Académique Documents
Professionnel Documents
Culture Documents
43Mfr-.-<;
COLLEGE ALGEBRA
BY
BY
HENRY
B.
FINE
^^^oy
c t
f t
PROU.S.A.
PKEFACE
In this book
I
algebraic processes in as elementary and informal a manner as possible, but connectedly and rigorously, and to present the processes themselves in the form best adapted to the
purposes of practical reckoning. The book is meant to contain everything relating to algebra that a student is likely to need during his school and college
and the effort has been made to arrange this varied material in an order which will properly exhibit the logical
course,
its
me
two
parts,
a preliminary part devoted to the number system of algebra and a principal part devoted to algebra itself. "I have based my discussion of number on the notion of
cardinal
first
siderations of a theoretical nature in favor of this procedure But experience has coninto which I need not enter here.
vinced
me
is
method
the best.
that from a pedagogical point of view also this The meaning of the ordinal definition of
an irrational number, for example, can be made clear even to a young student, whereas any other real definition of such a
number
is
by
advanced students.
My
discussion of
elaborate.
But
IV
PREFACE
of statements which require demonstration. I hope the details of the discussion will interest the more thoughtful class of
all
is
the
that the general student need be asked to ordinal character of the real numbers
and of the relations of equality and inequality among them, and that for all numbers, real and complex, the fundamental
operations admit of definitions which conform to the
tative, associative,
commu-
and distributive laws. In the second or main part of the book I begin by observing that in algebra, where numbers are represented by letters, the laws just mentioned are essentially the definitions of the fundamental operations. These algebraic definitions are stated in detail, and from them the entire theory of the algebraic processes and the practical rules of reckoning are subsequently
derived deductively. I shall not attempt to describe this part of the book minutely. It will be found to differ in essential features from
the text-books in general use. I have carefully refrained from departing from accepted methods merely for the sake of novelty. But I have not hesitated to depart from these methods
when
this
seemed to me necessary
consistency, or when I saw an opportunity to simplify a matter of theory or practice. I have given little space to special devices either in the text or in the exercises. On the other
Thus, instead of relegating to the latter part of the book the method of undetermined coefficients, the principal method
of investigation in analysis, I have introduced
it
very early
and have subsequently employed it wherever this could be done to advantage. This has naturally affected the arrangement of topics. In particular I liave considered partial fractions in the chapter on fractions.
They belong
there logically,
in ele-
and when adequately treated, supply the best practice mentary reckoning that algebra affords.
PREFACE
mation and
Again, I have laid great stress upon the division transforits consequences, and in connection with it have
introduced the powerful method of synthetic division. This earlier chapters on equations will be found to contain a pretty full discussion of the reasoning on which the solution of
equations depends, a more systematic treatment than is customary of systems of equations which can be solved by aid of the quadratic, and a somewhat elaborate consideration of the graphs
of equations of the first and second degrees in two variables. The binomial theorem for positive integral exponents is
treated as a special case of continued multiplication, experience having convinced me that no other method serves so well to
convey to the student the meaning of this important theorem. I have introduced practice in the use of the general binomial theorem in the chapter on fractional exponents, but have
deferred the proof of the theorem itself, together with all that relates to the subject of infinite series, until near the end of the book.
more important properties of resultants. These subjects do not belong in an elementary course in algebra, but the college student who continues his mathematdiscussion of the
ical studies will
need them.
The
and of the chapter on properties of continuous functions with which the book ends. The ideas which underlie the first part of the book are
ters
on
infinite series
those of
Kowan Hamilton, Grassmann, Helmholtz, Dedekind, But I do not know that any one hitherto
has developed the doctrine of ordinal number from just the point of view I have taken, and in the same detail.
tions
to
In preparing the algebra itself I have profited by suggesfrom many books on the subject. I wish in particular
acknowledge
my
vi
PREFACE
The book has been several years in preparation. Every year since 1898 the publishers have done me the courtesy to issue for the use of the freshmen at Princeton a pamphlet
containing what at the time seemed to me the most satisfactory treatment of the more important parts of algebra. With the assistance of my colleagues, Mr. Eisenhart and Mr. Gillespie,
good and
of times.
new trial to select what had proved what had proved unsatisfactory. As a consequence, much of the book has been rewritten a number
I endeavored after each
to discard
;
No doubt subsequent experience will bring to light further possibilities of improvement but I have hopes that as the book stands it will serve to show that algebra is
many
not only more intelligible to the student, but also more interesting and stimulating, when due consideration is given to the
reasoning on which
its
processes depend.
Princeton University
June, 1905
HENRY
E.
FINE
K-^
CONTENTS
PAET FIRST NUMBEES
I.
Counting,
II.
III.
16 27
IV.
S9
70
V.
II.
III.
Preliminary Considerations The Fundamental Operations Simple Equations in one Unknown Letter
. >
79
93110-
IV.
.... ....
127155-
176 Factors of Rational Integral Expressions VII. Highest Common Factor and Lowest Common Multiple 196 213 S^III. Rational Fractions
VI.
IX.
Symmetric Functions
Evolution
Irrational
Functions.
245
252
XL
XII.
260
Radicals
and
Fractional
XIII.
271 Exponents 298Quadratic Equations XIV. Discussion of the Quadratic Equation. Maxima and 304. Minima XV. Equations of Higher Degree which can be solved by
Means of Quadratics
vii
309-
viii
COXTENTS
XVL
PAGE Simultaneous Equations which can be solved by Means of Quadratics 317
340
.
XVII. Inequalities XVIII. Indeterminate Equations of the First Degree XIX. Ratio and Proportion. Variation
342
347
354 357
362
364
374
XXIV. Logarithms XXV. Permutations and Combinations XXVI. The Multinomial Theorem
XXVII. Probability XXVIII. Mathematical Induction XXIX. Theory of Equations
303
408 409
424
425
483
492
. .
520
XXXIII. Operations with Infinite Series 639 XXXIV. The Binomial, Exponential, and Logarithmic Series 563 XXXV. Recurring Series 660 XXXVI. Infinite Products 564 XXXVII. Continued Fractions 666 XXXVIII. Properties of Continuous Functions 677
INDEX
691
A COLLEGE ALGEBRA
PAKT FIRST NUMBERS
I.
The fingers of a hand, a herd of cattle, the angular points of a polygon are examples of such groups of things. think of certain things as constituting a group, when we distinguish them from other things not individually but
We
make them
let
For convenience,
Equivalent groups.
of letters
ABC
and
DEF
all their elements in pairs by matching elements of the one with elements of the other, one element with one element. Thus, we may match A with D, B with E, and C with F.
Whenever
it is
possible to
match
shall
all
say that the groups are equivalent ; and the process of matching elements we shall call bringing the groups into a one-to-one relation, or a relation
1
we
of one-to-one corresjyondence.
".\:'
'.,
(|;p;iXi:GE
algebra
to
Thcoiem:^
ffrottp,
J/,
iwoypyjt^^^^
to
the
same third
For, by hypothesis,
we can
into one-to-one correspondence with the third But group. the two groups will then be in one-to-one correspondence
with each other, if we regard as mates every two of their elements which we have matched with the same element of the third group.
Cardinal number. may think of all possible groups of things as distributed into classes of equivalent groups, any two given groups belonging to the same class or to different
classes, according as
it
is,
We
or
is
them
same
class,
the groups
ABCD
and
The property which is common to all groups of one class, and which distinguishes the groups cf one class from those of another class, is the number of things in a group, or its cardinal
number. In other words,
in a group, or its cardinal number, is to the group itself and every be brought into one-to-07ie correspondence
is
common
may
:
The cardinal number of a group of things that property of the group which remains unchanged if we rearrange the things within the group, or replace them one by one by other things"; or again, "it is that property of the group which is independent of the character of the things themselves and of their arrangement within the group."
is
Or we may say
"
For rearranging the things or replacing them, one by one, by other things will merely transform the group into an equivalent group, 2. And a property which remains unchanged during all such changes in the group must be independent of the character of the things and of their arrangement.
when
say that a first group is ^ijmrt of a second group the elements of the first are some, but not all, of the elements of the second.
We
ABC
it
is
From
If the
second a
this definition
first
ptf^'H't
of three groups he a part of the second, and the of the third, then the first is also a part of
the third.
Finite and infinite, groups.
blage
is finite
when
it is
it
is
i7ifinite
when
equivalent to certain of
its
parts.*
Thus, the group ABC is finite; for it cannot be brought into one-toone correspondence with BC, or with any other of its parts.
But any never-ending sequence of marks or symbols, the never-ending for example, is an infinite assemblage. sequence of numerals 1, 2, 3, 4,
We
assemblage
up a one-to-one relation between the entire and that part of it which begins at 2, namely,
1, 2, 3, 4, 5,
between
(a)
and
by matching
1
2,3,4,5,6,...,
in (a)
(b)
with 2 in
(b),
we
2 in (a) with 3 in (b), and so on, there may choose to name in (a) a corresponding
(a) is
equivalent to
its
part
(b).
Therefore
(a)
Let
two
finite
groups.
1.
It
or or
2.
3.
N is
equivalent to a part of M.
* Of course we cannot actually take account of all the elements, one by one, an infinite group or assemblage, as it is more often called. We regard such an assemblage as defined when a law has been stated which enables us to say of every given thing whether it belongs to the assemblage or not.
of
4
In the
first
4,
A COLLEGE ALGEBRA
case
we
say that
or equal cardinal numbers in the second case, that the cardinal number of is less than that oi N; in the third,
number,
number
group
of
is
3f
is tlie
of letters c6c,
and
N the
group
de/pr,
then
JIf
equivalent to a part of iV, to the part de/, for example. Hence the cardinal number of is less than that of iV, and the cardinal
number
of
is
M.
9
no
Note.
definition of finite group, 7, that there is ambiguity about the relations "equal," "greater," and "less" as
It follows
from the
here defined.
Thus, the definition does not make it possible for the cardinal number to be at the same time equal to and less than that of iV, since this would mean that 3/ is equivalent to and also to a part of iV, therefore
of
that iV
is
equivalent to one of
its
parts, 3,
and
is infinite, 7.
10
Corollary.
be less
than
is
the second, a nd the second less than the third, then the first also less than the third.
For
M
11
cardinals,
is
denote any groujis of things of which these are the therefore equivalent to a part of N, and iV to a part of to a part of P, 3, 0. equivalent
if
M, N,
is
The system of cardinal numbers. By starting with a group which contains but a single element and repeatedly "adding"
one new thing,
:
we
list
of the cardinal
I,
which contains
2. The cardinal number of a group like II, obtained by adding a single element to a group of the first kind. 3. The cardinal number of a group like III, obtained by adding a single element to a group of the second kind.
4.
And
one,"
"
two,"
"
three,"
1, 2, 3,
'
number
12
group a finite cardinal, we make the following observations regarding the list of cardinals which has just
any
finite
been described.
First.
lent, 7
For the group is finite, since it has no part to wliich to be equivaand each subsequent group is finite, because a group obtained
;
to a finite
group
;
is itself finite.*
Thus,
II
is finite
is
III
is finite
because
II
is
and so
on.
Second.
Every
finite cardinal is
finite
contained in the
is
list.
number of some For, by definition, every But we can construct a group of marks finite group, as M. equivalent to any given finite group M, by making one mark for each object in And this group of marks must have a last mark, and therefore be M.
cardinal
the cardinal
III
I
included in the
therefore
itself,
list
it
and with
M, be
infinite, 7.
Third.
No two
groups
is
I,
II,
III,
are finite
and
it is
may
prove this as follows (G. Cantor, Math. Ann., Vol. 46, p. 490) and e a single thing, the group Me, obtained
:
by adding
For
If
let
G=
e to
M,
is
also finite.
G and
to
H are equivalent.
of its parts, 7.
'
which is matched with e in Me, and represent Let f donote the element of the rest of P by Pi. and e =/, we have 3/= Pi. since Then P\ f But this is impossible, since 3/ is finite and Pi is a part of M, 7. does contain p. (2) Suppose that is matched with e in Me, for then the rest of P, It cannot be tliat e in which is a part of M, would be equivalent to M. But suppose that c in is matched with some other element, as g in Me, and p is matched with/ in /'. that e in If P oe true on this hypotliesis, it must also he true if we reoombine the elements e. /, g so as to mat<-h r in P with r in Me, ami /' in P with g in Me. But, as 'just shown, we sliould then liave a pnvt of eiiuivaieut to M. Hence this liypothesis also is impossible
Me is not finite, it must be equivalent Let P denote this part, so that Me = P. not contain e. (1) Suppose that P does
some one
Me=
Me =
A COLLEGE ALGEBRA
THE NATURAL SCALE. EQUATIONS AND INEQUALITIES
13
call the signs 1, 2, 3, oi " " one," two," three," positive integers or natural numbers. Hence
We
their
names
"
A
14
natural number
scale.
is
The natural
an order
corresponding to that already given the cardinals which they represent, 11, we have the never-ending sequence of signa
1,2,3,4,5,...,
or "one," "two," "three," "four," "five," ., which the Jiatural scale, or the scale of the natural numbers.
we
call
15
Each sign in the scale indicates the number of the signs in that part of the scale which it terminates.
Thus, 4 indicates the number of the signs
of signs in turn,
1, 2, 3, is
1, 2, 3, 4.
I,
same as the number of groups the same as the number of marks in the
4
is
the
II,
lill,
last
group,
lill,
8.
And
16
so in general.
The
The natural
scale,
by
itself
considered, is merely an assemblage of different signs in which there is a first sign, namely 1 to this a definite next follow;
ing sign, namely 2 to this, in turn, a definite next following and so on without end. sign, namely 3
;
In other words, the natural scale is merely an assemblage of different signs which follow one another in a definite and
known
order,
and having a
first
Regarded from this point of view, the natural numbers themselves are with merely marks of order, namely of the order in which they occur
respect to time
is
recited.
17
common with
all
other assem-
blages whose elements as given us are arranged in a definite and known order, has the following properties :
1. may say of any two of its elements that the one " and the other " " follows," and these words precedes precede " and " follow " have the same meaning when applied to
We
"
any one pair of the elements as when applied to any other pair. If any two of the elements be given, we can always deter2. mine which precedes and which follows. If a, h, and c denote any three of the elements such that 3. a precedes h, and b precedes c, then a precedes c.
An assemblage may already possess these properties when given us, or we may have imposed them on it by some rule of arrangement of our own choosing. In either case we call the
assemblage an ordinal system.
Instances of the
of events in time
;
first
kind are
(1)
(2)
a sequence
a horizontal
line.
left to right
is
" " assemblage may also have coincident elements. Thus, group of events two or more may be simultaneous. We call such an assemblage ordinal when the relations 1, 2, 3 hold good among its non-coincident elements it being true of
An
18
in a
If
c.
a coincides with
b,
and
with
c,
then a coincides
then a precedes
with
5.
If a coincides with
b,
and
b precedes
c,
c.
by numbers indicate
It is
19
relations
of greater
and
less
among the
cardinal numbers.
is
greater wliose
number occurs
:
And the relation " if the first of three cardinals be less than the second, and the second less than the third, then the first is less than the third," is represented in the scale by the relation "if a precede b, and b precede c, then a
:
precedes
c."
8
In
fact,
A COLLEGE ALGEBRA
other method than this for comparing do not compare the cardinal numbers of groups of things the by the method of 8. On the contrary, we represent them by natural numbers, and infer which are greater and which less
cardinals.
directly,
We
appropriate from the relative order in which these natural numbers occur in the scale. The process causes us no conscious effort of thought, for the scale is so when any two of the natural numbers vividly impressed on our minds that, are mentioned, we instantly recognize which precedes and which follows. and B, that the population of A is Thus, if we are told of two cities, and that of B, 125,000, we immediately conclude that B has the
120,000,
we know
20
word
21
In what follows, the Numerical equations and inequalities. " number " will mean natural number, 13 and the
;
letters a,
h, c
will denote
we wish to " coincide " in the natural scale, ber, or read " a a
When
any such numbers. indicate that a and h denote the same num-
we employ
the equation
= b,
equals i."
22
to indicate that
we employ one
''
a <h, read
h
is less
"
;
> a,
read "
b is
greater than
<z."
23
Of course, strictly speaking, these words, "equal," <'less," and " greater," refer not to the signs a and b themselves, but " Thus, the phrase, a to the cardinals which they represent.
is less than ^," is merely an abbreviation for, "the cardinal which a represents is less than the cardinal which i represents." But all that the inequality a < b means /or the signs a and b themselves is that a precedes b in the scale. From 17, 18 and these Rules of equality and inequality.
24
Tf
If
=b a <
a
and
ft
and
and
3.
li a = b
= b < b <
r,
then a then a
c,
= <
c.
c.
c,
then a <c.
25
existing among the natural numbers, and with certain operations by which these numbers may be combined.
The operations of arithmetic have their origin in counting. To discover what the cardinal number of a given group of objects is, we count the group. The process is a very familiar one. We label one of the objects "one," another "two," and so on, until there are no
Counting.
26
objects left
"two,"
being careful to use these verbal signs "one," without omissions, in the order of their occurrence
in the scale, but selecting the objects themselves in any order whim or convenience and the sign or label with
;
is
what we
seek,
the name
of the
of the group itself. For owing to the ordinal character of the scale, this last sign indicates how many signs
number
have been used all told, 15, and therefore how many objects there are in the group, 8. Thus, the process of counting may be described as bringing the group counted into one-to-one correspondence, 2, with a
part of the natural scale namely, the part which begins at " one " and ends with the last number used in the count.
counting
Observe that the natural numbers serve a double purpose in (1) We use a certain group of them as mere counters
:
in carrying out the process, and (2) we used to record the result of the count.
employ the
last
one so
We
it is
we
:
This
may
be proved as follows
Theorem.
The
group of
objects is
27
which we
Suppose, for example, that the result of counting a certain group were 99 when the objects are selected in one order, P, but 97 when they are selected in another order, Q.
10
A COLLEGE ALGEBRA
consists of the first 97 objects in the ordeT
would then be equivalent to the entire group in the order Q, for, by hypothesis, both have been matched with the first 97 numbers of the natural scale, 3. But this is impossible, since it would make a part of
the group equivalent to the whole hypothesis, a finite group, 7.
;
is,
by
the
28
We may make
:
number
of a finite group,
namely
The cardinal number of a finite group of things is that property of the group because of which we shall arrive at the same natural number in whatever order we count the group.
This
led
if
is
we choose
make
number to which we are naturally the natural scale, defined as in 16, our starting
ADDITION
29
Definition of addition.
To add 3
number,
8,
to 5 is to find
what number
We may
find this
scale,
forward in the
beginning at
thus
6, 7, 8.
We
And
writing 5
+ 3 = 8.
add & to a is to find what number occupies the bth place after a in the natural scale. Since there is no last sign in the scale, this number may always be found. We call it the siim of a and b and represent it in terms of a and b by the expression a -\- b.
in general, to
30
Note. The process of finding a + 6 by counting forward in the scale corresponds step for step to that of adding to a group of a things the elements of a group of b things, one at a time. Hence (1) the result of the latter process is a group of a finite cardinals, so also does a +
+
b.
6 things, 8,
and
(2) if
a and b denote
See footnote,
p. 6.
11
and so
all
1,
-\-
2,
and
31
numbers
1,
-\-
2,
denotes
which follows a. Hence any given number after a may be expressed in the form a + d, where d denotes a definite natural number. The process. To add large numbers by counting would be very laborious. We therefore memorize sums of the smaller numbers (addition tables) and from these derive sums of the " laws " of addition larger numbers by applying the so-called
that portion of the scale
32
explained in the following sections. The laws of addition. Addition is a " commutative " and an
" associative " operation
;
33
that
is, it
law.
-{-
b
is
-\-
34
The
a
to b.
result of
adding b
a
^o
the
The
associative law.
(1) -{-
c)
= (a +
b)
-\- c,
The
result of first
to a,
is
adding
c ^o
OAid then
adding the
^o
sum so
a and
obtained
the
then adding c
to the
sum
so obtaiiied.
Note. In practice, we replace the expression (a + 6) + c by a + 6 + c, our understanding being tliat the expression a + & + c + represents
36
tlie
result of
adding 6 to
a, c to
the
sum
so obtained,
and so
on.
We may
lata
.3
37
The comm%itative
-\-
-\-
a.
2 and 2
are equal.
first
For
.3
2 represents the
number found by
counting
scale.
off
three
Thus,
(a)
1, 2,
1, 2,
.3,
4, 5,
1, 2.
.3,
(b)
But as there is a one-to-one relation between the groups of signs (a) and and every one-to-one relation is reciprocal, 2, we may interchange the roles of (a) and (b) that is, if we make (b) the group counted, (a) will
(b),
;
12
Hence finding 3
A COLLEGE ALGEBRA
+2
is
(b)
is
1,2,
1,2,3.
(c)
But as (b) and (c) consist of the same signs and differ only in the manner in which these signs are arranged, the results of counting them
are the same,
27
that
is,
3.
b.
Second.
For
The
associative
law
a +(b
-\-
c)
= (a
-\-
b)
-}- c.
+ b, and then to the cth sign after this, namely to {a + b) + c, we count b + c signs all told, and hence arrive at the (6 + c)th sign after a, namely at a + (& + c).
in counting to the 6th sign after a,
namely
to a
The notion of cardinal number is involved in the proofs just But addition may be defined and its laws established independently of this notion, as is shown in the footnote below.*
given.
* The Italian mathematician Peano has defined the system of natural numbers without using the notiou of cardinal number, by a set of "postulates" which we may state as follows where " number " means " natural number." 1. The sign 1 is a number. 2. To each number a there is a next following mmiber call it a +. 3. This number a + is never 1. 4. If a += 6 + tben a= &. 5. Every given number a is present in the sequence 1, 1 +, (1 +)+, .
The numerals 2, 3, are defined thus 2=l + ,3=2 + ,---. The sum a+b is to mean the number determined (because of 5) by the series of formulas a+l = a+,a + 2=(a+l)+, . The series of formulas just written is equivalent to the single formula 6. a + (6 + l)= (a + b) + l. From 6, by " mathematical induction," we may derive the laws of addition: 7. a+ (b + c)= {a + b) +c. 8. a + b=b + a. If 7 is true when c= k, it is also true when c= k+1. First. For, by 6 and 7, = [a + (6 + k)] + 1 a+[b+{k + l)]-^a + [(b + k) + 1] = [{a + b) + k] + l= {a + b) + {k + l). But, by 6, 7 is true when c= 1. Hence 7 is true when c= 2, when c= ^, .. when c= any number, by 5. Second. We first prove 8 for the particular case 8'. a + 1 = 1 + a. If 8' is true for a=k, then {* + 1) + 1 = (1 + A") + 1 = 1 + (A; + 1), by 6. Hence if 8' is true for a k, it is also true for a= k+ 1. for a= 3, Hence, since 8' is true for a = 1, it is true for a = 2, Finally, if 8 be true for b = k, it is true for b = k + 1. For, by 7 and 8', a + {k + I) = (a + k) + 1= \ + (a + k) = 1 + (^ + a) = (1 + A-) + a= ik+l) + a. for 6=3, ... Hence, since 8 is true (by 8') when b= 1 it is true for b=2,
..
.-.
.-.
See Stolz and Gmeinor, Thcoretiache Arifhnipfik, pp. 13 ff., and the references to Peano there given; al.so Hnntiiiitfou in liiilli't'ni of the American Mathematical Society, Vol. IX, p. 40. H. Grassmanu (Lehrbuch der Arithmetik) was the first to derive 7 and 8 from 6.
13
38
it
The sum of any finite number of numbers will be the same, whatever the order in xohich we arrange them, or whatever the
For
a4-fe
+ c + d=a + c + 6 + d. + c + = a + (6 + c) + = a + (c + 6) + d =a+c+&+
cZ
cZ.
(Z
35
34
35
39
Rules
of equality
and inequality for sums. First. From the and the ndes of 24, it follows that
If a If If
2.
3.
= b, a < a > b,
b,
then a
-{-
then a
then a
=b +c<b c > b
c
-\-
-\- c.
-{- c.
-\- c.
Here
If
1 is
We may
b,
and
(6
c)
d, 34, 35,
.-..
>b +
c.
From
4.
5.
6.
1, 2,
3
c
it
li a
If
-\-
=b
Z
then a
b.
If
b
+c< + +^>^+
then a
c,
then a <ib.
c,
then
a>
b.
Thus,
Fox-
it
-\-
c,
b.
otherwise
or else
we must have
either
a<b
and therefore a
c
;by
2),
a>b
and therefore a
Ob +
1, 2,
+ c<6 +
(by 3).
Third.
7.
It also follows If
from
c
3 that
8.
9.
If a
If a
= <
b, b,
and
and
and
>
b,
= d, c < d, c > d,
b
then a
then a
-^ c
-\-
then a
and
if c
b b
b
-\-
d. d.
d.
-{-
-\-
Thus,
if
a
c
=
b
b,
then a
d.
+c=
c,
then b
+c=
+ d.
Hence a
14
A COLLEGE ALGEBRA
MULTIPLICATION
40
Definition of multiplication.
To multiply a by
is a.
6 is to find the
sum
We
terms of a and
express
it
in
Hence, by definition,
41
-\-
to
b terms.
42 43
We
and a and
The process. To find products by repeated addition would be very laborious. We therefore memorize products of the smaller numbers (multiplication tables), and from these derive
tion
products of the larger numbers by aid of the laws of addiand the laws of multiplication explained in the following
sections.
44
The laws of multiplication. Multiplication, like addition, is a commutative and an associative operation, and it is " dis-
"
that
is, it
conforms to
45
ab
=
a,
ba.
The
result of multiplying
a.
by h
is
the
tiplying b by
Thus, 2
and 3 2
=
a
6.
46
The
associative law.
(l)c)
= (aF)
c,
The
is
the
same as
Thus, 2
(3
4)
12
24
and
(2
3)4
24.
36.
In practice
we
Compare
47
The
distributive law.
(b -f c)
=
is
ab
+
c,
ac,
The
result
ing a by the
sum
so
obtained,
the
15
the products
Thus, 3
(4
5)
27
and
12
15
27.
We may
48
The +
ac
distributive law: ab
{a + a a + a +
-\-
ac
(a
=
+
a(b
a
-\-
c).
(1)
41
For ah
= =
ioh terms)
a
c
+
c)
to (6
+
ah
terms
41
Hence
Thus, a{b
a(J)
-\-
+ )=
+
c)
-\-
ac
+
-\-
(2)
+c+
have
d)
a(b
+
be
ad
ah
+
b)
ac
c.
ad
by
(1)
and
35
We
also
ac
+
+
= (a +
-j
(3)
to c terms) {a
Eor ac
he
= =
{a
(a
+ +
+ 6) +
a
to c terms)
6')
(a
to
(h + h + terms =
b)c.
38
Second.
ba.
= = =
(1
l-h
6
1 + + l-b +6+
to a terms) b
-]
to a
terms
by
(3)
to a
terms
ba.
41
Third.
The
(ab) c
= ab + ab + = a{b + b +
to c
terms
terms)
41
to
{be).
by
(2)
and
41
General
theorem regarding
products.
finite
These
laws
can
be
49
number
of factors.
Thus,
The j)roduct of any finite number of factors is independent of the order in which the factors are multiplied together.
Rules
1.
of equality
and inequality
for products.
4. 5.
These are
be,
50
b.
If a If a If
a,
2.
3.
= b, < b, >
b-
then ac
then ac
= be.
<
be.
If ac If ac If
then a
then ac
> be.
6.
then a <,b.
then a ">
b.
16
Here
1 is
A COLLEGE ALGEBRA
obvious, since iia
b,
We may prove 3 as follows, and 2 similarly. n a>b, let a = b + d. Then ac = (b + d)c = bc + dc,
The rules 4, 5, 6 are the converses of the reasoning used in 39.
1, 2,
From
that
1, 2, 3,
in 39, it follows
bd.
If a
li a
If
b <b
and
c c
d,
then ac
and
and
a> b
< d, c> d,
then ac
then ac
II.
51
Subtraction. To subtract 3 from 5 is to find what occupies the 3d place before 5 in the natural scale.
number
We
writing 5
read "minus,"
2.
And,
in general, to subtract b
^th
from a
is
to find
what number
52
place before a. call this number the remainder obtained by subtracting b from a, and represent it in terms of a and b by the expression a b. We also call a the minuend and b the subtrahend. Addition and subtraction inverse operations. Clearly the third number before 5 is also the number from which 5 can be obtained by adding 3.
occupies the
We
we may describe the remainder a b either before a, or as the number from which a can be obtained by adding b, that is, as the number which is
And,
in general,
as the
bth.
number
53
(a
b)
-Jr
a.
17
Again, since saying that 7 occupies the 3d place after 4 is equivalent to saying that 4 occupies the 3d place before 7, we have 4 + 3 3 And so, in general, 4,
=
b
-\-
(a -\-b)
b=
a.
54
;
Since a
since a
-\-
b = a, b a,
^ 54,
and
55
53, addition
undoes subtraction.
2iiQ
We
inverse opera-
The complete scale. The natural scale does not fully meet the requirements of subtraction for this scale has first number, 1, and we cannot count backward beyond that number.
; di,
56
it is
2.
there
is
merely a system of signs arranged in a definite order, no reason why we should not extend it backward by
:
placing a new ordinal system of signs before it. therefore invent successively the signs place before 1 1, which we place before
We
0,
;
which we 2, which
and so on.
create the complete scale
-4,-3,-2,
-1,0,1,2,3,4,5,...,
last sign or "
to count
forward, to any extent whatsoever. Observe the symmetry of this scale with respect to the sign 0. As 3 is the third sign after 0, so 3 is the third sign
before
;
and so
of
in general.
Meaning
the
new numbers.
signs,
58
be said to have a cardinal meaning. Thus, counting backward from 3 corresponds to the operation of removing the elements of any group of 3 things, one at a time.
namely
0,
may
This operation
may
18
A COLLEGE ALGEBRA
been removed, and we may call the sign of the cardinal number of the resulting " group " of no elements. therefore as one of the natural numbers. often regard
We
But
1,
2,
3,
the other hand, all these new signs have the same ordinal character as the natural numbers. Every one of them occupies
On
a definite position in an ordinal system which includes the And we may consider it defined by this natural numbers also.
position precisely as we may consider each natural number defined by its position in the scale. regard this as a numbers. sufficient reason for calling the signs 1, 2, 3
We
59
2,
the old
3,
distinguish the new numbers 1, as a class from the old, we call them negative,
To
])ositive.
60
The numbers of both kinds, and 0, are called integers to distinguish them from other numbers to be considered later. Let a, b, c denote any numAlgebraic equality and inequality.
bers of the complete scale. According as a precedes, coincides b. b, a b, or a with, or follows b, we write a
<
>
61
Since by definition the complete scale is an ordinal system, thus. 17, the rules of 24 apply to it also
;
If a
<
b and b
is,
<
c,
then a
<
c.
62
When
< b,
that
when
than
b,
Observe that the words "less" and "greater," as thus used, mean this and nothing more. in the complete scale " 20 " 20 is less than - 18 " means 18." precedes merely Thus,
63
and use the symbol 3 to represent = 3. Similarly for any negative number. it, writing | 3| The numerical value of a positive number, or 0, is the number
itself.
Thus[3|
= 3.
19
64
to,
<, =,
or
>
-
|b|.
is
it is
numerically greater
3,
it
than
is
numerically
We
65
combined by addition, multiplication, and subtraction. We call these operations by the same names, and indicate them in the same way, as the operations with natm-al nimibers
to
Employing
a,
as in 60, to denote
plete scale, but a and h to denote natural may define these new operations as follows
Definitions of addition and subtraction.
1.
These are
after
a.
66
2.
3.
a
a
& is to
Z*
mean
number
is
to
mean
number before
the same
a.
4.
5.
+ and a a + ( V) is to a ( h) is to
are to
mean
number
as a as a
as
a.
&.
mean
the same
number
i.
In other words, adding a positive number h to any number is to mean, as heretofore, counting b places forward in the scale; subtracting it, counting b places backward: while adding and subtracting a negative number are to be equivalent respectively to subtracting and adding the corresponding positive number.
20
Thus, byl,
A COLLEGE ALGEBRA
3
2
by by
2,
2=
5
1,
since
1 is
the 2d
number
after
3. 2.
= ==-
3,
since
is
the 5th
number
before
4,
5,
by
- 5 + (- 6 - (-
2)
2)
2 2
==-
7 (by 2).
4 (by 1).
:
67
Definition of multiplication.
1.
This
is
and
J)
are to
mean
0.
2.
a (
and (
h)
is
a) b are to
mean
ab.
3.
a)
to
mean
ah.
factors, neither of
which
is
be positive or negative according as the factors have the sam,e or opposite signs. And in every case the numerical
is
x X
by
2 2
= - G, = 6.
and
= - 6.
68
the statements of ^%
Observe that and 67 are neither assumptions nor theorems requiring demonstration, but what we have called them definitions of new ojjerations.
(^()
The
it would be absurd to attempt to prove that 2 ( 3) = 2 3 with nothing to start from except the definition of multiplication of natural numbers, 40, for the obvious reason that 3 is not a natural number. The phrase "2 taken 3 times" is meaningless.
Thus,
But why should such operations be invented ? To make the negative numbers as serviceable as possible in our study of relations among numbers themselves and among things in the world about us.
The new operations have not been invented arbitrarily on the contrary, they are the natural extensions of the old operations to the new numbers.
;
tion as
In dealing with the natural numbers, we first defined addiand then showed that 2i counting forward process
21
the results of this process have two properties tvhich are independent of the values of the numbers added, namely
:
1.
+
:
b=^b
-\-
a.
2.
(1}
c)
= {a + b) +
c.
Similarly
properties
3.
we proved
4.
ah
= ba.
(be)
= (ai)
c.
5.
a(b
-\-
c)
ab
ac.
When we employ
letters to
1-5 become to all intents and purposes our icorking definitions of addition and multiplication for, of course, we cannot then actually carry out the processes of counting forward, and so on.
;
Clearly if corresponding operations with the new numbers are to be serviceable, these " definitions " 1-5 must apply
to
them
also.
:
And
the problem
To make such an extension of the meanings of addition, multiand subtraction that sums and jjroducts of any numbers of the complete scale may have the properties 1-5, and that
plication,
subtraction
Thus,
(1)
may
when we
forward, and subtracting it as counting backward, the old definitions of addition and subtraction.
(2)
From
if
But
~-b
+
6.
we must have
+ (
6)
This suggests the definition a 6. 6) = a ( (3) If our new addition and subtraction are, like the old, to be inverse ( 6) = a + 6. Operations, we must also have, as in 66, 5, a (4) Again, to retain the old connection between addition and multiplication, 41,
we must
(
have, as in
67, 2,
a) 6
= =
a a
+ () + to 6 terms a to ^^ terms =
ab.
(5)
If
=^
ba
is
6)
6)
= 6a =
to hold good,
a6, as in 67, 2,
22
(6)
A COLLEGE ALGEBRA
Similarly,
to
a terms
it (7) Finally,
follows from
(6)
that
is
= -a-0 =
0.
But
if
the distributive
6)
law
to
we must
also
have
(_a)(-6 +
we
69
0,
{ a)(b)
as ab, as in 67, 3.
The operations
ciative,
just defined conform to the commutative, assoIt remains to prove that the new-
with the laws which operations are in complete agreement suggested them.
To begin
with,
we have
a
+ (6 + c) = a + 5 + c, - + c) = a - J (Z.
c,
3i^l,-b=&-b + h = a,
as follows
from the definitions of addition and subtraction as counting forward and backward, by the reasoning in 37, 52.
I,
+ b = b + a.
d
First,
-a + b = b+(-a).
let
Forifa>6,
Then
b.
31,34
-a + b = -{d + b) + h
b
and
by
(2)
and
(3)
66, 4
by
(2)
Second,
YoY
II.
+ (
b)
=b
b)
-{-
i^-
{-
b)
= -{a +
-{b
( a). + a)= - b + {-
a),hy
(2)
and 66,4.
The
associative law, a
First, a
+ [^ + (let
+ (b + c) = (a + b) + = a + + (- c). t)]
/^
c.
rorif6>c,
c.
31,34
23
+
a.
[6
(-c)]
a a
+[d +
c c
+ {-c) =
a a
+ +
d,
d.
by
(3)
and
66,
Second, a
+ [(
I
Z<
b)
c^
= & + ( b) +
a
c.
just considered.
Third, a
^')
(-
c).
c)
and
66, 4, since
+ (
(&
+ c).
ba.
First,
{ a)b^b{
a).
For
45; 67,2
Second,
For
(-
a)
(-
b)
= (-
b)
(-a)(-6) =
The
a6
(- a). = &a=:(-6)(-a).
45; 67,3
IV.
= (ab)c.
a)
a)
First,
(-
a)
[(a)
b)
(6)
0)]
= [(= (=: c)
&)]
(ahc,
c).
For and
(-
[(-
(-
c)]
6c
c)
=-
46
67, 2,
ai-
(-
=- a&c.
67,2,3
may
The
distributive law, a (b
c)
ab
+ ac.
First,
a[b
+{ c)] = ab
c)]
-{-a
c).
.
For
[6
+ (-
Hence
to a terms + (- c)] + [6 + (- c)] + = 6 + 6 + -to a terms +( c) + ( c) + -to a terms ^ha + {-c)a. II and III 41 67, 2 a[6 + ( c)] = a6 + a( c) by III
[6
'
a)&
+(-
a)
(-
c).
24
70
A COLLEGE ALGEBRA
result. As has already been observed, 68, in arithmetic or algebra, the laws a -\- b b + a, and so on, are equivalent to definitions of addition and multiplication, even when the letters a, b, c denote natural numbers. And we
The general
literal
have
now shown
we may change
expression without affecting its value, whatever numbers of the complete scale the letters involved in the expression may
denote.
Thus, whether
a, 6, c,
(a
6) (c
d)
= (a + = ac +
6) c
hc
+ (a + 6) d + ad + bd.
71
We may
prove by
According as
so
is
< = <, =
,
or or
> b, >b+c
and conversely.
Hence
72
or
it is
numbers
alike that
is
added
to
both sides,
73
Rules of equality and inequality for products. Observe that changing the signs of any two numbers a and b reverses the
order in which they occur in the complete scale,
Thus, we have
57.
2.
<-
2,
but 3
>2
< 2,
but 5
>it
From
this fact
50
follows that
According as
so
is
<, =,
<, =,
>, =,
but
a(
Hence
c)
and conversely.
25
74
same
positive
number leaves its sense imchanged. But multiplying both sides of an inequality by thv same negative number changes its sense, from <.to'>, or vice versa.
From
cation
the
0,
first
of these rules
a
by
:
namely
1.
theorem
If
= b,
2.
If ac
= be,
= be. = b, unless e = 0.
75
The exceptional case under 2 should be carefully observed. = 3 0, of course it does not follow Thus, from the true equation, 2
that 2
3.
Zero produets.
If a product be
0, either a
0,
one of
its
factors must be
0.
76
Thus,
if
ab
or b
0.
For, since
is
also equal to 0,
we have
and therefore
Numerical values
ab
= a =
b,
0,
unless b
= 0.
75
77
of products.
is
= |-24|=24;
24.
The numerical value of a sum of Numerical values of sums. two numbers is the sum of their numerical values when the numbers are of like sign, but the numerical difference of these values when the numbers are of contrary sign.
Thus,
78
|-3 +
(-5)|
But
|2-|-(-5)|
= |-3I =
8.
26
A COLLEGE ALGEBRA
THE USE OF INTEGRAL NUMBERS
IN
MEASUREMENT
79
Measurement.
We
results of counting groups of distinct things, but also to indicate the results of measuring magnitudes, such as portions of
80
81
We
lar
If the
times exactly,
we
call this
number
its
measure.
In particular, we
call the
we
find that
it
contains the foot rule exactly three times, that is, its measure
we say
is 3.
that
82
The usefulness
scale,
of the natural
numbers
in
measurement
is
in the natural
83
We often Application of the negative numbers to measurement. have occasion to make measurements in opposite " directions " from some fixed " point of reference."
Thus, we measure time in years before and after the birth of Christ, longitude in degrees loest and east of Greenwich or Washington, temperature in degrees below and above zero.
then distinguish measurements made in the one in the other by the simple device of representing the one by positive numbers, the other by
direction
We may
84
-32-1
P-^
P-, P_i
P_4
Pi
Pj
Ps
P^"
DIVISION
AND FRACTIONS
27
Here the fixed point of reference, or origin, is 0, the unit are such that OPi, and the points P^, P3, , P_i, P_2, = P_iO = P_2^-i = OPi = P1P2 = A^3 = Above these points we have written in their proper order the numbers of the complete scale, so that comes over O. The distance of each point P from 0, that is, the length
is
of the segment OP, is then indicated by the numerical value of the number written above it and the direction of P from O
;
is
3 over
P_ 3
indicates that
0.
Moreover, the order in which the points occur on the line is indicated by the order in which the corresponding numbers
occur in the scale.
Points used to picture numbers.
to-one relation,
0,
1,
Inasmuch
as there
is
a one-
85
between the system of points , P_2, P_i, and the system of numbers , 2, 1, 0, Pi, P2, 2, , either system may be used to represent the other.
2,
III.
DIVISION
AND FRACTIONS
of division.
There are
tivo
operations to which
86
name
one may
division is applied in arithmetic and algebra. The be described as repeated subtraction, the other as the
inverse of vmltiplicatio7i. There is a case in which the coincide. call this the case of exact division.
two
87
We
To divide 7 by 3 two questions 1. What multiple of 3 must we subtract from 7 remainder which is less than 3 ?
:
in the first
to obtain a
2.
What
is
this remainder ?
28
A COLLEGE ALGEBRA
may find the answer to both questions by repeatedly 4 and 4 3 3 Thus, since 7 subtracting 3. 1, we must subtract 3 twice, or, what comes to the same thing, we must
We
subtract
3x2.
And
the remainder
is
is 1.
to addition.
are connected
by
7=3-2 +
And so in general, if a and b are any two natural numbers, to divide a by b, in the sense now under consideration, is to find two natural numbers, q and r, one of which may be 0, such that
88
89
^ bq
-^ r
and
<
b.
We
Note.
call
a the dividend,
and
the remainder.
90
When
if
88
may always
be found.
a < 6, we have q = and r = a. Thus, If a>6, it follows from 31, 35 that we can continue the sum until it either equals a or will become greater than a if we add +6+
another
b.
And
if
we
q
shall
Again,
when a and
numbers
and
satisfying 88 exists.
we should have
bq'
r',
q')
r'
r.
But
but b{q
would be numerically
b.
less
than
6,
q')
91
b,
Exact
as
division.
If the dividend a
6
when a
12 and
= 3,
92
qb
a.
DIVISION
AND FRACTIONS
29
93
Hence when a is exactly divisible by b, we may also define the quotient q as the yiumber which, inidtiplied byh, will produce a. In this case, furthermore, we may indicate the division
thus, a
-i-
94
b,
in
terms of a and
as well as qb
by
=j
= a.
95
that
^b
X b
a,
a7id a
x b
-=-
a.
hi
93 and
87 respectively.
Theorem
2.
96
divisor by the
same number
a
?
For
if
qb,
then
then
am =
q
bm.
50, 4G
That
IS, if
= -
b
=
bm
94 97
Theorem
with
3.
Exact
resjject to
that
is,
a,b =
I
+
c
b
>
a7id
,a
c
b
b
c
For
if
= =
qc,
and
q'c,
we have
Hence
+
c
qc
+
,
q'c
=
c
{q
q') c.
39, 47
b d + b a ^ = Q+g' = -+. c
94
And
Thus,
1^
3
+^=
3
+ 9; and 1^ ^
3
2-1
9.
Formulas
for adding
These are
be
98
ad
-\-
be
'
ad
bd
l^d^
bd
'b~d^
30
A COLLEGE ALGEBRA
a
c
^
And
ad
be
ad
be
18
Thus,
[-
10 = ^ + 2 = 8; 6
5
^ and
18
10 3
^^^
3-5
=
120
15
8.
99
Formula
This
is
ac
I'd^hd
For
if
qh,
and
g'd,
we have
c a d
ae
,
gg'
6d.
50, 45, 46
_
.
Hence
15 6 2
Z)
= =
g'
= _.
bd
^. and
ac
94
Thus, '
,^
15
15-6 3-2
90 = ,^
6
15.
100
Formula
is exact.
by
,
another,
,
when
this division
This
is
a
h
a a
d
also q
he
=
Q
_ad
he
For
if
g&, c
g'd,
and
a
r
g"g',
we have
and
-^
~ d
-^
<i'
<l"i
94
Thus, '
g".
96,94
-=-
'
24-5 6-10
=2.
120
60
101
Exact division for negative numbers. Evidently the definition of quotient given in 93 has a meaning for negative numbers also, whenever the numerical value of the dividend is exactly divisible by that of the divisor. Expressing these quotients
as in
94,
we have
4.
102
Theorem
If
a.
~b~^~b'
rb~~b'
a_a
^~b"
a_a
DIVISION
For
if
AND FRACTIONS
a
31
73, 67
qb,
we have
Hence
= ( q)b. q=
94
And
On
is
93
meaningless
103
For g X
number
is
may
;
denote.
(2)
Hence
is
(1)
every
and
there
no number
which multiplied by
In other words, according to the definition of 93 and 94, the Symbol every number, and a/0 no number whatsoever.
is.O,
ber b which
fact,
For, according to 94, 0/6 should denote the number which multiplied and is that number (and the only one), since 0-6 = 0. by b gives
;
FRACTIONS.
DIVISION
of division
86
is
the gener-
93.
It requires that
be introduced into the number system. We seek an new numbers, like that given the
56.
c,
negative numbers in
theorem, in which
a, b,
Theorem 5. When a is exactly divisible bijh, and c by di, the quotients a/b and c/d occur in the natural scale in the same relative order as the products ad and be / that is,
a/b <, =,
1.
or
>
"> be.
For
if
-6d =
b
-d6.
50
93,94
But
-6
a,
and -d
a
c.
Hence
ad
= be.
32
A COLLEGE ALGEBRA
can show in a similar manner that
if
And we
If
2.
conversely, that
If
ad
be,
then a/b
c/d.
For otherwise we should have either (1) a/b<c/d, and therefore ad<hc, or (2) a/b>c/d, and therefore ad>bc. And we can show in the same way that
If
if
105
But the relative ordei Enlarging the ordinal number system. of ad and be in the scale is known, whether the values assigned
a, b,
c,
d be such
as
make a
divisible
by
b,
and
by
d, or not.
b,
of w^hich
and
w'ith
-, or a/b.
b'
exactly divisible by b, let a/b denote, as heretofore, the natural number which is the quotient of a \yj b; but if
is
If a
a/b for the moment merely as a new symbol, read whose relation to division is yet to be given, 122. give to all such symbols a/b, c/d, and so on, the property of order already possessed by those which denote natural numbers, by supposing them arranged in accordance
not, regard
6,"
a/b
according as ad precedes, coincides with, or follows be. Or, employing the signs <, =, >, as heretofore, to
"
precede,"
mean
106
>
or^
be.
is
4/5<7/8,
since
to
1
;
0-3<2-
107
To such
of the symbols
a/b
as denote natural
numbers
this rule assigns their proper places in the scale itself; while to the rest it assigns places between consecutive numbers of
the scale.
DIVISION
Note.
AND FRACTIONS
33
108
any particular symbol a/b with numbers of the scale, we have only to reduce a to the form = bq + r, where r<6, 88. Then if r = 0, so that a = bq, our rule a makes a/b comcide with q. But if r is not 0, our rule places a/b between q and q + I.
find the place thus given
To
respect to the
The entire assemblage of symbols a/b thus defined and like the natural scale which forms part of it arranged is an ordinal system.
109
For
it
has
all
in 17, 18.
Thus,
if
For
if
we have
and cf<ed.
by
106
the corre-
Multiplying the sides of the first of these inequalities sponding sides of the second, we have
adcf<bced.
50
Hence
and therefore
Fractions.
af< be,
a/b<e/f.
50
106
110
When a/b
fraction ; and
we
call it
Su
does not denote a natural number, we call a its numerator, b its denomi-
nator,
b its terms.
Hence
a symbol of the form a/b, defined by its position in an ordinal system which includes the natural numbers.
A fraction
From an
we
are justified in
34
111
A COLLEGE ALGEBRA
Negative fractions.
We
or
whosB numerintegers,
ators,
-1
denoiniuators,
b a
both,
are
nerjatlve
as
-J defining
them ordinally
a
as follows:
^a
2.
irrecede 0.
3. Negative fractions shall be arranged with respect to one another {and 7iegative integers) In accordance with the rule :
a -<,=,
112
or
> -)
according as
ad <,
=,
or
> be.
fractions alike
To distinguish integers and numbers. from other numbers which we have yet to conAnd we call the system sider, we call them rational numbers. which consists of all these numbers the rational system. This system possesses an important property which does
The system
of rational
not belong to
113
its part,
dense;
that
is,
<a
,
We can prove
be
ad
...
lies
between - and
^
a
b
c
--
2bd
Since -
OL
106
b
1.
If
we add ad
ad<bc, we have, by
2ad<bc +
2.
ad,
.:
a{2bd)<b{be
o
ad),
'''
be
-{-
ad
b'^~2bd
If
we add
be to
bc
+ ad<2be,
3
(be
+
,
ad)d<c{2bd),
4- 5 2
.-.
+ ad c <- 2bd d
we have
+
4
3 -6 o
38 7^ 4o
= ^" 24
19
DIVISION
AND FRACTIONS
35
114
Hence, when speaking of rationals, one must carefully avoid such expressions as the " next number greater or less than a To each integer given number for no such number exists.
''
;
such a next integer, but between any rational and a rational assigned as the next, there are always other rationals. In what follows let a, b, c, d denote Operations with fractions.
there
is
115
any given integers, positive or negative. In 98-102 we proved that, when a/b and c/d denote integers, we have
a
'
ad
-\-
be
^
b
ad
ad = -
be
be
bd
J
d
-
bd
.
d
-7-
^
3.
a c -- = ac
b
4.
a
b
be
c d
when
ad
be
-; is
an integer.
But the second member of each of the equations 1, 2, 3, 4 has a meaning even when a/b and c/d are not integers. Each of them is a definite fraction of the kind defined in 110, 111.
Hence
1, 2, 3,
ings of addition, subtraction, multiplication, and division which will make these operations applicable to fractions, namely The sum of two fractions a/b and o/t? is to mean the fraction
:
116
{ad
bc)/bd.
difference obtained by subtracting the fraction c/d is to mean the fraction (ad bc)/bd.
The
from
the
117
The product
fraction ac /bd.
of
two
fractions
is
to
mean
118
The quotient resulting from dividing the fraction fraction c/d is to mean the fraction ad /be.
a/b by the
119
Observe that these definitions are equivalent to the rules for reckoning with fractions given in elementary arithmetic.
The commutative,
associative,
120
mu
ac
Ca
poiid
orv
36
121
A COLLEGE ALGEBRA
rules of equality and inequality, 71, 73, also hold
The
good
Thus,
For
if
.r
a
b
if
f
f
ad
c e -;-^'
^,
then
a r
b
--
f
f
d
ce6/.
^.f = -.?,
b
then
aed/=
118,100,111
73, 106, 111
Hence
cb,
d and therefore -
122
The
fraction a/I>
may
now
be described as the number which multiplied by b tvill produce a, that is, as the number which is defined by the equation
123
'^b b
Tor
a.
-.6 = -. - =
b
6
1
-- = -- = a.
1-6
1
106,111,118
119,
it
124
From 118,
c
follows that
T'^d^d^H
""'"^
b^d^d'V
in other words, that multiplication and division, as defined in 118, 119, are inverse operations. Compare 55.
For, by 118, 119 and 106, 111,
we have
c c
a
b
_ad
be
adc
bed
_a
b
dc cd
_a
b' b
__ac bd
_ acd _a
bdc
b
rd de
_a
b
Hence we may
125
by
deseril)e the
and say
find a mimher tvhich multiplied
is to
c/d
tvill
produce a/b.
By
made
it
introducing fractions into our number system, we have possible always to find such a number, except when the
divisor
c/d
is 0.
DIVISION
This
is
AND FRACTIONS
37
algebra.
the usual meaning of division in arithmetic and the generalization of exact division, 93. Irreducible fractions. Reducing a fraction to its lowest terms.
It is
126
If the numerator
factor,
and denominator of a fraction have a common we can remove it from both without changing the value
of the fraction.
For
=
bin
-, since
b
am
-b
a- bm,
106.
When
tion
is
all
such
common
its
said to be in
7/"
Theorem.
other fraction
tiples
a/b
be
is
an
irreducible fraction,
to
it,
and
a'/b'
any
127
which
equal
then
equimul-
of a and b
a' /b'
resjiectively.
For since
ab',
is
a factor of
b.
a'b.
common
with
Hence a must
have a' = ma, where m is some integer. But substituting ma for a' in a'b = ab', we have mab = db', and therefore
We therefore
b'
mb,
50.
Corollary.
128
ators
must
be equal,
and
IN
MEASUREMENT
129
The definition of length given in 81 Fractional lengths. only applies to such line segments S as contain the unit segment s a certain number of times exactly.
But even
or
its
if
exactly,
it
may
still
be
commensurable with
that
is, it
some other
aliquot part of s
:
may
length as follows
If a given line segment co7itains the hth part of the unit segment a times exactly, we say that its length is the fraction a/b.
130
38
A COLLEGE ALGEBRA
s exactly 7 times, the
(in
length of
131
Note.
Observe that
For
if
is the length of S in terms of s according every fraction of the form ma/mh. contains the 6th part of s exactly a times, it will contain the
if
a/h
is
ma
times.
132
are the lengths of S and T in terms of s, so also 131 that is, the bdih. part of s is contained in S
;
T
or
exactly 6c times.
Hence
that
is,
S<, =, S<, =,
It
or
>T, >T,
according as
ad<, =,
or >6c,
according as a/6<,
or >c/d.
106
133
is
hardly need be said that the definition of length here given equivalent to the definition ot fraction given in elementary arithmetic,
Note.
lesser fractions are there defined as fractions v?hich
134
may
3
I
line, 85.
_4
I
S-2
T
I
1
I
I
1
I
253
I I
L^
>
that
Thus, to construct a point, P, which will picture 7/3 in the same way A pictures 1, we have only to start at the origin and lay off the third part of the unit OA seven times to the right.
is the picture of 7/3. P', the corresponding point to the left of 0, "We proceed in a similar manner in the case of any given fraction,
positive or negative.
135
All such points are arranged along the line in an order corresponding to that of the rationals which they picture. With this in mind we often speak of one rational as lying to the left or right of another rational, or as lying between two
other rationals.
IRRATIONAL NUMBERS
IV.
39
IRRATIONAL NUMBERS
PRELIMINARY CONSIDERATIONS
Definitions.
square
"
;
aaa ...
The product aa is represented by a^, read "a. " the product the product aaa, by a^, read " a cube to w factors, by a", read " the wth power of a."
;
136
In the symbols
exponents; a itself
a^,
is
a^
a",
the numbers
2, 3,
n are
called
Finding
finding
a",
a^
from a
is to
finding
a^,
cubing a
the wth power. The operation which consists in raising a given a given power is also called involution.
raising a
number to
137
If,
as
we
a"
5.
This equation suggests two new problems To assign values to 7i and b, and then find First.
:
a.
Second.
Thus,
To
n
b,
n.
(1) let
and
&
9.
and we
find that a
(2) let
Again,
=3 =2 =
3
or
and
-3 6 =
;
for both 32
8.
2"
and we
find that
for 2^
8,
When
1.
= 5,
b,
138
is
and
is
expressed in terms
of "
by the symbol V^, the simpler symbol V^, read square root of b," being used when n = 2. 2. n is called the logarithm ofh to the base a, and is expressed in terms of a and b by the symbol log b.
n and
40
A COLLEGE ALGEBRA
9,
Thus, since 3^ = 9 and ( 3)'- = 9, both 3 and and both may be written Vl) but see 139.
;
Again, 2
is
that
is,
logs 9.
139
Note.
V9, we 3, by
Instead of representing both the square roots of 9 by the symbol may represent the positive one, 3, by Vg, and the negative one,
V9.
This
is
the usual
method
it.
we
shall follow
140
V6
is
found,
when n and
log^Z* is
found,
when a and
b are
Both
'i.41
called ^'fZiw^ the logarithm ofh to the base a. these operations are inverses of involution, 55,
124
involution has two inverses, while addition and multiplication each has but one, will be seen by comparing the three
Note.
equations
\.
-\-h
c.
2.
ab
c.
3.
a^
c.
+ + a, ba, the problem : Given c and 6 in 1 or Given c and a, find b. 2, find a, is of the same kind as the problem But since a^ is not equal to b", the problem Given c and b in 3, find
Since a
b b
and ab
a, is
wholly different
in
Given
and
a, find b.
143
needed.
We
;
in detail
importance to the four fundamental operations only. But the now concerns us is this: 27ie// necessitate further
extensions of the
number system.
at
In fact,
it is
Va
number
Thus, to cite the simplest of illustrations, neither denote a rational number. For
1.
V 1
nor
V2
can
num ber
1
is
positive,
no rational
exists
1.
Hence
No
rational
number
exists
whose square
we can show
IRRATIOXAL NUMBERS
Suppose p/q
to be a fraction in its lowest terms,
41
such that
{p/(/)^
2,
orpVg- =
2/l.
But
this,
since p'^/q^ is in its lowest terms, 492, 2, it would follow 128, that p- = 2, which is impossible, since pj is an integer.
from
Therefore
It
2 cannot denote a rational number. can be .shown in tlie same way that if a/b be any fraction in
its
lowest terms, ^a/b cannot denote a rational number, unless both a and b are nth powers of integers.
We are to make good this deficiency in our number system by creating two new classes of numbers the irrational numbers, of which v2 is one, and the imaginary numbers, of which
:
V 1
We
is
one.
numbers
in the chapter
which
follows.
The rational numGeneral properties of the rational system. bers constitute a system which has the following properties 1. It is an ordinal system.
:
143
2.
It is dense
that
is,
of the system, a
and
b,
there
system.
3. The sum, difference, product, and quotient of every two numbers of the system are themselves numbers of the system, the quotient of any number by excepted. By the definitions which follow, we shall create a more extended system which possesses these same three properties, and which includes the rational system. 1. The number ^ separates the Separations of the first kind. remaining numbers of the rational system into two classes
:
144
42
A COLLEGE ALGEBRA
than)
than)
the one class consisting of all rationals wliicli precede (are less J, the other of all rationals which follow (are greater
A.
C'g
respectively.
pictures of
In the figure, the half line to the left of the point \ contains the pointall numbers in the class Ci, and the half line to the right the
all
point-pictures of
numbers
From
1. 2.
109, 111, and 113, it immediately follows that Each number in L\ precedes every number in Co.
is
There
no
last
number
in Cj,
and no
first in Cj.
it
Thus, were there a last number in Ci, there would be numbers between and 1/3, 113, which is impossible since, by hypothesis, all rationals
than
1
less
/3 are in C\.
145
2.
three parts
Ci'
C'j, \, Co, we may join ^ to C'l, so forming a class made up of C\ and J, and then say The number 1 separates the entire rational system into two
:
parts, Cj'
1. 2.
and
is
C.^,
such that
Each number
There
in C\' precedes
a last
number
in
C'l',
is
no
first
number
146
3.
Or we may join \
:
and
then say
parts, Ci
1. 2.
such that
Each number
There
in Ci precedes every
Ci,
number
is
in
a,
C2'.
no last number in
J.
but there
first
number
in Co',
namely
It is
numbers defines
147
Conversely, if we are able, in- any way, to separate the entire rational system into two parts, Bi and B^, such that each
IRRATIONAL NUMBERS
number
43
either a last
in B^ precedes every number in Bo and that there is number in B^ or a J77'st in B^, the separation will
numbers and,
Thus,
let
serve to distinguish this last or first number from all other in that sense, to define it.
us assign the negative rationals to Bi and the remaining is the first There is then no last number in i>i, but
rationals to B^.
number
in Bz-
And
zero
is
distinguished from
number
in Bo, as perfectly as
Note.
in B2.
11.3,
Obviously there cannot be both a last number in Bi and a first For there must then be rationals between these two numbers,
148
But we can
also, in various
149
ways, separate the entire rational system into a part Ai in which there is no last number, and a part A^ in which there is no first number.
every rational
Thus, since no rational exists whose square is 2, 142, is either one whose square is less than 2, or
one whose square is greater than 2. Let A2 consist of all positive rationals whose squares are greater than 2, and let A^ consist of all the other rational numbers. Then
1.
Each number
let ai
number
in A^.
be any number in Ai, and a^ any number in A2. Evidently ai<a2, if a\ is negative or 0; and if ai is positive, a-^Ka-i^,
For
and therefore
2.
ai
is
< a^.
no
last
There
number
in A^,
and no
first in A^.
than than
2, 2,
we can always
;
last in
Ax.
whose square is whose square is also hence no number can be assigned which is the 183, 2 (.3) Similarly no rational can be assigned which is the ^rs^ in A^.
positive rational, ai, has been assigned
V2.
The
relation
150
numbers, A^ and A^, is therefore precisely the same as that between the classes Ci and C2 in the separation
of
corresponding to
^,
in 144.
44
A COLLEGE ALGEBRA
But no rational number exists which can be said to correto be defined by it. spond to the separation A^, Ao, or For since every rational belongs either to ^i or to A^, no rational exists which lies between A^ and Az, SiS ^ lies betweenCi and Cg. And since there
is
A 2, no
rational exists
in
as
i corresponds
to the separation
of
145, or to the
146.
(Compare
147.)
Hence this separation .lu .1 2, creates a place for a 7iew ordinal number, namely, a number which shall follow all numbers
in
A I and
We
sent
it
by the
a,
9.
letter a
later,
we
shall
find that a^
2,
symbol V2, 182. We then define this new number a as that number which lies between all positive ratlonals whose squares are less than 2
significant
by the more
and
all
2.
We may
where
a-^
by the formula
<a<
<
is
a2
and
a>2,
means "precedes."
of the
152
Note.
tions of
Observe that
tlie
this definition
same kind
as the defini-
Like negative and fractional numbers given in 56, 110. these numbers, a is a symbol defined by its position in an ordinal .system of symbols which includes the natural numbers. It therefore has precisely
same right as they to be called a number. Our reason for inventing this and similar numbers is also the same as our reason for inventing negative numbers and fractions. They serve a useful purpose in the study of relations among the numbers which we already possess, and among things in the world about us.*
the
* We may add that there would bo no ohjection from an ordinal point of view to our inventing more tliaii one inuiilxT to corrpspond to the seiiaration the formula o, <a< b < Oj. A^, A^, say two numbers, a and b, detined ordinally by But there are objections of another kind to our inventing more than one such number. See page 67, footnote (3).
.IRRATIONAL NUMBERS
45
153
The irrational numbers in general. The real system. The particular separation of the rational system which we have been considering is but one of an infinite number of possible separations of a similar character. For every such separation we invent a new number, defining ordinally with respect to the numbers of the rational system V2 in 151, precisely as we have defined the number a
it
To
bers,
distinguish these
rational
num-
irrational numbers, or simply irrationals. Again, to distinguish the rational and irrational numbers
call
we
them
alike
consider,
we
call
them
rQ.al
numbers.
Finally, we call the system which consists of all the rational and irrational numbers the sijstem of real numbers, or the real
system.
Hence, using a to denote any irrational number, following general definition of such a number
:
we have
the
An irrational number, a, is defined whenever a law is stated which will assign every given rational to one, and but one, of two classes, Aj, Ag, such that (1) each number in A^ precedes
every number in A and (2) thei'e is no last number in A^ and no first number in Aj tlie defiiiition of a then being : it IS the one mimber which lies between all numbers in Ai and
;
154
all in Ag.
It is
A2;
also that
here implied that there are numhers in both the classes Ai and Ai and Ao together comprise the entire rational system.
An
irrational
it
number,
a,
is
155
according as
that is, the numbers The real which constitute it are arranged in a definite and known order, For the definition of each irrational indicates how it 17. lies with respect to every rational: and from the definitions of any two given irrationals we can at once infer how they lie
:
156
46
A COLLEGE ALGEBRA.
let a
Thus
1.
then
If every rational which precedes a also precedes b, and every rational which follows a also follows b, the numbers a and b occupy the same position relative to the nmnbers of the
rational system.
therefore,
By
154, a
our definition of an irrational number, and b denote one and tte same number.
We
indicate this
by the formula:
2.
If
among
b,
which precede
We
a).
<
b or b
>
a.
among the rationals which precede a there are some which follow b, then a itself must follow b (or b precede a).
3.
If
We
157
>
b or b
<
a.
different real numbers which precedes and which follows; also, that we may always draw the following conclusions with respect to three given real numbers, a, b, c
It thus
appears that
are given,
we can
at once infer
If a
b,
and b
= < <
c,
then a
then a
= < <
c.
If a
If a
< b, = b,
and b
and
b
c,
c.
c,
then a
c.
158
system is dense. For there are rational numbers not only between any two unequal rationals, 113, but also between any two unequal irrational numbers, and between
real
The
any
is
rational
irrational,
The
real
system
first
is
continuous.
possesses the
IRRATIONAL NUMBERS
47
system enumerated in 143. But it possesses an additional property not belonging to the rational system, namely
:
If the entire veal system be separated into two j)(^^'ts, Ri and R2J such that each nmnher in E-i jjvecedes everij number
in Rjj there
not both.
is
Rj or a
For in separating the real system into the parts i?i and E^ we separate the rational system into two parts, Ai and Ao, the one part consisting of all the rationals in Ei, the other of all the rationals in R2.
Every rational belongs either to ^1 or to An, and each rational in Ai precedes every rational in A^. Let a be the number which the separation Ai, A^ defines, 147, 154. Then either a is a rational namely the last number in Ai or the first
in A2, 147,
or, if there
be no last number in
154.
Ai and no
first in
A2,
is
1.
is
the last
number
in Ai,
it
Ei; for
it
were
a,
there any number in Ei after a, there that is, rationals in Ai after a, which
2. 3.
7?2-
would be
is
rationals
between
and
impossible.
in A2-,
it is
Similarly,
If If
if
is
the
it
first
number
also the
first in 2^2-
is
irrational,
i?i,
a belongs to
rationals in
must, by hypothesis, belong either to 7i or to for were there any it is the last number in Ei
;
number
that
if
is,
Ai
after
it
and
a, 158,
in like
manner,
a belongs to E^,
it is
the
first
number
in E^.
Finally, there cannot be both a last number in Rj and a first in JB2, since there would be rationals between these two numbers, 158, that is,
rationals belonging neither to
Ai nor
to
Ao
which
is
impossible.
indicate that the real system is dense and at the same time possesses the property just described, we say that it is
To
continuous.
Theorem.
real number,
a,
is
160
defined, ivhenever
real system
may
stated by means of which the entire be separated into two parts, Rj, R2, such that
a law
is
this
number,
a,
This
48
A COLLEGE ALGEBRA
APPROXIMATE VALUES OF IRRATIONALS
161
Given any irrational number, a, defined as in 154. By method illustrated below we can find a pair of rationals, the one less and the other greater than a, which differ from each other as little as we please. Such rationals are called
the
approximate values of
a.
Let a be the irrational, V2, which lies between all positive rationals whose squares are less, and all whose squares are greater than 2. 1. We may find between what pair of consecutive integers a lies by
1,
2,
3,
successively, until
we reach one
We see
2.
at once that 1^
lies
< 2 and
22
> 2.
Hence a
computing
than
2.
between
then find
. . .
We may
l.l^, 1.22,
and 2, or 1 <a<2. between what pair of consecutive tenths a lies by successively, until we reach one which is greater
1
We
3.
= 2.25.
Hence a
between
1.4
and
1.5, or
1.4<a<1.5.
By
a similar procedure
we
find, successively,
1.41
<a< 1.42,
1.414
<a< 1.415,
number
4.
Let
rti
Then
ai<a<a2 and
a2
Oi
1/10",
and by choosing n great enough we can make 1/10" less than any positive number, as 5, we may choose to assign, however small. 5. We call 1.4, 1.41, 1.414 the approximate values of a = V2 to the first, second, third place of decimals; and so on.
Evidently the
to
process
thus
illustrated
may
be applied
number, a; for all that the process requires is a test for determining whether certain rationals are less or greater than a, and the definition of a, 154, will always supply such a test. We therefore have the following theorem
any given
irrational
IRRATIONAL NUMBERS
Let a denote any
number, as
8,
49
If any positive small, we can
162
rjiven irrational
it
number.
be assigned,
matters not
how
such that
ai
<a<
a2 aiul ao
ai <
8.
Evidently this theorem is true of rationals also. Tlius, if a denote a given rational number, and
tto
ai
=
ai
1/10",
a^
1/10",
2/10" as
small
a.s
we
remains to give the real system the third of the propersystem enumerated in 143. For this we
theorem
Theorem.
1.
2.
Let Ai and Ag be two classes of ixitionals such that Each number in Aj is less than every number in Ao,
163
There is no last nmvber in Aj and no first in K^, For every positive number, 8, that may be assigned, it matters not how small, ice can find i?i A^ a number ai, and in A2 a number a2, such that
3.
a2
ai <
8.
We may
That there
lies
one
is at least
one such
number
follows from
and
2,
by
154.
That there cannot be more than one such number follows from 3. For suppose that between every ai in ^1 and every ao in ^2 there were the two rationals d and d\ as indicated in the figure
:
Then
we should have
aoXi', and
and therefore
which
is
aa
it
ai
>d'
- ai> d,
3.
d,
73,121
39, 121
impossible, since
contradicts
50
A COLLEGE ALGEBRA
of
which are
irrational^
for between these two numbers there lying between every ai and a^ would be two rationals also lying between every ai and ao, 158, which
we have
164
Note.
154, in
just
shown
to be impossible.
differs
This theorem
that
it is
from the
definition of an irrational
number,
lies
either in
^i
or in A2.
165
Addition.
rational or irrational,
and
l>i,
h^,
<a<
is
a^,
and
last
^i
(1)
no
number
or ^1, and no first number of the kind denoted by a^ or and that if any positive number, as 8, be assigned, it matters not how small, we can always choose a^, a^, and 7*1, h^, 162, so that both ! < S and h.^ h^ < 6. rtj (2)
by
ffi
Z2
When
from
116
by
121, that
i
^>l
<
i8
<
^2-
not, it follows
^'1
<
^2-
(3)
to define the
:
sum
of a
and
b,
when one
166
The sum of
which
as
lies
and
b,
written a
b.
is
to
+ \\.
In other words,
a-i
the
bi
and
bi
<
ao
a 4- b
<
bj
ao
+ ba,
where
bo denote
ai
<a<
and
IRRATIONAL NUMBERS
To
1.
51
is
we must show
that there
such number a
Each
There
(iTi
+ b. + 6i is
no
bi
163
for
than every a2
5i
+ +
b^.
2.
is
last ai
and no
a^
first a-z
+
;
b^.
Thus, ai
+ +
cannot be the
last ai
6i
is
no
last
a-i
and no
3.
last 6i,
we can choose
6i
and
bi so
that
ai>ai and
choose
6i>6i',
and
therefore ai
If
> a\' +
6i'.
any
positive rational,
5,
be assigned,
we can
ai,
a^-,
&i, &25
so that
a2-ai<5/2
and therefore
{a^
and
62
-&i<5/2,
hi)
102
121
60)
(cli
<
5.
as in Definition of a, i, a^ have the same meanings 165. Considerations like those in 165 lead us, when a is
a.
Let
167
irrational, to define
a as follows
The symbol
is to
168
a2 <C
where
ai, a2
a <C
a^,
denote
"\ a2.
is
It
a
1.
number
Each
There
2.
az,
3.
is less than every ai, since ai<a2. 73, 111 no last ao and no first ai. Thus, were there a last there would be a first a2 but no such number exists. We can always choose ai, a^, so that
az
is
Subtraction.
ao)
ao
< 5.
from
is, a,
162
b,
is
to
subtracthir/ h
written
169
+ (
b)
that
a-b = a+(-b).
The meaning
It follows
of a
+ (
ai
b) itself
is
known from
166, 168.
from
60
<a
< a2
61,
where
52
170
A COLLEGE ALGEBRA
Let a and b be any two
h.j,
any
piositive rationals
<a<
cr2
and
^i
<b<
a, b :=
J
b^.
(1)
it
When
(1),
and b are
rational, say a
,
y8,
follows from
by
121, that
^ n^
and
in every case
it
follows that
a^bi
<
a2^2-
(2)
are therefore led, when one or both of the are irrational, to define their product thus
:
We
numbers
a,
171
mean
The product of two positive numbers a and b, ivritten ab, is to that number which lies between all the numbers aibi and In other words, ab is the number defined all the numbers aobo.
ajbi
by the formula
< ab <
any
a^bo,
where
ai,
ao,
bi,
b2 denote
ai
such that
<a<
a.2
and
It follows
from
is
1G3 that
there
is
ab
for
1.
Each
There
aihi
is
less
2.
no
last aibi
a^bifirst a-zbo-
(Compare proof,
ai,
a-z,
100, 2.)
?'2i
3.
Any
we can choose
fri^'i
6i,
so that
(''-2
<
5.
For
a^bo
Ui,
Oi&i
b\)
bi (a2
fii),
a-z,
&i, &2,
1^2, so that
(1) (2)
&i
and therefore
Oo (62
and ff2 ai < 5/2 ?^i, < 5/2 61) + bi (uo c/i) < 5.
ffl2
We may
ai, 02, so
make such
a clioice of
ai,
rto,
?);,
60,
as follows:
621
First take
of the kind
as
b^',
that
ao- ai<d/2b/.
IRRATIONAL NUMBERS
Next, using the a^ thus found, choose
62
61, ^o,
53
so that
&i<5/2a2,
62'
as in
it
(1).
Since &i
therefore 5/2
02
follows from
(3)
that
(1).
i<5/2 6i,
0.
as in
Let a and b
172
Then
b)
and (
a) b are to
mean
ab.
2.
3.
(
a
a)
b) is to
mean
ab.
0.
and
a are to
mean
Definition of 1/a.
!, O25
Let a be any given positive number, and any positive rationals vi^liatsoever such that
i
173
<a<
Os-
165 lead
us,
when
a is irra-
1/a
as follows
is to
mean
bij
the
formula
174
l/2< l/a<
where
ai, ao
l/oi,
denote
any j^ositive
Oi
<a<
is
It
number
1/a
1.
for
Each l/ao
There
is
is less
2.
3.
no
last
Any
positive 6
than every 1/ai, 106. l/ao and no first 1/ai. (Compare proof, being given, we can choose ai, ao so that
168, 2.)
1/ai
- l/a2<5.
if
For
l/fli
if
- l/2<5,
a2
a2- ai<5-ai2Oi,
106,117
But
tti,
ai
we can choose
ao so that ai
> a/ and
fli
therefore
< 5airt2175
Definition
of
l/(
a).
number.
Then l/(
a)
to
mean 1/a.
54
176
Division.
A COLLEGE ALGEBRA
The quotient of a a 1 /b, that is, a
by b
1
(b
not 0)
is to
mean the
number
^b-
of a 1/b itself is known from the preceding definitions. a and b are positive, it follows from 171, 174 that we may also define a/b by the formula
The meaning
When
ai/62<a/b<a2/6i,
where
ai, a^, 6i, &2
6i
177
The operations just defined are extensions of the corresponding operSubtraction continues to be the ations for rational numbers.
The commutative,
associative,
inverse of addition, and multiplication of division. Finally, additiofi and multiplication continue to conform to the commutative, associative,
and
and
c
distributive laivs.
are any three positive
Thus,
170,
if
a,
b,
numbers defined, as
in
by the formulas
!
bi<h<b-2,
Ci
we have
For by
166, 171, a (b
ai('>i
ffli^i
c)
=
+
ab
ac.
c)
and ab
And
the
since aj (bi
+ ci)<a(b + c)<a2(62 + C2), + ttiCi < ab + ac < a2&2 + 2C2. + Ci) aibi + aiCi and a2 (62 + c^) = aobo + a2C2,
(1)
(2)
120,
numbers defined by
and
(2)
178
The
for
as just defined,
namely
According as
so
is
a<, =,
a
-t-
or
or or
<, =,
also,
ac
<, =, >, =,
but
ac
or
if
if c
> 0, <
0.
IRRATIONAL NUMBERS
Thus,
if
55
For
Ci
let
b,
and choose
a,
166.
(1)
so that Ci
< c < Ci +
a.
Then, since a<d and c<Ci + a, we have a + c<d + Ci + and since d + a < b and Ci < c, we have d + a + Ci < b + c, But from (1) and (2) it follows, 157, that a + c<b + c.
(2)
The proof that, if a < b and c > 0, then ac < be, is similar. But in this case we choose Ci so that Ci<c<Ci(l + a/d).
From
c
these rules
it
<
d,
then a
+c<b+
a, b, c,
d are positive,
39, that if a < b and and so on also, as in 50, when that if a < b and c < d, then ac < bd,
follows, as in
d,
179
values.
1.
180
we can
state the
theorem of
162 as follows
is
2ko,
Whe7i any irrational a is given, and any jiositive rational 8 assigned, however small, we caii always find rationals, aj and
which will
For, by 162,
than
8.
we can
it
and a^ such that a\<a.<a<2, and a2 ! < 5. ai<a2 ai, and therefore follows, 178, that a
find ai
Oi
< 5.
<
cti,
we prove
that a2
Thus,
161,
we have V2 -
1.41
a < 5. - V2<.01.
We
represents a with an
In practical reckoning we employ approximate values of 2. irrational numbers more frequently than the numbers themselves.
If !
and
hi
and b respec-
approximate value of the sum a + b. But to insure that the error of a-^ -\- h^ shall not exceed 8, we must ordinarily choose a^ and b^ so that their respective This follows from the proof in errors shall not exceed 8/2.
tively,
then
a-i
+ bi
166.
ab,
Similar rules for finding approximate values of a b, and a/b with errors not exceeding 8, may be derived from
56
A COLLEGE ALGEBRA
INVOLUTION AND EVOLUTION
181
Powers.
rationals,
182
Roots.
aa, aaa,
by
a-, a^,
positive
number
b,
written
;
Vb,
that
is
is,
to
mean
is
that positive
number whose
Vb
by the formula
b. ( Vb)'" must show that one, and but justify this definition, we accomplish one, such number as it implies actually exists.
To
We
this as follows
183
Theorem.
tive real
1.
the
mth
If b is the
is
the theorem
Thus,
2.
if
obvious.
8/27
2/3.
power of a rational, its mth. root is that real number a which lies between all positive rationals, than b and all positive rationals, !, whose mth powers are less
If b is not the mth.
a^,
whose
??ith
b.
Compare
151.
It follows
from
since (1) every positive rational is than every a.i, and (3) there is no last a\ and no
one, and but one, such number a, either an rti or an a^, (2) each ai is less
is
first 03.
We may
If
Then
since
p < b,
Let one of them be p + 5. We have only to show that we can find a rational q>p such that g"' <p'" + S, or q'" p'" < S for we shall then have p"'<q"' < b, so that p is not the last rtj.
between
p"'
and
b.
But
q"*
-p'^
{q
-\
qjy"-^
+ p'""')
308
a2,
if 02'
he any particular
S/mao.'"'-^.
iiq^p +
is
We
no
first a^.
This established,
may
= "v b.
IRRATIONAL NUMBERS
For, since
57
171, 181
02"*.
But b
Hence
Rules
is
the only
ai<a<a2, we have ai < a < 02. number between every Ui"' and every
b,
a'"
that
is,
= V b.
Let a and
h
of equality
and inequality.
denote any
184
and
any positive
integer.
Then
According as
so is
a <, =, or
a"'
<, =,
or
b,
b"',
(1) (2)
and
Va<,
prove
(1)
if
=, or
Vb.
We may
Thus,
by repeated use
of 179.
is,
a<b, then a-
a<b
-b,
that
a^
< 6-
and so on.
We derive (2) from (1). Thus, if a Va < or > V 6, we should have a < or > b.
Rules
bers,
1.
6,
then
Va = v6
for
were
of exponents.
rn
and
a'"
Let a and b denote any two and n any two positive integers. Then
real
num-
185
a"
= a'" +
a?
a!^
2.
acta
('")''
'"".
3.
(aby
a-'^
= a^b"
urn
Thus,
aa
aaaaa
a^
+-
177
(a2)3
= =
2.a2.a2
a?>
=a2 + 2 + 2
=a2.3
byl
177
(a6)3
a6
aft
aaa bbb
a^
b^
And
similarly for
any other
and
n.
A
real
theorem regarding
roots.
Let o and
186
numbers, and
any positive
integer.
Then
Va Vi =
For
Va/>.
(V6)'
(Va
V6)'
= V^)
(
a&
182, 185, 3
and
Hence
and therefore
(Vaby^
=
=
ab.
182
(Va
m
V5)'
(va6)'
in
m^ ^ Va v 6 = Va6.
,.
184, (2)
58
A COLLEGE ALGEBRA
VARIABLES AND LIMITS
187
Variables.
bers,
We
num
such as
^35
'
' '
^H>
'
known, if the value of every particular term a is known, or can be computed, when the index n which shows its position in the sequence is given. We often have occasion to consider variables which are such given but never-ending supposed to be running through
is
given or
sequences of values.
Thus, 1, f, |, such a variable
is,
is
if
is such a given never-ending sequence, and ;-^, we suppose it to be running through this sequence,
that
J,
188
Limits.
As x runs through
1,
how small, assign any positive number, as 8, it matters not x will ult ately become and remain less the difference 1 than the number so assigned. Thus, after x reaches the 100th
we
will
remain
less
than
.01.
We
express
all
this
the sequence
general
i, , f,
189
variable x, xvhvsh is supposed to be running through a given is said to approach the number
remain numerically
less
than every
number
that
we
may
also
assign.
it is
;
not enough that a - x become less than 5 it must to approach a as limit. the difference Thus, if X run through the sequence i, 0, f 0, f 0, - X will become less than every 5 that we can assign, but it will not
Observe that
remain
less, if
is
remain
less
than this
5,
and x
as limit.
In particular, &
may become
that
is,
may
reach
its
limit a.
IRRATIOXAL NUMBERS
59
190
To indicate that x is approaching the limit a, we write either x ^ Sl, read " x approaches a as limit," or lim x a., read " the limit of x is a."
191
entirely on the character of the sequence of values through which it is supposed to be running.
Thus, while x approaches a limit when it runs through the sequence f plainly it does not approach a limit when it runs through the
)
i) f
>
sequence
1, 2, 3, 4,
-,
or the sequence
1, 2, 1, 2,
the other hand, remains always less than it appjvaches a limit. And this limit
192
either c
or some
number which
is
less
than
c.
For by hypothesis there are numbers which x will never Assign all such numbers to a class R^, and all other numbers, that is, all numbers which x will ultimately exceed,
exceed.
to a class
^'^
i?i.
thus obtain a separation of the entire system of real numbers into two parts, Ri, R^, so related that each number
in Ri
is less than every number in T?.,Obviously there is no last number in Ri.
We
Hence,
a.
160,
there
is
first
number
in R^.
Call this
number
As x
it will approach a as limit. For however small 8 may be, if only positive, a S belongs to the class of numbers i?i, which x will ultimately exceed. Hence x will ultimately remain between a 8 and a, and therefore differ from a by less than 8.
increases,
it
may
be demonstrated that
x continually decreases, but, on the other hand, If remains always greater than some given number c, it approaches a limit. And this limit is either c or some number tvhich is
greater than
c.
193
60
194
A COLLEGE ALGEBRA
It is not necessary, however, that x Regular sequences. should always increase or always decrease, if it is to approach a limit.
Thus, X is sometimes increasing and sometimes decreasing, as it runs J, |, 5, 15, but it approaches as hmit. through the sequence
shall prove that x will or will not approach a limit, -, a, through according as the sequence of values a^, a^, which it runs, has or has not the character described in the
We
following definition
195
The sequence
eve7'i/
ai, aj,
, a,
is
said
to be regular,
if for
positive test number 8 that may be assigned a corresponding term a^ can he found, which ivill differ numerically from
less
than
8.
(1), 161, is
regular.
For the difference between the first term, 1.4, and every subsequent term is less than 1/10; that between the second term, 1.41, and every subsequent term is less than 1 / 10- that between the nth term and every subsequent term is less than 1/10". Now, however small 5 may be, we can give n a value which will make 1/10" smaller still and if k denote such a value of n, the kth term of will differ from every subsequent term by less than S. 1.4, 1.41, Thus, if we assign the value 1/500000 to 5, we have 1 /lO^ < 5, so that the sixth term of 1.4, 1.41, will differ from every subsequent term by
; ;
less
2.
5.
also regular
'
J 5
/0\
17...
1,
{0\
-h
-h
tV,
(4)
is
2,
1,
1,
(5)
Observe that
in (2)
a lesser term, in (4) We sometimes encounter regular sequences like (5), all of whose terms after a certain one are the same. Evidently a variable which runs through such a sequence will ultimately become constant, that is, will
reach
3.
its limit.
each term
2,
3,
4,
...,
(6)
I, i,
i,
i, ..-.
(7)
IRRATIONAL NUMBERS
For
in (6) the difference
61
'*
between a term and a subsequent one may always be' indefinitely great, and in (7) it may always be |, and therefore not less than every number, I for instance, that we can assign.
Formulas
relation
1.
We may
indicate
the
a^,
196
by the formula,
\ap
2.
vi^ill
63
af.\<8
a,,
for every
p>
k.
(1)
between
a^.
lie
between
a^
3.
8
ttp
may be which are > a^. and a^ + 8, and any which are < a^. will and a,., we may also write
Z<
<
Qj.
for every
if
p>
k.
(2)
It follows
from
(2) that
are less,
a^,
may
exceed
8,
but not 2
But we can always find a term, a which corresponds to 8/2 as a^. corresponds to 8. The difference between every two terms after , will then be numerically less than 2(8/2), or 8 that is, the relation between every two of these terms
;
by the formula
^
foi'
|p
Theorem
to
^.jI
<
every
>q>
I.
(3)
2.
The
variable
197
sequence of values
aj, ao,
a^,
through tvhich
it is
supposed
For there are numbers to whose right x will ultimately remain as it rnns through the sequence a^, ^2, , ,. (1)
Thus,
if S
will
remain
to
the right of
a<.
5 after it
a^.,
196
(2).
Assign
bers
all
that
such numbers to a
all
class,
Tii,
and
all
other
num-
is,
numbers
to
whose right x
to a class, R^.
62
A COLLEGE ALGEBRA
thus obtain a separation of the entire system of real into two parts, Ri, R^, so related that each number
less
a,
if
We
in Ri
numbers
is
By
160, a definite
number,
Thus,
constitute
exists at
which
i?i,
the sequence be \, J, \, ^^, , the negative rationals and the positive rationals, Eo. and a itself is 0. but
;
it
number,
S, it
matters not
a,,
how
(3),
regular,
we can
find a term,
196
[ttp
But since a
for every
to
/?i, all
2^>
q> m.
+
(2)
lie to the right of a 8/2 8/2. And since a belongs to i?2) among these values there will be some after a, which lie to the left of a 8/2 for otherwise a 8/2 would
belong to
Thus,
if
i?i,
since x
the sequence be
Jg, lie
|,
i,
|,
j-\,
and
between a
5/2 and
x.
5/2, that
between
Then
a-S/2<<<a + 8/2,
or
|a-<|<8/2.
From
(2)
(3)
and
1^
>
m,
it
<^;,|
<
for every
>
</'.
In other words, after x reaches the value a^ the difference a X remains numerically less than 8.
Therefore x approaches a as limit,
198
Conversely, if
189.
a,
x
,
is
approaching a
limit,
the sequence of
values
ai, ao,
a^,
through ivhich
it is
supposed
to
run,
must be regxdar.
IRRATIONAL NUMBERS
For since the difference
remain numerically
a.
63
will ultimately
become and
8,
less
a/,
189,
1^
we can choose
''^I
<
S/2 and
|a
a^,|
<
a^.|
8/2
for every
p>
j^
k;
k.
whence
[a^
aj, a^,
<
S
is
for every
>
regular,
196
(1).
We may
a limit
The sufficient and necessary condition that a variable approach is that the sequence of values through which it is supposed to run he a regular sequence.
199
it
imme-
200
2.
x approaches
as
limit;
and conversely.
the limit 0, as
it
Thus X approaches
and
1
|j
whose limit
yi
is
is
called an infinitesimal.
Theorem
ically less
and A and B remain numerIf than some fixed number^ c, as x and y vary, then
and y
= 0,
201
Ax
-V
By = 0.
positive
number,
5, it
matters not
how
small.
<S/2c.
200,1
64
Since y
A COLLEGE ALGEBRA
= Q,
will ultimately
remain numerically
< 5/2 c.
200, 1
..
Hence
Ax + By
will ultimately
remain numerically
1.
< 2c
2c
<
5,
as limit, 200,
= 0,
may
and
then (xy
3)
?/
0.
202
Note.
variables.
This theorem
readily be extended to
of
Thus,
if
= 0, =
2/
= 0,
then
Ax + By + Cz:^ 0.
203
limit of the sum, difference, product, quotient of two variables ivhich approach limits is the sum, difference,
2.
Theorem
The
product, quotient of these limits : that the limits a and b respectively, then
1. 2.
a-
is,
if x
and y approach
+y= + :^ a
ft 1/
5.
3.
xi/==ab.
b.
4.
x/ij
a /b, unless
= 0.
For, since a - x
and
- yz^O,
200, it
follows from
201 that
(1)
A{a-x) + B(b-y) = 0.
The formulas
1. 1, 2, 3,
4
{X
may
be derived from
(1).
.:
Thus,
a
is,
y)
that
2.
a
is,
{X
y)
= (a- X) + (b - y) + y = a + 6. = (a x) {b y)
~
y
i 0,
^ 0,
0,
by(i)
.-.
that
3.
x
ab
=a
-\-
b.
xy
{a
x)b
{b
y)x
.-.
that
is,
xy ==
x
ab.
'
that
= /a
x\
/x
x\
\1
,,
b-y
is,
{b-b)'-{b-y)
^''-^^b-^'-'^by--'
x/y :^a/b.
If
3.
:s.
204
205
Corollary.
=
is
a,
then x"
= a".
is,
Theorem
The
approaches a limit
If X
= a,
then
V x =L Va.
IRRATIONAL NUMBERS
1,
65
When
a
0,
Since x ==
< 5".
200, 1
Hence
V x will
a
is
< 5.
184
Therefore
2.
is
Vx==0.
notO.
It follows
When
from a
later section,
always exactly divisible by Vx Va, and that the quotient Q does not approach the limit when x =L a. It therefore follows from 203 (1), by setting A = 1/ Q and B = 0, that
Vx - V^ =
(X
a) /
Q = 0,
that
is,
V2
Va.
200,
unit.
If a
206
segment S be incommensurable with the unit segment s,that is, if, as when S and s are diagonal and side of the same
square,
is
we can prove that no aliquot part of s, however small, the definition of length given in contained in S exactly 130 does not apply to S.
But there
is
a,
which
s fall
stands in the following relation to S The segments which are commensurable with the unit
into
two distinct classes, those which are less than S and those which are greater than S. The rational nimibers which are their lengths, 130, fall into two corresponding classes, which we may call Ay and JjEvery positive rational belongs either to Ai or to
A^, each
number
in A^ precedes every number in A 2, and, finally, there is neither a last number in A^ nor a first in A^*
lies
There is, then, 154, a definite irrational number, between all numbers in Ai and all in A 2. We
a,
which
call this
* For were there a last number in ^,, then among the segments commensurable with s and less than S there would be a greatest, say S'. But no such seffniont exists. For according to the Axiom of Archimedes, explained in the following footnote, we could find an aliquot part of s which is less than S S' and the sum of S' and this part of s would be commensurable
;
with
s, less
S'.
66
A COLLEGE ALGEBRA
a the length of S.
:
number
207
s,
We
S,
a,
definition
which are lengths of segments less than S and which are lengths of segments greater than S.
Thus,
Vi
is
208 209
we
write S
as,
and
134,
As
in the figure of
as origin; also take any right line and on it a fixed point some convenient unit, s, for measuring lengths. And by the of any point P of the line, understand the distance from
OP
in
terms of
s,
130, 207.
choose as the picture of any given number, a, that point, P, is the tiumerical value of a., of the line tchose distance from the point being taken to the right or left of 0, according as a is
positive or negative.
is a rational number, we can actually construct P, 134. the contrary, if a is irrational, we usually cannot conthen assume that P exists, in other words, that struct P.
We
If a
On
We
is
all
points
than a and
all
which picture
not the place for a discussion of the axioms of geometry; hut of their relation to the subject of
of
Axiom
ice
Archimedes.
ji ml
S<S,
2.
can always
an
such that
ms>S.
of continuity. //" all the points of a right line be separated into two classes, K, and K.,, siirh tliat earli point in R, //>.s" to the left of every point MiR there is either a last point in R, orajirst in R,. (i) The Axiom of Arcliimedcs is involved in the assumption tliat every For the first steji in measurinuj S in terms of line segment can be measured. is. S is to find an iiitefjer, )/;, such tliat (i 1) (2) The axioms 1 and 2 cnnlile us to prove the assumption in 209 that for
Axiom
s<S<
every given irrational a there exists a corresponding point, P. For a separates the rational system into two parts, which we may name B and C respectively. Call the points corresponding to the numbers in each the
IRRATIONAL NUMBERS
67
210
Conversely, when P is given, we can find a, at least approxi sign to mately, by measuring OP and attaching the + or the result, according as P is to the right or left of O. Thus, if P is to the right of O, and we can lay s along OP Jive times, the tenth part of s along the part left over seven times, and the hundredth
part of s along the part still left over six times, then 5.76 will be the value of a to the second place of decimals.
ence, 2,
In this manner we set up a relation of one-to-one correspondbetween all the real numbers and all points on the
;
211
line
and
if a
real
numbers, and
P and Q
the corresponding points, P will lie to the left or right of Q according as a is less or greater than b.
Thus, if a and b are positive and a < b, and if c denote a rational lying between a and b, and R the corresponding point, we have, 206,
OP<OR
and
OR<OQ
and therefore
OP<OQ.
are to show that there is in the JS-points and the C'-points respectively. line a definite point, P, which separates all the iJ-points from all the C'-points. First assign the ii-points and, all intermediate points to a class i?, and all denote the point which this points to the right of these to a class R,, and let separation defines, by 2. Next assign the C'-points and all intermediate points to a class S^ and all points to their left to a class S^, and let Q denote the point which this separation defines, by 2. The points F and Q must coincide. For if not, let denote the line segment between them. By 1, we can find an integer, m, such that
We
PQ
mPQ>s,
this is impossible. c such that to b and c respectively, we
and therefore
select
PQ>s/m.
But
For we can
have
number
cb<l/m. And if L
LM<s/m, and PQ<LM, and therefore PQ<s/m. It is this one point, P or Q, that corresponds to a according to 209.
{?>) Finally, observe that corresponding to 2 the system of real numbers has the property described in KiO, and corresponding to 1 the property //'a and b are any two positive real numbers, we can always find an integer,
:
178,
at least not without a real system would not possess this property sacrifice of some of its other properties were we to invent more than one irrational for a separation of the rational system of the kind described in 154. Thus, if every rational is either an , or an a, and Oj<b<c<02 for every a^, 178 and proof of 1()3. ctj, fflj! we should have c-'b<a^ But however small a positive number, S, we might assign, we could find no b)>S. integer, m, so great that )/((c For it would then follow that c b>6/m, which is impossible since
The
b<a2 Oj and we
can choose
a^, a^
so that
a^a^K^/m,
68
212
Theorem.
A COLLEGE ALGEBRA
If the length of S in terms of T is a, and that of is b, then the length of S in terms of s is ab.
in terms of s
1.
When
and b are
and b
rational.
Let a
Since
itself
= a/b
c/cZ,
where
a, 6, c,
S contains
T
(1)
a times, that
is,
Similarly
6S = dT =
aT.
cs.
(2)
But from
(1)
and
(2) it
hdS
atZT,
and adT
acs,
and therefore
That
2.
is,
6dS
acs.
the length of
in
terms of
s is
(XC
(X
or
b
hd
130
Wlien
and
b,
and
let ai, 02
terms of T, so that
aiT and S2
208
such
that
Ti<T<T2,
and
let 61, 62
s,
so that
Ti
61S, .and
T2 =
ios,
where 6i<b<62-
Then
since
Si
1,
OiT, and
T>Ti, and Ti = M,
Si > ai6is.
we
have, by case
Similarly
S2 < ajb^s.
fli^is
Hence
and therefore
We
and
the
numbers
lengths, in terms of s, of segments respectively less and greater than S. Therefore ab, the one number which lies between all the numbers ai6i
0262, 171, is the length of
itself in
terms of
s,
207.
/
IRRATIONAL NUMBERS
Corollary.
b respectioely,
69
213
If the lengths of S and T in terms of s are a and then the length of S in terms ofT is a/b.
For
jf s
is
let
Then
the length of S in terms of T be x. since the length of S in terms of T is x, and that of b, the length of S in terms of s is xb, 212.
in
terms
in
a,
terms of
s is a.
Hence
and therefore
xb
x = a/b.
variable.
The continuous
is
One
of our
214
to be
to
According to the assumption of 209, the segment AB contains a point for every number between a and b, through which, of course, P must pass in its motion from ^4 to B. This leads us to say that as P moVes continuously from A to B, x increases
or that X varies continuously
from the value a to the value b through all intermediate values, from a to b.
Of course
it is
impossible actually to trace the variation of this x, since its values there is no next following value. If we attempt to reason about x mathematically, we must content ourselves with defining it thus (1) x may take every given value between a and b,
to
and (2) if p and q denote any given pair of these values, and p < 5, then X will take the value p before it takes the value q. We may add that x is
often called a continuous variable
is
when only
the
first
of these properties
attached to
Ratio.
it.
Let
and
By the measure of M in terms of N, or the ratio of M to iV, we mean the very same numbers which we have defined as lengths in 81, 130, 207, when M and N denote
same kind.
line segments.
215
70
A COLLEGE ALGEBRA
Hence the theorems of 212, 213 regarding lengths hold good for the measures or ratios of any magnitudes of the same In particular, kind.
216
If the measures of
and
in terms
to
of the same
a./^).
tinit
are a
and
o/M
is
V.
217
The real system does not contain the even roots of negative numbers for the even powers of all real numbers are positive. Thus the real system does not contain the square root of 1. To meet this difficulty, we invent a new system of signs
;
called
218
The simplest of these new signs is i, called the unit of With this unit and the real numbers, a, we form imaginaries. signs like ai, which we then regard as arranged in the order We in which their " coefficients," a, occur in the real system.
thus obtain a
new continuous
which we
we may
rirst
scale of imaginaries
i,
i,
i,
0,
i,
z,
i,
system by introducing imaginaries with and finally into a, continuous system by introducing
deji.se
imaginaries with irrational coefficients. Here 2 i is merely the name of one of our
new numbers.
Its
only
property
is
But when we ordinal system. shall see that 2 i also represents the product
new
ai.
Hence we write for i. i as 0. In particular we shall define Observe that is the only number which is common to the real system
and the system
of pure imaginaries.
71
219
equations
1.
For these new numbers we invent operations which we call They are defined by the following "
:
ai
+ bi = (a
-\-
b)i.
3.
2.
a-bi
= bi-a = abi.
ai -bi = ab.
Thus, 3, the product of two pure imaginaries, ai and bi, is to mean the real number, ab, obtained by multiplying the
coefficients of ai
and
bi together
the result.
We
define
power as
in 136.
Thus, (aiy
= ai
ai.
The system of pure imaginaries contains the square all negative numbers in the real system, namely :
roots
of
220
V 1 =
For
i'^
and
i-i
V =
a'-^
ai.
= \i-\i=-l.
219,3
The refore, i is a square r oot of 1, 138. We indicate this root by V 1, and thus have i = V 1. In like manner, it may be shown that i is a square root of 1. We indicate this root by V 1.
Similarly, since (ai)2
ai- ai
a^,
we have
ai
a^.
COMPLEX NUMBERS
To secure a number system which will contain the higher even roots of negative numbers, we invent comjilex numbers. These are expressions like a + bi, formed by connecting a real
number,
a,
221
bi,
by the sign +.
They
Since
<z
and
bi
bi,
real
222
72
223
A COLLEGE ALGEBRA
We regard the complex numbers as arranged in rows and columns in such a manner that all numbers a + bl which have the same b lie in the same row and are arranged in this row from left to right in the order of their a's while all numbers which have the same a lie in the same column and are arranged in this column from below upward in the order of
;
their
6's.
And we may
its
ber defined by
ment
and b. and b.
-2 + 2i -2 + -2 2 i
-I + 2i -1 + -1 1 i
i
2i
i
1 l
+ +
2i
i
+2i
2+i
2
2
-2-2i
This arrangement
-l-2i
may
-2i
i
l-2i
2-2i
also be described as
whose elements are rows (or columns), each system of signs of the form a + bi.
of
224
Definition of equality.
erjual
are said to be
the same position in the two-dimensional ordinal arrangement jiist described. Pleuce,
225
If a
bi
=c
-f di,
In particular, iia+bi
Of two unequal complex numbers, like 2 + Hi and 8 + i, we cannot that is, precedes or follows say that the one is less or yreater than the other, since complex numbers do not form a simple ordinal system.
226
The sum^
are to
and product of two complex numbers a + hl^ c + di, mean the complex numbers which form the second
of the following equations
:
members
73
(a
+ ^-0 + (c + (U) =.(a + c) + (b + ^i. - (c + dl) = (a -c) + (b -d) {a + hi) = (ac hd) + (ac? + he) (a + ^0 (c + di)
i.
i.
According to 1 and 2, addition and subtraction are inverse operations. In particular, by 1, (a + i) + (0 + 6i) = ( + 0) + (0 + 6) z = a + f'i that of a and hi, according to the definition 1. is, a + hi is the sum These definitions are in agreement with tlie commutative, associative,
;
and
distributive laws.
In fact,
we
arrive at
Thus,
hi) c
-\-
hi) (c
di)
{a
-\-
{a
hi)
di di
since
i^
1.
Corollary.
A product
+
&i) (0
227
For
(a
+ ^) =
0)
(a
+
-\-
0)
0.
Division.
We
hi
hj
-{-
di as the
228
complex number which multiplied hy c + di will give a + hi. When c + di is not 0, there is one and but one such number,
namely, that in the second member of the equation
a
c
rf2
But when
+ di
is 0,
no determinate quotient
of 220.
exists.
For the product of the right member of a + hi, as the reader may easily verify by aid
this equation
by
di
is
We
If
number
is
+
a
a
h.
di will give a
hi, let it
be
Then
or
{ex
(x
+
-\-
yi) {c
+ +
di)
= = =
+ +
hi.
(1)
(2)
- dy)
=
{dx
cy)i
cij
hi.
and therefore ex
dy
a and dx
(3)
225
x and
y,
.
we
obtain
.,
ac
hd
5
he
c2
c-2-i-d^
ad
J
unless c^
+
,
a^
j.>
a 0.
h\
(4)
tZ^
74
Moreover, since
A COLLEGE ALGEBRA
(4)
is
the only pair of values of x and y which will number x -\- yi is the only number which
l>i-
c^ -\- d- = our definition of quotient and d 0, since otherBut if c- + d^ = 0, both c = meaningless. and wise we should have a positive number equal to 0. And if c = d = 0, the divisor c + di is 0.
evident from
(4)
that
when
is
229
The commutative, associative, and distributive laws. The operations just defined evidently include the corresponding operaLike the latter, thei/ conform to the tions with real numbers.
commutative, associative, and distributive
Thus,
(a
(6
laivs.
a'i) (b
b'i)
b'i)
ah
a'b' b'a'
+ +
{aV
(h'a
a'h)
ba')
i,
(]
and
+
+
{a
+
+
a'i)
ba(2)
{b
i.
(2)
of (1)
and
are equal,
b'i)
by
177.
a'i) (b
b'i)
{a
a'i).
And
230
may
be established similarly,
Rules
Let
a
a, b, c
If
If If If a
3. 4.
1.
= b, a -f c = b + a = b, ac = be,
a
then a
c,
then then
+c=b+ a = b.
ac
c.
= be.
= b,
=c+
b'.
then
and
b'i,
a
c
unless c
c'L
= 0.
For
let
a'i,
If
a
a
a
{a
= b + b'i, + a'i = b +
and
then
= =
b
b
a'
225
+
c)
178
225 226
If
a a
c
c)
we have
and therefore
3 and
4.
c-|-(
-'
=b+ =b+ = b.
c,
+ (
c),
byl
226
The proofs
and 2
respectively.
75
231
its
by the reasoning
of 76.
The
positive real
number
232
Va'-
+
is
6^ is called
+ bi
and
represented by \a
bi\.
Hence, by definition,
Thus,
|2 4-
il
= V4 +
^ Vs.
definition of numerical value reduces to that already For a geometrical interpretation of this given for real numbers, 63. definition see 239.
6
0, this
When
We
also say of
first is
numer-
233
or greater than the second, according as the absolute value of the first is less than, equal to, or
ically less than, equal
For
|2
Theorem
The
234
members
is eqxial to
= ah = =
i,
226
we have
|ab|
V(a6
a'6')-
(^'
+
it
a'b)^-
232
will be
found that
a'b'Y
^
+
+
{ab'
{(ib'
+ +
a'b^
a'by^
=
=
(a^
a'^) (62
h"^).
a'b')
= Va^ +
|a|-|b|.
a'2
Vt^+b^.
186
|ab]
Theorem
2.
The
235
Employ
Then
if
the
same notation
Va2 +
a'2
a2
-j-
62
+ 6'2^ V(a + 6)2 + (a' + + 2 V(a2 + a'2) (62 + 6'2) > a2 + 62 + a'2 + 6'2 + 2 {ab + a'b')
a'^
Vfts
(1)
6'2
184
76
.-.if
if
A COLLEGE ALGEBRA
V(a2
a262
.-.
a'W^
178
2 a6a'6'
(2)
is
184 178
.-.
if
a'262
> 2 aha'h'
/.if
{a6'-a'6)2>0.
(2) is
178
positive
But
(or 0).
always true since the square of every real number which proves our theoi'em. Hence (1) is always true
Thus,
[2
+
+
i|
= V5 +
1.
and
|1
3i|
= V]L0.
Vs + VlO.
of a
But
(2
i)
(1
3
1
5,
and 5 <
236
(a
Powers and
roots.
The
?ith
lyoxver
hi,
written
hi)'",
is
the product of n factors each ot which It follows from 226, 3, that this product is a
to
c
mean
di.
185, that the laws of exponents hold good for powers of complex numbers as thus defined. If {a + hiy = c + di, we call a + hi an 7ith root of c + di, 2.
It
may
be proved, as in
it
thus,
di.
We
shall
c
number
prove subsequently that every given complex di has n such wth roots in other words, that in
:
+ di.
1 are
- 1/2 = f, the number Thus, since (l/Vi + i/v^)- = 1/2 + 2 i/2 I/V2 + i/Vi is a square root of i, and therefore a fourth root of 1.
The remaining
three fourth roots of
-\/V2 +
237 system
is
i/V2,
I/V2 -i/V2,
-I/V2 -i/V2.
number
General conclusion.
For, as has just been pointed out, 226, 236, the complex system meets all the requirements of the four fundamental operations and evolution. And while certain
necessary.
other operations with numbers have a place in mathematics, among them the operation of finding logarithms of numbers, these operations admit of definition by infinite series, 140,
if
like
III
-f M2
+ W3 +
numbers
and
sum
is
a complex number.
77
Complex numbers may be pictured by points in a plane, the points being called the graphs of the corresponding numbers.
Take any two
the origin
s for
1.
;
238
also
right lines X'OX, Y'OY intersecting at right angles at some fixed unit segment
measuring lengths. We represent each real number, a, by that point A on X'OX whose distance from 0, in terms of s, is |a|, 209, taking A to the right or left of O, according as a
is
positive or negative.
2.
We
by that point B on Y'OY whose disB above or tance from is \h\, taking below O, according as 6 is positive or
hi,
negative.
A
in
represent each complex number by the point P, which is obtained by the following construction Find A and B, the graphs of a and 6i, as in 1 and 2. Then through and B draw parallels to Y'OY and X'OX respectively. The point T which these lines meet is the graph of a + bi.
3.
We
6t
We
By
call
X'OX
and
Y'OY
the axis of
pun
imaginaries.
this
of complex
numbers
into a relation of one-to-one correspondence, 2, with the assemMoreover we obtain a complete blage of all points in a plane.
part
The
its
graph from
number
is
the distance of
239
are a
and
78
240
a
A COLLEGE ALGEBRA
of the siim
The graphs
a'i,
=^+
b'i,
may
Given
and B,
the graphs of
a and b
respectively.
Join
OA
and
is
OB
the
the
parallelo-
Then C
b.
BD,
a,
a',
Then
are the
lengths of OE,
respectively,
gles
EA, OD,
DB
ODB
a
a'
congruent.
OF = OE + EF = OE + OD = EC = FG + GC = EA + DB =
Therefore
+ b. + b\
in length,
in length.
When
drawing
Since
AC OC<OA + AC,
the graph of a + 6 + (a'+ b')i, or a -|- b, 220, 1. are in a straight line (and always) C may be found by 0, A, equal in length and direction to OB.
is
The graph
of the difference a
-|-
b|<|a|
|bl.
+ (
b).
Given A and B, the graphs of a and b, and let I denote the graph of 1. Join OA, OB, lA, and on OB construct the triangle OBC similar to OIA and such that were OB turned about O
until
it
lay along
OX, OC would
is
lie
along OA.
Then C
the
graph of ab.
from
it
for constructing
graphs of quotients and porvers. When b = i, OC is OA turned 90 " counter-clockwise " about 0.
241
tions
may
Hence
imagi-
nary numbers
may
express relations
among
real things.
- a)/ 2 shows that the diagonals Thus, the identity (a + b)/2 = a + (b of a parallelogram bisect each other; for the graphs of (a + b)/2 and in tlie first figure, 240. a + (b - a)/2 are the midpoints of OC and
AB
PRELIMINARY CONSIDERATIONS
242
number
the same as the product of the second by the first. In many algebraic discussions we find it convenient to make
is
the following distinction between two letters which have this meaning, as between the letters b and x in the expression x -{-b. First. regard the one, b, as having had a paHicular
We
value, but
it is
we
any that we please, assigned it at the outset, which Such a letter then to retain throughout the discussion. shall call a known letter or number, or a constant.
Second.
On
the contrary, throughout the discussion we x, as free to take every possible value and
Such a
letter
we
243
Unknown
letter
letters.
But
employed
be found.
to denote
Such a
we
shall call
are not at liberty to assign any value we please to an unknown letter, as we are to a constant or variable letter.
Thus,
in the
We
equation 2 x
0,
is
an unknown
letter
whose value
In the expression 2x readily find to be 5/2. any value we please, but in the equation 2x 5 =
we
79
80
244 The
A COLLEGE ALGEBRA
choice of letters.
is
The only necessary restriction on onr that no single letter be made to represent two different numbers at the same time.
choice of letters
It is customary, however, to represent known numbers or constants by the earlier letters of the alphabet, as a, b, c unknown numbers and variables by the later letters, as x, y, z.
we sometimes use
a',
letters
affected
a", a'"
read
"a prime,"
a^,
a^
245
On reckoning with
letters, as a,
h, c,
letters.
we can
them by the operations of arithmetic. Thus, to add b to a merely means to form the expression a + b, which we therefore call the sum of a and b. Similarly, the product of a by
Z
is
the expression ab. Inasmuch as the literal expressions thus obtained denote numbers, we may reckon with them by the operations of
arithmetic.
But
in
can merely connect the expressions by the the form of appropriate signs of operation and then simplify the result by changes which we know will not affect its value,
not given.
no matter what this
We
may
be.
Now, as we have seen, 68, all the changes that can be made in the forms of sums and products without affecting
their values are
1.
embodied
in the following
formulas
3.
It may therefore be said that the formulas 1-5 are practi-^ of addition and multiplication that we cally all the definition either need or can use when combining literal expressions and
;
the like
is
PRELIMINARY CONSIDERATIONS
Thus, to add 2x
to
81
+ Sy and 4 x + 5 ?/ merely means to find the simplest which the expression 2x + Sy + {ix + by) can be reduced by We thus obtain applying formulas 1-5, and adding given numbers.
form
2/
2/
= 2x + 3?/ + 4x + 6j/ = 2x + {4x + 3y) + 52/ = (2x + 4x) + (3?/ + 5y) = 6 X + 8 y, the sum required,
by by 2 and
2
1
by 2
by 3 and 5
may
regard
246
addition, subtraction, multiplication, division, involution, evolution as defined for algebra by the following rules
and and
formulas, which
we
fundamental
rules
numbers whatsoever, and the sign of equality, =, means "represents the same number as."
finite
Addition.
The
-\-
b.
We
call this
247
a value, and but one, for any given values of a and b. In particular, a-\-0 = 0-\-a = a. Addition is a cnmm^itative and an associative operation that
;
248
is, it
+b=b
If
-\-
a,
-\-
(b
-\-
c)
= (a
-\-
b)
c.
39
249
a
-{-
b,
-\- c,
then a then
If a
Subtraction.
c ^= b
+c=b a = b*
-{- c.
there
is
250
we
82
A COLLEGE ALGEBRA
call this
from which a can be obtained by adding b. We bar the remainder which results on subtracting
from
we represent
it
by the expression a
(a
b)
-\-
= a.
b
a,
numand
b.
Hence, by definition,
In particular, we represent
251
Multiplication.
of
by
b.
The
result
multlphjlng a by i
is
the
a by b. It has a value, and but one, for any given values of a and b. = a = 0, whatever finite value a may In particular, a
expression
ab.
We
have.
When
252
Zi
is
Multiplication
and it is distributive with respect to addition forms to the three laws, 45-47
:
that
is, it
con-
ab
= ba,
If
a (be)
= (ab)
c,
(b
-}-
c)
ab
ac.
253
75,
76
then ac then
then
If ac
If ac
be, = 0,
is
= be. a = b, unless a = 0, or
c c
= 0.* = 0.
^
254
Division.
This
12A.
Given
any two numbers, a and b except when b is 0, there is always a number, and but one, from which a can be obtained by We call this number the quotient which multiplying by b. results on dividing a by b, and we represent it by the expression - or a/b.
Hence, by definition,
e
255
* Later
- \b
a,
except
when
0.
Involution. This is a case of repeated multiplication. to n factors by a" and represent the continued product a a call it the nth power of a.
We
we
good when
c is infinite.
PRELIMINARY CONSIDERATIONS
In the symbol
83
a", n is called the exponent, and a the base. Involution, or raising to a power, conforms to the following thi'ee laws, called the laws of exponents, 185
:
256
a"'
n a"
am + n
(')
a""",
(ab)"
a^b""
The following
If If
184
257
= a" =
a
b,
b",
then a"
then
= a = b,
ft".
or a
b.
The second
is
of these rules and the general rule of which a particular case will be demonstrated later.
it
Evolution.
This
is
258
Given any jjositive number a, there is a positive number, and but one, whose nth power is equal to a. We call this number the p)'>'^ncipal nth root of a, and we represent it by v a,
140.
or,
when n
2,
by va.
Hence, by definition,
(^y =
a.
But this positive number, Vo, is not the only number whose n\\\ power is equal to a. On the contrary, as will be shown subsequently, there are n different ininiV)ers whose th powers are equal to a and this is true not only when a is
;
when a
is
When
is
positive
and n
is
va.
reversibility of the preceding rules.
tlie
On the
certain of
We
have called
259
rules just enumerated rules of equality ; call the rest rules of combination.
we may
Observe that
all
equality for sums are reversible, but that the rules of equality for products and powers are not completely reversible.
Thus, according to the distributive law, a{h -\- c) = ah + ac, which is one of the rules of combination, we can replace a{b + c) by ab + ac, or reversely, ab + ac hy a (b + c).
84
Again,
if
A COLLEGE ALGEBRA
a
if
=
if
6,
that a
c i= b
c,
and
reversely, that,
c,
then a
b.
But while,
trary,
c is
if
= 6, we may
=
ac
be,
we can conclude
6
it
always conclude that ac = bc; on the conthat a = b only when we know that
not
0.
And
260 The
while from a
b or
=
b
b^,
from
a^
is
b^ it
rules of inequality.
The formula a
real numbers, a
^
and
means
" a
not
equal to b."
b,
the one
is alge 62.
braically the greater, the other algebraically the lesser, If a is the greater and b the lesser, we write
"> b or b <. a.
In particular, we have a
261
positive or negative. For any given real
>
or a
<
c,
0,
according as a
is
numbers
a,
b,
we have the
rules,
178, 184:
1.
If
If
If
2.
=b =b a < b
a
as
and
and
and
= < b <
b
c
c,
then a
then a
=
<. <.
c.
c,
c,
c.
then a
c.
According
a <, =, or
so
is
<, =, or
and
but
3.
ac <, =, or
ac
>, =,
or
r,
When
according as
so
is
a <, =, or
a"
<, =,
or
or
and
Va
<, =,
This
c,
is
out, the rules under 2 and 3 hold good of imaginary numIf a = also true of the rule and b = c,
ft
then a
which we may
of equality.
PRELIMINARY CONSIDERATIONS
ADDITIONAL ALGEBRAIC SYMBOLS
85
262
employed
Various
sir/ns
employed above, and [ ], |, to indicate that the expression included by them is to be used as a single symbol. 2. The double signs , read "plus or minus," and q:, read "minus or plus."
Thus, in a
6 =f
c,
-\-
c,
the upper
.*.
than
for not greater than ; <^ for not Also, . for smce ; ^ for greater or less than.
less
ALGEBRAIC EXPRESSIONS
Any expression formed by combining letters, or letters and numbers, by the operations just described, is called an algebraic
expression.
Note.
263
The number
an expression 1 + X + x^ +
may
of times that an operation is involved in such be limited, as in 1 + x + x^, or unlimited, as in supposed to be continued without end. In the one
264
we say that the expression is finite, in the other, infinite. present we shall have to do 'vith finite expressions only.
case
For the
It is customary to classify algebraic expressions as follows, according to the manner in which the variable (or unknown) letters under consideration occur in them
:
265
expression is called integral if it does not involve an indicated division by an expression in which a variable letter
occurs
;
An
266
fractional, if it does.
86
Thus,
if
A COLLEGE ALGEBRA
X and y are the variable
axletters,
but
a, &, c
constants,
then
+
X
1
hx
and b
+ Vx
are integral,
.
but
2/
and
1
+x X
are fractional.
..
267
An expression is called rational if it does not involve an indicated root of an expression in which a variable letter
occurs; irrational,
Thus, a
if it
does.
+ Vbx
1.
is
rational, but
V^ + y/y
is
irrational.
268
Notes.
we suppose
it
reduced to its simplest form. Thus, Vx^ + 2 xy 4- y'^ is rational, since it can be reduced to the rational form x -\- y. 2. The terms integral, rational, and so on, have nothing to do with the numerical values of the expressions to which they are applied. Thus, X + 2 is a rational integral expression, but it represents an It represents a fraction for integer only when x itself represents one.
every fractional value of x, and an irrational value of X.
number
269
When
an algebraic expression
is
with the signs immediately preceding them are called the terms of A.
connected by
or
signs, these parts
made up
of certain parts
a+
a^c-{b +
c)
[d
+ e]- {f+g} +
+ i+j\- n +p
+m
are a, a-c, {b + c), and so on, those of the terms which themselves consist of more terras than one being enclosed by parentheses or some
1.
270
Integral expressions are called monomials, binomials, trinomials, and in general polynomials, according to the number of
their terms.
271
is
At
the
same
time,
it is
proper to
PRELIMINARY CONSIDERATIONS
In every monomial the coefficient should be written
coefficient is expressed,
it is 1.
87
"When no
first.
Thus,
1 is
272
2 x'^y
like terms.
is
The
degree of a monomial
the
sum
273
the variables are x and y, the degree of 4 ab'^x^y* 6, zero (see 595).
is
seven
that
The degree of a polynomial is the degree of its term or terms of highest degree and the degree of any integral expression is that of the simplest polynomial to which it can be reduced.
;
274
cy^
dx-
ey
+ / is
three ;
and the
It is convenient to arrange the tenants of a polynomial in the order of their degrees, descending or ascending, and if there are several terms of the same degree, to arrange these in the
275
274.
polynomial
said to be homogeneous
when
all its
terms
276
are of the
same degree.
Thus, 5 x^
x'^y
xy"^
+ y^
is
homogeneous.
Rational integral expresPolynomials sions in a single variable, as x, are of especial importance. They play much the same role in algebra as integral numbers
in arithmetic. In fact we shall find that they possess many properties analogous to those of integral numbers. They can always be reduced to the form of a polynoviial in x, that is, one of the forms
:
in a single variable.
277
UqX
or, as
-}-
!,
CTqCc^
aiX
a.,,
a^pc^
:
a^x"^
2^
+ ^3?
>
we
form
88
A COLLEGE ALGEBRA
where n denotes the degree of the expression, and the dots stand for as many terms as are needed to make the entire
In particular, any of them except a^ the polynomial then being called incomplete.
may be may be 0,
Observe that in each term the sum of the subscript of a and the exponent of x is the degree of the polynomial.
Thus, in 5 x^
as
x'
1,
04
2, as
+ 2 X- + X 3, we = 1, tte = 3.
or
have n
= 6,
ao
5,
ai
= 0,
a-
Oo
= 0,
278
x^
Functions.
2 or
+ y,
more
We call X -\- 2 & function of x because its value depends on that of x in such a way that to each value of x there corresponds a definite value q/" x + 2.
able.
call x^ + ?/ a function of x and y and, every algebraic expression z.f\inction of all the variables which occur in it.
in general,
we
call
279
What we have just termed integral or fractional, rational or irrational expressions in a-, x and y, and so on, Ave may also term integral or fractional, rational or irrational functions of
X,
X and
?/,
and so
on.
280
often represent a given function of x by the symbol /(), read "function of a-." We then represent the values of the function which correspond to a; 0, 1, i, by
shall
We
/(O), /(I),
/(ft).
6 + 2. And, 3, f{b) Thus, if /(x) = X + 2, we have /(O) = 2, /(I) in general, if /(x) represent any given expression in x, /(6) represents the result of substituting h for x in the expression.
When dealing with two or more functions of , we may represent one of them by f{x), the others by similar symbols,
as F{x^, <t>(x), </'(^)In like manner, we
ables,
may
X and
y,
by the symbol
PRELIMINARY CONSIDERATIONS
EXERCISE
1.
I
89
X,
the degree of x-yz^ + 2 x^y^z^ + 3 x'^y'^z^ with respect to y, and z separately ? with respect to y and z jointly ? with respect
is
What
to X, y, z jointly ?
2.
What
is
the degree of (x
3.
Given 3 x'^
n, ao, ai,
If
/(x)
2 x3
{x2
If
/(x) /(x)
=
= =
?/)
- x2 + - 3x + +
3, find /(O),
/(
1),
/(3), /(8).
2)/{2x
5),
find/{0),/(- 2),/(6).
6.
If
+ V;^ +
3,
7.
8.
If /(x)
If /(x,
2X
what
X
/(x
2) ? /(x^
1) ?
:
x^
?/
8, find
1),
the following
/(I,
1),
/(0,0),
/(I, 0),
/(O,
/(-2, -3).
281
258,
we say that A is identically equal to B. The notation A = B means ".4 is identically equal
Thus, X
(X
to
B"
2)
is
identically equal to x2
2 (x
2).
For
X (X
2)
+ 4 = (x2 +
2 x)
= x2 +
(2x
4)
= x2 +
2(x
2).
248,252
We
An
slo7i,
call
=B
Hence
282
identical equation
A=B
zs
283
3_8 + 2 = 4 + T-14
in
which no
90
A COLLEGE ALGEBRA
is
implied in
282.
If aox"
then
+ aix-i + + ao = bo,
a,
= aj =
boX"
bi,
+
,
bix"-'
a
+ b,
b^.
Thus,
If,
if
ax2
3X
- 3 = 2 x2 +
6x
c,
then a
2, &
3, c
=-
3.
instead of being constants, the coefficients ao, fli, , bo, 6i, denote algebraic expressions ivhich do not involve x, it follows from the
^oX" + ^ix"-! aix''-i + identity aoX" in other words, that the expressions denoted
that ao
= bo,
Ui
= bi,
by corresponding
coefficients,
ao and
bo,
identically equal.
285
similar theorem holds good of two identically equal polynomials whose terms are products of powers of two or more variables with constant coefficients.
Thus,
if
bx
+
b
cy
dx"^
exy +fy^
+
d'x'^
e'xy
+fy^ +
a',
b',
c',
d', e
e',f =/',
286
we
7/"
A=
B, then
B=
A.
A may
be transformed into
is
reversible
Thus, we
may
x2
For
2 (x
2)
= x2 +
(2
4)
248,252
If
B.
by Theorem
Hence
A=C
C ^B, and
therefore
A = B.
PRELIMINARY CONSIDERATIONS
Thus, since
91
248, 252
248, 252
x (x
x2
,
and
we have
x(x
+ 4 = x2 + + 2) = x2 + + 2) + 4 = x2 +
+
2)
2x
2
2 (X
+ 4, + 4, 2 (x + 2).
X
identity
Theorem
operation
is
3.
An
identity remains
its
an
when
the
same
ijerformed on both
members.
Thus,
A = B,
then
A + C=B
is
-\-
C,
and so on.
of
As 286, 2, and B to the same form C. The following theorem supplies another very useful method.
into B.
two given expressions, not necessary actually to transform shows, it is sufficient, if we can reduce
To prove
287
If from a sujjposed identity, A = B, a known identity, C = D, can he derived by a reversible process, the supposed identity
288
A=B
is
true.
is
is
reversible,
true,
-\-
C = -D
b b
is
derived from
C = D.
Example.
If
it
Prove that a
identically equal to a.
(1)
b.
we suppose
[(a
+
-
b=a + b=a +
and the step
b)
b]
(2)
249
But
(2) is
known
identity, 250,
Therefore
(1) is true.
That it is not safe to draw the conclusion A = B unless the process from ^ = i?to C = Z)is reversible may be illustrated thus
:
If
it
we suppose
but
=x x^ = (- x)2.
x
it
(1)
(2)
(1) is true,
is false.
Here
(2) is true,
(2) is
since
And
in fact, (1)
It is important to remember that Identity and equality. identity is primarily a relation of form rather than of value. At the same time.
289
If A and
which
may occur in
them.
92
A COLLEGE ALGEBRA
A
into B by a limited number of For, by hypothesis, we can transform = b + a, and so on. But a + b and b + a applications of the rules a -\- b have equal values whatever the values of a and b and so on.
;
The reason
later.
in
and
A =B.
have equal values for all values of the letters This will be proved subsequently.
Hence in the case of finite expressions we may always replace the sign of identity of form, =, by the sign of equality B. shall usually of value, =, and when A B, write A
We
is to
325.
ON CONVERSE PROPOSITIONS
290
Consider a proposition which has the form
If A, then B,
fully expressed then a certain other statement, B,
or,
:
(1)
is
more
If a certain statement, A,
is
true,
also true.
it is
Thus,
If
If
a figure
is
a square, then
a rectangle.
1
1,
then X
0.
291
Interchanging the hypothesis, A, and the conclusion, B, of (1) we obtain the converse proposition
If B, then
A*
:
(2)
a figure
is
a rectangle, then
it is
a square.
0,
then x
1.
292
As
the
first
true jyroposition
may
be false.
*A proposition like If A ami B, thou C, wliich has a doiihlp hypothesis, has two converses: namely, If C and B. then ^4, and If A and (\ then B. Simiare three converses; and so on. larly, if there be a triple hypothesis there
93
293
But the converse of a true proposition: If A, then B, is always true when the process of reasoning by which the conis derived from the hypothesis, A, is reversible; for by reversing the process we may derive A from B, in other
clusion, B,
words, prove If B, then A. The method of proving a proposition by proving its converse by a reversible process is constantly employed in algebra. An
illustration of this
in 288.
When
a proposition
If
yl
then B,
is true,
we
1.
call
suffi-
294
cient condition of B,
and
B
If
a necessary condition of A.
x
Tims, the
proposition
is
1,
then
(cc
(x
1)
(a;
2)
=
2)
is
true.
Hence x
(x
= 1) (x
sufficient
is
condition
that
1) (x
0,
and
2)
When
If B, then A, are true, we say that A sary condition of B and inee versa.
;
is
295
Thns, both
(1)
If
1,
then x
is
= 0,
(2) If
0,
then
1,
are true.
that X
XL
Sum
and remainder.
Let
A and B
296
sum of A and B, and by the remai7ider to be found by subtracting B from A, we shall mean the simplest forms to which the expressions A + B and A B can be
expressions.
By
the
reduced by aid of the rules of reckoning, 247-258. Some useful formulas. In making these reductions the
fol-
297
a a
3.
+ b c = a c b. c)= a + b -]-(b a (b 5.
-{-
2.
c.
4.
c')
ab
(b c) = a b c. (b c) = a b + ac.
-{-
c.
94
A COLLEGE ALGEBRA
These formulas may be described as extensions of the commutative, associative, and distributive laws to subtraction. We may prove 1 and 2 by aid of the rule, 249
:
Two
the
by adding
same expression
1.
a-\-b
c^=a
c
c-\-b.
For the
Thus,
result of adding
[(a
to each
member
is
i.
h)
c\
h,
250
and
(a-c)-\-b
2.
+ c-{a-c) + c +
c.
a-\-b.
248, 250
(b + c) = a b
b
c)
c)
-\-
c to
a,
each
member +
is a.
(6
c)]
(6
(6
+ +
250
and
a-6 c +
prove
= a 6-C + = a-6 + 6 =
:
a.
248, 250
We may
Since
3, 4,
5 as follows
h={b-c)+c,
3.
250
we
have,
4.
5.
+ 6 c = a + [(6 c) + c] c = a+{b-c) + e -c = a+{b-c). by 1 and a b + c = a \_{b c c']-\= a(b c)c-{-c = a-(b-c). ab ac = a[^(b r) + r] ac = a(b c)+ ac ac = a(b-c). by land
a
() -{-
248
250
by 2
250
252
250
Observe that
it
follows from
may
be
performed in
95
+ +
b
c c
= =
Rules
of sign.
+ c~d + e = a + c - (b + d) + e = a + c + e b d.
" " rules of sign
e
'
+
(b
e,
byl
d),
-}-
by 2 and
by 2
The
298
-\-
3.
a (
1, 2,
=a = ac. c)
c)
:
c.
2.
(
a)
c)
-\- c.
4.
= c)
ac.
"We obtain
3 at once
by setting
in 297, 3, 4, 5 respectively.
We may
prove 4 as follows
297, 5
=
Rule
2, 3, 4,
oc)
ac.
by
and 3
297,
of parentheses.
From
the formulas
248 and
:
299
rule
Parentheses
may
be introduced
in
same
rule.
Thus,
a-\-b
a-\-b
{c
e).
\b
[c
{d
e)]\
a a a
h b
+ +
[c c
{d (d
e)']
= =
e)
e.
6-l-c
Of course the parentheses may be removed in any order but by beginning with the outermost one (as in the example) we avoid changing any sign more than once.
96
300
A COLLEGE ALGEBRA
From
Rules for adding and subtracting integral expressions. the formulas of 248, 252, 297 we derive the rules
:
To odd
coefficients,
add
and
affix the
common
To subtract one polynomial from another, change the sign of every term in the subtrahend aiid add.
Example
1.
Add
ab"^
and
(i
5 a6-
also subtract
aJfi
from 4
ab"^.
We have
and
4 a62
4 a62
2.
+ (- ba) = - - 5 ab"^) =
(
Example
Add
x^
+
+
ax'^y
248
297, 5
We have
x^
axT-y
= x3 + = + = x3 +
a;3
ax^y ax-y
(a
+ + +
5)
2 ab^ 2 a63
+
+
{bx'^y
5 ab^)
bx"y
-bab^
5 ab^
299
248
bx-y
x^y
ab-
+ 2 a6' - 3 ab^
b^
252, 297, 5
Example
3.
Subtract 2
a^
a% a-b
a-'-b
from
a^
a-b
b^.
We
have
+ + -
+
+
b^ 63
(2 a-b
ab'^
= =
a' a^
2 a26
ab"-
+ ft?) - b^
299
a-b
ab-.
252, 297
When
terms,
it is
the polynomials to be added (or subtracted) have like convenient to arrange these terms in columns and
Add
a*
+ a% -
a"-b"-
and
ab'^
a"-b'^
a^b,
and
a^b"^
ab^
from the
a*
result.
We have
a^b
2 S
a%aW-
+
a6^
6*
-a^b + gi
5 a%"-
+
-I-
a6
- 4 a262
2a65
- &*
97
EXERCISE n
-Qax-y,
5 bx'^y,
1.
2. 3.
4.
Add4ax2y,
and
Sbx-y.
+ 2 a - 62, Sa + b^--2a^, and b^-4a-ia^. Add3x2-5x + 6, x2 + 2x - 8, and - 4x2 + 3x - 7. i a^ 5 a* - 6 a62 - a% Add 4 a3 + a"-b - d b\ + 10 6% and 6 6' - 15 a62 _ 4 a26 - 10 a^.
Add
7 a2
5.
6.
Subtract
4a 26 + 6c from
3a
c.
Subtract 2 x2
5x
from x^
6 x2
5.
7. 8.
What must
From
x^
be added to a^
y^
5 a"b to give a^
+
+
6^ p
6x + 5y + + ~
\
take the
sum
2 x2
of
-2x2-6x + 72/-8
9.
and
a
x-^
+
a
2/
9.
Simplify
Simplify
{a
b)
{2
b)\
6 (a
4b).
10.
11.
12.
?/
13.
To what should
To what should
xx*
14.
8x
5 be
added
to give x^
y"^
7 ?
9 x2
3 y be added to give
7 ?
MULTIPLICATION
Product.
By
shall
and
B,
we
mean
the jjroduct of two algebraic expressions, A the simplest form to which the expression
301
AB
1. 2.
Of
can be reduced by means of the rules of reckoning. especial importance in such reductions are
:
3.
The commutative, associative, and distributive The law of exponents a'"- a" = a m + n The rules of sign
:
laws.
a(
b)
= (
a)b
= ab;
a)
b)
= ab.
303
1. To find the Rules for multiplying integral expressions. product of two monomials, multip)ly the product of the numerical
98
A COLLEGE ALGEBRA
the latter by factors hy that of the literal factors, simjMfijing adding exponents of powers of the same letter.
or
monomials
or
have
2.
each pohjnomial, multiply each term of the multiplicand by term of the multiptlier and add the products thus obtained.
The
law
first
rule follows
from the commutative and associative laws and The second rule follows from the distributive
thus,
{a
c)
{m
-{ n)
The
first
monomials.
rule applies also to products of more than two When an odd mimber of these monomials have
;
the sign of the product is otherwise it is +. product of more than two polynomials may be found by repeated applications of the second rule.
signs,
A
Example
1.
We have - 4 a:^b^x^
2 6x*
b"^ -\- a^. Example 2. Find the product of a 2 6 and ah For convenience we arrange both factors in descending powers and choose the simpler factor as multiplier. We then have
of a,
(a2
+ ab-
b^) (a
2 6)
= =
a^
+
_
a^-b
ah^
-2a"-b-2 aW- +
268
a3
aPh
ab'^
b^.
303
The degree of the product with respect to any letter (or set of letters) is the sum of the degrees of the factors with respect to that letter (or set of letters).
This follows from 302,
degree in any product the factors.
is
1, and the fact that the term of highest the product of the terms of highest degree in
1) (x^
99
is
When
304
homogeneous.
For if all the terms of each factor are of the same degree, all the products obtained by multiplying a term of the one by a term of the other are of the same degree. Hence the sum of these products is a homogeneous polynomial.
When both factors are polyArrangement of the reckoning. nomials in x or any other single letter, or when both are
homogeneous functions of two letters, it is convenient to arrange the reckoning as in the following examples.
Example
2
n
3
a;2
305
1.
Multiply 2x3
A.
r.
- x^ +
_
I
X
J
.3
'
2
_ 2
"
1 f;
P^i'tial
products
"
corresponding to
''^"'^^ '"^' ^^ *^" multiplier, placing them so that like terms, that
'^^'
is,
Finally
terms of the same degree, are in the same column. we add these like terms by columns.
2.
2/2
Example
x2
Multiply
x"^
y'^
xy hy 2 y
-\-
x.
+
4-
2 x?/
2
?/
X
x^
+
+
2 ofiy 2 x"y
x^
4 x-y
We
2y^ 2y^
of x,
coefficients. In the reckoning illustrated in 305, the terms are so arranged that their positions suffice to indicate what powers of x occur in them. may make
Detached
1,
306
Ex.
We
use of this fact to abridge the reckoning and writing the coefficients only, and it
by suppressing x
is
always worth
while to do this
coefficients.
If either polynomial
to
100
Example.
A COLLEGE ALGEBRA
Multiply x^
2
3 x2
by
x^
+ 3 x2 -
2.
1
X -r
S4-O + "^
"''
.-.
We
Ex.
1,
arrange the reckoning as in 305, but write the coefficients only, indicoefficients.
o'
-r
v/
d
d
U y
+ +
2+6
^^
^
^j^j^
+ 09+4-
result
is
ing the appropriate powers of x in the final beginning with x^ since the sum
six
we
.six, is
also indicated
by the number
277.
of
1+0-9 + + 12-0-4,
This is called the method of detached coefficients. It applies not only to polynomials in a single letter, both arranged in but also to descending or ascending powers of that letter,
homogeneous polynomials in two letters. For in arranging two such polynomials in descending powers of one of the letters, we at the same time arrange them in ascending powers of the
what powers
307
other letter, so that the position of any coefficient will indicate of both letters go with it.
of detached coefficients.
Con-
Example
1.
+ a% +
a262
afts
54)
(_?,)
?,5_
We
1
1
perform
the
1
1
first
member by detached
,.
.
and
,.
r
1
m
.
descendmg powers
b.
of
a and
m
.
ascendnig
powers of
"^ We know in advance that the degree of the in product is five, which is also indicated by the number of terms, six, Hence the product is the final result.
a5
+
2.
a^6
a''62
a26
0- ah*
b^,
or a^
b^.
Example
identities
(aS
a^-b
-ab + + ab"- -
&2) (a
b^) {a
+ +
6)
b)
= =
flS
a*
+ -
63. b*.
(1)
(2)
101
we haTS
1
(1)
(2)
1+1
1+1
1-1+1 1-1+1
1
1-1+1-1 1-1+1-1
i.e.
1,
a^
6^
+ 0-1,
i.e.
a'
- .
By
we may prove
the truth of the following identities, of which the examples are special cases, namely
:
For
value of
?i
we have
308
(a"-i 4-
a"-^
-\
h ah"-^
+ b"'^) (a-b)=a'' b\
7i,
we have
309
b)
a''-H
-\
ab"-"-
+
-
b"-^) (a
a"
+
-
b\
And
we have
310
b".
(a"-^
a^-^b
-\
[-
ab"-^
b"'^) (a
+ b) =
a"
Powers
of a
of a binomial.
We
311
As the coefficients of the multiplier are always 1 + 1, it is only necessary to indicate for each multiplication the partial We thus obtain products and their sum.
(1)
1
102
partial "products"
A COLLEGE ALGEBRA
and
so'
312
To any coefficient in a power already found add the which precedes it ; the sum will be the corresjyonding
in the next power. All the coefficierits of this next power, excejit the first can he found by this rule ; these are 1 and 1
Thus, the coefficients of
3
(4)
coefficient
coefficient
and
last,
which correspond
obtain 4
to 3, 3, 1 in (3) are
or
4,
+
5.
3 or 6,
3 or 4.
to (4),
we
or
5,
4 or 10, 4
-t-
6 or
4 or
(a
Hence
6)5
a5
5 a*b
10 a^b^
10
a"-b^
+ 5a +
.
b
Evidently the coefficients of any given power oi a be obtained by repeated applications of this rule.
Example.
Find successively
of
(a
-\-
can
6), (a
+ Wi
the
[c-
^)^-
313
Products
two binomial
factors
of
first
degree.
The
kind by inspection.
{x
We
b)
have
+ a) {x +
aj)
(^0^:;
X'
{a^x
^i)
=^ o^o-^^
(1)
-f a^b^.
(2)
In the product (1) the coefficient of x is the stim and the final term is the product of a and b. In the product (2) the first and last coefficients are products
of the
first coefficients
and of the
coefficient is the
sum
(x
5)
(a;
8).
+
2.
8)
x2
-f
(5
(x
8)x
40
(x
x2
- 3x -
40.
Example
(X
3y)
+
7/2
lOy).
Sy)
(X
10
2/)
x2
(3
10)X2/
30
^i
+ l3xy +
30?/2.
103
(2
+
3
3)
(4x
+
3
7).
3)(4x
4.
7)
2 -4x2
+
+
(2
+
a
4)x
8x2
+ 26x +
21.
Example
By (x-10){x-
{3a
46)
(5
-66),
x.
(7
y)(5x
Sy).
314
a^) (h^x-
bo)
i^o)^"*
+ (ai^2 +
The product
is
2^1
a polynomial in x whose degree is the sum And the coefficient of each term
may
K)
be obtained by the following rule, in which a^, denotes Oq, a-^, 02, ((3, and b^. one of the numbers
and
I^ind the diffei'ence between the degree of the product the degree of the term, and then form and add all the k equals this difference. products ajjb^ in xvhich h
we
2,
or
3,
aih^,
a-zbi,
03601
these being
all
the products
X of the form
cates
This rule applies to the product of any two polynomials in + , and ^ya-" + -\- b. It also indiOo-^"' +
any particular coefficient of the product when the factors have numerical coefficients.
Example
(aox^s
1.
how
to obtain
Find the
product
-\
\-
aja;74
aux +
ays) (&oa;o
6ix59
659X
660).
The degree of the product is 75 + 60 or 135 and 135 - 100 = 35. Hence the coefficient of x^oo is 00635 + 01634 + + 03461 + 03560.
;
x^
is
040650
O41659
+
+
+
-
O74626
O75635.
Example
2.
Find the
coefficient of x^ in the
product
6X
3).
(3 X*
2 x3
x2
is
8 X
2)
7) (2 x3
5 x2
The required
coefficient
(-
(-
3)
+ (-
8) 5
2,
or
14.
104
Example and of x23_
3.
A COLLEGE ALGEBRA
In the product of Ex.
1,
find the
coefficients
of a;"*
Example
4.
2,
and
x.
315
The following Products found by aid of known identities. formulas or identities are very important and should be carefully memorized.
(a
(a
(1) (2)
(a
b)(a
-b)=a^- b\
(3)
To
this list
may
'
be added
311
{a-\-bf
a^
+ ^ a%
b
-f-
3 ai^
yz^
(4)
may be replaced by any algebraic expressions whatsoever, these formulas supply the The simplest means of obtaining a great variety of products.
Inasmuch
following examples will
as the letters a
and
make
this clear.
Example
(3x
1.
5?/)2.
5?/)2
=
2.
{Zxf
3x
5?/
(5
{x2
(/)2
9x2
sqx?/
252/2.
by
(2)
Example
(x2
+
y'^)
xy
y-) (x^
xy
+ xy +
2/2)
-xy +
2/2)
= =
[{x2
(X2
2/2)2
2/2)
+ ?/2). - xy] +
y4.
a;-2y2
by
(3)^ (1)
Example
(X
3.
2/
2) {X
-y
+ z)(x + y - z) {x - y - z) = [x + (y + z)] [X - + z)] [X + - 2)] [X = [X2 - + 2)2] [X2 - - 2)2] = [(x2 - 2/2 - 22) -2yz]- [(x2 _ _ 22) + 2 yz] = [x2-(2/2 + z2)]2_4 2/222 = X* - 2 X2 (2/2 + 22) + (2/2 + 22)2 _ 4 y222 = X* + + 2* - 2 x22/2 - 2 y2z2 _ 2 22x2.
(2/
(2/
(2/
z)]
(2/
(2/
2/2
2/*
105
Observe in particular that by this method we may derive from (1) and (4) the square and cube of any polynomial.
Thus, we have
(a
+ +
+ +
c)2
= =
[(a a(a
+ 6) + c]2 = (a + + 62 + + 2 ab +
c'-
6)^
2 (a
6) c
c2
2 ac
2 6c.
(a
c)3
= =
6)3
+
c3
3 (a
hy^c
3 (a
6) c2
c^
a3
+ 63 +
Generalizing the
first
we have
the theorem
jjolynom'ial
equal
to the
sum of
the
316
squares of all its terms together with twice the products of every two of its terms.
1.
(a
(1
2c
Zd)\
2.
3.
y^)^.
Powers
of
monomial products.
By
the
7ith.
algebraic expression, A, we shall mean the simplest form to which the expression A" can be reduced by the rules of
('")"
a""'
and (ab)"
a"b''
we
To raise a monomial
numerical of each
ing as
literal factor
exjyression
to the
nth
2)oive)',
raise its
^18
coefficient to the
by n.
If the sign of A be
or , accord-
is
even or odd.
Thus,
2 ax-^if)*
= (
2Ya*xy"-^
16 a^x^yss.
a"6
(o6)"
we have
(-
2 ax2y7)4
= (_
2)4a4 (x2)^
{a'")"
(?/7)4,
of tlie
law
a'""
we have
(-
= lea^xV^.
106
A COLLEGE ALGEBRA
EXERCISE
m
also the identities of 315.
In the following examples perform each multiplication by the most In particular employ detached coefficients expeditious method possible.
where
1.
this
Multiply 3x5
Multiply 5 x3
_ -
2x*
xs
7x2
_ 6x +
a^
_|-
5 by 2x2
- 3x +
2 a2.
y.
1.
2.
3.
3 ax2
2 a-x
by 3 x2
ax
Multiply x5
4.
5.
Multiply 3 x3
Multiply
7
.X
6.
Multiply a2
Multiply X*
ax +
6x
x2
by
i)
x.
x'-^
7.
8.
5x2
Multiply 2 x"
3 x -2
- X. by 3 + x -3 by x-2 _ ^n-s. 5
x*
9.
Multiply a2 Multiply x
Multiply
10.
11. 12.
Multiply
Multiply
13.
14. 15. 16. 17.
18.
19.
Multiply x
Find the
Also
Also
(x2
{x
+
X
7?/
+ +
1) (x2
+ +
1) (X*
x2
+ +
1).
z)
{- X
z) {x
z) (x
- z).
+
x
powers of x2
+ +
1.
Continue the table of coefficients of successive powers of a as far as the tenth power.
20.
21.
Find
(4 {x
3 y)2 and (4 x
3 y)^
22.
23.
3 z)\
24.
Find the
coefficients of x29
and of
x^^ in the
product
(aox2'
aix26
a26X
a^i)
{bcfic^^
bix^^
-\
6i8X
619).
107
Find the
and
x* in the product
(2x6
26.
-3x5 +
4x*
7x3
2.
3.
4.
+ y3 ^ z3) = 3(y + z){z+x){x + y). - ay)'^. (a^ + ft2) {x2 + 2/2) = (ax + byf + (6x - 62) (a;2 _ ^2) = (a^ + 5^)2 _ (j^; + ayy.^ (a2 (a + 6 + c)3 = a3 + 63 4. c3 + 3 a2 (6 + c) + 3 62 (c + a)
(X
2)3
(x3
+
27.
3c2(a
6)
6a6c.
(_x5y829)7^
(a26'"c3)2,
:
(a'"6"c2")".
28.
a62c3) (a36)2
ac3)5,
_
(
2 x'^y^f (axSyiija,
DIVISION
algebraic expressions the quotient of A divided by B, we shall mean the simplest form to which the fraction A / B can be reduced by the rules of reckoning.
of
Quotient.
Let A and
319
which
is
not equal to
By
Formulas.
320
namely,
ac
1.
2.
=
a"
a""",
when
a
m>
ti
;
=
a"
_,^^ a"
>
when w
a
> wi.
3.
a
4.
a +b=d
d
rule,
We may
Two
expression
prove
1, 3,
253
any third
0) are equal.
108
For
in 1 the
A COLLEGE ALGEBRA
product of each member by be
is ac.
Thus,
bc =
be
~ be ac; and b
~ b
b
ac.
254, 252
Again, in 3 the product of each member of the first equaby b is a, and the products of each member of the second and third equations hj b are a and a respectively.
tion
Thus,
-^6 = -a;
b
and
(--\b = -^b=-a.
\
b/
298,254
a. -f- ^
member by
~ d
c?
is
"^^d =
d
b;
(^ -^^\d ^^^d +
\d
d/
d
b.
254, 252
The formula 2
Thus,
if
is
1.
m > n,
= a-"
a"'"
a"
a".
256
Hence
a"
a"*".
by
321
Rules
for simplifying
A/B.
The formulas
1, 2,
and 3 give
A/B.
2.
Cancel all factors covimon to numerator and denominator. When numerator and denominator involve different jyoicers
letter (or expression^ as factors, cancel the lower subtract its exponent from that of the higher power. Give the quotient the -{-or sign, according as the
of the same
power and
3.
signs.
^
ca-
6a5-2
ba\ and
-^ = _ J_ = _ 1.
a'
a"-2
a*
322
Rules for dividing by a monomial. From the definition of division and 320, 4, we derive the following rules.
To divide one monomial bij another, form a fraction by 1. writing the dividend over the divisor, and simplify.
109
To divide a polynomial hy a monomial, divide each term of and add the quotients so obtained.
a^h'^c -^
6 alPd
8 a%'^c
Gad
4 a-c
,
Hb^d
of signs.
by cancelling the
common
factor 2 ab-
Again,
(ax^
4 a^x-)
ax
=
ax
= x^
ax
4 ax.
But when
regard (a
common with a and b, we cf "has no factor in b)/d as a simpler form of the quotient than
a/d +
b/d.
If
A and B
quotient
are
is
323
common
factors,
the
A/B
x'^
reduces
when
Thus,
if
^ = x2 -
?/2,
B=
2
+ 2xy +
y^,
the quotient
is {x
- y)/(x + y).
For
^_
B
x2
x2-y2
^
y-^
^x
y) (x
y)
x x
xy
(X
y)2
y y
The process
-^ b x c means j c, means a /be. In the chapter on fractions we shall consider complex expressions in which a number of indicated multiplications and divisions occur. In particular we shall find that
Observe that a
324
(b
c),
ax(bxc-hd)=axbxc^d.
a-h(bxc-r-d)=a-^b-v-cxd.
In (1) the signs
(1)
(2)
x and -=- within the parentheses remain unchanged when the parentheses are removed but in (2) each X is changed to -;-, and each -f- to x
;
.
(6
-a){c-
SIMPLE EQUATIONS
the "
111
unknown
letter
" x
is
to satisfy.
of restricting x to values which satisfy this condition, being 4 and a; 6 are the same and true when the values of 3 a;
then only.
Similarly x
-\- ij
is
unknown
letters
x and
y,
When
the exjjressions
A
B
and
A=B
B
326
a conditional equation. This equation means : " A and And it restricts the are supposed to have equal values."
is
variable letters in
is true.
and
to
The
letters
=B
thus restricts
are called the xmknoivn letters of the equation. In what follows, the word '' equation " will ditional equation."
If the only letters in an equation are the
mean
" con-
unknown
letters,
327
as X, y,
z,
we
call it
known
Thus,
letters, as a,
2x
32/
= 5isa numerical,
but ox
?*?/
c is
literal
equation.
A literal equation does not restrict the values of the known letters.
A and B are rational and integral with respect to unknown letters, the equation .1 = J5 is said to be rational and integral. But if A or B is irrational or fractional, the
If both
328
the
equation
is
No account
Thus,
taken of numbers or
c is
known
v2x + y/h =
simplest form,
In the case of a rational integral equation reduced to its 340, the degree of the term or terms of highest
is
329
degree
The degree
b is five. Thus, the degree of ax^ + 6x = c is two ; that of x^z"^ + y* is measured with respect to all the unknown letters, but
112
330
A COLLEGE ALGEBRA
first degree are often called simple or linear those of the second, third, fourth degrees are called
Equations of the
equations
;
equations respectively.
letter, as x, restricts
a;
331
An
finite
equation in
07ie
unknown
its
to a
number
of values.
We say that
solutions or roots.
Hence
332
A root of
which,
if
an equation in x is any number or knoicn expression substituted for a;, will make the equation an identity.
_^
Thus, 1 and 2 are roots of the equation x^ and (-2)2 + (-2) = 2. 6 is a root of x + 6 = a for (a Again, a
;
for
P+ 1=2
6)
+ 6 = a.
it
333
An
for
may
state a condi-
finite number can satisfy the equation x + 2 = a; + 3. In every equation in x which has roots, x is merely a sym.bol for one In fact the equation itself is merely a disguised or other of these roots.
Thus, no
2.
obtained by replacing
X by each root
Thus, x2
in turn.
and (-2)2
+ X = 2 is merely + (-2)^2.
P + 1 =2
ON SOLVING EQUATIONS
334
To
solve
roots, or to
is
is
illustrated
Example
is
1.
3x
4=x +
fi.
Starting with the supposition that x has a value for which this equation true, we may reason as follows
:
If
then
or
4),
(2)
and therefore
Hence,
if
3x
= =
6,
then x
5.
(4)
114
conclude that c
is
A COLLEGE ALGEBRA
a
root
;
and
it is
made up of reversible
336
It
is
ste2)s.
important to remember that the mere fact that a from an equation by the
must be
rules of reckoning does not prove it to be a root. reversible to Avarrant this conclusion. Thus, from
The process
it
follows that
and hence
x - 2 = 0, - 2) (x - 3) = 0, that either x = 2,
(x
(1) (2)
253 253
or x
3.
(3)
But we have no right to draw the absurd conchision that 3 is a root For when x = 3 we cannot reverse the process, tliat is, divide (1). both members of (2) by x 3, since the divisor x 3 is then 0. On the other hand, when x = 2 we can reverse the process, since x 3 is then not but 1 and 2 ts a root of (1).
of
;
TRANSFORMATION THEOREMS
337
In the light of what has just been said we may regard any correct application of the rules of reckoning to an equation as a legitimate transformation of the equation and if such a
;
transformation
the roots of
we may conclude that it leaves the equation unchanged. Hence the following
is reversible,
theorems.
338
Theorem
leave
1.
its
1.
roots
259, to each
member
separately.
2.
members, or subtracting
3.
Adding any expression which has a finite value to both it from both. Multiplying or dividing both members by the same constant
all
{not 0).
For
reversible, 259.
We
may
and 3 as follows
SIMPLE EQUATIONS
If
115
A and B denote expressions in a;, the roots of the equation A = B numbers which substituted for x in A and B make A = B, ^ 332. But any value of z which maizes A = B and C finite will make A + C = B + C, and conversely, 249 hence the roots oi A = B are the same as those oi A + C B + C. Again, if c denote any constant except 0, any value of x which makes A = B will make cA = cB, and conversely, 253 hence the roots oi A = B are the same as those of cA = cB.
are
;
Thus, in
334,
Ex.
1,
the equations
a;
(1)
(-a;
4),
(2)
(3) (4)
10,
5.
have the same root, 5. Here (2) is derived from (1) by the transformation 2, (3) from the transformation 1, and (4) from (3) by the transformation 3.
(2)
by
of an equation
339
sign changed,
occur
from one
may
tVi
both members.
Changing
For 3 is equivalent to multiplying both members by 1. And 1 and 2 are equivalent to subtracting the term in question from both members of the equation.
Thus,
if
we subtract we obtain
The
to transpose second.
a
a
-\-h
a
c
+ + +
6
b
a. (1)
(1)
(2)
both members of
first
and
a,
with
its
member
to the
By
340
integral equation in x may, without changing its roots, be reduced to the standard form
aoX"
+ aiX"-^
-\
h a_iX
+ a =
0.
116
A COLLEGE ALGEBRA
its
suppose such an equation reduced to this form when The like is true of rational is measured, 329. integral equations in more than one unknown letter.
degree
Thus, x2
We
3a:
= x2 4x +
2 = 0.
Its
degree
is
341
Theorem
When
A, B, and
AC = BC
has the same roots as the two equations
A=B
For any value
or
of x
and C
0.
which makes
AC = BC
must make
either
=B
C=
A = B or
C=
will
make
AC = BC,
251, 253.
In this proof it is assumed that A, B, C have finite values for the This is always true when, as is here supposed, values of x in question. A, B, C are integral but it is not always true when A, B, C are
;
fractional.
In particular, lohen
same
AC =
has the
two equations x
Thus, the roots of the equation x^ = 3 x are the same as those of the = 3 and x = 0, that is, 3 and 0.
Similarly the roots of (x
1) (x
two equations x
and x
and
2.
342
Hence the
effect of
B by the same integral function C is to gral equation A introduce extraneous roots, namely, the roots of the equation
C=
0.
factor
is to lose
AC = BC,
0.
the other hand, in a, fractional equation, it is usually the case that no extraneous roots are introduced when both members are multiplied by the lowest common denominator of all the fractions.
On
Thus,
if
the equation be 1/x = 1 /(2x 1), and we multiply both 1 = x, whose root is 1. As 1 is 1), we obtain 2x
1)
0,
root.
SIMPLE EQUATIONS
Corollary.
117
the
The
integral equation
as the equations
A=B
A
B
and
same
roots
343
For A- = B^ has the same roots as A"^ B- = 0, 339. And since A^ - B^ = {A- B) {A + B), the equation A'^ - B- = has the same roots
the
Thus, the roots of the equation (2x 1)2 = (x 2)2 are the same as those of the two equations 2x l = x 2 and 2 x 1 = (x 2), that
is,
and
1.
Hence the
effect of squaring
344
to introduce extraneous roots, namely, the roots of the equation A B. Conversely, the effect of deriving from
is
=B
the single equation A i? is to lose certain of the B. roots, namely, the roots of the equation A
A"^
B'^
Since A^ - 5" =
it
(.4
- B) {A^~^ + A^'-^B
345
A^-^B
x
-\
h i?""'
=
=
1
jointly.
1=
(x
1) (x^
+
x-
1),
the equation x^
has the
same
and
jointly.
The theorems just demonstrated, 338-345, hold good for equations in more than one unknown letter if the word root be replaced by the word solution, 355.
Thus, by 339, the equation x + 2?/ 3 tions as the equation x 2?/ + 3 (2), that
346
=
is,
(1)
has
tlie
same
solu-
X and y which
and conversely.
same
When two or more equations have the Equivalent equations. roots (or solutions), we say that they are equivalent.
Thus,
3, the
347
equations
A =B
and
is
Again,
341, the
equation
AC = BC
A ^ B and C = 0.
But
root 3.
xJ2
(1)
and x
For
(1) also
3 (2) are not equivalent although both have the 3, which (2) does not have.
118
A COLLEGE ALGEBRA
SOLUTION OF SIMPLE EQUATIONS
348
From
in one
we may
unknown
To Then
1.
solve
a simple equation in
form ax
= b.
2. 3.
If If
= = i/" a
a,
and b
b
9^ 0, the equation
0, a7id
0,
the equation is
an
identity.
it is
usually best
by multiplying both members by the lowest common denominator of these fractions. This process is called clearing
the equation of fractions.
We then reduce the equation to the form ax = h hj transposing the unknown terms to the first member and the known terms to the second, and collecting the terms in each member,
To
verify the result, substitute
it
for
2x
Example
1.
Solve
X
(4
x).
To
or
members by
the Led.,
6.
Then
Transpose and
Therefore
Verification.
4x - 3(x -
- = - 2
q.
(4
5).
Example
2.
Sol'^e
mx + n px +
p)x = q n. Transpose and collect terms, {m Hence if m ?t p, the equation has the single root
If
(q
n)/{m p).
If
satisfies
it.
m=p m =p
and q and q
-^
11,
it
has no root.
is
n,
it
SIMPLE EQUATIONS
Example
Expand,
Cancel
x^,
3.
119
Solve
x^
(x
(a
+ a) (x + 6) = + 6) x + a6 =
(3a
(x
x-
a)^.
ax
a?.
collect terms.
Then
and therefore
6)x
a^
a6,
=
3a +
6
inspection.
349
The equation
tion, for it
then completely solved if it be a simple equacan have no other root than the one thus found.
is
Example.
Solve (x
is
a)2
(x
6)2
(a
6)2.
Evidently this
identity (6
when x =
is 6.
it
reduces to the
a)2
6)2.
Hence
its
root
roots of an equation of the form AB = 0, in which A denote integral expressions of the fii'st degree in x, can be found by solving the two simple equations ^ = and
The
350
and
B 0, 341. In like manner, when A, B, C are of the BC and A'^ B'^ 0, degree, the roots of ABC be found by solving simple equations, 341, 343.
first
AC =
x
may
Example
1.
Solve (x
is
2) (x
+ 3)
(2
5) (3
2)
= =
0.
This equation
x-2 =
Hence
its
roots are
2.
0.
Example
Solve 4 x2
lias
3 x2
7 x.
This equation
the
EXERCISE V
Solve the following equations.
1.
2. 3.
'
120
^
A COLLEGE ALGEBRA
X
5.
6.
SIMPLE EQUATIONS
tlie
121
single equation connecting the expressions thiis obtained. This equation is the statement of the problem in algebraic symbols.
We
solve
it
for x.
If the
352
For the problem not an admissible solution of the problem. may be one which imposes a restriction on the character of
the
unknown numbers,
tion in
as that they be integers, and the equaX into which the statement of the problem has been
translated does not express this restriction. Having solved the equation in x, therefore,
we must
notice
number
is
an impossible one.
Example 1. The sum of the digits of a certain number of two digits is 12. If we reverse the order of the digits we obtain a number which is 4/7 as great. What is the number ?
Here there are four unknown numbers, namely, the tens
digit,
the
units digit, the value of the number as it stands, and the value when the but all four can be readily expressed in terms of digits are reversed either units or tens digit.
;
Thus,
let
X
12 10 X
Then
(12
x= x) = +
X
10 (12
x)
By
we have
(1)
less
Solving this
-x) + x=:
than
12
10, is
The
like
is
true of
X or
4.
is
84.
Thus,
Notice that with a slight modification the problem becomes impossible. if we require that reversing the digits shall double the value of the
(1),
the equation
(12
-X) + x:=2[10x +
-x)].
(2)
122
A COLLEGE ALGEBRA
is
And solving (2) we obtain x 32/9, which being fractional admissible solution of the problem.
353
not an
When dealing with a problem which has to do with certain magnitudes, as intervals of time, remember that the letters
used in stating the problem algebraically are to represent
not the magnitudes themselves, but the numbers which are Care their measures in terms of some given iniit or units.
must
of the
also be taken to express the measures of all magnitudes same kind, whether known or unknown, in terms of the
same
unit.
Example 1. A tank has a supply pipe A which will fill it in 3 hours, and a waste pipe B which will empty it in 3 hours and 40 minutes. K the tank be empty when both pipes are opened, how long will it be before
the tank
is full ?
Let X denote the number of hours required. Then 1 /x is the part filled in one hour when both A and B are open. But were A alone open, the part filled in one hour would be 1/3. And were B alone open (and water in the tank) the part emptied in one hour would be 1/3| or 3/11.
TT Hence
X
x
113 =
3
or
11
33
Therefore
= 33/2
crew can row 2 miles against the current in a certain river in 15 minutes with the current in 10 minutes. What is the rate of And at what rate can the crew row in dead water ? the current ?
Example
2.
Let X
the rate of the crew against the current minute, in dead water it would be 2/15 + x.
As
2/15
in
miles per
And
is
would be 1/5 -
x.
Hence
whence
-j!^
+
-\-
x,
x
-^^
3.
=
=
and
= number
two o'clock
SBIPLE EQUATIONS
123
Since the minute hand starts at XII it will then have traversed x minute spaces. The hour hand starts at II, or 10 minute spaces in advance of the minute hand, but it moves only 1/12 as fast as the minute hand. Therefore when the minute hand is at x minute spaces past XII, the hour hand is at 10 + x/12 minute spaces past XII.
But by the conditions of the problem, at the time required the minute hand is 30 minute spaces in advance of the hour hand.
Hence
or solving,
x
x
= Ao +
) + 30,
43j''j-
43/3-
minute spaces.
minutes after
Therefore the hands point in opposite directions at two o'clock, or I6j\ minutes before three o'clock.
Sometimes
in the statement of a
a, b,
a, h,
c.
problem the known numThe value found for x c y^Yach. may represent an
354
admissible solution of the problem for certain values of these The discussion of the followletters, but not for others.
ing problem,
this point.
known
Example.
in the
Two
couriers
and
same
are traveling along the same road and n miles an hour respectively.
if
B
so,
is
now d when ?
miles in advance of A.
Let X
the
number
will then
of hours hence when they will be together. have traveled mx miles, and B nx miles and since
;
is
now d
miles in advance of A,
we have
mx = nx +
whence
and therefore
1.
d,
(1) (2)
(3)
{m,
n)x x
= =
d,
m n hours hence.
must be
positive
;
If
is
and
since
by hypothesis
m> n.
d, ?n, n all denote positive numbers, Which corresponds to the obvious fact that
this requires
if
that
is
to overtake
B, he must travel faster than B does. 2. At the same time we can interpret the negative value which x takes if we suppose and B were together d/(n m) as meaning that
m<
hours ago.
124
A COLLEGE ALGEBRA
m
m
= n, we cannot, properly speaking, derive (3) from (1), since 3. If n wliich is 0. Bat we can derive the process involves dividing by differs at all from n, it matters not how little. And if (3) from (1) if
in (3)
we regard m as a variable, which while greater than n is continually n) becomes a variable approaching equality with n, the fraction d/{in which continually increases, and that without limit, 510. All of which
corresponds to the obvious fact that the smaller the excess of A's rate will never over B's, the longer it will take A to overtake B, and that
overtake
4.
if
Finally,
0,
the equation
(1) is
satisfied
if
by every value of
Which corresponds
same
rate
A and B
will
always be together.
EXERCISE VI
of the digits of a certain number of the order of the digits be reversed, the number
1.
The sum
is
two
is
digits
is
14.
If
increased by 18.
What
2.
the
number
the remainder 2
3.
There are two numbers whose difference is 298. And if the greater What less, the quotient and remainder are both 12. are the numbers ?
be divided by the
4.
The tens
digit of a certain
1
number
of
two
digit.
be added to the tens digit and 5 to the units digit, the number obtained is three times as great as if the order of the digits be first reversed and then 1 be subtracted from the tens digit and 5 from
if
And
the units
5.
digit.
What
is
the
number
If 2
number and
the remainder be
multiplied by 4, the same result is obtained as if twice the number and half a number one less be added together. What is the number ?
is now four times as old as his son. If both he and his 20 years longer, he will then be twice as old as his son. What are the present ages of father and son, and how many years hence will the father be three times as old as the son ?
6.
father
son
live
7.
tank can be
in 2 hours,
tank
if it
filled by one pipe in 3 hours, and emptied by a second and by a third in 4 hours. How long will it take to empty the start full and all the pipes are opened ?
SIMPLE EQUATIONS
;
125
8. A and B can do a certain piece of work in 10 days but at the end of the seventh day A falls sick and B finishes the piece by working alone for 5 days. How long would it take each man to do the entire
piece,
9.
working alone
At what time between eight and nine o'clock do the hands watch point in the same direction ? in opposite directions ?
10.
of a
How
?
angles
11.
In a clock which
is
it is
66 minutes.
12.
What
is
B
C.
receives f as
How much
A man
of the remainder
leaves | his property and .f 1000 besides to his oldest son ; and .flOOO besides to his second son I of the sum
;
remaining and $1000 besides to his youngest son. remain, what is the amount of the entire property ?
still
If
$3500
still
//^
14.
If 2
feet
be added to both sides of a certain square, What is the area of the square
its
area
is
15. The height of a certain flagstaff is unknown ; but it is observed that a flag rope fastened to the top of the staff is 2 feet longer than the staff, and that its end just reaches the ground when carried to a point 18
feet distant
16.
staff.
What
is
purse contains a certain number of dollar pieces, twice as many If the total value of half-dollar pieces, and three times as many dimes.
the pieces
17.
is
$11.50,
how many
invests $5000, partly at 6% and partly at. 4%, so that the average rate of interest on the entire investment is 5J%. What sum does he invest at each rate ?
18.
A man
30
26
cts.
cts.
In what proportions should two kinds of coffee worth 20 cts. and a pound respectively be combined to obtain a mixture worth
a pound
?
and copper contains 2 parts of copper must be melted with this alloy to obtain one which contains 3 parts of silver to 7 of copper ?
19.
A pound
silver to 8 of copper.
How much
126
20.
A COLLEGE ALGEBRA
If a certain quantity of
liquid, it contains
added,
it
water be added to a gallon of a given if twice this quantity of water be How much water is added each time,
;
A train whose rate of motion is 45 miles per hour starts on its trip
train
from Philadelphia to Jersey City at 10 a.m., and at 10.30 a.m. another whose rate is 50 miles an hour starts on its trip from Jersey City to Philadelphia. Assuming that the two cities are 90 miles apart, when will the trains pass each other, and at what distance from Jersey City ?
22.
If two trains start at the times mentioned in the preceding examand pass each other at a point half way between Jersey City and Philadelphia, and if the slower train moves | as fast as the swifter one, what are their rates, and when do they pass each other ?
ple
is now a distance equal to 50 of her leaps ahead of a fox pursuing her. How many leaps will the rabbit take before the fox overtakes her if she takes 5 leaps while the fox takes 4, but 2 of
23.
A rabbit
which
is
in air
in
water
sum
of
money
in
his pocket
besides.
At
and each day spent \ of what he began the day with and $2 the end of the third day his money was exhausted. How
at the outset ?
much had he
26.
of a certain pyramid is a square, and the altitude of each of the triangles which bound it laterally is equal to an edge of the base. Were this edge and altitude each increased by 3 inches, the area of the
The base
inches.
What
of
is
the area of
The sum
number
two
digits is a.
If
the order of the digits be reversed, the number is increased by b. is the number? Show that the solution is admissible only when
What
and 9a
9a^6
Is
if
by
18.
Two
?
persons
there a time
so,
when
and B are now a and b years old respectively. was or when A will be c times as old as B, and
a, b, c,
when
as in 354.
127
unknown
letters, as
x and
y,
355
will be satisfied
letters.
We
is
The
like
pairs of values of these call eveiy such pair a solution of the equation. true of an equation in more than two unknown
by
infinitely
many
letters.
?/
(1) is satisfied if
we
give
2x
to y.
Thus, x
of
(1).
+ (3 - 2 6) = 3. = 0, y = 3; x =
for y,
l,
\; x
2,
?/
are solutions
under consideration, the 2, and y any value whatsoever ; in other words, the equation x = 2 then has an infinite number And the like is true of any equation which involves but of solutions. one of the unknown letters.
Note.
equation x
When two unknown letters, x, y, are = 2 means that x is to have the value
356
It is therefore natural to inquire whether there may not be pairs of values of x and y which will satisfy two given equaSuch pairs usually exist. tions in these letters.
357
when x =
Thus, both the equations 2 x + ?/ = 3 and 4x + Sy = 5 are 2 and ?/ = - 1; for 2-2 + (- 1) = 3, and 4-2 + 3(-
satisfied
1)
= 5.
358
Simultaneous equations.
certain
Two
or
unknown letters are said to be simultaneous when each unknown letter is supposed to stand for the same number in
all
the equations.
Thus, the equations 2x + y = S (1) and 4x + 3?/ = 5 (2) are simultaneous if we suppose x to denote the same number in (1) as in (2), and
similarly.
It is
equations. in X and y.
not necessary that all the unknown letters occur in every one of the Thus, x = 2, y = 3 constitute a pair of simultaneous equations
128
359
A COLLEGE ALGEBRA
are
Generally speaking, the supposition that certain equations simultaneous is allowable only when the number of
is
equations
letters.
equal
to,
number
of
unknown
since x
Thus, the two equations x = 2 and x = 3 cannot be simultaneous, must denote different numbers in the two.
360
solution of a
of values of the
system of simultaneous equations is any set unknown letters which will satisfy all the
2,
2/
= -lisa
2x +
3,
ix + 3y =
6.
361
To
solve a
is
362
process depends
is
similar to
and y we begin by supposing that x and y actually have values which satisfy both equations. On this supposition the equations may be
treated like identities and the rules of reckoning applied to them. By aid of these rules we endeavor to transform the
and
equations into one or more pairs of equations of the form X = a, y = b. If the process by which such a pair x = a, y = b has been derived is reversible when x, y have the values
a, 6,
we may
;
at once conclude
that
is
a, b is
reversible if
consists of
is
the following
.-
Principle of substitution. Tf from the. supposition that all the given equations are actually satisfied it follows that the values of a certain pair of expressions, A and B; are the same, the one expression may be substituted for the other in any of the
equations.
129
(1)
(2)
2x + y =
2/
12,
8.
the supposition that x and y actually have values which satisfy both equations it follows that the value of ?/ in (2) and therefore in (1) is 8.
From
(1),
we obtain
(3)
whence
Therefore,
if (1), (2)
2x+8 = =
a;
12,
2.
(4)
is
=
to
is
a solution of
(1), (2),
x = 2, y = 8. inasmuch as the
(^4),
(2) is reversible.
follows from
(4),
and then
(1)
from
is
(3), (2).
Note
364
rules of equality, 240, 253, 257, and of the general rule of equality, If a = 6, and b c, then a = c, ^ 261.
Thus,
follows
If
7/
:
right to
make
=
if
then
y4-2x =
+ 2x, or2x +
and
2
= 2x +
7/,
249.
And
Note
2X
y,
?7
12,
then 2 x
12, 261.
can be applied only when we have a right to suppose the given equations to be simultaneous. Thus, from x = 2 and x = 3 we cannot draw the absurd conclusion
2.
Of course
this principle
365
3,
because
we have no
right to suppose x
2,
3 simultaneous.
TRANSFORMATION THEOREMS
In view of what has just been said we may regard any correct application of the rules of reckoning to a pair of equations as a legitimate transformation of the pair and if
;
366
such a transformation be reversible, we may conclude that leaves the solutions of the pair unchanged.
tions in
it
Hence the following theorems, which hold good any number of unknown letters.
for equa-
Theorem 1. The solutions of a pair of equations remain unchanged when the trarisformutions o/ 338, 339 nre applied
to the
367
equations sej^arately.
130
For the solutions
such transformations.
A COLLEGE ALGEBRA
of the individual equations
remain unchanged by
2 = 2y =
?/
1
1
and y and
2x
?/
= =
5 5
2 x.
368
Theorem
2.
X,
f(x,y)=0
f(x,X)=0.
y
Here
y,
= X,
X denotes any expression in x alone (or a constant), /(x, y) any expression in x and and/(x, X) the result of substituting X for y in
/(x,
2/),
280.
is
The theorem
= x + 2 and = x + 2 and
3x 2y
3
.x
2 (x
2)
1
1.
369
Theorems.
A=
has the same
solntiotis as the
B,
C
C
=D
=
D.
as
pair
D,
A+C=B+
For
A=
B^
C=
I)
has
the
-\-
.same
C,
.solntions
A + C=B +
~
C,
C= +
y
D
5
?/
(x
and therefore as
= = 2x =
?/)
and x
y
y
and x
= =
\
1,
.
and x - y =\.
370
Corollary.
3G9 we may,
equations
constants
without changing their solutions, multiply both the given that is, both members of each equation by any
we
please, except 0.
Hence
If k and
A=
B,
C
C
=D
= D.
kA IC = kB ID,
131
371
When
*
-d
A, B, and _a
= =
0,
A = 0,
For Ali
i^
and B
0.
J.
^ has the
same
solutions as the
two equations
are the
and
jointly, 341.
of the pair
AB = 0,
0,
C=
same as those
0,
C=
and
iJ
jointly.
= = =
and x
and x
+
+
y
y
?/
and x
= =
2
2,
2.
Two systems of simultaneous equations Equivalent systems. are said to be equivalent when their solutions are the same.
Thus, the pair of equations x
pair
372
3x +
2/
5,
4x +
3?/
10,
2?/ = 5, 2x + ?/ = 4is equivalent to the both pairs having the same solution 1, 2. = 0, ?/ = 2 is equivalent to the two pairs x
ELIMINATION.
Elimination.
To
eliminate an
unknown
letter, as x,
from a
373
pair of equations is to derive from this pair an equation in which X does not occur.
We proceed to explain the more useful methods of eliminating X ov y from a pair of simple equations in x and y, and of deriving the solution of the equations from the result.
Method
of
of substitution.
This method
is
374
368.
Solve
Example.
32/
52/
3x +
Solving
(1) for
= =
= = =
3,
l.
(1) (2)
in
terms of
y,
x
3 (3
3
1.
?/.
(3)
Substituting 3
?/
for x in (2),
?/)
?/
(4)
(5)
Solving
(4),
2/
2.
= - 3.
(6)
132
Hence
For,
solution,
A COLLEGE ALGEBRA
the solution,
(1), (2) is
= 3,
?/
2.
bj'
367,
3(J8, (1),
a;
namely:
solution of
(6), (5) is
= 2.
directly
from
362.
For the
(2)
to (5), (6)
is
reversible.
(1)
Verification.
-3 +
3-2
= 3,
(-
3)
1.
(2)
Here
(4)
substitution.
'
To eliminate an ^inhnown letter, as x, from a pair of equations by substitution, obtain an exjyression for x in terms of the other letter (or letters) from one of the equations, and then in the other
equation replace x by this expression.
375
The following example illustrates a sjDecial form of this method, called elimination by comparison.
Example.
Solve
x
X
+ +
?/,
5?/
7,
(1) (2)
67/=:8.
Solving both
(1)
and
(2) for
x in terms of
= l~by,
(3)
a;
x,
?/
2/
Substituting 1 for y in
= 8-G7/. = 8 6 y. = 1. = 2.
(7)
Hence the
solution of
2,
?/
1.
376
Method
of addition or subtraction.
This method
is
based on
2x-62/ =
7,
Multiply
Multiply
(1) (2)
by
3,
Subtract
(4)
by 2, from
(3),
Whence,
Substitute
- 1/2
iov y in (1),
2x
Whence,
of (1), (2) is x
2,
2/
1/2
133
For hy 367, 368, 370, the following pairs of equations have the same namely (1), (2) (1), (5) (1), (6) (7), (6) (8), (6) and the solution of (8), (6) is x 2, y = 1/2.
solution,
: ;
Verification.
- 6(- 1/2) = 7,
subtraction.
(1)
+ 4((1), (2)
l/2)
= 4.
(2)
We
y by addition.
Multiply Multiply
by eliminating
by by
2,
4x
12 y
12 y
3,
9x +
13x
x
letter,
= =
14.
12.
(9)
(10) (11)
Add
(0)
and
(10),
=26.
Whence, as
before,
=2.
(12)
To eliminate an unknown
as x,
coefficients
Then subtract or equations equal numerically. as these coefficients have like or unlike signs.
Exceptional cases.
add according
377
Let
y.
A =0, B
Q,
B=
Q has
solution
and but
one,
unless the expressions A and B are such that in eliminating x we shall at the same time eliminate y. This can occur in the
.1
.1
.-1
= =
kB, where k
and
B=
Evidently
of
if
A =
0,
and
= ^ =
is 0,
a solution
B=
Q has
infinitely
many solutions. = x + 32/-5 = Thus, let^=2x + 6?/-10 = (1), andB Here ^ = 2 2?, so that ^ = and B = are not independent.
if
(2).
that
(1)
to eliminate
x we multiply
(2)
we
2.
at the
y.
If
A and B
A = kB +
I,
where k and
denote constants,
not
0,
we
and
^=
^ =
134
A COLLEGE ALGEBRA
make 5
= 0, B = = will
(3), a.nd
has no solution
make A
for any
I,
not
0.
Here
= 2x + Qy-9 = A = 2 B + 1, so that ^ =
(3), (4)
and
eliminate x from
we
shall at the
378
solution.
x,
c,
We may
(1)
simple equations in
y to the form
a'x
ax-\-by
b'y
c',
(2)
where
a,
denote
known numbers
or expressions.
By 377, the pair (1), (2) has one solution, and but one, ka and b' unless a constant k can be found such that a' kb,
and therefore
ab'
a'b
=k(ab
ah)
0.
To obtain
y and x independently by
results are
(4)
376.
The
- a'b) x = b'c~ be', (3) {ah' - a'b) y = ac' - a'c. (ab' a'b ^ 0, the solution of (1), (2) is Therefore, if ah' _ ac' a'c _ b'c be' "^^ y^
ab'
-a'b'
ab'-a'b
^^
easily
remembered
.
if
written
(6) ^ -^
a'b -^ 0, the
in advance that the pair (1), (2) has a solution when argument here given would only prove that if the pair
it is
(1), (2)
(5).
135
5?/
3|,
L5x + 3y =
g
1.65.
M
10.
^
f
4x-32/ =
4(6 2/-2x)4-3.
fax
lbx
+ +
by
ay
= =
aa^
+ 2a + b^, + 2b + b^.
az
11.
+
y
by
c,
px =
(x
12.
(a-6)x +
(a
^
I
qy.
+
^
'
5,
jC-y
14.
13.
lUx
1
-2y + Q-0.
V
f b
2/
x a
15.
w
f
6
_,
c
X a
16.
+
+
= 1+ X,
1
X
a'
y
b'
2_
c'
X V
L
+y.
17.
Show
Ilx -2i,y
18.
10,
Gx - lOy =
15.
In Ex. 15 assign values to a, b, c, a', b', c' for which the equations are (1) not consistent, (2) not independent.
PAIRS OF EQUATIONS NOT OF THE FIRST DEGREE WHOSE SOLUTIONS CAN BE FOUND BY SOLVING PAIRS OF SIMPLE EQUATIONS
pair of equations which are not of the first degree with respect to x and y may yet be of the first degree with respect to a certain pair of functions of x and y. can then solve the equations for this pair of functions, and from the result it
379
We
is ofiten
Example
1.
1,
10 =
5.
y
1
Both equations are of the first degree with respect to 1 /x and Solving for 1/x and \/y, we find l/x= 1/3, \/y= 1/5.
X
/y.
Hence
3,
2/
5.
136
A COLLEGE ALGEBRA
2.
Example
Solve 3a;
6,
7x
=
I,
1.
we
find x
y/x =
3.
Hence
a;
1,
3.
380
Given a pair of equations reducible to the form AB = 0, A'B' = 0, where A, B, A', B' denote integral expressions of the It follows from the theorem of 371 first degree in x and y.
all the solutions of this pair can be obtained by solving the four pairs of simple equations ^=0, A' ^=0, 0. B' 0, A' 0, B' 0; B 0; B
that
Example.
Solve
(x
x--2xy = 0, + y-l){2x+y-S)=0.
+ + x + 2x +
X
2X
;
(1)
(2)
This pair
is
= 0, ^ 0, x-2y = 0, x-2y = 0,
a; a;
2/
?/
= =
0,
(3) (4)
0,
y-l=0, y -3 = 0.
we
(5)
(6)
(3), (4),
1
(5), (0)
namely
x,
= 0,
0,
381
And,
in general, if
ABC
and A'B'C
m and
ABC
A'B'C
can be found by solving the mn pairs of simple equations obtained by combining each factor of the first product equated to with each factor of the second likewise equated to 0.
If all these pairs of simple equations are both independent solutions, that is, the number
is
EXERCISE Vin
Solve the following pairs of equations.
2x
2.
3y
(\
10x
5,
15x
in =
8.
137
y X
-'-
2(3
-y)
X
r ,
3
)
+
X
= ^y-^
X
2)
1-
1.
2
{x
0,
4. 5. 6. 7.
8.
xy
0,
+ 2y-l){Sx-y +
3x
y)
8?/
0.
xy-y =
0.
x(x~y){x +
9.
10.
= 0, x+2y -5 = 0. = 0, (X - 2) - 3) = 0. (x-l)(y-2) = _ 1)2^ 2x + 3 2/-7 = 0. (2x + 2/)-2={x-3 + 5)2, (x + 2/)2 = 1. (x-52/ + 8)(x + 32/ + 5) = 0, (2x + + 5)(5x +
(2/
2/2
(a;
2/
?/
?/
14)
0.
IN
y.
TWO VARIABLES
It is
x and
convenient to
382
lines,
In the plane select as axes of reference two fixed straight X'OX and Y'OY, which meet at right angles at the point O, called the origin; and choose ^
for
measuring
B
if
Then
follows
:
I I
I
ues be X
h,
proceed as
(
1
I
On X'OX and
x-
negative, measure off a segment, OA, whose length is \a\, the numerical
value of
a.
.Y'
Similarly on
is
or below 0, according as b
is
whose length
parallels to
We
138
A COLLEGE ALGEBRA
a,
b.
= a, y b by the symbol (a, b). We call the number a, or one of the equal line segments OA or BP, the abscissa of P and b, or one of the equal segments OB or AP, the ordinate of P. And we call the abscissa
It is convenient to represent both the value-pair x
its
and
graph
We
Y'OY the
also call A''OA' the x-axis or the axis of abscissas, y-axls or the axis of ordinates.
this
and
Observe that
correspondence,
pair
is
x,
2, v?ith
is
that
is,
(a, b)
there
one value-pair
(a, 6)
one point P, and reciprocally for each point P there found by measuring the distances of P from Y'OY
appro-
of
a;-axis,
of (a, 0)
Example.
pairs
(4, 4),
(-3,
3),
(-
5,
4), (3,
(-4,
0),
2).
Carrying
accompanying
figure.
how
303
an equatk in x and y. If, as is commonly the equation ini x and y has infinitely many real solutions, there will usually be a definite curve which contains the graphs of all these solutions and no other points.
of
,
The graph
case, a given
We
call this
139
But the graph of an equation may consist of more than one curve. Observe that we here include straight lines among curves.
Theorem.
of the
letters
The graph of every simple equation in one or both x and y is a strair/ht line.
384
On
this
readily convince himself of the truth of the theorem by selecting some particular Y " " equation and plotting a number of its
solutions.
(-L6)
=
3,
2 x
i.
1,
2,
'(0,4)
2,
0,
2,
And
(2,
0),
plotting these value-pairs (0, 4), (1, 2), 2) as in the accompanying (3,
(L2)
figure,
we
graphs
all
lie
in the
same
straight line.
X'i
:
jJ2^
(3*2)
We
1.
When
a, or
h.
'
Example.
2.
This equation
satisfied
by the value 2 of x and every value of y, 356. Hence the graph is a parallel to the ?/-axis at For this line conthe distance 2 to its right. tains all points whose abscissas are 2, and such
points only.
(x=2)
X^
'^
x a so, in general, the graph oi a parallel to the ^/-axis at the distance to the right or left according as a is |a|
And
is
distance
as h positive or negative. In particular, the graph of
is
is
\h\
ic
the y-axis.
140
2.
A COLLEGE ALGEBRA
When
= mx.
Example.
2z.
The graph
the point
(1, 2)
the right line which passes through the origin (0, 0) a,nd for this line contains every point whose ordinate is twice
its
abscissa,
And
so,
general,
the
p = mx is the right line which passes hrough the origin and the point (1, m).
3.
graph
of
When
-\- c.
2/
= mx
Example.
2x
3.
Evidently we shall obtain the graph of this equation if we increase the ordinate of every But that comes to the same thing point of the graph oi y = 2 x by S. as shifting the line y = 2x upward parallel to itself until its point of intersection with the
j/-axis is 3 units
general, the graph of a right line parallel to ?fta; and the graph of y meeting the 7/-axis at the distance \c\ from the origin,
in
is
And so, = mx + c
is
posi-
385
To
As any two
of its
points suffice to determine a right line, we may find the graph of any equation, ax
in the following example.
+ by + c =
0,
as
Example.
First,
a;
0.
Second,
points
when x =
and
0,
then y
that
is,
6.
(2, 0)
(0, 6),
the points
meet the axes, and draw the line which these points
This method
when
b,
mx.
2.
We
the equation has one of the forms x = a, then find the line by the methods explained in
384, 1
and
141
386
Graph of the solution of a pair of simultaneous simple equations. This is the point of intersection of the two lines which are the
for this point, and this point only, is the graph of a solution of both
equations.
Thus, the solution of2a; 3y + and 3 X + 2/ - 6 = (2) is a; = 1, 2/
7=0
=
3.
(1),
And, as the figure shows, the graphs and (2) intersect at the point (1, 3).
of (1)
When
lines
consiste?it,
the given equations are not 377, 2, their graphs are A'-
387
which have
common,,
that
is,
coincident lines.
2/
2/
The graph
tion of the
of an equa-
388
form
AB =
jointly
;
and
B=
Example.
(1)
(2)
142
The graph
A COLLEGE ALGEBRA
of (1) consists of the lines PQ and BS which are the graphs 7=0 and 3x4-2?/ 17=0 respectively.
of4x + y
The graph of (2) consists of the lines PS and QE which are the graphs ofx 27/-f5 = and 2x 3?/ 7 = respectively. The points P, Q, E, S in which the pair PQ, RS meets the pair
PS,
QR
(1), (2),
namely,
1),
389
Graph
of
We
find a
number
and
of the solutions of the equation, plot these solutions, then with a free hand draw a curve which will pass
all
through
"near" enough
the points thus found. By taking the solutions together, we can in this way obtain a curve
which
differs
as little as
we
con-
panying
figure.
When
0,
1,
2,
3,
9,
4,
0,
1,
1,
4,
16,
2,
3, 9,
4,
we have
Taking the
1,
4,
16,
unit of length, plot the corresponding points (0, 0), (1, 1), (2, 4) ( 1, 1), A few of them suffice to indicate the general character of ( 2, 4) the graph, the curve in the figure, except between x = 1 and x = -I- 1.
wholly above the x-axis, extending upward indefinitely and it symmetrically with respect to the y-ax\s, the same value of y corre a. sponding to X = a and x
It lies
;
lies
143
(-
7,
1),
(-
4, 3), (5,
9).
2.
0,
?/
0,
2?/
0, 0,
3(/
7x + 32/-18 =
3.
a;y
+ x = 0, 3x-42/ =
1
z
24.
?/
0,
=
4.
(x
2/
- 3)
(X
2/)
0,
x2
0,
x2
2/2,
x^
2/2
0.
cal
Find the solutions of the following pairs of equations by the graphimethod and verify the results algebraically. rx
(1)
\
5.
+ 2/-3 = x-2?/ =
+
6)(x
0,
0.
= 0, f 32/ + 2x + 19 \ ^'l22/-3x + 4 = 0.
^A
of the following pairs.
Do
the
j'(x-42/
6.
32/
6)
=
5)
0,
J
0.
(y
2)x
2)2/
U3x + 2y-10)(2x-2/ +
Find the graphs
2/
l(2/-x +
V
= 0, = 0.
of the following
two equations.
=-
(ic
1)-,
'
SYSTEMS OF SIMPLE EQUATIONS WHICH INVOLVE MORE THAN TWO UNKNOWN LETTERS
Method
letters.
of solving a
system of n simple equations in n unknown pair of equations in three unknown letters will
390
many
solutions.
;
Thus, the pair x = 2z, 2/ = z + l has infinitely many solutions for both equations are satisfied if we assign any value whatsoever, as 6, to z and the values 2 h and 6 + 1 to x and y.
But a system
unknown
391
letters ordinarily has one, and but one, solution, be obtained as in the following example.
which may
Example.
(1) (2)
3y
2 2
7.
(3)
144
A COLLEGE ALGEBRA
:
by
by
3,
Multiply
4,
Add
Again,
(1) is (3) (7)
+ 8x + 20y17x + 14y
9x 6 y
12 2
39
(4)
(5)
122
=3
(6), (9),
=-36
(6)
(7)
Multiply
Subtract
by
2,
(8),
from
(8) (9)
by
4,
68 x
63x
(10),
+ 56 y = + 56y=
12
7
(10)
(11) (12)
5x
=5
x
y
Hence
Substituting x
Substituting x
= =
1
1,
in (9),
we
find
(1),
1 in
we
find 2
= 1. = \. = 2.
- 1, Therefore, 362, (2), (3) has any solution, it is x = 1, y = But the process by which we have derived x = l, y = 1,2 = 2 2 = 2. from (1), (2), (3) is reversible. In fact, it may readily be traced backward step by step. Hence x = 1, y = 1, 2 = 2 is the solution of (1),
if (1),
(2), (3).
We may
of (1),
It is
l,
y=
1,
2 is the solution
(2), (3).
evident by 368 that x = 1, y = - 1, 2 = 2 is the solution of (12), We therefore have only to prove that the system (12), (9), (1) (9), (1). has the same solution as the given system (1), (2), (3).
posed to the
with the
^ = 0,
It will
(1)
B = 0,
(2)
C=
(9)
(3)
and
(12)
were derived,
that
we may express
(1)
thus:
0.
A=0,
will
-^ + 2C =
(9)
x, y, z
19^ 4-16B-14C =
that
(12)
makes
14
J.
make
Conversely,
also 19 .4
then
C = 0. C=
= 0, B = 0, C =
;
and when
(1),
namely,
x=l,2/ = -l,2 =
145
392
In the case just considered, from the given system of three unknown letters, x, y, z, we derived a
tivo
system of
cr,
ij,
this system, a single equation in one letter, x. And, in general, if we start with a system of n simple equa1 oi these steps, wg tions in n unknown letters and take n
the form ax
Then,
in
ufiless
h = Q. a = 0, the
may
system has one, and but one, solution, is b/a and the values of the other
This
unknown
may always
be
On
nitehj
=
ft
inji-
0.
in 394.
equations
method of solving a system of simple given in the chapter on deterviinants. In certain cases labor may be saved by special devices.
is
A much
393
Example.
Solve
Add And
we
Hence
(1)
2/
tt
7j
48.
16.
(5)
?/
turn from
(5),
B Q denote a system Let 0, C Exceptional cases. b of simple equations in x, y, z, and, as in 392, let ax denote the equation obtained by eliminating y and z.
A=0,
394
1.
If a
and
=
/>
0, it
tions A, B,
= kB +
be expressed in terms of the other two, thus and I denote constants. We then say IC, where
C may
=:
0,
0,
146
A COLLEGE ALGEBRA
the identity
From
tion oi
A = IB
IC
it
B=
and C
is
a solution of
0.
Hence
if
B= A=
and C
0,
=
C
are consistent,
377,
2,
B=
0,
will
have
infinitely
many
^ = 3x-2y + 4z-13 =
i?=:2x
0, 0, 0.
Eliminating z between
,
and
(2),
3^1
+4 5 = 17 x +
(1)
14?/
-3 = 0.
(4)
Eliminating z between
and
(.3),
^ + 2C = 17x+ 14y-3 = 0.
Eliminating y between
(4)
(5)
and
(5),
2J.
4B-2C=
=
0-x
-0 =
0.
(6)
Here the
deriving
it,
final
we
+ i - 2 C 0, or C -\- 2 B. identity 2 And, in fact, we see on examining (1), (2), (3) that by multiplying by 2 and adding the result to A.
has the form x = 0, and in equation ax 5 = find that the expressions A, B, C are connected by the
=A
C may
be obtained
has infinitely
many
solutions.
If
and
h =h 0, it will
tions A, B,
C may
A = kB -^IC
where
k,
I,
+ m,
0.
denote constants,
not
the
are not consistent. equations A = 0, B = Q, C = !^rom the identity A kB -\- IC -\- m it follows that .4 = 0, B = 0, C = Q have no solution. For any values of x, y, z that
make S
and C
will
make A
m, not
.4
0.
(1) (2)
(3)
147
we
obtain
0.
2^ + 45-2C = 0x-2 =
Hence the
-4, i?,
C = J. +
to
2 = 0,
And,
and
25+L
From
in fact,
find
(.3)
tliis
be the case.
has no solution.
Systems
discussion
of
the preceding
395
we may draw
simjde equations in n unknown Ordinarily a system of = n, infinitely many solutions has one solution when when n. n, no solution when
letters
m<
m m>
Whenever exceptions
377, 394.
to this
B=
In particular, a system of three simple equations, A = 0, 0, C = 0, in two unknown letters, x, y, has a solution
the form
when, and only when. A, B, C are connected by an identity of A = kB + IC and B = 0, C are consistent.
Thus, the system x
solution
;
\ (1),
1 (2),
3x - y = 10
(3)
has no
4,
2/
3,
does not
satisfy (3).
On the other hand, the system x-?/ = l(l),x + y = has a solution for (4) is satisfied by x = 4, y = 3. 3 X - 2/ - 9 = 2 (X - 2/ - 1) + (X + ?/ - 7).
;
7 (2), 3
He
EXERCISE X
Solve the following systems of equations.
'
x
2/
^2
+ y= +2= +X=
'x
x X
+ + 2 = 1, + 2y + 3z = + 3 + 7z=
?/
2/
4,
13.
3.
5x -22/
4.
X
X
13.
+ +
2/
7 z
= -33, = 13,
10.
32/
=-
148
'
A COLLEGE ALGEBRA
X +2?/
2 X
2/
2/
4z
1
4z = = 0, = 0.
2 z
9,
ll,
6.
.
'3x-5 =
(x
2(x-2),
6 __ 4 z
1
Z
1
32 _
y
-+
X
5,
8.
x Lx
4.
+ + + +
2
2
2/
2/
+ = 4, + M = 3, + M = 1, + 2 = 10.
z
'i{
4x 3z +
5?/
9.
<^
9,
ex
17,
&2/
10.
62/
+.az
az
8.
ex
= = =
m,
n.
11.
ax -{-by
= my = }i2, + cz =
d.
-c
14.
2x
32/
6z,
Show
identity
that the following systems are not independent and find the which connects the equations of each system.
X
13.
2/
- = 3, - z = - 5, - X = 2.
2/
'3x
82/
+ 72 =
2/
10,
2x +
PROBLEMS
396
The following problems can be solved by means of simple How equations in two or more unknown letters, a,s x, y, . many of these letters it is best to employ in any case will
depend on the conditions of the problem. But when a choice has been made of the unknown numbers of the problem which X, y, are to represent, and the remaining unknown
numbers,
if
ters, it will
any, have been expressed in terms of these letbe found that the conditions of the problem still
y,
i/,
In
number
of equations, the
149
if less,
an infinite number
395.
of
The remark
of the problem
may impose on
also.
unknown
In a certain
of the first
number
and third, the sum of the second and third digits is 8, and if the first and third digits be interchanged, the number is increased by 99. Find the number.
Let
X
the
sum
hundreds
is
digit,
tens digit, z
units digit.
Then
number
100 x
-{-
\0 y
z.
we have
(1) (2)
x
2/
+ z = y, + z = 8, +
x
100 2
+
is
10?/
= =
100 a:
2,
?/
10?/
5,
Solving
we
find x
+z+ = 3.
99.
(3)
Hence
the
number
2.
253.
After walking a certain distance a pedestrian rests for continues his journey, but at | of his original rate, and on reaching his destination finds that he has accomplished the entire distance, 20 miles, in 6 hours. If he had walked 4 miles
Example
30 minutes.
He then
further at the original rate and then rested as before, the journey would have taken 5& hours. What was his original rate, and how far from the
starting point did he rest ?
Let X
and
let
= number
of miles
(1)
the
2/12^-,^ 2
X
'
7x/8
i,
^'
|MJ
X
find
2^7x/8
16^^
]
'
^'
Solving
(1), (2)
for
y/x = 3/2,
/x
1/4.
Hence x
6.
Example
water.
40%
of alcohol
3. Two vessels, A and B, contain mixtures of alcohol and mixture of 3 parts from A and 2 parts from B will contain and a mixture of 1 part from A and 2 parts from B will
;
150
A COLLEGE ALGEBRA
What
are the percentages of alcohol in
A and B
Let X and y denote the percentages of alcohol in Then by the given conditions we shall have
^ '
and
B respectively.
5
(1), (2)
100
100
'
Solving
we
find x
52/100, or 52%, y
22/100, or 22%.
EXERCISE XI
Find three numbers whose sum is 20 and such that (1) the first plus twice the second plus three times the third equals 44 and (2) twice the sum of the first and second minus four times the third equals 14.
1.
If the first number be divided 2. The sum of three numbers is 51. by the second, the quotient is 2 and the remainder 5 but if the second number be divided by the third, the quotient is 3 and the remainder 2. What are the numbers ?
;
digits from the following data (1) twice plus three times the second equals 37 ; (2) if the order of the digits be reversed, the number is diminished by 9.
3.
:
the
first digit
4.
B owes
$3000.
;
could pay
all his
debts
if
besides his
own money he had | of B's and B could pay all but $100 How much money of his debts if besides his own money he had h of A's.
? 5.
has each
Find the fortunes of three men, A, B, and C, from the following data A and B together have p dollars B and C, q dollars C and A, What conditions nuist p, q, and r satisfy in order that the r dollars. solution found may be an admissible one ?
: ;
6.
sum
of
money
at simple interest
amounts
to $2556.05 in 2 years
and
to $2767.10 in 4 years.
?
What
is
the
sum
of
rate of interest
7.
sum of money partly in 4% bonds at par, partly in 5% bonds at 110, and his income from the investment was $050. at 80 and the 5% bonds at 110, his income from If the 4% bonds had been
invested a certain
A man
he invest
Find the area of a rectangle from the following data if 6 inches be added to its length and 6 inches to its breadth, the one becomes f of the other, and the area of the rectangle is increasetl by 84 square inches.
151
money
left.
as
A 10. A
B as much money as B had then B gave A as much had left finally A gave B as much money as B then had then had $16 and B $24. How much had each origmally ?
gave
;
and
and C,
in 4| days.
work
it
in
5}-
days
it
for 2 days
when
take
A drops out and B and C finish it in If'j days. each man separately to do the piece of work ?
11.
circle
How
long would
Two
points
move
whose length is 150 feet. When they move in opposite senses they meet every 5 seconds when they move in the same sense they are
;
What
whose lengths are 240 yards and 200 yards respectively 25 seconds to pass one another when moving in opposite directions but were the trains moving in the same direction, it would take the faster one 3J minutes to pass the slower one. What are the rates of the trains in miles per hour ?
would take two freight
;
13.
Two
steamers,
cities
C and
D which
are
200 miles apart. The steamer A can start from C 1 hour later than B, overtake B in 2 hours, and having reached D and made a 4 hours' wait What are the rates of there, on its return trip meet B 10 miles from D.
and
14.
In a half-mile race
can beat
B by
20 yards and
C by
30 yards.
beat
15. and B run two 440-yard races. In the first race gives B a start of 20 yards and beats him by 2 seconds. In the second race gives B a start of 4 seconds and beats him by 6 yards. What are the rates of and B ?
A A
passengers together have 500 pounds of baggage. One pays If the bag$1.25, the other .$1.75 for excess above the weight allowed. gage had belonged to one person, he would have had to pay $4. How
16.
Two
much baggage
17.
is
Given three alloys of the following composition A, 5 parts (by weight) gold, 2 silver, 1 lead B, 2 parts gold, 5 silver, 1 lead C, 3 parts To obtain 9 ounces of an alloy containing equal gold, 1 silver, 4 lead. quantities (by weight) of gold, silver, and lead, how many ounces of A, B,
:
A and B are alloys of silver and copper. An alloy which is 5 parts A and 3 parts B is 52% silver. One which is 5 parts A and 11 parts B is 42% silver. What are the percentages of silver in A and B respectively ?
152
A COLLEGE ALGEBRA
A
19. marksman who is firing at a target 500 yards distant hears the An observer distant 600 yards bullet strike 25 seconds after he fires.
from the target and 210 yards from the marksman hears the bullet strike 2yV seconds after he hears the report of the rifle. Find the velocity of sound and the velocity of the bullet, assuming that both of these
velocities are constant.
by two pipes, A and B, and emptied by a third and all the pipes be opened, the tank will be emptied in 3 hours if A and C alone be opened, in 1 hour; if B and C If A supplies 100 more gallons a minute alone be opened, in 45 minutes. than B does, what is the capacity of the tank, and how many gallons a
20.
tank
is
supplied
pipe, C.
If the
tank be
;
full
We
to the subject matter of algebra itself. The inquiry may arise with regard to
some particular funcCan this ftmction be reduced to a certain specified form and, if so, what are its coefficients M'hen reduced to this form? The following example will illustrate the method of attacktion of the variables under consideration,
4z +
6 be reduced to the
if so,
form
of
1,
and,
what are
its coef-
The most
a{x + \)^ Hence,
general expression of the form in question may be written b{x + I) + c, where a, b, c denote constants. if the reduction under consideration is possible, we must have
x-^
or
x^
+ ix + + 4x +
(1)
c).
(2)
be an identity when, and only when, the coefficients of like powers of x in (2) are equal, that is, when a = 1, 2 o 4- b = 4, a + 6 + c = 6, or, solving for a, b, c, when a 1, 6 = 2, c = 3
By 284,
(2)
and hence
Hence
x^
-}-
4x
-f-
= (x +
1)2
2(x
1)
3-
153
sion of
Observe that we set the given expression equal to an exprestlie required form but with undeteniiined coefficients.
then lind that to make this supposed identity true, the
We
must satisfy certain conditional equations. And by solving these equations we obtain the values of the
coefficients
coefficients.
The following is a more general kind of problem including that just considered. Certain conditions are stated and the question is then asked, Does any function of a certain specified form exist
which
will satisfy these conditions, and, if so,
398
what are
its
coefficients ?
To
form
we
These coeffiand the given conditions yield the system of equations which they must
with undetermined
unknown numbers
of the problem
If this s^'stem of equations has a single solution, we if obtain a single function satisfying the given conditions the system has no solution, no such function exists if the
satisfy.
; ;
system has
infinitely
many
solutions, the
problem
is
indeter-
minate, there being infinitely many functions satisfying the It is here supposed that the function under given conditions.
discussion
is
ajinife expression,
If possible, find a
264.
in x, of the
polynomial
I
second degree,
when x =
and when x
3,
4.
in question must have the form axby the conditions of the problem
The polynomial
a
bx
c.
And
0,
9a +
36
+
is
2, 6
= 0, 16a + = 8, c = 6. 2 x^ 8 x + 6.
c
4&-f-c
6.
the problem been to find a polynomial of the first degree satisfying the given conditions, there would have been no solution had it been to find one of the third degree, there would have been infinitely many
;
Had
somtions.
154
399
A COLLEGE ALGEBRA
The method illustrated in the preceding sections is called the method of undetermined coefficients. It is the principal method of investigation in algebra and we shall often have occasion to
apply
>^
it
as
we
proceed.
EXERCISE
XII
1.
x^
a;
5 as a
polynomial in x
2.
2. 3.
+ 8x + 7asa
ax^
polynomial in 2 x
3.
Find/(x)
=
=
6x
such that
/(l)
ex
/(-1)=11,
4.
= -5,
/(5)
6.
rind/(x)
ax3
ftx^
+
1)
+
1,
d such that
/(I)
/(O)
5.
=
=
5,
/(-{-
=
c
9,
/(2)
31.
rind/(x,
y)
ax
by
4,
-\-
Such that
/(O, 0)
6.
/(4, 4)
0,
/(I, 0)
1
6.
two
of
whose solutions
+
4,
6y
2/
+c= = -l;
8.
is
4, 5).
Determine
of3x +
2/
will pass
through
the point
2, 3).
10. Find two simple equations, ax + 6y + 1 = (1), a'x + h'y such that both are satisfied by x = 2, y = 3 and also (1) by x and (2) by x = 3, ?/ = 7.
+\ =
(2),
7,
x^
6x2
ex
2, 1,
and
3.
12.
the solutions
13.
%=
which has
2y
3 in the
form
where
a, 6,
2)
c{x
+ 2y-3),
155
of
In 319 we defined the quctient Preliminary considerations. .1 by B as the simplest form to which the fraction A/ B
400
when A and B are polynomials in the same and the degree of A is not less than that of B.
that
It is then possible that iJ is a factor of ^, in other words, can be reduced to the form
A
where Q
is
QB,
x.
(1)
an integral function of
We
that
then have
the quotient of
is
=
B
Q,
is,
^ by 5
is
exactly divisible
by B.
if^=x3 + 4x2-2x-5 and i? = x2 + 3 x - 5, it will be found that x^ + 4x2 2x 5 = (x + 1) (x^ + 3x 5), an identity of the form (1), Q being x + 1. A = x3 + 4x2-2x-5 = x + 1. Hence B x2 + 3x-5
2.
But
We
show,
happen that B is not a factor of A. to the form QB; but, as we shall 401, we can reduce it to the form
it
will usually
A = QB +
where both Q and
of
R,
x,
(2)
is
We
then have
+ ^'
156
that
tion,
is,
A COLLEGE ALGEBRA
the quotient of A by B is the sum of an integral funcand a fraction, B/B, whose numerator is of loiver
Q,
degree than its denominator. In this case we call Q the ijitegral part of the quotient, and
the remainder.
B = x2 + 2x +
2)
2,
we can
at once
2i2
i? is
+ 3x +
x
= x(x2+2x +
is
(x
3),
where Q
is
x and
3,
which
Hence
^ = x3 + 2x2 +
B
x2
2x + +
+
3x +
2
+
,
x2
x + 3 + 2x +
401
The
reduce
division
.4
transformation.
It
remains to show
how
to
to the
form QB
R, where
is
and may be
is
plished
The process by which this is usually accom0. called the division transformation, or " long division."
2 x*
and B = x'^-x + l. and the problem is to find an integral function, Q, such that the remainder, R, obtained by subtracting QB from A, shall be of t\].e first degree at most and for if such a function, Q, be found, we shall have may be
Let
.4
+3
ic^
+ 4 ^2 + X - 2
is ttvo,
QB =
R, and therefore
is foiir
A = QB -^ R. and that of R is to
be not
we
subtract QB.
157
tract
Evidently we shall cancel the leading term of A if we subany multiple of B which has the same leading term that
The simplest multiple of this kind is 2x^B, where has. the multiplier 2 x^ is found by dividing the leading term of A, namely 2 x*, by the leading term of B, namely x^. Subtracting 2x'-B from A, as above, we have
A-2x''B^^x^ +
2x-
+ x-2.
(1)
cancel the leading term of the remainder (1) thus obtained, and with it the second term of ^, by a similar process. The quotient of 5 a;* by x^ is 5 a- and multiplying fi by 5 x
;
We may
- (2
a;2
+5
ar)
5=
7 x2
- 4 x - 2.
(2)
Finally,
(2),
we
and with
the multiplier,
7, is
found by dividing 7 x^ by
5X
The
result is
(3)
it
.4
- (2 x^ +
+ 7) B = 3 X -
9.
The remainder
and we obtain
by subtracting (2x"-^ + 5x + 7)fi from A. Hence the polynomials Q and R which we are seeking are
2 x2
5X
7 and
7?
3x
9.
And
A =(2x2
we have A
in the
form
of
is less
than that of B.
We
when A and B
Arrange both A and B according to descendiyig poivers of x. Divide the leading term of A by the leading term of B ; the
quotient will be the first term of Q.
Multijjly
from A.
158
A COLLEGE ALGEBRA
K QB
or
K..
illustrated above.
in multiplication.
customary to arrange the reckoning in the manner We can then use detached coefficients, as
Example
1.
Given
4; find
J.
2 x^
6 x*
7 x^
8 x2
19 x
20,
and
5 = x2-3x +
Q and
B.
20
159
At the same time it is customary to call the operation by which Q and R are found "division," and also to call Q the
"
"
quotient
is
when R
not
instead of the " integral part of the quotient " even and we shall usually follow this practice.
;
^" does not then mean, as in 254, which multiplied by B will produce A, but finding, first, what multiple of B we must subtract from A to obtain a remainder which is of lower degree than B, and, second, what this remainder is. Compare 87. 3. The steps by which the integral expression A is reduced to the integral form QB + R may be taken whatever the value of X. Hence A and QB + R have equal values for all values of On the other X, even those values for which B is equal to 0.
But "dividing A by
finding an expression
hand, neither
Thus,
if
A / B nor Q
x2
+ R /B
X+
=
\
when
= 0.
(1)
A=x'^ +
401,
and
5=
-
x
1)
1,
we
find,
by
+
1
(x
2) (x
3,
and therefore
Here B = 3, which
X i = x +
when x
is
(2) ^ '
1.
true, but
3/0 =
we have
that
of A to the form QB + ^ is unique, there exists but one pair of integral functions Q and R (of which it is of a lower degree than B) such that
4.
The transformation
is,
A
QB +
But
but (Q
QB
+ R.
we should have
q')B = R'
R= qB + R'
and therefore
R'
{Q-
R.
would be
B
403
The
effect of
1. If we multiphj the dividend by any constant, as multiply both quotient and remainder by c.
we
F-or if
A = QB + R,
then
cA
= cQ-B + cR.
160
A COLLEGE ALGEBRA
we
divide the quotient by
c,
For
3.
if
A = qB +
')n\dtiply both
c,
R, then
A = ^-cB-\- R.
If we
the remainder by
For
if
A = QB +
cA
Q,-cB
cR.
4. If at any stage of a division tra7isformation we midtiply an intermediate remainder or the divisor by c, the final remainder,
if changed at
all, is
1
merely multiplied by
and 2 and
402, 1.
c.
The student
particular case.
Thus, we
first
may
2
by
4
B=
+
6
X
1
=4x2_).6x
B=
+ +
4-2
8
1
1
1
4-2
8
1
.-.
8-4
404
.-.
Q = 2x + R = b.
4,
8-4
Q= R=
x
b.
2,
We may
:
also find
Q and R when
1.
.4
and
Example
Divide
^=
2x*
3x3
4x2
x-2byE
= x2-x +
l.
Since the degree of A is four and that of B is two, we know in advance that the degree of Q is two and that the degree of -K is one at most.
Hence
let
Q =
CqX^
c^x
Ci
and
R=
dox
di,
to be
coefficients Cq,
di that
3x3
4x2
X-2 = (CoX2 +
CoX*
CiX
Cj) (X2
1)
cfoX
di
++
Co X3
Ci
C0
161
make
(1)
an
identity,
2.
we must
have, 284,
Co=
= 3+2 + ci = 3, .-.01= 3 + Co + C2 = 4, .-.02= 4 Co + Ci = 4 2 + 5 rfo = 1 - Ci + C2 = 1 - 5 + 7 C2 + Cio = Cl 1, - 2, =-2-7 Cj + di = di = - 2 - C2 Hence Q = 2x'^ + bx + 1 and jB = 3x - 9, as in 401. Example 2. Divide 6x5 + 13x* - 12x3- 11x2 + llx-2 by
Co
Ci
Co
.-.
= 5. = 7. = o. = - 9.
2x2 + x-
2.
denote polynomials in x with For literal coefficients, and suppose the degree of B to be 7?i. the division of .4 by B to be exact, the remainder R must equal
Exact
division.
Let
A and B
405
identically.
all
the coefficients of
be
0.
277, and evidently these coefficients are functions of the coefficients of A and B. Hence
is
??i
1, it
has
m coefficients,
In order that a
jjohjiioviial
hy a
I>, tlie
coefficients
of
and
by x2
3x2
6x
2 exactly
We
have
x3
+ +
3x2
ax2
+ +
^^
+21 x2 + gx +
|x
+
a)
1)
(3
a)
(3(3
a)x2
g) x2
(6
Hence a and
6
must
a2
1 3a +
2.
0; whence a
and
Zx
3.
Example
so that 2x3
3x2
+ m ^ay
be
Let Dividend and divisor arranged in ascending powers of x. divisor arranged in ascending
406
162
.A COLLEGE
ALGEBRA
We may then obtain an begins with. A in terms by B by the process of If A is exactly cancelling leading terms described in 401. divisible by B, this result is the same as when A and B are
of lower degree than
arranged
divisible
in
descending powers
is
but
if
is
not
exactly
by B, the result
3x2
entirely different.
The following
clear.
+ 3x + + X
2x 2x
x3
163
we
please.
Hence
denote polynomials arranged in ascending powers
T/'A and
of X, A not being exactly divisible by B nor beginning with a term of lower degree than B begins with, we can reduce the
qiiotient
of A. by
to the
form
B-^
and R' beginning with a
If the
still
+B'
where Q' and R' are integral functions arranged in ascending j^owers of x, Q' ending with as high a power of x as we please
higher power.
is n, we call Q' the quotient of n terms, and R' the corresponding remainder. "When the value of x is small (how small will be shown
number
of terms in Q'
by
to
subsequently)
we can make
please by taking n great enough that is, we can find a polynomial Q' whose value will differ as little as we please from that of A / B. On this account the polynomials Q' are
we
the frac-
A /B.
1
Thus "dividing"
1 1
by
to
n "steps," we obtain
h
= 1+X +
x2H
X"-!
+
1
1,
X"
If
tliat
we
1
give x
less
than
4-
1/(1
x).
Thus,
differs
from
from 1/(1^ x) by only 1/18. Similarly x) by only 1/54 and so on. /(I
;
1/3, then
xV(l -
x)
Quotients to n terms found by the method of undetermined coefficients. proceed as in the following example.
408
We
Example
Let
1.
(3
x)/(l
aq,x"
X - X + 2 x2 = ao +
3
ttix
a^x^
(1)
164
Multiplying both
A COLLEGE ALGEBRA
members by
3
1
+
a;
2 x^
= ao +
tti
a.2
ao
165
and R is of lower degree than B with respect to that letter. But the form of this expression will depend on what choice is
made of the
Example.
(1)
letter
of arrangement.
Divide 4x2
6x2/
!/2
by 2x
y.
we have
4 x2 4 x2
+ +
6 xy 2 xy
2/2
+ 2 X +
2X
?y
2y
Hence
4x2
43^2/+
y2
4xy+2y2
_
(2)
y1
2x + y
Choosing y as the
y
letter of
arrangement,
we have
y2
y2
+ 2x
1G6
11.
12.
A COLLEGE ALGEBRA
Divide 2 x2 Divide 2
13.
14. 15.
16.
+ -
5 xy
b-
3y'^
- 5x + 13y - 12by
6c^
6'^
ab
Divide 8 x^
27 y^
by
?/*
2x
Also X*
4 X2/3
o*6
17.
18.
Also 6
a5
+ - a362 +
3
by X
?/.
n a^b^ +
5 a6*
4b^ hy 2a^
ab
62.
in ascending
powers of
is
x, find
when dividend
3x
5 x2
is 1
3 x2 and divisor 1
3x
5 x2.
21.
By
the
method
of
undetermined
2
x).
four
3x
4 x2) /
(1
We
the
degree whose leading coefficient is 1. Consider the result of dividing o-^ + (h^^
first
h.
2a^
by
UqX^
a^x"^
a.,x
-\-
a^
+ (n^b + ai)x + (a J? + cixb + a-) + ai)x^ + a^x (aj)'^ + aib)x (a J) + fti)x'^ as {ajj^ + /> + a^)x {ay + (>\b + 2)-^ - {(iJ>^ + (Ub"^ + Cob) =R a^b^ + a-J)^ -\-aj)-\-a^
rtpx^
(a^b
-\-
167
and
are
Observe
t hat
the
first
of these coefficients
tliat
is
the leading
coefficient of the
dividend and
the rest
:
may
be obtained.
__
Multiply the coefficient last obtained by b and add the next unused coefficient of the dividend.
Thus,
ao6-
+
+
a-ih
a^
= =
{aoh
{aah-
aj) 6
+
+
a^,
and
aoh^
ai6-
a^h
as
aib
a2)h
a^.
This rule applies whatever the degree of the dividend maybe. For since the coefficient of the leading term of the divisor
is 1,
each
new
coefficient of
will
Like that leading coefficient of the remainder last obtained. coefficient, therefore, it is found by multiplying the preceding coefficient of Q by J and adding a new coefficient of the dividend.
And
we
shall obtain R, if
we multiply
Q by
b and the diviHence, when the divisor has the form x dend the form a^x" + Oja;""^ -f + , we can find Q and R as follows, where Cq, Ci, c_i denote the coefficients of Q.
411
tto
a2
Cjfi
a_i
c_2b
^n-l>
\b
_ c^
Cq
Cl
_^
c_ib
-^
C2
We
first
write
down
know
to be the
same as
ai,
aQ.
We then multiply
Cq
by
b,
under
add,
and so obtain c^. In like manner we multiply aj, add, and so obtain c^.
Ci
by
b,
product
CjS
under
168
A collegp: algp:bra
And we continue thus, multiplying and adding alternately, until all the coefficients Uq, a^, a are exhausted.
Example.
Divide 3 x*
5 x^
4 x^
3x
by x
2.
We
have
1-2-1,-4
x2
Hence
Q=
3 x^
2x
and
E = - 4,
This very compact method is called synthetic division. The student should accustom himself to employ it whenever the
divisor has the form x
h.
112
Remarks on
the dividend
this method.
is
1.
powers of
h),
by
coefficients.
Since x
-{-
h :=
we may
divide synthetically by
hhy
va.
Example
Here
1.
Divide x*
by x
1.
(
I-
1),
we have
1-1+1-1,
Hence
3.
+1
-1
+1
and
Q=
x*
x^
x-
E=
0.
To
:
(3,
write
it
thus
a(x
(3
Then divide synthetically by x /3/a, and let Q and R represent the quotient and remainder so obtained. The quotient and remainder corresponding to the divisor
/ex).
ax-
13
will he
2.
Q/a and
3(x
li,
403,
18 x
2.
Example
Here 3x
Divide 3 x^
2
11 x^
by 3 x
2.
2/3,
we have
-11
+18
--_6
12,
-3
|2/3
J 3-9
_8
6
169
(3x2
_ 9x +
by x
12)/3, or z^
Sx +
4:
3.
Divide 5 x^
Divide x^
x'
+ +
+
x
3.
4.
5.
6 x^
Divide 2 x^
3 x^
+ 6 by x + 3. + 8 x - 12 by 2 x 11
3,
The Remainder Theorem. When a poly 7iomial in si is divided b, a remainder is obtained which is equal to the result of iy X
substituting b fur x in the dividend
;
413
dividend, f (b)
ivill
The demonstration of this theorem is given in 410 for it there shown that if we divide a^^ + a^x^ + a^ -\- a^ by X b, we obtain the remainder ajb^ + a^b^ + aj) + a^, and, in + by general, that if we divide f{x) = a^x"^ + ajx""^ +
is
b,
aj)"'^
:
+ a,
or f(b).
be the dividend, x
401,
h the divisor,
(x)
B.
the
remainder, then,
/(x)
<l>
(X) (X
6)
E,
where B, being of lower degree than x 6, does not involve x at all and therefore has the same value for all values of x. The two members of this identity havs equal values whatever the Hence value of X. In particular they have equal values when x = b.
f{b)
<p{b}(b-b)
R.
<p
But 6 Hence
<p
0; and since
(x) is
integral,
(6) is finite.
(6) (6
6)
0,
and therefore
/(6)
R.
414
ing that, when b and the coefficients of f(x) are given numbers, usually the simplest method of computing the value of f(b) is to divide the remainder thus b synthetically f(x) hy x
170
Example
1.
A COLLEGE ALGEBRA
What
=: 5 X*
is
the value of
12 x3
f{x)
1.
20 x2
43 X
6,
when x
=4?
By
the
4*
method
12
.
of direct substitution
42
we have
/(4)
=5
43
- 20
- 43
-20
32
12
2.
By
the
method
5
of synthetic division
-12
20
8
-43
48
5,
6[4 20
26 =/(4).
Example
415
2.
Given /(x)
3 x^
x^
5 x2
8x
4.
Find, by syn-
/(-
2), /(4),
/(-
2/3).
a.
= b,
is
then f (x)
is
For, by
hy X
413, f{b)
ft,
Thus, /(x)
division.
= x3 - 3 x2 + 2
is
vanishes
when x =
1,
Hence f{x)
Again, /(x)
exactly divisible by x
1,
as
may
= 1-3 + 2 = 0.
by actual
Example
value of
1.
= x" 6 is exactly divisible by x 6, since /(6) = 6" 6" = 0. If x^ + 3x2 jg exactly divisible by x 2, what is the
;,
m?
2^
We
must have
2.
is
2^
m=
6" is
0, or
m 20.
by x
Example
but not
if
Show
even.
that x"
exactly divisible
&
if
is
odd,
416
Corollary 2.
vanishes
when two of
If an integral function of two or rnore variables these variables, as x and j, are supposed
y.
For the function may be reduced to the form of a polynomial in X with coefficients involving the other variables. By It is therehypothesis, this polynomial vanishes when x = y. fore exactly divisible by x y, 415.
Thus, x2 {y x) -f 7r (x z) + xp- (z 2) stituting y for X, we have y- (y
y) vanishes
(2
when x = y
for sub-
2/2
y)
2^ {y
y)
= 0.
171
by x
We may
(y (y
y.
z)x2
(2/2
22)x
{y^z
z^y)
y
{y
2)a;2- {y^
-z)x-
{yz
- 22)
z^)x
^ (y z
z^y)
(x
x.
y)^ +
f
{y
z)^ +
{z
x)^
is
exactly divisible
Theorem.
whe7i X
If a poli/tiomial
then f (x)
is
(x) vanishes
when x
b,
exactly divisible hy (x
is
a) (x
417
by hypothesis, f(x)
if
exactly divisible
<^i
a,
415, and
we
</>!
(x),
f(x) = (x
If in (1)
a)
(x),
where ^i (x)
is
integral.
we
set
b,
we have
(2)
f(b)
But by
and
=^ 0.
its
Therefore, since
factors
must vanish, 253, it follows from (2) that ^i (b) = 0. But if ^1 (b) = 0, then <^i (x) is exactly divisible hj x
b,
415, and
if
we
we have
is
^1 (x)
= (x
(x),
where ^2 (^)
integral.
(3)
we have
(4)
f{x)
= {x-a)(x-b)is
4>2
(x),
exactly divisible
by
a)(x b). (x
is
may
= a,
,
.
a) (x b) (x c)
2X
then f(x)
exactly
418
Thus, 2 x3
3 x2
-3
vanishes
when x
1,
for 2
0,
and when x = - 1, for - 2 + 3 + 2 - 3 = 0. Hence 2x3 + 3x2 2x 3 jg exactly divisible by x2 1, as may be verified by actual division.
(x
1) (x
1),
or
172
Example L
X
A COLLEGE ALGEBRA
Find a polynomial /(x), of the second degree, which when x = 2 and when a; = 3, and the value 6 when
4.
is
Since /(x)
417,
it
may
2)(x 3), of the second degree and is exactly divisible by (x be expressed in the form/(x) = ao{x - 2) (x - 3), where Uq
=
6,
denotes a constant.
And
since /(4)
we have
2) (x
Hence f(x)
3 (x
3)
whence ao
3.
Example 2. Find a polynomial /(x) of the third degree which will when x = 2 and when x = 3, and will take the value 6 when x = 1 and the value 18 when x = 4.
vanish
(aox
ai) (x
2) (x
3)
where
are constants.
6,
and /(4)
= -
18,
3),
we have
or or
ao 4 ao
18
= = =
(ao
ai) (1
2) (1
(4ao
ai)(4
-2)(4-3),
ao
+ +
ai
Oi
= =
x
3,
9.
(1)
(2)
Solving
(1)
and
(2),
we obtain
1) (x
Hence /(x)
(2x
2) (x
= 2, ai = 1. - 3) = 2x3 -
Ox^
6.
419
Theorem.
A polynomial f (x).
whose degree
is n,
cannot vanish
For if f{x) could vanish for more than n values of x, it would be exactly divisible by the product of more than n factors of the form ic a, 41 8, which is evidently impossible
since the degree of such a product exceeds n.
420
Theorem. If we know of a certain polynomial f(x), whose degree cannot exceed n, that it will iJanish for more than n values of x, we may conclude that all its coefficients are 0.
For if the coefficients were not all 0, the polynomial could not vanish for more than w values of x, 419.
We
421
<^(x),
it
vanishes identically.
Theorem.
If tiro polynomials of the nth degree, f (x) and have equal values for more than n values of x, their
173
= a^x"" + aia;"-* h + ^i^;""^ H h^; ^(a;)= = f(x) (x) (x) = (ao - ^o) X- + (ai - b,) a;"-' + . + (- J).
f{x)
-\
^o-"
\p
<fi
Then
xf/
(x) is
<^
(x) are
the same, and by hypothesis these values are the same for more than 71 values of x.
\j/(x)
= (ao
bo)x"
+(a
b),
whose degree does not exceed n, vanishes for more than n values of x, and therefore, 420, all its coefficients are 0.
Therefore Uq
whence
that
is,
bo = 0, a^ = b^,
=
2 x2
Ui
= 0, Oi = b^,
bi
-,
b =0, = b,
a.
<^
Thus,
values
f{x)
+
6,
wheu x
2, 4,
+ Sx c = 3, and
-\-
have equal
c
5.
EXERCISE XIV
1.
Divide x*
3 x^
x^
11 x
4 by x
4 synthetically.
2.
3.
Also 5x5
-6x*+ x2
8x3
7x2
2.
+ 6x + 3by
3.
Also 3x*
4.
5. 6. 7.
8.
Also3x3 + 16x2
Also 3x3 Also x^
+ - 13x by X
X
by 3x
1.
6x2
+
b
by 3x
-f-
1.
{a
c) x-
{ab
ac
-{
bc)x
abc by x
a.
Also 2x*
-x^y
Given/(x)
10 y^ by x
the
2 y.
method
3),
of 414, find
6).
1),
/(-
/(-
By
may may
be exactly divisible by x
+ 7nx2 - 20x + 3.
I
10.
and
4).
so that 2 x^
x^
-f ix
+m
be exactly divisible by (x
2) (x
174
11.
A COLLEGE ALGEBRA
By
416, show that 2 n, also by a + 3 b.
Sbm + am
that
an
6bn
is
exactly divisible
by
m
12.
By 416,
417,
show
exactly divisible by (a
13.
b){b
a{b
a).
c)^
c) (c
b{c
- af +
c{a
b)^
is
when
14.
a;
Find the integral function of x of the third degree which vanishes = 1, 4, 2, and takes the value 16 when x = 2.
when
x
a;
Find the integral function of x of the third degree which vanishes = 2, 3, and takes the value 6 when x = and the value 12 when
l.
15.
Show
that
X'''
ax
x.
-ii-
and
x^
+ 5x +
A and B
x,
of higher degree
than B.
Divide
the remainder
then
(1)
not of lower degree than B, divide the quotient Qi, the remainder Ri then
If
is
Qhy B and
call
Q
if
is
Qi5
R^.
(2)
not of lower degree than B, divide Qi by Qi Similarly, and call the quotient Q2, the remainder R^; then
Qi
Q,B
R^.
(3)
Suppose that Q2
is
We then
have
A = QB + R
by by by
(1)
(2)
(3)
the coefficients
Q^,
R2, Ri,
than B.
175
is
of higher degree than B, and we continue the process just described until a quotient is reached which is of lower degree
A=
Q,_,B'-
i?,_i'-i
---
R,B-\-R
than B.
Q,.-i
all
being of
Reduce
4 to the form of a poly1 with coefficients whose degrees are less than two.
x^
4 x*
x-^
-.'
Using detached
coefficients,
we may arrange
1-4+3-1+1+4
1
-5+2-1 -5-5-5
176
A COLLEGE ALGEBRA
EXERCISE XV
1.
By
z^
1 iu
terms of x^
4. i.
2. 3. 4.
5. 6.
+ -
2 x^
+ +
4 x2 2 x^
+
+
6 in terms of 2 ^
1.
3 x6
5 X*
x2
6 in terms of x^
x^
3.
+ x3?/2
+
+ +
4-
a;22/3
y^ in
terms of x^
xy
x
H-
y"^.
By the method
Also x5
3 X* 9 x2
3 x2
of 423 express 2 x^
8 x^ +
7 in
in
terms of x
2.
-3
6 x3
2 x2
3X
terms of x
7.
8.
Also x'
Also x'
+
+
27 X in terms of x
3.
1 in
terms of x
1.
VI.
424
or
Factor.
Any rational integral function of these variables which exactly divides A is called factor of A. Hence in order that a given function, F, may be a factor of
vl, it is
more
and necessary be rational and integral with respect to the variables of which A is a function.
sufficient
1.
That
2.
GF, where
also
is integral.
Example
1.
Since 2x2
2xi/
= 2x(x
y),
both x and
x~y
are
factors of 2 x2
2 xy.
Example
2.
Since
3x2
2/
2^/2
V3 X + V2 V^
2/
and
3.
V3 X V2
are factors of 3 x2
both
Example
-I-
Althoiigh x
Vy and Vx Vy
y.
factors of x
177
425
The
coefficients of a factor
kind.
Therefore, since x^ y = {x + Vy) (j; V^), the expression x^ y, regarded as a function of both x and y, cannot be factored but regarded as a function of x alone, it has the factors x + V^ and x Vy. And the like may be said of other expressions which involve more than
;
On
numbers or expressions
of
any
one
letter.
Note
2.
Except v?hen dealing exclusively with functions having customary to include a mere "numerical
''st
426
for
if,
as here,
we do not require the coefficients of an integral any mere number (or constant) whatsoever may
is
For a
like reason, if
cF
also a factor of
or that
c is any constant (not 0), and cF as essentially the same list of the factors of A. would be equally correct to say that 2x and x y,
;
a factor of A, and
x,
Theorem.
is
a factor of A, then
427
For, by
424,
A and B A
= GB =
and
B = HF,
where
G and
H are
integral.
Hence
that
is,
A
is
GHF= GH-F,
a factor of ^,
424.
Prime, composite.
An
integral function
may have no
other
428
In that case we
call it conijjosite.
call it prime.
But
if it
have other
factors,
we
Thus, X
+ 2/2
and x
2y
yi
js
composite.
composite function, A, of the nth. degree, is the product of not less than two nor more than n prime functions, B, C, -.
429
factors of A.
178
430
In what follows
1. 2.
A COLLEGE ALGEBRA
Ave shall
assume that
Two or viore of these prime factors may be equal, but A 3. can be expressed in only one way as a product of j^owers of its
different pjrime factors.
A has hnt one set of prime factors. A are products of these jJrime factors.
in 484,
These theorems, of which 2 and 3 are corollaries to 1, will be proved 485 for the case in which ^ is a function of a single variable, and they can be proved generally.
- 2 x^^i - xxyyy (x - 2 y), the prime factors of Thus, since x^y^ - 2x2;/* are x, x, y, y,y,x-2y. Its other factors, as x^, xy, and x^y^ Its different so on, are products of two or more of these prime factors. and it can be expressed in but one way as factors are
x'^y^(x
2y).
is it
431
Factoring. factor a given function, A, completely "resolve it into its prime factors," that is, to reduce
To
to to
the form
functions.
A=BCD---,
We
But ordinarily we do not attempt to discover these prime endeavor first to resolve A into factors at the outset. of some tivo of its factors, as F and G, next to a
product
resolve
reached.
And
and G, and so on, until the prime factors are even the first step in this process may be
.4.
called "factoring"
Factoring
involves two
is
main steps
(1)
law, in order to replace (a + tion of like terms in the result thus obtained.
(1)
A multiplication usually a number of applications of the distributive (2) the combinab)c by ac + be, and so on
;
we must separate the terms thus combined the diihculty of factoring consists and then
(2)
To
it
law in order to replace ac + be by (a + b) c, and so on. It must not be supposed that every composite function can be actually factored. Thus, while it can be proved that x& + ax^ + bx- + ex + d is it can also be proved that tlie factors of this expression cannot
composite,
is,
by applying, a
finite
number
of
179
432
ab
-{-
ac
-\-
ad
-\-
^
+
(b
-\-
-{-
-\-
).
Example
1.
Factor 2 a^c
2 abc
4 ac^
6 acd.
"
Separating
"
it,
we have
2 a5c
4 ac2
d)
6 acd
2 ac (a
2 c
3 d).
Example
+ b{d c). Both terms have the factor c d. Separating it, we have ~ d) + b{d c) = a{c - d) b(c - d) = {a b){c a{c
2.
Factor a{c
d).
Some
a
433
can be reduced to
as
have
common
Example
factor.
1.
Factor ac
bd
ad
be.
Combining ac and ad, also be and bd, we obtain a(c a binomial whose terms have the common factor c + d.
d)
5(c
d),
Hence
ac
bd
ad
bc
{a
b) {c
d).
Observe that the parts into which we separate the expression in applying this method must all have the same number
of terms.
jExample
2.
Factor a^
ab
bd
ad
ac
cd.
must have either two groups of three terms, or three groups of two terms. Four of the terms involve a, namely, a^, ab, ad, ac, and To obtain groups the remaining two involve d, namely, bd and cd. which have the same number of terms we combine the term ad with the d-terms, and have
a'^
We
ab
ac
- ad-bd -
cd
a{a
c)
d{a
c)
= {a-
d)(a
b-ir c).
180
A COLLEGE ALGEBRA
EXERCISE XVI
6 x*yH^
12 x^y^z
6
8 x^yS.
2.
3.
a6
- L
8.
A.
6.
8.
+ (n - 3) . mx - nx mn +
2 n^
n^.
5.
3xi/-2x-122/ +
x32/2
lOxy
x*
2/2
+ 6x +
By.
7. 9.
x^ys
2 x2y
- 2 xy^.
ac
M - (6c + ad).
ce
10.
12.
11.
ad
bcZ
+ ae + cd + 6e.
ad.
products, as (a
can be reduced to the form of one of these products, we can The following sections will write down its factors at once.
illustrate this
method
of factoring.
435
This name
is
given to expressions
which have one of the forms a^ 2 ab + b'^. can be factored by means of the formulas
:
Such expressions
b^
Observe
that
in
of the exti-eme terms, and that the factors, which are equal, are obtained by connecting the principal square roots of the
To
Factor 9x2
12x1/
+
2
4^2,
v9x2
vij/^.
of the middle term
And
is
since
V9x2 =
3 x,
Vij^ =
?/,
-, we have 9x2
\2xy
4y2
(3a;
22/)(3x
-2y) =
(3x
22/)2.
181
Factor a^
b'^
c^
2 ab
2 ac
2bc.
We
a2
can reduce
:
this to the
form
of a trinomial square
by grouping the
terms thus
2 a6
62
2 ac 4- 2 6c
c2 :=
(rt
=
Example
1.
(a
+ +
6)2
2 (a
6) c
c2
c)2.
3.
x2
14 X
49.
2.
9
x2 a2
3. 5.
9x2^/2
64 a8
+ 30x1/ + - 48 a* + 9.
25.
4.
6.
difference of
it,
reducible to
two squares. Expressions of this form, or can be factored by aid of the formula:
436
a-'-P=(a +
Thus,
X2
b) {a
b).
2/2
22
2/2
= X2 = X2 - (X +
(?/2
(2/
2).
very useful device for reducing a trinomial expression to this form is that of making it a perfect square by adding a suitable quantity to one of its terms and then subtracting
this quantity
Thus,
x22/2
2/*
a;4
= =
Example.
1.
+ 2 x^y"^ + - ^^v"^ _ x2y2 (X2 + ^2)2 (x2 + 2/2 + xy) (x2 + 2/2
2/*
xy).
X*
?/.
6 a3
6 a62.
3.
6.
12
x*
aH^ - 75 axy^.
4.
25x2n_49x2m.
36x*-l.
Sx'^y'^
y*.
A sum
unit
i
of
two
1,
squares.
By making
sum
437
be b"^ can of squares a^ reduced to the form of a difference of squares and then factored
218, 220, a
= V
by
For since
P=-
1,
we have
a^
Hence
a^ + 6^ =
182
A COLLEGE ALGEBRA
seen, 219, 220,
i
As we have
ing
it
conforms to
to
all
the ordinai-j
rules of reckoning.
that
i^
1.
438
Sums and
First.
diiferences of
like powers.
Whether
?i
is
everi,
a--bSecond.
a"
= {aWhen
b) (a"-'
+ a"-^ +
a^-'b
+ ab' + b"-').
b"-').
(1)
is even,
^.
= (a + b) (fl^-i n
is
- + ab"-^ ab^'^
(2)
Third.
a"
When
odd,
+ b"=(a +
b is b is + b" is
b)
(a"-^
a'^-H
:
+
h.
b"'').
(3)
a"
a
always
divisible by
a,
divisible
by
a, -{a.
3. 4.
a"
divisible by
h when n h tvheji n
consists
is even.
is odd.
of the terms
b"-i
a-2b-ab"-2
connected by signs which are all when a and hut are alternately
when a
b
'
is
the divisor,
+bw
the divisor.
a;2
Thus,
1.
2.
3.
8 o3
_ 1 = (X - 1) (x5 + X* + x3 + + X + 1). - 1 = (X + 1) (X5 - X* + X3 - X2 + X - 1). + 27 6c3 = (2 a)3 + (3 bc)^ = (2 a + 3 6c) [(2 a)2 - (2 a) (3 be) + (3 6c)2] = (2 a + 3 6c) (4 a2 - 6 a6c + 9 62c2).
x6 X6
2.
Example.
1.
64x8-1252/3.
27x^
1.
3.
Wx*-8ly*.
imme-
439
Tlie following theorems are an diate consequence of 438, (1), (2), (3) and 436.
When
1.
is
composite.
If n
is
divisible by a"
multiiile
b^.
of any integer,
p,
then a"
b"
is
exactly
183
a;2y2
,/4).
2.
7/*
is
h" ts
exactly divisible by oP
Thus,
X6
b^.
2/6
(X3)2
(2/3)2
= (x3 +
3.
2/^)
(x3
- 2/3).
a"
an odd multiple of any integer, p, then 7/" exactly divisible by a^ + b^, whether n itself is odd or
is
is
even.
Thus,
X6
2/6
= =
(a;2)3
(2/2)3
(a;4
(X2
^2)
^ZyO
2/*).
4.
i/"
is
a power of
2,
then a"
+ b"
^8
2/*
- V2 x22/2).
Again,
X*
y*
+ V2x2i/2 = =
and so on.
As each
o"
6" is
When
or a"
is
composite,
it is
b",
into
two
factors
best to begin by resolving a" + 5", whose degrees are as nearly equal as
jjossible.
It will
y^ given
under 2
is
the best.
Continu-
ing,
y6
(a;3
(x
2/2)
(X
y) (x2
+ xy +
y^).
184
Example.
1.
A COLLEGE ALGEBRA
Factor the following expressions.
X*
y'^.
2.
x^
2/8.
3.
x9
y^.
440
The theorems
the form
integer p.
Thus,
a'"
of 438, 439 also apply to expressions of h" when and n are multiples of the same
m
= =
X6
yl*
{X2)3
(?/5)3
(X2
2/5)
(X*
+ x2!/5 +
2/10).
EXERCISE XVII
In the following examples carry the factorization as far as without introducing irrational or imaginary coefficients.
1.
is
possible
x3y/
20
x22/2 3.
+
x2
25 xy^.
2.
28
te2
63
ty^.
2/2
9 z2
4 x?/
12 yz
6 zx.
4.
5.
(7 a2
2 62)2
(2 a2
+
6. 8.
7 62)2.
(7x2
X*
4-4x-3)2-(x2 + 4x + 3)2.
x2
4(1
a*
62
a6)
a2.
7. 9.
+
+
1.
-& a-W +
6*.
a*
4 a2
16.
10.
9 x*
11.
4 (a6
cd)2
(a2 4. ^2
c2
+ Ibx^y^- + _ ^2)2.
14.
17.
16 y*.
12. 15.
576xV-9
xw
2/^^-
13.
16.
x^
x5
2/9.
x^'^
y^^
2/".
2^w.
32.
x'^
Sometimes the terms of a polynomial in x can be combined which have some common factor, as F. This
is
Compare
433.
1.
Example
Factor x^
+ 3 x2 -
2x
6.
first
two,
we have
x8
3 x2
2X
= =
x2(x
(x2
3)
2 (x
3)
2) (X
+ 3) = (X + V2) (X - V2) (X +
3).
185
Factor x3
2x2
+ 2x +
l.
like coeflBcients,
l)
we have
X (X
2x2
+ 2x +
= = =
x3
1) 4- 2
1)
(X
1)
(X2
+X+
1) (X
1).
this can be accomplished by first separating one of the given terms into tw^o terms.
Sometimes
Example
3.
Factor x^
4 x^
5x
+
^2
6.
We
have
X3
4X2
+ 5X +
G =:X3
+ +
3x2
_|_
-)-
3X
-f
2X
=
=r
x2 (X
(x2
+
X
3)
X (x
3)
2 (X
3)
2) (X
3).
Example
4.
Factor x*
2 x^
3 x2
1.
We
x*
have
2x3
+ 3x2 + 2x +
3. 5.
2.
1.
4.
?/.
6. 8.
x*ij-
^-y^
+
+
2 x2
3X
2.
y{. x5
4x3
10x2
12x
9.
This
is
442
a) (x
+ h) =
oc"^
we
shall
know
the factors of x^
h,
numbers, a and
such that a
find
two
186
A COLLEGE ALGEBRA
Two such numbers always exist, 444, though they are seldom rational. But when rational they are integers, 454, and may be found by inspection, as in the following examples.
Example
1.
Factor
x"^
+ lSx +
42.
seek two integers, a and b, whose product is 42 and sum 13. As both ab and a + b are positive, both a and b must be positive. Hence among the positive integers whose product is 42 namely, 42 and 1, 21 and 2,
14 and
3, 7
We
and 6
is
13,
6).
and find
and
6.
Hence
13 x
42
(x
7) (x
Example
their
2.
Factor x^
b
13 x
22.
;
sum
negative.
for their product is positive and Hence, testing as before the pairs of negative inte2
(x
;
must be negative
for
11
13.
- 2).
Factor x2
is
9x
22.
Here, since ab
6 is
22
must have opposite signs and numerically gi-eater must be 22 x 1 = 11 x 2, and, testing as
;
is
Example
1.
x2
4.
x2
+ 3x + 2. + X - 30.
x2-16x + x2 + 20 X +
15.
3.
96.
6.
x2-4x-12. x^ - 21 x + 80.
443
The quadratic ax^ + bx + c factored by inspection. This is sometimes possible, when a, h, and c are integers. By multiplying and dividing by a, we may reduce ax'' -\-hx-\-c
to
h (ax) -{- ac'\ / a, and then factor the the form \_(adcY bracketed expression with respect to ax by the method just
whose product
is
and sum
1.
b.
Example
Factor 2 x^
3.
We have
2x2-|-7x
(2x)2
7(2x)
= <?^^M^^ll = (x +
2
3)(2x
l).
187
Factor
abx'^
We have
abx^
.o
ro^
oro
(a2
b^)x
ab
^ _
{abx
^^
ab
g-) (abx
6^)
ab
= {bx +
Example
3.
a) {ax
b).
Factor 16 x^
it is
72 x
63.
In this case
we have
16x2
+ 72x-63 = =
(4x)2
+ 18(4x)-63
(4x
+ 21)(4x-3).
2. 4.
Example
1.
4.
3.
5.
6x2-13x + 6. 14 x2 + X - 3.
49 x2
105 X
44.
6.
be.
The quadratic x^
pleting the square.
+ px +
q or ax^
bx
c factored
by comin
444
perfectly general.
Since
(^"^9) =^^+/'^ +
't"
we can make
x^ -\- px a perfect square by adding --' ^^^^ the square of half the coefficient of x. This process is called comjileting the square of x^ px.
i^,
1.
We
a;"
+ ^;x +
?? if
we both
add and
siibtract jo^/4.
But by
this
x'^+px
+ q = x^+px + ^ + q
(
.(.-fft^2?S)(.tS-^E5>. (
188
A COLLEGE ALGEBRA
Since
ax^
2.
bx
-\-
= a{
y
x^
-\
x a a
-\
)>
we may
b /a for
2^
obtain the factors of this expression by substituting and c/a for q in (1). Simplifying the result, we
have
ax^ + bx
+c=
1.
a[x
+ -+
^^
+ -j[..
^^J-
(2)
Example
Factor
x^
We have
Example
2.
Factor x^
x^
We have
8x
Here we
this
sum
obtain a
sum
4,
i-,
The
factors
are imaginary.
Example
3.
Factor 3x2
5x +
1
1.
We have
3x2
-^
5x +
3 rx2
- -x +
^1
-(^-^f)(^-^f)
Example
1.
4.
x2
x2
+ 10x +
3. 5.
+ 45. 2 x2 + 3 X + 2.
12 X
6.
x2-10x + + X + 1. x2 - 4 ax x2
24.
62 4.
8 a6.
189
445
two variables.
The methods
form
to quadratics of the
+ hxy + cy^.
+
x'^
Example
1.
Factor x^
x2
8x2/
14
2/2.
We
have
xy
+Uy^ =
= =
-8xy +
(4
16
y'^
2/2
(x-42/)2_22/2
[X
+ V2) y]
2.
[X
(4
- V2) y]
Example
2.
1.
x?/
4 y-.
x2
xy
2/2.
Such
446
functions are ordinarily prime. But when composite, they may be factored as in the following example.
Example
If
1.
Factor
=x'^
+ 2xy -
8y'^
+ 2x +
Uy - S.
the product of two polynomials of the first Moreover its terms of the second degree, x2 + 2 xy 8 y'^, must degree. be the product of the terms of the first degree in these polynomials.
is
composite,
it
is
We
that
find
by inspection that
x^
2 xy
8 y^
(x
i y) {x
I
2 y).
Hence,
if
is
and m, such
we
x^
shall
have
+ 2xy
-8y'i
Im.
(1)
But
to
l
make
(1)
285,
+m=
(2)
2 (2),
(3)
find
I
From
for
and
we
= 14 (3), 2 + 4 = 1, m = S. And
Z
??i
Im
=-
(4).
=-
3.
Therefore x2
Note.
are.
If,
(x
+ 4?/-
l){x
-2y + S).
The example shows how exceptional these composite functions 3, leaving A otherwise unchanged, we replace the last term,
becomes prime
190
Thus, to factor x^
A COLLEGE ALGEBRA
+ xy 2
+ 2xy - Sy"^ + 2xz + Uyz - Sz"^, we set - 3 22 = (X + 4 + Za) (x - 2 + 4- 2 xz + 14 7/z = 3. and then proceed as above, again finding = 1, Example 2. Factor 2 x^ - 7 xy + 3 y^ + 5 xz - 5 yz + 2 z^. Example 3. Show that x- y- + 2x + y 1 is prime.
x2
8 y2
2/
2/
mz)
??i
447
Polynomials
of the
nth degree.
aox"^
The like is true of polynomials X of every degree, though no general method exists for
Every polynomial in
f (x)
Theorem.
x,
aox"
+ aix-i
-\
a_ix
+ a
is,
product ofn factors of the first degree; that x /3^, such that n binomials, x jSi, x (3^, ,
is the
there are
f (X)
ao(x
/?i)
(X
^2)-
(x
- ^).
The
448
Corollary.
y, of the
A homogeneous polynomial in two variables, x and nth degree, is the product of n homogeneous factors in
+
aixh/
X and
a^xy'^
a^y^
may
a^x"^
aox
as
y^.
But by
if
447, aoX^
OiX^
a^x
03
= ao (x ^1) (x
we
substitute
x/y
by
y^,
we
obtain
aox3
aix~y
a^xij'^
azy^
EXERCISE XIX
Factor the following expressions.
1.
x2
14 X
48.
2. 4.
6.
x2
21 X
120.
.3.
5.
5x2-63x-22.
64x2-21x +
2.
16x2
12x2
191
X*
x2
13 x2
3X
+
3.
36.
8.
x-^y
x'^y"-
3.
18 xy^.
11.
13.
15.
10.
3 x2
X-
2X
12.
14. 16.
6 ox
9 62
_
6x
18 ah.
^2
6d
dx
cx2
cdx.
+ 152/2 + 2x-
4?/-3.
SYNTHETIC DIVISION
On finding factors by aid of the remainder theorem. Let f{x) denote a polynomial in x. By the remainder theorem, 415, if h denote a number such that y(i) = 0, then cc is a factor of
ft
44S
f{x)'
We
~
by inspection.
Example.
"*"
"^
1 1
1
^
4,
L
X2
5x + 4. Since /(I) = 1 + 4 = 0, x 1 is a factor of /(x). 1, we obtain the quotient Dividing /(x) by x + X-4. - 1) (x2 + X - 4). jjgj^^g ^^^^ = (x
x^
-5
Note.
sum
450
If asked to factor a Polynomials with integral coefficients. polynomial, /(x), with integral coefficients, it is usually best to look first for any factors of the first degree with integral These ma}' always be found by coefficients that it may have.
451
coefficients,
o/f(x).
Thus,
/(x),
let
/(x)
Oox^
+ =
4-
aix2
a^x
+
+
03.
If
is
to be a factor of
we must
have, 415,
/(6)
ao63
+
+
ai62
a^b
as
and therefore
Therefore, since O062
ai6
{a^ffl
aib
aa) 6
= 0, = az.
6 is a factor of 03.
ag denotes
an integer,
192
Hence
all
A COLLEGE ALGEBRA
such factors x
b may
3a;<
ing example.
Example.
Factor
/(a;)
3a;6
ISx^
lli2
lOx
6.
and b
test
We
_ 3 _ 1,3 _ 11 _ 3+ 6+ 7+ 3 36^^ 1^ 4 3
3
10
4
6
Since /(I)
5^ 0,
1 is
not a factor.
01
013
We therefore begin by testing X ( 1) or X + L The division proves to be exact, the quotient being Qj = 3x* 6x3 7x2 4 X 6, and the remainder
H^^ce X + 1 is one factor of /(x), ^^^ ^^ '^ ^^^ product of the remaining
^'
factors.
We
X
X
Qi.
It also
+ +
1,
To
1,
the quotient being Q2 = 3x^ - Ox^ + 2x - 6. factor Q2, whose constant term is also 6,
but in each case obtain a remainder which Hence none of these are factors. But testing x 3, we find that
2, 2,
test successively
is it
not
0.
divides
3x2
2.
Therefore
= (x+
l)2(x-3)(3x2
2).
452
(- a, with integral polynomial f{x) = a^x^ + a^x"-^ H coefficients, may have a factor of the form ax /3, where a and (i denote integers which have no common factor. But
if so,
a must
be a factor
of
ao,
and
fi
a factor of
a.
This
451.
03.
is to
Thus, let /(x) = flox' + aix2 + ojX + be a factor of /(x), we must have,
If
ax -
/3,
or
a (x -
/3/a),
415,
^ - + as = 0, a = 0. and therefore Oq/S^ + aiP-a + a2/3n-2 + asa^ (1) From (1) we obtain ao^^= (ai/32 + uo^a + 3a-2)Q-. (2) Therefore, since 01^32 + OoPcr + a^a"^ is an integer, nr is a factor of ao,8'But ct has no factor in common with 2. Hence a is a factor 492,
/(D
i3\ *
flo
,
i33
a*
^2
ai
^, a^
a2
/3'',
of ao, 492,
1.
Again from
(1),
(aoi32
ax^a
a2f^^)/3
=/3
azo^,
is
(3)
03.
a factor of
193
such factors ax
/3 may
5 x^
be found as in the
following example.
Example.
If
Factor f{x)
6 x*
3 x^
3x
2.
ax /3 is to be a factor of /(x), a must have one of the values 1, 2, 3, 6, and /3 one of the values 1, 2 therefore p/a must have one of the values 1, 2, 1/2, 1/3, 2/3, 1/6. We may test ax ^ for these various values of ^/a by dividing /(x)
;
by X /3/a
If the division is exact and Q denotes the synthetically. ,8 is a factor of f{x) and Q/a is the product of the quotient, then ax
2,
+ 2,
none of
+ -
+ 3-3-2-1-1/2 - 1 -
But
x+
1/2
^^gg^ 2x
^j^g
quotient
4.
We
ax
/3
is
we
of the values 1, 2, 1/3, 2/3. But 1, x + 1, x 2, x + 2 are not factors, since already know that x 1/3, x + 1/3, we find that they are not factors of /(x). Testing x neither of them divides Qi/'2 exactly; but x 2/3 does, the quotient
being Q2
3x'-
+ 3x +
fix)
3.
Hence 3x
is
is
(^2/3
x-
(2
1) (3
2) {x2
1).
is
often becomes evident before a division by x completed that the division cannot be exact.
Note.
It
6 or
^/a
453
suffices to
prove that x
8 exactly
;
2 will not
divide 5 x^
2,
4 x-
for since
the
"divisor"
Q already
found,
namely
6,
194
454
A COLLEGE ALGEBRA
h a^, It follows from 452 that a polynomial f (x) = x H whose leading coefficient is 1, the rest being integers, cannot vanish for a rational fractional value of x.
For
if
/(/3/a)
0,
therefore
a must be
then f{x) must be exactly divisible by ax - /3 pnd a factor of 1, which is only possible when a = ^. 1.
It follows
455
from
350 that the problem of resolving a polynomial f{x) into its factors of the first degree is essentially the same as that oi
solving the equation f(x)
=
=
0.
Example
1.
Solve fix)
2 x*
+
x
a;3
17 x2
16 x
12
= 0.
By
(2
1) (x
2)2 (x
3).
2x-l = 0,
Therefore the roots
Example
2.
Solve x^
Transposing,
Factoring,
x^
EXERCISE
Factor the following expressions.
1.
XX
i3-7x +
X*
3.
5.
7.
2. x3 + 6x2 4- llx + 6. 6. + 35x2 _ 50x + 24. 4. x* - 2 x2 + 3 x - 2. 6. 2x3 - 5x2y - 2x2/2 + 22/8. 6x3-13x2-14x-3. 8. 4x6 - 41x* + 46x2 - 9. 2x*-x3-9x2 + 13x-5. 9. 6x5 + 19x* + 22x3 + 23x2+ 16x + 4. - 4. 10. 5x6 - 7 x5 - 8x* - x3 + 7 x2 + 8 X
10x3
0.
12. 14.
24.
15. 17.
x8-9x2 + 26x =
16.
18.
195
in the
is
possible without
6 xy
a3 (a
3. 5.
+ 15 X - 4 - 10. - 6) + h^ {b - a).
2/
2.
a"-bc
ac"-d
- obH +
hcd"^.
4.
6. 8.
a^
81 a6*.
a*6
a-b^
+ aW -
a.
3 abx^
(x2
6 axy
+ bxy-2 y\
7.
9.
3x6-192 2/6,
64x62/3_yi5.
+ x)3-8.
10.
12.
x'^-{a-b)x-ab.
x2-3x"-18.
3x*
2 xc
+ 3x3-24x-24. - a2 + x2 - 2 a6 +
&2--5(f
14.
c2
72.
62,
iq
18.
(^2
20)
+ +
64.
a2-2a6 +
6x2
7a;2/
52/2
+ 56 + G. - 4x - 2?/.
x*
10 x2?/2
(a2
9 y*.
^20.
22. 24.
b^)x'^
a262.
21.
23.
25. 27.
14x2
xy^
(x2
+ 19x-3.
a*
+ 19 y - 66 - 18 a262c2 +
+
2/)''
2/2.
81 6<c*.
8 (X
ftx
27 (X
2/)3.
{x
-2y)x^ -
- 2x)yK
29.
31.
x2
x5
+ -
a2
x''
_ -
a6
2 ax.
x^
6*
y^
(:x
yf.
X-'
2 x2
1.
32. 34.
36.
33.
35.
37. 39.
+ 9x + 22/-5. 25
0^
+ 62 + + 4.
i.
4 a*
x^
+ +
3 a262
96*.
8 ax
40 a6
62.
38.
l)2.
x*
1.
41.
x3
40.
42. 44.
(ax
x^
62/)2
(6x
a2/)2,
bx^
a^x
a36.
43. 45.
47.
a2-962+ 126c-4c2. x* - 2 x3 + 3 x2 - 2 X +
x2
1.
46.
48. 50.
(ax
4x5+4 x-37x3-37x2+9x+9.
49.
51.
abcz^
(a262
c^) x
ahc.
52.
196
53.
A COLLEGE ALGEBRA
(a
6) x2
2 ax
55.
x*-6x3 +
57.
58.
7x2
3x5
5x*
aJ^bc
54. 56.
x^s
yis.
4x3-3x-l.
iqx
3x2
8.
15x2
59.
60.
acd
X*
VII.
456
Let A, B, denote rational, inteHighest common factor. gral functions of one or more variables, as x, or x and y. have no factor in common, we say that they are li A, B, prime to one another. But if they have any common factor,
is
highest
we
common factor
Thus, x2
(H.C.F.).
+ 2/2 and x + ?/ are prime to one another. But 4 xi/z^, 8 xz*, and 4 x'^yz^ have the common factors xz2, xz^, and their highest common factor is xz^.
457
Notes.
2.
1. We here ignore common numerical factors. sometimes said of two or more functions which are prime to one another that their H.C.F. is 1. 3. The numerical value of the H.C.F. of A and B is not necessarily the greatest common divisor of integral numerical values of A and B. Thus, the H.C.F. of (2x + l)x and (x l)x is x. But when x = 4, the values of (2 x + l)x and (x l)x are 36 and 12, and the greatest common
It is
divisor of 36
and 12
\.
is
not
4,
but
12.
458
The H.C.F. of K, B, is the product of all the different common jmnie factors of K, B, , each raised to the loivest 2'>ower in irhich it occtirs ij2 any of these f met ions.
Theorem
The truth
of this theorem
is
obvious
if
we suppose each
form of a product of
197
assume
different
in 430,
Thus, the different common prime factors of xyz^, xz*, and x^yz^ are X and z, and the lowest powers of x and z in any of these functions are x and z^. Hence the H.C.F. is xz^.
Observe that
than one
if
it
way
in
terms of
more
to various results
corresponding to the various ways and there might be more than one common factor
When
may be
459
Example
We
and
have
6 x*y^
X*
?/
is
x^y (x
3 y).
Example
2.
1.
2.
3.
2/2,
x^
6,
2 xy
y^,
and
x^
x2
6X
8,
and
x^
+ +
t/^.
4.
9x2
x + 6. + 7x +
6.
prime factors of one of the functions A, B, are we can find by division or the remainder theorem known, which of them, if any, are factors of all the other functions. The H.C.F. may then be obtained by aid of 458.
If the
460
Example
f{x)
1.
= x2 we
^
and
<;&
(x)
By
have/(.c)
(x
1) (x
= x* - 3 x^ + 5 x2 - 8 x + 2). Testing x = 1
Hence the H.C.F.
is
5.
and x
x
=2
(1)
0,
but
<}){2) jt 0.
1.
Example
/(x)
2.
x2
+ 4x +
and
(^(x)
x*
Sx^
9x2
+ 8x +
4.
Since /(x) = (x + 2)2, we must find not only whether x + 2 is a factor of Dividing <^ /xi bv a: ^ a (xjj but whether it is a factor once or twice.
198
A COLLEGE ALGEBRA
obtain Qi
(synthetically) we Qi by X + 2 we obtain
i^
Q^^x^ +
2)2.
of/(x) and0(x)is
(x
Example
1.
x"^
3.
2. 3.
x2
(X
+ X-6 and 2 x^ + 7 x^ + 4 x + 3. + 5x + 6 and x*4- 6x3 + 13x2 + 16x + 12. - l)2(x - 3)3(3x + 1)2 and x< - Sx^ + x2 + 21x 2.
18.
461
Theorem
Let
and
and
denote two given integral functions, any two integral functions or constants. Then
is
A and B
every
a factor of
MA + NB.
B.
Por
F denote
common
factor of
A and
Then
where
A=
MA
GF
and
B=
HF,
Hence
where
A^ + NH
F is
a factor of
is
integral.
Therefore
MA + NB,
424.
462
By aid of this theorem the Applications of this theorem. in x, whose problem of finding the H.C.F. of two polynomials
of factoring a degrees are the same, may be reduced to that polynomial of a lower degree. single
Example
Hence,
1.
^ = x2 + 2 x -
4
\.
and
B=
x^
3.
Subtracting
B from
x
A, we obtain
A B = x
461,
1 is
common
factor of
and B.
1 is
not a factor of A.
Example
.4
1.
2.
= 2x3-3x2-3x +
begin by multiplying
and
We
and
Then
3
subtracting 2
5 x2
^ - 2B=-
5X
10
=-
6 (x2
2)
=-
5 (X
1) (X
2).
and x
2.
199
we
and B.
+S=
5x3
5x2
l)(x-2).
It is at
not a factor of
2.
or B, so that as before
we have
only to test x
3.
and x -
Example
1.
X*
-x^ + 3x2 _ 4x -
and
x*
x3
2.
6x3
25x2
^ 5x +
and 4x3
15x2
+ 2x2 + 3x - 2x + 8.
22.
Theorem
the
3.
related that
A = QB +
the
If the four mtegral functions A, B, Q, R are so R, the common factors of A and B are
463
same as
have
We
A
from
= QB + R,
(1)
(2)
and therefore
It follows
A- QB= R.
(2),
by 461, that every common factor of A and Bis Si factor of R and therefore a common factor of B and R. And, conversely, it follows from (1), by 461, that every common factor of B and i2 is a factor of A, and therefore
a
of
common
factor of
A and
B.
factors of
A and B
are the
same
as those
B and
R.
The general method for finding the H.C.F. of two polynomials When one polynomial in x is divided by another, the dividend, divisor, quotient, and remainder are connected by Hence it follows from 463 that the identity A ^ QB -{- R.
in X.
464
The common factors of dividend and divisor are always same as those of divisor and remainder.
mials in X
the
By making use of this fact, the H.C.F. of any two polynomay always be found. The method is analogous to that employed in arithmetic to find the greatest common
divisor of
where
yl
two integers. It is described in the following rule, and 5 represent the given polynomials, A the one of
if
higher degree
200
465
Rule.
A COLLEGE ALGEBRA
Divide
K
B
hy
^ and
and
call
the qriotient
q and the
qi
remainder Rj.
Next divide
^
by
E.^
call
the
quotient
and the
E-i
new remainder by
I
by Rgj *'^^ *o on, continually dividing each the one last obtained, until a remainder is
x. is
If
this final
remainder
not 0,
and
have no common
the H.C.F. of
factor.
If
B.
it
is 0, the divisor
which yielded
it is
A and
For suppose, for the sake of definiteness, that the final Then according as (1) R^ = c, where c is R^denotes a constant not 0, or (2) R^ = 0, we shall have
remainder
(1)
(1)
In
qB
or
(2)
B =
B =
For it follows from the identities (1), by 463, that A and have the same common factors as B and Ri B and Ri, as Ri and i?2 ^i and Ro, as i?2 and c.
Hence the
Ro and c, But as
factor.
(2)
all
c
pairs of functions
R^,
R^ and Ro,
have no common
the H.C.F. of
and B.
of
q2.R2j every factor of R^ is a common factor and R^ itself is the common factor of highest
degree.
But
as the
common
factors of ^1
those of
A and
and Ro is also the factor of highest degree common Hence R^ is the H.C.F. of A and B.
common to R^ to A and B.
201
X'
and
x^
+ 2x +
3.
Writiug divisors at tlie left of dividends, we have B = x'^ + x + l\x^ + x:' + 2x + S\x- q X a;^ + g^ +
Ri=
3|a;2
X?-
+ x+ l \x-2 =
+
3X
1
qi
- 2x +
-R2
-2x -6
=
+
x
7
As
i?2, is
not
0,
x^
and
x^
x^
2x
have no
common
2.
factor.
Example
x2
2X
in
and
1,
x3
2 x2
3X
2.
B = x3 +
x2
Rx
x^
X
i?2
|x3
X3
+ x2 + 2x + + X2 2x +
2|x
2lx2
X?
+ x |x/2 +X
i?3=
Here the division by
numerical factor 2 in
i?2 is
i?2,
0.
Here
we have an
tions of a
single variable
that
actual proof
for
f unc-
466
if
have any common factor, they have a highest common factor; for in 463, 465 it is not assumed as in 458 that an integral function can be expressed in but one way in terms of its
prime
factors.
465
it is
shown that
list
467
E,2)
Every two consecutive functions in the have the same H.C.F. as A and B.
2.
A, B, Ei,
A and B
is
of A and B.
Abridgments
of
of this
or
iJ
are obvious
468
202
factors
A COLLEGE ALGEBRA
and then
find the H.C.F. of the resulting expressions.
The result thus obtained, multiplied by such of the factors removed at the outset as are common to A and B, will be the H.C.F. of A and B, 458. The same course may be followed with any two consecutive since every two such functions in the list -1, B, R^, R^, functions have the same H.C.F. as A and B, 467.
Thus, A=x^ + x^ + 2x'^ + 2x and B = x* + 2x^ + Sx'^ -\- 2x obviously have the common factor x. Removing it we have x^ + x- + 2x + 2 and to be x + 1 (see a;3 -I- 2x2 + 3x + 2, whose H.C.F. we have just found and B is x(x + 1). 465, Ex. 2). Hence the H.C.F. of it cannot be a Again in Ex. 2, since x is a factor of Ri, but not of B, factor of the H.C.F. of B and i?i, and is therefore not a factor of the H.C.F. of A and B. Hence we may discard this factor x of Ri and divide B by the remaining factor x + 1, so lessening the number of divisions.
2.
divisor or dividend or
ical factor
;
In any of the divisions we may multiply or divide the any intermediate remainder by a numerfor this will affect the subsequent remainders
by
a numerical factor at most, 403, and therefore the H.C.F. not This device enables us to avoid fractional coefficients at all.
It is
Example
^ = X* +
Multiplying
2
2x2
+ 3x +
and
J?
2x3
5x2
1.
.4
coefficients,
we have
+ 5-1 -l|2 + 6+ 4+ 2 + 5- 1 -
+ +
2[l
2
1+ 5+
2
+ 10+
5 )5
14
+
+ +
2+ 5- 1-1
+
15 3
4[1
5
1+
Hence the H.C.F.
is
l|2
2
of
^
1.
and
x2
3X
-1-3-1
203
arranged, as follows;
1
+ 5+6+
204
A COLLEGE ALGEBRA
;
For (1), D^ is a factor of Zi and C and D^ is a factor of A and B. Hence, 427, D^ is a factor of ^, B, and C. And (2), every common factor of A, B, and C is a common
factor of
>!
and
C,
D^, 467.
Hence
Dg
the highest common factor of A, B, and C. The same conclusion follows from 458.
is
Example.
oi
7
A = x^ +
+
2,
x^
x"
x^
2,
B=
By
2 x*
5 x^
2 x2
and
C=
3 x*
x2
2.
465,
we
H.C.F. oi is x - 1.
A
x
a,nd
B is Bi-
x-
2;
and
of
A, B, and
is
1.
471
The polynomials in more than one variable. general problem of finding the H.C.F. of two such polynomials But the H.C.F. of is too complicated to be considered here.
The H.C.F.
of
two polynomials which are liomogeneons functions of two variables, as X and y, may readily be found by aid of the rule
given in
465.
EXERCISE XXn
Find the H.C.F. of the following.
1.
10
(a
2/*
x^yV,
4 x^yz^, 6 x'^yH^,
and 8
x*2/^z*u.
2.
3.
6)2 (a
2/2
6), (a
+
-
+
a2
and y^
2
+ 1,
h) {a
\.
6)2,
and
a?b
ahK
4.
5.
6.
a2
1,
and
ax
a^
+
1.
1.
x^
X*
and
x^
ax^
+ -
7.
8.
9.
x2
(X
5 X
4.
x3
(x2
and
1)2 (X
x^
4 x2
4 x
5.
+
-
1)2
and
(x^ (x2
(X
l)2{x
2)2
and
2 x3
z8
3 x2
11 X
and 4
+
x2
3 x2
9X
1.
2.
2 x2
2 X
3 and 2 x8
205
3 x3
a;*
+
x3
2 x2
19 X
+
+
and 2
x3
+ +
x2
13
a;
15.
3 x2
a;
2 and 2 X*
3 x3
X'-
^ -
6;
3X
1.
16.
17.
3 x3 3 x3 6 x5
+ + -
13 x2
8 x2
7 x*
18. 19.
20.
3 X.
21.
22. 23.
3 x3
x3
- x2 ax2
12 X 3X
+ 4, x3 - 2 x2 - 5 X + 6, - 3 a, x3 - x2 - 3 X + 3,
4
xy'^
and
7 x3 4-
19 x2
8X
- 4.
and
7 x*y
- 6 xV _ 18 x2;/3 +
1) (x3
and 14 x^y
+ - 19 x^-y^ - 32 xy^ + 28
x^ x2
3x
3.
?/*.
X (x
4 x3
8 x2
A
is
common multiple
of
two or more
is
472
exactly divisible by each of the functions A, B, . Among such common multiples there is one whose degree is lowest. call this the lowest common multiple (L.C.M.) of
We
A,B,....
Theorem
A, B,
,
A, B,
The L.C.M. of tioo or more integral functions, the product of all the different prime factors of each raised to the highest power in ivhich it occurs in
1.
473
is
at
occurs in that function, hence all the And the common multiple
is,
the L.C.M.,
is
no factors besides
these.
206
A COLLEGE ALGEBRA
Here, as previously, we ignore numerical factors and assume that an integral function can be expressed in only one way as a product of powers of its different prime factors.
474
we may
different prime factors, each raised to the highest occurs in any of the functions, are x^, y*, z^.
is x^y^z'^.
power
iu
Example
"We have
2.
xhj'^
xV -4x2/2 + 4 2/2=
is 2/2 (x
(x
2)2 (x
- 4 xy- + 4 ?/2 and x'^y -iy. _ 2)2 and x22/ - 4 = (x -2)(x4-2). + 2).
?/ ?/
475
is
Theorem
2.
and B,
For
let
A and
B,
and
let A-^
A and B by
and B^ D, so that
A = A^D and
Then
if
B= B^D.
For evidently a common multiple of A and B must contain the product of all prime factors common to A and B, namely D, (2) the product of all prime factors of A not belonging to B, namely Ai, (3) the product of all prime factors of B not belonging to A, namely 7?i and the lowest common
(1)
;
476
is
Corollary.
iiitegral functions,
and B,
equal
to
and
their H.C.F.
477
General method for finding the L.C.M. of two polynomials in x. from 4G5, 475 that the L.C.M. of two such polynomials, A and B, may always be obtained by the rule
It follows
:
and B, divide
by B.
hij
the H.C.F. of
A and
B,
and
207
Observe that this is equivalent to multiplying B by all those prime factors of A which are not already present in B.
Example.
X*
3x3
2x'-
By
465,
we
+ 5x- ~x-l. + 3x + 1. -x- i)/(x2 + 3x+ l) = 2x-l. - 1). (x* + 3x3 + 2x2 + 3x + 1) (2x
+ 3x +
1
and 2x^
x2
The L.C.M.
of
x.
This
may
be
478
on.
The
This follows from the fact that each step in the process is equivalent to multiplying the L.C.M. last obtained by those prime factors of the next function which are not alreadypresent in that L.C.M.
3 x3 + 2 x2 + 3 x + 1, = 2x3-3x2 + 2x-3. As we have just shown, 477, Ex., the L.C.M. of A and B is Ml = (X* + 3 x3 + 2 x2 + 3 X + 1) (2 X - 1).
Example.
^ = x* +
B=
2 x3
5 x2
1,
and C
have next to find the L.C.M. of Mi and C. we find that 2 x 1 is prime to C, and by 465 we find that the H.C.F. of x* + 3 x3 + 2 x2 + 3 x + 1 and C is x2 + 1.
We
By
division
Furthermore,
of
2x
3.
therefore of
(X*
3 x3
2 x2
3X
1) (2
1) (2
A, B, C, is X - 3).
Observe that we do not multiply the factors of proceeding to find the H.C.F. of Mi and C.
Mi
together before
EXERCISE XXra
Find the L.C.M.
1.
of the following.
i^
3x (a
1,
9x2
and 9x2
(a
i.
2.
6) (aS
65)
and
b) (ofi
lf>).
208
3.
A COLLEGE ALGEBRA
a^
-^ a"^
a, a^
o^,
(X*
and
a^
a'.
4.
5. 6.
_ y3) (X (x3
x2
y)\
y*) (x
6,
yy, and
x^
(x2
3.
^2)2.
3X
(2/
2, x2
z)2,
5X
and
4x
{x
+
+
x2
y"--{z
92/2,
1
x)2,
and
z^-
y)^
7.
8. 9.
2x2
x^
+ ^xy x2
3x2
and
+ g^y - 3?/2, x^ - x2 + x - L
and 6x2
- Uxy + 3y\
2 a2x
10.
11.
8x3
x3
x6
+ 2 x2y + 3 y2x + 3 a"-y and (2 x2 - 3 a^^) + (2 a2 - 3 y-)x. - 18x2/2, 8x3 8x22/ - 6x^2, and 8x2 -2xy - 152/2,
?/
_|_
2/')
^^
y^i
and
x*
+
4
x22/2
5,
?/*.
12.
13.
14.
1,
3x3
5x2
- 3x
36x2
and
x*
1.
81,
and 6x2
8
a"^,
4 a-, x3 2 X, x2
2 ax2
+ 4 a2x +
+
2
6,
6x
2X
and
aK
+ 3x + 2) (x2 + 7 X + 12) and (x2 + 5x + 6) (2x2 - 3x - 5). - 8)(27x3 + 1) and (2x3 + 5x2 + lox + 4) (x3 - x2 - x - 2).
4-
18. 19.
x3-6x2
x*
11 X
7x- 2,
+ 4X
and 2x3
+ x2 -
13x
+ 6.
5 x2
4X
2 X*
x3
3.
10 x2
5,
and
2 X*
20.
2x*
X*
+ x3 + 7 x2 + 3 X + - x3 + 2x2 + 3x - 2,
+
3 x3
2x*
3x3
4x2
i3x
6,
and
x2
5X
6.
denote polynomials in
to
x.
479
Fundamental Theorem.
tions,
and N, can
be
MA + NB =
1.
For if we apply the method of 465 to A and B, we final remainder a constant, c, different from 0.
shall obtain as
209
and use the
c is
A =qB +
B
Ei
=5ijRi
Ri,
/?2,
c,
and therefore
4.
5.
2.
3.
= q2R2
+ +
6.
Substitute in 6 the value of Eo given by 5, collecting the i?i and the B terms, and in the result substitute the value of Ri given by 4,
collecting the
and
c
terms.
~
We
thus obtain
Divide both sides of this last identity by c, and for (1 + q\q-i)/c and + 52 + Q'hq'i) / c, which are integral functions since c is a constant, write and N. AVe obtain
(?
and are integral functions. where, as just said, And we may demonstrate the theorem in the same
stant remainder,
c, is
l=MA+ NB, N
way when
the con-
Conversely, If
MA + NB =
B.
of
where
and
are integral,
480
then
is 2)riine to
MA + NB^
The following theorems are some of the more important consequences of the fundamental theorem just demonstrated.
Theorem
divisible
1.
If
is
prime
is
to
B,
and
the product
AC
is
481
by B, then
divisible by B.
For since
is
prime to
i?,
we can
find
M and N,
-l,
479,
such that
MA
and therefore
+ NB =
B.
M AC + NC
AC and
If
B=C.
Hence
it is
But
is
a factor of both
2.
to
a factor of C, 461.
Theorem
it is
prime
fu7ictions
B and
C,
482
210
For since
A COLLEGE ALGEBRA A
is
prime
to B,
we can
find
M and N,
479,
such that
MA
and therefore
Hence,
461.
if
+ NB=l,
it
MC A + NBC=C.
and
is
BC
had a common
factor,
is
would be contained
to C.
in C,
But
this
impossible, since
prime
483
Corollary.
If
is
prune
to
and
so on,
it is
B C D
it is
BCD
484
and so on.
3.
Theorem
and
prime factors.
For
If
let
its
degree.
P is composite it has
;
some factor A.
If y1. in turn, is
composite,
it
Continuing thus, we mu.st ultimately come upon a prime function for the degrees of the successive functions P, A, B, cannot fall below 1. begin with the finite number n, decrease, and Let F denote this prime function. It is one of the prime factors of P,
has some factor B.
427,
and we have
if
M= F'M',
Similarly
P= FM,
where
is
integral.
is
and therefore
P = FF'M',
where M'
is
integral.
Continuing thus, we reach the conclusion that a series of prime functions F, F', F", exists, whose number cannot exceed n, such that
P = FF'- F"
Hence P has at least one set of prime factors. Moreover P can have but one such set of factors.
P = FF'F"
where G,
so, it
(?',
= G-G'
G"
if
G",
also denote
prime functions.
Then G cannot
of P.
for,
would be prime
product P,
483,
whereas
it
is
a factor
Then G is not prime to F, for example. Suppo.se, therefore, that G and F have a common factor. But G and F are prime functions, and two prime functions can have no factor in x but themselves in common. Hence G differs from P by a numerical factor, as c, at most, and we have
G = cF.
211
= GG'G"
F'F"
= cG'G"---,
from which it follows by a mere repetition of our reasoning that G' differs from one of the functions F', F", by a numerical factor at most.
Continuing thus, we reach the conclusion that the set of functions G, differs from the set F, F', F", at most by numerical factors G\ G'\ or in the order in which they are arranged.
Corollary.
way
composite function can be expressed in only one as a product of powers of its different prime factors.
485
This follows at once from the identity P = F- F' each set of equal factors in the product F F' F" ing power of one of these factors.
F"
if
we
replace
by the correspond-
Irreducible factors.
By
function with rational coefficients, we usually of lowest degree with rational coefficients.
486
+ V2,
Thus, while the prime factors of (x 1) (x^ 2) are x the irreducible factors are x 1 and 3fl 2.
1,
Vi,
487
From
469,
it
reducible integral function with rational coefficients can be expressed in only one way as a product of powers of its different
irreducible factors.
488
We
A
than
shall
employ the
letters a,
"
b,
and so
on, to represent
integers, positive or negative (not 0), and shall factor of a any integer which exactly divides a.
mean by a
489
prime number
itself
is
and
1.
212
490
a
If
is
A COLLEGE ALGEBRA
two integers, a and said to be prime to h.
h,
have no
common
factor except
1,
491
he
Theorem,
ma +
6, if
nb
1.
prime to
divisor,
we apply
the usual
1
greatest
common
we
shall obtain
We
may deduce
we have
the theorem from this fact by the reasoning of 479. Thus, let a = 325, b = 116. Applying the method for finding G.C.D.,
116|325[2 232
ri=~93]ll6U.
93
i.e.
325
116
= = =
116
+
+
93, or 93
= = =
325
116
93
-2
1
116
93
(1)
r2=~23]93[4
92
rs
1-
93 23
23, or 23
(2)
93
1,
or
23
(3)
Hence, starting with (3), and substituting first the value of 23 given by (2), and then the value of 93 given by (1), we have
1
93
23
Therefore
Hence we have
= 5-93 -4. 116 = 5-325- 14-116. 5 325 + (- 14) 116 = 1. found two integers, m = 5 and m-325 + n- 116 = 1.
14,
such that
And
492
Example.
Corollaries.
Find integers
m+
125 n
1.
fimdamental theorem we may derive for integral numbers theorems analogous to those derived for integral functions in 481-485, and by the same reasoning.
this
From
In particular we
1.
may prove
by
to
b.
that
the product ac
is divisible
If
7/"
a.
is
prime
to b, aiid
hy b,
then c
2.
is divisible
is
prime
b and
c,
then a
is
prime
to be.
composite number can be expressed in one loay, and but one, as a jyroduct of powers of its different prime factors.
3.
RATIONAL FRACTIONS
VIII.
213
RATIONAL FRACTIONS
REDUCTION OF FRACTIONS
Let A and B denote any two algebraic expreswhich B is not 0. The quotient of A by B, expressed in the form A / B, is called a fraction ; and A is called the numerator, B the deno?ninator, and A and B together the terms
Fractions.
sions, of
493
of this fraction.
When
fraction.
both
A and B
are rational,
A/B
is
called a rational
494
When
fraction
fraction.
;
both
but
if
A and B are integral, .4 /B is called a simple 495 A or B is fractional, A/B is called a complex
496
simple fraction is called a proper or an improper fraction, according as the degree of its numerator is or is not less than that of its denominator.
_
Thus,
x-y
X2 -f
2/2
and
2x2_3
x^
are proper,
2x2+1
X2
and
x^
X2
3
1
improper. ^ ^
An
improper fraction whose terms are functions of a single sum of an integral expression
fraction,
497
and a proper
expression.
400.
1
-
This sum
is
called a mixed
X
X2
_ Thus,
2x2
X2
+1= +1
X2
x3-3 = X2 + 1
+ +
.3
.
These depend on
498
ator
The value of a fraction remains unchanged tvhen its numerand denominator are multiplied or divided hy the same
lue may change the signs of both numerator and denominabeing equivalent to multiplying both numerator and denominator Changing the sign of the numerator or of the denominator alone
itself,
by
1.
will
320, 3.
214
If the
A COLLEGE ALGEBRA
numerator or denominator be a polynomial, changing
all its
its
sign
is
terms.
c
a
Thus,
If the
b
b
c
c
a
c
ab b+c
a
b
+bc ca
we may change
fraction.
numerator, denominator, or both, are products of certain factors, the signs of an even number of these factors but chang;
number
{b
(e
of
them
d) g)
will
Thus,
(a
(e
b)
jc-d) _
{g
-/)
h)
-a){c-/) {h -
(b
(/-
-a)(d- c) - h) e) (g
499
Reduction of fractions.
factors
To
When
said to be in its
We
common
factors are, or
show that
look there are none, by the methods of Chapter VII. first for common monomial factors and other common factors
We
which are obvious by inspection or which can be fotmd by aid of the remainder theorem, and when these simpler methods fail we apply the general method of 465.
will illustrate
1.
Simplify {aec
aec
(c
^^
ade) / {bde
^^
ebc).
have
ae
be {d
d)
c)
'-
=
+
a(c
b{c
d) -= d)
a
b
Example
2.
{x^
x^
G)/{x-
+ 3x +
2).
+ I and x + 2. numerator and denominator have any common factor, it must be one of these. Testing by synthetic division, we find that the numerator is not divisible by x -|- 1, but is divisible by x + 2, the quotient being x2 - X + 3. X2 - X -f- 3 X3 -I- X2 -f- X + 6
By
inspection, the factors of the denominator are x
if
Hence
Hence
x2-|-3x-t-2
RATIONAL FRACTIONS
Example
3.
215
6).
Simplify (x^
+ -
10) / (x3
we have
10
(x3
6)
2 (X
2).
if
the numerator
2, 461.
and denominator have any common factor, it But the numerator does not vanish when x = 2.
ah'eady in
its
lowest terms.
Example
4.
^^
"
^^
^'^
^^^
Simplify
(a
"") +
^'^
(^
^)
b)(b
c) (c
a)
Here the only possible common factors are a b, b c, and c a. = 6 in the numerator, we have b-{b c) + b-{c b), or 0. Hence, 417, the numerator is divisible by a b. And we may show in the same way that it is divisible by 6 c and c a. Therefore the numerator is exactly divisible by the denominator. But the two are of the same degree, namely three, in a, 6, c. Their quotient must therefore be a mere number and since the a- terms in the two, c) and a- (6 c) respecwhen arranged as polynomials in a, are a- (6
Setting a
;
tively, this
is
equal to
1.
Simplify (2x3+
By
is
2x2
465, x+
we
1.
2 x2
1,
we
And
13 x2
2x*
5x3
- 6X + 7 _ x + 7 + 8x2-2x + 5~x2 + 3x +
EXERCISE XXIV
5'
_ 4 a;3y5 _ 2 xhj^
(x^
'
y^) (x
'
(x3
y^) (X*
+ y) - y*)
2'
'
x2 x2
- 4 X - 21 + 2x-63' 3 x2 - 18 6x + 27 62 2 x2 - 18 62
(x2-25)(x2-8x +
(x2
3x2
_8x-3
+ 7x +
3x2
5x^
5x2
15)
10)
'
+ 6ax + a^ + 2ax-3a2'
9) (x2
15x2-46x + 35 10x2-29x + 2r
216
A COLLEGE ALGEBRA
(x
(X3
, ,
+ +
2/3)
(Z3
2/3)
X^
11,
(1
x;/)"-^
2/)2
-I
x2
1<2.
2x3 2x3
13.
+ 7x2_7x- 12 + 3x2-14x- 15
4
::
,^
14.
+ 2 Xy 2 "^^ ~ '"y 12 ?ix + 6 ny Gmx 3my 2nx + ny x3_8x2 + 19x-12 2x3- I3x2+ I7x + 12 +
^/i
2^
15.
2x*
lb.
x*
(g
"
x3_2x2-x-6
+ 3x3 + 8x2 + 8x + 8 - 6)3 + (b- c)3 + (c - a)3 - 6) (6 - c) (c - a) (a
c2)2
4 b2x2
52a;2
X*
4 6x3
'
c4
To add
or subtract fractious,
we
first
Reduce
and
ah
to a lowest
common
denominator.
6c
ca
The L.C.M. of the given denominators is ahc. To reduce a /he to an equivalent fraction having the denominator
abc,
we must
multiply both
Similarly we of c/ah by c.
6,
Thus,
501
To reduce two or more fractions to eqtiivalent fractions having a lowest common denominator, find the lowest comtnon mnltiple of the given denomijiators. Then in each fraction rej^lace the denominator by this lowest
common
thtis
new factor
RATIONAL FRACTIONS
Addition and subtraction.
217
502
contained
abca-\-b
d
Hence
sum
of
two or more
fractiorf^,
If necessary, reduce them to a lowest common denommator. Connect the numerators of the resulting fractions by the signs
common denominator.
Filially, simplify the result thus obtained.
of one or
This rule applies when integral expressions take the place more of the fractions for such an expression may
;
be regarded as
It is
its
lowest
terms, unless a factor wliich would thus be cancelled occurs in one of the other denominators.
Care should
lowest
bvi
common denominator
is
frequent mistake
in the lowest
and
a,
and
to introduce both of
which them
common
denominator.
It is often better to
pairs.
Example
1.
Simplify
~b
a'-
6^
is
a-
h"^,
and we have
2h
a^ b"^
_ a-h
a"^
a-b
b-
a'^
+ -
6
62
a'^
26
62
a-6 + a + 6-26
a2
2a -26
a2
62
62
its
+b
terms,
Observe that the denominator of the sum, when reduced to may be but a factor of the lowest common denominator.
lowest
218
A COLLEGE ALGEBRA
1
J.3
Example
Since the
2.
Simplify x
1
3x4-1
X''
1,
1
is
common
^
denominator is denominator is x^ 1.
first
(x
1),
the lowest
We
therefore have
~
1
_x
3x
X2
Examples.
Here
it is
^ + ^ ^
1
2
^
^_^
^^^
pairs.
simpler to
1
Thus,
4
+
2
+ 2-(x-2) _ x2 4
X'-
x+1
4
x2-l
12^
_ x2-l-(x2-4)_
(x2-l)(x2-4)
2x2
x2_4
x2-l
x2-l
x*-5x2 +
'
+ 3x-2 + Example 4. Simplify ^,^^,_,^ 2x3 + x2 + 3x-2 -1 aid of the remainder theorem, 415, we find that x
By
factor of the
is
a common
two terms
is
denominator.
+ l)/(x3 + x^ + 2x). + 2x and 2x3 + x2 + 3x - 2 is x2 + x + 2 and x3 + x2 + 2x = (x^ + x + 2)x, 2x3 + x2 + 3x 2 = (2 x 1) (^2 .y ^ + 2). 1 Before reducing to a common denominator, we inquire whether 2 x AVe find that it is, and 3 x - 2. is also a factor of the numerator 2 x2 + fraction to (x + 2) / (x2 + x + 2). cancelling it, reduce the second 2x2 + 3x-2 x2-l - Hence ?r; + x4 + x2-2x 2x3-fx2 + 3x-2
7,
x+1
X (x2
x+2
+ X + 2) X2 + X + 2 2x _ x2 + 3x + l x2 _ ~ ~x+l+ + X3 + X2 + 2 X X3 + X2 + 2 X
603
The product of two or more fractions Multiplication. be found by applying the following theorem.
may
RATIONAL FRACTIONS
219
The product of two or more fractions is the fraction whose numerator is the product of the numerators of the given fractions, and
,
its
ac
bd bd
is
of each
member by
and
b
ac (see 253).
bd
bd = ac;
its
252,254
multiply
The
fraction
its
lowest
terms before the multiplications indicated in and denominator are actually performed.
Example
,. We
,
numerator
1.
Multiply
x3
(x^
1 1
have
x3
X
1
+ -
(x
1) 1)
l)/(x
x2
1). 1
= + x2 2)
4- 1
Example
2.
Multiply
(
\
\
(x
2) / (x2
+
x2
x2
by
(x
+
X
2) /x.
We
have
+ x-2
:
x-1
^"^
Involution.
To raise a fraction to any given power, raise both numerator and denominator to that power.
Thus,
fy=^".
/a\" (-) = \b) ~h'b
For
_,
OCT
ton
factors
6
a-a--6
to
to
factors
factors
a"
6"
Example.
ab'^c^
/efg^.
We have
Division.
To
505
220
A COLLEGE ALGEBRA
To divide one fraction by another, rmdtiply the dividend by the reciprocal of the divisor.
a
Thus,
c
a d
ad
Z^l^Z-^JT' be d b c b
of each
c c
member by c/d
a
b
,,;
is
a/6
(see 253).
Thus,
7^:;X:; = d d b
-and - = adc =
c
ad
be
a
b
bed
252,254,503
multiply
Divide
-X2/
x2
+
xy
2/2)
/(x2
X*
2/2)
by
(X*
x2y2
x"^
V*
C UitVC
2/^
+ x2y2 +
y4
+ yi) / {x* xy +
y"^
yi).
x*
y^
RATIONAL FRACTIONS
221
better to divide
In simplifying a complex fraction, A j B, it is sometimes ^ by at the outset by the rule of 505, sometimes better first to multiply both A and B by the L.C.M. Before taking either of of all the denominators in A and B.
these steps
it is
A and B
h
1
separately.
Examplel.
StapHfy
a
(|5
+
,
,\ l)
+ /(^^ + )
a
b
We
have
+ +
6 b b
a-\-h
\-
+
b
2a
a
a
a
2a
+ a +
b
a
b
_ 2a a b
+
2a
__a
a
+b +b b
Observe that when the terms of a complex fraction are simple fractions we may cancel any factors which are common to the numerators or to
the denominators of these fractions.
we may cancel 2 a.
Example ^
2.
Simplify (
\a-b
;
/ + by'\a +
J
4-
a-b^
We may proceed as in Ex. 1 but a simpler method is to begin by - b). We thus obtain multiplying both terms by (a + b) {a
a a
a
b
- b~
a
+
b
_ "
a{a
a{a
+ -
b) b)
b(a
b{a
b) b)
_ "
a"^
a'^
+
-
ab
ab
ba ba
+ b^ _ + V^
b
3.
Example
Simplify
b
_ ~^
e
a
,
cf
df+e
Complex
fractions.
222
A COLLEGE ALGEBRA
EXERCISE
XXV
2a-3&
o 2.
2a + Sb
4a2-962
1
X+1
+-
X2-1
+
X3
+
1
3.
_ ^ x2-3x + 2
1 ^
+-
x2_5a; +
+ -
x2-4x +
3)
(3
4-
(X
5.
1111
1) (X
x+3
x) (1
(2
X) (X
- X)
+ g)
b b
a
(a
"^
6)
(a
c)
(6
c) (6
"^
a)
(c
a) (c
"
6)
j/z (x
+
(X
1
zx (y
z)
a) x)
{z
xy{z
a)
(x
7/)
{y
-z){y -
x) (z
y)
3" -I
3-2x
+
8x*-33x 8x3-27
2x + 6 4x2 + 6x +
'
9'
9.
(x
{a
l)(x,
l)
+ 5)3-c3 a+ b c
x2-4
11.
x3-
3x2
-x + 6
.
12.
x*-4x2-x + 2
...,',
2x* + 3x3-2x2-2x + l
13.
(a4-l)^(a--). \ a/ \ a*/
(i \a3
1+ a^
l)(a* a/
a').
15
16
x2 + x-6 x2-6x + 6 x2 + 7x + 12 x2-4x-5 x2 + 3x-4 x2-8x + 15 1 rx-1 1/x-l (x-2)(x-3) \-|1 [i 2\x + l /J J L X X L x(x + 1)
RATIONAL FRACTIONS
17.
223
19
^g_
(^,_^,_,,^22/z).^-^
'
+^
x-y
a"^
o- \
a2
62/
20
1_ +
+
1 _^
'^
a;
]_ +z
_
a ^^
6
&
1_1
a*
6*
1
a
z
6a\a6/
+
^
_b
/I
\2
22.
23.
x-2 X
x+
x-2
INDETERMINATE FORMS
Limits. Suppose the variable x to be taking successively the values 1/2, 3/4, 7/8, 15/16, and so on without end; then evidently x is approaching the value 1, and in such a
508
difference 1
will ultimately
we can
how
may
, it
be.
We
approaches 1 as limit.
x denote a variable which is supposed to be running through some given but never-ending sequence of values, and if there be a number a such that the difference a x will ultimately become and remain numerically less than every positive number that we can assign, tve say that x approaches this number a as limit.
in general, if
And
To
x
indicate that x
or a
is
a,
we write
= a,
= lim x.
It will
meaning than in 242. Whether or not a variable of the kind here under consideration approaches a limit depends on the sequence of values through which
224
it is
A COLLEGE ALGEBRA
supposed to be running.
Thus,
if
the sequence be
1, 2, 1, 2,
the
variable will not approach a limit. full discussion of variables and limits will be found in 187-205, which the student is advised to read, at least in part, in this connection.
509
Theorems respecting
variables x and
ij
limits.
In
203
approach
limits,
lim (x
..
lim -
= lim X
t:
lim y
it
unless lim y
if
IT
rv
0.
From
these theorems
follows that
given rational function of x, and F(a') its value when x = a, then r(x) will approach F(a) as limit whenever x approaches a as limit, that is,
limF(a-)
F(a),
is
I(2x2-3x + x=a
l)
= 2a2-3a +
l.
510
Infinity. Evidently if x be made to run through the never ending sequence 1, 2, 3, 4, , it will ultimately become and remain greater than every number that we can assign.
and remain
is
numerically greater than every number that we can assign said to approach infinity.
infinity
we employ
the
symbol
a;
oo,
and,
is
approaching
infinity, write
== oo, or
00.
511
It is important to notice that oo, as thus defined, does not denote a definite number, and that the rules for reckoning with numbers do not apply to it. Illustrations of this will be found below.
Note.
KATIONAL FRACTIONS
The phrase "x
"X
is is
225
approaching infinity"
is
a variable which will ultimately become and remain greater than every number that we can assign." When, as is sometimes convenient, we write lim x = oo, of course the
word
limit
it
in 508.
is
Theorem.
a constant
512
and
its
denominator a variable.
as limit, the fraction will If the denominator approach approach oo; and if the denominator approach co, the fraction
will approach
as limit.
1
/x.
.01,
X approach by running through the sequence of values 1, .1, .001, , then 1/x will run through the sequence 1, 10, 100, 1000,
and
will therefore
if
approach
co.
X approach co by running through the sequence 1, 10, 100, 1000, , then 1/x will run through the sequence 1, .1, .01, .OOl,---,
And
and
will therefore
approach
as limit.
And
so in general.
Indeterminate forms.
A
=
rational
fraction
of
the
form
513
f(x)/<{>(x) has a definite value for any given value of x 0. But when <f>(x) 0, the fracexcept one for which <f>(x)
0/0
or a/0,
It is convenient, nevertheless, to ically meaningless, 103. assign a meaning to the fraction in both of these cases. The form 0/0. The fraction (x^ 1) takes the 1) / (x
514
1.
Now,
excep)t
when x
1,
we can
divide
x"^
by x
1,
and have
This
is
(x^-l)/(x-l) =
true
1.
however
little
may
differ
1,
from
1.
Hence
if,
we make
it
approach 1
1)
/(^
1)
= lim
(a-
1)
2.
Thus, while the rules of reckoning give us no meaning for 1) / (x 1) when x = 1, they enable us to prove that (x^
226
this fraction
A COLLEGE ALGEBRA
always approaches the definite limit 2 when x
approaches
1 as limit.
Now we have just shown, 509, that F(a) = l^^F(3c) whenever F(x) represents a rational function and F(a) has a meaning. It is therefore convenient when, as here, the rules of reckoning give us no meaning for F(ti), to assirjn to it the value ^^^^F(x)-, in other words, to make the formula
= ^^^^F{x) our definition of F{a). We therefore give {x- 1) / 1) the value 2 when x 1. 1) / {x 1) = x -^ 1 for all values of We then have
F{a)
(a;
{x"^
ic,
And
in the
<f)
(x (x
515
The form a/ 0.
0.
The
by
0,
we can divide
0.
as
we
please from
as limit, 512, that if x be made to approach then 1 /x will approach (X). therefore assign to 1/0, and in general to a/0, when " value " a 0, the co, writing
have shown,
We ^
= "form
And
f(x)/(x-a)<f>{x),
where /(a-) and <l>(x) are integral and /(.r) is not divisible by X a, we assign the value " a: when x = a; our meaning
*'
RATIONAL FRACTIONS
being that the fraction will always approach to approach a as limit.
co
227
when x
is
made
Conclusion as to the values of a fraction. From 514, 515 we draw the following conclusions regarding a simple fraction
516
its
lowest terms,
it
values of x which
infinite
numerator f (x) vanish, and become for values of x which make its denominator (x) vanish.
its
(f>
make
For
both
2.
all
it
4>{.^)
a,
it
and
71
times and
<l>(x)
contains
a when
m>
n,
become
both
and
oo
= 2, we have - 2)3 -2 ^ X+ 1 - 2) - 2)2 1 (X (x = U, =: = = = x-2 CO, x(x-2) U, x(x-2)2 CO, x(x-2)2 2 x+ 1 The form oo/oo. It is often important to know what limit
Thus when x
cc
-
(a;
'
517
the value of a fraction /(a;) / <^ (a?) approaches nitely increased, that is, when x == ao.
when x
is indefi-
512, that
1/x,
/x^,
0,
when x
oo.
Hence,
=
.
oo,
X2-X + 3 = 1-1/X + 3/X2 = 2x2 + x-4 2 + 1/X-4/X2 X + 2 1 + 2/x _ x2 + x + 5""x + l + 5/x x2 + x-7_x + l-7/x ~ 2x + 3 2 + 3/x
a;
1/2,
And
in general,
when
(aox^ +
a.x"'-'
b)
approaches the limit ao/bo, if, as in (1), the degrees of numerator and denominator are the same the limit 0, if, as in (2),
228
A COLLEGE ALGEBRA
;
the degree of the denominator is the greater the limit cc, if, as in (3), the degree of the numerator is the greater. And in each case the limit is called the " value which the
fraction takes
when x
=
oo
oo,"
that
is,
when
The forms
co
and
co /cc.
co.
But the expression can then be reduced particular value of x. to one which will take one of the forms 0/0, a/0, or co/oo already considered.
1.
Thus,
1,
(x2
1)
co
when x
1.
But, except
when X =
we have
I)
=
1
and therefore
lim
r(a;2
1) '
="==11
x-lJ
^=ix-
1) '
2.
Hence we assign
1 2.
when x
=
0.
1.
Again,
x(x
0,
+
^
cc
<x>
when x =
,
But,
except
when x
we have
X
lim r^
='='>Lx
"I
X (X
=
2)
X (x
and therefore
2)
lim
x(x
2)J
=^-0x{x
2)
lim_i_^l. ^-"x + 2 2
1/2 when x
Hence we assign
0.
519
General conclusion.
variable, as
(-''),
proceed as follows
Reduce the exjiression to its simplest form, and then find what limit its value apjjvoaches wlien x is made to approach a as limit. Call this limit the value which the function has when X = a.
520
This method is restricted to functions of a single variable, as Note. For the reason that the method yields definite results is this the F{x). value of "j" F (x) depends solely on the value of a and not on the values which X may take in approaching a and the like is not true of functions
:
of
variable.
RATIONAL FRACTIONS
229
fraction
Thus, suppose that x and y are unrelated variables, and consider the x/y when x = and y = 0. and ?/ The limit, if any, which x/y will approach when x 0,
depends on the sequences of values through which x and y may run. as limit if it runs through For example, a variable will approach either of the following sequences
:
(1);
(2)
X runs through
the sequence 2, 3,
and y through (1), 1 /4, and approach 0. Therefore, if x and y are unrelated and y lutely indeterminate when x =
will run through (1), and y through (2), then x/y and approach co. But if x runs through (2), 4, then x/y will run through the sequence 1/2, 1/3,
variables,
we regard x/y
so in general.
as abso-
0.
And
Infinity in relation
to
If
we
take
521
we must
state the
a
If
li
3.
-\-
= 0, ac = he, c = h
unless a
then a
c,
-\-
then
= h, a = h,
unless
unless
c c
= oo. = or
=
cc.
oo.
important to keep these exceptional cases in mind when applying the rules to the solution of equations.
It is
-x
1.
When
1,
the second
oo, it
1, is
but as the
first factor, 1
/(x^
i),
is
then
does
is 0.
The product
is 1
/2
in fact, 518.
Instead of saying, as
we have
522
been doing, that the equation a- + 2 = a- + 3 and other simple -x = b, have no equations which will reduce to the form
root,
we sometimes say that they have the root cc. For however small a may be, if not actually 0, ax = b has the root b/a. And if, keeping b constant and different
from
512.
its
as limit, b/a will approach oc, In other words, as arc = 6 approaches the form Ox =b, It is therefore quite in root b/a approaches the value oo.
0,
we make a approach
230
A COLLEGE ALGEBRA
when ax
is CO,
agreement with the practice explained in 515 to say that =^ b has the form Ox = b, it has the root oo.
Observe that if we regard x-|-2 = x + 3asa true equation whose root we are not driven to the absurd conclusion that 2 = 3. For since X = CO we have no right to infer that the result of subtracting x from both members is a true equation, 521, 3.
523
In like manner,
instead of saying of a system of inconsistent simple equations, 377, 2, 394, 2, that it has no solution, we sometimes say
it has an infinite solution ; for from such a system we can obtain by elimination a single equation of the form Oa; = b, and, by 522, this equation has the root oo.
that
Thus,
we may say
1 (2)
x=
(1),
jc
asm = \, of the
pair
- mx =
(3),
-x =
l (2).
The
(3), (2) is
l/(??i
1),
= m/{m
1),
both l/(m 1) and m/ (m 1) approach infinity. and when The same thing may be shown by the graphical method, 386, 387. For, when m=:l, the graph of (3) approaches parallelism with that of
m == 1,
of intersection of the two graphs recedes to an infinite (1), and the point distance in the plane.
EXERCISE XXVI
Assign the appropriate values to the following expressions.
3.
5.
x^-5x + 2. 6^ ^^^^ ^ ^ 2. ^:-f^+f when x = 1. x3-2x + l x2-6x + 8 x2 x2 1 , - 2ax + a2 when x=a. = 4. whenx = l. 2x + l x-' x2-(a + 6)x + ab (3x + l)(x + 2)^ ,whenx = -2. (x2-4)(x2 + 3x + 2) X3 _ x2 - X + 1 =
,
when X
1.
c3-3x2-|-3a;-l
RATIONAL FRACTIONS
3x^-x + 5 x^ + (2x^ + l)(x3-5)^ when 1^ _3^^ - 6) x2 + 1 X + 6X - 7 (X* + 1) (X -X 1 X 2 when x = 3. 8. X (X - 3) x2 _ 9
^
^
231
x
= oo.
2 x2
1 9.
2
1
X
x2
10.
-i
(x
when x
1.
1)
X
X
X
,
,
X2
x-1
X
-2 whenx = 2.
2
,, 11.
x-lx +
l 3x + -
X
r
1
when x
,
= oo.
X2
FRACTIONAL EQUATIONS
On
tion
solving a fractional equation. be transformed into one
its
Any
524
may
plying both
which
We call
341,
of fractions.
It
follows from
is
which
equation, and, if it has any roots besides these, that they and so may readily be must be roots of the equation i
Example
1.
3 r, Solve ,
x
1
x(x
+ 13 = ^ 1) 0.
...
(1)
D = x (x
(x
1).
We
obtain
(2),
3 (x
1)
6x
Solving
is
not a root of
= 0. x = 2. D = x (x - 1) = +
13)
(2) (3)
0, it is
a root of
(1),
Example
2.
Solve -
1 1)
X (x
xj^ ^ ^
1)
^^^
Clearing of fractions, 3 (x
6x
(x
5)
Solving
(2),
= =
0.
1.
(2)
(3)
232
A COLLEGE ALGEBRA
As 1 is a root of Z) = a;(x - 1) = 0, it is not a root of (1). In fact, when X = 1, the first member of (1) has the form 3 + 6/0-G/O; and by not 0. 518 we find that its value is 8,
Hence
(1)
has no root.
We
525
may sum up
To
ing
solve a fractional equation, clear of fractions hy multiplythe lowest common denominator of all both members
by D,
The
roots
D
526
vanishes
Note.
are the
of this equation
except
those, if
We may
the represent the result of collecting all the terms of = will be the then given equation in one member and adding them fractions. integral equation obtained by clearing of
Let
N/D =
1.
If
N/D is
;
in
its
the same
for a fraction in
Thus, in '
524, '
Ex.
1,
x
^
13
x-1
2.
x(x-l)
8(x-2) = = ,f
x(x-l)
in its
iV/D =
the
same
0,
namely,
2.
If
N/I)
is
not in
its loioest
terms,
N=
will
N/D =
to
does not have, namely, the roots for which the factor
iV +5 = 8(x-l) = -
common
N and D vanishes.
3
Thus, in
524,
Ex.
2,
x
;
x-1
x(x-l)
x{x-l)
Here
is
N/D
is
not in
its
not a root of
N/D =
1,
N/D
X
;
Evidently
1. is
2. is
exceptional.
xx-lx(x-l)
this
+ a =
-
U-
except
is
in
its
lowest terms
RATIONAL FRACTIONS
3.
233
It
^=
X
(x
just
D = 0.
X
Thus, consider the equation
2
1
=
1)
0.
Here
N/B = (x
1.
l)2/x(x
ij
when X =
N = {x
1) ==
1)
and, by
51G, this
l)'
x (x
In applying the rule of 525, care must be taken not to introduce extraneous factors in the expression selected as the
lowest
If
527
common
denominator.
any fraction in the equation is not in its lowest terms, begin by simplifying this fraction, unless the factors thus
cancelled occur in other denominators.
Before clearing of fractions it is sometimes best to combine certain of the fractions, or to reduce certain of them to mixed
expressions.
x2
Example
1.
Solve
6X
x2
11
x'--8x +
15
6x-2x2
Here the terms of the first fraction have the common factor x 5, and those of the second the common factor x. Cancelling these factors, we have 1 11 X X 1 X 11
x-3
Solving,
6-2x
or
x-3
10 x
= = =
2(x-3)
Clearing of fractions,
10
5x
22 x
8.
66.
Example
, 2.
-, Solve
x+lx+6 = x+2,x+5
x+2
1
x+7
x+3
-
x+6
X+2
minus
signs,
x+2
1111 1111
to a
x+7
x+3
x+6
may
be connected by
x+3
x+G
x+7
234
A COLLEGE ALGEBRA
of each
1
member
separately,
1
_ ~
x2
x2
Clearing of fractions,
Solving,
+ 5x + 6 x^ + 13 x +
42
The given equation may be solved by clearing stands, but that method is much more laborious. 528
of fractions as
it
Simultaneous fractional equations. The general method of is to clear the several equations of frac-
and then
except
371.
The
any, which make denominators in the are the solutions of these equations, given equations vanish
if the equations are of the form described in 379, or they can be reduced to this form, they should be solved by the method explained in that section.
But
if
Example
y
1.
2/
xy
2x
x
4y
2y'^
xy
X
Solving
-y + 1 = 0,
x
2,
+ 2y-S = = S.
0.
Therefore, since none of the denominators in the given equations vanish when x = 2 and y = 3, these equations have the solution x = 2,
= 3.
Example
2.
x, y, z.
+
xy
5 6
yz
,
=
z
3
1
2 (z
x)
xz
0.
379,
1_5
y
(j
1__2
z
1__1
x
2
Solving for 1 /x, 1/y, 1/z, we find Hence the required solution is x =
l/x=
2,
1/2, 1/2/
z
=
1.
1/3, l/z
= -l.
= 3,
RATIONAL FRACTIONS
EXERCISE XXVn
Solve the following equations for x.
^
235
'
6x 3x +
4a;
2x
7_ ~
5
1
2.
6 4
3.
4.
,31
2x +
3
'
x-2
x-4
x2-6x +
8
x-5
+
2
2x2-7x-15
2
= 0.
3
5.
(x
+ l)(x-3)
x2
(x-3)(x +
2)
(x
2)(x
^
l)
6.
-1
x2-l
x
+ 4x-5
2x
5
+
x2
'
+ 6x +
5
.0
5
,j
3x+l
x
b{x
^^
5-6x
x + b , r o (X + a)
:
+ 9x-18x2
a
b
+a + + b)
= +rab
^x^ + 9'
l 1
x^-l
x
20.
10.
x2
X
11.
x-o
^^"^
x
x-4~x + 8
+ +
Q
a:
s'
12
+ +
1
6
2
,16.
x3
x-2
X
14
x-2
1
3x2
1
X2
+x^
X*
15.
1_^ x-2
-
16.
+d= q
+
,
6.
236
x2
17.
A COLLEGE ALGEBRA
+ 7x-8
x'^
1
+
X
+
::
S
h
Xax
2
b^
2x2-a;+7_ X + 3
z
4 X.
18.
X
19.
+ 2bx-2ab + a
,
26 +
^
-x"^
3c2
X
-
0.
2c
(X
(x
(^- 6) (x -
"^^
+
c)
(x-6)2
(x
+
a) (X
c)2
c) (X
a) (x
6)
20.
3x+2
X2
x-5
X2
x-2
X2
+
2
X
0.
a
21.
x-2
+6 = x2-4
X
?/,
and
r3x +
22.
2/
-1 a;-2/ + 2 X + 9 _ X + ~ +4 +
y
2/
6
7
-2 x-3
y
2
x2
23.
3 3
^
= 2 X + xy 2y + xy 0,
9
^
xy x
24.
+y
yz
a.
x
25.
<
+ 2y y + z
+22/ + 2z
5
z
=
7
3,
+
zx
-\-
&,
- 3x
2
_Z
3x
= -1.
2?/
PARTIAL FRACTIONS
529
It follows
from
of
single variable, as x, can be reduced to the form of an integral function, or a proper fraction, or the sum of an integral function and a proper fraction.
it
is
when a proper
fi'action
A/B
is
the
sum.
The
and so
tions of X.
RATIONAL FRACTIONS
Theorem
tions
1.
237
530
A/B
and
C/D
For
Since
A C = ADBC
B
A
is
BD
AD
is
of lower degree
than BD.
And
since
is
BC
is
of lower
Hence
AD BC
2.
is
BD.
Theorem
Let
and
V denote
integral functions,
and Kf^
531
Ifl +
But I
A'/B'
A/B =
I'
A'/B', then 1
I
I' =
For, by hypothesis,
A'/B'.
I' denotes an integral function (or A / B denotes a proper fraction (or 0).
= - A/B = A/B =
and,
530,
Therefore, since an integral function cannot be identically equal to a proper fraction, we have
/
or
/'
and A'/B'
0,
/=/'and
Theorem
3.
A'/B'.
denote a jjroper fraction whose denominator has been separated into two factors, P and Q, which are jxrime to one another. This fraction can be reduced to a sum of two proper fractions
Let
A/PQ
532
of the forms,
B/P
is
and C/Q.
prime to P, we can
find, 479,
For, since Q
functions
1
two integral
Hence
If
PQ
A
and
PQ
(1) ^ ^
is
AM/P
AN/Q
already demonstrated.
238
If
A COLLEGE ALGEBRA
fractions, reduce
them
to
sums
let the
results be
AN ^/ + pand
B
C ^A- + -.
(2)
have
PQ
But
since
^B P
A
C
Q
^
it
A'
and
_B
C
q'
PQ^ P
as
was
Note.
to be demonstrated.
533
The
fraction
A/PQ
A =B PQ P
sum
(1) ^ '
B/P +
C/Q.
For suppose
where B'/P and
= P Q
C
B'
C
,
Then
But
wise
and therefore ^^
^'^
=C'-C.
(7 = 0.
(2) ^ '
(2) is
impossible unless
BB')
B'
=Q
P
and
For other-
(2)
is
is
prime to
534
Partial fractions.
C/Q, whose
just proved, jxirtia I fractions of A / PQ. To resolve a given fraction of the form A / PQ into its parand C/Q, it is not necessary to carry out tial fractions
existence
we have
B/P
532
we may apply
1)
the method of
undetermined
Example
fractions.
1.
example.
two partial
Resolve
(2
x"^
\)
/ (x^
into a
sum
of
This
factors,
is
and
x^
1,
RATIONAL FRACTIONS
Hence, 532, (2x2 + l)/(x' tions whose denominators are x
239
sum of two proper fracand x^ + x + 1 respectively. The numerator of the first of these fractions nmst be a constant, that of the second an expression whose degree is one at most. Hence we must have
1) is
equal to a
2 X-
+ 1 Z = a + 6x + c Z x3-l x-lx2 + x +
n) '
^
where
a, b,
and
denote constants.
clear (1) of fractions.
To
find a,
b, c,
We
or
obtain
2 x2
2x2
+ +
+
1
1
(2)
As
(2) is
an
284.
(3)
Hence
or, solving (3),
?>
a
2x2
= 2, I,
a
6
1
6 + c = 0, = 1, c = 0.
X
h
l,
_ Therefore
.
X3-1
2.
X-1
Example
Resolve (5x
+X+1 + 4)/(x* + x^ +
X2
x2
x) into
sum
of
two
partial fractions.
From
the theorem
535
532 we Let
may draw
A /PQR
R are prime to one another. This fraction can be reduced to a sum of the form
4
PQR
where
B /P, D / Q, and
PQR E
/R
^B
prime to QR, 482, A /PQR is the sum of two proper fractions of the form B / P -{- C/QR, 532; and since Q is prime to R, C/QR is itself the sum of two proper
For since
is
fractions of the
form
D/Q + E/R,
532.
The like is true when the denominator is the product of any number of factors all prime to one another. 2. Consider the proper fraction A / PQ^ in which P is prime to Q. By 532 it can be resolved into the sum
ABC
P
PQ^
Q"
240
A COLLEGE ALGEBRA
cannot apply the theorem of
since
532 to the fraction C/Q^, Q are not prime to one another. it
We
But
is
422, to a polynomial in
can be reduced,
C
where
CiQ2
c^Q
Cs,
Ci, C^, and C^ are of lower degree than Q. Dividing each member of this identity by Q^, we have
sum
A
where
=^+
C\
-T;^
+ TT +
C<,
C,
is
degree than
And
so in general
when
once in the denominator. We therefore have the following theorem. Suppose that the denominator of a given proper fraction has
some occurring once, some, it may been separated into factors \ohich are all j^rime to one another. he, more than once The fraction itself can then he resolved into one and hut one
a single fraction of the form B/P, and ivhich occurs r times, there is a group (2) for each factor, Q, + C,/Q^ where of r fractions of the form Ci/Q + Cj/Q^ H are all of lower degree than Q. , C, Ci, Cj,
536
.
It can
is
the product of factors of one or both the types x a and x^ + px + q, where a, p, and q are real, but where the factors of x^ + px + q have imagicoefficients
nary
coefficients.
it
Moreover
follows from 469, 532 that, if the numerator and the factors into which its
RATIONAL FRACTIONS
241
denominator has been separated have real coefficients, so will the numerators of the corresponding partial fractions. Hence,
by
534,
'proper fraction
coefficients is
Every
have real
whose
to
a definite sunt of partial fractions related as follows to the factors x a and x^ q o/* px
equal
its
denomiiiator.
1.
fraction of the
2.
form A/(x
is
a group
of
A^/(^
a)
A,/(x
_ a)2 +
+ AJ(x -
a)^
where Aj, Aj, A^ are real constants. 3. For every factor x'^ -|- px + q occiirring once there is a single fraction of the form (Dx + E)/(x^ + px + q), where D
and
4.
px + q form
is
q)
+ (D,x +
E,)/(x2
px
q)^
The fractions here described are usually called the simplest partial fractions of the given fraction. They are best found by the method of undetermined coefficients.
Example
fractions.
-r,
<.
537
1.
Resolve
x2
(a;
into
its
simplest partial
By
536,
ro^
we have
1) (X
+ x-3 - 2) (x obtain
3)
= A + B x - 2 x - 1
C -
,_
3
(1) ^ '
of fractions,
we
(2)
A, B, C by arranging the second member of (2) according to powers of x and equating coefficients of like powers but, since A, B, C are constants, the same results will be obtained by the following method, which is simpler.
find
;
We may
242
In
(2)
A COLLEGE ALGEBRA
letx
let
let
next
finally
x
x
= 1, = 2, = 3,
x'^
and we have
-1=A{- l)(-2),
3
= jB( = C 2
1)
1,
.-.
1,
.-.
C=
9/2.
9
Hence
+ x-S
3)
=
X
(X- l)(x-2)(xExample
2.
2(x-l)
x-2
2(x-3)
Resolve
x(x
+ 1 1)3
By
636,
we have
x
x(x-l)3
and therefore
In next
(1) let
= - + x-1 + X
I)-'
(X
+
1)-
(x-l)3
1)
= J. (x -
let
= =
0,
1,
Bx.
(1)
.-.
and
in (1),
transpose to the
first
mem-
(x
1)^
x3-3x2 + 2x
Dividing both members of
(2)
= x(xby x
(x
1)'^
+ Cx(xwe have
C.
1).
(2)
1),
x-2 = B(xEquating
1
1)+
7J
(3)
coefiicients of like
powers of x
C,
.-.
in (3),
we have
jB
and
=-5+
Hence
x+1 =
x(x-l)3
(1)
11
1
x-1
1,2
\-
and C
=-
1.
(x-l)2
x,
(x-l)3
obtaining
Or we may arrange
X
according to powers of
C)x2
+ {3A + B - C + D)x - A.
we have
l.
Equating
powers of
x,
we
find, as before,
Resolve
(a;2
5 x2
4X
4-
16
-a;
into
its
l)2(x-3)
simplest partial
The
factors of x^
being imaginary,
we
-:; a;2
have, 536,
Ax + B
(x2
_X+
Cx
1)2
-^ - X + 1 + X - 3*
+D
E are constants.
RATIONAL FRACTIOXS
Clearing of fractions,
243
1) (
3)
^(X2
-X+
1)2.
(1)
may
find
A, B, C, D,
by arranging
(1)
simpler.
In
(1)
let
3,
and we have 49
= 49 E,
.-.
E=
1.
term
(x2
-f 1)2
first
member,
by X
simplify,
3.
"We obtain
(x3
x2
5)
= ^x + 5 +
of (2)
by
x2
(Cx x
+ +
D)
1.
(x2
1).
(2)
"We obtain
(3) ^ '
By
and
.
2x +
X2
= Ax + B + 1 X2 - X
+
therefore
^ ^ Cx + D.
(3)
531,
are equal.
Hence
Cx + D, and
C=
3
1)2
1,
D = 2,
x
2x~S=Ax+
B, and therefore
A = 2,
x2
^, Therefore
2x +
(X2
-X+
= 3. +2 -X+1
j5
1-
When
(x
a, 423. Similarly when the denominator has powers of a; the form (a;^ +joa; + qY, the factors of x"^ -\- px + q being imaginary, we express the numerator in powers of x" + px + qExample. ^
fractions.
ay,
the denominator of the given fraction has the form it is best to begin by expressing the numerator in
538
^ Eesolve
,
x*
its
.,
^.
simplest partial
By
a;4
423,
we
find
(X
+ a;3 - 8 x2 + 6 X - 32 =
I8(x
- 2) - 28.
28
(x
+ x3-8x2 + 6x-32
(x-2)5
x-2
-,
(x-2)2
^ + (x-2)3 + (x-2)* ^
:
18
- 2)^
539
If given an improper fraction, we may first reduce it to the simi of an integral expression and a proper fraction and then resolve the latter into its partial fractions.
244
A COLLEGE ALGEBRA
Apply
this
Example.
method
to the fraction
x3-2x2-6x-21 X- 4 X 5
have We
,
x3-2x2-6x-21 = x + x2_4x-5
,
+ +
7X-11
x2-4x-5
7x (X
=
and proceeding as above we
find
11
+
4
1) (X
5)
(x
7x-ll + 1) (X -
3
5)
+
1
EXERCISE
XXVm
Resolve the following into the simplest partial fractions whose denominators have real coefficients.
2
1.
.r,
4-1
6X
SYMMETRIC FUNCTIONS
IX.
245
SYMMETRIC FUNCTIONS
In the expression x^
x"^ -\z"^ y"^ -\-
y^
z^
the
let-
540
manner that
is
them be interchanged
or
x"^ -\-
namely, y"
-f z-, or
z'^ -{y'^
+ x^,
z"^
tf.
To
indicate this,
we say
that x^
y'^
+
is
is
said
to be symmetric with respect to these letters when every interchange of two of the letters will transform the function into
an identically equal
functioi-.
+ y) + z) {y + z) with respect to x, y, z, + a) (x + 6) (x + c) with respect to a, 6, c. On the other hand, x + y z is not symmetric for if we interchange and z we obtain x + z y, which is not equal to x + y ~ z.
{xy
+
b
XZ
yz)/{x
(x
(a;
and
We
call 2 x'^y
541
to the variables x, y,
namely,
x'^y
and
because the variable parts of these terms, can be transformed into one another by
x, y, z.
And
a7i
so in general.
The
sufficient
integral finic-
542
symmetric with respect to terms of the same type shall have the
be
same
coefficients.
This implies that if a symmetric function contains one term of a certain type, it must contain all terms of that type that is, all terms that can be derived from the term in question by
;
of the letters.
be symmetric, we must have a = b = c. Again, if a symmetric function of x. y, z contains the term x'^y, it must contain all the terms x'^y -\- y-x -{ x^z + z^x -f yH + z'^y.
ax-
by-
cz^ is to
246
543
A COLLEGE ALGEBRA
This theorem will indicate the general form of a symmetric function of given degree with respect to a given set of letters.
Thus, the general form of a symmetric function of the first degree with respect to x, j/, z, u is a (x + y + z + it) + 6, where a and 6 denote
constants.
first, 1. 3.
and homogeneous functions of the Again, the most general symmetric second, and third degrees with respect to x, y, z are
a (X
a
(x^
2.
a (x2
y^-x
^2
22)
2-3)
5 (x2y
x^z
z^x
544
the
On expressing a symmetric function. The notation 2^ means sum of all terms of the same type as x^ that is, if x, y, z are
;
^x'y
selectgiven symmetric function may be represented by and writing the symbol 2 ing from its terms one of each type, before their sum.
Any
Thus,
S (2 X 545
x32/2)
= 2x + 2y + 2z- xV -
V^^^
- ^^^"^ -
^^^'^
V^^^
2^2/^-
writing out symmetric functions at length, it is best The to arrange the terms in accordance with some fixed rule. following examples will indicate a convenient rule for the
When
arrangement of
iyitegral
symmetric functions.
and by the
Suppose that the letters under consideration are a, 6, c, d, normal order of these letters understand the order a, 6, c, d.
We shall then write Sa6 and 2a6c as follows Sa6 ab + ac + ad + be + bd + cd, Xabc = abc
:
+abd + acd +
bed.
Observe that in each term we write the letters in their normal order. In forming 2a6 we take each letter a, b, c in turn and after it write each subsequent letter. The terms of ^.abe are derived in a similar manner
We shall
I,a"'b",
when
-^ n,
as follows
Sa263
a263
b'^a^
a^c^
c^-a^
cW + d^c^
;
Observe that we keep the order of the exponents fixed then under the exponents we write the letters of the first term of 2a6 in both the orders ah and ba, and so on.
SYMMETRIC FUNCTIONS
In like manner
247
we may
write
It follows
from the
546
symmetry, 540, that a symmetric function will remain symmetric when its form is changed by the rules of In particular, reckoning.
definition of
The sum, difference, product, and quotient of functions are themselves symmetric.
tivo
symmetric
By
we may
ing given symmetric functions by algebraic operations without It is only necessary to actually carrying out these operations.
result.
Find (Sa)2
z=
(a
)^
Evidently the required result is a homogeneous symmetric function of the second degree consisting of terms of the two types a^ and 2 ab.
Hence
(Sa)2
2.
Sa^
+
7/2
2 Sa6.
Example
Hence
Find 2x2 ^x
.
(x2
z^)(^x
z).
is
sum
of terms of the
two types
x^
and
x^y.
2x2 2x
Example
3.
Find
y^z
z'^y.
x, y, z.
We
543,
+ 2)3 =
6 (x2?/
+ yH + x22 +
6,
c,
To
values to
equations.
which
x
x
Thus, putting
= 0, we have 1 = a. = 0, we have 8 = 2 o + = 3a + 1, y -I, 2 = 1, we have 27 Finally, putting x Solving (1), (2), (3), we obtain a = 1, 6 = 3, c = 6. Hence (2x)3 = 2x3 + 3 "Lx^y + 6 'Zxyz.
Again, putting
= = =
1,
2/
1,
2/
= 0, = 1,
2 2
(1)
6.
(2)
-f c.
66
(3)
248
547
A COLLEGE ALGEBRA
ters
In the expression x^y Cyclo-symmetry. y^z z are involved in such a manner that if 0-, y,
2/,
+ ^ic
we
the
let-
replace x
by
x^y
y by
z,
and
y'^z
by
a;,
sion,
namely,
y, z,
z'-x
+ y'^z + s^jj
aj,
is cyclo-synimetric,
letters
x, y,
is
And,
in general,
an expression
said to be cyclo-symmetric,
or cyclic, with respect to certain letters arranyed in a given order, if it is transformed into an identically equal expression when we replace the first of these letters by the second, the
second by the third, and so on, and the last by the first. Such an interchange of the letters is called a cyclic interchange.
548
Observe that the terms of x'^y -\- y'h + z-x are themselves arranged eyelidy ; that is, so that the first changes into the second, the second into the third, and the third into the first,
when we
Cyclic expresreplace x hy y, y by z, and z by x. sions are of frequent occurrence and reckoning with them is
greatly facilitated by arranging them cyclicly. Evidently every symmetric function is cyclic, but not every
549
changes
550
As the example shows, a cyclic function will ordinarily not contain all the terms of a given type, but such of these terms as it does contain will have the same coefficients.
Theorem.
551
The sum,
difference, product,
and
quotient of two
{x'^y
y'^z
z-x) (x
z).
Moreover it concyclic but not symmetric. x'^yz, each once, and terms of these types only.
is
y'^z'^
y^z
+ z^ + xhj^ +
z^x^
+ x'^yz + yHx +
z^xy.
SYMMETRIC FUNCTIONS
EXERCISE XXIX
1.
249
2/4
2^
4 (X3
2/3) (?/3
23) (X^
z'^)
is
symmetric.
2.
Write out in
I,a^b^,
full
b, c.
Sa36*,
b)c,
'L{a
3.
S(a +
S(a
+ 26 +
3c).
Show
that (a
b, c
;
6) (6
c) (c
a) is
respect to a,
4.
also that (a
6)2(6
a)'^ is
symmetric.
Is
c,
(a
6)2 (6
c)2 (c
d)2 (d
a)^
a,
6,
6.
x2,
22
y2^ x^
z^
cC^bc,
obiP,
acH,
b'^cd
c).
tZ
c) (6
a),
(a
c) (c
6),
(a
6) (6
6.
Write out
in full
6,
c,
whose
first
terms are
a63c2,
7.
a{b-
c),
(b
2c){a
d),
a^/(a
6) (a
c).
identities.
2a*
2a36
2a6 2a
2a-'6
3 2a6c.
By
552
explained it is often possible to factor a complicated symmetric or cyclic function with comparatively little reckoning.
Example
1.
Factor x^ (y
z)
-\-
y^ (z
z,
x)
+
-
z^ (x
y).
when y =
for
23 (x
z)
is
2)
Z3
(z-x) +
2
= 0.
;
exactly divisible
exactly divisible
by
product (y x) (x y). 2) (2 Both dividend and divisor are cyclic and homogeneous, and their Hence the quotient must be degrees are four and three respectively.
250
A COLLEGE ALGEBRA
a cyclic and homogeneous function of the first degree and therefore have Hence -\- y + z), where k denotes some constant.
-z) + y^(z-x) + z^{x-y)~k{y - z) {z - x)yx ~y,{z + y ^ z) (1) To find k, assign to x, y, z any set of values for which the coefficient
Thus, putting x
2,
1,
0,
we have
&
Qk, ov k
\.
Or we may find k by equating the coefficients of like powers of x in the two members of (1) arranged as polynomials in x. Thus, the x^ term in the first member is x^ (y z), and in the second member it is kx^{y z), whence as before k = \. We therefore have
x^iy
{y
z) {z
x){x
y){x
z).
Example
Factor
(x
z)^
when x
x^ ys = y, for
z^.
y + zf + xj^ -y^ -z^ = 0. Hence the function is exactly divisible by x + ?/ and for a like reason it is divisible by ?/ + z and by z + x, and therefore by (x + y) (y -^ z) (z + x). As the dividend and divisor are symmetric and homogeneous and of the fifth and third degrees respectively, the quotient must be of the form
{-y +
ft(x2
2/2
z'^)
+
{y
i(X2/
+
{x
2/z
+
y
zx), 543.
-\-
Hence
= ( +
Putting
Putting
2/)
+
1,
z) (z
2/
+ +
1,
z)^
x^
y^
z^
X) \k (x2
= x =
2,
2/
I,
= =
zx)].
A;
+ y + )5 - 5 - - z5 = 5 (x + + z) (z + x) (x2 +
?/5
2/) (2/
2/2
22 ^. a;y
2/z
zx).
Example
(x
3.
2/
Factor
z)3
(2/
- (2/ 0. (2/ z)3 {y + zf Hence the function is exactly divisible by x or x and for a like reason it is divisible by y and by z and therefore by xyz. Since both dividend and divisor are of the third degree, the quotient is some constant, k. Hence
;
- (z + x = 0, for + zf - (z - 2/)3 z
x)3
2/)^
(x
2/
- zf.
(X
+ +
2/
Putting X
(X
2/
y)3
24,
2/)8
(X
+ +
2/
z)3
= kxyz.
=
(2/
z)3
24x2/2.
SYMMETRIC FUNCTIONS
The method just explained is sometimes useful
cyclic fractional expressions.
a^
h^
c3
251
in simplifying
553
Example.
Simplify
(a
6) (a
c)
(6
c) (6
c.
a)
(c
-a){c- h)
This expression is cyclic with respect to a, 6, The lowest common denominator is 'h c) (c
a) {a
h).
On
common denominator we
obtain as the
Hence, by cyclo-symmetry, c). b). h^{c a) and c'^{a Adding these three numerators and factoring the result, 552, Ex. 1,
(a
we have
&
c) (6
c) (c
a) (a
b).
to a
c.
EXERCISE
Factor the following expressions.
1.
XXX
2. 3.
y- (z
z'^
4.
6.
x{y
6.
7.
8.
xY + z{x y)^. 2)3 + y{z zY + 2/^ (z x)3 + z^ (x yY. X* (2/2 - 22) + yi (22 _ a;2) + 2" {x2 - 2/2). + z)-^ x^ 2/3 2^. (x +
x^ (y
2/
(2/
2)6
2/
(2
x)5
(2/
+
2
(x
yf.
9.
(x
(2/
+
-
+
(2/
2)5
+
(z
x)6
(z
2/)^
(x
2/
2)^-
10.
11. 12.
X (2/
x5
(2/
2)3
X) (Z
X)3
2/)^.
2/
-b)(a
5^
+
c)
'^ (6
c) (6
+
a)
(c
'*
a) (c
6)
14.
(a
x+a -
6) (a
+
c)
x+6
(6
+
a)
(c
x+c
a) (c
c) (6
6)
252
, _
A COLLEGE ALGEBRA
a'^
:
16.
be
-
(a 16.
(a
17.
-h)(a(& + c)2 - 6) (a -
c)
V^ ca +- - (b c){b
(c
^
+
a)
[c
c2
ah
a){c
~ -
b)
a)^
(a
^
6)2
6) c2
(c
c)
(6
c) (6
a)
62
(c
a) (c
(a
a2
b){a
c) {X
:
+
a)
{b
c) {b
-; -
a){x
b)
7-
a){c~
b) (x
c)
X.
554
(a
+b^c+
d) (e
+f -^ g) (h ^
+b+c+
d by each term of e+f+g, then multiply every product thus obtained by each term of A + ^, and finally add the results of these last
of a
multiplications. Hence we shall obtain one term of the product if we select one term from each of the three given factors and multiply And we shall obtain all the terms of these terms together.
the product if we make this selection of terms from the three given factors in all possible ways. Thus, selecting b from the first factor, g f-om the second, and k from the third, we have the term bgk of the product
;
and so
on.
term from a-\-b-{-c-\-d in four ways, g in three ways, and a term from h -\- k in two ways, the number of terms in the complete product is 4 3 2, or 24. And so, in general.
Since
select a
-\-
we can
a term from
e -\-f
The product of any number of i^olynomials is the sum of all the prodiicts that can be obtained by selecting one term from each factor and multiplying these terms together. and
terms, the second n, the third p, if the first factor has number of terms in the comjylete product before is mnp , like terms, if any, have been collected
so on, the
And
253
555
multiplication.
This theorem supplies a useful check on the correctness of a It may be applied to a product in which like
terms have been collected, provided its terms represent sums of terms of like sign and without numerical coefficients, the
coefficient of a
terms
it
represents.
c) {a
c)
should have
3 or 9 terms,
which are
b
all
-\-
of the
b'^
same
sign.
{a
c) (a
c)
a'^
+
1
c'^
+
1
2 ab
+ +2+
2 or 9 terms, as
it
Similarly, the product (a + b) (a + b) (a + b) should have 2 2 2 or 8 terms. But this product when simplified is a^ + 3 a-b + Sab- + 6", which means, as it should, an uncollected product of 1 + 3 + 3 + 1 or 8 terms.
this
556
Thus, the student has proved, p. 249, Ex. 7, 2a6 2a = 2a26 + 3 l,abc. To test this formula, suppose that only the letters a, b, c are involved. Then 2a6 has 3 terms, 2a has 3 terms, 2a2& has 6 terms, and 2a6c has term and 3-3 = 6 + 3-l, asit should.
;
first degree.
The theorem
557
to obtain the
-}-
hy inspection.
Thus,
(x
h,) (x
x"^
Selecting in all possible ways x's from two of the factors and a b from the third, we have the terms b^x'^, b^x"^, b^x^.
Selecting in all possible ways x from one of the factors and from the other two, we have the terms ^1^2^? b-fi^x, b^b^x. Selecting b's from all three factors, we have the term b^b^b^.
&'s
the coefficient of x
is
the
254
A COLLEGE ALGEBRA
is
the product
(x -f bi) (x
62, b^.
^2) (^
^3) is itself
Observe also that since (x + ^1) (x + Sj) (^ + ^3) is symmetric with respect to bi, bo, b^, we may obtain the product by finding one term of each type, as x^, b^x-, b^b^x, bxbj)^, and
then writing out
all
558
{x
By
the terms of these several types. we may prove the general formula
+ *i) {x + ^2) {x + b^)---(x + b,) = a;+ ix"-^ + 5oX"-2 H B, = + ^2 + where 5i = S^i b^, + + bj}^ + + b_ib^, Bi = -Zb^bi = b^b^ + = '$bJ)J>3 = bibj)^ + bib^bi ^ h ^_2^-i^> B3
\-
^-^
^3 -|
\-
b,b.,
that
is,
Bi
is
^1) ^2>
^n) a^i^d
the sum, and B is the product of all the letters the intermediate coefficients are B.2, the sum
:
of the products of every tivo of these letters; 3, the the products of every three ; and so on.
sum
of
we
select i's
from
tJiree
of the factors
and
a;'s
from the
rest.
Making the
bib2f)sx"''^,
selection in all possible ways, we obtain the terms and their sum is B.^x"~^. bibibiX"'^, -,
B
bi, b^, , b^.
559
(x
In like manner,
we have
5i) {x
\-
{-
lYE,,,
where
Bi, B^,
-B
558. and
255
and +, the the signs connecting the terms are alteiuately when n is even and last sign, that of ( 1)"5, being when 11 is odd.
We
all
the letters
obtain this formula by merely changing the signs of For this h-^, b^, b in the formula of 558.
leaves unchanged every B whose terms are products of an even number of b's, and merely changes the sign of every B whose terms are products of an odd number of b's.
Example.
1.
By
the
method
3.
(x+2)(x-3)(x +
{x
4).
4.
y){x
+ 2y){x
Sy){x
4y).
terms in the sums 2bi, Sbibo, . Let i, Wr,, denote the number of terms in 2/>i, "^b^bz, respectively. 1. Since 2&i + Kf evidently ni = n. ^i /*2
The number
of
560
of the other n
^i, b^j , ^ ^J each obtain n (n 1) products all told. But these n (n 1) products are the terms of "^bji^, each counted twice. Hence n^, the number of terms in S^i^g*
2.
If
we multiply each
of the n letters
1 letters in turn,
we
is
n(n
l)/2,
or
n
Thus, we have
&1&2,
n(n
1)
hh,
'
hb,,
b^bi, bzbs,
b2bn
There are n groups of these products, and n group, hence n{n l) products all told.
;', Mi,
&n&2>
Mn-l-
products in each
in the
But the product bibo here occurs twice, namely once and once in the form 62&1 and so on.
;
form
bib^
3. Again, if we multiply each of the ??2 terms of ^bjt^ by each of the n 2 letters which do not occur in that term,
we
obtain ng (n But these n^ {n 2) 2) products all told. products are the terms of '^b^bj)^, each counted three times. 2)/ 3, or Hence n^, the number of terms in '^b^b^bg, is 112 (n
^^-2
_ n( 7i-
l)(n-2)
256
Thus, we have
A COLLEGE ALGEBRA
hhh, iihh,
hb2b',
hhh, hhh,
, hh^n
bn-lbnbu bn-lbnb2,
, 6,-lMn-2.
There are no groups of these products, and n 2 products in each 2) products all told. group, hence n2 {n But the product 6162&3 here occurs three times, namely in three forms, And similarly every term of 2616263 here 6162 &3) ^1^3 ^2) ^2^3 ^1occurs three times, once for each of the three ways in which a product of three letters may be obtained by multiplying the product of two of
letter.
By
we can show
1)
(71
that
n,^^=
i
3 _n(n
2) (n
3)
'
1.2.3-4
to r factors
.
n(n =
l)(n 2)
^^
2 3
r
letters 61, 62, 63, 64,
taken one,
=4,
712
4-3 1-2
0,
713
3-2. =
+
1-2 -3
4,
714
= 4-3.2-17 =
1.2.3.4
558, namely,
---
1.
561
Binomial theorem.
(x + b,) (x +
b,)--- (x
If in the formula of
J)
.x
B.x'^-'
+ B^x'^-'- + +
+ B,
we
b^ by the same replace all the n different letters bi, b^, letter b, and x by a, the first member becomes (a b)". Again, since each of the n terms of B^ becomes b, and each
of the
Tij
terms of B^ becomes
''
b'^,
and so
!L_^
on,
we
*>
have,
560,
"
^^ ^^
-^
-tig
o ^3 , ....
to the following
''
+ by = a" +
'^
-L
a"-'b
^''
J.
a"-'b'
^ "^
where
iris's
"^
'
257
The number of terms on the right is n + ! The exponents of a. decrease by one and those ofh increase by one from term to term, their sum in each term being n. 3. The first coefficient is 1, the second is n, and the rest of them may be found by the folloiving rxde :
Multiply the
the term
coefficient
the exponent of b increased by 1 ; the result will be the coefficient of the next term.
and divide by
is
This formula
theorem.
known
Thus, '
{a ^
^ +
6)3 I
a3
^ +
=
Since a
-\-
a3
3 a^b
symmetric with respect to a and b, it follows from 542 that the terms of the same type in the expansion of (a + b)" namely those involving a" and b", a"~^b and ab"~^, must have the same coefficients. But these are and so on the first and last terms, the second and next to last terms, and in general every two terms which are equally removed from the beginning and the end of the expansion.
b \b
562
Hence the
on.
last
term
is
Z",
1,
When there ?i is even, two when n is odd. two middle terms, they are of the same type and have the same coefficients. And by what has just been said, the coefficients of the terms which follow the middle term or terms are but in reverse order. the same as those xvhich precede them
middle term when
are
It
may
also readily be
coefficients increase
563
up
to the
258
564
A COLLEGE ALGEBRA
the sign of b in the preceding formula and obtain
^ ^
By changing
simplifying,
(a
we
by
= a" - na^-^b +
a"-^^*
1T2T3
(2
*"
''
where the terms which involve odd powers of b have and those which involve even powers have + signs.
Example.
signs^
y^)^.
Substituting 2 x for a, and y^ for b, in the formula, and remembering that the last three coeflBcients are the same as the first three in reverse
order, 562,
(2 X
we have
(2 x)^
2/3)6
^ =
- 6 (2 x)52/3 +
6 (2
X)52/3
(2x)* (y^)^
^^^
(2 x)3 (y3)3
+ 15 (2 X)* {?/3)2 20 (2 X)3 (2/3)3 + 15 (2 X)2 (7/3)4 6 (2 X) (2/3)5 + (y3)6 = 64x6 - 192x52/3 + 240xV - 160x32/9 + 60x22/^^ - I2xy^^ + 2/I8.
(2 X)6
565
The general term. From 561 it follows that the term in the expansion of (a + b)" is
(r
l)th
n(n -^
l)(7i
2)
1.2.3...r
to r factors " ^
^-
This, with a minus sign before it when r of (a b)". l)th term in the expansion (r
is
odd,
is
also the
Example
Here n
1.
in the
7.
expansion of (x
yY^.
is
16 and r
8,
or r
_ 16.15.14.13-12-11.10
1.2.3-4-5-6.7
Example
x2o ?
2.
^_
^^^
in the
expansion of
(x3
l/x)i2 contain
,
Let r
+
1
and
Then, since n
12,
x^,
^-r5r
12(a;3)
r/xr
x36- 4
-^
X20.
259
satisfied
if
36
9
x^o
4r
20, or r
4.
x^",
and substituting
in the
formula we
10
= 495 x^o.
1-2. 3-4
EXERCISE XXXI
Expand
1.
(3x
(2
2/)3.
2.
5.
{a
bf.
3.
6.
(1
2x2)7.
4.
1/x)*.
(1
(x-3/x)6.
8.
(x/y-y/x)^.
7. 9.
2 x2)4.
(1
(a2
+ ax- x^f.
x/2)".
6) 12.
{a^
2 bcyo.
in (1
x)^.
Find the
Find the
14. 15.
16.
+ x)8, 2 x)^. +
x
x^)^.
2 x)9
(1
- 2x)".
17. 18.
19.
in (x
l/x)i2.
x''-
in (2
/x)
i^.
Find
Find
{x
+
+
b
2y){x
2) (x
3y)(x
5y) by inspection.
20.
21.
(x
is
3) (x
4) (x
5)
by
inspection.
What
{a
the
number
c
+ d){f+g +
h){k
l)
{m
+ n+p +
q)?
22.
1.
+ X2 + X3 + X4)3.
is
(1
+ X2)2 (1 + X +
?
3.
3x3)2,
23.
"What
the
sum
symmetric
d when expanded
Sn*
Sa2 Sa.
2.
Sa6c.
Sa6 2a6c.
24.
25.
Show that the sum of the coefficients in the expansion of (a + 6)" is 2".
Show
that in the expansion of (a
sum
260
A COLLEGE ALGEBKA
XI.
EVOLUTION
It is possible
566
Perfect powers.
a perfect nth ijower ; in other words, that a second If so, this rational rational function (2 exists such that P Q".
that
is
function Q will be an
function
-nth root of P.
:
A rational In the present chapter we consider the problem P being given, it is required to determine whether or not P is a perfect ?ith power, and, if it is, to find its wth root Q.
We
567 an
suppose n to denote a given positive integer. Roots of monomials. Let P denote a rational monomial
its
reduced to
Tith root of
If P is a perfect wth power, simplest form. P may be obtained by the following rule.
literal factors
a7id multiply the result by the principal nth root cal coefficient ofV.
of P by n, of the numeri-
318.
Thus, (aW/C")"
a^b'^/c"'"
a^"6''' /<;"'",
of a'"'V''/c'"", 566,
and
it
is
by
n.
568
The
is
(compare 258). We shall mean this root wh^n we speak of tJie nth root of P, or when we use the P. symbol
Example
1.
We have
Example
1.
8a36s__ 2ab^
3xy^
2.
roots.
64
-x
2.
v81xV2^'^-
V o,
,.
3.
6/32x^2/30
9
is
569
Roots of polynomials.
Example
1.
ix^
I3x-
6x +
EVOLUTION
261
K this is a perfect
form 2 x^
4a;*
-\-
square, evidently
it
px +
q,
= 4x* + ijjx^ +
which requires,
284,
{p-
4 3)x2
:
+ 2pqx +
32
(f,
that
p and
Vi
p^
(2)
;
+ Aq =
find
2p? =
g
-6
(3),
(4).
we
p=-
3,
i^ its
Example
If this
6x5
21x*
it
44x3
63x2 f 54x
27.
We
will
x6
6x5
21x*
= x6 +
which requires,
+ 63x2 + 54x + 27 = (x2 + px + + 3(p2 + q)xi + (p3 + 6p3)x3 + 3 (i)2g + q1) x2 + ^pq2y^ + ^3^
+
44x3
g)*
3px5
284, that p and q satisfy the six equations 3p = 6 (1), g3^27 3(p2 + g) = 21 (2),... From (1) and (2) we obtain p = 2, q = ^. And these values
:
(6).
of
p and
will
be found to satisfy the remaining equations (3)(6). Hence x^ + Gx^ + + 54x + 27 is a perfect cube, and
its
cube root
is
x2
+ 2x +
the
3.
By
method
illustrated in these
examples
it is
always
possible to determine whether or not a given polynomial in x is a perfect nth. power, and, if it is, to find its wth root.
a^x""
aiS-"'"^
,.
If this be
a perfect ??th power, its degree m, must be a multiple of n so = kn, where k is an integer and it miist have an th that root of the form ax'' + .4^., where (x denotes the AiX^~'^ +
, A,, are principal wth root of ao, and A^, call this root the principal nth root.
unknown
constants.
We
To determine whether
and
aox'"
a^x""
a,
we
set
aia:""i
...
+ =
(ax*
+ ^iX^- +
A^Y.
262
/
A COLLEGE ALGEBRA
reduce the second member to the form of a polynomial in x, and then equate its coefficients to those of the like powers of X in the first member. We thus obtain a system of nk equations in ^1, A^, , A^. The first k of these equations will A and this set of give a single set of values for .4 1, A^, values must satisfy the rest of the equations if aQX'" a^
^.
is
Example
3.
12a;5
I8x*
13x3
9x2
_ 3^ +
i.
570
If a polynomial
is
a perfect
square root
may
also be obtained
by the following
method.
As
denote a polynomial in x
= (a-\-b-{-c + -- -y. powers of x, so that P The problem is, knowing P, to find a, b, c, . Now, whatever the values of a, b, c,- may be, we have + bf = 0^ + 2 ab + = + {2 a + b)b, + cy=(a + by + 2 (a + b)c + c" {a + = a^ +(2 a + ^*)& + [2(a + -f c]c, c ia-\-h -^ dy = a" + i2a -^b)h -^\_2{a + b)+ cy + [2(a + i + c)+fZ]c?,
{a
b''
of even degree and arranged in descending powers of x. Let us suppose that P is a perfect square and that a, b,c, denote the terms of its square root arranged in descending
a"-
^-)
-\-
each
and so on, a new group of terms being added on the right with new letter on the left, namely, a group formed by adding the new letter to twice the sum of the old letters and multijilying
the result by the new letter. Therefore, since by hypothesis
P=
(a -f i
+c+
)'^,
we
have
P=
a"
+ (2 a +
b)b -\-[_2{a
^[2(a
+ b) + cy + b + c) + d-]d + .--,
EVOLUTION
263
where the leading terms of the several groups on the righi, namely, a^, 2 ah, 2 ac, 2 ad, -, are all of higher degree in x than any of the terms which follow them.
From
1.
this identity
we may
find a,
b, c,
as follows
Evidently a is the square root of the leading term of P. 2. Subtract a^ from P. As the leading term of the remainder, Ri, must equal 2 ab, we may find b by dividing this term by 2 a.
Having found b, form (2 a + b)b and subtract it from R^. the leading term of the remainder, R^, must equal 2 ac, we may find a by dividing this term by 2 a.
3.
As
4.
is
reached.
If this final remainder
is,
as
square and
+c+
P
is
to the
form
P=(a + b + c +
that
is,
..-y+R,
which
It is
is
and an integral function of lower degree than a. convenient to arrange the reckoning just described as in
of a perfect square
to the
sum
4 z^
13
a;2
6x
9.
&)
13 x2
x2
6X 6X
+ +
12 x2
9
9
12x2
-6x +
is
is 0,
2x2
3.
Compare
569,
square
divide
root.
Observe that as each new remainder Pj, R^, is found we its leading term by 2 a and so get the next term of the
Then
new term
of
264
the root.
A COLLEGE ALGEBRA
We
multiply this
sum by
the
new term
of the root,
subtract the result from the remainder under consideration, and thus obtain the next remainder.
Example.
40x3
46 x-
- 24x +
9.
571
This method
more
first
applicable to a polynomial P which involves than one letter, provided it be a perfect square.
We
powers of one of the letters, with coefficients involving the rest, and then proceed as in 570, it being understood that x now denotes the letter of
arrange
in descending
arrangement.
572
Approximate square
roots.
We may
method
to a polynomial in x arranged in ascending powers of this will then be arranged in ascending letter. But a, b, c,
powers of
and the degrees of the successive remainders will increase. Hence, 570, 4, if P is not a perfect square but has a constant term, we can reduce it to the form
x,
P = (a^b + c +
that
is,
..-y
R<,
of a perfect square and a polynomial, R', whose lowest term is of as high a degree as toe jilease. For small values of x we can make the value of i^' as small
to the
sum
Hence carrying this reckoning far enough. we call a -\-h, a -{- h -\- c, the approximate square roots of P to two terms, three terms, and so on. It should be added that these approximate roots are found more readily by the method of 569.
as in this case
we choose by
Example L
x to four terms.
By
569,
we
Squaring,
Hence,
or solving,
284,
Therefore the
Vl + x =\ \-px + qx- + rx' + 1 +x=l +2px + (p^ + 2q)x^ + 2{i:>q + 2p=l, p^ + 2q~0, pq + r = 0, r=l/16. p = l/2, q = l/8, x-/8 + x^/W. required result is 1 + x/2
r)x^
Example
2.
x'^
to three terms.
EVOLUTION
Square roots of numbers.
also derive the ordinary
265
From
method
the formulas in
570 we
573
of a ymmher.
Example.
of 53361.
Let a denote the greatest integer with but one significant figure, whose Its significant figure will be the leading square is contained in 533G1, We find a as figure of the root and its remaining figures wiil be O's.
follows
:
Remembering that for each at tlie end of a there will be two O's at the end of a-, we mark off in 53361, from right to left, as many j^eriods of two figures as we can, thus 5'33'61. at the end of a, and the Each of the periods 61 and 33 calls for one
:
than
5.
Hence a
200.
Having found
a,
we proceed
a polynomial. This is denotes the second figure of the root multiplied by 10, and c the units The scheme at the right gives the reckoning as abridged in figure.
quite as when seeking the square root of indicated in the scheme below at the left, where b
common
practice.
a
'5' 33'
61 200
1
+ b + + 30+
c 1
5' 33'
61 231
1
4 00 00
2 a 2
=
==
a^
=
=
400 430
1
1
33 61 29 00
Ri
43
1
1
2 (a
2 (k
6)
+ 6) = +c=
460 4 61
461 4 61
= = =
(2a Ba
[2{a
+
+
b)b
33 29
4 61
461 4 61
b)
=R
a',
i?i
c]c
"We
first
subtract
the remainder
by 2
a,
then find the significant figure of 6 by dividing next find Ro hy subtracting (2 a + b)b from Ri,
by dividing R2 by 2 (a + b). simplest way of accomplishing all this, as indicated in the abridged scheme at the right, is to omit final O's and to bring down one period
and
finally c
The
Then, as each new remainder is obtained, we write at its twice the part of the root already found as a "trial divisor," obtain the next figure of the root by dividing the remainder by this trial divisor,
at a time.
left
divisor
by
it.
We
then multiply
the complete divisor by the new figure of the root, subtract, and so obtain the next remainder. If too large a figure is obtained at any stage In the process, that is, a figure which makes the product just described
we
4 47
EVOLUTION
Therefore, since by hypothesis
267
(a
P=
+ b + c-\---
)',
we
have
P=
a''
+ (3
a^,
a""
right,
, are all of
From
1.
this identity
we may
find a,
i, c,
as follows
Evidently a is the cube root of the leading term of P. As the leading term of the remain2. Subtract a^ from P. der, Ri, must equal 3 a^b, we may find b by dividing this term by 3 a\
3.
from
4.
equal 3
we may
find c
and subtract it must by dividing this term by 3 a^. a remainder is reached which is of
-{-
3 ab
-\-
b^)
is 0,
then
is,
as
was supposed, a
^ c perfect cube and its cube root is a If this final remainder is not 0, P is not a perfect cube, but we shall have reduced it to the form
.
+ + +
P=
where
(a
+b+c+
a^.
---y
+ R,
:
is
Example.
6x6
+
x6 x^
21x*
44x3
63x2
543;
27.
|x2
3 a2
3 X*
6 x5
21 X*
44 x3
63 x2
54 X
+ 2x + + 27
27
= 3x* 6x5 + 21x* + = 6 x3 + 4 x2 3x2 2 X + 3x* + 6x3 + 4x2 6x5 + 12x*+ 3(x2+ 2x)2 = 3x* + 12x3+ 12x2 9x* + 9x2 + 18x + 9 3(x2 +2x)3 + 32= 3 X* + 12 x3 + 21 x2 + 18 X + 9 9x^ +
3(x2)2 ( 2 x)2
44x3
8x3 36x3
63x2
+ 54x +
= =
Bi
+ +
63x2 63x2
+ +
54
a;
27 27
R2
30x3
54x
=B
268
A COLLEGE ALGEBRA
is 0,
x^
6 x^
54 x
cube and
its
cube root
is x'^
2x
3.
Compa''e
569,
+ 27 is a perfect Ex. 2.
Observe that as each new remainder R^, R^, is found, we its leading term by 3 a'^ and so get the next term of Then at the left of the remainder we write the sum the root.
divide
of three times the square of the part of the root previously obtained, three times the product of this part by the new term, and the square of the new term. multiply this
We
term, subtract the result from the remainder under consideration, and thus obtain the next remainder.
576
This method is also applicable to a polynomial which involves more than one letter, if it be a perfect cube (compare 571). The method may also be applied to a polynomial in x
arranged in asceiiding powers of this
lack a constant term.
letter,
if it
does not
we
577
If the polynomial is not a perfect cube, thus obtain approxiinate cube roots (compare 572).
We may
575.
a
8 000 000
-ir
=
= =
i?i
= iV-a3
900
138900 4167_000 = 158700 320 168
1380
4
(3 a2
i?2
3 a6
62)
3{a
+ 6)2 + 6)c= c2 =
-ZV'
(a
6)3
[3 (a
6)2
3 (a
6
=i? = iV-(a +
In order to find
a,
+ 6) c + + cf.
c?] c
the greatest
significant figure
whose cube
figures in
is
contained in N, we begin by marking off periods of three from right to left (also from the decimal point to the right
:
when
12' 487' 168. there are decimal figures in iV), thus Each of the at the end of a, and the remaining periods 168 and 487 calls for one
EVOLUTION
cube
period, 12, calls for the initial figure 2, 2 being the greatest integer is contained in 12. Hence a = 200.
269
whose
rest of the reckoning is fully indicated above. Observe that each new figure of the root is found by dividing the remainder last obtained by three times the square of the part of the root
The
already found ; thus, we find tlie significant figure of b by dividing Ei by 3 a"-^, and c by dividing R^ by 3 (a + b)'^. If too large a figure is thus obtained, we test the next smaller figure.
same way
cubes
Approximate cube roots of numbers which are not perfect may also be found by this process (compare 574).
578
is
which
of polynomials. The fourth root of a polynomial a perfect fourth power may be obtained by finding the square root of its square root similarly the sixth root of
Higher roots
a polynomial which is a perfect sixth power by finding the cube root of its square root.
may
be obtained
It is also possible to develop special methods, analogous to those of 570, 575, for finding any root that may be required. But the general method of 569 makes this unnecessary. In
we have given the special methods for square and cube roots explained in 570, 575 only because of their historic interest and their relation to the problem of finding square
fact
XXXn
125
aV2
\625c2d3
'^
'
By
4. 5.
6.
569 or
X*
x2
-2x3 + 3x2-2x+ - 2 X* + 6 x3 - 6 X +
x6
9,
4 x6
12 x^y
9 x^y'^
4 x^y^
6 x^y*
y^.
7.
4x2-20z +
13
+ 30/x +
9/x2.
270
8.
9.
A COLLEGE ALGEBRA
49 -84a;
x8
{X2
-34x2 + 60x8 +
25x*.
5 x2
2 x^
x6
x*
6 x3
4X
4.
10.
11.
12.
16x2
24x?/
16.
X2/2/2
+2+
2 X2
2 y2
x22/2.
-2x.
14.
4-x + 3x2.
+
1.
By
15.
569 or
x6
3 x5
X*
7 x3
6 x2
3X
27x12
8 x6
6 x*
90 aH^
135 aH^
+ 3x2- i_ - 81 a^x + +
135 a*x2
27
7.
a^.
xVy^ +
x2.
2/Va;^
3x2/2/2
3y2/x2
+ 6x/y +
62//X
19.
1
-X+
20.
By
x8
4x^
10x6
16x5
+ + +
i9a;4
16x3
iox2
_ 4x +
1.
21.
By
fifth
root of
xio
5x9
+ +
15x8
30xT
45x6
45x*
+ +
51 x^
30x3
15x2
+ 5x +
1.
22.
To make
x* to
Ox^
11 x2
ax
5 a perfect square,
what values
must be assigned
a and
h ?
of the following
numbers.
25. 583.2225. 28.
27889.
24. 27.
2313.61.
4149369.
.00320356.
9.024016.
Find
30.
55.5.
31.
234.561.
1800867.
33.
167284.161.
34.
1036.433728.
IRRATIONAL FUNCTIONS
XII.
271
Roots.
tive
a, b,
579
numbers
values.
Again, va will denote the principal wth root of a, that is, the positive number whose nth power is a in other words, the
;
positive
number which
is
a.
Finally,
when
a,
71
is
odd,
V a will denote
a.
root of
namely
is
we
shall
mean principal
root.
a restricted use of the word root ; for any number whose Note. nth power equals a is itself an nth root of a, and there are always n such numbers, as will be proved subsequently.
Thus, since 2^
of 4.
580
4 and (- 2)2
4,
both 2 and
We
by -Vi.
the
is odd and a is real, one of the nth roots of a is real and of same sign as a, and the rest are imaginary. When n is even and a is positive, two of the nth roots of a are real, equal numerically, but of contrary sign, and the rest are imaginary.
When n
When
is
even and a
is
In the higher mathematics va usually denotes any nth root of a, not, as here, the principal root only.
Radicals. Any expression of the form Va^ or h Va is called a radical; and n is called the index, a the radicand, and b the coefficient of the radical.
581
When
b
both a and
b are rational
numbers or expressions,
Va
is
Thus
vi
is
is 3, its
radicand
4,
and
its
coefl&cient 5.
272
582
A COLLEGE ALGEBRA
The rules for reckoning Formulas for reckoning with radicals. with radicals are based on the following formulas, in which
m,
n,
4.
/ {ydf =
71/
^aJ'\
5.
m/ V Va =
n
r-
mn ,
Va.
not changed
to both
Observe in particular that, by 1, the value of a radical is if its index and the exponent of its radicand are
if
similarity of this thus, rule to the rule for simplifying a fraction is obvious. These formulas may be proved by aid of the definition
is
cancelled
Va.''
Va^.
The
(.")"
""",
(aby
a"b",
Two
any
like
powers of these
Va'"
(
For
2.
Va"'i^)"^
a'"''
and
(Va'")"-^
(aP'Y
a"'''.
'Vab
(
= Va
Va6)"
For
Q
'
=
,
and
Va
N^ti)"
(Va)
Vft)"
ab.
\/-
=
Vb
/
^^
since their th
^ For
4.
"fax"
(VJ=^;
(
J and /
(-)=
Va \ "
Va)"
=-.
Va)'"
For
(Va"')"
a'";
and [(Va)'"]"
a"'.
IRRATIONAL FUNCTIONS
m
5.
/
273
Va
(
= Va,
since their
;
mnth powers
are equal.
(va) = a. and [^ y^a) Va)""' = a The following examples will show the usefulness
For
of these
formulas.
1.
-v^S
= V^ = V2.
VSc
2.
3.
^1^
(
VSc
de2
3
,
4-
V V V32
,
Vd^
X1&2/5
5.
- On That form of a radical is regarded simplifying radicals. as simplest in which the radicand is the simplest integral Hence for simplifying radicals we have expression possible.
the following rules, which are immediate consequences of the formulas just demonstrated.
1.
If the radicand
common with
index.
be a poiverivhose exponent has a factor in the index, cancel that factor in both exponent and
Thus,
2.
is
//
Vl6x"2/9
V2%x3y8y
2 xy'^ Vx^y.
its
numerator and
render
it
tvill
pos-
^,
Thus,
-yA-^-x
s/xw
14
\
= Vn
xyz
1
^
4 xyz.
Similar radicals. Eadicals which, when reduced to their simplest forms, differ in their coefficients only are said to be
similar.
584
Thus,
VTx^
2
and
VsTx^
9 x^y
"^''xy,
are similar
forms,
namely
Vxy and
274
585
A COLLEGE ALGEBRA
the coefficient of a radical under the radical sign.
Vjfri, the coefficient of a radical
if its
On bringing
Since b
V^ =
EXERCISE
Reduce each of the following
1.
XXXm
Vl8.
2.
V588.
-^-272.
^^374.
11. 14.
4.
^-
1000.
5. 9.
V372.
6.
V372.
10.
7.
8.
V?7l6.
V25 asfeiocisde.
Vl28aW.
Va"62"c3.
16. 18.
v'Sx^z/^zis.
13.
Vx22/2
V'xG
a;'2;22.
V(x2
-t/a^^*
?/2)
(X
17.
xSyS.
O'^d^
19.
4t^. \32a62
a/ 1
20.
J-^
X2 21.
+
^
1
-
\a-h
<f^ + \9(x
-V
l)2
22.
23.
A/
24.
sign.
Bring the
25.
SaVs^.
26.
^^-L^a/"^'.
Show
28. 30.
2/3)
(X
y)
and Vx*y2
a;22/*.
We
To reduce the algebraic sum of two or more radicals to its simplest form, simplify each radical and then combine such of them as am similar by adding their coefficients.
IRRATIONAL FUNCTIONS
Example.
275
Vi/2.
3 V2, and
We
cannot be
Thus, we cannot have Vx + Vy=^ Va; + y except when a; or ?/ is we have x + y + 2 y/xy = x + y, .-.2 Vxy = 0, .-. xy =
or
2/
0,
0.
Reduction of radicals to a
common
index.
It follows
from the
587
we can always reduce two or more radicals to equivalent radicals having a common index. The least common index is the least common multiple of the given
formula
Va"'* that
indices.
V a'" =
Example.
to their least
common
index.
8, is 24.
The
6.
least
common
8,,
And
V a5 = U, Va2o
and V63
= Vfts.
24
We make the reduction to a Comparison of radicals. index when we wish to compare given radicals.
Example 1. Compare v 16, V6, and v3. The least common multiple of the given indices,
15 y
common
588
V 16 = V162 = V256
Therefore, since 256
30
30/
10
;
V6 = V& = V216
have
30,
15, 10, 6, is
30,
30,
30
and
30;
15
6,
Example
2.
Bringing the coefficient of the first radical under the radical sign, 585, and then reducing both radicals to the common index 6, we have
2
^41
2
= VlP = ^1681.
VJT.
Vs >
From
71
y
the formulas
589
n r~
\/a-^b
we
= -\ab
ni
and
:
^a/^h = ^a/b
ni
276
A COLLEGE ALGEBRA
if necessary
To multiply or divide one radical by another, reduce them Then to radicals having the least common index.
coefficients
and radicands
Example
1.
Multiply 4
2
Vxy by
Va:'-?/^.
We have 4 Vxy
Example
2.
V^- = 8 Vx^^
2
-v^V =
VxV = 8
a;?/
Vxy.
Divide 6 Vicy by
6
vxy.
We
590
have
Vxy /2 Vxy =
Involution.
From
the formulas
(V^)'
= Va^
'v'a'"'^
= Va"'
the
we
to
mth power,
dancel
any
factor tvhich
may
be
common
to
and
radical,
Example.
We
(2
have
V^ =
/
591
Evolution.
From
n,
the formulas
m/ V >/ = Va
nmi
and Va'"^
np
= Va"
we
To
any
radical of the form Va, cancel fiiid the mth root of a and the exponent of the factor tvhich may be common to
Example
1.
We have
Example
2.
\xy^
We
have
-^Uay^h = ^S^
^VTm = 3 vT^.
IRRATIONAL FUNCTIONS
Simple radical expressions.
277
592
By
we
call
shall
Thus, Va + Vo is a simple radical expression. We such an expression integral when it involves no fraction with a radical in its denominator.
only.
By
powers of simple integral radical expressions can be reduced to algebraic sums of simple radicals. In 607 we shall show
that the like
is
true of quotients.
-\-
But ordinarily a
'vb,
root of
cannot be reduced to
Multiply
We hare
(3
Ve +
V5)
(2
Veo -
Vso
Example
2.
Square
V2 + Vi.
2
We have
(V2 +
\/4)2
V2 V4 + vT6 =
V^4-2 v^.
EXERCISE XXXIV
Reduce the following
6/
1.
to their least
common
3,
2.
index.
i.
10,
IS,
Va^,
6,
Compare the
3.
Vi
and 2
4.
Reduce each
5.
V35 - VTJl.
10
- V5.
.
7.
.
4
2 2
- V'2.
V3
-^
8.
Ve VlO
VT5.
9.
v^
-v^S
a/qo
vTs.
10. 13.
3 \^.
11.
14.
V2-^/2--^.
Va^b^c^
12.
-!-
v^.
'
y^a^c\
15.
17.
V^
'V^.
.
16.
V^ - V^.
Va26c2
Va62ci.
278
18.
A COLLEGE ALGEBRA
Va
Va.
15
,
19.
Va/fe-Vo/d.
20.
21.
va&2
Vafts
-4-
(VaT59
22.
Vai26")
23. (2 Vx2/223)6,
26.
(Vl2)3.
{\/a2)8.
24.
VV^V v^.
V V2..^2.
25.
vV8.
^a366/c9.
27.
28.
V2V2.
2
29.
V2V2.
/-"/an , \nmp "~/2n -\n
30.
31.
Vva"
m/
32.
(V
Va)
V^.
Vti.
34. 36.
38.
VSOO-^
los
+ vTTi. - Vax^ -
37.
39.
6 ax2
9 ax
4 a2x2
4 a^x.
w Vx + 2/
41.
43.
\x
Ve.
w y
(
\x2 \x^
y''
(Vi + Vs+Ve)-
4S..
Vo + VlO + Vl4)
-=-
V2.
(V6+V5)(V2+Vl5).
(1
44.
V5 + 2V2 V5 - 2 Vi".
(Va + Va+l)(Va- Va + 1)-
45.
+ V3)3.
46.
In many cases reckoning with radicals is greatly facilitated by the use oi fractional exponents. Thus far we have attached a meaning to the expression a" The rules for reckononly when n denotes a positive integer.
ing with such expressions, namely,
1.
n a" a"
= f.m + n
a'
2.
(a^Y
^"'"5
3-
(^)"
= *"*">
among the simplest in algebra. Can we find useful meanings for a", in agreement with these rules, when n is not a positive integer ?
are
It is therefore natural to
inquire
IRRATIONAL FUNCTIONS
The
definition &v/(i
279
Va^. Takfi_af_for_matance. wish, a meaning for this symbol which will be in agreement with the rules 1, 2, 3. But, to be in agreement with 1, we must have
if possible, to find
We
594
(a^y
that
is,
= a^.a^ = J^'^ =
a^
a,
We We
must mean either Va or Va. choose the more convenient of these two meanings, and
a^
define a^ as
va.
we wish
a^ as
a^
Va,
Va^,
and
of
is,
aP,
is
Observe that since a* = Vo^ = Vo^ = a'", the value of a not changed when ^9/3- is replaced by an equivalent fraction.
Thus, a^
a^
a^] also
a"^
a^
a*.
The
definition
a^=
1.
l,^
595
we must have
and therefore
a
= a^+"' = a, = a^/a*" = 1.
596
We are
The
1,
definition
we must
have,
595
or' a'
and therefore
We
It
/a^,
a~^
/a^
/ \^aP.
remains to prove that the meanings thus found for a, a**, and a~* are in complete agreement with the rules of exponents.
Theorem
values of va.
1.
The and u.
lav; sJ^
Si
= sJ^^'^
597
280
Let p,
1.
q, r, s
A COLLEGE ALGEBRA
denote any positive integers.
Then
When m = p / q
-
and n
SI
7'/s,
we have,
gs,
582,
2.
a^-a' = y/aP Va'" = V a^" Va'" P ps + _ -<JaP' + = .'' = at \ When m = p/q and n = r/s, we have, by Case
7/
qr
r
qs
I
qsi
'f
1,
a "-a
'
1 = ^^-r =
a"
a'
=
r/s,
-^r
a
-'-+(--) "
'.
3.
When m =p /q
T> -
and n
Si
=
qsi
and p/q
> r/s,
we have
T '
ai-a
q, = <Ja^ / <J =
a''
-y/a'"
Va^"-'"'
4,
/ Wa"''
qs I
qr
">'
When
p
m=p/q
3,
and n
=
1
r/s,
and
p/q<r/s, we
have, by Case
1
9 ..
?+(-0
598
Theorem
2.
The law
(a"")"
a"^"
vahies of ni and n.
For, let
1.
Then
When
(a)
is
a"*
a piositive integer,
a'"
we
have,
'"
597,
to
factors
+ '" +
<''>"
^ ^mn^
we
2.
When n = p / q,
1,
where
integers,
have, by Case
p (a"")?
= V(a"')" =
7/
s,
7/
Va""'
mp
P m-.
3.
When n =
1, 2,
where
s is
any positive
-""
rational,
we have,
by Cases
(a^\-' ^ ^
11 = =
(a.'")"
a"'^->.
a""
IRRATIONAL FUNCTIONS
Theorem
values of n.
1.
281
599
3.
The law
(ab)"
p and
Then
(aby
2.
= -^(aby =
s,
-^a^b"
Vop
-Vb^
Let n
where
denotes
whether integral or
fractional.
'
{aby
Applications.
aflf
will
illustrate the
600
and negative exponents. A complicated piece of reckoning with radicals often becomes less confusing
use of fractional
when
Example
1.
Simplify
^ a/^a. V a/
3)*
6
We
have
2.
\a/y/a-.{aa
4
(f(3)5
Z,
ai
= Va.
Example
Simplify
^ ah^ v a^b
Va-ft^
h-
^a-b^.
We have
Va6^
va^6
-^
a^b^
a%^ a~%~^
12 ^
Example
3.
Expand
(x^
(x'
y~^^.
We
have
y-iy^
(xS)3
+ 3(a;3)2y-i + 3xi(y-i)2 +
Z
(j^-i^s
x~
-\-
x^" + 3 x%~ ^
^
-\-
y~^.
Example
4.
Divide x
?/
by x^
x^y^
+
2
y^.
in 401,
1
we have
x^
+
y
11
x^y"^
+ y^
x?w3
x'w' x^y^
x^y^
x'
3.
282
A COLLEGE ALGEBRA
EXERCISE XXXV
'^.
2.
V^.
b
3.
aV^.
4.
Vb*
^.
7.
a'^.
6. 8.
c-i-5.
(d')-6.
(e-^)-i
a- 1/6'0-2.
10. 12.
x~^Vy^.
a;-2
11.
(l/Vx^)-4.
Vj^/2/-2 Vx^.
6-1 c-2
a-i(6-i
a-2(6
+ +
c-i)
c)
6-
+ +
c-i
Reduce each
14.
17.
(3i)l
8-^.
(a^6)^a56*.
8li.
16.
(-27)?.
a^oThr-^^.
{ai)K
ah^K
a6-2/a-36.
20. 23.
21.
24.
(a-i6-2c3)-2.
(-32aio)l
{_ a66-9)-3.
26. 28.
6-iv^--6-i
V6=^.
(a-W6^)3.
Va'^ (6c1)-2.
(8a-iV^125a3)-3.
30.
32. 34. 35.
Va-i%'a3.
[(a;'^)^]''.
31.
"v'a?
33.
(x**
+ ^yy!''+^)=^+!'.
(X'
r)/(x-i
y"'*).
Multiply x*
+
6^
x^y'
y^
by x'
6'.
x'j/^
y.
36.
37.
Divide a^
by a^
yh^)*.
Expand
(x'
39.
x^
4x^y^
40.
+ 3 x^
Simplify [{^ + e-=^)2 - 4)]i ^ ^xy + 6x%^ + 'I2y^ + 9x-^y*. + 6 x + 7 + 6 z-i + 3 x-2 + x-s.
38.
IRRATIONAL FUNCTIONS
THE BINOMIAL THEOREM FOR NEGATIVE AND FRACTIONAL EXPONENTS
If in the binomial expansion,
283
561,
ti(n
J.
(a
by
= a" + na"-^b + V
T) ^
a^b^
we
we
for
601
1) /2, will then be 0. n(n shown further on that ii b < a the sum
;
of the
terms of this series will approach the value of (a + b)" as limit when 7n is indefinitely increased in other words, that, by adding a sufficient number of the terms of this series, we
may
we
please to
the value of (a
+ by.
is
This
tive
is
what
meant when
iu
is
+ by
when n
fractional or nega-
and
<
1.
a.
Example
Putting
71
Expand
1/3, a
(8
8, 6
= x~^
in the formula,
we have
(8
+ x~*)^ = 8' -f ^
S-'x"^
+ 5l_^8-8(x~5)2
V
;
2-3
. =2
X
-f
X
288
5x
'
.
12
20736
expansion of
1
Example
(a^
2.
in the
/ (a^
-{-
x^)^ or
x^)-2.
Putting n
term, 565,
2,
o^, 6
= x^,
l)th
we have
,
^
' ^
(-2)(-3)(-4)(-5)(-6) 1-2-3-4.6
,
'
^_
^ ^_,^^_
284
Example
Since
3.
A COLLEGE ALGEBRA
Expand
x
vT+x to
four terms.
Vl +
=
x
(1
x)^ we have n
|,
1, 6
x,
Hence
^\
\-
-\ ^
=
1
^x
X
2
^-
x2
+
.
2-3
X2
X3
16
The
result is the
4.
same
as that obtained in
572,
Ex.
1.
Example
of
v 10.
We
and
have
VlO =
i)^
(1
\)\
3(l.,)^.3[l.^.l.^(i;
2-3
216
Q-]
+
.00025
3888
.00462
.16666
3.1623 nearly
EXERCISE XXXVI
Expand each of the following to four terms.
1.
(1
x)i
i
2.
(a^
+ x-^)-l
_i
=)-*.
3.
4^(27
-2x)2. 3V?/)-6.
4.
(a^
x)"*.
5.
(a-i-6
8.
2
10.
+ 3x
^
-^(1
e.
(Vx +
9.
x)2
(^ Vi'y Wi +3
in (1
in
x)-3,
11.
12.
(x-2
2 y*)^.
in (1 in
x^)*.
13.
14.
x~^
(2
x~^)-3.
By
the
method
approximate values
of the following.
1.
V99.
2.
V^.
3.
V3I.
IRRATIONAL FUNCTIONS
RATIONALIZING FACTORS
285
the product of two given radieach of these expressions is called a rationalizing factor of the other.
Rationalizing factors.
cal expressions is rational,
When
602
Thus,
izing factor of
Hence
Va + Vft
is
a rational-
It can be proved that every finite expression which involves The following simple radicals only has a rationalizing factor. sections will serve to illustrate this general theorem.
Every expresand integral with respect to v .c can be reduced to the form A -\- B V^, where A and B are rational and integral with respect to x and A -\- B Vic has, with
sion which
is
603
A B
'wx, obtained
by
this
+ 3x(Vx)3 may
be written 2x2
a;2
x'^
Hence
We may
which
is
obtain
a rationalizing factor
of
an expression
and integral with respect to any finite number of square roots, as Va;, Vy, V^, by repetitions of the For we shall obtain a result which process just explained.
rational
is
completely rational
its
if
we multiply
rationalizing factor with respect to Va-, the product by its rationalizing factor with respect to V^y, and so on.
by
Example.
We We
or
have
(1)
Multiply
by
obtain
Multiply
(3)
by
We
is
obtain
ofl+Vx + Vy + + V^ + Vx (1 + 2 ^y). 1 + Vy- Vx(l + 2 Vy) _ x (1 + 2 V^)2, (1 + V^)2 1 -x + y -ixy + 2V^(\ -2x). 1 - x + y - ixy - 2 Vy {1 - 2 x) - 4:Xy)" - 4 (1 - 2 x)2. (I -x + y
. .
Vxy.
(1)
(2)
(3)
(4) (5)
(2)
?/
is
and
(4)
286
604
Rationalizing
A COLLEGE ALGEBRA
factors
of
binomial
radical
expressions.
The
form
Va V6
Example.
v a + v6.
We have
Va + V6 =
+
a'
6^
(a^)'^
(63)5.
(1)
rational expression
(63)1
(2)
(a^)^
(b^)'^
Hence
(2) is
(1).
605
On
Any
irrational
/B, in which B involves simple radicals only, may be reduced to an equivalent expression having a rational denominator by multiplying both A and B by the
expression of the form
rationalizing factor of B.
j
Example
1.
We u have
-m-
1 =
3
a*
3
1
a^ V = ^a/a.
, '
Example
2.
VxVx2
+ a^ + Vx2 + a2 - Vz2 _
a'^
a2
We have
Vx2"+
a2 a2
Vx2
+ Vx2 _ _ Vx2 -
a2
_
(
a2
Vx2 +
a2
_
606
x2
+ Vx* -
a*
cal expression
In computing an approximate value of a fractional numeriwhich involves radicals, one should begin by
Much
unnecessary reckoning
thus avoided.
Example. Find an approximate value of
(1
+ V8) / (3 v^)
.
+ v^ .
which
is
Wehave
V2)
2.414
...
IRRATIONAL FUNCTIONS
287
607
To divide one radical expresDivision of radical expressions. sion by another, we write the quotient in the form of a fraction and then rationalize the denominator of this fraction.
Example.
"We have 4 + 2V5 + V2 + V5
Divide 4
+ 2V5byl-V2+ V5.
3
(1
+ 2V2-V5 + VT0 V2 - 1
^/^
^
l)
^^
^/^_
It follows from 592 and 607 that every General result. expression which involves simple radicals only can be reduced to an algebraic sum of simple radicals.
608
EXERCISE XXXVn
Find rationalizing factors of the following.
1.
\/^.
2.
y^cC^
V63.
3.
x^
x^
xl
4. 7.
9.
Va + Vbc. Vx + Vy - Vi _
x3 x^ 3
3
5.
vx + Vy +
8.
Vz.
6.
Vxy+Vyz+Vzx.
1.
Vii.
Vx + Vx +
11.
+ yK
+
,
10.
13.
Va 1 1
_|_
v62.
x* x^
1
- yK
12.
2/1
x^yK
14.
+
3
x^
1.
15.
18.
- Vs.
3
,
16.
+ V2 +
V3.
3
.
17.
3
I
+ \^.
1^
1.
V9 + V3 +
its
19.
VI2+V6+V3.
number
Keduce each
or expression for
20.
denominator.
21.
-J.
V^4^
^-
^^.
a-V6
2^.
22.
^-^
2V3 + 3V3
23.
5 4.
..
^/^^ + ^^^
Vx +
?/
V62
a2
- vx
^
25.
L^i+^.
1_V2+V3
26.
I+V2 + V3 + V6
288
-\-
A COLLEGE ALGEBRA
xVy y Vx Vx + Vy + Vx+
y
1
Vs -
\/3
^.
^125
30.
^-2^.
V?
31.
^^^"^ V2 +V3
IRRATIONAL EQUATIONS
609
On
rule.
Next, solve the resulting rational equation. in the given equaFilially, test all the solutions thus obtained tion and reject those which do riot satisfy it.
the denote the given equation, and For, let P rational equation obtained by multiplying both members of hy R, the rationalizing factor of P. By 341, the roots
PR =
P= of PR =
are those of
P=
and R
= P=
jointly.
We
dis-
by testing them
Example.
Solve x
- Vx -
.5
0.
members by
(X
+ vx
6.
7)2
x^
Solving, by 455,
- (X - 5) = 0, - 15x + 54 =; 0. x = 9 or 6. x - 7 - Vx- 5 = 0, we
a root.
6
have 9
- 7 - V9-5=0,
is false.
which
6
is
is
true.
Hence
is
- V6 -
0,
which
Hence
obtained
is
'- 5
= 0,
Vo^ 5 =
true.
IRRATIONAL FUNCTIONS
289
An
may
be ration-
610
one member and the remaining terms in members to the n\h power.
By
from
equation which involves be completely rationalized. It follows 345 that this method is equivalent to that described in
may
609, but
it
Example
1.
v Vx + a =
Vft
Vx +
a x
= 6^ = (&^
a)2.
we^nd
it
to be a root.
Example
2.
Solve
Vx +
Transposing,
Squaring,
Simplifying,
+ Vx Vx x Vx + x +
= = = = = =
9.
4
4 5 5
81
5.
Vx + 0. - 18 Vx +
+x+
5.
Squaring,
Solving,
25.
20.
which
9,
Notes.
Observe, as in Ex.
1,
that
respect to the
unknown
letter
only and
611
which do not involve this letter. 2. Observe also that a n irra tional equation may have no root. Thus, the equation Vx + 5 Vx 4 = 9 has no root. For
attempt to solve
3.
it
if
we
we
We may
form
x = 20 and V25 Vl6 = 9 is false. add that the simplest method of rationalizing equation
of the
is
VI + V^ + Vc + V^ =
it
(or
V^ + V^ + Vc +
J5;
0)
to begin
by writing
thus
{otVa
and then to square both members. The resulting equation but two radicals and it may be rationalized as in Ex. 2.
290
612
A COLLEGE ALGEBRA
To
solve a system of such
equations we
may
first
the resulting rational system, and finally test the results thus obtained in the given system. But if the equations are of the form described in 379, they should be solved by the method there explained.
Example
1.
Solve
Vx
(1) (2)
Squaring
or
(1),
25
+ 'Z\'xy,
(3)
Squaring
Solving
(3)
and simplifying,
(4), (2),
Sub.stituting x
is true.
Hence
2.
= = =
Vxy.
5.
(4) (5)
which
Example
Solve
Vx +
2 V'x
6
6
+ 2/Vy = + 6/\/7/ =
4,
9.
(1)
(2)
Solving for
Vx +
(3)
And from
and
(2).
we
and
(3)
(1)
EXERCISE
Solve the following equations for
1.
XXXVm
=
2.
x.
X*
4.
2.
x~^
3.
3.
x^
8.
4.
(V2x-l)^ = V3.
5.
a/2
+ VS + Vx =
X
6.
V^ + Vftx + V^ =
Vx +
Vx +
4
d.
7.
9.
V4 x2 +
10
2X
1.
8.
+ Vx
X
-I-
-f
11
7.
10.
-6 =
0.
11.
+ 3x -
- Vx-^-x- 1 = 2.
12.
Vx +
Vx + 2 + Vx - 2 = Vx +
+ Vx 14.
1.
13.
:^^+^ Vx + 3
-v^ 6
2.
--i
Vx+1
Vx-1
^+-^^ = Vx2-1
0.
IRRATIONAL FUNCTIONS
Solve the following for x and
^
f
291
y.
Vx +
y/y
15.
+ Vy-2 = Vx + 5 + Vy + e, -X- Vs - X + Vy - 3.
17
^^
lb.
-j
17.
Show Show
Vx +
+ Vx + 6 + Vx +
h
+ Vx +
will reduce
that
Vax +
reduce to a
QUADRATIC SURDS
Surds.
Numerical radicals
is
like
V2^and Vo,
in
which the
are called
613
radicaud
surds.
is irrational,
surd
is
as its index
is called quadratic, cubic, and so on, according two, three, and so on.
is
614
Suppose that when the surds have been reduced to their Va and Vi. The
product of
Va
and
V6
is
'wab,
and this
is
a surd unless ab
is
a perfect square. But ab cannot be a perfect square, since by hypothesis a and b are integers none of whose factors are square numbers, and at least one of the factors of a is different from every
factor of
Thus,
b.
3 Vio.
Theorem
2.
615
This
is
obvious
let
are similar.
Hence
Va
and
V6
292
Suppose,
if
A COLLEGE ALGEBRA
possible, that
Va + Vo =
(1)
c,
(1)
where
c is rational.
and transposing,
(2)
which
c^
is
impossible
614, while
h is rational.
3,
i/"
616
Theorem
+ Vb = c + Vd,
b
Vb
and
Vd
are
surds, then a
and
^.
For, by hypothesis,
Vo
Vrt
Vt^ impossible unless Vo since otherwise Vi V(^ would be irrational, to c a, which is rational.
But
this
is
=c
a.
and
0,
615,
and equal
Hence
617
d and a
c.
V^ ^yY =
-\- ]/
We have 2 Vxy.
Hence if a + 2 VT* denote a given binomial surd, and we can find two 2>os'itioe rational numbers x and y such that
X
then
-\-
^=
a and xy
h,
\x +
+ 2 Vi and v^ Vy
When
Example
such numbers
x,
inspection.
1.
of 37
20
20 v3.
Reducing But
to the
Hence
V37 2.
37
300
Example
+ V5/6.
13
We
13
;
V30
6
have
2 V.30
"*"
\6
12
12
IRRATIONAL FUNCTIONS
Since 30
293
2
10
3 and 13
10
3,
we have Vi3 +
Hence
Note.
^ /l3 Xl
V12O + V36
12
:
V12
V30
-\
.
We may
618
(1)
By hypothesis
and
Vx + Vy = v +
2 Vft,
(2)
(3) (4)
But
Solvmg
(3), (4),
y x
= =
a.
y^
square.
Observe that these values are rational only when a^ 4 6 is a perfect Hence in this case only is the square root of a + 2 Vft a binomial surd.
EXERCISE XXXIX
Find square roots of the following.
1.9+ V56.
4.
2.
20
V96.
3.
32-2
Vl75.
1+2(a
-.
5.
7-3V5.
8.
6.
8V2 + 2V3O.
7.
+ Va2-62).
Vab
"V^17
12 V2.
10.
V 9 + 4 V4 + 2 V3.
619
The symbol
V 1
is
294
2.
A COLLEGE ALGEBRA
Symbols of the form
ai,
where a
is real,
imaginaries.
3.
hi,
where a and
+ hi = c + di,
then a
and
= d.
The sum, difference, product, or quotient of two complex numbers is itself a complex number (in special cases a real number or a pure imaginary) which may be found by applying The the ordinary rules of reckoning and the relation i^ = 1.
5.
any positive integral power of a complex number, to n factors. since by definition (a (a-\- bi) {a hi) biy
like is true of
+
3
i)
Example
1.
Add
5
5
3
We
have
2.
+ +
+ +
and 2
i.
(2
4
i
(5
2)
i.
(3
4)
i.
Example
Subtract 6
3
2
i
+
(6
from 3
i)
We have
Example
3.
+
3
2
i
+2 = (3 - 6) +
1
(2
2)
=-
3.
Multiply 2
(2
+
(1
by
i)
i.
We have
Example
4. (l 5.
i)
=2+3 +8 + = 2 + 3 + 8i i
12
^2
12
=-
10
11
1.
Expand
(1
i)2.
We
have
i)2
= l+2i +
2/i) i
i2
= l4-2i-l = = = +
+
2i.
Example
Find
(x
real values of x,
+ +
+
+
we have
2/)
(2/
2)
(X
4)
(X
(X
y)i.
Equating the
real
and the imaginary parts, 619, 4, -{y -\-2) = x-\-y and x + 4 = x - i/,
x
or, solving,
6,
4.
In 238-241 we have given a method for representing complex numbers by points called their grajths, and rules for
obtaining from the graphs of two complex numbers the graphs
IRRATIOXAL FUNCTIONS
of their
to Exs.
295
sum and
1, 3, 4.
product.
Two complex numbers like a + hi Conjugate imaginaries. and a bi, which differ only in the signs connecting their real and imaginary parts, are called conjugate imaginaries. The product of two conjugate imaginaries is a positive real number.
Thus,
(a
620
621
bi) {a
hi)
a^ hi)
hH^
(c
a2
Ifi.
Hence a
fraction, as (a
di),
may be reduced
its
to
622
the form of a complex number by multiplying both by the conjugate of its denominator.
Example.
Divide 5
terms
+ 7iby2
4i.
(' -^
(2
'^
We
have
'-^ =
2
')
i)
(^
-4 -18 +
20
(2
+ +
^') 4
z)
34i
9
10
17.
10
1 it follows that The powers of i. From the equation i^ 1 or 1, and the odd powers the even powers of i are either
623
either i or
Thus,
i^
i.
i"^- i
i; i^
i"
i^
i-
i"^
1
-^
and so
on.
any given value of n, divide n by 4. Then, according as the remainder is 0, 1, 2, 3, the value of t is 1, j, 1, i. Thus, i^* = (i*)^ = 1 i^ = i-* i = i; and so on.
find the value of
for
;
To
Even
The number
for
two square
roots
2)"'^i^
( 2 ly
root,
= (
= 4. We
2i and 2i;
and write
V 4 = 2i and
4 has
the
624
V 4 = 2
=
i.
Similarly the principal square root of any given negative a is \ai, that is, v a number Vai.
From
if
this definition of principal square root it follows that b are any two negative numbers, then and
625
V a V b =
For
^\/ab.
V a V 6 = Vai
Vbi
= i^y/aVb^ y/ah.
296
A COLLEGE ALGEBRA
negative numbers
Thus, while the product of the principal square roots of two a, b, is one of the square roots of their product ab, it is not, as in the case of real numbers, the principal
square root of this product.
When
mind
it is
Va Vb
of confusion
avoided
if
at the outset
we
replace every
symbol
V a
1.
by Vai.
Example
Simplify
We have
V^
2.
i)^
(V5)Tii3
+
(1
V^ by
+ V-1.
(2
We have
626
numbers.
V^)
+ ^^) =
/) {I
i)
=-
bi.
roots of negative
is 1
+
-I-
for
i)4
i^
i3
i^
=-
4.
627
As will be proved farther complex numbers. on, all roots of complex numbers are themselves complex numSquare roots of
bers.
We may
have
if
We
+ denote a given complex number in which Hence, and we can find two positive numbers x and y such that
X
then
h is positive,
-y = a,
will
bi.
(1)
and
Vxy =
b,
(2)
Vx + i V^
may
be a square root of a
bi,
and
vx i Vy
will
be a
square root of a
We
By
and
find such
numbers x and y as
follows.
(3)
hypothesis
Multiplying
(3)
by
(4),
(4)
(6)
IRRATIONAL FUNCTIONS
But, by
(1),
297
(6)
z-y = a.
(5)
Hence, solving
and
(6),
and
= a+ Va2 +
+
i.
62
And
Jfi
> a.
Example.
Here
=-
and Va2
Hence
Therefore
= {1
of
b'^
(4
(1
V5)2
9.
9)/2
5.
V_
Vs i =
+ Vs
i.
EXERCISE XL
298
A COLLEGE ALGEBRA
XIII.
QUADRATIC EQUATIONS
may
-{-
628
one unknown
letter, as x,
ax^
bx
c =^
i),
where
If,
a, h,
and
denote
6
known numbers.
as
may happen,
;
0,
the equation
is
called a pure
quadratic
\l b =^ 0, \i is
629
The
ax"^
roots found
by
^ are those particular values of x for which There are two c vanishes, 332. the polynomial ax"^ -\- bx of these roots.
bx
+c=
If
ax"^
the factors
bx
-{-
of ax"^
-{-
bx
-\-
roots
of
Q are also known, for they are the values of x If for which the factors of ax"^ + bx -\- c vanish, 253, 341. the factors are x a and x (3, the roots are a and ^.
c
Example
1.
We
have
{x
factor x
2 vanishes
3 and
2.
Example
2.
Solve
abx"^
{a^
b-)
+ (a^ -
b-)
0.
Factoring, by 443,
[ax
(a
x'^
(a
b)] [bx
Hence the
roots are
b)/a
0.
In particular, since
q =(x
the pure quadratic x^ q = are Vy and V^. Again, since ax^ + bx = (ax + b) x, the roots of a quadratic of the form ax^ -\- bx = are b /a and 0.
Thus, the roots of 4 x2
are
Vy)
(x
9 are
3/2 and -
3 /2
the roots of 2 x2
-x=
and
/2
^^
q and
0.
630
Conversely, to obtain the quadratic whose roots are two given numbers, as a and ji, we form the product {x a) (x ft)
and equate
this product to 0.
and
/3
is (x
+ 2)
(x
- 1/3) = 0,
QUADRATIC EQUATIONS
Example
1.
1.
299
x2
+ 2x-8 =
2.
2x2-7x +
0.
3.
(2x-l)(x-2) =
x2
2.
4.
(x-l)(x-3) = (2x-l)2.
3.1/4,0.
-{-
Example
1.-2/3,-3/2.
be factored
ax^
-{-
2.a,-a.
But ax^
in
bx
+ may
c
always
631
for, as
G
was shown
444,
bx
+
X
= a\
-b-^
^p _ 4 a(r\
;^
r
\
-b- v^rri^-i
+ bx-Jt- c =
-\-
(1)
bx -^ c
x
or, as
b+
-\/b-
- 4ac
za
and X
we
300
Example.
A COLLEGE ALGEBRA
Solve 3 + 56 x - 220 = 0. = 28, and substituting in (3) we - 28 V282 + 3.220 ^.
a;2
Here 6/2
have
10
x=
633
that
IS,
or
22.
given quadratic may also be solved by applying directly the process of completing the square, 444, as in the But since this method involves needless following example.
Any
it
to
reckoning,
it
is
not to be recommended
except
when the
formula of
Example.
Solve
3z2-6x +
0.
Transposing the
x2-2x = -2/3.
Completing the square of the
member,
x2-2x + 1 = 1/3.
Extracting the square root of both members,
a;
= V3/3,
whence x
(3
V3)/3.
634
The methods
equation which
just explained enable one to solve 2aij fractional yields a quadratic when cleared of fractions.
But
see 524-527.
1.
Example
Solve
x+1
x+2
X
x+3
x+4
11
+ lOx + = -5 V3
2x2
= 0.
Both of these values of x are roots of the given equation, for they cause none of its denominators to vanish.
Example ^
2.
Solve
X2-1
^^ + X2-X + ^^ = ^^ +
X2
0.
x(x2
Clearing of fractions by multiplying by the lowest common denominator 1), and simplifying, we obtain
3x2
+ 2x
0,
whence x
or
5/3.
cannot be a root of the given equation, since its first two denominators vanish when x = 1. Hence 5/3 is the only root of this
But
equation.
QUADRATIC EQUATIONS
EXERCISE XLI
Solve the following equations.
1.
301
3.
x2
2.
4x2
9x2
-4x
+
6X
=3.
4.
6.
= 0.
5.
+ 3x-4 = 0.
252
(2x-3)2 =
12x2
15x2
8x.
7.
9.
+ 9x-
0.
8.
8x2
-82x +
1
207
0.
lo.
12.
2) (X
11.
x2-3x- + V3 = 0.
(x
x2-(6 +
2i
= 0.
13.
2)2(x
7)
(X
3) (x
6).
14.
16.
-^ + ^+^2 =
2.
15.
^1 + 1=
X
"
x~l
I
17.
_i
2x +
l
4x-2
1-4x2
^
2x
=!
8
^3_
x-2
4
19
_^+i
x(x-2)
2x-2
L_ + JL = Q.
20.
X 21
4
a;
4{x
22.
+ + 2)(3x-l) +
3
2x
17x
4)
3(3x-l)(x +
6(x
+7 + 4)(x +
0.
^^
2)
23.
25.
+ {9a- l)x-3a = 0. + c (a - 6) X - a6 = 0.
6 acx
62)x2
24. 26.
x2
- 2ax + 4 ax
a2
62 =; o.
x2
4 a2
62
= q.
27.
28.
x2
(a2
+ -
a2(9 c2 2 (a2
62)
= o.
a2
62)x
62
0.
29.
0.
30.
(x-)^-(x-6)2 ^J_a6_^^^ o) (x 6) a2 62 (x
302
A COLLEGE ALGEBRA
EXERCISE XLH
1.
is
506.
is
2.
3.
481.
is 91.
4.
pairs
5.
587.
:
Find a number of two digits from the following data the product of the digits is 48, and if the digits be interchanged the number is
diminished by
6.
18.
The numerator
itself
exceeds
reciprocal by 24 / 35.
cattle
Having
lost 4 of
dealer bought a certain number of steers for $1260. them, he sold the rest for $10 a head more than they
entire transaction.
cost him,
How many
steers
did he buy ?
8. A man sold some goods for $48, and his gain per cent was equal to one half the cost of the goods in dollars. What was the cost of the goods ? 9.
If
is
inherits $25,000, but after a certain percentage has been 10. deducted for the inheritance tax and then a percentage for fees at a rate
A man
one greater than that of the inheritance tax, he receives only $22,800.
What
is
tl;ie
man bought a certain number of $50 shares for $4500 when they 11. were at a certain discount. Later he sold all but 10 of them for $5850 when the premium was three times the discount at which he bought them.
How many
12.
The
circumference of a hind
fore wheel
by 8 inches, and
in traveling 1
13. A square is surrounded by a border whose width lacks 1 inch of in being one fourth of the length of a side of the square, and whose area in inches square inches exceeds the length of the perimeter of the square by 64. Find the area of the square and that of the border.
QUADRATIC EQUATIONS
14.
303
is
The corners
whose side
in such a
way
What
is
the length of a
Having filled up the cask with water, he draws the taining 63 gallons. same quantity as before and then finds that only 28 gallons of pure wine remain in the cask. How many gallons did he draw each time ?
16. A man travels 50 miles by the train A, and then after a wait of 5 minutes returns by the train B, which runs 5 miles an hour faster than the train A. The entire journey occupies 2| hours. What are the rates
of the
17.
two trains
Had pedestrian walked 6 miles in a certain interval of time. the time been 1/2 hour less, the rate would have been 2 miles per hour
greater.
rate.
18. pedestrian walked 12 miles at a certain rate and then 6 miles Had he walked the entire farther at a rale 1/2 mile per hour greater. distance at the greater rate, his time would have been 20 minutes less. How long did it take him to walk the 18 miles ?
19.
From
two
straight roads
which cross
and B', set out simultaneously, on the one at right angles, two men, road at the rate of 3 miles per hour, B on the other at the rate of 4 miles
per hour.
20.
If
After
how many
on the roads just described, but at the rates of 2 and 3 miles per hour respectively, and A starts 2 hours before B, how
long after
21.
If
and
B walk
from a height of a
feet a
body be thrown
vertically
upward with
per second, its height at the end of t seconds The corresponding formula is given by the formula h = a -\- ht \(i f^. when the body is throvra vertically downward i?. h a bt \Q t-.
an
(1)
If a
body be thrown
vertically
upward from
initial velocity of
of 16 feet ?
(2)
32 feet per second, when will it be at a height of 7 feet? Will it ever reach a height of 17 feet ?
is
body
with an
initial velocity of
height of 36 feet ? (3) If a body be dropped from a height of 36 feet, the ground ?
when
will
it
reach
304
A COLLEGE ALGEBRA
XIV.
635
denote
The
bx
-y/b'^
>
ax"^
-\-
bx
-\-
c =^ 0.
When
a,
the coefficients
a, b, c are real,
siffii
p
1. 2.
is
indicated by the
of the discriminant.
Thus
When b^ 4 ac is jJositioe, the roots are real and distinct. When b^ 4 ac is 0, the roots are real and equal. When
b'^
3.
4 ac
is
naries.
It should also be observed that
1.
2.
When ^^ 4 ac = 0, then ax'^ + ia; + c is a perfect square. When a is positive and c is negative, the roots are always
4 ac
is
then positive.
If a,
b, c
and
only when,
i^
4 ac
Show
a perfect square.
a;^
Example
1.
6x +
6)-
10
1
=
10
are imaginary.
They
4 ac
= (
4.
Example
equal
?
+ 3x +
We
must have
3.
7n
0,
that
is,
m=
9/8.
Example
0,
If possible, factor y^
xij
2x^
llx
\2.
it
Arranging the polynomial according to powers of y and equating we have y'^ + {x + l)y - (2x2 - 11 x + 12) = 0.
Solvuig,
.
.
to
y
y
(X
5^
1)
V9x2-42x +
49
.
that
is,
4,
or y
2X
-\-
S.
305
Hence, 681, y'^ + xi/ - 2x'^ + U x + y - 12 = {y - x + i) {y + 2 x ~ S). Observe that the factorization is possible only because the radicand 9 x^ _ 42 X + 49 is a perfect square.
If
a and
(3
denote
636
-i-
bx
-\-
-\-
c
-\-
0,
we
have,
631,
ax^
bx
a (x
a) (x
/3).
Dividing both members of tliis identity by a and carrying out the multiplication in the second member, we have
x^
-i
X =
b
c
-\
x^
(a +
B)x
aB.
Since this
in its
is an identity, the coefficients of two members are equal, 264, that is,
like
powers of x
a
This
;8
/a and a^
/ a.
may also be proved by adding and multiplying the values of a and ft given in 631. Therefore, since a, /3 are
the roots oi x^
-{-
bx / a
-\-
/a
0,
+
to
px
of X
tvith its
siff7i
changed
is
equal
the
sum of
and
sum
Thus, in the quadratic 6x- + x = 2, that is, x^ + x/6 of the roots is 1/6, and their product is 1/3.
1/3 =
0,
the
Example
1.
Solve 9x2
lOx
+
,
0.
is 1
for 9
10
+ 1=0.
is
1/9
-^ 1
Example 2. 3x2 + 8x + 5
Let a and
0.
/3
Then
a+
/3
x2_ (3a +
9a/3
= x24-8x +
15
= 0.
637
of the roots. The expressions a + /8 and symmetric functions of the roots a, jS, 540. All other rational symmetric functions of a and /3 can be expressed
Symmetric functions
aft are
306
A COLLEGE ALGEBRA
a ^ and a^, and. rationally in terms of these two functions, therefore rationally in terms of the coefficients of the equation.
For every such function can be reduced to the form of an integral symIf an metric function or to that of a quotient of two such functions. term of the type kaP^P + ', it must integral symmetric function contaijis a
But 5427>K^herefore kaPpP {ai + ^i). readily be showrHixsuccessive applications of the binomial theorem that ai + /39 can be expresSed4n terms of powers
also contain the
aP^p
(a^)P,
and
may
ot
-\-
p and
a/3.
Thus, since {a
Similarly
we
+ ^Y=a^ + 2 n-^+/32, we have a2 + /S^ = (a + /3)2 - 2 a/3. + /33 = (a + /3)3 -Zap{a + /3). The roots of x- + px + g = being a, express 1 / a + 1 /
lind a^
jS,
/3
in
terms oi
l//3
a/33
p and
q.
Wehave l/a +
and
a^iS
638
Infinite roots.
Suppose
coefficients of ax^
+ Z'x + c =
are variables.
We
can then
show
that
co
if
;
approach
as limit, one of the roots will a approaches and if both a and h (but not c) approach 0, both
oc.
a=
+ V62 2a
4 ac
1
/3
-b - V62 -
4 ac
2a a hy b
\''b-
4 ac
and both
2c
+ V6-
--,
4 ac.
We
b
obtain
a =
b
2c
/3
+ V&2 _
if
4 ac
V62 _
b.
4 ac
By
and
203, 205,
Therefore
if if
co,
/3
==
00.
It is
customary
infinite
when a and b
(but not c)
be a function of x^
278.
It
639
may happen
value, m,
that, as
and then begin to decrease, or that y will decrease and then begin to increase. We then a maximum value of y and m' a minimum, value.
2/
Thus,
being
(x
if
1)2
4 has a
minimum
less
value
when x =
1
1,
this value
4.
For
than
and
increase, (x
1)^
and then
(x
increase.
Similarly
?/
1)^
has a
maximum
value, 4,
c
when x =
1.
ax'^
ix
640
maximum
or a
minimum
value,
which may be
of
?/
x^
+6x
7.
By completing
Example
shall
2.
Hence when x =
has a
16.
Let 2 a denote the length of the given segment, x and 2 a of the parts, y the area of their rectangle.
x the lengths
Then
(2
x)
2 ax
x^
a^
(a
x)^.
Hence y has a maximum value when x = a, that is, when the given ment is bisected and the rectangle is a square whose area is a"^.
seg-
The maximum and minimum values of quadratic trinomials and of certain more complex functions may also be found by
the following method.
Example.
2/
641
Find the
values,
if
any,
of
= (4x2-2)/(4x-3).
Clearing of fractions and solving for
we have
1) (y
^ *~ y Vy^
2
3y
~ ^
V(y 2
2)
By hypothesis, x is restricted to real values. Hence y can only take values for which the radicand (y 1) (y 2) is positive (or 0), that is, the value 1 and lesser values and the value 2 and greater values.
It
is
maximum and
2 a
minimum
value
of y.
308
A COLLEGE ALGEBRA
For observe that as y increases to 1, the two values of x, namely 3 2/ + 2)/2 and {y + V?/2 _ -^y ^ 2)/2, respectively increase and decrease to 1/2. Hence, conversely, as x increases through 1/2, y first increases to 1 and then decreases.
{y
- V2/2 _
642
Given y
= ax^ +
we
Sx
c,
where a
is
positive.
By completing
the square,
obtain
mum
As X
+ 00
b^)/4:a
x2
h'^)/4:a. jT 00 to + co,
3
'
= b /2 a, this minifrom
?/
00.
Thus,
(X
1)2
4.
As X
increases from
00 to
00 to
+00,?/
first
decreases from
from 4 to Moreover
that
is,
00.
when ?/ = when x = 1 or
x^
3.
?/
0,
1,
is
posi-
it
3,
when
it
When
(-1, 0)
0,
1,
2,3,4,
5,...
we have
?/= 12,
5,
0,-3,-4,-3,0,5,12,....
?/
,
(1-41 '~
^
''
a curve through them, as in 389. Observe that to the zero values of y there
correspond the points where the graph cuts the x-axis, and that to the mininmm value of y there corresponds the lowermost point of the graph, which is also a turning point of this curve.
EXERCISE XLHI
1.
(?n-|-
2)
x2-
mx + 1=0 equal?
when
2.
What
?n
m) x^ + 3 mx
-2 =
m=-
1 ?
when
309
If possible, factor
3x-
+ 5xy can xq
2yy"^
4.
5.
- 5x + 4y - 2. + mx + by 6 he
a and a and
/3,
factored
The
roots of x^
+ px +
a
in
being
express (a
/3)2,
a*
/3*,
and a/^
(3/
terms of
p and
g.
6.
The
roots of 2
x'^
3x
+
2
being
p, find the
values of
a/^'^
7.
i3/a:2
and a^^
a/S^.
The
roots of x-
+x+
jS
;
being
;
a,
l/a, 1/^3
2 a, 2 P;
a and j3, a+
1.
Find the
3.
a;2_8x +
x/{x2
9.
2x2
-X + 4.
+ 4x-x2.
+
1)/
4.
l).
l/x + l/(l-x).
(x
(2x2-1).
;
in
a given circle
man who is in a boat 2 miles from the nearest point on the 10. shore wishes to reach as quickly as he can a point on the shore distant If he can row 4 miles an hour and 6 miles from that nearest point.
walk 5 miles an hour, toward what point should he row
11.
?
What
height will a
body reach
if
thrown
vertically
See
p. 303,
Ex. 21.
643
we can resolve A into factors of the first or second degrees, we can find all the roots of ^ = by equating the several factors of A to zero and solving the resultis equivalent For ii A = BC , then ^ = ing equations. toB = 0, C = 0, , jointly, 341.
in the
form A
0.
If
Example
1.
Solve
x"*
By
436,
X*
+ +
x2
x2
+ +
= =
0.
(x2
1) (x2
1).
310
Hence
x*
A COLLEGE ALGEBRA
+
,
x^
=
.
is
x2
+X+
and
x2
0.
,.
= -l-i-iVS'liiVs or
5 x^
Example
2.
Solve x*
x^
12
0,
451,
we
find that
x3 x^
Hence
x*
_ 5x2 - 7x + 12 = (X - 1) (X - 3) (a;2 + 3x + 4). 5 x^ 7 x + 12 = is equivalent to the three equations x-l = 0, x-3 = 0, and x^ + 3 x + 4 = 0,
1, 3,
and ( 3
V7)/2.
Example
^
1.
3.
6x3-
11x2
+ 8x-2 =
0.
2.
X*
5x3
0,
x2
+ llx +
0.
644
function of
for
=
is
-\-
bu
-\-
when
solved
this equation
equations u
a and u
au^
-{-
-\-
/?,
for, 631,
a(u ft). a) (ti Hence to solve au^ + bu + c =^ for x, we have only the two equations u = a and u = (3 for x.
Example
Hence
1.
bii
to solve
Solve 3x<
x2,
10x2
Solving for
x2
= 0. = 2/3
or
4.
x
2.
= V(3/3
or
2i.
Example
Solve x^
x',
Multiplying by
Solving for x^
Hence
Example
3.
Solve (x2
+ +
5)
= =
6.
We
may reduce
form
3x)
14
0.
+
+
1,
9 (x2
Solving for x2
3 x,
x2
we
3X
=2,
2,
and
x2
+
i
3x
=-
7,
and { 3
Vl9)/2.
311
Solve (x
1) (x
2) {x
3) (x
4)
120.
multiplying together the first and fourth factors, and the second and third, we reduce the equation to the form
By
(x2
+ 5x + +
5
4) (x-
+ 5x +
6)
120,
3.
in the
same way
Example
5.
Solve x*
10 x^
31 x^
first
30 x
+
5
=
0.
0.
By completing
Solving for x^
x)'^
6 (x2
5x,
x2
we obtain
Example
6.
Solve 8
^ -
x2
+
1
-=:
11
= 0.
x2
2X
tiply
Observing that the second fraction is the reciprocal of the first, we mulboth members of the equation by the first fraction, thus obtaining
\
\x2-l/
1),
,
we
1
+ 2x =
1
^x2 and
x2
+ 2x -1
3
-,
1/2
and 3, 1/5.
All the values of x thus found are roots of the given equation since they cause none of its denominators to vanish.
Example
1.
7.
3x*
(X
-29x2 +
a) (X
0.
2.
x<
-6x3 +
a*.
8x2
+ 3x =
2.
3.
2 a) (x
6)
3 a) (x
4 a)
24
4.
(4 x2
2 x)/ (x2
(x2
0)/
(2 x2
x)
0.
These are equations which remain Reciprocal equations. unchanged when we replace x by l/x and clear of fractions.
If
645
we arrange
x,
powers of
the
the terms of such an equation in descending first and last coefficients will be the same,
and next
to last,
and so on
or each of these
contrary signs.
312
Thus,
A COLLEGE ALGEBRA
2x*
and
jc-2x*
to the quadratic
Reciprocal equations of the fourth degree may be reduced form and solved as follows.
1.
Example
Solve 2x*
3x3
4x2
- 3x +
0.
Grouping the terms which have like reduce the given equation to the form
coeflBcients
and dividing by
x^,
we
2(.
Since x^
4 = 0. i)-3(x + l) +
2,
/x-
(x
+ 1 /x)- +
we may reduce
form
2(x
Solving for X
^)-3(x + ^)
1
0.
and x
(3
X
i,
13 =
2
-,
i,
and
V7)/4.
Every reciprocal equation of odd degree has the root 1 or 1 and if the corresponding factor a; 1 or x 4- i he sepaHence rated, the "depressed" equation will also be reciprocal reciprocal equations of the third and Jifth degrees can be solved
;
by aid of quadratics.
Example
2.
Solve 2 x'
3 x2
3x
0.
to
2 (x^
1)
3 (x-
4- x)
0.
by x
1, it is
equivalent
and
2-x2
5x
0,
1, and
Solve x^
2,
1/2.
Example
3.
Grouping terms,
Dividing by x
whose
roots
1, we find that this equation is equivalent - 1 = and x* - 4 x* + 5x2 - 4 x + 1 = 0, are 1, and (1 i V3)/2, (3 V6)/2.
X
313
4.
a;3-2a;2
+ 23.
=0.
2.
x* X^
4x^
x5
x-*
X^
+ 5x^ - 4x + = 0.
0.
Binomial equations. Tliis name is given to equations of the x" + a = 0. They can be solved by methods already a can be resolved into factors of the first or given when x"
646
form
second degrees.
Example
Since x^
to the
1.
1
Solve x^
0.
1),
(x
1) (x2
+x+ = =
32
the equation x^
x^
is
equivalent
two equations
X
and
1
Solving,
X
2.
or
0.
+ x + 1 = 0. {- 1 i V3)/2.
vfe obtain y^
Example
Solve x^
32
=
is
From x5 By 438,
=
1
0,
643, y^
2/
by setting x
1
= V32 y -2y,
?/3
0.
and
1,
2/4
+
j
2/2
2/
+ 1 = 0,
2
Solving,
2/
(-
V5 + Vio
z
V5)/4,
or
(-1
Hence
x
2y=z2, (-
Vs + i ViO
i
V5)/2,
or
(-1
By
V5- ViO
2/"
2 V5)/2.
can
0.
Example
1.
3.
x3
X*
+1=
0.
3.
x6
0.
These examples
n nth
roots.
illustrate the
theorem
1 is
647
any number which satisfies the equation x^ = l\ and in Ex. 1, we found three such numbers, namely 1, ( 1 + i V3)/2, and ( 1 t V3)/2. Irrational equations. If asked to solve an irrational equation,
Thus, a cube root of
But, as will be ordinarily begin by rationalizing it, 609. illustrated below, certain equations admit of a simpler treat-
648
we
ment than
this.
Whatever method
is
used, care
must be taken
letter
for the
unknown
before
314
Example
1.
A COLLEGE ALGEBRA
Solve
Transposing,
Testing these values of x in the given equation, we find that 2 but that 7/6 is not a root.
is
a root
We may also rationalize the given equation by the method of G03. We thus discover that 4 (6x'^ 19x + 14) is identically equal to (V2x-3 - V5x - 6 + V3x- 5)(V2x - 3 + Vox - 6 - Vs x - 5). (V2X-3 + Vsx - 6 + V3x- 5)(V2x-3 - Vsx -6 - VSx -5).
which the product 6x2 _ 19a; + 14 right vanishes for one of these Hence there is no value of x values, 2; the second, for the other, 7/6. for which the third or fourth factor cau vanisli.
Tliere are but
two values
first
of x for
can vanish.
The
factor on the
Example
2.
Solve
V4x +
+ V] 2 x +
= V24 x +
10.
form of a quadratic with respect to this radical. then begin by solving for the radical.
Example
3.
We
Solve 2 x2
6x
Vx^
3x
-5 = 0.
Observing that the x terms outside the radical are twice those under the radical, we are led to write the equation in the form
The second
Talue.
to the
convention made in
Va
Squaring the
first
equation,
we
5
obtain
0,
x2_3x-l =
whose roots are
Testing 5 and are roots.
and
2.
we
them
Example
4.
Solve 2 x2
14 x
Vx2
10
+ 18=0.
315
Sometimes an equation may be reduced to a form in which, both members are perfect squares or one member is a perfect
square and the other
is
a constant.
x
Example
5.
Solve ix-
+ 2x Vsx^+x =
X
9.
We may
The
first
2X
Vs x'^ +
x2
9.
member is a perfect square, and both members, we obtain the two equations
Va x2 +
= =
3,
1,
and
V3 x- +
9/2, or (5
we
roots of the
Sometimes
all
the terms,
common
irrational factor.
6.
Example
Here the
Solve
Vx^ -
ax
first
also x
2 a.
we
equivalent to
2 a.
Vx 2a =
we
and
equation.
Example
7.
Solve
V3x2- 5x-
12
- V2x2-llx +
15
3.
If one or more of the terms of the equation are fractions with irrational denominators, it is often best to rationalize these denominators at the outset.
Example
8.
Solve
we have
3X
Solving,
= 2 X - 6. = 3 or 4.
of the given equation.
1
Testing,
we
9.
Example
Solve
Vx -
- Vx +
1)
Vx -
+ Vx +
1)
- 3.
316
A COLLEGE ALGEBRA
EXERCISE XLIV
4a;*
(x2
-17x2 +
18
9)
= =
0.
2.
4.
3x?(2x2
x<
4x^
=
x2
7.
3.
5.
4) (x2
7 x2.
-x-3)(3x2 + x -2)2=0
2 x3
+ -
x3
x2
3 X
0.
6.
0.
2X
= 0.
7.
8. 9.
7)
12
0.
12 x3
4 X*
X*
X*
4 x3
+ 33 x2 + 18 X - 28 = - x2 - X - 2 = 0.
2 x2
10. 11.
12. 13.
2 x3
x3
x5
x5(1
15.
17. 18.
+ 36 x3 = 0. + x)3 = (1 - x)3.
11 X*
+ -
2 X2
X + 1 = + 1 = 0. - 36 x2 +
2
0.
11 X
-1=
0.
243
14.
(2x(X-
1)8
= 1.
0.
16.
2)*
-81 =
(a (a X2
+ x)3 + - x)* + 3X +
(6
(6
1
+ -
x)3
x)*
(a (a
=
x2
+ -
6
?))
+
(a
2 x)3.
2 x).
^^
4x2
20.
+ 6x-l
x2
_3
4x2
+ 6x-l _^^^^ + 3x + l
x2 + - = a2 + -l.
a2
21. 22.
3x2
4 x2
-2x-5V3x2-2x + 3 + 9 = 0.
24.
25.
27.
^
30.
- Vx2 + 3x = 0. 6 /x - 2
28.
v^ -6Vi + 6-v^ =
0.
'^^-"^ = x-3. Vx - + Vx +
1 1
31.
V5x2 - 6x +
- V5x2 + 9x -
= 6x -1.
S17
^
+
0.
33.
^ + ^^^2^ =
3.
2.
34.
35. 36.
(X
+ a)5 + (X + 6)' + (X + - 1) (X - 2) (X - 3) =
a)2
c)'
.
0.
(X
4 (X
rt)
Vx =
a-
4 a
\^.
XVT.
A PAIR OF EQUATIONS IN X, Y, ONE OF THE FIRST DEGREE, THE OTHER OF THE SECOND
649
f(x,
cf)
(x,
= y) = y)
ax^
a'x
+ bxi/ + f/ + + b'y -f =
c'
dx
+ ei/ +/= 0,
may
Example.
x2
2x
+
X
2 y
2/
+ -
4
7
?/
= 0,
= 0. = 2x-7. = 0. = 13/3 or 3. = 5/3 or - 1.
(1)
(2)
From
(2),
(3) (4)
(5)
Substituting
Solvins: (4),
(3) in (1),
3 x^
22 x
39
X
(3), (1), (2)
Substituting in
(6)
The
solutions of
13/3,
= 5/3;
(2);
3,
= -l.
(7)
(2)
to (4),
(2);
(4),
(2)
to {bi
(5),
(2)
to (7).
We may
5/3;
3,
1.
y.
Care
tions.
Ordinarily such a pair of equations will have tivo finite soluBut if the group of first-degree terms in ^ (x, y), namely
650
318
A COLLEGE ALGEBRA
a^x Vy, is a factor of the group of second-degree terms in f{x, y), namely ax^ bxy + cy-, while <^ {x, y) itself is not a factor of f(x, y), there will be only one finite solution or no
such solution.
be infinitely
And
Solve
if ^ (x,
y)
is
a factor of f(x,
y),
there will
many
solutions.
y^
Example
1.
x^
2x
+ 2y +
y
= 0,
0.
0.
(1)
Eliminating
If wt2
y,
0, (3)
(m2
mx = - 1) x2 + 2 (m + 1) x + 4 =
finite roots,
(2) (3)
5^
has two
and
But ify- mx is a factor of y- - x-, that and (3) does not have two finite roots.
Thus,
if
(1), (2)
two
1,
finite solutions.
is, if ??i
then
in^
-1=0
=
0,
m = 1,
(3)
reduces to x
=
if
0,
And
m=-
finite root,
(3)
reduces to 4
both roots being infinite, 638. Hence if (2) has the form y - x = 0, the pair (1), (2) has but one finite And if (2) has the form y + x = 0, the solution, the other being infinite. pair (1), (2) has no finite solution, both solutions being infinite.
finite root,
which has no
Example
2.
Solve
y"^
Eliminating
y,
a;2
+ 2x + 2y = 0, y + x = 0. - x^ + 2 x - 2 x = 0.
x"^
(1)
(2)
(3)
But (3) is an identity and is satisfied by every value of x. Hence every pair of numbers x = a, y = - a is a solution of (1), (2). The reason for this result is that y + x is a factor ot y'^ - x- + 2x + 651
2y.
When
A, B,
AB ^
and
0,
This principle and 649 enable us to solve two integral equations when(x, y)= /(x, y) = 0, ever /(x, ?/) can be resolved into factors of the first or second
0,
5=
C= 0, C =
are integral functions, the pair of equations is equivalent to the two pairs .4=0, C
371.
cf>
degrees and
Example.
<^ (x,
Solve x^
+ xy- is
Hx
0,
0.
(1)
{2x-y){x + y-l) =
This pair of equations
X
(2)
X
a;2
= =
0, 0, 0,
2/2-5
x2
-f- ?/2
= =
0,
= 0, - 1 = 0, 2x-2/ = 0, - 1 = 0. X -h
2X
2/
(3)
'
(4) (5)
2/
(6)
319
be
0,0;
0, 1;
1,2; -1,
-2;
2,
the solutions of
(1), (2) to
pair of integral equations in x, y can be solved by means of quadratics only when it has one of the forms described in 649, 651 or when an equivalent pair which has one of these
652
it.
x + \ = 0^ y = x"^, Thus, the pair of equations of the second degree, ycannot be solved by quadratics. For there is no simpler method of x + 1 = 0, an equasolving this pair than to eliminate ?/, which gives x* tion of the fourth degree which cannot be solved by quadratics.
The preceding sections illustrate the truth of the following important theorem
:
653
A pair
of
y)
0, <^ (x,
y)
= 0,
whose
degrees are
and n
respectively,
xy-
has
x
mn
-
solutions.
-bx =
0, (2
y) (x
1)
has 3
2,
See
381 also.
It should be added, however, that if the groups of terms of highest degree in f{x, y) and <^(a;, y), but not f{x, y) and ^ {x, //) themselves, have a common factor, there are less than
654
mn finite
which
is
solutions.
common
<^ (.r,
to the groups of
Thus, for every factor of the first degree terms of highest degree in
is
/(r, y) and
y) there
at least
one
infinite solution
for
every such factor which is also common to the groups of terms of next highest degree there are at least two infinite solutions;
and so
on.
If /(x, y)
and
^(.r, y)
themselves have a
common
many
solutions.
- y2 _ i = o (2) cannot Thus, the pair x^ -xy"^ + xy - y"^ -y = (1), x2 have more than three finite solutions for there are 3 2, or 6, solutions all told, and at least one of these is infinite since x -\- y is a common factor of the groups of terms of highest degree in (1) and (2), namely X" xy'^ and x^ ?/-, and at least two others are infinite since x y is a common factor of the groups of highest and next higliest degree in (1), (2),
namely
x'
xy^,
x"^
y"^
and xy
y'^,
0{x
y).
320
A COLLEGE ALGEBRA
EXERCISE XLV
i2x-Sy =
4.
'
b.
L.3x-5?/
3
'
2.
\3x + 6y =
l.
0,
5y"~
5.
-Sx-7y = 0,
0.
1 x + 3y =
fx^
-i
r2x^-xy-3y = 0, [jx-6y-i =
f2x + 3y =
8.
-J
0,
37,
rx^
'
l{3x
_^^
jj
(X
13.
Determine
4iX
0,
= mx
14.
shall be equal.
Determine
and
.cy
7/2 _|_ a;
0,
?/
vix
shall be infinite.
15.
By
the
method
2/2
of 650, Ex. 2,
show that 2x
7/
+ 4isa
factor
of 2 x2
16.
xy
10 X
12.
Show
that the
finite solutions,
+ =
y
1
,
=
x^y
y2
2 has not
more
than four
finite solutions.
When both equations are linear with respect to some pair of functions of x and y. begin by solving the equations for this pair of functions by tlie methods of 374-376.
We
Example!.
Solve
2x2
82/2
=_
53,
(1)
^2)
3x2
y2
=111.
321
we
obtain x"
whence,
= 25, = 5,
is
?/2
= 36, = G. =
6,
By
y
6;
x= 5,
2.
?/
6; x
5,
?/=
Hence the
solutions of
(1), (2)
are 5, 6
= 5, = 6. 5, 6 5, 6
?/
;
5,-6.
Example
.5x2
3/7/2=
120.
When
possible
when
This is always one of the equations can be factored. the equation in question has the form.
ax^
656
bx]/
cj/^
=i=
0,
and, in general,
when
it is
au-
+
a;
ft -f- c
=
2
0,
x, y.
Solve x2
y2
_
5
^^
_
G
x2
xy
2/2
= ^
?/,
0,
0.
(1) (2)
Factoring
(2)
by solving
for x in terms of
x
or
X
Solving
(1),
= =
all
22/,
(3) (4)
(1), (2),
?/.
namely
4,
2;
2.
-2/5, - 1/5;
Solve 2 x2
(3)
and
(1),
(4),
we obtain
;
the solutions of
3, 1
-3/5, 2 y2
_|_
1/5.
3y
2/2
Example
+
:
4 xy
3^
_
+ -
2 5
3 x2
32
= =
= =
0,
0.
0.
(1)
(2)
We may
Solving,
write
(1)
thus
2 (x
2/)2
3 (x
y)
2
?/
x
x
or
+ +
]/2,
2.
(3)
2/
=-
(4)
(1), (2),
Solving
(2),
(3)
and
(2),
(4),
;
we obtain
all
the solutions of
namely
1,
/2
3/29, 23/58
3, 1
- 41 /29, x
X
1
17/29.
Example
3.
+ X2/ - 6 = 0, U2_5a; + 6 = 0.
r
x2
x'-'
x-y _ 26
x
2.
-y
'
2/2
-2x2 =
1.
322
657
A COLLEGE ALGEBRA
the given equations
When
may
be combined
by
addition or sub-
an equation which can be factored. This is always possible when both the given equations are of the
traction so as to yield
form ax^
+
1.
^xy
cy^
d.
Example
Solve
6 -c2
xy
?/2
5 x2
a;y
y2
= =
55, 49.
(1)
(2)
We combine
Multiply
Multiply Subtract
Solve
(5)
(1)
and
7,
(2)
(1) (2)
(4)
by
42 x2
by 8, from
40 x2
(3),
for X,
or
(6) (7)
(1), (2),
Solvmg
(2), (6)
and
(2),
(7),
;
we
obtain
2
all
the solutions of
namely 3 V35/IO,
V35/5
V2T/3, T V21/3.
And, in general, we obtain an equation which can be factored when the given equations are of the second degree, and can be combined by addition or subtraction so as to eliminate (1) all
terms of the second degree (2) all terms except those of the second degree (3) all terms which involve x (or y) or (4) all
; ;
?/).
Solve 2 x2
4 xy
2 x
?/
0, 0.
(1)
(2)
3x2
Gx?y
-x +
3y =
(1)
Here we can eliminate all terms of the second degree by multiplying by 3, and (2) by 2, and subtracting. We thus obtain
4X
9y
0.
(3)
all
Solving
(2), (3),
we
obtain
3. 2;
2,
the
See
654,
0,
0.
Example
Here
3.
Solve
- 3 xy + 2 ?/2 + 4 x 4- 3 y _ 1 = 2 x2 xy + y2 + 8 X + 2 y - 3 =
x2
r>
(1) (2)
the terms which involve x can be eliminated by multiplying We thus obtain (1) by 2 and then subtracting (2).
all
3y2
Solving
(1), (3),
+ 4y +
7
0.
(3)
(1), (2),
we
obtain
all
namely
1.
1/3,
- l/3j -
lG/3,
1/3; (-
V57)/2,
323
also.
0,
0.
Solve
x'^
xy
+
2/2
2y^
xy
Multiply
(2)
2/
+ 3x = +1 =
(1)
(2)
by 2 and add
{x
to (1).
We
obtain
(3)
(4)
(5)
(1),
2y)^
3{x
Solving
or
(3),
2y) + 2 x + 2y
x
= 0. = -l, + 2y = -2.
all
Solving
(2),
(4)
and
1
(2),
;
(5),
we
3
obtain
(1
the solutions of
(2),
namely
- 3
5.
V2,
T ^2
^5,
^^5) /2.
Example
r2x'^
^
1x2
+ xy + 5y=0, + 2/2 10 = 0.
_,.
rx-^+y"--\S=0,
'
2/
Xxy + y - x = -l.
658
When
factored.
we
obtained by eliminating x or y can be pair of equations of the second degree can eliminate x or ij by the following method. The result-
the
equation
From any
ing equation will ordinarily be of the fourth degree and not solvable by means of quadratics. But if we can resolve it
into factors of the first or second degrees, so obtain the solutions of the given pair.
we can
^
solve
it
and
Example
1.
Solve
lOx^
-^
5y^
27 x
x2
y2
~ iy + - 3X 5
5
2/
0,
0.
(1)
(2)
First eliminate y^
by multiplying
(2)
by
result
from
(1).
We
obtain
2/
5x2-12x +
Substituting
(3) in (2),
0,
or y
=- 5x2 + 12x-5.
6
(3)
5x*
Factoring, by 45l,
(X
0.
(4)
x
2,
2) (5 x2
(9
9X
3)
= 0.
(5)
(6)
Solving
(5),
643,
?/
Substituting
(6) in (3),
= -
1, 2, 1,
V21)/10.
(7 dz
3 V2r)/10.
(7)
(1), (2).
The
(6), (7)
Example
Solve
x2
x?/
2x2
X2/
+ 2 2/2 + 3 x - 3 y = - 2/2 + I - 2 + 3 =
2/
0.
324
A COLLEGE ALGEBRA
EXERCISE XLVI
x^ +
7x2
2/2
j/2
Sy"-
-2
X2/
2/2
= Sl, = 10.
6
2 5
r36/x2
1^1/2/2
1/2/2
_4/j.2
i/2
= 18, = 8.
+ _
-
+ -
y
x2
= 0, = 0. =
0,
2/2
fx2+
13x2
17
-3x + 22/-39 = 0, = 0. xy + 10
2/2
14 X'
.2
4 xy
4X
2/
3.
7.
2/
^^
+ 3x2/ -42/2 = 25, + 24x2/ -31 2/2 = 200. -x2-3xy + 3 = x22/2, = 12x2^2. .7x2 -10x2/ + 4
2x2
15x2
2/2 2/2
(rx(x
'
3?/)
1x2 -52/2
.x2
= 18, = 4.
2/^ ?/2
^^
'
1x2
rx'-^
+ -
X2/
X2/
+ +
= =
38,
14.
^^
x2-x2/
X2/
2/2
21(x-2/),
^g
20.
0.
sometimes advantageous
;
them by multiplication
or division
then be taken not to introduce extraneous solutions nor to lose actual ones (see 362, 342).
660
form
AB = CD
(1),
B=D
(2),
where
A, B, C,
by
= CD, B = D which is evidently equivalent to the two pairs A = C, B = D and D = 0, ^ = 0. We may obtain the pair A = C, B = D by dividing
D
each
if
x, y,
we may
replace
member of (1) by the corresponding member of (2); but we then merely solve this pair A C, B = D,wg lose some
325
of the solutions of (1), (2), except, of course, when either B or i) is a constant, so that the pair B has no solution. 0, D
Example
Dividing
Solving
2, 1
1.
Solve
x*
(1)
by
(2),
= x^ + xy + y'^ = x^ - xy + y^ = +
x-y-
y*
21,
7.
(1)
(2)
(3)
(1), (2),
3.
(2), (3),
we
2
;
obtain
all
2,
2.
1,
1,
namely
2.
See
x^
654.
2/^
Example
Dividing
Solve
x^
xy
(1)
by
(2)
(2),
= - 3 (x + = X + 1. x-y = -Zy.
1) y,
(1)
(2)
y'^
(3)
finite solutions as
the pair
(2), (3)
and
+ 1 = (5) jointly. And the solutions of - 2/3, 1/3; - 1, (1 i V3)/2. (2), (3) and (4), (5) are 2,-1; = x, Example 3. Solve (x + y)" (1) X2_^2 = _6y. (2) + y) / {x - y) = - x/Q y. Dividing (1) by (2), (3) Clearing of fractions, x^ + 5 xy + y^ = o. (4) - 2 9/4, - 3/4. The pair (2), (4) has the four solutions 0, 0, 4, +
y-
(4),
l^x
(4)
(1), (2) is
reversible
when
v^alues 0, 0.
3/4, but not when they have the Hence this reckoning only proves that 4,-2 and 9/4,
4,-2
(1), (2), 362. is a solution of (1), (2); but it obvious by inspection that 0, should be counted only once as a solution, not twice as in the case of Tliis follows from the fact that (1), (2) can have but three finite (2), (4).
-3/4
are solutions of
It is
solutions, 654.
It
may
x
also
tution y
tx (5).
We
yield
obtain
And
0,
be shown thus In (1), (2) make the substi= x (1'), (1 - f^)x'^ = - 6tx (2'). (1 + ty^x"^ y = once and but once.
:
EXERCISE XLVn
rx3-r/'
1
'
63,
\x-y = S.
+ x22/2 + 2/4 = 931, 1x2 + x?/ + ?/2 = 49, f{x + y)2{x-y) = 3xy+6y, la;*--l/2 = x + 2.
rx*
2.
+
-
2/)(x2-2?/2)=_70,
2/)
\
^
(X
(x2
2/2)
14.
rx^
- 3xy +
-2/2-_
2y^
5y.
6z,
1x2
326
A COLLEGE ALGEBRA
SYMMETRIC PAIRS OF EQUATIONS
661
A pair of equations in x, y is said to be symmetric remains unchanged when x and y are interchanged.
Thus, the following pairs,
(a)
if it
and
(6),
are symmetric.
^^^
'
'
,^ r2x2
22/2
a;j/
IxV ^.
+ 3x + 3y = + 1 = 0.
of
0,
rx2
l2/2
= 2x + 32/, = 2 y + 3a;.
which
like (a) in
the individual equations remain unchanged when x and y are interchanged, and those like ih) in which the two equations change places when x and y are interchanged.
662
Symmetric pairs
metric
equations
The simplest
+ =
?/
a,
xy
b.
solved as in
is
method.
Example.
Solve
Square
(1),
x^
Multiply Subtract
(2)
(4)
by 4, from
(3),
x"^
Hence
or
?/
(1)
(2)
(3)
xy
y'^
(4)
(5).
(6) (7)
From
663
(1), (6),
3,
and from
(1), (7),
2,
3.
If given a
637,
given equations
solve for
ti
and
we may set x = k -\- v, y = u v and then The second method is essentially the same
=u
-{-
V,
y
1.
V,
we have
ti
+y = (x
a;
and x
y;
for, since
-\-
y) /2, v
2/
=(x
y) /2.
(1)
Example
In
Solve
x2
(1)
and
(2) for
0,
0.
(2)
327
(3)
2/)2
Eliminating xy,
Solving,
or
Hence, from
or
(3),
(4)
(6)
(6)
(7)
(8)
(9)
And
X,
(6), (8)
and
(7), (9)
for x, y,
we have
= 2ViT, 2tV41; (_12V3)/3, (-1t2V3)/3. x* + y* = 97, Example 2. Solve x + y = 5. x = u + v, y = u v. In (1) and (2) set - u)* = 97, We obtain (u + 1;)* + (u
and
Eliminating
Solving,
w,
(1)
(2)
(3)
(4)
2u- 5.
16 v*
600
1)2
151
= 0. = l/2
(6)
(5)
or
iVlEl/2.
(6)
Substituting
=:
V,
y
X,
=u =
2/
of v in the formulas
3;
3,
2; (5
iVT51)/2, (5TivT51)/2.
/?
Evidently
pair,
if
=
a
a,
is
= p,
\^
another solution.
;
^
y,
as for
x = p, y ^ and namely a
-\-
but xy and x
+y
a,
ft
and the
corre-
a, differ
only
in sign.
Hence
the values of xy or x
pair will be less numerous than the values of x or y, that is the degree of the equation in xy or x -{- y derived from the
pair by elimination, as in Ex. 1, will be less than the degree As for of an equation in a; or y similarly derived would be. c must be the equation in ay, if e is one of its roots,
another root.
Hence
this
powers of x
~ y,
as in Ex. 2, or only
constant term.
328
664
A COLLEGE ALGEBRA
The methods just given are applicable to pairs of equations Note. which are symmetric with respect to x and y or some other pair of functions of x and y. Tlius, x* + 2/* = a, x -y = b may be written X* + (- y)* = a, x + (- y) = b.
665
Symmetric pairs
y^
Adding
(1)
and
(2)
(2),
x^
(1),
Subtracting
from
x^
+ -
2/^
7/3
(1)
(2) (3)
,
(4)
By
and
341, (3) is
(5)
-xy +
X
y-
= = =
10.
(6)
Similarly (4)
is
2/
0,
4. (6), (8),
(7)
(8)
and
x"
+
(1
xy
;
y"
And
-2,
solving
(5), (7)
(5), (8)
((3)^(7)
we
obtain
0,
2,
EXERCISE XL VIII
Solve the following pairs of equations.
= 5, JX + \xy + 36 = 0.
2/
7.
10.
12.
14.
329
A system of three equations in three unknown letters can be solved by means of quadratics when one of the equations is of the second degree and the other two of the first degree also
;
666
when
possible to reduce the system to one or more equivalent systems each consisting of one equation of the second
it is
The like is true of a degree and the rest of the first degree. system of four equations in four unknown letters, and so on. If A, B, C are integral functions of degrees 7n, n,p in x, y, z,
and no two of them have a common factor, the equations will have vmp solutions. But some of :=0, B = 0, C
these solutions
may
be infinite.
z-
Example
1.
Solve
+ 3x -2?/ + 2z +
xy
a;
z 2
= = =
0, 0,
0.
z,
Solving
(2), (3)
for x
and y
in
terms of
x
2/
(4) (5)
2z
57
2
(6)
Solving
=-3
or 19/7.
Hence, from
x,
330
A COLLEGE ALGEBRA
EXERCISE L
present chapter.
r7x^-6xy =
8,
rt
\2x-Sy = b.
"' ^
-y -a,
=
?/
Ic ,x-2/
3.
1.
xy
fX
(62-a2)/4.
+ = a2 +
i.
62,
2/
X
6.
+
a
=
+
+
6
-\-
6,
x^
5.
1x2
2/2^0.
1
1-1=12.
X2/
2/2
= 1.
a
^x
+ +
_1001
'
1-02
62
x2
7^^y^~ l25
1
2/2
17 = XJ?,
4
X2
8.
?/2
+
+
n
5
=2.
xy
^y-
rX2/
10.
a (X
2/)
& (
- 2/) =
0,
11.
40x?/
x2
= 21{x2-2/2) = 210(x +
3
X2/
2/).
1-4x2-25 2/2
12
2/2
=
2/2
9,
l2x2-10y2_32/ =
7
13.
4.
x2x/2/
13x2/
21
= -9.
14
j'a;2-
2/2-29
0,
lx2-6x2/ + 92/2-2x+62/ = 3
H'
15.
2 (x2
X2/2
x22/
6.
16, LX2/2 xy
1x22/
= X2/2 =
^^
a,
6.
-{
21
23.
x2
(x
2/)/(x-2/)
=
2/)
5/3,
(2x + 32/)(3-2
= 110a2.
r3(x3- 2/3) = LX - y = 1.
21 (x
2/)
13x2/,
'X* r^"
-r ?/*: </"
2/
"")
Ix +
22
X + = 21,
7.
a.
+ 2/2 =
x2
X2/
=
2/2
2/
2/.
24.
x2
2/-
y\
25
+ X2/ +
{ X + Vxy + =
2/
26
12(x-2/),
27.
x2
28
2/2
29.
28 (pfi r28(x6
Ix
+ y5) = =
:
61 {x3
+ y^),
30
331
+ 1)3 + (y _ + = 2.
?/
2)3
19,
f 2/2
33.
- xy - yz = 3, + iy + z = U, x-2/ + 2z = 0.
x
{y
(z
34.
-l
35.
?y
36.
xy
EXERCISE
1.
LI
is
The
two numbers
their difference
2. The square of the sum of two numbers less their product and the difference of their cubes is 189. What are they ? 3.
63,
is 11, and the product by one whose numerator and denominator exceed its numerator and denominator by 3 and 4 respectively is 2/3. Find the
The sum
of this fraction
fraction.
4. Separate 37 into three parts whose product is 1440 and such that the product of two of tliem exceeds three times the third by 12.
5.
The diagonal
than
it is,
of a rectangle
is
13 feet long.
If
feet longer
What
6. The perimeter of a right-angled triangle is 36 inches long and the area of the triangle is 54 square inches. Find the lengths of the sides.
7.
The hypotenuse
of a right-angled triangle
is
8. Find the dimensions of a room from the following data its floor a rectangle whose area is 224 square feet, and the areas of two of its side walls are 126 and 144 square feet respectively.
:
is
rectangle is surrounded by a border whose width is 5 inches. of the rectangle is 168 square inches, that of the border 360 square inches. Find the length and breadth of the rectangle.
9.
The area
332
10.
A COLLEGE ALGEBRA
In buying coal A gets 3 tons more for $135 than B does and pays than B pays for 5, Required the price each pays
$7
per ton.
certain principal at a certain rate amounts to $1248 in one year Were the principal $100 greater and the rate li times as great, the amount at the end of 2 years would be $1456. What is the principal and what is the rate ?
11.
at simple interest.
12.
all.
A man
The
children receive
leaves $60,000 to his children and grandchildren, seven in of it, .which is $2000 more apiece than the J
grandchildren get.
children,
13.
less
How many
At
time than
time
downstream would be only 1 hour less than his time upstream. is his usual rate in dead water and what is the rate of the current
14.
What
?
Three men A, B, C together can do a piece of work in \ hour, 20 minutes. To do the work alone it would take C twice as long as A and 2 hours longer than B. How long would it take each man to do the
work alone ?
15.
Two
bodies
and
are
moving
at constant rates
makes one
and in the same whose length is 20 feet. time than B, and A and B are
What
On two straight lines which meet at right angles at the points A and B are moving toward at constant rates. A is now 28 inches and B 9 inches 2 seconds hence A and B will be 13 inches apart, from
;
will
be 6 inches apart.
At what
rates are
Three men A, B, and C set out at the same time to walk a cei'tain A walks 4| miles an hour and finishes the journey 2 hours before B. B walks 1 mile an hour faster than C and finishes the journey
distance.
in 3 hours less time.
What
is
the distance ?
start sinniltanoously
tively
from P and Q respectoward each other. When they meet A has traveled 12 miles farther than B. After their meeting A continues toward Q at
18.
Two
couriers
A and B
and
travel
the
same
What
is
333
X,
Y
667
^x.
2
Example
1.
y2/
(1)
Solving for
Vx.
(2)
negative, y is imaginary ; when x is when x is positive, y has two real values which are equal numeri0, t/ is Hence the graph of (1) lies entirely to the cally but of opposite signs. right of the y-axis, passes through the origin, and is symmetric with respect to the x-axis.
(2) it
From
;
follows that
when x
is
When
we have
x=:0, 1/4,
?/
1/2,
1,
2,
3,
4,...,
0,
1, V2, 2,
V2,
Vs, 4,..-.
We
solutions
obtain the part of the graph given in the figure by plotting these 1), and passing a curve through the (0, 0), (}, 1), (],
Com-
pare
389.
It
touches the
called a
i/-axis.
This curve
parabola.
It
is
consists of
to
points
2/2
graphs of y
?/=x
1.
l{3),
2/
= x Z (2), = x + 3(4)?
(1),
;
The
1,
(1,
solutions of
(2)
are
points
2.
Hence,
as
is
386, the
graphs of
(1), (2)
intersect in the
shown
Tlie solutions
equal,
namely
1,
2;
1,
2.
Hence
meets the graph of (1) in two coincident points at (1, 2). Tins means that the graph of (3) touches the graph of (1) at (1, 2), as is
the graph of
V2
i.
l2V2i, (1), (4) are imaginary, namely Hence, as the figure shows, the graphs of (1), (4) do not meet.
334
Example
3.
A COLLEGE ALGEBRA
Find the graph
2/2-2x2/
of
2/
Solving for
y,
we have
+ 2x2-6x + 2 + = X 1 Vi X ?/
0.
(1)
(2)
x2.
The values
x(4
or
Hence the graph of (1) lies between the lines x = and x = 4. X = and when x = 4 the values of y given by (2) are equal: namely 1,-1 when x = 0, and 3, 3, when x = 4. Tliis means that the
4).
x), is
of y given by (2) are real when the radicand 4 x x^, or and 4 (or is positive (or 0), that is, when x Ues between
When
graph of
the point
line
(1)
(0,
tmdies the
line
at
1)
and the
See Ex.
line
2, 2.
=4
The
at the ptiint
xpoints of tangency, for when 4x vanishes, (2) gives the same values for
y that y = x \ gives. For each value of x between and 4 the equation (2) gives two real and distinct values of ?/, obtained by increasing and diminishing the value
of X
(4, 3).
by that
of
V4 x
x^.
Hence
two points of the graph of (1). They are most readily obtained by drawing and then increasing and diminis hing its ordinate for
tlie
value of
1,
v4 x
2,
1,
x^.
Thus, when
=
=-
0,
1, 1,
3, 4,
2, 3,
3.
we have
and
?/=
?/
0,
V3,
2,
V3,
shows the oval-shaped curve which the points thus found It is called an ellipse. determine. By solving (1) for x and applying the method of 641, we may show that the highest and lowest points of the curve are (2 + V 2, 1 + 2 V 2)
The
and
(2
V2,
4.
2 a^).
?/-
Example
x-
+ 2x +
2 y
0.
(1)
=-
V(x + l)(x-3).
(2)
The radicand (x + 1) (x 3) vanishes when x 1 and when x = 3, and in both cases (2) gives equal values of ?/, namely 1, 1. This
335
=
line
1 at
the point
1
(1,
1)
passes through
3) is positive when and only when x < 1 (x + 1) (x For every such value of x the equation (2) gives two real and distinct values of y and therefore two points of the graph of (1). These points may be obtained by drawing the line y 1 and then increasing and diminishing its conThe radicand
or X
>
3.
stant ordinate
by the value of
indicated
(1)
V(x
l)(x Hence, as
3).
is
in
the
figure, the
graph of
consists of
two
infinite branches, the one touch 1 and extending ing the line x = indefinitely to the left, the other
3 and extend-
ing indefinitely to the right. This curve is called an hyperbola. There are two straight lines called
asymptotes,
branches of
touch at
infinity.
x,
and y
hyperbola tend to touch, and which they are said to These lines are the graphs of the equations y = x 2 which we obtain as follows. Compare 650, Ex. 1.
(1)
Eliminating y between
we
obtain
(m"^
1) x^
2 (vie
?/i
1)
(c-
= mx + c, 2 c + 4) =
0,
0.
(3)
0.
(4)
Both roots of
if
(4)
mis, if
that
m=
y
1,
1= = 2,
or
and mc
or
if
+m+1= m = 1, c =
if (3)
= x-2
(30
(3')
=-
x.
and
(3")
two
Example
that
5.
Solving for y,
is,
?/2
3 X or
?/
4 xy + 3 x2 + 6 x Vx^ 2 x + 1, = X + 2.
?/
0.
(1)
(2)
= 3x
and
=X+
2.
1 = 0, the equaExcept when the radicand vanishes, that is, when x it two real and distinct values of y. But when x 1 =
336
gives
A COLLEGE ALGEBRA
two equal values
of y,
namely
3,
3.
Hence the
line
meets the graph of (1) in two coincident points at (1, 3). cannot mean that the line x - 1 = touches the graph of means that the points coincide in which the line x 1 =
lines,
2/
x 1 Of course
meets
tlie
=
this
It
= 3x
6.
and y
2,
two which
(1).
Example
their mtersections, of
x2 + ?/ ^ 25, XV16 + 2/V9-2.
(1>
(2)
The graph
center
is is
at
5.
radius
ellipse.
circle
(4,
3),
points (4, 3), (-4, these points being the graphs of the solutions of
7.
3),
(1),
(-4, -3),
(2).
Example
y
2/
From
(1)
we
obtain
- 2 = 0, + 3 = 0. y = 2/ (x -
(1)
(2)
3).
(3)
Here there
is
y,
of the graph,
When
= -3o, -1,
0,
1,
2|,
_ t, - |, _ l, - 2, - 8, oo, 8, 2, 0. we have y = 0, And plotting these solutions, we obtain an hyperbola whose two
3^, 4, oo,
infinite
branches are indicated by the unbroken curved lines in the figure, and whose asymptotes are y = (found as in Ex. 4), and x 3 = (since,
when X =
3,
then y
(x>).
cated by
tlie
the asymptotes y
The equations
(2)
have the
1, ?/ 1, single finite solution x the remaining three solutions being infinite, 654.
The hyperbolas which are the graphs of (1), (2) meet in the single
finite
point
(1,
1).
337
common asymptote y =
0,
infinitely distant coincident intersections at (ac, 0) ; and since they have o 2 =: 0, they are regarded as the parallel asymptotes x and x
(0, oo).
668
x,
y given,
'
+ 2 hxy +
by
bif
2 gx
y,
2fy
+c=
0.
(1)
not
0,
we have
(2)
be).
where
R = (A^ _
(2) we obtain two real values of y for each value of x which the radicand R is j)ositii)e. Corresponding to these two values of y there are two points of the graph which may be found by drawing the line
From
for
by=-(JiX-^f)
and then increasing and decreasing its ordinate for the value of X in question by the value of '\li /b. See 667, Exs. 1, 3, 4.
The form
factors of R.
1.
When
(hf
byy
is
- (h^ -
ab)
(f ~
be)
0.
a perfect square, 635, and the first member of (1) can be resolved into factors of the first degree, If these factors have real coefficients, the graph 635, Ex. 3.
In this case
i?
of (1)
2.
is
See
667, Ex.
5.
When
(hf
In this case,
reduced to the
a and
^8
- (h^ - ab) (f^ - be) > 0. bg)^ ab = 0, the radicand R can be unless form R = (h^ ab) (x a) (x /S) (3) where
h"^
a<
/3,
635.
338
If h^
A COLLEGE ALGEBRA
ab
<
0,
is
positive
when X
between a and /3. Hence the graph of (1) will be and x ^ = 0, a closed curve lying between the lines x a = See It is therefore an ellipse (or circle). it touches. which
lies
> 0^
or
when
x<a
x>
the product (3) is positive when and only Hence the graph will consist of two ft.
fi extending to its left, the other touching the line x and extending to its right. It is therefore an hyperbola. See
and
667, Ex.
If
4.
W -ab = 0, we have
0,
R=
- be), where
x:> (f^ be) /2(hf bff). of one infinite branch lying entirely to one side of the line It is therefore 2 (A/ hff)x + {f^ be) = 0, which it touches.
2i
Jif-^hgi^
and
this
is
positive
when and only when we have Hence the graph will consist
parabola.
3.
See
667, Ex.
1.
- (Ir - ab) (f^ - be) < 0. brj)^ are conjugate imaginaries, In this case the roots of 72 = we can reduce R 635, and if we call them X + fxl and X to the form R = {h'' - ab) [(a; - X)^ + /x^], (4). If h^ ab>0, the product (4) is positive for all values of x.
When
(hf
ixi,
Hence
which
lie
1. therefore an hi/perbola. Thus, v/^ x" If h^ ab< 0, the product (4) is negative for all values of X. Hence the graph of (1) will be wholly imaginary. Thus,
the graph of (1) will consist of two infinite branches It is on opposite sides of the line &y [hx +/)
x^
+f+
0.
In the preceding discussion it is assumed that b^O. Biit it b = 0, while a ^ 0, and we solve (1) for x instead of y, we and b = 0, the If both a = arrive at similar conclusions.
graph of
(1) is
straight lines of
to the y-axis.
an hyperbola, as in 667, Ex. 7, or a pair of which one is parallel to the a;-axis, the other
339
y2
x2
2/2
__ + -
8 X.
2.
x2
_
2/2
9.
3.
(y
_ +
0.
a;)2
= = =
x. 0.
4.
6. 8.
2 X2/
2/2
= 8.
5.
?/2
_ -
4 ^y
2 xy
x"^
2x2
2 x2
10. 11.
7.
9.
2/2
y"^
1 = - 3x +
+ -
3 x2
4x
0.
=
0.
0.
4x2-12x2/4-
92/2
+ 3x -6y =
Find the graphs of the following pairs of equations and their points of
intersection.
12.
fxy
^
l, '
Ux-5?/ =
r 2/2
13.
2.
X2/
2 x2
2 X
14.
0.
fx2
2/2
3,
ly^=2x.
2/
L2/2-X2/-2x2 +
^g
17.
0.
+ |(x-2 2/)(x t (X - 2 (X
2/)
2/)
2/)
+ x-32/ = 0, + 2 X - 6 = 0. of x2 + 2/2 6x
2/
22/
+ l=0 and +
y)
its
points of
x and
y.
Show
axes.
y)-
2{x
touches the x
and y
19.
Show
2/
= 3 x + 5 touches the
graph of 16 x2 + 2/2 16 =
3 touch the graph of
at the point
20.
(-3/5,
16/5).
= mx +
x2
2 y2
6 ?
c will
the line
7x iy +
1 =:
touch the
Show
2/
graph oi xy
2 y^
= +
and x
6
2y +
0.
23.
Find the asymptotes of the graph of the equation 2x2 + 3x2/-2 2/2 + x + 22/ + 2 = 0.
24.
a parabola
is
the graph of x2
\xy
2/^
x an ellipse ?
340
A COLLEGE ALGEBRA
XVII.
INEQUALITIES
absolute
669
"
Single
inequalities.
An
inequality
all
is
is
one
like
y''
>
;
letters involved
a conditional inequality
one like
cc
>
670
which does not hold good for all real values of these letters, but, on the contrary, imposes a restriction upon them. The principles which control the reckoning with inequalities
are given in or sign
261.
From
these principles
it
connecting the two members of an inequality will remain unchanged if we transpose a term, with its sign
changed, from one member to the other, or if we multiply both members by the sa^me positive number but that the sign > will
;
>
<
if
Example
1.
Prove that
We have
that
is,
a2
and therefore
Example
2.
Prove that
a2
h62
-\-
d^
We have
52
> 2 ah,
c2
-\-
he
c2
we have
a'^
Example
restriction
3.
ca.
is,
that
find
what
it
x.
Transposing terms,
2x
whence
x
4.
Example
Solve
(x
x2
2 x
- 3 < 0.
3)
Factoring,
1) (x
To
negative.
one factor must be positive and the other Hence we must have x > 1 and < 3, that is, 1 < x < 3.
671
Simultaneous inequalities.
qualities of the
A
hy
ine-
form
ax
+c>
INEQUALITIES
may
be solved for the variables
is
341
method which
Draw
385.
the straight line which is the graph of ax + bi/ -\- c = 0, Then for all i)airs of values of x, y whose graphs lie
on one side of this line we shall have ax -\- by -\- c> 0, and for all pairs whose graphs lie on the other side of the line we shall have ax -\- by -\- c < 0.
Thus,
?/i
let
{xu
c)
l/i)
{mx +
c)
so that
(inxi
= 0.
Then,
if
^2
c)
+
7/3
tliat
if 7/3
=
A:2
the point (xi, 2/2) lies below > 2/1 so that the point (xi, 2/3)
-f c)
> 0.
Example.
ki
= X - 2y + K0, ki =
A-2
+ y-5<0,
0, ks
k3
= 2x -
1>0.
0,
0,
The
inequality
<
is
satisfied
by those pairs of values of x, y whose graphs lie on the side of the line toward the origin for when ^2 = X = 0, 2/ = 0, we have ko = 5, that
;
< 0. It may be shown in a similar manner that the inequalities ki<0 and kz>0 are satisfied by the paira of values of x, y whose graphs lie on
is
0,
/cs
A>
< 0, ko< 0,
ks
>
are satisfied
by y whose graphs
lie
EXERCISE
imequal positive numbers.
1.
LIII
a, b, c
+ &/n> 2.
b) (a^
2.
3. 4.
Prove that
(a
b"^)
>
(a2
+
+
62)2.
Prove that a^
Prove that
+
+
6^
>
a'^b
062.
a'^b
b"a
b^-c
d^b
d^a
+ a^o 6 abc.
342
5.
6.
A COLLEGE ALGEBRA
Prove that
Solve x
a'^ + b^ + c^ > 3 abc. + 7>3x/2 - 8.
7.
8. 9.
Solve2x2
Solve (X
Solve
4x>x-^
1) (X
3) (x
+ 6x + 8. - 6) > 0.
X
?/-x 2<0,
3<0,
+ l>Oby
10.
11.
Also y Also x
x>0, +
2/
2x<0.
2/
+ 3>0, 2/-2x-4<0,
+ 2x + 4>0.
all
12. 13.
Prove that x^
Solve x2
2/2
+ 2x + _ 1 < 0,
>
is
true for
values of x.
2/2
4X
<
by a graphical method.
XVIII.
673
form
ax
a, b, c
-\-
01/
where
factor.
We
denote integers, of which a and h have no common seek an expression for all pairs of integral values
of X and y which will satisfy this equation; also such of these pairs as are positive as well as integral.
673
Theorem
1.
All equations ax
by
prime to one another, by the method explained two integers p and q, positive or negative, such that ap -{ hq \ and therefore a (pc) + b (qc) = c, and this proves that x = pc, y = qc is & solution of az + by =: c.
h are in
we can
find
674
=
-f-
x,
= y^
c,
by
all
of
its
Xo
+ bt,
= yo -
at
to
when
mDETERMINATE EQUATIONS
First,
343
Xo
bt,
yo
at
is
always a solution of ax
hy
c.
(1)
For, substituting in
or, simplifying,
(1),
a (Xo
-\-bt)
b {yo
oxq
= + byo =
at)
c, c,
which
true since, by hypothesis, x = Xo, y = yo is a, solution of (1). Second, every integral solution of (1) is given hy x = xq + bt, y = yo
is
at.
For
let
Xi,
Then
axi
byi
and axo
b{yi
is
whence, subtracting,
+ byo = c, yo) =
a{xi
Xo).
(2)
From
and
Xi
(2) it
follows that 6
Xo.
Therefore, since b
prime
call
tained in Xi
Xo, 492, 1,
and
Xo
if
we
Xi
= W, =
or Xi
Xo
bt'.
(3)
And
substituting (3) in
(2)
2/1
and simplifying, we
2/0at'.
also
have
(4)
673, 674 it follows that every equation ax + by = c of the kind just described has infinitely many integral soluWhen a and b have contrary signs there are also tions. but when a, and b have the infinitely many ^:(os/f iye solutions
From
675
is
3y
18
is
3,
2/
4.
Hence the general solution isx = 3 + 3i, y=:4 2<. The positive solutions correspond to ^ = 1, 0, 1, 2 and are
3,
x,
1/
0,
6, 2
9, 0.
The theorem
of
676
integral solution of an equation of the kind under consideration as soon as a particular solution is known. particular solution may often be found by inspection. Thus, one solution
= 4. A particular solution may a; + 3 ?/ = 12 is a; = 0, y always be found by the method indicated in 673 also by the method illustrated in the following example.
of 10
;
Example.
19
2/
213.
x,
(1)
and reducing,
(2)
we have
x
= ^li:ii^ = 30-22/ +
'-^.
344
Hence
integer.
if
A COLLEGE ALGEBRA
x is to be an integer when y is one, (3 5 2/) / 7 = u. Call this integer u, so that (3
5
5?/)/ 7 must be an
Then
Treating (3) as
7 u
3.
(3)
we have
just treated
(1),
we have
...
3-7m ^5
Set (3
3-2u
5
to
v.
Then
Treating
(5)
5u
3.
(5)
as
we have already
M
treated (1)
and
(3),
we have
(o)
=
2
=:l-2v + ^-
2
When v-\
integral value
(1-d)/2
1.
Substituting u
=-
1 in (4),
we
obtain y
2.
= 25. Substituting y = 2 in (2), we obtain x Hence the general integral solution of (1) is
x
25
19,
2/
= 2-7*.
to
<
respectively,
=-
and
are merely the successive remainders occurring in the process of finding the greatest common divisor of the given coefficients 7 and 19. We finally obtain the remainder,
namely
5, 2,
The like will factor. 1, because 7 and 19 have no common be true if we apply the method to any equation ax + by = c in which a and b have no common factor. Hence the method will always yield a solution of such an equation. it is seldom necessary to complete the reckoning above But in
or coefficient,
practice
indicated.
will
make
(3
= Thus, having obtained (4), we might have observed that u - 2 u) /5 integral, which would have at once given us y
=
25,
1 =2
and therefore x
by
(2).
677
Observe that an equation ax + bi/ = c with integral coefficients of which a and b have a common factor, as d, can have no integral solution unless d is also a factor of c. For if x and ax + by and therefore y were integers, d would be a factor of of c. Thus, 4 X + 6 ?/ = 7 has no integral solution.
INDETERMINATE EQUATIONS
Simultaneous equations.
trate a
345
will
illiis-
678
method
with integral
Example.
coefficients.
of
22, 28.
(!)
(2)
3x + 6y -2z = 5x + 8y -(iz =
First eliminate z
We
obtain
5y
=
=
19.
(3)
Next
"We obtain
x
(4) in (1)
=7+
^t, y
2t.
(4)
Next substitute
result.
(5)
We We
obtain
2z - St =
5.
Next
obtain
=1-
Su,
=-
2u,
(6)
We
which
obtain
is
(7)
The
y
0,
namely x
2,
S,z
1.
tion
Observe that the given equations will have no integral soluif either of the derived equations in two variables has
none,
677.
in a similar
We proceed
example
integral
manner
if
variables.
The following
679
method
a
solutions
of
Example.
of
5x + Sy + 192 =
5
50.
(1)
Solving for X,
Set
(3
10
-y-Sz-
^11^.
(2)
4 z)/5,
which must be
integral, equal to u.
346
Then
Solving for y,
Set (2 m
A COLLEGE ALGEBRA
3y
42
= 5 w. =u-
(3)
-\
(4)
u.
Then
Substituting Substituting
(5) in (4),
z-2u-Zv.
y
(5)
(6)
= =
10
-{
Av.
(5)
and
(6) in (2),
- 6u +
u, v
V.
(7)
The formulas
which
integral values
whatsoever, constitute the general solution required. Substituting u = 2, v = \ in the formulas (5), (6),
positive solution of (1),
(7),
we
obtain a
namely x
3,
2, z
\.
EXERCISE LIV
Find the general integral solutions of the following
integral solutions.
1.
;
3.
5.
2. 4. 6.
9.
22/
7 2
l3y10.
= =
42,
2.
2x + 3y + 4z =
17.
11.
3x + 7y =
12.
sum
A man
all
buys calves at
.$7
He
spends in
14.
$110.
How many
first part,
Separate 23 into three parts such that the sum uf three times the twice the second part, and five times the third part will be 79.
5, 7,
15.
9 will give
the remainders
16. Two rods of equal length are divided into 250 and 253 equal divisions respectively. If one rod is laid along tlie other so that their ends coincide, which divisions will be nearest together?
347
XIX.
VARIATION
In arithmetic and algebra it is customary to extend Ratio. the use of the word ratio, 215, to numbers and, if a and b denote any two numbers, to define the ratio of a to i as the
;
680
quotient a /b.
(Compare
a to
i is
216.)
denoted by a /b ovhj d'. b. In the ratio a:bwe call a the antecedent and b the conseqxient. Since ratios such as a 5 are fractions, Properties of ratios.
ratio of
:
The
681
Hence
its
The value of a
Thus, a:b
ratio is not
terms
same number.
= ma mb = a /n
:
b/n.
On
is
the other hand, except when a b, the value of a:b changed when both terms are raised to the same power, or
the same
b,
when
number
is
added to both.
In particular,
If a, adding
and
to
are positive, the ratio a:b is increased by b decreased, when a > b. both a and b ivhen a
<
^ For
and
a
b
;
+m
-\-
a
b
m {b
a) /bip +
m)
is
<6
or a
> 6,
682
Proportion.
When
a, b,
c,
the ratios a
and
four
numbers
d are said
to be in proportion, or to be
projiortional.
This proportion
may
a/b
It is
= c/d,
is to i
or as
:
a:b
= c:d,
d,
or a
::
d.
read " a
c is to <Z."
In the proportion a
the extremes,
=
c
and
and
the means.
Again, d
is
called the
fourth proportional to
a, b.
and
c.
848
683
Theorem.
A COLLEGE ALGEBRA
In any proportion the product of the extremes
is
equal
to
that of the
means ; that
:
is,
If
For from a/h
Example.
a b
d,
then ad
be.
= c/ d we
first,
obtain ad
be
The
.second,
/2
5.
Hence
or, solving for X,
5-l/2=:-3 x, x = 5/6.
684
to
Conversely, if the 2}roduct of a first pair of numbers be equal that of a second pair, the four numbers will be in prop)ortion
in
when arranged
let
the pairs
be.
we have a/b c/ d.
and
c
Hence
(2)
a:b = c:d
turn,
(1)
members
and
and
of
= ad =
:
a
be
6.
by
cd, a6,
and ae
in
-.b
= b:d =c a
:
(3)
(5)
d:c 685
= b:a
(7)
and
= = b:a =
6
:
c,
C'.a
d:b,
d:c.
Allowable rearrangements of the terms of a proportion. From it follows that if a, b, c, d are in proportion when any one of the orders (l)-(8), they will also be in
proportion
particular,
1.
when arranged
in
orders.
In
may
be inter-
changed.
Thus,
2.
c.
may
be interchanged. Thus,
if
-.b
b:d.
349
alter-
686
1.
-{-
-\-
d.
2.
3.
a-\-b: a
:
= =
4,
ma
tub
= nc
nd.
6.
b = c d d. d. c-{-d: c 5. ma nb = mc nd.
a"
^z"
c"
d".
For
ad
in 1
we have
Hence
6d
= 6c + 6d,
ad
be,
which
is
true since a
-.b
d.
1 is true, 288.
The
truth of 2-6
may be proved
in a similar
manner.
The transformations
rtjspectively.
of
a:b
Example.
Solve x2
By
3,
Hence
or by 4,
5,
Solving
(1)
and
(2),
Theorem.
In a
series
is to its
687
consequent as the
sum of
a\:hi
ni
:
sum of
Thus,
then
if
bi
a-2 .b2
as:
63,
:
ai
ao
as
61
62
&3-
For have
let r
denote the
common
We
then
ai/bi
Hence
or,
Oi
adding,
ai
mu * Therefore
ai
+ ao + 03 + 02 + a3 = +h 61 +
?>a
= at
W
350
Example
then
1.
A COLLEGE ALGEBRA
If
{b
y"^
c)yz
x^
z'^
= =
y.
0.
{c
a)zx
{a
h)
xy,
y,
For multiplying the terms of the first ratio by x, those of the second by and those of the third by z, and then applying our theorem, we have
x^
(6
_
xyz
(c
y"~
c)
_
xyz
(a
?/2
z2
_
xyz
0.
a;2
y2
_|_
g2
a)
h)
2-
in the proof just given w^ill be found useful in dealing with complicated problems in proportion.
=x = x3 + 2 2/3 xy^. Set a/h = x/y = r, so that a = rb and x = ry. Then /a62 = (rW + 2 6^) /y^? = (a^ + 2 = {rhj^ + 2y^)/ ry^ = and (x3 + 2 2/3) / X2/2
Example
Prove that
aS
if
?/,
then
2 6^
a62
ft^)
(r^ {r^
+ 2)/r, + 2) / r.
are said to
688
Continued proportion.
The numbers
a, h,
c,
d,
b -.c -. be in continued proportion if a:h c:d If three numbers a, h, c are in continued proportion, so that a:b b c, then b is called a mean jjroportional to a and c,
and
c is called
b, c
b.
i/a,
are
= ac.
b-.c,
we have
6^
^c, 683.
EXERCISE LV
Find a fourth proportional to 15, 24, and 20; a third proportional to 15 and 24 a mean proportional between 5 a%'^ and 20 aft^ a mean proportional between V'l2 and VtS.
1.
; ;
2.
If 3
2/
=
+
5
2/-
2/,
find x :y
0,
fijifi
also
2/
;
+
2/
y x
:
:
y.
3. 4.
If
2x2
5 xy
6?/
=
z
also
6'2/
x.
If
ax
+
:
cz
:
then
5.
X
If 62 j^
= =
and
be'
a'x
:
+
ca'
+
c'a
c'z
:
0,
is
b'c
ab'
a'b.
d,
then
aft
cd
mean
proportional between a?
c^
and
6.
(^2
If
(pfi
62)
cd
{c-
+ ^2) a6,
c:(Z or
a:6 = d:c.
351
d.
If
=
^
b
rf,
then
Va + \^
a^
b^
^^a
-{-
= Vc
b
i-
Vd Vc +
:
8.
If^
a
= ',then^- +
c !, a2
^V = 3i^+A+^'. + +
c^
(a
c)3
9.
If the
numbers
) ;
i>i,
:
b^,
b,,
/i, Zoi
ia
lb is
:
intermediate
6i, a-2
60,
&
10.
If
b:k =
0.
c:l
Iz
a: m, and
a, b, c
+ +m=
I
11.
If
X
y2
mz ny = y
22
61
-.nx
-.ly
mx, then
Ix
+ my +
is
nz
and x2
12.
+
:
= =
1^
0.
If ai
ao
:
62
03
^3,
equal to
(hal
hal
+ halY (hK + W, +
x"^
hbl)n.
13.
By
ax
ax
(1)
x"^
a
a
2 x2 2 X-
+
2
rt
2x3-3x2 + 2x +
(2)
3x3 3x3
2x3-3x2-2x-2
A
and
14. 15.
Two
casks
:
mixed
in
A in the
ratio 3
5, in
in the ratio 3
7.
be taken
from each cask to form a mixture which kind and 12 gallons of the other kind ?
one
VARIATION
One independent
related
variable.
If
that
however
their values
is
two variables y and x are so may change their ratio said to vary as x, or y and x are
689
said to vary proportionally. More briefly, y is said to vary as x when y / x where c denotes a constant. The notation y (X)x means " y varies as cc."
c,
ov
= ex, = ex
690
and
If given that y varies as x, we may at once write y if also given one pair of corresponding values of
x and
352
y,
A COLLEGE ALGEBRA
we may find c. The equation connecting ?/ and x is then known, and from it we may compute the value of y which corresponds to any given vahie of x.
of y
Example. when x
"We have
If
y varies as
?
x,
and y
y
12
when x
2,
what
is
the value
20
=
is
ex,
satisfied
c
when y =
is c
12,
2.
Hence
Therefore
12
Hence when x
20 we have y
= = =
2,
that
6.
(>x.
(]
20
120.
691
Instead of varying as x
itself,
may
+
y
Given that y is the sum of a constant and a term which varies inversely as x also that y = \ when ic = \, and y = 5 when x = 1. Find the equation connecting x and y.
Example.
By
a + b/x, where a and b are constants. hypothesis, y Since this equation is satisfied byx = -l,y = l, and by x
,
1,
5,
we have
Hence a
a r and 5
+
,
i,
o.
3, 6
2,
is
2/x.
//
692
Let x and
denote
variables which are independent of one another. If a third variable z varies as the product xy, so that z cxy, we say that z varies as x a7id y Jointly ; and if z varies as the quotient c X y, we x/i/,so that z / say that z varies directly as x and
inversely as y.
Tiius, the area of a rectangle varies as the lengtlis of its base
tude jointly
and altiand the length of the altitude varies directly as the area and
693
is
Theorem.
If irhcn x
is
constant z varies as
x,
then
select
of vnlnes of
x and
?/,
such as
Xi, yi
z,
X2,
2/2
and
so that
VARIATION
^1,
353
(1) (2)
Vu
zi
X2, ^2, Z2
(3)
are sets of corresponding values of the three variables. Then since the value of x is the same in (1) as in (2), and, esis, for any given value of x, z varies as y, we have, 689,
21/2/1
by hypoth(4)
23/2/2-
Similarly since
1/2
is
common
to (2)
and
(3), vpe
have
(6)
(4)
Z3/Xl=Z2/X2.
Multiplying together the corresponding members of
Zl/-ClUl
and
(5),
Z2/X2y2;
(6)
that
is,
EXERCISE LVI
1.
If
y varies as
7 ?
x,
and y
=~
x^,
when x =
5,
what
is
the value of y
when X
2.
If
y varies inversely as
and y
when x
2,
for
what values
of X will y
3.
S?
X-
Given that y is the sum of a constant and a term which varies as y = I when x = 1, and y = when x = 2. Find the equation connecting x and y.
;
also that
4.
If
and
5.
6.
=
If If
2,
y varies directly as x- and inversely as 2*, and y = 1 when x what is the value of y when x = 3 and 2 = 1 ?
y varies as
x,
show
that x-
y- varies as xy.
X,
show
7.
the square of y varies as the cube of 2, and 2 varies inversely as that xy varies inversely as the square root of x.
3
The wages of
for
men
for 4
men work
8.
$135?
of a circular disc varies as its thickness
The volume
its
face jointly. Two metallic discs having the thicknesses 3 and 2 and the radii 24 and 36 respectively are melted and recast in a
of the radius of
single disc having the radius 48.
9.
What
is its
thickness
right-circular cone whose altitude is a is cut by a plane drawn far is the plane from the vertex of the cone parallel to its base. when the area of the section is half that of the base ? How far is the plane
How
it
354
A COLLEGE ALGEBRA
XX.
694
of
ARITHMETICAL PROGRESSION
This name
derived,
is
Arithmetical progression.
given to a sequence
from a given number a by repeatedly adding a given number d, tliat is, to any sequence which may be written in the form
a,
-{-
d,
-\-
2d,
+ (n
T)d.
(I)
the difference between every two consecutive terms of (I), it is called the comvion difference of this arithmetical progression.
Since d
is
Thus,
2,
2, 5, 8, 4,
11
is
an arithmetical progression
in
which d
1,
7 is
= 3, = 3.
and
695
Observe that in
(I)
2 c?, a is one less than the number of the terms a, a c?, of the term. Hence the general or mth term is a (' l)f^ and if the entire number of terms is n and we call the last
term
I,
= a + {n-l)d.
(II)
Example. The seventh term of an arithmetical progression is 15 and its tenth term is 21; find the first term a and the common difference d; and if the entire number of terms is 20, find I.
We
have
a
(Z,
Hence
+ 6d = a = =
l
15 and a
3,
+ 9d =
41.
21.
2.
19,2
696
The sum.
be written
term,
I
Evidently the next to the last term of (I) d, the term before that, I 2 d, the
may
first
(71
if
l)d.
(I),
Hence,
we have
l)d^, l)d2.
S S
ARITHMETICAL PROGRESSION
Adding the corresponding members obtain 2 S = n(a + I). Therefore
S
of these
355
two equations,
we
^(a
l).
(Ill)
of six terms
Example. Find the sum of an arithmetical progression whose first term is 5 and whose common difference is 4.
Since n
6,
we have
+
(5
Hence
Applications.
S=
If in
a,
25)
= 25. = 90.
697
of the live
numbers
and (III) enable us to hnd the other two. The only restriction on the given numbers is that they be such as will lead to
positive integral values of n.
Example.
Given
iZ
1/2,
=
3
-
3/2, S a
-i
= - 15/2
find
a and
n.
Substituting in
(II), (III),
n
^
(1)
15 16
n/
3\ 3
,.
Eliminating
Solving
(3),
a,
n^
- 7n -
30
/I
= =
0.
(3)
10 or
3.
obtain a
is
The value n = 3 is inadmissible. Substituting n = 10 in (1), we = 3. Hence n = 10, a = 3, and the arithmetical progression - 3, - 2|, - 2, - 11, -1,-1, 0, I, 1, U.
Arithmetical means.
If three
698
number
mean
is
one
a, x, b
is
Hence
and therefore
a
x
= =
x,
6)
(a
/2,
356
A COLLEGE ALGEBRA
all
may
and
"
last
always possible to insert or interpolate any number of such means between two given numbers a and b.
Example.
Interpolate four arithmetical mean.s between 3 and
5.
terms.
"
asked to find the intermediate terms of an arithmetical progression in which a = 3, I b, and 7i == 4 + 2 or 6. Substituting i = 5, a = 3, n = 6 in (II), we have
5
We are
5d, whence d
2/5.
Hence
the required
means are
EXERCISE LVII
1.
first
twenty terms of
3, 6, 9, ; of 2.
-3, -li,
0, .
to
n terms
of
1, 2, 3,
of
1, 3, 5,
of 2, 4,
3.
first
n numbers
of the
form 6r
1,
where r
denotes
or a positive integer.
of ten terms
whose
fifth
term
is
means between
;
and 2
6.
Given
Ji
=
= =
10, a
1,
I
7.
8.
Given n Given n
12,
2,
7,
i
9.
Given a
=
=
= 4/3 find and S. 0, = -1, d = - 5/3 find a and S. a = - 5/3, = 311 find d and S. = - 23.^, .S = - 559 find n and d.
d
I
;
10. 11.
Given n
a
cZ
=
=
3/7, S
1
45
9?
;
find
d and
/.
Given a
Given
?i
4,
/5,
find
;
n and
S.
I.
Given
Given Given
= 9, d = - 4, .S = 135 find a and n = 10, = - 2, S = 115 find a and d. d = 5, = - 47, S = - 357 find n and a. a = - 10, d = 7, S = 20 find n and
Z
;
I.
Show
c),
that
if a-, b-, c^
1/(6
l/(c
a),
l/(n
GEOMETRICAL PROGRESSION
17.
if
357
is divisible
Show
that the
sum
of
by
n,
n be odd.
18.
terms
is
Find an arithmetical progression such that the sum of the first three one half the sum of the next four terms, the first term being 1.
is
Their sum 19. Three numbers are in arithmetical progression. and the sum of their squares is 83. Find these numbers.
20.
15
all
which are
multiples of
21.
If
sum
end
a person saves $130 a year and at the end of the year puts this how much will his savings amount at the
22.
miles apart to walk toward one another. miles an hour, while B walks 2 miles the
they meet?
XXI.
GEOMETRICAL PROGRESSION
This name
is
Geometrical progression.
given to a sequence
699
numbers which may be derived from a given number a by repeatedly multiplying by a given number r, that is, to any sequence which may be written in the form
of
a,
(I)
ratio of the geometrical progression (I) and say that the progression is increasing or decreasing according as r is numerically greater or less than 1.
call r the
We
common
sions in
8 are increasing geometrical progres4, respectively; while 1, 1/2, 1/4, 1/8 is a decreasing geometrical progression in which r 1/2.
Thus,
1, 2, 4,
8 and
1,
which
and
2, 2
Observe that the exponent of r in each term one less than the number of the term. Hence in a geometrical progression of n terms whose first term is a and
700
of (I)
whose
ratio is
r,
is
= ar"-i.
(II)
358
701
A COLLEGE ALGEBRA
Let S denote the sum of the geometrical pro-
The sum.
gression
(I).
Then
and
S
rS
= a + ar + ar^ = ar ar^ +
-\
-{-
ar".
Subtracting the second of these equations from the a ar'\ Therefore obtain (1 r)S
first,
we
s=
^J^.
it
(Ill)
thus:
From
(II)
:
written thus
Example.
terms, find
I
= a(?'-l)/(r-l). we obtain rl = ar". Hence (III) may also be S = (a rl) /(I r), ot S = (rl a) / (r 1).
2,
4, 8,
16,
to eight
and
<S.
Here
Hence, by
and, by >
'
a
(II),
i
= =
2,
8."
^ (III),
"
8=
2^^^^^=-'^. 3 l-(-2)
702
If in a geometrical progression any three of the five numbers a, I, r, n, S are given, the formulas (II) and Moreover these two numbers (III) determine the other two.
Applications.
can be actually found by methods already explained, except when the given numbers are a, n, S or I, n, S. If one of the
unknown numbers
is n, it
but
this is always possible if admissible values have been assigned to the given numbers, since n will then be a positive integer.
Example
1.
Given r
3,
6, -S
= =
728
find a
and
I.
we have
36
243
a,
and 728
a
a^^^ = 364a.
3-1
486.
2,
GEOMETRICAL PKOGRESSIOX
= 6, n = 5, = 2/27 find r and S. = 6r*, whence r* = 1/81, or r = 1- (1/3)5 242 Therefore, by (III), if r = 1 /3, then S = 6 27 1-1/3
Example
2.
359
Given
By
(II),
2/27
1/3.
and
if
=-
1/3, then
122
27
which a
= 6, n =
5,
Example
3.
Given a
= -S,
={a
46875,
S
{I
S =
rl)
n.
Therefore, by
39063
(II),
= nl + i^^IL^
1-r
whence
=-
5.
_ 46875 = - 3(-
5)-i, or
(- 5)-i
5)6.
= 15625.
Example
4.
Given a
93
=
1
3,
5, -S
93
find r
r2
and
I.
By
(III),
3^^ = 3(1 +
r
r*
+
0.
r3
r4).
Hence
r^
r^
30
problem involves solving an equation of the fourth degree when a, n, S are the given numbers, to find r we must solve an equation of the degree n 1. In this particular case, however, we may find one value of r by the method of 455. It is 2. Substituting r = 2 in (II), we have Z = 3 2* = 48.
this
;
Thus
and, in general,
Geometrical means.
If three
703
number
mean
a
The geometrical mean of any two numbers a aiid b square root of their product.
For if X be the geometrical a geometrical progression.
is
mean
of a
and
6,
the sequence a, x, h
is
Hence
x/a =
/x and therefore x
all
Vab.
the intermediate terms
first
may
and
last terms.
360
A COLLEGE ALGEBRA
insert
We may
numbers a and
Example.
It is
required to find the intermediate terms of a geometrical progression in which a = 18, Z = 2/27, and Ji = 4 + 2 = 6.
Substituting the given values in
(II),
we have
r
2/27
Hence the means are
18
r*,
whence
1/3.
6, 2,
2/3, 2/9.
704
the form
^^^.
_^ ^^,,
_^ ^,,._i
We call an ^
. .
expression of
^^^
.^
supposed continued without end, an infinite geometric series. By the formula (III), the sum of the first n terms of (1) is
a(l-.)/(l-r).
Suppose that
increased,
/"
<
numerically.
Then, as n
is
indefinitely
will
a(l /) will approach a /(I ?"")/(l r) as limit. this limit the su))i of the infinite series (1). Hence, if
the
approach
as limit,
AVe
call
S denote
sum
of (1),
we have
Example
Here
1.
Hence
1/8-1
Example
2.
recurs
may
be written
1-
1000
100000
, '
and, by ^
'
(2), ^
"
sum
is
.023
23
or
.01
990
Adding
6.1
.7,
does not recur, we obtain for the value of the ^ given decimal
358
495
EXERCISE
1.
LVIII
first five
Find the
fifth
terms of the
geometrical progression
2.
6,
18, .
first
geometrical progression
GEOMETRICAL PROGRESSION
3.
361
:
infinite
geometric series
12_6+3-...;l-i + i----;f + l + A +
4.
---.
:
.0567272
.03, r
8.45164516
I
5. 6.
Given a Given n
Given a Given n
== = = =
7, 1
= =
10,
i
?i
6
;
find
and
S.
48,
7.
8. 9.
/ 16, r
2,
Z
= 3/4 =8
i
find r find
and
S.
S.
n and
5, r
= 3, =
81
find a
;
and
S.
I.
Given a
54, r
10. 11.
12. 13. 14. 15.
16. 17.
=: 4,
find
;
n and
and
find r
I.
781/144;
;
find
I.
n and
r.
find a
and
a.
r.
Given r Given
?i
= =
3/2,
4,
5,
Z
301,
-^
= =
54/25, S
48,
=
;
544/25
find a
find
r.
a and
Given n
S = 93
and
means between
and
405.
is
18.
The
is
term
19.
the
first
Find a geometrical progression of four terms in which the sum of and last terms is 133 and the sum of the middle terms is 70.
20.
sum
shall
Find three numbers in geometrical progression such that their be 7 and the sum of their squares 91.
Three numbers whose sum is 36 are in arithmetical progression. 43 be added to them respectively, the results are in geometrical Find the numbers. progression.
21.
If 1, 4,
22.
first
The sum of progression and the last three in geometrical progression. the first and fourth is 16 and the sum of the second and third is 8. Find
the numbers.
23. What distance will an elastic ball traverse before coming to rest if be dropped from a height of 15 feet and if after each fall it rebounds to 2/3 the height from which it falls ?
it
362
A COLLEGE ALGEBRA
XXII.
HARMONICAL PROGRESSION
This name
is
705
Harmonical progression.
given to a sequence of
numbers whose reciprocals form an arithmetical progression, that is, to any sequence which may be written in the form
1/a, l/(a
Thus,
1,
d),
l/(a
2rf),
l/[a+(ft-l)(/].
1/2, 1/3,
progressions.
Example. c =2 a b
Prove that
:
if
a, 6, c
c.
is
l/6-l/a = 1/cHence
c (a
6)
(6
c),
is
c.
706
To find any particular term of an harmonical progression, we obtain the term which occupies the same position in the
corresponding arithmetical progression and invert
it.
Example. Find the tenth term of the harmonical progression 3/5, 3/7, 3/9,..-.
By
is
695,
the tenth term of the corresponding arithmetical progression is 23/3. Hence the tenth term of 3/5, 3/7, 3/9,
3/23.
707
Harmonical msans.
gression, the middle
If three
numbers are
is
in
harmonical pro-
number
mean
Again, in any harmonical progression all the intermediate terms may be called the harmonical means
of the extreme terms.
Example 1. Find the harmonical mean of a and 6. K this mean be x, then 1/a, 1 /x, 1 /6 is an arithmetical progression. Hence 1/x - 1/a = 1/6 - 1/x, or 2/x = l/a + 1/6.
Therefore
2 a6 / (a
6).
is
Prove that the geometrical mean of two numbers a and 6 also the geometrical mean of their arithmetical and harmonical means.
Example
2.
Let A, G, and 77 denote respectively the arithmetical, geometrical, and harmonical means of a and 6.
HARMONICAL PROGRESSION
Then
Henco
363
A=
'^^, G
2
= V^b, H "' ~ =
ah
2 ah
AH =
2
= G\
b b are positive,
Example
3.
A > G> H.
Wehave
H ='l^ - -^^
2
= i^^. + b 2 {a + b)
Therefore, since (a
2,
And
since,
by Ex.
A>G>n.
1.
A>H.
we have
EXERCISE LIX
Continue the harmonical progression 3/5, 3/7, 1/3 for two
terms.
2.
3.
4.
5.
15.
and
5.
and
their harmonical
mean
6.
15/4.
mean
7.
16/5.
is
Show
c),
that
a, b,
a/(b
b/{c
a),
c/{a
Three numbers are in harmonical progression. Show that if half of the middle term be .subtracted from each, the results will be in geomet8.
rical progression.
9. Show that if x is the harmonical mean between l/{x-a) + l/(x-b) = 1/a + 1/6.
a and
b,
then
meets bisector of the vertical angle C of the triangle at D, and the bisector of the exterior angle at C meets are in harmonical progression. produced at E. Show that AD, AB,
10.
The
ABC
the base
AB
AB
AE
11. The point P lies outside of a circle whose center is 0, and the If tlie line PO meets the tangents from P touch the circle at T and T. is the harmonical mean circle ut A and B and TT' at C, show that
PC
between
PA
and PB.
36-1
A COLLEGE ALGEBRA
XXIII.
we
obtain a sequence called the first order of differences of the new sequence in a similar given sequence; if we treat this of the given manner, we obtain the second order of differences
sequence
Thus,
if
and so
on.
l^,
1,
2%
8,
.S^,
we have
given sequence
first
differences
7,19,37,61,
12,
6,
second differences
third differences
6,
all
0.
709
This name is given Arithmetical progression of the rth order. to a sequence whose rth differences are equal, and whose
subsequent differences are therefore
Thus,
1^, 2^, 33, 43,
.
0.
is
for, as just
shown,
its
An
of
its
ordinary arithmetical progression, 694, is of the ^rsi order, each first differences being the common difference d.
order.
710
The nth term of an arithmetical progression of the rth Given any arithmetical progression of the rth order
!,
a^,
d,.
3)
di,
*)
^H + ij
)
its
y^)
and
let
c?i,
d^,
denote the
first
terms of
successive
orders of differences.
We
d^,
terms of
and
The
first
a^
!,
as
a2,
a^
as,
On + i
(2)
METHOD OF DIFFERENCES
The
first
365
its first,
term of
(2) is di,
and the
first
terms of
for the first differences differences are d^, d^, second, of (2) are the second differences of (1), and so on.
term of
Hence when we have found an expression for any particular from it an expression for the corre(1), we can derive
:
by
d^,
d<^,
d^
(3)
d^.
Now
follows
:
since
d^
a^
!,
2>
Starting
a^,
Og,
as
We
have
(3),
02
= 01+
^i
di
di
-{-
Hence, by
2 =
= Og = ^4 =
^3
d^
Adding,
Hence, by
(3),
+2 +
3
r/i
a^
di
Oi
c?i
Adding,
+ d^ + 2 dz + ds + 3 + d^
(/o
and so on
indefinitely, the reckoning, so far as coefficients are concerned, being precisely the same as that given in 311 for Therefinding the coefficients of successive powers of a + b.
fore,
by
561,
(?t
= ! +
1) di
,^
^ (-l)(.-2)
Example.
^
do
^ fa-l);-(-^) ,_.
1^,
(I)
Compute
ai
2-^,
by
this formula.
Here, 708,
1,
7,
Hence
a^ =
14
= 12, 141S ^
ch
= +
dr
6.
12
3375.
Sum
of the first n
terms
of
an arithmetical progression
.
of the
711
rth order.
Let
S,,
a.
o4.j, ,
(1)
and
di, d^,
d,.
710.
366
A COLLEGE ALGEBRA
the sequence of which (1)
:
Form
ences,
0,
is
the
first
order of differ-
namely
rti
!,
+
is
Then
term
are
arithmetical
is 0,
6?i,
+ a^ + as,---, a^ ao + a, (2) + l)th term of (2), and since (2) is an progression of the ( + l)th order whose first
a2,
ai
-\-
-\-
the
(?i
and the
d^,
first
terms of
have, by
its
!,
d,.,
we
(I),
Example.
terms of P,
2^, 3^,
6.
Here, 708, n
15,
ai
=
^ 7
7,
^2
=
, 12
12, ds
d,.
Hence
712
Si5
= ,^,15.14 15H
2
15
14
13
2-3
2-3-4
G = 14400.
1.
To
find the
number
of shot
when
2 shot, the of
+
if
3,
and so
1, 3, 6,
on.
Hence,
numher
of shot
is
the
sum
n terms
of the sequence
The
5,
1, 3, 6,
is
1,
di
2, (^2
1.
Therefore, by
(II), .S
+ M_2Lzi}
1-2
+ Mn-\){yi-2)
1.2.3
_
Thus,
2.
in
n(n
+ l)(n +
1.2-3
2)
21
22/6
1540
.shot.
To
find the
number
of shot
base.
when
of a
as before,
we
The first differences are 3, 5, 7, and the second are 2, 2, Hence 12, 22, 32, is an arithmetical progression of the second order
, .
.
in
which
ai
1, cZi
3,
1/2
2.
METHOD OF DIFFERENCES
Therefore, by (II), S.
367
2
1-2-3
Thus, when n
3.
21
41
/6
2870 shot.
To
find the
number
of shot
p,
Again enumerating the shot by courses, we obtain tlie sequence 2(p+l), 3(p + 2), 4(p + 3),.... The first differences are p + 2, p + 4, p + 6, and the second differ
ences are
2, 2,
Hence
p, 2 (p
in
1),
3 (p
second order
which Oi
(II), S,.
progression of the
Therefore, by
= _ ~
np
+ ?^i^> (p +
2)
+ ^i^-^n--'^)
n(?i
l)(3p
+ 2n-2)
shot
is
1-2-3
Thus, when n = 20
andp = 5,
the
number of
20
21
53/6 = 3710.
713
tion of the
An examinatheorem respecting arithmetical progressions. formula for the nth term of an arithmetical
progression of the rth order, 710, (I), will show that if we carry out the indicated multiplications and arrange the result
n,
we can
reduce
it
to the
form
where the
coefficients
b^, ^1,
^r
^^^ independent of n.
Thus, when r
=
a
2,
we have
ai
+
.
(n
.
l)di
-^
1-2
-d^
di
^2
n2+
{di
3 '-d2)n
(ai
d^).
Therefore the terms of any arithmetical progression of are the values for 71 = 1, 2, 3,
,
^v''^
^i?^'""'
with respect to n
is r.
We
are going to
368
714
Theorem.
A COLLEGE ALGEBRA
If
(f)(x)
obtained by
1, 2, 3,
progressio7i of the
rtli
is
an arithmetical
of
numbers
/>(3),
is
.^(2),
<^(4),...
(1)
and we
are to prove that all of its rth differences are equal. Evidently the first differences of (1), namely
(2)
But ^{x
nomial in
For, by
X.
be reduced to the form of a poly1) </>(a^) Ji^ay Its degree is r 1. Call this polynomial ^i {x).
<^ (a;
1)
<^ (cr)
for
a;
1, 2, 3,
.the
binomial theorem,
561,
we have
(6oX'-
0(x
l)-<?i(x)
6iX'-i
--.)
-
(M''+M'"-^ +
Similarly,
^l
if
we
</>l
write
(-^O
{X
+ 1) -
<^2
G^),
<^2
(--K
1)
<^2 (')
</>3
(a^),
a;
and
2, 3,
cj>.,(x),
^3(.r), ,
^,. (.r)
for
1,
But
(f>,.(x)
rth differences of (1). a constant and the ?'th differences of (1) are
,
For the degree of <^2(^) is (r therefore equal. 1) 1, or 3 and finally that of <{>r(x) is r 2 that of ^3(.r) is r
r,
or
0.
For example,
<Pi (x)
if
(x)
^2{x) 03
(x)
(X
1)
mp:thod of differences
Hence the values
are the
firKt,
369
of
6x2
+ 6x +
1,
l2x
12,
12 for x
1,
2, 3,
2x'
second, third differences of tlie corresponding values of and the third differences are equal, all being 12.
Thus, for X
1, 2, 3, 4, 5,
we
15,
find
62,
2x3-x + l= 6x2 + 6x + l =
12
2,
(1) (2)
13,
37, 73,
12x4-12=24,36,48,
72,
(3)
(4)
12,
12,
12,
12,-...
And by comparing
Corollary
1.
find that (2), (3), (4) actually second, tliird differences of (1).
we
The
form an
715
For
1'",
2'',
3'',
rth degree
(x)
x''
are the values of the rational integral function of the '. for x = 1, 2, 3,
The products of the corresponding terms of two arithmetical progressions, the one of the vth order and the other of the sth order, form an arithmetical progression of
Corollary 2.
716
the (r
s)^/i
order.
of a rational integral function of the rth degree by one a rational integral function of the (r -f s)th degree.
EXERCISE LX
1.
first
twenty terms of
the sequence
2.
first
eighty terms of
the sequence
3.
gressions.
(1) (3)
3,
0,
1,
0,
3, ,
(2)
(4)
1332,
..
the
Also find the eighteenth term of (1), the twentieth term of twelfth term of (3), and the tenth term of (4).
4.
(2),
What
?
is
the order of
of its first
3,
4, 3
5,
What
is its
nth
term
sum
4 22, 2 6
32,
3 8 42,
370
5.
A COLLEGE ALGEBRA
Find the number of shot
in a triangular pile of fourteen courses. shot are there in the lowest course ?
fifteen courses six courses
How many
6.
If
how many
7.
if
be removed,
the
shot are there in a rectangular pile whose top contains 9 balls and whose bottom course contains 240 balls?
11.
How many
row
Show
Show
that 13
that
+
+
2-
+
+
+ +
?x^
(1
+
{n
n)2.
12.
\*
2*
n* ^ ~
30
^){2n
\) (3
n^
+ Sn-
1).
13.
What
we
is
the order
?ith
of the first n
terms of the
progression whose
14. in
If
term
1 ?
n {n
1)
()i
2)
/6
write
down
first
order
to
which d
1, 2, 3,
terms,
we
numbers
1,
3,
6,
10,
1,
4,
9,
16,
1,
5,
12, 22,
Show
that in the
A;tii
of the first
respectively.
2)/2 and
n(ji
1)
{kn
S)
/O
15. Show that the order of an arithmetical progre.ssion of any order is not changed by adding to its terms the corresponding terms of an arithmetical progression of a lower order.
16.
Show
that
if
in
a polynomial of the nth degree. f{x), we substitute any aiithnietical progression of the first order,
;
we obtain an arithmetical progression of the nth order and, in general, if we substitute for x successive terms of any aritlimetical progression of the rth order, we obtain an arithmetical progression of the ?irth order.
that
METHOD OF DIFFERENCES
INTERPOLATION
371
Interpolation. Suppose that y is known to depend on x in such a manner that for each value of x between a and h, y has a definite value. Suppose also that the values of y which
717
This process is employed when the general expression for y in terms of x is unknown, or if known is too complicated to be conveniently used for reckoning out particular values
of
y.
:
Briefly stated, the process is as follows we set y equal to the simplest integral expression in x which will take the given
values and then derive the values of y which we seek from this Of course the values thus obtained will ordinarily equation. be only approximately correct.
Method
of
undetermined
coefficients.
We may
that y
proceed as in
718
For x
2, 3, 4, 5 it is
known
b,
4,
7,
34
when x = 5/2.
Since the simplest polynomial in x which will take given values for four given values of x will ordinarily be one of the third degree, we
assume that
2/
6o
bix
hnx^
?)3x',
and then
as follows.
5
Since y
Since y
Since
Since
2/
=b =4 1 y = - 34
when x =
when x = when x = when x =
2, 3, 4,
5,
1, &i
34
= 6o + = 60 + = 60 + = 60 +
2, 62
=
+
2/
+ 4 &2 + 9 ^o 4 61 + 16 62 5 61 + 25 62 = 4, 63 = 1.
2 6i
3 61
+ + + +
63.
63.
63.
27 64
125 b.
Hence
Therefore,
?/
=1-
2 x
4 x2
x^.
when
= 5/2 we
have
= l-5 + 25-
125/8
= 43/8.
372
A COLLEGE ALGEBRA
in general, if r
?/
And
values
?/i,
y2,
x^^y respectively,
, we
h,
+ 1 values of ij are known, say the = x^, x^, , Vr+v corresponding to x assume that
y
find
bo,
h,x
hx^
+ K^%
(1)
bi,
,
Xi,
employ
719
(1) as
by the method just illustrated, and then a formula for computing y for values of x interb,.
mediate to
a-o,
a;^
i-
Method
of differences.
When
are consecutive
integers, the
formula (1) of
,
718,
(.X
may
2/
2/1
Xo)
"^
+
where
di, ck,
1.2.-.r
denote the
iji,
?/2.
'^''
.2\ ^^
, ^^^
first
orders of diiferences of
; Vr
For since
values otbo
ical
Xi, X2,
x^ +
bix
6,.x^
namely
?/i,
ITU.
in
y.2,
by substituting n
,
1, 2,
2/r + 1, form an arithmetHence we may also obtain 2/1, the formula, 710, (I), namely
^2,
2/
yi
+ (n-l)---(n-r)
I
1.2...
in this
formula
member
x.
of (2)
and that of
(1),
718,
have equal
values of
Hence they
=
7
5, 4,
2/3,
5,
-1,
- ^^
10.
7,
34,
We
have
First differences
1, - H, 27.
Second differences
Third difference
Substituting in
ds
(2), Xi
2/
10,
6.
=-
6,
we have
METHOD OF DIFFERENCES
Example.
v'33
373
3,1748,
Given
3.2075; find
2/1,
2/2,
Vao = V3U3.
2/4
3.1072,
V3l =
3.1414,
V32 =
and
2/3,
==3.1072,
3.1414,
.0342,
3.1748, 3.2075.
.0334,
.0327.
First differences
Second differences
Tliird difference
.0008, .0001.
.0007.
Substituting in
da
(2) Xi
30, X2
31, X3
32,
?/i
3.1072, di
.0342,
=-
.0008, da
.0001,
and x
31.6,
we have
3.1072
(1.6) (.0342)
+ ILMl^ (_
.05472
.0008)
+ li:^M:M:iji) _
(
.0001)
3.1072
.000384
.0000064
3.1615 +
The formula (1) of 718 Lagrange's formula. reduced to the following form, due to Lagrange
:
may
also be
720
_ ^^
C-^'
- ^2) (x - a-a)
^2) (^1
~"
(a;
x,^,)
^r + l)
(^1
^3)
(^1
3^^4,1)
+ ^2
_l_
_
{x
Xi)
(a*
(a;
^3)
(a;
x-j)
(x
X2)
(x
x^)
For the right member of (3) is an integral function of x of ?'th degree and its values for x = x^, Xo, a-,.^! are t/i, Thus, if we set x = x^, every term except the y-2i ) Vr+ifirst vanishes and the first term reduces to ?/i. Hence, 421, the right member of (3) and that of (1), 718, are equal for ? + 1 values of x and are therefore equal identically.
the
=
(X
(2
2,
3, 4,
5,
let
=
5)
5,
4,
7,
34.
Sub-
we
obtain
_ ^~
^
3) (X
4) (X
-3)
(2
-4)
(2
-5)
(
^
(x-2)(x-4)(x-5) _
(3-2) (3-4) (3-5)
will
which
reduce to y
=1
4 x^
x^,
as in 718.
374
A COLLEGE ALGEBRA
EXERCISE LXI
1.
For X
=-
3,
2,
- 1,
it is
known
/2.
12,
that
?/
=-
20, 6, 0, 4
find y
when X = 5/2,
also
when x =
2. Given that /(4) = 10, /(6) /(x) and then compute /(12).
=-
/(7)
=-
20, /(8)
=-
18
find
3.
= 676,
272=729;
find 26.542
by the method
of differences.
4.
Given that
23
8,
3-3
27, 43
64,
53
method
5.
of differences.
.04546,
1/23
.04348,
1/24
.04167,
and
1/25
6.
by the method of
differences.
Given that
20.8566,
V435 =
ences.
7.
V432 = 20.7846, V43 3 = 20 .8087, V434 = 20.8327, V436 = 20.8806 find V435.7 by the method of differ;
By
XXIV.
LOGARITHMS
Theorem
1.
1,
and
p,
For a >
aP
> 1,
'I,
.-.
Va''
>
>
1.
'J
1,
.-.
a'l
> 1,
261.
722
Theorem
and
r,
s r
//"a denote any real number greater than any two rationals such that r > s, then a"" > a'*.
2.
.-.
1,
For
- s>0,
3.
a''-*>l,
.-.
a'--'-
a^>a^,
.-.
723
Theorem
//"a
>
and n
be integral, then
1
a"
00.
For since a > 1, we may write a Then a" = (1 + d)'', and since (1 +
Therefore, since
^''"
=
00,
d)"
(1
nd)
a"
00.
LOGARITHMS
Theorem
For
let
375
4.
//O
/6,
<a<
6
1,
and n
he intearal,
1
n
a"
0.
724
where
/ /n=
> 1,
since a
< 1.
Then lii n = oo
^^"^ 6
0,
li 512, since = oc 6
n
lii^i
, 723.
1
Theorem
1.
5.
i/'ii
Ze
integral,
v^ =
7i
li^ii^
1
=X
a"
1.
725
is
When
> 1, we have
.-.
> 1,
721,
so that a"
=1+
dn,
where d
l)/n.
lip^
li Therefore, since (a
1
- l)/n =
0, 512,
we have
0.
Hence 1 a n = 00
2.
7i
li|" (1 =X
When 0<a<l,
Then
11.
let
a"
Theorem 6. If h he a rational number and x he a variable which apjjroaches b through rational values, then ^^^ a'' = a^. 1. The theorem holds true when h, the limit of x, is 0. For in this case we can select a variable n which takes integral values \/n<x<\/n and that when only and such that we shall always have
X
'26
lie
a
i
",
722,
and since
lim
2.
an=
lim a
1,
725,
we have 1
a^
a\
true
For since a^
a^
a^-*,
when 6 t^ 0. we have 1 a^ x=
aP
1'"^ a:= b
a^-*
a^,
by
1.
Theorem
able
tvJiicli
7.
727
apj)roaches b through rational values, then aj^ ivill b, and the value of this limit is approach a limit as x independent of the values ivhich x takes in approaching b.
The reasoning
is
the
1,
but to
fix
the ideas
we
X
sequences of rational values through which approaching b as limit. From among them select some particular increasing sequence, and represent x by x' when supposed to run through this sequence. Then as x' == b, the variable a^' continually increases, 722, but remains finite less, for instance, than a'', if c denote
There are
infinitely
> 1. many
may run
in
376
A COLLEGE ALGEBRA
b.
Hence
a'-''
approaches a
limit, 192.
Call
this same limit i if x only remains to prove tliat a^ will approach b through any other sequence of rationals than that through which But a^ = a^' a^-^' and therefore lim a-^ = lim a^' lim a* *' = L, x' runs.
approach
1,
726.
728
Irrational exponents. employ the symbol a'' to denote the limit which a^ will approach when x is made to approach Hence by a*, when b through any sequence of rational values.
b is irrational,
'
We
we
shall
mean
^^^^^ a"".
x=
729
a""
we can
a*
that
a^
a*.
730
Theorem
exjjonents.
8.
For
only.
1.
let b
approach
a''
and c denote irrational numbers, and x and y variables which and c as limits. We suppose x and y to take rational values
a"
a'>
'^.
For since
W^a^'
a^ + i>,
we have
lim
a^'a'J
But
and
2.
(o*)'^
lim
a'^ay
lim a^ + y
= =
lim a^ lim ay
a^"" (-+
y^
203, 728
203, 728
a*"^.
For
(a^)^
Hence Hence
3.
(aft)'-
^i (a^)"
^'
= =
=
a^^^"^
a""'!
or
or
(a'')"
*"
"^^S
y=c^
(a'")"
^""
y=c
'"',
(a'')'^
= aK
728, 729
a'-b'^.
For
{ab)y
Hence
That
is,
\\m
{ab)!i
= = = (a6)<-
"&'.
V\m
Wbv =
Vim av
Ximbv
203
a'b^.
728
LOGARITHMS
LOGARITHMS.
Logarithms.
or
377
Take
number
of reference.
any positive number except 1, as a base We have shown that every real power
In a
731
of a, as
a'^,
subsequent section we shall show conversely that every positive number, m, may be expressed in the form a^, where /a is real.
to the base a and If a*^ = m, we call /* the loyarithin of Hence the logarithm of uirepresent it by the symbol log7. to the base a is the exponent of the power to which a must be raised to equal m, that is 0^""'" = m.
Thus, 3*
732
81,
.-.
log381
2-3
1/8,
.-.
-3 =
log.2l/8.
Since a"
=
>
1,
we always have
logl
0;
and since
a^
a,
733
= 1.
When
1, it
follows from
increase in the
;
number
734
m
if
is
greater than 1,
lies
its
loga0, its
between
and
logarithm
Again,
/x is
negative.
> 1, we have, 723, = oo, and lim a-'^ = lim 1 /a'^ = 0. We therefore say, when a > 1, that logco = x, and log0 = 00.
when a
lim a^
735
Theorem 1. The logarithm of a product to any base sum of the logarithms of the factors to the same base. m = af^, that is = loga m, For let n = a'', that is v = logafi. and mn = a'^a" = + ", Then that is, logam= + v = logam + logan.
fj.
is
the
736
a'-'
/j.
Theorem
2.
the dividend
The logarithm, of a quotient is the logarithm of minus the logarithm of the divisor.
737
For
that
if
m=
rn/n
loga rn/n
a^^
and n
a",
we have
is,
= =
a'^/a"
a*^-",
fj.
loga i
log
378
738
A COLLEGE ALGEBRA
3.
Theorem
is
the
power.
For
if
m = ai^,
ni'"
we have
that
is,
logaJn""
739
Theorem 4. The logarithm of any root of a number is the logarithm of the number divided by the index of the root.
For
if
we have
that
is,
?i
740
is due to the properLogarithms of numbers to the base 10 have been computed and arranged in tables. If we avail ourselves of such a table, we can find the value of a product by an addition, of a quotient by a subtraction, of
The
ties established in
736-739.
Thus,
log
6-
l^^
= log Vo + =
(log 5)
log
Ve -
log 325
1 736
3.
737
/7
(log 6)
/8-
25 log
738,739
have only to look up the values of log 5, log 6, and log 3 given in the table, then to reckon out 25 log 3, and finally to look up in the. the value of (log 5) /7 + (log 0) /8 table the number of which this value is the logarithm.
7,- 8^ v5 V()/3-5, we
EXERCISE LXII
1.
log2.125,
2.
logaV^,
If logio2
LOGARITHMS
379
4. Express the logarithms of each of the following expressions to the base a in terms of log6, logaC, logad.
(1)
bh'^/dK
logs
(2)
'
-v^a-Vft'^^VftWa"^31 / 18.
5.
Prove that
V 81 "^^729
X
9"
:=:
6.
+ Vx'-i - 1 = - Vx'-^ - 1
log,, (x
+ Vx2 - 1).
COMMON LOGARITHMS
For the purposes of Computation of common logarithms. numerical reckoning we employ logarithms to the base 10. Tliese are called common logarithms. In what follows log m
will
741
mean
logi?.
;
We have 10 = 1, .\ log 1 = 102 ^ 100, also log 100 = 2, 10-2 = .01, ..log.01=-2, .
.-.
10^
10"'
= =
10,
.-.
log 10
.1
.1, .-.
log
= =-
742
gers
Hence, for the numbers whose common logarithms are we have the table
:
ifite-
The numbers
their logarithms
.001,
.01,
.1,
1,
10,
1,
100,
2,
1000,
3,
2,
1,
0,
3, .
rical
Observe that in this table the numbers constitute a geometprogression in which the common ratio is 10, and the
common
743
The numbers
common
in this table are the only rationals whose all fractional powers of
10 are irrational. But, as we proceed to show, every positive number has a common logarithm, and the value of this logarithm
as
may be obtained correct to as many may be desired. If we extract the square root of 10, the
and
places of decimals
380
each case to the
Jable
:
A COLLEGE ALGEBRA
fifth
decimal figure,
we
= = = =
3.16228,
1.77828, 1.33352,
1.15478,
10^"
10"
10"^*
and so
we proceed (compare
725).
on
the left are the logarithms of the corresponding numbers on the right.
744
By
we may compute
the
common
loga-
Divide 4.26 by the next smaller number in the table, 3.16228. The quotient is 1.34719. Hence 4.26 = 3.16228 x 1.34719.
Divide 1.34719 by the next smaller number in the table, 1.33352. The quotient is 1.0102. Hence 4.26 = 3.16228 x 1.33352 x 1.0102.
last
number
in the table.
we
method an expression for 4.26 in the form taken from the table and g, the result being
4.26
of a product of
by this n numbers
= =
it
x g
n terms
108
lO^h
=
1
10^
+ s + 2h o
^^
n terms also always a part of the infiwhose sum is 1, 704, Ex. 1. Hence nite series 1/2 + 1 /4 + 1/8 + it approaches a limit which is some number less than 1, 192. Represent this limit thus 1/2 + 1/8+1 /250 + For each quotient as limit. Again, as n increases, q approaches 1 lies between the divisor used in obtaining it and 1, and as the process is
increases, the
increases.
As n
But
/8
1/256
to
1,
since
it is
Hence
l'.'"
tcrn,B^_^
= 10^ + s + sis +
and there-
l'/2"+
1/8 +
/256
.6294 ....
LOGARITHMS
381
745
From the common logarithms of the numbers between 1 and 10 we may derive the common logarithms of all other positive numbers by the addition of positive or negative integers.
Example.
1.
of 42.6
and
.426.
We
have
382
749 In like manner,
A COLLEGE ALGEBRA
if in
we
that
is,
if
we multiply n by
we add
1
/x
to the
char-
acteristic of log n.
log .0426
=-
log 4.26
1.6294,
In practice we find
characteristics
1, 2,
in the
before the mantissa, and the place the positive part 9, 8, 10 after it. Thus, instead of T.6294 we write 9.6294 10.
Hence the
If
rule
n<l,
is
negative.
to
To obtain
it,
the right of the decimal point in n, then write the result before the mantissa,
O's
number of
immediatelg
and
10 after
it.
7.6294
10, log
.000000426
3.6294
10.
less
decimal point, subtract their number from 19 and write the result before the mantissa and 20 after it and so on.
;
Example.
Given log 2
number
of digits in 2^.
750
A table of logarithms.
The accompanying
numbers of
three figures computed to the fourth place of decimals and arranged in rows in the order of their magnitude, the decimal
points before the mantissas being omitted. From this table we may also derive mantissas for numbers
of
figures
is
When
a number
very small
LOGARITHMS
383
The student will given to more than four places of decimals. find it easy to procure a five-, six-, or seven-place table.
To
find the
table.
We proceed
;
751
Example
1.
N
is
look up the first two significant figures, 58, in the column headed in the table, then run along the row to the right of 58 until the column
We
reached which
is
figure, 9.
We
The char-
10,
749.
Hence
log .00589
7.7701
-10.
46.
460 respectively.
of 800
and
.9031,
log46
Example 3. Find the logarithm of, 4673. The mantissa is the same as that of log 467.3.
:
It
must therefore
lie
between mant. log 467 and mant. log 468. From the table we find mant. log 467 = 6693 and mant. log 468 = 6702, and the difference between these mantissas is 9. Thus if we add 1 to 467, we add 9 to mant. log 467. Hence if we add .3 of 1 to 467, we should add .3 of 9, or 3 approximately, to mant. log 467. Hence mant. log 467.3 = 6693 + 3 = 6696, and therefore
log .4673
9.6696
10.
is
introduced
we omit
the decimal point which properly belongs before the mantissa. The method illustrated in Ex. 3 for finding the mantissa of
the logarithm of a number of more than three figures described as follows
:
may
to
be
the
and
the next
greater mantissa.
Multi2Jly d
bij
decimal
2)oint before
the remaining part of the number with a it, and add the integral part of the product
is .5
or more) to m.
384
A COLLEGE ALGEBRA
LOGARITHMS
385
386
752
A COLLEGE ALGEBRA
To find a number when its logarithm is given. We have merely to reverse the process described iu the preceding
section.
Find the number whose logarithm is 5.9552 10. row marked 90 and iu the column marked 2. Hence the required sequence offigures is 902. But since the characteristic is 6 10, the number is a decimal with
Example
1.
We
5,
number
Example
Looking
is
in the table
we
find that the given mantissa 5520 lies between 5527 corresponding to 356 and 357 respectively.
The
lesser of these mantissas, 5514, differs from the greater, 5527, by 13 and from the given mantissa, 5520, by fi. Thus, if we add 13 to the mantissa 5514, we add 1 to the number 356. Hence if we add 6 to the mantissa 5514, we should add 6/13 of 1, or .S
approximately, to 356. Hence the required sequence of figures is 3565, and therefore by the rule for characteristic, 748, the required number is 35650000.
therefore have the following rule for finding the sequence of figures corresponding to a given mantissa which is not in the table
:
We
Find from
m from
and divide
the remainder
hy
d,
obtained.
753
Cologarithms.
The cologarithm
of a
number
is
the logarithm
= log 1 /? = log 1 log = log m, 733, Since colog we can find the cologarithm of a number by merely changing the sign of its logarithm. But to avail ourselves of
737,
the table
positive.
we must keep
We
all
logarithms
LOGARITHMS
Example
"We have
1,
38T
Find colog
89.2.
log
and
log 89.2
= =
10
10
1.9504
Hence
colog
2.
89.2= 8.0496-10
Example
We
and
have
=
=
10
10
10 .0320
log 929
.
9.9680
Hence
colog. 929=
Hence we may find the cologarithm of a number from its logarithm by beginning at the characteristic and subtracting each figure from 9 until the last significant figure is reached,
which figure must be subtracted from 10. do or do not afiix 10 according as 10
to the logarithm.
To
is
this result
is
we
not or
affixed
In this
may
Computation by logarithms.
serve to
ucts,
The following examples will show how expeditiously approximate values of prodquotients, powers, and roots of numbers may be obtained
754
740).
5.26.
Log
But
(.0325
X .6425 x
5.26)
log .0325
log .6425
log 5.26.
log .0325
8.5119
10
log
log
log of product Therefore the product
.6425= 9.8079-10
5.26=
.7210
Hence
=
is
19.0408
20
9.0408
10
.1099.
Example
2.
Log But
(46.
72/. 0998)
log 46.72
log 46.72
log
-;
388
A COLLEGE ALGEBRA
write log 46.72, that
its
it.
We
to
is
1.6695, in the
form 11.6695
10 in order
is
make
10 which
to be
subtracted from
Example
3.
--
806.
806)
log 295
log .05631
colog 806.
= 2.4698 = 8.7506-10
7.0937
colog
806=
Hence log
of required result
=
is
18.3141
-10 - 20 =
8.3141
10
.02061.
= =
9.8992
10
Hence
Therefore (.7929)6
log (.7929)0
59.3952
60
9.3952
10
.2484.
Example
5.
Log
But
-s/.
00898
(log .00898)
7.
log .00898
= =
7.9533
- 10
-10
7)67.9533-70
Hence
log V. 00898
9.7076
Therefore V. 00898
.510.
Observe that when as here we have occasion to divide a negative logawe add to its positive and negative parts such a multiple of 10 that the quotient of tlie negative part will be 10.
rithm by some number,
Negative munbers do not have real logarithms to the base 10 since all real powers of 10 are positive numbers. If asked to find the vahie of an expression which involves negative
factors,
we may
first find
Thus, if the given expression were 456 x ( - 85.96), we should first find the value of 456 x 85.96 by logaritiims and then attach the sign to the
result.
LOGARITHMS
EXERCISE LXni
Find approximate values
1.
389
of
llie
390
Example.
A COLLEGE ALGEBRA
Find the logarithm of .586
to the base 7.
^"g^"-^^^
logioT
= 9.1619-10 ^ _
.8451
2321
8451
^_
^^^^^
We reduce 9. 7679
756 757
is
10 to the
form
of a single negative
number, namely
.2321,
final division
by logarithms.
When
l/log^b.
is
The only base besides 10 of which any actual use a certain irrational number denoted by the letter
is
made
whose
approximate value
connection.
2.718.
Logarithms to
this
base are
in another
758
the
Equations in which Exponential and logarithmic equations. unknown letter occurs in an exponent or in a logarithmic
expression
may sometimes
1.
be solved as follows.
Example
=
x
14^+".
5) log
13
(x
7) log 14.
7 log 14 2
5 log 13 5
=
14
2.4532
1.0817
= +
2.268.
_^
2 log 13
2.
-log
Example
Vx
21
\ log x
= 1.
By
736, 739,
we can reduce
log
form
Hence
Solving,
25 or
4.
Example
3.
=
+
10 x.
1.
Taking logarithms,
Solving for log
x,
Hence
= = x =
log x
1
or
1/2.
1/VlO.
10 or
759
Compound Suppose that a sum of P dollars is put at compound interest for a period of n years, interest being compounded annually and the interest on one dollar for one
interest.
first
year will be
it
P + Pr
or
P (1 +
r),
will
LOGARITHMS
I
391
be P(l + r) (1 + r) or P(l + ry, and so on. Hence, if denote the amount at the end of the wth year, we have
A = P (1
If interest be
if
ry.
quarterly,
= P (1
= P (1 +
-{-
r/2)-";
We
call
we can
find
oi A.
li A, n,
and r be given,
(1
formula
P=
^4
r)~".
We
have
log^ =
log 2500
18 log 1.04
3.7039.
Hence
A = $5057,
approximately.
Example
of $120
is
sum compound
At the beginning of each of ten successive years a premium 2. What is the worth of the paid on a certain insurance policy. of these premiums at the end of the tenth year if computed at 4%
interest ?
is
by
^
* 701,
120 x 1.04 x
..
^(i.o4)io],
1
1.04-1
-=-
By
logarithms, (1.04)io
1.479.
Hence the required value is 120 x 1.04 x .479 puted by logarithms, gives $1494, approximately.
Annuities.
.04;
and
this,
com-
sum
of
money which
is
to be paid at fixed
760
an annuity of
dollars payable annually for n years, beginning a year hence, the interest on one dollar for one year being r. The present worth of the first payment is ^ (1 r)~\ that
of the second
and so on. Hence the present worth of the whole is, 701,
payment
is
(1
r)~^,
lit;
"^
(1
+
ry
"^
(1
+ ry\
^
7L^
~
(i
If the annuity hejwrpetual, that is, if n cc, then (1 r)" and the formula for the present worth reduces to A /r.
+ ry\' = oo,
392
A COLLEGE ALGEBRA
Example. What sum should be paid for an annuity of $1000 payable annually for twenty years, money being supposed to be worth 3% per
annum ?
The present worth, P,
is 1
.03
(L 03)20 J
By
logarithms,
we
1.803.
Hence P
1000 r, 1 .03 L
1000 X .803
.03
1.803J
X 1.803
"^
EXERCISE LXIV
1.
Find
2.
729.
(2)
a^-
+2
a^^.
(3)
213^
516-^+4.
3.
log X
log (X
3)
1.
(2) log x2
(3)
log
(1-2 xY -
log (3
x)3
6.
(4) xios^
+ log z = 2. = 2.
5% compound
4.
.
in thirty-five years at
5. Find the amount of $5500 in twenty years at the interest being compounded semiannually. 6.
3% compound
interest,
Show
that a
it
sum
of
will increase
money will more than double itself in fifteen more than a hundredfold in ninety-five
compound
sura will
interest.
What
amount
to
$1250
if
A man
all
annum on
What
10.
and
c
denote the length of the hypotenuse of a right-angled triangle the lengths of the other two sides, b = V(c + a){c ~ a). Given 586.4, a = 312.2, find b and the area of the triangle, using logarithms.
If c a, b
393
If
ft,
6,
(rt
Knd
12.
a) (s c)/2, the area of the triangle is ^s{s the area of the triangle in which a 410.8, b = 424, c
b) (s = 25.68.
c).
of
Find the area of the surface and the volume of a sphere the length whose radius is 23.6 by aid of the formulas S 4^ 7tr'\ V = i ffr^/3, assuming that tt = 3.1416.
XXV.
of
letters to be given, as a, b,
k,
761
any kind.
Any
order,
more
called a combination of the n letters v at a time, or, briefly, an v-combination of the n letters.
shall use the
We
combinations.
Thus, the'2-combinations of the four letters
ab, ar, ad, be, bd, cd.
a, &, c,
There are
Ci
6.
in a definite order in a
the other hand, any arrangement of > of these Ji letters row is called a permutation of the n letters, r at a time, or, more briefly, an T-pe7'rmitation of the
letters.
On
We
symbol
P".
to denote the
number
d are
of such
permutations.
Thus, the 2-permutations of the four
ab,
ac,
letters a, b,
c,
There are twelve of these permutations, that is Pg = 12. Observe that while ab and ba denote the same combination, they denote
different permutations.
a. b,
In what has just been said it is assumed that the letters -, k are all different and that the repetition of a letter
394
A COLLEGE ALGEBRA
is
not allowed.
This where
762
We
had
ocf^asion to
ways, and if, when it has If a certain thing can be done in been done, a certain other thing can be done in n ivays, the entire number of ways in which both tilings can be done in tlie order
stated is
mn.
:
We reason thus
there are n
first
way
ways of doing both things, for in ways of doing the thing there are mn ways of doing both things.
More generally, if a first thing can be done in m ways, then a second thing in n ways, then a third in p ways, and so on, the entire number of ways in which all the things can be done
in the order stated
is in
-n-p
-.
can be
We may clioose any one of the nine digits for the first figure of the number, then any one of the remaining eight digits for its second figure, and finally any one of the seven digits still remaining for its third figure. Kence we may form 9 8 7, or 504, numbers of the kind required.
763
The number
of r-permutations of
n different
letters.
By
the
reasoning employed in the preceding example that this number Pj! is given by the formula
we
readily prove
w (
1) (n
2)
to r factors.
(1)
forming an r-permutation of n letters we may choose any one of the n letters for its first letter, then any one of the remaining n \ letters for its second letter, then any one of
in
For
the n
2 letters still remaining for its third letter, and so on. Hence, 762, the entire number of ways in which we maj choose its first, second, third, rth letters, in other wordf
395
number
n
of
ways
in
letters, is
1) (n
t^jB
2)
?'-permuto r factors.
c,
letters a, 6,
d, e
one, two,
Pi
5,
P|
4,
P-^
3,
P^
2,
P^
1.
Evidently the
is
11
(r
2)
may be
(2)
written
P = n{n- 1) (n -v2) ...(ii-r + 1). = n, the factor n r \ is n n + \, When r V) n. 2 -1, ov 1 -2 we have PI = n{n 1) {n
-\!
or
1,
and
con-
The
n is called factorial n and is denoted tinued product 1-2 by the symbol n or \iu Hence the entire number of orders in which we can arrange n letters in a row, using all of them in each arrangement, is given by the formula
Pl
= nl
later, 775,
1.
(3)
the meaningless
is
Example 1. How many different signals can be made with four flags of different colors displayed singly, or one or more together, one above
another
?
There
will
flags
taken
1, 2, 3,
or 4 at a time.
is
Pj
P^
P3
P\, or 64.
Example
taken
(1)
all
2.
Of the permutations
word fancies
at a time,
How many
first
place may be filled in 4 ways, then the last place in 3 ways, then the intermediate places in 5 ways. Hence the required number is 4 -3 -5!, or 1440.
!
The
(2)
How many
in the
odd places
even places ? even places in 3 ways, the conways, and each arrangement of vowels may
in the
in the
number
is
!,
or 144.
396
(3)
A COLLEGE ALGEBRA
How many
c is
do not have c as their middle letter ? the middle letter in (i of the permutations, for the remaining letters may be arranged in all possible orders. Therefox'e the number of the permutations in which c is not the middle letter is 7 6 !,
Evidently
! !
or 4320.
Example
Example
3.
Show
If
that
P^
Pl,
and that
P'g"
P^.
4.
5.
P\"
Example
How many
In
Example
taken
all
G.
letters a,
?
e, i, o,
u, y,
stand together
Example
ments
With
each can be formed in which the odd places are -i occupied by consonants ?
of five letters
V,
Example
form P^'
8.
Show
1, 2,
it is
possible to
figures.
Example
Example
if
all
told
row row ?
764
Circular permutations.
The number
of different orders in
which n
ference of a circle or any other closed curve is (re 1)!. For the relative order of the w letters will not be changed if we shift all the letters the same number of places along the
curve.
Hence we
11
of the
of the letters fixed in position and the r(;maining n 1 then arranged in all possible orders. ])Ut these n 1 letters can be arranged in (ii 1)!
letters if
we suppose one
orders,
7G3, (3).
7
!,
or 5040,
Example
n
1.
Show
that the
number
of circular r-permutations of
different letters is
P^/r.
397
2.
show that (n
)i
1)
fact that a circular ring will come revolved about a diameter through an angle !/2 different necklaces can be formed by stringing
together
Example 3. In how many ways can a party of four ladies and four gentlemen be arranged at a round table so that the ladies and gentlemen
may occupy
alternate seats?
of different letters
Permutations
when
765
With
11
different letters
we can form
if
n''
arrangements or per-
For in forming a permutation of this kind we may choose any one of the n letters for its lirst letter, and then again, since repetitions are allowed, any one of the n letters for its second letter, and so on. Hence, 702, the entire number of to ways in which we can form the permutation is n -n- n
r factors, or
??'".
1, 2, 3,
we can form
all
numbers
Example
1, 2, 3, 5, 7 ?
Example
boys
if
2.
each boy
The number
different.
of
all
766
Let us inquire how many distinguishable permutations can be formed with the letters a, a, a, h, c{l), three of
alike, all the letters being
which are
which are
different.
take any one of the permutations of (1), as abaca, and, leaving b, c undisturbed, we interchange the a's, we get nothing new. But if we treat the corresponding permutation of (2),
If
we
namely
aba'ca'\ in a similar
manner, we obtain 3
distinct per-
mutations, namely aba'ca", aba"ca', a'ba"ca, a'baca", a"haca', a"ba'ca. Hence to each permutation of (1) there correspond
398
A COLLEGE ALGEBRA
(2).
the reasoning here emjjloyed we can prove in general that the number of distinguishable w-permutations of n letters
By
of which
and so
on, is given
by the
formula
N-.
''
2)\q\---
Example
1.
In
how many
?
different
letters of the
word
independence be arranged
Of the 12 letters in this word 4 are Hence the required result is 12 !/4
e's,
!
word Antioch be arranged without changing the relative order of the vowels or that of the consonants ?
Example
2.
In
From
it is
arrangements is the same as it would be and the four consonants were the same.
evident that the required number of if the three vowels were the same
Hence
it is
!/3
!,
or 35.
Example
3.
How many
?
functions of five
terms has each of the following symmetric variables x, y, z, w, v, namely, 'Ex^y'^z, 'Zx^y'^z^, Zx^yzu,
and Zx^y^zhi^
terms of Zx^y-z once each if, leaving the expowe write under them every 3-permutation of the letters x, y, z, u, v. Hence the number of the terms is P^, or 60. If we apply the same method to "Ex^y-z^, we obtain the term x^y-z^ twice, once in the form x^y-z- and once in the form x^z-y'^. Similarly every term is obtained twice, once for each of the orders in which its letters under the equal exponents can be written. Hence the number of
nents
3, 2, 1 fixed in position,
We
terms
in
Zx^-Z"
is -P;]/2,
or 30.
!,
P]/3
or 20, terms,
!2
!,
Example 4. In how many ways can five jjennies, six five-cent pieces, and four dimes be distributed among fifteen children so that each may
receive a coin
?
In a certain district of a town there are ten streets running north and south, and five running east and west. In how many ways can a person walk from tlie southwest corner of the district to the northeast corner, always taking the shortest course ?
Example
5.
399
767
n different
letters.
This num-
^ (n = P"^-i-rl = -^
,
77
1)
^-
(n
f
r + 1) =^
^-
(1)
For evidently
if
we were
to
form
all
letters of each
combination in turn
in all possible orders, >\e should obtain all the r-permutations. But since each combination would thus yield rl permutations, 763, (3), all the combinations,
yield rl
P"^ -^
rl
c,
Thus, the numbers of combinations of the three, four, five at a time are
letters a, b,
d, e one,
two,
r.._5
^1
r5_5-4
^1-2
^,,_ ^
5.4.3
1-2.3
,_ 5.4.3-2
*
5.4.3.2-1
"
1.2. 3. 4
1.2.3.4-5
This expression for C". is the coefficient of the (r + l)th term in the expansion of (a + b)" by the binomial theorem, 565.
is
merely
768
another proof of the formula (1). If in the expression just obtained for C"
r)
we multiply both
obtain the more
!,
we
'"
r\{n-r)\
(2^ ^ ^
From this formula (2) it follows that the number of the /-combinations of n letters is the same as the number of the
(?i
769
?')-combinations. = (^n
r)\{7i
{n
'-
=
r)']\
{ii-r)\r\
C".
This also follows from the fact that for every set of r-things
taken, a set of n
Thus, C\\ found in this
r things
.
is left.
- CV =
14
13/1
91.
Observe
readily
Cjj is
way than by
a direct application of
400
A COLLEGE ALGEBRA
Example 1. There are fifteen points in a plane and no ttoee of these Find the number of triangles which points lie in the same straight line. can be formed by joining them.
Evidently there are as
many
Example
2.
In
(I)
A?
(2)
so
as always to exclude
A?
from the (1) The other two members of the committee can be chosen remaining nine pensons in Cl], that is, 9 -8/1 2, or 36, ways. (2) The entire committee can be chosen from the remaining nine
persons in C^, that
is,
9- 8-
7/1 2a,
e,
3,
or 84, ways.
Example
3.
i,
6, c, d, /,
of letters can be
The vowels for the arrangement can be chosen in C^ ways, the consonants in C3 ways then each selection of vowels can be combined with Hence every selection of consonants and the whole arranged in 5 ways. the required result is Cg C3 5 !, or 7200.
; !
C]'^
or
1,
way.
Hence,
7G2, the
C'g**
C'^?
Cji,
or 18!/(6!)-i.
sets of 6
how many ways the 18 books can be distributed into three books each, we nnist divide the result just obtained by 3 !, which for here the order in whicli the three sets may chance gives 18 /(e) !)3 3 to be arranged is immaterial.
find in
!
!
To
Example
5.
With the
?
different sele(^tions
letters of the word mathematical how many and how many different arrangements of four letters
As
all
different
we cannot
f(uC",.
by single applications
Uv.ncii
of the formulas
;
and
??(,
in
t,t; h,
e, i, c,
we may
classify
:
selections
and
arrangements as follows
401
Those having three like letters. 1. Combining the 3 a's with each of the seven other letters in turn, we or 28, arrangements. obtain 7 selections and 7 4 !/3 Those having two pairs of like letters. 2. 2 or 18, such arrangements. There are 3 such selections and 3 4 / 2 3. Those having two letters alike, the other two different.
!,
!,
Of such selections there are 3 Cg, or 63 of arrangements there are 63-4I/2!, or 756. 4. Those having four different letters. Of selections there are 0% or 70 of arrangements, 70 4 or 1680. Hence the total number of selections is 7 + 3 + 63 + 70, or 143 of
; ;
!,
arrangements, 28
18
756
1680, or 2482.
Cjl, C^9,
Example
6.
and
Cj^.
Example
Example Example
7.
CJ
CI',
find n.
8.
If 2
q=5
the
q,
find
ji.
9.
How many
lie in
no
four of which
same plane
Example 10. How many parties of five men each can be chosen from a company of twelve men ? In how many of these parties will a particular
man
A be
included
will
A be
excluded ?
Example 11. Of the parties described in the preceding example how many will include two particular men A and B ? How many willinclude one but not both of them ? How many will include neither of
them
?
From twenty Republicans and eighteen Democrats how 12. many committees can be chosen, each consisting of four Republicans and
Example
three Democrats ?
Example 13. With five vowels and fourteen consonants how many arrangements of letters can be formed, each consisting of three vowels and four consonants?
Example 14. In how many ways can a pack of fifty-two cards be divided equally among four players A, B, C, D ? In how many ways can the cards be distributed into four piles containing thirteen each ?
Example
15.
How many
numbers, each of
five figures,
can be formed
2, 3, 4, 2, 5, 2, 3, 6, 7 ?
402
770
A COLLEGE ALGEBRA
561,
bers,
we set a we obtain
=i=
C\
If in the formula for {a + &)", and then subtract 1 from both mem-
Q+
,
+
of
C:
2"
1.
Hence the
total
number
combinations of n different
n at a time, in other words, the total number of ways in which one or more things may be chosen from n things, is 2" 1.
things taken one, two,
This may also be proved as follows Each particular thing can be dealt with in one of two ways, that is, be taken or left.
:
Hence
2 2
the total
-.to
number
of
ways
762.
left,
factors, or
all
2",
in
which
Example
1.
How many
different
sums
of
Example 2. By the reasoning just illustrated, show that the total number of ways in which one or more things can be chosen from p + g + things of which p are alike, q others alike but different from the p things,
and so on,
^.^
is (j)
Example
3.
of
money can be
paid with
two dimes,
five quarters,
771
In the expression for C,", namely l)/r!, the r factors of the numerator Hence for decrease while those of the denominator increase.
Greatest value of C.
n(n
V)-
(?i
a given value of n the value of C" will be greatest when the next greater value of r will make
{n
r 4-
l)/r
<
1.
From this
r
it
= n/2
readily follows that if n be even, C" is greatest when" and if 7i be odd, C" is greatest when r l)/2 (^n
-\-l) /2, the value of C" being the same for these two values of r, 769.
or r
=(n
Example.
What
is
403
772
Let us inquire in Combinations when repetitions are allowed. can select three of the four digits 1, 2, S, 4
111, 112, 124,
which
all three,
two, none
respectively,
we obtain
123, 124,
3-combinations tvithout repetitions of the digits 1, 2, 3, 4, 5, 6. And a little reflection will show that if we make out a complete list of the selections like 111, 112, 124, arranging the digits in each so that no digit is followed by one of less value,
and then
add
0, 1, 2,
we
shall
obtain once and but once every one of the 3-combinations 1) or 6 digits 1, 2, 3, 4, 5, 6. without repetitions of the 4 (3
The number of the latter combinations number which we are seeking. The same reasoning may be applied
is
Hence 0% is the
to the general case of r-combinations, with repetitions, of the n numbers 1, 2, , n. And since the numbers may correspond to n different things of any kind, we have the theorem
:
The number of the x-comblnations loith repetitions ofn different things is the same as the number of the r-comblnations tvithout
repetitions o/ n
n(n
-f
l)---(n
r 1 different + r - l)/r!.
-{-
things, namely,
C"+r~S ^^
Example 1. How many different throws can be made with four dice ? As any one of the faces marked 1, 2, 3, 4, 5, 6 may turn uppermost in the case of one, two, three, or four of the dice, the number of possible throws is the number of 4-combinations with repetitions of 1, 2, 3, 4, 5, 6,
namely, C'^j^ or 126.
Example 2. How many terms has a complete homogeneous polynomial of the rth degree in three variables x, y, z? Evidently it has as many terms as there are products of the rth degree whose factors are x's, y's, or z's. Hence the number is = C7'+2 = C"+2 = C^ +
;:-i
(^
1) (r
2)/2.
404
773
A COLLEGE ALGEBRA
of combinations.
The correspond-
The following relations are of special ing algebraic identities. interest and importance.
C
C'V'
C",!].
(1)
Por we may distribute the ^--combinations of 7i letters into those which contain some particular letter, as a, two classes, and those which do not contain this letter. We shall obtain all
first class,
once each,
if
l)-combination of
the remaining n
;
we form every
and then add
is C'"z}.
letters
The combinations
the remaining n
(jm+n
is C"~'^.
c-;!
-f c,"!^
CI
c.'n^
QH
\-
<^'T
<",-i
C';..
(2)
For take any group of m -f n letters and separate it into two groups, one of ??i letters, the other of n letters. We shall
take account of
each, if
all
???
+ n letters,
once
we classify them as follows. They consist of The (a) The r-combinations of the letters of the i-group. number of these combinations is C"'. 1 letters of the (b) The combinations which contain m-group and one le.tter of the ?? -group. As we can choose the 1 letters in C^!!!i ways and the one letter in C" ways, the number of combinations of this kind is C^'"i C". 2 letters of the (c) The combinations which contain r and two letters of the n-group. As we can choose m-group
1
the r 2 letters in C^2 ways and the two letters in ways, -^.nd so the number of combinations of this kind is C,.l2 Qon, until last of all we reach the ?--combinations of the letters
C^..
8i and Cl
C'lC^
C^C^
C^
10
-f
40
30
-|-
84.
774
connecting
vi,
n,
r.
The
show that
405
when m,
n, r
algebraic identities,
and n are concerned, they are true holding good for all values of these letters.
we have
2)
This
may
nn-i
'-^
(n
1) (n
(n
r)
(n
1)
(n
2)
(n
1)
l-2.-.r
(n
1) (n
1
.
2)
(?i
1 )
l-2...(r-l) n r'
(r
1) 1)
['-^]
^
^
n (n
1)
(n
-r+
1.2.--r
But it is not necessary to make such a reduction to prove that these formulas are true identities. Thus, when expressed
in
terms of m,
n,
r,
each
member
function of vi and n whose degree with respect to each of these These two functions must be identically equal, letters is r.
since otherwise, were we to assign some particular integral value to m, thus making them functions of n alone, they could not be equal for more than r values of w, 421, whereas it has
already been shown that in reality they would be equal for integral values of n.
all
EXERCISE LXV
1.
four from
2.
roads leading from P to Q, two from Q to i?, and to S, by how many routes can a person travel from P to iS?
a
In
company
six
numbered
3.
If eight
the
first,
won
all
4.
In
men and
5.
four-oar crew be chosen from ten oarsthese crews be arranged in the boat ?
soldiers
three
how many
pickets of
6. Five baseball nines wish to arrange a schedule of games in which each nine shall meet every other nine three times. How many games must be scheduled?
406
7.
all
A COLLEGE ALGEBRA
In
the digits
1, 2, 1, 3, 2, 1, ?
5 be arranged,
letters in the
all
at a time, (1)
(2)
first
how many
how many
9.
begin with a vowel and end with a consonant? do not begin with/? (3) how many have vowels in the
three places ?
In
how many
ct,
o, i ?
of the permutations just described do the vowels In how many do the consonants retain the order
many do both
word
the vowels
With the
?
letters of the
resident
how many
first,
permutations of
five letters
third,
and
fifth letters
are vowels
11.
In
candidates of
whom
?
and nine
In
is
even be chosen
10
How many
numbers
formed
1, 2, 3, 4, 5, 6,
when
the digits
may
be repeated?
How many
can be
digits
How many odd numbers without repeated digits are there How many of these are divisible by 5 ? ?
how many ways can how many ways can
a person invite one or more of
between
In
five friends
to dinner ?
17.
In
fifteen apples
be distributed
among
three
In
six positive
and
five
written in a row
19.
characters
20.
1, 2, 3, 2, 3, 4, 2, 4, 5, 3, 6,
seven
From fifteen French and twelve German books eight French and German books are to be selected and arranged on a shelf. In how
this be
done ?
407
From
done
In
?
five are
to be selected
which
this be
In
22.
four
men
be chosen from
Americans
(2)
and
least
six
one Englishman?
at
23.
lines is
How many
met by another
parallelograms are formed when a set of ten parallel set of twelve parallel lines ?
24.
line,
except
Given n points in a plane no three of which lie in the same straight m which all lie in the same straight line. Show that the
of lines obtained
number
25.
is
C'^
C +
1.
can be formed by stringing pearls, six like rubies, and five like diamonds.
ten persons be arranged at two round
In
tables, five at
27.
In
of
game
28.
for
how many ways can six ladies and five gentlemen arrange a lawn tennis, each side to consist of one lady and one gentleman ?
In how many ways can fifteen persons vote to fill a certain office which there are five candidates ? In how many of these ways will the vote be equally divided among the five candidates ?
29.
men two of whom are qualified to and one on the bow side only. In how many
How many baseball nines can be chosen from eighteen players of ten are qualified to play in the infield only, five in the outfield only, and three in any position?
30.
whom
31.
Show
all at
taken
of permutations of six different letters of the letters are excluded each from a
!
particular position,
32.
is
!.
How many
letters p, q,
r, s, t, v,
when
33.
34.
How many
How many
Show
different
five dice ?
IfX'^y'^z'^u, I,x^y^z'^u^v,
number
of terms in a complete homogeneous function of the nth degree in four variables is {n 1) (n + 2) (n + 3) /S!
35.
that the
number
408
A COLLEGE ALGEBRA
XXVI.
775
Multinomial theorem.
nomial, and n a positive
^
+b+
k denote any
poly,
integer.
Then
Kl
al pi
"where the
sum on
/3,
the right contains one term for each 'set of n, , k that can be selected from 0, 1, 2,
being understood that when a = 0, a! -s to be replaced by 1, and the like for (3, , k. For (a -\- b -\+ k)" denotes the continued product
/S
+ K = n,
it
(a
-{-
b -{-'
k) (a-\- b
-\-
1-)
to
factors,
partial products obtained by actualty carrying out this multiplication, without collecting like terms, has the
a letter from the first parenthesis, times a letter from the second, times a letter from the third, and so on. But since the letter selected from each parenthesis m^ay be
:
any one of the letters a, b, , k, a list of the products as thus written would also be a complete list of the 7i-perniutations of
the letters a,b,---, k when repetitions are allowed. a,ft,---,K denote any particular set of numbers 0, 1, ,
And
?i
if
sum
is n,
list as
many
/>'s,
products in
of
whose which
A-'s,
a of the
of
them
, k
them
as
there are n-permutations of n letters of which a are alike, /3 others are alike, and so on, namely, 7i /n- /3 And k !, 76G.
is
equal to a%^
k",
their
sum
is
n'
.
a
j3l
0'
al
Ki
is
e)*
a* with the coefficients 4 !/l 1 1 1 I, or types abed, a^bc, a'^b'^, 24; 4!/2!l!l!, or 12; 4!/2!2!, or 6 4!/3!l!, or 4; 4!/4!, or 1, terms of the same type, we have respectively. Hence, uniting
!
a%
(a
(i
e)*
2a4
2a6
Sa'-ft^
12 -Ea-bc
24 Eabcd.
PROBABILITY
Example. Find the
coefficient of x^ in the
409
expansion of (2+3x4-4x2)8.
8!
is
or !/3
!
2''3^4>' x3 + 2 v, !7 are those for which 8 (1), and the terms required where a + ^ + 7 = 5 (2). complete list of the solutions of (1), (2) in positive |3 -f- 27 Hence the required 4, 3, 1; 5, 1, 2. integers orO's is a, )3, 7 = 3, 5,
coefficient is
'
23
35
3151
-^
2*
33
4!31v
-^
5!2!
25
42,
or 850,752.
EXERCISE LXVI
1.
+
6
+
c
-\-
d)3,
same
2.
3.
d)^.
Find the
and
a^hH"^ in the
expansion
of (a
4.
5.
6.
+c+
Find the
Find the
expansion of (a
dy^.
expansion of (a
(1 (1
36
2 cf.
expansion of expansion of
7.
in the
+ x + x2 + x^)!". x + 3x2)9.
XXVII.
PROBABILITY
SIMPLE EVENTS
Consider any future event which, if given a an opportunity to happen, must happen or fail to happen in one of a limited number of ways all equally likely, that is, ways so related that there is no reason for
Probability.
776
trial,
that
is,
The turning expecting any one of them rather than any other. of the ace uppermost when a die is thrown is such an event.
there
For one of the six faces of the die must turn uppermost, and is no reason for expecting any one face to turn rather
than any other.
410
A COLLEGE ALGEBRA
can happen or
Calling all the equally likely ways in which such an event fail the possible cases with respect to the event,
the ways in which it can happen the favorable cases, and the ways in which it can fail the unfavorable cases, we say
:
The probability or chance of the event is the ratio of the number of favorable cases to the entire number of possible cases, favorable and unfavorable.
Hence if m denote the number of possible cases, a the number of favorable cases, and p the probability, we have by
definition
p
m
a I m.
when a
die
is
Thus, the probability that the ace will turn up = 6 and a = 1. 1/6 for here
;
thrown
is
Again, the chance of drawing a white ball from a bag white and two black, is 3/5.
1.
known to
contain
777
Corollary
is 1
;
If an event
is certain to
happen,
;
its probabiliti/
if
its
it is
in every other
case
probability is a positive
fraction.
the event is certain to happen, there are no ways in can fail; hence a and a/m 1. If the event is certain to fail, there are no ways in which it can happen hence a = and a/m = 0. In every other case a is greater
if
it
For which
=m
than
and
less
than
7n,
so that
a/m
is
a positive proper
fraction.
778
Corollary 2.
ivill
happen
is
p.
For if a of the possible cases favor the occurrence of the a cases favor its non-occurrence. event, the remaining Hence the probability that the event will not happen is
m =
(m
779
a) Odds.
/m =
If the
a/m, number
and
certain event is a
we
say,
when a
> b,
PROBABILITY
event
;
411
when
when a
b,
b a, that the odds are b to a against the event; In the first that the odds are even on the event.
>
b), is
greater
it is
less
than 1/2
in the third it
equal to 1/2.
Thus, if a ball is to be drawn from a bag containing five balls, three white and two black, the odds are 3 to 2 in favor of its being white, and
3 to 2 against
its
being black.
Expectation,
li of
a certain
sum
money M,
denote the chance that a person will win the product Mp is called the value
780
sum
is
concerned.
is
Thus, the value of the expectation of a gambler who throws an ace with a single die is 12 x 1/G, or $2.
to
win $12
if
he
Examples
of probability.
must be taken
781
The following examples will to such as are equally likely. illustrate the need of this precaution.
Example
1.
If
two coins be tossed simultaneously, what is the chance two tails? one head and one tail?
:
We
the
result
first result,
is
might reason thus There are three possible one the second, one the third hence
;
cases,
tlie
1/3.
false, since the number of equally likely For if we name the coins and B possible cases is not three but four. tail, B tail ; head, B head respectively, the equally likely cases are And since one of these cases favors head, B tail; tail, B head.
the result two heads, one the result two tails, and two the result one head and one tail, the chances of these results are 1/-1, 1/4, and 2/4
respectively.
2.
"What
is
i,
Here the number of equally likely possible cases is 6 0, or 36, for any may turn up with any face of the other die. We have a total of eight if the faces which turn up read 2, 6 or 3, 5 or 4. But there are two ways in which 2, 6 may turn up, namely 2 on the
412
die
A COLLEGE ALGEBRA
A and 6 on the die B, or vice versa. Similarly there are two ways in which 3, 5 may turn up. On the contrary, there is but one way in which Hence there are five favorable cases. Therefore the 4, 4 can turn up. chance in question is 5/36.
Example
dice
if
3.
What
is
up
The number
We
or
of equally likely possible cases is 6 6 6, or 216. have a total of eight if the faces which turn up read 1, 1, 6 or
1, 2,
1, 3, 4.
But
6
1, 1,
numbers
1, 1,
may
turn up in 3 !/2 !, or 3, ways, 766 for the be distributed among the three dice in any of the
may
orders in which
1, 1, 6 can be written. Similarly 1, 2, 5 and 1, 3, 4 may each turn up in 3 !, or 6, ways. Hence there are 3 + 6 + 6, or 15, favorable cases. Therefore the chance in question is 15/216, or 5/72.
Example
balls.
4.
An
urn contains
six white,
(1) If four balls are drawn, what is the chance that all are white ? There are as many ways of drawing four white balls as there are 4-combinations of the six white balls in the urn, namely C^. Similarly, since the urn contains twelve balls all told, the total number of possible drawings is C'4. Hence the chance in question is C^l/ C^^, or 1 /33. (2) If six balls are drawn"what is the chance that three of them are white, two red, and one black ? The three white balls can be chosen in Cg ways, the two red balls in C* ways, the one black ball in Cf ways. Hence the number of ways in which the required drawing can be made is C3 Cg Cj. The total number of Hence the chance in^^uestion is C'l- C^- C{/ C^g, possible drawings is C'^.
or 20/77.
Example
(1)
5.
is
the chance that neither king nor queen is Aside from the king and queen there are eleven cards.
What
Therefore C'g sets of three cards which include neither king nor queen. the probability in question is C\J / C]^, or 15/20. (2) What is the chance that king or queen is drawn, one or both?
Hence
the probability in question is 1 - 15/26, or 11/26, 778. (3) Wliat is the chance that botli king and queen are drawn?
We obtain
if
we combine each
every set of three cards which includes both king and queen of the remaining eleven cards in turn with king and
in question
is
queen.
ll/C'g, or 1/26.
PROBABILITY
413
782
On the various meanings of probability. 1. The fraction ajm, which we have called the probability of an event, 776, means nothing so far as the actual outcome of a single trial, or a small number of trials, of the event is concerned. But it does indicate the frequency with which the event would occur in
the long run, that
series of trials.
is,
in the course of
an indefinitely long
Thus,
if
great number of times, say a thousand times, he will find that as the number of throws increases the ratio of the number of
times that ace turns up to the total number of throws approaches the value 1/6 more and more closely.
life is
There are important classes of events the duration of one to which the definition of 776 does not apply, it being impossible to enumerate the ways, all equally likely, in which the event can happen or fail. But we may be able to determine the frequency with which events of such a class have
2.
occurred in the course of a very great number of past trials. If so, we call the fraction which indicates this frequency the Like 1/6 in the case of probability of an event of the class.
the die,
class
it
may
indicates the frequency with which events of the reasonably be expected to occur in the course of a
very great number of future trials. Thus, if we had learned from the census reports that of 100,000 persons aged sixty in 1880 about 2/3 were still living
in 1890, we should say that the probability that a person now sixty will be alive ten years hence is 2/3. 3. But we also use the fraction a Jm to indicate the strength
of our expectation that the event in question will occur on a s'luijle trial. The greater the ratio of the number of favor-
number
frequency with which, to our knowledge, events of a similar character have occurred in the past, the stronger is our expectation that this particular event will occur on the single trial
under consideration.
414
A COLLEGE ALGEBRA
speak of the probability, in this sense, of any kind Thus, before a game between two football and B, we hear it said that the odds are 3 to 2 in
We may
of future event.
teams, A favor of A's winning, or that the probability that will win is 3/5. This means that the general expectation of A's winning is about as strong as one's expectation of drawing a white ball
which are
EXERCISE LXVII
1. The probability of a certain event is 3/8. Are the odds in favor of the event or against it, and what are these odds? What is the probability that the event will not occur ?
2.
The odds
is
What
3.
is
The odds
What
his expectation ?
:
4. The French philosopher D'Alembert said "There are two possible cases with respect to every future event, one that it will occur, the other that it will not occur. Hence the chance of every event is 1 /2 and the
How should
he be answered
An
If
urn contains sixteen balls of which seven are white, six black,
red.
is
and three
(1)
it
is
white?
black? red?
two balls be drawn, what (2) one white and one red ?
If
(3)
what is the chance that one white, one black, one red ? (4) If four balls be drawn, what is the chance that one the rest not ? two white and the other two not ?
If
all
are red?
none red
is
white and
(5)
If ten balls
be drawn, what
?
is
black,
6.
What
What
is
with
>
three dice?
7.
is
Show
that this
the chance of throwing a total of seven with two dice? is the most probable throw.
PROBABILITY
8.
415
What
?
is
two dice
9.
J
the chance of throwing at least one ace in a throw with and but one ?
One
banter.
10.
What
taken at random from each of the words factor and same letter is taken from each ?
box contains nine tickets numbered 1, 2, 9. If two of the drawn at random, what is the chance that the product of the numbers on tliem is even ? odd ?
tickets be
Vil.
1, 2,
(3)
If five tickets be drawn from this box, find the chance (1) that and 3 are drawn (2) that one and but one of 1, 2, and 3 is drawn that none of these numbers is drawn.
;
12.
If
is
four cards be
of fifty-two cards,
?
what
the chance that they are ace, king, queen, queen, and knave of the same suit?
and knave
ace, king,
lis. What
is
hand
trumps and
What
is
dice ?
15.
is
together
'
Two
or
more events
783
them
rest.
is
the
not affected by the occurrence or non-occurrence of In the contrary case the events are said to be
interdependent.
Thus, the results of two drawings of a ball from a bag are independent is returned after the first drawing, but interdependent if the
if
the ball
Theorem 1 The prohahilitij that all of a set of independent events will occur is the produc t of the probabilities of the single
.
784
.
events.
'"^
For consider two such events whose probabilities are i/mi and a^]m^ respectively.
416
Tlie
A COLLEGE ALGEBRA
number
event
of equally likely possible cases for and against for and against the second vx^, and since
the
first
is ?i,
the events are independent any one of the m^ cases may occur with any one of the m^ cases. Hence the number of equally
for likely possible cases
the same reasoning, a^a^ of these Therefore the probcases favor the occurrence of both events.
events
is ^riimz.
And by
is
711
~
>
that
is
-^
>
as
im2
nil
^2
of
is similar.
applies only to events of the kind described in 776, but for the reasons indicated in 782 we may apply the theorem itself to any kind of future event, as
in Ex. 2 below.
The demonstration
Thus, the chance of throwing ace twice in succession with a single die
is
Again, the chance of twice drawing a white ball from a bag which contains five white and four black balls, the ball first drawn being returned
before the second drawing,
is
5/9 x 5/9,
or 25/81.
785
If the probability of a first event is pi, and if nfter this event has happened the probability of a second eveiit is P2, the j^robability that both events will occur in the order Theorem
2.
stated is piP2.
And
tivo events.
This theorem
may
be proved
in
the
same manner
as
Theorem
1.
It evidently
Thus, after a white ball has been drawn from a bag containing five white and four black balls, and not replaced, the chance of drawing a
second white
ball
ball is
4/8.
when
the one
1.
first
Hence the chance of twice drawing a white drawn is not replaced is 5/9 x 4/8, or 5/18.
the chance that ace will turn up at least once in
?
Example
"What
is
Ace will turn up at least once unless it fails to turn in every throw. The chance of failure in a single throw being 5/6, the chance of failure in all three throws is 5/6 x 5/6 X 5/6, or 125/216. Hence the chance of
at least
one ace
is 1
125/216, or 91/216.
PROBABILITY
Example The chance
problem
2.
417
problem
is
The chance
that
it
A
is
3/4.
?
that
will solve
if
2/3.
What
is
it
both
and B attempt
independently
The problem
A's
is
failure
solved
1/4, of B's failure 1/3. Therefore the chance that the problem will be --11/12.
is
and B fail. The chance of Hence the chance that both fail
There are two purses, one containing five silver coins and 3. one gold coin, the other three silver coins. If four coins be drawn from the first purse and put into the second, and five coins be then drawn from the second purse and put into the first, what is the chance that the gold coin is in the second purse ? in the first purse ?
Example
is
C\/ C% or 2/3,
is
781,
taken from the first purse and put Ex. 5. The chance that it is then
is
second purse
it
^ / C\^ or 2/7.
is
in the
second purse
is
is
4/21, or 17/21.
is
Example
4.
If eight coins
them
will
the chance
If are hunting quail. Example 5. Four men A, B, C, and gets on the average one quail out of every two that he fires at, B two five out of every seven, out of every three, C four out of every five, and wliat is the chance that they get a bird at which all happen to fire
simultaneously
6.
An urn A contains five white and four red balls. A secExample ond urn B contains six white and two black balls. What is the chance of drawing a white ball from A and tlien, this ball having been put into B, of drawing a white ball from B also ? 'o
Example
B, 5/6.
alive ?
7.
What
A will be alive five years hence is 3/4; the chance that five years hence both A and B will be
dead,
alive,
B dead ? A
alive ?
both dead
If two or more events are so them can occur, they are said to be
786
mutually exclusive.
Thus, the turning of an ace and the turning of a deuce on the same throw of a single die are mutually exclusive events.
>
-1
418
787
A COLLEGE ALGEBRA
3.
Theorem
of mutually exclusive
'probabilities
will
occur
is
For consider two mutually exclusive events A and B. The possible cases with respect to the two events are of three kinds, all mutually exclusive, namely, those for which (1) .4
happens, B fails (2) .1 fails, B happens (3) A fails, B fails. Let the numbers of equally likely possible cases of these three kinds be I, ni, and n respectively. Then
;
(a) ^ ^
either
6v
happens ^^
I
is
I
}n
n possible and
+m
1
+m+
r7i)
favorable cases.
;
The chance
is
+ (m +
For since A never happens except when B fails, the I cases in which A happens and B fails are all the cases in which A happens, and the m + n cases in which A fails and B happens
or both
A and B
fails.
B is
m + (^ + ^0
sum
But
^l^L_ = ^_^i + ^+ m + + m n + (? n) (^
I -\-\-
-^
?i)
ov
B happens
is
the
of the chances of the single events A and B. The proof for more than two events is similar.
Thus, if one ball be drawn from a bag containing four white, five black, and seven red balls, since the chance of its being white is 1/4, and that of its being black is 5/16, the chance of its being either white
or black
directly
is 1/4 + 5/16, or 9/16. Of course this from the definition of probability, 776.
result
may
be obtained
may
3.
Care must be taken not to apply this theorem to events which are not mutually exclusive.
will
Thus, if asked, as in 785, Ex. 2, to find the chance that a problem be solved if both A and B attempt it, A's chance of success being
PROBABILITY
419
3/4 and B's 2/3, we cannot obtain the result by merely adding 3/4 and 2/3, since the two events A succeeds, B succeeds are not mutually excluThe mutually exclusive cases in which the problem will be solved sive. are A succeeds, B fails A fails, B succeeds A succeeds, B succeeds.
:
The chances of these cases are, 784, 3/4 x 1/3 or 3/12, 1/4 x 2/3 or 2/12, 3/4 x 2/3 or 6/12; and the sum of these three chances, or
11/12,
is
Example
An urn A contains ten balls three of which are B contains twelve balls four of which are white.
random and a
?
white,
If
is
one
the
ball
drawn from
it,
what
is
white
We
(2)
exclusive events
are required to find the chance of one of the following mutually and then drawing a white ball from it; (1) choosing
:
choosing B and then drawing a white ball from it. The chance of choosing A is 1 / 2, and the chance when A has been chosen of drawing a white ball is 3/10. Hence the chance of (1) is
1/2 X 3/10,or3/20.
(2) is 1
/2 x4/12,orl/6.
3/20
1/6, or 19/60.
Example
to have
2.
What
is
who
is
any two coins he may draw at random from a purse which contains five dollar pieces and seven half-dollar pieces?
of his expectation so far as it depends on drawing two dollar $2 x C'^/C'| = $2 x 5/33 = .30; on drawing two half-dollar pieces, $1 x C^/ Cj" = $1 x 7 /22 = .32 on drawing one dollar piece and one half-dollar piece, $1.50 x 5 7/ Cf = $1.50 x 35/66 = $.80. Hence the total value of his expectation is $.30 -|- $.32 -|- $.80, or
The value
is
pieces
$1.42.
Example
at a time
balls
3. Two persons A and B are to draw alternately one ball from a bag containing three white and two black balls, the drawn not being replaced. If A begins, what chance has each of
being the
first
to
draw a white
ball ?
The chance that A succeeds in the first drawing is 3 / 5. The chance that A fails and B then succeeds is 2/5 x 3/4, or 3/10, for when B draws, the bag contains four balls three of which are
white.
fails,
when
white.
is
3/5
-f
3/10.
420
Example
chances of
4.
A COLLEGE ALGEBRA
A
and B
In the drawing described in Ex. 3 what are the respective if the balls are replaced as they are drawn ?
On
the
first
their chances
will be the
round A's chance is 3/5, B's 2/5 x 3/5, or 6/25; and on every later round, of which there may be any number,
as these.
same
Hence
"^
3/5 6/25,
:
or 5
that
is,
Example 5. In a room there are three tables and on them nine, ten, and eleven books respectively. I wish any one of six books, two of which If a friend are on the first table, three on the second, one on the third. select a book for me at random from those in the room, what is the
it is
chance that
one of those
of
wish
running horses enters for a certain race two ^^ What horses whose chances of winning are 1/2 and 1/3 respectively. is the chance that he will obtain the stakes ?
Example
6.
An owner
Example
which
is
7.
and
B throw
the one
alternately with
two dice
for a stake
to be
won by
who
first
if
winning
What
are
788
The following theorems Repeated trials of a single event. are concerned with the question of the chance that a certain
event will occur a specified number of times in the course of a series of trials, the chance of its occurrence on a single trial
being known.
789
Theorem 4. If the jjrohahUlty that an event tvill occur on a single trial is p, the probahilitij that it ivill occur exact/// r times in the course ofn trials is C"p'"q"~'', where q 1 p.
it
ular set of r trials and fail on the remaining n 1 - />, 784. 2f{l -py~'', ov 2'''(f-'', if 7
any
partic-
r trials is
??.
we may
select this
ways which, of
C'^.p''q"~'',
course, are
in question is
will turn
787.
out of
is
five
(1)2
up exactly twice in five throws with dice thrown simultaneously two and but
two
C^
(^y, or 625/3888.
PROBABILITY
Observe that
sion of
(j}
421
C^p'if''' is the
y)"
Theorem
at least
r
r+
5. The probability tliat such an event will occur times in the coarse of w trials is the sum of the first 1 terms in the exjjansion of (^ -\- q)", namely,
790
pn
C-Jp^-'q
C5p-2q2
C^'l.pV^
it
at least r times if
occurs exactly
r,
number
and
C",./''y"~'"repres'^'it
^
[
five tlirows
Thus, the chance that ace will turn up at least twice in the course of with a single die is
U)'
5 id' f
10 (l)MD^
10 (D^
(p\
or ^^,%%
in which A's skill is win as many as three such games in a se'. of five ? A's chance of winning a single game is 2/3, of losing 1 /3. Hence the chance that A will win as many as three of the five games is the sum of the
Two
persons
and
are playi;*g a game which cannot twice B's. Wliat is the chance that
will
first
three terms of (|
2.
J)5,
that
is,
Q)^
5 {V}* i
1
10 (|)3
is
{i-)2,
or 64/81.
Under the conditions of Ex. Example will win three games before B wins two ?
what
The chance in question is that of A's winning at least three of the first four games played and this chance is (?)* + 4(1)^ J, or if. And, in general, the chance of A's winning m games before B wins n is
;
the
same
winning at
least
of the first
m+
games played.
Example 3. Ten coins are tossed simultaneously. What is the chance that exactly six of tliem turn heads up ? that at least six turn heads up ? Example
4.
If
four dice be thrown simultaneously, what is the chance up? that at least three turn ace up?
in
Example
that
5.
Ex.
what
is
the chance
will
win at
6.
games played ?
is
Example
A will
wins one
422
A COLLEGE ALGEBRA
J
EXERCISE LXVni
five
1.
balls.
On
the understanding that one ball is drawn at a time and replaced as soon as drawn, what are the chances of drawing (1) first a white, then a red, then a black ball ? (2) a wldte, red, and black ball in any order whatsoever ?
2. What is the chance of obtaining a white ball in the first only of three successive drawings from this bag, balls not being replaced ?
e
3.
What
is
who
is
allowed to
dom from
(t
a purse containing
?
ai
tF'it a certain door is locked is 1/2. The key to one of a'liunch of eight keys. If I select three of these keys at random and go to the door, what is the chance of my being able to open it ?
The chance
is
the door
5. There are three independent events whose chances are 1/2, 2/3, and 3/4 respectively. -^Vhat is the chance that none of the events will occur ? that one and but one of them will occur ? that two and but two
will occur ?
6.
that
occur
in
Find the odds against throwing one of the a single throw with two dice.
7.
What
8.
What are the odds against throwing a total of ten with three dice ? are the odds in favor of throwing a total of more than five ?
numbered
is
Three tickets are drawn from a case containing eleven tickets 11. AVhat is the chance that the sum of their numbers 1, 2, twelve ? What is the chance that this sum is an odd number?
9.
if
Two
is
seven
gamblers A and B throw two dice under an agreement that thrown A wins, if ten is thrown B wins, if any other number is
to be divided etjually.
Compare
their chances.
that
is
to
win if he throws six before B throws seven, and that B is to win if he throws seven before A throws six. A is to begin and they are to throw
alternately.
11.
Compare
their chances.
Three gamblers A, B, and C put four white and eight black balls into a bag and agree that the one who first draws a white ball shall win.
PROBABILITY
If
423
when
they draw in the order A, B, C, what are their respective chances the balls drawn are not replaced ? when they are replaced ?
12.
What
is
having
13. balls
and twenty
of
.$5 ?
into
A bag A contains five balls one of which is white, and a bag B six none of which is white. If three balls be drawn from A and put B and three balls be then drawn from B and put into A, what is the
is
in
A?
balls
The bag
contains
a of which are white, and the bag B Is the chance of obtaining a white from one of these bags chosen at random
all
and a
drawn
In a certain town
first.
January January
16.
What
is
five deaths occurred within ten days including the chance that none of the deaths occurred on
first?
on the average two persons out of three aged sixty live to be seventy, what is the chance that out of five persons now sixty at least three will be alive ten years hence?
If
17.
boy
is
If eight
able to solve on the average three out of five of the probproblems are given in an examination and five
is
18.
is
to receive a dollar
if
first
throw
What
wins on the average three games out In playing tennis with B, What is the chance that he will win a set from B by the score of four.
of six to three ?
What
is
Under
chance has
against
21.
the conditions described in the preceding example what of winning a set in which the score is now four to two
him ?
Two
gamblers
.$32.
and B are playing a game of chance and each They are playing for three points, but when A has
one they decide
to stop playing.
How
should
424
A COLLEGE ALGEBRA
XXVIII.
MATHEMATICAL INDUCTION
A
number
of
791
Mathematical induction.
the
formulas cou-
tained in recent chapters may be established by a method of It is illustrated in the proof called mathematical induction.
following example.
Example.
Prove that the sum of the
first
n odd numbers
is n^.
We
are asked to
show
1
that
-..
+ (2n-l) =
n2.
(1)
We see by
k, so
inspection that
Suppose that we
that
true for certain values of n, as 1 or 2. have thus found it true when n has the particular value
(1) is
1
is
...
(2A;
1)
A;2
(2)
known
A;
to
1,
be true.
or 2
to
Adding the next odd number, nan\ely, both members of (2) and replacing A;^ + 2 A; +
l
2
1
(A;
1)
1,
by
[k
1)^,
we
obtain
...
(2A:
l)
(^-
l)2.
(3)
But (3) is what we get if in (1) we replace n by A: + 1. We have therefore shown that if (1) is true when n has any particular value A;, it is also true when n has the next greater value k + \. But we have already found by inspection that (1) is true when k has the particular value 1. Hence it is true when n = l + l, or 2 hence when n = 2 + 1, or 3 and so on through all positive integral values of n, which is what we were asked to demonstrate.
;
true for n
formula involving n has been found and we can demonstrate that if true for n = k it is also true for n = k -{- 1, we may conclude that it is true for all positive integral values of n. For we may reason: Since it is true when n = 1, it is also true when n = 1 + 1,
And,
in general, if a
or 2
hence when w
1,
or o
and
so
on through
all posi-
As another
illustration of this
find
6)
a"
C'i'a"-'i)
Cj'.a"
'
6'-
(1)
THEORY OF EQUATIONS
Multiplying both
(a
426
members
i
of (1)
by a
-\-
b,
we
obtain, 773, 1,
C';.
6)
a +
+ + +
C'l
a"b
C!;
an-ib^-...
art-r
+ ljr
CI
(2)
But (2) is the same as (1) with n replaced by n + 1. Hence if (1) is true when n = A;, it is also true when n = k = 1. It is therefore true when (1) is known to be true when n or 2 therefore when n = 2 + 1, or 3 and so on.
; ;
+
?i
1.
But
1
1,
Since the formula C" + C^"i = C"\^' can be proved independently of the doctrine of combinations, 774, the proof of the binomial theorem nere given is independent of that doctrine.
EXERCISE LXIX
Prove the truth of the following formulas, 701, 712, by the method of mathematical induction.
1.
2. 3.
4.
+ ar2 + + 32 + 13 + 23 + 33+ .. 1 + 3 + 6 + +
ar
12
+
+
ar -1
=
{n
(1
r")
(1
r).
2-'2
= =
2)/3!.
XXIX.
THEORY OF EQUATIONS
of
the
792
Every rational integral equation involving a single unknown letter, as x, and of the nth degree with respect to that letter, can be reduced to the standard form
a^x" + aiX"-^ H a_iX + a = 0. When the coefficients a^, !,, a are given numbers,
\-
(1)
(1) is
called a numerical equation, but when they are left wholly undetermined, (1) is called the general equation of the Tith
degree.
426 The
A COLLEGE ALGEBRA
final coefficient is often called
call
an equation of the form (1) complete or incomplete according as none or some of the coefficients ai, a^,---, a are 0. Observe that in a complete equation the number of the terms
We
n + 1. In what follows, when all the coefficients Oo, !, a are real numbers, we may and shall suppose that the leading one
is
fto is positive, and when they are which have no common factor.
we reduce
it
to the
Kx--^
j^^a;
coefficient is 1,
(2) is
and
+ j^ = 0, h^ = ai/a^,
(2)
equation. In the present chapter it is to be understood that /(a-) denotes an equation of the form (1) or (2).
=
are
"'
793
Roots
of equations.
The
7^oots
the
sometimes convenient to
794
the equation the roots of the polynomial. From the definition of root it follows that when is one of the roots of /(a-) is also that an equation/(a-) all of whose coefficients are positive can have no positive
root,
whose
coefficients
Thus, 2x3
2
a;3
x2
x2
Q cajj jjave no positive root since the polynomial is positive and 2 X'' x^ + 3 x - 1 =
;
x^
+ Sx
795
Theorem
by X
divisible by
root of f (x)= 0, then f(x) is exactly conversely, iff(x) is exactly divisible 0. h, then b is a root of f (x)
1.
// b
b;
is
and
THEORY OF EQUATIONS
For, by
427
bis f(b).
is
413, the remainder in the division of f{x) by this remainder But when i is a root of f{x)
f{b)
0,
is
exactly divisible by x
a;
b;
and conversely, when f(x) is exactly divisible by remainder f{b) is 0, so that 6 is a root of f{x) =
Example.
1
^',
the
0.
Prove that 3
9
is
a root of /(x)
-2 +0 -9[3
3
Dividing x^
ically, 411,
is 0.
= x^ - 2 x^ - 9 = 0. - 2 - 9 by x - 3
x'^
synthet-
3
3,
we
3
is
0=/(3)
Hence
a root of /(x)
is
0.
If & is a root of
\)j
b,
f(x) =
0,
so that f{x)
</>
exactly divisible
796
and we
call the
quotient
{x),
we have
f(x)
= (x-b)<f>(x).
are the values of x
;
Hence the remaining roots of f(x) = for which the polynomial (x) vanishes
<f)
(f>
0,
341.
Example.
1 1 1
3 x^
5x
= 0.
_3 +5
_
2
311 ~
3
ing x3
Z2
+3, "~0
1, 1
depressed equation x^
0.
The
roots
of this quadratic,
found by
631, are 1
I
Vi, and
V2. V2.
i
Hence the
roots of the
We
shall
797
(x)=
795 we deduce the following theorem, often called the fundamental theorem of algebra.
From
this
assumption and
Theorem
2.
degree, as
a
798
0,
f (x)
aox"
aix-i
+ a_ix +
roots.
is
By
Call
it
797 there
)8i.
Then f{x)
Hence
(1)
= {x-p,){a,x^-' + ). f{x)
428
A COLLEGE ALGEBRA
the same reasoning, since there is a value of which the polynomial a(^x"~^ + vanishes,
a-,
By
/3o,
call it
for
we have
(1),
a,x"''
(a;
^2) {a,x"-'13,)
f(x)
= (X == a, (X
(X
(2)
f{x)
(30
Pn)-
(3)
have thus shown that n factors of the first degree exist, p,, x fi,,- ,^ A* of which/(a;) is the product namely, x and by 419, /(./) can have no other factors than these and
;
We
when one
it
Hence,
793, the
when n
it
-,
/3
and
From this theorem it follows that the problem of solving an is equation f{x) = essentially the same as that of factoring the polynomial f(x). Also that to form an equation which
shall
its roots,
we have merely
x,
numbers
in turn
from
and then to
equate to
Example.
It is (X
Form
2)
(.c
2,
1/2,
/2)
(cc
1)
(a;
0)
0,
or 2 x*
3x5
1, 0. - 3 a;2 + 2x = q
800
Multiple roots.
Aj
13,
we
Observe that two or more of the roots /3i, two or more of them are equal to
And
in general, r, we call y8 a , double root, a triple root, in general, a root of order r. A simple root may be described as a root whose order r is 1. Evidently it follows from 798 that
tlie
roots equal to
two, three,
(3 lie
py
a root of order
but not by (x
0/
f (x)
is
that
/8)'
+^
THEORY OF EQUATIONS
429
When we say, therefore, that every equation of the nth. degree has n roots, the understanding is that each multiple times. It is of course not root of order r is to be counted
/
different
Thus, x^
since x^
3 x2 + 3 X 1 = 3x2+3x l = (x
the
rational
is
1)^,
but
'_
On
f(x)
finding
roots
.
numerical
equations.
Let
801
a^yX"
ttix""^
a,,
integral coefficients, and let b denote an integer and b/c a, rational fraction in its lowest terms. It follows from 451,
0,
is
then
i is a factor of a
is
a root, then S
a factor
and
c is
a factor of
a^.
Hence, in particular,
if a^
Ij
/c cannot be a root unless c = 1, that is, unless b /c denotes the integer b. Hence the following theorem, 454
:
An
ai, ,
+ a 0, where ajX''"' equation of the form x" a denote integers, cannot have a rational fractional
root.
It follows from what has just been said that all the rational roots of an equation with rational coefficients can be found by a limited number of tests. These tests are readily made by
803
synthetic division.
Example.
8 X*
x2
12 X
1,
12
14
2,
4 [2 4
[2
We
it is
by inspection that
is
not a root.
2
I
equation
3x^-2 x3-4x2-7x-2=0.
is
- 1/3
We
find that 2
a root of this
3x'
4x2
+ 4x +
0.
depressed equation also and obtain the second depressed equation This equation can have no positive root since
430
all its
A COLLEGE ALGEBRA
Testing 1, we find that it is not a root a root and obtain the third depressed
is
Hence the
remaining
roots,
Its rational roots of the given equation are 2, 2, 1/3. found by solving x'^ + x + 1 0, are ( 1 i v3)/2.
803
if
The reckoning involved in making these tests will be lessened one bears in mind the remark made in 453 also the fact
;
that a
cannot be a root of one of the depressed equations, finally the following theorem
:
796
and
If b
is 2)ositive
and
b synthetically are plus, f (x) result of dividing f (x) iy x can have no root greater than b ifh is negative and the signs
;
and minus,
b.
f (x)
can
less
than
For
it
effect of increasing b numerically will be to increase the numerical values of all coefficients after the first
coefficient, that
without changing their signs, so that the is, the remainder, cannot be 0.
final
Example
2
2
1.
Showthat 2x3+3x2 4 a; + 5 =
Dividing by x
cients
1.
+3 4 +5[1
5
1
_
2
If
only.
1.
we
+5
we
+1,
than
divide by x
2,
we
all
positive,
namely, 2
2.
10, 25.
3.x''
Example
3 3
Showthat
2
I
+ 4x2-3x + l = +
+4 - 3 +1 6 +4 2
If
Dividing by x
we
obtain coeflBclents
+1,-1
divide by x
which are alternately plus and minus. there is no root less than 2.
3,
Hence
we
we obtain
coefficients
just
12,
35.
804
add that any number which is known to be algeis called a braically greater than all the real roots of/(cc)=
We may
THEORY OF EQUATIONS
known
is
431
is
superior limit of these roots, and that any number which to be algebraically less than all the real roots oi f(x)
called an inferior limit of these roots.
is
of
that
is
EXERCISE LXX
1.
Form
2.
-b,a + b. (2) 3, 4, 1/2, 3 is a triple root of the equation Show that x4 + 8x3 + 18x2 -27 = 0.
(1)
1/3,
0.
3.
Show
that 1
of the equation
23 x?
33 x2
17 X
0.
the method of 803 find superior and inferior limits of the real roots of x^ 5 X* 5 x' + 4 X- 7 X 250 = 0.
4.
By
5.
Show
that 2 x*
3 x^
4 x^
10 x
Each
them.
6. 8.
x^
x2
14 X
24
l
= =
0.
7. 9.
x^
2 x2
25 x
3x3-2x2 + 2x +
X*
0.
2x*
7x3
2x2
+ 50 = 0. - x = 0.
10. 11.
12.
4 x3
8 x2
8X
=
6
0.
2x*
3x*
x5
+ +
7x3
+ 4x2- 7x-
=
6
0.
11x3
9x2
+ 11X+
71 x2
=
+
0.
9 X*
2 x3
81 X
70
4
=
0.
0.
2x5
x5
-8x4 +
7x3
+ 5x2- 8x +
16.
17.
12 x*
0.
x5-7x4 +
2 x<
18.
17 x3
25 x2
432
19.
A COLLEGE ALGEBRA
4 x5
x5
x''
6 x2
13 X
20. 21.
8x*
.3x'5
80x2
+ = 0. _ y2x + 240 =
0. 0.
2 x5
22. 23.
24. 25.
6x*
+ -
11 X*
89x3
X-
10 x<
3Gx*
12x5
2 x6
+ +
41 x^
46 x2
20 X
=
6
0.
108x3
20 X*
+ 107x2- 43x +
29x3
0,
77x2
+ 69x +
18
0.
26.
27. 28.
+
+ -
7 x5
8 x*
7 x3
2 x2
-14 x -
12
0.
2x6
5 x
11x3
7 x5
0.
whose
form,
roots are
^Si, /^o,
/3 is
798,
-\
(.3),
becomes
h
h^x"-'^
h^x"-j80 {X
h^x"'^
\-
= {x-
p,) (x-fi,)... (X
p).
Carry out the multiplications indicated in the second ber and arrange the result as a polynomial in x, 559.
memThen
equate the coefficients of like powers of x in the two members, We thus obtain the following relations between the 284. coefficients h^, lu, and the roots ^y, ji^,--, (S,,: h,^
-^i = A + A + A + --- + h = M. + A/Sa + + A/?3 + + A,-! A,> -h = A/82/83 + M:P, + A,-2 A,-l/8n,
)8,.
(1)
(2)
(3)
(-l)''&
/3,/3,/3,...A
(n)
where the second members of (2), (3), represent the sum of the products of every two of the roots, of every three, and so on,
THEORY OF EQUATIONS
433
and the sign before the first member is plus or minus according as the number of the roots in each term of the second member Hence the theorem is even or odd.
:
Theorem.
to
the form
f-
806
bix"-i
box"-^
-\
b =
0,
the coefficient bi of the second term, with its sign changed, is equal to the sum of all the roots; the absolute term b, with its sign changed or not according as n is odd or even, is equal to
the product of all the roots ; and the coefficient b^ of each intermediate term,, loith its sign changed or not according as r is
odd or even,
the roots.
is
equal
to the
Before applying this theorem to an equation whose leading coefficient is not 1 we must divide the equation by that If the equation be incomplete, it must be rememcoefficient.
0.
Thus, without solving the equation 3 x" G x + 2 = 0, we know the Reduced to the proper following facts regarding its roots /3i, p^, l^sform for applying the theorem, the equation is x^ + x^ 2 x + 2/3 = 0.
Hgiicg
/3i
*
/32
^3
0,
/3i,'3o
^1^3
^2/33
=-
2,
/3i^2/33
=-
2 /3.
find the
If all but one of the roots of an equation are known, remaining root by subtracting the sum of the
we can known
if
807
roots
its sign changed from b^, or by dividing b, with odd, by the product of the known roots.
is
Example.
Two
is
of the roots of 2 x^
3 x^
23 x
12
are 3 and
4.
What
The remaining
- 3/2 -
[3
+ (-
4)]
= - 1/2
or again,
it is
6-3(-4) = -l/2.
When
cients.
the roots themselves are connected by some given must exist among the coeffifind this relation
808
To
we apply
the theorem of
806.
434
Example
1.
A COLLEGE ALGEBRA
Find the condition that the roots
of x^
+ px^ +
gx
+r=
a / p,
h
/3
a, a/3,
we have
(X
/3
I-
a+
a/3
= p,
ct'^
a^
a-/3
q,
ex a a^ =
r.
The
r,
whence a
0.
r.
first,
q^
ph
+
8 x^
substituting
a=
V r in the.
Example
that
it
5x
50
0,
having given
we have
2 a/3
a+ ^=-
8,
a2
5,
a2/3
60.
= - 1 /3, ^ = - 22/3. The values a = 5, = 2 satisfy the equation a2/3 = a = 1/3, = 22/3do not .satisfy this equation. Hence the required roots are 5, 5, 2.
a=5,
/3
Solving the
first
and second
/3
a and
50,
/3,
we
obtain
and a
/3
809
Symmetric functions
roots to
of
the roots.
The expressions
in the
sym-
metric functions of the roots, 540. It will be proved in 868 that all other rational symmetric functions of the roots can be
expressed rationally in terms of these functions, and therefore rationally in terms of the coefficients of the equation.
Example
1.
2x3-3x2- 4xa^
0.
)3,
7,
we have
Example
is
px^
+
?
gx
are a,
/3,
7,
what
we have
-P' = ^1 + 7a + a/3 = 7, g' = /37 7a + 7a a/3 + a/3 ^7 = a/37 (a + + 7) = (- r) (- p) = - r' = Py ya ap = (a/37)2 = r*.
/3
ri),
is
x^
qx"^
+ prx r* = 0.
THEORY OF EQUATIONS
EXERCISE LXXI
1.
435
Two
The
;
of the roots of 2 x^
- 7 x2 +
10 x
are 1
find the
third root.
2.
progression
(1)
3.
8x3-14x2-21x4-27 =
The
find them.
(2)
x^
x2
+ 3x +
27
0.
progression
(1) x3
6x2
+ 7x if
0.
(2) x3
9x2
+ 23x -
15
0.
4.
Show
that
one root of x^
+ px2 +
qx
he the negative of
another root, pq
5.
r.
x^
+ px- +
qx
shall
be
Solve X*
4.x3
10x2
12x
0,
it
has
two double
7.
its
roots.
13x2 - 18x +
0,
that two of
Solve the equation x* - x^ - 56 x2 + 36 x + 720 = 0, having given 3 and that the difference between its roots are in the ratio 2
:
is 1.
If a, p,
+ j9x2 +
qx
0, find
the equa-
tions
(1) (3)
(5)
(2)
(i)
l//i,
1/7.
a +
k,
+ k,y +
k.
p\ y\
/3,
(6)
_
x2
l/a2, -1//32,
I/72.
10.
+ 1 = 0, find the values of (2) a3 + /33 + 7^(1) a2 + ^1 + 1 / nr/3. (4) a^2 + ^^^2 + ^.^,2 + ^^i 4. ^^a + a27. (3) 1 / ^7 + 1 / 7a + of x^ 2 x2 + x 3 = 0, find the values of 11. If a-, 7 are the roots
7 are the roots
-y2.
of 2 x^
4x
/3,
(1)
a//37
(/3
i3/7
7//3
(2)
(4)
a/3/7
(^2
/37/
+ 7/^a2) (a2
(3)
7) (7
a) (
/3).
^2) (^2
+ ^\
436
A COLLEGE ALGEBRA
TRANSFORMATIONS OF EQUATIONS
810
Some important transformations. It is sometimes advan = into another tageous to transform a given equation f{x) in some given relation to the roots equation whose roots stand The transformations most frequently used are of f(x) = 0. the following
:
811
into another To transform a given equation f (x)= o/f (x)= with their signs changed.
ivhose
for x in
/8
for y in
/(
For substituting any is /( y) = 0. f(x) gives the same result as substitutHence, \if{x) vanishes'when x = p, y).
?/
/3;
that
is, if
/3
is
a root of
= 0. Therefore every root of f{x) = 0, with its sign changed, has no other roots and /( y) = is a root of /( y) = are of the same and /(- y) = than these, since f(x) =
/(y)
;
a^x-"-^
+ a^x"-^
-\
= 0, = 0,
=
o(- vT
(-
y)""'
^2(-
2/)"--
o^n
iy""'
(>2i/"-'
+ (-
!)"
0.
Hence the required equation may be obtained from the given one by changing the signs of the terms of odd degree when n is even, and hjj changing the signs of the terms of even degree, including the absolute term, when n is odd.
We may use x instead of y for the unknown letter in the for /( y) = 0. transformed equation, and write /( x) =
THEORY OF EQUATIONS
Example.
437
Change the
4x5
9x3
6x2
- 18x +
= 0.
Changing the signs of the terms of even degree, we have 4x5 - 9x3 - 6x2 - 13x - 6 = o.
In fact, the roots of the given equation are
of the transformed equation are
1
/2, 3/2,
2,
i-
2,
i,
and those
1/2,
3/2,
(x)
into another
whose
k.
812
The required equation is f(y / k) = 0. For when X = P:f{u/k) will vanish when y /k =
y
=:
/(a;) vanishes
that
is,
when
k(i
(compare
811).
a^x"
+ aia;"-i + a^x^-^
-\
1-
= 0,
kaiy--^
+ k'^a^rf-^ +
k^a,,
0.
Hence the required equation may be obtained by multiplying the second term, of the given equation by k, its third term by k^, and so on, taking account of missing terms if any. When A- != 1 this transformation reduces to that of 811.
Example.
divide
x-*
+ +
2x3
4 x3
=
48
by
2.
Also
them by
first
is
The
x*
8x
0,
and since
0.
dividing by 2
x*
illustrates
an important application
813
now under
consideration.
Example. Transform the equation 36x3 another whose leading coefficient is 1 and
integers.
18x2
+ 2x +
into
its
remaining coefficients
Dividing by 36,
we have
x3
x2/2
+ x/18 +
l/4
0.
(1)
438
A COLLEGE ALGEBRA
k,
A:xV2
fc2x/18
+fcV4 =
0.
(2)
all
"We see by inspection that the smallest value of k which will cancel the denominators is 6. And substituting 6 for k in (2), we have
x3
3 x2
2X
54
= 0,
by
(3)
The
6 are the
814
of those of
f (x)
= =
The required equation \sf{\/y) = 0. For when X = /3, fi)-/>j) will vanish when 1/y =
If the given equation is
aoX"
if
f(x) vanishes
p, that is,
when
aix"-^
-\
+ a_iX + a =
0,
o/y'
l/y"'
+
"
+ an-l/y +
= =
0,
-i//"
iy
f'o
0.
Example.
rocals.
x^
3x
= +
2
by
their recip-
x^
0.
815
An
is,
equation like 2 x^
this
a;"^
a;
2
is
0,
which remains
to
it,
unchanged when
transformation
its
api^lied
that
when the
order of
If ^ is a root of such an equation, reciprocal equation, 645. Hence when the degree of the l/)8 must also be a root.
equation is even, half of the roots are the reciprocals of the The same is true of all the roots but one when other half.
the degree
is
which
is
its
own
odd; but in this case there must be one root reciprocal, that is, one root which is either
THEORY OF EQUATIONS
1 or
439
and
- 1.
+ 3a;2-3x-2 = 0isl
2 and
1/2.
816
it follows that when and the leading coefficient vanishes, 07ie of the roots becomes infinite; when the tivo leading the roots become infinite; and so on. coefficients vanish, two of
From
Example
becomes
Applying
1.
infinite
Show when
3a;2
2x + l=0
(1) (2)
vanislies.
814 to
mx^
x*
we
obtain
If the roots of (2) are
+ 3 x'- - 2 x + 1 = - 2x2 + 3x + m =
^3,
0,
0.
/3i, /Sg,
those of
if
By
roots
806,
/3i,
/3i/32i33
m.
Hence,
/32, /Ss
must
also
(1),
corresponding root of
approach namely, 1
as limit,
//3i,
the
Example
become
2.
infinite
vi^-
=
(1) (2)
Applying
814 to
we have
Also,
^iftft
It follows
+ m'^x'^ + x + 1 = + x^ + jn^x + m =
^2, ft,
0, 0.
/3i,
those of
(1)
=and
iSiiSa
^ift
+
jSi,
ftft
= m\
.,
,.
(3)
/3i,
from
(3)
0,
that
if
m approach 0,
these roots be
two
00.
ft, ft
must
also
approach
if
must approach
f
(x)=
into another
whose
817
0,
as k.
The required equation is/(.?/ + A;) = 0. For if /(x) vanishes when X = fi, f{y + k) will vanish when g + k = (3, that is, when ?/ = y8 k.
If the given equation is
f(x)
a^pc"
aix"-^ H
[-
a_-iX
= 0.
f(g
k)
a,(g
+ ky +
a^{y
k)"-'
+ a =
0,
440
A COLLEGE ALGEBRA
which, when its terms are expanded by the binomial theorem and then collected, will reduce to the form
</>(y)
=
Ci
^oy"
^1^""'
c,-^y
= o,
where
= Cq
(,,
?ikci^
i)
and
so on.
laborious.
from f(x) is usually very much more expeditious, at least when the coefficients oif(x) are given rational numbers. li X = y -\- k, then y =z x k, and we have
This method of obtaining
(j>(}/)
is
that
Co(x
is,
k)"
-\
h c_i (x
k)
aoX H
\-
a_iX
If both
members
by
a;
+ a. k, if
k, and again the quotients thus obtained be divided by x so on, the successive remainders yielded by the first member, namely, c, c_i, , will be the same as those yielded by the
second member.
follows: Divide
f
k,
(y)
from f(x) as
by X
k,
Cn-u
Ci,
(compare
423).
The
Example L
First method.
7 x^
3x +
3 (y
=
+
4)
by
4.
2 x3
7 x2
Second method.
in 423,
we have
1[4
4
5
.-.
Ca
= =
= =
5.
.-.
C2
37.
2,
17
.-.
ci
17
and
Cq
2,
17y2
+ 37y +
o.
THEORY OF EQUATIONS
Example
2.
441
3 =:
4 x^
by
4.
To
- 4.
the
same thing
as to diminish
them by
Hence
ing 2/ be 2/3
required equation may be obtained either by substitut4 for X or by dividing synthetically by 4. It will be found to
tlie
2/2
17
?/
0.
By aid of 817 we can transform a given equation into another which lacks some particular power of the unknown
letter.
818
Example 1. Transform the equation x^ 3x2 + 5x + 6 = another which lacks the second power of the unknown letter.
Substituting x
into
must have Sk
x3
= y + k, we have y'^ + {Sk 3)y'^ + -. Hence we 3 = 0, that is, k 1. And diminishing the roots of - 3x2 + 5x + 6 = by 1, we obtain x^ + 2x + 9 = 0. 5x- + 8x 1 = into Example 2. Transform the equation
x*^
first
power
of
tlie
unknown
letter.
+ k, vfe have - 5) 2/2 + (3 - 10 + 8) + = 0. 2/3 + (3^ - 10 + 8 = 0, that is, = 2 or 4/3. Hence we must have 3 5 x2 + 8 x 1 = by 2, we obtain Diminishing the roots of x^ x3 + x2 + 3 = 0.
y
A:2
A;
2/
fc2
A;
fc
= by k, we obtain If, when we diminish the roots of f(x) an equation <^(a:)=0 whose terms are all positive, k is a = 0, 804. For in superior limit of the positive roots oi f(x) Hence any this case <^(.r)=0 has no positive root, 794. may have become negative when positive roots that f(x) =
diminished by
k.
819
They
process of synthetic division is such that it is possil)le by inspection and trial to find the smallest integer k for which I'n most cases this will be positive. all the terms of <^(a-)=
The
can be accomplished with comparatively little labor. We may obtain an inferior limit of the negative roots of
f(x) = For if
0.
is
0,
then
A;
is
442
Example.
tion /(x)
A COLLEGE ALGEBK.^
Find superior and inferior
X*
6 x3
14 x2
48 X
trial
all
121
0.
A:
We
find
by inspection and
(x)
if
that neither of
nor
A;
2 will give
=
+
positive,
but
we diminish
78 x
=
is
by
3,
we
(x)
x*
6 x^
14 x^
0.
a superior
x)
= =
is
x*
6x^
+ Ux^ - 48x -
121
0.
We
roots.
by inspection and trial that 3 is a superior limit of its positive Hence 3 is an inferior limit of the negative roots of /(x) = 0.
820
On
If
we eliminate x
x,
between
the
equations f(^x)
=
in
and
wq
obtain
= x.
we
811
This
equation of the form y = (x), where <^ (x) is rational, we shall obtain an equation F(y)= whose roots are connected with those oif(x) = by the relation y 4^{x), so that if the
roots oif(x)
<I>(J3.^,
are ^i,
</)(^).
those of F(y) = are cf> (/?i), The transformations of 812, 814, 817
p.,,
/3,
In the
is
first
of
these transformations the equation y z= (^(x) second it is ?/ = 1 /a-, and in the third it is 3/
as in these cases, // <^ of X is readily effected.
(a-)
x,
the elimination
Example
roots of X*
1.
+ px2 +
0.
(p{x) is
= x^.
y/y for
And substituting Solving y = x'^ for x, we liave x = VyX in the given equation and rationalizing, we obtain
y^
+ (2q- p2) y2 + +
(g2
_2pr)y -r^ =
0.
If the roots of x^
px^
+ gx +
are a,
j3,
7, find the
a/3.
We
a/3 in
first endeavor to express each of the proposed roots /37, ya, terms of a single one of the given roots a, /3, 7 and the given
THEORY OF EQUATIONS
coefficients p, q,
r.
443
This
is
^y
= a^y/ a =
r/ a,
ycx
= a^y / =
jS
r/ ^,
Hence each root y of the required equation is connected with the corre r /x. sponding root X of the given equation by the relation y r/x for x, we have x = r/y. Solving y =
And
fying,
substituting y^
we have
qy'^
r/y +
for x in x*
+ px^ +
which
is
qx
i:>r?/
7-2
0,
EXERCISE LXXII
1.
Change
3 x*
2 x^
2.
+
10
x^
4 x2
6x
+ =
6x
by
+7= - 2.
0.
Also
divide
3. 4.
them by
In 5 x^
3.
X* +
3 x^
9X
+
x^
=
x*
its
reciprocal.
3 x^
by
2.
Also increase
them by
5.
x^/S
x2/4
+ x/25 -
1/48
=
1.
x*
36 x'
+x
into another
which
7.
x3
0x2
+ 9x + +
/32,
10
=
x
0.
(2)
x3
x2
0.
8.
If the roots of x*
x^
=
r
are a,
^3,
7, 5, find the
equation
a2,
72, 52.
If the roots of x*
+
/3
3 x^
+ +
2 x2
5,
are a,
5,
/3,
y,
5,
tt
find the
/3
a+ 7+
a +
^3
5,
7.
px-
qx
are a,
/3,
7, find the
equa-
tions
^7
(1)
a^
1
/37
1
ya
(<i)
i3
a
1
^
1
11.
If the roots of x^
+7 + 2x2 + 3x + 4 =
''/3
i32.
7+aa+i3
7),
tions
(1)
^2
+
+
y2^
(3)
By
-, ytx+
111
yi
0-2,
a:2
(2)
ex
(^
+
a
^ (7
,
),
7 (
,
^)-
a-/3
-.
(4)
444
12.
A COLLEGE ALGEBRA
Find superior and
inferior limits of the real roots of the following
equations.
(1) x* (3) X*
+ 3x3 -13x2 -6a; + 28 = 0. (2) 2 x^ - 120x2 - 38x + - 26 x^ + 60x2 _ - 29x2 + 50x + 12 = 0. (4) 2x5 - 14 x3 + 44 x2 + 28 X - 92 = 0. (5) ji 77x2-50x-110=.0. (6) 3x-35x3 +
27
92
= =
0. 0.
IMAGINARY ROOTS.
821 Theorem.
cients.
-//"
If a
it
ib is
if
= denote an equation with real coeffif (x) has imaginary roots, these occur in pairs; that is, a roof, a ib is also a root.
Let
a root of fix)
;
= 0, then f(x) is divisible by and we shall prove that a ib is a root (a ib), or, if we can show that f{x) is also divisible by a; what comes to the same thing, if we can show that f(x) is
For
4- ih is
X (^a
ib),
795
divisible
by the product \_x (a + ib)][x (a This product has real coefficients, for, since i^
ib)^. = 1,
lx-(a +
ib)^ [.r
(a
ib)^
common
factor x
(a +
b'^)
have
ib),
factor.
cannot be x (a + ib), since whereas the highest common factor of two polynomials with real coefficients must itself have real coefficients, 469. Hence the highest common factor - 2 ax + (a^ + b"") in oif{x) and x^-2ax+ (a"- + b") is x^
ov
x"^
This highest
common
b'^).
2ax + (2 +
imaginary
But
this has
coefficients,
x"^
2 ax
+ (a^ +
is
b-),
as
was
Solve
One
root of 2
x''
T)
x^
46 x
87
i.
Since
the roots
is
+ 5 i is a root, 2 5 iis also a root. But 5/2, 806. Hence the third root is _5/2-(-2 + 5i-2 - 5i) = 3/2.
the
sum
of all
THEORY OF EQUATIONS
445
822
Corollary 1. Eoery polynomial f (x) xi'itli real coefficients is the product of real factors of the first or second degree.
For
to
c of
real factor
f{p:)
;
roots a
factor
a;2
oif(x),
795
and
the real
Corollary 2.
corre-
823
spond which
to the
f(x)=
x.
is
a function of x
is positive
it
For
may
be expressed as a product of factors of the form 821, and every such factor, being a sum of
Every equation
degree
824
odd has at
even,
For the number of its imaginary roots, if it have any, is 821, and the total number of its roots, real and imagiHence at least one root must be real. nary, is odd, 798.
them
Thus, the roots of a cubic equation with real coefficients are either real, or one real and two imaginary.
all
of
By
if
821
it
may
be proved that
825
are themselves rational, but Vi^ irrational. be an equation whose Irreducible equations. Let <^ (x) coefficients are both rational and real. say that this
826
We
equation is irreducible if ^{x) has no factor whose coefficients are both rational and real (compare 486).
Thus, x2
i-
and
x-
is
not irreducible.
Theorem.
Let
equation of the
be any equation and real, and let ^(x)= same or a lower degree.
f(x)=0
ichose coefficients
be
827
an
irreducible
446
A COLLEGE ALGEBRA
be a root of f (x) If one of the roots of ^ (x) are roots of f (x) = 0. of the roots of <f>(x)=
0,
then
all
This
f(x)
4> (a-)
be proved by the reasoning of 821. For if = have the root c in common, f(x) and <^ (x) have the common factor x c, 795, and therefore a
may
and
common factor which is either x c, some factor of which contains x c, ov (x) itself. But since, by hypothesis, (f>(x)=z is an irreducible equation, (fi(x) is the only one of these factors Avhich has real and
highest
(f>
(x)
(jt
rational coefficients such as the highest and (f> (x) must have, 469.
common
factor oif(x)
and
Therefore <^(x) is itself the highest common factor of f(x) <li(x); in other words, /(a-) is exactly divisible by
^(ir).
Hence f(x) may be expressed in the form f(x) = Q<j> (x), where Q is integral, and from this identity it follows that
f(x) vanishes whenever <f)(x) vanishes; in other words, that, is a root of every root of <f,(x) = f(x) = 0. Permanences and variations. In any polynomial f(x), or equation/(a-) = 0, with real coefficients a. 2ier7?ianence or continuation
of sign
sign,
is said to occur wherever a term follows one of like and a variation or change of sign wherever a term fol-
828
829
f (x) is exactly divisible bt/ x h, where b is the coefficients ofi(x) are real, the quotient (f>(x) will have at least one less variation than f (x) has.
Theorem.
If
positive
and
from the rule of synthetic divided by x - b, the coeffiquotient are positive until the first negative coefficient of f(x) is reached. If then or later one of them
division, 411, that cients of the
For since
b is positive, it follows
when /(a;)
is
becomes negative or zero, they continue negative until the next positive coefficient of f(x) is reached, and so on. Hence <^ (x) can have no variations except such as occur at the same
THEORY OF EQUATIONS
or earlier
447
is,
terms of /(x).
But
by hypoth-
<^ (.r)
<f>
must be contrary to the last (x) must lack the last variation
in /(a-).
1
+1
2
-2
+_6
1+3
1
+4
-10 -1 +12 -4[2 Thus, f(x) = x6 + x^ - 2 x* -10x3 -x2 + 12x - 4 isex+ 8 - 4 - 10 + 4 2, the 2 5 +2, actly divisible by x 2 x^ 5 x + 2. Observe that the (x) = x^ + 3 x* + 4 x'
of/
(x)
-1
2
T+1+3
disappear in
+1 +2 +
-7
G
+2\2 2
Again, /(x)
exactly
x*
divisible
by x
1, being 0(x) = x'' + xonly one of the four variations of /(x) is reproduced in 0(x), and we have an illustration of the fact that wlien inter)nedlute variations of /(x)
<p{x),
Theorem (Descartes's
cannot
830
have a greater number of positive roots than it has variations, nor a greater number of negative roots than the equation f ( x) = has variations.
A]
For
let
if
we
hy X
since
divide f(x) by x fS^, the quotient thus obtained and so on, we obtain a final quotient ^(.r) which (3.,,
r less variations
jSi, /So,
/8,.
0.
has at least
less
than f(x) has, 829. Therefore, than no variations, f{x) must have
at least as
is,
many
as f(x)
has
positive
2.
of/(.r)=0 become the positive roots And, as just demonstrated, /( x) = cannot have more positive roots than variations. Hence = cannot have more negative roots than /( x) = f(^x)
The negative
x)
roots
of
/(
0,
811.
has variations.
x^ x^ + x 1 = cannot have more Thus, the equation /(x) = x^ than three positive roots nor more than one negative root. For f(x) has three variations, and /( a') = 0, that is, x^ + x^ + x' x 1 = 0, has
one variation.
448
831
A COLLEGE ALGEBRA A
Corollary. cotnplete equation cannot have a greater number of negative roots than it has jjernianences.
For when f{x) = is complete its permanences correspond one for one to the variations of /( x) = 0, since of every two consecutive like signs in /(-x) = one is changed in /( x) 0,
81L
then
+ x*-6x3- 8x2-7x + l =0, (1) - X* - 6 x^ + 8 x2 - 7 X - 1 = 0. (2) In (1) we have permanences at the terms x*, 8_x2, 7x, and at the 7 x,; we have variations. 8 corresponding terms of (2), namely, cannot Since (1) has tivo variations and three permanences, /(x)
Thus,
if
/(x)
X)
/(-
= =
is
x5
x5
is
x**,
x'-,
have more than two positive roots nor more than three negative
roots.
832
equation
roots
we can frequently prove the existence of imaginary by aid of Descartes's rule of signs.
he an equation of the nth degree ^uhich has no Let f (x) zero roots, and tet v and v' denote the number of variations in
x)
least
f (x)
=
v'
negative roots,
cannot have more than v positive roots nor 830, and therefore not more
roots all told.
v' real
The
rest of its
n roots must
therefore be imaginary.
This theorem gives no information as to the imaginary roots of a complete equation, since v f' is equal to n in such an
equation.
Example.
Show
that x^
isx'^
In this case/(x)
=
')
Hence n
(u
+ x^ + 1 = has four imaginary roots. + x^ + 1 = 0, and/( x) = is x^^ x^ 1 = 0. (0 + 1) = 4, so that there cannot be less than
four imaginary roots. But since there are five roots all told and one of them is real, the degree of x^ + x^ + 1 = being odd, 824, there cannot be more than four imaginary roots. Hence x^ + x- + 1 = has exactly
THEORY OF EQUATIONS
EXERCISE LXXIII
1.
449
One One
root of 2 x*
x^
5 x-
13 x
= +
2
is 1
is
t.
Solve this
equation.
2.
root of 2 x*
11 x^
17
x-^
10 X
V2.
Solve
this equation.
3. Find the equation of lowest degree with rational whose roots are 5 + 2 i and 1 + v 5.
coefficients
twc
of
4.
5.
Find
tlie
whose roots
is
v'2
i.
conclusions regarding the roots of the following equations can be drawn by aid of Descartes's rule and g 8o2?
(1)
What
x^
(7) x5
+ + + -
0.
(2)
x*
2 x^
x5
+ +
x2
+
1
=
0.
0.
x^
X--
-X+ = 0.
x2
2 x'
as
all of whose roots are real has and as many negative roots as
permanenc:es.
7.
Given that
all
the roots of x^
3 x*
+
1
5 x^
35 x2
54 x
72
how many
how many
negative.
What
conclusions can be
X2'+1
9.
drawn by
aid
+ 1=0? X2"-1=:0?
X2'' + 1-1
0?
Prove that an equation which involves only even powers of x with positive coefficients cannot have a positive or a negative root.
10.
that the equation z^ + px + q = 0, where tive, has but one real root, that root being negative.
Show
p and
q are posi-
12. Show that an incomplete equation which has no zero roots must have two or more imaginary roots except when, as in x^ 3 x"^ + 1 = 0, the missing terms occur singly and between terms which have contrary signs.
13. Show that in any equation f{x) = with real coefficients there must be an odd number of variations between two non-consecutive contrary signs, and an even number of variations, or none, between two
450
A COLLEGE ALGEBRA
14. Prove that in the product of the factors corresponding to the negative and imaginary roots of an equation with real coefficients the final term is always positive, and then show that if this product has
any variations
15.
their
number
is
even.
Prove that when the number of variations exceeds the number of positive roots, the excess is an even number.
16.
Show
that x*
x^
x-
=0
tive roots
root.
is
17. Show that every equation of even degree who.se absolute term negative has at least one positive and one negative root.
Theorem
(^xj
1.
7/'f(a)
signs,
a root of
lies heticeen
a and
This may be proved as in the follov^ing example. statement of the proof will be given subsequently.
Example. and 2.
Prove that f{x)
A general
1
x^
3x +
The
sign of /(I)
is
of /(2)
is
plus.
By computing we
successively,
between
and
which /(x) has the .same signs as for x = 1 /(1. 5) =- .125 is miuu.s, and /(l.O) = .290 is plus. By the same method we find tiDO consecutive hundredths between 1.5 and 1.6, namely, 1.53 and 1.54, for which /(x) has the same signs as for x = 1 and X = 2 for /(1. 5:5) = - .008423 is minus and/(1.54) = .0,32204 is plus. This proce.ss may be continued indefinitely. It determines the two never-ending sequences of numbers:
;
;
(a)
1, 1.5,
1.53, 1.532,
(b)
2, 1.6, 1.54,
1.533,
-,
the terms of which approach the .same limiting value, 192, 193. this limiting value c. It is a root of /(.r) = 0, that is, /(c) = 0.
if
Call
run througli either of the sequences of For, by 509, values (a) or (b), /(x) will approach f(c) as limit. But since/(x) is always negative as x runs tiirougli the sequence (a), its limit /(c) cannot be positive and since /(.c) is always positive as x runs through the .sequence (b), its limit /(c) cannot be negative. IIence/(c) is zero.
x be
to
;
made
THEORY OF EQUATIONS
451
834
Theorem 2. If neither a nor \) is a root of f (x)= 0, and an lie between a and b, f odd number of the roots of f(x) = (a) and f(b) haae contrary signs; hut if no root or an even number of
roots lie between
a and
Conversely, if f(a) and f (b) liave contrary signs, an odd but (/" f (a) number of the roots of f (x) lie between a and b
and
same
sign, either
b.
of roots
between a a7id
< b and that of the roots of /(a-) = between ^2) divisible by (x ^i)
Suppose that a
ySi, yS,,
/3,.
is
a complete
if
list
and
b.
/3,.),
(a;
(x
is exactly-
we
call
the quotient
<^
(x),
we have
/30 (x
/(,r)
= (x-
/?,)
...(X-
/3,.)
<t>
(x).
(1)
Substituting first a and then b for x in (1) and dividing the first result by the second, we obtain
f(a)
f{b)
^ a- pi
(3.2
(f>
(3,
<f>(a)
b-(3i'b-/3,"'b~(3,.'
<}>
^^^ ct>(by
For In the product (2) the factor (a) / (b) is positive. and <^ (b) have the same sign, since otherwise, by 833, = and therebetween a and b there would be a root of (x)
<^
(a)
<f)
fore,
by
(1),
a root of f(x)
(3i,
(3---,/3,..
On
and
the other hand, each of the r factors (a (3i)/(b so on, is negative, since each of the r roots ^1, /So, ,
/3i),
/S^
is
b.
Therefore, when r is odd, f (a) /f(b) is negative, that is, f(a) and f(b) have contrary signs but when r is even or zero, f{fi)/f(b) is positive, that is, /() and f(l)) have the
;
same
sign.
Conversely,
signs, so that
;
f(a) /f(b)
follows from (2) that r is odd and when f (a) and f(b) have the same sign, it follows that r is even or zero.
is
negative,
452
835
A COLLEGE ALGEBRA
Observe that in the proofs of the preceding theorems, 833j, 834, no use has been made of the assumption that every equation /(a-) has a root. Notice also that in applying these
theorems a multiple root of order r is to be counted as r simple roots. From 834 it follows that as x varies from a to b, f(x) will change its sign ias x passes through each simple root or multiple root of odd order of /()= which lies between a and b, and that/(x) will experience no other changes of sign than these.
Thus,
to
5,
if
/(x)
(X
2) (x
3)2 (x
4)3,
and x be made
the sign of /(x) will be plus between x = 1 and x 2 and x 4 and x = 5. 4, and plus between x
2,
836
By
833
usually possible to determine between what pair of consecutive integers each of the fractional and irrational roots of
lies.
x*
6 x^
x2
12 x
= 0.
Descartes's rule of signs, 830, this equation cannot have more than three positive roots nor more than one negative root.
By
To locate the positive roots we compute successively /(O), /(I), /(2), until three roots are accounted for by 833 or until we reach a value of
X which
/(O)
is
= - 6, = 102.
2,
/(2)
=-
414,
we
find
10, /(.3)
=-
42, /(4)
=-
70, /(u)
= - 46,
2,
Hence,
8.33,
one root
lies
and
and
6.
between and 1, another between 1 and There cannot be more than one root
in
any of these intervals, since there are only three positive roots all told. Making a similar search for the negative root, we have /(O) = - 6, /(-l)=-10>/(-2) = 38. Hence the negative root lies between -land -2.
The mere substitution of integers for x in f(x) will of course not lead to the detection of all the real roots when two or more
of
them
lie
will be considered in
is
between a pair of consecutive integers. This case 844 and again in 8G4, where a method
how many
THEORY OF EQUATIONS
EXERCISE LXXIV
Locate
1. 3.
458
2 x3
x3
3 x2
9X
+
5
0.
2.
x^
3 x^
2X
0. 0.
4.
6.
2 x3
x3
- ix -2 = 0. + 3 x2 - 10 X - 15 =
x-
0.
5.
x3- 4x2-4 x +
x3
12
7.
9.
5 X
19
0.
8.
x*
x*-8x3 + 14x2 + 4x -8 =
X*
0.
10. x*
x2
- 13x - 7 = 0.
11.
-11x3 + 32x2-
4x- 46 = 0.
+
is
12.
13.
is
x5
+
its
2 X*
16 x3
tliat
24 x2
48 X
32
0.
Assuming
term
when x
that of
(1)
of highest degree,
show
that
6
in
+ Every equation x" + 6iX"-i + which ?i is even and b,i is negative, has
with real
coefficients,
at least
6) (x
c)
+
+
^2
(x
Hi- (x
a,
- c) (x - a) a) (x 6) b,
X (x
a) (x
6) (x
c)
=
being
lie
a and
I,
k,
it
14. Show that every equation of the form x^ + (x 1) {ax 1) = 0, where a>3, has two roots between and 1, namely, one between \/a and 1 1/a and one between 1 1/a and 1. 15.
Show
that x^
(x
1) (2
1
1)
(ax
1
between
2/a,
837
numer-
The most expeditious of equations can be computed. these methods is due, in its perfected form, to an English mathematician named Horner. It may best be explained in
connection with an example.
454
Example.
1.
A COLLEGE ALGEBRA
Find the positive root of /(x)
x-^
x^
15 x
59
0.
By
the
method of 836, we find that the required root lies between Hence if it be expressed as a decimal number, it will have the
y, 5,
2.
(1)
by
3.
We
0(x)
2x3
19x^
+ 45X-41 =
0,
^2)
lying between
.1,
.2,
.3,
that 0(.6) is and ^(.7) is +. Hence the root of (2) lies between .6 and .7, that is, ^ is 6, and the root of (1) to the first decimal figure is 3.6.
in <p{x),
we
find
2+19
V2
20.2
1-2
+45
12.12
41 1^ 34.272
3.
(2)
by
.6.
We
f
57.12
12.84
6.728
(x)
2x5
+ 22.6x5 +69.96X-
6.728
0,
(3)
THEORY OF EQUATIONS
In the example we found the
transformed equation, that
first figure
455
839
successive decimal figures of the root of the given ec[uation, by the method of substitution. But the first figure of the root of each transformed equation
is, tlie
from the second on may ordinarily be found by merely dividing the absolute term of the equation, with its sign changed, by the This is called the method of trial divisor. coefficient of x.
Thus, consider the second transformed equation in the example
^p{x)
2x^
22.6a;-
69.90x
6.728
0.
(3)
This equation is known to have a root c wliich is less than .1. The second and higher powers of such a number c will be much smaller than
c itself.
Thus, even
(.00)'- is
but .0081. Hence, were c known and subtwo terms of the resulting numerical identity
22.6 c2
in
2 c3
+ 69.96 c -
6.728
=
last
two.
figure
6.728
=
is,
obtained by discarding the x^ and x^ terms in (3). But solving (3'), we have x = 6.728/69.96 = .09 +, that above, that the first figure of the root of (3) is 9.
we
find, as
This method cannot be trusted to give the first figure of the root of the fJrst transformed equation correctly. But it will usually give at least some indication as to what that figure is
and
so lessen the
number
of tests that
need
to be
made
in
applying the method of substitutions. Occasionally the method fails to give correctly the first figure of the root of even the
second transformed equation. But in such a case the error is readily detected in carrying out the next transformation for if the figure is too large, a change of sign will occur in the
;
absolute term of this next transformed equation, 838 if it \s too small, the first figure of the root of this equation will be of too high a denomination.
;
We may
840
812, before
456
transformation.
A COLLEGE ALGEBRA
This may be done by affixing one zero to the
second coefficient of the equation in question, two zeros to its third coefficient, and so on. We then treat the iigure of the root
employed
it
were an
integer.
Thus, the
837 was
(2)
.6
+.
0,
(20
Diminishing the roots of (2') by 6, the reckoning differing from that above given only in the absence of decimal points, we have
2 x3
226 x2
G990 x
6728
0,
(SO
22.6 x2
69.96 x
6.728
= 0.
(3')
(3)
The method
as above, .09
as the root of
and therefore,
841
We may now arrange the reckoning involved in computing the root of 2 x^ + x^ lox 59 = to the third decimal figure as
follows
:
59 [3.693 18
-41
THEORY OF EQUATIONS
tenth, thus, .9, .3.
457
Observe that here each figure obtained by the trial divisor method is a Had the last coefficient of the second transformed equa-
tion been
672 instead of
The and
6728, we should have had 672/6996 = .09 root as far as computed would then have
we should have
by
100. that
is,
multiplied the roots of this second transformed equation we should have affixed two zeros to its second coefficient,
four
to the third,
and
This process
may
be continued indefinitely.
But we soon
842
encounter very large numbers, and after a few decimal figures of the root have been obtained we can find as many more as are likely to be required, with much less reckoning, by the
following contracted method.
reckoning above given
The
last
transformed equation
2
a;'5
in the
is
2314 x^
740706 x
247082
0.
(4)
Instead of affixing zeros to the coefficients in order to multiply the roots of (4) by 10, we may substitute x/10 for x in (4), 812, thus obtaining
.002 x^
23. 14
X'i
74076.6 x
247082
0.
(4')
Ignoring the decimal parts thus cut off from the coefficients as being too small to affect the next few figures of the root, but adding 1 to the corresponding integral part when the decimal part is .5 or greater, we
23 x^
74077 x
247082
0.
(4")
:
We may
74077
247082
|.
003332
69
222438
458
A COLLEGE ALGEBRA
off figures as
and thus reduce (5) to the simple equation 7422 x 24644 = 0. The next two figures of the root, namely, 3, 2, are then obtained by merely dividing 24644 by 7422 by a contracted process which consists in cutting off figures at the end of the divisor instead of affixing zeros at the end of
the dividend.
before
843
Negative roots.
To
0,
/( x)=0 and
x^
then
change
its sign.
Example.
Here/(
of /(x)
x)
x^
lOx
0.
found by
844
If the given equation has two roots Roots nearly equal. lying between a pair of consecutive integers, they may be found as in the following example.
Example.
roots,
1,
if
any, of /(x)
= x3 + x2 10x+9 = 0.
/(2) = 1, /(3) = 15, and the reckoning shows that 3 is a superior limit of the roots, 803. Hence, 834, either there is no positive root, or there are two such roots both lying between
We
find that/(0)
9,
/(I)
and
1,
3.
differ less
from
3.
may
between 2 and
We
= +
1
by
1,
obtaining
x3
4 x2
5X
if
and
for x
is
f(x)
= =
0,
of
+ and
(/)(.3) is
(x)
we
find that
<f)
(.2) is
is
and another between .7 .8. By Horner's method we find that these roots are .25560 and .77733 approximately. Hence /(x) = has the two positive roots 1.2556 and 1.77733.
.3
and
+. and
Hence ^{x)
845
On
greater than ten, we may employ the followfor finding the figures of its integral part.
THEORY OF EQUATIONS
To
obtain the
459
compute the values of f(x) for = 100, 200, , and so Thus, if we found that/(400) and /(500) on, applying 833. had contrary signs, so that the root lay between 400 and 500, To find the remaining figures, the first figure would be 4.
first figure,
10, 20,
or, if
necessary, for x
successive transformations of the equation, as when Thus, in the case just cited we finding the decimal figures. should diminish the roots of /(a-) by 400 and so obtain an and 100. If we having a root between equation <^(a-)=
make
found that
between 70 and
80, the
second figure
We
=
9,
root between
having a by 70 and so obtain an equation i}/(x) = and 10. If we found that this root lay between
should have shown that the integral part of the
8 and
we
root of f{x)
was 478.
If asked to find uil the real
On
846
roots of a given numerical equation f(x) 0, it is best, at least when the coefficients are rational numbers, to search first
for rational
roots
by the method of
802.
This process
will yield a depressed equation <^(x)= whose real roots, if locate these roots by the method of any, are irrational.
We
and then
find their
approximate values by
Horner's method.
It may be added that a fractional root may also be found by Horner's method, exactly when the denominator involves only the factors 2 and 5, approximately in other cases.
EXERCISE LXXV
Compute the
1.
x^ 4-
X
2
=
20
2.
3.
2. 3.
x^ x^
root between 2
2
;
and
4.
5.
+ 10 X 3 x^ + 5 X 40 = x3 + 10 x2 4- 8 X +
6 x2
root between
and
3.
1.
root between 2
r=
;
and
120
3.
460
6.
A COLLEGE ALGEBRA
2 x3
x^ x^
X*
x'*
7.
8.
9.
10.
11.
x^
_ 9X + 1 = root between 1 and 2. 5x 1 = root between 1 and 2. root between 5 and 6. 2 x2 23 X + 70 = root between 3 and 4. 10 x2 4 X + 8 = root between - 2 and 6 x3 + 12 x2 - 11 X - 41 =
a;2
;
x2
3.
3 x2
4X
13
3.
Find
12.
14.
to the third decimal figure all the roots of the following equations
x3
x^
3 x2
3X
4X
+
0.
10
0.
13.
x3
X*
+ +
x2
2X
0.
15.
5 x3
x2 -
13 X
0.
16.
By
to the equation x^
17
compute
Vl7
17.
By
By
the
V3 and V87
decimal figure.
18.
aid of 845
x^
2500
0.
By
20.
Find
TAYLOR'S THEOREM.
847
Derivatives.
MULTIPLE ROOTS
n,
the exponent of x, and then diminish that exponent by 1. obtain n(ix"~^, which is called the derivative of ax", or,
precisely, its derivative
We
more
with respect to
a,
x.
In particular, the
derivative of a constant
that
is ax'^, is 0.
of the derivatives of the terms of a polynomial f(x) is called the derivative of f(x), or, more precisely, its ,/irst
derivative,
The sum
and
is
represented by
of /'(.r)
is
/"'(x).
The derivative
f(x), and
is represented by f"(x), and so on. Evidently every polynomial f(x) of the nth. degree has a series of n derivatives, the last of vvhich, f^"^ (x), is a constant.
THEORY OF EQUATIONS
Thus,
if
461
4,
f{x)
we have
/'
{x)
/" (x)
/"'{x)
/""(x)
= x* - 8 + 4 x2 - X + = 12 x^ - 24 + 8 x - 1, = 36 x2 - 48 X + 8, = 72x-48, = 72.
.S
a;3
x'-
the
f(x)
f (x).
a^,x"
a^x"-'^
-\
we
848
we obtain
f(x
-f
h)
a,(x
(x
70" (x
By expanding
//)",
a.
the bino-
mial theorem and then collecting terms, we can reduce this We shall show expression to the form of a polynomial in h.
that the result will be
f(x
+f"(:x)
^,
(I)
For when the result of expanding (x- + //)'" by the binomial 561, is multiplied by a constant a and written in the form
a (x
//)'"
= ax'" -f max'" -^
l)ax"'~^,
m(jn
1)('"
2) ax'""',
the derivative of the one which immediately precedes Hence, if we arrange the expansion of each term of
it.
f{x
//)
of the leading terms in these several expansions will be f(x) the sum of the second terms will be
;
=a,{x sum
+ hy +
a, (x
7/)"-'
462
A COLLEGE ALGEBRA
of the leading terms, or
f'(x)h\ the sum of the third terms will be A^/2! times the
sum
and
so on.
of the derivatives of the second terms, or /"(a;)/i^/2! In other words, we shall have
f{x
Thus,
f{x)
we have
/(x
h)
a',
h)-
ao (x
^)
Uz
849
850
THEORY OF EQUATIONS
that/' (a;)
is
463
is
by (x
f'{x)
it
ay
is
Again, iif(x)
it
divisible
(1)
follows from
that
/(a) =/'()
general,
ii
=0
is
but/"(a)^0, and
by
a*
divisible
a but not by (x
f(x)
divisible
by (x
a)''
And
in
/''"'
but follows from (1) that /(a) =/'() /<'-!> (a) ^0, and therefore from (2) that f'(x) is divisible by )'". a)''^^ but not by {x (x Therefore, by 800, we have the following theorem.
= (a)
a)'''^^,
is 7iot a root of i' Theorem. J simjyle root of f(x) (x):=0 ; is a simple root but a double root of f (x) 0, aiid, q/f'(x) ts a root of in general, a inultiple root of order r of f (x)
851
order \
\ of f (x)
'
0.
roots of /'(x)
= -
x"
8
x2
~ 8x +
12
are 2, 2,
3,
and the
are 2,
4/3.
therefore have the following method for discovering the Seek the highest 0, if there be any. common factor oif(x) a,ndf'(x) by the method of 465. If
We
852
that /() and /'(a;) are prime to one another, But if we find that /(a;) and has simple roots only. f'{x) have the highest common factor <^(a;), then every simple root of ^ (.) = is a double root of 0, every double root f{x)
we thus hud
f(x)
of
if
(f)
(x)
(.r)
is
= 0,
and
so on.
is
</)
(x
is
divisible
For,
850,
divisible
by
+^ a)'"
Observe that
of
F(a-)=0
Example.
if the quotient /(a;)/ <^ (a-) be F(x), the roots are those of /(a-) 0, each counted once.
roots,
x-^
if
highest
and /' (x) to be <^ (x) = x^ 3x 2. The roots of ^(x) = may be found by 802 and are 1, 1, Hence /(x) has the triple root 1 and the double root 2, that is,
Here /'(x)
0.
by
common
its
roots are
1,
1,
1,
2, 2.
464
Observe that
that
A COLLEGE ALGEBRA
f(x)
f'{x)
= =
and that
F{x)
if
+ +
1)^ (x
1)" (X
2).
853
any
it
and
i/^(x)
common,
factor
may
x^
common
oif{x) and
x*
3 x^
0,
and the Evidently the two roots mentioned are common to /(x) = x) = x* + x* Sx^ 4 x 4 = 0, and may therefore be equation /{ obtained by finding the highest common factor of /(x) and/( x). By 465 we find this highest common factor to be x^ 4. Hence
the roots mentioned are
2,
2.
Dividing /(x) by x^
0,
we
two
EXERCISE LXXVI
1.
Find the
first,
second,
derivatives of 2x5
4 x*
4-
x^
20 x.
2. 3.
Given /(x)
~x*
2x^-\-\, find/(x
h)
by Taylor's theorem.
Using the formula 849, (II), express (1) x* + x^ - 32 in powers of x - 2 (3) (x^^ + 1) / (2) x5
;
+
(x^
in
1)
powers of in terms
of X
4.
roots.
Solve them.
0.
5.
6.
= x3-3x -2 = 0. (2) 9x3 + 12x2 -llx + 2 = 0. (3)4x^ + 12x2 + 9 = 0. (4) x*-4x3 + 8x + 4 = 0. (5) 2x* -12x3 + 19x2 -Gx + 9 - x3 - 4x2 - 3x - 2 = 0. (6) x6 - 2 x3 - x2 - 4 X + 12 = 0. (7) x< - x - 2 x3 + 2 x2 + X - 1 = 0. (8) x6 = 0. (9) 3x5-2x* + 6x3-4x2 + 3x- 2 Show that x" a" = cannot have a multiple root.
If the
equation x^
12x +
a=0
+
a.
7.
Determine a and
b so that 3 x^
may have
a triple
root,
and
THEORY OF EQUATIONS
8. 9.
465
Show
that x*
qx"^
+s=
tliat x^
cannot have a
triple root.
px- +
may
laave a
double root.
?
10. 11.
What
is
the
form oif{x)
if it is
and
x*
x^
The equation
the roots of x*
13.
x^
one of
Show
that
if
Show that if an
equation of the fourth degree/(x) = with rational must be rational unless /(x) is a
perfect square.
15.
Prove that
if
a
0,
is
a root of /(x)
0, of
order
r, it is
a root of
all
f"(x)
0,
/('-i)(x)
0.
Let
x,
f (x)
854
and let b denote the numerical value of its leading coefficient and g that of its numerically greatest coefficient. The leading term of i(x') ivill be numerically greater
powers of
than the sum of the remaining terms for are numerically less than b/(b + g)Eirst, let /(a-)
let x'
all values
of x which
b^x
boX^
so that b
= [b^],
and
x.
Then
gx'
^a-^"
is
to)
gx''^
or g (x'
x'<
1, it is less
x') is less than b when x' <b / (b g). so that b =\b-^\. We + b^x^ -f b^x^ Second, let/(ir) = then have \b^x'^ + b^x^ < |6ix| when \b2X + b^x"^ + \< \bi\, that is, when x' < b / (b + g), and so on.
But
gx' / (1
-{-
than gx' / (1
+ -
), 235,
x'),
704.
ftia:
-\
-\
Thus,
if
x'<5/(5 +
9),
/(x) = that
|3x2
9x*|<|5x| when
466
855
Theorem.
A COLLEGE ALGEBRA
Let
f (x)
and let a denote the numerical value of its leading coefficient and g that of its numericalbj greatest coefficient. The leading term o/f(x) ivill be numerically greater than the sum of the remaining terms for all values of x which
ing jiowers of x,
are numerically greater than (a
g)/a.
h a, so that a
x.
ic"
For let/(a-)
let x'
a^x"" +
ajic"-'
-^
= |ao|,
1-
and
Hence
But, that
(a
+ a^/x
-\
a"/x").
1-
when
|
|
a \>\ai/x-\
\-a/x^\.
854, ||>|ai/a; H
a/x"\
is,
when
7)/
x'
>
when l/x'<a/(a
2]
g),
(a
+ g) / a.
when
Thus, if/(x)
3x3
+ x2- 7x +
'>(3
3,
that
is,
From this theorem it evidently follows that the number (a + g)/a is greater than the absolute or numerical value of any root of the equation = 0, whether the root be real or imaginary. f{x)
856
If & is a root of f (x) = 0, the values of{(x) and have contrary signs when x is slightly less than a, and the same sign when x is slightly greater than a.
Theorem.
f'(x)
For express /(.r) and /'(a-) in powers of x then divide the lirst expression by the second.
simple root, so that f{a) reduced to the form
a,
849,
and
but/'()
^ 0,
the
The numerator and denominator of the fraction on the right are polynomials in a; a. Hence for all values oi x a which are small enough to meet the requirements of 854 their signs wull be those of their common term and the
leading
f (a),
according as
a;
<a
or
ic
> a.
sign of f(x)/f'{x) will a and therefore minus or plus But when the sign of f(x) ff (x)
The
THEORY OF EQUATIONS
is
46?
of f(x)/f' (x)
minus, /(a*) and/' (a;) have contrary signs, and when the sign is plus, /(a-) and/'(a:;) have the same sign.
Wlien a
f(x)
^
is
we have,
850,
f'(x)
a)
from which the theorem follows by the same reasoning as when a is a simple root.
Rolle's theorem.
roots of f (x)
there
is
For
a
let
and
^S.,
number Then
be the roots in question, and let c denote and d a number slightly less )8i
/B^.
than
p., so
that
(Si
<c<d<
same sign as f(c), 856, and /(c) has the same sign as f(d), 834 but f(d) has a different sign from that of /' (d), 856. Hence /' (c) and /' ((/) have contrary Therefore a root of f'(^x)0 lies between c and d, signs. that is, between ft^ and ^o, 833. is 2 x - 3 = 0. The Thus, if f{x) = x2 - 3 X + 2 = 0, then /'(x) =
f'(c) has the
;
roots of /(x)
=
2.
are 1
and
2,
=
+
9
is
lies
between
and
lOx
Since /(I) = 1 and./"(2) = 1, there are two roots or none between 1 and 2. If there are two roots, /' (x) = must also have a root between 1 and 2, and this root must lie between the two roots of /(x) = 0. 10 = But /'(x) = 3x- + 2x has a root between 1 and 2, for 5 and /' (2) = 0. Solving, we find that this root is 1.5 approxi/' (1) = Moreover /(1. 5) = .375 is minus. Therefore, since both /(I) mately. and /(2) are plus, /(x) = has two roots between 1 and 2, namely, one between 1 and 1.5, and another between 1.5 and 2.
Theorem. If the variable x is increasing, then, as it passes through the value a, the value o/f (x) is increasing if i' (a,) > 0,
but decreasing i/f'(a) but f"(a) i/'f'(a)=
< 0. ^ 0,
(a) is
maximum
f"(a)
when
f"(a)
<
0,
minimum
value
when
>
468
For by
A COLLEGE ALGEBRA
849 we have
- a)+/"()(x of 12 + .... /(x)-/(a) =/'() (X is a polynomial in x a, and for all The second member values of x which make x a small enough numerically to
!
of
We
sign of the entire expression and therefore that of /(a-) /(a). Then shall suppose x restricted to such values.
1.
If
has the same sign as (x a). Therefore, since x a changes from minus to plus as x passbt. through a, the same is true of f(x)f(a'), that is, /*(.t) is then increasing from a value less than f{a) to a value greater than f{n).
2.
f{a)
>
0,
f{a){x
a),
If /' (a)
<
0,
/' {a) (x
fl)
and {x
1,
a)
have contrary
that f{x)
is
we conclude
the sign
a.
If
and therefore that of /"(a); for a)'^ is positive whether x < a or x > a. Hence when (x /"() < 0, we have/(a') <f(a) just before x reaches a and also just after x passes a, which proves that /'(a) is a maximum
that oi
ay /2
but/"()^0,
oif{x)-f{a)
is
value of /(cc), 639. And the same manner when/"(fl) f(a) is a minimum, value of /(a-).
>
0,
we may show
that
It
is
may
2).
be added that
if
/"(a)=0 but/"'(a)^0,
/(a)
not a
maximum
And,
or
minimum
Ex.
in general, if all
l)th, vanish
when x
a,
f{a)
r
is
when
+ 9x
?
is
when
even.
Example.
values of/(x).
Ls/{x)
x^
Gx'^
increasing or decreasing as x,
Find the
We find/'
is
(.x)
3x2
- T2x +
is
3(x
1) (x
Hence/' (2)
2.
=-
negative.
Therefore /(x)
THEORY OF EQUATIONS
469
We have/' (x) = when x=l and when x = 3. Moreover/" (x) = G x 12, Hence is negative and/" (3) = 6 is positive. and therefore /"(I) = Lis a minimum /(I) = 3 is a maximum value of /(x), and /(3)=
value.
Let us now consider how the value of a Variation of f (x). polynomial f(x) with real coefficients varies when x varies oo to + cc. (jontinuously, 214, from
2x2 x + 2. Example 1. Discuss the variation of /(x) = x^ The roots of /(x) = are - 1, 1, 2, and/(x) = (x + 1) (x - 1) (x - 2). oo, /(x) = co when x is between oo and 1, Hence, when x = = 1, /(x) ?-0 when x is between 1 and 1, /(x) is negative when x = 1, /(x) = when x is between 1 and 2, /(x) /(x) is positive when x when x is between 2 and oo, /(x) is when x = 2, /(x) = is negative
859
= co. = 3x2-4x -1 = are (2 V7)/3, or - .2 and 1.5 When x < (2 - V7) / 3 and when x > (2 +^7) / 3, /' (x) when x is between (2 V7)/3 and (2 f V7)/3, /'(x) is =
/(x)
is
Therefore,
to
858,
to
(2
+ V7) / 3
It
to
00.
follows from this, 639, that /(x) has a maximum value when V7) /3, and a minimum value when x = (2 + V?) /3. This is in 4 is negative when x = (2 V7)/3, agreement with 858, for/"(x)= 6 x
(2
and positive when x = (2 + v7)/3. The variation of /(x) will be exhibited to the eye if we put y f(x) and then construct the graph of this
the
accompanying
figure.
The
points A, B, C at which the curve cuts the X-axis are the graphs of the
roots
-1,
1,
2of/(x)=:
0.
The por-
correspond to positive values of /(x), those below to negative values. The uppermost point on the curve between and B corresponds to the maximum value of /(x), the lowermost point
between
and C
to the
minimum
value.
4T0
As X
from an
varies
A COLLEGE ALGEBRA
from
infinite distance
to oo the corresponding point on the curve moves below the x-axis upward and to the right through
above the x-axis. term of /(x), the graph of y =/(x) would be shifted vertically upward and the points B and C would at first approach coincidence in a point of tangency and then disThe corresponding roots of /(x) = would at first become equal appear. and then imaginary.
to
infinite distance
an
we were gradually
Example 2. Discuss the variation of /(.x) = x* 2 x-^ -f- 2 x 1. The roots of /(x) = are - 1, 1, 1, 1, and/(x) = (x -)- 1) (x - 1). Hence when x = oc, /(x) = oo when x = 1 /(x) = when x < 1 and when x >1, /(x) is positive when x is between 1 and 1, /(x) is
;
, ;
negative.
are
between
1/2
= 4 x^ - 6 x2 + 2 = 2 (2 x + 1) (x - 1)2, and the roots of 1 /2, 1, 1. When x < 1/2, f (x) is negative when x
;
and
and
also
is
positive.
Therefore,
858, /(x) is continually decreasing as x varies from co to 1/2, and continually increasing as x var es from 1/2 to 1 and from 1 to oo. Hence /(x) has a minimum value when
1/2, but
is
it
agreement with
1 /2, /" (x) is positive bi;t when /" (X) = and /'" (x) it 0.
The graph
of y
indicated in the accompanying figure. The point where the curve merely
cuts the X-axis corresponds to the root 1 of /(x) = 0, and the point B where
860
in these examples, so in general, iif(x) is of odd degree, oo leading coefficient being positive, when x varies from to +00, /(a-) increases from oo to the first maximum value,
its
As
first
on,
and
finally
THEORY OF EQUATIONS
increases from the last
471
It is possible,
minimum
value to
+ oo.
however, that there are no maximum or minimum values, for the equation /'(a:) = 0, being of even degree, may have no real
The graph of y =f(x) extends from an infinite distance below the x-axis to an infinite distance above the cc-axis. It once at least. crosses the a*-axis an odd number of times,
root.
On
if
f{x)
is
decreasing from
minimum value and ends by increasing from the last minimum value to + oo. In this If case the graph of ij =f{x) need not cross the a; -axis at all. it does cross the axis, it crosses an even number of times.
oo
to the first
In most cases we can obtain a sufficiently accurate representation of the graph of y =f(x) by the method of 889, which consists in assigning a series of values to x, computing the
corresponding values of y, plotting the pairs of values of x, y " thus found, and passing a '' smooth curve through all these Such a curve will indicate roughly where the true points.
a;-axis
and where
its
But
we must
positions of the
0. the equation /' (a-) To every multiple root of /(.r) there corresponds a point of tangency of the graph with the If the order of the multiple root is odd, the graph X-axis. also crosses the a;-axis at this point.
EXERCISE LXXVII
1.
Discuss the
its
variation
of
finding
- 2)2 = a;3 - 3a;2 + 4, /(x) = (x + 1) (x values if any, and draw the graph of
y=f{^).
2.
(3) x3
(5)
(7)
(2)
(4)
(6)
2).
(8)
x3
(j;
(j;
-I
472
3.
A COLLEGE ALGEBRA
Find the graphs
of each of the following fractional equations
a;
1,
by
1/2,
0,
1/2,
1,
, 4.
(1).=
(j;
't:'\. - 2) (X - 3)
i^)y(X
"^'-'^
1) (X
3)
STURM'S THEOREM
361 be any equation which has Sturm functions. Let f(x) = no multiple roots, and let/i(a;) be the first derivative of /(x). Divide f(x) by fi(x) and call the quotient qi, and the remainder, with its sign changed, /g (^)' Again, divide /i (a;) by fo{x), and call the quotient q.^, and
its sign changed, /s (a;). so on, modifying the ordinary process of finding the highest common factor of f(x) and fi (x) in this respect only the sign of each remdlnder is changed, and care is taken to
And
make no
Since /(a;)=
other changes of sign than these. has no multiple roots and therefore /(cc) and
fi(x) have no common factor, 851, we shall finally obtain a remainder which is a constant different from 0, 465. Call this remainder, with its sign changed, f,^.
The sequence
of functions
consisting of the given polynomial, its first derivative, and the several remainders in order, each with its sign changed, is called a sequence of Stiinn or a sequence of Sturm functions.
862
Relations
among
the
Sturm
functions.
These functions
are,
by
definition, connected
f{x)^q,f,{x)-f{x),
Mx) =
/2 (x)
= =
(2)
(3)
fm-ui^)
(^
- 1)
THEORY OF EQUATIONS
From
1.
473
these equations
we conclude
that
Two
same
value of
X.
if
Thus,
from
is
(2).
c,\t follows both/i(a;) and /a (a;) vanish when x that fz{x) also vanishes; therefore, from (3), that
is 0.
But
this
contrary to hypothesis.
2.
When
functions
fj (x), i^
(x),
-,
f^-i
(a;)
it
vanishes, the
fmctions ivhich
immediateli/ 2^recede
and folio iv
Thus,
it
if
/2(c)=0,
/i
a,
(c)
= f (c).
be
Sturm's theorem.
Let
and h
863
of
the sequence
r
,
f m
the
=
b,
which
lie
between a and
b.
To
tinuously from a to
214.
b,
As X
varies
from a
to
unchanged, and the only changes whrchr are possible in the signs of the remaining functions, and therefore in the number
of variations of
^n in the sequence
occur
0,
are such as
may
equations f(x)
1.
fi(x)
0,
and so
on,
835.
But
The mimber of variations in the sequence is neither increased nor diminished when any function except the first^
f (x),
changes
its sign.
474
A COLLEGE ALGEBRA
e is
a root of f^ (x)
and that
changes from plus to minus as x passes through c. Since c is a root of /g (*) = 0, it cannot be a root of either
f^ (x)
lies
number h
And if Ave take a positive or fs (x) 0, 862. so sniall_.that__rio r oot of fijx)_^_0 or of fsix)^
c
between
7* and
c,
or
between
g_
and
its
-\-
h,
neither of
sign as x varies
7^,
c, f^^ix) and/3(.r) have opposite signs, 862. Suppose that fi(c) is plus; then fz(c) is minus, and we have the following scheme of the signs of f-^{x), f^ix), /^(x) for values of x between c h and c -\- h
:
But when x
+ =
flix)
f2{x)
fslx)
=c
THEORY OF EQUATIONS
475
In other words, the sequence has a variation between f{x) and/i(.r) just before x reaches c, and this variation is lost as
X passes through
c.
of
varia-
But, on the other hand, it loses a variation each time that x passes through a root of. f{x) 0,
from a
to
and in
is
the
number
oif(x)=
is
x passes which lie between a and b. In the number of these roots, as was to be
we
rnultijAe voots,
which has apply the method of 861 to an equation f{x) = we obtain a sequence/ (x),/i(x), ,/,(x), (1) the last term
Divide all the highest common factor of all the terms, 465. thus obtain a sequence of the form 0(x), the terms of (1) by /,(x).
of
which
is
We
01
(x),
1, (2)
which, as
is
easily
shown, possesses
all
the properties on
(x)
of roots of
0,
6,
1
852,
the
number
and
and in <p{b), 0i(6), , 1. And this difference between the number of variations in /(a), /i(a),
/i
(^)i
by fm
{a)
fm (b) and /,
for multiplying the sequences </> (a), variations. {b) respect ively will not affect their
864
many
It
different
also
enables one
find
how
many
of these roots lie between any pair of consecutive integers and therefore in every case to solve the problem of locatBut this method of locating roots is very ing the roots. laborious
fails.
and
is
836
476
Example
1.
A COLLEGE ALGEBRA
Locate the roots of x^
3 x^
4x
+1=
0.
- 4x + 1 and/j (x) = 3x2 + 6x - 4, Here/(x) = x-' + 3x2 Arranging the computation of the remaining functions as in have
468, 3,
we
3+ 6-
Hence
/(x)
/i(x)
/2(x)
/3
3
1,
15
defined in
Observe that the fi (x) and /s here obtained are not the fi (x) and /s 861, but these functions multiplied by positive constants* We should have lessened the reckoning had vFe divided /i(x) by
/2(x)
2(x
.5)
synthetically.
It
is
method in the final division. 1. "When X is very great numerically, the sign of a polynomial Hence the following table. of its term of highest degree, 855.
X X
that
=
=
f{z)
/i(x)
Mx) h
THEORY OF EQUATIONS
.477
The only property of the final function / of which any use made in the proof in 863 is that its sign is constant. Hence if, when computing the Sturm functions oi f(x)= in order to find the number of roots between a and b, we come upon a function j^ (x) which has the same sign for all values of X between a and h, we need not compute the subsequent
is
For it follows from the proof in 863 that the functions. required number of roots will be the difference between the number of variations in f(a), ,fp{a) and in f(b), ,fp(b).
Example
Here
/i
3.
(a;)
Sx^
How many real roots has f{x) = x^ + x'^ + x + l = 0? + 2 x +1, and, since 2^ < 4 3, this is positive for
all
The
signs of /(x),
Hence we need not compute /o(x) and/3. - 00 are -, +; for x = 00 they are fi{x) f or x =
has one real root.
+.
Hence /(x)
EXERCISE
LXXVm
By
1.
aid of Sturm's theorem find the situation of the real roots of the
following equations.
x3 x3
G x2
5X
+
0.
13
= =
0.
2.
x3 x3
4 x2
10 X
41
8
0.
3.
+ 5x +
4.
0. 0.
0.
5.
a;3_a;2_i5a;4.28
6.
7.
9.
8.
3x2
+ 8x +
10.
X*
2 x3
6 x2
8X
0.
By
11. 13.
number
4x3
x"
0.
0.
12.
14.
X*
+1=
X*
+ x3 + x2 + X + 1 = 0. - 6 x3 + x2 + 14 X - 14 =
0.
15,
roots.
be an equation of the nth degree without multiple Let /(x) Show that the condition tliat all the roots of /(x) = be real is that
there be n
and that
16.
all
+ 1 terms in its sequence of Sturm functions /(x), /i (x), -,/ the leading terms of all these functions have the same sign.
By
aid of the theorem in Ex. 15 prove that the condition that + px + g = be real and unequal is that
4p3
27 q2 be negative.
478.
A COLLEGE ALGEBRA
SYMMETRIC FUNCTIONS OF THE ROOTS
865
Theorem
are
^^
^,,
,^,,so
h b
(x
= - ft,),
(1)
Substituting x
+ h for x in (1), we have - /30 + A] [(x - ft,) + h-\ [(X - ft,) + A], f(x + h) = [(x
can reduce each
h,
(2)
We
nomial in
the
first
second by continued multiplication, as in 5ob. Since (2) is an identity, the coefficients of like powers of h in the two polynomials thus obtained must be equal, 284.
+ h)=f(x) +f(x)h^
is
ft.,)
{x
- ft.^ {X -
Hence
fts)
(x
f'(x).
ft.,)
- A) (^ - A)
^^
so on.
+ ix-fti)(x-ft.,)
x-ft,x-ft,x-ft./ ft?^ =zf(x)/{x ft^, and
since (x
ft,)
(x
This reasoning is applicable to an equation of any degree n. In the general case there are n factors in the second member of (2), and when this member is reduced to the form of a polynomial in }l, the coefficient of h is the sum of the products x y8, taken ?i 1 at a of the binomials x ft,, x ft,,
,
time, 558.
i)roducts
fti
may
ft,),
THEORY OF EQUATIONS
Thus,
if
479
2) (x
/(x)
/'
we have
and
(x
2) (x
= x^ - 6 a;2 + 11 X = 3 x2 - 12 x + 1 - 3) + (x - 3) (x - 1) +
(x)
1
=
-
(X
1) (x
3),
(x
1) (x
2)
3 x2
12 x
11.
Theorem
equation
2.
The sums of
f (x)
roots of
866
coefficients.
is
f{x) = Let
a,
ft,
ic^
b^x"
b.x
&3
0.
(1)
s^, S2,
s^
have
the meanings
Si
(3
y,
sj
'
/3'
y',
s.
a'-
yS'-
y^
are to prove that s^, s^, in terms of the coefficients ^1, ^>o,
1.
We
By
tlie
preceding tlieorem,
865,
we have
X
^^
Since f{x)
is
a
.r
a-,
fi
a?
y
and x
divisible
by
;8,
y,
each of
the fractions in (2) represents a polynomial in x which can be found by the rule of 410. Applying this rule and then
adding the
results,
we have
0^2
^-i)
+ {a + ^ + (' + ^^1^ + ^2) fix) /(x-(3)==x-' + (fi-^b,)x+ {P' + b,fi + b,) - y) = x^ + (y + b,)x-{-(y' + b,y + b,) f(x)/(x f (x) =3x'+ (s, + 3b,)x+ {s, + b,s, +
f{x) / (x
a) =
b,)
(3)
But by
definition, 847,
we
also have
f'(x)=3x^-
2bix
b2.
(4)
we have
(5) (6)
&i
&i,
..
Si
S2
^i-si
+ 3 62 =
Z>2,
.'.
= bi, 2 &2S2 = ^1
480
2.
A COLLEGE ALGEBRA
From
a,
Si
and
s^
:
we can
obtain
s^, s^,
/3'
(7)
(8) (9)
Adding
these identities,
^3
we
obtain
0,
^182
+ boSi + 3 ig =
(10)
which gives s^ rationally in terms of bi, b^, b^, s^, s.2 and therefore, by (5), (6), rationally in terms of bi, b^, b^. Next multiply the identities (7), (8), (9) by a, yS, y respecWe obtain tively and add the results.
Si
biSs
b^Sz
b^Si
0,
(11)
which by aid of
i>i,
terms of
b^, b^.
And
by
a"^,
in like
/?-, y^,
manner,
/3^,
if
we multiply we
Se
y^,
respectively series of
results,
^32
^l4
^2^3
0,
^1^5
^'2-^4
^'3^3
0,
which give
S5, Sq,
rationally in
terms of
By
may
be proved for an
If a, 1. 7 denote the roots oi x^ - 2 x- + ^ x + 2 = 0, find Sl/an=l/a + l/^ + l/7, 2 l/a2 = 1/^2 4. 1/^2 4. 1/^2^ 2 1/a'' = l/a3 + 1/^3 + 1/73. Applying the transformation x = \/y, and dividing the transformed y + / 2 = 0. equation by the coefficient of y^, we have y^ + 2 y'^ For this equation, by substituting 61 = 2, 62 = 1, &3 = /2 in the formulas (A), (6), (10) above, we obtain Si = 2, S2 = 6, S3 = 31/2. - 2, S l/a^ = 6, S l/aS = _ 31/2. Therefore, 814, S 1/a =
Example
/3,
THEORY OF EQUATIONS
Example show that
Si
2.
481
b^x"^
=
Ss
x*
6ix^
63X
64
0,
Si
+ +
61
61S3
62,
61S2
85 4- &1S4
+ +
62S1
62S3
+ +
4 63
&3S2
= 63, + &4S1 =
0,
and compute
in terms of
The preceding formulas also show that s^, s^, s^, are inteyraZ functions of the coefficients of the equation when, as in is 1. (1), the leading coefficient
Theorem
2.
867
roots
868
of an eqiiation f(x)=
of
its coefficients.
be
or,
/?,
y,
v.
of a,
(3,
can be
2a:'',
and so on, 544. Hence it is only necessary to prove our theorem for functions of these several types. This was done for the type 2<;i:'' = s^ in 866, and we shall now show that Str''^, and so on, can be expressed rationally
in
s^^.
-\
.
a''^'
+
)
ji"a''
The product
is
/?'
(a'
+^
(1)
the
sum
of the
(2)
(3)
aPpJ
+ p'\x'>^
and (2) are rational functions of the coefficients, Hence (3), or 2a''/8^, which may be obtained by sub 866. tracting (2) from (1), is also such a function. Since (1) is s^s^ and (2) is s^+g, we have the formula
(1)
2a''iff^
But
= V,-P +
-z-
(4)
When p = q,
becomes 2
'S^a^pp.
We
then have 2
Set''/?"
= s^ s^^.
482
2.
A COLLEGE ALGEBRA
The type ^a^^^Y
(a''/3''
a'^jS^'f
-\
/3''aY H
The product
is
+ /S"a''
) ('"
/T +
y-"
(5)
the
sum
(6) (7)
J^
(8)
But we have already shown that (5), (6), and (7) are rational Hence (8), or 2a''/3V5 which functions of the coefficients. be obtained by subtracting the sum of (6) and (7) from may
(5), is also
such a function.
the group (8) becomes 2 '%aPj^Y'i
When jj = q,
it
when^ = q = r,
becomes 6
3.
"^aP/S'^Y'-S.a''l3''y't',
The types
and so
on.
coeffi-
We may prove
cients
by repetitions of the process illustrated begin by multiplying 2a''j8''y'" by a" + ^8" + y"
Example.
and
2.
We
Show
that
EXERCISE LXXIX
1.
aox^
aix^
a^x
a3
=
qx
find S3
and
s^ in
terms
of Uo,
2.
2i
ots-
If a,
/3,
+ px^ +
r =: 0, find
21/ a^,
S 1 /a^, and S
3.
terms of p,
q, r.
Find the equation whose roots are the cubes of the roots of
0.
x3-2x2 + 3x-l =
4. If a,
/3,
x^ + 3 x + 4 = 0, 7 denote the roots of the equation x^ symmetric functions of these roots by the
(2)
methods
(1)
of 866, 867.
Si, S2, S3, Si.
Sa'^.
(3)
2as^7.
(4) 2a3i327.
(6)Sl/a4.
(6)
Sa2^/7.
CUBIC
483
found exactly or approximately, and it is possible to extend the methods there employed to the complex roots of such equaAs hardly need be said, these methods are not applitions. To solve such an equation we must cable to literal equations.
obtain expressions for its roots in terms of its coefficients. say that an equation can be solved algehraically when its roots can be expressed in terms of its coefficients by apply-
We
ing a finite
number
We have already proved, 631, that the general quadratic equation has such an algebraic solution, and we are now to prove that the like is true of the general cubic and biquadratic But general equations of a degree higher than four equations.
cannot be solved algebraically. In 646 we showed that the equaCube roots of unity. tion x^ = 1 has the roots 1, (- 1 -f t V3)/2, (- 1 - i V3)/2.
870
Hence each
will be
of these numbers is a cube root of unity. The third found to be the square of the second. Hence if we
w^.
represent the second by w, we may represent the third by 1 =.0 lacks an x^ term, we have, 805, 1 -f w + w^ Since x^
= 0.
Similarly every number a has three cube roots, namely, the a. If one of these roots be three roots of the equation a;^
a.
Cardan's formula.
By
the method of
818
871
px
+q=
(1)
which the
a;'^
term
is
lacking.
484
In
(1)
A COLLEGE ALGEBRA
put x
= y + z.
(2)
We
or
obtain
y'
0.
(3)
As the variables y and z are subject to the single a second condition upon them. (3), we may impose
condition
We
suppose
(3), y^
^yz
-\-
and therefore, by
z^
From
(5), (4),
y^
+ p = 0, + q = 0. + z' = -q,
y^z^^-j?l21.
(4)
(5)
(6)
and from
(7)
Therefore,
+ y_ ^3^27 =
0.
(8)
Solving (8) and representing the expressions obtained for the roots by A and E respectively, we have
y%
These equations
namely,
s,
870,
= Vl,
3/
3/
3/
(10)
a)Vl, w^VZ,
But by
(4),
= V^, yz = p/3,
0,-^, 0)2-^.
(11)
of y, z in (10), (11)
y,
which
-y/li
;
v'J,
w v'j
w'-'
^/^
w^ -v^,
w V^.
we
namely,
where
.=_|,.Vfr|,B=-|_V^|-
(12)
CUBIC
Example
1
.
-ISS
Solve x^
+ 6x q
0.
By
+ py +
The
v^4 + V2, Vi w + V2 0)2, Vi w2 Hence the roots -of the given equation are
2
+ V2 w.
3
3
+
2.
3^3 V4
+
V2, 2 +
V4 0, + V2 ^2,
+ V4 ai2 + V2 w.
6x
Example
x"^
0.
When /> and q are real, the charDiscussion of the solution. acter of the roots depends on the value of y^/4 -\- p^ Jll as follows
:
872
1.
7/"
q^/4
+ p^/27 >
0,
one root
are real.
is
imaginary.
For
in this case
A and B
Hence
x^ is real,
and
x^,
870.
If q^/4
For
and
by
ccg
= 0, all the roots are real and two equal. = 2 V^/2, in this case A = B = q /2. Hence = = (w + or) V///2 = V^/2, since w + w^ = 1,
-f-
p'/27
a-j
3/
r-r
3/
a;,,
870.
3.
If (^ /^
p'/27
<
0, (7^^
and unequal.
This
l>ut
may be proved by Sturm's theorem (see p. 477, Ex. 16). when q"^ / ^ -\- p^ /27 < 0, A and B are complex numbers,
and though the expressions for a-j, X2, x^ denote real numbers, they cannot be reduced to a real form by algebraic transformations. This is therefore called the irreducible case of the
cubic (see
885).
is
873
its
vanishing
is
the condition
635).
486
874
A COLLEGE ALGEBRA
The general biquadratic. Ferrari's solution. By the method of 818 every biquadratic equation can be reduced to the form
,
X*
-\-
ax^
+ bx + c =
0.
(1)
a view to transforming the first member of (1) into a difference of squares, we add and subtract xhi ?*^/4, where
With
is
to be found.
We
thus
or
4- ?/^/4
-\-
(2)
To make
we must have
or
M^
(3)
Let
is
i denote
in u.
When m
replaced by Uy in
is
the square of
V?<i
aX
b /2 Vmi
x'
a,
and
(2)
two quadratics
+ V^^T^ ^ + f^' \
-^
, 2
,-
Vmi
a^ J
0.
(4)
x'^-V^^;:^ax+('-^
\^
+ jt==.\=0.
2 "Vi
aJ
(5)
We may
and
(5).
1.
Example
Here a
Solve x*
6
1,
4, c
+ =
M^
x'^
+
u2
4x
.3
0.
Is
3,
12 w
x^
28
=
x
0.
One
is 2,
and setting mi
and
= 2 in = 0.
(4), (5),
we have
we
obtain x
Example
2.
= ( 1 V5)/2, (1 -t i 4 x'' + x^ + 4 x + 1 = 0.
vll)/2.
As
the cubic (3) has three roots any one of which may be the method (5), it would seem that
CUBIC
487
above described might yield 3-4 or 12 values of x, whereas the given equation (1) can have but four roots. But it is not
difficult to
prove that the choice made of u^ does not affect (4), (5) combined, but merely
and
(5).
Reciprocal equations.
875
or not a given reciprocal equation, 815, has either of the roots 1 or 1, and if it has, derive from it a depressed equation
^ (Vr) = which has neither of these roots. 815 that this depressed equation <^ (a;) =
aocc^"'
from
must be
ao
of the
form
+ aia;2'''- +
'
a,a;"'
+ a^x +
0,
(1)
is, degree must be even, and every two coefficients which are equally removed from the beginning and end of ^{x) must be equal. We proceed to show that by the substitution z x -{-\ Jx
that
its
this equation ^ (j-) may be transformed into an equation in z whose degree is one half that of It will then <f)(x)= 0.
follow that
if
we may
the degree of (j>(x)= be not greater than eight, by aid of 631, 871, 874.
Dividing
by
x"'
X"
find that
y(-;)-(-+.-^-)^<>
and
if
=x+
x^
1/x, and
in (3)
we
set
1, 2,
succes-
sively,
we have
+ x^ = z' - 2, \
x'
+ \ = z'-3z, x^
(4)
cc*
+ ^, = z*-4.z' + 2,-..,
488
that
A COLLEGE ALGEBRA
cc^ l/x^ an expression of the ^th is, we obtain for Substituting these several expressions in (2), we degree in z. obtain an equation of the ?/ith degree in z, as was to be demon-
strated (compare
645).
Example L
This
is
Solve 2x8
-xT_
12 x^
14x5
_ HxS +
12x2
0,
1.
2x6
Dividing by
x^,
a;5
lOx*
13x3
10x2
0.
Hence by
or
Solving,
(4),
2 (z^
3z)
(22
2)
lOz
13
2z3-z2_i6z +
z
15
= =
0, 0.
= l, -3, or 5/2. x + l/x = l, -3, or 5/2, Hence and therefore X = (1 iV3)/ 2, (-3V5)/2, 2, or - x^ + x* + ^2 - x + 1 = 0. Example 2. Solve x"
876
1/2.
a;"
may
be reduced to the
x Vay, the result reciprocal form by aid of the substitution 1 of this substitution being y" (compare 646, Ex. 2).
+ =
877
amplitude.
Expression of a complex number in terms of absolute value and In the accompanying figure, P is the graph of the
complex number a
in 238.
-\-hi,
constructed
as
The length
sent
it
of
OP
is
hi,
Vo-^
239.
h\ the
ahsolute value of a -f
Repre-
by
r.
We call 6 the amplihide of a + hi. We call the ratio b/r the sine of
6,
= sin
$.
CUBIC
489
We call
= cos 6.
We
thus have
a
a
-\-
and therefore
bi
= r cos 0, r (cos
and a
b
-[-
i
r.
sin 6).
When
cosO
0,
then
Hence
sin
= 0,
and
878
1.
The
circular
measure of 360
is
tt,
is
given equally by
r,
r,
niTT,
where
^ + 2 tt by m denotes any
?,
tt
and, in general, by
integer.
2 mir.
Theorem.
The
879
numbers
tude
is
and
its
ampli-
For
r (cos 6
sin 6)
0'
r'
(cos
6'
sin
6')
=r
r'
[(cos 9 cos
i
sin 6 sin
6')
6')']
+ +
$')
(9')
sin $ sin
5', ^'.
cos 6 sin
240
is
By
repeated applications of
r'
879 we have
880
r (cos e
sin 6)
(cos
6' ^'
sin
6')
sin 6")
J, ^',^.'f
[cos(^
^"
+ ...) + sin (^ +
i
^'
Corollary 2.
in 880,
we
:
r' r" ^" , and Setting r obtain the following formula, known as Demoivre's
+ ^" + ^ = e' =
)].
881
theorem
^
[?'(cos 6
-\-
sin 0)]"
490
882
Corollary 3.
A COLLEGE ALGEBRA
For a quotient we have the formula
.
.
r (cos ^
-^ r
For
+ i sin ^) = r^ ,^ ^ ,^ , l / /j 7 sm ^) y' [cos {6 -e') + i sin (e - e')]; + ~ O')']- r' (cos 6' i sin 6') ^ [cos (6 $') + sin {6 = r(cos^ + isin^). 879
..
.
.
(COS ^
-\-
883
Corollary 4.
,
The n
ni\i
roots of a
(-
sm
?i
1).
when
For
A;
is
0, 1, 2,
.
.
(w
r
\_
?"
V
\
cos
2k'jr
,
n
i(sin 6
\- 1
e sm + 2^-7^^~l"
11
yj
t
= r [cos (^ + 2 A-tt) +
884
+ 2 k-K)'] = r(cos ^ +
sin 6).
881, 878
The n wth roots of r(cos 6 Binomial equations. the w roots of the equation a;" r(cos 6 + i sin ^) in particular, the n roots of the equation a;" r real and therefore 6 is 0, are
Vr(cos 2
kir fn
sin 2 kir/ii), k
0, 1,
?i
1.
are Thus, the roots of the equation x^ 1 = cosO + isinO, cos2 ;r/3 + i sin2 7r/3, cos4 7r/3
isin4 7r/3,
(-l + iV3)/2,
(-l-iV3)/2.
885
In Trigonometric solution of the irreducible case of the cubic. the irreducible case of the cubic x^ 0, 872, 3, the px q For expressions .4 and B, 871, are conjugate imaginaries.
+ =
q'^/4:
+^^^/27
is
negative,
we have
-^-i-V-M'^-l-W-M?
Hence the expressions for ^4 and B in terms of absolute value and amplitude, 877, will be of the form
(1)
CUBIC
491
where
\ J.
A.
//
//
(2)
and
r
jp
and q are given we can find the value of Q from tliat of cos B by aid of a table of cosines. In the formulas for the roots of a;^ + ^x + g- = 0, 871, (12), substitute the expressions (1) for A and 5, and the The results when expressions for w and w'^ given in 884.
simplified are
jTi
When
o =2
i
/'
^ COS -
o
r
a-2
o 2
r^
cos
+
o
27r
>
ccj
r'
cos
+ 47r
o
And,
to
and
Q being
known by
(2),
Here
(^
/^
r-1/ V27,
Hence by
Xi
cos ^
=-
and simplifying, we find V3/2, and therefore d = 150. and cosines we obtain
= -^ cos 50 =
V27
Xs
7422
x^
1.
1371
.3949.
V27
EXERCISE LXXX
Solve equations
1. 3.
x3-9x-28 = 0.
x3 x3
X* x*
x3
9x2
3X
== 0.
4.
6.
x''
+ 9x -8 = - 6 = 0.
7
0.
5.
7. 9.
+ 9x-l = 0. + 6 X + 1 = 0. + 12x-5 = 0.
3x2
x2
3x='-9x2 + 14x +
X* x*
0. 0.
8.
4 x3
x2
+1=
x
10.
8x3
+ 12x2-l]
3
0.
11.
Solve
0.
492
12. 13. 14.
A COLLEGE ALGEBRA
Solve 2x Solve 6 x"
on which, by
875,
the solution of x^
-1=
t- c
=--
depends.
15.
all
the roots of x^
3 ax^
3 6x
be
real.
16.
and
of x^
+1=
0.
17.
x3
3 X
0.
(2)
is
x3
_ 6x -
0.
18.
base
is
a right prism with a square the volume of the prism is 27, what is its altitude ?
V3
19.
The volume
of a certain right circular cylinder is 50 tt is 105 tt/ 2. Find the radius of its base
and
and
its
its
The altitude of a right circular cone is 6 and the radius of its base In this cone a right circular cylinder is inscribed whose volume is four ninths that of the cone. Find the altitude of the cylinder.
20.
is 4.
XXXI.
886
Inversions. Odd and even permutations. When considering the permutations of a set of objects, as letters or numbers,
we may
normal
fix
order.
Any
many
inversions as
it
followed by one which in the normal order And precedes it. is called odd or even according as the number is odd or even (or 0).
Thus, if the objects in normal order are the numbers 1, 2, 8, 4, 5, the permutation 45^,12 lias the eight inversions 43, 41, 42, 53, 51, 52, 31, 32. Hence 45312 is an tven permutation.
493
887
Theorem. If two of the objects in a permutation are interchanged, the number of inversions is increased or diminished by
an odd number.
two adjacent objects are interchanged, the number of is increased or diminished by 1. Thus, compare ApqB (1) and AqpB (2), where A and B denote the groups of objects which precede and follow the interchanged objects p and q. Any inversions which may occur in A and B and any which may be due to the fact that A, p and q precede B are common to (1) and (2). Hence the sole difference between If joj' is (1) and (2), so far as inversions are concerned, is this an inversion, qp is not, and (2) has one less inversion than (1) but if p)(l is not an inversion, qj) is, and (2) has one more For
if
inversions
inversion than (1). But the interchange of any two objects may be brought about by an odd number of interchanges of adjacent objects. Thus, from ^ja^^y we may derive qahiJ by five interchanges of
adjacent
letters.
We
first
interchange
2^
and we
then interchange q with each preceding letter, obtaining aqbp, There is one less step in the second part of the process qabp. than in the first part, because when it is begun q has already been shifted one place in the required direction. Had there been ^l letters between p and q, there would have been ft + 1
steps in the
(/x 4-
first
1)
/A,
or 2
yu,
in the second,
and
Therefore, since each interchange of adjacent objects changes the number of inversions by 1 or 1, and the sum of an odd
number
theorem
of
is
is
1 or
1 is odd, our
It will
Thus, if in 21457368 (1) we interchange 4 and 6, we get 21657348 (2). be found that (1) has^?;e inversions and (2) eight, and 8 5 is odd.
Of the n\ permutations of n objects taken all at a time, For from any one of 763, half are odd and half are even.
888
494
A COLLEGE ALGEBRA
these permutations we can derive all the rest by repeated As thus obtained, the permutainterchanges of two objects. tions will be alternately odd and even, or vice versa, 887.
are odd
Therefore, since 7i is an even number, half of the permutations and half are even.
!
889
In what follows
we
shall
with subscripts, as a^, a.^, h^, b^, , and so on. Having chosen any set of such symbols in which all the letters and
subscripts are diiferent, arrange and then find the sum of the
letters
letters
and of inversions of the subscripts. If this sum is even, it will be even when the symbols are arranged in any other order if odd, odd. For when any two of the symbols are interchanged the inversions of both letters and subscripts
;
are changed by odd numbers, 887, and therefore their sum by an even number. In particular, the number of inversions of the subscripts when the letters are in normal order and that of the letters when the subscripts are in normal order are both odd or both
even.
Thus,
in a^hzCi the
number
in c^a^hs the
sum
is
of the
number number of
four.
890
Definition of determinant.
We may
in general n-
numbers
in the
3^,
! 02
495
With the elements of such an array form all the products that can be formed by taking as factors one element and but one from each row and each column of the array.
(letters)
In each product arrange the factors so that the row marks are in normal 07'der, and then cotmt the inversions of
the
column marks
(subscripts).
0),
give the product the plus sign; if odd, the minus sign. The algebraic sum of all these p)lus and minus products
is
called the detei'minant of the array, and is rejjresented by the array itself with bars written at either side of it.
Thus,
496
893
894
895
A COLLEGE ALGEBRA
The products above described, with their proper signs, are called the terms of the determinant.
To expand a determinant
length.
is
to write
out
its
terms at
The diagonal
of elements
is
a-Jj^c^,
is
The leading term enclosed by bars, thus, \ai used as a symbol for the determinant itself.
896
b^ Cz
is
],
often
The number of the terms of a determinant of the nth order n and half of these terms have plus signs and half have minus signs.
is
!,
Por, keeping the letters in normal order, we may form n\ permutations of the n subscripts, 763, and there is one term of the determinant for each of these n\ permutations, 890.
888.
we have
It
or 6 terms
for
4,
we have
or 24.
897
must be remembered that the letters and subscripts are mere marks of row and column order. Any other symbols which will serve this purpose may be substiOther notations.
an
Thus, the elements of a determinant are often represented by a single letter with two subscripts, as 023, the
first
indicating the
Thp
symbol
023 is
898
A rule for expanding a determinant of the third order. To obtain the three positive terms, start at each element of the first row in turn and, so far as possible, follow the direction
of the leading diagonal, thus
aib^c^, 2^3^i>
:
(^abiC^.
497
To obtain the three negative terms, proceed in a similar ^manner but follow the direction of the other diagonal, thus
:
3
1
Tims.
third.
This rule does not apply to determinants of an order higher than the Tims, it would give but eight of the twenty-four terms of a deter-
EXERCISE LXXXI
Expand
the following determinants.
r
1
V
s
P
q
r
s
<1
p
s
498
A COLLEGE ALGEBRA
PROPERTIES OF DETERMINANTS
899
Theorem 1. The value of a determinant is not changed if its rows are made coluvins, and its columns rows, without changing
their relative order.
ai ^2 On
Thus,
900
901
902
499
the rows of (1) may be transformed into a term of (2) with factors arranged in the order of the rows of (2) by inter-
changing
its first
and
last factors
and conversely.
But
this
interchange will increase or diminish the inversions of the subscripts in the term by an odd number, 887, and therefore,
since the
(2), it will
is
123 in both
(1)
and
Thus, UibsCi
of
(2).
a term of
(1)
and
Cib^ao is the
corresponding term
Cibsa-z
For
UobsCi the
Example.
determinants
Corollary.
and
(2).
If two of the rows (or columns) of a determinant are identical, the determinant vanishes.
903
For
902,
let
An
;
inter-
it
will change
into
is,
unchanged
but,
by
D.
Therefore
D^- D, that
b
c e
i)
0,
or Z)
0.
a a d
Thus,
b
c
abf
aec
dbc
aec
abf
dbc
0.
f
If
the
all
Theorem
multiplied plied by k.
4.
bt/
the elements of a
k, the
row
(or column')
is
are
904
same number, as
determinant
multi'
For of the elements thus multiplied "by k one and but one occurs as a factor in each term of the determinant, 890.
The evaluation
of a determinant
may
often be simplified by
6 8 15-20
3
2
5
-3
= 2-5
1 1
1 1
4-1
480.
-1
500
905
A COLLEGE ALGEBRA
Corollary. If the corresponding elements of two columns (or rows^ are pi'oportional, the determinant vanishes.
ra a d
Thus,
906
907
501
Thus,
5-4
-2
5
7.
since
4,7,7
2(5,
-4,
2)
+ 3(-
2,
5,1).
Theorem
integral functions of some variable, as x, vanishes a,. the determinant is divisible bi/ x
For the determinant when expanded may be reduced to the form of a polynomial in x. And since this polynomial vanishes when X = a it is divisible by a; a, 415. The factors of a determinant may often be found by aid of
this theorem.
1
Example.
Show
that
502
1
A COLLEGE ALGEBRA
4.
Prove that
5.
Prove that
503
911
Corollary. If e denotes the element in the ith row and Vth column of A, the sum of all the terms of A which involve e is
(-
+
l)'
''eA,.
For we can bring e to the position of leading element without disturbing the relative positions of the elements which lie outside of the row and column in which e stands, namely, by
first interchanging the row in which e stands with each preceding row in turn and then interchanging the column in which e stands with each preceding column. In carrying out these successive interchanges of rows and columns we merely
902.
shall
Hence,
if
form,
we
have
A'
= (this
+ ^1)'
A = (- ly + ^-A.
which involve e is For the minor of
By
eA'g.
910, the
in
sum
Hence
l)'
^eA^.
e in
is
the same as
its
minor
ai 62 cs
in A'.
(Z4
Thus,
i
in the case of
A=
|
the element
1
(Z3,
for
:
which
= 4,
3,
may
a\ a^ as at
dz di d^ di as ai a^ a^
h h
C3
62 C2
h
Ci
(3)
Ci
By
first
interchanging the fourth row of (1) with the third, second, and in turn, we obtain (2) before which we place the minus sign because
is, i
1,
interchanges of rows.
of (2) with the second
Then by interchanging the third column first in turn, we obtain (3) before which we
before
(2)
and
place the
same
sign as that
is, fc
1,
of ds in (1) is the same as its of all the terms of (1) which involve d^ is
The minor
minor
ds
Theoreifi.
determinant
may
he expressed as the
its
913
504
A COLLEGE ALGEBRA
and minus,
or
2A^
az^^
a^^^
sum
which involve
a^, a 2> as, a.. And, by 910, 911, the "Sj which involve a^ is aiA the sum of all cloA. and so on. t2tA2,
,
In like manner,
A=-
b,\^
b^\^
hz\^
<'Ac^
= i^oi 913
Cofactors.
It
is
h\ +
+ h\^
diAi^,
and so
on.
sometimes more convenient to write the preceding expressions for A in the form
A=
i.4i
^2-4 2
-i-
asAs
a^A^
and so
call
-^"1
on,
-^21
where A^
A^, yl2
Aj,
and so
on.
We
then
1,
the cofactors of
a^, a^,
Thus, in
ai
914
605
expressed
912,
we have
tt
tZ 2
OE' c
bi
bz
ba
506
is !, 0, 0,
A COLLEGE ALGEBRA
. Hence this determinant is equal to a^ times its minor, which is the determinant of the (n l)th order (3). Observe that each element of (3) is obtained from the cor-
column of
(1).
we have
2-1
13-2
2
1
1
2-1
3-2-2
507
(1)
and
905.
And
(4) vanish, since their cohimns are proportional, simplifying (2) and (3) by aid of 902, 904, and
^l?2
508
3
A COLLEGE ALGEBRA
2 6 4
5
6
3.
9
12
509
! ^2 3
510
If
A COLLEGE ALGEBRA
A and
all
the determinants \k
h^
C3\,
\ai
Cg],
\ai h^
m\
vanish, the equations (2), (3), (4) impose no restriction on the values of iCi, x^, x^. In this case the given equations (1) are not
independent.
which
(2),
(3),
This follows, by 394, from the manner in (4) were derived from (1), unless all the
minors A^, A,
may
And if all the minors vanish, it vanish. readily be shown that the three equations (1) differ only by constant factors, so that every solution of one of them is a solution of the other two.
These results are readily generalized for a system of n equa71
tions in
unknown
letters.
921
Homogeneous
equations in
linear
equations.
When
= =
I
7n
0,
the
equations (1) of
x^, x^,
ajxi
+ a^x^ + ^3X3 =
I
(1)
^1^1
^22*2
~r ^3'^3 ^^
'
919 become
Ax^
Aa-i
0,
A:r2
0,
0.
(2)
= X2 = Xs = 0,
and
it
0.
is
A=
But
A=
0,
rAi, Xo
rAo, x^
rA^,
(3)
where r
may
we have
UiAi
a^A^
a3'43
0,
b^A^
CiAi
C2.42
C3A3
+ b^Az = 0,
+ b^As =
0,
one because
A=
0,
the
:
as follows
we solve the second and third of the equations (1) for Xi and X2 in terms of x^, we obtain Xi/ Ai = X2/A = X3/ A^, or, if
511
rA^.
And
as just
shown,
if
A=
0,
satisfy the first of the equations (1). From this second proof it follows that
a-i
:
when A
=
:
.Tj
X3
= .li
J2
-^ 3
^1
-62
i?3
Ci
Cj
C'3,
that
is,
the minors
rows of
922
It is assumed that these minors are not 0. are proportional. From the system of three non-homogeneous equations in x, y
/>3
(1')
we may
Hence
921 by suba
stituting X
= a-i/cTg, y = x^/x^ and clearing of fractions. A = is the condition that the equations (1') have
EXERCISE LXXXIV
common
solution.
2x + 3?/-52 =
1.
3,
x-2?/ + z = 0, ^ Ux + y + 3z = 7.
(
2x + 4y-3z 3x - 8y + 6z
!
= = =
+ + + -
3,
1,
4.
az
-\-
by
cz
d,
4.
3.
a?x
IP-y
(f-z
d^.
= -3, = -l, 3 = 9, = 2. 7
4 5
Show
them
that the following systems of equations are consistent, and solve for the ratios x-.y-.z.
x
+ 2y-2 =
0,
3x-t/ + 4z =
,4x-|-?/
7
.
0,
0.
<!
ttoX
.
+ 3z =
asx
+ +
hoy
&32/
-f (A-a2
(A-as
+ +
^^2) 2
^^s) z
= =
0,
0.
4X
2/
Xx,
Zx - Ay +
X
-\-
1 z
1 y
Qz
= =
\y,
\z.
512
A COLLEGE ALGEBRA
RESULTANTS
933
By the resultant of two equations /(x) = and meant that integral function of the coefficients of /(x) and ^ (x) whose vanishing is the necessary and sufficient
Resultants.
(^
(x)
is
condition that/(x)
and 0(x)
have a common
root.
aiX
aih
+
-|-
ag a2
!&
a2
for
when
6.
a^ffi -f
= = 0,
(1)
and
.x
=
(1)
(2)
is
the equations
and
(2)
have the
common
root
924
any two equations f(x) = 0, <^ (x) = may be obtained by eliminating x by the following method due to
The resultant
of
Sylvester.
To
fix
the ideas,
let
ttox'^
f(x)
cf,
= = (x)
-r aix^
b,x^
+
1,
b,x
+ a^x + + ^2 = 0.
(2)
Og
0,
(1)
(2)
x,
and
by
x"^,
and
0, 0,
1 successively.
We
obtain
bfp:*
linear equations in the five quantities x*, x^, x"^, x, 1. 921, they cannot have a common solution unless
^0 ^1
rt ^'2
fli
^'2
homogeneous Hence,
'^3
*^
^3
0.
Z>
/>0
/^l
(3)
^'i,
^>i
b.
b^
bo_
,
b^
Hence
common
fifth
(3) is the necessary condition that (1) and (2) have a For to the root. It is also the sufficient condition.
column of
add the
first
x*,
513
determinant,
/(ar),
X respectively. We thus transform D into an equivalent 907, whose fifth column has the elements a-/(.T),
(x),
X(f>
x^^
(x),
cji
(x).
Hence,
if
denote the
column of D, we have,
Q1
"^
'
D = {fi^x +
from
fio)f(x)
(fi.sx'^
4-
fJ^iX
/xs) <^
(x)
It follows
= 0. /3 oi f(x)
must be a
factor of
fj-iX
+ /aj) ^(a^),
/(x) is of the third degree and /xgo;^ + fiiX + fi^ is of only the second degree, that at least one factor x (3 oif(x) must be
a factor of
^(j^"),
in
f(x)
fx^r'-fH-i ^^^
^ot
all zero.
are 0,
it
can be proved
root.
and
<^(.r)
any
925
row
of D,
it
follows from
when
D=
I
',
',
X X
Ao
A3
A4
A5,
= =
A2/A3
A4/A5.
Therefore
when
have a common
root
is
and when the degrees of f(x) = and n respectively, the resultant D will be a determinant of the (in + n^ih order whose first n rows consist
In the general case
(li(x)
Ai/Ao.
are
926
of the coefficients of f{x) and zeros and rows consist of the coefficients of <i>{x)
as ii} 924, (3). Hence in the terms of D the coefficients of f(x) enter in the degree of ^ (x) and vice versa. Example. By the method just explained show that the equations and x + 1 = have a common root and find this root, x'^ + 3z + 2 =
.13 2
Here
D = 110 =1 + 2 3 = 0,
1
so that there
is
common
of
root.
The values of the cofactors of 1 and 3 in the first row and 1. Hence the common root is 1 1, that is, 1.
:
are 1
514
927
X,
A COLLEGE ALGEBRA
By the preceding method either of the unknown letters y may be eliminated from a pair of algebraic equations of
cf>
(x,
y)
= 0.
j/^
x2
x
?/
2 x2
?/2
= =
0,
0.
(1)
(2)
We may
1,
regard
(1)
(2)
and
1,
2y",
and
in x, (1)
2, 0,
Zy.
Hence the
result of eliminating
1
is
515
= <f}(x)
let
and ^S^. any root of denote any root of f(x) = There will be 77in differences of the form a^ fi,^;
(3/,)
n (a^
is
a^
is
(3/^)
the necessary and sufficient conbe equal to one of the roots (3^-
yS^.,
and
<^(a;)
= 0,
of
867, 868.
Hence,
</,
if
B (/,
923,
</>)
f(x)
and
(x)
= 0,
we have
R (f,
The product
11 (a-
<f>)
= U (, -
/?,).
^^)
may
be written
(i
- A) ('^1 - A)
-
(i
(^n),
(2-/8i)(2-y32)---(a2-/8),
/5l)
{<^>:,n
('l-,,.
/32)
("-,
- ^).
(a;
/3i) (x
row
is
^82)
(3),
the product
row
<^(a2).
Hence
= (x
n-j)
(x
a,^)
is
-{x
column
column (
l)"'/(i32),
and so on.
( l)'"/(/8i), Hence
n {a, - A) = (When
f{x)
that
i)"'"/(A) -/(A)
-/(A,).
oX"'
...
+ , = 0,
<t>{x)
box-
^...
= 0,
the leading coefficients are not 1, the product of is, the factors in the first row is (j>(ai)/l)Q, and so on; and the product of the factors in the first column is ( l)"'/(/?i)/a(
when
and so
on.
make
11 (a^
/3^)
an integral
516
A COLLEGE ALGEBRA
and
<^(a;)
we must
multiply
it
by
ao^'u".
We
then have
929
(x)=
(}>
0, <^ (x)
=
and
(x)
0,
to the first degree. involving the coefficients of /(a;) = thus have another proof that the determinant D described
For the product ^ (ai) (^ {a^) <^ (a-,) contains factors, to the first degree each involving the coefficients of <^ {x) = and the product /(/3i) -fifi^) fW} contains n factors, each
We
in 924,
is,
926
is
that
D^ R(f,
and
</>(a;)
0,
that
930
<j>
The sum of
(x)
ofi(x)=
and
in each term of E.
is
mn.
4>(x)
For, by 812, if we multiply each coefficient of f(x) and by the power of r indicated by its subscript, we obtain
two equations
/i (x)
cf)i
^
-|
\-
r'"a
r^b^
-\
f-
= 0, = 0,
whose roots are r times the roots of f(x) = and (x) = 0. Each term of R (f, 4>i) will be equal to the corresponding term of R (f (j>) multij^lied by a power of r whose exponent is the sum of the subscripts of the coefficients off(x) and (f>(x) which occur in the term. Hence our theorem is demonstrated But if we can show that in every term this exponent is mn. since there are m?i factors in the product EI (a, (3/^), we have
(f>
R (f,
4>,)
alb'l
n (/-a, -
7'/3,)
= r- R (/,
<^).
517
\-a^z=0
931
= a^pf^
-\
that integral function of the coefficients of /(x) whose vanishhave ing is the necessary and sufficient condition that/(a:')
0,
then
D = R{f,f')/a,. has a finite multiple root when and For, by 851, f(x) and /' (x) have a finite root in common. only when f(x) By 928, the condition that/(a') = and /' (x) = have a root
in
common
is
R(^f,f')=
0.
But
a^
is
a factor of
in the
a^,
R (/,/'),
as
may
of
be shown by expressing
924.
Hence R
(/, /')
R (f, f)
when
to /(a-)
determinant form
0.
which
816.
in this case is
common
co is
and/'(x)
is oo,
Therefore, since
we have D
ao
-f^
(/>/') /^o-
UiX
0,
D=
518
same degree
''1
\
A COLLEGE ALGEBRA
in Wj, x^^ x^.
I
Thus, x^
"
+ xy + y
-{-
becomes
^l'''^"2-'^3
^2-^3
3^3
in
Observe also that a homogeneous equation of the th degree determines n finite values a-2, x^ which is not divisible by x^
Thus, from x|
x^/x^
3 x^x^ + 2
(x,
a;|
we obtain
934
Let
f{x, y)
and
<^
y)=0
and 7i respectively. denote two equations whose degrees are If they involve the terms x'" and x'\ then by substituting
x
Xi/xg,
X2/xs,
terms,
collecting
= = Xs)
aoXT
boxl
+ +
iaT
-'
b.x'r'
coefficients Uq, !,
+ ,
=
(3
+ +
0,
(1)
bn
=0,
(2)
bo, b^,
denotes a
homogeneous fimction of Xo, x^ of the degree indicated by its subscript. Hence R, the resultant of (1) and (2) with respect to
Xi, is
By
a homogeneous function of x^, Xg of the degree mn, 930. 928 the necessary and sufficient condition that (1) and
(2) be satisfied
Xi is that
(3)
i?
R=
If
0.
is satisfied
by
mn
If
these values, the equations f(x, j8) 0, <^ (x, common root, and if this root be a, then x
a,
fi
is
solution of /(, y) 0, <l>{x, y) (compare 927). Moreover it can be shown that to each simple root of 72 there thus
0,
<f>
=
y)=
(x, y)
and
there correspond
=
(f>
0, all different or
some
finite
(x,
have 7nn
If
R
fi
is
mn
divisible by
xJ^,
then
R=
y,
finite values of
x^/x^ or
519
have only itui ^ finite solutions. But since and y = x^/xs, when x^ = either x ox y or both Xi/X'i X and y are infinite. We therefore say in this case that = have /x infinite solutions. f(x, y')= 0, (x, y)
y)
<f)
by a substitution of the form y = y' -\- ex, into equations of the same degrees which have these terms. By what has just been proved the transformed equations in x, y' will have mn solutions. But if x = a, y' = ^ be any one of these solutions, then x = a, y = ft -\- ca is a, solution ofJ\x, y) also have
7iin
If the given equations f(x, y) 0, and x" terms, we can transform them,
<^
{x,
}/)
lack the
ic"*
0,(f> (x,
y)
= 0.
Hence /(a-,
y)
0,<f> {x,
y)
solutions.
In the preceding discussion it is assumed that R does not vanish identically. If R does thus vanish, f(x, y) and <^(cc, y) have a common factor and therefore f(x., y'^ = 0, (f> (x, y) =
have
infinitely
many
solutions.
:
We
tively
(fi
a7id n resjiecIf i(x, y) and <^(x, y) are of the degrees and have no common factor, the equations f (x, y) = 0, (x, y)0 have mn solutions.
EXERCISE LXXXV
1.
By the method of
924, 925
show
and 2x3
2. 3.
x2
9x
have a
common
aiX
that the equations 6 x2+5 x-6=0 root and find this root.
a2
Form
+
ex
=
d
and
^oX^
+ =
biX
1.
60
0.
bx'^
and
x"
4.
By
the
method
q
(1) x^ 5.
+ px +
0.
(2)
ax'
By
aid of 931
this root.
show that
x^
x2
8x
+ 6x2 + c = 0. - 12 -- has a
double root
and find
6.
3 x?/
2/2
16 X
x2
xy
2 y2
5X
- 28 = 5y =
?/
0,
0.
520
A COLLEGE ALGEBRA
XXXII.
935
Infinite series.
If
u^, u^,
??,
''2
is
704).
For
?/i
?/o
we may write
2?/,
read "
sum
of
ti
to
infinity."
The
real, positive
when
all its
when all its terms Vi, Vo, are terms are positive. In what follows
we
A
Thus,
- v7i/(ft +
by means
the series
of
is
1),
m.
Sometimes such a formula is indicated by writing the first three or four terms of the series. Thus, in 1/2 + 1 3/2 4 + 1 3 5/2 4 6 4 we - l)/2 4 have M = 1 3 2 ji. (2 n
936
71
Let
---,
.S'
denote the
S-^
sum
S
of the first
and
Ui
= v^
S,^
+ Vo + + Vo
so that
-{-
= u^,
u^
+ v^,
-\-
As
increases,
u^, ?<i
+
.S'
no
"s,
will take successively the values i, , and one of the following cases must
present
itself,
namely
will approach
some
finite
number
as limit,
or
or
will be indeterminate.
In the
gen,t,
first
and lim
case the series Ui -f 2 is said to be converis called its suth. In the second and third .s;^
is
said to be divergent.
When Ml +
lim
5,
Mj
is
convergent
by S and write S
= Ui + Ui +
we may
its
sum,
we may
521
number
S.
is convergent Thus, the geometric series 1/2 + l/i + 1/8 + 1/16 + and its sum is 1. For here, as n increases S takes successively the values
and, as is proved in 704, it approaches Observe that here, as in every convergent series, lim w = 0.
as
The
values
series 1
1, 2, 3,
+1+1+
is
divergent.
For
oo.
jS
The
series 1
l+
-.
the values
1, 0, 1, 0,
It is therefore
We
An
937
the
its first
n terms approaches a
Otherwise
it is
finite limit
to
when as n is
sum of
indefinitely
increased.
said
he divergent.
The
limit of the
convergent
This
is
new
the
result of a finite
means
of additions performed consecutively; here it the li7nit of such a result. It must therefore not be assumed that
number
the characteristic properties of finite sums, namely, conformity to the commutative and associative laws, always belong to these infinite sums
(see 941, 961).
gent, a finite
In determining whether a given series is convergent or divernumber of its terms may be neglected.
of the neglected
938
terms
will
have a
+ 2 +
any
(1) is
sum
S,
939
and
is
c is
finite
number, then
+ c2 +
if
(2) is a con-
But
(1) is divergent, so
For if the sum of the first n terms of n terms of (2) is cS and lim CiSi = c lim
;
(1)
/S
be S, the
sum
of the first
cS.
The sum of a convergent series will not be changed if its terms are combined in groups without changing their order.
940
622
Thus,
if
A COLLEGE ALGEBRA
,
and gi, g^,--- denote the sums the given series heui + u^-] two terms, its next four terms, and so on, the series gi + g2 will have the same sum as Wi + ifo + For if M denote the last term in the group g,,,, we have
of
its first
-\
Pi
f/2
\-
Om
Ui
U2
Mn,
is
same
limit as in
and
may
be
shown
its
941
We may therefore introduce parentheses at will in a conIt is also allowable to remove them unless, as vergent series. in the following example, the resulting series is divergent.
The convergent
(11^
series 1/2
+
.
1/4
.
+
1^
1/8
936,
it
may
be written
1)
-I-
(1|
1)
(11
1)
-I-
..
But here
is
not allowable to
IJ
1+
li
1+
is
divergent.
942
It
is
removal of parentheses.
Thus, the
sum
Sn
of 1/1
+
-
1/2
1/3
+
1)
is 1.
For
+ - + 2-3 1-2
-
+
71(71
1223
Hence
Example.
71
71
S=
lim S
lim (1
\
71
=L
is 1
of the series
/n (n +
2).
943
Remainder
after n terms.
If the series i
''2
(1)
is
be convergent,
is
sum
(1).
of (2).
It
called
the
0.
523
A positive
series Uy
<2 -\-
is
convergent
if,
as
944
increases, S
remains always
less
than
som.e finite
number
c.
For since the series is positive, S^ continually increases as n increases. But it remains less than c. Hence, 192, it
approaches a limit.
Therefore,
is
convergent.
Theorem
2.
-|-
Let u^
^2
and
1.
let
a^
convergent.
The
(1) denote a given positive series, (2) denote a positive series known to be series^l) is convergent in any of the cases:
+ U2 +
945
When
is
less
of
(2).
2.
When
Herm of
3.
the ratio of each term of (1) to the corresponding than soine finite nnmbei* c.
tlie
2)receding term
1.
?'i
less
ratio of each term to the immediately than the corresponding ratio hi (2).
For
let
J
i'
of
the
first
n terms of
-I
+ "2 H If Ux < i,
^1 4- W2
a.nd.
U2
< 2.
is
A denote the sum of the series ^i + 2 we shall always have < A. Hence
S'
convergent,
944.
2.
For
if
< <
c,
c,
then Ui
<
ca^, u^
< cag?
a\
a^
+ ra, H
1.
is
convergent,
+
6.
W2
is
convergent, by
ror
II
Ux
ax
^l^
W2
*2
f^h
^3
lis
Uo
then
It follows
u^/a^, Us/as, fore t<i W2
Wo ^4 Us <,<;<,.
a^ ax
as Og a^ as
from these inequalities that each of the ratios Thereis less than the finite number Ux/ax.
is
convergent, by
2.
524
It follows
A COLLEGE ALGEBRA
from 938 that the same conclusions can be any one of the relations 1, 2, 3 holds good for all but number of the terms of the series (1) and (2).
Prove that
1
drawn
a finite
if
Example.
is
+ +
1/2
1/2
+ +
1/2
(1)
convergent by comparing
1
it w^itli
1/2
1/2
1/2
(2)
the second is less than the corresponding convergent, by 1, Second. The ratios of the terms of ( 1 ) to the corresponding terms of (2) are finite. Hence (1) is convergent, by 2. namely, 1, 1, 2/3, 2 2/3 4,
(1) after
is
,
by each
1, 2, 3.
Third.
The
-.
Hence
(1) is
convergent, by
3.
946
Theorem
3.
Let Ui
b,
+ Uo
-|
(1) denote
a given positive
series,
to
and
let bj
known
be
divergent.
The
any of
the cases:
1. When each term of (1) is greater than the corresponding term of (2). 2. Wlien the ratio of each term of (1) to the corresponding term of (2) is greater than some positive numher c.
3.
When
preceding term
in (1) the ratio of each term to the immediately is greater than the correspo7iding ratio in (2).
The proof
of this theorem,
which
is
947
945, is left to the student. Test series. The practical usefulness of the preceding tests, 945, 946, evidently depends on our possessing test series
known
to be convergent or divergent.
is
the geometric series a ar -\- ar^ -\which has been shown, 704, to be convergent when r
-,
<
1,
and which
is
1.
Another very
948
The when p
series
+ 1/2p +
1/3p
-\
\-
> 1,
1/nP H
is
convergent
divergent
when p <
1.
525
1/2'',
p>
1.
the four terms beginning with 1/4^, the eight terms beginning with 1/8^, and so on, we obtain the equivalent series, 940,
i
+ U7; + -^ +
ii;
i^
+ .7^+^
(1)
Evidently each term of (1) after the corresponding term of the series
first is less
than the
that
is,
less
+
is 1-
|+^+
p>
1,
---'
^' l
2^ri
But
2.
since
---.
(3)
geometric
series (3)
2^
convergent.
Hence (1) is convergent, 945, 1. Combining the two terms ending with 1/4, the
four terms ending with 1/8, the eight terms ending with 1/16, and so on, we obtain
^Ml-d<l-l-'^^l>
Evidently each term of (4) after the second
the corresponding term of the series
is
(4)
greater than
that
is,
+ 2 + 4 + 8 + "-'^-^ + 2 + 2 + 2 +
divergent.
12
111
is
---
^^^
But
3.
is
(6) is
Therefore (4)
divergent,
946,
-|
1.
p <1.
1/2^
1/3^'
divergent since
terms are greater than the corresponding + 1/2 + 1/3 + , which has just been
proved to be divergent,
946,
1.
526
949
A COLLEGE ALGEBRA
The following exam-
ples will serve to illustrate the usefulness of the theorems of 945, 946.
Example
It is
1.
Show
that 1/1
1/2
1/3
4 H
is
convergent.
convergent because its terms after the first are corresponding terms of the convergent series 1/2^ + I/32
less
than the
I/42
945,1.
is divergent. Example 2. Show that 1 + 1/3 + 1/5 + 1/7 + The ratios of the terms of this series to the corresponding terms
of the
+
all
2.
1/2
1/3
1/4
namely,
1
1,
Hence
1/3
1/5
is
Example
or
is it
3.
Is the series in
?
which m
^
1
{2n
1
+
2
l)/{n^
+
_
n) convergent
divergent
Here
^
11,,-^'"'^^= n^ + n
n^
+ l/n ^ + l/n"
Hence the ratio of u to 1/n^ is (2 an expression which is finite for all values of n, and which approaches the finite limit 2 as n increases. But l/n^ is the 7ith term of the convergent series
1
1/2-
1/3- H
is
convergent,
945, 2.
950
method employed in Ex. 3, it may be proved that if u has the form ?/ =^f(n)/(j>(n), where f(n) and <^(?2.) denote
By
the
is
convergent
;
when
the
degree of
that
it is
<^
by more than 1
otherwise,
divergent.
1.
Example
Show
(3)
11
a(a
b)
{a
b){a
+
2b)
{a
1
1
* .....
-|-
b) (a
36)
Example
2.
Show
V2
V3
Vi
+ 2 V2
+ 3 V3
I+4V4
527
Example 3. Write out the first four terms of the series in which u has each of the following values and determine which of these series are
convergent and which divergent.
(1) ^ '
=
(n
2n-l
+
1)
._^
(n
(2) ^ '
Un
V;^
ri-
.,
2)
(3) ^ '
l
*' Theorem 4. The positive series u^ M2 convergent if the ratio of each of its terms to the immediately preceding term
951
is less
which
itself is less
than
1.
+ u^ + Uz + (1) the ratio of each term to the immediately preceding term is less than the corresponding
For
ratio in the geometric series i
+ v-^r + i?-^ +
(2), since in
is always less than r, while in (2) it (1) the ratio in question 1. Therefore But (2) is convergent since ? is equal to r.
<
1,
the series
is
divergent
0.
Corollary. If as n increases the ratio u + i/Uj, approaches a 1, divergent definite limit X, the series is convergent when A
953
<
when X
1.
>
1.
For
if
<
1,
take any
number
X,
r such that X
<r<
1.
<r
X,
189,
n and there-
Hence the series is convergent, 938, 951. r. + i/ If X 1, after a certain value of 71 we shall always have
<
>
+ i/^'n
>
1-
Hence the
series
is
divergent,
951.
the series is + i/ (m + i/u) 1 and lim(w_^i/?/) 1, no divergent; but when + i/'' conclusion can be drawn from the theorem of 951.
1
When
>
and lim
= 1,
<
Example
1.
Show
3 that 5
3
5
10
3 5
5-7
,^ 10
,^
15
is
convergent.
5 n, and The 7zth term of this series is 3 5 7 (2 n + l)/5 10 15 the ratio of this term to the term which precedes it is (2 n + l)/5 n. - 2/5, which But since (2 n + l)/5 n = 2/5 + 1/5 n, lim (2 n + l)/5 n
is
<
1.
Hence the
series is convergent.
528
A COLLEGE ALGEBRA
2.
Example
,
.
When
?
is
X being positive
u + i
_
l
and
tlieref ore
convergent, 6 2 t2 i 4- S t3 + ^ i + zx + dx X" + 1/n + nx" x" + i(l + 1/n) + l/ft + (i + l)x" + -JLtl _ lim
1 1
\-
1 ^
-I- 3-
-4-
Hence the
series
is
<
'^
1,
that
is,
when x > L
Q
1
Example
Example Example
3.
Show
that 1 1
'
1-4
4-7
1
'-
1-
is
convergent.
4.
Show that ~
22525
-\
13
1
13
*
-|
\-
is
convergent. ^
5.
When
"When
is
- H
1
Example
positive ?
6.
is
+ xi+x
1
+x
convergent, x being
953
Series in which lim (Uj, ^ ^/Un) =1. ratio u^ + i/n can be reduced to the
*n
In a series of form
+ i/n
where lim
= (o-/n)
0.
that
if,
as
becomes and remains greater than some number which is itself greater than 1, the series is convergent but that if a ultimately becomes and remains less than 1, the series is divergent.
increases, a ultimately
;
1.
have
<T
n,
which we may
call k,
we
Then
=
1
^-
Un
<
1
a/n
+
+
I
(1
when n^k.
a)/n form
i],
i<~[nUn
k
{n
l)u +
I
when n>fc.
(1)
set
=
1
k,
+
+2
1,
+
Ujt
successively,
inequalities.
MA-
We
+
obtain
rii
+
+
+
I
i<-[kUk
{k
+l)u/,+
of the
less
,].
(2)
I
It follows
from
(2) that as
increases the
sum
first
terms of
finite
Uk + 2
remains always
than the
529
There-
is
convergent, 944.
is convergent, 938. fore the complete series Wi + Mj + 2. Suppose that when n > A; we have < 1
Then
But
^^^^^
=
1
^-
Un
is
>
a/n
+
^
1
when n>k.
1/n,
the ratio of the corresponding terms of the divergent 1/(1 + 1/n) for l/(n + 1) ^ 1/n = 1/(1 + 1/n). series 1 + 1/2 + 1/3 +
Hence
tlie
given series Ui
ug
is
divergent, 946,
3.
If a remains greater than 1 but approaches 1 as limit, the preceding test will not determine whether the series is conver-
But in this case " can be reduced to the form a = 1 + y8/, where lim /8/?i = and if /? remains less than some finite number h, the series is divergent.
gent or divergent.
;
For since
/3
< 6, we have
1 1 1 ~ ^ ^ + <:r/n + 1/n + /3/n2 1 + l/,i + b/n'^' But 1/(1 + 1/n + h/n-) in turn is greater than the ratio of the corre sponding terms of tlie divergent series 1/(1 6) + 1/(2 6) + 1/(3 h)-\
Mn + l u
_ ~
For
(n
1
=
+ l)-&
n-h
n
smce
111
=
6
n-h
(n-6) +
n
1
=
l
1
+
-\
l/(n
6
1
62
<
6) l
l/n
+ ft/n^
=
b/n
1 n
(-
n^
n^
U2
is
divergent, 946, 3.
which
?/
series in
954
is
convergent when
a'
>
1,
divergent
when
its
a'
a<^
1.
this fraction
by
numerator, we have
Un
ni>
a'nP
_ ~
1
(a'
- a)/n +
Pn/n'^'
where ^
is finite.
Example.
g-^
1-7
is
+ l)/3(^+l)
/3
a(a + l)(a +
2)^(;3
1-2 -37(7
1) (7
> 0,
divergent
when 7
+ 2) + 2) a ^ 0.
+
l)(^
)3
630
A COLLEGE ALGEBRA
EXERCISE LXXXVI
series are
convergent or divergent.
2+1
2
+
22
+1
23+1
+
4.
1-2
1-
1-3
1.2-3 1.3.5
f-'
4
-\
2.3
1
3-4
1
3
+
4
1
5.
V3
i
2
8.
V3
+T- +
V3
2-4.6
4-7 10
4
6.
i+
4
.
+
a2
+2
+
4
7
+
2
fi
ji
10
(3
+
1)
+
+
3 6 3 4
4
1
4
1 2-
4^
+
1.3 2T4
2_^
+
+
2.3-4--.
(71
1)
4.5.6.
(n
S)
9.
13.5. -(2n-l) +
2.4.6...2n
Write out the first four terms of the series in which u has the following values and determine whether these series are convergent or divergent.
n
10. u
n{n
12.
+1 + 2)
1
.. 11.
Vn
u
VnF+1
-n
Un
= Vjl2 +
VnF+l +
Determine whether the
series in
n
which w + i/u,i has the following
u +
j3
Wn +
i_
2
2n
14.
Un
n+ 3
For what
15.
,
3 5
1
3-6 5.8
3.6.9
5-8.11
X-2
16.
1
+
1
+
X2
1
+
X3
1
X"*
+
,
+
X*
17.
Show
is
a
that
1
+
,
a(a
'
\)
a(a
1-2
l)(a
2)
1-2.3
Vv,,
+
,
is
divergent
when a
18.
positive.
If for all
values of n
Ui
we have
is
< r, where
is
positive
it
and
less
than
1,
show that
+
r^
Un
+
r''
convergent by comparing
with the
convergent series r
531
By
definition, 937,
if 5'
an infinite
approaches
955
a finite
But, 195, 197, S^ will approach a limit if the sequence of values through which it runs as n increases, namely, S^, S^,
Ss,
number
which
less
possesses the property that for every given positive 8, however small, a corresponding term S^ can be found
differs
8.
than
numerically from every subsequent term S/^.^^ by If this condition is not satisfied, .S' will not
approach a
Since
limit, 198.
and
we have
.S-^.^^
n,,
-j
h A.+p,
-\
\-
u^^^.
:
test of convergence
convergent if for every given j^ositive number 8, however small, one can find a term u^ such that the sum of any number of the terms after w^ is numerinfijiite
Any
series Ui
Ug
is
ically less
than
8 ; in other words,
such that
|% +
for
it is
^^k
2H
H%+p1<
all values
of
p.
If the
divergent.
in particular,
cannot be cona series + 2 + = 0. But this single condition is not sufficient for convergence. We must also have lim (?< + ti+i) = 0, + + + = 0, and so on. lim(i< -f
Hence
?<i
?/
,)
^
/
is divergent altliough lira u = lim \/n = 0. 1/3 + For in this series the sum of the k terms which follow the term \/k is
Thus,
1/2
1/2.
-1
A;
+ iyi: +
Hence
A;
1-
111
Thus,
>
k-\-k
2k
tI
2k
1-
,,, to terms,
+
i.
.^1,1 > 2k
e.
or 2
Uk + ic'is
less
than every
divergent (compare
94:8, 2).
532
956
is
A COLLEGE ALGEBRA
1.
Corollary
convergent.
let
its
given series, and u'l + u'n 4(2) be the same series with the signs of all Then negative terms changed.
let
For
Ui
+ 2 +
(1)
be
the
|%+i
I't+o.
-\
+ ifk+p\ <
li'k+i
+ 'a-+2 H
we can make
h u\+p.
Hence,
if
+ %+^J.
Therefore (1)
957
also shows that a series with imaginary terms is convergent if the series whose terms are the absolute values of u^, ?/, , 232, namely, the series ''i + "2 + is convergent.
111
>
Thus, i/l
is
iV22
i-VSS
is
convergent since
+ 1/22 +
I/32
convergent.
958
Corollary 2.
and
negative
is
numericaUg
less
than the
and if the
For
let
positive.
where
a^,
a<,,
are
we here have
We
in the
can write
%. + i
(a,.
form
form
and
in the
a^ +
Since
a^.^.,
>
cik+i^
%+3
->
parentheses in (2) and (3) is (2) that (1) is positive, and from (3) that (1) is algebraically less than o^. + i, and therefore from (2) and (3) combined that
numerically less than a^_^^. since lim a = 0, we can choose h so that a^^i<8. Therefore Oi ag + 3 is convergent, 955.
(1) is
But
533
959
is
Absolute and conditional convergence. convergent real series said to be ahsohiteli/ convcnjent if it continues to be convergent when the signs of all its negative terms, if any, are changed
;
conditionallij convergent if it
Thus,
series 1
Theorem.
bi/
themselves
In an absolutely convergent series the j^osltiue terms form a convergent series, and in like manner the
960
be
negative terms by themselves. And if the sums of these two series P and N. respectively, the sum of the entire series is P
But in a conditionally convergent series both the series of positive terms and the series of negative terms are divergent.
For
infinite
let Ux
^2
of positive and negative terms. Of the first n terms of this series suppose that^^ are positive and q negative. Then if S denote the sum of all n terms, Pp
number
the
sum
of the
j^
negative terms,
we
and
N^ the N^. p^
sum
of the q
When
is
indefinitely,
and since
indefinitely increased both p) and q will increase S,^ will approach the finite limit S, one
of the following cases must present itself, namely, either (1) both Pp and N^ will approach finite limits which we may call
and N, or
both P^ and N^ will approach infinity. In the first case lim 5 = lim (P^ N^ = lim P^
(2)
is,
lim
203, that
P N.
is
N^,
The
iV,
In
sum
Por
of the series
P+
is
first
obtained by changing the signs of the negative terms, lim S' = lim (P^ + N^) = oo.
we have
534
961
A COLLEGE ALGEBRA
The terms of a condltionalhj convergent series may Corollary. be so arrancjed that the simi of the series will take any real
value that
may
be assigned.
Eor, as just shown, in a conditionally convergent series the positive terms by themselves and the negative terms by themselves each constitute a divergent series the limit of
is 0.
Hence, for example, if we assign some positive number c, and then, without changing the relative order of the positive terms or that of the negative terms among themselves, form 5 by first
adding positive terms until the sum is greater than c, then negative terms until the sum is less than c, and so on indefinitely, the limit of this S, as
?i
is
c.
Hence
EXERCISE LXXXVII
1.
series are
convergent or divergent.
^''3
2.
for
For what real values of x are the following what values are they divergent ?
(1)
^'l_x
X
^
'
Ill
h
l
series convergent
and
+ 2x
X'
l-3x
Z5
1
+
l
1 h
H-(-l)"Jix
X2n-1
absolutely convergent, and Oi, ai, Os, denote any sequence of numbers all of which are numerically less than some finite number c, prove by the method of 956 that the series
3.
If
Wi
3x6
nx2''
is
aiMi
4.
a<2,U2
azUz
is
also convergent.
If
S denotes
sums
S.
the
sum
show
that the
less
oi, ai
Oo, ai
ao
-\-
a^,
than
535
This name
a^x
is
a^pc^
962
ao, !,
are constants.
The values
if
of x and
Ui,
By
jao]
may
+ \aix\ + \a2X^\ +
is
+ \a.^x"\ +
(2) is convergent.
is
When
(2)
(1)
(compare 959). Whether (1) is depend upon the value of x. Hence the importance of the
following theorems.
Theorem
If when x
=b
every term of ao
ajX
+
c,
is
963
numer'ically less than some finite positive x|<|b| the series is absolutely convergent.
number
when
For since
a,fi"
<c
|
for every n,
we have
(;"
= |ai"|of |ao|
<
for every n.
+ [^i^I + 1^2^^ +
|
(1) is less
than
x -
X
c
(2).
But
that
(2),
being a geometric
\x\
series,
converges when
\x/b
< 1,
is,
when
l
when
945,
1.
Thus,
+ 2x +
1
.
x'^
---
converges
is
when
|x]
< 1.
Corollary
If
a,Q -\-
aiX
+
|x|
convergent
when x
b, it is
964
absolutely convergent
when
< |b|.
is
This follows immediately from 963. For since ao + i^ -l convergent when x = b, all its terms have finite values when
b.
Corollary'' 2.
7/* ao
also divergent
when
is
divergent
when x
= b,
it is
965
536
For were a^
A COLLEGE ALGEBRA
is
966
to converge for a value of x which + a^x + numerically greater than b, it would also converge for x = b, 964. Limits of convergence. It follows from 964, 965 that if
all positive values of x for which converges and to a class A^ all for which it diverges, each number in A^ will be less than every number in Ao. Hence, 159, there is either a greatest number in A^ or a least in A^. Call this number X. It represents the Ihnit of the series being convergence of a^ + a-^x + absolutely convergent when [a-] < A, divergent when |cc| > A.
we
Gq
assign to a class A^
a^x ^
Thus, in both x + xV2 + xV3 + (!) and x + x'VS^ + x^/2,^ + (2) the limit of convergence X is 1. Observe that (1) diverges and (2) con= X = 1. It is possible to construct a series in which verges when X X = for example, the series x + 2 x^ + 3 x^ + .
is more frequently called the radius of the circle of convergence. For if we picture complex numbers by points in a plane in the manner described in 238 and draw
at the origin and whose radius is X, the series will converge for all values of x whose graphs lie within it will lie without diverge for all values of x whose
is
graphs
967
Theorem
2.
If in
fi,
ao -f ajX H
fi is
a definite limit
For, by
then
\a^\
^-\a^x\ ^
converges
|.r|< lim a n +
a..
when
lim
*n +
-^
ax"
|c^,)|
<
1,
that
is,
when
Similarly
+ |ffix| +
3-5
5-10
diverges when
> \x\
l)
lim
..
'rt
+1
Example
.
1.
series
X2
+ 3.5--.(2n +
5- 10- -.5/1
,
omce
a +
i
=
2n +
3
=
2
we have u =
lim
2
= -.
+ 3/Ti
2
3/ft
537
2.
Find the
limits of
+ 23x-3 + 22a;-2 + 2x-i + 1 + a;/3 + ccVS^ + x^/3^ + . Here 1 + x/3 + x-/'S'^ + is a geometric power series in x which converges when |x|<3, for o/a + = 3. On tlie other hand, 2x-i + 2'^x-^ + 2^x-^ + is a geometric power
series in
x-i or 1/x which converges when lx-ij<l/2, and therefore when |x|> 2. Hence tlie given series converges when 2 <|x|<3.
Example
x/(l
3.
For what
x)
real values of
+ 3 xV(l + x)3 + converge ? This is a power series in x/(l + x) which converges when |x/(l + x) |< 1, for lim a/a + = lim n/{n + 1) = 1. But |x/(l + x)|< 1 for all positive values of x and for negative values which are greater than /2. Hence the series converges when x > 1/2.
+
2
xV(l
x)2
The binomial, exponential, and logarithmic series. We proceed to apply the preceding theorem to three especially important
968
power
1.
series,
990, namely,
x"
x^
nl
x.
is
convergent for
For here
all finite
values of
=
a+i
n\
lim
-.
(n
(ji
+
1)
=n+
1)!
1.
Hence
2.
= lim
x, that
is, /i
= oo.
The logarithmic
-i
series,
992, namely,
o
1,
1
;
n
divergent
is
when
|x|
> 1.
^or here
I.
3i
=
1
=
1
+
n
a +
n
lim
+
n
Hence lim
~ =
ftn
=1,
lim
+l
+ -")=:11/
1,
that
1.
The
series converges
when x =
958, diverges
when x
1,
948.
538
3.
A COLLEGE ALGEBRA
The binomial
series,
namely,
where
is
is
1,
divergent
For here
a
when
>
1.
_7n(in
i
\)
(m
1)
m(m
lira
I)
(??i
n)
a +
12--n
lira
Hence
=
a +
1
lira
VI
l-2---(n+l)
mn
fj.
-t11
X 1001, Ex.
When
When
if
2).
a;
if
m > 0,
a we
l)(
diverges
if
m < 0.
=
1
I
by setting
"^"^^-^^
I
I
m=
(^
may
^)
I
1-2
1-2-3
Evidently from a certain term on all the terms are of the same sign, so that the test of 954 is applicable, 956.
But here
Hence,
or, since
954,
- =
u +
i
+ n~l =
n
if
+ {a-l)
^
(a
it
1)
>
1,
that
if
is, if
m,
it
converges
if
in
> 0.
But
diverges
(a
a > 0, 1) < 1,
that
is,
if
m < 0.
EXERCISE LXXXVIII
Determine the
1.
limits of
series.
+mx + m2xV2! +
+
3(2x)3
2)
.
...
2.
2(2x)2
+
,
2(2x)*
TO (ni
!^
3.
mx +
m (m 5
2!
x2 +
+ 3(2x)5 + 2) (m 4)
'-
x3
3!
For what
real values of
converge
3x
4.
1/ 3x
\--(
\2
+
X
2^x4-4/
/
)+-( 3Vx
/
1/ 3x
\8
)+ + 4/
\3
'
"
\2
x2TT
6.
"^
+ 2x +
(2x)2
(2x)8
539
When
sum
is
which we may represent by = a^ + a^x + In what follows when f(x), writing f(x) we write f(x) = a^ + i^ + , we apsume that a^ + (ii^ + has a limit of convergence A whicl is greater than 0, and
+ a^x +
is
convergent,
its
969
suppose that
|a-|
<
A..
Given that ^ (x) = aiX + ajX^ -{-..., and that when X has the positive value b every term of ^(x) is numerically less than some finite positive number c.
Theorem
970
|^(x)|
If any positive number 8 be assigned, hotvever small, then < 8, whenever |x| < b8/(c + 8).
For, as was
shown
in the proof in
2
963,
when
ja;|<5,
\<i>{x)\<c
and therefore
<
'
,,
that
is, '
\x/b\
< 7^^,^-|^|
8,
704
Hence
that
is,
\(i>(x)\< 8 ^ '^'
when
c\x\ '- ^.
1x1
<
when
i^
f (x)
Corollary,
s* c + = ao + aiX H
I^:^! '
'
<
th en
li^ f (x)
)
= a = f (0).
200.
971
li]i^
{a^x
a^x^
-\
0,
Theorem
2.
//'
the series ag
it
+ ajX +
have
972
converges, then ao
at once
For setting x
= 0,
we
Hence
for every value of
a^x
+ a^x^ +
a^x^
-|-
= 0, ai = a^ = 0. =
(1)
',
x for which
it
converges.
.>''
540
If
a;
A COLLEGE ALGEBRA
0,
we may
ai
x.
Hence
a^x
ag.x'^
=
;
(2)
for
a;
a^ =^
x for which it converges, except perhaps for were follows from this that Oi =
so small (without
making
it
0) that
\a^x
would not
be
satisfy (2), as
just
shown
it
must.
it
Hence
that a^
a^
0,
cts
= 0,
And by
and
a
may
shown
so on.
6x'
The
like is true of
ex
dx'
different
series.
such
vanishes for every value of x for that ao + aix + converges contains more than is required for the proof that For the reasoning above given shows that if ^i, ao = 0, ai = 0, -denote any given never-ending sequence of numbers such Pni Piivanishes when x = /3i, ^2, that lim /3 = 0, and if ao + otiX + A.,
The hypothesis
it
which
then ao = 0, ai be rational.
0,
/Si,
^2,
may
all
973
= bo + bix + bzx'^ H Theorem 3. 7/* ag + aix + agx'^ -| for every value of y. for which these series converge, the coefficients = bo, a^ = bi, a.^ = bj, of the like powers ofiL are equal ; that is, a.^
and
so on.
Tor subtracting the second series from both members of the given equation, we have, by 974,
(ao
a;^
= =
series converge.
a^
Hence,
that
is,
972, a^
flo
S^ = 0, i b^ = = ^i = ^u fh =
^0)
b^
0,
^2)
This theorem
cients.
is
coeffi-
x cannot be expressed in more than one way as a power series in x (compare 421).
It asserts that a given function of
541
many
series only, it
with such
important to establish rules for reckoning These depend upon the following theorems, 974, 976, which we shall demonstrate for infinite series in
series.
general.
Theorem
converge
(Uj
1.
If the
the
series
Uj
U2
(md
the
v^
Vj
+ +
974
and have
sums S
ajid
respectively,
the series
T.
Vi)
+ (u2 + Vo) +
203,
converges
and has
sum S
For,
lim
= lim (i = S + T.
\-
The like is true of the series obtained by adding the corresponding terms of a.ny Ji.7iite number of infinite series. Hence the rule for adding any finite number of functions defined by power series in x is to add the corresponding terms of these
series, that is, the
975
x.
.
Thus,
then
if
/(x)
=1+
/(x)
3 x3
when
when |x|<
1.
number of series whose sums are and we add the corresponding terms of these series, we ordinarily obtain a divergent series, even when the series But in the case described in the is convergent. +- 7' +following theorem we obtain a convergent series by this process, and its sum is 5 + T +
If there be given an infinite
T,
S,
.S'
Let Ui + U2 + denote a convergent series each of ivhose terms is the sum of an absolutely convergent series,
Theorem
2.
976
namely,
Ui
u^P
u<'>
(1),
U2
u^P
+ uT +
(2), ....
542
A COLLEGE ALGEBRA
denote the sums of the series obtained Again, let Uj', U2', their absolute values, by replacing the terms of (1), (2),
by
so that
and
so on.
series Ui' is convergent, the several series U2' obtained by adding the corresponding terms of (1), (2), ,
If the
namely, the
series u<J)
u<p
-\
u^ +
u^^)
-j
and
so on^
and if
their
sums
we
Ui
For
let
Uo
...
+
by
U3
= Vi + V2 + V3 +
. .
. .
943, so that
ci)
U,
(i)
7.(1)^
U,
= u^'p + (|) +
series
+ ) + R^^\
i<*f
-|
,
Each
and
of the
is
column
u'-\^
fp
?<<f
-\
convergent since each of its terms is numericallyless than the corresponding term of the convergent series
Ui + U^' -{-, 945, denoted by V V
If
so on,
1.
V, R,,
we add the corresponding terms of these n we obtain the original series Ui + U^ since n is finite, we have, 975,
series,
-\
.
column
Therefore,
U,+
C^2
---+ f4
---=
Fi+
72.
To prove our theorem, therefore, we have only to show that when n is indefinitely increased lim R = 0. But if the remainder after k terms in R'^]^ + 7?^;' + be denoted by 6<*>, we have R = R^l^ -j- R^^J h R^t? + S^n^
-\
Let
it
matters not
how
small.
7?<;f
-\
543
U^'
-(b),
the remainder
after k terms in (a) is numerically less than the corresponding remainder in (b). But since (b) is convergent we can so choose
k that the latter remainder will be less than 8/2. can so choose k that tvhatever the value ofn have 5^*' < 8/2 numerically.
Hence we
be,
may
u^\'
we
shall
But
w^f
again,
u^:^
row
series
+ ^P +
>
is
will ultimately become and remain R^]^, R^l\ 7?*^^^ numerically less than 8/2 k, and therefore the sum of these + jR^J,'^ will become and remainders, namely, ii^,^ + -R',f + remain less than (8/2 k) k, or 8/2.
remainders
Therefore, as
will
?i
increases, /?
7?'P
R^;^
-\
\-
R^^^
S^';^
ultimately
less
than
8/2
8/2, or
8.
Hence
liui iZ
= 0,
200
and therefore
as
was
to be demonstrated.
series
Ui
U^
is
itself
an
x)
xV(l
x)2
a;V(l
x)3
(1)
which converges for all real values of x which are greater than 1/2 1/2). (also for imaginary values of x whose real parts are greater than Is it possible to transform (1) into a power series in x, that is, into a series which will converge for any value of x except ? When |x|<l, each term of (1) is the sum of a power series which loay
be obtained by the binomial theorem,
988.
Thus,
--
x)
(1
3 x*
(2)
first
,
of these series
is
by
its
absolute value,
we
obtain
|x|
jx^l
[x^] 4-
whose sum
|x|/(l
)x|).
544
A COLLEGE ALGEBRA
we
is
. .
obtain series
|x|)2, \x^\/{l
W+
1/2.
Us'
-]
here
.,
\X\)
|X2|/(1
|XI)2
_ lx|)3 +
|x|
<
Therefore, when |x|<l/2, the power series obtained by adding the - 2 x^ + 4 x^ - 8 x* + corresponding terms of the series (2), namely, x converges and is equal to the given series (1) that is, when x < 1/2 we have - x7(l + x)-^ + xV(l + x)3 = X - 2x2 + 4x3 - 8x4 + . x/(\ + x)
977
We
in a position to demonstrate that the terms of an absoat pleasure without lutely convergent series may be rearranged the sum of the series. changing
now
1.
We may
let ui 4- M2
let Ui
and (1) denote any absolutely convergent series, denote the same series with its terms rearranged. of Again, let 8 denote the sum of the first n terms of (1), and S'm the sum the first m terms of (2). Assign any value to n then choose m so that the first n terms of (1) are
For
u./
(2)
to
be found
first
among
the
(2)
first
terms of
(2);
and
the
finally
first
choose
so that
(1).
the
terms of
are to be found
among
+ p terms of
Then S; - S is made up of terms in S +p S, that is, of terms the sum u + 1 + + 2 + + +/. Hence |S,' -S|<|u + ]| + |h + 2| H + |Un + p|+ |Mn + /)|) = But since (1) is absolutely convergent, lim (|tt + i| +
in
0-
S|
0,
that
is,
lim S'm
lim S,
We
may break
number
(finite or
infinite)
which occur
in the
same order as
original series from every such set of series of the theorems of 974, 976.
Thus,
if
by applying one
we form one
indices,
974,
of those
we
liave,
by
Wl
"2
U3
U4
(Ui
W3
"6
("2
W4
"6
545
:
u^
Ua
as follows
+ + +
U3
Ws Ms
series.
+ +
Me M9
The sum
of Ui
tlie
Mo
+ Ms +
is
equal to the
sum
of
Mio
the terms of
scheme added by rows, 940. the sum by rows is equal to the sum by columns,
And
976.
.
Hence
Mi
+ M2 +
M3 H
(mi
M2
+ M4 +
(M3
M5
) 4
And
3.
^2
ih-\
2.
may
Products of power
defined, when
<^ (a-)
I I
If the functions/(a3)
and
(f>
(a*)
are
978
< X, by the power series/ (x)=a(^-{- a iX-\ (1), = ig + bix + (2), their product /(cf) (x) will be defined, when \x\ < A, by a power series derived from (1) and
<f>
(2)
314).
Thus,
ao
aix
546
A COLLEGE ALGEBRA
sions for y,
repeatedly multiplying (2) by itself, we obtain expresin the form of power series in x which y"^, y^, converge when (2) converges. If we substitute these expres-
By
of (1),
we
2
obtain a series of
H
)
the form
-!
(^o
+ ^i^ H
+ 2 (^o +
bf^b^x
(3),
and
are
(o
this,
when
)
collected,
the terms which involve like powers of x becomes a power series in x of the form
i^o H
+ (i^i + 2 azMi
values
of
-1
)^-\
(4).
This
as
(1)
final
for
x such that
|Jia-|
< A.
For
976
is
satisiied
by the
being
con-
doubly
|o|
series
(3),
the
series
C/j'
Uo'
+ |^il(K^| + i^i'^i +)+ ' which < ^ verges when l^o] + l^r^I +
by
hypothesis
980
fraction
series, as
ajx
-1
)/(/>
b^x H
),
may be transformed into a power series which conwill converge for all values of x for which a^ + a^^x -f
where
b^ =^ 0,
verges and
For
let
\bix\
+ [b^x^l -{-<
y
1
\bf,\.
biJ:
Then
6o
111
+
box'^
-\
.
(1)
bix
62x2 H
bo
ij
60
-f-
y/bo
,2
bo\
since,
60
Oo
by hypothesis and 2-32, |y|<|5ia;| + \b-zx"\ + <|6o|' We In (2) replace y by its value (1) and then apply 979. transform (2) into a power series in x which converges when
\bix\
shall thus
\box'''\-\
<|6o|.
Multiply this power series by Uo + ciix + a^jfi +, 978. The result will be a power series in x which will converge and be equal to the given fraction for all values of x for which a^ -\- a\X 4converges
and
15ix|
162x21-1- _-<|&ol-
547
The quotient series may be obtained to any required term by the process of cancelling leading terms described in 406, or by the method of undetermined coefficients, 408.
Example.
terms.
Expand
(1
+ 2^2
4-
.)/(i
x^
to
four
Using detached
coefficients,
we have
1+2+4+8+ 1+1+1+1+
1
1+1+1+1+
+ +
3
1
+ +
7
1
+ +
---
548
Example
2.
A COLLEGE ALGEBRA
Expand
(1
x)/(x2
4.
x,
We
have
l-x
x2h-4x3
11-x
5x + 20x2-80x3+
80 X
2X
Example
2 x2
x3
3.
Expand
(2 x^
3)/(x3
+
X
4) in
powers of 1/x.
_ ~
1
'
2
1
2X
+ +
1/x
3/x2
2/x
4/x3
I/2 X\
_?_A
X2"'/X
^-?_A_^
X2~x3"
982
Reversion of
series.
From the
it is
defining y in terms of x,
form X
x in terms of y. , defining l\y b^ij^ cess is called the reversion of the given series a^x
The
a^x^^
pro
It will be
a^,
that i
0.
It can be
proved that
has a limit of convergence greater than 0, aiX -f a^piy^ the like is true of the reverted series biy ^2^^ H
+
.
Example.
2x~
Zx^
---.
(1)
Assume
biy
boy-
63?/- H
from the given equation by the method Computing ?/2, y^, and substituting the resulting series in (1), we have
X
of 978,
6iX
2 61
(2)
361
462
63,
...
(3)
By
y
the
same
method,
,
from
an
a^
b^
(y
+- bo
(y
And from an
983
equation of the form y = a^xr +can derive two others of the form x b^y^ + b^y
a^^
0.30;^
we
b^y^-
An algebraic equation of Expansion of algebraic functions. the form f(x., y) = which lacks a constant term is satisfied
when X
and y
0.
Hence,
if
we suppose /(x,
y)
549
solved for y in terms of x, one or more of the solutions must be expressions in x which vanish when x vanishes. It can be proved that these expressions may be expanded in series
in increasing powers of
In ordinary cases these series may be obtained greater than 0. to any required term by the method illustrated in the following examples.
Example
1.
The equation
the equation
2/2
?/2
?/
2a;
of
0.
When
0,
2/
(1)
becomes y^ + y = 0. Since one and but one of the roots of this equation is 0, one and but one of the solutions of (1) for y in terms of x vanishes
wlien X vanishes.
Suppose that when this sohition is expanded powers of x its first term is ax'^, so that y = ax^ +
in a series of increasing
(2)
Substituting
(2) in (1),
a2a;2fi
we have
h ax'^
2x
0.
(3)
an identity, the sums of the coefficients of its terms of like degree must be 0. Hence there must be at least two terms of lowest degree and since /u is positive, these must be the terms 2 x. Therefore /^ = 1 and a 2 = 0, or a = 2. ax'^ and We therefore assume that
Since by hypothesis
(3) is
;
2/
=
c
2X
6x2
+
(4
cx^ H
^20
Substituting Hence 4 + 6
(2')
in (1),
we
obtain
0,
0,
46 +
,
is
4 x2
= + c)x3 + = 4, c = 16, + 16 x^ +
0.
Example
2.
Find the expansions of the values of y in terms of x which xy + x2 = and vanish when x = 0. y'^
the equation
2/3
When
becomes
=
=
0,
xy
x2
=
0.
(1)
y"^
0, all
find three expansions of the kind required. Let ax^ denote the leading term in one of these expansions, so that
2/
ax'^
+
+
(2)
Substituting
(2) in (1),
we have
ax*^
a%3fi
+1
x2
0.
(3)
550
By
A COLLEGE ALGEBRA
the reasoning of Ex.
1,
less
3/i,
/a
1,
in (3).
/u,
= 1/2. This is an admissible value of Setting 3/x = ^ + 1, we find = 1/2, both 3 and m + 1 are less than 2. since when + 1 = 2, we find = 1. This also is an admissible value of Setting = 1, both + 1 and 2 are less than 3^. since when 2/3. But this is not an admissible value Setting 3/i = 2, we find = 2/3, 3/a and 2 are greater than m + 1. of since when
/i
;u /x
fi,
/^
/x
ij.
ytt,
p.
Hence
fx.
When
it
/u
must have one of the values 1 or 1/2. ax~ + = 1, (3) becomes a'^x^ +
.
.
follows that
x^
0,
from which
from
+1=
(3)
0,
or a
=
0,
1.
.
When
which
it
II
1/2,
becomes a%'^
-j-
_
a
ax?
yii
follows that a^
or, since
0,
that a
x^ = 0, =L
2/
form
.
= x^ + 6x +
cx^ H
And
o o = X + x2 + 3x3 +
, ,
...,
2/
3x? +
-
..
In this method
it is
assumed that
is
oa-^,
if
powers of
a-'.
method
general. Having found the leading term ax^ of an expansion as in the examples, x*^ set y It becomes an equa(a v) in the given equation.
fails.
x. From this equation find the leading term of the expansion of v in powers of x, and so on.*
tion in V
and
Thus, in Ex.
2, setting
x"^
2/3
(1 + v) in - xy + x2 =
(1)
and
.simplifying,
we have
v^
Sv^
+ 2v +
x^
-0;
(2)
,
whence v= X-/2H
To
find
and therefore 2/=x- (1 x-/2 + ) = x' x/2H the next term, set v = x^ ( 1/2 + v') in (2), and so on.
,
* For a fuller iliscnssion of the inctlioils of thi.s section and the use in connection with tlictii of Newton's iiaralli'loL;r;ini see Clirystal's Alr/ehra, II, also Fro.st's Curve Tracini/ and Johnson's Curve Tracing pp. 349-371
;
551
<
A,
when
984
f(x
It follows
ag
aj
(a;
+
|
h)
a:;
]
-\-
a^^x
A
|
+ hy +
from
< X we may
transform
(a
+ hy,
power
series in
By
result will be
+ h)=f(x)+f(x)h+f"{x).~ +
(x),
...+f''^(:x).^
+ ...,
where /'
f" (x),
first,
second,
given series a^
-\-
>
so on.
we replace x by a and h by where |a + |a5 a] < A, we obtain the expansion oif(x') in powers of x a, namely,
985
f(x)=f(a)+f'(a)(x-a)+..-+r(a)^^^^ +
From
f(x)
this
last
....
expansion and
971
if
it
follows that
if
986
a^-{-
UiX
when
|a;|
< A,
and
[rt|
<
A,
then
]}iy(x)=f(ay
EXERCISE LXXXIX
1.
Show
Show
that
that
(1
X
x
X-
+
+
) )^
= =
+ 2x +
+ 3x + +
1
Sx"^
+
-^
4x^
".
2.
3.
(1
x"
6x^
lOx^
-.
Show
that
(1
x2
x^
)/(!
+ x- +
UiX
---)
= l-x + x^+
-.
4.
Assuming
that
(1
2 x^)'
aoX^
as, Ui
973.
552
5.
A COLLEGE ALGEBRA
By
(1)
a similar
method
(8-3x)5.
+ x-x2)i.
powers of x
6.
Expand each
2
+ x- 3x2+ 5x3 1 + 2 X + 3 x2
+ 5x2-x3 - X + x2 - x3
powers of x
coefficients.
7.
Expand each
3x2
(1)
1
+
+
x3^
x2
^^^
5x*
x3
2 X*
3 x5
fifth
8.
term by the
method
example
in 981
limits of con-
(x2
9.
^^-^^ - 4) (X -
'^ + ^
(2) ' '
3)
(2x
3)(x
1)2
Expand each of the following fractions to the fourth term in descending powers of x. For what values of x will the first of these
expansions converge
?
2 x2
10.
15
'
'
x^
x3
x2
= X + x2 + x3 + X* +
+
(2)
2/
a;
|A
+^-^+ o 4
11.
From
series for
12.
= 1 + X + x2/2 2/ x in powers of y 1.
x2
x^/Z
From y =
?/^.
powers of
13.
By
the
method
0.
which
and
vanish
when x =
(1) x2
2/2
2,
3X
0.
(2) x3
2/3
x?/
0.
14.
By
"^
show
that
X2 X3
^
X
1
- X2 +
3x3
1
"^
_ ~
(1
x3
X)2
(1
x2)2
(1-X3)2
553
XXXIV.
The binomial
When m
is
a positive integer,
''
m
(^(i
is
(m
1)
?"
.
!)('l-'Z'd
2)
x^
(1)
a finite or terminating series and its sum is ( 1 + a; )"*. is not a positive integer, (1) is an infinite series, but one which converges, that is, has a sum, when |a'|<l,
When m
'
We proceed to demonstrate that if m has any rational 968. value whatsoever, this sum is (1 + .r)'". The series (1) is a function of both x and m, but since we
are
its
relation to
we
shall
represent
that
= VI (ni
if vi
1)
{yn
r+
(1), so
I)//-
!.
Then
= 1 + m.^x + ninx"' + TTizX^ + = 1 + n^x + tioX' + n^x^ + (n) = 1 + {yn + n)ix + (m + n).,x^ + <^ {m + n) We can prove that (w) (ii) = {ni + n). For when < 1, so that (2) and (3) converge, we <i>{m)-4>{n)=l + nu r + vu + i
^ (w)
<^
(-)
(3)
.
have
(4)
</>
<f}
<^
|.r|
554
A COLLEGE ALGEBRA
repeated applications of
(6),
By
we have
<f>
<i>(m)-cf>(n)-<f,(p)= (f>(m
-{-
v)
(p)
4>(m
-\-
p),
finite
number
of factors of the
form ^(w))
all
We
988
are
now prepared
theorem for
//'
he
the
sum
,
,^(m)=l
iw
jx+ \
(1
m(m
,^
x^
.,
2)
^x^+---,
when\x\<
Notice
1, is
+
1
x)"".
first
of all that
-m
when
7n
1,
+
+
a".
Hence
1.
<^(0)=
7n be
and ^(1)=
a-.
(1)
Let
a positive integer.
<j>
Then
(m)
= = =
<^
(1
+1+
<f)
to
to
w
?w.
terms)
factors
<t)(l)
(1)
[cf>{i)r=a
+ ^r,
by(i),
(2)
Let
Then
(p/Q)y
= = =
4>
U'l'l)
<^
{pI'1)
to
(^
factors
^ {p/'l
<^(/0
+ P/<1
p
to ? terms)
= (l+a-)^
^O''-
by
(2).
Therefore
= (1 + <^(;V'?)
(2)
For it follows from the equation [^ (/'/'?)]'' = (^ + ^Y ^"^ 986 that the values which <f>(p/q) takes for all values of x such that |a-l < 1 must be the corresponding values of one and the same (/th root of (1 + a*)''.
555
qih. root,
this root
is
namely,
the only one of the qth. roots of (1 xf (1 which has the same value as ^ {p/q) when x 0.
;
for this
3.
Let
Since
we have
m be any negative rational number = (- 5 + s) = (0) = (s) (- s) = l/<^ (s) = 1/(1 + x)' <^ (- *) = (l+a-)
+ xy
s.
</,
</>
<^
<^
1,
by
(1)
by (3)
(4)
for
of the exponent.
Example.
We
have
(1
Expand + 2x +
(1
2x
=:,
+
[i
a;^)^
in ascending
powers of
x.
3x2)3
(2x
3x2)]3
=l+ +
.
i-(2x
3x2)
+ ilzil
3x2)3
(2
3 x2)2
^
^
3M^^(2X + o
'
...
2x
5x2
68x3
The expansion converges when 2|x| + 31x2|<l; therefore when 9|x2| + 6|x| + l<4; therefore when 3|x|+l<2; therefore when |x|<l/3.
Corollary.
If to.
is
rational
and \x\
<
[a,],
we have
989
-\
= (a + x)
For (a
on ^a.'^-^x^
a;)"*
1 -f
m(m
7H(
1) x^
~\
[.
(1)
\-
7Ti
1^ ^-a"'-^x^-\
(2)
and therefore
<
1,
or
\x\
< \a\.
556
990
The exponential
1
A COLLEGE ALGEBRA
series.
We
liave already
+ VI + ^'72! + x'/3\ +
all finite
its
+
968.
x"/nl
(1)
converges for
values of x,
ic
Let
denote
e
sum when
1,
so that
...
= 2.71828- . 1/2! 1/3! We are to prove that the sum of (1) for any real value of x
l
+l+
is e'.
For
let
/(a;)
(1), so
!
that
f(x) = 1
/(2/)= 1
+
...
x'^/n
..-,
+ yV;i!
+....
Then by the
/(^)-/(y)=l+(a:
+ 2/)+
X |. 2!
y. + ^y + 2! ^ f'
+ (^ + ^^ + Tl^ + 2! l"^12
V3!
+
f-')
= -11/ + y) + 1 +(^
I
(^ y)^ ^^-2j
+
,
(-^
+ y)' 31^ +
From
this result
it
follows that
y)
=/(^ +
f{z)=f(x + y +
z),
e,
and so
on.
and /(I)
vi,
we may prove
988, that
When
ic
is
a positive integer
/(m) = [/(!)]'" =
2.
6-.
When
is
3.
When
a;
is
a negative rational
s,
557
that
is,
rational
!
we have f(x) = e%
+ X + x72 + a'V3! +
is
+ x^/n + .
!
(2)
made
to
approach
denote any given irrational number and x be h as limit through a sequence of rational
we have f{x)
e""
and
therefore ^^^^f(x)
li"i
e^
that
is,
+ ^ + *V2 C-\
(2) is also
^ =
g&^
728.
There-
eK
is,
has a
Hence
x.
!
(2)
may
!
be used to define
e-^
for imagi-
Thus, by
definition,
h i"/n\
=1+i+
iy2
iV3
.
991
real
number.
e^iSe, 730. e^^^e", 732, we have a^ Therefore, substituting x log^a for x in the series,
Since a
990, (2),
we have
a'^
=l+x
log^a
+ x\\og,ay/2 +
!
+ x" (\og,ay/nl + -.
The logarithmic
series.
a""
992
we
replace a by 1 -f a and x by
we have
988,
when
|a;|
<
1,
By carrying out the indicated multiplications and collecting terms we can transform (2) into a power series in y. The coefficient of y in this series will be
'
"
558
Equate
A COLLEGE ALGEBRA
this to the coefficient of y in (1).
We
-.-.
obtain, if
(3)
\x\<\,
'
log,(l
a;)
= x-a;V2 + a;73-a;V4 +
when \x\<
1.
This series
is
In
968 we
proved that
it
converges
In the proof just given we have assumed that the series (2) remains equal to (1 + xyj after it has been transformed into a power series in y.
But this follows from 976 when and y respectively, the series TJi
I \
|a;|<l.
-\-
For
if x'
and
y'
denote
|x|
Uo
of 976, corresponding to
^'^'"
^
..., '
and
this series is
< 1,
its
sum
being
(1
x')-"', 988.
the truth of the binomial theorem only for rational values of the exponent, it may be observed that (3) follows from (1) and
(2)
As we have proved
even when y
of 972).
is
remark at the
end
Example.
"We have
Show
that
^\"^
("
e.
logyi +
^y = nlog,(l+-)=n(l- +
1
...)
("l
= l--^ + + -V =
e^
....
V=
1,
and therefore
e.
limeio?e
ei''"(ioge),
726-729, 731.
993
Computation
of natural logarithms.
e
The logarithms
:
of
num-
are called their natural logai-ltJims. of natural logarithms may be computed as follows
A table
We
have
and therefore
,
.
(1)
(2)
Subtract (2)
from
(1).
Since
a;
log, (1
+
-\-
a;)
j-^^'
\
(3)
we
obtain
log,
X?
-^
+ ^+-j.
x^
559
:;
=
1
and therefore x
=
1
7;
We
obtain
n+l_^f
or log^(7i
>
^^^
+
^
1)
= ^^^ +
+ + 3(27 + 1)^ 5(27TT7"^-/ = 1 in (4), Setting w - .6931 log,2 = 2(1/3 + 1/3* + 1/5 35 + = 2 in (4), Setting = 1-0986 = log, 2 + 2 (1/5 + 1/3 5" + log, 3
(,277+1
71
and so on
rithnis of
to
n.
Modulus.
By
nnmbers
Hence the loga755, logri log,?i/log,a. to any base a may be obtained by multi-
994
plying their natural logarithms by l/log^a. We call l/\og^a or its equivalent, 756, loge, the mochdus of the system of
logarithms to the base
a.
EXERCISE XC
1. 2.
lege 5
!
figure.
Show
that e-i
2/3
6/7
3.
Show by
multiplication that
4.
5. 6.
Show
Show
bmomial
7. 8.
l - xV2 + xV4 - xV6 + . = x- xV3 + xV5 - xV7 + Show that the (r + l)th term in the expansion of (1 x)-" by n(n + 1)- --(n + r - 1) theorem is
that (e^
that (&^
+ -
e-
'^)/2
'
e- '^)/2
the
^^
x'".
r!
in the
expansion of
(8
(1
x)^
which involves
x*.
x^.
560
9.
A COLLEGE ALGEBRA
For what values
X
(12
x2)3 in
(1
10. 11.
Expand
Find the
x.
x*.
'
first
in
powers of
12.
three terms of the expansion of (8 + 3 x)^ (9 2 x)~ For what values of x will the expansion converge ?
lim
'^'
e^^-e-^' 3x
U^
e^.
(i
13.
Prove that
^(\+ ^X' =
1
^ x=0 (e'^
^^''-=0(i
14. 15.
Prove that
x)^
e^.
Expand logc(l 4- x + x^) in powers of x to the term For what values of x will this expansion converge ?
16.
involving
x*.
Show
that
ioge=^?^^-v^^^^y+V'^^^y----.
7l2
17.
18.
Show
Show
n
that logg
n2
=
1
1
1
1
\-
[-.
n2
2 n*
3 n6
that
+1
2(n+l)2
3(n
l)3
2n2
Sn^
of the
form
where pi, p^, , Pr are constant for all values of n, is called a recurring series of the xth order, and the identity is called its
scale of relation.
Thus, in
3x
5 x^
7 x3
(2
71
1)
x"
(1)
- 2a_i +
3
5-2-3 + 1 =
0,
7-2-5 +
ix
a<ixP-
562
A COLLEGE ALGEBRA
the example
it
Fiom
when 2
way
as a
recurring series of the rth order, and that when 2 r are given it may be continued in infinitely many
1 terms
ways
as a
l)th order.
998
The generating function of a recurring power series. Every recurring power series of the rth order is the expansion of a proper fraction whose denominator is of the rth degree (comThis fraction is called the generating function pare 996).
of the
series.
It is the
sum +
of the series
when
the series
is
convergent.
Thus,
let
ao
a^x
^2x2
ax"
(1)
Set
+ pan-i + +
\-
qar,
_2
= 0.
+ pa-iX'
(2)
a_ix-i
2)ffl_2X"-i
..{l
+ px + qx^)Sn = ao+{ai +
pao)x
-\-{pan-i
qa-2)x'>-
+qa-ix"+^
the remaining terms on the right disappearing because of (2). When (1) is convergent, as n increases Sn will approach S, the of (1), as limit, and x" will approach 0.
sum
Therefore
that
is,
(I
+ px +
qx'^)
S=
ao
+
1
{ai
+ poo) x,
^
'
g^ao + {m+pao)x^
+ px +
gx2
of a recurring
999
power series. This may be generating function is known, by the method illustrated in the following example.
obtained,
when
the
Example. Find the generating function and the general term of the a_i 2 a_2 = and whose first two recurring series whose scale is a terras are 5 + 4 x.
Here
Therefore, by
p=
1,
=
1
2,
ao
5,
ai
4.
998, (3),
S =
^~\ x2 X 2
=
,,
(1
^ x) (1 + ^j/
,, , 2 x)
(1)
RECURRING SERIES
Separating
(1)
563
^~-^
(l
+ x)(l-2x)
1+x
= l-2x
-
2(l
a;)-i
+ 3(l-2x)-i.
+ (_l)na;n + ...]. But if |X|<1, 2(1 + X)-' = 2[1-X + X2 And if |x|<l/2, 3(l-2x)-i = 3[l + 2x + 4x2 + ... + 2"x + ].. Therefore the general term is [(- 1)2 + 3 2"]x''.
EXERCISE XCI
1.
If the first
_ 3x +
2.
+ 2a_i
a_2
+ 3a_3 =
in
0,
and
fifth
terms.
each of the
following
(1) (2)
2
1
(3)
3.
5X
x-'^
..
Find the generating function and the general term of each of the
:
following
(1) (2)
4.
Prove
whose
scale of relation
tto
function
(ai
is
of the third order, + diX + + pa_i + ga_2 + ra_3 = 0, the generating + pao) X + (a2 + pcn + 9o) x^ 1 + j)X + gx2 + rx^
5.
By
and
+ 2x +
a
11x2 _i.24x3
85 x*
+ 238x5+
+
(a
..
+ +
1
(a
ci)
(a
2 d) x2
its
3 d) x^
is
a recur-
generating function.
that 12
22x
+
2
32x2
+
+
4flx^
is
is (1
+
4
x)/(l
5 x^
x)^.
that
3x
4 x2
is
recurring
and
find its
sum when
convergent.
564
A COLLEGE ALGEBRA
XXXVI.
1000
Infinite products.
INFINITE PRODUCTS
This name
is
form
n(i
in
,)
= (1
4-
ai)(i
.),
which the number of the factors is supposed to be infinite. Such a product is said to be convergent or divergent accordas
(1
ing
a-i)
(1
a^
(1
+
is
a)
approach a
1001
Theorem.
finite limit as
indefinitely increased.
a^
If
all
the
numbers
are positive,
the
infiiiite
product
n (1 +
a^) is
First,
S.
Then
a-
<
e"^
when x
positive, 990,
e"^
we have
that
is,
a^) (1
e"^---
e"-,
(>s^
Hence, as n increases, (1 + a^) (1 + Oo) (1 + ) increases It therefore but remains less than the finite number e'^.
approaches a limit,
192, that
is,
11
(1
a,.)
is
convergent.
Second, suppose that la,, is divergent. In this case lim (^i + Oo + + ) = co.
But
(1
i) (1
a)
(!
+
-\-
fl)
>
1 4(ai
Oo
+
co,
Therefore lim (1
Oj) (1
a^)
(1
a)
+ ).
is,
that
n (1
-t- a,.)
is
divergent.
l/nf)
If
is
Thus,
n (1 +
convergent when
p>
1,
divergent
when p < 1.
all
Example L
less
Sa^
is
than
1,
show that
11 (1
a^)
0. \
Since a^
<
and
aj) (1
Hence But
(1
1, (1
we have
lim
(1
Therefore
n (1 -
a,)
+ ay) (1 = lim (1
ar< 1/(1 + a,) numerically. + a^) (1 + ag) (1 + a). + ch) (!+ a) = co. - ay) (1 - 02) - a) = 0. (1
a,)
<
1/(1
INFINITE PRODUCTS
Example
verges
if
565
2.
Show
that
when x =
if
m + 1 > 0,
X
but diverges
m + 1 < 0.
l)
When
l
=
,
....
^ ' (1)
In this series
m-n +
n
have
|
__ J
m+l
'
Un
Hence,
But
For
that
if
u,, + \/Ua > 1, and (1) is divergent, 951. a certain term the series will be of the kind
\
described in
if
m+
= 0.
I, it
the terms of
numerically.
Therefore
converges
if
limu
But
m
Un +
1
7n
(-)-0-^)(>-"^)-(-'^)and
it
approaches
EXERCISE XCII
1
1.
ci,
Show
3 . that 2
*!,
5
'^
17
,5 10 and
4 9
17
26
16
16
25
are convergent.
2.
convergent
x3\ 32/
<)
"(^ +!-:)=(
that lim
<'l<'
-D('-f )('-!;) +
^)---'
3.
Show
or
ni<^
n)
, according as
a>
a<o.
566
A COLLEGE ALGEBRA
XXXVII.
1002
Continued fractions.
CONTINUED FRACTIONS
This name
,
is
form a
written.
+d
or
+
,
+
,
We
These
+
,
,
;
where
a^
is
a positive integer
-.
partial
quotients of the continued fraction. infiAccording as the number of these quotients is finite or
nite, the fraction is called
terminating or no7iterminatlng.
1003
Evidently every terminating simple Terminating fractions. continued fraction has a positive rational value, for it may be
reduced to a simple
Thus,
2
fraction..
3^^
4 - = 30 -;
13
3^'
be converted Conversely, every positive rational number may This will be into a terminating simple continued fraction.
evident from the following example.
Example.
Applying
29)67(2
58
integers to 67
divisor of
two
9)29(3
27
a2
67 ^ 29
a3
9 _ = 2+;^ = 29
+ -.
29/9
(1)
2)9(4
=
...f
a4
3+?=3 + ^.
2"
(2)
1)2(2
=^+
^^^
CONTINUED FRACTIONS
Substituting
67
(2) in (1),
1
'
667
and
(3) in
the result, 3
1
we have
-,as
2
1
=+2
29
=2 +
1
required. ^
4+1
2
Since 29/67
1 -^
67/29,
we
also have
1 1
1
29_
67~2 +
Convergents. are called the
1
2"
1
Tlie
first,
fraction
:r'i"l
Uo
?!
+ 4- ^5 o + Oo "3
c-2
11
1004
second, third,
Oi H 2
3 H
When
aj is 0, the first
convergent
is
written
Theorem 1. Each odd convergent is less and each even con- 1005 vergent is greater than every subsequent convergent.
This follows from the fact that a fraction decreases when
its
denominator increases.
Thus,
1.
ai<at +
! H
2.
11
>ai
1
-\
since
.11
>
a2
,
a^
a2
02
3.
ai
-\
02
<ai-\
1
since ^2 H
.11
+
> a2
as
az-\
02 H
as
as
by 2
and so
on.
Reduction of convergents.
On reducing
the
first,
convergents of aj H
fractions,
we
obtain
1
.,
+ + +
to the
form of simple
^'l^2
1
,
<'\^l2^!\
^3
,
....
/^N
aj
"2*^3
(1)
568
A COLLEGE ALGEBRA
denote tne numerators, and q-^, q^, q^, Let pi, P2, pz, the denominators of the convergents as thus reduced, so that
Ih
<1\
= i? I5
a.^,
?i
Pt
<l-l
= i2 + = 2,
^,
Pz
= a^a^a-z + ^i + = 23 + 1, qs
3,
(2)
(3)
Since
a^,
a.3,
(3) that as
are positive integers, it follows from (2), increases j9 and q continually increase, and
00 if
By examining
i>3
(2)
3i>2
and
found that
+ Pi and
03*72
qi.
(4)
This
1007
is
Theorem 2. The numerator and denominator of any convergent are connected with those of the two preceding convergents by
the formulas
Pn
a^Pa-l
+ Pn-2>
qn
anq _ i
q_2.
For suppose that these formulas have been proved to hold good for the kih. convergent, so that
Vk
(^kPk
-l+lh-2,
Pk
qk
__^
<lk
^klk -l
7k-2,
(1)
^2)
(hPk-\
(^k'lk-i
+ Pk-2 + 'Ik-i
be derived from the
The
{k
l)th convergent
may
Hh
by
1004.
a^., it
^ + ^/ak + ^)Pk-i +Pk-2 + 1/X- +l)'Jk-l + 9k-2 ak+i(fhPk-i + Pk-2) + Pk-i ^ ak + iPk+Pk-i ^ ak + i(ak7k-i + 9k-2) + qk-i O'k+iUk + lk-i
((^k
{('k
by a)'
that
is,
Pk+
-\Vk+Pk- V
<ik
+ i=-- t + 1 Ik
7),,
q^
We have thus proved that if the formulas = a^q_i + q_2 hold good for any particular
+ 1k-i= a^_
+Pn-t>
convergent,
CONTINUED FRACTIONS
569
But we have they hold good for the next convergent also. already shown that they hold good for the third convergent. Hence they hold good for the fourth, hence for the fifth, and
so on to every convergent after the thh-d (compare
791).
Example.
Since 3
92
of 3
2+3+4+5
3,
1/2
7/2,
we have pi =
p2
7,
qi
1,
2.
+ 3 = 24, j)4 = 4 + 1 = 7, 54 = 4
1
24
7
^37
+7= +2=
24
7
,
103, ps
30, 55
>
=5 =5
103
30
+ 24 = 539, + 7 = 157.
103
539
157
30
Theorem
3.
Pnqa-l
Pn-iqn
= (-
!)"
The formula holds good when n 1006, we have poqi 2^\1'i (i2 + 1) ~ f'1'^2 Moreover we can prove that if the formula holds good when n = k, it also holds good when n = k + 1.
For
pi^+iqj,
= 2, For, by = 1 = ( 1)^
-ihik+i
Hence,
tl^en
if
p^ + iqk- Pk^k +
is
!)'
true for
= 2,
it is
true for
to
1 or 4,
and so on
any
Eueri/ convergent
q
-p^^/q^
is
an
it
For
if p,^
and
had a common
1
factor,
the relation ^'7_ jo_j2^ =( 1)" that this divisor of ( 1)", which is impossible.
570
1010
Corollary 2. the fornmlas
A COLLEGE ALGEBRA
For
the differences between convergents
we
have,
1.
Pn qn
Pn^l
qn-1
(-1)
qnqn-1
is
"' 2
E^ qn
j^'
qn-2
(~^)"^^
qnqn-2
rela-
The
tion
first
1
formula
JMn -
- Pn -!'/=({'^nPn 1
fact, 1007,
^In
1008, that
+ Pn - 2)
=
The theorem
1011
of
(i>-iy-2-2^n-2g'n-l)
= (-
~ 1)"
^.
Theorem 4. The nth com-ergent of a nonterminating simple continued fraction approaches a definite limit as n is indefinitely
increased.
form a For, by 1005, the odd convergents Pi/qi, Ps/qs, never-ending increasing sequence, every term of which is less than the finite number p<i,fp_. Hence, 192, a variable which
as limit.
even convergents
Similarly a variable which runs through the sequence of will decrease toward some p^^/q^^ Pi/q^j
fj.
number
as limit.
But/A-=A,since/*-A=
lim
[i^-P?jn^~]
= Vim
(-1)''"'
5'2ni9'2m-l
Therefore a variable which runs through the complete will approach sequence of convergents Pi/qi,P2/q2, Ih/qs, X as limit.
1012
is
By
meant the number l\m (p/q)' It follows from 1003 that n= this number is always irrational. The value of a terminating fraction is that of its last convergent,
1004.
CONTINUED FR ACTIONS
571
that
when
1013
a continued 1014
less
fraction and
l/^nqn + i
than
For
let
fix
the
is
odd.
We
then have
^'
'la
1005, 1013
+2
'Li
Hence
and
-^"
?
<^-^^^
? +
!
-^S
?
.-.
< -^
'/,/In
1010,1
1010,2
rela-
,.,^Jbi^.
Qnln +
i
Evidently l/iq +
tion y
_^2
<
'^-/'il,
+ + 2A/7 + 2^ !/'/(// +
+ oy +
1
'i,
1007,
\)-
may
'L,
Hence
X and p/q
is less
q +
to
i)-
Each convergent is a closer approximation value of the fraction than is any preceding convergent.
Corollary 3.
the 1015
For, by 1014, if A. denote the value of the fraction, the difference between X and 7?,yy is numerically less than l/qqn + n
while the difference between X and 2^n-i/q-\ and l/qq + 1 < greater than (/ + ,/y _ ,y +
^
;
is
((
numerically
+ i/y _ ,7,, +
since
q-i<
('n
+ \qn^
1006.
Tlie convergent Pn/<ln ^^ ^ closer approximation to the value of the fraction than is any other rational fraction
Corollary 4.
1016
q.
ajb is a closer approximation to the value of the fraction than p^/q,, is, it must also, 1015, be a closer
For
if
?-
CONTINUED FRACTIONS
In general,
if
573
+
!
+
i
'
we
have,
1
1007,
re
+
i
(tk
+x
= Pk^+Pk-i
q^x
'
y^._i
and therefore
Since
it
is is
has one and but one positive root, and this root
of the fraction.
Again,
if
-f-
i
i
a,.-\- a,
+1 H
1-
rt,
i.
we
as above,
and then
have,
1007,
2/
13r^+ 7^-1
q^x
a^
+x
q^
On converting
Every positive
irrational
numbers
into
continued
fractions.
1018
irrational
number
is
nonterminating simple continued fraction which may be obtained to any required partial quotient by the following
process. If h denote the
number
a-^ -\-
1/ii,
where
b^ is
J)i
is
some
irra-
Next
l/^2>
than
b^.
Then
^^
where
some
irrational
number
greater than
1 +=
b-^
1.
b^
ai
a^^
11 =
(?o
...
a^-]
,11
Oo
<^2
3 H
It can be proved that when J is a quadratic surd the continued fraction thus obtained is a recurring fraction.
574
Example.
A COLLEGE ALGEBRA
Convert vll into a continued fraction.
The
VTl
3
is 3,
and,
603,
Vll =
The
+ {VII-3) =
+ -I
Vll +
3 3)/2
i
+
3)/2
(1)
VTI +
2
is 3,
{Vll and
3
^5i_3=3 + ^^ii:^ = 2
2
3+^l = +
(Vll
3
3)
_l_.
VTT+3
?
(2)
The
Vll +
6
is G,
and
VTl +
+ (Vn -3) =
=^
Vll +
3
=6 +
(Vli
(3)
3)/2
The last fraction in (3) is the same as the last in (1). Hence the steps from (3) on will be (2), (3) repeated indefinitely that is, tlie partial quotients 3 and 6 will recur. Hence, substituting (2) in (1), and (3) in the
;
result,
...
^3 + +
1019
given irrational number can be expressed in only one icay as a simple continued fraction. This follows from the fact
that two nonterminating simple continued fractions cannot be equal unless their corresponding partial quotients are equal.
For if a + a = c + 7, where a and c denote positive integers and a and 7 denote positive numbers which are less than 1, then a = c, since otherwise it would follow from a c = y a that an integer, not 0, is
numerically
less
than
1.
Hence,
Ci, C2, Cs,
if
aiH
a2
+
as H
Ci
-\
C2
=
1020
11
+
cs H
.-.
ao H '
11
...
+
where
Cg
>
we ^ have
1
a\
Ci,
.-.
11
"2
("2
C2
as
Cg
.'.0-1
C2,
and so on.
we compute the continued fraction to which a given irrational number h is equal as far as the wth partial quotient, we can find its ?ith convergent j'nlln^ ^^^ ^l^is rational fraction
If
Vnl^ln will
1/y,^,
express h approximately with an error less than 1014. Moreover p^/qr, will be a closer approximation
is
to h than
1016.
COXTIXUED FRACTIOXS
Thus, the
3 10 -, 1 3
first
575
3
,
four convergents
ofvll =
111
6
are
1
3 H
63
,
19
190
,
^ and
190
60
60
^^
Given 1021 any equation of the form ax + htj = e, where a, b, c denote If integers of which a and h have no common factor, 672.
Solution of indeterminate equations of the first degree.
we convert a/b
of this fraction will be a/b itself, and if the convergent next This fact to the last be j^/o^, "^e have aq bp =-kl, 1008. makes it possible always to find a pair of integral values of x
b//
c.
The method
is
illustrated
As
in 1003, Ex.,
we
find
=2 + 1111 + + +
93 ^ 4
_
1
of 205 x
+ dSy =
_
8
_.
1.
97
,
44
205 93
Hence
or,
205 44
93
97
multiplying by
7,
205 (- 44
7)
93(97
7)
= - 1, = 7.
Therefore x
=-
The general
Similarly
=-
we may show
that 205 x
93 y
= 308,
679.
EXERCISE
Compute the convergents
1.
XCIII
:
of the following
2.
3+1 4
1. 5
1
1
^
1-
1
+
10
1.
12
Convert each of the following into a continued fraction. Eor each of the last three compute the fourth convergent and estimate the error made in taking this convergent as the value of the fraction.
3.
!2. 12
576
A COLLEGE ALGEBRA
Convert each of the following into a recurring continued fraction and fifth convergents and the corresponding errors for the first four of them.
compute the
11.
Vl7.
Vl05.
3 Vs.
12. 16.
V2t5.
13.
17.
V6.
Vl9.
21.
14. 18.
V.38.
15. 19.
I/V23.
20.
V7I.
l)/(
(Vl0-2)/2.
(V2 +
V2 -
1).
fractions.
22.
1
i
-..
i
+ 2+3 + +
1
3
23.
i
2
24.
---
3+{ + 5 + 2+.. 4
^
25.
1 4
26.
1
2
...
1
+2 + 1 +
27.
Show Show
Show
that
Va^ +
Vu^
-a +
=
a
2a + 2a +
a
28.
that that
+-
+ 2a +
c)
+ ia +
29.
ill
a
_ ~
+
{abc
+ a-h +
y/(abc
c)^
2 (a6
30.
+ 1) +x-1=
...
into
a continued
fraction. 31.
Show
Show
that
^"=^ + -i
9
32.
that
11
+
made
qi
3i?2
=
,
111
q-2qz
L+
llll)l'.
qn-iqn
.
a2
33.
as H
qiq2
q-iqz
qsqt
less
What
denominator
ratio of the diagonal of a square to its side ? in taking this fraction as the value of the ratio.
will express
tt
Find the simplest fraction which with an error which is less than .000001.
35.
3.14159265-
error
36.
37.
38.
2.71828
214 y
=
=
-
6.
412 y
x
10.
solutiiTn of 517
323 y
81.
577
XXXVIII.
we
call
is
write y =f(x), we shall mean that it is a one-valued function; in other words, that to each value of x there corresponds but
one value of
y.
And f{a)
a function of x
x"^
y which
is
if
when y =
\.
But a
relation
of X may be one which cannot be expressed by an equation. Thus, y is a function of x if ?/ is 1 for all rational values of x and 1 for all other
values of
equation.
x.
But
this relation
call y a function of x even when there are exceptional values of x which the given relation between y and x fails to determine y, 1024. Sometimes y is defined as a function of x only for a certain class of values of x or only for values of x which lie between certain limits. Tims, the equation ?/ = x + 2x2 + 3x^ + , by itself considered, determines y for those values only of x which are numerically less than 1.
foi
We
Let f(x) denote a given function 1023 Continuity of a function. of x. a, say that /(a;) is continuotis at a, that is, when x
We
if
what follows the notation ^^^f(x) = f(a) means that f(x) will approach f(a) as limit whenever x approaches a as limit, that is, no matter what the sequence of values may
Here and
in
^^^ /(a) has a definite finite value, and if f{x)=f(a). In the contrary case we say that f(x) is discontinuous at
a.
be through which, x runs in approaching a as limit. In the case of a function y defined by a given equation 1024 y=f{x) it may happen that the expression /(x) assumes an
578
A COLLEGE ALGEBRA
when x
indeterminate form
a,
513-518.
The equation
i/
y =f(^x) by itself considered then fails to define when x = a. But if l^^/(ic) has a definite finite value, we assign this as the
value of f((i),
519,
at a.
If
is
li;/(a;)=
00,
we
515; f(x)
then discontinuous at
nate,
we
Finally, if ^^^ f(^) is indetermihave no reason for assigning any single value to
/(a).
^i^ Evidently we can assign none for which f{^) =/(^)In this case also f(x) is discontinuous at a.
1. Thus, every rational function/(a;) is continuous except perhaps the denominator of some fraction occurring in /(a;) vanishes.
r
; '
when
(x
or
1.
X
l)/(x2
and
lifi
/(x) -/(a),
509.
When
form
x
0/0.
1,
the expression (x
But] /(x)
l)/(x2 1) assumes the indeterminate - l)/(x2 - 1)] = J [l/(x + 1)] = 1/2, [(x
we make
for
and by assigning
/(x) continuous
when
= l. When
2.
1,
/(x)
is
discontinuous
:
/(x)
oo.
/W - -\ +
2^
3 _
1
3/2"^
i
1/2
_]
i -
3
1
2^+1
1/2^
1/2
Here/(0) has the indeterminate form co/co, 517. But if we write /(x) in the second form and then make x approach
1
through posifjue values, we have lim 2-^ = co, and therefore lim/(x) = 1. If we write /(x) in the third form and then make x approach through
negative values,
Finally,
positive
if
we have lim 2 ^ = oo, and therefore lim/(x) = 3. we make x approach through values which are alternately
for
and negative, /(x) will not approach any limit. Hence /(x) is discontinuous at 0. No value can be assigned to/(0) which
''
]/(x)=/{0).
1025
Prom
follows,
1023
it
immediately
579
and necessary condition that f (x) he continuous that f (a) have a definite finite value, and that for every positive number S luhich can he assigned it shall be possible to
at a
is
find a corresponding
|f
j^ositive
number
such that
|x
(x)
(a)|
<8
whenever
a|
<
e.
Thus in the neighborhood of a value of a-, as a, at which f{x) is continuous, very small changes in the value of x are accompanied by very small changes in the value of fix), and the change in the value of x can be taken small enough
to
of f{x) in an interval containing a value of x at which fix) is discontinuous. See the examples in 1024.
as
Theorem
tinuous at
also of f
1.
f (x)
</>
a,
same
is
(j>
true of f
= (a)
a,
(x)
(x)
0.
Ifi(x)
continuous at
the
same
is true
of vf
(x).
This follows immediately from the definition of continuity and the theorems of 203-205, according to which lim \f(x) + ^(^)] = lim/(a;) + lim ^{x), and so on.
at a, 1023,
Real functions.
that
is,
one which takes real values only, and f{x) will denote a real function of x, that is, one which has real values when x
is real.
Number intervals. The practice of picturing real numbers by 1028 points on a straight line, 134, 209, suggests the following convenient nomenclature.
Let us call the assemblage of all real numbers between a and b, a and h themselves included, the number interval a, b, and represent it by the symbol (a, b'). Moreover, it being understood that a Kb, let us call a and b the left and right extremities of the interval (a, b). Also, if
580
c
A COLLEGE ALGEBRA
say that
(a, h) is
= (a + i)/2, let us
(a, c)
divided at
on.
into the
two
equal intervals
Thus,
(1, 7) is
and
(c,
b)]
and so
and
at 3
and 5
1029
We
say that the function f{x) is continuous throughout the if it is continuous for every value of x in this
1030
Theorem
a7id f (a)
Xfj
2.
Ifi (x)
is
such
til
0.
To
fix
Divide
the ideas, suppose that /(a) is + and that/(&) is into any number of equal intervals, say into the (a, b)
.
two equal intervals (a, c) and {c, b). Iif{c) =0, our theorem is proved, c being Xq. But iif(c) ^ 0, it must be true of one of the intervals (a, c) or (c, b) that /(a?) is + at its left extremity and at its right. Thus, if f(c) is this is true of (a, c), and if /(c) is +, it is true of (c, b). Then Select this interval and for convenience call it (a^, bi). is + and /(b^) is -. /(ai) Deal with this interval (a^, b^) as we have just dealt with and so on indefinitely. We shall either ultimately come (a, b),
,
for
which f(x)
0,
which
is
then
we
,
are
such that
,
. .
f(a),
.
and
It follows
.,f(b),
are -.
from 192, 193 that as 7i increases a and b approach the same number as limit. For remains less than b and never decreases, and ^ remains greater than a and never a) = lim (b a)/2" = 0. increases, and lim (b,^
Call this limit
Xq.
Then /(xq) =
0.
581
continuous at
x^,
lim/(o)
= lim/(^) =f(x^.
since /() is always positive, its limit /"(xo) cannot be negative, and since f{h,^ is always negative, its limit f(x^
But
cannot be positive.
Thus,
if
Therefore
a;V2
!
/(j;'o)
!
is 0.
fix)
=1-
a;V4
x^/Q
shown
that /(I) is positive and /(2) negative. vanish for some value of x between 1 and 2.
Maximum
2 11
and
minimum
11
is
...
values.
1031
/A^
2 21
23
27
... (V,\
In (A) there
number
On
among
the other hand, while there is no least number in (A), the numbers which are less than those in (A) there is a
namely 1. Similarly among the numbers which are greater than those in (B) there is a least, namely 3. The like is true of all infinite assemblages of finite numbers,
greatest,
that
is,
h.
of
numbers which
In other words,
lie
finite
numbers
a and
Theorem 3. Let ai, a^, , a, (A) denote any infinite 1032 assemblage of finite numbers. Then 1. Either among the different numbers in (A) there is a
greatest or
is
among
the
a
2.
least.
or
among
Either among the different numbers in (A) there is a least the numbers less than those iti (A) there is a greatest.
1 assign all
all
To prove
a class
/?2,
and
i?i.
to a class
numbers greater than those in (A) to other real numbers, including those in (A), Since each num})er in E^ will then be less than
every number in R^, there will be, 159, either a greatest numwhich means either a greatest ber in Ri or a least in R2,
/"
582
A COLLEGE ALGEBRA
the different numbers in (A) or a least
among
among
the num-
bers which are greater than those in (A). By similar reasoning 2 may be proved.
1033
among the different numbers of an assemblage there is we call that number the maximum number of the assemblage if a least, its minimum number. The superior limit of an assemblage is the maximum number, if there be one. If not, it is the least number which is greater than every number in the assemblage. The inferior limit of an assemblage is the minimum number, If not, it is the greatest number which is lesa if there be one. than every number in the assemblage. An assemblage like 1, 2, 3, 4, which contains numbers greater than every assignable number is said to have the supeIf
a greatest,
4,
is
1, 2, 3,
cc.
Evidently,
either \
is its
an assemblage has a
or
differ
maximum number
we can
blage
numbers which
from X as
little as
we
please.
1034
" By the " values of /(cc) in (a, h) we shall mean those which correspond to values of x in (a, h). And if this assem])lage has a maximum or a minimum value, we shall call it the abso-
Tnaximum or yninimum. value of f(x) in (a, b). The maxiand minimum values defined in 639 may or may not be the absolute maximum and minimum values.
lute
mum
1035
Theorem
(a, b),
it
4.
If
f (x)
is
continuous throughout
the interval
has an absolute
(a, b).
maximum and an
absolute
minimum
value in
For since the values oif(x) in (a, b) are finite, 1023, they have finite superior and inferior limits. Call these limits X and {x respectively. We are to demonstrate that in (a, b) there is a number K ^^^^^ ^ number Xi such that f(x^ = /*. Xq such that f(^o)
583
As the proofs of these two theorems are essentially the same, we shall give only the first of them. Divide (, b) into any number of equal intervals, say into
two such
intervals. Evidently A will be the superior limit of the values of f(x) in at least one of these half intervals. For convenience call this half interval (ui, bi).
(a, b),
b^)
as
we have
We
intervals,
(a, b),
in
each of which A
is
the
superior
limit
of
the
values
of f(x).
As 71 is indefinitely increased, a and 5 approach the same number as limit (see 1030). If we call this limit .r^, then /(.r,,) = A.
For
if not,
difference
so that
since both /(x^) and A denote constants, their must be some constant, as a, different from 0,
X-f{Xo)=a.
(1)
(a, b)
Since f(x) is continuous at x^, we can make the interval so small that for every value of x in (a, Z) we have,
1025,
\f(x)-f(x,)\<a/2.
(2)
And
(a, b),
since A
we
the superior limit of the values of f(x) in can choose x in (a, b^^) and (2) so that, 1033,
is
A -f(x)
< <
a/2.
(3)
(3) that (4)
is
But
it
from
(2)
and
a.
A -/(xo)
false
that
is,
\ /(.To)
0, or
= /{xq),
as
was
to be proved.
584
1036
A COLLEGE ALGEBRA
If f (x) is continuous throughout the interval (a, b), Corollary. will have in (a, b) every value intermediate to its maximum and minimum values in this interval.
it
For
let c
function /(.t)
If/(.r|,)
denote the value in question and consider the c, which is continuous in {a, h), 1026.
maximum and
mini-
c is + and c is /(Xq) f{xi) Hence, 1030, between Xq and x^ there is a number, call it x^, such that/(j-2) c 0, ov f(xn) = c, as was to be proved. 1037 Oscillation of a function. By the oscillation of f(x) in (, h)
mum
values of f{x) in
is
inferior limits
1038
Theorem
5.
(a,
b).
If
any positive number a be assigned, however small, it is possible to divide a finite member of equal inter^'als in each of (a, b) into which the oscillation of iix) is less than a.
For divide (o, A) into any number of equal intervals, say into two such intervals, each of these in turn into two equal
The process must ultimately yield interintervals, and so on. vals in each of which the oscillation oifix) is less than a.
if not, there must be in {a, b) at least one half interval which the oscillation of /(.t) is not less than a in this, in turn, a half interval in which the oscillation of f(x) is not less than a and so on without end. Let this never-ending sequence of intervals within inter-
For
in
vals be
(a,b), (i,
Z>i),
(ao,b^), , (a,b),
,
it
lim b,, and, as in 1030, let lim a Since f(x) is continuous throughout
lute
.r,,.
(a, b),
has an abso-
maximum and
an absolute
intervals (a,
minimum
(a, b), ,
maximum and
(a, b).
the
minimum
value of /(.r) in
585
But this is impossible. For since a and /3 are and lim a = lim b = x^, we have lim a- = lim /S =
Therefore, since /(.r) is continuous at x^, lim/(a-) that is, lim/(n-) 0, .". not > a. lim/(/3)
in (a, b),
a-^.
= lim/(/3);
We
h) will
mean
when
a function of x
when X = 1, ?/ = 2, we have u /(I, 2) = 1 4 + 1= 2. The note at the end of 1022 applies, mutatis mutandis, here
and y
if
n =/(x,
t/)
x^
_ 2y +
1.
Here,
also.
Let f(x, y) denote a given 1040 Continuity of such a function. function of x and y. say that /(a;, y) is contimwus at a, b,
We
that
is,
when x
\i
value and
= b, ii f(a, b) has a definite finite f(x, y) will always approach /(, b) as limit when
a and y
X and y are made to approach a and b respectively as limits. In the contrary case we say that f(x, y) is discontinuous at
a, b,
that
is,
when x
a and y
b.
From
The
ous at
this definition
and
189,
it
sufficient
a,
and necessary
(a,
is
that f
b) have a definite finite value, and that for S tvhich can be assigned it shall be possible
t
such that
e
|f
y)
f (a, b)|
<
whenever
|x
a] <
and
[y
b|
<
e.
586
1042
A COLLEGE ALGEBRA
1.
Theorem
f (x,
(x,
y) and
<}>
(x,
y) are
continuous at
y)
</)(x,
true of f (x, y) <^(x, y) tmless <^(a, b) y), also ofi(x, y)/</>(x, y),
is
same
and
0.
If
f (x, y) is
continuous at
a, b,
the
same
is
true of
Vf (x,
y).
1043
to be understood that
X and
ij
1027). 382, pairs of values of x and y may be Evidently, if employing this pictured by points in a plane. method we draw the lines wh. :h are the graphs of the equa-
As
is
shown
in
tions X := a, X
^ h,
c,
d,
by these
X,
lines will contain the graphs cf all pairs of values oi such that a<x<h, c<y<d. With this rectangle in mind,
we
c,
dy
We
b
;
(a,
c,
d) if
it is
in this region.
1045
Theorem
c, d), it
2.
has a
maximum and
IfHy^, y) ^' continuous th roughout the regio7i(a,,hf a minimum value in this region.
Since f{x, y) is continuous throughout the given region, its values within this region have finite superior and inferior limits, 1032, 1040. Call these
limits X
and
/u..
77
77,
Cn
c,
We
^^^^ ^i"'^" '^ that/(a:'o, ^o) lar reasoning will show that there
such
is
x-^,
y^
such
that/(.ri, yi)
/A.
EFGH
For construct the rectangle which pictures the number region (a, b c, d), 1043.
;
587
EFGH"
all
we
shall
mean
the values
pairs of values of x, y in
(a,
c,
d).
EFGH into four equal rectangles as in the figure. Evidently A will be the superior limit of the values oi f{x, y) Call this quarter in at least one of these quarter rectangles.
Divide
rectangle E^F^GiH^. Deal with the rectangle EiFiGiHi as
we have
just dealt
with
EFGH, and
so
on
indefinitely.
We
EFGH, E,F,GJI
in each of
E^F^G^H^, .,
(1)
which A
is
f(^, y)-
Let
and
c its
ordinate.
As
and
is c
proved in
If lim a
1030,
when n
limits.
c^
is
indefinitely increased a
approach definite
= Xq
and lim
A
= X.
(2)
For
if not, let
f{x^,
a.
Since f(x, y) is continuous at .Tq, ?/, we can so choose E^F^G^H^ that for every pair of values of x, y in this rectangle
we
have,
1041,
(3)
\f{x,y)-f{xy,)\<a/2.
And
since A
E^F^G^H^, we
the superior limit of the values oif(x, y) in can so choose x, y in E^F^G^H^ and in (3) that
is
- f(^;
!/)
<
a/2.
(4)
From
(3)
and (4)
it
- fi^o,
y^
<
(5)
Hence (2) is false (5) contradicts (2). A, as was to be demonstrated. f(xo, yo)
But
and therefore
588
A COLLEGE ALGEBRA
THE FUNDAMENTAL THEOREM OF ALGEBRA
We
1046
are
now
We
^ (z) =
1-
kz", ivhere
a complex
1.
variable;
it
is
always
|<^(z)|<
For
and
b in
terms of absolute
877, be
(S).
Then
= p (cos 9 + i sin 6), b = \b\- (cos (3 + i sin bz"' = p"'\b\- [cos (mO + /3)+i sin (md + 13)^ 879,881
77id
(3
tt.
(1)
Then
since
bz""
= p^l^l- (cos tt +
cos
tt
sin ir)
and
= p'"\b\, sin tt = 0.
877,
878
Next
choose p so that,
+i
1,|p
854,
(2)
...+17,1p<|^,|p'<1.
which
|
of 6
<f>
(z^)
<
1.
For since
cz^
+'
+ kz^,
we
have,
235,
...
\<f>(z,)\<l
+\k\p", ..<!, by
(2).
1047
if
Corollary.
f (z)
f (z)
az"
aiz"
'
h a
when
= b,
v)e
z so
that\i{z)\<\i(h)\.
For in f(z) set z = b -\- h and develop by Taylor's theorem, It may happen that certain of the derivatives f'{z), vanish when z = b; but they cannot all vanish f"{z), and so on
848.
589
= n\aQ.
f{b
first
Then
and therefore
h)
=f(h) +f"(h)
l
+f'%b) ^,
fib)
f{h)
n\
Given
f (z)
of 1048
For in
=x
-{-
iy,
real,
and
to the form f(z) = ^ (x, y) + /i/^ (.r, y), where <^ (x-, y) and (x, y) denote real polynomials in x, y, and therefore have, 232,
ij/
a^^x + iy)"~^, by aid of the all the real terms in the result, and the imaginary terms. We may thus reduce /(5;)
iy)'\
By
roots
less
855,
we can
0, if
find a positive
number, as
c,
oif(z)=
c
;
them numerically
than
and
if c'
than
c for all
values of
x,
c', c') the expression region ( c', r' [_(f>(x, yy i}/(x,yyy is a continuous function of x and y, It therefore has a minimum value in this region, 1042.
But
in this
number
1045, say
If z,
make
Therefore |/(o)I = 0, since otherwise, |/(s;)|<|/(o)|. 1047, we could so choose z that |/(2)| < |/(.~o)f. Hence |/()|, and therefore /(), vanishes when z = Zq]
is,
= Xq, y = y^. then |/(^)| - \_4>i.^o. l/oY + ^(^o, 3/o)']^ = 0. For since |/(^o)| is the minimum value of |/()], we cannot
when x
x,
iy,,
that
Zq is
= 0.
INDEX
Numbers
refer to pages
Abscissa, 138 Addition of integral expressions, 93 of numbers, 10, 19, 35, 50,71, 72 of radicals, 274
of rational expressions, 217 of series, 541
11,
74
187,
300
Annuities, 391
Approximations, 48,
55,
453
Assemblage,
infinite, 3
Asymptote, 335
Converse, 92
Base of power, 39 of system of logarithms, 377 Binomial theorem, 256, 283, 554
Binomials, products of, 102, 253 Biquadratics, 112, 480
Coordinates, 138
Correspondence, one-to-one,
Cosine, 489 Counting, 9
Cube
root.
See Roots
Cyclo-symmetry, 248
592
Determinant, 494 bordering a, 605 cofactors of, 504 diagonals of, 490 elements of, 494 evaluation of, 505 minors of, 502
order
of,
A COLLEGE ALGEBRA
Equations, fractional, 111, 231, 300
identical, 89
495
of,
products
506
terms
of,
496
of,
111
Differences,
method
364
logarithmic, 390
Discriminant, 517
of cubic, 485
of quadratic, 304
Divergence of
infinite series,
520
simple,
ll:i,
118
by factorization,
of, 114, 129,
by aid
of
undetermined
transformation
436
Errors of approximation, 55
Elimination, 131, 143, 317
Evolution, 39, 56, 70, 83, 200, 276.
by determinants,
Ellipse, 334
508, 514
490
Expectation, value
of,
411
Equality,
3, 8, 32,
34, 45, 72
algebraic and numerical, 18, '^5 rules of, 13, 15, 24, 36, 54, 57
integral
exponential, 390
common, 196
INDEX
Factor, irreducible, 211
Infinitesimal, 63
Infinity as limit, 224, 229
Interest,
593
compound, 390
Interpolation, 371
Inversions, 492
489 Lagrange's formula of interpolation, 373 Length, 26, 37, 66 Limit of variable, 58
Limits,
continued, 566
improper, 213
irreducible, 37
partial,
236
proper, 213
reciprocal, 219
of
Logarithms, 39,^377
characteristic of, 381
expansion
551
common, 379
mantissa
of, of,
381
integral, 85
modulus
559
rational, 86
symmetric, 245
natural, 558
of algebra,
Fundamental theorem
427, 588
Maximum,
Groups of things,
equivalence
finite
Minimum,
3
of, 1
and
infinite,
of series, 545
Number,
cardinal, 2, 10
complex, 71 fractional, 33
imaginary, 70 integral, 18
irrational,
See Equality
46
594
Number,
natural, 6 negative, 18 positive, 18
rational, 34
real,
A COLLEGE ALGEBRA
Quadratics, 112, 298, 304 simultaneous, 317
45
Number
Numbers, theory
Odds, 410
Ordinal,
7, 33,
211
simple, 271
Radicand, 271
45
Ratio, 09, 347 Rationalization, 285
Parabola, 333
Parentheses, rule
of,
95
Permanences of
sign,
446
Permutations, 393
irrational,
453
87
number
of,
427
rational, 429
of,
Power
series, 535
of,
symmetric functions
535
of,
305,
convergence
products
quotients reversion
of,
of, of,
545
546
548
of,
square, 262
transformation
Probability, 409
545
39, 56, 76
trigonometric
expression
of,
490
Scale, complete, 6
364
geometrical, 357
natural, 17
Sequence of numbers, 58
regular, 60
INDEX
Series, alternating, 632
5%
of
Subtraction
rational
expres-
sions, 217
doubly
infinite,
543
of series, 541
geometric, 360
248
Term, 86
absolute, 426
Simultaneous,
tions Sine, 488
127.
See
Equa-
128
infinite, 230,
318
integral, 342
number
of,
517
numbers,
of radicals, 274
RETURN
TO^^
GENERAL LIBRARY
U.C.
BERKELEY
BDODflblSfll
V
'~f
V
y^i #r-^
^3
4
: