divided
by
x
-
2.
l_6
+
2
+
15 + 0+
7
2-8-12
+ 6 +
l'J
1-4-6+
'
Thus the
required quotient
is
the remainder
being
19.
The above
process
is called the method of
Synthetic
Division. The method can
easily
be extended to the
case when the divisor is a multinomial
expression,
but
this extension is not needed for our
present purposes.
472. The actual values of b
,
6
1(
b
2
,
&c. in terms of a
,
a
lt
a
2
,
&c.
and X can be at once written down
; they
are
b =a
,
fc^
i.
J
=
a.
J
+
Xa
2
6
n
_
1
=
a*.!
+
Xa
B
_
and .
Thus
iS2
=
a^*-
1
+
(^
+
Xa
)
x
n
~
2
+ . . .
X A
From the above we can obtain the formula of Art. 439.
For,
if
o, 6, c,
... be the roots of the
equation
f(x)
=
;
then
=
...
'
x-a x-b
=
{ agx*-
1
+
(oj
+
aa
)
x
n
~
2
+
(a.,
+
oo!
+
o
2
a
)
x*~
3
+ . . .
}
1
+
(0!
+
ba
)
x-
=7ia a;
n
-
1
+
(naj
+
o
Sa)
+
(na2
+
But
/' (x)
=na x*~
1
+
(-!)
o
l
ar"-
2
+
(n
-
2) a^""
3
+ ...
592 THEORY OJF
EQUATIONS.
Equating
the coefficients of like
powers
in the two
ezpansions,
we have
na =na
,
(n
-
2) a,
=
na
s
+
a
a
2o
+
a 2a
s
,
Whence the
required
result follows at once.
473. We have
already
seen
[Art.
442,
in.]
that in
order to diminish each of the roots of the
equation f(x)
=
by
X,
we have
only
to substitute
y
+
X for x in
/(#).
Let the
equation
whose roots are those of
a^K
n
+
diX*-
1
+
Ojaf"
8
+
. .. +
a
n
=
0,
each diminished
by
X,
be
b
y
n
+
b
t
y
n
~
l
+
bjy
n~*
+
. . .
+
6
n
=
0.
Then,
since
y
=
x
\,
the last
equation
is
equivalent
to
b
(x -\)
n
+
b
1
(x- X)
11
-
1
+
... +
&_!
(x
-
X)
+ b
n
=
0.
The
equation
last written must be identical
vfithf(x)
=
0.
Hence we have
identically
f(x)
=
b
(x X)
n
+
6j
(x X)"""
1
+
. . . 4- b
n i
(x X)
+
b
n
.
From the form of the
right-hand
member of the above
identity,
it follows that if we divide
f(x) by
x
\,
and
then divide the
quotient by (# X),
and so
on,
the suc-
cessive remainders will be the
quantities
6
n,
6
n
_
1> ..., &x,
6
9
.
Ex. 1. Find the
equation
whose roots are those of
each diminished
by
2.
Using
the method of Art. 469 to
perform
the successive
divisions,
the whole
operation
is indicated
below,
the successive remainders
being printed
in black
type.
1-2+ 0+
3-6
2006
100
3+1
248
1 2 4
11
2 8
1 4
12
2
1 G
THEORY OF
EQUATIONS.
593
The first division
gives
the
quotient
a?
+
3 with remainder 1
;
the
second division
gives
the
quotient
x
2
+
2z
+4 with remainder
11;
the
third division
gives
the
quotient
x
+
4 with remainder
12,
and the
last division
gives
the
quotient 1,
and remainder 6.
Ex. 2. Find the
equation
whose roots are those of
z
s
-x
2
-x+4=0,
each increased
by
3.
The divisor is here a: +
3,
and the work is as under.
1_ i_
1+
4 i_ i_
1+4
-
3
+
12-33
i
1_ 4
+ 11-29
1-
7|
+ 32
1-10
1- 7 + 32
-
3
1-10
Thus the transformed
equation
is
We shall in future write the
operation
as on the
right,
the multi-
plication
and addition
being performed mentally,
and the result
only
being
written down.
474. In order to
multiply
all the roots of the
equation
a,<F
n
+
a^x"-
1
+
o^""-
2
+
. . .
+
On
=
by
ten,
we must substitute
^r
for x in its left-hand
member. If we then
multiply throughout by
10
n
,
the
transformed
equation
will be
"
1
+
lO'a"-
2
+
... +
10
n
a
n
=
0.
Thus in an
equation
with numerical coefficients the roots will be
multiplied by
10
by affixing
one
nought
to the coefficient of x
n
~
l
,
two
noughts
to the coefficient of **~
2
,
and so on.
For
example,
the
equation
whose roots are those of
each
multiplied by ten,
is
x*
-
20s
3
+ 5000* + 80000
=
0.
475. Homer's method of
approximating
to the
real roots of
equations
with numerical coefficients.
Having
found
(by
trial or
by
Sturm's
Theorem)
two
consecutive
integers
between which a real
positive
root of
s. A. 38
594 THEORY OF
EQUATIONS.
the
given equation
must
lie,
the first
step
is to diminish
all the roots
of
the
given equation by
the smaller
of
those
integers.
Then,
by supposition,
the transformed
equation
will have a root between and 1. We now
multiply
all the roots
of
the last
equation by
10
by
the
process
of
Art.
472,
so that this new
equation
has
by supposition
a
root between and
10;
now find
by
trial between what two
integers
less than 10 the root
lies,
and diminish the roots
of the
equation by
the smaller of these
integers.
Then
again multiply
the roots
by
10,
and continue the
process
until the
required degree
of
accuracy
is attained.
After the roots of the
given equation
have been di-
minished
by
the
integral part
of the
required
root,
the
roots are
multiplied by
10 in order to avoid decimals in
the
work,
the next
integral
root found must therefore
originally
have been so
many
tenths. After
again
multi-
plying
the roots
by
ten,
the next
integral
root found
corresponds
to hundredth^ in the
original equation;
and
so on.
By
the above
process
it is clear that we are con-
tinually approximating
to the root
sought;
care
must,
however,
be taken that we do not
pass beyond
the
root,
which would be shewn
by
the
change
in
sign
of the
constant term.
The
negative
roots can be found
approximately
in a
similar manner after
changing
a; into x.
Ex. 1. Find to two
places
of
decimals,
the
positive
root of the
equation
There can
only
be one
positive root,
and
by
trial this must lie
between 2 and 3. First diminish the roots
by 2,
and the transformed
equation
will be found to be x
8
+
6x
s
+ 9x-2=0.
Multiply
the
roots
by
10 and we have the
equation
x
8
+ 60**
+ 900x
-
2000=0,
which will be found to have a root between 1 and 2. Diminish the
roots of this last
equation by 1,
and the transformed
equation
will
be x
+ 63x
s
+1023x- 1039=0.
Multiply
the roots of this
equation
by
10,
and the
resulting equation
will be found to have a root
between 9 and 10. Diminish the roots of the last
equation by
9,
wid the
resulting equation
is X
s
+ 657x
1
+113883x- 66541
=
0,
which
could be used to obtain a more accurate result.
THEORY OF
EQUATIONS.
595
The work is written as
under,
lines
being
drawn to indicate the
completion
of each
stage
of the
process.
0-3
-4
(
2-19...
2 +1
|
-2000
596
THEORY OF
EQUATIONS.
at each successive
stage,
as in the
ordinary
method of contracted
division.
476.
Imaginary
roots. The numerical values of the
imaginary
roots of an
equation
can
theoretically
be obtained in the
following
manner,
but the work would
except
in
very simple
cases be
very
laborious.
Ex. Find the numerical values of the
imaginary
roots of the
equation
x
4-
3x- 1=0.
Put a +
ifi
for x in
/(x),
and
equate
the real and
imaginary
ex-
pressions separately
to
zero;
then we shall have
a
3
4-3a-l-3oj3
2
=0 and 3a
3
/3-/3
3
4-3)3=0.
Rejecting
the factor
/9=0,
which
corresponds
to a real root of the
given equation,
we have
by eliminating /3
the
equation
Now a must be a real root of the
equation
last
written,
and this
real root will be found to be
-
-16109
Then
/3*=3 (a
2
4-l),
whence
/3
is found to be 1-75438.... Thus
the
required
roots are
-
-16109... 1-75438...
J~^l.
EXAMPLES XLVI.
1. Solve the
following equations
:
(i)
x
3
-
12x4-65=0.
(ii)
x
s
-9x+28=0.
(iii)
x
3
-48x- 520=0.
(iv)
x
3 -
21x
-
344
=
0.
(v)
x
3
-2x
+ 5=0.
(vi)
x
3
-6*-ll=0.
2.
Solve the
following equations
:
(i)
x*4-2x4-14x+15=0.
(ii)
x*-12x-5=0.
(iii)
x*-12x
a
+ 24x4-
140
=
0.
(iv)
4x*+4x-7x
3
-4x-12
=
0.
3.
Apply
Sturm's Theorem to find the number and
position
of the
real roots of the
following equations
:
(i)
x-3x
+ 6=0.
(ii)
x
3
-&-
(iii)
2X
4
-*
9
-
10x
+ 3=0.
(iv)
**-14z
2
4-16x4-9=0.
(v)
x
4
-7x
a
4-3;c -20=0.
4. Find all Sturm's functions for the
equation
x
s
4-3px
a
4-3gx4-r=0,
and hence shew
that,
if
p*<.q,
there must be two
imaginary
roots.
6. Prove that the roots of x
3
4-p**4-r
=
are all real if either
p
or r
be
negative,
and
-
4p*r
be
greater
than 27r
2
.
THEORY OF
EQUATIONS.
597
6. The coefficients of the
algebraical equation
/(z)
=
are all
integers.
Shew
that,
if
/ (0)
and
/ (1)
are both odd
numbers,
the
equa-
tion can have no
integral
roots.
7. Shew that one root of the
equation
x
3
-
2x
-
5 =0 is 2-09455148.
8. Find the real
positive
roots of the
following equations,
each to
4
places
of decimals :
(i)
x
s
-7x + 7
=
0.
-(iv)
a^
+
x
3
-
4x2-16=0.
(ii)
x-8x-40=0.
(v)
x
4
-
14x2
+
16*
+ 9
=
0.
(iii)
x-6x
2
+ 9x-3
=
0.
(vi)
x-2=0.
9. Find the number and
position
of the real roots of
(i)
x
4
+
2x
3
-23x
j
-24x
+
144
=
0,
and
(ii)
ar
1O. Prove that the
equation
x
6
-
Is? + 15x
2
+
3x
-
4 =0
cannot have more than 4 real roots
;
prove
also that these roots must lie
between 1 and
-1.
1 1. Solve the
equation
2x
-
7x* +6X
3
-
llx
2
+
4x
+
6
=
0,
having given
that the real roots are commensurable.
12. Find the
equation
whose roots are the
square
of the roots of
x
3
-3px*-
3
(l-^)x
+
l
=
0,
hence shew that the
given equation
has three real roots for all real
values of
p.
13. Prove that the
equation
x
3
-
3px*
-
3
(1 -p)
x
+
1 =0
has three real roots for all real values of
p.
Prove also
that,
if these roots be
a,
/3, y
then
/3(l- 7)=7 (l-a)=a(l- 18)
=
l,
or
/3(l-a)
=
7 (l-/3)
=
a(l-y)
=
l.
14. Shew that the
equation
whose roots are the sum of
pairs
of
roots of the
quintic
x
8
+px
+
q
=
is
x
10 -
Spz
6
-
llgx
5 -
4p
2
x
2
+
4pgx
-
q
2
=
0.
15. Prove that the
equation
x
4
+
4CUC
8
+
6a
s
x
2
+
4ax
+
1
=
has no real roots unless
l>a>J,
and that the
equation
has two real roots if a
3
is between these limits.
598 THEORY OF
EQUATIONS.
16. Prove
that,
if a be the root of the
equation
x*
+
ax*
-
Gar
8
-
ax + 1
=
0,
1
+a
BO also is = .
l-o
Prove also that the other two roots are
1
, a-1
and
;
.
a
a-t-1
17. Prove
that,
if
a,
/3, y,
... be the roots of
x- +
Pl
x*-* +
p^~ +...+!>= 0,
then Sa
8
07
=
-
p^pz
+
3p^p4
-
5p6
.
18. Shew
that,
if a
l
,
a
2)
...be the roots of
x
8
-
Spx
3
+
5p
s
x
-
q
=
0,
then 2ff
4
^
3
a
2
a
+
MISCELLANEOUS EXAMPLES.
1. Find the factors of the
expressions
:
(i)
a
2
(b
-
c) (c
+ a
6) (a
+ b
c)
+ 6
2
(c a) (a
+ b
-
c) (b
+ c
a)
+ c
2
(a
-
b) (b
+ c
-
a) (c
+ a
-
b).
(ii)
abed
(a
2
+ 6
2
+ o
2
+ rf
2
)
-
6W
-
c
2
<2V
-
dW
-
aW.
(iii)
2
(a
8
+ b
9
+ c
3
)
+ a?b + a?c + We + b*a + c*a + c
2
6
-
3a6c.
(iv)
a
4
(6
+
c)
+ 6
4
(c
+
a)
+ c
4
(a
+
6)
+ a
3
(6
+
c)
2
+ 6
s
(c
+
a)
2
+ c
8
(a
+
6)
2
+ 2a6c
(6c
+ ca +
ab).
2. If a
2
-
a
2
=
6
2
-
/8"
=
c
2
-
y
2
,
prove
that
by c(3
ca
-
ay
aft
-
ba
-
H
--
;
-
^
+ u.
a a 6
p
c
-
y
3. Shew
that,
if
yz
+ zx +
xy
=
a
2
,
then
1 1 1
___
__ i
_
^
___
7/3
(a
2
+ a;
2
)
20;
(a
8
+
7/
2
)
a;?/
(a
2
+ z
2
)
___
~
xyz V{(
2
+
^) (a
2
+
2/
2
) (a
2
+
#)}
'
4.
Shew
that,
if
yz
+ zx +
xy
=
0,
then
(y
+
z)
a
(*
+
*)
3
(*
+
2/)
a
+ 2a
^/
222
=
x*
(y
+
a)
8
+
t/
4
(
+
a;)
8
+ z
4
(a;
+
y)*.
5. Prove
that,
if
/(#), any
rational
integral
function of
x,
be divided
by (cc
-
a) (cc
-
6),
the remainder will be
J
:: .
a 6 a
-
6
6. Prove that 2
{(6
-
c)
4
+
(c
-
a)
4
+
(a
-
b)
4
}
is a
perfect
square.
7.
Find the
square
root of
3
{(j
_
c
)
+
(
c
_
a
)
+
(
a
-
6)
8
-
2
(a
2
+ 6
2
+ c*
-
be
-
ca
-
ab)*\.
8. Shew that
(b
-c)
7
+
(c
-
a)
7
+
(a
-
ft)
7
=|(6-
c) (c-a)(a-6)
2
(6 -c)
4
.
600
MISCELLANEOUS EXAMPLES.
9. Shew
that,
if
a, b,
c be
positive quantities
and not
all
equal
to one
another,
be
(a-b)(a c)
+
ca(b c) (b a)
+ ab
(c
-
a) (c
-
b)
will be
positive.
10. Shew
that,
if a + b + c
=
0,
then will
a(a-&)
8
(a-c)
2
+&(&-c)
2
(6-a)
a
+
c(c-a)
3
(c-&)-
+ 27 abc
(bc
+ ca +
ab)
=
0.
11. Shew
that,
if a+b + c + d
=
Q,
(i) (V)
1
=
22a +
Sabcd,
(ii)
**
=
^
.
^
.
Shew also that
Sa
2""1
"
1
is divisible
by
2a
s
,
if n be a
positive
integer.
12. Shew
that,
ita + b + c + d
=
Q
=
a* + b* + <? +
d*,
then
8 8 8
o
13. Shew
that,
ifa + J + c + d+e
+/=
0,
and a
s
+ &
8
+ c
8
+ </
s
+e+/'=0,
then will
(a
+
b)(b
+
c)(c
-f
a)
4-
(rf+ e)(e +/)(/+ <^)
=
0,
and
= 2a
7
=
=
2a
8
x
-
2a.
7 o 2i
14. Shew
that,
if one of the three
quantities
ax +
by
+
cz, by
+ ex +
az,
cz +
ay
+ bx
vanishes,
the sum of the cubes of the other two will be
equal
to
(a*
+ b' + c
1
-
15. Prove
that,
if
w
=
x +
y
+
z+a(x
+
y 2z),
then 27a*
(a?
+
y*
+ z
-
Sxyz)
=
u*
+
V
s
+ ^
-
MISCELLANEOUS EXAMPLES. 601
16. Shew
that,
if
X ax +
by
+
cz,
A
=
ax +
cy
+
bz,
Y=cx +
ay
+
bz,
B=bx +
ay
+
cz,
Z
=
bx +
cy
+
az,
C
=
ex +
by
+
az,
then will
=
(Z-A)(Z-B}(Z-C)
=
17. Prove
that,
if 6* < 4ac and b'
3
<
4aV,
then
(be
1
-
b'c) (ab'
-
a'b)
<
(ca
1
-
ac').
18.
Having given
that
b,
be 1
3
and abc + 2lmn -at
2
bm? en*
are all
positive, prove
that
a, c,
ac m* and ab n? are all
positive,
all the letters
denoting
real
quantities.
19. Shew
that,
if
a, b,
c be
unequal,
and
a
(b c)
x + b
(c a) y
+ c
(a b)
z
=
0,
a
(6 c) yz+
b
(c
-
a)
zx+ c
(a b) xy= 0,
tiien x
^
y
===
z,
/
20. Shew
that,
if
aa? -I-
by*
+ cz
2
+
2/y
+
20205
+
2&cy
=
(Ix
+
my
+
nz) (I'x
+
m'y
+
n'z),
wi
(mri
-
m'n) (gh of)
=
(rd' ril) (hf- bg)
=
(lm'-rm)(fg-ch).
21. Eliminate
(i)
x
y y,
z and
(ii)
a, b,
c from the
equations
cy
+ bz
=
az + ex
=
bx +
ay
-
ax +
by
+
cz.
22. Shew
that,
if
xa
yb_xb
ya
yb
xa
ya
x-b'
then either a
=
b or x
= <
G02 MISCELLANKOUS EXAMPLES.
23. Shew
that,
if a + b + c
=
0,
then
,
a
8
-6-c , /_ 1\
r
-
.
r
-
b c a
1
\
a
24. Prove that the
necessary
and sufficient conditions that
the roots of
X
(cue
2
+ bx +
c)
+
p. (a'x*
+ b'x +
c')
=
may
be real for all real values of X and
p.
are that
6
a
-
4oc>
and
(6c'
-
6'c) (06'
-
a'b)
-
(ca'
-
c'a)
2
> 0.
25. Find the condition that the solution of the
equations
2a
(1
-x)
+
(b
+
c)
x= 26
(1 -y)
+
(c
+
a) y,
2d
(1
-
x)
+
(e +f)
x= 2e
(1 -y)
+
(
may
be indeterminate.
26.
Having given
that x +
y
+ z
=
0,
and
a?
y
3
2
s
.
i
-
+
-
+
-
r
=
0,
6 c c-a a-o
prove
that 2
(6
-
c) (6
+ c
-
2a)
2
/a^
=
0.
27. Prove
that,
if
then
(a
+ 6 4
c) (a'a;
+
6*y
+
0*2)
=
(6c
+ ca +
ab) (bcx
+
cay
+
abz).
28. Solve the
equations
:
(i) J2*?-1
..... /z-a\
/x-b\* /x-e\
(111) (
- -
)
+
( i }
+
(
- -
)
\ ->. j. * i
\x
+
b/ \x
+
cj
(x-a)(x-b)(x-e)
T ft
j-r-T
r- 1
=
U.
(x
-*-
a) (x
+
b) (x
+
c)
MISCELLANEOUS EXAMPLES. 603
(iv)
ax +
by
+ cz
=
0,
tfx* +
b*y*
+ cV
=
3
(6 c) (c a) (a b).
(v)
2yz =y +z,
(vi)
x
(x
+
y
+
z)
+
yz
=
a
2
,
2zx
=
z +
x,
y (x
+
y
+
z)
+ zx
=
6
2
,
2xy
=
x +
y.
z
(x
+
y
+
z)
+
xy
=
c
8
.
(vii)
a?
(a?
+
4)
=
y
2
(y*
+
4)
=
z
z
(z
2
+
4)
=
bxyz.
(viii)
ayz
+
by
+ cz
=
bzx + cz +
ax
=
cxy
+ ax +
by
=
a + b + c.
(ix)
x
y
=
2,
xz
yw
=
3,
xz
z
yv?
=
5,
and
xz
3
yvt?
=
9.
(x)
x +
y
+ z
0,
ft
77,
IA
. ~n, -\
L >
(c
-
a)
x
(a b)y (b c)z
b c c a a b
_
1 1 1
x
y
z a b c
cy
+
bz_az
+
cx_bx
+
ay_
*
'
y
+ 2 + a;
o j. f
(xii)
yz
+ icaj.4-
a6y
+ caz
=
0,
2c + aia; 4-
cay
4- 602;
=
0,
xy
4- cox 4-
6cy
4- 062
-
0.
29. Shew
that,
if
x
and r ,
a4-y 04-y c4-y
x
y
z
a/3y
then will
_
+
1,
+
_
-
1,4.
_TJ-
.
a b c aoc
604
MISCELLANEOUS EXAMPLES.
30. Prove that the three
equations
a b a b' .
-rx-j-fy
+ +
=
0,
a b
y
x
y
b c b' c'
A
y-
2+ +
=
0,
b
*
c
y
z
c a c a'
-fZ
--
7
x + +
=
0,
c a z x
are inconsistent unless a + 6 + c
=
0.
31. Eliminate
<c, y,
z from the
equations
p
=
ax +
cy
+
bz,
q
=
ex +
by
+
az,
+ cz
32. Determine those
pairs
of
positive integers
whose
product
is twelve times their sum.
33. Prove
that,
if
Xi
:
y1
: z
1 ; x^
:
y
t
:
zf, x,
:
yt
:
z,
be the three solutions of the
equations
x
9
+
y
3
+ 2* +
axyz
=
0,
Ix +
my
+ nz
=
0,
then x
1
x
t
x
t
+
y^ ys y,
+ z
}
z
t
z
8
=
0.
34. Shew
that,
if
x+
y+ z=pi, yz
+
zx+xy=p,, xyz
=
p3 ,
and if
a^-f-yz
=
a, y*
+ zx
=
b,
z* +
xy
=
c;
then a + 6 + c
=
pf pt ,
bc + ca + ab=
p*p
t
-
and abc
=
pfp, Gpiptps
+
pf
+
8/)s
2
.
35. Shew
that,
if
x
1
=
x + x* +
xf
+ . . . ad
inf.,
a:,
=
!
+
Zj*
*-
Xj*
+ ... ad
inf.,
then nx
n
=
nx
%
+
n
8
^
1
+
n'a^*
+ . . . ad inf.
-
MISCELLANEOUS EXAMPLES.
605
36. Find four numbers
x,
y
y z,
u such that x
1,
y1,
2+1,
u + 1 are in arithmetical
progression,
as are also x
3
,
y
2
,
z*
and
a?, a?,
w
2
.
37. Shew that
any
number can be
expressed
in the form
where
plt p2 ,
... are
positive integers,
and
p
t
<
2,
j9
2 <3,
and
so on.
Express
999 in this
way.
38. Find JV^so that the
arithmetic,
geometric
and harmonic
means of the first and last terms of the series
25, 26, 27,
...... N-2
t N-l, N,
may
be terms of the series.
39. The result of
multiplying
a whole number of three
digits
in the scale of r
by (r
1
)
is to
interchange
the first and
last
digits
and to increase the middle
digit by
the difference
between the first and last. Find the number.
40. Shew that the number of
permutations
of n
things
r
together
when two of the
things
are excluded from
having
definite
positions
is
41. Shew
that,
if n be a
positive integer,
the series
(2ro-l) (2n-l)(2n-2)
~-
is
equal
to
(
-
!)*- (2n
-
2) (2n
-
3)
...
42. Sum the series
1 1 1
j_
I ,Tk I f\
[n-1 \l
n-3
43.
By
means of the
identity
{1-
a;
(2 -)}-*=(! -x)-\
606 MISCELLANEOUS EXAMPLES.
prove
that
|2n |2n-2 |2n-4
_
+
'
=2"
o
~
i o i . i
|n
[n |J.
[n-1 |n-2
|2
[n-2 |n-4
44.
Shew
that,
if
a, 6, c,
e be the coefficients of
any
four
consecutive terms in the
expansion
of
(1 +)*,
then
(a
+
6) (c
2
bd)
=
(c
+
d) (6
s
ac).
45. Prove that the coefficient of of in the
expansion
of
(1 +)"(!- )
-"
is
ln
+ r-1 ln +
r-2
re/'jr-l^
ln + r-3
9*
I
on
i I
=.
i
\ /
on 2
I
,
<
T
~
n . a 5 ^ ";
T
|w-l |r
ln-2
46. Sum the series
(l
a
+
l)
|j_+(2
8
+l)
[2
+
(3
8
+l)
(3^+...
+(7i
a
+
l)[n.
47. Shew
that,
if n be
any positive integer greater
than
2,
to n terms is
equal
to zero.
48. Shew
that,
if
then
1-CKC 1
-.
5
7-5 S-T
--
=-=-
=
1
1 + ax 1
-
(o
2
+
2)
a; + aV
x
being
such that both series are
convergent.
49. Prove
that,
if x <
1,
x a? . x
9
2g* 3s
8
MISCELLANEOUS EXAMPLES. 607
50. Prove that
fl
i
a
I
b
I
608 MISCELLANEOUS
EXAMPLES.
57. If
2n
+1
+... ad
inf.,
prove
that
and deduce that
58. If
n
H
r
denote the sum of all the
homogeneous pro-
ducts of r dimensions of the letters
a, b, c,
. . .
;
shew that
the sum of all the
homogeneous products
of r dimensions
wherein no letter is raised to so
high
a
power
as the
with,
will be
59. Shew that the sum of all the
homogeneous products
of
a, 6,
c of all dimensions from to n is
60. A man addresses n
envelopes
and writes n
cheques
in
payment
of n
bills;
shew that the number of
ways
of en-
closing
within each
envelope
one bill and one
cheque
in such
a manner that in no instance are the enclosures
completely
correct,
is
61. In a
plane
n circles are drawn so that each circle inter-
sects all the
others,
and no three meet in a
point.
Prove that
the
plane
is divided into n* n + 2
parts.
"
MISCELLANEOUS EXAMPLES. 609
62. Shew that
n" n
(n-\)
n
n(n-l) (n-2)"
x + n laj + w-l 1.2 oj + n-2
63. Test the
convergency
of the
following
series :
2 3 n
F*I y+T
-
+
^TT
+
,..
x
23 w
*~
+
~^
4 +
^
+
--
2
2
-2+l
2'. 4' 2 . 4
s
. 5
3
. 7
s
y~Ty
+
3 . 3
s
. 6
s
. 6
s
+ "
2. 4
8
. 5
8
. 7
8
...3?t-
3
3
. 3
s
. 6
s
. 6
3
...(3w)
3
(3w)
3
/ a \
w
64. Find the condition that the series whose rth term is
(TO
+
n)
r
(2m n) (3m
-
2n)
. . .
(rm
r-l
n)
(m
ny
(2m
+
n) (3m
+
2n)
...
(rm
+ r 1
n)
may
be
convergent.
65. Shew that the limit when n is infinite of
|"
.
J V
1.2.3...n
66. Shew that
-
^-
-
n + 4w + TO
8. A.
39
610 MISCELLANEOUS
EXAMPLES.
67. If
a 1 a* 1 a
+ * +
a a
then will
x<f>(x)
x+ x+l +
x + 2 +
"
68. If
pn
/q
n
ftnd
pn
-i/q
n-i
be the last and last but one
convergents
to
11 11
a + b + ... + k+ I
'
prove
that
11 1111 1
1
_Pg
r
+
/
>
-ig-i
a+6+ ...
-i- A+ / + + & + ... + -!-
9
r
s
+Pn9'-i
69. Shew that the nth
convergent
to the continued fraction
123 ?t
.
V|r_
2
_
3
-
4
-
...
-
n + 1
-
...
1S
1 +
2,"
[r
'
70. Find the nth
convergent
to the continued fraction
!
*
JL
n>
2-5- 10-...-n
J
+ l-....
71. Prove that the nth
convergent
to the continued fraction
is <r
w
/((r
n
+ 1
),
where
o-,,
=
Oj
+
a^j
+
Ojaja8
4-
... to n terms.
72. Prove that the continued fraction
_
I
__^^ -^_ J_
1-a 1
-
a
*
1 a*
MISCELLANEOUS EXAMPLES. 611
73. Prove that a
being greater
than
1,
the rtth
convergent
to the continued fraction
2a
-
1 + 2a + 2a + 2a +
is
equal
to
<
a
-
1)
{g-l
1
(2a-l
~
2a l^a-l
+ -
+ =
f
.
(no,
-
1) (n
+ 1 a
1)
)
74. The difference of the first
convergents
of the
square
roots of two consecutive odd
integers expressed
as continued
fractions is
1,
and the difference of their third
convergents
is
497
;
find the
integers.
75. Shew that, in the
ascending
continued fraction
a
l
a., a
s
a
t
the n
convergent pn
/q
n
is
given by
the laws
Hence shew that the value of the fraction
1+2+3+4+
2 3 4 B
is
unity.
76. Prove that the continued fraction
!_
x
x+1 2
(x
+
2) (n-l)(a?
+ n +
l)
1 a:+l-a;+3 x + 5
"
x+2n 1
is
equal
to
77. If
2,.'=a+l
I 1
a
a
+
3
+ + a
r
shew that
^
. 2
8
n
=
V
. 2
l
n_!
.
392
612 MISCELLANEOUS EXAMPLES.
78. Prove
that,
if
p
r
/q
r
be the rth
convergent
to the con-
tinued fraction
a, a,
o
1 + 1 +
"
+
1
'
and
p
r
'/q
r
'
be the rth
convergent
to the continued fraction
-i
^
1
+
1 +
"
+ 1
'
then
qn=p'*-i
+
<l'n-i> <I*-i
=
q'n-i
and
#,_!
=
Oi^'.-a.
79. Shew that the value of the continued fraction
)' (m+2)
*-
H c
4-
ad inf. is TH
-
2m--5
i^
80. Shew that, if n be
any positive integer,
N
n
can be
expressed
as a
non-recurring
continued fraction with unit
numerators,
and in the
particular
case when N
=2,
w
=
3,
prove
that
4 5
.29
3' 4' 23
are con
vergents
to it.
81. Find the
greatest
value of
(a
-
x) (b
-
y) (c
-
z) (ax
+
by
+
cz),
where
a, b,
c are known
positive quantities,
and
a-x,
b-y,
c-z
are also
positive.
82. If
a, b,
c be three
positive quantities
such that the sum
of
any
two is
greater
than the
third,
and if
ax +
by
+ ex
=
0,
prove
that
ayz
+ bzx -t-
cxy
is
always negative
for real values of x.
MISCELLANEOUS
EXAMPLES. 613
83. If TO and n be
positive integers
and m
>
w,
shew that if
x be
positive
1 + x
+ y? + . . . + of
~
x
> 1 + a; + as
8
+ . . . + JE*~*
ra
*
according
as a;
<
1.
84. Shew
that,
if a be
any positive quantity
and
p
>
q,
unless
p
> >
q.
85. Prove
that,
if a; be
any positive quantity,
as"
1
> x~
x
> 1 + x x
3
.
86. Prove
that,
if all the letters denote
positive quantities,
. m
m
. n
n
m + n
and deduce the minimum value of x +
y
when
x
m
y
n
=
a.
87. If
a, b,
c be
positive, prove
that the least value which
(x
+
y
+
zy+*+
can take for
positive
values of
x, y,
z occurs when
x
=
a, y
=
6,
z
=
c.
88. Shew
that,
if
a, b, c, d, a, ft,
y,
8 be
positive quantities
such that
a>o,
b>
(3, c>y,
c?>8,
then
Sa&crf +
8ay8
>
(a
+
a) (b
+
/?) (c
+
y) (d
+
8).
89. If
Oj, a.,, a,, ...,
a
w ; &!,
6
2 ,
6
3 , ...,
b
n ; Cj,
c
2 ,
c
3 , ...,
c
n
be three sets of
positive quantities
which in each set are
arranged
in
descending
order of
magnitude ;
then will
o^Ci
+
o^aCa
+
<h^z
+ -
Sc^ 2&i
2^
n n 7i
'
n
614
MISCELLANEOUS EXAMPLES.
90. Prove
that,
if
a, b,
c be
positive quantities,
(\
U
and
(b
+
c)
a
(c
+
a)
6
(a
+
b)
e
<
jjj
(a
+ b +
c)Y
'.
91. Prove
that,
if a is
any prime
number
greater
than 1
9,
then a
18
1 is a
multiple
of 9576.
92. Prove
that,
if the radix of a scale of notation be the
product
of the different
prime
numbers m +
1,
n +
1, p
+
1,
the
(rnnp
+
l)th power
of
any integer
will end in the same
digit
as
the
integer
itself.
93. Shew that the number of solutions in
positive integers,
zero
included,
of the
equation
x +
'2y
+ 3z;= 6n
is
94. Shew
that,
if n be
any positive integer,
(n+ l)(+2)(n
+
3)...(M
+
n)
will be a
multiple
of 2".
95. Prove
that,
if
q
be a
prime
number and
p prime
to
q,
then
P
(P
+
g) (P
+ 2
g) (P
+
^Tg)
n
will,
when reduced to its lowest
terms,
have for its deno-
minator a
power
of
q.
96. Shew
that,
if
p
be
any prime
number,
the sum of the
products p-2 together
of the numbers
1,
2,
3,
......
,
p-\
will be divisible
by p*.
97. Shew
that,
if
p
be
any prime
number,
the sum of the
rth
powers
of the numbers
1,
2, 3,
......
,
p
-
1 will be divisible
by p,
if r be
any positive integer ^> p
2.
MISCELLANEOUS EXAMPLES. 615
98. Shew
that,
if
p
be
prime,
|2-2|Lp
is divisible
by p*.
99. If a white balls and b black balls be
placed
in a
bag,
and be withdrawn one at a
time,
without
replacing
them
;
shew
that the
probability
that all the black balls will have been
drawn in the
process
of
drawing
a balls is
a
being
not less than b.
100. A
bag
contains 10
coins,
and it is known that 2 of these
are
sovereigns.
Two coins are drawn out and neither is a
sovereign;
shew that the
probability
there were
only
two
sovereigns
in the
bag
is one-third.
101. A
bag
contains
any
number of
balls,
which are
equally
likely
to be white or black
;
one is drawn and found to be
white. Shew that the chance of
drawing
another white
one,
the first ball not
being replaced,
is two-thirds.
102. Prove
that,
if
a, ft, y
be the roots of the cubic
then the
equation
whose roots are
(a-/3)(a-y), (0
-
y) (/?
-
a),
is *
8
103. If S
r
denote the sum of the rth
powers
of the roots of
the
equation
x
m
+ aa? + b
=
0,
shew
that,
if m >
5, -S^m-i
=
0.
104. Two roots of the
equation
x*-Qx>+ ISar
2
-
30* + 25
=
are of the form a
+
ifi, ft
+ to..
Find all the roots of the
equation.
616
MISCELLANEOUS EXAMPLES.
105. In the
equation
ar"
1 *2
+
(2
-
a)
ac*
+L
+
(b
-
a +
1)
a* +
(a
+
2nb)
x
+ a +
(2w- 1)6
=
0,
prove
that two of the roots are
equal,
and that of the rest the
sum
one, two, three,
<kc.
together
are in arithmetical
pro-
gression.
106. If n be an even
integer,
and
p
, plt p
a
...be all
positive,
and if H be the
greatest
of the ratios
Pi Pi Pn-i
Po' Pi*
'
Pn-a'
and Kbe the least of the ratios
Pj
P* Pn
.
JPi' P*
'"'
P-S
then will all the real roots of the
equation
=
lie between Hand
K,
and all the roots of the
equation
will be
imaginary
unless H be
greater
than K.
107. Shew that
1 1 1
1 1+a 1
to n rows and columns
108. Shew that
1+a 1 1
1
1+a 1
1 1 1+a..
to n rows and columns
=
a
1*
109. Prove that
1111
a b c d
a' b' c' d'
aa' 66' cc dd
-(a-c)(b-d)(a'-b')(c'-d').
MISCELLANEOUS EXAMPLES. 617
110. Shew that
a
2
+ b
1
+ c
2
,
be + ca +
ab,
be +ca + ab
be + ca +
ab,
a
2
+ 6
2
+ c
2
,
be + ca + ab
be + ca +
ab,
be + ca +
ab,
a* + 6
2
+ c*
111. Solve the
equation
(x-a)*, (x-b)*, (x-cY
(x
+
a)
s
,
(x
+
b)
s
,
(x
+
cj*
112.
Shew that
=
(a
3
+ b
3
+ c
3
-
3a5c)
2
.
=
0.
b*<P, be,
b + c
c
2
a
a
, ca,
c + a
a
a
6
J
, ab,
a + b
=
0.
113.
Simplify
1, 0, p,
1,
b +
c, be,
1,
c +
a, ca,
(p +c?)(p
+ a
2
)
1, a+b, ab,
(p
+ a
z
)(p
+
b*)
114. Shew that
-6V
115.
Shew
that
b' +
c*+l,
c'+l,
6 +
c,
c +
a,
-o
2
6
2
V+l,
b + c
a' +
1,
c + a
a* + 6* -f-
1,
a+b
a +
b,
3
=
(be
+
ca +
ah)*.
ANSWERS TO THE EXAMPLES.
EXAMPLES I.
1. 4x. 2.
-
2x
-
Gy
-
4z.
555
Qm (^
-|-
^
>
-f- C *4 fl A
. 2m
a
+2mB+27i
a
.
8.
13 8
1O. -a~-
!r
b
+
-c. 11. x
a
-x-9.
u o
12. -6a
4
+3a
3
6-3a5
8
+55
4
.
13. 2s;
1
-
7xy
+
ly\
14. -bc +
lca
+
lab. 15. -3a*
+
2&
16.
x-y.
17.
-oy-3z.
18.
19. 6
+
d. 2O.
y.
21. 4a. 22.
-
23. -4n
+
4m. 24. 20. 26. -20.
EXAMPLES H.
82
4. x
+
y*.
6. x
4
-!. 6.
y'-x
5
.
7.
x
4
-a;-
+
4x-4. 8.
9. x
8
+
.r
4
+ l. 1O.
ANSWERS TO THE EXAMPLES.
G19
12. 4x
13. 6a
+
lla
B
6
-
16a
4
Z>
2
--
20a*b*
-
29a
2
6
4
+
15a&
5
-
36
6
.
14. 2a
6
3
6 -
3a
5
z
s
t/
2
+ 8a
4
arV
-
lla
8
*
V
+
fia'^^
8
+
20aa:t/
10
-
10y
12
.
16. 2o-3a
2
+a
3
+6a
4
-5a
5
-18o
+
44a'
16.
a+b
s
+c
3
-3a6c. 17.
18. 8a
3
+ 276
3
-c
s
+18a6c. 19.
2O. x
8
-256j/
8
. 21.
32. a;
u
-3^+3a;
4
-l. 23.
24. a
8
+
8a6
2
+
48a
4
6
4
+
128a
2
Z>
+
2566
8
.
25.
(i)
a
2
+
4&
2
+9c
2
+4a&-6ac-126c,
(ii)
a
4
-
2a
3
6
+
3a
2
6
2
-
2a6
3
+
b
4
,
(iii)
6
2
c
2
+
c
2
a
2
+
a
2
6
2
+
2a
2
6c
+
2a6
2
c
+
2a&c
2
.
(iv)
!-
(v)
26.
(i)
a
8
+
ft
3
+
c
3
+
3
(a?b
+
aft
2
+
a
2
c + ac
2
+
6
2
c + 6c
5
)
+
6a6c.
(ii)
8a
3
-
276
3
-
8c
3
-
36a
2
6 -f 54a6
8
-
24a
2
c
+
24ac
2
-
27. 8z.
30. x*-2a?x-2ax>-a
l
. 41. 0.
6O. Sa
3
-3Sa6c.
EXAMPLES in.
1.
x-3j/.
2. z
2
+
4y
2
. 3.
4.
-3x-2y.
6.
6.
8.
9.
1O. l
+
x*
+
x*
+
x
6
. 11. l-
12. 2-3x+
2z
2
.
13. 2a5
2
620 ANSWERS TO THE EXAMPLES.
14. x*-
16. *
2
-
18.
3O.
33.
34.
36.
38.
39.
15.
17. a
+
2b-3c.
19.
21.
33.
36.
2az-(3fe-4c)y.
37.
EXAMPLES IV.
3.
6.
7.
9.
11.
18.
16.
17.
10.
31.
38.
34.
36.
37.
38.
3o-26)(4-
8(o-6)(a
24*
(*-l) (*+!)*
(a -4) (a
+
2).
(l-21x)(l
+
3x).
ab(a-b)(a-3l).
(b+e-a)(b
+
c-Sa).
(x-2)(x+2)(x-5)(x
3.
4.
6.
8.
10.
13.
14.
16.
18.
3O.
33.
(2ar-3aZ>) (2x
+
B
(2y
+
2z-x)(2y-2z
+
x)
16
(a
-
6) (a
+
6) (a*
+
&*).
16x(2-3*).
(4-*)(3
+
x).
-4(a-l).
0*6
(0+6) (a+4&).
(5x- -y) (2x
+
y).
(*-6)(*+b-2a).
36.
(x
-
o) (x
+
2y
+
a).
(a
+
b
+
c-d) (a
+
b-c +
d) (a-b
+ c
+
d)(
-a + b +
c+d).
ANSWERS TO THE EXAMPLES.
621
EXAMPLES V.
3.
0.
7.
9.
11.
13.
16.
17.
18.
10.
21.
22.
23.
25.
27.
29.
31.
32.
33.
30.
30.
38.
4O.
(a-b)(c+d)(c~d).
(ax
+
b)(cx*
+
d).
(a
+
1) (a
-!)(&
+
!) (6-1).
(a;
2
*
-
1) (fz
-
1).
2.
(a+b)(c-d).
4.
(ax
+
by) (cx
+
dy).
6.
2(a-d)(a
+
b
+
c
+
d).
8.
(a-6)
2
(a
2
+
a&
+
&
2
).
1O.
(*
+
*) (x-z)(y
+
z) (y-z).
12.
14. a;
+
x- 3
16.
(x
+
a
+
b)(x-a-b)(x
+
a-b) (x-a+b).
(a;
2
-2y
2
z
2
)
2
. 2O.
(x
-
2b
+
ab)
(x
-
2a
-
ab).
24.
(2
26.
(2
28.
(1
-
ax) (1
-
ax
-
ex'-).
3O.
(6
+
c) (c
+
a) (a
+
6).
(x+y-Bz)(x-
(a-36+c)(a
+
6-3c).
(l-ax)(l
+
ax
+
ba?).
-(6-c)(c-a)(a-6).
(o-6)(a-c)(6
+
c).
(x*-xy+y*){x(a+l)+y(b
+
l)}.
(xy
+
ab) (ay
2
+
b*x)
. 34.
(2* -)(*- j/)
8
.
(x
2
-
yz) (t/
2
-
x) (z
2
-
xy).
(a: +4) (a;
+
2)
(a: -!)(*- 3).
37.
(a;
+
4) (x+2) (a;
39.
(x
+
2) (x
+
6) (a;
2
+8a;+10).
622
ANSWERS TO THE EXAMPLES.
.
EXAMPLES VI.
1.
3(y-z)(z-x)(x-y).
3. 5
(y
-
z) (z -x)(x- y) (x*
+
y*
+
z*-yz-zx- xy).
3.
(b+c)(b-c)(c
+
a) (c-a) (a+b)(a-b).
4.
(b-e)(e-a)(a-b)(a+b
+
c).
6.
(6-c)
(c-a)(a-
-9abc).
6.
-(b-c)(c-a)(a-b).
7.
-
(6
-
c) (c -a) (a
-
b) [6V
+
cV
+
a*6
2
+
a6c
(a
+ 6
+
c)].
8.
-(6-c)(c-a)(a
9.
-(6-c) (c-a)(
+
abc).
11.
80abc(a*
+
b*
+
c*).
13. iabc.
13. 2abc. 14. 4a6c. 16. -4
(6-c) (c-a) (a
-6).
17.
18.
-(6-c) (c-a) (a- b).
19. -2
(6-c) (c-a) (a-6) (a+
6
+
c).
30.
-(6-c)(c-a)(a-6)(3a+36
J
+3c*--56c + 5<ja +
5a6).
31.
(6
+
c)(c
+
a)(a+6).
33.
-(6-c) (c-a) (a-6) (a
+ 6 +
c)
8
.
23.
(x
+
y
+
t)(yz
+
tx+xy).
34.
(6
+
c) (c
+
a) (
36.
Vxyz(x
+
y
+
z).
36. -3
(6-c) (c-a)
(a- b).
37. 16
(6
-
c) (c
-
a) (a-b)
(d-a) (d-b) (d-c).
33. 27a
1
J
I
(a
+
6)
1
. 34.
(a*
+
6
1
)
2
(c
1
+
d*)'.
36.
(6-c)(e-a)(a-6)(a-d)(6-d)(e-d).
37.
-(6-
ANSWERS TO THE EXAMPLES. 623
EXAMPLES VII.
1. a
-
b.
624
ANSWERS TO THE EXAMPLES.
a + b
11.
;
r
r-7-
r .. 12.
yz
+
zx
+
xy.
(
a _5
+
c)(_^i-i-i-i-/'\
13.
-l(y-t)(z-x)(x-y).
14.
15.
i"
+
** +
*y. 16.
x+y+z
"
ii-
l8-
3=fr-
1*.
r^Lp.
20.
x-2a
48a
'
(x
+
a)(x
+
3o)(x
+
5a)(a;+7a)'
'
x
(x*
-
a?) (x
s
-
la?)
'
23. 0. 24. 0. 25. 1. 26. -1.
27.
'. 28.
0. 29. 1. 30. 2.
31. a
+
b
+ c. 33.
33.
(a
+
6
+
c)
1
. 34. a
+ b+c.
35.
-
x .-
0-
(a+6-c)
37. 16 . 38.
-
-
EXAMPLES X.
1.
2a-6,a-26.
2.
l,.
3.
0,
.
'b b-a
4. a
-26,
6-2a. 5.
1. 6.
7.
1,
-8.
8. 1.
9.
0,
5^2.
10.
6,
-8J.
-
ANSWEIIS TO THE EXAMPLES.
625
60
'
29'
13.
0,|
14.
15. 5.--. 16.
v/6.
4
/
5
17.
iv/afc, J-ab.
18.
0,
-5.
20.
31.
a,
-
T.
32.
-
23 34
o+& ab cd
28.
0,
Ji^b.
3O.
8,
-5. 31.
2,
3.
32. 1. 33.
0,
a
2
-Z>
a
. 34.
a,
6.
36.
,
. 36.
0, 4(
+
&).
37.
-a,
-6.
38.
i(a-&).
39.
i^/afr.
4O.
0,
2
Jab.
41. 33. 42.
43. 0. 44. adb&.
45.
-
'
'
~
'
S. A. 40
626
ANSWERS TO
THE EXAMPLES.
48.
^-,J{2a*b*c*(a*
+
b
:
'+c*)-b*c*-c
t
a*-aW}
&CIOC
(-
bc
+
ca+ab) (bc-ca
+
ab) (bc
+
ca-ab)\.
4
49. Values between 3 and
-
5
.
62. Values between 3 and
-
.
63. x lies between
-
2 and
8,
and
y
between
-
9 and 1.
64. x between
-
2 and
10,
and
y
between
-
1 and 5.
66.
g.
11.
13.
15.
18.
20.
21.
23.
25.
27.
29.
30.
ANSWERS TO THE EXAMPLES.
0,
-5,
^(-5=^-15).
12.
a, -9a,
-
la, -8a, (-1* ^-167).
14.
-
4,
-
6
,J
(
-
15
t i
627
21
3,
-.
-1,
-1, 2, 3,
-4.
a, 6,
c.
16. 17.
19.
3, ,
(_
24.
1, (
-
26.
9,
-3^
9, -6,
^
(3 7^215).
28.
a, 6,
^(a+b).
a, 6;
(
o
+
&)
a, 6, (a
+
6),
31.
a, i,
(a
+
6).
32.
a, b,
33.
34.
a, 6,
a,
b. 35.
o, 6,
and roots of
(a-x)(x-6)
=16
(a-6j
l
.
37.
o-26, 6-2a,
-
O
38. Boots of
x(a-x)
=
402
628 ANSWERS TO THE
EXAMPLES.
39.
-
6 a
41. a
,,
a+b a+b
42.
|{-6^(t
+
4)},
|{-aV(a'
+
4)}.
43.
3
44.
-(a
+
ft
+
c),
-5
(a
+
6
+
c)g^/(Za
:
-
2
i
45. a
+
6
+
c,
5
(a+6
+
c)-^/(2a
5
-7S6c).
46.
a, 6,
e.
*7-
*
\/{-
d
-^
b
+c+
(
d
+d)
} ^x/r^-r^r
EXAMPLES
XII.
=
1,
y=-l.
3.
x=3,
y
=
6. 4. x=
j/=3.
p
. x
=
6,
y
=
a. 6. x
=
ai>,
y=
-a -b.
7. x
=
a
+
fc, y
=
a ft. 8. x
=
y
=
<i.
9. ar
=
o,
y
=
6. 1O. x
=
o
(a-6), y
=
6(a -6).
II.
x=-3,
y
=
3,
*
=
1. 13.
|,
f
i,
-i.
13. x
=
y=z=l.
14.
*=6+c-o,
y=c
+
a-6,
z
=
a
+ fc-c.
16. x=fc
+
c, y
=
c
+
a,
z
=
a
+
6.
18. x=
,
,
ANSWERS TO THE EXAMPLES.
629
_
a
_
b
_
c
'~(a-b)(a-c)'
y
~(b-c)(b-a)'
*~(c- a) (c- b)'
20. x
=
a,
y
=
b,
z=c.
21. x=-a+
b
+
c,
y
=
a-b
+
c,
z
=
a + b-c.
22. x
=
a(b-c), y
=
b(c-a),
z
=
c(a-b).
23. x
=
l,
y
=
0,
z
=
0.
24. x
=
abc,
y=bc
+
ca
+ab,
2
=
a
+
6
+ c.
m
(ni
-
b) (m
-
c)
25. z
=
^ rf ^,
&c. 26. x
=
a,
y
=
6,
z=c.
a(a-b)(a-c)
a (a
+
b
+
c)
27. x
=
b
+ c, y
=
c
+
a,
z =
a
+ o. 28. j= . -. . . &c.
(a -b)(a- c)
29. x
=
-(ro
+
7i), y
=^(n
+
l), z=-(l
+
m).
30. x
=
y
=
z
=
l
2
+
m?
+
'n? mn nl Im.
31. lx
=
my=nz=l.
32
x=^
a)(
;.
+
.
6)(a
.
+
C)
.&c.
33. ,
=
J
(a -p)(a-y)
6
34. x
=
a6cd,
y=
-(bcd
+ cda + dab +
abc),
EXAMPLES
XIII.
1.
12,
11. 2.
1, 1,
1
4.
2,
3;
yx/3, Ty^S.
5.
12,7; -7,
-12. 6.
a,
b;
2a-b,2b-a.
2ab 2ab a
/
-
b
7
^ao ^a
e>
a^
a2+
o
o a
o+o
o
a
630 ANSWERS TO THE EXAMPLES.
sl
11.
7, =p5; 5, T
7.
13.
64, 8; 8,
64.
15.
i(lJ^TT), 1(1--
16.
1, 1;
2
^7,
S
18.
2,
1
;
1,
2.
3O.
-
,
2.
33.
6, 6;
-|(1
33.
34.
ab
36.
4,2; 2,4;
39.
1,-; -1,
-2.
31.
13.
9, 4; 4,
0.
14.
-6*^/30, 6^30.
2, -1; -1,
2.
18.
2a, 2ft; -a,
-b.
31.
6,
a.
35.
8,4; 2,
4.
r^B.
38.
3, -6; -i, r.
A A
+^ b a
ij2
>
83.
b,
a; b, -;
^,
a;
$,
EXAMPLES
XIV.
ca a& be ca ab
''*
o' ft' c*
2.
a
'
b c
x
y
t
ANSWERS TO THE EXAMPLES. 631
2abc 2abc 2abc
'bc+
ca
+
ab' be-ca
+
ab' bc
+ ca-ab*
Q*
U, U Ul .
' "
,
:
"
b
+
c-a c
+
a-b a
+ b-e
6. a;
=
y
=
z=2.
7>
__*
y
=
z
=
1
b+c-a
c
+
a-b a
+
b-c
*J(2a
+
2b
+
2c)
*
be ca ab
8. o=t
-
u
=
c== .
a b c
*
y
*
=
i
.1! 7/ K
10
y
a(-a
+
b
+
c) b(a-b
+
c) c(a
+
b-c)
=
-7T-
b
+
c-a c +
a-b a
+
6-c
1
,y {(6
+
c-a)(c+a-fe) (a
+
fr-c)}
'
a: 11 x
12.
13.
ft+c-a
c
+ tt-b a
+
b-c
(fe
+
c-a)
8
+
(c
+
a-i)'
2
+
(a
+
b-c)
8
'
a:=1
\/-
14. x
=
(
1*
1. 2, 3; , ,
.
17.
1, 2,
3. 18.
3, 5,
7. 19.
0, 4,
5.
|, |,
20.
3, 3,4.
21.
0, 0,
0; , ,
.
632 ANSWERS TO THE EXAMPLES.
33.
-a, b, c; a, -b, c; a, b,
-c;
(l=W-7), (
24. x
=
25. z=-^-
(
-
+
-
)
,
X
26.
(c
+
a)
2
-
(a
+
b) (b+c)
1
27.
a, 6, c;
28. x=
j-r-(bc-ca
+
ab) (bc
+
ca-ab),
&c.
39.
a,0. 0; 0.^,0; 0.0.
C
;
3aa6 S.fccc 3
30.
0.0.0;
--a,-. -; -, --^. -; -. -,
--c
t
=
_
(
-c c-a a-b
ax
by
cz I -abc
-c c-a a-
EXAMPLES XV.
1.
20,
30. 3. ^
10,
B
15,
C 25. 3. 20
years ago.
4. 2*. 6.
5, 15,
30. 6. 5
days.
7. 1800. 8. 58. 9. 30 miles.
1O. 120 Ibs. 11. 4
days.
12.
3G, 9,
12,
15.
ANSWERS TO THE EXAMPLES.
633
13. 48 miles.
14. 15 miles. 16.
54, 81,
108.
17. A
450,
B
225,
C 237.
10s.,
D
87. 10s.
18. 5.
19.
38,83.
2O. 18 miles.
31. At 1
o'clock,
15 miles from
Cambridge.
22. A
10,
B
5,
C 1000.
23. 25. 24.
9, 7; 8^/2, ^/2.
25. 50miles.
26. 576.
28. 3 miles an hour. 29. 3 hours.
3O. 253. 31. 2
gals,
from the
first,
and 12
gals,
from the second.
32. 15 minutes
past
10. 33. 9
o'clock,
30 miles from
Cambridge.
34. 45 and 22
J
miles an hour. 35. 3.
36. 450 miles. 37. 30 miles.
EXAMPLES XVI.
3. a + b +
c
+
abc=Q. 4.
(b
+
c-a)
(c
+
a-b) (a
+
b-
c)
=
8.
6.
~- -.
=,1; 7.
(b-d)
2
(o
+
a)
2
8. o
s
+ 2c
3
-3a6
2
=0. 15.
1. aP +
bm?
+
cn
z
+ lmn
=
4abc. 17.
18.
19.
2O.
EXAMPLES XVIL
1. *6. 2. 1. 3. y.
4. 1.
5.
x-y.
6.
a^+l +
ar-
4
. 7.
8. x$-x~^' 9. o^ + a"a;T
+ a^x +
oa;
+x5. 10.
12. 4x
2
4
3x
+
2
-
Sz-
1
. 13.
634 ANSWERS TO THE EXAMPLES.
14. o$:r~S.f
a3;iri
+
tt~ixi
+
a~3x
5
.
15.
x-y-
3
*-x$y
20.
(i)
aS-ai&
(ii)
a^x^
-
a* x^""
y^
+
a-x^y
-
a$x%
y*
+
a* x
y*
-
t/*.
(iii)
a?
+
b-x$ +c*.T$-bcx-cax*-abx*.
(iv) {x^ +y*
+
2*
-
y3z'
-
z^x^
-
x^y^} {(x
+
y
+
)*
EXAMPLES XVm.
1. 2-
v
/3. 3.
5-JI5.
3.
-s/2.
* 52-
6. 0. 6.
14*.
7.
V
3 + 3
N/2~N/
3Q
e.
V
30 +
V
3
-
3
N
/2
.
1
10
13.
4/2-1.
16.
5-^3.
19. 2-
N
/3. 2O. 1.
23.
26.
3
ANSWERS TO THE EXAMPLES.
635
1O. x$-2xi-xi. 11. x&
-
'2x$
+
x*
. 12. a~%xt-x*-at.
13. x-8. 14.
x*-xy+y*.
16. l-3a;
2
16.
2(bc
+
ca
+
ab).
17. x^-x
(y
+
z) -yz.
18. a?+b\
21. 4
=
20, J5=68,
(7=
-44;
or 4
=
52,
B=
-68,
C=76.
23.
af=gh,l
=
lt
EXAMPLES XX.
5. x=3. 6. a : b : c=2 : 3 : 4.
17.
(i)
-
,
1.
(ii)
0,
1.
(iii)
oo
,
0.
o
EXAMPLES XXI.
2.
2, 4, 6,
8; -2, -4, -6,
-8. 4.
2,4,8.
14.
6, 12, 24,
&c. 17.
3, 9,
15.
EXAMPLES XXH.
1. 31. 2.
(r-l)(r-l)
,1000....
4.
lib., 21b.,
10241b.
7. 46. 8. 6. 13. 502 or 361. 15.
288,
289 or 290.
16. 2775 or 2525. 17. 135. 18.
a=8, 6=0,
c=6.
10. 7.
20. 1089. 23.
142857,
285714.
24.
166,
199.
EXAMPLES XXIII.
1.
[20/{|4}
B
. 2.
185.
3. 3
12
;
5. 260. 7.
*7t(n-l).
8.
O.
i{n(n-:
630 ANSWERS TO THE EXAMPLES.
m+n-
2
11.
gn(n-l)(n-2)(n-3).
12. 3m. 13.
j^y^j.
17.
|n [
p
-
n +
1
' 2O.
14.
g(-4)(n-6).
23.
2(mn
+
m
+
n-l).
24.
22a+ 22a6
-
2
(n
-
1),
where n is the number of
given
diameters.
EXAMPLES
XXIV.
2.
3.
4.
6.
6.
l-6z
2
+15x
4
-20.|8+15x
8
-6z
10
+
x
12
.
x
10
-
10xV
+
40x<y
-
8Qx*y
9
+
80xy
2
-
[20
10.
-
13.
16.
16.
22.
1.
4.
7.
9.
1O.
_]
[B_|39
3
r
x*
3
1
4
4
x
16
.
11.
70X
4
. 12.
14.
. 924a;>.
-
x
10
and
jgl
[11(10^
(n
924x.
(-I)l2/||n.
*/ u
r
[n-r
i/
2
...
-
945x
2
(2y)
18
-f
45x
(2y)
14
-
(2j/)
!
17. 6435x
7
,
6435x.
23. 7 or 14. 24. 7.
EXAMPLES XXVI.
2.
Convergent.
3.
Convergent.
Convergent.
Convergent.
6.
Convergent.
e.
Divergent.
Divergent.
c.
Convergent
if x > 1
; Divergent
if x i-
Divergent
if x= 1
;
Convergent
if x
4= 1.
Convergent
if x> 1
;
Divergent
if x i
1.
1
ANSWERS TO THE EXAMPLES.
637
11.
Convergent
if x
%
1
; Divergent
if x > 1.
12.
Convergent
if x
:j>
1;
Divergent
if x > 1. 13.
Divergent.
14.
Divergent.
15.
Convergent
if m < 1
; Divergent
if m
4-
1.
16.
Convergent
if m < 1
;
Divergent
if m
-^
1. 17.
Divergent.
18.
Divergent.
19.
Divergent.
2O.
Divergent.
31.
Divergent.
22.
Convergent
if x <
1, Divergent
if x > 1.
If x
=
1,
then
Convergent
if k > 1 and
Divergent
if k
^-
1.
23.
Convergent
if x < 1
; Divergent
if x
<j:
1.
24.
Convergent
if x
^
1
; Divergent
if x > 1.
26.
Convergent
if x <
1,
Divergent
if x > 1. If
x=l,
then
Convergent
if m <
J
and
Divergent
if m
<
^.
27.
Convergent
if a; <
1,
Divergent
if > 1. If
x=l,
then
Convergent
if A < and
Divergent
if k
-
.
EXAMPLES XXVII.
.
(ii) 4(r
+
l)(r
+
a)^. (iii)
_
2.5.8...(3r-l)
_
2.1.4...(3r-5)
M 1J
3.6.9...ar 3.6.9...3r
,.5.2.1.4.7...(3r-8)
.
.. 3.8.
13...(5r-2)
W (
-
X)
-
3.6.9...3V
-"
(^
l.*.*...r_f'
(viii)
_8.8.8.18...(5r-7)
jrj
(il)
g (g+f) (g
+
2p)...(g+i^I. p
5.3.1.1.3.5...(2r-7) r
(xi)
--
j
-i--
x
r
a^
r
,
r
>
3.
2.
(i)
The ninth
term,
(ii)
the
eighth
term.
3. The 39th term. 4. The first and second
terms.
638
ANSWERS TO THE EXAMPLES.
5. After the 12th term. 8.
(i)
X ' 4 ' "
(3r
~
2)
^
(ii)
2a~
r
x
r
.
(Hi)
4ra-
r
x
r
.
(iv)
' '
"'*
-
a~*x* when n is
2.4. 6...n
1.3.5...(n-2)
.
,,
even,
;
-
a~
n
x
n
when n is odd.
2.4.
b...(n- 1)
(v) (2r
a
+
2r
+
l)a-'-.c
p
.
(vi) (
-
l)
r
16
(r
-
1)
a?~
r
x
f
.
EXAMPLES XXVHL
1.
(i)
2.
(ii)
y^.
(in) 4^4. (iv) ^27-2.
(v)
1.
(vi) ^4. (vii)
/A.
(viii)
1.
11
(-l)r-i^"
""" -/ \"
/. 23 L 25
\r-l
31.
-245/8.
32.
246,792.
33. 462. 34. 35.
35. Coefficient of x*
r
is 3
3p
2-
Sf
-
s
a-
3r
-
s
,
of a^*"
1
is
-
3
3r+1
2~
3r-'
-
and of :r
5r+a
is 0.
38.
(
+
!) (
n +
2)(n+3)(n+4)(n+5).
39. 2
+r
|3n
+r-l r
,'3n-l.
EXAMPLES
'
'
4(*
+
3) 8(x
+
5)
8
(a: -hi)'
4
* * + !*'
7 18 1 4*-8
"
l"' '
3(x-2)""4(x-3)'
'
5
(S+ 2)
ANSWERS TO THE EXAMPLES. 639
_J_
1
' '
'
T *
__
'
1-10*
3(1
+
3x)
3 21 21 7
10"
2
(1
-
3z)
3
+
8
(1
-
3x)
J
+
32
(1
-
3x)
+
32
(1
+
a:)
*
2 4
'
2 11 llx-4
13.
-TW
+
14.
5
(a -I)
2
25(*-l)
25(z
2
+
4)'
311
x+2
1 1 1 17 1 3
1 1 11 1 1 3
>
4
(x
+
2)
6
(i
+
2)
2
144
(x
+
2)
9
(x-
17.
(-ir{2--3-'-i.
18.
l-
20.
1(8--1)-|{(
-!)-!}.
21.
I{9
+
5+
2
-2.3+
2
-2->-*}.
1
3
EXAMPLES XXXII.
1. 1-262. 2. 1-48169. 8. 1146-74.
5.
742. 19s. 6d. 7. 785. 10*.
8.
1979. 5s. &d. 9. 1735
nearly.
1O. 122-58.
.
EXAMPLES XXXIII.
!
iJJ_
1
l.fi
^-
'
8(3.7 (4n+3)(47i
+
7)f
'
168'
3.
nn+l3n
2
i-
23 +
46).
4.
640
ANSWERS TO THE
EXAMPLES.
6.
(n +!)-(
+
2).
6.
(
+
!)
180 12
(2
+
1)
(2n
+
3) (2~n~+ 5)
'
S(0
~
180
*
5
3"
+ 5
j '
36
6(n + l)
'- ov '- -' > ""
2;t+5
_5
4
2(n
+
l)(7t + 2)
;
^"-J'
11.
J-
8
8(2n
+
6"
a
+ 27ra
+ 29
29
-
*
36
6(n
+
l)(n
+
2)(n+3)
13 S=
16. Ha
!
+
n(n-l)a&
+
(n
-
1)
n
(2n
-
1)
&
3
.
3
1 1
+ -n
(n
-
1)
a*b
+
-
(n
-
1)
n
(2n
-
1)
o6
+
1
n
8
(
fl
-
1)6.
18.
^n(4n
j
-l).
19. n
(16n
2
-
12n-
1).
23.
gn(n
+
l)(n
+
2).
34.
n(n
+
l)(4n
-
1).
36.
_
4
M
?-
3 /3\*
'
-
* '-
ANSWERS TO THE EXAMPLES. 641
EXAMPLES XXXIV.
.....u+ ....
_
w
2.5.8...3n +
2)
~
'
4.
7 . 10...
(3
5.7...
.
....
L
~~
'
^
(1V)
n
2.
(i)
2
+
3(?t-l)(n-2);
2
+
n(M
-!)(-
2).
(ii) 7n-(-l)(n-2);
jj*(+l)-gn(n-l)(n-3).
(tii)
2
n
+!-n-2;
2+
2
-
1-^
(re
+
2) (n
+
3).
2
(iv)
2
+1
-Ji(i,
+
l)-n;
2+
2
-
4-
^n (n
+
1)
(w
+
2) -(
+
!).
O ti
(v)
^
n(
+
l)(
+
2) (71
+
3);
A
n
(+ 1)
(
B
+
2
) (n
+
3) (n
+
4).
(vi) (n-2) (Ti-1) n(n
+
l)
+
(n-l)
n-n
+
2;
3.
(i) :
'- 4* i- 2* 1
-
fe
1
15
-
14a
4.
(i)
2
H
+
1
-2;
2
n+2
-27i-.
(iii)
i
(3
-
(
-
1)"}
;
(3
n
+
1
-
3}
when ?i is
even,
and
{3"+!
-
1}
when n is odd.
6 a
=
l,
6
=
4,
c
=
l,
d
=
0.
2-3.T-a;
3
' '
> ^
abc...l
i^)-
S. A.
642 ANSWERS TO THE EXAMPLES.
32.
(i) Divergent, (ii)
Divergent, (iii) Convergent, (iv) Convergent.
(v) Convergent
if
-y>a
+
/3, Divergent
if
-y<a+/J.
EXAMPLES XXXVL
7
"-1
1
.
w+2
10-
&-jJ>
2
-
(*..
+
1
V.-i) P*-i
-
a
A-i (
6
A-i
+
) 7>
8
.-
EXAMPLES XXXVH.
1111
(1I1)
6+
11111111 -1
1_
(
+ "
16.
(n
s
+ 3i. 26.
EXAMPLES XXXVIIL
6. 266. 10.
8, 7, 9, 11, 13,
19. 13. 504n-G.
EXAMPLES XL.
1.
(i)
x
=
2,
y
=
3.
(H)
j-
=
l,
y
=
10;
x
=
U,
y
=
2.
(ui)
x
=
4,
y
=
8;
x=13,
y
=
l.
(iv)
696, 3; 625, 18; 651, 33; ; 57,
13*.
ANSWERS TO THE EXAMPLES. 643
3.
(i)
a;=
(iii)
x=15m-7,
y
=
17m-10.
(iv)
x
=
4.
(i) 3, 1, 2; 5, 2, 1; 2, 4,
1.
(ii)
1, 21, 1; 5, 14, 1; 9, 7, 1; 3, 13, 2; 7, 6, 2; 5, 5, 3; 3, 4, 4;
1, 12, 3; 1, 3,
5.
(iii)
2, 8,
3.
(iv)
8, 38, 50; 19, 44,
35
; 30, 50, 20; 41, 56,
5.
5.
(i)
1325, 2; 441, 3; 101, 8; 77,
10; 33, 21; 25, 27; 5, 112; 1,
333.
(ii)
5,3.
(iii)
8,5.
(iv)
6, 1; 13,
14.
7.
195,
121;
52,
264.
8. 3. 9. 20. 10. 3. 11. 3. 14s.
6d.,
4. 5s. Gd.
12. 2s.
Id.,
2s.
10d.,
2s.
lid.,
3s.
Id.,
3s.
2d.,
3s.
3d.,
3s.
id.,
3s.
5d.,
3s.
6d.,
3s.
Sd.,
3s. 9d. and 4s.
13.
11,12,15,24,30.
14.
15, 55; 25, 65; 35,
75. 15. 21.
1-
67' 07'
2>
33397' 33397' 33397'
117
2
pence.
IS.
-
(
. 16. -. 21. ,
42
y
EXAMPLES XLIII.
1.
(i) q. (ii)
27?. (iii) -2p. (iv)
-
3g. (v) 2p\ (vi)
3g.
(vii) -2p2. (viii) 3j)
2
.
(ix) -p3. (x) p/q. (xi) j/2
2
.
(xii) -p^KSq^+p
3
).
2.
(i)
0.
(ii) -3p. (iii) -4gr.
3.
(i) 3p
3
-16pq
+
6ir.
(ii) (2
3
-
4^3
r
+
Sr^/r
3
.
(iii)
4.
(i)
28,
-24.
(ii)
44,
-252.
5.
(i)i>1
2
-2^2
.
(ii) 3p1?2 -p1
3
-3p3
.
(iii) (pn
(iv) Pi+P-1 (2p.2 -p1
2
)/iJn
.
(v)
(vi)
644
ANSWERS TO THE
EXAMPLES.
0. x
3
-
10x2
+ Six -31
=
0. 7. G. 8.
(i)
x^-qx^+prx-r-^Q.
(ii)
x
3
+
2p
(iii) x*(r-
(iv)
x
3
-
(v)
Eliminate x
between
giveu equation
and
y
=
(p
+
a:)
2
+ .
x
( vi)
j/
3
-
(3?
-
p*)
1,2
+
(3^2
-
qp*) y
+
rp
-
qS
=
0.
9.
(i)
Substitute
-
(y
+p)
for x.
(ii)
Substitute
-
^
(y +p)
for x.
(iii)
Substitute
-
(p
z
-2q-y)
for X
s
.
10.
EXAMPLES XLIV.
. 2
s
/3,
-3iV2.
2.
.Si
3.
,
^/2 v
/5.
4.
5. x*-
16x2
+
4
=
0.
e.
8.
1V^T,
12
N/^T.
9.
10.
;}
J
-
4pq
+
8r
=
and
(p
s
+
4
2) (36g
-
lip
2
)
_
1600s.
=
0.
11.
5, 1,
-3.
12.
4, 1, -2,
-5.
13.
(i)
3.
(ii)
+4,
-4.
(iii)
~.
14.
8,
i-
*
2
3,7,
-10.
is.
3, 9,
-*.
o
16.
^2,
-2
^7.
16.
17.
3, 7,
-
10.
20.
(i)
r
!
-p^
=
0.
(ii)
EXAMPLES XLV.
11 1 1
(iv)
4, 4, 4,
--.
a.
(6c-od)
2
=4(&
2
-flc)(c-W).
9.
(i) y*-
(ii) (y
10.
(y
-
10j>2
+
24g)3
_
243 (y
-
+ 64
(4/>r
-
*) (y
-
Wp
n
-
+
24q)
-
512
(2p*s
-
3qs
+
2^)
=
0.
ANSWERS TO THE EXAMPLES.
G45
EXAMPLES XLVI.
5 3
/
1.
.
(i)
-5,
-
w-4w
2
,
-o>
2
-4w,
i.e.
-5,
-;r */-3.
64
(ii)
-4,
--3w
2
,
-w
2
-3w.
(iii)
10,
2w
+
8w-,
2
2
+
8u>.
(iv)
8,
o>
+ 7w
2
,
w
2
+ 7w.
(v)
-2-094...,
-
1-703. ..to- -391... w
2
,
-
1-703. ..w
2
~
-391... w.
(vi)
3-0913,
2-1699+-9214o/
!
(
2-1699w
2
+-9214w.
2. .
(i)
-1, -3,
l2i.
(ii) li^
2
.
-
1:t 2
N
/^L
(iii) 3J^5, -3J^T. (i
3.
(i)
One real root between
-
3 and
-
2.
(ii)
One between
-
7 and
-
6,
one between 1 and
2,
and one
between 5 and G.
(iii)
One between and
1,
and one between 1 and 2.
(iv)
Two between 2 and
3,
one between and
-
1,
and one between
-
4 and
-
5.
(v)
One between 2 and
3,
and one between
-
3 and
-
4.
8.
(i) 1-3569,
1-6920.
(ii)
4-1891.
(iii) -4679, 1-6527,
3
8793.
(iv)
2-2317.
(v) 2-1622,
2-4142.
(vi)
1-1487.
9.
(i)
3, 3, -4,
-4.
(ii) 3, 3,
-3=4=^/8.
11.
1, 3,
-g,
^^7.
-
MISCELLANEOUS EXAMPLES.
1.
(i) -(6-c)(c-a)(a-6)(Sa)
2
.
(ii) (ab
-
cd) (ae
-
db) (ad
-
be).
(iii) Sa(2S
2
-26c). (iv) (b
+
c) (c
+
o) (a
+
b)
2a
2
.
B.
{(b c)
2
+
(c a)
2
+
(a fc)
2
}
2
. 7. 3
(b-
c) (c a) (a b).
28.
(i)
-2,
(-Si^S). (ii) a, b,
& &
(iii)
0,
i^/bci^cai^/aft
(all signs being positive
or
only
one
positive).
(iv)
*
=
-tL
= =
.*
.
(v)
0,
0, 0; 1, 1,
1.
v/
6-c c-a a-b
%abc
(vi)
2a6c x
=
(c
2
a
2
+
a
2
6
2
-
6
2
c
2
),
&c.
(vii)
0,0,0; 4, 4, 4; 1, 1,
1.
[Signs
in
ambiguities
must be taken so that
xyz may
be
positive.]
a
+ b+c
ill-
J.) J.}
J.
| I
.
a-b-c b-c~a c-a-b
646 ANSWERS TO THE EXAMPLES.
(ix) 1, -1,
1, 2; 1, -1, 2,
1.
(x)
a(b-c),b(c-a),c(a-b);
-*-
=
-1L
_
z
=
(
b
JL
e
+
c
~
a
+
!LT
b
\f(* +
1
-
*\
a-b b-c c-a
\a-b
b-c
c-aj/\a
b
cj'
(xii)
0, 0,
0; a-fcc, b*-ca,
c?-ab.
31.
p*
+
q*+r
3
-qr-rp-pq
=
0.
32.
13, 156; 14, 84; 15, 60; 16, 48; 18, 36; 20, 30; 21, 28; 21,
24.
36.
49, 225,
1225. 37.
Digits 1, 0,
r- 1.
38.
7,13,17,23;
-|. J, -\. \.
30. l+2+23+4+25+6. 43.
(2-2)/n!
46.
61.
64.
{(
55.
llog.2-.
83.
(i)
D.
(ii)
C.
(iii)
If
a>l,
D. If
a^l,
C.
(iv)
D.
(v)
C.
S"?^!)
2
64.
m<n<m(N/2
+
l).
7O.
_2_1_. 74. 63 and 65.
81.
(a*+6*
+
c
2
)
4
/256abc.
1O4. 2
1 13.
2|> (6
-
c) (c
-
a) (a
-
b) (be
+
ca
+
ab).
CAMBRIDGE: I*KINTBD BV
J.
* c. K.
CLAY,
AT THE UNIVERSITY
PRESS.
PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET
UNIVERSITY OF TORONTO LIBRARY
QA
155
S65
1900
C.I
PASC