Peackock - A Treatise On Algebra Vol 2 (1845)
Peackock - A Treatise On Algebra Vol 2 (1845)
Peackock - A Treatise On Algebra Vol 2 (1845)
-
- A-
STAY*
LIBRARY
A TREATISE
ON
ALGEBRA
VOL. II,
ON SYMBOLICAL ALGEBRA,
AND ITS
APPLICATIONS
TO THE
GEOMETRY OF POSITION.
BY
CAMBRIDGE :
writers on Algebra.
812262
IV PREFACE.
and error, yet I have adopted the only course which was
signs which
are symbolized by the roots of 1 : we are thus
could be very
Algebra, without whose aid the two sciences
imperfectly separated from each other.
have not entered further into the general theory of
I
CHAPTER XI.
PAGK
ON the operations of addition and subtraction in Symbolical Al
gebra 1
CHAPTER XII.
CHAPTER XIII.
CHAPTER XIV.
On the reduction of algebraical expressions to their most simple
equivalent forms 52
CHAPTER XV.
Formal statement of the principle of the permanence of equiva
lent forms 59
CHAPTER XVI.
The theory of indices. ...*. 61
CHAPTER XVII.
On the extraction of square roots in Symbolical
Algebra: origin
of ambiguous roots and of the 67
sign */-!
CHAPTER XVIII.
CHAPTER XIX.
On the general theory and solution of quadratic equations... 77
V1U CONTENTS.
CHAPTER XX.
PAGE
On the solution of equations of higher orders than the second,
which are resolvible into simple or quadratic factors ............ 96
CHAPTER XXI.
The binomial theorem and its applications .............................. 110
CHAPTER XXIII.
CHAPTER XXIV.
On the symbolical properties of different orders of the root of 1. 123
CHAPTER XXV.
On the general principles of the interpretation of the signs of
which are symbolized by the roots of 1 .................. 135
affection
CHAPTER XXVI.
On the representation and measures of angles ........................... 144
CHAPTER XXVII.
On the theory of the sines and cosines of angles ..................... 166
CHAPTER XXVIII.
On the tangents, cotangents, secants, cosecants and versed sines
of angles ........ ............................................................ 181
CHAPTER XXIX.
On the construction of a canon of sines and cosines, tangents
and cotangents, secants and cosecants ................................. 188
CHAPTER XXX.
On Demoivre s formula and the expression, by means of it, of
the roots of 1 and of a b J~^l ....................................... 194
CHAPTER XXXI.
On the representation of straight lines both in position and
magnitude, and the application of Algebra to the theory of
rectilinear figures ............................................................ 207
CONTENTS. IX
CHAPTER XXXII.
PAGE
On logarithms, and logarithmic tables and their use 234
CHAPTER XXXIII.
CHAPTER XXXIV.
On the use and application of subsidiary angles 260
CHAPTER XXXV.
On exponential and logarithmic series 268
CHAPTER XXXVI.
On the limits of the values of series proceeding according to
CHAPTER XXXVII.
On the series and exponential expressions for the sine and co
sine of an angle 284
CHAPTER XXXVIII.
CHAPTER XXXIX.
On the decomposition of rational fractions with compound deno
minators into partial fractions 308
CHAPTER XL.
CHAPTER XLI.
CHAPTER XLII.
CHAPTER XLIII.
PACK
CHAPTER XLIV.
On the solution of simultaneous equations, and the theory of
**66
elimination
CHAPTER XLV.
On the algebraical solution of binomial equations
391
CHAPTER XLVI.
Sources of ambiguity 42^
ERRATA.
It is feared that very few of the errors have been noticed, no sufficient time
544. But the symbols, which are thus employed, do not The as-
convey, either to the eye or to the mind, in the same manner as
digital numbers and geometrical lines, the limitations of value to limited
which they are subject in Arithmetical Algebra : for they are the symbols
equally competent to represent quantities of all kinds, and of all f^i Jesfhe
relations of magnitude. But if we venture to ascribe to them necessary
a perfect generality of value, (upon which a conventional limi- oft!Je?nde-
tation was imposed in Arithmetical Algebra), it will be found P e
to involve, as an immediate and
n u
s ^ ^ j
symbols.
Thus, in the expression a b, if we are authorized to assume
a to bi either greater or less than b, we may replace a by the
tivequan-
connection with other symbols, are called positive and negative
(Art. 32) symbols, or positive and negative quantities
titles.
: such sym
bols are also said to be affected with the signs -f and . Positive
symbols and the numbers which they represent, form the subjects
of the operations both of Arithmetical and Symbolical Algebra :
but negative symbols, whatever be the nature of the quantities
which the unaffected symbols represent, belong exclusively to
the province of Symbolical Algebra.
Assump- 546. The following are the assumptions, upon which the
de
!nrmbo -
rules of operation in Symbolical Addition and Subtraction are
lical addi- founded.
subtrac- ^ s ^* Symbols, which are general in form, are equally gene-
tion. raj i n representation and value.
2nd. The rules of the operations of addition and subtraction
in Arithmetical Algebra, when applied to symbols which are
respectively, when a becomes equal to zero : this is another mode of deriving the
conclusion in the text.
547- Proceeding upon the assumptions made in the last Rule for
*1
Article, the rule for Symbolical Addition may be stated as fol-
SdUion
lows :
"Write all the addends or summands (Art. 24, Ex. 1.) in the
same line, preceded by their proper signs, collecting like terms
(Art. 28) into one (Art. 29): and arrange the terms of the result
or sum in any order, whether alphabetical or not, which may be
considered most symmetrical or most convenient."
be understood that negative (Art. 545) as well as positive Negative
It will
3a + 5a = 8a (Art. 29).
Add -
(2) together 3 a and 5 a.
Add together
- 3 a and - 5 a.
(4)
-3a-5a=-8a (Art. 31).
-4*2
2
Qa: 4>abc
4
(10) -3a-4>b + 5c
-4,a + b + 8c-lld-3e
10a*
- 6,r
2 -
in^ymbo -
by virtue of the assumptions in Art. 546, from the
derived,
lical Alge-
corresponding rule in Arithmetical Algebra : it may be stated
as follows.
" To the minuend or minuends, add
(Art. 54?) the several
terms of the subtrahend or subtrahends with their signs changed
from -f to - and from to + ."
Examples.
The following are examples of Symbolical subtraction.
40 -4 Wa - 10a
(6)
a + b (7) a-b (8) a +b
a-b a+b a+b
2b -26 2a
(6) a + b - (a - b), or a + b - a + b = 2 b.
3a
6
-4>y
= 9x- ISy.
The between brackets and pre
several subtrahends included
ceded by the sign , are written in the same line with the
minuend ; the brackets are subsequently removed and the signs
of the several terms which
they include changed, in conformity
with the Rule.
x s -xy+y a .
cessively removed, and the like terms, which they involve, are
obliterated or reduced into one. See the Examples in Art. 33.
^ *n tne ex P s iti n an d exemplification of the
preceding
used in rules, we have felt it to be unnecessary to repeat definitions and
Arithmeti- ^. , . , , , , ,. . . ,
cal and
Symbolical
Algebra such are the ordinary uses and meanings of the signs
:
In Arith- + and , of coefficients (Art. 25), of like and unlike terms (Art. 28),
of indices and powers (Arts. 38 and 39), and the methods of
the defini-
denoting the ordinary operations (Art. Q).
tions deter
mine the 551. The use however, of the same terms in these two sciences
operation :
will by no means imply that they possess the same
meaning in all
"
j? ^y^ their applications. In Arithmetic and Arithmetical Algebra, ad-
licalAlge
bra the dition and subtraction are defined or understood in their ordinary
sense, and the rules of operation are deduced from the defini-
opeTation
determine tions: in Symbolical Algebra, we adopt the rules of operation
which are thence derived, extending their application to all the mean-
values of the symbols and adopting also as the subject matter of
^rations
our operations or of our reasonings, whatever quantities or forms of themselves,
symbolical expression may result from this extension but, inas- results. :
552. The
rule of subtraction, derived rigorously from the Considera-
definition or assumed meaning of that operation in Arithmetical different
^
Algebra, directs us to change the signs of the terms of the caf es of .
subtraction .
subtrahend and to write them, when so changed, in the same in the tran-
line with the terms of the minuend, incorporating, by a proper
AhthmetiU
rule, like terms into one. In the application of this rule, three cal to Sym-
.11 , i . , , bolical
cases will present themselves, which it may be proper to con- Algebra.
sider separately.
2nd. Where the subtrahend may or may not be less than when the
the minuend, according to the relation of the values of the subtrahend
... , may be less
symbols involved. than the
If the minuend be 3a + 46 and the subtrahend 2a + 5b, then
^nof
3a + 4>b (%a + 5b) = 3a + 4ib 2a 5b = a b. necessarily
so.
If a be greater than b, this is a result of Arithmetical Al- Examples.
gebra, but it ceases to be so when a is less than b: as long
therefore as the relation of the values of a and b remains un
determined, it is uncertain whether it is a result of Arithmeti
cal or of Symbolical
Algebra it is one of an infinite number
:
the terms occupies the first place (Art. 22) and the several operations in-
succeed
dicated are Lpossible*. Thus a +b is equivalent
*
to b + a: a-b
each other.
+ c is equivalent to a + c- b or c+ a b or c - b + a*. The
same rule is transferred to Symbolical Algebra, without any
values of a and b: 3a + 5a is
equivalent to 5a 3a: a b + c
is
equivalent to a + c b or c + a b or c b +a or b + a+c
or b + c + a, and similarly in all other cases.
the results obtained are or may be made* identical in those mus t be as-
two sciences in which are equally within their pro- Jf Jnter ^re
all cases
555. The
addition of a symbol preceded by a negative Conditions
sign is equivalent to the subtraction of the same symbol pre- tetion terpre
of positive
3 a positive sign and conversely.
ceded by J
and nega-
Thus a+ -b = a-b = a
556. The signs plus and minus, when prefixed to symbols The signs
denoting quantities of the same kind, cannot denote modifica- ^ed inde
pendently,
"
*
Thus if a be greater than b, the symbolical result - b + a is convertible into
{* ^on"-
a b, which is, under this form, a result of arithmetical algebra. vertible af-
t These formulae express the rule for the and unlike signs fections of
concurrence of like
in symbolical addition and subtraction, which! is as follows: "when two unlike ma g nitude
signs come together, they are replaced by the single sign and when two like;
signs come together, they are replaced by the single sign +."
VOL. II. B
10
Possible
557. Quantities and their symbols
J are said to be real or
andimpos-
sible quan- possible, when they can be shewn to correspond to real or
Problem.
Its solution
bytheprm-
ciples
bra.
of
theory.
departure and
"A traveller
northwards for b miles
of departure ?"
Let A be the point of
the dis-
expressed in mag-
nitude by the symbol a, to
which he
AB
travels southwards
_
--
moves southwards for a miles, and then returns
what is his final distance from the point
:
t
S
wards: his final distance AC, when he stops, from the point
of departure, is the excess of AB above AC, and is
correctly
- we suppose c to represent the distance
represented by a b : or if
As long as AB
greater than BC, or a is greater than b, the
is
but inasmuch as, in this case, AC the final distance of the tra
veller to the north of A is the excess of BC above AB, or of
b above a, it willbe correctly represented in Arithmetical Alge
bra by b a : and if b = a + c, we shall have
b a = c.
There are therefore two distinct cases of this problem, when TWO dis
^ 48
le s ot
in each case expresses the absolute magnitude of the final dis- ? -
,
-
tance only, and not its quality or affection, whether south or cal Alge-
bra
north.
In the geometrical solution of the problem, the distances are its geome-
al solu "
expressed and exhibited to the eye, both in quality and mag- |F^
nitude and there is no such interruption of continuity in passing
:
a b =a (a c)
=+c= c, and in the other, when b =a+ c,
a b =a (a + c) = -c.
The signs
559. We are enabled to conclude, from the discussion of
+ and -
preceding problem, that the signs + and , when thus used
j-jjg
symbols
nitude which the symbol c expresses, where the nature of the
denoting relation between them is determined by the conditions which
lines, sym- . .
r , . , .
direction?
been considering, it is
opposition of direction, and opposition of
and oppo- direction alone which these signs can correctly express: for if
S e d
t io ns only.
we suppose AB (a} to be the distance c
to which the traveller proceeds south- \
the other, will not express the final distances of the traveller
either in magnitude or direction, and consequently will admit
of no interpretation which is relative to the problem proposed*.
560. The same problem may be
variously modified in form, The same
without altering the essential conditions upon which its solution P robl em
by the principles of Symbolical Algebra depends: thus instead variously
of estimating distances northwards and southwards of a point
of departure, they may be estimated, with reference to the lead to tne
L
same con-
_
, t. . .
i 7 r
observer s position, as lines to the right or left, up and down, elusion.
to and from, or as reckoned upon any straight line intermediate
in position to these three fundamental directions t : and in all
these cases the quantity and direction of the movement to or
from the same point, or the magnitude and direction of the geome
trical lines which express them, will be correctly symbolized
by the signs + and the steps of the processes by which these
:
JJJf
6
in a straight
suppose the traveller to move from
A to B upon the curved or zigzag
line ACB through a distance -^ ^-^"fr
equal to a and then return upon A
the same path from BtoC through \ /\ /\
C
a distance equal to b, his final \s \A A v
distance from A, reckoned upon the same
path, will be expressed
a - and
by b as the conditions
: of the problem and of Symbo
lical Algebra will admit of every possible relation of value of a
and b, we may equally replace b by a c or by a + c and the :
pulls, or a force which attracts and a force which repels, may a nd affec-
n by
be expressed by lines in opposite directions whose lengths are
[| s
a b =a (a + c)
= c:
a+ ( c)
=a- c and a ( c )
= a+ c,
plication.
they are as follows:
1st. When the multiplicand and multiplier are mononomials.
2nd. When the multiplicand is a polynomial and the mul
tiplier a mononomial.
3rd. When both the multiplicand and multiplier are polyno
mials.
568. In Arithmetical Algebra, the rule for the concurrence The rule for
Rule for
5yo.
_
We now_ proceed to exemplify,
*
in their order, the three
symbolical v/r .. _ ../ . .
multiplica- aitterent
Cases ot symbolical multiplication.
f0
CASE 1. When the multiplicand and multiplier are mono-
Case l
nomials.
RULE. " In
finding the mononomial product we must de
termine its sign; secondly, its coefficient; and
first, lastly, its
literal part."
(2) ax b = ab or x b = + ab.
(4)
-a x b = a b or - a x + b = a b.
These four examples merely express the rule for the con
currence of like and unlike signs. (Art. 569-)
(5) 5
(6)
-
(7)
_ -143 a 6V _
llabc -13a 3
6 c
4 3 -
(Q\ x
12 44 528 48
signs, in
any order which may appear most symmetrical or
most convenient."
" If the
sign of the mononomial multiplier be negative, the
signs of all the terms of the
multiplicand will be changed in :
every other respect the rule agrees with that which is given in
Arithmetical Algebra (Art. 50).
20
a(a-b) = a ~ab.
2
(1)
Rule for
574. CASE 3. When both the multiplicand and the multi-
Case 3.
(1)
x-b
x z - ax
bx + ab
x*a + bx + ab
(2) 3a-5b
-21a 2 +35ab
(3) (x + a) (.r
2
- ax + a 2) . x3 + a 3 .
(a -b +2bc-c )(-a +
2 2 2 2
(4)
-b 4 +2b c -c\ 2 2
(5) (x +ax + b) (x
2
- ax + c)= x*- ( 2 - b - c)
2
x 2 - (ab -ac)x
+ bc.
21
to the powers of some one letter, which may be called the letter
or symbol of reference: thus x is the symbol of reference in Exam
578. CASE 1. When both the dividend and divisor are Rule for
mononomials.
RULE. " The sign of the quotient is positive or negative,
J
according as the signs of the dividend and divisor are the same are mono-
or different."
"The coefficient of the quotient is found by dividing the
coefficient of the dividend by that of the divisor."
" The literal part of the quotient is found by obliterating the
symbols or theirpowers, which are common to the dividend
and divisor : and retaining those which are not thus suppressed/
(Art. 78).
Examples. 579. The following are examples:
(1) ax + x = a: or,
x
=.
/~\ ax
(2) ax^- x = a: or,
ax
(3) ax+ x=a: or, -a.
_3 x 3a 2 x 2 x x _3x
4i x 3a 2 x 2 x a 4>a
s 2
suppressing the factor 3a x , which is common to the divi
dend and divisor.
When the CASE 2. When the dividend is a polynomial and the divisor
di nd
a ;l1; n o-iSamono " omial -
(x
3
i 3
ax*-a x +a x 2 -a x
-3
ax
ax2 + ax
4
a*jc
a 3)
2
= a3
# 4 jr by
x*+ ax
ax,
a^x + a
a z x + ax 2 x3,
RULE."
Arrange the divisor and dividend according to the
powers of some one symbol or letter of reference, (Art. 576), or
as much as possible in the same order of succession, whether
quotient multiply this term into all the terms of the divisor and
:
bra, (Art. 84), and the examples which are given in illustration
of it
(Art 86), are examples also of Symbolical Division.
Divide - 2# + 3x z
3 2 2 2
(1) y- 6xy + 9/, by x +3y .
24
2 - (a-
by a;
3&),r- 3a6.
a:
4 -
^-(fl-SfeJjr-Sfl^
l-x) a (a +ax +a
a ax
ax
(4) Divide x + a by x + b.
x + b) x + a (I +-(a-b) - (a-b)b ~
+ (a-b)b*
.
&c.
x xIT
1
x3
25
ax
ax f a\ (b a)x (a b}x
-
~rr* j -j-
(a x (& b} x~
6) ~
b~ ~br
first remainder is found : x and -7- are like terms (Art. 28),
jl
I or its ^ y -
for the coefficient of x in the re-
equivalent
II mainder.
(S)
x-b x x*
All the terms of this series, after the first, are negative.
(6)
x-a -
T
x+ b
1 - (a
v - +
x
b}-
+ 1
(a + b}b
x
--
f4
2
J-
(a +
x
b} b
r-
3
2
+ &c.
VOL. II. D
a-x_a
~ (a + b)x
~ ~^
(a + b) x*
where k = a 2 pa+q.
In all these Examples it will be observed that the series
tients in divisor and the dividend, in the order of their magnitude, pro-
tic *
duces, in the case of interminable quotients, a series of im
perfect quotients, which approximate more and more nearly to
the true quotient the further the operation is continued: thus
if we divide 1 by 3, we get the interminable or recurring
x+a+ +
x x2
whose successive terms, after the third, grow less and less con
tinually, inasmuch as x is greater than a: but if we admit indif
x+ a H 1
5- +
x x2
but whose successive terms go on increasing continually, in
stead of diminishing.
ot its terms.
which convergent or divergent according as x is greater or /ero
is
a x H ----- - -f T- - &c. . . .
a a2 a*
-
greater or less
than
is
a; : it follows,
a2
-
a +x
+ x*
that
, we must
though the fraction
adopt, in Arithmetical
x + a
we
22number
Conse- 588. If J any
multiply of the terms (n) of the
quencesof
- - be
ing in cer- quotient corresponding to , by the divisor x a, it will
tain cases
), according
infinity (co as a is less or greater than x, or in other
x + a H H 5-
2
+ &c.
x x
as equally significant for all relations of the values of the symbols:
but under such circumstances, the processes which we must fol
low for the purpose of effecting this transition, must not be de
pendent upon the arithmetical aggregation of any number of the
terms of the series, but upon methods which are altogether inde
pendent of the relative or absolute value of the symbols involved.
29
590. When
one of the two symbols in the product ab is When one
an abstract number, it
simply means that one factor is to be factors*^
is an abs
multiplied by the other in its ordinary arithmetical sense, the
. /.i T i
, . . . , . ,tract num-
units of the product being identical with those of the concrete her.
factor thus, if a denote a line of 6 inches in length, and b the
:
591. When both the factors are concrete magnitudes, whe- When both
$
ther the same or different, their interpretation, when possible, are eon -
must be made in conformity with the following principles. crete -
1st. The product must be identical with the arithmetical pro- Principles
duct, when the factors are replaced by numbers. taticST^"
For if not, the operations in Arithmetical Algebra would not
be identical in meaning with the corresponding operations in
arithmetic.
The product is always the same in meaning and magni
2nd.
tude, when the factors incorporated are so. For otherwise, the
factors would not determine the
product.
Upon the same principle, if the concrete factor be affected with the sign -
h
the product will be affected with the same, and the negative units of one will be
identical in meaning with the negative units of the other thus if a denote a dis
:
tance of 3 miles to the south of a given point, and b the number 5, their pro
duct ab or - 15 will denote a distance of 15 miles to the south.
30
will correctly
represent the area of the rectangle which those
lines contain.
Their pro- For, let us suppose the lines AB (a) and AD (6), which con
duct is the .
xl _ jrtnr\ i ,1
rectangular
tam tne rectangle ABCD, to be
-i
in the
^ n
area which of the numbers 5 and and
proportion
r 4,
they con-
tain. also to be represented by them : it we
divide AB into 5 equal parts in the
points a, b, c, d, and AD into 4 equal
/ A a b v
* A parallelogram, not rectangular, would satisfy the first condition, but not
the second for if the sides
: (5) and AB
AD (4) of the oblique parallelogram -, -, ,
^
/ / / /
guity exists with respect to the rectangle, where the angle included by the
containing sides is determined by the definition of the figure.
31
BA
.
The pro-
595.If we extend our inquiry to the determination of the
ductofthree of the product a b c of three factors a, b, c, which seve
symbols re
meaning
presenting rally represent geometrical straight lines, we
lines, re
shall find that it
presents the may correctly represent the
solid paral- volume or solid content of the rectangular G
lelopipedon
which they parallelopipedon, of which a, b and c are
contain.
three adjacent edges: for if (a),
AB
(b), AC
AD (c), the three adjacent edges of such a
S f e
596 There are onl y two si ns of tne product of three fac-
*
roduct of
three fac- tors, though they may arise from eight different combinations of
tne s ^ ns ^ tne component factors; they are as follows:
may be
positive or /1X
negative. (1) ix +b x = abc.
+ c
(2) + a x - -b x = abc.
c
(3)
- a x +b x c = abc.
(4) a x b x + c = abc.
(5) + a x + b x c = abc.
(6) + ax bx+c = abc.
(7) -ax + bx+c = abc.
(8) a x - b x - c = a b c.
33
only in common, have the same sign: whilst those pairs which
have only one point in common, and all whose three edges have
therefore different signs, have also different signs.
in symbols only.
Linear re- 600. When, however, lines are assumed to represent quan-
tkmsof titles like velocity (v) and time (), which are different in their
nature > anc* tnere ^ore admit of no comparison with each other in
ferenUfods
perfectly respect of magnitude, the first assumption of them must be per-
ary *
fectly arbitrary :
thus, if v denoting a certain velocity, be repre
sented by an assumed line, v denoting any other velocity, would
be represented by another line bearing the same proportion to
the former, that v bears to v : and in a similar manner, one line
may represent a time t, and another line any other time t 3 if they
bear to each other the proportion of t to i but the magnitudes v
:
Numerical 601. The same remark applies, and for the same reasons,
to ^ ne representation of essentially different quantities by means
foctlVarbi-
trary. o f numbers, the values of their primary units being perfectly
quotients, when
the dividend and divisor are assigned both in
|
same kind, the quotient is an abstract number or a numeri
cal fraction (Art. 590).
If the dividend be an area and the divisor a line, the quotient
is a line which contains, with the divisor, a rectangular area
equal to the dividend, (Art. 592).
If the dividend be a solid and the divisor a line, the quotient
is the rectangular base of an equal rectangular parallelopipedon,
Explana-
tionofthe
(JQ4 THE determination of the highest common divisors
,
common tne processes for finding the greatest common measure and the
multiples least common multiple of two or more numbers are involved
of algebra- L
.
i)etection
606. In the first place, there may exist a common numerical
of simple divisor of the coefficients of all the terms of the numerator and
numerical .
of the fraction
Examples.
37
equivalent form
3x + 5y
101 lW
+ 12436
is reducible to the more simple form
607- In the second place, there may exist a simple symbol, Detection
or powers or products of symbols, by which the several terms
of the numerator and denominator may be divided without intro- braical
considered
,
. .
*
a
which is
equivalent to
- ax
and
. a
a+ x
x
x-a
which more simple than the former with respect to the dimen
is
sions, though not with respect to the number, of its terms but :
98). It is as follows :
"
Rule. Arrange the numerator and denominator, or the ex
pressions whose common divisor is required to be found, ac
cording to the powers of some one symbol of reference (Art.
576), and divide one of them by the other, making that
39
expression the divisor which does not involve the highest power
)f the symbol of reference continue the division until the highest
:
610. The application of this rule will very generally intro- intrusion
r!
coefficients of the symbol of reference, which are foreign
puce eo*g ^ u i
|;o the most simple form of the compound divisor which is sought tipliers of
jcbr,
and which a subsequent modification (Art. 6ll) of the pou nd fac-
to s de er -
fcorocess will enable us to exclude the following; is an example
:
r F |
mined by
pf
the occurrence of such a coefficient. Let it be required to the Rule in
Art 609
the fraction
[-educe
3
x +4>x* + 5x + 2
x*+4,x*+5x + 2 (x -
-x + x + 2
2
-x - 5x-4>
2
4>x +4
may
^
likewise to be absolutely prime, if it be not resolvable into rational
factors.
614. The following is the general form of the process for Form and
finding the highest common divisor of two algebraical expres- ofthe"
sions ; its proof will readily follow from the Lemmas. P rocess. preceding
B) A
B) Aa (p
Cc) B
C) Bb
Dd) C
D) C(r
rD
can only disappear through division thus, if a = ex and b = dy, then ab = cdxy,
:
or in other words, c, d, x and y or all the separate factors of a and b are found in
VOL. II. F
42
x -a ) x a (x
a?x - a
2 3
(Dividing by a )
2 -
x a) x a 2 (x + a
ax -a*
ax -a 2
* The process assumes that a, b, c, d... are so chosen, that the quotients
p, q... may not take a fractional form : otherwise it would not necessarily follow
that D would measure pB and q C, when it measures B and C. 612.)
The highest common divisor, therefore, of x -a* and x
3 2
-^
is x a and the reduced fraction is
:
x2 + a x + a 2
x+a
2
x ~3x-70} x 3 -39x + 70 (x + 3
Ix + 70
3x - Qx- 210
2
- lO.z-70
traction is ---
x-W .
* Other
examples are
(1)
a
3
_
+a3
- =
-- 2
: the common divisor is x + a.
4 2
* Q *
(2) 3
.r
44
4z4- 9x 3+2x 2+ 3x
2
a -t- 9z- 8
- 4
- 3x2-54a;+57
- 19
57x 2+1026rr-1083 (3
57 x 2- 126*+ 69
2
replaced by one, if we had multiplied by 4 or 16 in one case,
and by 19 2 or 36l in the other: the process would then stand as
follows :
16
1 444 x
3 - 3249 x 2 + 722*4-1083 (^*-
1444x 3 -3192x2 +1748x
-1152* +1152
*
The determination of a series of remainders in conformity with this rule, is
made the foundation of Sturm s process for finding the number and limits of the real
roots of numerical equations the
process, however, becomes extremely laborious,
:
even for equations of the 4th or 5th degrees, in consequence of the rapidity with
which the numerical coefficients increase.
Other examples of the same class are
3*-4*-15
the common divisor is - 3.
x
x 4 - 3x 3 - 18a 2 + 32x + 96 x3 + x 2 - 14r - 24
_
4x3 -9x 2 -36x + 32 4x 2 +7x-8
the common divisor is x 4.
x 3 -x 2 -21x + 45 = x 2 +2x-15
(3) ~ = :
3x2 2x 21 3x + 7
the common divisor is x 3.
2
15x 3 +35x 2 +3x + 7 5x +l
(4) = :
(5)
(4) To reduce
18 9
!*_?
3
"
3
to its lowest terms.
I8x2 -6aa;-l2a 2
- 2a 3 (x
18ax 2 +3a 2 x- 6a 3 (a
s 2
(Dividing by 3a ) 9a x+6a a
^ a "
x a
2 2
-(b + c)a-b -2bc-c
(Dividing by 2 a)
6 +a+ c, or a + b + c,
,~^
(6)
Make a
To
, ,
reduce - c a
^, to
ae + af+be + bf
the symbol of reference
=
- its
,
lowest terms.
(c + d) a + be + bd\ (e +/) a + be + bf {
a*!*-!,* ~ax-b
the common divisor is ai + b.
a 2 x 2 -2aca2-ft 2 j/ 2 +c 2 2 2
/9^ _ ai-by-rs
~
a*x*+ Zabxy+tfy^c**
1
ax + by +cz
the common divisor is ax + by cz.
a 3 + (\ + ax + x2 +x
W a-
a)
-a 2
_
*"a--.i
a
:
Therefore -
ae +
-
ab + ad+bc + bd --
r^ =
c + d
af+ be + bf e+f
2-
a z +3ab
49
a
2
+6ab + 9b -
2
_a + 3b
2
(b + c)a + 3 (b + bc)~ T+~c~
and the original fraction, in its lowest terms, replacing the
4
factor is
fj-j,
4fl (a+ 3 b)
7x*(b + c)
the highest common divisor being
617. The preceding process will always succeed in detecting The com-
the common factors of two algebraical expressions, when exhibited "/i^afee
under a rational form, whenever they exist.
T^ i f i
X or in the first place, when the same power only of each under a
11
braical ex
pressions
n
symbol presents itself (Ex. 6, Art. 6l6), the common factor, if }jjj| ^y 1
6xy + Bx + 9y + 12
(1)
+ yz + 6x + 3y + 2s + 6
"
^24.3^ + 2^ + 6 and yz + Zy + 2s + 4:
the secondary coefficients of y in the primary coefficients of x are 2 + 3 and
c + 2, which are not identical but the secondary coefficients of 2 derived from
:
-J-
Thus the terms of the fraction
(a - 6 2 )a 2 -
2
(2a
2 - ab - 2 ) bx + a (a - 6) 6 2
fc
(a
- ft) 2 * 2 - (2a
2 - 3ab + b*)x + atP-ab*
are arranged according to the primary symbol of reference x we first find a b, :
Rule for 618. If the highest common factor of three algebraical ex-
^ P ressions A, and C is required, we find X the highest common
B
hfglieft
common divisor of A and B; and then the highest common divisor of X and
more alge- same ^ mutatis mutandis, as of the Rule for finding the greatest
pressions. common measure of three numbers (Art. 106) ; and similarly
for any number of such expressions.
(a + ft) a? -aft
"
(a b) a; aft
where a and ft are prime (Art. 613) to each other, then -- = Ab = Ba, is the
lowest common multiple (M) of A and B : for if not, let m be a multiple of A and B,
whose dimensions are lower than those of M, and let M=my = Ab = Ba : it follows,
therefore, that m = Ax = Bx ;
or in other words, that y is a common divisor of
a and ft,
which contrary to the hypothesis.
is
a?)
=J+ ax- ax - a
(2) The lowest common multiple of x
3
3x~ + lx- 21 and
a-
4
- 4-9 is
14-7.
x a) x"*
a", x~ (a + b} x + a b, x 2 - b
2
x b
) x 4- a, x b , x- b*
x + a, 1 , x + 1).
equivalent form.
The lowest common multiple of the denominators is their
2 2
b :
product
- 6) = (a
- (a
- = a - 2a b + b
2
(a ft) b)
Their sum = 2 a 2 + 2b 2
*
See Appendix.
53
___j -__
x+6 x+2
(2) To reduce the expression to it
Their sum
Dxx ,
Q 2 3
.
_ Q
1 -0?
x + x2 - x*
_x (I + x x
1 + x - x3 x4 1 + x --x 3 - ,
-
its
(4) To reduce
D x-3 v v ,, 2
-r^4i = 5(a-
4-
I)
2
(^4- 3)- 5* 3 4- 25^4-35*
40 4-
I) (x
3
4-
3) (x + I) (x
3
+ 3)
which admits of no further reduction.
In the preceding reductions we have adhered strictly to the
general rule, though the same result can frequently be obtained
by shorter and more expeditious processes, which it is not ne
cessary to notice they can only be safely employed by a student
:
*
The following are examples of the same class :
/1N a + b a b 4ab
624. The rule for the multiplication and division of fractions
MultipH-
11 am
given in Art. 143: and it will be seen, by 1
is e J a reference to the S? l ! .
division of
.
-
.. which is the form of the product of T and -. in one case,
cd b d
must be its form likewise in all others.
ft (*
- become or m, ]
~
4(x-3)" 8(a - - 9a 2
r 3
5) a
56
(1) Multiply x
+
9
4>
and ~^
x+3
.
l) * 3 -.r 2 -9.r + 9 x2
(* + 4) (*+.)- *2 + 7
when reduced to its lowest terms.
=
T6 a 3 + a 2 b-ab*~b 3
when reduced to its lowest terms.
= 2
_
~
x* +2 x+ 1 (a;
+ 2) (x + 2) j?
3
+ 2,r
2
+ 2 a- + 4
when reduced to its lowest terms.
fA\ r\ i a2 + b* . a b
Divide
(4)
^-^ by ^
Division by
628. The
J-equirt- quotient of the division of a by ~ is ai, and
lenttomul- i
,,
tiplication the product of a and is
^ ^ (Art 144,): it
follows, therefore,
57
"
a.... for iUl.
r+i-i+i-*
1 a
l l
IT ~~" U/ 1 -L
J. "T" U>
T
, 1 + j: 1 1
_ 1 + #-1 = _^_
1 + ^r
/ A \
(rf*
\i + x/
-
- for-
1 1 +x
and, therefore,
x
+ x.
multiplying the numerator and denominator by 1
by
l
= a:
2 - lx + 12 : for
(9)
-4 ^2 -7^ + ^ 2 - 7^ +12
run
a:- 3
"T" "T"
+
!
+
~ 12
^_ 493?
3^ 8 -
or + .S oT?l a: - 8
+-
x+ -
^ l+x2
0^1 + !
1+^ + ^
x
(12) x
1 +*
i +x+x
CHAPTER XV.
631. This principle, which is thus made the foundation Its formal
statemenl
of the operations and results of Symbolical Algebra, has been
called "The principle of the permanence of equivalent formst",
and may be stated as follows :
" Whatever
algebraical forms are equivalent, when the symbols
are general inform but specific in value, will be equivalent likewise
when the symbols are general in value as well as in form"
It will follow from this principle, that all the results of and con-
Arithmetical Algebra will be results likewise of Symbolical
Algebra: and the discovery of equivalent forms in the former
science, possessing the requisite conditions, will be not only
their discovery in the latter, but the only authority for their
existence for there are no definitions of the operations in Sym
:
633. THE
continued product of a number or symbol a into Definitions
"
itself, repeated as a factor n times, is expressed by (Arts. 38 anJTpowers
and 39) and the expression itself is called a power (the n th ) of a, m A nth -
and n exponent or index: and it is easily shewn to be a neces- Algebra,
its
am
~n "
greater than n, : for the product of the divisor and
m n m~n =a n+ m -n
am
the quotient ~", or a x a we conclude there
m ~ n is
fore that a the quotient required, inasmuch as, when mul
tiplied into the divisor, it produces the dividend.
635. Observing that the indices m and n in the expressions The equa-
which constitute the equation
633 gen*?*
^m n__ m+n ralized by
3
the prin-
le of
though specific in value, are general in form, we are authorized c
^P
to conclude by "the principle of the permanence of equivalent nence of
forms" (Art. 631), that in Symbolical Algebra, the same expres-
sions continue to be equivalent to each other for all values of those
indices: or in other words, that
am x a n = a m + ",
whatever be the values of in and w.
62
Principle This equation, which forms one of the most important pro
of indices.
called "the prin
positions in Symbolical Algebra, is sometimes
ciple of indices." We shall proceed to notice some of the numerous
conclusions which are deducible fromit, beginning with examples
negative numbers.
The product of a* x al x al = al + l + l = a 1
= a: and the product
of f]a x ^a x If a = a, where /a is the cube root of a : it follows,
05 x as x al x as x as = a** * = a = ^ 4 4
x 4
x
4
x ^
4
x
4
^a ^a ^* :
a~-a a~ =
a
-
2
Examples.
2
a
And, conversely,
1 o 1
a = ^ : a = :,
~
.
a a
ok-
a~~-i
2
64
For ^-
m =a
m-m =
a: and since =1, it appears that a= 1,
a am
whatever be the value of a.
(O" fl -.
m n
For, in this case, (<2 ) denotes the continued product of
m m
a into itself, where a is repeated as a factor n times, and the
index of the powe* which is equivalent to this product is equal
to the sum of the indices of the component factors (Art. 632),
which \$mn or n times the index of one of them: and since
n and m are general in form, though n is specific in value, it
will follow from "the principle of the permanence of equivalent
forms" that the proposition is true for all values of m and n what
soever.
It follows, therefore, that
Examples 1
of the re- (l)
2
x -3 = a~- 3 = a~ = -1 l
.
duction a a
of expres- j i i
+ 1 5
8 a
sions in
(2) H Xfl sstf -fl 6 .
which in- ,
(4) fl
3
x fl~ 4 = fl s ~* = al2=^_".
a*
(5) (ay = ai
(6) {(fyi
_1 2
2 5
(9) (cfb c )
/ b
(
VVV
_ 5
\-L
\40
(10) |flAV
13 14 151 25 73 I 5 73
VOL. II. i
66
J(a
2
-* 2
)
,
(Art. 66.) .
. ax ax
(6)
a-J ax
fj ax ax
^/uu/.tu-
= a-^
it, a, i , , ,,
Jax = 2
i f.t
i -
For -^ , and, therefore,
a-^ax a-Jax
i ax _ ajax-ax + ax _ a J~ax a Jx
*J a
~ \l x
a Jax a Ja x a Ja x
2
ax 2 ax
~ ax
3
a3 x
(v it n ^ ^\~: ~3
+ "5= 1-
CHAPTER XVII.
645.IT will follow, from the Rule of Signs. (Art. 569.) There are
i o i T i AI T/y. always two
that,, in Symbolical Algebra, there are always two roots, diner- roots, with
ing from each other in their sign only, which correspond to the s gn S
same square: thus a 2 may equally arise from the product ax a which pro-
7N2 f *. A ducethe
and a x - a : (a
/
+ b) IT
may equally arise from the product same
<.
S 1 uare -
(a + b)x(a + b), and -(a + b) x -(a + b) : (a-b)* may equally
arise from the product (a b)x(a b) and (b a) x (b a),*
and similarly for all other squares. It follows, therefore, that
in passing from the square to the square root, we shall always
find two roots, which only differ from each other in their sign ;
but it is the positive square root alone which is recognized in Arithmeti-
quare
Arithmetical Algebra, and which may therefore be called the ^
arithmetical root.
646. We
have already had occasion to notice these am- Occurrence
2 2
fl -2aor + a? , which are identical in their arithmetical value, calAlge-
ra>
though the arrangement of their terms. If the
different in
relation of the values of the symbols x and a be known, the
Rule for the extraction of the square root of these expressions,
which is given in Arithmetical Algebra, would require the terms
The sym- 647-In extracting the square root, we follow, as in all other
roots de- operations, the same process both in Symbolical and in Arith-
rived by metical Algebra, assuming the proper relation of the symbols:
process as and the negative is at once found from the positive root, by
1"
mary term of the root, making the quotient the second term of the
root: add this second term, with its proper sign, to double the
primary, to form the divisor: multiply the last term of the root
into the divisor, and subtract their product from the remainder
Jirst.
Examples 649. The following are Examples in which the process ter-
minates.
square
* We do not assume, in Arithmetical Algebra, that x a or a x, are equally
the roots of x8 2ax + a 2 when the relation of values of x and a is unknown, but
2
that x a is always the root of x -2aa; + a2 , and a x the root of a 2
%ax + x2 :
and it is only when the relation of values of x and a is unknown, that there is
nothing to guide us in the selection of one of those forms of the square in pre
ference to the other.
t The Rule for extracting the square root in Arithmetical Algebra has not
been formally stated apart from the corresponding Rule in Arithmetic (Art. 218.)
69
2
a 2ab + b* (a-b: the second root is
2a-b) -2ab + b 2
-(a-b) orb -a.
We obliterate a
2
and make its square root a, the primary
term of the root we double a (2 a), and we divide - 2 a b by it
: :
the quotient b is the second term of the root we add (Art. 547) :
a - 2a x
4 3
+ 3a~x 2 -2ax 3 + x*,
Za 2 -ax) -2a 3
x + 3a 2 x 2
-2ax + x 4 (a?~ax + x
3 2
: the
3 2 2
20 a?4- a .r second root is
2a 2 -
2
In forming the second divisor, we consider a ax as the
single primary term of the root, the double of which is
2a 2 -2ax: we divide the first term 2a*x* of the second re
**
tf + x2 (a+~2-* + 16V
_^__^_ _^_
\
+ _ + rf__
a 4a 3 l6a 5 8a
256a 1
8
5 a:
* CD.
*
Other examples are
71
^ (a
(
a "~~~ x* x* x6 5x*
an indefinite series, where all the terms after the first, are
negative.
This series is convergent or divergent, according as a is greater The square
or less than x : in the first case, both the square and its posi- *^ 2
tive root are arithmetical quantities, and we can approximate bo tn wnen .
61
indefinitely to the value of the latter by the aggregation ofandfess
the successive terms of the series*: but in the second case, thanx *
neither the square nor its roots are recognized in Arithmetical
2
651. But if a, in the expression a 2 a: , be less than x, Square
2
we may replace x by a~+b
2
which gives us
,
negative
a*-x* = a 2 - (a 2 + b 2) = - If, "> ra bols.
and therefore,
* a = 5 and =
Thus, if x 1, we find
which Is the same magnitude, though affected by different signs, and those
1
to express
s
^
ns express affection or quality, and not magnitude (Art. 556) :
2
the symbo- thus, if + b be the square contained by the sides + b and + 6,
2
root o^-l?. or ^ an d b, then b will express the equal square con
tained by +b and b, or by b and + b. (Art. 593.)
2
In the second place, the squares b 2 and b are respectively
2
equivalent to the symbolical products of 1 and b and 1 and ,
b*: for
2
1 X b =lb*=b* (Art. 30), and -1 x ,
and
It thus
appears that J- 1 is the sign, which is determined
by the rules of Symbolical Algebra, as characteristic of the
(b J~l) = b C/^l) = 6
2 2 a 2
x - 1 =- b
2
.
is
653.
arithmetical:
The
-iii
2
square roots of a are a and
and the square roots of a 2 are a ,J I and
#, the first
/
-
of which
and imagi
nary.
for the symbolical meaning of v 1, or of the square root of 1 is, that, if
it be multiplied into itself, it will reproduce - 1.
73
-
served that the several roots , , a J^\, -aj ^T differ
a =- I x a)
impossible or imaginary.
654. It appears therefore, that the introduction of the addi- The ex-
. . i- . traction of
tional sign >J 1, will enable us to give to the inverse operation higher roots
of extracting square roots the same general application that
gquJJ^wiii
we were enabled to give to the inverse operation of subtraction, render the
, , ,, , ,. .
^introduc- . .
by the use of the sign , as distinct from its use as a sign of turn of
operation and it will be found generally that the inverse
:
VOL. II.
CHAPTER XVIII.
-
Symbols affected by the sign J 1, may be further affected,
like any other symbols, with the signs + and but it will :
Examples.
^ $a + ^b J - 1 + 40 + 3b J^-l = 7 + 5b
(4) (fl
+ 6V-~l)-( - fl
(5) (a + b J- 1 ) + (c +
(6)
(7)
(8)
(9)
(10) (aj~^ = a\
It will be observed that the series
is
equivalent to the series
thus the 1st, 5th, Qth terms are 1 the 2nd, 6th, 10th terms :
are J- 1 : the 3rd, 7th, llth terms are - 1 : the 4th, 8th, 12th
terms are J 1.
J - 1) (a - b J - 1) =
2 2
(12) (a + b + b : the common process
of multiplication applied to this Example, would stand as follows;
-abj^l+b*
a*+b 2
(13) (a + bl) (c +
76
- b -d = ac-bd-(ad + bcy- 1
(14) (a lTi) (c l)
2
+ 2fl6/=~l.
6^n)=fl -6
s
(15) ( +
(16) ( fl + 6 x/^"l)
3
= 3 - 36 8
+ (3a 2 b - b^J^l.
(17) ^(-rf+aa*-P)-*(-fcW^l-
(18) J(4 a*-l2abJ~^l-9b*)=(2a-3bJ~=i).
f
roots are -a -
J b and - c J- d, and are not required to be
considered in this Example.
I I 2a
a+bjl a-b J -I a2 + b2
- 1 a + bj-l
CHAPTER XIX.
656. THE
process of extracting the square root of a number
Binomial
, quadratic . , . .
x- -4 = 0,
may present itself in connection with another term which in- equations.
volves the simple root, as in the following forms (Art. 383) :
x*-px = q (1)
q (2)
-q (3)
-q (4).
q
= (6)
(8).
The three first of these equations have been considered in
Art. 384 : the last belongs exclusively to Symbolical Algebra.
It is obvious that these four equations are convertible with
each other, by merely changing the signs of p and q : and it
will follow (for the same reason) that the symbolical solution
of one of them will furnish the symbolical solution of all the
others.
--
equations.
Since
and since x^ px q, it
follows that
we change
If the sign of p and retain that of q, as in equation
(6), we get
we retain
If the sign of p and change that of q, as in equation
(7), we get
*
This process is, in every respect, similar to that which is followed in extract,
ing the square root :
thus, if it be required to find the square root of
79
s
is always negative.
Examples.
Thus in the equation
a?- x -90 = 0,
and, therefore, x
--= ,
and x 10 or 9*
7 -px-q
2
consequently, if x 2
px-q -0, which the conditions of the equation require, we
/ \ 2
p p
have
I*-^J =4+9^
and, therefore, a:-
1 =V ($
In the equation
4.T-21 = 0,
\ve find (x + 2)
2
- 21 + 4 = 25,
and a: = 3 or 7-
Inequa- 660. Again, it appears from the formulae (11) and (12)
both roots of solution of the equations
*-*>* + 9 = (7)
possible: /oN
inequation X 2 +pX + q = (8),
always less than -, and therefore the signs of the two roots
rn
L-
are determined by the sign of .
7 1
and, therefore, x -- = -,
and x = 4 or 3.
we find (.r + 6)
2
=36-35 = 1,
and x =J or 5.
If, in the same formulae (11) and (12), is less than q, then
^-
2
- and the roots of equations and
q is negative, (7) (8) will
x 2 - 8x + 25 = 0,
we find (x
- a
= 16 - 25 = - 9,
4)
4
9_97
4 36 y
and, therefore,
661. The roots, which are thus determined, bear a very Composi-
simple relation to the given quantities of the equations to which Coefficient
they belong ; their symbolical sum being always equal to the
of the se
coefficient of the second term with its sign changed, and their and of the
and
HI
VOL. II.
82
simple
and, therefore,
= ,r 2 ^;
+ q.
2
3
px + q = x2 px + q (a
2
pa + q) = a.
2 - a? p(x a) :
and, therefore,
-
= x +a p x (p a) :
s
it follows, therefore, that x a is a factor of x px + q: the other factor is
x-(p-a).
t In Example 1, Art. 389, the second or negative root is 3, and in Example 3,
it is 14 the two roots in Example 5 are both negative, one being
: , and
the other 4 : the solutions in Examples 2 and 4 are ambiguous, and both the
83
(1) Let - - 3 --
x
- -1
x+
- . Examples.
x 4- 2 1
\ 6J 36 3 36
_7_ :k
17
6~ ~6~
and therefore
x 4, or
- .
There is
only one arithmetical root: but the negative root
- fully satisfies, as may be found by trial, the symbolical
.
1-x x 10
The several processes of reduction and solution follow the
same order as in the last example.
7 3
~ =
x= 5, or
o o *
27Y 729 91
27 i
-T--*!
4 4
IS
and ~
T
* 2 -4* 3* 2 +ll
(4) Let + + 8 = 3*,
J
- 172* =-559,
producing
^65. We shall endeavour to connect the solution of the
quadratic with the
following problems producing quadratic equations,
and inter- illustration of some of the more common principles of inter-
a n0f
P retin S the results of symbolical
their re. operations.
suits.
* In the
examples of quadratic equations which are given in the Note, Art.
389, the Symbolical, as distinguished from the Arithmetical, roots, are in Exam
ple 1,
- 10 : in Example 2,
- 7 : in Example 4,
- : in Example 5,
- 1 : in
j--
-
Example 6, 10 : in Example 8,
-
^ : in Example 11 ,
2 + J ^T, or 2 - J H"jE
85
and, therefore, x - 10 x =
2
24.
we (x - = 25 - 24 =
2
Completing the square, get 5) 1,
and, therefore, x 5 = 1,
and x = 6, or 4.
-5) = 25 -25 =
2
(x 0,
x - = 0,
ft
x= 5.
(.r~5)
2
= 25-26 = -!,
* The product
of the parts into which a number is divided, will be the greatest
when those parts are equal for if 2n be the number, and n + x one of those parts,
:
- =
= (5 + V 37) x (5
J^l) = 25 + 1 26*.
75 = |42-3(n- l)j 5,
I5-P=50,
n 2 - 15;? =-50,
225 25
15
w= 10 or 5.
* Another
equation of the same class is given in
Example 7, Art. 410 : the
student particularly recommended to study the problems leading to quadratic
is
equations, which are given in that and the following articles, with especial reference
to the more enlarged views of the interpretation of symbolical results which are
given in this Chapter.
87
(3) Given one line, to find another such, that the rectangle A geome-
which they form, shall be equal to the square of their difference, jj}
Let a be the given line, and x the line to be determined generally :
ax = (x-a) 2 or (a-x)
2
(1),
x*-3ax = -a 2 .
3a a J5
X =
~~2 ~2~
=
3a
~ ~~
aj5
, or .382 a, nearly.
In the problem proposed, it is uncertain whether the line Its solution
,. !
include* . .
7
to be determined is less or greater than the given line: ana tne two cases.
of
symbolical solution, being coextensive with the enunciation
the problem, includes both cases.
The first case is equivalent to the following problem. Its first case
stated and
solved.
88
"
To divide a given line into two such parts that the rectangle
contained by the whole and one of the parts shall be equal to
the square of the other."
If we call x the greater of the two parts, and, therefore,
a x the other, the equation will become
a(a-x\=x* (2),
*-
(v/5-1) (v/5 +
L a and - **
>
1)
a>
one being positive and the other negative: the solution is there
fore not ambiguous, there being only one positive root. But
if x be taken to represent the less of the two parts into which the
line is divided, the equation becomes
ax = (a x)
z
,
rectangle
T
formed
P int in a given line P rocluced > sucn tnat tne
find a
-
x) (3),
~
the two roots of which are -a, and - "/ a, differing
the less of the two roots being excluded by the conditions which
it imposes.
It thus appears that the same problem lead to one
may
equation whose solution is unambiguous, and to another which
is ambiguous but in the latter case, the conditions of the pro-
:
89
(1).
and x + c = ,J(c
z
+ I
2
),
or x = J(c +2 2
J(c + I ) c.
/
)
c and 2 2
Its
and it is found that the only arithmetical value of x, which
syn
2
thesis. will satisfy this equation, is N/(c +/ 2 )-c.
The preceding process of investigation may be called the
analysis of the problem proposed, and immediately leads to
its synthesis or construction, which is as follows.
Bisect the chord BA in E: draw perpendicular to BF AB
and equal to CD or
join EF, which is equal to
/:
J(EB* + BF 2
)
= J(c* + I 2 ) = c + x = EP :
BA
directs us to bisect BF
in E: to draw
perpendicular to
EB and equal to CD: to join EF: and in BA produced to take
EP equal to EF and we finally assert that
: the tangent PT
drawn to the circle from the point P thus determined, isequal to
the given line CD. We
then subjoin the proof, by which the
correctness of this construction, or synthesis, is established.
Theprocesses of exposition in Geometry are generally syn-
The pro-
and - (Art. 659), and shew that there are two points on dif
/?
ultimately equal to *.
a + b-c
If, in equation (l), we 1 + e by 1 - the values of
replace e,
x will be expressed by
which are both of them negative, shewing that there are two
points on the left of the point Z), which answer the conditions
of the problem but if we reverse the directions, and, therefore,
:
the signs of a, b, c and x, the two roots, under the same cir
result may be otherwise obtained from the general expression for a? (4): for if we
take the negative sign of the square root, and make t = -- c, we get
extracting the square root of 1 + -~ by the process given in Art. 650, Ex.
we get
ab
~ b-c
-A = a + b-c-ec, and
{
The following are particular examples of this problem.
T + 8) <> + 6)
(1) Let a = =
8, ft
6, and ( c = 3, and let
(,r + 3),r
93
The problem, whose solution we have been considering, was The pro-
known to the ancient geometers under the name of the Deter- Determi-
minate Sectio?i, and is said to have branched out, with its kindred nate Sec-
B P C D P
I | I I I I I I I I I I I I I I I I I I I
A
If we change the signs of a, b, c, and therefore the directions of the lines
they represent, we merely change the signs, and therefore the directions of the
values of x.
(3)
ratio either to the rectangle formed by a given line and another of these inter
(1) *-p = 0,
(2) x*-px + q = 0,
px + qx r = 0,
2
(3) x3
2
(4) x*-px*+ qx rx + s= 0:
and similarly for equations of the fifth, sixth and higher orders.
Any equa- 668. Any equation, may be solved, by processes which have
tionmaybe -, -,
, .
,, 1-11-
. .
i i
solved already been considered, it it be resolvible into simple or quad-
rat * c ^actors f r those values of the unknown symbol (x), which
:
resolvible
into simple
severally reduce those factors, and therefore their product, to
1 *
ratio zero, and no others, will satisfy the conditions of the equation :
factors. and it is
merely necessary to solve the several equations which
arise from successively equating those factors to zero, in order
to discover all the roots of the equation*.
It is not our intention, in the present Chapter, to
attempt
to prove the necessary existence of such simple or quadratic
factors in all equations reduced to the ordinary form, or the
possibility of discovering them, whenever they exist: but we
value of x different from these, will make u the product of three significant factors
and, therefore, not =0.
97
primitive equation*.
Inasmuch as the equation x s l =0 gives x3 = 1, and there
fore x = yi, it follows that there are three cube roots of 1, which
~ 1 + ~~
l
"
are 1, \ and
-l^V3 v/-l\3
r
- 1 3 V 3 V - 1 + 9 T 3 V3 x/ - 1 8 Their pro-
v )r /
= =
\~ ~2 j ~~8~~ 8
The square of one of these imaginary roots is equal to the other : for
and -
-/ 4 2
2
If, represents one of these cube roots, a will represent the other,
therefore, a.
2
and the three cube roots of 1 may be represented by 1, a, a or they may be :
2 3
represented by a, a a , since a
3
,
= 1.
Again, since a x a = a
2 3 = 1, it follows that a
2 = ,
and a= : and conse
where a 3n = 1 ;
a 3n + 1
= a 3"
x a = 1 x a = a ;
and a 3n + 2=a3nx 2 = lx a2 =a 2 .
VOL. II. N
98
are con
properties of the cube and higher
The roots of 1
x3 +l=0.
Since x
3
+ = + z -x+ values of x, follows
1), for all
1 it
(x 1) (x
that
x 3 +l = 0,
For x 6
1= (# 3 3
1)(# + l) for all values of ,r, and the roots
of x - =
6
1
(Art. 668,) are therefore those of
x3 - I = and x 3 + 1 = 0.
e
The same expression x l may be likewise resolved into
the three quadratic factors
2
x*-l, x + x+ 1 and x 2 -x + l:
* If a be one of these
imaginary roots, the three roots may be expressed
torm in
Art. 667.
673. The roots of the equation considered in the last Article, The same
may be exhibited under other and equivalent forms, which are bite.
1
under
(J eren
sometimes susceptible of more complete arithmetical reduction. m
f r
For if we divide the original equation
2
by a: , we get
A
2 _
and therefore
100
and therefore
-<?) (2).
and therefore
and, consequently,
(3)
(4)
(6)
101
we get
If the equation be
If the equation be
.r
4 - 20 or + 64 = 2
0,
we get
*==V(5 + 4)* x/(5-4)
= 3 l =4 or - 4, 2 or - 2.
and its roots determined by the solution of two quadratic equa- root of
. their first
tlons -
member as
For if we express by X the first member of the equation, arranged in
and by u the terms which are obtained in the root before the
process of extracting its square root terminates, then Xu? is
X- u *
= -v 2 ,
we have
X= u 2
v
2
= (u + v) (u v),
*
The condition, upon which the success of this process of resolution depends,
may be easily discovered by extracting the square root of
x4 + px 3 -\- qx* + rx + s,
(which is the general form of the first member of a biquadratic equation), as follows :
The last remainder but one, with its sign changed, or (%r - q\ rx-s,
This
p = =
last condition will always be satisfied, if and r as in the
0,
Examples in Art. 674.
The preceding method will succeed in
detecting u and , whenever they
are of the form x 2+ ax and bx + c or of the form x* + ax + b and c but not
:
:
And -
~-+ 24,r + 256 is a complete square, since
3x Sx 3x
u+ v = x2 +-- + -- +
44 16 = x 2 +
2
+ 16 = 0,
and u v = x2 -\
-----
4 4
l6 = x* 16 = 0:
3 -
,
4 and - 4 *.
2,r
2 - x) -2
10 T 3
(1) a:
4
+-39x-
The factors are a
2
+ -^ +9 and x 2 - 9 : and the roots are 3, - 3, and
o
13 J- 155
2
8r 10.r
The factors are x + 1 and x
a: ^~
;r- + 1, and the roots are
-3, 4, 3 and -.
104
We thus get ii
2
= (x
z
-x + I)
2
and v
2
=4, and therefore
u + v =x 2
x+ 3= 0,
u v = x2 -x\ =0.
The four roots are ^ and .
/ /o
(3)
Let the equation be
mine a right-angled triangle, the sum of whose sides containing the right angle
is a, and the perpendicular from the right angle upon the hypothenuse is 6."
If a: be taken to represent one of the sides, and therefore a x the other, the
well-known properties of the right-angled triangle will readily furnish the equation
{.r
2 - ax + b V(a
2 -
2ox + 2a 2 )} (r 2 - ax - b V(a 2 - 2ax + 2;r )}
2
= x4 - 2ai3 + (a 2 - 26 s ) i 2 + 2o6 2 x - a 2 6 2 = 0.
2 2
The two roots
^^V JT+ & ~^\/( a2 + b )l belong to the first of these
equations, and express the two sides of the triangle which are required to be
determined the other two roots
:
belong to the second equation, and are totally foreign to the problem proposed.
This problem is given in the Arithmetica Universalis of Newton (Prob. vn.
Cap. II. Sect, iv.) a juvenile work, but which is every where pregnant with marks
of the extraordinary genius of its author the whole section on the Reduction of :
Geometrical Questions to Equations will richly repay the most careful perusal of
the student.
105
(*
2
-7-r-24)
2
-.r
2
+7*-18 (a;
2 - 7* - 24 - 1
2 2
2( r -7,r-24)-i> -.r
l ) +7*
169
4
Therefore (x 2 - lx - 24 - A) 2 = i^ ,
2
and ,r -7,r-24-l==b .
-7
=*=
/173
of values 9 and - 2, and .
7 =*= /173-
the two others - belong to the factor
x 2 - lx - 24 - j(x 2 - lx + 18) f
= 0,
*
Equations, such as that in the text, involving square roots of the unknown
symbol or of expressions involving it, may be generally rationalized, by trans
posing all their terms to one side, and by multiplying together the several factors
which arise from changing the signs of the quadratic surds: but it must be kept in
mind, that there are generally, though not always, roots of the rationalized, which
are not roots of the proposed equation, and that there are many equations, in
the first root only which belongs to the proposed equation x - J%x= 4;
OCCttT. it is
we form the two factors, which are the products of (1) and (2) and of (3)
first
* 2
11148
-625-* + W
44388
= ,
(?)
7 v
625
The first of these roots is the proper root, and corresponds to the primitive
equation (1), or to the rationalizing factor (5) the second is a root of solution, :
and corresponds to the factor (6), or to the subordinate factor (4) it may be :
further observed, that the subordinate factors (2) and (3) are satisfied by neither
of these roots, and that they are equations which admit of no solution whatsoever.
In a similar manner, the equation
18
of whose roots 9 and - , the first is the proper root of the proposed equation
(1), and the second a root of solution only: also the proper root of the equation
T - 5 ) ~ ^(l + 45 ) ~ i
14568980
is 20, and its root of solution : and in both these cases there are irrational
but there are no values of x which will satisfy its rationalizing factor
x - 2 V* 2 + * + 5)
- 14 = 0.
107
l6- 2
+3M = 12 (2),
or x z Qx + 16 = 16,
and x = or 6,
j (x + 3)
2
+ (x + 3)}
2 - 7 (* + 3)
2
= lx + 71 1
{(x + 3)
2
+ (x + 3)}
2
-7{(x + 3)
2
+ (x + 3)}
= 690,
u 2 - 7^ = 690.
The values of u are 30 and -23: those of x are 2, -9,
- ,
which are the four roots of the proposed biqua
dratic equation t.
In the equation
a?+x +
x
5
+
v/(^
2
5
+x+
__ 5)
116
~~25x
* If we u
take the negative value of or 1, the resulting roots 3/^/ 6,
if we make 2
; = w, the roots of the resulting equation
v/( <2? ~T" <X? ~h O )
w2+5M =~
H6
25
4 29
are - and , of which the second must be rejected.
5 5
The roots of the equation
u= -
,,
2 ^ = -5K ,
>J(x +x+ 5)
20
are 4 and - , of which the second must be rejected, being
u = a* = 27, or - 33,
- 5 x 243, or - .
V39135393 t-
(I) ( 1)
per solu-
affected by a common denominator of the index of the unknown
n svmbo1
equation
or its powers : for if the highest power of this symbol
was x ^ > tne process of solution would give us the values of x~z ,
byadeno-
initiator of and the
number of those values would not be increased by the
the index of
to admit equally all the forms which are proper to the different
5
symbolical values of # (which will be found hereafter to be 5
in number,) we should have 5 equations and 30 roots, of which
6 only are the proper roots of the proposed equation, the rest
being roots of solution.
It will appear, likewise, that the reduction of the indices of
the unknown symbol, when one or more of them are frac
tional, to a common denominator, will convert roots of solution
into proper roots, unless the change from one form to the other
is accompanied by a limitation of the roots to be extracted,
x* + x* =6 (2),
x 3 ~6J-l6 = (1),
The bino 679. IN Chap. viii. (Art. 486...), we have proved, when
mial
theorem the index n is a whole number, that
when the
2
index is a ^3
whole (1 + = 1 + nx + n (n
- +n (n
- 1) (
-
2) + - :
*)" l) ^-^-g
number.
and it be seen, from an examination of this series, and of
will
the law of its formation, that the powers of x and their divisors
are independent of n, and that the coefficients of
x2 xs xr
TTz iTzTs i.2...r
are n, n (n - -r+
n(n-l), n(n-l)(n-2), . . .
, 1) . . .
(n 1),
th
being, for the (1 r) term, the continued product of the de
+
scending series of natural numbers from n to n r + 1 .
The same n
680. This series for (1 + x) is perfectly general in its form,
series is
equivalent though n is
specific in its value, and it will continue therefore,
when the "the
index is by principle of the permanence of equivalent forms" (Art.
B
perfectly 631) to be equivalent to (1 + j?) , when n is general in value
general in
value as as well as in form: and it will consequently admit, in virtue
well as in of this equivalence, of being immediately translated into the
form.
whole series of propositions respecting indices and their inter
pretation, which are given in Chapter xvi *.
The pro- 681. Thus "the general principle of indices" (Art. 635)
^ws that
1 + M+ (
- 1) + (- l) (
_ ft)
~L-^
+ &c.,
* See
Appendix.
Ill
1 +n x+ - 1) + - _ 2) - + &c.
( l) (
I
+ (n 4- n (n +n - 1) (n + - 2) + &c
)
-^-^ >
(1 + x}~ n = I - nx + n (n + l)
^ - n (n + 1
) (n + 2)
^-^ + & c :
-
product n(- n-l)(-n- ) by - n(n + l) (n + 2), and simi
larly for the subsequent terms.
683. The series for (1 #)" is deducible, in virtue of the The series
r
same principle (Art. 631), from that of (1 4 #)", by changing _
^ T v.
the signs of those terms which involve the odd powers of x:
we thus get, as in Art. 491,
3.4.T 2 3.4.5.t 3
112
it follows that
Examples. Thu ^ A + Au = a + rf = a * *
+ we have
1
(1 )
if
(l
l
+) ,
= a* and u= -,
CL
d An - 2"
w2 = u3 =
n
= a2 " +n - 1)
^+
2
therefore (a + x*) 1 .
(
.
4
j
2 "
multiplying a into every term of the series.
If A + Au = a*-ax = a 2 we get
(2) (l-^,
A = a*,u=
and therefore
Series for 685. Since (1 + x^ means the square root of 1 +x (Art. 636,,)
(1 +x)a. it follows that the corresponding series \ replacing n by J in the
n
series for (l + x) },
we get
1x1 x2 1x1x3 x3 1x1x3x5
--- .
*
2
_f +
8
2
16
_ ^128
113
6 s
+ a?_ _x^_ + x ~~
5x
20 8a 3 16 a* 12Sa 7
These results are identical with those given in Art. 650,
Ex. 1 and and the student will not fail to observe the readi
2,
ness with which the terms of the series are formed and the law
of their formation ascertained,
compared with the ordinary pro
cess of extracting the root.
square
1.2.5.8
1.2.3 34 "1.2.3.4
p-9*
--
,
p-2(/
,
p-3 9 .... we shall find
+ -
VOL. II.
114
2
Examples. Thus, in the expansion or developement of (a + ax), where
a
we get
2
3
lo = a5
3 /
+
Q j,
--
^7 /r
2
+
3.7*17 3?
(a +ax) ^l
.
-^ j-^ .
2
uF"" 1 .2 . 3a 3
\
3.7.17.27 i*^
4
+
10 4 1 . 2 3 4 a
. . .
/
*
= 5
+ .3 x
5-=-
M.
_ 10 5 , x
fit
+ .0595 x ^-
a 5
-0401625 x -^ +
fl *
-r-
. . .
by multiplying
5
into every term and reducing the several
numerical coefficients.
fit
gives
2
1 x 1.11 ,r 1.11.21
= +*
.
f
+ .1 x - + .055 x + .0385 x . . .
distinction in the
origin of the factors of the entire coefficient,
by calling one of them the exponential, and the other the ter
minal coefficient of the series thus, the
exponential and terminal :
coefficients of the
(1 + r)
th
term of the series for (1 + x) n are
*
Other examples of the same kind are,
35 x 4 63 x 5
"128 ~~25b~
+-
= a2 + .4 x a
2 - .12 x + .064 x ^ - .0416 x + ...
115
- -
M ( 1) ... (n-r+1) and -
J. X ,c
-
X o , . . 7*
respectively,
3 x _ 7 x _ 17 x -27_ 3 x 7 x 17 x 27
4
10 4 10
1x2x3x4 ,
3 x 7 x 17 x 27
and the complete coefficient required is r^,
1 x2x3x4xlO
or .0401625.
76
x6x13x20x27x34
xlx 2x3x4x5x6
j ^ t
a
.
_i a s
the 4th term of (\Jx V )
4
, where A= x*
i
, n= -- -
,
is
x*
1x5x9
116
The limits
verse ratio
of any two
consecutive (l
coefficients
much
+
690.
coefficients
r)
th
,
The
of the
will
be expressed by
series for
values of r are
r
th
and the
= --
TZ +
r
l
1
and
1
* : and inas-
(or n, when n
a positive whole number), is it follows that the
limits of the values of this ratio are n and 1, and that, when
the series does not terminate, it
perpetually approximates to the
latter.
-
The limits 691. In a similar manner it will appear that the inverse
n
verse ratio ratio of any two consecutive terms of the series for (1 + x)
of any two
n_r+ j /n + I \
consecutive is
expressed by x = ( - - - 1 ) x, the extreme limits .
terms of the r \ r J
n
^ wn i cn are nx an d % > and the series for (] + x) will there-
The series
f re sooner or later, when it does not terminate, approximate
>
sooner or
later to a which converges or diverges according as x is greater or less
than 1- (Art. 432, and Note.)
mial series."
i nverse ratio ( -- - x of two of
J
its consecutive terms: if
(n
-
+ \)x,
T, according as
--
n + l
1 is
positive or negative: thus,
1
* For the complete L^!LL
coefficient of the r* term is
1 ,6 . O . >
( 7* ly
" ~
and of the (1 + r)* term is
w( * + - ?
and the quotient
\"^~J^ ,
(n +-- ,\ + --
1
Tf
If * = we get r = (n l) lT
but
/n+l
f-~ \
{ 1J
1,
p- ;
if
ij
x=-l,
we get r = -
?
3 9
n ~ a X~ n r *s tne next wn l e number, which
2 To
45
is
greater than , and therefore the series becomes convergent
-j-
= 22,
12
13
and the convergency begins from the 23rd term : if n and
693. It appears, therefore, that the series for (1 + #)" and The series
^^
ve]
or whenever the binomials, whose powers are developed, are n
and x is greater than 1, we may obtain positive values of r from both these
formulae, which shews that a point of convergency is in such a case followed
by a point of divergency.
*
It is in the inverse operations of Arithmetic and Arithmetical Algebra,
such as Division and Evolution, that we meet with interminable results, and
in which the arrangement of the terms of the expressions, which are the subjects
of the operations, in the order of their magnitude,is absolutely necessary to enable
Thus (1 + j;)
2 = l + 2.r + x 2 and (x + I)
2
=x 2 + 2x + I, and the results
2 2
1 + <2x + x and x + 2x + 1
are identical in their arithmetical value, though not in their arrangement, whether
x be less or greater than 1 but ( 1 + x)~ 2 = 1 - 2x + 3x 2 - 4a 3 + ... and
:
For if p = 1 -- -
, it will follow, since 1 increases as r increases,
,
-
that the
T
sum of the
n
of the series for (1 + x) ;
but the sum of the geometrical series
T-Tx+Tx 2
-Tx*+...
T -
which is , will be greater than o- since p is necessarily less than 1, when r is
y
4 x 42 2 x 4 x 43 43 x 4*
3
2 x 4 x 4
5
we should find
(7 = 2.236026 nearly,
T= .0000534 nearly,
= .000032 1 nearly,
T = 4 T- =
.0000427 nearly,
1 +x 5
a- + -^ = 2.2360687 nearly,
1+3
and therefore s differs from 2.2360687 by a quantity less than
.0000321.
-
-,
1
T
+ px
and -j
1
-T
+x
that <r +
,
T
1 + x
will be =
- less and <r + .
I
T
+ px
will be greater,
T T (l-p)Ta;
+ /ox I + .T
CHAPTER XXIII.
The signs 695. THE signs of affection, which we have hitherto used,
1011
-
hitoto are +, -, J^l and -^=1; the two first of which + and
required -
may ^ e conveniently replaced by + 1 and 1, if we consider
nized, are them, like the two last, as factors of the symbols which they
Ur
Thus + a = + 1 x a and - a = - 1 x a, in the same manner
f
affect*.
biqu a-
dratic roots
that fl
J~^ = ^^ x and - a J~^l =- J^l x a (Art. 652).
Of these signs, the two first + 1 and - 1, or 1 and 1, are the
- 1, + -
J 1 and 1, be J
easily shewn to be the four may
biquadratic roots of 1 t.
1 and - 1 : those of .r
2
+ 1 = 0, are >J^l and - J~^l.
121
inasmuch as we have shewn (Art. 66 9), that the three cube roots of l .
roots of 1 are
it will follow that there are also three cube roots of a 3, which are
fore
699.
3i
A
- -
In a similar manner,
= s/
1 x fl :
we have - a = - I x a 3 and there-
i i
3
IT Of the cube
rootsof- 1.
it will follow that there are also three cube roots of a 3, which
are
- fl, a,
K-j*
none of which are arithmetical.
roots of 1.
(1):
and inasmuch as
*-! =(*-!) (**+!),
for
and also
VOL. II.
122
it follows that
v a" = tf 1 x a and J -a
n
= l] -I x 0,
f-
= (x~ i) (*- + *-- +...*+!)
1
-jir
which becomes zero when x is 1 but there is no arithmetical :
^"^^Q
value of x, which, under any circumstances, can make the second
n~ ~
factor of this product x + x n 2 + . . x + 1 equal to zero.
l
.
2
704. The
roots of the equation a: 1=0, which are the The square
ro
square roots of 1, are 1 and 1, one of which only is arith-
metical.
If we represent the unarithmetical root or 1
by a, the two Their pro-
2
roots of the equation will be expressed by a and a : for
= (-!) =!.
3
705. Theroots of the equation a; 1=0, which are the The cube
c , roots of 1.
cube roots ot 1, are
and since all its terms are positive, there is no positive or arithmetical value
Thebiqua-
dratic roots
706. The roots of the equation
_x*-l =
J-l
. _ and
0,
,
which are the
.
ofl.
biquadratic roots of 1, are
(Art. 695, 1, -1, -*J-1
Note). The two first are also the square roots of 1 : the two last
are the square roots of 1, or of that square root of 1, which
- I or
is not arithmetical: and if a represent 3 the J -J-1
four roots will be expressed by a, a 2, a 3 and a 4 .
9 1 1
or x 2 + ~2 -f 1 + - + 1 = 0;
the solution of this quadratic equation gives the values of x in the text.
+ For if a = .309 and b = .951, we find
2
V(a + b 2) = J .999882 = .9999 ...... = 1 nearly :
or if a = and ^v , we get fl
125
a, a2 , a , a 4, a 5 ;
3
a prime
3
a, a?, a ...a" number.
,
(l).
In the first place, no power of a, whose index is less than n,
can be equal to 1.
Vfa* + b 2
)
= J .999639 = .9999 = 1 nearly:
orifa= v
5+l //5_^5x 2 2,
=
3 + V5 5-V5-=1
-+-
and&=\/ \ O-^-)/ , we get a + b
=-
where a _i.
and b =
^ : we therefore get V(
2
+ 2
& ) = V(? + i) = 1-
a 3 = -.809 -.5
-1,
which are also the quinary roots of 1 : it will be found likewise, if the series be
continued, that
a 6 = a 5
Xa=lxa = a,
7 = a 5 X a2 = 1 X a2 = a 2 ,
have been determined, affords a reasonable presumption that they exist likewise
in all other cases.
126
ap 1 =
and a" 1 = 0, for the same value of , it follows
that they must have a common factor, which becomes zero for
that value: but if n be a prime number, a 1 is the only common
"
factor of a p 1 and 1, and which can only become zero
(a
3
y=) 3
=(l) =l,
3
() = (!)=!,
it follows that a 2 , a 3 ..... a", satisfy the requisite conditions
of the equation equally with a and since they are all of them :
naryroot^of t
is P eriodic For a" +
I
= a" x a = 1 x a = a,
a" +3 =a n x a3 = 1 x a3 =a 3
may be called its base : and it is obvious that there are as many
bases as there are roots of the equation different from 1.
It will be farther shewn,, in the articles which follow, that if
equation x
pq 1 = 0.
Again, all the terms of this product are different from each
~ -
r
other: for, if not, let a {3 and therefore a r r = /3* t, =a p r
,
we get /3
s
=
-37^
=
TJ~,
= /3- <- ), which is a root of &- 1 = 0, and different
r
f3 a root of x q I =0, whatever be the value of r: and there
fore a
r
j3
r
is, by the last Article, always a root of x
pq 1 = 0.
Again, all the terms of this period are different from each
other.
r- s~ r *-^-^
we thus get a = fi = fi 1
6
The roots of 713. Thus the
roots of the equation # -l=0, are found
1 =
z
by multiplying together the three roots of x and the
2 -
two roots of x 1 =0: they are therefore
-l + x/^x/- "l
-I-JSJ^I I-j3j~l I
The same series of roots, though not in the same order, form
the terms of the period whose base (a /5) is
^-1=0 and x q - 1 =
become identical with each other, and no new values or roots
129
x^- 1 =0,
will be found to be the p th roots of those of the equation
xp -l = 0.
of a, we shall find
that the p" roots of the equation xp2 1 =0
r- ~
r and s are less than p: and since a = /3p(r \ it follows that
s
^o>+i)
o-o_ 1 k u t since r
.
s cannot exceed p 1, it follows
that (p+l)(r-s) cannot exceed (p + l)(p-l), or p~-l, and
(p+1) - s)
therefore /3 cannot be equal to 1:
Cr
it will follow therefore
2 p2
that the terms of the period a/3, (a/3) , . . .
(/3) form all the
roots of the equation x p
1 = 0.
716. Thus the roots of the equation x* 1 =0 are the cube The roots of
*
roots of the roots of the equation x
3
1 = :
they are therefore
the cube roots of
-
2 2
The method
pressions, such as
-
of extracting the cube and higher roots of ex-
-
-
2
-- ,
and -
-- *-
/*
3 - ,
will be
The ro 9 ts
717. prime factors of n, which are dif-
If p, q, r be three
ferent from each other, and if a, /3, j be severally roots of the O ftheequa-
x p
-l=0, ^-1=0, and of- 1=0, which are dif- t n s i
equations a ?
_ 1 = 0> r
ferent from 1, then the period of n terms, whose base is a/3y, and
will form all the roots of the equation .r" 1 = 0.
3
shall find that the period of p terms, whose base is a (3 7, will
equation x
n
1 = 0, when ?i is any composite number whatsoever.
n
General It appears, therefore, that the roots of the equation x 1 =0
I0n *
will form, in all cases, periods of n terms, which are the suc
cessive powers of a common base : and it is this remarkable pro
The roots of The roots of the equation #"+1=0, which are the
yjg.
n roots of 1, are included amongst the roots of the equation
a, a 2, a 3, ... a 2 "- 1
, a 2n ...
For "
= -!, (")= () = (-!) = -!:
Again, all the terms, and therefore the odd terms of the
period
are different from each other : and since they are n in number,
they form all the roots of the equation #"+1 = 0.
The roots of
4
a: +l = 0,
719.
amon g s t
Thus the
- =
1 = are included
-
_
J? 1 = 0. 1 : and if
y^ - be
A/2
one of the roots of x 4 + 1 = 0, all its roots are expressed by
the period
For ,v2-l = -
131
j* ^ ~jr- jr
The series formed by the even powers of the same base
(Art.
1 +v ~ l
- -
710) , which are J~l, 1, J~I\ and 1, are the
v^
roots of the equation x* - I =0. (Art. 695).
720. We have shewn (Arts. 708 and 710), that when the The order
index n a prime number, the same period of n terms
is
f suc e
may -
f
f"
arise from (n
1) different bases, which are the (n 1) first terms of a
likewise be found that the terms of all these periods follow a a c.hange .
of its base.
different order: thus, if in the period
a, a 2, a 3 , a\ a 5 , a 6, 1 (1),
to x 1 - = 0, we 2
corresponding 1
replace the base a by a we ,
a 2, a\ a 6, a, a 3, a 5, 1
(2),
whose terms are the 2nd, 4th, 6th, 1st, 3rd, 5th and 7th terms
of the period from which it is derived : if in the same period
(1) we replace a
3
by a , we shall get the period
a 3
,
a 6, a 2, a 5, a, a 4, 1
(3),
whose terms are the 3rd, 6th, 2nd, 5th, 1st, 4th and 7th of
the same period and it may be observed generally, that if in
:
* For if one term (the p th ) of the series of r th terms coincides with a final
we must have p r = q n or = ,
and since n is prime to r, the fraction cannot
q r r
be reduced to lower terms: and since p and q are the least possible numbers,
which will answer the required conditions, it will follow that p = n and q r: or
in other words, it is the n th term of the new period which coincides with the final
term of the r th period of the periodic series.
Again, all the terms of the period thus produced are different from each other :
a, a 2, a 3 ,
a 4, a 5 ,
a6, 1 (l),
a 2, a 4 , a 5, a, a3 ,
a 5, 1 (2),
a5 ,
a3 , a, a 6, a 4, a 2 ,
1 (3),
x-l
where n is =0,
a prime
X- 1
them less than n, it follows that s t is less than n : but we have shewn above,
a nr .
r
that the least power of a which ,
is equal to 1, is
It may be observed, that the proposition is true of all powers of a base whose
indices are prime to n, whether n be a prime number or not.
133
Q Q2 q3 q4 y q5 06
*/ y il>
y CJ y *-*
<-?} *-*?
are
3, 2, 6, 4, 5, 1*:
and if the terms of the period
a, a i a , a , a , a
corresponding to
= 0, be distributed in the order of these
residuals, as follows,
(2),
(3),
*
These residuals are successively formed by multiplying the preceding resi
dual by the primitive root, or 3: thus, 3 being the first residual, the second is 2,
which is the remainder from dividing 3 x 3 by 7 the third is 3 x 2 or 6 the fourth : :
afterwards recur in the same order for ever. If we take the second primitive root
of 7, which is 5, we shall get the series of residuals 5, 4, 6, 2, 3, 1, the five first
terms of one series being the five first terms of the other series in a reverse order.
If n = 13, there are 4 primitive roots corresponding, which give the following
periods of residuals :
2>
2 and 7 j
\ 7, 10, 5, 9, 11, 12, 6, 3, 8, 4, 2, 1
these periods form pairs, in which the first 11 terms follow severally an inverse
order and it may be observed, that the terms of the second pair are the 5th terms
:
student will find the theory of such primitive roots discussed at consider
The
able length in Art. 531, and those which precede and follow it.
134
lar arrangement with those of the period (2), each term being
3
one place in advance: if we replace a by a or a 2 , we get
a 6 , a 4, a 5 , a, a 3, a 2 (4),
period
a 4, a 5 , a, a 3, a 3, a 6 (5),
Cyclical
Such periods of the imaginary roots of 1, which cor-
723.
periods :
respond prime values of the index, may be properly called
to
their great \ , . . , ,.\3 . . .
*
They form the basis of Gauss well-known researches respecting the geo
metrical division of the circle, and of Lagrange s theory of the solution of the
724. WE
have explained
r in a former Chapter (xxm.) the The use of
the roots
introduction and use of the roots of 1, as the recipients of the O f 1 as the
../,!_
signs of affection, which the application of the general principle ^g^"^
of the "permanence of equivalent forms" (Art. 631) renders ne- affection
, , , . i .
i . i . in Symbo-
cessary in algebraical operations, and more particularly in those }j ca Alge-
]
725. We
have subsequently investigated (Chapter xxiv.) Recapitu-
the more important of these properties, with a view to the dis- theTrsym-
covery of the conditions which must be made the basis of their bohcal pro
perties.
interpretation: we have shewn, when n is a prime number,
symbolical form only, or to the quality of magnitude which its symbolical con
ditions are competent to express: all the values of (l)^r are equal in mag
nitude, but different in affection or quality.
136
Research of
726. If r, therefore, be a specific magnitude to which the
the condi-
their inter-
sl
Sn C1 )"" ^s attached, or into which it is multiplied, forming
,
a n r or r (l),
which are all of them symbolically different from each other.
The sue- 727- It will be observed that the successive terms of this
termslf a P er id bear the same symbolical relation to each other, the ratios
period pos-
sess the -
a2 r
ar
,
a3r
5
a*r
,
a*r
a r
^
j ,
. . .
a
an r
, l,
r
,
samesym-
bohcal re-
lation to
k emfife identical with each other, and with a and it will follow,
each other: therefore, that
r
...
capable of symbolizing, when applied to the specific magnitude
r, must equally symbolize, the affection which connects a 2 r
it
ma Stlfde
^ nasmucn as tne successive terms of this period are symbolically
different from each other, they are also different in their af
fections if we enter upon a second period, the same terms
: but
will re-appear, and in the same order: and so on from period
to period, however far the series of them may be extended.
Their geo- 728. The relative position of equal lines in Geometry, mak-
* n e(l ual angles with each other, and which are the quotients
interpreta-
tion.
Quantities may be of the same kind though different in quality or affection :
t For, if not, let the magnitude of the quantity expressed by ar, considered
without regard to its affection, ber(l + c) then since a 2 r bears the same relation
:
4
r(l + c) ... and that of a"r, ... r(l + c)
a
but a n r = r, and therefore r = r (1 + c)"
: :
or in other words, c = 0.
137
pretation.
Let a circle be described with a radius AB(r), and let its
circumference be supposed to
be divided into n equal parts in
the points 19 B B B ... B _ 2, 3 n 2 ,
B n _t ; the radii AB AB AB 13 2 , 3,
AB n _ 2, AB _ and AB being
n 1}
ABn _, = a x AB _ n z
= an - r,
l
AB =a x AB _ n
= an r or r,
a 2
aV a n-
a n
the line by the last term
ar, r, r, r,
. . . represented
VOL. II. B
138
Choice of 729. We
have assumed, in the preceding investigations, a
S
which cor- to be the appropriate sign of affection, which when multiplied
tKast
** mt ^ B
* T ^^
denote the mag tude and position of 1
m AB
2
with respect to AB, and therefore a to be the
angle of
transfer.
gign for AB ^ ^ for AB ^ and ^ for AB^.
corresponding
but masmuc h as a
may express any one of the (n-l) roots of (1), which are
different from 1 (Art. 708), there is apparently no reason why
one of them should enter into the expression which designates
the relative position of V
with respect to AB
corresponding AB
to the least angle of transfer, in
preference to any other: it
will be found, however, when we are enabled, in the
progress
139
severally
T
make -i i -^
the same angles with the primitive line
^ .*! :
/
for we
the same
determi-
angle BABV
with the primitive line AB. The operation itself is
the transfer of the line AB
through the determinate angle BAB V
:
is r or ; AB
and inasmuch as l
is transferred through 2 right AB
angles in passing from the position l
to AB, corresponding AB
to the multiplication of ar by a, it coincides in position and
biquadratic 1>
~ ^ J~~^ and ~ \/~-^ ( Art 706) : if we make J- 1 (a) (Art.
roots of 1.
7] ^ thebage ^ the period will be J^i _i, -J~l, 1. t
3r
or r jj 1 would have denoted AB l , and o 4 r would have denoted AB or r.
141
by -J-l, will be
tiplication its from AB to AB3
transfer ,
If we draw AB, AB AB
1} Z dividing the circle into three
equal parts, or making angles
BAB^B.AB^ B AB with each Z
therefore, that if EB l be
further considered with re
ference to AB or r, it will be represented both in magnitude
thenuse 19 AB
the three sides being considered in magnitude and
also in position with respect to the primitive line AB: and in
a similar manner, the symbolical sum of and Z , when con AE EB
sidered both in magnitude and position with respect to AB, is the
is equilateral, and the perpendicular J^E bisects the base it follows likewise :
that ^2= AE 2
+ EBi 2
, and therefore EBj = AB, - 2 2
AE = AB - \ AB* = % AB 2
S 2
,
and
The 735. THE theory of angles, their measures, and the peri-
f
Goniorne-
dical ratios which determine them, constitutes the science of
try, Trigo-
Goniometry, whilst the specific application of some of its results
and Poly- to the determination of the sides, angles, and areas of triangles
gonometry, w^id De properly termed Trigonometry, and to rectilineal
with refer-
may possess every magnitude between zero and two right
they
their mag- angles, which are their limiting values for lines which contain :
and from a ^ nes (Deluding those which are in the same straight line), according as the lines
given posi- which form them are drawn in the same or in opposite directions, or according as
tion. they form angles with each other, which are an even or an odd multiple of two
right angles, zero being included amongst the former : this classification of parallel
lines is not recognized in Geometry, where angles cease to exist at the extreme
limits of zero or two right angles.
145
Relation
movement be continued from between
AP to AQ, the arc BP will be arcs and
the angles
increased by PQ, and the an they sub
tend.
gle BAP by QAP: and it
will follow from a well-known
proposition*, that the angles
BAP and BAQ
will bear to
each other the same propor
tion with the arcs and BP BQ, by which they are subtended
in the circumference of the same circle.
quadrant BC
to be divided into 90 equal parts, the right angle
will be divided
by the radii which pass through these points
into 90 equal angles, each of which is called a degree: and
ifthe arc subtending a degree be divided into 60 equal parts, -phe sexa .
each of them will correspond to, or subtend, an angle which gesimal
*
Euclid, Book vr. Prop. 26.
t The sexagesimal division of the circle has prevailed since the time of
Ptolemy or the astronomers who preceded him, (see the article Arithmetic" in the
Encyclopaedia Metropolitana, p. 401), and is so intimately associated with our
Anglesthus 739- It thus appears that the angle generated will increase
* n *^e same proportion with the corresponding arc which is
cjTaSeof
indefinite described or passed over,, by the extremity P of the revolving
radius AP: and consequently if one of them admit of indefinite
angle BAP l
is greater than one right angle and less than two,
bearing to the right angle the BAC
same proportion that the
arc BCPi bears to the quadrantal arc BC: the revolving radius
AP will reach the position Ab, after describing two right angles
BAC and CAb, and it is in this sense that the lines AB and A b
which are drawn in opposite directions, are said to make with
each other an angle equal to two right angles*: if the motion of
AP be continued, it will reach the position AP 2 after describing
two right angles together with the angle bAP 2, or an angle
of the cen- computation adapted to the decimal scale is tending rapidly to replace the sexa-
tesimal
gesimal by the decimal, or rather by the centesimal, division of the degree thus :
"
13 27 42 in the sexagesimal scale is equivalent to 13. 4633 nearly in the
- -
thTdegree.
decimal or to 13 degrees 46 (centesimal) minutes and 33 (centesimal) seconds
nearly.
The French T ie French, simultaneously with the establishment of their Systeme metrique
division of decimale, proposed to divide the quadrant into 100 degrees, the centesimal degree
the qua- i nu m
j nto JQQ tes, and the centesimal minute into 100 seconds, and so on, and
this division was adopted in the Mecanique Celeste of Laplace and other co-
temporary scientific works. The change however from the nonagesimal to the
Causes of centesimal degree, was attended with no advantages sufficient to compensate for
its failure, the great sacrifices of tables and records which its adoption rendered necessary,
and its use was speedily abandoned, even in France. If the proposed change had
been limited to the centesimal division of the nonagesimal
degree, it could hardly
have failed, when the authority of the great men who proposed it is considered,
to have been readily and universally adopted.
* At this the contained AB and the revolving radius
point geometrical angle by
ceases to exist, inasmuch as the lines which contain it are in the same straight
line : no regard is paid in Geometry to the directions of lines when considered
per se.
740. The angles which are thus generated, and which are Goniome-
capable, like other magnitudes, of indefinite increase, may be
termed goniometrical, as distinguished from geometrical, angles : angles.
position of AP
with respect to AB, implying the additional con
dition that this position had been attained in the 1st, 2nd, 3rd,
th
4th, . (M + l)
. . revolution of the generating radius AP.
It may be observed that goniometrical angles, which are less In what
than 180, or two right angles, coincide both in magnitude and coincide*^
in the position which they express, with the corresponding
597. 11 44"- 2
. x 360 = - 122. 47 .
16",
or to the geometrical angle 122. 48 16", or formed below
. BAE
the primitive line AB, or bAB: the goniometrical angle 1104
corresponds to the symbolical angle
- 1104+ 3 x 360 = - 24,
which is equivalent to the geometrical angle 24, or BAF formed
likewise below* AB: and similarly in other cases.
as well as other relative terms of a similar class, such as high and up and down,
low,
right and left, to and from, have generally some reference, not merely to a zero
point or line, but likewise to relations of position, which are real with reference to
the writer or observer thus the terms above and below, are used in their ordinary
:
meaning, when applied to the lines and angles formed in the figure given in the
text, when viewed by a reader in his ordinary position.
"t The relations of magnitudes of the same kind are numerical only, and may
be replaced by commensurable or incommensurable numbers (Art. 165) it is the :
common unit of magnitude expressed by one of these numbers, when they are
commensurable, which is said to measure the magnitudes which they express, a
notion of a measure which is quite distinct from that which is considered in the
text.
151
angles: but if the radius of the circle in which the arcs are
reckoned, is not given or is indeterminate, then a determinate
value of the arc will not correspond to a given or determinate
746. If, however, the arc alone be an indeterminate, the ratio The arc
of the arc to the radius of the circle on which it is taken, will
be a determinate, measure of the corresponding goniometrical of the circle
angle, whatever that radius may be*: it is this ratio which is j s taken is
a
assumed as the measure of angles in the articles which follow. ,
sure of the
corre-
sP ndin s
747. The and
ratio of the arc to the radius is numerical,
. . angle.
if its value can be determined for the whole or any definite If this ratio
*
For if the arcs EC and bc subtend the same angle A in the circles whose
radii are AB and Ab respectively,
then by a well known property of
EC be
the circle, we have = : and
AB Ab
if we take the angle BAD in the
circle whose radius is AB and the
angle bAc in the circle whose radius
is A b, then we find
L BAC: L BAD:: EC ED
:
EC ED bc^
ED
:: ::
AB AB Ab AB
and since these angles increase in
15 ri
the same proportion with the corresponding ratios, it follows that or its
be
equivalent -77- is the proper and determinate measure of the corresponding
A I)
angle.
152
The ratio Thus, the ratio of the whole circumference to the diameter,
cumference though they are not commensurable, has been approximately
determined by various methods, some of which will be con-
pressed by sidered hereafter. This ratio, which very frequently presents
itself in analytical formulae, is approximately expressed by the
number 3 .
14159, and
also generally represented by the symbol
is
2
Unity is it be
If required to determine the angle whose measure is 1,
the mea- or ^ n other words, the arc whose length is equal to the radius,
|
sure-sanf"
an gl es an d their measures t, by which they may be converted
very often 22
designated Archimedes, in his Ku/cXou Me-rjorjcris, assigned j-
as an approximate value
6
^ 7r> w ^ cn differs in excess from its true value by less than ^o tn P art f tne
symbol^
and the diameter it is this value, which
:
generally used by workmen, in their estimation
is
same deno- o f c i rcu i ar work. Peter
Metius, by a similar process, found the remarkable ratio
mmations. 355
jYg 5 which is correct to 5 places of decimals, and which may be very easily remem
bered, by observing that if we write each of the three first odd digits twice in suc
cession, as in the number 113355, the three last digits form its numerator and
the three first its denominator. Later researches have assigned its value correctly
to 208 places of decimals, a prodigious approximation, which is effected by pro
cesses which are extremely simple and expeditious, and well calculated to shew
how much the most complicated calculations may be shortened by a judicious
selection of formulae. See Phil. Trans. 1841, p. 281.
unity, and its centesimal subdivisions grades, minutes, seconds, we should have
the French division of the quadrant, or of the right angle, with its scale of units
increased one hundred-fold.
There is no geometrical mode of determining an absolute unit of any other
magnitude. Thus a linear unit must be sought for in some assumed standard,
such
153
promptly and easily into each other, yet in the case of certain
periodical ratios, which will be considered in the next
Chapter,
which are equally determined by them, it is usual to
represent
them by common symbols, and to call them by common denomi
nations :
thus, is considered equally as the representative of
^
a right angle, andits measure: and
similarly 6 or (p (which are
symbols very generally employed for such purposes) and any
other symbols are applied indifferently to designate both
angles
or their measures.
if or , (Art.
VOL. II. u
154
geometrical of one line with respect to another, we may add to, or subtract
from, its measure, any multiple of 2 TT, or of the measure of
4 right angles : this conclusion follows immediately from Art.
742.
2w + and
-
6 corresponds to the same geometrical
TT 6 : =t
( 1)
3 IT ( 4 n + 3) TT
Great im- 753. The propositions in this and the following Chapter
Fhe fun<la will be found to constitute the grammar, as it were, of the Ian-
mental pro-
m
rr e o f
ffua &
& Trigonometry, and will be made the foundation of a
positions
Trigono- very extensive analytical theory, admitting of the most varied
" try "
*
The whole theory of music is dependent upon the properties and conversions
of a very small number of numerical ratios :
yet how vast and complicated is
the superstructure which is raised upon them !It should be the first lesson
of a student, in every branch of science, not to form his own estimate of the
importance of elementary views and propositions, which are very frequently
repulsive or uninteresting, and such as cannot be thoroughly mastered and re
membered without a great sacrifice of time and labour.
CHAPTER XXVII.
drawpm perpendicular to AM
or to ,4Mpro-
755. If we form the right angle MAm, drawing Pm per- The sine of
the definitions of the sine and cosine which are given in the last
Article, that -j-= is the sine, and -j= is the cosine of the angle
PAm, or
g-0:
but since Pm=AM and PM = Am, it follows
that
_
~ 3 i --_
and therefore the cosine of the angle -- is the swe of the angle
by the abbreviated expressions sin and cos 6*: and the pro- the sine and
cosine of e -
position in the last Article would, therefore, be symbolically
expressed by the equations
and cosine did not determine the angle when the radius was not given, and no
trigonometrical formula was complete unless the radius (when not assumed to
be 1 ) was introduced as one of its elements the definition in the text, which
:
Expres- 7^7- Again, it follows, from the definitions given in Art. 754,
sides of a
r
h6that if ABC be an y right-angled triangle,
right-an- we shall find, since
gled trian
gle in terms ,, BC
BAC =-177
. ~ AB
_
. .
Geometry.
mental"
^ Inasmucn as sin =
~Tp
( Fi g- Art 754), and COS 6 =
relation of
the sine and
AM
__ jt follows that
cosine. AP
Sin
2 a
+ COS 2an = 1
/\
sine in (3),
coTne^and
c^ ennes tne fundamental relation between sin 6 and cos 6, and
conversely, enables us to express one of them in terms of the other: we
thus get
sin 6 = 7(1- cos 0)
2
(4),
2
cosO-^l-sin ^) (5).
every value of cos 6 there are two values of sin 0, and conversely,
which differ from each other in their signs
only, one being posi
tiveand the other negative.
Detailed
examina- 760. The course and succession of these changes, however,
changes, will appear much more distinctly, from tracing them in detail
159
to 1 : for sin d = -
(Art.
0=
PB is
754) is when -^ 0:
At>
and when 6 is 90, then PM
becomes CA, and
= 0.
AB
In the second quadrant, the values of sin0 decrease from
1 to 0: for if 6
represent the angle BAQ, then sin0 is 1, when
Q is at C, and when Q
is at b, where Q?n is 0: the correspond
cose = cosBAQ = ^,
which is at C, and equal to 1 at b } where Am coincides with
Ab: for since AM and Am are drawn in opposite directions from
the
A, it follows that if
-^ be positive, then -^ is negative :
tract the same number repeatedly from another, as 1 from 4, we get a series of
remainders
3, 2, 1, 0, -1, -2, -3,
which
160
BAp is still
negative, the sine is also negative, and diminishes
The sine 7^1. The same position of one line with respect to another,
and cosine or primitive line, and therefore the same geometrical angle 0,
trieal an-
(whether positive or negative, less or greater than 90), and con-
V
sequently the same sine and cosine* will correspond to the series
differ by
multiples of o f 2 ?r + 0, 4 ?r + 0, 2 w TT + 0, which
goniometrical angles Q, . . .
the same as differ from each other by multiples of 2 TT only (Art. 752) and :
6} (8).
which increase both ways from : but it is sometimes very incorrectly said that
1 is less than 0, forgetting that the operation called subtraction changes its
meaning (Art. 555) when the subtrahend is greater than the minuend thus :
therefore, which correspond to the same geometrical angle, have the same sine
and cosine.
1G1
angles, but with different signs (Art. 74-1): and their sines
PM and
.
pM
*- m
^
At) At)
are equal in magnitude, but different sign:
763. The sine of an angle is identical with the sine of its The rela-
8
(Art. 751), both in magnitude and sign:
supplement but the^^J
cosine of an angle is equal to the cosine of its in cosines of
supplement
angles and
magnitude, but differs from it sign. m their sup-
For, if the angle BAQ (Fig. in Art. 760), be the supplement plementa.
spect to the line BAb : but the cosine of the angle BAP or -^
is equal in magnitude to the cosine of its supplement BAQ.,
which is
-jj , but differs from it in sign, since AM and Am
are drawn in opposite directions from A.
The symbolical enunciation of this proposition is
VOL. II. x
162
cos (u) -
H--""
For ^ + is the supplement of ^ - 0, and therefore
from the
The same conclusion may be deduced geometrically
figure given in Art. 760.
of all goniometrical
The rel a- 765. Again, since the sines and cosines
S the
sine and angles, which differ only by multiples of STT or 360, are identical
C Sil e with each other (Art. 761), it follows that
J and0
sin (TT + 0)
=- sin (TT
-
0)
=- sin (15),
= cos 0.
other
Other simi- The sameprinciple will enable us to convert many
expressions of a
similar kind into others which are equivalent
to them.
Thus, sin
(^ -0V- cos (17).
\ * /
For sin [ -? - )
= sin (-? - - STT )
= sin - f^ -}
Similarly,
we find
(18),
B1 n (20).
163
= sin ( ? =
sin
(^ + 0\ + 0\ cos 0,
cos (2 TT + (?)
= cos 6,
cos f ~+ J
= cos f
^+ J
=- sin 0,
COS (3 TT + 0)
= COS (TT + 0)
=- COS 0,
cos
^ + 0\ = cos
(ir
+
0)
= sin -
766. The general methods which are adopted for deter- Angles
mining and registering in tables the numerical values of the ^nd cos ne s i
of45orj.
45 =
3 2
sin 45 + cos 1 (Art. 758),
we get 2 sin
2
45= 1, and therefore sin 45 =- = cos 45: the
v~
positive sign is taken, since 45 is less than - (Art. 7^0).
and if DE
be drawn perpendicular to AB, it
will bisect it in E: it follows therefore that
AF
cos BAD = cos 60 = - = = sin 30 :
cos 60
2
+ sin
2
60 = 1,
we 2
cos 60 1
get, by replacing by ,
2
or sin 60=f,
The sine
768. The sine of 18 is and the cosine of 18 is
and cosine
of 18 or
7T ^(10 + 2^5)
10 4
therefore that
BAG and also the chord EC: the angle BAD is therefore 18,
~
it follows that sin I8
9
=" 5 l
, and
769. It will readily follow from the three last Articles, and
1. Sin 135 =
-^:
v^
2. Sin 225 = sin 315 = - -i .
v*
3. Cos 135 = - = cos 225.
-^
4. Cos 315 = -?-.
5. Sin 120 =.
6. Sin 210 = sin 330 = - .
8. Cos
330=^.
9- Sin
150=^.
/3
1 0. Sin 240 = sin 300 = - ^- .
16.
4
and
ence.
differ
Let the angle BAG = 6, CAD = 6 , and therefore BAD = 6 + 6 :
and also
and therefore
DE = BC+DF= AD sin cos + AD cos sin 6 ,
and
AE = AB - CF = AD cos cos - AD sin sin .
We thus find
=
DF = + cos sin
sin (0 + ) -jjtr
sin cos (a),
and
and also
and AB=ACcosO;
16?
shall get
and therefore
DE = BC - CF - AD sin cos d - AD cos sin ,
and
^4 = .4 B + DF = ^1D cos cos + AD sin sin V.
We thus find
- =
DF = sin -
sin (0 )
.
^
AJJ
cos cos sin 6 (c),
and
A ZT1
cos (0 - )
= ---= = cos cos + sin sin (</).
771. Of
the four formulae (), (6), (c) and (rf),
which are Three of
given in the preceding Article, the three last may be easily mu i ffi ; n the
=sn ~
and
168
an equation which has been shewn to be correct when 6, 6 and 6+6 are
less than - : the formula (a) assumed, therefore, is correct in this case, since
we
^+ ^+ and assume the formula (a)
Again, if replace 6 by 6, 6 by 6 ,
sin O+ 6 + 6 ) = sin
(^
+ 6\ cos
(^ + 6 \ + cos
^ + 0) cos
(^
+ e
)
sin (6 + )
= cos sin 6 sin 6 cos :
sin (0 + )
= sin cos + cos sin (a),
a correct equation.
+ ,
and assume its correctness for this case also, we shall get
= sin + + cos + (~ +
sin I + + 0) cos
j+ 0) sin
(
(TT (TT :
-^ J j J
- +
-^ +
if we now replace sin f + by cos (0 6 ) (Art. 765), sin
(TT + 0)
J
cos (0 + )
= sin sin cos cos ,
cos (0 + )
= cos cos - sin sin (b),
a correct equation.
It is obvious that AVC may apply the same process to the verification of the
fundamental formula sin (0 + ) = sin cos + cos sin , for all values of 0,
,
and + whatsoever, including those in which or become equal to
tigation.
169
and therefore
sin (0 - 6 = sin cos - sin cos
) (b).
sin
(|
- -
)
= sin
(^
- 0] cos - cos
(-
- A sin & :
we shall get
) (d),
- and -
observing that cos is replaced by cos , sin by
- sin .
sin (0 + )
= sin cos - cos sin (b),
sin (0 + )
+ sin (0 - ) = 2 sin cos (e),
cos (0 - )
= cos cos + sin sin (d),
VOL. II.
170
we get
sin + sin = sin (s + + sin (s - d) = 2 sin s cos d,
d)
cos - cos r
= cos (s + d)
- cos (,?
- d)
= - -2 sin s sin d :
p i
/y p A
and if we replace s by and d by , we find
/0 + \ -
sm + sin = 2 sin -
cos
(0
I
0\
I
(z),
J J
r
/0 4. \
cos - cos r
=-2 sin (
- sin
J
cos (0 + )
+ cos (0 - = 2 cos cos (g),
)
other, from the left-hand side of the equation to the right, they
will assume the form
sin (0 + =2 sin cos - sin (0 -
) ) (),
cos (0 + )
=2 cos cos - cos (0 - ) (o).
to.
be given, we
can find by these formulae the sine of double the
angle and also, if either the sine or cosine of an angle be given,
:
2 4
cos 36 = 2 cos
2
18- 1=1-2 sin 2 18 9
Given the
sine or co
sine of an
776. The same formulae will likewise enable us to express &
*n es to .
Inasmuch as
S
sin =2
.
sin - cos - ,
and
1 = sin
2 - + cos 2 - (Art. 758)
we get, by adding
/l f\ f\ /j
= - +2 - cos cos -
2 2
1 + sin sin sin -t- ,
and, by subtracting
f\ n
If sin- be greater than cos-, the terms of the complete
srafl = sin-
in + C os-
S
2 2
e e
- sin 0) = sin
.
- - cos -;
J(\
+ sin ~ - sn
the squares
e e e
9
- +2 - cos - + cos2 - , and sm 2,e- - 2 - cos - + cos"2 -
,
2 8
and therefore
cos - =
i J(l + sin 0) + 4 7(1 - sin 0) (3),
f cos 0.
Since
n a
cos = 2 cos
2
--1 = 1-2 sin
2 - ,
s\
2 sin -
2
= 1 - cos 0,
nd therefore
cos- =
the limits 2-rr and 4?r, GTT and STT, 10 TT and 12 TT, then sin -
1 - cos
25899,
-^2
cos 15 = A ^(1 + sin 30) + i V(l
- sin 30)
-96609.
5 + ^5
(2) Given the cosine of 1 8 = /( = .897, to find the
J
sine and cosine of 9.
.9729-
*
Or, in other words, to find all the values of 0, which satisfy the equation
sin = a, where a is any positive or negative number less than 1.
175
It will follow, therefore, that all the angles in the two series
In the first place, all angles are equicosinal or have the same equicos
780. The propositions in the two last Articles deserve the Great im-
most careful consideration of the student, as furnishing the tne theory
The periods 78!. Thus the angles in the series 2wr + are equisinal with
of the sines th
b
B: but the sines of a series of angles, which are - part of those
muhfples of
equisinal
angles. in the series 2w7r+ 0, form periodical series, the sines of the first
m terms being different from each other, and afterwards re
curring periodically in the same order.
Thus, the equisinal series being
and the
m ,
27T
mm-
sines of the
+-
,
4-7T
m
+
first
,
. -
2(w-l)7T +
m
terms are different from each other
, A if H--
m j
n a
th
but the + term is 2-7M which is
equisinal with the
-
(7H l) ,
h - which ----
(m + 2Y term is 2-7T+ is equisinal with
m ,
m ,
consisting of m terms.
Again, if we
take the second series of equisinal angles, in
cluded in the formula (2w + l)w 0, which is
7T-.0
m
STT-d
m ,
-
the corresponding series of their submultiples by
-- 57T-0
m
...
(2m-l)7T-0
m , ,4
m
TT H
will be
7T
m
-
. . .
and the sines of the m first terms are different from each other ;
th
that of the (rw+l) term being equisinal with the first, of the
th
+
(m 2) with the second, and so on for ever. The sines there
fore of the second series of submultiple angles or measures of
/27r 0\ Cf2m-lW 6\
j(2m-l)7r 0) /47T
/4-Tr 0\ f2(wi- 2)
cos -)=:cosiI m ^-=F->, cosf -)=cos^-^
I
\w mj m] \m mj [
m
J,
thus reducing the number of different values of the cosi
cosine
signs only.
The Example!.
783. following are examples:
a
(1) Given sin 0, to find sin - .
0, 27T + 0, 47T + 0,
VOL. II. z
The corresponding series of semi-angles are
e e e
, 7M-- , 27T + -, ......
sin-, smf 7r
-J,.and
0\ /37T
~ \^ ~/J <w
The corresponding
*g,
006 series of semi-angles is
*j,
2vr -, ......
cos-, c
of which cos
00 = cos - and cos (
/
TT
6\ = -
-
J
cos - .
TT STT 5-7T 77
3~3~ !T~3 ~3~ ~3 3
_, sm-+-, sm-
and
Sir
-8* - &
- **
~ +- ..v 9
but since - of - - -
. ,
is the supplement of - , , and
3 o o o o o o
- + - of - -
rejecting 2-Tr),
it follows that the values
o i> o o (after
of the sines of the first period are identical with those of the
second: and there are consequently only three such values,
which are different from each other*.
a
TT Sir
-
= 2 cos
. 20 - .0
sin sm sin
j -g- -^
= 4 cos 2
00- -
- sin sin - /
I
replacing sin
20
by 2 cos
0\
- sm - .
3 sin
.0 -
4 sm d3 ,
/ 6\ -
/2jr_^
"~ ai in
a cubic equation, whose roots are sin 3 sin
3j
("3 \ 3 3
-,
a /TT 0\ /V 0\ /
cos-,
cos(^--J, cos(^-f-J, cos^r--
which alone are different from each other, and of which the
fourth and third only differ from the first and second in their
sign*.
Caution re- 784. It should be observed, that the definitions of the sine
an ^ cosine apply to all geometrical angles, and therefore to
theapjfli-
cation of those which are negative as well as positive, extending thus
pie of equi- beyond the limits of Arithmetical Algebra: and whatever be
1
t ^le P ro P os iti ns J which come within the range of these defi
forms
nitions, they must be established by means of them, and not by
the generalizations of Symbolical Algebra: and though it is by
the aid of Symbolical Algebra alone that we are enabled to
form and express negative lines, negative arcs, and negative
angles, yet as soon as we have given to them a consistent inter
pretation, they come within the proper province of Geometry,
and are subject to the same definitions and propositions, as apply
to the geometrical quantities which they represent in affection
as well as in magnitude: it is for this reason that in demon
Algebra.
I
n f\
* = -
The corresponding biquadratic equation is cos 8 cos4 8 cos 2 -7 + 1.
CHAPTER XXVIII.
785. THE ratio of the sine to the cosine of an angle is called Definition
its tangent : and the reciprocal ratio of the cosine to the sine
of an angle is called its g ent
cotangent*.
*
786. The tangent and cotangent of an angle determine the They de-
* The tangent and cotangent were formerly defined as lines, and not as
ratios :
thus, if J3P be the arc of a circle sub
tending the angle at the centre, and if BT
and C t be drawn from B and from the extremity
C of the quadrant, tangents to the circle meet
ing AP produced in Tand t respectively, then
BT was called the tangent, and Ct the cotangent
of the angle BAP or 0.
182
r or tan
.
=
sin
and, cot = -:
cos 6
. ,
jvc
and the first ratio
s is
i,
the
cos sin
_, sin - - sin
For tan - = = -
5-= - tan
.,
-
a 0, and
cos cos
sm
Ar
For tan - - (
A= "-OH
H2 V =.cos0 =
. -. cot 0:
\2 /
)
/TT A sm
si
cos
(H~
and cotfg-flU
\2
f \
/
COS
(HA= -
v
/TT
;"
v j _._
-
-^
cos0
n
= tan0.
sm
(-)
Tan (TT - 0) = - tan and cot - =- cot 0.
(4) (TT 0)
sin CTT 0} sin ,
For tan - 0) = \ -^
= -=-tan0: and
(TT
cos - - cos
(TT 0)
COt
m -r
- 0) COS (TT 0)
=
COS
. = - COt 0.
(TT } ) - sf .
sm (TT 0) sm
By a similar process it may be shewn that
+ 0\ = - - tan 0.
(7) Tan (?-~- cot and cot
(^ + 0\=
(8) Tan (TT + 0)
= tan and cot (TT
+ 0)
= cot 0.
-
J(l + tan 6) =
2
(10) and N/(l + cot 0)
2
= -,_ *.
For 1 + t-n = 1 +
^
sin
2
B
= cos
-^
3
+ sin 6
-,
2
= _ I
and there-
1 + Cot 6 = 1 +
cos d
.
sin- d
. ,
sin 6
= ---:-
2
+
sin- ^
r-
cos
2
=
sm* 6
I
1
and therefore J(l + cot a 0)=-r
sin0
Reciprocally, we find
1 tan
C Sd=: "~ andsm0n =i
tan2 0)
(11)
^
Tan0 + cot0=-
01
sin 20
2 2
sin cos sin + cos 1
-P
For tan + cot = -, +
cos sin cos sin \ sin 2 1
2
t.
788. The tangent and cotangent may have every possible Changes of
value between zero and infinity, whether positive or negative,
vajie^thc
tangent and
-
. TT
S111 cotan s ent -
sinO TT 2 1
For tanO= =_ = 0: and tan-= = -=oo, or mfi-
.
cosO 1 2 TT
cos-
2
t = 2 cosec 26 (Art. 798), and the proposition in the text may be put
sin2tf
under the following form, tan + cot =s 2 cosec 20.
184
and differ
ence.
_ sin (0 =fc 6
_ sinfl cos0 / =fccos0 sin
fi fi
)
~ )
" :
cot (0*P) cos0cos0=Fsin0siiT0
sin sin
7
cos0 coslJ tan tan 6
"
sin sin 1 =F tan tan
cos cos 6
In a similar manner it
may be shewn that
cot0 cot0 =F 1
cot (0 6 ) =
cot 6 cot
*
The signs of geometrical angles change in passing through zero and two
right angles : the signs of the sines and cosines change in passing through zero
only.
185
tan 20a = - --
2tan0
~ tan^.
Q
tan20in
terms of
tan 0, and
"
If we solve this equation with respect to tan Q, we get verse
tan e =- cot 2 J(l +
2
cot 2 0).
and the tangents of all the angles in the first series are identical
with tan 0, and in the second series with cot and it will -
be found that
tan = - cot 20 + 7(1 + 2
cot 20)*
- cot 0=- cot 20 -^/(l +cot 2 20).
(1)
(2)
= cosec20-cot20,
tan0
cot = cosec 20 + cot 20.
l-V3tan0
for tan 60 = V3 and tan 120 = - V3.
VOL. II. AA
186
(*) "
cos sin
sin + sin 6 + 0^
= tan (d
(6)
cos + cos d \FV
sin 0- sin 6 /$ -6\
__
~~
(7) tal1
cosd + cos6 \r2~J
. tan
sin Q
.
,
+ sin
(8) ~;
sin sin
tan
Definition
793. The reciprocal of the cosine of an angle is called its
cant and" secant, and the reciprocal of the sine of an angle is called its
cosecant.
CQsecanL
Their
794. The changes of sign of the secant and cosecant are those
changes of
sign. of the cosine and sine respectively. Likewise the secant of m-
* The secant and cosecant were formerly defined as lines, and not as ratios :
thus, if BP (Figure in Note, Art. 785), be the arc of a circle subtending the
angle at the centre, and if BT and Ct be respectively drawn from B and from the
extremity C, of the quadrant BPC, tangents to the circle, then AT was called
the secant, and At the cosecant of the angle BAP or 0.
187
795. The relations of the tangent and cotangent with the The rela-
secant and cosecant are expressed by the equations (Art. 787,, [JJUgent ami
No. 10), cotangent
tan 0) = sec0,
+ 2
cot 6) = cosec 0.
796. The versed sine of an angle is the abbreviated expres- The versed
_ . sine of an
sion tor 1 - cos 6 it is written vers 6.
:
. .
angle.
If BAP be an angle, and if PM be drawn perpendicular
AB BM
to (Fig. in Note, Art. 785), then the ratio r-p
is the versed
AJj
sine of the angle BAP *.
It never changes its sign, and the limits of its values are
To find the 798. As the basis of our enquiries, we shall begin with the
sine and calculation of the numerical values of the sine and cosine of 1 .
30
sin = i 7(1 + J-)
- i 7(1
- i)
= .258819 = *
189
But
30
2
17
30
=~
x
~~^~
60 x I
7
~~
15 x 15 x 1
=
225
256
X li and 1
sin
^
. x 1 =
^ sin 1 = .000255625,
and therefore
sin 1 = -
256
x .000255625 = .0002908882 nearly.
gives us
cos 1 = 9999999577-
.
sin(rc+ 1)
=2 cos sinw0-sin(w- 1)0,
when is very small, cos0 is very nearly equal to 1, and therefore the
where,
sine is almost exactly doubled when the angle is doubled, and conversely:
QAO
sin = .0010224959,
^-
ono
sin
^ =.00051 12482,
sin = . 0002556254,
|
from whence it will be seen, for very small angles, how nearly the sines are
sin 2 =2 cos I
7
sin 1 = .0005817764,
sin 3 =2 cos I
7
sin 2 - sin 1 = .0008726645,
sin 4 = 2 cos l sin 3 - sin 2 = .001 1635526,
sin 5 =2 cos 1 sin 4 - sin 3 = .0014544406,
Not neces- In this manner we might proceed as far as the sine and cosine
P roceec^ n g beyond which we shall find the same values of
"
13 ^ 45 *n
ceed be
yond 45. sines and cosines which have been previously determined for inas :
much as sin (45 + 0) = cos (45 - 0) and cos (45 + 0) = sin (45 - 0),
it will follow that the sines of the angles in the series ascending
from 45 will be severally equal to the cosines of the complements
of angles in the descending series which have been already
determined, and conversely it thus becomes necessary to construct
:
cos 5
2
= .999997884601 1 1 872929,
sin
2
5 = .000002 11 539745892836,
and
2
cos 5 + sin 5
2
= .99999999999857765765,
a number which differs from 1 by a quantity less than
.0000000000015,
a discrepancy which is referrible to the influence of terms in
the calculated values of cos 5 and sin 5 , which are necessarily
th
omitted, as being beyond the 10 place of decimals, to which
the registered values are limited.
* We thus get
sec =i + + cot f 45 + .
|tan ^45 -) -)}
Formulae of 805. Formulae of verification are equations between the sines
tion. an(l cosines, tangents and cotangents of different angles, which
if satisfied by their values, as given in the canon, or not, would
verify their correctness, or the contrary.
Such are the equations:
(1) Sin = sm(60 +0)-sin(60 -0).
(A r ,79i )} .
* These equations are easily verified by developing the sines and cosines
of the sums or differences of the angles which they involve, and substituting
the numerical values of the sines and cosines of 30 and 18, or their multiples.
VOL. II. BB
CHATTER XXX.
we shall find
(cos (p
+a sin 0) (cos +a sin 0)
= cos cos 6 +a 2
cos cos 6 +a (cos sin 6 + sin </>
cos 6).
2
If we replace a by J 1, and therefore a by 1, this equa
tion becomes
(cos </>
+ J 1 sin
0) (cos 6 + J 1 sin 6) = cos cos 6 sin sin 6
(Art 770.)
If we make (p
= ft, we get
- cos 40 + ^ 1 sin 4 0.
J- m J
.
i /
cos 1 sin I sin
m
-f
v m (
\ in)
)
:
and since
/ m
, 0\
cos +
,
J- 1 sin = cos - + J- 1
.
sin -
(
\ m in)
;
and therefore
1
fi f)
(cos + J- \ sin 0)
" = cos + J- I sin ,
it follows that
= (cos + J 1 sin 0) m .
Also, since
and therefore
n *
"
(cos + */~l sin 0)
" = (cos
- J- 1 sin 0) .
we replace 6 by
- 6, we get
But cos
m
4- J-
v 1 sin
m
= cos (0 + J- 1 sin 0)
"
"
= cos (0-^-1 sin 6) :
n6 , nd .
J-
.
consequently
cos
m J~^- I sin
m
= (cos 6 =F J- 1 sin 0)" .
cos + J- I sin = a6
cos J 1 sin0 = ff-0,
* The demonstration in the text is dependent upon the properties of indices,
by the aid of which we are enabled to shew that
are in every respect convertible into each other general properties : and as the
of indices are referable for their authority to the "principle of the permanence
of equivalent forms," (Art. 631), so likewise must the formula under consider
ation be ultimately referrible to the same principle for its establishment but :
the properties of indices being once admitted as algebraical truths, we refer to them
as furnishing the immediate authority for other truths deducible by means of
them, and not to the fundamental principles upon which they rest in common.
197
a -a
~
cos n d J 1 sin n 6 = a n&
we should find, in like manner,
cos jl0 = ,
and
_ - nQ
an 9 a
sin n 6 7 ^= .
2J-1
808. The exponentialexpressions for the cosine and sine
of They
an angle which are given in the last Article, will be found to fund
0,
satisfy
J the fundamental equations for those quantities, whatever **le<i\ia.-
tor the tions
be the value of a : for .-ine and
cosine for
all values
of a.
4 4
and also
cos 6 = - - = cos 0,
a -e_ a e
sin - = ==- = - sin 0, Art. 7<
809. It should be observed that the definition of the sine The value
metrical angle, and therefore will remain the same whether tne metisu re of
measure of the goniometrical angle be the ratio of the arc to the gyj^J.
radius, or of the arc to the diameter, or any other ratio whatever nary
im -n-
and it is for this reason, that if the value of B, for a given geo-
metrical angle, be indeterminate, the value of a will be indeter- e vrr >
198
angles, the ratio of the arc to the radius, as adopted in Art. 746,
the analytical value of a (for in no case does it possess an arith
metical value) will be shewn in a subsequent Chapter to be
determinate and equal to e^~ where = 2. 7182818 is the base
-
}
,
e
of Napierian logarithms :
replacing a by we shall find
this value,
cos 6 = , sin 6 =
There are
n different
values of
810. If in the equation
cos n 6 + ,J
_ I sin ?id = (cos 6 +
_ 1 sin 0)
n
*J
give the the values of nQ are to be determined from the values of its
but
ferent
dif-
different values of
COS + J ~l
_ sill 6,
values of
cos6jt
and which equally satisfy the equation.
V-lsin0. p OY} inasmuch as the terms of the two series
and no others, are cquisinal, (Art. 780), and those of the two
series
cos n +J 1 . sin n 0,
and no others, which give different values of cos 6 and sin 0, one
or both of them, and therefore different values of
cos 6 + x/- 1 sin 0,
but which give the same values to both the terms of the ex
pression
cos nd + J I sin nd:
(cos 6 + x/ 1 sin 0)
n
= cos n 6 + J I sin n 0.
(1)
cos 6 + ,J~^l sin0,
(2) c
/ 2 CM
- 2) TT ) 2 (w - 2) TT
(
- 1) cos
I -^ - +
n
I
+ V-
/
1 sm /
{
-
^ -+ ,)
}
,
expression
cos nQ + tj- 1 sin nd
it follows that
cos
n
/2r?r
r-rr \ /2r7T , \ -
= cos -
0)-J-lsm
v
- -0);
.
w / \ /
200
(
cos + J- 1 sin )
= 1,
\ n n )
and therefore
* B -1 =0.
The cube 812. Thus, let it be required to express the cube roots of 1.
(2)
201
p 2 IT
(4) Cos "*/- * sm
^~
=cos
-J--\/-
* sm -T- = -V - L
(2) Cos + J^l sin = cos 72 + J^l sin 72=.309 + J^l x .95 1 .
\ / p? \i C *
+
(4)
Cos^ x
/-,sin^^cos^ +N/-^-n^
=- r
25. 43 - .893 + J ~l x .450.
cos 25. 43 + J~l sin
_ 8 7T 8 7T 7T . 7T
-
, .
cos - - J-
J-
/ /
(5) Cos - + 1 sm
.
- =- ,
1 sin ...
~-.89S-.J~l x.450.
VOL. II. C c
202
47r
-
- , 4-7T 3-7T /
- . STT "
(6) Cos
y J- ismy "COSy-V-lsm-y
= - 222 -
.
,l x .975.
(7)
Cos^-y^lsin y = ................ 628-7^1 x .782.
2 /* 7T
+
\
+ //-
/O Y* rjr
-+
\ \ n
0H =cosn6 + J-
. _
{/ cos/
0J
1 sinf l sin 7*0,
--+ lW + J -
-
;f cos ^(2r
i
1
(2r+l)
sin -
.
- J
and therefore
cos - K J
in which r may be 0, 1, 2 ...( 1) successively, giving n re
sults and no more: these are the n roots of 1, or the n roots
of the equation
xn + 1 = 0.
He n roots 817- The n roots of-1 are the 2nd, 4th, 6th and 2w th terms
of ~ a e
f of the period of 2w roots of 1, arranged in their order, or the
J.
2(w-lW + J - ,
1
.
sin
2(w
-
v
203
or
cos - 4- J 1 sm - , cos
n n n
818. Let
it be
required to express the complete values of To express
a n and - n
considered as derived respectively from a and - a, p ietevahies
( a) ,
(-ay=(-iy
i
p
= {cos(2r + l) rnr + J^l. sin(2r + I) 1111} p,
0, 1, 2,...p-l:
beyond these limits, whether the series be continued backwards
n m
or forwards, the same values of a or ( a) recur.
If n be an irrational number
(Art. 245), every different value
of in the ascending series, will give a different value of ()"
r,
If we make cos 6 = .,
-
, and
204
we get
and
cos - =l sin =
fl
2
J(a +b )
a
{cos0 + J^~l sin 6} =a+
and
J(a + b
2
) {cos
- 7- 1 sin 0} =a- b
J- 1.
Consequently, if we make (a
a
+ 6
s
)
a = p, we get
= cos
(a
-6N i)
B
= {cos w (2r w+ 0)
- ^/l sin w (2r7r + 0)} p.
If we suppose 6
= and therefore = 0, these formulae dege
nerate into those which are given in the last Article.
The ex- 820. The conclusions, which are given in the preceding
S 1 1
Articles, are of very great importance, and deserve the most
cos e +
*fi sin careful attention of the student.
may m all j n t]le fi rst p] ace ^ j n expressions such as
for, inasmuch as
.
a - b J
_
In the second place
- 1 are equal
it
to
has been shewn that
J (a
2
+ 2
b ) (cos + J
_ a+bj-l
.
- 1 sin 6)
and \A> 2 + & 2 )
and mon
is the com-
ar.th-
^J(a
2
+b )(cosd- J^l
2
and consequently that
/
sinO) respectively, value
ot_
a + bj^l and a-bj^l respectively express the same arith- +6v - ]
n
2 2
metical or geometrical magnitude x/( +6 ), affected by the s\ga _ hij
cos 6
_
+J - 1 sin 6
r
\ where cos 6
a
= -77 ^ 7^ and sin 6 =
b
r^
-JT-^z + b
\ which dif-
2 } fer in
af-
a [
?
1
,J(a*+J?) J(a )] fection O ni y
r7r + d) + f
j^l sin n (Zr-rr + 6)}
and similarly
xA a= /2r*7r + 0\
J- 1)3 + (a -b ,J-
/ 7 /
(a + b l)s %p cos (
)
\ J /
2
where p is the arithmetical value of (a* + b )^ , and r is any
term in the series 0, 1, 2, 3, ...
and
a
03 r
where is greater than ,
we should find, making
~U 27 4/ ~W/ VA3
-^i.^-CiV- ~
and cos =
T
and it will be observed that if be less than - , there is no term in the formula
|-
=
which involves J~l :
or, in other words, b 0.
CHAPTER XXXI.
are expressed by the n values (and there are no more, Art. 811)
of the formula
cos - r-7T
-f J
/
-
1
.
sin - ,
n n
inclusive.
824. In the application of the successive terms of the period The base a
" correspond-
1, a, a 2 ,,., a 71 1
, ing to the
least angle
formed by the n roots of 1, as signs of affection to denote the
of transfer,
in the
uncertainty however will no longer be found to prevail
208
a = cos
27T
J/
+v
- sm 1
. STT
n
1, a, a 2 ,... a n ~\
we shall find
a 2 =(cos
/ 27T /
- . 27T\
2
=
4>TT 4-7T
+J-1 sin
J
cos
3 GTT
/ STT 27r\ 6-7
6-7T
-lsm
, .
a3 = (cos + J-lsm
.
=cos
\ n n / )
n n
STT
=cos
an
~l
= ( cos f-
J 1 sin
J
= +-
-
where the successive angles of transfer , , ,...
n n n n
follow the order of the successive powers of a : but if we had
assumed
4?r . 4-7T
a = COS
/
-
+J 1 Sin
n n
STT . STT
a 2 =COS
n
+J-1 /
Sin
n
,
127T . 127T
a 3 = COS /
+ J-1 Sin ,
n n
a""
1
= cos ^ --
4(w-l)-7r
-f J- I
/
sin
,
4(rc
^
where the order of the powers of the base of the period does not
follow the order of the angles, and where we must pass twice
round the circumference of the circle before we return to the
209
2 r TT ,
cos- -+ -
by the formula
2-7T , 2?r
+ J-1
.
cos sin
n n
27T
where is the angle of transfer, and n is the denomination
of the root.
(2)
, ^ .
(3) pcos -
+ x /-lsm ,
f
2(w-lW- + /-ls /
.
(n) pJcos- x
--
p
r
/
(
\
cos --
2W7T
n
J
- h
/
-
1 sm
.
-
2W7T\
n /
)
= p,
(
2(n+l}7T
-~ - -
- .
2(W+1)7T1 / 2 IT - . 27T\
J- 1 sm -"^-\ = P (cos V- 1 sm
, /
-f -f
p n j,
jcos
VOL. II. Dn
210
2 (n + r) TT 2 (n +
^
.
f
J- sm -
-+ r}
1
= rP (cos
n
+ J- 1 sin -
n J
)
;
tionsgene-
ralized.
an angle 6, with the primitive
(cos 6 + J- I
-
.
line, then
sin 0) p
cos + J- 1 sin
(
\
cos
n
+ J- 1 sin -\ r p
n-/
will represent a line in magnitude and position (Art. 820) as near
as we chuse to that which is assumed to be
expressed by
5
J_ JL ^9 397
27 109 136 1333" 10800
with a primitive line AB, is expressed, both in magnitude and p retat ons j
of *
position with respect to it, by ,-^
for R iven
(cos J^l
+v sin 0) p,
values off fl
,
6.
212
it follows, that if 6 or if AC
be perpendicular to AB, then
= 90,
AC is
expressed by p J^l (Art. 733 ) if be
6 c : ,
pressed by 5 ^
by
(A*.
position by
a and magnitude and perpendicular to A 13,
to p sin 6
h */-!, its then the hypothenuse ^4C of the triangle ADC
nusewillbe is equal to p, and makes an angle 6 with AB,
-
expressed an(j g therefore represented in magnitude and
for ^4c is
equal in magnitude to jo,
and makes an angle 6 with
and sine
and also that the symbolical difference of two lines AD and DC,
or a b J
be the hypothenuse of the right-angled triangle
1, will
formed by AD and DC, where DC is equal and opposite to
2
DC, and where Ac is equal in magnitude to ^/(a^+b ) and
makes an angle with the primitive line or a whose cosine is
829. It appears, from the last Article, that if AC and Ac(p) The sum or
06
be equal to each other, and make the angles C oftwT
and respectively with the primitive /\ equal lines
\
line
in
AB, they
magnitude and position by
,
are
p (COS 6
n
severally represented
/
+ ,J-
- 1 Sin 0)
/i\
A/_
\
V
-\ ,5 ^B
//
-^Son
as well as
magnitude
are the
diagonals
and
,
\\ /
/
of the
_
- J-
\/
c
rhombus
which they
form.
p (cos 6 1 sin 6).
If we add and
subtract these expressions for and Ac, we AC
get, for their symbolical sum, 2 p cos 6, and for their symbolical
it is
expressed both in magnitude and position by 2p sin ,J 1 :
symbolical difference.
lelog^am
m
tne P r i itive line respectively, then AB
considered
they are represented in magnitude and
in position . .
,
as well as position by
magnitude, , - .
/
is the p(cos0 + j-
f 1
Sin0)
diagonal ,
p cos
-
p cos + (p sin
-
p sin 0^ J- 1 :
\^ O y [*
JJ
and therefore
Ad = AE + Ae p cos + //cos .
Again,
CE = AC sin = p sin 0, and ce= CJV= ^4c 85110 = p sin ;
and therefore
NE = Z)rf = p sin + j/ sin .
) ^7 1
and
Cn = p sin 6
- f
p sin :
d(cos $ +*/- 1
sin0),
it follows that (Art. 828)
p - cos + p cos ~~
and cos </>
~~j
would be
p cos - p cos
and cos cf> -7,
832. The
interpretations which form the subjects of the pre- Applica-
,. . ,
. , ,. .. tion of these -,-, .
The sum of 834. The sum of the three sides of a triangle taken in order
the three
sides of a
is equal to zero.
triangle, For, if CD be drawn perpendicular to AB, we shall find
taken in
order, is a smB = CD = b sin A: and AB = AD+DB=b cosA + n cosB=c:
zero.
and it follows therefore that the sum of the three sides of the
called
for A+B+C = *.
j-
The same conclusion follows from the
AB : and since AD
equal to CB andis
a sin B = b sin A .
J-
C^7r
835. It will follow, as an immediate consequence of the The propo-
proposition in the last Article, that the sum of the sides of any Ar^sS
rectilineal figure, taken in order, when estimated both in position extended to
the sum of AD
and DE, or of AB, BC,
CD and DE is the last side of the figure
AE and, finally, the sum of EA and AE
:
an equation which is
equivalent to the two equations*
satisfy.
VOL. II. EE
218
" =
a + b cos 6 + c cos &+d cos 6" + e cos (a),
" =
b sin + c sin +d sin 0" + e sin (6).
angle"/.
A+B + C + D + E = 3. (c),
we get
"
if we now replace 0, , 0", in equations (a) and (b), by
these values, they will assume the form
i
sin AI - a 2 sin (A + A^) +...+(- !)"_!
l
sin (A 1
+A a +...A n_ )=0 (6), l
Again, since
A! + A2 + ... ^ n _! = (n 2) 7r-A,
^+ A 2 + . . . A H _ 3 = (n - 2) TT -
the equations (a) and (6) may be put under the form
a sinA
l l
-a 2 sin(A + A 2 ) + ...+(-
l
l)"-
1
a n _ a sin(A + A n _J
The equations (a) and (6) may be called equalio?is ofjigure, Equations
inasmuch as they are not satisfied unless the lines, from whose offigure
-
symbolical sum
these equations are derived, either form a
figure,
or are capable of forming one, when
arranged in consecutive
order : if their sum is not equal to zero, but to a + (3 J 1, then
+P J 1 will express the line, in position and magnitude, which
completes the figure.
838.
Thus, if a be the side AB of a regular hexagon Expression
g "
ABCDEF, then its several sides are re-
nl^of a
presented by a, a (cos 60 + ^^T sin60),
hexa^n
a (cos 1 20 + J~^l sin 120), ^ c
- ,
- a (cos 60 + J~^~\ sin 60),
A B
-a (cos 120+ J- I sin 120),
since the three last sides are parallel to the three first, but esti
30.
The sum of AB, BC and CD or
/3
whose cosine is
^ and sine ,
and therefore of 60 : this line
3a
~~~ H
| aJSj^l ~~
which is equivalent to
projections of AP : and if QR
be a line
equal and parallel to AP and estimated
in the same direction with it, and if
Qq and Rr, Qq f and Rr be drawn perpendicular to AX and
AY respectively, then qr and q r are its orthographical pro
jections, which are also respectively equal to those of A P.
2
N/(a+ /3 2 ) being the length of the line, and </>
the angle which it
841. Again, the perpendiculars let fall from a point P upon Relations
graphical projections of the line AP, which is drawn from the graphical
a
cons idered in their absolute as well as in their relative positions :
illation
of Algebra it is a theory full of important consequences and which will
figure, when the equations in Art. 837, were verified for if the :
th th
including the angle of transition from the w to the (w + l)
line or from a n _ to a was 2-rr or any multiple of %TT, and if the
l
answer the conditions of such a definition with the regular and equations
of angles.
completely bounded pentagon of Geometry the essential distinc
:
(2) A+
(3)
224
(4) A+B+C
(5) A + B + C + D + E = -57r.
In the case Thus, in the figures corresponding to the equations (1) and (4),
of pentago- , , . , . , , , .
nal move- the several interior angles are equal to each other, but with dif
ferent signs, and the figures are formed on different sides of the
The - or 60
Tn the case 845. angle of transfer in a regular hexagon is
3
:
of hexago-
ineats
Ve ~
^ we snou ^
suppose the sum of the exterior angles doubled, the
angle of transfer would be doubled or become 120, in which case
the corresponding hexagon would degenerate into an equilateral
or regular triangle and the describing point would pass twice
over each of its sides in six changes of movement: if we supposed
the angle of transfer tripled, the figure corresponding would be
come a straight line described by an oscillating movement: if
quadrupled it would become the regular hexagon of Geometry,
in a reversed position to the former, in which each of the in
octagonal
trans fer f ~ anc* f * ts successive multiples, would be distri-
* The
general theory of such periodic movements requires that the sums of
the orthographical projections of the n first lines described should be equal to
zero, where , or any of its successive multiples, is the least angle of transfer :
n
or if we express this angle of transfer by 0, then the equations of the figures
+ sin 20 + sin (n
- 1)0} = 0. ( )
p {sin
e 9
a + a~ becomes
If we replace cos0 by (Art. 807), the first series (s)
(Ar , 429 ).
9
_9
_ -(n-i)0 + a *- a
/a (n-*)0 \
s= 1 r"
i I
\ a2-a~2
- _-
- 6
a 2_ a by 2 N/^Tsin-,
-
n6
cos
(n-l)e
^ ^
{e\ sin(n-.e + , s.n / ^ __
sin-| ~
e I L*
sm-
sin
| | 2
VOL. II. F F
226
. 6
sm-
p sin TT cos
= 0.
. 7T
sin
n
p simr sm
* "
s = = 0.
. IT
sm
n
p sinm-Tr cos I
-
In- 1\
1 WITT
sm -
rmr
p sin m TT sm I I m ir
*
=0.
. rmr
sin
n
p sin n IT cos (n 1 ) TT
_
sin-n-
sin TT
Under this form, these expressions are indeterminate but it will appear by :
that the movement of the point, when = 2r, is one of continual progression
in the direction of its first motion.
* from the
It appears first equation of figure (a) Art. 836,
a - al cosA + l
a2 cos(^i +^ 2 ) ... +( I)"-
1
a n _! cos(^i + ^ 2 + ^-i) = 0,
that
227
0, sin A - aa
l
sin (A + A a ) + a 3
l sin (A + A 2 + A 3) =
l (2),
A + A + A 2 + A 3 = 27r
1 (3).
If we combine equation (3) with (l) and (2) (Art. 836), we get
a !
cos A +a
1 2 cos (A^ + A^) a 3 cos A= (4),
a sin
l A l
a 2 sin (A t
+ A2 ~)
a 3 sin A= (5).
A+A ~A l l
+ A 2 - A 2 + A 3 = A 3 + A = TT ,
we find
A = 7T
l A y A 2 = TT A! = A, A 3 = 7T A2 = A lf
that no one side can be greater than the sum of all the others, and from the
relation of the exterior and interior angles, that no angle of the figure can be
greater than two right angles.
* The
opposite sides of regular polygons of 2m sides are parallel, but it i-
not usual to call them parallelograms, unless the figures are also quadrilateral.
228
a sin A l
a sin (^A l + A^ - a = 0,
we thus get
1 = cos AL - cos ( AL
+ A 2) = 2 sin ^ (2 AL + A 2) sin ^ A2 ,
Consequently
-l=tan(2^, + ^ 2 ),
and therefore
its
angles are right angles.
Definition A
square is commonly defined to be a four-sided figure which
as uare:
(
has all its sides equal, and all its angles right angles: but if the
^ rJi
850. The definitions of regular and other figures which are Definition
the ratios , , ,
. .
^^ are the same in all similar figures,
0, a2 a3 a n _j,
and as the angles
therefore given, as well
* if the four angles be right angles, the opposite sides are necessarily
For
equal the equality of two opposite sides is not, but of two adjacent sides is,
:
an independent condition.
f One of the principal inconveniences which results from the introduction of
is the necessity it imposes of
superfluous conditions into the definitions of figures,
are satisfied
shewing that the unessential as well as the essential conditions
and de
whenever such figures present themselves in geometrical constructions
are satisfied thus,
monstrations, in order to prove that all the required conditions
:
_ i
_ 2
" _ n-2 _
~
a al fl
a a n_ 2
and therefore
a = ae, !
= !, a a =a 2 e, .. . a n _ z =a n _ z e:
but, inasmuch as
a + (cos ^, + - 1 sin 0,) + . o n _ 2 (cos n_ 2 + /v/^T sin
!
^/ . .
n
= J? (cos + ,/
- 1 sin 0),
and
a + a^cosa^^^Tsin^) -i- ... an
=x r
we shall get
ex cos <p
=x f
cos d> ,
and therefore
ex = x and fy
- </>
.
all
figures are similar, which have these conditions in common: Of nodes.
without enquiring into other consequences of the fulfilment of
these conditions, we shall content ourselves with pointing out
one of them, which furnishes a simple and immediate geome
of the similarity of rectilinear figures.
trical test
of the complement of AD
and DE
to DE, that is of AE to DE
complement of ad and de to de,
is the same as that of the that
ABC, ACD, ADE, and a be, acd, ade, similarly formed in each
figure, are similar to each other: and the same conclusion will
equally follow, whatever be the number of the sides of the
figure : it follows, therefore, that all rectilinear figures are similar
to each other, in which all the triangles similarly formed in each
are similar to each other.
VOL. II.
CHAPTER XXXII.
A system of 854. A
of logarithms is the series of indices of the
logarithms.
y
system
same base, which correspond to the successive values of n such :
There are 855. Though our general reasonings have reference to sys-
only two terns of logarithms calculated to any
J base whatever, there are
systems of .
logarithms two systems only which are commonly employed in analysis : the
^ rst * s t *lat ^ taDu ^ ar logarithms above referred to, and which are
comniorSy
employed, exclusively used in numerical calculation the other is the system
:
" *
and the" whose base is 2-7182818. . , which is called
.
Napierian, from the
Napierian. name o f the great man to whom their invention is due, and
which is almost exclusively used in analytical formulse.
a = a x+x = nn where x + x
x x r
(2)
= ax ~ x = , where x 1
x is the logarithm of : in
(3) (a
x p
)
= apx = n v, where px is the logarithm of n
p
in other :
th
words, the logarithm of the p power of a number is p times
the logarithm of the number.
- - - x -
(4) (a
I p
)
=a p = n p , where - is the logarithm of n p : in other
ih
words, the logarithm of the p root of a number is -th part of
a, ^
(
1, then a ;
.
860. The logarithms of all numbers which are not integral JjjJjjyL,
wlm-h are
powers of the base, involve a decimal part.
For if a be the base of the system, then the logarithms ofJJiJJnfJF
a, a~,
3
,
4
, or
of the integral powers of a are the indices of a, * ba
and therefore whole numbers: the logarithms of all numbers decimal
part.
236
10 ... 100
100 ... 1000
1000 ... 10000 3 ... 4
10000 ... 100000 4 ... 5
tal opera- Evolution, and the order of succession in which they are thus
The order 863. If we compare the operations with numbers with the
two unities,
with numbers corresponds to that of Addition with logarithms,
237
log 295316
=: 5.4702870
9
(3) To find the value of 9 , or of the ninth power of nine.
9
The logarithm of 9 is 9 times the logarithm of 9.
Log 9= -9542425
9
log 387420489
= 8.5881825
387420489 = 9
9
or .
log 260.6587
= 2.4160723
Consequently
^67943 = 260.6587...
log ^/67943
= 1 log 67943 = 1.6107149
log 40.80515
= 1.6107149
Consequently
4/67943
= 40.84515...
log 4.901245
= .6903064
Consequently
7/67943 = 4.901245...
This is an operation, which is extremely laborious and diffi
NX N
mal, expressed by the formulae 10" and, or in other
log ^ n
1UUU
98 = log 96.498
= 1 .9845228,
log lS = 1
S 9-6498 .
.9845228,
=log
*
For log IV x 10" = logJV + loglO B = n + logIV,
log N
= - log 10" = -n
log JV
:
log
In the three last cases the sign - is placed above, and not
Incon- 867. If, however, the base of the system was not coincident
of table^of
W^ ^e ra(^ x or kase of our system of arithmetical notation,
logarithms
to a base
the logarithms of all numbers included in the formulae 10"x N
jy
different
anc*
To"
anc* w^ cn are therefore expressed by the same succes
sion of significant digits, would not be known from one opening
of the tables: thus if the base was the number 2.7182828 (Art.
N to Nx N
logarithm of that of 1 0" or without an additional
*
Some modern writers are accustomed to subscribe the base to the prefix log,
in order to designate the particular system of logarithms which is employed :
thus Iog 10 means tabular logarithms log e means Napierian logarithms, whose base
:
firstfour digits of the number, the fifth being written in the same
line with it at the head of the successive columns. In the
column headed 0, are written the mantissce of the logarithms
of 26500, 26510, 26520, , the three first digits being sup
cated, in the place where it first occurs, by writing the four last
proportion
^ un ^ by actual reference of the Tables, to continue nearly the
with the same for the whole series of numbers between 26500 and 26800 :
of?nnin^ number of more than 5 digits (and therefore beyond the range
ber of more O f t ne
ordinary tables), we should proceed as follows,
places. We should find from the tables, the logarithm of the number
*
This proposition will be easily shewn to be an immediate consequence of
the logarithmic series, which will be given in Chapter xxxv.
243
871. Again, if it was required to find the number corre- To find the
sponding to a non tabulated logarithm, we should proceed as correspond-
follows :
tional parts, we find the number which is equal to, or next below,
the difference thus found and the digit opposite to it is the 6th
property assumed in the text, corresponding to the two next inferior units in the
6th and 7th places respectively, and therefore - and -- will be these dif
ferences, if a and b be the digits in those places : the first is found in the column
of proportional parts opposite the digit a : the second, removed one place to
the right, is found in the same column opposite the digit f>.
244
logarithms ^73. A
very little consideration will shew the great con-
of the nu-
venience, for the purposes of calculation, of increasing the logar-
values of ithms of the goniometrical quantities, as recorded in the tables,
6
cosines &c ^y the
number 10: for the natural values of the sines and cosines
are in- are included between and 1 and the characteristics of their
10. logarithms are therefore necessarily negative: thus
245
sin 1 = 0002909:
. its unaltered logarithm is 4.4637261,
sin 1=. 0174524; ......................... 2.2418553,
sin 50 = .7660444 ; ...................... ... 1.8842540,
we should say that the tabulated logarithms were those of the sines, J^^f
cosines, &c. in a circle whose radius was 10 or 10000000000 logarithms
A + B+C=7r* (3).
and therefore
c = AB = AD + DB=bcosA + acosB, or A
c a cosB - b cos A= 0, (1)
which is the first equation. And again, since CD = b sin A - a sin B, we get
AB = c = AD BD = b cos A + a cos B,
or c-acosB-&cos^ = 0; (1) Ac
and again, since
CD = ,4 C sin /I = a sin ,4
Of the six elements which present themselves in every tri- Two only
angle, namely, the three sides and the three angles, there are
five involved in the first of the
preceding equations, four in the Bangle
second, and three in the third. The last equation, which is the specific
equation of angles, determines the third angle when two of them
11 11 *
fo",^ ,?
are given : it follows, therefore, that two only of the three
angles termina-
ll
can be considered as dependant upon the specific* conditions of
the triangle.
_=
a sin B determine
sin 5 ~sm~Z~
sin A=T
b
sin B, sin B - sin A.
a
, . a sin B
we replace b by
J .
A , we get
sm A
c =a cos B+a sin B A
; cos
sm A
- **(
sm ^ cos B+ sin B cos A)
A sin
_asin(A + B) _a sin C
sin A sin A
877- The conclusions in the last Article may be exhibited The sides of
, , ,,
a triangle
under the torm are pro-
portional to
a __
sm A4 a _ sin A i
b
_ sin B
r>
the sines of
the oppo
b ~sinjB c slrTC" c suTC*
site angles.
* The which are not common to all triangles,
specific conditions are those
but peculiar to that which is under consideration : the three sides and two
angles be given, they are not the angles of a triangle, unless their sum be 180
:
but if two of them only, whose sum is less than 180, be given, there is always
a triangle of which they may form two of the angles.
248
c
2
= a 2 (cos2 B + sin
2
B) +
2
b (cos
2
A + sin
2
A)
+ 2ab (cos A cos B - sin A sin B)
cosC: (),
B + sin B=\, cos A + sin A =
for cos
2 2 2 2
1 ,
isB (c),
Expression 879. If we severally solve the equations (a), (b), and (c)
cosine of
with respect to cos C, cos A, and cos B, we shall find
an angle
a + ; ~
2 2 _ 2
c
of a triangle
in terms of
r< _
% a fr \a)
( ^
its sides.
cos B=
2ac
which are expressions for the cosines of the several
angles of a
triangle in terms of its sides.
Expression
for the sine
of an angle 880. Corresponding and more symmetrical expressions may
e fu n(l for the sines of the angles of a triangle in terms of its
249
= (a + b)
2
-c 2
= + c - b) (b + c -
(a a), (h).
Replacing 1 cos
2
C by 2
sin C, dividing both sides by 4># 6
2 2
,
If
olll
we make
b
\_x
+
-
c -a =
.9 = -
a
b
+
+
b
c
+
+a
c
,
we
2a
find a
= %s
+ o
2a = 2
+ c =
(s
2s.
a),
and therefore
rinC-M (
-^p
*>< --?
(0-
we
In a similar manner,
sin A= Is/fc
find
r) C
--
ac
VOL. II. 1 1
250
,f be sin A ^
AA ^by
.
It in -
, we replace sin the expression given in
2
the last Article, we get the area of the triangle (k) J\s (s a)
(s b)(s c)}, or it is
equal to the square root of the continued
product of the semi-sum of the sides and of its several excesses
above the three sides of the triangle.
The pre 882. If it be granted that the value of an angle may be both
ceding ex
determined and calculated from the value of its sine, cosine, tan
pressions
are ade
gent, &c. (and the limits of the values of the angles of triangles
quate for
the solu will be found to remove all
ambiguity when the value of the
tion of all
is not to be determined from the value of the sine*), the
the cases of angle
Trigono expressions given in the preceding Articles will enable us to
metry.
calculate the sides, angles, and areas of triangles whenever the
data are adequate to determine them.
We shall now proceed to consider the different cases of data
which are sufficient to determine a triangle, and to exemplify
their logarithmic computation.
*
For the sine of an angle is identical in magnitude and sign with the sine
of its supplement ;
but the cosine, tangent, and cotangent of an angle are seve
rally equal to the cosine, tangent and cotangent of its supplement in magnitude,
but different from it in sign.
251
sin sin
and C.
In the first place, to find C,
we have
7T-180
^= 37-14
B= 67.27
A+B 104.41
C= 75.19
To find b, we have
log 749-6= 2.8748296
log sin 67.27
= 9-9654582
12.8402878
log sin 37. 14
= 9-7818002
log 74,9.6
= 2.8748296
log 1144.16
= 3.0584876
= 3.0786092 or c = 1197-42.
log 1197.42
252
sides.
sought for is opposite to the less of the two given sides, then
B is less than A and therefore
necessarily less than 90 in this
case, therefore, there is no ambiguity*.
But if b be greater than a, or if the angle, whose value is
* For the
greater side of a triangle is opposite to the greater angle and
conversely, a conclusion which is proved in Geometry (Euclid, Book i. Prop. 11.)
and may be easily deduced from the fundamental equations in Art. 875 for if :
a be greater than b, then sin A is greater than sin B if A be greater than 90, :
then B is less than 90 and therefore less than A, inasmuch as the sum of
the angles A and B is less than 180 but if A be less than 90, then B must
:
plement TT B must be less than 90, and therefore less than A, since sin (TT B)
is less than sin A but if-Tr-B be less than A, then IT is less than A + B,
:
11.2320359
log 93.451 1.9705840
11.2320359
log 17.09 1.2327420
and both these angles will equally answer the conditions of the
problem.
log 17.09
- 1.2327421
arith. comp. log 93.451
= 8.0294160
ing to the technical usage of the term, may admit, in many cases,
of more rapid computation, by mixed or even by merely arith
metical means, than one which is so: the selection therefore of
one or the other, when more than one method are within our
reach, must be determined by the judgment and experience, and
sometimes by the taste of the computer.
Solution of The introduction of subsidiary angles will enable us to adapt
the pro-
tne expression
posed prob
lem without 2 ,
2
the use of
angles!^
or an y other, whose terms are connected by the signs + and ,
8 n
Since j = !
-4 (Art. 877), we have (Art. 792, No. 8)
U Sill. JLJ
tan -
a + b
~
sin A+ sin B
a b sin A- sin B
tan
tnr)
255
. A +- 13 TT C . .
- known
is known, and therefore is :
consequently
b = 27.04,
and therefore
a-b = 47.63 and a + b = 101.71,
and therefore
A+B=
31.20 ,
- log
B +
log tan tan \ + log (a-b)- log (a A),
(^-^) (^-
Iogtan31.20 = 9-7844784
log 47.63
= 1.6778806
11.4623590
log 101.71
= 2.0073637
Iogtanl5.55
= 9-4549953
A+B=
31.20 ,
2
A~B
256
Therefore A= 4*T.15 ,
and =15.25 .
Also c = -~ -.
j- and therefore
sin^I
log c
= log 74.67 + log sin 1 17.20 - log sin 47-l 5
log 74.67
= 1.8731462
11.8217314
Iogsin47.15 9-8658868
The cosines (Art. 879) of the three angles of the triangle are
expressed by
cos A=
cos BD =
a 2 +b 2 -c*
cos C =-
and where the factors are formed therefore by a very easy and The ex-
The log sin A, which is here sought for, is the tabulated loga-
of.
rithm, exceeding by 10 the proper logarithm of ^ .
but if not, this equation will be satisfied, (assuming that the arith
metical and logarithmic processes are correctly performed), by
the
taking that value of the equisinal (Art. 776) angle opposite
greatest side,
which is greater than QO".
Let the three sides of the triangle be 107-9, 193.4- and 217-12 Example.
to find the three angles of the
yards: and let it be required
triangle and its area.
VOL. II. KK
258
a -107-9
b = 193.4
c= 217-12
2) 518.42
5=259-21 :
log*
= 2.4136518
a= 107.9
5 = 259-21
6=193.4
= 259-21
- = 1.6241789
s42X)9 :
log (* c)
2J 8.0359902
log A
= log 10423.06 4.0179951
10 + log 2 = 10.3010300
14.3190251
2.2844565
-log 193. 4
=
log 6
=
log c log 217-12
=
2.3366998
+ log c :
4.6211563
log 6
= log sin ,4
= 9-6958686
log sin 29. 46
Again,
10 + log 2 + log k = 14.3190251
= 4.3697212
loga + logc
4.3194779
Therefore A = 29.46
B = 62.51
C= 87.23
180
C
1 + cos C=2 cos 2 = 2s(s-c)
2 ab
I - cos C= o sm~ -
2
.
=
2 ab
and therefore
If therefore we make
C R B R A --
R
tan = ---- tan = -- tan = -
s - c -
, , ,
2 2 s b 2 s a
in the
expressions which are not only more easily calculated than the expressions
" Geometrical Problems and Analytical
text, but are also free from ambiguity. See
Formula?, with their application to Geodetical Problems ;" p. 21, a very ingenious
and original work by the late Professor Wallace, of Edinburgh.
CHAPTER XXXIV.
if we make 2
tan 6 =- *, where 6 is the subsidiary angle, we shall
find
a + b = a(l+ tan 6)
2
=a sec
2
= -^2 :
cos 8
and therefore
log (a + b) = log a + 20
- 2 log cos
* = -
Therefore log tan 6 10 + (log 6 log a).
261
,
{.97 cos 14 + 410.7 sin 6?. 10
log 3.97
--
2.5987905
12.5856946
log 410.7
=
2.6135247
= 12.5780849
= 2.5780849
log a (rejecting 10)
log b log a
= 2) .0076097
= 10.0038048
log tan 45. 15
= log + = 2.8829215
log 770.56 b)
889 Let the expression be a - b, where both a and b are In the case
, of a- / .
which gives
= 10 - J -
log sin 6 (log a
:
log b)
we thus get
a - b =a (1
- 2
sin 0) =a cos
2
6,
and therefore
- = - 20.
log (a h) log a + 2 log cos
262
a l-a(l cos 6)
2
=a sin 6,
2
and therefore
log (a
- b)
= log a + 2 log sin 6 - 20*.
0-8
a+ b b i + tan
1 +-
a
Thus the expression (Art. 886),
tan
fA-B\
- - = a-b tan fd
- + B\
(
V 2 / a+b \
7 (
2 /
*
Thus, if two sides a and b and the included angle C of a triangle be given,
we find
a
_ sin A _ sm{Tr-(E+C)} _ sin(E + C)
b sin B sin B sin B
and therefore, dividing by sin C,
& sinC
;md consequently
cot B= ? - - cot C
b sm C
=
alb ~
cos C\
& sin C\ a /
1 f we make -
= cos 0,2
we get
a
cotB =
6 sin C b sin U
b sin 2 6
cotyl = .
a sin C
This a very convenient and expeditious method of solving a triangle when
is
becomes
if we make tan 6 =- .
^(a - = a J(i -
2
6
2
)
2
cos 0) = fl sin 6.
= J{a 2 +<2ab + b
2 - 2ab + cos C)\
(1
( 2
C
4>ab cos
y\r-trnr
cos 0)
2
= (a + b) sin 0,
.cos-
making; cos 6 = }
a +b
In this case
- log cos -
log cos + log 2 + \ (log a + log b] log (a + b)
log c
= log (a + b) 4-
log sin 010.
2 2
892. Let the expression be J(a + b ). In the case
b
If we make tan - =- , we get
z
J(a + b
2
)
= a/J(l+ tan 0)
2
=a sec B
a
cos
264
cos C)
2*6(1- cos C)
(a -6)
COS0
Sm
2
making tan = ,
or a^ j(a
f
2
+b 2
).
-= sin 6,
a
we get
= 2a cos
2 -
fl + cos 0) ,
- = 2a 2 -
a - =a (1
cos 0) sin 1.
which gives
c
(a-bcos-
= & cot-,
a + b = asec d,
2
if tan
2
=
|.
a sec 2
sec
2
by a", we find
Similarly, replacing
2
a +b + c + d = a" + d = a"sec d"
2 2
= sec 0sec
2
sec 0",
and if we make
-^ = sin 6,
P
the two roots will be expressed by
- 2
pcos
2
and psin -.
we make
2V? _
= tan 0,
and if nuiKX
;
1
P
the two roots will be expressed by
r\
VOL. II. L L
266
=
^=a+b+ =
2
if tan 0"
c
When the 895. Let the expression consist of series of converging terms
convergent
connected alternately with the signs + and -, such as
and the
terms con- a o + c a + ...
nected \vith
the signs In the first place
+ and -.
a-b = a cos
2
=a ,
if -= sin
2
0.
a
Again
2
a a cos
= 7=
if
a! a-b
If we include a fourth term, we get
2 2
a cos cos B"
ab+cd=a- d = a cos
cos
2
if . = =s in2 0".
a a o + c
or (a
- b) (a
-b f
) (a"
- b").
a + b or
a a a
and we get
267
sin
2
= -, sin
2
= -,, sin 0"=^,..-.
a a a
and we get
- f
- V) (a"
-
b")
= aa a" 2 2
cos 6 cos 6 cos 0".
2
(a b) (a
if we make a = J(a2 + b
2
cos and - sin 0, we
Thus, ),
^= ^
get
a sin A b cos A a sin 6 sin A sin cos A
= sin (A 0)
= J(a* + b
2
)
sin (A 0),
and
a cos A 6 sin A= cos cos A a sin sin y4
2
- a cos (A=F0) = J(a* + 6 ) cos (A^Q).
An form
expression of the
a sin
2
A- b,
2
2 b =a sin 0,
where 6 is less than a sin ^, becomes, by making
asinM-a sin
2
= - sin (4 + 0) sin (4-0),
x
Develope- 898. THE exponential expression a , the series into which
mentofa*. |t may j^ d eve i O p ec^ an(j t ne various logarithmic and other
series which may be deduced from it, enter very extensively
into analytical enquiries, and deserve the most careful examina
tion of the student. We shall begin with the developement of
ax into an equivalent series.
x
If, in a 3 we replace a by 1 + (a 1), making
-{!+ (a -I)} ,
which, when severally divided by the terminal (Art. 688) coefficients 1x2,.
,r .T ,T
... (3),
-+ 3A,-s+ .
(4),
are obtained implies that they are identical in all those terms
in which x and y, one or both, are similarly involved: we are
thus enabled to obtain a series of equations, expressing the
conditions of identity, by which the form, or analytical values
of the successive coefficients are determined.
a =i+A J ,
.
+ _
k*x*
+ __s+ s
k x
3
...
(),
rical value
of.T.
-
we shall find, by making x = K ,
aLi + i+L
+
.2
+
1 1
1
.2.3*
a rapidly converging series, from which its numerical value
may
be calculated to any required degree of accuracy: the aggre
gation of 14 of its terms gives us
a~k = 2.7182818,
* The
process for this purpose is very simple and expeditious: divide 1, and
the successive quotients which thence arise by the successive natural numbers :
2) 1.
2.71828180
k
Modulus
Again, since a = we find
of 904. e ,
a system of
logarithms. a x
= ^ x = n,
where kx is the Napierian logarithm of n, whilst x is its
logarithms by j-
or --- , we shall obtain the corresponding
o
or replacing a by w,
2
log w
= (n
-
1)
- >(w-1); + (n-lV
v (n-1)*
; - v y
+ &c. (c),
g g ^
a series of great importance in analysis, and which expresses
the Napierian logarithm of a number in terms of the number
itself.
10* -1 =2.162277,
10* - 1 = .778279,
10 5 -1 = .333521,
lOw-i = .154781,
10 F2 -1 = .074607...
- 1 = .000000000536112,
32
which, multiplied into 2 *, gives
8
^(lO* -!)* 2.3025851,
a result which expresses the accurate value of the
Napierian
logarithm of 10, as far as it goes: for it will be found that
EPS
=
2.3025851
" 43 *2 9 4481 *&
isthe modulus (Art. 904) of tabular logarithms, or it is the factor
l_
10*
54
- 1 = .000000000000000127819149320035 :
the same work contains the tabular logarithms of the natural numbers as far
we - x, we
910. If, in this equation (e), replace x by get
fore ,r =^ i we get
(A).
q p+q
When
P= g + l.
fore
If we make p = a+l, we get
p+
-=s
q
-
2^+1
, and there-
or
_1
log ( 9 + 1)
= log , + 2
{^ + J
( T?
+
&c.[
().
275
2 -
= -
p = x and q = x + and therefore
2
Thus, let 1 (x 1) (x 1),
_ *
: we find log =2 log *-log (*+!)- log (*-l),
p + q 2.r -l
2
q
and therefore
log (x + 1)
= 2 log x - log (x - 1)
l 1
"2 f
3
+ \J
+3 i
2
-l)l2^=T"J. (2,r
series.
converging
Again, we assume
if
= 2 - 2 2
= x*- 98 4
+ 2401 x 2
p x (x 7) (x + 7) ,
q
= (x-3)(x + 3)(x-5)(x+5)(x-8)(x + 8)
= x 6 - 98 x 4 + 2401 x 2 - 14400,
p-q = 7200
we find 6 - 98 x4 + 2401 x 2 - 7200
a:
and therefore
= 2 log + + 2 log x + 2 log -
log (x + 8) (x 7) (x 7)
7200
^ _ 8^ + 2401^-7200
276
log (n + 1)
= log n + log fl +-
and therefore
Again, since (
+ or)
= (!+ we find
j,
log (n + x)
= log n + log fl + - ,
j
and therefore
X X* XS
+ x )-log n = -- i+ 3
-...
Napierian, then
and
log (n + x)
- log n = p
(^
-
^ +
&c.^
log (n + x)
= log n
will furnish an approximate value of log (n + x) it will be at :
In a geo-
metric
916. If we make x = -
a + ^, the sum of the geometric series
series.
a + ax + ax* + ax + .. .
is equal to a + 8.
^, we get
is less than -
a+
s <
S 1
sav x =
2
-.
(+))
- ^I , will make s differ from a
918. It will follow therefore that the limit of the value of The limit
the series
a + ax + ax + ax+ .. . series pro-
is its first term or a : for it has been shewn that a value of x cording to
a + ay + ay 2 + ay 3 + . . .
ceeding ac- if r be the greatest inverse ratio of any two consecutive co-
cording tO /
of efficients.
powers
we have = and
sequent ratios of a similar kind : then r,
aa
therefore a l = ar : is less than r, and therefore a 2 is less
a
-^, and not the first , be the greatest, and therefore equal
than ar: less than r, and therefore 2 I GSS tnan a i r> anc^
a
fortiori less than ar 2 , since a l is less than ar: and so on, until
n ~
we come to a n _ lt which is less than ar
l
the next ratio
:
n
= r, and therefore a n = an _ l r, which is less than ar , since
~1
_! is less than ar tl
: and similarly for all subsequent coeffi
cients: it will follow, therefore, that the terms of the series (a),
after the first, are severally greater than those corresponding
to them in the series (/3).
th
If the successive ratio. . , , as far as the 7/ ratio
a al a2
and (/?)
are equal to each other: but all the subsequent terms
of the series (a) are severally greater than those corresponding
to them in the series (ff).
281
+ 5 , and
therefore x=
+ 1)
r (a
- g
^ , the sum of the series (a) is a + , and
therefore the sum of the series (/?) is less than a + B : and the
limit of value (Art. 917)
its is a, or the first term of the series.
2 + 2x3,r + 2x 3V + 2 x 3V+ . . .
(a),
posed (/3)
: if x= - - ^ , the sum of the terms of the series
(/3), will differ from its first term 2 by a quantity less than g.
s
unknown,
^ j
than the third; then, if the first and third series have the same is^Vi h l
in value be
first term or the same limit, the first term or limit of the second
VOL. II. NN
282
tween two series willbe necessarily equal to it it being assumed that the :
others .. />
which have
whi^h inverse ratio of any two consecutive coefficients of those series
the same is finite.
limit.
always
*
Thus, if the three series representing the three quantities be
a + diX + a 2 x z + a a x 3 + ... (1),
s 3
b + b1 x + b2 x + ba x +... (2),
2 3
a + Cj x + c a x + ca x + ... (3),
a b + 8 8j and b a + 8j
- 88
rf + g_g 1
and -d + ^-Bg:
and inasmuch as these expressions are necessarily positive, it
will follow that, in the first case, 8-8, is greater than d, or 8
The series 923. IT has been already shewn (Art. 808) that the value
_ _____ k 3 x3 k*x*
1.2 1.2.31.2.3.4
where k = loga: we thus find
2 1.2 1.2.3.4
a* -a-* I
(, k3 x 3 k b x* \
sm * = ** + + +
-^l=7=TV
2 TT8T8 i. a . 8 .4.5 -)
If we further replace k by cj-l, these series become,
COS X = 1 -
1.21 .2.3.4
sin x = ex
Themea- 924. It may be easily shewn that these series will satisfy
e fan
the equation (Art. 758)
ang i e
which is .
2
assumed in sin * + cos 2 * = 1
assume, as we have already done (Art. 74-6), x or the measure Art 746)
of an angle, to be the ratio of the arc which subtends it to the determines
the value
radius of the circle in which it is taken, then it may be demon- of the sym-
ucj
strated that the only value of c, in the series for sin a: and
1 is
jjjgy inl
cos x, which will satisfy the conditions to which it leads. volve.
we get
sin x c
5
x4
= c
1.2.3 1.2.3.4.5
sin a?
where c, which is the limit of (for it is the limit (Art. 918)
2 cos sin
and therefore
CP = cos-:
(Art. 775)
chord BC
2 sin - 2 sin -
-
It appears, therefore, that is
interposed in value between
sin x c
3
x2 c
5
,
x 1.2.3 1.2.3.4.5
cosj;=1
_^ + _^__... I
greater than be, and therefore Bt + Ct, and therefore a fortiori BT, greater than
Bb + be + cC : in a similar manner, if we bisect the arcs BE and CE, and from
theirmiddle points draw tangents meeting Bb, be and Cc, then the sum of the
tangents thus formed will be less than Bb + bc + Cc, and therefore a fortiori
than BT
by continuing this process, we should increase the number and diminish
:
the magnitude of these small tangents, until their sum, which is necessarily less
than BT, shall differ from the arc BEC by a quantity or line less than any
that may be assigned or in other words, the arc BEC is necessarily less than
:
22 , 2"+^
2rsin|, rsin^ sin^-f^
will express the chord BC and the several sums of the successive inscribed chords :
th
the ratio of the n th term of the second series to the n term of the first is
a
1
+i _^ cos
s in
EL ^
COS
2" <6"~ri
and the limit of this value is 1 : this is another mode of arriving at the conclusion in
the text.
This relation of the chord, arc and circumscribing tangent, is one of fun
damental importance in the application of Algebra to the theory of curves:
it applies to the arcs of all curves of continuous
curvature, as well as to arcs
of the circle.
287
cos x = 1 -
1.2 1.2.3.4
sin x=x
1.2.3 1.2.3.4.5
where x is the measure of an angle, which is formed by dividing
the arc which subtends it at the centre,
by the radius of the
circle in which it is taken.
c = 1 and * = c 1 = n :
we find
and therefore
a* -a
sin #=
= = cos a
cos a:
+v/-lsinr.
and
we get, by adding
and by subtracting
cos x J 1 sin ,r = e~ * ^^ .
r *
2r?r
f7r 7 v - 1
= f\\
J- J-l =
/ / vr~
cos +v 1 sin v e (
l )
n n
288
and if the series be continued, the same values recur in the same
order.
cos- +
7T -
J -Ism- =ez
/ V^T . 7T ""
or
it follows that
and therefore
. .
/-I
Conclusions of this kind are symbolical only, and cannot be
made the basis of calculation and it is only in those cases in :
and
cos n6 + 7^ sin n0 = en ^~ = (e ^~
l e l
}
n
= (cos + J~l sin 0)":
and if in
~x ^~
cos x J 1 sin x e l
and
~ ne V:TI
cos nQ-J~\ sin n0 = e = (e -o V~)n = ( cos d-J^lein 0)".
a
=
sin A = sin (B + C) = cos B sin C
D 5 COS C 4-
-T
6 sin B
. ^
sin B sin Bo:
,
and therefore
sin B b sin C
cos B a b cos C
if we replace the sines and cosines their equivalent expo
by
nential expressions, we get
and therefore
log 6
2B V^T .
log (a-be-c ^ - log (a
- be c ^)
~
a
or
For if
L= S.
f
we get
^= -
,
and conversely. Art. 296.
VOL. II. o
290
or
=- sin C+ 5
2
sin 2 C+ d
sin 3 C+ . . .
a 2 3o
This series converges rapidly if b is small compared with a.
and therefore
- m-1 - 2<x V-l
1
,
/ m+ 1
m- 1 _
from whence we derive, as in the last Article,
we readily find
or
x = tan ,r - J tan 3 j? + 1 tan 5 # - . . .
291
v3
we get
TT
= 1,1111
+(l
11 +
6 ~j3 -3 3 5"3*-7 3
3-
a series which converges with great rapidity.
933. Among many other conclusions which are deducible The sines
from the series for the sine and cosine of an angle, it may
angles are
nearjy P r -
J shewn that the sines of small angles are very
be readily J nearly
J
. portionalto
proportional to the angles themselves, a proposition whose truth the angles
themselves
we have assumed in a former Article (798) for if 6 arid be :
-
sin
are very small, or differ very little from each other. Thus, let
the angles measured by 6 and 6 be I and 2 respectively: inas
much as an arc, which equal in length to the radius (Art.
is
1 2
747), contains 57.3 = 4438 nearly, we get 0= and 6 = - ,
44oo 44?o
and therefore
sin 1
=A X 1.000000025
sin 2
=l nearly.
and subtract from their sum the series in which tan x = =7- ,
the result will be
the expression for j : it was by this method that the remarkable approximation,
Formulae 934. Again, the ordinary formulae for the solution of triangles
solution of sometimes give results, which are less accurate than those which
8
are commonly obtained, when they involve the cosines of angles
volye the which differ little from zero or 180, or the sines of those which
angle?
differ little from 90, inasmuch as, under such circumstances,
which dif-
change slowly for considerable changes in the value
their values
from zero, of the angle the series, however, for the sines and cosines of
:
those such extreme cases, and where the minutest changes of value
which dif- , , .., . .
fer little m . .
from 90.
omitting all terms of the series for cos 6 beyond the second, as
being too small to affect the result within the limits of the
recorded places of figures ; by extracting the square root, we get
c = (a TAJ,
+ b) < 1 -
(
-
abd *
3
2(a +
r^2 V
b)
I
)
nearly.
= 35.999424 nearly.
Consequently
c sin A
sin C
+
0-^
a+ 2
= _^L |!-^ + (
/) l
a + j3 I
ft j
Ca ,
6-J*,,, +
a very convenient and very accurate formula for expressing one side
of a triangle
in terms of the two others and of the measures of the adjacent angles, provided
those angles are small.
CHAPTER XXXVIII.
Meaning of 935. THE terms infinite and indefinite are frequently used
LlnSTand indiscriminately by mathematical writers, though, if due regard
indefinite. was p aij to propriety of language, they should be distinguished
from each other: they are negative terms whose meaning must
be denned by that of the terms Jinite and definite, which are
respectively opposed to them.
A
Jinite number, a Jinite line, a Jinite space, a Jinite time,
would denote any number, line, space, or time, which is either
assigned or assignable: whilst the term dejinite could properly
be applied to such of those quantities only as were already
assigned or determined: in other words, the term Jinite is more
comprehensive than dejinite, being limited only by the power
possessed by the mind of conceiving the relations which the
magnitudes, to which it is
applied, bear to other magnitudes
of the same kind.
An infinite number, an infinite line, an infinite space, an in-
* The phrase for ever, though properly expressing infinite duration of time,
is commonly applied to denote infinite repetition as well as infinite time thus :
the processes which never terminate are said to be continued for ever : the terms
295
used to denote those which are infinitely small: they are con
which, when
multiplied into zero or an infinitely small number
(fractional or decimal) will produce a finite product there is :
duce a finite result, but it does not follow that it must do so: symbol.
00
= =00 or infinity.
of a series which are said to be continued in injinitum, are also said to be con- thecircum-
tinued for ever but it should be observed, that the notion of infinity of time stances of
:
th er origin
is closely associated in the mind with all our notions of indefinite repetition.
are known.
296
Example. But though the symbol oo might not have the same origin
in the expressions a x oo and b x oo , yet if the precise symbolical
conditions of their origin in both cases were known, the inde-
termination of their relation to each other might be removed :
2
expression j when a?=l, we should find
x (1 -x) (when
^ *. !): b x
297
a x oo 2
ax + ax 2 + ax 3 + .. .
x
whose value is a when x is zero, it will follow, that if we denoted
the first zero by and the second zero by 0, we should find
o-*
or, in other words, a definite relation would thus be shewn to
exist between the zeros denoted in this particular instance by
and 0: but all traces of this relation would be obliterated
if and were replaced, as is usual, by the common symbol
as the common and only representative of zero, and if no reference
was made to the circumstances of their origin.
Again, the sum (s) of the series
ax 2 + ax3 + ax 4 + . . .
- = ax + ax* + ax 3 + ...
x
VOL. II. Pr
298
when x we - we
which is zero is zero : if again divide by x, get
s
-5 = a + ax + ax + .. .
x*
0"
the zero corresponding to x by 0, we shall find = , and
0"
therefore = 0, whilst =a : in other words, the relation be
TJT
tween the two first zeros is zero, but between the two last is a :
a
- + a + ax + ax + .. .
x
is infinite when x = : and similarly the sum s of the series
a a
5 H \- a + ax + ax + ...
x2 x
s = -, when x = 0,
oo oo 1
oo = and therefore =- ,
oo
should have lost all traces of the relations which existed amongst
them.
*"" fl
As an example, the fractional expression , becomes
- Examples.
x -a o
when x = a: but if we obliterate the common factor x-a (which
becomes when x d) of its numerator and denominator, the
fraction assumes the equivalent form
a2 + a x + x 2
a -f x
which becomes = , when x = a : its value, therefore, is no
longer indeterminate.
Again, the expression
- ax 3 - 3a*x* + 5a 9 x - 2a 4
becomes - , when x = a\ but if we obliterate the common factor
- 3
of its numerator and denominator, the original fraction
(x a)
943. But if
different powers of a common factor exist in When dif-
infinity.
Thus, a-2x is found to be a common factor of the nume
rator and denominator of
3a
3 - 10a 2 .r
300
When the 944. In many cases, however, the expressions which become
expressions ^
which be- _ or for S p ec ifi c values of the symbols which they involve,
come oo
O
or are do not present themselves under a rational form, and the ordinary
irrational
therefore, for their common factors ceases to
or involve
exponential
rule,
_ i.
be applicable.
,11-11detecting
the absence
Iii
^i
or the more prompt and
-,
certain
logarithmic
or gonio-
me thods of finding their determinate values, under such cir-
metrical cumstances, whenever they exist, which will be given in a sub
sequent Chapter of this work, we may very frequently succeed
in determining them by the following method.
TJ /A .
which becomes
when h = 0*:
Let the expression be
a
n
-xn
log a log ,r
_
log a
- log (a - A)
a 2a 2
= wa", when A = or x a.
the common radical sign, which is thus found, is the common factor of the
we get
x~ x3
= w, when A r= 0*.
An expression which assumes, for specific values of its symbols, the form
x co becomes by inverting the expression which becomes co thus
,
^ :
l^?f
IT
^^~^ = TT ,
^
when a: =
-7T
-r ;
cot a: 2
2
and its determinate value is . In a similar manner, an expression such as
x 1
~
x- 1 log x
which becomes co -
co , when x = 1,
assumes, by reducing the fractions to a com
mon denominator, the equivalent form
x log T a; + 1
(x-l)log*
^+
passes through all values between ~- h and h, including -~ , its cosine is
cumstances, and we regard zero as the cosine, and infinity as the tangent of
pose examples of a transition through zero and infinity, in which no such change of
affection or sign takes place : thus if a curve
PCp touch the line ACE in C, and if it be
symmetrical with respect to it (such as the
arc of a circle) on each side of the point C :
If we suppose b
r
= b, and a not equal to a, we get
x = 0.
If we suppose a =a, and b greater than b, we get
07= 00 .
=
0-
In the first case, 0, and in the second, oo, are the only values
of x, which will satisfy the proposed equation, in conformity
with those symbolical properties of the symbols and oo , which
distinguish them from ordinary symbols.
In the third case, all values of x will equally satisfy the con
ditions of the equation, which is therefore absolutely indeter
minate* :
(or oo ) is the only form, which the general
and
\ /
expression for x could assume, which would not give a de
terminate value, whether Jinite, infinite, or zero : the interpre
(l),
c (2),
_ Vc-bc
a c-ac
_mbc mbc
mab mab ~0*
__mac mac
y ~ mab-mab ~Q
which shew that x and y are indeterminate as far as these ex
pressions are concerned : for the equations
(1) and (2), under such
circumstances, are not independent equations (Art. 391), the second
being derived from the first, by multiplying each of its terms by
the same symbol m y and it furnishes, therefore, no new con
dition for the determination of the
symbols which it involves.
The values of a; and y, in the equation
ax + by = c
are separately,
though not simultaneously, capable of all values
whatsoever, and are, therefore, when separately considered, abso
lutely indeterminate: but if we consider them as connected in
the proposed
equation with each other, the value of one of them
determines the other, and one of them only can be considered as
absolutely indeterminate with reference to the other.
al value, we
obtain diverging series, whose terms are alter- arithmeti-
VOL. II. QQ
306
equivalent results?
These questions, which are of fundamental importance in
many of the most delicate and difficult applications of analysis,
have given occasion to much dispute and controversy, and have
not hitherto been satisfactorily settled. It will be prudent there
fore for a student, whenever he is called upon to consider the
infinite values, whether of series or of expressions under a defi
The values
947. The symbol oo will sometimes present itself in ex-
ofCGSOO . i.ii
which have periodical
T i
when
i
the result will
MI
and sin oo. pressions, values,
2
(I-*)
and when x is greater than 1, the first series is an alternating, and the second a
progressing divergent series : but and are, under these cir- <
a )8
307
recur, when other resources are within our reach: they form
the value is assumed to be unique, is the only value which the doctrine of
fmproper!
dimensions of the symbols in their numerators are lower or higher
1 + x4
and 2 is an improper algebraical fraction.
larly
J = is reduced to x 2 - 1 + where x 2 -! is an integral
1 + x2 1+x2^
o? + 1 A B
x*-7x + 12 x-3 x-4f
where A and B are unknown and are required to be determined.
A + B=1,
g+1 ~
-4 5
,r
2
-7a? + 12 ,r ~3 JT
-4 "
f
- - is not merely one of equality, but also
MM
we make
-
-^i. =
- r-pr =
N (a + bx)Q=
A -- P
a + bx
-
h -77
Q
=
we
,
get
* For if not, by subtracting the equal expressions (1) and (2) from each
other, we get
where x can be neither zero nor infinity, unless a a and the other coefficients
are also zero.
311
(*+!)(** 2) (* + )
_
into partial fractions.
Assume
A
+ J, ^ 3
(* + 1) (ar + 2) (a?
+ 3) x+ I x+2 x+3
we
M
Q
thus get
-TS^5
x2 mil
ml )
(
(.r-f l)(,r2) t
Consequently
x* 1 4
H-
N
* It will be found that P=- , under the same circumstances : for it may
and therefore
M=A + A2 + bx) + A + bxy 1 P(a + bx)-
+ -
r
Q-
l (a . . .
r (a =Q :
A =
m
A> -.
M- A = A 2 (a + bx) + . . A + bx) r ~l
+
P
+ bx) r
-ft l
.
r (a -^.(a ,
-Q-
2 r (a -Q (a ,
m
which gives Az .
Assume
x* + 2x A l A,
which gives
-
and therefore, making x 2 = or x = 2, we get
. m 12
313
_Q
and therefore,
M
x-2
J
___~
x ^
3
2x ---
36
(x+3)(x-2)~
making x = 2, we
o
find
we
x+
get
5
_
x*+2x+-
3
18
5
3
~~</~ ~T25*
B= _ 33
Consequently,
12 58 92 33
A, A,
x- a-
Ax
and assume
M~ Ax + B A3
N x* - 2 a x + a a + (J* + x - a
VOL. II. R R
314
If we make N= (x
z
-2ax + a a
-f/3
2
)Q, and replace x by
(3 J
_ i f when M becomes m + m^J -I and Q becomes
-i we
[
2
We thus find
a?
-4
M= Q=
_
= a? - 4. Replacing -2-
where a? and 2
a;
by V" 1
we = -2, = -!, m = -2, m^ = -l = -6, = -l, and
get a /? 3 q qi
therefore
^zi an dB = |.
If, in the second place, we make M= x, Q=x a
+ 4>x + 5 and
x= 4, we get
we therefore find
x 4 -4>x+5
we assume
M - A.x
~~~+ B, ~~
A 2 x + B2 x+B
A~~ r r
+
P
^T _ (TO + CTI V - 1 ) (g -
315
which gives
we then determine A l
and E^ as in the last case, and subsequently
making
^
}
=M
_.,f
, we get
~-
we then determine A 2 and Be as before, and proceed in the same
manner as far as A and B
r r.
4 fa +5) 25 (x +-5) 25
CHAPTER XL.
Direct and 955. THE methods which are employed for the determination
in
^ e 9 u i v a^ ent forms, may be considered as constituting two great
^thods f
determin- classes,according as they are direct or indirect: direct methods
va?en?
U1
are tnose m
which the transition from the primitive to the equiva-
forms. } en t form, is effected by means of defined or definable operations,
such as multiplication, division, the raising of powers, and the
extraction of roots; whilst indirect methods are generally resorted
meaning.
Example. The series corresponding to the fraction
A + Bx
a + bx + ex*
= A + A >* + A > x * + Ax *
+ &c
317
where^ ,
A lf A 2 A 3 , ... are indeterminate* coefficients, whose
values are required to be assigned: for this purpose we multiply
both sides of this equation, instead of dividing A + Bx, by
2
a + bx + ex , which gives us
A +Bx = A a + A ax + A ax
l 2
2
+ A a ax + A 4 ax* +
3
...
+ A bx + A^x 2
+ A 2 bx + A 3 bx*+
3
...
4- A cx* 4- A ex + A
l
3
cx* + ...
A ia + A b = B,
A2 a + A^b + A c = 0,
A3 a + A2 b + ^!c = 0t,
A
A -
--,
B Ab
a-lS
Bb Ab 2
Ac
~
2
+
a ~a^~~tf~
Bb* Ab 3 Abe
a* a* a*
* The term
indeterminate, which is generally used in such cases, might be
more properly replaced by undetermined or unknown: there is generally nothing
indefinite in its character when thus applied.
f Such from the ordinary process of division in algebra, are Scale of
series, resulting
re
sometimes called recurring series, and the multipliers employed in the formation
of the successive coefficients, connected with their proper signs, constitute what
is called the scale of relation: thus the scale of relation in the series under con
will extend to all those which follow, and the series may thus
be continued at pleasure.
and therefore
Again, assuming
sin a
1 <4
e sin (a
~
COS O
-~ = A sma
+6
b)
Q + eA l
sin (a + b) + eM 2 sin (a + 20) + . . .
- 2 e^ sin a cos 6 - 2 A l
sin (a+) cos o- . -
+e A 2
sin a.
A sin a sin a;
and therefore A =1 :
A l
sin (a + b) 2 sin a cos b = sin (a b),
and therefore -^ = 1 :
we thus find
sina
1 -
e sin (a
2e cos b
7
-~
+ e2
b} .
= sin a + e
.
sin (a
, ?.
+ b) + e*2 sin (a
/
+ ni\
2 6) + ...
e 2
+ 2 b)
sin (a = 20 cos 6 x e sin + ft) e
(a
2
. sin a,
esin(a + 26) -
3 e2
e sin(a + 36) =2ecos& x sin(a + b).
and therefore
s(l - 2ecos& + e2
) = sina esin(a b),
319
957. The form of the series in the last Example is peculiar, Thispro-
and such as could not have been assumed arbitrarily, or without
vfo^fby
some previous knowledge of the relation existing between the an err -
generating expression* and its developement but in the absence sumption :
ohhe^
appear likely to lead sometimes to the assumption of series, series.
whether they possessed a necessary existence in the form as
signed to them, or not or, in other words, whether the opera
:
tions which, on one side of the equation are indicated and not
+X
=
A
X
+A + A, X 4- A 2 X* + ...
Examples
of series
assumed,
with super-
,
1 +X
sin a esin(a 6)
:
where s is the expression which generates the series in the text it is only when :
--
e is less than1 that this reasoning can be considered as strictly correct, mil.
a term, infinitely distant from the beginning, may be equally neglected, whether
it be infinitely small or infinitely great.
rating function.
320
theassump- essential to the correct determination of the series itself, yet it will
series. always, more or less, tend to simplify the process for that purpose,
by lessening the number of quantities to be determined, and
by not encumbering the equations, whose solutions are required,
with unnecessary symbols: it is and other reasons, that
for this
itbecomes important, in all such researches, to avail ourselves
of any considerations which may serve to guide us in the form
which must be assumed for the series.
duct of such a series with a given expression, or the result of any other alge
braical operation performed with it, is a definite symbolical result, when the
series, which is thus employed, though general in its form, is assumed to be
fact at least as difficult to reject divergent series, when viewed either as the
results or the subjects of operations, as it is to admit them when viewed as the
representatives of arithmetical or other magnitudes.
Thus the expression or generating function (Art. 956)
sin a e sin (a
~~ b)
the two first cases, it will likewise express the arithmetical value or sum of the
series continued in infinitum ; in the third case, when e is greater than 1, it
will equally generate the series, though it will not express its arithmetical sum
under such circumstances, is incapable of arithmetical expression: but
/hich,
it may be considered equally equivalent for some symbolical purposes at least, if
not for all, whether it is arithmetically equal to it or not.
n
( 1 + x)
* we should assume the existence of a series for pro
Thus, if ,
powers of x, would be o" but the : first term of the series, for the same expres
- would be 2a n inasmuch as
sion, proceeding according to powers of , ,
milar manner, the first terms of series for cos T, proceeding according to powers of i
inasmuch as cos.i
or Of -_.r, would be 1 in one case, and in the other,
is 1 or 0, according as ,r = or .
VOL. II. S s
322
An equi-
valent
959. A series, in
, ,
order to be completely equivalent to the
. . . , , .
. . . .
for (1 + #)*
is
4
be expressed by
* 4
x 3x* 3 7x 3.7-H.r .
+3 4 +
be replaced by cos ^
27-77-
+
/
_
biquadratic root, which may also /Jl
sin , whose several values are 1, 1, J I, and J I.
Similarly, we find
developement.
other ex-
subsequently reappears in the equivalent series into which it is
developed, it should be
kept in mind that, whatever be the value admitting
which it is assumed to possess on one side of the sign of equality, tlp le
j[jjjgg
it must continue to retain it on the other thus, in the series : sesses in
one term of
, a series, it
v r retains in
*
1
4-
H xv
to possess in -
I
- j-
+ Jx
, the same is
1,
The
consideration of this subject, which is one of the most
A a x* + A a x3 + . . .
x = a y + a 2 y* + a a y 3 +
l ...
measure of an angle in terms of its sine, the series for the sine
sin x = x ~, a o + o Q
" ^
in terms of
1.X.0 i
1 . X . . 4. /;"
its measure.
we may assume, for the inverse series
,r = ^[ sin ,r + ^3 sin
3
x+ A 5 sin
5
x+ . . .
(2) :
x= A sin x+ A3 sin
3
x+ A 5 sin
5
x+ . . .
A 3A A3
3 3 2 5
x3 sin x sin x
"l.2.3~ 1.2.3 1.2.3
x5 A n
5
sin
5
x
1.2.3.4.5 1.2.3.4.5
sm,r =
A 3
=
1
therefore Aa
2 3
J 2 J
+
J 6
= and thCTefore ^
and therefore
sm 3 x 9 sm 6 x
sin * + + + ---
T72^ r273T4r5
The law of formation of the coefficients of this series is not
made manifest by the number of its terms which have been thus
determined, and it
appears that the extension of the process to
other terms would involve the formation of the successive powers
of
A sin x + Aa sin
1
x + A b sin
5
x + ...
325
962. It has been shewn (Art. 778) that all measures of angles The inverse
included in the expressions Qnr + x and (2r+l)7r x are equi- s n:r son ly i i
th
sinal, and we may conclude that all such values of x are equally j^
6 f "
J
included in the equation measures of
the equisi-
appear :
for, if we make x y+-
3
, we get
2
a 2 11 a
3
X3 3
=y + ay + + ,
~j-
2
~a
ab
+ bx = + by + ,
3
c.
327
2 3
a
or, if we replace
- 7
b
,
by and
, 2a
- ab +
-
c by we
q, r, get
y_^-r = 0, (2)
ory = qy +
3
r*. (3)
equation (3), a
is a symbolical root of
equation (4), and
conversely: the same process therefore which determines one
of these roots will determine the other also, by passing from
equation (3) to equation (4) or conversely, the negative roots
of one equation being the positive roots of the other, and
conversely it is usual therefore, in the solution of cubic and
:
process, whenever such a process can be found : they are also Distinction
pression involves the sign J\, and which are called unreal roots>
or imaginary roots.
*
Thus, the equation
t/
3
+ 6y 2 + 9y + 4 -
is reduced, by replacing y by x 2, to
*3 -3x + 2 = 0.
The equation
is reduced, by replacing x by y 7, to
y -i,- 1=0.
The equation
2
3.v 11 j,
91
y +
~T + l6 + m~
-
is reduced, by replacing y by x $ , to
328
equations, have seen (Art. 669), three symbolical cube roots of a, which
are
_ 1\ /- l - l\
>
which are the three roots of the binomial cubic equation
x3 -a = 0. (1)
,r
3
+a= 0. (2)
The root p in
equation (l) p in equation (2), are and
called real roots, in conformity with the conventional language
adopted in the last Article: the other roots, which are affected
or x to be
equal to the sum of the cube roots of s and t,
that such values, whether arithmetical or
3
symbolical, may be
assigned to *J S and t, as will J
satisfy the same equation for
allvalues of q and r: for we thus
get
(2);
3 1] st .x + s + t,
and
qx + r
are identical
expressions, and therefore
3
S J~si = q, and s + t
or
It will follow therefore that s and t are the roots of the given
quadratic equation
-s-
t-L- /( l
r f
We thus get
~2 VU 27
we must replace q by
3
6, r by
---
2/ 3
+c, and add -, which
j
gives
* For if and
s t be the roots of the equation
u2 - au + b = 0,
VOL. II. TT
330
a3 2a*b 2
3 3
ab c /a c abc b
fl_^ +
27 6 3
. /^
Y \27 +
2a
27
2
6
2
The values 969. The values of s and / are derived, as we have seen,
of v/ A and from the quadratic equation
Ut are 3
deter- *__* 4- JL- O
rained by 27
the solu
tion of an
an(j those of 1/s lit from the binomial cubic equations
and of v
equation u.
cquttiiuii of
six dimen
j
sions. v
2
and
3vz = q (2);
o3
or v
6
-rv 3 + = 0,
J=:
ll,
12 V/^_il
4 27\
sf
patible with the conditions which the sum of the cube roots of s the root
and / are required to satisfy for if we
represent the arithmetical extJane^us
:
cube root of
of 1
by
s by <r, and that of t
a 2 (Art. 669,
1, a,
by T, and the three cube roots
Note,) then a-, a a-, aV are the three not roots
^
cube roots of s, and r, ar, aV are the three cube roots of t, and
proposed
one of the first triplet of values
may be combined with one of equation.
the second in nine different
ways, as follows:
f
1. <r + T.
<[
2. ao- + a 2 r.
13. aV + ctT.
I 4. <r + a T.
j
5. ao- + T.
[ 6. a*<r + a 2 r.
7. a- + aV.
j
< 8. aV + T.
9 a o- + a T.
But it should be observed that the equation of condition
9y*y*=q
restricts the selection of the combinations of such cube roots to
971. The three triplets of combinations of the cube roots The nine
x*-qx-r = 0,
an- the
roots of an
X aqx - r = 0, equation of
>-Vf--r-6,
and the nine values of those combinations will be the roots of the
equation
332
xs qx r =
is a factor : and we are not authorized to conclude, from this
x 3 -qx- r = 0,
if - be
,
greater than
L- , then
-
a- and T
cal.
(Art. 968) are arithmetical and real, and the three roots are
(1) T + T,
(2)
-
2 2
ncrtarith-
"
and T are ne g at i ye an d real: in this case there is only
metical. on e real root, which is not arithmetical the other roots are :
imaginary.
Where one If q be negative, then a- and T are real, but with different
8
real which signs there is, therefore, only one real root, which is arithmetical
:
is anthme- or no t
tical or not,
according as r is positive or negative the other roots :
less than $
r
It- o 973. If q be positive, and , then the two
negative. 4 27
Case in
which all roots s and t of the reducing quadratic equation
the roots
only, or at
Se
S
arith
are i ma^ nar y : tne y nave therefore no real and arithmetical cube
metical. roots.
333
we shall find
x = (c +d J- 1)3 + (c - d ^/-"i) 3 :
and if we make
r
it =2 (c
2
+ <f )* cos
|
,
or 2 (c +
2
dj cos ^ ,
or
and , inasmuch as (c
2
+ d 2 )* = ( %- V = the three values of
/|,
x are
* =2 Acos-
g COS 27T +
3 3
These three roots of the equation are real, and one at least,
but not more than two, of them, are arithmetical for their
sum :
is equal to zero*.
=
For * + *" + *" 2
V3 J
cos
5
+ cos
2
334
s+ ^- isin and
manner, the values of the other two roots will be the symbolical
sums of AP and Ap , and AP" and Ap", which are respectively
r
algebraists, in the
solution of cubic and biquadratic equations, is given by Mon-
tucla in his Histnire des Mathematiques, Tom. I. Pars in. Livre in.
335
ing equation is
positive, and two are negative but if 6 be :
a
between 90 and 180, or if its cosine be negative, then - and
%j
x +y + z = a,
xy + xz + yz = b,
xyz = c.
y+2-a x,
or x (a- x) + yz = b,
or y z = b a x + x* :
or x " ax* + bx -c =
3
0.
negative, and the symbolical sum of the two others, which are
imaginary, is
equal to it with a different sign.
of a cubi x3 ax 2 + bx - c = 0,
equation.
bear to evident likewise, from the result of the
its roots, it is
*
The same remark applies to the values of the unknown symbol in an
equation, which results by elimination from any system of equations which are
symmetrical with respect to the several unknown symbols which they involve :
xs + y
3 = 35,
r + y= 5,
where x has two values, which are those of a; and y in the proposed system
of equations for a; and y are obviously interchangeable with each other.
;
X3 + y
3 = 35,
x-y = l,
which are not symmetrical with respect to x and y, the equation resulting
from the elimination of y would have been
a cubic equation, in which x has one real value only, the other two being
imaginary.
t For, if a be a root of the equation
as -aa 2 +ba - c = 0,
then
a s_ ax 2 + bx -c = x - 3 a3 - a (i - 2
a2 ) + b (x - a),
which is divisible by r a :
337
term is the sum of the roots, the coefficient of the third term
(b) is the sum of their products two and two, and the last
term is their continued product: and it will therefore follow
that if the second term disappear, the sum of the roots is neces
(1) Let
*>-^ + ^=o,
Sine* = , ,, we get
!___-
1~27~
and therefore
iVG-
-581 .y
,_
#asi)- 2
S
-581- 733731 1\ = -7-739
//-581-V337311X
Consequently
.
= <r + T = 7?
4"r )
(T
j;" =_( +
2
,_ (IT
+ T) (^z
2 2
a tentative of the
under a finite form, as determined by process tinilinj,
n ,, . , . , the cube
following kind. roots of
binomial
it thus appears that
-
and since a, ft, and y are roots of this equation,
can b **
x - /3, and x -y are factors of it, and therefore there exJ9t under
thus <rPt ll <inite
thUS gCt " -- *)fx-B)(x-y) form.
= *3 - (a
or
Again, since
we get
ff{(a
and inasmuch as it would follow from the hypothesis which we
z z
have made, that x and y are whole numbers, the value of
2x 2 + 2^ 2 is necessarily equal to the sum of the two nearest
2
integral values i, / of ff{(a + Jb) R} and y{(a-Jb)*R}, one of
them being taken in defect and the other in excess : we thus find
* 2
be not a perfect cube, or not resolvible into factors which are
If a b
if a - b
2 = 54 = 2 x 27 = 2 x 3 3 then R = 4 and x 2 - y 2 = 6
, :
-
these expressions to and from each
other, we get
/+/+2c
and x = v/0
.
2a?
2
=
and therefore
y( nearly: say 175.
We thus get
88,
= 49 and x = -
=39 and y =
340
Therefore
Let # 3 +8tf-9 = 0,
(2)
and therefore
(3) Let
t = -7 -722 = -11.6904,
T = - 7ll.6904 = - 2.2696;
o- + T = - 3.5904,
(4) Let o;
3
-]8cr 2 +87^-70-0.
If we make x-6 = u or x = u + 6, we get the transformed
equation
u 3 -21u + 20 = 0,
2 10
If we make cos 6 = .
3
= ; and therefore
/. 75
V 27
COS (TT
- 6)
= ;
7!
we find
and therefore
= TT - 57. 19 12"= 122. 40
. . 48".
\/ 3
log 2
= .3010300
ilo7 .4225490
.7235790
- = = 9-8784812
log cos o log cos 40. 53 36" .
Iogw =log4
=
.6020600, omitting 10.
ff. g 7T 6
To find the value of u" = cos
A/| 3"
.7235790
l
og cos ?^1? m iog cos 790. 6 24" = 9-2764210
.
= -0000000
logM"=logl
Ml
.7235790
" = .6989700
-Iogu = log5
we get
uv = 2x + 3.
we thus get
r
4 - 14,r
2 - 12,r = 0,
or dividing by x,
x3 - 14,0: -12 = 0*.
o3 r2
it is not difficult to prove that in this case is necessarily less than j. This
-^=
3 2
O T
In this case *- is
greater than ,
and the process followed
27 4<
x" =- .9118,
"= -
x 3.201 5.
PBxPE = PC,
we get
PC PBxPE
PD
but since
PC 2
= AP 2
+ AC 2 - ZAP x AC cos (TT - 0),
we get
= x2 + r
2
+ 2 r o: cos
hich gives
or x3 2
3r x 3
2 r cos 6 = 0,
3
x\ - Sx\
- 3 f - - 2 cos e = 0.
) J
*
See Appendix.
344
Consequently
s = cos + J - 1 sin 0,
t = cos 6 J- 1 sin 0,
x"
= 2 cos 100 =- .34740,
r
negative in problem of the trisection of an angle has been reduced, but for
metrical
wn i n no construction, which can be effected by the rule and
solution of compass only, has been hitherto discovered and it is presumed, :
VOL. II. Xx
CHAPTER XLII.
biquadratic
equation. x* - ax 3 + bx 2 - CX + d = (l),
therefore
3
3a*u* a u a*
x* = u* + a u3 H h TT + -
8 16 256
3a 2 u 2 3a 3 u a4
~ax 3
-au* T*-- ^7
4 16 64
2bau ba 2
+ b.ar + bu + - - + ^~
4 lo
ca
ex CM
+ d + d.
and if we replace
3a 2 3
-
.
b
,
b ~~-ba
a
347
and
4
3a ca
_ba? _
we get
-qu -ru-s = 0.
4 2
?4
(2)*
shall now proceed to apply one of the most common and suc
cessful of the expedients which are there exemplified, which
consists inadding or subtracting such terms to or from both
their members, as may make them complete squares.
* Thus the
equation
43 75 851
The equation
t/
4 - 20i/
3
+ 148t/
2
+ 464* + 480 =
x4 - 2x 2 + 16x - 15 = 0.
The equation
y
4
+ 7t,
3
-19 2/-20 =
becomes, by putting x ^
for y,
(x
2
+ w)
2
= (2 u + q) x
2
+ r x + u2 + s (2).
or
2w 3 + qu 2 + 2su + qs = ,
or
(3).
+ - =
+!+* (>
(3).
x2 - J%u +q . x+ u- Jit 2
+ s = (4),
or
x2 4- J% u + q . x + u + Jit 2 + s = (5),
according as we
use the upper or lower sign: the solution of
these two quadratic equations will jjive us the four values of
x in the proposed biquadratic equation t.
* For if (ax + 6)
2 - a2 ! 2 + 2afe.r + b
2
, then 4a 2 x 6 2 = (2a&) 2 .
j-
It was F. Ferrari of Milan, a disciple of Cardan
who, about the middle of
s,
u3 + 3u 2 + 11 u- 255 = 0,
and
,r
4
-25.r 2 + 60.r-36 = 0.
3
w + 12 . 5w +36w =
2
0,
and
x~ + 5x -6= 0, whose roots are 1 and 6.
2
For it
appears that 25 x 60x + 36 is a complete square, and
.r
2 - 3.r +2= 0, whose roots are 1 and 2 :
and
xz 4- 3x 18 = 0, whose roots are 3 and - 6.
x - 2
4>x + 3= 0, whose roots are 1 and 3 :
and the product of the first and second of which is equal to 6."
This problem leads immediately to the biquadratic equation
x
4
+ Qx 2 - 60* + 36 = 0,
and
.r
2
-f 4,r- 12 = 0, whose roots are 2 and -6.
2 2
In this case, we add -Qx + 20.25 to 25,r -60,r + 36, which
2
produces the complete square l6.r -60,r + 56.25.
The roots of this biquadratic equation, as well as those of
its
reducing cubic, are all real, and the different solutions cor
respond to the three different ways of resolving it into quadratic
factors, involving three different combinations of the roots; it
is
only when all the roots of the biquadratic equation are real
that all those of the reducing cubic are real also.
biquadratic
com pl ete: f r ^
su ch a biquadratic equation be put under the
equations, form
all whose
px = qx + rx +
3 2
terms are x* S ; (1)
complete.
we shall find
*
2 - + = x*-px 3 + (- + 2
+ M
2
2u)x -pux
1
I ,
2 MJ
4 3
which becomes, by replacing x px by qx* + rx + s,
or if
<2 u-9 -
x+ u -T4),
351
3 __ 2
53 u 225 _
IT ~8~~
l 2
~2
The two numbers required are, therefore, 1 and 2.
+ (a + (3 +7+ B)
X 3 + (a/3 + ay + aS 4- /3 7 + j3l + 7 S) X
2
+ (a/?7 4-
afid + aye + /3y%) x + afiyd
= x4 - aj;
3
+ bx -cx + d:
2
or the sum of all the products of the roots taken two and two :
C = aftj + a/33 4-
ay% + /3y%,
or the sum of all the products of the roots taken three and three:
is zero*.
x+y+z+v=a (1),
xy + xz + xv +yz+yv + zv = b (2),
xyzv = d (4).
y+ z + v =a x.
x (y + 2 + v) +yz+yv + zv -b,
and therefore
= x (a = b ax + x
2
yz y + 2v b .
4. i) x]
and therefore
c-x (b-ax + x ) = c - bx + ax - x\ 2 2
yzv =
=-= C bx 4- CX 2
X
and therefore
* The statement of this proposition supposes the terms of the original equation
a
4 3
+ px + 2
qx + rx + s = 0,
we should have found p equal the sum of the roots with their signs changed, and r
equal to the sum of their products three and three with their signs changed.
353
a/3 7+ a/31 + ay I + (3 76 = C,
aft^b = d*.
984. If we resume
the consideration of the quadratic factors Value of u
(4) and (5) in Art. 980, into which the equation in terms of
the roots of
w + *Ju a + s = a/3,
u- Ju 2
+ s = yt,
_ a/2 + 73
and there are no more^ : we thus discover the reason why the
values of u, upon which the determination of the roots of the
VOL. II. Y Y
354
which a
factors*, by whose solution its roots are found: we shall now
quadratic , . .. , , L . .
-qx -rx-s =
4 2
x (2),
z
x*+ ax + b) x*qx rx- s (x? ax + a 3
b q
2 2
b
(a q)x +(abr)x s
(a
2
-b ~q)j; + (a ~ab-aq)ju + b(a 2 -b- q)
2 3
b(a*-b-q) + s = (4).
* If 2
+ ax + b is a factor of x4 2
rx continues to be a factor for
i 7a s, it
*-l- f -*^5j-
The substitution of these values in equation (4) gives us
are
ax +
* Lib.
III. t Ibid.
356
x" + 4>x + 2 = 0,
x a - 4<x-3 = 0,
J* + Jt +>Jt"= (
+ /?) + (a + y) + (0 + 7)
= 2(a + /3 + 7)=-2S,
V*
-
Jf-Jt"=* + ft- ( - 7) - (ft + 7) =~ 2 7 ,
u*-2qu*+ (q
2
+ 4ts)
u - r
2
= 0,
x*-qx
2 - rx-s = (1),
**
side, get
(3).
4f
solution.
9, 16 and 25.
358
Therefore
- +2 + = 6,
3 5
w 3
-156w- 1600 = 0,
- 8 QJ
whose roots are 16, 4- 1, 8 6 J 1
Therefore
= i {4 + (1 +3 V^T) + (1 - 3 V^T)} = 3,
!) + (1
- 3 =-2- 3
1)}
or
(a + /3)( 7 + 8), (a + 7) 08 + 8), (a +8) 08 + 7),
359
or
07 -
a + /? - (7 + g) a + 7 .- (/9 + 8) a + c - (/3 + 7)
multiple values shall express all the roots of that equation and no
more: such equation being supposed to possess the most general
form of its order, or such other form as is deducible from it by a
enable us, in all cases, to determine its roots to any required |jjj?Jf
degree of accuracy, when its general coefficients are replaced equations.
*
These combinations form the basis of a very ingenious solution of a biquad
ratic equation which is given in the 1st volume of the Cambridge Mathematical
Journal, a publication which is justly distinguished for the originality and elegance
of its contributions to almost every department of analysis.
t See Waring, Meditationes Algebraicee, p. 139.
360
always kept in mind that the general and the numerical solution
of equations are problems of a different nature, and almost entirely
indifferently all its roots, and no more for no such formula has
:
~
990.
, -ill
IN the absence of general methods of solving equations
ot degrees superior to the fourth, there are
h ch
i
many cases where
Recurring
and other
equations
y
equations present themselves with particular relations of their ^ \
coefficients, which admit of reduction to binomial and other sed to
x 5 + px* + qx 3 + qx 2 + px + 1 - 0,
and
1=0,
the first of which is of odd, and the second of even dimensions.
by x + 1, we get
x* + (p-l)x 3 + (q-p + l)x 2 +(p-l)x+l = 0,
= 0.
VOL. II. Zz
362
1=0
dimensions cn ,. ,
3
this equation by x , and combine together the first and last
term and also those which are equidistant from them, we shall
get
and if we make
1
x+ - = w,
we get
and
u 3 -3u + p(u
2
-2) + qu + r = 0,
or
-
2p =
+ pu 2 + - +
?j*
(q 3) u r 0,
u4 + pu
3
+ (q - 4) it* +(r- 3p) u + s-2q + 2 = 0.
The roots
993. further be observed that whatever be the value
of recurring
jt may
equations of u m the depressed equation, there are two values of x,
corr esporiding to it, in the primitive equation, and which are
values are the reciprocals of each other: for if u = a, we get
the recipro
cals of each i
other. x H = a,
x
and therefore
363
or
x= - + / I
- 1 I or
tion, then - , T ,
- ... are also roots of it: such equations have,
a b c
on this account, been sometimes called reciprocal equations.
2
r
x* + px 3 + qx 2 + rx + = 0,
P
where the last term is the quotient which arises from dividing
the square of the coefficient of the last term but one by the
square of the coefficient of the second term: for if we divide
2
the several terms of this equation by x and ,
if we make
r
x H = M.
px
we get the quadratic equation
u* + pu + q--- = 0.
Of the same class is the equation
x + px + qx + rx + qsx + ps*x + s = 0,
6 35 2 4 3
r>
equation
f+Py + Q-^=o.
Thus, if the proposed equation be
comes
Other 996. Other equations, which are not quasi recurring equa-
\?hoseroots ti ns > ^ ut wn ose coefficients possess a particular relation to each
areexpres- other, may be solved by formulae similar to those which occur
finite for- in the general solution of a cubic equation. Thus if,
mulae.
then
x"-ax =-s 5 5
,
*
Cambridge Mathematical Journal, Vol. I. p. 254.
365
and
and therefore
u* = a*+ + 5,
- a + 5s (u
3 - 3su) + Ws*u,
which becomes
u 5 -5su 3 + 5s*u- a = 0,
2 5
a
- /fa
/ \~I~ S
\
]
res pectively, and if a be a root of --
x 1
= 0, then
THEORY OF ELIMINATION.
symbol, when the others are not assigned, but are considered
as perfectly arbitraryand independent of each other: in other
words, the dependent symbol can only become known and de
termined, by assigning specific values of the several symbols,
in terms of which it is expressed.
(as is most commonly the case), that the first letters of the
alphabet a, b, c, denote quantities which are known or deter
minate, then x is the dependent, and y the independent symbol :
367
ax + by c 0.
The term function is not only used to denote the de- Presumed
1000.
The exhi- 1001. When more than one unknown or indeterminate sym-
bitionofthe .
symbol
w ^
res P ect to an y one f them, and therefore the exhibition
upon the of its dependence upon the others must be effected by the general
equation,
methods which have been taught in the preceding Chapters,
and must be limited by the limitation of those or other methods :
x=9-y,
where the actual dependence of x upon y is exhibited: in a similar
y*
2
+ -jy-JL^QQ-J y
leads to the equivalent equation
,.Zi*8Z-jt:
2 2 y "
and
to the equation
thine tothe
^ one svm kl upon one or more other indeterminate symbols
determina- leaves them necessarily as indeterminate as in the primitive
tion of their . , , .
i -i
* Such is the condition, considered in Chap, ix., which restricts the values of
the unknown or indeterminate symbols to whole numbers.
369
Chap. v.
gives us
_ 100-lOy
~~3~~
= 700-70y,
97 y = 679,
#=7,
and therefore ^M = I^ = = ia
Vot. II. sA
370
and therefore yy = 3 or - .
2
7
The corresponding values of x are 5 and - .
(3)
^-4^- i-oL
=
x +y + z- 12 0j
respect to x, we get
_ 5y -4>E + 24
8
or 13 + 4z - 72 = 0.
and the processes which are employed for that purpose, which
are very various in form, are generally identical with those em
... ..,.
by one: and a very little consideration would shew that if n was equations
diminished
the number of the primitive equations involving any number of by the
is
A - A 2 = 0,
l At - A3 = 0, A -A4 = 0,
l
...... A - A n = 0.
l
thus,
A, -A ^=(A -A )~(A ~A.,)
3 l 3 l =0,
A3 -A ^(A -A )-(A -A
5 l 5 l 3} =0,
A n _, - A H = (A,
- A H ) - (A - A n_,} m l 0,
in other words, the equation corresponding to
any such com
bination will always arise from subtracting from each other some
two of the (n 1) equations which resulted from the first series
of those combinations which it was considered most proper and
convenient to select.
(n + m)~(n + m - 1) (m + 1)
1 72777 .7~
in number : the values of the unknown quantities which would
thus be obtained might be different for different combinations,
and therefore m out of the n + m equations are at least super
fluous, if they are not inconsistent with each other. (Arts- 398,
402, 403).
Equations When we speak of equations as independent of each other,
dentofeach we m ean such as severally contain conditions for the determina-
other. tion of the unknown quantities which they involve, which are
not supplied by the other equations, nor derivable from them :
case, the two equations will have a common measure (Art. 1005):
factors, we institute
upon and E2 , arranged according to
E l
ing values of x, and we shall thus be enabled to form all the sets
or pairs of proper roots of the equations E 1
= and E a
= 0.
and so on, and therefore the the first case, the two
last divisor in
last divisors in the second, and so
become equal to zero, on, will
for the corresponding
proper value of y these cases and their :
and the process instituted again with the quotients which result
from its exclusion, which, when made equal severally to zero,
become the independent equations which are the proper objects
of consideration.
(1) ax + by-c = = E l ,
ax 4-
by
- ax + (3y
- 7
c)
(a/3-ab)y-(ay-ac) = F.
376
-
If F=0, we find y= CLO
-j- t
and therefore ax + by c
Gfj
= ax + fay~ ac\ - =
,
b {
-, )
\aj3-ab)
c 0, which gives us x =
j3c
5
by
afi-ab
- .
(2) x -y = 5
13
*-13 (x-y + 5
-(y-5)x+y*-13
2/- 10^+12= Y.
x -f 2 - 5 = 0, and therefore x = 3.
(3) tf
+# =fl or x +y -a = = E 1 ,
x +y = b
3 3
or x +y -
3 3
b = - E 2,
2 2
-(y-a)x -(y-a) x
(y
- a) 2 x + (y - a) 3
377
If Y=0, we find
3
a a //4>b-a \
y = 2n +
f(4,b-a*\
/ (\ - 120 and %= -- /(
V
A
/
) >
.-.
2 V \
)
12 fl /
,
- 3
a 3
\
and *.* = + //46-a-).
/
\
L2a r
) ,
2 V
A 1
\ 12 )
3 3
by dividing x + y by x + y and 6 by a (Art. 976 and Note) the :
(5) xy + xy*-12 = =
x + xy 3 -18 = =
-
12y*+30y-12 = Y=0.
Therefore
/-|Ul=0.
If ^ = 2, then x = 2.
If # = i, then #=16.
In this case, the coefficients of the first terms of the divisor
3
and dividend have a common factor 1
+y and
}
therefore y(l +y )
VOL. II. 3 B
378
(6)
(y-l) x 2 -3
y(y-l) x*-
- 2
7)
l6y*-96y + 144^ = = F.
s
2
This final equation is reducible to the form l6y (y 3) =0:
if ?/
= 3, we have .r = 2: but if y = which is also a value of y
in F=0, we no corresponding value of x: for under such
find
circumstances and E2 become
E l 7 # + 2 and - x
2
3x 2,
which have no common measure: but E and yE 2 become, under l
foreign^
Y Q, would necessarily contain the values of y which severally
the final
give common factors, not merely of E and E2 in their original
l
only, willnever affect the degree of the final equation: for the
values of x which belong to the factor whose introduction
may
be necessary in this case, will make the first term of the divisor
equal to zero, and will leave therefore no term in which x exists :
(7) *9 + (2y-7)*+y-7y-8=0,
x s + (2y- 5) x +y*- 5y - 6 = 0.
If we subtract these equations from each other, we get
-2,r-2?/-2 = 0.
Therefore
J + (2y - 5) x + y 2 - 5y - 6 (x
(y-6)x+y*-5y-6
(y-6)x+y*-5y-6
equations.
a common factor, and therefore x and y are indeterminate ; but
if we take the equations
x+y-6 = 0,
x*-y* + lx = Q,
r
quantities, to the one which has been employed in the preceding examples:
thus let it be proposed to find the values of x, y, and z in the
following equations :
381
x-y+z- 3=OJ
+ =
xy-z* 10 1;
-y + z-3)yx-z*+10
zx Sx
or -z-
If we now proceed to eliminate y from F, = and Y = 0, we
2
y*-(z-3)y-z +10 =
z
y=-: 4 4
the substitution of the corresponding pairs of
,,8 or -S or
4
3 s
9-r 31 x 147 45
"-I7 -T +
^*-T
3
9V 3iy 147 45
(a + \a )x + (6 + A6 )j/
= c + Ac :
(4)
b
we make + \b =
and if b or
\=j-, we get
ba \ be
or (ab a b)x =b c bc
f
.
(2)
(3)
tiplied by ca ,
(a i C*
- 2 Ci)a? + (6. c.
-
b^y - (c<, /^
- d ky ) = 0. (4)
383
and
a x + Vy = c , (b)
b c-bc
3
ab -a b
The pro- 1015. If the coefficients, one or more of them, involve one
elimination or more unknown symbols, the equation of condition (of) will
Ua
Sons of
be tne final equation to which the elimination of x and y, by
higher this process, will lead: and it will be found to be very easy,
0, (1)
0, (2)
ax 3 + bx2 + cx = 0, (3)
3 2
and if we consider x , x , and x as distinct and independent
symbols, like z, y, x, the number of equations, thus formed, will
exceed by 1 the number of symbols to be eliminated.
z
If we eliminate x from equations (3) and (4), we get the
equation
(a b -ab )x 2 + (a c - ac )x = 0, (5)
or Ax + Bx =
z
0,
replacing a b ab by A and a c ac by B.
2
If we now from equations (5) and (1) in one
eliminate x
case, and from equations (5) and (2) in the other, we get
*
Cambridge Mathematical Journal, Vol. n. p. 232 and 276.
385
-S(y -l)x+y - 3 y - 6 = 0,
2 2
2,r
-
252/ - 440/ +
4
430# 1 50 y +112 = 0,
It is this reduction
of the general problem of elimination
to that of eliminationfrom equations of the first degree only,
which gives additional importance to the latter problem, and
to the research of the most easy and expeditious rules, both
for the expression of the unknown symbols which they involve,
or for the formation of the final equation, under its most simple
form, which results from their elimination, when the equations
exceed by one the number of such symbols the discussion of :
ax -f
If in the expressions lf z, N N N
3 we replace k l} k z , Jc 3 by
by d lt k z by dz k3 by d3 then we
, , shall find, from the three first
_-n,n
and substituting these values in the last equation (4), and sup-
pressing fractions, we get
(#4;?! + 64 w2 + c4 w s - d4 D) u-atNi-btN?- c4 N3 + k4 D = 0,
and therefore
_ciiN
~
+ l
b 4 N<i-{-c4 N3 -k 4 D
d4 D
387
and if N ,
N" f
with their proper subscript numbers be suc
N",
cessively taken to represent the values of the numerators of
the expressions for any other three amongst the four unknown
quantities in the three equations (l), (2), (3), namely, of
g jy w + 64 jyg
w+ c N" f
- D"r
_
~ 4 t 4 3 k,
a.Pt + ln P 2 + c n P3 +... ln P n. r
-kn g
expressions
for every additional unknown quantity or additional equation,
for the n
unknown Again, the number of terms, when they are completely ex-
n^a system halted, in the numerator and denominator
of the expression for
ofnequa- xn is 1 x 2 x 3 x n, or is equal to the number of permu
. . .
just investigated:
are formed
we find,
...
by the general law which we have
in a similar manner, from the n
than the
number
equations.
of
last equations,
D~~D
__
or ND
which is the final equation required.
a x + by c = 0,
a x +by - c = 0,
al c"- aV c -a bc"+a r
b"c + a"bc - a"c = 0.
1018. WE
have shewn in a former Chapter, that the roots The gonio-
of the binomial equation
S ionof
--i-o
i n - 1=0
are completely expressible by means of the different values of does not
formula
COS -
2rvr
-f J i
1
.
Sin -
2r7r
e p
solution
al
so
, properly
u u called.
0, 1, 2, 3, ... (r-1);
the same roots recurring, in the same order, if additional terms
of this series are taken.
But this solution, though complete, is not algebraical, in the
strict sense which is attached to that term, for reasons which
have been explained before (Art. 990) ; but it will be found that
the symbolical properties of those roots, and of the cyclical
periods (Art. 722) which they form, will be sufficient to deter
mine them algebraically or rather arithmetically in all cases.
For this purpose it will be necessary to investigate some
propositions, which are equally applicable to the roots of bino
mial and other equations, and which will enable us to bring
the various methods of solution of cubic and biquadratic equa
tions, which we have already considered, under the operation
of a common theory*.
equation
#n 1 =0 assumed
form of its
root
* The substance of this theory is given by Lagrange, in notes xui. and xiv. to
the Resolution des Equations Numeriques.
392
1 x 2 x 3 x . . . n:
for if we suppose
1, a, a2 ... a"
1
1 x 2 x 3 x ... (n - 1)
th
groups of n roots each, whose w powers are identical : for if
~l
t = 2
p + ap! + a p 2 + ... an p n _ ly
then also
at = ap + d p + a 3 p 2 + l
l
. . .
p n -i,
o?i = a z p + a? pi + a*p 2 + .. .
ap n_ lt
~l 2
an t = a"
l
roots of the reducing equations in the solutions of cubic and biquadratic equa
tions : should be kept in mind that the propositions which follow have
and it
reference to the specific form assumed for t only or to such as are reducible to
it, and therefore do not close the inquiry as far as other forms, which may be
th
1
the number of different values of is -
only the n part of the The lowest
dimension
number of and consequently the values of
values of t, are 1 ^ncrali
expressible by means of an equation of
reducible.
dimensions.
We may assume, therefore,
1, a, a2 ... a"
1
1, r lt r a ... /_!,
6 = A + aA + l
2
a A* + ... a""
1
A n_,
by
proved that all symmetrical combinations of the roots of an equation are expressible
in terms of its coefficients, and are therefore given, when the equation is given.
VOL. II. 3D
394
that
or,
+
Again, since
we get
or
"Je + rr 1
y0i + ^n " 1
A + - - -
^:l ^Ci -
^ i=
"N/^O
+ n"- 1 A + r,- 1
~~^r
A + ^:{ yg^
Pn-l
=
For
395
the n roots of
x n -l =0:
and it is in this sense, and in this sense only, that the general
resolution of equations may be said to be dependent upon that
of binomial equations of the same degree.
But p 2 +p 2
and 2pp , or A and A^ are symmetrical com
binations of the roots, and are expressible therefore in terms of
the coefficients of the proposed equation ; we therefore find
We thus get
eo =p
e^
and therefore
~_ JQo
~
p
Again, inasmuch as
Q = A + A, + A.,+ ... A n^ =pn ,
we get
A =p n - A!- A*- ... A n_,:
and if we substitute this value of A in
= A + a A + a? A? + ... a n ~ A n ^
l
l
,
6 t =fT + (r a - 1) A + (r,/ - 1) A. +
l . . .
(r^ - 1)
if we make
t = p + up +l
Ave get
where
397
and therefore
We thus get
Art. 9^8.
*
The complete when the coefficients are q, v, are us follows
:
/>,
expressions
P =-3/77 -9r,
+ 81 .
398
l x"-
2 - . . .
= 0,
when n is a
prime num- , . i /
number, we
i
where n is a prime find
ber, isde-
analysis,
upon an which is given, as we have shewn in Art. 1019j by an equation of
equation of
X X3 " x2 x3 x ...
?n -2)
1 (n-l)
dimensions.
Dimensions : but if we form the equation
u n-l __ p u n- 2
+ p^ u n-3 __ . . . ^ p ^
= Q,
- = - -
q (n 1) 1 x 2 x 3 x ... (w 2) (w 1),
it follows that
q
= 1 x 2 x 3 x ... (n - 2) :
it
appears, therefore, that whilst the values of 6 are expressed
by an equation of Ix2x3x ...(n-l) dimensions, the values
Ai t A^ ... A n _i, in terms of which it is expressed, will be de
pendent upon an equation of ] x 2 x 3 x (n 2) dimensions. . . .
1023.
In the case If n be 5, or if the proposed equation be of 5 dimensions, the
tion^f the
va ^ ues f ^ wn depend upon
"l an equation of or 24 1x2x3x4
5th degree,
dimensions, and the coefficients of the equation which expresses
the values of A 1} A^ y A 8 A4 , will depend upon an equation of
1x2x3 or 6 dimensions, exceeding by 1
degree the dimensions
of the proposed equation.
When the 1025. If the number which expresses the dimensions of the
which ex- proposed equation be not prime but composite, then it will be
presses the found that the process under consideration may be very greatly
399
simplified: for if n=>qm, where m is not greater than q, then the dimensions
of the equa
series
tion is not
2
1, a, a ... a""
1
prime, but
Of the roots of composite.
*"-] -0
p 2m + ap 2m+l
and if we make
pi
we shall find
t = X + aX l + a 2 X2 + ... a"- 1
^.
Considering therefore X Xlf X , 2 ... Xm_ l as the roots of a
new equation of m dimensions only, we shall find
= tm = A
(a- -
1
...
Ill
400
is
1 x 2 x 3 x . . n
(1 x2 x 3 x ... q)m
1, a, a2 . . . a" 1
.-<?)"
Again, the number N of different values of is
only the ]
ith
part of the number of different values of t
(Art. 1019), and
m
therefore
Lastly, inasmuch as
B= rr+(a-l)A + (a*-l)A + l 9 . .
(a*
-
m 1 m (m - I
)
_ Ix2x3x...
m
~ra(w-l)(lx2x3x ... q)
1027. Let us apply the formulae in the last Article to the Applica-
solution of the biquadratic equation preceding
theory to
x*- px 3 + qx 2 rx + s = (1). the solution
ofabiquad-
In this case we have m = q = 2, and or the root of equa "
[f^
x2 - 1 =0,
which is different from 1, is - 1 : we thus find
/ = X l
+ a X 8 , where X l
= p + p z and X2
= pi
+ pz :
and
u3 -Pu*+Qu-R = (2),
Inasmuch as
M /3
-PV+QV-fl = 0, (3),
=p^-2A = p
2 /
l l
-4,q + 4 u
f :
Again, since
Xi = P + P*,
we may consider
x a
-X x + \ = l
x L+ jW-4x)
f= and ^g.
and similarly for p l and p 3 .
equS?on
tlC
tion wmch expresses the triple values of 6, we should have
*
processes of solution of biquadratic equations, which we have considered
The
in XLII., are equivalent to the corresponding processes for forming the
Chap.
reducing cubic equations, whose roots are such symmetrical combinations of those
of the proposed equation as have triple values only: such are (p + p 2 ) (PI + PS)
2
or (pp 2 + p l p 3 ), or (p + p 2 p 1 ,/o 3 ) , and others which may be formed: such
processes are always greatly simplified, by supposing the coefficient of the second
term of the biquadratic equation to be zero thus the equation, whose roots are :
(p + p 2 P!
- p 3 ) 2 , investigated in the Article which follows, when all the terms
of the proposed equation are complete, is
it is only in this latter form that the particular process of transformation (if so
it may be termed) which is commonly Jcnown as Euler s solution (Art. 986), is
practicable and it is
: to this form of the original equation exclusively that Des
found
if we farther make u + v = pp 2 ,
and u - v = p p^, the
l
two quad
ratic factors of the proposed equation would become
x* - ^ (p + t)
x 4- u +v = 0,
x 2
^ (p ) x+u- v 0,
jj _
* 1 - 2 - - 2=
p x + {2u + (p* / )} a? (pu + lv)x + u* v :
pu 4- tv = r,
If Q lt a ,
63 be the three values of 6 in this equation, we
get
Kncyclopa-dia Uritaniii
...
Article "Equations" in tin-
Ivory.
404
0, x0 2 xd 3 = (p
3
and therefore
c should find
dimensions. t
p + ap + a 2 p 2 + a 3 p 3 + a*p t
l ,
2 3 4
where a, a ,
a ,
a are the roots of the equation
x 5 - 1=0,
2 3
which do not change upon the substitution of r", r &c. ;-
, ,
406
n ~2 3
for r: tor if we replace t
by a t, a"~ t ...at, successively,
we get
~3 ~2
an t = a"~ 3 r + an ra + r
2
4- ... a*"
4
^""2,
B "2
at = ar 4- a2 r 4- a3 r 2
4- r ;
"- 2 - = a"-
2
*,
-3 /
=a t,
ra""
2
+ a r + a2 r 4- ... a"~ ra"~
2 3
= a/ ;
r- =(ay- = (a0-=...(a^r
are the same, whatever be the term of the series
+ aA + l ... an A n_ 2,
by
P+Qr + Q,r 2
+ ... Qn_ 2 r"-
2
(i),
r + ra + ra
2
+ ,.. t r rt"-
2
=_ 1
eo = P-Q + (p -Q l l )
+ (P2 - Q)+
2 . . .
(Pn_ 2 _ Q n _J,
0, =P- Q+ (P,
- Q,) + a,
2
(P2 - Q)+ "- 2 -
ttl 2 . . .
ai (P_, Q_*),
= P- Q+ (P,
- QO + a./ (P 2 - Q.) + .,
"-2
a2 - Q B _2),
2 2 .
(P^,
C, = P - Q + - QO + aL 2 2 - Q 2 +
n _2 (Pi (P )
. . .
::S (Pn _, - Q n _ 2 ),
where i
= a, a 2 = a 2 a 3 = a 3
,
. . . a n _ 2 = a"~2 .
We thus find
r =
n- I
rf,
_
M- 1
m m+1 2
a 8 /*""" +
"
+ r<i + ar 4-
1
+ fafe-D" + ar" (v
4- . . . a"
1
/-" ;
and if we assume
X =
similarly^ by replacing r by r
2
, X becomes X3 X^ becomes X4 l ,
.
X^z
. becomes
. X and similar consequences will follow from
l
:
8 4- 3 X, t ;
2
when r is
replaced by r , it will become
~l
X 3 + aX 4 +a 2
X 5 4- . . . am X z
= a^t :
and when r is
replaced by r*""
1
, it will become
Xm + aX l + a 2 X2 4- ... a"
1 1
Xm^ l
= at.
6 = r = (a t}m - (a 2
*)"
1
= . . .
(a"
1 1
/)
^4 , ^4 U v4 2 . . .
^^_ 2 will be rational expressions involving X,,
X Xm in such a manner as not to change, when Xj is changed
2 . . .
Oo P
= + Q** + (P + l Q l s + ...... + (Pm_,
#1 = P
-1
Q + S + a, (P, + Q lS) + . . . + aj"
1
(P^
2
=P Q + Q + 2 (P,
s +a
=
X3
;
m-l l
1
X l
= r +r
2
=r+-: (Art. 532)
X2=ra+ ^>
If we make therefore
r"=cos ^^+
n
\dsin -^-,
n
we find
X l
=2 cos
n ,
X2
= 2 cos
n
X8 = 2 cos n
VOL. II. 3F
410
tf-1 =0,
where A A 1} A% ,
A q-l are reducible to the form P + Qs,
. . .
x 5 - 1=0.
of the or
equation 2 4 3
a;
5 - 1=0. r> r r > *
or that root of a;
2
1 = 0, which is different from 1, we find
t = Xi + a X% ,
where Xi = r + r and 4
X
2
=r + 2
r
3
.
Consequently
where
A = 2X,X, = 2
l (r + r2 + r
3
+ r4 ) = - 2 :
and A = s
z
-A l
(Art. 1022) =3: therefore
411
X ^V e
oJ ei
=
-
2
But inasmuch as
.3d = r + r
4
and X =r + 4
r
3
,
/ = r -f a r4 ,
and therefore
= /
/2
= r* + /-
3
+ 2a = X, + 2a = X 8
- 2.
2 2
X g
X.
If we replace A r
j
and X 2 by. their numerical values, we get
-1-J5-J(- 10-2^/5)
n^r~
These will be found to coincide with the values given in
Arts. 707 and 814.
412
I)
~3
)
r
I
a
)
r e ) r*3 r s
x ,=>
1
we get
/ = X, + aX 2 ,
where
X 1
X 2
Therefore
d = t*=A + aA l ,
where
J, = SXtX, = 2 (3 + r + r
3
+ r
2
+ r
6
+ r
4
+ r
5
)
=2 + = 6 + 2* = 4, since * = -^ 1
,
(3 *)
and therefore
A = s*~ A = l-4, = -3. l
We thus find
= -3 + 4a,
and therefore
0: = -7.
Consequently
V ~~
^-N/^l.--
2
Again, since q
- 3, if we make a a root of the equation
x T1 = 0,
413
and consider r, r
2
, r*, which are involved in X lt as the roots
of a new equation, we shall find
and therefore
e> = t
3
=A + aA^ + JAv,
and therefore
We thus get
and therefore
3 5 6
The values of r , r , and r may be found in a similar manner.
1C 14 3 7 4 !2 H "
414
and assume
where
X l
= r + r+ r ls
+ r
15
+ r
16
+ r
8
+ r
4
+ r
2
:
Therefore
where
and
We thus get
and therefore
= _ 2 5615528 .
Y A
= A! Y2 = ms =
4
(2m) = 4mxm =
2
for there are (n 1) terms in their product, forming
^
*+_Sr!. ft
and therefore
415
F2
= r
9
+ r
15
+ r
8
+ r
2
.
We thus get
and therefore
Thus if n = 5, we find
If n = 13, we find
- 1 yi3
we shall find
and
B = ms + 2m + 1 = - m + 2m + 1
71+1
If n = 11, we find
where
A =2Y
l l Yt =
and
Therefore
^
13-X/17 9 _
1= -2~~
and also
=
X 1
2 4
_
4
Z4 Z6 Z
, , fl ,
Z7 Z 8
, :
taking the same value of a, we assume
here
We thus get
P-P*-Jl
where
J", = 2Z,Z, = 2 (r 3 + r
5
+ r
14
+ r
12
)
=2 F 3,
and
* The value of A" is more readily found from Zf + Z2Z than from
j
2 - A \.
417
Therefore
0", = 4+y,
and
= 1.864944,
Yl ~J e" l
=. 184537.
^
= 1.478018,
c =r + -
-- .547326,
77 = r + -4 = -1.700434,
Zs = r
11
+ 4r = - 1.205269.
17 !-
.092368 .995734 V
12
The roots of the equation n
2 - Zs u + I =(),
5
or r and r ,
are
- .602635 1
.79801? J ^-
^~ * sm 1i
/
cos +
Tf
being a radius of the circle which is
equal in magnitude to 1,
SOTC^ equation
which pos- Xn 1 =
sess the
same pro- may be similarly solved
by the aid of quadratic equations with
real terms, and the subdivision of the circle into n or 2 m + 1
equal parts may be effected by arithmetical and geometrical
means but a very little consideration will be sufficient to shew
:
tn
that 2 + 1 cannot be a prime number, unless m is of the form
419
binomial equation
xn ~ 1 =0
may be solved, and a circle divided into n equal parts, by pro
cesses which involve the solution of A equations of a dimensions,
of n equations of b dimensions, of equations of c dimensions ,
// . . .
n = 13, we have = 22
and a regular polygon of 13 sides may
n x 3,
be inscribed by the solution of two quadratic and of one cubic
equation. It is this important proposition which forms the
t For 223a + 1 or the 32nd term of the series, has been shewn by Euler
tobe divisible by 641 ;
all the terms in the series which precede it are prime
numbers.
SOURCES OF AMBIGUITY.
2
of 1 +x is 1 + 2x + x , and the only recognized
are^quTllytne only square
2
square root of l + 2x + x is l+x:
unique. the only square of (l-#) is
2
l2x + x , and the only recognized square root of 1 %x + x
z
are derived in such a manner that one of them may replace the
other indifferently, which can only take place when the relations
of value of 1 and x are unknown, that the square roots of either
one of the two, or of both, will equally admit of being considered
as real and arithmetical results.
1040. By selling sugar at $d. per Ib. I lose as much per An ambi-
cent, as the sugar cost me. What was the prime cost of the guo u s
, ,
*
*. ..
problem,
sugar ?
x-9 __ _*_
x ~100
or
100,r-900 =
2
.r ,
or
100,r-,r 2 =900.
2
Subtracting both sides from 50 or 2500, we get
2500 - 100 x + x*= 1600,
or
50 - x = 40,
which gives
x= 10,
It
appears, therefore, that the problem proposed is really am
biguous, and that the processes of Arithmetical Algebra equally
adapt themselves to the determination of the two values of the
unknown symbol which is required to be found. Such an
ambiguity in this and in similar cases could only be made to
disappear, by supposing the range of its values to be limited by
*
See a similar problem, No. 7, Art. 409.
422
it was
A problem 1041. required to determine a right-
Thus, suppose
theHmita
an ^ whose hypothenuse was given, as well as the
ec^ triangle,
tions of sum of the other two sides and of the perpendicular let fall from
value of the , . , , -,
-,
,
(1).
H: z = b, ( 2 ).
Thus
in the solution of the following problem :
" A bought Problem.
pears for l%d., and found that if he had bought six more for the
same money, he would have paid id. less for each. How many
"
did he buy ? we readily find, by Symbolical Algebra, the two
roots of the resulting equation, which are 18 and -24: and
inasmuch as no meaning relative to the problem, in the form
in which it was proposed, can be given to the
negative root,
we derive no information, from the application of the principles
of Symbolical Algebra, which those of Arithmetical Algebra,
in such a case, would not equally supply*.
The inter 1043. Again, though the multiple values which Symbolical
pretations
of symboli Algebra assigns for the solution of a problem, whether they be
cal results
must be re
may admit of interpretation, yet
positive, negative, or imaginary,
lative to the they are not applicable unless they likewise satisfy the geo
problems metrical or other conditions which the problem imposes upon
in whose
solutions them thus symbols with different signs, if applied to lines,
:
they ap
pear.
would denote, generally, lines in opposite directions but such :
negative root of one solution would become the positive root of the other :
for this purpose, in the enunciation of the problem, we change buying into
selling, more into less, paying into receiving; securing in every instance such
a change of the relations of the magnitudes involved as may be correctly
symbolized by a change of their sign, from + to ,
and conversely the ;
EF = AE*EB = AC - CE 2 2
,
= AC 2 - (DE - 2
CD 2
= AC 2
+ CD 2 - DE 2
) ,
we get
ex = a~ + b
2
x 2,
and therefore
If DE
be expressed by the positive value of a-, the negative
root will not, in this case, be
expressed by a line drawn in the
direction of ED
produced for the direction of the line : is DE
estimated with respect to that of the given line EF, which is
is that in which DE = - | + /( j+
2
+ #*
J,
where EF is within
VOL. II. 3H
426
which are
2
i^g&c + c
0, ).
than Cd, then there are no points E and e within the circle
which fulfil the conditions of the problem, and two of values
of x, expressed by
If the as- 1045. If the assumptions which are made with respect to
sumptions fa e representation of lines or other magnitudes be arithmetical,
one part of as opposed to symbolical, in one part of the solution of a problem,
100
tne results which are obtained may be either erroneous or in-?
nfYpro-
blem be same principles of representation be not applied
consistent, if the r
anthmeti-
cal, they throughout. Ihus, supposing a given chord r
tTnueST
DE P asses through a given point C of the
throughout, diameter AB of a circle, and it is required
to determine the segments CD and CE into
which it is divided in C ; then if we should
denote AC by a and CB by b } representing
lines drawn from C by
in opposite directions
symbols with the same signs, then the same principle of repre
sentation and interpretation would be found, in the result, to
extend to the lines CD and CE } which are also drawn in op
posite directions from C: for if we denote CD by x, and DE
by c, we should find
x (c .r)
= ab }
or
where the values of x are both of them positive and express the
segments CD and CE of the chord and as the assumptions made :
mon with D
Alembert, and others who preceded him, that such interpretations
of signs were absolute and not relative, and that the opposition of direction
which they symbolized, was to be sought for, not in a different but in the
same straight line thus CE and Ce, in the figure in the text, would be legi
;
which is likewise
positive and arithmetical: the lines AD and
AE, as well as the lines AC and AB are represented in this
solution by positive symbols.
430
two sides is 35, and of which the perpendicular let fall from
the right angle upon the hypothenuse is
equal to 12."
If we denote one of the two sides by x, the conditions of the
The roots of the first equation are 15 and 20, which satisfy
the primitive equation of condition, and will be found to solve
the problem: they express the two sides of the right-angled
triangle.
The roots of the second equation, which are 47-4 and 12.4
only, and not the primitive equation, and are therefore roots of
solution only, and are not relative to the problem proposed.
a -x + c + x~ = 0,
x? - 7x + 2 -
lx + - 24 =
J(x 18)
(x
2 - 7x - 24) 2 - (x
2 - 7x +
18)
= 0,
x a - lx 4-
(x*
- 7* +
18)
- 24 = :
proposed in Soah?
the last Article, no ambiguity would occur in the equation different
terms of the
8 same equa-
-
, tionor
or in the series
Jx + 3x 5xjx f
35 x*
"
,Jx)~ 2 8 16 128
where the specific radical Jx would retain throughout its
single
arithmetical meaning : but if the equation was
+?-;?-
X X
where x* admits of double and x^ of triple values, and where
it is not assumed that the values either of ar or of or are
simultaneous on both sides of the equations, it would be equally
capable of 36 different forms, and if rationalized would become
an equation of 6 dimensions, all whose roots would be equally
admissible as proper roots of the equation.*
in which the specific values of the radical terms involved in the JjJlJJS m
*
expressions
generating expressions are, in all cases, assumed to be transmitted derived
to the results to which they lead : thus if it was required to
byTnown
operations.
*
Appendix.
VOL. II. 3 I
434
multiply 1+^
into 1 + o?*, the specific values of x* and or*
*
and in the developement of a series such as (1 -# 2 ) which gives
1*2 14 a 35
an expres- the
symbols involved in it which are general in form but specific
sion and -1-1.1 -I.T "
the ordi- m
value, and which are subsequently generalized by the prm-
of its
c
deve- ^
e ^ tne permanence of equivalent forms," or by means of
lopement some proposition to which the application of that principle leads,
have not .
we sometimes adopt a form
, V /. i t
or the series whose value is unique
the same
number of
wnen that of the expression in which it originates and to which
it is assumed to be equivalent, is multiple : thus the value of
the series
1 + nx + * (n
- 1)
~ +n (n
- 1) (n
-
2)
^^ g
+ . . .
...
l g g
we should find
(1 +*) = !(! + *),
and
n
(!+#)"= r(l+jr) =l".*,
where 1" is considered as (Art. 724) the recipient of the multiple
values of (1 -f
#)": and inasmuch as it
appears (Art. 810) that
435
We thus get
or
m in
1055. Let it be required to find a series for (2 cos 6) Series for
cosines of
therefore multiples
m offl.
(2cos0) =:r.p,
if 2 cos 6 be positive, and
(2 cos 0)"
= (-!)>,
if 2 cos 6 be negative.
(2 cos 0)"
= l
m
(e
e
^~+ e~ ^r> e
J^l sinm0,
= cos _ + -
(
OT
4) a yri s in (m 4)
436
Therefore
cos (TW _ 4) (2
+ J- 1 {sin 77Z
(2 r TT + 0) + m sin (7/2
- 2) (2 r ?r + 6)
?W v
(Vj 1 yJ y
sm (?w
- 4) x ..
(2r-7r + ^) + . .
.}
= Cr + 7^ *r,
m =(
If we take the second case, in which (2 cos 0) 1
)"*/, we
shall find, in a similar manner
The + term of c = _
(p !)
1L<2L^2r!) cos (m 2p} 9
=t cos (m
- 2p) 6.
The (m p)
ih
term of c = t cos (m 2m + 2p)
= t cos (m - 2p) 6.
The terms of the series for c being, therefore, the same from
the beginning to the end, it follows that
*
CQ = 2 {cos mQ + m cos (m 2) 6 + - cos
(m 4) + ...}
1x2x3...
|
which is a modified form of the half of the middle term of the
m
binomial (1 + l) (Art. 489): and in the second case being
1 x 3 x 5 x ... m 2
x 2 cos 6.
1 x 2 x 3 x ...
m+
-
I
p- cosZmrir smZmr-
438
"
coswz (2r + 1) TT sinTH (2r + 1) TT
tion in its lowest terms of the form - , and also that r does
n
not exceed n 1.
and r
2r
44 n
or
3n
, we get
-
n
pn or Spir
2prir =-
-
2 2
, and therefore
cos
sm -*
n
-
2or7r
= 0.
In these cases, p is
expressible by a series of cosines only,
and p = c or cw .
pariter par.
Thirdly, if p be odd, w even and of the form 4z
and r =
n-2-44 or
3n-2
- .
A
we nnd
,
cos ~ +-
+ 2 or
p(2r l)7r
n
z*m-
= 0.
Since 2 N/ r l smd = ee ^~ - e - e ^~ l l
(Art. 926), we get
(2 V^Tsinar=r {e
me ^~ l - me< m - 2 ^~ + l
= cosm(2r-7r + 0) ?H
cos(m 2)
-m sin(7w-2) (2
....-
the sign + or being used according as m is of the form 4t
or 4z + 2.
2"
1 "1
(sin 0)
m = {sin w - m sin (w - 2) + . . . to
^ terms},
=
^ 4
or -c-
4
=
cos m (2 r + --) TT sin m (2 r + -f-)
TT
If m= -
(in its lowest terms), p and n being odd numbers,
then
f^^n+i or =fcs 3n+1 ,
4 4
very general terms : for it will very generally be found, that the
more comprehensive is the form in which a problem is stated
441
1061. The same principles will likewise find their appli- Theory of
cation in the general theory of symbolical as distinguished from ^S^ 1
1062. If m
be a fraction in its lowest terms, with an even
n p -i T-- L. value of the
denominator of the form -, there is one value of 1 which is logarithm
2w " con-
O fl-
o
to r = n: in this case we find
equal to -1, corresponding
of a loga-
m p = I
i
VOL. II. 3K
442
we find
log a
m=
log 1
m
+ log a m,
m
where p isthe arithmetical logarithm of a .
m
When we seek for the general logarithms of a , we presume
that a is viewed as a symbolical and not as an arithmetical quan
tity, and we replace it by 1 x a, where 1 is made the recipient of
the affections of a, whilst the other factor a, is considered as an
arithmetical magnitude merely : this is a distinction of funda
mental importance, both in this and other theories, and which
our ordinary notation is not competent to express, a defect which
is a constant and
very embarrassing source of ambiguity*.
= TH
(2 r + 1) TT
J^l + p.
Mr D. F. Gregory, in a very able memoir " On the Impossible Logarithms
*
of Quantities," which is given in the first volume of the Cambridge Mathe
matical Journal, has proposed to distinguish these arithmetical and symbolical
values, by a and by a, where the sign + takes the place of 1, in the
-I-
notation adopted in the text, as the recipient of the affections of a, and where
+m is considered as equivalent to 1 such a use of the sign + , as the :
views which they take of the principles of mathematical reasoning, and gave
ample promise of the valuable results which could hardly have failed to have fol
lowed from the full developement of his powers.
443
If m = 2, we find The
2
= (4r + 2) TT
_
J- + 1 2 log a.
loga-
rithms of
2
(- a) and
log (- a)
a 2 , are not
It appears, from the last Article, that identical.
log a
2
= 2 r-K J^\ + 2 log a.
2
and though the logarithms of ( a) are always included amongst
2
those of a the converse proposition is not true it consequently
, :
log
- am = log (- + log l m + log am
1)
= (2r + 2?wr + 1) TT
J^i + p.
This is the most general form, in which the logarithm of The loga-
a m can be exhibited, and it remains to consider, whether it nega ti ve
admits, in any case, of an arithmetical value. quantity
* Mr
Gregory, in the memoir to which I have before referred, has con
cluded from the equation
that - = + +,
from 3
and he infers thence that + is the general representation of the
sign : but it might, with equal propriety have been taken, as the general
representation of the sign + ;
for the equation
+2 = +
is equally true with the equation
444
is not admissible*.
2""
For one of the values of I is, as we have already shewn,
-(-1) (-!)- %
a form is necessarily excluded, inasmuch as it is assumed
which
m
that x a) m is equal to
(1) (1 a m and not a : we may con
clude, therefore, generally that there is no possible logarithm of
a negative quantity.
y= <?>
x x
they replace y by 1 x y, and a* by 1* x ax, assuming ax in l x a
to be arithmetical: we thus find
*
The existence of a possible logarithm of a negative number was asserted
in my first work on Algebra, page 569, and the mistake was pointed out by
Mr Gregory, in the memoir to which I have before referred.
445
log 1 + log y
t
= log 1"" + log a*,
and therefore
_\ogy
On the
Pa e 12 > Art 558
- - IN Arithmetical Algebra, zero and infinity
are the extreme limits of the values of the symbols which we
limits of the
the^ymbols employ, though the circumstances of their usage will very gene-
lth ~
ra^ v con fi ne them within a much less extensive range.
mefic
Ariihmeti- Thus, in the expression a -b, a may have every value be-
Symbolical
tween infinity and b, whilst b may have every value between a
Algebra. an(j Q the expression a - b itself may have every value between
:
and infinity.
But in Symbolical Algebra, a, 6, and a-b may severally
have every value between positive and negative infinity, zero
being included in their number and if, as in the case of lines
:
geometrical representation.
It may be further remarked that Arithmetical and Symbolical
the primary units, whatever they may be, that we can approxi
mate, in this science, to the representation of continuous mag
nitude.
tudeTrepre-
^
^ n e contrary, the symbols of Arithmetical or Symbolical
A1 ebra are ^competent to represent, like those of Arithmetic,
bols of discontinuous magnitudes : for whilst in Arithmetic,
magnitudes
11 "
are represented at their limits only, however numerous
caTancf they may
Symbolical be, in the other science,
they are represented between their
Algebra .. .
Page 52, Art. 622. We have assumed the rules for the ad
dition and subtraction of numerical fractions as the basis of the
corresponding rules for those operations in Symbolical Algebra :
P P P
+
rp sp _m + n
P P
it appears, therefore, that when m = rp and n = sp, we get
m n m +n
+
~p p~~p~
and inasmuch as this formula or expression represents the sum
448
m n m n
p p p
as
Considering therefore such fractions, in the first instance,
the of the division of the numerator by the
expressing quotient
denominator, we should proceed to interpret their meaning, as
in the case of the results of all other symbolical operations, with
reference to the symbolical conditions which they are required
to satisfy; we be readily enabled to arrive at the in
shall thus
Symbolical
me diumf symbols which are general in form but specific in
Algebra, value, or by rules which are general in their form, though ap
plied to quantities which are specific in their value, are assumed
to be true likewise when the
symbols which represent such mag
nitudes are equally general in their form and
representation.
In Arithmetical
Algebra it is the definition of the operation
(whether expressed or understood) which determines the result,
and also the rule for
obtaining it in Symbolical : it is
Algebra
the rule which determines the
meaning of the operation, such rule
being determined by the principles of Arithmetical Algebra,
when the symbols, though general in their form, are
yet so
specific in their value, come under the operation of its
as to
definitions: the rules of same in Arithmetical
operation are the
and Symbolical Algebra, and therefore the results are the same
as far as
they proceed in common: but it is at the point of
transition from Arithmetical to
Symbolical Algebra, when the
449
^ ^
The case is
very different, however, with the results of Sym- C0 nven-
bolical, as far as they are not common likewise to Arithmetical, tional -
which
i
is
i-i
universally
extended
to the sign
Algebra, when placed between the primitive expression and the the primi-
^^J
but in Symbolical Algebra, in cases which are not likewise com- from it.
operation (using this term in its largest sense, Art. 632) which
is indicated on the other side of it and not
performed: this
view of its general meaning will include as a consequence, arith-
VOL. II. 3L
450
m n
developement of binomials such as (1 + x) and (l + x) , by the
general theorem considered in Chap, xxi, we find
m = m -
(1 + x) l
{l + mx + m(m- 1)
-
CO,
2
x
(1 + x)
n
= l
n
{1 + nx + n (n
- 1) (2),
and it
appears that for all values of x, m, and n, we have
+ m n
= m+
(1 x) x (1 + x) (1 + ar) ",
and likewise that the product of the series (1) and (2), if they
are multiplied together, term by term, according to the ordi
l
m+n
[I + (m + ?i)x + (m + n) (m + n - 1) + &c. . . .
},
Page 106, Art. 680. The existence and form of the series for
(1 + #)", is proved, when n is a whole number, by considerations
derived from the fundamental operations of Arithmetic and Arith
metical Algebra : and the same series, exhibited under an inter
minable form, is extended to all values of the index, by the
principle of the permanence of equivalent forms.
The primary assumption of indices in Arithmetical Algebra
leads directly to the propositions
words, their product will be that series which arises from putting
m + n in the place of m in one series, or of n in the other and :
in the same
Euler had drawn the same conclusion, nearly m
of the series for (1 + x) *, though
manner, in his celebrated proof
of a prin
he at the same time denied the universal application
ciple equivalent
to that of the permanence of equivalent forms,
which alone could make it valid: he produced, as a striking
truth, the very remarkable series
exception to its
other values.
A little consideration, however, will be sufficient to
shew that
is not ap
the principle of the permanence of equivalent forms
)
|
(l-<)(l-fl"-p(l-a
)
[ ^
is not given, or, in other is no statement or definition
words, there
of the operation, by which we
pass from m, on one side of the
the results, which are general in form, but specific in value, are
derived by processes which are definable and recognized, that
of this prin
they become the proper subjects for the application
ciple.
1111
to be symbolically equivalent to
n n
(+ l) x a or 1" x , and n
( l)
n
i
x o
+
i
n
where ,
1 x
* Mr
Gregory, in an article above referred to, p. 442, distinguished + a
and a from each other, the former as belonging to Symbolical Algebra and
considered as opposed to a, and the other as a simple unaffected symbol in
Arithmetical Algebra: such a distinction, however, would rarely enable us to
remove ambiguities, if it could be preserved, and it is farther opposed to the
very important and fundamental rule, which allows us to suppress the sign +
when it is connected with a symbol which has no other symbol before it: it
is contrary to all just views of the relations of Arithmetical and Symbolical
Algebra, to retain any distinctions between them which the ordinary rules of
operation would lead us to suppress.
Mr Gregory further proposes
)
n
instead
of ( 1)" :
considering the use of 1 and 1, as the recipients of signs of
affection, asobjectionable, in consequence of their tending to recal arithme
tical notions, when the circumstances, in which such signs originate, present
themselves exclusively in Symbolical Algebra, and may, therefore, according to
his views, be altogether independent of arithmetic but there may be a deri
:
vative though not an immediate dependance of one of these sciences upon the
other, and I believe that no views of the nature of Symbolical Algebra can be
correct or philosophical which made the selection of its rules of combination
JWW^t
are not correctly stated in the text, inasmuch as no notice
is
equation
x* -par
5
-
qx
2
-rx-s =
is still more completely developed in Art. 1028.
*
Published by Professor Ilolmboe at Christiania, 1839.
455
Academy.
u 5 3
5su + 5s u 2
a = 0,
x5 + x + e = 0,
THE END.
7 DAY USE
RETURN TO DESK FROM WHICH BORROWED
ASTRONOMY, MATHEM*":CS-
STATIS.TiCS LIBRARY. c
This publication is due on the LAST DATE
stamped below.
..
MAR 1 1965
RB 17-60m-8, 61
(Cl641slO)4188 University of California
Berkeley
4153
P4
v.
HIil:
IS!
iiii
,