HIGHER ALGEBRA A SEQUEL TO ELEMENTARY ALGEBRA FOR SCHOOLS BY H. 8S. HALL, MA, FORMERLY SCHOLAR OF “CHRIST'S COLLEGE, CAMBRIDGE, MASTER OF THR MILITARY AND ENGINEERING SIDE, CLIFTON COLLEGE ; AND 8. R. KNIGHT, B.A., FORMERLY SCHOLAR OF TRINITY COLLEGE, CAMBRIDGE, LATE ASSISTANT-MASTER AT MARLBOROUGH COLLEGE. FOURTH EDITION. Dondon: MACMILLAN AND CO. AND NEW YORK. 1891 { . u (he Right of Translation is reserved.\LIBRARY OF THE _ -ecAND STANFORD JUNiG2 \ UNIVERSITY. 4 A 310% First Printed 1887. Second Edition with corrections 1888. Third Edition revised and enlarged 1889. Reprinted 1890. Fourth Edition 1891.“PREF ACE.. THE present work j is intended as a sequel to our Elementar Algebra for Schools. The first few chapters are devoted t a fuller discussion of Ratio, Proportion, Variation, and th Progressions, which in the former work were treated in a elementary manner; and we have here introduced theorem and examples which are unsuitable for a first course < reading. From this point the work covers ground for the mos part new to the student, and enters upon subjects of speci importance: these we have endeavoured to treat minutel and thoroughly, discussing both bookwork and example with that fulness which we have always found necessary i our experience as teachers. It has been our aim to discuss all the essential part as completely as possible within the limits of a singl volume, but in a few of the later chapters it has been im possible to find room for more than an introductory sketch in all such cases our object has been to map out a suitabl first course of reading, referring the student to special treatise ‘for fuller information. In the chapter on Permutations and Combinations w are much indebted to the Rev. W. A. Whitworth for pe mission to make use of some of the proofs given in hi Choice and Chance. For many years we have used the proofs in our own teaching, and we are convinced Yoeh!vi PREFACE. part of Algebra is made far more intelligible to the beginner by a system of common sense reasoning from first principles than by the proofs usually found in algebraical text-books. The discussion of Convergency and Divergency of Series always presents great difficulty to the student on his first reading. The inherent difficulties of the subject are no doubt considerable, and these are increased by the place it has ordinarily occupied, and by the somewhat inadequate treatment it has hitherto received. Accordingly we have placed this section somewhat later than is usual; much thought has been bestowed on its general arrangement, and on the selection of suitable examples to illustrate the text; and we have endeavoured to make it more interesting and intelligible by previously introducing a short chapter on Limiting Values and Vanishing Fractions. In the chapter on Summation of Series we have laid much stress on the “ Method of Differences” and its wide and important applications. The basis of this method is a well- known formula in the Calculus of Finite Differences, which in the absence of a purely algebraical proof can hardly be con- sidered admissible in a treatise on Algebra. The proof of the Finite Difference formula which we have given in Arts. 395, 396, we believe to be new and original, and the development of the Difference Method from this formula has enabled us to introduce many interesting types of series which have hitherto been relegated to a much later stage in the student's reading. ‘We have received able and material assistance in the chapter on Probability from the Rev. T. C. Simmons of Christ’s College, Brecon, and our warmest thanks are due to him, both for his aid in criticising and improving the text, and for placing at our disposal several interesting and original problems. It is hardly possible to read any modern treatise on Azalytical Conics or Solid Geometry without some know-PREFACE, vii ledge of Determinants and their applications. We have , therefore given a brief elementary discussion of Determi- nants in Chapter xxxuu., in the hope that it may provide the student with a useful introductory course, and prepare him for a more complete study of the subject. The last chapter contains all the most useful propositions in the Theory of Equations suitable for a first reading. The Theory of Equations follows so naturally on the study of Algebra that no apology is needed for here introducing pro- positions which usually find place in a separate treatise. In fact, a considerable part of Chapter xxxv. may be read with advantage at a much earlier stage, and may conveniently be studied before some of the harder sections of previous chapters. It will be found that each chapter is as nearly as possible complete in itself, so that the order of their succession can be varied at the discretion of the teacher; but it is recom- mended that all sections marked with an asterisk should be reserved for a second reading. In enumerating the sources from which we have derived assistance in the preparation of this work, there is one book to which it is difficult to say how far we are indebted. Todhunter’s Algebra for Schools and Colleges has been the recognised English text-book for so long that it is hardly possible that any one writing a teat-book on Algebra at the present day should not be largely influenced by it. At the same time, though for many years Todhunter’s Algebra has been in constant use among our pupils, we have rarely adopted the order and arrangement there laid down; in many chapters we have found it expedient to make frequent use of alternative proofs; and we have always largely sup- plemented the text by manuscript notes. These notes, which now appear scattered throughout the present work, | have been collected at different times during the lost twent HH. A. vviii PREFACE. years, so that it is impossible to make definite acknowledge- ment in every case where assistance has been obtained from other writers. But speaking generally, our acknowledge- ments are chiefly due to the treatises of Schlémilch, Serret, and Laurent; and among English writers, besides Todhunter’s Algebra, we have occasionally consulted the works of De Morgan, Colenso, Gross, and Chrystal. To the Rev. J. Wolstenholme, D.Sc., Professor of Mathe- matics at the Royal Indian Engineering College, our thanks are due for his kindness in allowing us to select questions from his unique collection of problems; and the consequent gain to our later chapters we gratefully acknowledge. It remains for us to express our thanks to our colleagues and friends who have so largely assisted us in reading and correcting the proof sheets; in particular we are indebted to the Rev. H. C. Watson of Clifton College for his kindness in revising the whole work, and for many valuable suggestions in every part of it. May, 1887. H. S. HALL, 8S. R. KNIGHT. PREFACE TO THE THIRD EDITION. Iw this edition the text and examples are substantially the same as in previous editions, but a few articles have been recast, and all the examples have been verified again. We have also added a collection of three hundred Miscel- Janeous Examples which will be found useful for advanced students. These examples have been selected mainly but not exclusively from Scholarship or Senate House papers ; much care has been taken to illustrate every part of the subject, and to fairly represent the principal University and Civil Service Examinations. March, 1839,CONTENTS. CHAPTER I. ratio. PAGE Commensurable and incommensurable quantities . » @ Ratio of greater and less inequality . . 8 1 a_c_e@_ _ (partge”+ren+...\™ 4 aa fF \ po edt erfet...] * . Ay + Og t+ Ogt... $y conan b+, 40, +..45, lies between greatest and least of fractions bo Be Cross multiplication . . e a fl 7 ‘i c 8 Eliminant of three linear eruatons 5 9 Examples I. . . 10 CHAPTER II. proportion. Definitions and Propositions 13 Comparison between algebraical and geometrical d definitions . 16 Case of incommensurable quantities : . 17 Examples II. 19 CHAPTER III. variation. It Ac B, then A=mB . 21 Tnverse variation . 22 Joint variation . 28 It Aa B when C is constant, and AG “2£eo when Bis | constant, then A=mBC . . . . . . » @wW Illustrations, Examples on joint variation 5 2S Examples IIT, . b&x CONTENTS. CHAPTER IV. ARITHMETICAL PROGRESSION, PAGE Sum of n terms of an arithmetical series. : poner 5 - 28 Fundamental formule . . eee) . . . . +. 29 _Insertion of arithmetic means. : : 5 ‘i 5 . . 81 Examples IV. a. . : . . . . 81 Discussion of roots of ant + e- a n- Qs= o e . . . . 33 Examples IV. b. . 5 fi a o c 7 . 35 CHAPTER V. GEOMETRICAL PROGRESSION. Insertion of geometric means 5 c c 5 5 - 38 Sum of n terms of a geometrical series . ; a : a O . 39 Sum of an infinite geometrical series. 5 : _ 5 : » 40 Examples V.a. . : a a » 421 Proof of rule for the reduction ofa , recurring ; decimal o a O » 43 Sum of n terms of an arithmetico-geometric series 5 : O - 44 Examples V.b. . 5 : . : : a D 5 . . 45 CHAPTER VI. HARMONICAL PROGRESSION, THEOREMS CONNECTED WITH THE PROGRESSIONS. Reciprocals of quantities in H. P. are in A. P. 7 7 : 7 » 47) Harmonic mean . re) Formul connecting AM., @: M., H. Mu. 5 fe . . 5 » 49 Hints for solution of questions in Progressions. . a 5 » -49 Sum of squares of the natural numbers : 7 7 . . . 50 _ Sum of cubes of the naturalnumbers . . 2. ww wsCOL Z notation . 5 . : 0 . . 5 - . - 52 Examples VI. a. . . o . 5 » 659 Number of shot in pyramid or ona panera base. : : : f . 54 Pyramid on a triangular base e 5 a : 5 : . Bf Pyramid on a rectangular base. 5 fi I 5 . . . B44 Incomplete pyramid. : S i : : : c . . 55 Examples VI. b. . . . . . . . ; ‘i . 56 CHAPTER VII. scares oF NOTATION, Explanation of systems of notation 5 : a : : o . 57 Examples VII, a. o S : 5 » 59 £xpression of an integral number i ina propoeed scale so . » 59 expression of a radix fraction in a proposed seale. 9. swCONTENTS. ifference between a number and the sum of its digits is divisible yr-1. eee ee of rule for «casting out the nines’? fi c 5 i . of divisibility byr+1 . . . : ‘i 5 “6 . ‘ples VIL. b. . eee . CHAPTER VIII. surps AND IMAGINARY QUANTITIES. aalising the denominator of aalising factor of Ya+{/b ‘e root of a+ /b+/c+r/d : : ‘i . 5 root ofat/b . _ . . . : : . nr) ples VIL a. nary quantities ‘x /-b=- ‘ 4b=0, then a=0, b . tb=c+id, then a=c,b=d . 5 lug of product is equal to eae of moduli . root of a+ib . sofi . . roots of unity ; ltotet= 0. sofw . . . . ples VII. b. a Nbt+Je+/d xi PAGE RSs HAPTER IX. THE THEORY OF QUADRATIC EQUATIONS. Aratic equation cannot have more than two roots ‘tions for real, equal, imaginary roots of roots= — 2 3’ product of roots=<¢ ation of equations when the roots are given . o tions that the roots of a quadratic should be (1). oan in magni. ade and opposite in sign, (2) reciprocals . ‘ a ples IX. a. 2al values of z the expression az be+e has i in general tho same ign as a; exceptions . . . . . S . ples IX. b. . . . . tions of function, variable, rational integral function 5 tion that az?+ 2hry+ by?+ 292+ 2fy+c may be resolved into two near factors . . . . tion that az?+br+e=' 0 and tava =0 may have a common ales Ke a : . _ A . . . . . Rk 888 88 2RRB &xii CONTENTS. CHAPTER X. MISCELLANEOUS EQUATIONS, Equations involving one unknown ee Reciprocal equations. . 7 Examples X. a. Equations involving two unknown quantities Homogeneous equations . . . . . a Examples X.b. . O 5 5 a . . Equations involving several unknown quantities . cS . . Examples X.c. . . . 5 . . Indeterminate equations; ey numerical al oxamples Examples X. d. . : CHAPTER XI. PERMUTATIONS AND COMBINATIONS. Preliminary proposition . . . . . . Number of permutations of n things r at a time . . 5 Number of tombinations of n things r at a time The number of combinations of n things r at a time is equal to the number of combinations of n things n-r at a time Number of ways in which m+n+p+... things can be divided into : classes containing m, n, p, ... things severally Examples XI. a. . Signification of the terme “ike? and‘ unlike’ : Number of arrangements of n things taken all at a time, when P things are alike of one kind, q things are alike of a second kind, &e. . Number of permutations of n things r at a time, when each may be The total number of combinations ofn things a c a a To find for what value of r the expression *C, is greatest Ab initio proof of the formula for the number of combinations of n things r atatime . Total number of selections of ae qtrt+.. . things, whereof 7 pare alike of one kind, g alike of a secund kind, &c. ac O 7 . . Examples XI.b. . : o 0 : c CHAPTER XII. marwematicaL INDUCTION. astrations of the method of proof Product of n binomial factors of the form z +a @xzamples XI. PAGE 100 101 103 104, 106 107 109 111 113 115 115 117 119 120 122 124 125 126 127 127 128 129 131: ; 138 134 . weCONTENTS. xiii CHAPTER XIII. sBINoMIAL THEOREM. POSITIVE INTEGRAL INDEX. PAGE Expansion of (z+a)*, when n is a eee cs . . 187 General term of the expansion. * 139 The expansion may be made to depend aronl ‘the case in n which the frst term is unity . . . . . . . . . . 140 Second proof of the binomial theorem . c e 141 Examples XIII. a. . a 142 The coefficients of terms equidistant from the begining and end are equal. . . . . . + 148 Determination of the greatest term 143 Sum of the coefficients. : . 146 ~ Sum of coefficients of odd terms is equal to sum of coefficients of even terms. . . . . . . . . » 6 Expansion of multinomials e 5 5 146. Examples XII. b. oO eaneee) 147 CHAPTER XIV. BINOMIAL THEOREM. ANY INDEX. Euler’s proof of the binomial theorem for any index 160 General term of the expansion of (1+=)* 153 ‘Examples XIV. a. . D 155 Expansion of (1+ 2)" is only arithmetically intelligible ‘when Z <1 155 The expression (x+y)" can always be ae by the binomial theorem . . . . 0 0 157 General term of the eepaaeionl of a- a" 157 Particular cases of the expansions of (1 -)-" 158 Approximations obtained by the binomial theorem 159 Examples XIV. b. c : 7 161 Numerically greatest term in the expansion of (ta). 162 Number of homogeneous Ce of r dimensions formed out of n letters . . 164 Number of terms in the expansion ‘of a multinomial 165 Number of combinations of n things r at a time, repetitions being allowed 166 Examples XIvV.c.. : 167 CHAPTER XV. MULTINOMIAL THEOREM. General term in the ume of (a+ ba+ca*+dz3+...)?, when p isa positive integer . . General term in the expansion of (eta teats at. * aro w isa rational quantity . . . . . 5 . wu Examples XV. yxiv CONTENTS. CHAPTER XVI. tocarirums. Definition. N=alogeN . Elementary propositions Examples XVI. a. Common Logarithms Determination of the characteristic by inspection . Advantages of logarithms to base 10 Advantages of always keeping the mantissa positive Given the logarithms of all numbers to base a, to find the logarithms to base b. log.bxlogaa=1 Examples XVI. b. . PAGE 175 176 178 179 - 180 181 182 183 183 185 CHAPTER XVII. EXPONENTIAL AND LOGARITHMIC SERIES. Expansion of a*. Series for e . eis the limit of (145); when 1 is infinite Expansion of log, (1+ 2) . Construction of Tables of Logarithms O Rapidly converging series for log, (n + 1) - ee n The quantity e is incommensurable Examples XVII. CHAPTER XVIII. rterEsT AND ANNUITIES. Interest and Amount of a given sum at simple interest. . Present Value and Discount of a given sum at simple interest Interest and Amount of a given sum at compound interest . Nominal and true annual rates of interest Case of compound interest payable every moment Present Value and Discount of a given sum at compound interest. Examples XVIII.a. . c o ‘ : p o 7 Annuities. Definitions. 5 Amount of unpaid annuity, simple interest : Amount of unpaid annuity, compound interest a . Present value of an annuity, compound interest . : : : } amber of years’ purchase Present valine of a deferred annuity, compound interest. Fine for the renewal of a lease ener Ysamples XVIII, b, 187 188 191 192 194 195 195 198 198 199 201 202 203 203 204CONTENTS. xv CHAPTER XIX. requatitizs. PAGE {Elementary Propositions . : . . 208 -‘VArithmetic mean of two positive quantities i is frosty than the geometric 1 mean . 209 ‘The sum of two quantities being given, their product is greatest ‘when they are equal: aa being cont the sum is least when they are equal . 210 ‘The arithmetic mean of 8 number of positive quantities is greater than the geometric mean 5 . . 211 Given sum of a, b, ¢,...; to find the greatest value of am bre? are » 212 Easy cases of maxima and minima. . 5 5 5 5 . 212 Examples XIX. a.. e 218 The arithmetic mean of the me powers of a * aumber of poaitive . quantities is greater than m‘* power of their arithmetic mean, except when m lies between 0 and 1 a . a 5 + 214 Ifa and b.are positive integers, and a>b,: (1+ 2\"> (ust) .. 216 Itz "/ity If 1>2>y>0, Jt > /Pt ee BIT wo > (Sry ee ee 2 317 Examples XIX. b. . 5 . : ‘ . . . . . . 218 CHAPTER XX. LIMITING VALUES AND VANISHING FRACTIONS. Definition of Limit . . 5 . « 220 Limit of a) +0,2+ 4,27 age... is oe is zero. : 222 By taking z small enough, any term of the series Gy +ayE-+0qt* +... may be made as large as we please compared with the sum of all that follow it; and by taking large enough, any term may be made as large as we please compared with the sum of all that precede it . oo + 222 Method of determining the Limits of vanishing fractions . 224 Discussion of some peculiarities in the solution of simultaneous equations . . o . « 226 Peculiarities in the solution of qusarati enuaton : : : . 227 Examples xx, . . a 0 . . 5 . 228 VCHAPTER XXI. CONVERGENCY AND DIVERGENCY OF SERSES. { Cage of terms alternately positive and negative . D : + Series ia convergent if Lim — ig less than 1 . . . 6 oO! ‘a -1xvi CONTENTS. PAGE Comparison of Zu, with on auxiliary series Zu, . . 5 9 - 234 eas peer 1 The auxiliary series + 4 2+ 2 theese e : : o : 5 235 (8 Application to Binomial, Exponential, Logarithmic Series . . . . 287)" { Limits of ee andnz*whennisinfnite . . . . . . 238f Product of an infinite number of factors 0 e 5 eer . aad Examples XXL. a,. q 5 . . . . . . . . 241) weeries is convergent when v-series is convergent, if ra >aa - 243 1 On Series is convergent if Lim Jn ( - 1)f>1 . . I : . 244 Ml Beries is convergent if Lim (n log Fass Ste. ee 45 +1. Series Z¢ (n) compared with series Za"¢(n) . : s c : - 247 on a 1 The auxiliary series = adognP * : a : : a o . 248 Series is convergent if Lim [f= (=. - 1) - 1} tog n] ol... 24g . +1 Product of two infinite series : . . oo. . . . 249 Examples XXI. b.. c c ‘i 9 0 a : G 5 . 252 CHAPTER XXII. UNDETERMINED COEFFICIENTS. If the equation f(t)=0 has more than n roots, it isan identity . » 254 Proof of principle of undetermined coefficients for finite series. . 254 Examples XXII. a. . 5 . 256 Proof of principle of undetermined coefficients for “infinite series . . 257 Examples XXII, b. . . . . . er) : 5 + 260 CHAPTER XXIII. ParTIAL FRACTIONS. Decomposition into partial fractions . . . ae . . 261 Use of partial fractions in expansions . : . . A 5 . 265 Examples XXIII. . 5 . . : : ‘i 5 : o - 265 CHAPTER XXIV. recurRING SERIES. Scale of relation . . . oo . : a . . . 267 Sum of a recurring series. s . . - : 5 ‘ + 269 Generating fanction . . «6 ww te 2808 Examples XXIV, , . . . . : : ee ew BADCONTENTS. xvii = CHAPTER XXV. coNnTINVED FRACTIONS. PAGE }) Conversion of a fraction into a continued fraction : 273 , Convergents are alternately less and greater than the continued fraction 275 Law of formation of the successive convergents . ‘i a 0 - 275 Padi Paadn=(~-W®- se ww ee 28 Examples XXV. a. . « 277 The convergents gradually approximate to the continued fraction. . 278 Limits of the error in taking any convergent for the continued fraction 279 Each convergent is nearer to the continued fraction than a fraction with smallerdenominator .. . . oe cr) » 280 PP 2 ra aq” or <2’, according a> or a , 7 5 G a : + 281 Examples XXV. b. : 5 0 : 5 : o : +» ot. 281 CHAPTER XXVI. INDETERMINATE EQUATIONS OF THE FIRST DEGREE. Bolution of az-by=c . . a : - «284 Given one solution, to find the general solution 5 5 5 eI . 286 Solution of az+by=c oo. er - 286 Given one solution, to find the general solution 0 5 5 . . 287 Number of solutions of az+by=c. . . . . . + 287 Solution of az+by+cz=d, darvyteond : . a . + 289 Examples XXVI. . : 5 5 : o 5 . 290 CHAPTER XXVII. RECURRING CONTINUED FRACTIONS. Numerical example . . . . 292 A periodic continued fraction i is equal to a quadratic surd o : - 293 Examples XXVII. a. . . . : » 204 Conversion of a quadratic surd into a continued fraction 5 : . 295 ‘The quotients recur. 5 02 6 of 9 oer) The period ends with a partial quotient “20, oe . + 297 The partial quotients equidistant from first and last are vega . + 298 The penultimate convergents of the aan 0 5 . . ~ 299 Examples XXVIL b. : : o . . . . . 30: CHAPTER XXVIII. INDETERMINATE EQUATIONS OF THE SECON DEGREE. Solution of ar7+ 2hay + by" + 29x + 2fy+c=0 : . : : o The equation z4- Ny?=1 can always be solvedxviii CONTENTS. PAGE Bolution of s*-Ny?=-1 . . . : 5 5 : : . 305 General solution of 2?-Ny?=1 . z c c ‘i . : . 3806 Solution of z?-nty?=a c . 5 a : : : . 808 Diophantine Problems... . . . : : i ‘ . 309 Examples XXVIIL. . 5 : . . 5 oO . . Bil CHAPTER XXIX. suMMATION OF SERIES. Summary of previous methods. . . . . . . . 812 1, the product of n factors in A. P. . ‘ : 5 . 814 u, the reciprocal of the product of n factors i in A. P, c C . 816 Method of Subtraction. . . ee ee 818 Expression of u,, as sum of faotoriala . . 48 . . « 818 — Polygonal and Figurate Numbers . : Fi c . : . - 819 Pascal’s Triangle . . . . oe . . : _ . 3820 _ Examples XXIX,a, . nr) . : . . : 5 - 3821 Method of Differences . : . 322 Method succeeds when 1, is a , rational integral fanetion ofn n _ 826 If a, is 8 rational integral function of n, the series Za,2* is a recurring series . . . . . . . . « 827 Farther cases ‘of recurring series. 7 . : 5 : . 829 Examples XXIX.b, «te See ee BBB Miscellaneous methods of summation . soe . . a » 334 Sum of series 1° +2743" +...+n". cee se - « 836 - Bernoulli's Numbers . . : see ee 887 Examples XXIX.c, . és 5 : : : ji : 5 . 338 CHAPTER XXX. THEORY OF NUMBERS. Statement of principles . I : 5 : c : 5 I . 341 Number of primes is infinite . 5 G 5 . - 342 No rational algebraical formula can represent primes only 5 ‘ . 343 A number can be resolved into prime factors in only one way =. - 342 Number of divisors of a given integer . . . . 348 Number of ways an integer can be resolved into two factors . . +, + 343 Sum of the divisors ofa given integer . . . 54 . . 344 Highest power ofa prime contained in Jn... / ww. 845 Product of r consecutive integers is divisible by |r . . B45 Fermat's Theorem N-1—1=. =™() ae is prime and N prime stop 347 Examples XXX.a, . ; - 848 nition of congruent . 5 eee ee BRD eeCONTENTS. xix } PAGE ‘$f a is prime to 5, then a, 2a, 3a,...(b-1)a@ when divided by b leave different remainders. 7 5 7 7 a . . . 350 G(abed...)=$(a) (I) be) OA) ee BER 1 1 1 o(w)=N (1-3) ) (1-3)... See ee 858 Wilson’s Theorem: 1+|p~1=M (p) where pisa prime . a . 864 A property peculiar to prime numbers . a a a a : . 854 Wilson’s Theorem (second proof) . 5 . . 5 . 5 . 3855 Proofs by induction . . : oo : . . a . 356 Examples XXX. b. . : S 5 . 5 : 5 . . 3867 CHAPTER XXXI. THE GENERAL THEORY OF CONTINUED FRACTIONS, Law of formation of successive convergents . . . . . . 359 1 bt hae a definite value if Lim Smt Set oe ee BER qt at The convergents to a a ane BFE. positive proper fractions in ascend- ing order of magnitude, ifa,t1+b, . e . 863 General value of convergent when a,, and b,, are constant D . 864 Cases where general value of convergent can be found . co a . 365 a te ini it On a4 af is incommensurable, if os <1. - fi : . 366 Examples XXXL a. © - zi - c _ - 367 Series expressed as continued fractions . c . 7 7 . 369 Conversion of one continued fraction into another : : . 871 Examples XXXI.b. . : : c aC . a 7 - . 372 CHAPTER XXXII. prozasinity. Definitions and illustrations. Simple Events . : : : . 373 Examples XXXII.a, . : . . : . : : ‘i . 3876 Compound Events : fi . 377 Probability that two independent events will both happen i is pp’ : . 878 The formula holds also for dependent events s : - 3879 ce of an event which can happen in matually exclusive ways . 881 Examples XXXII. b. : . eanr) I . 383 Chance of an event happening exactly r ‘times inn . trials ‘ : . Wo Expectation and probable value. : 7 5 : . : . IW “Problem of points”. anne . . an BWxx CONTENTS. Examples XXXII.c. a 0 : : : 5 : . Inverse probability a . Fi : 7 ‘i FI c Statement of Bernoulli’s Theorem . : : ; 6 —_PrPr Proof of formula eS (eP) 5 : 1 0 ‘I c Concurrent testimony . 7 0 . Traditionary testimony . . 7 : 0 : 5 . O Examples XXXII. d. . . : 5 : _ . Local Probability. Geometrical methods 5 5 : : ‘i c Miscellaneous examples c 5 Examples XXXIL e. CHAPTER XXXIII. pererminants. Eliminaut of two homogeneous linear equations . 7 : _ - 409 Eliminant of three homogeneous linear equations . o eee +» 410 Determinant is not altered by interchanging rows and columns 0 . 410 Development of determinant of third order . D 411 Sign of a determinant is altered by interchanging two adjacent rows or columns . . . 412 If two rows or olunnets are identical, the determinant vanishes e . 412 A factor common to any row or column may be placed outside. . 412 Cases where constituents are made up of a number of terms. . » 418 Reduction of determinants by simplification of rows or columns . - 414 Product of two determinants . . . : 0 . . - 417 Examples XXXIII.a. . . . : . 419 Application to solution of simultaneous ‘equations. . . . - 422 Determinant of fourth order . : 0 5 zi 5 : 5 - 428 Determinant of any order. zi 7 : : amr : . 423 Notation Za bcd, . ‘i 0 _ _ ‘i . : + 1. 495 Examples XXXIII.b. . 7 c ‘ 0 : : 0 . 427 CHAPTER XXXIV. MiscELLANEOUS THEOREMS AND EXAMPLES, Review of the fundamental laws of Algebra . 5 5 Fi 0 . 429 F(z) when divided by 2-a leaves remainder f(a) . : . . « 482 Quotient of f (x) when divided by z-a . . O : : - 483 — Method of Detached Coefficients . . : a . i : - 434 | Horner’s Method of Synthetic Division . : : : : . 434 Symmetrical and Alternating Functions . 0 . c 5 . 435 Zxamples of identities worked out : ee 48 List of nsefal formule . : ; ee ARS.' CONTENTS. xxi PAGE Examples XXKIV.a, . . . 5 . - 438 Identities proved by properties of cube roots of anity a : 7 . 440 Linear factors of a? +08 +c - 8abe . . ar) . 6 . 441 Value of a®+0"+e"whena+b+c=0 . . . .« « « . 448 Examples XXXIV.b. . . . . . . . . . - 442 Elimination... : ee ee Elimination by symmetrical functions o ao a 7 a c . 444 Euler’s method of elimination =. . S O a 5 « 445 Sylvester's Dialytic Method. 6 ww ewww Bezout’s method . . eae : . . - 446 Migcellaneous examples of elimination . 5 5 c a 6 ' . 447 Examples XXXIV.c. . : z 5 s 5 5 8 - . 449 CHAPTER XXXV. THEORY OF EQUATIONS. Every equation of the n‘* degree has n roots and no more. a . 452 Relations between the roots and the coefficients . . . . . 452 These relations are not sufficient for the solution. . . . « 464 Cases of solution under given conditions . 5 O : a . 454 Easy cases of symmetrical functions of the roots . 5 : a . 455 Examples XXXV. a. e : 5 O 5 : . 456 Imaginary and surd roots occur in pairs . f 0 ce . . 457 Formation and solution of equations with surd roots . . « 458 Descartes’ Rule of Signs . 2 5 e G9 5 . 459 Examples XXXV.b. . . 5 : : . _ . 460 Value of f(z+h). Derived Functions . . . . . . . - 462 Calculation of f(2+h) by Horner's process . . . . . - 468 J (z) changes its value gradually . . 464 If f(a) and f(b) are of ce signs, f. (y= 0 acl a root between aandb. . . . : » 464 An equation of an odd degree has ¢ one real root . Smee 465 An equation of an even degree with its last term nce has two real roots. . - 465 It f (z)=0 has r roots eqnal to a, f " @)= 0 hes re 1 roota equal to a. 466 Determination of equal roots . . a 5 : e . 467 f@)_ 1 1 1 Fiayzaateseteeet te 888 Sum of an assigned power of the roots . : c 7 : 5 . 468 Examples XXXV.c. . .. ao A . D : : » AW, Transformation of equations. . a : . AT Equation with roots of sign opposite to ‘those of f (x)= “0 a : . ATA uation with roots multiples of those of SF (x) =0 - ~ At yxxii CONTENTS. Equation with roots reciprocals of those of f (z)=0 Discussion of reciprocal equations . O Equation with roots squares of those of f (= Oo. Equation with roots exceeding by h those of f (z)=0 Removal of an assigned term. o Equation with roots given fanctions of those of (= 0 Examples XXXV.d. . a . Cubic equations, Cardan’s Solution : G 5 7 Discussion of the solution . . : . . Solution by Trigonometry in the irreducible ¢ case . Biquadratic Equations. Ferrari’s Solution . Descartes’ Solution . eee Undetermined multipliers . Discriminating cubic; roots all real Solution of three simultaneous equations —— Examples XXXV. e. Miscellaneous ee Answers ye aan toon tapith aeHIGHER ALGEBRA. CHAPTER I. RATIO. ‘1. Derivrtion, Ratio is the relation which one quantity bears to another of the same kind, the comparison being made by considering what multiple, part, or parts, one quantity is of the other. The ratio of A to B is usually written 4: B, The quantities A and B are called the terms of the ratio. The first term is called the antecedent, the second term the consequent. 2. To find what multiple or part A is of B, we divide A by B; hence the ratio A : B may be measured by the fraction $ , and we shall usually find it convenient to adopt this notation. In order to compare two quantities they must be expressed in terms of the same unit. Thus the ratio of £2 to 15s. is measured . 2x20 8 by the fraction ip 5° Nore. A ratio expresses the number of times that one quantity con- tains another, and therefore every ratio is an abstract quantity. 3. Since by the laws of fractions, a ma 5 mb? it follows that the ratio a : 6 is equal to the ratio ma \ mb that is, the value of @ ratio remains unaltered if the antecedent and the consequent are multiplied or divided by the same quantity. 4 FT. H, A rV2 HIGHER ALGEBRA. 4, Two or more ratios may be compared by reducing their equivalent fractions to a common denominator. Thus suppose = ang 2. 5 bby’ yy by the ratio a : 6 is greater than, equal to, or less than the ratio a : y according as ay is greater than, equal to, or less than ba. a:band2: y are two ratio, Now 3 hence 5. The ratio of two fractions can be expressed as a ratio of two integers. Thus the ratio ¢ 5 is measured by the bid a fracti & ad . we . raction a) OF 53 and is therefore equivalent to the ratio d ad : be. 6. If either, or both, of the terms of a ratio be a surd quantity, then no two integers can be found which will eaaetly measure their ratio. Thus the ratio ,/2:1 cannot be exactly expressed by any two integers. 7. Derinition. If the ratio of any two quantities can be expressed exactly by the ratio of two integers, the quantities are said to be commensurable; otherwise, they are said to be incommensurable. Although we cannot find two integers which will exactly measure the ratio of two incommensurable quantities, we can always find two integers whose ratio differs from that required by as small a quantity as we please. 1 2. ve a AB _ POBO68 559017... V5 _ 559017 559018 - and therefore “ft 7 1000000 and < 1000000’ so that the difference between the ratios 559017 : 1000000 and /5 : 4 is less than (000001. By carrying the decimals further, closer approximation may be arrived at. 8. Derinition. Ratios are compounded by multiplying to gether the fractions which denote them; or by multiplying to gether the antecedents for a new antecedent, and the consequent for a new consequent, 4zanple, Find the ratio compounded of the three retion a 2a: 3b, Gab : 58, cra) RATIO. 3 A - 2a 6ab_¢ The required ratio= 55 x pax se _4a Re 9. Derinition. When the ratio a:b is compounded with itself the resulting ratio is a’ : b*, and is called the duplicate ratio of a:b. Similarly a’: b° is called the triplicate ratio of a:b. Also at : 64 is called the subduplicate ratio of a : b. Examples. (1) The duplicate ratio of 2a : 8b is 4a? : 987. (2) The subduplicate ratio of 49 : 25 is 7: 5. (3) The triplicate ratio of 2x : 1 is 823 : 1. 10. Derinrrion. A ratio is said to be a ratio of greuter inequality, of less inequality, or of equality, according as the antecedent is greater than, less than, or equal to the consequent. ll. A ratio of greater inequality is diminished, and a ratio of less inequality ts increased, by adding the same quantity to both tts terms. Let i be the ratio, and let 2+* be the new ratio formed by b+a adding x to both its terms. Now @ ate _ av—be 6 b+au b(b+z) _2(a-2), “ebre)) and a—6 is positive or negative according as a is greater or less than b. 7 @ ata Hence if a > 8, Bbee? ; a ata, and if a kb, 5 sk; o. a, > kb; 5, and so on; ., by addition, A, +O, t,t veers +a,>(b, +b, 48+ 2. Fb) KS G++ Oy+ rere +O, 7 oo * 548, 48,4 ne +b,” i that is > 5°. Similarly we may prove that +4, +0, +». 6, +0, +0, + where Z is the greatest of the given fractions. In like manner the theorem may be proved when all the denominators are negative. 15. The ready application of the general principle involved in Art, 12 is of such great value in all branches of mathematics, that the student should be able to use it with some freedom in any particular case that may arise, without necessarily introducing an auxiliary symbol, at eee Example 1. Ue Syenaadanb ate , “rove that atyte_x(y+a)t+y (e+2)+2(2+y) atb+c 2 (ax + by +cz)RATIO. . 0 sum of numerators Each of the given fractions = of denominatora Jenominat att ty te FDR be Te teen eeeeeeentenee (1). Again, if we multiply both numerator and denominator of the three given fractions by y+2,2+2, z+y respectively, . a (y +2) = y(z+2) = z(z+y) each fraction= 73) @4e=a) ~ Fa) (Faq) ~ (ery) (@4d—<) sum of numerators ~ gum of denominators a (y+2)+y (e+2) +2 (e+y) 2az + Qby + 2cz .. (2). «*. from (1) and (2), tyte 2 (ytaty (eta)te(z+y) at+b+c_ 2 (ax + by +¢z) = y —— Esample 2, Ut fae a aeee e 1 m Prove that 2 (by+e2—az)~ y (ce-+az—by)— ier) cz)" 2 y 2 7 m™ n Wehave = pypnenla nerlamb latmb—ne ye a ~ 2a =two similar expressions; nytme lene _mo+ly a@ ob ¢° Multiply the first of these fractions above and below by z, the second by y, and the third by z; then nay tmae _ lyst ney _ moet lye 1 = (by+cz— Twa) = ye baee by) # Gaatysay8 HIGHER ALGEBRA. 16, If we have two equations containing three unknown quantities in the first degree, such as az+by+e2=0 ag+by+ez=0... we cannot solve these completely; but by writing them in the form a, () +, () +e, =0, a, (2) +b, () +¢,=0, we can, by regarding = and ¥ as the unknowns, solve in the. ordinary way and obtain w _ dyc, — bye, Y _ Cry C40, | % ab,-a,b,’ 2 a,b,—-a,b,’ or, more symmetrically, —*_._ ¥ = bc,- 8c, ¢4,-¢,@, a,b,- a,b, °°" (3). . It thus appears that when we have two equations of the t; represented by (1) and (2) we may always by the above form write down the ratios x:y:2 in terms of the coefficients of the equations by the following rule: ‘Write down the coefficients of x, y, z in order, beginning with those of y; and repeat these as in the diagram. . b, ¢, a, 6, b, C, a, b, Multiply the coefficients across in the way indicated by the arrows, remembering that in forming the products any one obtained by descending is positive, and any one obtained by ascending is negative. The three results bye, ,¢,, 64, —6,0,, 4,5, — a,b, are proportional to 2, y, 2 respectively. This is called the Rule of Cross Multiplication,RATIO, Example 1, Find the ratios of z : y : z from the equations Ja=4y +82, 82=122+1ly, By transposition we have 72 - 4y - 8z=' 122+ 1ly -32=0. Write down the coefficients, thus -4 -8 7 -4 ll -3 2 1, whence we obtain the products (+4) x(-8)-11x(-8), (-8)x12-(-8)x7, 7x11~12x(-4), or 100, -75, 125; oY 100° -75 125’ i tly _* that is, G7 cat Example 2. Eliminate 2, y, z from the equations act by +ez=0.. Agr + bay + gz Ago + Dgy +0 g2 =O... ceeccessstreeeeeeeeeeeee (8). From (2) and (3), by cross multiplication, z y byey— dye Cog — Cgg 2 : * aby a5, jenoting each of these ratios by k, by multiplying up, substituting in (1), ind dividing out by k, we obtain % (byey— Dyes) + dy (Cyi4g — C504) +4 (Agby — A403) =0. This relation is called the eliminant of the given equations, Example 3. Solve the equations az+by+cz=0. ft yt z= bea + cay + abz=(b—¢) (6-4) (A—D)..eeererererreerrerrers (3). From (1) and (2), by cross multiplication, = =_Y _ *# _, - b=ee-aa—b ? SUPPOBES “. e=k(b-c), y=k(c—a), 2=k (a—b). Substituting in (3), k {be (b-¢) +. ca (c — a) + ab (a ~ b)} =(b-c) (ca) (a-d), k {-(b-¢) (¢- a) (a -1)} = (0c) (¢-a) (2-2) we kal z=c-b, y=a-¢,z=b-a. whence10 HIGHER ALGEBRA, 17. If in Art, 16 we put 2 =1, equations (1) and (2) become azt+by+c,=0, a,e+by+e,=0; and (3) becomes ee a be, 7 be, - oa, 7 Ca, - a,b, 7 a, , Oi = 240, ™ “aha, Hence any two simultaneous equations involving two un- knowns in the first degree may be solved by the rule of cross multiplication. Ezample. Bolve 52 -8y-1=0, «+2y=12. By transposition, 5a-3y- 1=0, w+2y-12=0; ye “* 3642 -1+60 1043’ whence ont, ya, EXAMPLES, I. 1, Find the ratio compounded of (1) the ratio 2a : 36, and the duplicate ratio of 95? : ab. (2) the subduplicate ratio of 64 : 9, and the ratio 27 : 56, 2 (3) the duplicate ratio of ee p Ghd , and the ratio 3ax : Qby. 2. Ifv+7 : 2(¢+14) in the duplicate ratio of 5 : 8, find 2. * 3, Find two numbers in the ratio of 7: 12 so that the greater exceeds the less by 275. 4, What number must be added to each term of the ratio 5 ; 37 to make it equal to 1 : 3? & Ife: y=83 ; 4, find the ratio of 7r—4y : Br+y. 6 Ifis (224 —y3) = Tay, find the ratio of 2 + y.RATIO. 11 . @ ie e 1. If Bacup Qatb? + Baret—5etf a “BO5+ Bb? — BFF ~ BA 8 If $737 3 prove that S is equal to b'c+ a4 + Bcd?” ee oe qt+r-p r+p-q ptq-r’ shew that ‘(q-1) a+(r-p)y+(p-g)2=0. to, If 2 —24* _” find the ratios of x: y: & a-z 2 y’ ll. If yte atu _ aty shew that _2@+9+2) _ (bte)w+ (cta)y+(atd)z at+bte be+ca+ab 12, If ey Lt show that St@ , #40 | tte _(otyte)P+(atb+e) apa * AEB FLA” (ety tet (arbre? 13, If ay tSe— 22+ 20— y tyne z shew that oe = 4. 4 2b4+2c-a I+2a—b Ba+2b—c° WA, Tf (a2 4 b2 +o) (x8 +9229) = (an+dy + c2)%, shew that w@ra=y:b=z:0 WY 45, If U (my +nz—Le) =m (nz+Le— my) =n (le + my — 02), ytene _atany _ety-2 om TH prove 16. Shew that the eliminant of ax+cy+bz=0, or+by+az=0, br+ay+cz=0, in @+B+40-3abc=0. 17, Eliminate , y, 2 from the equations arthy+gz=0, he+by+fe=0, gurfy+c=0.12 HIGHER ALGEBRA. 18, If w=eythz, y=az+cx, z=br+ay, a yf a ee 1a 1-8 7 1ra° 19. Given that a(y+z)=2, b(z+2)=y, e(x+y)=2, prove that be+ca+ab+2abe=1. Solve the following equations : 20. 32—4y+7z= 0, 21, aty= 4 Qe-y-2= 0, 3r-Qy+1liz= 0, 328 - Y+8=18. w+ 3y8 + 228 = 167. 22, ‘Ty2 + Ben day, 23, 3uvt—2y?+528=0, Qlyz - 32r=42y, ‘724 — 3y?— 1528=0, w+2y+3z=19. 5a—4y+72=6. t m n m4. If Yan ns b= sca sc= Jame n Jann - Jerwe * Sera” t m n shew that< — ———— = —————____ = —________,, (a—b)(e—ab) (b—c)(a—WVbc) (c—a) (b- Vac) Solve the equations: 25. an+ by+cz=0, bex+ cay + abz=0, wy2+ abe (aix + by + 8z)=0, 26. az+by+ce=aiz+ Py+c2=0, w@+y+2+(b—c)(c—a) (a—b)=0. 27. If a(y+s)=2, b(2+2)=y, c(e+y)=z, at x Ea a(1—be) ~b(l=ca a) ~¢(1—ab)" 2 If axthy+gz=0, het byt+fe=0, grtfy+cz=0, prove that prove that q) ef 8 be— ft capi ab (2) (bo-F?) (ca - 9*) (ab— 23) = ( fy — ch) (gh—af)) (hf - bg).CHAPTER II. PROPORTION. 18. Derinition. When two ratios are equal, the four quantities composing them are said to be proportionals. Thus if ; =) then a, b,c, d are proportionals, This is expressed by saying that a is to b as c is to d, and the proportion is written :d; or id. The terms a and d are called the extremes, b and c the means. 19. If four quantities are in proportion, the product of the extremes is equal to the product of the means. Let a, b, c, d be the proportionals. Then by definition whence ad = be. Hence if any three terms of a proportion are given, the fourth may be found. Thus if a, ¢, d are given, then b = = Conversely, if there are any four quantities, a, b, c, d, such that ad = be, then a, 6, c, d are proportionals ; a and d being the extremes, b and c the means; or vice versa, 20. Derinition. Quantities are said to be in continued proportion when the first is to the second, as the second is to the third, as the third to the fourth; and so on. Thus a, b, c, d,...... are in continued proportion when a a 6 oe aor14 HIGHER ALGEBRA, 1f three quantities a, 6, ¢ are in continued proportion, then a:b=b:c; / ac =". [Art. 18.] Tn this case 6 is said to be a mean proportional between a and ¢; and c is said to be a third proportional to a and 6. 21. If three quantities are proportionals the first is to the third in the duplicate ratio of the first to the second. Let the three quantities be a, b, c; then ; =. Now a_a,b e be mom amar “oO that is, aicaa':B. Tt will be seen that this proposition is the same as the definition of duplicate ratio given in Euclid, Book v. 22. Ifa:b=e:d and e: f=g:h, then will ae : bf=cg : dh. For po gand oa $s uo RF ah or ae: bf=cg: dh. Cor. If a:b=e:d, and b:aadiy, then Gix=ery. This is the theorem known as ex equali in Geometry. 23, If four quantities a, 6, c, d form a proportion, many other proportions may be deduced by the properties of fractions. The results of these operations are very useful, and some of them are often quoted by the annexed names borrowed from Geometry. .PROPORTION. 15 (1) If a:b=c:d, then b:a=d:c. [Znvertendo.] For 2 therefore 1 + ba a that is or 6: (2) If a:b=0:4d, the For ad = be; therefore e =bid, [Alternando.] that is, or a:c=b:id. (3) Ifa:b=c:d, then a+b:b=c+d:d. [Componendo.] For =p therefore Ftla=S+hs F a+b_c+d, that is Sr or atb:b=ct+d:d. (4) Ifa:b=e:d, then a-b:b=c-d:d, [Ditidendo.] a c For i = 53 therefore #-1=5-1; f a-b_ c-d that is, er ae or a-b:b=c-d:d. (5) If a:b=e:d, then a+b:a-b=c+d:c-d at+b_c+d For by (3) “yg -~b_c-d and by (4) +r pone a+b c+d . by division, ab era? or a+b:a-b=c+d:e-d. This proposition is usually quoted as Componendo and Div 0. Several other proportions may be proved in a similar way.16 HIGHER ALGEBRA. 24. The results of the preceding article are the algebraica equivalents of some of the propositions in the fifth book of Euclid and the student is advised to make himself familiar with then in their verbal form. For example, dividendo may be quoted a follows : When there are four proportionals, the excess of the first abov the second ia to the second, as the excess of the third above th fourth is to the fourth. 25. ‘We shall now compare the algebraical definition of pre portion with that given in Euclid. Euclid’s definition is as follows : Four quantities are said to be proportionals when if any equt multiples whatever be taken of the first and third, and also an; equimultiples whatever of the second and fourth, the multiple o the third is greater than, equal to, or less than the multiple of th fourth, according as the multiple of the first is greater than, equa to, or less than the multiple of the second. In algebraical symbols the definition may be thus stated : Four quantities a, 8, c, d@ are in proportion when po=g according as pa = qb, pand q being any positive integers whatever I. To deduce the geometrical definition of proportion fron the algebraical definition. Since 5 = 5) by multiplying both sides by?, we obtain’ pa _ pe, gb gd? hence, from the properties of fractions, pee q@ according as pa=qb, which proves the proposition. II. To deduce the algebraical definition of proportion frox the geometrical definition. Given that pe = gd according as pu = 6, to prove ¢ 7 oaPROPORTION, 17 1S is not equal to oa one of them must be the greater. Suppose oo ; then it will be possible to find some fraction £ which lies between them, g and p being positive integers. P Hence Seles saecereeteecees 1 : a7) Q), and See ccecccceetetaneteees 2). oe @) From (1) pa>qgo; from (2) peG, therefore G>H; that is, the arithmetic, geometric, and harmonic means between any two positive quantities are in descending order of magnitude, 66. Miscellaneous questions in the Progressions afford scope for skill and ingenuity, the solution being often neatly effected by some special artifice. The student will find the following hints useful. 1. If the same quantity be added to, or subtracted from, all the terms of an A P., the resulting terms will form an A.P. with the same common difference as before. (Art. 38.] 2. If all the terms of an A.P. be multiplied or divided by the same quantity, the resulting terms will form an A.P., but with a new common difference, [Art. 38.] 3. If all the terms of a G.P. be multiplied or divided by the same quantity, the resulting terms will form a G.P. with the same common ratio as before. [Art. 51.] 4. Ifa, b,¢,d... are in G.P., they are also in continued pro- portion, sizice, by definition, abe zye-= Sinvolad Gaaae sim Conversely, a series of quantities in continued proportion may be represented by a, x7, ar",...... 5 Example 1. If a’, 27, c? are in A.P., shew that b+c, ct+a, a+b are in H.P, By adding ab +ac + bc to each term, we see that @+abt+tactbe, b?+bat+be+ac, c?+ca+cbh+ab are in A.P.; thatis (a +B) (a+¢), (b-+e)(b-+a), (c+) (¢+b) are in A. P. «'., dividing each term by (a +b) (b+¢) (¢+a), 1 1 1 F ye’ aya’ appre AP.; that is, b+¢,c+a, a+b arein H.P. AHA 450 HIGHER ALGEBRA, Example 2. If 1 the last term, d the common difference, and s the sum of n terms of an A. P. be connected by the equation 8ds=(d +203, prove that d=2a. Since the given relation is true for any number of terms, put n=1; then a=les, Hence by substitution, 8ad=(d + 2a)*, or (4-2a)*=0 o. d=2a, Ezample 8. If the p*, q",r,s* terms of an A.P. are in G. P., shew that P-% 4-7, 7-8 are in G.P. With the usual notation we have [Art. 66. (4)]; .. each of these ratios _ {a+(p-1) 4} -{a+(q-1) 4} _ {a+(q-1) d} - {a+(r-1) 4} {a+ (q-1) 4} - {a+ (r—1) a} {a+ (7-1) a} - {a4 (6-1) a} Hence p-q, 9-1, 7-8 are in G.P. 67. The numbers 1, 2, 3,...... are often referred to as the natural numbers; the n‘” term of the series is n, and the sum of the first n terms iss (n +1). 68. Zo find the sum of the squares of the first n natural numbers. Let the sum be denoted by S; then Sal 4 24 34 000... +n'. We have n? — (n—1)'= 3n*- 3n+1; and by changing n into »— 1, (n- 1)? — (n — 2)° = 3(n— 1)? -3(n-1) +1; similarly (n—2)°— (n — 3)? = 3(n — 2)*— 3(n— 2) +1; 3° 27 =3.3°-3.34+1; -1?=3.2°-3.24+1; _ 1-0? =3.17-3.14+1.THE NATURAL NUMBERS. 51 Hence, by addition, n? = 3(17 + 274+ 3% +... +m) —-3(142434+...4¢n) tn =35- 2+), n+ n(n) =n(n+1)(n—1 +3); _ (n+) (2n4+1) ae. o. 88=n— a 69. To find the swum of the cubes of the first n natural numbers. Let the sum be denoted by S; then Sa 14243 +. tn% We have n‘— (n—1)* =4n? — 6n? +4n-1; (n—1)* - (n- 2)* =4 (n— 1)? - 6 (n— 1)? +4 (n—1)-1; (n— 2) — (n— 3)*=4 (n— 2)? - 6 (n—2)° +4 (n—2)-1; 3*— 24= 4. 37-6. 3°44.3-1; 8 14=4,2°-6,.979+4.2-1; 1*-0*=4,1°-6.1°+4.1-1. Hence, by addition, n= 45-6 (194+ 2%+...4n%)+4(1424+.. tn)—n; o 4S=nt+nt6 (184+ 2°40. 4n")-4(1424...4¢0) =nt+ntn (n+ 1) (204 1) —2n(n+1) =n(n+1) (nt —-n+14+2n4+1-2) =n(n+1)(n*+n); ; gam (ntl _ {ree Dy at coerce eno aeaiia Thus the sum of the cubes of the first n natural numbers is equal to the square of the sum of these numbers. The formule of this and the two preceding articles may be applied to find the sum of the squares, and the sum of the cubes of the terms of the series C @ 240, a420,......... 4—252 HIGHER ALGEBRA. 70. In referring to the results we have just proved it will be convenient to introduce a notation which the student will fre- quently meet with in Higher Mathematics. We shall denote the series 14+24+3+4...+0 by 3n; 174294 3% +... +n" by Sn’; 1°4+2°4+3°+... 40" by Sn’; where % placed before a term signifies the sum of all terms of which that term is the general type. Example 1, Sum the series 1.24+2.3+43.4+...t0 n terms, n term= eee =n*+n; and by writing down each term in a imine form we shall have two columns, one consisting of the first n natural nambers, and the other of their squares. .. the sam=2Zn*+ In n(n+1)(2n+1) n(n+1) sae __ %(n+1) (n+2) 2: Ezample 2, Sum to n terms the series whose n™ term is 2*-! + 8n° - 6n*, Let the sum be denoted by S; then S=Z 2" 48En3-6En" 2-1, San 41)? _ Gn (n+) Qn+]) 2-1 4 6 ~ Lin (n+1){3n(n+1)- (an+1)} =2"~14n (n+1) (2n?-1). EXAMPLES. VI. a. 1, Find the fourth term in each of the following series: (1) 2 8, 34,.. (2) 2, 2$, 3,... (3) % 2b, 3h... 2, Insert two harmonic means between . and ne 3 Insert four harmonic means between 2 3 and 7 5EXAMPLES ON THE PROGRESSIONS, 53 4. If 12 and 9% are the geometric and harmonic means, respect- ively, between tao tunsbers find them. 5. If the harmonic mean between two quantities is to their geo- metric means as 12 to 13, prove that the quantities are in the ratio of 4 to 9. 6. Ifa, 6, ¢ be in H. P., shew that . a:a—b=ate:a-c 7. If the m" term of a H. P. be equal to n, and the n term be equal to m, prove that the (m+n) term is equal to ws : 8. If the p", ¢**, r** terms of a H. P. be a, b, ¢ respectively, prove that "q-n)be+ (r—p) cat (p—g) ab=0. , 9. If6 is the harmonic mean between a and ¢, prove that 1 1 11 b-atb—cmate: Find the sum of n terms of the series whose n“* term is 10, 3n?-n. 11. n43 n, 12, n(n+2). 13, n*(2n43). 14, 38-2, 15. 3(4"42n?)—4n3, 16. If the (m+1)", (n+1)*, and (r+1)" terms of an A. P. are in G. P., and m,n,r are in H. P., shew that the ratio of the common difference to the first term in the A. P. is ~2. 17. If 1, m, mare three numbers in G. P., prove that the first term of an A. P. whose 2, m, and x“ terms are in. P. is to the common difference as m+1 to 1. 18. If the sum of n terms of a series be a+bn-+cn’, find the n term and the nature of the series. 19, Find the sum of x terms of the series whose n“ term is An (n3 +1) —(6n?+1). 20. If between any two quantities there be inserted two arithmetic means A,, A,; two geometric means @,, @,; and two harmonic means H,, Hy; shew that GG, : H,H,=A,+A, : H,+Aj. 21. If p be the first of » arithmetic means between two numbers, and g the of x harmonic means between the same two numbers, prove that the value of g cannot lie between p and a Dp. 22. Find the sum of the cubes of the terms of an A. P., and chew that it is exactly divisible by the sum of the terms.54 HIGHER ALGEBRA. Pires or SHor anp SHELLS. 71. To find the number of shot arranged in a complete pyramid on a square base. Suppose that each side of the base contains n shot; then the number of shot in the lowest layer is n*; in the next it is (n—1)*; in the next (n—-2)*; and so on, up to a single shot at the top. wo. Santt(n-1)+(n—-2)+...41 = n+ Gn+) [Art 68,] 72. To find the number of shot arranged in a complete pyramid the base of which is an equilateral triangle. Suppose that each side of the base contains n shot; then the number of shot in the lowest layer is n+ (m—1)+(m-2)+...... +1; that is, = (us 1) or pnt +n). In this result write n — 1, —2,...... for m, and we thus obtain the number of shot in the 2nd, 3rd,...... layers. o. S=}h (Sn? + Bn) _n(nt+1)(n+2) sao [Art. 70.] 73. To find the number of shot arranged in a complete pyramid the base of which is a rectangle. Let m and n be the number of shot in the long and short side respectively of the base. The top layer consists of a single row of m—(n—1), or m—n+1 shot; in the next layer the number is 2(m—n +2); in the next layer the number is 3 (m —n + 3) ; and so on; ~ in the lowest layer the number is n(m—n +m).PILES OF SHOT AND SHELLS. 55 o. S=(m—n+1)4+2(m—n + 2)4+3(m—n+ 3) +... +n(m—n+n) =(m—n)(1+24+34...4n)+ (19+ 294+ 3%+... 40%) _(m—n)n(n+1) n(m+1) (2n+1) QB = ae) {3 (m—n) + 2n4 1} _ 2 (n+1) (3m-n+1) =e 74. To find the number of shot arranged in an incomplete pyramid the base of which is a rectangle. Let a and 6 denote the number of shot in the two sides of the top layer, n the number of layers. In the top layer the number of shot is ab ; in the next layer the number is (a + 1) (6 + 1) ; in the next layer the number is (a + 2) (b + 2) ; and so on; in the lowest layer the number is (a +n — 1) (b+ -1) or ab + (a + b)(n-1) + (n—1)*. o. S= abn + (a+b) 3% (n—1)+ 3 (n—- 1)? = abn U(r), (mon@ noi s)) =F (Gab + 3 (a + 8) (n- 1) +(n—1) (2n— 1} 75. In numerical examples it is generally easier to use the following method. Ezample. Find the number of shot in an incomplete square pile of 16 courses, having 12 shot in each side of the top. If we place on the given pile a square pile having 11 shot in each side of the base, we obtain a complete square pile of 27 courses; . 27 x 28 x 55 _ “pe 11x12x23_ 3 = and number of shot in the complete pile= 6930; [Art. 71.] algo number of shot in the added pile= 506 ; +, number of shot in the incomplete pile = 6424.56 HIGHER ALGEBRA. EXAMPLES. VI. b. Find the number of shot in 1, A square pile, having 15 shot in each side of the base, 2, A triangular pile, having 18 shot in each side of the base. 3. A rectangular pile, the length and the breadth of the base con- taining 50 and 28 shot respectively. 4, An incomplete triangular pile, a side of the base having 25 shot, and a side of the top 14, 5. An incomplete square pile of 27 courses, having 40 shot in each side of the base. 6, The number of shot in a complete rectangular pile is 24395; if hah are 34 shot in the breadth of the base, how many are there in its ngth 7 7. The number of shot in the top layer of a square pile is 169, and in the lowest layer is 1089; how many shot does the pile contain ? 8. Find the number of shot in a complete rectangular pile of 15 courses, having 20 shot in the longer side of its base. 9. Find the number of shot in an incomplete rectangular pile, the number of shot in the sides of its upper course being 11 and 18, and the number in the shorter side of its lowest course being 30. 10, What is the number of shot required to complete a ea pile having 15 and 6 shot in the longer and shorter aide, respectively, of its upper course? 11, The number of shot in a trian pile is greater by 150 than half the number of shot in a square pile, the number of layers in each being the same; find the number of shot in the lowest layer of the tri- angular pile. 12, Find the number of shot in an incomplete square pile of 16 courses when the number of shot in the upper course is 1005 less than in the lowest course. 13. Shew that the number of shot in a square pile is one-fourth the number of shot in a triangular pile of double the number of courses. 14, If the number of shot in a triangular pile is to the number of shot in a square pile of double the number of courses as 13 to 175; find the number of shot in each pile, 15. The value of a triangular pile of 16 lb. shot is £51; if the value of iron be 10s. 6d. per cwt., find the number of shot in the lowest layer. 16. If from a complete square pile of courses a triangular pile of the same number of courses be formed ; shew that the remaining shot will be just sufficient to form another triangular pile, and find the _ number of shot in its side,CHAPTER VII. SCALES OF NOTATION. 76, The ordinary numbers with which we are acquainted in Arithmetic are expressed by means of multiples of powers of 10; for instance 25=2x10+5; 4705 =4 x 10°+7 x 10°+0x10+5. This method of representing numbers is called the common or ~ denary scale of notation, and ten is said to be the radix of the scale. The symbols employed in this system of notation are the nine digits and zero, In like manner any number other than ten may be taken as the radix of a scale of notation ; thus if 7 is the radix, a number expressed by 2453 represents 2x 7°+4x79+5x7+3; and in this scale no digit higher than 6 can occur. Again in a scale whose radix is denoted by r the above number 2453 stands for 2r° + 4r*+57r+3. More generally, if in the scale whose radix is r we denote the digits, beginning with that in the units’ place, by a,, a,, a,,...a,; then the number so formed will be represented by arta, +a, r+... +4artart+a,, where the coefficients a,, a,_,,...a, are integers, all less than 1, of which any one or more after the first may be zero. Hence in this scale the digits are + in number, their values ranging from 0 to r—1. 77, The names Binary, Ternary, Quaternary, Quinary, Senary, Septenary, Octenary, Nonary, Denary, Undenary, and Duodenary are used to denote the scales corresponding to the values two, three,...twelve of the radix,58 HIGHER ALGEBRA. In the undenary, duodenary,... scales we shall require symbols to represent the digits which are greater than nine, It is unusual to consider any scale higher than that with radix twelve; when necessary we shall employ the symbols ¢, e, 7’ as digits to denote ‘ten’, ‘eleven’ and ‘twelve’. It is especially worthy of notice that in every scale 10 is the symbol not for ‘ten’, but for the radix itself. 78. The ordinary operations of Arithmetic may be performed in any scale; but, bearing in mind that the successive powers of the radix are no longer powers of ten, in determining the carryiny Jgures we must not divide by ten, but by the radix of the scale in question. Example 1. In the scale of eight subtract 371532 from 530225, and multiply the difference by 27. 530225 186473 871532 27 136473 1226235 7 276166 4200115 Explanation, After the first figure of the subtraction, since we cannot take 3 from 2 we add8; thus we have to take 3 from ten, which leaves 7; then 6 from ten, which leaves 4; then 2 from eight which leaves 6; and so on. Again, in multiplying by 7, we have 3 x T=twenty one=2 x 8+5; we therefore put down 5 and carry 2. Next 7x7+2=fifty one=6 x 8+3; put down 3 and carry 6; and so on, until the multiplication is completed. In the addition, 84+6=nine=1x8+1; we therefore put down 1 and carry 1. Similarly 24+6+41=nine=1x8+1; and 6+1+1=eight=1x8+0; and so on. Example 2. Divide 15et20 by 9 in the scale of twelve. 9)15¢t20 1¢e96...6. Explanation. Since 15=1x T+5=seventeen=1x9+8, we put down 1 and carry 8. Also 8 x T +e=one hundred and seven=e x 9+8; ve therefore put down e and carry 8; and 80 on.SCALES OF NOTATION. 59 Ezample 8. Find the square root of 442641 in the scale of seven. 449641 (546 34 1384/1026 602 1416/12441 12441 EXAMPLES. VIL a. Add together 23241, 4032, 300421 in the scale of five. Find the sum of the nonary numbers 303478, 150732, 264305. Subtract 1732765 from 3673124 in the scale of eight. From 34756 take 2e46¢2 in the duodenary scale. . Divide the difference between 1131315 and 235143 by 4 in the scale of six. 6, Multiply 6431 by 35 in the scale of seven. 7. Find the product of the nonary numbers 4685, 3483. 8. Divide 102432 by 36 in the scale of seven. 9. In the ternary scale subtract 121012 from 11022201, and divide the result by 1201. 10. Find the square root of 300114 in the quinary scale. 11. Find the square of tttt in the scale of eleven. 12. Find the G. C. M. of 2541 and 3102 in the scale of seven. 13. Divide 14332216 by 6541 in the septenary scale. 14. Subtract 20404020 from 103050301 and find the square root of the result in the octenary scale. 15. Find the square root of ee¢001 in the scale of twelve. 16, The following numbers are in the scale of six, find by the ordi- nary rules, without transforming to the denary scale: (1) the G. C, M. of 31141 and 3102; (2) tho L. C. M. of 23, 24, 30, 32, 40, 41, 43, 50. oP epr 79. To express a given integral number in any proposed scale. Let W be the given number, and r the radix of the proposed scale. Let a,, a, @,,...a, be the required digits by which WV is to be expressed, beginning with that in the units’ place; then Naas" +a," "'+... +a +ar+a, We have now to find the values of a,, @,, dg)...60 , HIGHER ALGEBRA, Divide NW by r, then the remainder is a,, and the quotient is a +O, + FO + Oy. If this quotient is divided by r, the remainder is a, ; if the next quotient ...... eeeeerenl te seteeesssceessscenees Ogi and so on, until there is no further quotient. Thus all the required digits a,, a,, a,,...a, are determined by successive divisions by the radix of the proposed scale. Example 1, Express the denary number 5218 in the scale of seven. 7)6218 1)744......5 7)106......2 Die 1 Bl Thus ° 6218 =2 x 44+1x P41 x P4+2x7+5; and the number required is 21125. Example 2. Transform 21125 from scale seven to scale eleven. attas «*. the required number is 820¢. Explanation. In the first line of work 21=2x7+1 = fifteen=1xe+4; therefore on dividing by e we put down 1 and carry 4. Next 4x 7+1=twenty nine=2xe+7; therefore we put down 2 and carry 7; and so on. Example 3. Reduce 7215 from scale twelve to scale ten by working in scale ten, and verify the result by working in the scale twelve, 7215 12 86 In scale of ten 2. 1033 12 12401 Thus the result is 12401 in each case, Explanation. 7215 in scale twelve means 7 x 123+ 2x 1274+1x12+65 in scale ten. The calculation is most readily effected by writing this expression in the form [{(7x 12+ 2)}x12+1]x12+5; thus we multiply 7 by 12, and add 2 to the product; then we multiply 86 by 12 and add 1 to the product; -an 1083 by 12 and add 5 to the product.SCALES OF NOTATION. 61 80. Hitherto we have only discussed whole numbers; but fractions may also be expressed in any scale of notation ; thus ap 2,5. tae. 25 in scale ten denotes 75 + 53; +25 in scale six denotes + é 8 ‘25 in scale r denotes 2 ao ror Fractions thus expressed in a form analogous to that of ordinary decimal fractions are called radix-fractions, and the point is called the radix-point. The general type of such fractions in scale r is 4 8, 4 fry 7a 4 28 ; ett Bt ee A where 4,, },, b,,... are integers, all less than 7, of which any one or more may be zero. 81. To express a given radix fraction in any proposed scale. Let F be the given fraction, and r the radix of the proposed scale. Let 5,, 5,, 6,,... be the required digits beginning from the left ; then 6 4 ra OO ‘We have now to find the values of 6,, 3,, 3,,...... Multiply both sides of the equation by r; then a , 4, rF=b, ++ 4+ seeaee 3 Hence 5, is equal to the integral part of r¥'; and, if we denote the fractional part by F,, we have Multiply again by +; then, as before, 8, is the integral part of r¥; and similarly by successive multiplications by r, each of the digits may be found, and the fraction expressed in the pro- posed scale,62 HIGHER ALGEBRA. If in the successive multiplications by r any one of the products is an integer the process terminates at this stage, and the given fraction can be expressed by a finite number of digits. But if none of the products is an integer the process will never terminate, and in this case the digits recur, forming a radix- fraction analogous to a recurring decimal. Example 1, Express i as a radix fraction in scale six. . . 4.5,1,38 .. the required fraction=5 + gt gat gi = 14513, Ezample 2. Transform 16064-24 from scale eight to scale five, ‘We must treat the integral and the fractional parts separately, 5)16064 124 5)2044...0 5 54 TH 5)71... 5 5)18 er i 5 +e 5 024 After this the digits in the fractional part recur; hence the required number is 212340-i240, 82. In any scale of notation of which the radix: is r, the sum of the digits of any whole number divided by r—1 will leave the same remainder as the whole number divided by r — 1. Let WV denote the number, a,, a,, @,,......a, the digits begin- ning with that in the units’ place, and S the sum of the digits; then Naa,tartagr +... a_ tag"; Sasa, +0, +, 4 04, ni + @, : W-S=a, (r-1) +a, (r*—1)+ sta,_, ("= 1) +a, (r"-1).SCALES OF NOTATION. 63 Now every term on the right hand side is divisible by r— 1; V-S . “Sry =an integer ; f NV Ss that is, mat , where J is some integer ; which proves the proposition. Hence & number in scale r will be divisible by r— 1 when the sum of its digits is divisible by r—1. 83. By taking +=10 we learn from the above proposition that a number divided by 9 will leave the same remainder as the sum of its digits divided by 9. The rule known as “ casting out the nines” for testing the accuracy of multiplication is founded on this property. _ The rule may be thus explained : Let two numbers be represented by 94+6 and 9c+d, and their product by P; then . P= Blac + 9be + 9ad + bd. P a bd Hence , has the same remainder as > and therefore the sum of the digits of P, when divided by 9, gives the same remainder as the sum of the digits of bd, when divided by 9. If on trial this should not be the case, the multiplication must have been incorrectly performed. In practice 5 and d are readily found from the sums of the digits of the two numbers to be multiplied together. Ezample. Can the product of 831256 and 8427 be 263395312? The sums of the digits of the multiplicand, multiplier, and product are 17, 21, and 84 respectively; again, the sums of the digits of these three numbers are 8, 38, and 7, whence bd=8x8=24, which has 6 for the sum of the digits; thus we have two different remainders, 6 and 7, and the multiplication is incorrect. 84. If N denote any number in the scale of r, and D denote the difference, supposed positive, between the sums of the digits in the odd and the even places ; then N-—D or N+D is a multiple of r+l.64 HIGHER ALGEBRA, Let a,, a,,@,,......@, denote the digits beginning with that in the units’ place; then Nea, tartar tar t+... tar +ar. o. N-a,+4,-a,+4,-... =a, (r +1) +a, (r*-1) +4, (° +1)+...; and the last term on the right will be a, (r"+1) or a, (r*— 1) according as ” is odd or even. ‘Thus every term on the right is divisible by r+ 1; hence N-(a,-4,+4,-4,+......) 7 oT an integer. Now a,—G@,+4,-@,+......=%D; NwxD. ‘ are rH is an integer ; which proves the proposition. Cor. If the sum of the digits in the even places is equal to the sum of the digits in the odd places, D = 0, and W is divisible by r+1. Ezample 1. Prove that 4°41 is 9 square number in any scale of notation whose radix is greater than 4, Let r be the radix; then 4.1 1\° aatadyS 45 = (2+) : thus the given number is the square of 21. Ezample 2, In what scale is the denary number 2°4375 represented by 2:13? Let r be the scale; then 1 3 _ous75=0e, 245+ {= 24975= 875; whence Tr* -16r — 48=0; that is, (Tr +12) (r- 4) =0. Hence the radix is 4. Sometimes it is best to use the following method. Example 3. In what scale will the nonary number 25607 be expreased by 101215? etl The required scale must be less than 9, since the new number appears the greater; also it must be greater than 5; therefore the required scale must be 6, 7, or 8; and by trial we find that it is 7,SCALES OF NOTATION. . 65 Ezample 4. By working in the duodenary scale, find the height of a whose volume is 864 cub. ft. 1048 cub, in., and the area of whose base is 46 sq. ft. 8 sq. in. The volume is 364} $4 cub. ft., which expressed in the scale of twelve is 264-734 cub. ft. The area is 46,%, sq. ft., which expressed in the scale of twelve is 8t-08, ‘We have therefore to divide 264-734 by 3-08 in the scale of twelve. 3t08)26473-4(7-e 22248 36274 36274 Thus the height is 7ft. 1lin. SE ANMSAPepy 10. 11. 12. 13. 14. 15. 16. 17. 18, EXAMPLES. VII. b. Express 4954 in the scale of seven. Express 624 in the scale of five. _ Express 206 in the binary scale. Express 1458 in the scale of three. Express 5381 in powers of nine. Transform 212231 from scale four to scale five. Express the duodenary number 398¢ in powers of 10. Transform 6¢12 from scale twelve to scale eleven. Transform 213014 from the senary to the nonary scale. Transform 23861 from scale nine to scale eight. Transform 400803 from the nonary to the quinary scale. Express the septenary number 20665152 in powers of 12. Transform ttteee from scale twelve to the common scale. Express 2 as a radix fraction in the septenary scale. 10 Transform 17:15625 from scale ten to scale twelve. Transform 200211 from the ternary to the nonary scale. Transform 71-03 from the et to the octenary scale. Express the septenary fraction 1552 eis adenary vulgar fraction 4 in its lowest terms. 19, 20. 21, Find the value of -4 and of -42 in the scale of seven. In what scale is the denary number 8 denoted by 2227 In what scale is the denary fraction 2°. [58 5 denoted by (03023 HH A, . s66 HIGHER ALGEBRA. 92. Find the radix of the scale in which 554 represents the square of 24, 23, In what scale is 511197 denoted by 1746335? 24, Find the radix of the scale in which the numbers denoted by 479, 698, 907 are in arithmetical progression. 25. In what scale are the radix-fractions ‘16, ‘20, ‘28 in geometric progression? 26, The number 212542 is in the scale of six; in what scale will it be denoted by 17486? 27, Shew that 148-84 is a perfect square in every scale in which the radix is greater than eight. 28. Shew that 1234321 is a perfect square in any scale whose radix is greater than 4; and that the square root is always expressed by the same four digits. 29, Prove that 1:331 is a perfect cube in any scale whose radix is greater than three. 80, Find which of the weights 1, 2, 4, 8, 16,... lbs, must be used to weigh one ton. 81. Find which of the weights 1, 3, 9, 27, 81,... lbs. must be used to weigh ten thousand lbs., not more than one of each kind being used but in either scale that is necessary. 32, Shew that 1367631 is a perfect cube in every scale in which the radix is greater than seven. 33, Prove that in the ordinary scale a number will be divisible by 8 if the number formed by its last three digits is divisible by eight. 34. Prove that the square of rrzr in the scale of s is rxrg0001, where 9,7, 8 are any three consecutive integers, 35, If any number J be taken in the scale r, and a new number WV’ be formed by altering the order of its digits in any way, shew that the difference between WV and WV’ is divisible by r—1. 36, If a number has an even number of digits, shew that it is divisible by r+1 if the digits equidistant from each end are the same. 87. If in the ordinary scale S, be the sum of the digits of a number WN, and 3S, be the sum of the digits of the number 3, prove that the difference between S, and S, is a multiple of 3, 38, Shew that in the ordinary scale any number formed by writing down three digits and then repeating them in the same order is a multiple of 7, 11, and 13, _ 39, In a scale whose radix is odd, shew that the sum of the digits of any number will be odd if the number be odd, and even if the number be even. 40, If n be odd, and a number in the denary scale be formed by writing down n digits and then repeating them in the same order, shew that it will be divisible by the number formed by the n digits, and also by 9090...9091 containing n- 1 digits, .CHAPTER VIII. SURDS AND IMAGINARY QUANTITIES. 85. Inthe Elementary Algebra, Art. 272, it is proved that ‘the denominator of any expression of the form Jord can be rationalised by multiplying the numerator and the denominator by ,/b— ,/e, the surd conjugate to the denominator. Similarly, in the case of a fraction of the form orgerga , where the denominator involves three quadratic surds, we may by two operations render that denominator rational. For, first multiply. both numerator and denominator by b+ J/e—J/d; the denominator becomes (/b+,/c)*—(,/d)* or b+c-—d+2,/bc. Then multiply both numerator and denominator by (6+¢-d)—2 Jbc; the denominator becomes (5 + ¢ - d)*— 4be, which is a rational quantity. Example. Simplify EE 5 . 12 (84/5 -+2n/2) The expression =G+ are ae 12 (844/58 + 2n/2) - 6+6/5 2 B+n/5+ 2n/2) (5-1) (V5 +1) (5-1) _242n/5-+2/10~ 2/2 ater Ne =14/5-+/10- 4/2. 5—268 HIGHER ALGEBRA, 86. To find the factor which will rationalise any given bino- mial surd. Case I. Suppose the given surd is 7/a—,/b. Let Y/a=a, Yb=y, and let n be the t.c.m, of p and g; then a" and y are both rational. Now 2*—y" is divisible by 2—y for all values of n, and a my" = (e—y) (a +a" yrey+ een +y""). Thus the rationalising factor is a T* + of 8y + oP 8g? eo... +95 and the rational product is a" — y". Case II. Suppose the given surd is 2/a + 2/b. Let a, y, n have the same meanings as before; then (1) If n is even, x" — y" is divisible "ve a+y, and ay" = (at y) (a — a ty to. ay"*—y""). . Thus the ae factor is ay te tayo; and the rational roel is 2" — ‘¥. (2) If m is odd, x" + y* is divisible by x+y, and at y= (ety) (ee a fy to, — ay" +"), Thus the rationalising factor is and the rational product is a" + y*. Example 1. Find the factor which will rationalise ,/3 + 2/5. 2 1 Let <=3?, y=5%; then 2‘ and y® are both rational, and a8 y= (wy) (28 — aly tay? — ay? + nyt 8); thus, substituting for z and y, the required factor is B42 8 2 3 3 1 3? — 82, 534 3? . 58 38, 584.87, 53 — 58, 5 2 3 3 laos or. 82-9. 53437, 58-15 438, 535i 6 6 4nd the rational product ia 3 - 53 = 3? - 5°=2,SURDS AND IMAGINARY QUANTITIES, 69 14 as Example 2, Express (83498) = (5i_98) as an equivalent fraction with a rational denominator. 1 Sty; then since 2!—yt=(—y) (2+ fy +2y74+y3) 1 1 1 To rationalise the denominator, which is equal to 5?-84, put 5?=2, 2, 212 2 3 8 the required factor is 53+ 53. 34+ 53. 844. 34; 4 4 and the rational denominator is 5?— 84=5?- 8=22. (Bro) a8 shad. del . the expression = 5+ — +57, 3443 4 see WH 203 3 _ 57+2. 57, 8442. 57. 3442.57. 84434 ~ 22 3 1 2 1 3 _ 14459, 8445 . 874 57. 34 - ll 87. We have shewn in the Elementary Algebra, Art. 277, how to find the square root of a binomial quadratic surd. We may sometimes extract the square root of an expression contain- ing more than two quadratic surds, such as a+ /b+ /ce+ /d. Assume Ja+ Jb+Je+ Jd=Ju+Jyt J; atlb+ for fdaatryte+2,/ay +2 Jaat2 Jyz. Ifthen 2 ,Jzy=Jb, 2fzz= Jo, 2./yz= Jd, and if, at the same time, the values of a, y, z thus found satisfy a +y+z=a, we shall have obtained the required root. Example. Find the square root of 21 — 4,/5 + 8,/3 - 4,/15. Assume = 21 - 4.5 + 88-415 =z + /y-W/23 we 21-4) /5 +8)/8 - 4/1 =a ty +2 +2n/ zy — Qn) xe — 2n/ yz. Put 2n/ zy =8)/3, Infzz =4/15, 2n/yz= 4/5; by multiplication, zyz=240; that is /zyz=4,/15; whence it follows that ,/z=2,/3, /y=2, /#=,/5. And since these values satisfy the equation x+y +2=21, the required root is 3/3+3-/5. 2 ,70 HIGHER ALGEBRA. 88. If Jat /b=x+ Jy, then will t/a Jb=x- Jy. For, by cubing, we obtain a+ Jb =a? + 3a? Jy + 3ayty Jy. Equating rational and irrational parts, we have aaa + 3xy, /b=32" Jy+y Jy; oe @— b= 2? — 32? /y + 3xy-y Jy5 that is, Ja- Jo=a- Jy. Similarly, by the help of the Binomial Theorem, Chap. XIII., it may be proved that if Ya+ Jb=a+,/y, then Ja—Jb=x-,/y, where is any positive integer. 89. By the following method the cube root of an expression of the form a+ ,/b may sometimes be found. Suppose Jax fo=a+ Jy; then Va - Jb=a- Jy. LP RB a af HY eee eeecccrteteeteeeees (1). Again, as in the last article, GHP BEY. eect erect steteeseees (2). The values of « and y have to be determined from (1) and (2). In (1) suppose that {/a*—b =c; then by substituting for y in (2) we obtain a =a + 3a: (a* —c) ; that is, 42° — 3cex = a. If from this equation the value of x can be determined by trial, the value of y is obtained from y=" — c. Nore. We do not here assume ,/r+,/y for the cube root, as in the extraction of the square root; for with this assumption, on cubing we should have a+ fb=a/z+Ba/y + 8yJ/zt+ yy, and since every term on the right hand side is irrational we cannot equate . 4ational and irrational parts.SURDS AND IMAGINARY QUANTITIES. 71 Ezample. Find the cube root of 72 - 82/5. Assume 872-325 =2-Jy; then YiB+ 826 = 2+ Jy. By multiplication, 3/5184—1024x5=a?-y; that is, Again 72 — 32,/5 = 23 — 32°, /y + Bay —y/y whence 72=23 + Bay (2). From (1) and (2), 72 = 25 + 8x (x?- 4); that is, a3 —8e=18. By trial, we find that s=3; hence y=5, and the cube root is 3 - ,/5. 90. When the binomial whose cube root we are seeking consists of two quadratic surds, we proceed as follows. Example. Find the cube root of 9,/3 + 11,/2. JaJrT= x/ ay (8+ 54/3) =v3a/ 8+ ie we By proceeding as in the last article, we find that 57, tl 2 2 84 3/5} + 9/33 .*. the required cube root =,/8 (a + J 3) =N8 +2 91. We add a few harder examples in surds. Example 1. Express with rational denominator woeri 5 The expression = —— 33-3341 4 ( 3 + 1) (3841) (3841) _ (341) ; F | =33 B71 =85 41,72 HIGHER ALGEBRA, Ezample 2. Find the square root of Rle-1) 4/9 Ted, The expression = 4 {80-842 /@e+1)(e-4)} = H(2z4+1)+(e-4)+2,/Gz Fi) eH}; hence, by inspection, the square root is a (Jaz41+ J/2-4). Example 3. Given ,/5=2-23607, find the value of v3= Jo __ N24+/7-3)5° Multiplying numerator and denominator by ,/2, the expression EXAMPLES. VIIL a. Express as equivalent fractions with rational denominator : 1) 2 — v2 * THJ2—,/3 ° " FO4+J/3-5° 1 4 aVatl Jat b+Va+b Va-1-V2a+Va+1 | J104+./5-/3 g, W8+v5) (J5-+4/2) N3+,/10-J5° ° J24+/84+)5 * Find a factor which will rationalise: 1 7. §3-/2. 8 54-32. 9. ahh IQ f3-1. 11, 24A/7. WwW Yo- 43.SURDS AND IMAGINARY QUANTITIES. 73 Express with rational denominator : ¥3-1 99-98 J2.2/3 13. W341" gorge 1B. F342" w3 8+ 9/4 27 1 BSI: ms: 1 5: Find the square root of 19, 16-2,/20-2,/284+2,/35. 20, 2444,/15-4,/21-2,/35, 21, 6+,/12—./24-/8. 22, 5-./10-J15+/6. 23, a4+3b+44+4J/a—4/3b-24/3ab. A, 214+3./8-6/3—-6,/7— 2/24 - /564+2,/21. Find the cube root of 25. 10+6,/3. 2%. 384+17/5. 27. 99-70 /2. 2B, 38J14-100/2 29, 54/3+4175. 30. 135,/3-87,/6. Find the square root of B81. ate+V/2art+2. 32. 2a—»/3a?— 2Qub — b. 2 ay 33, l+a%+(1+a%+at)?, 34 14+(1-a*) % 35.. Ifa= find the value of 7a? + 11ab — 78°. 1 4.1 a—J3? 924,73? 36, If one, 9433 , find the value of 32°- 5ay + 3y*. Find the value of 26-15 ./3 6+2V3 a 5 J/2—-V38+5/3 33-19 /3 2 2 39, (28- 10/3)! (744 3) 40, (26415 J/3)3— (26 + 15/3) 3 41, Given ,/5=2-23607, find the value of 10/2/10 +18 V18-V84+N5 J84+V3— 5 Divide 2341430 92 by e—1+2/2. Find the cube root of 9ab? + (8? + 24a”) »/6?— 3a%. BB ati _ 1 44, Bvaluste Ra when Be=Jat7, ’74 HIGHER ALGEBRA. IMAGINARY QUANTITIES. 92. Although from the rule of signs it is evident that a negative quantity cannot have a real square root, yet imaginary quantities represented by symbols of the form ,/—a, ./—1 are of frequent occurrence in mathematical investigations, and their use leads to valuable results. We therefore proceed to explain in what sense such roots are to be regarded. When the quantity under the radical sign is negative, we can no longer consider the symbol ,/ as indicating a possible arithmetical operation ; but just as ,/a may be defined as a symbol which obeys the relation ,/a x ,/a =a, 80 we shall define ,/—a to be such that J=ax./-a=-a, and we shall accept the meaning to which this assumption leads us. It will be found that this definition will enable us to bring imaginary quantities under the dominion of ordinary algebraical rules, and that through their use results may be obtained which can be relied on with as much certainty as others which depend solely on the use of real quantities. 93. By definition, J—-1x/-1=-1. we fa J=1x Ja. J-1=a(-1); that is, (Ja..J/=1)*=-a. Thus the product ,/a../—I may be regarded as equivalent to the imaginary quantity /—a. 94. It will generally be found convenient to indicate the imaginary character of an expression by the presence of the symbol ,/—1; thus Jha fT) 2 FT. J=Ta! = J 7a? x (=) =a /7 HT. 95. We shall always consider that, in the absence of any statement to the contrary, of the signs which may be prefixed before a radical the positive sign is to be taken. But in the use of imaginary quantities there is one point of importance which ‘serves notice,SURDS AND IMAGINARY QUANTITIES. 75 Since (- a) x (—b) =a, by taking the square root, we have [ax f=ban fab Thus in forming the product of ,/— a and ,/—6 it would appear that either of the signs + or — might be placed before ,/ab. This is not the case, for JaaxJ=b=Ja. J=1x¥b.J/=1 = fa (=I) =~ fab. 96. It is usual to apply the term ‘imaginary’ to all expres- sions which are not wholly real. Thus a+6,/—1 may be taken as the general type of all imaginary expressions. Here a and b are real quantities, but not necessarily rational. 97. In dealing with imaginary quantities we apply the laws of combination which have been proved in the case of other surd quantities. Ezamplel. a+b)/-14(c+d./-1)=axc+(bad)/—1. Example 2. The product of a+ b./—i and c+d./-1 =(a+b/=1) (e+4,/=7) =ac—bd + (be-+ad) J =1. 98. Ifa+bJ-1=0, thena=0, and b=0. For, if a+bJ/-1=0, then b/-i=-a; o -Baa’; - &+8=0. Now a? and 6° are both positive, therefore their sum cannot be zero unless each of them is separately zero; that is, a=0, and b=0. 99. Ifa+b/—-I=c+d/-l, thena=c, andb=d. For, by transposition, a —¢ +(b—d),/-1=0; therefore, by the last article, a—c=0, and b-d=0,; # that is a=c, and 6=d,76 HIGHER ALGEBRA. Thus in order that two imaginary expressions may be equal tt is necessary and sufficient that the real parts should be equal, and the imaginary parts should be equal. 100. Derinirion. When two imaginary expressions differ only in the sign of the imaginary part they are said to be conjugate. Thus a—b,/—1 is conjugate toa+6,/—1. Similarly /2+3,/—1 is conjugate to /2—-3,/—1. 101. The sum and the product of two conjugate imaginary expressions are both real. For a+b J-T+a-b,J=1 = 2a, Again (a+6,J=1) (a-6,/=1) = a? (-8°) =a +6% 102. Derinition. The positive value of the square root of a’ +0 is called the modulus of each of the conjugate expressions a+b J=T and a-bJ—1. 103. The modulus of the product of two imaginary expres- sions is equal to the product of their moduli. Let the two expressions be denoted by a+b/=1 and e+d,/—1. Then their product =ac—bd+(ad+be),/—1, which is an imaginary expression whose modulus = Mer 8) (+ a) =JaaB x (Ord; which proves the proposition. 104, If the denominator of a fraction is of the form a+6 1, it may be rationalised-by multiplying the numerator and the denominator by the conjugate expression a—b ai.SURDS AND IMAGINARY QUANTITIES. 77 For instance e+dJ/=1 _(¢+dJ=1)(a-b,/-1) a+bf-1 (a+b,J/-1)(a-b/=1) _ac+bd+ (ad ~ be), J=1 ~ at +b ac+bd ad-—be ;,—~ abt ate Thus by reference to Art. 97, we see that the sum, difference, product, and quotient of two imaginary expressions is in each case an imaginary expression of the same form. 105. To find the square’root of a +b,/—1. Assume Ja+bV¥—l=x+yV-l, where x and y are real quantities. By squaring, a+b —laa'—y'+2ay,/-1; therefore, by equating real and imaginary parts, woyaa.... 2ay =b o (+ yf = (ey) + Ray) =a' +b; oo tees [F4B., From (1) and (3), we obtain Jetta J@+b—a a= 3 y . ona eee, you {feel Thus the required root is obtained. Since x and y are real quantities, 27+ y? is positive, and therefore in (3) the positive sign must be prefixed before the quantity ,/a?+ b%, Also from (2) we see that the product zy must have the same sign 88 b4 hence x and y must have like signs if b is positive, and unlike wigna it b ia negative.78 HIGHER ALGEBRA. Example 1, Find the square root of — 7-24 ,/=1. Assume J-1-%4)-l=ety /71; then -7-24,/—1=29-y2422y,/=1; tay? =-7, and ary = — 24, oe (att y= (a? - 78+ (2ay)? =49 +576" =625; - weet +yt= O56. From (1) and (2), 2?=9 and y2=16; 2a £3, y= 4, Since the product zy is negative, we must take a=3, y=-4; orz=-3, y=4. Thus the roots are 8-4,/—Land -34+4,/-1; that is, J =7-24,f—1= #(3-4,/ 71). ” Example 2. To find the value of 4/— 64a‘, STU [tet JA = /2 J 2J=1. It remains to find the value of /+,/—1. Assume At JnisetyJ/ Tis then tf lato? 4 2ay J=T; 22 y?=0 and Q2y=1; 1 1 1 1 whence . tay Y= TR oo ae aaa /al o Va )=T=4 ltd). Similarly ota a 3a) and finally nf ~ Bata 2.20 (14/=1).SURDS AND IMAGINARY QUANTITIES. 79 106. The symbol ,/— 1 is often represented by the letter 4; but until the student has had a little practice in the use of imaginary quantities he will find it easier to retain the symbol ,/—I. It is useful to notice the successive powers of ,/— 1 ori; thus W/=Ty=/-1, isd; (J=1)*=-1, * WJ-Te=-J=1, #=-85 (J=i)'=1, a1; and since each power is obtained by multiplying the one before it by ./=1, or i, we see that the results must now recur. 107. We shall now investigate the properties of certain imagi- nary quantities which are of very frequent occurrence. Suppose a= 2/1; then 2*=1, or 2 -1=0; that is, (@—1) (a? +%+4+1)=0. * either w—-1=0, ora*+a+1=0; whence a=1, or ga ats=§, It may be shewn by actual involution that each of these values when cubed is equal to unity, Thus unity has three cube roots, pris =8 t1e=3, ? Oma ’ 2 > two of which are imaginary expressions. Let us denote these by a and 8 ; then since they are the roots of the equation w'+a+1=0, their product is equal to unity ; that is, aB=1; @B=a'; that is, B=a', since a’ =1. Similarly we may shew that a =p. 108. Since each of the imaginary roots is the square of the other, it is usual to denote the three cube roots of unity by 1, ©, o.80 HIGHER ALGEBRA. Also o satisfies the equation a*+a2+1=0; l+a+a’=0; that is, the sum of the three cube roots of unity is zero. Again, ow = =1; therefore (1) the product of the two imaginary roots is unity ; (2) every integral power of w* is unity. 109. It is useful to notice that the successive positive integral powers of w are 1, , and w*; for, if n be a multiple of 3, it must be of the form 3m; and w*=w™= 1. If n be not a multiple of 3, it must be of the form 3m +1 or 3m + 2. If n=3m+1, ow =o" a0", If n= 3m+2, oe = ot = 110. We now see that every quantity has three cube roots, two of which are imaginary. For the cube roots of a® are those of a*x 1, and therefore are a, aw, aw*, Similarly the cube roots of 9 are 2/9, w ¥/9, w* 2/9, where 2/9 is the cube root found by the ordinary arithmetical rule. In future, unless otherwise stated, the symbol Ya will always be taken to denote the arithmetical cube root of a. =I) — Eaample 1. Reduce @+3/=1) to the form 4+ BZ. 24+J-1 The expression 479412, /-T 2+4,/-1 _(-5+12,/=1) (2-,/=1) @+ /-1)(2-/-1) _ -104+12+29,/=1 441 2,29 =gtgvii which is of the required form. Example 2, Resolve x*+-y8 into three factors of the first degree. eo) B+y= (ety) (o-2y+y) vB +ys=(e+y) (z+ wy) (z+ wy); for w+o'= -1, and w=1,SURDS AND IMAGINARY QUANTITIES, 81 Example 3. Shew that (a+ «b+ we) (a+ wd + we) =a? +b? +409— be -ca—ab. In the product of a+wb+uw%e and a+w%d+we, the coefficients of b? and c? are w*, or 1; the coefficient of be =u8+ot=u?+0=-1; the coefficients of ca and ab=w?+w=-1; os (4+ 0b + we) (a+ wb + we) =a? + L240? - be - ca — ad. Example 4, Shew that (L+o- 0%) ~ (1-04 ut)'=0. Bince 1+w+w*=0, we have (1+ 0) - (1-0 + 0) = (— 20%) — (- Qu)? = - 88 + But =-8+8 =0. EXAMPLES. VIII. b. r Multiply 2./—34+3/—2 by 4./—3-5/—3. Multiply 3 —7-5/—2 by 3\/—7+5—2. Multiply eV=1 46-V-1 by eV=1—e-N=1, Multiply x—2+¥=3 by #-tov=8 9 po 3 . Express with rational denominator: 5. 1 6 BV —-242f-5 * 3-7-2" * g7-2-27-5° 7 3427-1 3-2-1 a eteval_a-aWV=1 * 9-57=1 245/-1- a-aV¥-1 atay-1 9, (EVERY (w-W=I! yg (atV=T—(@-V=TP teal eb ad (at TP (@= 1h 11, Find the value of (—4/—1)'"*3, when 7 is a positive integer. 12, Find the square of /9 +40 /—1+/9—40N 1. BEA 682 HIGHER ALGEBRA. Find the square root of . 13 -54127-1, 1 -11-60V7-1. 15, -474+87-3. 16. -8/-1. 17. a§-142a/71. 18 4ab—2(a?- b*) /-1. Express in the form 4+7B 3450 J3-iJ2 1+é 19, 223i" 20, 2/312" 21, Tor (1+7)? (a+? (a-ib? 2 S-i B. G-ih ~ at If 1, w, w? are the thrce cube roots of unity, prove %, (1+0%)t=o. 25. (l-@+e?) (l+o—o%)=4, 26. (1-a) (1-") (1-04) (1-05) =9. 27. (2+50+202)' = (2+ 2w +50?) = 729. 28, (1-w +o) (1—w? +0) (1—o!+0.).., to 2n factors = 2%, 29, Prove that B+ P+ 3 —Bny2=(e+y+z) (etyo + 20%) («1 +yo?+z20). 30. If a=atb, y=aot+be’, z=aw*+ bo, shew that (Ql) ayz=ai +B (2) #+y?+2=6ab. (3) #8 +y3+23=3 (a3 +09), Sl. If axteyt+be=X, cxtbyt+az=Y, brt+ayta=Z, shew that (a? +0? +o —be—ca—ab) (x+y? +2?—ys—20—ay) =X24 24 77- YZ-XZ-XY.CHAPTER IX. THE THEORY OF QUADRATIC EQUATIONS. 111. Arter suitable reduction every quadratic equation may be written in the form Gat? +b CHO oe eecce cee e seen (1), and the solution of the equation is ~, rere tpg —b=/b*- 4uc “= =e Pte peeees eee seeeeesseaese (2). ‘We shall now prove some important propositions connected with the roots and coefficients of all equations of which (1) is the type. 112. A quadratic equation cannot have more than two roots. For, if possible, let the equation ax*+ba+c=0 have threo different roots a, B, y. Then since each of these values must satisfy the equation, we have a + bat CHO eceeccctteteeeeeeeeeee af*+bB+c=0. ay'+by+e=0.. From (1) and (2), by subtraction, a (a? — B*) + b(a—f)=0; divide out by a— 8 which, by hypothesis, is not zero; then a(a+f)+b=0. Similarly from (2) and (3) a(B+ y)+b=0; .. by subtraction a(a-y)=0; which is impossible, since, by hypothesis, a is not zero, and o is not equal to y. Hence there cannot be three different roots. 6—284 HIGHER ALGEBRA. 113. In Art. 111 let the two roots in (2) be denoted by a and B, 80 that : _-b+,/6*—4ac pare veda, — 2a , 2a , then we have the following results : (1)_ If 8*-4ac (the quantity under the radical) is positive, a and £ are real and unequal. (2) Jf b*-4ac is zero, a and B are real and equal, each reducing in this case to — 5a" (3) 1£b*—4ac is negative, a and f are imaginary and unequal. (4) If b"'—4ac is a perfect square, a and f are rational and unequal, By applying these tests the nature of the roots of any quadratic may be determined without solving the equation. Example 1. Shew that the equation 2z?-6x+7=0 cannot be satisfied by any real values of x. Here a=2, b= -6, c=7; so that UW - 4ac=(-6)?-4.2.7= -20. ‘Therefore the roots are imaginary. Example 2. If the equation 27+ 2 (k+2) 2+9k=0 has equal roots, find k. The condition for equal roots gives (k+2)9=9k, K-5k4+4=0, (k-4)(k-1)=0; . k=4, or 1, Example 3. Shew that the roots of the equation 23 —Qpz+p*—q?+2gr-r?=0 are rational. The roots will be rational provided (--2p)*~4(p*-q?42qr—r%) is a perfect square, But this expression reduces to 4 (q?-2qr+r%), or 4(q—r)%. Hence the roots are rational. 114. Since Soe ERT pa wai fae fn tee we have by addition a4 pares MBH dae — b~ f= dae 2a SBT Tg ert rereeereees (1);THE THEORY OF QUADRATIC EQUATIONS. 85 and by multiplication we have (<3 + J/6¥ = 4a0) (— b— ,/6* = 4ac) a8 = tai _ (-2)*- (0 - 4ac) 4a’ 4ac_¢ i i (2). By writing the equation in the form at+ a +n 0, a a these results may also be expressed as follows. In a quadratic equation where the coefficient of the first term is unity, (i) the sum of the roots is equal to the coefficient of a with its sign changed ; (ii) the product of the roots is equal to the third term. Nore. In any equation the term which does not contain the unknown quantity is frequently called the absolute term. A b ¢ 115. Since ~g70tB, and a7 the equation a*+ S at < =0 may be written (at B) FOB HD oo. ecreeeeeeeeee (1). Hence any quadratic may also be expressed in the form a’ — (sum of roots) a+ product of roots=0......... (2). Again, from (1) we have (@— a) (2 — B) =O ..eeeeeeeeeeeeeteeeeeee (3). We may now easily form an equation with given roots. Ezample 1, Form the equation whose roots are 8 and — 2. The equation is («- 8) (2+2)=0, or a-2-6=0. When the roots are irrational it is easier to use the following method,86 HIGHER ALGEBRA. Ezample 2, Form the equation whose roots are 2+,/3 and 2 -,/3, ‘We have sum of roots=4, product of roots=1; .*. the equation is a-424+1=0, by using formula (2) of the present article. 116. By a method analogous to that used in Example 1 of the last article we can form an equation with three or more given roots. Example 1, Form the equation whose roots are 2, - 8, and ee The required equation must be satisfied by each of the following sup- positions ; 2-2=0, 2+8=0, a-1=0; therefore the equation must be (22) (248) (2-3) =05 that is, (# - 2) (+8) (52-7) =0, or 5a3 — 227-372 4+42=0, Example 2. Form the equation whose roots are 0, +a, ; a The equation has to be satisfied by ¢ #20, t=a, 2=-4, 225 therefore it is #(e+a)(2~a) ( ~$)=0; that is, & (2? - a?) (ba -c)=0, or bat — ex3 - aba? + aex=0. 117. The results of Art. 114 are most important, and they are generally sufficient to solve problems connected with the roots of quadratics, In such questions the roots should never be considered singly, but use should be made of the relations ob- tained by writing down the sum of the roots, and their product, in terms of the coefficients of the equation. Example 1. If a and £ are the roots of «*-px+q=0, find the value of (1) a?+ 6%, (2) a3 + 6%. We have | at+p=p, ap=q. +. al+ p= (a+8)?- 208 =p*—2q.THE THEORY OF QUADRATIC EQUATIONS, 87 Again, aS + B*= (a+) (a*+ B* - a8) =p {(a +6)? — Bag} =p (p?-39). Ezample 2. If a, B are the roots of the equation lz?+mz+4+n=0, find the ane aB equation whose roots are 2 — We have sum of roots =* + BL ids , Boa ap product of roots =< .. by Art. 115 the required equation is 2 2 (3) z+1=0, ap or aBa? — (a? + 6°) 2+ aB=0. 2 As in the last example ayp=™ -. and ap=4- Aah n m?-2nl on .*. the equation is [e- et 7=% or nlz? — (m3 —2nl) x+nl=0. Example 8. When 2StbV =} )find the valuc of 223 4229-72472; and shew that it will be unaltered pbc belt NV! - 11 substituted for 2. Form the quadratic equation whose roots are Ee vai, the sum of the roots =8; the product of the roots = a, hence the equation is 2a? -62+17=0; .*. 22%-62+417 is a quadratic expression which vanishes for either of the values Now 2a342c%—Tx+72=2 (229-G2+17) +4 (22?-6x+4+17)+4 =2x0+4x044 =4; which is the numerical value of the expression in each of the supposed cose88 HIGHER ALGEBRA. 118. To find the condition that the roots of the equation ax’+bx+c=0 should be (1) equal in magnitude and opposite in sign, (2) reciprocals, . The roots will be equal in magnitude and opposite in sign if their sum is zero; hence the required condition is - e =0, or 6=0, a Again, the roots will be reciprocals when their product is unity ; hence we must have ¢ -=1, orec=a, a The first of these results is of frequent occurrence in Analyti- cal Geometry, and the second is a particular case of a more general condition applicable to equations of any degree. Example. Find the condition that the roots of az?+bz+c=0 may be (1) both positive, (2) opposite in sign, but the greater of them negative. b e ‘We have a+p=-c, op=s. . Qa the roots are both positive, a8 is positive, and therefore ¢ anda have like signs. Also, since a+ is positive, e is negative; therefore b and a have unlike Hence the required condition is that the signs of a and ¢ should be like, and opposite to the sign of b. (2) If the roots are of opposite signs, af is negative, and therefore c and a have unlike signs, . Also since a+ has the sign of the greater root it is negative, and there- fore HB is positive; therefore 6 and a have like signs, Hence the Ra ae condition is that the signs of a and b should be like, and opposite to the sign of c. EXAMPLES, IX. a. Form the equations whose roots are none one p-q _ptg Loa om 3 pry?" p-4 4 7425. 5 +2,/3-5, 6. —p+2V2q.THE THEORY OF QUADRATIC EQUATIONS, 89 J -345%. 8. -a+tib, 9 +7(a-6). 2 1 a 2 10. -3, 5) 3 ll. 97% -G- 12, 2+,/3, 4. 13. Prove that the roots of the following equations are real : (1) 2#-2ar+a2—B?-=0, (2) (a—b+c)2+4(a-b)2+(a-b—c)=0, 14. If the equation 2*~ 15 -m(2¢-—8)=0 has equal roots, find the values of m, 15, For what values of m will the equation a8 — 2x (14+3m)+7 (3+2m)=0 have equal roots? - . 16. For what value of m will the equation w—br_m-1 az-c m+1 have roots equal in magnitude but opposite in sign? 17, Prove that the roots of the following equations are rational: (1) (at+e-6) x*+2ce+(b+c—a)=0, (2) abe®z* + 3a%or + beex — 6a? —ab+2b?=0. Tf a, B are the roots of the equation az?+bx+c=0, find the values of 1,1 a p\t ate 19. aff’ +a7pt. 20. a-a)* Find the value of 21, 28422-2422 when #=1+4 2%. 22, «8-322 -8x+15 when x=3+7. 18. 23, a3 ax?+2a%r + 403 when a=1-/=3. 24. If a and £ are the roots of 2°+px+q=0, form the equation whose roots are (a - 8)? and (a+8)*. 25, Prove that the roots of (7—a)(7—6)=/? are always real. 26. If .x,, x, are the roots of a2*+bx+e=0, find the value of (1) @my+8)-2+ (ary +b)-%, (2) (aay +8) ~9+ (aay +8)-*.90 HIGHER ALGEBRA. 27, Find the condition that one root of ac*+be+c=0 shall be n times the other. 28. Ifa, B are the roots of az*+br+c=0, form the equation whose roots are a?+ 8? and a~?+,-% 29, Form the equation whose roots are the squares of the sum and of the difference of the roots of QF+2 (m+n) r+ m= +n2=0, 30. Discuss the sigtis of the roots of the equation pe+gr+r=0. 119. The following example illustrates a useful application of the results proved in Art. 113. © Example. If 2 is a real quantity, prove that the expression ee can have all numerical values except such as lie between 2 and 6. Let the given expression be represented by y, so that 420-11 Q@-8 4 then multiplying up and transposing, we have 224 2x (1-y)+6y-11=0. This is a quadratic equation, and in order that 2 may have real values 4(1-y)-4(6y—11) must be positive; or dividing by 4 and simplifying, y? — 8y +12 must be positive; that is, (y—- 6) (y—-2) must be positive. Hence the factors of this product must be both positive, or both negative. In the former case y is greater than 6; in the latter y is less than 2, Therefore y cannot lie between 2 and 6, but may have any other value. In this example it will be noticed that the quadratic expression y’ — 8y + 12 is positive so long as y does not lie between the roots of the corresponding quadratic equation y* —8y + 12 =0. This is a particular case of the general proposition investigated in the next article. 120, For all real values of x the expression ax*+bx +c has the same sign as a, except when the roots of the equation ax*+bx +c =0 are real and unequal, and x has a value lying between them. Case I. Suppose that the roots of the equation ax’ +bz+e=0 are real ; denote them by a and f, and \et « be the greater,THE THEORY OF QUADRATIC EQUATIONS. 91 1 2 Then at+betena(ats ret 3) a a, =a {a?-(a+ B)x+aB} =a(2—a) (2). Now if is greater than a, the factors «—a, 2-8 are both positive ; and if « is less than £, the factors x —a, «— are both negative ; therefore in each case the expression (a — a) (a — f) is positive, and az*+bx+e¢ has the same sign asa, But if x hasa value lying between a and £, the expression (x — a) (x—-) is negative, and the sign of az* +b. +¢ is opposite to that of a. Case Il. Ifa and B are equal, then aa" + ba +c=a(x—a)*, and (#—a)* is positive for all real values of x; hence ax*+ba+e has the same sign as a. Casz III. Suppose that the equation az*+bx+c=0 has imaginary roots; then at ebevenale sine waf(e+ 5) + 5"}. 2a, 4a? But b°-4ac is negative since the roots are imaginary ; hence _B Bae is positive, and the expression 4a* b\% 4ac— 0? (+3) + mene 2a, 4a? is positive for all real values of 2; therefore ax*+bx+c has the same sign asa. This establishes the proposition. 121. From the preceding article it follows that the expression az" +bx+c will always have the same sign whatever real value x may have, provided that 6° — 4ac is negative or zero; and if this condition is satisfied the expression is positive or negative accord- ing as a is positive or negative. Conversely, in order that the expression ax*+ba+c¢ may be always positive, °—4ae must be negative or zero, and a must be positive ; and in order that ax’ + ba+c may be always negptivay 6° — dae must be negative or zero, and a must be negative, ~92 HIGHER ALGEBRA. Example, Find the limits between which a must lie in order that ax*~72+5 b29-Ta+0 may be capable of all values, « being any real quantity. ax? -T2+5 Pat Bi Tera then (a - 5y) 22-72 (1-y)+ (5-ay)=0. In order that the values of x found from this quadratic may be real, the expreasion 49 (1 - y)? - 4 (a — 5y) (5 - ay) must be positive, that is, (49 - 20a) y? +2 (2a*+1) y+ (49 — 20a) must be positive; hence (2a* + 1)?- (49 - 20a)* must be negative or zero, and 49 - 20a must be positive, Now (2a?+1)*— (49 — 20a)? is negative or zero, according as ; 2 (a?- 10a + 26) x 2 (a? + 10a — 24) is negative or zero; that is, according as 4 (a — 5)? (a +12) (a— 2) is negative or zero. This expression is negative as long as a lies between 2 and -—12, and for such values 49 — 20a is positive; the expression is zero when a=5, — 12, or 2, but 49-20a is negative when a=5. Hence the limiting values are 2 and - 12, and a may have any intermediate value. EXAMPLES. IX. b. 1, Determine the limits between which 2 must lie in order that the equation Qax (ax+nc)+(n?—-2)A=0 may have real roots. ed must lie between 1 and — es 2. If be real, prove that s—sey0 it: w@—e+1 e+e+1 3. Shew that lies between 3 and B for all real values of x. @434¢—71 aoe 7 om have no value between 4, If.xbe real, prove that 5 and 9. 5. Find the equation whose roots are —va__, JatVa—b 6. Ifa, A are roots of the equation 2? — pr+4¢=0, find the value of (1) o2(at8-1—A) +6" (arta), (2) (a-p)-*+(B-p)-*.THE THEORY OF QUADRATIC EQUATIONS, 93 7. Ifthe roots of lz*+nx+n=0 be in the ratio of p : g, prove that [Pp qg n Py ft *=0, St Jit Z (z+m)?—4mn 8. If « be real, the expression 2(e=n) admits of all values except such as lie between 27 and 2m. 9. If the reots of the equation a2?+2bz+c=0 be a and f, and those of the equation 47?+2Bxr+C=0 be a+8and 8+, prove that B-ac_ Bt-AC “a AF ener will be capable of all values when is real, provided that p has any value between 1 and 7. 10. Shew that the expression . @+2 . 11, ; Find the greatest value of i +3036 for real values of 2. 12, Shew that if x is real, the expression (a - be) (2 —b—o)~1 has no real values between 6 and ¢. 13, If the roots of az?4+2bc+c=0 be possible and different, then the roots of . (a+) (ax® + 2bz +c) =2 (ae — b*) (4241) will be impossible, and vice versd. (ax —b) (dee) 14. Shew that the expression (b= a) (en —d will be capable of all — a) (ex —d) values when w is real, if a?- 0? and & “a (ere the same sign. *122. We shall conclude this chapter with some miscellaneous theorems and examples. It will be convenient here to introduce a phraseology and notation which the student will frequently meet with in his mathematical reading. Derinition. Any expression which involves x, and whose value is dependent on that of a, is called a function of x. Functions of x are usually denoted by symbols of the form /(2), F(@), $(@). Thus the equation y=/ (x) may be considered as equivalent to a statement that any change made in the value of 2 will pro- duce a consequent change in y, and vice versd. The quantities x and y are called variables, and are further distinguished 2a the independent variable and the dependent variable.94 HIGHER ALGEBRA. An independent variable is a quantity which may have any value we choose to assign to it, and the corresponding dependent variable has its value determined as soon as the value of the inde- pendent variable is known. *123. An expression of the form Poe + pe + pa +... +p, ety, where n is a positive integer, and the coefficients p,, p,, py-..p, do not involve 2, is called a rational and integral algebraical function of x. In the present chapter we shall confine our attention to functions of this kind, *124. A function is said to be linear when it contains no higher power of the variable than the first ; thus ax + } is a linear function of # A function is said to be quadratic when it contains no higher power of the variable than the second ; thus ax*+ba+ce is a quadratic function of a Functions of the third, Jfourth,... degrees are those in which the highest power of the variable is respectively the third, fourth,.... Thus in the last article the expression is a function of x of the n degree. *125. The symbol f(z, y) is used to denote a function of two variables x and y; thus aa +by+c, and ax*+ bay +cy*’+da+ey+f are respectively linear and quadratic functions of 2, y. The equations f(x) =0, f(2, y) = 0 are said to be linear, quad- ratic,... according as the functions f(x), f(x, y) are linear, quad- ratic,.... *126. We have proved in Art. 120 that the eapression ax*'+be+c admits of being put in the form a(x—a) (a—-), where a and £ are the roots of the equation aa* + ba+c=0. Thus a quadratic expression aaz*+bx+c is capable of being resolved into two rational factors of the first degree, whenever the equation az*+be+c=0 has rational roots; that is, when 0° — dac is a perfect square. *127. To find the condition that a quadratic function of x, y may be resolved into two linear factors. Denote the function by f(a, y) where S (% y) = at + hay + by? + 2ga+ fy +eTHE THEORY OF QUADRATIC EQUATIONS. 95 Write this in descending powers of x, and equate it to zero; thus . aa® + 2x (hy + g) + by? + Wy+e=0. Solving this quadratic in « we have aay +9) = Shy + 9) = @ by" + By 6) a , or aw hy + g=+ Jy (i ab) + 2y (lig af) + (g*= ae) Now in order that f(a, y) may be the product of two linear factors of the form px+qy+7, the quantity under od radical must be a perfect square ; hence (hg — af = (h — ab) (9° - ac). Transposing and dividing by a, we obtain ye A abe + 2fgh — af* — bg* — ch? =0; , i which is the condition required. oF NS This proposition is of great importance in Analytical Geometry. *128. To find the condition that the equations ax'+be+c=0, a'x*+b'x+0c'=0 may have a common root. Suppose these equations are ‘both satisfied by x-:a; then ao'+ba+c=0, aa?+ Date’ =0; +. by cross multiplication af a it bc'-b'e ca’—c'a ab’—ab® To eliminate a, square the second of these equal ratios and equate it to the product of the other two; thus a® a® 1 (ca’ —ca)*~ (be —0’c) * (ab’—a'b) by? « (ca’— ca)’ = (be! — bc) (ab’ — a’b), which is the condition required. It is easy to prove that this is the condition that the two quadratic functions ax* + bay + cy? and a’x* + b'xy + cy? way have a common linear factor.96 HIGHER ALGEBRA. *EXAMPLES, IX. c. 1, For what values of m will the expression P+ Qvy + 20+ ny -3 be capable of resolution into two rational factors ? 2. Find the values of m which will make 2z*+ may +3y-5y-2 equivalent to the product of two linear factors, 3. Shew that the expression A (ey) — ay (B- 0) always admits of two real linear factors. 4, Ifthe equations @tprt+g=0, a+p'r+q/=0 have a common root, shew that it must be either p¢-Pd 4 I-F q-¢ PP 5. Find the condition that the expressions lx2®+may+ny*, Ux8+m'ay+n'y* may have a common linear factor. 6. Ifthe expression 3a2 +4 2Pay + 2y* + 2ax—4y+1 can be resolved into linear factors, prove that P must be one of the roots of the equation P?+ 4aP + 2a?+6=0. 7. Find the condition that the expressions an? + Qhay + by®, a’x?+ Way + by? may be respectively divisible by factors of the form y—mx, my +x. 8, Shew that in the equation a2 — Bary + 2y%— 20 — By —35=0, for every real value of x there is a real value of y, and for every real value of y there is a real value of 2. 9, If. and y are two real quantities connected by the equation 9x? + 2ay +y? — 92x — 20y+244=0, then will « lie between 3 and 6, and y between 1 and 10. 10. If (ax*+br+c)y+a'x?+b'a+e =0, find the condition that « aay be a rational function of y. ‘CHAPTER X. MISCELLANEOUS EQUATIONS. 129, In this chapter we propose to consider some mis- cellaneous equations ; it will be seen that many of these can be solved by the ordinary rules for quadratic equations, but others require some special artifice for their solution. Example 1. Solve sa — a —68. . Multiply by z™ and transpose; thus i} a 8a" ~ 6322" - 8-0; Let 2aby? - 6a*y — b*y + 8ab=0; (2ay — 2) (by - 8a) =0; b 8a y=ay * Fi 2 9a? oS gg 8 Bet & gat that is, w= ah pee HA, -!98 HIGHER ALGEBRA. Example 8, Solve (x-65)(«- 7) (x+6) (x+4)=504. - ‘We have (2? - @— 20) (23 - @ - 42) = 504 ; which, being arranged as a quadratic in 2*- x, gives (a2 - x)? - 62 (a?) +886=0; -. (#3? -@-6) (2? -2- 56)=0; .. w—2-6=0, or 2-2-56=0; whence 2=8, -2, 8, -7. 130. Any equation which can be thrown into the form as! +be+e+pJax+bete=q may be solved as follows. Putting y= ast + bx + ¢, we obtain y+ py-q=0. Let a and £ be the roots of this equation, so that JaFilete=a, Jattbete=B; from these equations we shall obtain four values of a. ‘When no sign is prefixed to a radical it is usually understood that it is to be taken as positive; hence, if a and B are both positive, all the four values of « satisfy the original equation. If however a or f is negative, the roots found from the resulting quadratic will satisfy the equation aa + b+ c—p Jaa + be+ce=4, but not the original equation. Example. Solve x?-52+2,/27—6a+8=12, Add 8 to each side; then 22- 604+842,/29- be+8=15, Putting /z?-62+8=y, we obtain y*+2y—15=0; whence y=8 or — 5. Thus /z?—62+3= +3, or Jz?—b2t5= —5. Squaring, and solving the resulting quadratics, we obtain from the first a=6 or —1; and from the second paoaNlisy The first pair of values satisfies the given equation, but the second pair satisfies the equation 2—be—2 /8— be +8=12.MISCELLANEOUS EQUATIONS. 99 131. Before clearing an equation of radicals it is advisable to examine whether any common factor can be removed by division. Ezample. Solve ./z*— Tax + 10a? - ./z* + az — 6a*=2 — 2a, We have A (@= 2a) (@— 6a) - ,/(@— 2a) (e+ Ba) =x - 2a, The factor ,/z— 2a can now be removed from every term; o. faba- Ju4+8a=/z-2a; 2z—ba+z+8a-2 /(z— 6a) (@+8a)=2-2a0; 2=2,/2? -2az —1ba?; 323 — 8ax — 60a?=0; : (z — 6a) (82+ 10a)=0; a=6a, or -&. Also by equating to zero the factor Nz — 2a, we obtain z=2a. On trial it will be found that z=6a does not satisfy the equation: thus the roots aro - 22 and 2a. The student may compare a similar question discussed in the Elementary Algebra, Art. 281, 132. The following artifice is sometimes useful. Example. Solve ,/3a?— 42434 +/8a9—4@—11=9 oe eeeeeeeee (vy. We have identically (8x3 - 42 + 34) — (327 - 4-11) =4! Divide each member of (2) by the corresponding member of (1); thus nl Bat = ae 84 — [Ba Le TL HB oe seeeeseeee (3). Now (2) is an identical equation true for all values of z, whereas (1) is an equation which is true only for certain values of x; hence also equation (3) is only true for these values of 2. From (1) and (8) by addition /8a2 — 404 84=7 ; whence x=8, or -;100 HIGHER ALGEBRA. 133, The solution of an equation of the form ase! te bar? ae car me ba + a = 0, in which the coefficients of terms equidistant from the beginning and end are equal, can be made to depend on the solution of a quadratic. Equations of this type are known as reci equa- tions, and are so named because they are not altered when 2 is changed into its reciprocal 7 For a more complete discussion of reciprocal equations the student is referred to Arts. 568—570. Example. Solve 12z4— 562% +8929 - 562-+12=0. Dividing by x? and rearranging, 12 (#+3) — 56 (e+ 3) +89=0. 2 x, 1 1 Put += 85 then 2° + 5=2*-2; w+ 12 (2% - 2) - 562 +89=0; whence we obtain r=, or, pba opl8 Veto =H Ors. . . 132 By solving these equations we find that «=2, a°3°5° 134. The following equation though not reciprocal may be solved in a similar manner. Example. Solve Gat — 252 + 12244 2524+6=0. ‘We have 6(2+3) - 25 ( -})+12=0; 2. @, 1\3 1 whence 6(#-3) ~25 (2-2) 424-0; @, x. 2 (#-3)-8=0 or 8(#-3)-0~0; @, ©, :, 1 1 whence we obtain ,2=2, - 8, - z 135. When one root of a quadratic equation is obvious by inspection, the other root may often be readily obtained by making use of the properties of the roots of quadratic equations proved in Art. 114,MISCELLANEOUS EQUATIONS. 101 Example. Solve (1 -a*) (z+a) - 2a(1-27)=0. This is a quadratic, one of whose roots is clearly a. Also, since the equation may be written 2az* + (1- a?) z- a (1+a%)=0, ie, l+a? EXAMPLES. X. a. Solve the following equations : 1, 2-3-2758, 2 9+a0-4=10c-% 1 3 1 1 3. 2fet+2e 3=5, 4, Gxt= Teton 4, 2 1 1 1 5. a 46=52", 6. 32%—g8-2=0, 3 le a er 1. 6/247, [5-254 8. [2+ f= 1 6 9, 6J/x=be #13, 10,. 14+825+9,423=0. 11, 3%+49=10.3% 12, 5(5#+5-*)=26. 13, 2+841=32. 2% 14, 2% +3_57=65 (2"— 1 3 Be . /2® =2, . a 15. + e=2 16, ie =Byy. 17. (@—"7) (7-3) (v+5) (w7+1)=1680. 18, (%+9) (7-38) (# - 7) (+5) =385. 19. % (+1) (x — 2) (20 -3)=63. 20, (20-7) (x8 - 9) (20 +5)=91. D1, 2242/29 +6r= 24 — 6x. 22, 348-404 /37—42—6=18. 93, 308-743 /3a9— 16x +21 = 162. 24, 8+9,/(Bx—1) (7-2) =322- 7x. — 2 Ben + oem wepg— Et,102 8 8S REEBBRBESBBR B 8 41, HIGHER ALGEBRA. te VSP _ (Baye). Jie —Ta — 15 — J 28-80 = a8 9. Jil 024443 /e-1= Jaap Sia. 208450 —7 + 8 (a Te +6) — 728 — 6x —1=0. Jaeiaa— 8a — Japan a= /Fatp San — oa, V2 + Ba —2— 208+ 62—9=1. /8a8 8249+ /328— 22—4=13. (3a Hi — , JS Oe Fda. 3x8 — Ta — 30 — /22*—Ta-5= 2-5. wh+-08 — 49 +0-4+1=0, ws Sotp 1a e8482, 37. a1 3 (23-42) = 22%, 10 (24-41) — 63x (22-1) +5222=0. zt J/a-2 = etl, 40. at2e+ Jat— 4% on " g-Jida-2 Ja- ” a+2n—Ja—4ee a et JB _o-JPH1_ 9 amare, v-Je=-1 «+ Je-1 Nee eo e 43, Stinet /. , oe; 228 +1, 45. a®(a?+1)=(a+a7)a. 8Je-5 _ 80-7 7: a7, 18(77-3) 250 ,/22+1 “Br-7 «ae —8 * “eeF1 ~ 3/7 -3 | (a+a44(a—2)i=s (aa, . Nat-tar—1— /+br—1=Ja- ob. at Je=1 2-1 e—Nal 2p fast t4— 2234 2 = 380. BA Wa 4 WyeMISCELLANEOUS EQUATIONS. 103 136. We shall now discuss some simultaneous equations of two unknown quantities. Example 1, Bolve 2+2+y+8+a/(@+2)(y+8)=89. (+2) + (y +3)? + (+2) (y+8) =741. Put 2+2=u, and y+8=v; then uto+,/uv=39... (1), u+09+u0=741.. (2), hence, from (1) and (2), we obtain by division, UD UV HID eee ccceceesteeesesesenees (8). From (1) and (3), u+v=29; and ‘uv =10, or uv=100; whence u=26, or 4; v=4, or 25; thus #=28, or 2; y=1, or 22. Example 2. Solve ah + y*= 82, -(1), (2). Put z=ut+v, and y= then from (2) we obtain v=. Substituting in (1), (w+1)4+(u-1)4=82; os 2 (wh+ 6u? +1) =82; 0; us+6u?— 40=0; whence w=4, or —10; and u= £2, or + /-10. Thus 2=8, -1,14,/-10; y=1, -8, -1+,/-10. Ezample 8. Solve ne - arya seseesueessetesseesessvessseeed Tat BY =29 weeereseee need From (1), 15 (227+ 82y +y?— 82% + day — y?)=38 (827 + 2xy — y*) ; wt. 12929 - 292ry — 88y2=0; wt. (Be - 2y) (482 +19y) =0. Hence 8e=2y .... or 482 = -19y104 HIGHER ALGEBRA. From (3), *. 2=2, y=3, Again, from (4), aL iy let =-% by equation (2), 561 1247 Hence 2=2, y=8; ores - 55 9a Ezample 4. Solve 4234 8aty +y3=8, 223 - Bx2y + 2y?=1, Put y=mz, and substitute in both equations. Thus 23 (44+8m+m)=8 . 23(2-2m+m)=1 . . 4+8mtm °"2-am+mi”? m3 — 8m? +19m -12=0; that is, (m—- 1) (m- 8) (m- 4)=0; «. m=l1, or 8, or 4, (i) Take m=1, and substitute in either (1) or (2). From (2), @=1; .. e=1; and y=mr=z=1, (ii) Take m=8, and substitute in (2); thus Stal; a= 4/53 and y=me=82=3 a/5: (iii) Take m=4; we obtain and y=mr=MISCELLANEOUS EQUATIONS. 105 Hence the complete solution is asl al 2=1, Je Js s/l ay. yn ag/t, an/d. Nore. The above method of solution may always be used when the equations are of the same degree and homogeneous. Example 5. Solve 8127y?-7y4-1122y+64=0 xt Try +4y2+8=0...... From (2) we have ~8=23-7zy+4y?; and, substituting in (1), BLaty?— Tyt+ Lay (24 - Tay + 4y!) + (a8 ~ Tey +-4y")?=0; oe BLaty! — Ty + (x? — Tay + Ay?) (1day + 29 — Tay + 4y?) =0; oe BLx2y? — Ty* + (a2 + dy)? — (7ay)?=0; that is, ah — 10z%y? + 9y4=0 w+ (ty) (@— 99") =0; hence w= ty, ora= + By. Taking these cases in succession and substituting in (2), we obtain gays +2; ee-yea J; a= £8, y= +1; Norz. It should be observed that equation | (8) is homogeneous. The method here employed by which one equation is made homogeneous by a suitable combination with the other is a valuable artifice, It is especially usefal in Analytical Geometry.106 HIGHER ALGEBRA. : aty + 1 This equation is a quadratic in 4) , from which we easily find, + (2) =2orl; whenoe 7+¥_8 or1; o-y, a-y . Te=9y, or y=0. Combining these equations with (2), we obtain 2=9, y=7; or a=¥, y=0. EXAMPLES. X. b. Solve the following equations : 1, 3e-2y=", 2 5e-y=3, 3 4x —3y=1, ary =20, x - 622=25, 12ry +13y?=25. 4, aA+ otyt+yA—931, 5. w+ ay +y*=84, w- xy +=19. a —Jay+y =6. 6. x +J/ay+y =65, T vty =T4+A/zy, w+ xy +y2=2275. a4 y2= 138 - ay. 8. 3x? 5y?=7, 9. 5y%—7e*=17, 10. 32%+165=162y, Bay — 4y?=2, bay — 62° =6, ‘Tay + 8y?= 132. i. 3x4 +ay+y*=15, 12, x+y?-3=32y, Bay — 3x9 — 5y?= 45, 222-6 +y2=0. 13, «t+y*=706, 14. at+y4=272, 15. 25 -y5=992, aty=8. a—y=2 a—y=2. a 2 ye _9 7 YL 1 245-1, m5 er a 1. 5+5=5 4 3 2,5 5 yt 5 =. aay ty 6" ,o4 ,o4 19. «+y=1072, 20. xy?+ya*=20, 21, w+yi=5, ees Bo aboot 2+ Y= 16, +yi=6d. Sle try ASS.MISCELLANEOUS EQUATIONS, 107 2, Jary+Jz—y=4, 23, y+ /e-1=2, an Natl Je-1=Jy. Jet v.00 25, Venvy , VatWy a Naty” Je=Jy~ Y eyo, ayy =106 244y?-152=10(8y-8), xy=6. wy +400=4lzy, y*=bay— 42%, 4a? + 5y=6 + 20ay —25y2 +22, 7e—1ly=17. 927+ 33x —12=12xy —4y?+29y, a? -2y=18, (—y?) (@-y)=l6xy, (at—y) (2? y") = 64022? 2x2 — wy +y=Qy, 2x + doy = by. Bap ay _ = G@ryp* @=ye8 ? wh dey. 494-0, 2 (y? - dary + 228) 4+8=0. 328 — 8ry?+y84+21=0, a2 (y—z)=1. x (409-108) = (28-998), 22° + 92y+y?=108, 624 +22? + 16=22(laet+y’), 2+ay-y~=4 a(at+z)j=y(b+y), axt+by=(xet+y). ay+ab=2ac, 2%? + ab? = 2bYy?, ® BSBRBRB ba—Ty=4, S$ 8S88R8 8 b23 = 10a*bx+ Baty, ay’ =10ab y +3b%x. 41, 2a(= -2) $a dot -G= = 137. Equations involving three or more unknown quantities can only be solved in special cases. We shall here consider some of the most useful methods of solution. Ezample1. Solve 2+y +2 =13. att yt+22=65. wy=10. From (2) and (3), (z+y)?+2=85. Put u for z+y; then this equation becomes ul+ 22=85,108 HIGHER ALGEBRA. Also from (1), ute=18; whence we obtain w=7 or 6; «=6 or 7. Thus we have aty= 7, zty= 6, wyai0 8d 830 Hence the solutions are z=b, or 2, a=84n -1, y=2, or 5,\ or y=89V-1, 2=6; = 2=7, Example 2. Solve (c+y)(2+2)=80, (y+s) (y+2)=15, (2+2) (e+y)=18. Write u, v, w for y+2, 2+2, x+y respectively ; thus vw=80, wu=15, w=18 . Multiplying these equations together, we have who? = 80 x 15 x 18= 15" x 6?; . uvw= £90, Combining this result with each of the equations in (1), we have u=8, v=6, w=5; or u=-38, v=-—6, w=- 5; oe yte =8, yte=-8, ris=6 } or +2= a at+y=5; z+ whence a=4, y=1, 2=2; or e=-4, y=-1, 2=-2. Example 8. Solve ptyet2=49. 24254e=19 e+ay+y2=89 Subtracting (2) from (1) yt a+2(y — 2)=80; that is, (y —2) (ey +2) 80 ieee se eeeeee eel (4). Similarly from (1) and (8) (f= 2)(2 YH 2) ENO ee cccccssecsesneeensed (5). Hence from (4) and (5), by divisionMISCELLANEOUS EQUATIONS. 109 Substituting in equation (3), we obtain 28-822 +827=18, From (2), e+ 22+ 22=19. Solving these homogeneous equations as in Example 4, Art. 186, we obtain = +2, 2= +38; and therefore y= “fi 1 or mAh and therefore y = =$—, “a a Example 4, Solve 29-yz=a", y*-ex=b3, 22—2y=c3, Multiply the equations by y, z, x respectively and add; then C7 + PY +B HO oo. ceeeceseteeeeeseesseeeees (1). Multiply the equations by z, x, y respectively and add; then Be+e%y+a%—=0 . From (1) and (2), by cross multiplication, (2). ae z aR aah aman = * eUPPOM. Substitute in any one of the given equations; then B (a® +084 c8 - 8a%%e%) =1; y oe eee eee B= a aa Nato ct Bao EXAMPLES, X. c. Solve the following equations : 1. 9x+y—8e=0, 2 3e+y-22=0, 4a —8y+72=0, 42 —-y—32=0, yore t+ ry =47. B+y+8=467, 3 w-y-2=2, 4. v+2y—2=11, wi+y?—A=22, a — 4y8 4 8=37, ay =. we=24, 5, a+y?-2=21, 6 w+2y+22=18, Bre + By2— Lay =18, Pt+yetye+12=0, a+y—2=5. A+cr+z2y=30, 7. a+ 2ry+302=50, 8. (y—2)(2+2)=22, | iy! + 92 +9010, (+2) (e-y)=33, | 8A +4424 2 =10, (w—-y) (y-2)=8.110 HIGHER ALGEBRA. 9. a%yfetu 12, oytut=8, 2tyeut=1, Baytut=4, 10, y%=12, ay =54, alyt=72, lL ayt+at+y=23, 12, Qay-4e+y=17, x2+e+2=41, ~ 3yz+y —62=52, yetyt+e=27. 602 +32 + 22 =29. 13, wety=T2, ye+e=82, et+yt2=12. 14, 8+y4+3=a8, e+y2+2=a%, r+y+e=a. 15. 492+ 2=y2+er+ay=al, 3e-—y+2=a,/3. 16. a8+y?+2=21a}, yz + 2r—xy=6u?, 8x t+y—-22=3a. INDETERMINATE Equations. 138, Suppose the following problem were proposed for solu- tion : A person spends £461 in buying horses and cows; if each horse costs £23 and each cow £16, how many of each does he buy? Let a, y be the number of horses and cows respectively ; then 23x + 16y = 461. Here we have one equation involving ¢wo unknown quantities, and it is clear that by ascribing any value we please to 2, we can obtain a corresponding value for y; thus it would appear at first sight that the problem admits of an infinite number of solutions. But it is clear from the nature of the question that 2 and y must be positive integers; and with this restriction, as we shall see later, the number of solutions is limited. If the number of unknown quantities is greater than the number of independent equations, there will be an unlimited number of solutions, and the equations are said to be indeter- minate. In the present section we shall only discuss the simplest kinds of indeterminate equations, confining our attention to posi- tive integral values of the unknown quantities; it will be seen that this restriction enables us to express the solutions in a very simple form, The general theory of indeterminate equations will be found tn Chap. Xxv1., INDETERMINATE EQUATIONS. 111 Example 1. Solve 7x+12y=220 in positive integers. Divide throughout by 7, the smaller coefficient ; thus aey+ aaiss; 5 aty+ 3 nat Since z and y are to be integers, we must have and therefore that is, y-2.. and therefore a =integer= p suppose, i. y-2=Tp, or YHTDFD ececccccerccseeeete cee cae eee (2). Substituting this value of y in (1), t+ 7p+245p+1=381; that is, If in these results we give to p any integral value, we obtain corresponding integral values of z and y; but if p > 2, we see from (3) that « is negative ; and if p is a negative integer, y is negative. Thus the only positive integral values of x and y are obtained by putting p=0, 1, 2. The complete solution may be exhibited as follows: p= L 3 z=28, 16, 4, y= 2, 9, 16, Nore. When we obtained we 5 _ ntoger, we multiplied by 3 in order to make the coefficient of y differ by unity from a multiple of 7. A similar artifice should always be employed before introducing a symbol to denote the integer. Example 2. Solve in positive integers, 14x -11ly=2! Divide by 11, the smaller coefficient; thus Ba 7 at+ii-y= a+ th? B2-7 A T=? ~@+y =integer ; (1).112 : HIGHER ALGEBRA. 122-28 hence Ti = integer; that is, 2 242° integer ; », 27S eintegar=p suppose; o. e=lip+6 } and, from (1), y=ld4p+5 This ‘is called the general solution of the equation, and by giving to p any positive integral value or zero, we obtain positive integral values of x and y; thus we have 2=6, 17, 28, 89, y=5, 19, 88, 47 the number of solutions being infinite. . Ezample 8, In how many ways can £5 be paid in half-crowns and florins? Let 2 be the number of half-crowns, y the number of florins; then 52 -+4y=200; P=0, 1, 2 By ccseecssrees and y=50-5p. Solutions are obtained by ascribing to p the values 1, 2, 3, ...9; and therefore the number of ways is 9. If, however, the sum may be paid either in half-crowns or florins, p may also have the values 0 and 10. If p=0, then «=0, and the sum is paid entirely in florins; if p=10, then y=0, and the sum is paid entirely in half-crowns. Thus if zero values of z and y are admissible the number of ways is 11. Example 4. The expenses of a party numbering 48 were £5. 14s. 6d.; if each man paid 5s., each woman 2s. 6d., and each child 1s., how many were there of each? Let 2, y, 2 denote the number of men, women, and children, respectively ; then we have Eliminating z, we obtain 82+3y=143. The general solution of this equation is 2=8p+1, y=45-8p;INDETERMINATE EQUATIONS. 113 Hence by substituting in (1), we obtain z=bp-3. Here p cannot be negative or zero, but may have positive integral values from 1 to 5. Thus p= 1, 2, 8, 4, 5; z= 4, 7,10, 18, 16; y =87, 29, 21, 18, 5; z= 2, 7,12,17,22. . EXAMPLES. X. d. Solve in positive integers: 1. 3c4+8y=103. 2, Sat+2y=53. 3. Te-+12y=152. 4, 13¢+lly=414. 5, 237+25y=915. 6, 417+47y=2191. Find the general solution in positive integers, and the least values of « and y which satisfy the equations: 1 5a—Ty=3. 8 6%—-13y=1. 9, 82-21y=33. 10. 17y-18¢=0. 11, 19y-282=7. “12 77y-300—=295. 13, A farmer spends £752 in buying horses and cows; if each horse costs £37 and each cow £23, how many of each does he buy ? 14. In how many ways can £5 be paid in shillings and sixpences, including zero solutions ? 15, Divide 81 into two parts so that one may be a multiple of 8 and the other of 5. 16. What is the simplest way for a person who has only guineas to pay 10s. 6d. to another who has only half-crowns ? 17. Find a number which being divided by 39 gives a remainder 16, and by 56 a remairider 27. How many such num! are there? 18. What is the smallest number of florins that must be given to discharge a debt of £1. 6s. 6d., if the change is to be paid in half-crowns only? 19. Divide 136 into two parts one of which when divided by 5 leaves remainder 2, and the other divided by 8 leaves remainder 3. 20. I buy 40 animals consisting of rams at £4, pigs at £2, and oxen at £17: if I spend £301, how many of each do I buy? 21, In my pocket I have 27 coins, which are sovereigns, hali-crowns, or shillings, and the amount I have is £5. 0s. 6d.; how many coins of each sort have I? AHA. aCHAPTER XL PERMUTATIONS AND COMBINATIONS. 139, Eacu of the arrangements which can be made by taking some or all of a number of things is called a permutation. Each of th€groups or selection\which can be made by taking some or all of a numbe ig8 18 called a combination, Thus the permutations which can be made by taking the letters a, b, c, d two at a time are twelve in number, namely, ab, ac, ad, be, bd, cd, 7 ba, ca, da, cb, db, de; each of these presenting a different arrangement of two letters. The combinations which can be made by taking the letters a, b, c, d two at a time are six in number: namely, ab, ac, ad, be, bd, cd; each of these presenting a different selection of two letters. From this it appears that in forming combinations we are only concerned with the number of things each selection contains ; whereas in forming permutations we have also to consider the order of the things which make up each arrangement; for instance, if from four letters a, 6, c, d we make a selection of three, such as abe, this single combination admits of being arranged i in the following ways: abe, ach, bea, bac, cab, cba, and so gives rise to six different permutations.PERMUTATIONS AND COMBINATIONS, 115 140. Before discussing the general propositions of this chapter there is an important principle which we proceed to explain and illustrate by a few numerical examples. If one operation can be performed in m ways, and (when it has been performed in any one of these ways) a second operation can then be performed in n ways; the number of ways of per- forming the two operations will bem x n. If the first operation be performed in any one way, we can associate with this any of the m ways of performing the second operation : and thus we shall have n ways of performing the two operations without considering more than one way of performing the first; and so, corresponding to each of the m ways of per- forming the first operation, we shall have n ways of performing the two; hence altogether the number of ways in which the ,two operations can be performed is represented by the product mMmxn, Example 1. There are 10 steamers plying between Liverpool and Dublin; in how many ways can a man go from Liverpool to Dublin and return by a different steamer? There are ten ways of making the first passage; and with each of these there is a choice of nine ways of returning panes the man is not to come back by the same steamer); hence the number of ways of making the two journeys is 10x 9, or 90. This principle may easily be extended to the case in which there are more than two operations each of which can be per- formed in a given number of ways. Example 2. Three travellers arrive at a town where there are four hotels; in how many ways oan they take up their quarters, each at a different hotel? The first traveller has choice of four hotels, and when he has made his selection in any one way, the second traveller has a choice of three; there- fore the first two can make their choice in 4 x 3 ways; and with any one such choice the third traveller can select his hotel in 2 ways; hence the required number of ways is 4 x 8 x 2, or 24. 141, To find the number of permutations of n dissimilar things taken r at a time. This is the same thing as finding the number of ways in which we can fill up 7 places when we have n different things at our disposal. { The first place may be filled up in n ways, for any one of the ® things may be taken; when it has been filled wp in any one ot 82116 HIGHER ALGEBRA. these ways, the second place can then be filled up inn —1 ways; and since each way of filling up the first place can be associated with each way of filling up the second, the number of ways in which the first two places can be filled up is given by the product m(n—1). And when the first two places have been filled up in any way, the third place can be filled up in n-2 ways. And reasoning as before, the number of ways in which three places can be filled up is n (n— 1) (n— 2). Proceeding thus, and noticing that a new factor is introduced with each new place filled up, and that at any stage the number of factors is the same as the number of places filled up, we shall have the number of ways in which r places can be filled up equal to ; n(n—1)(n—2)...... to r factors ; and the r* factor is n-(r-1), or n-r+l. | Therefore the number of permutations of » things taken r at a time is n(n—1) (n—-2)..... (n-r+1). Cor. The number of permutations of n things taken all at a time is n(n—1)(n—2)...... to m factors, or n(n—1)(n—2)......3.2. 1. It is usual to denote this product by the symbol |n, which is read ‘factorial n.” Also 7! is sometimes used for |n. 142. We shall in future denote the number of permutations of n things taken r at a time by the symbol "P_, so that (@)gP.=n(n-1) (n-2).....0=r41); also fy’ 2 "PL=|m In working numerical examples it is useful to notice that the suffix in the symbol "P, always denotes the number of factors in the formula we are using. 143, The number of permutations of » things taken r at a, time may also be found in the following manner. Let "P, represent the number of permutations of n things taken r at a time.PERMUTATIONS AND COMBINATIONS. 117 Suppose we form all the permutations of » things taken r—1 at a time; the number of these will be "P__,. With cach of these put one of the remaining n—1+ 1 things. Each time we do this we shall get one permutation of n things r at a time; and therefore the whole number of the permutations of n things r at a time is *P__, x (n—r+1); that is, “P.="P_.x(n—r4+1). By writing r—1 for r in this formula, we obtain “Py ="P,_, x (n—7 + 2), similarly, "P_ P_,x(n—r+3), "P,="P,x(n— 2), "P,="P,x(n-1), "Pi=n. Multiply together the vertical columns and cancel like factors from each side, and we obtain "P,=n(n—1) (w—-2)...... (n-r+1). Example 1. Four persons enter a railway carriage in which there are six seats; in how many ways can they take their places? The first person may seat himself in 6 ways; and then the second person in 5; the third in 4; and the fourth in 3; and since each of these ways may be associated with each of the others, the required answer is 6x 5x 4x8, or 360, Example 2. How many different numbers can be formed by using six out of the nine digits 1, 2, 3,...9? Here we have 9 different things and we have to find the number of per- mutations of them taken 6 at a time; -". the required result =°P, “9 x8x7x6x5x4 = 60480. 144. To find the number of combinations of n dissimilar things taken r at a time. Let *C, denote the required number of combinations. Then each of these combinations consists of a group ot + dissimilar things which can be arranged among themselves in |r waya [Art, 142.]118 HIGHER ALGEBRA. Hence "C, x |r is equal to the number of arrangements of n things taken r ata time; that is, "C_x|r= =n(n—-1) (n—2)...(n—r+1); 4 _n(n—1)(n- 2)... (aert)) . c, a Cor, This formula for *C, may also be written in a different form ; for if we multiply the numerator and the denominator by [n= + we obtain n(n—1) (n—2).. (n- r+1)x|n—r ———Tr[n=rt—~—S The numerator now consists of the product of all the natural numbers from 7 to 1; - (1). ont “Tina It will be convenient to remember both these expressions for *"C,, using (1) in all cases where a numerical result is required, and (2) when it is sufficient to leave it in an algebraical shape. Nors. If in formula (2) we put r=n, we have noe = 2 Cae as ee but *C,,=1, so that if the formula is to be true for r=n, the symbol {0 must be considered as equivalent to 1. Example. From 12 books in how many ways can a selection of 5 be made, “wr when one Pere book is always included, (2) when one specified book is always excluded ? (1) Since the specified book is to be oe in every selection, we have only to choose 4 out of the remaining 11 Hence the number of ways="C, —11x10x9x8 “TX8x3xt =330,PERMUTATIONS AND COMBINATIONS. 119 cot £2 Sinoe the specified book is always to be excluded, we have to select 5 books out of the remaining 11. Hence the number of ways=""C, _11x10x9x8x7 ~~ 1x8xBx4xb = 462, \ 145. The number of combinations of n things r at a time ts equal to the number of combinations of n things n—r at a time. In making all the possible combinations of n things, to each group of r things we select, there is left a corresponding group of n—r things; that is, the number of combinations of n things r at a time is the same as the number of combinations of n things n—r at a time; o "C=", The proposition may also be proved as follows : “0 im Toe n—(n-r) nm In — 4 |r ="C,. Such combinations are called complementary. [Art. 144,] Norge. Put r=n, then "C,="C,=1. The result we have just proved is useful in enabling us to abridge arithmetical work. Example. Out of 14 men in how many ways can an eleven be chosen? — The required number="C,, =, _ 14x18 x12 ~“Txax8 = 864, If we had made use of the formula 4C,,, we should have had to reduce ox expression whose numerator and denominator each contained 11 factors.120 HIGHER ALGEBRA. 146. To find the number of ways in which m +n things can be divided into two groups containing m and n things respectively. This is clearly equivalent to finding the number of combi- nations of m+n things m at a time, for every time we select one group of m things we leave a group of n things behind, Thus the required number = me : 7” Nore. different ways of subdivision is PRE for in any one way it is possible to —- the two groups without obtaining a new distribution, number of ways in which m +n +p things can be divided into three groups containing m, n, p things severally. First divide m+n+p things into two groups containing m and n+p things respectively; the number of ways in which this m+n+p If n=m, the groups are equal, and in ie number 0} 2: can be done is |m |n+p- Then the number of ways in which the group of n+p things can be divided into two groups containing ” and p things respec- n+p [Pe Hence the number of ways in ‘which the subdivision into three groups containing m, n, p things can be made is m+n +P. m+ +p im + + Pp jm [exp * np’ EEE Nore. If we put n=p=m, we obtain Em _, mi but this formula regards a tively is as different all the possible orders in whit e three groups can occur in any one mode of subdivision. And since there are |3 such orders cor- responding to each mode of subdivision, the number of different ways in which subdivision into three equal groups can be made is Pees Exzample.( The ni in which 16 recruits can be divided into [15 three equal ‘oups is cEER'™ and the number of ways in which theyPERMUTATIONS AND COMBINATIONS. 121 148, In the examples which follow it is important to notice that the formula for permutations should not be used until the suitable selections required by the question have been made. Example 1. From 7 Englishmen and 4 Americans a committee of 6 is to be formed; in how many ways can this be done, (1) when the committee con- tains exactly 2 Americans, (2) at least 2 Americans? (1) We have to choose 2 Americans and 4 Englishmen. The number of ways in which the Americans can be chosen is ‘C,; and the number of ways in which the Englishmen can be chosen is 7C,. Fach of the first groups can be associated with each of the second; hence the required number of ways=‘C, x7C, 4 1% =)a]2* 74] 22 il ae =210. 2 /2(8 Tos (2) The committee may contain 2, 3, or 4 Americans. ‘We shall exhaust all the suitable combinations by forming all the groups containing 2 Americans and 4 Englishmen ; then 3 Americans and 3 English- men; and lastly 4 Americans and 2 Englishmen, The sum of the three results will give the answer. Hence the required number of ways 40, x 70g 440g x 705 +4C, x70q 4 it 4. “GEE 8” BeBe =210+140+21=371. In this Example we have only to make use of the suitable formul# for combinations, for we are not concerned with the possible arrangements of the members of the committee among themselves, CEzample 2. Out of 7 consonants and 4 vowels, how many words can Le ma Cont g 3 consonants and 2 vowels? The number of ways of choosing the three consonants is 7C,, and the number of ways of choosing the 2 vowels is ‘C,; and since each of the first groups can be associated with each of the second, the number of combined groups, each containing 3 consonants and 2 vowels, is 7C, x 4C,. Farther, each of these groups contains 5 letters, which may be arranged among themselves in |5 ways. Hence the required number of words="C, x ‘C, x |5 at xis [4° BO =5x|7 ¢ =25200.Y 122 HIGHER ALGEBRA, E. How many words can be formed out of the letters article, so that the vowels occupy the even places? Here we have to put the 8 vowels in 8 specified places, and the 4 conso- nants in the 4 remaining places; the first operation oan be done in |8 ways, and the second in |4. Hence — the required number of words = [8x |4 =14, In this Example the formula for permutations is immediately applicable, because by the statement of the question there is but one way of choosing the vowels, and one way of choosing the consonants. EXAMPLES XI. a. 1, In how many ways can a consonant and a vowel be chosen out of _ the letters of the word courage? 2, There are 8 candidates for a Classical, '7 for a Mathematical, and 4 for a Natural Science Scholarship. In how many ways can the Scholarships be awarded? 3, Find the value of ®P,, ™P,, *C,, C,. 4, How many different arrangements can be made by taking 5 of the letters of the word equation ? 5. If four times the number of permutations of n things 3 together is equal to five times the number of permutations of n—1 things 3 together, find x. 6. How many permutations can be made out of the letters of the word triangle? How many of these will begin with ¢ and end _with e? 7. How many different selections can be made by taking four of the digits 3, 4, 7,5, 8,1? How many different numbers can be formed with four of these digits? 8. If *C, : *C,=44 : 3, find n, 9, How many changes can be rung with a peal of 5 bells? 10. How many changes can be rung with a peal of 7 bells, the tenor always being last? 11, On how many nights may a watch of 4 men be drafted from a crew of 24, so that no two watches are identical? On how many of these would any one man be taken? 42) How many arrangements can be made out of the letters of the \ word draught, the vowels never being separated? \\ PERMUTATIONS AND COMBINATIONS. 123 13. In a town council there are 25 councillors and 10 aldermen ; how many committees can be formed each consisting of 5 councillors d 3 aldermen ? yi Out of the letters A, B, C, p, g, r how many arrangements can be a (1) beginning with a capital, (2) beginning and ending with a cap 15. Find the number of combinations of 50 things 46 at a time. 16. If "C,,="C,, find *Cj,, #C,. Ac, In how many ways can the letters of the word vowels be arratiged, if the letters oe can only occupy odd places ? * 18, From 4 officers and 8 privates, in how many ways can 6 be oan (1) to include exactly one officer, (2) to include at least one cer’ 19. In how many ways can a party of 4 or more be selected from 10 persons ? 20. If 8C,="C,,2, find "C,. 21, Out of 25 consonants and 5 vowels how many words can be formed each consisting of 2 consonants and 3 vowels? 22, Ina library there are 20 Latin and 6 Greek books; in how many ways can a group of 5 consisting of 3‘Latin and 2 Greek books be + placed on a heli \et n’how many ways can 12 things be divided equally among 4 mst dvidergoning beelle ares tag tec ory 24, From 3 capitals, 5 consonants, and 4 vowels, how many words can be made, each containing 3 consonants and 2 vowels, and beginning with a capital ? 25. At an election three districts are to be canvassed by 10, 15, and 20 men respectively. If 45 men volunteer, in how many ways can they be allotted to the different districts? In how many ways can 4 Latin and 1 English book be placed on a shelf so that the English book is always in the middle, the selec- tion being made from 7 Latin and 3 English books? S “ST A boat is to be manned by eight men, of whom 2 can only row on bow side and 1 can only row on stroke side; in how many ways can the crew be q 7 WY There are two works each of 3 volumes, and two works each of 2 volumes ; in how many ways'can the 10 books be placed on a shelf so that volumes of the same work are not separated? = 29, In how many ways can 10 examination papers be arranged 80 that the best and worst papers never come together’124 HIGHER ALGEBRA. 30. An eight-oared boat is to be manned by a crew chosen from 11 men, of whom 3 can steer but cannot row, and the rest can row but can- not steer. In how many ways can the crew be arranged, if two of the men can only row on bow side? : ] . Prove oo eased number of ways in which p positive and x negative signs ma: laced in a row so that no two negative signs shall be together is ?*+!€, "32, If P_,,: “Pg 430000 21, find r. , 33, How many different signals can be made by hoisting 6 differ. “ently coloured flags one above the other, when any number of them may be hoisted at once? 34. If*C,, : *C,,4=225 : 11, find r. 149, Hitherto, in the formule we have proved, the things have been regarded as unlike. Before considering cases in which some one or more sets of things may be like, it is necessary to point out exactly in what sense the words like and pa are ‘When we speak of things being dissi fj , UN- y-distiniguishable—from h other—‘On the other hand we = always usé the term Tike things to denote such as are alike to the eye and cannot be distinguished from each other. For instance, in Ex. 2, Art. 148, the consonants and the vowels may be said each to consist of a group of things united by a common characteristic, and thus in a certain sense to be of the same kind; but they cannot be regarded as like things, because there is an individuality existing among the things of each group which makes them easily distinguishable from each other. Hence, in the final stage of the example we considered each group to consist of five dissimilar things and therefore capable of [5 arrangements among themselves. [Art. 141 Cor.] 150. Suppose we have to find all the possible ways of arrang- ing 12 books on a shelf, 5 of them being Latin, 4 English, and the remainder in different languages. The books in each language may be regarded as belonging to one class, united by a common characteristic; but if they were distinguishable from each other, the number of permutations would be {12, since for the purpose of arrangement among them- selves they are essentially different. kPERMUTATIONS AND COMBINATIONS. 125 If, however, the books in the same language are not dis- tinguishable from each other, we should have to find the number of ways in which 12 things can be arranged among themselves, when 5 of them are exactly alike of one kind, and 4 exactly alike of a second kind: a problem which is not directly included in any of the cases we have previously considered. oi. To find the number of ways in which n things may be arranged among themselves, taking them all at a time, when p of the things are exactly alike of one kind, q of them exactly alike of another kind, v of them exactly alike of a third kind, and the rest all different. Let there be n letters ; suppose p of them to be a, g of them to be 6, r of them to be c, and the rest to be unlike. Let x be the required number of permutations ; then if in any one of these permutations the p letters a were replaced by p unlike letters different from any of the rest, from this single permutation, without altering the position of any of the remainin, letters, we could form |p new permutations. Hence if this change were made in each of the « permutations we should obtain x x lp permutations. Similarly, if the g letters 6 were replaced by g unlike letters, the number of permutations would be xx |p x |g. In like manner, by replacing the r letters ¢ by r unlike letters, we should finally obtain a x |p x |g x |r permutations. But the things are now all different, and therefore admit of |n permutations among themselves. Hence wx px lax |e= [es that is, a= fe -s plgk’ which is the required number of permutations. Any case in which the things are not all different may We treated similarly.126 HIGHER ALGEBRA. Era 1, How many different permutations can be made out of the letters rh word assassination taken all together? : ‘We have here 18 letters of which 4 are s, 8 are a, 2 are i, and 2 aren. Hence the number of permutations [18 “Eee =18.11.10.9.8.7.8.5 = 1001 x 10800= 10810800. Example 2. How many numbers can be formed with the digits 1, 2, 3, 4, 8, 2, 1, so that the odd digits always occupy the odd places? The odd digits 1, 3, 8, 1 can be arranged in their four places in 4 Pe WHY iS)sessecesscrstcsressccrtcsressccrreser? (1). The even digits 2, 4, 2 can be arranged in their three places in 8 (2 WAYB ooo ccc ccc ccececceeccesecssceesneeneeeeel (2). Each of the ways in (1) oye associated with each of the ways in (2). 8 Hence the required namber= 9a x nox 3=18. Ay\ 2, To find the number of permutations of n things r ata time, when each thing may be repeated once, twice,...... up tor times in any arrangement. Here we have to consider the number of ways in which r laces can be filled up when we have n different things at our —Trposal, each of the n things being used as often as we please in any arrangement. The first place may be filled up in n ways, and, when it has been filled up in any one way, the second place may also be filled up in n ways, since we are not precluded from using the same thing again. Therefore the number of ways in which the first two places can be filled up ismxn or n*. The third place can also be filled up in m ways, and therefore the first three places in n? ways. Proceeding in this manner, and noticing that at any stage the index of n is always the same as the number of places filled up, we shall have the number of ways in which the r places can be 4illed up equal to n’.PERMUTATIONS AND COMBINATIONS. 127 Example. In how many ways can 5 prizes be giver! away to 4 boys, when each boy is eligible for all the prizes? . Any one of the prizes can be given in 4 ways; and then any one of the remaining prizes can also be given in 4 ways, since it may be obtained by the boy who has already received a prize. Thus two prizes can be given away in 4? ways, three prizes in 48 ways, and so on. Hence the 5 prizes can be given away in 45, or 1024 ways. aes To find the total number of ways in which it is possible to selection by taking some or all of n things. * Each thing may be dealt with in two ways, for it may either be taken or left; and since either way of dealing with any one thing may be associated with either way of dealing with each one of the others, the number of selections is 2x2x2Qx2d...... to n factors. ~~ But this includes the case in whith all the things are left, therefore, rejecting this case, the total number of ways is 2"-1. This is often spoken of as “the total number of combinations” of n things. . Example. Aman has 6 friends; in how many ways may he invite one or > more of them to dinner? + He has to select some or all of his 6 friends; and therefore the number of Says is 26-1, or 63. # This result can be verified in the following manner. : The guests may be invited singly, in twos, threes,...... ; therefore the * number of selections =°C,+°C,+5C,+5C,+5C,+5C, =6415+20+4154+6+1=63, 154, To find for what value of r the number of combinations “of n things v at a time is greatest. . Since 0, -2 eV n- =r+2)(u-r+)) | - 0,="0,, x Bott}, The multiplying factor acrt} may be written meta, { which shews that it decreases as r increases. Hence a8 7 recekves128 . HIGHER ALGEBRA. the values 1, 2, 3...... in succession, "C, is continually increased. until = —1 becomes equal to 1 or less than 1. n+l Now ——-1>], r 1 8o long as ne >; that is, eens We have to choose the greatest value of r consistent with this inequality. (1) Let x be even, and equal to 2m; then n+1 Qm+1 al 2 2 +93 and for all values of r up to m inclusive this is greater thanr. 7” e we find that the greatest number of Hence by putting r= m= combinations is "C. 3 (2) Let n be odd, and equal to 2m +1; then n+l I%m+2 2° 2 and for all values of r up to m inclusive this is greater than r; but when r= m+ 1 the multiplying factor becomes equal to 1, and = "Ons, ="C,,; that is, "C.,,="C, 3 =mt+1; ‘nm “T and therefore the number of combinations is greatest when the things are taken ae , or a at a time; the result being the nd same in the two , 155. The formula for the number of combinations of n things r at a time may be found without assuming the formula for the number of permutations. Let "C, denote the number of combinations of n things taken rata flame and let the m things be denoted by the letters a, 8, o, a. Qt. _-->sPERMUTATIONS AND COMBINATIONS. 129 Take away a; then with the remaining letters we can form ""¢._, combinations of n—1 letters taken r—1 at atime. With each of these write a; thus we see that of the combinations of n things r at a time, the number of those which contain a is *"C_,; similarly the number of those which contain bis""C,_,; and so for each of the 7 letters. Therefore n x 1 18 equal to the number of combinations r ata time which contain a, together with those that contain }, those that contain ¢, and so on. But by forming the combinations in this manner, each par- ticular one will be repeated times. For instance, if r= 3, the combination abe will be found among those containing a, among those containing 5, and among those containing c«. Hence n ” °C 2} C.=""C_, xe By writing n ~1 and r—1 instead of nm and r respectively, - n-1 C=" Cg X TF ear - = n-2 Similarly, "30, ="C,_, x roe? n-r+2 mrtg THO. x and finally, "HC =n-rtl. Multiply together the vertical columns and cancel like factors from each side ; thus t n(n—1)(n—-2)......(n—r+1) r (r-1) (r—2) ...... T . 3 3 "C= ‘ Lis To find the total number of ways in which it is possible ti a selection by taking some or all out of p+ q+tt...- things, whereof p are alike of one kind, q alike of a second ki alike of a third kind; and so on. The p things may be disposed of in p+1 ways; for we may take 0, 1, 2, 3, ...... p of them. Similarly the q things may be disposed of in g+1 ways; the r things in r+ 1 Ways; 8 so on. EEA, Q ”Hence the number of ways in which all the things may be disposed of is (p+1)(q¢+1)(r+]1)...... . But this includes the case in which none of the things are taken; therefore, rejecting this case, the total number of ways is (p +1) @4+1) (+1)... -1, 157. A general formula expressing the number of permuta- tions, or combinations, of n things taken r at a time, when the things are not all different, may be somewhat complicated ; but a particular case may be solved in the following manner. ‘zample, Find the number of ways in which (1) a selection, (2) an ar- ent, of four letters can be made from the letters of the word proportion. There are 10 letters of six different sorts, namely 0, 0,0; p,p; 7,7; t; i; n. In finding groups of four these may be classified as follows: (1) Three alike, one different. (2) Two alike, two others alike. (3) Two alike, the other two different. (4) All four different. (1) The selection can be made in 5 ways; for each of the five letters, p, 7, t, i, n, can be taken with the single group of the three like letters o. (2) The selection can be made in °C, ways; for we have to choose two out of the three pairs 0,0; p,p; 7,7: This gives 8 selections, (8) This selection can be made in 3x 10 ways; for we select one of the 8 pairs, and then two from the remaining 5 letters, This gives 30 selections. (4) This selection can be made in °C, ways, as we have to take 4 different - ~_.letters to choose from the six 0, p, 7, t, i, n, This gives 15 selections, Thus the total number of selections is 5+8+480+15; that is, 53, In finding the different arrangements of 4 letters we have to permute in all possible ways each of the foregoing groups. if (1) gives rise to 5x ‘=, or 20 arrangements. iB iB ives ri 4 (2) gives rise to 3x Pe , or 18 arrangements. 4 (8). gives rise to 30x E , or 360 arrangements. (4) gives rise to 15x |4, or 360 arrangements. Thus the total number of arrangements is 20+ 18+ 360+ 360; that ia, 758.PERMUTATIONS AND COMBINATIONS. 131 EXAMPLES. XI. b. 1, Find the number of arrangements that can be made out of the etters of the words (1) independence, (2) superstitious, (3) institutions. 2. In how many ways can 17 billiard balls be arranged, if 7 of shem are black, 6 red, and. 4 white? 3, A room is to be decorated with fourteen flags; if 2 of them are olue, 3 red, 2 white, 3 green, 2 yellow, and 2 purple, in how many ways zan they be hung? 4, How many numbers greater than a million can be formed with the digits 2, 3, 0,3, 4, 2, 37 “ §, Find the number of arrangements which can be made out of the letters of the word algebra, without altering the relative positions of vowels and consonants. 6. On three different days a man has to drive to a railway station, and he can choose from 5 conveyances; in how many ways can he make the three journeys ? 7. Ihave counters of n different colours, red, white, blue,......; in how many ways can I make an arrangement consisting of r counters, supposing that there are at least r of each different colour? 8. In a steamer there are stalls for 12 animals, and there are cows, horses, and calves (not less than 12 of each) ready to be shipped ;- in how many ways can the shipload be made? . ) In how many ways can » things be given to sons, when theres no restriction as to the number of things each may receive? 10. In how many ways can five things be divided between twol persons ? . 11. How many different arrangements can be made out of the letters in the expression a%°ct when written at full length? . 12. A letter lock consists of three rings each marked with fifteen different letters ; find in how many ways it is possible to make an unguccessful attempt to open the lock. C13/ Find the number of triangles which can be formed by joining gular points of a quindecagon. a A lib: has a copies of one book, } copies of each of twe * books, ¢ copies of each of three books, and single copies of d books. Ir how many ways can these books be distributed, if all are out at once? 15. How many numbers less than 10000 can be made with the>- eight digits 1, 2, 3, 0, 4, 5, 6,7? 16. In how many ways can the following prizes be given owmay to & class of 20 boys: first and second Classical, first and second Mekne- matical, first Science, and first French? Qa132 HIGHER ALGEBRA. 17. A telegraph has 5 arms and each ay is capable 4 ey itions, including the position of rest; what is the total number ot Nignals that can be made? , In how many ways can 7 persons form a ring? In how many wa} n 7 Englishmen and 7 Americans sit down at a round table, no two Americans being together? 19, In how many ways is it ible to draw a sum of money from a bag containing a sovereign, a half-sovereign, a crown, a florin, a shilling, a penny, and a ing? 4 20, From 3 cocoa nuts, 4 apples, and 2 oranges, how many selec- frioas of fruit can be made, taking at least one of each kind? 21, Find the number of different ways of dividing mn things into n equal groups. 22, How many signals can be made by hoisting 4 flags of different colours one above the other, when any number of them may be hoisted at once? How many with 5 flags? 23, Find the number of permutations which can be formed out of the letters of the word series taken three together? : Q ‘4. ‘There are p points in a plane, no three of which are in the same straight line with the exception of g, which are all in the same straight line; find the number (1) of straight lines, (2) of triangles which result joining them. 25; There are p points in space, no four of which are in the same 6 with the exception of g, which are all in the same plane; find fow many planes there are each containing three of the points. 26. There are n different books, and p copies of each; find the number of ways in which a selection can be made from them. ’ 27, Find the number of selections and of arrangements that can be made by taking 4 letters from the word expression. 28, How many permutations of 4 letters can be made out of the fetters of the word examination F AUB Find the sum of all numbers greater than 10000 formed by using the digits 1, 3, 5, 7, 9, no digit being repeated in any number. _30, Find the sum of all numbers greater than 10000 formed by using the digits 0, 2, 4, 6, 8, no digit being repeated in any number. 31. If of p+q+r things p be alike, and q be alik id th different, shew that the total amber of Combinations ae a © rest (p +1) (+1) 2"-1. 32, Shew that the number of permutations which can be formed from 2n letters which are either a’s or ’s is greatest when the number of a’s is equal to the number of b’s. 33, Ifthe n+1 numbers a, 5, ¢, d,...... be all different, and each of them a prime number, prove that the number of different factors of the expression abed...... is (m+1) 2"—1,CHAPTER XII. MarTHematicaL INDUCTION. 158. Many important mathematical formule are not easily demonstrated by a direct mode of proof; in such cases we fre- quently find it convenient to employ a method of proof known as mathematical induction, which we shall now illustrate. Ezample1. Suppose it is required to prove that the sum of the cubes of the first n natural numbers is equal to tS 4 : ‘We can easily see by trial that the statement is true in simple cases, such as when n=1, or 2, or 8; and from this we might be led to conjecture that the formula was true in all cases. Assume that it is true when m terms are taken ; that is, suppose 2 184284 33+, abt} : Pa to n torms=| Add the (n +1)" term, that is, (n+1)* to each side; then a 194.274894+...... tonth terms = [> @+1)} +(n+1)3 =(n+1) (F+n+1) _ (n+)? (n9-44n+4) —__ s = jortintay, = a ; which is of the same form as the result we assumed to be true for n terms, n+1 taking the place of n; in other words, if the result is true when we take a certain number of terms, whatever that-number may be, it is true when we increase that number by one; but we see that it is true when 8 terms ere taken; therefore it is true when 4 terms are taken; it is therefore true when 5 terms are tsken; and soon. Thus the result is true universally.184° HIGHER ALGEBRA. Example 2, To determine the product of n binomial factors of the form ata, By actual multiplication we have (2+a) (c+b) (c+¢)=23 +(a+b+c) 2+ (ab+be+ca) x+abe; (2+ a) (+b) (w+c) (1+d)=24+ (a+b+c+4d) a +(ab+actad+be+bd+ced) x* + (abe + abd+ acd + bed) 2 +abed, In these results we observe that the following laws hold: 1, The riumber of terms on the right is one more than the number of binomial factors on the left. 2. The index of x in the first term is the same as the number of binomial factors; and in each of the other terms the index is one less than that of the preceding term. 8, The coefficient of the first term is unity; the coefficient of the second term is the sum of the letters a, }, c,......; the coefficient of the third term is the sum of the products of these letters taken two at a time; the coefficient of the fourth term is the sum of their products taken three at a time; and so on; the last term is the product of all the letters. Assume that these laws hold in the case of n-1 factors; that is, suppose (a+4) (2@+)...(e+h)= 2") + pe? + pyr 8 4 poe 44 tas where Py=atb+et+...h; Pa=ab+act+...Fah+bet bdt...... 3 Ps=abe+abd+ 3 Pa-=abe...h, Multiply both sides by another factor x +k; thus (v-+a) (2-40)... (24h) (e+) Sat (py +k) 214+ (pat pk) 8+ (Dy + Pak) 293+... + Pyake Now Dtk=(atbt+c+...+h)+k =sum of all the n letters a, b, ¢,...4; Pat Pk=pytk(at+b+...+h) =sum of the products taken two at a time of all the n letters a, b, c,... 3 Pat pak=pgtk(abt+act...t+ah+be+...) =sum of the products taken three at a time of all the n letters a, b, c,... &; Pn-1* =product of all the n lettera a, b, ¢, ... keMATHEMATICAL INDUCTION. 13 If therefore the laws hold when n-1 factors are multiplied togethe they hold in the case of n factors. But we have seen that they hold in tl case of 4 factors; therefore they hold for 5 factors; therefore also for factors; and so on; thus they hold universally. Therefore (z+a) (c+b) (e+e) ... (c+k) =a" +82") +4 Sy2"-9 + Syr*-F 4+ +S, where S,=the sum of all the n letters a, b,c... k; S,=the sum of the products taken two at a time of these n letters. S,=the product of all the 2 letters. 159. Theorems relating to divisibility may often be est: blished by induction, Example. Shew that 2*-1 is divisible by z-1 for all positive integr values of n. 7 a1 wa 1 By division gel + zc1i if therefore 2“! — 1 is divisible by z- 1, then 2*-1 is also divisible by z- But 2?-1 is divisible by x~1; therefore 2-1 is divisible by x-1; ther fore z‘— 1 is divisible by z - 1,and so on ; hence the proposition is establishe Other examples of the same kind ‘will be found in the chapter on tl Theory of Numbers, 160. From the foregoing examples it will be seen that th only theorems to which induction can be applied are tho: which admit of successive cases corresponding to the order « the natural numbers 1, 2, 3,...... nN EXAMPLES. XII. Prove by Induction : 1, 14+34+5+......4(n-1)=n% Q VERY tnt} n(nt1)(2n+)). 3 2424984,..,..499=2(2"—1), 1 1 1 a 4 pataetyate to m terms = 5 - 5. Prove by Induction that #*—y* is divisible by «+y when even.CHAPTER XIII. BinoMIAL THEOREM. PositivE INTEGRAL INDEX. 161. Ir may be shewn by actual multiplication that (2+ a) (a+b) (e+e) (e+), =a'+(atb+e+d) 2+ (abt+act+ad+be+bd+cd) x? + (abe + abd + acd + bed) 2+ abd oo... .eececeeeeteeeeeeeeee (1). We may, however, write down this result by inspection ; for the complete product consists of the sum of a number of partial pro- ducts each of which is formed by multiplying together four letters, one being taken from each of the four factors. If we examine the way in which the various partial products are formed, we see that (1) the term 2x* is formed by taking the letter x out of each of the factors. (2) the terms involving x* are formed by taking the letter x . out of any three factors, in every way possible, and one of the letters a, 6, c, d out of the remaining factor. (3) the terms involving «* are formed by taking the letter « out of any two factors, in every way possible, and two of the letters a, 6, c, d out of the remaining factors. (4) the terms involving # are formed by taking the letter x out of any one factor, and three of the letters a, 6, c, d out of the remaining factors. . (5) the term independent of « is the product of all the letters a, b, ¢, d. Example 1. (@- 2) (2 +3) (7-5) (2 +9) =att(— 243-549) 294 (-6+10-18- 15427 —45) 2? "4 (80-54 +90 — 185) +270 = 244 5258-47 a%- 6924210, ©BINOMIAL THEOREM, POSITIVE INTEGRAL INDEX. 137 Example 2. Find the coefficient of 2* in the product (%—8) (x +5) (w-1) (2+ 2) (x-8). The terms involving z* are formed by multiplying together the z in three of the factors, and two of the numerical quantities out of the iol: maining factors; hence the coefficient is equal to the sum of the products of the quantities ~3, 6, —-1, 2, —8 taken two at a time. Thus the required coefficient = -154+38-6+424-5+410-40-2+8-16 39. 162. If in equation (1) of the preceding article we suppose b=c=d=a, we obtain (c+)! =a' + 4ac? + 6a'a* + data + af, The method here exemplified’ of deducing a particular case from a more general result is one of frequent occurrence in Mathematics ; for it often happens that it is more easy to prove a general proposition than it is to prove a particular case of it. We shall in the next article employ the same method to prove a formula known as the Binomial Theorem, by which any binomial of the form %+a can be raised to any assigned positive integral power. 163. To find the expansion of (x+a)" when n is a positive integer. Consider the expression : (w+a) (+b) (@+0) 0... (a+), the number of factors being . The expansion of this expression is the continued product of the n factors, +a, 2+6, w+, ......0+, and every term in the expansion is of m dimensions, being a product formed by multi- plying together 7 letters, one taken from each of these factors. The highest power of 2 is 2*, and is formed by taking the letter # from each of the n factors. The terms involving 2*-' are formed by taking the letter « from any n—1 of the factors, and one of the letters a, b, c, ... & from the remaining factor; thus the coefficient of 2*~' in the final product is the sum of the letters a, b, ¢, ...... k; denote it by S,. The terms involving 2*~* are formed by taking the letter x from any n—2 of the factors, and two of the letters a, b, ¢, ...k from the two remaining factors; thus the coefficient of x * in the final product is the sum of the products of the \ettera a, b,c, ... & taken two at a time; denote it by S,.138 HIGHER ALGEBRA. And, generally, the terms involving «*~" are formed by taking the letter « from any n—r of the factors; and r of the letters a, b, c,...k from the r remaining factors; thus the coefficient of x"~" in the final product is the sum of the produgts of the letters a, b, ¢,...k taken x at a time; denote it by S. The last term in the product is abc... 4; denote it by S,. Hence (@+a)(e+ 2 (@+e) .... © +k) =a" + Sat" + Sot t+... 4 Soh +... +8 +8. In S, the number of terma is n; in S, the number of terma i is the same as the number of combinations of n things 2 at a time; _ that is, "C,; in S, the number of terms is *C,; and so on. Now suppose 8, ¢, ... k, each equal to a; then S, becomes "Ca; S, becomes *C,a*; S, becomes "C,a*; and so on; thus (2+ a)" =a + "CO aa + *Catat* + "Carat +... + "Cla"; . substituting for "C,, *C’ ‘+++ We obtain (e+a)"= "+ naa?" + art) ata" mn?) ah 4,40", the series containing +1 terms. This is the Binomial Theorem, and the expression on the right is said to be the expansion of (a + a)". 164. The Binomial Theorem may also be proved as follows : By induction we can find the product of the m factors ata, 2+b, w+c,...0+k as explained in Art. 158, Ex. 2; we can then deduce the expansion of («+ a)" as in Art. 163, 165. The coefficients in the expansion of vA +a)" are very conveniently expressed by the symbols "C,, "Cy, +. “"C,. We shall, however, sometimes further abbreviate ita by omitting n, and writing C,,¢,, C,, ... C,. With this notation we have (a +a)" = 0" + Cac" + C,a%a* + Carat? +... + 0,0". If we write-—a in the place of a, we obtain (a —a)"= ar" + C, (—a@) w"'4 C, (—a)*a**4+-C,(— a)?" +... +0, (-a)" = 2" — Cau" + O,a%a"* — Cau" +... + (-1)"C,a". Thus the terms in the expansion of («+a)" and (a— a)" are numerically the same, but in («—a)" they are alternately positive and negative, and the last term is positive or negative according 4s 2 is even or odd.BINOMIAL THEOREM. POSITIVE INTEGRAL INDEX. 139 Example 1. Find the expansion of (x+-y)*. By the formula, (e+ y)®=28 + 90; why + 9Cyxty? + 6C, a8y8 + °C,a2y6 + Cerys + *Cyy® = 25+ Gaby + 1bxty* + 2028y3 + 152%y* + bry® + y8, on calculating the values of °C, , °C,, °C5,.........+ Example 2. Find the expansion of (a - 2z)’. (a - 2)? =a? —7C, a8 (22) +70, a® (22)? —7C, at (2a)? + to 8 terms. Now remembering that "C,="C,_,, after calculating the coefficients up to 7C,, the rest may be written down at once; for 7C4="C; 7C,=7Cq; and 80 on. Hence (a-22)?=a7 - Tas (22) 47-8 $ a (22)°— ae a 5 gO Oz) + oe =a? ~ Ja (22) + 21a5 (22)*—85a* ‘oom 85a (22)4 —2la? (2x) + 7a (22)8 — (22)? =a’? ~ 14a%e + 84a5x* — 2800423 + 560a%xt — 672a%5 + 448ax6— 12827, Eszample 3. Find the value of (a+ Ja?—1)'+(a- J/a*=1)’. We have here the sum of two ‘expansions whose terms are numerical: y the same; but in the second expansion the second, fourth, sixth, and eight! terms are ’ negative, and therefore destroy the corresponding ‘terms of the first expansion. Hence the value =2 {al +2108 (a®— 1) 4+ 85a (a? — 1)?+7a (a- 1)3} = 2a (64a8 ~ 1120*+56a-7). 166. In the expansion of (x +4)", the coefficient of the second term is "C,; of the third term is "C,; of the fourth term is "C, ; and so on; the suffix in each term being one less than the number of the term to which it applies; hence *C’, is the co- efficient of the (r+1)" term. This is called the general term, because by giving to r different numerical values any of the coefficients may be found from “C,; and by giving to x and a their appropriate indices any assigned term may be obtained. Thus the (r+ 1) term may be written "Cgt'a, or n (n—1) (n- i «(n- rt) nye, . In applying this formula to any C atiouar case, it should be observed that the index of a is the same as the suffix of ©, oma Hast the mem of Sha indices 7x and a ie n,140 HIGHER ALGEBRA. Ezample 1, Find the fifth term of (a +22), The required term =, a! (2z3)4 _17.16.15.14 ~~ 1.2.8.4 =88080a) 212, Example 2, Find the fourteenth term of (3 - a)®. The required term =50,, (8)? (-a)'3 =15¢, x (—9al4) [Art, 145.] = - 945033, 167. The simplest form of the binomial theorem is the ex- pansion of (1+2)". This is obtained from the general formula of Art. 163, by writing 1 in the place of » and « in the place ofa, Thus (lta)=14"Ce+ "Cr+... + "Cait ..4+"O x" alae MD sy seeeee +2"; x 16a 238 the general term being n (n=1) (n= 2) (n-r+1),, The expansion of a binomial may always be made to depend upon the case in which the first term is unity; thus @e+yr={e (1+ xy =a%(1 +2)", where at Example 1, Find the coefficient of x'® in the expansion of (2 — 22)!°, We have (2*—22)!=2% (: -3)" and, since z® multiplies every term in the expansion of (1 - 2" we have in this expansion to seek the coefficient of the term which contains i. Hence the required coefficient ="C, ( - 2)4 _10.9.8.7 “7.2.8.4 = 8360. Jn some cases the following method is simpler. x16 —_-sBINOMIAL THEOREM. POSITIVE INTEGRAL INDEX. 141 Example 2. Find the coefficient of x” in the expansion of (#+ zy. Suppose that <* occurs in the (p +1)" term. The (p+1)" term =", (22)8-P Gy . ="Cp zp, But this term contains 2, and therefore 2n —5p=r, or p= =F . Thus the required coefficient="Cp="C,,_, “o- ln 5 (n+) - z (nn) is 8 positive integer there will be no term containing 2" in 2n-r Unless E the expansion. 168. In Art. 163 we deduced the expansion of (a +a)" from the product of factors (w+) (a+b)... (a+), and the method of proof there given is valuable in consequence of the wide gene- rality of the results obtained. But the following shorter proof of the Binomial Theorem should be noticed. It will be seen in Chap. xv. that a similar method is used to obtain the general term of the expansion of (a+b+et...... Ne 169. To prove the Binomial Theorem. The expansion of (x +)" is the product of m factors, each equal to x+a, and every term in the expansion is of m dimen- sions, being a product formed by multiplying together n letters, one taken from each of the n factors. Thus each term involving a;"~"a’ is obtained by taking a out of any r of the factors, and x out of the remaining n—r factors. Therefore the number of terms which involve x*-‘a" must be equal to the number of ways in which r things can be selected out of n; that is, the coefficient of a""'a’ is "C,, and by giving to r the values 0, 1, 2, 3, ... min succession we obtain the coefficients of all the terms, Hence (% +a)" =a" + "Coa + "Op" 8a? +... + "CO aa +... ta", since "C, and "C, are each equal to unity.142 HIGHER ALGEBRA, EXAMPLES. XIII a Expand the following binomials: 1, (#-3)% 2 (Br+2y)t, 3. (Qe-y). 4, (1-324), 5. (at +a). 6. (1—zy). 1. (0-Fy. 8 (20-5). 9, (+8). ww. (j2-2). ve (G+2)". 12, (-3)" Write down and simplify : 13. 15, The 12" term of (2¢—1)%. 16. The 28t term of (57+ 8y)™. 10 17. The 4 term of & + 98) . bys 18 The 5“ term of ( - 3) rth 5\? 19. The 7” term of ee 3 5B y 20. The 5“ term of (5 a t) : at Ba Find the value of QL. (@+a/2)8+ (a -./2)4 22, (a8 =a? +2) — (ox? a? 2), 23. (J24+1)®—(/2-1)% WA, (2—V1— a) +(24 J1—a, +25. Find the middle term of S45)" . Find the middle term of ( -3)" : " . 3a\ 27, Find the coefficient of 2'8 in (24% ‘Me 28. Find the coefficient of x!8 in (ax — bz), 29, Find the coefficients of 2°? and 2-1” in (#- 3) - 30. The 4% term of (7-5). 14, The 10" term of (1-27). Find the two middle terms of (sa -§) c ee ee ceBINOMIAL THEOREM. POSITIVE INTEGRAL INDEX. 148. . : » (3 1\9 31. Find the term independent of « in G ae Pa 5 \ eee 32, Find the 13% term of (97-,—-) . 3,/z, 33, If x" occurs in the expansion of (#43) ; find its coefficient. 3a 34, Find the term independent of x in (= 3) 5 Qn 35. If x? occurs in the expansion of (+3) » prove that its co- efficient is — : fs (n—p) |g (2n-+p) 170. In the expansion of (1 + x)* the coefficients of terms equi- distant fromthe beginning and end are equal. The coefficient of the (r +1) term from the beginning is "C,. The (r +1) term from the end has n+1—(r+1), or n-r terms before it; therefore counting from the beginning it is the (n—7r + 1) term, and its coefficient is “C__,, which has been shewn to be equal to "C,. [Art. 145.] Hence the proposition follows. 171. To find the greatest coefficient in the expansion of (1 +x) The coefficient of the general term of (1+2)" is"C_; and we have only to find for what value of r this is greatest. By Art. 154, when x is even, the greatest coefficient is "C,3 and when n is odd, it is*C,_,, or ™ a a+) these two coefficients being equal. ; 172. To find the greatest term in the expansion of (x + a)". We have (w+ a)" = 2" ( + 5) therefore, since x" multiplies every term in ( + “y, it will be sufficient to find the greatest term in this latter expansion.bl 144 HIGHER ALGEBRA. Let the r* and (r+1)*® be any two consecutive terms. The (r+1)" term is obtained by multiplying the r* term by n-rt+1l a : m+] a ret; that is, by (e- 1) 8. [Art, 166.] The factor pues decreases as 7 increases; hence the ie (r +1)" term is not always greater than the r term, but only until (2 - 1) ‘ becomes equal to 1, or less than 1. Now met 1) 251, r a n+1 a so long as F -1l>-; : n+1l_ a that is, ——>-+l1, r a or a a (1). % ~4+1 a n+1 ; ; f If oe be an integer, denote it by p; then if r=p the =+1 a multiplying factor becomes 1, and the (p + 1)" term is equal to the p; and these are greater than any other term. If peal be not an integer, denote its integral part by q; 241 a then the greatest value of r consistent with (1) is g; hence the (q+ 1)* term is the greatest. Since we are only concerned with the numerically greatest term, the investigation will be the same for (x—a)"; therefore in any numerical example it is unnecessary to consider the sign of the second term of the binomial. Also it will be found best o work each example independently of the general formula.BINOMIAL THEOREM. POSITIVE INTEGRAL INDEX. 145 Keample 1, If s=5, find the greatest term in the expansion of (1++4:)* Denote the r** and (r+-1)'* terms by T, and T,,, respectively; then 8-rt+1 Tra Ff 4exT, 9-r 4 Say xx tes henea Tra>T,, 9-r 4 so long as pe gehs that is 36 -4r>38r, or 86>7r. The greatest value of r consistent with this is 6; hence the greatest term is the sixth, and its value 5 5 ="0.x(5) =8C, x (3) =e Example 2. Find the greatest term in the expansion of (3—2z)° when z=1, (8-22)P=8° (0 -3) thus it will be sufficient to consider the expansion of (2 . $): Here Tyga : ex T,, numerically, = os x ax Ty3 hence Try>Tpy so long as nr 21, r 8 that is, 20>5r. Hence for all values of r up to 3, we have 7, > Tes but if r=4, then T,4,=T,, and these are the greatest terms. Thus the 4’ and: 5 terms are numerically equal and greater than any other term, and their value : = 89x °C, x (5) = 88 x 84 x 8= 489888,4 146 wy HIGHER ALGEBRA. wn. 173 ‘0 find the sum of the coefficients in the expansion of i+ xy. In the identity (1+2)*=1+C0+Cp'+CaP+... +02", put «=1; thus =14+0,4+0,+0,+..4+0¢, =sum of the coefficients. Cor. 0, +0, 4+0,+...4+0,=2"-1; that is “the total number of combinations of things” is 2"—1. [Art. 153,] 174. To prove that in the expansion of (1+ x)", the sum of the coefficients of the odd terms is equal to the sum of the coefficients of the even terms. In the identity (1+a)"=1+C,e+Cp°+Cp+...+ 02", put «=-1; thus 0=1-C,+0,-C,4+0,-O,4 0... 5 - 140,40, 4+ 0.0... =0,4+0,4+0,+ «00. =} (sum of all the coefficients) =o) S175. The Binomial Theorem may also be applied to expand expressions which contain more than two terms. Example, Find the expansion of (2*+22- 1), Regarding 2z — 1 as a single term, the expansion = (2) +B (22)? (2x — 1) +822 (2x - 1)?-+ (20-1)? = 28 + 625+ 924 — 423 — 9234 62—1, on reduction. 176. The following example is instructive. Example, If (1-+2)*=09+ 00+ Cy? + ......+0y 2%, find the value of y+ 2¢, + Bcq-+4ey+......+ (+1) Cy, and CP + 2e,9+ Bc2+...... +My? ... : The series (1) = (Cy +¢, + C+ ......-+0q)+ (Cy + 2Cq+ Bey+ ......+10,) =4n {i+-y+O-9G-9 , =O 4n[141)2 = 4n,2"-1,BINOMIAL THEOREM. POSITIVE INTEGRAL INDEX. 147 To find the value of the series (2), we proceed thus: C+ 2e,27 + Begz8+...... +N, 2 =n frree-n 24 ODE-9 ary ae +o} =nz (1+2)"-1; hence, by changing z into : , we have Cy | Beg , Bey nm, 71 1\"> atat gts + 3-2 (143) sevesseeneee (B)o Also Cote LHC +....., Cyt = (LEZ)... cece (4). If we multiply together the two series on the left-hand sides of (3) and (4), we see that in the product the term independent of 2 is the series (2); hence the series (2) =term independent of z in - S (lta) (143) is = term independent of x ink Sate) =coefficient of 2" in n(1 we =n x10, 2n-1 “(a-1[n-1" EXAMPLES. XIII. b. In the following expansions find which is the greatest term : 1, («-y)™ when c=11, y=4. 2. (Qu —3y)*8 when x=9, y=4. 3, (20+) when a= 4, b=5. 4, (8422)! when «= 5 ‘ In the following expansions find the value of the greatest term : 5. (1+2)" when a=3, n=6. 1 1 6. (a+2)" when a=5, %=35 n=9 W—2ee 148 HIGHER ALGEBRA. 7. Shew that the coefficient of the middle term of (1+)™ is equal to the sum of the coefficients of the two middle terms of (1+ 2)™-1 8 If A be the sum of the odd terms and B the sum of the even terms in the expansion of (x+«a)", prove that A?— B!=(2%— at)", 9, The 2", 3", 4 terms in the expansion of (#+y)" are 240, 720, 1080 respectively ; ; find X,Y, 10, Find the expansion of m +22 — 2), 11, Find the expansion of (327—2ax+ 3a*)3. 12, Find the r* term from the end in (7+a)*. an 13, Find the (p-+2)" term from the end in (#-3) o In the expansion of (1+) the coefficients of the (2r+1)" and the eae terms are equal; find r. ¢ ) 15. Find the relation between r and n in order that the coefficients of the 3r* and (r+ 2)" terms of (1+.7)" may be equal. 16. Shew that the middle term in the expansion of (1+.7)™ is 1.3.5... (2n—1) —a ana", If oy a Cg ++ Cy denote the coefficients in the expansion of (1+.7)", prove LT. C+ 26+ 30g + oes tM = NM. 2P-2, 41% ey _ OAtt—1 Bt ht tnt A=. 19, 4424, 34 ly _n(n+l) | mu cae 20. (ote) (+e). qn ty) = Sea I 2c, , We. | Be, antIo, Bntl_1 2. 2,424 sat 2 . mn ote tg tat. tar n+l 2. 3. 2. 2 = 22. ok + oF +074... +6.) Tae |2n ZB. OCA Cr Aap gt creeee 40 — y=CHAPTER XIV. BinommaL THEOREM, ANY INDEX. 177. In the last chapter we investigated the Binomial Theorem when the index was any positive integer; we shall now consider whether the formule there obtained hold in the case of negative and fractional values of the index. Since, by Art. 167, every binomial may be reduced to one common type, it will be sufficient to confine our attention to binomials of the form (1 +x)". By actual evolution, we have ; Vize le! 1 (l+2)?= T+e=lt5a—gattiae meee ees 5 and by actual division, (laa) t= qigyal + tes Sate dats eee ; [Compare Ex. 1, Art. 60.] and in each of these series the number of terms is unlimited. In these cases we have by independent processes obtained an 1 expansion for each of the expressions (1+) and (1+a)-". We shall presently prove that they are only particular cases of the general formula for the expansion of (1+.2)", where n is any rational quantity. This formula was discovered by Newton. 178. Suppose we have two expressions arranged in ascending powers of 2, such as m(m—1) Ll+me+ T.3 at a oe (1), n(n-1),, »(m—-1) (n—2) and Line ++ “Teg pence (2).150 , HIGHER ALGEBRA. The product of these two expressions will be a series in as- cending powers of x; denote it by 1+ Aa +B? + Cx? + Dat+...... ; then it is clear that A, B, C,...... are functions of m and n, and therefore the actual values of A, B, C, ...... in any particular case will depend upon the values of m and x in that case. But the way in which the coefficients of the powers of x in (1) and (2) combine to give A, B, C,...... is quite independent of m and x ; in other words, whatever values m and n may have, A, B, C, _ preserve the same invariable form. If therefore we can determine the form of A, B, C,...... for any value of m and , we conclude that A, B, C,...... will have the same form for all values of m and n. The principle here explained is often referred to as an example of “the permanence of equivalent forms ;” in the present case we have only to recognise the fact that in any algebraical product the form of the result will be the same whether the quantities in- volved are whole numbers, or fractions ; positive, or negative. We shall make use of this principle in the general proof of the Binomial Theorem for any index. The proof which we give is due to Euler. 179. To prove the Binomial Theorem when the index is a positive fraction. Whatever be the value of m, positive or negative, integral or JSractional, let the symbol f(m) stand for the series m (m — 1) m (m— 1) (m—2) . Ta Pt gg ts L4ma+ then /(n) will stand for the series n(n—1) n(n—-1)(n-2) Lena Sy att gg att vee If we multiply these two series together the product will be another series in ascending powers of x, whose coefficients will be unaltered in form whatever m and n may be. To determine this invariable form of the product we may give to m and m any values that are most convenient; for this purpose suppose that m and x are positive integers. In this case /(m) is the expanded form of (1 +)", and f(n) is the expanded form of “1 +z)"; and thereforeBINOMIAL THEOREM. ANY INDEX. 151 F(m) xf (n) = (1 +2)" x (L+ a) =(L+a)"", but when m and x are positive integers the expansion of (1 + x)"** is 1+(m+n)a+ (m+) (mtn) Be : This then is the form of the product of f(m)x/(n) in all cases, whatever the values of m and m may be; and in agreement with our previous notation it may be denoted by /(m +n) ; there- fore for all values of m and n SF (m) xf (n) =f (m+n). Also S(m) xf (m) x F(p) =f (m +n) xf (P) =f(m+n+p), similarly. Proceeding in this way we may shew that , JS (m) xf (n) xf (p)..-to k factors =/(m +n +p +.,.to k terms). Let each of these quantities m, x, p, ...... be equal to where hf and & are positive integers ; {r()=70: but since A is a positive integer, f(t) = (1 +2); ava {0 a+afasQ)s but f (i) stands for the series hh 1+ he + eG) We. ; Ez 1) . ae (+ajfe1 star thts sees > which proves the Binomial Theorem for any positive fractional index.152 HIGHER ALGEBRA. 180. To prove the Binomial Theorem when the index is any It has been proved that S(m) xf (2) =f (m+n) for all values of m and n, Replacing m by —n (where n is positive), we have F(-0) xf (2) =f(-n +m) =f(0) eis since all terms of the series except the first vanish ; 1 oe rice (-"); but f(n) =(1 +2)*, for any positive value of n; i G+ay =f) or (l+a)*=f(-n). But /(—2) stands for the series Le (enyer CRD, pestees 5 (142) = 14cm) ee COD gry esses ; which proves the Binomial Theorem for any negative index. Hence the theorem is completely established. 181. The proof contained in the two preceding articles may not appear wholly satisfactory, and will probably present some dif- ficulties to the student. There is only one point to which we shall now refer. In the expression for f(m) the number of terms is finite when % is a positive integer, and unlimited in all other cases. See * ", 182, It is therefore necessary to enquire in what sense weBINOMIAL THEOREM. ANY INDEX. 153 are to regard the statement that f(m) x f(n)=/(m+n). It will be seen in Chapter xx1., that when «<1, each of the series /(m), ST (n), f (m+n) is convergent, and f(m +n) is the true arithmetical equivalent of f(m) x/(n). But when #>1, all these series are dwergent, and we can only assert that if we multiply the series denoted by /(m) by the series denoted by /(n), the first + terms of the product will agree with the first r terms of f(m+n), whatever finite value r may have. [Nee Art. 308.] 3 Example 1, Expand (1-2)? to four terms. 3/8 3/3 8 : 8 (84 8(8_4\ (39 ooahaay8e-—) EOD ae 1G) G4), a 3.38 1 slogrtgetiget peeed Example 2. Expand (2+8z)~ to four terms. (2-482)-4= 9-4 (2 +5) A140 ) 2559 GG) MGPLEM EY] 1 45, 185 = 5g (1-884 GF t-Fos see ). 182, In finding the general term we must now use the formula n(n-1)(n—2)...... (n-r+1)_, a written in full; for the symbol "C, can no longer be employed when 7 is fractional or negative. Also the coefficient of the general term can never vanish unless one of the factors of its numerator is zero; the series will there- fore stop at the r* term, when n—r+1 is zero; that is, when r=n-+1; but since r is a positive integer this equality can never hold except when the index 7 is positive and integral. Thus the expansion by the Binomial Theorem extends to n+1 terms when nis a positive integer, and to an infinite number of terms in wl | other cases.154 HIGHER ALGEBRA. 1 Ezample 1, Find the general term in the expansion of (1+2)?. The (r+1)" term= = U-1)(-8)(-5) oy alr The number of factors in the numerator is r, and r—1 of these are nega- tive; therefore, by taking - 1 out of each of these negative factors, we may write the above expression a The (r+1)" term = (=nzyr rT 1(t-n) (1-9n)......(l-7-1+m), a, we (ap ES Borah 2 (0—1) (@n=1)......=Ten-V) =(-1)"(-1) __ (w-1) (Qn-1).....: (F=1-n-1) , ea since (-1)'(-1) 1 =(-1)741=-1, Example 3. Find the general term in the expansion of (1 — <)-’, The (r+1)!* term = (= 9(= 4-5) 5 «2, (4) (749), 1.2 ’ ” Temoving like factors from the numerator and denominator.BINOMIAL THEOREM. ANY INDEX. 155 EXAMPLES. XIV. a. Expand to 4 terms the following expressions: L (4a) 2. (+a. 3 a-a. 4 (1+2%)-% 5. (1-323. 6. (1-308 3 1 (1420), 8. (148). a (143). 1 1)" MW. (2+2)-3 12 (9422)! 0. (1452) : . (2+2)-8, (9422), 13. (8+12a). 14. (9-62) 3, 15. (40-82) 2, Write down and simplify: -t 16. The 8% term of (14+2z) *. n 17. The 11" term of (1 — 224)?, 18 18, The 10% term of (14 32%). 19, The 5" term of (3a — 26)-1. 20. The (r+1)" term of (1—2)~%, 21, The (r+1)" term of (1—2)-4. 1 22. The (r +1)" term of (1+) n 23, The (r+1)" term of (1+.2)5. 3 24, The 14” term of (210— 2%x)3, n 25. The 7 term of (3°+ 64z)4. 183. If we expand (1—2)"* by the Binomial Theorem, we obtain : (1 -—a)~*= 14 24 3a*%+ 4aP+ o.. ; but, by referring to Art. 60, we see that this result is only true when « is less than 1. This leads us to enquire whether we are always justified in assuming the truth of the statement (taya1+ noe ON ry bese ; i156 HIGHER ALGEBRA. and, if not, under what conditions the expansion of (1 +2)" may be used as its true equivalent. Suppose, for instance, that n=—1; then we have (l-a)sltaetat+ tats (1); in this equation put a =2; we then obtain (- 1TH 1424284 2424 This contradictory result is sufficient to shew that we cannot take n(n-1) , l+net Le + as the true arithmetical equivalent of (1 +<)" in all cases. Now from the formula for the sum of a geometrical pro- gression, we know that the sum of the first r terms of the a l-2# series (1) “Ile = 1 a . “Ta” 1-2’ and, when 2 is numerically less than 1, by taking r sufficiently large we can make tC 2 small as we please ; that is, by taking a sufficient number of terms the sum can be made to differ as little as we please from ea But when a is numerically greater than 1, the value of a increases with r, and therefore 1 no such approximation to the value of % =a is obtained by taking any number of terms of the series Lemtartet i. It will be seen in the chapter on Oonvergency and Diver- gency of Series that the expansion by the Binomial Theorem of (1+«)" in ascending powers of 2 is always arithmetically in- telligible when a is less than 1. But if « is greater than 1, -then since the general term of the series n(n-1) L+nat Ty et eeeeesBINOMIAL THEOREM. ANY INDEX. 157 contains a’, it can be made greater than any finite quantity by taking r sufficiently large; in which case there is no limit to the value of the above series ; and therefore the expansion of (1 +2)" as an infinite series in ascending powers of @ has no meaning arithmetically intelligible when is greater than 1. 184. We may remark that we can always expand (x +y) by the Binomial Theorem ; for we may write the expression in either of the two following forms : wv x. w(i+ty, x (145); and we obtain the expansion from the first or second of these according as « is greater or less than y. 185. To jind in its simplest form the general term in the expansion of (1-x)"*. The (r+ 1) term a) 2-1I)Cen-2)...(-n-r+]) (ay ir = iy Ret (ot 8) (wer) (-1)ye n (w+ 1) (n+ 2)... (n+r—1) e fr = a wv. =" From this it appears that every term in the expansion of (1 —2)~* is positive. Although the general term in the expansion of any binomial may always be found as explained in Art. 182, it will be found more expeditious in practice to use the above form of the general term in. all cases where the index is negative, retaining the form n(n- en = (n-r+1) x Ir only in the case of positive indices,158 HIGHER ALGEBRA. + . 1 Ezample. Mind the general frm ine eeanion ot i= wt a)", ai The (r +1)" term ey If the given expression had been (1+ 82) * we should have used the same formula for the general term, replacing 82 by - 82. 186. The following expansions should be remembered : (Lea) tsltaetattart ct tee (1 - a)" = 1+ 204 Baht dot + (rt lat oe (1a) 1 + B24 Gat + 1008 +... + CPD) or 187. The general investigation of the greatest term in the expansion of (1+.)", when m is unrestricted in value, will be found in Art. 189; but the student will have no difficulty in applying to any numerical example the method explained in rt, 172 Zeample. Find the greatest term in the expansion of (1+2)-* when ant , and n=20. We have Tri= pT ex T,, numerically, 80 long as that is, . 88>r. Hence for all values of r up to 87, we have T,,,>T,; but if r=88, then 74, =TZ,, and these are the greatest terms. Thus the 38 and 89" terms , Vagual t numerically and greater than any other term.BINOMIAL THEOREM. ANY INDEX. 159 188. Some useful applications of the Binomial Theorem are explained in the following examples. Example 1. Find the first three terms in the expansion of 1 a (1+82)3 (1-22)73, Expanding the two binomials as far as the term containing x?, we have (eet) (: +3etget.-) 8.8 2 9 =l+a a*g) +°(5+ a3 -3) een 138 55 al+Gatige If in this Example «=-002, so that x= -000004, we see that the third term is a decimal fraction beginning with 5 ciphers. If therefore we were required to find the numerical value of the given expression correct to 5 places of decimals it would be sufficient to substitute -002 for z in 1 +3 2, neglect- ing the term involving 2°. Example 2, When z is so small that its square and higher powers may be noglested, find the value of -s 1422 +/4 420 (4+2)% Since 2? and the higher powers may be neglected, it will be sufficient to retain the first two terms in the expansion of each binomial. Therefore 6 1422) 42 143) the expression = the term involving 2° being neglected,160 HIGHER ALGEBRA. Eample 8, Find the value of | 7 to four places of decimals. a 1 ante 1 a\2 Jaren *=(P-2) F751 1 1,381,5 =a(ltptg gta ept ) 1,1,81,51 aatatg pty at To obtain the values of the several terms we proceed as follows 7)1 1 7) T2867... esses fone Ts 7) 020408 1 7) 002915... eeeeeeteees =a 7) 000416 1 00059... sesseerseessee = iBS and we can see that the term : 4 is a decimal fraction beginning with 5 ciphers. — 291! “Tat +142857 + -002916 + 000088 ='14586, and this result is correct to at least four places of decimals, Example 4, Find the cube root of 126 to 5 places of decimals. 1 1 (126)8= (58-+-1)# “(3 2 =54l 12 1 97 =5+-018333 ... — 000085 wehe. = 501329, to five places of decimals,BINOMIAL THEOREM. ANY INDEX, - 161 EXAMPLES, XIV. b. Find the (r+ 1)" term in each of the following expansions: 1 (+2), 2 (1-2)-5 3. (1422). 4, (ay, 5. (1+2%)-3, 6. (-20)3, 7. (a+bx)-), 8, (2-2)-2. 9 Wea. 1 1 10, re 1. wa Find the greatest term in each of the following expansions : 4 7 =— 13, (1+.)-? when x: is¢ a 14. (1+2)? when 2= 91 bo ay 1 15, (1—‘7x) * when w=5. 16, (2¢+5y)* when 2=8 and y=3. 17, (5-4a)-7 when aay. 18, (322+ 4y3)-" when #=9, y=2, n=15. Find to five places of decimals the value of 19, 98. 20, 4/998. 21, 1003. 22, «2400. 1 1 3 =: 1 3 25. (630) 4. 26. 8 23. Fig 24. (140) (630) 3128. If 2 be so small that its square and higher powers may be neglected, find the value of an. (L—Ta)8(14 22), 28, Jie. (2-9) 2 2 \-5 1 (8+32) mo ( 1+3*) x (4432)? (24+ 32) V4—52 uaa x, HA, VA162 HIGHER ALGEBRA. a 32. N6482—Vi-a 2 a\? (1+52)5+ 442) EH 33, Prove that the coefficient of 2” in the expansion of (1- 4.) ? i, 1& an -«@ n(n+l) /1-2\? +2 1.20 \W4+a) 0°" . 3A, Prove that (tepase ft o* + 35, Find the first three terms in the expansion of eee eeeeee (+a) V1+4e- 36, Find the first three terms in the expansion of 3 (Lt 0+ VI+ 62 | (1-2)? 37. Shew that the x‘ coefficient in the expansion of (1—.2)-* is double of the (n— 1), 189, To find the numerically greatest term in the expansion of (1+ x), for any rational value of n. Since we are only concerned with the numerical value of the greatest term, we shall consider « throughout as positive. Case I. Let m be a positive integer. The (r+1)" term is obtained by multiplying the 7** term n-r+1 ; n+l by .%; that is, by CE -1) 2; and therefore the terms continue to increase so long as (2 - 1) z>1; fe that is, (+ De >lt+a, or ae liaBINOMIAL THEOREM. ‘ANY INDEX. 163 It (n+1)z l+a multiplying factor is 1, and the (p+ 1)" term is equal to the p™, and these are greater than any other term. be an integer, denote it by p; then if r=p, the (n+1)a - lea then the greatest value of r is g, and the (q +1)" term is the greatest. be not an integer, denote its integral part by q; Case II. Let m be a positive fraction. As before, the (r +1) term is obtained by multiplying the > term by CH 1)e. (1) If @ be greater than unity, by increasing r the above multiplier can be made as near as we please to — a; so that after a certain term each term is nearly x times the preceding term numerically, and thus the terms increase continually, and there is no greatest term. (2) If @ be less than unity we see that the multiplying factor continues positive, and decreases until r>n+1, and from this point it becomes negative but always remains less than 1 numerically ; therefore there will be a greatest term. As before, the multiplying factor will be greater than 1 (n+ 1) so long as Tee" . Tf +) = be an integer, denote it by p; then, as in Case L, the (p +1) term is equal to the p", and these are greater than any other term. If rte be not an integer, let ¢ be its integral part ; then the (q + 1)" term is the greatest. Case III Let be negative. Let n=-m, so that m is positive; then the numerical value of the multiplying factor is eee . 2; that is (*2++ 1) r Ww SD164° HIGHER ALGEBRA, (1) If be greater than unity we may shew, as in Case II., that there is no greatest term. (2) If x be less than unity, the multiplying factor will be greater than 1, so long as (24 +1)e>1; r that is, (m- He >1l-2, (m1) lj or Ice (m-1)a Tf cz be a positive integer, denote it by p; then the (p+ 1)* term is equal to the p™ term, and these are greater than any other term. it Ne De be positive but not an integer, let g be its inte- gral oy ‘then the (7 + 1)" term is the greatest. f m—e be negative, then m is less than unity; and by writing the multiplying factor in the form ( -3")s we see that it is always less than 1: hence each term is less than the preceding, and consequently the first term is the greatest. 190. To find the number of homogeneous products of r dimen- sions that can be formed out of the n letters a,b, c, ...... and their powers. By division, or by the Binomial Theorem, we have =—_ =ltaxtara+arr+...... ; l—ax 1 2, 8, Toby 7 1 + be + Dat + Bia? + eee , toggcltent oar + ar +.. ————BINOMIAL THEOREM. ANY INDEX, 165 Hence, by multiplication, 1 1 4 Trae * T=be* Tee = (1 +a + a%e* +...) (1+ bat batt...) (ltcatctets...)... Hlta(atbtet...)+a% (a t+ab+actb?+bete%+...) +... =14+8e+827 + Se +...... suppose ; where S,, S,, S,, ...... are the swms of the homogeneous pro- ducts of one, two, three, ...... dimensions that can be formed of A, by Cy veers and their powers. To obtain the number of these products, put a, }, ¢, ...... each equal to 1; each term in §,, S,, S,,...... now becomes 1, and the values of S,, S,, S,,....+6 so obtained give the number of the homogeneous products of one, two, three, ...... dimensions. 1 1 1 Also Toa T=te Toa 1 =. becomes a-ay or (l—2)™. Hence S, =coefficient of x” in the expansion of (1 —2)"* _ (nt 1) (m+2) (n+7r-1) = nn jetr-1 “pt 191. Zo find the number of terms in the expansion of any multinomial when the index is a positive integer. In the expansion of (G+ 4, +0,+ 0.00 +a,)", every term is of 7 dimensions ; therefore the number of terms is the same as the number of homogeneous products of n dimensions that can be formed out of the r quantities a,, yy 6 Oy and their powers ; and therefore by the preceding article 1s equal to r+ n—1166 HIGHER ALGEBRA, 192. From the result of Art. 190 we may deduce a theorem relating’ to the number of combinations of n things, Consider 7 letters a, , ¢, d,...... 3 then if we were to write down all the homogeneous products of x dimensions which.can be formed of these letters and their powers, every such product would represent one of the combinations, 7 at a time, of the n letters, when any one of the letters might occur once, twice, thrice, ... up to ¢ times, Therefore the number of combinations of n things r at a time when repetitions are allowed is equal to the number of homo- geneous products of r dimensions which can be formed out of x IRAPaT o aenig, Feat That is, the number of combinations of » things r at a time when repetitions are allowed is equal to the number of com- binations of n+r—1 things r at a time when repetitions are excluded. letters, and therefore equal to 193. We shall conclude this chapter with a few miscel- Janeous examples. (1-22 (ra) The expression = (1 - 42+ 42%) (1+pyr+pyt"+...+p,2"+...) suppose, Example 1, Find the coefficient of x” in the expansion of The coefficient of z* will be obtained by multiplying p,, p,-1, Pps by 1, — 4, 4 respectively, and adding the results; hence the required coefficient =p, - 4p,_,+4p,_.. But Pr=(-1) (tay r+?) : [Ex. 8, Art. 182.] Hence the required coefficient = (=p EEF (ay FD yg (ay cn [r+ 1) (+2) + 49 (+1) +4r (r-1)] a& 7 (972 4+-87+2).BINOMIAL THEOREM, ANY INDEX. 167 Example 2. Find the value of the series 5 | 6.7 5.7.9 [tag tet ot 3.5 8.5.7 1,8.5.7.9 1, . 1 Tho expression =347 3° «gt Tg + gt Te B 85 8 7 2°2m 3 2° tat 8 8 5 132,29 =lty-3+ z Ezample 8, If nis any positive integer, shew that the integral part of (8 +,/7)" is an odd number, Suppose I to denote the integral and f the fractional part of (3 +,/7)". Then 14+f=8*4C,8"-1,/7+ 0,8"-9, 74+.058°-3 (V/7)3 + (1). Now 38-,/7 is positive and less than 1, therefore (3-,/7)" is a proper fraction; denote it by f’; ate f= B% — C891 IT + CyB. T+ Cg B38 TBF cesceseeeseesee(Q)e Add together (1) and (2); the irrational terms disappear, and we have I+f+f'=2 (8"+C,3"?.7+...) =an even integer. But since f and f’ are proper fractions their sum must be 1; .. I=an odd integer. EXAMPLES. XIV. c. Find the coefficient of 1, 2! in the expansion of Dey . Q=2)" f F 4+42a-a? 2 era 2. a’? in the expansion of W+ays 3. «in the expansion of ee .168 HIGHER ALGEBRA, 4, Find the coefficient of 2* in the expansion of ete. 5. Prove that pb 1,18 1 1.8.5 1 +227 of. 2°2° 94°28” 274.6° 93° 974.6.8° 3 6. Prove that V8= 149+ pe Soe econ 7. Prove that 1452 4 8 Gnt8) Qn (2n+2) (2n+4) | 3.6 3.6.9 0°" _ n(n+1) n(m+1)m+2) | nor fig 3.6 +—37¢.9 + $9 tt . 8. Prove that _ nm n(m—1) , a(n-1)(n-2) mfg a geo aas } fa nm n(n+1) , n(n+1)(n+2) =4 {i454 20+ +o ae tee 9. Prove that approximately, when = is very small, 10. Shew that the integral part of (5+2./6)" is odd, if » be a positive integer. 11. Shew that the integral part of (8+3./7)" is odd, if n be a positive integer. 12. Find the coefficient of 2” in the expansion of (1 - 20+ 32? - 4a°4+...... ines : 4 13, Shew that the middle term of (« + *) = equal to the coefficient r of 2 in the expansion of (1 — any 4), 14. Prove that the expansion of (1—.°)" may be put into the form om 6 — 2) Sms (1 — a)n-94 32 OH=3) gy — aymem egBINOMIAL THEOREM. ANY INDEX. . 169 15. Prove that the coefficient of 2* in the expansion itera is 1, 0, —1 according as x is of the form 3m, 3m—1, or 3m+1. 16. In the expansion of (2++c)8 find (1) the number of terms, (2) the sum of the coefficients of the terms, 17. Prove that if m be an even integer, 14,2 41 1 _gr-t Tpat* Beas * beet Pa 18. If cy, cy, Cg »..... , are the coefficients in the expansion of (1+.)", when x 1s a positive integer, prove that me (1) y+ 0g 03 Hove + (= 1)%e,=( 1)" ———— Flore (2) Cy — 2c, + Bey — 465+ ...0..+(—1)*(0-+1)eq=0. (8) ot —e,2 +e2—-c,2+...... +(-1)"¢,2=0, or (— 1)3c,, according as 7 is odd or even. ; 19, If s, denote the sum of the first » natural numbers, prove that (1) (1-2) ~3=8, + sg7-+857? +...... +8y0*-14,., (2) 2(8,8an +82 Sen—1 + 20. If ¢,.= pee ae ie Gn 1) | prove that 1 (1) Gant FL1FanF Gallon — at v0 FGu—1Gn tat Indus 1= 5 * (2) 2 Gan — Q1Gon—1 + Gada at oooeee F(— I) Ga 19u4 1} = gat(— 1)" 190? 21, Find the sum of the products, two at a time, of the coefficients in the expansion of (1+.)", when 7 is a positive integer. 22, If(7+4./3)"=p+B, where n and pare positive integers, and 8 a proper fraction, shew That (1-8)(p+8)=1. 23. If cq, Cy, Cg +++. 0, are the coefficients in the expansion of (1+.)", where nis a positive integer, shew that Cy , ©; —1)*~¢, 1,1 1 an gt gree +s f =.CHAPTER XV. MuLTINOoMIAL THEOREM. 194. We have already seen in Art. 175, how we may apply the Binomial Theorem to obtain the expansion of a multi- nomial expression. In the present chapter our object is not so much to obtain the complete expansion of a multinomial as to find the coefficient of any assigned term. Example. Find the coefficient of a‘b?cd° in the expansion of (atb+e+d)*, : The expansion is the product of 14 factors each equal to a+b+c+d, and every term in the expansion is of 14 dimensions, being a product formed by taking one letter out of each of these factors. Thus to form the term a‘b*c8d°, we take a out of any four of the fourteen factors, b out of any two of the re- maining ten, c out of any three of the remaining eight. But the number of ways in which this can be done is clearly equal to the number of ways of ar- ranging 14 letters when four of them must be a, two 6, three c, and five d; that is, equal to iu e 4213/5" This is therefore the number of times in which the term a‘b%c*d° appears in the final product, and consequently the coefficient required is 2522520. 195. To find the coefficient of any assigned term in the ex- pansion of (a+b+c+d+ ...)?, where p is a positive integer. The expansion is the product of p factors each equal to a+b+c+d+..., and every term in the expansion is formed by taking one letter out of each of these p factors; and therefore the number of ways in which any term a*bcrd* ... will appear in the final product is equal-to the number of ways of arranging p letters when a of them must be a, 8 must be 6, y must be c; and soon, That is, [Art. 151.] the coefficient of a*bferd® ... is eee vhere a+B+ytd+ =p.MULTINOMIAL THEOREM. 171 Cor, In the expansion of (a+ bx + cat + dat...) the term involving a*bfcrd? ... is __ iP Eee or (2 — arbicrd® .., aBt2y+8+... , eB i... where a+B+y+8+...=p. a (ba (cat) (cat. This may be called the general term of the expansion. Example. Find the coefficient of 2° in the expansion of (a + bz + cz*)’. The general term of the a is \9 at Yah th | een" Perel (Qs where a+8+y=9. We have to obtain by trial all the positive integral values of 6 and y which satisfy the equation 8+2y=5; the values of a can then be found from the equation a+p+y=9. Putting y=2, we have B=1, and a=6; putting y=1, we have B=3, and a=5; putting y=0, we have 8=5, anda=4, The required coefficient will be the sum of the corresponding values of ‘the expression (1). Therefore the coefficient required 9 nner, 2 ow = jen + Bis ihe = 252a%be? + 504a5)%e + 126a4b5. 196. Zo find the general term in the expansion of (a+ bx + ox? + dx? + ...), abe + where n is any rational quantity. By the Binomial Theorem, the general term is n(n-1) at ss (np +) at"? (be + ox? + de® +...)", where p is a positive. integer.172 HIGHER ALGEBRA. And, by Art. 195, the general term of the expansion of (ba + car + da? +... p_ Berd? ... aBt2yt+38+... , cara where B, y, 8 ... are positive integers whose sum is p. Hence the general term in the expansion of the given ex- pression is n(n—1)(n—-2) ... (n—p+1) e EDe-- where B+yt+8+... SP OBords .., gh tty+38+..., Pp. 197. Since (a+ bx + cx*+da*+ ..)" may be written in the form a( nar =) a a a it will be sufficient to consider the case in which the first term of the multinomial is unity. Thus the general term of (1 + ba + ca? +dx’+ ...)" ; n(n—~1) (n—2)... (n—p+1) 5, dd... tty +384... is Bye ovd® ... aB+2y+: : where B+yt8+...=p. Example. Find the coefficient of z* in the expansion of a (1-82 - 20? + 6x8, ‘The general term is 2/2 5(3-1 We have to obtain by trial all the positive integral values of 8, y, 5 which satisfy the equation 8+27+35=3; and then p is found from the equation ~=8+y7+48. The required coefficient will be the sum of the corresponding ‘tes of the expression (1). ptt) (-3)° (2) @P PHY (a), ely |eMULTINOMIAL THEOREM. 173 In finding §, +, 8, ... it will be best to commence by giving to 3 successive integral values beginning with the greatest admissible. In the present case the values are found to be s=1, y=0, B=0, p=1; 3=0, y=1, B=1, p=2; 3=0, y=0, B=3, p=8 Substituting these values in (1) the required coefficient 2 1 4 -() 0+(8) (Sean Mw hel 198. Sometimes it is more expeditious to use the Binomial Theorem. Example. Find the coefficient of 2‘ in the expansion of (1 - 27+ 82*)-3, The required coefficient is found by picking out the coefficient of 2 from the first few terms of the expansion of (1-2z-8z%)-* by the Binomial Theorem; that is, from 148 (2x — 82%) +6 (22 — 82%)? + 10 (22 — Bx?)8 + 15 (2 - 82?)4; we stop at this term for all the other terms involve powers of x higher than 24. The required coefficient=6 . 9 +10. 8 (2) (-8) +15 (2)! = - 66, EXAMPLES. XV. Find the coefficient of 1. a%b%ctd in the expansion of (a-b-c+d)", 2. a®b'd in the expansion of (a+b-—c-d). 3. a%b%c in the expansion of (2¢+6+3c)’. 4, .«x%A in the expansion of (ax — by +cz)°. 5, 23 in the expansion of (1432 - 22%), 6. x in the expansion of (1+2r+ 32%)!” 7. 2 in the expansion of (1 +27 -.2*)6, 8, 28 in the expansion of (1 - 27+3x4— 425), a174 HIGHER ALGEBRA. Find the coefficient of 9, «in the expansion of (1-204 32%— xt —25)8, 10, 2 in the expansion of (1 -20-4328)3, 11, <3 in the expansion of (1—-27+322- 429)}, 12, x8 in the expansion of ( aL + 5) 7 13, in the expansion of (2 -47+3z2)-2, 3 14. 2 in the expansion of (1 +422+ 102442028) 4, 15, !* in the expansion of (3 — 1525+ 182*)-1, 1 16, Expand (1—22r—2z°)4 as far as 2. 2 17. Expand (14-322—623) 3 as far as 25. ‘ 18, Expand (8— 925+ 1824) as far as 28 WO Tf 1+ 2+28+ 00... +2? Pay taytt age? +... p0"?, prove that (1) ag+ay+agt...... +Onp=(p+1)%. (2) a, +2a,+3a,+ ND Og =h np (p+) 20. If ao, a,, a), a3... are the coefficients in order of the expansion of (1+2+2°)", prove that Og? — Oy? + 4 — s+ oe (= 1tah_.= hay {1-(-1)*4,}. 21, If the expansion of (1+2+.2°)" be At ayetage?+ ... $0,207 +... +g t™, shew that Ay +A, +O, +... =A +A, +z + ... =Agtag+agt .., =3"1,CHAPTER XVI. LoGaRITHMS. 199, Derrvrrion. The logarithm of any number to a given base is the index of the power to which the base must be raised in order to equal the given number. Thus if a*=J, « is called the logarithm of to the base a. Examples. (1) Since 3‘=81, the logarithm of 81 to base 8 is 4. (2) Since 10'?=10, 10°?=100, 10°=1009....... the natural numbers 1, 2, 3,... are respectively the logarithms of 10, 100, 1000,...... to base 10, 200. The logarithm of WV to base a is usually written log, J, so that the same meaning is expressed by the two equations a’=N; w=log,N. From these equations we deduce N= gt an identity which is sometimes useful. Example. Find the logarithm of 82 {/4 to base 2,/2. Let x be the required logarithm; then, by definition, (2,/2)%=32,0/4; 2 1 a. (2. 29) = 28, 25; Boats, hence, by equating the indices,176 HIGHER ALGEBRA, 201. When it is understood that a particular system of logarithms is in use, the suffix denoting the base is omitted. Thus in arithmetical calculations in which 10 is the base, we usually write log 2, log 3,...... instead of log,,2, log,,3,...... Any number might be taken as the base of logarithms, and corresponding to any such base a system of logarithms of all numbers could be found. But before discussing the logarithmic systems commonly used, we shall prove some general propositions which are true for all logarithms independently of any particular base. 202. The logarithm of 1 ts 0. For a°=1 for all values of a; therefore log 1=0, whatever the base may be. 203. The logarithm of the base itself is 1. For a' =a; therefore log.a=1. 204. To find the logarithm of a product. Let MV be the product ; let a be the base of the system, and suppose w=log,M, y=log,N; so that W=M, a’=N. Thus the product MN =a" x a =a"; whence, by definition, log, MV=a+y =log, M+ log WV. Similarly, log, MNP = log, M + log, V + log, P; and so on for any number of factors. Example. log 42=log (2x 3x 7) =log 2+log 3+log 7. 205. Zo find the logarithm of a fraction. Tet 7 be the fraction, and suppose w= log.M, y=log.N; 80 that a=, av=N,LOGARITHMS, “177 Thus the fraction aie e Naw =a": 3 whence, by definition, log, Wat-Y + log, Af - log, NV. , 30 Ezample. Jog (47) =log 7 =log 30-log7 log (2x3 x 5) -log7 slog 2 + log 3 + log 5 — log 7. 206. Zo find the logarithm of a number raised to any power, integral or fractional. Let log,(Z") be required, and suppose x=log,M, so that a*= MU; then M= (ay =a; whence, by definition, log,(f") = px; that is, log, (M*) =plog, M. 1 Similarly, log, (d") = log, I. 207. It follows from the results we have proved that (1) the logarithm of a product is equal to the sum of the logarithms of its factors ; (2) the logarithm of a fraction is equal to the logarithm of the numerator diminished by the logarithm of the denominator ; (3) the logarithm of the p* power of a number is p times the logarithm of the number ; (4) the logarithm of the r root of a number is equal to ith of the logarithm of the number. Also we see that by the use of logarithms the operations of multiplication and division may be replaced by those of additior and subtraction ; and the operations of involution and evolutia by those of multiplication and division, HH, A, Ww178 HIGHER ALGEBRA, Ezample 1. Express the logarithm of we in terms of log a, logb and log c. 3 la? 2 3 = log a? — log (c5b*) = § log a~ (log e+ log 24 =} log a— 5 loge— 2 log. Example 2, Find x from the equation a* .c~*=)**"1, Taking logarithms of both sides, we have x log a — 2z log c=(3z+1) log b; «. @ (log a—2 logc — 8 log b) =log b; log b “ ®*foga- Blog ¢—8 log b” EXAMPLES. XVI. a. Find the logarithms of 1. 16 to base ./2, and 1728 to base 2./3. 2, 125 to base 5/5, and ‘25 to base 4. 3. ahs to base 2,/2, and ‘3 to base 9. ‘0625 to base 2, and 1000 to base ‘01. 5. 0001 to base -001, and “i to base 9/3. re 6. a, > a ? to base a. a 7. Find the value of 1 1 Jog, 128, log, a6? logy, a logs4349. \ Express the following seven logarithms in terms of loga, log}, and loge. 8 log(V ab), 9, log(d/a? x 40%). 10, log(Va-%5),_ll. 13. 15. 16. 17. LOGARITHMS, 179 log (Va-% x A/ab=3), 12, log(Wa-" J+ 83 Ja). ab-te-? be-2\-3 b-16\5) log ————,. 14, log }( 5. +(s=3) ft Sr (ea)"= Ge) 75.21 2, 4 2 Shew that log vis. 2 i log5 — 3 log2 3 log3. 4 oT = Simplify log Ni 729 J 9-1.278, Prove that log z —-2 log 2 +log 32 sa37 log2. Solve the following equations : 18, Fe 8B RB 8 a* = cb*, 19, a*, be=. at 21, a. bv m5 err a, by mf * If log (a*y)=a, and loge =8, find logs and logy. If a=, b= ar+5, b8, shew that slog (2) =loga. Solve the equation (at — 2a%b? + b4)*-1 = (a — b)**(a+b)-*. Common LocaRriTHMs. 208. Logarithms to the base 10 are called Common Logar- ithms ; contemporary of Napier the inventor of logarithms. this system was first introduced, in 1615, by Briggs, a From the equation 10*= J, it is evident that common logar- ithms will not in general be integral, and that they will not always be positive. For instance 3154 > 10° and < 10‘; . log 3154=3 +a fraction. W—2®180 HIGHER ALGEBRA. Again, 06 > 107° and <107'; «. log 06 =— 2 +a fraction. 209. Derinition. The integral part of a logarithm is called the characteristic, and the decimal part is called the mantissa, The characteristic of the logarithm of any number to the base 10 can be written down by inspection, as we shall now shew. 210. To determine the characteristic of the logarithm of any number greater than unity. Since 10'=10, 10*= 100, 10°= 1000, it follows that a number with two digits in its integral part lies between 10' and 10°; a number with three digits in its integral part lies between 10° and 10*; and so on. Hence a number with n digits in its integral part lies between 10"~' and 10*. Let WV be a number whose integral part contains » digits ; then V= 10-0 +8 fraction, . log V = (n— 1) +a fraction. Hence the characteristic is n— 1 that is, the characteristic of the logarithm of a number greater than wnity is less by one than the number of digits in its integral part, and 18 positive. 211. To determine the characteristic of the logarithm of a decimal fraction. Since 10°=1, 1 1 ane 10" = 45 =1 1 10°? = 75501, 10%= i 001,LOGARITHMS. 181 it follows that a decimal with one cipher immediately after the decimal point, such as 0324, being greater than -01 and less than -1, lies between 107* and 10-'; a number with two ciphers after the decimal point lies between 107° and 107%; and so on. Hence a decimal fraction with n ciphers immediately after the decimal point lies between 10-“*” and 107*. Let D be a decimal beginning with n ciphers; then D=107**) +8 fraction ; *. log D=—(n+1) +2 fraction. _~ Hence the characteristic is — (n+ 1) ; that is, the characteristic of the logarithm of a decimal fraction is greater by unity than the number of ciphers immediately after the decimal point, and is negative. 212, The logarithms to base 10 of all integers from 1 to 200000 have been found and tabulated ; in most Tables they are given to seven places of decimals, This is the system in practical use, and it has two great advantages : (1) From the results already proved it is evident that the characteristics can be written down by inspection, so that only the mantissz have to be registered in the Tables. (2) The mantisse are the same for the logarithms of all numbers which have the same significant digits; so that it is sufficient to tabulate the mantisse of the logarithms of integers. This proposition we proceed to prove. 213. Let W be any number, then since multiplying or dividing by a power of 10 merely alters the position of the decimal point without changing the sequence of figures, it follows that Wx 10’, and V+ 10, where p and q are any integers, are numbers whose significant digits are the same as those of WV. Now log (Wx 10°) =log V+ p log 10 slog Nt Peeeeeeccecceceee ceeeeeeeenees (1). Again, log (W~-10") =log V —q log 10 log Mag vecsessessestestestesessees (2). In (1) an integer is added to log , and in @ an iy a subtracted from log V7; that is, the mantissa or decimel of the logarithm remains unaltered,182 HIGHER ALGEBRA. In this and the three preceding articles the mantisse have been supposed positive. In order to secure the advantages of Briggs’ system, we arrange our work so as always to keep the mantissa positive, so that when the mantissa of any logarithm has been taken from the Tables the characteristic is prefixed with its appropriate sign according to the rules already given. 214. In the case of a negative logarithm the minus sign is written over the characteristic, and not before it, to indicate that the characteristic alone is negative, and not the whole expression. Thus 4-30103, the logarithm of -0002, is equivalent to.—4 + ‘30103, and must be distinguished from — 4°30103, an expression in which both the integer and the decimal are negative, In working with negative logarithms an arithmetical artifice will sometimes be necessary in order to make the mantissa positive. For instance, a result such as — 3:69897, in which the whole expression is negative, may be transformed by subtracting 1 from the characteristic and adding 1 to the mantissa. Thus — 369897 = — 4+ (1 — 69897) =4:30103. Other cases will be noticed in the Examples. Example 1. Required the logarithm of -0002432. In the Tables we find that 3859636 is the mantissa of log 2432 (the decimal point as well as the characteristic being omitted); and, by Art. 211, the characteristic of the logarithm of the given number is - 4; +. log 0002432 = 43859636, Example 2. Find the value of 2/-0000016 » given log 165 =2-2174839, log 697424 = 5-8434968, Let x denote the value required; then 1 log z= leg (00000165)° = ; log (-00000165) 1 = 5 (62174889); the mantissa of log 00000165 being the same as that of log 165, and the characteristic being prefixed by the rule. Now £(G-2174899)=5 (70-+4-2174889) =2-8484968LOGARITHMS. 183 and -8434968 is the mantissa of log 697424; hence z is a number consisting of these same digits but with one cipher after the decimal point. [Art. 211.] Thus 2='0697424, 215, The method of calculating logarithms will be explained in the next chapter, and it will there be seen that they are first found to another base, and then transformed into common loga- rithms to base 10. It will therefore be necessary to investigate a method for transforming a system of logarithms having a given base to a new system with a different base. (26) Sappes that the logarithms of all numbers to base a are ‘known and tabulated, it is required to find the logarithms to base 6. Let NW be any number whose logarithm to base 0b is re- quired. Let y=log,N, so that b=; log, (8) = log, ; that is, y log,b = log. NV ; 1 Y= Topp * 08, Se 1 or log, V = ogb poplOgEN geeenercrestniceesesstsn (1). Now since V and 6 are given, log,V and log,b are known from the Tables, and thus log,V may be found. Hence it appears that to transform logarithms from base a to base 6 we have only to multiply them all by Tood this is a eb j constant quantity and is given by the Tables ; it is s known as the modulus, 217. In equation (1) of the preceding article put « for V; thus 1 a 1 San logy = jogs” OB. foey) log,a x log.b = 1.184 HIGHER ALGEBRA, This result may also be proved directly as follows: Let a=log,b, so that a*=b; then by taking logarithms to base b, we have a log,a = log,b =1;. «. log.b x loga=1, 218. The following examples will illustrate the utility of logarithms in facilitating arithmetical calculation ; but for in- formation as to the use of Logarithmic Tables the reader is referred to works on Trigonometry. ‘ 5 Example1, Given log 3=:4771213, find log {(2°7)® x (‘81)5+(90)*}. The required value =3 log #2 + ; log ia - 4 log 90 =8 (log 8*— 1) +5 (log 8¢~ 2) -§ (og 3*+1) 16 5 86 = (945-5) os (345 +5) 97 = jg 108 3 - 545 = 4°6280766 — 5°85 = 2-7780766. The student should notice that the logarithm of 5 and its powers can always be obtained from log 2; thus log 5 =lo WP og 10 log 2=1 — log 2. Example 2. Find the number of digits in 875", given log 2 ="3010300, log 7 = "8450980. log (87515) = 16 log (7 x 125) =16 (log 7+ 3 log 5) , =16 (log 7 +38 -3 log 2) =16 x 2°9420080 =47-072128; ence the number of digits is 48. [Art, 210.)LOGARITHMS, 185 Example 3. Given log 2 and log 3, find to two places of decimals the value of x from the equation 68-42 , 4245 — 8, Taking logarithms of both sides, we have (3 - 42) log 6 + (x-+5) log 4=log 8; «. (3 —4z) (log 2+ log 3) + (w+ 5) 2log 2=3 log 2; os @(—4 log 2-4 log 34-2 log 2) =3 log 2-3 log 2 - 8 log 8 - 10 log 2; 10 log 2+3 log 8 2log 2+4 logs EXAMPLES, XVI. b. 1, Find, by inspection, the characteristics of the logarithms of 21735, 23°8, 350, -035, °2, 87, ‘875. 2, The mantissa of log’7623 is ‘8821259 ; write down the logarithms of 7°623, 762°, -007623, 762300, “000007623. 3. How many digits are there in the integral part of the numbers whose logarithms are respectively 430108, 14771213, 3°69897, ‘565157 4, Give the position of the first significant figure in the numbers whose logarithms are 37781513, 6910815, 5°4871384. Given log 2=:3010300, log3=-4771213, log 7 =-8450980, find the value of 5, log 64. 6. log 84. 7. log ‘128, 8. log 0125. 9, log 14-4. 10. log 43. 1, log’/T3, 12, log, / 2 18, log D108. 27° 14, Find the seventh root of 00324, having given that log 44092388 = 7°6443636, | 15. Given log 1946445 = 22896883, find the eleventh root of BLP.186 HIGHER ALGEBRA. 16, Find the product of 37:203, 3°7203, 0037203, 372030, having given that log 37-203=1°5705780, and log1915631 = 6-2823120, 325" 17. Given log 2 and log 3, find log AG: a) 18, Given log 2 and log 3, find log (4/48 x 108-176), 19. Calculate to six decimal places the value of a/ (Bax (Ba * 125 jE 42 x 32 given log 2, log 3, log’7; also log 9076-226 =3-9579053. 20. Calculate to six places of decimals the value of (330+ 49)'+- 4722 x70; given log 2, log 3, log 7; also log11=10413927, and 1og17814'1516 = 4:2507651. 21, Find the number of digits in 31? x 28, 21\10 22, Shew that Gs) is greater than 100, 23, Determine how many ciphers there are between the decimal 1000 point and the first significant digit in () Solve the following equations, having given log 2, log 3, and log 7. 24, 3-2=5, 25, 5*=10%, 26, 59-3942, 27, 217 = 241, 57, 98, 9%, G%- t= 52%, 71-2, 29, 2=+y=6" 30, 3i-t¥=4-7 Bid eee gtz-1 —33y-2f* 31, Given log,)2=°30103, find logs, 200. 32,. Given log,)2= 30103, logy) 7 ="84509, find log;,/2 and logy27.CHAPTER XVII. EXPONENTIAL AND LOGARITHMIC SERIES. 219. In Chap. xvi. it was stated that the logarithms in common use were not found directly, but that logarithms are tirst found to another base, and then transformed to base 10. In the present chapter we shall prove certain formule known as the Exponential and Logarithmic Series, and give a brief ex- planation of the way in which they are used in constructing a table of logarithms. 220. To expand a* in ascending powers of x. By the Binomial Theorem, if n>1, ( + x) w Line, 54 ERD) 5 ee) ot 1 1 2 none) eae), “ee a. By putting x=1, we obtain (a +2) 21 +14 Ge Coos) But (14 “ - {(Q+3) ts188 HIGHER ALGEBRA. hence the series (1) is the 2" power of the series (2); that is, 1 1 2 #(-5) 2-5) Gos) l+at Tr + Heceeee EDD =J14+1+ aaa and this is true however great n may be. If t therefore n be indefinitely ineroase we have ne BB a The series 14+1 ai eeeeee 2 “B *g is usually denoted by e; hence e= liste fe Beceeete / Write cx for x, then > Mears Now let e'=a, so that c=loga; by substituting for ¢ we obtain 2 (log.a)" | «* (log.a)* 2 ete ; This is the Zaponential Theorem. a®=1+aloga+ - Cor. When x is infinite, the limit of ( + 3) = [See Art. 266.] Also as in the preceding investigation, it may be shewn that when 7 is indefinitely increased, 147Jelte4 4 Bye ; (1stferter +S SesEXPONENTIAL AND LOGARITHMIC SERIES. 189 that is, when n is infinite, the limit of (: + =) =e. , By putting == -4, we have (-)-G+2)-{0-2)T- n, mM, ™, Now m is infinite when 1 is infinite; a\" thus the limit of (2 - 2) =e, Heneo the limit of (11 y= o 221. In the preceding article no restriction is placed upon the value of ; also since . is less than unity, the expansions we have used give results arithmetically intelligible. [Art. 183.] But there is another point in the foregoing proof which deserves notice. We have assumed that when x is infinite 1 2 -1 In Ir Sor all values of r. Let us denote the value of “(--) 6-3) ie Then . ie ( Tot) =f 545. r ron It is clear that the limit of u, is iz hence the limit of u, is 2 is that of u, is 3 and generally that of u, is i190 HIGHER ALGEBRA. 222. The series eel ath Et esses , which we have denoted by e¢, is very important as it is the base to which logarithms are first calculated. Logarithms to this base are known as the Napierian system, so named after Napier their inventor. They are also called natwral logarithms from the fact that they are the first logarithms which naturally come into consideration in algebraical investigations. When logarithms are used in theoretical work it is to be remembered that the base ¢ is always understood, just as in arithmetical work the base 10 is invariably employed. From the series the approximate value of ¢ can be determined to any required degree of accuracy ; to 10 places of decimals it is found to be 2°7182818284. Ezample 1, Find the sum of the infinite series eatat pt seneee Be Be 1,1 ‘We have e=14+1+ ere Pt eee ; a BE and by putting z= -1 in the series for e*, etel-1+5 " ebet=2 (+8 hence the sum of the series is 3 Leese), Example 2. Find the coefficient of x* in the expansion of toase = Lent a -as—aye* momar fae fy Ed eB aEXPONENTIAL AND LOGARITHMIC SERIES. 191 The coefficient required = a(t _(-ite_(-r* fee (“IF 1 parr (r—1)}. ac 223. To expand log, (1 + a) in ascending powers of x. From Art. 220, 2 . 3 w= +ylog.a+ 2 Oech , Loma aeons In this series write 1+ for a; thus (1 +ay =1+ylog,(1+a)+ ia tes. (t +2) + fg teed +a) + ...(1). Also by the Binomial Theorem, when x <1 we have (saya tye MOS Dar HOT Dar poeees (2). Now in (2) the coefficient of y is +) CVC) 5, 1-2) (-3) Fi a+ tog et Trg tt peers 5 ; a? a ast that is, e-g+g-qt Equate this to the coefficient of y in (1) ; thus we have a? at log, (l+a)=a—-p+3- 7+ peel This is known as the Logarithmic Series. Example. If «<1, expand {log, (1+2)}* in ascending powers of x. By equating the coefficients of y? in the series (1) and (2), we see that the required expansion is double the coefficient of y? in (y-1 1 2) 1 2) (y - 3) a as Nt pn? a ee date that is, double the coefficient of y in yok WM {e=4) y-}) y =) y-8) Traet 7 wt 1. sy Same ae Thus Hog (.+2)}4=2 fp gag (14 Dad (145 +B)a-192 HIGHER ALGEBRA. 224, Except when a is very small the series for log,(1 +) is of little use for numerical calculations. We can, however, deduce from it other series by the aid of which Tables of Logar- ithms may be constructed. By writing + for x we obtain Jog," 3 hence ieeiiee! log, (m + 1) —logya= = — gat gam eee (1). By writing = for x we obtain log,” = ; hence, by changing signs on both sides of the eauation log,n — log,(n — 1) == From (1) and (2) by addition, 1 1 1 log,(n +1) — log,(n - 1) =2 ( tant get -) bese (3). From this formula by putting n=3 we obtain log,4 —log,2, that is log,2; and by effecting the calculation we find that the value of log, 2 = -69314718...; whence log,8 is known. Again by putting n= 9 we obtain log,10—log,8; whence we find log,10 = 2°30258509.... To convert Napierian logarithms into logarithms to base 10 we multiply by ac aa 0 which is the modulus [Art. 216] of the common system, and its value is , or *43429448...; FAOEREBONT we shall denote this modulus by p. In the Proceedings of the Royal Society of London, Vol. xxvi1. page 88, Professor J. O. Adams has given the values of e, p, log, 2, log, 3, log,5 to more than 260 places of decimals, 225. If we multiply the above series throughout by yw, we obtain formule adapted to the calculation of common loyarithms. # f Thus from (1), 4 log,(n + 1) — p logan =" -% +h sm ee§EXPONENTIAL AND LOGARITHMIC SERIES. 198 that is, 1og,.(0 +1) —logn =H — Po Bm eee (1). Similarly from (2), log, — log, (m= 1) = Ee Ee opt sessseeee (2) From either of the above results we see that if the logarithm of one of two consecutive numbers be known, the logarithm of the other may be found, and thus a table of logarithms can be constructed. It should be remarked that the above formule are only needed to calculate the logarithms of prime numbers, for the logarithm of a composite number may be obtained by adding together the logarithms of its component factors. In order to calculate the logarithm of any one of the smaller prime numbers, we do not usually substitute the number in either of the formule (1) or (2), but we endeavour to find some value of n by which division may be easily performed, and such that either n+ 1 or m—1 contains the given number as a factor. We then find log(n+1) or log(m—1) and deduce the logarithm of the given number. Example, Calculate log 2 and log 8, given »=-48420448. By putting n=10 in (2), we have the value of log 10-log 9; thus 1-2 log 8=-048429448 + -002171472 + 000144765 +-000010857 ++000000868 + “000000072 +--000000006 ; 1-2 log 8=-045757488, log 8=-477121256. Putting n=80 in (1), we obtain log 81-1log 80; thus 4 log 8-8 log 2 — 1 =-005428681 — -000088929 + 0000002838 — 000000008 ; 8 log 2=+908485024 — -005895082, log 2=-801029997. In the next article we shall give another series for log, (n+ 1)—log,n which is often useful in the construction of Logarithmic Tables. For further information on the subject the reader is referred to Mr Glaisher’s article on Logarithms iw Yve Encyclopedia Britannica. , HA, WW194 HIGHER ALGEBRA. 226, In Art. 223 we have proved that a log, (1 +2)= 2-5 + Br changing x into —, we have a of log,(1-2)=-a— 37-37 By subtraction, ] a 2 ee ogg? (#+5 +5 +): ; we thus obtain l+ae n+l 1 Put -— = ——, so that a ore | l-« n 1 1 }og.(m +1) — logem = afecit SGasiy “serypt~}- Nore. This series converges very rapidly, but in practice is not always 80 convenient as the series in Art, 224. 227. The following examples illustrate the subject of the chapter. Ezample 1. If a, 6 are the roots of the equation az?+bz+c=0, shew 2th? abt pe that log (a— ba +ex*)=log a+ (a+f) 2 -— > — a+ 3 a Since a+p= -2, op =< » we have a-be+cx*=a {1+4(a+f) x+aBz7} =a (1+az) (1+8z). *. log (a - bx + cx?) =loga+log ie 38 x5 2 =logataz~ "2" +H + pe E>, Be a+ 6? , oF +f gt ge a =loga+(a+f)2—- Example 2. Prove that the coefficient of x" in the expansion of log (1+2+24) is — : or E according as 7 is or is not a multiple of 3. 1-3 log (1+2+2%) =log j— 7 gala (t — x5) — log (1-2) a of a a x ar S85 grec ret tg t gt tet).EXPONENTIAL AND LOGARITHMIC SERIES. 195 If n is a multiple of 8, denote it by 3r; then the coefficient of 2" is -} from the first series, together with é from the second series; that is, the coefficient is mori or = nin If n is not a multiple of 3, z* does not occur in the first series, therefore the required coefficient is 2, 228. To prove that © is incommensurable. For if not, let e= ° where m and 1 are positive integers ; m 11 1.1 then ni tltpt gt tnt pelt multiply both sides by |n; m|n-1 1 =inte; ~ _ —_1 mina © = integer +1 * (atlynt) * Getlyned)(m+d) * 1 1 1 But TI *@al@s® * @ely@eDmrsy* and less than the is a proper fraction, for it is greater than x 1 geometrical progression ae + 1 oy, re n+l (n+l) (n+ 1? that is, less than i, hence an integer is equal to an integer plus a fraction, which is absurd; therefore ¢ is incommensurable. EXAMPLES. XVII. 1. Find the value of 1,1 1,11 l-5+5-qtg-gte 2, Find the value of 1 1 1 1 1 a neta Tate 1196 HIGHER ALGEBRA. 8, Shew that a af og,(n+-a) —log,(n - a) = a(3+ 4+ 8+ 4 If a shew that =y+ +0 5, Shew that <=) ae (eS a “cate log,b. 6, Find the Napierian logarithm ot 1 pal correct to sixteen places of decimals. _ 2,3 7. Provethat e t= (Gt gtat Ds 8. Prove that 29 {2% at Fd Jog,(1 +.2)!*#(1 — 2)! 29( 5+ Sateet): "9, Find the value of 1 1 —yt4 (at = (48 — wa Pr ig V+ 3 PY) t+... 10, Find the numerical values of the common logarithms of 7, 11 and 13; given »="43429448, log 2=-30103000. 11. Shew that if az? and % za ore each less unity 1 a 1 a 1 a (+3) -§5 (#+3) +5 (#43)---= Hoe. (1+ast+0t4S). 12. Prove that 5a2 | Oa8 17x log, (1 + 84+ 228) = B-type and find the general term of the series. 13. Prove that oe 5a® 3523 654 dogg 88g FQ tes and find the general ‘orm of the series, © in a series of ascending powers of x, 14, Expand “sEXPONENTIAL AND LOGARITHMIC SERIES. 197 15, Express Fetee) in ascending powers of x, where i=J/—1. 16. Shew that log, (+ 2h) =2 log, (a +4) — 1 bs a ie log, (w+ 2h) = 2 log, (a +h) — ee (eent sort aeap ty: 17. Ifa and B be the roots of x*- px+q=0, shew that log.(1+-p2+gat)=(a+ pa 24 yay PHP go 18. If «<1, find the sum of the series 1 2 3 4, gut got ght ge Pees 19. Shew that tog. (1+-2)*=1- et oe. Te 2.3 +1f 3.441“ 1 A . 7 2. If logy tates be expanded in a series of ascending powers of x, shew that the coefficient of 2* is -3 if n be odd, or of the form 4m+2, and . if n be of the form 4m. 21, Shew that 3 4 Ets +] [a t= 5e. 22. Prove that 1 1 1 2 log, n— log, (n + 1) - log, (n-1)=23+ sat zt 1 1 1 n+it2(+iet 3@+ipt loi ii] =n ond t Bs 23. Shew that 24. If log, 2 = -a log 35 = -b, log. h=6, shew that log, 2=7a—2b+ 3c, log,8=1la—3b+5c, log,5=16a—4b+7c; and calculate log, 2, log, 3, log,5 to 8 places of decimals.CHAPTER XVIII. INTEREST AND ANNUITIES. 229. In this chapter we shall explain how the solution of questions connected with Interest and Discount may be simplified by the use of algebraical formule. We shall use the terms Interest, Discount, Present Value in their ordinary arithmetical sense; but instead of taking as the rate of interest the interest on £100 for one year, we shall find it more convenient to take the interest on £1 for one year. 230. To find the interest and amount of a given sum ina given time at simple interest. Let P be the principal in pounds, r the interest of £1 for one year, n the number of years, / the interest, and M the amount. The interest of P for one year is Pr, and therefore for n years is Pnr; that is, th ire ecerneeceteceneesrreartares (1). Also M=P+TI; that is, M=P(1+nr)... ++-(2). From (1) and (2) we see that if of the quantities P, n, 7, I, or P, n, r, M, any three be given the fourth may be found. 231, To find the present value and discount of a given sum due in a given time, allowing simple interest. Let P be the given sum, V the present value, D the discount, r the interest of £1 for one year, n the number of years.INTEREST AND ANNUITIES. 199 ( Since V is the sum which put out to interest at the present time will in n years amount to P, we have P=V(l+nr); __P ~T+ar" P And D=P-V=P- 5 l+nr _ Par ~L+nr’ Nore. The value of D given by this equation is called the true discount. But in practice when a sum of money is paid before it is due, it is customary to deduct the interest on the debt instead of the true discount, and the money so deducted is called the banker’s discount; so that Banker's Discount=Pnr, Pnr I+ar’ Ezample. The difference between the true discount and the banker's discount on £1900 paid 4 months before it is due is 6s. 8d.; find the rate per cent., simple interest being allowed. Let r denote the interest on £1 for one year; then the banker’s discount 1900r True Discount= is eal , and the true discount is l+5r 1900r . 19007 ~8~ 1 a 3 l+3r whence 1900r?=8+r; _14:,/1+22800__ 14161 ~~~ 8800 ~~ 8800 * as F 152s Rejecting the negative value, we have r= 3800 = 35+ .. rate per cent.=100r=4. 232. To find the interest and amount of a given sum in a given time at compound interest. Let P denote the principal, 2 the amount of £\ in one year, n the number of years, / the interest, and M the amount.200 HIGHER ALGEBRA. The amount of P at the end of the first year is PR; and, since this is the principal for the second year, the amount at the end of the second year is PR x R or PR*. Similarly the amount at the end of the third year is PA’, and so on; hence the amount in n years is PR" ; that is, M= PR’; I=P(R*-1). Nore. If, denote the interest on £1 for one year, we have R=1+4r. 233. In business transactions when the time contains a fraction of a year it is usual to allow simple interest for the fraction of the year. Thus the amount of £1 in } year is reckoned 1 +53 and the amount of P in 42 years at compound interest is PR (143°). Similarly the amount of P in n+ years is PE (1+). mm ™, If the interest is payable more than once a year there is a distinction between the nominal annual rate of interest and that actually received, which may be called the true annual rate; thus if the interest is payable twice a year, and if r is the nominal annual rate of interest, the anfount of £1 in half a year is 1 ee 5 2 and therefore in the whole year the amount of £1 is (2 +5) 5 or lered; so that the ¢rwe annual rate of interest is a r+7: 234. If the interest is payable g times a year, and if r is the nominal annual rate, the interest on £1 for each interval is = and therefore the amount of P in n years, or gn intervals, is P(1+5)", y In this case the interest is said to be “converted into principal” “mes 4 year.INTEREST AND ANNUITIES. 201 If the interest is convertible into principal every moment, then q becomes infinitely great. To find the value of the amount, put ==, so that g=rx; thus the amount = P ( ay" = Pal + ae P {a + =} q a, a, = Pe", [Art. 220, Cor.,] since 2 is infinite when q is infinite. 235. To sind the present value and discount of a given sum due in a given time, allowing compound interest. Let P be the given sum, V the present value, D the discount, & the amount of £1 for one year, x the number of years. Since V is the sum which, put out to interest at the present time, will in n years amount to P, we have P=VR*; P 7 Vege PR, and D=P(1-R"). Example. The present value of £672 due in a certain time is £126; if compound interest at 44 per cent. be allowed, find the time; having given log 2=-30108, log 8=-47712. =4t_i Eee Here T= 99 = ga! BI R=3;- Let n be the number of years; then a 672=126 G) ; 25 672 i nlog 53> 108 356° -, 100 16 or nlog 5 slog si «*. m (log 100 — log 96) =log 16 — log 8, __4log2—-log3 "=9-Blog2—log8 aqang 41, very nearly , thus the time is very nearly 41 years. .202 . HIGHER ALGEBRA. EXAMPLES. XVIILa. When required the following logarithms may be used. log 2=:3010300, log 3=-4771213, log 7 =*8450980, log11=1-0413927. 1, Find the amount of £100 in 50 years, at 5 per cent. compound interest; given log114:674= 20594650. 2. At simple interest the interest on a certain sum of money is £90, and the discount on the same sum for the same time and at the same rate is £80; find the sum. 3. In how many years will a sum of money double itself at 5 per cent. compound interest ? 4. Find, correct to a farthing, the present value of £10000 due 8 years hence at 5 per cent. compound interest; given log 67683°94 = 4:8304856. 5. In how many years will £1000 become £2500 at 10 per cent. compound interest ? 6. Shew that at simple interest the discount is half the harmonic mean between the sum due and the interest on it. 7. Shew that money will increase more than a hundredfold in a century at 5 per cent. compound interest. 8. What sum of money at 6 per cent. compound interest will amount to £1000 in 12 years? Given log 106=2°0253059, log 49697 = 4°6963292. 9, A man borrows £600 from a money-lender, and the bill is renewed every half-year at an increase of 18 per cent.: what time will elapse before it reaches £6000? Given log118=2°071882. 10. What is the amount of a farthing in 200 years at 6 per cent. compound interest ? Given log 106=2°0253059, log 115-0270 = 2°0611800, ANNUITIES. 236. An annuity is a fixed sum paid periodically under certain stated conditions; the payment may be made either once @ year or at more frequent intervals. Unless it is otherwise stated we shall suppose the payments annual. An annuity certain is an annuity payable for a fixed term of years independent of any contingency; a life annuity is an annuity which is payable during the lifetime of a person, or of the survivor of a number of persons.INTEREST AND ANNUITIES. 203 A deferred annuity, or reversion, is an annuity which does not begin until after the lapse of a certain number of years; and when the annuity is deferred for m years, it is said to commence after n years, and the first payment is made at the end of n+1 years. If the annuity is to continue for ever it is called a perpetuity ; if it does not commence at once it is called a deferred perpetuity. An annuity left unpaid for a certain number of years is said to be forborne for that number of years, 237. To find the amount of an annuity left unpaid for a given number of years, allowing simple interest. Let A be the annuity, r the interest of £1 for one year, n the number of years, M the amount. At the end of the first year A is due, and the amount of this sum in the remaining n—1 years is 4 +(n—1)74; at the end of the second year another A is due, and the amount of this sum in the remaining (n—2) years is 4+(n—2)rA; and so on. Now & is the sum of all these amounts ; oe M={A + (n-1) rd} + {A+ (n-2) rd} +... +(4+74)+4, the series consisting of n terms ; wt Mand +(14243 4.0... +n-1)rd =nd nnd, 238. To find the amount of an annuity left wnpaid for a given number of years, allowing compound interest. Let A be the annuity, R the amount of £1 for one year, the number of years, M the amount. At the end of the first year A is due, and the amount of this sum in the remaining n—1 years is AR""'; at the end of the second year another A is due, and the amount of this sum in the remaining n — 2 years is AR*-*; and so on. o Ma AR + ART 4+ +AR*+AR+A =A(1L+R+ Rit... to m terms) Rr-1 =4z-204 HIGHER ALGEBRA. 239. In finding the present value of annuities it is always customary to reckon compound interest; the results obtained when simple interest is reckoned being contradictory and un- trustworthy. On this point and for further information on the subject of annuities the reader may consult Jones on the Value - of Annuities and Reversionary Payments, and the article Annuities in the Encyclopedia Britannica. 240. To find the present value of an annuity to continue for @ given number of years, allowing compound interest. Let A be the annuity, 2 the amount of £1 in one year, n the number of years, V the required present value. The present value of A due in 1 year is 4 R'; the present value of A due in 2 years is AR‘; the present value of A due in 3 years is AR~*; and so on. [Art. 235.] Now V is the sum of the present values of the different payments ; . VaAR"4+ AR + AR +4...... to n terms _ ypu l-k™ =4R" j= 1-R™ =4RIT° Nore, This result may also be obtained by dividing the value of M, given in Art. 238, by R*. [Art. 282.] Cor, If we make n infinite we obtain for the present value of a perpetuity v= = A A R-1 7° 241. If mA is the present value of an annuity A, the annuity is said to be worth m years’ purchase. . In the case of a perpetual annuity mA = : 3 hence 100 1 m= ~ rate per cent. ’INTEREST AND ANNUITIES. 205 . that is, the number of years’ purchase of a perpetual annuity is obtained by dividing 100 by the rate per cent. As instances of perpetual annuities we may mention the income arising from investments in irredeemable Stocks such as many Government Securities, Corporation Stocks, and Railway Debentures. A good test of the credit of a Government is fur- nished by the number of years’ purchase of its Stocks; thus the 23 p. c. Consols at 96} are worth 35 years’ purchase; Egyptian 4p. c. Stock at 96 is worth 24 years’ purchase; while Austrian 5 p. c. Stock at 80 is only worth 16 years’ purchase. 242, To find the present value of a deferred annuity to commence at the end of p years and to continue for n years, allow- ing compound interest. Let A be the annuity, # the amount of £1 in one year, V the present value. The first payment is made at the end of (p+1) years. . [Art. 236.] Hence the present values of the first, second, third... pay- ments are respectively AR-?*, AR-@*®, AR-@*®, we Ve ARMs ARP 4 ARPS to n terms 1-R™ = — (pty Ab = ARE _AR? AR?™ “R-1 R-1° Cor. The present value of a deferred perpetuity to commence after p years is given by the formula y= AR? ~Rk-1° 243, A freehold estate is an estate which yields a perpetual annuity called the rent ; and thus the value of the estate is equal to the present value of a perpetuity equal to the rent. It follows from Art. 241 that if we know the number of years’ purchase that a tenant pays in order to buy his farm, we obtain the rate per cent. at which interest is reckoned by dividing 100 by the number of years’ purchase.206 HIGHER ALGEBRA. Ezample. The reversion after 6 of a freehold estate is bought for £20000; what renf ought the p ser to receive, reckoning compound interest at 5 per cent.? Given log 1-05 =-0211893, log 1-340096 =-1271868. The rent is equal to the annual value of the perpetuity, deferred for 6 years, which may be purchased for £20000, ~ Let £4 be the value of the annuity; then since R=1-05, we have +05)-8 20009 = 06-*, =. Ax (1:05)-$= 1000; log A - 6 log 1:05=8, log A =8-1271858=log 1840-096, ~. 4=1840°096, and the rent is £1840, 1s. 11d. 244, Suppose that a tenant by paying down a certain sum has obtained a lease of an estate for p+gq years, and that when q years have elapsed he wishes to renew the lease for a term ptm years; the sum that he must pay is called the fine for renewing 7 years of the lease. Let A be the annual value of the estate; then since the tenant has paid for p of the p+ years, the fine must be equal to the present value of a deferred annuity A, to commence after p years and to continue for years ; that is, AR? AR?™ the fine= 5-5 -BI7- [Art. 242.] EXAMPLES. XVIII b. The interest is supposed compound unless the contrary is stated. 1, A person borrows £672 to be repaid in 5 years by annual in- stalments of £120; find the rate of interest, reckoning simple interest. 2. Find the amount of an annuity of £100 in 20 years, allowing compound interest at 44 per cent. Given log1-045='0191163, log 24-117 = 1-3823260. 3. A freehold estate is bought for £2750; at what rent should it be let so that the owner may receive 4 per cent. on the purchase money? mn 4 A freehold estate worth £120 a year is sold for £4000; find the of interest,INTEREST AND ANNUITIES, 207 5. How many years’ purchase should be given for a freehold estate, interest being calculated at 34 per cent. ?. 6. Ifa perpetual annuity is worth 25 years’ purchase, find the amount of an annuity of £625 to continue for 2 years, 7. If a perpetual annuity is worth 20 years’ purchase, find the annuity to continue for 3 years which can be purchased for £2522. 8, When the rate of interest is 4 per cent., find what sum must be paid now to receive a freehold estate of £400 a year 10 years hence; having given log 104=2°0170333, log 6°75565 = "8296670. 9. Find what sum will amount to £500 in 50 years at 2 per cent., interest being payable every moment; given e~!=-3678, 10. If 25 years’ purchase must be paid for an annuity to continue n years, and 30 years’ purchase for an annuity to continue 2n years, find the rate per cent. 11, A man borrows £5000 at 4 per cent. compound interest; if the rincipal and interest are to be repaid by 10 equal annual instalments, End the amount of each instalment; having given log 1:04=-0170333 and log’675565 =6:829667. 12, A man has a capital of £20000 for which he receives interest at 5 per cent.; if he apenas £1800 every year, shew that he will be ruined before the end of the 17" year; having given log 2='3010300, log 3='4771213, log '7="8450980. 13. The annual rent of an estate is £500; if it is let on a lease of 20 years, calculate the fine to be paid to renew the lease when 7 years have elapsed allowing interest at 6 per cent.; having given 1og106= 20253059, 1og4°688385 = "6710233, log3°118042 = -4938820. 14. If a, b, c years’ purchase must be paid for an annuity to con- tinue n, 2n, 3n years respectively; shew that a@—ab+b=aec. 15. What is the Present worth of a perpetual annuity of £10 payable at the end of the first year, £20 at the end of the second, £30 at the end of the third, and so on, increasing £10 each year; interest being taken at 5 per cent. per annum ?CHAPTER XIX. INEQUALITIES. 245. Any quantity a is said to be greater than another quantity 6 when a—6 is positive; thus 2 is greater than — 3, because 2—(- 3), or 5 is positive. Also 5 is said to be less than a when b—a is negative; thus —5 is less than —2, because —5—(-2), or —3 is negative, In accordance with this definition, zero must be regarded as greater than any negative quantity. In the present chapter we shall suppose (unless the contrary is directly stated) that the letters.always denote real and positive quantities. 246. Ifa> ), then it is evident that at+e>b+e; a-c>b-c; ac > be; ab e7e? that is, an inequality will still hold after each side has been increased, diminished, multiplied, or divided by the same positive quantity. 247. If a-c>b, by adding c to each side, a>b+e; which shews that ix an inequality any term may be transposed from one side to the other of tts sign be changed. If a>, then evidently 5b b, then a-6 is positive, and b-a is negative ; that is, -a—(—6) is negative, and therefore -a<-b; hence, if the signs of all the terms of an inequality be changed, the sign of inequality must be reversed. Again, if a > b, then —a < —b, and therefore -ac <—be; that is, if the sides of am inequality be multiplied by the same negative quantity, the sign of inequality must be reversed. 248. If a,>b,, a,>b,, a>, ...... a, >, it is clear that G,+4,4+0,+...4+0,>6,+6,40,4+...4+3,5 and a,a,0, ...a,, > b,b,b, ... b,,. 249. if a>b, and if p, £ ore positive integers, then ./a>,/b, or at > B; 3 and therefore at > a; that is, a* > 6", where n is any positive quantity. Further, Le

0; - &+b* > Qab. Similarly “34 > Jay; that is, the arithmetic mean of two positive quantities is greater than their geometric mean. The inequality becomes an equality when the quantities are ‘equal. 251. The results of the preceding article will be found very useful, especially in the case of inequalities in which the \etvera are involved symmetrically. 4H A, 5 Vs210 HIGHER ALGEBRA, Example 1. If a, b, c denote positive quantities, prove that a? +d3+ c8>be+ca+ab; and 2 (a? + b9 +8) >be (b+c)+ca (c+a)+ab(a+b). For BPH C8 > BWe oo. ce eee eee eee (1); c%+a*> 2a; a?+b'>2ab; whence by addition a+)? +03>be+catad. It may be noticed that this result is true for any real values of a, b, c. Again, from (1) Be (2); o. B+e>be b+c) + (3). By writing down the two similar inequalities and adding, we obtain 2 (a? + b* +3) >be (b +c) +ca (¢ +a) +ab (a+b). It should be observed that (3) is obtained from (2) by introducing the factor b+c, and that if this factor be negative the inequality (3) will no longer hold. Example 2. If z may have any real value find which is the greater, +1 or 22+. a 4+1-(x?+2)=23- 2? - (2-1) = (22-1) («-1) =(2-1) (e+). Now (x - 1)? is positive, hence ®B+1> or or < -1. If «= —1, the inequality becomes an equality. 252. Let a and 6 be two positive quantities, S their sum and P their product; then from the identity 4ab = (a + b)* - (a—6)', we have 4P = S* -(a—6)*, and S*=4P+(a-b)*. Hence, if S is given, P is greatest when a=); and if P is given, S is least when a=b; that is, if the sum of two positive quantities is given, their product ts greatest when they are equal ; and if the product of two positive ~antities is given, their sum is least when they are equal,INEQUALITIES, 211 253. To find the greatest value of a product the sum of whose Sactors is constant, Let there be n factors a, b,c, ... k, and suppose that their sum is constant and equal to s. Consider the product abc... k, and suppose that a and 6 are any two unequal factors, If we replace the two unequal factors a, b by the two equal factors ost ’ _ the product is increased while the sum remains unaltered ; hence so long as the product contains two unequal factors it can be increased without altering the sum of the factors ; therefore the product is greatest when all the factors are equal. In this case the value of each of the n factors is < , and the greatest value of the product is (:); or (eres ee + n Cor. If a, 6, ¢,... & are unequal, >pabe...k; n (Pathe 7 that is, 1 arbor nthe aay BF, n By an extension of the meaning of the terms arithmetic mean and geometric mean this result is usually quoted as follows : the arithmetic mean of any number of positive quantities is greater t the geometric mean, Example. Shew that (1°+2°+3"+...-+n7)*>n*(|n)"; wi r is any real quantity. I+ or+ar+... tn” : 1 3 > (IT. OF. BT. wy; ; ; ince i ; | “ (FRE) aro a nf, that is, >(|n)Fs n . wh Jace we obtain the result required, af . VAD212 HIGHER ALGEBRA. 254. To find the greatest value of a™bc... whenatb+c+t... és constant; m, n, p,... being positive integers. Since m, n, p,... are constants, the expression a™b"c’... will a\" (b\" (c\?_— fi be greatest when (2) () () ... is greatest. But this last expression is the product of m+n+p+... factors whose sum is m(2)+n(F +p *) + -., Or @+b+c+..., and therefore con- stant. Hence ab"c’... will be greatest when the factors anna mon? pr are all equal, that is, when Thus the greatest value is mon, a+b+ot ...\mintet. mn"? ...( ———__— M+ M+ P+. Example. Find the greatest value of (a+)* (a—2)‘ for any real value of « numerically less than a. ‘at+a\§ (a-z 8 4 *7*) is greatest; but the sum of the factors of this expression is 3 ()+ 4 (7) ,or 2a; <== The given expression is greatest when ( hence (a+ <) (a - z)‘ is greatest when wate, or -} -_ 68. 8 Thus the greatest value is 7 a’. 255. The determination of maximum and minimum valufles may often be more simply effected by the solution of a quafd- ratic equation than by the foregoing methods. Instances flof this have already occurred in Chap. 1x.; we add a furtl{ & illustration. ,. Example. Divide an odd integer into two integral parts whose prod| 4 is & Maximum. Denote the integer by 2n+1; the two parts by z and 2n4+1-2; the product by y; then (2n+1)2-2z*=y; whence 2n=(2n+1)+ /(Qn+1)?—4y;INEQUALITIES, 213 but the quantity under the radical must be positive, and therefore y cannot be greater than 3 (2n-+1)%, or meneys and since y is integral its greatest value must be n?+7; in which case z=n+1, or n; thus the two parts aren andn+1. 256. Sometimes we may use the following method. Example, Find the minimum value of (ete) O42), Putc+x=y; then —(a-e+y) (b-c+y) 7 y —@-9 0-9) — Yy the expression +y+a-ctb-c = (AGE Wy) 4a-c4d-042 J (@-¢) (6-6). Hence the expression is a minimum when the square term is zero; that is when y=,/(a—c)(b—c). Thus the minimum value is a-c+b-c42 /(a—<) (b-<); and the corresponding value of x is ./(a—c) (—e) —c. EXAMPLES, XIX. a. 1, Prove that (ab + xy) (ax+by)>4abzy. 2, Prove that (b-+c)(c+a)(a+b)>8abe, 3. Shew that the sum of any real positive quantity and ita reciprocal is never less than 2. 4, If a?4+B=1, and 2*+y?=1, shew that ar+by<1. 5. If a?+b8+c!=1, and 22+y?+22=1, shew that ax +by+cz<1. 6. If a>, shew that a%}’> ab4, and tog b 92y%2?, 8, Find which is the greater 3ab? or a3+26%. 9, Prove that ab +ab3 abc (a+b+c).214 HIGHER ALGEBRA. 12. Which is the greater 2° or 2°+2+2 for positive values of x? 13, Shew that 2°+13ate> Saz"+9a°, if x>a, : 14. Find the greatest value of x in order that Tat+11 may be greater than 2°+17z. 15. Find the minimum value of «?—12%+40, and the maximum value of 247-8 — 92%, 16. Shew that ((n)?>n", and 2.4.6...20<(n+1)*. 17, Shew that (7+y+z)* > 27xyz. 18. Shew that n*>1.3.5,...(2n—1). 19. If n be a positive integer greater than 2, shew that 2 > 140 /30-1, 20, Show that (|n)?< wr (*E2)". 21. Shew that (1) (wt+y+2)8> 27 (y+2—2) (2+"-y) (x@+y—2). (2) aye> (y te- *)ete— ~Yety- -*. 22, Find the maximum value of (7- oh (2+) when z lies between Tand -2. of O+H)8+2) 23, Find the minimum value l+ae *257. To prove that if a and b are positive and unequal, a™+b™ /a+b\™ 5 me : > Cr) , except when m is a positive proper fraction. We have anyorn(238 aot) (P-+ i and since“ is less than 7% expressions in ascending powers of 2—* . (Art. 184.] » we ~~ expand each of these a+b" fatb 4 m(m=-1) 1) (a+b ‘a — b\*

(24 b\™ 2 2 . (2) If m is positive and less than 1, all the terms on the right after the first are negative, and therefore a™+b" (at 2" 3s ‘ee . (3) If m->1 and positive, put m=) where » <1; then 2 22 a+ B"\™ (a+ b™\" SS) Cry 1 (SEE) ELOY, ay ey, 1 . ae ost » (feb. amon (227 lama meioa aoe Hence the proposition is established. If m=0, or 1, the inequality becomes an equality. *258, If there are n positive quantities a, b, ¢,...k, then am+b™+o™4...4k™ (ebsstet ee en n n unless m is a positive proper fraction. Suppose m to have any value not lying between 0 and 1. Consider the expression a"+6"+c"+...+k", and suppose that a and bd are unequal ; if we replace a and 6 by the two equal quantities — = , the value of a+b+e+...+% remains un- altered, but the value of a" +5" +c"+ ... +" is diminished, since a” +b">2 oy .216 HIGHER ALGEBRA. Hence so long as any two of the quantities a, }, c,...k are unequal the expression a"+b"+c"4+...4+k" can be diminished without altering the value of a+b+¢+...+; and therefore the value of a™+b"+0"+...+%" will be least when all the quantities a, b, ¢,...k are equal, In this case each of the quantities is equal atb+e+...+k, aaa and the value of a” +6"+c"+... +4" then becomes (eee ty nm | ———_- . n to Hence when a, 4, ¢,...4 are unequal, a+b" +o" +... +k" fatbtot...+h\™ —jr SS ‘Ne If m lies between 0 and 1 we may in a similar manner prove that the sign of inequality in the above result must be reversed. The proposition may be stated verbally as follows : The arithmetic mean of the m“* powers of n positive quantities is greater than the m” power of their arithmetic mean in all cases except when m lies between 0 and 1. *259. If a and b are positive integers, and a>b, and if x be a : . > positive quantity, (i +3) (1 +5): For a\* 1\ 2" 1 2\ a* (2 +2) =1 +a+(1 -aet ( -)0 ~ a) pts the series consisting of a +1 terms; and x 1\ 2 1 2 a ( +5) ~1+0+ (1-5) i+ (1-5) (0-5) Ft the series consisting of b+ 1 terms. - After the second term, each term of (1) is greater than the corresponding term of (2); moreover the number of terms in (1) 8 greater than the number of terms in (2) ; hence the proposition tablished,INEQUALITIES, 217 *260. To prove that Jt JE fx and y are proper fractions and positive, and x > y. *fl+a "+ For [-t® futy ila lee I-y , according as tog} > or < leet 7 But blog 772 =2(1 +9 +340 4..), [Art 226); . 1, lty of, .¥.¥ and piepetn2 (149 +Es...). 1, l+a2 1 l+y 3 8T > 8ity and thus the proposition is proved. ¥261. To prove that (1+x)**(1—x)'*>1, ff x<1, and to att deduce that ath? > *) . . Denote (1 +2)'** (1—a)'"* by P; then log P = (1 + 2) log (1 +) + (1 — 2) log (1 — x) = 2 {log (1 +) — log (1 — x)} + log (1 + 2) + log (1 —2) a ott a ott at ~2e(24 5+ 5+.) -8(54 94 gt) a9(@,% 2 ). 5 2*3.4*5.6* Hence log P is positive, and therefore P>1; that is, (1+2)'**(1—2)'*>1,218 HIGHER ALGEBRA. In this result put 2= 2 » where u>z; then u (-9-e te fy —a\er* C) (“) >1*, or 1; uu uw we (w+ 2)" (uz) > u™, a+b. Now put w+z=a, w—z=6, so that w= —>— ; a+b . a> (45°) ‘ *EXAMPLES, XIX. b. 1, Shew that 27 (at+b4+ct) > (a+b+e)4. 2. Shew that 2 (n+1)? <8 (19+ 28+38+...+n3). 3. Shew that the sum of the m powers of the first » even num- bers is greater than 2 (n+1)", if m>1. 4. Ifa and 8 are positive quantities, and a> 8, shew that 1\¢ 1\8 (45)"> Gta)” Hence shew that if n>1 the value of (45) lies between 2 and 2°718... 5. If a, 6, ¢ are in descending order of magnitude, shew that ()"< Gey. a—6, b—«, + DF CH th othteh a ty ‘ n. 8 If nis a positive integer and x <1, shew that L-antt _ 1-a n+l nnINEQUALITIES. 219 9. Ifa, 6, care in H. P. and n> 1, shew that a*+c* > 2b". 10, Find the maximum value of x (40 —x)s if x is positive and less than 4a; and the maximum value of a(- x when « is a proper fraction. 11, If 2 is positive, shew that log (1+) <« and > et 12. If r+y+z=1, shew that the least value of atyte is 9; and that (1-2) (1—y) (1—2)>8xyz, "18, Shew that (2+b+¢+d) (+++) > (a+b 484d), 14. Shew that the expressions a (a—b)(a—c)+b (b—c) (b—a)+¢ (c—a) (e—b) and a3 (a—b)(a—c) +02(b—c) (b- a) +08(c— a) (e—b) are both positive. 15. Shew that (c™+9™)" <(a"%+y")", if m>n. 16, Show that ab < (ey. 17. Ifa, b, ¢ denote the sides of a triangle, shew that (1) @(p-9g) p—-7) +P (g—7)(g—-P) +E r—P)(r-9) cannot be negative; p, g, 7 being any real quantities ; (2) a®yz-+b%e0+-cny cannot be positive, if r+y+2=0. 18, Shew that |1 |3 [5 |2n—1 > (|n)* 19, If a,b,c, d,...... P positive integers, whose sum is equal to n, shew that the least value o! [a |b eld... is ([g?-*(\Q+)5 where g is the quotient and r the remainder when » is divided by p.CHAPTER XX. LIMITING VALUES AND VANISHING FRACTIONS. 262. Ir a be a constant finite quantity, the fraction = can be made as small as we please by sufficiently increasing x; that . a ; is, we can make z approximate to zero as nearly as we please by taking a large enough ; this is usually abbreviated by saying, ‘when w is infinite the limit of © is zero.” Again, the fraction < increases as x decreases, and by making a as small as we please we can make < as large as we please ; thus when « is zero < has no finite limit; this is usually ex- pressed by saying, “when « is zero the limit of ‘ is infinite.” 263. When we say that a quantity increases without limit or is infinite, we mean that we can suppose the quantity to become greater than any quantity we can name. Similarly when we say that a quantity decreases without limit, we mean that we can suppose the quantity to become smaller than any quantity we can name. The symbol co is used to denote the value of any quantity which is indefinitely increased, and the symbol 0 is used to denote the value of any quantity which is indefinitely dimi- nished. : : YLIMITING VALUES. 221 264. The two statements of Art. 262 may now be written symbolically as follows : if x isco, then = is 0; if zis 0, then © is @. But in making use of such concise modes of expression, it must be remembered that they are only convenient abbreviations -of fuller verbal statements. 265. The student will have had no difficulty in understanding the use of the word mit, wherever we have already employed it; but as a clear conception of the ideas conveyed by the words limit amd limiting value is necessary in the higher branches of Mathematics we proceed to explain more precisely their use and meaning. 266. Derinition. If y=/(x), and if when x approaches a value a, the function f(a”) can be made to differ by as little as we please from a fixed quantity 6, then 6 is called the limit of y when =a. For instance, if S denote the sum of terms of the series Lag + get get eos then S=2- i [Art. 56.] 37% ae Here S is a function of n, and _ can be made as small as we please by increasing n; that is, the limit of § is 2 when n is infinite. ‘ 267. We shall often have occasion to deal with expressions consisting of a series of terms arranged according to powers of some common letter, such as G, +a,0+ Og +a, +...... where the coefficients a,, a,, a,, @,,... are finite quantities independent of a, and the number of terms may be limited or unlimited. It will therefore be convenient to discuss some propositions connected with the limiting values of such expressions under certain conditions,222 HIGHER ALGEBRA. 268. The limit of the series a +axtaxt+ax?t...... when x is indefinitely diminished is a,. Suppose that the series consists of an infinite number of terms. Let 6 be the greatest of the coefficients a,, a,,@,,...; and let us denote the given series by a,+S; then S Uy > Uy > Uy> Uy reer The given series may be written in each of the following forms : (tu, = Uy) + (tig — Uy) + (Uy — Uy) F ceeceeee sevens (1), U, — (Uy — Uy) — (Uy, — Uy) — (Ug — Uy) — ore seeerreee (2). From (1) we see that the sum of any number of terms is a positive quantity ; and from (2) that the sum of any number terms is less than u,; hence the series is convergent.CONVERGENCY AND DIVERGENCY OF SERIES. 231 281. For example, the series 1,1 1,11 l-g+g-gtgr gt is convergent. By putting x=1 in Art, 223, we see that its sum is log, 2. Again, in the series 2 3 at 5 6 T°2°3°4°5~ each term is numerically less than the preceding term, and the series is therefore convergent. But the given series is the sum of ois 1,1 1,1. 1 l-5+g-atg et eeemes piesa (1), and : 1-14+1-1+1-1+...... Meee: (2). Now (1) is equal to log, 2, and (2) is equal to 0 or 1 according as the number of terms is even or odd. Hence the given series is convergent, and its sum continually approximates towards log. 2 if an even number of terms is taken, and towards 1 + log,2 if an odd number is taken. 282. An infinite series in which all the terms are of the same ; sign is divergent if each term is greater than some finite quantity however small. For if each term is greater than some finite quantity a, the sum of the first m terms is greater than na; and this, by taking n sufficiently great, can be made to exceed any finite quantity. 283. Before proceeding to investigate further tests of con- vergency and divergency, we shall lay down two important principles, which may almost be regarded as axioms. I. If a series is convergent it will remain convergent, and if divergent it will remain divergent, when we add or remove any finite number of its terms; for the sum of these terms is a finite quantity. IL If a series in which all the terms are positive is con- vergent, then the series is convergent when some or all of the terms are negative; for the sum is clearly greatest when all the terms have the same sign. We shall suppose that all the terms are positive, unless th contrary is stated.232 HIGHER ALGEBRA. 284, An infinite series is convergent if from and after some fied term the ratio of each term to the preceding term is numerically less than some quantity which is itself numerically less than unity. Let the series beginning from the fixed term be denoted by WU, + Uy + Uy + Ut vrrereene 8 uU, uU, u, and let — ; Here we have a case of a convergent series in which the terms may increase up to a certain point and then begin to decrease. 99 1 00° then i=am 100, and the terms do not to decrease until after the 100 term. For example, if #=CONVERGENCY AND DIVERGENCY OF SERIES. 233 286. An infinite series in which all the terms are of the same sign ts divergent if from and after some fixed term the ratio of each term to the preceding term is greater than unity, or equal to unity. Let the fixed term be denoted by u,. If the ratio is equal to unity, each of the succeeding terms is “equal to u,, and the sum of n terms is equal to nu, ; hence the series is divergent. If the ratio is greater than unity, each of the terms after the fixed term is greater than u,, and the sum of n terms is greater than nw, ; hence the series is divergent. 287. In the practical application of these tests, to avoid having to ascertain the particular term after which each term is greater or less than the preceding term, it is convenient to find the limit of “= when n is indefinitely increased ; let this limit be denoted by x If <1, the series is convergent. [Art. 284.] If A>, the series is divergent. [Art. 286.] If A=1, the series may be either convergent or divergent, and a further test will be required; for it may happen that Me 2) but continually approaching to 1 as its limit when n is iidifinitely imereased, In this case we cannot name any finite quantity r which is itself less than 1 and yet greater than A, Hence the test of Art. 284 fails. If, however, 3 tinually approaching to 1 as its limit, the series is divergent by Art, 286. We shall use “Lim ” as an abbreviation of the words “the limit of Pad when » is infinit a (n+ (te Example 1. Find whether the series whose n‘* term is is con- vergent or oe234 HIGHER ALGEBRA. hence if z<1 the series is convergent; if z>1 the series is divergent. If z=1, then Lim aah and a further test is required. me Example 2. Is the series 1+ 2r+ 87294 4225 +...... convergent or divergent? tn nig Here Lim ra =Lim a @-ie* Hence if z<1 the series is convergent; if z>1 the series is divergent. If c=1 the series becomes 17+ 27+ 87+ 47+ ..., and is obviously divergent, Example 8. In the series a+ (at+d)r+(a+2d) r+...4+(a+n—-1.d)r +... on Yn pam tt(n— ld Lim tg a4 (nnd) d thus if r<1 the series is convergent, and the sum is finite. [See Art. 60, Cor.] Tar; 288. If there are two infinite series in each of which all the terms are positive, and if the ratio of the corresponding terms in the two series is always finite, the two series are both convergent, or both divergent. Let the two infinite series be denoted by Uy + Ug t Uy HUF ceecceees : and UU; Heer The value of the fraction U+U,+U,t... my Me vy ae ghee a [Art, 14] and is therefore a finite quantity, L say ; U+Uz+Ust...+u= Lv, +v,+0,+...+4,). Hence if one series is finite in value, so is the other; if one series is infinite in value, so is the other; which proves the ‘oposition.CONVERGENCY AND DIVERGENCY OF SERIES. 235 289. The application of this principle is very. important, for by means of it we can compare a given series with an auziliary series whose convergency or divergency has been already esta- blished. The series discussed in the next article will frequently be found useful as an auxiliary series. 290. The infinite series lela tet pet get, ts always divergent except when p is positive and greater than 1. CasE I. Let p> 1. The first term is 1; ‘the next two terms together are less than 3 the following four terms together are less than 4 ; the fol- lowing eight terms together are less than iS ; and soon. Hence 8 the series is less than 14344 Bo that is, less than a geometrical progression whose common ratio 7 is less than 1, since py >1; hence the series is convergent. CaszII. Let p=1. - 1,1,1,1 The series now becomes 1 +5 +gtqt5t ae The third and fourth terms together are greater than 7 or o the following four terms together are greater than or 3 the following eight terms together are greater than 5 or 7 3 and 80 on. Hence the series is greater than 1+ a + a + a + z + see bee ee , and is therefore divergent. [Art. 286.] Case III Let p<1, or negative. Each term is now greater than the corresponding term in Case II., therefore the series is divergent. Hence the series is always divergent except in the case whe p is positive and greater than unity.236 HIGHER ALGEBRA. Example. Prove that the series 2,8 4 n+1 qtatgtet arte is divergent. ; . A 1,1 1 Compare the given series with l+g+ gt te PEP eee Thus if u, and v, denote the n™ terms of the given series and the auxiliary series respectively, we have hence Lim ot 1, and therefore the two series are both convergent or both divergent. But the auxiliary series is divergent, therefore also the given series is divergent. This completes the solution of Example 1. Art. 287. 291. In the application of Art, 288 it is necessary that the limit of “s should be finite; this will be the case if we find our auxiliary ‘series in the following way: Take u,, the n term of the given series and retain only the highest powers of x. Denote the result by v,; then the limit of = is finite by Art. 270, and v, may be taken as the n term of the auxiliary series. . « /ar-1. Example 1, Sh@w that the series whose n‘* term is —~7~———— is A/Bn842n45 divergent. As n increases, u, approximates to the value oe 55 TBE Yas ga My Mans? 8" ct 3 Hence, if m=4, we have Lim ta? | which is a finite quantity; a % V8 therefore the series whose n* term is S may be taken as the auxiliary : nu series, a this series is divergent [Art. 290]; therefore the given series is “xgent.CONVERGENCY AND DIVERGENCY OF SERIES. 237 Ezample 2. Find whether the series in which : u,= Ynt+1—n is convergent or divergent. Here mon (,/ 145-1) But the auxiliary series 1,1,1 1 piptgt gt. is convergent, therefore the given series is convergent. 292. To shew that the expansion of (1+x)" by the Binomial Theorem is convergent when x <1. Let u,, u,,, represent the r and (r+ 1)" terms of the ex- pansion ; then 4,4, n-r+l1 art x. u, r When r>n+1, this ratio is negative; that is, from this point the terms are alternately positive and negative when x . is positive, and always of the same sign when & is negative. Now when + is infinite, Lim Mest = numerically; therefore since a <1 the series is convergent if all the terms are of the same sign ; and therefore a fortiori it is convergent when some of the terms are positive and some negative. [Art. 283.] 293. To shew that the expansion of a* in ascending powers of x is convergent for every value of x. Here “== zee 3 and therefore Lim “s <1 whatever be — n- n=) the value of 23 hence the series is convergent. a238 HIGHER ALGEBRA. 294, To shew that the sxpaneion of log (1+) in ascending powers of x i oowergont selon x is memmorically lve than 1 n-1 Here the numerical value of “2 = 2—-. a, which in the limit is equal to x; hence the series is Scavermene when 2 is less than 1. If w=1, the series becomes 1-5+5- it vy and is con- vergent. [Art. 280.] If w=—1, the series becomes —-1—,;-—5-7- vey and is divergent. [Art, 290.] This shews that the logarithm of zero is infinite and negative, as is otherwise evident from the equation e"=0. 295, The results of the two following examples are important, and will be required in the course of the present chapter. Example 1, Find the limit of ee when « is infinite, Put c=e¥; then loge _y ee Levey +t “T2428 ~+ltptiet.. y BB also when 2 is infinite y is infinite; hence the value of the fraction is zero. Example 2. Shew that when n is infinite the limit of nz*=0, when <1, Let emi, so that y>1; also let y*=z, so that nlogy=logz; then wi 1 loge__ 1 loge my a logy logy’ Now when n is infinite z is infinite, and 12849; also logy is finite ; therefore Limnz*=0, 296. It is sometimes necessary to determine whether the product of an infinite number of factors is finite or not. Suppose the product to consist of n factors and to be denoted by UU Ug reves Uy then if as m increases indefinitely u, <1, the product will ulti- mately be zero, and if u,>1 the product will be infinite ; hence in order that the product may be finite, u, must tend'to the limit 1.CONVERGENCY AND DIVERGENCY OF SERIES. 239 Writing 1+, for u,, the product becomes (1+4,) (1+%,)(1+4,) 0... (1+4,). Denote the product by P and take logarithms ; then log P=log (1 +9,) + log (1 +¥,) +... + log (1+) ...... (1), and in order that the product may be finite this series must he convergent. Choose as an auxiliary series O+%,4%,4.. w.-(2). 1. . Yu yp Uatee Now rim 28040) im ( = = since the limit of v, is 0 when the limit of u, is 1. Hence if (2) is convergent, (1) is convergent, and the given product finite. Example. Shew that the limit, when n is infinite, of 133557 2Qn-1 Qn+1 2°2°4°4°6°6°"" Qn" Qn is finite. The product consists of 2n factors; denoting the successive pairs by U1, tty, Us,... and the product by P, we have Paty Uy tg. Uys a _2m-1 Qntl_, 1 “eon an an but log P=logu, + log ug +log uy +...+log uy ......... (), and we have to shew that this series is finite. 1 1 1 Now. log u,=log (1- a) =- me" ae therefore as in Ex, 2, Art. 291 the series is convergent, and the given product is finite. 297. In mathematical investigations infinite series occur so frequently that the necessity of determining their convergency or divergency is very important ; and unless we take care that the series we use are convergent, we may be led to absurd conclusions, (See Art, 1837240 HIGHER ALGEBRA. For example, if we expand (1—)~* by the Binomial Theorem, we find (L—a)* = 1+ a+ 308 + daP +... But if we obtain the sum of n terms of this series as ex- plained in Art. 60, it appears that 1-2" na” a Le 14204 327+ ... +o" = Qa T-3 whence 1 a1 a nx" aaa *adaayt ics’ By making n infinite, we see that aaa can only be re- garded as the true equivalent of the infinite series 1+ Qar + Ba? + dak+ .... ao" ne" fs when a-= +Tme vanishes, If m is infinite, this quantity becomes infinite when z= 1, or >1, and diminishes indefinitely when «<1, [Art. 295], 0 that it is only when 2<1 that we can assert that 1 ; Cite eaeeaaeaae tenes to. inf. ; and we should be led to erroneous conclusions if we were to use the expansion of (1 —2)~* by the Binomial Theorem as if it were true for all values of a In other words, we can introduce the infinite series 1+ 2” + 3a*+... into our reasoning without error if the series is convergent, but we cannot do so when the series is divergent. The difficulties of divergent series have compelled a distinction to be made between a series and its algebraical equivalent. For example, if we divide 1 by (1—2)*, we can always obtain as many terms as we please of the series 1+ Qe+ 3a + 4aP4+ o.. whatever « may be, and so in a certain sense ac may be called its algebraical equivalent ; yet, as we have seen, the equi- —-lence does not really. exist except when the series is con-CONVERGENCY AND DIVERGENCY OF SERIES. 241 vergent, It is therefore more appropriate to speak of as the generating function of the series 14 2x4 Saft... being that function which when developed by ordinary alge- braical rules will give the series in question. The use of the term generating function will be more fully, explained in the chapter on Recurring Series. i 1 (= EXAMPLES. XXI. a. Find whether the following series are convergent or divergent,: 1 1 1 1 1 zg aeat aia aesat” a and a being positive quantities.” 1 1 1 1 Veataateatast 171 1 l 3 Fy GNF) * @42) 94%) @FB) HBT wand y being positive quantities. x a a a rateatsatrete z a wz at Tatsatsetret 2 3 4 6. l4pt Bt geo i 2, 4 ie Sat J 3+ Sit [$+ 8. 14304 527+ 7254 90A+...... 2,3 ,4,65 ptgotgnt+ gt aeees n co “an ata tigte tart n-1 a 2, 4, 5. 10. 1+ 3 8 15 11, ate io? + Wt wet HH, A. W242 HIGHER ALGEBRA. 2 1 1 14+5 B7+5 62414 Wott ee HF osenee 17 12. Gta +i pt ap teres Be 403 (w+1) 2" We et Et Go beg tosses w. (i-7) +@-g) +(G-3) + Baty 16. ltats +e 33° 17. Test the series whose general terms are Q) Ja+i-n (2) Jnt+1- J/nt—1. 18. Test the series 1 1 1 a) Hesitaa test aa , 1 ® Stattantecataat a being a positive fraction 19, Shew that the series +Etpt gt teense is convergent for all values of p. : 20, Shew that the infinite series ~ Uy + Ug + Ug + Ug t+ eeeeee is convergent or divergent according as Lim 1. 21, Shew that the product 22446 Q%m-2 mm-2 In 1°3°3°5°5 2n—3° 2n—-1° In-1 is finite when 7 is infinite. 22. Shew that when «=1, no term in the iesreneen of (1+-2)* is infinite, except when n is negative and numerically greater than unity.CONVERGENCY AND DIVERGENCY OF SERIES, 243 *298. The tests of convergency and divergency we have given in Arts. 287, 291 are usually sufficient. The theorem proved in the next article enables us by means of the auxiliary series Beat + u Het E + pret Rete tote to deduce additional tests which will sometimes be found con- venient. *299. If uy, vz are the general terms of two infinite series in which all the terms are positive, then the u-series will be con- vergent when the v-series is convergent if after some particular term os = ; and the u-series will be divergent when the v-series is 1 Van divergent if = >a Va-1 n-1 Let us suppose that u, and v, are the particular terms. Case I. Let “<%, MO eres 3 then , UU, %, that is, <2 +4, +0, + ode 1 Hence, if the v-series is convergent the w-series is also con- vergent. U, », wu, v, CaseII. Let #>-#, > ...... ; then uy, %, uy »%, U, + Ut Uyt overs uU, uU, u, =u, (1+? +—. 4+ ) Uy 2 uy,244 HIGHER ALGEBRA. that is, >A ey+oto, + see) r Hence, if the vseries is divergent the weeries is also di- vergent, "#300. We have seen in Art. 287 that a series is convergent or divergent according as the limit of the ratio of the n term to the preceding term is less than 1, or greater than 1. In the remainder of the chapter we shall find it more convenient to use this test in the equivalent form : A series is convergent or divergent according as the limit of the ratio of the n* term to the succeeding term is greater than 1, or less than 1; that is, according as Lim vs +h Similarly the theorem of the preceding article may be enunciated : The wseries will be convergent when the v-series is convergent provided that Lim vergent when the w-series is is dveigent provided that >1, or <1. Ys 3 and the wseries will be di- v, Lim —- 1 the auxiliary series is convergent, and in this case the given series is convergent if J CA or (1 +2); U, ‘n+ that is, if Mn 14 Py ees >), a; ta, n or n( wn -1)> sptlga ), eeets nt that is, if Lim {nu os ep __ a SNCONVERGENCY AND DIVERGENCY OF SERIES. 245 But the auxiliary series is convergent if p is greater than 1 by a finite quantity however small ; hence the first part of the proposition is established. When p<1 the auxiliary series is divergent, and by proceed- ing as before we may prove the second part of the proposition. Example. Find whether the series 2,1 a8 1.8 2 1.8.5 ita 3 *3.4°s *3.4.6° is convergent or renee +. “18 Here Lim =3; hence if «<1 the series is convergent, and if s>1 the series is divergent. If z=1, Lim —“» =1, In this case ett _2n(2n+1) tata @n—1) (20-1)? ; n (ee : 1) =n (6n—1), and ata @n-iy! Li ve =3, » Lim | n ( ) =33 hence when z=1 the series is convergent. *302. The series whose general term is u, is convergent or di- vergent, according as Lim (a log =) >1, or <1. nt Let us compare the given series with the series whose general . 1 term is —. n When p>1 the auxiliary series is convergent, and in this case the given series is convergent if Me aC + iy ; [Art 300.] 7 », or if gt oP Fie ...5 “a ‘ath that is, if246 HIGHER ALGEBRA, that is, if Lim (nog “.)) oo ter Hence the first part of the proposition is established. When p<1 we proceed in a similar manner ; in this case the auxiliary series is divergent. Example. Find whether the series Be? BS dtet Bat Ep et Bt is convergent or divergent. Here My tet (nelyettentt __on® dl, Ue |B In+1 (n+1)*2 (143) o Lim tn a3, [Art. 220 Cor.]. Untr e% Hence if ot the series is divergent. If at, then un ce Me 142 Be n, Un 1 o. log “*-=loge—al (143) og *-=log e— n log (1+ = =l-n ae re ~ n Sn" Bnd +. Li Sam ee eae Lim (nlog *-) = 5; hence when a=} the series is divergent. *303. If Lim -“* =1, and also Lim {n ie 1)} =1, the tests given in Arts,"300, 301 are not applicable” To discover a further test we shall make use of the auxiliary 1 ; n (log ny’ In order to establish the convergency or divergency of this series we need the theorem ved in the next article. series whose general term isCONVERGENCY AND DIVERGENCY OF SERIES. 247 *304. If p(n) te positive for all positive integral values of n continually diminishes as n increases, and if a be any positive integer, then the two infinite series (1) + $(2) +. $(3) +... + O(n) +...5 and acp(a) + a'p(a*) + a°p(a") +... ta"p(a*)+..., are both convergent, or both divergent. In the first series let us consider the terms o(@ +1), o(a+2), p(a+3), ...... Pa) eee (1) beginning with the term which follows $(a’). The number of these terms is a**'- a’, or a’(a—1), and each of them is greater than ¢(a'*’); hence their sum is greater than a'(a—1) p(a'"'); that is, greater than = x al" (ab*?), By giving to & in succession the values 0, 1, 2, 3,... we have (2) + f(3) + (4) +o. +(a)> x adp(a); 1 a-1 a a $(a+1)+$(a+ 2) + $(a+ 3)+...... + $(a*)> an aba"); therefore, by addition, _S, - ¢(1) > a Sis where S,, S, denote the sums of the first and second series respec- tively; therefore if the second series is divergent so also is the first. Again, each term of (1) is less than ¢(a'), and therefore the sum of the series is less than (a — 1) x a'p(a’). By giving to & in succession the values 0, 1, 2, 3... we have $(2) +4(8) +6 (4) + soon. h(a) < (a—1) x (1); p(at+1)+6(a+2)+$(a4+3)+...... + $(a’) <(a-1) x ag(a); therefore, by addition S,- $(1) <(a-1) {8,+ 6}; hence if the second series is convergent so also is the first. Nors. To obtain the general term of the second series we take ¢ (hdl general term of the first series, write a” instead of h and multiply Wy248 HIGHER ALGEBRA. + *305. Tha serie hoe general torm is ts convergent ¥f p> 1, and divergent if p=1, or p<1. By the preceding article the series will be convergent or divergent for the same values of p as the series whose general term is . 1 1 ox a” (log a*)?? °r (nm log a)?’ or (log a)? a The constant factor ‘sey is common to every term; there- fore the given series will be convergent or divergent for the same values of p as the series whose general term is op Hence the required result follows. [Art. 290.] *306. The series whose general term is u, is convergent or di- vergent according as Lim [ {fn (2 ~1)=1} tog »] >1, or <1. Uy 1 Let us compare the given series with the series whose general term is n(logay . When p>1 the auxiliary series is convergent, and in this case the given series is convergent by Art. 299, if My (n+ 1) {log (2 +1)? a7 log nh eee (1). Now when n is very large, log (n+ 1) =logn+ log (1 +3) =logn +2, nearly ; Hence the condition (1) becomes % (14+ gy; u, n, nlogn, at that is, is > (1 +2) 1+—? Me w, n, nlogn, ‘vt is, Me] Pear p_, n nilogn’CONVERGENCY AND DIVERGENCY OF SERIES, 249 or n (Ss -l)>1+ P_; ts, logn or {n (a - 1) - i} logn>p. Hence the first. part of the proposition is established. The second part may be proved in the manner indicated in Art. 301. Example. Is the series 2 28.43 28, 42, 6? lth a t 5 Bt. o8 t arpa eeaieee convergent or divergent? (2n-+1)9_ 1 Here os eae aie d Fe (. .. Lim “* =1, and we proceed to the next test. Ut From (1), n(™= ~-1)=142 (2) , te BSR ag eee 7 tim {n (= - 1)} =1, and we pass to the next test. a )- } = en, F (2), I} log n = : rom (2) t (a, °8 4n stil o(-4)-}ses J be [Art. 295]; hence the given series is divergent. since Lim *307. We have shewn in Art. 183 that the use of divergent series in mathematical reasoning may lead to erroneous results. But even when the infinite series are convergent it is necessary to exercise caution in using them. For instance, the series a te ya 78" ya” ye is convergent when a=1, [Art. 280.] But if we multiply the series by itself, the coefficient of 2” in the rome’ is 1 1 1 : 1 4 f2n * in =i * Yim * Nr. Tent q 1-24”250 HIGHER ALGEBRA. Denote this by a,,; then sine 1 vr. Yr. ine an-r > Coy a OF > 1 we a, > a , and is therefore infinite when 1 is infinite. If e=1, the prota t becomes Gy — Oy +O, — By + 0s + gy — yyy + Boggy — oo and since the terms a,, a, are infinite, the series has ‘ant? anaes no arithmetical meaning, This leads us to enquire under what conditions the product of two infinite convergent series is also convergent. *308. Let us denote the two infinite series Qt Get ag + ag? +... +00" 4+... b, +0 + bc + baht... + Bot... by A and B respectively. If we multiply these series together we obtain a result of the form a,b, + (a,b, + a,b,) # + (a,b, + a,b, +.a,b,) 2+... Suppose this series to be continued to infinity and let us denote it by C; then we have to examine under what conditions C may be regarded as the true arithmetical equivalent of the product AB. First oe that all the terms in A and B are positive. Tet A,,, B, ,, denote the series formed by taking the first 2n +1 terms any B, C respectively. If we multiply together the two series A,,, B,,, the coefficient of each power of x in their product is equal'to the coefficient of the like power of x in C as far as the term 2“; but in A, B,, there are terms containing powers of » higher than a", whilst " is the highest power of a in C,,; hence A,B, > Cup nan If we form the product A,B, the last term is ab"; but (,, includes all the terms in the” product and some other terms besides ; hence CL>AB,CONVERGENCY AND DIVERGENCY OF SERIES. 251 Thus C,, is intermediate in value between A,B, and A,B,,, whatever be the value of n. Let A and B be convergent series ; put A,=A-X, B,=B-Y, where X and Y are the remainders after terms of the series have been taken; then when n is infinite XY and Y are both indefinitely small. *. A,B,=(4—-X)(B- Y)=AB-BX-AY+XY; therefore the limit of A,B, is 4B, since A and B are both finite. Similarly, the limit of 4,,B,, is AB. Therefore C which is the limit of C,, must be equal to AB since it lies between the limits of 4,B, and 4, B,,. Next suppose the terms in A and B are not all of the same sign. In this case the inequalities 4,B,.>C,,> A,B, are not necessarily true, and we cannot reason as in the former case, Let us denote the aggregates of the positive terms in the two series by P, P’ respectively, and the aggregates of the negative terms by J, WV’; so that A=P-N, B=P-N. Then if each of the expressions P, P’, V, N’ represents a con- vergent series, the equation AB=PP'- NP'- PN'+ NN’, has a meaning perfectly intelligible, for each of the expressions PP, NP, PN', NN’ is « convergent series, by the former part of the proposition ; and thus the product of the two series 4 and B is a convergent series, Hence the product of two series will be convergent provided that the sum of all the terms of the same sign in each is a con- vergent series, But if each of the expressions P, NV, P’, N’ represents a divergent series (as in the preceding article, where also P’ =P and V'= J), then all the expressions PP’, VP’, PN’, NN’ are divergent series. When this is the case, a careful investiga tion is necessary in each particular example in order to ascertains whether the product is convergent or not.252 HIGHER ALGEBRA. *EXAMPLES. XXI. b. Find whether the following series are convergent or divergent : 1 of 1.3.5 of 1.3.5.7.9 af Lo143-a+e 406° 8t24.6.8.10° 1st" 3 3.6.9 3.6.9.12 ae ot es itr 0+ FiO Ts 10.18” *7, 10.13.16" t at 93, 42 28, 43, 6? 3 + ats a ee" t34.6.6.7.8 2. 3, 4, eg By Be ‘er er ee . 12 12,38 12, 33, 52 gat org? tar ar git to 1. 142059) a) +Gteed- pe a) (@ta(+e ot (1 — a) (2-a) (3— (B-a) 18,933? @ being a proper fraction. a+e , (a+2e)? | (a+30)? 2 8 a.B al (a+1)B B+1) 22 Ty Ty Gh) 4 vlat)) (a+2) B(B+1) (B+2),, 1.2.3. y(ytly+2) * 10, x? (log 2)? +23 (log 3)* + 2 (log 4)"+...... a(a+1) 4 H@t1) (a2) | 1.2 1.2.3 9, 1+ ll. 1l+a+ ur _nk+ An* 1+ Brk-24 Cn 34... 12, ae at an T+ bn? + on +... integer, show ‘that the series ty + tig tts aseees is convergent if 1 is positive, and divergent if A-—a—1 is negative or zera » Where & is a positiveCHAPTER XXII. UNDETERMINED COEFFICIENTS. 309. In Art. 230 of the Elementary Algebra, it was proved that if any rational integral function of x vanishes when z= a, it is divisible by c—a. [See also Art. 514. Cor.] Let Doe + pa + pg t+ ee +P, be a rational integral function of x of n dimensions, which vanishes when 2 is equal to each of the unequal quantities ,, Gy Gy w- O. Denote the function by f(a); then since f(x) is divisible by a—a,, we have S (&) = (@—4,) (pg + eee » the quotient being of »—1 dimensions. Similarly, since f(a) is divisible by x —a,, we have Pg + veveee = (w— ay) (pyr? + os »° the quotient being of n —2 dimensions ; and per +... Proceeding in this way, we shall finally obtain after n di- visions S@)=Ple—a,) (ea) (2-4) oon. (=a). 310. Ifa rational integral function of n dimensions vanishes Jor more than n values of the variable, the coefficient of each power of the variable must be zero. Let the function be denoted by f(x), where S (2) = py" + pet + ppt + ee +P,3254 HIGHER ALGEBRA. and suppose that f(z) vanishes when 2 is equal to each of the unequal values a,, @,, G, ...... a,; then S(@)=P, (=a) (4) (4) «..... (@—a,). Let c be another value of a which makes /(x) vanish; then since f(c) = 0, we have P, (0-4) (0a) (¢= 4) «11... (0-4) = 05 and therefore p,=0, since, by hypothesis, none of the other factors is equal to zero. Hence f(a) reduces to pe + pat? + ptt os. +2,: By hypothesis this expression vanishes for more than n values of a, and therefore p, = 0. ‘ In a similar manner we may shew that each of the coefficients Pes Dg serves p, must be equal to zero. This result may also be enunciated as follows: If @ rational integral function of n dimensions vanishes for more than n values of the variable, it must vanish for every value of the variable. Cor. If the function f(x) vanishes for more than n values of x, the equation f (x) = 0 has more than n roots. Hence also, if an equation of n dimensions has more than n roots tt is an identity. Example. Prove that (2) (ze) , (@-c) (z@-a) | (e-a) (2—2)_, (a—6) (a-c) © (b-c) (b—a) “ (c-a) (c-b) ~~ This equation is of two dimensions, and it is evidently satisfied by each of the three values a, b, c; hence it is an identity. 311. If two rational integral functions of n dimensions are equal for more than n values of the variable, they are equal for every value of the variable. Suppose that the two functions Det" + pe" + part i. +25 et + G0"! + gt oe. +s ‘we equal for more than x values of x; then the expression 9.) 2" + (D,- 9) B+ (Pa 9.) BPH vee + (D-H)UNDETERMINED COEFFICIENTS. 255 vanishes for more than n values of x; and therefore, by the preceding article, Po~ =} P,— I= Py—Gy= 04 2 PLT = 5 that is, = Gor Pr= Us Pa= U9 °° Pa=Qn° Hence Po two expressions are tdentical, and therefore are equal for every value of the variable. Thus ¥f two rational integral functions are identically equal, we may equate the coefficients of the like powers of the variable. This is the principle we assumed in the Elementary Algebra, Art. 227. Cor. This proposition still holds if one of the functions is of lower dimensions than the other. For instance, if Pw + p+ pet + pg +... +p, = Qe + gg t+ we. +4,, we have only to suppose that in the above investigation g, = 0, q, = 0, and then we obtain Po=% P,=9, Py=To Pa= Yar vre-s P= Ger 312. The theorem of the preceding article is usually referred to as the Principle of Undetermined Coefficients. The application of this principle is illustrated in the following examples. Example 1, Find the sum of the series 1.243.848.4400 +n(n+1), Assume that 1.24+2.84+8.44...+2(n+1)=4+Bn+Cn?+ Dri+ Ent+.. where A, B, C, D, £,... are quantities independent of n, whose values have to be determined. Change n into n+1; then 1.24+2.8+4...+n(n+1)+(n+1) (n+2) =A+B (n+1)+C (nt+1)?4D (n+-18+E (n41)4+ By subtraction, _ (ntl) (nt 2)= =B+C (2n+1)+D(Bnt+Bnt1) +H (dn+ 6+ dnl) +... ” This equation being true for all integral values of n, the coefficients of the respective powers of n,on each side must be equal; thus Z and all succeeding coefficients must be equal to zero, and 8D=1; eee ; D+04+B=2; whence p=}, c=1, Bai. a256 HIGHER ALGEBRA. Hence the sum Ae a he bat To find A, put n=1; the series then reduces to its first term, and 7 2=A+2, or 4=0. Hence 1.242.84+8.44...4n(n+l)=on (n+l) (n+2). Nore. It will be seen from this example that when the n“ term is a rational integral function of n, it is sufficient to assume for the sum a function of n which is of one dimension higher than the n‘* term of the series. Example 2. Find the conditions that a°+ pz" +qx+r may be divisible by a+an+b. Assume + part qetr= (atk) (22+ax+). Equating the coefficients of the like powers of x, we have kta=p, ak+b=q, kb=r. From the last equation hae ; hence by substitution we obtain a b that is, r=b (p—a), and ar=b (q-b); which are the conditions required. +a=p, and Gtb=as EXAMPLES. XXII. a. Find by the method of Undetermined Coefficients the sum of 194-37 + 524 72+ ...to m terms. 1.2.342.3.443.4.5+...to n terms. 1,224+2.37+3. 4244. 52+...t0 terms. 134-33 + 53+ 73+.,..t0 n terms, . 144 24 344 444... to m terms, Find the condition that 28-3px%+2g may be divisible by a factor of the form 2*+2ax+ a, 7. Find the conditions that a2 + ba*+cxr+d may be a perfect cube. 8, Find the conditions that a%#+b23+cx!+dxe+/f? may be a A~perfect square, 9. Prove that ax%+2bry+cy?+ Idx+2ey+f is a perfect square, if P=ac, Pas, Pmof. Pe PE QnUNDETERMINED COEFFICIENTS. 257 10, If ax*+b2*+cr-+d is divisible by 2*+ A, prove that ad=be, lL If 2—5gxr+-4r is divisible by (x—c)', shew that g=r', 12, Prove the identities : qQ a? (x —b)(x—c) B(x—c)(e—a) | A (z-a)(n—b) _ a (a—0)(a—a) * =a (—a) * (e-a)(e—¥) ~*~ a) 2=e=0)(e-d) , (2-0) (2- (22) (a=8)(a—e)(a=d) * (6=c) (B—d) (=a) 4, 2D lena) (=) (2-2) (2-2) (w-9) (e=d) (e=a) (e=8) * (d=a)(d=8)(d-<) 13, Find the condition that axt+ hay + by? + Igx + Wy +e may be the product of two factors of the form prtgytr, prtgytr. Ww If f=le+mytnz, n=netlyt+mz, (=me+ny+tlz, and if the same equations are true for all values of «, y, 2 when &, , ¢ are inter- changed with 2, y, z respectively, shew that B+2mn=1, m?+2ln=0, n?+2lm=0. 15. Shew that the sum of the products »—r together of the n quantities a, a%, a,,..a" is (@ttt—1) (ar? ~1)..(@*=1) Hn- niin. (a= 1) (a*—1)...(a"-*—1) 2? 313. If the infinite series a, +0,x +0,x* +a,x°+...... is equal to zero for every finite value of x for which the series is convergent, then each coefficient must be equal to zero identically. Let the series be denoted by S, and let S, stand for the ex- pression a, +ae+ ac" + ; then S=a,+28,, and therefore, by hypothesis, a, + #S,= r all finite values of x. But since S is convergent, S, cannot exceed some finite limit; therefore by taking @ small enough aS, may be made as small as we please. In this case the limit of S is.a,; but S is always zero, therefore a, must be equal to zero identically. Removing the term a,, we have «8, = 0 for all finite values of x; that is, a,+aetag*+...... vanishes for all finite values of x Similarly, we may prove in succession that each of the coefficients a,, @,, @,).....- is equal to zero identically. H. HA. XU 4258 HIGHER ALGEBRA. 314, If two infinite series are equal to one another for every JSinite value of the variable for which both series are convergent, the coefficients of like powers of the variable in the two series are equal, Suppose that the two series are denoted by a, +ae+a+ag°+.. and A,+Aet+ Ag+ Ag+. then the expression a,—A,+ (a, -—A,) 2+ (a,—A,)a*+(a,-A,) 2 4+...... vanishes for all values of x within the assigned limits; therefore by the last article a,-A,=0, a,—A,=0, a,-A,=0, a,—A,=0,...... that is, a,=Ay a,=A, a,=A, a,=Ay...... ; which proves the proposition. Example 1. Expand as Z in a series of ascending powers of a as far as the term involving 2°. 2 Let Pet mote taget tages... where a, 4, 4, 43,... are constants whose values are to be determined; then 2498 (142-2) (ag+a,2 +027 + a234+...), In this equation we may equate the coefficients of like powers of z on each side. On the right-hand side the coefficient of <* is a,+ a, and therefore, since x? is the highest power of « on the left, for alt values 3 of n>2 we have Ag+ Oy_1— On—g=05 this will suffice to find the successive coefficients after the first three have been obtained. To determine these we have the equations %=2, a, +4)=0, a,4+0,-a)9=1; whence a)=2, a= ~-2, a,=5. Also y+, —a,=0, whence a= —7; @,+03—,=0, whence a,=12; . and @+@,-a,=0, whence a,= -19; thus Bet = 2-2e4 bat Ta 4 18et— 19eh+.. itz-UNDETERMINED COEFFICIENTS, 259 Example 2, Prove that if » and r are positive integers woman ED ay MOOR D oye. is equal to 0 if r be less than n, and to |p ifran. a x xt i ‘We have e-e(s+gtetiat ae ) =2 + terms containing higher powers of z...(1). Again, by the Binomial Theorem, -1 (4 -1)*=e— nes 4 OD aens eee pier (2). By expanding each of the terms e**, e*-)*,... we find that the coefficient of 2? in (2) is aT (R= 1)" | m(n—l) (n— 2)" _n(n-1) (n-2) (n-3)° re" EEE BE and by equating the coefficients of x* in (1) and (2) the result follows. 7? Ezample8. It + yrartdz3+ecx8+...,.. e express z in ascending powers of y as far as the term involving y*. Assume z=pytgy?tryst...... . and substitute in the given series; thus y=a(pytgyrt ryt...) +b (py tgy?t PHO (PYF GH.) torre Equating coefficients of like powers of y, we have ap=1; whence p=4. "aq +bp?=0; whenoe g=~ 2, 2b? ar+2bpq+ep*=0; whence rete Git Thus ant Me, Oe ade, This is an example of Reversion of Series. Cor, If the series for y be given in the form y=k+ax+ba?+ca3+... put y-k=z; then z=antbatteat+...5 from which # may be expanded in ascending powers of z, thatinofy—k. ~ yd260 : HIGHER ALGEBRA, EXAMPLES. XXII. b. Expand the following expressions in ascending powers of « as far as 23, 1+2z¢ 1-8 l+a L aw z 1-a2- 62?" 3 Q+at+ar 4 34+2 5. 1 " Q-g-2" * l+ax-axt— xs" at+be 6. Find a and b so that the n™ term in the expansion of —~—, may be (3n— 2)a*~1. 7. Find a, b, ¢ so that the coefficient of 2" in the expansion of oe may be n®+1. 8. Ify?+2y=«(y+1), shew that one value of y is (=a? 9. If cx’+az—y=0, shew that one value of x is aa Hence shew that «=-00999999 is an approximate solution of the equation #3+1002—-1=0. To how many places of decimals is the result correct ? 10, In the expansion of (1+) (1+am)(1+a%r) (1+a%z)......, the number of factors being infinite, and a< 1, shew that the coefficient of 1 ae (1 -a) (l—a@?) (l-a)...... (l-ar) . 11, When a <1, find the coefficient of «* in the expansion of 1 (1-2) (1—a%x) (1—a%z)...... to inf." 12. If nis a positive integer, shew that af is (q) meetin apes OD (n- 21-bit; (2) m—(n+1) (n— 1+ 4 Oe ayn rf the series in each case being extended to n terms ; and (3) 1 nan BOD ge... (1) [ns () (nspp—ninep-1p 48) 4he series in the last two cases being extended to n+\ terms. k (n+p -2)r—CHAPTER XXIII. PartiaL FRACTIONS. 315. In elementary Algebra, a group of fractions connected by the signs of addition and subtraction is reduced to a more simple form by being collected into one single fraction whose denominator is the lowest common denominator of the given fractions. But the converse process of separating a fraction into a group of simpler, or partial, fractions is often required. For 3-52 1-424 32* ing powers of 2, we might use the method of Art. 314, Ex. 1, and so obtain as many terms as we please. But if we wish to find the general term of the series this method is inapplicable, and it is simpler to express the given: fraction in the equivalent form 1 Toe tics Each of the expressions (1—2)~' and (1 — 32)" can now be expanded by the Binomial Theorem, and the general term obtained. example, if we wish to expand in a series of ascend- 316. In the present chapter we shall give some examples illustrating the decomposition of a rational fraction into partial fractions. For a fuller discussion of the subject the reader is referred to Serret’s Cours d’Algébre Supérieure, or to treatises on the Integral Calculus. In these works it is proved that any rational fraction may be resolved into a series of partial fractions; and that to any linear factor «—a in the denominator there cor- responds a partial fraction of the form — 3 to any linear factor x—6 occurring twice in the denominator there correspond two partial fractions, as and Pa . VE ab occurs three 5 ond Ga times, there is an additional fraction B,. amd so on. TS @ Ya262 HIGHER ALGEBRA. any quadratic factor x*+px+q there corresponds a partial fraction of the form sete if the factor 2*+ px +q occurs twice, there is a second partial fraction wy 3 and so on. Here the quantities A,, B,, B,, B,...... P,Q, P,, Q, are all independent of a. We shall make use of these results in the examples that follow. Example 1. Separate moss a into partial fractions. Since the denominator 2z?+2-6=(x+2) (2x - 3), we assume be-ll 4, B Ietye—6 2421 Be" where 4 and B are quantities independent of z whose values have to be determined. Clearing of fractions, Br-11=A (22-8)+B (2+2). Since this equation is identically true, we may equate coefficients of like powers of x; thus 944+B=5, -3442B=-11; whence 4=3, B=-1. a ae dat a2-6 242 2a-3° +n 7 7 Example 2. Resolve a @53) into partial fractions. matn _ A B aoe (e-a)(2+b) za aad oe mMa+N=A (2+0) +B (t—-a).. eee (1). We might now equate coefficients and find the values of A and B, but it is simpler to proceed in the following manner. Since 4 and B are independent of x, we may give to x any value we please. In (1) put z-a=0, or c=a; then putting 2+b=0, orz=-), B= metn — 1 maken . mb —w * (e-a)(z+b) at+d\ a-a © xxbPARTIAL FRACTIONS, 263 282-112? Ezample 8. eT ae 280112 No Assume = @z-1)(@+2)(@-2) ~Se-1* B+a* iS Bel 282 —11z9=A (84-2) (8-2) +B (2x- 1) (8-2) +0 (2-1) (8 +2). By putting in suocession 2s-1=0, 3+2=0, 8-z=0, we find that A=1, B=4, C=-1. ., Sbe-Met _ 1 | 4 “* @e-1)@-2) 32-11 8+ 8-2" Be84+2-2 * ass, . Ezample 4. Resolve (ay (1 — aa) ine? partial fractions, B28+2-23 A B c Assume = (@~a)"(1-2a) 1-32 *z-a* @-a?* wt. Bate B= A (c-2)"+ B (1-22) (x -2)4+C (1-24). Let 1-22=0, then a=-}, let -2=0, then C=-4, To find B, equate the coefficients of z*; thus 8=4-9B; whence D=-°. Bt 42-2 1 5 4 * @-3) (1-22) 3 (1-22) 8(e-2) (z- 2)?" 42-192 Example 6. Resolve @re-) into partial fractions. 42-192 Azt+B Cc @rie-H 7 war tz 42 -192= (42 +B) (z- 4) +C (2241). Let 2=4, then C=-2; equating coefficients of x, 0=4+C, and A=2; equating the absolute terms, 42= -4B+C, and B= -11, 42-192 2a-11 2 @+De-)> w+ 2-4 Assume 317. The artifice employed in the following examps Wik sometimes be found useful.264 HIGHER ALGEBRA. Qa. 24x + 48a . e ; Example. Resolve e-3)"@+1) into partial fractions. Qx8-24294 482 Af (2) : Assume =)" (e41) =aei*@-3 where 4 is some constant, and f (x) a function of « whose value remains to be determined. 9a — 24.29 + 480= A (xe - 2)4-+ (2 +1) f (2), Let c= -1, then A=-1. Substituting for 4 and transposing, (e+) f (x) =(e- 2)"-4 923 — 2423 + 48x 270 HIGHER ALGEBRA. from which we see that although the generating function a, + (4, — pa) & 1 — px — ga" may be used to obtain as many terms of the series as we please, it can be regarded as the true equivalent of the infinite series G+ ABH OLE + 0.0, only if the remainder (pa, +944) 2" + 90,_,2"" 1 — px — qa* vanishes when 1 is indefinitely increased; in other words only when the series is convergent. 328. When the generating function can be expressed as a group of partial fractions the general term of a recurring series may be easily found. Thus, suppose the generating function can be decomposed into the partial fractions eee l-ax” 1+bx ” (1—ca)*" Then the general term is {da’ + (— 1) Bo’ + (r +1) Cedar’. In this case the sum of n terms may be found without using the method of Art. 325. Example. Find the generating function, the general term, and the sum to n terms of the recurring series 1-72 - 2*-4823-...... Let the seale of relation be 1-p2-qz*; then -1+7%p-q=0, -434+p+7q=0; whence p=1, g=6; and the scale of relation is 1-2- 62%, Let S denote the sum of the series; then S=1-Te- 2°-4823-...... -2S= - c+7a%+ a+ ~ 62°S = — 628 + 4223+ .. (1-2- 621) S=1-82, which is the generating function. ereRECURRING SERIES. 271 If we separate aes into partial fractions, we obtain ee 1-2-6 1+ whence the (r +1)" or general term is {(-1prartt — 3ry ar, Putting r=0, 1, 2,...0-1, the sum to n terms = {2 — 2% Bx8 — + (— 1) 821} - (14 824 8228+... B8 1 eA}) _24(-1s tae en 1-328 - 142% 1-32 * 1 : im 329. To find the general term and sum of n terms of the recurring series a,+@,+a@,+...... » we have only to find the general term and sum of the series a,+a,7+4,2°+...... , and put a= 1 in the results. Ezample. Find the general term and sum of n terms of the series 1464244844. The scale of relation of the series 1+ 6x + 24a7+ 84234... is 1— 50+ 623, and the generating function is 1- 62+ 627° This expression is equivalent to the partial fractions 4 8 1-82 1-22" If these expressions be expanded in ascending powers of x the general term is (4.8°-8. ar) 2", Hence the general term of the given series is 4.8"—3. 2°; and the sum of n terms is 2(3"-1)-38(2"-1). 330. We may remind the student that in the preceding article the generating function cannot be taken as the sum of the series 1 + 6a+ 240" + 84a7+...... except when has such a value.as to make the series convergent. Hence when #=1 (in which case the series is obviously divergent) the generating function is not a true equivalent of the series, But the general term of 14+64+244+84+...... ts independent of x, and whatever value x may have it will always be the coefficient of 2” in 1+ 6a + 242° + 84a°+. We therefore treat this as a convergent series and find ite general term in the usual way, and then put «=1.272 HIGHER ALGEBRA. EXAMPLES, XXIV. Find the generating function and the general term of the following series : 1, 145249294 1325+....... 2 2—-x+502— T+ ....... 3. 24+3r+522+905+.,....., 4. 7-6v+92°+2704+....... 5. 34+6x+ 1422+ 36x53 + 98244 276a5+....... Find the n™ term and the sum to x terms of the following series : 6, 24+54+134+35+.. 7 -1+629+3025+....... 8. 24+'7e+2522+9125+, . 9. 14+ 2x4 6x?+ 2023 + 6624+ 212254....... 10, —3+24048+4....... 11. Shew that the series 12+ 224-324 48+ 0.0... +n’, 134-284 33+ 484.00... +n, are recurring series, and find their scales of relation. 12, Shew how to deduce the sum of the first n terms of the re- curring series Ay +O, + Mgt? + ase +... from the sum to infinity. ; 13, Find the sum of 2n+1 terms of the series 3-14+13-9+441—53+....... 14, The scales of the recurring series Opt aye + dye? + agt8+. by + bye + bar? + byx3+, are l+pxe+qa’, 1+rx+az*, respectively; shew that the series whose general term is (a,+5,)* is a recurring series whose scale is 1+ (ptr)a+(q+stpr) 2? + (gr+ps) 25+ gsr, 15. If a series be formed having for its n* term the sum of n terms of a given recurring series, shew that it will also form a recurring series whose scale of relation will consist of one more term than that of the given series,CHAPTER XXV. CONTINUED FRACTIONS. is called a b d c+— e+e. continued fraction; here the letters a, b, c,...... may denote any quantities whatever, but for the present we shall only consider 331. An expression of the form a+ the simpler form a, + — » where a,, a,, @,,... are positive tt. integers. This will be usually written in the more compact form 11 a +— a,+ a+ pb ee 332. When the number of quotients a,, a,, a,,... is finite the continued fraction is said to be terminating ; if the number of quotients is unlimited the fraction is called an infinite continued Sraction. It is possible to reduce every terminating continued fraction to an ordinary fraction by simplifying the fractions in succession beginning from the lowest. 333. To convert a given fraction into a continued fraction. Let = be the given fraction; divide m by n, let a, be the quotient and p the remainder ; thus m ea, +P =a,+ n n 3 rg138te274 HIGHER ALGEBRA, divide n by p, let a, be the quotient and q the remainder ; thus mo 4%e ade pp erp? q divide p by q, let a, be the quotient and r the remainder ; and so on, Thus If m is less than n, the first quotient is zero, and we put m1 <= slat and proceed as before. It will be observed that the above process is the same-as that of finding the greatest common measure of m and n; hence if m and m are commensurable we shall at length arrive at a stage where the division is exact and the process terminates, Thus every fraction whose numerator and denominator are positive integers can be converted into a terminating continued fraction. 2 re 5 Ezample, Reduce Es to a continued fraction. Finding the greatest common measure of 251 and 802 by the usual process, we have 5 | 251 | 802 | 3 6 6) 49] 8 1 and the successive quotients are 3, 5, 8, 6; hence 251 _ i i1oi1i 802° «34+ .5+ 8+ GC 334. The fractions obtained by stopping at the first, second, third,...... quotients of a continued fraction are called the first, second, third,...... convergents, because, as will be shewn in Art. 339, each successive convergent is a nearer approximation to the true value of ‘the continued fraction than any of the seceding convergents.CONTINUED FRACTIONS. 275 335. To shew ‘that the convergents are alternately less and greater than the continued fraction. Let the continued fraction be a, + a a,+ a+ The first convergent is a,, and is too small because the part = + ..- is omitted. The second convergent is a, +4, and is too great, because the denominator a, is too small. The third convergent is a, +o , and is too small because ath is too great ; and so on. a ° ‘When the given fraction is proper fraction a, =0; if in this case we agree to consider zero as the first convergent, we may enunciate the above results as follows : The convergents of an odd order are all less, and the convergents . of an even order are all greater, than the continued fraction. 336. To establish the law of formation of the successive con- vergents. Let the continued fraction be denoted by 1 1 1 a+— -—— — a, ; then the first three convergents are a, @,a,+1 a,(a,a,+1)+a,, 1’ a, ’ @,-4,+1 ” and we see that the numerator of the third convergent may be formed by multiplying the numerator of the second convergent by the third quotient, and adding the numerator of the first con- vergent ; also that the denominator may be formed in a similar manner. Suppose that the successive convergents are formed, in a similar way ; let the numerators be denoted by p,, p,, p,,-.., and the denominators by 9,, 9s Ya»-+- Assume that the law of formation holds for the n convergent: that is, suppose P= Fy Put + Pumas In = On Innr* Inv? : a WW276 HIGHER ALGEBRA. The (n+ 1)" convergent differs from the x only in having the quotient a, + in the place of a,; hence the (n+ 1)* con- a, nth vergent 1 (ats eathes — Gus (Ou Pps + Pas) + Pay Bry = Garr Pat Par py supposition. Oa Get Gey If therefore we put Parr =n: Put Paar Unt = Uno Ue + Una we see that the numerator and denominator of the (n +°1)" con- vergent follow the law which was supposed to hold in the case of the n. But the law does hold in the case of the third con- vergent, hence it holds for the fourth, and so on; therefore it holds universally. 337. It will be convenient to call a, the x" partial quotient; the complete quotient at this stage being a,+ : Guy Fag t - We shall usually denote the complete quotient at any stage by k. ‘We have seen that Pa _ %ePastPrs, Te Inn + Guns let the continued fraction be denoted by x; then a differs from Be only in taking the complete quotient & instead of the partial quotient a,; thus a PPa t Pane Qn. + Gans 338, If . be the n“* convergent to a continued fraction, then Pa Gai ~ Pada = (— 1). Let the continued fraction be denoted byCONTINUED FRACTIONS. 277 then Pa Gu=1~ Par Ta= (e Pat + Pat) Gant ~ Pam (4n a1 + Yas) =(Y) (Py, Gee Pa Ges) (-Y)* (Pais Ga-a Poa Yas? ia = C 1" (a— ri na But ppm —- 1% = (414,41) —4.0,=1=(-1)'5 hence Pu Qua Pa a= (— 1)". ‘When the continued fraction is /ess than unity, this result will still hold if we suppose that a,=0, and that the first convergent is zero. Note. When we are calculating the numerical value of the successive convergents, the above theorem furnishes an easy test of the accuracy of the work. Cor. 1. Each convergent is in its lowest terms ; for if p, and qg, had a common divisor it would divide p, 9,_,—?,._19,) oF unity ; which is impossible. Cor. 2. The difference between two successive convergents is a fraction whose numerator is unity ; for Pa, Pant Pa Gus ~ Por G 1 Ge Una Wn Un-r Qn In EXAMPLES, XXV. a. Calculate the successive convergents to L 8g> uy ig ae 2222313 11 " 2+ a 3+ 14+ 44+ 2+ 6° 1 1 ie 3. 34+ —— z rs + 14 2} 2+ 14+ 9 Express the following quantities as continued fractions and find the fourth convergent to each, 253 x 832 1189 729 * 179° 59° 8. 3987" 7. 338° og 8 37. 9 1-139, 10, -3029. NY 2a,278 HIGHER ALGEBRA. 12, A metre is 39°37079 inches, shew by the theory of continued fractions that 32 metres is nearly equal to 35 yards. 13, Find a series of fractions converging to ‘24226, the excess in days of the true tropical year over 365 days, 14. A kilometre is very nearly equal to *62138 miles; shew that 23 «64 5 18 the fractions 55 99° 37° 103 ratio of a kilometre to a mile, 15, Two scales of equal length are divided into 162 and 209 equal parts respectively; if their zero points be coincident shew that the 31" division of one a coincides with the 40 division of the other. are successive approximations to the nt+ni— + + A . 16, If ate is converted into a continued fraction, shew that the quotients are n—1 and +1 alternately, and find the suc- cessive convergents. 17, Shew that q) Pa+i—Pa- =Pa, Gnt1-Gn-1 In” © Ge JOR) Geyer. 18, If ori is the n convergent to a continued fraction, and a, the corresponding quotient, shew that Put2%n-2~ Pn-29n+2=In+9+ Fn41+ Unt One gt Oy. 339. Hach convergent is nearer to the continued fraction than any of the preceding convergents. Let « denote the continued fraction, and Pu Pati Pats Ge Tarr” Tere three consecutive convergents; then x differs from Pass 7 only in Gus taking the complete (n + 2)" quotient in the place of « 44g) denote this by &; thus a ee Se Pa & Pas Ge ~ Panes) _ & % (Ma +9) Ee 40) , and Patt 2 gy me Puts Gn ~ Pu ders — Lots Garr Mga +9 dann eaCONTINUED FRACTIONS. 279 . Now & is greater than unity, and q, is less than g,,,; hence on both accounts the difference between 2+) and a is less than the att ?, difference between <* and x; that is, every convergent is nearer to the continued fraction than the next preceding convergent, and therefore a fortiori than any preceding convergent. Combining the result of this article with that of Art. 335, it follows that the convergents of an odd order continually increase, but are always less than the continued fraction ; the convergents of an even order continually decrease, but are always greater than the continued fraction. 340. To find limits to the error made in taking any. convergent Sor the continued fraction. Let ee, ne 5 a be three consecutive convergents, and let ke denote the complete (n + 2) quotient; AP tP the ea tls " Mart Pan *® ~ + + ™ - a Ge Un Kner + %0) 1. (ues? Now & is greater than 1, therefore the difference between a and Po ig less than , and greater than —— % Ge Gus 4, Ines + Ge) Again, since q,,,>9,, the error in taking Po instead of x is 1 1 . less than —, and greater than 5-;—. qos aq 341. From the last article it appears that the error in aaa 1 taking Ps instead of the continued fraction is less than —-— , ns Int 1 or —_—_—_——_, ;_ that is, less than 5 hence Ye None 94 (4,61 Le + Int)” , Onde 2,,, 48, the nearer does 2 ay approximate to the continued tracksI 280 HIGHER ALGEBRA. therefore, any convergent which immediately precedes a large quotient ts a near approximation to the continued fraction. Again, since the error is less than os , it follows that in order to find a convergent which will differ from the continued fraction - by less than a given quantity 7 we have only to calculate the successive convergents up to a » Where q,* is greater than a. 342. The properties of continued fractions enable us to find two small integers whose ratio closely approximates to that of two incommensurable quantities, or to that of two quantities whose exact ratio can only be expressed by large integers. Example. Find a series of fractions approximating to 8-14159. In the process of finding the greatest common measure of 14159 and 100000, the successive amen are 7, 15, 1, 25,1, 7,4, Thus a sudoo=8+ rs apy Te Bee TE TE ‘The successive convergents are 8 92 883 855 I’ 7’ 106’ iis’? this last convergent which precedes the large quotient 25 is a very near sprrorimation the error being less than iBx eT" and therefore less than aH x 100)?” or 000004, 343. Any convergent is nearer to the continued fraction than any other fraction whose denominator is less than that of the convergent. Let « be the continued fraction, Pa Pn-1 two consecutive " amt r : ‘ 7 convergents, 3° fraction whose denominator ¢ is less than ¢,. Tf possible, let 7 be nearer to « than Pa then 7 must be 8 q 8 nearer to « than 2+-1 [Art. 339]; and since a lies between Ps and amt S Pani , it follows that : must lie between 2* and 2, tah a ayCONTINUED FRACTIONS. 281 Hence Poot that is<—!_, that is, an integer less than a fraction; which is impossible, Therefore ”* must be nearer to the continued fraction than . 344, Fy Pr a be two consecutive convergents to a continued fraction x, thon e ts greater or less than x*, according as i ts greater or less than F,. Let & be the complete quotient corresponding to the con- kp’ +p kg +q? vergent immediately succeeding E ,; then a= aT wera? (hq! + 9)* - a9 (kp' + p)* Yad — pa) (p= P'9) | ag (hy + 9)" The factor k‘p'q’ — pq is positive, since p’>p, q'>9, and k> 1; hence o > or <2%, according as pq’ —p’q is positive or negative ; ’ that is, according as ‘ >or< o Cor. It follows from the above investigation that the ex- pressions pq'—p’q, pp’ — 97x", p'- gx", q°x*—p" have the same sign. EXAMPLES. XXV. b. 1, Find limits to the error in taking = yards as equivalent to a metre, given that 4 metre is equal to 1-:0936 yards.282 HIGHER ALGEBRA. 2, Find an approximation to May be Te oF aE which differs from the true value by less than :0001. 3, Shew by the theory of continued fractions that ee differs from 7 . 1 1:41421 by a quantity less than T1830" a+ 6a?+13a+10 ; ‘i 4, Express a4 608+ l4at+ 16047 * continued fraction, and find the third convergent. 5. Shew that the difference between the first and n“ convergents is numerically equal to 1 1,1 (-19 pp N92 GAs W% In-1In 6, Shew that if a, is the quotient corresponding to ee 5 In 1 1 1 1 1 1 1) Pe sa,4—— —— —— a) Pry Ayyt Ayigt Oyigt “Ag+ Gat a,” 1 1 1 1 1 2) — =a,+ eee Se ee aaa Sioa OG tat Gat Rat Gt a A rt 1 1i1éi21 7. In the continued fraction —- —- — —...... , shew that, a+ a+ a+ at (1) Pa? +P nt1=Pa-1Pa tit PPnvar (2) Pr=Qa-1- 8 If a is the x™ convergent to the continued fraction a+ b+ at b+ at b+ ~ a shew that Gen=Pontis Yon-1=F Pon- 9, In the continued fraction shew that Puta ~(4b4+2) Pyt+Pu-2=9, Guva-(b+2) Gat qn-a=0,CONTINUED FRACTIONS. 288 . 10. Shew that 1 . a @+a5 -to 2n quotients) 11 1 ; aan tee ant 2+ sse0e-t0 2n quotients, MPR lL If>, as are the n™, (n—1)'*, (n — 2) convergents to the continued fractions 111 111 ieee Gt yt ast? Ob Ot GR SF at agt respectively, shew that M=a,P+R, N=(a,a,+1)P+a,R. 12, ute is the n™ convergent to rn 11 #21 ae kay shew that p, and g, are respectively the coefficients of «* in the expansions of 2 a arta T-ae— 2 °° Tar Hence shew that Pant where a, 8 are the roots of the equation ¢-at-1=0. 13, If 2" is then convergent to In 1 1 21 «2 ae be at Be? shew that p, and q, are respectively the coefficients of x* in the expansions of v+bat— x3 and ax + (ab+1)22— aA 1 (b+ Datpat T= (ab +2) 08+ ok” Hence shew that P= bm =a HF, +1 Batl_ (gre Paver dye OP where a, 8 are the values of x? found from the equation 1-(ab+2)a2 + A=0,CHAPTER XXVI. INDETERMINATE EQUATIONS OF THE FIRST DEGREE, 345. In Chap. X. we have shewn how to obtain the positive integral solutions of indeterminate equations with numerical co- efficients; we shall now apply the properties of continued fractions to obtain the general solution of any indeterminate equation of the first degree. 346, Any equation of the first degree involving two un- knowns a and y can be reduced to the form ax+by=+c, where a, b,c are positive integers. This equation admits of an unlimited number of solutions ; but if the conditions of the problem require x and y to be positive integers, the number of solutions may be limited. It is clear that the equation az+by=—c has no positive integral solution ; and that the equation ax — by =—c is equivalent to by — az =c; hence it will be sufficient to consider the equations ans by=c. If a and 6 have a factor m which does not divide ¢, neither of the equations ax + by=c can be satisfied by integral values of x and y; for ax by is divisible by m, whereas c is not. If a, 6, ¢ have a common factor it can be removed by division; so that we shall suppose a, b, ¢ to have no common factor, and that a and 0 are prime to each other. 347. To find the general solution in positive integers of the equation ax—by=c. Let ¢ be converted into a continued fraction, and let denote the convergent just preceding $ 5 then agq—bp ==. (Sse. 338)INDETERMINATE EQUATIONS OF THE FIRST DEGREE. 285 I. If ag—bp=1, the given equation may be written az — by =¢ (aq — bp); . @(2—- 0g) =b (y~cp). Now since a and 6 have no common factor, x—-cq must be divisible by 5 ; hence 2 — cg = bt, where ¢ is an integer, + S24 _y_Y-®. “ 6 a’ that is, w=bi+eq, y=at+ep; from which positive integral solutions may be obtained by giving to ¢ any positive integral value, or any negative integral value numerically smaller than the less of the two quantities 4, z; also ¢ may be zero; thus the number of solutions is unlimited. Il. If ag—bp=-1, we have ax — by =—¢ (aq — bp) ; a (a+ 0g) =b (y + ep) ; aay ute =t, an integer ; 6 a hence a=bt—cqg, y=at—cp; from which positive integral solutions may be obtained by giving to ¢ any positive integral value which exceeds the greater of the two quantities 4 2; thus the number of solutions is unlimited. III. If either a or 6 is unity, the fraction ; cannot be con- verted into a continued fraction with unit numerators, and the investigation fails, In these cases, however, the solutions may be written down by inspection ; thus if 6=1, the equation becomes ax—y=c; whence y=ax—c, and the solutions may be found by ascribing to a any positive integral value greater than < 0 Nors. It should be observed that the series of values for & wd y Lor éwo arithmetical progressions in which the common diferenves wie b wos respectively,286 HIGHER ALGEBRA. Ezample. Find the general solutiort in positive integers of 29x ~ 42y=65. In converting fe into a continued fraction the convergent just before 2 is & 3 we have therefore 99 x 18 ~42x9=- 1; wt. 99 x 65 — 42x 45=— 5; combining this with the given equation, we obtain 29 (2 +65) =42 (y +45); +, +65 yt45 . . 497 ="o9 75 an integer ; hence the general solution is = 42t-65, y=29t— 45. 348. Given one solution in positive integers of the equation ax — by =¢, to find the general solution. Let h, k be a solution of aw—by=c; then ah—bk=c. «. aa—by=ah— bk; a(w—h)=b (yh); — Fi a = — =4, an integer ; aw=h+bt, y=k+at; which is the general solution. 349. T'o find the general solution in positive integers of the equation ax + by=c. Let be converted into a continued fraction, and lett be the convergent just preceding ; ; then ag—bp=+ 1. I. If aqg—bp=1, we have ax + by =c (aq — bp) ; a (eq—x) =b (y + ep) ; eq—a_ ytep 60a w. @=eg—bt, y=at—cp, =t, an integer;INDETERMINATE EQUATIONS OF THE FIRST DEGREE. 287 from which positive integral solutions may be obtained by giving to ¢ positive integral values greater than # and less than ¢. Thus the number of solutions is limited, and j if there is no no integer fulfilling these conditions there is no solution, II. Ifaqg—bp=- 1, we have ax + by =—c (aq — bp); a (+ 0g) =6 (cp—y); wtey_ p-y 6 a a= bt-—cq, y=cp—at; from which positive integral solutions may be obtained by giving to ¢ positive integral values greater than 4 and less than 2 . As before, the number of solutions is limited, and there may be no solution. = t, an integer ; III. If either a or 6 is equal to unity, the solution may be found by inspection as in Art. 347. 350. Given one solution in positive inteyers of the equation ax + by =¢, to find the general solution. Let h, & be a solution of ax+ by=c; then ah+bk=c, ax + by =ah + bk; a(z—h)=b(k—y); an an integer ; e=h+bt, y=k-at; which is the general solution. 351. To find the number of solutions in positive integers of the equation ax +by=c. Let ; be converted into a continued fraction, and \et ve Wwe convergent just preceding 3 then aqg— bp =" 1.288 HIGHER ALGEBRA. I. Let ag—bp=1; then the general solution is a=cq— bt, y=at—cp. [Art. 349.] Positive integral solutions will be obtained by giving to ¢ positive integral values not greater than 4, and not less than 2, a (i) Suppose that < and i are not integers. Let Poms, gan+g, where m, 7 are positive integers and jf, g proper fractions ; then the least value ¢ can have is m+ 1, and the greatest value is n; therefore the number of solutions is -mad_ Py piga a s- n-ma= att g agtt g- Now this is an integer, and may be written 5 +a fraction, or ca ab number of solutions is the integer nearest to i greater or less a fraction, according as fis greater or less than g. Thus the according as /f or g is the greater. (ii) Suppose that i is an integer. In this case g = 0, and one value of 2 is zero. If we include this, the number of solutions is ot J, which must be an in- teger. Hence the number of solutions is the greatest integer in < +lor = , according as we include or exclude the zero solution. (iii) Suppose that Sis an integer. In this case f=0, and one value of y is zero. If we include this, the least value of t is m and the greatest is m; hence c the number of solutions is m—m-+\, o¢ G-g*\. Twos theINDETERMINATE EQUATIONS OF THE FIRST DEGREE. 289 number of solutions is the greatest integer in Sel or 5 ac- cording as we include or exclude the zero solution. ¢ é In this case f=0 and g=0, and both 2 and y have a zero value. If we include these, the least value ¢ can have is m, and the greatest ism; hence the number of solutions is n—m+1, or (iv) Suppose that < and ; are both integers. 5 +1. If we exclude the zero values the number of solutions is c woh II. If aq—tp= —1, the general solution is = bi— eg, y= cp — at, and similar results will be obtained. 352. To find the solutions in positive integers of the equa- tion az + by + cz=d, we may proceed as follows. By transposition az + by=d-—cz; from which by giving to z in succession the values 0, 1, 2, 3,...... we obtain equations of the form aa +by=c’, which may be solved as already explained. 353. If we have two simultaneous equations axt+byt+cz=d, wa+by+cz=d, by eliminating one of the unknowns, z say, we obtain an equation of the form 4x +By=C. Suppose that z=, y=g is a solution, then the general solution can be written w=f+ Bs, y=g—As, where ¢ is an integer. Substituting these values of « and y in either of the given equations, we obtain an equation of the form Fs+Gz=H, of which the general solution is 8=h+Gt, z=k- Ft say. Substituting for s, we obtain w=f+Bh+ BGt, y=g- Ah—AGty and the values of 2, y, 2 are obtained by giving to t suitebie integral values. HHA. \S290 HIGHER ALGEBRA. 354, If one solution in positive integers of the equations ax+by+ez=d, au+by+cz=d, can be found, the general solution may be obtained as follows, Let f, g, h be the particular solution ; then aft+bgtch=d, af+b'g+ch=d'. By subtraction, a (e@—f) +b (y—g) +e (2A) =0, a’ (x—f) +b (y—g) +e (z-h)=0; whence af _ y-g _ %-h t be’ — bc = ea da ~ a —a'b ~ Be where ¢ is an integer and & is the u.c.F. of the denominators be’ — be, ca! — ca, ab’—a'b, Thus the general solution is waft (bBo) 5, y=9+(ca'—ca) z, 2=h+ (ab'- 8) EXAMPLES, XXVI. Find the general solution and the least positive integral solution of 1, 7752-Tlly=1. 2 4052-519y=1. 3, 436r-393y—=5, 4. In how many ways can £1. 19s. 6d. be paid in florins and half- crowns ? 5. Find the number of solutions in positive integers of lle+15y=1081. 6, Find two fractions having 7 and 9 for their denominators, and such that their sum is 1}. 7. Find two proper fractions in their lowest terms having 12 and 8 for their denominators and such that their difference is. 2 8, A certain sum consists of # pounds y shillings, and it is half of y pounds « shillings; find the sum. Solve in positive integers : 9 Gr+7y +42=122) WO, 12x Uy442=02) 11x + 8y —62=145) * —Aedy dye 2=VIY"INDETERMINATE EQUATIONS OF THE FIRST DEGREE, 291 11, ya aaa 12 13r+112=103) dy+ 42=34J° Te by= it 13. 72+4y+192=84. 14, 2382+17y+112=130, 15. Find the general form of all positive integers which divided by 5, 7, 8 leave remainders 3, 2, 5 respectively. 16. Find the two smallest integers which divided by 3, 7, 11 leave remainders 1, 6, 5 respectively. 17, A number of three digits in the septenary scale is represented in the nonary scale by the same three digits in reverse order; if the middle digit in each case is zero, find the value of the number in the denary scale, 18, If the integers 6, a, b are in harmonic progression, find all the possible values of a and B. , 19, Two rods of equal length are divided into 250 and 243 equal parts respectively; if their ends be coincident, find the divisions which are the nearest together. 20. Three bells commenced to toll at the same time, and tolled at intervals of 23, 29, 34 seconds respectively. The second and third bells tolled 39 and 40 seconds respectively longer than the first; how many times did each bell toll if they all ceased in less than 20 minutes? 21, Find the greatest value of ¢ in order that the equation 7a+9y=c may have exactly six solutions in positive integers. 22. Find the greatest value of c in order that the equation 14”+1ly=c may have exactly five solutions in positive integers. 23, Find the limits within which c must lie in order that the equation 192+14y=c may have six solutions, zero solutions being excluded. 24, Shew that the greatest value of c in order that the equation ax+by=ec may have exactly n solutions in itive integers is (n+1)ab-—a—6, and that the least value of ¢ is (n—1)ab+a+6, zero solutions being excluded. AQ—DCHAPTER XXVII. RECURRING CONTINUED FRACTIONS. 355. We have seen in Chap. XXV. that a terminating con- tinued fraction with rational quotients can be reduced to an ordinary fraction with integral numerator and denominator, and therefore cannot be equal to a surd; but we shall prove that a quadratic surd can be expressed as an injinite continued fraction whose quotients recur. We shall first consider a numerical example. Ezample. Express ,/19 as a continued fraction, and find a series of fractions approximating to its value. 3 W19=44 (/19- A= 44 oa Yl9+4_o V19-2 5, 5 as Lee CETE M1942, /19-8 1, 2 gait 514 Figgas i948 19-8 _ 8 =8 489 =84 Jiggs! Ji9-2 4, 8 3 =1+Fi943! J194+2 19-4 1. Smee =2+Tig¢a) 19-4484 (/19 — 4) 84 vecccesssesseee after this the quotients 2, 1, 3, 1, 2, 8 recur; hence ti1daiaida W440 Te ay EF TEES It will be noticed that the quotients recur as soon as we come to a quotient which is double of the first. \n Art, 86\ we shell prove that this is always the case. {19+ mrRECURRING CONTINUED FRACTIONS, 2938 [Ezplanation. In each of the lines above we perform the same series of operations. For example, consider the second line: we first find the greatest integer in vise, 3 this is 2, and the remainder is Y19+4_ 9, that is V9 -2 + We then multiply numerator and denominator by the surd eonjognto 1 4/19 2, so that after inverting the ronal ag 4.9% begin a new line with a rational denominator. ] The first seven convergents formed as explained in Art. 836 are 4 9 18 48 61 170 1421 I’ 2’ 8’ Ii’ 14’ 39° 326° ‘The error in taking the last of theee is less than aa and is therefore 1 less than wa or 309400" and a@ fortiori less than °00001. Thus the seventh convergent gives the value to at least four places of decimals. 356. Every periodic continued fraction is equal to one of the roots of a quadratic equation of which the coefficients are rational. Let « denote the continued fraction, and y the periodic part, and suppose that 1 +! y’ 1 111 and Y= Mt— wee Seabee n+ utory where a, }, Fe k, m, n,...u, v are positive integers. Let 2 ep P be the convergents to # corresponding to the quotients }, & respectively; then since y is the complete quotient, we have 2=?2./+ P, Po y+ ge-p 3 whence y= Let ., 2 be the convergents to y corresponding to the 7 rytr quotients w, v respectively ; then y= fyte Substituting for y in terms of « and simplifying We Qotekn, 3 quadratic of which the coefficients are rations.294 HIGHER ALGEBRA. The equation s'y*+ (s—1r')y—r=0, which gives the value of y, has its roots real and of opposite signs ; if the positive value of y be substituted in a=P¥ = , on rationalising the denominator the value of x is of the form a +8 where A, B, C are integers, B being positive since the value of y is real. 7 1$1i1iéda21 Example. Express lta, 34 2334" . a8 8 surd. 1 Let x be the value of the continued fraction ; then z—- lage aa@= ire=n? whence 227+ 22-7=0, The continued fraction is equal to the positive root of this equation, and is therefore equal to V25-1, EXAMPLES. XXVIL a. Express the following surds as continued fractions, and find the sixth convergent to each: Ll V3. 2 Jb 3% V6. 4/8. 5. AJL. 6 13. 7 Ji4. & 22. 9% 2,/3. 10. 4,/2. ll, 3/5. 12 4,/10. 1 1 (J 7 1 7,4, since b,_, is a positive integer ; oo IN +a,>r1- But N-af=r7,_,3 ow JN -a, 1 . ang VV +% when it recurs; then are two consecutive complete quotients ; therefore a,+a,=7, b,, 7,7, =N—a’; but V—a,*=7,; hence r,=1. Again, a,-—4,<7,, that is <1; hence a,-a,=0, that is a,=a,. Also a,+a,=7,6,=5,; hence b,=2a,; which establishes the proposition. *362. To shew that in any period the partial quotients equi- distant from the beginning and end are equal, the last partial quotient being excluded. NV Let the last complete quotient be denoted by v = on ; then r,=1, a,=a,, 6,=2a,. We shall prove that Tray BiH 2, == 5,5RECURRING CONTINUED FRACTIONS. 299 We have Ta="%1=N-a!=N-as=r,. Also Qa + a, = Ga + a, = Tet a = " b. A and a,+a,=1,),; a,—a,,=7, (6,-2.4); a, = Gar = 6, — 6,_,=zero, or an integer. 1170-1 | Which is less than at unity ; thus a@,—a,_,= 6; hence a,_,=4,, and b,.,=6,. Similarly T=" Gs=,, 6,,=6,; and so on. *363. From the results of Arts. 361, 362, it appears that when a quadratic surd ,/ is converted into a continued fraction, it must take the following form a+e5 ToT tte eee *364, To obtain the penultimate convergents of the recurring Let n be the number of partial quotients in the recurring period ; then the penultimate convergents of the recurring periods are the n4, 2ntt, 3nt4,...... convergents; let these be denoted by Po Pm Pan . , SS, Sy. respectivel, Gn” In? Yn” a 11 #421 1 1 Now VN = 0+ 555.4 peeees 5.430407 so that the partial quotient corresponding to Pasi ig 2a,; hence ne Poss _ 20; Pat Pont Gurr 201 In + Gun The complete quotient at the same stage consists of the period300 HIGHER ALGEBRA. and is therefore equal to a, + ./; hence (6+ VN) Pat Par VN G4 IN) Qn Gos’ Clearing of fractions and equating rational and irrational parts, we obtain Oy Dy + Paa= Nas Bi Unt Yuna = Dy cer veereeeee (1). Again P ra can be obtained from a * and Pat by taking for the quotient be 9 a 1 to a? which is equal to a, +22, Thus +2) B+ Pac ¥q,+® me Pa q, a, pu (orgs Pes Im (a +2) a yen. GT) ttt Pato Ge Bu — 5 (Be 4 Te) ee (2). In like manner we may prove that if Pe is the penultimate » from (1); lon convergent in the ct recurring period, , Pex + Pont = Neus 1 Yon + Yeni = Pons and by using these equations, we may obtain Pon =, Pin ise suc- Gn” Tn? cessively. It should be noticed that equation (2) holds for all multiples of n; thus Pron 5 (2 Mdm) dam 2 Gem” (Pon the proof being similar to that already given. *365. In Art. 356, we have seen that a periodic continued fraction can be expressed as the root of a quadratic equation rith rational coefficients,RECURRING CONTINUED FRACTIONS. 301 Conversely, we might prove by the method of Art, 357 that an expression of the form os where A, B, C are positive integers, and B not a perfect square, can be converted into a recurring continued fraction. In this case the periodic part will not usually begin with the second partial quotient, nor will the last partial quotient be double the first. For further information on the subject of recurring continued fractions we refer the student to Serret’s Cours d'Algébre Supé- rieure, and to a pamphlet on The Expression of a Quadratic Surd as a Continued Fraction, by Thomas Muir, M.A., F.R.S.E. *EXAMPLES, XXVIL b. Express the following surds as continued fractions, and find the fourth ape to each: Na+. 2 Ja—a. 3. at—1, Jisi. 5. /at+e. 6. Ja-s. a 6 n 7. Prove that 1 1 1 1 ‘9a? — ——_ wat —— N9a?+3=3a+ Qa+ bat Yat bap” and find the fifth convergent. 8, Shew that 22131d21ié21 —=—_ P+it peeps iF cece BEE Ap. 9. Shew that 1 1 1 1 1 Pe (tap a+ pqayt )-em tea Pat 94,4 10. If /a?+1 be expressed as a continued fraction, shew that 2(+1)Gn=Pa-1t Pati, Pa=In-1tIn+1- Eee teagan 11. If 2= at Gt at at 1 1 1 1 I> 9a, + Ya, Ya, 2ag+ po ~ Ba, + 3a,+ 3a,+ 3a,+ °°” 4 shew that x (y?—22) +-Qy (28 — 2%) +32 (a8 — 9%) =0.302 HIGHER ALGEBRA. 12, Prove that hana Gana \= b+ a+ b+ a+ b+ a+ b+ at") 4a ei até we ate b+ 1 =b+ 1 9m °F OF at BF OF" shew that (ab?+a+b) 2—(a%+a+b)y=a?-B, 14. If a be the n** convergent to ,/a?+1, shew that In Pot pe + +P a1 _ PatiPara—PiPa G+IS+ AG a+: In41%nr9-~ 192 15. Shew that Aad \ (e+ Lid \ litte apbnar" Bx ah on") iad" J5+1 2 16. If a denote the r convergent to , Shew that It Pst Ppt---+Pr-1=Pin— Pas Ist Ipt »++ + Yan—1= Yan — Fa 17. Prove that the difference of the infinite continued fractions eles eines at b+ e+"? be ate+ "” is equal to 2" I+ 7 . 18, If JJ is converted into a continued fraction, and if n is the number of quotients in the period, shew that Ym=2PnQus Pm = Ipyrt(—1)9*1, 19, If ./W be converted into a continued fraction, and if the pen- ultimate convergents in the first, second, ...4" recurring periods be denoted by 7,, 73;...%, respectively, shew that mt+/N (aise) m%-JN \n,-JN)**CHAPTER XXVIII. INDETERMINATE EQUATIONS OF THE SECOND DEGREE. *366. The solution in positive integers of indeterminate equations of a degree higher than the first, though not of much practical importance, is interesting because of its connection with the Theory of Numbers. In the present chapter we shall confine our attention to equations oe the second degree involving two variables. *367. To shew how to obtain the positive integral values of x and y which satisfy the equation ax’ + 2hxy + by’+ 2gx + 2fy+c=0, a, b, ¢, f, g, h being integers, Solving this equation as a quadratic in 2, as in Art. 127, we have axe + hy +9 = J) WF ab) +2 (hig = af) y + (97 a0)...(1). Now in order that the values of z and y may be positive integers, the expression under the radical, which we may denote by py’ + 2qy+ 1, must be a perfect square; that is py’ + 2qy + r= 2", suppose. Solving this equation as a quadratic in y, we have py +q=e JP — prt pe; and, as before, the expression under the radical must be a perfect square ; suppose that it is ae to #; then { —pe=q'— pr, where ¢ and z are variable, and p, q, f are constants,304 HIGHER ALGEBRA. _ Unless this equation can be solved in positive integers, the equation does not admit of a positive integral solution. Wo shall return to this point in Art. 374. If a, b, & are all positive, it is clear that the number of solutions is limited, because for large values of x and y the sign of the expression on the left depends upon that of az*+ 2hay + by* [Art. 269], and thus cannot be zero for large positive integral values of x and y. Again, if h*~ab is negative, the coefficient of y* in (1) is negative, and by similar reasoning we see that the number of solutions is limited. Example. Solve in positive integers the equation 23 Any + by? — 20 — 20y =29, Solving as a quadratic in z, we have eady +14 (IT By— By But 30+24y — 2y?=102 - 2 (y-6)?; hence (y - 6)? cannot be greater than 61. By trial we find that the expression under the radical becomes a perfect square when (y—6)?=1 or 49; thus the positive integral values of y are 5, 7, 18. ‘When ya w=21 or 1; when y=7, c=25 or 5; when y=13, z= 29 or 25, *368. We have seen that the solution in positive integers of the equation as: + Zhay + by® + 2gu + 2fy +c=0 can be made to depend upon the solution of an equation of the form ait Ny? =a, where WV and a are positive integers, The equation 2*+Wy'=—a has no real roots, whilst the equation a*+Ny’=a has a limited number of solutions, which may be found by trial; we shall therefore confine our attention to equations of the form 2* — Vy? = *369. To shew that the equation x*— Ny*=1 can always be solved in positive integers. Let JN be converted into a continued fraction, and let 2 be any three consecutive convergents; suppose that 2 vy’ q areINDETERMINATE EQUATIONS OF THE SECOND DEGREE. 305 ” JN +a, is the complete quotient corresponding to 7; then r,(pq' — p'q) = Nq" -p" [Art. 358]. But r, = 1 at the end of any period [Art. 361]; Pp" -NY*=p'q-py , being the penultimate convergent of any recurring period. If the number of quotients in the period is even, P is an even convergent, and is therefore greater than ,/N, and therefore greater than os thus p’qg—pq’=1. In this case p*- Ng*=1, and therefore z =p’, y=q’ is a solution of the equation a* — Vy*=1. . Since 2 is the penultimate convergent of any recurring q period, the number of solutions is unlimited. If the number of quotients in the period is odd, the penultimate convergent in the first period is an odd convergent, but. the penultimate convergent in the second period is an even convergent. Thus integral solutions will be obtained by putting z=p', y=’, where % is the penultimate convergent in the second, fourth, sixth,...... recurring periods. Hence also in this case the number of solutions is unlimited. *370. To obtain a solution in positive inteyers of the equation x*-Ny’=-1. As in the preceding article, we have p" — Nq"= pq py. If the number of quotients in the period is odd, and if . is an odd penultimate convergent in any recurring period, < i ' and therefore p’q — pq’ =—1. In this case p?—Ng"=-—1, and integral solutions of the equation x*— Vy’=—1 will be obtained by putting e=p’, y =q’, , where 7 is the penultimate convergent in the first, third, fifth... recurring periods. HH. A, W806 HIGHER ALGEBRA. Example. Solve in positive integers z*- 18y?= +1, We can shew that 1111421 NIB“St ie iy Te ie er Here the number of quotients in the period is odd; the penultimate con- vergent in the first period is &; hence 2=18, y=5 is a solution of 2) -18y?= -1, By Art. 364, the penultimate convergent in the second recurring period is 1/18, 6 -, 649 5 (Ft 75% 18) that is, $395 hence 2=649, y=180 is a solution of z*- 13y?=1. By forming the successive penultimate convergents of the recurring periods we can obtain any number of solutions of the equations 22—18y= -1, and 2?-18y?= +1. *371. When one solution in positive integers of «*— Vy*=1 has been found, we may obtain as many as we please by the following method. Suppose that «=, y=k is a solution, A and k being positive integers; then (*— NV#*)"=1, where n is any positive integer. Thus a* — Ny* = (h? — Ne)". oe (@t+yJN) (e@-yJ/N)=(hth JN) (h—k, JN)". Putet+yJ/N=(h+k JN)", e-yJ/N =(h-k/N)'; wy Qn (h+k/N)" + (h-k JN)"; Qy JN = (h+k JN) —(h-k, JN)". The values of « and y so found are positive integers, and by ascribing to 7 the values 1, 2, 3,..., as many solutions as we please can be obtained. Similarly if «=A, y=k is a solution of the equation a? - Ny =—1, and if n is any odd positive integer, at ~ Ny? = (W? — Nie)". Thus the values of « and y are the same as already found, but n is restricted to the values 1, 3, 5,....... *372. By putting e=az’, y= ay’ the equations a*— Vy*= +a? become «*— Vy*=+1, which we have already shewn how to solve,INDETERMINATE EQUATIONS OF THE SECOND DEGREE. 307 *373, We have seen in Art. 369 that p'— Ng" =~, (pq' — pg) =*",. Hence if a is a denominator of any complete quotient which occurs in converting ,/.V into a continued fraction, and if 2 is the convergent obtained by stopping short of this complete quotient, one of the equations *- Ny’= 8mn (m2 —n%), or m3(m — 8n) + 2n?n? + 8mn3 +nt>0; which condition is satisfied if m>8n. If m=9, n=1, then 2=3362, y =2880, and the numbers are 482, 8362, 6242. The sums of these taken in pairs are 3844, 6724, 9604, which are the squares of 62, 82, 98 respectively.INDETERMINATE EQUATIONS OF THE SECOND DEGREE. 311 *BXAMPLES. XXVIII. Solve in positive integers: 1, 529 l0sy + 7y8=77. 2. 722 2ry +3y8=27, 3. y?—4ry+52*-10xe=4, 4, wy-Qr-y=8. 5. 8r+3ry—4y=14, 6. 42%-y?=3l5. Find the smallest solution in positive integers of 7. a2-14y3=1, 8. 2 19y8=1, 9, wt=4lyt—1, 10, 29-61y?+5=0, ll. 2-7y!-9=0. Find the general solution in positive integers of 12, 28-3y8=1. 13, at—by?=1, 4 f-17y¢=-1. Find the general values of « and y which make each of the following expressions a perfect square : 15, 22-ry+3y% 16, 22+ Qryt2y% 17, Batty, 18. Find two positive integers such that the square of one exceeds the square of the other by 105. 19, Find a general formula for three integers which may be taken to represent the lengths of the sides of a right-angled triangle. 20, Find a general formula to express two positive integers which are such that the result obtained by adding their product to the sum of their squares is a perfect square. 21, “There came three Dutchmen of my acquaintance to see me, being lately married; they brought their wives with them, The men’s names were Hendriek, C! and Cornelius; the women’s Geertruij, Catriin, and Anna; but I forgot the name of each man’s wife. They told me they had been at market to buy hogs; each person bought as many hogs as they gave shillings for one hog; Hendrick | bought 23 hogs more than Catriin; and Claas bought 11 more than Geertruij; likewise, each man laid out 3 guineas more than his wife. I desire to know the name of each man’s wife.” (Miscellany of Mathematical Problems, 1743.) 22, Shew that the sum of the first 7 natural numbers is a perfect square, if n is equal to & or #2—1, where & is the numerator of an odd, and # the numerator of an even convergent to ./2.CHAPTER XXIX. SUMMATION OF SERIES. 380. Examples of summation of certain series have occurred in previous chapters; it will be convenient here to give a synopsis of the methods of summation which have already been explained. (i) Arithmetical Progression, Chap. IV. (ii) Geometrical Progression, Chap. V. (iii) Series which are partly arithmetical and partly geo- metrical, Art. 60, (iv) Sums of the powers of the Natural Numbers and allied Series, Arts. 68 to 75. (v) Summation by means of Undetermined Coefficients, Art, 312. (vi) Recurring Series, Chap. XXIV. We now proceed to discuss methods of greater generality ; but in the course of the present chapter it will be seen that some of the foregoing methods may still be usefully employed. 381. If the r* term of a series can be expressed as the dif- ference of two quantities one of which is the same function of r that the other is of r—1, the sum of the series may be readily found. For let the series be denoted by Uj +Ut+U+...+%,, and its sum by S,, and suppose that any term zw, can be put in the form v,—v,_,; then 5, = (0 ~%) + (Uq—%,) + (Vg %,) + + FO. 9, VO HP) =%,- 0%,SUMMATION OF SERIES. 313 Example. Sum to x terms the series 1 1 1 (+2) 08s) + 42a) 48a) * eee eae If we denote the series by Uy ty ttt... + Uys 1/1 1 we have “=2(iGge-izae)’ a (i - =) e\lenz 1ynti.c/’ Eee 1/1 1 *. by addition, Sa=5 (i: - i) _ n ~ (142) (l+n¢1. 2)" 382. Sometimes a suitable transformation may be obtained by separating u, into partial fractions by the methods explained in Chap. XXIII. Ezample. Find the sum of — to, (1+2)(l+az) © (1+az)(1+a%z) © (1+a7x)(1+a5z) ant A (rez) (14 ar2) = Taiz * Tate PPOs +. 12 A (1+a%2) +B (1+a% 2), By putting 1+a"-'z, 1+a"z equal to zero in succession, we obtain +... to n terms, The n* term= a a AqTma BHT 1 a ae “=52a(ipe- Tee)" aa 1 a a ane “= [rallies Trai)? _ 1 a" a” maT aa a*iz Tyre )* a® re tom): BR q “ ile i “-s" “| + [m3314 HIGHER ALGEBRA, 383. To find the sum of n terme of a series each term of which is composed of r factors in arithmetical progression, the first factors of the several terma being in the same arithmetical progression. Let the series be denoted by 2, + %g+ Us t ....0. +Ug, where Uy = (a+ nb) (at+n+1.b)(atn+2.b)... (a+n+r—1.b). Replacing n by n—1, we have Up. = (a+ n—1.b) (at+nb)(a+nt1.b)...(at+n+r—2.b); o (@4+n=1. b) ty = (a+ n+ r—1 . b) Uq_1= Op, BAY. Replacing x by n+ 1 we have (a+n4r.b) Un=Unyr3 therefore, by subtraction, (7 +1) b. y= One Similarly, (9 +1) Bb. Uy =m — May (r+1)b. u,= v3 —%,, (7 +1)b.u,=%,-% By addition, (r+1)b. Sy= My — %45 that is, Sy = Crs atntr.b\u, -! Gate 1% Ys where C’ is a quantity independent of n, which may be found by ascribing to n some particular value. The above result gives us the following convenient rule : Write down the n™ term, afc the next factor at the end, divide by the number of factors thus increased and by the common differ- ence, and add a constant. i = It may be noticed that C= @+hb~-@+hs however better not to quote this result, but to obtain C as above *ndicated, u,; itisSUMMATION OF SERIES. 315 Example, Find the sum of n terms of the series 1.8.5643.5.74+5.7.9+...... The n‘* term is (2n - 1) (2n+1)(2n+8); hence by the rule 5,02) (an) (an 8) (n+) 6 To determine C, put n=1; then the series reduces to its first term, and 13.5.7 15 8 we have 165 +C; whence C=—; +, g, = (28-2) n+ 1) (2n+3)(2n+6) | 15 oo Sy ORD Ont) Ont 8) Gn 46) , 18 =n (2n3 + 8n?+7n — 2), after reduction. 384. The sum of the series in the preceding article may also be found either by the method of Undetermined Coetlicients [Art. 312] or in the following manner. We have u, =(2n— 1) (2n +1) (2n +3)=8n" + 12n*— 2n— 3; oS, = 83n? + 123n* — 23in - 3n, using the notation of Art. 70; * S,= Qn* (n+ 1)° + In (m+ 1)(2n4+1)—n(n4+1)-3n =n (2n? + 8n?+7n — 2). 385. It should be noticed that the rule given in Art. 383 is only applicable to cases in which the factors of each term form an arithmetical progression, and the first factors of the several terms are in the same arithmetical progression. Thus the sum of the series 1.3.64+2.4.64+3.5.74+...... to n terms, may be found by either of the methods suggested in the preceding article, but not directly by the rule of Art. 383. Here u,=n(n+2)(n+4)2n(nt1 +1)(n+242) =n(n+1)(n+2)4+2n (n+ 1)+n(n4+2)4+2n =n(n+ 1) (n+2)4+3n(n+1)+3n. The rule can now be applied to each term ; thus S,=4n (n+1) (n+ 2)(n+3) +n (m+ 1) (n+ 2) 4+ 3n(nt y+*q =4n(n+1) (n+4) (n+ 5), the constant being zero.316 . HIGHER ALGEBRA, 386. To find the sum of n terms of a series each term of which is composed of the reciprocal of the product of r factors in arith- metical progression, the first factors of the several terms being in the same arithmetical progression. Let the series be denoted by 2, + tq + ts +...... + Ug, where 1 qt latnb)(atn+l. b)(a+n+2.d). ~1.b). Replacing n by n- 1, ch =(a+ma1.8)(a+nb) (04m 2)...(a+nerad.6); m2 “. @tntr—-1.b)u=(a +n—1.b) Wq_1=U,_, Bay. Replacing n by n+ 1, we have (a+ 7B) Uy =Pnyr5 therefore, by subtraction, (7 = 1) b. Uy = % — agre Similarly (r—1)b.u,=0,-%. By addition, (r-1)b. 8, =%- 413 that is §,= 2a = 6 - oa o, where C is a quantity independent of n, which may be found by ascribing to some particular value. 1 1 (r-1)8' (atne1.b)... (atntr—1.b) Hence the sum may be found by the following rule : Write down the n™ term, strike off a factor from the beginning, divide by the number of factors so diminished and by the common difference, change the sign and add a constant. Thus S,=C— vw, a+rb The value of C = ——.— = —— (1b @-Ne “ ach case to determine C by ascribing to n some particular veloc, u,; but it is advisable inSUMMATION OF SERIES. 317 Example 1, Find the sum of n terms of the series 1, 2 1 i9.8.4°9.8.4.6°3.4.5.60° 1 n(n+1)(n+2)(n+8)’ hence, by the rule, we have The n‘* term is 1 a= C- gin ri) ra) +8)” 1 1 1 Put n=1, then 103.947 °-sca9 a? whence C= igs ee 1 7+ 816 ~ 8 (n+1) (n +2) (n +8)" By making n indefinitely great, we obtain S, =4 5 Example 2, Find the sum to n terms of the series 8 et + 1.2.4°3.8.6 "8.4.65" Here the rule is not directly applicable, because although 1, 2, 8, ......, the first factors of the several denominators, are in arithmetical progression, the factors of any one denominator are not. In this example we may proceed as follows: “= n+2 _ (n+ 2)? w'n(n+1)(n+8) n(n+1) (n+2) (n+8) _ _n(nt+1) +8n44 ~ n(n +1) (n+ 2) (n +38) _ 1 + 8 x 4 ™ (n+2) (n+ 3) (n +1) (n +2) (n +8) © n(n +1) (n+2)(n+3)° Each of these expressions may now be taken as the n‘* term of a series . to which the rule is applicable. : 1 8 4 n+3° 2(n+2)(n+8) 8(n+1) (n+2) (n +3)? we Spa - put n=1, then318 HIGHER ALGEBRA. 387. In cases where the methods of Arts. 383, 386 are directly applicable, instead of quoting the rules we may always effect the summation in the following way, which is sometimes called ‘thi Method of Subtraction.’ : Example. Find the sum of n terms of the series 2.545.848. 11+11.14+...... The arithmetical progression in this case is 2, 5, 8, 11, 14,. In each term of the given series introduce as a new factor the next term of the arithmetical progression; denote this series by S’, and the given series by S; then 8'=2.5.84+5.8.1148.11.144+......4(8n—1) (3n+-2) (9n+5); o. S'-2.5.8=5.8.114+8.11.14411. 14.17+.., to (n-1) terms. By subtraction, ~2.5.8=9[5.8+8.11411.14+...to(n—1) terms] - (Sn - 1) (8n-+2) (3n-+5), -2.3.8=9[9~2.5]—(3n—1) (8n+2) (8n-+5), 9S =(8n—1) (3n-+2) (8n45)-2.5.84+2.5.9, S=n (8n"+6n+1). 388. When the x‘ term-of a series is a rational integral function of m it can be expressed in a form which will enable us readily to apply the method given in Art. 383. For suppose ¢ (n) is a rational integral function of 2 of p dimensions, and assume $(n) =A + Bnt+ Cn(m+1) 4+ Dn(n+1) (n+ 2)+.....5 where A, B, C, D,...... are undetermined constants p+1 in number. This identity being true for all values of n, we may equate , the coefficients of like powers of n; we thus obtain p + 1 simple equations to determine the p + 1 constants. Example. Find the sum of n terms of the series whose general term is nd + 6n + bn, Assume nt+6n3+5ni=A + Bn Cn (n+1)+Dn(nt1) (n+2)+ En(n+1)(n+2)(n+3); it is at once obvious that 4=0, B=0, E=1; and by putting n=- 2,n=-3 successively, we obtain C=- 6, D=0. Thos n+ 6n8 + 5n2=n(n4+1) (n+2) (n+8)-Onlard)-SUMMATION OF SERIES. 319 Hence Sq=2n (n+) (n+2)(n+8)(n-+4) — In(n4.1) (a +2) =}n(n+}) (n+2) (n+ 7+). PotygonaL AnD Ficurate Numbers. 389. If in the expression n+ 4n(n—1)}, which is the sum of » terms of an arithmetical progression whose first term is 1 and common difference 6, we give to b the values 0, 1, 2, 3, ..., we get n,in(n+1), n', 3n(38n—-1),...... 5 which are the n terms of the Polygonal Numbers of the second, third, fourth, fifth,...... orders; the first order being that in which each term is unity. The polygonal numbers of the second, third, fourth, fifth,...... orders are sometimes called linear, triangular, square, pentagonal,...... 390. To find the sum of the first n terms of the r“ order of polygonal numbers. The n‘* term of the r“* order is n + 3n(m— 1) (r— 2); o S,= nt} (r—2)3(n-1)n =4n(n+1)+4(r- 2) (n-1) n(n+1) [Art. 383] =4n (n+ 1) (7-2) (n— 1) + 3}. 391, If the sum of x terms of the series be taken as the n'" term of a new series, we obtain aan Seay eee n(n+1) If again we take , Which is the sum of n terms of the last series, as the nt term of a new series, we obtain 1, 3, 6, 10, 15, ....... By proceeding in this way, we obtain a succession of series such that in any one, the n'™ term is the sum of n terms & The preceding series. The successive series thus formed exe KNOWS as Figurate Numbers of the first, second, third, ... otders-320 HIGHER ALGEBRA, 394. Zo find the n‘ term and the sum of n terms of the r= order of figurate numbers. The n‘* term of the first order is 1; the n term of the’ second order is x; the n‘" term of the third order is 3m, that is 3n(n+1); the n' term of the fourth order is (n+) » that is 1,2 aint) e+) the " term of the fifth order is 3 (nel) (2) 1.2.3? that is see erat) 3 and 50 on. Thus it is easy to see that the n‘* term of the r** order is = n+r—2 we (ntr 2) or jotr . n-1|r-1 Again, the sum of n terms of the 7 order is n(n+1) (n+2)... (nt+r—-1) ln ; which is the n't term of the (7 + 1)" order. Nore. In applying the rule of Art. 383 to find the sum of n terms of any order of figurate numbers, it will be found that the constant is always zero, 393, The properties of figurate numbers are historically interesting on account of the use made of them by Pascal in his Traité du triangle arithmétique, published in 1665. The following table exhibits the Arithmetical Triangle in its simplest form Yodo1.21271721¢293 «23 «20°22 1 2 3 4 5 6 7 8 9... 1 3 6 10 15 21 28 36... 1 4 10 20 35 56 84... 1 5 15 35 70 126... 1 6 21 56 126... 1 7 2 8&4... 1 8 36.. 1 9... ~SUMMATION OF SERIES. 321 Pascal constructed the numbers in the triangle by the follow- ing rule: Each number is the sum of that immediately above it and that tmmediately to the left of it; thus 15=5410, 28=7+21, 126=56 +70. From the mode of construction, it follows that the numbers in the successive horizontal rows, or vertical columns, are the figurate numbers of the first, second, third, ... orders. A line drawn so as to cut off an equal number of units from the top row and the left-hand column is called a base, and the bases are numbered beginning from the top left-hand corner. Thus the 6th base is a line drawn through the numbers 1, 5, 10, 10, 5, 1; and it will be observed that there are six of these num- bers, and that they are the coefficients of the terms in the ex- pansion of (1 +2)*, The properties of these numbers were discussed by Pascal with great skill: in particular he used his Arithmetical Triangle to develop the theory of Combinations, and to establish some interesting propositions in Probability. The subject is fully treated in Todhunter’s History of Probability, Chapter 11. 394, Where no ambiguity exists as to the number of terms in a series, we have used the symbol % to indicate summation ; but in some cases the following modified notation, which indicates the limits between which the summation is to be effected, will be found more convenient. Let $() be any function of x, then Se (x) denotes the sum = of the series of terms obtained from ¢ (a) by giving to « all posi- tive integral values from / to m inclusive. For instance, suppose it is required to find the sum of all the terms of the series obtained from the expression (p-1) (p-2) ... (p-7) Ir by giving to p all integral values from r+ 1 to p inclusive. HLH. A. Ww