CHAPTER V COMPLEX NUMBERS Preliminary. In order that the operation of taking the square root of a number may be always possible, we extend the number system so as to include numbers of a new class known as imaginary numbers. 1. Consider the equation x*-2x 4-5 = 0. Proceeding in the usual way, we obtain the formal a?=l%/^T, x= or solution, l2j~^l, which at present has no meaning. Suppose (i) the symbol possess the following properties to i It combines with itself and with real : numbers according to the laws of algebra. (ii) We may substitute ( -I) for t 2 , wherever this occurs. The reader can verify for himself that (at any rate so far as addition, subtraction, multiplication and division are concerned) reckoning. under these rules will not lead to results which are mutually inconsistent. Ex. We 1. Show that if x = l+2i, then x* x a = l + 4t + 4i a = l + 4i-4 = -3+4i; have .*. x 2 -2z + 5= Immediate consequences -3 + 4t-2(l+2t)+5=0. of the supposition are (i) (x + ly) + (x + iy') = (x + x (ii) (x + 1 iy)-(x' f (x (iii) 1 + itf f (x 4- + iy) (x f iy') (x f ) + iy') = (x-x') + i(y-y ). + iy') = xx' -yy' + i (xy + x'y). f f - f iy ) x'* + y'* x'* + y'* ' This reckoning is at present meaningless, and can only be justified by an extension of the number system. THE FUNDAMENTAL OPERATIONS A Complex Number 2. x+ iy, where y are x, represented by an expression of the form is The sign + does not numbers. real 53 indicate addition as hitherto understood, nor does the symbol i (at present) denote a number. These things are parts of the scheme used to express numbers of a new class, a and they signify that the pair of real single complex (x, y) are united to form number. The complex number x + number x, and we write and numbers be regarded as identical with the real lO is to iQ = x x+ in particular = tQ -f (A.) (B) Thus the system of complex numbers includes all the together with new numbers which are said to be imaginary. For shortness we write _ Q real numbers, + and + in particular tl /QX = *, (D) an abbreviation for + il, the symbol i denotes a complex number. Thus two real numbers are required to express a single complex number. This may be compared with the fact that two whole numbers are required so that as to express a fraction. In order that expressions of the form x numbers, we have to say what is + ly meant by be regarded as denoting equality and to define the may * ' fundamental operations. We 3. Definition of Equality. x+ ly = x' + iy' if say that and only if x = x' and y = y'. In applying this definition, we are said to equate real and imaginary parts. It should be noticed that the terms greater than and less than -have no significance in connection with imaginary numbers. ' 4. ' The Fundamental Operations. by the formal reckoning of Art. 1, In defining these we are guided and the equations (i)-(iv) are taken as defining addition, subtraction, multiplication 5. ' ' The sum Addition. of the and division. complex numbers x-f iy and x' + iy' is defined as (x and we write (x + iy) It follows that so that x + iy is the + ') + i(y + y') + (v' + iy') = (x+x') + i(y+y') (E) x+ iy = (x + 40) -f (0+ iy), sum of the numbers denoted by x and (P) iy. MULTIPLICATION AND DIVISION 54 Subtraction 6. is defined by the equation Hence subtraction is the inverse of addition, for x-x' + i(y-y') number to which, if x' + iy is added, the result is x + ly. The meaning of - (x + ty) is defined by the equation is the r Hence = (0 - a + i (0-y) aB (- a? + 4(-y) ................ (H) = 0, we have . ) In particular, putting ) .) -y=*(-y) 7. Multiplication is by the equation defined x'y) .................. (J) It follows that multiplication Putting #' = &, 2/' = is commutative. in the last equation, number, we see that if k is a real + iy)k = xk + iyk = k(x + iy) ......................... (K) (x we put x=x = and y = y' = l, we have 2 or i 2 =-l .......................... (L) (0 + il) =-l f if Again, 8. Division is the inverse of multiplication, that is to say, the equations x + mean and ly From the same thing. Equating real and imaginary the last equation parts, Xx'-Yy' = x mi and we have and Xy'+Yx' = y. v x *- f Therefore unless x' we have y' are both zero. Thus the equation defining division is x + iy _xx' + yy' x' + iy' excepting the case in which x' Putting x'=k, y' =0, we x'* yx'-xy' + y'* x'* + y'* ' ...................... ( ' + iy' =0. see that if A; is ,)= a real number different from zero, +i ................................. (N) ZERO PRODUCTS 55 It is easy to see that addition, subtraction, multiplication and division above) are subject to the same laws as the corresponding operations for real numbers. (as defined We have therefore justified, and attached a definite meaning of reckoning described in Art. Zero Products. 9. is zero, then where the process 1. For complex, as for one of the factors of the product 1 xx -yy' real is zero. numbers, if a product For let? Then numbers. x, y, x' , y' are real to, + i (xy + x'y) = 0. f Hence, by the rule of equality, .\ xx'-yy'~Q and 2 2 and z(*' + 2/' ) = zy' y(x'* + x'y Q; + y'*)=Q. + y' 2 is not zero, it follows that x = 0, y = 0, and therefore x + ty = 0. If x' 2 + ?/' 2 = 0, then, since x', y' are real they must both vanish, and therefore z' + 4j/' = 0. If a;' 2 10. Ex. 1. Examples. Express (1 + 24) 2 /(2 + (2+4) 2 2 in the farm = = ~4+4i-l"" X + iY. 3+4i~~ (3+40(3~4t) 24 7 -9 + 16 + 244 Ex. 2. Solve the equation x3 = 1 . The equation can be written (#-l)(a; 2 + a; + l)=0, giving # = 1 or If a;2 + a; + 1 =0, we have in the usual way x= Hence Ex. 3. the numbers Fen/y that 1, J( J( - 1 11. 2 V3) - 1 + K/3) are ca//ed JAe ^reg cw6e roofe of unity. M a roo CD, 3 o/ x = 1. we have = |{ - 1 +34^3 -3(tN/3) 2 + (^3) 3} = |( - 1+34^3 + 9-34^3) = Geometrical + a: + 1=0. - Denoting the given number by to 8 a; Representation of 1. Complex Numbers. In naming line-segments and angles the order of the letters will denote the direction of measurement. Thus AB denotes the distance travelled by a point in moving from A to B, and we write (1) AB + BA-0, and EA--AB. ARGAND DIAGRAMS 56 Again, we use L BAG must turn in order that it ' ' to denote the least angle through which may AB along AC. lie Thus L CAB = - L BACy and we shall suppose that -7T< = x + iy, we say that the number z is represented by the point whose coordinates referred to rectangular axes are x, y. (2) If z When there is no risk of confusion, this will be called the point z. A diagram showing points which represent complex numbers is called an Argand diagram. Let point be the polar coordinates of the 6) (r, then z, = x + iy = r(cos + 1 sin 5), z = rcos#, y = rsin0, tan = y/x. r = J (x 2 + y 2) z where and hence Here r, which denoted by z | is or \ It follows that if The angle 6 is is essentially positive, called the modulus of z and is by mod z. Thus modz = = z \ \ then z | z\ = 0, = r = J(x 2 + y 2 many ). y = 0, and consequently called the amplitude of angle has infinitely of 6 such that is FIG. 2. , z, and is z = 0. denoted by values differing by multiples of 2rr. am z. This The value called the principal value of the amplitude. With the convention am z is L XOz. of Unless otherwise stated, amplitude of (1), it will amz will be seen that the principal value of mean the principal value of the z. and Subtraction. z = x + ty, z' = x' + iy' Draw 12. Addition (1) Let '. parallelogram Ozsz', then s represents z For then if (X, Y) are the coordinates of the + z'. 5, X ==sum of projections of Oz, zs on OX = sum of = + X'. 05 projections of Oz, Oz' on OX FIG. 3. ADDITION AND SUBTRACTION Y = y + y', Similarly that therefore s represents z + z' = the z length Os, | | Now Os<0z + 0z', the sign Therefore i.e. if am z = am z'. To prove this algebraically, where each root (x is and + z'. am (z + ') 57 It is to be observed = L XOs. of equality occurring we have to if LXOz = L.XOz' show that to have its positive value. This will be the case + z') 2 + (y + y')*<x* + y 2 + x' 2 + y' 2 + 2>/(x2 + y2 ) (x' 2 if + j/ /2 ), i.e. if or if (x or if (xy' (2) ^en Draw the - x'j/) 2 >0, which parallelogram is the case. Oz'zd d represents z-z' (Fig. 4), Y . by the last construction, the number = represented by d + z' z. For, Observe that = the length am (z - z') = z. Jf Od. | and (3) it is = // s n Zt - 2' 2 | + z2 -f ... 4-z w , FIG. 4. wAere zl5 required to find the point s n ~O z'z, 2J 2> z n are given complex numbers, . FIG. 5. Draw in succession z^, sense as 022 Oz3 Oz4 , 2j + 22 + 23, This is , , ... , s 2s3 , 5 35 4> then s 2 , s3 , equal to, parallel to, and in the same ... are the points representing 2 1 + 2a , etc. a continued application of the first construction of this article. PRODUCTS AND QUOTIENTS 58 The modulus of the sum of any number of complex numbers (4) or equal to the sum For in the Ozl9 21$2> S2S29 of the moduli of the numbers. the moduli of z v z 2 last figure, > an(^ ^at Osn Therefore | f sn < Oz *n The sign of equality is the same straight line and Os n z2 +| to be taken l than eng^s Also . + 2^2 + s2s3 + l |<K| XOTE. are in 1s ^e are 2s > is less | . . . 4-... v + sn _ r +| * n and only |. tS che points O, z l9 * 2 occur in this order, that is, if z l9 z t ... z n have the if, if, , same amplitude. To prove this algebraically, | and sn | ^ | sn ^ | + we have by \z n I < I (1) of this article, s n _2 z n _i -h I 1 -h I \z n I , so on. 13. (1) Products and Quotients. Let z = x + iy =r(cos^H-tsin0), z' By = x' + iy' =r'(cos 0' + 1 sin 0'). the definition of multiplication, == rr'{cos cos 0' - sin sin & + t(cos sin 0' + cos 0' sin 0)} ; ^^rr'tcos^ + ^-ftsin^ + fl')} ........................................ (A) /. Since division is the inverse of multiplication, 4 = ^{cos(0-0') + 'sin(0-0')} ........................ (B) For this last is the number which when multiplied by z' produces z. Hence also ) z l *=x l If with similar notation for z2 (A) , + iy^ = fj (cos 23 , ... zn , l -h i .............. (C) sin 0j), by continued application of equation we have ^ 2 ---n = ^2---M COS (^l^^+---+^n) In (D) put Z1 ...(D) =z2 ==...~2 w = z = r(cos0 + tsin0), then zn where n + ^sin(01 4-02 +...+0n )}. rn (cos n0 + 1 sin n0), ........................... (E) any positive integer. Similarly, from (C) it follows that is ) ........................... (P) DE MOIVRE'S THEOREM Show Ex. 1. We have few 1 - that is nearly equal to . + i)=*/26(cos0 + i sin 0), where tan = 1/5, 4= 676(cos 40 + 1 sin 40), by (E). (5 + i) 2 = 4 = 476 + 480* hence we have (5 + 1) (24 + 10t) (5 and therefore But ; sin 40 = cos 40 = 476/676, tan 40 = 1, nearly. and 480/676, 40=rr/4 approximately. .*. Modulus and Amplitude. From (D) we the complex numbers zl9 z2 ... z w (2) of 59 |Z|=V2 ...rw = KN*2 |...|*n|, axnZ and but note that the --- last if Z the product is , , , see that ..................... (0) + amz n (H) equation does not give the principal value of unless am Z -7T<01 +02 +. ..+0n < 77. Again, from (B) we have M am = and am 0' z 21 .(I) am z' (J) , although the last equation does not give the principal value of amz/s' unless -7r<0-0'<7T. important to notice that the amplitude of z/z' through which Oz' must be turned in order that it may and that, with the convention of Art. 11, (1), It is is the angle lie along Oz; ft the principal value of Again, if z, a, a' are am - = L z'Oz. any numbers, z-a am and (3) the length az the length a'z : z-a = L aza. z-a , . De Moivre's Theorem. FIQ. 6. From (E) and (F) it follows that if n is a positive or negative integer (cos This, with + i sin 0) n an extension to be given = cos n9 + 1 sin n0. later (Art. 16), is known as De Moivre's Theorem. B.C.A. COMPLEX ROOTS OF EQUATIONS 60 14. Conjugate Numbers. z=x + iy and // (1) = x-iy z' 9 then are called conjugate numbers. z, z' Their product is a? 2 + y2 and polar coordinates of the point , point z' are r - 0. , 6 are the if r, z, those of the Hence O am(x + ty) = = -am(x-iy). (2) a polynomial in is // /(z) with z real = X + 1 F, where coefficients, and f(z) =f(x + iy) X, Y are real, then f(x - iy) = X - 4 Y. We have /(# 4- iy) = a + a t (x + iy) + a 2 (x + iy) 2 -f FIG. . . . where a 7. ax , , are . . . real. We can therefore obtain expansions of the form f(x + iy) = 2axm (iy) n f(x ) - ty) = Zaxm - iy) n ( , and m, n are positive integers or zero. of f(x + iy) in which n is even or zero is real, and occurs with the same sign in/(x - iy). Every term of f(x + iy) in which n is odd is imaginary, and occurs with where every a Every term is real the opposite sign inf(x-ty). Hence the result follows. (3) // an equation with real coefficients has complex = 0, where a, roots, then these occur in conjugate pairs. Let a + tjS be a root of f(x) the polynomial /(x), are Let /(a + ij8) A = 0, B = 0. j8, as well as the coefficients of real. = A + iB, where A B 9 are real. Since /(a + 1/?) == 0, we have Hence Therefore a - ijS is also a root of f(x) = 0. EXERCISE VII COMPLEX NUMBERS 1. Find the modulus and amplitude of 2. The condition that 3. Find 4. Express in the form JC 2 so that 2(3 (i) 2~+3i (a + t6)/(a' -f f i4) 4- ib') s/3 may =2 -f 3*. + 1 7, iJ + 1, V3 - 1, - <s/3 -f *, - \/3 ~ (iii) be real is ofe'-a'6=0. . COMPLEX NUMBEEb x 4- iy> express If 2 5. form in the X + iY Z 6. Find the moduli of 7. What (i) If (ii) If 8. Let 9. | z (i) if (ii) if (2 + 3i) 2 (3-40 3 2 2 | | = = | | z' | z' = 1 and am z = 2 x + ty where differ by 2|z-l| = |z-2|, then 3 (z -f y ) = 4z 22- 1 = 2-2 then # 2 4- y 2 = 1 | a z describes Prove that 12. 22 cos a 4-2* |<1 <2 2 cos a Ifz=x + iy and Z = Jf + 1 7, show that | 22 cos a 4- 2 1 / | | 4- 1 Given that 1 + 24 is find the other three roots. From cos w0 = cosn sin if ___ one root of the equation, x 4 - 3a;3 -h 8x 2 -7a? + 5=0, .^ +1 the formula cos nO that deduce integer, 14. 2 | .. 13. . < 2r 4- r < 1 2 z then ,,, , 2 | |2|<O41. if z =r, then | ' circle. 11. 2 -C\ cos n0=(7? cos*- 1 1 sin n^ = (cos 9 + n ~ 2 ^ sin 2 sin sin 0) n , cos n + <7 -C?" cosn ~ 3 1 tf ~4 where ^ sin 4 + ... sin 3 ^ - a positive ri is ... , , ( tannf. 15. By putting 2=r (cos = -f i sin 0) in the identity prove that 1 -f t coM+r 2 cos 26 + 2 r sin 04- r sin 26 ? ; Prove that |24-2'| 2 4-|2-2'| 2 =2|2| 2 4-2|2 | 6 . |, | 50) . then z'= -z. TT, sin . 2 2 | =- 1 4- + 4t) 3 Prove that y are variables. x, z' 10. | (2-f 3t) /(3 (ii) ; - am 2', then am 2, am z' and | Thus, in either case, the point [Let * 2~ 1 9 the principal value of the amplitude of (cos 50 is | (i) 61 ... +rn ~ l cos (n - 1)0 I-rco&O - rn cos n6 -f rn+1 + ... -t-r - 1 sin 11 (rc ~ 1)0 cos ' rsin0-rn 8inn0-f rn " " t 1 sin (n~ (TI - 1 ) ROOTS OF COMPLEX NUMBERS 62 = cos 16. If z and n sin 4- 1 zn is + z -n = 2 a positive integer, prove that - z~n = 2i zn cos nd, sin n0.. Hence show that ~ 2n 1 cosn 6 = cos n0 + C%coa(n -2)6 + 0% cos(?i-4)04- and if n is and if n - 2"- 1 sinn 1)2 odd n-l n-l (-1) 2 2 = cos ?i0 -Of cos (n -2)0 + 0% cos (n -4)0 - Binn 0==8mn0-C%8w(n By similarly for (*-ar n l ) -l the method of Ex. 16, prove that 16 cos 8 18. If cos a + cos 0= - cos sin* 50 - cos 30 4- 2 cos 0. + cos y=0 and sin a 4- sin + sin y=0, prove cos 3 + cos 3jS 4- cos 3y = 3 cos (a 4- ]8 4- y) /? that 8in34-sin3)54-8in3y = 3sin(aH-)S4-y). and [Put a = cosa4-t sin and a 8 4- 6 8 + c3 = 3a6c, (19) . is ...; 17. , even n ( ... From a, b = cos j3 4- 1 sin c j5, = cos y4-t sin y, then etc.] the identity (x-b)(x-c) t (x-c)(x-a) " (a-6)(a~c)" t (ft-cXft-a)"*" (x-a)(x-b) (c-a)(c-6) deduce the identities - sin (0 sin (a B) sin (0 j5) - sin (a y) y) [Put x = cos 20 4- i sin 20, a = cos 2a 4- 1 sin 2a, Roots of Complex Numbers. 15. any number whose nth power is r>0 and root of r. is 4- 1 n is a positive integer, called afc n-th root of is sin 0), Consider the values of s any integer or zero. un = r {cos (0 from which it follows that z. Let ^r denote the positive arithmetical nth -7r<0<7r. > where & (1) If equal to z z = r (cos Let where etc.] u By -I- is + L sin n Art. 13 2for) + 1 sin I n ) we have (0 4- Zkn)} an nth root of z. z ; PRINCIPAL VALUES 63 and only if, the corresponding That this may be the case, the angles must differ by a multiple of 2ir and the corresponding values of k must differ by a multiple of n. Therefore u has n distinct values, namely Again, two values of u will be equal angles have the same sine and cosine. if, y those given Thus by & = 0, 1,2, the n-th roots cos- ofz = r(cos 6 + - n lvalue of (2) , n . cos- _ 1, 2, ..., , where &=A 0, . 0\ i sin ~i is , n-1. , called the principal ^z, To find the points representing the find the length represented Find 1. points values of the representing where %Jz, we must be able to V# into Q lt Q2 the n values of by $r and to divide the angle n equal parts. Ex. J /n + | n-1. sin 6) are the values of \ + 1 sin v^. v ,,^ v ^,,, i ..., Q^ , /z, z Here and Oz=r=\/5 + 3~= ^Oz =tan 8 ^= tan Also fyr= */2* an(* so Ci ^2 ^3 are and points on the circle with centre radius */2, such that 8. General Form of De Moivre's Theorem. If x w one o/ the values of (cos 6 + isin 6) x cos x0 + M*n 16. then FIG. ^ The case in is rational, . which 2 is a positive or negative integer has been considered in Art. 13. Let x~p/q where p, q are integers and q (cos 0/q is positive, then since + 1 sin 0/q) q = cosO + t sin 0, therefore i cos 0/q also, since j? is an + 1 sin 0fq is a value of (cos + t sin 0)* ; integer, (cos 0/q + 1 sin 0/q) p = cos pO/q + 1 sin j?0/# therefore ; ,, cos p0/q 4- 1 sin jo0/j is a value of (cos + 1 sin 0) . ROOTS AND FACTORS 64 The n-th Roots 17. +1 then since cos of Unity. = 1, sin 2k-rr r-1 and 0=0, let the n-th roots of unity are the values of 2kn -hsm~ n n cos In Art. 15, where #7 = 0, . , n 1, 2, ... 1 1. The principal n-th root, given by A = 0, is 1. 1 is given by k If w is even, the root \n. ~2 -1 n n = 1 (x 1 Since x +x + + 1 ), it follows that (x 71 . . . ) other than of x"~- 1, 71 The - +1)77 (2* VL- . sm -ft .n - ( If n 19. (1) n-th root of odd, the root is root, (2A These are the values of 1 ). + 1)7T T where k = n Q, n , This follows from Art. 15 by putting r = The principal any is +an ~ 2 +...+<* + 1=0. n-th Roots of cos a 1=0, then a -1 18. if ( - ( - 1) is 1) is cos Factors of x n -1 and x n + O 1, 2, ... 1 n~l. 6 = 77. l, to k = 0. n corresponding & = -|(n - 1). -+ n given by 1 1 sin , 1. n The n factors ofx -l are x - {cos 2r7r/n + 1 sin 2rrr/n} where = 0, r 1, 2, ... n-1. This follows from Art. 15. are given by r = 0, r = n/2. The remaining (n - 2) factors can be grouped in pairs as follows Since 2r7r/n + 2(n-r)7r/n = 27r, the factors corresponding to r and (i) If n is even, the factors x-1 J.nd x +1 : n - r are x- f { cos 2rn -- [ And n f- sm 2f7rl . * the product of these j x- f { I cos 2rn -n 1 -- 2f7T - cos 2f7T\ = x^ - 2x cos 2f7T + sm^2 n n / n . a; . and sm 2r<jr}/ n . ] is 2 / \ n '/ ) h 1 ; therefore n / where 77 denotes the product of the factors as indicated. - 1 is = (ii) // n is oddy the factor x given by r 0. The remaining factors can be grouped in pairs, and (A) n-1 (B) CUBE ROOTS OF UNITY x-{cos // n (i) 15 (2r //" n + l)77/n l = + ism +1 (2r are + l)7r/n}, where 'r = 0, 1, 2, ...(n-1). even x + (ii) factors of xn n the (2) Similarly, 65 is n^ n - 2) C (x*-2x \ os&-^ + l) ............. n (0) / ocW, (D) 20. The Imaginary Cube Roots of Unity. of x 2 4- x -f 1 = 0. Their actual values are cos .277 277 i sin j Denoting either of these by Moreover, l when Again, except cube roots of 1, and o co, ( +w+ co 2 = /Q -1, the other . ^3). for (a> 2 ) 3 o> 2 , is = (6t> 3 2 = l. ) .................................. (A) r is a multiple of 3, co r and o> 2r are the imaginary so l-f-o/4- The following -, or These are the roots o> 2r = .................................. (B) identities are important. x*-y* = (x-y)(x-o)y)(x-aj 2 y) ....................... (C) + y* = (x + y)(x + ajy)(x + a> y) ....................... (D) x2 + y2 + z 2 - yz - zx - xy = (x + coy 4- o> 2 z) (x + o> 2 y 4- ojz) ......... (E) 2 z? x 3 + y3 + z3 - Sxyz = Ex. Prove *Aa* 1. (x* (x -f- y + z) (a; + coy + +y* + z* -3xyz)* = X*+Y* + Z* X=x z +2yz, 2 co z) (x + co -3XYZ, 2 y -h coz) ....... (F) where Y= This follows from (F) on observing that and Ex. If (l+x) 2. n =c Q + c 1x + ctf(;* + ...+cnxn and o/ the series being continued as far as possible, show that the values of are |(2 \ n + 2 cos ~ Putting * 1, o>, o> 2 ) , where r = n, n - 2, n + 2, Slt #2 , 3 respectively. / for #, in succession, (l-fl) n -c + Cl -f...-fc r 4-...-fcn ..................................... (A) n ...-hc n a> , 4-...+c n cu ........................... (B) 2n .......................... (C) COMPLEX ROOTS 66 r 3, 1-f co not a multiple of If r is we have Now -i 1 hence, adding (A), (B), (C) together, n 2 n +o>) + (l+o> ) n +c3 +c 6 4-..0=2 + (l 3(c XT + o>2r =0; +co = l i 2rr.27r . 7r/7r. TT\ + cos -^- + 1 sin -5- =2 cos- cos^+tsmO /), O O O O rt \ and, since AI i Also 1 cos ~ =i, we have 01l 4-cos 27T -i sin +aj 2 = . O Again, multiply (B) by co a , 2-7T , , O by (C) 1 -f o> = cos - + 1 sin - ,, and +o> 2 ) n ,., (1 cos and add to (A) co, Ttrr o ; . -tsui Al a> 277 mn / (l+eo) =( cos -^--4 2 /i . sm 27T\/ (n-2)ir - ^ If . tJ then ; n 2 n n 2 8(^+04+07 + ...)=2 +co (l4-a>) +o)(l+a> ) Also nrr TlTT cos- - W7T\ +tsm-~ . 1 (n-2)ir -^~ . ^ - , and 3 Finally, multiply (B) by o>, (C) by o> 2 , and add EXERCISE to (A) ; it will be found that VIII COMPLEX ROOTS ^-1 1. Prove that the values of 2. Find the values of ^(1 + 1) and *J(l-i). 3. Use Art. 19 to are ~^(1 v^ t). show that 1), ) 4. Prove that =a; 2 (a; where a, j3 are the roots of z* + z- 1=0.] ' (#- 2x cos^-H l). COMPLEX ROOTS 67 Give a geometrical construction to find the points 5. the values of z l9 z 2 corresponding to *Jz. [If 01 is the unit of length, Oz l bisects the angle XOz, and is a mean proportional to 01 and Oz ; and 2 2 is on z-f) produced, so that Oz a =2 1 O.] If a, b are complex numbers, show that 6. a [Let 2 1 =a + */a I *i I a-*Ja*-b 2 then by Exercise VII, Ex. 10, 2 + *$J 2 =i *!+** 2 + i *i -*2 2 =2 a )t + 2 a*-6> -6 2 , z2 , I ! I I I I . | | | a -6 | -f | a -6 ; 2 | } = {|a + 6| + |a-6|}M Solve the equation 7. f + 2(l+2i)z -(11 + 20=0. - (11 + 20. Verify that the sum of the roots is -2(1 + 20 and the product [Put z~x + iy, equate real and imaginary parts to zero and solve for x, y.] Prove that, with regard to the quadratic 8. z2 (i) if (ii) if + (p + ip')z the equation has one real root, then the equation has two equal roots, then and p*-p'*=:4q If a [(i) Eliminate In (ii) is a a. this case (p + ip')*=4:(q + i,q'), If z ~x + ly =r (cos 6 + 9. rule of equality a 2 by the real root, 1 etc.] sin 0), prove that Jz=z--={'Jr + x + i*/r-x} according as y Also them if is y>Q - -f l sin J then 0<^<?r, and and if [The roots are -J( lie on the line l +i cot x + i=0.] cos 6 = - y<0 we the roots of zn = (z + l) n , are collinear. points {Vr -f x or - i\lr - x} 9 positive or negative. */r ( cos [\/z pp'ty'* J, ; have -7r<8<0 9 etc.] and show that the points which represent where r=0, 1, 2, ... n-1. The corresponding COMPLEX FACTORS 68 r = 0, that the roots of (1-f 2) n = (l -z) n are the values of ... n - \ but omitting n/2 if n is even. Show 1, 2, 12. i tan-^, where n , Prove that * 2W -2xn cos (i) 194- 1 *n + ar* - 2 cos (ii) cos (iii) rz< = 1 2 /Trio" (* - 2x cos = /T^I JJ - cos nS = 2n - 1 77^o "* (a? ~* x m - 2#n cos 6 + 1 ~~ {xn - (cos 4 t sin Art. 15. (iii) Put x =cos <f> + 1 sin ^, and for n is odd and not a multiple of + l) n -xn -l. - 1, co, where [Put x=Q, cos put 3, cos - cos < n- 0)} {z [(i) 13. If + ar * - 2 + - (cos 1 ~ sin 0)}. Now use n0.] prove that x(x + l)(x* + x + 1) is a factor of (x 14. If ( 1 + x + x2 ) n=a -f a; is an imaginary cube root of unity.] a^ -f a 2 + a: 2 . . . -f 2 n^ 2n > prove that aQ + a 3 + a 6 + u 15. If v=x+y + z + a(z w=x + y + z + a(x + y-2z), 2 3 3 3 Sa^z) prove thUt 21 a (x 4- y -f z (a where a> = w 3 -f v 3 - Suvw. , r 4- 2 2 2 4- z 3 2 [For o>=J( then where 3 and w are the imaginary cube roots of unity, prove that [IT] If 2^ real, 1^ + w')- 1 -f (6 -f co')- 1 + (c + co')" 1 + (d + w')- 1 - 2ft/- 1 1 [Consider the equation (a -f a)"" -f (6 18. If -f- - 2; 2 2;3 1 4-t\/3), *- ^^s = 0, l -f (c 4- z)- 1 4- (d 4 x)- 1 = 2ar 1 .] /. ZB - zt = being an imaginary cube root of unity, and + oA+ <o c)= 2 V(a 4- co 6 + we) = </( x)~ - Zfa - 2^2 = 0, prove that x 4- a>2:2 - 4- 2 a, 6, c are J{> D = s/( provided that 6>c. If 6<c, the sign of A must be changed. [Use Ex. 9. For a positive number x 9 >Jx denotes the positive square root.] PRODUCTS AND QUOTIENTS 21. Points representing the Product Let z = r(cos0 + *sin0), Given Numbers. Construction for (1) the point 69 and Quotient of z' = r' (cos 0' Two + 1 sin 0'). representing the product zz'. be the point on OX which represents Draw the triangle Oz'P unity, so that 01 =1. directly similar to the triangle Olz. 1 L^t P Then represents the product zz' z . For by similar triangles OP Oz -, OP r IB, Lz'OP^ LlOz = 0, also But zz' Therefore P = rr f {cos (0 represents Construction for (2) ..OP = rr r-, . that i FIG. 9. -f i ; LlOP = :. + 0') , sin (0 + 0')}. zz'. the point representing the quotient z/z'. Draw the triangle the triangle Oz'l Then Q directly similar to . i represents the quotient by the For, OzQ FIG. 10. z/z'. last construction, (number represented by Q) // k (3) is constant, z-a z-a and .z' z. z varies so that "k, then the point z describes a circle of which a, a! are inverse points; unless k~I, in which case z describes the perpendicular bisector of For at d, let d f . aa f . FIG. 11. the bisectors of L aza' meet aa Then, by Art. 13, Therefore ri, (2), az : d' divide aa' internally r a'z =k : 1. and externally and are fixed points. Therefore z describes the circle on dd' as diameter. > in the ratio k : 1, DISPLACEMENTS AND VECTORS 70 Also f , we have ca.ca' = cd2 ; hence are inverse points with regard to a, a' the the mid-point of dd c is if circle. If &= 1, then az = a'z\ and on z lies the perpendicular bisector of aa'. if Conversely, which circle of the point z describes a a, a' are inverse points, Z ~Q> then z-a For, since ca . ca' the same ratio, k to Hence, : =cd 2 then , where <f> on is dd' is divided internally and externally in say. 1, a'z = k : 1 .* If z varies so that (4) circle az FIG. 11. K. , am z-a a constant angle, then the point aa', containing an angle z describes an arc of a segment of a <f>. \a FIG. 12. FIG. 13. For by Art. 13, La'za = on which the segment lies. <f>. The Thus sign of <f> is <f> determines the side of aa' positive in Fig. 12 and negative in Fig. 13. 22. Displacements and Vectors. In connection with the geometrical representation of complex numbers, we introduce the notions of displacement and directed length or (1) vector. Displacements in a given Plane. Let P, Q be two points in the plane OXY. The change of position which a point undergoes in moving from P to Q is called the displacement PQ. (2) * See Elements of Geometry, Barnard and Child, p. 316 ' : Circle of Apollonius.' ADDITION OP VECTORS If any sense as 71 drawn equal to, parallel PQ, the displacements PQ, P'Q' are straight line P'Q' is and to, in the same said to be equal. To specify completely a displacement we must know PQ : (i) its magnitude, (ii) its direction (iii) its i.e. denoted by the sense, we draw OL equal and that L ; ; the letters, and arrow. If PQ the length if order of - necessary by an parallel to XOL is the angle which PQ PQ and FIG. 14. in the same sense, we say makes with OX. This angle determines the direction and sense of An (3) Vectors. with reference to line PQ) used to denote a line-segment and sense, the actual position of the expression (such as length, direction its being indifferent, is called a vector. Quantities which can be represented by lines used in this way are called Velocities and accelerations are vector quantities. vector quantities. A force can be represented action of the force. by a vector ' localised ' to lie in the line of Quantities (such as mass) which do not involve the idea of direction are called scalar. Connection with Complex Number. (4) If z=x + t,y and P the point is a one-to-one correspondence exists between the number z of the following (ii) the displacement OP (i) the point P (x, y), : and any ; ; (iii) the vector (or directed length) OP. Any to be one of these three things represented by may therefore be said to represent %* z. 23. Addition of Displacements and of Vectors. Q R Let P, be any three points. If a point moves from P to Q and then from Q to R, the resulting change of position is the (1) same as to R. if We 9 it had moved therefore displacements as follows * Some from P addition of directly define the Flo : writers use an underline instead of an overline. 15 z, or ZERO AND NEGATIVE VECTORS 72 The result of adding QR expressed by writing This equation PQ, QR, If to PQ is defined to be PR\ and this is PQ+QR=PR .................................. (A) also taken as defining the addition of vectors. is RS are any three displacements or vectors (Fig. 17), Q FIG. FIG. 16. 17. The Commutative and Associative Laws hold for placements and vectors. (2) In Fig. 16, complete the parallelogram (i) and PQRS\ the addition of dis- then by Arts. 23, (1), 22, Hence the commutative law (ii) holds, and PR is called the sum of PQ and QR. In Fig. 17, therefore and the NOTE. associative law holds. In Fig. 16, PQ + PS = PR t a fact which is expressed by saying that dis- placements and vectors are added by the parallelogram law. 24. Zero and Negative Displacements and Vectors. If after two or more displacements the moving point returns to its initial position, we say that the resulting displacement is zero. Thus we write This equation the meaning of is also written in the form -PQ = QP, which defines - PQ. These equations are also taken as defining the meaning of zero and negative vectors. DISTRIBUTIVE LAW 73 For displacements and vectors the meaning of 25. Subtraction. PQ-QR is defined by PQ-QR - PQ + (-QR) = PQ + RQThus OP-OQ -OP + QO = QO + OP = QP. 26. Definition of Multiplication by a Real Number. To multiply a displacement or a vector PQ by a positive number k is to multiply The resulting displacement its length or vector by is A, its direction unaltered. denoted by JcPQ or by we Further, leaving define PQ . k. - k)PQ by the equation ( In particular, ( So that to multiply a vector by 27. The Distributive number then ( - 1) is to Law. turn We it through two right angles. shall prove that if k is a real , Let k be positive. Along PQ set off PQ' = kPQ. in R'. By similar triangles, to QR to meet (i) Draw PR and and Q'R' Q'R' parallel = kQR\ kPQ + kQR^PQ' Hence the theorem holds for positive -k), we have numbers. (ii) For a negative number and ( IG * ( hence the theorem holds for negative numbers. Thus the distributive vectors by real law holds for the multiplication of displacements and numbers. Fig. 18 is drawn for a value of k greater than unity : the student should see that the same result follows from a diagram in which k is less than NOTE. unity. The diagram of VECTORS AND COMPLEX NUMBERS 74 Complex Numbers represented by Vectors. 28. It will now be seen that, so far as addition, subtraction and multiplication by real numbers are concerned, complex numbers are subject to the same laws as the vectors which represent them. This fact is fundamental in theory and very useful in practice. It should be noticed that if a number AB and am z the length makes with the directed line Ox. then | z | is Theorem 1. If C divides is AB in z is represented by a vector AB, the angle which the directed line the ratio n : m and is AB any point, then (m + n)OC - mOA + nOB. For mOC^mOA+mAC, Also mAC = nCB = - nBC whence the Theorem in the ratio result follows 2. If n:m, by ; O addition. FIG. 19. z z point which divides the straight line joining v 2 then the corresponding numbers are connected by the z is the relation This follows from Theorem 1, in accordance with the principle stated in this article. In particular, if z Ex. is the mid-point of then z = %(z l + z 2 ). that // a, b are complex numbers, prove geometrically 1. A B be the points which represent Bisect AB at C, then Let z^ 9 a, 6. OA+OB=20C OA-OB^'BA^Z'CA. and Therefore a + 6 and a -b are represented by 200"and2CJ; Now, since C hence is mid-point of base A B, SYMBOLS OF OPERATION Ex. 2. // OA, OB, OC are connected by the relation = 0, ? Men A, B, C [This a -f b where -f c = 0, are collinear. the converse of is 75 Theorem We of this article. 1 + c)(W==a.OA+c.OC'; hence (a The Symbol a have AB^c . . BC.] as an Operator.* Along two straight lines at consecutively, equal lengths OP, OQ, OP', OQ' in the 29. i right angles set off, positive direction of rotation. Let the symbol i applied through a it of turning operation a vector denote the to angle right in the positive direction of rotation. To bring our language into conformity with that of algebra, we say that to multiply a vector by is i turn to it through a right angle in the FIG. 21. positive sense. Thus in Fig. 21 OQ=i OP , OP' = Therefore where i*OP Thus i 2 an abbreviation is 1(1 OP) - 1 2 OP, Hence for i(iOP). t and -1 denote the same operation, and OQ' = OQ' = iOP' = i(-l)OP and Again, Either of these results to OP' = i OQ. and multiply a vector by (- is i) turn it OP - ( ( - 1)OP. in this sense -OQ = written in the form is to 2 - 1) ( we write -l)iOP. OP^OQ', so that through a right angle in the negative sense. Again, 3 . i 2 OP is taken to mean t 3 (i 2 OP), it is obvious that b are complex numbers, find numbers so that the points z, z' and a, 6 may be opposite Ex. z, z' if i 1 . //a, corners of a square. Let c be the mid-point of a6, then Oc + icb Similarly, * 2'=!- (a ; + &)+* (a- 6). For the moment, the reader should forget FIG. 22. his conception of I as denoting a number. B.C.A. PRODUCT OF COMPLEX NUMBERS 76 30. The Operator cos0-MSin0. Draw two equal straight lines OP, OP' inclined at an angle 0. Draw P'N perpendicular to OP. Along NP set off NQ equal to NP'. Then OF = ON + NP' = ON + iNQ. ON = ^ n OP - cos Also OF - cos /. . OP + * . OP, sin OP, . which we write in the form OF-(cos0+ism0)OP, and we say that to multiply a vector by cos + i sin is to turn it through the angle 0. and Division of a Vector by a Complex 31. Multiplication Number. say that to to multiply Here In accordance with Arts. 26, 29 and 30 of this chapter, multiply a vector OP by the complex number r (cos -h i sin 0) length by r its r is and turn applied If z, OQ we and cos -ft each other, and the is the vector obtained by multiplying OP by the complex number write also say that the ratio of Division is the turning factor. order in which they are sin indifferent. is OQ^zOP we is the resulting vector through the angle 0. the stretching factor These are independent of we OQ arid to resulting vector through the angle result is the number the inverse of multiplication, so that is the result of dividing OQ by z. Therefore to divide a vector OQ by The OP OQ/OP^zi is ( - if OQ = zOP z is to divide its length y then OP by r and tnr^ the 0). the same as that obtained 32. Product of z. by multiplying OQ by I/z. Complex Numbers. OQ^z'OP and OR^zOQ, Let then we write where zz' applied to OP denotes that the operators in succession, in this order. z', z are to be applied AEGAND DIAGRAMS 77 Since the stretching and turning factors may be applied in any order, OP may be transformed into OR by multiplying its length by rr' and turning the resulting vector through the angle (6 + 6'). Hence the operations denoted by and rr' {cos z'z zz', (0 + 0') + i sin (0 + 0')} are equivalent. Again, if we take (cos 0-h i sin 0) n to +1 (cos is to be applied n times, the result is mean that the operation sin 0) the same as that given by the operator cos n0 -f i sin 0) n = cos n0 + sin n0. In this sense then (cos + i i sin w#. It will be seen that complex numbers used as operators on vectors conform to the laws of algebra. EXERCISE IX ARGAND DIAGRAMS VECTORS : If z = 3 + 2i, z' 1 -f i, mark the points z, z' in an Argand diagram, and by geometrical construction, the points representing 1. Z + Z', Z-Z', ZZ\ find Z/z'. z, a, b be complex numbers of which a, b are constant and z varies. is given in terms of z by one of the following equations, it is required to find 2. Let If Z the point Z corresponding to a given point z. Explain the constructions indicated in the diagrams, 01 being the unit of length. FIG. 24. (ii) (iii) Z= (iv) Z=tz where Z~az+b. t is real, CENTROIDS 78 find tho point Z corresponding to a given point z t the point Z is on the y-axis. Z=(l + z)/(l-z), 3. If and show that if | = 1, numbers and are given complex b 4. If a, z | J^ 6^, find the point z corresponding to value of and prove t, -oo to + 00, z that varies i the entire describes real any given as from which line passes through a, b, the segment ah corresponding to 1 of t. to the values from [Along the ab set off the length az = t.ab, line Oz=0a+az~0a + t then ab; . :. z=a + (b-a)t, etc.] 5. If z=a(l + it) where t is a real number, prove that as tho line through the point a perpendicular to Oa. 6. If c, a are given numbers, a being z (i) if real, = c+a(cos< + isin<), show that as t varies from with centre c and radius a. is real, t once the circle 7. If A, B, C, that this theorem 8. If and O To G is G is any GO to is [Put . an immediate consequence of the identity m m w 3 ... at A A A& ... = (m + m z + m 3 + ...)OG m OA l + msOA 2 + maOA 3 4- ... point, then l9 29 . . . are not 9. If z is the centroid of particles of all of the three coplanar Moreover, same m m 29 2, l9 2 and non-parallel pOA +qOB + rOC = Q, OA, vectors where s , ... p 9 : . at z l9 z 29 z a , . . . OB OC 9 ... , then . are connected by a q 9 r are real numbers. p:q:r=& OBC A OCA A OAB : + 3 3 2 we may suppose acting at A l9 A% ... sign, 2, l9 m w m m^ + m z + w z mass ...)s relation of the form 19 , 1 l m m z, 19 to be the centre of parallel forces, proportional to Any the point z describes , = tan J<^.] t . the centroid of particles of mass include cases where 10. + oo D are any four points in a plane, then AD BC^BD AC + CD AB. . Show ; z=c + a(l + it) I (I - it), if where varies z describes and, find the point z corresponding to a given value of cf> (ii) t 9 where the signs of the area# are determined by the usual rule. Also the points A, B C are collinear if p + q + and conversely. rQ 9 [For and p 9 q, let G p:q:r= &OBC AOCA : : 9 &OAB, be the centre of parallel forces, acting at A, B, r then G coincides with O, and pOA let 9 (7, and proportional to . VECTORS AND COMPLEX NUMBERS 11. (i) of the //a, j8 are non-parallel vectors andp, q, 79 p' 9 q' are real numbers, adequation form involves the two equations If a, (ii) , p=p' q=q'. 9 y are non-parallel coplanar vectors connected by the equations = Q and p'a + q'fi r'y = 0, p<x + qfi + ry p' q', r' are real, then these are one and the same equation, that -fc- where p 9 q, r, 9 is [These theorems follow at once from Ex. 10.] 12. three complex Any form where p, q, r numbers z l9 z 2 , z 3 are connected p:q:r~ A0z 2z 3 : the signs of the areas being determined as usual. Also the points z l9 z 2 z a are collinear if p + q + r Prove this algebraically, and deduce Ex. 11. Q, , =x + iy l whence the l9 z2 =z 2 + iyl9 zs =z 3 + [In Art. 28, (i) we have B C are collinear, and is any point, OA EC + OB CA + OC AB^O. Theorem 1, m n m + n=CB AC AB.] A, . : Zj, z 2 , z a then 9 . Let it/ 3 , and conversely, results follow immediately.] 13. If the points 14. relation of the are real numbers. Moreover, [If z l by a . : : : be complex numbers, no two of which are equal, then If the points z l9 z 2 z 3 are collinear, , l \ 2|2~2l Z 2~*3\ ! *8hl-2l| =- the point z l lies between z 2 and z 3 , the ambiguous signs are both minus. (ii) If the above equation holds, then either z l9 z 2 , z 3 are collinear, or else O is the centre of a circle which touches the sides of the triangle Z 1 z 8z 3 . Also, if 15. If A i^4 2^4 a is an equilateral triangle, the vertices occurring in the positive direction of rotation, prove that s ~ 4- 1 sin an imaginary cube root of unity. numbers corresponding to A l9 3 X -f wz 2 + cj*z 3 = 0, and consequently 2 2 z^ 4- z a + z 3 - z^ - Z& - Z& = 0. where o> is Also, if z l9 z 2 , z 3 are the 16. If numbers A X YA'X'Y' a, a' 9 is a regular hexagon and then the numbers represented by where 6 has the values ~, 3 ~ 3 - . A X 9 9 A t9 A B9 prove that A' represent given complex X' Y Y' are given by 9 9 TRANSFORMATIONS 80 The 17. triads of points A B C and X, 9 9 triangles if the corresponding by the relation x(b - c) + y(c [The triangles are directly similar 18. If CA, AB, ABC Y, Z are the vertices of directly complex numbers and a, b, c - a) + z(a - b) =0. if = AC XZ i.e*. , if a triangle and triangles BCX, CA Y, similar to one another, the centroids of directly [By 1 J Ex. is x - a x, y, z are - c-a ABZ similar connected =.] z-x are drawn on BC, XYZ &ndABC coincide. = y--- = z br. Hence show that# + yv + z=a + & + c.lJ b-c c-a a-b c 17. i drawn on the sides of a given triangle ABC, all inwards. Prove that their centroids form an equilateral triangle. be the centroids of the triangles drawn outwards on BC, CA, AB. 19. Equilateral triangles are outwards or [Let P, Q, all R Prove that _ _ I QA = CA \/O (cos 30 + L sin 30), __ __ A R =AB i . *Jd 30 - (cos L QR \CB + ^ (CA -AB), and that RP has show that EQ = EP (cos 60 + sin 60), and use Ex. 15.] Hence show that Hence sin 30). a similar value. L ' TRANSFORMATIONS ' and 20. If Z, z are connected by any of the relations in Ex. 2 given curve s, then Z will describe a curve S. if z describes a Explain the following, where a, b are given complex numbers. = z -f a, S can be obtained from 8 by a translation. (i) If Z Z=tz where t is real, the curves s, S are similar and similarly situated, the centre of similitude. In this case we say that S is a magnification being (ii) of If s. (iii) If Z = (cos a-f tsin a) z, Zaz S through L a. (iv) If can be obtained from s by a rotation about -f 6, S can be obtained from s by a Z = l/z, S is the reflection in a magnification and a rotation, translation. (v) If OX of the inverse of s, O being the centre of inversion. 21. Show except that in circle C, 22. that each of the substitutions in Ex. 20 converts a circle c into a Show (v), if c passes through O, then [Z L or a circle into = ~> H--a is a straight line. that the substitution a'z converts C a more of those ? a circle or, in -; az + br? in Ex. 20.] 9 an + b' exceptional case, into a straight line. therefore the transformation is to one equivalent ^