This is a reproduction of a library book that was digitized

by Google as part of an ongoing effort to preserve the

information in books and make it universally accessible.

https://books.google.com
UC-NRLF

$B 175 749
 LIBRARY
OF THE

UNIVERSITY OF CALIFORNIA.

Class
258
 Clarendon Press Series

A TEXT BOOK OF ALGEBRA

1

ALDIS
 London
HENRY FROWDE

DOMIMINA
NVS TIO
ILLYMEA

OXFORD UNIVERSITY PRESS WAREHOUSE

AMEN CORNER, E.C.

 Clarendon Press Series

TEXT BOOK OF ALGEBRA

BY

W. STEADMAN ALDIS, M.A.

"

UNIVERSITY COLLEGE, AUCKLAND, NEW ZEALAND

Oxford
AT THE CLARENDON PRESS

M DCCC LXXXVII

[ All rights reserved ]

 ㅏㄷ

General
45423
 PREFACE.

THIS book is the outcome of lectures delivered during

some years to students in the College of Physical Science
at Newcastle-upon-Tyne. The Author has endeavoured to
place the subject on a foundation of strict reasoning, and
acomparatively large amount of space is therefore devoted
to the discussion of first principles. It is hoped that on
' this ground the book may be of service to persons who
have not the opportunity enjoyed by students in the old
Universities of access to large libraries, or intercourse with
other mathematical scholars.
The Author desires to thank his brother, Mr. T. S.
Aldis, M.A. , of Trinity College, Cambridge, for revising
the proof-sheets, a task which the remoteness ofthe Author's
present residence rendered it impossible for himself to
fulfil.

UNIVERSITY COLLEGE,
AUCKLAND, NEW ZEALAND.

45423
 CONTENTS .

The numbers indicate the Articles.

CHAPTER I.
PAGE

1
ARITHMETICAL NOTIONS
Addition; Subtraction; multiplication; commutative, associa-
tive and distributive laws, 21-26. Fractions, 29-40. Division,
41-46. Division offractions,47,48. Zero, 51.

CHAPTER II .

ALGEBRAICAL LAWS . 33

Summary of formulae, 53. Negative expressions, 54, 55. Rule

of signs, 56-58. Powers, indices, index law, fractional and nega-
tive indices, 62-70. Extensions of formulae, 72-80.

CHAPTER III.

ADDITION AND SUBTRACTION 52

Definitions, 81-88. Addition, 89-94. Subtraction, 95-97.

Brackets, 98-101 .

CHAPTER IV.
61
MULTIPLICATION

Single terms, 103. Rule of signs, 104, 105. General rule of

multiplication, 106-110. Special examples, 114-137 .
viii
ム
CONTENTS. スース C
くー
CHAPTER V.
PAGE

DIVISION 81

Preliminary, 138-148. Examples ; definition of quotient and

remainder, 149-158. General theorem for division by a binomial,
163-172.

CHAPTER VI .

HIGHEST COMMON DIVISOR AND LOWEST COMMON MULTIPLE 100

Definition of Highest Common Factor, 175. Monomials, 177.

General process, 180-185. Examples, 186-189. Definition of
Lowest Common Multiple, 191. Monomials, 192, 193. General
method, 195, 196. Examples, 197, 198.

CHAPTER VII .

FRACTIONAL FORMS 116

General definition ; repetition of laws, 199, 200. Reduction of

fractions, 203. Addition and subtraction, 205. Multiplication
and division, 208-210. Compound fractions, 211 .

CHAPTER VIII .

INDICES 128

Four laws, 214. Incommensurable quantities, 215-225. Inter-

pretation of fractional and negative indices and proof of the four
laws with this interpretation, 227-248 .

CHAPTER IX.

SURDS 146

Definition, 249-251 . Simplification of surds, 252-256. Binomial

surd, 257. Rationalising factor, 258. Product ofbinomial surds,
261-263. Square root of binomial surd, 264-266. Impossible
or operational quantities, 270. Interpretation, 272-276. Cube
roots of unity, 277-283.
 CONTENTS . ix

SECTION II. —EQUATIONS .

CHAPTER X. PAGE.

SIMPLE EQUATIONS WITH ONE UNKNOWN 166

General principles, 284-287. Examples, 288-296. General type,

297, 298. Problems, 299-304.

CHAPTER XI .

SIMPLE EQUATIONS WITH MORE THAN ONE UNKNOWN 183

Example, 306, 307. General type of equations with two un-

knowns, 308. Elimination, 309. Three methods, 310-316.
Equations with three unknowns, 321-323. Problems, 324-330.

CHAPTER XII.

DETERMINANTS 202

General formulae for solution of two equations with two unknowns,

334-336. Definition of determinant, constituent, row, column,
337-351 . Evaluation of determinant of third order, 344. Inter-
changes, 345, 346. Minor determinants, 348. Properties of
determinants, 349-354. Solution of equations with three un-
knowns, 355-358 . Inconsistency or independence of three
equations, 361-364.

CHAPTER XIII .

SQUARE ROOот 224

Aninverse process, 366. General method, 367-369. Examples,

370-376. Approximate square root, 377, 378. Square root of a
number, 379-385. Examples 386-388. Approximation by
division, 389-392. Square root of fraction, 394-400.
CHAPTER XIV.

QUADRATIC EQUATIONS 244

Definition, 401. Type form, 402. Preliminary theorem, 403.

General method, 406-409. Discriminant, 410-413. Sum and
product of roots, 414. Resolution of quadratic expression into
factors, 415. Another investigation of method of solution,
416-421. Equations with given roots, 424. Symmetrical
functions of the roots, 426, 427. Equations like quadratics,
428-434- 1
X CONTENTS.

CHAPTER XV.
PAGE

CUBE ROOT AND CUBIC EQUATIONS . 268

General method for cube root, 435-439. Examples, 440-442 .

Cube root of a number, 444-449. Approximate cube roots,
450-451 . Cubic equation, 452-459.

CHAPTER XVI .

SIMULTANEOUS EQUATIONS ABOVE THE FIRST DEGREE . 283

Equations in which one unknown occurs only in the first power,

462. Homogeneous equations, 464. Symmetrical equations,
466-469. Two equations of second degree, 471 , 472. Elimina-
tion, 473. Problems, 474.

SECTION III .-SERIES .

CHAPTER XVII .

PERMUTATIONS AND COMBINATIONS 297

Definition of series, factorial, 475-477 . Permutations, 480-484.

Combinations, 485-498 . Arrangement of n things some of which
are alike, 499.

CHAPTER XVIII .

ARITHMETICAL AND HARMONICAL PROGRESSIONS 316

Definition of Arithmetical Progression, 503. Formulae, 505,

506. Problems, 507-509. Arithmetic means, 510, 511. Reci-
procal, 515. Harmonical Progression, 516. Means, 517, 518.

CHAF

GEOMETRICAL PROGRESSION 327

Definition, common ratio, 5 læ, 521, 522. Means,

523-527 . Sum to infinity, epeating decimals, 532
533. Interest, 534-538. A 541 .
 CONTENTS . xi

CHAPTER XX.
PAGE

BINOMIAL AND POLYNOMIAL THEOREMS 345

Product of unlike binomial factors, 543. Binomial theorem, 544,

545. General term, 547. Examples, 548, 549. Polynomial
theorem, 550-553. Mathematical induction, 554. Another
proofofBinomial Theorem, 555-557. Greatest term, 558.

1
CHAPTER XXI .

GENERALISATION OF THE BINOMIAL EXPANSION 360

Explanation of the equivalence of the power of a binomial and

the infinite series, 560-572. Convergency of series, 573-576.
General term, 577-579. Approximate roots, 581. Extension of
polynomial theorem, 582-584.

CHAPTER XXII.

INDETERMINATE COEFFICIENTS AND RECURRING SERIES 377

General principle of undetermined coefficients, 585-587. Recur-

ring series, scale of relation, 588-590. Resolution into partial
fractions, 591. General term, 593, 594. Sum to n terms and to
infinity, 595. Homogeneous products, 597, 598.

CHAPTER XXIII .

SUMMATION OF SOME SPECIAL SERIES 389

Series inwhich each term is the product of factors in arithmetical

or harmonical progression, 599-604. Sums of powers of natural
numbers, 605-609. General term, any rational integral function
of n, 610. Method of differences, 611 .

CHAPTER XXIV .

LOGARITHMS AND EXPONENTIAL SERIES 404

Limit defined, 612. Examples, 613. Value of €, 614-616. Ex-

ponential theorem, 617, 618. Logarithm defined, 619. Series
for logarithm, 622, 623. Properties of logarithms, 624-626.
Base, characteristic, mantissa,628-632. Different bases,633-635-
Modulus, 636. Other series, 637, 638. Calculations, 639-643 .
xii CONTENTS .

SECTION IV.-ARITHMETICAL APPLICATIONS .

CHAPTER XXV. PAGE

424
RATIO, PROPORTION AND VARIATION
Definition of ratio, 644-646. Proportion, 647. Algebraical ex-
pression of, 648-650. Derived results, 651-655. Rule ofThree,
656. Compound ratio, 657. Duplicate ratio,658. Sum ofante-
cedents has same ratio to sum of consequents, 659. Variation,
definition of, 660-662. Inverse, 663. Joint, 664, 665. Double
rule of three, 666.

CHAPTER XXVI.
CONTINUED FRACTIONS 440

Reduction of fraction to form of continued fraction, 667-669 .

Convergents, 670-677. Quadratic surd, 678-680. Repeating
continued fraction, 681, 682.

CHAPTER XXVII .

INDETERMINATE EQUATIONS 454

Statement of problem, 683-685. Methods of solution, 686-689 .

Use ofcontinued fractions, 690-692. Number of solutions, 693,
694. Three unknowns, 695. Special equation of two dimensions,
697.

CHAPTER XXVIII.

INEQUALITIES 466

Definitions, 698, 699. Arithmetic and geometric mean, 700, 705.

Maxima and minima values, 701-704, 706. Other theorems,
707-709 .

CHAPTER XXIX.

THEORY OF NUMBERS 474

Modulus and residue, 710. Scale, radix, digits, 711-713. Ex-

amples of change of scale, 714-717. Numbers divisible by
certain factors, 718-722. Natural numbers arranged in sets ;
prime numbers, 723-730. Divisors of a given number 731 , 732 .
Numbers less than given number and prime to it, 733-736.
Repeating decimals, 738-740. Extension to any radix, 741-745-
Fermat's theorem, 746, 747- Product of any n consecutive
integers divisible by In, 748. Second proof of Fermat, 749, 750.
 CONTENTS . xiii

CHAPTER XXX.
PAGE

PROBABILITIES 497

Measure of chance 751-756. Compound chances, 757-762.

Inverse chances, 763, 764. Value of expectation, 765, 766. Life
assurance, 767, 768. Present value of assurance, 769, 770. Of
life annuity, 771. Commutation Tables, 772, 773- Annual
premium, 774-

ANSWERS TO EXAMPLES 513

MISCELLANEOUS EXAMPLES 553

ANSWERS TO MISCELLANEOUS EXAMPLES 575

1

1
 ERRATA.

Page 20, line 11 , for multiple read multiplier.

" 28, " 2, for Art. 45 read Art. 44.
" 34
, " 19, for laws read law.
" 45, " 10, omit of.
" 51, " 8, for bread √b-c.
" 52, " 2 from bottom, for the value read these values
" 77, " 9,for y + x read y + z.
"
81, "
9 from bottom, insert the before highest power.
" 91, 4th line from bottom,for 161 read 160.
" 106, line 24,for x³- 3x- 12 read x³- 3x +2 .
 ALGEBRA .

CHAPTER I.

ARITHMETICAL NOTIONS .

1. THE laws which regulate the relations of different

numbers to each other constitute the science of Arithmetic.
The application of these laws to practical purposes con-
stitutes the corresponding art, an art which in too many
cases conceals from its possessors the science which under-
lies it. The laws of Algebra are in most cases identical,
in form at least, with those of Arithmetic, and for this
reason it is desirable to obtain clear ideas of the latter
before proceeding to the former.
2. A single thing of any kind is called a unit or one, and
is denoted by the symbol 1. If another thing of the same
kind be placed with the first, we say that there are two of
them, and denote the word two by the arbitrary sign 2.
To indicate the juxtaposition of two things we use the
sign +, and can thus also denote the two things when
placed together by the combination of symbols 1 + 1 .
The fact that this symbolical representation is equivalent
to the former is expressed by the use of another sign = .
Thus we have what is called an Arithmetical equation,
1 + 1 =2,
the sign = meaning that the numbers represented by the
symbols on either side of it are equal.
B
2 Arithmetical Notions. [3 .
3. If another thing of the same kind be placed with
the former two it is said that there are three of them, and
their number is denoted by the arbitrary symbol 3. It is
equally denoted by either combination of previously known
signs 2 + 1 or 1 + 1 + 1 . We have thus
2+ 1 =3

or 1 + 1 + 1 = 3.

4. By a similar process the ideas represented by the

symbols 4, 5, 6, 7, 8, 9 are obtained, each of these in
succession being defined as equal to its predecessor with
one more put to it. Thus 4 = 3 + 1 , 5 = 4 + 1 , 6 = 5 + 1 ,
and so on. At this point the introduction of new arbitrary
symbols to represent new numbers ceases, and the remaining
numbers are represented by placing the symbols already
named in different positions. The result of adding one to
nine, which may be denoted by 9 + 1, is, it is true, called
by a new and arbitrary name ten, but it is denoted by writing
the sign for one in a different position, so that it shall
indicate not one unit but one ten. To effect this alteration
of position another symbol 0, signifying nought or nothing
and spoken of as nought or zero, is used, and ten is denoted
by the combination of 1 and 0, thus, 10 ; the combination
indicating that the number consists of one ten and no
units.
5. All succeeding numbers are similarly represented by
writing down the number of tens they contain in one
place and the number of units over to the right. Each
number is still defined as being equal to its predecessor with
a unit added to it. When a number which contains ten
times ten is reached, a new name, a hundred, is given to it .
and a new place, the third from the right, is reserved for
it. Ten hundred is called a thousand, and the sign which
represents the number of thousands in a given number is
8.] Arithmetical Notions. 3

put in the fourth place from the right. This process can
be continued indefinitely. Thus in the symbol 2456789
the 9 means nine units, the 8 means eight tens, the 7
means seven hundreds, the 6 six thousands, the 5 five ten-
thousands, the 4 four hundred-thousands and the 2 two
thousand-thousands, to which latter accumulation of
numbers the name of a million is given.
6. The symbol 2456789 thus carries on its face, so to
speak, the manner, or rather a manner, in which the
number which it represents is composed. It does not
however tell all the different ways in which the number
might be made up.
7. If three things of a certain kind be lying on a table
and four more of the same kind be placed with them, the
result is denoted by the combination of signs 3 + 4, the
sign + merely signifying that the number represented by
the symbol which follows it is to be put to, or added to,
that represented by the symbol which precedes it. If the
original three things be removed, there will evidently
remain four, and the result of replacing the three will be
properly represented by 4 + 3. Hence
3 + 44 + 3 .

This proof does not in any way depend on the particular

numbers 4 and 3, but it will similarly follow that if a and b
be taken as symbols to represent any two integers whatever,
a + b = b + a.

8. It is also tolerably evident that if there be three

groups of things, one containing three, the second four,
and the third five, and the three groups be all collected
into one, the total number of articles in the combined
group will be the same whichever of the three groups we
bring to the other and in whatever order we move the two
that are moved.
B2
4 Arithmetical Notions. [9.

The number of things in the final group will be

represented by one or other of the sets of symbols,
3 + 4 + 5, 3 + 5 + 4,
4+ 3 + 5, 4 + 5 + 3,
5 + 3 + 4, 5 + 4 + 3,
and we see that any one of these sets is equal to any
other.
The same thing would be true if any number of different
numbers be combined by addition ; the order in which the
additions are effected is a matter of perfect indifference.
The result is known as the Commutative law in addition.
The student can easily convince himself that the proof of
the truth of the principle does not in any way depend on
the particular numbers 3, 4, 5, that have been employed,
but that it is equally true for all numbers.
Thus if a, b, c represent any three integers, the value of
a + b + c is the same as that of either b + c + a, a + c + b,
b + a + c, c + a + b or c + b + a.
9. When two groups of similar articles, the number in
each group being known, are placed together, the process
of discovering the number of articles in the combined group
is called addition, and is denoted by writing the symbol +
between the numbers which represent the groups. The
resulting number is called the sum of the two original
numbers, and can always be discovered by a process of
reasoning depending on the definitions of the original
symbols ; thus if the number in one group be 3 and the
number in the other 4, the number in the united group is
3 + 4, or 4 + 3.
But 3 = 2 + 1 = 1 + 1 + 1
Therefore 4+ 34 + 1 + 1 + 1
= 5+ 1 + 1
= 6 + 1
= 7
11.] Arithmetical Notions. 5

The collection of symbols 4 + 3 can thus be replaced by the

single symbol 7 .
10. If from a group of articles containing a given
number we take a part away, the number of the latter
being also known, the process of ascertaining the number
left is called subtraction. The symbol for the operation of
taking away one number from a greater is 一.
Thus if
there be originally 5 articles, and 2 be taken away, the
operation is denoted by the combination of symbols 5 -2 .
The process of subtraction is the reverse of that of addition :
the result of a subtraction has to be obtained by remember-
ing known results of addition. Thus when we have
ascertained from addition that 3 + 2 = 5, it follows that 3
is the number left when we take 2 from 5. Hence we
may write 5-23 .

The equation 5 = 3+ 2 shows also that if 3 be taken from

5, 2 will be left. Thus also
5-3 2.

If the results of addition of all possible pairs of numbers

were tabulated, the table would also give the remainder
after subtracting any number from any larger number.
By the help of memory and certain rules depending on the
ordinary system of decimal notation the necessity for the
actual formation of such a table is obviated.
11. A collection of symbols such as 4 + 3 or 5-2 will in
future be called an expression. The expression 4+ 3 may
be always replaced by the number 7, and the expression
5-2 by the number 3 ; but it is sometimes desirable to
retain the expression as a record of the manner in which
the number was obtained. When it is required to perform
any operation on the number which is represented by the
expression, the latter quantity is frequently enclosed in a
6 Arithmetical Notions. [12.

bracket thus (4+ 3) or { 4 + 3 } or [4 + 3]. Thus the

compound expression 6+ (4 + 3)
implies that the number represented by 4 + 3 is to be
added to 6 ; the compound expression
8-(5-2)
means that the number represented by (5-2) is to be
subtracted from 8.
The parts of an expression which are connected by the
sign + or are called the terms of the expression.
12. The expression (4 + 3) denotes a group formed by
putting three things along with four things of the same
kind. The expression 6+ (4 + 3) denotes a group formed
by putting the group (4 + 3) along with a group of 6
things. It is clear that the final result will be the same if
the four things be first carried to the six and the three
remaining things fetched afterwards. The result of this
process would be denoted by the expression 6 + 4 + 3.
Hence
6+ (4 + 3) = 6 + 4 + 3
Similarly it follows that
6+ (4 + 3 + 5) = 6 + 4 + 3 + 5,
for the one expression denotes the result of taking three
groups of 4, 3 and 5 things respectively and putting them
to a group of six all at once, while the other represents the
result of taking them in succession ; which operations
must lead to the same final result.
Hence if an expression all whose terms are preceded by
a + sign be added to a number, the bracket may be
removed and the terms of the expression added to the
number in succession.
13. The expression
9- (4 + 3)
means that from a heap of nine things the number of
things denoted by the expression (4 + 3) is to be taken.
15.
] Arithmetical Notions. 7

Clearly the result will be obtained equally well by first

taking four of the things, and then going back and fetching
three more. The result of taking the four things away is
denoted by the expression 9-43; the result of taking the
three other things by 9-4-3 .
Hence 9- (4 + 3) = 9-4-3 .
Similarly 9-(4 + 3 + 2) = 9-4-3-2,
the former denoting that (4 + 3 + 2) things are taken away
at once, and the latter that 4 are first removed, then 3, and
then 2.
Hence if an expression all whose terms are connected by
the sign + be subtracted from a number, the bracket may
be removed provided the sign - be written before each
term.

14. Again, consider the expression

8+ (5-2),
which means that (5-2), that is, the remainder after taking
2 from 5, is to be added to 8. Let the numbers represent
pounds sterling : the process required to be effected is to
pay £(5-2) to a person who already has £8. If a
payment of £5 be made, the payee has evidently received
too much by £2, which he must give as change to the
payer. The net result is that the number of pounds the
payee has is 8 + 5-2. Hence
8+ (5-2) = 8 + 5-2.
15. Suppose that a person A employs a messenger B to
collect debts of £3 and £4 due to A and on the same
errand to pay a bill of £5 due by A. B receives (3 + 4)
pounds and pays 5 pounds. He thus brings back to A a
number of pounds denoted by (3 + 4-5) and if A originally
had £8, he now has 8+ ( 3 + 4-5) pounds.
IfA had gone on the errands himself, after receiving
the first payment he would have 8 + 3 pounds ; after
8 Arithmetical Notions. [16.
receiving the second, he would have 8 + 3 + 4 pounds.
He then pays 5 pounds, and consequently has 8 + 3 + 4-5
pounds left. Clearly however his final financial state is
the same as when he employed the messenger. Hence
8+ (3 + 4-5) = 8 + 3 + 4-5 .
By this and similar instances it can be seen that when an
expression is to be added to a number the bracket may be
removed and all the terms written down with their
original signs.
16. It may be noticed here that when A went to collect
his debts himself it would make no difference if he paid the
sum of £5 before he received either or both of the sums of
money due to him. Thus all the expressions
8 + 3 + 4-5 , 8-5 + 4 + 3
8 + 3-5 + 4,
8-5 + 3 + 4,
8 + 4-5 + 3
8 + 4 + 3-5 } (1)

since they merely denote the result of the same operations

conducted in different orders, are identical.
17. In the expressions in (1) we see that a term which
has a + sign before it indicates a payment made to A,
while a term with a - sign before it indicates a payment
by A to another person. Here we have a first hint of a
wider interpretation than that with which we started of
the signs + and -, a suggestion that besides indicating
the two opposite processes of addition and subtraction,
they may indicate two opposite qualities in the quantities
before which they stand.
18. Suppose now that A keeps his capital in a bank,
and that he pays £5 by a cheque on his banker, while
the debtors pay £3 and £4 respectively into the bank to
A's credit. Provided that A's original balance exceed £5,
the order in which the payments are made to and from
the banker will make no difference to the final state of
20.] Arithmetical Notions. 9

affairs. If then we regard the sign - to denote that the

number before which it stands is to be subtracted from
some number which need not be stated, while + denotes
a similar process of addition, the operations denoted by
+ 3 + 4-5, 4-5 + 3 , -5 + 3 + 4
+ 4 + 3-5, 3-5 + 4 , -5 + 4 + 3
all give the same result, namely in this case an increase
of 2.

19. With this understanding as to the meaning of the

signs + and -

when they stand without any preceding

terms, the theorem of Articles 7 and 8 may be extended,
and it may be asserted that the value of any expression
such as those which we have been considering is inde-
pendent of the order of the terms. In writing the ex-
pression the sign + is usually omitted if it occur before
the first term ; the sign -, on the other hand, must be
carefully inserted. A term with a sign + before it is
called a positive term, one with the sign - preceding is
called a negative term.
20. The expression 8- (5-2) denotes that the expression
(5-2) is to be subtracted from 8. If 5 be subtracted from
8, the result is denoted by 8-5. Evidently however this
result is too small, because more has been taken from 8
than ought to have been ; to obtain the correct result we
must add that which has been subtracted in excess, that
is, 2. Hence the final result is 8-5 +2, or
8- ( 5-2) = 8-5 + 2 .
Suppose again that a person A employs a second person
B to pay his debts and to receive the money due to him.
A owes £3 to C, £4 to D, while E owes £5 to A. In order
that B may be able to discharge the debts it will be
necessary for A to give him the excess of what he owes
to Cand Dover what E owes to A ; that is, A must give
10 Arithmetical Notions. [21 .

B (3 + 4-5) pounds. EvidentlyB in this case can go first

to E and receive £5 and will then have enough to pay C
and D. If A have £11 to begin with, he will have
{ 11- (3 + 4-5)} pounds when he has paid this single sum
to B. Suppose however that A goes to pay the money
himself. When he has paid C, A has (11-3) pounds left.
After paying Dhe has ( 11-3-4) pounds. He then goes
to E and receives £5, after which he has (11-3-4 + 5)
pounds. Hence
11-(3 + 4-5) = 11-3-4 + 5.
Comparing this result with that of Article 13 we derive
a general rule, that if any expression is subtracted from
a number the bracket may be removed from the expression,
provided the sign + before any term is changed into -,
and the sign – is changed into + .
21. If a number be added to an equal number the re-
sulting number is called twice the first. If a third number
equal to either of the two former be added to this sum,
the result is called three times the first, and so on. Thus
a group which is formed of seven equal groups contains
seven times as many articles as either of the groups singly.
The resulting number can be obtained by a series of ad-
ditions. It has been found useful however to construct
a table, known as the multiplication table, which gives
the numerical values of such sums up to twelve times
twelve things, or in some cases up to twenty times twenty.
Familiarity with the construction and use of this table
is strongly to be recommended to all would-be mathe-
maticians. The process of adding together several equal
numbers is called multiplication, and the process of taking
a number, as six, any number of times, as seven, is de-
scribed as multiplying 6 by 7. The result of the process
we shall denote as 6 x 7, the sign x signifying ' multi
23.] Arithmetical Notions. 11

plied by.' The number 6 is called the multiplicand, while

7 is called the multiplier.
22. The process may be graphically represented by
drawing six stars or dots, as in the top line of the ac-
companying diagram, and repeating the process seven
times.
*

* *
* * *

* *
* * *

* * *
* * *

*
* * * *

*
*

We thus obtain seven horizontal rows, each containing six

stars. The total number of stars is therefore 6 x 7.
This number may also be counted by taking together
those in one vertical line. In each of these there are
seven stars, and, as there are six vertical lines, the total
number is obtained by taking seven stars six times over,
and is therefore represented by 7 x 6. Hence
6x7 = 7x 6.

The result obtained by multiplying one number by

another is called the product of the two numbers ; the two
numbers are called the factors of the number which is the
product. The product is often called a multiple of either
of its factors.
23. The result of the last Article can also be obtained
in a slightly different manner.
Let a rectangle ABCD be constructed of which one side
AB contains six units of length and the adjacent one AD
contains seven. Through each point of division of AB
and AD let lines be drawn parallel to the sides AD and
AB respectively. The rectangle ABCD will thus be divided
into a number of squares, each having the unit of length
12 Arithmetical Notions. [24.

for its side. The number of these in any horizontal row

is evidently the same as the number of units of length in
AB, in the supposed case, six. The
A B
number of horizontal rows is the
same as the number of units of
length in AD, in this case seven.
The total number is obtained by
taking six seven times, or is 6 x 7.
By counting the squares in a
vertical column, in each of which
in the figure there are seven, and
then counting the number of ver-
D C
tical columns, in this case six, the
total number of squares is seen to
be seven taken six times, or 7 x 6 .
Hence, as before, 6x7 = 7x 6 .

The student will see that nothing in either proof depends

on the particular numbers 6 and 7, but that they would
each be equally valid if any other numbers had been em-
ployed. It may be laid down then as a general principle
that the value of the product of two numbers is unaltered
by a change in the order of the factors. Thus if a and b
represent any two integers, a × b = bxa.
This law is known as the Commutative law in multiplica-
tion.
24. The law can be easily extended to the product of
three numbers .
Let a right solid be formed whose three edges AB, AC,
AD contains respectively 3, 4 and 5 units of length.
Through each point of division of AD draw planes
parallel to that which contains AC and AB. These planes
will divide the right solid into as many slices, all equal,
as there are units of length in AD. Again let planes be
drawn through each point of division of AB parallel to
24.] Arithmetical Notions. 13

CAD, and other planes through each point of division of

AC parallel to BAD. The right solid will be thus divided
into a number of cubes, each
D
having a unit of length for its
edge. The number of these cubes
in the lowest slice is evidently,
as in Article 22, the product of
the numbers of units of length
in AB and AC, and may be re-
presented in this figure by either
3 x 4 or 4 x 3. There will be the
A
same number of cubes in each B

horizontal slice. The total num-

ber is therefore obtained by mul- C
tiplying this product by the
number of slices, in this case 5, and may be denoted by
(3x4) x 5 or (4 x 3) x 5.
By counting the cubes in a different order, as for instance,
counting those first which lie in a slice between any two
planes parallel to CAD, and multiplying by the number
of such slices, the total number of cubes is easily seen to
be represented by
(4x5)x 3 or (5 x 4) x 3.
Similarly it may be seen to be equal to
(3x 5) x 4 or (5 X
x 3) x 4.
Hence, whichever pair of the numbers we multiply first,
we get the same result when we multiply that product by
the third. Also, since by Article 22,
(5 x 3) x 4 = 4 x (5 X
x 3),
we may, if we please, multiply either pair together and,
using that as a multiplier to operate on the third number,
still obtain the same result.
14 Arithmetical Notions. [25.
As a particular and important case it follows from the
equality of these different forms that
(3 x 4) x 5 = 3 x (4 X
x 5)
where, as previously, the enclosure of two figures and a
connecting sign in a bracket implies that the whole set
of symbols within the bracket is to be treated as one
number.
This last result is sometimes known as the Associative
law in multiplication. The student will easily see in this
case too that no point in the proof depends on the par-
ticular numbers employed, but that it holds equally what-
ever numbers may be used. Thus if a, b, c represent any
three integers, it follows that
(axb) x c = a x (bxc) = (a x c) X
x b.
25. The expression (3 + 4) × 5 means that the sum of
the two numbers 3 and 4 is to be multiplied by 5. The
process can be graphically represented. Place three dots
or stars in a row. A little to the right of them in the
same straight line place four more stars,
* *
* * * * *

* * * *
* * *

* * * *
* * *

* *
* * * * *

* * *
* * *

There will then be 3 + 4 stars in that line. Underneath

place an exactly similar row and repeat the process until
there are five of them. The total number of stars is five
times (3 + 4) or is represented by ( 3 + 4) x 5.
By counting the stars in the two separate groups it will
be seen that in the left-hand group there are three stars
taken five times, the number of which is ( 3 X
x 5), while in
the right-hand group there are four stars taken five times.
26.] Arithmetical Notions. 15

the number of which is (4 x 5). The sum of these two

numbers must equal the total number, or
(3 + 4) x 5 = (3 x 5)+ (4x 5). (1)
It follows by the commutative law that
5x (3 + 4) = (3 + 4) x 5.
Therefore 5x (3 + 4) = (3 x 5) + (4x5) by (1)
= (5 x 3) + (5 x 4) ; (2)
since by the commutative law
3x5 = 5x3 and 4x5 = 5 x 4.
The two results (1) and (2) are examples of what is called
the Distributive law.
By exactly similar reasoning it can be seen that a simi-
lar result holds for the multiplication of the sum of any
number of numbers by any number; that, for instance,
(3 + 4 + 5 + 6) × 7 = 3x7 + 4x7 + 5 × 7 + 6x7,
and consequently by the commutative law that
7x (3 + 4 + 5 + 6) = 7x3 + 7x4 + 7 × 5 + 7x6.
26. The expression (7-3) x 5 means that three is to be
subtracted from 7, and the remainder to be multiplied
by 5.
This process can be graphically illustrated. Let seven
stars be written in a row, and let a line be drawn cutting
off three of them to the right.
*
* * * * * *
* * * * *
* *

* * *
* * * *
* * * *
* * *
* * *
* * * *

The number of those on the left of this line, will therefore

be (7-3). Let an exactly equal row of stars be placed
under the first, and the process repeated until there are
five rows altogether. The number of stars to the left of
16 Arithmetical Notions. [27-
the vertical line is evidently (7-3) x 5. The whole num-
ber of stars is equally evidently (7 x 5), and the number of
stars to the right of the vertical line is (3 X x 5). If the
latter stars be removed from the diagram the number left
will therefore be represented by (7 X
x 5) - (3 X
x 5). But there
are now left only the stars to the left of the line. Hence
(7-3) x 5 = 7 × 5-3 × 5.
By applying the commutative law it will follow that
5x (7-3) = 5 × 7-5 × 3 .
Thus the distributive law is seen to hold when the mul-
tiplier or the multiplicand contains a negative term as well
as when all the terms are positive.
Similarly it can be shown that
5x (6 + 7-3) = 5x6 + 5 × 7-5 x 3,
and the general result may be stated that when an expres-
sion consisting of several terms is to be multiplied by a
single number the result may be obtained by multiplying
each of the terms by that number and writing down the
successive products with the same signs as those of the
original terms. This is called the Distributive law in
multiplication.
27. It has been seen that the process of taking six
seven times is described as the process of multiplying six
by seven. It will be sometimes convenient to speak of the
same process as that of multiplying seven into six. In
most parts of higher mathematics it is convenient and
desirable to mention the operator before the subject
operated on. Thus (7) is the operator which multiplies,
(6) the subject which is multiplied, and thus the new
phraseology will be more in accordance with future
generalisations than the old.
28. When we have ascertained the value of the product
of two given numbers, as for instance that of 6 and 7,
29.] Arithmetical Notions. 17

namely 42, it follows that 42 contains 6 seven times, and

also contains 7 six times. We express these two results
otherwise thus, 42 divided by 6 gives 7, and 42 divided
by 7 gives 6. The number of times which one number
contains another is called the quotient of the first divided
by the second. The symbol is used to express the
words divided by. Thus the equations
4267 ,
427 = 6 ,
are equivalent to the verbal statements previously made.
The number which is divided is called the dividend, the
number by which it is divided is called the divisor.
The process of division is an inverse one, and the
answer to the question, how many times does one number
go into another, depends on the results, previously calcu-
lated, ofmultiplication.
For instance, if it be required to divide 45 by 6-that is,
to find how many times 6 is contained in 45-we must
refer to the table of successive multiples of 6. In this
table it is found that 6x7 = 42, 6x8 = 48. Hence 45,
which lies between 42 and 48, contains 6 more than seven
times and not quite eight times.
The quotient is in all cases a number which, when
multiplied by or into the divisor, gives a result equal to the
dividend. If no number can be found such that this shall
be the case, the divisor is not contained an exact number of
times in the dividend, but, as in the examplejust stated, we
can find two consecutive numbers such that one of them
multiplied into the divisor gives a less, and the other
multiplied into it a greater product than the dividend.
29. Hitherto the numbers considered have been formed
by placing together one after another things of the
same kind ; they have expressed, that is to say, a number
of whole things. They are on this account often called
C
18 Arithmetical Notions. [30.
integers, integral numbers, or whole numbers, the adjective
being transferred from the things represented to the
numbers which represent them. We have now to con-
sider the method of numerically representing parts of
things. In this case too the adjective is transferred from
the thing itself to the sign or number which represents it ;
and as the thing itself has to be broken or divided into
parts, the sign which represents any of these parts is called
a broken or fractional number, or more shortly a fraction.
30. To represent the parts of a given whole, two signs
are required, one to represent the kind of parts, which is
therefore called the ' namer ' or ' denominator,' and the
other to signify how many of them are taken ; this latter
is therefore called the ' counter ' or ' numerator.' The
parts considered are always supposed to be equal parts.
And the denominator tells into how many equal parts the
whole is divided, and therefore tells also what kind of a
part, or how big relatively to the whole, each of them is.
The two numbers are written one above the other with
a line between, the numerator being uppermost as thus
5/7, or more usually and conveniently . Here the unit,
whatever it may be, is divided into seven equal parts, each
of which is called a seventh of the whole, and five of these
parts are supposed to be taken.
31. The most important property of fractions is that
the value of a fraction is unaltered if its numerator and
denominator be multiplied by any the same number.
Let, for instance, the whole be the straight line AB, and
let it be divided into seven equal parts marked as in the

A C B

figure. If AC contain five of these, AC is of AB. Let

each of the seventh parts be divided into any number of
32.] Arithmetical Notions. 19

equal parts, as 4, the total number of these parts in AB is

(4 x 7) while the number of them contained in AC is 4 x 5 .
Hence AC is also obtained if AB be divided into 4x7
equal parts, and 4 × 5 of them be taken; that is, AC is also
4x5
represented by the fraction of AB.
4x7

54x5
Hence
7 4x7

The student can easily convince himself that the proof

does not in any way depend on the particular numbers
employed. It of course follows that the value of a fraction
is not altered by dividing numerator and denominator by
any common factor.
32. In the use either of integers or fractions we tacitly
understand that they always refer to a whole number or a
fraction of some thing which we may call the unit. Thus
the numbers 4, 5, 7, mean four things, five things,
seven things of some particular kind, and three quarters of
some thing respectively. It has been seen that all the
properties of addition, multiplication and so on which
these numbers possess are quite independent of the par-
ticular thing to which they refer. We can in fact abstract
the properties of the numbers from those of the things
which they represent ; and when this is done they are called
abstract numbers. Still for certain purposes it is impor-
tant to remember that the symbol 2 does represent that
there are two of some particular kind of thing. If the
kind of thing be represented by a general symbol, as a, the
number two will be represented reasonably by the symbol
2. a, where the dot, which we shall find later on may be
omitted, denotes that 2 indicates an operation performed
on a, namely that of taking two of it.
Now this operation performed on the unit is exactly
C2
20 Arithmetical Notions. [33.
analogous to the operation we have called multiplication
performed on a number. Thus, in conformity with the
above notation, it would be reasonable to denote the
operation of multiplying 6 by 7 by the symbol 7.6
instead of 6 × 7 as we have already done. The new symbol,
in the exact form in which we have written it, would be
misleading owing to its resemblance to the point used to
denote a decimal fraction in arithmetic. We may make
it unambiguous and safe by writing it thus 7. (6) which
has the advantage over the former notation 6x7 of
expressing clearly that 7 is a multiple which is employed
to operate on 6 .
33. By analogy with the notation that 2. a means the
result of the operation of taking two of the quantity
represented by a, and that 2. (6) means the result of the
process of taking two of a set of groups of six units each,
the symbol .a will denote the process of taking of a,
and the symbols 후 . (6) and . ( ) will denote the result of
the process of taking of a group of 6 units, and of of
a unit. Just, too, as taking two of any thing or number is
called multiplying that thing or number by 2, so we may
call taking of any thing, multiplying that thing by .
We will agree to make this extension of the word ' multi-
plying ' to fractional multipliers, and will investigate a few
important consequences of the extension. The two or more
fractions multiplied together will still be called factors, and
the result of the multiplication will be called the product
(Art. 22).
34. To multiply a fraction by an integer, as for instance
to find the value of 3. (4), means to take the fraction three
times over. But the fraction itself means that the unit
having been divided into 7 equal parts, 5 of them are
taken. Hence the result of taking three times the
fraction will be to take 3. (5) or 5 × 3 of the same parts, a
36.] Arithmetical Notions. 21

3. (5) 5x3
result which is denoted by or Thus a frac-
7 7

tion is multiplied by an integer by multiplying its

numerator by the integer.
35. The multiplication of an integer by a fraction, as to
multiply 3 by , means to take of 3 units, and is denoted
by .(3).
Let the unit be represented by a square ABCD, and let
two other equal squares DCFE, EFHG be placed in
succession below it. Divide AB into
A P B
seven equal parts, and through each
ofthe points ofdivision draw straight
lines parallel to AG or BH, thus di-
viding the whole set of three squares
into seven equal narrow strips. If
PQ be the line through the fifth D
R
C

point of division, by definition the

area APQG represents the quantity
후. (3), since ABHG represents 3 units.
But PQGA is made up of three equal
parts APRD, DRSE and ESQG, each
S
F
of which contains five parts of a size E

such that seven make up a unit square ;

that is, each of them is & of a unit.
The whole PQGA therefore contains
three times as many, or 3. (5) parts,
of which 7 make a unit; that is,
G Q H

PQGA is 3.(5) of a unit.

7

We thus see that . (3) = 3. (수), or the commutative

law holds when one of the factors of a product is a fraction.
36. Again, let the square ABCD represent the unit. If
AB be divided into seven equal parts, P being the fifth
point of division, and through all the points of division
22 Arithmetical Notions. [37.
lines be drawn parallel to AD or BC, the area APQD
contains five of the equal strips of which seven make up
the whole unit ; consequently it is
A P B
represented by the fraction . Let
AD be now divided into four equal
parts and lines be drawn through
these points of division parallel to
R
S T AB or DC ; let RS be the line drawn
through the third of these points of
D Q C
division cutting PQ in S. Then the
rectangle ARSP evidently contains
three equal rectangles of which ADQP contains 4, and is
therefore of ADQP, or of & of the square ABCD, which
is the unit.
If RS meet BC in Tit is obvious that RTBA is simi-
larly of ABCD and that APSR is of RTBA ; that is,
of of ABCD. Whence it follows that of of ,
or, using the notation ofArt. 33,
(후) = 후 . ( ) ;
that is, the commutative law holds for the product of two
factors which are both fractions.
37. The sum or difference of two fractions which have
the same denominator is evidently obtained by taking the
sum or difference of the numerators for a numerator, and
retaining the denominator.
Thus 5 3 5+ 3
+ ,
7 7 7

since either result means that 5 + 3 parts are to be taken of

a kind such that 7 make a unit.
Similarly 5 3 5-3
,

7 7 7

since either expression means that 3 parts, of a size such

that 7 make a unit, are to be taken away from 5 parts of
the same size.
38.] Arithmetical Notions. 23

The sum of two fractions with different denominators

as and can be expressed as + . By means of the
theorem of Art. 31 it can be simplified by reducing the
two fractions to a common denominator. If the unit be
divided into 7 x 4 or 28 parts the fraction will contain
5 x 4, or 20, of these parts, while contains 3 x 7, or 21 ,
of these parts. Hence the sum of the two fractions con-
tains 20 + 21 , or 41, of these parts, and must therefore be
represented by the fraction ofa unit.
Similarly the difference of two such fractions may be
represented by , and as the larger of these fractions
contains 21 parts such that 28 make the unit, and the
smaller contains 20 parts of the same size, the remainder
left when the latter is taken from the former is one part of
1

that kind. Hence the remainder is 5 of a unit.

These two results can be expressed symbolically in the
following equations :
5 3 (5 x 4) + (3x7) _ (5 x 4) + (3 x 7)
=

(1) +4 =

(7x4) (7x4) (7x4)

20 + 21 41
= = ;
28 28

3 5
(3x7) (5x4) _ (3x7) - ( 5 x 4)
(2) 4-7 (7×4) (7×4) = (7x4)
21-20 1
=

28 28

It is easily seen to follow that, in adding two fractions,

as in adding two integers (Art. 7), the commutative law
holds; for evidently
5 3 3 5

7 4 4 +デ
7
5

38. The value of either of the products .(수) or 후. ( ) is

easily expressed as a single fraction.
By the two sets of parallel straight lines the unit square
has been divided into a series of equal small rectangles, the
24 Arithmetical Notions. [39.
number of which is evidently (7 x 4) or (4 x 7), the product
of the denominators of the two fractions. The number of
these small rectangles contained in APSR is evidently
(5 x 3) or (3 x 5), the product of the numerators of the two
fractions.
Hence as APSR contains (3 × 5) equal parts of such a size
that (7 x 4) make up the unit or ABCD it can be expressed
by the fraction (4
(3 ×x 7)
5) of a unit. Hence since APSR is
5 3
also either . ( ) of a unit or 후. (1) of a unit it follows that
3 5 5 3
=
(3x5)
4 7
(4x7)
The bracket round the second fraction may be omitted if
we please, as there is no other symbol with which confusion
could be made.
39. In Art. 24 it has been shown that in the product
of any three integers the order of multiplication is in-
different, and that
(3x4) x 5 = 3 x ( 4 X
x 5),
or, with the notation of Art. 32,
5. (4. (3)) = (5. (4)) . 3 .
The same law holds in the product of three fractions.
Thus 6
( ( )) ,
which means that of is to be calculated, and then I of
that result taken,
6 (5 x 3) 6.(5 x 3) 6. (5)} . 3
= = =
11 (7 × 4) 11. ( 7 × 4) { 11. ( 7) } . 4
3
=

4
{ 11. (7)}
5 3
56
={
11 }
754
Whence the associative law holds also in the product of
three fractional factors.
40.] Arithmetical Notions. 25

It follows that the product of any number of numerical

factors, whether integral or fractional, is a definite number
quite independent of the order in which the factors are
taken.
40. Let it now be required to evaluate

음 ( +)
11

= 6 (5 × 4) + (3x7)
By the last article this 11
(7x4)
=
6. {(5 x 4) + (3 x 7)} byArt. 37
11. (7 x4)
6.(5x4) +6 . (3 x 7)
=
byArt. 25
11.(7 x4)
=
6. (5 x 4) 6. (3 x 7)
11.(7x4) + 11.(7x4)'
since this last expression means that 6. (5 × 4) parts have
to be taken, of which 11. (7 x 4) make a unit, and then
6. (3 x 7) of the same kind of parts, which is the same thing
as if 6. (5 x 4) +6.(3 x 7) of those parts are taken at once.
But this last result

=
{6. (5)} x 4 {6. (3)} x 7
{11. (7)} x 4 + {11.(4)} x7
by the associative law
=
6.(5 )+ 6. (3)
11. (7) 11. (4)
6 5 6

= 음. ( ) + 음. ( ) .
11 11

Hence 6 6 5 6 3

( + ) = 음. ( ) + 음.( ) . (1)
7

11 11 11
26 Arithmetical Notions. [41 .

By an exactly similar process the student can easily

prove that
6 3 5 6 3 6 5

음 ( - )=
11 7 ( )- 11 ( ),
and thus it follows that the distributive law (Arts. 25, 26)
holds when any or all ofthe numbers involved are fractions.
41. A number expressed in the form of a fraction may be
really an integer. Thus the fraction 4 means that 42
parts are to be taken such that 7 of them make a unit.
Evidently these 42 parts can be arranged in 6 sets each
containing 7 parts. Each set of 7 parts makes a unit.
Hence the whole 42 parts make 6 units. Hence
42
= 6.
7

But 6 was shown in Art. 28 to be the quotient when 42

is divided by 7, and the equation 427 = 6 was proved
from the definition of division and quotient.
Hence 42
= 427.
7

42. Similarly the fraction

45 42+ 3 3

7 7 = 6 +7
In this way any fraction whose numerator exceeds its de-
nominator can be reduced to the sum of an integer and a
fraction whose numerator is less than its denominator. For
we can parcel off from the whole number of parts taken, a
number of parts represented by the denominator, which
will make up a whole unit, and repeat this process until
either none are left or a number of parts less than the
denominator.
A fraction whose numerator is less than its denominator
is called a proper fraction. It evidently represents a quan-
tity less than the unit. A fraction whose numerator is
45.] Arithmetical Notions. 27

greater than its denominator is called an improper fraction,

and evidently represents a quantity larger than a unit.
43. We are now able to return to the subject of division
treated of in Art. 28.
The word division is used in most treatises on Arithmetic
in two entirely different senses which should be carefully
distinguished as logical processes, although they lead to
results identical in appearance and numerical representation.
In the sense already explained it is said that 42-6 = 7
when it has been ascertained that 42 units of any kind can
be assorted into 7 sets of 6 units. The quotient 7 is strictly
a number of times.
There is an entirely different sense in which the same
numerical result 7 can be obtained. Suppose a line 42
inches in length, and let it be divided into 6 equal parts :
the length of each of these parts is easily seen to be 7 inches,
since 6. (7) = 42. Here a process of division performed on
42 with the help of the number 6 gives the same number
7 as before, but the 7 now means a length of 7 inches and
not a number of times.
44. In this second sense we can represent the process of
division of any number, as 45, by any number, as 7, as equi-
valent to the operation of taking of 45, so that we may
say that 45 divided by 7 is of 45, or = . (45) (Art. 33).
1 45 3
But =
= 6+ 7 , (Art. 42).
7 (45) 7

That is to say, that if a line of 45 inches long is divided

into 7 equal parts, each part will be six inches and of an
inch in length.
45. Referring back now to the former meaning and
taking 45 as the number to reason upon, let the 45 units
be continually parcelled off into sets of 7 units. When 6
of these sets have been taken, there will be 3 units left and
28 Arithmetical Notions. [46.
it may be said that 45 contains 7, 6 times with 3 remain-
ing. The number 6 occurs here as in Art. 45, but here
as strictly a quotient, a ' how many times.' The number 3
occurs here as a remainder after the complete sets have been
taken away. It is desirable however for many purposes to
make the two results agree formally in all cases, as it has
been seen they do when the quotient is an integer. This
can be done by a somewhat similar enlargement of the word
' times ' to that which was made of the word ' multiplica-
tion.' By the definition of a fraction, 3 units is of 7
units, since if a length of 7 units taken as a whole be
divided into 7 equal parts, each of them is a unit, and if 3 of
these parts be taken the result is 3 units. It may be agreed
to say that 3 contains 7, of a time, and then it will follow
that 45 contains 7, (6+ ) times, and that therefore we may
write in this sense also,
3

457 = 6 + 7
46. The important result is however that the result of
division, when multiplied into or by the divisor, always
gives numerically the dividend, and as far as numerical
results are concerned it is perfectly true that in all cases
45 3
= 45-7 = 6+
7

47. It may be required to determine how many times.

with the above extension of the word ' times,' one fraction
of a unit contains another, or in other words to divide one
fraction by another.
Suppose it be required to divide by . The fractions
when reduced to a common denominator become re-

spectively (6×7) and (5×11) .

(11 x 7) (11 × 7) '
that is to say, the first contains (6 x 7) or 42 parts of a size
such that (11 x 7) of them make up a unit while the other
48.] Arithmetical Notions. 29

contains (5 x 11) or 55 parts of the same size. Hence the

first fraction is of the second or may be said to contain
42

the second times. Hence we may write

6 5 42 6 7
=(6x 7) X
=

11 5 (Art. 36),
11 7 55 (11x 5)
so that the number of times one fraction contains a second
is obtained by inverting the second fraction and using it as
a multiplier.
Here too it is easily seen that the result of division
whenmultiplied by the divisor gives back the dividend.
For (6 x 7) 5 (6 x 7) x 5
=

(11 x 5) x 7 (11 x 5) x 7
6
6 x(7 x 5)
= =
(Art. 31) ;
11 x (5 x 7) 11
since the value of a fraction is not altered by multiplying or
dividing both numerator and denominator by any integer,
and (7 x 5) = (5 X
x 7) (Art. 22).
48. We can now give a meaning to an expression such
6
IT
as that is, an expression assuming the form of frac- a

tion, in which the numerator and denominator have been

replaced by fractions. Such an expression is often called
a compound fraction, and as no interpretation has yet been
given to it, it is permissible to assume it to have any
meaning consistent with former results.
It has been seen that whichever of the two possible 45

meanings be given to the word ' division ' 457 = 7 .

By analogy it may be assumed that

6 6 5
TT
5 11 7

This meaning will obviously be consistent with previous

definitions in any case in which the fractions, which form
30 Arithmetical Notions. [49.
what by analogy with the definition of a fraction are called
the numerator and denominator of the compound fraction,
become really integers.
49. From this definition of the meaning of a compound
fraction it will follow that the value of such a fraction is
not altered by multiplying its numerator and denominator
by any the same number whether integral or fractional.
Thus if we multiply the numerator and denominator of
the fraction in the last article by the fraction , we have
3 6
X
(3x 6)
4 11 (4 x 11) (3 x 6) (3 × 5)
+
3 5
(3×5) = (4×11) (47) (Art. 48)
4 ×7 (4x7)
(3 × 6) (4x7)
(3 x 5) (
= X
Art. 47)
(4 x 11)
=
(3x6)x(4x7) (Art. 36)
(4 × 11 ) x ( 3 × 5)
(3 x 4) x (6 x 7)
=

(4 x 3) x (11 x 5)
(Art, 24)

=
(6x 7)
(Art. 31)
(11× 5) since (4x3) = (3 x 4)
6 7 6 5
==

11 5 11 7

=+ 7

50. The addition of compound fractions is most easily

effected by reducing each of them to a simple fraction as
in Art. 47, and then using the method of Art. 39. The
particular case in which the fractional denominators of the
two compound fractions are the same demands a little
consideration.
51.] Arithmetical Notions. 31

Let it be required to add the two fractions

5
7
이 and
5
7
We have 6
+ 6 = x + x
II

= ( + ) by Art. 40
( + )
=
6
fi
byArt. 48.
5
4 7
Similarly 6 6 6

II TI II

51. In Art. 4 reference has been made to a symbol o

used in the ordinary system of decimal notation to fill up
a place corresponding to which there is no actual number.
This symbol is one ofgreat importance and one with which
the student should early familiarise himself. In its strict
use it signifies that absolutely none of the quantities con-
sidered are taken. Hence each of the expressions ( 2 × 0),
(0x2), which signify respectively that the number 2 is
taken no times and that no part of two units is taken, give
the same result, namely nought or zero.
Similarly the expression 7. (0) which directs that one
seventh part of nothing is to be taken, must equal nought,
and the expression 07, whether it is to be taken to mean
the number of times that 0 contains 7 or the value of one
seventh part of nothing, has also a null or zero value.
Hence if in any product one of the factors is zero, the
product must also be zero.
It obviously follows that if the product of any number of
arithmetical factors is zero, one of the factors multiplied
together must be zero.
32 Arithmetical Notions.

EXAMPLES .

2 1 31 1
1. Simplify (-6.) × ( +++
X
48 +

2. Divide 22557 days 19 hrs. 30 min. 48 sec. by 3432, and

also by 57 min. 12 sec. Explain clearly in what respects the
two processes differ and what is the difference in the nature
of the results.

3. Add together ,, ; also 3 , 5, 7, and multiply

their difference by 1 .
4. Find the simplest form in which 68. 9d. can be expressed
as a fraction of £1 . Express also 6s. 9d. as a fraction of
11. Os. 3d. Show in each case the size of the parts into which
the whole has been divided.

5. Illustrate by taking the example of a pound sterling as

the unit, that of =
of =
6. If a = 10, b = 2, c = 3 , d = 4, illustrate the truth of
the equations
a- ( b + c - d) = a - b - c + d,
a . (b + c - d) = a.b + a.c - a.d.
7. If a = 1 , b = 2, c = 3, d = 4, find the values of
d
a+
a+ b + c
(1) b+ c-d; (2) C
 CHAPTER II .

ALGEBRAICAL NOTATION .

52. In the last chapter a proof has been given of the

principal laws which govern the combinations of different
numbers with one another. In proving each, definite
numbers have been used, just as in the figures used in geo-
metrical propositions one particular triangle is made to do
duty for all possible triangles. The essence of the proofin the
arithmetical proposition, as in the geometrical, is indepen-
dent of the particular instrument employed, and thus all the
propositions of the last chapter are true for all numbers
which satisfy the conditions supposed to exist in each case.
It is possible by means of a new notation to enunciate
the propositions in a more general form.
53. Let the letters a, b, c, d, p, &c. be employed to denote
numbers whether integral or fractional, then, retaining the
meanings already given to the signs + , - , × , ÷ , . , = ,
which indicate operations, or are equivalent to abbreviations
for words, the following laws have been proved to hold :-
( 1) a + b = b + a,
the commutative law in addition ; (Arts. 7, 39)
(2) a + b + c = b + c + a = c + b + a
= a +c+b = b +a +c = c +a + b ; (Art. 8)
(3) a + b c = -c + a + b = ...... ; (Art. 18)
(4) a + (b + c) = a + b + c ; (Art. 12)
(5) a + (b − c) = a + bc ; (Art. 14)
(6) a-(b + c) = a -b -c ; (Art. 13)
D
34 Algebraical Notation. [53 .
(7) a-(b −c) = a- b + c; (Art. 20)
(8) bxa = axb ; (Art. 22)
(9) a.b = b.a, (Arts. 32, 35, 36)
which last two equations are different ways of stating the
commutative law in multiplication. In the last form the
dot between the letters is usually omitted and ab written
for the product of b and a. Thus a third way of stating
the commutative law is
ab = ba.

(10) a . (bc) = (ab) .c ; the associative law.

(Arts. 24,38)
From this it follows that either of these products may
be written abc and the factors taken in any order. Thus,
similarly, the product ofany number of factors can be shown
to be a definite quantity quite independent of the order in
which the factors are taken.
(11) a(b + c) = ab + ac ; (Arts. 25, 40)
(12) a (b −c) = ab -ac ; (Arts. 26, 40)
These two equations constitute the distributive laws in
multiplication.
a pa
(13) চ
=
; (Arts. 31, 49)
a C a+ c
(14) ㅎㅎ =

; (Arts. 39, 50)

C a- c

(15) = ; (Arts. 39, 50)

ca

(16) == (Art. 37)

As a particular case of this
C ac C

a.a = a = a.a ; (Arts. 34, 35)

( 17 ) = a ÷b ; (Arts. 41 , 42, 45, 48)

54.] Algebraical Notation. 35

(18 ) b. = ab ÷ b = a ; (Arts. 28, 46, 47)

a C ad
(19) =

bc
; (Art. 47)
(20) ax0 = 0 xa = 0 ; (Art. 50)
(21) 0 = 0. (Art. 50)
To these algebraical relations may be added the four
following axioms or general principles of operation.
(1) If equals be added to equals the wholes are equal.
(2) If equals be subtracted from equals, the remain-
ders are equal.
(3) If equals be multiplied by equals the products are
equal.
(4) If equals be divided by equals the quotients are
equal.
It may be noticed that the single letters a, b, c, &c. in
the above algebraical formulæ may stand for the sum or
difference of any number of numbers, may in fact represent
any expression (Art. 11) formed by combining the symbols
for any numbers.
a C a+ c
A relation such as + b
=

b , or any of those in
this article, is called an algebraicalformula.
The signs , > , < , which have not yet been introduced
will be occasionally used in the following senses : -
(a) ab means the numerical difference between a
and b obtained by subtracting the smaller from the larger ;
(β) a > b means that a is greater than b ;
(y) a < b means that a is less than b.
54. In the equations of the last article no limitation is
imposed on the values of the letters a, b, c, except in those
formule which involve subtraction. It will easily be seen
that the proofs given of these will only hold good on the
D2

A
36 Algebraical Notation. [55.
supposition that the subtrahend, or number subtracted, is
numerically less than the number from which it is sub-
tracted. It will be desirable to remove this restriction
before proceeding to deduce any consequences from these
laws. This can be done by somewhat enlarging and
making more precise the definition of a negative quantity
hinted at in Articles 17, 18, 19.
Let it be agreed that the sign – written before a term
shall signify some quality of an opposite kind to that
which is indicated by the term when standing alone or with
the sign + before it. For instance, if the term alone or
with a sign + signify an asset, a term with a sign will
-

represent a debt. Thus let the sum of a person's assets be

represented by a and the sum of his debts by b. If a be
greater than b, a- bis the net amount of the balance in
the person's favour. If, on the other hand, 6 be greater
than a the balance of debt is evidently b - a, and with the
understanding that a debt is to be indicated by placing the
sign - before its numerical representative, the state of the
man's finances is indicated by saying that he has a balance
-( - a).
Now if c be any quantity numerically larger than b - a,
с- (b - a) = c - b + a (Art. 20)
= c + a- b (Art. 18)
= c + (a−b). (Art. 14)
Hence as far as regards the effect of placing them along
with any other number c, the two expressions
-(b - a) and + a-b, which may be written a - b, (Art. 19)
are identical. Hence whether a be greater or less than b
the expression a - b will indicate, by its numerical magni-
tude the amount, and by its sign the nature, of the balance
for or against the person considered.
55. Again, assume that a positive term a represents a
distance measured in one direction along a straight line OX.
56.] Algebraical Notation. 37

Let OA represent this length and from A let a length b

represented by AB be measured in the same direction.
The distance of B from O is evidently a + b as in (1).

X
이 A B
(1)
B A X
(2)

B A X (3)

If the length b supposed smaller than a be measured from

A in the opposite direction the distance of B from O is
a - b, and B is obviously to the right of O as in (2).
If b be larger than a the point B will come to the left
of O, and B will be finally at a distance ba to the left
from O as in (3). If distances measured from right to left
be estimated as negative it will be sufficient to say in (2)
that B is at a distance a - 6 from O, and in (3) that B is at
a distance - (b -a) from O. But, as in the last Article,
-(b - a) = a -6. Hence the statement will include both
cases, namely, that by travelling a distance a to the right
and then a distance b to the left a point will be finally
reached at a distance (a−b) from O, the magnitude of
(a-b) telling the actual distance, and the sign telling on
which side it is measured.
56. Illustrations of this kind show that the use of the
signs and + to indicate opposite qualities in the terms
to which they are prefixed, while not in any way incon-
sistent with the primary meanings of these signs will give
a generality to algebraical formulæ which they would not
otherwise possess.
With this extension of meaning the formulæ (5), (6), (7)
and (12) are easily seen to hold good whatever be the relative
38 Algebraical Notation. [56.
magnitudes of the letters involved. For instance, the
formula (12)
a(b - c) = ab -ac
was proved on the supposition that a, b, c were numbers,
and that b was a larger number than c. If b be a smaller
number than c, (b−c) is what we have agreed to call a
negative number, and is equal to (c -b).
But since c is greater than b
a(c-b) = ac -ab ,
this equation meaning that the numerical values of the
two sides are equal, whence it follows that
-a(c -b) = -(ac-ab).
Again a . (b − c) = a . { -(c - b)} ,
that is, a . (b -c) means that the negative quantity (b-c)
is to be repeated a times. Evidently a collection of
negative quantities must be a negative quantity, the word
negative indicating some property common to them all.
The number of the quantities is not affected by the quality
(negative) of each, and the number of the quantities in the
given case is therefore a(c-b). Since all the quantities
are negative the whole group must be represented by the
symbol -a(c -1).
Hence a . (bc) = -a(c-b)
= -(ac-ab)
= -ac + ab (Arts. 15, 20)
=ab - ac.

Since b is less than c, ab is less than ac, and therefore

ab-ac is a negative quantity. Hence a positive number
multiplied into a negative quantity gives a negative
product.
57. Multiplication by a negative number has at present
no meaning. We may therefore define the process in any
manner we like. It will be advisable to do so in such a
59.] Algebraical Notation. 39

manner as to allow the general laws of Art. 53 still to

hold good for such an operation. Thus if 6 be a positive
quantity, and -a a negative multiplier, it will be desirable
that the meaning put to the process ( -a) .6 shall be such
as to allow the commutative law (9) to hold. So that
( -a) . b = b . ( -а).
But b . (- a) has been shown to be - ba or -ab. Hence
it will be desirable to agree that the operation indicated by
( -a) . b shall give a result -ab, that is, that to multiply a
quantity b by a negative number -a shall give the same
product numerically as that of b multiplied by a, but with
an opposite sign.
58. Another case of multiplication to which no meaning
has yet been given is that of a negative quantity or
number multiplied by another negative number, the
operation indicated by ( -a) . ( - ).
We have seen that consistently with the general inter-
pretation of the sign - as indicating an opposite quality
to + a . ( - b) = -ab.
It seems reasonable to assume that multiplication by
-a shall in all cases give the same product as that given
by a multiplier + a, only with an opposite quality. This
assumption is consistent with all the previous formulæ, and
extending it to the case in point it follows that (-a) . ( -6)
must give a product ab with an opposite quality to -ab.
That is, we have
( -a) . ( - b) = + ab .
59. The results of Articles 56, 57, 58 give one theory of
what is called the rule of signs in multiplication. They
may be summed up in the statement that quantities of like
signs when multiplied together give a positive product,
and quantities of unlike signs a negative product. Another
view of this subject will be given hereafter (Art. 104).
40 Algebraical Notation. [60.
60. It may be assumed in future that if, in using the
formulæ of Art. 53, a negative result is obtained, such
result is capable of interpretation by giving to it an opposite
quality to that which it would have if it were positive.
Thus if a positive result indicates an asset, a negative one
will indicate a debt. If a positive quantity indicate time
future, a negative one will represent time past. When a
positive number denotes a line measured in one direction the
same number taken negatively is to be interpreted as an
equal line in a direction exactly opposite, and so on.
61. The laws of Art. 53 have been proved on the sup-
position that the letters involved represent numbers either
integral or fractional, and by the wider interpretation now
given to the sign - they have been extended to apply to
negative quantities. 1

Logically, it would perhaps be more correct to reverse

this process, and to assume that the laws in question hold
good for all numbers, and from that assumption to reason
out the interpretation that ought to be given to a negative
quantity if such should arise in the course of the work.
It must not be forgotten, however, that the laws of
Art. 53, and therefore all consequences deduced from them,
may apply to other things than numbers. The letters
a and b when placed before another letter even now may
be regarded as representing not a mere number, but a
direction to perform a certain operation called multipli-
cation on another quantity. It may easily happen that
other operations, as well as this particular one of multipli-
cation, may be indicated by the letters a or b, and provided
that with that meaning the laws of Art. 53 hold good, so
will all the consequences deduced from these laws also
be valid.
It will be found, too, that a and b and other letters may
be sometimes employed to represent magnitudes not capable
62.] Algebraical Notation. 41

ofbeing expressed exactly by number ; but, again, provided

the fundamental laws hold, the consequences will be dedu-
cible, and will give valid results in all cases in which the
results are interpretable.
62. It has been seen that the product of three factors, as
a, b, c, has a single definite value in whatever order the
factors be taken as multipliers. This product is written as
abc, the dots being omitted because there is no possibility
of confusion with the symbol for anything else. It is easy
to see that the same must be true for the product of any
number of factors.
If the factors multiplied together are equal, we get a
series of such products as
α, αα, ααα, aaaa, aaaaa
according as there are one, two, three, or more factors.
Such products as these are called powers of the letter a, a
itself being the first power of a, aa the second, and so on.
It is clear that a much simpler symbol than that formed
by writing all the factors in succession can be devised for
expressing any power of a letter. To do so completely two
things must be shown ; first, the letter of which the power
is taken ; and, secondly, the particular power which is to be
taken. The letter a suffices for the first object, and along
with it must be placed in some position the number which
indicates what power it is, hence called the index of the
power. The number is usually written slightly above the
letter and to the right hand. Thus the symbol as indicates
that a is the letter of which a power is taken and that it
is the fifth power which is taken. Hence as is an abbre-
viation for the much longer symbol aaaaa, or, in words, as
means the product of five factors each of which is a.
The second power aa or a² is often called the square of a,
and the third power aaa or a³ is often called the cube of a.
(Compare Art. 86.)
42 Algebraical Notation. [63.
63. If one power of a be multiplied by another power of
a the product is a third power of a, such that its index is
the sum of the indices of the former two. Thus
a5 x a³ = (aaaaa) × (aaaaaaa)
=(aaaaaaaaaaaa)
by repeated application of (10),
or a5 xa = a12.

Similarly, if m and n be any positive integers

a™ xa" = (aaa... to m factors) x (aaa... n factors)
=(aaa ... to (m + n) factors)
by repeated application of (10),
or am xa" = am + n (22)
This law is called the index law in multiplication.
64. The law proved in the last article differs from those
of Art. 53 in the fact that the letters m and n are bound
to be positive integers. In the general investigations of
algebra it is desirable that the formulæ shall not be thus
limited, and we must seek for a method of extending the
formula (22) to the case when mand n are fractions or
negative quantities.
To such a symbol as as no meaning has as yet been
assigned. As in previous cases we may therefore assign
any convenient meaning, and the most advantageous pro-
ceeding will be to give such a meaning to this and similar
expressions as will make the law (22) universally valid.
If this be assumed it follows that

a × a = a + = a² = a,
that is a must be such a quantity that its second power,
or square, is equal to a. In particular cases such a quantity
may be discoverable and expressible numerically. For
instance, if a have the value 4, since 2 x 2 = 4, it follows
that 2 is a value which at may have. It does not follow
that it is the only value.
67.] Algebraical Notation. 43

A quantity whose square or second power is equal to

agiven quantity is called the square root of that quantity.
Thus we have shown that as is the square root of a .
65. Similarly as may be interpreted. Assuming the
truth of(22)
a xa xa = a + xa

= a ++

= a².

Hence as is a quantity whose third power, or cube, is

equal to a³, or with an analogous definition to that of the
last article, as is the cube root of a² . P

66. In a similar way the more general expression a " ,

where p and q are any positive integers may be interpreted.
Assuming (22) to hold, it will easily follow that
P
2+ 2 + 2 + .. to q terms.
4 9 ५
xa x ... to q factors =
=a

But +++ ... to q terms

9 9 9

p + p + p + ... to q terms
=
(Art. 39)
9
qP
=
=p . (Art. 28)
9
P P P
9
Hence a xa xa x ... to q factors = a",
P

that is, a must be such a quantity that its qth power

P

(Art. 62) is a", or a must be defined as the 9th root of a".

67. The square root of a given quantity a, which we
have seen is devolved by a¹, is also frequently represented
by the symbol a or √a. The cube root of a is similarly
44 Algebraical Notation. [68.
3
denoted by th root of ap will be analogously
a and the qth
represented by Var. Hence we may write
P

a ap

where the sign = means that the expressions connected by

it are two different ways of representing the same thing.
It is merely a question of convenience as to which of the
two notations is employed.
68. If any integral power of a as a be divided by a
lesser power as a³,
a5
a5a3 = by (17) ofArt. 53 .
aaa.aa
=

aaa

= aa
by (18) of Art. 53.
= a² = 5-3.

Hence, if m and n be any integers of which m is

greater than n
am a = am- .

This law is an immediate consequence of (22), for by (22)

am- n xa" = am- n + n = am,
and therefore by (18) of Art. 53
aman am-n ...... (23)
69. It may happen in deducing results from the for-
mulæ already obtained that such a symbol as a-3 is met
with. It will, as before, be convenient to assume that the
meaning given to this expression shall be such as to make
the law (22) valid in this case.
Assuming (22)
a5 xa-3 = 5-3 = a2 .

But a5 = a2 + 3 = a² x a³ ,
71.] Algebraical Notation. 45

therefore (a² x a³) x a-3 = a²,

or a² x (a³ xa-3) = a²,
whence a xa = a² 2
by (18) and axiom (4)
= 1, by (27)
3
Hence a = 1 ÷ a³,
1
=
......
by(17)
1
The fraction is frequently called the reciprocal of a.
a

Hence we have arrived at the conclusion that a-3 must

denote the reciprocal of a³.
70. Again, take the symbol of a-" and assume that law
(22) holds good.
Therefore am + nxa- n = am + n- n
m
=a ;

or amxanxa- " = am .
Therefore a " xa " = am÷ am

= 1 ;
u-D 1
whence
ani
or a-" must denote the reciprocal of a".
The symbol aº must have the value unity, if law (22) is
tohold : for
am xa = am + 0 = am,
Therefore a = amam = 1 .

71. It has been shown that the only possible meaning

p
9

that can be given to the symbol a so as not to disobey

the law (22), is that it is the 9th root of a". It remains
to be shown that this meaning will in all cases be
such as to obey (22). As, however, there will not be much
occasion at present to employ fractional or negative indices,
46 Algebraical Notation. [72 .
the fuller consideration of this subject will be deferred until
later on in the book (Chapter VIII).
72. By law (22)
amxam am + m = a2m, since 2 = 1 + 1, (Art.2)
amxam xam a2m xam = a3m, since 3 = 2 + 1, (Art. 2)
or, as these results may be written,
(am)² = a2m ;
(am)³ = a³m.
More generally it follows that if n be any positive integer
(am)n = am xam xam x ... ton factors ;
= am + m + m + ... to n terms ;
(by 22)
= anm ; by definition of nm.
Thus if m and n be any two positive integers
(am)n = amn.
It of course follows equally that
(an)m= amn = (am)".... (24)
For instance

(a³)5 = a³ x a³ × a³ × a³ × a³ = a³ + 3 +3+ 3 + 3 = 5. (3)

= 3. (5)
= (a)5+ 5 + 5 = a5 x a5 x a³ = (a³)3:
The laws (22), (23), (24) will be afterwards shown to hold
good for fractional and negative indices interpreted accord-
ing to the preceding articles.
73. By the definition of an index
(ab)² = ab.ab
= aabb
by (9) and (10)
= a²b .

Similarly (ab) = ab.ab.ab

= aaabbb
= a³%3.
And (al)"= (ab). (ab).(ab) ... n factors,
= aaaa ... n factors.bbb ... n factors,
or
(ab)" = a*b" .... (25)
75.] Algebraical Notation. 47

This law also will be shown later on to hold for all indices,
negative or fractional, defined as in Articles 67 and 69.
74. It has been suggested that other interpretations may
be given to a, b, &c. in the formulæ, and that valid results
may be deduced from these formulæ by means of such in-
terpretations. An illustration which may be useful here-
after will now be given.
Let the symbol a denote the transference or step of
a moving point through a given space as OA, in a given
direction, and let b denote a transference through another
given space in some direction not B
necessarily the same, as AB. Then C

a + b will denote the effect of making

first the step a, and then the step b.
By the two steps the point is moved Ο A

from O to B. Through O and B

draw OC and CB parallel to AB and OA respectively. A
transference from Oto Cis exactly the same in magni-
tude and direction as one from A to B, while the step
from C to Bis exactly the same as that from O to A.
Hence a step first from O to Cand then from C to Bis
denoted by b + a. The effect of these two steps is a trans-
ference from O to B. Hence F C B

a + b = b + a,
and either of these expres-
sions denotes a transference
アー

over OB. A A

75. If AO be produced to
۶

A' making OA' equal in

length to OA, and CO to C',
E C D
making OC' equal in length
to OC, and through A' and C' straight lines be drawn
parallel to CC' and AA' so as to complete a parallelogram
BDEF, very elementary geometry will show that OE and
48 Algebraical Notation. [75-
OB are in the same straight line and equal in magnitude,
and similarly OD and OF.
The step OA being denoted by a, the step OA' is, in
accordance with Art. 54, denoted by -а. Hence a
transference from O by C to Fis denoted by b-a, while a
transference from O to F through A' is denoted by -a + b.
Hence b - a = -a + b.
Similarly, a transference from O to E may be denoted by
either of the equivalent symbols -a- b or -b - a, and a
transference from O to D by either of the equivalent sym-
bols a - bor -b + a.
76. Again, a transference from O to E is exactly equal in
magnitude, but opposite in direction to the transference
from O to B. Hence with the recognised interpretation
of the sign -,
-(a + b) = -a - b.
Similarly, the transference from O to Dis equal in mag-
nitude, but opposite in sign to that from O to F. Hence
a - b = - (b -a) .
Similarly, if c represent a third step or transference it is
easy to show that the formulæ (1) to (7) will all hold.
77. In the numerical interpretation previously given to
the equations of Art. 53, a. (1)-that is, the result of per-
forming the operation of multiplication indicated by a on
the unit has been written briefly as a. If the same be
done with the new interpretation of a as a transference, the
symbol a will denote equally the transference of the point
or the product of (i.e. the thing produced by) the trans-
ference that is, the straight line OA.
The symbol b . a will therefore reasonably indicate the
result of performing the operation 6 on the line a. The
product of, or thing produced by, the transference of the
line a over the distance indicated by b is evidently the area
of the parallelogram OACB.
79.] Algebraical Notation. 49

The product of a . b is evidently the same thing. Hence

b.a = a.b
and either of these products may be taken to represent the
area ofthe parallelogram of which a and b are adjacent sides.
This result agrees with and includes the former numerical
result in the case when a
G
and b are at right angles to
each other.
F
C
78. In a similar manner if
c represent a transference not
H
in the same plane as the
A
former two, let a parallele-
piped be described with OA, E

OB, OC as adjacent edges.

The product ab representing B

the area of the parallelogram

OAEB, the product c.ab will represent the thing produced
by the transference of this parallelogram over the step OC,
that is, the volume of the paral-
c'
lelepiped. But this same quantity
can be also represented by a.bc or B

c.ab. Hence
C
a.bc = b.ca = c.ab = (abc).
79. If a step OB be represented
F'
by 6, the step BC by cand the step
OAby a, the step OC is represented H

by b + c. Hence if the parallelogram A

OCFA be completed, it is repre- F

sented by a . (b + c).
But, assuming that a, b, c are all in the same plane, if
BHbe drawn equal and parallel to OA or CF, and FH, AH
be joined, it is easily seen geometrically that
OAFC = OAHB + □ BHFC ;
..a (b + c) = ab + ac.
E
50 Algebraical Notation. [80.

If CB be produced to C' and BC' be made equal to BC,

OC' is evidently represented by b- c, and if the parallelo-
gram AOC'F' be completed it is represented by a . (b - c).
But AOC'F' = □ OAHB− □ BHF'C',
... (b − c) = ab - ac.
80. These illustrations have been given now, not with
the intention of following up the interpretation thus
suggested, but that the student may see more clearly that
the equations of Art. 53 admit of still greater generality
of meaning than has been given even by the introduction
of the negative quantity, and that Algebra is something
very much wider in its scope than a mere substitution of
letters for numbers to aid in the solution of general arith-
metical problems.

EXAMPLES .

1. If a = 4 , b = 3 , c = 2 , d = 1 ; find the values of

(1) a + b-(c+d) ;
a+ 2c a- c

(2) b+ d_b-d ';

(3) (a + b + c) (b + c + d) -(a + b)².
2. With the same values of the letters, show, by finding the
arithmetical value of each of the quantities involved, that
(1) a (b + c + d) = ab + ac + ad ;
a b ad
‫ب‬ =
(2) C be
;

a b ad +be
=
(3) C
+d cd

3. Find the values of

(a + b + c) ; (a + 2c) ; (b + d) ,
a, b, c, d having the same values as before.
80.] Algebraical Notation. 51

4. Illustrate the formulæ

(1) (ab)" = " b" ;
(2) amx am+ n ,

in the cases when a = 2, b = 3, m = 4, n = 5, by finding the

numerical value of each side in each equation.
5. If a = 7, b = 5, c = 4, find the values of
(1) √(a + b + c) -2a + c- 2b ;
3

(2) Na+ √b-c- 26 + 2 (a− c)} ;

(3)
√2a-b+ 2
/2c-a
(2a-b) -(2c-a)
6. If x = 3, y = 2, find the values of
(1) (x² -xy + y²)² ;
(2) x² - 2xy + 3x²y² - 2xy + y² ;
(3) (x² + xy + y²)² ;
(4) x² + 2x²y + 3x²y² + 2xy + y² .

E2
 CHAPTER III .

ELEMENTARY OPERATIONS. ADDITION AND SUBTRACTION .

81. Any combination of symbols, such as a, b, c, when

connected by the signs of operation x , . ,, + , - , & c.,
is called an algebraical expression. Those parts of an
expression which are connected by the signs + and are
called the terms of the expression (Art. 11).
82. An expression, such as
a²b + a³ + 5ax²-2x3,
which does not assume a fractional form, is called an
integral expression.
If none of the terms involve factors raised to fractional
indices the expression is called a rational expression.
5a- b
If the expression assume a fractionalform, as it
a+b '
is called afractional expression.
83. It may of course happen that an integral algebraical
expression may, for some values of the letters, have a
fractional value, while the fractional expression may have
an integral value. Thus if a have the value , and 6 the
value , the integral expression a - 2b has the value
-2x ; that is, -1 or . On the other hand, the
a

fractional expression has for the value of a and b the

value ,which is 3. The words ' fractional' and 'integral,'

86.] Addition and Subtraction. 53

as applied to algebraical expressions, thus apply to theform

of the expression and not to its value for particular values
of the letters involved.
84. A single term may consist of one factor, as a, or
may be the product of two or more factors, as a², abc, a² c.
If one of the factors is a definite number, as the factor 3
in the term 3 abc, that factor is usually written first, and is
called the coefficient or sometimes the numerical coefficient
of the term.
85. Terms which involve the same combination of
literal factors, as 3abc, 7abc, are called like or similar
terms.
Terms which involve different combinations of literal
factors are called unlike or dissimilar terms ; as 5 a³, 7 a²b .
86. A term involving a single literal factor, as 5a, 7 b, c,
is called a term of one degree, or of one dimension, or a
linear term. The last two designations arise from the fact
that single letters, as a, b, c, are often employed to denote
the lengths of lines.
A term involving the product of two literal factors,
as a², ab, bc, is called a term of two dimensions, or of the
second degree. In Articles 23 and 77 it is seen that a
product, such as ab, may represent an area, having two
dimensions, length and breadth.
A term involving the product of three literal factors
is similarly called a term of three dimensions or of the
third degree, as a³ , a²b, abc. In Articles 24 and 78 it is
shown that a volume having three dimensions-length,
breadth and thickness-may be represented by the product
of three factors each of which represents a line.
Most people are unable to conceive of actual physical
quantities possessing more than three dimensions, but the
phrase is a convenient one and in analogy with the
language used up to this point ; a term containing four
54 Elementary Operations. [87.
literal factors is called a term of four dimensions or of the
fourth degree, and so on.
Thus the terms 7a362 and 11a²bc4 are of five and nine
dimensions respectively, for a³ 12 is merely an abbreviation
for aaabb, and a²bc4 is an abbreviation for aabbbcccc, and
these terms contain five and nine literal factors respectively.
87. An expression consisting of one term is called a
monomial: one consisting of two terms, as a + b, or 2a- 36,
a binomial ; and the words trinomial, quadrinomial, polynomial
are applied to expressions containing three, four, or many
terms respectively.
88. An expression of which all the terms are of the
same dimensions is called a homogeneous expression. Thus
a + b + cis a homogeneous expression of one dimension or
the first degree or a linear expression ; 3a² + 2ab - 502 is a
homogeneous expression of two dimensions or of the second
degree.
A homogeneous expression will be spoken of as an
expression of four, five, &c., as the case may be, dimensions.
89. Before proceeding to explain and illustrate the
processes of addition and subtraction as applied to Alge-
braical expressions, it will be well to reproduce those
formulæ of Art. 53 which will be chiefly employed. They
are, retaining the numbering of that Article,
(4) a + (b + c) = a + b + c,
(5) a + (b −c) = a + b — c,
(6) a-(b + c) = a - b - c,
(7) a-(b - c) = a -b + c,
(9) ab = ba,
(11) a(b + c) = ab + ac,
(12) a(b - c) = ab - ас.
90. The addition of two algebraical expressions, if none
of the terms are like, can only be represented by connecting
the second expression with the first by means of the
93.] Addition and Subtraction. 55

sign + . Thus if it be required to add a + b - ctop + q + r

the process and the result can only be represented in the
form
p + q + r + (a + bc),
or, by (4) and (5), p + q + r + a + b - c.
91. If however some of the terms in the one expression
are like to some in the other, a considerable simplification
may be effected. Thus let it be required to add
3x² + 4x + 5 to 2x² - 3x-2,
the result is expressed as
2x²- 3x- 2 + (3x² + 4x + 5)
= 2x² - 3x - 2 + 3x² + 4x + 5
= 2x² + 3x² + 4x- 3x + 5-2
= (2 + 3) x² + (4-3) + (5-2) by (11) and (12)
= 5x² + x + 3 .
Two like terms can be replaced by a single term like to
them, its coefficient being the sum or difference of the
coefficients of the original terms, according as those terms
have the same or opposite signs.
92. The above process is more conveniently represented
by writing down the expressions to be added in two
horizontal lines, like terms being placed in the same
vertical line, thus 2x² - 3-2
3x² + 4x + 5
5x² + x + 3

The sum is then obtained by adding or subtracting the

coefficients of like terms .
93. It is convenient for many purposes to alter somewhat
the phraseology hitherto employed in regard to subtraction.
The expression a- b has been said to denote the process
of subtracting b from a : it will in future be more usually
spoken of as the result of adding -b to a, and will be
called the algebraical sum of a and -b.
56 Elementary Operations. [94.

Thus a + b-c- d is the algebraical sum of a, b , c

and -d.
94. With this modified phraseology the rule for the
addition of two or more expressions becomes : The co-
efficient of any term in the sum of the expressions is the
algebraical sum of the coefficients of the like terms in the
separate expressions.
The practical method of adding any number of expressions
is, as in the case of two, to write them down in separate
horizontal lines so that as far as possible all like terms shall
come in the same vertical line. The addition is then
easily performed by taking the algebraical sum of the
coefficients of each set of like terms and writing that as
the coefficient of the same kind of term in the sum. As an
example, the sum of the four expressions ³ - 3x² + 3x - 1 ,
23-3, 3x² - 4x + 5 and x² - 2x - 5 will be best found
in the following manner.
All the expressions should be written in the same order
either of descending powers of a (that is, with the highest
power first, and so on) or in the opposite order. By (2)
and (3) of Art. 53 the order of the terms may be changed
at pleasure.
x³- 3x² + 3x- 1
2x3 3

+ 3x² - 4x + 5
2x3 + x2 -5

x² + x² -4
The algebraical sum of the coefficients of 2³ is 1 + 2-2
or 1, that of those of ² is -3 + 3 + 1 or 1 , of those of a
is 3-4 which, by Art. 54, is -( 4-3) or -1 , while the
algebraical sum of the terms independent of a is
-1-3-5-5 or -4 .

95. The subtraction of one expression from another can

be similarly effected.
97.] Addition and Subtraction. 57

Let it be required to subtract

x³- 3x² + 3x- 1 from 23 - 5 x² + 4x - 3.
The result of the process can be expressed as
2x3 - 5 x² + 4 x 3 - (x³ - 3x2 + 3x- 1 ),
which

= 2 x³- 5x² + 4x - 3x3 + 3x² - 3x + 1 by (6) and (7)

= ( 2-1) x3 + ( - 5 + 3) x² + (4-3 ) x - 3 + 1 ,
in which form it will be seen that the coefficient of each
term is obtained by changing the sign of every term in the
subtrahend, or expression to be subtracted, and then
proceeding to take the algebraical sum of the coefficients.
96. This process is more easily performed by writing
the subtrahend below the expression from which it is to be
subtracted, mentally going through the process of changing
the sign of each coefficient in the subtrahend and taking
the algebraical sum of the coefficient so changed and that
of the like term in the other expression as the corresponding
coefficient in the resulting difference.
Thus the above example will be written
2 x³ - 5 x² + 4x - 3
x³- 3x2 + 3x - 1
x² - 2x² + x - 2
The coefficient of x³ will be 2-1 or 1 ; that of a² will
be - 5 + 3, which = - (5-3) or -2 (Art. 54) ; that of a
is 4-3 or 1, and the numerical term is -3 + 1 or -2.
97. Again, let it be required to subtract
x²- 2x² + 3x - 1 from x² - 2x² + 2 .
Here the terms in each expression are not all like to a
term in the other, but those which are like can be written
as before in a vertical line,
x²- 2x3 +2
x² - 2x² + 3x - 1
x² - 3x3 + 2x² - 3x +3
58 Elementary Operations. [98.
There is no term 24 in the subtrahend. Hence the
coefficient of x in the difference is 1 ; that of r³ is -2-1
or -3 ; that of æ² is + 2, that of x, -3, and the numerical
term 2 + 1 or 3 .
98. The validity of these rules for addition and sub-
traction depends on the formulæ repeated in Art. 89 with
the implied extension of (4), (5), (6) and (7) mentioned
in Articles 15 and 20, that whatever be the number of
terms within a bracket, the bracket may be removed and
the signs left unaltered, if the bracket be preceded by a
sign + ; while, if it be preceded by a sign -, the signs
of all the terms within it must be changed when the
bracket is removed.
99. The enclosure of expressions within a bracket
(Art. 11) implies that the whole included expression is to
be treated as one quantity and subjected to the operation
indicated by the symbol preceding it. The same notation
sometimes occurs in a more complicated form. Thus as
the result of an algebraical calculation an expression of the
following form may be met with :
a - [ { b- (c - d)} - e].
This means that the whole expression within the square
brackets [ ] is to be subtracted from a. That expression
is the result of subtracting e from the expression within
the brackets { }, while the latter again is the result of
subtracting (c-d) from 6 .
By means of the formulæ (4), (5), (6) and (7) these
brackets may be removed in succession, beginning with
the interior one. Thus

a- [ { b- (c- d) } -e] = a − [ {b − c + d} -e]

= a - [b - c + d - e]
= a- b + c - d + e.
101.] Addition and Subtraction. 59

100. As another instance take the expression

2a- [a - b- (a − b)} ] .
Removing the brackets one at a time, beginning from
within, this expression becomes in succession equal to
2a- [a- { b - a + b } ]
= 2a- [a - b + a_b]
= 2a - a + ba + b
= 26,
since the coefficient of a becomes 2-1-1 or zero.
101. The brackets may also be removed in the opposite
order, thus
2a- [a- { b- (a − b) } ]
= 2a - a + { b- (a − b) } by(7)
= 2a - a + b- (a − b) by(5)
= 2a - a + ba + b by (7)
= 26 as before.
An additional notation for brackets, which is sometimes
useful and frequently used, is the placing a line above the
terms which are to be treated as in a bracket. Thus the
notation
a - b- c

is often used to denote the same thing as a -(bc).

EXAMPLES.

1. Add together a + b + c, a + b - c, a - b + c, c + b - a.
2. Add together x³ - 3x² + 3x- 1 and x² + 3x² + 3x + 1 .
Find the value of each of these expressions when a has
the value 3, and verify that the sum of these values is the value
oftheir sum.

3. Add together 2x² - 3x²y+ 4xy² -y³, x² + 2x²y- 3y²,

5xy² - 3x² + 4y³, and x² +x²y- 9xy².
4. Subtract x³ - 3x² + 3x- 1 from x² + 3x² + 3x + 1 , and
taking 3 as the value of x verify that the difference of the
values of the two expressions is the value of their difference.
60 Elementary Operations.
5. Add together x² + 4x² + 4 and x² + 3x -2x² - 3,
and from their sum subtract x - x² - 4x² - 4x.
6. Simplify
(a + b + c + d) -(a - b + c - d) + (a - b - c - d) -( a + b - c + d).
7. Remove the brackets from
a-[b - c-(d- e -f)} ].
8. Remove the brackets from
7a- [ 2a + { b- (4a + a - b)} ]
and reduce the resulting expression to its simplest form.
9. Add together
-10a + 56-6c, 7b + 9a - 3c, 8c - a - 96 ,
and verify the result when a = 4 , b = 3 , c = 2.
10. Simplify 5a- [2b- { a + ( 26-4a)}].
11. Add together
a² -2ab + b², b² - 2bc+ c², c² - 2 ca + a², and
2ab + 2bc + 2ca -a² -b² -c².
12. If a = 3 , b = 2, c = 1 , find the values of
(1 ) (a² - 2abc + b³) + (63-2abc + c ) + (c3-2abc + a³) ;
(2) (a + b + c) {(a² - 2ab + b²) + (b²- 2bc + c²) + (c²- 2 ca + a²)}.
 CHAPTER IV.

ELEMENTARY OPERATIONS . MULTIPLICATION .

102. The laws which govern the process of multiplication

of algebraical symbols are, still retaining the former
numbering,
(9) ab = ba,
(10) a . (bc) = (ab) . c = abc,
(11) a(b + c) = ab + ac,
(12) a(b − c) = ab - ac,
(22) am xa" = am+n , (Art. 64)
(23) ama" = am- n, , (Art. 68)
(24) (am)" = amn, (Art. 72)
(25) (ab)" = a*b", (Art. 73)
in the last four of which the letters m and n are at present
supposed to be restricted to mean whole numbers.
103. Two single terms involving the same letters can
be multiplied together by taking the product of the
numerical coefficients as coefficient of the product, and by
writing each of the letters with an index equal to the sum
of the indices of that letter in the two given terms .
Thus 3 abcx 5a2b3c
= 3 × 5.abc a²b³c
= 15 aa²bb³сс
= 15 a3b4c2 by (22)
the various steps depending on the facts concisely indicated
in (9) and (10) that the value of the product of any number
of factors is a definite quantity quite independent of the
order of the factors. (Art. 62.)
62 Elementary Operations. [104.

Similarly 2a²b x 7 ab³c² = 14a3b4c2 ;

-5abc × 3ab³ = -15a²b5c.
The distributive laws (11) and (12) show that a single
term can be multiplied into an expression containing many
terms by multiplying it into each term in succession, and
writing down the products with their proper signs.
Thus (2a2b - ab² + 4ac²) × 3abc = 6a3b2c - 3a²bc + 12a²bc3 .
104. Let it be required to multiply (a +b) by (c +d).
The process is indicated by the symbol
(c + d) . (a + b) .
But by (11)
(c + d) . (a + b) = (c + d)a + (c + d)b
= a(c + d) + b(c + d) by(9)
= ac + ad + bc + bd by (11)
Similarly
(c + d) (a − b) = (c + d) a-(c + d)b by (12)
= a(c + d) -b(c + d) by(9)
= ac + ad- (bc + bd) by (11)
= ac + ad-bc - bd by (6) ofArt. (53)
Again
(c-d) . (a− b) = (c - d)a- (c-d)b by(12)
= a(c - d) -b(c-d) by(9)
= ac - ad-(bc-bd) by(12)
= ac- ad - bc + bd by (7) ofArt. (53)
Comparing these results we see that the product of one
binomial expression into another is obtained by taking
the product of every term in the one by each term in
the other, and placing a + sign before a product of two
positive terms or two negative terms, and a - sign before
the product of a negative into a positive term.
105. We have thus from a different point of view
arrived at the same rule of signs for multiplication as was
obtained in Articles 56, 57, 58.
107.] Multiplication. 63

By this second method we apparently avoid the difficulty

of explaining the meaning of the multiplication of a
negative quantity by a positive one or of two negative
quantities together. The evasion is only apparent if it be
wished to have perfectly general formulæ; for the quantities
a- b, c - d in the last article may either or both be negative
unless the results are to be accepted as true only when a is
greater than b, and c than d.
The last of the three formule of the Article 104 will
evidently give the same value for (c-d) (a -b) as for
(d - c) (b - a), which again leads to the theorem that the
product of ( -a) . ( -b) = ab .
106. Similarly
(a + bc) . (p +q- r)
= (a + b− c)(p + q) -(a + b - c)r by(12)
= (a + b−c) p + (a + b - c)q- (a + b - c)r by ( 11)
= p(a + b −c) + q(a + b -c) -r( a + b -c) by (9)
= pa + pb -pc + qa + qb -qc- (ra + rb - rc)
by (11) and (12)
= pa +pb -pc + qa + qb - qc - ra - rb + rc
by (6) and (7) of Art. (53)
From this result we see that the rules of multiplication of
Art. 104 apply to the product of two trinomial algebra-
ical expressions, and by a similar process those rules can be
shown to hold for the product of any two algebraical
expressions, whatever may be the number of terms in each.
The two algebraical expressions are called thefactors of the
result formed by multiplying them, and the resulting
expression is called the product. (Art. 22.)
107. If the terms in the expressions be unlike, no farther
reduction can be made in the product, but ordinarily some
of the terms in this latter are alike and can therefore be
replaced by a single term as in addition.
64 Elementary Operations. [108.
108. It most frequently happens that the multiplier and
multiplicand contain the same letter or letters, and can
each be arranged in descending or ascending powers of one
of the common letters. Let it for instance be required to
multiply x³ + 2x² + 2x + 1 by x² + 3x + 2. The process is
best conducted in the following manner. Write the
multiplier beneath the multiplicand, putting the highest
powers of the common letter æ in the two expressions in
a vertical line
x² + 2x² + 2x +
x² + 3x + 2

5 + 2x² + 2x² + x²
+ 3x² + 6x3 + 6x2 + 3x
+ 2x + 4x² + 4x + 2
4
x5 + 5 x² + 10x³ +11 x² + 7x + 2

Drawing a horizontal line below the multiplier, write

immediately under it the sum of the products of the first
term of the multiplier into all the terms of the multipli-
cand, these products being obtained according to the rule
of Art. 103. In the line below this write the products of
the second term of the multiplier into the different terms
of the multiplicand, and so on, until all the terms of the
multiplier have been used. Then adding together the
coefficients of like terms the final product is obtained and
written below in its simplest form.
It will be noticed that the successive lines in the product
are written continually more and more to the right hand.
This is merely a matter of convenience, in order to make
like terms come in the same vertical line.
11.] Multiplication. 65

109. Another instance will be given in which some of

the terms are negative.
x²- 2x3 + 3x² - 2x +
x² + 2x² - 2x- 1

x²- 2x + 3x5-2x² + x3
+ 2x - 4x5 + 6x4 - 4x3 + 2x²
-2x5 + 4x² - 6x + 4x² - 2x
-

x² + 2x³ - 3x² + 2x-

27 -3x5 + 7x - 7x3 + 3x² -1 .

Here the successive rows of multiplication are the results

of multiplying the multiplicand in succession by the terms
of the multiplier, regard being had to the rule of signs of
Art. 104.
110. The product of two algebraical expressions can
thus in all cases be found by a direct process. When the
two expressions to be multiplied contain any common
letter, the term of the product which involves the highest
power of that letter is the product of the terms in the
multiplier and multiplicand which respectively contain the
highest powers of the same. This remark will be found to
be of importance when the subject of the inverse operation
of Division comes to be considered.
111. When a single term is multiplied by a single term,
the dimensions of the product (Art. 86) are the sum of the
dimensions of the factors. This follows since the number
which expresses the dimensions of any term is merely the
number of literal factors involved. Thus, in the example
of Art. 103, the term 3abc is of three dimensions, since
it contains three literal factors a, b, c ; the term 5a²b³c,
which means 5aabbbc, since it contains six literal factors,
is of six dimensions. The product 15a3b+c² containing
(3 + 4 + 2) literal factors is ofnine dimensions.
F
66 Elementary Operations. [112.
112. If a homogeneous expression (Art. 88) be multiplied
by another homogeneous expression, the product will be a
homogeneous expression whose dimensions are the sum of
the dimensions of the two original expressions.
This follows since, by Art. 106, the product of the ex-
pressions is the sum of all the products of any term in the
first by any term in the second, and, by Art. 111 , each
single product is a term of dimensions equal to the sum of
the dimensions of the two factors .
If therefore in multiplying two such expressions together,
the student discovers that he has obtained terms of different
dimensions, he may be sure that there is an error in his
work, and the error will be in those terms whose dimensions
are not the sum of the dimensions of the two original
factors.

113. The expressions used in the examples of multiplica-

tion in Art. 108 and 109 were not homogeneous. Any
one of them could easily have been altered into a homo-
geneous expression by introducing a letter a, the value of
which is taken as unity. For instance, the second multi-
plier x² + 2x² - 2x- 1 could be written as
x² + 2 ax² - 2a²x- а³,
with which it is identical if a have the value unity. The
multiplicand could similarly have been written as
x²- 2ax³ + 3a²x² - 2a3x + a4 .
The product of these expressions is found to be
x² - 3a²x5 + 7a3x² - 7a4x³ + 3ax²-a ,

a homogeneous expression of seven dimensions, which

reduces to the result of Art. 109 by replacing a by unity.
114. One or two examples of multiplication are of
sufficient importance to be specially examined.
(a+b)². This means, by Art. 62, (a + b) (a + b).
117.] Multiplication. 67

Hence (a + b)² = (a + b) (a + b)
= (a + b) a + (a + b) 6 by(11)
= a (a + b) + b (a + b) by (9)
= aa + ab + ba + bb by(11)
or (a + b)² = a² + 2ab + b² by (9) and Art. 62 .
115. Again
(ab)² = (ab) (a − b) by Art. 62
= (a − b) a- (a − b) b by(12)
= a (a − b) -b (a − b) by (9)
= aa - ab- (ba-bb) by(12)
= a² - ab- ab + b²
by Art. 62 and (7) ofArt. 53
or
(a − b) ² = a² - 2ab + b².
This result can be deduced from that of Art. 114, if
we assume that the latter holds universally, by substituting
-6 for b. Thus we get
2
{a + ( - b) } ² = a² + 2a ( -b) + ( - 6)2
= a² - 2ab + b² by Art. 56 and 58.
116. Another result of very great importance is the
following :
(a+ b) (a − b) = (a + b) a (a + b) b by(12)
= a (a + b) -b (a + b) by(9)
= aa + ab- (ba + bb) by(11)
= a² + ab - ab - b2
by Art. (62) and (6) ofArt. 53
or
(a + b) (a− b) = a² -b² .
117. The reasoning of the last three Articles does not
depend on the values of the quantities a and b, and the re-
sults will consequently hold whatever a and b may represent.
Thus, if a represent 2x + 3y, and b represent 4z, a + b is
(2x + 3y) + 4z, and a-b is (2x + 3y) -4z.
Hence the formula of the last article gives
{ (2x + 3y) + 4z } {( 2x + 3y) -4z} = ( 2x + 3y)²-(42)2 .
F2
68 Elementary Operations. [118.
But again, by the formula of Art. 114,
(2x + 3y)² = ( 2x)² + 2 (2x) (3y) + (3y)²
= 4x² + 12xy +9y² by Art. 103
and (42)² = 1622 by Art. 103.
Hence
(2x + 3y + 4z) (2x + 3y- 4z) = 4x² + 12xy + 9y² - 16 22.
In this manner the results of Art. 114, 115, 116 can
frequently be used to abridge the regular process of multi-
plication.
118. An example of some importance is afforded in the
product of the four expressions
a + b + c, b + c - a, c + a - b, a + b -c.
The value of the product is not affected by the order in
which they are multiplied together, by (10) and Art. 62.
Taking the product of the first two we have
(a + b + c) (b + c- a) = {(b + c) + a } { (b + c) -a}
= (b + c)² - a² Art. 116
= b² + 2bc + c² - a² Art. 114.

Again, the product of the last two is

(c + a - b) (a + bc) = { a-(b − c) } { a + (bc)}
by (7) ofArt. 53
= a²-(b −c)2 Art. 116
= a² - b² - 2bc + c²) Art. 115
= a²-b² + 2bc-c² ;
by (6) and (7) ofArt. 53 .
Hence the product of all the four factors, or
(a+b+c) (b +c-a)(c +a−b) (a + b −c)
= (b² + 2 bc + c²-a²) (a² - b² + 2bc-c²)
= { 2bc-(a² - b² - c²)} { 2 bc + (a² - b²_c²)}
by (6) and (7) of Art. 53
= (2bc) -(a² - b² -с2)2 Art. 116
= 4 b²c² - {(a² - b²) - c2} 2 Art. 103
2
= 4b2c²- { (a² -b²)²-2 c² (a² - b²) + (c²)2} Art. 115
119.] Multiplication. 69

2
= 40²²-(a² -b²)² + 2 c² (a² -b²)-c
(6) and (7) ofArt. 53, and (24) ofArt. 102
= 4b2c2- (a² - 2a²b² + b²) + 2c2a² - 2 c2b2 -c
Art. 115 and (24) ofArt. 102
= 402c2-26²c² - a + 2a²b² - b² + 2 c²a² - c4
by (6) and (7) ofArt. 53
= 2b2c² + 2c2a² + 2a2b2 - a* - * - c
by (2) and (3) of Art. 53.
119. In Art. 114, 115 formulæ have been found giving
the values of (a + b)² and (a − b)². Later on a formula
will be proved by which any power (Art. 62) of a
binomial expression (Art. 87) can be written down. The
form of the square (Art. 62) of any polynomial can be
determined now.
Let the polynomial be a + b + c + d + ... , where the symbol
+ .. means that it is convenient not to specify the exact
number of terms. The square of this expression is obtained
bymultiplying it by itself.
Let the two factors be written down in lines one below
the other, thus
a + b + c + d + ...
a + b + c + d + ...

a² + ab + ac + ad + ...
ab + b2 + bc + ...

Σ(α²) + 2 Σ (αδ.)
Then the product is obtained (Art. 106) by multiplying
each term in the upper row by every term in the lower row
and adding all the results.
The two terms multiplied together must either be like or
unlike (Art. 85). The product of two like terms gives
aa, bb, cc, ... or a², b2, c², ... The sum of all these may be
denoted by the convenient symbol ∑ (a²), where ∑ may be
70 Elementary Operations. [120.
taken as an abbreviation for the words ' the sum of all such
quantities as.'
The product of two unlike terms gives a term of which
ab is the type. This particular term will occur twice,
namely as the product of the a in the lower row into the b
of the upper row, and the b in the lower row into the a of
the upper row.
Hence the sum of the products of unlike terms may be
written as 22 (ab), where has the same meaning as before.
Hence on the whole
(a + b + c + d + ...)² = ∑Σ (a²) + 22 (ab).
120. As an example, suppose it be required to find the
value of
( 1 + x + x2 + x3 + x4)²,
Σ (α²) = 12+
12 (x)²)2 + (x2)² + (x3)2+ (x4)2
= 1 + x2 + x + x + x8 by (24) of Art. 72 .
The value of 22 (ab) is most easily obtained with accuracy
by taking first the product of the first term into all that
succeed it, then the product of the second term into all that
follow it, and so on. Thus
22(ab) = 2x+ 2x² + 2x² + 2x*
+ 2x² + 2x² + 2x5
+ 2x² + 2x
+ 2x7
and since 2 (a²) = 1 + x² + x + x +x
it follows that
(1 + x + x² + x² + x*) ² = 1 + 2x + 3x² + 4x + 5x + 4x + 3x + 2x² + x* .
121. A similar formula holds for the cube (Art. 62) of any
polynomial. Let the three factors to be multiplied together
be written in separate horizontal lines one below the other :
a + b + c + d + ...
a + b + c + d + ...
a + b + c + d + ...
121.] Multiplication. 71

Then, by an extension of the rules of Articles 104, 106,

the final product must be obtained by multiplying each
term in the third row by every product of one term in the
second and one term in the first row. Three cases may
arise :
(1) The three terms multiplied together may be all
alike. This will give aaa, bbb, ccc, ... or a³, b , c , ... ; and the
sum of all these terms may be denoted, as in Art. 119, by
Σ(α³).
(2) Two of the terms taken may be alike and the
third different. This will give a term of the type a²b.
This particular term may arise in three ways, namely by
taking the unlike factor b out of the first row and the two
like ones out of the second and third, or the factor 6 out of
the second row and the two factors a out of the first and
third ; or finally the two like factors out of the first and
second, and the factor 6 out of the third row. Thus the
sum of all this class of terms will be denoted by 32(a²b).
(3) All three terms taken may be unlike. The type
of such a term is abe. This particular term abc occurs in
six ways. For if a be taken out of the first row, 6 may
be taken out of the second and c out of the third, or bout
of the third and c out of the second. If a be taken out of
the second row, there are similarly two ways of taking b
and c, and two more if a be taken out of the third. Hence
the sum of all this class of terms is 6 2 (abc).
Thus on the whole

(a + b + c + d + ...)³ = Σ (a³) + 32 (a²b) + 62 (abc).

122. As particular examples
(1) (a + b)³ = (a³ + b³) + ( 3a²b + 3b2a)
= a³ + 3a2b + 3ab² + b³,
arranging the terms in descending powers of a.
72 Elementary Operations. [123.
In this example no terms exist of the type abc, since
the expression only contains two different terms.
(2) (a + b + c)³ = a³ + b³ + c3
+ 3a2b + 3a²c + 3b2a + 3b2c + 3c2a + 3c2b
+6abc.
(3) (1 + x + x²)³ = 1 + x + (x2)3
+ 3x + 3x²
+ 3x² + 3(x²)x²
+3(x2)² + 3(x2)2 20
+ 6x3
=1 + 7x3 +26
+ 3x + 3x²
+ 3x² + 3x+
+ 3x² + 3x5
= 1 + 3x + 6x² + 7x³ + 6x4 + 3x5 + x6 .
123. In applying the formulæ of Art. 119 and 121,
the rule of signs in multiplication has to be carefully
remembered. Thus in finding the square of (a + b + c) by
Art. 119, the square of -c being the product of two
negative factors will be positive (Art. 58) while the product
of + a and -c will be negative (Art. 56, 57). Thus
(a + b −c)² = a² + b² + c² + 2ab - 2ас - 2bc.
124. The cube of (a + b −c) can be deduced from the for-
mula of Art. 121 with the same precautions. Thus
(a + b − c)3 = { a + b + ( - c)} 3
=a³ + b³ + (-c)3
+ 3a2b + 3a² ( c) + 3b2a + 3b2( -c) + 3( -c)2a + 3( c)2
+ 6al(-c)
Now ( c)² = ( c) x ( c) = + c² (Art. 58)
a² ( c) = -a²c
(Art. 56, 57)
( c)3 = ( - )( - )( -c) = (- c)(c²)
by (10) of Art. 102
= -3
Art. 56, 57.
128.] Multiplication. 73

Hence (a + b- c)³ = a³ + b² - c³ + 3ab - 3a²c

+ 3ab² - 362c + 3ac² + 3bc²- 6 abc.
125. It has been seen that the product of two negative
factors is positive. The product of three negative factors,
since it may be obtained by multiplying the third factor
into the product of the first two, is therefore negative.
The product of four negative factors, since it is obtained
by multiplying the fourth factor, a negative quantity, into
the product of the first three, also a negative quantity, is
positive. Proceeding in this way, it is easily seen that the
product of an even number of negative factors is positive,
while the product of an odd number of negative factors is
negative.
126. Any even number is a multiple of 2, and can con-
sequently be expressed in the form 2n where n is a positive
integer. Any odd number is obtained by adding unity to
some even number, and can consequently be represented by
the general form 2n + 1. Hence as particular cases of the
last article we have
(-1 )2 = +1
(-1)2 + 1 = -1
since the numerical value of the product of the 2n or 2n + 1
factors, each of which is negative unity, is unity.
127. The symbol - a can be written ( -1)a
Hence ( -a)" = {( -1)a } "
= ( -1)"a", by(25) ofArt. 73 or 102 .
If n be even, the value of this is + a", and if n be odd it
becomes -а " .
128. The multiplier ( -1)" is a convenient symbol to
attach to a term which depends on n in such a manner that
the term is positive when n is even and negative when n is
odd. Examples of such terms will frequently occur in the
expansions given by the Binomial Theorem.
74 [129.
Elementary Operations.
129. It frequently happens in algebraical investigations
that there is one letter of more importance than the rest,
which may be called the dominant letter of the expressions
involved. When this is the case, it is usually best to collect
all the terms which involve the same power of that letter
in each expression, and by means of (11)and (12), Art. 102,
to arrange them as one term. Thus let it be required to
multiply a + a by x + b, a being the dominant letter.
Arranging the process after the manner of Art. 108, 109,
it is seen that in the two lines of multiplication there are
two terms involving & to the first power, namely ax and br.
x+a

x+b

x² + ах
+ bx + ab

x² + (a + b)x + ab.
Instead of writing these terms down separately in the
result as ax + bx, they are more conveniently written in
what, by (11), is the equivalentform, (a + b)x ; and the whole
product is written as x² + (a + b) x + ab.
In this form, by an extension of the definition of Article
84, the quantity (a + b) is termed the coefficient of x in the
product.
130. The result of the last article can be written

(x + a) (x + b) = x² + (a + b) x + ab. (a)
This being true whatever values may be given to a and
b is valid if -6 be written throughout for b ; we thus get,
since x + ( - b ) =x - b,
(x + a)(x - b) = x² + (a -b)x - ab, (β)
since a( -b ) = - ab .
Again in (a) replace a by -a as well as 6 by -6 ; the
132.] Multiplication. 75

formula then becomes

(x -a)(x-b) =x²- (a + b)x + ab, (y)
since (-a - b) = (a + b) by (6) of Art. 53
and (-a)( - b) = ab. Art. 58
The student who feels doubtful of the logical force of the
deduction of (s) and (y) from (a) should work out the multi-
plications involved for himself, and compare the process with
that given in the last article. He will see that each step
in his work could be deduced from the corresponding one in
the investigation of(a) by making the change of b into - b
for (3), and a and b into -a and -b respectively for (y).
131. Taking the product of the two factors (x + a) (x + b)
and multiplying it by a third factor a + c, it is seen that in
the two lines of multiplication there are two terms in-
volving a², namely (a + b) x² and cr².
x² + (a + b) x + ab
x + C

x² + (a + b) x² + abx
+ cx² + (ca + cb)x + abc
x² + (a + b + c)x² + (ab + ca + cb)x + abc.
The sum of these, or (a + b)x² + cx² by (11) can be written
(a + b + c)x². Similarly there are two terms involving 2,
namely abæ and c(a + b)x, which by (11) has been written
(ca + cb)x. The sum of these, or abx + (ca + cb)x, again by
(11) is equal to (ab + ca + cb)x. Hence the whole product
is x² + (a + b + c) x² + (ab + ac + bc)x + abc. The quantities
(a + b + c) and (ab + ac + bc) are, as in the last article, called
the coefficients of a² and a respectively in the product.
132. The result of the last article,
(x + a)(x + 1) (x + c) = x² + (a + b + c)x + (ab + ac + bc)x + abc,
is one of considerable importance, to which reference will
be made in a later part of the book.
76 Elementary Operations. [133.
133. One or two more examples of multiplication may
begiven
(a) (x² -ху + y²) x (x + y) = x³ + y3
(β) (x² + xy + y²) x (x −y) = x³ - у³.
The working of these is given below
(a) x² -xy + y²
x + y

x³-x²y +xy²
+ х²у-ху² +у³
+y³. (a)
(β) x² + xy + y²
X -

x³ + x²y + xy²
-х²у-ху²-у³

-y³. (β)
These results will be seen in the next chapter to be par-
ticular cases of general formulæ. It may be noticed that
the second, (β), can be deduced from the first, (a), by sub-
stituting -y for y.
134. Another important example of multiplication is
afforded by the product of x² + y² + 22-yz-zx-xy and
x+y + z.
It is very easy to work this by the ordinary method of
Art. 108, 109. The student can do this for himself.
Here it will be taken as an example of the method of
Art. 129, taking a as the dominant letter. The two
expressions arranged in descending powers of a are repre-
sented as below, where in each expression the terms
involving any power of a are placed in a bracket and are to
be treated as one term.
136.] Multiplication . 77

x²-(y+ z) x + (y -yz + z2)

x + (y+ z)
x³-( y + z) x² + (y²- yz +22) x
+ ( y + z) x²-(y² + 2yz + z²) x + (y3 + 23)
23 -3yzæ + (y³ + z³).
The student will have no difficulty in following the first
line of multiplication, which is obtained by multiplying
the multiplicand by a. The second line is the product of
(y+ ),treated as one quantity, into the multiplicand.
The first term in the latter gives evidently (y + z)x² . The
second term gives -( y + z) (y + z) x or -(y + z)2x, which
by Art. 114 is - (y² + 2yz + 22)x. The third term gives
(y + z) ( y -yz + z²) or, by (a) of Art. 133, y³ +23.
In adding, the term involving a² evidently goes out ;
the coefficient of a is the algebraical sum (Art. 93) of
y²-yz + z² and -(y² + 2yz + z²) which is equal to
y²-yz + z2-(y² + 2yz + z²) or y²-yz + z2-y2-2 yz- 22
or finally to 3yz. Hence the whole product becomes
what is placed below the line.
135. The formula obtained in the last article may be
written symmetrically thus
(x² +y² + z2-yz-zx - xy) (x + y + z) = x² + y² + 23 - 3 xyz,
a result which is frequently useful.
136. It is often required when multiplying together two
expressions involving the same dominant letter, not to dis-
cover the whole of the product, but to ascertain the coeffi-
cient of some particular power of the dominant letter.
For instance, in the product indicated by the symbols
(1 + 2x + 3x² + 4x³ + 5x+) ( 1-2x + 3x2 - 4x3 + 5 x4),
let it be required to find the coefficient of 5 .
To ascertain this, it is necessary to consider what terms
in the first expression multiplied by what terms in the
second give a product involving 25. These are clearly the
78 Elementary Operations. [137-
second term in the first, multipliedinto the last in the second;
the third in the first multiplied into the last but one in the
second, and so on. The algebraical sum of all these pro-
ducts is
2 x x 5 x² + 3x² × ( -4x3) + 4x3 × ( 3x²) + 5x^( - 2x)
or
10x5-12x5 + 12x5-105, which is equal to zero.
Hence there is no term involving as in the product.
137. It is evident that the highest power of x in the
product is the product of the highest powers in the two
factors, or is x x x², or x8 .
Hence if by a method similar to that of the last
Article the coefficients of successive powers of a beginning
with aº or 1 and going up to 2º be found, the whole product
will be ascertained
Thus the term independent of x is 1 x 1 , or 1 .
The term involving a is 1x ( -2x) + 2x ( 1 ), or (0).
The term involving æ², is
1 × 3x² + 2x ( -2x) + 3x² × (1 ), or 2 x².
The term involving ³, is
1 × ( -4x³) + 2xx (3x2) + 3x² × ( -2x) + 4x3 x (1 ), or 0.
x

The term involving æ , is

1x ( 5x ) + 2xx ( -4x³) + 3x² × ( 32)
+ 4x3 x
× ( -2x) + 5x* x (1 ), or 3x4.
The term involving æ , is
2x(5x*) + 3x²( -4x3) + 4x³(3x2) + 5x*( - 2x), or 0.
The term involving æ , is
3x² . (5x ) + 4x³ . ( -4x³) + 5x4 . (3x2), or 14x6.
The term involving 7, is 4x³ . (5x4) + 5x4 . ( -4x³), or 0.
The term involving a³, is 5x4 . (5x4), or 25x8.
Hence we finally obtain
(1 + 2x + 3x² + 4x3 + 5x4)(1-2x + 3x2 - 4x3 + 5x4)
= 1 + 2x² + 3x² + 14x6 + 25x8,
a result which the student can verify by going through
the process of multiplication in the ordinary way.
 Multiplication. 79

EXAMPLES .

1. Multiply 2x by 5x²; 4ab2 by 3a²b ; 6abc² by 5abc.

2. Multiply 4a³ - 3a²b + 2ab² -b² by 5a²b².
3. Multiply 3x² -6xy + 2y² by -7xy.
4. Multiply 2x+ 3y by 3x+ 2y.
5. Multiply 7x- 4y by 4x- 7y.
6. Multiply 4x² - 6xy + 9y² by 2x+ 3y.
7. Multiply 9a² + 6ab + 462 by 3a-2b.
8. Multiply x² - 3x+ 2 by х- 3.
9. Multiply x² -6x² + 11x- 6 by x- 4.
10. Multiply x²- 3x+ 2 by x² + 3x+ 2.
11. Multiply x² -2x+ 2 by x² + 3x- 2.
12. Multiply x+ y + z by x-у-2.
13. Multiply x² - y² - 2yz-22 by x + y- z, and the product
by x-y+z.
14. Multiply 4x³ - 3x²y + 2xy² -y³ by x² - 2x²y + 3xy² - 4y² .
15. Multiply x²-5x²- 3x- 18 by x² + 3x- 4.
16. Simplify (x-a) (x-b) (x- c)
- {bc(x - a) - [(a + b + c)x - a (b + c)]x } .
17. Simplify a² + b² + c²-(a -b + c) (a + b −c)
-(b - c + a) (b + c-a) -(c - a + b) (c + a − b).
18. Multiply x² -7x² + 5x+ 1 by 2x²- 4x+ 1.
19. Simplify (x- 2) (x-3)- (x- 7) (x- 1) + (x- 1) (x - 2).
20. Prove that (a² + ab + b²)²-(a² + ab - b²)² = 4ab²(a + b).
2

21. Prove that (a² + ab -b²)²- (a² - ab - b²)² = 4ab (a² -b²).
22. Show that (1 + xz)2 (1 + yz)² - { (1 - xz) ( 1 - yz) + 2xyz }2
= 4 (х + у - ху) (xyz³ + xyz² + z).
23. Prove that (a + b + c)² + a² + b² + c²
= (a + b)² + (b + c)² + (c + a)².
24. Show that (x + a)* = x² + 4x²a + 6x²a² + 4xa³ + a*.
25. Multiply (x−y)² + (y + 2)² + (z − x)² by x + y + z.
80 Elementary Operations. Multiplication.
2

26. Prove that (x² -yz)² -(y −zx) (z² -xy)

= x(x + y + z) (x² + y² + z² - yz-zx- xy).
27. Prove that (a-b) (b -c) (c-a)
= a² (c- b) + b² (a - c) + c² (b - a).
28. Prove that (a + b) (b + c) (c +a)
= bc(b + c) + ca(c + a) + ab (a + b) + 2abc.
29. Simplify (x + y)² + 4(x + y)³ (х - у)
+6(x + y)² (x− y)² + 4(x + y) (x − y)³ + (x − y)*.
 CHAPTER V.

INTEGRAL FORMS. ELEMENTARY OPERATIONS, DIVISION.

138. In the last chapter a method has been given for

ascertaining the product of any two algebraical expressions,
whether consisting of one or many terms.
The problem of finding the product when the factors are
given has been completely solved. The inverse problem
is- Given the product and one factor, to find the remain-
ing factor,' and the solution of this problem is effected
by the methods of Division. Those methods solve a
rather more general problem ; namely, ' Given two
algebraical expressions involving the same dominant letter,
which occurs in one of them to a higher degree (Art. 86)
than in the other, to ascertain whether the second is a
factor of the first, and if not, what is the expression of
lowest degree by which the first can be increased or
diminished so that the second may be a factor of it.'
139. It may be useful to define the degree of an expres-
sion as the index of highest power of the dominant letter
which occurs in it. Thus the expressions æ³ + 2x² - 3 and
x - 1 are of the third and fourth degree in a respectively.
The expression a³ + a2 b + al² is of the third degree, if a be
regarded as the dominant letter, but only of the second
when b is taken in that capacity.
Thus the degree of an expression may depend on the
letter which is chosen as dominant, and may therefore
differ in the case of a homogeneous expression from the
G
82 Integral Forms. [140.
number of dimensions (Art. 88). Thus a³ + a²b + ab² is a
homogeneous expression of three dimensions, while, if 6 be
the dominant letter, it is only of the second degree.
In the investigations in this and the next chapter the
degree of the expressions will be chiefly considered with
regard to some particular letter taken as dominant.
140. It has been seen (Arts. 28, 46, 47) that, in the
case of numbers whether integers or fractions, the result of
division is always such a quantity as, when multipliedby
the divisor, gives back the dividend..
This may be taken as a definition of the process of
division in algebra. Thus the result of division in any
case can only be ascertained by remembering or proving a
corresponding result of multiplication.
141. In Art. 68 it has been shown that if a higher
power of any letter be divided by a lower power the result
is obtained by subtracting the index of the lower power
from that of the higher. Thus
aa2 = 6-2 = a4 .

Obviously this is a consequence of the corresponding

fact in multiplication that
a² . a² = a² + 4 = a6 .

142. A single term as a³be involving powers of more

than one letter can be exactly divided by a single term
involving the same letters, or some of them, provided
that the power of each letter in the divisor is lower than
that of the same letter in the dividend. The result of
division is obtained by subtracting the index of each letter
in the divisor from its index in the dividend. Thus
a3bca²b = abc.

The proof of this result lies in the fact thatby Art.

103
a²b . abc = a²bc.
145.] Elementary Operations. Division. 83

143. If the single terms one or both contain numerical

factors, or coefficients (Art. 84), the coefficient of the
dividend must be divided by that of the divisor and the
quotient taken as coefficient of the result. Thus
15a32c3a²b = 5 abc,
this equation again depending on the corresponding fact
in multiplication that (Art. 103)
3a²b.5abc = 15a3b2c.

144. If the single term of the divisor contain any letter

to a higher degree than it occurs in the dividend, the
result of division can be written in one of two ways. First,
in a fractional form : thus
a3b2 c
a3bcabc2 = by (17) of Art. 53
ab³ c²
ab2c.a²
=

abc.bc by (9) and (10) of Art. 53

a2
bc (a) by (13) of Art. 53 .
==
Secondly by the use of negative indices and by taking the
formula (23) of Art. (68) as perfectly general.
Then abcabc² = a² 2-3 1-2
= a²b-1c-1. (β)
With the interpretations given to a negative index (Arts.
69, 70), the two results (a) and (3) are evidently identical.
145. If the dividend be a polynomial and the divisor a
single term the division can be effected by dividing each
term of the polynomial by the divisor, and placing the
several quotients together with their proper signs. Thus
the result of the division of
6a3b2c- 3a²bc + 12a²bc³ by 3abc is 2a²b - ab² + 4ac2.
The proof again lies in the fact that this last expression
multiplied by 3abe gives back the dividend.
G2
84 Integral Forms. [146.
This process is conveniently conducted by writing the
divisor to the left of the dividend with a line to separate
them, and a line being drawn below the dividend, writing
the successive terms of the result below the terms of the
dividend from which they arise : as illustrated below in the
example already worked.
3abc ) 6a3b2c- 3a²b³c + 12 a²bc3
2 a²b -

ab² + 4 ac²

146. In the more difficult problem of the division of one

polynomial by another, the same principles must be used.
The result of division must be an expression such that
when multiplied by the divisor it will give the dividend.
If the two expressions do not involve the same dominant
letter, the division can not be effected as an algebraical
operation. It may of course be effected as an arithmetical
process, provided special values be given to the letters, but
the result can only be exhibited in algebraical symbols by
a fractional form of which the dividend is the numerator
and the divisor the denominator. See (17) ofArt. 53.
It is farther obvious that the division cannot be
algebraically effected unless the dominant letter occur in
the dividend to at least as high a degree as in the divisor,
and that if it occur to the same degree in both, the result
of division must be a number, or an expression not con-
taining the dominant letter.
147. A farther difficulty requiring removal is found in
the fact that, as has been seen in Art. 136, the term in-
volving any particular power of the dominant letter in the
product of two factors is obtained by adding together the
products of several pairs of terms in the factors.
The one exception is the case of the term involving the
highest power of the dominant letter, which, we have seen,
is formed by the product of the highest terms in the
 150.] Elementary Operations. Division. 85

two factors (Art. 110). A similar statement applies to the

product of the two lowest terms.
By the phrases ' lowest ' and ' highest ' terms will in
future be meant the terms involving the ' lowest ' and
' highest ' powers respectively of the dominant letter.
148. It becomes thus of fundamental importance to the
process of division to arrange the dividend and divisor both
in descending, or both in ascending, powers of their common
dominant letter. This arrangement was found to be
convenient in multiplication. It is practically indispen-
sable in division.
149. Let the dividend be 2x² + 7x + 3 and the divisor
2x + 1 . The highest term in the other factor required
must (Art. 147) be such as, when multiplied by 2x, the
highest term in the divisor, will give 2x², the highest
term in the dividend ; that is, it must be x. To ascertain
whether this is the whole of the required factor, the divisor
is multiplied by : the product is 2x² + x which is not
equal to the whole of the dividend. Hence something has
to be added to the term æ in order to give the full result
of division. This ' something ' must be such as, when
multiplied by the divisor, will give the rest of the dividend.
Hence the next step is to subtract 2x² + x from the dividend.
This leaves a remainder 6x + 3 .
The next term in the required result must therefore be
such as when multiplied by 2x shall give 6x, that is, it
must be 3. Multiplying the whole of the divisor by 3, we
get 6x + 3, which is equal to the rest of the dividend.
Hence it has been ascertained that +3 when multiplied
by 2x+ 1 gives 2x² + 7x + 3, whence we may write
(2x² + 7x + 3)÷ (2x + 1 ) = x + 3 .
150. The actual working of the above process is usually
written more concisely as follows.
86 Integral Forms. [151.

2x + 1 ) 2x² + 7x + 3 ( x + 3
2x² +

6x+ 3
6x+ 3

The dividend and divisor are arranged in descending

powers of their common dominant letter and the divisor
written to the left of the dividend with a line separating
them. A similar line is drawn to the right.
The first term in the required result must, as before
explained, be x. This is written to the right of the right-
hand curved line. The divisor is multiplied by this term
and the product written down under the dividend, the same
powers of the dominant letter being for convenience written
in a vertical line. This product is subtracted from the
dividend. The remainder now becomes the dividend, for
the problem now to be solved is to find what has to be used
to multiply the divisor in order to produce this remainder.
151. In the last article the student will see that the
product of æ into 2x+ 1 has been first subtracted from
the dividend and afterwards that of 3 into 2x + 1 . On
the whole the product of x + 3 into 2x + 1 has been
subtracted and there is no remainder. The process has
therefore not only ascertained what the factor must be
which, when multiplied by 2x + 1, shall give the dividend,
supposing there to be such a factor, but has also shown
that +3 satisfies completely the required condition.
In this case 2 + 3 is called by analogy with numerical
results (Art. 28) the quotient of 2x² + 7x + 3 divided by
2x + 1 .

152. Suppose that it had been required to divide

2x² + 7x + 5 by 2x + 1. The process up to the last stage is
identical with the previous one.
153.] Elementary Operations. Division. 87

2x + 1 ) 2x² + 7x + 5 ( x + 3
2x2 + x

6x+ 5
6x+ 3
2

It is still obvious that if any factor can be found which,

when multiplied by 2x + 1, shall give the dividend, the
terms of that factor must be in succession & and 3. On
subtracting the products of these terms into the divisor in
succession from the dividend it is found that there is a
number 2 still left over. It is obvious that no expression
of integral form when multiplied by 2x+ 1 can give
2
the number 2. A fractional form would do so
2x + 1

(Art. 53 (18)), and the result of division might therefore be

2
written as a +3+
2x + 1

It is usual, however, to speak of 2 as the remainder after

the division has been effected, while the part of the result
x + 3, which assumes an integral form, is called the quotient.
153. As another instance let it be required to divide
3x² - x² + 2x² + x + 1 by x² -x + 1 .
Arranging the divisor and dividend as in Art. 150, the
process is represented below
x² -x + 1 ) 3x4 - x² + 2x² + x + 1 ( 3x² + 2x + 1
3x - 3x3 + 3 x²
2- x² + x + 1
2x² - 2x² + 2x
2 -x + 1

x2 x+ 1

The highest term in the quotient, as it may now be called,

88 Integral Forms. [154-
must be such as when multiplied by 22 shall give 3x4, that
is, it must be 3x². The product of the whole divisor by
this term is written below the dividend and subtracted
from it in accordance with the rules of Arts. 95, 96. The
remainder is written below, and it now becomes obvious
that the highest term in the rest of the quotient must
be such as when multiplied by æ² shall give 2æ³, that is,
it must be 2x. The whole of the divisor is multiplied
by this term, written below the first remainder and
subtracted from the latter. There is still left x² - x + 1 of
the dividend, and it is obvious that the only remaining
term required is 1. The divisor being multiplied by 1 and
the product written down below and subtracted from the
second remainder there is nothing left.
154. As in the previous example it may be noticed that
the products of 3x², 2x, and 1 into the divisor have been in
succession subtracted from the dividend: that is, on the
whole the product of ( 3x² + 2x + 1) into the divisor has
been subtracted. There being no remainder the process
has proved that
(3x² + 2x + 1) . (x²-x + 1 ) = 3x4 - x3 + 2x2 + x + 1 ,
or that (3x - 23 + 2x2 + x + 1) ÷(x² -x + 1 ) = 3x² + 2x + 1 .
155. As another example let it be required to divide
3x -x³ + 2x² + 5x- 2 by 3x² + 2x + 1 .
3x² + 2 x + 1 ) 3 x - x + 2x² + 5x - 2 ( x² -x + 1
3x² + 2x² + x²

-3x + x² + 5x
-323-2x2 - x

3x² + 6x- 2
3x² + 2x + 1
4x- 3

The first term in the quotient is evidently 2². The

157.] Elementary Operations. Division. 89

remainder after the product of ² into the whole divisor

has been subtracted is - 3x3 + x² + 5x - 2. The last term
of these has not been written down in the lower line as
experience has now taught us that it will not be affected
by the next stage of the working.
The next term in the quotient must be such a term as
when multiplied by 3x² shall give -3x³. It must there-
fore be -x (Arts. 56, 57), and in multiplying the divisor
by this term the rule of signs must be attended to.
The third term in the quotient must evidently be 1. When
the product of the divisor by this term is finally subtracted
there remains over 42-3. Evidently no integral form
when multiplied by the dividend can produce 4x- 3, and
either the result of division must, as before, be written in
4x- 3
the form x²-x + 1 +
3x² + 2x + 1 , or it may be said that
the quotient is x²-x + 1 and that there is a remainder
4x- 3.

156. The process of the last Article has shown that if

x²-x + 1 be multiplied into 3x² + 2x + 1 and the product
subtracted from the dividend, there will be a remainder
4x- 3 .

It follows that if 4x- 3 be added to the product of

x²-x + 1 into 3x² + 2x + 1 the result will be the dividend.
157. Let any expression be divided by another so as to
give a quotient and a remainder. Then if the quotient be
multiplied into the divisor and the remainder be added to
the product the result will be the dividend.
In symbols, if A denote the dividend, B the divisor,
Qthe quotient, and R the remainder
BQ + R = A,
where, as in Art. 67, the symbol = is an abbreviation
for the words, ' is identical with' or ' is the same thing as.'
90 Integral Forms. [158.
158. The remainder' in any division will always be of
lower degree in the dominant letter than the divisor. For
the process of finding successive terms in the quotient can
always be carried on until the highest term in the re-
mainder is of lower degree than the highest term in the
divisor. If the coefficient of the highest term in the
divisor be not unity it may happen that the coefficients of
some of the terms in the quotient will be fractional.
159. If the divisor and dividend involve more than one
letter it may be a matter of indifference which of them is
taken as the dominant letter. Whichever is selected at
the beginning must be retained throughout the operation.
160. As an example, let it be required to divide
x³ +y³ + z³- 3xyz by x +y + z.
Here, the expressions being perfectly symmetrical with
respect to x, y, and z, it can make no difference algebraic-
ally which of the three is taken as the dominant letter.
From alphabetical habit it is usual to take x. Hence the
dividend must be written in the order below ;
x + y + z ) x³- 3yzx + y³ + z³ ( x² -xy -xz + y² -yz + z2
x³ + x²y + x2 z
-x²y-x²z- 3 xyz
-х²у-ху² - xyz
-x²z + xy²-2xyz
-x2z ryz-xz2
+xy². -

xyz + xz² + y + z³
+ху² + y³ +y²z
-

xyz + xz²-y2z + 23
-

xyz -y2z-yz2
+ xz2 +yz² + z3
+ xz2 + yz² + 23
 162.] Elementary Operations. Division. 91

The process is sufficiently indicated. After the first

subtraction has been effected there are two terms involving
22 in the remainder. It is a matter of indifference which
of these is taken as the leading term, inasmuch as by
Art. 110 there obviously must be a term in the remaining
part of the quotient which, when multiplied by æ, shall
give each of these two. Similarly, after the third step of
subtraction, there are three terms which involve a to the
first power, and these terms may be taken in succession in
any order.
161. The only point of importance to notice is that every
term in the successive remainders which contains a higher
power of the dominant letter must be taken into account
before proceeding to any terms of a lower degree.
162. The example of Art. 160 is, however, more ele-
gantly worked by a process analogous to that given for
multiplication in Arts. 129, 134. The dividend and divisor
will be written, and the process of division conducted as
below.

x + (y + z) ) x³ - 3yz . x + (y + z³) ( x²-( y + z)x + (y -yz +z2)

x³ + ( y + z) x²
-( y + z) x² - 3yz.x
-( y + z) x² - ( y² + 2yz + z²) x
(y - yz + z2) x + (y³ +23)
( y2 - yz + z2) x + ( y³ +23)
Here, in multiplying the divisor by the successive terms
in the quotient, (y+ z) is treated as one term. Thus after
each subtraction all the terms involving the same power
of a, which in Art. 161 appeared as distinct terms, are
now merged into one. In the second subtraction the
coefficient of the term in the remainder involving a is
obtained by changing the sign of the term to be subtracted
92 Integral Forms. [163.
(Arts. 95, 96) and adding it algebraically to the term in
the upper line. The resulting coefficient of x is thus seen
tobe
y² + 2yz + z2 - 3yz or y² -yz + 22.
The results of multiplying ( y+ z) the second term in
the divisor by ( y + z) and again by ( y -yz + z²) the suc-
cessive coefficients of terms in the quotient are obtained
from Arts . 114 and 133.
163. The discovery of the factors (Arts. 22, 106) of
algebraical expressions is a matter of great importance.
The following theorem is often useful in helping to the
discovery of such factors. If any rational integral algebraical
expression of which the dominant letter is x assume the
value zero (Art. 51) when any particular value a is given
to æ, then x-a is a factor of the given expression.
The proof follows in the next three Articles.
164. Let the given expression be denoted by
Poan + P₁xn-1 +P2 xn−2 + ... + P n−2X2 + P n-1x + P n (a)
where the suffixes to the coefficients of the different powers
of a do not denote any relation between the magnitudes
of Po , P1 , P2 , ... like the indices of Art. 62, but merely the
position which each occupies in the series. Thus py means
the seventh coefficient not reckoning the first, and is the
coefficient of an-7. The number which denotes the co-
efficient increases by unity from term to term, that which
denotes the power of a, or in other words the index of x
(Art. 62) decreases by unity from term to term. The sum
of these two numbers is therefore always the same and
equal to n the degree of the expression (Art. 139). It is
thus clear that the coefficient of x² will be pn-2 , and so on.
The dots which separate the two + signs denote the terms
omitted between those which are written down, the number
of which omitted terms, as it depends on n, must be left
undetermined if the proposition is to be perfectly general.
167.] Elementary Operations. Division. 93

165. The expression (a) is the most general form of a

rational integral expression of the nth degree in a. It will
be sometimes conveniently denoted by the symbol f(x),
where f is an abbreviation for the words ' a function of,'
the term ' function ' being used in mathematics to express
any quantity which depends on another in a definite
manner.

166. Let, then, the expression (a) be divided by x-a with

remainder R and quotient Q. By Art. 158, R is of lower
degree than x- a in x ; that is, R cannot contain & at all,
and is in fact a definite function of a and the coefficients of
the expression (a). This will be obvious to those who have
followed the examples in division previously given. Then,
byArt. 157,
Q (x-a) + R = pox™ +p₁x -1 ... + Pn−1x + Pm•
This identity exists quite independently of the particular
value of æ, and will be true for all values. Let the value
a be given to x. R which is independent of a remains
unchanged by this substitution, Q will assume some special
value, and x- a will assume the value zero. Consequently
(Art. 51) whatever may be the particular value of Q, the
product Q (x -a) certainly is zero. Hence it follows that
R = poa" + p₁a"-1 + ... +Pn-1a + Pn ;
that is, the remainder after dividing f(x) by x-a is the
value which f(x) assumes when for a the value a is
substituted.
167. Hence iff(x) be zero when a is put for a, the
remainder after dividingf(x) by x-a is zero, or in other
words f(x) is divisible by x-a without remainder, that is
z-a is a factor off(x) (Arts. 138, 151, 154).
Conversely if x- a be a factor of f(x), f(x) must vanish
when the value a is put for a ; for in this case f(x) is
divisible by x-a without remainder, and consequently R,
orf(a), as we may write it, is zero.
94 Integral Forms. [168.
This second result follows also from Art. 51 since the
factor æ-a has the value zero when a is put for æ, and
therefore the product must also be zero.
168. The expressions 2" - a" and " + a" are of frequent
occurrence. Suppose a to be a positive integer, whatever
the value of n may be, x" -a" vanishes when a is put for x.
Hence for all integral values of n, r" -a" has a factor x- a.
If n be even and -a be substituted for æ, х -а"
becomes ( -a) " -a", or by Art. 127, since n is even, a"-a" ;
that is, zero.
Hence, when n is even x" -a" has also a factor x-( -a)
or x + a.
In the expression 2" +a" let -a be put for x. The
value becomes ( -a) " + a", which is 2a" if n be even and 0
if n be odd (Art. 127). Hence, if n be odd " + a" has
a factor x + a.
The results of Articles 116 and 133 are particular cases
of the theorems of this Article .
169. As another instance, take the frequently occurring
expression a²b - ab² + ac² -a²c + b²c-bc², (a)
which may be written in either of the equivalent forms
a² (b -c) + b² (c-a) + c²(ab), (β)
or bc (b -c) + ca(c -a) + ab(a −b), (7)
as the student can easily verify by multiplying out.
Regarding a as the dominant letter and putting for a
wherever it occurs in (a) the value b, the expression
becomes 63-63 + bc²-b²c + b²c-bc² or zero .

Hence a - bis a factor of (a).

Similarly by putting for a the value c, the expression
becomes be²- b²c + c³ -c³ + b2c-bc2 or zero .

Hence a - cis a factor of the expression (a).

170.] Elementary Operations. Division. 95

Regarding now b as the dominant letter and putting for

b the value c the expression (a) becomes
a²c-ac² + ac² - a²c + c² -c or zero .

Hence b-c is a factor of (a).

Now since each of the three factors (a -b), (a-c), (bc) is
(Art. 88) a homogeneous expression of one dimension, their
product (Art. 112) is a homogeneous expression of three
dimensions, that is of the same dimensions as (a).
Hence a can have no other factors involving the letters
a, b, c than the three we have found.
It may possibly contain some numerical factor independent
ofa, b, c. If this be denoted by k, it follows that
a²b - ab² + ac² - a²c + b²c -bc² = k ( a −b) (ac) (bc) (8)
This identity must be true whatever values be put for
a, b, c. If to a we give the value 2, to b the value 1 and
to e the value 0, the left hand side becomes 4-2 or 2,
while the right hand side becomes k × 1 × 2 × 1 or 2k.
Hence 2 = 2 kor k = 1 so that the identity (8) becomes
a²b - ab² + c² -a²c + b²c -bc² = (ab) (ac) (b−c) (ε)
The student can satisfy himself by substituting any
other values, as 1, 2, 3 in (8) for a, b, c, that he will always
obtain the same value of k.

170. The form of the quotient Q in Art. 166 deserves

consideration. It is clear from the examples of division
given in Arts. 152-155 that Q must be a rational integral
function of a degree lower by unity than the dividend.
Hence its general form must be
2n - 1
+ 91x - 2 + q2x -3 + ... + 9n-2x + qn-] (a)
where 90, 91, 92...9n-1 are coefficients at present unknown.
If, then, (a) be multiplied by x- a and R be added to the
product, the resulting expression must be the dividend.
The multiplication can be written in the usual form
 96 Integral Forms. [171 .
(Art. 108), and the two lines of the product added as in
Arts. 129, 131, 134.
Loxn-1 + q₁xn-2 +920-8 + ... + In-2x + In-1
х -а

-2
Lox™ + q₁xn-1 +q2x²+ ...
+ In-2x + qn-1X
2

-aq.xn- 1 -aqx - 2 ... -aqn -9x - aqn_2x - aqn -1

qox² + (q₁ - aq )x -1 + (q2 - aq₁ )xn-2 + ... + ( qn-2 - aqn -3)x²
[+ (9n-1 - aqn-2)x - aqn-1-
The resulting expression with R, which does not contain
a, added to it must be the original dividend or
Pox" +P₁xn- 1 + P2x 2 + ... + Pn-2x² +Pn-1X + Pn
Hence arise the series of identities
or
Lo Po Lo = Po
Ax 1.
11 -aqo = P₁ or adding aq, to both sides, 91 = aqo +2 {Art. (53)
92 -aq1 = P2 " aqı "
92 = aq₁ +P2,

In-2 -aqn-s = Pn-2 " aqn-3 ‫وو‬ In- 2aqn-3 + P -22

In-1- aqn-2 = Pn-1 " aqn
-2 " In-1 = aqn2 +Pn-1 )
R - aqn-1 = Pn ‫دو‬ aqn-1 ‫وو‬
R = aqn-1 + Pn
Hence the coefficients 91 , 92 , 93... can be formed in
succession by multiplying the previous one by a and adding
(algebraically) the corresponding coefficient in the dividend,
the first coefficient qo being the same as po.
171. As an example let it be required to find the quotient
and remainder when the dividend is
a² - 2x6 + 5x5 + 3x² - 6x³ + 7 x² - 11 x + 2,
and the divisor is x- 2. Writing the divisor to the left of
the dividend, as in Art. 150, the quotient can be written
below, the coefficients being calculated in succession by the
rules of the last Article.
x- 2) x²- 2x6 + 5x5 + 3x² -6x3 + 7x² - 11x + 2
x + 0x5 + 5 x² + 13x3 + 20 x² + 47x + 83
Remainder = 168 .
173.] Elementary Operations. Division. 97
Here the first coefficient is 1,
‫دو‬
second ‫وو‬ 2x 1-2 = 0,
"
third ‫دو‬ 2x 0 + 5 = 5 ,
‫دو‬
fourth 2x 5 + 3 = 13 ,
‫دو‬
fifth ‫دو‬ 2 × 13-6 = 20,
‫دو‬
sixth ‫دو‬ 2x20 + 7 = 47,
‫دو‬
seventh ‫دو‬ 2 x 47-11 = 83 ,
and the remainder is 2 x 83 + 2 , or 168.
The student can easily verify that 168 is the value
which the original expression assumes when a has the
value 2.
172. As a second example let it be required to divide
x5 +y5 by x + y. Here for a must be substituted -y, and
it must also be remembered that the dividend is not a
complete expression of the fifth degree, because the terms
involving 24, 2³, 22 and x are wanting. It can be put into
the form of a complete expression of the fifth degree by
writing the missing terms with zero coefficients. The
process of division will then be indicated thus :
x + y ) x5 + 0x²y + Oxy² +0 x²y³ + Oxy +y5
04- х³у + xy²- ху³ +у+
and the remainder is zero.
Here the coefficient of the highest power of a is 1,
that of the next highest is -yx1 + 0 or -y,
‫دو‬ is -y . ( -y) +0 or y2,
‫دو‬

‫وو‬ ‫دو‬ is -y.y² +0 or -y³,

‫دو‬ ‫دو‬ is -y -y³) +0 or y4,
and the remainder is -y.y + y5 or zero .

173. The process is one which with a very small amount

of practice can be very rapidly performed: incidentally it
often affords the most expeditious means of calculating the
value which a function of the supposed kind assumes, for
any given value of its dominant letter.
H
98 Elementary Operations. Division.

EXAMPLES .

1. Divide 16abc by 2a²b ; 25a²x²y² by 5a²x².

2. Divide 24ab2-18a²b by 6a²b².
3. Divide 8a³ - 6a²b + 10ab by -2a.
4. Divide 18a*x³ - 24a³x² + 30a2x² + 42ax® by 6ax².
5. Divide x² - 3ax + 2a² by x - a ; x² - 3ax + 2a² by x- 2a.
6. Divide x³ - 6ax² + 11a²x - 6a³ by x- a.
7. Divide x³ - 6ax² + 11a²x- 6a3 by x² - 5ax + 6a².
8. Divide x -y by x-y.
9. Divide x -y by x²-y².
10. Divide x² + x²y + xy² + x²y + xy² + y by x+ y.
11. Divide xy² +
x² + x²y² +y by x² + xy+ y².
12. Divide x² + x²y² + y² by x² -xy+ y² .
13. Divide 27a³ -86 by 3a- 2b.
14. Divide x - 5x² + 4 by x³ - 3x+ 2.
15. Divide x² - 9x² + 12x- 4 by x² + 3x- 2 .
16. Divide x - 2x² - 22x² - 7x² - 42x + 72 by x² - 2x- 24 .
17. Divide x² -y² - 2yz-z² by x + y + z.
18. Divide 4x - 11 x²y + 20x²y² - 30x²y +20x²y
-11 xy² + 4y by 4x³ - 3x²y + 2xy²-у" .
19. Divide x² + x²y - x²z- xy²- 2 xyz - xz² -y -y²z + yz² + 23
by x+y- z.
20. Divide x + y + z² - 2x²y² - 2xz² - 2yz
by x² -y - 2yz-z².
21. Divide a² (b - c) + b² (c- a) + c² (a−b)
by a² -a (b + c) + bc.
22. Divide a² (b + c) + b² (c + a) + c² (a + b) + 2abc
by a² + a (b + c) + bc,
23. Divide (a³ - b³) x² -(a² + ab -ab² -b²) x + (a*b -ab*)
by (a-b) (x-a).
24. Divide (x+ y + z) (xy + yz + zx) -xyz by x + y.
 99
Elementary Operations. Division.
25. Divide 1 + y + z - 3yz by 1 +y + z.
26. Find the quotient and remainder when
x²- 6x + 4x5 + 3x² + 2x² + x² + 2x - 1
is divided by x- 1.
27. Find the quotient and remainder when
x - 5x + 4x - 3x² + 2
is divided (1) by x - 1, (2) by x²- 1 .
28. Find the quotient and remainder when
x² + 5x + 3x + 7x² + 9x + 4
is divided by x - 1. Deduce the remainder when 15030794
is divided by 9.
29. By the method of Art. 170 find the quotient when
x -5x -x² + 26x² - 2x+ 31 is divided by x² - 7x+ 10.
30. Divide ³ + 3ax² + 3a²x + a³ -b³ by x + a - b.
31. Resolve (x² + y² -22 -и²)² -(2xy- 2zu) into four factors.
32. Resolve 4(xu + yz)² - (x² -y - z² + u²)² into four factors.
33. If 2s = a + b + c,
prove that s (s - b) (8 - c) + s (s- c) (s -a)
+ s (s - a) (8-6) - (s -a) (s - b) (s -c) = abc.
+8

34. Prove that

(a + b + c)+ (b + c) (c + a) - (a + b)² + a² + b² + c²
= 36abc (a + b + c).

H2
 CHAPTER VI.

ELEMENTARY OPERATIONS . INTEGRAL FORMS . HIGHEST

COMMON FACTOR AND LOWEST COMMON MULTIPLE .

174. If it were always possible to discover all the factors

of any expression, it would be merely a matter of com-
parison to see what factors were common to any two
expressions.
Thus we know that x² - a² = (x-a) (x + a) (Art. 116),
and we also know that x - a³ = (x − a) (x² + ax + a²)
(Art. 133). It is obvious that the only factor common
to the two expressions x²-a² and x³ -a³ is x - а.
175. If there be more than one factor common to two
given expressions, it follows from the associative law in
multiplication that the product of all the common factors
will also be a common factor. This product will be of
higher dimensions (Arts. 86, 111) or of higher degree
(Art. 139) than any of the common factors which produce
it. Hence it has the highest degree of all the expressions
which are common factors of the two given ones. It is
therefore called the highest common factor.
176. Strictly speaking, if the two given expressions have
a common numerical factor, the introduction or omission of
this factor will not affect the degree or dimensions of the
product of the common factors. It is usual however to
include such numerical factors as part of the highest
common factor, and the usage has its advantages for some
practical purposes, although not without some counter-
balancing disadvantage in producing a confusion in the
178.] Elementary Operations. Integral Forms. 101

student's mind between the highest common factor of two

algebraical expressions, a term which refers to algebraical
form, and the greatest common measure of two numbers, a
term which refers to numerical magnitude.
177. If the two expressions be monomials (Art. 87) the
factors of each are obvious, and the highest common factor
of the two is the product of the highest powers of all
common letters which are contained in each.
Thus the highest common factor of the two monomials
a3b2c and abc³ is abc. For a is the highest power of a
which is a factor of both, 2 the highest power of b, and c
that of c.
If the two monomials have numerical coefficients with a
common numerical factor or factors, it is usual to multiply
the highest common factor of the literal part by the largest
numerical common factor of the coefficients, and to call the
result the highest common factor or, more incorrectly, the
greatest common measure of the two monomials. Thus
the highest common factor of the two monomials
15a3b2c and 12 ab2c3
will be said to be 3ab²c .
As far as regards its dimensions (Art. 86) this factor is
no higher than abc. It is of course numerically larger.
178. The numerical magnitude of an expression may or
may not increase as the dimensions are increased. Thus if
a have the value 2, the expressions a², a³, a4, ... of successively
higher dimensions have the successively larger numerical
magnitudes 4, 8, 16 ... If, on the other hand, a have
the value , these same expressions have the successively
smaller values 1, 1, 1 .... The student must therefore
remember that in investigating the form of the highest
common factor we are not necessarily discovering the
magnitude of the largest numerical factor for any
particular values of the letters involved.
102 Highest Common Factor. [179.
179. The process for finding the highest common factor
of two polynomial algebraical expressions is identical in
form with that for finding the greatest common measure of
two numbers in Arithmetic. The proof of the validity of
the algebraical process also obviously includes that of the
arithmetical one, for if all the algebraical factors represent
positive integers the two problems are identical in substance
as well as in form. The scope of the algebraical theorem
is however much wider than that of the arithmetical one.
180. The process may be briefly described thus. Let a
and 6 represent the two polynomials, and let a be that one
of the two which is not of higher degree in the common
dominant letter than the other. If there be no common
letter, obviously there can be no common algebraical factor.
Take a as divisor and 6 as dividend and carry on the
process of division until the remainder is of lower degree
than a (Art. 158). Let the quotient be denoted by pand
the remainder by c.
a) b (p
μα

c) a ( q
qe

d) c ( r
rd

Reference to the examples of division in Arts. 152-156

will show that, whether the quotient consist of one term
or many, the divisor a has been multiplied in succession by
the different terms of 1, and on the whole pa has been
subtracted. Hence (Art. 157)
b = pa + c ; (1)
or subtracting pa from these equals,
b -pa = c. (2)
181.] Highest Common Factor. 103

The remainder c is now to be taken as divisor and a as

dividend. Let q be the quotient and d the remainder,
which must be (Art. 158) of lower degree than c. As
before, we get the two relations
a = qc + d, (3)
a - qc = d. (4)
This process canbe carried on until one of the remainders
divides exactly the preceding one. Whatever the degree
of a or b may be, since the degree of each remainder is less,
by unity at least, than that of the previous one, the process
must come to an end either by one of the remainders
dividing exactly into the previous one, or by the last
remainder being a number, independent of the dominant
letter. In the latter case this remainder will algebraically
divide exactly into the previous one if fractional numerical
coefficients be admitted, and therefore in any case the
process will terminate by one remainder dividing exactly,
from an algebraical point of view, into the previous one.
Suppose that d is the last remainder and that it divides
exactly into e with quotient r.
It obviously follows that
c = rd. (5)
181. It can now be proved that d is the highest common
factor of a and b.
For the proof two consequences of the general laws of
multiplication are necessary.
(1) Every factor of any expression is also a factor of any
multiple of that expression.
For if a be a factor of 6, 6 must equal qa, and therefore
by Axiom (3) of Art. 53, pb, any multiple of 6, must equal
pqa. Hence a is a factor of pb.
(2) Every common factor of two expressions is also a
factor of their sum or difference.
For if a be a factor of b and also of c, b must equal pa
104 HighestCommon Factor. [182.
and e must equal qa. Hence, by Axioms (1) and (2) of
Art. 53, b + c must equal pa +qa which by (11) of Art. 53
is (p + q)a, and b-c must equal pa-qa which by (12) of
the same Art. is (p-q)a. Hence b + c and 6-c are both
multiples ofa.
182. By (5) of Art. 180, d is a factor of c. Hence,by
(1) of Art. 181, d is a factor of ge, and therefore by (2) of
Art 181, d is a factor of ge+ d or a by (3) of Art. 180.
Hence d is a factor ofpa and therefore of pa + c or of b ,
by(1) ofArt. 180.
Hence d is a common factor of a and b. Again, every
common factor of a and b is a factor of b -pa or c, by (2)
ofArt. 181 ; and therefore of qe, and therefore of a - qe or d.
But clearly no expression ofhigher degree than d can be
a factor of d. Hence d is itself the highest common factor
required.
183. It has been noticed in Art. 158 that in carrying
on the division to the point required in the different stages
in the foregoing proof, it may happen that the coefficients
in the quotient and remainders are fractional. This will
make no difference to the degree of the resulting expression,
but will produce it in a form not perhaps the most
convenient. In order to avoid such fractional coefficients,
certain precautions are adopted. In the first place, it is
desirable before beginning the operation to discover and
remove any numerical factors of either of the given
expressions. If there be any common numerical factor,
this can be restored at the end to the Highest Common
Factor, as in Art. 176.
Secondly, if at the end of any stage the remainder is
found to have a numerical factor common to all its terms,
this factor may be removed. For clearly, if the first
precaution has been taken, this will not be in any sense
part of the Highest Common Factor.
184.] Highest Common Factor. 105

Thirdly, if at any stage the division cannot be carried on

without introducing fractional coefficients and the second
plan is not available, the dividend may be multiplied by
such an integer as to make the coefficient of its first term
exactly divisible by that of the first term of its divisor.
184. The student who has followed and understood the
proof in Arts. 180-182 will see that neither of the plans
suggested in the last Article will affect the algebraical
form of the resulting Highest Common Factor. Suppose,
for instance, that c, in Art. 180, contains a numerical factor
x, so that c = ac'.
Let c' be taken as divisor, and suppose that in order to
avoid fractional coefficients in the next division it is neces-
sary to multiply a by some number y, y not being a factor
of c.

The process will be as indicated in the accompanying

form-

a) b ( p
pa

c= x

c) ya ( q
qc

d) c ( r
rd
...

and the relations ofArt. 180 will be replaced by

b-pa = c = xc', or b = pa+ xc
ya- qe = d, or ya = qo' + d
d = rd.
From these, as in Art. 180, it will follow that dis a
measure first of c and then of qc' + d or ya. Now as it
has been agreed to take away all numerical factors at each
106 Highest Common Factor. [185 .
stage of the process, in accordance with the second
precaution of Art. 183, d cannot have any factor common
to y, hence d must be a factor of a. Therefore d being
a factor of a and c', is a factor of pa and ac', and therefore
of pa + xc or b.
Again, all common factors of a and b are factors of pa
and b, and therefore of b -pa or æc'. But a and b have no
common factor which is a factor of æ by supposition.
Hence every common factor of a and b must be a factor
of c', and therefore of qe', and therefore of ya - qe' or d.
Whence, as before, d is the highest common factor.
185. The proof in the last Article also shows that it is
allowable at any stage of the process to remove any literal
factor, provided that it is not a factor common to the two
original expressions. The only condition required to be
satisfied by the factor a removed from e in the last Article
was that it was not a common factor of a and b.
186. As an example, let it be required to find the H. C. F.
or Highest Common Factor of x³- 3x + 2 and 4-1 .
Taking the former as first divisor and arranging the
division in the usual way, a quotient & and a remainder
3x² - 2x - 1 are obtained. The latter must be taken as
divisor, and, in order to avoid a fractional coefficient in the
quotient, x³ - 3x12 is multiplied by 3 before it is taken as
dividend.

x³ - 3x + 2 ) x* -1 ( x
x²- 3x² + 2x

3x² - 2x - 1

3 x2 - 2x - 1 ) 3x3-9x + 6 ( x
3x² - 2x²-

2 | 2x² - 8x + 6
x²- 4x + 3
187.] Highest Common Factor. 107

x² - 4x + 3 ) 3x2 – 2x - 1 ( 3
3x² - 12x + 9

10 10-10

-1

x - 1 ) x2- 4x + 3 ( x - 3
02-

-3x + 3
-3x + 3

At the end of the first step of the second division, the

remainder is 2x² - 8x + 6 , and as a factor 2 goes into all
the terms of this expression it is better to remove it before
proceeding. Ordinarily it is better to carry on each
division until the remainder is of lower degree than the
divisor. In this case however it will avoid fractional
coefficients or the introduction of a new multiplier to take
this simplified remainder at once as divisor. We then
obtain a quotient 3 and a remainder 10-10 . The factor
10, which is not a factor of either of the original expres-
sions, may be removed, and 2-1 taken as the next divisor,
when it is found to go exactly. The H. C. F. required is
therefore -1 .
187. It may be worth while to go through the proof in
the particular example last worked, that x- 1 is the
H. C. F.
The division shows that x- 1 is a factor of x² - 4x + 3 .
It is therefore a factor of 3 (x² - 4x + 3 ) or 3x² - 12x + 9,
and being a factor of 10x- 10 it is a factor of their sum or
3x² - 2x - 1 .
Hence x- 1 is afactor of 2x² - 8x +6 and of 3 x³ - 2x² -0,
and therefore of their sum or 3x3-9x+ 6, that is of
 108 Highest Common Factor. [188.
3 (x³ - 3x + 2). But 2-1 is evidently not a factor of 3.
Hence it must be a factor of x³- 3x + 2, and therefore of
x (x³ - 3x + 2 ) or x - 3x² + 2x ; and therefore of the sum
of this last expression and 3x² - 2x- 1 or of 4-1 .
Hence x - 1 is a common factor of the two given
expressions.
Again, every commonfactor of the two given expressions
divides 4-1 and x (x³- 3x+ 2) or x²- 3x² + 2x, and
therefore divides the difference of these, or 3x² - 2x- 1 .
Hence it is easily seen from the second division that every
such common factor divides 2x² - 8x + 6 or 2 (x² - 4 x + 3).
But no common factor of them divides 2. Hence every
common factor must be a factor of x² - 4x + 3. From the
third division it similarly follows that every such common
factor is a factor of 10x- 10 and therefore of x- 1, since
none of them can be a factor of 10.
Whence 2-1 is the Highest Common Factor.
188. If there be more than one letter involved in the
two expressions it may be desirable or even necessary to
arrange them, as in the multiplications of Arts. 129-135
or the division ofArt. 162..
Let it for instance be required to find the H. C. F. of
x² + y² + z² + 2yz + 2zx + 2xy and x³ +y + z³ - 3xyz.
Taking a as the dominant letter, the expressions will be
written as in the form annexed :

x² + 2(y+ z)x x³-

)) x² + 2(y +3yzx
+ (y² + 2yz + z²),
+ (y + z )
z)x² + (y² + 2yz + z²)x
(x-2(y+z)
-2(y + z)x²-( y² + 5yz + z²)x + ( y + z )
-2(y+ z)x²- (4y² + 8yz + 4z²)x- ( 2y³ + 6y2z + 6yz² + 2z )
3 | (3y² + 3yz + 3z² )x + (3y³ + 6y2z + 6yz² + 3z')
(y + yz + z²)x + ( y³ + 2y²z + 2yz² + z )
190.] Highest Common Factor. 109

At the end of the first stage a remainder is arrived at of

the first degree in æ, every term of which contains a
numerical factor 3. This may be removed.
A more formidable difficulty to proceeding is however to
be found in the still remaining coefficient of æ, namely
y² +yz + z². This is evidently not part of the H. C. F. ,
since it could not divide an expression arranged in powers
of a unless it were a factor of the coefficient of each power,
which it clearly is not in the case of the two given
expressions. Fortunately the term independent of æ in
the remainder, namely y³ + 2y² z + 2yz² + z³, is tolerably
easily seen to be the product of y + z into y² + yz + z2.
Hence the latter factor goes into both terms of the re-
mainder, and not being part of the H. C. F. required may
be removed (Art. 185). There is left x + (y + z), which
must be taken as next divisor, and which is easily found to
be the H. C. F. required.
189. In any similar case the student must examine
whether a coefficient, such as y² + yz + z², occurring as that
of the first term in any expression which has to be taken
as divisor cannot be removed in virtue of its being a factor
of all the other coefficients. If it cannot, further pursuit
of the H. C. F. is probably hopeless unless the factor in
question goes exactly into the coefficient of the highest
term in the corresponding dividend.
190. The discovery of the highest common factor of two
expressions, when any such exists, leads to the resolution, in
part at least, ofeach of them into its component factors.
When this resolution can be completely effected it gives
the complete solution of a problem almost of equal import-
ance with that of finding the Highest Common Factor,
namely, that of discovering the Lowest Common Multiple of
two expressions. Any expression which contains another
as a factor is called a multiple of the latter (Art. 22).
110 Lowest Common Multiple. [191 .

191. By the Lowest Common Multiple of two expres-

sions is meant the expression of lowest degree or dimen-
sions of which they are both factors. An infinite number
of Common Multiples can be found, but the Lowest Common
Multiple, for which words the letters L. C. M. are usually
written as an abbreviation, is a perfectly definite quantity,
of which all the other Common Multiples are multiples.
192. In the case of two or more monomials (Art. 87)
the L. C. M. is discoverable by inspection. It will be ob-
tained by taking the product of the highest powers of each
of the letters involved that occur in any of the monomials.
Thus the L. C. M. of the monomials

abc, a²bc², c³, a² 14

is evidently a³l4c3. For no lower powers than a³, 14 or c³
will contain the powers of a, b, c in all the monomials as
factors.
193. If there be numerical coefficients to the mono-
mials, it is usual to find the Arithmetical Least Common
Multiple of these and to multiply the Lowest Common
Multiple of the literal factors by this number. The result
is then called the L. C. M., which letters may stand in
that case for either Least or Lowest Common Multiple of
the given expressions.
Thus the L. C. M. of
15a²bc, 12a2b3c², 1874c³, and 20a274
is said to be 180a343, because 180 is the least common
multiple of the numbers 15, 12, 18, 20 and a³³ is the
lowest common multiple of the literal factors.
The remarks of Art. 178 apply here also mutatis mutandis.
194. If the factors of any number of algebraical
expressions can be completely determined, their L. C. M.
can be written down by inspection, as in the case of
monomials.
196.] Lowest Common Multiple. 111

Let it be required, for instance, to find the L. C. M. of

x²- 2ax + a², x² + 2ax + a², and x²-a². By Arts. 114,
115, 116 these expressions can be written as
(x -a) , (x + a)², (x - a) (x + a),
and evidently the L. C. M. is (x − a)2 (x + a)², which can be
multiplied out ifdesirable.
195. If the factors are not obvious, the plan of the last
Article cannot be applied, but the following method can
beused.
Let a and b represent the two expressions and let a be
their Highest Common Factor, found by the methods
already explained.
It is then evident that a must equal pæ and b must
equalqx, where p and q have no commonfactor. For ifpand
q had any common factor y, yx would be a common factor
of a and b, which is contrary to the supposition that æ is
their Highest Common Factor. Hence no expression can
contain both a and b as factors unless it contain p, q and 2.
The L. C. M. of a and b is therefore pqx.
But pqx = p.qx = p . b.
But since px = a. p is the quotient when a is dividedby
a
a and =

a
Hence the L. C. M. of a and b = 2.6.
х

b
.a.
Similarly it can be shown to be X

The verbal expression of this rule is that, to find the

L. C. M. of any two expressions, their H. C. F. must be
found and one of them divided by this. The product of
the quotient into the other will give the L. C. M. required.
196. If the L. C. M. of more than two expressions be
required, that of the first two must first be found. The
resulting L. C. M. must be a factor of all common multiples
112 Lowest Common Multiple. [197.
of the first two (Art. 191). Hence if the L. C. M. of this
L. C. M. and the third expression be found, the resulting
expression is the L. C. M. of the first three expressions.
This process may be extended to find the L. C. M. of any
number of expressions .
197. As an example let it be required to find the
L. C. M. of
x² - 6x² + 11x- 6 and ³ - 12x² + 47x- 60 .
The process of finding the H. C. F. ofthese two expressions
is shown below :
x² - 6x2 + 11 x − 6 ) x3 - 12x² + 47x - 60 (1
6x² + 11x- 6
- 6 6x² + 36x- 54
x²- 6x + 9
-

x2 − 6 x + 9 ) x3 – 6 x2 + 11 – 6 ( x
x²- 6x² + 9x

222-6

2-3

x - 3 ) x² - 6x + 9 ( x - 3
x²- 3x

-3x + 9
-3x + 9

The H. C. F. is thus 2-3 . It may be noticed that at

the end of the first stage of division the factor - 6 has been
used as a divisor. The sign is prefixed in order that the
sign of the first term of the next divisor may be positive.
This is a matter of convenience but not of necessity.
Taking 2-3 as divisor, and dividing 23-62² + 112-6
by the method of Art. 170 the quotient is found to be
x²- 3x + 2 .
x - 3 ) x³ - 6 x² + 11 x— 6
2-3x + 2
198.] Elementary Operations, etc. 113

Hence the L. C. M. of the two original expressions is

(x² - 3x + 2) (x3 - 12x² + 47x - 60).
198. The process of finding the Highest Common Factor
in the last question has shown that 2-3 is a factor of
x³- 12x² + 47x
47x-- 60 . The other factor is easily found by
division to be x² - 9x + 20 .
Hence the L. C. M. in the last Article can be written
2
(x² - 3x + 2) (x - 3) (x² - 9x + 20).
Andby comparison with the result (y) ofArt. 130, namely
(x -a) (x - b) = x²- (a + b) x + ab,
it is easy to see that
x² - 3x + 2 = (x - 1 ) (x - 2),
and that x² - 9x + 20 = (x - 4) (x - 5).
Hence the two original expressions are respectively
(x -1) (x - 2) (x- 3) and (x-3) (x - 4) (x - 5);
and their L. C. M. is
(x - 1) (x- 2) (x-3) (x - 4) (x -5),
which form is for many purposes more convenient than
that of the last Article.

EXAMPLES.

1. Find the Highest Common Factors of

(1) 15a²bc, 25a²bc², (2) 36xyz², 24xy²z .
2. Find the Highest Common Factor of
18abc and 24a²bc- 30 abc².
3. Find the Highest Common Factor of
x²-4x+ 3 and x² -5 + x6.
4. Find the Highest Common Factor of
x³- 3x² + 3x- 1 and x² - 6x² + 11x- 6.
Find the Highest Common Factor of the pair of expressions
in each of the twelve following examples.
5. x -7x+ 6, x² - 4x² + x + 6 .
6. x -x²-8x + 12, x² + 4x³ - 3x- 18.
I
114
Elementary Operations, etc.
7.9x² - 13x²y² + 4y², 15x³ - 19x²y + 4y² .
8. 11x² - 9ax -a²x² -a , 13x²- 10 ax³ - 2a²x² - a*.
9. x - 2x² -x -x² + 2x + 2, x -x -x- 1.
10. x² - 4x² + 7x² - 11x + 7, 5x - 18x² + 19x- 6.
11. x² + 4x² + 16 and x² + x² + 8x + 8.
12. x² + 3px²- (1 + 3p), px³ - 3 ( 1 + 3p) x + 3 + 8р.
72

13. хт -ут, х" -y" ; (1) when mand n are both odd
numbers ; (2) when m is odd and n is even ; (3) when m
and n are both even.

14. x² + (a + b) x² + 2abx² + ab (b + c) x + abc, and

x² + (b + c) x² + 2bcx² + bc (a + b) x + abc.
15. 6a + 15ab - 4a³c² - 10a²bc², and
9a3b - 27 a²bc- 6abc² + 18bc3.
16. a³ (b−c) + b³(c -a) + c² (a − b), and
bc (b -c³) + ca(c³-a³) + ab (a³-b³).
Find the Lowest Common Multiple of the expressions in
each of the nine following examples :-
17. ах-х², а²-x².
18. x² -у², х + у, х² + ху.
19. x³ - 3x + 2, x³ - 4x + 3.
20. x² + 67x² + 66, x² + 2x² + 2x² + 2x + 1 .
21. x³ -y³, x + y²³, ху² -у³, x³-ху².
22. x² + 3x + 2, x² + 4x + 3, x² + 5x + 6.
23. x²- (a + b)x + ab, x²- (b + c) x+ bc, x²-(c + a) x + ас.
24. 9x* -13x²y² + 4y², 15x³ - 19x²y + 4y³.
25. x²- 9x- 10, x² - 7x- 30, x² + 4x + 3.
26. Find the Highest Common Factor and Lowest Common
Multiple of 6x -x²-2x- 15 and 2x²- 5x² + x + 3.
27. Also of the expressions
6x -xy- 3x²y + 3xy -y and 9x - 3x²y- 2x²y² + 3xy -y*.
28. Also of the three expressions
x² - 6x² + 11x- 6, x³ - 8x² + 19x - 12, x² - 9x² + 26x - 24.
 Elementary Operations, etc. 115

29. Find the Highest Common Factor of

x² + x + x - 3x² + x - 7x² - 3x- 15, and
x -x + x² + 3x - 3x² + x - 2x² + 6x- 12
30. Find the Highest Common Factor and the Lowest
Common Multiple of x² + a²x² + a* and x² - 2ax + ax - 2a .
31. Find the Lowest Common Multiple of
2-3ax + 2a², 3x² - 19ах + 28a², and x²- 5ах + 4а².
32. Find the Lowest Common Multiple of
x² + 3x² -6x- 8, x² -2x² -x + 2, and x² + x- 6.

12
 CHAPTER VII .

FRACTIONAL FORMS .

199. The meaning of a fractional form, as , in any case

when a and b are positive integers, has been explained in
Art. 30. It has been shown in Arts. 43-46 that the
a
original definition leads to another, namely that is the
a
result of dividing by b, so that may be taken to be
a a

quantity which when multiplied by b gives a. InArt. 48

this second meaning has been taken as the definition of
a
the form ,
when a and b are not integral numbers but
fractions.
a

The algebraical form b7 can thus be defined as a quantity

which when multiplied by or into b gives a. The quantity
a is still called the numerator, and the quantity b the
denominator, although the original reason for those names
no longer applies.
200. With this meaning, the following are the principal
laws which govern operations on Fractional forms. The
numbering of the equations of Art. 53 is retained.
α
pa
6 (13)
α C a+ c
+
し
=

6
(14)
a C a-c
(15)
203.] Fractional Forms. 117

ac ac ca
; (16)
and as a particular case,
C ac C
a. == a.
d

a
= ab. (17)
a ab
b. = abb = a. (18)
66
a C ad ad
= =

bc (19)
201. These laws have been proved to be true when a, b,
c, d, p are integral or fractional numbers. If they hold
good for other meanings of a, b, c, d, p, all their conse-
quences will also hold good for those other meanings.
202. The most important problems to be considered in
dealing with fractional forms are, first, the reduction ofany
single one to its simplest equivalent form ; and secondly, the
reduction of an expression involving several of them to one
single equivalent form.
The first of these problems is usually known as that of
reducing a fraction to its lowest terms ; the other involves
the processes commonly called addition, subtraction, multi-
plication, and division of fractions.
203. In virtue of (13) it follows that the value of a
fractional form is not altered by introducing any common
factor p in numerator and denominator. It equally follows
that any common factor already existing may be removed
without altering the value of the fraction.
Hence if the highest common factor of the numerator
and denominator of a fraction be found and the numerator
and denominator be each divided by it, there will result an
equivalent fraction the degree of both expressions of which
is lower than in the original one. The resulting fraction
118 Fractional Forms. [204.
is then said to be in lower terms than the original one, and
as the remaining numerator and denominator have now no
common factor, no farther reduction can be made, and the
fraction is in its lowest terms.
204. As an example take the fraction
x²- 6x2 + 112-6
11 x —
x³ - 12x² + 47x- 60
In Art. 197 the highest common factor of the numer-
ator and denominator has been ascertained to be a- 3 . Re-
moving this factor from both, the fraction is reduced to
x² - 3x + 2
the form
x²-9x + 20 (see Art. 198), and is incapable of
farther algebraical reduction.
205. By means of (13) any two fractions with different
denominators can be changed into others with the same
x+ 1
denominator. For instance, the two fractions and
x- 1

-1 can be reduced to fractions having the denominator

x+ 1
(x +1) (x - 1) which is the L. C. M. of their denominators.
For, in virtue of (13),
x+ 1 (x + 1) (x + 1) (x + 1)2 x² + 2x + 1
= = =

2-1 (x+ 1) (x-1) x²- 1 x²- 1

(Arts. 114, 116),

2-1 (x -1) (x - 1) (x - 1 )2 x² - 2x + 1
and =

x+ 1 (x - 1) (x + 1) x²- 1 -

(Arts. 115, 116).

Hence
x+ 1 -1 x² + 2x + 1 x² - 2x + 1
+ +
2-1 x+ 1 2-1 x²- 1

=
(x² + 2x + 1 ) + (x² - 2x + 1)
x²- 1 , by ( 14),
2x² + 2
=

2-1 (Art. 91).

207.] Fractional Forms. 119

x+ 1 -1 x² + 2x + 1 x²-2x + 1
Similarly, 2-1 x+ 1
=

x²- 1 x²- 1

=
(x² + 2x + 1 ) - (x² - 2x + 1)
2-1 , by ( 15),
40

x²- 1
,
(Art. 95).
206. In a manner similar to that of the last Article the
algebraical sum (Art. 93) of any number of fractional forms
can be reduced to a single fraction whose denominator is
the L. C. M. of the denominators of the different fractions.
In doing this it will be most convenient to include the
numerical factors, if there be any, of the denominators in
the L. C. M., as suggested in Art. 193. This will have the
advantage of avoiding fractional numerical coefficients to
the different terms in the numerator or denominator.
207. Let it for instance be required to reduce the
expression
a2 a2 2 ах
2 2+
(x-a) (x + a)² x²-a2
to a single fraction.
In Art. 194 it is shown that the L. C. M. of the deno-
minators is (x- a)² (x + a)² .
a2 a2
a² (x + a)2
Hence ,
by(13),
(х -а)2 = (х - а)2 (x + a)2 '
a a

a² (x² + 2 ax + a²)
(Art. 114),
(x-a) (x + a) '
a²x² + 2a3x + a+
=

(x-a) (x + a)²
a2 a2
а² (х-а)2 a² x² - 2a3x + a *
Similarly, (x + a)2 = (x-a) (x + a)2 = (xa) (x + a)2
2 ах 2ах (х + а) (х-а) 2ax³- 2a3x
=
and =
2
x² -a2 (x-a) (x + a)² - (x - a)² (x + a)2
(Art. 116).
120 Fractional Forms. [208.
a2 a2 2 ах
Hence 2
+
(х-а)
a
(x + a)2x² -a2
a2x² + 2a3x + a * a²x² - 2a3x + a * 2 ax³- 2ax
-

+
(x-a) (x + a) (x - a)² (x + a) (x − a) 2 (x + a)²
a²x² + 2a3x + a* -(a²x² - 2a3x + a ) + 2ax³- 2a3x ,

(х-а)2 (x + a)²
by (14) and (15),
2a3x + 2ax³
=

(x - a) (x + a)²
208. The multiplication or division of one fractional
form by another is conducted by the laws (16) and (19).
It is usually undesirable to perform the multiplications
indicated, in the numerator or denominator, until it is
ascertained that none of the factors involved are common
to both. If there be any common factors these should be
removed, as explained inArts. 203, 204, so that the result
may be given in the lowest possible terms.
209. Let it be required to multiply ах
by
x² +a³
x³ (x²-a²)
The product is, by (16), ax ( ³ + a³)
But, by Art. 116, x² - а² = (x - a)(x + a) ; and, byArt. 133
(a), x³ + a³ = (x + a) (x²- ax + a²).
Hence there is a factor x + a common to the numerator
and denominator: there is also obviously a common factor
r: both of them may be removed, and the product in its
lowest terms becomes (х-а)
, in which the multi-
a (x²- ax + a ) '
plication indicated may now be performed, if it be desired.
210. As another example let itbe required to reduce to
its simplest form the expression
a a2

{( -a) - (c(x ++a)²

a) + } + { +a + a+a} •
210.] Fractional Forms. 121

In Art. 207 it has been shown that the dividend can be

2a3x + 2ax3 2ax (a² + x²)
reduced to which =
(x- a) (x + a)2 (x - a) (x + a)2
x+ a x-a
The divisor +
х-а x+ a

(x + a)2 (x-a)2
=

(x-a)(x+ a) + (x + a) (x-a)' by (13),

(x + a)² + (x − a)2
=
by(14),
(x-a)(x + a) '
x² + 2 ax + a² + x² - 2ax + a2
=

(x-a)(x+a) , (Arts. 114, 115),

2x² + 2a² 2(x² +a²)
= =

(x-a) (x + a) (x - a) (x+a)
x) _ ах x+ a
Hence {( a) -( + ) + 2 }+{ + + х-а x+ a
}
=
2аx (x² + a²) 2 (x² + a²)
+
2

(x - a) (x + a) (x -a) (x + a)
2 ax (x² + a²) (x - a) (x + a)
=

(x-a) (x + a)2 x 2 (x² + a²) by (19),

2 ах(x² + a²) (х -а) (x + a)
2 (х-а)2 (x + a) (x² + a²)
by (16).
And removing, in virtue of (13), the common factors 2,
x² + a², x-a and a + a from the numerator and denominator,
ах ах
this finally reduces to (х-а) (x + a) or x² - a2
One very commonly occurring case is that of the multi-
plication of a fractional form by an integral form.
This is included in the general case of (16), since an
integral form may always be treated as a fractional form
whose denominator is unity, for
a
a = a1 =
1 by (17), (or see Art. 34).

L
122 Fractional Forms. [211 .
b ab
Hence a-
C C

Thus
x- 1 (x - 1 ) (x-3)
x- 2
× (x - 3 ) = x- 2

If the multiplier be a multiple of the denominator the

result is an integral form.
x² + a2 (x² + a²) (x²- a²)
Thus x (x² - a²) =
х-а х- а

(x²+ a²) (x− a) (x +a), Art. (116),

х- а

= (x² + a²) (x + a) by ( 18) of Art. 200.

This last case will be found to be of great importance in
the solution of equations (Chapter X).
211. The reduction of a compound fractional form to a
simple one is effected by previous methods.
Thus, for instance, the compound form
X
+
x+ 1 x² + (x + 1)²
r+ 1 2 x(x + 1)
= ,
(Arts. 205, 206),
+1 X
(x + 1) -2
x x+ 1 x(x + 1)
=
+(1+ 1)
x(x + 1)
( x(x
+ 1)²-
+ 1)
2, by (17) ofArt. 200,
r² + (x + 1 )² x(x + 1)
=
by (19),
x(x+ 1) × (x + 1) -
x² + (x + 1)²
=
(Art. 208),
(x + 1 )² - x
2² + 22 + 1
=

2 + 1

As a somewhat more complicated instance take the fraction

1

1
2+
‫ند‬
212.] Fractional Forms. 123

1 1
This = ,
by(19),
1 X
X
x² + 1 x² + 1
х
1

x(x² + 1) -x ' (Arts. 205, 206)

x² + 1
x² + 1
x (x² + 1) -x
,
by (19),
x² + 1
=

212. An expression in a fractional form can be reduced

to a partly integral and partly fractional form in any case
when the degree of the numerator is as high as that of the
denominator.
x² - 3x + 2
Thus the form in Art. 204, x² - 9x + 20' can be written

as
(x² - 9x + 20) + 6x- 18 ,

x² - 9x + 20
x²- 9x + 20 6x- 18
which + by (14), Art. 200 ,
x²- 9x + 20 x²- 9x + 20 '
6-18
= 1+
-9x + 20

The numerator can always be divided by the denominator

until the degree of the remainder is lower than that of the
denominator (Art. 158). If the fractional form be denoted
N
byD'
and be the quotient and R the remainder when N
is divided by D, it follows from Art. 157 that
N = QD + R.
R
Hence N= QD+R
D
= QD
D ++ D by(14),
R
= 2+
124 Fractional Forms.

The converse process, of reducing a mixed form to a

purely fractional one, can be conducted by exactly reversing
this operation.
1
Thus x- 1 +
x- 2

= (x-1)(x-2) +12
=
(x- 1) (x- 2) +1
x- 2

x²- 3x + 3
-2

EXAMPLES.

x²- 3x+ 2
1. Reduce to its lowest terms
x²- 4x + 3
Reduce to their lowest terms each of the fractions in
the twelve following examples :
x²-a² a (x³-a³) x - 7x+ 6
2. 3. 4.
ax+ a² x² + a²x² + a* x²-4x² + x + 6

5.
9x - 13x²y² + 4y² 6.
x² + 3px² - (1 + 3p)
15x³ - 19 x²y + 4y px³- 3 ( 1 + 3p) x + 3 + 8p
x² + 4x² + 16 a² (b -c) + b² (c -a) + c²(a−b)
7. 8.
x +8 a³ (b- c) + b (c- a) + c²(a−b)
x² + 4x² + 16 2x² - 2x + 6x²- 6x
9. 10.
x² + x + 8x + 8 3x²-6x² + 3x
x² + 7x + 7x²- 15x
11.
x²- 2x² - 13x + 110

12.
ab (a − b) (a² + b²) + bc (b − c) (b² + c²) + ca (c-a) (c² +a²)
a²b² (a − b) + b²c² (b -- c) + c²a² (c- a)
13.
a³ + ( 1 + a) ab + b²
a²-b²
 Fractional Forms. 125

Reduce to their simplest forms the expressions in the

following examples :
a+ba- b 4ab
14. +
a- ba+ b a²-b²
2x 1 40
15.
2x- 1 2x + 1 4x² - 1

16. 2(2-1)+2-2 + 2(2-3)*

3 2 2
17. +
x- 1x- 2 x- 3

6 6 5
+
18. -12-3
19. x+ y 2x
x -x²y
+ - x+y+yy
y³ -x²y
2
1 a x+ a
20. + 3
x- 2x² - 8a³ x² + 2ax + 4a²
a -a²b + ba+ b a-b 1
21. +
a -b a³-b³ a³+b³ 2
a²-b²
a³-b³ a-b 1 a+ b 1
22.
(a² + b² -

)
2

a -ba² - b² a+ b

11 x+ y
23. х
++ ( -3)2
x²
Simplify-
24. Simplify X
2

25. Simplify a-b +ab

x+y )
26. Simplify 1++++
1 (x-y) } { + 2
}
2

y
27. Simplify ++++ х-у)
126 Fractional Forms.

a a
+-- 2 -+- + 2
a х a X
28. Reduce to its simplest form x- a
+
x+a

29. Reduce to its simplest form

(x -b) (x−c)+ (x- c)(x-a) + (x-a) (x-b).
(a − b) (a - c) (b − c)(b - a) (c - a) (c - b)

30. Simplify (a+b+a+b) + ( - )

x + ax + bx + c 3 (x + a) (x + b) (x+c)
31. Simplify + +
x-a x-b x- c (х -а) (x - b) (x − c)
1 1 1 1 1 1 1 1 1
+ +
ca ab bc ab+bc
ab ca be ca ab
+ +
32. Simplify (b + c)-- a (a + c)² - b² (a + b) ²- c²
1
33. Simplify 1 1 1
х- x+
1 1
x+ X
X X

34. Simplify
3 2 2 6 6 5

x- 3 ) ÷ ( x- 1 x-2
+ +
x- 1 x-2 x 3).
35. Reduce to its simplest form
1 1
18

1
X
+
1 1
1+
X X

1 1
1
X X

1 1
1+
X x

1 1 1

36. Simplify (a−b)(a - c) +(b −c) ( -a) + (c-a)(c-b)

 Fractional Forms. 127

37. Reduce
(ab) (ac) + (b −c) (b -a) + (c-a) (c -b)
x+ 12x + 11
y+ 11 y + 12
38. Simplify
y+ 12 y+ 11
x+ 11 x + 12
and find its value when x = 1 and y = 2 .
39. Reduce to its simplest form
х- а x-b x- c
+ + c_3 (x-a) (x-b) (x−c)
x + ax + bx + c (x + a) (x + b) (x + c) ,
X X X x² + (bc + ca + ab)x᾿
x+a + x+b + x+ c-3 (x + a) (x + b) (x + c)
X

40. Reduce to its simplest form X

2
X
2-
х
2
2-x
 CHAPTER VIII .

FORMS INVOLVING FRACTIONAL OR NEGATIVE INDICES,

OTHERWISE IRRATIONAL FORMS .

213. In Art. 66 it has been shown that the only mean-

P

ing that can be given to a symbol such as a¹ , consistent

with the general index law (22), is that it must be such
a quantity as when raised to the qth power shall give
ar as the result.
214. It remains to show that with this meaning given
to the symbol the laws (22) Art. 63, (23) Art. 68, (24) Art.
72, and (25) Art. 73 will hold good for all fractional and
negative values of m and n.
These laws are
am xa" = am +n, (22)
ama" = am- ", (23)
(am)n =amn = (a")m, (24)
(ab)" = a"l". (25)
They have been shown to hold when m and n are any
positive integers, the only limitation to their values being
that in (23) m is supposed to be numerically larger than n.
215. Before proceeding to the proof it is important to
consider what meaning can be attached to a symbol such
as at in a case where it directs the performance of an opera-
tion on a which cannot be exactly numerically performed.
For instance, if a represent 2, the symbol 2 denotes a
quantity such that the square of it-that is, the product of
itself into itself-shall be 2. Now no such number exists
216.] Fractional or Negative Indices. 129

in the form either of an integer or of a fraction whose

numerator and denominator are integers.
The proof of this fact is contained in the next three
articles.
216. It has been shown, in Art. 23, that if the sides of
a rectangle AB and AC contain a and b units of length
respectively, a and b being integers,
C D
the area of the rectangle contains
ab unit squares.
The same proposition holds good
A B
ifAB contain unit lengths and AC
9
r
contain ‫و‬
where p, q, r, s are positive integers. For
8
9
r gr
is the same thing as ps and 8
is the same thing as ".
qs qs

Hence if a length of which (gs) make a unit be taken for

a new unit, AB contains ps of these and AC contains qr.
Hence the area ABDC contains ps.qr squares, of which
the side is the new small unit length.
But since the old unit length contains qs of the new one,
the square whose side is the old unit contains (qs . qs) or
(q2s²) of the squares whose side is the new.
Hence the area ABDC contains pqrs squares of such a
size that the unit square contains q2s² of them, that is
ABDC is representedby the fraction Pars
92 82
of a unit square
pqrs pr
(Art. 30). But =
Pr, removing the common factors
92 82 qs
qs from numerator and denominator, or the area ABDC
T

contains , that is 8
unit squares.
qs 9

Hence if a and b represent the two sides of a rectangle, ab

represents the area, whether a and b be integers or fractions.
K
130 Forms involving [217.
A similar extension of Art. 24 can be made by the
student.
217. Let now ABCD be a square the length of whose
side is represented by a.
D
Let AC be joined. Then, by Euclid I.
C
47, the square on AC is equal to the sum
of the squares on AB and BC; that is, the
Square on AC is double the square on AB.
Hence if a represent the length AC, in
virtue of the last article we must have
B

x²= 2 a² .

Hence, in virtue of Axiom (4) of Art. 53, dividing these

equal quantities by a²,
x²
=2;
a2

or, in virtue of the law (16) ofArt. 53,

(*) = 2.
Whence by the definition of a fractional index a
=2 .
218. By the last article it follows that 2 can only be
expressed in the form of a fractional number provided AB
and AC can be both exactly expressed
C D
as integral multiples of the same
E
length. It can be shown that this
G
is not possible.
F Join AD and from AD cut off a
length AE equal to AB. Draw EF
A B
perpendicular to AD to meet BD in F,
and join BE.
Then, since AB is equal to AE, the angle ABE is equal
to the angle AEB. But the whole angle ABF is equal to
the whole angle AEF, both being right angles. Hence the
219.] Fractional or Negative Indices. 131

remainder, the angle FEB, is equal to the angle FBE.

Therefore BF is equal to EF.
Again, the angle DEF is a right angle, and EDF is half
a right angle, therefore DFE, the remaining angle of the
triangle DFE, is half a right angle, and equal to FDE.
Hence DE is equal to FE and therefore to FB.
Thus if there be any line such that it exactly measures
both AB and AD, DE their difference must also contain
this line an integral number of times, and therefore also
DF, which is the difference of DB or AB and BF or DE.
Thus if AB andAD have any common measure, that is
any length in terms of which they can both be measured
by integers, that length must also be a common measure of
DE and DF.
But DE and DF are the side and diagonal of another
square much smaller than ABCD, and by repeating the
same reasoning, any common measure of DE and DF must
also be a common measure of the side and diagonal of
a square smaller again. It is evident that this process may
be repeated until both the side and diagonal of the final
square are less than any finite length whatever.
Consequently no finite length whatever can be found
such that both AB and AD can be expressed exactly as
integral multiples of it.
Hence such a symbol as 2ª cannot be replaced by any
equivalent fraction whose numerator and denominator are
integers. P

The same statement applies to any such form as a

where a is a number, integral or fractional, which is not
exactly the qth power of some other integer or fraction.
219. Quantities such as those represented by 2 are

called incommensurable or irrational with regard to quan-

tities represented by numbers such as were employed in
Chapter I. The word incommensurable denotes that the
K2
132 Forms involving [220.
magnitudes represented by the two kinds of numbers
respectively have no common measure, the word irrational
that their ratio to one another cannot be expressed by
a single simple numerical relation. To the latter subject
we shall return presently (Arts 644, 645).
220. It has been shown, by taking a specific instance
in Article 218, that incommensurable lengths do exist, and
that, if such lengths are to be represented by symbols of a
numerical kind, the symbols representing one class cannot
be expressed exactly in terms of those representing the
other. Such a symbol as 24 or √2 (Art. 67) does how-
ever represent a clear idea, when it has been connected
with the geometrical figure of Art. 218, and any other
symbols involving the index can similarly have their
representative lengths geometrically constructed. There
is no simple equivalent construction applicable to such a
symbol as 2ª, but nevertheless the student will probably be
able to convince himself that the symbol does, or may be
taken to, represent a length intermediate between two
lengths which can be represented by numbers commensur-
able with unity.
221. That lengths must exist which can only be
represented by incommensurable numbers seems tolerably
evident from reasoning of the following kind.
Let AB represent a straight line, and suppose a point to
travel along it from A to B. The distance of this point from
PQ B

A changes in an absolutely continuous manner. Between

any two points P and Q in AB, however near to one another
they may be taken, there is an infinite number of interme-
diate positions of the moving point. Hence if PQ be th 2

of AB, and AP containr parts equal to PQ, and con-

223.] Fractional or Negative Indices. 133

sequently AQ contains (r + 1) such parts, between the

r+ 1
two fractions - and of AB there must be an infinite
n n

number of numbers required to express all the lengths

intermediate between AP and AQ. This is true however
large n may be. Hence it seems reasonably to follow that
some of these lengths cannot be expressed exactly as any
fraction of AB, where numerator and denominator are
integers, that is, there must be lengths expressible only
by numbers incommensurable with unity.
222. It must be understood that for practical purposes all
lengths can be sufficiently nearly expressed by numbers
commensurable with unity. If PQ be taken so small that
the human senses are incapable of taking account of it, all
the lengths between AP and AQ will be, as far as human
consciousness is concerned, represented with sufficient
accuracy by the same number. That fact is however only
aconsequence of the imperfection of human faculties and
would cease to hold good as a fact in the case of beings whose
eyes had the power of an infinitely magnifying microscope.
For the purposes of strict mathematical reasoning,
as distinguished from practical calculation, it is necessary
to take account of quantities represented by numbers
incommensurable with unity.
223. Although it is not possible to find any fraction
whose numerator and denominator are integers, which shall
be exactly equal to any such number incommensurable with
unity as can be denoted by a symbol a , where p and q are
positive integers, it is possible to find two fractions with
any arbitrary denominator and numerators differing by
unity, one of which shall be greater and the other less
than any such number.
For let n be any integer whatever, and let a series
134 Forms involving [224.
123
of fractions -, - , - ... be formed. Let each of these
nnn

fractions be raised to the power q. The resulting powers

will form a series of magnitudes each successively greater
than the previous one, and by taking the numerator
sufficiently large the value of the power can be made
to exceed any number desired. Thus there must be powers
of two consecutive fractions, one of which is less than a",
while the other is greater than ar : for the supposition
P

that as is incommensurable with unity shows that none of

these powers can be equal to a". Hence there must be an
+19
integer such that () isless than a"and that N
is

greater than a". Hence the number whose qth power

P

is equal to a", that is, the number denoted by as, must lie
r r+ 1
between and
n n

As n may be taken of any magnitude whatever, it is

possible to find an ordinary fraction which represents the
value of an incommensurable number within any required
degree of accuracy.
224. It will, for instance,be capable of proof that, taking
n to have in succession the values 10, 100, 1000, &c., the
4 15
quantity 2 is greater than 1 and less than 16, or that it
142 1415
lies between and 18, or between 1
100 1000 and 1 , and
so on (Art. 397).
225. There are incommensurable numbers of a different
P

kind from those represented by such symbols as a. The

ratio of the circumference of a circle to its diameter and
the base of the hyperbolic system of logarithms are
instances in point. About all such numbers however two
228.] Fractional or Negative Indices. 135

things may be assumed: first, that there is always some

length which they may be taken to represent, any given
length being unity; and secondly, that, if the unit length
be divided into any number n of equal parts, an integral
number r can be found such that the length represented by
the incommensurable number will contain more than rand
less than +1 of these parts; that is, that the incom-
r

mensurable number will be greater than the fraction n

- and
r+ 1
less than the fraction ofunity.
n

226. The phrase ' incommensurable number' is con-

veniently used as an abbreviation for the more strictly
accurate phrase ' number incommensurable with unity.'
From the fact that incommensurable as well as commen-
surable numbers may be taken to represent lengths on any
given scale, all numbers of both kinds are sometimes called
' scalar quantities,' or, more shortly, ' scalars.' The quantities
i and w, which we shall meet with later on (Arts 271,277)
and which cannot represent lengths on a scale, are called
' non scalar ' or ' operational quantities . The adjectives
impossible, imaginary, and unreal are often used to denote these
latter quantities.
227. Let now a and b one or both of them represent
incommensurable numbers. It is necessary to interpret the
symbol a . b.
One method of interpretation is analogous to that used
in the discussion of the multiplication of fractions
(Arts. 33-36). In this method a.b is considered to
denote the result of an operation on the line b exactly
similar to that which, when performed on the unit line,
produced the line a.
228. As an instance, let AB represent the unit line and
let AC be the diagonal of the square described on AB ;
then AC is represented by the number 24.
136 Forms involving [229.
If then on AC a square be described and AD be its
diagonal, AD will be represented by 2. (21).
But a perfectly easy geome-
trical proof shows that AD is
equal to twice AB, or AD is
represented by the number 2.
Hence 2. (2 ) = 2, as ought of
course to be the case (Art. 64).
229. To take an instance in
which the two factors are not
equal.
A Let AB represent the unit
B

line and let it be produced to

F, making BF equal to AB. On AFdescribe the equilateral
triangleAFG, and draw AE perpendicular on FG. Then
AF contains 2 unit lengths, or AF
is represented by the number 2, FE
G
by the number 1 ; and if AE be
represented by the symbol æ, it
follows from Euclid I. 47 that
H
E
22 = 12 + x² , or x² = 3 ; consequently
AE must be represented by the
symbol 3 (Art. 64).
A B F If BG be joined and in BG, BH
be taken equal to AB, and AH be
joined, AH is represented by the symbol 2 .
Let EG be produced toKand EK made equal to EA and
AK joined. Then AK is derived from AE exactly as AH
is derived from AB. Hence AK will be denoted by the
symbol 2. (3 ).
Again, join HF, HK and BE. By the construction it is
easy to see that AH is equal to HF, and also that the angle
AHFis a right angle, while the angle AKE or AKF is half a
right angle. Hence Hmust be the centre of the circle which
231.] Fractional or Negative Indices. 137

passes round AKF. Therefore KH is equal to AH, and

the angle HAK to the angle HKA.
But the angle KAE is equal to the angle HAB, each
being half a right angle, and therefore the angle KAH is
equal to the angle EAB. Therefore the triangle KAH is
equiangular to the triangle EAB. Hence AK is derived
from AH by an exactly similar construction as that by
which AE is derived from AB, that is, AK is properly
represented by the symbol 3. (2 ).
Hence 2. (33) = 33. (23).
Hence the commutative law holds with regard to these
two numbers.
230. The student can easily see that in any case if the
triangle AEK is similar to the triangle ABH, it will easily
follow that the triangle AHKis also similar to the triangle
ABE. Hence if a, b be any symbols representing the
lengths of the lines AE, and AH when AB is the unit,
AK will equally be represented by a . borb.a. Thus the
commutative law holds for the multiplication of any two
numbers incommensurable with unity, assuming that all
such numbers can be represented by lines (Art. 225).
It should be noticed here that the symbols a, b only
refer to the lengths of the lines, not, as in Arts. 74-80, to
their directions.
The student can probably convince himself in a similar
way that the distributive and associative laws, and there-
fore the index law (Art. 63), hold for incommensurable as
well as commensurable quantities.
231. Another method of arriving at the same conviction
may be indicated.
The product ab may be interpreted in all cases to
denote the area of a rectangle whose sides are represented
by the symbols a and b. This has been shown to be true
when a and b are integers (Art. 23) and when a and b are
138 Forms involving [232.
fractions (Art. 216). It is reasonable to assume it when
a and b are incommensurable quantities. The results of
Arts. 223-225 could be used to show that any other
supposition as to the area than this would lead to an
absurdity.
With this assumption the commutative and destributive
laws obviously hold, and if the farther closely connected
assumption be made that the volume of a right solid is
represented in all cases by the product of the three num-
bers which represent its edges, the associative law will hold
also (Art. 24). It follows that the index law holds for
incommensurable numbers when the indices are positive
integers.
232. It can now be assumed that quantities such as a ,
even if really incommensurable with unity, may be treated
by the same laws as commensurable numbers.
P

It has been shown (Art. 66) that as must be a quantity

such that when raised to the qth power it shall give ar.
It follows at once that

a = (a ) , (a)
since either of these quantities when raised to the 9th power
gives a".
It may be noticed once for all that in dealing with
fractional indices there has been given as yet no indication as
P

to whether such a symbol as a may or may not have more

than one value (Art. 64). This point will be considered
later on. At present such an equation as the last given
can at the most be taken as an assertion of the fact that
P

one of the values ofas is the same as one of the values

of (ar) 1
233. Again, let a = aq .
235.] Fractional or Negative Indices. 139

Hence, raising both sides to the power q, which may be

done in virtue of axiom (3) of Art. 53, it follows that
x = a.

Again, raising both sides of this last equation to the

powerp, (x )² = a²,
or (x ) ² = a², by (24), Art. 214 .
Whence x is a quantity which when raised to the power q
gives a", that is, P

x = a² ;
or, replacing for æ its value,
1 P

(a²)p = a
a².. (β)
P

234. Again, let y = ar .

Therefore, by definition,
y =a ;

and raising both these to the power r, where is any

integer, gr
yar = apr.
pr

Therefore y = aqr, (Art. 66)

P pr
or aq = a .

235. Hence it follows that p, q, r, s being all integers

Pr ps rq

aq . a = aqs.a84, (Art. 234)

=
(a ) (a ) , (Art 233)
I

= (a )ps+ra , (22) (Art. 63, 214)

ps+rq
8q
= a , (Art. 233)
= aqa( + ) (2), (13) and(14) ofArt. 200.
Hence (22) holds when m and n are positive fractions.
140 Forms involving [236.
r
236. Again, if be greater than 8 ,

9
T
ps rq
P
s qs
aa = als a
I ps

=( ) a ( )
I ps-rq

= asq
ps-qr

SQ
= a

= a( - ), (13) and (15), Art. 200

that is, (23) holds when m and n are positive fractions, the
restriction of m being greater than n being still retained.
I pr

237. Again, let x = a², y = a² , z = (a² )' , p, q, r, s being

still supposed to be positive integers.
Hence y = xP, Art. 233
T

z = ys
Therefore z = y = xP", by (24), Art. 214
pr
_pr
= a

pr
Art. 233
= a² .
Hence (2*)2 = apr, Art. 66
or 282 apr
(24), Art. 214
pr

Hence z = aq ; Art. 66

pr

or, finally, (a) = a98. (8)

240.] Fractional or Negative Indices. 141

So that (24) of Art. 214 holds for positive fractional

values of m and n. PP

238. Again, let x = a² . 6º , p and q being positive in-

tegers.
Then w = (a . )
= (a ) ( ) , by (25), Art. 214
= a² . br, Art. 66
=
(ab) . (25), Art. 214
p
Therefore x = (ab)² , Art. 66
PP P
or
a . b = (ab) , ... (ε)
that is, law (25) holds when n is a positive fraction.
239. The laws (22), (24), (25) have been now shown to
hold for any positive values of m and n commensurable
with unity, and also (23) provided m is greater than n.
In Art. 69 it has been shown that the only admissible
interpretation of a symbol like a consistent with those
laws is that
1
a
an

It will be now shown that, with this interpretation, the

three laws will hold good for negative indices, and also
that the restriction on the relative magnitudes of m and n
in (23) may be removed.
240. With the assumed meaning it follows that
1
am.a am
an

am
a"
by (16) of Art. 200
142 Forms involving [241 .
Now if m be greater than n the factor a" may be
removed from the numerator and denominator of this
fraction, and it becomes am- " by (23). If, on the other
hand, n be greater than m, the factor am can be removed
from both numerator and denominator and the fraction
1
becomes But with the assumed meaning for a
an-m

negative index,
1

an-m
= a (n-m) = am-". (Arts. 54, 55)
Hence am xan = am-n = am + (-n) in all cases ; that is, the
law (22) holds when one of the indices is negative.
241. If both indices be negative ,
1 1
a-mxa- n
am an

am.a"' by (16) of (Art. 200)

1

am+n
= a-(m + n )

= a- m-n ;

whence the law (22) holds if both indices be negative.

242. By the use of negative indices the value of law (23)
is very much diminished. To divide by a" becomes in fact
the same operation as to multiply by a-", and obviously
with this interpretation (23) becomes a mere special case
of(22). At the same time it is made more general by the
removal of the restriction as to the relative magnitude of
m and n.
243. It easily follows that
1 1
mn.
(1) ( ) * = (a ) = a = a--1,
7727
247.] Fractional or Negative Indices. 143

1 1 1
a-mn
(3) (a-m)" = ( m)"= (am)"
am 77616

1
(1) (a-m)-" = (am ) *
1
73

=(a
m

1
amn = a (-m) (-n).
1

amn

Hence law (24) holds good when either or both of the

quantities m or n are negative.
244. Again,
1 1 11
= a-" b-".
(ab)-" = (ab)" an on an on

Hence law (25) holds when n is a negative quantity.

245. It may be noticed as an extension of (25) that
n

() = (ab-1)"
= a" (b-1)", by(25)
= a" b ", Art. 243 (β)
an
(26)
and that this result holds for all commensurable values of n ,
positive or negative, integral or fractional.
246. The value of aº, namely unity, has been deduced
in Art. 70. It is obvious that this value will enable the
symbol aº to satisfy the laws (22), (25), in any case when
it enters into combination with other symbols.
247. It will be desirable to some extent to enlarge the
idea of lower and higher powers which has been involved
in some of the processes of division of integral expressions.
Any power of a letter is higher than a second power of
144 Examples. [248.
the same letter when the index of the second power sub-
tracted from the index of the first leaves a positive
remainder.
5

Thus at is a higher power than as, because when

subtracted from 4 leaves a positive remainder . On the
other hand, a-5 is a lower power than a-3 because - 3
when subtracted from -5 leaves a remainder - 5 - ( - 3),
or -5 +3, or -2 (Arts. 54-55).
248. Thus the quantities
11 1
... a³, a², a, 1,,
a
,
3
a
, ...

each of which is obtained by dividing the previous one by

a, are a series in descending powers of a, for they may be
written with the notation of negative indices,
... a³, a², a , a , a-1, a2, a-3, ...
and the successive indices
3, 2, 1 , 0, -1 , -2 , -3
are obviously each obtained by subtracting unity from the
preceding one.

EXAMPLES.

1. Find the product of a , a , a , a , a , and as.

2. Multiply together x, x 4, 4, 4, 4, 4,
1
bc ca

3. Reduce to its simplest form (ab) . ((b ) ( )

a

4. Divide a boca by a b cहे.

5. Multiply 2x + 4x -x-1 by 3x-- 6x + x.
6. Multiply x -x² + 4x-1-8 by x² +x² + 4x + 8 .
7. Multiply a + ab + ab + by a -b .
8. Multiply a -ab + b by a + ab + b , and the product
by a-ab + b.
 Examples. 145

9. Divide x-y by x -y .
10. Divide 64x-1 + 27y2 by 4x + 3y .
11. Divide x + x + 1 by x + x + 1, and divide the quotient
by x -x + 1 .
+ y-zz by x +y +z .
12. Divide x + 2xy +y-
13. Divide

2yz + 2zx + 2xy -x -y -z by x + 2xy + y -z.

14. Find the Highest Common Divisor of
xy + xy+ x + x+ yz + yt zł-1
and xy + xy + x -xyz -yt zł - 1 .
15. Find the L. C. M. of
x-a, x + xas + a , x + x +as and x -a .
(x²-6x² + 11x-6)
16. Simplify
(x - 1) (x- 2) (x -3)
17. Reduce to its simplest form
1 1 1

(a+ b) +(a− b) (a + b) (a−b)

18. Simplify +
(a+b) -(a−b) (a+ b) + (a−b)
1

19. If a = m, a = n, and (a²) = m"n", prove that xyz= 1 .

20. Reduce to its simplest form
a+ (a² -b²) a- (a² -b²)
+
(a+ b) +(a−b) (a+ b) -(a-b

L
 CHAPTER IX.

SURDS AND IMPOSSIBLE QUANTITIES.

249. In the last chapter it has been shown that such

quantities as 3 , 25, cannot be expressed exactly in the form
of any fraction whose numerator and denominator are
integers and are therefore incommensurable with unity.
Incommensurable quantities of this class are called surd
quantities, or more simply surds.
250. The other notation (Art. 67) for 3ª, 2 , &c., namely
3

√3, 2/2, is more frequently used, and will be adopted

in the discussions which immediately follow. The student
should remember that the two symbols 3 and √3 are
strictly equivalent, and that the use of one rather than
the other is merely a matter of convenience. When dif-
ferent roots (Arts. 64, 67) of the quantities involved are
employed, the index notation is the more useful ; but when
there is only one kind of root, as the square root, the
symbol is sometimes easier to work with.
251. A surd quantity involving only the index , or the
square root, is called a quadratic surd.
252. In the case of quadratic surds, equations (25) and
(26) of the last chapter become
(ab) = a ,
a

() = ;
or, with the other notation,
√ub = √a . √o, (a)
 Surds and Impossible Quantities. 147

La Sa
(β)
Thus √6 = √2 . √3,
2√2 =

3
√3
253. The equation (a) of the last article can often be
employed to simplify a quadratic surd when the number
under the square root is not a prime number, that is when
it is resolvable into two or more factors .
Thus 12 = 4 x 3.

Hence √12 = √4x 3 = √4 . √3

= 2√3.
Similarly √75 = √25 . √3 = 5√3.
Hence √75 + 12 = 5√3 + 2√3 = 7√3.
254. A quadratic surd when reduced so that no factor
of the number under the root sign is an exact square is said
to be in its simplest form.
Surds other than quadratic can be similarly reduced and
simplified. Thus
3

24 = 8x3 = 8x /3
= 2 /3.
255. A fraction whose denominator is a surd can be
replaced by an equal one with a rational denominator, that
is with a denominator commensurable with unity.
2 √2 √3 √2
Thus == by (13), Arts. 53, 200,
3
√3 √3 √3
√6
=

3 by (a), Art. 252 ;

or generally, =
da =
do da =
Sab
b

256. In a later chapter (Arts. 397, 398) will be given a

L2
148 Surds and Impossible Quantities. [257-
method of finding a commensurable fraction whose square
approaches within any required difference, to any given
integer ; or, in other words, to find an approximate value of
a surd quantity such as Va.
It is evident that the methods of Art. 253, 255 will
enable approximations to be found to such quantities as
√75 +12 and with much greater facility than would
be the case were these quantities left in their original forms.
257. An expression involving two terms one of which at
least is a surd is called a binomial surd. Thus √3- √2,
1 +√5 are binomial surds.
258. An expression can always be found such that when
it is multiplied into a binomial surd, the product shall be
rational ; that is, commensurable with unity(Arts. 219, 255).
1

The most general form of a binomial surd is ab ,

where a and b are rational.
1

Let n be the L. C. M. of pand q: then (a )" = a , and

12

(6)" = b . (Art. 233. )

n n
But"and"areboth integers, since nis acommonmultiple
P 9

of p and q. Hence (a ) and(6 )" are both rational.

1 1

If a be denoted by a and be by y, the binomial surd is

denoted by x +y.
In Art. 168 it is shown that " -y" is divisible by x-y
always, and by x +y when nis even; also that " + y"
is divisible by x +y when n is odd. By the method of
Art. 170 the quotients can be ascertained in these cases,
and it will be found that
(xn -y") = (x -y) (xn- 1 + xn- 2y + xn-3y² + ... + xyn- 2 +yn-1),
(x −y ) = (x + y) (xn- 1 -x - 2y + x -3y² - ...
+xyn-2 -yn-1) if n be even,
(x +y ) = (x + y) (xn-1 - xn-2y + x -3y² -...- xym-8+ m-1)
ifn be odd.
260.] Surds and Impossible Quantities. 149

Now with the assigned meanings of x and y the left-hand

sides of these identities are rational. Hence the second
factor in each case is the factor required to rationalise
x -y or x + y.
259. The most important case is that of a binomial
quadratic surd, as√a + √b.
Here the formula ofArt. 116 is sufficient, and since by
that formula it follows that
(√a+ √ ) (√ā- √b) = a − b, (a)
the factors necessary to rationalise va+ √o and √a-do
are da-√ and √a + √orespectively.
260. A fraction whose denominator is a binomial surd
canhave its denominator rationalised by multiplying the
numerator and denominator by the factor indicated by the
processes of Arts. 258, 259.
√3+ 1
Thus the fraction
√3-1

(√3 + 1) ( √3 + 1)
=
by (13) Art. 200
(√3 + 1) ( 3-1)'
= (√3+ 1)2, (a) Art. 259
(√3)2-1
=
( √3)² + 2√3 + 1 , Art. 114
3-1

3 + 2√3 + 1 4 + 2√3
2 2
= 2 + √3.

A repetition of a similar process will often enable us to

rationalise the denominator of a fraction containing more
than two terms, some of which are surds. For example,
1 √2
+ √3- √5
√2 + √3 + √5 (√2 + √3 + √5) (√2 + √3 - √5)
150 Surds and Impossible Quantities. [261 .

=
(√2+ √3 - √5
(√2 + √3) -( 5)
=
√2+ √3- √5
5 + 2√6-5
√2+ √3- √5
2/6

=
(√2+ √3- √5)(√6)
12

2√3+ 3√2 - √30

12

261. A binomial surd can be multiplied by another

binomial surd by the ordinary rules of multiplication.
Thus ( √a + √o) ( √c+ √a)
= √a . Ve+ √a . √ā + √b . √c+ √b. √a
=
√ac + √ad + √bc + √bd.
The product is no longer a binomial surd. In two very
commonly occurring cases the product will however be a
binomial surd, the first being when each of the original
binomials involve only one quadratic surd, which is the
same in both.
For instance, let the two binomials be a + b and
c + d√x, where a, b, c, and d are rational ;
then (a + b√x) (c + d√x) = ac + ad √x + bc √x + bdx
= (ac + bdx) + (ad + bc) √x.
262. The same thing is also true of the product of two
binomial surds involving two quadratic surds, the same in
both.

Thus (a√x+ b√y) (c√x+ d√y)

= acx + ad √xy + be√xy +bdy
= (acx + bdy) + (ad+ bc)√xy.
264.] Surds and Impossible Quantities. 151

In this case the single surd involved in the product is

no longer the same as either of those in the two factors.
263. As a particular instance it follows that the square
of a binomial quadratic surd, whether it have only one or
both terms surds, is itself a binomial quadratic surd.
Thus (a + b√x)² = a² + 2ab √x + b²x
= (a² + b²x) + 2ab √x ; (a)
(a√x + b√y)² = a²x + 2ab √xy + b²y
= (a²x + b²y) + 2ab√xy. (β)
It may therefore happen that any given binomial quad-
ratic surd of which one term is rational (Art. 258) is
the square of some other binomial surd. The next few
articles lead to a criterion as to whether this be the case,
and, if it be, to the discovery of the binomial surd of which
the given one is the square.
264. As a preliminary proposition it is necessary to
prove that one quadratic surd cannot be either the sum of
two quadratic surds whose simplest forms (Art. 254) are
different from each other, or the sum of a rational quantity
and a surd.

For suppose it possible that ve = a√x+ b√y, where

✓e and √y are reduced to their simplest form and are
different. Then by axiom (3) of Art. 53
2

(√c)² = (a√x+ b√y)²,

or
c = a²x + b²y + 2ab√xy,
or
c- a²x -b²y = 2ab √xy by Axiom 2 (Art. 53).
Now c, a²x, by are commensurable with unity by
supposition ; consequently c- a²x-by is so also. Hence
2ab √ry is commensurable with unity. But by supposition
x and y are different numbers and neither of them contains
a factor which is an exact square. Let p be the G. C. M.
152 Surds and Impossible Quantities. [265.
of x and y, so that x = pu, y = pv, where u and v have no
factors in common.
Then xy = p²uv,
and therefore √xy = p√uv.
But all the factors in u are different from any in v, and
no factor in either u or v is an exact square. Hence uv
cannot be an exact square, and therefore ✓uv, and con-
sequently day, is incommensurable with unity.
Hence the supposition was untrue ; and in exactly the
same way it can be shown that de cannot be equal to
a+ b√y.
265. It follows that if such an equation as the following
hold between x, y, a and b,
x + √y = a + √ ,
where y and b are not exact squares, then a must equal a
and y must equal b.
For the equation can be written, subtracting a from
both sides (Axiom (2) Art. 53),
(x - a) + √y = √T,
which makes √o equal to a rational quantity together with
a surd, unless the rational quantity vanish. Hence x- a
must be zero, or a must equal a, and thereforey must equal b.
It further follows that if

x + √y = a + √b,
then x- √y = a - √i.
266. Let now a +√o be the given binomial surd, and,
if there be a binomial surd whose square is equal to a + √b,
let itbe √x + √y.
Then (√x + √y)² = a + √b,
or x +y + 2√xy = a + √b. Art. 263
Hence x +y = a
(a). Art. 265
2√xy= √IS
267.] Surds and Impossible Quantities. 153

From these equations by squaring each side of each, a

legitimate process in virtue of Axiom (3), (Art. 53),
x² + 2xy +y² = a² ,
4 ху = b.
Hence in virtue of Axiom (2), (Art. 53), subtracting the
lower equals respectively from the upper ones,
x² - 2xy +y² = a² - b,
or, by Art. 115,
(x −y)² = a² - b. (β)
Hence if a² - b be not an exact square, there can be no
binomial surd whose square is a + √b, and further investi-
gation is needless. If however a² -b be the square of
some number k, it follows that (3) is satisfied if we take
x-y = k.
And since x +y = a,
by adding these equals (Axiom 1, Art. 53) we obtain
2x = a + k,
therefore x = (a + k) ;
and by subtracting the upper equals from the lower ones,
2y = a - k,
or y= (a - k).
Hence the values of x and y are determined.
267. Let the given binomial surd be 83 + 12√35. The
process of the last article gives us
(√x + √y)² = 83 + 12√35 ;
whence x + y = 83 ,
2√xy = 12/35,
or, squaring both of these,
x² + 2xy +y² = 6889,
4ху = 5040,
and subtracting,
x²- 2 xy +y² = 1849,
or (x −y)² = ( 43)2 ;
154 Surds and Impossible Quantities. [268.
which is satisfied if
х-у = 43 ;
but x +y = 83 .
Therefore, adding, 2x= 126 or x = 63 ,
and subtracting the upper equals from the lower,
2y = 40, or y = 20 .
Hence the square root required is √63 + √20, which
can be simplified by the method ofArt. 253 into the form
3√7 + 2√5.
268. The student can easily verify, by comparison of
Arts. 114, 115 and 263, that if the square of √2 + √y
is a + √b, that of √2-√y will be a -√b.
269. The cube of a binomial surd of which one term is
rational is also a binomial surd. This follows from a com-
parison of Art. 261 and equation (a) ofArt. 263, but it can
also be seen by means of example (1) ofArt. 122. Thus
3
(x + √y)³ = x² + 3x²√y + 3x ( √y)² + ( √ )3
= x² + 3x²√y + 3 xy +y√y
= x3 + 3xy + (3x² + y)√y.
Occasionally it is possible, by means of this formula, to
find values of x and y which shall make (x + √y)³ equal to
a given binomial surd.
270. It has been seen that the formal laws of algebra
however interpreted require the assumption that ( -a)
(-6) = + ab (Arts. 58, 105). Consequently ( -a)² = + a² ;
that is, the square of a negative quantity as well as of
a positive quantity is positive.
Hence such a symbol as ( -a²) involves quite a different
sort of difficulty from that which was met with in the
symbol 3. In the latter case there is a number, repre-
senting a perfectly definite length, whose square is equal
to 3, and an approximation to the value of this number in
272.] Surds and Impossible Quantities. 155

fractions of unity can be obtained, by a process explained

hereafter (Arts. 397, 398), to any required degree of
accuracy. But no number whatever can have a negative
square. Hence such asymbol as (-a²) cannot denote a mere
number, and it is therefore frequently called an impossible
or imaginary or unreal quantity. A few articles must be
devoted to the consideration of its possible interpretation.
271. Since the formula (ab)" = a" b" has now been
established for all indices (Arts. 238, 244), such a symbol as
(-a²) may be written as ( -1) . (a²) , or ( − 1) . a.
Similarly (-a) may be written (-1) a ; and as the
symbols a and as have received an interpretation, it only
remains to give to (-1) an interpretation consistent with
the laws of operation proved to hold in the case of numerical
quantities.
Itwill be convenient to denote this symbol by the letter i,
where i is determined by the condition
= -1 . (a)
272. It is clear that i cannot represent a number, nor
indeed in any case the same kind of quantity as the
symbols a, b, c, ... hitherto employed. As it only occurs as
amultiplier of one of these other symbols it must denote
some operation performed on the quantity denoted by these
symbols ; imay therefore be called an ' operational quantity '
(Art. 226).
Thus ia must denote the result of some operation per-
formed on a, and ira must similarly denote the result of
performing this operation twice.
But ia - a.
Hence the operation denoted by i must be one which
when once repeated reverses the sign of a. It has been
seen that if a denote a line measured in one direction, - a
156 Surds and Impossible Quantities. [273.
denotes a line of equal length measured in the opposite
direction, a line which could be obtained by rotating the
first line in any plane through two right angles. Hence
the operation i might, so far consistently, be taken to
denote the operation of turning the line a through one
right angle in any plane which contains its original
position.
273. Let then OX denote the direction in which a is
measured, and let the plane of the paper be that in which
the rotation is supposed to take
B place. Measure a length a
along OX, and measure a length
b perpendicular to OX. Then if
OA be denoted by the symbol a,
A X
OB will be denoted by the
symbol ib.
Let the rectangle OACB be completed, then, as in Art. 74,
a being taken to denote a transference from O to A, ib will
denote a transference from A to C, and therefore a + ib
denotes a transference from O to C.
Let the numerical length of OC ber : then, by Euclid
I. 47, r² = a² + b² .
It is also clear by Euclid VI. 4 that the values of the
a b
fractions r
and depend solely on the angle COA, and not
r

on the lengths either of a or b. If these ratios be denoted

by the letters cand s, the following relation holds good :
a + ib = cr + isr
= (c + is) . r.
Here c + is is an operator depending solely on the angle
between OX and OC; and since the absolute length of OC
is r, the multiplier c + is, or, as it may be conveniently abbre-
viated, cis, must denote the operation of turning a length r
274.] Surds and Impossible Quantities. 157

from the direction OX through an angle corresponding to

the values of c and s.

274. As a particular illustration, let the angle COA be

the angle of an equilateral triangle. Then, if the triangle
OCB be completed, OA is half the C
base and therefore is half of OC.
OA
Hence in this case, c = OC
= .
Also OC² = OA +AC².
And since OC = 20A, OC² = 4OAº ;
whence 4042 = OA² + AC², A B

or 30A2 = AC ,
3
or
4
OC² = AC2;
AC2 =
3
whence
OC2 4'
AC √3
or
OC = 2'

since the squares of these two quantities are equal.

Hence 83. 2

Therefore c+ is = 1+i√3. 2

1+ i√3
Thus the multiplier 2
ought to denote a rotation
through one third of two right angles. Hence the multi-
1 + i√3
plier (1+ 13 ) ought to denote a rotation through two
2

right angles.
But, applying to the symbol i the ordinary rules of
algebraical multiplication, we find, remembering that
i = -1, and consequently i³ = −i,
158 Surds and Impossible Quantities. [275-

(1 +1 /3)3 = (1 + i√3)3_1 + 3i√3 + 3 (i√3)2+ (i√3)3

2 8 8
... (Art. 122)
1 + 313-9-31√3
8

= -1 ,
as ought to be the case.
275. If OC make an angle of two thirds oftwo right angles
with OX, OA is measured to the left, and consequently
C
√3
cis- , while s is still 2

-1 + √3
Hence c + is = •

-1 + i√3
The multiplier 2
X
A
ought therefore to denote
a rotation through one third of four right angles. The
multiplier (-1 + √3)32
should thus denote a complete
revolution.

But, as before, (-1 + √3) 2

=
( -1 )3 + 3 ( −1 )2i√3 + 3 ( −1 ) ( i√3)² + (i√3)
8

-1 + 3i√3 + 9-3i√3
8

= 1.

In both these cases we see that the interpretation given

to i leads to a result consistent with the previous meanings
of the signs + and
It is not possible to develope the subject fully without
some knowledge of trigonometry : enough has probably
276.] Surds and Impossible Quantities. 159

been given to show that the symbol i of the equation (a)

of Art. 271 is susceptible of interpretation, and that, with
that interpretation, it obeys the laws of other algebraic
symbols. No fear need then be entertained of the truth of
the results of investigations in which it occurs, provided
they are interpretable.
276. If, in the figure of the last article, CA be produced
to a distance AD equal to CA, and OD be joined, OD is of
equal length with OC; but since AD is C

measured in an opposite direction toAC,

AD is symbolically represented by -AC,
i√3
OC.
that is, by 2

Hence for the line OD, c + is becomes A X

-1-i√3
2

A rotation through two thirds of four

right angles, that is, from OX through D

OC to OD, is thus represented by the

-1-i√3
multiplier 2
2

But ((-1 + √32 _ (- 1)² + 2 ( -1)

2
=

4
i√3 + (i√3)²

-1-213-3-1-13
4 2

Here, again, a result is obtained consistent with the

interpretation assigned, since a rotation through the angle

COXtwice repeated is represented by (-1 + √3)) ,

2 and
-1 - i√3
one through twice the angle COX by , and
2
160 Surds and Impossible Quantities. [279.

these two symbols ought therefore to be equivalent, as they

have just been shown to be.
-1 + i√3
277. Let the symbol 2
be denoted by w, then
-1 - i√3
the symbol is w².
2

It has been shown that w³ = 1. Hence it follows, since

(ω²)3 = ( ω³) , (24) Art. 72, that (w²)³ = 1 .
Each of the three quantities 1 , w, w² has therefore the
property that its cube (Art. 62) is unity. They are called
the three cube roots of unity.
Any power of w as w" is a cube root of unity, since
(w")³ = ( ³)" = 1 ;
but the powers of w beyond w² only repeat the same series
6
ofvalues 1, ω, ω² in virtue of the facts that w³ = 1 , w = 1 ,
and so on.
Thus ω4 = ω ω = ω,
ω5 = ω ω²2 = ω²,
46
= ( 3) 2 = 1 ,
ω = ωω = ω,

ω = ωσω = 2.

278. Since there are three cube roots of unity, that is,
three different symbols such that when cubed they give
unity as the result, there are three cube roots of any
number a .
For a can be written as 1 xa, and therefore each of the
symbols απ, ωα , wat will produce a when cubed, where as
may be taken to denote the number, commensurable or not,
whose cube is a.
279. Similar considerations show that if a be a symbol
denoting rotation throughth part of a complete revolution,
a will denote rotation through four right angles, and must
280.] Surds and Impossible Quantities. 161

therefore be unity. Anypower of a, as a', is also a quantity

whose pth power is unity, for
(a ) = (a )* = 1* = 1 .
Thus all the series of quantities
1, a, a² ... ap-1
may be called pth roots of unity. If higher powers of a
than a -1 be taken, the same values will be repeated in
virtue of the facts that a" = 1 , a2 = 1 , and so on. Hence
it may be asserted generally that there are p symbolical
forms which when raised to the pth power give unity, or,
more concisely, that the number ofvalues of (1) isp (Arts.
67, 232).
280. The multiplier 1 denotes a rotation through any
1

multiple of four right angles. Hence (1) must denote a

rotation through one qth part of some multiple of four
right angles, or, if we use the symbol 2 to denote four
1

right angles, ( 1) must denote a rotation through some one

of the series of angles
2π 4π 6π (29-2) π 2απ (29 + 2) π
, , ...
, , ...
;
999 9 9 9

and it is evident that the geometrical positions denoted by

2απ
these angles up to or 2 are all different, but that the
9

next set of q angles merely repeat these with 2 added, and

give therefore geometrically the same positions. Hence
it follows, as in the last article, that (1) has q and only q
different values.
Aspecial case of great importance is (1) , which must
2π 4 π
denote a rotation through 2 2
..., thatis through π, 2 π...
Thus the two values of (1) are - 1 and + 1 .
M
162 Surds and Impossible Quantities. [281 .
Similarly - 1 denotes a rotation through any odd mul-
1

tiple of two right angles. Hence ( -1) must denote a

rotation through a qth part of some odd multiple of two
right angles, that is through one of the angles denoted by
π 3π 5π (29-1) π (29 + 1) π
, ...
, ......
qqq q 9

and, as before, only q of these are geometrically distinct.

It follows that either of the symbols (1) or( -1) , since
it denotes a rotation through some angle, must be capable
of being replaced by a symbol of the form c + is, where c, s
are numerical quantities (Art. 273). This last result will
be true even if q be incommensurable with unity.
281. The symbols i, w and similar ones can be used
sometimes to give a formal resolution of algebraical expres-
sions into factors. Thus, for instance, the expression x² +y²
can be written x² - i²y², i having the meaning ofArt. 271 .
Hence x² + y² = (x + iy) (x - iy).
The student can exercise himself by trying to give a
geometrical interpretation to this identity by the help of
Art. 273 .
282. Again, it has been proved in Arts. 134, 160 that
x +y + z is a factor of x³ + y + z³- 3 xyz.
But the latter expression can be written in either of the
forms
x3 + (ω )3 + ( w2z)3 – 3 x . ω . ω² ,
2
or
x3 + (w²y)3 + ( wz)3 – 3x . w²y . wz,
since w³ = 1 ; from which forms it follows that it must con-
tain both x + wy + w²z and x + w²y + wz as factors. Since the
original expression is only of three dimensions in x, y, z
(Arts. 86, 88), it cannot be the product of more than three
linear expressions (Arts. 86, 88, 112). Hence the only other
283.] Surds and Impossible Quantities. 163

factor possible must be a numerical or symbolical one not in-

volving a, y, z. Ifthis factor be denoted by k, we must have
x3 +y³ + z³ - 3xyz = k (x + y + z) (x + wy + w²z)(x + w²y + wz),
and by comparing the term involving a³ on both sides it is
seen that k is unity. Hence
x³ +y + z³ - 3 xyz= (x + y + z)(x + wy + w²z)(x + w²y + ωz).
283. In the last article it may be noticed that, by
exactly similar reasoning, the four expressions
wx + y + w²z, w²x + y + wz, wx + w²y + z, w²x + y + z
must also be factors of x³ + y + z³- 3 xyz.
These factors however are each equal to one of the latter
two given in Art. 282 multiplied by some power of w.
Thus wx + y + w² z = wx + w³ y + w²z, since w³ = 1 ,
= ω (x + w²y + ωπ),
and the resolution in the last article is a complete one,
although the factors may be put into different shapes.
This first section of the book has been occupied in
deducing the laws of algebraical operation and in explaining
the various symbolic forms which may arise in applying
those laws.
The next section will be devoted to the application of
these laws to the solution of algebraical equations.
EXAMPLES .

1. Reduce to their simplest forms each of the surds

3
√32, 16, 18, 64.
2. Reduce to its simplest form
2 + √54-24.
3. Given √2 = 1.4142, calculate the values of
1 3+2
/2
(1) (2)
√2+1' 3-2-2

to four places of decimals.

M2
164 Examples.
3

4. Simplify
5. Rationalise the denominators of
4 4
(1) ;
(
2) 1+√2+√3
√3+ √2

6. Simplify (5+4-√3
2/3)* x (2-3)2
√3+1 .
( √2+ √3)( √3+ √5)(√3+ √2)
7. Simplify
(√2+ √3+ √5)²
(7-25) (5 + √7) (31 + 13√5)
8. Simplify
(6-27) (3 + √5) (11 + 4√7)
2√3 ( 1 + √3 + √6)
9. Simplify (√2 + √3) (√3 + √6) (√6 + √2)
10. Simplify
4 5 9 13 + 4√5
+
1+ √2 + √3 + 1+ √6√3 + √2 1 + 2/2 + 2√5
11. Find the difference between the sum and the product of

(1-3)(1+2√2-√3), 2-1 , 1.
√2+√3 √3
√a² + 3b2 - a + b √a² - b² + a + b
12. Simplify
a + b + √a²-b² a -b - √a² + 36²
13. Find the square roots of
(1) 5 + 2/6; (2) 55-12/21 ;
(3) + √2; (4) +√2.
14. Find the square roots of
(1) a + √a² - b² ; (2) ab + cd + √(a² - c²) (b²-d²).
15. Rationalise the denominator of the fraction
7-2/10
✓7-2/10+ 5-10
 Examples. 165

16. Find the square root of

1 + n² + √1 + n² + n².
17. If i be determined by the equation = -1, show that
( 1 + 3) + (-1-13)" is equal to 2, if n be any mul-
2 2

tiple of 3, and is equal to - 1, for all other values of n.

1+ i8
18. With the same meaning of i show that
√2 )= 1, and
give a geometrical illustration.
19. If x, y, z be quantities satisfying the equations
x + y + z + 6xyz = a,
3 (x²y + y²z + zx) = b,
3 (xy + yz + z0²) = c,
and if u have the meaning attached to it in Art. 277, show
that x + y + z = (a + b + c) ,
x + wy + w²z = (a + b + cw²) ,
x + w²y + w z = (a + bω² + cω) .
 SECTION II .- EQUATIONS.

CHAPTER X.

SIMPLE EQUATIONS INVOLVING ONE UNKNOWN QUANTITY.

284. Such a relation as x²- a² = (x + a) (x -a) may be

termed an identity ; it is true whatever values may be
given to either æ or a.
On the other hand, suppose that it were given that
2x + 5 = 7, (1)
it can easily be shown by the use of the axioms of Art. 53
that a must have the value unity. The value unity for 2
is said to satisfy the equation (1), or to be a root of the
equation (1).
285. Equations of condition such as that of the last
article occur in all the applications of Algebra to Geometry
or Physics. The discovery of the values of the unknown
quantity or quantities, which satisfy the given equation
or equations, is called the solution of the equation or
equations.
286. Equations are divided into classes according to the
number of unknown quantities involved. If there be only
one unknown quantity to be determined, one equation of
condition is sufficient; if there be two unknown quantities,
two equations will be required, and, generally, the number
of equations of condition must just equal that of the
unknown quantities to be determined.
Equations are also divided into classes according to the
highest power of the unknown quantity or quantities
Simple Equations involving one Unknown Quantity . 167
involved in them. Equations which when reduced to their
simplest form only contain terms of one dimension in the
unknown quantities are called simple equations ; those in
which one term at least is of two dimensions in the
unknown quantities are called equations of the second
degree or quadratic equations. If the term of highest
dimensions involved be of the third degree, the equation in
which it occurs is called a cubic equation or an equation of
the third degree, and so on.
The easiest equations are those of the first degree in-
volving only one unknown quantity; and the solution of
such equations is the first problem to be examined.
287. For the reduction of all classes of equations the
four axioms of Art. 53 are the main bases of operation.
They may be repeated here.
(1) If equal quantities be added to equal quantities the
sums are equal.
(2) If equal quantities be subtracted from equal quan-
tities, the remainders are equal.
(3) If equal quantities be multiplied by equal quantities,
the products are equal.
(4) If equal quantities be divided by equal quantities, the
quotients are equal.
These axioms might be all replaced by one to the effect
that if equal quantities be similarly operated on, the
results are equal ; but it is easier to take the four special
cases enunciated separately as above.
288. Take the equation ofArt. 284, namely,
2x + 5 = 7 .
By Axiom (2), subtracting 5 from the equal quantities,
it follows that 2x7-5 = 2 .

Hence, byAxiom (4), dividing these equals by 2 we have

x = 1,
which gives the only root of the equation.
168 Simple Equations [289.
The student should observe that the process of solution
is a strict process of syllogistic demonstration, as rigorous
as the propositions ofEuclid.
289. As another instance take the equation
x- 2 x- 3 x+ 1 x+ 5
5 4 2 3

The two expressions which are given equal will, by

Axioms (3), be still equal when multiplied by any the
same quantity. Let them both be multiplied by 60, which
is chosen because it is the Least Common Multiple of the
denominators of all the fractions occurring in the equation.
We thus obtain

60 (x - 2) 60 (x - 3 ) 60 (x + 1 )
5 4
= 60( +1)
2
60 (x + 5) ,
3

or, simplifying the apparently fractional forms by removing

common factors from their numerators and denominators,
12 (x - 2)-15 (x - 3) = 30 (x + 1) -20 (x + 5),
or, multiplying out and removing the brackets,
12x - 24-15x + 45 = 30x + 30-20x - 100 ;

and collecting terms on each side,

-3x + 21 = 10x - 70 .

The object now to be attained is to get all the terms

which involve a on one side of the equation, while those
which do not contain æ, that is the merely numerical
terms, shall go on to the other. This can be done by first
adding 70 to both the equal quantities (Axiom 1), which
gives 70-3x + 21 = 10x,
and then by adding 3x to these equals, whence is obtained
70 + 21 = 10x + 3x,
or 91 = 13x.
291.] involving one Unknown Quantity. 169

Then by Axiom (4), if these equals be divided by the

coefficient of x, the quotients will be equal, or
7 = x,
which of course may be written
= 7.

290. The last example has illustrated the general

method of proceeding. The two expressions which are
given equal must be multiplied by such a number as shall
cause all the fractional forms to disappear. The least
number that will effect this object is the L. C. M. of the
denominators of all the fractions occurring on both sides of
the equation. By Axiom (3) the resulting expressions are
equal.
When the fractions have thus been got rid of, the
expressions on the two sides must be simplified, and the
equation will always assume the form
ax + b = cx + d ; (a)
where a, b, c, d are numbers.
By Axiom (2), subtracting cæ and 6 from both of these
equal quantities, it follows that
ax-cx = d-6, (β)
or
(a - c) x = d - b ; (2)
and therefore, by Axiom (4), dividing these equals by
(a-c), it follows that
d-b
=
a-c
(δ)
291. In the deduction of (3) from (a) it is seen that the
terms cæ and 6 which in (a) occur on the right and left
hands respectively with a + sign, occur in (3) on the left
and right hand respectively with a - sign. It may
happen that one of the numbers as b is really negative, so
that (a) should be written
ах-b = cx + d.
170 Simple Equations [292 .
In this case b must be added to both sides, and the result
corresponding to (3) is
ax-cx = d + b ;
and the term b which was on the left-hand side with a
sign-, appears on the right hand with a sign + . Hence
a general rule arises from the use of Axioms (1) and (2),
that any term may be taken from one side of an equation
to the other provided its sign be changed.
As a special case a term which occurs on both sides with
the same sign may be removed from both.
292. Examples may occur in which the fractional forms
involve the unknown quantity in their denominators.
The method already given (Art. 290) still applies, provided
the L. C. M. of the denominators be taken to include both
numerical and literal factors (Art. 193). As an example,
take the equation
x+ 1 x- 1 17-02
2 (x - 1 ) x+ 1 2 (x² - 1 )
Here the L. C. M. of all the denominators is 2 (x² - 1), and
multiplying the two equal quantities by this L. C. M. we
obtain, after reducing the fractional forms to integral ones,
(x + 1)2-2 (x- 1 )² = 17 - x²,
or, multiplying out and removing the brackets,
x² + 2x + 1-2x² + 4x- 2 = 17 -x²,
or 6x - x - 1 = 17 - x² .
Adding x² + 1 to each of these equal quantities, there results
6x = 17+ 1 = 18,
and dividing these equal quantities by 6,
x = 3.

293. It is sometimes desirable to simplify the expressions

on the two sides of the equation before proceeding to get
rid of the fractions by multiplication ; or occasionally, by
transposing some of the fractional forms from one side of
 294.] involving one Unknown Quantity. 171

the equation to the other and combining two or more of

them together, a considerable reduction may be effected
and some long operations avoided.
294. As an instance take the equation
1 1 1. 1

2-1 -2 x- 3 x- 4

The L. C. M. of the four denominators is

(x - 1) (x - 2) (x - 3 ) (x -4),
and if the two equal expressions were at once multiplied by
this L. C. M., the reduction of the equation would involve
the multiplication of three binomial factors for each original
fraction. By combining the two fractions on each side
into one, before proceeding to multiply, a much simpler
process is required. The equation thus becomes (Art. 205)
(x -2) (x - 1) (x - 4) - (x - 3)
= ,
(a)
(x-1) (x-2) (x-3) (x-4)
-1 -1
or
(β)
(x -1 ) (x- 2) (x - 3) (x - 4)
Now, multiplying these equals by the L. C. M. of their
denominators, we obtain
-(x-1)(x-2)(x-3)(x-4)
(x- 1) (x-2)
- -(x- 1)(x-2)(x-3)(2-4)
(x - 3) (x- 4)
, (γ)
or
-(x -3) (x - 4) = - (x - 1) (x - 2) ; (8)
whence multiplying out,
-x² + 7x - 12 = -x² + 3x - 2 ;
and transposing terms, as before, the two terms - 22
disappear and we get
7x- 3x = 12-2 ,
or 4x = 10 ;
and by Axiom (4),
10 5
=
4 2
172 Simple Equations [295.
295. The student will probably be able to pass at once
from the form (3) to the form (8) without writing down the
intermediate form (y). It is sometimes also more con-
venient to change the sign when both equal quantities are
apparently negative. Thus if a = b, it follows, either
by multiplying the equals by ( -1) (Axiom 3), or by
adding a + b to them both (Axiom 1), that a = b.
296. As another example, take the equation
11 + 12 7x+ 4 3 (x ++)
=

11 (2x + 1 ) -2 (4x + 1) 4 ( 2x + 1)
Here the L. C. M. of the denominators is
44 (2x + 1) ( 4x + 1).
Before multiplying the given equals by this, it is best to
combine together the two fractions which have the literal
factor in their denominators the same. This is to be done

by adding
3 (x + 11) to both sides, or (Art. 291) transposing
4 (2x + 1)
that fraction from the right hand to the left hand, and
changing its sign. This gives
11x + 12 3 (x + 11) =
7x+ 4
+
11 (2x + 1 ) 4 (2x + 1) -2 ( 4x + 1)
Combining the two fractions on the left-hand side into
one, with the denominator 44 (2x + 1), the expression
becomes
4 ( 11 x + 12) +33 (x + 1 )
44 (2x+ 1)
44x + 48 + 33x + 18
=

44 (2x + 1)
77x + 66 7x + 6
=

44 (2x + 1) = 4 (2x + 1)' by (13) of Art. 200.

Hence the equation becomes
7x + 6 7x+ 4

4 (2x + 1) -2 (4x + 1)
298.] involving one Unknown Quantity. 173

Now multiplying these equals by 4 (2x + 1) (4x + 1 ),

which is the L. C. M. of these denominators, we obtain
(4x + 1) (7x + 6) = 2 ( 2x + 1) (7x + 4) ;
or, multiplying out,
28x² + 31x + 6 = 28x² + 30x +8 .
Whence 31x - 30x = 8-6,
or x = 2.

297. If the coefficients of the terms be represented by

letters instead of definite numbers, the principles of solution
are the same.
For instance, let the equation be
х-а -26 x- 2c
+ + = 1.
2a 36 4c

The L. C. M. of the denominators is 12abc. Multiplying

the two equal quantities by this, it follows from Axiom 3
that the products are equal, or
6bc (x - a) + 4ac (x - 26) + 3ab (x - 2c) = 12 abc,
where the fractional forms have been reduced to integral ones.
Multiplying out, and collecting all the terms containing
æ into one term, as in Art. 129, we obtain
(6bc + 4ac + 3ab) x - 20 abc = 12 abc ;
adding 20 abc to both sides,
(6bc +4 ac + 3ab) x = 32abc ;
and dividing these equal quantities by the coefficient of x,
32 abc
X
6 bc +4 ac + 3 ab
298. Considerable reductions can occasionally be made
in equations by which long operations of multiplication can
be avoided. For instance, take the equation
x³ + 7 x² + 24x + 30 2x3 + 11x² + 36x + 45
x² + 5x + 13 2x² + 7x + 20
174 Simple Equations [299.
If the fractions be at once got rid of by multiplication,
it will be found that the coefficients of all the powers of x
above the first are the same on both sides, and the equation
thus reduces to a simple equation. The process is however
long, and the reduction may be more easily effected as
follows.
By division we obtain
x+ 4 2x + 5
x+ 2 +
x² + 5x + 13 = x + 2 + 2x² +7 +20'
x+4 2x + 5
whence
x² + 5x + 13 2x² + 7x + 20
x² + 5x + 13 2x² + 7x + 20
Therefore = ,

x+ 4 2x+ 5

9 15
whence x+ 1 + =x+1 +
x+ 4 2x+ 5'
9 15
and
x+ 4 2x+ 5
By multiplication
18x + 45 = 15x + 60 ,
or 3x = 15,
so that X = 5.

Any simple equation with one unknown quantity a can

thus, by the use of axioms (1), (2), (3), be reduced to the
form ax = b ;
whence, by Axiom (4), it follows that
し
X
a

299. A large number of Arithmetical questions can be

easily solved by the help of a simple equation. In these
cases the chief difficulty is the statement of the question
in Algebraical language. A few illustrative examples
follow.
302.] involving one Unknown Quantity. 175

300. A sum of £550 is to be divided between three

persons A, B and C, so that for every pound sterling that
A gets, B shall have two pounds and C fifty shillings.
In all these so-called problems it is necessary to take
a symbol as a to represent some unknown number which is
required to be found. Here it will be convenient to assume
x = number ofpounds sterling that A has.
Then 2x = number of pounds sterling that B has,
5x =
and number ofpounds sterling that Chas.
2

5x
Hence x + 2x + 2
= total number of pounds = 550 .
Multiplying these equal quantities by 2 to get rid
of fractions, it follows that
2x + 4x + 5x = 1100,
or 11x = 1100.
Therefore x = 100 .

Hence A receives £100, B receives £200 , and C £250.

301. Students should carefully notice that a must
represent a definite number, and in their solutions of problems
must carefully state of what it is assumed to be the number.
Thus in solving the above problem beginners very commonly
state,
Let x = A's money,
or
Let x = what A gets,
without explaining whether the money is reckoned in
pounds or shillings or pence, or, in the latter statement,
that it is money of any kind.
302. As another instance take the problem :-
Two persons A and B working together can do a
certain work in 4 days. A working with a different
companion C can do it in 3 days, while B and C work-
ing together do it in 54 days How many days will A, B
and C each take to do it alone ?
176 Simple Equations [302.
Let x = the number of days A takes to do the whole
work alone.
That is, in a days A can do the whole work.
1
Therefore in 1 dayA can do X
of the whole work.
4
Therefore in 4 days A can do X
of the whole work.
But in 4 days A and B together do the whole.
Hence in 4 days B does ( 1 of the whole.

Therefore in 1 day Bdoes (1-1) of the whole. (a)

3
Again, in 3 days A does 25 of the whole work.
X

33
Therefore in 3 days Cdoes (1-38) of the whole.X

Therefore in 1 day Cdoes 3 (1-0)1) of the whole ;

5
or, as it may be reduced, 1s (1-8) of the whole.
18 (β)

Hence in 5 days, that is in days, B does

36 1
7 (1 4) of the whole,
36 5
18
and in 54 days Cdoes (1- 5x) of the whole.
But in 5 days B and C together do the whole work.
36
Hence 7
(1 )+ (1-1 ) = 1 ,
whence & can be found.
Reducing the compound fractions, the equation can be
written
10 18
(1 )+ 7
(1-1) = 1,
303.] involving one Unknown Quantity. 177

or, multiplying both sides by 7 and getting rid of brackets,

36 36
9 +10 =7;
х

and transposing terms

36 36
9 + 10-7 = + ,

72
or 12 = •

Therefore 12x = 72,

whence x = 6.

And from (a) it appears that B does of the work in

one day ; that is, B will take 12 days to do the whole, while
from (3) it similarly appears that C will take 9 days.
303. At what time between one and two o'clock will
the hour and minute hands of a clock be exactly in opposite
directions to each other ?
It is evident that at one o'clock the minute-hand is
exactly five minute-spaces behind the hour-hand.
Let x be the number of minutes after one o'clock when the
required event happens. Then a is the number of minute-
spaces over which the end of the minute-hand has
moved. Hence, since the minute-hand moves over sixty
spaces while the hour-hand moves over five, the hour-hand
X
will have moved over minute-spaces in this time, and
12

consequently be directed at the instant to a mark at the

distance (5 + 12) minute-spaces from the mark for 12
o'clock. But it is now exactly 30 minute-spaces behind
the minute-hand.

Hence
x = (5 + 1) +30
X
=
= 35+
12

N
178 Simple Equations [304.
Therefore 12 x = 420 + x,
or 11 x = 420 .
420
2
Therefore x= 38 11
11

304. Two travellers A and B set out from two places

and travel in opposite directions along the road joining
them. If c be the number of miles between the places, and
a and b the number of miles which A and B respectively
walk per hour, find how far each will have travelled when
they meet, supposing A to start two hours before B.
In problems of this kind it is sometimes useful to assume
x to represent not the number actually required but some
other number from which the one sought can be easily
found. Here, for instance, a good assumption is :-
Let x = the number of hours A has travelled before
he meets B.
Then x- 2 = the number of hours B will travel before
he meets A, since B starts 2 hours later
than A.
Hence, since A travels a miles in one hour, he travels xa
miles in a hours. Similarly B travels (x - 2)6 miles in
(x-2) hours. Altogether A and B have travelled the
whole distance e between the two places.
Therefore ха + (x - 2) b = c,
or
(a + b)x - 26 = c,
or
(a + b)x = 26 + c.
26 + c
Whence x=
a+b
Hence x is the number of hours A travels, and therefore xa,
the number of miles A travels, is a (2b + c) . The number
a+ b

of miles which B travels is (x - 2) 6, that is, b(c- 2a),which

a+ b
can also be obtained by subtracting from e the number of
miles A travels.
305.] involving one Unknown Quantity. 179

305. Examples of such problems can be multiplied ad

libitum; the foregoing specimens will furnish some hints :
for the rest the student must rely on his own ingenuity,
sharpened by practice. The first thing necessary is to
obtain a very clear understanding of the problem itself, and
to recognize the one thing unknown, a knowledge of which
is the key to all. The numerical value of this quantity is
generally the best assumption for x. The next thing is to
have a distinct appreciation of the fact that the solution
of the problem is a strict process of reasoning, and that the
methods of algebra are not a sort of conjuror's box into
which symbols may be thrown at random with a certainty
of their producing any result which the spectators may
desire.

EXAMPLES .

Solve the equations :

1.2x = 5-3x.

2.7x = 4 + 3x.
3..5x+ .25x = 1.5.
7x-5 x+ 3
4. = 2+
3 5

7x
5. (x-9) - (5 -x) = -2
x+ 5
6. (x + 1) - =
= 1.
1
7

7. 50-24 + 0-2 - 40-4.

3 7 2

8. 3-5 + 50-8-
6 4

x+ 3 7x-2 5x- 1
9. + =
+ 10x+8) .
4 5 { 2 9

N2
180
Simple Equations
7x + 59x + 4 11x + 1 7- x
10. +
12 13 4 6

6x + 72x+ 5 8x+ 1
11. + = 3-
13 7 9

7x+ 6
12.
13
+ 50+ 4 = 3- 4x-
9 3
1

1
13. ++
x- 1 x- 3 x+ 2 x- 6

1
14. + = +
x- 5 x- 3

8
15. +++ ) - (3-2x)
6

= 4-14 + 0+ 10+ 40+4. 3 10 6

x- 1 2
16.
2
+ x- 1 = 3-5 +13 6

17. -a
b
- a²+b²_w-b. a

3x + 1 7x+ 2 5
=
18.
2x- 1 8x + 3

19.
(1 - x) ( 1-2x) +
( 1 + x) ( 1 + 3x)
2 3

=
(1 + 2x) ( 3x − 1) +
( 1 + x) ( 4x - 1)
6 4

3x
20. +
4x+ 1 422 = 1 .
8x+ 7 2x
21. = 2
5x+ 4 5x+ 1

x+ 1 x+1) -2 x+3)
22. *+
* - { - 1} =
3 2 5 - {* -*+3} +2

x + 4a + b 4x + a + 2b
23 . + = 5.
x+ a + b x+ a- b
 involving one Unknown Quantity. 181

24. (x + b + c) (x + c + a) (x + a + b) = (x + 2a) (x + 2 ) (x + 2c) .

25.
3x² + 12x² + 44x² + 185x² + 8x + 98 3x² + 44x + 2
3x² + 18x² + 26x² + 15x + 14 3x² + 6x + 2
x² + 2x- 2 x²- 2x- 2 2x² -6x+ 2
26. + =

x- 1 x+ 1 x- 3

x² + 10x + 38 x² + 65x +43 5x3 + 46x² + 147x+ 163

27.
x² + 6x² + 13x + 10 5x² + 26x + 38
x+ 2a (x+ a)2
28. =

x-26 (x-b)2
4x² + 4x² + 8x + 1 2x² + 2x+ 1
29. =

2x² + 2x+ 3 x+ 1
ac bc
30. = a + b.
bx ax

31. From a place A a messenger goes to a place B, 21 miles

distant from A, and immediately returns, going at the rate of
4 miles an hour; and simultaneously with the messenger's
departure from A, another messenger starts from B at the
rate of three miles an hour, goes to A and immediately returns ;
find the distance between the two points at which they cross one
another.

32. A farmer bought equal numbers of two kinds of sheep,

one at £3 each, the other at £4 each. If he had expended
his money equally in the two kinds, he would have had two
sheep more than he did. How many did he buy ?
33. Two persons A and B divide equally a sum of money
consisting of half-crowns, shillings, and sixpences, the values of
the several parts being respectively in the ratio of 15, 4, 1.
It is found that each has 60 coins. What was the sum ?

34. A cask A contains 10 gallons of wine and 4 gallons of

water. A second cask B contains 12 gallons of wine and 2
gallons of water. How many gallons must be taken out of
each so that the mixture may contain 11 gallons of wine and
3 gallons of water ?
182 Simple Equations.

35. At what time between noon and one o'clock are the two
hands of a clock exactly opposite to each other ?
36. Find two numbers whose sum is 18, such that when
one is divided by 4 and the other by 2 the sum of the quotients
may be 6.
37. A certain fraction is equal to : when its numerator is
increased by 5 and its denominator by 9 it becomes . Find
the fraction.

38. The sum of the fourth, third and twelfth parts of a certain
number when subtracted from the number leave a remainder
which exceeds the fourth part by 5. Find the number.
39. Find a number whose third part exceeds its fourth part
by 5.
40. A sum of £ 10 was distributed in prizes in a mixed school
of boys and girls. Three times the sum given to the girls
was equal to twice that given to boys. How much was given
to each ?

41. A person invests £35000 partly at 4 per cent. and partly

at 3 per cent. The income derived from the two parts is equal.
How much is invested in each ?
 CHAPTER XI .

SIMPLE EQUATIONS . TWO AND THREE UNKNOWN

QUANTITIES .

306. The next problem to be considered is that of the

solution of two simple equations involving two unknown
quantities. An example will best elucidate the process.
307. Let the unknown quantities be denoted by x and y,
and let the equations be
4x- 2 4y- 5x_x +y
=

3 2 5
; (a)
3x + 1 2х-у =
2у-х
7 2 8
(β)
The first process is to reduce each equation separately,
by methods similar to those of Art. 288 and onwards, so
that on one side there shall only be terms involving æ and
y, and on the other only a numerical term.
Multiplying (a) by 30, the L. C. M. of the denominators
of the fractions involved, it becomes
10 (4-2) -15 (4y - 5x) = 6 (x + y),
or
40 x - 20 - 60y + 75x = 6x + 6y,
or, transposing terms (Art. 291),
40x + 75x - 6x - 60y - 6y = 20 ;
and collecting like terms
109x- 66y = 20 . (7)
Treating (3) in a similar way, it becomes
8 ( 3x + 1 ) -28 (2x −y) = 7 ( 2y -х),
or
24x + 8-56x + 28y = 14у- 7x,
184 Simple Equations. [308.
or 24x - 56x + 7x + 28y - 14y = 8,
or
-25x + 14y = -8,
or, changing the signs of both sides (Art. 295),
25x- 14y = 8 . (8)
308. Any two simple equations in two unknowns & and
y can always be reduced by the processes of simplification
explained in the last chapter to the form (y) and (8). This
form may be represented generally by the equations
a₁x + b₁y = C1 ,
a2x + b2y = C2 ;
where the suffixes 1, 2 do not indicate any relation between
the magnitudes of the letters a1 , a2, and so on, but merely
that a letter with a suffix 1 occurs in the first equation,
a letter with a suffix 2 in the second (compare Art. 164) :
while either quantity denoted by a is the coefficient of the
unknown æ, either denoted by b is the coefficient of y, and
either e denotes the numerical term on the other side of
the equation.
309. Taking now the reduced equations (y) and (8),
109x - 66y = 20,
25x - 14y = 8 ,
the next process is to deduce from these a single equation
involving only one of the unknown quantities as a.
The various methods of doing this are known as methods
of eliminating y between the two above equations. Three
methods are usually described.
310. The first of these consists in multiplying the two
given equations by such numbers as shall make the coeffi-
cients of y the same in both. Thus, if both sides of the
first equation be multiplied by 7, and both sides of the
second by 33, by Axiom (3) it follows that
763x - 462y = 140,
and 825x- 462y = 264 .
313.] Two and Three Unknown Quantities. 185

If, to use the common phrase, the upper of these equa-

tions be subtracted from the lower, or more strictly if the
left-hand side of the upper be subtracted from that of the
lower, and the right-hand side of the upper from that of
the lower, by Axiom (2) the remainders must be equal, or
(825x - 462y) - ( 763x - 462y) = 264-140 ;
that is 62x124.

This equation involving only one unknown quantity a,

by Axiom (4), gives
2.

311. Taking the value x = 2, and considering the second

equation
25x - 14y = 8 ;
it follows that
50-14y = 8 ;
therefore 50-814y ; (Art. 291)
or 42 = 14y.
Therefore y = 3. Axiom (4)
Hence the original equations are completely solved.
312. A second method of eliminating y is, by the methods
of the last chapter, to deduce from either of the given
equations a value ofy in terms of a and known quantities.
When this value of y is substituted in the other equation,
there results an equation involving a only, from which the
latter can be found. When æ is known, y can be deter-
mined by substituting for a in the expression found for y.
313. Thus the second of the given equations is
25x - 14y = 8 .
Transposing terms, it follows that
25x- 8 = 14y ;
250-8
and by Axiom (4), y=
14

But the first equation gives

109x - 66y = 20 .
186 Simple Equations. [314 .

Whence 109 x
66 (25x- 8) = 20 ,
14

or 109x
33 ( 25-8 ) = 20 .
7

Multiplying both sides of this equation by 7, it follows

that
763x- (825x - 264) = 140,
or -62x + 264 = 140 .
Therefore 62x = 124,
or x = 2, as before.
25x- 8
Also, since y = 14 ,
and a has the value 2, it follows
50-8 42
that y = == 3 .
14

314. A third method of eliminating yconsists in finding

a value for y in terms of æ and given quantities from each
of the equations : by equating these two forms an equation
is obtained involving a alone
Thus from the second of the two equations it has been
shown to follow that
25x- 8
y=
14

Similarly, from the first,

109x - 66y = 20,
it follows by transposition that
109x - 20 = 66y,
and therefore by Axiom (4) that
109-20
y= 66

Hence, since things which are equal to the same thing

are equal to one another, it follows that
250-8 109x- 20
,
14 66

an equation for determining a.

318.] Two and Three Unknown Quantities. 187

315. Multiplying both sides of this last equation by

66 x 7, the L. C. M. of the denominators of the fractions
involved, it follows that
33 (25-8 ) = 7 ( 109x - 20),
or 825x- 264 = 763x - 140 ;
which, by transposition and reducing, leads to the old
result
62x = 124,
or x = 2,
and then, as before, it will follow, taking either value of y
in terms of æ, that y = 3 .
316. In this example the student will notice that all
three methods lead to the same result. If a student trying
different methods on the same pair of equations should
chance on differing results, he may be pretty sure that there
is a mistake in the working of one method or the other,
possibly, not to say probably, of both.
It may be also noticed that not only are the final results
the same, but a great many of the steps in all three pro-
cesses are also identical. In each method the products of
25 and 33, 7 and 109, 33 and 8, 7 and 20 have to be found.
Thus as a rule the three methods are about equally long,
and in nine cases out of ten the student will find it best to
keep to the first method. The second and third are chiefly
useful when the coefficient of y in either or both of the
equations is either unity or a very small integer.
317. It hardly needs to be pointed out that the equations
can be solved with equal ease by eliminating & in the first
instance and so obtaining an equation containing y alone.
It is usually most convenient to eliminate that one of the
unknowns whose coefficients are the smallest numbers .
318. As another instance, let the two equations be
5x + 4y = 14 ;
17x - 3y = 31 .
188 Simple Equations. [319.
Using the first method, and multiplying the first pair of
equal quantities by 3, and the second pair by 4, it follows
that 15x + 12y = 42 ,
68x- 12y = 124 .
Here the coefficients of y in the two equations are nu-
merically equal but have opposite signs. When this is the
case the elimination of y is effected, not by subtracting each
of one pair of equal quantities from one of the other as
in Art. 310, but by adding them together. This process
gives
(15x + 12y) + (68x - 12y) = 42 +124,
or 83x = 166 ,
whence x = 2.

Substituting this value in the first of the two given

equations, it follows that
10+ 4y = 14 .
Therefore 4y = 4 ;
or
y = 1.
Hence the equations are completely solved.
319. Examples may occur more complicated and difficult
than those which have been solved, but the principles which
underlie their solution will be the same. Each equation
must in ordinary cases be separately reduced to its simplest
form by the methods of the last chapter. (Art. 308.) One
of the unknown quantities must then be eliminated by one
of the methods explained in Arts. 310-315, and the value
of the other determined from the resulting equation.
Either of the given equations can then be used to deter-
mine the value of the remaining unknown.
It may be repeated that, as in the case of a single equa-
tion with one unknown, the process of solving a pair of
simultaneous equations is a strict process of reasoning,
leading, by no accident, but by a rigidly directed purpose,
to the values of the unknowns.
321.] Two and Three Unknown Quantities. 189

320. Occasionally instances occur in which the processes

of solution may be abridged by somewhat deviating from
the strict order of the methods previously indicated.
Suppose, for instance, the equations to be solved are
+
2y
= 1 ;
a

20 y
+ = 1;
a ៖
the methods previously used would suggest the multiplica-
tion ofboth equations by ab, so as to get rid of fractions, as
a preliminary to elimination by either of the three methods.
The following is a shorter process of solution. From the
two given equations it follows, by Axiom (2), that
2y
( + )-( + )=
= 1-1 ,
X
or
b
= 0,
a

X
whence =

b a

Hence from the first equation it follows that

20
+ = 1,
a a

3x
or = 1,
a

or
3x = a, multiplying the equals by a,
a
or =

3
,
dividing the equals by 3.
0 b
And, since b = , it is easily shown that y = 3

321. Similar principles govern the solution of a set of

three simple equations containing three unknown quantities.
Each equation must be reduced, by getting rid of fractions
and transposing terms so that on one side shall be only
terms containing the three unknowns, and on the other
190 Simple Equations. [322.
only known numbers or quantities. If the three unknowns
be denoted by x, y, z, the three equations in their reduced
forms can be represented by
a₁ x + by + c₁ z = d₁,
a2 x + b2 y + C2 z = d₂,
A3 x + b3 Y + C3 2 = d3 ;
where a1 , 61 , C1, d₁ , &c. represent known numbers, the
notation being in every respect similar to that explained
inArt. 308.
322. When the three equations have been reduced to
this simplified form, one of the unknowns, as z, must be
eliminated between any two of the equations by one of
the methods of Arts. 310-315. There will thus result an
equation involving only x and y of the form
px + qy = r, (a)
where p, q, r are numbers.
Let another pair of the original three equations be taken,
and z be eliminated between these. There will be obtained
another equation involving only x and y of the form
p'x+q'y = r', (β)
where p', q', r' are known numbers.
From (a) and (3) x and y can be found as in the early
part of this chapter. The third unknown, z, can be ascer-
tained by substituting these values of x and y in either of
the three given equations.
323. As an instance, let the three equations be
5x + 7y + 3z = 28 ; (1)
7x- y + 11z = 38 ; (2)
11x + 6y- 42 = 11 . (3)
Multiplying ( 1) by 11 and (2) by 3, there results
55x + 77y + 33z = 308 ,
21x - 3y + 33z = 114 ;
subtracting the lower of these pairs of equal quantities each
324.] Two and Three Unknown Quantities. 191

from the corresponding upper quantity, z disappears and

we obtain 34x + 80y = 194 ;
and, dividing each of these equal quantities by 2, it follows
from Axiom (4) that
17x + 40y = 97 . (4)
Again, multiplying (1) by 4 and (3) by 3,
20x + 28y + 122 = 112 ,
33x + 18y- 122 = 33 .
And, adding equals to equals, z disappears and there results
53x + 46y = 145 . (5)
From (4) and (5) x and y must be determined. Multi-
plying (4) by 23 and (5) by 20 the coefficients of y will
become the same, and they give
391x + 920y = 2231 ;
1060x + 920y = 2900 .
Subtracting the upper pair of equals from the lower, it
follows that
669x = 669,
thus x = 1.

Hence, from (4),

17 + 40y = 97 ;
therefore 40y = 80 ;
or
y = 2.
Substituting these values ofx and y in ( 1), that equation
becomes 5 + 14 + 3z = 28 ;
therefore 3228-5-14 = 9,
or z= 3 .

Hence the equations are completely solved.

324. Arithmetical problems of considerable intricacy can
be solved by the method of this chapter. As an instance,
the problem of Art. 302 can be more expeditiously solved
by assuming three unknown quantities, instead of one, as
was done in that article.
192 Simple Equations. [325
Two persons, A and B, working together, can do a certain
work in 4 days. A working with a different companion,
C, can do it in 3 days, while B and C working together
do it in 5 days. How many days will A, B, and C each
take to do it alone ?
Let x = the number ofdays whichA takes to do the whole
work alone.
y= ‫دو‬ ‫وو‬ ‫دو‬ ‫دو‬
B ‫وو‬ ‫دو‬

2= ‫وو‬ ‫دو‬ ‫دو‬ ‫دو‬

C ‫دو‬ ‫دو‬ ‫دو‬

1
Hence in one day A does X
of the whole work,
1
" ‫دو‬ ‫دو‬
B ‫وو‬ ‫دو‬ ‫دو‬ ‫دو‬

Therefore in oneday AandB together do ( + ) of the

whole. But as they take 4 days to do the whole, in one
day they must do of the work.
1 1 1
Therefore + (1)
X y 4

By similar reasoning
1 5
+ = 23
=
18 (2)
1 1 1 7
+
2 57 36
(3)
y

325. From these three equations x, y, z may be found.

Their solution is best obtained without getting rid of
fractions, as it may be noticed that x, y, z only occur in the
111
particular forms -, -, 2
-, and it is easier to find the values
ху
of these latter quantities and then to deduce those of
x, y, z.
It will be noticed that one of the equations (1) only
1 1 1
involves and Between the other two must be
2
y
326.] Two and Three Unknown Quantities. 193
1
eliminated. The coefficient of being the same in both,
2

this elimination is effected by a simple process of subtrac-

tion, from which there results
1 1 3 1

0 36 12 (4)
y
1 1 1
But by ( 1) +
4
y

Adding these equations, there results

2 1 1 4 1
+ =
X 12 4 12 3
whence 6 = x, multiplying by 3x,
or x = 6.

Again, subtracting (4) from (1),

2 1 1 2 1
,

y 4 12 12 6

whence y = 12.
And then from (2) it follows that
1 5 1 5 1 2 1
=

2 18 X 18 6 18 9

Therefore z = 9.

326. In attempting the solution of equations such as

those in Art. 324 students often fall victims to a devo-
tion to the letter of the rule which blinds them to the
spirit of the principle underlying the rule. They proceed
to get rid of the fractional forms and thereby introduce
beside x, y and z, the products xy, xz, yz. In strictness
the equations (1), (2), (3) are not simple equations (Art.
286) in regard to x, y, z, but they are so in regard to
111
х 2
And the values of the latter quantities are
y

therefore the best objects for research.

0
194 Simple Equations. [327.
327. As another example we may take the following
problem.
There is a number of three digits whose sum is 11 .
The sum of the first and last digits is less by unity than
twice the middle digit. If the first and second digits be
reversed the number is diminished by 90. Find the three
digits.
If x, y, z represent the digits in the units, tens and
hundreds place respectively, the number is represented by
100z + 10y + x. The three given conditions expressed in
terms of x, y, z become
x + y + z = 11 , (1)
x + z = 2y- 1 , (2)
(100y + 10z + x) = ( 100z + 10y + x) -90.
The last of these can be reduced by transposition of terms
to 90-90-90,
whence y- z = -1 . (3)
Replacing a + z in (1) by its value 2y- 1 derived from
(2), it follows that
(2y - 1) + y = 11 ,
or
3y = 12 ;
thus y = 4.
Then from (3) 4-2-1 .
Therefore zz = !5 .

And then, from (2), x = 2y - 1 -2 = 8 - 1 - 5 = 2 .

The number in question is therefore 542 .
328. Two persons, A and B, set off together to walk
from Cambridge to Ely. When A has gone 4 miles, he is
obliged to stop for refreshment. B being in haste goes on,
and walks faster that he did when with A. After half an
hour A goes on, but walks as much slower than his former
pace as Bdoes faster, and arrives at Ely two hours and six
minutes after B, and an hour and a half later than they
328.] Two and Three Unknown Quantities. 195

would both have arrived had they gone on together at

their first pace. B arrives twelve minutes later than he
would have done had he gone at the faster pace all the
way. Find the distance from Cambridge to Ely and the
rates at which A and B walked.
Let x= the number of miles walked per hour at the first
pace.
x + y = the number of miles walkedper hour by Balone.
Then x-y = ‫دو‬ ‫دو‬ ‫دو‬
A‫دو‬ ‫دو‬ ‫دو‬ ‫دو‬

Let z = the number of miles between Cambridge and

Ely.
2 the number of hours in which A and B would
Hence =

X
have walked the whole distance together.
4
Snumber of hours which they take to walk
x { the first four miles.
2-4 number of hours which B takes to walk the
x+y
{( rest of the way.
2-4
= (number ofhours which A takes to walk the
х-у rest of the way.
Hence the number of hours actually occupied by B in
4 2-4
the whole journey is + x+y and the number of hours
,

he would have taken had he gone at the faster pace all

2

theway is The last condition then gives,

x+y
4 2-4 2 1
+ +
X
x+y x+y 5

since twelve minutes make a fifth part of an hour.

2

By subtracting from these two equal quantities this

x +y
equation becomes,
4 4 1

х 5
(1)
x+y
02
196 Simple Equations. [329.
Again, the whole number of hours occupied by A is
4 2-4

2
++
2
Hence the first condition gives
х-у
4 1 2-4 4 2-4

X
+
2
+
х
+ +210,
х-у x+y
4 1
or, subtracting = + 2 from these equal quantities,
X

2-4 2-4 8
+
5 (2)
x -y x+y

Again, the second condition gives

4 1 2-4 2
+ + +1 ,
2
х-у
4 1
or, subtracting X
+ 2 from both sides,
2-4 2-4
+1 . (3)
х-у х

329. The three equations (1), (2), (3) are sufficient to

determine z, x, y. Strictly speaking they are not simple
equations according to the definition of Art. 286, but, like
the equations of Art. 324, they can be solved by similar
methods.
2-4
Equating the two values of given in (2) and (3), it
х-у
follows that
2-4 8

X ++ 5'
+1 = x+y
or, transposing terms,
2-4 2-4 3
=
X x+y 5

Dividing these equal quantities by (z- 4), we get

1 1

X
x+ y = 5 (2-4)
 330.] Two and Three Unknown Quantities. 197

But from (1), dividing both sides by 4,

1 1 1
.

х
x+ y 20

3 1
Hence = ,

5 (2-4) 20
and multiplying these equals by 20 (z - 4),
122-4.
Therefore z = 16.

330. To determine x and y there are now left ( 1) or

1 1 1

x+ y 20'

and the result of substituting for z its value 16 in (3),

which easily gives
1 1 1

х-у X 12

Getting rid of fractions, the first of these becomes

20 y = x (x + y), (a)
and the second gives
12 y = x (x −y) ; (β)
whence by Axiom (4)
20y_x (x +y)=

12yx (x -y) '

5 x+ y
or reducing 3
=

х-у
Therefore 5 (x −y) = 3 (x + y),
or
5x - 5y = 3x + 3y,
and transposing 2x = 8y,
or
x = 4y. (7)
Hence 20y = 4y (4y + y) from (a).
Andtherefore, dividing by 20y, y = 1 ; and from (y) x = 4.
Hence, they start to walk at 4 miles an hour, and after
they part B walks at 5 miles an hour and A at 3 .
198 Simple Equations.

331. Cases may occur of problems involving more than

three unknown quantities. In these, if the solution be
possible, there must always be as many conditions as there
are quantities to be determined. If there be thus, for
instance, four equations involving four unknown quantities,
the process of solution will be to eliminate one of these
successively from three pairs of the original four equations,
and thus obtain three equations involving the other three
unknowns, which can be solved as before.
Care must be taken that the three pairs chosen to
eliminate from shall be independent pairs. Thus, if one
result is obtained by elimination from (1) and (2), these
denoting the numbers of the equations, a second may be
obtained from (1) and (3), and a third from either (1) and
(4), (2) and (4), or (3) and (4). It will not do to get the
third from (2) and (3), as any result obtained from them
would be derivable from those already given. The only
necessary point to observe is that each of the four given
equations must be used in one or other of the eliminations.
332. The process can be extended to the solution of five
equations with five unknowns, or to more. The only limits
are those imposed by the finite nature of the energy and
skill of the worker and the time at his disposal.

EXAMPLES .

Two Unknowns .
Find the values of x and y from the following pairs of
equations.
1. 2x + 3y = 7, 8x - 5y = 11 .
2. 5x+ 7y = 19, 7x + 4y = 15.
3. 11x- 12y = 9 , 4x + 5y = 22.
2x 5y 7x 5y =
43
4. = 3,
3 4 4 3 3
 Two and Three Unknown Quantities . 199
5. x+ 8y = 17, 7x- 3y = 1 .
7x 6y 5x 2y
6. + = 32, = 1.
6 7 4 3
2x
7. + = 3, +
2y = 3.
a b a b
2x Y 2y x
8. = = c.
a b a

9. (a + b) x + (a-b) y = 2a, (a -b) x + (a + b) y = 2a.

x
Y
10.
a+ b + a=b = 2, x + y = 2a.
7x- 5y8x + 1
11. (4x + 3y) = 4
+
12
,
16x-5y = 27 .

12.
8x- 5у+ 11y- 4x
x- 5y 17x- 13y 2x
= 7.
7
= 4, +
5 5 3

7x- 5y y - 3х7у-
5у- 3x + 5х = 6.
13. 5х-
11
3у + 12
= 8,
7 12
2a b 26 a
14. = 1, = 1.
х Y X
Y
a+ ba -b 34
15. + = 2a, -+- = 7.
x х
Y y
2 2y
16.
3y х

( + )= b a
,
х- у = a - b.

9a² - 13ab + 462 1 1

17. = + , ax-by = a² - b².
36 abxy 4by 9ax
18. ax + by = bx + ay = c.
1 1 4
19. + =
2-x 1+y 3 2 + 1+ y = 3.
2-x1

Three Unknowns.

Find the values of x, y, z in the following examples.

3x + 2y + 5z = 21 ,
20. 2x- 3y + 4z = 11 ,
x + 3y + 7z = 20 .
200 Simple Equations.

10x- 2y + 4z = 30,
21. 3x + 5y- 3z = 9,
{2x + 3y- 2z = 5.
20x + 2y + 5z = 39,
22.
{ 23x + 7y- 6z = 19,
16x + 3y- 7z = 1 .

23. + 1, + 1, + = 1.
a C a

24. cy+ bz = a, az + cx = b, bx + ay = c.
a C a
25. + 2, + 2
2, 2
+ X
= 2.
Y Y

2x3y 4z 4x 2y 3z 3x 4y 2z
= 5.
26. + =1 , + =3, +
a b C a b C a b C

4 5 9 6 10 3 8 15 12
27 . + = 0, + = 4, + = 5.
X 2 2 X 2
Y Y Y

28. Find z from the following equations in terms of a, b, c,

d, e, f, g :
ax = p - r, b ( y + z) = s - q,
су = p - s, d (z − x) = q - r,
ez = r - s, f(x + y) = q - p + g .
29. A person walks from A to B, a distance of 7½ miles, in
2 hrs . 17 minutes, and returns in 2 hrs. 20 min. His rate
of walking up-hill, down-hill, and on a level road being 3,
34, and 3 miles respectively. Find the length of level road
between A and B.

30. There are two fractions such that the fraction formed with
the sum of their numerators for numerator and the sum of their
denominators for denominator is of the greater ; and the
fraction similarly formed with the difference of the numerators
and denominators is ; also the sum of the numerators is twice
the difference of the denominators. Find the fractions.

31. Three persons, A, B, and C, can do a certain piece of work

4

in 81's days when all work together. After two days' working
 Two and Three Unknown Quantities. 201

all together C goes away, and A and B finish the work in

83days more. IfB had stayed for four days, A and C together
could have finished the rest in 8 days more. Find the number
of days in which each could do the work alone.
32. A certain number consists of two digits ; by the addition
of 27 the digits are reversed: find the number.
33. A fraction becomes by the addition of 2 to its numer-
ator and 3 to its denominator. If 2 be subtracted from the
numerator and unity from the denominator it becomes . Find
the fraction.

34. Acertain number when divided by a second has a quotient

7 and a remainder 4. If three times the first number be divided
by twice the second number, the quotient is 11 and the remainder
4. Find the numbers.

35. A cistern can be filled in 5 hours by two pipes A and B.

Both are left open for 3 hrs. and 45 min. and then A is shut,
and B takes 3 hrs. and 45 min. longer to fill the cistern. Find
inhow many hours each pipe would fill the cistern alone.
36. There is a number consisting of three digits. If the digits
be reversed the new number exceeds the old by 99. If the
number be divided by 3 the quotient is a number of two digits
the sum of which is equal to the sum of those of the first
number, and the digit in the tens' place is double that in the
tens' place of the first number. The sum of the first and last
digit exceeds the middle one by unity. Find the number.
 CHAPTER XII .

ON GENERAL FORMULAE FOR THE SOLUTION OF

SIMULTANEOUS SIMPLE EQUATIONS .

333. THE last chapter has explained the methods by

which simultaneous simple equations can be solved. In
the present chapter we have to consider certain general
formulae by which the results of that solution can be
expressed. For this purpose the typical forms given in
Arts. 308 and 321 will be employed.
334. Taking the equations of Art. 308, namely
a₁x + b₁y = C1 , (1)
a2x + b2y = C2 , (2)
the elimination of y can be best effected by the first
method. Multiplying (1) by b₂ and (2) by b₁, it follows
that
a₁b2x + b₁b2y = b₂C1 ,
a₂b₁x + b₁by = b₁02 .
And subtracting the lower pair of equals, each from the
corresponding quantity in the upper row, y disappears and
we get
(a₁b2 - a2b₁)x = 6201-61022
whence b21-6102
x =
(a)
a1b2-a201
335. Similarly multiplying (1) by a₂ and (2) by a₁, there
result
aa2x + a2b₁y = a2C1,
a₁ax + a₁b₂y = a1C2 ,
 Solution of Simultaneous Simple Equations. 203

and subtracting the upper pair of equals, each from the

corresponding quantity in the lower row, we obtain
(a₁b2 -a2b1) y = a₁C2- A2C1,
whence A1C2-a2C1
y= (β)
a1b2- a2b1
336. In any particular case the formulae (a), (β) will
give the solution of a pair of equations by substituting the
proper values for a₁ , b1 , C1 , a2 , 62, C2.
Thus taking the equations solved inArt. 318, namely
5x + 4y = 14,
17x - 3y = 31 ,
and noting that here
=
a1 5, ₁ = 4, C₁ = 14,
a₂ = 17, b₂ = −3 , C₂ = 31 ,
it follows from (a) that
-3 × 14-4 × 31 -42-124 166
х =2 ;
5x ( -3 ) -17 x 4 -15-68 83

and from (3) that

5 × 31-17 × 14 155-238 83
y= = 1.
-15-68 83
5 × ( -3) -17x4
The student will notice by comparing this process with
that of Art. 318, that not only are the results the same,
but the multiplications which have to be effected are also
identical : that in fact the method of this article merely
condenses and reduces to a rule the more diffuse operations
ofArt. 318.
337. It will be noticed that the denominators of the
values of x and y in (a) and (3) are the same, and farther
that this common denominator depends merely on the
coefficients of x and y in the two given equations.
This denominator a₁b2- a2b, is called the determinant of
204 General Formulae for the Solution of [338.
the four quantities a₁ , 01 , a2 , 62, and is usually and con-
veniently denoted by the symbol

The numerator of the value of x in (a) can similarly be

denotedby

C2362
since it can be obtained from thedenominator by replacing
each a by the corresponding c. Hence (a) becomes
C1, 61
C2, 62
x= (γ)
41, 61
a2, 62
Similarly (3) becomes, with the new notation,
A1 , C1
A2, C2
y= ; (8)
a1, 61
a2 , 62
where the numerator can be obtained from the denominator
by replacing each 6 by the corresponding c.
338. The results in the form (y) and (8) are more easy to
remember and more convenient to use than those of (a) and
(3) when once the student has mastered the definition of a
determinant, to the consideration of which we shall return
in the next article.
The denominator of the value of either & or y is obtained
by writing the four coefficients of x and y in the two
equations
a₁x + b₁ y = C1,
a2x + b2y = C2,
just in the positions in which they occur, and drawing ver-
tical lines on each side of the square thus formed to indicate
that the determinant of these four quantities is meant.
341.] Simultaneous Simple Equations. 205

The numerator of the value of æ is obtained by replacing

each a, that is, each coefficient of æ, in the denominator,
by the corresponding e or numerical term.
The numerator of the value of y is similarly obtained
by replacing each 6, that is, each coefficient of y that occurs
in the denominator, by the corresponding e, the numerical
term in the same equation as the b belongs to.
339. The determinant of the four quantities a₁ , 61 , a2 , 62 ,
a , has been defined as being a₁b2 - a2b1.
written asa, 2

The four quantities a₁ , 61 , a2 , b₂ are called the constituents of

thedeterminant. Those constituents, as a₁ , b₁ , which form a
horizontal line are called a row ; while those, as a1 , a2 , which
form a vertical line are called a column of the determinant.
A determinant of four constituents is called a determinant
of the second order .
340. Determinants can be formed of nine, sixteen or any
square number of quantities, and are called determinants of
the third, fourth, etc. order. The determinant of the nine
quantities a₁ , b1, C1 , A2, B2, C2, a3 , 63 , C3 is represented by
the symbol
α1, 61 , C1
a2, 62, C2
Az, bz , Cz

As in the last article, the quantities a₁ , 61 , C1 , A2, 62, C2,

a3, b3, C3 are called constituents of the determinant. The
assemblage of constituents which are in a horizontal line is
called a row, and the constituents which are in a vertical
line form a column.
341. The value of the determinant of any number of
quantities arranged in a square, all the quantities in each
row or in each column having the same suffix, is defined to
be the algebraical sum (Art. 93) of all the products that
can be formed by multiplying together one constituent out
206 General Formulae for the Solution of [342.
of each row and one out of each column. The first term of
the determinant is formed by multiplying together the
constituents which lie in the diagonal of the square going
from the left-hand at the top to the right-hand at the bottom .
This term is to be reckoned as positive, and all the other
products are to have a sign + or prefixed according as the
suffixes in them can be formed from those of the first term by
an even or odd number of interchanges of consecutive suffixes.
342. The definition of the last article is perfectly general
for determinants of all orders. The student will probably
understand it better by applying it to the simplest case,
that of the determinant of four quantities,
αι,

A21 b2
The first term of this determinant is the product of the
constituent in the diagonal which begins with a₁, namely
a, b . The only other possible product accordant with the
definitions of the last article is a₂ b₁ , and since the suffixes
of this term 2,1 are obtained from those of the first term 1,
2 by one interchange, this term must have a sign - prefixed.
of a1
1,
Hence the value of according to the definition
az, ba
of Art. 341 , is ab₂ - ab₁, the same as in Art. 337.
343. The following transformations of a determinant
of four quantities are easy to follow. They are chiefly
important from their analogy to similar transformations of
determinants of the third and higher orders.
a, b a1, a
=
(a)
a2, 02
for each of the determinants has the value a₁b - ab
b1, 21 αι
= b₁a2 - ba₁
201 = - (a₁b2 - a2b₁) =
1 (β)
b, a a , b2

a = a -ar = 0 ; (γ)
1,
 Simultaneous Simple Equations. 207
344.]
or a determinant of four quantities with two columns
identical vanishes.
b α1, 01
par = pa₁b₂- pa2b₁ = p (a₁b2 - a2b₁) = p a2, b2 ,
(8)
b2
a1 + b1 , C1
= (a1 + 0₁)C2-(a2 + b2) C1
a2+b2 , C2
A1, C1 61, C1
=
(a₁C2-a2C1) + (b₁₂-b20 ) = + (6 )
A2, C2 2

344. The determinant ofnine quantities, or

41 , 61 , C1
a2, 02, 02 ,

аз, 3, C3
can now have its value discovered. The first term is the
product of all the constituents in the diagonal beginning
with a₁ , that is a₁6203.
The number of terms is evidently the same as the
number of possible arrangements of the suffixes 1, 2, 3, since
each term in the determinant must only contain one factor
out of each row, that is, only one factor with each suffix,
and must also only contain one factor out of each column,
that is, one factor denoted by each of the letters a, b, c.
Hence a₁ can only occur when multiplied by b₂03 or b3 C2, A2
when multiplied by bac₁ or b₁C3 , and a3 can only occur in
combination with b₁₂ or b₂01. There are thus six terms
altogether whose numerical magnitudes are
a1b23,5a₁b3C2, a2b3C1, a2b1C3, asb₁C2, a3b2C1.
The first of these has a sign + . The second is obtained
from the first by one interchange, namely of the suffixes
2 and 3. Hence the second term has the sign - . The
fourth term ab₁c3 can be also obtained by one interchange
and therefore has the sign The third term a2b3c₁ can
一 .

be obtained from the fourth by one interchange, namely

of 1 and 3 ; hence it is obtained from the first by two
interchanges, and therefore it has the sign + .
208 General Formulae for the Solution of [345.
The fifth term can be derived from the first by first
interchanging 2 and 3, thus giving a₁b302, and then
interchanging 1 and 3. Hence the term a3b1c2 has a
sign + , being obtained from a₁b2c3 by two interchanges ;
while a3bc₁ requires another interchange of the 1 and 2, and
has thus a sign - . Hence, on the whole, the determinant
is the algebraic sum of the six terms
+a1b2c3 , -a1b3c2 , + a2b3C1, -a2b₁C3 , + a3 b₁C2 , -a3b₂ C₁;
and we may therefore write
41, 61, C1
a2 , 62 , C2 =a₁₂C3 - a₁b3C2 + a2b3 C₁ - a₂b₁ C3 + a3 b₁₂-a3b2C1.

345. In Art. 341 the sign of each term has been deter-
mined by the number of interchanges of consecutive suffixes.
The word ' consecutive' may be omitted; for any two
suffixes can be interchanged by an odd number of inter-
changes of consecutive suffixes. Thus, let the order of the
suffixes in the first term of a determinant of the 9th order be
1, 2, 3, 4, 5, 6, 7, 8, 9. (a)
By interchanging the 1 with its successor 6 times, the order
becomes 2, 3, 4, 5, 6, 7, 1 , 8, 9 ; (β)
and by interchanging the 7 with its predecessor 5 times
the order becomes
7, 2, 3, 4, 5, 6, 1 , 8 , 9 , (7)
which is the same as (a) with 1 and 7 interchanged. The
whole number of consecutive interchanges is 6 + 5 = 11 ,
an odd number.
346. To interchange any two suffixes we need only, as in
the last article, interchange that one which comes first with
its successor till it is just one place in front of the other.
Let this require pinterchanges. It will clearly require one
less than this number, or p - 1 interchanges of the second
suffix with its predecessor to bring it into the position which
the first did occupy. The number of consecutive inter
347.] Simultaneous Simple Equations. 209

changes required is therefore p +p - 1 or 2p -1, which is

always an odd number.
Hence an odd number of interchanges of any kind
can always be effected by an odd number of consecutive
interchanges, and similarly an even number of the one by
an even number of the other. Hence the word ' consecutive'
in Art. 341, although useful in giving a clear idea at
first, may be omitted without altering the results.
347. From Art. 344 it follows that
a₁ b₁₁
a2 b2 C2 = a₁ (b2C3-63C2) -a₂(b₁C3-63C₁) + a3(6₁C2-62C₁)
A3 63 C3
= 62 C2 - 2
+
+a3
C3

This result, and a corresponding one for determinants of

any order, can be obtained otherwise, thus :-
The terms which involve a, cannot, by the definition of
a determinant, contain any other factor a or any other
factor with suffix unity. Hence the multipler of a₁ must
be obtained by taking all possible products of one out of
each column and one out of each row of that part of the
determinant which is formed by omitting the first row and
column. Hence the terms which multiply a₁ are the
same numerically as the terms of the determinant 62 C2
b3 C3
Also, since in all terms which involve a₁ the interchanges
of suffixes are merely those of the b's and c's, their signs will
be obtained from that of the first, by the same rule as
C2
are obtained from its first
b.2 C3
those of the determinant 63
term. And since the first term a₁b2c3 is positive in both
b C2
cases, the whole set of terms involving a₁ is a₁
63 C3
P
210 General Formulae for the Solution of [348.
b
Similarly the coefficient of a can be shown to be

and that of a to be
02 C2
348. The determinants
B2 Ca
, and similar ones formed
C3

by cancelling any row and column of the original deter-

minant, are called minors of the original determinant.
349. It is easily seen that
a b C1 A1 A2 A3
a ba C = 66
A3 3 C3 C1 C2 C3

For by working out each of these determinants the same

expression is obtained as in Art. 344.
Or it may be deduced from the fact that the first term
of each is a₁b263 , and that any other term, as abic , can be
obtained from this either by two interchanges of suffixes so
as to obtain in succession a₁ba , azbic ; or by two inter-
changes of the letters, leaving the suffixes unaltered, so as
to obtain in succession b₁ac3 , b12a3.
Thus the value of a determinant is not altered if rows be
changed into columns, and vice versâ (Art. 340).
350. If the position of any two columns be interchanged,
the determinant retains the same numerical value but
changes its sign.
Thus a1 b1 C1 Ca
abc = -

A3 3 C3 33 3

For any term in the second, as cha , can be obtained

from the first term c₁bas , by one more or one less inter-
change than is required to bring it from abc3 , the first
term of the first determinant ; since abc, is derived from
abac , or will give a₁bac3 , by one interchange. Hence every
352.] Simultaneous Simple Equations. 211

term which can be obtained from a₁ b203, by an odd number of

interchanges will require an even number to come from a3b2c1.
Thus each term in either determinant is numerically equal,
but of opposite sign, to a term in the other. Hence the two
determinants are equal in magnitude but of opposite sign.
351. From the last article it follows that a determinant
ofwhich any two columns are identical must vanish.
For if a₁ = C1 , A2 = C2 , A3 = C3 , the two determinants
a1 b1 C1 C1 61 1
a2 b2 C2 and C2 b2 a2 are also identical.
C3 63 3
But by the last article they are of equal numerical value
with opposite signs. Hence the numerical value of each
must be zero.
This can also be seen from the consideration that,
a1 b1 C1
corresponding to any term in a2 62 C2 , as a₂ 61 c3, there is

another term a3b₁c, whose sign must be opposite to that of

the first since it is obtained from it by one interchange
of suffixes. Hence if a₂ = c₂ and a3 = C3 , the algebraic
sum of these terms vanishes, and thus the whole deter-
minant must vanish.
352. The following results easily follow from the form for
ab₁₁
a2 b2 c₂ given in Art. 347 :

a1 +a b₁ C₁
,
41 61 C1 ab₁₁
a2+a2 b2 C2 =
A2 b2 C2 + a2 b2 C2 (a)
A3 63 C3 аз вз Сз
pa1 b1 C1 a1 b1 C1
pa2 b2 C2 = pa b₂ C2 (β)
раз 63 C3
P2
212 General Formulae for the Solution of [353 .
353. From Arts. 349 and 352 it follows that

a₁ +ab + bí atá ab₁₁ ai bí cí

=
a2 b2 C2 A2 b2 C2 + a2 b2 C2 ,

аз b3 C3 A3 63 C3
pa pb pc ab₁₁
a2 b2 C2 = pa2 V2 C2
A3 63 C3 A3 63 C3

354. Another easy deduction from Art. 350 and (3) of

Art. 352 is that

pa, qb₁ re₁ ab₁₁

paz qb2 rc2 = pqr a2 b2 C2 (2)
раз qb3 rc3 A3 03 C3
For, by (3) of Art. 352, the given determinant
a1 qb₁ re re₁ qb1 a1
=paz qba re₂ = -p re₂ qbz az2 by Art. 350,
a3 qb3 re3 rC3 qb3 аз
C1 gb a
=-pr c₂ qb, a (3) of Art. 352,
C3 903 A3
01 Ca
=
pr qb2 C2 a₂ Art. 350,
QV3 C3 A3
bca
=
pqr baca a 2 (3) ofArt. 352,
03 C3 A3
ab
= -pqr a₂ C2 02 Art. 350,
A3 C3 03
ab
= +pqr az b₂ Ca Art. 350.
A3 3 C3
355-] Simultaneous Simple Equations. 213

355. The theorems of the last few articles can now be

used for the formal solution of the three typical equations
ofArt. 321 , namely
a₁ x + by + c₁ z = d₁ , (a)
a2x + b2y + C2 z = d₂ , (β)
a3x + b3y + C3z = d3 . (y)
C2
and
Multiplying (a) by aC3 (8) by - 63 C3 ,

(y) by they become

62 C2 +
b2 C2 62 C2 = d022
24103 63 +201/6303 C3

-20-20-20-
C3
=
2

14 C1 61 C1 =
хаз + ZC3 d3
02 C2 62 C2
The sum of the three expressions on the left-hand sides
of these equations must, by Axiom I, be equal to the sum
of the three expressions on the right-hand side. If the
addition be performed, the coefficient of x on the left-hand
side becomes
62 C2 -

a1 a2 +a3
63 C3 63 C3 4362 C2
a1 b1 C1
but by Art. 347 this is a₂ b2 C2
A3 3 C3
The coefficient of y becomes
C2 C1
2 + ,

63 C3 63 C3 B2C2
1444
which by Art. 347 is 62 62 C2 which vanishes by Art.
351.
214 General Formulae for the Solution of [356.
C11 C1
Similarly the coefficient of zis C2 b2 C2 , which also
vanishes. C3 63 C3

The sum of the expressions on the right-hand side is

62 C2 d2 + d3 ,

63 C3 362 C2
d₁₁₁
which by Art. 347 is d₂ b₂ C2
d3 63 C3
Hence the addition finally gives
41 b1 C1 d1.61 C1
0
A2 b2 C2 | = d2 b2 C2
d3 63 C3
d₁ 61 C
d2 b2 C2
d3 b3 C3
whence a
41 61 C1 (δ)
a2 02 C2
аз 63 C3
356. Similar formulae can be obtained for the values of
yand z. The three equations (a), (β), (γ) can be written
b₁y + a₁x + c₁ z = d₁ ,
b2y + a2x + c2 z = d₂, 2

b3y + a3x + C3z = d3.

Whence, by a process identical in form with that of the
last article,
d1 a1 C1
d2 a2 C2
d3 A3 C3
y=
61 21 C1
b2 A2 C2
63 A3 C3
358.] Simultaneous Simple Equations. 215

But, by Art. 350, if the first and second columns of the

determinants in the numerator and denominator be inter-
changed, the determinants will have the same numerical
values with opposite signs. Hence the interchange will
not affect the value ofy. Thus, finally,
a₁ d₁ C₁
a2 d2 C2
a3 d3 C3 •

y= (ε)
a1 61 C1
a2 b2 C2
A3 63 C3
By an exactly similar process it can be shown that
a₁ b₁ d₁1

a2 b2 d2
a3 b3 d3
2=
(5)
a b C1
a2 b2 C2
A3 63 C3
357. The formulae (8), (6), (5) are easy to remember. The
denominator in the value of each of the three quantities
x, y, z is the same, and is the determinant of the nine
coefficients of the unknown quantities written in the order
in which they occur. The numerator in each case is
obtained from the denominator by replacing each coefficient
of the particular unknown whose value is to be given by
the term which stands on the right-hand of the corre-
sponding equation. Thus, in the value of x, each a must be
replaced by the corresponding d, while in the value of y it
is the b's for which d's are substituted ; and in the value
of z, c's are similarly exchanged.
358. The student can always solve any set of simul-
taneous equations with three unknowns by the formulae
ofArts. 355, 356. In the case of equations with numerical
or unsymmetrical coefficients there will often be little
216 General Formulae for the Solution of [359.
advantage gained by this method over that suggested in.
Art. 322 and exemplified in Art. 323. When the coeffi-
cients of the unknowns are reasonably symmetrical with
reference to the letters involved, the determinant method
often presents great facilities.
359. The following equations will serve as an instance :
ax + by + cz = A,
a²x + by + c²z = A²,
a3x + by + c3z = A³.
By (8) of Art. 355,
Abc
A2 b2c2
43 33
=
a bc
a² 62 c2
α3 3 3
al C 111
The denominator a² 62 2 = abc a b C , Art. 354.
a2 62 2
111

Now if, in a bc , a were given the value b, the

a² 2 c2
determinant would vanish by Art. 351. Hence a- b
must be a factor of the determinant by the theorem of
Art. 163. Similarly a- c and b -c must be factors,
and therefore the product of these three, or (a − b) (a- c)
(b - c). Now the determinant is a homogeneous function
of three dimensions in a, b, c, since each term is the product
of one factor out of each row. Hence it can contain no
other factors involving the letters a, b, c. Hence
111
a bc = k (a − b) (a − c) (b - c),
a2 2 2
where k is some numerical factor independent of a, b, c.
360.] Simultaneous Simple Equations. 217

Hence the denominator of the value of x is

kabc(a - b)(a- c(b -c).
But the numerator is obtained from the denominator by
replacing a by A. Hence the numerator is
kAbc (A - 6)(A - c)(b - c).
Hence x =
kAbc (A - 1) (A − c) (b − c) _4(4-6) (4- c)
=

kabc (a − b) (a - c) (b − c) - a(a − b)(a - c)

The value of k is easily seen to be - 1, by comparing the
first term in the determinant with the known value of the
product of (a - b), (a - c), (b - c) ; but, as is shown above,
its discovery is not necessary for the evaluation of x. The
values of y and z can be written down by symmetry.
360. As another example take the equations
x + y + z = a + b + c,
bx + cy + az = ab + bc + ca,
cx + ay + bz = ab + bc + ca.
Hence the value of x is given by the equation
a+ b+c 11

ab + bc + ca ca
ac + ab + bc ab
X
111

bca
cab
By an extension of (a) of Art. 352 the numerator of a
can be written as the sum of the three determinants,
a 11 611 C 11

and = abca + bcca + caca

aca b aba b cbab

111 111 111
=abca +bcca +caca
cab aab bab
111

=abca ,
since the last two determinants vanish,
cab byArt. 351.
218 General Formulae for the Solution of [361 .
111

a bca
cab
Hence x = = a.
111
bca
cab

Similarly y =b, z = c.
361. If three unknown quantities x, y, z be connected by
two relations of the form
a₁ x + by + c₁ z = 0 , (1)
ax + by + c z = 0 , (2)
it is not possible to determine from these two the values
of x, y, z. It is however possible to determine the values of
X
the fractions and " .
2 2

The two equations can be written

X
+ 1 ? = -1
2

y
a2
. + 0 = - .

These forms are obtained from the original ones by dividing

by z and transposing the terms c₁ and co.
From these latter equations,by Art. 337, it follows
that
-

1 11
- b Ca
=

ab
; by Art. 343 ( 3) and (8)
0101
ab az ba
1 1

y =
Ca
=
ab ab
a b ab
362.] Simultaneous Simple Equations. 219

From the first of these equations, multiplying by the

product of the denominators, it follows that
2

61 C1 41 61
b2 C2 a2 b2
From the equation giving 2
it similarly follows that
2
y

C1 1
a2 b2
C2 A2
And these two relations can be symmetrically written
X 2
y
=
;
61 1 C1 A1 41 61
62 C2 C2 A2 a2 b2
or, working out the determinants,
X 2
y
(a)
61C2-6201 C1A2-C2A1 a1b2- a2b1
362. If x, y, z be further connected by a third relation,
A3 x + b3 Y + C3z = 0, (3)
and if each of the fractions in (a) be assumed equal to a
quantity u, the substitution of x, y, z from (a) in this new
equation gives
C1 1
{3 C2 +02 + C33 C2 A2 b2 } = 0,
u

and assuming that u is not zero, that is, that the three
equations (1), (2), (3) are satisfied by values of x, y, z,
different from zero, it follows that

аз
C1
+ C1 1
+C3
a, b, = 0. (β)
b2 C2 C2 A2 a2 b2
By working out the determinants it is easily seen from
Art. 344 that this is equivalent to
a161 C1
a2 62 C2 = 0. (γ)
A3 3 C3
220 General Formulae for the Solution of [363 .
Hence this latter is the condition that the three equations
(1), (2), (3) may be simultaneously satisfied by values ofx, y, z
differing from zero.
363. The deduction of (y) from (3) can be also effected in
the following manner.
By (a) and (3) of Art. 343, (β) can be written
61 62 A1 A2 a1 a2
аз 3 +C3 = 0,
C1 C2 C1 C2 61 62
which by Art. 347 gives
A3 A1 A2
b3b1 b2 = 0,
C3 C1 C2

whence (y) follows by Arts. 350 and 349.

364. In solving the system of simultaneous equations
of Arts. 321 , 355, it occasionally happens that the de-
nominator in the values of x, y, z, namely the determinant
a161 C1
a2 b2 C2 , vanishes. When this is the case it indicates

either that the three equations are not independent or are

inconsistent.
The last article shows that the vanishing of the deter-
minant in question implies that if two of the expressions
a₁x + by + c₁z, a₂x + b₂y + C₂z, azx + b3y + C32
vanish for any values of x, y, z, the third will do so also.
Two quantities pand q can be found to satisfy the two
conditions a3 = pa₁ + qa₂ , b₃ = pb₁ + qb2 ;
whence

azx + by + C3z -p(ax + by + c₁z) -q (a2x + b2y + C22)

= (a3 -pa₁ - qa₂) x + (b₃ -pb₁ - qb2) y + (C3 -PC1 - QC2) 2,
= (C3-PC1- QC2)z.
Hence if a₁x + b₁y + c₁z, a₂x + by + coz vanish for any
364.] Simultaneous Simple Equations. 221

values α, β, y of x, y, z it follows that this final expression

must vanish. Thus we must have
C3 = PC1 + qC2;
and therefore
a3x + by + c3z = p (a₁x + b₁y + c₁z) + q(a₂ x + b₂ y + C22).
Thus the third of the three equations of Art. 355 will be
derivable from the first two if d₃ = pd₁ + qd₂ , and will be
inconsistent with them if this be not the case.
By the help of the last article the student will see that
the three conditions
a3=pa₁ + qa2, b₃ = pb₁ + qb2, C3 = PC1 + qC2,
lead to the condition
a1 b1 C1
a2b C2 = 0 ;
A3 63 C3
while the combination of the condition d3 = pd₁ + qd, with
any two of these leads to the vanishing of the numerators
of the values of x, y, z given in Arts. 355, 356.
Thus if the denominator alone vanish, the equations are
inconsistent ; if the numerators also vanish, the equations
are not all independent.
Similar remarks apply to the solutions in Arts. 334, 335,
of equations with two unknowns.
The methods of this chapter can be applied to the solu-
tion of any number of simultaneous equations with the
same number of unknowns. For the development of these
methods the student is however referred to special treatises
on determinants.

{
222 General Formulae for the Solution of [364.

EXAMPLES .

1. Solve the equations

ax+by = f,
bx+ cy = g,
and explain the solution when 62 ac.

2. Nolve by determinants the equations

(1) 2 + 3y = 7, 5x- 4y = 6 ,
(a) 3 + 7y = 23, 8y- 5x = 1 .
M. Find the values of the determinants,
bu
u ис ис

awv

4. From the last question deduce the value of wbu

vue

acb
5. Find the value of the determinant cba
bac

6. Reduce the determinant x²-yz, 22-xy

200

7. Reduce to its simplest form the sum of the two deter-

minants
a + x, b + y, c + z а -х, в - у, с- 2
b + x, c + y, a + z + b - х, с - у, а - z .
c + x, a + y, b + z с - х, а - у, b -z

8. By the determinant method solve the equations

3x+ 2y + 5z = 21 ,
2x- 3y + 4z = 11 ,
x + 3y + 7z = 20 .

9. By the determinant method obtain the solution of the

equations
2x- 3y + 4z = 5, 4x+ 3y + 2z = 6, 8x + 15y- 2z = 8 ;
I explain the result at which you arrive.
364.] Simultaneous Simple Equations. 223

Solve the equations : -

10. ax + cy + bz = cx + by + az = bx + ay + cz
= a³ + b³ + c³ - 3abc .
(b - c) x + (c - a) y + (a − b) z = 0,
11.
a(b -c)x + b (c - a) y + c(a - b)z = - (ab)(b - c) (c - а),
a(b² -c²) x + b (c²- a²) y + c (a² -b²) z = 0 .
x +y + z = 1,
ax + by + cz = m,
12.

{
}
l- a +++ 1-c
2 = 1.
lx + my + nz = mn + nl + lm,
13. x + y + z = 1 + m + n,
(m - n) x + (n - 1) y + (1 -m) z = 0.
}

1
 CHAPTER XIII .

SQUARE ROOT .

365. In the process of solving quadratic equations

(Art. 286) it is frequently necessary to discover the square
root (Art. 64) of a given number or algebraical expression.
The present chapter will be devoted to explaining the
manner in which this root can be found.
366. The problem of division was solved by noting the
results obtained in multiplication. Similarly the discovery
of the square root of an expression is effected by com-
parison with the known forms of the squares of different
expressions.
Thus the square of a + b is known to be a² + 2ab + b²
(Art. 114). Hence the square root of a² + 2ab + b² is a + b .
367. The first term in the square root, namely a, is
evidently the square root of the first term in the given
expression, namely a².
In any given case, the first part of the square root can
thus be determined at once by inspection :
a² + 2ab + b² ( a + b
a2

2a + b2ab + b²
2ab + b²

If a², that is, the square of the first part of the root, be
subtracted from the given square, the remainder is 2ab + b².
It is known that the second part of the square root is b.
This quantity can be obtained by dividing the first term
370.] Square Root. 225

of the remainder by 2a, which may be called a trial

divisor.
To verify that the term thus obtained is the remainder
of the square root, it must be added to the trial divisor so
as to form 2a + b, the complete divisor.
If this complete divisor be multiplied by 6 the result is
2ab + b² ; and when this is subtracted from the remainder
there is nothing left.
368. The process in the last article is sometimes
perplexing to students because they feel that it teaches
them nothing that was not previously known, and that, as
from the beginning it was clear that a + b was the square
root required, the whole labour expended is useless.
The object of the article is however to show the manner
in which the expression a + b can be deduced from the given
expression a² + 2ab + b². By applying an identical process
in a case in which the square root is not known, the
unknown square root can be found.
369. It is important to notice that at the end of the
first step, a² has been subtracted; and at the end of the
second step (a + b)2 has on the whole been subtracted. The
fact that there is finally no remainder proves that the
original expression is the square of a + b, or that a + b is
a square root of the given expression.
370. Let it be required to find the square root of
9x² + 12xy + 4y².
Here the first term is 922, which is the square of 3x.
Hence the first term in the square root, corresponding to a
in Art. 367, is 3x :
9 x² + 12 xy + 4y² (3x + 2y
92

6x + 2y 12xy + 4y²
12xy+ 4y²
Q
226 Square Root. [371 .
Subtracting the square of 3x from the original expression
the remainder is 12 xy + 4y². The trial divisor, or 2a, is
62. The quotient corresponding to b, obtained by dividing
the first term in the remainder by 6x, is 2y. The complete
divisor, 2 a + b, becomes 6x + 2y. If this be multiplied by
the term 2y, corresponding to b, the product is 12xy + 4y²,
which leaves no remainder when subtracted from what is
left of the original expression.
Hence 3x + 2y is the square root required.
371. The student will see that, in the process of the
last article, different portions of the square of 2x + 3y
have been in succession subtracted from the given expres-
sion. As there is finally no remainder, the process proves
that the square of 2x + 3y is exactly equal to the given
expression, or that 2x+ 3y is the square root of the
latter.
372. Let it be required to find the square root of
a² + b² + c² + 2ab + 2ac + 2be.

As in division (Art. 148), it is important that an ex-

pression whose square root is required should be arranged
in powers of some letter. If a be taken as dominant letter,
the expression may be written
a² + 2ab + 2ac + b² + 2bc + c² ( a + b + c
a2

2a + b2ab + 2ac + b² + 2bc + c²

2ab +62

2a + 2b + c2ac + 2bc + c²
2ac + 2bc + c²

The square root of the first term is a. Subtracting a², as

before, the remainder is written down in a lower line.
Taking 2a as trial divisor the quotient may be taken as
374.] Square Root. 227

either b or c, according to which term involving the first

power of a is taken first. If 2 ab be chosen, the complete
divisor becomes 2a + b. Multiplying this by b and sub-
tracting, there is still a remainder, 2 ac + 2 bc + c².
Now at this point a + b is the part of the square root
obtained and (Art. 369) the square of a + b has been
subtracted from the original expression. Hence a + 6
occupies exactly the same relation to the present remainder
that a did to the remainder at the end of the first step.
Thus 2 (a + b) or 2a + 20 must be taken as trial divisor.
The quotient is evidently c and the complete divisor is
2a + 2b + c. If this be multiplied by e and the product
subtracted from the remainder, there is nothing left.
Hence (a + b) + c is the whole square root of the given
expression.
373. In the last article the two terms 2ab + 2 ac,
occurring in the remainder after a has been subtracted,
might be written as one term in the form 2a (b + c). The
quotient of this by the trial divisor would be b + c, and this
might then be taken as the remaining part of the square
root. The process would be as indicated below :
a² + 2a (b + c) + b² + 2bc + c² ( a + (b + c)
a2

2a + (b + c) 2a (b + c) + (b² + 2bc + c²)

2a (b + c) + (b + c)²

The complete divisor is 2a + b + c, and when this is multi-

plied by b + c and the product subtracted, there will be no
remainder, since (b + c)² = b2 + 2 bc + c² (Art. 114).
The process of Art. 372 is however for some purposes
more instructive than this second method.
374. Let it be required to find the square root of
x - 6x5 + 15x4-20x³ + 15x² - 6x + 1 .
Q2
 228 Square Root. [375-
The process is indicated below :
x²- 6x + 15x - 20x3 + 15x² - 6x + 1 (x³ - 3x² + 3x - 1
26

2x³- 3x² 6x5 + 15x

-6x + 9x

2 (x³ - 3x²) + 3x + 6x -20x² + 15x2

= 2x²- 6x² + 3x
+ 6x - 18x² + 9x²
2(x³ - 3x² + 3x) -1 -

2x3 + 6x² - 6x + 1
= 2x² - 6x² + 6x- 1
2x² + 6x² - 6x + 1

The square root of the first term is evidently w³. The

first trial divisor is therefore 20³ and the quotient is - 3x² .
The complete divisor is therefore 2x³ - 3x². When the
product of this into - 32 has been subtracted, there is
still a remainder, 6x4-20 x³ + 15x² - 6x + 1. At this
stage the square of x³- 3x² has been subtracted, and
x³- 3x² occupies therefore the position of a in Art. 367
after the first step.
Hence 2 (x³ - 3x²) must be taken as the next trial
divisor, and the process carried on by similar steps until
there is no remainder.
375. The student can easily verify that at the end of
each complete step in the last article the square of that
part of the root which has been ascertained has been sub-
tracted. This is obviously true after the first step. Suppose
it to be true after any step, and that a is the part of the
root already ascertained, so that a² has on the whole been
subtracted ; then 2a is the trial divisor, and if b be the
quotient 2ab + b² is now subtracted. That is, on the whole,
a² + 2ab + b², or (a + b)², is the total subtrahend when a + b
is the part of the root discovered. That is, if the statement
in italics be true after any one step in a process conducted
377.] Square Root. 229

according to the law explained, it will also be true after

the next step.
But the statement is true after the first and second step,
therefore it will be true after the third step, and so on,
till the process comes to an end.
376. It follows that, if at last there is no remainder,
the process proves that the expression obtained is the
square root of the given expression.
For the square of the root obtained has been subtracted
from the given expression, and there is no remainder, that
is, the square of the deduced expression is equal to the
given one, which is, by definition (Art. 64), the condition
that the former is the square root of the latter.
377. Let x - 10x3 + 14x² - 11x + 13 be an expression of
which it is required to ascertain whether it have an exact
square root, and if there be one, to find its value. The
process may be conducted as in the previous examples, and
is represented below :
x² - 10x3 + 14x² - 11x + 13x² - 5x - 11 2

2x² - 5x -10x³ +14 x²

10 x³ +25 x²

2x² - 10x_11 - 11 x² - 11 x + 13
2

- 11 x² +55 +121 4

66-69 4

It is evident that, if there be a square root, x² must be the

first term. Reasoning as in previous articles, it is equally
evident that the next term must be - 5x, and the next
term- . At the end of the indicated process, the square
of x² - 52- has on the whole been subtracted, and there
2

is a remainder, -66x - 9, which is of lower degree in x

230 Square Root. [378 .
than the next trial divisor. Hence the working cannot be
continued, and it is evident, first that the given expression
is not an exact square, and secondly that the square of
x²- 5x- differs from the given expression by a quantity
2

of lower degree in x than that of any other expression.

Hence in a sense x² - 5-12 may be called the nearest
2

approximation to the square root of the given expression.

378. In a manner similar to that of the last article
the approximate square root of any algebraical expression of
which a is the dominant letter can be ascertained ; subject
to the two conditions, (1) that the index of the highest
power of æ involved is even, and (2) that the numerical or
literal coefficient of this highest power is an exact square.
These two conditions are obviously necessary that the first
term in the expression shall be an exact square of some
quantity a.
If, when the process is carried as far as possible, there is
no remainder, it shows that the original expression is an
exact square ; and in that case the exact square root has
been arrived at. If there be a remainder the expression
given is not an exact square and the deduced expression is
only an approximate root.
379. The investigation of the square root of a number
depends on that of an algebraical expression and is con-
tained in the following articles .
380. In the ordinary system of decimal notation any
number is expressed in the form (Art. 164)
Poxn +P₁x - 1 + P2x²-2 + ... + Pn-1x + Pm (0)
where a has the value 10.
The quantities Po , P1, P2 Pu-1 , Pn in this case are called
the digits, p, being the digit in the units' place, pn-1 that
in the tens' place and so on. All these numbers are nu-
merically less than 10, they cannot be negative although any
 382.] Square Root. 231

of them may be zero. If the highest power of 2 involved

be the nth, the number of digits is (n + 1) .
381. The restriction that the digits must always be less
than 10 introduces the chief difference between the mani-
pulation of a number decimally expressed and an algebraical
expression.
Suppose for instance that in multiplying two numbers
together a term qr" is arrived at, where q is greater than
10 or x. Let q contain ap times with remainder r, so that
q = px + r, where we may suppose p and reach less than a.
Then
qxn = (px + r) xn
=pxn + 1 + rx",
and the term qr" is replaced by two terms, one of which,
px" +1, will have to be added to the next higher term in the
number, if there be one, or otherwise set down as the
highest term by itself.
This is the well-known process of ' carrying ' in
arithmetical addition and multiplication, the part p which
gives the number of a's, or tens, in q ( = px + r) being
' carried ' one place to the left so as to be a multiplier of
+1 instead of x" .
382. The square of any number as (a) in Art. 380 will
thus probably assume a different form from that which it
would primarily take as an algebraical expression.
Treating it algebraically, the terms involving the
highest powers of a are easily found by multiplication in
the ordinary method to be
Po²x2n + 2PoP₁x2n−1 + (2PoP2 +P₁²) x2n-2 + ... (β)
Now po may have any of the values 1, 2, 3 up to 9 inclusive.
2
Hence po may be only unity or may be as great as 81, and
2
part ofpo may have to be carried to a higher place. Thus
the square of (a) may involve x2n+1. It cannot involve a
higher power of a than this, because (a) which contains no
232 Square Root. [383.
higher power than 2" is certainly less than 2n + 1 * , and
therefore the square of (a) is certainly less than 22n+ 2.
Hence the square of (a), a number involving n + 1 digits,
may contain 2n or 2n + 1 digits.
Thus the square of a number of one digit will contain
either 1 or 2 digits ; that of a number of 2 digits will
have 3 or 4 digits, and so on.
383. The second term in (β), 2pop₁x2n-1, may similarly
give something to be carried to the term involving 22",
and thus the terms, which in the Algebraical process of
finding the square and the inverse process of finding the
square root were distinct, become in the arithmetical square
of a number mixed up and indistinguishable. This fact, as
we shall see, sometimes gives rise to a little uncertainty in
ascertaining the second figure of the square root.
384. The usual method of ascertaining the number of
figures in the square root of a given number, is to put a
dot over every other digit beginning with the units' digit.
These dots will be over the digits which multiply the unit,
102, 104, 106, ...and every even power of 10. If there be
n + 1 dots the last of them will stand over 102%, and the
given number contains either 102 or 102n+ 1 as the
multiplier of its highest digit. Hence its square root
must contain 10" as multiplier of its highest digit. The
square root therefore contains as many digits as there are
dots placed over the number.
Also if the first digit of the square root be po, the multi-
plier of 102" in the square will be at least po² and may
exceed this value (Art. 383).
Thus po will be the integer whose square is next below
the multiplier of 102" in the given number ; that is, next
* The greatest number containing six digits, for instance, is 999999, and
this is less by unity than 10,000,000 or 107. Thus the greatest number
containing n + 1 digits is less than the least number containing n + 2 digits ;
that is, is less than 10 +1.
386.] Square Root. 233

below the number formed by the figure under the left hand
dot and the one to the left of it, if there be one. The first
term of the square root is therefore po10". The first trial
divisor will be 2po10".
385. At this point a difficulty comes in. In the alge-
braical process, the trial divisor being 2a, the terms in
the remainder 2ab + b² are quite distinct and the quotient
b is obtainable by dividing the first of them by 2a. In
the number the terms are mixed together, and the whole of
the remainder must be taken as dividend. Occasionally
the value of b thus deduced is too large, and when the
complete divisor 2a + b is multiplied by b, a number is ob-
tained greater than the remainder of the given square. If
this happen, a value of the second digit smaller by unity must
be taken, and the process repeated until the correct digit has
been found. After the second digit has been found there is,
as will be shown (Art. 391), no further chance of error.
386. As an example let it be required to find the square
root of 1522756 .

1522756 ( 1000 + 200 + 30+ 4 = 1234

1000000

2a = 2000 522756
b= 200

2a + b 2200 440000

2a = 2400 82756
b= 30

2a + b = 2430 72900

2a2460 9856

b= 4

2a + b = 2464 9856

Placing dots over each alternate digit, beginning with

234 Square Root. [387 .

the units' figure, the last dot falls on the left-hand figure
which multiplies 106. Hence the first digit in the square
root is 103. The trial divisor is 2000, and the quotient
must lie between 200 and 300. Hence 200 is the next
part of the root. The process in this case, since there is
finally no remainder, shows that the square of 1234 is the
given number, or that the square root of the latter is 1234.
387. As another example let it be required to find the
square root of 31371201 .
31371201 ( 5000+ 600 + 1 = 5601
25000000

2a10000 6371201
b= 600

2a + b 10600 6360000

2a = 11200 11201

b = 1

11201 11201

Here the last dot stands over the last digit but one.
The highest even power of 10 involved is 106. Hence the
highest power of 10 in the square root is 103, and the digit
multiplying this power in the square root must have its
square less than 31, the multiplier of 106 in the given
number. Hence the first term in the square root is 5000.
The rest of the process is similar to the former examples.
388. In working out such examples as those of the last
two articles, it is not usual, nor is it necessary, to write
out all the noughts. Neither is it necessary at each stage
to bring down more than two figures of the remainder,
except in the last stage of the second example, where if only
the figures 12 were brought down the trial divisor would
not go into the part brought down. This is indicated by
the ' nought ' occurring in the result 5601. The operations
of the last two articles are concisely represented below.
390.] Square Root. 235

1522756 (1234 31371201 ( 5601

1 25

22 52 106 637

44 636

243 827 11201 11201

729 11201

2464 9856

9856

The student should carefully compare these conciser forms

with those previously given.
389. Suppose that the square root required is represented
by an expression such as (a) in Art. 380, so that 10" is
the highest power of 10 involved. Let us further suppose
that the first 8 + 1 digits have been found by the process of
the last few articles. The whole number of digits being
n + 1 there remain n- s undiscovered. The highest power
of 10 involved in these remaining digits is therefore
10-8-1 (Art. 380). And the number represented by the un-
discovered digits is less than 10"-* (Art. 382, note).
390. If the part of the root already ascertained be de-
noted by a and the unknown part by b, the whole square
root is a + b. The original number is therefore a² + 2ab + b²
from which at the supposed stage of the process a² has
been subtracted (Arts. 369, 372, 375). The remainder is
therefore 2ab + b² and the trial divisor is 2 a.
The total quotient of the remainder by the trial divisor
is thus 62

2a

It has been shown that bis less than 10"-", while a is

greater than 10".
72 102-28
Hence is, on both accounts, less than or less
2a 2 x 10"'
thanx 10"-28, and a fortiori less than 10-28.
236 Square Root. [391 .
62
That is, 2a
contains n - 2s digits at most (Art. 380),
62
while b contains n - s, that is, at least s more than 2a
72
Hence the highest s digits in band b + 2 a will be the
same. Thus when 8 + 1 figures have been obtained by the
ordinary process, the trial divisor will give the next s
figures accurately.
391. Hence, for instance, when two figures have been
ascertained by the regular process, one more can be ac-
curately obtained by division by the trial divisor.
If one figure only has been obtained, it is uncertain
whether the trial divisor will give even one more figure
with accuracy. Trial must be made of the first figure of
the quotient, and if this turns out to be too large, a number
smaller by unity must next be tried until the correct one
is reached.
392. As an example in which this difficulty occurs let
the given square be 346921 .
3469215
250000

1090 96921

98100

Pointing it in the usual way, it is seen that there will be

three figures in the root, and that the first of these is 5,
that is, the first part of the square root is 500. Subtracting
the square of this from the given number, the remainder is
96921. The trial divisor 2a is 1000, and the quotient
is 90. Adding this to the trial divisor, the complete
divisor becomes 1090, and when multiplied by 90 thus
gives 98100, a quantity greater than the remainder of the
given number. Hence 9 is too large for the second figure.
Replacing the 9 by 8 the square root is found, as in former
393-] Square Root. 237

examples, to be 589. The process, omitting needless figures,

is represented below :
346921 ( 589
25

2a = 100 969

b = 8

2a + b = 108 864

2a1160 10521

b= 9

2a + b = 1169 10521

393. Let it be required to find the square root of

200000000.

Pointing in the usual way, it is clear that there are five

figures in the square root and that the first of them is unity.
200000000 (14142
1

2a = 20 100

b= 4

2a + b = 24 96

2a = 280 400

b= 1

2a + b = 281 281

2a = 2820 11900

b= 4

2a + b = 2824 11296

2a = 28280 60400

b= 2

2a + b = 28282 56564

3836
238 Square Root. [394.

In this process, it may be noticed that in the second step

the quotient of the remainder 100 by the trial divisor 20
is 5. If 5 were taken as second figure in the square root
the product of the complete divisor unto 5 would be too
large. This example thus gives another illustration of the
possibility of error in the second figure.
It will be further noticed that there is, finally, a re-
mainder. Thus the given number is not an exact square.
The number 14142 is the largest number whose square
does not exceed the given number 200000000, and is
frequently called the approximate value of the square root
of the latter.
394. The square root of a fraction can be obtained by
taking the square roots of the numerator and denominator
separately ; thus
Sa
b
=
(Art 252 (β))

and this is true whether a and b are algebraical expressions

or numbers.
When both the numerator and denominator are exact
squares, an exact square root of the fraction can be found.
When the numerator or denominator are, either or both
of them, not exact squares, only an approximate square root
can be found.
It is always possible to multiply the numerator and
denominator by such quantities as to make the denomin-
ator an exact square. When this is done, the fraction
whose numerator is the approximate square root (Arts.
378 , 393) of the numerator and whose denominator is the
square root of the altered denominator of the given fraction,
is called the approximate square root of the given fraction.
a Sa Jab
Thus
6
396.] Square Root. 239

And if a be the approximate value ofVab, is called the

a

approximate value ofN

395. The most important class of fractions are those
which are expressed as decimals, that is, fractions having
some power of 10 for their denominators.
The denominator, in the case of a fraction or mixed
number expressed as a decimal, is the power of ten whose
index (Art. 62) is the number of digits in the decimal part
of the given number. In order that the denominator may
be an exact square, there must therefore be an even number
of digits to the right of the decimal point. If there be
originally an odd number, a nought can be added at the
end to make the number of decimal places even. The
square root of the denominator will then be a power of ten
whose index is half the number of decimal places in the
given number.
To find the square root, exact or approximate, of the
fraction or mixed number, all that has now to be done is
to find, by previous methods, the square root of the
numerator, and from that to mark off half as many decimal
places as there were in the original number.
396. Let it for instance be required to find the square
root of 3137-1201 .
This really represents a fraction whose numerator is
31371201 and whose denominator is 10000. The square
root of the former is (Art. 387) 5601, and that of the latter
5601
is 100. Hence the square root of 3137-1201 is 100, or
56.01 .

It thus appears that the square root of 3137-1201 can be

obtained by the same process as that of 31371201. It is
clear that since there is an even number of decimal places
240 Square Root. [397-
the dots used in pointing (Art. 384) will fall over the same
digits whether we begin over the unit figure 7 in 3137-1201
and place a dot over each alternate figure both to right and
left, or place a dot over the unit figure 1 in 31371201
and point backwards, as in Art. 387.
The former method is usually more convenient, and the
rule for finding the square root of a mixed number becomes
Place a dot over the unit figure of the mixed
number and over every alternate figure both to the right
and the left. If the last dot to the right does not come
over the last decimal figure, add a nought, which will
make the number of decimal places even, and a dot will
fall over this nought. Then find the square root of the
whole, disregarding the decimal point, and mark off in the
result as many decimal places as there are dots to the right
of the decimal point.
397. The approximate square root of a number can now
be found to any number of decimal places.
Thus 2 can be written as 2.00, 2.0000, or in an infinite
number of such forms, adding noughts ad libitum. The
approximate square root of 200 has been found (Art. 393)
to be 14. Hence an approximation to the root of 2 is 1.4.
The first three steps of the working in Art. 393 give the
approximate square root of 20000 as 141. Hence a second
approximation to the square root of 2 is 1.41. Thus the
whole of the work done in that article gives 1.4142 as
an approximate square root of 2 .
398. The successive approximations 1.4, 1.41, 1.414,
1.4142 are numbers whose squares differ from 2 by succes-
sively smaller quantities.
For by Art. 375 it follows from the working in
Art. 393 that the squares of 14000, 14100, 14140, and
14142 differ from 200000000 by the numbers 4000000,
1190000, 60400, and 3836 respectively. And dividing all
400.] Square Root. 241

the numbers by the square of 10000, that is 100000000, it

follows that the squares of 1.4, 1.41, 1.414, 1.4142 differ
from 2 by the respective fractions 04, 0119, 000604, and
00003836 .

399. The example in Art. 393 has thus given what is

called the approximate square root of 2 to four places of
decimals. If eight more noughts had been added to the
number 2 at the beginning of the operation, the whole
square root would have contained nine figures. Of these, five
have been found, and therefore, in virtue of Art. 390, four
more can be obtained merely by dividing the remainder by
the trial divisor.
The part already found is 141420000 ; the trial divisor
is 282840000, and the remainder is 383600000000. Hence
the four figures required are to be obtained by dividing
38360000 by 28284.
28284 ) 38360000 ( 1356
28284

100760

84852

159080

141420

176600

169704

6896

Hence 1356 are the last four figures, and the approx-
imate square root of 2x1016 is therefore 141421356.
Hence the square root of 2 approximately to eight places
of decimals is 1-41421356.
400. The examples of numerical square roots in this
chapter have not been in all points worked out in the
briefest manner possible. The object has been rather to
R
242 Square Root.

exhibit the reason of the process than to give rules for

expeditious operation for which the student can refer to
any treatise on Arithmetic : although a little practice will
soon lead to all useful abbreviations when once the prin-
ciple of the method is understood.

EXAMPLES.

Find the square roots of

1. 4x² + 20xy + 25y².
2.9a²- 12ab + 4b2 .

3. x² - 4x² + 4.

4. 49x² - 42ху +9 y².

5. x² - 4x²y +6x²y² - 4xy² +y² .
6. 4a - 12a²b + 29 a²b² - 30ab3 + 256 .
7. 16x + 24x²y + 9x⁴y² - 16x²y³ – 12x²y + 4yº.
2

8.4x - 4x²y² + 12x²y + x²y+ − 6 xy +9 y .

9. x - 12x + 60x² - 160x² + 240x² - 192x + 64.
225 45 111 33 121
3
10. x²+ 23y- x²y- xy²+
256 32 96 24 y.
144
22 2x 2
11. ++++ 3 .
Y

12. ( +2) + 2 + 2) + (Y+ 2).

13. x + y + z- 2√ry- 2√yz+ 2√xz.
14. 174-5041, 2732409, 90643-1449 .
9 12
15. x² - 4x+ +10.
a

16. (a² + b²) (c² + d²) - (ad - bc)²

 Square Root. 243

17. a²b + bc¯² + c²a² + 2ac-1 + 2cb-1 + 2ba-1.

18. 16a² (a + b + c) + 16a²bc + 4a2 (b² + c²) + 4abc (b + c) + b²c².
19. x² + 2x² + x + 1 - 2x (x + 1).
20. (x + x² + 2x + 2x + x + 1 ) (x - 1 ) (x- 1).
21.1-2x + x + 6x - 6x² + 9x3.

22.
x² + x² + 2x + 2x + x + 1
x + 5x + 8x + 4

23. Find the value of √3 to eight places of decimals.

24. Find the value of √24 to six places of decimals.

R2
 CHAPTER XIV .

QUADRATIC EQUATIONS .

401. An equation which involves the unknown quantity

to the second and no higher degree is called a quadratic
equation (Art. 286). We shall at present deal only with
equations involving one unknown quantity.
402. It is evident that by processes of reduction similar
to those employed in solving simple equations (Arts. 290,
291) any quadratic equation can be reduced to consist of
three terms at most, two of them involving respectively
x² and æ, and the third being independent of x. The
general type of a quadratic equation is therefore
ax² + bx + c = 0 , (1)
where a, b, c are scalar quantities (Art. 226) independent
of x.
403. The following preliminary theorem is useful in the
1
solution of such equations.
From Arts. 114 and 115 it follows that
x² + 2 ax + a² = (x + a)²,
2-2axах + a² = (x -a)².

If in these identities be written instead of a, they

become
x² + kx + ( k) = (x + k)² ; (a)
x² - kx + ( k)² = (x - k)2 . (β)
Hence, if ( 1 ) be added to either of the expressions x² + kx
or x² - kx, the sum will be an exact square.
404. Let the quadratic equation be
x²- 4x + 3 = 0 .
405.] Quadratic Equations. 245

Subtracting 3 from both sides of this equation, it follows

that x² - 4x = -3 .
On the left-hand side there is now an expression of the
form x²-kx, k having the value 4. Hence if ( k) , that is,
in this case, (2)², or 4, be added to both sides, the expression
on the left-hand side will be the square of x - k, or x - 2 .
Thus x² - 4x + 4 = 4-3 = 1 ,
or (x - 2)² = 1 .
Hence x - 2 is a quantity such that its square is 1, or, by
Art. 280, x- 2 must be either +1 or -1 .
Thus the given quadratic equation resolves itself into the
two simple equations
x- 2 = + 1,
-2-1 .

The former gives x = 3, the latter ax = 1 .

Hence the quadratic equation has two roots (Art. 284) ;
that is, two values of x, namely 3 and 1, satisfy it.
405. Let the equation be
х x+ 1 13
+
x+ 1 6

Multiplying the equal quantities by 6x (x + 1) so as to

get rid of fractions, it follows by Axiom 3 (Arts. 53, 287)
that
6x2 + 6 (x + 1)² = 13 x (x + 1 ) ;
or, multiplying out,
6 x² + 6 x² + 12x + 6 = 13x² + 13x ;
whence transposing and collecting terms,
-x2 -x + 6 = 0,
which is of the form (1) of Art. 402.
It follows, by transposition, that
x² + x = 6 .
Here the left-hand expression is of the form x² + ka,
k having the value unity. The addition of (1) to both
246 Quadratic Equations. [406.

sides will, by Art. 403, (a), make the left-hand member

the square of x +
2
Thus
x²+x+ ( )² = 6 + ( ) ;
25
or
( + ) = 6 +1=2 4 4

25
Hence x + is a quantity whose square is 4
; that is,
5 5
x+ must (Art. 280) be either + 2 or
2
The quad-
ratic equation gives therefore the two simple equations
1 5
x+ = ,
2 +2 '
1 5
and x+
2 2

From the former we obtain

5 1 4
x= = 2 ;
2 2 2

and from the latter

5 1 6
= -3 .
2 2 2

The two roots are therefore 2 and -3 .

406. The solution of the general type equation of
Art. 402, namely
ax² + bx + c = 0 ,
can be conducted in a similar manner.
Subtracting e from each of the two quantities given
equal, it follows that
ax² + bx = — с.

Dividing these equal quantities by a, the quotients are

equal by Axiom 4, or
b
x² + - x
a a
406.] Quadratic Equations. 247

16
Ifthe square of 2a be added to these two equal quan-
tities, the sums will be equal by Axiom 1, or
b 162 162 c
x² + a
- 2
-

32 C
=

4a2 a

b2-4ac
4a²

By Art. 403 the left-hand member of this equation is

16
the square of (x + 2 a Hence the last equation can be
written
2
62-4 ас
( +20) =

4a²
b 62-4 ac
That is, x + 2 a is a quantity whose square is 4a²
,

b
whence x + must be either
2a

+ √b2-4ac - √b² - 4ac

or
2a 2a

Thus the quadratic equation is satisfied by a value of x

which satisfies either of the two simple equations
b + √b - 4ac
x+ ,

2a 2a

b -b²- 4ac
or x+
2a 2a

These equations give respectively

b √b - 4ac -b+ b² - 4 ас
+ ,
2a 2a 2a

b² - 4ac
4а с
(a)
b √b²- 4ac -b- √b²-
and
2a 2a 2a
248 Quadratic Equations. [407.
The two values of æ are usually expressed by the single
formula
-6 + √62-4ac
,
2a (β)
which evidently gives the same values as (a), if first the
sign + and then the sign - be taken. With either sign
the value of x in (3) satisfies the given equation.
407. The formula (3) of the last article can be em-
ployed to find the roots of any quadratic equation in the
reduced form.
Thus let the equation be
x²- 4x + 3 = 0 ;
comparing this with the general type equation
ax² + bx + c = 0,
it is seen that a has the value unity, bis - 4, and cis 3 .
Hence the values of æ, as given by the formula (3), are
х
- ( - 4) + ( - 4)² - 4 × 1 × 3
2

=
4+ √16-12
2

4+2 6 2
or
2

2 2 2
= 3 or 1 ,
the same results as were obtained in Art. 404.
408. Taking, again, the equation (Art. 405)
x² + x - 6 = 0 .
Here a = 1 , b = 1 , c = −6 ; the values of a are con-
sequently given by the formula
− 1 + √(1)² - 4 x 1x ( -6)
2

-1+ 1 + 24
2
 410.] Quadratic Equations. 249

-1 + 5 4 -6
or
2 2 2

= 2 or -3 ;
agreeing with the result of Art. 405.
409. The formula (3) ofArt. 406 thus contains implicitly
the solution of every quadratic equation when reduced
to its simplest form. The student is recommended to
make himself thoroughly familiar with this formula so
as to be able to apply it at once to any given equation.
The solution of any particular equation may of course
be conducted on the lines of the working in Art. 406 and
the roots investigated independently, as has been done
in Arts. 404, 405, but the quotation of the formula (3)
is usually the easier way. The general method of pro-
ceeding may be verbally stated thus :
Reduce the equation by getting rid of fractions and
transposing so that the terms involving 2 and a shall
be on one side, and the term or terms independent of a
on the other side of the equation. Then divide both sides
by the coefficient of x², so that its coefficient shall become
+1 .

To both sides of this reduced equation add the square of

half the coefficient of x. The left-hand side then becomes,
by Art. 403, an exact square, and the square root of this
expression must be equal to one or other of the numbers
whose square is the number or expression on the right-
hand side. Thus two values of a are obtained.
410. In the process of solution of Art. 406 the following
equation is deduced from the given quadratic,
2
b2-4ac
(x + 2)
2a = 4a

The sign of the number (62-4ac) comes here to be of

great importance in discriminating between the character of
the roots in different cases.
250 Quadratic Equations. [411 .
First let 62-4ac be positive. Then it has been already
explained (Arts. 220-227) that a length can be geome-
trically constructed such that the square on it is represented
by 62-4 ac whether the latter number be an exact square
or not. Thus √b2-4ac in this case represents a real or
scalar quantity (Art. 226), and if this quantity be denoted
-b +k -b-k
by k, the two values of æ, namely and are
2a 2a

also real or scalar quantities.

411. Secondly, let b2-4ac have the value zero. Then
62 b
(x + 2 a is also zero. Hence + is zero, and there-
2a

b
fore a can only have one value, namely - In this
2a

case the two values of x in the general solution, namely

-b + b² - 4 ас -b - b² - 4ас
ac
and
2a 2a

b
each reduce to the same value, 2a

Hence if 62 -4ac is zero, the two roots of the original

quadratic equation become coincident and equal.
412. Thirdly, let b2-4ac be negative. Then 4ac-b²,
which = -( 2-4 ac), is positive. Let 4 ac - b² be denoted
by the symbol k², where k is a scalar quantity. It follows
that b2-4ac = − k² = i²k² , where i is the operational or
imaginary quantity introduced in Art. 271 .
Hence in this case
2
i2 k2
(x+2 ) =40
b + ik
And thus x + must have one of the values or
2a 2a
-ik -b + ik -b-ik
, that is a must be either 2a
or
2a
2a

In this case the two values of a are operational or im-

aginary.
414.] Quadratic Equations. 251

413. The roots of the equation

ax² + bx + c = 0
are thus real and unequal if b² - 4 ac be positive, they are
real and equal if b2-4ac be zero, and operational or
imaginary if ² - 4ac be negative.
The expression 62-4ac is often called the discriminant of
the expression ax² + bx + c.
414. Whatever may be the character of the roots, the
formula (3) of Art. 406 holds good. In fact the results
of Arts. 410-412 can be readily deduced from that
formula.
Let the two values of x in (a) or (β) of Art. 406 be
denoted by 21, 22 , so that
-b + b²- 4 ac
,
1=
2a

-b- b²- 4ac

2=
2a

By addition
-b + b² - 4ac - b - b² - 4 ас
X1 + x2 = 2a

-26 b
=

2a a
(1)
By multiplication
-b + b² - 4ac -b - √b² - 4 ac
X1X2 2a 2a

= (-b)²- ( b2-4ac)2
4a2
,
(Art. 116)
= 62
b² -(62-4 ас)
4a²
4ac C
=

4a² a
(2)
Since the ordinary laws of multiplication and addition
hold for operational or imaginary quantities as well as for
252 Quadratic Equations. [415.
real or scalar quantities, the results (1) and (2) hold in all
cases whatever be the quality of the roots of the original
equation.
415. From these results it follows that the expression
ax² + bx + c can be resolved into factors in terms of x1
and 22 .
Thus, by (1 ) , b = -a (x1 + x2),
by(2), c =
A X1 X2 .
Therefore
ax² +bx + c = ax²-ах (x1 + x2) + ax1x2
= a { x² -x (x1 + x2) + X1X2}
= a (x-x₁) (x -x₂). (Art. 130) ( y)
It may be noticed that this last result follows from the
theorem of Art. 163, on the assumption that the quadratic
equation ax² + bx + c = 0
has two roots, 21 and 22 .
Because a₁ is a root of the equation it follows that the
expression ax2 + bx + c vanishes when a₁ is put for 2. Hence
x-x₁ is a factor of the expression. Similarly x-a₂ is a
factor, and as the product of (x - 1) (x - 2) is of the second
degree in æ, there can be no other factor involving a. The
numerical factor a is evidently necessary to make the coef-
ficient of x² the same in both. Hence

ax² + bx + c = a (x -x₁) (x - x₂).

416. The whole subject may now be treated from a dif-
ferent starting-point.
The solution of the equation
ax² + bx + c = 0

means the discovery of the values of a which make the ex-

pression ax2 + bx + c equal to zero, these values being called
the roots of the equation.
By the theorem of Art. 163 it follows that this dis-
covery will be effected if the factors of the expression
417.] Quadratic Equations. 253

ax² + bx + c can be found. Because if ₁ be a root of the

equation, the theorem referred to shows that x-x₁ is a
factor of the expression ; and conversely, if x- x₁ be a factor
of the expression, a₁ is a root of the equation. We must
proceed to investigate the factors of the given quadratic
expression.
417. It will not limit the usefulness of the investigation
if a be supposed a positive scalar. With this restriction it
can be shown that ax² + bx + c can be always reduced to one
of the forms u² + v² or u² - v², where u is an expression
involving a to the first degree and v is a real quantity
independent of x.
For
b
ax² + bx + c = a (x² + x + )
a

b 62 62
Ξα x² + -

a 2a 2a

62 ас

= a {(x + 2 a } (Art. 403)

2
b -4 ас
(x + 4a
(1)
First, let 2-4ac be positive. Then a scalar quantity v
62-4ас
can be found such that = v². And if
4a

u = √a (x + 2a ),
equation (1) gives
ax² + bx + c = u² -v². (2)
Secondly, let b² - 4ac be negative, then 4ac-b² is posi-
tive, and it will be a legitimate assumption to suppose
2
4 ac-b2
4a where v₁ is a scalar quantity.
254 Quadratic Equations. [418.
Making the same assumption as before, as to the meaning
of u, ( 1) becomes 2
ax² + bx + c = u² + v₁² . (3)
418. The factors of u² -v² are u - v and u + v (Art. 116).
Hence if b² - 4 ac be positive,
b
ax²+bx+ c = { √ā(x + 2 ) --4ac
{√ (x + 2 ) + -4aca

Hence in order that ax² + bx + c may be zero, one or other

of its factors must vanish ; that is, & must satisfy one or
other of the simple equations
√b² - 4ac
√a(x + 2a
2√a
= 0,

b √b - 4ac
= 0.
or
√a(x + 2a
+
2√a

Dividing these equations by Va, and transposing, they

give respectively
√b - 4ac
x= + ,
2a 2a

b √b - 4ac
and x= + ,

2a 2a

the results (a) of Art. 406.

419. Next assuming that 2 - 4 ac is negative
ax² + bx + c = u² + vì².
But (Art. 281 ) u² + v₁² = (u + ir) (u - iv).
Hence
√4ac-b²)
av²+bx +c = {và (x + 2 ) + √A2√a }
√4 ac-12
{ x+
2a 2 a
420.] Quadratic Equations. 255

In order that ax2 + bx + c may be zero, a must have such

a value as to make one of the factors zero, that is, a must
be a root of one or other of the simple equations
b √4 ac- 2
= 0,
√a (x + 2 ) -i 2√a
√4ac-62
or
√a(x + 2a)) + i 2√a
0.

Dividing these equations by va and transposing, the

values of æ are given by
b i√4ac-62
=
+ ,
2a 2a

b i√4ac-b²
and x =
-

2a 2a

The values of a are in this case (compare Art. 412)

operational or imaginary quantities. These values are
formally included in (a) of Art. 406 ; for since i² = -1
(Art. 271),
22
i√4ac - b² = √i² (4ac -b²)
= √ (4ac-b²)
= √b²- 4ac.
The forms (a) and (3) of Art. 406 are thus universal what-
ever be the sign or magnitude of b² - 4 ac.
420. Suppose, thirdly, that 2 - 4 ac has the value zero.
Then the relation
2
b 62-4ас
ax² + bx + c = a (x + 2 ) - 4a

reduces to
2
b
ax² + bx + c = a (x + 2 ) . a

The expression ax² + bx + c thus reduces to the single

term u². The two factors which in previous cases were

‫יין‬
256 Quadratic Equations. [421 .
different now become identical, and the two simple equa-
tions, either of which might satisfy, become identical
also, each of them reducing to
b

√ (x + 2 ) = 0 .
a

Whence there is deducible the one value of x,

b
х
2a

421. Since in the last case the expression ax² + bx + c

has two equal factors, it is usual to say that the equation
ax² + bx + c = 0
has two equal roots .
The condition
b2-4ac0
is equally the condition that the expression ax² + bx + c
shall be an exact square, and that the two roots of the
equation ax² + bx + c = 0
shall be equal.
In many applications of Algebra to Geometry and
Physics the student will discover the importance of the
difference between the statements that an equation has
only one root, and that an equation has two equal roots.
422. From the last few articles it is clear that if x1, x2
be the roots of the equation
ax² + bx + c = 0 ,
the expression ax2 + bx + c is identically equal to
а (х - х ) (х -х ).
For, in Art. 418 it is shown that if - 4 ac is positive,
ax² + bx + c
ас √b²- 4acl
= √ (x+ a)-- 2α 2 {√a(x + 2 ) + √ 2√a

b 1 √b² - 4ac
= √a
ax + √b -4ac) Va
ax + +
2α 2a 2a 2a
424.] Quadratic Equations. 257

b b b - 4 ас
=a{x- (- 20+ 2a
x-
2a ac)}
= a (x -x₁) (x - 2),
21 and 2 being the two values obtained for a in Art. 418.
Again, in Art. 419, when b² - 4 ac is negative,
ax² + bx + c
b i√4ac-b² i√4ac- b²)
=
2a }{ va(x + 2 ) 2√a
b i√4ac - b² b i√4ac-b²
=a{x- (- 20 2a -}} { x-
2a
+
2a

= a (x - x₁) (x -x₂),
where 21 and 22 are the values obtained for a in Art. 419.
If b²- 4 ac is zero, the two values of a are equal, and
b
each is It has been already shown that in this case
2a
62
ax² + bx + c = a
a(x + 2 )
= a (x- r₁)²,
where ₁ is the value of the single root.
But if a be equal to x1, (x -x1) (x-x₂) becomes (x-x1)2.
Hence in all cases, if x1 , x2 be the values of the roots of
the equation ax² + bx + c = 0, the expression ax² + bx + cis
identically equal to a (x -x1) (x-x2).
423. The converse proposition, that if ax² + bx + c be
identically equal to a(x-x1) (x- 2) then 21, 22 are the
roots of the equation ax2 + bx + c = 0, follows at once from
Arts. 51, 167.
424. By means of the last article it is possible to write
down a quadratic equation whose roots are any two given
numbers. For if these numbers be represented by 21 , 2
the quadratic equation required is
(x -x₁) (x -x2) = 0 ,
S
258 Quadratic Equations. [425.
or, multiplying out,
x²-(x1 + x2) X + X1X2 = 0 .
Thus the quadratic equation whose roots are 1 and 3 is
(x - 1) (x - 3) = 0 ,
or x²- 4x + 3 = 0.
This last result can be compared with Art. 404.
425. The relations (1) and (2) of Art. 414, namely,
b
X1 + x2 = a
,

C
,
X1X2 =a

which can also be deduced from the identity of Art. 422,

sometimes simplify the solution of a given quadratic equa-
tion.
Thus the equation
x² + 2001x - 2002 = 0

is obviously satisfied by the value x = 1. Hence one root

being unity and the product of the two roots being
-2002, it follows that the other root must be - 2002 .
The two roots are thus 1, – 2002 .
426. The same relations can be used to express any
symmetrical function of the two roots in terms of the
coefficients of the given equation.
A symmetrical function means an expression involving
the two quantities in such a manner that its value is not
2 2

altered when they are interchanged. Thus ₁₂ + x2, x₁² + x₂²,

3 3
xx + xx are instances of symmetrical functions of 21
and 22 .
b
427. Since a₁ + x = a
squaring these equal quan-
tities it follows that
62
² + 2₁₂ + ² = 2'
1
a
427.] Quadratic Equations. 259

2
62
or
x₁2 + x¸² = 22-2012
2

72 62-2ac
2
a2 a a2 (1)

Also, cubing the same quantities, it follows {Art. 122 (1)}

that
2
63
x₁³ + 3x₁2 x2 + 3x1 x2 + 2 = -
63
Whence x1 + x2 = 3 - 3x12(x1 + x2)
- =

3
-3 (- )
=
3bc63
a2 a3

3 abc-b3
=

a³ (2)

Again, by squaring the two equals in (1),

4 (b² - 2ac)2
1 + x2 + 2x12x2 = a

2
b - 4ab2c + 4a²²
Therefore x1 + x2 = 2
a4 a2

b4-4ab2c + 2a² c²
a+ (3)

In a similar manner the sum of any equal powers of the

two roots can be obtained, and by processes not more
difficult the values of any symmetrical functions of the
roots.

428. Any equation involving only two powers of the

unknown quantity, the index of one of which is double
that of the other, can be solved in the same manner as
a quadratic equation.
S2
260 Quadratic Equations. [428.
Thus if the equation be
x - 9x + 8 = 0,
the assumption x3 = y reduces it to
y²-9y + 8 = 0 ,
the roots of which by the processes previously explained
are found to be y = 1 and y = 8 .
Whence x³ = 1 or x³ = 8 .
The scalar values of a are therefore x = 1 , x = 2 , and
the operational ones are ω, ω², 2ω, 2ω², where w has the
meaning given to it in Art. 277 .
429. Again, let the equation be
1 13
+ 1=
6

Multiplying both equals by a and transposing, it all

becomes
13 1
х x + 1 = 0.
6

Assume x = y, then x = y², and it follows that

13
y2 6
y+1 = 0;

2 3
whence y = or 2

2 3
Hence x =
3 or 2
2

and x = (3) or ( ) ,
4 9
= or
9 4

430. Equations of a more complicated nature sometimes

occur, the solution of which can be reduced to that of two
quadratic equations. For instance, the equation
x² - 6x² + 10x² - 3x- 2 = 0
431.] Quadratic Equations. 261

can be written
4

(x² - 6x3 + 9x2) + (x² - 3x) - 2 = 0,

or
(x² - 3x)² + (x² - 3x) -2 = 0.
Thus x²- 3x is the only formal combination in which
x occurs in the equation. Assuming
x² - 3 x = y,
the equation becomes
y² + y- 2 = 0 ;
whence it easily follows that
y = -2 or 1 .
Hence a must satisfy one or other of the equations
x² - 3x = 2,
or x² - 3x = 1 .

The values of a derivable from these equations are re-

spectively x = 1 or x = 2

3 + √13
and
2

This method of simplification can often be adopted in

cases when the unknown quantity occurs in some one
formal combination only.
431. One or two instances of the reduction of equations
in which apparently surd forms occur will be useful to the
student.
As a first example let the equation be
x + √x + 1 = 5.
This may be advantageously solved in either of two
ways.
(1) Assume x + 1 = y².
Then, since x = y² - 1, the equation becomes
y² - 1 + y = 5,
262 Quadratic Equations. [432.
or y² + y- 6 = 0 ;
whence y = 2 or -3.
Hence x + 1 = 4 or 9,
therefore x = 3 or 8 .

(2) Transpose the term a to the other side so as to

have the surd form x + 1 alone on the left hand. Thus
√x + 1 = 5 - х.
Squaring these equal quantities, a process which is
merely an application of Axiom 3 of Arts. 53, 287, it
follows that
x + 1 = (5 - x)²
= 25-10x + x².
Whence, transposing,
x² - 11x + 24 = 0,
the solution of which gives
x = 3 or 8 .

432. It may be noticed that only one of the values of a

obtained in the last article, namely x = 3, is really a solu-
tion of the original equation. The other solution x = 8
satisfies the equation
x - √x + 1 = 5,
a fact which is indicated in the first method of solution
by the negative sign of one value of y.
It sometimes happens in similar cases that neither of
the roots obtained satisfies the given equation. Then if it
be required to solve the equation
x + 3 + 3√x + 1 = 0,
the solution conducted by either method gives the values
x = 0, x = 3 .
Neither of these values satisfies the given equation ; both
of them satisfy the relation
x + 3-3√x + 1 = 0.
433.] Quadratic Equations. 263

The given equation has in fact no solution possible, a state-

ment which must be entirely distinguished from that of
Art. 412 that the equation there discussed has imaginary,
or as they are often called impossible, roots.
433. As another example of reduction let the equation
be √x² + ax - 1 + √x² + bx - 1 = √a+ √b.
The student will notice at once that the value unity
put for a satisfies the equation. This observation will
simplify the solution later on.
The two surd forms involving a must be got rid of. To
do this two processes of squaring are required.
Transposing the terms so that one of these surd forms
stands alone on one side of the equation, we have
√x² + ax - 1- ( √a + √b) = -

x²+bx- 1 .
Squaring these two equal quantities (Axiom 3), it follows
that
2
x² + ax - 1-2 ( √a + √6) √x² + ax - 1 + (√a + b)²
= x² + bx- 1 .

Again, transposing the terms so that the single surd form

now left involving & may stand by itself on one side, the
terms x²- 1 disappear and the equation reduces to
(a− b) x + ( √a+ √5)² = 2 ( √a + √ ) √x² + ax - 1.
Since (Art. 259) a - b = ( √a- √ ) ( √a+ √), the ex-
pressions on both sides of this equation are divisible by
√a+ √b, and, effecting the division, it follows by Axiom 4
(Art. 53) that
(√a- √x + (√a + √b) = 2√x² + ax - 1 .
Squaring these equal quantities we obtain
(√a- √2x² + 2 (a − b) x + ( a + b)² = 4x² + 4ax - 4,
264 Quadratic Equations. [434.
or transposing,
x² (√2- √6)2-4} -2 (a + b) x + ( √a+ √b)² + 4 = 0.
a

This equation can be now solved by the ordinary process,

or the formula of Art. 406. As however it has been
observed that one of the roots is unity, and by Art. 414
the product of the two roots is ( a+ b)²+ 4 , it follows
(√ √6)2-4
without any further work that the two roots must be 1,
and (√a + √6)²)2 + 4 •

(√a - √6)2-4
b C
434. The relations x1 + x2 = a
,
X1X2 =

a
, and the

results of Art. 427, can be often used to discover equa-

tions whose roots are related in some given manner to
those of the original equation. Thus let it be required to
1 1
find the equation whose roots are ,
1

1 1 1 a2
Since 2 2 2
xx
1 01222
X

1 1 x² + x² 62-2 ас
and +
=
= ,

X1
2
X2
2
XX
1 2
2
C2 (Art. 427)
it follows that the equation required is
62-2 ас a2
x². x+ = 0,
C2 2

or
c² x² - (b² - 2 ac) x + a² = 0.
 Quadratic Equations. 265

EXAMPLES.
Solve the equations :-
1. x² + 16 = 17x. 2. x - 12x+ 27 = 0 .
3. x - 3x + 4 = 0 . 4. x²-x- 2 = 0.
5.5x² - 4x- 1 = 0. 6.7x² - 8x + 1 = 0.
7. 20x² - 41x + 20 = 0 . 8. 12x² - 23x = 77 .
7
9. (2-4) = 2x-2 10. - 2x = -9 . 2

11. abx²-(a² + b²) x + ab = 0.

2 13 2x- 1
12. 3
2+ = 13. + = 2.
6 3 2x- 1
1 1 9x- 1
14. 55
+ =
15.
x-2 x- 4 3 (x-3) 1 6
x-
X

1 1 1 x+ 1 x+2 13
16. + = 17.
х- а x-b +
х- с x+ 2 x+ 1 6
a² + b² 1 1 11
-
x-b + х-а
18.- ab
19.
x-a +x-ba + o
20. (x + a) (x + mb) _ (mx + a) (x + b)
(х-та) (x - b) (х -а) (mx - b)
21. a(c -d) + d(a − b) _b(c-d) + c (a−b)
=

x+ a x +d x+ b x+ c
22. x² - 5x² + 4 = 0.
23. √a+x + √a-x = 2√x.
24. 2x²-√x²- 2x- 3 = 4x + 9.
25. x + a = (x + a) .
26. √x− 3 + √x+ 4 = √7.
27- (x² + 8x + 16x- 1 ) -x = 3.
28. (x - 1) + (2-1) = x .
29. 2 (x+ a) {x-(x²-a²) } = a².
30. 6x²- 35x² + 6x² - 35x + 6 = 0.
31. (x-3) (x- 4) (x- 5) (x- 6) = 24.
266 Quadratic Equations.
1 5x+ 81xx - 9
6x+4)
32. 2-3 + (5x- 5x+ 1
=x + 3x-
+4+ ++
14 37 5

a 12x 16x a
33. +
a- 2x а- 3x 4x-aa- 6x

34.
(x + 1)(x + 1 + 0+1) +1 = 0.
35. (x + 1 )² = (x + 2) (x² + 2) .
4
36. ( -a) +( + ) = √3
37. If the equations x²- 7x + c = 0 and x² - 9x + 2c = 0
have a root in common, solve both of them.
38. Find the condition that the equations ax² + bx + c = 0
and px² + qx + r = 0 may have a common root.
39. If x²y = z(x + y- z) , prove that
y = z or yz = (x- 2)².
40. If a, ß be the two roots of the equation x² + px + q = 0,
form the equations whose roots are respectively
11
' ; a²,β²; and a²+β, β² + α.
41. If x1 , x, be the two roots of the equation
a β Y
+ + = 0,
a-x b-x c-x

where a ++ y = 1 ; and if y1, y2 be the corresponding

values of y derived from the equation
a² a Βεβεγ 6-

y= + +
a -x b-x с-х

then x₁ + y2 = a + b + c = x2 + Y₁ •
42. If the roots of the equation x² + px + q = 0 be a ± √B,
1 1
show that the equation whose roots are a +
√B
is (p² -4q) (p²x² + 4px) = 16q.
 Quadratic Equations. 267

43. If the roots of the equations

x² + p₁x + q = 0, x² + p₂x + q = 0, x² + px + q = 0
b a

be respectively a, b ; ma, -; , mb ;
m m

2 2

prove that p₁ ( Pr²- 4q) + (P2 + P3) q = Pi² Pa P3

44. If one root of the equation x² +px + q = 0 be a root
of the equation x² + ax + b = 0, its other root is a root of
x² + (2p-a)x +p² -ap + b = 0.
45. Find a number which shall exceed its square root by 110.
46. Two kinds of pears are sold in the market, two more of
one kind being given for a shilling than of the other; a score
of the inferior sort costs sixpence more than a dozen of the
superior sort. Find the price of the pears.
47. There is a number of two digits whose sum is 11 ; the
square of the smaller digit exceeds the larger digit by 9 ; find
thedigits.
48. The united ages of a father and son amount to 64. Twice
the father's age exceeds the square of the son's age by 8. Find
their ages.
49. The length of a rectangular field exceeds its breadth
by 33 yards. Its area is an acre : find the dimensions of
its sides.

50. A messenger starts from A at 8 a.m. to go to B; after

waiting an hour at B, he returns walking at one mile per hour
less than his pace in going, and arrives again at A at 4 p.m.
Had he walked half a mile an hour faster at first he would have
reached B at 20 min. to 11 a.m. Find his pace going and
returning, the time of reaching B, and the distance from A
toB.
 CHAPTER XV.

CUBE ROOT AND CUBIC EQUATIONS.

435. THE extraction of the cube root (Arts. 65, 67) of an

algebraical expression or number is not such a frequently
occurring or important process as that of finding the square
root which was explained in Chapter XIII. Some ex-
position of the method is however desirable.
436. The cube of a + b is known (Art. 122) to be
a³ + 3a2b + 3ab² + b³. The object of the present investi-
gation is to discover a method by which a + b may be
recovered from the latter expression supposed given .
437. The first term, a, is evidently the cube root of the
first term of the given expression. If the cube of a be
subtracted from this latter, the remainder is
3a²b + 3ab² +63.

The trial divisor which will give b is evidently 3a², that is

three times the square of the term already discovered.
It only remains to verify that b, the quotient obtained
by dividing the first term of the remainder by the trial
divisor, is a quantity such that (a+b)³ is equal to the
given expression.
438. To do this, various methods of building up the
expression 3a²b + 3ab² + b³ from the discovered values of a
and b are adopted.
The most usual and on the whole the easiest is to
calculate separately the term 3a², which is the trial divisor,
440.] Cube Root and Cubic Equations. 269

and the term 3a +b. Multiplying the latter by b and

adding to the former, the expression
3a² + ( 3a + b) 6 or 3a² + 3ab + 2
is obtained. This may be regarded as the complete divisor,
and the product of this into b, which is 3a²b + 3ab² + b³,
ought to be equal to the remainder of the original
expression.
439. The operation is usually written in the following
manner.

a³ + 3a2b + 3ab² +63 ( a + b

a3

3a2 3a²b + 3ab² + 63

+(3a + b) b
3a² + 3 ab + b² 3a²b + 3ab² +63

The student who has carefully followed the method of the

square root (Arts. 367-369) will have no difficulty in under-
standing the meaning and use of the present method.
440. As an example let it be required to find the cube
root of 8x3-36х2у + 54ху² - 27y³. The cube root of the
first term is 22. The trial divisor is consequently 3 (2x)²,
that is, 12x². The remainder after subtracting (2x)3
being -36x²y + 54 xy² - 27y³, the quotient of the first
term of this by the trial divisor is - 3y. Hence the cube
root is 2x- 3y. The process is indicated on p. 270 (*).
The values of a and b having been determined, the re-
mainder of the process consists in the calculation of
3a²b + 3ab² + 03.
When this is done and the result is subtracted from the
remainder of the given expression, it is found that nothing
is left.
Thus the process proves that the cube of 2x- 3y is
equal to the given expression, that is, that 2x- 3y is the
cube root of the latter.
 (*) 8x3
( 674yx
53y³
+
2x²y
-
xy²
823
270

a
3
y³
-
xy²
2x²y
-+
5-36
b= 74x3y
;1)²3(2x²
b
+
3a 3
-
6
=xy
b+
)3( a 9+
-18xy
=y2
=
6+
32ab
3a² 12x²
18
-
y²
+9
xy 3+(= 2x²y
-
xy²
7y³
a²b
64
5)b³ab²

(†) x®
1²505x²
6x3
-
2(xx²
1-
+x5
x

3
= x4
3a² 2
1
- 5x
+0x3
-6x
23-
,-x²
b=
+ax2x
4−+
b)= a6x3
3( x2
+
4-
6
= x
ab
b+²x²
33a² 8x3
1
+2x4
-6x5
-
Cube Root and Cubic Equations.

+
-
)²= 2x2
x
32(x3 ²2x3
13a2 1+
16
-
-5x²
3x x
2x3
+1(3a
'b)'= 3x²
6x
1-
+
6+
-
3x=
1+3bα'b
3a² 5x2
²x
'+42x3 +
1-
+
-x
3x²
65x²
2x3
[440.
442.] Cube Root and Cubic Equations. 271

441. At the end of the operations indicated in the last

two articles the cube of a + b has been subtracted. If
there be any terms left from the original expression, a + b
must be considered as the first part of the cube root, and
since the cube of a + b has been subtracted it occupies
exactly the same position as a did at the end of the first
stage.
Hence 3 (a + b)² must be taken as trial divisor and the
operations repeated in exactly the same manner.
442. As an example let it be required to find the cube
root of
x - 6x5 +15 x4 - 20x3 + 15x² - 6x + 1 .
The process is indicated on p. 270 (†).
The cube root of the first term is ² . The first trial
divisor is consequently 3x4 and the quotient of the first
term in the remainder by this is -2x, which must con-
sequently be taken as b. When the expression
3 a² + 3 ab + b²

has been formed and multiplied by b, and the product

subtracted, there is still a remainder.
At this stage, on the whole, the cube of x²- 2x has
been subtracted from the given expression. Consequently
x² - 2x occupies the position of a in the general investi-
gation of Arts. 437-439. In the above working it is
denoted by a'. The second trial divisor or 3a2 is easily
formed, and the next quotient is unity. Calling this b'
and forming the expression 3a^2 + 3a'b' + b², multiplying
this by b and subtracting the product, there is no re-
mainder.
Hence on the whole a'³ + 3a2b + 3a'b'² + b³ or (a + b)³
has been subtracted, and there being no remainder, the
process proves that the given expression is (a +b)³, or
(x² - 2x + 1)3 . Hence x²- 2x + 1 is the cube root re-
quired.
272 Cube Root and Cubic Equations. [443-
443. The second trial divisor is
3 (a + b)² or 3a² + 6 ab + 362.
In the operation of forming the first complete divisor
it will be noticed that the expressions (3a + b) b and
3a² + 3ab + b² are formed in succession. The sum of these
is 3a² + 6ab + 262. If to this sum bº be added, the result
is the second trial divisor. This is usually as easy a
method as any of deducing the trial divisor in any stage
from the results obtained in the previous one.
444. The application of this process to the calculation
of cube roots of numbers is easy after the full explanations
given in relation to the process of the square root in the
articles from 379 onwards.
The cube of 1 being 1 and that of 10 being 1000, it
follows that the cube root of any number between 1 and
1000 must lie between 1 and 10, that is, must have one
digit only. Since the cube of 100 is 1000000, it follows
similarly that the cube root of any number between 1000
and 1000000 must lie between 10 and 100, that is, must
have two digits. Hence it follows that the cube roots of
all numbers with one, two or three digits have one digit,
those of numbers with four, five or six have two digits, and
so on.

445. Again, if any number be represented in the form

Pox" + P₁xn - 1 + ... + Pn−1 x + Pn , (a), (Art. 380), where x has
the value 10, the cube of this number will assume the
form
2
Po3x3 + 3PoP₁x3n-1 + ....
Now 23 may be greater than 10 but must be less than
103, since po is less than 10. Hence if the highest power
of 10 in any number be 10", the highest power of 10 in
its cube may be either 10m, 103n +1, or 103n +2.
If then in any number whose cube root is required a
dot be placed over the unit figure, and in succession over
447.] Cube Root and Cubic Equations. 273

each third figure going to the left-hand, these dots will lie
over digits which multiply 103, 106, 109 ... , that is, powers
of 10 of the form 103m. The number of these dots will
thus indicate the number of digits in the cube root re-
quired, since the nth of them will lie over the digit which
multiplies 10° (n-1).
446. Again, po³ must be the principal part of the
coefficient of 23" in the cube. Owing to the process of
carrying it will probably not constitute the whole of this
coefficient. When the dots have been placed as suggested
in the last article, the number formed by the figure under
the left-hand dot and the figure or figures to the left of
it, is the coefficient of x³". Hence po will be the greatest
number whose cube is less than this coefficient. Thus
pox", the first term of the cube root required, is easily
found. The remaining terms must be discovered by a
process exactly equivalent to the algebraical one, the only
difference being that there is some little doubt as to the
second and third figures of the root, for a similar reason
to that which made the second figure in the square root
uncertain.

447. One example in illustration of this not very im-

portant subject will suffice. Let the number be 259694072 .
Placing dots over the unit and every third figure, there
are seen to be three figures in the cube root.
The first digit must be the number whose cube is nearest
below 259, that is 6, and the principal part of the cube
root is therefore 600. The first trial divisor is 3 x (600)2
or 1080000. The first digit in the quotient of the re-
mainder, after subtracting (600)3, by this trial divisor is 4,
but it will be found on forming the complete divisor with
this number that 4 is too large, and accordingly 3 must
be taken. The different steps in the process are indicated
pretty clearly in the following diagram :-
T
274 Cube Root and Cubic Equations. [447.
259694072 (600 + 30 + 8 = 638
216000000

3a² = 1080000 43694072

b = 30
3a + b = 1830
(3a + b) b = 54900
3a² + 3ab + b² = 1134900 34047000 = 3a²b + 3ab² + b³
3a² + 6ab + 26² = 1189800 9647072

b² = 900

2
3a 1190700
b= 8
30' + b' = 1898
(3α' + b') b' = 15184
3a² + 3a'b' + b² = 1205884 9647072 = 3a²b' + 3a'b'² + b'

A great number of unnecessary noughts have been

written down in order to make the principle of the process
clearer. With a little practice the student will discover
for himself how to omit needless figures.
448. Suppose that the whole number of digits in the
cube root is n+ 1, and let s + 2 of these have been found
by the above process. Let a represent the number formed
by these digits, and let b represent the remaining part of
the cube root. Then, since the whole cube root contains
n + 1 digits, the first of its digits is the coefficient of 10 " ;
while the first digit in the part denoted by b is the
coefficient of 10%-8-2, and b contains n -8-1 digits.
Hence a must be greater than 10", and b must be less
than 10-8-1. It is here assumed that any number with
pdigits, which consequently contains no higher power
than 10-1, must be less than 10". The Arithmetical
student who is familiar with the fact that 9999 is less
by unity than 10000, will have no difficulty in allowing
 449.] Cube Root and Cubic Equations. 275

the general truth of the above statement (see note, Article

382).
449. The remainder of the given number, after a has
been found and a³ subtracted, is 3a²b + 3ab² + b². The
trial divisor is 3a². The complete quotient of the re--
62
mainder by the trial divisor is b + a
+ 3a²
Now b is less than 10"-8-1
Hence 6º is less than 102n - 28-2
Also a is greater than 10".
62 102n- 28-2
Thus is less than or 10-28-2.
a 10"

Similarly, 3 is less than 1034-38-3, and a² is greater

than 102 .
103n-38-3
Therefore is less than or 10-38
a2 102n -3

62 63
Hence + is less than 10"-28-2 + 1.10-38-3, that
a 3a2
1 1

1 + 3.108+2 ) ,which is evidently

is, less than 10-28-1 (10
1 1
less than 10"-28-1, since + is a proper fraction.
10 3 x 108+2
62 33
Hence
a + 302, being less than 10"-28-1, contains
at most n - 28-1 digits, while b contains n - 8-1 , or
(n - 28-1) +8, digits. Thus the highest & digits of the
62
complete quotient, b + + , are the same as those
a 3a²
of b.

Thus when (8+ 2) digits have been found, the trial

divisor will give the next s digits accurately.
For instance, when three figures have been obtained the
trial divisor will give with certain accuracy the fourth.
T2
276 Cube Root and Cubic Equations. [450.
There will be a little necessity for guessing in ascertaining
the values of the second and third digits.
450. All the statements (Arts. 378, 393, 394, 397) made
about the approximate square roots of numbers apply,
mutatis mutandis, to approximations to cube roots.
The cube root of a fraction is obtained by finding the
cube root of the numerator and dividing by that of the
denominator. If the fraction be not an exact cube it is
more convenient to multiply its numerator and denomin-
ator by such a number that its denominator shall become
an exact cube. The approximate cube root of the numer-
ator can then be found, and this, divided by the cube root
of the denominator, is called the approximate cube root
of the fraction.
451. Any number can be expressed in the form of an
improper fraction with any power of 10 as denominator,
by the simple process of adding noughts to the number
and taking it so altered for the numerator. If the index
of the power of 10 be a multiple of 3 the denominator
will be an exact cube. The approximate value of the cube
root will thus be a fraction with a power of 10 as
denominator, and in this way an approximate value of
the cube root of a number can be obtained to any required
number of decimal places.
2000 2000000
2= = ,
Thus, 1000 1000000

and if the approximate cube roots of 2000, 2000000 be

obtained, these values, divided by 10 and 100 respectively,
will give the approximation to the cube root of 2 to one
and two places of decimals respectively (compare Arts.
397-399).
452. The solution of a cubic equation-that is, an equa-
tion with one unknown quantity, in which the unknown
quantity occurs to the third and no higher degree-can
454.] Cube Root and Cubic Equations. 277

be effected by the help of the previous articles and

Art. 282.
The most general form of a cubic equation is
x³ + ax² + bx + c = 0 ;
since by getting rid of fractions, transpositions, and
division of the whole equation by the coefficient of x³,
every equation involving the third and no higher power
can be reduced to this form.
453. The expression æ³ + ax² + bx +c can be reduced to
a

the form u³ + qu + r where u is equal to x + 3

and q and r
are constants.

For (x + 3) = 3 + 3x² + 3 9 + 27', Art. 122, (1)

a2 a3
= x² + ax² + 3 x + 27
Hence
a_3

3
a2 a3
x³+ ax² + bx + c = (x + 2) 3
х- + bx + c
27+
4_3

a3
= (x + 3) + ( - ) +c- 27

a3
= u³ +
+ ( - ) ( - ) +c-
ab 2a3
= x² + (b -

3 ) +c 3 + 27

= u³ + qu + r,
a2 ab 2a3
where q = b- and r = c- +
3 3 27

454. The solution of the general cubic equation will

therefore be effected if that of

u³ + qu + r = 0 (1)
is effected.
278 Cube Root and Cubic Equations. [454-
The latter problem depends on the resolution of the ex-
pression u³ + qu + r into factors, since (Art. 163) if a be
any value of u which satisfies (1), u-a is a factor of
u³ + qu + r ; and conversely if u- a be a factor of this last
expression, u = a is a solution of ( 1).
In Art. 282 it is shown that

x3 +y3 + 23-3 xyz = (x + y + z) (x + wy + w²z) (x + w²y + wz),

where w is one of the cube roots of unity.
In this identity, let -u be written for æ, and the sign
of each side be changed; it becomes
и³-
u3 - 3yzu -y³ - 23 = (u - y - z) (u - wy - w2z) (u - w2y - wz).
If then by a proper choice of y and z, the expression
u³ + qu + r can be made identical with u³ - 3 yzu - y³-23,
the resolution of the former into its factors and, conse-
quently, the solution of (1) will have been completely
effected.

455. The conditions of identity of the two expressions

are 3yz = -q,
-y³ - 23 = r;

93
whence y323 = 27'

y³ + 23 = -r.
Hence (Art. 414), it follows that y³ and z³ are the two
values of t in the quadratic equation
93 = 0.
t² + rt -

27

But these values are

2 2 3

- 2 ++
4 27
and 2 +L.
27

(Art. 406) (a)

457.] Cube Root and Cubic Equations. 279

Hence the values ofy and z are

r2
3
r r

(- ++ ) and (--
2 4 2 4 +27)
456. The three factors of the expression u³ + qu + r
being, with the above values of y and z, u-y-z,
u - w²y - wz, u - wy- w²z, it follows that the three roots
of the equation ( 1) are y +z, w²y + wz and wy+w²z.
Hence the equation is completely solved.
457. If 72 + I be a positive quantity, the values of y
4 27

and z are scalar quantities. Hence, in this case, one root

of (1) is scalar and two are operational.
7.2
If + 93 be negative, y and z are both operational.
4 27

Nevertheless all three values of u in this case are scalar.

For y³ and 23 assume (Art. 271) the respective forms
a + bi and a - bi, which, by Art. 273, can be written
in the forms r (c + is) and r (c- is). Hence y and z, or
(a + bi) and (a-bi) can be written as rs(c+ is) and
rs (c- is) respectively. Now c+ is denotes (Art. 273) a
rotation through some angle which we may call a, c-is
denotes a rotation through an equal angle in the opposite
direction. The multipliers (c + is) and (c-is) must denote
a

rotations through angles of one third the amount, or 3, also

in opposite directions. Hence (c + is) can be replaced by
same multiplier c' + is', where c' and s' are geometrically
determinable functions of cand s, and (c- is) can be re-
placed by c'-is' .
Hence y + z = rs (c' + is) + rs (c'- is)
= 2crd,
which is scalar.
-1 + i√3 -1 - i√3
Again (Art. 277), ω = 2
,
ω
2
280 Cube Root and Cubic Equations. [
457-
Hence
wy+ ω²z
=rt (c+is) (=1+ √3) + ( - 8) (11-13)
2 2

= r . -c+ci√3-18-813
2

-c-c'i√3 + is' -8' √3

+ .
2

= -ro (c + 8√3),
which is also scalar.
Similarly, the expression w²y + wz is reallyscalar, although
like the other two roots it assumes an operational form.
458. The actual deduction of the scalar values of the
2

roots of the cubic in the case when 4

+ I is negative, is
27

beyond the powers of elementaryAlgebra, and must be left

to the more advanced subjects of Trigonometry and Theory
ofEquations.
459. It is scarcely necessary to remark that the solution
of any equation in which only two powers of the unknown
quantity occur, the index of one of which is three times
that of the other, can be made to depend on that of a
cubic equation.
Thus the equation
4x + 3x - 7 = 0
becomes, by the assumption x = u,
4u3+ 3u - 7 = 0,
3 7
or u³ + 4 - 4 = 0.

Comparing this latter equation with (1) ofArt. 454,

3 7
==
4 4
459. ] Cube Root and Cubic Equations. 281

Hence

= ++ 4
= + 8
50 = 7+5√2
64 8

3
r 9.2 7 50 7-5/2
23 +
2 4 27 8 64 8

The student can easily verify that these are satisfied by

the values
1+ √2 1-2
y= and z= ;
2 2

whence y + z = 1, which is the only scalar value of x : the

two operational roots can easily be written down.
It may be noticed that, in this case, the deduction of the
value of y and z from those of y³ and z³ is as difficult an
operation as that of the solution of the original equation. It
canin fact be reduced to depend on the latter. In any case
2

in which
4 + I is not an exact square a similar difficulty
27

is met with, so that for practical purposes the general

formula is of very little value.

EXAMPLES.

Find the cube roots of :

1. 27 x² + 135x²y + 225xy² + 125y .

2.64a³- 48a²b + 12ab² - 6.
3.8x - 12x + 6x7-37x + 36x - 9x + 54x² - 27x² - 27 .
4.8x -36x + 54x² + 21x - 144x² + 108x4
+ 96x - 144x² + 64 .
5.8x - 12x² - 6x + 23x - 9x - 12x² + 11x - 3x + 1 .
6. x² - 6x²y + 3x²z + 12xy² + 3xz² - 8 y³
12y²z- 6yz² + z3-12xyz.
7. 1879080904, 66775173193, 258474853 .
282 Cube Root and Cubic Equations.

Solve the equations :

8. x² - 2x + 4 = 0.

9. x - 3x² + 5x- 3 = 0.
10. x² + 2x² - 3x - 6 = 0.
11. If each of the equations
x² + 3px² + qx + r = 0, x² + 3px² + qx + r' = 0
have two equal roots and the third common to the other
equation, the unequal roots not being the same in the two
equations, show that
q-d r-r 3
3
3 (p -p') = = (p +p') and that 99 = r.
p+p' r+ r 9pp
'
12. If a, b, c, be the roots of the equation
x² + 3px² + 3qx + r = 0 ,
prove that those of
(x+p) (x² + 3px² + 3qx + r) - (x² + 2px + q)² = 0
bc + ca w + ab w² bc + ca w² + αφω
are ,

a+ b + cw² a+ bw² + cω
where w is one of the imaginary cube roots of unity.
13. If the equations ax³ + 3bx² + d = 0, and bx + 3dx + e = 0
have a common root, prove that
(ae- 4bd)³ = 27 (ad² + b²e)² .
14. Prove that if
X = ax + cy + bz, Y = cx + by + az, Z = bx + ay + cz,
(a³ + b³ + c³- 3abc) (x + y + z³ - 3xyz) = X + Y + Z - 3 XYZ:
 CHAPTER XVI .

SIMULTANEOUS EQUATIONS OF A HIGHER DEGREE THAN

THE FIRST.

460. THE solution of two equations containing two

unknown quantities to a higher degree than the first,
involves, as a general process, the solution of a single
equation, with one unknown, of a higher degree than any
within the province of elementary Algebra. There are,
however, many pairs of equations in which, by particular
artifices, the solution can be made to depend on equations
similar to those which have been already considered. The
present chapter will be devoted to the consideration of
some of the most commonly occurring forms.
461. The primary object to be attained is usually the
elimination of one of the unknown quantities and the
deduction of a single equation involving only the other.
Sometimes it is advantageous to express one or both of the
given unknowns in terms of some other quantities, the
values of which may be more easily determinable in the
first instance, from which those of the original unknowns
can be deduced.
462. The first class of equations for which a general
method of elimination can be suggested, consists of those in
which one of the unknowns occurs, in one of the equations,
only to the first degree.
In such a case, the value of this unknown can be
determined from this equation in terms of the other ; this
284 Simultaneous Equations higher than the First. [463.
value being substituted in the other equation, there results
an equation with only one unknown. If it so happen
that this final equation is either a quadratic, or reducible
to one, the solution can be completed.
463. As an example, suppose the pair of given equations
tobe
3x² + xy = 18 , (1)
2xy-y² = 3, (2)
Here y occurs in (1) and x in (2) only to the first degree.
The method may be applied either by finding the value
of y from (1) and substituting in (2) ; or the value
of x from (2) and substituting in (1). Taking the latter
plan ; it follows from (2) that
2xy = y² +3,
y² +3
whence x
(3)
2y
Hence from (1),
y² +3
3 ( 2y+3) + 2y
y = 18 ,
2
or, multiplying these equals by 4y², so as to get rid
of fractional terms,
3 (y++ 6y2 + 9) + 2y² (y² + 3 ) = 72y² ;
whence, transposing and collecting terms,
5y4 - 48y² + 27 = 0.
This equation, involving only y² and y¹, can be solved as
a quadratic as y² (Art. 428). By the formulae of Art. 406
it follows that

48 + √(18)² - 4 × 5 × 27 48+42
y2 10 10

= 9 or .

Hence y = +3 or ± √3. 5.

Corresponding to each value of y, (3) gives one value of x.

465.] Simultaneous Equations higher than the First. 285
The four pairs of values are therefore
9+ 3
y= 3, = = 2,
6

9+ 3
=
y = -3, = -2,
-6
3
+3
y =√ , x=
2√
= 3√ ,

+3
y= -√ , x= = -3 .
2

464. The second class of equations for which a general

method can be suggested are those in which the terms
containing the unknown quantities are homogeneous
(Art. 88), or nearly so, as regards those quantities. The
equations solved in the last article are a case in point.
The process in this case consists in assuming a new
variable v connected with the two original unknowns
y
x and y by the relation v =
, or
oryy = να. If, in the two

given equations, væ be substituted for y, the condition of

homogeneity will be found to make it easy to eliminate a
and to deduce a single equation in the unknown quantity v.
When v can be determined from this, either of the original
equations will give the corresponding values of æ ; and,
v and a being known, væ or y is easily derived.
465. Taking again the equations ofArt. 463, namely
3x² + xy = 18,
2xy-y² = 3,
the substitution of væ for y reduces them to
3x² + να² = 18,
2x² - v²x² = 3 ,
or to x² ( 3 + v) = 18 ,
α² ( 2υ - ν²) = 3 .
286 Simultaneous Equations higher than the First. [465.
From the first of these it follows that
18
x2 ,
3+ 0
and from the second that
3
22
20-02
whence, equating the two values of x²,
18 3
,

3+v 20-02

an equation involving only the unknown v.

Getting rid of fractions, and dividing the two equal
quantities by 3, the equation reduces to
6 ( 2υ - ν²) = 3 + v ;
which, by transposition and reduction, becomes
602-110 + 3 = 0 .
The roots of this, by Art. 406 , are
11 + 121-72 11 + 7 18 4
or
12 12 12 12
3
= or .
2

18 18
3
If v = 2 , x² = =
3
= 4,
3+ v 3+

whence x = +2,
and y = vx = x2 = ± 3 .
18 18 27
2 ,
If v = 1 , x² = = =

5
3 +v 3+

whence ± 3√ ;
x = +3
and y = vx = ± x 3 =+
= ±√
The solutions obtained by this method are of course
identical with those given by the former.
466. A third class of equations consists of such as are
symmetrical in form with respect to the two unknown
quantities ; that is, which remain unaltered when a is
changed into y and y into x.
468.] Simultaneous Equations higher than the First. 287
In this class it is often useful to make the substitutions
x +y = u,
ху = v.

By means of processes such as those given in Art. 427

the given equations, being symmetrical with respect to x
and y can be expressed in terms of u and v. It is frequently
much easier to solve the resulting equations with respect
to u and v than the given ones in x and y.
Suppose that from these equations the values u = a,
v = β are deduced. Then, since x +y = a, and xy = ẞ, it
follows from Art. 414 that x and y are the two values of
t in the quadratic equation
t - at + β = 0 .
Thus x and y can be completely determined.
467. As an example of this class of equations let us take
the pair
x² + x²y² +y² = 133,
x² -ху +y² = 7 .
Making the assumptions x +y = u, xy = v, it follows that
u² = x² +y² + 2xy,
whence x² + y² = น²- 2 v.
Squaring these equals
x² + 2x²y² +y = u² - 4 u² v + 402,
whence, since x²y² = v², subtracting these equals from the
former two, it follows that
x* + x²y² +y = u² - 4 u²v + 3 v² .
Hence the original equations become
κ²- 4 κ² + 30² = 133,
²- 30 = 7 .

468. In the latter of these equations v occurs to the

first degree only. Hence the method of Art. 462 can
beused.
288 Simultaneous Equations higher than the First. [468.
u²- 7
Substituting for v in the first equation its value 3

derived from the second, we obtain

24 4 и² (и² - 7) + ( 2-7)2 = 133,
3 3

or, multiplying both sides by 3, and multiplying out,

3u4-4u4 + 28 и² + u² - 14u² + 49 = 399,
whence 14u² = 350,
or u² = 25,
whence u = ± 5.
u2-7 25-7
Hence v, or , = = 6.
3 3

Taking u = 5, v = 6, x and y are the two roots of the

equation
t - 5t + 6 = 0 ;

that is, x and yhave the values 2 and 3, x being equal to 2

and y to 3, or x to 3 and y to 2.
If u = -5 and v = 6, x and y are the roots of the
equation
t + 5t + 6 = 0 ,
where x and y are - 2 and 3. On the whole there are
four pairs of values of æ and y, namely x = 2, y = 3 ;
x= 3, y= 2 ; x = −2 , y = − 3 ; x = −3 , y = -2 .
469. The equations of the last article can also be
solved by the method of Art. 464. They can be solved
perhaps even more simply by noticing that
x² + x²y² + y² = (x² + xy + y²) (x² -ху + y²),
whence the first equation gives, by the help of the second,
x² + xy +y² = 19 .
Also, since x²-xy +y² = 7 ,
it follows by subtraction that
2xy= 12,
or
xy= 6.
471.] Simultaneous Equations higher than the First. 289
Hence, adding ay to the first equation,
x² + 2 xy +y² = 25,
whence x +y = ± 5 ,
and the rest of the solution canbe conducted as in the last
article.
470. The solution of simultaneous equations involving
more than two unknown quantities must be effected by
similar methods to those which have been indicated in the
case of two unknowns. General rules are less to be relied
on than special artifices, which can only be learnt by practice
and close attention to the results of algebraical transforma-
tions.
471. The elimination of one unknown between two
equations each of the second degree leads in general to an
equation of the fourth degree.
The most general form of an equation in x and y of the
second degree may be represented by
ax² + by² + 2hxy + 2gx + 2fy + c = 0 .
The second equation may be taken as
a'x² + b'y² + 2h' xy + 2g'x + 2f'y + c = 0.
These equations may be written in the forms
by² + 2 (hx +f)y + ax² + 2gx + c = 0 ,
b'y² + 2 (h'x +f') y + a'x² + 2g'x + c = 0 .
or by² + Py + Q = 0 , (1)
V'y² + P'y + Q' = 0, (2)
where P = 2 (hx +f), Q = ax² + 2gx + c and P' and Q'
represent similar expressions in the second equation.
Multiplying (1 ) by band (2) by b, and subtracting these
results (bP -bP') y + bQ - bQ′ = 0. (3)
U
290 Simultaneous Equations higher than the First. [471.
Multiplying (1) by Q' and (2) by Q and subtracting, it
follows that

(bQ' − b'Q) y² + (PQ' −P'Q) y = 0 ;

or dividing by y,
(bQ' - b' Q) y + (PQ′− P'Q) = 0. (4)
Equating the values of y obtained from (3) and (4), we
obtain
VQ - bQ_PQ -P'Q .
=

P -6P Q- Q
whence, getting rid of fractions and transposing,
(PQ' - PQ) ( P - bP") + (b'Q - bQ')² = 0 . (5)
Now P, P' are expressions of the first degree in æ, and
Q, Q' are of the second. Hence PQ' - P'Q is of the third
degree, UP - 6P of the first, and b Q - bQ' of the second
degree. Thus the equation (5) will be of the fourth
degree.
The student will easily verify for himself that equation
(5) can be written in the determinant notation as
[b´P-bP' VQ-BQ' = 0 ;
and he will recognise that each of the constituents of
this determinant is itself a determinant.
Further acquaintance with the theory of determinants
will show him that the whole process might be effected by
the investigation of a determinant of the fourth order,
which can easily be reduced to the above determinant of
the second order.
472. In one particular case the result has a special
interest, namely, when the equations are of the simplified
forms
ax² + by² + 2hxy = 0,
α'x² + b'y² + 2hxy = 0 .
473.] Simultaneous Equations higher than the First. 291
Here P = 2hx, Q = ax², P' = 2h'x, Q' = a'x², and the
equation (5) reduces to
(2hx xa'x² - 2 h'x xax²)(2hb'x - 2h'bx) + (ab'x² -a'bx²)2 = 0,
whence, dividing by a¹, we obtain
4(ah - ah') (bh - bh') + (al' - a'b)² = 0. (a)
This is therefore the condition that the two given equa-
tions may be satisfied by any common values of x and y
differing from zero.
Each of the two equations may be written as a quadratic
equation in , and the equation (a) is the condition that
х

they shall be satisfied by a commonvalue of ". 0

473. It is theoretically possible to eliminate any number

of quantities, as n, from a set of equations, provided the
number of the latter exceed by unity the number of the
quantities to be eliminated. The general method is to
find the values of the n quantities from n of the equations
and substitute these values in the remaining one. There

thus results a relation independent of the n quantities.

For example, let it be required to eliminate x and y from
the equations
x + y = a, x³ + y³ = b³, x5 + y = c .
The first two equations will give by the method of
Art. 466
a³ - b³
x +y = а, ху ;
За

whence x and y can be determined and their values

substituted in the third equation. It happens, however,
that x +y can be readily expressed (Art. 427), in terms of
xy and x +y, and the further investigation of the values of
x and y is unnecessary.
U2
292 Simultaneous Equations higher than the First. [473-
Thus, x5 + y = (x² + y²) (x³ + y³) -x²y² (x +y)
= { (x + y)² - 2xy} b³ -x²y²a
= a
2 (a³ - b³) ( 3-13)
-a
3a 9a2

Thus the final result is

3 3
2
C5 = 63 (a³ - b³) ,

3a 9a
3
or
9ac5 = 313 (a³ + 263) - (a³ - b³)2
= 50 + 5b3a³ - a6 .
For the solution of problems of this class the student
must learn to rely on special artifices, only to be acquired
by great practice and considerable familiarity with differ-
ent algebraical expressions and their transformations and
relations to each other.
474. Problems are frequently proposed similar to those
which have been solved in Arts. 299-304 and 324-330,
the solution of which depends on that of equations of a
higher order than the first. There is no difference in the
principle of the methods to be adopted from those formerly
described. The conditions of the question must be carefully
stated in algebraical language, one or more of the unknown
numbers required being represented by the letters x, y, and
so on. The resulting equations must then be solved by
such of the previous methods, or such special artifices as
may be most applicable.

EXAMPLES .

Solve the equations :

1. x² + 3xy = 34, xy + 4y² = 110.
2. x + y = 7, x² + y² = 25.
3. 3x- 4y = 2, 4x² -ху = 14.
4. x² + ху = 12, xy+ y² = 4.
 Simultaneous Equations higher than the First. 293
5. x² -ху = 6, ху-у² = 2 .
6. x² + 3y = 7, y² + 3x = 7.
7. 4x² + 7xy = 78, 5xy + 9y² = 66.
х- у x+ 3y
8. x- = 4, y- = 1.
2 x+ 2

9. x² + y² = 65, xy = 28 .
10. x² + xy = 12, ху- 2y² = 1 .
2
x²y
11. + X
= 18, x + y = 12.
y

12. х -у = 2, ху+ 5x- 6y = 120 .

13. x (y + 1 ) = , y (x + 1) =
1 1
14. 3x- 4, 9x² + = 40.
y y
a

15. ax + by = c, +
+ = d.
X y

16. x + y = a, x² + y² = b².

17. z - y = 12, +9 X
= 34.
18. x² + y² = a² = xy (x² + y²).
19.
{°x²yx²++y²
xy² = 180,
= 189 .

20.
{ x²-ху + y² = 13,
+ y² + xy = 103.

21. 16ху,
(x − y ) (x² - y²) = 640x²y².
22. xy = c(x + y), zx = b (x + z), yz = a( y + z).
23. x² -yz + a² = y² -zx + b² = x² -xy + c² = (x² + y² + z²).
x + 3y + 5 3x + y + 4
+ = 4,
24. x+ y+ 1 x + y- 1
{(x + 2y)² + (y + 2x)² = 5 (x + y)² + 4y.
294 Simultaneous Equations higher than the First.

25. (x² + y²) + (2xy) + √(x² + y²) - (2xy) = √37,

x+ y = 287.

26. { (x+ y)(x+ 2) = a²,(z + x) (y+ z) (y + x) = b ,

(z + y) = c².
27. a²b-2x-1yz¯¹ = xy¯z¯¹ = a^2b^2xyz = 1 .
28. ху = 8, yz = 28, zx = 14.
a 62 (a+ b)c
=
+ ,

29.
x+a y+ ba + b +c
α² ²
+- = c.
Y

ax + cy- bz_by + az - cx_cz + bx- ay = ,

30. a 62 C2

x² + y + z² = bc + ca + ab.
x + y + z = a + b + c,
31.
x² + y² + z² = a² + b² + c²,
++ = 3.
C

32.
{(4y-3x)(x+y)
(3-5) (x − y) ==2,
++)
(x - 1 ) (x - 2).
a ( y + z-x) = (x + y + z)² - 2by,
33.
{c(x+
b (z + x− y) = (x + y + z)²- 2cz,
y-z) = (x + y - z)² - 2ax.
34.
y+ z_z + x x + y_x² + y² + z²
a b C a² + b² + c²
x + y + z = 0,
35.
{
x² + y² + z² = 2 (a² + b² + c² -ab -ac - bc),
x + y + z = 3 (b −c) (c - a) (a− b).
36. Eliminate x, y, z from the equations
x + y + z = a,
x² + y² + z² = b² + c²,
x + y + z = 3abc,
xyz = - a³.
 Simultaneous Equations higher than the First. 295
Solve the equations :
х

37. ++
++
2
= 3+ = ++ ,
X 2
y
x + y + z = 40.

38. Eliminate x and y from the equations

x² + xy +y² = a², x² +x²y² + y² = b , x² + x*y* +y = c .
39. Eliminate x and y from the equations
x² -xy + y² = a², x -x²y² + y² = b , x -x²y + y = c³.
40. Eliminate x, y, z from the equations
x- z = a (x-y),
1 1 1

X 2

=()
41. A person spends five shillings in the purchase of eggs.
He spends part of the money on one kind and part on another.
The number of pence he spends on the cheaper kind added to
the number of eggs of the dearer kind is 52. He gets as
many for a shilling of the dearer as he does of the cheaper, and
he has twenty more of the cheaper than of the dearer eggs.
How many does he buy of each kind ?
42. A number consists of three digits whose sum is 6. If
the digits be reversed the number is increased by 99. The
sum of the squares of the digits is 14. Find the number.
43. A regiment is drawn up in marching order. If the first
ten ranks be taken off and the remaining men be arranged so
that there is one fewer in each rank, but the same number of
ranks as before, there will be ten men left out. The number
of ranks exceeds the square of the number of men in each rank
by 19. Find the number of men and their arrangement.
44. Two persons, A and B, run a race to go five times round
acertain course. When A has gone three laps, B is 150 yards
behind him. A then slackens speed and goes at B's rate, while
296 Simultaneous Equations higher than the First..
Bquickens his rate and goes at A's first rate. A wins by 30
yards. Find the length round the course, and compare the
original speeds of A andB.
45. The sum of three numbers is 6, the sum of their squares
is 14, and the sum of their cubes is 36. Find the numbers.
46. Three persons, A, B, C, are engaged to do a certain work.
When working together they can do it in 6 days. If A only
work as hard as B, while C works as before, it will take them
84 working days to finish. The sum of the number of days
which A and C would respectively take to do it alone is double
the number which B would take by himself. Find in how
many days each of them will separately do the work.
47. The owner of £20,000 divides it into three portions, which
he puts out at different rates of interest. The total interest
he receives is £950. Three times the interest on the first por-
tion is equal to five times that on the second and also to ten
times that on the third. The sum of the rates per cent on the
last two portions is double that on the first. Twice the first
rate added to the third is equal to four times the rate on the
second portion. Find the sums of money and the rates per
cent.

48. There are three numbers whose sum is 54. The square
of the middle one exceeds the product of the other two by 4,
and the sum of the first and last is double the middle one.
Find the numbers.
49. A cistern can be filled by three pipes A, B, C. Twice
as much water passes through B per minute as through A.
The three when open together fill the cistern in one hour, and
B alone will take one hour longer to fill it than Calone. Find
the time in which each alone will fill the cistern.
 SECTION III .- SERIES.

CHAPTER XVII.

PERMUTATIONS AND COMBINATIONS .

475. THE third section of this book will be devoted to

the subject of Series, that is, to the consideration of the
properties of algebraical expressions consisting of a number
of terms formed according to some law, the value of each
term depending in some manner on its position in the
series.
In some of the series which will have to be considered,
the coefficients depend on the number of arrangements
which can be formed out of the different letters involved.
It is therefore necessary in the first place to investigate
the arithmetical formulae which determine this number in
different cases.
476. The number of possible arrangements of three
quantities among themselves has already been incidentally
shown to be 6 (Arts. 121, 344), that is, 3 x 2 x 1. The
number of arrangements of n things among themselves can
similarly be shown to be the continued product of all
integers, beginning with n and going down to unity, or, as
it may be written, n (n - 1 ) (n - 2 ) ... 3.2.1 .
477. This product occurs so frequently in mathematical
investigations that it has received both a name and an
appropriate symbol. It is usually called factorial n, and
denoted by the symbol n.
298 Permutations and Combinations. [477-
Thus factorial 6, or 6, means 6.5.4.3.2.1 , or 720 ;
factorial 5, or 5, means 5.4.3.2.1 or 120.
From the law of formation it is clear that | 6 = 65.
More generally, n = n (n - 1 ) (n - 2 ) ... 3.2.1,
n + 1 = (n + 1 ) n ( n - 1 ) (n - 2) ... 3.2.1 .
n + 1 thus contains all the factors of in and one factor,
n + 1 , additional. Hence
| n + 1 = (n + 1) n.
If in this last result n have the value zero given to it,
there appears the anomalous result
|1 = 0 .
Now the value of 1 is obviously unity, while the sym-
bol 10 has at present no meaning. It will be convenient
to give it such a meaning as to make the relation
| n + 1 = (n + 1 ) ❘n as general as possible, and it is there-
fore usual to consider 10 as merely a symbol for unity.
It will be seen later on (Arts. 546, 550), that this assump-
tion renders slightly more general certain forms obtained
in the multiplication of binomials and polynomials.
478. In recent years the notationin has by some writer's
been replaced by n !. The latter symbol has the disad-
vantage that the second part of it is in most minds already
associated with other ideas, and a page in which such signs
frequently occur gives at first sight an impression of
astonishment beyond even that due to the most brilliant
achievements of the mathematicians who introduced it.
The notation in will accordingly be exclusively adopted
here.
479. Suppose that there are n pupils in a class among
whom a number of prizes have to be distributed. As the
first prize may be awarded to any one of the pupils, there
are a different ways of giving the first prize.
481.] Permutations and Combinations. 299

When the first prize has been assigned to a particular

pupil A, there are n- 1 pupils left to contend for the
second. Thus, A having the first prize, there are (n - 1)
different ways of giving the second prize. Similarly if B,
another pupil, have the first prize, there are n - 1 ways of
giving the second, and so on for each of the n pupils.
Hence the total number of different selections of a first
anda second prizeman is n times n - 1 or n (n - 1 ) ; different,
that is to say, in one at least of the two prizemen, each
from every other.
Again, when the first two prizes have been assigned in
any particular manner there are n-2 pupils left as
competitors for the third prize, and consequently (n- 2)
different ways ofawarding it. Hence, since the number of
different ways of awarding the first two prizes is n (n - 1 ),
and with each of these ways there are n - 2 different ways
of awarding the third, there must be n (n- 1 ) times (n - 2)
ways of awarding the first three prizes ; or the number of
choices of three out of n is n (n - 1)(n - 2).
480. The student will notice that this method of reckon-
ing counts ABC as a different choice from BAC, or CBA,
or any other arrangement of the three pupils denoted by
ABC.
The number of different choices when this is the case is
called the number of permutations of n things, in this case
3 together, or if r be the number chosen, r together.
481. It has been shown that the number of choices of
three things out of n, the order of choice being a matter of
importance, as in the case of choosing a first, a second, and
athird prizeman out of a class ofn, is n (n - 1) (n - 2). When
any set of these have been chosen, there are (n- 3) left, and
consequently (n - 3) ways of choosing a fourth. Hence,
under the condition stated in italics, the number of different
300 Permutations and Combinations. [481 .
ways of choosing 4 things out of n is n (n - 1 )(n - 2) times
(n - 3), or n (n - 1)(n - 2)(n - 3).
Thus the number of choices is so far represented by the
product of as many factors as there are things to be chosen,
the first factor being n and each factor being less by unity
than the previous one.
482. It is obvious that this law will continue to hold
for the number of choices r together, r being any integer
less than n.
For let P, denote the number of distinct choices of r
things out of n. When any particular set of r things have
been taken, there are (n - r) left, and the number of ways
in which one more thing can be chosen and placed at the
end of this set of r things is (n - r). Hence the total
number of ways of choosing ( +1) things, that is, one
more than r, is „P, times (n - r), or nP, (n - r).
Hence Pr + 1 = nPr . (n - r). (1)
This shows that in deriving the number of permutations
of n things (r + 1) together from the number r together,
one additional factor is introduced. Further, asr increases
by unity, this additional factor (n - r) decreases by unity.
Thus for each additional thing chosen the number of per-
mutations has a new factor less by unity than the last
factor in the previous number. Thus the number of
factors being at first the same as the number of things
chosen, must continue to be so, and the successive factors
diminish by unity.
Also (1) shows that the last and least factor in „Pr+ 1 is
(n - r). Hence the last and least factor in „P, must exceed
this by unity or be (n - r + 1).
Thus, finally, it follows that
„P, = n (n - 1) (n - 2) ... (n - r + 1),
T (2)
where the dots represent the intermediate factors, which
 2haw in touc "

with Ana face

485.] Permutations and Combinations. 301

cannot be written down without specialising the value

of r.
483. The formula (1) of the last article can be used in
a slightly different manner to deduce (2).
InArt. 481 it has been shown that the number of permu-
tations of n things 4 together is n(n - 1) (n - 2) (n - 3 ), or
nP4 = n(n - 1)(n - 2) (n - 3).
Hence by (1), nPs = P (n- 4)
= n(n - 1 ) (n - 2) (n - 3) (n - 4).
Similarly,
„P = P (n - 5) = n ( n - 1 ) (n - 2) (n - 3) (n - 4) (n - 5 ),
6

and so on.
Thus each additional thing chosen introduces a new
factor less by unity than the last factor of the former pro-
duct. The number of factors and the number of things
chosen increase therefore pari passu, and being equal at
first must remain equal. Also the second factor is n - 1 ,
the third n - 2, the fourth n - 3 , and so on. Hence the
th factor, which is the last in „P , must be n - r + 1 .
Thus, as before,
„P, = n (n - 1) ( n - 2 ) ... (n - r + 1).
484. A particular and very important case is that of the
total number of distinct arrangements of then things.
This is evidently the number of permutations of n things,
n together, and will therefore be obtained from „P, by
giving tor the value n. The number required is therefore
the product of n factors beginning with n and diminishing
from factor to factor by unity. The last or nth factor must
therefore be unity, and the required number of arrange-
ments is n (n - 1 ) (n - 2) ... 3.2.1, or the product which
has been denoted by the symbol n.
485. Suppose that from a body, such as a Town Council
consisting of n persons, it is required to select a committee
for a special purpose consisting of r persons. The number
302 Permutations and Combinations. [485.
of distinct ways in which this can be effected is evidently
quite a different thing from the number of distinct choices
on the supposition hitherto adopted, that the order of
choice is an important element.
For instance, if three persons have to be chosen, the
number of permutations is by the previous investigation
n(n - 1) (n - 2). This reckoning counts all the different
arrangements of any one set of three persons, A, B, C,
as distinct. In whatever order these three may be chosen
they will, however, still form the same committee, and
this set of three must therefore in the new problem only be
counted once.
The number of orders in which this same set of three
persons would be chosen is by the last article 3.2.1 , or
13. Hence each distinct committee of three persons fur-
nishes 3 different permutations. The whole number of
permutations, three together, must be therefore 3 times the
number of distinet committees. The number of the latter
1
must therefore be of the number of permutations, or
13

mustbe
n(n - 1) (n - 2)
3

486. The number of distinct choices of rout of things,

when the order of choice is not considered, is called the
number of condirations of a things together. It is often
denoted by the symbol„C..
487. As a particular illustration, let there be five letters
a. 1, c, d, e, and let it be required to find the number of per-
mutations and combinations of these three together. The
combinations are easily seen to be

This comprises all that can be found. the first six being
 い "

488.] Permutations and Combinations. 303

all that contain a and two of the other four letters, the next
three all that contain b and two of the letters which follow
b, and the last being the only remaining possibility.
The whole number of permutations can be found by
arranging each of these combinations in all possible ways.
The permutations will thus be
abc, abd, abe, acd, ace, ade, bcd, bce, bde, cde,
acb, adb, aeb, adc, aec, aed, bdc, bec, bed, ced,
bca, bad, bea, cad, cea, dae, cbd, cbe, dbe, dce,
bac, bda, bae, cda, cae, dea, cdb, ceb, deb, dec,
cab, dab, eab, dac, eac, ead, dbc, ebc, ebd, ecd,
cba, dba, eba, dca, eca, eda, deb, ecb, edb, edc.
There are obviously 3, or six, times as many permu-
tations as combinations. The number of permutations
being 5.4.3, that of the combinations must be one-sixth
5.4.3
ofthis, or 3
488. Suppose now that any one combination of r out of
n letters is represented by abc ... k. The letters which form
this combination can be arranged among themselves in r
ways (Art. 484), each of which arrangements will count as
a different permutation.
If all the C,different combinations be similarly treated,
each of them will yield r permutations. All the permu-
tations given by one combination will be different from all
those given by any other combination, since they must
differ in at least one letter. Hence the „C, combinations
producer times that number of different permutations.
But since every possible permutation of n things r
together does contain some particular combination of r
letters out of the n, the above process must produce all the
possible permutations. Thus r „C, must be the number
304 Permutations and Combinations . [488,
of permutations of n things together. By Art. 483
this number is n(n- 1) (n - 2) ... (n - r + 1).
Hence
[r . „C, = n (n - 1) (n - 2) ... (n - r + 1) ;
whence C = n (n - 1 ) (n - 2) ... (n- r + 1)
r

489. The proof of the last article may be put in rather

a different point of view.
Supposing all the possible permutations of n letters r
together to be written down, it is clear that any one per-
mutation being taken, a large number of others can be
found which contain the same combination of r letters.
The number of permutations containing this particular
set of r letters is evidently the number of different ways
in which these r letters can be arranged, or r. Let the
r permutations containing the particular combination be
set aside.
Any other permutation being taken, which must contain
a different combination from all of the former ones, T

permutations will be found containing this second com-

bination. In this manner the permutations may be
arranged in sets, every permutation of each set containing
the same combination of r letters, and those in one set
having a different combination from those in all the
other sets.

Thus the whole number of permutations may be arranged

in as many sets as there are different combinations, each
set containing ❘r permutations. Hence the number of
permutations is ❘r times that ofcombinations, or
P = r.nC ;
whence, as before,
C= P. n (n - 1)(n - 2)...(n + 1) (Art. 482).
‫تا‬
=

r
492.] Permutations and Combinations. 305

490. The formula for „C, can be put into a more con-
cise form. Multiplying both numerator and denominator
of the fraction by the product of all integers beginning
with n - r, that is, a number less by unity than the least
factor in the numerator, and going down to unity, the
value of the fraction is unaltered, and

C =
n (n - 1) (n - 2) ... (n - r + 1 ) . (n - r) ... 3.2.1
r . (n - r) (n - r - 1) ... 3.2.1
The numerator is now the product of all integers from n
down to unity inclusive, or is n (Art. 477). The product
of the factors introduced may be written as n - r, and the
n

value of C, becomes r n- r

491. The result of the last article shows that the number
ofcombinations of n things together is the same as the
number n - r together.
n n

For Cn-r = nCr.

n- rr
n - rn- (n - r)
This can however be even more easily seen from the
consideration that for every different set of r letters taken
away from n, there is a different set of n - r letters left.
Hence the number of distinct sets of r letters out of n
must be the same as the number of distinct sets of (n- r)
letters.
492. The fractional form which represents the value of
„C. contains an equal number, r, of factors in numerator
and denominator. Each additional letter taken in the
combination introduces an additional factor in both numer-
ator and denominator.

Thus C-1 = n (n - 1 ) ... (n + 2) ,

1.2.3 ... (r - 1)
X
306 Permutations and Combinations. [493-

there being -1 factors in both numerator and denom-

inator.

nCr =
n (n - 1 ) ... (n - r + 2). (n - r + 1) ,

1.2.3 ... (r - 1) .r
an additional factor (n - r + 1) being introduced in the
numerator, and an additional factor r in the denominator.
n- r + 1
Hence C= nr-1 (1)
T

493. The result of the last article is of considerable

importance. An independent proof of it follows.
Let abcd ... h represent any combination of r - 1 letters
out of n. There are consequently {n-(r - 1 ) }, or (n -r + 1),
letters which are not included in this combination. By
placing each of these letters in succession along with the
given combination of (r- 1) letters, (n-r + 1) combinations
of r letters will be formed.
From each of the different combinations of r - 1 things
the same number of combinations of r could be formed.
On the whole there would thus be produced (n-r + 1) . „Cr-1
such combinations.
These would not all however be different. Each com-
bination of r letters, as abc ... hk, would appear r times, by
the combination of each letter with the other r- 1 . Thus,
on the whole, the process would producer times the number
of combinations of n things together ; or it would follow
that
(n -r + 1) .nCr-1 = r.nCr ,
n- r + 1
whence C= •nCr-1
T

494. This formula, obtained as in the last article, will

give another proof of the number of combinations of n
things r together.
495.] Permutations and Combinations. 307

For replacing byr- 1 , -2, &c., in succession, it follows

that
n- r + 2
C -1 = 7-1
nCr-2,

n- r + 3
Cr-2 = 7-2
nCr-31

n- 2
C3 = 3
nC2 ,
n- 1
nC2 =

2
nC1,

and C₁ evidently is n.
Hence, by successive substitution,
n-r+ 1 n -r + 2 n-r+ 3
C =
r 1-1 r- 2
...
n-2. -1.n ;
3

which is the same as the formula of Arts. 488, 489, the

factors being written in an opposite order.
495. The numbers of the combinations of n things taken
one, two, three, and so on, together, form a series (Art. 475),
the terms of which are n, n (n
1.2
- 1), n (n 1.2.3
- 1) (n - 2), and
so on.

The rth term of this series „C, is derived from ther - 1th
n- r + 1
term C-1 by multiplying the latter by the factor r

(Arts. 492, 493).

Hence for all values of r for which this factor is greater than
unity, the rth term of the series is greater than ther - 1th ;
on the other hand, for all values of r for which this factor
is less than unity, the 7th term is less than the r - 1th.
If for any value of r, as p, this factor has the value unity,
the pth term will be equal to the p - 1th.
X2
308 Permutations and Combinations . [496.
If r have the values 1, 2, 3, 4 ... n, given to it in suc-
n- r + 1
cession, the numerator of the fraction r
continually
decreases fromn to 1 ; the denominator, on the other hand,
increases from 1 to n. In virtue of both these facts the
value of the fraction decreases, having, to begin with, the
n 1
value and ending with the value Thus for thesmaller
1 n

values of r the multiplier by which „C, is derived from

n

C -1 is greater than unity, and C, is consequently greater

than C -1, while for the latter and larger values of r the
reverse is the case. The terms in the series under con-
sideration increase in the first part and decrease in the
latter part of the series.
496. Let p be a quantity such that the fraction
n -p + 1
= 1 ; it follows that n -p + 1 = p , or n + 1 = 2 p,
P
n+ 1
whence p = 2
Thus, if n be an odd number, and n + 1
consequently even, p is an integer and, as above explained,
the pth term is equal to the p - 1th. Also when r is less
than p, the considerations of the last article show that the
n- r+ 1 n -p + 1
factor is greater than , or unity, and thus
r
p

for all values of r less than p, nC, is greater than C.-1 ;

while similar reasoning shows that for values of r greater
than p, C, is less than „Cr-1. Hence, when n is odd, the
number of combinations of a things r together increases
asr increases from 1 to (n + 1), remains unchanged as r
passes to ( n + 1), and decreases as assumes successively
greater values.
497. If, on the other hand, n be even, no integral value
n -r+ 1
of r makes equal to unity. Ilence no two terms
r
498.] Permutations and Combinations. 309

of the whole series are in this case equal. By the last two
articles it follows that if r be any integer less than p, that
n+ 1 n 1
is, 2
or
2 +2', this multiplier is greater than unity.
n
Hence the terms of the series increase, until ris 2
After
that point they will decrease, since the next and all suc-
n+ 1
ceeding integral values of r, being greater than 2
orp,
n- r + 1
must make the multiplier r
less than unity. Hence
n

the terms of the series increase until r = 2, forwhich value

the term is greatest, and then decrease.
498. Whether n be even or odd the terms during the
latter or decreasing part of the series are equal respectively
to those in the former or increasing part in a reverse order.
(Art. 491.)
As an instance of the last article, suppose n to be 6. The
terms of the series are

6.5 6.5.4 6.5.4.3 6.5.4.3.2 6.5.4.3.2.1

6, ,
;
1.2 1.2.3' 1.2.3.4 ' 1.2.3.4.5' 1.2.3.4.5.6
or
6, 15, 20, 15, 6, 1.
Here the third term is the greatest, and the terms increase
up to that point and then decrease.
Again, if n be 7, the terms of the series are
7.6 7.6.5 7.6.5.4 7.6.5.4.3 7.6.5.4.3.2
7, , , , ,

1.2 1.2.3 1.2.3.4 1.2.3.4.5 ' 1.2.3.4.5.6

7.6.5.4.3.2.1
1.2.3.4.5.6.7
or 7, 21 , 35, 35, 21, 7, 1.
Here the third and fourth terms are equal, the terms of
the series increasing up to the third and decreasing from
310 Permutations and Combinations. [499.
the fourth onward. This result agrees with that of
Art. 496.
499. A problem of great importance in many applica-
tions of this subject is that of the number of distinguishable
permutations, all together, of n things which are not all
unlike one another.
Suppose that there are n letters, of which p are a's, q are
b's, r are c's, and the rest are distinct from each other and
from these. Let any permutations of these letters be
represented by
aabbaabccabcdaef. (1)
Suppose that each of the different letters represented by
a has a distinguishing number affixed to it. This permu-
tation will then be written
aabbaza bcca bcda ef, (2)
where the a's are now distinguishable. By altering the
positions of the pa's among themselves without changing
those of the other letters, p distinguishable permutations
can be found out of this single one. Thus if there were
a distinct permutations, when the a's were all alike, there
will be ptimes that number, or x . p when the a's are
made distinguishable.
Suppose that (2) represents any one of this latter set of
permutations, and let q letters o be made distinguishable-
by numerical suffixes : the permutation (2) can then be
written
aabba a bocca bacdaef. (3)
By interchanging the qb's among themselves without
altering the positions of the other letters, this one permu-
tation will give a permutations. Thus the c . p permu-
tation will give q terms that number, or, when both the
a's and b's are distinguishable among themselves, there
willbe x . p q distinct permutations.
500.] Permutations and Combinations. 311

Similarly, if there's be made distinguishable by suffixes,

each of these will give❘r permutations. Thus, on the
whole, there will be xp 9 permutations when all the
letters are distinct from each other.
But on this latter supposition the number is known to
ben (Art. 484). Hence
xpqr = n;

whence x=
r
PIT
500. The result of the last article may be obtained in a
somewhat different manner.
The number of arrangements required is the same as the
number of distinguishable ways in which n things whereof
pare a's, q are b's, r are c's, and the rest unlike, can be
arranged in n places. The number of different sets of p
places that can be selected for the pa's is the number of
combinations of n things p together, or (Art. 490) is
n

pn -p

There are now (n -p) places left unoccupied, and q of

n -p
these can be selected for the q b's in ways .
qn -p - q
Hence the number of different ways in which the pa's and
q b's can be placed in p + q of the places is the product of
n
n -p
these two numbers, or X ,that is,
p n -p q n -p - q
n

is
Pqn - P - 9
There are now (n -p- q) places left vacant, and r of
n -p- q
these can be taken for there's in ways.
r
n - p-q-r
312 Permutations and Combinations. [501 .
Hence the total number of distinguishable ways of assigning
(p + q + r) places to the pa's, q b's and rc's is the product
of this into the former number, or
n
n-p- q
X
P9 n -p-q r n - p -q-r ;
n

that is, r
PIP n -p - q - r

If all the remaining letters be unlike, they can be

distributed in the remaining (n -p -q- r) places in
n -p - q - r different ways. Hence the total number of
distinguishable arrangements is
In n

xn -p- q-r , or
r
P9rn- p - q - r P91

501. The student will easily see that, by either method

of proof, if there had been s quantities of another kind
alike, an additional factor 8 would be introduced in the de-
nominator of the fraction, and the number of permutations
n

would be
r 8
PIPS
502. There is an almost infinite number of problems of
a similar nature to those which have been discussed in this
chapter. The preceding articles contain those which are
most important for algebraical purposes, and the principles
which have been employed will serve as a guide to the
methods to be employed in any similar questions of some-
what greater arithmetical difficulty.
 Permutations and Combinations. 313

EXAMPLES. 6.2 .

1. Find the number of words of three letters that can be

formed out of the English Alphabet.
2. In how many different ways can three prizes be awarded 1.

to a class of fourteen scholars ?

3. How many committees of four can be formed from a

council consisting of 13 members ?
4. A Board of Education consists of nine members. In
how many different ways can a chairman and vice-chairman be
selected ?

5. In how many ways can a party of twelve be selected 7

from a company of 100 soldiers ? In how many of these will )
a particular soldier be found ? ' 2
6. How many words each containing one vowel and two
consonants can be made out of the letters of the word number ?
How many will there be if the vowel is to occupy the
middle place ?
7. Prove that the greatest number of combinations that can
be formed with 2n things, each containing the same number,
is always double the greatest number that can be formed with
2n- 1 things.
8. If the number of combinations of 2n things taken n - 1
together be the number of combinations of 2(n - 1) things
taken n together as 132 is to 35, find n.
9. If the number of permutations of 2n things 3 together is
equal to twice the number of permutations of n things 4 together,
find n.

10. If the greatest number of combinations of n things r

together be of the greatest number of combinations of
n- 1 things, find n.
314 Permutations and Combinations.

11. If m denote the number of combinations of n things

taken 2 together, prove that the number of combinations of m
things taken 2 together is equal to three times the number of
combinations of n + 1 things taken 4 together.
12. In how many different ways can a party of six people
form a ring ?
13. Find the number of arrangements of n people at a round
table ; one of them being supposed to occupy a particular chair.
14. A gentleman invites a party of m + n friends to dinner,
and places m at one table and n at another, both tables being
round. Find the number of ways in which he can arrange
them among themselves.
15. Out of six ladies and eight gentlemen, how many different
parties can be formed, each consisting of three ladies and four
gentlemen ?
16. Find the number of committees that can be formed out
of a House of Representatives containing 45 Liberals and 50
Conservatives, each committee to have 9 Liberals and 10 Con-
servatives .

17. What is the number of distinct arrangements of the

letters in the word precipitate ?
18. Find the same thing for the words Mississippi, Papa-
toitoi, Ngahauranga.
19. How many words containing two vowels and three con-
sonants can be formed out of 21 consonants and 5 vowels ?
How many will there be if the vowels are to occupy the even
places ?
20. Prove that 2n = 1.3.5.7 ... (2n - 1) . 2 " n . Hence
show that the number of combinations of 2n things n together is
1.3.5 ... (2n - 1) .2"
n

21. Prove that the number ofways in which p positive signs

and n negative signs can be placed in a row so that no two
 Permutations and Combinations. 315

negative signs shall be together is equal to the number of com-

binations of ( +1) things taken n together.
22. If C, denote the number of combinations of n things
r together, prove independently that
r.nC = n.n-1 Cr-1 .
From 6 ladies and 5 gentlemen in how many ways could
you arrange sides for a game of croquet, so that there should
be two ladies and one gentleman on each side ?
23. If the permutations of the things a₁ , a2, a3 , ... , an be taken
all together, and Q, be the number of ways in which no one
of the suffixes indicates the place which the corresponding
thing holds in the permutations, show that
Qn+1 = n ( n + On-1).
55 55
24. If Cr+1 264 +1Cr =

40-1
C,, find nand r.
 CHAPTER XVIII .

ARITHMETICAL AND HARMONICAL PROGRESSION .

503. A SERIES of terms such that each is greater, or less,

than the preceding by a constant difference, is called an
Arithmetical Progression.
The first term and the magnitude and sign of the common
difference between each term and the preceding being given,
it is obviously possible to determine by repeated additions
or subtractions the value of any succeeding term. It will be
shown that it is also possible to express the sum of any
number of terms in terms of the same two quantities and
the number of terms.
504. The series
1 + 2 + 3 + 4 + ...

is an arithmetical progression, the first term being unity

and the common difference also unity.
The series
3+2 +2 +1 + ...

is also an arithmetical progression, the first term being 3

and the common difference- . The common difference
is always taken to be the algebraical quantity obtainedby
subtracting any term from the next succeeding term.
505. If the first term be denoted by a, and the common
difference by d, the second term is a + d, the third a + 2d,
and so on. The series of terms is represented as below.
2, 5, 6
1, 3, 4, ......

a, a + d, a + 2d, a + 3d, a + 4d, a + 5d......

506.] Arithmetical and Harmonical Progression. 317
The number in the upper of these two lines is the number
of that term in the series which is written below.
The coefficient of d increases by unity in passing from
term to term, since each term is obtained by adding d to
the previous one. The number which denotes the term
also similarly increases by unity from term to term. In
the terms written down, the coefficient of d is less by unity
than the number which denotes the position of the term
in the series. As these two numbers increase at the same
rate, the difference between them must remain always the
same. Hence in the nth term the coefficient of d must be
n - 1 , and the nth term must be a + (n - 1) d. If this nth
term be denoted by l, we have
l = a + (n - 1 ) d ; (1)
whence it also follows that
a = l- (n - 1 ) d.
506. If s denote the sum of the first n terms of the
series,
s = a + (a + d) + (a + 2d) + ... + { a + (n - 1) d}.
Writing the terms in the reverse order, and remembering
that as any term is obtained by adding d to the one before,
so it can also be obtained by subtracting d from the
following one,
s = 1+ (l-d) + (1-2d) + ... + { l- (n - 1)
Adding the equal quantities on both sides of these two
equations, it follows that
2s = (a + 1) + (a + 1) + (a + 1) + ... to n terms
= n (a + 1).
n
Hence 8 = 2 (a + 1) ; (2)
or, since l = a + (n - 1) d,
n

8 = { 2a +(n- 1) d} . (3)
318 Arithmetical and Harmonical Progression. [507.
507. The formulae (1) and (3) suffice for the solution of
all problems relating to arithmetical progression. Such
problems are of two classes. In the first class, the values
of a and d are given, and by means of (1) and (3) every
other required result about the progression can be deter-
mined. In the second class, some two conditions which the
progression is to satisfy are stated : from these conditions,
by means of (1) and (3), the values of a and d can always
be obtained, and then any further result can be investigated
as in the former class. 1

508. An important problem of the first class is that of

the determination of the number of terms of a given
progression whose sum is a given quantity.
Here a, d are known since the progression is given, and
8 is also known ; n is required.
From (3), multiplying out,

s = na + n (n - 1) d
2

= na + nad_nd
2 2

d d
= n
n (a -

) + . n².
2

From this equation two values of n can be determined

by the ordinary process of solving a quadratic equation.
If either value of n be a positive integer this gives the
required number of terms. A fractional or negative value
of n is irrelevant, regarded as a solution of the problem
considered, though an interpretation may be found for
such a value by seeking some other problem to which it
applies.
509. Let it, for instance, be required to find how many
terms of the series 11 + 9 + 7 + ... amount to 32. If n be
510.] Arithmetical and Harmonical Progression. 319
the required number, the formula (3) of Art. 506 gives,
since a = 11 , d = -2 , 8 = 32,

32 = { 22+ ( n - 1 ) ( - 2)}

= 11n - n (n - 1 )
= 12n - n² ,
whence n² - 12n + 32 = 0 ;
which gives n = 4 or 8 .
Here both values are admissible. The first four terms
are 11 , 9, 7, 5, the sum of which is 32 ; while the first eight
terms are 11, 9, 7, 5, 3, 1, -1 , -3, the sum of which is
also 32.
510. As an example of the second class of problems, let
it be required to insert n arithmetical means between two
given quantities a and b. The problem means, to find n
quantities x1, x2 , ... , such that the whole set, a, x1 ,
X2, X3 , ... Xn , b shall be in arithmetical progression.
Here a being the first term, b is evidently the (n + 2)th.
Hence if d be the common difference

b = a + (n + 2-1) d, by (1),
= a + (n + 1) d ;
b- a
whence d= (a)
n+1

Hence d is known, and the value of x1, x2 can be

determined by (1). Thus æ, being the r + 1th term of the
series,
x, = a + rd

= a +
r(b -a)
n+ 1

(n - r + 1 ) a + rb
n+ 1
(β)
320 Arithmetical and Harmonical Progression. [511.
511. A special case of the last article is the discovery of
a single arithmetic mean between a and 6, that is, a quantity
x such that a, a, b are in arithmetic progression.
The formula (a) and (3) of the last article, putting
b- a a+b
n = 1 , give d = and x =
2 2

This latter result can be obtained more easily from the

consideration that if a, a, b be in arithmetic progression,
it follows from the definition of Art. 503 that

x - a = b -x ;
whence transposing,
2x = a + b,
a+ b
or x=
2

512. As another problem of the same class, let it be

given that the pth term of an arithmetic progression is
and the qth term is y, required the nth term and the sum
of n terms. Here p, q, x, y are supposed to be known
quantities.
Let a be the first term and d the common difference of
the progression. The pth term will be a + (p- 1) d, and
the qth term a + (g - 1) d.
Hence, by the given conditions,
a + (p - 1) d = x,
a + (q - 1 ) d = y.
From these two equations the two unknown quantities
a and d can be found. Subtracting the second from the
first,
(p - q) d = x -y ;
whence d= ".
P- 9
 513.] Arithmetical and Harmonical Progression. 321
Substituting in the first equation
a + (p- 1) - =x;
P- 9

whence a = x-(p - 1) -
P- 9

=
x (p -q) -(p - 1)x + (p - 1 ) y
P- 9

= (p - 1)y-(q- 1)
P- 9

Hence the nth term, which, by (1), is a + (n - 1) d,

-(9-1) + (n- 1)P-- 9
= (p -1) P-q
= (p-n)yP-9
+ (n- q)x ;
and, by (3), the sum of n terms can similarly be deter-
mined.
513. The nth term of any series is conveniently denoted
by a symbol such as un, the suffix indicating the position
of the particular term in the series. In all cases of series
proceeding according to any law, un is a function of n
(Arts. 165, 475).
In the case of an arithmetical progression, the nth term
has been shown to be a + n- 1d. Hence in this case,
un = a + (n - 1 ) d
= (a- d) + nd.
Thus un , regarded as a function of n, is an expression
of the first degree, or a linear expression in n (Arts. 88,
139).
Conversely, an arithmetic series can always be discovered
whose nth term shall be represented by any expression of
Y
322 Arithmetical and Harmonical Progression. [514.
the first degree in n. For let the given expression be
p + qn. Then if d be taken equal to q and a -d to p, this
becomes identical with (a -d) + nd.
Hence a-d = p
d = q.
Therefore a = p + q.

The required series has therefore p + q for its first term,

and q for the common difference between consecutive
terms.

514. The sum of n terms of a series is conveniently

denoted by sη . Hence in the case of an arithmetic
progression
d
8 = (a - 2 n + n². (Art. 508)

Thus the sum of n terms of an arithmetic progression is

an expression of the second degree in n (Art. 139).
Conversely, an arithmetic series can always be found such
that the sum of n terms shall be represented by any
expression of the form pn + qn² .
For let a and d be taken so as to satisfy the two
conditions
d
2
= q

d
a- = p;
2

whence a = p + q, d = 2q.
2
d
Then
pn + qu² = (a ) n + n²,
2

or pn + qn² is the sum of n terms of an arithmetic

progression whose first term is a or p + q and common
difference d or 29.
517.] Arithmetical and Harmonical Progression. 323
515. The subject of Harmonical Progression is closely
connected with that of Arithmetical Progression. Before
entering on it a preliminary definition must be given of
the term ' reciprocal.'
Two numbers whose product is unity are called reci-
procal numbers and each is said to be the reciprocal of the
other.
Thus if ab = 1 , a is the reciprocal of 6, and b is that
of a.
1
From the relations ab = 1, it follows that a = and
1
b = Thus the reciprocal of a may also be defined as
a

1
being-. a

1 4 5

As examples the reciprocal of2 is 2

5,' and that ofis 4

516. A series of numbers is said to be in harmonical

progression when their reciprocals are in arithmetical
progression.
Thus all problems relating to quantities in harmonical
progression can be solved by taking the reciprocals of
these quantities and using the formulae relating to arith-
metical progression.
517. The discovery of n harmonical means between two
given quantities a and b depends on the investigation of n
1 1
arithmetical means between and
a চ
If the harmonic means be called 21, 22, 23 , ... 2 , so that
the (n + 2) numbers A, 21, 22, 23... Zn, b,
are inharmonical progression, the (n + 2) numbers
1111 11
,, ...
a Zn
Z1 Z2 Z3
will be in arithmetical progression.
Y 2
324 Arithmetical and Harmonical Progression. [518.
If d be the common difference of this latter, by ( 1) of
Art. 505, 1
= + (n + 1) d,
a

1 1 1
whence d=
n + 16
1
Hence, since Z is the r+ 1th term of the series,
1 1
+ rd
Zr a

r 1 1
= +
a

(n - r + 1) 6 + ra
(n + 1)ab
Hence (n + 1)ab
+ = (n - r + 1 ) b + ra
518. The particular case of inserting one harmonical
mean z between two quantities a and b may be either de-
duced from the last article by putting n = 1 , r = 1 , whence
2ab
2 =
a +b ; or may be independently investigated thus. 1 1 1
If a, a, b are in harmonic progression, ㅎㅎㅎ are in
arithmetic progression. Hence by the definition of Art. 503
1 1 1 1
- =
2 a 6

2
whence
= +;a

and, getting rid of fractions,

2 ab = z (a + b),
=
2ah
or
a+ b

519. There is no general formula for the sum of a terms

of a harmonical progression.
 Arithmetical and Harmonical Progression . 325

EXAMPLES.

Find the nth term and the sum of n terms in each of the
following series :-
1.1 +++ .... 2.1 + 3 + 5 + ....
3.2 + 2 + 3 + .... 4. + 1 + 2 + ....
5.11 + 9 + 7 + ....
6.15 + 2 + 16 + ....
7.4 + 3 + 2 + .... 8.1 + 1 + 0 + ....
9.0 + 1 + 2 + ....

10. (a + b)² + (a² + b²) + (a − b)² + ....

11. (n - 1) + (n-2) + (n - 3) + ......
a b 2b- a
12. + + ....
a- b + a-
a- b a-b
1 a + 1
13 . + + ....
a-b a²-b² a+ b
14.13 + 9 + 5 + .... 15. - 5-3-2 -....
16. - 2 + 10 + 6 + ....
17. Find how many terms of a series, whose first term is
unity and third term is 5, amount to 169.
18. The fourth term of an arithmetical progression is 7,
the eighth term is 15; find the nth term and the sum of n
terms.

19. The third term of an arithmetical progression is e, and

the fifth term f; find the thirteenth term and the sum of
n terms.

Find the nth term in each of the harmonical series :-

20.1 +++ .... 21.1 + 1 + 2 + ....
22. +++ .... 23.1 +4-2 .
24. 1, 2, 0, -2 .
25. The first term of a harmonical progression is unity, the
third term is ; find the tenth term.
326 Arithmetical and Harmonical Progression.

26. The first term of a harmonical progression is unity, the

sum of the first three terms is 11,
6,
find the progression.
27. How many terms of the series 11 + 9 + 7 + ... amount
to 20 ?

28. If the mth and nth terms of an arithmetic progression

be M and N, find the pth term.
29. If 2x²+ yz then z is a harmonic mean between
,

Y 2y² + zx
x and y.
1
30. Find 17 harmonic means between 1 and To
а-х a-y a- z
31. If ‫و‬ and p, q, r be in arithmetical
12
px qY
progression, then x, y, z are in harmonical progression.
32. If s be the sum of any number of terms of the series
1 + 2 + 3 + ... , prove that 88+ 1 is always a square. How
many terms of the series must be taken to make 21 ?
33. Find the arithmetic series whose fourth term is 3, and
the sum of seven terms is 21 .

34. S1 , S2, S, are the sums of three series of n terms each

in arithmetical progression, the first term of each being unity
and their common differences in harmonical progression ; prove
that
2SS -S, S -S2 S3
n =

S₁+S - 2S
35. A and B go round the world; A goes east one mile the
first day, two the second, and so on in arithmetical pro-
gression. B goes west at the uniform rate of twenty miles
aday. Find when they will meet, and interpret the negative
answer. The circuit of the world being 23661 miles.
 CHAPTER XIX.

GEOMETRICAL PROGRESSION .

520. A SERIES of numbers such that each is equal to the

preceding multiplied by a constant factor is called a series
ofnumbers in Geometrical Progression.
Anticipating here the definition of the ratio of a number
a to another number 6 (Art. 644) as being measured by the
a

fraction , the common factor is the ratio of each term to

the preceding one. Hence this factor is usually called the
common ratio of the series.
521. Let a represent the first term, and the common
ratio. Then since each term is obtained by multiplying
the preceding term by r, the second term is ar, the third ar²,
the fourth ar³, and so on. The index of r increases by
unity from term to term and is always less by unity than
the number of the term. Hence the nth term is ar"-1 ; or,
denoting the nth term by un ,
Un = arn-1. (1)
522. Let the sum of the first n terms be denoted by sn.
Hence 8 = a + ar + ar² + ar³ +
n
... +arn-1.
Then rsn = ar + ar² + ar³ + ... +arn- 1 + ar" .
Subtracting the second of these equations from the first,
all the terms on the right hand except the first in the first
row and the last in the second row disappear, and
8. ( 1-1) = a - ar",
1-
or 8 = a
1- r
(2)
If r be less than unity, rs, is less than s„ , and the sub-
traction gives a positive result on each side. If r be greater
328 Geometrical Progression. [523 .
than unity it would be more convenient to subtract the
upper equation from the lower, and the result would be
sn (r - 1) = ar" -a,
r"-1
or Sn a
r- 1
(3)
The form (2) is more convenient when ris less, and the
-1
form (3) when ris greater than unity, but since is
r-1
1-
algebraically equal to 1- r , either form maybe safely used
in either case.
523. The formulae of the last article give the means of
solving any problem involving a finite number of quantities
in geometrical progression. (Compare Art. 507.)
Let it be required, for instance, to insert n geometrical
means between a and b ; that is, to find n quantities
Y1 , Y2 Y3 , ... Yn such that the whole series
AY1 , Y2 , Y3 , Y4 , ... Yn, b 12

may be in geometrical progression.

If r be the common ratio, b is the n + 2th term of the
series, and must therefore, by (1), be equal to arn+ 1.
Hence b = ar* +1 ;
b
therefore rn + 1 = - ,
a

6n + 1
or ↑ = ,

whence yp, the pth of the means which is the (p + 1)th term
of the series, is given by the equation
P

6n + 1
Yp = a ( ) a

n-p +1 p
n+1
= a +1 ‫و‬

= (an-p + 1 ZP)n + 1 (a)

526.] Geometrical Progression. 329

524. The particular case of the insertion of one geome-

trical mean y between a and b, may be deduced from (a) of
the last article by putting nand peach equal to unity. We
thus get y = (ab) . (β)
The same result may be obtained independently from the
consideration that if a, y, b are in geometrical progression
the quotient of each term by the preceding must be the
same (Art. 520). Whence in this case
b
y
a
y'
or, multiplying these equals by ay, ‫له‬

y = ab,
whence y = (ab) .
525. In Arts. 511 and 518 it has been shown that if
x and z be the arithmetical and harmonical means be-
tween a and 6,
a+ b
,
2

2ab
2=
a+ b
a+ b 2 ab
Hence (Art. 524)
xa+ 6 = ab = y²
x2 =
2

y being the geometrical mean between a and b.

Hence, since x2 = y2,
2
Y
=
,

or x, y, z are in geometrical progression.

526. The connection between x, y, z, a, b may be geome-
trically illustrated.
Let the line OA represent a, and OB represent b.
IfAB be bisected in C, since AC = CB, it follows that
OC- OA = OB- OC ; whence OA, OC, OB are in arith
330 Geometrical Progression. [526.
metical progression, and OC is the arithmetical mean
between OA and OB.
With Cas centre and CA as radius describe a semi-circle
on AB as diameter ; and draw OP from O to touch this
circle. Join PA, PB,
P
and draw PNperpen-
dicular on AB.
Since OP touches
the circle and OAB
AN C B cuts it, the square on
OP is equal to the
rectangle contained by OA, OB, or, replacing the lines
by the numbers which represent them,
OP² = OA.ОВ,
OA OP
whence =

OP OB

or OP is the geometrical mean between OA and OB.

Again, the angle OPA is equal to the angle PBA in the
alternate segment, which again is equal to the angle APN,
because the triangles ANP, APB are similar. Hence PA
bisects the angle OPN; and PB, which is perpendicular to
PA, since the angle in a semicircle is a right angle, must
bisect the angle between PN and OP produced.
NA NP
Hence, by Euclid VI. 3, 40 =

PO
NB NP
and, by Euclid VI. A, BOPO
NA NB
whence =

40 BO'
ON - OA OB - ON
or ,
04 OB

ON ON
or -1-1-
04 OB
527.] Geometrical Progression. 331

and, dividing these equals by ON,

1 1 1 1

04 ON ON OB
1 11
whence OA' ON' OB are in arithmetical progression,
and therefore OA, ON, OB are in harmonical progression
(Art. 516), or ON is the harmonic mean between OA
and OB.
If CP be joined, since the angle OPC is a right-angle,
the triangle ONP is similar to OPC, and therefore
ON OP
OP =

OC; whence ON, OP, OC are in geometrical pro-

gression.
527. The following is another method of comparing the
conditions that three quantities shall be in arithmetical,
geometrical, and harmonical progression respectively.
If a, b, c be in arithmetical progression, by the definition
of Art. 503, a - b = b - c, which may be written
a- b a

b- c a
(1)
If a, b, c be in geometrical progression, by the definition
a b
of Art. 520, = C
, or ac = b2 ; whence, subtracting each
of these equals from ab,
ab -b² = ab - ас,
or
b (a − b) = a (bc),
or, dividing these equals by (bc) b,
a- b a

b - c6 (2)
If a, b, c be in harmonical progression, by the definition
ofArt. 516,
1 1 1 1
--

6 a C
332 Geometrical Progression. [528.
a- b b- c
or ;
ab bc

whence, multiplying these equals by ab and dividing by

b - c,
a- b ab a

b- c be C
(3)
a- b
Hence in the three cases the fraction is equal to
b- c
a a a

' ' respectively.

The condition (3) is sometimes given as the definition of
harmonical progression.
528. The sum of n terms of a geometrical progression
1 - rn
has been shown to be a
1- r (Art. 522)
Supposing to be less than unity, the second term in
the numerator, r", diminishes as n increases. Hence the
sum of the series to n terms continually approaches nearer to
a
the value
1- r

Since S =
α(1 )
1-7
a arn
,

1- 1- r
a arn
the difference between su and 1- is 1
-r
, or r" S The

difference continually diminishes as n increases, as has been

already observed .
It may be further asserted that by making n sufficiently
large the value of r"s can be made less than any number how-
ever small.
If this be granted, it follows that the sum of the
geometrical series carried on absolutely to infinity must
a
be
1- r The symbol , read infinity, is frequently used
530.] Geometrical Progression. 333

to denote a number to the magnitude of which no limit

can be assigned, and thus in this case we may write
a
S (4)
1 r

529. The truth of this equation depends upon the two

a
statements in italics. For if so be not equal to itmust
1- r

differ from it by some finite quantity . Then, in virtue of

preceding statements, by taking n sufficiently large, s, can
a
be made to differ from by less than 2, however small
1-r

x may be. Also, as n is increased s keeps approaching

a a
nearer to Hence 80 must be nearer to than sn
1- 1- r
q
is, that is, 8 cannot differ from 1- r
by so much as x.
a

Hence the supposition that 80 differs from 1- r

byany A

finite quantity is false, and 80 must therefore be absolutely

a

equal to 1- r

530. The theorem of the last article is so important that

it is perhaps worth while to illustrate the second statement
in italics by a few particular examples.
1
Let, for instance, r = 2 Then the first ten powers of r
111 1 1 1 1 1 1 1
are
32 64' 128' 256' 512 and 1024
, ,
2'4 8 16
1 1
Hence r10 < <
< .001, the symbol < being
1000 103
used as an abbreviation for the words ' is less than.' Con-
1
sequently lon < 03n; and by taking n sufficiently large,
this latter quantity, which is decimally written as 3n - 1
334 Geometrical Progression. [531 .
ciphers and a unit to follow, can evidently be made less
than any given measurable or conceivable quantity.
Suppose, again, that r = 5, the successive powers of rare
6

5 25 125 625 1
, of which the last is less than Hence
6'36' 216' 1296 2
1
< ; whence by the previous calculation
1、10 1 1
40 < ( 0 < 103
2
and 40n <
103n

Hence, as in the former case, by taking m sufficiently

large, rm can be made less than any conceivable quantity
however small.
The nearer r approaches to unity, the higher power of it
will have to be taken before its value comes below any
given small quantity, but by taking a sufficiently high
power, the object can always be obtained provided ris at
all less than unity.
The symbol < has been used for the words ' is less than.'
The symbol > is similarly frequently used as an abbrevia-
tion for the words ' is ' or ' are greater than.'
531. The formulae (1) ofArt. 521, (2) and (3) ofArt. 522,
and (4) of Art. 528 are the fundamental formulae of the
subject of Geometrical Progression. They may be repeated
here
Un = arn- 1, (1)
1- γη "-1
n = a 1-1 =a 7-1 ,
(2) and (3)
a
8 = , where r < 1 . (4)
1-r

532. A repeating or circulating decimal is an instance of

a geometrical progression carried on to infinity. Thus the
decimal 43527 is a concise expression for the series
43 527 527 527
102 + 1055+ 108 + 1011 + ... ad. inf.
533.] Geometrical Progression. 335

The terms of this after the first form a geometrical series

527 1
whose first term a is and whose common ratior is
105 103

Hence the value of the repeating decimal is

527
43 105
+ ,

102 1
1-
103

43 527
=

1022+ 102 (103-1)'

=
43(103-1) +527
102 (103-1)
43000-43 + 527
= ,

100 x 999
43527-43
,
99900

a result which agrees with the ordinary rule given in

treatises on Arithmetic.
533. In the general case we may suppose that the non-
repeating part I contains p figures and the repeating part
Q contains q figures.
P
The value of the non-repeating part is therefore 10P , that
Q
ofthe first period 10p+q, and, since each of the repeating
periods begins q places after the first figure of the previous
one, the successive periods really represent
Q Q
10P+2q' 10P+3 '
and so on. Hence the whole value of the fraction is
P Q Q
+ + + ... ,
10P 10p+q 10P+29

ofwhich the terms after the first are a geometrical pro

336 Geometrical Progression. [534-
Q
gression whose first term is and whose common ratio
10P+9
1
is ; the sum of these terms to infinity is therefore,
109

by (4), Q
10P+q Q
or

1
1 10P (10-1)
109
P Q
Hence the whole fraction = +
10P 10P (10-1)
=
P(10-1) + Q
10P (10-1)
(P 10º + Q) -P
=

10º(10-1)
Now P10 + Q is, in accordance with the decimal sys-
tem of notation, the number obtained by writing down
the p figures of Pand after them the q figures of Qas
one number. Hence the numerator is the result obtained
by writing down in order the non-repeating part of the
decimal fraction and the figures in one period as one
number, and subtracting the non-repeating part.
The denominator is the product of 10" ( 10-1 ) ; but
10º- 1 is a number less by unity than 10º, and this is
well known from Arithmetic to be a number made up
of g nines, the multiplication of which by 10º adds p
noughts at the end.
Hence the denominator contains as many nines as there
are figures in the repeating part of the decimal followed
by as many ciphers as there are figures in the non-repeat-
ing part.
This is the rule given in treatises on Arithmetic.
534. Another important illustration of the subject of
Geometrical Progression is afforded by the calculation of
the present values and amounts of annuities.
535.] Geometrical Progression. 337
४
If P be a sum of money, the amount paid for its use
for a year is called the Interest on it for one year. This
interest is usually reckoned at so much for each hundred
pounds, and the number of pounds paid for each hundred
pounds is called the rate per cent. If a be the rate per
x
cent., 100 will be the fraction of any sum, as £1 , paid
for the use of that sum, or the interest on £1 . This
fraction is usually denoted by r.
Hence £Pr is the interest on EP for one year. If this
interest remain in the hands of the borrower the amount
lent during the second year is really
£(P+ Pr) or £P( 1 + r).
The borrower must therefore pay interest on this sum and
not on P only. Hence the interest during the second
year will be £rP( 1 + r), and the total number of pounds
sterling due from the borrower at the end of two years
is P( 1 + r) + rP( 1 + r), or P( 1 + r) ( 1 + r), or P( 1 + r)².
Similarly if the interest be left unpaid for another year,
the total amount due at the end of three years is
P(1 + r) : and, by a continuation of the same reasoning,
the amount due at the end of n years is P( 1 + r)".
535. This sum is usually called the amount of £P at
Compound Interest in n years.
The factor 1 + r represents the amount of £1 in one
year at the given rate of interest. The single letter R
is usually written instead of 1 + r. Thus, if £M represent
the amount of £P in n years at the rate of interest Er
for £ 1 ,
M = P(1 + r)" = PR". (1)
The amounts of £P in 1 , 2, 3, ... n years are thus a
series of terms in geometric progression whose common
ratio is R.
Z
338 Geometrical Progression. [536.
536. The difference between the amount of £P in
n years and £P is the total sum paid for the use of £P
for n years, and is called the interest on ₤P for n years.
If this be denoted by £I,
I = PR - P

= P(R - 1). (2)

537. A sum of money due in some years' time is not
so valuable as an equal sum of money paid at present.
The present value of a sum M duen years hence is a
sum such that its amount in n years at the obtainable
compound interest will be M. Hence if P be the present
value, and R be the amount of £1 in one year at the
standard rate of interest,
PR = M.
M
Hence P= .

(3)
R

The present value of a sum M due one year hence is

M M
therefore due two years hence is and so on. These
R' R2
present values obviously form a geometrical series whose
1
common ratio is
R
The difference between the sum M due n years hence
and its present value, is called the discount. It is the
amount which the person to whom the money is due
n years hence may forego without loss in return for having
the amount paid at once. If £D be the discount on £M,
D = M-P

=M-M
R

=M (1 ) (4)
539.] Geometrical Progression. 339

538. The interest on £M for n years is £M(R" - 1).

If this be called £I', it follows from (4)
DR" = I' .
Hence I'−D = D(R - 1).
But £D (R - 1) is the interest on £D.
Hence the interest on a given sum Mfor n years ex-
ceeds the discount on the same sum for the same time by
the interest on the discount.
539. An annuity for a fixed number of years consists
of a payment, at the end of each year, of a certain sum.
Let this sum be £A. Then the present value of the pay-
A
ment at the end of one year is, by Art. 537, &R , that
A
of the second payment is £ and so on. Hence, if
R2
II, denote the present value of the annuity, the latter
being supposed to continue for n years,
A A
Π = + + ... +
R R2 R

Thus I is the sum of a geometrical series whose first

1
term is and common ratio
R R
A
(1 )
Therefore Π =
1
1
R

A
= (1- R ). (5)

The present value of an annuityA to continue for ever

is given by the formula
A A
Π. =
8
R- 1 r
(6)
Z2
340 Geometrical Progression. [540.
This last result is obvious, for it is merely the statement
that £A is equal to £r , the interest to be annually paid
for the use of £Π. (Art. 534).
540. The sum £I is a sum of money which if put to
interest would just supply the payment of £A at the end
of each of n years. Obviously this is a necessary con-
dition to be satisfied by the present value of an annuity,
and the value of I can be calculated from this condition.
The amount of £In at the end of one year is £IR.
When £A is paid there remains £( П„R- A). The amount
of this at the end of the second year is £(Π„R - A)R.
After the payment of the second sum of £A there remains
£(П„R² - AR -A). At the end of the third year there
will similarly remain £( П„R³ - AR² —AR2 – A) ; and by
similar reasoning at the end of n years there will be left
£(ΠR" - AR -1 – AR"-2 ... AR2 - AR - A). If this be
-

zero, I will be the required present value.

Hence

П₂R" = AR -1 + ARn-2 + ... + AR² + AR + 4. (a)

The right-hand member of this equation is ageometrical
1
series whose first term is AR"-1 and common ratio R
Hence its sum is
1
AR"−1 (1 -

R 4R (1- )
or
1 R- 1
1-
R

AR (1- )
Therefore ΠΡ" = ,

R- 1
1
4(1- )
or Π72 = ;
R- 1
the same value as before.
542.] Geometrical Progression. 341

541. The equation (a) of the last article is equivalent

to the statement that the amount of I, allowed to accu-
mulate for n years must exactly equal the sum of the
amounts of A due at the end of one year, A due at the
end of two years, and so on, all allowed to accumulate
to the date of the last payment A at the end of the nth
year. This is evidently a condition which the present
value ought to satisfy.
542. The present value of an annuity to commence after
pyears and last for 9 years can be calculated in a similar
manner to that adopted in Art. 539. For the present
values of the 1st, 2nd, ... 9th payments due in p + 1 , p + 2, ...
p + q years will be respectively
A A A

Rp+1 ' Rp+2 *** Rp+q ;

and the sum of all these will be the present value re-
quired. But this sum is that of q terms of a geometric
A 1
progression whose first term is Rp+ and common ratio
1 R

Hence the present value required is (Art. 522)

1 1
1 1-
A R A R
RP+ 1 1 RP R- 1
1-
R

EXAMPLES.

Find the nth term and the sum of the first 10 terms of each
of the following series in geometrical progression :-
1.1 + 1 + 1 + .... 2.1 - + ....

3 .- + -.... 4.1+ 2 + 4 + ....

5.5 + 6 + 7 + ....
342 Geometrical Progression.

Find the sum of each of the following series to n terms, and

in all possible cases to infinity :-
6.1 - + .... 7.2-1 + ....

√3-1
8.3 + √3 + 1 + .... 9. ( √3 + 1) +1+ 2
+ ....

10.5-25-125 -.... 11.6 + 3 + 1 + ....

62
12. a + b + - + ....
a

13. Find the values of the repeating decimals -428571

and -0127.

14. Find a geometrical progression whose first term is unity

and whose third term is .

15. How many terms of the series 1 + 2 + 4 + ... amount to

1023 ?

16. If the pth, qth, rth, sth terms of an arithmetical progres-

sion be in geometrical progression, prove that p - q, q - r, r - s
are in geometrical progression.
17. There are two geometrical series whose common ratios
are x and y. It is found that the differences of their first three
terms are in arithmetical progression, and the sums of those
terms in harmonical progression. Prove that
(x + y) (1 -x) ( 1 - y) = 0 .
18. The sum of n terms of the series 15, 53, 5, ... is of
the sum to infinity. Find n.
19. The sum of 2n terms of a geometrical series whose
common ratio is ris 3 (r" +16) ofthe sum of the first n terms,
2

and the sum of the first, third, and fifth terms is 10 of the sum
of the second and fourth terms : find rand n .

20. If a, b, c, d, ... l be n terms in geometrical progression,

prove that abed...1 = √ul.
 Geometrical Progression. 343

21. Prove that the arithmetic mean of the arithmetic and

geometric means between any two quantities is equal to the
arithmetic mean between their square roots.
22. S, S and ∑ denote the sum of a geometrical series
ton terms, 2n terms and infinity respectively. Prove that
Σ- Σ
S-S S is independent of n.
23. If 81 , 82 , 83 be the sums of a geometrical progression to n
terms, 2n terms and 3n terms respectively; shew that
81 (83-82) = (8-8₁)2.
24. If a series of terms in arithmetical progression be
collected in order into groups of n terms and the terms of each
group be added together, the results form an arithmetical
progression whose common difference is n²x original common
difference .

If this process be repeated on the second series to form a

third, and so on, and if a,, b, be the first term and common
difference of the rth series, then ab , a₂- b , ab , ...
form a geometrical progression whose common ratio is n.
25. The sum of six terms of the series 1 - x -1 -x² + ... is
65 times the sum to infinity. Find x.
26. The arithmetic mean between two numbers exceeds the
harmonic by 1, and twice the square of the arithmetic mean
exceeds the sum of the squares of the geometric and harmonic
means by 11. Find the numbers.
2 1 2

27. If a, b, c be positive integers and ab, bac, co be in

2 1 2

geometrical progression, shew that a ", ba"c", ct" are also in

geometrical progression.
28. If an arithmetical, geometrical and harmonical progression
have the same first and second terms, and their third terms be
a, b, c respectively, prove that
b
( -3) +4( + ) = 12.
344 Geometrical Progression.
1
29. A man lays by annually th of his income which he
m

invests. Prove that if his income for any one year be £I, his
12
r

income n years later will be £I(1 + -) , £r being the interest

m

of £1 for one year and compound interest being assumed.

30. Interest is payable m times a year. Assuming that the
interest is invested at once, find the amount of £P in n years.
31. What is the real rate per cent. when the interest is
nominally 5 per cent. per annum paid quarterly, assuming that
the interest is invested immediately ?
32. Find the present value of an annuity of £ 100 a year paid
quarterly to last for 10 years, interest being reckoned at 4 per
cent. per annum.
33. Find the present value of an annuity of £A per annum
payable m times a year to last for n years, £r being the interest.
on £1 per annum.
34. If the arithmetic means between the pairs of quantities
(a, x) (b, y) (c, z) be in arithmetical progression, the geometric
means in geometrical progression, and the harmonic means in
harmonical progression, prove that
1 1 2
by a
+ C
7) = 2b -a- c, or that by = ac.
 CHAPTER XX.

ON THE SERIES FORMED BY THE PRODUCTS OF

BINOMIAL FACTORS .

543. IN Arts. 129 and 131 the following relations have

been proved,
(x + a) (x + b) = x² + (a + b) x + ab,
(x + a)(x + b) (x + c) = x² + (a + b + c)x² + (bc + ca + ab)x + abc.
Here the product of two or three binomial factors, the
first term in each of which is æ, have been expressed in
series arranged according to descending powers of x. The
product of any number of binomial factors of the same
form can be also expressed in a similar series.
In Arts. 104 and 106 it is shewn that the product of
any two algebraical expressions is the algebraical sum
(Art. 93) of all the products which can be formed by mul-
tiplying any term in the first expression by any term in
the second.
Hence it follows that the product of three or more ex-
pressions is the algebraical sum of all the products that
can be formed by multiplying together one term out of
each of the factors.
Let there then ben binomial factors, the first term in
each ofwhich is æ, and let the factors be represented by
x + a , x + a2 , X + A3 , ... x + an .
Each term in the product will be the product of one term
out of each of these. The highest power of a will occur
when all the a's are multiplied together, and, since there
346 On the Series formed by the [543.
are n of these, it will be x". A number of terms can be
obtained involving 2"-1, namely, the product of n - 1 x's,
out of any set of (n - 1) of the binomials, into the second
term of the remaining binomial. Thus, on the whole, the
coefficient of 2"-1 must be a₁ + a2 + A3 + ... + an , which may
be concisely denoted (Art. 119) by the symbol (α).
The terms involving 2"-2 must similarly contain the
product of two of the terms a₁, A2, A3, A4, ... an , and the
coefficient of xn-2 may be denoted by the symbol 2 (a₁ a₂),
since a term will certainly occur containing the product
of any two of the quantities A1 , A2, A3, ... an . The terms
involving "-" must contain r of the second terms as
factors: and by the same reasoning, the coefficient of "-
must be (a1 a2 ... a.), the symbol meaning the sum of
all possible products of r out of the n quantities
A1, A2, A3 , ... an .
The only term independent of a will evidently be
A1 •A2 A3 ... An .
.

Thus, on the whole,

Σ
(x + a₁) (x + a₂) ... (x + an) = x² + xn-12(α) + x -22(α₁₂) + ...
... + -Σ (α₁₂ ... a ) + ... + a₁ . A2 ... An .
544. If all the quantities a1 , a2, a3 , ... an be equal, and
each be denoted by a, the left-hand member of the above
identity becomes (x + a)".
The number of terms in 2(a) is n, and, since the value
of each term is a, the value of 2(a) is na.
The number of terms in 2(a₁ a₂) is the number of
distinct pairs that can be formed out of the n quantities
n (n- 1) As each of
A1 , A2, A3, ... an, that is (Art. 488), 1.2
these terms will have the value a², it follows that

Σ(α1 α₂) = n (n
1.2
- 1)a2.
 347
545.] Products of Binomial Factors.

The number of terms in ∑ (a₁ a₂ ... a.) is the number of

distinct sets of rout of the n quantities a1 , A2 , A3 , ... Ans
that is (Art. 488), n (n - 1 )...(n - r + 1 ) , while each of the
r

terms, being the product of r factors equal to a, is a'. Hence

Σ(α1 αρ...α.) = n (n - 1)... (n-r + 1) ar.
a
r

The last term, a₁ a₂ ... an, becomes a".

Hence, on the whole, it follows that

(х + а)" = x² + пах"-1+ n (n-

1.2
1) a²xn-2 + ...
... +
n (n - 1) ... (n - r + 1 ) ax - + ... + a".
T

This result is known as the Binomial Theorem.

545. The following slightly different proof may be given
of the Binomial Theorem. If then binomial factors be
written on a vertical column as below,
x+ a (1),
x+ a (2),
x+ a (3),

x+ a (n),
their product is obtained by taking the sum of all possible
products of one letter out of each of the n binomials.
The product of the nx's is 2", that of the na's is a" .
These are consequently two terms in the required result.
Every other term will contain some factors 2 and some
factors a. If in any term there ber factors a there must
ben- r factors a, since the total number of factors is n. L

This term is therefore represented by a' x"-". This product

can be formed in a great number of ways ; namely, the
number in which n things of which rare a's and n- r are a's
348
On the Series formed by the [545-
can be arranged in n places ; for an a or an a has to be
taken from each binomial. This number is (Arts. 499,500)
n

rn r
Hence the sum of all the terms involving
n
x - a is xn-rar.
r n- r

Now r may have any of the values 1, 2, 3, ... (n - 1 ).

Hence, collecting all the different terms, it follows that
n n

(x + a)" = x² + xn- 1a + xn-2 a² + ...

1 n- 1 2 n- 2

... + xn- a + ... + a " ;

rn- r

which, by taking out factors common to numerator and

denominator in the coefficients (compare Art. 490), gives
the same result as before; namely,
(n - 1 ) n

(x + a)" = x² + nxn-1a + 1.2 xn -2 a² + ...

... + n (n - 1)... (n - r + 1 ) an-ra + ... + a". ( 1)

Pr

546. If the meaning attached to 10 in Art. 477 be ac-

n

cepted, the formula x"-" a" includes the first and

r n- r

last terms.
n

For, giving tor the value zero, we get On x" aº, which
n

is æ" ; and, giving tor the value n, we get n0

xa", or

a".
n

547. The expression rn -r

x - a is called the general

term of the series. If the terms be arranged in descending

548.] Products of Binomial Factors. 349

powers of x, and consequently in ascending powers of a

(Arts. 108, 148), this term is the r + 1th term.
The coefficient can be simplified by dividing all the
factors in In - rout of numerator and denominator. The
r + 1th term thus becomes
n (n - 1 ) (n - 2)... (n - r + 1) xn-ra .
r

The whole series is called the expansion of (x + a)".

548. If a have the value unity, the formula of Art. 545
becomes

(1 + a)" = 1 + na + n (n- 1) a3 + n (x-1.2.3

1.2
1) (n- 2) a² + ... , (2)
ther + 1th term of the series being
n ( n - 1 ) (n - 2) ... (n - r + 1 ) a".
r

If the value unity be also given to a, the formula gives

n (n -1) n ( n - 1 ) (n - 2) + ... ;
(1 + 1)" = 1 + n + 1.2 + 1.2.3

or, as it may be written, using the notation ofArt. 486,

2" = 1 + nC₁ + nC2 + ... + nCn.
Hence nC1 + C2 + ... + C = 2" -1, (3)
or the whole number of possible combinations of n things
is 2"-1 .
Similarly if the value - 1 be given to a, we get
(1-1) = 1 - C1 + nCz - nC3 + ... ,
or, nC1+ nCz + nC5 + ... = 1 + nC2 + nC₄ + ... ;
3 (4)
that is, the whole number of combinations of n things
when an odd number of things occur in each, exceeds by
unity the whole number when any even number is taken.
By (3) it follows that each of the equal quantities in (4)
is = x 2 " = 2 "-1.
350 On the Series formed by the [549.
549. Either of the formulae (1) or (2) can be used to
expand, that is, to find the expansion of, any binomial.
Thus, let it be required to expand (2x -3y)7.
The formula (1) will give the required expansion by
making the substitutions 2x for x, -3y for a, and 7 for n .
Thus,
7.6
(2x- 3y) = ( 2x)² +7( 2x) ( -3y) + 1.2 (2x) (-3 )2
7.6.5
+ 1.2.3 (2x) ( -3 )3
7.6.5.4 7.6.5.4.3
+ 1.2.3.4(2x) (-3 ) + 1.2.3.4.5 (2x)² (-3y)5
7.6.5.4.3.2 7.6.5.4.3.2.1
+ 1.2.3.4.5.6(2x) ( -33) + 1.2.3.4.5.6.7(-3 ) .
= 128x7 - 1344 x²y + 6048 x5 y² - 15120 x²y³ + 22680 x³y
- 20412x²y + 10206xy - 2187y7 .
The formula (2) can also be applied, for
17
3y
(2x- 3 ) = {2x (1-2 )} 2x

3y7
= (2x) (1-3 ) . 2x

3y7
The expansion of (1- 31
) can be obtained from (2) by
2x

3y
the substitutions of -
for a and 7 for n. Thus
2x

3y 7.6 3y2
(1-3) =1 +7(- ) + 1.2
7 % (- ) + ...,
when the terms of this series are reduced it gives the same
result as the former process.
550. The general term in any integral power of a
polynomial, as (a + b + c + d + ...)", can be obtained by a
process similar to that of Art. 545.
550.] Products of Polynomial Factors. 351

Let the n polynomial factors be written one below the

other,
a + b + c + d + ... (1),
a + b + c + d + ... (2),
a + b + c + d + ... (3),

a + b + c + d + ... (n).
The product of the factors is the sum of all possible pro-
ducts formed by taking one term out of each row.
The type of such a term is abcd , ... where the sum
of the indices p + q + r + 8 + ... must be n, since every term
is the product of n factors.
This particular term will occur in as many ways as the
number in which it is possible to select pa's, qb's, re's, and
so on, out of then polynomials, or the number of ways in
which pa's, qb's &c., can be arranged in n places. This
n

number is (Art. 499) r s

PI ...

Hence the general term of the polynomial expansion is

n

abcd ....
Pqrs ..
…
where p + q + r + s + ... = n. (1)
By giving p, q, r, s all values consistent with the relation
(1), all the different terms in the expansion can be ascer-
tained.
The formula holds good even if some of the indices,
as s, be zero, provided it be understood that | 0 = 1 (Art.
477). In this case the particular letter whose index
vanishes does not occur in the term considered.
352 On the Scries formed by the [551.
551. If the polynomial be itself a series arranged in
ascending powers of a, any power of it will be also a series
of the same kind.
Thus, if the polynomial be
a + bx + cx² + dx³ + ... ‫و‬

the general term in the expansion of its nth power will be

n

ax (bx) (cx2) * (dx³)* ...

Pqrs...
n
=
a b c d ... xq +2 +38 + ***.
Pqr8 ...

The quantities p, q, r, s must satisfy the condition (1), that

p+ q+ r + 8+ ...
= n ; and if it be required to find the
whole of the term involving any particular power of a,
as x™, such values of q, r, s ... must be taken as also satisfy
the condition q + 27 + 3s + ... = m.

552. Particular cases of the general theorem of Art. 550

have been worked out in Arts. 119 and 121. As an example
of the process of Art. 551 let it be required to determine
the coefficient of x in the expansion of
(1 + 2x + 3x² + 4x3 + 5 x4)5.
Here the general term is
5
1P
1º (2x) (3x2)* (4x³)* (5 x4)
Pqr st
5
=
2.3.4.5.x²+2 +38+ 48, (a)
Pqrst
where p + q + r + s + t = 5. (1)
As only the terms involving a7 are required, we must
alsohave
q + 27 + 3s + 4t = 7 . (2)
55.4.] Products of Binomial Factors. 353

The greatest allowable value of t is unity, and with this

value of t, s may be unity and qandr consequently zero
from (2), and then from (1) p must be
3. All the sets of values allowable are P9 r S t

given in the accompanying table, the 3001 1

1 0 1
method of obtaining them being by 21
1 3 00 1
giving to t, s, successively, lower values

21100

00000
210 0
until all the possibilities are exhausted, 2021
no value of any letter being greater 0 4 01
than 5. 1 130
03 2
There are thus eight terms which
contain æ , and the values of these
terms can be obtained by putting in (a) the sets of values
in each horizontal line of the table. They thus become
15 5
4x5 +
311 21 112x3x5 + 1 312x5
5 5 5
+ 2 x 42+ 32 x 4 + 24x 4
212 221 41

5 15
2 x 33+
+1132 3 223x32,
which reduces to
400 + 1800 + 800 + 960 + 1080 + 320 + 1080 + 720 ,
or 7160 .

The term in the expansion involving a is 7160x7.

553. The expansion of a polynomial can be completely
determined by means of the last three articles. The
different terms can be arranged in classes as in Arts. 119
and 121, but the terms cannot be easily arranged in any
definite order as in the expansion of a binomial in Art.
545 .

554. The general formula for (x+a)" can be arrived at

by what is known as the method of Mathematical Induction.
Aa
354 On the Series formed by the [555.
This method consists of three steps. The first step is the
proof that the theorem holds in some simple case. The
second step is the proof that if the theorem holds in
any one case it holds in the next case. The third step is
the deduction that the theorem holds in all cases following
that for which it has been shown to be true in the
first step.
555. The theorem for the expansion of a binomial given
in (1) of Art. 545 holds when n has the values 2 or 3.
For (Arts. 114, 122)
(x + a)² = x² + 2xa + a² = x² + 2С₁ха +2 С₂а²,
(x + a)³ = x² + 3x2a + 3xa² + a³
= x² + 3 C₁x2a + 3C2xa² + 3C3а³.
This gives the first step in the induction.
556. Again, if it be true that
(x + a)" = x² + nC₁xn-1a + nCzxn-2a² + ...
... + n Can-ra + ... + nCna", (1)
the expansion of (x + a) +1 must be obtained by multiply-
ing the right-hand member by x + a.
In the product a term involving a" as a factor will arise
from the multiplication of Cx -ra" by æ and also from
that of C -10- +1a -1 by a. Hence the term in the
product will be
(nCr +n Cr-1)xm- r+1a",
but n Cr + nCr-1 = n+ 1Cr. For n+1 C, means the number of
combinations of (n + 1) things r together. The number of
these in which one particular thing occurs, is the number
of combinations of the remaining n thingsr - 1 together,
or Cr-1. The number in which that same thing does not
occur is the number of sets of r that can be obtained from
the remaining n things, or „Cr. Since the particular
thing either does or does not occur in every combination,
it follows that
nC, + nCr-1 = n + 1Cr .
558.] Products of Binomial Factors. 355

Hence the term involving at in (x + a)" is on the given

supposition n +1Cran +1-ra .
Thus, on the supposition that the expansion of (x + a)" is
correctly given by the form (1), that of (x + a)*+1 will be
given by the precisely similar form
xn + 1 + n +1C₁xna + n + 1 C2xn-1a² + ....
557. From the last two articles it follows that since
the theorem (1) is true when n has the value 3 it must be
true when n is 4 ; and thence its truth follows for the
value 5, and so on for any integral value of n.
558. The rth term of the expansion of (x + a)" is
n ( n - 1 ) ... (n - r + 2) - +1a -1 .
r- 1

The r + 1th term is similarly

n (n - 1 )... (n - r + 2) (n - r + 1) mm-rar .
r

The latter of these can be obtained from the former by

multiplying by the expression
n- r + 1 a.
r х

If this factor be greater than unity the r + 1th term will

be greater than the rth, if it be less than unity the r + 1th
will be less than the rth.
n- r + 1
Obviously, as in Art. 495, the factor decreases
r

asr increases.
The greatest term in any given expansion will therefore
be ascertained by finding for what value of r the expression
(n- r+ 1) a first becomes equal to or less than unity. If
τα

there be an integral value of r, as p, which makes this

factor equal to unity thepth and p + 1th terms will be equal,
Aa2
356 On the Series formed by the [559.
and greater than all that precede or that follow. If there
be no integral value of r for which the factor becomes
unity and p be the least value of r which makes it less
than unity the pth term will be the greatest. If p be
greater than n + 1 the terms continue to increase to the
end, so that the last term is the greatest.
559. The Binomial Theorem can be often used to prove
relations similar to those of Art. 548. Suppose, for in-
stance, that the coefficients in the expansion of (1 + x)" be
denoted by ao, a1 , a2 , ... , so that
(1 + x)" = a + a₁x + a2x² + ... + a,x² + ... + anx . (1)
It follows, replacing a by -x, that
(1 - x)" = а - а₁x + a2x² + ... + ( −1 )′ a,x" + ...
... + ( - 1)" anx". (2)
Hence (1 + x)" x (1 -x)" or ( 1 -x²)" is the product of
the two series on the right-hand side. But replacing & in
(1) by -², we have
(1 - x2) = a - a₁x² + ... + ( -1) a,x² + ...

... + ( - 1)"anxin. (3)

Hence ( -1) a must be the coefficient of 2² in the
product of the two series on the right-hand sides of (1)
and (2).
But this coefficient must be obtained by multiplying
together all terms, one in each series, the sum of the
indices of x in which is 2 r, and will be seen to be
Apazr -A1A2r-1 + a2a2r-2 -a2r-101 + a2av .
-...
(4)
Hence
(-1) a = apa - a1a2r-1 + ...-a2+-1a1 + a2+a0
= 2 (aa2 - a1a2r-1 + ... + ( - 1) -1a -1ar +1)
2

+ (- 1 ) a,² ;
the last reduction being obtained by noticing that there is
an odd number of terms, 2 + 1, in (4), and that the last r
are the same as the first in the reverse order.
 Products of Binomial Factors. 357

EXAMPLES .

1. Expand (a- x) ; (2a + 3b) ; (a - 16) . 3 20

x7
2. Expand to five terms ( 2- ) ; (x² X
)
1.7
3. Expand(x + 1); (a−b) .
3 20

4. Find the term involving y15 in (x² -

)
х
12
1
5. Find the r + 1th term in (x+ =)
1.2n+1
6. Find the terms in (x + =) which involve 2 and x
х

respectively.
1.2 1

7. Find the terms in (x +2) which involve and 2

respectively.
8. Four consecutive terms of a binomial expansion are 14,
84, 280, 560, respectively : find the expansion.
9. Find the greatest term in (a +b)" when a = 2, b = 3 and
n = 7.

10. Find the coefficient of x in the expansion of

(1 + 2x- 3x² + x³) .
11. Find the coefficient of x in (2 -x + x² - 3x3) .
12. If A be the sum of the odd terms and B that of the even
terms in (x + a)", prove that
(x² - a²)" = A2 -B2.
13. If the whole number of combinations of (m -n) things
1
is of the whole number that can be made out of (m + n)
17
15
things and of the number that can be made out of
128

2 (m - n) things, taking an odd number together in each com-

bination; find m and n.
14. Iff(n) be the (r+ 1)th term of (a +b)", prove that
{f(n) }2 n (n -r+ 1)
{{f(n)}
f(n) } ² -f(n - 1) .f (n + 1) r
358 On the Series formed by the

15. If the (m + 1)th and (m + 2)th coefficients in (a + b)" be

equal, and also the (m + 3)th and (m + 4)th terms, then
(m + 3) a = (m - 1)b.
16. If (x + a)" = pox" + P₁ x²¹ a + P₂ x²-2a² + ... ,
and
(x + a) = qox ™ + q₁ xm-¹ a + ... ;
find the value of Poqr + P19r +1 + P2 qr+2 + ... •
17. If ( 1 + x)" = a + a₁ x + a₂ x² + ... + a, x² + ... + an x",
prove that
(1) a = an_r
2n

(2) aa + a₁ar+1 + A2 Ar +2 + ... + an_ran = n - rn + r

;

(3) ao ar -a1 a2r -1 + A2 A2r-2 - + ( - 1) -¹ar_1ar+ 1 + ( - 1 ) 1a²

n
(-1)
2 Ir n- r

(4) aa - a, ar+1 + a2 ar +2 .. + ( - 1) -an_ran

n-r
n
2

= ( -1) *
(n - r) (n + r)
In (4) examine the nature of n and rand find the value of
the expression on the left-hand side when this condition is not
satisfied.

18. Prove that if S, denote the sum of the products r together

of the n quantities X1, X2, X3, ... Xu,
2 2

(x²- 1) (x² - 1) (x² - 1 ) ... (x² -1 )

= (S + Sn2 + ...)² -(S -1 + Sn3 + ... )2.
19. If p + q + r = p' +q + r'= p" + q" + r" = a positive in-
teger n, and q + q + 2 (r + r') = 2q" + 4r",
1 1 1
Σ
then ( -1) "

ppqqrr = » Σ
Show also that if p + q + 2r and also p + q' + 2r' be in-
variable and equal to two numbers whereof the former exceeds
the latter by unity, then
29.9 " 29.9
Σ Σ
3
PI2r' +1 pq2r
 Products of Binomial Factors. 359

20. In the expansion of (a + bx + cx²)" if the coefficient of x

vanish that of x2nm will also vanish .
If the coefficient of x² vanish, prove that 2ac= -(n - 1 )b².
21. Show that if for any positive integral value of n the
value of (2 + 3)" is calculated, its integralpart will be an odd
number.

22. Prove that the integral part of ( a² + 1 + a)" is even if

n be odd and odd if n be even.

23. If a, be the coefficient of x in (1 +x)", prove that

72
n

a² -a² + a - a² + ... =(- 1) 2 or 0,

2-2
)

according as n is an even or odd integer.

 CHAPTER XXI .

ON THE INFINITE SERIES 1 + nx +

n (n - 1) x² + ....
1.2

560. The expansions of the last chapter, and the con-

verse summations of series in Chapters XVIII and XIX,
with one exception, have all contained the expression of
the arithmetical identity of two different forms, each of
which might be employed to calculate the value of either,
though one form might be sometimes more convenient
than the other.
Thus the value of the sum of an arithmetical progression
may be obtained in any given case either by adding all
the separate terms, or by calculating the value of the ex-
pression to which this sum is proved to be equal (Art. 506).
The latter process is usually the easier.
Similarly the value of (x + a)" for any given values of
a and æ, may be obtained either by multiplying n factors
x + a together, or by computing the values of the separate
terms in the equivalent expansion of Art. 545.
561. The one exception alluded to has been the theo-
rem of Art. 528, in which it is proved that the sum of a
a

certain series carried on for ever is equal to Here

1- r

it is impossible to go through the processes of calculating

the separate terms absolutely to infinity and adding the
result. The identity arithmetically of the series
a + ar + ar² + ... ad inf.
a

and the quantity is therefore one which cannot be

 n (n - 1)x²+ㅏ .... 361
563.] On the Infinite Series 1 + nx + 1.2

verified by actual computation of their equivalence. The

proof of the equality of the two expressions depends on
the fact that the supposition of their inequality leads to
an absurdity. There are many other series which when
carried on absolutely for ever can be shown to be equiva-
lent to finite algebraical forms. The numerical equivalence
of the finite and infinite forms in cases where it holds
good must be established by considerations analogous to
those which have been employed in the treatment of the
infinite geometrical series.
562. The series of which ther + 1th term is

n (n- 1) (n - 2) ... (n-r + 1) . ",

r

where r is, from the nature of the case, integral, is for

general values of n an infinite series. It is true that if
n have any positive integral value all the terms after the
n + 1th will contain a factor n - n or zero. The value of the
series will in that case reduce to that of its first n + 1 terms,
but regarded as a general formal function of n (Art. 165)
the series still contains an infinite number of terms. The
next few articles will be devoted to an investigation of
the properties of the series in question, namely
1 + nx +
n (n + 1) x² + + n (n - 1) ... (n - r + 1 )
...

1.2 r

+ ... ad inf.

563. The series considered depends for its value on two

quantities n and 2. It may therefore be regarded as a
function of either n or x or both (Art. 165). For present
purposes it will be treated as a function of n to which
different values will be given while the value of a remains
unaltered. For shortness the series will be denoted by
(n) : and the quantity n will be called the parameter of
the series.
362 On the Infinite Series [564.
564. By the theorem (2) of Art. 548, it follows that
the value of $ (n) when n has any positive integral value
whatever, is ( 1 + x)".
It will have been observed by the student that two
algebraical forms which are equivalent for any general
series of values of the letters involved are also equivalent
for all other values.
It will probably seem hardly conceivable that the two
forms (n) and (1 +x)" should be equivalent for all in-
tegral values of n and that the equivalence should fail
for values intermediate to the integers. The equation,
when n is any quantity whatever,
n
(n- 1) x² + ...
1 + x)" = $ ( n) = 1 + nx + 1.2

may be perhaps accepted at once as a particular instance

of a general principle of the permanence of equivalent
algebraical forms.
The assumption involved in this method, an assumption
which really lies at the bottom of all proofs of the theorem,
is perhaps rendered more easily acceptable in the develop-
ments of the following articles.
565. Let it be required to multiply a series
1 + a₁x + a2x² + a3x3 + ad inf. ...

by a second series,
1 + b₁x + b₂ x² + b3x3 + ....
By conducting the multiplication in the ordinary method
of Arts. 108, 109, or by that of Art. 136, the product can
be ascertained to be
1+ (a₁ + b₁)x+ (a2 + a₁₁ + b2)x² + (a3 + a2b₁ + a₁b2 + b3) x3 + ... ;
that is to say, it is a series of the form
1 + C₁ X + C2 X² + C3 x3 + ... ,
where C1 , C2, C3 , ... are definite functions of the coefficients
a1 , a2 , az ... , b₁ , 62 , 63 ....
 n (n - 1) x² + .... 363
566. ] 1 + nx +
1.2

The point of chief importance to notice is that the

quantities C1 , C2, C3 ... involve the coefficients a1 , a2 , a3... ,
b1, b2, b3 ... in exactly the same manner whatever be the
values of these coefficients. t

566. Thus, if the series $ (n) or

1 + nx +
n (n- 1) x² + ... ad inf.
1.2

be multiplied by the series (m), or .

1 + mx + m(m- 1 )x² + ... ad inf.,
1.2

the product will be a series of ascending powers of x, the

coefficients of the different terms in which must involve
mand n in the same manner whatever be the values of
mand n. The coefficients of each power of a cannot be
a different formal function of m and n when there are
positive integers from what it is when they are fractional
or negative.
The first few terms of the product are easily ascertained
tobe

1+ (m + n) x + m(m -1) + mn + n(n- 1) }w² + ...

1.2 1.2

The coefficient of z² in this series

m (m - 1 ) + 2mn + n (n - 1)
1.2
m² + 2mn + n² - m - n
1.2

=
(m + n)²-(m + n)
1.2

=
(m + n) (m + n - 1)
1.2

Hence for the first three terms the product is

1+ (m + n) x + (m + n) (m+ n-1) 22 +...,
1.2
 364 On the Infinite Series [567 .
which agrees with the first three terms of the series
ф(m + n).
✔ Any number of coefficients in the product could be cal-
culated and the law would be found to be followed as
far as the calculation went. The universality of the result
¿ can be proved by the following considerations.
567. When m and n are any positive integers whatever
the two series $ (m) and (n) are equivalent to (1 +x)
and (1 + x)" respectively (Art. 548). Hence their product
is equivalent to (1 + x)m × (1 + x)" or (1 +x) +", which
again (Art. 548) is equivalent to the series $ (m + n).
Thus, when m and n are any positive integers
φ (m) X
× φ (n) = φ (m + n).
But the forms of $ (m) and $(n) are the same whether
m and n be positive integers or not ; hence the coefficients
of all the terms in the product must be of the same form,
regarded as functions of m and n, whatever values these
letters may have (Art. 565). Hence, since the product
assumes the form of the series $ (m + n), when m and n are
positive integers it must be always of the same form ; that
is, for all values of m and n
ф (m) × (n) = $ (m + n). (1)
568. The equation (1) of the last article is an instance
of what are called functional equations. In all cases where
a functional form & satisfies this equation it can be shown
that the value of $ (n) must be expressible in the form k
where k is some quantity independent of n. In the par-
ticular case before us it is evident that, since
φ (n) = (1 + x)"
when n is a positive integer, the value of k must be 1 + x.
Another familiar instance of a function satisfying the
general equation (1) is afforded by the Demoivre Function
cos 0+ i sin 0, which is shewn in treatises on Trigonometry
569.] 1 + nx + n (n1.2
- 1) x2 + .... 365

to be equivalent to ei , e being the base of hyperbolic

logarithms which will be investigated in a future chapter,
and i being a quantity such that i² = -1 .
569. Returning to equation (1), it follows that
ф (m)× (n) X× (p) = φ (m + n) X × (p)
$ = (m + n +p).
Similarly the product of any number of series of the
form 4 (n) can be represented by a series of the same form,
the value of the parameter in the product being the sum
of the parameters in the several factors.
Thus let s series, the parameter in each of which is n,
be multiplied together, the product will be a similar series
whose parameter is sn, or
{ $ (n)} * = φ (ns). (2)
If n be any positive quantity commensurable with unity,
8 can be so chosen that ns is a positive integer. Let this
r
integer be r. Then ns = r, whence n =

Hence
{$ ( )} = $(7).
Φ

But sincer is a positive integer the series (r) is equal

to (1 + x) , (Art. 548). 8

Hence
{$( )} = (1 +x)*;
that is, $ ( ) is a quantity which when raised to the sth

power is equal to (1 + x) ; that is (Art. 66), Φ( ) can be

otherwise expressed as (1 + x) .
Hence when n is a positive fraction, the series (n) or
1 + nx + n(n- 1)x2 + ... ad inf.
1.2

is one of the values of (1 + x)".

366 On the Infinite Series [570 .
570. In Art. 279 it has been shewn that any such
"

expression as (1 +2) has 8 different equivalents. It is

obvious that if the value of the series (1) be calculable,
it can only give the value of one of these equivalents. It
is usual however to reverse the equation proved in the last
article, and to write it
(1 + x)" = 1 + nx +
n(n-1) x² + ... ,
1.2

in which, when n is a fraction, it must be borne in mind

that the meaning is that the series on the left hand is
one of the values of (1 + x)".
571. The equivalence of the last article may be ex-
tended to negative values of n.
For, let m be any quantity numerically greater than n.
From the equation (1) of Art. 567, replacing n by -n, it
follows that

ф(m) × ( -n) = (т -п).

Also by the same law
(т - п) X
× (n) = (m -n + n) = φ(т).
Hence ф(m) × ( -n) × φ(n) = φ(m);
or, dividing these equals by $(m),
φ( -n) × φ(n) = 1 .
1 1
Therefore φ( -n) =
(1 + x)" (Art. 570)
=

ф (п)
= (1 +x)-".
Thus the equivalence of the series (n) and ( 1 +x)" holds
when n has a negative value given to it.
572. The series (n) regarded in the light of its equiva-
lence to (1 + x)" is frequently called the expansion of
(1 +x)", which may be now enunciated in the following
 1 + nx +
n (n- 1) x² + .... 367
573-] 1.2

general form. If n be any number, integral or fractional,

positive or negative, the series
1 + nx +
n(n - 1) x² + n(n- 1) (n-2) x3 + ...

1.2 1.2.3

... + n(n - 1)... (n -r + 1 ) + ... ad inf.

r

is one of the values of (1 + x)". And the fact of their

equivalence is known as the Binomial Theorem.
573. The equivalence which has been proved to exist
between the form (1 +x)" and the series (n) does not
always lead to any verifiable result of numerical equality.
An instance of failure to do this will be afforded by the
expansion of ( 1 -x)-1.
This can be deduced from the general formula by the
substitutions of -x for æ, and of -1 for n .
The r + 1th term in the series will be

(-1)(- 2) (-3) ... (-1-7 + 1) ( 2) .

r

In the numerator of the coefficient there are r negative

factors ; each negative factor being taken as the product
of - 1 into the numerical value of the factor, the numerator
becomes ( -1) .1.2.3 ... r, or ( -1) . ", while ( -x) is
(-1) .x (Art. 127). Hence the whole term becomes
(-1) r

r
( -1 ) x = ( -1)2r. x = x . (Art. 126.)
Hence the series for (1-x)-1 becomes
1 + x + x2 + x³ + ... ad inf.
Now in Art. 528 it has been shewn that the sum of
1
this series is arithmetically equal to 1 X
or (1 - x)-1 on
condition that x is less than unity.
368 On the Infinite Series [574.
The sum of r terms of the series can in all cases be
1-x 1
written as , and therefore differs from by
1-x 1-x

1-

This latter quantity continually diminishes, as has been

shewn (Arts. 528-530), as r increases if a be less than
unity, but increases without limit with r, if a be greater
than unity. Hence the equivalence given by the Bino-
mial Theorem in the latter case cannot in any ordinary
sense include the notion of arithmetical equality.
574. The equation
(1 -x)-1 = 1 + x +x² + ... ad inf.
is thus only true as an arithmetical equation when a is
less than unity. The equation
(1 −x)−1 = 1 + x + x² + ... + x −1 + 1 −x
is true for all values of x. The term may be called
1-2

the remainder after r terms of the expansion of (1 -x)-1.

In treatises on the Differential Calculus methods are
given by which the remainder after r terms ' of the ex-
pansion of any function can be ascertained. If in any case
this remainder diminishes indefinitely as r increases, the
expansion to infinity may be regarded as an arithmetical
equivalent of the function. If the contrary be true, the
equivalence of the function and its expansion must be
regarded as a merely formal truth without arithmetical
significance.
575. The following theorem is useful to distinguish
between the cases in which the binomial expansion gives
an arithmetical equality from those in which it fails to
do so.
576.] 1 + nx+ n(n-
1.2
1 ) ² + .... 369

If, from and after a fixed term in a series whose rth term
is denoted by u, (Art. 513), the value of the fraction
4 +1 is always less than some quantity which is itself
Ur
less than unity, then the sum of the terms of the series
carried on for ever is a finite quantity.
For let up be the fixed term after which the given
condition holds good, and let y be the quantity less than
unity. Then (Art. 530), using the symbol < as an
abbreviation for the words ' is less than,'
Up+1 Up + 2
< y,
Up + 3 <
< y, ... ,
< y,
Up Up+1 Up+1
whence Up + 1 < Yup , Up +2 < YUp + 1 , Up +3 < YUp +2 , ... ,
or Up + 1 < Yup , Up +2 < Y² Up , Up + 3 < y³up , and so on.
Hence
Up + Up + 1 + Up + 2 + ... ad inf. < up ( 1 + y + y² + ... ad inf.)
<
Up
(Art. 528.)
1-y
Hence the sum of all the terms of the series, including
and after up, is less than a certain finite quantity. The
terms up to up being each finite, and there being a finite
number of them, must have a finite sum. Hence the
proposition is true.
A series for which it can be shown that the sum of
its terms carried on for ever is a finite quantity is called
a convergent series. A series for which the reverse is true
is called a divergent series.
576. The Binomial Theorem can obviously only afford
an arithmetical equivalence when the series given by the
expansion is convergent. In this series
u =
n (n - 1 ) ... (n - r + 2) x -1,
T
r- 1

Wr+1 = n(n -1 )... (n -r + 2) (n- r + 1) .

r

B
370 On the Infinite Series [576.
Ur+1 n- r + 1
Hence
U r

=( +11)
=
-(1 +1)x.
r

The latter form is the more convenient when r is greater

than n + 1, that is, in the later terms of the series.
n+ 1
Here as rincreases 1- becomes more and more
r

nearly equal to unity, but is always less than unity.

Hence the value of "r+1 is always numerically less than x.
Ur

If therefore a be itself less than unity the binomial series

will be convergent and the sum of its terms carried on
for ever will be a finite quantity. If then a be less than
unity the arithmetical value of the expansion (n) is
finite, and the Binomial Theorem gives an arithmetical
equality.
If a be equal to, or greater than unity, the value of the
series is indefinitely large, and the theorem cannot be used
for any practical purpose of calculation.
577. The form n (n - 1 ) ... (n - r + 1 )
-th

r
x gives the r + 1

term of the expansion of (1 +x)". In particular cases this

general form is susceptible of considerable simplification.
One instance of this has been given in Art. 573. One
or two more may be useful as examples.
578. Let the expansion required be that of (1 -x)-3.
Here a must be replaced by -x, and n by -3. The value
th

of the r + 1 term consequently becomes

(-3) (- 4) ... (-3-7 + 1) (−∞)",
T
 n(n - 1) x² + .... 371
579.] 1 + nx +
1.2

which, as in Art. 573, reduces to

(-1) .3.4.5 ... (r + 2)
r
(-1)".
= (-1)2.3.4.5 .... ( + 1) ( + 2)
1.2.3.4 ... r
.
The factors 3.4.5 ... r are common to the numerator
and denominator and may therefore be removed, while
(-1)2 " is + 1 (Art. 126).
-th
Hence ther + 1 term of the expansion becomes
( +1)1.2
( +2) .
579. Again, let the expansion required be that of
1

(1-2x) . Here 2 must be written for n and - 2x for x.

-th
Ther + 1 term is consequently
1

2 ( -1) ( - 2) ... ( + 1)
r
(-2x) .

In the numerator all the factors after the first are negative.
Each may be replaced by a factor - 1 multiplied by its
numerical value. Thus, since there are r factors altogether
and consequently (r - 1) ofthem are negative, the coefficient
can be written
-1
11 135 (2-3)
(-1) 2 22 2 2

1.1.3.5 ... (2-3)

= ( -1) *-1 2.

Also ( -2x) = ( -1) . 2.x. Hence the whole term

becomes
1.1.3.5 ... (2-3) 20
(-1)2-1 1.1.3.5 ...(27-3)
2
2.
r

since (-1)2-1 = -1 (Art. 126).

Bb2
372 On the Infinite Series [579.
In either of these cases by giving tor the values 1, 2, 3 etc.
in succession all the terms of the expansion can be obtained.
Thus, in the last example,
22 3x3 3.5 3.5.7
25
(1-2x) = 1 - x 2 3 4 5
...

x2 203 5x
= 1-8- ....

2 2 8

580. The expansion of the Binomial Theorem can be

used in the case of a binomial whose first term is not
unity.
Thus
(a +b) = {a(1+ )}
=a* (1+ )
As many terms as are required can be found of the
b
expansion of (1+ )"by substituting for a in the a

general formula. Multiplying the series thus obtained

by a", the expansion of (a + b)" is obtained.
581. As an example of the numerical application of the
Binomial Theorem to purposes of practical calculation, let
it be required to compute the value of the square root
of20.

This square root can be written as (25-5) , which by

5
the last article (25) (1-2 ) = 5.(1-1) . Also the
=

25
1
expansion of ((11 ) can be obtained from the general
1 1
formula ofArt. 570 by writing for n and -

5
for æ, or it
may be deduced from the expansion of the last article
1
by putting for 2x the value , and consequently the
1
value for x.
10
 1 + nx +
n(n-1)x² + .... 373
582.] 1.2

Thus,
51
(1- 1 1
) = 1- 10 11

2102
11

2103 8104
71

8105

The terms in this series after the first are all negative and
rapidly diminish in numerical value. The sum of a small
number of them will therefore differ but slightly from that
of the whole series. The calculation may be conducted in
the following manner :
1
= 100000
10

1 1 1 1
= X = .005000
2 102 20 10
1 1
Therefore = .000500
2 103
5 1 1 1 1
=
X = 0000625
8 104 8 2 103
7 1 7 5 1
= X 00000875
8 105 50 8 104
•10557125

The succeeding terms will not affect the first four or five
places ofdecimals. Hence to four or five places of decimals
1 894429 .
(1-3) 5

1
And therefore (20) =

5 (1-0) 4.472145 .

The value is really true to four places of decimals, as can

be verified by the process of Art. 397. By taking in a
few more terms of the series a still closer approximation
can be obtained.
582. A formula has beengiven(Art. 550) for the general
term of the expansion of any positive integral power of
any polynomial. The theorem of Art. 572 can be used to
give a similar formula for the general term when the
374 On the Infinite Series [582 .
index is fractional or negative. The most important case
is when the polynomial is itself a series, finite or infinite,
arranged in ascending power of some letter æ, in which case
it may be denoted by a + a₁x + a2x² + ....
Then

(a + a₁x + a2x² + a3x3 + ...)" = ao " (11+ a₁ x + a2x²

do
+

a₁x + a2x + ...

Treating as one quantity, thep + 1th term
do

in the expansion will be

α ".
n (n - 1 ) ... (n - p + 1 ) (a₁x + a2x² + ...)P ,

p ap

which may be written as

n (n - 1 )... (n - p + 1)
an-p (ax + ar² + ...) ”.
p

Now p being a positive integer the general type of the

P
terms in the expansion of (a₁x + ar² + ...) is by Art. 550,
p

r 8 ...
(ax) (a2x²) ( 3x3)* ...
91
p
=
aaa ... x² + 27 + 38 + *** ,
qr8 ...
where q + r + s + ... = p .

Hence the general type of all the terms in the expansion

of the polynomial will be,
p
n (n - 1)...(n - p + 1) a -Paaa
ao 2
7 2
... x + 2 + 3 8 + ...
1
P Is...

=
n (n - 1) ... (n − p + 1 ) a - aaa ...x2 + 2 + 3 8 + ...
Ir 18 ...

583. In the particular case of n being a positive integer,

n -p is also a positive integer. By multiplying the
 584.] 1 + nx + n (n- 1) 22 + ....
1.2
375

numerator and denominator of the coefficient by n -p the

term becomes
n

an-Paaa ...xq +2 + 38 + ...,

n-pqr8...
which agrees with the formula of Art. 551, with the change
of (n - p) for p ; for since q + r + 8+ ... = p, it follows
that (n-p) + q + r + 8 + ... = n.
584. The calculation of the coefficients of particular
terms or particular powers of x in the expansion can be
conducted exactly in the manner indicated in Art. 552.

EXAMPLES.

1. Expand to four terms (1 -x) and (1 +x)- and find the

eighth term of each.
2. Expand ( 1 -x ) to five terms.
2x 4
3. Expand (1 + 3)* and find the first negative coefficient.
4. Find the first four terms and ther + 1th term of each of
the following expansions,
(1-2x) , (1-2x)- , (1 -x)-3, (1 -x)-5.
5. What term of (1 -x) is equal to is of the same term
13

of ( 1 + x) ?
1+x
6. Find the general term of
(1 - x)
7. Find the coefficient of art in the expansion of (1
(1-2)3
+ x)
8. Find the values of √24 and 5
/31 each to five places of
decimals.
9. Find the value of √17 to four places ofdecimals.
10. Expand, to the second power of x,
(1 + x) + (1 -x)-
(1 -x²)
376 On the Infinite Series 1 + nx +
n(n - 1) x² + ....
1.2

11. By ordinary multiplication find the square of the series

for (1 -x) , and shew that the coefficients of x², x , x , x , x
all vanish.
12. From the identity
-1
x

1 +x= ( 1 1+x
= 1 + x (1 + x) + x² (1 + x)-2+ ... ,
prove that if r be any integer greater than unity
( -2) ( -3)
1-(r - 1 ) + 2
-...
... = 0,

the series being carried on until a term vanishes.

1 1
n +r
13. If (r)= { n + r n- r r- 1 n - r +1
+ 1

(r - 1) ( -2) 1
+
2 r- 2 n- r+ 2 +...} ,
prove that 2 { $ ( 0) + ( 1 ) + ... Φ (n - 1 ) } + $ (n) = 3".
14. Expand, to the 4th power of a, (1-2ax + a²) .
15. Find the coefficient of x in the expansion of
( 1 + x -x² - 3x² -x²) 3.
16. Find the coefficient of x in the expansion of
(1-2x + 3x² - 4x² + ...) .
17. Show that the coefficient of x in the expansion of
(9a² + 6ax + 4x²)-1 is 23 ( 3a)-3-2 and that every third term
vanishes.

18. Find the coefficient of x in (1 + 2x- 3x² -x ) .

19. Prove that

+x)² = ( 1 + x)" + nx ( 1 + x) -1 + n(n + 1)

1.2 x² ( 1 + x) -2 + ....

20. If f(x) =
2x (n -x) ( n )
n² 2 n { n -xx } 2
n- 1
prove that f(0) + f ( 1) + f (2) + ... +f(n) = 2n- 1
1 13 135
21. Prove that √2 = 1 ++ + + ....
4 48 4 8 12
 CHAPTER XXII .

UNDETERMINED COEFFICIENTS AND RECURRING SERIES .

585. For the treatment of several problems connected

with the series which are next to be considered the
following proposition, usually known as the principle of
undeterminate or undetermined coefficients, is useful.
If two expressions of the nth degree in æ, of the forms
A + A₁x + A2x² + ... + Anx" and a + a₁x + a2x² + ... + anx",
have equal values for more than n different values ofa they
must also be equal for all values of a, and the coefficients of
the one expression must be severally equal to those of the
other, namely, =
A 0 ao , A₁ = a1, ...
An a
From the given conditions it follows that
(A + A₁x + A2x² + ... +A„x")
-(a + a₁x+ a2x² + ... + anxn) (1)
must have the value zero for at least n + 1 different values
ofx. Let these be a₁ , A2 , A3 , ... An, An +1. Then arranging
the expression (1) in descending powers of æ, it becomes
(An - an) x + (An-1 - an-1) x -1 + ...
... + (4₁ - a₁) x + 4o - a ; (2)
and since this expression vanishes when a is a₁ it follows
(Art. 163) that x-a, is a factor of the expression. Simi-
larly x -a2 , ..., x- a, are factors of the expression; and this
latter being of n dimensions only in a cannot contain any
more factors involving a.
378 Undetermined Coefficients [585.
Hence the expression must identically equal
(An - an) (x - a₁) (x - 2) ... (x - an).
But, by hypothesis, this expression also vanishes when
x = an +1, which is impossible unless An-an vanishes, since
an + 1 is not equal to any of the quantities a1, a2 , ... an.
Hence An must equal an, and thus the expression (2)
becomes really only of the n - 1th degree. Applying the
same reasoning to this, it follows that An-1 = an-1, and so
on, until we arrive at A = a .
Thus the coefficients of the one series are severally equal
to those of the other, and the two expressions are identical
with one another for all values of x.
586. If the two expressions be infinite series, n is
indefinitely large, and n + 1, the number of values for which
the expressions are given equal becomes also infinite. It
is usual to assume, as a limiting consequence of the last
article, that if two infinite series,
A + A₁x + A2x² + ... ad inf. and a + a₁x + a2x² + ... ad. inf.,
be equal in value for an infinite number of values of x, then
A = a , A₁ = a1 , A2 = a2 , ... •
Instead of the words ' an infinite number of values ' it is
not unusual in the above enunciation to write the words ' all
values of æ.
587. The chief difficulty in the recognition of the truth
of the theorem of the last article is the meaning of the term
' value' as applied to an infinite series. A numerical value
can only be imagined for an infinite series, when it is con-
vergent. When this is the case, the proposition admits of
strict proof in the latter form-namely, that if the series
A + A₁x + A2x² + ... ad inf.
have the same value as the series
ao + a₁ x + a2x² + ... ad inf.
for all values ofa, then must
A = a , A₁ = a1 , A2 = a2, ....
587.] and Recurring Series. 379

Assuming that the series are convergent for some finite

values of æ, it follows (Art. 575) that for such values
Ar+ 1
4,
x and "Ar+1 x must be each less than unity. Hence the
fractions 47A+ 1 and ar +1 must, in all cases that need be con-
ar

sidered, be finite quantities. Thus we may assume that in

neither of the series will the coefficients become infinitely
large as we proceed.
Let then Kand k be quantities numerically larger than
the largest coefficient in the two series respectively, and let
x be further supposed to have a value less than unity.
Then
A₁x+A2x² + ... < + K(x + x² + ... ad inf.)
Kx
< + (Art. 528);
1-х
Kx
but by diminishing & sufficiently, can be made less
1-x
than any assignable quantity.
Hence the sum of A₁x + A2x² + ...ad inf., and similarly
that of a₁x + a2x² + ... ad. inf., can be made less than any
assignable quantity. But for all values of a
A +A₁x + A2x² + .. = a + a₁x + a2x² + ....
Therefore, transposing,
A -a = (a₁x + a2x² + ...) - (4₁x + A2x² + ...) .
The value of the left-hand member remains unaltered by
any change in the value of x. That of the right-hand
member can be made less than any assignable number.
Hence A - a, must be zero, for if it had any finite value
the right-hand side could be made less than this value by
sufficiently diminishing 2.
Thus A =a .
380 Undetermined Coefficients [588.
Hence it further follows that for all values ofx,
A₁x+A2x² + ... = a₁x + a2x² + ... ;
or, dividing by a,
A1 + A2x + ... = a₁ + a2x + ... ,
whence, as before, A₁ = a₁.
This process may be carried on indefinitely, and it will
thus follow that the coefficients of the one series must be
severally equal to those of the other.
588. In Art. 573 it has been shewn that
(1 -x)-1= 1 + x + x² + ... ;
writing pæ for æ it follows that
(1 -px)-1 = 1 +px +p2x² + ....
In Art. 582 a method has been given by which the
coefficients of successive powers of x in the expansion of
(a + a₁x + ...)" can be obtained. The method of undeter-
mined coefficients will prove more effective in the particular
case of n = -1 .
589. Let the polynomial be represented by 1 -рx -qx²,
and let the expansion of ( 1 -px- qx2)-1 be assumed to be
a + a₁x + a2x² + azx³ + ... ad inf.
Hence for all values of æ,
1
1 -px - qx2
= a + a₁x + a2x² + ...ad inf.;
or, multiplying both sides by 1 -px-qx2,
1 = (1 - px - qx²) (ao + a₁x + a2x² + ... ad inf.).
Multiplying the infinite series by 1-px- qr2 and
collecting the terms with the same powers of a, this
identity becomes
1 = a + (a₁ - ap) x + (a₂ -pa₁ - qa )x²
+(a3 -pa - qa₁) x³ + (a -pa - qa₂) x² + ... ;
2

the coefficient of a" on the right-hand side being evidently

an-pan-1- qan-2
591.] and Recurring Series. 381

Hence, equating coefficients of like powers of a on both

sides, ao = 1,

a1-pao = 0,

a2-pa - qa = 0,
a3-pa - qa₁ = 0 ,

an-pan-1- qan-2 = 0.
From these equations it easily follows that
a = 1, a₁ = p, a₂ = p² + q, az = p³ + 2pq, ... ,
and the values of as many of the coefficients as are wanted
canbe obtained. The expansion thus becomes
(1 - px - qx2)-1 = 1 + px + (p² + q) x² + (p³ + 2pq) x³ + ....
590. A series, like that of the last article, in which there
is a relation of the form
an - pan-1 - qan-2 = 0
between any number of consecutive coefficients, is called
a recurring series.
It is pretty obvious from the last article that any such
series could be produced by the expansion in a series of
b+ cx
some fraction of the form , where band e can be
1 -px - qx2
chosen so as to give any required values of a, and a₁ . The
sum of any such series to infinity can be therefore always
represented by such a fraction as the above.
The method of undetermined coefficients can be employed
however to give expressions in a general form for the value
of an and also for the sum of n terms of the series.
591. A fraction whose denominator consists of any
number of factors of the form 1 - ax, 1 -bx, 1 -cx, and
whose numerator is of lower degree than the denominator,
canbe replaced by a number of fractions whose denominators
382 Undetermined Coefficients [591 .
are the factors of the given denominator and whose
numerators are determinable constants.
Let the fraction be
p + qx + rx2
(1 - x)(1 - x) (1 - x)
then constantsA, B, C can be found such that this fraction
shall be equal to the sum of the fractions
A B C
+ +
1-ах 1- bx 1- сх

For the latter sum can be expressed as a fraction with

the denominator (1 - ax) ( 1 - bx) ( 1 -cx) and with an ex-
pression of the second degree in a for its numerator, namely
AА (1 - bx) (1 - сх) + В ( 1 - сх) ( 1 — ах) + C ( 1 - ах) (1 - x).
If, by properly choosing the values ofA, B, C, this latter
expression can be made equal in value to p+ qx + rx² for
three values of æ, it must (Art. 585) be equal to it for all
values.
111
Let the three values of x chosen be -,
Putting
C
1
a for x in the two expressions and equating their values,
the terms involving B and C disappear, as they involve
a factor zero (Art. 51) ; and there results
r
+
4 (1 ) (1-2) = p + a

whence A=
pa² + ga + r
(a-b) (a-c)
Similarly the equality of the two expressions for the
1 1
values and ofæ is ensured by taking
C

B=
pb² + qb + r
(b-a) (b−c) '
C=
pc²+qc+ r
(c-a)(c-b)
593.] and Recurring Series. 383

With these values ofA, B, Cthe sum of the three fractions

will therefore be equal to the given fraction.
592. The process of the preceding article is known as that
of the resolution of a fraction into its partial fractions. For
fuller developments the student is referred to treatises on
the Integral Calculus. It will be now used to investigate
a general formula for the coefficient of the nth term in
a recurring series.
It may be noticed that it has been assumed that the
numerator of the given fraction is of lower degree than the
denominator. If this be not the case the denominator can
be divided into the numerator, giving a formally integral
quotient and a remainder of lower degree than the de-
nominator (Art. 158).
593. A recurring series
ao + a₁x + a2x² + ... + anx" + ... ,
where the relation between successive coefficients is
an -pan-1 - qan-2 = 0,
can, as we have seen(Art. 590), be produced by the expansion
of a fraction of the form
b +cx
1 -px - qx2
The values of b and c determine the first two coefficients of
the series and the general relation determines the succeed-
ing ones.
If the factors of 1 -px-qx² be 1 -ax and 1-3x,
6+ сх
the fraction can be replaced by the sum of two
1 -px- qx2
A B
fractions + , where A and B are constants to
1-ax 1 -βα
be determined as in Art. 591. But if these two fractions
be separately expanded in series, they give (Art. 588),
A ( 1 + ax + a²x² + ...) + B ( 1 + 3x + β2x² + ...)
= A + B + (Aa + Bβ) x + ... + (4a" + Bẞ")x" + ...
384 Undetermined Coefficients [593.
But this series is identically equal to
a + a₁x + a2x² + ... + ax + ....
Hence (Art. 587),
a = A + B, a₁ = Aa + Ββ, ... , an = Aa" + Ββ" .
The last result
an = Aa + Bβ"
gives a general form for an . The values of a and 3 are the
two roots of the quadratic equation in t,
t -pt - q = 0 ,
for ( 1 - x) ( 1 - x) = 1 -px - qx2 by assumption. And
1
writing t
for æ, it follows that
β
(1-2) (1-2) = 1-t t

or, multiplying both sides by t²,

(ta) (t - 3) = t2 -pt -q.
The values of A and B can be determined from those of
the first two coefficients ao , a₁ , since
A+ B = a ,
Aa + BB = a .
Hence in any recurring series whose first two terms are
given, as well as the scale of relation between successive
coefficients, an expression can be obtained for the nth term .
594. A recurring series in which four successive
coefficients are connected by a relation of the form
an-pan-1 - qan- 2 - ran- 3 = 0 ,
can similarly be derived from the expansion of a fraction
a + bx + cr²
,

1 - px - qx - ra
and the value of an willbe of the form Aa" + Bẞ" + Cy",
where a, β, y are the roots of the cubic equation
t -pt - qt - r = 0 .
 596.] and Recurring Series. 385

The values of A, B, C will depend on those of the first

three terms of the series.
595. The sum of the first n terms of a given recurring
series can be obtained when the value of an is known.
Let the relation between coefficients be

an -pan-1 - qan-2 = 0 , (1)

and let S be the sum required.
Then
S = a + a₁x + a2x² + ... + an-12-1.
Thus
-px Sn =
-рах -рах² ...- pan -2x - 1. pan-1x
-qx2 S = -qa x2 ...- qan-3x -1 - qan -2xn - gan -1Xn + 1.
Adding these different results, it follows that
S (1 -px - qx2) = a + (a₁ -pa ) x
- (pan-1 + qan-2)x -qan-1Xn + 1,
since the coefficients ofx², x³ ... , 2"-1 in the sum of the three
expressions on the right-hand side vanish in virtue of the
general relation (1).
Hence
a + (a -pax- pan -1 + gan_2)xth-gan-12n+ 1
=
Sn
1-px-qx2
If a be a quantity less than unity, so that " and 2n+ 1
diminish indefinitely when nis increased indefinitely
(Art. 530), the series can be summed to infinity, and we
have
S8 = a + (a₁-pa )x ,

1 -px-qx2
a result agreeing with the assumption of Art. 590.
596. It may be observed that a geometric series is
merely a particular case of a recurring series, the relation
being an -pan-1 = 0.
The nth term, an-12-1, is therefore app"-12"-1, and the
cc
386 Undetermined Coefficients [597 .

sum of n terms is ao-pan-10" or a (1 -px") , results with

1-рх 1 -рх
which the student may compare those ofArts. 521 , 522.
597. The expansion of Art. 588 can be used to determine
the sum, or at any rate the number, of all the homogeneous
products of any dimensions (Art. 88) of any number of
letters a, b, c, ....

1
For = 1 + ax + a²x² + ... ,
1 ax

1
= 1 + bx + b²x² + ... ,
1-6x
1

1-сх
= 1 + cx + c²x² + ... ,

Hence
1

(1 - ax)
x) (1-6x)
(1 ( 1 - x)... = (1 + ax + a²x² + ...)
(1 + bx + b²x² + ...) (1 + cx + c²x² + ...) ....
On the right-hand side the index of x in any term on the
product will evidently be equal to the sum of the indices
of the letters a, b, c, ... which occur as cofactors. For every
term must be of the form
a xx bºxx cx" ... ,
or abc" ... xp+ q+ r+ ....

Hence the coefficient of any power of a, as a", on the right-

hand side will be the sum of all possible products of powers
of a, b, c, ... the sum of whose indices is n; that is, will be
the sum of all the homogeneous products of the nth degree
of the letters a, b, c, ....
1
If the fraction be resolved into
(1 - x) ( 1 - x) ( 1 - сх)...
partial fractions by the method of Art. 591, into the form
A B C
+ ....
1 +1
1-bx + 1-сх
598.] and Recurring Series. 387

the coefficient of a" is Aa" + Bb" + Cc" + ..., and from this
form the sum of the homogeneous products of n dimensions
can be easily calculated.
598. The number of the homogeneous products is the
value which the sum assumes when each of the quantities
a, b, c,... is made equal to unity, since in that case each
such product will be unity.
But if a, b, c, ... be each unity the fraction
1 1
becomes , or (1 - x) ",
(1 - x) ( 1 - x) ( 1 - сх)... (1 - x)"
if there ber of the letters a, b, c....
Hence the number required is the coefficient of " in the
expansion of ( 1 -x)-". But by the formula of Art. 572
the n + 1th term in this expansion is
(-7) (-7-1)...(--n + 1) (−x)» ,
n

which reduces, as inArt. 578, to

\2nr (r + 1) ... (r + n - 1 • 2".
(-1 )2 " n

And since ( -1)2* is always + 1, the coefficient of a" in

the expansion is
r (r + 1 )... (r + n − 1),,

which is the number of homogeneous products required.

CC2
388 Undetermined Coefficients and Recurring Series.

EXAMPLES.

3-2x
1. Resolve into partial fractions 1-3x+ 2x²
3-2x

1-3x+ 2x² by the method of

2. Expand to five terms
Art. 589, and also by the help of the last example and the
binomial theorem. Find the nth term of the series.
3-2x
3. Resolve into partial fractions
(1 - x) ( 1-2x) ( 1-3x)
1
4. Resolve into partial fractions 1-x³
5. Find the nth term and the sum of n terms of the re-
curring series 3 + 7x + 15x² + 31x² + 63x* + ....
6. Find the scale of relation in the series
a+ (a + b) x + (a + 2b) x² + (a + 3b) x² + ... ,
and hence sum the series to n terms.
7. Find the scale of relation in the series
a² + (a + b)²x + (a + 2b)²x² + (a + 3b)²x² + ... ,
and sum the series to n terms .

Deduce the sum of the squares of the numbers

1, 2, 3, ... n .
 CHAPTER XXIII.

SUMMATION OF SOME SPECIAL SERIES.

599. Ir the general, or nth, term be represented by un ,

and the sum of the first n terms bySn, both un and Sn, will
be functions of n (Art. 165).
Also, u, being the first term,
Sn-1 = U1 + U2 + ... + Un-1
S = U1 + U2 + ... + Un -1 + Un.
Therefore Sn- Sn-1 = un (a)
Hence if u can in any given case be resolved into the
difference of two terms one of which is the same function
of n that the other is of n - 1, that is to say, one ofwhich
involves n in the same manner as the other does n - 1 ,
the former term will be a possible value for Sn.
600. There are two large classes of forms of un forwhich
this resolution is possible.
The first of these is when u is the product of a fixed
n

number of factors in arithmetical progression, the first

factor of each term being the same as the second of the
previous term.
The general form of un will thus be
(an + b) { a (n + 1) +6 } {a (n + 2) + b} ... {a (n + m - 1) +6 } ,
where m is the fixed number of factors, a is the common
difference of successive factors, and b is a constant.
The term may be denoted by the symbol Un, m the
letter m being introduced to mark the number of factors.
 390 Summation of some Special Series. [600.
The symbol un, m + 1 will consequently denote the product of
m + 1 factors of which the first is an + b, or
Un, m + 1 = (an + b) (a.n + 1 + b) (a . n + 2 + b)...
...(a.n + m - 1 + b) (a.n + m + b)
= {a (n + m) + b }un, mo (1)
Again Un- 1, m+1 is the value of un, m+ 1 when n - 1 is
written for n, or
Un- 1, m + 1 = (a.n - 1 + b) (an + b) (a.n + 1 + b)...(a.n + m - 1 + b)
= (a.n - 1 + b) Un, m. (2)
Hence, subtracting (2) from (1),
Un, m + 1 - Un-1, m + 1 = [ { a (n + m) +6 } - {a (n - 1) +6} ]Un, m
= (m + 1)a . Un, mo
Hence Un, m+1 Un-1, m +1 .
Un, m = -

(3)
(m + 1) a (m + 1 ) a
Comparing this equation with (a) of the last article, it
is evident that S, may have the value (mUn,+m+1 ; but since
1) a '
equation (3) is not altered if any quantity C be added
to both terms on the right-hand, so that it may be written
=
Un, m+1 - Un-1, m + 1
Un,m
(m + 1) a (m + 1) a
it is obvious that all that can be fairly inferred is that
Un, m +1 + C,
S = (4)
(m + 1) a
when C must depend on the point in the original series
from which the summation is commenced.
If the series is to be summed from 21, m inclusive, the
value of S must be zero when n = 0. Hence
0 =
Uo, m + 1
+C;
(m + 1) a
602.] Summation of some Special Series. 391

Uo, m+ 1
whence C=
(m+1)a'
and Un, m + 1 - Uo, m + 1
Sn =
(5)
(m + 1) a
601. Equation (5) may be derived from equation (3)
somewhat differently ; thus-
Writing in succession the values 1 , 2, 3, ...
n for n, (3)
gives
U1, m = U1, m + 1-40, m + 1 ,
(m + 1) a
=
U2, m + 1 - U1 , m +1 ,
U2, m
(m + 1) a
U3, m = Из, т +1 - U2, m+ 1 , ,

(m + 1) a

и =
Un, m +1 - Un-1, m +1
Un, m
(m + 1) a
Adding all these results, the sum of the terms on the
left hand is S. while that of the terms on the right is
Un, m +1-4o, m +1 , since the second term in the numerator of
(m + 1) a
each fraction is the same, with opposite sign, as the first
term in the previous one. Hence, as before,
S = Un, m +1-40, m +1
(m + 1)a
=
(an + b)(a.n + 1 + b)...(a.n + m + b) - (a + b)(2a + b)... (ma +b)
(m + 1) a
602. As an example take the series
1.2.3 + 2.3.4 + ....

Here m = 3 and un, 3 = n (n + 1) (n + 2), a = 1, b = 0 .

Consequently un, 4 = n (n + 1 ) (n + 2) (n + 3), Vo 4 = 0.
392 Summation of some Special Series. [603.

Hence S₁
Sn = n (n + 1) (n + 2) (n + 3).
4

603. A second very important class of algebraical series

are those of which the general term is the reciprocal of un, m
as defined in Art. 600, so that the series is
1 1 1 1
+ + + ... +
U2, m u
U1, m Из, т 71, 77
%

1 1
The expression differs from by containing
Un, m-1 Un, m
m- 1 factors in its denominator instead of m.
1 1
Hence =

Un, m-1 (an + b) (a.n + 1 + b) ... (a.n + m −2 + 0)

a (n + m - 1 ) + b
=

(an + b) (a.n + 1 + b) ... (a.n + m - 2 + b) (a.n + m - 1 + 0)

a (n + m - 1) +6
Un, m
Similarly
1 1

Un+1, m-1 (a.n + 1 + b) (a.n + 2 + b) ... (a.n + m − 1 + b)

an + b
=

(an + b) (a.n + 1 + b) ... (a.n + m −1 + 6)

an + b
=

11, m
Hence
1 1
(a.n + m − 1 + b) (an + b)
Un, m-1 Un +1, m- 1 Un, m

= a (m- 1)
Un, in
Replacing n by n - 1, it follows that
1 1
a (m -1) ,

Un-1, m-1 Un, m-1 Un-1, m

604.] Summation of some Special Series. 393

and so on, 1 1
a(m - 1) ,

Un-2, m--1 Un-1, m-1 Un-2, m

1 1
= a (m - 1)
U1, m- 1 U2, m-1 U1, m

Adding all these results, the second term in each expression

on the left-hand side goes out against the first term on the
previous expression, while the right-hand side becomes the
sum required multiplied by a (m- 1).
Hence, if S denote the sum required,
1 1
a (m - 1) S = ,

U1; m-1 Un+1, m-1

or
Sn = (m-1)
1)a {( Ui, m-1 Un+1, m-1 }
604. As a special example, let the series be
1 1 1 1
+ +
1.2.3 2.3.4 3.4.5 + ... + n (n + 1) (n + 2)
Here a = 1 , b = 0 , m = 3 ; consequently
Sn =

21.2-(n +1)(n+2)}
This value of S can ofcourse be obtained independently of
the general formula by applying the method of the last
article to the resolution of the general term
1

n (n + 1) (n + 2)
Thus it is easy to verify that

n(n+1)(n + 2) = 2 {n (n+1) (n+1)(n+2)}'

whence the required result is obtained for S.
394 Summation of some Special Series. [605.
605. The series of which the general term is nº can be
summed by a process not dissimilar in principle to that of
Art. 601. Thus, by the Binomial Theorem (Art 544),
(n + 1)2+1 = NP + 1 + ( p + 1) n² + (p+1.21)Pnp-1 + ...
... + ( p + 1 ) n + 1 ,
or

(n + 1) + 1_n + 1 = ( p + 1) n² + (p1.2
+ 1)Pnp-1 + ...
... + ( p + 1) n + 1. ( 1 )
Replacing n in succession by n - 1, n - 2, ... 2, 1,
nP + 1 - (n - 1 ) +1 = ( p + 1) (n - 1)
(p + 1)p
+
1.2 (n - 1) -1 + ... + (p + 1) (n - 1) +1, (2)
(n - 1) + 1-(n - 2) +1 = ( p + 1) (n -2)
(p + 1)p
+
1.2 (n-2) -1 + ... + ( p + 1) (n - 2) +1 , ( 3)

2P+ 1-12 + 1 = ( p + 1 ) 1
+ (P+1)P.
1.2
12-1 + ... +(p + 1) . 1 + 1. (n)

Adding all these results (1) to (n), we obtain

(n + 1) + 1-12 + 1 = (p + 1 ) Ση
En
+ (p1.2
+ 1)P2n -1 + ... + (p + 1)Σn + n... , (a)
where a symbol such as En means the sum of the series
1* + 2* + ... + (n - 1)" + n*.
606. Let the values 1, 2, 3, be given successively to
...

pin (a).
607.] Summation of some Special Series. 395
It follows that
(n + 1)² - 1 = 2n + n,
(n + 1)3−1 = 3 Ση² + 3n + n ,
(n + 1)*− 1 = 4 Ση³ + 6 2 n² + 4 Ση + n.
From the first of these, it follows that
Ση = (n + 1)2-1 - n_n (n + 1)
= ,
(a)
2 2

a result deducible from (3) ofArt. 506. The second gives,

by transposition and substitution ofthe value of En,
3 Ση² = (n + 1 )3 − 1 − n − 3Ση
32n
3n (n + 1 )
= (n + 1) { (n + 1)² - 1 } — 2

(n+ 1)
=

2
{ 2n² + 4n + 2-2-3n}
n+ 1
=

2
(2n² + n)
n (n + 1) (2n + 1)
2

Hence En² + n (n + 1) (2n + 1) 6

....
(β)
From the third similarly,
4 Ση³ = (n + 1)² - 1 - n - 62 n² - 4 Ση
= (n + 1 ) { (n + 1)3-1 } - n (n + 1) (2n + 1) -2n (n + 1)
= (n + 1) n { n² + 3n + 3- ( 2n + 1) -2 }
= (n + 1) n. (n² + n)
= n² (n + 1)².
Hence Ση³ =
n² (n + 1 )²
4
= { Σ (n) } 2. (γ)
The values of En4, Ens can be deduced in succession.
607. The general relation (a) of Art. 605 shows that
En" will involve n to the power p + 1 and to no higher
power.
396 Summation of some Special Series. [607.
This can be deduced by giving to p in succession the
values 1, 2, 3, ... in succession. It has been seen that the
highest powers of n involved in Ση, Ση², Ση³, ... are the
2nd, 3rd, 4th, ... powers respectively, and it is clear that as
pis increased by unity the highest power of n on the left-
hand side of (a) increases by unity, whence it follows that
the index of the highest power of n in En² increases as
fast as p, and must therefore always be ( p + 1 ) .
With this assumption the method of undetermined
coefficients (Art. 585) affords another means of arriving at
the value of Ση" .
For assume that

1º + 2 + ... + n² = Ap + 1 np + 1 + Apn² + Ap -1nP- 1 + ...

... + A₁n + A , (1 )
where A0 , A1 , A3 , ... Ap + 1 are quantities independent of n.
This assumption is a legitimate one since the right-hand
side of (1) is the most general form of a rational integral
function of n of the p + 1th degree (Art. 165).
It follows, by substituting (n + 1 ) for n, that
1² + 2º + ... + n + (n + 1) ² = Ap +1(n + 1)p+ 1
+ A,(n + 1)² + ... + A₁ (n + 1) + A. (2)
0°

Subtracting ( 1 ) from ( 2 ),
(n + 1 )² = Ap +1 { (n + 1)² + 1 - np + 1 } + Ap { (n + 1)² - n "} + ...
... + 41 .
By Art. 544, this identity can be written,
+
p ( p - 1) nP -2 + ... + pn + 1
1.2

= Ap+1 {(p + 1) n²NP+ (1+ 2) P np-1 +...+(1) + 1) n + 1}

p

+
p ( p - 1 ) NP-2+
A₂{Imp−1 +
P
1.2
...
+p
+ ...

+41 .
608.] Summation of some Special Series. 397

This identity being true for all values of n, it follows

(Art. 585) that the coefficients of the successive powers
of n on the two sides must be respectively equal. This
gives
1 = ( p + 1 ) Ap + 1 ; (a)
(p + 1)p
p=
1.2 Ap +1 +Pp ; (β)

p ( p - 1) _ ( p + 1) p ( p - 1)Ap+1+
A. p (p- 1) Ap
1.2 1.2.3 1.2

+( p -1)4-1. (γ)

from which the values of Ap+1) Aps Ap-1 ... 4₁ can be

successively deduced.
Since (1) must be true for all values of n, it follows by
giving to n the value zero that A, must be zero.
608. It follows from (a) of the last article that
1
Ap+1 p+ 1

Hence whatever positive integral value p may have,

1
1P + 2P + ... + n = - • np+ 1 + An² + ... + 40 .
p+ 1
Hence
1 + 2 + 3 + ... + n 1 Ap Ap-1 A
+ + + ... + +1 °
NP+1 n n2
p+ 1

If n be supposed indefinitely increased the values of all

the terms after the first on the right-hand side continually
diminish and their sum can be made less than any assign-
able quantity. This can be proved in a similar manner to
that adopted in Art. 587.
1 + 2 + ... + n
Hence the value of the expression nP+1
ap-
1
proaches more and more nearly to as n is increased,
p+ 1
398 Summation of some Special Series. [608.
1
and can be made to differ from by less than any
p+ 1
assignable quantity when n is made sufficiently large.
1
This is usually expressed by saying thatp+ 1 is the limit,
1 +2 + ... + nº
or the limiting value of the expression NP+1
,

when n is indefinitely increased.

609. The special values of En² and En³ can be easily
deduced from equations (a), (β), (γ) of Art. 607. For if
12 + 22 + ... + n² = Azn³ + A2n² + A₁n + 40 ,
it follows from those equations that
1 = 3 A3 ,
2 = 343 + 242,
1 = A3 + A2 + 41,
while in all cases A = 0 . Hence
1 1 1
A3 = 3
,
A2 = 2 , 4₁ = 6 1

n3 n2 n
and 12 + 22 + 32 + ... + n² = ++
3 2 6

=
n (2n² + 3n + 1)
6

=
n ( 2n + 1)( n + 1)
6

as in Art. 606 (β ).
Again, if
13 + 23 + ... + n³ = A₁n² + Azn³ + A₂n² + A₁n + A0 ,
3

A = 0, as before, and A4, A3 , A2, A₁ are determined by the

1

equations, 1 = 444 ,
3 = 644 +3 A3,
3 = 4 A4 +3A3 + 2 A2,
1 = 44 + 4 + 4 +4₁;
 610.] Summation of some Special Series. 399

1 1
whence A =
4
,
3
=

2 42 = 41 = 0 ;
n4 n3 n2
whence 13 + 23 +33 + ... + n³ = ++
4 4

=
n4 + 2 n² + n2
4

n² (n + 1)²,
4
the result previously obtained.
610. Any series of which the general term is a rational
integral algebraical function of n can be summed by the
method of Art. 607, or the sum required may be deduced
by separating the given series into a number of series one
for each power of n contained in the general term.
Thus if the general term of a series be n² + 3n + 2, the
series itself to n terms is

{ 12 + 3.1 + 2} + {22 + 3.2 + 2 } + ... + { n² + 3n + 2 },

which can be otherwise written as

{12 + 22 + ... + n2 } + 3 { 1 + 2 + 3 + ... + n }

+ 2 (1 + 1 + 1... to n terms) ;
which therefore is equal to

n(n + 1) (2n + 1) + 3n (n + 1) + 2n
6 2
n
=
{ (n + 1) (2n + 1) +9 (n + 1) +12}
2

= { 2n² + 3n + 1 + 9n + 9 + 12}
n

= { 2n² + 12n + 22 }
n (n² + 6n + 11)
3
400 Summation of some Special Series. [610.
This can be written in the form
(n³ + 6 n² + 11n + 6) -6
3

=
(n + 1) (n + 2) (n + 3) -2 ,
3

and in this form the result will be recognised as a

particular case of Art. 601, since the given general term
n² + 3n + 2 can be also written as (n + 1) (n + 2).
A series whose general term is a rational, but fractional
function of n can sometimes be summed by special artifices
similar to that of Arts. 603 or 604.
2n +3
Thus, if the general term be the
n (n + 1) (n + 2) (n + 3)'
series can be summed by taking the general term as the
sum of two terms,
2 3

(n + 1) (n + 2) (n + 3) + n (n + 1) (n + 2) (n + 3)'
and applying to each of the terms the result ofArt. 603.
The general term can also be resolved into either of the
two forms
1 1 1(1 1 1 1
or- +
n (n + 2) (n + 1 ) (n + 3) 2 ( n n+ 2 n+ 1 n+ 3

and the sum can be obtained from these forms by the

principle ofArt. 599 .
611. Series are sometimes given for summation of which
the law of formation is not obvious, and the general term
ofwhich requires some skill for its discovery. For instance,
let the given series be
2, 7, 14, 25, 44, 79, .... (1)
By subtracting each term from its successor a new series
is obtained,
5, 7, 11 , 19, 35, ... , (2)
611. ] Summation of some Special Series. 401

of which the law is still not evident. Repeating the

process, a third series is found,
2, 4, 8, 16, ... , (3)
which is evidently a geometrical series of which the nth
term is 2". The nth term of (2) is evidently obtained by
taking the sum of n- 1 terms of (3) and adding 5, the first
term of (2). Hence the nth term of (2) is
5+ 2 "-2 or 2 " +3 .

The nth term of (1) is similarly obtained by adding the sum

of (n - 1) terms of (2) to 2, the first term of (1), and is
therefore 2+ ( 2 " -2) + 3 (n - 1 ) or 2" + 3 (n - 1). The sum
of n terms of (1) can now be obtained by the help of
Arts. 506 and 522 .
The series (2) is called the series offirst differences of ( 1 ) ;
and (3) the series of second differences. In all cases where
the nth term of the original series can be expressed algebrai-
cally as an integral function of n, a series of differences will
be at last reached which assumes the form either of a
geometrical or arithmetical progression .

EXAMPLES.

1. Sum ton terms 2.3 + 3.4 + 4.5 + ....

2. Sum ton terms 3.5.7 + 5.7.9 + ....
3. Sum to n terms
2.2.3 + 4.3.4 + 6.4.5 + 8.5.6 + ....
1 1 1
4. Sum to n terms + + + ....
2.3.4 3.4.5 4.5.6

1 1
5. Sum to n terms + + ....
3.5.7 5.7.9
Dd
402 Summation of some Special Series.
1 1 1
6. Sum to n terms + + + ....
1.3 3.5 5.7

1 1 1
7. Sum to n terms + + + ....
22-132-142-1

8. Find the nth terms and the sum to n terms of

1.2 + 2.3 + 4.5 + 7.8 + 11.12 + ....

9. Sum to n terms n + 2 (n - 1) +3 (n - 2) + ....

10. Find the nth term and the sum of n terms of the series
2 + 3 + 5 + 9 + 17 + 33 + ....

11. If every term of a series be subtracted from the following

one, find the nature of the resulting series (1) when the original
one is a geometrical series, (2) when its nth term
= An + BnP-1 + ....
Find the nth term and the sum of n terms of the series
6, 16, 34, 68, 142, 328, 842, ....
12. Find the nth term, and the sum to n terms, of the series
1, 2, 4, 5, 7, 8, 10, 11 .
13. Find the nth term and the sum of n terms of the series
4, 6, 13 , 18, 31, 48, 85, ....

1 1 1
14. Sum to n terms + + + ....
1.2 2.3 3.4

15. Sum to n terms 2.3 + 4.5 + 6.7 + ....

16. Sum ton terms 1.2.3.4 + 2.3.4.5 + ....
1 1 1
17. Sum ton terms + + + ....
2.55.8 8.11

1 1 1
18. Sum ton terms + + + ....
1.5.9 5.9.13 9.13.17

19. If s₁ be the sum of n terms of the arithmetical pro-

gression a, a + b, a + 2b, ... , and s₂ that of n terms of c, c +d,
c+ 2d, ... , then, if S be the sum of n terms formed by mul
 Summation of some Special Series. 403

tiplying together the corresponding terms in each of the former

two, prove that
n²(n - 1) bd.
nS-8182 = 12

20. Find the nth term and the sum of n terms of the series
2 5 8
+ + .....

2.3.4 3.4.5 4.5.6

21. A person discharges a debt in n years by paying at the

end of each year sums in the ratio of 1, 2, 3, ... ; what was
the first payment ?

Dd2
 CHAPTER XXIV.

EXPONENTIAL AND LOGARITHMIC SERIES .

612. THE word ' limit' has been introduced in Art. 608.
As it will be necessary to make some use of this term in
the following articles it will be as well to repeat more
formally the definition.
The limit of a function (Art. 165) of any variable
quantity, for any assigned value of that variable, is a
quantity towards which the value of the function con-
tinually approaches as the variable approaches its assigned
value and from which the value of the function may be
made to differ by less than any definite magnitude, by
making the variable approach sufficiently near to its
assigned value.
613. Thus when r is less than unity the limit of the
sum of the geometrical series a + ar + ar² + ... to n terms is
a

1- r
,
when n is indefinitely increased (Arts. 528, 529),
1 + 22 ... + n
and the limit of when n is indefinitely in-
np+ 1
1
creased is 1 (Art. 608).
p+ 1
614. We have now to investigate the value of a very
1."
important limit, namely, that of (1 + when n is in-
n

definitely increased.
By the Binomial Theorem,
2

(1 + 1)* = 1+ (1) + (n-1) (1)² +...

n 1.2 n

... + n (n(n-- 1)1 )...

r
... (n - r + 1 )
() .... yw,
 Exponential and Logarithmic Series. 405

This is true whether n be integral or fractional. In

the latter case the series on the right-hand side contains
an infinite number of terms, but if n be greater than
1
unity, and consequently n
less than unity, the series is
convergent (Art. 576), and the equation therefore implies
an arithmetical equivalence for all large values of n.
The series on the right-hand side may be somewhat
differently written, thus
1
1-
(1+ 1
= 1+1+ n (1 n )(1-3)
+ + ...
n 1.2 1.2.3

1 2 r- 1
(1 n (1- n
... ( 1- n
... + + ... , (a)
1.2.3 ... r

the r + 1th term being reduced to this new form by dividing

each of ther factors in the numerator of its coefficient by
one of the factors n in the denominator, n".
The series on the right-hand side of (a) is always less
than the series
1 1 1
1+1+
1.2
+
1.2.3 + ...
...

r
+ ... ad inf. (β)

Further, the right-hand member of(a) approaches nearer

to (3) the larger n becomes ; and by making n sufficiently
large the difference between the right-hand member of (a)
and the series (3) can be made less than any assignable
1
quantity. Hence the limit of (1 + " when n
nis in-

definitely increased is the series (3) .

615. The numerical value of the series (3) is usually
denoted by the letter e. It follows from Art. 576 that (3)
must be convergent, since the right-hand member of (a)
406 Exponential and Logarithmic Series. [616.
is so for all values of n greater than unity; or the follow-
ing special proof may be given
1 1 1
1+ 1 +- + + + ... ad infin.
1.2 1.2.3 4

1 1 1

< 1 + 1 + 2 + 2 + 23 + ... ,
since each term, after the third, of the first series, is less
than the corresponding term of the second, and the first
three terms are identical. Hence the series (3)
1
<1 + - (Art. 528),
1-1
2

< 1 + 2,
< 3.

Hence the value of the series (3), or e, is a finite

quantity less than 3 .
616. The value of e can be approximately calculated to
any required number of terms. Thus,
1+ 1 = 2.0000000
1
= •5000000
1.2

1
1.2.3
=
•1666667 dividing the line above by 3
1
= -0416667 ‫وہ‬ " ‫در‬
4
1.2.3.4
1
= 0083333 "
5
5
1
= -0013889 "
6
6

1
= -0001984 " "
7
7
1
= -0000248 " " .
‫و‬
8
18
1 = -0000028 9
" "
9

1 -
0000003 ‫دو‬
10
10

2-7182819
617.] Exponential and Logarithmic Series. 407

The last figures in the several rows have been increased by

unity when the first figure omitted is greater than 4.
The sum of the remaining terms of the series
1 1
+ + ...
11 12

is less than 1 1 1

11 +++...}
1+
11 112
1

11 1
or < <
1 1010
1
11
1
<
36288000

< .00000003,
so that the above result is probably accurate to seven
places, certainly to six places, ofdecimals.
1 n

617. Since the limit of (1 + when n is indefinitely

n

increased is e, it follows that the limit of (1 + 1nx which n

,

1
is equal to {(1 +
n )"}, is e under the same conditions.
But, by the Binomial Theorem, 2
1nx nx

(1 + 1) ²= 1 + no. 1 + no(nx- 1) (1)

n
nx
n 1.2
3
пх (пх- 1)
+ nx (mx-1) (n -2)(1) + ... 3

nx

... + nx (n - 1)... (no - r + 1) (1) + ...

r N

1 1 2

n
= 1 +x+ + + ...
1.2 3

x(x - 1) (x- ) ... ( - )

... + + ....
r
 408 Exponential and Logarithmic Series. [618 .
And, as before, the limit of the series on the right-hand
side, when n is indefinitely increased, is
x²
2 23
1 ++++ ... +++ ....
2 r

Hence, finally,
x2
1 + x+ + +... + + ... ad inf.
2 3 r

This equation is frequently known as the exponential

theorem.
618. A somewhat more general relation can be deduced
from the theorem of the last article.
Writing cæ for æ, it becomes
cx
ex = 1 + cx + + ... + + ....
+
2 r
2

Hence, if e = a, (1)
this gives
c2x2 C3x3
ax 1+ cx+
2
+
3
+ ... +
r
+ .....
(2)
1
This relation gives an expansion of a in ascending
powers of a, the only new quantity involved being the
quantity c, which is connected with the given quantity
a and the known number e by the relation (1).
This quantity e is of so much importance for many pur-
poses, both theoretical and practical, that it has a name
given to it. It is called the logarithm of a to the base e.
619. The general definition of a logarithm may be thus
given.
The logarithm of a number to a given base is the index of
the power to which the base must be raised to be equal to the
:
number.
Thus in (1), e = a. Hence c is the index of the power
(Art. 62) to which e, the base, must be raised in order
621.] Exponential and Logarithmic Series. 409
that the result may be a ; c is therefore the logarithm
of a to the base e, or, as it is concisely written,
c= loga.
In the general way if a = m, æ is the logarithm of m
to the base a, or x = logam.
The two equations a = m,

x = logam,
must be regarded as merely two different ways of stating
one relation between the three quantities a, m, x.
620. The equation (2) ofArt. 618 can now be written
a² = 1 + x loga +
(xloga) + (xloga)3 .....
2 3

621. The relation of the last article can be obtained

by another method which leads to some other important
results.
By the Binomial theorem,
a* = ( 1 + a − 1 )*
= 1 + x (a- 1) +
x(x-1)
1.2 (a-1)2
x (x - 1 ) (x - 2)
+
1.2.3 (a - 1)3 + .... ( 1)
The coefficients of the successive powers of (a-1) in-
volve different powers of ax. The whole expansion can
therefore be arranged in powers of x. The only term in-
dependent of a is the first term, while the first power of a
occurs in every succeeding term. The coefficient of this
first power is thus an infinite series, the successive terms
of which involve successively higher powers of (a - 1).
The term of this series arising from the r + 1th term
of(1), or x (x - 1) ... (x - r + 1 )
r
(a- 1) ,
is ( -1) ( - 2) ... (- r + 1) (a- 1)
r
or (-1) - (a- 1)"
(
410 Exponential and Logarithmic Series. [621.
Hence the coefficient ofx in (1) when rearranged is
a-1- (a-1)² + (a- 1)3... + (-1) - (a - 1) + .... (2)
r

Let this be called c, and let the coefficients of x2, x3, ...
be denoted by C2, C3 , ... , these quantities being functions,
at present undetermined, of a and independent of x. Hence
a* = 1 + cx + c2x² + C3x3 + ....
It follows that
a"= 1 + cy + c2y² + C3 Y³ + ... ,
and that
ax+ v= 1 + c (x + y) + C2 (x + y)² + C3 (x +y)² + ... ;
since C, C2, C3 ... are the same whatever be the value of x.
But ax+ y = a x a" for all values of x and y.
Hence

(1 + cx + C2x² + C3x3 + ...) (1 + cy + C2y² + Czy³ + ... )

= 1 + c (x + y) + c₂ (x + y)² + c3 (x + y)3 + ... ,
for all values of x and y.
Therefore, by Art. 587, the coefficients of the same powers
of æ on the two sides must be equal. Equating the co-
efficients of the first power of x,
c (1 + cy + c2y² + c3 y³ + ...) = c + 2 C2 Y + 3 C3 Y² + 4 C473 + ....
This identity must be true for all values of y; hence
equating the coefficients of the different powers of y on
the two sides,
C2
C2 = 2C2, therefore c₂ =
C3
CC2 = 303, C3= 2.3
C4
CC3 =404, C4= ;
2.3.4

cn
C'n-1= NCn, Cn
2.3 ... n
 623.] Exponential and Logarithmic Series. 411

Whence
c2x2 cn xn
a 1 + cx + + + ... + + .... (3)
2 3 n

1
By giving to a the particular value , which makes cx
C

equal to unity, we get

1
1 1 1
a = 1 +1+ + + ... + + ....
23 n

This is the quantity previously called e. Hence

1

a = e,
or a = e,

whence c = loga,
and equation (3) becomes
a* = 1 + x log a +
(xloga)2 + ... + (xloga)" + ... ,
2 n

the relation ofArt. 620.

622. Since c is equal to the expression (2) of the last
article, it follows that
1
loga = a- 1- 2(a-1)² + (a- 1)3...+ (-1) - (a- 1)" + ... ;
r

or replacing a by 1 + z,
1 z
log ( 1 + 2) = z − 222+ ...+ (- 1) -1 + ... , (a)
a series from which the value of log.(1 + z) may in some
cases be calculated.
623. The series for e is only certainly convergent when
a- 1 is less than unity, or a is less than 2. Thus the
proof of Art. 621 is strictly only applicable when a is less
than 2. The proof for cases when a is equal to or greater
than 2 can be easily deduced. For if a be any number
and A any number less than 2 and greater than unity, it
is always possible to find a number y such that
A = a.
412 Exponential and Logarithmic Series. [623.
Hence a = Av ,
and since the theorem of Art. 621 has been proved for
all numbers less than 2,
(yx log.4)2 + ....
Av = 1 +yx log.A +
e
2

Now if e = A, c = log.A.
Also ev = A", and cy= log. A",
= log.a.
Hence a (ylog A)2x² + ...
Av*= 1 + (ylog A) x + 2

= 1 + x log a +
x² (loga)2 + ... ,
2

which proves the theorem for all values of a.

The proof contained in Arts. 614-620 is independent of
any limitation to the value of a.
624. Other series can be derived from (a) of Art. 622
which are often more convenient for purposes ofcalculation.
Before entering on the discussion of these, it is necessary to
examine more fully the nature and properties of a logarithm
to any base.
Let m and n be any two numbers, and a any base. Let
x and y be such numbers as to satisfy the equations
a* = m, a = n,
so that, by definition (Art. 619),
x = logam, y = logan.
By multiplication
mn = a.a² = a²+ v .

Hence by the definition of a logarithm (Art. 619),

loga (mn) = x +y
= logam + logan.
It easily follows that
loga (mnp) = loga (mn.p)
= loga (mn) + loga (p)
= logam + logan + logap .
629 ] Exponential and Logarithmic Series. 413

Thus it follows that the logarithm of a product is equal

to the sum of the logarithms of the factors.
625. Similarly by division
m ax
α -ν,
n

whence by definition
m

loga n
= x -y,

= logam - logan, (2)

or the logarithm of a quotient is equal to the logarithm of the
dividend diminished by that of the divisor.
626. Again, with the same notation, if
m = a*,
it follows that mp = apx ;

whence again by definition (Art. 619),

loga m² = px,
= p logam. (3)
This formula holds whether p be integral or fractional.
627. The formulae (1), (2), (3) of the last three articles
are those on which the practical utility of logarithms
depends. If tables be calculated giving the logarithms of
all numbers to any base these will enable the operations
of division and multiplication to be performed by means
of addition and subtraction, and formula (3) will give the
means of raising to powers or taking roots by means of
multiplication or division.
628. The base to which tables of logarithms are usually
calculated is 10, the radix of the ordinary arithmetical
notation. This has some important practical advantages
in simplifying both the construction and the use of such
tables.
629. With this number as base, the only numbers whose
logarithms will be integers are 10, 102, 103, 104, ... , or
414 Exponential and Logarithmic Series. [629.
10, 100, 1000, 10000,.... The logarithms of all other
numbers will be fractions or mixed numbers. Thus, since
the number 7834 is greater than 1000 and less than 10,000
its logarithm must lie between the logarithms of these
two numbers, that is between 3 and 4. The integral part
of its logarithm is consequently 3. The integralpart ofthe
logarithm of a number is called the characteristic of the
logarithm : thefractional part when expressed as a decimal,
is called the mantissa of the logarithm.
Thus the characteristic of the logarithm of 7834 to the
base 10 is 3. Since any number containing p + 1 digits
in its integral part is greater than 10 and is less than
10 +1, it follows that the logarithm of any such number
lies between pand p + 1. Hence the characteristic of the
logarithm to the base 10 of any number containing p + 1 digits
in its integral part is p, one less than the number of digits.
The use of the base 10 thus enables the characteristic of
the logarithm of any number greater than unity to be
determined by inspection.
630. The logarithm of any number to the base 10 being
ascertained, those of all numbers obtained from the first
by merely changing the position of the decimal point can
be written down by the help of Arts. 624, 625.
Thus the logarithm of 52185 to the base 10 is 4.7175457 .
The numbers

5218-5, 521-85, 52-185, 5-2185, 52185, 052185 ...

are obtained from 52185 by dividing successively by 10.
Hence the logarithms of these numbers will be derived
from 4.7175457 by subtracting unity, the logarithm of 10,
and will all therefore have the same decimal part, namely
7175457, and for their characteristics the integers
3, 2, 1 , 0, -1 , -2 .
Thus, since any two numbers which are formed by the
632.] Exponential and Logarithmic Series. 415

same digits in the same order, and only differ in the

position of the decimal point, can be obtained from each
other by multiplying or dividing by some integral power
of 10, it follows (Arts. 624, 625) that the logarithms of any
two such numbers to the base 10 must differ by some
integer and must therefore have the same decimal part or
mantissa.

631. In tables of logarithms to the base 10 it is only

therefore necessary to give the mantissa of the logarithm
corresponding to any set of digits for the number. The
characteristic can be determined by the rule of Art. 629
when the number is greater than unity.
When the number is a proper fraction expressed as
a decimal, as 52185, the logarithm is a negative number.
In the last article it has been seen that this logarithm is
expressible as -1 +7175457, which is more usually written
1-7175457, the sign -, written over the characteristic,
indicating that it affects the characteristic only, and that
the mantissa is a positive quantity.
The logarithm of .052185 is similarly 2-7175457, and
the general rule is easily deduced that, the characteristic of
the logarithm of a decimal fraction is negative and numeri-
cally greater by unity than the numbers of ciphers before
the first significant digit.
The convention that the logarithms of a proper fraction
shall be always written in the form of a negative integer
with a positive fraction makes the rule of Art 630 abso-
lutely general.
632. The last three articles exhibit the great advantage
of 10 as a base for practical purposes. Most of the published
tables of logarithms contain full directions for their use, and
to these the student is referred for further exemplification
of details. One of the most convenient collections of tables
is the volume of ' Mathematical Tables' in Chambers'
416 Exponential and Logarithmic Series. [633-
Educational Course, published by W. and R. Chambers,
Edinburgh.
633. The fundamental series for the calculation of
logarithms is (a) of Art. 622. Other more advantageous
series will be deduced presently: but as the logarithms
found from these series will all correspond to the base e, it
is necessary to find a means ofderiving the logarithms to
the base 10.
634. Let a and b be two different bases, and let x and y
be the logarithms of the same number N, corresponding to
these two bases respectively. Thus
=
logaN3 . (1)
y = log N
Consequently, by the definition of alogarithm (Art. 619),
a = N,
b = N.

Hence a = bu, and therefore av = b,

y

and b = a.
Hence, by definition of a logarithm,
= loga b
y
(2)
х
= log a
Thus x = y loga b,
or, loga N = logo Nxlogab by (1). (3)
Similarly, y = x log a,
or, log N= loga Nx log, a. (4)
Incidentally it follows from (2) that
loga bx log a = 1 . (5)
635. As a particular example of (3) or (4) we have
log. N= log10 Nx log. 10 ;
1
whence log10 N = xlog.N.
log, 10
638. ] Exponential and Logarithmic Series. 417

Hence if the logarithms of all integers, including 10,

to the base e, be calculated, those to the base 10 can be
1
deduced by multiplying by log, 10
1
636. This multiplier, is often called the modulus
log, 10'
of the system of logarithms to the base 10.
637. The series (a) ofArt. 622 is
1 1

22+3 23
log. (1 + 2) = 2-22 ....
(a)
If for z we write - z, the formula becomes
1 1
log. (1-2) = -2-22 -23- ... ‫و‬ (β)
3

since the signs of the odd powers only of z are changed

(
Art. 126).
Hence, subtracting (3) from (a),
1 1
log (1 + 2) -log ( 1-2) = 2 {2+ 2 + 5 25
25++ ...}.
3

But by Art. 625 the left-hand side of this equation is

1+ 2
equal to log. 1-2
1+ 2 1 1
Hence, log. 1-2 = 2 2+
{ 2 +52 + ...} . (γ)

The series inside the bracket will be rapidly convergent

if z is small, and thus, if log.(1-2) be known, log. (1 + z)
can be easily computed.
638. Let a quantity n be taken so as to satisfy the
equation n+ 1 1+ 2
n 1-2

Whence, by solution as a simple equation with regard

to z, it easily follows that
1

2 = 2n+ 1
Ee
418 Exponential and Logarithmic Series. [639.
Hence (y) gives,
3

logo +1 = 2{2 +1 + (2n+1) +...} ,

whence
1
log.(n + 1) = log n+
n+ 2
2 2n + 1
2
+ (2n+1) +...}. (8)
From this series the logarithms of different integers can
be successively calculated.
639. Since aº = 1 (Arts. 70, 246) whatever a may be,
it follows that the logarithm of unity to any base is zero.
Hence if in (8) of the last article we put n equal to unity,
it gives
log. 2 = 2 +++...} .
3

The calculation to a few places of decimals can be easily

effected. Thus,
1 1

3
= 333333333 ; ..
3
= 333333333 ;

dividing by 9,
3
1 1

3 ()
= 037037037 ; 012345679
3

33
1 115 = 000823045
= 004115226 ;
35
1 117
二3

37
=
000457247 ; )=
(3 000065321

9
1 11
= 000005645
39
= 000050805 ; 9 ()
1 1 111
= 000000513
11 (3)
=
000005645 ;
311
11 13
1
= 000000627 ; = 000000048
313 13
(3)
1 1115
= .000000005
315
= .000000069 ;
15 ()
346573589
2

693147178
640.] Exponential and Logarithmic Series. 419
Hence with probable accuracy to eight places of decimals,
log, 2 = 693147178 .
Hence, since log 4 = 2 (log, 2) (Art. 626), it follows that
log. 4 = 1.386294356.
640. Again, putting 4 for n in (8) ofArt. 638,
1 113 115
log. 5 =
log4 + 2 + 39 ‫واة‬
The series in the bracket is more rapidly convergent than
the former one, and the calculation can be conducted in the
following way.
1 1
Thus, = 111111111 ; 111111111
9 9
1

92
= .012345679 ;
1

93
=.001371742; (3) = 000457247
39
1

94 = .000152416 ;
1 115
95
= 000016935 ; ( ) = .000003387
59
1
= .000001882 ;
96
7
1 1
= 000000030
9

= .000000209 ;
97 79

1
= 000000023 ;
98
1
= .000000003 ;
99
-111571775
2

-223143550

log4 = 1.386294356
Hence log. 5 = 1.609437906
Again, log. 2 = 693147178
2-302585084
E e.2
420 Exponential and Logarithmic Series. [641.
But (Art. 624) log 10 = log 2+ log. 5. Hence
e

log10 = 2.302585084 .
641. By Art. 635,
log, 2 -693147178
log10 2 = =

2-302585084
log, 10
and effecting the division there results
log102 = 30102999 ... ;
or, to seven places of decimals,
log102 = 3010300 .
The logarithm of 2 being determined, that of 3 can be
deduced by (8) of Art. 638, and so on for all succeeding
integers. It will not however be necessary to calculate
independently the logarithms of any but prime numbers
since the logarithm of a number composed of two or
more factors can be deduced from those of the factors by
Art. 624 .
642. The first figure in the value of the logarithm of 2
to the base 10 can be deduced by means of simple Arith-
metic.
Thus, 210 = 1024, and therefore 210 is very slightly
greater than 103. Thus 2 is very slightly greater than
10 , and log102 exceeds 3 by a very small amount.
643. It has been seen that for important practical
reasons it is desirable always to keep the mantissa of a
logarithm a positive quantity. Logarithms of proper
fractions are thus always the algebraical sum of a nega-
tive integer and a positive fraction. In any case where
such a logarithm has to be divided by an integer it is
convenient to add to both negative and positive part the
smallest number which will make the negative part exactly
divisible by the divisor. Thus the quotient comes in the
desired form of a negative integer and a positive fraction.
643.] Exponential and Logarithmic Series. 421

For instance, let it be required to divide the logarithm

2-8450980 by 3. The process adopted is really the fol-
lowing :
1 (2.8450980) = 1 ( - 2 + 8450980)
3
= 1 (-3 + 1-8450980)
= -1 + 6150327

= 1-6150327.

EXAMPLES.

1. Find the logarithm of 32 to the base 2.

2. Find the logarithm of 625 to the base 5.
3. Find the characteristics of log, 497 and log, 400.
4. Given log102 = 3010300 and log103 = 4771213 : find
the logarithms of 6, 8, 9, 24 and 36.
5. Given log105 = 6989700 : find log101.25.
6. Given log10 12 = 1.0791812 and log10 15 = 1-1760913 :
find the values of log10 2, log10 3, and log105.
7. From the logarithms given in question 4, find the logarithms
of √2, and 0125.
8. Given log10 7 = 8450980, log10 2 = 3010300 : find the
64
values of log10 ,log 9.8, log 350.
√1715
9. Find thevalues of the logarithm of 125 to the base 25, and
of the logarithm of 343 to the base 7 .
10. From the logarithms in question 8, find the value of
4
log100 (49) .
11. Given a = 7, b = 6, c = 5, s = (a + b +c), and the
logarithms in question (4) : find the logarithms of
(s - b) (s -c)
√s (s - a) (s - b) (s-c) and
8
(1-8) 8
422 Exponential and Logarithmic Series.
12. Shew that
1 1 1 1 1 1
log2 = 1-
1- 2 + - + ... = + + + ....
3 4 1.2 3.4 5.6

1 1 1 1
13. Sum the series + + 3.7 + 4.9 + ... to infinity.
1.3 2.5

1 1
1+ + + ...
2 4
e²+ 1
14. Prove that
e2-1 1
1 +++ .. ...

15. Prove that

10
1 1
loge (x + 1) = loge (x - 1) +2 +
3x
+
5x +...}.
16. Prove that
1 1 1
loge(x+1) = 21
2 logo - logo(x -1)- 2 + 2x t
+
3x +... } .
17. If a, b, c be consecutive integers, prove that
1 1

log. c = 2 log. b - loga - 2 {

(2ac+ 1
+
3 (2 ac + 1) +... }.
8 5 7 9
= + + + ad. inf.
18. Prove that loge e 1.2.3 3.4.5 5.6.7

19. A person has a capital of £A at interest. He spends

3
annually the interest on £A, taking the difference between his
expenditure and income, at the end of each year, from his
capital. At the end of five years he reduces his expenditure in
the ratio of 2 to 15, and invests at the end of each year his
surplus income. If at the end of 5 years more his capital is
9
increased on the whole by 200 A, find the rate per cent.

Given log10 13 = 1-1139434, log 1053 = 3-0224284,

log10543-0228106.
20. If yº = ", prove that plog, a = q log, a.
21. Prove that if loga a, loge b, logyc, ... be in arithmetical pro-
gression, loga a, logs 3, log, y, ... are in harmonical progression.
 Exponential and Logarithmic Series. 423

22. Having given that the product of the n quantities a, b, c, ...

is equal to r", and that the product of the n quantities α, β, γ, ...
is equal to p", prove that if loga a logsb, log,c, ... be all equal,
each of them is equal to log, r.
23. a, b, c, ... are (n + 1) quantities ingeometrical progression,
the common ratio being a. x, y, z, ... are the logarithms of any
number m with a, b, c ... as bases. His the harmonic mean of
the n quantities y, z,..., prove that
2x
H
n+ 3

24. Shew that money will increase more than fifty-fold in a

century at 4 per cent per annum compound interest ; having
given log2 = 301030, log 13 = 1-113943.
25. If a, b, c be in geometrical progression, and log,a, log, c,
loga b be in arithmetical progression, the common difference of
3
this progression is 2
 SECTION IV. ARITHMETICAL APPLICATIONS.

CHAPTER XXV.

RATIO, PROPORTION AND VARIATION.

644. THE term Ratio is usually defined as the relation

between two quantities of the same kind in regard to their
magnitude.
In Arithmetic this relation is usually expressed by
stating that a certain multiple of one of the quantities
is exactly equal to some multiple of the second. Thus,
if there be two lengths about which it is known that
three times the first is exactly equal to twice the second,
this fact includes every possible relation between the mag-
nitudes of the two lengths. It follows that the first length
is two-thirds of the second, and it is conventional to say
that the ratio, or relation between the magnitudes, of the
first to the second is expressed by the fraction .
645. If the two quantities of the same kind be incom-
mensurable it will not be possible to find two integers,
m and n, such that m times the first shall equal n times
the second. The ratio therefore of the first to the second
n

cannot in this case be expressed simply by a fraction m

The only way of estimating the ratio of two such

quantities is by considering all the multiples of the first
and all the multiples of the second, and noticing the
 Ratio, Proportion, and Variation. 425

manner in which the different multiples of the first exceed

or fall short of the different multiples of the second.
646. Thus, as in Art. 223, if there be two incommen-
surable quantities, and n be any integer whatever, it will
always be possible to find an integerr such that n times
the first quantity shall exceed r times the second and be
less than + 1 times the second. Each of these statements
is a relation between the two quantities in respect of
magnitude. The ratio between them must include all
such separate relations.
Suppose, for instance, that AB and CD represent the two
quantities,

A B B2 B3 B4

D D2 D3 D4 Ds

and let AB and CD be placed along side of each other,

A and Cbeing in a vertical line. Let them be produced
indefinitely to the right, and on AB produced let lengths
BB2, B2B3 ... be marked off each equal to AB; and simi-
larly on CD produced let lengths DD2 , D2 D3 ... be marked
off equal to CD. Then in the figure, since D₁ lies to the
right of B3 , it follows that four times CD is greater than
three times AB. Similarly it is seen that five times CD
is less than four times AB. Thus the manner in which
the points D2, D3 , ... lie to the right or left of the points
B2,B3 , ... involves all possible relations between the mag-
nitudes of the lengths AB and CD.
647. If there be four quantities the first and second
of which are of the same kind and consequently have a
ratio to each other; and the third and fourth also have
a ratio to each other : the condition of equality of these
426 Ratio, Proportion, and Variation. [647.
two ratios will evidently be that every relation which
subsists between the magnitudes of the first and second
must also subsist between the magnitudes of the third
and fourth.

E F F2 Fa F4 Fo

G H H2 H3 H H5 He H,

If AB, CD, in the figure of the last article, represent

the first and second quantities, and a similar figure be
constructed with EF and GH representing the third and
fourth quantities, the equality of the ratios of AB to CD
and of EF to GHwill be insured if the points D2 , D3, ...
lie among the points B2 , B3, ... exactly in the same
order as the points H2, H3 , ... lie among the points F₂,
F3, F4, ....
This condition is evidently equivalent to the definition
5 of the Fifth Book of Euclid, viz.-
The first of four magnitudes is said to have the same
ratio to the second, that the third has to the fourth,
when any equimultiples whatever of the first and third
being taken, and any equimultiples whatever of the second
and fourth, if the multiple of the first be less than that
of the second, the multiple of the third is also less than
that of the fourth, and if the multiple of the first be
equal to that of the second, the multiple of the third is
also equal to that of the fourth, and if the multiple of
the first be greater than that of the second, the multiple
of the third is greater than that of the fourth.
For, referring to the figures, Euclid's definition requires
that n and m being any integers AB shall be less than,
equal to, or greater than CDm, according as EF is less
than, equal to, or greater than GHm.
649.] Ratio, Proportion, and Variation. 427

Euclid's definition of the equality of two ratios is thus

a necessary consequence of the definition of ratio.
648. When two symbols a and b represent commen-
surable quantities the ratio of these is completely repre-
a

sented by the fraction 7.

b For from the knowledge of this
fraction every possible relation between the magnitudes of
the two things can be deduced.
Thus, if four quantities be represented by the symbols
a, b, c, d, of which a, b represent commensurable quantities
and c, d represent commensurable quantities, the equality
of the ratio of the first to the second, with the ratio of
the third to the fourth is completely determined by the
equation a
=

649. If the symbols a and b represent incommensurable

a

quantities, the fraction , may be still taken to represent

the ratio of these quantities. No meaning has, in fact,
as yet been given to a fraction of which the numerator
and denominator are incommensurable, the only assump-
tions made having been that the laws of operation were to
be the same whether a and b were commensurable or no.
Now if a and b represent incommensurable quantities,
it has been seen (Arts. 223, 225) that any integer n being
taken a second integerr can always be found such that
na > rb and < ( + 1) 6.
If c and d represent two other incommensurable quan-
tities, the ratio of which is equal to that of a to b,
according to the test of Art. 647, it follows that since
na > rb, nc > rd, and since na < (r + 1) b, nc < (r + 1) d.
Hence
a r C r r+ 1
>
n
< +1, and d is also
and <
n
>
n
and <
n
428 Ratio, Proportion, and Variation. [6+9.
This is true whatever be the magnitude of n. Thus
a

whatever meaning be attached to the forms and , they

must represent quantities whose difference is less than
1 however large n may be, that is, they must represent
n

equal quantities.
The equation a
=
C

is therefore in all cases a necessary and sufficient condition

for the equality of the ratios of a to b and e to d.
a

The fractional form b7 is thus a suitable measure of the

ratio of the two quantities represented by a and b, whether
commensurable or incommensurable.
650. The equality of two ratios is termed proportion, and
the four quantities involved are called proportionals.
The ratio of a to b is often denoted by the symbol a : b .
Thus the equality of the two ratios of a to b and of c to d
may be denoted
a : b = c : d.

In this particular case the symbol = for equality is

often replaced by the equivalent symbol :: ; the propor-
tion may be also written a : b :: c : d.
651. A number of consequences of a simple proportion
are deduced by Euclid from the definition quoted in
Art. 647.
They follow easily from the equation
a C
=
d (1)

For multiplying each of these equal quantities by bd

it follows that ad = bc . (2)
655.] Ratio, Proportion, and Variation. 429

Again, dividing each of these by cd, we get

a b
,
C

or a : c :: b : d ; (3)
that is, a, b, c, d are also proportionals when taken al-
ternately
652. Again, dividing the equals in (2) by ac,
db
一,

C a

or d : c :: b : a ; (4)
that is, a, b, c, d are proportionals when taken inversely
or in the inverse order.
653. Again, from each of the equals in(1) subtract unity.
a C
Then -1= -1 ,
b

or a-bc-d,
d

or a - b : b :: c -d : d. (5)
This is quoted by Euclid under the name ' dividendo.'
654. Again, add unity to each of the equals in ( 1),
a C
then +1= + 1,
b

a + bc +d
or =

d'
or a + b : b :: c + d : d. (6)
This is quoted by Euclid as ' componendo.'
a+ bc+d
655. Since ,

6 d

a-bc-d
and also =

6
430 Ratio, Proportion, and Variation. [656.
dividing the equal quantities in the first equation by those
in the second, it follows that
a+ b c+d
=

a-bc-d
'
or a + b : a - b :: c + d : c-d. (7)
656. Dividing each of the equals in (2) by a, we get
bc
=
d
a

Hence, if four quantities be proportional the fourth is

equal to the product of the second and third divided by the
first. This is the basis of the Single Rule of Three in
Arithmetic.
657. The ratio compounded of two or more given ratios is
defined to be the algebraical product of the fraction which
a C

represent the ratios. Thus, if , represent two ratios,the

ac

ratio compounded ofthese is represented by the fraction .

Euclid's definition leads to this same result. Let a, b be
any two quantities ; then a being any third quantity, a
fourth quantity y can be obtained such that
a : b :: x : y. (1)
Similarly from the quantities c, d, y another quantity z
can be found such that c : d : y : z. (2)
Euclid defines the ratio of a to z to be that compounded
of the ratios of a to band c to d. From (1) it follows that
a x
=

y
C

Similarly a
=
2
.

a C
Therefore =

ac 0
or =
bd 2
659.] Ratio, Proportion, and Variation. 431

658. If the two ratios be equal, the ratio compounded of

them is called the duplicate of either ratio. Thus a² : 62 is
the duplicate of the ratio a : b.
If c be a quantity such that
a: b :: b : c,
inwhich case c is called a third proportional to a and b and
b a mean proportional between a and c,
a b
c

a a a b a2 a
therefore =

ㅎㅎㅎ 이
인 or 62

Hence the ratio of a to c is the duplicate of that of

a to b.
Similarly the ratio compounded of three equal ratios is
called the triplicate of the given ratio.
If the ratio a to y be the duplicate of the ratio a to b ;
the latter ratio, a to b, is called the subduplicate of the ratio
x to y.
659. If any number of ratios a : b, c : d, e : f, ... be equal,
each of them is equal to the ratio of
a + c + e + ... : b + d +f+ ....
Let k denote the value of each of the equal ratios,
a C e
then = k, =
k = k, ...
Hence a = bk, c = dk, e = fk, ... ;
and a + c + e + ... = bk + dk +fk + ...
= k (b + d +f+ ...),
a + c + e + ... a C
or
b + d+f+ =k=z=a= ....
(1)
With the same conditions it follows that p, q, r, ... being
any multipliers, each of the ratios is equal to that of
pa+ qc + re + ... to pb + qd + rf+ ....
432 Ratio, Proportion, and Variation. [659 .
For, as before, a = bk, c = dk, e =fk, ....
Whence pa = pbk, qc = qdk, re = rfk, ... ;
and pa + qc + re + ... = pbk + qdk + rfk + ...
= k (pb + qd + rf+ ...),
a C
pa + qc + re + ... k=
or = === .... (2)
pb + qd + rf+ ...
A still more general result can be deduced.
For, since a = bk, c = dk, e = fk, ... ,
it follows that a" = " k", c" = d" k", e" = frk", ... ,
and pa" = pb" k" , qc" = qd" k", re" rfn k", ... ;
=

whence pa" + qc" + re" + ... = pb" k" + qd" k" + rf" k" + ...
= (pb + qd " + rf* + ...) k" ,
pa" + gc" + re " + ...
whence = k",
pb + qd" + rf + ...
11-2

pa + gc" + re" + ... ۱ a C

or (pa
pb +gd"
+ +rf* +...) =k
ба
....
(3)

660. When one quantity depends on another in such a

manner that if the value of the second be changed, that of
the first is changed in the same proportion, that is, so that
the two values of the first quantity and the two values of
the second form a proportion, the first quantity is said to
vary as the second.
Thus the wages of a man will change with the time
during which he works, so that the wages for two different
numbers of days will have the same ratio as the corre-
sponding numbers of days. The wages are then said to
vary as the number of days he works.
661. If A denote the value of one quantity correspond-
ing to B the value of another, the fact that the first
quantity varies as the second is oftendenoted by the symbol
x, and thus Ax, B.
663.] Ratio, Proportion, and Variation. 433

This appears to express a proportion by means of only

two quantities. The statement really however involves
four, namely, two values of the second, and two correspond-
ing values of the first quantity, and is equivalent to the
proportion
A1 B1= ,
A2 2

where B1 , B2 are any two values of the second quantity,

and A1 , A2 the two corresponding values of the first.
662. Thus if A₁, B₁ be any given pair of corresponding
values of the two quantities and A, B represent any cor-
responding values whatever, if the first quantity A vary
as B, we have
A B
=
,

4B
1
or A= B.
B1
But ¹ is a definite given quantity and does not change
B1
when.B, and consequently A, is changed.
A1
Thus, denoting by a symbol c, the relation
B1
Ax B
may be replaced by
A = CB,

where c does not change with A and B.

Thus if Wdenote the wages of a man for D days, we
know that Wx D, which may be replaced by W = CD,
where c is evidently the wages of a man for one day.
663. One quantity is said to vary as a second inversely
when, if the second is changed, the first is increased in the
same ratio as the second is decreased, and vice versa.
Thus, if A varies inversely as B, andB1, B₂ be two values
Ff
434 Ratio, Proportion, and Variation. [663.
of the second, and A1, A2 the corresponding values of the
first, we have A1:A2:: B2: B1 ;
whence A1B1 = A2B2 , (Art. 651, (2))
and =
A2B2
41 B
B1
If then A2, B2 be any given corresponding values of
A and B, while A1,B₁ are any other variable pair of
corresponding values, denoting A2B₂ by the symbol c, we
have
= .
41
B1
Thus, if A varies inversely asB, we may write, either
c
A= ,
B'
1
or Ax
B

In this latter form the result can be deduced directly

from the proportion
A1 : A2 :: B2 : B1 .
For, dividing the terms of the second ratio by B₁B2 ,
which will not affect the value of the ratio, it gives
1 1
,

41 : 42: BB
1
4x
or (Art. 660)
664. One quantity Ais said to varyjointly as two others,
B and C, when if B and C are changed in any manner, A
is changed proportionally to the product ofB and C.
Thus, if A varies as B and C jointly, and a be the value
of A corresponding to values b and c of the other two
quantities, A: a :: BC :bc.
665.] Ratio, Proportion, and Variation. 435

665. If a quantityA depend on two others, B and C, in

such a manner that when one of these, C, changes while B
is unaltered in value, A varies as C, while when C remains
the same and B changes, A varies as B, then under all
circumstances of change A varies jointly as B and C.
LetA be the value of the first quantity corresponding to
values B and C of the second and third.
Let A₁ be the value of the first quantity corresponding
to values B and c, and finallylet a be the value correspond-
ing to values b and c.
First quantity. Second quantity. Third quantity.
(a) A B C;
(β) 41 B c;
b
(7) a c.

The three lines (a), (3), (y) represent the three sets of
corresponding values.
In (a) and (3) the second quantity has the same value,
consequently, since by hypothesis under these circumstances
Ax C, it follows that
A: A₁ :: C : c,
A C
or =

c
41
Again in (3) and (y), the third quantity has the same
value, hence, since under this condition A & B,
A₁ : a :: B : b,
or
A1B. = ;
a

A C B
whence =

41 a 이언

or
ABC.
a
=

bc

But A, B, C ; a, b, c are any two sets of corresponding

Ff2
436 Ratio, Proportion, and Variation. [666.
values of the three quantities. Hence (Art. 664) A varies
jointly as B and C.
666. A good illustration of the last article is afforded
by the case of wages earned by different gangs of men,
working for different numbers of days. Evidently if the
number of men be the same in two cases, the wages earned
will vary as the number of days they are employed, while
if the number of days be the same fortwo gangs the wages
earned will vary as the number of men employed. Hence,
if W be the wages earned by M men working for D days
and w the wages earned by m men working for d days, it
follows that W : w :: MD : md.

This and the last article contain the theory of what is

called the Double Rule of Three in Arithmetic. It is
evident that if five of the quantities A, B, C, a, b, c be given,
the sixth can be determined from the equation
BC
,

a bc

and will in all cases be obtained by multiplying three of

the given quantities together, and dividing the product so
formed by the product of the remaining two.

EXAMPLES.

1. If a : b :: c : d, prove that
a² + ab : ab -b² :: c² + cd : cd -d².
2. If a : b :: c : d, prove that
a4 64 C4 d
a³+ -- :: c +
a C

3. Find the duplicate of the ratio 4 : 5, and the ratio com-

pounded of 2 : 3, 6 : 7, 14 : 5, and 5 : 8.
4. Find a third proportional to 4 and 6, and a mean pro-
portional between 9 and 16.
 Ratio, Proportion, and Variation. 437

5. Two numbers are in the ratio of 3 : 4, and if they be

increased by 6, they are in the ratio 4: 5. Find the numbers.
6. If a : b :: c: d, then
(a³ + a²b + b³) (a + b) : (c² + c²d + d³) (c + d) : : a² + b² : c* + d*.
7. If a + b : a -b :: b + c : 2 (b −c) :: c + a : 3 (c -a),
prove that 8a + 96 + 5c = 0.
y+ z z+ x x+ y
8. If =

3b- c 3c-a3a-b'
then x + y + z_ax + by + cz
=

a+b +c a²+ b² + c²
2x-y2y- z 2z- x
9. If = = , shew that
2a + b2b + c 2c+a

21 (a + b + c) (x + 2y + 3z) = ( 41a + 386 + 47c) (x + y + z).

2
10. If == , prove that
a b

(x² + y² + z²) (a² + b² + c²) = (ax + by + cz)².

11. If la² +(m -n) bc_mb² + (n - 1)ca_nc² +(1 -m)ab
= =

a b C

prove that
(la² + b² + nc²) (la³ + mb³ + nc³)
= (a² + b² + c²) (l² a³ + m² b³ + n²c³).
12. If
bx+ ay- czcy + bz - ax az + cx -by
a² + b²
=

b² + c²
=

c² + a² , prove that

x+y+ z =
ax + by + cz
a + b + cab + bc + ca
13. There are three numbers, such that the arithmetic means
of the two least and of the two greatest are in the same ratio
as the corresponding mean proportionals. Moreover the sum of
these mean proportionals is double the sum of the two least
numbers, while the difference of the mean proportionals is 12.
Find the numbers.

14. If 2x - Y = 5y - x 2y+ x
, shew that each fraction=y
x+ 3y x - 3y2 (x - 1)
and solve the equations.
438 Ratio, Proportion, and Variation.
x 2
y
15. If
y+ z+ u- 6 z+ u + x- 1 u+x+ y+ 9
5

x + y + z + 14_x + y+ z + u + 4 '
shew that each of these fractions = 1 or - 5 ; and find the values
of x, y, z, u.
16. If x ( y + z - x) _y(z + x − y) _ z (x+ y-z)
= = ,

logx logy logz

prove that yz" = 2x = x²y".
17. Taxes are laid upon alodging-house in proportion to the
rent of the rooms ; they were one-seventh of the rent and are
raised to one-sixth. In what proportion must the rent be
raised to meet this additional charge ?
18. If n arithmetic means A1, A2, ... An be taken between
a and b, n geometric means G1, G2, ...Gn; n harmonic means
Η , Η , Η , ...H.; prove that
A1.A2.A ...An: G1G2 G3 Gn: : G₁G₂G3... Gn: H.H2.Η ... Η .
3 3

19. Ifx + y + zxx + y- z and x² + y² + z² x x² + y² -z², shew

that each of x and y varies as z.
20. A is walking along a road and passes B, when finding he
haslostsomething, he turns back andmeets Bt hours afterwards;
having found what he had lost he overtakes B again t' hours
after he met him, and arrives Thours too late at his destination.
Compare their rates of walking.
21. Given that the volume of a sphere varies as the cube
of its diameter and that a sphere of 14 feet diameter has a
volume of 1433-74 cubic feet : find the number of cubic feet in
a sphere 2 feet in diameter.
22. If 8 men working 9 hrs. a daydig out 400 cubic yards in
5 days, how much can 15 men working 74 hours a day dig out
in 11 days ?
23. If the sum of two quantities vary jointly as the difference
of their cubes and a third quantity ; and the difference vary
jointly as the sum of the cubes and the third quantity: shew
that the one of the two quantities varies as the other.
 Ratio, Proportion, and Variation . 439

24. Supposing that the velocity of a steamer varies inversely

as the area of its greatest section when the tonnage is constant,
and inversely as the tonnage when the area is constant, and that
a steamer whose section is 200 sq. ft. and tonnage 1000 goes 15
miles per hour; find the velocity of a steamer whose section is
250 square feet and tonnage 1200.
25. Divide 111 into three parts such that the products of each
pairmay be in the ratios 4: 5: 6.
26. Two coins of the same bulk, whose values are as 25 : 4 and
whose weights are as 9 : 8, are each composed of silver and
copper. Bulk for bulk silver is as heavy again as copper :
weight for weight silver is 42 times as valuable as copper. Find
the proportions of silver and copper in each coin.
27. A certain number is added to each term of the ratio 3 : 10,
and the same number is also subtracted from each term, and it
is found that the resulting ratio in one case is the duplicate of
the resulting ratio in the other. What is the number ?
28. There are two vessels each containing mixtures of two
liquids in the ratios of r : 1 and s : 1 respectively. A mixture
consisting of certain quantities from the two vessels is com-
posed of the two liquids in the ratio : 1, and one consisting of
the same quantities each taken out of the other vessel is composed
of the two liquids in the ratio y : 1 : shew that
(x+ 1) (y+ 1) _xy-1
=

(r + 1) (s + 1) rs- 1

29. The sum of fourquantities in proportion is 55, the sum of

the extremes exceeds that of the means by of the first term,
and the sum of the squares of the extremes exceeds the sum of
the squares of the means by 165 of the square of the first term :
64
find the four terms.
 CHAPTER XXVI .

CONTINUED FRACTIONS .

667. THE term ' continued fraction ' will be employed to

denote such forms as
1 1
1+ , a+
1 1
2+ 6+
1 1
3+ c+
4 1
d+
e

where all the numerators of the partial fractions are unity

and all the letters a, b, c, ... denote positive integers.
For economy of space the continued fraction is usually
written in the form 1 111
a+
b + c + d + e

668. Any positive rational fraction in its lowest terms

can be expressed as a continued fraction.
N
For let the fraction be where N and D are positive
D'
integers having no common factor,
D ) N ( 91
91D
1 ) D ( 92
9271

2)11( 93
932

73...
669.] Continued Fractions. 441

Let the process of finding the Greatest Common Measure

(Arts. 179, 180) of the two numbers N and D be carried
on, and let the successive quotients be 91, 92, 93, ... and
71, 72 , ... the corresponding remainders. The last of these
must obviously be unity, since Nand D have no common
measure.

Then, by the nature ofdivision (Art. 157), the following

identities hold
N = 91D + 1 ,
D = 921 + 12 ,
1 = 932 + 3 ,

N 1
Whence = + = +

D 1
= 92 + 2-=
= 92+ ,
1 1 1

T2
1
71 73
= 93 + = 93+
T2 T2 72
73

Whence, substituting in succession,

N 1 1

D
91+
92+ 93 + ...
355
669. Let the fraction be ,

113

113 ) 355 ( 3
339

16) 113 ( 7
112

1) 16 (16
16

...
442 Continued Fractions. [669.
355 1 1
Hence = 3+
113 7 + 16
323
As another example take 117'
117 ) 323 ( 2
234

89 ) 117 ( 1
89

28 ) 89 ( 3
84

5 ) 28 ( 5
25

3) 5 ( 1
3

2) 3 ( 1
2

1) 2 ( 2
2

...

323 1 1 1 1 11
Hence = 2+
117 1 + 3+ 5+ 1+ 1+ 2
N
670. If a fraction be reduced to the form
D
1 1
91+
12+ 13+ ...
it can be shewn that the quantities obtained by taking
one, two, three, of the quotients will be alternately less
...

N
and greater than the value of D
N
For q₁ is evidently less than D since the fractional part
is omitted.
671.] Continued Fractions. 443

1 N
On the other hand, 91 + is greater than D since the
92

latter exceeds q₁ by a fraction whose denominator is greater

1 1 N
than q2 . Similarly 91 + is less than
92+ 13 D
The fractions obtained by taking one, two, three, of ...

the quotients are called convergents to the value of the

continued fraction.
671. The successive quotients being 91 , 92, 93 , ... the
corresponding convergents are
1 1
91, 91+ , 21+ 1
, ... ;
92
92+
13

or, reducing to the form of simple fractions,

911 , 9291 + 1 , 939291 + 91 + 93
92 9392 + 1
If these successive convergents be denoted by
N3
N1 , N2 ,
‫ و‬...

1 2 3

it is evident that

N3 = 93 (9291 + 1) + q1 = Q3N2 + n₁ ,
d3 = 93 92 +1 = 93d2 + d₁ .
This law of formation will be shewn to hold universally,
so that if p be any integer and qp , np , dp denote the pth
quotient and the numerator and denominator of the
corresponding convergent,
n
NpP = QpNp- 1 + Np -21
(1)
dp =
Ipdp-1 + dp-2
For assume that this relation holds for any value ofp, so
that the pth convergent "p = IpNp-1 + Np-2 .
dp qpdp-1 + dp-2
444 Continued Fractions. [67г.
Since the p + 1th convergent only differs from the pth by
1
having q,+ for qp , it follows that this convergent, or
Ip+1
1

Np + 1
(2p + Ip +1) Пр-1 + Np -2
1
dp+1
(2p +Ip+1)dp-1 +dp-2
=
Ip + 1 (Jpnp -1 + Np-2) + Пр-1
Ip+1(9pdp-1 + dp-2) + dp-1
Ip +1Np + Np-1 .
=

Ip+1dp + dp-1
Hence we may assume
Np + 1 = Ip + 1Np + Np-1 ,
dp +1 = qp +1dp + dp-19
that is, if the assumed law of formation hold for the pth
convergent, it will hold for the p + 1th.
But the law does hold for the 3rd convergent, and con-
sequently it holds for the 4th and so on, for all the con-
vergents.
672. Multiplying the two equations (1) in the last
article by dp -1 and np-1 respectively, and subtracting the
second from the first, we obtain
npdp-1 - np- 1dp = np-2dp-1 - np-10-2
= ( -1) (np -1-2- Np-2dp-1).
Similarly,
Np-1dp-2 - Np- 2dp-1 = ( -1) (np -2Ap-3 - np-3dp-2)
=

n3d2 - n2d3 = ( -1) (nad₁ - n₁₂).

There are altogether p - 2 of these equations. Hence by
continued substitution,
npdp-1 - np- 1dp = ( -1)p-2 (n₂d₁ - n₁₂).
2
675.] Continued Fractions. 445

Referring to the value of n₁ , d₁ , n₂, d₂, we see that

n2d1 - n1d2 = (9192 + 1 ) -9192 = 1 .
Also ( -1 ) -2 = ( -1 ) (Arts. 126, 128).
Hence, finally,
npdp-1 - np-1dp = ( -1 ) .
673. From the result of the last article it follows that
any convergent, as n , must be in its lowest terms. For if
d P

n, and d, have any common measure it must (Art. 181) be

ameasure of npdp -1 - np- 1dp or unity. Hence they have no
n
common measure or is in its lowest terms.
dP

674. The difference between any two consecutive con-

vergents, can be expressed in a similar form by means of
the relation in the last article.
n.
Thus,
np
P
Np -1_npdp-1 - Np-1dp _ (- 1) .
=
=

dp dp-1 dpdp-1 dpdp-1

Np Np-1
From this we see that if p be even P-1, while ifp be
ddp-1
odd the reverse is the case.
Hence the convergents are alternately greater and less
than the previous one.
n
675. The convergent is produced from the preceeding
dp
one by the introduction of a new quotient qp . The whole
N
fraction D
will be produced from "p-1 by introducing
dp-1
1 N
9p+ instead of qp . Thus can be deduced from
Ip+1 + ... D

" , if for q, we substitute q, +f

-1 by the same rule as dp
dp-1
1 1

whenfis a proper fraction denoting 9p +1 + Ip + 2 + ...

446 Continued Fractions. [675-
N
Hence =
(qp +f) np-1 + Np-2
D (qp +f) dp-1 + dp-2
= np +fnp-1 .
dp +fdp -1
N
Hence -
np_ (np + fnp -1) d-(dp +fdp-1) np
=
Dd (dp +fdp-1)dp
_f(np-1dp - np7p-1)
dp(dp +fdp-1)
= - (-1) f
(1)
dp(dp +fdp-1)
N np np
Hence if p be odd is positive, and dp is less
D
-

d.
P

N
than , while ifp be even, the reverse is the case. Thus,
as was seen before, the first, third, fifth convergents are ...

less than the original fraction, while the second, fourth ...
are greater.
Also since as pincreases, d, obviously increases, the
N n
difference between and diminishes. That is, each
D dp
convergent is nearer in value to the given fraction than
the previous one. N
It is evident from (1) that the difference between D and
n 1
is numerically less than d2
dp p

676. A slightly closer limit can be assigned to the error

N
in taking for D'
dp
1 1 1
for f= < or
1
Ip+1 F > Qp+1°
Ip +1 +
Ip+2 + ...
N (-1) f (-1)
Hence Np = =

D 1
dp d (dp+fdp-1)
678.] Continued Fractions. 447

N n 1 1
Hence ~ < <
D
1
dp
P
dp (qp +14 + dp-1)
1
dp.dp+1
Again, which = qp + 1 + < 1p +1 + 1 .
Ip+2 + ...
N ~ 1
Hence
D No > do {dp-1 + (Ip+1 + 1) dp}
dp
1
>
dp (dp+1+dp)
N
Thus the error involved in taking " instead of D lies
1 1 dp
between and
dpdp+1 dp(dp +1 +dp)
677. The successive convergents to a given fraction are
thus fractions, expressed by smaller numbers than the
original one, whose values are continually closer and closer
approximations to that of the original one.
For instance, the successive convergents to the fraction
355 22
in Art. 669 are 3 and , the value of the latter
113 7

exceeds very slightly that of the given fraction.

678. The method of continued fractions thus gives a
means of approximating to the values of any commen-
surable fractions. It may also be used to approximate to
the value of a quadratic surd.
Thus 18 can be expressed as a continued fraction in
the following manner,
2 1
√18 = 4+ ( √18-4) = 4 + = 4+ ,

√18 + 4 √18+4
2

√18 + 4 √18-4 1
= 4+ = 4+ ,

2 2
√18+ 4
2
√18 + 4 = 8+ √18-4 = 8 +
√18+ 4
and the process will repeat itself indefinitely.
448 Continued Fractions. [679.
111 1
Thus √18 = 4 +
4 + 8 + 4 + 8 + ... ad inf.
The quadratic surd is thus expressed in the form of an
infinitely repeating continued fraction. It is proved in
treatises on the subject that if a be the integral part of
✓N, the continued fraction will begin to repeat when a
quotient 2a is obtained. The proof is long and somewhat
tedious. Any particular example can be worked out like
the above.
679. As an example of the use of this method of ap-
proximation we may take √2,
1
√2 = 1 + √2−1 = 1 + ,

√2 + 1
1
√2 + 1 = 2 + √2-1 = 2 + ,

+1
√2

1 1
Hence √2= 1 + 1.
2 + 2 + 2 + ...

The convergents, obtained by the rule of Art. 671 , are

3 7 17 41 99
1, ,
.... ;
2'5' 12 29 70

the last of these expressed as a decimal fraction gives

1.41428 ... , which agrees to four places of decimals with
the value found for √2 in Art. 399.
It is greater than the real value, as we should expect
from Art. 675 .
239
The next convergent is 169 Hence the error in taking
99 1

70
for √2 is (Art. 676) less than 70 x 169
and greater
1 1
than or or the error lies between
70 x (70 + 169) 70 x 239
1 1
and or between 000084 ... and .000059 ....
11830 16730
681.] Continued Fractions. 449

680. Any fractional form involving a single quadratic

surd can be reduced to a repeating continued fraction.
3+√5
For instance, let the fraction be 4
,

√5 is slightly

greater than 2, hence the integral part of of3 + √5 is unity.

4

Thus,
3 + √5 √5-1 5-1 1
= 1+ = 1+
4 4 5+1'
4 ( √5 + 1 ) = 1 + √5+1
5-4 1
√5 + 1 = 3 + √5-2 = 3+ = 3+ ,

√5+ 2 √5+2
1
√5 + 2 = 4 + √5-2 = 4+ ,

√5 + 2
and the process will now repeat itself indefinitely.
Hence, finally,
3 +√5 1 1 1
= 1+ ad inf.
4 3 + 4 + 4 + ...

681. Conversely, any infinitely repeating continued

fraction may be expressed in terms of a quadratic surd.
For let the fraction be
111111 1
a+
b + c + p + q + r + s + p + ... '
where the quotients from p to s repeat indefinitely.
Let the value of the whole fraction be denoted by æ, and
1 1 1 1
let the infinite repeating fraction p +
q + r + 8+ p + ...
be denoted by y.
Hence 1 11
x = a+
b+ c+ y
1 111
y =p+
q + 7+ 8+ Y
Gg
450 Continued Fractions. [682 .
Reducing these continued fractions by the method of
Art. 671, we obtain results of the form
x =
A +By
A' +B'y'
P + Qy ;.
Y = P'+ Q'y'
where A, B, A', B' depend only on a, b, c ; and P, Q
P
', Q' contain only p, q, r, s.
The second equation is a quadratic equation in y, and
consequently gives y in terms of a quadratic surd. The
first equation then gives a in terms of the same.
682. Thus, taking the repeating fraction
1 1 1 1
1+
3 + 4 + 4 + 4 + ... '
1 1 1
x = 1+
3 + 4 + 4 + ... '
1 1
y = 4+
4 + 4 + ...
1 4 + 1
Then y= 4+ - = , (1)
y y

11 4 +1
x = 1+ =
(2)
3+ y 3y+ 1
From (1), y²- 4y = 1 or y = 2 + √5.
Evidently the value of y must be positive and a little
greater than 4, hence the sign + must be taken, and
y = 2 + √5.
4 +1 9 + 4√5
It follows that x = =
;
3y + 1 7+ 3√5
and rationalising the denominator by multiplying numer-
ator and denominator by 7-35, the fraction becomes
3 + √5
4
 Continued Fractions. 451

EXAMPLES.

Reduce each of the following fractions to the form of a con-

tinued fraction and find the successive convergents :- 314159 ?
13 7 355 31459
1. 2. 3. 4.
23 17 113 10000

2723 617 148

5. 6. 7.
1799 1839 263

1 1 1 11
8. Find the convergents to 1 + 2 + 3 + 4 + 5+ 6

9. Two bells begin to ring together ; the one rings 12 times

in 7 minutes, the other 17 times in 9 minutes. What strokes
most nearly coincide in the first half-hour ?
10. Express 27-321661 days, the average period of the moon's
revolution, by the nearest equivalent fraction whose denominator
is less than 1000.

11. If the arithmetic means of each two consecutive con-

vergents be formed, prove that they will be alternately greater
and less than the continued fraction.
NN' N"
12. If D' D' D' be three consecutive odd (or even) con-
vergents to a continued fraction, and p be the product of the
quotients corresponding to the two convergents preceding the
N" -N' (p + 1 ) N' -N
last of these, then
D
'' −D
' ¯ (p + 1) D'−D
13. If the quotients 92 , 93 , 97+1 corresponding to the con-
vergentsN2,N3 Nr+1 , be all equal, then
DD 3
.....

Dr+1
NN - NN
+1
N2
N₁² + N2
N₂² + .. + N2
=
2

Dr+ D - DD_ D₁² + D²₂ + ... +D,

Gg2
452 Continued Fractions.

14. If P" be the nth convergent to the continued fraction

In
a1 A2 A3
,

b + b2 + b + ...
prove that p = bnPn-1 + anPn-2 , qn = bnqn-1 + anqn-2
15. Convert √11 into a continued fraction.
16. Express 18 as a continued fraction, and find the limits
of the error made by taking the fourth convergent.
17. Express 19 in the form of a continued fraction.
18. Find the value of the fraction
1 1 1 1 1
1+ ad inf.
3 + 2 + 1 + 3 + 2 + ...
19. Find the value of the fraction
1 1 1 1 1 1 1
ad inf.
1 + 2 + 3 + 1 + 2 + 3 + 1 + ...
20. Find the value of the fraction
1 1 1 1 1
a+ ad inf.
1 + b + a + 1+ b + ...
21. Shew that the value of the fraction in the last question
divided by the fraction
111 1 a+ 1
b+ ad inf. is
1 + a + b + 1 + ... 6+ 1
22. Convert the positive root of the equation 3x² + 8x- 7 = 0
into a periodic continued fraction, and find the first six con-
vergents.
23. Find the sixth convergent to each root of the equation
5x² - 20x + 14 = 0.
1 1
24. If x = y+ ad inf. ,
2y+ 2y+ ...
1 1
prove that y=x ad inf.
2x- 2x...
N 1 1 1 1
25. If be the rth convergent to a + ,shew
D, b + a + b + ...
that D2n= N2n+1 and Nan: Dan -1 :: b : a.
 Continued Fractions. 453

1 1
26 Prove that √a² + 1 =a + ad inf.
2a+ 2a+ ...

27. If In be the nth converging fraction to √a² + 1, prove that

Pn (a + √a² + 1) + (a- √a² + 1)"
= √a²+ 1
In (a + √a² + 1) -(a - a² + 1)"
28. If √x + a be expressed as a continued fraction in the
a a
form x+
2x+ 2x+ ...
and P"
In
be the nth convergent ; provethat
Pn²-(x² + a)qn² = ( -a)"
PnPn+1-(x² + a) qnqn+1 = x ( -a)".
 CHAPTER XXVII .

INDETERMINATE EQUATIONS .

683. THE definite determination of the values of two

unknown quantities requires, as we have seen, two equa-
tions. If the two equations be of the first degree, inde-
pendent and not inconsistent, there will be only one pair
of resulting values.
There are arithmetical problems the solution of which
may be made to depend on that of a single equation
between two unknowns, the second equation being re-
placed by some condition of a general nature which the
unknown quantities must fulfil.
The most common of these conditions is that the values
of the unknowns must be positive integers, zero being
sometimes admissible and sometimes not.
684. Let x and y be subject only to the conditions that
they are positive integers and that they satisfy the equa-
tion
5x + 7y= 24. (1)
Dividing by 5, we obtain
4
x+y+ 2
5
= 4+ ,

or x + y- 4 =
4-2y_2 (2 -y) .
=

5 5

Hence 2 (2-3) must be integral and therefore 2-ymust

5 5

be integral. If the integral value of this be denoted by t,

2-y = 5t,
or y= 2-5t. (a)
 Indeterminate Equations. 455

Substituting in the original equation,

5x + 7 ( 2-5t) = 24 ,
or 5x = 10 + 35t ;
whence = 2+ 7t. (β)
The only integral value of t which will allow both x and y
to be positive integers is zero.
Hence x = 2, y = 2 are the only values of a and y
which satisfy the given conditions, and the problem is
in this case perfectly determinate.
685. The general form of the problem is the discovery
of values of x and y which satisfy an equation of one of
the forms
ax +by = c, (1)
ax-by = c, (2)
and which are at the same time restricted to be positive
integers.
We may premise that the equation is supposed to have
been previously freed from fractions, so that a, b and c
are all integers. There will obviously be no limitation in
supposing that a and b are positive integers.
The solution will obviously be impossible if a and b
have any common factor which is not a factor of c. For
any measure of a and b must measure ax + by if x and y
be integers (Art. 181).
Since any factor common to a, b and c may be removed,
we may assume that a and b are positive integers which
have no common factor. In the case of form (1) no solu-
tion is possible unless c be also a positive integer.
686. A very useful practical way of solving such equa-
tions is exemplified in Art. 684.
Suppose that a is the smaller of the two numbers a and
b, and let the whole equation be divided by a ; we thus
get b C

x+ a
y= a
- (a)
456 Indeterminate Equations. [686.
Now each of the numbers 6 and e when divided by a
must give a quotient and a remainder, if these be q, q',
r, r', respectively, (a) becomes

* + (2 + 1) = 1 + a

r - ry
or x +qY-q= a

Hence
r'-ry
- ry must be an integer, not necessarily posi-
a

tive. Let this be called z. We then have

r = ry + az. (β)
In this equation y and z are integers, and ris obviously
less than a.
Divide by r, and by a similar process a new equation
of the form
ru + 2
(γ)
is obtained, where u and z must be integers.
Now from the process of formation the coefficients of
the successive equations continually diminish. Hence at
last the coefficient of one of the unknowns will become
unity. Suppose that in this last equation r"= 1 .
Hence -ru .
2=

And therefore from (β),

ry = r - az = r' - ar" + aru. (8)
But by the law of formation r" or unity is the re-
mainder after dividing a by r. Hence a - 1 is divisible
by r. Also r -r" is for a similar reason divisible by r.
Thus r - r"-(a - 1)r", or r - ar", is also divisible by r,
and (8) gives
r - ar"
y = β + au ( 1) where β r
688.] Indeterminate Equations. 457

Substituting in (a), we get

=
c-b3-bau
a

c-bß
= -
bu
a

r- rß
= q -qß + -bu.
a

r' -ar"
But β = •

Hence r' - rẞ = ar", and finally,

r

x = q - qß + r" -bu,
= a-bu. (2)
The value of u is only restricted to be integral. Hence
the solutions of the original equation will be obtained by
giving to u all integral values including zero, which make
both a-bu and ẞ + au positive integers.
687. The solution of the equation
ax-by= c
can be conducted on the same plan as that of the last
article.

888. Sometimes one solution can be discovered by in-

spection. When that is the case the general solution can
be deduced.
Thus, let x = a, y = ẞ be one solution of the equation
ax + by = c.
It follows that
aa + b = c,
and if x = x1 , y = y₁ be any other solution whatever,
wehave ax₁ +by₁ = c.
Subtracting the former from the latter,
a (x1-a) + b ( 1 - β) = 0,
or
a (x1 -a) = 6 (β -31) ;
a
В-У1
whence
= x -a
458 Indeterminate Equations. [689.
a

Now, by supposition, is a fraction in its lowest terms.

Hence, since ẞ-y₁ and a₁ - a are both integers, it follows,
as will be strictly proved hereafter (Art. 727), that B -41
and x₁- a must be equimultiples of a and b.
Thus we must have, if u be some integer,
β-y1 = au,
x₁ - a = bu,
or
x1 = a + bu,
Y₁ = β - au ;
and the whole assemblage of solutions will be obtained by
giving u all integral values consistent with making a₁
and y₁ positive integers.
689. In a similar way, if x = a, y = ẞ be one solution
of the equation
ax-by = c,
the general solution will be
x = a + bu,
y= β + au.
It is obvious that while in the equation in the last
article, u can only have some integral value lying between
-and + in the present case, u may have any positive
a

integral value whatever, as well as possibly some negative

ones, and therefore the number of solutions of the equation
ax- by = c is unlimited.
690. The discovery of one solution can always be
effected by means of one of the theorems of the last
chapter.
b
Thus let a
be converted into a continued fraction, and
let d be the convergent next preceding b.
n

a
691.] Indeterminate Equations. 459

Then it has been shewn (Art. 672) that

an - bd = +1 ,
b
the + or - sign being taken according as a
is even or
odd in the order of convergents.
Multiplying by c,
anc - bcd = c.

If the sign be +, this can be written

a (nc - bu) + b (au - cd) = c ;
and evidently x = nc -bu,
y = au - cd,
give a solution if u be so taken that nc - bu and au - cd
are both positive integers.
These results in fact give the general solution, as may
be seen by comparison with Art. 688.
If, on the other hand, the sign be -, we shall have
a (bu - nc) + b (cd -au) = c :
whence a solution is given by
x = bu-nc,
y = cd - au,
u again being taken so that these shall both be positive
integers.
691. The solution of the equation
ax-by = c
can be conducted in the same way and will lead to the
results, x = nc-bu,
y= cd-au ;
or
x = bu- nc,
y = au - cd ;
b
according to the order in which comes in the con-
a

vergents to itself expressed as a continued fraction.

460 [692.
Indeterminate Equations.
692. The student will notice that, in the method of
Art. 686, the process is really gone through of finding
the greatest common measure of a and b leading at last
to a remainder unity. The same process has to be gone
b
through in expressing as a continued fraction. The two
a

methods, though apparently unconnected, do thus depend

essentially on the same fundamental process. The method
of Art. 686 generally gives the solution in a more con-
venient form than that of Arts. 690, 691 .
693. It has been seen (Art. 689) that the number of
solutions of the equation
ax -by = c
is unlimited.
It has been also shewn that if a, ß be any values of
x and y which satisfy the equation ax + by = c the general
solution is given by x = a + bu,
y = β - au.
Hence u may have any integral positive value such
that au is not greater than ẞ, and any integral negative
value such that bu is not numerically greater than a ;
u may also have the value zero.
Let then β= m +f, where f is a proper fraction, and
a
α

m an integer; and similarly let 7 = n +f

', where n is an
integer and f a proper fraction. The number of values
admissible for u, or the number of distinct solutions will
be m + n + 1.
β a
Now
a + 7 = m+ n +f+f'.
Hence m+n+ 1 = + -f-f + 1
bβ + aa
=

ab
-f-f + 1 .
But x = a, y = β is one solution of the original equation.
 461
696.] Indeterminate Equations.

Hence aa + b = c, and therefore

C
m+n+ 1 = +
ab

If therefore f+f exceeds unity, the number of solutions

is the integral part of , but iff+f is less than unity,
C
the number of solutions is the integral part of ab +1.
694. If either of the quantities for f' is zero, asf; it
β
follows that
a
is integral, and when u = a
y = 0. If the
nature of the question render a zero value for y inad-
missible, this solution is excluded, and the number of
C
admissible solutions is the integral part of ab
a

Iff' also vanishes the value u = makes x = 0, and

if this be excluded by the nature of the question the

C

number of solutions will be ab

- 1, abeing in this case
easily seen to be an integer.
695. Problems sometimes occur requiring the discovery
of all the positive integral values of three quantities x, y, z
which satisfy a single equation of the form
ax +by + cz = d.
The method of solving such a problem depends on the
methods already elucidated, and will be best illustrated
by one or two examples.
696. Let the equation be
2x + 3y + 42 = 36 .
It is obvious that z, whose coefficient is the largest, has
the least range of possible values. The greatest possible
value of z is evidently 9, and if z have this value æ and y
must be zero.
462 Indeterminate Equations. [696.
This one solution is
x = 0, y = 0, z = 9.
Let z have given in succession the values, 8, 7, 6, ... 2, 1 ;
thus we get the several pairs of equations
z = 8, 2x + 3y = 4 ;
z = 7, 2x + 3y = 8 ;
z = 6, 2x + 3y = 12 ;
z = 5, 2x + 3y = 16 ;
z = 4, 2x + 3y = 20 ;
z = 3, 2x + 3y = 24 ;
z = 2, 2x + 3y = 28 ;
z = 1, 2x + 3y = 32 ;
z = 0, 2x + 3y = 36 .
Then each of the equations with two unknowns can be
solved by previous methods, and the whole series of values
ascertained.
As another example take
2x + 3y- 4z = 4.
Here z may be zero or any positive integer. Con-
sequently there is an infinite number of sets of values
of x, y, z, derived from the series of equations
z = 0, 2x + 3y = 4 ;
z = 1 , 2x + 3y = 8 ;

z = n, 2x + 3y = 4 (n + 1) ;
where n is any positive integer.
In this case the complete solution can easily be shewn
to be contained in the formula
x = 2 (n + 1 ) -3t, y = 2t, z = n,
n and t being both positive integers, and the only restric-
tion being that 2 (n + 1) must be greater than 3t.
697.] 463
Indeterminate Equations.
Problems of this kind have been met with in Arts.
551, 552.
697. Indeterminate equations of higher degree than the
first present too many difficulties for an elementary treatise.
One class of such admits however of easy treatment.
Let the equation be
axy + bx + cy + d = 0 .
Multiplying by a and transposing, this becomes
a² xy+ abx + acy = -ad,
or, by adding be to both sides,
(ax +c) (ay + b) = bc - ad.
The number be- ad can ordinarily be resolved into two
factors in several ways. If a, ẞ be one pair of these we
mayhave ax + c = a, ay + b = β,
and if these give positive integral values of x and y, there
will be one solution. By taking all the possible pairs of
factors of be- ad, all possible solutions can be obtained.
Thus the equation xy- 2x-y + 14 = 0 can be written
(x - 1) ( 2-3) = 12 .
The pairs of factors of 12 are 1, 12 ; 2,6 ; 3,4 ; 4,3 ;
6,2 ; 12, 1 ; of these only
x - 1 = 6, 2 -y = 2 ;
x- 1 = 12 , 2 -y = 1
are admissible ; thus x = 7, y = 0 ; x = 13 ; y= 1 are the
only solutions.

EXAMPLES.

1. Solve in positive integers the equation 7x + 5y = 31 .

2. Find the general integral solutions of 21x- 13y = 95.
3. Find the general solutions of 8x + 7y = 56.
4. Find all the integral solutions of 5x + 7y = 62 .
464 Indeterminate Equations.
5. Solve the equation 13x+ 17y = 116.
6. Find the number of positive integral solutions of
5x+ 3y = 74.
7. What is the least value of m that 5x + 3y = m may have
six solutions in positive integers ?
8. Find the number of solutions in positive integers of
11x + 15y = 1031 .
9. Find the general solution of 10x + 11y = 1001 .
10. Find the simplest way in which a person who has only
guineas canpay 10s. 6d. to a person who has only half-crowns.
11. A has 15 florins and 8 half-crowns, and B has five
shillings and 4 half-crowns, and A owes 12s. 6d. to B. In how
many ways can A pay B ?
12. The mint price of standard gold is 31. 17s. 101d. per
ounce; find the least integral number of ounces that can be
coined into an integral number of sovereigns and guineas; and
how many of each sort may be taken.
13. Aman was born in a certain century and died in the next,
the last two digits in the years of his birth and death being the
same . His age when he died consisted of two digits, the first
of which equals the second digit in the year of his birth, and the
other equals the first digit in the sum of the years of his birth
and death, which ends with three digits the same. When was
he born ?

14. A certain number consisting of two digits is multiplied,

and thus becomes greater by one than the number formed by
inverting the digits : what is the number and the multiplier ?
15. A and B begin to read simultaneously two works each
consisting of a great number of volumes. A's volumes contain
89 pages each, and B's 100, and A reads 7 pages while B reads
11. Will they ever finish the same page at the same instant,
and if so, when ?
16. In how many different ways can a volunteer shooting at
a target score 16 in 9 shots, a bull's-eye counting 3, a centre 2,
an outer 1, and a miss 0 ?
 Indeterminate Equations. 465

17. There is a number of two digits, which if its digits be re-

versed, becomes less by unity than its half. Find the number.
18. Find three numbers such that twice their sum is a
number whose digits are the greatest and least of the numbers,
and three times their sum, one whose digits are the least two of
the numbers. Shew that there are two such sets of numbers
and that in each case six times their sum is a number whose
digits are the greatest two of the numbers.
19. Find the greatest integer which can be formed in 9
different ways and no more by adding together a positive integral
multiple of 5 and a positive integral multiple of 7.
20. Find the general form of numbers which when divided by
7 and 9 leave remainders 5 and 7 respectively.
21. Find a number of two digits such that the product of the
digits is less by 100 than twice the number.

Hh
 CHAPTER XXVIII.

ON INEQUALITIES AND ARITHMETICAL MAXIMA

AND MINIMA .

698. In the chapter on Permutations and Combinations,

Arts. 495-497, an instance has been given of the deter-
mination of the greatest value of an algebraical expression,
nCr, where r was restricted to have only integral values.
Another instance is given in Art. 558.
There are a few general propositions which enable the
maxima and minima values of some algebraical expressions
to be determined, without any restriction as to the integral
character of the values of the letters involved.
699. An algebraical quantity a is said to be greater
than another 6 when the expression a-b is a positive
quantity.
Thus a > b,
and a- b is positive,
are interchangeable and equivalent assertions.
700. The product (a − b) (a-b) is always positive
whether a-b be positive or negative (Art. 59), con-
sequently a²- 2ab + b² is positive,
or a² + b² > 2 ab,
a² + b²
or > ab.
2

This is true whatever scalar quantities a and b may

represent. Hence for a we may substitute √ , and for b
x+y
we may write √y. Thus we get 2
> √xy. Or (Arts.
Inequalities andArithmeticalMaxima andMinima. 467

511 , 524) the arithmetic mean of two quantities, æ and y,

is greater than their geometric mean.
The only exception is when a- b is zero, or when a is
equal to b, and consequently a is equal to y. In this case
the arithmetic mean is equal to the geometric mean.
701. It is not infrequently required to find the least
value of an expression which can be put into the form
A
x+ ,

where a is the only variable part.

By the last article,
A
x+
X A
> XX ,
2 2

> √A
,
A
or x+
X
> 2√4,
A A
except when x = , in which case x +
x
becomes equal
to 24.
A A
Hence x + has its least value when x = or x² = A,
X X

or x = √A, and this least value is 2√A.

702. The result of the last article can be obtained in
another way.
A
Let x+ = y,
x

whence x²-xy + A = 0.
Solving as a quadratic in æ,
y ± √y²- 4A
X
2

Hence if a be real or scalar, y² - 4 A must be positive, or

at any rate cannot be negative, thus y² must not be less
Hh2
468 On Inequalities and [703-
than 4A, or y cannot be less than 2√A. The least value
A
of x+ is thus 2A, and when this is the value of y,
X

2√A
that of a is
2
, or √A.

703. The theorem ofArt. 700 can sometimes be applied

conversely to determine the greatest value of a given
algebraical expression which can be resolved into two
factors.
Let the expression be 9x - 2x² - 9. By the processes of
Chapter XIV this can be resolved into (3 -x) (2x- 3).
Now (3 - x) (2x - 3) = ( 6-2x) (2x - 3).
And by Art. 700,
(6-2x) + (2x - 3 ) < 3>
√(6-2x) ( 2x - 3) < 2

unless 6-2x = 2x- 3 .

Hence the greatest value of (6-2x) ( 2x - 3) is given
when 6-2x = 2x- 3 or when a = 1, and this greatest
value is 1.
Hence the greatest value of 9x- 2x² - 9 is , and is
obtained by given to a the value .
704. The general principle underlying the result of the
last article can be stated that the product of two expressions
whose sum is constant is greatest when the expressions are
equal.
The corresponding principle for Art. 701 is that the sum
of two quantities whose product is constant is least when the
quantities are equal.
705. If there be n quantities a1 , a2 , a3, ... an equal or
unequal, the value of a1 + a2 + ... + an" is called their Arith-
n
1

metic mean, and the value of (a1a2a3 ... an) is called their
Geometric mean.
 706.] Arithmetical Maxima and Minima. 469

These terms are derived by analogy from the case of two

quantities.
It can be proved as an extension of Art. 700, that the
arithmetic mean of n quantities, which are not all equal,
is greater than their geometric mean.
For it is evident that when the quantities are all equal
their arithmetic mean is equal to their geometric mean.
If any two, as a, and a,, are unequal, let them be replaced
by two equal quantities each equal to half their sum or
(a +a ). This will not affect the value of the arithmetic
mean of the n quantities, but, by Art. 704, will increase
the value of the geometric mean. Thus by repeated
substitutions of this kind, as long as any two remain
unequal, the arithmetic mean is unaltered while the
geometric mean is increased. In the final result the two
means are equal : hence previously the geometric mean
must be the less of the two.
706. This principle can occasionally be used to determine
the greatest value of an algebraic expression which is
resolvable into factors .
For instance, let it be required to determine the values
of x and y which give its greatest value to the expression
(c-x) (c-y) (x+y-c). (a)
By the theorem of the last article,
C

{(c-x) (c −y) (x + y- c) }$ < (c-x) + (c- y) + (x + y -c) < 3',

3

unless the three quantities c-x, c -y and x +y-care all

equal. Thus (a) will have its greatest value when a and y
satisfy the conditions
c - x = c -y = x + y - c
2c
whence
x =y = z •
470 On Inequalities and [707
707. There are two propositions with respect to ratios
or fractions which are sometimes useful. The first is, that
a fraction greater than unity, or as it may be called, a ratio
of greater inequality, is diminished by adding the same
quantity to its numerator and denominator. On the other
hand, a fraction less than unity is increased by the same
process.
a+x a
Thus > or < ,

b+x

as (a + x) 3 > or < a (b + x),

as ab + bx > or <ab + ax,
as bx > or <ах,

as b > or <a ;
a
that is, according to whether the fraction is less or greater
than unity.
708. The second proposition is that a fraction whose
numerator is the sum of the numerators of a number of
fractions, and whose denominator is the sum of their
denominators, is intermediate in value between the
greatest and the least of the given fractions.
Let the fractions, arranged in descending order of
magnitude, be
A1 A2 A3 An
, , ...

61 62 63 bn
1
Then
61 = , therefore a₁ = 61.

6201
‫وو‬ a₂ < b₂ 2 ,
"

An
< " an < b .
Un 61 61
709.] Arithmetical Maxima and Minima. 471

Adding these results,

a₂ + a₂ + a3 + ... + a,, < (01 + 02 + 03 + ... + 0 ) ;
whence a1 + a2 + a3 + ... + an < 1 .
61 + b2 + b3 + ... +6 61
a1 + a2 + ... + an >
An
Similarly
61 +b2 + ... +6bn
709. In some of the foregoing articles the following
principles have been assumed.
If one quantity is greater than a second, any positive
multiple of the first quantity is greater than the same
multiple of the second.
If one quantity be greater than a second the sum of the
first and any third quantity is greater than the sum of
the second and the same third quantity.
If one quantity be greater than a second the remainder
after taking away any quantity from the first, is greater
than the remainder after taking the same quantity from
the second.
If one quantity be greater than a second, any positive
even power of the first is greater than the same power of
the second.
These may be assumed as axioms, extensions of those in
Art. 53. They may also be proved, if the student prefer,
by means of the definition of Art. 699.
Thus a > b, when a-b is positive. But in this case
p (a - b) or pa-pb is also positive ; therefore pa > pb,
which proves the first principle, and the others can be
treated similarly.
472 On Inequalities and

EXAMPLES.

X 2 U
1. Prove that ++ + X > 4 unless x = y = z = u.
Y 2 U

a²+ b²
2. Prove that ab + (a - b) x -x² is never greater than 2

3. Find the greatest value of ax-x².

1 1
4. Find the sum ton terms of the series + 2.3+... , and
1.2

hence shew that (n + 1)"-1 < ( n)².

3m (3m + 1)2 > /
5. Prove that
4
3m.
6. Prove that (1 + x)" (1 + x") > 2n+1 x", n being positive.
7. Prove that xyz > (y + z - x) (z + x −y) (x + y- z), all the
factors being positive.
4
8. Prove that (1) rin + 2nrin-1 > 1 + 2np2n+1
n- 1 n+ 1 n+ 1

(2) r 2 (r2 + n) > nr 2 +1 .

9. Prove that x² + y² + z² > yz + zx + xy.
n³ (n + 2) or (n + 1 ) (n + 3)
10. Which is the greater ?
(n + 1) (n - 1) (n + 2)³n
x² +y² + z² bc+ ca- ab
11. Prove that > 2 , the first fraction
yz + zx + xy a²+ b²+ c²
being positive.
12. Prove that 2n- 1 < | n ( 2n) -1.
13. Prove that (2x+ a)√a-xhas its greatest real value when
a
X
2

14. Shew that pxq + qx -p + rxp- > p + q + r unless x = 1

orp = q = r.
n

15. Shew thatthe limit of (x + a) (x + ar)... (x + ar²-1) when

n is indefinitely increased, r being less than unity, is x.
 Arithmetical Maxima and Minima. 473

16. Supposing that the consumption of tea is forty millions of

pounds when the duty is 2s. 6d. per lb., and that the increase
of consumption varies as the decrease of duty, so that if there
were no duty, the consumption would be sixty millions : find the
consumption when the duty is 2s. per lb., and what would be
the most productive duty.
17. Prove that if x, y, z, ... , a, b , c, be allpositive quantities
96

210

a b C
+ ... +
y
++
..} can never be less than n².
18. Prove that (x² + y² + z²) (a² + b² + c²) > (ax + by + cz)² unless
2
y
==
b C

x+ 4
19. Find the values of a which give the fraction
its maximum and minimum values. (x+5) (x + 6)
x² + 3x + 2
20. If x be real prove that the expression can have
1
x² + 3x
no real value between and 1.
9

21. Prove that

[n (n +1
+
{ 8 > ( n) .
 CHAPTER XXIX.

ON NOTATION AND NUMBERS .

710. Ir a number Nbe divided by a smaller number m,

with a quotient q and a remainder ro, the following identity
holds (Art. 157) , N = qm + ro.
The number m being called the modulus and ro the
residue, any two other numbers are said to be congruent with
respect to the modulus m when they have the same residue.
Thus 12, 17, 22, 27... all leave the same remainder when
divided by 5 and are therefore congruent to each other with
respect to that number as modulus.
The word number throughout this chapter will be always
understood to mean positive integer.
711. If q be a number greater than m, we can similarly
obtain an identity of the form
q = 91m + r1.
Whence, by substitution,
N = q₁m² + r₁m + ro .
If q₁ be still greater than m, we may write
91 = 92m + 2 ,
whence again
N=
= q2m³ + r2m² + r₁m + ro.
This process may be repeated until a quotient qn-1 is
arrived at which is less than m, and finally we obtain
N= qn-1m² + n -1 m²-1 + ... + r₂m² + r₁m + ro ;
or, writing rm for In-1 , since qn- 1 will be also the remainder
after dividing In-1 by m,
N= rnm" + rn-1 m²-1 + ... + 12m² + r₁m + ro . (1)
 On Notation and Numbers. 475

Thus any number can be expressed in a series of powers

of any modulus which is less than itself. The coefficients
ofthese powers, beginning with the lowest, are the successive
remainders after dividing N and the successive quotients
bym.
712. When a number N is expressed in terms of a second
number m in the manner of the last article,m is called the
radix, and the numbers rn, "n-1, ... 2, 1, o, are called
the digits.
These digits are all, from the nature of the process by
which they are found, less than m. Any of them, with
the exception of rn, may have the value zero.
713. In the ordinary system of Arithmetical notation,
the number 10 is employed as radix, and the coefficients
only of the different powers of the radix are written down,
the power of 10 belonging to each digit being indicated by
the position of the digit.
Thus 416732 is an abbreviated way of writing
4 x 105 + 1 × 104 + 6 x 103 +7 × 102 + 3 × 10 + 2 .
Numbers expressed in this manner are said to be expressed
in the scale of the radix 10.
Similarly the number N in (1) of Art. 711 is said to
be expressed in the scale of the radix m.
714. A number expressed in any scale can be expressed
in any other scale by the help of 7 ) 673521
the foregoing articles. Thus, let 796217-2
673521 be a number expressed in 713745-2
the ordinary decimal scale, and let 71963-4
it be required to transfer it to the 7280-3
scale 7. 740-0
The process is represented by 5-5
the diagram: the given number
when divided by 7 leaves a remainder 2. The quotient
divided by 7 again leaves a remainder 2, and so on. By
476 On Notation and Numbers. [715 .
the theory of Art. 711 it follows that the given number
is equal to
5 x 76 + 5x75 + 0x74 + 3 x 73 +4 x 72 + 2x7 + 2 ;
or expressed in the ordinary way may be written as
5503422, it being remembered that the shifting of a
figure one place to the left multiplies its value by 7 and
not by 10.
715. To transfer a number from any other scale to the
scale 10, the student will probably find it easier to work
out the several powers of the radix, multiply by the
successive digits, and add the results.
Thus the number 5503422 in the scale of seven means
5 x 76 + 5 x 75 + 3 x 73 + 4 x 72 + 2 x 7 +2,
and 2 = 2
=
2x7 14

4x 72 = 196

3 x 73 = 1029

5 x 75 = 84035

5 x 76 588245
673521

Hence the number 5503422 in the scale 7 is represented

by 673521 in the scale 10.
716. The transformation in the last article can be also
effected in the manner of the previous one, the only difficulty
being that the student will not
13 ) 5503422 be so familiar with the formal
13 400235-1
1325431-2
results of the multiplication
table in the scale 7 as in the
13 1651-5
13 124-3
scale 10, and thus the division
6-7
will present greater difficulties :
the process is indicated by the
diagram. In the scale 7 the number 10 is evidently
denoted by the symbol 13.
 718. ] On Notation and Numbers. 477

Thus the number is represented in the scale 10 by the

figures 673521 .
The student will at first find the above process facilitated
by turning the portion of the dividend operated on into
the scale 10. Thus 55 in the scale 7 means 5 x 7 + 5, or 40
in the scale 10, whence the quotient by 10 is 4 and the
remainder zero.
717. A number may thus be expressed in any scale, and
transferred from that to any other. The digits in the case
of scales whose radix is not greater than 10 may be expressed
by the ordinary symbols. For scales whose radix is greater
than 10, additional symbols will be required. Thus for the
scale with radix 12, two symbols will be needed for the
digits 10 and 11 which may occur. The letters t and e are
usually employed for this purpose. Thus the number
56t7e43
in the scale 12 means
5 x 126 + 6 x 125 +10 × 124 +7 x 123 + 11 x 122 + 4 x 12 + 3.
It is hardly necessary to add that the processes of multi-
plication, division, square and cube root, &c.,can be conducted
in any other scale on the same principles as in the scale of
10; the only requisite is a familiarity with the expression
of the multiplication table in the scale in which the opera-
tions are carried on.
718. If N be a number expressed in the scale m with
digits ro, 1 , 2 , ... so that
N= ro + r₁m + r2m² + ... ,
we may write N in either of the forms
N = ro +11 + 12 + ... + r₁ (m - 1) + r₂(m² - 1) + r3 (m³ - 1) + ...,
2

or N= "o- 1 + 2-3 + ... + 1₁ (m + 1) + r2 (m² - 1) + r3(m³ + 1) + ... ;

from the first of which it follows that Nwill be divisible by
m - 1 if ro + r1 + 12 + ... be so divisible; and from the second
that N is divisible by r + 1 if ro- 1 + 2-3 + ... be divisible
by r + 1.
478 On Notation and Numbers. [719.
In the scale 10 it follows that any number is divisible by
9 if the sum of its digits be so divisible ; and by 11 if the
sum of the even digits and the sum ofthe odd digits leave
the same remainder when divided by 11 .
719. It has been previously assumed that in the scale of
10 any number with n digits is less than the smallest
number with n + 1 . This is true in any scale. For if N
be the number with n digits in the scale of m, and r denote
the greatest digit, it follows that cannot be greater than
rm - 1 + rm - 2 + rmn- 3 + ... + rm + r ;
that is, cannot be greater than
r (m - 1)
m- 1

Now ris not greater than m - 1. Hence N is not

greater than m" -1 and is therefore less than m", which is
the least number represented in the scale of m by n + 1
digits.
720. Taking 2 as the modulus (Art. 710) any number
must have a remainder 0 or 1. Thus any number can be
expressed in one of the forms 29 or 29 + 1 . All numbers
of the first form are called even numbers, those of the second
are called odd numbers. Thus all even numbers are con-
gruent with regard to the modulus 2, and similarly all odd
numbers.

721. Any two numbers are congruent with respect to a

given modulus when their difference is divisible by that
modulus.
Let N and N be the numbers and m the modulus, and
let NN be divisible by m.
Let N = mq+ r, N'= mq' + r' where both randr' are less
than m .
Then N- N' = m (q - q') + r - r'.
But N- N = mp by hypothesis where p is an integer.
Hence r - r' = m (p - q + q).
724.] On Notation and Numbers. 479

But rand r', and therefore r - r', are numerically less than
m, and the only way in which this equation can hold is that
both sides shall be identically zero. Hencer must equal r ',
or N and N ' are congruent with respect to the modulus m.
The converse of this proposition is obviously true.
722. The problem sometimes occurs to find an expression
for all numbers which are congruent with one number with
respect to one modulus and to a second number with respect
to another modulus. Thus, let it be required to find an
expression for all numbers which are congruent with 5 the
modulus being 7, and with 8 the modulus being 9. The
numbers in question must evidently be equally represented
by the symbols 7p + 5 and 9q + 8. Thus the problem is to
find all possible values of p and q to satisfy the equation
7p + 5 = 9q + 8,
or 7p- 9q = 3 .
The solution of which by the principles of Chapter XXVII
is given by p = 3 + 9t,
q = 2 + 7t,
where t is any integer. Consequently the numbers required,
which are represented by 7p + 5, can be expressed in the
general form 26 + 63t.
723. The series of numbers 1, 2, 3, 4, 5,...ad inf. is called
the series of natural numbers. With reference to any
modulus m the natural numbers may be divided into sets,
each containing m consecutive numbers, which have the
numbers 1 , 2, 3, 4, ... m-
m - 1,0 as residues in the same order.
Thus if 5 be the modulus, the natural numbers may be
arranged in sets, thus
1, 2, 3, 4, 5 ; 6,7,8,9,10 ; 11, 12, 13, 14, 15 ;
and each set has as residues with respect to the modulus 5
the numbers 1, 2, 3, 4, 0.
724. A number which has no factors except itself and
482 On Notation and Numbers. [728.
728. It follows from the last article that if a numberp
which is prime to q divide aq it must divide a.
Another proof can be given of this result. Let the
ordinary process of finding the G. C. M. of q and pbe
carried out, and let 1, 2, 3 , ... be the successive re-
mainders. Since q and p have no common measure higher
than unity, the last of these remainders must be unity.
Let X1 , X2, X3 , ... be the corresponding quotients. Then
we know (compare Art. 180),
9
p
=
= x + 1,
P
whence aq = ax + p
P
arr1 ; (1)
ar
ar1
a=
p= 112+ 12, ‫وو‬
a= P
2+ P
; (2)
ar ar2 ar3
11 = 123 + 13 , ‫دو‬
=
X3 + ; (3)
P P p

ar
Since "
p
is an integer it follows from (1) that P1 is so
also, and from (2) that ara2, and so on. Since the last of the
P
a

quantities 1 , 2, 3 , ... is unity it finally follows that P

- is
integral.
From this the result of the last article can be deduced.
729. A number prime to each of two others is prime to
their product.
Let p be prime to each of q andr ; then pis also prime
to qr. For, if possible, let pand qr have a common factor
s, so that p = st and qr = su, where u and t are prime to
u
each other. Hence = -

But q and rare each prime

8 r

top or st and therefore to s, so that 2 is a fraction in its

8

lowest terms. Thus (Art. 727) u and rare equimultiples

of qand s ; orris both prime to s and a multiple of s,
which is absurd. Thusp must be prime to ab.
733-] On Notation and Numbers. 483

730. Any number must obviously be either a prime

number, or resolvable into factors which are prime. Let
N be the number, and let abcd ... be its prime factors,
of which some may be equal. If possible let N also
equal aẞyd ... , where α, β, γ, 8 are also prime numbers.
Then a divides N or abcd .... Hence a must divide one
of the factors a, b, c, d. For if not, being a prime number,
it must be prime to each of them and therefore to their
product (Art. 729). Let a divide a. Since a is a prime
number it has no factors except unity and itself. Hence
a must equal a. Similarly β, γ, ... must each be equal to
one of the factors b, c, d, ... , that is, the two resolutions into
factors are identical.
731. The most general form of resolution ofa number is
abuc ... where a, b, c, ... are different prime numbers and
x, y, z, ... are any integers including unity.
732. When a number is expressed as in the last article
it is evident that all the possible divisors of it will be the
different terms of the product,
(1 + a + a² + ... + a*) (1 + b + b² + ... + b )
(1 + c + c² + ... + *) ... ;
and the sum of all these divisors is therefore,
ax+ 1-1 bu + 1 -1 +1-1
a- 1 6-1 c- 1
...
(Art. 522).
The number of these divisors is evidently the number of
terms in the above product, or
(x + 1) (y + 1) (z + 1) ... ;
or, if unity be excluded,
(x + 1) ( y + 1) (z + 1) ... -1 .
733. From Art. 723 it follows, that any set of m
successive natural numbers will leave the same residues in
the same cyclical order with respect to the modulus m,
these residues being the integers 1, 2, 3 ... (m - 1 ), 0 .
1i 2
484 On Notation and Numbers. [733-
If any m equidistant integers be taken the same pro-
position holds good provided the interval between the
integers be a number which is prime to m. Thus if any
five integers, increasing in succession by 4, be taken, as
3, 7, 11, 15, 19, the residues of these relatively to the
modulus 5 are 3, 2, 1, 0, 4. The cyclical order of the
residues is different to what it is when the interval between
the numbers chosen is unity.
The general proposition is, that if n be any number
prime to m the residues of the numbers
a, a + n, a + 2n, ... a + (m - 1) n,
with reference to the modulus m will be in some order, the
m numbers 0, 1, 2, 3, ... (m -1) : and the cyclical order of
the residues is independent of the starting point a.
In the first place the residues must be all different. For
if two of them, as those of a + pn and a + qn, were the
same, then (Art. 721),
(a + pn) (a + qn), or ( p ~ q) n,
~

must be divisible by m. But m is prime to n, consequently

(Art. 728) m divides pq, which is impossible, since p
and q, and afortiori pq, are less than m.
Hence, since there are m different residues, these must be
the m integers 0, 1, 2, (m- 1).
...

For the second part it is evident that whatever a may be

the difference between any two consecutive residues must
ben - m or m - n according as norm is the greater.
Hence if any two particular residues are consecutive in the
cyclical order when a has one value they will also be con-
secutive for any other value of a, that is, the cyclical order
is independent of the value of a.
734. Among the numbers which are less than any given
number which is not a prime, some will be prime to it and
others not. A very simple formula gives the number of
the former.
735.] On Notation and Numbers . 485

Let the number N = a* where a is a prime. This is the

simplest case of the general formula ofArt. 731.
Then the numbers not greater than N and having a
common measure with it, are
a, 2a, 3a, 4a, ... a -1 . a.
N
The number of these is ax- 1 or Hence the number
a

of numbers less than N and prime to it is

N
N- N
a or (1 )
735. Let now φ (n) represent the number of numbers
less than n and prime to it ; and similarly let (m)
represent the similar quantity for another number m.
Further, suppose that n is prime to m. Then, (mn)
representing similarly the number of numbers less than
mn and prime to it, it can be shewn that
ф (тп) = ф (т) . φ (п).
For the mn integers from 1 to mn inclusive can be
divided into m sets of n consecutive integers.
Taking the first number out of each set we get the
series 1 , n + 1 , 2n + 1, ... (m - 1) n + 1 (a). The residues
in this series with regard to the modulus m will by the
last article be the same as those of the series
1, 2, 3, ... (m - 1 ), т (3),
only in a different order. The same number of the
numbers in (a) will therefore be prime to m as of the
numbers in (β). That is, the number of numbers in (a)
which are prime to m, must be ф (m) .
Similarly in the series formed by taking the second
number out of each set, namely
2, n + 2, 2 n + 2, ... (m - 1) n +2 ,
there will be $ (m) which are prime to m, and so on for the
486 On Notation and Numbers. [736.
different series formed by taking the third, fourth, &c. out
of each set.
Now all the numbers in each such series will be prime
to n or not according as the first terms, 1, 2, 3, ... n- 1 are
prime to n or not. The number of series in which all the
numbers are prime to n is therefore (n). Hence the
number of those numbers which are prime both to m and
tonis ( n) . φ (m), that is,
ф (тп) = ф (п) .ф (т).
736. Let now Nabuc ....

Then with the notation of the last article,

φ (Ν) = φ (α*) . φ ( ...)
= φ (α*) . φ ( ") . φ ( ...)
= φ (α*) . φ ( *) . φ ( *) ...

= a² (1-1). (1-7). (1-1)... (Art. 734)

a

= a² ... (1 ) (1 ) (1-1)...

= N(1-1) (1-1) (1 ) ... ,

which is the formula required.
737. It follows from Art. 723 that any fraction whose
denominator is m, whether improper or proper, can always
be reduced so that its fractional part shall have for its
numerator one of the m - 1 integers 1, 2, 3, 4, ... , (m - 1 ) ;
unless the denominator be a factor of the numerator, in
which case the fraction is really an integer.
If m be a prime number, all the fractions
1 2 3 m- 1
, , ,
m m m m

will be in their lowest terms .

739] On Notation and Numbers. 487

738. The student of Arithmetic is familiar with the

process of reducing a vulgar fraction to a decimal one, that
is to a fraction whose denominator is some power of 10.
The process really depends on the principle that the
value of a fraction is not altered by multiplying both
numerator and denominator by any number.
Thus any fraction P 10"p where n is any number
= ,

9 10"g
whatever. If, by taking n sufficiently large, 10" can be
made to contain q as a factor, and the quotient of 10"
10"p rp
divided by q be r, the fraction 10"
, which is
= =

9 10" q
the required decimal form.
It is evident that this process is only possible when q
contain no factors except those of 10, namely 2 and 5.
The only fractions then which can be reduced exactly to
decimals are such as have denominators of the form 28.5",
where s and tare positive integers. The denominator of
the resulting decimal will be 10 or 10 , according as s or t
is the greater.
739. Any fraction,,, can however be reduced approxi-
9

mately to a decimal fraction by the following process,

10p 1
a+
P 10p = 1 =
=
a1
=

9 109 10 10 10 + 10
109q
where a₁ is the quotient and r₁ the remainder after divid-
ing 10p by q.
Again, 10/1 72
a2+
1 10/1 =
2

109102q 1
102 = 10 = 10 + 1029
10g'2 2

a₂ and 2 being the quotient and remainder after dividing

10r, by q.
P 12
Thus, =
10
++
102 1029
9
488 On Notation and Numbers. [740.
The numbers a₁ , a₂ are thus the first two figures in the
decimal fraction equivalent to 2.
9

This process can be carried on indefinitely. If any of

the quantities 1 , 2 , 3 , ... vanish, the series will terminate
and the fraction will reduce to a finite decimal. This, as
in the last article, can only be the case when q is of the
form 28.5 .
740. If the quantities 1, 2 , ... none of them vanish
they must after a time recur in the same order. All the
numbers 1 , 2 , ... are less than q. Hence they have only
q- 1 different possible values. If at any stage of the
process any remainder, as 12, occurs again, the whole series
of operations will recur from that point, and thus the
decimal fraction obtained will be in some part at least a
repeating decimal (Arts. 532, 533).
741. The process by which a fraction can be expressed
9
as a series of fractions with the successive powers of 10 as
denominators may be applied to any other radix. Thus, if
m be the radix and the fraction,
9
mp 1
a1+
P mp 9 A1
==== m
+ 1,
9 mq m m mq

mri
1 mr1 a2 12 ,
= +
mq m² q m2 m² m²q
mra
mr2 9 A3 13
=
3
=
3
+ ,

m²q m³q m³ m³ m³q

and so on ; a1 , a2 , a3 , ... being the quotients and 1,2,3 , ...
the remainders after dividing mp, mr₁ , mr₂ , ... by q.
743.] On Notation and Numbers. 489

Thus, finally,
p a1 a2 A3
+ 2
+ + ... ‫و‬
m m² m³

the series being finite or recurring according as q contains

only the same factors, as m, or some different factor.
742. Returning to the case of the radix 10, let us
suppose that q is a prime number, not either 2 or 5, and
consequently also prime to 10. In this case any proper
fraction,, will reduce to a recurring decimal repeating
9
from the first place, and the number of figures in the
period will be either (q- 1) or some factor of (q- 1 ).
743. The student will notice, in the process of Art. 739 ,
that the numbers a1 , a2 , as, ... which constitute the
figures in the final decimal equivalent to P are the
quotients, and 11, 12, 13, ... the remainders after dividing
10p, 101, 102 , ... by q. It may also be observed that
1, 2 , 3 , ... will obviously also be the remainders after
dividing 10p, 10²p, 10³p, ... by q. Assuming then that
q is prime to 10 and also a prime number, each of the
fractions 10/1 1 , 10/22 , ... is in its lowest terms, since the
9 9

quantities 1, 2 , ... are all less than q, and there is con-

sequently no factor common to numerator and denom-
inator.

Thus, if it be required to reduce the fraction 12, for

instance, whose denominator is q and whose numerator is
any one of the series of residues 11 , 12 , ... , the process will
be the same as that of reducing 9
only beginning at a
later point, and in this case the figures of the equivalent
decimal will be a3 , a4 , ....
‫و‬
490 On Notation and Numbers. [744-
Suppose then that a3 is the first figure that recurs in the
decimal equivalent to 2. Let n be the number of figures
9
in the repeating part, the figures themselves being there-
fore a3 , a4, ... An+2. Thus in reducing to a decimal the
9
figures will recur from the beginning and will be a3 , a4 , ...
an + 2. In reducing 93, they will also recur from the first
and be a4, a5 , ... An + 2, a3 , and so on. Thus in reducing
9
to a decimal, n fractions whose denominator are q, and
whose numerators are different integers 12 , 13 , ... n + 1 , will
have been reduced, the period of each being the same, only
the figures beginning at a different point.
r
744. By Art. 533 it follows that the fraction 92 can be
N
expressed in the form where N is the number
10"-1'
formed by the repeating figures. By Art. 727 it follows
that N is a multiple of r₂ . Thus if N = r₂N', we have
1 N'

2 10"-1 , that is (Art. 533), 1 can be expressed as a

repeating decimal of n places recurring from the first.
Multiplying both sides by p, it follows that
P
=
PN'
9 10"-1

Hence P can also be expressed as a similar repeating

L

decimal. Thus, if q be any prime number, all the fractions

1 2 3

2
,

qq
1-1 will be convertible into repeating
decimals of the same number of places recurring from the
commencement .

745. Again, if the number of places in the period be n,

we have seen (Art. 733) that in working out the period for
any one of the above fractions we really work out the
748.] On Notation and Numbers. 491

period for n of them. Thus q- 1, the whole number

of fractions, must either be n or a multiple of n, since the
fractions must either all have the same figures in their
repeating decimal, or must be divisible into a number of
sets of n, all of each set having the same figures. Thus n,
the number of places, must be either q- 1 or a factor
of q - 1 .
746. It follows that the process of reducing to a
9
decimal must always repeat itself after q - 1 places. Hence
by the observation in Art. 743 it follows that the re-
mainder after dividing 10-1.p by q is p. Thus
109-1 . p = Mq + p,
where Mis some integer ; or ( 10º−1−1 ) p = Mq.
But since p divides Mq and is prime to q it follows that
pdivides M. Let Q be the quotient.
Hence
109-1-1 = Qq. (a)
747. The conditions under which the equation (a) holds
good, are that q is a prime number and also prime to 10.
If any other number Nbe taken as radix the reasoning of
Arts. 742-746 holds good, provided q be a prime number
and prime to N. Hence under these conditions, it follows
that
Nq-1−1 = Qq, (β)
where is some integer.
This result (3) is known as Fermat's Theorem. Another
proof can be given depending on the proposition in the
next article.
748. The product of any n consecutive integers is
divisible by n.
This is evident from the fact that the fraction

(m + 1) (m + 2) ... (m + n − 1) (m + n)
n

represents the number of combinations of (m + n) things

492 On Notation and Numbers. [748.
n together (Art. 488) and must therefore be integral, that
is, the product (m + 1) (m + 2) ... (m + n) of any n con-
secutive integers is exactly divisible by n. A more direct
proof can be given.
Let the above fraction be denoted by the symbol I(m, n).
m+ n
The fraction can also be written as (Art. 490),
mn

and is thus symmetrical with respect to m and n.

Then I(m, n) = (m + 1) (m + 2) ... (m + n - 1) .m

n

+ (m
(m ++1)(
1)(m +2)...
+ 2) ... (m + n - 1).n
n

= m (m + 1) ...n(m + n - 1 ) + (m + 1) ...n-(m1+ n - 1) ,
or
I(m, n) = I(m - 1 , n) + I(m, n − 1). (1)

Now I(m, 2) = (m + 1) (m + 2),

2
and this is always

integral since one of the two numbers m + 1 , m + 2 must

be even (Art. 720).
Consequently I(2, n) is also integral for all values of n.
By the above relation (1), we have
I(m, 3) = I (m- 1, 3) + I(m, 2).
Hence, since I(m, 2) is integral, I(m, 3) will be integral if
I(m - 1, 3) is so. But I(2, 3) is integral, consequently
I(3, 3) is so, and therefore I(4,3) and I(5, 3), and so on,
to any value of m or I(m, 3) is an integer.
Again by (1), I(m, 4) = I (m - 1, 4) + I(m, 3), and since
I(m, 3) is integral it will follow that I(m, 4) is so if
I(m - 1, 4) is an integer. Hence, as above, I(m, 4) is seen
to be always integral : and in a similar way I(m, 5) and so
on to I(m, n), where n is any integer.
750.] On Notation and Numbers. 493

749. It follows that all the coefficients in the expansion

of the binomial (a +x) ", where n is an integer, are integral.
The coefficient of the r + 1th term is
n
or
n (n - 1 ) ... (n - r + 1)
r n- r r

If n be a prime number, since none of the factors in the

denominator can divide n, this coefficient will be divisible
by n. Hence all the terms of the binomial expansion
(a+ x)", when n is a prime number, are divisible by n
except the first and the last, namely a" and ".
750. The coefficient of any term in the expansion of
n

(a + b + c + d + ...)" can be written as ,

where
PIr ...
p + q + r + ... = n (Art. 550).
This coefficient is obviously an integer from the method
of its formation. Hence, if n be a prime number, this
coefficient will, as before, be always divisible by n except in
the cases when one of the quantities p, q, r, ... is equal to
n and the others zero. The coefficient in this case is unity
and the corresponding terms are a", b", c", ....
Hence

(a + b + c + d + ...)" = a + b + c + ... + M. n,
where M is some integer.
Now let each of the numbers a, b, c ... be unity, and let
the number of them be N. Hence
N = N + Mn,
or N - N = Mn,
or N(N -1-1 ) = Mn.
If then N be prime to n it follows that the other
factor N -1-1 is divisible by n, or
Nn- 1-1 = Q.n,
where is some integer.
Thus we again arrive at the theorem of Art. 747 .
494 On Notation and Numbers.

EXAMPLES .

1. Express 2345 in the scales of 5 and 9.

2. Express 174.26 in the scale of 5.
3. Transform 1234-56 from the denary to the septenary scale.
4. Express in the common scale and in the scale of 8 the
number denoted in the scale of 9 by 723.
5. Transform 119716 into the scale 7 and extract its square
root in that scale. Transform the square root back into the
scale of 10.

6. Multiply 1t923 by t15 in the scale 12; divide the product

by 3 and extract the square root, and transform the result to
the scale 10.
7. Extract the square root of 112123-0213 in the scale 6,
and transform the result to the scale 8.

8. Shew that any number of six digits in the decimal scale

formed by the repetition of three digits in the same order is
divisible by 7, 11 and 13.
9. Find a number which is expressed by the same two digits
in the scales of 7 and 9.

10. The difference of two numbers having the same digits is

35453221, in what scale are they expressed ?
11. If S1 , S2, S3 be the sums of every third digit in a number,
beginning at the units', tens', hundreds' place respectively, and
in each series making the alternate digits negative; then the
number is divisible by 7 if S₁ + 3S + 2 S is so.
12. The successive digits of a number expressed in the decimal
scale beginning with the units' place are a , a₁, a2, a3 , ... ; shew
that the number will be divisible by 77 if
a2-a + ag ... =
a -a + ag ...
a₁ - a₄ + a - . = - 4 5

13. The number 1865 when expressed in a certain scale

becomes 12345; find the radix of the scale.
 On Notation and Numbers. 495

14. Find the scale in which a certain number is expressed by

1111 when were the scale doubled it would be expressed by 125.
15. A certain odd number is expressed by 6 digits in the
scale of 3 andby the last three of those digits in the scale of 12 .
Find the number.

16. Anumber is denoted by 4.440 in the quinary scale and

by 5.54 in a certain other scale. What is the radix of the
latter ?

17. Shew that a number in the decimal scale is divisible by 3

ifthe sum of its digits is so divisible.
18. Shew that a number is divisible by 8 if the number formed
by its last three digits is so divisible.
19. The square of every number is of the form 3m + 1or3m ;
and the square of every prime number greater than 2,diminished
by unity, is divisible by 8 without remainder.
20. Every cube is of one of the forms 4m or 4m + 1 .
21. If n be a prime number different from 2, then in the
scale ofnotation whose radix is 2n, any number ends with the
same digit as its nth power.
22. If (r - 1 ) be a prime number, then in the scale of r, of the
numbers 121 , 12321, none but the last is divisible by r - 1 ,
and that is divisible by (r- 1)², the last digit in the quotient
being unity.
2

23. If n be an integer, shew that 7n³ - 3n² - 4n is divisible

by 6.
5
24. If m be any even number, m7 - m -m³ + m is divisible
by 90.
2n
25. Prove that if m be a prime number n + 1n
is an integer.

26. If m be a multiple of 5, m (m² +89) is divisible by 30.

1
27. If be converted into a circulating decimal with p- 1 re-
P
curring figures a1 , a2 , a3, ... ; shew (1) that p is a prime number ;
496 On Notation and Numbers.

(2) that the recurring period being multiplied by 2, 3, ... p - 1 ,

will reproduce its own digits in their own order ; (3) that
=

a1+ ap + 1 a2 + ap + 3 = ... = ap-1 + ap-1 = 9 .

2 2 2

28. How many numbers will divide 800 exactly ?

29. What is the number of factors of 720 ?

30. How many numbers are there less than 720 and prime
to it ?

31. If p be a prime number, not a factor of a and not equal to

(a- 1), then the sum of the remainders when a, a², a³, ... as are
divided by p is less by unity than a multiple of p. Also shew
that the coefficients of (1 + x) -1 will differ from multiples ofp
by unity in excess or defect alternately.
Shew also that ifA , A1, A2 , ...
be the coefficients in (1 + x)-²,
then A - 1 , A1 + 2, 42-3, A3 + 4, ...will be multiples ofp.

2
32. If n be a prime number and pany integer, prove that
(n²p² - 1) -1 + 1 and (np - 1)*-1+ (n + 1) -1 have the same
remainder when divided by n.
33. If m and n be any two prime numbers, prove that
(m"-1 + nm-1) divided by mn leaves a remainder unity.
34. If n be a prime and m not divisible by n, shew that
2n - 1
m² -1-1- (m- 1) is divisible by n.
nn - 1

35. The difference of the squares of any two prime numbers

greater than 3 is divisible by 24.
36. Shew that three prime numbers, each greater than 3,
cannot be in arithmetical progression unless the common
difference is a multiple of 6.
37. If in the scale of 12 a square number ends with a single
cipher the preceding digit must be 3, and the cube of the square
root ends with 60.
 CHAPTER XXX.

PROBABILITIES OR CHANCES .

751. ALL persons are conscious that their expectation

of any undecided event may vary from an almost certain
anticipation that the event will happen to an almost equal
certainty that it will not. Thus expectation in such cases
is susceptible of being greater or less, and therefore must
be capable of quantitative measurement. It can therefore
also be represented by a numerical symbol. The number
which represents the expectation of a well-informed person
in regard to a given undecided event, is called 'the chance'
of that event happening. The unit is taken to represent
that degree of expectation which is called certainty, and any
less expectation than this must therefore be represented by
a proper fraction.
Thus the numerical measure of 'the chance ' of any
uncertain event is always a proper fraction.
752. Suppose a halfpenny is tossed up. As it falls it
must have either ' head ' or ' tail ' uppermost. The expec-
tations of these two events are equal. As one of them must
happen, the sum of the expectations is certainty. The sum
of the two equal chances of ' head ' or ' tail ' is therefore
unity. Hence each of these chances is represented by .
753. Similarly, if a six-faced die be thrown, the chances
of the different numbers being uppermost are all equal.
The sum of the chances is unity, hence each of them, as for
kk
498 Probabilities or Chances. [754-
instance the chance that the number 2 shall be uppermost,
is .
754. The chance that either 1 or 2 shall be uppermost
is similarly or , since of the six ways all equally likely
two are favourable.
755. The general rule is that, if any event can happen
in a ways and fail in b ways, the a + b ways being the only
possible cases and all being equally likely, the chance of the
a

event happening is a+ b and the chance of its failing is

b

a+ b
For, evidently, if the chances of the event happening and
failing respectively be a and y, we have the two conditions
x : y :: a : b,
x +y = 1 ;
the former equation being obvious from the data, and the
latter expressing the fact that the event must either happen
or fail.
From these equations we easily obtain
a b
= •

a+b ' y = a+b

756. The result of the last article is the foundation of
the mathematical treatment of chance. The student who
desires further elucidation of the logical basis of this result
can consult with great advantage the ' Essayon Probabilities'
of the late Professor De Morgan. For full information on
the earlier history of the subject, Dr. Todhunter's ' History
of Probabilities' may be referred to.
757. Let p represent the chance that the event A will
happen ; then 1 -p represents the chance that A does not
happen. Similarly, if q be the chance that another inde-
pendent event B happens, 1 -q will be the chance that B
fails.
758.] Probabilities or Chances. 499

With these assumptions the chance that

(1) A happens and B fails is p ( 1-9) ;
(2) A fails and B happens is (1 -p) q ;
(3) A happens and B happens is pq ;
(4) A and B both fail is (1 -p) (1-9) .
For, reverting to the notation ofArt. 755, let a be the
number of ways in which Acan happen and 6 the number
ofways in which it can fail, while a' and b represent the
number of ways in which Bcan happen or fail respectively.
Then each of the ways of A happening may be combined
with each of the ways of B failing, and there are thus ab'
ways altogether in which Amay happen andB fail; that is,
ab' is the number of ways in which the compound event
described in (1) may happen. But the total number of
possibilities is (a + b) (a + b'). Hence the chance of ( 1)
must be ab
(a+ b) (a + b) '
a
or
a+ ba+ b'
that is, p (1-2).
Similarly, the chances of (2), (3) and (4) may be shewn
to be as stated.
Hence the chance of the concurrence of two independent
events is the product of the chances of those events happening
separately.
758. As an instance of the last article, suppose that
there are two bags, one containing three black and four
white balls, the other containing five black and three white
balls, and let it be required to find the chance of drawing a
black ball out of each at the first drawing.
There are eight possible draws out of the second bag, and
seven out of the first ; any one of the eight may be com-
Kk2
 わこ

500 Probabilities or Chances.f [759-

bined with any one of the seven ; that is, on the whole there
are 7 x 8 distinct possible results of drawing one ball out
of each bag. All these possibilities are equally probable.
There are three different ways of drawing a black ball from
the first, any one of which may concur with either of the
five possible ways of drawing a black ball from the second ;
✓ that is, there are 3 × 5 different ways of drawing a black ball
out of each bag. Thus the chance of drawing a black ball
3x5
out of each bag at the first trial must (Art. 755) be 7x8 ог
3 5
as it may be written 7 × 8 ; that is, the product of the
chances of drawing a black ball at the first trial out of each
bag separately.
759. By reasoning similar to that of Art. 757 it will
appear that if P1, P2 , P3 , ... be the chances of any number
of independent events, the chance of the concurrence of all
these events will be the product of these separate chances,
or P1 , P2 , P3 , ....
760. The investigation of the measure of the chance of
any event simple or compound is thus reduced to depend
on the calculation of the number of ways in which the event
may happen or fail. In the case of a compound event, this
is again reduced to the consideration of the number of ways
in which each of the separate events can happen or fail.
761. Let pbe the chance of a given event happening
and q that of its failing in any one trial ; then p + q = 1 ,
since the event must either happen or fail. It is required
to find the chance of the event happening exactly r times
in n trials.
The chance that it will happen in one particular set of
trials out of the nand fail in all the others is evidently
p' q"-". The number of different sets of rout of the n trials
n

is (Art. 490) equal to Hence the chance that

n- r
762.] Probabilities or Chances. 501

the event will happen in any set of r trials out of n, and fail
n

in all the rest, is r n- r

pq- .

This is evidently the r + 1th term in the expansion of

(q + p)" (Art. 547).
It follows that the chance of the event happening at
least r times in n trials is the sum of all the terms in the
above Binomial expansion beginning with the r + 1th, while
the chance that it will not happen so often asr times is
the sum of all the terms preceding the r + 1th.
Since q + p = 1, we have (q + p)" = 1, whence it may be
noticed that the sum of the chances of the event happening
at least r times, and not so often as r times, is unity ; as of
course it ought to be, since one or other of these events
must happen.
762. As an example, let it be required to find the chance
of throwing head exactly seven times in ten tosses of a
coin.
1
Here the chance
1
of throwing head being 2 that of
failing is also 2: thus the chance required is equal to
10 3 7 10 10
=

2 37
37
The chance of throwing head at least seven times is the
10

sum of the last fourterms ofthe expansion( + ) , or is

1218 10 9 10
| 10 1317 | 10
+ +
37 (2) + 28 9 2

= ( ) 10.9.8 + 10.9+ 10+ 1};

110 176 11 11
=
= ( ) { 120 + 45 + 10 + 1 } = 210 26 64
502 Probabilities or Chances. [763 .
The chance of throwing head less than seven times is
the sum of the first seven terms of the expansion, or
9 10.918 2 10.9.817 13
( +10 ( ) ( ) +
+ + 1.2 1.2.3 () () 2

10.9.8.716 14 10.9.8.7.615 15
+
1.2.3.4 ( ) ( )* + 1.2.3.4.5 ( ) ()
6
10.9.8.7.6.5/14
+ 1.2.3.4.5.62 ()
110
=

() {1 + 10 + 45 + 120 + 210 + 252 +210 }

2

848 53
= =

210 64

11 53
The sum of the two chances and is of course unity.
64 64

763. The preceding articles have all referred to problems

of what may be called direct chances. We have next to
consider briefly the subject of inverse chances.
Suppose that a given event has happened. It may have
been produced by any one of a certain number of causes.
It is required to find the chance that it shall actually have
arisen from any particular one of these causes.
Let P1 , P2 , ... P be the probabilities of the existence of
each of the several causes, estimated before the given event
has been observed to happen. Let P1, P2 , P3 , ... Pn be the
probabilities that the event would happen on the supposition
of the existence of each of the possible causes. Then the
chance of the compound event that the first cause should
exist and the event happen as a consequence of it is, by
Art. 757, P1P1. Similarly, the chances that it will happen
as a consequence of each of the other causes are
PaP2, P3P3, ... PnPn
Pr
Now it is in accordance with the ordinary behaviour of
764.] Probabilities or Chances. 503

reasonable beings to assume that the chance of the existence

of any one cause as the efficient agent in producing the
givenevent is proportional to the chance, estimated before
the event has been observed, that the event would happen
as a consequence of the given cause.
Hence, if X1 , X2 , X3, ... , be the chances required that the
event actually was produced by the 1st, 2nd, nth causes ...

respectively, we have
X1 X2 Xn x1 + x2 + ... + Xn
= =
...

PiP1 PaPa
2 PaPn PP1+PaP2+... +PnPn
(Art. 657).
But the event has happened, and must have been produced
by one of the causes. Hence
X₁ + X2 + X3 + ... + x = 1 ,
and therefore
PP1 P1 P2P2
X1 ‫ و‬.. ‫وه‬
P₁P1+PaP2+ ...+ Pa Pa = 2 (PP) ' 2 = 2(PP)
and so on ; where ∑(Pp) is an abbreviation for the sum of
all the products PiP1, Pap2 .....
764. Suppose, for instance, that a bag is known to con-
tain six balls which may be either black or white. A ball
is drawn and proves to be black, it is required to find the
chance that it is the only black ball.
Apriori there are seven possible and equally likely cases,
namely, that the bag may contain 6black; 5 black, 1 white ;
4 black, 2 white ; 3 black, 3 white ; 2 black, 4 white ;
1 black, 5 white ; or 6 white balls. The chances of the
existence of each of these causes, or the values of the
1
quantities P, P., P3, P4, P., Po, P are therefore each 7
The probability of the event happening if the first state
exist, is evidently certainty or 1, so that p₁= 1 . Similarly,
5 4 3 2 1
P2= 6P3 =6 P4 =
= 6 P5 = 6P = ,
P = 0.
504 Probabilities or Chances. [765.
Hence
1 6 1 5 1 4
X
RP₁ =
= 7
X
6
PaP2 = 7
PSP3 = 6

1 3 1 2 1 1

PAP4 = 7 X
6
,
PoPs = PoP = 7X PP = 0.
6 6
Hence X1 = =

6+5+4+3+ 2+1 21

5 4 3 2 1
=
X3 = X =
X2 21 21
4= 21 = 21 21

The chance therefore that the ball drawn is the only

1
black ball is
21

765. One very important application of the theory of

chances is to the calculation of the value of the expectation
of a sum of money, the obtaining of which is contingent
on some as yet undecided event.
The general rule will be best elucidated by considering a
particular case. Suppose that there is a lottery with a + b
tickets altogether, a of which give a prize of £c, while the
remaining b are blanks.
The whole sum of money to be divided among the
purchasers of tickets is thus Lac. Any person who chose
to purchase all the (a + b) tickets might therefore pay this
sumwith a certainty of not losing. As all the tickets have
an equal chance of winning the prize, and as the fair price
for the (a + b) tickets is seen to be Lac, it follows that a fair
price for one to a man, to whom the possible loss is a matter
ac
of no importance, must be ₤-
a+ b

Thus the value of such a man's expectation may be

measured by the above sum. But the sum he will obtain
a

if successful is £c, and a + b is the chance of his winning a

768.] Probabilities or Chances. 505

prize. Hence the value of his expectation is measured by

the product of the chance of his getting a prize into the value
of the prize he may get.
The general principle to which this example leads is :
The mathematical value of the expectation of a con-
tingent gain is represented by the product of the sum of
money which is possibly attainable into the chance ofgetting it.
766. The proviso that the loss of the sum of money paid
for the ticket is a matter of no importance to the speculator
is avery important one, as distinguishing between the moral
and mathematical values of expectation. In the book
' Choice and Chance,' by W. A. Whitworth, will be found
a very full and clear discussion of the difficulties of this
subject.
767. One of the most important practical applications of
the foregoing theories is afforded by the subject of Life
Insurance. A Company for the Insurance or Assurance of
Lives is a society which, in consideration of regular annual
payments during the life of the assured person, undertakes .
to pay over to his representatives a certain sum after his
death.
The primary basis of the operations of such societies is the
Life Table. From a considerable number of observations of
the mortality among persons of the class who will be clients
of the society it is found that of a given number lo born
at any time, a definite number I will be alive at the end of x
years. The series of numbers lo, 11 , 12 , 13 , ... l ... to the
last year at which it is found that any persons survive,
constitutes what is called the Life Table. It is assumed
that such a table will represent the actual mortality in the
future among persons similarly situated to those who were
the subjects of the observations from which the Life Table
was formed.
768. It follows that of le persons alive at the age of a
506 Probabilities or Chances. [769.

years, only +1 will survive to the age of a + 1 years.

Hence the chance of any one person aged 2 years living
for one year more, is estimated by the fraction 2+1 . The
chance that he will die within the year is therefore
1-12+1 , or +1 .
Similarly the chance that a person aged 2 years will live
for n years is +", while the chance that he will die before
lx+n 2+18
the expiration of n years is 1 - しょ
x x

The latter result may also be seen from the consideration

that l - len
x+n is the number of persons out of the who die
within the n years considered.
769. The first problem is to find the formula giving the
mathematical value of the expectation of a sum of £A to
be paid at the end of the year in which death takes place
to the representatives of a person aged 2 years.
The present value of a sum of EA certainly to be paid
after n
years is, by Art. 537, L
4. The chance of the
R

payment being made at the end of n years exactly is

1+ 1-12+ . Hence, by Art. 765, the present value of
x

the expectation of the payment of £.1 at the end of n years

exactly is R
Lx
770. The above expression can be put into another more
convenient shape. Letz, n denote the chance that a person
1エル
aged a will live for a years ; then Px, n =

1
Let also be substituted for Then the value of the
R
771.] Probabilities or Chances. 507

expectation of £A being paid at the end of n years, which

is £ x +n-1- lz + n. A becomes £( Pz, n-1 -Px, n) v™A.
Rn
The sum of the values of this expression for all values of
n, from unity to the last age in the table, is the present
value of the assurance. Thus the present value of an
assurance of £1 , which is the sum usually taken for con-
venience of calculation, may be denoted by
8

(Pz, n-1 -P2, n) v" , (a)

where the symbol of summation refers to n.

771. As the payment to the assurance company is not
usually made in one lump sum but by a series of annual
payments during the life of the assured, it is necessary to
investigate also the present value of such a series of pay-
ments, that is, of an annuity of a fixed sum payable during
the life of the assured.
The present value of the payment of £1 at the end of one
year will evidently be £(VPz, 1) or vlx +1 . The sum of the

present values of all the payments of £1 at the end of one,

two, three, ... years, will be evidently

£ vlx + 1 + v2lx + 2 + v3lz+3 + ...

(β)
la

the summation being carried on to the end of the ages

included in the Life Table.
Let this sum be denoted by az, and the present value of
the assurance given in the last article be denoted by A..
Then, if the annual premium required to assure £ 1 at
death be Luz , and the first payment is to be made at once,
the present value of the payments will be
£ (Ux + aUz),
508 Probabilities or Chances. [772.
and that of the assurance is £A₂ . These must in fairness
be equal. Hence Ux ( 1 + ax) = Ax ,
whence, ax andA having been calculated from the formulae
(a) and (3), u is found.
772. The calculation of an is much facilitatedby a slight
alteration in the form of its expression.
Thus,
ax
Vlx + 1 + v2lx+ 2 + ....,

vx+ 1 lx+ 1 + v + 2 x +2 + ... ,

by multiplying numerator and denominator by v .

Let the value of the product vl be formed for each age
ofthe table, and let it be denotedby D. If these products
be arranged in a column ; in an adjacent column can be
placed opposite each age the number obtained by adding the
values of D for all greater ages than the one considered.
The number opposite the age x in this second column is
denoted by N.
Thus N = Dx+1 + Da+2 + ... to the end of the table.
Hence aI =
2x + 1lx + 1 + 2x + 2lx+2 + ...
v2l

=
Dz+1+ D +2 + ...
D
N
=

D
so that when the D and Ncolumns have been computed the
value of az is derived by a simple division.
773. The formula for the present value of an assurance
can be similarly reduced. Let us denote by de the number
of persons who die in the ath year of their age: thus
dx = x-1-
775.] Probabilities or Chances. 509

Hence (Arts. 769, 770),

A =
(Ix- x + 1) + (1 + 1-12 + 2) ² + (12+2-1 +3) 23 + ...
20

=
dx + 1V + dx +2v2 + dx + 3 v³ + ...
la
= vx+ 1dx + 1 + va+ 2dx+2 + ....
υα
multiplying numerator and denominator by v*.
Let now the series of products de be formed for each
age of the Life Table and arranged in acolumn. Let the
numbers in this column be treated as those in the Dcolumn
were to form the N column, and let M. denote the sum of
all the numbers opposite the ages x + 1 , x + 2, ... to the end
of the table. Thus
Mx = vx + 1dx + 1 + vx + 2dx + 2 + ....
Hence, evidently,
M
A =
D
774. The annual premium u required to assure £1 at
death can be expressed in a simple form.
It has been shewn (Art. 771) that
A
Ux
1+ ax
M
D M M
=
;
N Na+D N -1
1+
+D
since Nx-1 = Na+D
775. The calculation of the present values of life annuities
and assurances on lives is thus reduced to the computation
of the values of D, N, and Mfor all ages in the Life Table
and for all reasonable rates of interest. For fuller infor-
mation on the subject the reader may be referred to Jones
on 'Annuities and Reversionary Payments.'
510 Probabilities or Chances.

EXAMPLES.

1. Abag contains four white and three black balls. Find

the chance, if two balls are taken out, that one is of each
colour.

2. If the balls in the last question be taken out one by one

until all are exhausted, find the chance that the first four drawn
out will be white. Find also the chance that the first three will
beblack.

3. Four persons draw each a card from an ordinary pack.

Find the chance (1) that one card is of each suit; (2) that no
two cards are of equal value ; (3) that one is of each suit and
no two of equal value. 1

4. Find the probability of throwing 10 with three dice.

5. Four persons cut a pack of cards in succession. The first
spade wins. What is each person's chance of winning ?
6. Two persons throw a pair of dice. Find the chance that
the first person's throw shall exceed that of the second by at
leastunity.
7. A and B play for a stake which is to be won by the one
who makes the highest score in 4 throws of a die: after 2
throws A has scored 12 and B9, what is A's chance of winning ?
8. A has backed a horse to win at evens for £1000 ; after the
raceB tells A that he has heard from C that the horse has won :
supposing B speaks truth 7 times and C 9 times out of 10, what
is the value of A's expectation ?
9. What is the probability that a number consisting of seven
digits, the sum of which is 59, will be exactly divisible by 11 ?
10. There are n dice in the shape of regular tetrahedrons
having each one side marked with 2, two sides with 1, and
the fourth side blank. If they are thrown on a table, shew that
the chance of the sum of the numbers on the uncovered sides

amounting to 3n is n
2" 1.3.5 ... (2n -1) •

4
n
 1,6,3
4
Probabilities or Chances.‫کرا‬4 4 511
2,533,
11. A bag contains six shillings and two sovereigns. Find
the chance of a person's drawing a shilling at the first, second,
or third time respectively but not before, the coins not being
replaced. What is the value of his expectation ifhe be allowed
to draw until he draw a sovereign,keeping all the coins he draws ?
12. There are mn balls which are distributed into m bags, n
balls being placed in each. Find the chance that two specified
balls will be found in the same bag.
13. Abag contains an equal number of black and white balls.
Find the chance that any drawing will consist of an equal number
of each colour.

14. A pack of cards consisting of four suits of 13 cards, is

dealt among four players. Find the chance (1) that a parti-
cular player has no card of a particular suit; (2) that he has no
card of some one of the suits.
15. There are three balls in abag which may be either white
or black with equal probability. Aball is drawn and found to
be white. Find the chance that it is the only white ball.
16. If the ball in the last question be replaced, find the
chance that the next drawing will give a black ball.
17. There are three balls in a bag each of which may be white
black or red. A ball is drawn and proves to be white. It is
replaced. Find the chance (1) of all the balls being white : (2)
that the next drawing will give aredball, it being assumed that
all arrangements of colours among the balls are equally probable.
18. There are five balls in a bag, known to be some black and
some white, all states consistent with this being equallyprobable.
Two are successively drawn and replaced and found to be each
white. Find the chance that the next drawing will produce a
black ball .

19. Two bags A andB contain, each of them, 3 balls ; 2 white

balls are drawn from A and 2 black ones from B, and are put in
BandA respectively. If one ball be now drawn from each, what
is the probability of the two being of the same colour ?
20. A bag contains five coins which may be either shillings or
sovereigns, any state of things being equally probable. Two are
512 Probabilities or Chances.

drawn and found to be shillings : what will be the value of

another drawing of two coins, supposing those first drawn not
to be replaced ?
21. A man is known to have in his pocket, half-a-crown
in small silver. A coin taken from it at random is found to be
a shilling ; shew that the chance of his having another shilling
490
is •

963

22. A bag contains n balls, but nothing is known about their

colours, that is they may be all of different colours or any
number of the same colour. A ball is drawn and found to be
black : it is replaced, and then a second drawing made with the
same result : supposing the ball drawn the second time to be
also replaced, shew that it is 3 (n + 1 ) to n -1 in favour of a
third drawing giving a black ball.
23. A speaks truth three times out of four, B four times out
of five ; they agree in asserting that from a bag containing nine
balls, all of different colours, a white ball has been drawn. Shew
96
that the probability that this is true is 97
24. Adopting De Moivre's hypothesis that out of 86 persons
born, one dies every year until all are extinct : shew how to
obtain an expression for the present value of an annuity which
A, who is now 86 -m years old, will receive, if he be alive, after
the death of B, who is now 86 - n years old.
25. With the notation of Art. 772, shew that the present
value of a life annuity of £1, the first payment of which is to be
N
made at the end of (n + 1) years, = Dr
26. Shew that the annual premium which ought to be paid in
order to secure the annuity in the last question, the first payment
being made at once and the others annually until the reversion
N +n
is entered upon, = N -N2+18
 ANSWERS TO EXAMPLES.

CHAPTER I.
61
1.
140

2. (1) 6 days 13 hrs. 44 m. 49 sec.; (2) 567889, the first

answer is a fractional part of the original length of time, the
second denotes the number of times the smaller time is contained
inthe larger.
646 150 8 27 1
3. ,
385'385' 55 80 3

7. ( 1 ) 6 ; (2) 1 .

CHAPTER II .

1. ( 1 ) 4 ; (2 ) 1 ; (3) 5. 3. 3; 2; 8.
5. ( 1 ) 3 ; ( 2 ) 0 ; (3 ) 2.
6. (1 ) 49; (2) 49 ; ( 3) 361 ; (4) 361.

CHAPTER III .

1.2a + 2b + 2c. 2.2x + 6 ; 8; 64.

3. x . 4. 6x² + 2 .
5. 2x + 3x² + 2x² + 1 ; x + 4x + 6x + 4x + 1 .
6. 0 . 7. a -b + c -d + e -f. 8.10-26.

9. -2a+ 3b-c; value of sum is -1 in case supposed.

10. 2a. 11. a² + b² + c². 12. (1 ) 36 ; (2) 36.
Ll
514 Answers to Examples.

CHAPTER IV.

1. 10 ; 12ab³ ; 30abc³.
2. 20ab2-15 a* b* +10 ab * -5a²b .
3. -21x²y +42x²y²- 14ху³. 4. 6x² + 13xy+ 6y².
5. 28 х² - 65ху + 28y². 6.8x² + 27 .
7.27 a³- 863. 8. x² - 6x² + 11x- 6.
9. x² - 10x + 35x² - 50x + 24. 10. x² - 5x² + 4.
11. x - 9x² + 12x- 4 . 12. x² -y - 2yz - z2 .
13. x² + x²y- x²z - y²x- 2xyz-xx² -y -yz + yz +23 ;
x +y + z - 2y² z² - 2 z² x² - 2 x²y².
14. 4x - 11 xy + 20 xy² - 30 xy² + 20x²y - 11 xy² + 4y .
15. x - 2x - 22x² - 7 x² - 42x + 72 . 16. x .

17. 2(ab + ac + bc). 18.2x - 18x² + 39x² - 25x² + x + 1 .

19. x² + 1. For 20, 21, and 22 use Article 116. 29.16x4.

CHAPTER V.

1. 8abc ; 5ay². 2.4a- 3b.

3. -4a² + 3ab - 5b2 . 4. 3α²x- 4α²x² + 5ax² + 7x .
5. x- 2а ; х - а. 6. x² - 5ax + 6a² .

7. x- a . 8. x² + x²y + xy² + x²y + xy² +y² .

9. x² + xy² + y². 10. x² + xy² + y².
11. x² - ху + y². 12. x² + xy+ y² .
13. 9a² + 6ab + 462. 14. x² + 3x + 2 .
15. x² - 3x + 2 . 16. x² + 2x - 3 .
17. х- у- z. 18. x² - 2x²y + 3 xy² - 4y².
19. x² - y² - 2yz- z2 . 20. x²-y² + 2yz-z .
21. b - c . 22 b + c. 23. (a² + ab + b²) (x − b).
 Answers to Examples. 515

24. xy+ yz + 20 + 22. 25.1 + y² + x² -y-z -yz.

26. x - 5x -x² + 2x + 4x² + 5x + 7 ; 6.
27. ( 1) x² + x - 4x - 4x - 3x- 3 ; -1 .
(2) x - 4x - 3 ; -1.
28. x + 6x + 6x + 9x + 9x² + 16x + 25 ; 29; 2. The second
part is obtained by giving to x the value 10.
29. Divide first by x- 2 andthe quotient by x- 5 ; the result
is x² + 2x + 3x² + x + 3 .
30. x² + ( 2a + b)x + a² + ab + b² .
31. (x + y + z + и) (x + y - z - u) (x - y + z - u) (x - y-z + u).
32. (x - y + z + u) (x + y - z + u) (x + y + z - u) (y + z + и -х).
33. The expression on the left-hand side can be written out as
383-2(a + b + c)s² + (bc + ca + ab)s - s-(a + b + c)s2
+(bc + ca + ab) s - abc } = 28-( a + b + c) s² + abc = abc.
Or the problem may be solved indirectly by means of Art.
167. It is easy to see that if a have the value zero, the
expression on the left-hand side vanishes. Hence it must
contain (a-0) or a, as a factor. Similarly it contains b and c
as factors. Hence, since it is of only three dimensions (Art.
112), in the letters it must equal kabc, where k is some number
independent of a, b and c. By giving to a, b, c any particular
values, as for instance unity, the value of k is seen to be unity.
34. As in 33. shew that a, b, c, a + b + c are factors and
deduce the numerical coefficient by giving particular values to
a, b, c.

CHAPTER VI.

1. (1) 5a²bc ; (2) 12xyz². 2. 6abc. 3. x- 3 .

4. x- 1 . 5. x- 2 . 6. x² + x- 6.
7. 3x² - 5xy+ 2y². 8. x- a.

9. x -x - 1 . 10. x- 1 . 11. x²-2x + 4.

12. x- 1 . 13. (1) x -y; (2) x -y; (3) x² -y².
L12
516 Answers to Examples.
14. x+b. 15. 3a²- 2c². 16. (a-b) (b -c) (c -a).
17. x (a² -x²). 18. x (x² -y² ).
19. (x - 1) (x - 2) (x - 3). 20. (x² + 1) (x + 1)² (x² + 66).
21. ху² (x -y). 22. (x + 1) (x + 2) (x + 3).
23. (x-a) (x -b) (х- с).
24. (9x² - 13x²y² + 4y*) (5x + 2y).
25. (x + 1) (x + 3) (x - 10).
26.2x- 3 ; (2x- 3 ) (3x² + 4x + 5) (x² - 2x - 1 ).
27. (3x² - 2xy + y²) ;
(2x² + ху-у²) (9x - 3x²y- 2x²y² + 3xy -y').
28. (x - 1 ) (x - 2) (x - 3) (x- 4). 29. x² + 2x² + 3.
30. x² -ax + a² ; (x² + ax + a²) (x− 2a) (x + a ).
31. (x-a) (x -2 ) (x- 4 ) (3x- 7a).
32. (x² - 1) (x- 2) (x + 3) (x + 4).

CHAPTER VII .

x- 2 x-a
a(x-a)
1. 2. 3.
x- 3 a x² -ax + a²
x² + 2x- 3
4.
x²- 2x-3
Б. 3x² + 5xy+ 2y²
5x+ 2y
x² + 2x+ 4
6. x² + (3p + 1)x+ 3p + 1 . 7.
px²+px- 3-8p x+ 2
1 x² + 2x + 4
9.
8. a+ b+ c x² + 3x+ 2

10.
2(x² + 3) 11. x (x²+ 2x-3)
3(x² + x-1) x²- 7x+ 22
a² + b² + c² a+ b
12 .
bc+ ca + ab
13.
a-b
14. 2(a-b).
a+ b
2x- 1 2x²-8x+ 7
15. 16.
2x+ 1 (x- 1) (x - 2)(x - 3)
 Answers to Examples. 517

3x² - 17x + 20 7x - 39x + 44

17.
(x - 1) (x - 2) (x - 3) 18. (x- 1)(x-2)(x-3)
19.
Y Зах + 7a²
20.
x+ y x³- 8a³
2ab
21. 22. 0.
(a+ b) (a³-b³)
23.
(x +y³)
xy(x -y)²
2
24. (х- у) (x² -ху + y²).
a² + b² (x+y)²
25. •

26.
a
ху(х-у)
2ху 2
27. 28.- . 29. 1.
x²-ху +y² a

a²+b²
30. 31. 4x {(a + b + c) x + bc + ca + ab }
2ab
(x-a) (x - b) (x-c)
3 2
32. 33. 3
abc (a + b +c)
3x-5
34.
7x- 11 (see examples 17 and 18). 35. 2-x²

6
36.0. 37. a +b + c. 38. (x + 11) (x + 12) ;
( +11) ( +12)
39. 2. 40. x(8-8x+ x²)
16-20x+ 5x²

CHAPTER VIII.

1. a². 2. 1. 3. ac 4. disc .
5. 2x + 4x - 12x - 19 + 12x + 6x - 3x-
6. x + 7x-4-64. 7. a + a - b -b .
8. a + ab + b ; a² + ab + b². 9. x + xy + y .
10. 16x-3-12x- y + 9y- .
11. x -x + x -x + 1 ; x -x + 1 .
518 Answers to Examples.

12. x + y -z . 13. z + 2xy -x - у.

14. xy +1 . 15. x²-α².
1 .
16. (x-1) (x- 2) (x -3) . 17.-
х

2a

18. 어
b
20. (a + b) .

CHAPTER IX.

1.4√2 ; 22; 3√2; 22. 2.2/6.

3

3. 4142 ; 33-9704. 4. 1812

5. 4(√3- √2); √2(1 + √2 + √3).

в. 2-3. 2
7. √3+√22

29
8.
2
Simplify bymultiplying together the binomials which
involve the same surd (Article 261).
9. √3+ √6-2√2-1 . 10. 2√6+ 5.
2/6-5
11. The sum and product each equal
√3
√a² + 362
12.
b
1
13. (1) √3+ √2; (2) 2 /7-3√3; (3) √3 + √ ; (4) 1+
a+b a- b
14. (1 ) √ ++ 2 2. ;
(2) /(a+c)(b + d) + (a−c)(b-d).
15. ( 5-1) (√3-
4
√2)
 Answers to Examples. 519.

1-n + n²
16. 1 + n + n² +
2 2

17. See Articles 276, 277.

1+ i
18. can be shewn to denote a rotation through half a
√2
right angle.
19. See Articles 283 and 121 .

CHAPTER X.

2. 1. 3.2.. 4. 2.
1. 1.
6. 2. 7.5. 8.2.
5. 2.
10. 1 . 11. 1. 12. 1.
9. 1.
15. 6. 16. 3.
13. 2. 14. 31.
11 15
17. 2 (a²+b²).
ab
18.
34
19.
13

21. 1. 22. 11 . 23. -b.

20. 2.

24. 8abc-(b+
a²+
c)(c+ a) (a+b). 25. 4.
b² + c²-bc - ca - ab
2ab
27. 4. 29.2.
26.. 28. a-b
30.
(a-b)c
ab
31. The two places ofmeeting are 12 and 6 miles from A.
32. 48 of each kind. 33. £10.

34. Seven gallons out of each.

10
8 37.
35. 32
,
past noon. 36. 12, 6. 15
11

38.60. 40. £6 to boys, £4 to girls.

39.60.
41. £20,000 in the Three per Cents., 15,000 at 4 per cent.
 520 Answers to Examples.

CHAPTER XI .

1. x = 2, y = 1 . 2. 1 , 2 . 3. 3, 2.

4. 12, 4. 5. 1, 2. 6. 12, 21 .
7. a, b. 8. ac, bc. 9. 1, 1.

10. a + b, a - b . 11. 2, 1 . 12. 3, 2.

13. 19, 17.

11
14. Take- , as the unknowns, the value of x is a and that
ху
of y is b.
15. x = 1 , y = 1 . 16. a, b. 17. a - b, b - а.
C C
18.
a+b' a+ b
1 1
19 Solve for and ; the values of these are found
2 -x 1+y
to be 1 and , whence x = 1, y = 2 .
20. x = 3, y = 1 , z = 2 . 21. 2, 3, 4.
22. 1, 2, 3. 23. a, b, c.
b²+ c²- a² c² + a² -b² a² + b² -c²
24. , ,

2bc 2 ca 2ab

25. a, b, c. 26. a, b, c. 27. 2, 5, 3.

28. z =
g(cd-ab)
abc + abd + abe + abf + acd + ace + adf + aej + bed + bcf + bde + bef+ cde + cdf + cef+ def
5 3
29. 43 miles. 30. ,

2016
31. A in 18 days, B in 27, C in 36.
14
32. 36 . 33 . 34. 39, 5.
17

35. A in 7 hrs., B in 15 hrs. 36. 243 .

 Answers to Examples. 521

CHAPTER XII.

1. x =
cf-bg ag-bf
ac-62 Y = ac-ba ; if ac-b² = 0 the equations are
inconsistent unless ag- bf = 0 also ; when it will easily follow
that of- bg = 0, and the equations are identical since
a
b_f
==

2. (1 ) x = 2, y = 1 ; (2) x = 3, y = 2.
3. wu - bv, bc -u², cw- uv.
4. abc + 2uvw - au² -bv² -cw².
5.3abc- a³ -b³ -c³.
6. z (3xyz- x -13-23).
7. Use Article 352 (a) : the result is
a, b, c,
2 b, c, a or 6abc- 2a³ - 26-2c .
c, a, b
8. x = 3, y = 1 , z = 2.
9. The third equation is derivable from the other two.
10. x = y = z = a² + b² + c² - bc- ca - ab.
11. x = a, y = b, z = c.
12. x = (1-a)(c-a)
{12-1(b +c)+m} and symmetrical values for
(a -b)
yand z.
m+ n n+ l 1 +m
13. x y=
2 2 ,≈= 2

CHAPTER XIII.

1. 2x+ 5y. 2. 3a- 2b. 3. x² - 2 .

4.7x- 3y. 5. x² - 2xy+y².
6. 2a²- 3ab + 562. 7. 4x² + 3x²y- 2y .
8. 2x²-xy² + 3y²³. 9. x²- 6x² + 12x-8 .
522 Answers to Examples.
15 3 11 2
10. 11. +1+ .
16 x²+ ху- 12У².
4
Y х

12. X
++
2

13. √2- √ - √ . 14. 13-21 ; 1653; 301-07.

3
15.x- 2 + - 16. ac +bd.
х

17. ab-1 + bc-1 + ca-1. 18. 4a² + 2a (b + c) + bc.

19. x + x - 1 . 20. x²- 1 .
x+ 1
21. 1 -x + 3x . 22.
x +2
23. 1.7305080 . 24. 4.898979 .

CHAPTER XIV.

1. 1, 16. 2. 3 , 9 . 3. 2, 4.
1 1
4. 2, -1 . 5.1,
5 6. 1,7
45 11 7
8. 9. 3, 1 .
7.4 3' 4

9 ab 4
11. 12. 3,
3,
10. 3,2 δ'α 3

13. 2, 2. 14. 1, 5. 15. 5, -11 .

16. c + √(c-a) (c -b). 17.1 , -4.

2ab
18. 0, a + b . 19. a + b,
a+ b
cd- ab
20. 0, + √ab. 21. 0, a + b -c-d
4a
22. +1 , +2. the former not admissible.
23.0,5 ,
3 7 3
24. √²- 2x- 3 = -1 or the latter gives x = 2'
2' 2
 Answers to Examples. 523

25. 0, a (1+2√2) where i = -1.

3

26. 3 . 27. -4, -7.

28. Square, and assume x -x -x² + 1 = y, it easily follows

that y= 1, whence x = 0 or 1+√5
2
5a
29. •

30. Divide by x², the equation can be written

( +1) 35
35 x² + 1
х
-1 = 0,
6 х
whence
x²+ 1
х
= 6 or
1
6
whence x can be found.
31. Assume x² - 9x = y, the equation can be written
(y+ 18) (y + 20) = 24 ;
whence y= -14 or -24, the former gives x = 2 or 7, the
latter imaginary values.
243
32.0, 33. 0, 0.
193

34. x (x + 1) +1 = 0, or x (x + 1) +2 = 0 .
7 x+ a 1
35. 36. = 3 or whence x = 2a or - 2a.
4 х- а 3

37. The two expressions must have a common measure. This

C

can only be 2x-c, whence the common root is 5,and by

substitution it follows that e must = 0 or 10. If c = 0 the
roots of the first are 0, 7, and of the second 0, 9. If c = 10
the roots of the first are 2, 5 and of the second 4, 5.
38. By working out the H. C. F. of the two expressions and
making the last remainder vanish, the condition becomes
(br- qc) (aq- bp) = (ar-pc)².
39. Solve as a quadratic in y.
40. qx² + px + 1 = 0, x²- (p² - 2q) x + q² = 0 ;
x²-(p² -p- 2q) x + (q² + q -p³ + 3pq) = 0 .
524 Answers to Examples.

41. By division,
xa x²β x²y
y = aa + ax + + bβ + βx +
a -x β-x + cy+ x+ y
β γ
= aa + b + cy + x + x² + + );
β-у y- 2

whence substituting x₁ for x, y₁ = aa + b + cy + x₁ ; but from

the given equation x1 + x2 = a (b + c) + β (c + a) + (a + b),whence
the result will follow.

42. a = -
P p²-4q whence the result will easily
=- , B = 4
,

follow.
b =
a
43.
q = ab, -P₁ = a + b, -p₂ = ma + , -P3 + mb.
m m

Hence P₂ +p₃ = -(m+ 1)(a+ b) = -(m + )P1

1 1
Also pap₃ = a² + b² + ma)ab = pi²-2q+
+(m² + m². (m² + 2) ;
2

=
1
and since
(m+ ) m² ++
m
2, the result follows.

44. See examples 37, 38. 45. 121 .

46. 6 and 8 for a shilling ; or 8 and 10 for a shilling.

47. 4, 7. 48. 54 yrs. 10 yrs. 49. 55 by 88 yards.
50. He goes at 4 miles perhour, reaches B at 11 a.m.; returns
at 3 miles per hour : distance is 12 miles.

CHAPTER XV .

1. 3x+ 5y. 2. 4a-b. 3.2x -x²- 3.

4.2 x³ - 3x² + 4. 5.2x -x - x + 1 .
6. x-2y+ z. 7. 1234; 4057 ; 637.
8. x = -2 or 1 + i. 9. x = 1 or 1 + i√2.
10. x = -2 or +√3 .
 Answers to Examples. 525

11. The expression on the left-hand side of the first equation

must be identical with (x-2) (x-3), while that in the other
equation must be identical with (x- a) (x- 3)².
12. The expression x³ + 3px² + 3qx + r must be identical with
(х-а) (x - b) (x − c) ;
whence a + b + c = -3p, bc + ca + ab = 3q, abc = -r ;
applying the formulae of Art. 414, and referring to Art. 282
and the value of a given in Art. 277, the result is easily
arrived at.

13. Work out the highest common divisor of the two ex-
pressions and equate the last remainder to zero. If the equa-
tions have a common root, the expressions must have a common
factor (Art. 454), or see Art. 473.
14. It is easily seen that X+ Y+ Z = (a + b + c) (x + y + z)
and then use the result of Art. 282.

CHAPTER XVI.

1. x = 2, y = 5 ; x = −2, y = -5 ; x = 34, y = −11 ;

x = -34, y = 11 .
2. x = 3 , y = 4 ; x = 4, y = 3.
28 55
3. 2, 1; 13' 26
4.3,1 ; -3, -1 . 5. 3, 1 ; -3 , -1 .
-3 + √37
6. x = y = 2
; or x = 1 , y = 2 ; x = 2, y = 1 .
56
7. x = 3, y = 2 ; x = -3, y = -2 ; x = 22, y = - ;
56
x = -22, у= 3

8. x = 2, y = 6 ; x = 5, y = 3.
9. x = ± 7, y = +4 ; x = ± 4, y = ± 7.
8 1

6' y = ± √
10. x = ± 3, y = ± 1 ; x = ± √6
526 Answers to Examples.
11. x = 8 , y = 4 ; x = 4, y = 8 .
12. x = 12 , y = 10 ; x = -9, у = -11 .
1 1 4 3
13. x = , x = - -

2 =3 3 Y= 2
1 2 1
14. x = 2, y = 2 ; x = 3Y =,

6
cd + c²d² - 4abcd C
15. x = -, y = dx
2ad

a + √262 - a² a+√26²- a²
16. x = , y=
2 2

17. x = 16, y = 4 ; x = 4, y = -16 ; x = 6 + 3√13,

y = -6313.
a

18. x = y = 19. x = 4 , y = 5 ; x = 5 , y = 4 .
24
20. x = 3 , y = 4 ; x = 4, y = 3, and imaginary values.
21. x = 9 , у = 3 ; x = 3 , у = 9 .
2abc b²+ c²
22. x = 23. x = ,
&c.
ab + ac-bc √2 (a² + b² +c²)
19 3
24. x = -2 , y= 7 11 , y = 0.
25. х = 147, у = 140 .
a²b + a²c² -b²²
26. x = ,
&c.
2abc
a2
27. x = ab, y = a¯¹b³, z = b
28. x = ± 2, y = ± 4, z = ± 7 .
29. x =
a (a +b) b(a+b)
, y=
C C

Y √bc + ca + ab
30. =

a b C
√a² + b² + c²
х-а y-b 2- C
31. = = = 0
a (b -c) b (c -a) c(a −b)
or 2 a² (c -b) + b² (a -c) + c² (b-a)
a² (c - b)² + b² (a -c)² + c² (b-a)2
 Answers to Examples. 527

2 5
32. x = 1, y= 1 ; x = 7 , Y = 7; put y = vx.

х y
33.
3bc + 2ca + ab + b² - c² 3 ca + 2ab + bc + c² -а²
2
=

3ab + 2bc + ca + a² -b² '

a+ b + c
and each of these factions = 0 or 18(bc+ca+ab)
х 2
y
34.
b+ c-a_c + a - b_a + b - c
2 (a² + b² + c²)
= or 0.
3 (a² + b² + c²)-2 (bc + ca + ab)
35. x, y, z must have the values ± (b − c), ± (c - a), ± (a − b).
36. 7a³ = 3a (b -c)².
37. The values of x, y, z are 16, 16, 8 or 10, 10, 20 in any
orders.
38. b (a* b*)² = 2a* (b −c ).
39.2 (b ± √36 -c²) = (a² + √3a² -b²)².
1-a -1
40. ( ) = ( = ) 1

41. 36, 16 ; 24, 4. 42. 132 ; 213.

43. 900 men in 100 ranks, or 20 men walking in single file.
44. 300 yards; A runs six yards for B's five.
45. 1, 2, 3 . 46. A takes 11 days ; B22 ; С 33 .
47. 10,000 at 5 per cent.; 7500 at 4 per cent.; 2500 at 6
per cent.
48. 16, 18, 20.
49. A in 6 hrs., Bin 3 hrs., Cin 2 hrs.

CHAPTER XVII.

1.26.25.24. 2. 14.13.12.
13
3. 4. 72.
310
528 Answers to Examples.
100 99
5. 6. 72, 24.
1288 ' 11-88

7. See Arts. 496, 497. 8. n 6.

9. n = 8 . 10. n = 5.

12. 5, or if right-handed and left-handed arrangements are

5
counted as identical, 2

13. n - 1 orn - 1 according to the interpretation of the

word different arrangements.
|m + n
14.
mn
m - 1 | n - 1 , or one quarter of this number.

16 8 4550
15. X 16 .
33 44 91036 40

11
17.
( 2)
11 10 11
18. ;
( 4) 2 (12)5 (12)2.14
21 15
19. 15 ; 21.20.19.5.4.
318 32

65
22 . 24. n = 11 , r = 7 .
(12)33

CHAPTER XVIII.
2
5- n 9n-n²
1. ;
8
2. 2n- 1 , n².
4

5+n 11n + n²
3.3+n 7n+n²
2
,
4
4.
4
,
4

5.13-2n, 12n + n² . 6. 29 +n 59n + n²

2
,

4
 Answers to Examples. 529

9n-n² 3 -n 5n- n²
7.5 - n, 2
8. ,
2 4

n²-n
9. n - 1 ,
2

10. a² + b² - 2 (n - 2) ab, n {a² + b²-(n -3)ab}.

n²-n
11. 0, 2 "
12. a (2 -n) + (n - 1) 6 n { a (3 - n) + (n - 1) 6 }
,
a-b
2 (a-b)
13. a-(n-2) 6 n {2a-(n-3)6} ,
a²-b² 2 (a²-b²)
14. 17-4n, 15n- 2n² .
3n - 13 3n²- 23n
15.
2 , 4

25-39 25n² - 53n

16. ,
6 12

17. 13 . 18. 2n - 1 , n².

n

19. 5f- 4e, {(n - 5)f+ (9 - n) e} .

1 4 2
20. 21. 22.
n 5- n 3 +n

4 2 1
23. 25.
7-3n 24.3 10

26. 1 +++
2 3
...,or 1 + 3-2 ... 27. 2 or 10.

28. N(p -m) + M (n-p). 29. Solve for z.

n -m

11 1
30.
2'318
а- ха -у a - z
31. It easily follows that х
,
2
are in A. P.
y
Mm
LIBPA
OF THE
UNIVERSIT
530 Answers to Examples.

32. (2n + 1)², 6. 33.0 + 1 + 2 + 3 + ....

35. 280 days or -229 . The negative answer implies that

229 days previous to starting they would have been together had
the arithmetic progression been continued backwards and A's
motion consequently been westward.

CHAPTER XIX.

1 1023 (-1) -1 14762

1. 2. ,
2n-1' 512 3-1 19683

1 3-1 11605
3.3(- ) ;
1536
4. 2-1, 1023.

6-1 610-510 2

5.5-2 ,
58
6.
1-(- )} ; 2.
2 3་
4 1" 4
7
.--
1

8.
3√3( √3 + 1) (1-1), 3√3(√3 + 1) .
2
32
2

9.1+2√31-(
√3
3-1)}, 4+2√3
√3
5
10. 6
{ 1 - ( - 5)"} , 11. 121-1 , 12.
2
3 7
12. (1- ); a-b ifb < a. 13.
7550
,

14. 1 ++
5
.... 15. 10 terms .
25

18. 6. 19. r = 2 , n = 3 .
22. The expression = Σ. 25. x 2.

26. 2, 6.
 Answers to Examples. 531

27. The condition gives (ac) , =b , and ac must= b since

a, b, c are positive integers, whence the result follows.
mn

30. £P(1+ ) ,
100r being the nominal rate per cent. per
annum.

31. 100 (1 +400) -1 } .

CHAPTER XX.

1. a² -7ax + 21ax² - 35a* x + 35ax* -21a²x + 7ax® - x² ;

32a5 + 240 a²b + 720ab² + 1080 a²b + 810ab* + 2435 ;
3 1 64
a²- 2ab + ab²- ab +
2 ab³ + 16
448x 672x² 560x³ 280x
2. 128 + + -

...
;
3 9 27 81
9
-20xy³ + 190xy --1140x³¹y
1140x³ y + 4845 x28y12 - ....
35 21 7 1
3. x² + 7x + 21x² + 35x + + ++ 3
0

a - 8a²b + 28ab² - 56a²b + 70a* b* -56a²b + ....

20 n
4.
515
X25y15 5.
rn- r
22-2

2n + 1 2n+ 1
6. x; х.
n- 1n + 2 nn + 1

2n 2n 1
7. x² ;
n- 1n + 1 n-1n + 1x²

8. (1 + 2)7. 9. 35 x 28 x 34, the 5th term.

10. 3 . 11. -48 .

12. (x + a)" = A + B and (x-a)" = A - B.

13. Use Art. 548 (3) and (4); prove 2m-" = 16, m = 6, n = 2.
Mm2
532 Answers to Examples.
n

14. By Art. 558 f(n) = an-rbr

rn- r

n+ 1
Similarly f(n + 1) = an-r+18,
rn - r + 1

whence f(n + 1) n+ 1 f(n) n

f(n)
=

n- r + 1
⚫a; similarly f(n- 1) =

n- r
a,

whence the result can be easily deduced.

15. Use Arts. 492 (1), and 558.
16. See Art. 559 : the coefficient of x in
m +n
(a+ x)m+n r m+ n- r

17. Use Art. 559. In (4) n and r must be both even or both
odd, otherwise the series = 0 .
18. Compare Art. 543 and Example 12.
n

19. (1 + x + x )" = Σ -x²+ 2r, ( 1) if p + q + r = n ;

PIT
n ( -1 )
(1 − x + x²)* = Σ x '+2 ', (2) if p +q + r' = n ;
PIr
n

(1 + x² + x*)*= M x2q"+4r ", (3) if p" +q"+ r" = n.

P" q" r"
Also (1 + x² + x*) = ( 1 + x + x²) (1 -x + x²) ; and equate co-
efficients of any power of x in (3) to coefficient of same power
in product of (1) and (2). For second part take (1-3x + 2x²)*
and put x = 1. Compare Art. 548 (2).
bx b
20. aa ++ bx
bx ++ ccx² = a (1 + b + c ) = (1 ++ y²)
a a
ac

if y= xN
whence the result follows .
;
 Answers to Examples. 533

21. Let (2 + √3)"= I+F where I is an integer and Fa

1

proper fraction. Then, since 2 -√3 = 2+ √3' (2-3)" must

be a proper fraction. Call this F. Hence
(2 + 3) + (2 - √3)" = I+F+F'.
But it is easily seen that the sum of the expansions on the left
is integral and even. Hence F+ F' must = 1 and I must be
an odd integer.
1
22. √a²+ 1 - a = and solve as 21.
√a²+ 1 +a

CHAPTER XXI .

3 5
1.1 ---1 ; 1- 2 +
2 8 16 - 16 ;
33
211
x; 429
211
.
0 7
2.1 2
9 81 20.
7x 119 21.17.13.9.5.1.3
3.1+ + x² + ... ;
2 24 477

4. Ther + 1th terms are

1.1.3.5 ... (2-3) 1.3.5 ... (2-1 )
-x ; x;
r r
(3+4)(1+ 4)
(r + 1) (r + 2) (r + 1 ) ( +2) (r + 3) (r + 4) .
x;
1.2

5. The 8th term. 6. (r + 1 )2.x .

7.
(1-x)
= ( 1-3x + 3x²- x³) ( 1 + x) ². Hence the coeffi-
(1 +x)²
cient required = 1 x coefficient of x in (1 + x)-2-3 x coefficient
of x in (1 + x)-2 + 3x coefficient of x² in (1 +x)-2-1 x coeffi-
cient of x in (1 + x) ² = 5 + 12 + 9 + 2 = 28.
534 Answers to Examples.
1
8. √24= 5 (1 25
= 4.89898 ;

5
31 = 2(1-2) = 1.9873.
9. 4-1231 . 10.1 + x + x² + ....

12. The series is the coefficient of x obtained by picking out

the term involving x in each of the binomials on the right-
hand.

13. Use the identity

(1 + x + x²)" = { 1 + x (1 + x) }"
n(n - 1)
= 1 + nx ( 1 + x) + 1.2
x² (1 + x)² + ... ,
then (r) is the coefficient of x : the final result is obtained
by putting x = 1 .
a2 4

14. 1 + x + (3x - 1)+ (5x -3x) + a*( - 12x² + x*).

15. 33x . 16. 0.

17. The given expression = ( 3a - 2x) (27a³ - 8x3)-1.

875630
18.
6561
X -72

19. (1 + x)2 = (1 + x)" (1 1+ x

|n- 1
20. Coefficient of y -1 in (1 + y)"-1 is ; coefficient
x- 1| n -x
n- 1
of y" -2-1 is xn -x- 1
; the sum of the products of these for

all values of æx from 1 to n- 1 is the coefficient of y"-2 in

2n- 2 n- 1 2n
(1 + y)2n-2, or = n- 2n
•

; whence
2 (2n- 1) nn

n- 1 2n
2Σ ( n- 1)2 n- 1
1 ,
xn -x - 1x - 1n -x 2 (2n - 1) nn

whence the result follows.

21. √2 = (1-2)
 Answers to Examples. 535

CHAPTER XXII .
4 1
1.
1-2x 1-x

2.3 + 7x + 15x² + 31x3 + 63x² + ....

1 21 8
3. +
22 (1
( 1 --x)
x) 2 (1-3x) 1-2x

1 1
4. 151 + + where w is one of the cube
31 -x 1 -ωχ ' 1- ω³ )
roots of unity (Art. 277). The last two fractions can bejoined
so as to give a real form, and the result becomes
1 2+x
3 (1 - x) * 3 ( 1 + x + x²)
4(1-2" x") 1 -x"
5. (2 +1-1) x -1 ; 1-2x 1-x

6. an- 2an-1 + an-2 = 0,

a- (a - b) x-(a + nb) x² + (a + n - 16) xn+1
(1 -x)²
7. an- 3an-1 + 3an -2 - an -3 = 0 ;
n
a2 - x
-b2)+(a − b)²x2-x (a+ b)²+ xn+1 { 2a² +( n-2) ab+ (2n2-2n-1)62} -xn+2(a+n-16)
-(1- x)3
For the second part put a = b = 1 and x = 1 -h, and
expand all the terms in the numerator by the Binomial theorem
in powers of h. The terms involving lower powers of h than
h³ willbe found to vanish, while that involving h³ reduces to
n (n + 1)3(2n + 1) 3.
The denominator becomes h³; hence the fraction reduces to
n (n + 1) (2n + 1)
3
+ terms involving h,

and making x = 1, and therefore h = 0, the required sum is

obtained.
536 Answers to Examples.

CHAPTER XXIII .

1. (n + 1) (n + 2) (n + 3) -1.2.3
3

2. (2n + 1) (2n + 3) (2n 4

+ 5) (2n + 7) -1.3.5.7

3. n (n + 1) (n + 2) (n + 3) .
2

1
4. 151 1.
2 ( 2.3 (n + 2) (n + 3)
5.15 1
23.5 (2n+3)(2n+5)}
3n2 + 5n
6.1-2 +1 7.
4 (n + 1 ) (n + 2)

8. Un =
(n² - n + 2) (n² - n + 3).; S = n (3n² + 60
25n² + 92) . :
4
use Art. 610.
9. n (n + 16) (n + 2) 10. 1 + 2 *-1 ; n + 2 " -1 .

11. ( 1) A geometrical series ; (2) a series whose nth term is

of the form A'np − 1 + B'np - 2 + ... ;
(3) un = 1 + 2 + 2 r² + 35-1,

$n = n + n (n + 1) + " (n + 1) (2n + 1) + 35-1.

3n2
12. Un = 3n-1_1+(-1)"; 8 = 3² + 1-(-1)":
73
2 4 4 8
use Art . 611.

3n² + n 1 -(- 1)"

13.2 " -1 + 3n - 1 - ( - 1 )" ; 2 " -1+ 2
+
2

14. 15. n(n + 1) (4n + 5)

n+ 1 3
 Answers to Examples. 537

16.
n (n + 1 ) (n + 2) (n + 3) (n + 4)
5
n 2n² + 3n
17. 18.
2(3n+ 2) 5 ( 4n + 1) ( 4n + 5)
3n- 1 (2n + 1) n
20. นn =
(n + 1)(n + 2) (n + 3) = 3 (n + 2) (n + 3)
21.
DR" (R- 1)2
&D being the debt and £R the
Rn+1-(n + 1) R+ n
amount of £1 in one year.

CHAPTER XXIV.

1. 5. 2. 4. 3. 3, 3.
4. 7781513, 9030900, -9542426, 1.3802113, 1.5563026.
5.0484550 .

6. log10 12 = 2 log10 2+ log103, log10 15 = log10 3 +log10 5

and log10 5 + log10 2 = log10 10 = 1, whence the values of log10 2,
log103 and log10 5 can be found: the values are given in examples
4and 5.

7. 0-1505150 ; 1-3656367.
8.0-1890480, 9912260; 2.5440680.
3 6
10. 1-8186440.
9.25
11. 1-1672269 ; 1-9119544.
12. In the series for loge (1 + 2) put z = 1.
13. 2 (1 -log. 2).
1 1
14. e + -1 = 2 (1 +++...
24 ):
e-6-1 = 2 (1 +++
3 5 ...
).
1
15. Put for z in Art. 637.
X
538 Answers to Examples.
1
16. Put x² for z, in (3) Art. 637.
2

n+ 1
17. ae = b - 1 : use the series for loge n
in Arts. 6, 38,
putting n = ac.
18. The general term
2n + 3 2 3 1
+ ,

(2n - 1) .2n . (2n + 1) 2n- 1 2n 2n+ 1

the sum ofwhich gives 3 loge 2-1 .
19. The capital at end of 5 years can be shown to be
3
A
G (1 + r) ) ;
2 2
that at the end of 10 years
10
1 13
= 4 這 +18(1
A
10 + r)* (1+r) . 5

209
Equating this last to 200 A, a quadratic equation in (1 + x) is
13 log 13
obtained which gives (1 + x) = 10' whence log (1 + r) = 5
,

and 1 + r can be obtained from the given logarithms.

20. Take logarithms and use Art. 634 (5).
21. Art. 634 (5).
22. If loga a = logs b = ... = x ; a = a , b = ẞ", ... whence
r= p .

23. The harmonic mean of n quantities is the reciprocal of

their arithmetic mean (Art. 705).
25. Assume a = m", b = m", c = m² : the two conditions
X 2 2
Y
give x- y = y- z and 2
=
, whence eliminating x
y Y
we obtain 4y - 3y²z- 3 yz² + 23 = 0, whence
y= z or 4y² + yz - 22 = 0 ;
from the latter
2y 2 3 2 2
-- 1 and 2y- z
------- 3 or
2 2 2 2 2
Y Y y
 Answers to Examples. 539

CHAPTER XXV.

a C
1. and 2. Assume = x. 3. 16:25 ; 1 : 1.
b d

4. 9, 12. 5. 18, 24.

6. As (1) and (2). For examples (7) to (12) useArt. 659 (1)
and (2) : thus in (8),
y+ z 2+x x+y p (y + z) + q (z + x) + r (x + y)
= =
;
36- c 3c- a 3a - b - p ( 3b - c) + q (3c- a) + r (3a-b)'
x+ y+ z
and by taking p, q, r each unity this becomes a + b + c while
'
by taking p = b + c - a, q = c + a - b, r = a +b - c, it becomes
ax + by + cz
after reduction,
a²+ b²+ c² ; the difficulty consists in the proper
choice of the quantities, p, q, r in each case.
13. 2, 8, 32.
14. Eachfraction = (2x - y) + (5y - x) - (2у + x)
= y;
(x + 3y) + (х- 3у) -2 (x- 1)
3 3
x = 4, y = 1 , x= 2 Y = 2
55 15 45
15. Use Art. 659 (1); 2, 3, 5, 6 or 4
,
2
, 5,
4
16. Each of the fractions

=
yx (x + y - z) + xy (z + x − y) =
2xyz
ylogx+ x log y ylog x+ xlogy
and so on with the other possible pairs.
17. In ratio 36:35.

18. It is easy to prove that A, Hn-r = ab.

19. The conditions are equivalent to
x+ y = kz and x² + y² = kz² ;
whence x = Az, y = μz ; k, k', λ, μ being constants.
20. Rate of A : rate of B :: t : t + 2t- T.
540 Answers to Examples.
21. 7223.04. 22. 1375 cubic yards.
24. 10 miles an hour. 25. 30, 36, 45.
26. vol. of silver in first : to vol. of copper :: 57 : 7; in the
second:: 1 : 7.
5 55 55
27. 29. 8, 12, 14, 21 ; or 0, 0,
3 22

CHAPTER XXVI .

1 1 11 14 1 11
1. 1 , 2.
1+ 1+ 3+ 3' 2 2+ 2 + 3
11
3.3+
7+ 22

1 1 1 1 1 1 1 1 1
4.3+
6+ 1 + 5+ 1 + 5 + 1 + 1 + 1 + 10
1 1 1 11
5.1+
1+ 1+ 17+ 1+ 6
1 1 1 1 11
6.
2+ 1 + 50+ 2 + 2+ 2
1 1 1 11
7.
1 + 1 + 3 + 2 + 16
3 10 43 225 1393
,

8. 1,2' 7'30' 157 972

9. The 49th of the first with the 54th of the 2nd : the first
bell rings 108 times in 63 minutes, the second 119 times.
248
10. 27, days. 11. Use Art. 675.
771

12. Use Art. 671 (1): it can be shown that

N"-N= { (p + 1) N-N} ;
N' N"
I
', q" being the quotients corresponding to D' D'
 Answers to Examples. 541

13. If q be the value of each quotient, it follows that

Nr+1N -NN -₁ = q N²,
whence the result follows.

14. Prove by induction as in Art. 671 .

1 1 1 1
15.3+
3 + 6 + 3 + 6 + ...

1 1 1 1 577
16.4+ , 4th convergent and the
4+ 8 + 4 + 8 + ... 136 '
1 1
error lies between and
136 x 1121 136 x 1257

1 1 1 1 1 1 1 1 1
17.4+
2 + 1 + 3 + 1 + 2 + 8 + 2 + 1 + 3 + ...

√37-4
18. 3+√377
19.
3

20. The positive root of

(b + 1) x²- (ab + a + b - 1) x- (a + 1 ) = 0.
21. The second fraction is the positive root of
(a + 1) y²- (ab + a + b - 1) y- (b + 1 ) = 0,
the roots of which are the reciprocals with the signs changed of
those of the former equation (see Arts. 414, 424), whence the
result follows.

11111 27 9 25 84
22 .
1 + 2 + 3 + 1 + 2 + ... 1,3' 10' 13' 36' 121
,

1 1 322391
23.3+
10+ 10 + ...' 6th convergent is 104030
;

1 1 1 1 1 1 4389
6th convergent
1 + 9 + 2 + 10 + 2 + 10 + ... ' 4830

24. The equation gives

1 1
x = y+
y+x'
whence y = x-x+y
542 Answers to Examples.

25. By means of the relation ofArt. 671 it is easily proved

that for any three consecutive even convergents
N2n-(ab + 2) N2n-2 + N2-4 = 0 ,
and for any three consecutive odd ones
N2n +1-(ab + 2) N2n-1 + Non-3 = 0 ,
similar relations holding for the denominators. Hence, by
Art. 593, it follows that each of the quantities N2n-1, Nan 2

Dan-19 D2n can be expressed in the form Ay₁" + By₂", where

A, B are constants and y₁, y₂ the roots of the equation
y²-(ab + 2) y + 1 = 0 .
The constants for each can be determined from the first con-
vergents and then the required results will follow.
26. Work as in Art. 678 .

27. Pn = 2apn -1+P - 2, whence, by Art. 593,

Pn = Ay₁ + By2",
where y₁ and y₂ are the roots of the equation y² - 2ay - 1 = 0 .
28. Use examples 14 and prove by induction.

CHAPTER XXVII.

1. x = 3 , y = 2 . 2. x = 7 + 13t, y = 4 + 21 t.
3. x = 7t, y = 8-8t. 4. x = 4, y = 6 ; x = 11 , y = 1 .
5. x = 5, y = 3 . 6. Five solutions.

7. m = 75 if zero solutions are allowed, otherwise m = 83 .

8. Seven solutions . 9. x = 99-11t, y = 1 + 10t.
10. He must pay 3 guineas and receive 21 half-crowns .
11. If x be the number of florins A pays to B and y the
number of half-crowns, while z and u are the numbers of
shillings and half-crowns B pays to A, the equation is
4х + 5 (у - и) -25 = 25 .
 Answers to Examples. 543

Here y- u is merely the balance of half-crowns that passes from

A to B, and may be treated as one quantity : the equation
thus becomes 4x + 5v - 2z = 25, which can be treated as in
Art. 696, itbeing borneinmind that the limits of x are 0 and 15,
of y- u, or v, - 34 and + 8, and ofz, 0 and 5.
Then 4x- 2z = 5 (5 - v). Hence 5 - v must be even and
can only have values between -3 and 9. Hence 5-0
may =-2, 0, 2, 4, 6, 8, and we have the series of equations
2x- z = -5, 0, 5, 10, 15, 20, the number of admissible solutions
of which can be ascertained ; total number 16.
12. 24 oz.; 11 sovs. 79 gs.; 32 sovs. 59 gs.; 53 sovs. 39 gs.;
or 74 sovs. 19 gs.
13. He was born in 1741. 14. 37 × 2 or 14 x 3 .

15. Depends on the solution of the equation

979x- 700y + 4z = 0,
whence x = 4t : results are, 71st page of 5th and 7th volumes,
38th of 13th and 18th, &c.
16. In 14 different ways: it is the number of solutions of
3x + 2y + z = 16 consistent with x + y + z not being greater
than 9 .

17. 52 . 18. 4, 2, 1 or 8, 4, 2.
19. 292 . 20.63t - 2 . 21. 78, 65, 50.

CHAPTER XXVIII .

1. Use Art. 705. 2. Art. 703.

a2 n
3. 4. : use Art. 705.
4 n+ 1

(3m)² (3m + 1)²

5.
4
= 13 +23 + ... + (3m)³; use Art. 705.
2

6. 1+ > * , 1+x* > .

2 2
x .
544 Answers to Examples.

(x + y-z) + (x-y+ z
7.
2
, or x, > √(x + y- z) (x-y + z).
24 n -1
8. (1) 2-1 = 1 + r² + r + ... + ran-2, use Art. 705.
-1
(2) r- 1 = 1 + r + r² + ... + r"-1, use Art. 705.

9. x² + y² > 2xy, y² + z² > 2yz, z² + x² > 2zx.

10. The former.

11. The first fraction > 1, the second is less than unity since
(a+ b −c) > 0.
12. (n + 1) (n + 2) ... (2n) ... < (2n)".
13. (2x + a) √a-x = √(2x + a) (2x + a) ( 4a - 4x).
14. A.M. of p quantities = x²- , and qquantities = x- , and
r quantities = x -q > G. M.
15. Use Art. 705. 16. 44 millions; 38. 9d. per lb.
х 2 n
xyz...
17. + + ... <
C abc ...

18. (x² + y² + z²) (a² + b² + c²)

= (ax + by + cz)² + (ay - bx)² + (az -cx)² + (cy-bz)².
19. x = -2-4 gives a minimum value, x = √2-4 a
maximum.

20. Equate the given expression to y and solve as a quadratic

equation in x.
1 1
1 + 2 + 3 + ... + 13 +23 + ... + n³
21.
n n
>( n)= {( n) } > ( n) .

CHAPTER XXIX.

1. 33340, 3185. 2. 644-112. 3. 3412-3630.

4. 588, 1114. 5. 1006012 ; 1003, 346.
6. 17241623 ; 2643, 4371 . 7. 241-43, 14146.
 Answers to Examples. 545

8. The number is divisible by 1001 which = 7 x 11 x 13.

9. 31 43 in scale seven = 34 in scale 9 .

10. The difference of two numbers having the same digits in

scaler is divisible by r - 1. Hence the sum of the digits in this
difference is divisible by r - 1 . Hencer - 1 is 5 or 25, orris
6 or 26 .

11. 10 = 7 + 3, 10² = 7 x14 + 2, 10° = 7 × 143-1 , and so

on, whence the result will follow.
12. The number is divisible by 11 if a - a₁ + a -a + ... = 0,
or a multiple of 11 (Art. 718), or with the notation of the last
question if S₁ + S - S₂ = 0. It is divisible by 7 if
S₁ + 3S + 2S = 0 .
Ifboth conditions hold good
S = S
,1
S₂ =
2
-

4 5

These conditions are sufficient but not necessary.

13. 6. 14. 4.

15. The last digit must be unity since the number is odd :
the others may be 0, 1 or 2 : since the number has six digits
when expressed in scale 3, it lies between 243 and 729. Hence
the first digit when expressed in the scale of 12 must be 2 : the
number is 289 = 201 in scale 12 = 101201 in scale 3.

16. 6. 17. See Art. 718.

18. 1000 is divisible by 8.

19. Every number is of one of the forms 3n, 3n + 1 , or

3n + 2 ; every square number is therefore of one of the forms
3m, 3m + 1or3m + 4, which last is 3 (m + 1) +1. Every prime
greater than 2 is of the form 2n + 1 . Hence we have to prove
(2n + 1)*- 1 or 4n (n + 1) divisible by 8, which is evidently
the case.
Nn
 546 Answers to Examples.

21. q being the last digit q -q = M.2n .

23. The expression
= n (n - 1) (7n + 4) = 3n² (n - 1 ) + 4n (n - 1) (n + 1 ).
24. By Fermat m²- 1 is a multiple of 5 and m²- 1 of 3.
(n + 2) (n + 3) ... 2n the numerator
25. The expression = n
,

contains n - 1 and nas factors.

26. If m = 5n, the expression

= 5n (25n² + 89) = 125n (n - 1) + 570n.
27. Use Arts. 742-745. 28. 18.

29. 30 . 30. 198.

ap-1-1 Mp
31. 1 + a + a² + ... + ap- = =

a- 1 a- 1

33. If Nbe a number prime to n and N' a number prime

to m, then
Nn-1-1 = qn, N'm-1-1 = q'm, thus (N -1-1)(Nm-1-1 ) = qq'mn,
whence the result follows. 32 is done similarly.
2n- 1 (n + 1) (n + 2) ... (n + n- 1)
34.
n- 1n
=qn + 1, whence
1.2.3 ... (n - 1)
the expression
= (m²n -1 -m") +m" -m +(m - 1) -( 1 + qn) (m - 1),
and is a multiple of n.
35 and 36. If m be any such prime number, m³- 1 is
divisible by 3 by Fermat, whence both results can be obtained.
37. The square root must have 6 for its last digit and thus
be of the form 12m + 6, whence by squaring and cubing the
required results are obtained.
 Answers to Examples. 547

CHAPTER XXX.

4 1 1
1. 2. ;
7 35' 35

( 13)* 48.44.40 36.22.10

3. (1) 17.25.49) (2) 51.50.49 ; (3) 51.50.49

4. The number of ways in which 10 can be thrown with

three dice is the coefficient of x¹º in (x + x² + x + x* + x + x®) ,
since it is the number of ways in which the number 10 can be
obtained by adding one out of each of three sets of the numbers
1, 2, 3, 4, 5, 6. Either by the method of Art. 551, or by putting
3

the polynomial in the form x (1 )= x (1 -x )* (1 - x) ,

and expanding after the manner of Example 7 of Chapter XXI
this coefficient is easily found to be 27. Hence the chance
27 1
required is 216 8
133 132 x< 12 13 × 122 128
5.
13-124 134-124; 134-124 13-124
545
6.
(36)
206

7. (36) ; for this question and the last seesolution of (3). 63

8. The chance that both have told the truth is ; chance
100
3
thatB tells a lie is 10 in which case, as there is no information,
1
A's chance of winning is 2 Hence A's whole chance of win-
63 3 78
ning is 100 + 20 100
,
thus the value of his expectation is
£780 .
4
9.
21; the condition that a number is divisible by 11 is
given in Art. 718.
Nn2
548 Answers to Examples.
3 3 1

11. 4 14 28 The value of the expectation in shillings is

the sum of the terms
6 5 652
( ×20 + ) + ( × 20 ++ 8 ( × 20 + )
65 365432
+ 20 ++ 87 65 ×20 + )
6 5 4 3 2/2
22 6 5 4 3 2 1
+ 20 +1) + 20,
7

which reduces to 2218 shillings.

mn

12. The total number of distinct distributions is If

( n)m
the two balls be called A and B and A be put into any one
box, the number of arrangements which put B into the same
mn
box is Hence the chance of their being
n- 2 | mn - n + 2

m ( n)m
together is n - 2 | mn - n + 2

13. If C1 , C2 , ... C be coefficients of x, x², x" in (1 + x)" the

...

chance required, n being the number of balls of each colour

2
c² + c² + ... + c
=

22-1

But 1 + c² + c² + ... + c² is easily seen to be the term inde-

13

(1 + x)2n
pendent of x in ( 1 + x)" (1+ ) xn
or Hence

2n
c²+ c² + ... + c² = 1.
nn

2n
1
nn

The chance required therefore = 22-1

( 39) 4( 39)
14. (1) 26 52'; (2) 26 52
 Answers to Examples. 549

1 2
15. 16.
6 9

3 1
17. (1) 10 (2 )
1 3
18. 19 . 20. 98. 7d.
3 8

24. This can be obtained by the principles of Arts. 769 and

onwards. The chance that after a years A will be living
86 -m-x x
and B dead is Hence the value of A's ex-
86-m 86- n

pectation of a payment of £1 at the end of a years is

£
(86 -m -x) x • v .
(86 -m) (86 -n)
By summing up these for all values of x the required result
is obtained.
i

1
:

1
MISCELLANEOUS EXAMPLES .
:
 MISCELLANEOUS EXAMPLES.

1. Simplify
(a + b + c) (x + y + z) + (a + b − c) (x + y- z)
+ (b + c- a) (y + zx) + (c + a − b) (z + x − y).
2. Prove that

a (a -x) (a - 2x) = ( ab) ( a - b - x) (a + 2b - 2x)

+b (b - x) (3a - 26-2x) .
3. Prove that
x (x + y) (x + 2y) = (x − y) (x - 2y) (х- 3у)
+ 9y(x − y) (x- 2y) + 18y² (x − y) + 6y³ .
4. Simplify
3 (a -b + c) -5 (a -2b + 3c) +4 (a - 3b + 2c) -2 (a- 76-2c).
5. Multiply 2a + 3b by 3a- 46, and divide x² - 4xy + 3y²
by x² - 2xy + y².
6. Shew that { 2bc (a − b) - (b² + c² - a²) (a -c)} 2
+(a- c)² (b + c- a) (c + a -b) (a + b −c) (a + b + c)
= 4abc { abc- (b + c -a) (c + a - b) (a + b −c)} .
7. Multiply a³- 7ab2-663 by a³ - 11ab2-663, and divide
the product by a³ - 6a²b + 7ab² + 66³.
8. Multiply
7a - 3a²b - 2ab3-26 by a + a²b² + b²,
and divide the product by a³ - b³.
9. Multiply
a* -3a²x - 9a²x² - 36ax² + 14x by 2a²- 3 ах + х² ,
and divide the product by a²-6x + 2x².
10. Multiply
4x² - 5x- 4-7x-1+ 6x-2 by 3x- 4 + 2x-1,
and divide the product by 3x- 10 + 10x-1-4x-2.
554 Miscellaneous Examples.
11. Prove that

= a + b + c.
( - ) +( - ) 2+( - ).
12. Simplify( - ) + ( - ). 9 + ( - )
2 2
a+ x X -

X (^
13. Prove that
(lx + my + nz) { (m - n)² yz + (n - 1)2zx + (1 -m)²xy}
-(x+ y+ z) { l (m - n)² yz + m (n - 1)² zx + n (l - m)²xy}
= { (1 -m) y + (1 - n)z} { (m -n) z
+ (m -1)x} {n -1)x+(n -m) y} .
14. Simplify
C 1 p- c 1 p+ c 1
+ 3
+ 3

p²x p²x 2р³ х-р 2p³ x+p

15. Simplify
(х-а) (x- b) (x- c) - {bc (x - a) - [(a + b + c) x - a (b + c)] x }.
16. Simplify
a² + b² + c² + (a -b + c) (a + b − c) + (b -c+ a) (b + c-a)
+ (c-a + b) (c + a− b).
17. Shew that
(1 + xz) (1 + yz)² - { (1 - xz) ( 1 - yz) + 2xyz}*
= 4 (х + у -ху) (xyz³ + xyz² + z).
18. Simplify
1

2(x-1)²-
2 (x- 1)244((x-1)
- 1) +4(x+1)
4 (x+ 1) - (x-1)
(x- 1)* (x(x++ 1)
19. Simplify
a b b2
1+ - + -१ 1+ -
१
a a
D D

a
2 b 62 b
a
-2 10

a a2 ba
+2 2
 Miscellaneous Examples. 555

20. Simplify
5a - 7 (b - c) - [6a-( 3b + 2c) + 4c- { 2a-(b + 2c-a)} ].
21. Divide
b (x³ - a³) + ax (x²-a²) + a³ (x-a) by (a+ b) (x-a).
22. Find the Highest Common Divisor of
c²x² + bx + c (a + b) x + b²x + ab
and bcx + (b² + c²) x + b (a + c) x + ac.
23. Find the Highest Common Divisor of
x - 3x + x - 4x² + 12x- 4 and 2x²- 6x + 3x² - 3x + 1.
24. Find the H. C. D. of
x² + 4x² + 16 and x - x² + 8x-8 .
25. Find the H. C. D. of
x -x²y + xy - xy + xy² -у
and x -x²y -x y + xy² + xy²-y .
26. Find the H. C. D. of
x² + 2x - 3x² - 4x + 4 and x² + 2x²-x- 2 .
27. Find the L. C. M. of
x² - 6x + 11x- 6 and x - 9x² + 26x - 24 .
28. Find the H. C. D of
12x + 30x + 60x² + 48x² + 30x²
and 18x - 9x + 9x5-63x² + 45x .
29. Find the Highest Common Divisor of
4x² + 3x- 10 and 4x + 7x - 3x- 15 .
30. Find the H. C. D. and L. C. M. of
2ab + 3ab² - 3a* b* − 2 a b + a²b - ab
and 2a²b + 3ab² + ab + 4a* b* - a²b + a * bº.
31. Find the H. C. D. of
5x - 3x²y + 13x+ y²- 10x y³ - 3x² у* -7ху³ - 11 у
and 7x - 6x5y + 20xy² - 20x³² y³ - 9 x²y* - 14xy – 22y .
 556 Miscellaneous Examples.

32. If x + y + z = 0, find the H. C. D. of x + y + z and

x +y +z7.
33. Find the H. C. D. of
4
22x - 52x+ y + 30x3y² + 2x²y² + 4xy - 6y
and 29x - 70x²y + 41 x³ y² + 3x²y + 6 xy² - 9y .
34. Find the H. C. D. of
190x235x² - 5x + 70x² - 21x + 1
and 121x² - 214x² + 115x² - 23x + 1 .

35. Find the H. C. D. and L. C. M. of

4a³ + 4a²b - 13ab² + 563
and 6a² - 5a²b - 13a²b² + 17 ab³ - 56*.

36. Find the H. C. D. of

5x² + 38x² - 195x - 600 and 4x² - 15x² - 38x + 65 .
37. Shew that

{ 2x + (x² - a²) } { 2x- (x²- a²) } = (x + a) -(x- a) .

38. Shew that

√5-13-5 43-5 √5-1 1

3 2 ) + 15( 2) +3 2
=

39. Simplify
2
{ (1 + x) 1 - x- ( 1 - x) 1 +x } - { 1 -x - 1 } .
40. Shew that

(1 + a) v1+b² ( 1 + b) √1 + a² 2(1 -ab)

a- b
(1 + a) √1 + b² + (1 + b) √1 + a
41. Shew that
1 a√2i + b a√- 2i + b
!!!√- 21
i √2i
-) = 2a + b, where 2-1 .

Solve the equations :

x + 3x + 4x + 5
42. + + = 16 .
2 3 4
 Miscellaneous Examples. 557

x+ 7x+ 2 3x+ 7 1
43.
2 3 12 4

1 1 1
44.-
x- 3 x- 5 x-2 x-4

1 1 7 1
45. + = +
x²- 1 x+ 1 8 1-x

x+ 1 5- x 7
46. =

x- 2 2 (x- 1 ) -3
3+ 4√x
47.x- 3 =
X

48. (x - 1 ) (x- 2) (x - 3 ) = 2.3.4.

49. (x- 3 ) (x - 4) (x - 5) (x - 6) = 1.2.3.4.
x² + 2x²- 1 2
.
50. = x+ 2 + --
x²- 1 x

51. (x -m) (x + m)² + (x - n) (x + n)² = x (x² + 5mn).

52.
(xy + 6 = 3y + 2x,
{ x + y = 5.
3a

53. { x+
x+ yy== '
ху= та².
х-у 20

54. x- y + Ny x+y x+y

x² + y² = 34.
55. x (y + z) = a² ; y (z + x) = b² ; z (x + y) = c².
√x-√y= 2√xy,
56.
{ x+y= 20.

√x+ y+ √x− y = √2a,

57.
{ √x² + y² + √x² -y² = a.
558 Miscellaneous Examples.

58.
√(x² + a² ) (y² + b²) + √(x² + b²) (y² + a²) = (a + b)²,
x+ y = a + b .
59.1 + 4x- 8x² + 2x² = 0.

12
60. 2-10x +3
6-x x- 2 5

3 8
61.5
5- x + 4
4-x = 2

ac be
62. = a +b.
bx ax

1 3

2
*-3-10
4 3
63.
5
+
2
--
14-20) = 10-
8 6

12a
64. √a + x ++ √a
√a - x =
= a.
5√a+x

65. x + 1 = (x² + 1).

√3

66. (x² + 7x + 5)² - 3x² - 21 x = 19 .

3x+ 1 2x- 7 5
67. = 0.
3(x-5) 2x- 8 2

68. a² + b²_b²+c²-
4x²
a² = (a + b −c) (a - b + c) .
2 2

9x+ 5 4x- 2 12x + 3 4x + 3 11

69 . =
+
12 7x- 1 16 7x+ 9 48

70. x² + xy = 66 , x² - y² = 11 .
28

Y 13
71.-+ = ,
25 xy= 6.
Y x 6

7
72. x + y = 2 , x + y =2
 Miscellaneous Examples. 559

36 18
+ = 4,
73.
x² + 2xy+ y² x+ y
12 2
= 2.
x²- 2xy+ y² х-у

x+yy + z z+u и-х

74. = and 3z + u- 2x- 4y = 8 .
3 5 8 4

75.
x + y + z = xyz = 6.

(b-c)√x+ (c-a) √y+(a −b) √z = 0, 2

(b- c)2 + (c-a) (a- b)² = 0,

76.
{ y- z 2-0
+
х-у
ax + by + cz = a³ + b³ + c³.
77. Find the value of x from the equations
y + z + u = 14, 2x + 3z + 2u = 33 , x + y + 2u = 18 ,
3x + 2y + 2z = 20.
78. Solve the equations :
cx - by + az = a² + c², bx + ay- cz = a² + b²,
-ax + cy + bz = b² + c².
79. Eliminate x, y, z, u from the equations
x + y + z + u = 0,
x² +y² -z- u² = a²,
x + y + z + u³ = b ,
x +y -z - u² = c ,
x +y +z + u = d .
80. Form the equation whose roots are the squares of the
sum and difference of the roots of
2x² + 2 (m + n) x + (m² + n²) = 0.
81. If
x* + px³ + qx² + rx + 1 = 0 and x² + rx² + qx² +px + 1 = 0
have a common root, shew that p + r = q + 2, the symbols being
supposed essentially positive.
560 Miscellaneous Examples.
82. If each pair of the equations
x² +P₁x + q₁ = 0, x² + P₂x + q₂ = 0, x² + Px + q = 0
have a root in common, shew that
P3
(Pa +Po-P₁)( +2 L2 )= (Po+P -P2)( + 91 - )
91
P1

P1 P2 P3
=(P +P2-P3)(11 +222).
13 92

83. Eliminate x, y, z from the equations

28

2
х
+ = a, ++ = b, x + y + z = c,
X y 2

xyz = c (( x² +y² + z²).

84. If x* + px + q = 0 and x² + qx + p = 0 have two roots
2

in common, shew that p³+p² + 7p − 1 = 0 , p and q being

positive.
85. If ax + bx + c be divisible by x² + hx +h², then will
ac = b2.

86. Prove that

(ah² + 2bhk + ck²) (ah'² + 2bh'k' + ck'²) - { ahh' + b (hk' + h'k) + ckk'} ²
= (ac - b²) (hk' -h'k)².
87. From the equations
(1 ) cx- 2by + az = 0 ; (2) h²x + 2hky + z = 0( ;
(3) h²x + 2h'ky + kz = 0 ; (4) ah² + 2bhk + ck² = 1 ;
( 5) ah'² + 2bh'k' +ck'² = 1 ;
deduce that (ac-b²) (hk'-h'k)² = 1 .
88. Eliminate x, y, z from the equations
yz y 20 22 ху
+ 2
= 1; + 2 = m; + = n.
yz x² 20
y ху 22

89. Ifthe two expressions x³ +px² + qx + r, x² + p'x² + qx + r',

have the same quadratic factor, prove that
r- r p'r- pr_dr-qr =

p-p q-4 r-r

 Miscellaneous Examples. 561

P-P
Shew also that the third factors are x + rand
r- r
P-P
x+ •r; and that the quadratic factor is
r- r

20²+ 9-1 x+ r-r

P-P P-P

90. If x² + ax² + bx + c be divisible by x²+px+ q, prove that

p ( p -a) = q- b.
91. If a + b + c = 0, then
(1) (a² +b² + c²)² = 2 (a² + b² + c²) ;
11abc
(2) a +b +c = 2
{a³ +b + c³ - 3ab² c² (a² + b² + c²)} .
a-c
92. If (a-b) (b- d)+(b−c) (c-a)+(d -a) (a-c) =

(a-c) (c-d) + (b -d) (d - a) + (d - b) (b −c) b-d'

shew that either a + d = b + cor (a - b) (c - d) = (b−c) .
93. If a +b + c = 0,
and a (by + cz- ax) = b (cz + ax -by) = c (ax +by- cz),
then x+ y+ z = 0.

94. Find the values of a and b in order that x² + 3xy + 4y²

may exactly divide
x + 7x y + 6x+ y² + 5 x²y + ax²y + bxy + 12y .
95. If x =
a+b-c
, then
a + bx²_ (a -b + c)² + 4ab
a+b+c b + ax (b - a + c)² +4ab
96. If (a + b)² = c (a-b), prove that
b (a + b + c) (a + b -c) + 2abc (a + b + c) -4abc (a + b - c)
-ac (a + b + c) (a + b -c) = 8abc².
97. If x, y, z be unequal, and if
2a- 3y = (2- x)² and 24-32= (x−y) , 2
Y

then will 2a- 3x = (y- 2)2-, and w + y + z = a.

00
562 Miscellaneous Examples.

98. Prove that the result of eliminating x, y, z from the

equations
ax² + by² + cz² = ax + by + cz = yz + zx + xy = 0,
is abc = (b + c - a) (c + a − b) (a + b - c).
99. Solve the equations :
(1) √√a+ √x + √√ - √x = √2√6 + 2√b + x ;
(2) ( )* + ( ) =a +b ;

(3) x + a = (x + a) .
100. If x1, x2 be the roots of the equation
A B C
+ + = 0,
х-а x-b x- c

find the value of (x -a) (x2 - a) without solving the equation.

101. Solve the equation
(a + x)-- 6 (a² -x²)- = -5 (a -x)-3.
102. The rational values of x which satisfy the equation
Y = xare given by the formula x = (2+1) , where p isany
P
integer positive or negative.
103. If x² + 2ay is a square, x² +ay² is the sum of two
squares .

104. If a and 3 are the roots of ax² + bx + c = 0, shew that

a³
the equation whose roots are and Bis
B2 a

a c² x² + (5a²bc² - 5abc + b²) x + a² c³ = 0 .

105. Solve the equations :
a² + b² + c² + d²
(1) ²+c² + b² + d² = (ab -cd)² + (ad + bc)2 ;
2

5 6
(2) x+ 3 + x+4 = 2;
 Miscellaneous Examples. 563
27
x* y* (x + y ) + x + y = 4
-x yt,
(3)
{ Y-x = 4;
15

x+ y y+ z
+
(x -2) (x −y) (x -2) (y−z) X
х-у a
2-0
(4
) x+ z z+ y
+
(x − y) (x-2) (x − y) (z -y)
x² = yz, x² + y² + z² = a² + ab + b².
106. If a and ẞ are the roots of the equation ax² + bx + c = 0,
prove that the equation of which the roots are
a* + * and -(α + β) (a³ + 3³)
is a³ x² - a c (2ac -b²) x + b² (6a² c³ - 14 a² b² c² + 7abc - b*) = 0.
(b+c) 62
107. Prove that C2 (c+a)² a2 = 2 (bc + ca + ab)³.
a² (a+b)²
1 +x 2 3 4

1 2+x 3 4
108. Prove that = x² (x + 10).
1 2 3 +x 4
1 2 3 4 +x

(b+c) a² a2
109. Prove that 62 (c+a)² 62 = 2abc (a + b + c) .
C2 c2 (a+b)²
110. Prove that the determinant
1111
αβγδ
α βε γε δε
α' β' γ' δε
= (α - δ) (3-8) (γ - 8) (α - γ) (β - γ) (α - β) (α + β + γ + δ).
111. Reduce the equation
a³ C

(a+1)3 (6 +1)3 (c+ a)3 = 0

(2a + 1)3 (26 + 1) (2c + 13)
to the form of a quadratic in λ.
002
564 Miscellaneous Examples.
112. Prove that
-mn , m² +mn, mn +n²
nl +l ,-nl n² +nl = (mn + nl + lm)².
,

l +lm, lm +m², -lm

113. Given x² + y² + z² - 2yz- 2zx- 2xy = 0 ,
prove that √x + √y+ √z = 0 .
114. Investigate the necessary independent conditions for the
coexistence of the four equations
bz + cy cx+ az ay+ bx
ax +by + c'z = 0 , c
=x+y + z ;

and shew that they can be put into the form

(aa') +( b) + (c ) = 0, ( )*+ * +(÷) = (ab )².

115. Two pedestrians start at the same time from two towns
and each walks at a uniform rate towards the other town.
When they meet, it is found that one has travelled 96 miles
more than the other, and that if they proceed at the same
rate they will finish the journey in 4 and 9 days respectively.
Find the distance between the towns and the rates of walking
perday.
116. Two boats A and B row a bumping race against a
stream flowing 1 mile per hour; in still water they row re-
spectively 13.2 feet per second and 266 yards per minute :
Bgains 44-3 feet on A but does not bump. A time race is
rowed down the stream over A's course. Byhow many seconds
willB win?

117. A farmer sold 10 sheep at a certain price, and 5 others

at 10s. less per head. The sum he received for each lot was
expressed in pounds sterling by the same two digits. Find
the price of each sheep.
118. Two persons A, B walk from Pto Q and back. A starts
1 hour after B, overtakes him 1 mile from Q, meets him 20
minutes afterwards, and arrives at P when Bis of the way
back. Find the distance from P to Q and the rates at which
they walk.
 Miscellaneous Examples. 565

119. In a match a person fired 7 shots at each of three

ranges. He made an equal number of centres at each of the
first two ranges and twice as many at the third, and he made
the same number of outers at the second and third ranges ;
the number of bullseyes and misses at the first and last ranges
were all the same, and he scored as much for bullseyes at the
second range as for outers at the first. A bullseye counts
4points, a centre 3, and an outer 2, and his total score was
56, that at the last range being 20. Find the number of
bullseyes, centres, outers and misses at each range.
120. A steamer, whose speed in still water is v₁ miles per
hour, starts at a certain hour up a river which it ascends
against the current in t₁ hours. N hours afterwards another
steamer (speed v₂) starts down the stream, and accomplishes
the distance in t₂ hours. In how many hours will they meet ?
Find also the length of the river and the velocity of the
current .

121. On a tidal river a boat's crew can row with the tide
half as fast again as they can in still water. They start with
the tide at 12 o'clock to row from A to B, but on reaching
a certain place, C, the tide changes and flows with equal
velocity in the opposite direction. The rest of their journey
occupies half an hour more than the time from A to C. If
the distance from A to B had been 10 miles more they would
have been three hours longer on the journey; or if they had
gone twice the distance they did before the tide changed, the
whole journey would have been completed in two-thirds of
the time actually consumed. Required the distance from A
to B and the time at which the tide changed.
122. The distance from London to Peterborough is 75 miles,
and from Peterborough to Grantham 30 miles. An up ordinary
train starts from Grantham at the same time that the down
express leaves London. The up train is delayed 30 minutes
at Peterborough, and then passes the express 10 minutes after
leaving Peterborough. Again, a down ordinary train leaves
London at the same time that the up express leaves Grantham,
566 Miscellaneous Examples.

each train travelling at the same rate as the other ordinary

and express respectively; but in this case, the express, delayed
only 6 minutes at Peterborough, arrives at the point where
the former trains met, when the ordinary train has only got
two-fifths of the distance of that point from London. Find
the speed of the trains.
123. A market-woman having bought equal quantities of
eggs at two different prices, sells them all at one price, giving
to her customers for twopence the sum of the numbers she
got for a penny at the two prices, and finds that she has lost
pence equal to the difference of these numbers. If she had
given one less for twopence, she would have gained pence
equal to the sum of the same numbers. If she had sold them
all for 70 pence, the price of each egg would have exceeded
the average price per egg at which she bought the whole by
half the price of one of the cheaper sort. Find the number
bought and the price.

124. A flock of s sheep is turned into a field of turnips

which would last them d days ; after d' days, s
' sheep are
added to the flock. In how many days will the remainder be
consumed ?

125. A number consisting of two digits is such that when

divided by the sum of the digits the result is the second digit ;
and if the digits be reversed the square of the number thus
formed is four times the cube of the sum of the digits. Find
the number.

126. What is the price of eggs per score, when 10 more

in half-a-crown's worth lowers the price 3d. per score ?
127. In the astronomical clock where the hours are marked
upon the dial from 1 up to 24; find the two times between
8 and 9 o'clock when the hands are at right angles.
128. An officer can form the men in his battalion into a
hollow square 4 deep and also into a hollow square 8 deep.
If the front in the latter formation contain 16 men fewer than
 Miscellaneous Examples. 567

in the former formation; find the number of men in his

battalion. Shew that the battalion can be formed into 3 other
hollow squares only.
129. Extract the square roots of
x² 4
x*+
4
+x² -x + 4x- 2 and 16+ 5√7.
130. Having given that x² - 6x + 4x² is the difference of
two perfect squares neither of which vanishes with æ, find
them.
131. Prove that
(x² + y² + z²) (a² + b² + c²)
= (ax + by + cz)² + (ay - bx)² + (bz - cy)² + (cx-az)².
132. Having given that
ay₁z + bzx + cx₁Y₁ = ay2z2 + bz2x2 + CX2 Y2 = 0, and
(bz₁ + cy₂) x + (cx₁ + az₂) y + (ay₁ + bx₂) z = ay₁₂ + bz₁₂ + CX1Y2 ,
(bz₂ + cy₁) x + (cx₂ + az₁) y + (ay₂ + bx₁) z = ay₂z₁ + bz2 x₁ + cx2Y1;
0-1 У-У1 = 2-21
shew that ;
х-х, У- У2 2-22
ах cz
and + by + = 0.
X1X2 Y1Y2 12
133. If 8 = (a + b + c), prove that
83 -( -a) - ( - b)³ - (sc)³ = 3abc.
134. Prove that
x (y + z)² + y(z + x)² + z (x + y)² - 4xyz = (y + z) (z + x) (x + y).
135. If n be an odd integer, prove that
(a + b + c)"- a -bn -c"
is divisible by (b + c) (c + a) (a + b).
136. Given aX+ by + cZ = 0, a'X+ b'Y + c'Z = 0 ;
where X= ax + a'x' + a", Y= bx + b'x' + b", Z = cx + c'x' +c":
shew that
X²+Y² + 22 = {a"(bc -b'c)+ b"(a'c-ac') +c"(ab'-a'b)}²
(bc - b'c)² + (a'c- ac')² + (ab' - a'b)²
568 Miscellaneous Examples.
137. Find the square root of 4738-027, and of
4x - 12xy+ 25x²y² - 24xy + 16y .
3√2 4√3 √の
138. Simplify +
√3+ √6 √6+ √21√2+ √3
139. If x* +px + qx² + rx + s be a perfect square, shew that
r² = p²s and p³ - 4pq + 8r = 0.
140. If a + √o+ √c+ √a+ √e+ √f+√g can have its
square root extracted in the form of a polynomial surd, shew
be
that de = of= bg and that e
must be rational, and find
the other condition.
Extract the square root of
8+ 2/6-2-3-2-2 + 4√2+ 4√3-4-1.
141. The first term of an arithmetical progression being 2,
the fifth being 7, and the sum being 63; find the number of
terms.

142. In an arithmetical series consisting of an odd number

of terms, the sum of the odd terms is 44 and the sum of the
even terms is 33. Find the middle term, and the number of
terms.

143. Find the sum of 4 + 3 + 2 + ... to nine terms and the

number of terms whose sum is 9.

144. Insert 7 arithmetical means between -13 and 3.

145. A complete polynomial of a dimensions in x and y is

a + bx + by + cx² + c'xy +c"y² + ... + hx + hx - y + ... + h' "y" :
shew that the number of terms is (n + 1)
1.2
(n + 2) .
146. The sum of a geometrical series whose greatest term
is 12, is 21 ; but when the series is continued through as many
more terms the sum is 189 ; determine the series.
147. Find an arithmetic series whose fourth term is 3 and
the sum of 7 terms is 21.
 Miscellaneous Examples. 569

148. If four quantities a, b, c, d are in harmonical pro-

b-d 2d b-d 26
gression, prove that a-b
=

ac-dc

149. Prove that (n 1+ 1) (n + 2) (n + 3) ... 2η = 2".

3 5 ... (2n - 1)
150. Let a, b, c be any three quantities ; if n harmonic
means be inserted between a and b, and m harmonic means
between b and c, find the condition that all these quantities
may be in harmonic progression.
2
151. Sum 18 terms of the series 3, -1, -23.....
152. If the 5th and 18th terms of an arithmetical pro-
gression be respectively 7 and 72, find the first four terms.
153. If two geometrical progressions having the same first
terms and continued to infinity are in the proportion of their
common ratios, shew that these common ratios are equal or
have their sum equal to unity.
154. Sum ton terms
Pq + (p - 1) (q - 1) + ( - 2) (q - 2) + ... .
Prove that
12 22 32 n2 (1 + x)x +1 - { n (x- 1) +x }²-x
+ + + +
х x" x" (x- 1)3
Deduce the value of 12 + 22 + ... +n².

155. Three persons A, B, C whose ages are in geometric

progression, divide amongst them a sum of money in amounts
proportional to the ages of each. Five years afterwards when
Cis double the age of A they similarly divide an equal sum,
A now receiving £17 10s. more than before, and B £2 10s.
more than before. Find the sum divided on each occasion.

156. Find the present value of an infinite series of annual

payments, the first payment being £1, the second £2, the
third £3, and so on.
570 Miscellaneous Examples.

157. A property now worth £A per annum increases every

year in the ratio 1 +p to 1 ; find the present value of the
enjoyment of it for n years. What does this become if p = r,
£r being the interest on £1 for one year ?
158. I borrow £ 1000 on condition that I repay £10 at the
end of every month for 10 years. Find an equation for
determining the rate of interest I pay.
159. There are 24 oranges at 7 for a shilling. How many
selections can be made in buying three shillings' worth ? In
how many of these will a particular orange occur ?
160. There are fifty different kinds of cakes at a con-
fectioners. A person orders two particular kinds to be sent
in, and any other three that the confectioner chooses. How
many different assortments may be sent ?
161. How many different sums can be made by taking three
coins from a purse containing a sixpence, a shilling, a florin,
a crown, a half-sovereign and a sovereign ?
162. Four ladies and four gentlemen are arranged to dance
a quadrille, two couples standing side by side, at one end of
a room, vis à vis to the others at the other end. How many
arrangements are possible ?
163. A crew for an eight oar has to be chosen out of eleven
men, five of whom can row on the stroke side only, four on the
bow side only, and the remaining two on either side. How
many different selections can be made ?
164. Fifty-two cards in four suits are dealt out to four
people in rounds; find expressions for the number of hands
which one person can have and for the number of possible sets
that can be dealt.

165. Find the general term of (a - 2x )- .

166. Find in their simplest forms the coefficients of
(1) x³ in ( -1 +x)10; (2) x in (1-7x) .
 Miscellaneous Examples. 571

167. Find in its simplest form the coefficient of x in

(1 + 3x) , and the first negative term.
168. Give the general term of (1 - nx)-
169. Write down the general terms of
1
76

(1)(1-5 )- ; (2) (3+ ) ; 2 (3) (1 )

170. If in (a +b)" the 7th and 8th terms are in the same
ratio as the 6th and 7th in (a+b) +1, find n.
171. If the two middle terms of (a+b)2n+1 be taken and
n = 0, 1, 2, ... in succession, the sum of all will be equal to
1
a+bs 1
2ab1-4ab
1
x

172. Find the coefficient of an in expanded in

1
ascending powers of x. x+
x

173. Find the coefficient of 23 " in (1 + x2)3

(1 -x )
174. If (1 +x)" = 1 + B₁x + B₂x² + B₂x² +
2
.., find the values
... ‫و‬
(1 -x)m
ofB1, B2 , B3 .
3 4 5
175. Shew that 2 =
4 + 8 + 16 + ... ad infinitum.
176. Find the coefficient of x20 in (x + x + x + x + xº)4.
177. Find the coefficient of x in (1 + x + 2x² + 3x²) .
178. If S₁, S2 , S3, S₁ be the sums of every fourth coefficient
in ( 1 + x + x² + x³)" beginning with the 1st, 2nd, 3rd, 4th,
respectively, prove that S₁ = S₂ = S = S.
179. If n be any prime number except 2, the integral part of
( 5+ 2)"- 2"+1 is divisible by 20n.
180. Iff(n, r) be the number of combinations of n things
taken r together, prove that
f(n, 1) . f(n, 2) +f(n, 2) .f(n, 3) + ...
... +f(n, n- 1) .f(n, n) = f(2n, n + 1) -f(n, 1).
572 Miscellaneous Examples.

181. Prove that the sum of the products of the first n

natural numbers three and three together
(n -2) (n - 1)n2
=
п² (n+ 1)² •

48

182. Prove that the sum of the products of the n quantities

c, c² c³, ...c" taken m at a time
m(m+1)
= (n-1) (c -1-1) ... (cn-m+1-1).c 2

(c-1) (c² - 1) ... (c - 1)

5 + 6x- 2x²
183. Resolve into its partial fractions
(3+ 2x)
1
184. Resolve into its partial fractions
x (x- 1)² (x² + 1)
x³ +2
185. Resolve into partial fractions.
(x-2)² (x²+ 1)
186. Prove that
1 12

= 1 + 4x + 11x² + ... + (2n+2 n − 3) x² + ....

(1-2x) ( 1-2x + x²)
1+ 7x-x²
187. Expand in a series of ascending
(1 + 3x)² (1-10x)
powers of x, and find the nth term of the series.
188. Shew that the series u + u₁ + u₂ + ... is convergent if
(un) is always less than a quantity which is itself less than
unity, however much n may be increased.
189. Transform the series
ax a2x² Anxn
+ + ... + + ...
1 + 6, x (1 + bx)"
(1 + b2x)²
into one ascending by powers of x, and find the nth term of the
series. Hence sum the series
2 3

14+
1 + ах (1 ) + (1 ) +... ad inf.
190. Prove that
(x-ων)" = Χ- ων, (x- w²y)" = (X-w² Y)
 Miscellaneous Examples. 573

where w³ = 1 and X and Y are rational functions of x and

y. Hence prove that (x² + xy+ y²)" can be put in the form
X²+XY+ Y², n being integral.
If n = 2, X = x² -y², Y = 2 xy +y²,
n = 3, X = x² - 3xy -y³, Y = 3xy (x + y).
191. Aperson has 3n homogeneous balls of diameters 1,2,3, ...
inches : the weight of the first n balls is an aliquot part of that
of the others ; how many balls are there and what is the aliquot
part in question ?
192. If 1º +2 + ... to n terms = (1 + 2 + ...)",where p, q, r
are integers ; find p, q, r.
193. The population of a country increases slowly byth
P
of
itself every year, and after every decennial census an epidemic
takes off 1th
th of the population; at the end of Nyears it is found
q
that the population is the same as it would have been in half
the time had there been no epidemics ; shew that approximately
१- 1

1+p (1 )
where n is the number of years elapsed since the last census.
194. Shew that
n (n + 1 ) ... (n + m - 1) - n
n (n + 1) ... (n + m - 4)
m m- 3

n (n - 1 ) n (n + 1 ) ... (n + m - 7) +
+ ...

1.2 m- 6

is zero if m > 2n and = 1 if m = 2n. Give the last term in

the series.

195. Prove that the expression (a-x) {x + √x² +b² } cannot

exceed (a² + b²) .
1 1 1 1 1 1 1
108.
196. If น
+ υ
= uta + v- b= u-a + = ,

thenwill f² (ab'-a'b)² = aa'bb' (a- a) (b -b').

574 Miscellaneous Examples.

197. There cannot in any scale be found three different

digits such that the three numbers formed from them by
placing each digit differently in each number shall be in arith-
metical progression unless the radix of the scale exceed by
unity a multiple of three. If this condition be satisfied and the
radix be 3p + 1, there are then (p- 1) such sets of digits ; and
the common difference of the progressions is in all of them the
same.

198. If x² + y² = z² where x, y, z are integers, prove that one

of the three x, y, z is divisible by 5.
199. If a³ + b³ = c³ where a, b, c are integers, shew that abc is
amultiple of 7.
200. If a" + b" = c", shew that abc is a multiple of (2n + 1),
provided 2n + 1 be a prime number.
 ANSWERS TO MISCELLANEOUS
EXAMPLES.

1. 4(ax+ by+ cz).

2. and 3. Reduce the expressions on the right.
4. 96. 5. 6a² + ab - 126² ; x² + 2xy + 3y².
6. Reduce both sides.

7. a - 18ab² - 12ab +77 a²b + 108 ab + 366° ;

a³ + 6a²b + 11 ab² +66 .
8.7a³ - 3a²b + 7ab² - 5a²b + 5a* b* -5ab5 - 2a²b - 2ab - 268 ;
7a5-3a²b + 7ab² + 2a²b + 2ab + 265.
9.2-9ax- 8a* x² - 48a3x² + 127a2x² - 78ax5 + 14xº ;
2a³ + 3a²x + 6a²x² - 18 ax² + 7x*.
10. 12x3-31 x² + 16x- 15 + 38x-1-38x-2 + 12x- 3 ;
4x² + 3x + 2-3x-1.
a² + ax- x² 14.
x- c
15. x .
12.
a(a+x) x² (x² -p²)
16. 2 (bc + ca + ab). 17. Use formula of Art. 116.

18. 0 . 19. 1. 20.2a - 5b + 3c.

21. x² + ax + a². 22. cx + bx + a.

23. x² - 3x + 1 . 24. x² - 2x + 4.

25. x -y . 26. x² + x - 2 .

27. (x-1) (x- 2) (x- 3) (x- 4).

576 Answers to Miscellaneous Examples.

28. 3x² ( 2x² + 3x + 5) . 29.4x- 5 .

30. ( 2a³ + 3a²b - ab² + b²) ab ;

a²b² (a² -b²) (2a³ + 3a²b - ab² + b²).
31. x² +y².
32. Putting -(x+ y) for z in each of the expressions the
H. C. D. will be found to be x² + xz + y² or (x² + y² + z²).
33. х²- 2xy +y². 34. x- 1 .

35.2a² - 3ab + b² ;
(2a + 5b) ( 6a - 5 a* b - 13a²b² +17 ab³ - 564).
36. x-5 .

37. √2x- (x² -a²) = √x + a - x-a (Art. 266).

38. √3-√5 - √5-1 (Art. 266).

2
=

39.2x²√1 -x² -x². 40. Use Art. 259.

41. It is easily seen that (1 + i)² = 2i and that

(1-1) = - 2i,
whence √2i = 1 + i, √-2i = 1 -i.
42. x 11 . 43. x 24 . 44. x = 3 .
5 1
45. x = 3 or 46. n = 4 or
7 5

47. The equation may be written in the form

x (x - 1 ) = 3 (x + x + 1).
The factor x + x + 1 is obvious, and we have the two equations
1
x + x + 1 = 0, or x - x = 3 .
48. The root x = 5 is obvious. Hence the others can be
obtained by removing the factor x- 5. They are impossible
quantities.
49. The two roots 7 and 2 are obvious.
50. x 2.
 Answers to Miscellaneous Examples. 577

51. The equation can be written in the form

(x² -m3 -n³ - 3mnx) + x² (m + n) -x (m + n)² = 0,
whence a factor x - m -n is obvious : the other values are
x = -(m + n) ± √3mn.
52. x = 3 , y = 2.
3 + √9-16m 3+√9-16m
53. x = a . , y= a . .

2 4

54. x = ± 5, у = ± 3 .

55. x² =(a²+b²-c²) (a²-b²+c²) &c. ,

2 (6² + c²-a²)
56. x = 10 + 4 /6, y = 10-4√6.
a

57. x = 5 , y = 0.
58. x = a, y = b or x = b, y = a : obtain a result xy= ab.
59. Can be written (1 + 2x- 2x²)² = 2x .
154 58
60. 1,
27
.
61.
13'
2. 62. (a-b)
ab
c.
3α 4α
63. 16. 64.
55
65. The equation can be written

2x -2x + 1 = 3(x -x√3 + 1),

and then, since (x²- x√3+ 1) (x² +x√3 + 1) = x - x² + 1, the
solution is reduced to that of the two quadratic equations
x²-x√3+ 1 = 0 and x²+ x√3+ 1 =
√3
66. Put x² + 7x+ 5 = y ; we thus get y²- 3 (y- 5) = 19,
whence y= 4 or -1; finally a = -1 , -7 or -7+3√5
2

59 bc
67. 7,7,8
15
68.x=
N
Pp
578 Answers to Miscellaneous Examples.
69. x = 3 . 70. x = ± 6, y = ± 5.
2 3 3 2

± 5 v= ±5
71. x = ± 5 ' y = ± z ; or x = °
3 1 1 3
72. x = y = 2; x=
2

3
73. x + y = 6 or ; x -y = 2 or -3.

74. x = 2, y = 4 , z = 6 , u = 10 .

75. The values of x, y, z are 1, 2, 3 .

76. From the first two, eliminating z, it can be shewn that
√x = √y. Similarly
1 a³ + b² + c²
√y= √3; ..

x = y = z = 3a+ b + c
77. x = 2 . 78. x = b + c, y = c + a, z = a + b .
2069 56° c
79. d = 0.
27a 3a²

80. y² - 4mny- (m² - n²) = 0.

81. The common factor must divide the difference of the
two expressions, and the common root must therefore be -1,
the symbols being all positive.
82. Assume the roots to be (α, β ) ; (β, γ) ; (γy, a) and use
the relations of Art. 414.

83. (a + b + 3) = ( 2 + 4 + 4b + ab) ( 7 + 2a +26).

84. By subtraction it is easily seen that the factor which
gives the common roots must be q'x² + x + 1) -p. Dividing
either of the given expressions by this the remainder must
identically vanish.
C

85. The other factor must be ax - ahx + , whence multi-

plying and equating the coefficients of x² and x, the result
follows.
 Answers to Miscellaneous Examples. 579

h
and
87. From (2) and (3) maybe regarded as the two
roots of a quadratic equation xt² + 2yt + z = 0, whence
h h 2y hh 2
= -;
xkk'
which may be put into the symmetrical form
kk hk +kk hh
-2y 2

Then by the last example the result can be deduced from the
other three equations.
88. 12+ m² + n² = Imn +4.
89. The common quadratic factor must be the difference of
the two expressions.
90. The remaining factor must be x + .
9

91. a + b + c = a + b (a +b)11. Expanding the latter,

and arranging the ten remaining terms in pairs, the factor
11 ab (a +b) is obvious and the remainder is easily reduced to
the required form.
92. Put a - b = u, b - c = v, c-d = w the given relation
easily reduces to u (v + w)² = w (u + v) , whence
u = w or uw = v².
93. Equate the last three quantities to k. Then we easily
k(b +c), whence the result follows.
get x = 2abc

94. a = 20, b = 85.

95. Substitute for æ and divide by factor a + b.
96. Substitute for cits value on the left-hand side and
reduce.

97. Subtracting one of the given relations from the other

we can deduce x² + y² + z² = 2 (yz + zx + xy) which being sym-
metrical the third relation must follow. From the first
2a =
3y² +(z−x)²
y

which can now be reduced to the form 2 (x +y + z).

Pp2
580 Answers to Miscellaneous Examples.
98. Eliminating x between (2) and (3) and comparing with
(1) we obtain ax² = (b + c- a) yz, and two similar equations.
a-b- 2√ab
99. (1 ) x = 0 , or 2
; (2) x = 1 or a*b ;
(3) x = 0 or a (1+2-2)' . 3
62
100. 4(a-b)(a-c)
A+B+ C
. 101. x = 0 or a.
63
1

102. Assume y = vx, then x = vv-1, and if this be rational

1

v- 1
must be an integer = p.

103. Assume x² + 2 ay² = (x + v)². Then it is easily seen

that v must be even and = 2z, whence x² + ay² = (x + 2)² + 2 .
7
105. ( 1 ) x = 1 ; (2) x = 2 or -

2
;

(3) the first equation can be written

1 1 2

(x + y + 2= x*y*)² = 9x1yt ,
20
whence canbe found. Among the values are
y
1 1
x=
4, y = 4 ; x = 4, yy =
=--
108. From Arts. 351 and 352 it follows that the value of
a determinant is not altered if the constituents of any column
be replaced by the differences between them and the cor-
responding constituents of any other column. By Art. 349
it follows that the same applies to rows. If the constituents
of all the rows except the first in this example be diminished
by those of the first, the determinant becomes
1 + x, 2, 3, 4 1 +x, 2, 3, 4
-х, х, 0, 0 -1 , 1 , 0, 0
=x
-х, о, х, 0 -1,0,1,0
-х, 0, 0, x -1,0,0,1-
which can be easily evaluated.
 Answers to Miscellaneous Examples. 581

109. Replace constituents of 2nd and 3rd columns by dif-

ferences between them and those of the first. A factor
(a+b +c) then becomes obvious.
110. Modify the first three columns by subtracting the con-
stituents of the fourth. The factors (a-d) (β-8) (7-8) can
then be removed. Repeat the similar process in the quotient
with the first two columns and the third.

111. By processes similar to those used in the last examples,

first replacing A by 2k, and taking the factor 8 out of the
third row, the determinant can be reduced to the form
3λο(α -b) (a - c) (b −c) {λ² (a + b + c) +31(bc + ca + ab) + 2abo}.
113. Use Art. 118 .

114. It is easy to prove that

20 2
y

a (bb' + cc - aa') b (aa -bb' + cc) c (aa + bb -cc)

whence, from the first equation, the first result follows by the
help of Example 113. The second result is obtained first in
the form
a (bb' + cc -aa') + b (aa' + cc -bb) + c (aa' + bb -cc') = 2abc ;
which by means of the former can be reduced to the required
shape.
115. 480 miles, 48 and 32 miles per day.
116. 2.4 seconds.

117. Price of dearest sheep was £3 2s. apiece, the solution

depends on an indeterminate equation.
118. 5 miles ; 4 and 2 miles per hour.
119. At 1st range bullseyes 0, centre 3, outer 4, misses 0.
At second range the numbers are 2, 3, 1, 1, and at the third
0, 6, 1, 0.
120. They meet after (N+t₂)
+ta
hours; length of stream
vo + v,
=
tt, miles ; velocity of current = vit, -vate miles per
to+t なな
hour.
PP3
582 Answers to Miscellaneous Examples.
121. 6 miles ; 20 minutes past 12.
122. 42 miles per hour and 30 miles per hour.
123. 84 eggs, 3 and 4 for a penny.
124.
s (d-d') 125. 45.
8+8

126. 15 pence per score.

127. 5'5 or 36′ past eight.
128. 288 men, 2, 12, or 16 deep.
х 2 5+√7
129. x² + ;
2
√2
130. (x² - 3x + 2)² -(3x - 2)².
132. By means of the first two relations x, y, z may be re-
placed in the second two by x - x , y - Y₁ , z -z, or by x- x ,
y-y2, z - z, the right-hand members becoming zero. Hence
the ratios of x - x , y - y₁ , z- z, to each other are the same
as those of x - X2 , Y - Y2 , Z- Z₂ . Also from the first two
a b C a b C
+ + = 0 = + + .
X1 У1 21 X2 Y2 Z2

C
Whence (x2 -x ) a
+
(y -y₁) b + (z - z₁) c = 0,
X1X2 Y1 Y2 2122

whence the result easily follows.

133, 134, and 135. Use Art. 163.
137.68.83 ; 2x² - 3xy + 4y² . 138.0 .

140.2 + √2 + √3 - √-1. 141. 9.

142. Middle term = 11 , 7 terms.

143. 0, 3 or 6 terms. 144. -11 , -9, etc.
146. 3 , 6, 12, .... 147. 0, 1 , 2 , 3 , ....

149. Multiply numerator and denominator by In .

150. (m + 1) c (a - b) = (n + 1) a ( b − c).
 Answers to Miscellaneous Examples. 583

151. -243. 152. -13 , -8 , -3, 2.

154. npq
-
n (n- 1)
(p +q) +
nn (n-
(n - 1) (2n - 1)
2 6

For the second example of 154 assume the sum to be S. Then

S
S- will depend on the series
х

1 3 2n- 1
+ + ... + ,

which can be treated in the same manner and the problem

reduced to the summation of a geometrical series. For the
deduction put x = 1 +h and find the value when his in-
definitely small,
155. £1045.
R

156. (R-1)2'› R being the amount of 1£ in 1 year.

157.
41-(
A R ) ; if R = 1 + p it becomes
nA

R 1+r
1-1+P
R

1
158.1- = 100r, £r being the interest of £1 for one
(1 + r)120
month.
24 23 48
159 . ; 160.
3 21 3 20 3 45

161. 20. 162. 4x 3.

163. 145. Take the different cases of selecting both the

indifferent men, one of them and neither.
52 52
164.
13 39 { 13}
2
165. - 5.7.9 ...(2 + 3)(*) .
a
5

.
r

5.2.9 ... (7-12)..

166. (1 ) -120-1 ; (2) r
584 Answers to Miscellaneous Examples.

167. The first negative term is the 7th. If r be greater than

5, the coefficient of x is

(-1) -5 . 13.10.7.4.3.1.2.5 ... ( 3r - 16)

r

168. (1 + n) (1 + 2n) ... (1 + r - 1n) 20.

r

7.9.11 ... (2r + 5 )

169. ( 1) r (5x)' ;
2

(2) (-1) .3-1.1.3.5 ... (2-1)

r 12
;

(3) (1 +n)(1 + 2 )...(1 + r-1 ).(1) .

r

170. n = -8 .
2n + 1
171. The sum of the two middle terms = a" b " . (a+ b).
nn + 1

172. ( -1) -1 (4n² + 1). 173. 2n.

174. B₁ = (m +n), B₂ = (m + n)²+(m - n),

2

B₁ = (m + n) {(m + n)² +3 (m- n) +2}

6

178. (1-1) = 1 +++++... +(+1) +.... 2r

176. 73. 177.726.

178. Give to the values - 1 and √-1 in succession.

179. Assume (√5+ 2)" - 2n+1 = I+ Fwhere Iis the integral
part and F a proper fraction. Then (2 - √5)" = -F', since n
is odd, where F' is a proper fraction. Thus
I+ F-F' = (2 + √5)" + (2 - √5) -2n+1,
which is an integer. Hence F = F', and it is easily seen by
expansion that Icontains the factors 22, 5 and n or is divisible
by 20n.
 Answers to Miscellaneous Examples. 585

180. ( 1 + x) " = (1 + x)" (1 + x)". Take the coefficient of

2 +1 on both sides. Or it can be done by considerations of
combinations.
181. Assume
(1 + x) (1 + 2x) ... ( 1 + nx) = 1 + P₁ x + P₂ x² + P₃x² + ...
where p₁ is the series we want; take logarithms of both sides
and equate coefficients of x, x, x³.
182. Assume
(1 + cx) (1 + c²x) ... (1 + c²x)
= 1 +P₁x + p₂ x² + ... +Pmx + ... +P x*.
Replace x by cæ thus,
(1 + c²x) ( 1 + c²x) ... (1 + c +1x)
mnm
= 1 +P₁ cx +P₂ c² x² + ... +Pm c™ x ™ + ... ,
whence
(1 + cx) (1 + p, cx +pac²x² + ... +Pmc ™ x + ...)
= ( 1 + cn+1 x) ( 1 + p₁x² + ... + Pm xm + ...),
and equating coefficients of x™ a relation is obtained between
Pm and Pm-1, whence the result can be deduced.
1 6 17
183. +
2 (3 + 2x ) (3 + 2x)² 2 (3 + 2x)
1 1
184. -
х 2-1 + 2(2-1) +2 (2 +1)
4 2 x+ 2
185. +
5 (x - 2) (x - 2)
2 ) 2 + 5 (x² + 1)
1 4 2 1
Take
186. (1-2x) (1 -x)² 1-2x 1-x (1 -x)2
coefficient of x" in the expansion of each of these fractions.
187. nth term = { 10-1 + ( - 1) -1 (n - 1) . 3-2 } x -1.
189. The coefficient of
(n
(n-- 1) (n
( -2)an-2bn -2-...
x = a-(n - 1) an-1 bn-1 + 1.2

+ (n - 1) ab -2 ( -1) -2 + ab -1 ( -1)"-1.
586 Answers to Miscellaneous Examples.

In the special example this reduces to (a -a)"-1 or zero, hence

the sum is ax.

190. See Art 277. w² can be deduced from w by changing

the sign of -3. For second part
(x - ων) (x - w²y) = x² + xy + y².
1
191. If the aliquot part beth
P
the equation between p and n
reduces to 9 ( 3n + 1 )² = ( p + 1 ) (n + 1)². Hence p + 1 must be
3 (3n + 1) 6
a square number greater than 9. Also or 9
n+ 1 n+ 1

must be an integer, thus n = 1, 2, or 5, and p + 1 = 36, 49

or 64.

192. By the method of undetermined coefficients it is possible

to prove that
NP + 1 np
1 +2 + ... + n = ++ 2M12
P -1 + ... (Art. 607).
P+ 1

By raising this to the rth power and equating corresponding

terms we get (p + 1) r = (q + 1 ), ( p + 1) = q + 1,
9 r (r - 1) тр
+
12 8 ( p + 1) -2 12(2 + 1) -1
whence can be deduced that either p = q or p = 1. If p = q,
r = 1, which gives an obvious solution whatever value p and q
1

may have. If p = 1 we can get rr-1 = 2, which is satisfied

by r = 2, and then q = 3.
n
5
N

193. The equation is (1-1) = (1-1

1 +p
). Expand
after taking the fifth root of both sides, keeping terms to the
n 1 1
first power of N and 1+p
and to the second power of 9 since
that is a larger fraction than the others.
194. (1 -x )" ( 1 -x) -" = ( 1 + x + x²)". Equate coefficients
of powers of x on both sides. The last term will be different
 Answers to Miscellaneous Examples. 587

according to whether m is of the form 3r, 3r + 1 or 3r + 2. If

In
23 .
m be of the form 3r the last term is ( -1)". rn -1

195. Equate the given expression to y and solve as an

equation in x. If the values of x be real (a² + b²) y² - 2y³ must
bepositive.
196. From the first equation
u (u + a) v ( v - b ) or น 22 นง
=

a b a চ = u+ v =
น υν 22 น υν ບ
Hence -- = 0. Similarly
Similarly -- =0;
a f
υν นง

whence u²( - )= (6-6), ( - ) = ( -a).

a

197. If r be the radix, x, y, z the digits in descending order

of magnitude, we must have
x- y = y-z - 1 = r - x + z,
whence r = 3 (x−y) +1. Hencer must be of form 3p + 1 .
Also x- y = p = y - z - 1 ; .. x = 2p + z + 1 ;
thus the greatest value of z is p- 1.
198. The residue of any number with respect to the modulus
5 must be 0, 1, 2, 3, or 4. Hence the residue of any square
number must be the same as that of 0, 1, 4, 9 or 16, that is,
must be 0, 1, 4, 4 or 1. Hence the equation x² + y² = z² must
assume the form
5p + a + 5q + b = 5r + c,
where a, b, c can only have the values 0, 1, 4. Thus residue of
a+ b must equal c. The only possible cases are a = 0, b = 1 ,
c = 1 , a = 1 , b = 0 , c = 1 ; a = 1 , b = 4, c = 0 ; a = 4,
b = 1 , c = 0. In every case one of the quantities x, y, z is a
multiple of 5.
199. With respect to the modulus 7 every cube number must
have a residue 0, 1 or 6. Hence the given equation can be put
into the form
7p + a + 7q + b = 7r + c' ,
where a', b'c' have each one of the above three values 0, 1 or 6,
and the residue of a + b' must be c'. This is only possible when
one of the three quantities a', b' or c' is zero.
588 Answers to Miscellaneous Examples.
200. By Fermat's theorem, since 2n+ 1 is a prime number,
if Nbe any number less than 2n + 1 , N2-1 is a multiple of
(2n + 1). Hence either N" - 1 or N" + 1 is a multiple of 2n + 1 .
Thus the residues of the numbers 0", 1", 2", ... (2n)" with
respect to the modulus 2n + 1 are either 0, 1 or 2n. It easily
follows that the residue of the nth power of any numbermust be
0, 1 or 2n. Hence, if three numbers be connected by the
equation a" +b" = c", the residue of one of the three numbers
a, b, c must be zero.

THE END .
 ‫‪1‬‬

‫‪:‬‬
‫دم‬
123,6
1,4.5

3, 3.4
2,
2,
4
36
33 4
4
 THIS BOOK IS DUE ON THE LAST DATE
STAMPED BELOW

AN INITIAL FINE OF 25 CENTS

WILL BE ASSESSED FOR FAILURE TO RETURN
THIS BOOK ON THE DATE DUE . THE PENALTY
WILL INCREASE TO 50 CENTS ON THE FOURTH
DAY AND TO $ 1.00 ON THE SEVENTH DAY
OVERDUE.

OCT 18 1932

OCT 11 1933
FEB 1941 M

SEP 16 1941 M
FEB 2 1935

SEP 30 193
MAR 31 1942 P
MAR 14 1936
SEP 25 1936
MAR 8 1943

DEC 11 1938 25Nov15 V

MAR 8 1943
REC'D LD

FEB 8 1939 NOV 10 1969

LD
REC 20ct62RA
APR 13 1949 SEP 18 1998
LD 21-50m- 8, 32
 YC169553

aldis
45423
QAIS3