You are on page 1of 313

THE BOOK WAS

DRENCHED

a:

DQ

DO

THE THEORY OF EQUATIONS.

VOLUME

I.

DUBLIN UNIVERSITY PRESS SERIES

TIIK

THEORY

EQUATIONS:

01-

INTRODUCTION TO THH THHORY OF BINARY


ALGKBRAIC FORMS.
BY

WILLIAM SNOW BURNSIDK,

M.A., D.Sc.,

LAIE SENIOR I-KLIOW OF TRINI1Y COII.EGK, DUBLIN;


AND ERASMUS SMUH's PROFESSOR OP MATHBMAIICS IN
IHK UNIVBRSI'IY OF DUBLIN;

AND

ARTHUR WILLIAM PANTON,


I'KIIOW

AND 1U1 OR,

M.A., D.Sc.,

IRINI'lY COLLEGE,

DONhOAl I.KCirRI<K

IN

VOL.

EIGHTH EDITION
DUBLIN: HODGES, FIGGIS, &

LONDON: LONGMANS, GREEN

Sc

1924.

DUBLIN;
MA1HEMA1ICS.

I.

(Second hsue).

CO.,

CO.,

NASSAU STREET.
PATERNOSTER ROW.

Printed by PONSOVBV

&

GIBBS, University Press, Dublin.

PREFACE TO SEVENTH EDITION OF VOLUME


HAVING been deprived
thought

it

of Dr.

inadvisable to

the introduction of

new

Pan ton's

to preserve the original design,


this

volume as indicated

co-operation, I have

make any change


matter, as

and

I.

in this

volume by

wished as far as possible


also the limits fixed for

in the Preface to previous editions.

In coining to this decision I have been influenced to a great


degree by the favourable reception already given to this work.

W.
June !//, 1912.

Dr. linrnside died llth March, 1920.


Dr. Panton died 18th December, 1906,

S.

F>.

vi

Preface.

under the

Modern

and

Higher Algebra,

Chapter on

our aim to make this Chapter

It has been

Determinants.
as simple

of

title

intelligible as possible to the beginner;

and

at

the same time to omit no proposition which might be found

For many

useful in the application of this calculus.

Chapter, as well as in other parts of the

this

in

examples

of the

work, we are indebted to the kindness of Mr. Cathcart, Fellow


of Trinity College.

We

have approached the consideration of Covariants and

Invariants
differences

the

of the

simplest

medium

the

through

roots of

the functions

of

the

of

This appears to be

equations.

and most attractive mode

of

the

presenting

subject to beginners, and has the advantage, as will be seen,


of enabling us to express irrational

We

terms of the roots.


to show

how

mode

this

oovariants

rationally

in

have attempted at the same time

of treatment

may

harmony with the more general problem

be brought into

of the linear trans-

formation of algebraic forms.

Of the works which have afforded us

assistance in the

more

elementary part of the subject, we wish to mention particularly


the Traite d'Alyebre of
late Professor

much

to

Young

M.

Bertrand, and the writings of the

of Belfast,

which have contributed so

extend and simplify the analysis and solution of

numerical equations.

In the more advanced portions of the subject we are


indebted

mainly,

among

published

works,

to

the

Lessons

Introductory to the Modern Higher Algebra of Dr. Salmon, and

the Theorie der bindren algebraischen Formen of Clebsch


in

some degree

to

the

Theorie

des

Formes

binaires

and

of the

yii

Preface.

We must record

De Bruno.

Ohev. F. Fad

also our obligations

in this department of the subject to Mr. Michael Roberts,

whose papers in the Quarterly Journal and other

and from whose professorial

lectures

the

in

periodicals,

University of

very great assistance has been derived.

Dublin,

the examples also are taken from Papers set

from

Many

by him

of

at the

University Examinations.

In connexion with various parts of the subject several


other

have

works

mentioned the

been

treatises

consulted,

among which may be

on Algebra by Serret, Meyer Hirsch,

and 'Rubini, and papers in the mathematical journals by Boole,


Cayley, Hermite, and Sylvester.

We
to

have added also in

what was contained

new Chapter on
Our aim has been

this

and the preceding

edition,

in the earlier editions of this work>

the Theory of Substitutions and Groups.


to give here, within as

narrow limits

as

possible, an account of the subject which may be found useful


by students as an introduction to those fuller and more

systematic works which are specially devoted to this depart-

ment

of Algebra.

in the

assistance

Cours

(PAlgebre

The works which have

preparation of this Chapter are

supMeure

Equations algebriques by
Netto's

Algebra

des

und

ihre

Serret's

Substitutions
(Paris,

Anwendung

et

des

1870);

auf

die

which there is an English


1882),
F.
N.
Cole
and
(Ann. Arbor, Mich., 1892)
by

(Leipzig,

Legons sur

Traitt

M. Camille Jordan

Substitutionentheorie

translation

afforded us most

of

la

Resolution algtbrique des Equations,

(Paris, 1895).

COLLEGE, DUBLIN,

May, 1904.

by M. H. Vogt

TABLE OF CONTENTS OF VOL

I.

INTRODUCTION.
A rt

Page

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

1.

Definitions,

2.

Numerical and algebraical equations,

3.

Polynomials,

CHAPTER

I.

GENEKAL PROPERTIES OF POLYNOMIALS.


relating to polynomials when the variable receives large values,
when the variable receives small values,

4.

Theorem

5.

Similar theorem

6.

Change

of

...

......
......
............10
.......

form of a polynomial corresponding


Derived functions,

to

an increase or diminu-

tion of the variable.

Continuity of a rational integral function,


8. Form of the quotient and remainder when a polynomial
binomial,
9. Tabulation of functions,
.
.
.
10. Graphic representation of a polynomial, .
11. Maximum and minimum values of polynomials,
7.

5
6

CHAPTER

is

divided by a
,,

12

.13
.

17

II.

GENERAL PROPERTIES OF EQUATIONS.


.
14. Theorems relating to the real roots of equations,
Existtjnce of a root in the general equation.
Imaginary roots,
.
16. Theorem determining the number of roots of an equation,

12. 13,

15.

17.
18.

Equal roots,
Imaginary roots enter equations in

pairs,

19. DescRrtes' rule of signs for positive roots,

.19
.21
.22

.....
...

25
26
28

Table of Contents.

Art.

30

20. Descartes' rule of signs for negative roots,

of Descartes' rule in proving the existence of imaginary roots,

21.

Use

22.

Theorem

relating to

30

the substitution of two given numbers for the


31

variable,

32

Examples,

CHAPTER

III.

RELATIONS BETWEEN THE ROOTS AND COEFFICIENTS OF EQUATIONS, WITH


APPLICATIONS TO SYMMETRIC FUNCTIONS OF THE ROOTS.
23. Relations between the roots and coefficients.

Theorem,

.35
36

24. Applications of the theorem,


26. Depression of an equation

when

a relation exists between

two of

its roots,
"

42

26.

The cube

roots of unity,

43

27.

Symmetric functions of the roots,

46

48

Examples,
28.

Theorems

relating to symmetric functions,


.

Examples,

CHAPTER

...

53
54

1Y.

TRANSFORMATION OF EQUATIONS.
60

29. Transformation of equations

60

30. Roots with signs changed,

31. Roots multiplied

32. Reciprocal roots

by a given quantity,
and reciprocal equations,

33.

To

84.

Removal of terms,

increase or diminish the roots

......

by a given quantity,

37.

The
The

38.

Homographic transformation

.64

68

73

71

cubic,

76

by symmetric functions,
whose roots are any powers

40. Formation of the equation

proposed equation,
41. Trans foimation in general,
42.

biquadratic,

89. Transformation

76
of the roots of the

.........
......*.

Equation of squared differences of a cubic

43. Criterion of the nature of the roots of a cubic,


44. Equation of differences in general,

Examples,

62

67

35. Binomial coefficients,

36.

61

78

80

.81

84
84

86

xi

Table of Contents.

CHAPTER

V.

SOLUTION OF RECIPROCAL AND BINOMIAL EQUATIONS.


Page
90

Art.

45. Reciprocal equations,

Propositions embracing their leading general

46-52. Binomial equations.

92

properties,

53.

The

54.

Solution of binomial equations by circular functions,

special roots of the equation #>

....

0,

96
98
100

Examples,

CHAPTER

VI.

ALGEBRAIC SOLUTION OP THE CUBIC AND BIQUADRATIC.


55.

On

66.

The

105

the algebraic solution of equations,


algebraic solution of the cubic equation,

108
109

67. Application to numerical equations,


68. Expression of the cubic as the difference of

59. Solution of the cubic

two cubes,

by symmetric functions of the

....

Examples,
60.

Homographic

61. First solution

62.

relation

between two roots of a cubic,

by radicals

roots,

.111
.113

Euler's assumption,

of the biquadratic.

114

120
121

126

Examples,
Second solution by radicals of the biquadratic,

127

63. Resolution of the quartic into its quadratic factors.

Ferrari's solution,

64. Resolution of the quartic into its quadratic factors.

Descartes solution,

133

135

roots,

139

65. Transformation of the biquadratic into the reciprocal form,

by symmetric functions of the

129

66.

Solution of the biquadratic

67.

Equation of squared differences of a biquadratic,

142

68. Criterion of the nature of the roots of a biquadratic,

144

Examples,

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

146

CHAPTER YIL
PROPERTIES OF THE DERIVED FUNCTIONS.
69.

Graphic representation of the derived function,

70.

Theorem

71. Rolle's

relating to the

Theorem.

maxima and minima

154

of a polynomial,

.155
167

Corollary,

72.

Constitution of the derived functions,

73.

Theorem

relating to multiple roots,

167
.

.168
159

74. Determination of multiple roots,


75. 76.

Theorems relating

the equation,

Examples,

to the passage of the variable

through a root of
161, 162

163

Table of Contents.

xii

CHAPTER

VIII.

SYMMETRIC FUNCTIONS OF THE ROOTS.


Art.

77.

Newton's theorem on the sums of powers of

78.

Expression of a rational symmetric function of the roots in terms of the


coefficients.

roots.

I.,

in terms of the coefficients.

sums of powers of

roots

169

Prop. III.,

80. Expression of the coefficients in terms of

sums

powers of roots,
Definitions of order and weight of symmetric functions, and theorem
of

170

173

relating to the former,


82. Calculation of

symmetric functions of the

83.

products,

Homogeneous

Page
166

167

Prop. II.,

79. Further proposition relating to the expression of

81.

Prop.

CHAPTER

roots,

.174

178

IX.

LIMITS OF THE HOOTS OF EQUATIONS.


180

84. Definition of limits,

180

85. Limits of roots.

Prop.

86. Limits of roots.

Prop. II.,

I..

181

87. Practical applications,


88.

Newton's method

89. Inferior limits,

of finding limits.

Prop. III.,

and limits of the negative

....

186

roots,

187

90. Limiting equations,

188

Examples,

CHAPTER
SEPAIIATION OF
91. General explanation,
92.

THE

X.

iiOOTS OF EQUATIONS.

..........

Theorem of Fourier and Bu dan,


.

from the theorem

96. Sturm's theorem,

97. Sturm's theorem.

Equal

imaginary

roots,

of Fourier

and Budan,

197

198

.'

.203

roots,

206

99. Conditions for the reality of the roots of an equation,


100. Conditions for the reality of the roots of
-

.192
194

98. Application of Sturm's theorem,

Examples,

189
189

93. Application of the theorem,


94. Application of the theorem to
95. Corollaries

183

185

a biquadratic,

.210
.211

,212

Table of Contents.

xiii

CHAPTER XL
SOLUTION OF NUMERICAL EQUATIONS.
Art.

Pa^e
215

and numerical equations,


102. Theorem relating to commensurable roots,
101. Algebraical

103. Newton's

104. Application of tbe


106.

216

metbod of divisors,

Method of

method of

limiting tbe

217

218

divisors,

number

of trial-divisors,

221

106. Determination of multiple roots,

222

Newton's method of approximation,


108. Horner's method of solving numerical equations,
109. Principle of the trial-divisor in Homer's method,
107.

....
....

112.

Homer's process,
Application of Homer's method to cases
Lagnwge's method of approximation,

113.

Numerical solution of the biquadratic by Descartes' method,

227
231

235

110. Contraction of
111.

226

Mliere roots are nearly equal,

238
241

242
246

Miscellaneous examples,

CHAPTER

XII.

COMPLEX NUMBERS AND THE COMPLEX VA1UABLE.


114.

115.

Complex numbers
Complex numbers

116. Multiplication

and

Graphic representation,
Addition and subtraction,
division,

117. Other operations on

llS.

The complex

complex numbers,

249

260

.......

'261

....

25?

variable,

119. Continuity of a function of the

.....

complex variable,

252
266

120. Variation of the amplitude of the function corresponding to the descrip-

tion of a small closed curve

by the complex variable,


number of roots comprised within a
.

266

121. Cauchy's theorem relating to the

268

plane area,
122. Proof of the fundamental theorem relating to the

number

of roots of the

269

general equation,
123. Second proof of fundamental theorem,
124. Determination of complex numerical roots

260
:

solution of the cubic,

261

125. Solution of the biquadratic,

266

126. Solution of biquadratic continued,

267

xiv

Table of Contents.

NOTES.

A..

Algebraic solution of equations,

B.

Solution of numerical equations,

C.

Tta

....<
........

proposition that every equation has a root.

Pa*
271

27
278

THEORY OF EQUATIONS.

INTRODUCTION.
1.

Definitions.

We

Any

mathematical expression involving a

called a function, of that quantity.


shall be employed mainly with such algebraical func-

quantity

is

and integral. By a rational function of a


meant one which contains that quantity in a rational
form only that is, a form free from fractional indices or radical
signs.
By an integral innnfriftn of a quantity is meant one in
which the quantity enters in an integral form only that is,
never in the denominator of a fraction. The following exprestions as are raj^oj^l

quantity

is

sion, for

and

example, in which n

is

a positive integer,
x

integral algebraical function of

ax* + bx"~ l

or"-

3
-i-

is

a rational

Jcx

/.

It is to be observed that this definition has reference to the


quantity x only, of which the expression is regarded as a function.
The several coefficients a, 6, c, &o., may be irrational or
fractional, and the function still remain rational and integral
in x.

function of x is represented for brevity by F(x), f(x) r (x),


or some such symbol.
The name polynomial is given to the algebraical function
to express the fact that it is constituted of * number of
<j>

Introduction.

containing different powers of x connected by the signs plus or


minus. For certain values of x regarded as variable one poly-

may become

equal to another differently constituted.


The algebraical expression of such a relation is called an equation; and any value of x which satisfies this equation is called a

nomial

The determination

root of the equation.

of all possible roots

constitutes the complete solution of the equation.

by bringing all the terms to one side, we


any
equation according to descending powers of x
may arrange
in the following manner
It

is

obvious that,

-f

w ~l

#i#

fl^tf"'

<in-\x

+ an

0.

The highest power of x in this equation being n, it is said- to


be an equation of the nth degree in x. For such an equation TVO
The suffix
ghall, in general, employ the form here written.
attached to the letter a indicates the power of x which each coefficient accompanies, the sum of the exponent of x and the suffix

An equation is not altered


of a being equal to n for each term.
if all its terms be divided by any quantity.
may thus, if
we please, dividing by Oo, make the coefficient of #* in the above

We

equation equal to unity.

make

this supposition

It will often be found convenient to

and in such

cases the equation will be

written in the form

*n + ptX*- + ptX"-* +
1

An

equation

is

+P*-\X

said to be complete

+pn *

when

it

0.

contains terms

involving x in all its powers from n to 0, and incomplete when


some of the terms are absent ; or, in other words, when some of
the coefficients p t9 p 2 &c., are equal to zero.
The term pn ,
which does not contain #, is called the absolute term. An equa,

tion is numerical or algebraical according as its coefficients are

number*, or algebraical symbols.

Vamerical and Algebraical Equations.

In many
and physical science the final
mathematical problem presents itself in the form of an equation
on whoe solution that of the problem depends. It is natural,
2.

researches in both mathematical

Numerical and Algebraical Equations.

therefore, that the attention of mathematicians should have

been

at an early stage in the history of the science directed towards


The science of the Theory of Equainquiries of this nature.
tions, as it now stands, has grown out of the successive attempts
of mathematicians to discover general methods for the solution
of equations of any degree. When the coefficients of an equation

are given numbers, the problem is to determine a numerical


value, or perhaps several different numerical values, which will
satisfy the equation.

progress has been

In

branch of the science very great


the best methods hitherto advanced

this

made and
;

for the discovery, either exactly or approximately, of the numerical values of the roots will be explained in their proper places

in this work.

Equal progress has not been made in the general solution of


The stuequations whose coefficients are algebraical symbols.
dent is aware that the root of an equation of the second degree,
whose coefficients are such symbols, may be expressed in terms
of these coefficients in a general formula and that the nume;

any particular numerical equation may be obtained


by substituting in this formula the particular numbers for the
symbols. It was natural to inquire whether it was possible to

rical roots of

any such formula for the solution of equations of higher


Such results have been attained in the case of equadegrees.
It will be shown that
tions of the third and fourth degrees.
discover

in certain cases these formulas fail to supply the solution of


substitution of the numerical coef-

a numerical equation by
ficients

for the general

symbols, and are, therefore, in this

respect inferior to the corresponding algebraical

solution of

the quadratic.

attempts have been


formulas for equations of the

Many

may now be regarded


analysts that

it is

made
fifth

as established

to arrive at similar general

and higher degrees

by

but

it

the researches of modern

not possible by means of radical signs, and

other signs of operation employed in common algebra, to ex'


press the root of an equation of the fifth or any higher degree
in terms of the coefficients.

Introduction.
3.

Polynomials.

From

the preceding observations

it is

plain that one important object of the science of the Theory of


Equations is the discovery of those values of the quantity x

regarded as variable which give to the polynomial /(a?) the


In attempting to discover such values of
particular value zero.
x we shall be led into many inquiries concerning the values

assumed by the polynomial for other values of the variable.

We shall, in fact,
a

see in the next chapter that, corresponding to


continuous series of values of x varying from an infinitely

great negative quantity (- oo ) to an infinitely great positive


quantity (+ oo ),/(#) will assume also values continuously varyThe study of such variations is a very important part of
ing.

the theory of polynomials.


The general solution of numerical
equations is, in fact, a tentative process ; and by examining the
values assumed

by the polynomial for certain

arbitrarily

assumed

values of the variable, we shall be led, if not to the root itself,


at least to an indication of the neighbourhood in which it exists,

and within which our further approximation must be

A polynomial is sometimes called a quantic.


to have distinct

degrees.

names

The terms

carried on.

It is convenient

for the quantics of various successive

quacfrgtic (or guadric), cujjic, biquadratic (or

are used to represent quantics of the


2nd, 3rd, 4th, 6th, oth, &c., degrees ; and the equations obtained
by equating these quantics to zero are called quadratic, cubic*.
quartic),

gum tic,

sextic, &c.,

tiquadratief &c., equations, respectively.

CHAPTER

I.

GENERAL PROPERTIES OF POLYNOMIALS.


IN tracing the changes

4.

of value of a polynomial correspond-

ing to changes in the variable, we shall


in the polynomial axe most important

first

inquire what terms


values very great

when

very small are assigned to

x.
This inquiry will form the
subject of the present and succeeding Articles.
Writing the polynomial in the form

or*

^
-- + ---+...+
_L

it is plain that its value tends to become


equal to a&P as x tends
towards QO
The following theorem will determine a quantity
such that the substitution of this, or of any greater quantity*
for x will have the effect of making the term a x" exceed the
.

sum

In what follows we suppose

of all the others.

to be

and

in general in the treatment of polynomials and


positive ;
the
equations
highest term is supposed to be written with the
positive sign.

Theorem.

the value

If in

1, or

a$
is

the polynomial

any greater

that one of the coefficients

a,,

value, be substituted for x, where

ak

whose numerical value

is

a,

a,,

greatest, irrespective of sign, the term containing the highest power


ofx will exceed the sum of all the terms which follow.

The

inequality

>

x*~ l

4-

a, of*-*

+ a^j x +

a*

General Properties of Polynomials.

by any value

is satisfied

of

x which makes

n ~l

a xn > a k (x

where #*

is

the greatest

without regard to sign.


the brackets, we have

<*
which
that

> a"

is satisfied if

-f

#""2

+x+

an ^, an
among the coefficients a #2
the
series
withip
Summing
geometric
ly

frp

or *" >

"

M^T) (x

a (x - 1) be > or = a k

x > or =

is

1),

1} '

1.

#0

The theorem here proved

is

useful in supplying,

when the

polynomial are given numbers, a number such


x receives values nearer to 4- oo the polynomial will

coefficients of the

that

when

If we change the sign of #,


preserve constantly a positive sign.
the first term will retain its sign if n be even, and will become

negative if n be odd ; so that the theorem also supplies a negative value of x, such that for any value nearer to - oo the

polynomial will retain constantly a positive sign if n be even,


and a negative sign if n be odd. The constitution of the poly-

nomial is, in general, such that limits much nearer to zero than
those here arrived at can be found beyond which the function
preserves the same sign ; for in the above proof we have taken
the most unfavourable case, viz. that in which all the coefficients

except the first are negative, and each equal to a& ; whereas in
general the coefficients may be positive, negative, or zero.
Several theorems, having for their object the discovery of such
closer limits, will be given in a subsequent chapter.
5. We now proceed to inquire what is the most important
term in a polynomial when the value of x is indefinitely diminished and to determine a quantity such that the substitution
of this, or of any smaller quantity, for x will have the effect of
;

giving such term the preponderance.

Theorem.

If in

the polynomial

Theorem.
-

the value

or any smaller value, be substituted for

where a*

a?,

the greatest coefficient exclusive of an , the term an will be

is

rically greater than the

sum of all
x =-

a*

the others.

To prove this, let


then by the theorem cf
being now the greatest among the coefficients a*, a^
;

without regard to sign, the value


an
y, will

*
^
that

nume-

1,

i-

or

Art. 4,
.

any greater value of

make

fl

is,

henee the value

> <V! - +
y

This proposition
follows:

is

or

f/ n _ 2

. .

any

less

tfo

-7.

0"

//

value

of,

x y will

jy

often stated in a different manner, as

Values so small

may

be assigned to

x as

to

make

the

polynomial

less

than any assigned quantity.


This statement of the theorem follows at once from the above

proof, since an may be taken to be the assigned quantity.


There is also another useful statement of the theorem, as

follows

When

the variable

receives a very small value, the sign

of the polynomial

is

the

same as the sign of its first term

This appears by writing the expression in the form

for

when a value

sufficiently small is

value of the term a n ^ exceeds the

sum

given to

^,

the numerical

of the other terms of the

expression within the brackets, and the sign of that expression


will consequently depend on the sign of

General Properties of Polynomials.

6. Change of Form of a Polynomial corresponding


an increase or diminution of the Variable. Derived
Functions. We shall now examine the form assumed by the
polynomial when x + h is substituted for x. If, in what follows,

to

h be supposed essentially positive, the resulting form will corre-

spond to an increase of the variable and the form corresponding


x will be obtained from this by changing the
;

to a diminution of

sign of h in the result.

When x is changed
+

a\ (x

to

h)""

x + h f(x) becomes /(a? +


y

+ # 2 (# +

n ~*

A)

. .

A), or

a*.i (x

h)

+ an

Let each term of this expression be expanded by the binomial


theorem, and the result arranged according to ascending powers
of h.
We then have

+
-f

(n

. .

1)

a^-*

-h

in.

n-1

...

+ #.2# +

(n

2) o

It will be observed that the part of this expression indepen(x) (a result obvious d priori), and that the succes-

dent of h is/

powers of h are functions of x of


degrees diminishing by unity. It will be further observed that
the coefficient of h may be derived from f(x) in the following
sive coefficients of the different

manner:

Let each term in f(x] be multiplied by the exponent


and let the exponent of x in the term be
diminished by unity, the sign being retained the sum of all
of x in that term,

the terms of f(x) treated in this

way will constitute a polynomial

This polyof dimensions one degree lower than those oif(x).


It is usual to
is called the first derived function oif(x).

nomial

represent this function

by the notation

(x).

The

coefficient

Derived Functions.
of

5 may be derived from/

by a

(x)

process the

same as that

(x) from /(#), or by the operation twice


employed in deriving
performed on/ (x). This coefficient is represented by/" (a?), and
is called the second derived function of /(#).
In like manner the

succeeding coefficients
tions of this character

indicated,

we may

may
;

be derived by successive opera*


employing the notation here

all

so that,

write the result as follows


f"(a\

f(* +
It

h)

may

-/(*)

+f (x) h +'L-

be observed that, since the interchange of x and h


^), the expansion may also be written in the

does not alter /(^ +


fortn

/(* +

K)

We

shall in general employ the notation here explained ;


but on certain occasions when it is necessary to deal with derived

functions beyond the

two or

first

be found more

three, it will

convenient to use suffixes instead of the accents here employed.


The expansion will then be written as follows :

hr

A2

f(* +

A) -/(*) +fi(x)h +/,(#)

y-H +

. . .

l.z

+/r()

j
1

<

o
o

~
. . .

'

EXAMPLE.
Find the result of substituting x
Here

+h

for

x
1

in the polynomial

7#

42s + 6s2

-7x + 4.

-h 4,

'-7,
>

and the remit

24

12,

is

1.2
The student may

verify this result

hy

4-

24

...

direct substitution.

Continuity of a Rational Integral Function of#.


If in a rational and integral function /(a?) the value of x be
7.

10

General Properties of Polynomials.

made

to vary, by indefinitely small increments, from one quana


to a greater quantity b, we proceed to prove that/(#) at
tity
the same time varies also by indefinitely small increments in
;

other words, that/ (a?) varies continuously with x.


Let x be increased from a to a + h. The corresponding incre-

ment

and

of f(x]

is

this is equal,

by Art.

G,

to

in which expression all the coefficients/' (#),/" (#), &c., are finite
Now, by the theorem of Art. 5, this latter expresquantities.

may, by taking h small enough, be made to assume a value


less than any assigned quantity
so that the difference between
f(a + h] and f(a) may be made as small as we please, and will
The same is true during all stages of
ultimately vanish with h.
the variation of x from a to b ; thus the continuity of the funcsion

tion/^)

is

established.

to be observed that

it is not here proved that / (x)


from /(a) to/ (b). It may either increase
or diminish, or at one time increase, and at another diminish
but the above proof shows that it cannot pass per saltum from

It

is

increases continuously

and

consequently, amongst the


values assumed by/(#) while x increases continuously from a to
The sign
b must be included all values between f(a) and /(>).

one value to another

that,

otf'(a) will determine whether/ (x) is increasing or diminishing


for it appears by Art. 5 that when h is small enough the sign of
"We thus
the total increment will depend on that of /'() h.
;

observe that when f(a)

when

8.

(a) is

is positive

negative f(x]

Form of

Polynomial

is

is

is

f(x)

increasing with x

diminishing as

the 4tuotit?nt and Remainder


divided by a Binomial. Let the
n ~l

n ~*
a<tx

-*

and

when

a^ix

an

quotient,
*

when
ai%

increases.

Form of
is

Quotient

and Remainder.

11

divided by x - h, be

This

we shall represent by

and the remainder by R.

Q,

have then the following equation

We

- h]
/(*) - (x

tf.

The meaning of this equation is, that when Q is multiplied


by # - A, and jR added, the result must be identical, term for term,
with/ (x). In order to distinguish equations of the kind here
explained from equations which are not identities, it will often
be found convenient to use the symbol here employed in place
of the usual

identity

The right-hand

symbol of equality.

side of the

is

Equating the coefficients of x on both sides, we get the followiug series of equations to determine &<>, b l9 Aa . 6n_i, R
,

&0
bi

0,

bji

di 9

by

bji

rt a

bs

bji

-f

^3,

These equations supply a ready method of calculating in


succession the coefficients J
#i> &c. of the quotient, and the
,

For this purpose we write the


the following manner

remainder R.
in

series of operations

In the

first line

are written

down

the successive coefficients

General Properties of Polynomials.

12

The

term in the second line is obtained by multiis equal to it) by A.


which
The product b h is
&o,
to obtain the
order
placed under i, and then added to it in
term 61 in the third line. This term, when obtained, is multiThe product is
plied in its turn by A, and placed under 8
added to a9 to obtain the second figure bt in the third line. The
of f(x).

first

plying a (or

the coefrepetition of this process furnishes in succession all


ficients of the quotient, the last figure thus obtained being the
few examples will make this plain.
remainder.

EXAMPLES.
1

Find the quotient and remainder when 3*4

by x

5**

10** +

Ix - 61

U divided

3.

The

calculation is arranged as follows

3-6

10

11

-61.

12

66

231.

77

170.

22

Thus the quotient


-f 77,
+
+
2. Find the quotient and remainder when
is 3s8

4*a

and the remainder 170.

22#

2
3# + 2 is divided by *
-f 6#
22-11.
Ans. Q
a? + Qx + 9,

a?

-f-

1.

3. Find Q and JK when x6 - 4#* + 7s* - lla? - 13 is divided by x - 5.


N.B. When any term in a polynomial is ahsent, care must be taken to supply
the place of its coefficient by zero in writing down the coefficients of /(*). In this

exai0}le, therefore, the series in the first line will he

1-470
Ans.

4.

-11

-13.

# = #* + a^+12^ 2 -f 60* + 289; J?1432.

Q and R when x9 -- 3s7 - lo#2 -f 2 is divided by x - 2.


Q = a* 4- 22* + la? + Ux* + 28* 4 + 56^ + 112s2 + 209* + 418; 22 = 838.
2
Find Q and R when tf6
a? - lOa; + 113 is divided
hy x 4- 4.
= -855.
Ans. Q = x*-4z*+ 16* 2 -63a? + 242;
Find

^ifc.
5.

9.

-I-

Tabulation of Functions.

The

operation explained

in the preceding Article affords a convenient practical method


of calculating the numerical value of a polynomial whose coef-

given numbers when any number


For, the equation
- h) Q -f R,
/(*) - (x

ficients are

since its

is

substituted for *.

two members are identically equal, must be

satisfied

13

Tabulation of Functions.

when any quantity whatever is substituted for


- h being = 0, and Q remaining
JB, #

x.

Let x

A,

Hence
the remainder when

then /(A)

finite.

the result of substituting h for x in f(x) is


- A, and can be calculated rapidly
f(x] is divided by x
process of the last Article.

by the

For example, the result of substituting 3 for x in the polynomial of Ex. 1, Art. 8, viz.,
3^4

_ 5^3 + 10^ + ii x _ 61,

170, this being the remainder after division by x


student can verify this* by actual substitution.
- 4 for x in
Again, the result of substituting

is

The

3.

* + 3*- 10^ + 113


is

- 855, as
appears from Ex.

5,

We saw in Art.

Art. 8.

as x receives a continuous series of values increasing

7 that

from -

oo to

* oo, /(a?) will pass through a corresponding continuous series.


If we substitute in succession for x, in a polynomial whose coefficients are

given numbers, a series of numbers such as

...-5,-4,-3,-2,-l,

0,

1,

2,

3,

4,

5,...,

and

calculate the corresponding values of /(#), the process


be called the tabulation of the function.

may

EXAMPLES.
1.

Tabulate the trinomial 2#2

+x

6, for

-4, -3, -2, -1,

0,

the following values of


1,

2,

3,

4.

a Polynomial.

In
on
any
investigating the changes of a f unotion/(a?) consequent
10.

Graphic Representation

of*

14

General Properties of Polynomials.

changes in the variable which it contains, it is plain


that great advantage will be derived from any mode of representation which renders possible a rapid comparison with one
.another of the different values which the function may assume.
series of

In the case where the function in question is a polynomial with


coefficients, to any assumed value of x will correspond

numerical

We

one definite value of /(#).

proceed to explain a

mode

of

graphic representation by which it is possible to exhibit to the


eye the several values of f(x] corresponding to the different
values of x.

Let two right


(fig.

1)

,angles,

lines

OX,

OY

cut one another at right

and be produced

nitely in both directions.


lines are called the axis ofx

indefi-

These
and axis

Lines, such as
measured on the axis of x at

of y, respectively.

OA,

"x

the right-hand side of O, are re*


garded as positive ; and those, such

OA', measured at the left-hand


Lines parallel
side, as negative.

.as

to

OF which are above

.and those below

it,

Pig.

XX', such as
such as
Tor

l.

AP or I?Q',

AP,

are positive;
These
are negative.

conventions are already familiar to the student acquainted with

Trigonometry.

Any arbitrary length may now be taken on OX as unity,


and any number positive or negative will be represented by a
line measured on XX'
the series of numbers increasing from
-to -f oo in the direction OX, and
diminishing from to oo in the
direction OX'.
Let any number m be represented by OA calculate f(m)
from A draw AP parallel to OF to represent f(m)
in magnitude on the same scale as that on which OA represents
m, and to represent by its position above or below the line OX
;

the sign of /(m).


.represented

by OA,

Corresponding to the different values of m


OB, OC, &c., we shall have a series cf points

JP, Q, It, &c. f which,

when we suppose the

series of values of

15

Graphic Representation.

indefinitely increased so as to include all numbers between


and 4- oo , will trace out a continuous curved line. This

oo

curve will, by the distances of its several points from the line
OX> exhibit to the eye the several values of the function /(a?).

The process here explained is also called tracing the function


f(x). The student acquainted with analytic geometry will observe
it is equivalent to tracing the plane curve whose equation
=
is y
/(#).
In the practical application of this method it is well to begin
by laying down the points on the curve corresponding to certain

that

small integral values of x positive and negative. It will then


general be possible to draw through these points a curve
9

in

which

will exhibit the progress of the function, and give a general


its character.
The accuracy of the representation will

idea of

number of points determined between


values
of
two
variable.
the
When any portion of the
any
given
curve between two proposed limits has to be examined with care,
of course increase with the

be necessary to substitute values of the variable


The following exseparated by smaller intervals than unity.

it

will often

amples will

illustrate these principles.

EXAMPLES.
1.

Trace the trinomial 2s*

The
the line

unit of length taken

OD in

In Ex.

1,

is

x -

6.

one-sixth of

fig. 2.

Art. 9, the values of f(x) are

given corresponding to the integral values


4 to + 4, inclusive.
of x from

By means

of these values

we

obtain

the positions of nine points on the curve;


seven of which, A, B, C, D, E, F, G, are

here represented, the other two corresponding to values of f(x) which lie out of the
limits of the figure.

The

it a useful exercise
more minutely between
and E in the figure, viz. by

student will find

to trace the curve

the points

calculating the values of f(x) corresponding

to all values of x between

$M

is

and

done in the following example.

separarod by

Fig. 2.

small intervals, say of one-tenth,.

General Properties of Polynomials.

16

Trace tho polynomial

2.

6.

17**

- 4 and 4.
already tabulated in Art. 9 for values of x between
Art.
that
this
function
retains positive
on
an
exercise
as
be
4,
observed,
may
values for all positive values of x greater
2*7, and negative values for all
This

is

It

values of x nearer to

curve

<xf

than

^a

The

2-7.

will, then, if it cuts the axis of

x at

all,

a point (or points) corresponding to


some value (or values) of x between - 2*7 and
cut

it at

2*7

so that if our object is to determine, or

approximate

to,

the positions of the roots of the

= 0, the tabulation may be conequation f(x)


2*7 and 2*7.
fined to the interval between
This

is

a case in which the substitution of

only of x gives very little help


towards the tracing of the curve, and where,
consequently, smaller intervals have to be ex-

amined.

We

X'

give the tabulation of the func-

tion for intervals of one -tenth between the

integers

I,

0,

1;1,2.

From these values

the positions of the corresponding points on


the curve may be approximately ascertained,
and the curve traced as in fig. 3.

Values of x

-22

-9

-8 -7

Y'
Fig. 8.

--6 !--

-15-96 -10-8 -6-46 -2-88|

-1

-4

2-24

5-04

3-9

5-72
**

5-04

4-32

3-5

1-3

1-4

1-6

54

(a

1-62

2-64

1-8

1-6

1-7

5-04

7-7

42

1-04

1-8

1-9

11 -04 15-12

20

The curve traced in Ex. 1 cuts the axis of x in two points


number equal to the degree of the polynomial) in other
:

words, there are two values of x for which the value of the given
2
polynomial is zero ; these are the roots of the equation 2# + x

-6

2, and 1*5.
Similarly, the curve traced in Ex. 2
outs the axis in three points, viz. the points corresponding^ to the
0, viz.

roots of the cubic equation 10#*

- 17x* + x + 6

0.

The

curve

Maxima and Minima.

17

representing a given polynomial may not cut the axis of $ at


all, or may out it in a number of points less tlian the degree of
the polynomial.
Such oases correspond to the imaginary roots
of equations, as will appear

more

For

fully in the next chapter.

example, the curve which represents the polynomial 2tf + x + 2


will, when traced, lie entirely above the axis of x \ in fact, since
this function differs

from the function of Ex. 1 only by the ad-

dition of the constant quantity 8, each value of f(x) is obtained


by adding 8 to the previously calculated value, and the entire

curve can be obtained

by simply supposing the previously traced

moved up

parallel to the axis of y through a distance


equal to 8 of the units. It is evident, by the solution of the

curve to be

8
0, that the two values of x which render
equation 2# + x + 2
the polynomial zero are in this case imaginary. Whenever the

of points in which the curve cuts the axis of x falls


short of the degree of the polynomial, it is customary to speak
of the curve as cutting the line in imaginary points.

number

11.

It

is

Maximum and Minimum Values of Polynomials.


apparent from the considerations established in the pre-

ceding Articles, that as the variable x changes from

oo to

+ oo

the function f(x) may undergo many variations. It may go


on for a certain period increasing, and then, ceasing to increase,

may commence

to diminish

it

then cease to diminish and

may

which another period of

commence again

to increase

diminution

arrive, or the function

example

may

after

of the preceding Art.)

may

(as in the last

go on then continually

in-

At a

stage where the function ceases to increase


creasing.
and commences to diminish, it is said to have attained a
maximum value ; and when it ceases to diminish and commences to increase, it is said to have attained a minimum value.

polynomial

may

have

several

such values

the

number

depending in general on the degree of the function. Nothing


exhibits BO well as a graphic representation the occurrence
of

such a maximjun or

minimum

value

as

well

as

the

various fluctuations of which the values of a polynomial are


susceptible.

18

General Properties of Polynomials.

A knowledge of the maximum and minimum values of a function, giving the positions of the points

where the curve bends

often of great assistance in tracing


a given polynomial. It will be
to
the curve corresponding
shown in a subsequent chapter that the determination of these

with reference to the axis,

is

points depends on the solution of an equation one degree lower


than that of the given function.

easy to show that maxima and minima occur alternately for, as the variable increases from a value corresponding to one maximum to the value corresponding to a second,
the function begins by diminishing and ends by increasing,
It

is
;

and therefore attains a minimum at some intermediate stage.


like manner it appears that between two minima one maximum must exist.

In

CHAPTER

II.

GENERAL PROPERTIES OF EQUATIONS.


12.

THE process of tracing the


may be employed for the

function /(#)

explained in

Art. 10

purpose of ascertaining apthe


of
a
real roots
proximately
given numerical equation ; for
when the corresponding curve is accurately traced, the real roots
the equation f(x) can be obtained approximately by
measuring the distances from the origin of its points of intersection with the axis.
With a view to the more accurate numef

rical solution of this

problem, as well as the general discussion of

equations both numerical and algebraical, we proceed to establish


in the present chapter the most important general properties of
equations having reference to the existence and number of the
roots, and the distinction between real and imaginary roots.
By the aid of the following theorem the existence of a real
root in an equation

may

often be established

Theorem.

If two real quantities a and b be substituted for


the unknown quantity x in any polynomial f (x) and if they furnish
results having different signs, one plus and the other minus ; then
must have at least one real root intermediate
the equation f(x) =
in value between a and b.
This theorem is an immediate consequence of the property
of the continuity of the function/(#) established in Art. 7 for
since f(x) changes continuously from f(a] to/ (6), and therefore
passes through all the intermediate values, while x changes from
a to b and since one of these quantities, /(a) or /(), is positive,
and the other negative, it follows that for some value of x inter,

mediate between a and b f(x) must attain the value zero which
intermediate between f(a) and/ (6).
y

is

20

General Properties of Equations.

The student will assist Ms conception of this theorem by


reference to the graphic method of representation.
What ia
here proved, and what will appear obvious from the figure, is,
there exist two points of the curved line representing tha
polynomial on opposite sides of the axis OX, then the curve
that

if

It will
joining these points must jout that axis at least once.
from the figure that several values may exist
between a and b for which / (x)
0, t. e. for which the curve
also be evident

For example, in fig. 3, Art. 10, x - - 2 gives a


- 2
gives a positive value (20) r
negative value (- 144), and x
and between these points of the curve there exist three points of
cuts the axis.

section of the axis of x.

Corollary. If there exist no real quantity which, substituted


x
makes f(x) = 0, thenf(x] must be positive for every real value
for
9

of x.

For it is evident (Art. 4) that a? - oo makes / (x) positive


and no value of #, therefore, can make it negative for if there
were any such value, the equation would by the theorem of
this Article have a real root, which is contrary to our present
;.

With reference to the graphic mode of representahypothesis.


tion this theorem may be expressed by saying that when the
=

has no real root, the curve representing the


must
lie entirely above the axis of x.
polynomial/*^)
13. Theorem.
Every equation of an odd degree has at least
equation f(x)

me

real root of a sign opposite to that of

This
Article.

an immediate consequence of the theorem in the last


Substitute in succession - oo , 0, oo for a? in the poly-

for x

x =

x=
is
-i

term.

is

nomial/^).

If a n

its last

The
0,

-I-

GO

results are, n being

,/(#)

is

negative

sign of f(x)
oo

,/(#)

is

odd

(see Art. 4),

the same as that of a n

is positive.

positive, the equation must


A . .
Mrtrt i
An n f,' TO rAA f .
,

have a real root between:


an d

if

41

jg

negative, the>

21

Imaginary Roots.
equation must have a real root between
14.

term

and oo, i.e. a real


therefore proved.
Every equation of an even degree^ whose last

The theorem

positive root.

Theorem.

is

negative, has at least two real roots, one positive

is

and

the

other negative.

The

results of substituting

oo

- 00,

+,

0,

-,

+ oo,

hence there

is

0, oo

are in this case

+;

a real root between - oo and 0, and another beoo ; i. e. there exist at least one real negative, and

and
tween
one real positive root.
We have contented ourselves in both
-i-

this and the preceding


existence
the
of
with
Articles
roots, and for this purpose
proving

very large positive or negative values,


It is of course possible to narrow the
done, for x.
limits within which the roots lie by the aid of the theorem of
it is

as

sufficient to substitute

we have

Art. 4, and still more by the aid of the theorems respecting


the limits of the roots to be given in a subsequent chapter.
15.

Existence of a

Imaginary Roots.
real root

in

the

case

Roo%in

We
of

the General Equation^

have now proved the existence of a


every

equation except one of an even


degree whose last term is positive.

Such an equation may have no


It is necessary
real root at all.
then to examine whether, in the
absence of real values, there may

not be values involving the, ima-

ginary expression */-

when

1,

which,

substituted for x, reduce the

polynomial to zero or whether


there may not be in certain cases
;

both -real and imaginary values


of the variable which satisfy th?emiation.

Fig. 4.

We

take a simple

General Properties of^Equations.

22

illustrate the occurrence of

example to

As

such imaginary roots.

already remarked (Art. 10), the curve corresponding to the

polynomial

above the axis of

lies entirely

has no real roots

f(x)

2^ + x

/(*) -

but

x,

it

+ 2

as in

fig. 4.

The equation

has the two imaginary roots

We

observe,
by the solution of the quadratic.
are in
there
that
in
of
values
the
absence
real
therefore,
any
this case two imaginary expressions which reduce the polynomial
as is evident

to zero.

The corresponding general

proposition
integral equation has a root of the form

a +
a and

that Every rational

v/=~l,

being real finite quantities.

/3

is,

This statement includes

both real and imaginary roots, the former corresponding to the


value j3 = 0. When a and
are numbers, such an expression
/3

called a complex

number ; and what

is asserted is that every


numerical equation has a numerical root either real or complex.

is

As

the proof of this proposition involves principles which


not
could
conveniently have been introduced hitherto, and
which will present themselves more naturally for discussion
in subsequent parts of the work, we defer the demonstration
For the present,
until these principles have been established.
therefore,

we assume

certain consequences
16.

Theorem.

the proposition, and proceed to derive

from

it.

Every equation of n dimensions has n

roots,

and no more.

We first

any quantity A is a root of the equaby x - h without a remainder.


evident from Art. 9; for if /(A) = 0, i.e. if h isa root

tion/^) This

is

nf ffa\

observe that

0,

then f(x]

72 miiflf.

if

is divisible

h =

23

Imaginary Roots.
Let, now, the given equation be
#* + j^tf"" 1

f(x)

-i-

p2 #"-* +

+ JE>_, a? + /> =

0.

This equation must have a root, real or imaginary (Art. 15),


which we shall denote by the symbol a
Let the quotient,
when f(x) is divided by x - ai, be fa (x) we have then the
t.

identical equation

/(*)-(*-i)$i(*).
- 1 dimensions,
Again, the equation ^i (x) = 0, which is of n
must have a root, which we represent by o a Let the quotient
obtained by dividing $\(x) by x - a 2 be fa(x). Hence
.

i>i

and

.-.

(#)

/(a?)

(a?

(x

- o s)
0a(),
- en) (a? - a a

0,(#)

where 2 (#) is of n - 2 dimensions.


Proceeding in this manner, we prove that f(x)
tf>

consists of the

product of n factors, each containing x in the first degree, and a


numerical factor $ n (x). Comparing the coefficients of #*, it is
plain that

f(x)

$ n (x)
(a?

1.
01) (x

Thus we prove the identical equation


- a a) (x - a 3) ..... (x- an-i)(#- a).

any one of the quantix\n the right-hand member of this equation


member to zero, and will therefore reduce f(x)

It is evident that the substitution of


ties ai,

n2 >

a n for

will reduce that


to zero

that

quantities

ai,

is to

aa

say, the equation /(a?)

03

-i>

n-

And

it

has for roots the n

can have no other roots

any quantity other than one of the quantities m, <* 2 . a*


be substituted in the right-hand member of the above equation,
the factors will be all different from zero, and therefore the profor if

duct cannot vanish.


each of the n th degree in x
cannot be equal to one another for more than n values of x without

Corollary.

Two polynomials

being completely identical.

difference be equated to zero, we obtain an equath


n
the
tion of
degree, which can be satisfied by n values only
of Xj unless each coefficient be separately equal to zero.

For if their

24

General Properties of Equations.

The theorem

solution of the equation /(#) -

the converse problem,

any n given

no assistance in the

of this Article, although of

i.e.

us to solve completely

0, enables

to find the equation

whose roots are

The

required equation is obtained by


n
the
simple factors formed by subtractmultiplying together
ing from x each of the given roots. By the aid of the present
quantities.

when any (one or more) of the roots of a given


are
known, the equation containing the remaining
equation
For this purpose it is only necessary
obtained.
roots may be
theorem

also,

to divide the given equation

mial

The

factors.

of the given binobe the required polynomial

by the product

quotient will

composed of the remaining

factors.

EXAMPLES.
1.

Find the equation whose

roots are

-1,

-3,

6.

4,

An*. **
2.

The equation
a*

has a root 5

-6** + 8**-

17a?-r 10

6*8 - 13s*

60

at

and

& - x* + 8* - 2 = 0.

16*8

+ 86* - 178* +

106

0,

7.

Ana. The other two roots are

Form

0.

division of Art. 8.

Solve the equation

roots heing 1

4.

=0

An9.

two

63*

find the equation containing the remaining roots.

Use the method of


3.

-I-

3, 6.

tke equation whose roots are

3
*'

~2'

f
Ant. 14** - 28** - 80* 4 9 m

6.

Solve the cubic equation

Here

it

is

evident that *

resulting quadratic.

6.

Form an

=* 1 satisfies

The two

0.

the equation.

roots are found to

Divide by *

1,

and solve tht

be

equation with rational coefficients which shall have for a nyrt the

irrational expression

Equal Roots.
This expression has four different values according to the different combinations
of the radical signs, viz.

The
(x

required equation

- </p - </g)

is,

therefore,

(*

or
2

(x

-p

or, finally,

must be observed that the n factors


of which a polynomial f(x) consists need not be all different
from one another. The factor x - a, for example, may occur in
17.

Equal Roots.

It

the second, or any higher power not superior to n. In this case


the equation /(a?) is still said to have n roots, two or more

being

now equal

to one another

tiple root of the equation

and the root a

is

called a

mul-

double, triple, &c., according to the

number

of times the factor is repeated.


reference to the graphic construction in Art. 10
will help to explain the occurrence of multiple roots.

(fig. 3)

We see

by an inspection of the figure that the two positive roots of the


8 2
17a? -f x + 6 =
and we may
are nearly
equation 10#

equal,
conceive that a slight addition to the absolute term of this polynomial, which is, as already explained, equivalent to a small
parallel movement upwards of the whole curve, would have the
effect of rendering equal the roots of the equation thus altered.

In that case the


distinct points,

curve

it is

line

OX would no longer

but would

touch

it.

out the curve in two

Now, when a

line touches a

properly said to meet the curve, not once, but in two


The student acquainted with the theory of

coincident points.

plane curves will have no difficulty in illustrating in a similar


manner the occurrence of a triple or higher multiple root.

form the connecting link between real and


We have just seen that a small change in the
imaginary
form of a polynomial may convert it from one having real roots
into another in which two of the real roots become equal.
A
further small change may convert it into a form in whioh the

Equal

roots

roots.

26

(General Properties of Equations.

Let us suppose that the above


by another small addition to the
then have a graphic representation in

two roots become imaginary.


polynomial

is

further altered

absolute term.

We shall
OX cuts the

curve in only one real point, viz.


that corresponding to the negative root, the two points of
section corresponding to the two positive roots having now dis-

which the axis

appeared.
3 17x* + x + 28,
Consider, for example, the polynomial 10^
which is obtained from that of Ex. 2, Art. 10, by the addition

The student can

of 22.

corresponding to
Divide by x
of x.

easily construct the figure ; the point


3 will now lie much above the axis
fig.

in

and obtain the trinomial l(Xr2 - 27o? + 28


which contains the remaining two roots. They are easily found
+

1,

to be

,_

391

We observe in this

27_ y/39l /-y


20
20^

example of Art. 15,


that when a change of form of the polynomial causes one real
root to disappear, a second also disappears at the same time, and
case, as well as in the

The reason
the two are replaced by a pair of imaginary roots.
of this will be apparent from the proposition of the following
Article.

Imaginary Roots enter Initiations in Pairs.


proposition to be now proved may be stated as follows

18.

The

If an equation f(x) = 0, whose coefficients are all real quantities?


have for a roof the imaginary expression a + j3 \/ 1, it must also
have for a root the conjugate imaginary expression a
have the following identity :

/3

*/

We

Let the polynomial f(x) be divided by the second member of


and if possible let there be a remainder lix + R\

this identity,

We

have then the identical equation


f(x)

where

is

(x

a)

/3

Q + Ex + ,R',

the quotient, of n - 2 dimensions in

x.

Substitute in

27

Imaginary Roots.

+ /3 \/- 1 for x. This, by hypothesis, causes f(x)


2
2
Hence
It also causes (x - a) + j3 to vanish.

this identity a

to vanish.

"!)

jR'

- Of

from which we obtain the two equations

Ba + jR' =
since the real

0,

JR/3

and imaginary parts cannot destroy one another

hence

o,

K=

0.

Thus the remainder Rx + R! vanishes


is divisible

0,

and, therefore, /

equationjhas, consequently, the root a

The

(a?)

without remainder by the product of the two factors

as the root a +

Thus the

j3

/3

v/-

1 a s well

<v/- 1-

total

number

with real coefficients

of imaginary roots in an equation

is always even
and every polynomial may
be regarded as composed of real factors, each pair of imaginary
roots producing a real quadratic factor, and each real root pro;

ducing a real simple factor. The actual resolution of the polynomial into these factors constitutes the complete solution of the
equation.

We

observed in Art. 17 that equal roots

may

be considered

between real and imaginary roots. This


statement may now be regarded from another point of view.
- 2 -f /, and
a}
Suppose a polynomial has the quadratic factor (x
let its form be altered by means of slight alterations in the
as the connecting link

value of

k.

When

pair of real roots;

when k

is

when

negative, the quadratic factor gives a


k = 0, this factor has two equal roots, a
;

the factor has two imaginary roots.


proof exactly similar to that above given shows that surd
roots, of the form a
v/y, enter equations whose coefficients are
is positive,

rational in pairs.

28

General Properties of Equations.


EXAMPLES.
Form a

1.

which

rational cubic equation

1,

have for roots

shall

+ 2\/~l.
Ans. x*

2.

Form

3.

Solve the equation

a rational equation which

*<

2a 3

shall

have

Ans.

a;

bx2

12# 3

0,

6* + 2

4-

two of

for

1x*

+ 19* -13 =

its

roots

72x 2

0.

- 312* + 676 =

0.

V-

1.

which has a root

V3.

Ans. The roots are - 2


4.

\/3,

Solve the equation

+
one root being

-f

v/

88

0,

7.

.4w*.

The

roots are 2
u

Descartes9 Rule of Signs Positive Roots. This


which enahles us, by the mere inspection of a given equalimit to the number of its positive
to assign a
superior^

19.
rule,
tion,

roots,

may

be enunciated as follows

No

equation can

have

more positive roots than


has changes of sign from + to -, ana
to -f, in the terms of its first member.
from
shall content ourselves for the present with the proof
which is usually given, and which is rather a verification than
it

We

a general demonstration of this celebrated theorem of Descartes.


It will be subsequently shown that the rule just enunciated, and

other similar rules which were discovered by early investigators


number of the positive, negative, and imaginary

relative to the

roots of equations, are immediate deductions


general theorems of Budan and Fourier.

Let the signs of a polynomial taken at


other in the following order

from the more

random succeed each

In

+_

+ _

--

+ +

this there are in all seven

changes from +

to -,

and from -

..f

<

changes of sign, Including

to +.

It

is

proposed to show

Deseartes* Rule of Signs.


that

29

polynomial be multiplied by a binomial whose signs,

if this

corresponding to a positive root, are + -, the resulting polynomial will have at least one more change of sign than the
original.

We write down
as follows

only the signs which occur in the operation

is placed
Here, in the third line, the ambiguous sign
wherever there are two terms with different signs to be added.

We

observe in this case, and it will readily appear also for


every other arrangement, that the effect of the process is to

introduce the ambiguous sign wherever the sign -f follows +, or


- follows -, in the
The number of variaoriginal polynomial.
tions of sign

is

never diminished.

one variation added at the end.


instance,

There
This

is

is, moreover, always


obvious in the above

where the original polynomial terminates with a varia-

terminate with a p^jrmnfinn of sign, it will equally


that
the corresponding ambiguity in the resulting polyappear
nomial must furnish one additional variation either with the

tion

if it

preceding or with the superadded sign.. Thus, in even the most


unfavourable case that, namely, in which the continuations of
sign in the original remain continuations in the resulting poly-

and we may conclude in


nomial, there is one variation added
of
the
the
effect
multiplication of a polynomial by
general that
a is to introduce at least one additional
a binomial factor x
;

change of

sign.

Suppose now a polynomial formed of the product of the


factors corresponding to the negative and imaginary roots of an
equation

x - a x ,

>

/3>

each

the effect of multiplying this


|3,

-y,

&c., is to introduce at least

so that

when

by each

of the factors

&c., corresponding to the positive roots

one change of sign for

the complete product

is

formed containing

30
all

General Properties of Equations.


the roots,

least as

we conclude

that the resulting polynomial has at

changes of sign as

many

it

has positive roots.

This

is

Descartes' proposition.
20.

Descartes 9 Rule of Signs

Negative Roots.

In

order to give the most advantageous statement to Descartes' rule


in the case of negative roots, we first prove that if - x be substituted for x in the equation f(x) = 0, the resulting equation will
have the same roots as the original except that their signs will

be changed.

This follows from the identical equation of Art. 16

f(x)

(x

a,) (x

- a a)

(x

a 9)

(x

a w ),

from which we derive

/(-

From

x)

- (- l) n

(x

a,)

(x

a a ) (x

e< 3

this it is evident that the roots of /(- x]


di,

cr 2 ,

,),

Hence the negative roots of f(x)


we may enunciate Descartes' rule

....

- an

(x

<*,,).

are

are positive roots of /(- a?), and


for negative roots as follows:

No

equation can hare a greater number of negative roots than there


are changes of sign in the terms of'the polynomial f(- or).

Use of" Ifescartes' Rule in proving the existence


of Imaginary Roots. It is often possible to detect the
21.

existence of imaginary roots in equations by the application of


Descartes' rule; for if* it should happen that the sum of the
greatest possible

number

of positive roots,

added

to the greatest

number of negative roots, is less than the degree of the


equation, we are sure of the existence of imaginary roots. Take,

possible

for example, the equation


x*

+ 10^ 3 + x - 4 =

0.

This equation, having only one variation, cannot have more than
one positive root. Now, changing x into - #, we get

-x-4 =

find since this has only one variation, the original equation cannot have more than one negative root. Hence, in the proposed

31

Theorem.
equation there cannot exist more than two real roots.
therefore, at least six

Descartes' rule

equations

for

imaginary

It has,

This application of

roots.

available only in the case of incomplete


is easily seen that the sum of the number of

is

it

variations in f(x) and /(- x) is exactly equal to the degree of


the equation when it is complete.
22.

Theorem.

If two numbers a and

substituted for

b,

the polynomial f(x), give results with contrary signs,

of real roots of the equation f(x] =


give results with the

same

ber of real roofs

beticeen them.

lies

lies betiveen

in

an odd number

them

sign, either no real root or

and

if they

an even num-

This proposition, of which the theorem in Art. 12 is a particular case, contains in the most general form the conclusions

which can be drawn as


signs furnished by
are substituted for

the proposition

its

to the roots of

an equation from the

member when two given numbers

first

We

x.
proceed to prove the first part of
the second part is proved in a precisely similar

manner.

Let the following


the equal i on J

(x)

a m and no others, of
01, a 2
between the quantities a and b, of

roots

lie

which, as usual, we take a to be the lesser.


Let (;r) be the quotient when ,/'(#) is divided by the product
</>

of the

factors

(.r

ai) (x

a 2)

We

(x- a m ).

have, then,

the identical equation

f(r) s

(x

a,) (x

a,)

Putting in this successively x


f(a) =
=
f(b)

Now

<f>(a)

(a
(6

a,} (a

,)

and $

different signs there

(b)

(i-

a.)
,)

....

(x

x =

- am
) $
b,

(x).

we obtain

.... (a- a m )
(a),
(i-a m )0(6).
</

....

have the same sign; for if they had


be, by Art. 12, one root at least of

would

the equation 0(#) =

between them. By hypothesis, f(a) and


have
different
hence the signs of the products
signs
f(b)
- 8
a,) (a
;

(a

(b

n,;

'/>

no)

(a

a*),

.... (6 - a m ),

32

General Properties of Equations.


but the sign of the second is positive, since all
;
are positive ; hence the sign of the first is negative ;
the factors of the first are negative ; therefore their

are different
its factors

but

all

number must be odd, which proves the


In

are counted a

number

proposition.

be understood that multiple roots

this proposition it is to

the degree of their

of times equal to

multiplicity.
It is instructive to apply the graphic method of treatment to
From this point of view it
the theorem of the present Article.

appears almost intuitively true ; for it is evident that when any


two points are connected by a curve, the portion of the curve

between these points must cut the axis an odd number of times
when the points are on opposite sides of the axis and an even
;

number

of times, or not at

all,

when

the points are on the same

side of the axis.

EXAMPLES.
1.

If the signs of the terms of

an equation be

all positive, it

cannot kave a

positive root.
2. If the signs of the terms of any complete equation be alternately positive
and negative, it cannot have a negative root.
3. If an equation consist of a number of terms connected by + signs followed
signs, it has one positive root and no more.
by a number of terms connected by

Apply Art. 12, substituting and oo and Art. 19.


4. If an equation involve only even powers of #, and
;

positive signs,

it

if all

the coefficients havr

cannot have a real loot.

Apply Arts. 19 and 20.


5. If an equation involve only odd powers of x, and if the coefficients have all
positive signs, it has the root zero and no other real root.
6. If an equation be complete, the number of continuations of sign in f (x) IB
the same as the
7.

When

number of variations

an equation

roots is equal to the


to the
8.

number

An

is

number

complete

of sign in
;

f(x).

if all its roots

of variations,

be

real,

the

number

and the number of negative

of positive

roots is equal

of continuations of sign.

equation having an even number of variations of sign must have

sign positive,

and one having an odd number

of variations

must have

its last

its last

sign

negative.

Take the highest power of a? with positive coefficient (see Art. 4).
Hence prove that if an equation have an even number of variations it must
have an equal or less even number of positive roots and if it have an od?t number of
9.

variations

it

must have an equal or

less

odd number of positive roots

in other

33

Examples.
when

words, the number of positive roots


from it by an even number.

less

than the number of variations must

differ

and

Substitute

Find an

10.

oo

and apply Art. 22.

inferior limit to the

a*

number

3*;*

of imaginary roots of the equation


1

0.

Ans. At

least

two imaginary

roots.

Find the nature of the roots of the equation

11.

x*

-\-7x-ll-0.

I5x~

-f

14, 19, 20.

Apply Arts.

Ans. One positive, one negative, two imaginary.

Show

12.

that the equation


a:

+ qx

-t

0,

where q awl r are essentially positive, has one negative and two imaginary roots.
13. Show that the equation

- qx +

x*

where

two

0,

q and r are essentially positive, has one negative root; and that the other

roots arc cither imaginary or both positive.

Show

14.

that the equation

$z

AX - a

where

a, b, c,

....

C2
.1-f

numbers

are

-f
T

....

4I

I?
=x
v ~-I

-m,

from one another, cannot have an

all different

imaginary root.
Substitute a

-f

$ V

and a

)8

in succession for x, and subtract.

get an expression which can vanish only on the supposition


16. Show that the equation

a"

is

and -

when n is even, two real roots, 1


1, and no other real root; and,
odd, the real root 1, and no other real root.

has,

We

0.

when n

This and the next example follow zeadily from Arts. 19 and 20.
16. Show that the equation

has,

when n

is

even, no real root

xn +

and,

when n

a 2
30 #

-f

is

odd, the real root

1,

and no

other real root.


17. Solve the equation
a:

This

is

-f

20 *r

r4

r4

Ans.

- q

=* 0.

equivalent to
(x

The

2qaP

+ qx

-f

2 2
)

0.

/--

different signs of the radicals give four


combinations,

here written involves the four roots.

02

+ vV +

r*.

and the expression

General Properties of Equations.

34

equation which has for roots the diffeient values of the expression

Form the

18.

where

+ v/11 + 0V7,

e \/~t

1.

made by the introduction of 0, this expression would


The \/7 must now be taken with the same sign where it occurs

If no restriction had teen

have 8 values.

under the second radical and free from


in

Form

\/34 -

v^ll^f-H

Am.

1.

Form an

20.

12u; 2

only four values


MX - 63 = 0.

the equation which has for roots the four values of

are, therefore,

# 4 - 8# 3 -

Am.

19.

where

There

it.

all.

x*

36x

20 Vlo7,

400 j; 2

3168* + 7744 -

equation with rational coefficients which shall have for roots

0.

all

the

values of the expression


0i

where

0i

\/p
= 1,

02 \/'y

-t

v r*

0- J

=1,

ft-

1.

Tl;ere are eight different values of this expression, viz.,

- %/J" - \/(? - y/r,

y/P + \/f -r V/^


'

Vp - V

- y/r,

'<i

i>

<i

+ v^^

Assume

Squaring,

* =

0i

2 (02

y/p

4"

02

\/^ +

03

\/r.

we have

*2 = p + q +

-f

A/V +

e 3 *i

\/^ +

01 02

Transposing, and squaring again,


2
{a;

-p-^-

r)

4 (0r

rp

py)

H0j0a0s

Transposing, substituting x ibr 0i -v/^


obtain the final equation free from radicals

{*

This
written.

yp
is

is

- 2* 2 (p +

r)

4-

p2 +

02

r*

V/

?; l? r

-f 02

2j?r

(0i

\/P +

\/# +

03

2rp

02

V^ r

an<* squaring,

2joy}

we

= *lpqr&.

an equation of the eighth degree, whose

Since

\/q

0i, 02,

yr

is

roots are the values above


have disajjpeared, it is indifferent which of the eight roots
assumed equal to x in the first instance. The final equation

that which would have been obtained

if

from X9 and the continued product formed,

each of the 8 roots had been subtracted


as in Ex. G, Art. 16.

CHAPTER

III.

RELATIONS BETWEEN THE ROOTS AND COEFFICIENTS OF EQUATIONS, WITH APPLICATIONS TO SYMMETRIC FUNCTIONS OF THE
'"*'

HOOTS.

Relations between the Roots and Coefficients.

23.

Taking for simplicity the coefficient of the highest power of .r as


unit}', and representing, as in Art. 1C, the n roots of an equation
a M we have the following identity
by ai, tig, a 3
,

xn + yM-"**

When

-f

- a,)
(s

;>,.r"--

+...

--

(JP

,,)

(.r

04) ----

(JT

- un
)

(1)

member of this identity are


the
multiplied together,
highest power of x in the product is .r";
~
the coefficient of ^ n l is the sum of the n quantities a^ - a^ &c.,
the factors of the second

the roots with their signs changed the coefficient of M ~*


the sum of the products of these quantities taken two by

viz.
is

two

jt

the coefficient of

by three; and

;r

n ~3

sum

the

is

of their products taken

term being the product of


all the roots with their signs changed,
liquating, therefore,
the coefficients of x on each side of the identity (1), we have the
three

so on, the last

following series of equations


p\

pz

= =

(HI

ay

(ci

riv

(- 1)"

-\

a,

-f

4-

a 2 aj +

n ),

....+ a n .\

o w ),

36

Roots ond

of Equations.

Coefficients

which enable us to state the relations between the roots and


coefficients as follows

Theorem.
whose highest term
its

sign changed

The

In every algebraic equation,


is

coefficient

the coefficient of
unity , the coefficient p^ of the second term with

is

equal

of

to the

sum of

the roots.

the third term is equal to the

sum of the

products of the roots taken two by two.

The

coefficient

equal to the

and

so

o)i,

p* of the fourth term with

sum of

its

sign changed

is

the products of the roots taken three by three

the signs of the coefficients being taken alternately neyatin-

positive, and the number of roots multiplied together in each


term of the corresponding function of the roots increasing by unity,
till finally that function is reached which consists of the product of

and

the n roots.

When

the coefficient a

must divide each term

of

is

then equal to

to
equal
^

and

is

not unity (see Art.

of the equation
'

roots

;r

the

sum

by

it.

The sum

1),

we

of the

of their products in pairs is

so on.

tfo

Cor. 1.
Every root of an equatioil is a divisor of the absolute term of the equation.
If the roots of an equation be all positive, the
Cor. 2.
coefficients (including that of the highest power of JT, will be

and if the roots be all


alternately positive and negative
will
be
all
the
coefficients
This is obvious
positive.
negative,
from the equations (2) [cf. Arts. 19 and 20],
;

24.

Applications of the Theorem.

(2) of the preceding Article supply


the n roots and the coefficients, it

it

Since the equations

distinct relations

between

might perhaps be supposed


that some advantage is thereby gained in the general solution
of the equation.
Such, however, is not the case for suppose it
were attempted to determine by means of these equations a root,
;

of the original equation, this could be effected only by the


elimination of the other roots by means of the given equations,
and the consequent determination of a final equation of which
di,

37

Theorem.
01 is

one of the

Now,

roots.

in whatever

way

this final equation

obtained, it must have for solution not only <n, but each of
the other roots a 2 a 3 ,
an
for, since all the roots enter in the
is

same manner in the equations (2), if it had been proposed to


determine a 2 (or any other root) by the elimination of the rest,
our final equation could differ from that obtained for ai only by
the substitution of a 2 (or that other root) for m.
The final
equation arrived at, therefore, by the process of elimination
must have the n quantities <n, a2
a n for roots
and can,

than the given equation.


the original equation itself, with

not, consequently, be easier of solution

This

final

equation

in fact,

is,

we

are seeking substituted for x.


This we shall show
for the particular case of a cubic.
The process here employed
is general, and may be
applied to an equation of any degree.

the root

JLet a,

/3,

7 be the roots of the equation


r

We

3
;

j>^'~

l>*

= -

Multiplying the

and adding the

<^>r

IB

The

*-

ft

+ ay

r /;,

0.

7),
i

/3y,

of these equations

three,

jhiC

which

"P7'

first

>
afi

"
]>*

a,

p&

have, by Art. 23,


;>i

by

4-

/'V<

7>,* +

we

find

= -

/; 3

--

p.,a

;> 3

by

2
,

the second

0,

the given cubic with a in the place of x.


student can take as an exercise to prove the same result

an equation of the fourth degree. In the corretreatment


of the general case the successive equations
sponding
~
~
~
of Art. 23 are to be multiplied by a w
a n 2 a w 3 &c., and added.
in the case of

Although the equations (2) afford, as we have just seen, no


assistance in the general solution of the equation,
they are often
of use, in facilitating the solution of numerical
equations when
relations
the
roots are known to exist.
any particular
among

They may

also be

employed to establish the

relations

which

Roots and

38

must obtain among the

of Equations.

Coefficients

coefficients of algebraical equations cor-

responding to known relations

the roots.

among

EXAMPLES.
1.

Solve the equation


a;S_

the

sum

of

two of

its roots

Let the roots be

a,

5:r

We have

7.

Taldnc:
01

ft

third

Thus the throe


2.

y=
we

0,

we

80

IGrr -f

0,

being equal to xero.

second

-f

theu

7 =

-f

16,

a#7 = -

80.

have, from the

obtain

first of these,

We

16.

roots are 5, 4,

6,

7 -

a7 4

find for

ft

and from either the

and 7 the values

and -

4.

Solve the equation


.r

3r2

0,

two of

its roots being equal.


Let the three roots be a,

a,

We have
2a +
2

and j8 = -

1.

a
from whioh
3.

we

find a

2,

2a/3

3,

0,

The

roots are 2, 2,

i.

The equation
9*

4# 8 -2a:2

12.*

+9 =

has two pairs of equal roots ; find them.


Let the roots be a, a, /3, ft we have, therefore,
;

=-

4,

+ 4a = -

2,

2a +
2

-f /8

2/3

from which we obtain for a and 3 the values


4

a;3

two

of

and

3.

Solve the equation

whose

93.2

14a

24

0,

roots are in the ratio of 3 to 2.

Let the roots be

o,

j8,

7, with the relation 2o

easily obtain

3/3.

By

elimination of a
(

60 + 2y =

18,

28,

30' + 6/87

we

39

Examples.
from which

we have

the following quadratic for


19j8

14

The

The

1.

three roots are

ficance of the value

all

6, 4,

~
of
iy

j8

0.

The student

1.

and the same

56

will here ask

difficulty

It will he observed that in

the previous examples.

never require

-f

the former gives for a and 7 the values 6 and

and

roots of this are 4,

90

what

is

the signi-

may have

presented

examples

of this nature

itself in

we

the relations between the roots and coefficients in order to deter-

mine the required unknown quantities. The reason of this is, that the given conmore relations amongst the roots. Whenever the equations

dition establishes one or

employed appear to furnish more than one system of values for the roots, the actual
roots are easily determined by the condition that they must satisfy the equation (or
equations) between the roots and coefficients which we have not made use of in
= 4 gives a system
detei mining them. Thus, in the present example, the value
satislying the omitted equation
a&-y

while the value &


to

=-

24

gives a system not satisfying this equation, and

is

therefore

be rejected.
6. Solve the equation
*

- 0^-f

23^-15-0,

whose roots are in arithmetical progression.


Let the roots be
a45; we have
a,

06,

3er

from which we obtain the three roots

5-

=-

at oru-o

9,

23,

1, 3, 5.

Solve the equation

6.

a-

whose

3a

-f 2a-

2 1* 3

- 22*

-f

40

0,

roots are in arithmetical progression.

Assume

for the roots

3$,

5,

5,

-f-

35.

'

Am.
7.

-6, -2,

1,

Solve the equation


27:r3

+ 42#2 -2S;r-8=*

0,

whose- roots are in geometric progression.

Assume
have a8
for p.

for the roots ap, a, -.

or a

From

the third of the equations

(2),

Art. 23,

we

2
-.

Either of the remaining two equations gives a quadratic

An,.

2-2
-_.

-2,-,

Roots and

40
8.

Solve the equation


Zx*

whose

- 40* s

-f

130*2

120a?

for the roots ^,


'
equations (2), Art. 23.

ap t

-,

3
a/>

0,

the second and fourth of the

Employ

^ ^ ^
^
3

Solve the equation

120s + G4

whose

27

-f

roots are in geometric progression.

Assume

9.

of Equations.

Coefficients

0,

Ans.

roots are in geometric progression.

8.

4,

2,

1,

10. Solve the equation

whose roots are in harmonic progression.


We have here the
Take the roots to be o,
7.

relation

1-

?..

hence

By + 70+ aB = 3yal &c.

An

11
*

*'

2'

11. Solve the equation


[

whose

roots are in

0,

harmonic progression.

22

"''

3'

9'

~3*

12. If the roots of the equation

xs -

i/x

-f

qx

be in harmonic progression, show that the mean root


13.

The equation

is

p-21 =

has two roots equal in magnitude and opposite in sign determine all the roots.
Take a-f = 0, and employ the first and third of equations (2), Art. 23.
;

Ans.
14.

v/3, - v/3,

V/ ^

vTi.

The equation
# 2 - 50# +12=0

has two roots whose product

is

find all the roots.


.

16.

One

of the roots of the cubic


a?

is

6,

double another ; show that

it

px z +

may

^a?

be found from a quadratic equation.

41

Examples.
16.

Show

that all the roots of the equation

can be obtained

when they

Let the roots be

a,

gives

are in arithmetical progression.

$,

-pi = na

2&,

a + (n

+2+3+

-f { I

The

1) 8.

(n

first

of equations

(2)

1) } 5

Again, since the sum of the squares of any number of quantities is equal to the
square of their sum minus twice the sum of their products in pairs, we have the
equation
2
/?l

2j0 2

a2

(a f 8)

-f

(a

25)

IN

~, /

ft)**

1 \

-i

(2)

we find

Subtracting the square of (1) from n times the equation (2),


of p\ and j2We can then find a from equation (1). Thus

all

8 2 in terms

the roots can be

expressed in terms of the coefficients p\ and p*.


17. Find the condition which must be satisfied by the coefficients of the equation
re

when two

of its roots

a,

2
px + qx

$ are connected by a

r = 0,
relation a

& =

0.

Ana. pq
18.

r = 0.

Find the condition that the cubic


#3

px* +

</x

its roots in geometric progression.


Ans. p 9 r q* = 0.
Find the condition that the same cubic should have its roots in harmonic
Ans. 27r 2 - pqr + 20 3 a 0.
progression (see Ex. 12).
20. Find the condition that the equation

should have
19.

JT* -f

pa?

gx~

+ rx + s =

hould have two roots connected by the relation a

+3 = 0;

ease two quadratic equations which shall have for roots

Ans.

pqr-p^ - r8 = 0,

(1)

px*

and determine in that

(1) a,

+ r = 0,

(2)

and

xz

-t

(2) 7, 5.

P
px + - =

0.

21. Find the condition that the biquadratic of Ex. 20 should have its roots connected by the relation
Ans. j? 3 4pq + 8r = 0.
+ 7 = a 5.
22. Find the condition that the roots a, /8, 7, 8 of
I

should be tonneoted by the relation oj8 = 78.


23. Show that the condition obtained in Ex. 22
'

the biquadratic are in geometric progression.

Ans. p*t - ra =
is satisfied

when

0.

the roots ot

Roots and

42

of Equations.

Coefficients

Depreggion of an Equation when a relation


between two of its Roots. The examples given

25.

exists

in the preceding Article illustrate the use of the equations con-.


necting the roots and coefficients in determining the roots in
particular cases when known relations exist among them.

We

in general, that if a relation of the form /3 = </*(a)


exist between two of the root* of an equation f(x) = 0, the equation
may be depressed two dimensions.

now show

shall

Let $(x) be substituted

for

x in the identity

-f <7,.r

x
/(# = (<>#)'* +
We represent, for convenience,

then

identity

by

hence a

satisfies

j3,

the second

member

of this

we have

the equation Fix] = 0, and it also


hence the polynomials /'(.r) a,ud
;

satisfies the
1*'

common measure x
or

Substituting a for ^,

F(x).

equation. /"(a?)

0(ci)

n~

(x)

have a

thus a can be determined, and from it


and the given equation can be depressed two dimena

sions.

Ex
1.

\MT'LKS.

The equation
-

has two roots whose difference

3_5.^_4.r+ 203

find

them.

= 3-f-a; substitute x + 3 for x in the given polynomial f(x)


Heie )3-o = 3,
it becomes x* + 4x z
7x 10 the common measure of this and /(.*:) is x - 2 from
which o = 2, = 5 the third root is 2.
2. The equation
;

6-0

has two roots connected by the relation

2ft

+ 3o =

find all the roots.

Ans.

1,

2,

v/--~2-

It may be observed here, that when two polynomials /(#)


and F(x] have common factors, these factors may be obtained
by the ordinary process of finding the common measure. Thus,

43

The Cube Roots of Unity.

we know that two given equations have common roots, we


can obtain these roots by equating to zero the greatest common

if

measure of the given polynomials.

EXAMPLES.
1.

The equations

have two common roots


2.

The

- 6x-

4-

5r 2

3^ +

= 0,

-ll*-150
An*

find them.

3.

1,

equations
ir

:<

:r

have two common root?

-i

px~
p'x

-f

qx
#'#

-f r'

= 0,
=

whose roots are these two, and

find the quadratic

-f

find also

the third root of each.

p-p
26.

r-r

p-p

The Cube Roots of

Unity.

r-r'

Equations of the

forms
xn -

-=

0,

xn

-f

0,

consisting of the highest and absolute terms only, are called


The roots of the former are called the n n th
binomial equations.

A general discussion of these forms will be given


subsequent chapter. We confine ourselves at present to

roots of unity.

in a

the simple case of the binomial cubic, for which certain useful
It has been
properties of the roots can be easily established.

already shown (see Ex.

5,

Art. 16), that the roots of the cubic


a*

are

*'

- 1 =

11

11
*
^

'

~2~2^~~ 6

'

If either of the imaginary roots be represented by w, the


2
is easily seen to be u> , by actually squaring; or we may

other

see the

must

same thing

also

as follows

be a root

for,

since

If
oi

w be a root of the cubic, w*


1, we get, by squaring,

44

Roots and

w8 =

1,

or (w

8
)

We

# - 1 = 0.
3

Changing

thus showing that


have then the identity

we

tf

1 a

the cubic

get the following identity also

(a?

satisfies

o>

1,

into -a?,

a;

of Equations.

Coefficients

1) (#

-i

o>) (a?

a> ),

which furnishes the roots of

&

-f

0.

Whenever

in any product of quantities involving the imagiof unity any power higher than the second
cube
roots
nary
it
can be replaced by o>, or or, or by unity for
presents itself,
;

example,

The

first or second of equations (2), Art. 23, gives the


of the imaginary cube roots
property
lowing
1 -f u> + or = 0.

fol-

the aid of this equation any expression involving real


quantities and the imaginary cube roots can be written in any
f a>Q,
+ urQ, <P + ^Qof the forms

By

EXAMPLES.
1.

Show

that the product


(o>w

is

2.

Prove the following

or//) (o>

wri)

Ans. m" -

identities

m 3 -f

0,3 _ W 3 ss

3.

Show

(m -f n)((am -f
(

WJ

mn+

w,)(

ww

a>

w)(&r

-I-

wn),

'*) (or ;>*

/>).

that the product


(a-f

is

rational.

o))8-t

a>

7)(a

+ co 2

rational.

-4m.
4.

(a 46.

<r

I-

& 2 + 72 - 0y

- ya

Prove the identity


j8 -f

7)(a -f

j8 -f

7)(

2
)8

+ W 7) =

a3

-f )8

-f

3
7 - 80^7.

Prove the identity


(a +

Apply Ex.

j8

2,

7)

+ (a+

2
/8 4-

7)

(2a

- 3 - 7) (2/5 - w - a)(2-y ~ a

3).

n2

45

The Cube Roots of Unity.


6.

Prove the identity


(a

a>j8

7)

(a

a>

Apply Ex. 2, and substitute


7. Prove the identity
s

a' 4-

w~w

for

j8'

4 wy s -

7'*

its

*/~$ ($ - 7) (7 -

value

3a'Y a

(a

\/
8

a) (a

- 0).

3.

4 7* -

SajS?)

',

where
2
a' s a 4-

8.

Form

'

7'==7 +

0*427a,

207,

iw

o>m

n,

orw,

^ns.
9.

Form

2aj8.

the equation whose roots are

whose

the equation

I 4-

4-

a;

- 3mi* - (w8

4-

n3 ) =

0.

roots are

Ant. x*

aw.

4-

4-

tm 4

w ? w,

3te* 4- 3 (/-

4 w? w 4

- mn) x -

(/

4-

cuw.

m8 4

- S^mw) = 0.

th
important to observe that, corresponding to the n n
th
The roots
roots of unity, there are n n roots of any quantity.

It

is

of the equation

xn - a -

are the n n th roots of

The

a.

three cube roots, for example, of a are

V3/

0>

fl,

V3/~

0)~

V3/-

<7,

where v^a represents the real cube root according to the ordinary
Each of these values satisfies the
arithmetical interpretation.
It is to be observed that the three
cubic equation x* - a = 0.
cube roots

may

be obtained by multiplying any one of the three

above written by

1, <u,

wa

In addition, therefore,

to the real cube root there are

two

imaginary cube roots obtained by multiplying the real cube


root by the imaginary cube roots of unity.
Thus, besides the
ordinary cube root
cube roots

3,

the

number 27 has the two imaginary

,-

as the student can easily verify


10.

Form

by actual cubing.

a rational equation which shall have

wV
for a root

Vhere

Compare Ex.

o>

8.

Q 4\/Q2 + f* +

3
2

-V& + ^

Ans. x* 4 3Pz - 2Q

0.

46

Roots and
11.

Form an

for a root,

of Equations.

equation with rational coefficients which shall have

where

Cubing both

Coefficients

0i

and

1,

fa

1.

sides of the equation

x=
and substituting

a:

for

its

0!

v/p +

0,

value on the right-hand side,

we

get

Cubing again, we have

Since

Q\

and

may

0%

each have any one of the values

1,

u>,

w2 the nine roots of


,

this equation are

We see

also that, since 9i and 2 have disapperaed from the final equation, it is
which of these nine roots is assumed equal to x in the first instance.
The resulting equation is that which would have been obtained by multiplying

indifferent

together the nine factors of the form


above written.

x -<\/P

\/Q

obtained from the nine roots

12. Form separately the three cubic equations whose roots are the groups in
three (written in vertical columns in Ex. 11) of the roots of the equation of the

preceding example.

"We can write these down from Ex.


then equal to

cu

1\

&\/ Q,\

and

8,

taking

first

finally equal to

and n equal

to^/P /Q',
9

o*~\/P, <jj"\/Q.

Ans.

x*

l/TQx - P~ Q
- P- Q =

**

27.

0,
0,

Symmetric Functions of the Hoots.

Symmetric

functions of the roots of an equation are those functions in


which all the roots are alike involved, so that the expression is
unaltered in value when any two of the roots are interchanged.

For example, the functions of the roots (the sum, the sum of the
products in pairs, &c.) with which we^ were concerned in Art. 23
.are of this nature; for, as the student will readily perceive, if
,in

any

of these expressions the root a b let us say, be written in

Symmetric Functions of

47

the Boots.

every place where a 2 occurs, and a 2 in every place where en


occurs, the value of the expression will be unchanged.
The functions discussed in Art. 23 are the simplest sym.

metric functions of the roots, each root entering in the

first

any term of any one of them.


without
We can,
knowing the values of the roots separately
in terms of the coefficients, obtain by means of the equations (2)
of Art. 23 the values in terms of the coefficients of an infinite
degree only in

It will be shown
variety of symmetric functions of the roots.
in a subsequent chapter, when the discussion of this subject is
resumed, that any rational symmetric function whatever of the

The examples appended to this


so expressed.
most of which have reference to the simple cases of the
cubic and biquadratic, are sufficient for the present to illustrate

roots can be
Article,

the usual elementary methods of obtaining such expressions in


terms of the coefficients.
It

is

letter

pression

usual to represent a symmetric function by the Greek


attached to one term of it, from which the entire ex-

be written down.

may

Thus,

if a, /3,

y be the

roots of

a cubic, Su~/3 represents the symmetric function


2

cr/3

+ a? 7 2 +

/3Y,

possible products in pairs are taken, and all the


terms added after each is separately squared. Again, in the

where

same

all

case,

Sa 2 /3 represents the sum


a

where

4 a 2 y 4 fi*y 4

13

/3

4 y*a 4 y 2 /3,

permutations of the roots two by two are


taken, and the first root in each term then squared.
As an illustration in the case of a biquadratic we take 2a 2 /3 2
all possible

whose expanded form

By
among

is

as follows

the aid of the various symmetric functions which occur


the following examples the student will acquire a facility

in writing out in all similar cases the entire expression


the typical term is given.

when

Roots and

48

of Equations.

Coefficients

EXAMPLES.
1

Find the value of 2a5

of the roots of the cubic equation

4 px* 4

tf

qx

=s

0.

Multiplying together the equations


a

+ & 4 y=

0y 4 ya 4
we

2a2 /3

obtain

hence

2crj8

^.

3r

2.

Find for the same cubic the value of

3.

Find for the same cubic the value

a2

-f

4 y\

hence, by Ex.

47 +

-I-

1,

2a = 3

4.

Find

We

for the

jt?

= - ^3 +

Zpq

2^^

3r.

V+7

'W

easily obtain

2cr/3

(ft

is

7)

?'S

2pr.

7) (7

3).

equal to

2a$7

2a--

-f

Ans.

Find the value of the symmetric function


o2 /3y

47

a)8

4-

o2/85
2

47

o3

cryS

47

4 ^ 2 a7 4 # 2 a5 4

^35

5*a

82 a7

of the roots of the biquadratic equation


ar*

4 px* 4 qx 1 4

Multiplying together

obtain

ra;

0.

a44748 = -;>,
a/87

hence

Find for the same cubic the value of

This

we

)3

6.

p*

same cubic the value of

from which
6.

2a-)3
3

2a'J

+ 7s
we obtain

Multiplying the values of 2a and 2o

Ans.

of

a+0
o8

y,

- pq\

3 a#y

4-

-JP,

a)85

4 a7^ 4

2a

j87

^878

r,

4 4a075 = pr ;

2a 07 = .pr 2

=-

4t

49

Examples.
7.

Find for the same biquadratic the value of the symmetric function
o8

S2 .

+ 72 +

4-

Squaring 2a, we easily obtain

2a2 = p2 8.

Find

for the

a2

a2 7 2 + a 2 5 2 +

Squaring the equation

we

2?.

same biquadratic the value of the symmetric function


2

2a0 =

2
7 +

+ y-5 2

5-

f,

obtain

2a2
hence,

by Ex.

+ 22a 2 /3y

-f

6a0y$ =

8
;

6,

- 2pr +

2*.

Find for the same biquadratic the value of 2a3 /8.


To form this symmetric function we take the two permutations a# and j8a of
We have similarly two
the letters a, ft
these give two terms a3 and /8 3 a of 2.
so that the symmetric functerms from every other pair of the letters a,
7, 5
9.

tion consists of 12 terms in all.

Multiply together the two equations

2a/3

2a2 = JP - 2?

y,

and observe that

s 2a

2a 2a

3
j8

2o

4-

/87.

convenient to remark here, that results of tho kind expressed by this last
equation can be verified by the consideration that the number of terms in both
members of the equation must be the same. Thus, in the present instance, since
[It is

2a2

contains 4 terms, and 2a 6 terms, their product must contain 24 ; and these
3
2
tlie 12 terms which form 2a
together with the 12 which form 2a #7.]
Using the results of previous examples, we have, therefore,

are in fact

2a 3 /8 = p*q - 2? 2 - pr +
10.

Find

for the

same biquadratic the value


a4

+ 74

fr

-f

4*.

of
54 .

2
Squaring 2a , and employing results already obtained,

2a 4 = J3 4 - 4p~q
11.

Find the value, in terms of the

-f

2g-

+ 4pr ~

4*.

coefficients, of the

sum

of the squares of the

roots of the equation

Xn +p&n- 1 + p2X~2
Squaring 2ai,

we

-f

. .

+ 22aiaa

+p n = 0.

easily find
2
jPi

2ai 2

hence

2ar =
12.

Fiad the value,

pi* -

2p z .

in terms of the coefficients, of the

the roots of the equation in the preceding example.

sum

of the reciprocals of

Roots and

50
From

the second
CX203

last,

... On

01

and
tt3

of Equations.

Coefficients

last of

CUCC2

....

by

~r

"r

02

l)*- ^,,-

= (-l)"jp

<*

we have

the latter,

= (-

a,i-l

....

01 0303

dividing the former

we have

the equations of Art. 23,

.... On +....+

~~~

,
~r

~t~

On

03

In a similar manner the sum of the products in pairs, in threes, &c., of the
found by dividing the 3rd last, or 4th last, &c., co-

reciprocals of the roots can be


efficient

13.

by the last.
Find for the cubic equation

the value, in terms of the coefficients, of the following symmetric function of the
roots o, 8,

2
08-7) +(7-a) +(a-/3)

N.B. It will often be found convenient to write, as in the present example, an


equation with binomial coefficients, that is, numerical coefficients the same as those
which occur in the expansion by the binomial theorem, in addition to the literal
Here the equation being of the third degree, the successive
coefficients a
i, &c.
,

numerical coefficients are those which occur in the expansion to the third power,
viz. 1, 3, 3, 1.

We

easily obtain
3

14.

Express

in

{(ft

7)

(y

a)

(a

0)

18 (a?

terms of the cooffirients of the cubic in the preceding example

the successive coefficients of the quadiatic

where

a,

/3,

are the roots of the cubic.

we have

Here, in addition to the symmetric function of the preceding example,


to calculate also the

two following

7)

+ P~ (7 -

Ann. (o2 15.

-f

x~

(a

~ #) 2

+ (o3-oi2)a?-r

(tfis

Find for the cubic of Example 13 the value in terms of the


(2a

Since

2
)

JB

7) (2jB

-7-

2a-/3-7=3a-(o4

a) (2 7

/8^ 7)

- j8)

= 3a-h

3ai

= 0.

coefficients of

51

Examples.
the required value

30]

-4w. ao

16. Find, in

of

- 7)

(2a

--

obtained by substituting

is easily

+ 3aa# +

(2)8

03

- 7 - a)

o (#

(#

- a - /3)
(2 7

for

(*

ft)

-7).

- 27 (o2

>

in the identity

-3

+ 2ai 8 ).

terms of the coefficients of the biquadratic equation


OQX* f 4ai a? 4 6a z a?

+ 4 3# +

4 *= 0,

the value of the following symmetric function of the roots


(ft

- 7) a (a -)*' +

(7

3
)

08

5)

(a

*)* (7

2.

Here the equation is written with numerical coefficients corresponding to the


expansion of the binomial to the 4th power. The symmetric function in question
is easily seen to be identical with

Employing the

<,-{(

7)

2
(

*)

examples G and

results of
2

+ (7-

(0

a
)

8,

+ (a- jB)*

we

find

(7

8
)

= 24

- 40, *s

4 3^ 2 2 ).

17. Taking the six products in pairs of the four roots of the equation of Ex. 16,
and adding each product, e.y. oj8, to that which contains the remaining two roots,
78, we have the three sums in pairs

7 4
it

is

ya

a5,

-f

a/8 -f

5,

78

required to find the values in terms of the coefficients of the two following

symmetric functions of the loots


(70 T 35) ;a

-f

76;

is

the

(a

(7 -f a5)
The former of these

sum

+ 78)

(7 +

(ya

j85) (aj8 -f

5)

(7 +

a3) (ya

+ /38),

78).

of the pi oducts in pairs, and the latter the con-

tinued product, of the three expressions above ^iven.


As these throe 1 unctions of
tbe roots are important in the theory of the biquadratic, we shall represent them

uniformly by the

letters X,

of the coefficients for

The former

is

(j.v

"We have, therefore,


and A/tv.

v.

/x,

I/A

2a'~'7, and

i.s

easily expressed as follows (cl.

~
The

latter

is,

when

3
375(a +

and we obtain

to find expressions in terms

A/u,

ix.

(4ais -

multiplied out, oqual to

+ 7 +8- +
:i

/3-

>

>

a-

^-7 -5 -( \

\ a-

^-

-\ H -[:)
5-/

7-

after easy calculations the following:


a<p\p.v

8 (Zauaj- - S'/oOuty

K 2

-I-

2r ^4).

6)

Roots and

52

of Equations.

Coefficients

18. Find, in terms of the coefficients of the biquadratic of Ex. 16, the value of
^he following symmetric function of the roots
:

{(7

-)(-)-(-*) (7-*)}

l(a

This is also an important symmetric function in the theory of the biquadratic.


prevent any ambiguity in writing this, or corresponding functions iii which the
differences of the roots of the biquadratic enter, we explain the notation which will

To

be uniformly employed in this work.


7, we have the three differences
Taking in circular order the three roots a,
and subtracting 5 from each root in turn, we have the three
a, a
7, 7
,

other differences a
OB

The symmetric

5,

-7)

/3

5,

We

5.

(-),

(7

combine these in pairs as follows

-*)(*-*),

(a

-3) (7-).

function in question is the product of the differences of these

three taken as usual in circular order.

Employing the values of

A,

/A,

v,

in the preceding example,

-M + "03-7)(-*). -" + A=(7-)(/3.--

We have,

we have

5),

therefore, to find the value of

(2A

-^-

v]

- A)
(2ft- v

(2v- A -

,i),

or

2a)3),

in terms of the coefficients of the biquadratic.

Multiplying this out, substituting the value of 2o#, and attending to the results
of Ex. 17,

The

we

obtain the required expression as follows

function of the coefficients here arrived at, as well as those before obtained

in Examples 13, 16, and 16, will be found to be of great importance in the theory
of the cubic and biquadratic equations.
19. Find, in terms of the coefficients of the biquadratic of

Ex.

16, the value of

the symmetric function

This

may be

represented briefly

by 2

(a

2.

Ans.
20.

Prove the following relation between the roots and


Ex. 16

dratic of

-a-

S)

(7+ a-

-5)

(a-f

-7-8) =32

coefficient?

Theorems relating

53

Symmetric Functions.

to

Theorems

relating to Symmetric Functions.


The following two theorems, with which we close for the present
the discussion of this subject, will be found useful in many in28.

stances in verifying the results of the calculation of the values


of symmetric functions in terms of the coefficients.
(1).

The sum of the exponents ofall the

roots in

any term of any

symmetric function of the roots is equal to the sum of the suffixes in


each term of the corresponding value in terms of the coefficients.
The sum of the exponents is of course the same for every

term of the symmetric function, and

The

all the roots of that function.

may be

called the degree in

truth of the theorem will be

observed in the particular cases of the Examples 13, 15, 16, 17,
&c. of the last Article ; and that it must be true in general appears from the equations (2) of Art. 23, for the suffix of each
coefficient in those equations is equal to the

degree in the roots

of the corresponding function of the roots hence in any product


of any powers of the coefficients the sum of the suffixes must be
;

equal to the degree in


tion of the roots.
(2).

When an

all

equation

the roots of the corresponding funcwritten with binomial coefficients, the

is

expression in terms of the coefficients for

any symmetric function of

the roots, which is a function of their differences only,


the algebraic

equal

sum of the numerical factors of all

is

such that

the terms in

it

is

to zero.

The

truth of this proposition appears by supposing all the


#> #i> a, &o. to become equal to unity in the general
written
with binomial coefficients, viz.,
equation
coefficients

a Q xn + naixn
for the equation then

equal

~l

~
1

becomes

a^'^ +

Z
(a?

+ l) w =

0,

,...
i.

e. all

=();

the roots become

hence any function of the differences of the roots must

in that case vanish,

and therefore

also the function of the coeffi-

equal to it ; but this consists of the algebraic sum


of the numerical factors when in it all the coefficients a a t , a8 ,
cients

which

is

&c. art

made equal

to unity.

27 we have instances of

In Exs. 13, 15, 16, 18, 20 of Art.

this theorem.

Roots and

54

of Equations.

Coefficients

EXAMPLES.
1.

Find

in terms of p, q, r the value of the

J_

_i_

&y
where

a, 3,

symmetric function

_i_

a&

ya

are the roots of the cubic equation

Xs

2
px + qx +

4-

0.

Ans.

3.

r
2.

Find for the same equation the value of

^4w.
3.

Here

Calculate the value of 2a 3

2a2a

= 2a s /8 3

-4-

/8

of the roots of the

a72a'0

5a 6

is

Find

for the

easily obtained

3
by squaring 2a

same equation.

2
%pqr + 3r

Ex.

(see

-f

7*

fi

2;>

1 2yy

1 2j3

8;/ 02

Spqr

60

2
2
7 4 a

a2

4- j3

2
'

+7

7+0

+
2^

Find for the same equation the value of


or

+ 7
a-f)

Am
7.

Art. 27)

3,

Find for the same equation the value of


2

6.

same equation the value of the symmetric function

Ans.
5.

- jo 3

hence, &o.

Ans.
4.

24r

Find for the same equation the value of

a3

27a -

&

2a/8

7*
*

4pr

20

55

Examples.
8.

Find the value of the symmetric function 5

--

\a +

for the

equation.

- 4?8 "

9.

Calculate in terms of p,

+ py?

x*
1

2aS a-2

Here

- 5-

4-

a/8

-f

qx*

+
1

and Sa2 -

S~

the value of

g, r, s

rx

+ 2

7-

Find the value of 5

~a

- 9r2

0.

gr

10.

2pqr

for the equation

-f

same cubic

Iq-s

~ pra f

of the roots of the equation

fl~

p n -i# + /> =

-f

11.

Find for the biquadratic of Ex. 9 the value of


(/3

Compare Ex.
12.

0.

7 -

a8)

7a -

5) (a/3

- 7 3).

^s.

22, Art. 24.

Find the value of 2 (oa

-f

i)

(j8

- 7) 2

in terms

cubic equation
tfo#

+ 3a 8 # +

3#i#-

0.
1

Find the value of the symmetric function 2

xn

-f

^la?"'

+ jt?2#n~2 --....

The given function may be written

in the form

W |I ,L +
+

12

(ai

1
(

2oi2
14.

Qear

!)_,
1

~ ,1

-f-~f"

03

ai

a;

a*

-f-a\

or

.. ..

Ill
- + - +

-f l^n-ia? 4- JP

<

hence, &c.

of radicals the equation


2

V/*~^~a~

-f

A/^-0

4 v/' - 72

5
^o*"

(a

2
i

)*-

of the roots of the

CC1A2

equation

of the coefficients of the

v4w.

13.

r2 ~jo2

0.

56

Roots and

and express the

of Equations.

Coefficients

coefficients of the resulting equation in

in terms of the coefficients

of the cuhic of Ex. 1.

Ans. 3* 2 - 2
15. If a,

2
(j

+ !)(* -f

(a

1)( 7

1)(5

+1) -

-p* +

2#)

7, 5 be the roots of the biquadratic of

Ex.

Substitute in turn each of the roots of the equation

a?

8jpr

0.

- r).

(p

prove

9,

- q +*) 2 4-

(1

4p*q

in the identity of

Art. 16, and multiply.


16. Prove the following relation between the roots and coefficients of the
th
general equation of the n degree
:

17.

Find the numerical value of

vhere

1) (a2

(ai

---- (a

1)

1)

(1

- p* + Pi -

. .

.)*

(p\

- P* +

)*

are the roots of the equation

a, 0, 7, 8

x*

Substitute in turn for

- 7s8 +

x each

8s2

6x

10

0.

root of the equation

x2 4 2 =

0,

and multiply.
Ana. 166.

If a, 0, 7, 5 be the roots of the equation

18.

0,

prove
o

(3

The

A/tv,

-f

7) (7

4-

+ W(a ^

)(

symmetric function in question

where

A,

/x,

equal to

is

have the values of Ex.

19. Calculate the value of the

Ex.

biquadratic equation of

Ifi

(Ca^as

(/* 4-

v}(v -f A) (A

-f A*),

or

17, Art. 27.

symmetric function

2 (a -

#)

of the roots of the

9.
4

Ans.

Show

5)(/3 -f 5) (7 -f 5)

when

S;;

I6p~g

20?

-f

+ 4pr -

16*.

written with binomial coefficients, as in


tfee biquadratic
Ex. 18, the value of the symmetric function of the preceding example may be
20.

that

is

expressed in the following form


a
21.

The

(a

4
)

{48

(a

ff2

distances on a right line of

the line are the roots

(a,

ax 2
prove that

1C>

and

(a', /8')

Zbx 4

3
i

two

(a

%-

4rt]

from a fixed
two quadratic equations

pairs of points

of the

0,

a'x*

2b'x

-f c'

origin

on

when one

pair of the points are the harmonic conjugates of the other


pair, the following relation exists :
ac'

22.

The

-I-

a'c

2##'

distances of three points A, B,

on the line are the roots of the equation

C on

0.

a right line from a fixed origin

57

Examples.
find the condition that

one of the points A,

C should

bisect the distance

between

the other two.

Compare Ex.

Ans. a*d- Zabe + 25 s =0.

15, Art. 27.

23. Retaining the notation of the preceding question, find the condition that the

four points 0,

A,

C should form a harmonic

J?,

division.

Ans. adz

3bcd+

2c s

= 0.

This can be derived from the result of Ex. 22 by changing the roots into their
reciprocals, or it can be easily calculated independently.
24. If the roots (a, 0, 7, 5) of the equation
e

be so related that a

8,

yS are

5,

harmonic progression, prove the relation

in

the coefficients

among

ace+ 2bcd-ad z -b*e-&**Q.

Compare Ex.

18, Art. 27.

equation whose roots are

Foifli the

25.

By +
a
3

where

1,

and

a,

/8,

twya

-|-

c*>0 -f

a>

7 are the

26.

Express

sum

of

-f

ya-f a>a0

ar0

-f-

w*v

of the cubic

-f 34^'" -f 3r^r 4- rf

-7 ~

ajB) (2 7 a

= 0.

- 187) (20 - ^7 -

aj8

two cubes.
Ann. ($7

Compare Ex.

4- a>

14, Art. 27.

(20?
as the

a0

roots*

ax*

Compare Exs. 13 and

co*7

5,

-f-

o>7<x 4-

<w'

Art. 26.

27. Express
3

z)

in terms of

jc

4 y

-f

and

-1-

(a?

4-

\vheie

*yj5,

f/ -4

o>

w'J s) 3

+ (j;

w-

4-

//

1.

A**.
28. If
r'

find

-3T,

r,

^ in terms

Apply Jbixample

3 *' y' *')

H-

F8 4- ^s -

'

of #, y,

#', y',

4, Ail. 26.

Am.
29.
(a

-f

7)

a3y

(07

+ ya -f

a0)

into threefactors, each of the second degree in a, 0, 7.

Ans.

Compare Ex.

18, Art.

24

(cr

- 07)

(0

- 7)

(7*

aj8)

Roots and

58

of Equations.

Coefficients

30. Eesolve into simple factors each of the following expressions


(1).

(-

(2).

(J

8~
Ans.

31.

(1).

(2a-0-7)(2-7-a)(27-a-/8).

(2).

-90S-x)(7-a)(a-#.

Find the condition that the cubic equation

Xs

px 1 -\-qx-r-Q

should have a pair of roots of the form a


the roots in that case.
If the real root
roots,

p2

*2q

we

is b,

The

b~.

and show how to determine

hy forming the sum of the

easily find,

required condition
(P*

2?) (q*

sq uares of the

is

2pr)

r*

0.

32. Solve the equation

a*

whose

roots are of the

7x*

form indicated

20#

in

24

0,

Ex. 31.

Ans.
33.

Roots

3,

and 2 + 2

V^l

Find the conditions that the biquadratic equation


x*

px*

among

rx

qx

aV

should have roots of the form a


conditions

1,

Here there must he two

1.

the coefficients, as there are only two independent quantities

involved in the roots.

Ans.

p*-

1q

r*

'Iqs

0.

84. Solve the biquadratic


x*

whose

roots are of the form in

4a- 3

Sx*

120s

4-

900

-f /3

0,

Ex. 33.
Ans.

35. If a

V^, -6 + 5 V~l.

be a root of the equation


x^

0s

0,

prove that 2a will be a root of the equation

Xs + g# - r =
36.

0.

Find the condition that the cubic equation


x9

should have two roots

a,

ft

-f

2
j# + ##

-f

connected by the relation

aft -f 1

Ana.

=
1 -f

0.

+ pr

4 r*

0.

59
Find the condition that the biquadratic

37.

#4 + px* + qx" 4 rx

-f

should hnve two roots connected by the relation aj8


The condition arranged according to powers of s
1

+ q + pr +

Find the value of 5

38.

xn
This

r1

+ (^ -f j?r - 2q 2

ao)

(ai

--

p\x

n'}

1) *

0.

is

(q

- 1) **-f * ^ o.

a3 04 .... cm of the roots of the

+pzz

n~2

equation

+pn = 0.

readily reducible to Ex. 13.

is

Am.

(-

39. If the roots of the


equation

'*
i

be in arithmetical progression, show that they can be obtained from the


expression

by giving to r

all

0, 2, 4, 6 ....

the values 1,3,5, ____


1,

when w

is

1,

when n

is

even

40. Representing the differences of three


quantities

follows

and

all

the values

odd.
a,

j8,

7 by

ai,

1,

71, a.

aiE=-7,

7isa-^3;

j8i-7-a,

prove the relations


3

oi

ai

ai

These

results

+
+
-f

-f

7i

-f

71*

3
/8i

4
)3i

fo*

-f

7i

3ai/3,7i,
2
2
J {ai + ^i

= f

can be derived by taking

* + qx is

2
/Si

01, 3i,

(where the second term

{ai

absent since the

-f

7r}%

4 71"

01/8171.

71 to be roots of the equation

sum

of the roots

= 0), and

calculating

the symmetric functions 2ai 8 , 2ai 4 , 2oi 5 in terms of and r.


The process can be
q
extended to form 2ai 6 , 2ai 7 , &c. The sums of the successive powers are, therefore,

capable of being expressed in terms of the product ai/3i7i and the sum of squares
2
-f
+ 71* ; the former being equal to r, and the latter to - 2 (fr 71 + 71 ai -f- a t 1),
i
or
2g. These sums can be calculated readily as follows :
By means of x* = r - qx,
all

oi

and the equations derived from this by squaring, cubing, &c., and multiplying by
x or ar2 any power of #, say xf, can be brought
by successive reductions to the form
A + Bx+ Cx*, where A B, C are functions of q and r. Substituting ai, 0i, 71, and
,

adding,
this

we

way

find 2ai*

7?

r,

= 3^t
5ai u

2^(7.

110r

The student can


3

(? -r*).

take as an exercise to prove in

CHAPTER

IV.

TRANSFORMATION OF EQUATIONS.
29.

We

Transformation of Equations.

instances, without

knowing

tion in terms of the coefficients, transform


substitutions, or
roots, into

can in

many

the values of the roots of an equait

by elementary

of the symmetric functions of the


equation whose roots shall have certain

by the aid

another

transformaassigned relations to the roots of the proposed.


tion of this nature often facilitates the discussion of the equation.
proceed to explain the most important elementary transfor-

We

mations of equations.

Hoots with Signs changed.

To transform an equabe equal to the roots of the


be the
given equation with contrary signs, let cu, a a a 3
30.

whose roots

tion into another

shall

roots of the equation

3? +

We
JC

nr
p X

n ~*

-*rp 2 X

p n ^X + pn =

have then the identity


2
+ p n ^X + p n m ( X 'Vp.,0*- +

+ p^-

cliauging x into
//"

'

-Piy

n~ l

y^

8)

we have, whether n be even

+ p*y n ~* - ...

The polynomial

Ol ) (x

0.

pn^ y T

pn

(//

i)

(x

dn )

or odd,

(y + a 2 )

(//

n).

y equated to zero is, therefore, an equation


- a w and to effect the
- a2
required
transformation we have only to change the signs of every alternate

whose roots are -

in

ai,

term of the given equation beginning

ivith the second.

EXAMPLES.
1

Find the equation whose roots are the

&+
with their signs changed.

7* 4

7o?3

roots of

- 8# 2
Ans.

-f

xb

-f

7# 4

1x*

8x* +*x

=0.

To Multiply Roots by a Given Quantity.


2.

Change the signs

61

of the roots of the equation


7

3# 6

-4

-f

8
it-

[Supply the missing terms with zero

x-

= 0.

Ix + 2

-f

coefficients.]

Am.

a7

4- 3.r

+ # 8 -f x* + 7x-

0.

To

Multiply the Roots by a (niven Quantity.


a
iiito anTo transform an equation whose roots are
a,,
31.

2,

i,

ma^ wa

other whose roots are

2,

many we change

the identity of the preceding Article.

x into

Multiplying by

in
in
n
,

we

have
n
n
y + mp\y

~l

m*p 2

n ~*
i/

+
-

(l/

mn p n ^y + mnpn
~l

men) (y

- ma-z ---)
(y

iwa n ).

to multiply the roots of an equation by a given quanwe


have only to multiply the successive coefficients, beginning
tity m,

Hence,

Kith the second, by m,

/;?.%

;w

3
,

mn

The present transformation is useful


moving the coefficient of the first term

for the

of

purpose of rean equation when it

not unity and generalty for removing fractional coefficients


from an equation. If there be a coefficient a of the first term,
we form the equation whose roots are aa\ tf 2,
the
oa n
transformed equation will be divisible by a and after such
division the coefficient of x n will be unity.
it is

When

there are fractional coefficients,

by multiplying the

roots

common

all

multiple of

by a quantity

we can

get rid of them


is the least

which

the denominators of the fractions.

In

many cases multiplication by a quantity less than the least


common multiple will be sufficient for this purpose, as will
appear in the following examples

EXAMPLES.
1.

Change the equation

into another the coefficient of whose highest term will be unity.


Am. x^ - 4.r + I2x~
multiply the roots hy 3.

We

2.

Remove

- 18*

-4-

27 =0.

the fractional coefficients from the equation

Multiply the roots hy

6.

dns

3?

3x~

+ 24#- 216 = 0.

62

Transformation of Equations.
Remove

3.

the fractional coefficients from the equation

6,7

By noting the factors which occur in the denominators of these fractions, we


observe that a number much smaller than the least common multiple will suffice to
remove the fractions. If the required multiplier be m, we write the transformed
equation thus :

evident that

it is

have only

if

= 6,

be taken

to multiply the roots

by

each coefficient will become integral

Ann.

Remove

4.

The student must be


terms with zero

13

we

&-\i>x*- Ux +

0.

77

careful in examples of this kind to supply the missing

coefficients.

The required

multiplier

Ans.

Remove

hence

the fractional coefficients from the equation


3

6.

6.

x*>

is

10.

30# 2

-f

520# 4-770

= 0.

the fractional coefficients from the equation


.
JF*

--6 #3 +
,

5
;r

-- 18

o2

0.

900

12

Ans.

- 25^-f 375r2 - 11700 =

0.

32. Reciprocal Roots and Reciprocal Equations.


To .transform an equation into one whose roots are the reciprocals

of the roots of the proposed equation,


identity of Art. 30.

we change x

into - in the

This substitution gives, after certain easy

reductions,
1

Pi

P*

Pn~\

or

hence,
1

if

in the given equation

we

replace x

by

-,

and multiply

resulting polynomial in y equated to zero


for roots the reciprocals of a 1? a 2 .... a n

by

y'

the*

*rill

have

63

Reciprocal Equations.

There

when x

is

a certain class of equations which remain unaltered

is

changed into

The

equafions.

These are called

its reciprocal.

conditions which must obtain

among

reciprocal

the coef-

an equation in order that it should be one of this class


has been just proved, plainly the following
what
by

ficients of

are,

Pn~\
- ~
2J n

The

- PL

Pn-i

- P*

- - P- ~1 -

;?i

*<>

lj n

Pn

J'n

last of these conditions gives

*
pn =

or

1,

P~

pn =

l.

procal equations are divided into two classes, according as


equal to + 1, or to -1.

In the

(1).

which give

first

case

we have

Reci-

pn

is

the relations

rise to thejfov^ class of reciprocal equations, in which

the coefficients of the corresponding terms taken

the beginning

from

a if (I end are equal in magnitude and have the same signs.


In the second case, when p n = 1, we have
(2).
Pn-i

= -Pi,

pn-2

giving rise to the second

&c,

-I**,

class

pi = -p,,_i

of reciprocal equations, in which cor-

responding terms counting from the beginning and end are equal in
magnitude hut different in sign. It is to be observed that in this
ease when the degree of the equation is even, say n- 2ui, one of
the conditions becomes

p m = -p m

equations of the second class,


term is absent.

>

or p m =

If a be a root of a reciprocal equation,

equation

identical with the proposed

reciprocal equation occur in pairs,

degree

and
or +

is

it is

a,

/3,

also be a root,

and the transformed


hence the roots of a

odd there must be a root which


in fact obvious

even, the middle

is

- must

for it is a root of the transformed equation,


is

so that in reciprocal

whose degree

is its

&c.

own

When

the

reciprocal

from the form of the equation that -1,

then a root, according as the equation is of the first or


second of the above classes. In either case we can divide off by
I

is

64:

Transformation of Equations.

the

known

factor

(x+

1 or a?- 1),

equation of even degree

and

and what

of the

the second class of even degree a?


tion may be written in the form
n-2

is left is

is

- 1 +

a reciprocal

In equations

first class.

of

factor, since the equa-

0.

- 1, this also is reducible to a


dividing by x*
reciprocal
of
first
class
of even degree.
Heuce all reciprocal
the
equation

By

equations
even,

and

may be reduced to those of the first class whose degree is


may consequently be regarded as the standardform

this

of reciprocal equations.
EXAMPLES.
1.

Find the equation whose roots are the reciprocals of the roots of
**

- 3^ +

7* 2

6*

Am.
2.

Reduce

to a reciprocal equation of
a
6

* +

K
8

22

=
2y

0.

5//

even degree and of

22

T* +T*

Ans.

To Increase or Diminish
To

Quantity.

effect this

3y

-f

0.

first class

*-i-o.
x*

-f

- x* - -- x* + - x +
6

33.

7?/

0.

the Roots by a Given

transformation

we change

the vari-

polynomial/ (x) by the substitution x = y + h the


resulting equation in y will have roots each less or greater by h
than the given equation in x, according as h is positive or negaable in the

tive.

The

resulting equation

f(h) +/' (A)y +

There

is

mode

is (see

Art. 6)

^-V -Q^yH-

... -

o.

which for

of formation of this equation

practical purposes is much more convenient than the direct calculation of the derived functions, and the substitution in them

of the given quantity h.

This we proceed to explain.

proposed equation be
.

+ an - x + a n }

Let the

To Increase or Diminish
and suppose the transformed polynomial
y =

since

a?

h, this is

65

the Roots.
in

be

to

equivalent to

which must be identical with the given polynomial. We conclude


if the given polynomial be divided by x - A, the remainder

that
is

A ny

and the quotient

A
if this

(x

+ AI (x -

A)*"

again be divided by x

quotient

*Q. o

7jN**~"^

f/yt
I***

/fry

_i_
1~

A)**"

h,

A)

the remainder

is

/)\H~^
/*
J

( t*

>4

+ An_ 2 (x -

<<cjLl I***

"f"

A^i

An

-i,

and the

_i_
~T~

4-

-Cl^j_2

Proceeding in this way, we are able by a repetition of arithmetical operations, of the kind explained in Art. 8, to calculate
in succession the several coefficients

formed equation the


;

last, A

An

An.^ &c., of the trans-

being equal to a

It will appear

in a subsequent chapter that the best practical method of solving numerical equations is only an extension of the process
employed in the following examples.

EXAMPLES.
1.

Find the equation whose


si

each diminished by

The

roots are the roots of

5# 8

7# 2

llx

-f

11

0,

4.

calculation is best exhibited as follows

12

12

20

60

65

11

11

17

66
-

Transformation of Equations.

Here the first division of the given polynomial by x - 4 gives the remainder
and the quotient x z - x* -f 3#
5 (cf. Art. 8).
Dividing this a^ain by

9 (= At) t

we get the remainder 55 (= A$), and the quotient x~ + 3# + 15. Dividing


we get the remainder 43 (= AZ), and quotient x + 7 and dividing this we get
A\=ll and AQ = 1 hence the required transformed equation is
x

4,

again,

55y - 9

Find the equation whose roots are the

2.

.#

each diminished by

4#

-#

0.

roots of

-f

11

0,

3.

15

The transformed equation


y
3.

-f

therefore,

is,

2
305y 4 507f/ + 353 =

3
15y + 94y

-f

0.

Find the equation whose roots are the roots of

4x*

each increased by

The

2# 3

multiplier in this operation

is,

by

7z

of course,

Am.
4. Increase

0,

2.

2.

303y-129 =

4i/s-4Qy*

7 the roots of the equation

0.
*

6.

Diminish by 23 the roots of the equation


6z 3

13# 2

- 12* +

0.

The operation may be conveniently performed by


20,

- 2876y -f 4058 = 0.

first

diminishing the roots by

and then diminishing the roots of the transformed


equation again by

3.

The

Removal of Terms.
calculation

may

67

be exhibited in two stages, as follows, the broken lines marking

the conclusion of each stage

332

Am.

Removal of Terms.

34.

-f

5//

One

7325?/

+ 53689 =

0.

of the chief uses of the

transformation of the preceding article

is

to

remove a

certain

Such a step often facilitates


specified term from an equation.
its solution.
Writing the transformed equation in descending
powers of

y,

we have

If h be such as to satisfy the equation naji + a =


l

formed

equation will

want the second term.

0,

the trans-

If h be either of

the values which satisfy the equation

5
JL

1)

0,

the transformed equation will want the third Jterm the removal
of the fourth term will require the solution of a cubic for h and
;

so on.
/(A)

To remove

0,

which

is

the last term

we must

the original equation

solve the equation

itself.

68

Transformation of Equations.
EXAMPLES.
1.

Transform the equation


3

ic

which

into one

shall

- 6^ + 4x -

want the second term.

naoh

a\

gives

2.

Ans. y* - 8y

Diminish the roots by 2.


2. Transform the equation
a^

which

into one

shall

65

= 0.

0.

+ Sa^-f #-5 =

want the second term.

Increase the roots


3.

15

-4s. y 4

2.

by

24^-

-f

65y

Transform the equation

which

into one

The

want the

shall

third term.

quadratic for h is

- 12A -

6/i 2

Thus there are two ways


Diminishing the roots by

Binomial

giving h

0,

=-

3r J

we obtain
y*+ 8// s - Illy- 196 =
we obtain
y*-fy4 17y-8 = 0.

1,

(2)

35.

I.

3,

(1)

Increasing the roots by

18

of effecting the trans formation.

Coefficients.

0.

In many algebraical pro-

found convenient to write the polynomial /(a?) in the


form
following
v (n - 1)
n(n - 1)
n"
-f
^-2^ + w^n-i^ + <*m
cesses it is

-*-

in

which each term

ficient,

*w

affected, in addition to the literal coef-

is

with the numerical

coefficient of the

corresponding term

w
expansion of (x + l) by the binomial theorem. The
student will find examples of equations written in this way on

in the

Examples 13 and

referring to Article 27,


to

The form

16.

is

one

which any given polynomial can be at once reduced.

We

now adopt

the following notation

Un = a&?

naiX

-f

n~ l

+
J.

#n

~2

+ na n . x + a n
}

/o

with the suffix n to represent the polynomial of


thus using
th>
the n degree written with binomial coefficients

Binomial

We have, therefore,

One advantage

69

Coefficients.

changing n into n

form

of the binomial

1, &o.,

that the derived

is,

The

functions can be immediately written down.

Un

function of

or n

Un

is,

first

derived

plainly,

so that the first derived function of a polynomial represented in this way can be formed by applying to the suffix
of
the rule given in Art. 6 with respect to the exponent of the
.i

variable.

Thus, for example, the

by multiplying the function by


unity

We

derived of U^

first

is

and diminishing the

4,

formed

suffix

by

4 Z73 as the student can easily verify.


to
now
proceed
prove that the substitution of y + h for

it is,

therefore,

x transforms the polynomial


n
n
#o# + nciix

~l

7W or
,

-:

flr

1 &

2 o?

n~ a

-f

na n

\x

into

A yn
Q

nAiif- +
1

-f

nin - 1)
n'2
+
\ >> - ^3//

^l-iy + -4,

where
-^oj -^ij -4 2,

are the functions which result


*7
i.e.

A* =

0n,

-^

=
i

o^

CT,,

Z7,,

^i,

by
.

-^2

-^-n-ly

-&n

substituting A for x in
.

Un- l9
o^

a
*

Un

',

2iA

+ aa , &o.

Representing the derived functions of /(A) by

suffixes,

as

70

Transformation of Equations.

explained in Art.
mation, viz. f(y +

/,

we may

6,

write the result of the transfor-

A), in the following form:

W,
Un

f (A) is the result of substituting h for # in


-4

its first

derived/i(A)

derived of this again


substitutions,
to write

is

we have

by

is,

n(n

A n_i\

1)

it

and

so on.

therefore,

is,

nA n.

above rule,

tlie

the

first

these

Making

the result above stated, which enables us

down without any

calculation the transformed equation.

EXAMPLES.
1.

Find the

Am.
The

<7

result of substituting

3 (onh

-f

i)

-I-

y+h

A2 +

x in the polynomial

for

2iA +

^2)

?/ -I-

^o^ 3

-f 3^!/<

stud(^nl will find it a useful exercise to verify this result

calculation explained in Art. 33,

-f

3^/t

<z s .

-f

hy the method of

which may often be employed with advantage

ii

the case of algebraical as well as numerical examples.


2. Remove the second term from the equation
3

3r/2#

-f

We

must diminish the

roots

o^

#3

0.

h obtained

by a quantity
-f

-f

from the equation

"i

= 0,

i.e. h

= --a,,

Substituting this value of

--

3 (a<ff2

h in A*, and A$, the resulting equation in y


2

r)

i*

o ^3

3. Find the condition that the second and third terms of the
equation Un should be capable of being removed by the same substitution.
Here A\ and A^ must vanish for the same value of A; and eliminating h

between them we find the required condition.

Ans.
4.

a^ci'i

Solve the equation


x*
its

fijr-

-4-

2r

second term.

by removing
The third term

is

removed by the same


y

substitution,

- 27 =

0.

which gives

- a{1 =

0.

The
The required

roots are obtained

71

Cubic.

subtracting 2 from each root of the latter

by

equation.
5. Find the condition that the second and fourth terms of the equation Un =
should be capable of being removed by the same transformation.
Here the coefficients A\ and A* must vanish for the same value of h ; eliminat-

ing h between the equations

=
we

ff

0,

/*

3#iA

Zctih

0,

obtain the required condition

3
i

0.

among the coefficients of a biquadratic equareducible to that of a quadratic for when the second term is
removed the resulting equation is a quadratic for y^ ; and from the values of y
N.B.

"When

this condition holds

tion its solution

those of
6.

oc

is

can be obtained.

Solve the equation


x*

by removing

its

-f

16s3 + 72s2

0.

in

is

y4_24y-

Solve in the same manner the equation

x + 20s 3
l

-I-

143s2

+ 430s

4-

Am. The
8.

and

129

second term.

The equation
7.

64s

462

0.

roots are

- 7, - 3, - 5

V'S.

Find the condition that the same transformation should remove the second

fifth

terms of the equation

Un =

0.

3#i

The

= 0.

account of their peculiar interest, we


shall consider in this and the next following Articles the equa36.

Cubic.

tions of the third

On

and fourth degrees, in connexion with the


When y + h is subarticle.

transformation of the preceding


stituted for x in the equation

we

0,

(1)

obtain

where AI,

0,

have the values of Art. 35.


If in the transformed equation the second term be absent,
-4 2 ,

^=

0,

or

A--^

72

Transformation of Equations.

Substituting this value for h in


Art. 35,

and A s we

find, as in

hence the transformed cubic, wanting the second term,


8
y +

3
2

(0 02

y+

01 )

00

The

T (00*03

3000102

20j

3
)

Ex.

2,

is

0.

00

functions of the coefficients here involved are of such

importance in the theory of algebraic equations, that


to represent
the notation

mary

them by single
2

0i

0o 08

-ff,

letters.

it is

custo-

We accordingly adopt

- 30Q0102 + 20j 8 =

and write the transformed equation in the form

If the roots of this equation be multiplied


s

3Hz + G =

by

it

becomes
(3)

a form which will be found convenient in the subsequent disThe variable, s, which occurs in the first
cussion of the cubic.
member of this equation, is equal to ,y or # + 0j ; the original
cubic multiplied

by

being in fact identical with

3H (a& +

as the student can easily verify.


If the roots of the original equation be

a,

transformed equation

(2) will

be

al

-i

cii
,

0o

i)

H-

-,

7 +

or, since

a + /3 + T1B

they

may

be written as follows

80!
-_
a

0j
J

p, y, those of the

73

The Biquadratic.

We

can write

down immediately by

the aid of the trans-

formed equation the values of the symmetric functions

S( 2a -j3-7)(2/3-7-a), (2a-/3- 7 )(2/3- T -a)(2 y --/3)


The

of the roots of the original cubic.

latter will be

found to

agree with the value already found in Ex. 15, Art. 27.
may here make with regard to the general equation an

We

important observation : that any symmetric function of the roots


a, j3, 7, S, &c., which is a function of their differences only, can

be expressed by the functions of the coefficients which occur in


the transformed equation wanting the second term.
This is
obvious, since the difference of

transformed equation

is

any two

roots

a',

/3'

of the

equal to the difference of the two corre-

sponding roots a, (3 of the original equation ; and any symmetric


7
function of the roots a', /3', 7', S , &c., can be expressed in terms
of the coefficients of the transformed equation. For example, in
the case of the cubic, all symmetric functions of the roots which
contain the differences only can be expressed as functions of
r/,,,

and G.

jET,

Illustrations of this principle will be

among

The Biquadratic.

37.

ing the second term,

is

r/o//

where
article

The transformed

.equation, want-

in this case

+ 6^4 2 // 2 + 4 A $ 4

-4 4

0.

and A z have the same values as in the preceding


z
and where A is given by the equation

641

The transformed equation


//

found

the examples of Art. 27.

Hy*
'

-f

"

G'/ +

We

'

is,

therefore,

(a*a

^ 1 ^8 + 6a

2
(ti

a2 -

3rt x )

0.

might if we pleased represent the absolute term of this


and G, and have thus three
equation by a symbol like
functions of the coefficients, in terms of which all symmetric
functions of the differences of the roots of the biquadratic could
be expressed.

It is

more convenient, however,

to regard this

74

Transformation of Equations.

term as composed of // and another function of the coefficients


determined in the following manner
We have plainly the
:

identity

This involves #

>

H, and another

function of the coefficients,

viz.

4^3
which

is

3fl 2

2
,

of great importance in the theory of the biquadratic.

This function

is

represented

by the

letter /, giving

The transformed equation may now bo

written

6H 40 + *'/-- 3jEP = A
r+ -Tjy
^ + T**
171
^0
#0
^0
,

We

J)

can multiply the roots of this equation, as in the case

the cubic of Art. 36,


s

by

of

and obtain

+ 6 Hz + 4 Gz + a<~I - 3II 2 =
2

0.

(2)

This form will be found convenient in the treatment of the


algebraical solution of the biqp.ajrflf.in.
same as in the case of the cubic, viz.

quartic multiplied
(a^x

f/!)

-f

by

The

a&

-f

variable

ft

is

the

the original

being in fact identical with

QH(a& +

i)

4G

(a<& + ai]

+ a<?I- 3ID.

Any symmetric function of the roots of the original biquadratic equation which contains their differences only can therefore be expressed by r/ ZT, 6r, and /.
,

If the roots of the original equation be a,

j3,

7, 8, those of

the transformed (1) will be, as is easily seen,

7 -S-), |(3 7 -g--/3), |(38-a-/3- 7 ).

The sum
dg

j; the

of these =

sum

the

sum

of their products t in pairs

of their products in threes =

r#0

and

for their

75

HomograpJiic Transformation.
continued product

we have

the equation

7 -$-a)(3 7
There

is

-p) (3S-a-/3-y)

attention, as

it

will be

found to he of great im-

portance in the subsequent discussion of the biquadratic.


the function arrived at in Ex. 18, Art. 27, viz.

This

is

denoted by the letter J.

shows that

we

another function of the coefficients to which

now to call

wish

--<

it

is

The example

It is

referred to

a function of the differences of the roots.

It

must, therefore, be capable of being expressed in terms of


have, in fact, the identity
H, G, and /.

tf

0?

We

HI- G*-H\

a<?J^a<?

which the student can easily

Or

this relation

verify.

can be derived as follows

Whenever a

a l9 a 2 &c. s is the expression of a


function of the differences of the roots, it must be unaltered by
the transformation which removes the second term of the equahence its value is unaltered when we change a^ into zero,
tion

function of the coefficients a

d 2 into

<7otf24

2,

# 3 into

+ 2<VWj -

z,

Thus

&c.

ffoffs

#i

<?2

s a*A ,A, - a^A-i -

A?

their values in terms of H, G, /, we


2 , A^
substituting for
easily obtain the above identity, which will usually be written
in the form

38. l!oiiiograjjiic

.Transformation.

The transforma-

tion of a polynomial considered in Art. 33 is a particular case


of the following, in which x is connected with the new variable

y by the equation

If A

ft

= - ^, A' =

0, [/

Solving fqr x in terms of

y,

= 1 we have y = x we have
,

A*

h, as in Art. 33.

76

Transformation of Equations.

This value can be substituted for x in the given equation,


and the resulting equation of the n th degree in obtained.
Let a, j3, 7, 8, &e., be the roots of the original equation,
and a', j3', 7', 8', &o., the corresponding roots of the transformed
;//

From

equation.

the equations

we

Aa +
Aa +

ju

= T7

\(3
it
&->
P = v/T +
ju
.,

/>

/u

>>

Ap

easily derive the relation

a -

A'//) (a

(\1JL

j3)

j3'

with corresponding relations for the differences of any other


If we take any four roots, and the four correpair of roots.

sponding

roots,
(a

('

we
-

obtain the equation


ft) (j

/)

(#'

y) =
~
8')

(a

(a

j3j

(7

7)

(/3

8)
Sj"

the roots of the proposed equation represent the


distances of a number of points on a right line from a fixed origin

Thus,

if

on the

line, the roots of the transformed equation will represent


the distances of a corresponding system of points, so related to
the former that the anharmonic ratio of any four of one system
is

the same as that of

It is

four conjugates in the other system.


in consequence of this property that the transformation is
tlieir

called holographic.

It

important to observe that the transformation here conand y are connected by a relation

is

sidered, in which the variables x

of the

form

Axy + Ex + Cy +

0,

the most general transformation in which to one value of either


variable corresponds one, and only oue, value of the other.

is

39.

Transformation

l>y

Symmetric Functions.

Sup-

required to transform an equation into another whoso


roots shall be given rational functions of the roots of the propose

it is

posed.

Let the given function be

involve all the roots,

7
or any number of them.
</>

(a, /3,

.),

where

We form all

may
pos-

77

Transformation by Symmetric Functions.


sible

combinations

down

and write

When

(afiy),

<f>

of the roots of this type,

(a/38), &c.,

the transformed equation as follows

expanded, the successive coefficients of

this product is

y will be symmetric functions of the roots a, /3, 7, &c., of the


given equation ; and may therefore be expressed in terms of the
coefficients of that equation.

EXAMPLES.
1.

The

roots of

+ px* +

s*

qx

-f

find the equation whose roots are a


Suppose the transformed equation to be

are a,

R=

3
y + Py* + Qy 4

then

- r=
and we have
tion.

We

to

- p2 -

in the

2cTj8

= a2

72

a2 /3 2 7 2 , of the given equa-

<<J

n
~

^+

2q]

(q'

2pr) y

same case the equation whose


y

a2 ^ 3 7 2

2/?r,

r*

therefore,

(p

^*.
If a,

2a 2 ^ =

2,7,

is,

3
2/

3.

j8

Q=

easily ohtain

the transformed equation

Find

+ 72

form the symmetric functions 2cr, 2a

2o-

2.

a?

+ j0# 3

-f

(p

-f

r*

roots are a

3pg

3r) y*

qx*

r#

(q*

0.
n

s
.

H=

0.

0>

024 + 2ai*4) =

ft.

3j>qr

3r2 ) y

7, 5 be the roots of

a*
find the equation

whose

roots are or,

-I-

0,

Let the transformed equation be

jy +

4
y +

%+

Q /y 2 +

xS

0,

then

- P = 2a 2
Compare Exs.

-R

= 2a2/3 27 2

^=

a 2 /8 2 7 2 5 2 .

8, 17, Art. 27.

Jim.
4. If a,

Q = 2a 2^3 2

y*>

2
(2?

- 2?) y 3

(g

-1-

2;?r -f 2*)

(r

2qs)

y+

*2

7, 5 be the roots of

ao# 4
find the equation

whose

-f

4i# 3

-f 6rta^

roots are x,

4-

o5,

/*,

7a

+ 43^ +

a4

viz.,
aft -f

-f j85,

75.

See Ex. 17? Art. 27.


4

<

(40i #3

- ao4) y

--8
-,

<*

(2o33 -

30

78

Transformation of Equations.
Show

5.

when

that the transformed equation,

the roots of the resulting cubic of

Ex. 4 are multiplied by i#o, and the second term of the equation then removed,
*

- Iz + 2J =

is

0.

Formation of the Equation whose Roots are any


rowers of the Roots of the Proposed. The method of
40.

effecting this transformation by symmetric functions,


plained in the preceding article, is often laborious.

as ex-

much

simpler process, involving multiplication only, can be employed.


It depends on a knowledge of the solution of the binomial equaThis form of equation will be discussed in the
tion xn - 1 = 0.

next chapter. The general process will be sufficiently obvious


to the student from the application to the equations of the

2nd and 3rd degrees which


-examples

be found among the following

will

EXAMPLES.

Form

1.

the equation whose roots are the squares of the r>ot* of

xn

To

+ piX

n~ l

+ pzX"--

changing x into

ap-pix"multiplying,
n

#,

we

+2>zti*-

evident that the

-even

~l

pix"~'~

we have

4-

+ .... 4

pn-l-r

powers of x

The

+ pn -

0.

the identity

+ pn =

(x

Ui)(x

02)

(x

cr, 4

derive, as in Art. GO,

Pn-\x +jj n

.^-(pia"-

first

member

we may then

(x

01) (x

a2 )

(x f

a,,)

4- jya

of this identity contains,

2
replace x

by

2jyj^3-f 2j> 4 )y-*+

N.B.

Pn~\ x

we have

+ p 2 JC-~ + piX n -* +

it is

u
p\ x

effect this transformation,

X"

(x

-f

and obtain

y,
.

B (y - oi 2

when expanded, only


finally

(//- a 2 ~)

(//- a

2
)".

member of this equated to zero is the required transformed equation.


This transformation will often enable us to determine a limit to the num-

first

ber of real roots of the proposed equation.


For, the square of a real root must be
positive; and therefore the original equation cannot have more real roots than tho

transformed has positive roots.


2.

Find the equation whose

roots are the squares of the roots of

r3 -

x-

-f

8* -

0.

Ana. y*

Tho

latter equation,

^positive root

by

15#

-f

52y

36

0.

Descartes' rule of signs, cannot have more than one

hencL the former must huvc a pair of imaginary

roots.

79

Examples.
3.

Find the equation whose

roots are the squares of the roots of

**--

+ *2

+ 2# +

Am.

y* +

0.

2y

2
6y* f 3y

- 2y - 9 =

0.

from Descartes' rule of signs that the original equation must hato

It follows

four imaginary roots.

by the method of Ex.

4. Verifj5.

Form

tho Examples

and

3 of Art. 39.

the equation whose roots are tho cubes of tho roots of

X n + plX>^ + P2X--

+p

iX

pn

-I-

0.

Ex. 1 the process consists in multiplying together


tho given polynomial, and f(x] : the variables involved in these being those
1 = 0.
which are obtained by multiplying x by the two roots of the equation x*
It will bo observed that in

/"(#),

In the present case we must multiply together /(#), /( wa Oi /(^ 2 #) the variables
1 = 0.
involved being obtained by multiplying x by the roots of the equation a; 3
'

The transformation may be conveniently represented


"Write the polynomial f(x) in the

(pn+pH-*& +

which we ropiesent,

.)

+ f(p n -i

for brevity,

H-^,,-4*

We

jinil

P, Q,

have

nre

all

P
P4

-f

xQ

+ x*R

(x

+ X-(p

ti

}>n-^* r

.),

x*R,

and

wxQ.

a-VQ

w-Ji-R

4 a**li

--.

HH

(tax

02)

t*)

----

(*

- a n).

(I)

x successively, we obtain

ai)(a>JC

(w^ - ai)(V-

03)
03)

(wx

(u'JC

a M ),

(2)

- O H ),

(3)

we
1

(1), (2), (3),

and attending to the

results

obtain
.r

3
tf

+ *JP - Za*PQR B

ai )(a;

a 2 3)

..(*-

a,

3
).

member

of this identity contains x in powers which are multiples of 3


can, therefore, substitute y for x" and obtain the required transformed

first

We

- oi)(-

Multiplying together both members of

only.

and P, being functions of # 3 are unaltered.

since P, Q,

The

functions of x*.

Changing, in this identity, x into tax

P-

then.

P4

of Art. 26,

by
P-f xQ

whorr

as follows

form

equation.
6.

Find the equation whose


&

roots are the cubes of the roots of

- *3 +

2xz

0.

Ans. y*

+ 3x

4- 1

3
2
14y + 50y + 6y

0.

<* s

0.

Verify by the method of Ex. 5 the result of Ex. 2 of Art. 39.


8. Form the equation whose roots are the cubes of the roots of
7.

ax*

Ans.

-f

Zbx*

Sex

+d=

0.

V+ 3(flM+ 9A- 9*0)y* + 3(<r4*+

9<:

- 9^)y +

80

Transformation of Equations.

41. Transformation in General.


In the general problem of transformation we have to form a new equation in ?/,
=
with
whose roots are connected by a given relation
(#, y)
=
0.
The transformed
the roots of the proposed equation / (x)
</>

equation will then be obtained by substituting in the given


equation the value of x in terms of y derived from the given
relation ^ (x y) =*
or, in other words, by eliminating x
;

= 0. For
between the two equations / (x) = 0, and
(x, y}
example, suppose it were required to form the equation whose
roots are the sums of every two of the roots (a, /3, 7) of the
<f>

cubic

pa? + qx r = 0.

or*

We have here
The equation

(x,

y)

is in

his case

=p - x

for

when x

takes the value a, y takes one of the proposed values; and when
x takes the values /j and 7, // ta,k s tne other proposed values.

The transformed equation is theiviore obtained by


- y for x in the given equation.
\)
E\
1.

If a, 0,

7 be the

substituting

VMPLES.

roots of the cuhn

xz

px

-4-

qx

0,

form the equation whose roots are

#7 +

-,

70 4

ajS-f -

-,

j8

Here
a#7

and the given relation


1

lubstituting

4-

is

xy =

for # in f(x)

Form,

Substitute

for the

same

for x.

-j-

ay,

is

then obtained by

0.

cubic, the
a;8

+ r; the transformed equation

-4w. ry 2.

e<i

(1

nation

a.& +

^7,

4- r) T/>

+ p (1 +

r)

q*)

r)

0.

- ^yr -

0.

(1

whose roots are


^87 4- ay.

Ana. y* - Zqy z + (pr

4-

-f

r1

Equation of Squared Differences of a Cubic.


Form, for the same cubic, the equation whose

Substitute

7 be the

a + 8

- 4pq +

-y

'

8r)y

-f

(JP

12r)^

(6r

roots of the cubic

a&
prove that the equation in y

3&r 2 + Zcx

-f

whose

7 -

-f <Z

0,

roots are

ya - &
y + a - 20'

or

+7 is

&'

for x.

Ans. (p*-4pq
4. If a,

a-

y +

roots are

_JL_ _

+7-a

81

2a

a& - yz
a +~j8

- 27

obtained by the homographic transformation

axy

42.

We

I (x

-f

-}-

y]

0.

Equation of Squared Differences of a CuMc.


now apply the transformation explained in the preced-

shall

an important problem, viz. the formation of the


whose
roots are the squares of the differences of every
equation
two of the roots of a given cubic. We shall do this in the first

ing

article to

instance for the cubic


a?

in which the second term

-f

qx

r = 0,

(1)

absent, and to which the general


Let the roots be a, /3, 7.

is

readily reducible.
equation
have to form the equation in y
is

We may

-f

We

whose roots are

here observe that the method of Art. 39 can be

applied to the solution of the general problem, viz. the formation of the equation whose roots are the squares of the
differences of every

when

two of the roots of a given equation

for

the product

{y-(a!-a 2 )

2
{

\y-(^-*tf\\y-(^-atf] ....

\y

(a a

a
ct 3

formed, the coefficients of the successive powers of y will be


symmetric functions of m, a 2 a*, 04, &c., and may, therefore, be

is

expressed in terms of the coefficients of the given equation.

In

82

Transformation of Equations.

the present instance, however, the method of Art. 41 leads more


This equation
readily to the required transformed equation.
"
may be called for brevity the equation of squared differences"
of the proposed equation.
Assuming y equal to any one of the
roots of the transformed equation,

y -

(]3

7)*

= a2 +

/3

e.

g. (j3

+ 72 -

7)

we have
-

a'"

also
2

a +

The equation

+ 72 =

/3

(#,

y)

cj3j

2<?,

r.

of Art. 41 becomes, therefore,


2;<

x
or
#3

4.

W +

- 2r =

20)3?
J /

(,,

;
r

we get

subtracting from this the proposed equation,

(?/ -f

q]y
^

x - or =

3r

x =

or

0,

2/

hence the transformed equation in

y
If

it

-f

6^///

9*/

y/

+ 4^ 3

//

is

4-

first

27 r2 =

0.

(2)

be proposed to form the equation whose roots are the

squares of the differences of tho roots

we

+^

remove the second term

(a, j3,

7) of the cubic

the resulting equation

is

the same as the equation of squared


differences of this latter, since the difference of any two roots
is unaltered by removing the second term.
can therefore

and the required equation

is

We

write

down

the required equation by putting

83

Equation of Squared Differences of a Cubic.

The

in the above.

result is

18-Zz

r^

a^

SljET 2

^o

#0

(to

which has

_~ (G*+ 4.ff

27

# +

,^V9

_.

0,

for roots

(4)

(-/3).

7 -a);

(/3- 7 ),

The equation

can be written in a form free from fractions

by multiplying the roots by a<?. It becomes then


& + 18Hu? + 81H*x + 27(G* + 4# 3 ) = 0,
whose roots are
2

We

(4)

can write

(/3-y)

a?(y-a)\

down from

(-/3)

(5)

2
.

an important function of the

this

roots of the cubic (3), viz. the product of the squares of the differences, in terms of the coefficients :
6

2
s
2
2
03-7) (7-) (-/3) = -27(G

-t4.H- 3 )-

evident from the identity of Art. 37 that


z
have in fact
as a factor.
a
contains
It

(6)
6r

is

477 3

We

G~

477 = a* {ftvds <5

-f

Tlie expression
cubic,

and

is

111

Grtutfi

brackets

represented by

&

-f

#2^3

+ 4^

^ 33

-f

4a^a 3 - 3a?aJ\.

called the discriminant of the

is

giving the identities

JW -

47P s rvA,

tfo,/-

A.

EXAMPLES.
1.

Form

the equation of squared differences of the cubic

s3

7*

0.

-4iw. x*
2.

Form

x 3 + 6x~
First

- 42# 2

-f

-f

7x

Form

0.

0.

0.

+ 6# 2 + 9x +

225s - 68

0.

^w.
What

-f

the equation of squared differences of


x*

4.

- 400 =

remove the second term.


Ans. x* - SO* 2

3.

441a;

the equation of squared differences of

iP

18a; 2

+ Six =

conclusion with respect to the roots of the given cubic can be

from the form of the resulting equation in the

last

example ?

0.

drawn

84

Transformation of Equations.

Criterion of the Nature of the Roots of a Cubic.

43.

We

can from the form of the equation of differences obtained


in Art. 42 derive criteria, in terms of the coefficients, of the
nature of the roots of the algebraical cubic. For, when the
equation (5) of Art. 42 has a negative root, the cubic (3) must

have a pair of imaginary roots, in order that the square of their


difference should be negative; and when (5) has no negative
root, the cubic (3) has all its roots real, since a pair of imagi-

nary roots of (3) would give rise to a negative root of (5).


In what follows it is assumed that the coefficients of the
equation are real quantities. Four cases may be distinguished
:

+ 4jff is negative, the roots of the cubic are all


(1). When G
real.
For, to make this negative II must be negative (and
4IP > (?") the signs of the equation (5) are then alternately
2

positive and negative, and, therefore (Art. 20), (5) has no negative root and consequently the given cubic has all its roots real.
is positive, the cubic has hco imaginary
(2). When G' + 4H*
For the equation (5) must then have a negative root.
roots.
z
+ 4ZP = 0, the cubic has two equal roots. For
(3). When G
In this case
the equation (5) has then one root equal to zero.
=
a
not
vanish.
that
does
We
assumed
it
A 0, being
may say,
;

therefore, that the vanishing of the discriminant (Art. 42) expresses


the condition for equal roots.

When

(4).

For the
tions

may

6r

0,

andH=Q,

the cubic has its three roots equal.

These equaroots of (5) are then all equal to zero.


as
can
in
the form
be
also be expressed,
easily seen,
#0

which relations among the

#2
'

al

a->

coefficients are therefore the condi-

tions that the cubic should be a perfect cube.

44.

Equation of Differences in General.

problem

The general

by the aid

of symmetric functions, of
roots are the differences, or the squares of the

of the formation,

the equation whose


differences, of the roots of a given equation,
Let the proposed equation be
follows:

f(x)

(0

i)

(x

(x

a,)

(a?

may

be treated

- an =
)

0.

as-

85

Equation of Differences in General.


Substituting x + a r for x
-

n, in succession,

/(.r +

(a?

aO - x (x +

+ a2)

3x

d!

and giving

r the values 1, 2, 3,

we have the equations


- a 2 ) (a? +

(x + a 2

cii

a 3)

(#+a!-a n ),

- a 3 ) ....
ai) (a? + a a

+ a a - a n ),

(x

(i)

Also, employing the expansion of Art. 6,


f(a r ) = 0, we find the equation

a r)

and observing that

-f (ar}

side of this equation by $ (x, a r ), and


of
the identities (1), we obtain
both
sides
multiplying

Denoting the second

tf>

(x, cii) ^>

(j?,

a a ) ----

<t>

(a?,

a w)

{#

(ai

a2)

2
}

{%*

(ai

a3)

2
.
}

To form

the equation of differences, therefore, we can mula 2 ), &c., and substitute


tiply together the n factors
(x, ai), ^ (a?,
for the symmetric functions of the roots which occur in the pro</>

duct their values in terms of the coefficients.

Or we may,

as

already explained in Art. 42, form directly the product of the


^n (n - 1) factors on the right-hand side of the above identity,

and express the symmetric functions involved in terms


coefficients.

degree in

a?

of the

of the resulting equation of the n(n - \) th


are equal in pairs with opposite signs.
Since in

The roots

x occurs in even power^ only, we may substitute


2
a
and
thus
obtain the equation of the \n (n - l) th degree
y
whose roots are the squared differences.
this equation

for

For equations beyond the third degree the formation

We

of the

shall give the


equation of differences becomes laborious.
'result in the case of the general algebraic equation of the fourth
degree in a subsequent chapter.

86

Transformation of Equations.
EXAMPLES.
1.

are a,

The

roots of the equation

form the equation whose


02

+7

roots are
2

2
7 +

2
,

o-

32

-f

^w*. y
2.

The

o,

8
-f-

2#2

form the equation whose

L + L.-L
^

JL

^T~^'

The

a, /?,

-f-

246y - 660 =

roots are

i_JL
^-07

^3

L-L

I +

^
y

12y

-f

/3'

172y

- 2072 =

form the equation whose roots are

+ &y + T2

ya

-f

-f

a2

a2 ,

-f a/3 -f

-4w. (y
4.

The

a,

+ pxz

qx

-f

a2

a2

(jt?

If a, 0,

7 be the

prove that

- y)(y -

(ft

0.

a2

4jP

-f

*.

8pr)
-f

8;>y -

(1

a) (a

+ + a 2 )* + 1 -f 3rt -f 3 2
- &} is a rational function

-}-

2fl

3 ==

Find the relation between


o#
its roots

and

3i^

are so related that

7, 5

0.

9(1

a-f a 2 ).

H of the cubic
+ 32^ +
7)

2
,

(7

o)

(a

#)

are in arithmetical

An*.
,

progression.
7. If a,

2
IGjo^r + 8r

of a.

Am.

when

roots of the cubic

#3 - 3

6.

qY

form the equation whose roots are


2

5.

-f

roots of the cubic

x9
being

0.

roots of the cubic

s3 + qx + r
are

28y

3tf -f 1

^s.
3.

roots of the cubic


a:

are

#2

-f

2jff 3

0.

be the roots of
;

0,

find the value of

(P - 7 2 ) 2 (' ~

52) 2

(7

- a2

2
)

(0

5 2) 2

-f

(a

- 0*)

(7*

- *2).
*.

0.

87

Examples.
Prove that,

8.

if

07 + 70 + 00

- 7)*( -

{(0

18

8)2

(0

(y

a) (0

8)

-*+ 8s

which has one root of the form

aV

other factors are

The roots

10.

(a?

and

1)

a,

3,

15

S)*}

(a

2 2
)

(7*

- $*)}

0,

for

we

find that

Hence the

2.

a must

factor

satisfy

x 2 - 2a? +

6,

3), as is evident.

(#

of the cubic

a G x*
are

8 2 )*

1.

Diminish the roots by 1 substitute a V


a4 - 3a2 - 4 = 0, and a* - 6o2 + 8 = ; hence a

The

0,

0)*( 7

a2 ) 2 (0 2

9*

78

-f

- 7*) 2 (a2 - V? +- ( 72 -

Solve the equation

9.

a& + 05

+-

30i a?

3a2 #

-f

form the equation whose roots are

+ 7,

-f a,

0.

We give here

This question has been already solved in Art. 41.

another solu-

the simplest, will be


Let the roots of the given equation be dimi-

tion which, although in this particular instance it is not

found convenient in many examples.

The transformed equation

nished by A.

aoy* +

whose

roots are o

A,

3-4 iy2

7-

/*,

h.

is

(Art. 35)

t>A z y

should have two roots equal with opposite signs.


Art. 24)
z

This equation

0,

"We express the condition that

This condition

this equation

is (see

Ex.

17,

0.

a cubic in h whoso roots are

is

for the above condition is


(0

h)

27*

(7

h}

0,

or
-f

7,

From the equation in h


0, 7 represent indifferently any two of the roots.
the required cubic can be formed by multiplying the roots by 2.
where

11.

The

roots of the biquadratic

0o2

are o, 0, 7, 8

+ 4#i# 3

+-

6#2#2

4- 4tf 3

0i

form the sextic whose roots are

7,

7 +

a,

Employing the method of Ex.

+-

0,

+-

5,

5,

7+8.

10, the required equation can be obtained

from

the condition of Ex. 20, Art. 24.

The

condition

is

in this case

A\AiA* - A\*Ai - aoA*2 =

0.

a sextic in h whose roots are J (0 + 7), &c., from which the required
equation can be obtained as in the last example.

This

is

88

Transformation of Equations.
Form, for the cubic of Ex.

12.

fly

- a

+7-

jya
7 + 0-2/8'

2a'

whose

10, the equation

y*

- 2y

roots are

'

A, and express the condition that the resulting cubic


in geometric progression (see Ex. 18, Art. 24).
The con-

Diminish the roots by


should have

its roots

dition is
0.

This will be found to reduce to a cubic in h

whose roots are the values above

written, since
,

Form

13.

2/3 7

Dimmish
should hare

same cubic the equation whose

for the

- gy

qj3

-f

7 - 2a

2y a
'

'

- 7)8
- 2y

701

4- ft

We

_
~

+
JJT^A
2/3 7

or

The equation

in h

7-

7 - 2a

found

is

^ 3 2 - SAiAzAa +
will be

A'

-a/3-7

/8 -f

The

*ci

the roots by A, and express the condition that the transformed cubic
roots in harmonic progression (see Ex. 19, Art. 24).
have
2

14.

its

a- A

which

2a

7
4-

T-

roots are

2Az*

0,

to reduce to a cubic.

roots of the biquadratic


(IQX* 4-

are

a,

7, 5

find the cubic

whose

ya ~ &$

0)8-7?

7-f a-jS-"5'

o-f/8-7-"

fry- aS

+ 7-0-^5'
Diminish the roots by
condition

is

A,

roots are

and employ the condition of Ex.

A?At which reduces


16.

to a cubic in

Tne

also

= - =

/(a)

0,

+ qx
by

-f

ratios of the roots of the cubic

0.

elimination.

the ratio of two roots;

h whose roots are the values above written.

general problem can be solved

equation, and p

f (pci)

a Q A<?

Find the equation whose roots are the


a?

The

22, Art. 24.

in this case

and the required equation

Let /(*)

then since f(B)


in p is obtained

be the given

0,

we have

by eliminating

89

Examples.
between these two

For the cubic in the present example the

latter equations.

result is

16. If a, 0,

7 be

the roots of

a?Hp*2 +

#-f r

= 0,

form the equation whose roots are

Ans. ^-2(p*
17.

Form

for the

same cubic the equation whose

&

y
-d(*.

18. If a,

3
rZtf

- (pqr -

7 be the

3r2

)#

roots are

a"

7'

- bpqr -f 3r2 + #3
-f (pV
- (jp* ?2 - 2^V +

a?

4jt?^r

- 2? 3 - r2 =
)

0.

roots of the cubic

a?

+ qx +

= 0,

form the equation whose roots are


la

-4*.
19. If a,

j8,

t/

7 be the

m#7,

-f

#J

my a,

-f

- wyy2 + (i2 ^ -f

$lmr)y

ly

-f

mojS.

- Pm q 1 - Mm2 qr - m3 ra = 0.

l*r

roots of the cubic

o#

+ 30i#2 +

30 2 # +

0,

form the equation whose roots are

(a-/3)(- 7),

(3-7)

(r-)<7--

(3 -a),

Ans. y

,
3

+ -r-

y*

-- i

o*

20.

Form,
(/8

for the cubic of

_ 7 )3 (2a _0_ 7 )2,

Ex.
(

19, the equation

7 -a)

required equation can be obtained


differences of the cubic (4) of Art. 42, since

21.

Form,

of squared

7 -a) -(a-/3) 2 =OB- 7 )(2a-j8-7).

for the cubic of

Ex.

a()8-7)

2
,

16, the equation

0(7 -a)

2
,

Ans.

Form, for the same


a

M = 0.

Q = # 3 - 9pqr -f 27r3 + pV,


= - r(4^ 3 + 27r2 -f 4p*r-p* q 2 - ISpqr).

cubic, the equation

+ 207,

-f

roots are

a.

P = py-9r,
J2

22.

whose

7 (a-)

Let the transformed equation be # 3 4- Px* + Qx

0.

(-0-'(27-a-jS).

by forming the equation

whose roots are

(2j8- 7 -),

The

5
6

whose

roots are

CHAPTER

V.

SOLUTION OF RECIPROCAL AND BINOMIAL EQUATIONS.


It has been

Reciprocal Equations.

45.

shown

in Art. 32

reciprocal equations can be reduced to a standard form,


in which the degree is even, and the coefficients counting from
the beginning and end equal with the same sign.
now
a
that
to
the
standard
reciprocal equation of
form
proceed
prove
can always be depressed to another of half the dimensions.

that

all

We

Consider the equation


m~l +
a x* m +
ai&*

a m xm

4-

-f

a\x +

r?

0.

m
Dividing by # and uniting terms equally distant from the
extremes, we have
,

Assume # -f - =

s,

Vp

and

let

We have plainly the

xp +
a?

be denoted for brevity by

relation

Giving p in succession the values 1, 2,


s
F, = Fi s - Fo = c - 2,
= F2 s- FV=s 3 - 3s,
3

F
F
F

fi

and

so on.

=
=

F s- F
F s~F
3

=
=

3, &c.,

-4s
-5^-f

we have

-\

2,

5s;

Substituting these values in the above equation, we


m th degree in s; and from the* values of

get an equation of the

x those of x can be obtained

by solving a

quadratic.

91

Examples.
EXAMPLES.
1

Find the roots of the equation


**

Dividing by

x+ 1

za

-f

+ x9 + x +

= 0.

we have

(see Art. 32),

& + x* + 1 = 0.
This equation

may

be depressed to the form


*2

- 1 = 0,

x + --i

whence

giving

# + - = -!,

and the roots of these equations are

~1

/Tls

/~^B

2
2.

Find the

Dividing by

roots of the equation

a?*

1,

which may be done

briefly as follows (see Art. 8),

1-3 6-6 3-1


1-2 3-2
-2
3-2

we have

the reciprocal equation


1

Substituting for Fi,

4s2

**

* -6z*
i

z2

whence

+2
+

and for

and

3,

(z

0,

2,

\/ 3,
/T:

roots of these equations are

y/^l

- \/3

Solve the equation


a:

Dividing by x

we

from which we obtain

have

y/

These roots are double roots of the equation

- 1=0.

(1)

we have

2
-3) =0,

y/3

3.

z2

V^

or

#+-

giving

and the

0,

(1).

the equation

92

and Binomial Equations.

Solution of Reciprocal
we have

Solving this equation,

the quadratics
~

0,

from which we ohtain

*
where

0*

= *{-

6</&

1 -f

(10

20

1.

This expression gives the four values of x.


4.

Find the quadratic factors of


x6

Transforming

this,

Z8

whence

The quadratic

0.

and

0,

32

=
z

0,

a?

0,

x -s/3

(!+*)

*(!+**),

0.

(l+*)* = a(

(2).

Reduce to an equation of the fourth degree in

-<**. (1

In

-f

Solve the equations


(1).

6.

factors of the given equation are, therefore,

x2
6.

we haye

46.

Binomial

this

and the following

<i)*

-f

(7

is

articles will

a =

thi

1,

and therefore
am

is,

The same
case

is

(4

a)= 0.

Properties.

be proved the leading


1

x*

0, then a

must

is

m =

(u

1, or (a

m
)

1;

a root of x - 1 = 0.
n

true of the equation %n + 1 = 0, except that in


be an odd integer.

47. PROP. II.

m - 1 =
equations x
unity.

a root,

that

3^)c

General

Ecjiiatioiis.

general properties of binomial equations.


PROF. I. // a be an imaginary root of
also will be a root, m being any integer.
Since a

-f

If

m
x n

0,

and n be pnmf^jQ, ftaeh other, the


1 =
have no common root except
.........

Binomial Equations.

To prove
numbers
and

we make

this

If m and n

is

use of the following property of

be integers

1.

turned into a continued fraction,

preceding the
r

Now,
equations

final restoration of

each other, integers a


For, in fact, when

is

r-

the approximation

._

be any

let a

if possible,
;

to

prime

mb - na -

b can be found such that

93

common

root of the given

then
a m = 1,
a mb =

therefore

and a w =

1,

a ('"*-'"*) = 1, or a

whence

and ana =
l

1,

or a = 1

only root common to the given equations.


48. PROP. III.
If k be the greatest common measure of two
and
m and the roots common to the
xm 1 -=

that

is,

1 is the

equations

n,

integers

xn -

0, are roots

To prove

0,

- 1 = 0.
of the equation #*

this, let

m = Aw',

n = kti.

and n are prime to each other, integers


may be found such that m'b - na = 1 hence

Now,

since m'

and a

mb
If, therefore, a

be a

- na =

k.

root of x m

common

-1 =

0,

and xn -

0,

which proves that a is a root of the equation #*-! = ().


49. PROP. IV.
When n is a prime number^ and a any
n
root
x
1 = 0, all the roots are included in the series
of
imaginary

For,
of

them be

by

sarily

11

'1
.

I., these quantities are all roots of the equaare


all different ; for, if possible, let any two
they

p
equal, a

= a?

whence
but,

by Prop.

And

tion.

a,

1,

p ~^

-=-

Prop. II., this equation is impossible, since

prime

to (p

q),

which

is

number

less

than

n
n.

is

neces-

Solution of Reciprocal

94:

50.

PROP. V.

When n

a composite number formed of the


- 1 = 0, a^-l==0,
of
n =
the equation x
0.
1

0, &c., all satisfy

is

the equations #?

factors p, q, r, &c., the roots


of

and Binomial Equations.

For, consider a root a of the equation a?


from which we derive

then a p

which proves the proposition.


51. PROP. VI.
When n is a composite number formed of the
are
prime factors p, q, r, &c., the roots of the equation #" - 1 =
the n terms of the product
(l

+ a + a*+... +

where a

is

a?-

a root of x*>

We prove this

r- l

)(l+(3 + ... + &- )(I+y+...+y )... y


- 1 = 0,
= 0, y of of - 1=0, &c.
/3 oftf-1

for the case of three factors JP,

q, r.

similar

a b c
Any term, e.g. a fi y , of the product
n
bn =
is evidently a root of the equation x - 1 = 0, since a*" = 1, fi
1,
a b c n =
=
And no two terms of the
1.
1, and, therefore, (a fi y )
7*"

proof applies in general.

a b c
product can be equal for, if possible let a [3 y be equal to
b ~b
c-c
a '~ a =
b
a
c
another term a 'fi 'y '; then a
p 'y '. The first member
of this equation is a root of xp - 1 = 0, and the second member
is a root of aP" - 1 = 0.
Now these two equations cannot have a
;

common root since p and qr are prime to each other (Prop. II.);
hence aa fi b y c cannot be equal to a a 'fi b 'y c'.
52. PROP. VII.
The roots of the equations? -1 = 0, where
n = p a q b r c and p, q, r are the prime factors of n, are the n products
,

vf the form a/3y, where a


c
and y ofof = 1.

is

a root of xp

1, /3

a root of x q

1,

an extension of Prop. VI. to the case where the prime


more than once in n. The proof is exactly similar.
n
n
n
such
Any
product aj3y must be a root, since a = 1, /3 = 1, y = 1,
b
a
c
n being a multiple of p q r
and a proof similar to that of
Art. 51 shows that no two such products can be equal, since
This

is

factors occur

We

r* are prime to one another.


have, for convenience,
stated this proposition for three factors only of n.
similar

proof can be applied to the general case.

From this and the preceding propositions


to derive the following general conclusion :

we

are 'now able

The Special n th Roots of Unity.

95

The determination of the n th roots of unity is reduced to


where n is a prime number, or a power of a prime number.

The Special Roots of the Equation

53.

the case

^-1

0.

Every equation x" 1 = has certain roots which do not belong


Such roots
to any equation of similar form and lower degree.
th
n
roots
we call special roots* of that equation, or special
of unity.
If n be a prime number, all the imaginary roots are roots of this
th
If n =*p a , where p is a prime number, any n root of a

kind.

lower degree than n must belong to the equation #p


-1=0,
1
since every divisor of p a is a divisor of p"" (except n itself) ;

hence there are p a ( 1

If,

=
again, n

],

- 1 =

which belong

to

no lower degree.

where p and'g are prime to each

and? 6

p)
x*

roots

p)

V
a

special roots

other, there

ofa^ - 1 =

0,

and

q/

0, respectively.

Now,

if

a and

/3

be any two special

n
is a special root of x - 1 =
for if
;
m =
not, suppose (a/3)
1, where m is less than n; we have then
m is a root of
w but a m is a root of a* a - 1 =
am
0, and

roots of these equations, a/3

/3~

-1

/3"

and these equations cannot have a common root


other than 1 as their degrees are prime to each other
consen
is
a
m
than
and
cannot
less
be
root
of
f
a/3
special
quently
xn - 1 = 0. Also, as there are
xq

0,

PJ

such products, there are the same number of special n th roots.


This proof may be extended without difficulty to any form of n.

All the

roots

n
0/# - 1 =

are given by the series 1, a, a 2 ,

th

a*" 1

any special n root. For it is plain that a, a &c., are


And no two are equal for, if ap = a 9 a^~ = 1 and
all roots.
th
therefore a is not a special n root, since p - q is less than n.

where a

is

ff )

When

one special n th root a

is

given,

we may obtain

all the other

th
special n roots of unity.

* The term

"

"

'
is here used in preference to the usual term
primitive root," since the latter has a different signification in the theory of numbers.

special root

96

and Binomial Equations.

Solution of Reciprocal
Since a

different

"1

a special root, all the roots 1, a, a 2 , ... a*


are
if
as
we
have
we
and
select
a
roots,
just proved
1

is

n th

root a? of this series,

where p

is

a?, a *,.

aC'-

to n, the roots

prime

a w^(=l)

)*,

are all different, since the exponents of a when divided by n give


different remainders in every case ; that is, the series of numbers

n - 1 in some order

0, 1, 2, 3, ...

whence

this series of roots is

the same as the former, except that the terms occur in a different
order.
To each number p 9 prime to n and less than it (1 inth
for a mp cannot
cluded), corresponds a special n root of unity
be equal to 1 when
is less than n, for if it were we should
;

have two roots in the

series

equal to

give all the roots in that case

and the

1,

therefore a?

is

series could not

not a root of any

binomial equation of a degree inferior to n that is, ap is a special


u th root of unity.
What is here proved agrees with the result
;

above established, since the number of integers

prime to

by a known property

it is,

when n=p

which

n
special roots of #

than n and

nil

as above proved, the

is also,

of numbers,

less

If

number

of

0.

EXAMPLES.
1.

To determine

Here, 6
it

the special roots of

= 2x3.

are roots of x*

and dividing x*

-f 1

by

the special roots of x 6

-c

0.

9
Consequently the roots of the equations x

-1

Now,

0.

or x

dividing x*

we have

1,

x~

0.

Solving this quadratic, the

oots are

also since

which may be

The

aai

ai

6
,

5
,

easily verified.

special roots are, therefore,


a,

a5

or ai 5 , 01

or a, -

by
-f 1

a?

0,

and

0,

which determines

we have

a:

97

Examples.
2.

To

discuss the special roots of

a?

M-

= 0.
12

Since 2 and 3 are the prime factors of 12, and -2

# -1 =

and s - 1 = 0,

12

12
~

8,

**

4,

the roots of

a -

12

1 by x*
1
are roots of x
1,
0,
;
now, dividing
and a? - 1, and equating the quotients to zero, we have the two equations
* 8 + xt + 1 = 0, and a* + 1 = 0, both of which must he satisfied hy the special
4
1
roots of # 12
1 =
therefore, taking the greatest common measure of oP + a;
and aft + 1, and equating it to zero, the special roots are the roots of the equation
4

-I-

x*

a?

The same
least

common

* 4 - x2 +

0.

would plainly have heen arrived at hy dividing a; 13 - 1 by the


4 - 1 and a?* - 1.
Now, solving the reciprocal equation
multiple of s

result

= 0, we have x 4- - =

V3

if

whence,

a and

ai

be two special roots

2
are the four special roots of * 12

We proceed now to express the


Since

+-

we

take aai

1 (as

ai

may

0, or (a -f ai)

ai

#8

0,

therefore be expressed

o8

by the

we have

a.

0,

assigned to a and ai)


8

1,

any one of them

aai/

consistent with the values

since a and ai are roots of

a, ai,

-f ai -f

= 0.

four special roots in terms of

and a

ai.

The

and

roots

series a,

a5, o7 , a11 , since o 12 *

1.

11
5
7
Further, replacing a by a a a , we have, including tbe series just determined,
the four following series, by omitting multiples of 12 in the exponents of a :
,

a5 ,

a,

a7 ,

a 11 ,

o,

a5 ,

a",

a7,

a6 ,

a,

a",

where the same roots are reproduced in every row and column, their order only
being changed. We have therefore proved that this property is not peculiar to any
one root of the four special roots and it will be noticed, in accordance with what
;

above proved in general, that


and less than it. "We may obtain

is

one of the four special roots

a,

1, 6, 7,

all

and 11 are

tbe roots of # 12

a 6 , a7 , a 11 , as follows

all

1
:

the numbers prime to 12,


=
by the powers of any

98

Solution of Reciprocal

Prove that the special roots of #16 -

>.

Show

4.

obtained

5.
a;

21

and Binomial Equations.


=

are roots of the equation

that the eight roots of the equation in the preceding example may he
roots of a* +
+ 1 = by the four roots of

by multiplying the two

Form the equation


= 0, and reduce it

of the 12th degree whose roots are the special roots of


one of half the dimensions.

to

*-*-6*

Am.

+ 6*' + 8^-8*+l =

0.

54. Solution of Binomial Equations by Circular


Functions. We take the most general binomial equation

#" = a + b

where a and

ft

a~R cos a,

Let

x =

then

now,

J-

1,

are real quantities.

It (cos a

we

be a root of this equation,


rn (cos w0

+ J- 1

R sin a

+ J- 1 sin a)

+J-1

r (cos

if

sin 9)

have, by

sin nO)

De

Moivre's Theorem,

R (cos a + J- 1 sin a)

and, therefore,
rn cos nO =
r"

Squaring these two

sin nO =

equalities,

r8n =

2
,

R cos a,
R sin a.

and adding,

n =
giving r

where we take R and r both positive, since in expressions of the


kind here considered the factor containing the angle may always
be taken to involve the sign.

We have then
cos

nO = cos

a,

sin

nO =

sin a

and, consequently,

nO = a

-f 2/C7T,

k being any integer; whence the assumed n th root

is

of the

99

Solution by Circular Functions.


general type
a +

Jit cos

2/C7T

p^1
+ J-

a + 2for\

sin

this expression any n oonseoutive values in the


numbers between - oo and + oo we get all the n th roots
and no more than n, since the n values recur in periods.

Giving to k in

series of

We may

write the expression for the n th root under the

form
*\T>{
Jltl cos-

If

we now suppose
n

becomes # =
root of 1

J-

ir

cos

R=

and a =

1,

or unity,

n J

the equation 3? = a + 5

0,

J-

+ ^/- 1

lT

Sill

--n

definite value, for instance zero,


1-7-

Jit
^

/
{

a
cos /ii

+ l>
J- 1

a\
sin .

'

one n th root of

+ b J~ 1.
The preceding formula shows, therefore,
of any imaginary quantity may be obtained by

is

is

fTT
----

we give k any

1 sin

the general type, therefore, of an n th

COS

If

+J~

nj}\

J-

-f

/-;
J-l sm-

-f

of them by

Taking

multiplying any one

tu
the n
roots of unity.

in conjunction the binomial equations

xn - a +

we

that all the n th roots

n
J-l, and x = a

J-l,

see that the factors of the trinomial

x 2n - 21t cos a x"


.

are

a +

cos

where k has the values

----

rir

0, 1, 2,

ls

/- 1 sin

3 ... n -

H2

It

-f

1.

100

Solution of Reciprocal

and Binomial Equations.

EXAMPLES.
1.

Solve the equation X 1 1, this is

Dividing by

Assuming

= x -f -, we

0.

reduced to the standard form of reciprocal equation.

obtain the cubic


*

z*

2z

0,

from whose solution that of the required equation


2.

Resolve (x

-f I)

-x -

obtained.

is

1 into factors.

Ans. 1x (x
3.

-4-

2 -f

1)(#

-f 1)*.

Find the quintic on whose solution that of the binomial equation x 11 -

depends.

Ans.

2*

- 4z3

- 3*2 + 3* -f 1 = 0.

4. When a binomial equation is reduced to the standard form of reciprocal


2 1, x -f 1, or x
equation (by division by x
1), show that the reduced equation
has all its roots imaginary.
(Of. Examples 15, 16, p. 33.)
6.
ss a:

When
+-

- 2 and

reduced reciprocal equation

this

show that the equation

is

transformed by the substitution

in z has all its roots real, and situated

For the roots of the equation in x are of the form cos a + V


Art. 54)

hence x

between - 2 and
6.

Show

+ -

is

of the form 2 cos

a,

and the value of

- x+

I)

is reciprocal,

- 27*2 (x -

a
l)

Am.
Exhibit

The

sin a (see

this is real

and

2.

that the following equation


4 (*2

7.

between

2.

all

the roots of the equation #9

and solve

it

0.

Roots:

2, 2,

J,

-I,

-K

= 0.

solution of tbis is reduced to the solution of the three cubios

where w, w2

are the imaginary cube roots of unity.

sented as follows

The nine

roots

may be

repre-

2
the other six roots are special roots of the given equation
Excluding 1, w, u>
and are the roots of the sextic
;

Xs
8.

0.

th
Reducing the equation of the B degree in Ex.

tion

= #+-, we

obtain

3, Art. 63,

by' the

substitu-

101

Examples.
prove that the roots of this equation are

Zcosii'.

2 cos?-*

2oosg,

2eo.jj,
Reduce the equation

9.

4# 4 - 85#3 4 357s2
to a reciprocal equation, and solve

340*

64

it.

= -+-.

Assume

Ans. Roots:

4,

1,

J,

16.

10. Solve the equation

x*

4- wjfliC

4 w 2 ##2

m*px

-f

4-

m = 0.
4

Dividing the roots by m, this reduces to a reciprocal equation.

an imaginary root of the equation x


prove the relation

11. If a be

number

(1

Show

12.

a)(l

a2 )(l - a n

(1

a"- 1 )

0,

where n

a prime

is

n.

that a cubic equation can be reduced immediately to the reciprocal


relation of Ex. 18, Art. 24, exists amongst its coefficients.

form when the

Show that a biquadratic can be reduced immediately to the reciprocal form


the relation of Ex. 22, Art. 24, exists amongst its coefficients.

13.

when

Form

14.

the cubic whose roots are

where a

is

an imaginary root of

a6

a3 4 a 4

x 1

o2 4 a5 ,

Ans. x* 4 x z - 2#

0.

0.

the roots of this cubic are known, the solution of the equation x 1
1 =
be completed by means of quadratics. For, suppose the three roots to be

When
may

#2, #3

a?i,

then a and a* are the roots of xz

and a 2 and o 6 of xz

x$x

are all real, and they

41 =

may be

0.

x\x

41 = 0;

a 3 and a 4 of x2

41 = 0,

x^x

It is easy to see that the roots of the cubic

readily found approximately

by the methods

of

Chap. X.
15.

Form

the cubic whose roots are

where a

As
are

is

a8

je

a2 4 a s

in the preceding example,

known, the

4-

an imaginary root of x n

o12 and
x*x

x\

o8
-*

a 11

0.

when

a 10

4-

0,

Let x\ y

#2, #s

a*

4-

x'6

a 9 and a8

x*

a9

4 a7

- 4x

4- 1

0.

the roots of the cubic (which are

# 13

may be

be the roots of the cubic.

a8 are the roots of xz

and

a6

a4

Ans.

solution of the binomial equation

solving quadratics.

4 a 12 4 o6

- x\x

4-

a7 of x1

x&

aa
-f

12
6
quadratics are solved, each pair of roots a, a ; a8 , a , &o.,
volution of another quadratic, as in the preceding example.

all real)

completed by

It is easily seen that

a11 and o3

= 0.
may be
#j

When

a10 of
these

found by the

102

Solution of Reciprocal

and Binomial Equations.

1 = 0.
to quadratics the solution of # 17
form
we
the
of
the
one
a
roots,
quadratic whose roots are
imaginary
Calling

16.

We

Reduce

ai

= a +

02

aiaz = 4 (ai

easily find

10

a 13

o+

15

-f

o 13 -f a 15

+ a 16

11

14

-f

-f

-f 02) =

and may he found by solving


1

a*

-f

-f

hence

ai

this quadratic.

a l6
a

a4

4-

4 a

and

-f-

-f

a4
a

12

+ a2

a6

-f

az are the roots of x*

~ a3
a

72

a8

-f

x-4

= 0,

Assuming, again,

71

-f

-I-

11
10

-f

a5

11

a 14

+ a 12
-f

a6 ,

1 = 0, and
0.
1
seen that 1, &-z are the roots of x
a\x
c^'
71, 72 of xSeparating again each of these into two parts, and forming the quadratic whose
roots are, for example, a + a 16 and a 13 -f a 4 the sums of the roots in pairs are
1

it is

obtained

and

finally the roots themselves,

by the

solution of quadratics, as in the

preceding examples.
This and the preceding two are examples of Gauss's method of solving alge- 1 = when n is a prime number. The solution
braically the binomial equation
of such an equation can be made to depend on the solution of equations of degree
1. When n= 13,
not higher than thw greatest prime number which is a factor in n
1 being = 3*2 2 in that case ; and
e.g. the solution depends on that of a cubic, n
4
1
when n
17, the solution is reducible to quadratics, n
being then = 2 . For
the application of Gauss's method it is necessary to arrange the n - 1 imaginary
roots in a suitable order in each case according to the powers of any one of them.

'J

number n

primitive root" of a prime

possesses the property that

when

raised

to n
2 inclusive, and divided in each case by
powers from
remainders are all different.
(See Serret's Cours d'Algel^e Superieure,

to successive

the n

There are several such primitive roots of any prime number


Gauss arranges
13, and 3, 5, 6, 7, 10, 11, 12, 14 of 17.
the imaginary roots so that the successive indices of any one of them, a, are the
2 of any primitive root of n.
successive powers from
to n
Taking, for example,

vol. n. sect. 3.)

e.g.

2, 6, 7,

and 11 of

the lowest primitive root of 13, and dividing the successive powers of 2
get the following series of remainders

and

12

11

10

by

13,

we

7;

powers of a in order when the indices

these, therefore, are the successive

which exceed 13 are reduced by the equation a13 = 1. If the lowest primitive
same way, we get the following series of remainders

root of 17 be treated in the


1

On comparing

10

13

5-

15

16

11

14

12

these series with the assumptions above made,

that in the former case, viz.

n=

13, the

it

6.

will be observed

twelve roots were divided into three sums

two sums of eight each. The method of


1
and it is not
any case depends on the nature of the factors of n
difficult to show in general that the product of any two such groups is equal to the
sum of two or more, as the student will have observed in the particular applications
of four each,
partition in

given above.

and in the

latter case into

103

Examples.

The lowest primitive root in any particular case is the only one necessary to be
known for the application of Gauss's method and this can usually he found with;

out difficulty hy

may

It

trial.

be observed that one or other of the three simplest

prime numbers 2, 3, 6, is a primitive root in the case of every prime number less
than 100, with the exception of 41 and 71, whose lowest primitive roots are
Methods of finding all the primitive roots are given in
6 and 7 respectively.

work above

the section of Serret's


17.

Find by

the equation # 19

trial

referred

to.

how

the lowest primitive root of 19, and hence show

to solve-

0.

It is readily found that 2 is a primitive root, and the remainders after division
19
are given in the process of trial.
Since 18 = 3 2 .2, the solution will be
by
effected by cubics and quadratics.
The first cubic is found by forming the
equation whose roots are

Show

+
+ au +

o8

-f

16

13 4
-f a

a 18

a1

a9

17

15

-f

a 11

4-

-f

a 12 ,

4-

a 5,

4-

a 10 .

whose degree is a prime number the


whose solution depends on quadratics .is 2* 57 -1=0.
The next prime number after 257 which satisfies the condition that n
1 is a
power of 2 is 65537. We have therefore the series 3, 6, 17, 257, 65537, &c.
and Gauss remarks (Disquisitiones Arithmetics, Art. 366) that the division of a
circle into n equal parts, or the description of a regular polygon of n sides, can be
effected by geometrical constructions when n has any of these values.
18.

that of binomial equations

lowest after # 17

19. If 01, 02, 03

...

4-

<*

pix~

form the equation whose

4-

jt)

a^-2

-f

p n~\x

4-

4-

pn -

0,

roots are

--1

ai H

1
,

as

4-

ai

We

be the roots of the equation

an

have here the identity


4-

and changing x into x

...

+ pn-\X

-f

pn s

(x

a\)(x

...(*-

+ jt?2*2 + Pi* +

Pn (*
\

--

(*

01 /

right-hand side take the

form x

-\

---

a+ -

--

a;**,

the factors on the

and assuming x

4-

2,

the

left-

by means of the

Art 45

Fimi the value of the symmetric function 2aa

equation

side can be expressed as a polynomial of the n th degree in *

20.

2/

Multiplying together these identities, and dividing by

relations of

On )

(see Art. 32),

hand

O2)

(7

0.

&)

of the roots of the

104

and Binomial Equations.

Solution of Reciprocal

This can be derived from the result of Ex. 19, p. 52, by changing the roots into
2

1 \
-

of the transformed equation, and multiplying

(1a
2
2
by a j8 V* which

equal to

is

p/

~
Ans

From

<7

2a 2 /8 2 (7~

5)

48 (a8 2

the values of the symmetric functions given in Chapter III. several others

can be obtained by the process here indicated.


2

Find the value of the symmetric function 5(ai

21.

a2) as 04

2
.

a,

of the

roots of the equation

oox*

naix*-

2#"1

We easily obtain
into their reciprocals

Show

- a.*f =

5(oi

- l)(i 2

a
(

na n -\x

oa)

-I-

an

0.

and changing the roots

we have

22.

"

aa)

ajW

2
.

a,,

fi

2
(

that the five roots of the equation


3

#5

v^

are

+ j/b,

Sjpjr

et/a
a4

-f

bp*x

2y

-f

-f

^v7 ^,

^2

-f

B %/!>,

& y~a +

%/a '+

^ yj,
0-

i,

where v^ a ^ = P> a 4 i = - 2^, and 6 is an imaginary fifth root of unity.


N.B. A quintic reducible to this form can consequently be immediately solved.
23.

Write down trigonometrical expressions for the roots in the preceding

example; and,
(1)

(2)
(3)

being supposed essentially positive, prove

when p b < q\
when p* > g 2
when p* = g*

the roots are one real and four imaginary

the roots are

there

is

all real

+ ft 4 7) ( +

--

a square quadratic factor.

24. Find the following product, where 6


(a

f4?) (a -I- *fi

an imaginary

is

+ &y)

(a

Form th* biquadratic


a

where a

if

2a*,

a6

Ant.
26.

fifth root

7)

(o -f

of unity

+ fry).

47*-&ay(a*-0y).

equation whose roots are


a2

an imaginary root of x*

2a8 ,
1

a8

2
,

a*

+ 2a,

0.

Ans.

*+3

3*

11

0.

CHAPTER

VI.

ALGEBRAIC SOLUTION OF THE CUBIC AND BIQUADRATIC.

On

the Algebraic Solution of Equations. Before


proceeding to the solution of cubic and biquadratic equations
we make some introductory remarks, with a view of putting
clearly before the student the general principles on which the
55.

algebraic solution of these equations depends. With this object


give in the present Article three methods of solution of the

we

and state as we proceed how these methods may be


extended to cubic and biquadratic equations, leaving to subsequent Articles the complete development of the principles
quadratic,

involved.
(1). First

form

method of solution

by assuming for a root a general

involving radicals.

Since the expression p + Jq has two, and only two, values


root involved is taken with the double sign,

when the square

this is a natural

Assuming,

form to take for the root of a quadratic.


= p -f Jq, and rationalizing, we have

therefore, x

x* ~ 2px

Now,

if this

4-

z
p -

q = 0.

be identical with a given quadratic equation


+ Px + Q = 0,

&

2p = - P,

we have
giving

-=

r=
+ Jq

p*

= Q,

-P

which is the solution of the quadratic.


In the case of the cubic equation we shall find that

Algebraic Solution of the Cubic and Biquadratic.

106

are both proper forms to represent a root, these expressions


having each three, and only three, values when the cube roots

involved are taken in

all generality.

In the case of the biquadratic equation we

Jp+Jq
are forms which

+ 7=r-p,

shall find that

Jqjr

represent a root, these expressions each


when the square roots

may

giving four, and only four, values of x


receive their double signs.
(2).

Let
its

Second method of solution by resolving into factors.


be required to resolve the quadratic #2 + Px + Q into

it

For

simple factors.

this

s?

and determine

PX + Q

+ e-

it

under the form

0,

so that

a?

may

purpose we put

be a perfect square,

i. e.

Px + Q
we make
4-

+-, or-whence, putting for 6

its

2
the quadratic to the form w -

Thus we have reduced


its

we have

value,

simple factors are n + v, and u v.


we
shall
reduce
the
cubic to the
Subsequently
+ m) 8

(fo

and obtain

its

(I'x

+ ;')%

or

u*

&

and

form

v\

from the simple equations


u - (t)V = 0, w - o> 2 # = 0.
0,

solution

u-v=

be shown also that the biquadratic


to either of the forms
It will

(lx*+
2

(x +

mx +
px

ri)*

(Fx*

2
-f-

q) (x

-i-

may

+ m'x + n')%

p'x +

(?'),

be reduced

107

Algebraic Solution of Equations.

by solving a cubic equation and, consequently, the solution of


the biquadratic completed by solving two quadratics, viz. in the
first case, Ix + mx + n=*
and in the second case,
n'}
(I'x* + mx
=
=
9? +px + q
0.
0, and x* + p'x + q
;

-t-

Third method of solution

(3).

by symmetric functions of the

roots.

Consider the quadratic equation # 2 + Px +


the roots are a, /3.
have the relations

Q,

0, of

which

We

- - P,

Q.

a/3-

we attempt to determine a and j3 by these equations, we


back on the original equation (see Art. 24)
but if we

If
fall

could obtain a second equation between the roots and coefficients,


of the form la + w)3 = /(P, Q), we could easily find a and /3 by

means

of this equation and the equation a +


in the case of the quadratic there

j3

Now

finding the required equation;

(a-0)

=P

-4Q;

for,

is

no

difficulty in

obviously,

2
and, therefore, a -/3 = JP - 4Q.

In the case of the cubic equation x* + Px*


require two simple equations of the form
/a

= - P.

fW/3

y=/(P,

Qx

0,

we

Q, R),

in addition to the equation a + /3 + 7 =


P, to determine the
roots a, /3, 7. It will subsequently be proved that the functions
(a

may

+ wj3 +

cu

7)

(a

-K

w 2 j3 + wjY

be expressed in terms of the coefficients by solving a quad-

equation and when their values are


the cubic <^ay be easily found.
In the case of the biquadratic equation

ratic

X*

we

+ p#3 + Qff +

Rx +

known

S=

require three simple equations of the form


la

w$ + nj

-f

rS = /(P, Q, P, S),

the roots of

108

Algebraic Solution of the Cubic and Biquadratic.

in addition to the equation

to determine the roots a,

/3,

It will be proved in Art. 66,

7, 8.

that the three functions

be expressed in terms of the coefficients by solving a cubic


equation; and when their values are known the roots of the
biquadratic equation may be immediately obtained.

may

56.

The Algebraic

Solution of the Cubic Equation.

Let the general cubic equation

be put under the form


s3 +

3Hz + G =

0,-

where
s

= ax +

To

6,

H<*ac-b*,

<?d - 3abc

2b* (Art. 36).

solve this equation, assume*

hence, cubing,

f~
therefore
*8

Now, comparing

coefficients,

*Jp*J<l=-ff,

we have

p + q=

from which equations we obtain

Thi solution is usually called Cardan's solution of th* eubie.


the end of the volume.

See Note

A at

109

Algebraic Solution of the Cubic.


-

and, substituting for

JOT

*Jq its

value

we have

=r,

IIP

as the algebraic solution of the equation


s

+ 3//s

It should be noted that

+6 = 0.

p be

if

replaced

by

q this value of *

unchanged, as the terms are then simply interchanged ; also,


since *Jp has the three values *Jp, wjj/?, <*>*l]p, obtained by
is

multiplying any one of its values by the three cube roots of


unity, we obtain three, and only three, values for s, namely,

HP +

pp=>

Jp

w HP + w ';r=>

^ HP +

J;>

<>

rr

J/>

the order of these values only changing according to the cube


root of /? selected.

Now,

if z

be replaced by

value ax +

its

ax + b =

,;

Jp

6,

we

have, finally,

H=

(where jo has the value previously determined in terms of the


coefficients) as the complete algebraic solution

ax*

the square root

-f

3bx*

and cube

-f

3c# +

tf

of the cubic equation

0,

root involved being taken in their

entire generality.
57.

Application to Numerical Equations.

tion of the cubic

The

solu-

which has been obtained, unlike the solution

of the quadratic, is of little practical value when the coefficients


of the equation are given numbers ; although as an algebraic

solution

it is

complete.
the roots of the cubic are all real, 6? 2 + 4jET s = -K* Y
an essentially negative number (see Art. 43) ; and, substituting

For,

for

when

and! q their values

110

Algebraic Solution of the Cubic and Biquadratic.

in the

formula *Jp + *Jq

a root of the cubic

we have

the following expression for

Now there is no general arithmetical process for extracting


the cube root of such
and consequently this
complex numbers,
formula is useless for purposes of arithmetical calculation.
But when the
rical value

may

cubic has a pair of imaginary roots, a

nume-

be obtained from the formula

-f 47P is
As a practical method,
positive in this case.
however, of obtaining the real root of a numerical cubic, this

since

process

of little value.

is

In the
can make

namely, where the roots are all real, we


Trigonometry to obtain the numerical values

first case,

use of

manner

of the roots in the following

272 cos $ = - G, and 2 It sin

Assuming

d> V
7
- lie

~, and

-,

>

we nave

and

p
=tan d>
T

also

finally, since

01

q =

= i (G 2

J-

= cos

YV
d>

>

ue

=
<j)

7f,

1
;

!T*)-= (- 7/)

-f

1 sin

the three roots of the cubic equation


s

viz.

jjj?

-i-

<?,

01

+ 377s +

y^

cu

Jj

<7,

0,

'"

?J/?

-f

(t

jjy,

become
2

(- JT) J cos

|,
o

-2

(- JT)i cos

^4^
o

from which formulas we obtain the numerical values of the roots

Expression of the Cubic as Difference of two Cubes. Ill

and cosines. This process


and
in
not convenient in practice
general, for purposes of
arithmetical calculation of real roots, the methods of solution of
of the cubic

by

aid of a table of sines

is

numerical equations to be hereafter explained (Chap. X.) should


be employed.

Expression of the Cubic as the Difference of two

58.

Cubes.

Let the given cubic


ax* + 3bx*

3cx + d s

4-

(x)

be put under the form


s

where

= ax +

3Hz

G,

4-

b.

Now, assuming
s

where

-f

--

and v are quantities

fj.

[ft

-v(

i,)

+ M ) 3 },

(1)

to be determined, the second side

when reduced,

of this identity becomes,

fiif

Comparing

(z

(IUL

v).

coefficients,

Hv = - H,

fzv

(/LL

v}

=-

therefore

+
where

a*A

<?

2
4-

G
*=-,

4^%

=-aJK
-,
.

,- v

//

42

as in Art.

--

(z

Whence, putting

for z its value, ax

also

ju)

(s

-f

v)

-f

-i-

~
6,

~ H.

(2)

we have from

(1)

which

is

the required expression for

two cubes.

By

(x}

as the difference of

the aid of

the*

identity just proved the cubic can be

112

Algebraic Solution of the Cubic and Biquadratic.

resolved into

its

We

simple factors, and the solution of the equation


proceed to obtain expressions for the roots of

completed.
the equation ^ (a?) =
in terms of
mial cubic the equation

(fJL

we

V) C?$ (X)

ju

(z

-f

p and

vY -

now

selected

Solving as a bino-

V (2 + fdY = 0,

find the three following values for z

If

v.

ax + b

and jjv be replaced by any pair


one from each of the two series
*Jn

of cube roots

will be seen that we shall get the same three values of , the
order only of these values changing according to the cube roots
selected.
It follows that the expression
it

has three, and only three, values when the cube roots therein are
taken in all generality. This form therefore is, in addition to
that obtained in the last Article, a form proper to represent a
root of a cubic equation (see (1), Art. 55).

The function (2) given


becomes, as

may

above,

when transformed and reduced,

be easily seen,

d- bc)x+

(bd-<*)}.

This quadratic, therefore, contains as factors the two binomials


ax -f b + n, ax + b + v, which occur in the above expression of ^ (x)
as the difference of two cubes.

113

Solution of the Cubic by Symmetric Functions.


59.

Solution of the Cubic by Symmetric Functions

of the Roots.
{a

when

/3

Since the three values of the expression

+ y + 0(a +

takes the values

w/3-f

a>

1,

o>, o>

are

y) +

(a

o> /3

it is

a, /3, y,

017)

plain that

if

the functions
9 (a + wfi +

w 2 y),

were expressed in terms of ^he

(a

a/*/3

o>y)

coefficients of the cuhic,

we

could,

by substituting their values in the formula given above, arrive


this
at an algebraical solution of the cubic equation.

Now

cannot be done directly by solving a quadratic equation


although the product of the two functions above written

for,
is

It
rational symmetric function of a, /3, y, their sum is not so.
will be found, however, that the sum of the cubes of the two

functions in question

is

a symmetric function of the roots,

and

can, therefore, be expressed by the coefficients, as we proceed to


show. For convenience we adopt the notation

We

-a

-f

oi/3

oTy,

-I

M= a

w 2 /3 +

-f

ojy.

have then
(OL)*

A Bu + <?or,

-f

(0W)

A + B<

+ Cu,

where

^=a

+/3 -fy +6a)3y, J?=3(a

/3

2
2
2
+ /3 y + y 8 a), <7=3(a/3 + /3y -f ya 2 J

from which we obtain

U M
-f

(Of.

Ex.

5, p.

2S

44; Ex. 15,

- 3Sa 2 +
= - 27
)3
12a/3y

p. 50.)

Again,
*~

whence

(a

r"

+ w/3 + o/y)

3
,

f-' i

(a +

/3

/*

+ wy)

are the roots of the quadratic equation

a3

Denoting the roots

a6

of this equation, viz.

"f-'
s

Algebraic Solution of the Cubic and Biquadratic.

114
by

ti

and

2,

the original formula expressed in terms of the

coefficients of the cubic gives for the three roots

It will be seen that the values of a,

/3,

7 here arrived at are

of the same form as those already obtained in Art. 56.


It is important to observe that the functions
-f

(a

w[5

o>

y)

(a

>

-f

c*r/3

tuy)

are remarkable as being the simplest functions of three quantities which have but two values when these quantities are interchanged in every way. It is owing to this property that the
solution of a cubic equation can be reduced to that of a quadand
Several functions of a, )3, 7 of this nature exist
ratic.
it will be proved in a subsequent chapter that any two such
;

functions are connected

by a

rational linear relation in terms of

the coefficients.

Having now completed

the discussion of the different

of algebraical solution of

the cubic,

we

modes

give some examples

involving the principles contained in the preceding Articles.

EXAMPLES.
1.

Resolve into simple factors the expression


(0

Let

-7

2
)

(x

a) -f (7

F= 08 -?)(*-),

- a)* (x - 0) +

- *)

(ZT +

Prove that the several equations of the system


()3~ 7 )3(^-a)3=( 7

have two factors common-

-a)

(x

- if.

W= (a- 8) (*- 7

F= (7 -)(*-,
Am.

2.

(a

(^-^ 3 -(a

).

115

Examples.
Example, we have

in the last

Making use of the notation


whence

i7

since

therefore
is

the

(ft

7)* (*

a)

+ F-f TFsO;
- a) 2 (* - jS) 2 +

(7

- j8)

(a

(*

7)'

common

quadratic factor required.


Resolve into factors the expressions

3.

These

(1).

(ft-

(2).

(/8

(3).

()8

(x

- 7 s (* ~
- 7 )7 (x -

a)

by

)
7

4-

a)

4.

(7

5
)

ZUVW\

f (Z7

(2)

(*

a)' (x

3
ft)

0)

(a
(

(a

/8)

(*

j3)'

(*

-7

0) (*

- yf
- 7) 5

7
)

at once from the results established in Ex. 40,


and replacing ai, $1, 71, in the example referred

1,

W, we obtain the following

F,
(1)

(x

a)

down

factors can be written

Z7,

Ans.

(7

+ (7 -

Using the notation of Ex.

p. 59.
to,

+F

W*)UVW\

(3)

i(Z7

V^+

-f

Express

-)(*-

(*-<*)(*
as the difference of

7)

two cubes.

Assume

- *V;

^i 3

(*-a)(*-(*- 7 )=
whence
Z7i

w [7,

we have

arid, therefore,

\
but AM*'

A +

/i -f

- 7 ),

(ft

- taTi

Z7j

Adding,

- Fi = A (x "
Fi = ft (a: -

/u

/8),

v (x

Aa

0,

a),

p (7

-f-

7).
/xy8 -f

a),

1^7

i>

=
^>

(a

ft)

whence

Substituting these values of A,

/x,

and using the notation of Ex.

r;

1,

whence

and

Z7i

5.

and Fi are completely determined.


are functions of the differences of the
L and

Prove that

We

L = a -f

have

for all values of A, since


cession,

we

aj8

-f

w-7 = a

o> 4- or*

obtain three forms for

-Similarly for

A + w

(/8

and giving

/*)

to

+ w2

(7

h the values

in terms of the differences

M.

17

roots.

h]

/3

a,

j8,

7, in suc-

- 7, 7 - a, a - ft.

116

Algebraic Solution of the Cubic and Biquadratic.


To

6.

express the product of the squares of the differences of the roots in terma

of the coefficients.

We have
Z4- Jf=2a--7, Z +
agam

and,

L- J/=

(ft

- 7)(w-

orZ- wJf =

2
),

Z3

- a2

(7 -a)(a>

eoL-tfM =

),

(a

- w 2 ),

)(

= - 3 <~Z (ft - 7)( 7 - a) (a - ft)

since

- Jlf s) 2 == (Z 3 4- 3/ 3
(Z

Z 3 + M8

have, suhstituting the values of


.

fl6 (

(Cl

= (27- a-/8)ar;

26,

we

-\-wHf

from which we obtain, as in Art.

and

3Jf=(2-7-a)a>,

7 )2 ( 7

. a (a
j2

2
)

and

- 4Z-W 3

27

0)2=

Z Jf obtained

in Art. 59,

Art. 42.)

7.

Prove the following

+ Jf3

Z8 These are

Jf 3

identities

_
i{ (2a ^_ 7

E V/^JOS -

easily obtained

).>

+
+

7)

^ 7 _ a )3

(2j8

(7

3
)

(a

:i

^8)

by cubing and adding the values of

&c.

M,

4-

Z-

Jf, &c.,

in the preceding example.

2
To obtain expressions for Z 2
&c., in terms of the differences
The following forms for Z 2 and 3/ 2 are obtained by subtracting

8.

(a

+ 7 2 )(l4 cy + w 2 )^ from (a+ <w)8+


- Z 2 = (ft - 7 2 + w2 (7 - a)'
- Jlf 2 = (j8 - 7) 2 + (7 - a) 2

In a similar manner, we

(7
2
(

difficulty,

2Z1T =

9.

There are

(ft

we have

-7

2
)

4-

4-

to form the equation

(aft

+ w

<*>'*& 4-

whose

and

(a

(a
2

4-

(a

)3)

+ ury) 2

y
/8)

- 7 - ) 2 + ^(a - J8) 2
- a) 2 -f w (a - j3) 2
(2^8- 7

)M2)8

3
7 - o)

the following forms for


(7

a)

(a

six functions of the type of

a +

7)

from these expressions

find

- Z 4 == (j8 - 7)M2 ~ 3 - 7) 2 +
- Jlf*= 03-7) 2 (2a-/8- 7 2 4
Also, without

of o, 0, 7.

7,

W7,

wa

4- )3 4-

or Jf, viz.,

7,

w2 a

4-

ft 4-

7,

w^a

4-

cu8

roots are these six quantities.

(2 7

Z Jlf and L 2 M 2

2
)

t*ft

wa +

(2 7

a7,
4-

7,

117

Examples.
These functions

be expressed as follows

may

X,

X,

M,

<*M,

hence they are the roots of the equation


(0

- X)(<J> - wZ)(f -

or

0>'

Substituting for

we have

and

jfcf

o>

- Jf )(<J> - a>M )(<j, - W2jf ) =

X)(<J>

(1,3 -f

&M* =

^+

If 3)

0,

0.

from the equations

this equation expressed in terms of the coefficients as follows

10. To form, in terms of L and H, the equation whose roots are the
squares of
the differences of the roots of the general culic equation.

Let

* =

(a

fi?

hence, by former results,

Rationalizing this,

we

obtain

27

which

the required equation.


In a similar manner, by the
is

aid

of

the results

of Ex. 8, the equation of

squared differences of this equation, or the equation whose roots are

is

obtained by substituting

- X 2 and ~

for

M and X,

respectively, in the last

equation ; and this process may be repeated any number of times.


Finally,
these equations may be easily expressed in terms of the coefficients of the cubic

means of the

relations

LM=-9~,
For instance, the

first

equation

*(*
<Cf. Art. 42.)

is

4P

and

T*

'

"

""

all

by

Algebraic Solution of the Cubic and Biquadratic.

118

11. If a,

y and

a',

',

y' be the roots of the cubic equations

ax*

a's'-f-

to

+ 3c

3baP

rf

-I-

0,

+ Zc'x + d' =

3'a?

form the equation which has for roots the six values of the function

$a

oa'

/3j8'

+ yy.

The easiest mode of procedure is first to form the corresponding equation for
the cubics deprived of their second terms, viz.,
**

+ 32Ts + Q =

s3

0,

+ 31T* +

and thence deduce the equation in the general case


transformed the corresponding function

4,0

s;

(aa

6?'

0,

for in the case of the cubics so

+ *)(V

Substituting for the roots of the transformed equations their values expressed

by

radicals,

we have

which reduces to

Cubing

this,

we

find
0.

Now,

substituting for
8

a?

we have

and

q,

p' and

+ Gx - jET3 =

a;

0,

their values given

#',

'#

-f

JB"'

by

the equations

0,

the six values of ^o given by the two cubic equations

aa'-v/^

((?^'

0,

where

and
Finally, substituting for

we have
be 8 - 1 e

^>

its

value

aa'<f>

together,

the required equation.

cubics

0,

sidered in Ex. 9.

<f>

-f

wj8
,

2
-f-

7,

It
&c.,

and multiplying these cubics


be noticed that if one of the

3*^',

may

which case has been already con-

Mr. M. Roberts, Dublin Exam. Papers, 1855.

119

Examples.
12.

Form the

equation whose roots are the several values of

where

p,

a-7
and

Since

/>

a,

are the roots of the equation <M? 3 + 3bx* + 3<a? + d ** 0.


involves only the differences, and the ratios of a,
,
7, the result

will he the
*

-f

32&

same

-f (7

if a,

whence, eliminating

S>

the required equation

i,

(p

!)(,,

2)(2p

Find the relation between the

roots

i,

z2)

1)

*s

of the equation

1)

is

3i'ir' 2

Q* (?

0.

coefficients of the cuhios

+ Ux* +Zcx

ay*

when

by the

replaced

have, therefore,

H=-

nd similarly

13.

7 be

j8,

We

0.

3e?V

4-

0,

d'

0,

the roots are connected by the equation

a ('

Multiplying by

w2

ot

7')

(7'

LM
Cubing, and introducing the

0.

becomes

this equation

+ 7 (a' -

a')

'

i'JIf.

we

coefficients,

find

the required relation.


14.

Determine the condition in terms of the roots and

coefficients

that the

cubics of Ex. 13 should become identical by the linear transformation

In

Eliminating p and

= pa +

-px-\-

g,

is

g,

q.

ft'

= p&

qt

-f

y*spy +

q.

we have

7'

which

x'

this case

fry + yof

7'a

aft'

a']8

0,

the function of the roots considered in the last example.

moreover,

is

unchanged

if for o, 0,

la

m,

Ift

JV

-I-

m',

l'&

a', 0', 7',

+ m,
4-

m',

we

ly

substitute

+ m,

/V +

w'

This relation,

Algebraic Solution of the Cubic and Biquadratic.

120

whence we may consider the cubics


2s

+ ZHz +

in the last

G=

0,

example under the simple forms

+ 3HV

z'*

G'

4-

0,

obtained by the linear transformations z = ax + b, z = ax' + V ; for if the condition


holds for the roots of the former equations, it must hold for the roots of the latter.

Now

putting z

= kz,

these equations become identical if

G'm&G;

H'eztfH,
whence, eliminating

is

&,

G^H* = Gfi H?

the required condition, the same as that obtained in Ex. 13.

become

that the reducing quadratics of the cubics necessarily

It

may

identical

be observed

by the same

transformation, viz.,
jy

77"'

(a'x'

b')

=-

(ax

b).

Homographic Relation between two Roots of a

60.

Cubic.

Before proceeding to the discussion of the biquadratic

we prove
cubic

the following important proposition relative to the

The

roots of the cubic are connected in pairs by a

relation in terms of

tJie

we have

Referring to Exs. 13, 14, Art. 27,


2

O - y)

flo

<?

(a (/3
\

(ft

y)

jY

+
-f

(7
(y

/3
2

/3

homographic

coefficients.

(y

a)
2

+ y

a)
2

a)

(a

+7

(a
2

(a

/3)

/3)

/3)

the relations

= 18 (a? - flofl,),
= 9 (aa z - a,a^ 9
- 18

a.a,).

Using the notation

multiplying the above equations by

and adding since


2
a - a (a -f ft) +

aj3,

- (a +

/3), 1,

respectively,

a/3

0,

we have

ft

ft

(a

|3)

aft

but

(see Art. 42)

whence

V ~f ("4^)

H $ + JTi(a + ^) +

0,

121

First Solution by Radicals of the Biquadratic.


and, therefore,

It is to be obis the required homographic relation.


served that the coefficients in this equation involve one irrational quantity, the second sign of which will give the relation

which

between a different pair of the roots.


61. First Solution by Radicals of the Biquadratic.
Euler's Assumption. Let the biquadratic equation
=

be put under the form (Art. 37)


s

where

= ax +

+ 6//s* + 4Gz + cfl -

3H

0,

b,

H - ac - b\

I=

ae

- 4bd

3c

4-

(7

= a 2 ^ - Sabc + 2^ s

To

solve this equation (a biquadratic wanting the second


Euler
assumes as the general expression for a root
term)
*

= Jy> +

J?

J>*-

Squaring,

^~P-g-r = 2(J,j Jr

+ J^J7)-

Jr Jp

Squaring again, and reducing, we obtain the equation


x*

-2(p +

q +

r)z*-8zjp Jq Jr+ (p + q

Comparing

p+

q + r

this equation

= - 3H,

q,

~4(qr+ rp+pq) = 0.

r}

we have

with the former,

qr+rp+pq =

and consequently jo,

3H

Jp Jq J r = -

r are the roots of the equation


(1)
v 7

\
or, since

6?

4 J? 3 - a2 J27 + <fj,

(Art 37),
.

where
t7 s ^r^

+ 2bcd

ad z

eb*

s
,

122 Algebraic Solution of


this equation

and

finally,

may

and Biquadratic.

be written in the form

putting

4r*

This

the Cubic

H*

we

0*0,

- laO +

obtain the equation

J=

0.

(2)

called the reducing cubic of the


biquadratic equation
will in what follows be referred to
by that name.
is

necessary to

we

make a

between equations

distinction

and

When

it is

and

(2),

(1)

shall refer to the

former as Euler's cubic.


- ac + a 2 6 if 13 2 3 be the roots of the
tm
V
since
Also,
reducing cubic, we have
2
b - ac + 2 2
r^b*
p ^ b* - ac + a*9 l9
;

and, therefore,

If this formula be taken to


represent a root of the biquadramust be observed that the radicals involved have not

tic in 2, it

complete generality for if they had, eight values of z in place


would be given by the formula. The proper limitation
;

of four
is

imposed by the relation

which

sight of in squaring to obtain the value of pqr]


requires such signs to be attached to each of the quantities
(lost

JP> J$>

r/*%

that their product

mined by the above equation

may

maintain the sign deter-

thus

(- J7>)

are

all

the possible

combinations of Jp, J#,

this condition, provided that

Jp,

throughout, whatever those signs

remove

all

(-J7) J^

Jg

may

Jr

fulfilling

same signs
may, however,

J~r retain the


be.

We

ambiguity as regards sign, and express in a single

First Solution by Radicals of the Biquadratic.


algebraic formula the four values of

by eliminating one

s,

123
of the

quantities J/>, Jq, Jr from the assumed value of z by means


of the relation given above, and leaving the other two quanti-

The

ties unrestricted in sign.

a formula free from

all

expression for z becomes therefore

ambiguity, since

it

gives four,

and only

four, values of z when Jp and Jq receive their double signs


the sign given to each of these in the two first terms determining that which must be attached to it in the denominator of
:

the third term.

given before,

And finally,

restoring to

q,

and

z their

values

we have

ax + b = J b* - ac + a*Q l + J #* - ac + az f) 2

G
as the complete algebraic solution
0i

and 9 2 being

of the

biquadratic

equation

roots of the equation

- laO +

J=

0.

To assist the student in justifying Euler's apparently arbitrary assumption as to the form of solution of the biquadratic,
we remark that, the second term of the equation in z
being
absent, the sum of the four roots is zero, or Sj + s 2 + s 3 + s 4 =
;

and consequently the functions


in o-eneral

are in this

which there are


six (the combinations of four
quantities two and two),
case reduced to three
so that we may assume
s2)

&c., of

Si)

from wbjch we have

(*i

==

s,, s 3 , s s ,

2 4 , included in the

Jp +Jq + J' 7

formula

Algebraic Solution of the Cubic and Biquadratic.

124

We
and

now

proceed to express the roots of Euler's cubic (1),


reducing cubic (2), in terms of the roots

also those of the

Attending to the
remarks above made with reference to the signs of the radicals,
we may write the four values of z = ax + b as follows
>

j3>

S of the given biquadratic in x.

7>

ay +

from which
sions for

7;,

may
#,

= -

be immediately derived the following expres-

r the roots of Euler's cubic

(T +

'

f3

(4

8)',

)*-

Subtracting in pairs the equations (3), and making use of


the relations above written between p g, r and 6/1, 6., 6/
we
y

;j

the following useful relations connecting the


differences of the roots of the cubics (1) and (2) with the diffe-

easily establish

rences of the roots of the biquadratic

4
4

(q

_
_

(,-

4 (;>-

Oi

r)
;;)

y)

4 ,/a

= 4

2
r/

= 4a*

(0,
(9,

0,)

= ~

rr (/3-

viz.,

- rr
00 =
(7
a) (/3
= - ^ 2 (a -/3)( 7 ft)
-

Finally, from these equations, by


= 0, we derive the values of
2 + ft

a, |3, 7, S,

7 )(a

aid
ft, ft,

of

8),
S)

(5)

S).

the

relation

ft in terms of

125

Examples.

EXAMPLES.
1

roots,

"When the biquadratic has two equal

reducing cuhic has two equal

oots, the

and conversely.

2. "When the biquadratic has three roots equal,


0.
cuhic vanish, and consequently
0,

J=

all

the roots of the reducing

/=

3. When the biquadratic has two distinct pairs of equal roots,


12H~ = 0.
of Euler's cubic vanish, and consequently G = 0, a-I

two

of the roots

'

4.

Prove the following relations between the biquadratic and Euler's cubic with

respect to the nature of the roots


(1).

are

all real

(2).

cubic are

When
and

the roots of the biquadratic are

all real,

tho roots of Euler's cubic

positive.

When

the roots of the biquadratic aie

all

imaginary, tho roots of Eulei's

two being negative and one positive

nil real,

When the biquadratic has two real and two imaginary roots, Euler's
(.").
cubic has two imaginary loots and one real positive root.
These results follow readily from equations (4) when the proper forms are subIt is to be observed that all possible
7, S in the values of;;, g, r.
1

stituted for a,

cases arc heie comprised, the biquadratic being supposed not to have equal roots.
It follows that the converse of each of these propositions is true.
Hence, when

Euler's cubic has

all its roots real

and

positive,

we may

conclude that

all

the roots

when Euler's cubic has negative roots, we conclude


of the biquadratic are real
that all the loots of the biquadratic are imaginary; and when Euler's cubic has
;

imaginary

roots,

we

conclude that the biquadratic has two real and two imaginary

roots.
5.

Prove that the roots of the biquadratic and the roots of the reducing cubic

are connected

by the following

relations

When

the roots of the biquadratic are either all real, or all imaginary,
(1).
the roots of the reducing cubic are all real ; and, conversely, when the roots of
the reducing cubic are all real, the roots of the biquadratic are either all real or all

imaginary.
(2). When the biquadratic has two real, and two imaginary roots, the reducing cubic has two imaginary roots and, conversely, when the reducing cubic has
two imaginary roots, the biquadratic has two real and two imaginary roots.
;

These results follow readily from the preceding example, since the roots of the
(1) and (2) are connected by a real linear relation.

two cubics
6.

When If is
For

7.

positive, the biquadratic has

imaginary

roots.

in that case the roots of Euler's cubic cannot be all positive.

When / is

negative, the biquadratic has two, real and

For the reducing cubic has in that

case

two imaginary

two imaginary

roots.

roots (Ex. 12, p. 33).

126 Algebraic Solution of the Cubic and Biquadratic.

H and

When

8.

are both positive, all the roots of the

are

biquadratic

imaginary.

J is

For, since

has a real negative root

positive, the reducing cubic

fore also Euler's cubic has a real negative root, since


tive

and

Ex.

this is case (2) of

Show

that the

H, and

a*Q

there-

is

posi-

It is implied in this proof that the


l^&djpg
substituted for /in the statement of the proposi-

4.

coefficient a is positive ; if a/be


tion no restriction as to the sign of a
9.

is

necessary.

two biquadratic equations

A X*
Q

4.A& + ^4 =

SAiX*

4-

have the same reducing cubic.


10. Find the reducing cubic of the two biquadratic equations

m3 + w 3

Sim* + 3 (\mn
Ans.

11.

2
)

0.

- 3mnO - (m*

Prove that the eight roots of the equation

{x*
are given

6lx*

(4ww

)}

64 (^

+ w3 +

nz

- Mum) x 2

by the formula
n +

''Compare Ex. 20,

v*

+ u~n

-f tarn

\/ 1

-f

4-

p. 34.)

12. If the expression

be a root of the equation


z4

determine H,

J in

I,

-f-

62Zir 2

terms of

13.

-f-

a zl

- 2H'Z =

0,

n.

m,

/,

Ans.

4^

JT= -

oo

/,

J=

12ww,

/=

- 4 (w 3 +

3
).

Write down the formulas which express the root of a biquadratic in the parwhen /= 0, and /= 0.

ticular cases

by the aid of the reducing

14. Express,

of the roots a,
15.

I and / in terms of the

differences

Express the product of the squares of the differences of the roots

in terms of

I and

By means
a* (3

of the equations (5) above given,

_ yy y -

What

is

a,

7, 5

/.

tain the result as follows

16.

cubic,

(See Exs. 16, 18, Art. 27.)

7, 5.

and the equation

(2), p.

82,

we ob-

a )2 ( a

)2

8 )2

_ 5)2 7 _ 5 yz =
(

2 56 (/*

- 27/2 ).

the quantity under the final square root (viz., that which occurs

under the cube root in the solution of the reducing cubic) in the formula expressing
a root?
Ans. 27J*-I*.
17.

Prove that the

coefficients of the equation of

4
biquadratic equation a #

jterms

o>

J7) /,

and

4ai# 3

-f

62#2 +

4^3 x

squared differences of the


=
may be expressed in

Second Solution by Radicals of


the second term from the equation,

Removing

^+

^+^
6JET

and changing the signs of the

4G

127

the Biquadratic.

we

obtain

we have

roots,

These transformations leave the functions

(a

#)

2
,

0.

&c.,

unaltered

hut

becomes

G, the other coefficients of the latter equation remaining unchanged;


thei ef ore G can enter the coefficients of the equation of squared differences in even

And by

powers puly.
ducing

aid of the identity of Art. 37, G"1

In a similar manner we

II, I, J.

of the differences of the roots a,

the function

7, 5

)3,

may

be eliminated, intro-

may

prove that every even function


be expressed in terms of a JZ", /, /,

may

of odd degree not entering.

Second Solution by Radicals of tbe Biqua-

02.

Let the biquadratic equation

dratic.

ax*

-f

+ ^dx +

6cx?

4-

4/>.f'

be put, as before, under the form


s

where

We

rt,r

4 CJZs 2 +

46% +

/ - 3// =

0,

^.

now assume

as the general expression for a root of this

equation
^

J> +

J?

J r Jy

+ J/ ; J?>

a formula involving three independent radicals,


Squaring twice, and reducing, we have
-

(z

- rp - pq)' = 4pqr (2z + p +

J/;,

J^, Jr.

qr

-f

r)

or
s

-2

(yr

-f-

rjt?

-f

j^g) s

Comparing

Spqrz

-f

this equation

(^r

=-

^,

^7^

+pq}

-4(p + q+

<tI

r are the roots of the equation


2
4-

(12//

r)pqr =

with the former equation in

easily find

whence, p,

-I-

a !)

f - 6//*

-f

0.

s,

0.

we

128

Algebraic Solution of the Cubic and Biquadratic.

This equation
or

making

may be readily

transformed into Euler's cubic,

directly the substitution

H- iftf
and putting
reduce

it

for 6r its value in terms of

4fl

JET,

- laQ +

J=

and

/,

form of the reducing

to the standard

J,

we may

cubic, viz.,

0.

important to observe that in the present method of


solution we meet with no ambiguity corresponding to that of
Art. 61 for the expression here assumed as the value of z has, in
It

is

virtue of the double signs of the ladicals contained in it, only


four values, while the form assumed for z in the preceding Article

This appears from the identical equation

has eight values.

which shows that the number

of distinct values of the radical

number

expression of the present Article is the same as the


2
values of (Jp + Jq + Jr) namely four.

of

In order

to express p,

the biquadratic,

we

q,

have, giving to x the four value?

Jq Jr - Jr Jp - Jp

*!

- aa +

z2

*ap

z,^ay +

= -

*4

r in terms of the roots a,

aS + b =

The student may

-Jq Jr

-\

Jq,

Jq Jr + Jr Jp

Jq.

0(/3

a z (fly - aS) + ab

((3

Jp

easily satisfy himself that

Si

y-a-g)

+ y - a -

8)

of

Jq,

Jq Jr - Jr Jp + Jp

the values of z 2 + s 3 -

y,

Jr Jp - Jp Jq,

of the signs of the radicals can lead to


these four.

From

/3,

a, /3, y, o,

s4 ,

any value

and

sa s3

= - 4jq Jr,

= 4

no combination

-z

different from

z^

we

obtain

Resolution of the Quartic into

From
tion

its

Quadratic Factors. 129

and similar equations we have, employing the rela= - 2pqr, the following modes of expressing p, q, r in

these

terms of the roots

a, /3, 7,

/3y-aS
'

8#

a* ( 7

- r = a

- 78
--- -

a/3

a +

~:

=:

/:>

- 3 /

"

86?

=
a* (a

4-

/3

Resolution of the Ciuartie into

63.

Factors.

8)

its

Let the quartio

be supposed to be expressed as the difference of two squares* in


the form
(ax?

and comparing it with


Multiplying the given quartic by
we have the following equations to determine
,

this expression,

M, N, and

M* = b - ac
2

Eliminating
4a

which

is

MN= be - ad + 2abO, N

+ art,

(ae

(c

+ 2a0) 2 -

ae.

M and N from these equations, we find

- 4bd +

3c'

aO + ace + 2bcd-ad* - eb~ -

the reducing cubic before obtained.


this equation we have three values of

From

(0,,

0,

2,

3 ),

2
with three corresponding values of J/ 2 MN,
and thus all
the coefficients of the assumed form for the quartic are deter,

* The reduction of the


quartic

to the diif erenre of

two squares was the method

employed for the solution of the equation of the fourth degree. This mode of
solution is due to Ferrari, although by some writers ascribed to Simpson (see note A).
first

The method explained

in the following Article, in

directly to the product of

two quadratic

factors, is

which the quartic

due to Descarte*.

is

equated

130 Algebraic Solution of the Cubic and Biquadratic.


mined in three

distinct

that to each value of

ways

it

moreover,

M corresponds a

should be noticed

single value of

N>

since

MN = bc-ad
The

quartio

+ 2a6)* -

may

plainly be resolved into the two quadratic factors

When

receives the three values

2,

1?

#3,

we

pairs of quadratic factors of the original quartic,


is completely solved.

obtain the three

aud the problem

In order to make clear the connexion between the present


and the solution by radicals, let us suppose that the

solution

roots of the quadratic factors iu the order above written are


and that the roots of the remaining pairs of
(3, 7 and a, 3
;

and

quadratic factors are similarly y> a

"We have,

j3,

a,

and

7, S.

therefore,

where
2

J/!

Jl>

-ac+(fO

l9

M* m Jtf

Jtf-ac+a^,

Subtracting the last equations in pairs, we find


/3

+7 _ a _S=4

and

'

since

obtain

'

..

a +

we

_S-8_4

7 +

..

^+7

+ S =

aa

= - J/

aft

ay +

ft

jl/!

J/!

- -

J/l

-f

a + 'fl-7-8-4

-4-,

M> +

- J/ 2
-t-

4-

J/"3 ,

1T3 ,

M-i

- J/3 ,

J/,

J/3

131

Resolution of the Quartic.

It appears, therefore, that the roots of the biquadratic are here


expressed separately by formulas analogous to those of Art. 61.
2
values of M'\ viz. Jfi , Mf^ M*> are in fact identical with
the roots of Euler's cubic in the preceding Article.
There

The

with regard to the signs of the radicals involved in


J/3 a restriction similar to that of Art. 61
since, in

exists also

J/2,

l9

virtue of the assumptions above made with respect to the roots


of the quadratic factors, we have the equation
a* (ft

+ 7 - a - S) (7 + a -

j3

S) (a

which implies the following relation

and by means
stricted in the

/3

S)

64Jtfi Jfa

Ex. 20,

(see

p. 52)

#3,

of this relation the signs of


19 Jf2 , Jfs are rein the previous Article.

manner explained

aid of the equation last written we can eliminate


3
from the expressions for the roots, and thus obtain, as in Art. 61,
all the roots of the biquadratic in a single formula, viz.,

By

which the radicals

iii

-are

/r

uc

and JT2 -

a~t) l9

b'

ao

-t-

a2

taken in complete generality.


EXAMPLES.

1.

Form

the equation whose roots are A,

7 4
Adding the

where

0i,

70

a5,

-1-

/*,

Exs.

viz.,

78.

last coefficients of the quadratic factors of the quartic,

07 + aS =

40!

2 -,
a

05=

402

2 -,

4^ +

2-,

7a

-f

0)8

+ 78=

62,*^ are the roots of the reducing cubic

An*, (ax
(Cf.

v,

a0

35,

4, 5, Art. 39.)

K2

we have

hence the required equation.


3

2o)

- 41 (ax -

2c)

+ 16J=

0.

Algebraic Solution of the Cubic and Biquadratic.

132

2. Express, by means of the equations of the preceding example, the roots of


the reducing cubic in terms of the roots of the biquadratic.

2c

Substituting for

its

value in terms of

1201

= 2\ -

120 2

2/i

120a

2r

fji

(7

-Xs

- \ - fi s

a,

a) (3

0) (7

8)

7) (a

5)

we

7, 5,

ft,

5)

(a

find

0) (7

OB

5),

7) (a

(7

immediately

),

a) (0

5).

(Cf. (6), Art. 61.)


3.

Ex.

Verify, by

5, Art. 61,

means of the expressions

and reducing cubic are


4.

a3)(j8

From

1,

the conclusions of
oi'

the biquadratic

related.

the equation whose roots are the functions

Form

(37

Ex.

for 0\, 02, 03 in

with respect to the manner in which the roots

+ 7 - a - 5),

7 a - 05)

+a-

(7

we

the quadratic factors of the quartic


4,l/i

5),

(a/8

a5

78) (a

+0 - 7

3)

find

r
2J\

7-a-5,

)8

also

JfiJVi

Ic

- ad + 2ab6\ =

the roots of the required cubic being represented

We

obtain, therefore, the required equation

reducing cubic.

Ans.
6.

Form

+ bc-

(a*<l>

ad)

az

<f>i,

by <i, </>2, ^3.


by a linear transformation of the
z
b'

l(a~(^

be

ad]

2W= 0.

the equation whoso roots are

ya - 35

7 - a5

7 +

7-0-5'

a-0-5'

a -T"0

- 78
-7-

denote any one of these functions indifferently, and


<f>
of the reducing cubic, we have, employing former results,
If

__

MN _

^~_

- ad + labe

Ic

r-ac +

Jf- ~~b'

5*

the corresponding root

'

a~ti

and thus we obtain the required equation by a homographic transformation of the


reducing cubic. This formula may be put under the more convenient form

by means of which we obtain the required cubic in the following form

26

(0<J> -f

3
)

+ (a-I-

12^2)

which, expanded and divided by

W^ +
(Cf.

Ex.

(a*e -f

3
,

(<ty

14, p. 88.)

-f b)

G*

0,

becomes

6^ - 9ac- 2W) ^ +
-f

+ bf - HG(a<l>

2 (afo

+ 2^2 rf - Zacd) t + Pe- ad* =

0.

133

Resolution of the Quartic.


Form

6.

the equation whose roots are


2

?-

These are the three values of

(ya

^03y-a8)

?5)-',

- 7 5) 2

(a

in the foregoing Article.

Representing, as

by <, we find that the required equation may be obtained


from the reducing cubic by means of the homographic transformation
before, one of these values

- ad~-

2bcd

Form

7.

<

labdd

the equation whoso roots are

07 -a5
- (a +

y) aS

7'

is

(7

a)

a&

#8

ya
5)

The required equation


rmq^un

tr an s f

eb 2

08 - (0

-f

8)

ya

(a

-f

78

78

"

(7

+ 5)

obtained from the reducing cubic by the homographic

--

cd

be

+ 2dO

20
r = d 2 -ce
+ ae8

may be derived from Ex. 5 by changing the roots into their reciproand making the corresponding changes ia the coefficients.

This result
cals,

64.

The Resolution of
Second

Factors.

the (fcuartic into Quadratic


Let the quartic

M ethod

<:\r

-I-

-f

cfo

be supposed to be resolved into the quadratic factors


a

We
p

have,

(a?

4 2px

by comparing

p'=2-

q) (#

these

we

fifth

could eliminate^, p',


the several values of 0.
fiftji

q,

%px

+ q).

pq'+p'g = 2

ti>

now we had any

The

two forms, the equations

+ q+ typ'= 6

Cl

If

and

-,
Cl

^=U

^.

fl)

equation of the form

q\

and thus find an equation giving

equation might be assumed to be jt?;/= 0, or q -f q'- $ ;


would be determined by a cubic equation,

in each case

<

since each of these functions,

when

expressed in terms of the

134

Algebraic Solution of the Cubic and Biquadratic.

roots of the biquadratic, has three values only.

venient, however, to

If

the two functions of p, p'9


second of equations (I).

q, q'

We

equations,
,

pq + pq .
and eliminating p,

It

is

more con-

assume

p', q, q',

2c

here involved being equal by the


easily find, by the aid of those

--

4abc - 2a*d

by means

8bd>

_;

of the identical relation

there results the equation

J
which

the reducing cubic obtained

is

0,

by the previous methods

of solution.

pp' 9 or q + q', we may complete the


the
resolution of
quartic by means of the equations (1).
The reason for the assumption above made with regard to

Having thus found

From a comparison
the form of the fifth equation is obvious.
of the assumed values of
with the equations of Ex. 1, Art. 63,
it

in the preceding Article and


foresee that the elimination of p, ;/, g, <?', must lead

appears that

therefore

we

is

the same as

identical with the reducing cubic before


to an equation in
In general, if represent any[function of the differobtained.
ences of A, /x, v, and consequently an even function of the differ<f>

<

Ex. 18, Art. 27), the equation whose


cannot involve any functions
roots are the different values of
of the coefficients except a, H, J, and J.
ences of

If

a, /3, 7, 8

(see

be assumed equal to any of the expressions in the second

of the following examples, the equation in


whose roots are the
different values of this expression is formed as in the above

instance

by the elimination

of

jp, jp',

g,

q.

Transformation of Biquadratic into Reciprocal Form. 135


EXAMPLES.
1

Resolve into quadratic factors

i*

Comparing

this

+ 4Gz

a 2/- 3.H*.

-f

form with the product


(*

we

6Ift 2

2p*

find the following equation for

$)(

2j>* -f ?'),

^""12") ^"^
(2r\

2==

and putting
this equation,

when

3
by a becomes

divided

4a s
2.

3
<f>

- lap

If a quartic be resolved into the

/=

-f-

0.

two quadratic

x + px 4 gt

x + p'x

factors
q',

determined by a cubic equation when


corresponding to each of the following types
prove that

is

<f>

has

it

all

possible values

(p

- p')\

(P

'

'

Pq

~P

and by an equation of the sixth degree when


PI

q,

P-P',

q-q'*

~p
q-q'

'

it

'

pq>

p-p'
p-p'
- p'} (q - <?'), (q -

has

- p'q?

(pq'

all

values corresponding to

p*-4q.

or

pq'-p'q*

Expressing these functions in terms of the roots, the number of possible values of
each function becomes apparent.

Transformation of the Biquadratic into the


Reciprocal Form. To effect this transformation we make
65.

the linear substitution x = ky + p in the equation

ax 4 + 4bz 3 + 6c# 2 + 4r/# +

0,

which then assumes the form


fl&y +

4&Vry +

6 z72 /t-y + 4ty;//

-f

u, =

o,

where
Z/i

* ap +

C^ =

J,

+ 2bp +

c,

Z73

tfp

+ 35p 2 +

S^jo

rf,

&c.

If this equation be reciprocal, we have two


determine k and /o, viz.,

(See Art. 35.)

equations .to

2
/o

Algebraic Solution of the Gubic and Biquadratic.

136

eliminating

and

we have

7r,

the following equation for p

since

there are two values of

/;,

equal with opposite signs, correspond-

ing to each value of p.

The equation

when reduced by

the substitutions (Arts. 36, 37)

a- IT*
*

- Z7? +

3J5TZ7;

6//^

/-

12// 2 )

4
i

+ G,
2

4G Z7

+ a 2 /-

becomes
2

which

is

is

J7,

(a

US -

6GHU

- G* =

0,

a cubic equation determining Ui = ap + b\ and

(1)
if

we put

determined by the standard reducing cubic


4rf

This transformation*

jfr/0-f

,7=0.

be employed to solve the biquadratic ; and it is important to observe that the cubic (1) which
here presents itself differs from the cubic of Art. 62 only in

may

having roots with contrary

signs.

We proceed now to

express k and p in terms of a, j3, 7, 8,


Since the equation in y y
the roots of the biquadratic equation.
obtained by putting x = ky + p, is reciprocal, its roots are of the

form y l9 y^

y*

o = *yi +

hence we

y
p,

/3

- kfo +

p,

may
7

write

ft

- + />,
2/2

8 =A

-+p

y\

* This method of
solving the biquadratic by transforming it to the reciprocal
S. S. Greatheed in the Camb. Math. Journ*, vol. i.

form was given by Mr.

Transformation of Biquadratic into Reciprocal Form.

137

and, therefore,

-O-^Cy-p)-*

(a-p) (-/>)

1
,

from which we find

/3y-a8
/

+7 _a~

-/3

r -g)_.
-_
_ ( 7 -a)(/3-S)(a-/3)(
-a Sr
(P + 7

An

important geometrical interpretation may be given to


the quantities k and p which enter into this transformation.

Let the distances OA, ()B, OC, OD, of four points A, B, C, D,


on a right line from a fixed origin
on the line be determined
the
of
roots
the
a, /3, 7, S,
by
equation
air

also let Oj, 0,,


tlie

4-

4fcp

y)

= 0;

l9

F{

F F

2,

'
;

J^, J^',

quadratics:^

(*-)

0,

(-)(a--/3j-0

We have

4r/^

systems of involution determined by the

three following pairs of

+ 6 co? 2 +

be the centres; and

foci of the three

(.f

(*

(*- 7 )(*-S) =

>

0.

then the equations

0,B. 0,C = 0^4

OiD =

O^

2
,

&o.,

which, transformed and compared with the equations


(/3

p) (y

- p) =

(o

p) (S

prove that the three values of p are


of the three centres of involution

p)

/r,

&c,,

00 00 00
2,

1?

3,

the distances

from the fixed origin

0.

Also

z
since OiFi = k* 9 k has six values represented geometrically

by

the distances
0,F,, 0,^,';

whore OiFi + OiFi =


opposite directions.

0, &o.,

8,

0,F,';

3,

as the distances are

3',

measured in

Algebraic Solution of the Cubic and Biquadratic.

138

We can

from geometrical considerations alone find the positions of the centres and foci of involution in terms of a, )3, 7,
and thus confirm the results just established, as follows
Since the systems [FiBFi'G] and [FiAFi'D] are harmonic,
,

and
0,

if

F or FI

x represent the distance of

from the fixed origin

we have
1
d?

j3

Solving this equation, we find

x = p

or
,

whence

p =

7f,

--+- OFS

OK-OFS =
-^

A =

EXAMPLE.
Transform the cubic
ax*

3#2 +

%cx

-f *?

to the reciprocal form.

The assumption x - ky

- G Z7i 3 +
The

values of

/>

p leads to the equation

-f

3 Jf 2

2
Z7i

+ H*

==

0,

where

U\

a$

ap

J.

are easily seen to be

)8

&y
4- 7

a2

ya

2a*

$2

7-fa-

2^*

-f

^ - 2y

The geometrical interpretation in this case is, that if three points ', J5', C" be
with respect to
is the harmonic conjugate of
taken on the axis such that
B and C, B' of .# with respect to C and y and C of C with respect to -4 and 2?

then

we have

the following values of p and k

_ 6U

4-

OA-

OA

For the values of OA', OB', OC'

in terms of a,

OA'

7, see

Ex.

13, p, 88.

Solution of the Biquadratic by Symmetric Functions.

39

Solution of the Biquadratic by Symmetric Func-

66.

tions of the Roots.

The

possibility of reducing the solution

of the biquadratic to that of a cubic by the present method


depends on the possibility of forming functions of the four
roots a, j3, y, 8, which admit of only three values when these
It will be seen on referroots are interchanged in every way.
ring to Ex. 2, Art. 64, that several functions of this nature
exist.

These, like the analogous functions of Art. 59, possess

an important property to be proved hereafter, viz., any two


such sets of three are so related that any one function of either
set is connected with some one function of the other set by a
relation in terms of the coefficients.

rational

homographic
For the purposes

of the present solution we employ the


functions already referred to in Art. 55, since they lead in the
most direct manner to the expressions for the roots of the bi-

We

proceed accordingly
quadratic in terms of the coefficients.
to form the equation whose roots are the three values of

"**

/a + 0/3 + 0*7 +

the roots are interchanged in every way, and


These values are

when

and

since

S
we

(a

]3)

- 3Sa 2 - 2X -

find the following values of ^, t^

12

whence

a2

'

2fji

t*

- 2v = - 48

~,

1.

140

Algebraic Solution of the Cubic and Biquadratic.

Again, since

Sfc -

and

i/)*

= 24-,,

we have

also

t}

Hence

equation whose roots are t^ t^

tlie

(rr/)

or,

3H(a

2
substituting for (?

which

is

(<tt

ty +

its

(zH*

t:i

becomes
-

value from Art. 37,

+ //) 3 - a 8 / (a 8 * + IT) + a 8 JT=

0,

transformed into the standard reducing cubic by the


z
eft + If = a &.

substitution

To determine

a, /3, 7, S

we have

-4 7 -

the following equations


=

2,

a+/3-fy-fS = -4-;

along with

tt

from which we

find

-i

Solution of Biquadratic by Symmetric functions.

We

Lave

also

from the above values

of Jtfi,

J^, Jt z

the

equation

which one radical can be expressed in terms of the


other two, and the general formula for a root shown to be the
same as those previously given.

by means

of

It is convenient, in connexion with the subject of this Article,


to give some account of two functions of the roots of the biquadratic which possess properties analogous to those established
in Art. 59 for corresponding functions of the roots of a cubic.

Adopting a notation similar

we may

to that of the Article referred to,

write these functions in terms of A,

ing form

/u,

v in the follow-

S (07 + nS) +

01

(70 + |3S) +

a,

(a/3

+ 78),

-f

ai

(a/3

+ 78).

By means of the equations of Ex. 1, Art. 63, these functions


can be expressed in terms of the roots of the reducing cubic in
the form
j-

L=

0!

OJ0 2

0>

S,

-J-

5f =

01

+ CU0a +

0.6/3.

also be expressed, by aid of the equation of the


present Article connecting t and 0, in terms of the values of
t 2> t 3 , as follows
tfi,

They may

The functions L and Jf are as important in the theory of


the biquadratic as the functions of Art. 59 in the theory of the
cubic.
The cubes of these expressions are the simplest functions
of four quantities which have but two values when these quantities are interchanged in every way ; they are the roots of the

reducing quadratic of the reducing cubic above written, and


underlie every solution of the biquadratic which has been given,

Algebraic Solution of the Cubic and Biquadratic.

142

EXAMPLES.
1

Show

that L and

Increasing
2.

To

a,

by

A,

a, 3, 7, 8.

and Jf remain unaltered, since

a,

7, 8.

)3,

the values of

1202 =

and

120!=
2

M in terms of
L-

M,

X+

2
jtf,

we

0i, 02, 0s,

0.

find easily

2
Jif=08-7)(a-S)(a> -

X-

Jf = (7

(3-

a)

8)

Again, from these equations, multiplying the terms on both


remembering that 0i, 02, 03 are the roots of
4

we

find in terms of the coefficients the product of the squares of the differ-

ences of the roots

From

M are functions of the differences of

7, 8

w),

a>),

sides together,

and

- J*0 + /=0,

find

- JIP -

VT3

()8

- 7)

(7

- a)

adding the squares of the same terms,

also,

= 24

^ = (3 - 7)
~

8
)

(a

0) (a

2
JP) ^

^ (7

(T?

5)

(3

8) (y

S)

we have
-

2
)

and, since

(X

3 2
)

(3

5)

- 4Z 3 Jlf

substituting for these quantities their values derived

(a

2
j3)

(7

from former equations, we

have finally
6

3.

(&

- 7? (7 ~

2
)

(a

Show by a comparison

- 3) 2

8) (0

- S) 2 (7 -

8)

- 256

(7

- 27,7*).

of the equations of Art. 59 with those of the present


may be extended to the biquadratic by changing

Article that the results of the former

-7, 7-0, a-/B


respectively

67.

dratic.

into

-05-7)(a-8), -(y-a)(3-), -

and, at the same time, If into

-- 1, and G into

(a

- ft)

(7

~ 5),

167.

Equation of Squared PifferenceL of a BiquaIn a previous chapter (Art. 44) an account was given

of the general problem of the formation of the equation of difIt was proposed by Lagrange to employ this equaferences.
tion in practice for the purpose of separating the roots of a
given numerical equation ; and with a view to such application

Equation of Squared Differences of a Biquadratic. 143


he calculated the general forms of the equation of squared differences in the oases of equations of the fourth and fifth degrees
wanting the second term (see Traiti de la Resolution de$ Equations

Numdriques, 3rd

ed., ch. v.,

and note

Although

in.).

for

practical purposes the methods of separation of the roots to be


hereafter explained are to be preferred yet, in connexion with
;

the subjects of the present chapter, the equation of squared


differences of the biquadratic is of sufficient interest to be given

We

proceed accordingly to calculate this equation for a


It will appear,
biquadratic written in the most general form.
Ex.
was
in
with
what
in accordance
17, Art. 61, that
proved
here.

the coefficients of the resulting equation can all be expressed in

terms of

and

a, If, /,

The problem

J.

equivalent to expressing the following product in terms of the

is

coefficients of the biquadratic


2
2
2
{>- (3- 7) }{4>-(7-) }{4>-("-0) {<*>-(- 5)^} {*-03}

The most convenient mode

of procedure is to

5)

{0-( 7 -5)'}

group these six factors in pairs,

to expiess the three products (which we denote by n


ITo, n 3 ) separately in terms
of the roots of the reducing cubic, and finally to express the product HI IIz Ila in
tt'irus of a, II, I, J.

and

s <p -

IIi

-7 2+

{ (ft

7)

(a

a
)

- 5)

we

and, by aid of the results of Art. 61


OB

(a

easily derive the following expressions for

:-

hence, without difficulty,


IIi

Introducing

now

16 JT

Oi becomes

-f

(801

<1>~+

4J

P,

- 480 2

Reducing the product ni

V 8 -f 3Q

^4Q</> +
2

pressed in terms of P, Q,

f* +

3P</>

-f

H\

4
--J $>+

4S0203.

Ila IIs

18 J2^)

(3P

J?,

167 B s^,

Q,

$*

P</>

+ Qe

^F,

3.

Finally, restoring the value of

16

for brevity the notation


2

80i

by the

V-

4
</>

(8720

we have

as follows
2 Q)

result of

-J-

Example

+ 12<22

2
<J>

18,

page 89,

we

the equation of squared differences ex-

(P

-f

obtain

+ 36Q2fy -f 27^ 2 = 0.

8PQ -

267?)

8
<f>

0.

144

Algebraic Solution of the Cubic* and Biquadratic.

The following

is

the final equation in terms of

(96JT

It should

+a

/)

4
<

/ 2 - 288*27)

a, JET, /,

J*

+ 32 (1281T + 160 IZJ- 13aV)


8

+1152 (2JZZ- 37)

Jty

he observed that the value ahove obtained for

III

quadratic function of Q\
calculation

by

<

aid of the equation 02 #3

= #i 2

might have been conducted by eliminating

0i

--

8
<f>

256 (I 3

- 27/ 2 = 0.
)

can be expressed as a

;,,

and the subsequent

between

this quadratic

and

the reducing cubic.

68.

Criterion of the Nature of the Roots of the

Biquadratic*.

Before proceeding with this investigation

it is

stated (Art. 43), that when


to
the
nature of the roots of an
condition
with
any
respect
algebraic equation is expressed by the sign of a function of the

necessary to repeat

coefficients,

what was before

are

these coefficients

numerical quantities.

It is

supposed to represent real

assumed

also,

as in the Article re-

ferred to, that the leading coefficient does not vanish.


Using as before A to represent that function of the coef-

the discriminant] which is, when multiplied by a


-numerical
factor, equal to the product of the squares of
positive
the differences of the roots, we have, from the results established
ficients (called

in preceding Articles, the equation

where
It will be

/ - 27 J\
3

found convenient in what follows to arrange the

discussion of the nature of the roots under three heads, accord-

A vanishes, or (2) is negative, or (3) is positive.


(1) When A vanishes, the equation has equal roots. This is evident
from the value of A above written. Four distinct cases may be
noticed
(a) when two roots only are equal, in which case / and J

ing as

(1)

do not vanish separately (]3) when three roots are equal, in which
case / = 0, and J = 0, separately (see Ex. 2, Art. 61)
(y] when
;

* The
equation of squared differences was first given in this form by Mr.
Roberts in the Nouvelles Annales de Mathematiques, vol. xvi.

M.

145

Criterion of Nature of Roots of Biquadratic.

two distinct pairs of roots are equal, in which case we have the
= 0, aU - 12H 2
It can be
(Ex. 3, Art. 61).

conditions

readily proved by means of the identity of Art. 37 that these


hence these two equations,
conditions imply the equation A =
;

along with the equation A = 0, are equivalent to two independent conditions only. Finally, we may have (8) all the roots
equal ; in which case may be derived
= 0, 7=0,
independent conditions

from Art. 61 the three


and J"= 0. These may

be written in a form analogous to the corresponding conditions


in case (4) of Art. 43.

When A

(2)

ginary

negative, the equation has two real

is

This follows from the value of

roots.

and

two ima-

in terms of the

for when all the roots are real A is plainly positive and
f
k J- 1,
k J 1,
the proper imaginary forms, viz., h
are substituted for a, |3, y, S, it readily appears that A is positive

roots

when
also

when all the roots are imaginary.


When A is positive, the roots of the

(3)

equation are either all

This follows also from the value of A, for


can show by substituting for u, /3 the forms h k J - 1 that

real or all imaginary.

we

is

when two roots are real and two imaginary.


therefore, when A is positive, this function of the

negative

In the

case,

coefficients is

not by

determine completely the


remains still doubtful whether the

itself sufficient to

nature of the roots, for

it

roots are all real or all imaginary.


necessary to discriminate between these

The
two

further conditions
cases

may, however,

In order
be obtained from Euler's cubic (Art. 61) as follows
that the roots of this cubic should be all real and positive, it is
:

necessary that the signs should be alternately positive and


negative ; and when the signs are of this nature the cubic cannot

have a real negative root. We can, therefore, derive, by the


aid of Ex. 4, Art. 61, the following general conclusion applicable to this case

When A

is

positive the roots of the biquadratic

are all imaginary in every case except when the following conditions

are fulfilled,

viz.,

negative,

case the rdots are all real.

and a2! - 12H* negative

in which

146

and Biquadratic.

Algebraic Solution of the Cubic


EXAMPLES.

Show

1.

that if

H be positive, or H~
if

& not

(and

a pair of imaginary roots.


be negative, the cubic will have
2. Show that if

unequal,
(2)

0),

its

two imaginary, according


- 4H*.
greater than

two equal, or

(2)

roots

as

(3)

the cubic will bave

G2

(1)

is

all real

and

(1) less than,

equal to, or (3)


3. If the cubic equation

ao*3

have two

o2 -

where

roots equal to

2
i

-I-

i2

2JETi,

3a 2x

s =*

prove

s &>

os
+

<w

4. If

Sai*

3te

Sex

is -

rf

(x

-Z
*

r)

be a perfect cube, prove


(ae
5.

to)r

(W -

-f

- 0.

c2 )

Find the condition that the cubic

may be

3ia:2 -f

aa?

3c# +

rf

capable of being written under the form


I

where

$2 ) r2 -f (ad

ai, ^i,

(x

+ m (x -

ai)

n (x - 7 i) 8,

71 are the roots of the cubic


cnx*

3M

2
-r

$cix

rfi

0.

Comparing the forms we have


as=

Also

ai ai*

-f

/ai

<f

/ai

33i ai

m;3i

4-

zj3

-f 3<?i

4-

01+^1 =

Whence, multiplying these equations by d it

we

0,

<S:c.

Zc\, 3bi t ai, respectively,

find the required condition


(adi
6.

If

7 be

3 (bc\

a\d)

0.

the roots of the cubic equation

4
rationalize the equation
4

. ......... .

V'a:

and express the

l\e)

4
~
- o4 v/^

result in terms of the coefficients


^4ff.

126^*4

* ~ 7 =
o,

i,

^2,

s.

and adding,

147

Examples.
If ai,

7.

and

/8i,

a\x*

HI 3

Let

4-

whose

find the equation

ai GI

fa be the roots of the quadratic equations

aj,

fa

2bix

0i

4-

2 a? 4-

roots are the four values of ai 02.

Hi m

2
,

a^c*

Ans.
N.B.

a 2# 8

0,

(ai 02<J>

- 2&i i 2

This and the two following Examples

may

<j>

4- ci cj)

- Hi BzQ*

0.

be solved by expressing

<j>

by

radicals involving the coefficients.

of Ex. 7,

Employing the notation

8.
_

four values of

Let

2^Tia

a\c^

a^c\

In

this

9.

In the same case,

Example the

several values of

roots are the

2b\ b z .

-f

Am.

Let

form the equation whose

ai 4 02

(2ai a 2

4-

2(i* 2 4is

resulting biquadratic

if $>

(ai

az)

2*i)0

such that

JT^)

-I-

G=

- JTi J5T2 =

0.

0.

form the equation whose roots are the

^>.

s= rt!/> 2

^lj.

2.?7i 2

2 *i,

(ia 2

-f

a^-j

i/ J2 ) 2

2jtf

2 ci
2
<j>

-f

10. Show that when the biquadratic has a double root, the cubic whose
roots
are the values of p (Art. 65) has the same double root ; and find what this
cubic
becomes when the biquadratic has three roots equal.

/be

11. If
and
both positive, prove directly (without the aid of Euler's
cubic) that the roots of the biquadratic are all imaginary.
It appears from the expression for
in terms of the roots (Ex.
19, p. 62) that
is positive there must be at least one
when
roots h
pair of
k

V^T

imaginary
diminishing all the roots by A, and dividing them by k (which transformations
will not alter the character of the other
pair of roots 7, 5, nor the signs of
and J }
the biquadratic may be put under the form

Now

1),
*

or

whence

-f

4j0#

/ =

^c

-f

p*t
4-

6cx~

%p*c

4jy#

-I-

q,

_ 4^2 +
q

- jo 2

(^ 4- 1)

where 6c = ^

-f 1

3^
<

<

(q

4^2

_ ^y

and therefore

proving that

7 and

5 are imaginary

when JSTand /are


I

positive

(of.

Ex.

8, Art. 61).

148

and Biquadratic.

Algebraic Solution of the Cubic

12. If the biquadratic has

two

distinct pairs of equal roots, prove directly the

relations

In

this case the biquadratic divided

where

by

assumes the form

OQ

a&c

a\,

a -

- =

and

whence, comparing the forms

we

3J?

find

from which the above


lish the identity

=-

A2 ,

0,

/ - 3 H* =

relations immediately follow.

ThejiUident will easily estabArt. 61.


Also it should be

of these relations with those of Ex.

noticed that in this case only one square root

is

3,

involved in the solution of the

biquadratic (coming from the solution of the quadratic (x

a) (x

/8)

Find the condition that the biquadratic may be capable of being put under

13.

the form
/

In

(#

this case the second

2px 4

#)

and fourth

4m

(x

4 2px

coefficients are

q)

n.

removed by the same trans-

formation, and the general solution involves only two square roots.

Am.
14.

Prove that

0.

vanishes for the biquadratic

m (x

w)

n (x

w)

15. If the roots of a biquadratic, o,


,
7, 5 represent the distances of four
points from an origin on a right line ; prove that when these points form a harmonic division on the line the roots of Euler's cubic are in arithmetic progression,
and the roots of the cubic of Art. G2 in harmonic progression.
16.

Form

the equation whose roots are the six anharmonic functions of four
by the equation

points in a right line determined

= 0.
The

six

anharmonic

ratios are
1

149

Examples.
\

5)

7)

'

-A

-\

(a-j8)(7~5)
(y

*8 ~
also the equation

is

whose

a)

01-02

fl

=
23

S)

02-0i'
3

*-,*

OS-'yXa-S)

01

-0 3 '

roots are

one of the cubics

aft*

I2a lt

16 v//

- 277 2 =

0.

The equation whose roots are the ratios, with sign changed, of the roots of either
of these cubics is

4A

2
fa>

<f>

- 27/V>(4> -

I)

A e J3 - 27/ 2

where

The

I)

(see

Ex.

16, p. 88),

are the six anharmonic ratios.

roots of the equation in


This equation
can be written in a more expressive form, as will appear from the following propo'itions

<

The

(a).

as follows

six

anharmonic

*'

From

may

be expressed in terms of any one of them,

"

*'

n~?

^i

-^r*

'

the identical equation


03

we have

ratios

- 7)

_
(

+ 7 - a

5)

(^

5)

3)

7 -

5)

the relations

01

02

1,

-f

03

03

1,

-f

-=

1,

0^

rf>l

which determine
(b).

w and

If

all the anharmonic ratios in terms of any one of them.


two of the anharmonic ratios become equal, the six values of
2
o>
each occurring three times and in this ease J = 0.

For suppose 0i =

are

we have then from


0i*

whence

0!

-I-

w, or

0i

and substituting either of these values

the second of the above relations

0,
o>

in (a),

for

2
;

we

find all the

Also, since

iT^V
A
'""

F^
A

'

or

2
5(M--) =

fA

we have
1

#004

4i3 + 32 = 0.
2

o,

anharmonic

ratios.

150 Algebraic Solution of


(e).

When

one of the ratios

occurring twice

and in

4>j

(<J).

These

this case

m_

/ (see

one of the factors of

<f>

harmonic, the six values of

/=

_H^

if

Show

are

1, 2, -,

each

for if

-u-

or 2\

if

0,

Ex. 18,

p. 62).

under the following form

-P{(* + I)(f-)(f 17.

<p

as well as the converse propositions,

results,

writing the sextic in

is

and Biquadratic.

the Cubic

(see

Ex.

= nf*{ <* + )(+

he proved by

may

12, p. 119)

*)}.

that the equation


2
/t?

14s

1\

_ *(*-

I)

\p*

b satisfied by

the solutions which follow

18. Express

5(

(7

- 8) 2

as a rational function of 0i, 62, 0s

and ultimately

in terms of the coefficients of the quartic.

96

Bi+
C2H\
19.

Express
(^

- 72

2
)

(a

- 52

4-

2
2
7 -a

2
)

()8

~ 52

2
-f-

(a

as a rational function of 0i, 62? 0s-

This symmetric function

20.

Form

is

equivalent to

the equation whose roots are the several products in pairs of the

roots of a biquadratic.

The

required equation

is

the product of three factors of the type

2 -

<p

CT

21.
a,

7,

Form

4</>0,.

tt

- 47$> 2 (a^2 - 2c^> -f e) +

the equation whose roots are the several

are the

+- -

values of

16/<J>

~-

= 0.
,

where

roots of a biquadratic.

The required equation

is

the product of three factors of the type

--?-')(-)

-":+*?"*:+;-

Ant. 4 (a^1

-f

2$> -f

c)*

-I(a<p -f

2fy>

+ e) + J =

0.

151

Examples.
22. Prove
-

From

_\2

7-

2 \

the expressions for a, 3, 7, 8 in terms of


1
(

a 2 0i 4 2JT

which may be expressed in terms of

we have

0i, 02, 03,

2 -f

a, JT, /, /, as

2JI

above.

if

24.

0,

and

form

of the

Here

sum

or

3j>

0,

being a positive integer.

1, JP

or difference of

/ ==

ace

4 2bcd

aU ss

(ax

and

or

SJP

Prove that

can be resolved into the

is

*'"*

23. Prove

by

(ac

a perfect square

eb 2

ad*

czf-

i2)

two squares

c3

0.

2
F)y 4

(00

2
y 4 2 (ad

be)

if

yz

(a<

(ae

fo)yz

4-

(o^

3
)J

c2 ) a2

if

7=0.
25. If a, #, 7, 8 be the roots of the equation

solve, in terms of the coefficients ao,

V#When

Va 4

is rationalized,

Now,

a 4 Var-

and the

i,

j8

N//3

&c., the equation

4 Vir - 7 4 Vi~T=

+ V^ +

\/5

0.

coefficients substituted, for a, 3, 7, 5,

nibstituting CT0| Ui, Ut,

3)

Ut for a

03, 04*

we have

and reducing, we find

152

and Biquadratic.

Algebraic Solution of the Cubic

26. To obtain the equation of differences of a biquadratic, the equation of


semi-sums (Ex. 21, p. 150), and to solve the biquadratic by one and the same

transformation.

Substituting

x'

p for x, and using the notation of Art. 65,

ax'*

We can suppose x

x'

it

4 Vix' 3

-f

Z72#' a 4-

4 U&'

and p to have suoh values as

ax' 4

from which

+ 6 Uix' 2

Ui

-t-

(Tix"

0,

Z7<

satisfy the

Z/3

we have

0.

two equations

any value of p there are two values of


is eliminated we find an equation of the

appears that correiponding to

2
equal with opposite signs, and when x'
To obtain the values of p and sf in terms of the roots
5, of the biquadratic, assume
, 7,

sixth degree for p.


a,

pi

x'i

The equation

a,

in

*',

#'i

pi

whence

0,

therefore, obtained

2pi=a +

by eliminating

2#'i

p, is

jS.

the equation of

and the sextic in p the equation of semi-sums. By the mode of


reduction of Art. 65 the latter equation can be readily expressed in the form
semi-differences,

4(72 3

To

we have from

solve the biquadratic,

root of the reducing cubic

- IUz + /=

the last equation

17*

where

00,

6 is a

hence

U^ap + b^y/M-X;
from which,

(Compare Ex. 21.)

0.

x'*

Z7i*

+ 3JT+

finally,
2
/-a 0-2J2"

an expression with only four values, in which the root of the biquadratic is expressed in terms of a single root of the reducing cubic.
27. Prove that every rational algebraic function of a root B of a given cubic
equation can in general be reduced to the form
Co

-f

C\e

Do + DI$'
.

Let the given function be


tions of $ of

these

may

any degree.

By

/\

~-A where

and

$> (0)

ty (9)

successive substitutions

be reduced to a quadratic.

are rational integral func-

from the given cubic each of

Hence the given function

is

reducible to the

form
1

CQ -f ciO -f ezff

Equating this to the form written above, and reducing by the given cubic, we
obtain an identical equation, viz. LQ + Ltf -f ZzO 2 m 0,
where Q, L\, L\. are

L\ =

0,

Xs

as

0, to

We

D\.
<7i, D
have, therefore, the three equaVions
determine the ratios of C O\ t 2>o, DI.

linear functions of Co,

Zo* O t

153

Examples.

28. Prove that the solution of the biquadratic does not involve the extraction

when any relation among the roots a,


7, 5 exists which can be
by the vanishing of a rational function of a root $ of the reducing cubic.

of a cuhe root

expressed

Any

rational function of 9 can always be depressed to the second degree, as in

Hence the determination of will not involve the extracthe preceding example.
tion of a cube root
and the formula of Ex. 26 shows that the expression for the
;

root of the biquadratic will not then involve

29.

when

by any

(2U

(l),

30.

(4)-',

_J =

(8)07-a5 = 0,

(5)

(5

+7=

(2)

may

0,

(3) (y

a) (0

8)

(7-a)08-5)-a>(a-j8)(7-5) =

be proved as follows
it

Putting

(a

0,

j8)

(7

(6), (7)

5)

0,

3- 7 = 0.

= in the values of /and /, and


A independent of a\ may be thrown

a\

readily appears that the part of

form
o#4 (#o#4

922 +
)'~

270<) #3" (2a

24

os 2

2a2

3
)-

A^

and substituting for the latter quantireplacing a z , 3 , en by At, A$,


the values of Art. 37, we obtain the result.
Mr. M. ROBEIITS.

Now,
ties

Prove the identity

expanding,
into the

of the following values of p

(3)0,

+7_

(1)
(4),

This

root.

the equation

is satisfied

Am.

any cube

Find in each case the relation which connects the roots of the biquadratic

31. When a biquadratic has two equal roots, prove that Euler's cubic has two
equal roots whose common value is

21

and hence show that the remaining two roots of the biquadratic in
real, equal, or

imaginary, according as

2HI

3aJ

is

this case are

negative, zero, or positive.

Prove that when a biquadratic has (1) two distinct pairs of equal roots the
two terms of the equation of squared differences (Art. 67) vanish, giving the
conditions A= 0, ZHI
3o7"=
and when it has (2) three roots equal, the last
three terms of this equation vanish, giving the conditions 1=0, J
and show
32.

last

the equivalence of the conditions in the former case with those already obtained in
Ex. 3, Art. 61,f and Ex. 12, p. 148. Prove also that the equation of squared differences reduces in the former case to < 2 (a z <p + 12JT) 4 , and in the latter case tx*

CHAPTER

VII.

PROPERTIES OF THE DERIVED FUNCTIONS.


69.

tion.

Graphic Representation of the Derived Func-

APB

Let

be the

curve representing the pothe


/(#), and

lynomial
point on
to

it

any value

ble x =

OM.

corresponding
of the varia-

We

determine the

proceed to

mode

of re-

presenting the value of f'(x)


at the point P.
Take a

second point

Q on the curve,

corresponding to a value of
x which exceeds
by a small quantity

OM

OM =

ar,

MN =

//,

The expansion

Thus

h.

ON = x +

PM = /(#), QN =f(x +

also

5.

Fig.

/*).

of Art. 6 gives

f(*+h) =

f(x)

+/'

(x)

h +

-^ h*+

or

But

Now, when h

(1)

_
"

is indefinitely

_
~

_ ton
~

^ *

_
~

diminished, the point Q approaches,


the chord PQ, becomes the
;

and ultimately coincides with,

Maximum and Minimum

PT

155

Values.

PEN

to the curve at P
the angle
becomes PTM<
tangent
Also all terms of the right-hand member of equation (1) except
the first diminish indefinitely, and ultimately vanish when h- 0.
;

The equation

becomes therefore

(1)

tan P2VJ/

from which we
function/'

=/(#)

conclude that the value assumed by the derived

on the substitution of any value of x is represented by


made with the axis
by the tangent at

(x)

OX

the tangent of the angle

the corresponding point to the curve representing the function

Maximum and minimum

70.

mial.

Theorem.

mum

minimum

f (x)

Values of a Polyno-

Any value of x which renders f(x) a maxi= 0.


a root of the derived^yuation
(x)
Let a be a value of x which renders f(x) a minimum.
or

is

We

proceed to

ment

prove that/'

or decrement of

(x)

h)

/(a),

0.

We

x.

/(a) </(a +

hence f(a +

/>),

Let h represent a small

have, since f(a)


also /(a)

and f(a -

h}

</(a

is

h)

~ f(a) are both

the following two expressions are positive

incre-

minimum,

positive, i.e.

is very small, we know (Art. 5) that the signs


of these expressions are the same as the signs of their first
terms hence, in order that both should be positive, /' (a) must

Now, when h
;

vanish

and, moreover,

similar proof shows that

andf"

(a) is negative.

minimum
tion

(x)

furnish a

/" (a) must


when / (a)

f"

when

An

maximum

Thus, in order to find the

exactly
(a)

0,

maximum and

values of a polynomial /(a?), we must solve the equaEach root will


0, and substitute the roots in /(#).

maximum or minimum,

these being the sign of /"


it

be positive.
is

(x)

(x) is positive,

is

(x)

the criterion to decide between

when

negative, the value

the value is

the root
is

a minimum.

is

substituted in

a maximum ; and when

Properties of the Derived Functions.

156

The theorem

of this

Ar-

follows at onoe from

ticle

the construction of Art. 69


for

it is

value of ./(#)
at

as

minimum,

is

P, P'

when the

plain that

maximum,
or

6),

(fig.

as at p, p'

the

tangent to the curve will


be parallel to the axis OX,
Fig. 6.

and, consequently,

Pig. 6 represents a polynomial of the 5th degree. Corresponding to the four roots of/' (x) = (supposed all real in this case),
viz. OM, Om, OM', Om', there are two maxima, MP, M'P' ; and

two minima, mp,

m'p'.

EXAMPLES.
1.

Find the max. or min. value of


f(x)

= ** +

/
1

= - - makes
4

(See

!,

,,

f(x)

/"(*) =

4.

-49

a minimum.

-,

fig. 2, p. 16.)

2.

Find the max. and min. values of

/ (x) /'

(at)

x
3.

makes f(x)

makes

68, a

./(#)

6 (2x

I).

maximum.

67, a

minimum.

Find the max. and min. values of


f(x)

Here

=-

4.

- 3#2 - 36* + 14.


- x - 6), /" (x) =
(ac*

2# 3

- 6

f(x)

W + x-6.

(x)

3#4

16s

4- 6.r

has only one real root,

48.r

and

7.
it

gives a

minimum

value,

345.

Find the max. and min. values of


/(*)

The

roots of

maximum

(x)

10#3

17# 2

are, approximately,

value, the latter a

minimum.

-f a: -f

6.

-0302, 1-1031.

(See

fig. 3, p. 16.)

The former

gives a

157

Rollers Theorem.

Between two consecutive real roots


71. Molle's Theorem.
there lies at least one real root of
a and b of the equation f(x) =
the equation

(x)

0.

For as x increases from a to J, /(#), varying continuously


from f(a) tof(b), must begin by increasing and then dimmish,
It must,
or must begin by diminishing and then increase.
or
minimum
maximum
at
least
one
therefore, pass through
This value (./'(a),
value during the passage from f(a) tof(b).
suppose) corresponds to some value a of x between a and b,

which by the theorem of Art. 70


=

a root of the equation

is

o.

The

figure in the preceding Article illustrates this theorem.


and
observe that between the two points of section

We

maximum or minimum values, and between the


B and C there is one such value. It appears also

there are three

two points
from the figure that the number

of such values

consecutive points of section of the axis

Two

is

between two

always odd.

consecutive roots of the derived equation

may

not comprise between them any root of the original equation,

and

Corollary.

never can comprise more than one.

The first part of this proposition merely asserts that between


two adjacent zero values of a polynomial there may be several
maxima and minima and the second part follows at once
from the above theorem for if two consecutive roots of f'(x) =
;

comprised between them more than one root of / (x) = 0, we


should then have two consecutive roots of this latter equation
comprising between them no root of
dictory to the theorem.

(x)

0,

which

Constitution of the Derived Functions.


of the equation f(x) be a lf a 2 a 3
an

72.

roots

f(x)

In

(x

<n) (x

a*) (x

Q 3)

this identical equation substitute

...

(x

y + x

a n ).

for x\

is

contra-

Let the
have

We

158

Properties of the Derived Functions.

where

e#

ql

qt

ai

(a?

+#

ai)

+#

aa

(a?

a2)

= (#-a 2 )(#-a s)

Oa

**

ai) (x

(a?

(0-a,,)

+ #

a 3)

a,) (#

We

- a2

(a

...

a,)

(0

-/(*) +/'

(*)

/'

(a?)

=
=

a)

(x
.

(x

(x

n)

- aj,

a w _i),

a w ).

y+

Equating the two expressions


(a?)

01) (X

8)

_i)

have, again,
a-)

(- a n

+ ... +

ni)(#-

(a?

+ (*

**n>

(x

(#

(ir

- a2
)

a*) (#

- a)

ai)

'-

for

(x

- a n ),

(a?

- an +
)

y>

f (y

4-

we

#),

+ y.
obtain

as above written,

"
(^)

1.2

= the similar value of q n ,* in terms of x and the roots,

The value of/'

(x)

may be conveniently written as follows

X
73.
carder

ai

equation /(#) =

of the

1 of the first derived equation

is

**oot

a multiple
(x]

an

Theorem.

Multiple Roots.

a^

of

thi

ihe ordef

0.

This follows immediately from the expression given for/'

(x)

for if the factor (x - ai) m occurs in


in the preceding Article
= az =
= a m we have
/(#), i.e. if en
;

/
Each term
first,

?^

in this will

which

is a factor in

will
(x).

have

X-

a m+1

x - aw

have (x - a^ m as a factor, except


m " as a
- ai) m ~
ai)
factor; hence (x

still

(x

159

Determination of Multiple Roots.

COB. 1. Any root which occurs m times in the equation


occurs in degrees of multiplicity diminishing by unity in
f(x) =

m-1

the first

is

derived equations.

Since/''

(x) is

from/(#),

it is

derived from/' (x) in the same manner as/" (x)


evident by the theorem just proved that j* (x)

m"2 as a factor. The next derived


function,
m~s and so on.
ai)
/"'(#), will contain (x
C OR 2. Iff(x) and Us first m-1 derived functions all vanish
m
a
value a of x, then (x-a) is a factor inf(x).
for
This, which is the converse of the preceding corollary, is
will contain (x

ai)

most readily established directly as follows


Eepresenting the
derived functions by/! (#)>/* (#)>
-/m-i (x ) ( see Art. *>)> an <l
- a for #, we find that/(#) may be expanded
substituting a +
in the form
:

from which the proposition

Determination

is

manifest.

multiple Roots. It is easily


inferred from the preceding Article that if f(x) and /'(#) have
"
- m will be a factor in
a common factor (x - a)"
a]
(x
74.

of*

f(x]

for,

by Cor. 1, the m 2 next succeeding derived functions vanish


as well as f(x] and/' (a?) when x =* a
hence, by Cor. 2, a is a
In the same way it appears that
root of f(x] of multiplicity w.
;

if

f(x) and f(x) have other

common

(s-r)'the equation /(a?) =


will have
to 7, r roots equal to 8, &c.

In

1
,

factors

(*-)",

&o.,

roots equal to

/3,

g roots equal

whether any proposed equation


has equal roots, and to determine such roots when they exist,
order, therefore, to find

we must
Let

common measure of f(x) and f (x).


The determination of the equal roots will

find the greatest

this be* 0(o?).

depend on the solution of the equation

(x}

0.

Properties of the Derived Functions.

160

EXAMPLES.
1

Find the multiple roots of the equation


20

The G.

C.

factor in /(#).

M. of /(a?) and/' (a?) is easily found


The other factor is # + 5.
after

Whenever,

remaining factors,

it

0.

to

be x

hence

(a?

2)

is

determining the multiple factors of /(a?), we wish to obtain the


will be found convenient to apply by repeated operations the

method of division of Art.

divide twice by x - 2, the

we

Here, for example,

8.

calculation being represented as follows

Thus

and

5 being the

two

coefficients left, the third factor is

5.

-f

This

operation verifies the previous result, the remainders after each division vanishing
as they ought.
2.

Find the multiple

and the remaining

roots,

x5

The G.
a factor in

C.

M.

f (x}.

of f(x) and /'

10#2

(x) is

factor, of the equation

found to be #

0.
2

- 2x +

(x

-I) 3 (a + 3* +

Hence

1.
1,

we

(x

I)* is

obtain

6).

Find the multiple roots of the equation


x*

Dividing three times in succession by x

/(*)=
3.

15*

-f

The G. C. M.
Hence
3.

_ 2# 3 - 11#2 4 12x + 36 =

of /(#) and /' (x)

is

a;

- x - 6.

The

0.

factors of this are

-f

2 and

/(.).( + 2)(*-3).
4.

Find

all

the factors of the polynomial


5
f(x) msc*- 5z

6**

+ 9#3 - Ux~ - 4x +

Am.

f(x)

(x

8.

1) (x -f I)' (x

2)5.

The ordinary process of finding the greatest common measure of a polynomial and its first derived function may become
very laborious as the degree of the function increases. It is
wrong, therefore, to speak, as

is

customary in works on the

Theorem.

161

of Equations, of the determination in this way of the


multiple roots of numerical equations as a simple process, and

Theory

one preliminary to further investigations relative to the roots.


is chiefly in connexion with Sturm's theorem that the
opera-

It

any practical value. The further consideration of


roots
is deferred to Chap. X., where this theorem
multiple
will be discussed.
It will be shown also in Chap. XI. that the
tion is of

multiple roots of equations of degrees inferior to the sixth can,


in any particular instance, be determined from simple considerations not involving the process of finding the greatest

common

measure.
75. This and the succeeding Article will be occupied with
theorems which will be found of great importance in the sub-

sequent discussion of methods of separating the roots of equations.

Theorem.

In passing continuously from a value a-hofx


to a value
than a real root a of the equation f(x]
a + h a little greater, the polynomials f(x) and f'(%) have unlike

little less

signs immediately "before the passage through th$ root,

and

like signs

immediately after.

Substituting

a-

A in f(x) and/'

(a?),

and expanding, we

have

/(a

A)

T^ow, since /(a) =

=/() -/'

(a)

+-

A'

- ....

the signs of these expressions, depending


When the sign of h is
terms, are unlike.
the same. The
the
become
of
the
changed,
signs
expressions
theorem is therefore proved.

on those of their

Corollary.

0,

first

The theorem remains true when a

root of any order of the equation

(oi)

Let the root be repeated r times.

a multiple

0.

The following functions

(using suffixes in place of accents) all vanish


/(a),/i(a),/,(a)

is

A. (a).

Properties of the Derived Functions.

162
In the

series for /(a

do not vanish

h)

and/'

(a

h) the first

terms which

are, respectively,

These clearly have unlike signs but when the sign of h is


changed the signs of the terms will become the same. Hence
;

the proposition

is

established.

76. Extending the reasoning of the last Article to every


consecutive pair of the series

we may

state the proposition generally as follows

Theorem.
root a, a value a

When any
little

little

f (x)

has an r-multiple

inferior to a gives to this series of r functions

signs alternately positive

and a value a

equation

and

negative, or negative

and

positive

superior to it gives to all these functions

the

moreover, the same as the sign offr (a) 9


the first derived function which does not vanish when a is substituted for x.

same sign; and

this sign is,

In order to give a precise idea of the use of this theorem,


let us suppose that/5 (a) is the first function which does not
vanish when a is substituted, and let its sign be negative
;

the conclusion which


for a value a

/3 ,/4,/*>

-h

of

may

be drawn from the theorem

x the signs of the

is,

series of functions/,

that

/i,/2,

are

+-+-+-;
and

for a value a

+ h

of x they are

for before the passage through the root the sign of /4 must be
from that of 6 ; the sign of /8 must be different from

different

that of /4, and so on;

and

after the passage the signs of all


It is of course assumed here

the functions must be the same.


that h

is

so small that

interval through

no root of /6

which x

travels.

(a?)

is

included within the

Examples.

63

EXAMPLES.
Find the multiple roots of the equation

1.

xi

f(x)

12# 3

-f

32#2

-f

- 24#

AT

^*.
Show

2.

0.

(*

/(*)

4 6*

2)*.

that the hinomial equation

z* - a

cannot have equal roots.


3.

Show

that the equation

x - nqx
will

have a pair of equal roots

4.

= r-

if q n

(n

1)

4-

Prove that the equation

x6

Apply the method of

bpi? +

-f

has a pair of equal roots when q 9


roots it mast have a second pair.
6.

5
4jt?

5j0

and that

if it

have one pair of equal

Art. 74 to determine the condition that the cubic


**

3 IIz

+ G=

should have a pair of equal roots.

The

last

remainder in the process of finding the greatest

6.

common measure must


Ans. G~

vanish.

Apply the same method

to

show that both O and

H vanish

-\

4 1/ 3 =- 0.

when

the cubic

has three equal roots.


7.

If o,

7, 5 be the roots of the

/'()+/'
<;au

biquadratic/^)

08)

0,

prove that

+/'(?) +/'($)

be expressed as a product of three factors.

Am.
8.

If

a, 0, 7, 5,

(a+ 3

-7-

&c., be the roots of/(j?)

5) (a -f

0,

and

7-

J8

5)(a

a', 0', 7',

-f

j3

&c., of/'

7).

(x]

0,

prove

and that each is equal to the absolute term in the equation whose roots are the
squares of the diil'ereuces.
9. If the equation

& + pi

have a double joot

a,

3C

n- 1

-f

p2 X"~

prove that a

is

-I-

....

-f-

p n -l X + pH

a root of the equation

SB

Properties of the Derived Functions.

164
Show

10.

that the

max. and miu. values of the cubic

ax*

3bx*

3cx

+d

are the roots of the equation

Vwhere

is

2fir P

+ A =

0,

the discriminant.

If the curve representing the polynomial f(x) be moved parallel to the axis of y
(see Art. 10) through a distance equal to a max. or min. value p, the axii of x wiU
will have equal roots.
Hence
become a tangent to it, i.e. the equation/ (x)
p =
the max. and min. values are obtained by forming the discriminant of f(x)
p, or
'
by putting d - p for d in G* + 111 = 0.
11.

Prove similarly that the max. and min. values of


ax*

4fcc* -f 6<?*2

4dz

are the roots of the equation

ay where

is

3(a

J- 9^2

3 (a/ 2

-f

p*

- l&HJ)

- A=

0,

the discriminant of the quartic.

12 Apply the theorem of Art. 76 to the function

f(x) s

x*

~-

7a-

-f

Ibx* - 13;r

-f

4.

"We have
-f

2 (6**

/<(*)

Here /a

is

(sr)

the

first

30ar

- 21* +

13,

16),

24.

function which does not vanish

when ^ =

and/a

(1) is

What

the theorem proves is, that for a value a little less than 1 the signs
-f -, and for a value a little greater than 1 they are all
of/, /j, /2, /a are +
are able from this series of signs to trace the functions/, /i, &c., in
negative.
the neighbourhood of the point x = 1. Thus the curve representing/^) is above

negative.

We

the axis before reaching the multiple point x = 1, and is below the axis immediately
and the axis must be regarded as cutting the curve in three

after reaching the point,

3
coincident points, since (x
I) is a factor in /(#).
Again, the curve corresponding
below the axis both before and after the passage through the point x = 1.
The curve representing fi(x] is above the axis
It touches the axis at that point.

to /i (x) is

before,

and below the axis

after,

the passage, and cuts the axis at the point.

CHAPTER

VIII.

SYMMETRIC FUNCTIONS OF THE ROOTS.


77.

IVewtou's

Theorem on the Sums of the

We

of the Roots.

I* o were

now resume

the discussion of symmetric


functions of the roots of an equation, of which a short account

has been previously given (see Art. 27)

and proceed

to prove

certain general propositions relating to these functions


PROP. I. The sums of the similar powers of the roots of an
:

equation can be expressed rationally in terms of the coefficients.

Let the equation he


xn 4

f(x)

x -

af^ 1 +

jW

ai) (x

aj,)

"'

(f

We

2
proceed to calculate Sa
the usual notation, *,, s s > . . s m

3)

Sa

We

+ ...+;?
.

3
,

(x

Sa m

a*)
;

or,

0.

adopting

in terms of the coefficients

have, by Art. 72,

x
(w

we

/()
x - a

find, dividing

= xn - 1 +

+ Pi

by
-f

the
a

4-

method

of Art. 8,

-f

3
(I

j0^,

166

Symmetric Functions of

If, in this
cii,

a2

we

f(x]

we

by adding

= nvn -1 4

Xn

np

all

*
-r

replace a

and put sp

n in succession,

have,

equation,

the Roots.

by each of the quantities


a wp
Sa p = a\ p -a/ +
.

-\

these results, the following value for

S2

whence, ooniparing this value of f'(x) with the former, we


obtain the following relations

=
i

0,

4 4;7 4

... 4

The

first

the second

We

(n

0.

in terms of p\, p*>


pn \
equation determines
the third 8 and so on, until n_j is determined*
6'

way
p\

4 2p 2 * -

64

= - Pi 6

Having shown how


terms of the

find in this

- Pi,

0,

coefficients,

2,

4/> 4 ,

(tf-p^pi

*3>

-i

we proceed now

-f

-jp

) ;

&o.

can be calculated in
to extend our results

167

Newton's Theorem.
to the

For

sums of

this

powers of the

all positive

roots, viz. s w , s n +i,

sm ,

purpose we have
'
Xm n f(x) a Xm

Keplaciug, in this identity, x by the roots a ly a 3 ,


sion,

a n , in succes-

and adding, we have

+ J0l*ro-i +

*m

Now, giving

and observing that

JVro-a

0.

the values n, n + 1, w + 2, &c., successively,


= w> we obtain from the last equation

jP a s-a

+ P^n-i +
+ JpiSn+i +

n+2

JV'in-n

Hence the sums

/>2*n

+ wp

0,

0, &C.

f;r$i

+ ^n

.S'

0,

of all positive powers of the roots may be


of the coefficients.
And by

expressed by integral functions

transforming the equation into one whose roots are the recia n and applying the above formulas,
procals of ai, aj, a 8
,

we may

PROP.

78.
root* of

of the

express similarly all negative powers of the roots.


II.
Every rational symmetric Junction of the
an algebraic equation can be expressed rationally in terms

coefficients.

theorem for integral functions


symmetric functions can be reduced to a

It is sufficient to prove this

only, since fractional

whose numerator and denominator are

single fraction

both

Every integral function of ai,


number of terms of the form
a numerical constant and if this

integral symmetric functions.


a a a 3 . . . a n is the sum of a
,

j^W

a-/ a s

r
.

where

is

symmetrical we can write it under the form


all the terms being of the same type.
ThereJ!VSa/ a/ a/
be
that
this
can
if
we
fore,
expressed rationally
prove
quantity
function be

in terms of the coefficients, the theorem will be demonstrated.


shall first establish the following value of the symmetric

We

function

fiai

p a a9

Saf a 2 q =

sp s

(1)

168

Symmetric Functions of

To prove

we multiply

this,

8P
sq

= af

-f

+ a 2*

cu

the Roots.

together sp and

a? +
+

a-f

/ +

sq ,

where

a,/,

<in

9
;

whence

or

which expresses the double function 2cti p a 2? in terms of the


single functions sp s q sp + q in the form above written.
,

We

proceed

now

to prove a similar expression for the triple

function, viz.,

f
Multiplying togetlier

we

^ai p a 2 9 and

sr ,

(2)

where

a fa!

obtain an expression consisting of three different parts,


Sn^a/, and Ea/a, 9 a3r

terms of the form

Sa^W,

viz.

Honoe

a formula connecting double and triple symmetric functions.


But, by (1),

Substituting these

Sa/a 2

the series

values,

we

find

the

triple

function

expressed as above in terms of single functions in


i,

$3 , ss>

&c.

In the same manner the quadruple function l&afafiai'a*

169

Newton?* Theorem.
can be

made

depend on the
$
&c.
and
a

to

ultimately on

s i9

*afa<paf 9 and

triple function

so on.

Whence,

finally,

rational symmetric function of the roots may be


in terms of the coefficients, since, by Prop. I., i,

can be so expressed.
The formulas (1) and (2) require to be modified
of the exponents become equal.

p=

if

Thus,

*$,

&o.,

when any

and the terms in (1) become


therefore Sa^a^ = ^Sa/a/; whence

- |

Zafaf
if

8,

a fa-? - a/<u*,

equal two and two

Similarly,

every

expressed

**q

-s

2ai'Wa

r in

by interchanging the roots in

ai

770
* o

(V

W8

2p ).

r
3

the six terms obtained

become

all

equal

hence

in general,

And,

repeated 1

exponents become equal, each term

if t
I

is

times.

EXAMPLES.
1.

Prove

2iPa2a3 r a4* =
2.

SpSgSr*,

- 2sp sqt>

+a

22^,^>r 4

-f

Prove

79. PROP. III.


P\>P*y

jn>

The

value

of

sr ,

in

expressed

terms

powers ofy, of -r log y

J\-\

Since

an +^?
putting

^"

for

4 p&P~* +

a?

4-

jt?

(a?

- m) (a?

in this identical equation,


*

of

coefficient of <f in the expansion, by ascending

-I-...

+ pnyn -

(1

we

a,)

find

(^

- a n ),

170

Symmetric Functions of the Roots.


the Napierian logarithms of both sides,

Now, taking

&c.

P^y+p%
1

...

-Pip*
1

-PIP*

Therefore, equating eoeiiioieiits of

whc re
j

Pr

is

in both expansions,

^/

the coefficient of y r in log

w
//

From

the above identical equation


than ^) involves the coeflicients

6> (r les.s

J.

it

p^

may

be seen that
.

f> 2 , jt> 3 ,

;> r

only

and, therefore, pr+i, p r **y


pn may be made to vanish without
nif noting the form of tiie expression of &y in terms of the coetficionts.

To express the

80.

powers

coefficients

in terms

of the sums of the

of the roots.

Sin oe
1 + piy

+ ^ 2 y2 +

. . .

+ p^T a

(1

ojy) (1

atf/)

... (1

- an
//),

we have
log (1 +jpjy +
and, therefore,

..

+p H t/)

- - yi -

*8

---- y

*r

- . ;

(1)

171

Newton's Theorem.
which becomes by expansion
1

"a*
r-

obtain values for

and we

see that

pr

the coefficients of the different powers of y,


p i9 p^ p^ . . . p n in terms of $,, s29

Now, comparing

we

...;

involves no

Bum

of powers

beyond

sr .

If the identity (1) be differentiated with regard to y, the

equations of Art. 77 connecting the coefficients and sums of


powers may be derived immediately from the resulting identity.

important to observe that the problem to express any


symmetric function of the roots in terms of the coefficients, or
It

is

terms of the sums of the powers of the roots, is


perfectly definite, there being only one solution in each case.
General expressions, due to Waring, for s m in terms of the

any

coefficient in

coefficients,

and

p m in terms of the sums of the powers of

for

the roots, will be given in a subsequent chapter.


EXAMPLES.
1.

Determine the value of


<f>

where

ai,

a2

as

...

(ai)

<J>

(n-z) -f

arc the roots off(x)

a,,

-h

0,

a ><)

and

is

integral function of x.

We haye

fM
ft

1
"

"T"

f(x)

and
"

"/>"J

J (X)

0,1

X-

X~

an

any

rational

and

172

Symmetric Functions of the Roots.

Performing the division, and retaining only the remainders on both sides of thii
we have

equation,

"

'

_.

x -

f(x)

on

4>

(an)

whence
IiV'-

+ -flu"' 2 +

-!

and, comparing the coefficients of

x n~ l

2<J>

on both

Ho 2.

3.

Prove that syp

is

(ai) (s

02)

(a;

...

03)

- o)

(a

sides of this equation,

5<p (ai).

in the quotient of the division of f(x)


the coefficient of --1'- 1

arranged according to negative powers of x.


Prove that *-pis the coefficient (with sign changed) of

ad'~

in the

same quo-

tient

arranged according to positive powers ot x.


4. If the degree of
(x) does not exceed n
<f>

where

sum obtained by giving

denotes the

2,

prove

values from

all

to

inclusive.

r=l

We

have, by partial fractions,


<J>

A\

(a;)

/,

f(x)

and, multiplying across by f(sc)


<>

(F)

<^>
:

^(ar)

.--

-p

^'(^2)

x - a*

01, 03,

_ (2J

<

^_

ai

4?

a>2

and putting x

(ai)

(ai)

a\

in succession,
1

(<*)
^_ B ^

.f.

f\o-n}

&n

whence

When

^> (a;) is

function of -,

We

it

of the degree

n - 2

expressing the

readily appears that there

is

first side

of the equation as

no term without x

as a multiplier.

have, therefore, comparing coefficients,

As f may be any

we have

rational

and integral function of degree not higher than


which are worthy of special notice

the following particular oases

-2

n-3
-

/^__^
(a)

0,,

(a)

a
... 3 ^rpr

(a)

0,,

2 JTJ/

(a)

0.

- 2f

Order and Weight of Symmetric Functions.


Given the following n

*.

2 *r =

0,

r-1

* .

rn

rn

r/

express the

2 equations between n variables #1, #2,

173

0*

Xr =

----

2a ^=
"

M" 3

0,

r-1

r-1

variables in terms of

two ntw variables X\, Xz.

Xl

-f <Xr

^2
.

81. Order and Weight of Symmetric Functions.


The degree in all the roots of any term of a symmetric function

of the roots [see Art. 28] is called the weight of the function.
highest degree in which each root enters the function is

The

The

2 8
weight, for example, of Sa/3 7 is six, and
It has been proved (Art. 28) that in the value

called its order.


its

order three.

in terms of the coefficients of


roots the

sum

We

of the function.
to

to the

pn

of

term

is

equal to the

of the

weight

now prove another

symmetric functions,

ficients pi y pt,

any symmetric function

of the suffixes in each

proposition relating
The degree in terms of the coef-

viz.

the value of

any symmetric function

is

equal

order of the symmetric function.

This can be readily inferred from the equations of Art. 23,


since the value of each coefficient in terms of the roots contains

any

power only, and therefore the highest


the coefficients will be the same as the degree of the

root in the

degree in

first

corresponding symmetric function in any individual root. The


value of Su 2 j3 2 , for example, is p^ - %pip* + 2p^
The degree
of this function of the coefficients is two,

which

is also

the order

of the symmetric function.

As the proposition just stated is of importance, we add


another proof, in which the symmetric function multiplied by a
suitable power of a is expressed as a homogeneous and integral
function of the coefficients

>

a*

#i> #2>

^ ae f

rm i n which

the result will usually appear in subsequent applications.


Replace the coefficients

Pn

a\

by

-,

ff a

,....

174

Symmetric Functions of

Now,

if

a2 ,

(cti,

a n ) denote any rational and integral

the Roots.

symmetric function of the

we have

roots,

where A is the degree, in the coefficients, of F(a^ a , a^ . . # n ),


a homogeneous and integral function of the coefficients, not
.

divisible

by

We require

now

show that h

to

For

the order of $.

is

this

purpose change the roots into their reciprocals, and, therefore,


a l9
a n into a n a n -i,
Whence

= F(a
(,,...)
fl/

0fA/>

n,

a^,

(1)

) ;

Cl 2

also

/I 1
7 T'

~
1

'

(oi, a.,

a3

the order of </>, and


an integral function not diviall
of
the
the
a n p being the
roots; (aia 2 a, .
by
product
lowest common denominator of all the terms.
Substituting in

where p

is

<//

sible

(1),

we have
af $

From

(ai, a,,

ficients

F(a n

a n _^

iu a

F(a n

h, ip

and

*>
,

# M _i,

(a h a 2

if it
)

).

p is equal to h for if p
a M ) would be divisible by the

this equation it follows that

wore greater than


product

a t P~ h

...

were

would be

less,

the function of the coef-

divisible

by

a n> both of

which

suppositions are contrary to hypothesis.


82.

Roots.

Calculation of Symmetric Functions of the

Any

rational symmetric function can be calculated

by aid of the proposition of Art. 78. In practice, however,


other methods are usually more convenient, as will appear from

We

shall show
the examples which follow the present Article.
also, when this subject is resumed in the second volume of this

work, that use

may

be

made

of

methods founded on the prin-

ciples there explained to facilitate in


^calculation of symmetric functions.

many

instances

the

175

Calculation of Symmetric Functions.

The number
is

of terms in

any symmetric function of the roots


For example, the number of terms in
- 2), this
is n
of the n th
(n

easily determined.

2!

a* a 3 of the equation

degree

(n

1)

number

of permutations of n things taken three


If the exponents of the roots in any term be not all

being the
together.

different, the

number

terms will be reduced.

of

for a biquadratic consists of twelve terms only (see

and not of twenty-four,

Thus Sa 2 /3y
Ex.

two permutations

since the

6, p. 48),

a/3y, ayj3

2
2
The student
give only one distinct term, viz., a /3y, in 2a /3y.
acquainted with the theory of permutations will have no diffi-

these

reductions in

culty

in

When

two exponents of roots are equal, the number obtained

effecting

any

case.

particular

all unequal is to be divided by


1.2; when three become equal this number is to be divided
by 1 2 3 and so on. In general, the number of terms in
of the equation of the n th degree, each term conSo? a 2 ? 3r

on the supposition that they are

taining

roots,

and

i/

of the indices being equal,

n(n-l)(n-2)
i

(n

w+

is

1)

.2. a ...v

When the highest power iu whiuh any one root enters into
the symmetric function of the roots is small, i.e. when the order
of

the function

Art.

(see

81) is low, the methods already


be employed with advantage for the

illustrated in Art.

27

calculation of the

symmetric function.

It

is

may

important to observe that when any symmetric function,

whose degree in all the roots (i.e.


in terms of the coefficients p h p^

its

weight)

is n, is

calculated

equation of the
n th degree, its value for an equation of any higher degree (the
numerical coefficients being all equal to unity) is precisely the

same

for it is clear that

no

p n for the

beyond p n can ei t u into


this value, and the equations of Art. 77, by means of which
the calculation can be supposed to be made, have precisely the
same form for an equation of the nth degree as for equations of
all

higher degrees.

It

is

coefficient

also evident that the value of the

symmetric function for an equation of a degree

same

m (lower than n)

176

Symmetric Functions of the Roots.

obtained by putting p m +\, j0 m , .


p n all equal to zero in the
calculated value for an equation of the nth degree, since the
is

th
equation of lower degree can be derived from that of the n by
putting the coefficients beyond p m equal to zero ; and the corre-

sponding symmetric function reduces similarly by putting the


a n each equal to zero.

roots crn-m

EXAMPLES.
1.

Calculate

2j 2 0203

of the roots of the equation

l
pi x"~

a?" 4-

4-

pi x-*

4-

. .

4-

p n ,\x

pn =

0.

M ultiply together the equations

In the product the term a\ 2 02 as occurs only once


from the product of ai by a? as 04, of

times, arising

and

of 04

by

the term 01 02 as a* occurs four

by ai3i)

0.1

of

0.3

by

ai

Hence

oia-jos.

2a>~ 02

a.i 4-

4^ai

therefore

2ai

aa ag

^^iy'a

(Compare Ex.

4pi.

If the calculation were conducted by the method of Art. 78,

2ar
which

leads,

03

a-i

= i2 s^ -

*i $3

J*2

4-

we

6,

Art. 27.)

should have

4,

on substituting the values of Art. 77, to the same result; but it is


much more simple, since the values of

clear that in this case the former process is


*i, 8 2 ,

2.

&o., introduce a

Calculate

number

2Sara2

of terms which destroy one another.

for the gem'ial equation.

Squaring
2ai as - pz,

we have
2oi

In squaring

a2

2
4-

22ai~ a2 03

4-

62aj ag as

= pz*

a-2 as ai will arise from the product of


0104 by 02 as; hence the coefficient of
ai 02 03 04 iii the result is 6, since each of these occurs twice in the square.
The
result differs from the similar equation of Ex. 8, Art. 27, only in having 2 before
the term ai 03 as 04* Hence, finally,

ai 02

by

as a $,

it is

evident that the term ai

or of 0103

by

as a*, or of

177

Examples.
3. Calculate

We have,

2ai 8a2 for the general equation.

as in

Ex.

9, Art. 27,

Hence, employing previous

results,

2ai a2

= p\ *pt

2
2/?2

- pi pa

-f

4/?4.

Calculate 2ai 2 02 2 <x3 for the general equation.

4.

The result will be the same as if the calculation were made for the equation of
the fifth degree.
To obtain the symmetric function we multiply together 20102 and 2010203 ; and
consider what types of terms, involving the five roots 01,02, an, en, as, can result.
The term ai 2 02 2 as will occur only once- in the product, since it can only arise by
multiplying 0102 by aitwas. Terms of the type 01-020304 will occur, each three
since ai 2 oj084 will arise

from the product of aia2 by 010304, of 0103 by


it cannot arise in any other
The term
way.
10203014 as will occur ten times in the product, since it will arise from the product
of any pair by the other three roots, and there are ten combinations in pairs of the
times

ia204, or of aiom

010203;

by

We have,

five roots.

and

then, for the general equation,

2ai022aia203

2ai 2 02 2 O3

-f

32ai 2 O203a4 + 102ai 02030405.

[We can verify this equation when n = 5, just as in Ex. 9, Art. 27 for the
product of two factors, each consisting of 10 terms, will contain 100 terms. These
2
are made up of the 30 terms contained in 2oi 2 a2 03 along with the 20 terms con;

tained in 2o) 2 a20304, each

taken three

tiniefc ?

and the term 0102030105 taken 10

times.]

Thus the
;

calculation

for

which we

of the required

2oi 201020304

Henee,

The

symmetric function involves that of

easily find

-Soi 02ai04 -f

620102030405.

finally, \ve obtain

process of Art. 78 would involve the calculation of *&; and

would be introduced through the values of

ti,

j a , Ace.,

many

terms

which disappear in the

result.

6.

Find the value of 2oi 2 a2 8 oso4 for the general equation.


multiply together 20102 and 201020304, and consider what types of terms

We

can arise involving the six roots

01, 02, 03, 04, 05,

Terms

a.

Theteiniai 2 ai; 2 a304 can occur

of the type o^agasoias will each occur four


times, this term
arising from fte product of oio 3 by 01030405, or of aio by 01020405, or of aia*
by

only once.

Symmetric Functions of the Roots.

178

by 01020304. The term 01020*040608 will occur fifteen


number of combinations in pairs of the six roots. Hence

or of 0105

ttie^asas,

this being the

2oi02 ^01020304

We have again,

2ai 2 02 2 a304

for the calculation of

2ai 20102030405

Hence,

finally,

2oi 2 a22 a3 04
2

Find the value of

6.

4-

42ai 2 a2a304O5

times,

6201030*040600.

5ai 02030405,

2ai 2 020s04a5

62ai 0208040500.

2oi o2 03

in terms of the coefficients of the general

equation.

2010203 2ai0203

from which

we have

2010203,

Squaring

we

2at

O2 os

4-

202010203040504,

find
2
2
2
2ai a2 o 3

83.

22ai 2 O22 0304 + 62ara20304a5

-*-

Homogeneous Products.

There

in

are,

general,
a n which
several symmetric functions of n quantities m, a 2
have the same weight, and amongst these may be included two
,

more which have the same order as well as weight. Of any


n letters there are, for example, the following symmetric
functions whose weight is four
or

The sum of all such symmetric functions of weight r is called


the " sum of the homogeneous products of r dimensions" of the
It is easy to see that
n letters. This sum we denote by Tl r
.

II r is

(1

the coefficient of xr in the following product of n factors

+ aiX -f

aiV + ...)(!

2^ + a a

The examples which

V + ...)...(!+ a x + a V +
n

.).

follow include the most fundamental

propositions connecting the sums of the homogeneous products


with the coefficients of the equation whose roots are ai, a 2 , . a,-.
.

EXAMPLES.
1.

Prove

n^
-"^Tw*
,__!

a"**" 1

We have
*
__
"
f(x)

(1

aiy)(l

'

(1

iV+.-.)( l
2

W eW
,

o 2 y) ...

4-

...

-I-

179

Examples.
Also

v> ^1

^/'M

whence

/(a:)

from which the

?L- m
/(*)

result follows

-'

y ^!L.
^f'(a)

a" 4^ 1

J
-.

^ /'(a)

^9

-ay

by comparing

coefficients of

y.

Express the sums of the homogeneous products of the roots in terms of the
coefficients of an equation, and vice versd.
2.

Since

we have

immediately, from the preceding example,


(1

+ Piy 4 Jfejf 2

...

4-

4 JW H )(1 4

Hi?/

U2y* + ...)!,

-4

whence
l>i -f

ni

=B

These equations
determine pi pi, .
.

By means
tric functions

pz 4 Uz 4 j^iOi

0,

which p\

(in

may be found

n r by

Representing by
ing,

we

the

du

n2

&c.

and

n,

7>)'

j2ni
are

0,

&c.

interchangeable)

vic^ versd.

aiy)(l

a" 4 1

/R

sums of the powers of the

a
= 2 --1

w =

also

We have,
(1 -f

&0>

roots.

a*y)

(1 -a,,y),

aSd

differentiat-

1 4-

Da,

*=

*2y

4-

...

*33r

ay

Oiy

-t-

H2 y2

therefore,

Uiy 4

n zy 2 4

.) (*i

Now, comparing the several


number of equations by means
,

jt?iH 2 4-

find

udy

n2

H2

an

- the product (1

III,

-f

in terms of the coefficients

a"

Express

+ Hs

jp a

and the preceding example the values of the following


symme-

of this

*7w'
5.

jt> 2 ,

0,

&c. and HI,

in terms of HI,

pn

4 *sy 2 4-

HI

+ 2n2y 4 3n3 y2 4

coefficients of the different

of

powers of

y,

..

we have

which the sums of the homogeneous products

... may be expressed in terms of

*i, *2,

&,

&c.

Prove the following formula for calculating the sums of the


homogeneoui
products in terms of the coefficients
4.

Differentiate both sides of


equation (1) in Art. 80,

the equation of Ex.

2.

and introduce Hi,

na

&c.,

by

CHAPTER

IX.

THE ROOTS OF EQUATIONS.

LIMITS OF

Definition of limits.

In attempting to discover the


numerical equations, it is in the first place advantageous to narrow the region within which they must be sought.
~We here take up the inquiry referred to in the observation at
84.

real roots of

the end of Art. 4, and proceed to prove certain propositions


relative to the limits of the real roots of equations.
superior limit of the positive roots is any greater positive
number than the greatest of these roots ; an inferior limit of the

positive roots is any smaller positive number than the smallest


of them.
superior limit of the negative roots is any greater

negative number than the greatest of them an inferior limit


of the negative roots is any smaller negative number than the
the greatest negative number meaning here the
smallest
;

number

nearest to - oo

When we

have found limits within which all the real roots


of an equation lie, the next step towards the solution of the
equation is to discover the intervals in which the separate roots

The

principal methods in use for this latter purpose will form the subject of the next chapter.
The following Propositions all relate to the superior limits
are situated.

of the positive roots

to which, as will be subsequently proved,


and limits of the negative

the determination of inferior limits


roots can be immediately reduced.

85.

Proposition

JL

In any equation

^x + pn
if the first negative term be

- p r %n~*', and

0,

if the greatest negative

181

Propositions.
coefficient be

jpk +

p*, then

a superior

1 ia

limit

of ike positive

roots.

Any

value of x which makes


Pk

a fortiori, make f(x) positive.


Now, taking a? greater than unity,

will,

this inequality is satisfied


*"""

by

the following

or
or

af

which inequality again

"J

(x

1)

is satisfied

by the following :

(a-lr'C*-!)-^*
since plainly

of"

We have, therefore,

>

(x

I)'"

finally
r

(a?- l)

= or>jf?

x - or >

or

jfe

4-

jpk*

Proposition II.

86.

ficient be taken positively,


coefficient*

If in any equation each negative coefand divided by the sum of all the positive

which precede it, the greatest quotient thus formed inis a superior limit of the positive root*.

creased by unity

Let the equation be

aX

a^' + fljtf"-1 - fl3#"~* + ....- a^^ +


1

in which, in order to fix our ideas,


ficient as negative,

general,

viz.

Let each

means

- ar

and we consider

. . . .

+ an

we regard the fourth

0,

coef-

also a negative coefficient in

positive

of the formula

- am

(x

term in

this equation

be transformed by

182
which

Limits of the Roots of Equations.


derived at once from

is

#"x

1
x

the negative terms remaining unchanged.


The polynomial f(x) becomes then, the horizontal lines of

the following corresponding to the successive terms off(x)

a?-

We

now regard

the vertical columns of this expression as


successive terms in the polynomial ; the successive coefficients
of

&-\

tf-\ &c., being


(x

Any

value of # greater than unity

is

1)

&o.

sufficient to

make

positive every term in which no negative coefficient #, ar , &o.


To make the latter terms positive, we must have
occurs.

(00

4-

0t

. . .

flr-i) (a?

1)

>

flr,

&0.

Hence
-.a
-f

00

And

to ensure every

term being made

positive,

1, dbo.

we must take

the value of the greatest of the quantities found in this way.


Such a value of x, therefore, is a superior limit of the positive
roots.

183

Practical Applications.

Practical Applications.

87.

The

propositions in the

two preceding Articles furnish the most convenient general


methods of finding in practice tolerably close limits of the
roots.
Sometimes one of the propositions will give the closer
limit; sometimes the other.

It is well, therefore, to apply

both methods, and take the smaller limit. Prop. I. will usually
be found the more advantageous when the first negative coefficient is

coefficients, so that r is

preceded by several positive

and Prop.

large ;
before the

first

when

II.

large positive coefficients occur

In general, Prop.

large negative coefficient.

We

will give the closer limit.

the numerical value givon

II.

speak of the integer next above

either proposition as the limit.

by

EXAMPLES.
1.

Find a superior

limit of the positive roots of the equation

x*
Prop.

I.

8+1,

gives

Prop. II. gives


2.

Find a superior

I.

40^2

3x- 4

Hence

or 6.

1,

6 is a superior limit.

a:

8# 2 - 5l#

and 6

f-

+ 34-1

1,

is,

and 12

this case Prop. I. gives the closer limit.

3.

Find a superior limit of the positive roots of


1

4x*

3.r

9x*

4.

+ 5# 4 -

0.

is

a limit,

0.

6*

4 fa - 8

lla: 2

fractions

1+4*
the third

18

therefore, a limit.

In

o.

or 9, as limit.

\/51 +

gives

Prop. II. gives

Of the

8r + 23

limit of the positive roots of the equation

#5 +
Prop.

5^3

is

11

6'

+ 4+5'

the greatest, and Prop. II. gives the limit 3.

Find a superior
#8

Prop.

I.

gives 5.

limit of the positive roots of

20*7

4^6

_ Ha^ _

120^*

13z

25

0.

An*. Both methods give the limit


6.

6.

Find a superior limit of the positive roots of


4s 5

8#*

22;e 3

98# 2

73# 4

Ans. Prop.

5=0.

I. trives

20.

ProD. II. gives

3.

184

Limits of the Roots of Equations.

It is usually possible to determine by inspection a limit


closer than that given by either of the preceding propositions.

This method consists in arranging the terms of an equation in


groups having a positive term first, and then observing what is
the lowest integer value of x which will have the effect of

The form

rendering each group positive.

of the equation will

suggest the arrangement in any particular case.


6.

The equation

of Ex. 2 can be arranged as follows


a-

a?

(#

8)

x (3* 3

+ a? +

61)

18

0.

any greater number, renders each group positive

3, or

hence 3

is

a superior

limit.

The equation

7.

x6
x

or

3,

of Ex. 4

(ar

may be

11)

-I-

arranged thus

20#* (x 5

6)

4# 6

any greater number, renders each group

+ 13# - 26 =
positive

0.

hence 3

if

a limit.

Find a superior limit of the roots of the equation

8.

* 4 - 4*3 +

33a:2

"2x

18

0.

This can be arranged in the form

xz (x* ~ 4x

Now

the trinomial x

4x

Hence x =

of x (Art. 12).

The

-f

5)

-f 5,

-f

28r (x

- &) +

having imaginary

18

0.

roots, is positive for all values

a superior limit.
way of a quadratic whose roots are imaginary, or of one

1 is

introduction in this

with equal roots, will often be found useful.

Find a superior limit of the roots of the equation

9.

bx*

7^4 - 10# 3 - 23#2 - 90s

In examples of this kind

among

7)

+ x9

(x*

317 =

- 10) -f x*

(*

methods give a very high

In

this case the general'

limit.

Find a superior limit of the roots of the equation

10.

#*

When

- x3 -

2x*

- 4x -

24

0.

there are several negative terms, and the coefficient of the highest term
convenient to multiply the whole equation by such a number as will

it is

enable us to distribute the highest term among the negative terms.


we can write the equation as follows :

plying by 4,

a? ( x

and 4

power of

- 23) + x (** - 90) + & - 317 - 0,

00 that 7 is evidently a superior limit of the roots.

unity,

0.

Here the equation may be written

the negative terms.


x* (x

it is convenient to distribute the highest

is

4)

a superior limit.

32 (s

8)

-f

(x

The general methods

16)

-f

give 26.

96

0,

Here, multi-

185

Propositions.
88.

Proposition III.

Any number

which renders positive

and all its derived functions fi(x\ /(#),


a superior limit of the positive roots of the equation f(x]

the polynomial f(x]


is

. .

/(#)
0.

This method of finding limits is due to Newton. It is much


more laborious in its application than either of the preceding
methods but it has the advantage of giving always very close
limits
and in the case of an equation all whose roots are real
;

the limit found in this

way is, as will be subsequently proved,


the next integer above the greatest positive root.
To prove the proposition, let the roots of the equation
be diminished by h ; then x - h = y, and
f(x) =
f(y +

A)

./(A) +/,(*) y +

If

now h be such

as to

make

<6

all

_L

ft

'

the coefficients

y cannot have a positive root that


x has no root greater than h hence h

positive, the equation in

to say, the equation in

is
is

a superior limit of the positive roots.


EXAMPLE.
/(*)

#*

2* s

3# 2

16*

3.

limits to any example the general mode


Take the smallest integer number which renders
and proceeding upwards in order to f\ (x), try the effect of substif*-i ( x) positive
When any function
tuting this number for x in the other functions of the series.
is reached which becomes negative for the
in
question, increase the integer
integer
successively by units, till it makes that function positive arid then proceed with

In applying Newton's method of finding

of procedure

IB

as follows:

the

new

integer as before, increasing

it

again

if

another function in the series

hould become negative ; and so on, till an integer is reached which renders all the
functions in the series positive.
In the present example the series of funt tions is

f(x)
/i

Ibx

4s - 6* 2 - 6* -

1$,

= x*

22?

2
3:r

(x)

*/(*)

6*2

*/s(*)

4* -2,

-6* -3.

3,

186

Limits of the Roots of Equations.

makes /a (x) positive. "We try then the effect of the substitution x = 1
makes ft (x) negative. Increase by 1 and x = 2 makes /2(#) positive.
Try the effect of x = 2 in J\ (x) it gives a negative result. Increase by 1 and
K = 3 makes f\ (x} positive. Proceeding upwards, the substitution x = 3 makes

Here x

It

in/a().

f (x) negative

and increasing again by unity, we find that x

Hence

4 makes /(a?) posi-

the superior limit required.


It is assumed in this mode of applying Newton's rule, that when any
makes all the derived functions up to a certain stage positive, any higher
tive.

is

number
number

make them positive so that there is no occasion to try the effect of the
on the functions lower down in the series. This is evident from the
number
higher
will also

equation
t>

(taking

<f>

(a+

(a)

$>

<p

(a)

for derived functions), whicli shows that if

and h
It

also positive,

may be

<j>

(a

-f

h}

must be

A2

+ </>"(

any function in the

to represent

(x)

h)

<f>

series,

and using the common notation

(a), ^>'(), <>"( a )

are

all positive,

positive.

observed that one advantage of Newton's method is that often, as in


it gives us a knowledge of the two successive integers between

the present instance,

which the highest root


x
3, and positive

f or

lies

between

and

Roots.

To

4,

in the present example, since f(z) is negative

we know

that the greatest root of the equation

find an inferior limit of the positive roots,


first

The

the

transformed by the substitution x =

-.

Find then a superior

4.

must be
equation
*

limit

Inferior Limits, and Limits of the Negative

89.

in y.

Thus

lies.

for

limit h of the positive roots of the equation

reciprocal of this, viz. -, will be the required inferior

for since

y <

.11.
h,

~ >
7,
h
y

i.e.

x>

1
-r.

To find limits of the negative roots, we have only to transform the equation by the substitution x - - y. This transforLet the
mation changes the negative into positive roots.
Superior and inferior limits of the positive roots of the equation
Then - h and - A' are the limits of he negain y be h and A'.
tive roots of the proposed equation.

187

Limiting Equations.
90.

Limiting liquations.
f

equation f(x)
the

could be found,

number of real

If
it

all

real roots of the

the

would be possible

roots of the equation f(x]

to

determine

0.

To prove this, let the real roots of /'(a?)


be, in ascending
order of magnitude, a', /3', 7', ... X' ; and let the following
series of values be substituted for x in/(#) :
-oo,

When any

a', /3', </,

successive

with different signs there

and by the

A', 4

oo.

two of these quantities give results


is a root of /(#) =
between them
;

Cor., Art. 71, there is only one

give results with the

same

root between them.

Thus each change

sign, there

is,

by

and when they

the same Cor., no

of sign in the results of

the successive substitutions proves the existence of one real root


of the proposed equation.
If all the roots of f(x)
are real, it is evident, by the
theorem of Art. 71, that all the roots of f(x) =
are also real,
and that they lie one by one between each adjacent pair of the

and by the same theorem,


=
it follows that the roots of /"(#)
0, and of all the successive
derived functions, are real also and the roots of any function
roots of f(x)

0.

In the same

case,

lie

severally between each adjacent pair of the roots

function from which

Equations of
degree of

it is

this

of the

immediately derived.

kind, which are one degree below the

any proposed equation, and whose,

roots lie severally

between each adjacent pair of the roots of the proposed, are


called limiting equations.

It

is

evident that in the application of Newton's method


when the roots of /(#) = are

of finding limits of the roots,

proceeding according to the method explained in


Art. 88, the function f(x) is itself the last which will be renall real, in

dered positive, and therefore the superior limit arrived at


integer next above the greatest root.

is

the

188

Examples.

EXAMPLES.
Prove that any derived equation ft*

1.

(x)

cannot have more imaginary roots,

may have more real roots, than the equation f(x) = from which it is derived.
From this it follows that, if any of the derived functions be found to have
imaginary roots, the same number at least of imaginary roots must enter the
but

original equation.
2.

Apply the method of Art. 90


3

jf

should have
3.

qx + r

When
q

>

When
q

- nqx +

<

or

=*

two

0.

real roots or none, according as

r"- 1 .

or
(x

<

r"- 1 .

1)

lias all its roots real

hence show, by forming

derived function, that the following equation has all

unequal, and situated between

and
n (n

1)

*-"&^' + -T^
Show

similarly

and

2^1) *"*""
n (n

1)

by forming the n th derived of (#* - 1)" that the following


and unequal, and situated between - 1 and 1

fi(n-l)L ~n-t.
X* - n
.

root* real

its

lias all its roots real

equation

__

_-

(*-!)' __ __J(w- l)(ff-2)(ft-3)


'_
L>
LJ
1.2 2(2n- 1) (2*- 2) (2n - 3)
v

.1.

2(2-l)
6.

l)

of the roots of the equation

odd, the equation has three real root* or one, according at

is

>

(a

even, the equation has

is

The equation xn
*

6.

=B

Determine by the same method the nature

An&.

4.

all its roots real.

&>

the

to determine the conditions that the equation

-*

&C.

0.

If any two of the quantities I, w, i in the following equation be put equal


sho* that the quadiatic to which the equation than reduces is a limiting

to zero,

aquation

and hence prove that the

(x
7.

-a) (x~l)

roots of the proposed are all real

(*-)- P {* - a) - iw2 (*- *)

'

(*- e) - Wmn

Discuss the nature of the roots of the equation


**

to?

2**

12* 4 p

0,

according to the different values of j.

Apply Art. 90. "When p


when p lies between - 7 and

is less

than -

7,

two roots

are real

and two imaginary ;

and when

p is greater than 9,
the roots are all imaginary. The equation has two equal roots when, p = - 7, and
two pairs of equal roots when p - 9.
9, all

the roots are real

CHAPTER

X.

SEPARATION OF THE ROOTS OF EQUATIONS.

lie.

equation
to

methods of the preceding chapter we are enabled


between which all the real roots of any numerical
Before proceeding to the actual approximation

By the

91.

to find limits

any particular

which

it is

roots.

root, it is necessary to separate the interval in

from the intervals which contain the remainpresent chapter will be occupied with certain

situated

The

ing
theorems whose object is to determine the number of real roots
between any two arbitrarily assumed values of the variable. It
plain that if this object can be effected, it will then be possible
to tell not only the total number of real roots, but also the limits
within which the roots separately lie.
is

The theorems given

for this purpose

by Fourier and Budan,

although different in statement, are identical in principle. For


purposes of exposition Fourier's statement is the more convenient, while with a view to practical* application the statement
Budan will be found superior. The theorem of Sturm, although

of

more laborious
that

it is

number
tities

in practice, has the advantage over the preceding


unfailing in its application, giving always the exact

of real roots situated

between any two proposed quan-

whereas the theorem of Fourier and

certain limit which the

number

Budan

gives only a

of real roots in the proposed

interval cannot exceed.


92.

a and

Theorem of Fourier and Budan.

b,

f(x) and

of which a

is

it$ successive

the

less,

Let two numbers

be substituted in the series formed by

derived functions,

viz.,

'(*),/.()./(*),.../.(*);

Separation of the Hoots of Equations.

190

the number of real roots which lie between a and b cannot le greater
than the excess of the number of changes of sign in the series when
a is substituted for x, over the number of changes when b is substituted for

x ; and when

This

number of

the

falls short of that difference,

it

will be by

real roots in the interval

an even number.

the form in which Fourier states the theorem.

is

It is to be understood here, as elsewhere, that,

speak of two numbers a and 6, of which a is the


both of them may be negative, and what is meant
nearer than b to - oo

when we

less,
is

one or

that a

is

We proceed to examine the changes which may occur among


the signs of the functions in the above series, the value of x
being supposed to increase continuously from a to b. The

following different cases can arise


(1). The value of x may pass through a single root of the
:

=
equation f(x)

0.

(3).

may pass through a root occurring r times in /(a?) = 0.


It may pass through a root of one of the auxiliary

(4).

functions/m (#) = 0, this root not occurring in either


m+1 (:r)=0.
/m
It may pass through a root occurring r times infm (x) = 0,

(2). It

_^=0or/

and not occurring in fm.


In what follows the symbol x

(x]

is

0.

omitted after

for con-

venience.

In the

(1).

first

case

it

is

evident,

from Art. 75, that in

= 0, one
passing through a root of the equation f(x)
change
for / and /\ have unlike signs immediately
of sign is lost
;

before,

and

like signs

immediately

after, the

passage through

the root.
(2).

In the second case, in passing through an r-multiple


= 0, it is evident that r changes of
are lost ;

root of f(x]
for,

by Art.

sign

76, immediately before the passage the series of

functions
J\\ fi> f*9

yr-i>

fr

have signs alternately + and -, or - and


after the passage have all the same sign

+,

and

iriimediately

Theorem of Fourier and Budan.

19/

third case, the root of /m (#) = must give to/m_i


and/m4>1 either like signs or unlike signs. Suppose it to give like
signs ; then in passing through the root two changes of sign are

In the

(3).

for before the passage the sign of m is different from these


like signs, and after the passage it is the same (Art. 76).
Supfor
is
of
no
then
it
to
unlike
lost,
;
sign
change
pose
give
signs

lost,

before the passage the signs of m -i fm> fm+i must be either


+ + -, or - - +, and after the passage these become

+ +. On the whole, therefore, we conand


,
clude that no variation of sign can be gained, but two variations may be lost, on the passage through a root of fm {} = 0.

(4).

In the fourth case x passes through a value

(let

us say

a)

to vanish.
which causes not only /m but also fm +i fm+*i
., /m+r-i
It is evident from the theorem of Art. 76 that during the passage
a number of changes of sign will always be lost. The definite
9

number may be

collected

The following
(a).

by considering the

series of functions

results are easily established

When fm-t (a) and/wm-(a)

have like signs:

If r be even, r changes are lost.


+ 1 changes are lost.

If r be odd, r
(b).

When/m_i(a) and/m+r (a)

have unlike signs:

If r be even, r changes are lost.


If r be odd, r - 1 changes are lost.

We

conclude, therefore, on the whole, that an even number


is lost during the
passage through an r-multiple root

of changes

of/m (4
It will be observed that (1) is a particular case of (2), and
= 1. Since, however, the cases (1) and (3)
of
(3)
(4), i.e. when r
are those of ordinary occurrence, it is well to give them a separate classification,

Reviewing the above proof, we conclude that as x increases


from a to b no change of sign can be gained; that for each

192

Separation of the Roots of Equations.

one change is lost; and


passage through a single root of f(x)
a
that under no circumstances except
passage through a root of
=
of
number
be lost. Hence the
an
odd
can
changes
/(#)
number of changes lost during the whole variation of x from
a to b must be either equal to the number of real roots of
in the interval, or must exceed it by an even number.
f(x) =

The theorem is therefore proved,


93. Application of the Theorem.
the theorem has been stated by Budan is,

The form

in

which

as has been already


observed, more convenient for practical purposes than that just
Let the roots of an equation f(x) =
It is as-follows
given.
be diminished, first by a and then by b, where a and b are an// two
:

numbers of which a is
between a and b cannot

number of

real roots

be greater than the excess of the

number of
number in

the less

then

the

changes of sign in the first transformed equation over the


the second.

This

is

evidently included in Fourier's statement, for the


(see Art. 33)

two transformed equations are


-

f(b) +/,(*) y +'

y*

from which, assuming the results of the


proposition is manifest.
The reason why the
in practice

is,

that

theorem in

we can apply

last Article, the

this

form

is

the expeditious

above

convenient

method

of

diminishing the roots given in Art. 33.


EXAMPLES.
1.

Find the

situations of the roots of the equation

x6

- 3^ -

24# 3 4 95# 2

- 46s -

101

0.

We shall examine this function for values of x between the intervals


-10,

-1,

0,

I,

10;
t

these numbers being assumed

on account of the facility

of calculation.

Diminution

193

Examples.
of the roots

equation

by

gives the following series of coefficients of the transformed

1,

-26,

2,

In diminishing the roots hy

-78.

65,

16,

10, it is apparent at the

lation that the signs of the coefficients of

very outset of the calcuequation will he all

the transformed

so that there is no occasion to complete the calculation in this case.


In diminishing the roots by - 10 and - 1, it is convenient to change the alternate signs of the equation, and diminish the roots by 4- 10 and -f 1
and then in

positive

the result change the alternate signs again.


equation when the roots are diminished by -

-8,

1,

In diminishing by
that the signs will be

-2,

- 10 we

coefficients of the transformed

are
60.

-291,

139,

observe in the course of the operation, as before,


in the result, i. e. when the alternate signs ar

all positive

will be
alternately positive

changed they

The

Hence we have the following scheme

and negative.
:

(-1)

+- +
+ --

(0)

-f

(1)

_,

(10)

+.

(-10)

+
-

_,
+,
the eqtutiioa

-,

-f-

itself,

These signs are the signs taken by/(#) and the several derived f unctions /i,/2,
on the substitution of the proposed numbers but it is to be observed that

/>> fi> ft

they are here written, not in the order of Art. 92, but in the reverse order, viz.

From
between
of sign

BO

we draw

these

10 and

is lost

between

between

is at least

the other two roots


roots

10

the following conclusions:

one real root

one real root

real root lies

are lost, there

-f

and

lies

and

lies

between ;

and

and between

one real root

All the real roots

10 and

between

0, since

and

but we are

1,

must

lie

since one change

one change of sign

is lost

changes of sign
in doubt as to the nature of

10, since three

left

whether they are imaginary, or whether there are three

real

10.

We might proceed to examine, by further transformations, the interval between


and 10 more closely, in order to determine the nature of the two doubtful roots
;

evident that the calculations for this purpose might, if the roots were nearly
This is the weak side of the theorem of Fourier and
equal, become very laborious.
Budan. Both writers have attempted to supply this defect, and have given methods

but

it is

of determining the nature of the roots in doubtful intervals ; but as these methods
we do not stop to explain them the more especially as the theorem

are complicated,

of Sturm effects fully the purposes for which the supplementary methods of Fourier
and Budan were invented.

194
2.

Separation of the Roots of Equation*.


Analyse the equation of Ex.

The

and

roots of this are all real,

lie

100, viz.,

1, p.

* + **

2*

between

0.

2 and 2 (see Ex. 5, p.

ICfu).

jkVhen-

ever the roots of an equation are all real, the signs of Fourier's function^&ffermine
the exact nuraher of real roots hetween any two proposed integers. We" obtjin thd

following result

The

roots lie in the intervals

(-2,-!);
3.

Analyse the equation of Ex.

Ans.

Two

(-1,0);

3, p.

(1,2).

100, viz.,

roots in the interval (- 2,

of the intervals (- 1, 0)
4.

1),

(0, 1)

and one root in each


;

(1, 2).

Analyse the equation

+ 1998** - 14937s + 5000 =

0.

The equation can have no negative roots. Diminish the roots hy 10 several times
obtain the
in succession till the signs of the coefficients become all positive.

We

following result

+ - + -+,
+ _ + + _,
- + +,
+
+ + + - +,
+ + + 4- +.

(0)

(10)
(20)

(30)

(40)

Thus, there

is

one root between

hetween 20 and 30.

and

10,

Between SO and 40

and one between 10 and 20; no root

two real roots, or there is


an indication of a pair of imaginary roots. That the former is the case will appear
by diminishing the roots of the thud transfoimed equation by units. This process
will separate the roots, w hich will be found to lie between (2, 3) and (4, 5) so that
cither there are

the proposed equation has a third real root in the interval (32, 33), and a fourth in

%e

interval (34, 35).

94.

Application

of the

Theorem

to

Imaginary

Roots.

Since there exist only n changes of sign to be lost in


oo to + oo , if we have any reason for
the passage of x from
that a pair of changes is lost during the passage of x
through an interval which includes no real root of the equation,
we may be assured of the existence of a pair of imaginary roots.

knowing

Circumstances of this nature will arise in the application of


Fourier's theorem when any of the transformed equations con-

For we can assign by the principle


of Art. 76 the proper sign to this coefficient, corresponding to
tain vanishing coefficients.

195

Example*.

values of x immediately before and immediately after that value


which causes the coefficient to vanish the whole interval being
;

so small that

equation /(a?)

be supposed not to include any root of the

may

it

0.

EXAMPLES.
1.

Analyse the equation

f(x)

We

shall

examine

z*

- 4#3 -

this function

3#

23

0.

between the intervals

0, 1, 10.

The

trans-

formed equations are

A/4 (0)

the

first

*'

+ */

(0) **

+ i/i (0) * + /i (0) x + / (0) =

0,

(10)z+/(10) =

0,

of these being the proposed equation itself.

Making the

by the method of the preceding Article, we find that


and we have the following scheme :

calculations

the coefficient /i (1)

= 0,

(0)

- +,

(1)

- - +,

+ + + +

(10)

+.

We may now
first

replace each of the rows containing a zero coefficient by two, the


corresponding to a value a little less, and the second to a value a little greater,

than that which gives the zero coefficients, the signs being determined by the
It must be remembered that in the above scheme
principle established in Art. 76.
the signs representing the derived functions are written in the reverse order to that
The scheme will then stand as follows, h being used to
of the Article referred to.
represent a very small positive quantity:

{-h
(0)
(

(1)

(10)

hi

\\~\
(l + A

+ - + -

+,

_ _ _ +

+ +

'

__+,
+ +

+,

h are determined by the condition that


where the signs corresponding to - h and
- h, be difthe sign of the coefficient which is zero when x =
must, when x
ferent from that next to it on the left-band side
and when x = 4- h these signs
h and 1 4- h are determined in
must be the*same. The signs corresponding to 1
-I-

similar

manner

O 9

196

Separation of the Roots of Equations.

Now since a pair of changes is lost in the interval (- A, + A), and since the
h and -I- A, we have proved the existence of s
equation has no real root between
Two changes of sign are lost between 1 + h and 10, so
pair of imaginary roots.
that this interval either includes a pair of real roots, or presents an indication of a
Which

pair of imaginary roots.

of these

still

doubtful.

be able to establish the existence of

This will appear from the following example

several pairs of imaginary roots.

*6

The

the case remains

is

we may

2. If several coefficients vanish,

h and

signs corresponding to

0.

-f

are,

-f

by the theorem

--

(_A)

(+ h)

+ + + + + + -.

-f

of Art. 76,

h and 4- A, and since 4 changes of sign


Hence, since no root exists between
to one very little greater, we
are lost in passing from a value very little less than
The other two roots
are assured of the existence of two pairs of imaginary roots.
are in this case plainly real (see Art. 14).

The number

of imaginary roots in

any binomial equation can be dete/fmined

in

this

way.
3. Find the character of the roots of the equation
x?

10.r 3 -f

-f

0.

In passing from a small negative to a small positive value of


following series of signs

x we

obtain the

(-A)

+ - +

-f

-f

-,

'

(0)

+ -,

The
Since six changes of sign are here lost, there are six imaginary roots.
remaining two roots are, by Art. 14, real One positive, and the other negative.
and 1.
The negative root lies between 2 and
1, and the positive between
:

4.

Analyse completely the equation

#8-3*2-*'+
There are two imaginary

In

this

way we

Whenever,

roots.

are comprised within small limits,

0.

as in the present instance, the roots

convenient to diminish by successive units.


find here a root between
and 1, and another between 1 and 2.

Proceeding to negative roots,

we

it is

find on diminishing

by -

that

1 is itself

a root,

and writing down the signs corresponding to a value a little greater than
observe an indication of a second negative root between - 1 and 0.
6.

1,

wo

Analyse the equation


x*

at

+ ** -

25*

36

0.
c

There are two imaginary roots

two

one

reai!

real negative roots in the intervals (- 3,

positive root

2), (- 2,

between 2 and 3
1).

and

Corollaries
95.

from

the

Corollaries

Budau.

Theorem of Fourier and Sudan.

197

from the Theorem of Fourier and

The method

of detecting the existence of

imaginary

roots explained in the preceding Article is called The Rule of


the Double Sign.
similar rule, due to De Gua> was in

This rule and


use before the discovery of Fourier's theorem.
Descartes' Rule of Signs are immediate corollaries from the
theorem, as
Cor. 1.

The
eessire

roots

rule

we proceed to show.
De Gua's Rule for finding Imaginary

may he stated

generally as follows

Roots.

When 2m

suo-

term* of an equation are absent, the equation has 2m imaginary


and when 2m + 1 successive terms are absent, the equation

2m +

2, or 2m imaginary roots, according as the two terms between which the deficiency occurs have like or unlike xigns.
This
follows as in case (4j, Art. 92, by examining the number of

has

changes of sign lost during the passage of # from a small negative value - h to a small
value h.
positive

Cor. 2.

Descarte* Rule of Signs.

When

is

substituted

for

x in

the

series

of

functions

fn (),fn-i (#),
/(#)> /i (#)/(#), the signs are the same as the
rt w-1
of
the
a n of the proposed
coefficients
a , a\, a t
signs
-

equation

arid

when +

number
that

is

This

oo is substituted

Fourier's theorem asserts that the


these limits, viz. the

number

the signs are

number

of positive roots, cannot exc<wl the

of variations lost during the passage

the

number

all positive.

of roots between

from

of changes of sign in the series a Q9

Descartes' rule for positive roots


for negative roots follows in the usual
is

a,,

to

a3

oc,

an

and the similar rule

way by changing

the

negative into positive roots.


Cor. 3.

When

Newton's Method of finding Limits.


a number A has been found which renders positive

the functions /(#), /-i(a),


/*(*)> fi(*) 9 /(#) ; since
also renders each of them positive, it follows from Fourier's

eacli of

oc

theorem that there can be no root between h and + oo , that is to


say, A is* a superior limit of the positive roots ; and this i

Newton's proposition (Art. 88).

198

Separation of the Roots of Equations.

96.

that

We have already shown (Art. 74)


by performing the common algebraical opera-

Sturm's Theorem.

it is

possible

tion of finding the greatest

f(x] and

common measure

of a polynomial

derived polynomial to find the equal roots ol


the equation f(x)
0.
Sturm has employed the same operation
its first

for the formation of the auxiliary functions which enter into


his method of
separating the roots of an equation.

Let the process of finding the greatest common measure oi


f(x) and its first derived be performed. The successive remainders will go on diminishing in degree till we reach finally
either one which divides that immediately
preceding without
remainder, or one which does not contain the variable at all,
i. e. which is numerical.
The former is, as we have already
seen, the case of equal roots.

equal roots exist.

It

is

The

latter is the case

where no

convenient to divide the discussion of

Sturm's theorem into these two

We

cases.

shall in the present

Article consider the case where no


equal roots exist
ceed in the next Article to the case of equal roots.

and proThe per-

formance of the operation itself will of course disclose the


which any particular example is to be referred.

class

to

The

by Sturm are not the


remainders as they present themselves in the calculation, but
the remainders with their signs changed.
In finding the greatest
auxiliary functions employed

common measure

of two expressions it is indifferent whether the


of
the
remainders
are changed or not : in the formation
signs
of Sturm's auxiliary functions the
is essential.

We
change
what follows that the sign of each
remainder is changed before it becomes the next divisor.
Confining our attention for the present to the case where no
shall suppose therefore in

equal roots exist, Sturm's theorem

Theorem.
for x

in the series

may

be stated as follows

Let any two real quantities a and b

ofn +

be substituted

1 functions

/(*). /(*)> /(*), /(*),

/-i(*), /.(*),

consisting of the given polynomial f(x), its first derived


/(#),

and

199

Sturm's Theorem.

the successive remainders (with their signs changed) in the proces*

offinding the greatest common measure of f(x) andfi(x)

then the

number of changes of sign in the series when


and the number when b is substituted for x
number of real roots of the equation f(x) =

difference between the

is

substituted for x

expresses exactly the


beticeen

a and

b.

The mode

of formation of Sturm's functions supplies the

following series of equations, in

which

<?,,

the successive quotients in the operation

ja,

qn-\

represent

(1)

These equations involve the theory of the method of finding


the greatest common measure ; for it follows from the first equation that if f(x] and fjx) have a common factor, this must be

and from the second equation it follows, by


like reasoning, that the same factor must occur in/3 (#)
and so
on, till we come finally to the last remainder, which, when/(#)
and fi (x) have common factors, will be a polynomial consisting
In the present Article, where we suppose the
of these factors.
given polynomial and its first derived to have no common
a factor in f*(x)

factor, the last

remainder /(#)

is

numerical.

It

is essential

for

the proof of the theorem to observe also, that in the case now
under consideration no two consecutive functions in the series

can have a

common factor

for if they

had we

could,

by reason-

ing similar to the above, show from the equations that this factor must exist also in f(x) and/t (o?) ; and such, according to our

In examining, therefore, what


hypothesis, is not here the case.
the series during the passage
can
in
of
take
place
sign
changes
of x fronf a to ft, we may exclude the case of two consecutive
functions vanishing for the same value of the variable

and the

200

Separation of the 'Roots of Equations.

different cases in

the following

when x

(1).

which any change of sign can take place are

/()
(2).

when x

(3).

when x

passes through a root of the proposed equation

o,

passes through a value which causes one of the


.
.fn-.i to vanish,
auxiliary functions /i,/2
.

passes through a value which causes two or more


of the series/, /,, . /n_i to vanish together, no two of
.

the vanishing functions, however, being consecutive.

When x passes through a root

(1).

of /(a?) = 0,

it

follows from

Art. 75 that one change of sign is lost, since immediately before


the passage f(x) and /(#) have unlike signs, and immediately

they have like signs.


x
Suppose to take a value a which

after the passage


(2).

fr (x)

0.

From

/r-i'tf)

we have

satisfies

the equation

the equation

/r-i(a)

= qrfr(%)
= -/r +i('

~/r+i(#),
,

which proves that this value of x gives to fr-\{x\ and /r+i(#) the
same numerical value with different signs. In passing from a
value a little less than a to one a little greater, we can suppose
the interval so small that it contains no root of /r-i(#) or/ (a*) ;
hence, throughout the interval under consideration, these

two

If the sign of fr (x) does not change


functions retain their signs.
will
in
the
(as
happen
exceptional case when the root a is re-

peated an even number of times) there


series of signs.

In general the sign

of

no alteration in the

is

r (x)

changes, but no

variation of sign is either lost or gained thereby in the group of


three ; because, on account of the difference of signs of the two

extremes /r-i(#) and

/r+

(a?),

there will exist both before

and

after

the passage one variation and one permanency of sign, whatever


be the sign of the middle function. If, for example, before the
after the passage they become
;
passage the signs were +
a variation and a permanency are changed into a
i. e.
permanency and a variation but no variation of sigrf is lost or

+ + -,

gained on the whole.

201

Sturm's Theorem.

Since the reasoning in the preceding oases is founded


on the relations of the function to those adjacent to it only
and since these relations remain unaltered in the present case,
(3).

because no two adjacent functions vanish together, we conclude


that if f(x) is one of the vanishing functions, one change of sign
is lost,

We

and if not,
hav

is

either lost or gained.

fore, that

when x

passes through a
no other

of sign is lost, and under


of sign either lost or gained.

root

Hence

sign lost during the variation of x


the number of roots of the equation

'.of

from

flTto b

\&j*fc&/m

Betwee
Befclfe pogy^Siiig to\he case of equal roots, we add a few
theorem.
sijnjple eSiwwfiTes to illustrate the application of Sturm's

It

is

convenient in practice to substitute

first

oo

0,

oo in

Sturm's functions, so as to obtain the whole number of negative

and of positive roots. To separate the negative roots, the inte- 1, - 2, - 3, &o., are to be substituted in succession till we

gers
reach the
of -

oo

&c.,

till

same series of signs as results from the substitution


and to separate the positive roots we substitute 1, 2, 3,

the signs furnished by + oo are reached.

EXAMPLES,
1.

Find the number and

situation of the real roots of the equation

f(x)

We find

fi(x)

3r>

Corresponding to the values

(-00)
(0)

Hence there

is

a;

- 2* -

fz (r) =

2,
oo

4* +

-f oo

of x y

- +
- -

0,

16,

/,(*)

=-

643.

we have
-,

+ -,
+ + + -

(+ 00 )
only one real root, and

it is

positive.

The

student often finds a difficulty in perceiving in what way a record is preferred in Sturm's series of the number of changes of sign lost, since the only loss takes
place between the

two functions, f(x) aud/i(#). It may tend to remove this


x increases from one root a of f(x) = to a second
although no alteration takes place in the number of changes of sign, the distribution
of the signs among f\(x] and the following functions alters in such a way that the
signs of /(*) and /i(tf), which were the same immediately after the passage of
first

difficulty to observe,

through

a,

that as

become again

different before the passage

through &.

202

Separation of the

Again, corresponding to values 1,

->-

(3)

+ + + -

real root, therefore, lies between* 2

2.

we hare

or,

(2)

(1)

The

3 of

2,

- + +
- + +

and

-,

3.

Find the number and situation of the


tc*

We easily obtain
/, (x)

/a

1x

3*

real roots of the equation

0.

7,

(*)** -3,

whence
(-00)

- + -

(0)

+
oo

(+

Hence

all

the roots are real

We have,

+,

+ + +

+.

one negative, and two positive*

-f

(-2)

- + + + + + -

(-1)

+ - -

+,

(1)

+,

(2)

+ +

(-3)

4 and

+,

further, the following results

(-4)

Here

+,
f,
-f,

+.

2 give the same series of signs as - oo and +


hence we stop at
;
root lies between
4 and - 3 ; and the two positive roots

The negative
between 1 and 2.
these.

This example illustrates the superiority of Sturm's method over that of Fourier.
The substitution of 1 and 2 in Fourier's functions gives, as can be immediately

yerified, the following series of signs :

From

Fourier's theorem

(1)

- +

(2)

+ +

+.

we

more than two roots between

two

roots

between

must, of course,
3.

and

make

2.

are authorised to conclude only that there tannot b*


and 2. From Sturm's we conclude that there art

If

we have

occasion to separate these

two

further substitutions iuf(x)

Find the number and situation of the


**

2s8

3#8

-f

real roots of the equation

lOx

-4 =

0.

roots,

we

Sturm? s'Theorem.

We

obtain,

removing the factor 2 from the derived,

= -8* -3,

/,(*)

1433.

In forming Sturm's functions it is allowable, as is evident from the


introduce or suppress numerical factors just as in the process of

[N.B.
equations

203

Equal Roots.

(1), to

finding the o. o. M. ; taking care, however, that these are positive, so that the signs
of the remainders are not thereby altered.]

We have the

following series of signs

- 4

(0)

(4

Hence
roots.

there are

To

two

GO

--

real roots,

--.

+ + +

one positive, and one negative, and two imaginary

find the positions of the real roots,

it is

sufficient to substitute positive

and negative integers successively in/(#) alone, since there

is

only one positive and

We

ne negative root.
easily find in this way that the negative root
2 and - 3, and the positive root between
and 1.

97.

Sturm's Theorem.

Equal Roots.

lies

between

Let the opera-

tion for finding the greatest common measure of /(#) and f(x]
be performed, the signs of the successive remainders being

changed as before. The last of Sturm's functions will not now


be numerical, for since f(x) and f(x) are here supposed to contain a common measure involving #, this will now be the last
function arrived at

by the

process.

Let the

series of functions

be:

/(#) /i(*)

/i(#)>

......

MX).

During the passage of x through any value except a multiple root


of /(a?) - 0, the conclusions of the last Article are still true with
respect to the present series, since no value except such a root
can cause any consecutive pair of the series to vanish. When x

passes through a multiple root of /(a?) - 0, there is,


Art. 75, one change of sign lost bet ween /and/! ;

ceed to prove that no change of sign


of the series,

viz.

/ /
t,

t,

tiple root a of /(*). It is

is lost

by the Oor.,
and we pro-

or gained in the rest

r
Suppose there exists an in-mulevident from the equations (i) of Art. 96,
.

204

Separation of the Roots of Equations.

that (# - a) w ^ is a factor in each of the f unctions f\if*>


fr*
Let the remaining factors in these functions be, respectively,
^i,

(y

....

^2,

*" 1

a)

By dividing each of the equations (1) by


a
series of equations which establish, by the
get

$r-

we

reasoning of the. last Article, that, owing to a passage through a,


no change of sign is lost or gained in the series $ 19 2
fa*
<f>

Neither, therefore,
/i>/2>

/;

is

lost or

any change

for tXe effect of the factor (x

of x from a value a - h to a value a + h

is

gained in the
- a)" in the

series

1 "" 1

passage

either to

change the

m - 1 is odd)

- 1 is even)
or of none (when
of the functions
2
1?
$ r and changing the signs of all
these functions cannot increase or diminish the number of
signs of all (when

variations.

We

have therefore proved that when x passes through a


multiple root of f(x) = one change of sign is lost between / and
/i, and none either lost or gained in any other part of the series.
It remains true, of course, that

when x

passes through a single


lost as before.
may thus

a change of sign is
We
theorem as follows for the case of equal roots
The difference between the number of changes of sign when a and b

root of /(a?)

state the

are substituted in the series


fy f\9
the

lad of

/r>

./

these being the greatest

common measure of f and f^

equal to the number of real roots between a and

6,

is

each multiple root

counting only once.

EXAMPLES.
1.

Find the nature of the roots of the equation


a?*

We easily obtain
/i

- 6^ +

9#2 - 7*

= 4s 3

(x)

15#

/ (*)-*- 2*+
/a (x) divides f\ (x) without remainder

-f

18$

0.

1;

hence in this cffiPSturm's

/*(*), thus establishing the existence of equal roots.

series stops at

205

Examples.
To

number

find the

The

4,

(4

+.

2.

real distinct roots

from the form of /2

which

(x),

substitute

oo

and 4

oo

result is

(-00)

Hence the equation has only two


root, as is evident

we

of real roots of the equation,

for x in the series of functions /, /i, /2.

but one of these

equal to (x

is

I)

is

a triple

Find the nature of the roots of the equation

* - 6*

- 12* +

13*2

0.

Here
/! (*)

is

the last Sturmian function

18*2 4 26* - 12,

-3*4

*2
/2 (*)

4* 8

(_co)
co

(+

There are only two


the roots
3.

1,

is

2;

so that the equation has equal root*.

-I-

4,

444.

real distinct roots.

In

fact, since /a (a;)

(x

1) (*

2),

each of

a double root.

find the nature of the

Here

roots of the equation

a*

2.r*

/,

4 *3 -

5* 4

8.r

x*

- 2* 2

4 3*

/ = 2a; 4 7^ 2 4 12z
/3 = - # - Ca; - 5.
3

2x

7,

0.

2,

Since

y*s

\Ve have

0,

a;

is

/4

=-*-

/6

1,

0.

common measure oif and/i,

and/(ar) hai a double root

1.

also

_-+,

(-00)

_ +

(4 oo

+ + +

__.

Hence there are two real distinct roots. The equation has, therefore, beside the.
double root, one other real root, and two imaginary roots.
4.

Find the nature of the roots of the equation


x*

Hwe

- 7#8 4

/i (x)

/a (#)
/s(s)

15**

- 40* 4 48* -

16

=* 0.

- 35* 4 4 60* 3 - 80* 448,

=r

6* 5

13**

5
ic

- 84a?3 4 192*- -

76* 4 48,

- 6* 4 12*- 8 n (*-2} ?
2

Ans. There are three real distinct

roots,

one of them being quadruple.

206

Separation of the Roots of Equations.

Application of Sturm's Theorem.

98.

In the case

oi

equations of high degrees the calculation of Sturm's auxiliary


It is important there.functions becomes often very laborious.

pay attention

fore to

what

to certain observations

which tend some-

to diminish this labour.

the final remainder when it is numerical,


(1). In calculating
since its sign is all we are concerned with, the labour of the last
operation of division can be avoided by the consideration that

fn,

must give opposite


general possible to tell without any

the value of x which causes

to vanish

It is in
signs to/n_2 and/*.
calculation what would be the sign of the result if the root of
Thus in Ex. 3, Art. 96, if
substituted in /-*(#).
A-i (a?) = were
o
the value - 5, which is the root of /(#) = 0, be substituted for

o
x in 9x* - 27% 4-11, the result
sign of

/n (x)

- 1433
given

is -,

for

and there

fn (x)

When it is possible

(2).

roots of

any one

is

in the

evidently positive; hence the


no occasion to calculate the value
is

example referred

in

any way

to.

to recognize that all the

of Sturm's functions are imaginary,

we need

not proceed to the calculation of any function beyond that one ;


for since such a function retains constantly the same sign for all
values of the variable (Cor. Art. 12), no alteration in the

number

of changes of sign presented by it and the following functions


can ever take place, so that the difference in the number of

changes when two quantities a and b are substituted is independent of whatever variations of sign may exist in that part of
the series which consists of the function in question and those
With a view to the application of this observation
following it.
it

is

(ax*

always well, when we arrive at the quadratic function


4- bx + c,
suppose), to examine, in case the term containing

and the absolute term have the same sign (otherwise the roots
could not be imaginary), whether the condition 4ac > b* is fulfilled ; if so, we know that the roots are imaginary, and tha
x*

need not proceed farther.


Similar observations apply to the case where tone
functions is a perfect square, since such a function cannot
calculation

change

its

sign for real values of x.

207

Examples.
EXAMPLES.
1

Analyse the equation

s*4 3*8 +

We find

/2 (*) = /8 (*) = -

29*'

+ 1*0,

78*

from -

makes

ft (x)

and two imaginary

roots.

real,

f-8. -1},

{-1, 0).

little

J,

14,

* given by the equation /$ (x) = 0,


hence /* (x) is negative.

see immediately that the value of

which differs
There are two

2.

10*

1086* -481,

N /*(*)-

Here we

7**

positive;

The

real roots lie in the intervals

Analyse the equation

- 4*

**

We

3*

23

0.

find

/ (*)= 12** + 9* -89,


/3 (*) =-491*+ 1371,
2

/<(*)=
Here

/,(*)

*=

gives

>

>

2-74

-j^-^j/* (*) positive hence the root of /s (*) makes it positive


There are two real and two imaginary roots.
;

The
3.

real roots lie in the intervals

{2, 3},

>

-,

and

=
|

makes

also.

(3, 4}.

Analyse the equation


2* 4

13*2 +

10*-

19

0.

Hero

/!W =
fi(x)

Sinoe

>

4 x 13 x 38

15 2 ,

4* -13*-|-6,

IS*2

15*

38.

the roots of fz(x) are imaginary, and

further with the calculation of Sturm's remainders.

Substituting

oo

we

+00,

0,

(-00)
(0)

+
-

+,

+,

+.

(+00)
There are two
4.

Here

real roots,

obtain

one positive, the other negative*

Analyse the equation

5**

+ *s - 4*2 - 3* -6 - 0.
+ 8*8 + 3*2 - Sx

/a(*)

6**

/8 (*)

/(*)

m ae6 +

/i (*)

2ar*

66**

116*2

+ 44* +
57*

110,

223.

we

proceed no

208

Separation of the Roots of Equations.

Sinee

>

4 x 116 x 223

substituting

oo

we may

(-OD)
(0)

(+

00

There are four imaginary


6.

stop the calculation here.

We find,

4.

and one

roots,

Find the number and situation of the


x*

- 2# 3 - 7*2

-f

Ans. The roots are

-,.

-,

real positive root.

real roots of the equation

10*+

10

= 0.

and are situated in the intervals

all real,

{-3, -2}, {-1, 0}, and two between


6.

{2, 3}.

Analyse the equation

xs
It will be

3*4

4.

2*s

3a .a

2*

0.

found that the calculation may cease with the quadratic remainder.

Ans. There
7.

or

+00,

0,

57 2 ,

is

only one real root

in the interval

{ 1

2}

Analyse the equation

=0.

181

We find

f*(x)

= 854* -

2751,

In some examples, of which the present is an instance, it is not easy to tell


immediately what sign the root of the penultimate function gives to the preceding
We have here calculated /af*), and it turns out to be a much smaller
function.

numoer than might have been expected from the magnitude of the coefficients in/s(#).
In fact when the root of ft (x) is substituted in/i (a?) the positive part is nearly equal
This is always an indication that two roots of th* proposed
to the negative part.
There are injthe present instance two positive roots beequation are nearly equal.
tween 3 and 4. Subdividing the internals, we find the two roots still to lie between
3*2 and 3'3

"We have here another

so that they are very close together.

tion of the continuity

which

illustra-

between real and imaginary roots (cf. Arts. 17,


the two roots would be equal; and if it were a small
exists

If fz(x) were zero,


18).
negative number, they would be imaginary.
8.

Analyse the equation

The

quadratic function

is

found to have imaginary

Ans. One

real root

between

{0,

0.

roots.
1

four imaginary roots.

209

Examples.
9.

Analyse the equation

& - 6*

We find

- 30* + 12* -

/(*) and as

this

has clearly

all

20**

0.

4-

roots, the calculation

imaginary

An s. Two

6**

real roots

may

in the intervals

stop here.
1

2,

{6,7}.

0),

{5, 6).

Analyse the equation

10.

2*

18**

60**

-I-

120*'

We find

- 30* 2 +

2
/2(*)-6**-f 220* +

and the calculation

may

--

0.

stop.

Two

Ana.

Examine how the

11.

18*

real roots

in the intervals

{-

1,

roots of the equation

2**+ 16* 8 -84*- 190 =


are situated in the several intervals between the

Here

The

/i(*)

x1
-

-f

bx -

numbers -

oo

7, 6,

-f

oo

14,

substitution of the above quantities gives

(_oo)

+,

(-7)

(6)

+,

-f

+.

(-f 00

Whenever,

as in this example,

-f,

any quantity makes one


the zero

of the auxiliary functions

be disregarded in counting the


number of changes of sign in the corresponding row for, since the signs on each
ranish (here

7 satisfies f\ (x)

0),

may

no

alteration in the

number

of changes of sign in the row


could take place, whatever sign be supposed attached to the vanishing quantity.

eide of it are different,

The

roots are all real.

tween 12.

and

There

is

one root between

oo

and

and two be-

6.

Analyse the equation


3**

- 6*3 -

8*

0.

We find

As/f#)

is

a perfect square the calculation

Am. Two real

may

cease.

roots; in the intervals

{-1, 0},

{1, 2).

210

Separation of the Roots of Equations.

Conditions for the Reality of tbe Roots of an


Equation. The number of Sturm's functions, including
- 1 remainders, will in
f(x) f(x) and tbe n
general be n + 1.
99.

In certain

cases,

function, some

owing

to tbe absence of terms in tbe proposed

be wanting. Tins can


occur only wben the proposed equation has imaginary roots for
it is clear that, in order to insure a loss of n changes of sign in
the series of functions during the passage of x from - oo to + oo
(namely, in order that the equation should have all its roots
of tbe remainders will

real), all

must

all

the functions must be present.


And, moreover, they
and alternating
take the same sign when x = + oo
;

when x - -

Since the leading term of an equation is


signs
a
positive sign, we may state the condition
always taken with
for the reality of all the roots of any equation (supposed not to
oo

have equal roots) as follows

In order that

all the roots

of an

th
degree should be real, the leading coefficients of
equation of the n

all

StUrm's remainders, in number n -

1,

must

be positive.

EXAMPLES.
1.

Find the condition that the roots of the equation


ax- + Ibx

=
Ans.

should be real and unequal.


2.

uc

>

0.

Find the conditions that the roots of the cuhic


^-t-Sflz-f

should be

When
which

bz

all real

<?

and unequal.

this cubic has its roots all real, it is evident that the general cubic

from

derived (Art. 36) has also its roots all real ; so that in investigating the
conditions for the reality of the roots of a cubic in general, it is sufficient to discui/
the form here written.
it is

We find

/iW =

22

4-

//,

[In calculating these, before dividing f\

(z)

by/2(z), multiply the former by the

2
positive factor 2.Zr .]

Hence the required conditions

are, J5~ negative

These can be expressed as one condition,


implies the former (cf. Art. 43).

and GP +

viz.,

G2

-f

4J5T 3 negative.

4IT 8 negative, since this

Conditions for Reality of Roots of a Biquadratic.


3.

211

Calculate Sturm's remainders for the biquadratic


**

We find

0.
GHz* 4- 4Gz + a 2 / - 35" 2
= - ZHz* -ZGz- (a2 / - 3JT 1 ),
= - (2J7J- 3a/)z - GI,

-I-

2 (z)

f (z)
/ (*)-/-27/
s

2
.

These are obtained without much

by

difficulty

aid of the identity of Art. 37.

Before dividing /i by/z, multiply by the positive factor 3.S" 2 and when the remainder is found, remove the positive factor a 2 . Before dividing^ by^g, multiply
;

by the positive factor


2
2
positive factor a j2" .

(2HI

3aJ)

2
;

and when the remainder

is

found, remove the

Conditions for the Reality of the Roots of a


Biquadratic. In order to arrive at criteria of the nature of
100.

the roots of the general algebraic equation of the fourth degree


by Sturm's method, it is sufficient to consider the equation of

Ex. 3 of the preceding Article. By aid of the forms of the


leading coefficients of Sturm's remainders there calculated, we

down

can write

the conditions that all the roots of a biquadratic

should be real and unequal in the form

negative,

2III - 3aJ

negative,

J 3 - 27 J

positive.

It will be observed that the second of these conditions

is

form from the corresponding condition of Art. 68.


To show the equivalence of the two forms it is necessary to
different in

when
is negative and A positive, the further con2HI 3aJ negative implies the condition 3Jr -12J9r2

prove that
dition

From the identity of Art. 37,


negative, and conversely.
2
- 12H 2 s a 2
written in the form (a-I
(2H1 - 3aJ) - 3<? ,

it

readily appears that

a 2! - 12 H~

we

H and

when

necessarily negative.
observe that when aj is positive

I is
when aj
since

positive
is

2HI - 3aJ are

And

is

2HI -

on account of the condition

2HI

3aJ

is still

negative

to prove the converse

3a J

is

negative,

positive ; and
negative, since the

negative
2HI exceeds the positive part - 3aJ, as may be
2
2
shown
readily
by the aid of the inequalities 12H > a / and
8
>
/
27J*.
negative part

The student will have no difficulty in verifying, by means


of Sturm's functions, the remaining conclusions arrived at in
the different cases of Art. 68.
p

212

Separation of the Roots of Equations.

EXAMPLES.
1.

Apply Budan's method

to separate the roots of the equation

**

16*3

69* a

70*

42

0.

Ant. Boots in intervals {-1, 0}, {2, 3}, {4, 6), {9, 10}.
2.

Apply Sturm's theorem to the analysis of the equation

6*

0.

In analysing a biquadratic of this nature which has clearly two real roots, when
a Sturmian remainder is reached whose leading coefficient is negative, the calculation

may

may

must then be imaginary, and the positions


be readily found by substitution in the given equation.

Two

roots imaginary

cease, since the other pair of roots

of the real roots

Ans.
3.

real roots in intervals

Two

Ans.

roots imaginary

Apply Sturm's theorem


**

21

2,

real roots in intervals

{-1, 0}, {3,4}.

to the analysis of the equation

+ 3* 3

-&-

3*

11

0.
all

imaginary.

Find, by Sturm's method, the number and positions of the real roots of the

equation

x*

Ant. Boots

all real

{1,
6.

0.

Ans. Boots
5.

0},

Analyse in a similar manner the equation

- 6* -

4.

{1,

0}

10*3

one

63

o.

m the interval {4,

3}; two in the interval

and positive roots in the intervals

{0,

},

3,

Calculate Sturm* s functions for the following equation, and show that

4}.
all

the

loots are real:

-5x-l - 0.
7.

Calculate Sturm's functions for the following equation, and

roots are imaginary

show that

four'

5*3

0.

This and the preceding example are instances in which, as the student will
common to two f Sturm's remainders which are not

easily see, there is a factor

consecutive.
8. Calculate

clusioni of

Ex.

Sturm's functions for the following equation, and verify the


with regard to the character of the roots :

23, p. 104,

x6

bpx*

+ 6>2* -f

1q

= 0.

213

Examples.
Prove that, if *

I.

lias

any value except


a?*-2c2 a;8

3
<?

unity, the equation

2*

-f

has a pair of imaginary roots.

Prove that the roots of the equation

10.

ip3

are

and solve

all real,

it

a2

when two

02)

2flfo

of the quantities

a, b, c

become equal.

Prove that when the biquadratic

II.

/(*)
has a triple factor,

it

may

a 3/(ar)3
12. Verify,
filled

when the

by means

ax*

4bx*

6<?s*

+ idx 4

be expressed in the form


{

ax

+ V^lf} 3

b- 3V^lf}.

{ax +

of Sturm's remainders, the conditions

biquadratic of the preceding

example

which mustbe

ful-

a perfect square, and prove

is

in that case

a\f(x)= {{ax

b)

+3H}

2
.

'

are present, the number of changes o


Prove that,
terms
is equal to the number of pairs of
of
the
the
coefficients
leading
among
igu

when all Sturm's functions

1 3.

imaginary roots of the equation.


14. If the signs of the leading coefficients of the first two of Sturm's remainder*,
for a quintic be
h, prove that the number of real roots is determined.
Am. One real root only.
15. If JJTand

imaginary

ooufficients

/ are

both positive, prove that

all

the 'roots of the biquadratic are

and that under the same conditions the quintic written with binomial
has only one real root.
(MR. M. ROBERTS, Dublin Exam. Papers, 1862.)

16. In the application of Sturm's theoiem, if any function be reached whose


signs are all positive or all negative, the number and situations of the positive
roota of the

original

Stunnians; and

if

equation

can be examined without the aid of the lower

a function be reached

whose signs are alternately positive and

.negative, the negative roots of the original equation

may be

discussed in a similar

manner.
are real, prove that all the roots of
17. If all the roots of any equation f(x] =
every one of Sturm's auxiliary functions aie also real.
This can be established by reasoning similar to that of Art. 96. Consider the
jfef* remainder
This and the m functions which follow
jfo, and let its degree be m.
constitute

series of

which no adjacent two can vanish together.


and negative, and when a? =

their signs are alternately positive


positive.

There

are, therefore,

m changes

of sign to be lost as

oo

=-

oo

they are all

x passes from -

+ oo and no change of sign can be lost except on the passage


fa = 0, which equation must consequently have m real roots.
to

When
-f-

oo

through a root of

Since a value of x which causes any of the functions to vanish gives opposite
signs to thje two adjacent functions, it is easily inferred that any equation of the
series is

a limiting equation with regard to the function which precedes

it.

214

Separation of the Roots of Equations.

18. If the real roots of any one, /(), of the Sturmian auxiliary functions be
known, prove that the number and positions of the roots of the original equation
may be determined without the aid of the functions below fm(x}Let the real roots, in order of magnitude, of/m (#) =
the
.
be a,
TJ, 9
,

oo to a value a little less


remaining roots being imaginary. As x varies from
than a, the function fm (x) cannot change its sign; and therefore in examining
the roots of f(x) =
which lie between these limits, the Sturmians which follow

fm (x) may be

disregarded.

greater than

to one a

The same

little less

than

holds true as x passes from a value a

iy

and similarly

for the

little

remaining intervals.

we examine separately the intervals { co 0}, {0, ij}, . .. {&, a},


the number of roots of the original equation which lie in each of these
legions can be determined without the aid of the lower Sturmian functions.
If therefore

{a, -f oo },

19. If any one of Sturm's auxiliary functions has imaginary roots, the original
(Mil. F. PURSER.)
equation has at least an equal number of imaginary roots.
This can be inferred from the preceding example by examining the greatest

possible number of changes which can be lost in the series terminating with fm (x),
oo to -f oo
remembering that, so far as the limited
during the passage of x from
series is concerned, a change of sign may be gained on the passage through each
;

rcul root of

fm (x) =

0.

Apply the method of Ex. 18 to the equation of Ex. 1, Art,


Disregarding the two lowest Sturmian remniiulers, we have

20.

/(a-)

aar'-f

/*(*)

7?i

The
and

E-

(0, 1).

of x- in

8.1

-f

4 14*

Ox*

29/-

7*- ^

78jf

10*+
-f

1,

10,

14.

lie in the intervals (- 3, - 2)


has
two
imaginary roots, since the coefficient
f(x)
The real roots, if any, must be negative. The three

roots of the equation ll\

The equation

4a-

98.

=
=

are easily seen to

E\ is negative.
functions above written are sufficient to determine the existence and situations of
roots in the intervals (-00,

3)

and (-2,

0).

It is at once seen that

roots of the original equation are situated in the latter interval.


It will be found possible in many examples to avoid in this

of the last

sary to

two Sturmian remainders

know

and

it

way the

will be observed that

it is

two

real

calculation

not neces-

the actual roots of the quadratic function, but only the intervals in

which they are

situated.

CHAPTER XL
SOLUTION OF NUMERICAL KQUATIONS.

There
Algebraical and Numerical liquations.
an essential distinction between the solutions of algebraical
and numerical equations. In the former the result is a general
formula of a purely symbolical character, which, being the general
101.

is

expression for a root, must represent all the roots indifferently.


must be such that, when for the functions of the coefficients
involved in it the corresponding symmetric functions of the
It

roots are substituted, the operations represented

signs J,
roots of

by the radical
become practicable and when the square and cube
these symmetric functions are extracted, the whole

*J

expression in terms of the roots will reduce down to one rc&t


the different roots resulting from the different combinations
2
J f square roots, and *J, w*J, o> Jj of cube root*. For a
:

simple illustration of what is here stated, we rel'er to the case


of the quadratic in Art. 55.
In Articles 59 and 66 we have
similar illustrations for the cubic and biquadratic.
It is to be

observed also that the formula which represents the root of an


algebraic equation holds good even when the coefficients are

imaginary quantities.
In the case of numerical equations the roots are determined
separately by the methods we are about to explain and, before
;

attempting the approximation to any individual root, it is in


general necessary that it should be situated in a known interval

which contains no other real

The

root.

real roots of numerical equations

mensurajble or inoompiawirable
integers, fractions,

may

be either cpwr

the former class including

and terminating or repeating decimals, which

216

Solution of Numerical Equations.


the latter consisting of interminable
;
roots of the former class can be found exactly,

are reducible to fractions

The

decimals.

and those of the latter approximated to with any degree of


accuracy, by the methods we are about to explain.
We shall commence by establishing a theorem which reduces
the determination of the former class of roots to that of integer
roots alone.

Theorem.

102.

term

first

is

unity,

An

and

equation in which the coefficient of the

the coefficients of the other terms whole

numbers, cannot have a commensurable root which


number.
if possible, let
j,

For,

a fraction in

its

is

not a whole

lowest terms, be

a root of the equation


#* + j^tf"- 1 +

n ~*
+ ....+
p x
2

p^x

+ pn -

we have then

from which, multiplying by

Now

an

not divisible by

is

side of the equation

is

an

b,

w~l

we

obtain

and each term on the right-hand

integer.

tion in its lowest terms equal to

Hence

We have,

therefore, a frac-

an integer, which

- cannot be a root of the


equation.

The

is

impossible.

real roots of

the equation, therefore, are either integers or incommensurable


quantities.

Every equation whose

coefficients are finite numbers, fraccan


tional or not,
be reduced to the form in which the coefficient
of the first term is unity and those of the other terms whole

numbers

(Art. 31)

so that in this

way, by the aid of a simple


commensurable roots

transformation, the determination of the

in general can be reduced to that of integer roots,

Newton's Method of Divisors.

217

We proceed to

explain Newton's process, called the Method


of Divisors, of obtaining the integer roots of an equation whose
coefficients are all integers.

Newton's Method of Divisors.

103.

Suppose h

an

to be

integer root of the equation

+ .... + an .ix + a n =
Let the quotient, when the polynomial

in which J

0.

(1)

divided by x -

is

&c., are clearly all integers.


as
in Art. 8, we obtain the following equations
Proceeding
,

#1,

flo

fln-2

be

A,

b 0y

bi

cii

= b n _ t - A6 n _3,

fl

- hb
n _!

a-i

b2

- hb l9 ....

= &_! - A^ n -,

= - hb n-i*

The last of these equations proves that a n is divisible by h,


the quotient being - J^. The second last, which is the same as

proves that the

sum

oond last coefficient


- i^a and so on.

of the quotient thus obtained


is

again divisible

by

A,

and the

se-

the quotient being

Continuing the process, the last quotient obtained in this


way will be - & which is equal to - a Q
.

we perform

the process here indicated with all the divia


sors of n which lie within the limits of the roots, those which
If

satisfy the
step,

and a

equation.

above conditions, giving integer quotients at each


final quotient

Those which

-a
equal to
at

any

are roots of the proposed

stage of the process give a frac-

tional quotient are to be rejected.

When

the coefficient a

1,

we know by the theorem of


determined in this way

last Artiple that the integer roots


all

the commensurable roots of the proposed equation.

the
are

If a be

218
not =

Solution of Numerical Equations.


the process will still give the integer roots of the equastands ; but to be sure of determining in this way all

1,

tion as

it

we must first transform the equation


one which shall have the coefficient of the highest term equal

the commensurable roots,


to

to unity.

104. Application of the Method of Divisors.


With a
view to the most convenient mode of applying the Method oi

we

write the series of operations as follows, in a


analogous to Art. 8

Divisors,

manner

~/^n-2

The

first

-//^-3

~//i,

'hi,

figure in the second line (-

b^}

is

obtained

by

This is to be added to a n_i to obtain the first


dividing a n by h.
in
the
line
third
(-hh^. This is to be divided by h to
figure
this to be
obtain the second figure in the second line (- & n _ 2 )
added to a^ and so on. If h be a root, the last figure in the
;

second line thus obtained will be - a

When we succeed in

proving in this manner that any integer


h is a root, the next operation with any divisor may be performed,
hut on those of the
not on the original coefficients a n > n-i>
>
second line with their signs changed, for these are the coefficients
of the quotient when the original polynomial is divided by x - h.

When any divisor gives at any stage a fractional result it is to


be rejected at once, and the operation so far as it is concerned
stopped.

The numbers
divisors of #m

and -

1, which are always of course integer


need not be included in the number of trial divisors.

more convenient before applying the Method of Divisors to


determine by direct substitution whether either of thesq numbers

It is

is

a root.

219

Examples.

EXAMPLES.
1.

Find the integer roots of the equation


Z*

By grouping
roots lie

and

-f

5.

commence with

The operation

we

a root

24

-2,

2,

is

a root

all

the

4.

3,

-2

38

-13

-6

32

-5

number

is

not divisible by 4, 4 cannot be a root.

38

-13

-2

-8

10

-1

-1

30

-3

-3

0;

coefficients of the

integer, 2,

we make

use, as

second line with signs changed:

8-10

-3

-1

-6-2
hence 2 also

0.

observe without difficulty that

and in proceeding with the next

above explained, of the

following divisors are possible roots

- 5
stops here, for since

proceed then with the

is

+ 38* -

13*'

-24

hence 3

The

-3,

-24

We

the terms (see Art. 87)

between

-4,

WQ

2*3

0;

-431

and we proceed with

6;

hence - 2

is

not a root, for

does not divide

it

does not divide 5.

is

plainly not a root, for

it

4.

[We might at once have struck out 3 as not being a divisor of the absolute
term 8 of the, reduced polynomial. This remark will often be of use in ^^^inhing
the

number

of divisors.]

220

Solution of Numerical Equations.

We proceed now to the last divisor,


-4

4.

-1

Thus - 4 is a root.
The equation has, therefore, the integer

roots 3, 2,

and the

last stage of

the operation shows that when the original polynomial is divided by the binomials,
x - 3, x
Hence the original
1 ; so that 1 is also a root.
2, x + 4, the result is a?
is

polynomial

2.

equivalent to

Find the integer

roots of

3**

The

- 23* 3

roots lie between

-1-

and

35**

+ 31* -

30

0.

we have

hence

only to test the divisow

2, 3, 5, 6.

"We find immediately that 6


For 5 we have

is

-30

hence 3

The

is

is

For

a root.

a root

quotient,

35-23

31

hence 6

not a root.

25

40

-16

we have

-6

-8

-1

-3

-3

-9

and we easily find that 2

when

8-3

0;

0;
is

not a root.

the original polynomial

divided by (x

is

5) (x

3), is,

from

the last operation,

3*2
of this 1 is not a root,

posed equation are

The

1 is

+ *-2;

a root.

Hence

all

the integer roots of the pro-

1, 3, 5.

other root of the equation

an integer,
3.

and -

is

Find

is -.

It is a

commensurable root

but, not being

not given in the above operation.

all

the roots of the equation


x*

Limits of the roots are

+ * - 2*2 + 4* -

24

0.

4, 3.

^^

Ant. Rooti

- S,

2,

V-f.

Method of Limiting
4.

Find

all

roots lie

- 2* -

between - 6 and

We find that
is

Number of Divisors.

221

the rooti of the equation


*

The

the

hence 2

2, 3,
is

19*2 + 68* - 60

0.

6.

5 are roots,

a douhle root.

and that the factor

The polynomial

(*-2)(- 3) (* +

is

left after

the final division

therefore equivalent to

).

In Art. 106 the case of multiple roots will be further considered.

105.
It

is

Method of Limiting the Number of Divisors.

possible of course to

determine by direct substitution

whether any of the divisors of a n are roots of the proposed equation ; but Newton's method has the advantage, as the aboveexamples show, that some of the divisors are rejected after very
little labour.
It has a further advantage which will now be
the number of divisors of a n within the limits
When
explained.
is important to be able, before proceeding
with the application of the method in detail, to diminish the
number of these divisors which need be tested. This can be

of the roots is large, it

done as follows
If h

and the

is

an integer root of f(x) =

0,

f(x)

by x- h,
was above

divisible

is

coefficients of the quotient are integers, as

If therefore we assign to a? any integer value, the


explained.
of
the
quotient
corresponding value of /(#) by the correspond-

We

h must be an integer.
take, for convenience,
the simplest integers 1 and - 1 ; and, before testing any divisor h,
we subject it to the condition that /(I) must be divisible by

ing value of x

1 - h (or, changing the sign, by h - 1) ; and that /(- 1)


be divisible by - 1 - h (or, changing the sign, by 1 + A).

must

In applying this observation it will be found convenient to


if either of thesa
Calculate /(I) and/(- 1) in the first instance
vanishes, the corresponding integer is a root, and we proceed
with the operation on the reduced polynomial whose coefficients
:

have been ascertained in the process of finding the result o


substituting the integer in question.

Solution of Numerical Equations.

222

EXAMPLES.

**

1.

The

between

roots lie

We have the

4,

281a?2

257*

440 -

0.

and 24.

10,

8,

6,

11,

22.

20,

easily find

=-840,

/(I)

We

+ 160*8 -

following divisors

2,

We

23ar*

/(-I) =-648.

and

therefore exclude all the above divisors, which,

when diminished by

1,

do

The first
not divide 840; and which, when increased by 1, do not divide 648.
condition excludes 10 and 20, and the second 4 and 22.
Applying the Method of
Divisois to the remaining integers 2, 6, 8, 11, we find that 5, 8, and 11 are roots,
Hence the given polynomial is equiand that the resulting quotient is x 2 + x + 1
.

valent to
(*

J3 -

2.

The

/(I)

f(

=0;
i)

We

as

- 2,

so

124

1 is

(*-

6)

29.i-

between

roots lie

Divisors

8)

(*-

11) (*

+ *+!).

- 31* 3 + 31* 2 - 32* + 60 =

0.

and 32.

2, 3, 4, 6, 6, 10, 12, 15, 20, 30.

a root.

and the above condition excludes

easily find that

The given polynomial

is

and 30 are

equivalent to

roots,

(x

all

the divisors except

and that the


1)

(x

2
30) (x + 2) (# -f

Determination of Multiple Roots.

106.

2, 3, 30.

final quotient is

t2 +

1.

1).

The Method

of Divisors determines multiple roots when they are commensurable.


In applying the method, when any divisor of a n which

found to be a root is a divisor of the absolute term of the reduced polynomial, we must proceed to try whether it is also a
root of the latter, in which case it will be a double root of the
is

proposed equation. If it be found to be a root of the next


reduced polynomial, it will be a triple root of the proposed ; and

Whenever in an equation of any degree

so on.

there exists only

*
one multiple root, r times repeated, it can be found in this
way
for the common measure of f(x) and f(x) will then be of the

form

(x

1*" 1

a)

mensurable

if

and the

coefficients of this could

a were incommensurable.

not be con*

223

Determination of Multiple Roots.

Multiple roots of equations of the third, fourth, and fifth


degrees can be completely determined without the use of the
process of finding the greatest common measure, as will appear

from the following observations


(1).

The Cubic.

In

must he comnot high enough to allow of two

this case multiple roots

mensurable, since the degree

is

distinct roots being repeated.


(2).

In

The Biquadratic.

this case either the multiple roots

are commensurable or the function

is

For the

a perfect square.

only form of biquadratic which admits of two distinct roots

being repeated

is

(x

a)'- (,r

/3,.',

The roots of the quadratic may


the square of a quadratic.
be incommensurable. If we find, therefore, that a biquadratic
has no commensurable roots, we mubt try whether it is a perviz.

i'ect

square iu order to determine further whether

incommensurable
The

it

has equa^

roots.

In this case, either the multiple roots are


the
function consists of a linear commenor
commensurable,
surable factor multiplied by the square of a quadratic factor.
For, in order that two distinct roots may be repeated, the
(3).

Qidntic.

function must take one or other of the forms

In the

latter case the roots

cannot be incommensurable

but the

correspond to the case of a commensurable factor


multiplied by the square of a quadratic whose roots are incommensurable. If then a quintio be found to have no commen-

former

may

surable roots,

it

can have no multiple roots.

If

it

be found to

have one commensurable root only, we must examine whether


the remaining factor is a perfect square.
If it have more than
one commensurable root, the multiple roots will be found
the commensurable roots.

among

Solution of Numerical Equations.

224

EXAMPLES.
Find

1.

all

the commensurable roots of

-31*2 +

2*3

The

roots lie

between the

therefore a root.

is

The

1, 16.

Proceed

0.

divisors are 2, 4, 8.

31

15

-2

120

-16

112

64

limits

112* + 64

now with

-8

0;

the reduced equation.

-16

-2

- 16
8

is

a root again, and the remaining factor

is

2#

-f 1

Am. /(*).(2*+l)(*-8)'.
Find the commensurable and multiple roots of

2.

xi

The

roots lie

30^2

limits,

-6,

12.

between the

x*

76a

56

0.

(Apply method of Ex. 10, Art. 87.)

Ant.

f(x)

(x

2}^ (x

7).

Find the commensurable and multiple roots of

3.

40*

10

-f

0.

The roots lie between the limits


2, 5.
The equation as it stands is found to have no
but it may slil]
ii^teggyroot
have a commensurable root. To test this we multiply me roots by 3 in order tc
;

get rid of the coefficient of x*.


x*

Limits

We
to (x~

12#

-x*- llx

120a?

-f

-f

15.

4 to be a double root of this,

find

following

4.

6,

"We find then

9) (x

4)

The

and the function to be equivalent

original equation is therefore identical with tin

Find the commensurable and multiple roots of


s*

12*

-f

32* - 24s

m 0.

225

Newton's Method of Approximation.


between - 12 and

1.
The only divisors to he tested are, therefore,
has no commensurable root.
the
proequation
2,
4,
This can be done by
ceed to try whether the given function is a perfect square.
find
extracting the square root, or by applying the conditions of Ex. 3, p. 125.

The

roots lie

We

We find that

1.

We

that

it is

has two

Find the commensurable and multiple

6.

roots of

8* 2 +

The

limits of the roots are

We find

that

-3

is

and that there

The only

is

4, 4.

4* 3 + 8*

no other commensurable

-f

(x*

/(*)

8**

~2*

-2)*(* +

An*.

The following equation has only two


**

In general

The

it

is

13**

-f

obvious that

h- as a factor,

67**
if

here

800**

this

f(x)m

18
(*-

0.

+ l)(*-2) (*-3) a

find

+ 210* - 108 =

them

0.

an integer root h occurs twice, the last coefficient


if the root occurs three times,
last A

22

33 .

Hence,

and h
if

-x+

102**

of the third last coef-

neither

That these are the roots

3.

The equation

has ecfual roots

In

+ 21* -

factor of the last, A z of the second last,


last coefficient

this

3).

different roots

171*'

and the second

the required roots must be 2 and


8.

when

roots of

22* 3 - 26*-

-f

therefore,

is,

found to be a perfect square, and we have

It is

/(*)

ficient.

0,

root.

Find the commensurable and multiple

must contain
h 3 must be a

case of possible occurrence of multiple roots

latter function is a perfect equaie.

7.

12=0.

282:

a root, and that the reduced equation 10

&

6.

Hence the given equation

2 (cf. Ex. 1, p. 163).


the square of *2 -f 8*
incommensurable.
both
of
equal roots,
pairs

is

nor

1 is

a root,

easily verified.

find all the roots.

example

it

i*w5aikyenient to

change the roots into their reciprocals before

applying the Method of Divisors.

Ant.

/(*)(10*-8)(6-l)(4*+l).

Newton's Method of Approximation. Having


shown how the commensurable roots of equations may be obtained, we proceed to give an account of certain methods of
107.

obtaining approximate values of the incommensurable roots,

The method

of approximation,

commonly

ascribed to Newton,*

which forms the subject of the present Article,


* See Note

at the

end of the volume.

is

valuable as

226

Solution of Numerical Equations.

being applicable to numerical equations involving transcendental

which involve algebraical functions

functions, as well as those

Although when applied to the latter class of functions


Newton's method is, for practical purposes, inferior in form to
Homer's, which will be explained in the following Articles, yet
in principle both methods are to a great extent identical.
In all methods of approximation the root we are seeking is
supposed to be separated from the other roots, and to be situated
only.

in a

known

interval between close limits.

be a given equation, and suppose a value a to


Let/(#) =
be known, differing by a small quantity h from a root of the
We have then, since a + h is a root of the equation,
equation.

/ (a +

h)

or

*+

A +

Neglecting now, since h


the first, we have
giving, as a

first

is

small, all powers of h higher than

approximation to the root, the value

Representing this value by &, and applying the same process


second time, we find as a closer approximation

By
t6

any

/w

repeating this process the

approximation can be carried

degree of accuracy required.


EXAMPLE.

Find an approximate value of the positive root of the equation


x* - 2# - 5 = 0.

The

root lies between 2

found to
cannot

lie

and 3 (Ex.

between 2 and

differ

2'2.

1,

Art. 96).

We take 2'1

from the true value a

-f

Narrowing the

h of the root by more than

easily
.

f'(a)

f'(2-l)

limits,

the root

as the quantity represented

11-23

00543

0*1.

by a.

We

if

It
find

Homer's Solution of Numerical Equations.


A

flrtt

approximation

is,

therefore,
2-

Taking

this as 4,

227

1-0-00643 -2-0946.

we obtain
and calculating the
fraction"^-,

--r = 2-09466148'
/

for a second approximation

and

I*)

so on.

The approximation in Newton's method


When, however, the root we are seeking
another nearly equal to

it,

the fraction

is,

in general, rapid.

is

accompanied by

-r- is

not necessarily

small, since the value of either of the nearly equal roots reduces
case of this kind requires special
f(x) to a small quantity.

We do not enter into any further discussion of


precautions.
the method, since for practical purposes it may be regarded as
entirely superseded by Horner's method, which will now be
explained.
108. Homer's

tions.

By

this

Method of Solving Numerical Equamethod both the commensurable and incom-

mensurable roots can be obtained.

The

root

is

evolved figure

the integer part (if any), and then the decimal


by figure :
part, till the root terminates if it be commensurable, or to any
first

number of places required if it be incommensurable. The process is similar to the known processes of extraction of the square
and cube

which

are, indeed, only particular cases of the


general solution by the present method of quadratic and cubic

root,

equations.

The main

principle involved in Horner's method


oessive diminution of the roots of the given equation

is

the suo-

by known
manner explained in Art. 33.
The great
the method is, that the successive transformations

quantities, in the

advantage of

are exhibited in a compact arithmetical form, and the root


obtained by one continuous process oorrect to any number of
places of decimals required.

This principle of the diminution of the


illustrated in the present Article

roots

by simple examples.

will

be

In the

228

Solution of Numerical Equation*.

some additional

principles which tend


to facilitate the practical application of the method will be

Articles which follow,

explained.

1.

Find the positive root of the equation


2* 8

The

first step,

flrtt figure of

- 85s 2 - 85* -

87

0.

when any numerical equation

is proposed for solution, if to find th*


This can usually be done by a few trials although in cei>

the root.

methods of separation of the roots explained in Chap. X. may have


to he employed.
In the present example there can he only one positive root and
it is found by trial to lie between 40 and 60.
Thus the first figure of the root is 4.
tain cases the

We now

The transformed equation

diminish the roots by 40.

and 10.

between

It is

found by

trial to lie

between 3 and

4.

will

We

have one root

now

diminish

the roots of the transformed equation by 3 ; so that the roots of the proposed equation will be diminished by 43.
The second transformed equation will have one xoot

and

between
that

its

1.

On

absolute term

diminishing the roots of this latter equation by -6, we find


reduced to zero, i. e the diminution of the roots of the pro-

IB

posed equation by 43*6 reduces


a root of the given equation.
as follows

is

its

absolute term to zero.

The

We

conclude that 43*6

series of arithmetical operations is represented

(43*

173
1

174

The broken lines mark the conclusion of eacri transformation, and the figures in
dark type are the coefficients of the successive transformed equations (see Art. 33).

Thus
2* 8 + 165* 2

+ 2716* - 11487

229

Example*.

b the equation whose roots are each less by 40 than the roots of the given equation,
and whose positive root is found to lie between 3 and 4. If the second transformed
equation had not an exact root *5 but one, we shall suppose, between '6 and -6, the
;

three figures of the root of the proposed


equation would be 43-5 ; and to find
the next figure we should proceed to a further transformation, diminishing the roots
first

-5

by

and so on.
Find the positive root of the equation

2.

31*

276

0.

We first write down the arithmetical work, and proceed to make certain observation*
On

it

-31

-276

66

210

35
24

35
210

-65
61-392

245

13-608
13-608

11-96

256-96
12-12

26908
3-08

272*16

614
2

61-6

We find by trial that the proposed equation has its positive


The

first

figure of the root is therefore 6.

g*

has therefore a root between

The

two

Dimmish

root

the roots

by

between 6 and
6.

1.

The equation

+ 59s + 246* - 66
and

1.

It is

found by

trial to lie

between

*2

and -3.

the root of the proposed equation are therefore 6*2.


Diminish the roots again by *2. The transformed equation is found to have the
root -06.
Hence 6*26 is a root of the proposed equation.
first

figures

of

This can

It is convenient in practice to avoid the use of the decimal points.

When

the decimal part of the root (suppose 'at* . . .)


Ks about to appear, multiply the roots of the corresponding transformed equation by
the right
10, i.e. annex one lero to the right of the figure in the first column, two to
easily be effected as follows

of the figure in the second column, three to the right of that in the third and so on,
Lf there be more columns (as there will of course be in equations of a degree higher
;

than the third).


n-kc

The

root of the transformed equation is then, not 'abt . . . , but


4.
The transformed equation has a root ***...

Diminish the roots by

Multiply the roots of this equation again by 10.

The

root becomes

*..., and

230

Solution of Numerical Equation*.

the process

is

To

continued as before.

In

omitting the decimal points.

all

illustrate this

we

repeat the above operation,

subsequent examples this simplification will b

adopted:
(6-26

6160

Find the positive root of the equation

3.

20z

The
7

ab

'

root is easily found to lie

equation

When

- 121* - HI =

121*'

between

the roots are diminished by

and
7,

8.

0.

It

is,

therefore, of the

and multiplied by

form

10, the resulting

is

20*8 + 2990** + 112500* _ 67000

0.

The positive root of this is a I . . and aR the root clearly lies between and 1,
therefore place zero as the first figure in the decimal part of
we have a = 0.
the root, and multiply the roots again by 10, before proceeding to the second trans.

We

formation.

6 is easily seen to be a root of the equation thus transformed.

Am.

7-06.

In the examples here considered the root terminates at an

When

the calculation

is of greater
length, if it
the
successive
were necessary to find
figures by substitution, the
labour of the process would be very great. This, however, is

early stage.

not necessary, as will appear in the next Article ; and one of


the most valuable practical advantages of Homer's method is,
that after the second, or third (sometimes even after the
figure of the root

is

found, the transformed equation

by mere inspection the next figure of the root.


this simplification will

now be

explained.

The

first)

itself suggests

principle of

231

Principle of the Trial-divisor.

Principle of the Trial-divisor.

109.

We have seen

in

Art. 107 thatwhen an equation is transformed by the substitution


of a 4 h for #, a being a number differing from the true root by
a quantity h small in proportion to a, an approximate numeri-

Now

the
h is obtained by dividing /(a) by f(a}.
successive transformed equations in Uorner's process are the
results of transformations of this kind, the last coefficient being
cal value of

f(a),

and the second

last

(see Art. 33).

f(a)

Hence, after two

or three steps have been completed, so that the part of the root
remaining bears a small ratio to the part already evolved, we
may expect to be furnished with two or three more figures of the
root correctly

by mere division of the

by the second last


We might thereHomer's operations, apply

fore, if

we

Newton's

last

transformed equation.

coefficient oi the final

pleased, at any stage o;


method to get a further approximation to the root.

In liorner's method

this principle is employed to suggest the


next following figure of the root after the figures already
obtained.
The second last coefficient of each transformed equation is called the trial-divisor.
Thus, in the second example of

the last Article, the


trial-divisor

number
In

2690800.

is

this

correctly suggested

by the

example, indeed, the second

figure of the root is correctly suggested by the trial-divisor


of the first transformed equation
although, in general, such
In practice the student will have to estimate
is not the case.
;

the probable effect of the leading coefficients of the transformed


he will find, however, that the influence of these
equation
;

terms becomes

less

and

less

as

the

evolution

of

the root

proceeds.

EXAMPLES.
1.

Find the positive root of the equation


x3

-I-

a;

4-

x - 100

correct to four decimal places.


It is easily seen that* the root lies

and proceed

to

make

observations on

between 4 and
it

6.

We write

down the work,

Solution of Numerical Equations.

232

(4-2644

13792
First diminish the roots

by

As the decimal

4.

part is

now

about to appear,

attach ciphers to the coefficients of the transformed equation as explained in Ex. 2,


Since the coefficient 130 is small in proportion to 6700, we may expect
Art. 108.

The figure to
that the trial -divisor will give a good indication of the next figure.
he adopted in every case as part of the root is that highett number which in th* pro-

cm

Here 2
of transformation will not change the sign if the absolute Urm.
In diminishing by 2 the roots of the transformed equation

if

tilt

proper figure.

**

ISO*2

the absolute term retains

its

absolute term would have

become

gone beyond the root.

+ 6700* -

sign (- 4072).

positive, the

We must take

16000 -

0,

we had

adopted the figure *, the


change of sign showing that we had

II

care that, after the

first

transformation (the

reason of this restriction will appear in the next example), the absolute term preserves its sign throughout the operation.
If we were to take by mistake a number

would show itself, just as in ordinary division or evolution, by


number being greater than 9. Such a mistake, however, will
The error which is most common if to take the number too large,

too small, the error

the next suggested


rarely be made.

233

Examples.
and

this will

show

In the above work

itself in
it is

the work

the corresponding tigure of the root


places is 4*2644.
2.

The equation

has one root between

10644

**

by the change

of sign in the absolute term.

evident, without performing the fifth transformation, that


is 4,

so that the correct root to four decimal

+ 4#3 - 4*8 - 11* + 4

and 2

find its value correct to four decimal places.

234

Solution of Numerical Equations.

We see without completing the


The

the root.

The

fifth

transformation that 9

is

the next figure of

root is therefore 1'6369 correct to four decimal places.

trial-divisor

becomes effective after the second transformation, suggesting


and all subsequent numbers. The first transformed equation

correctly the number 3,

We

its last two terms negative.


may expect, therefore, that the influence of
the preceding coefEueients is greater than that ol the trial -divisor, as in fact is here
The number 6, the second figure of the root, must be found by substituthe case.

has

We have to

tion.

determine what

the equation

few

& + 80*3 4
show that

trials

root lies between 6 and 7

subsequent

trials

we

is

the situation between

HOD* 2 - 3000*

- 60000

and 10 of the root

of

0.

Hence the
number of which we are in search. In the
numbers 3, 6, 9, in succession, which allow

6 gives a negative, aiid 7 a positive result.

and 6

is

the

take those greatest

In the first transformation, diminishing


its negative sign.
a change of sign in the absolute term.
The meaning of this
have parsed over a root lying between and 1, for gives a positive

the absolute term to retain


the roots by
is,

we

that

result, 4

1,

and

there

is

gives a negative result,

6.

In

all

subsequent transfonnations,

we keep helow

the root, the sign of the absolute term must be the same
This supposes of course that no
as the sign resulting from the substitution of 1.
This supposition we have
rout lies between 1 and that of which we are in search.
so long as

already

made

in the statement of the question.

have only two positive roots


one between 1 and 2.

When two

roots exist

one of them

lies

In fact the proposed equation can


and 1, and therefore only

between

between the limits employed in Homer's method,

i.e.

when

the equation has a pair of roots nearly equal, certain precautions must be observed
which will form the subject of a subsequent Article.
3.

Find the root of the preceding equation between

places.

Commence by multiplying by
1,

40,

400,

The

10.

and

to

four decimal

coefficient are then

11000,

40000;

the trial-divisor beoomec effective at once in consequence of the comparative small-

The positive sign of the absolute term must be


ness of the leading coefficients.
An*. -3373.
preserved throughout.
4. Find to three places ol decimals the root situated between 9 and 10 of the
equation

[Supply the zero

**-3*3 + 76*coefficient of #*.]

10000

0,

Ans. 9'886.

In the examples hitherto considered the root has been found


to a few decimal places only.
We proceed now to explain a
method by which, after three or four places of decimals have

been evolved as above, several more may be correctly obtained


with great facility by a contracted process.

110.

Contraction of

235

Hornets Process.

Contraction of

Homer s

Process.

In the

ordi-

nary process of contracted Division, when the given figures are


exhausted, in place of appending ciphers to the successive dividends, we cut off figures successively from the right of the
the divisor itself becomes exhausted after a

divisor, so that

number

depending on the number of figures it conresulting quotient will differ from the true quotient
In
figure only, or at most in the last two figures.

of steps

The

tains.

in the last

liorner's contracted

method the principle

retain those figures only

which are

is

the same.

We

effective in contributing to

the result to the degree of approximation desired.


When the
contracted process commences, in place of appending ciphers to
the successive coefficients of the transformed equation in the

before explained, we cut off one figure from the right


of the last coefficient but one, two from the right of the last
coefficient but two, three from the right of the last coefficient

way

but three; and so on.

The

effect of this is to retain in their

proper places the important figures in the work, and to banish


altogether those which are of little importance.
The student will do well to compare the first transformation

by the contracted process in the

first of the following examples


with the corresponding step in the second example of the last
He will
Article, where the transformation is exhibited in full.

how the leading figures (those which are most


in
important
contributing to the result) coincide in both cases,
and retain their relative flaces ; while the figures of little
then observe

importance are entirely dispensed with.

In addition

to the contraction

now

explained, other abbre-

viations of Horner's process are sometimes recommended ; but


as the advantage to be derived from them is small, and as they

we do not tiiink it necessary


The contraction here explained

increase the chances of error,

give any account of them.


of so

much importance

method

to
is

in the practical application of Horner's


of approximation that no account of this method is

complete without

it.

Solution of Numerical Equation.*.

236

EXAMPLES.
Find the root between

1.

and 2 of the equation in Ex. 2 of the

last Artiolk

correct to seven or eight decimal places.


result of the

Assuming the

Example

referred to,

we

commence the conThe subsequent

shall

tracted process after the third transformation hat bean completed.

work stands

as follows

26165788
18936

17649439
16213090

(1

-63691367ft

3156

3162
6

255733
285

-9161

2560181

76SO
1471
12SO

191

179
12

Here the

effect of the first cutting off of figures,

last coefficient, 14

altogether the

by 6 as

if

from the third

first coefficient

the coefficients

of a cubic equation.

1,

last,

namely, 8 from the second


last, ia to banish

and 062 from the fourth

of the biquadratic.

We proceed to

- 17649439 which

3150, 2516578,

diminish the roots

are left were thos

In multiplying by the corresponding figure of the root the


and account taken of the^jiujulijatJX)

figures cut off should be multiplied mentally,


t>e

carried, just as in contracted division.

After the diminution by 6 has been completed, we cut off again in the transformed
cubic 7 from the last coefficient but one, 68 from the last but two, and the- first

The work then proceeds as if we were dealing


with the coefficients 31, 255448, - 2336349 of a quadratic. The effect of the next
The subpiocess of cutting off is to banish altogether the leading coefficient 31.
sequent work coincides with that of contracted division. When the operation tercoefficient disappears altogether.

minates, the

two

QJ^

number

to

figures.
carried before the contracted process is

places required

*A th* 1nft
may be depended p" "p
which the evolution of the root must b*

of decimals in the quotient

The extent

three

commenced depends on the number of decimal


commences we shall be furnished, in
already evolved, with as many more as there are figures in

for after the contraction

addition to the figures


the trial -divisor, le&s one.
2.

Find

to

seven or eight decimal places the root of the equation

- 12*
which

lies

between 2 and

3.

237

Examples.

ThJ* equation can have only two positive roots - one lies between
and 1, and
For the evolution of the latter we have the following
the other between 2 and 3.
:

223180

BUJS

On

this

we remark,

that after diminishing the roots

by

2,

and multiplying the roots

we find that the trial-divisor 20000 will not


term 10000 we put, therefore, zero in the quotient, and

of the transformed equation by 10,

"go

into'' the absolute

multiply again by 10, and then proceed as before.


3.

Find the root of the same equation which

lies

between

nod

1.

Ans. -693685829.
4.

Find the positive root of the equation


a

- 67-613*- 3761*2758=0.

When the coefficients of the proposed equation contain decimal points, it will
be found that they soon disappear in the work in consequence of the successive
multiplications by 10 after the decimal part of the root begin* to apponr.
Ann. 11-1973222.
6.

Find the negative root of the equation

**-12* 2 +12'

-8 =

to aeven places of decimals.

When
And

a negative root has to be found,

it is

convenient to change the sign of x

find the 'corresponding positive root of the transformed equation.

Ant.

3-9073735.

Solution of Numerical Equations.

238

Application of Homer's Method to Cases where


Roots are nearly Equal. We have seen in Art. 107 that
the method of approximation there explained fails when the
111.

proposed equation has two roots nearly equal. Examples of


this nature are those which present most difficulties, both in

Ex.

their analysis (see

Horner's method

it

Art. 98) and in their solution.

7,

possible, with very little

is

By

more labour

necessary in other oases, to effect the solution of such


So long as the leading figures of the two roots are
equations.

than

is

the same, certain precautions must be observed, which will be


After the two roots
illustrated by the following examples.
the
have been separated,
subsequent calculation proceeds for

each root separately, just as in the examples of the previous


It is evident, from the explanation of the trialArticles.
divisor given in Art. 109, that for the same reason as that

which explains the failure of Newton's method in the case


(see Art. 107), it will not become effective

under consideration
the

till

first

or second stage after the roots

have been separated.

EXAMPLES.
I

The equation

*3

has two roots between

and 2

lx + 7

(see Ex.

Art. 96)

2,

find

each of them to eight

decimal places.

Diminishing the roots by


roots are multiplied

by

we

1,

find that the transformed equation (after its

10), viz.

x * + 30*2 _ 400*
must have two
3

and

4,

roots

between

and

and the other between 6 and

1000

0,

"We find that these roots

10.

7.

The

roots are

now

lie,

one between

separated,

and we

proceed with each separately in the manner already explained. If the roots were
not separated at this stage, we should find the leading figure common to the two,
and, having diminished the roots by it, find in what intervals the roots of the
resulting equation were situated
2.

and so on.

Am.

1-35689584,

1-69202147.

Find the two roots of the equation

& _ 49^2 +658*-! 379


which

lie

We

between 20 and 30.


complete work of approximation to the smaller of the two
and then make some observations which will )>e a guide to

shall exhibit the

roots to seven places

the student in all cases of the kind

239

Examples.
f

(23-2131277

206^91

The diminution
is

sent concerned.

by 20 changes the sign of the absolute term. This


between and 20, with which we are not at preof the first ti ausf ormed equation

of the roots

an indication that a root

The

roots

exists

a^ 1U

~ 102*+

181

are not yet separated, lying both between 3 and 4.

these numbers gives a positive result, so that

we have

The

substitution of each of

not here the same criterion

to guide us^in our search for the proper figure as in former cases, viz. a change of
have, however, a different criterion which enables
sign in the absolute term.

We

240

Solution of Numerical Equations.


by mere

us to find

substitution the interval within which the

diminish the roots of **

4 23a

x*
lie

-f

34a?4

between

and

Hence the two

we

we

If

4.

diminish

root* lie

by

4,

of sign.

two

4.
They are, therefore, not
The next transformed equation

200s*

found in the same way to have both

If

tie.

we

by 3, the resulting equation (as in


of changes of sign as the equation itself.

between 3 and

4-

root

the resulting equation if


Hence the two roots must

roots

its

number

proceed to diminish by 3.

&
is

181

13*0, which has no change

the above work) has the same

ind

102*

11**

yet separated;

- 900# 4 1000
roots

its

between 2 and 3

the diminution by

2 leaving two changes of sign in the coefficients of the transformed equation (as in
So far, then,
the above work), and the diminution by 3 giving all positive signs.
diminish again by 2.
the two roots agree in their first three figures, viz. 23*2.

We

The resulting aquation # 3 4


1

and 2

and 3

2060jr2

- 8800* 4 1261000 =

giving a positive, and 2 a negative result

The

3 giving a positive result.

roots are

now

has one root only between

its

other root lies between 2

We

proceed, as
the above work, to approximate to the lesser root, by diminishing the roots of this
To approxithe trial divisor becoming effective at the next step.
equation by 1
mate to the greater root, we must diminish by 2 the roots of the same equation,
;

separated.

in

taking care that in the subsequent operations the negative sign, to which the preThe second
viously positive sign of the absolute term now changes, is preserved.
root will be found to be 23-2295212.

So long

as the

two

may be obtained by

roots remain together, a guide to the proper figure of the root


dividing twice the last coefficient by the second last, or the

The reason of this is, that the proposed equation


lecond last by twice the third last.
approximates now to the quadratic formed by the last three terms in each transformed
equation just as in previous cases, and in Newton's method, it approximated to the
simple equation formed by the last two terms, this quadratic having the two nearly
and when the two roots of the equation ax 2 4 bx 4 c =
equal roots foi its roots
;

2e

are nearly equal, either of


i

them
o

above example, the number 3

is

given approximately by -^-

2x

is

suggested by

181
;

10 'Z

or.b

Thus, in the

lrt ,2x

and the number 2 by

1000
.

900

In this way we can generally, at the first attempt, find the two integers between
which the pair of roots lies. We shall have also an indication of the separation of
the roots by observing when the numbers suggested in this way by the last three
coefficients

3.

become

different,

i.e.

when

2c

suggests a different

number from

Calculate to three decimal places eacli of the loots lying between 4 and 6 of

the equation
at 4 8# 3

70x*

144*

4 936 =

0.

Am.
4.

2a

4*242

4-246.

Find the two roots between 2 and 3 of the equation

692s 2

4 649*
.

1446

Am. The

0.

roots are both

2-125.

Lagrangds Method of Approximation.

241

Here we

find that the two roots are not separated at the third decimal place.
diminish by 5, the absolute term vanishes, showing that 2*125 is a
and proceeding with this diminution the second last coefficient also vanishes.

When we
root

Hence 2-125

is

a double root.

When

an equation contains more than two nearly equal


roots, they can all be found by Horner's process in a manner
Such cases are, however, of
similar to that now explained.

The principles already laid down


rare occurrence in practice.
will be a sufficient guide to the student in all cases of the kind.
112. Lagraiige's Method ofApproximation.

Lagrange

has given a method of expressing the root of a numerical equaAs this method is, for
tion in the form of a continued fraction.
practical purposes,

much

inferior to that of

content ourselves with a brief account of

Homer, we

shall

it.

Let the equation f(x) = have one root, and only one root,
between the two consecutive integers a and a + 1. Substitute
a + - for x in the proposed equation.
in y has one positive root.

Let this be determined by trial to


and 6 + 1. Transform the equation

lie

between the integers

in

y by the substitution y -

The transformed equation

The

-.

positive root of the

found by trial to lie between c and c+ 1. Continuing this process, an approximation to the root is obtained
in the form of a continued fraction, as follows
equation in

z is

a + r1

EXAMPLES.
1.

Find

in the

form of a continued fraction the positive root of the equation

* - 2* - 6 =
The

between 2 and

3.

the transformation

root lies

To make

We

-,

we

then find
diminishing jthe roots by 2.
reciprocals of the roots of the transformed.

0.

first

employ the process of Art. 33,

the equation whose roots are the

Solution of Numerical Equation*.

242
The equation

in

is

way found

in this

10v

to

be

- 6y -

0.

This has a root between 10 and 11.

Make now

the substitution

The equation

in z

y=

This has a root between


in

and

which has a root between

94*'-

20*

Take

2.

0.

-.

-f-

+ 25w 2 - 89w -

and 2

61

0,

and so on.

have, therefore, the following expression for the root

10

2.

-.

is

54

We

is

61* 3

The equation

10

Find in the form of a continued fraction the positive root of


*

6a?

13

0.

Ant. 3

113.

Numerical Solution of the

II i quadra tie.

It is

proper, before closing the subject of the solution of numerical


equations, to illustrate the practical uses which may be made of
the methods of solution of Chap. VI.
Although, as before
observed, the numerical solution of equations is in general best

by the methods of the present chapter, there are cerwhich it is convenient to employ the methods
When a
of Chap. VI. for the resolution of the biquadratic.
leads
a
^cubic
to
which
has a
reducing
biquadratic equation

effected

tain cases in

can be readily found, and the


We proceed to solve a
solution of the biquadratic completed.
few examples of this kind, using Descartes' method (Art. 64),

commensurable

which

root, this root

will usually be

found the most convenient in practice.

243

Examples.

EXAMPLES.
Resolve the quartio

I.

a*

6**

3* 2

4-

+ 22* - 6

into quadratic factors.

we

easily obtain

p+p'-S, q+4+lpp'=Z,

pq'+p'q =

Making

the assumption of Art. 64,

also

/ and

/,

the equation for

q qf

6,

+ <f-l)\

*=\-pp'^\(q

and, calculating

\\,

<p is

*- T *- 226
T -*
Ill

Multiplying the roots by

we

4,
*

By

these,

0.

found to have a root - 6

is easily

PP'

t,

-450

lilt

the Method of Divisors this

From

if 4^>

have,

'

2,

- -

j/

1,

hence

5.

we

get

= -

6.

combined with the preceding equations,

p - -

1,

?'

Whn

the values of q and #'are found, the equation giving the value of p<f -I- p'q
determines which value of q goes with p 9 and which with p\ in the quadratic

The

factors.

quartic

is

resolved, therefore, into the factors

(x*-x+ !)(*- 2* -6).


By means
factors

in

of the other

we caa resolve
we can do the same

two values of

two other ways

or

the quartic into quadratic

<f>

ining by solving the two

quadratics already obtained.


2.

Kesolve into factors the quartic


f(x)

The equation

for

<f>

x*

is

12* 2

+ 60* +

63.

is

4^ which

8a^

found to have a root

195<j>

- 475 -

0,

6.

Ans. f(x)

R2

(*

- 2* -

2
3)(a;

6*

21).

Solution of Numerical Equations.

244
3.

Resolve into factors

-20*-fl.
The reducing cubic

found to be

is

217

_
or, multiplying the roots

by

8186

6,

-661* +3186 m

4* 3

0,

This has a root

hence

<t>

s
o
^4n*.

ss

./"(a?)

(r

4* 4

2)(ar

4ar

3).

Eesolve into factors


f(x)

The reducing cubic

m x4 -

6ar

- $x z + 66* - 22.

is

40 _

___

^_

hence
=

~2*
^w*. f(x)

(jr

- 11)(^ - 6* +

2).

6. Resolve into factors

f(x)

x*

- 8s8 + 2U-2

- 26* + 14.

4w*. f(x)
6.

ss

2*

(a?

-|-

2)(**

- 6* +

7)

Resolve into factors


3.

^*.
7.

(*

a;V6

+ 3+

V<J)(*

+ 3

- V6).

- J5) -

3\/7}.

*\/6

Find the quadratic factors of


4**

- 8z 3 -

and solve the equation completely

Ans. {*

12#

(see

+ 84* - 63 *

Ex.

0,

18, p. 34),

2#(2 + V/7)

-f

3 v/7}

{ic

2dr(2

245

Miscellaneous Examples

MISCELLANEOUS EXAMPLES.
1.

Find the positive root of

- 6* -

x*

13

0.

^lw. 3-176814393.
2.

Find the positive root of

- 2* -

-4J. 2-094561483.

correct to eight or nine places.


8.

The equation
2* - 660-8*3 + 6* - 1627

has a root between 300 and 400

find

it.

Ant. Commensurable root, 326


4.

4.

Find the root between 20 and 30 of the equation

+ 1896* - 457 =

4J 3 - 180# 2

0.

Ans. 28-62127738.
6.

Find

to six places the root

x*

between 2 and

+ ess* _

49*2

3 of the equation

1379 ^0.

Ans. 2-557351.
6.

Find to

six places the root

between 2 and 3 of the equation

-3 =

** - 12.r 2 ^ 12*

0.

Ans. 2*868053.
7.

Find the positive root

of the

equation

^ + 2#2 - 23* - 70 Ans. 6*13467872628.

correct to about ten decimal places.


8.

Ans. 8766.

Find the cube root of 673373097125.

Ant. 14.

9. Find the fifth root of 637824.

10.

Find

all

the roots of the cubic equation


3

JF

The equation

x*

x*

0,

3*

-I-

0.

of Ex. 7, p. 100, reduces to thia.

Am. The

sphere whose radius


11.

1-87988,

smaller positive root gives the solution of the problem

Find

all

is

1, 0.

To

1-6320*.

divide a hemi-

unity into two equal parti by a plane parallel to the bos*.

the roots of the cubic


x?

(See Ex.

0-34729,

100.)

**

- 2* -

- 0.

Ans. - 1*80194,

0*44604,

1*24698.

246

Solution of Numerical Equations.

12.

Find

to five decimal places the negative root

between

and

(see

Ex.

3,

p. 100) of the equation

a*

#*

- 4*8 - 3*2 4 3*

-I-

0.

Am. 3.

0*28468.

Solve the equation

- 316*' - 19684* 4 2977260

**

We find

that there

is

a root here between 70 and 80.

The deprt-M

found to be 78.

0.

By

two

d equation furnishes

Horner'a process

it IB

roots, \thich, increased

by

78, are the remaining roots of the cubic.

Am.
14

Find the two

78,

347,

11727*4 40386 =

0.

An*. 3-46692,
is

110.

real roots of the equation

x* -

This equation

given by

Mr

21-4S067.

G. H. Darwin in a Papei On the Prece^im of*

riseom Spheroid, and on the Remote History of the Earth. Phil. Trans., Part ii.
The roots are " the two values of the cube root of the earth's rota1879, p. 608.
T

tion for
1

which the earth and moon move round as a


Find

all

rigid

body."

the roots of the cubic equation

20** - 24*' + 3

Am.

0.

-0-31469,

0-44608,

1*06866.

This equation occurs in the solution by Professor Ball of a problem of Professor


in the Educational Times of Dec., 1878, to determine the deflection of

Towneend'e

beam unifoimly loaded and supported


16.

al its

two ends and point* of trisecion.

Find the positive root of the equation


14* s 4 12**

- 9* -

10

0.

An*. 0-86906.

The equations

17.

and the following example occui


beams supported by props.

in the investigation of

of this

questions relative to

Find the punitive root of the equation


7**

4 20*8 4 3*2 - 16* -

0.

Am.
18.

0-91*36.

Find to tec decimal places the positive root of the equation

& + 12** 4 69*

4 160* 1 + 201* - 207 -

0.

Am.

0-6486068038.

247

Miscellaneous Examples.
19.

Find

all

the commensurable roots of

x6

/(*)

2.r

- 149s 2 - 232* - 336 =

0,

and solve the equation completely.

Ant. f(x)

(x*

+ x+

3) (*

4)

(*

7).

20. Solve similarly the equation

f(x]

mx*~

32.t*

116s 8
<4fi.

21.

Find the condition

should have

its roots

16* 2

115a?

/(*)

(a;

1) (*

84

be positive,

When

when

Tand /are both

the biquadratic has two- roots equal to

23. If the equation f(x)


r'(x)

has

[f(x)Y

3) (9

HI +

ZaJ

3,

28).

Art. 99.

positive.

(since then

/ must

It is therefore easily inferred that the


positive (cf. Ex. 15, 5. 213).
a,

prove

-Gl

a *^

f(x)

(*

1)

and Jare both positive

JET

4
identity o Art. 37).

by the

biquadratic has no real roots


22.

when

is fulfilled

0.

the quadratic Sturmian remainder of Ex.

thai,

imaginary.

Ant.
This condition

has

all its

all

roots real, prove that the equation

its

roots imaginary.

24. If an equation of any degiee, arranged according to powers of x, have three


consecutive terms in geometric progression, prove that it roots cannot be all real.

These three terms must be of the form


be multiplied by x
absent, and

a.

The

A-J?

A'a.**"

-f

aV- 2

resulting equation will have

must therefoie have

two imaginary

at least

two

roots

Let the equation


terms

consc-t utive

but

aH the

roots of

this equation except a are roots of thi* given equation.

26. If an equation have four consecutive coefficients in arithmetic progression,


"

prove that its roots cannot be all real.


This can be reduced to the preceding example. Writing down four teiuw of the
1, it readily appeal's that the lesulting equation
proper form, and multiplying by x
ias three consecutive

terms in geometric progression.

26. Calculate the first

two

of Sturm's remainders for

second term, viz.


a-

/(a?)

Am.

ax*

bx*

JKi

a-

Jfc

a Ax* + BX+

2a,r

;i

3for

ex

+ d=

cx

a quintic wanting the

0.

6rf,

C,

where

- 40<? - 12n 8 - 45*2 ,

B&

5Qa ft -

SaH ~ 60^,

--4

A - 75M.

of the third
Retaining this notation, it is easy to calculate the coefficients D,
remainder
Dx -I- in terms oi a, b, ct tl, A, , C; and, finally, 4 in termt
3
of A, $ t j; D, ^.

Solution of Numerical Equations.

248
Remove

27.

coefficients,

the second term from the general quintic written with binomial
coefficients of the first two of Sturm'i

and prove that the leading

remainders for the resulting equation are

28. Calculate the leading coefficients of the first

an equation of the

n*h degree

* + ax*~*

No

coefficients

two Sturmian remainders

for

wanting the second term, viz.


4-

bx*~*

beyond those here given

c*"-4

-f

&c.

will enter into the K- quired values

we

readily find
J

29.

g1

s ~

s -

Remove

with binomial

2'/.r-

- 3kr- 3

2;

u'

4r*'-<

bnac

-f

9^

&c.
2
{

/?"-

&c.

the second term from the general equation of the n* h degree * ritten
and prove that tlie leading coefficients oi the first two

coefficients,

Sturmian remainders of the resulting equation are

- H,

-nHI+

l<

n - 2) o/.

These expressions are easily derived from the preceding example by aid of the
transformation of Art. 35 the values of ^2, ^3, ^i being given by the equations
;

G* being replaced by

its

value from the identity of Art. 37, and positive multipliers

omitted.
30. Calculate Sturm's functions for Euler's cubic (see Art. 61).

We find,

after

some reductions, and omitting

positive factors,

s 21* -f

All the conditions of Art. 68, with respect to the nature of the roots of the
may be derived from these results, by the aid of Ex. 4, p. 125. And
will be observed that the conditions for reality of all the roots as given in

biquadratic,
it

Art 100, as well as in the Article already referred to, are both obtained here
together ; for, in order that Euler's cubic should have all its roots real and positive,
the substitution of

*/- 12# 2 and

for

2JET/

a;

must give three changes of

- 307

should be both negative.

sign,

and

this requires that

CHAPTER

XII.

COMPLEX NUMBERS AND THE COMPLEX VARIABLE.

Complex Number* Graphic Representation.


In the foregoing chapters many examples have been met with
114.

the solutions of numerical equations of


- 1, involving the extraction of
/>J

of the occurrence

among

quantities of the

form a +

Such an expression,
the square root of a negative number.
consisting of a positive or negative real units, and b positive or
negative imaginary units, is called a complex number (see Art.

The imaginary

J-

1 is denoted for brevity by t.


Real and purely imaginary numbers are both included in the
expression rt + /^.the former being obtained when 6 = 0, and the
0.
latter, when a
Complex numbers may be submitted to all

15

!.

unit

and in the result


the ordinary rules of arithmetical calculation
i
such
calculation
of
any
beyond the first can
integral powers
;

of

always be reduced by the relation

We

explain a

proceed to

numbers geometrically, which

* -

t*

mode
will

1.

of representing

complex

be found very convenient in

the- treatment of functions involving quantities of this kind.

The

expression a + ib
/u

may
(cos

be written in the form

a +

sin a),

where
-

u =

The quantity
amplitude, of the

la

+ 6%

/u is

cos a

MM
a

= -,

sin a

called the modulus,

complex number a + ib.

b
= -

and the angle a the

The modulus is always

taken positively, the negative sign of the radical corresponding


to

an increase of the amplitude by

TT.

250

Complex Numbers and

the

OT

Let rectangular axes OX,


point

X OA =
We
=
IUL

11

such

a,

7) be taken,

and a

//.

OM

have then
cos a = a, and
sin o

(fig.

that

OA =

and

Complex Variable.

AM

=6.

The

if

expression -f ib may
therefore be represen ted graphically by

~~

fo
j

the right line drawn


from
to a point in

lg "

a plane whose coordinates referred to the fixed axes are a, b


the distance OA of this point from the origin being equal to
the modulus, and the angle
equal to the amplitude of
;

XOA

the complex number.


The magnitude of a complex quantity is estimated by the
When the complex quantity
magnitude of its modulus.

vanishes (that

is,

when

and

b separately vanish), its

modulus

vanishes; and, conversely, when thr modulus vanishes, since


9
then a -f b 2 = 0, a and b must separately vanish, and therefore

Two such quantities, a + ib and


the complex quantity itself.
=
a' i ib', are equal when a
a and b = b\ i.e. when the moduli
and when the amplitudes either are equal or

are equal
h

multiple of
Jii

what follows we
of a

mod.

Complex

IV

shall, for brevity, represent the

ib

OA

OA

',

-f

(a

amp.

ib},

(a

ib).

Addition and Subtraction.

umber*.

ib'

be represented by the

so that

= mod.

We proceed

modulus

by the notation

Let a second complex number a +


right line

by

~TT.

and amplitude

115.

differ

(a'

ib'),

XOA

to determine the

a +

ib

mode

+ a +

ib'.

- amp.

(a

4 iV).

of representing the

sum

and

Multiplication

Writing

sum

this

in the

251

Division.

form a +

a'

(b

6'),

we

in accordance with the convention of Art. 114, that

observe,

it

will be

represented by the line drawn from the origin to the point


whose coordinates are a + a', b + b'. To find this point, draw

AB
a\

BP

are equal to
and equal to OA' since AP,
the required point, and we have
- mod. \a + a' + i (b + V) J, JTOft = w/>. {a + *' + (6 + *')J
parallel

b',

OB

is

To add two complex numbers,


represent one of
the second (that

them

and, at

we draw

therefore,

its

OA

to

AB to

represent
equal to the modulus,
and the angle it makes with
equal to the amplitude, of
the second)
then OB represents the sum of the two complex
is,

so that its

extremity,
is

length

OX

numbers.

OB

Since

is

not greater than

OA

-f

AB,

modulus of the sum of two complex numbers


equal

to]

it

is less

follows that

tin

than (or at mo$t

the ttum of their moduli.

This mode of representation

may be extended to

the addition

any number of sucli quantities. Thus, to add a third a" ib"


represented by OA", we draw BC parallel anil equal to OA",
and join OC. Then OC represents the sum of the three, OA,
of

+-

OA\ UA'\

It is evident also that

that the modulus of the

we may conclude

in general

sum of any number of complex

quantities

than (or at most equal to} the ttum of their moduli.


Subtraction can be represented in a similar way.
Siuoo

is less

OA

OL

OA

and OA'
will represent the difrepresents the sum of
To subtract two complex numbers,
ference of OB and OA'.
^

we draw at the extremity of the line representing the


a line parallel and equal to the second, but in an opposite
an angle greater
direction (i.e. a direction which makes with
to the
by v than the amplitude of the second).
join
therefore,
first

OX
We

extremity of this line to find the right line which represents the
difference of the two given complex numbers.
116. Multiplication

two complex numbers a +


a 4 ib
fi (cos a + t sin

and
t*6,

a),

Division.
-+

a'

ib',

we

ib'

write

To multiply
them

j/ (cos a'

in the

4- 1

the

form

sin a')-

Complex Numbers and

252

We

have then, by

De

ib'}

(a

ib) (of

which proves that

the

Complex Variable.

Moivre's theorem,
{cos (a

JJLJJL

+ a)

sin (a

4- i

the product of two complex

plex number, whose modulus

is

(he product

of

a') },

numbers

is

a com-

the two moduli,

and

whose amplitude is the sum of the two amplitudes.


In the same way it appears that the product of any number
of such factors is a complex quantity, whose modulus is the

product of
all the

To

all

the moduli, and whose amplitude

amplitudes.
divide a + ib by a' +

-w

a +

ib

we have

,
cos (a - a +

fuL

= --

-7

il/ 9

sum

of

similarly
.

the

is

sin

- a, )},

-f

JJL

which proves that the quotient of two complex numbers is a complfx number, whose modulus is the quotient of the two moduli, and
whose amplitude
It

the difference

is

of the two amplitudes.

was assumed in the proof of the theorem of Art. 16 that

when a product

any nuniLer of i'aotors (real or imaginary)


must vanish. This is evident when
real.
From what is above proved the same

of

vanishes, one of the factors

the factors are

all

when the factors are complex for, in order


modulus of the product may vanish, one of its factors
must vanish, and therefore the complex quantity of which that
conclusion holds

that the

factor is the modulus.

117.

Other Operation* on Complex Numbers.

From

the foregoing propositions it follows that any integral power of


a complex number, e.g. (a + ib} m y can be expressed in the form
and
are real.
+ i, where
And, more generally, if in

any

rational integral function

a zn + a i*11

whose

coefficients are

"1

rtjss"-*

-f

complex (including

real)

nvmbers, a

The Complex
complex quantity a +

ib

253

Variable.

be substituted for

*,

the result can be

+ ill.
expressed in the standard form
It is not proposed in the present chapter to discuss any
functions of complex numbers beyond the rational integral
function of the kind hitherto treated in this work.

It

is

easy,

however, to show, by the aid of De Moivre's theorem, that


the remaining processes of numerical calculation
powers with

complex exponents, logarithms, and powers whose


base and exponent are both complex reproduce in every case a
fractional or

complex number as result. This is expressed by saying that


complex numbers form a system or group complete in themselves.

118. Tlie

Complex Variable.

In the earlier chapters of

the present work the variation of a polynomial was studied corresponding to the passage of the variable through real values

from - oo to + oo and the mode of representing by a figure


Such a mode of
the form of the polynomial was explained.
treatment is only a particular case of a more general inquiry.
Given a polynomial, rational and integral in s, whose coeffi;

numbers

cients are

f(z)

we may study
of

s,

where

both take

real or complex, viz.

a Q z n + di% n- 1 +
its

r/ 2

n "2

,_!*

-f

an

variations corresponding to the different values

z has the

complex form x + iy, and where x and y


This form x + iy is called the

all possible real values.

All possible real values of the variable are of


complex variable.
course included in the values of x + iy, being those values which
arise

by varying x and putting y =

principles of Art.

114 we

may

0.

In accordance

witli the

represent the complex variable

OP

to
(fig. 8) drawn from a fixed origin
the point whose coordinates are x, y.
Or we may say, x + iy is
Thus all possible values of x + iy
represented by the point P.

x +

iy

by the

line

will be represented

by all the points in a plane. Since, for any


value
of
-f
*,/(*) takes the form
particular
iB(Art. 117), the
values of f[z) may be represented in a similar manner by points

Complex Numbers and

254

a plane.

in

confine ourselves in the present Article to the


of the variable

We

variation of x +

conceive the

iy to take place

for

continuous manner;

example, by

Complex Variable.

We

representation
x + iy itself.

in

the

the motion of the

along a curve. If OP
and OP' represent two consecupoint

x, y>

we

tive values of the variable,

write the corresponding values


x -r iy> x' + //, as follows
:

smx + iymr (cos

sin 0},

z'

Fig.

myf +

?>/

8.

/ (cos 0' +

sin 9').

OP

and PP' (Art. 115),


represents the sum of
follows that PP' represents the increment of z and if *' = s +
Since

OP'

//

may

it

A,

be written in the form

sn
where p = PP', and $

The

the angle PP' makes with OX.


or /-r; the
variation of the modulus of z is OP'is

OP

variation of the amplitude of 2


of g itself is A or

/>

(cos

returns to

its

P'OP

to describe a closed. curve.

When

original position P, the modulus takes again its


and the amplitude takes its original value if the

original value
point O is exterior to the curve, or
;

or 9'- 0; the variation

sin ^), as just explained.

Let the point be supposed


it

is

is

increased

by

2ir if

is

interior to the curve.

If the complex variable describes the same line in two opposite directions, the variations of its amplitude are equal and of

From this we
is nothing.
opposite signs, i.e. the total variation
can derive a property of the variation of the amplitude of the
in our
complex variable, which will be found of importance

succeeding investigations.
Let a plane area be divided into any number of parts by lines
the amplitude
JBD, AF, EC, &c. (fig. 9) ; then the variation of

Continuity of a Function of the Complex Variable. 255


relatively to the perimeter

of the whole area

is

equal to the sum

to
oj*its variations relatively

the perimeters of the partial

the areas being


supposed to be described by
the variable moving in the
areas

all

same sense. This is evident;


for

when

--

the point is made


the partial

to describe all

the same

in

areas

sense,

Fig. 9.

each of the internal dividing lines will be described twice, the

two descriptions being

in opposite directions

perimeter will be described once

and the external

hence the total variation of

the amplitude relatively to the dividing lines vanishes, and the


variation relatively to the external perimeter alone remains.

AFD in

the figure.
When
the point describes these areas in the sense indicated by the
vanishes.
arrows, the total variation relatively to the line

Take, for example, the areas

ABF,

AF

119.

Continuity of a

Variable.
fixed value s

Function of the

Complex.

Suppose the complex variable s, starting from a


to receive a small increment h s pfcos^-t /sin $);

we have

then, if,/'*) be the given function, replacing x


the expansion of Art. 6,

/(*) = /(*

A)

/()

/'( Sc) )

h +

+
'-J^A'
1
.

and the increment

of /(s), being equal to/(s

In this expression the

coefficients of the

by

z in

&o.,

/o

//)

/(c

),

is

powers of h are

all

complex expressions of the usual form ; and if their moduli be


2
3
a, 6, c 9 &c., the moduli of the successive terms are a/>, fy> e/> &o.;
,

and

since,

sum

of the moduli,

of/()

by Art. 115, the modulus

is less

it

follows that the

of a

4-

is less

than the

modulus of the increment

than
ap

sum

1
bp* op + &o.

Complex Numbers and

256

Now

the

Complex Variable.

be assigned to p (Art. 5) such that for it


or any smaller value the value of this expression will be less
than any assigned quantity. It follows that to an infinitely
a value

may

small variation of the complex variable

(viz.

one whose modulus

small) corresponds an infinitely small variation of


the function; in other words, the function varies continuously at
is infinitely

the

same time as the complex variable

itself.

Variation of the Amplitude of /() corresponding to tbe description of a small Closed Carve by the
Complex. War i able. Corresponding to a continuous series of
values of z we have a continuous series of values of /(*), which
120.

can be represented, like the values of


plane.

* itself,

by points

in a

We represent these series of points by two figures (fig. 10)

0'

Fig. 10.

by side, which, to avoid confusion, may be supposed to be


drawn on different planes.
To each point P, representing
x + iy, corresponds one determinate point P' representing f(z}.
side

P describes a continuous curve, P' describes also a continuous curve


and when P returns to its original position
after describing a closed curve, P returns also to its
original
When

position.

Our present object is to discuss the variation of the amplitude


of /(*) corresponding to the description of a small closed curve
by P.

Let

be any determinate point whose coordinates are

Variation of Amplitude
</o>

oases

*-

*o

#b

We

*y<>.

ofj(z\

257

fyc.

divide the discussion into two

When

(1).

iy* is

-f

different

is

(2).

When

x +

not a root of f(z)

from

0,

1.0.

when/(a

zero.

i^ is a root of /(a)

0, or

/(s

0.

In the

first case, to the point


corresponds a point A'
is different from zero.
value
A'
the
of
and
0'
/'(a ),
representing
=
=
Let a s -f /*, where h p (cos -f i sin 0) and suppose P, which
Let P'
represents a, to describe a small closed curve round A.
(1).

AP'

represents the increment of /(a) corresponding to the increment ^4P of a. By (he previous Article
it
appears that values so small may be assigned to p, that the

represent /(a)

modulus

then

of the increment of /'(*)

namely

A'F may be

always
than the assigned quantity 0'A\ hence P may be supposed
describe round A a closed curve so small that the correspond,

less

to

ing closed curve described by P' will be exterior to 0'.


follows, by Art. 118, that corrcs]>ondiny to the description by

It

of

a small closed curve, which does not contain a point satisfying the

equation f(z)

0,

the total variation of the amplitude of f(z)

is

nothing.
(2).

In the second

=
equation /(a)

suppose

case,

repeated

ir

times, and

H-

is

///

a root of the

let

/()-(*-s,,r^(*);
then
/(a)

= h m iL(z} = p w
*

(cos mty

-f

sin ?m/>)i/> (s).

In this case Q'A.' -

round A, P' returns

and when

P describes

to its original position,

a closed curve

and the amplitude

of /(a) will be increased by a multiple of 2rr, which may be


From the above equation we have
determined as follows
:

amp.

f (a)

m$

+ amp.

i//

(a)

and the increment of amp. /(a) will be obtained by adding the


Now the latter
(a).
incremen^ of m<f> to the increment of amp.
i/>

Complex Numbers and

258

increment

is

nothing by

(1),

the

Complex Variable.

since the curve described

by P

and since the increbe supposed to contain no root of ^ (*)


of $ is 2ir in one revolution of P, the increment of fjty
;

ment
is

when

It follows that

2mir.

describes a small closed curve

containing a root of the equation f(z)

amplitude of f(z)
121.

is

two opposite

responding line in

0, repeated

times, the

increased by 2ni7r.

Cauchy's Theorem.

line in a plane in

its

When

describes

-s

the same

directions, f(z) describes the cor-

plane in two opposite directions, and the

amp.f(z) undergoes equal and opposite variations.

It follows

any plane area be divided into parts, as in Art. 118, the


variation of the amp. fix) corresponding to the description iu

that, if

the same sense by

% of all

the partial areas,

is

equal to the variaby % of the

tion of amp. f(z) corresponding to the description

Now let any closed perimeter in the


external perimeter only.
be described and suppose, in the first place, that it
plane

XY

contains no point which satisfies the equation / (z) = 0.


It can
be broken up into a number of small areas, with respect to each
of which the conclusions of (1), Art. 120, hold

been just proved,

it

and by what has

follows that the

variation ofamp.f(z) corresponding


by z of the closed

to the description

perimeter

is

nothing.

Suppose, in

th e second place, that t he closed peri-

meter contains a point which is a root


of the equation f(z) =
repeated m
Let a small closed curve
times.

PQRS

be

described

point.

The

variation of amp.

tion

by

z of the

variations

round

corresponding

ABCPSR, CDAltQP,

'

f (z) corresponding to the descrip-

whole perimeter,
to

Fi

this

the

PQ21S.

is

equal to the

sum

of its

of

the

areas

description

The two former

variations

vanish by what is above proved ; and the latter is, by (2),


Art. 120, equal to Zm-rr.
The total variation, therefore, of f(z)
is

2w7r.

Similarly,

if

the

area includes

additiohal points

Number of Roots of the General

259

Equation.

which correspond to roots repeated m', ;//', &o., times, the


Hence we derive the
total variation = 2(m + m' + m" + &c.) TT.
following theorem due to Caucliy

The number of

obtained by dividing by 2ir the total variation of the

ix

plane area,

amplitude of

this polynomial

tion by the complex variable

corresponding

of

by means

are enabled

to

the perimeter

Number of Hoot*

122.

We

of any polynomial, comprised within a given

roots

the complete descrip-

of the area.

of the General Equation.

of the principles established in the

preceding Articles to prove the theorem contained in Arts. 15


and 16 namely, Every rational and integral equation of the n th
;

degree has n roots real or imaginary.

Let
/(*)

ft

rtis

n~l

fljs"'

# n _iS

+ an

Without making any


be a rational and integral function of z.
the
of
roots
further than
existence
of /(*) =
supposition as to
that

cannot vanish for any infinite values of the variable,


z to describe in its plane a circle so large that

/ (a)

we can suppose

no root exists outside of


= z
/(*)

s',

whose modulus

4-

|f/

- zn

<i>

it.

(V),

If,

then,

a& + a&* +
where

a n z'}

-,

the reciprocal of the modulus of

is

will

*,

Jesmibe a small circle containing a portion of the plane corresponding to the part outside of the circle described by z ; and
uo root of

<p

will

(z')

Hence, corresponding
the variation of

amp.

be included within

f (z')

0,

variation of amp. f(z]

md
9 is

if

r (cos

)f

sin 9)

whole

circle

by

s,

and therefore

variation of amp. z n

or z

= r" (cos

nti

-f

sin w0),

2?r, and therefore amp. z is increased by 2u7r.


from Cauohy's theorem, Art. 121, that the number

increased by
It follows

this small circle.

to the description of the

roots comprised within the circle described

by

z, i. e.

the total

Complex Numbers and

260
number
is

of roots of the

Complex Variable.

the

equation/ (z)

0, is n

and the theorem

proved.

The proposition whose proof was deferred in Art. 15 is thus


shown to be an immediate consequence of Cauchy's theorem,
which

may

therefore be regarded as the fundamental proposiIt is proper (o observe, how-

tion of the theory of equations.

theorem of Art. 15, viz., that every numerical


equation has a numerical root, can be proved directly, and
independently of Cauchy's theorem, by aid of the principles
contained in Art. 119 and the preceding Articles, as we proceed
ever, that the

now

to show.

123.

Second Proof of Fundamental Theorem.

If

there be no value of* which rnakes/(s) vanish; and


let the value
,
represented by A, fig. 10, correspond to the
possible let

to the origin 0'.


It is
nearest possible position, A\ of
a
show
that
such
direction
to
be
to
the
may
proposed
given

increment h as to bring P' into a position nearer to the origin


have the following expansion (Art. 119)
than A'.

We

///

/(*o 4 *)

/'(s

does not vanish

derived functions, /'(

),

which does not vanish

+/()

Bj hypothesis

f'**Qj

&c.,

be/w (s

h 4

may
),

''

do

V'

but one or more of the

Let the

so.

and

let

(cos

am 4

first

of these

us suppose
i sin

a,,,),

with corresponding expressions for the coefficients which follow.'


w
Collecting all the terms which contain powers of h beyond A
into one

complex expression, we

4 A)

=/(s

may

write

4 n m p m {cos (m$ 4 a m ) 4 t sin (m$ + a


4 (cos ^ + i sin ),
IUL

where, by the proposition of Art. 115,

261

Second Proof of Fundamental Theorem.

from the theorem of Art. 5 that such a


m
Now the direcvalue may be given to p as to make p < ti m p
tion of the increment h can be so selected, viz. from the equation
It

is

easily inferred

in$ +

X'tfA' +

TT

(fig.

10), as to bring P', in virtue of the

second expression in the value of f(z -f A), through a distance


m nearer to the
Let S be the
origin in the direction AO'
fi m p
in
this
A'
P"
is
which
the
line
on
ff
to
way. The
brought
point
l)

+ //) is to move
=
P' from 8 to a point T at a distance 8 T p and whatever the
O'T
direction of this movement, i.e. whatever the amplitude
<
that
have
.4'
We
is
since 8T< SA'.
proved, therefore,
effect of the last expression in the value of f(z Q
;

&A

>

not the nearest possible position of P' with reference to the


and in the same manner it may be shown that no
origin

is

other value different from zero can be the least possible value
of the modulus of /(*).

In the proof here given it is only shown that the equation


must have a root, and the precise number of roots is not determined, as it is in the proof derived from Cauchy's theorem
but when it is proved that one root at least must exist, the
;

proof can be easily completed by the method of Art. 16.


It is important to observe that when /'(* ) does not vanish,
the limiting value of the ratio of the
for any particular point

increment of/(s

to// is

the constant ./*() ^fii (cos

It is easily inferred that the

ui

-t-

/sin en).

two increments are inclined

at a

constant angle, and their moduli are in a constant ratio.


This
is usually expressed by
that
the
described
figures
by
-saying

and P' are similar in their infinitely small parts.


The student is referred to Note
at the end of the volume

some further observations on the subject of this Article.


124. Determination of Complex. Numerical Roots.
Solution of the Cubic. Little attention has been given by
for

writers

on the Theory of Equations to the actual determination

the complex numerical roots of equations ; nor is it easy to


jive any account suitable to an elementary text-book of general
methods in existence for this purpose. Theoretically the problem
:>f

Complex Numbers and

262

presents no difficulty

the

Complex Variable.

for if the real

and imaginary parts of

iy) be equated separately to zero, and from the two resulting equations one of the variables eliminated, an equation
is obtained from which a real value of the remaining one can

f(x +

be calculated by Homer's process.

method

that this

We

is

It will be found, however,

of little practical value.*

and the following

confine ourselves in this

Articles to

cubic and biquadratic equations with real numerical coefficients,


and exhibit the calculation in these instances in what appears to

be the simplest form for practical purposes.


/(a?)

a?

+ pop

qx +

be proposed for solution. The roots


h - k, of which a is real.
a, h + k,

Let the equation

may

be assumed to be

The character of the

roots will appear in the process of calculation;


k being detei mined from its square, which may turn out to be
No preliminary analysis of the
either positive or negative.

remaining

k be substituted for ^, and the


If A
and odd powers of k equated separately to
in Ex. 26, p. 152, we find immediately the equation
- k* =
= 3A 4- 2p/ + q.
f(h)
is

equation

sums

necessary.

-I-

of the even

zero, as

We
the

get also, by the elimination of A, a cubic equation for


determination of h
but there will be no occasion to
;

best got from the relation


2A = -/>, a having been calculated in the first instance in
the usual way by Homer's method.

form

this

equation,

since

is

-f

The student desirous of information as to the attempts of mathematicians in


the direction of the calculation of imaginary roots of numerical equations may
refer to the following ^ orks :
Lagrange's Traite de la Resolution dez Equation*
numeriques

Murpliy's Theory of Algebraical Equations; Allgemeine Auffotung


by Simon Spitzer (Wien, 1851); Die Aufl'owng der hoheren

der Zahlen-Gleichungen,

numerischen Gleichungcn, hy P. C. Jelinek (Leipzig, 1865)

method for calcu-

by Emory M'Clintock (American


and 2)
and Methode pratique pour la

lating simultaneously all the Roots of an Equation,

Journal of Mathetnatict, vol. xvii., Nos.

Resolution numerique complete des Equations algebrique& ou tramcenda*te& by


y

Carvallo (Paris, 1896).

M. E.

263

Examples.

It will be necessary finally to calculate A, and with it the


remaining two roots, whether real or imaginary. For this

purpose the following mode of procedure will he found convenient


The value of S/' (a) in terms of the coefficients is
:

p*

30, viz.,

/'(a)

+/'(A +

/' ( h +

also

+/'(*-*) =

*)

+ /' (A

A)

A)

2
jt?

3?;

= 2/' (A) + 6k2

whence immediately,
/'(o)

+ 4*'=jt>-3 ?

from which A can be determined


the numerical value of

f (a)

last coefficient in the final

process already completed.

witli

very

little

labour, since

can be written down from the second

transformation in the work of Home r's


The character of the remaining two

depend on the sign of the number so found and the


roots themselves will be determined by taking the positive and
negative square roots of this number.

roots will

EXAMPLES.
1.

Solve the equation


3

tf

First

the

calculate

real

-f

positive

Homer's method, and obtaining


coefficients

2#2

23#

root,

70

0.

completing four transformations by

for the final trans formed equation the following

1,

-44341896.

7G609868,

17402,

Remembering that the roots have been three times multiplied by 10, we find
the values of /(a) and /"() by cutting off nine figures from the right in the former
It is well to carry
case, and six in the latter, and supplying the decimal point.
the approximation a couple of steps further by the contracted method, and thus get
a more accurate value of /'(a).
find, in this way,

We

/'(a)

Subtracting this

number from;? 2

4* 2

76-6286.

3#,

which

is

equal to 73,

we

find

3-6286.

tfcis is negative, we have proved that the remaining roots are imaginary.
the ascertained value of a, viz. 5*13457, the value of h is found immediately

Since

From

Complex Numbers and the Complex Variable,

264
to

be

3-6672, and dividing 3-6286

finally the

two complex

by

4,

and taking

roots of the equation as follows

its

we have

square root,

0-9624 -v/^T.

3-5672

Solve completely Newton's cubic (see Art. 107), viz.,

2.

;r

2# - 6 -

0.

Completing four transformations by Horner, and proceeding as


- 2-09466, and
example, we find a
/'(a)

- 11-16078;

*-

hence

the former

ii>

1-290195,

and the remaining two roots (thus proved imaginary) are found to be

- 1-04727 + 1-13594 x/ 1
Find the remaining two roots

3.

x-

We

find

/'(a)

64-0841,

k*

of the

x-

= -

example of Art. 109,

x - 100

-f

4-0646

p. 231, viz.,

0.

and the required

16-52102,

2-6322

1.

roots are

\/^T.

Solve ihe equation

Dividing by 20, ana applying Homer's process to find the root of the equa2 -^
s -15 =
1 2.r
and 1, we find a = 0-4460366, and
.r
lying between

tion

/'(a)

- 0-47364.

We
4#

henr-e

k2

'47841,

have therefore

= pi -

3q -/'(a)

1-44

+ 0*47364

and the remaining two roots are

real.

We

find

-37698

and adding and subtracting &, the other roots are found to be 1*06865 and
- 0-31469 id. Ex. 15, p. 246).
5.

Solve completely Lagrange's cubic

&-

1x + 7

0.

Change the signs of all the roots, and calculate the positive root a betweer
3 and 4 of the transformed equation f(x)
0, thus obtaining a = 3-0489173, and
hence k 2 = -0281575, and Jc = -1678. Also h = - 1-624458
/(a) = 20-88737
;

whence the values of h

-f

k and h -

Jc ;

and changing the signs of

all

the roots

thus found, the roots of the given equation are

3-0489,

The examples given

1-3666,

1-6922.

are sufficient to

(Of.

Ex.

1,

Art. 111.)

show in what way

this

be used to solve a given numerical cubic, without


any previous examination of the character of its roots. The
process

amount

may
of

work required

to decide in this

way whether

the two

265

Solution of the Biquadratic.

remaining roots are real or imaginary is usually very little


greater than is required in the application of Sturm's theorem;

and the additional labour necessary


of the roots

We

extremely small.

is

for the actual determination

proceed

now

to biquad-

ratic equations.

1^5.

When

Solution of the Biquadratic.

a biquadcan be solved in a

has real roots (two or lour), it


to that employed in the preceding Article.

ratic equation

manner analogous

In some examples the existence of a real root can be at once


recognized and when such is the case, the following process for
the complete solution of the equation can be used with advan;

Lot the proposed equation be

tage.

and

---

/*'/,

+ px*

.r

-f

qx

-f

rx +

0,

the remaining two roots

be reprek, no assumption being made as to the


Let a and j3 be both calculated
character of the latter pair.
its real roots a,

/3

may

sented by h + k and h -

by Horner's process, and the numerical values of /'( a ) and


f (fi) determined at the same time, as in the preceding Article.
Now, if h + k be substituted for x in /^), and the method of
f

solution of Ex. 26, p. 152, employed,

- 6/

f (a) f

(/3)

JTT

Again, we have, as

"

+ /'

is

we

find,

8
8
- 4A +3p/i +

fy/i

without

difficulty,

+r

easily proved,

(h +

k)

+/' (A -

*)

= -

;>

4 4

W - 8r,

and
/'(A

4-

-f

k)

f (h -

7c)

r 2f'(h) 4 f"' (A) k\

whence immediately
-

4#(4A +

p)

/'(a)

+/

(/3)

+ p s - 4pj +

8r.

This formula oan be used for the calculation of A3 the value


of A having been previously ascertained from the equation
,

1-

/3

4-

2A = ~jo by means of the calculated values of a and

The second

)3.

pair of roots will be real or imaginary according as


the resulting value of A2 is positive or negative.

Complex Numbers and

266

Complex Variable,

the

EXAMPLES.
Solve completely the equation

1.

- 3*3 + 7z 2 -

a,

l(te

therefore be a second,

we

which

is

found to

between

lie

0.

between

It is at once apparent that a real root exists

and

and

1.

There must

By Homer's process

2.

find

a
/'(a)

& =

0-107767,

= -8-59078,

1-923262,

= 12-09133;

/'(/3)

whence we have
/'(a)

z
+/'() + p - 4pq +

Also, from the values of

0, and^, h

a,

=-

8r

19*49945.

0-484485, and 4A

-f

p = - 1*06206;

therefore

_
~

19-49946^
1-OG206

It is now proved that the remaining two roots are imaginary, and their values
can be ascertained by calculating k from this formula. Logarithmic Tables will
The roots are found to be
assist in the calculation.

2-1424

0-4845
2.

Solve completely the equation of Ex.

&
We

- YLx

1.

2, Art. 110, via.,

0.

find

<*

a
/'(a)

0-59368,

= -

whence the pair of imaginary

j8

2-04727,

22-3180;

/\/3

11-1635,
roots

- 1-32048
3.

4-

\/ -

v/^T-

2-0039

Solve the equation


2* 4

13s 2

-i-

There must be two real roots: one


Divide by

2,

When
=

j8

(a)

19

is

a4

G-5:r 2

calculated in the usual

0, it is to

-f-

0.

positive,

and write the equation as follows


/(#)

of f(x)

10#-

5#

way by

and the other

first

9-5

/'(a)

changing the signs of the roots

last coefficient supplied

2-45733.
32-409,
.

6-473

Fl46?

and the imaginary roots


0-2866

by the

final

= - 3-03055,
-66-936;
/'()
ft

whence

~ ** "

1-0924

negative.

0.

be observed that, in order to get the value of

change the sign of the second


We find
in Horner's process.

(ft)

y7

/'(jS),

we must

transformation

267

Solution of Biquadratic.
4. Solve the equation

- 80*8 + 1998** - 14937s + 6000

There

lie

clearly a real root between


between 12 and 13 (see Ex. 4, Art. 93).
is

/'

(a)

and

18664,

/'

whence

1,

is

easily seen to

We find

0-35098,

=-

0.

and a second

(jB)

=
-

12-75644,

5286-7

413-3

...

63-785

The remaining two

roots therefore are real,

and are

found to he 32-0602 and

easily

34-8322.
All the root* of this equation have been calculated by Homer's method
by
Young (Analyst* and Solution of Cubic and Biquadratic Equations, pp. 216-221).
Our last two roots agree, to the number of places here given, with the values
arrived at

hy him.

126. Solution of

Biquadratic continued

When

the

roots of a biquadratic equation are all imaginary, the mode of


solution of the preceding Article of course fails.
In this case,

and

in general,

whatever be the nature of the

ing method may be uaed


its

Let the equation,


second term, be written in the form
f(x) a

rr* -f

qx*

+ rx +

roots, the followfirst

deprived of

0.

may be assumed to bo h k, - h //, no


made
as to their character, which will depend
assumption being
on the signs of k* and k' 2 when calculated.
Substituting h + k
for #, and
proceeding as before, we find
The

roots of this

__

+ qi

-h

r
>

57,

"/"'(*)

whence

- 4*1 =

4/<

+ 2q +

from which k can be found when h

is

known.

eliminated between the two equations of Ex. 26,


sextic in h reduces to the cubic

of whioft

4/i

is

a root.

When
p.

is

152, the

This cubic must have one positive

Complex Numbers and

268

Complex Variable.

the

the remaining two may be both positive, both negative^


or both imaginary, according to the nature of the roots of the

root

The equation

given biquadratic.

cubic (with roots multiplied

is,

in fact, Euler's reducing

4) for the biquadratic

by

under

consideration (see Ex. 4, p. 125).


Let the positive root of the
cubic be calculated by Homer's process (if the three are positive,

any one of them will do).


it h\ and the full solution
is

Thus 4A

determined, and from

is

of the proposed biquadratic equation

given by the two formulae

EXAMPLES.
1.

Give the complete solution of the equation

+ x + 10 =

0.

This equation

is used by Murphy (Theory of Equations, p. 126) to illustrate his


propo&ed method of determining the imaginary rodts of equations by means of
find readily the reducing cubic
recurring series.

We

3
y - 40y -

0,

3
and, by Horner'n process, the positive root 6-3370184: hence the value of A and
,

from

it

/*

We

1-2586.

find then

0'7945,

according as the positive or

fy

used. In either case the quantity under the square root


and
the
are therefore all imaginary.
roots
negative,
They are easily found to be

negative sign of

is

1-3362

1-2686
2.

Solve the equation


x*

This example
thungen, p. 16).

is

v/^L,
4- 9a?

treated

by Spitzer
The reducing cubic is

whose positive root

is

6tf

1-2686

1*1771

is

\/-T.

0.

(Allyemeine

Auflowng for Zahlen-Qlei-

found to be 0*51094249; hence

0-36740.

The

numerical value of r divided by h is found to be 16-7878 ; and whether h be taken


with positive or negative sign, the quantity under the square root is negative, and
therefore all the roots imaginary.

0-3674

0-6663

The four

V'-T,

roots are

0-3674

+ 2-9706 \/^l.

269

Examples.
3.

Solve the equation

* - 2s 3 -

by

7^ +

10*

-HO

0.

To remoye the second term, multiply the roots by 2, and then diminish
The reducing cubic of the transformed equation is easily found to be

roots

2
3
y - 68y

320y

- 256 =

0.

by 10, and find immediately that the transformed


equation has a root between 6 and 7, which is found by Homer's process to be
6*29838. Hence 4A 2 = 62-9838, and h =
3-968.
Whether A is taken positively
Divide the roots of this

or negatively,

found that the quantity under the square root is a positive


case.
We find 4 A 2 = 9*04840,

it is

number, and therefore all the loots are real in this


4#2 = 0-98400 hence k = + 1*604, k' = 0*496
;

two transformations which weiemade


four roots as follows

5, p.

4.

-2-236.

-0-732,

2-236,

the product of the factors x*

is

and & 2

it

is

2x

easily seen that

2 (compare also

208).

Solve the equation

# 4 - 7a s
This example

d leichtmgen,

-f

Tt- -

7a- 4

0.

by Jelinek (Die Aujtosung tier holier en numerischen


remove the second term, multiph the roots by 4, and

discussed

is

To

p. 29).

then diminish by

We find in this way


& - 182^ - 16242- - 3069 =

7.

whose reducing cubic

0,

is
3
3/

To

leeults, in this instance, can be readily verified, foi

the given function

Ex.

whence, taking account of the


removing the second term, we have the

2-732,

The

in

- 364y 2 4 45360^ - 2637376 =

0.

find the situation of the positive loot, it is well to divide the roots

when readily appears that the tiansformed equation lias a root between
8 =
205-91, and h
By Homer's process it is found to be 2'0591 whence 4//
it

When

by

100,

2 and 3.
f 7*17.

taken positively, the quantity under the square root is found to be


and uheu it is taken negatively, the quantity
positive; hence two real roots;
under the squaie root is negative, and gives a pair of imaginary roots. Taking
account of the two transformations employed to remove the second term, we find

is

the four roots of the proposed equation as follows


5-993,

1*091,

- 0*042

l'033v/~l.

6. Solve the equation


a?

- 80*8 +

1998a--

- 14937s + 5000 m

0.

We

solved in the preceding Article.


repeat
H Young's equation, already
solutioti here by the method of the present Article, in order that the student

This
its

is

270

Compkx Numbers and

may have

an opportunity of comparing the amount of labour required in the two

methods.

When

instance), or

the second term

when

the second term

Complex Variable.

the

removed

(as is

the case in the present

is

easily

is

already absent in an equation,

be found that the method of the present Article

is

it

will usually

the more expeditious of the two.

Diminishing the roots by twenty, we find

- 402s* + 983* + 25400 =

z*

whose reducing cubic

0,

is

- 804y 2 + 59764y - 966289 =

0.

We get, by Homer's process, 4/< 723-21038, and therefore h


The quantity under the square root is found to be positive whichever
2

taken, and for the four

roots

we have

we have

hence, adding 20 to each root,

- 3^ +

10000 =

0-18748

Solve completely the equation Ex.

The

76*

10-2609,

2, p.

- 4s3 - 3*

**

0.

-f

9-927

207, viz.

23

\/~T.

0.

roots are

2-0526,

3-7853,

0-9189

Solve the equation of Ex. 4, p. 212, viz.

*4
Multiply the roots by

3s3

surable root

= 180

-*3 -3*+

1-4545

hence A

it is

11

3 \/5.

found that the

The

four imaginary roots expressed as follows

Find the imaginary roots

0,

solution

When Homer's

latter equation has


is

method

a commen-

easily completed,

and the

of the equation of

**-

\/^~l.

and remove the second term.

4,

applied to the reducing cubic,

9.

32-0603.

34*8321,

roots are

9-8860,

is

</l '92090;

Solve completely the equation of Kx. 4, p. 234, viz.

The

8.

the four roots of the proposed equation as

0-3511,

**

7.

12-7665,
6.

is

the two formula)

- A i v/38-47390,

follows

13-4462.
sign of h

11727* + 40385

Ex.

14, p. 246, >iz.

0.

Am. -

12-4433

197596y/::

NOTES.
NOTE

A.

ALGEBRAIC SOLUTION OF KQl'ATIONS.

THE
and

solution of the quadratic equation was known to the Arabians,


found in the works of Mohammed Ben Musa and other writers

is

In a treatise on Algebra by Omar


Alkhayyami, which belongs probably to the middle of the eleventh
century, is found a classification of cubic equations, with methods of
The
geometrical construction, but no attempt at a general solution.
study of Algebra was introduced into Italy from the Arabian writers
by Leonardo of Pisa early in the thirteenth cent ury and for a long
period the Italians were the chief cultivators 01 the science. A work
styled UArte Maggiore, by Lucas Paciolus (known as Lucas de Burgo),
published in the ninth century.

was published

in 1494.

of cubic equations,

This writer adopts the Arabic classification

and pronounces their solution to be as impossible in

the existing state of the science as the quadrature of the circle.


At
the same time he signalizes this solution as the problem to which the
attention of mathematicians should be next directed in the development of the science. The solution of the equation a? -f mx = n was

by Scipio Eerreo ; but nothing more is known of his discovery


that
he imparted it to his pupil Plorido in the year 1505. The
than
attention of Tartaglia was directed to the problem in the year 1530, in
effected

consequence of a question proposed to him by Colla, whose solution


depended on that of a cubic of the form a? + px* = q. Florido, learning
that Tartaglia had obtained a solution of this equation, proclaimed his
of the solution of the form a? + mx = n.
Tartaglia,

own knowledge

doubting the truth of his statement, challenged him to a disputation

272

Notes.

and in the meantime himself discovered the solumx = n. This solution depends on assuming
x an expression *]t - 3Jw, consisting of the difference of two radi-

in the year 1585;

tion of Ferreo's form a? +


for

and, in fact, constitutes the solution usually known as Cardan's.


Tartaglia continued his labours, and discovered rules for the solution
cals

of the various

forms of cubics included under the classification of the

Arabic writers.

Cardan, anxious to obtain i knowledge of these rules,


After
applied to Tartaglia in the year 1539, but without success.

many

imparted to him a knowledge of these


from him, however, the most solemn and sacred prosecrecy.
Regardless of his promises, Cardan published in

solicitations Tartaglia

rules, receiving

mises of

1545 Tartaglia's rules in his great work styled Ar& Magna.


been the intention

own.

He commenced

in 1559, before he

As

of Tartaglia to publish hie rules in a

his work, therefore, contained no

came

work

the publication of this

had reached the consideration

It

work

had

of his

in 1556, but died

of cubic equations.

mention of his own

ruies, these

time to be regarded as the discovery of Cardan,


and to be called by his name.
The solution of equations of the fourth degree was the next
rules

in process of

problem to engage the attention

of Algebraists

the case of the cubic, the impulse

and here, as well

was given by

to the learned the solution of the equation x*

-f

6# 2

Colla,
4-

who

36 = 60#.

as in

proposed
Cardan

appears to have made attempts to obtain a formula for equations of


hut the discovery was reserved for his pupil Ferrari. The
this kind
method employed by Ferrari was a transformation of such a nature as
;

make both sides of the equation perfect squares, a new unknown


quantity being introduced which is itself determined by an equation
It is, in fact, virtually the method of Art. 63.
of the third degree.

to

This solution

is

who published it in
known as Simpson's,

sometimes ascribed to Bombelli,

his treatise on Algebra in 1579.

which was published much


tially different from that

The

solution

later (about 1740), is in no respect essen-

In the year 1637 appeared


treatise, containing many improvements in algebraical
science, the chief of which are his recognition of the negative and
" Eule of
His exSigns."
imaginary roots of equations, and his
of Ferrari.

Descartes'

pression of the biquadratic as the product of two quadratic factors,


although deducible immediately from Ferrari's form, was an important

contribution

to

the study of this quantic.

Euler's

Algebra was

273

Notes.
published in 1770.

His solution

of the biquadratic (see Art. 61)

is

important, inasmuch as it brings the treatment of this form into


harmony with that of the cubic by means of the assumed irrational

The methods of Descartes and Euler were the


attempts made to obtain a general algebraic solution of

form of the root.


result

of

Throughout the eighteenth century many mathematicians


occupied themselves with this problem but their labours were unsuccessful in the case of equations of a degree higher than the fourth.
In the solutions of the cubic and biquadratic obtainea by the older
equations.

we

observe two distinct methods in operation the first, illusthe


assumptions of Tartaglia and Euler, proceeding from an
by
assumed explicit irrational form of the root the other, seeking by the
analysts

trated

aid of a transformation of the given function to change its factorial


In Art. 55
character, so as to reduce it to a form readily resolvable.

these two methods are illustrated; together with a third, the concepwhich is to be traced to Vandermonde and Lagrange, who

tion of

published their researches about the same time, in the years 1770 and
The former of these writers was the first to indicate clearly
1771.
the necessary character of an algebraical solution of any equation,
viz. that it must, by the combination of radical signs involved in it,
represent any root indifferently when the symmetric functions of the
roots are substituted for the functions of the coefficients involved in

the formula (see Art. 101). His attempts to construct formulae of this
character were successful in the cases of the cubic and biquadratic,

but failed in the case

of the quintic.

Lagrange undertook a review of

the labours of his predecessors in the direction of the general solution


of equations, and traced all their results to one uniform principle. This
principle consists in reducing the solution of the given equation to
that of an equation of lower degree, whose roots are linear functions
of the roots of the given equation and of the roots of unity.
He shows
also that the reduction of a quintic cannot be effected iu this

equation on which

its

way, the

solution depends being of the sixth degree.

All attempts at the solution of equations of the fifth degree


having failed, it was natural that mathematicians should inquire

whether any such solution was possible at all. Demonstrations have


been given by Abel and Wantzel (see Serret's Cours cFAlgelre Superieure. Art. 516) of the impossibility of resolving algebraically equa-

tions unrettricted in form, of a degree higher than the fourth.

274

Notes.

transaandental solution, however, of the quintic has been given by


in a form involving elliptic integrals.
Among other

M. Hermite,

contributions to the discussion of the qumtic since the researches of

Lagrange, one of leading importance is its expression in a trinomial


form by means of the Tschirnhausen transformation. Tschirnhausen
himself had succeeded in the year 1683, by means of the assumption
2
in the reduction of the cubic and quartic, and had
4- Q,x + #
y

imagined that a similar process might be applied to the general equaThe reduction of the quintic to the trinomial form was published
tion.

by Mr. Jerrard in his Mathematical Researches, 1832-1835, and has


been pronounced by M. Hermite to be the most important advance
in the discussion of this quantic since Abel's demonstration of the

In a Paper published by the


impossibility of its solution by radicals.
Kev. Ilobert Harley in the Quarterly Journal of Mathematics, vol. vi.,
p. 38, it is shown that this reduction had been previously effected, in
Of equal importance
1786, by a Swedish mathematician named Bring.

by means of
which the quintic is expressed as the sum of three fifth powers a
form which gives gieat facility to the treatment of this quantic.
Other contributions which have been made in recent years towards the
discussion of quantics of the fifth and higher degrees have reference
For an account
chiefly to the invariants and covariants of these forms.
<*>
W**
of these researches, additional to what will be found in the second
volume of this work, the student is referred to Clebsch's Thtorie der
lindren algebraischen Formen, and to Salmon's Lessons Introductory to
the Modern
to Bring's reduction is Dr. Sylvester's transformation,

275

Notes.

NOTE

B.

SOLUTION OF NUMKKICAL EQUATIONS.

THE

first

attempt at a general solution by approximation of numewas published in the year 1600, by Vieta. Cardan

rical equations

had previously applied the rule


to the

"regula aurea")
method were of little value.

cubic;

of "false position" (called

by him

but the results obtained by this

It occurred to Yieta that a particular

numerical root of a given equation might be obtained by a process


analogous to the ordinary processes of extraction of square and cube

and he inquired in what way these known processes should be


modified in order to afford a root of an equation whose coefficients are

roots

given numbers.

Taking the equation

(x)

Q,

where Q

is

a given

number, and/(#) a polynomial containing different powers of x, with


numerical coefficients, Vieta showed that, by substituting in / (x] a

known approximate value

of

the root, another figure of the root

When this
(expressed as a decimal) might be obtained by division.
value was obtained, a repetition of the process furnished the next
It will be observed that the principle
figure of the root ; and so on.
of this

method

is

identical with the

main principle involved in the

methods of approximation of Newton and Homer (Arts. 107, 108).


All that has been added since Vieta's time to this mode of solution of
numerical equations
afford facility

is

the arrangement of the calculation so as to


of evolution of the root.

and security in the process

Hjow gcaat has been the improvement in this respect may be judged
by an observation in Montucia's ffistoire des Mathtmatiques, vol. i.,

of

where, speaking of Vieta's mode of approximation, the author


regards the calculation (performed by Wallis) of the root of a biquad-

p. 603,

ratic to eleven decimal places

labour.

The same

as a

calculation can

work

of the

most extravagant

now be conducted with

great ease

by anyone who has mastered Korner's process explained in the text.


Newton's method of approximation was published in 1669; but
before this period the method of Vieta had been employed and simAfter the period of
plified by Harriot, Oughtred, Fell, and others.
Newton, Simpson and the Bernoullis occupied themselves with the

276

Notes.

same problem. Daniel Bernoulli expressed a root of an equation u


the form of a recurring series, and a similar expression was given by
Euler; but both these methods of solution have been shown by
Lagrange to be in no respect essentially different from Newton's
solution (Traite de la Resolution des Equations numeriques}.
Up to
the period of Lagrange, therefore, there was in existence only one
distinct method of approximation to the root of a numerical equation
;

and this method, as finally perfected by Horner in 1819, remains at


the present time the best practical method yet discovered for this
purpose.

Lagrange, in the work above referred to, pointed out the defects
methods of Yieta and Newton. AVith reference to the former

in the

he observed that

it

required too

many

trials

and that

it

could not be

depended on, except when all the terms on the left-hand side of the
equation /(#) = Q were positive. As defects in Newton's method lie
signalized

first,

its failure to

give a commensurable root in finite

secondly, the insecurity of the process which leaves doubtful


the exactness of each fresh correction; and lastly, the failure of the

terms

method

an equation with roots nearly equal. The


" Etant
problem Lagrange proposed to himself was the following:
donnee une equation nuuierique sans aucune notion prealable de la
in the case of

grandeur ni de 1'espece de ses racines, trouvcr la valeur numerique


exacte, s'il est possible, ou aussi approchee qu'on voudra de chacune
de ses racines."
Before giving an account of his attempted solution of this problem,
necessary to review what had been already done in this direction,

it is

in addition to the

methods

of

approximation above described.

Harriot

discovered in 1631 the composition of an equation as a product of


Vieta
factors, and the relations between the roots and coefficients.

had already observed this relation in the case of a cubic but he


This
failed to draw the conclusion in its generality, as Harriot did.
;

discovery was important, for it led to the observation that any integer
must be a factor of the absolute term of an equation; and Newton's Method of Divisors for the determination of such roots was
root

Attention was next directed towards finding limits


of the roots, in order to diminish the labour necessary in applying the
a natural result.

method

of divisors as well as the

in existence.

methods of approximation previously

Descartes, as already remarked,

was the

first to

recog-

277

Note.
nise the negative

and imaginary roots of equations

commenced by him

as to the determination of the

and the inquiry


of real and

number

any given equation was continued by Xewton,


and
others.
DeGua,

of imaginary roots of
Stirling,

that, in order to arrive at a solution of the

Lagrauge observed

problem above stated, it was first necessary to determine the number


of the real roots of the given equation, and to separate them one from
For this purpose he proposed to employ the equation whose
another.
roots are the squares of the differences of the roots of the given equation.

Waring had

previously, in

1762, indicated this method

of

separating the roots but Lagrange observes (Equations numeriques,


Note iii.) that he was not aware of Waring' s researches when he
;

It is evident that when


this subject.
the equation of differences is formed, it is possible, by finding an
inferior limit to its positive roots, to obtain a number less than the

composed his own memoir on

least difference of the real roots of the given equation.


By substituting in succession numbers differing by this quantity, the real roots
When the roots are sepaof the given equation will be separated.
rated in this way, Lagrange proposed to determine each of them by

the method of continued fractions, explained in the text (Art. 112).


This mode of obtaining the roots escapes the objections above stated
to Newton's method, inasmuch as the amount of error in each successive approximation is

known

and when the root

is

commensurable,

the process ceases of itself, and the root is given in a finite form.

Lagrange gave methods

also of

obtaining the imaginary roots of

equations, and observed that if the equation had equal roots, they
could be obtained in the first instance by methods already in
existence.
Theoretically, therefore, Lagrange's solution of the problem

he proposed to himself
is

almost useless.

which

As

a practical method, however, it


The formation of the equation of differences for
is perfect.

equations of even the fourth degree

is

very laborious, and for equa-

becomes well-nigh impracticable. Even if


the more convenient modes of separating the roots discovered since
tions of higher degrees

Lagrange's time be taken in conjunction with the rest of his process,


still this process is open to the objection that it gives the root in
the form of a continued fraction, and that the labour of obtaining
it in thfs form is greater than the corresponding labour of obtaining it

278

Notes.

by Homer's process

in the

form

of a decimal.

also that the latter process, in the perfected

has brought
above stated.

it, is

free

from

It will be observed

form

to

which Horner

the objections to Newton's method

all

Since the period of Lagrange, the most important contributions to


the analysis of numerical equations, in addition to Homer's improvement of the methods of approximation of Yieta and Newton, arc those
of Fourier,

Budan, and Sturm. The researches of Budan were puband those of Fourier in 1831, after his death. There

lished in 1807

no doubt, however, that Fourier had discovered before the publication of Budan's work the theorem which is ascribed to them conjointly
is

in the text.

The

researches of

methods of separation

Sturm were published

in 1835.

The

of the roots

explained in Chapter X.

By
we have now

proposed by these writers are fully


a combination of these methods with

a solution of Lagrange' s problem far


simpler than that proposed by Lagrange himself. And it appears
impossible to reach much greater simplicity in this direction. In
that of Horner,

extracting a root of an equation, just as in extracting an ordinary


square or cube root, labour cannot be avoided and Homer's process
;

appears to reduce this labour to a minimum. The separation of the


roots also, especially when two or more are nearly equal, must remain

a work of more or less labour.


tion

by

This labour

may admit

of

some reduc-

the consideration of the functions of the coefficients which

play so important a part in the theory of the different quantics. If,


for example, the function? H> I, and 7J are calculated for a given
the character of the roots
quartic, it will be possible at once to tell
Mathematicians may also invent in process of time
(see Art. 68).

some mode of calculation applicable to numerical equations analogous


simple roots. But at the present
time the most perfect solution of Lagrange's problem is to be sought
in a combination of the methods of Sturm and Horner.
to the logarithmic calculation of

All that has been said applies only to the real roots of numerical
have referred, in a foot-note on p. 262, to the chief
equations.

We

works in which attempts have been made to give general methods of


calculation of the imaginary or complex roots; and in Ajts. 124,
125, we have shown how these roots may be calculated most expethird and fourth degrees with
ditiously in the case of equations of the
'

real numerical coefficients.

Notes.

NOTE

0.

THE PROPOSITION THAT EVERY EQUATION HAS


IT

is

what

A ROOT.

important to have a clear conception of what is proved, and


it is possible to prove, in connexion with the proposition dis-

cussed in Arts, 122, 123.

If in the equation

ti^r*

the coefficients

fli**""

+ ...+

a l9
a n are used as mere algebraical symbols
without any restriction that is to say, if they are not restricted to
denote either real numbers or complex numbers of the form treated
,

XII. then, with reference to such an equation it is not


proved, and there exists no proof, that every equation has a root.
The proposition which is capable of proof is that, in the case of any

in Chapter

rational integral equation of the n th degree,

whose

coefficients are all

numbers, there exist n complex numbers


which satisfy this equation so that, using the terms number and
numerical in the wide sense of Chapter XII., the proposition under

complex (including

real)

consideration might be more accurately stated in the form Every


numerical equation of the n th degree has n numerical roots.
As regards this proposition, there appears little doubt that the

proof is one founded on the treatment of


or
imaginary expressions
complex numbers of the kind considered in
XII.
first
idea
of the representation of complex numbers
The
Chapter
in
due
a
is
to
Argand, who in 1806 published anonyby points
plane

most direct and

scientific

mously in Paris a work

entitled JEssai sur une mani&re de representer

This
quantity imaginaires dans les constructions geomttriques.
writer some years later gave an account of his researches in Gergonne's

les

Annales.

Notwithstanding the publicity thus given by Argand to his


attracted but little notice, and appear to have been

new methods, they

discovered independently several years later by Warren in England


and Mourey in France. These ideas were developed by Gauss in his

works published in 1831 and by Cauchy, who applied them to the


proof of the important theorem of Art. 121. "With reference to the
;

280

Notes.

proposition now under discussion, the proof which we have given in


Art. 123 is a modification of a proof found in Argand's original
memoir, and reproduced by Cauchy in his Exercices & Analyse.

proof in many respects similar was given by Mourey.


Before the discovery of the geometrical treatment of complex
numbers, several mathematicians occupied themselves with the prob-

lem

of the nature

researches
riques.

is

The

An

of the roots of equations.

account of their

given by Lagrange in Note IX. of his Equations numtinquiries of these investigators,

among whom we may

mention D'Alembert, Descartes, Euler, Foncenex, and Laplace, reand the object in
ferred only to equations with rational coefficients
view was, assuming the existence of factors of the form x - a, x - ft,
&c., to show that the roots a, /?, &c., were all either real or imagi;

nary quantities of the type a + b J 1 ; in other words, that the


solution of an equation with real numerical coefficients cannot give rise
an imaginary root of any form except the known form a -f b J - 1,
For the proof of this proposition,
in which a and b are real quantities.
the method employed in general was to show that, in case of an
to

equation whose degree contained 2 in any power k, the possibility of


its having a real quadratic factor might be made to depend on the

an equation whose degree contained 2 in the power k - 1


only, and by this process to reduce the problem finally to depend on
the known principle that every equation of odd degree with real coefsolution of

Lagrange's own investigations on this subject,


given in Note-X. of the work above referred to, related, like those

ficients has a real root.

of his predecessors, to equations

founded ultimately
root in an equation

.on

with rational

coefficients,

and are

the same principle of the existence of a real

of v odd degree

with real

coefficients.

As
an

resting on the same basis, viz. the existence of a real root in


equation of odd degree, may be noticed two recently published

methods

of

considering this problem

Clifford (see his Mathematical Papers,

sophical Society's Proceedings, II., 1876),

>one
p. 20,

by the late Professor


and Cambridge Philo-

and the other by Mr. Malet

(Transactions of the

Royal Irish Academy, vol. xxvi., p. 453, 1878).


th
with
an
Starting
equation o*f the 2m degree, both writers employ

method of elimination to obtain an equation of the


on
whose solution the existence of a root of the
1)
is
proposed equation shown to depend and since the number m (2m - 1 )
Sylvester's dialytic

degree m(2m -

281

Notes.
contains the factor % $*<**
is

I****

^^

f.him

f^a number 2m. the problem


methods above mentioned,

reduced ultimately to depend, as in the

on the existence of a root in an equation of odd degree. The two


equations between which the elimination is supposed to be effected are
m and m - I and the only difference between the two
of the
degrees

of proof consists in the manner of arriving at these equations.


In Mr. Malet's method they are found by means of a simple transfor-

modes

mation

of the

by equating
polynomial

proposed equation while Professor Clifford obtains them


the coefficients of the remainder when the given
;

to zero
is

The general forme


divided by a real quadratic factor.
found among the Miscellaneous Examples

of these coefficients will be

appended to the chapter on Determinants in the second volume of th^


work and it will be readily observed that the elimination of /? from
the equations so obtained will furnish an equation in a of the degree
;

m(2t* - M.

fSee Kx. 38.

D. 64,

Vol. II.

INDEX TO VOLUME

I.

[The figures appended refer to the pages.]

Abel, 273.

Algebraical equations,

2,

Biauadratic

u a stormed

215.

r,

lo reciprocal, 135.

their solution, 105.

solved

solution of cubic, 108.

equation of differences, 142.


nature of roots, 144, 211.

of biquadratic, 121.
historical note, 271.

by symm.

functions, 139,

Bombelli, 272.

Alkhayyami, 271.

Bring, 274.

Amplitude of complex number, 249.

Budan, theorem

of, 189, 278.

variation of, 256.

Appi oximation to numerical root*


Kewton's, 225.

Homer's, 227.

Cardan, solution of cubic, 108.


his relations with Tartaglia, 272.

Cauchy, his theorem, 258.


Clifford, on the proposition that every

Lagrange's, 241.
Arabians, 271.

equation has a root, 280.

Argand, 279.

Colla, 271, 272.

Commensurable

roots

Complex numbers,

Ben Musa,
Binomial

271.

solution of, 90.

leading properties
solution

by

of,

92.

circular functions, 98.

Gauss' method of solving, 102.


Biquadratic, 73.
Euler's solution of, 121.
Ferrari's, 129.

Descajtes', 133.

theorem, 216.

graphic representation of, 249.


addition and subtraction of, 250.

coefficients, 68.

Binomial equations

22, 249.

multiplication

and division

of, 251,

other operations on, 252.

Complex

roots of numerical equations,

261.

determination of, for cubic, 262.


for biquadratic, 265, 267*

Complex

variable, 253.

continuity of function of, 256.

284

Index.

Continuity of rational integral function,


9.

Fourier: his theorem, 189,278.


applied to imaginary roots, 19<

Cube

roots of unity, 43.

corollaries from, 197.

Cubic, 71.

equation of differences, 81.


criterion of nature of roots, 84.

Cardan's solution
as difference of

of,

108.

Fundamental theorem, 259.


derived from Cauchy's, 260
second proof

of,

260.

historical note on, 279.

two cubes, 111.

solved by symmetric functions, 113.

homograpbic relation of

roots, 120.

Gauss, binomial equations, 102.

Graphic representation

of polynomial, 13.

Darwin, G. H., example solved, 246.

De Gua,

rule for imaginary roots, 197.

Derived functions,

of derived functions, 164.


of complex numbers, 249.

Greatheed

8.

on the biquadratic, 136.

grapbiu representation of, 154.


in terms of the roots, 157.

Descartes, rule of signs, 28, 30, 197.


solution of biquadratic, 133.

improvements in Algebra, 272.


Divisors, Newton's method of, 217.

Harley, 274.

Hermite, 274.

Homogeneous

products, 178.

Homographic transformation,

75.

relation of roots of cubic, 120.

Homer
Equal

roots, 25.

his

method

of solving

nume-

rical equations, 227.

condition for, in cubic, 84.

contraction of the process, 235.

in biquadratic, 144.

his process applied to cases where


roots are nearly equal, 238.

determination

by method
Equation

of,

159.

his

of divisors, 207.

of squared differences

improvements

in the solution of

numerical equations, 277.

of cubic, 81.
of general equation, 84.
of biquadratic, 142.

Equation whose roots are powers of


roots -of given equation, 78.

Euler: solution of biquadratic, 121.

Imaginary

roots, 21.

enter in pairs, 26.

determined for a cubic, 262.

and for a biquadratic, 265, 267.

his reducing cubic, 122.

Sturm's functions for his cubic, 248.


publication of his Algebra, 272.

Lagrange

on equation of differences,

142.
his approximation to roots

by con-

tinued fractions, 241.


Ferrari

solution of biquadratic, 129,

272.

Florido: 271.

on solution of equations, 273.


his treatise on Numerical Equation*
referred to, 143, 276*277, 280.

285

Index.
1

Quartic, see "Biquadratic.'

Leonardo, 271.
Limits of roots

definitions, 180.

Quintic

propositions on, 180, 181, 185.

Newton's method, 185.


inferior limits, and limits of nega-

special

248.
impossibility of its solution, 273.

tive roots, 186.

Quotient and remainder: when polynomial is divided by binomial, 10.

Limiting equations, 187.


Lucaa de Burgo, 271.

Beality of roots
Malet, on the proposition that every
equation has a root, 280.

Maxima and minima,

17, 155.

of Divisors, Newton's, 217.


Modulus, of complex numbers, 249.

Multiple roots, 25, 158, 159.


of,

of cubic, 84.

in general, 210.

tions, G2.

solution of reciprocal equations, CO.

reciprocal form,

theorem on sums of powers

bis

Robei ts

two

of roots, 165.

mode of approximation, 225.

of,

67.

on an equation derived from

cubics, 118.

identical relation, 153.

example on quartic and quintic, 213.

Numbers, complex, 22, 249.


Numerical equations, 2, 215.
commensurable roots

of biquadratic, 144.

divisors, 217.

Rollers theorem, 157.

Roots: theoiems relating

216.

to, 19.

imaginary, 21.

multiple roots of, 159, 222.

methods of approximation to

number
roots,

of, 22.

equal, 25.

225, 227, 241.

note on solution

to

on equation of squared differences

finding of limits, 185, 197.

method of

biquadiatic
135.

Reducing cubic, 122.

Removal of tc-ims,
Newton

of

transformation

by method of

222.

divisors,

of biquadratic, 144, 211.

Recipiocal roots, and reciprocal equa-

Method

determination

form solved, 104.

Sturm's remainders calculated for


one wanting second term, 247,

of,

275.

note on the fundamental proposition


as to roots, 279.

Descartes' rule for positive, 28.


for negative

and imaginary, 30.

relation of, to coefficients, 35.

cube roots of unity, 43.


Order, of symmetric functions, 173.

symmetric functions

of, 46, 165.

multiple, 158, 222.


limits of, ISO.

Polynomials

general properties, 5, 6.

change of form

continuity of, 9.

graphic representation

maxima and minima,


Puraer,

<Ai

separation of, 189.

commensurable, 216.

of, 8.

of, 13.

17.

Sturm's functions, 214.

approximation to, 225, 227, 241.


Cauchy's theorem on, 269.

complex roots determined, 261.


every equation has a root, 260, 279.

Index.

286
Eule

Symmetric functions

Descartes', of signs, 28, 197.

De

expressed by coefficients, 167.

Gua's, 197.

order and weight of, 53, 173.

of the double sign, 197.

calculation of, 48, 174.

Salmon, 274.
Scipio Ferreo, 271.

Tabulation of functions, 12.

Separation of roots, 189.

Tartaglia, 271.

Serret, 273, 274.

Transformation

of equations, 60.

Simpson, 272.

of cubic, 71.

Special roots of binomial equations, 95.

of biquadratic, 73.

Sturm:

homographic, 75.

his theorem, 198.

'

for equal roots, 203.

by symmetric functions,

application of theorem, 206.

in general, 80.

exercises

on

theorem,

213,

76.

214,

248.

Sums

of powers of roots

Yandermonde, 273.
:

Variable, change of form of polynomial


with change of variable-, 8.

Newton's theorem on, 165.


in terms of coefficients, 169.

complex, 253.

coefficients expressed by, 170.

Symmetric functions

Vieta, 276.

**

j^f

definitions, 46.

theorems relating

to, 53.

Wantzel, 273.

transformation by, 76.

Weight

K*S>

OF

1*

of

symmetric functions, 63, 171.