INTRODUCTION TO TENSOR CALCULUS, RELATIVITY AND COSMOLOGY DF Lawden Solutions to ExercisesEXERCISES} fo. In the inertial frame 6» xy, the pol i ave (ry o64)" and the polar Pesslutes of ane equation of rotion argo afr - r(eb'] = f,. mle) + 2F(6ebi] = fy T rhe polar components of the particle's velocity relative to the Frame Ox“y” are (Vs Vp) = (Fr, rd). ihe soba “e C }. Hence, the abeve equations can be rearranged into mfp - rt) = f+ dmov, + mut, {re + 28) ¢, - omy, ~ mb. Since ® r . ape Be a vy = Sing, these are the equations of notion asked for. A force ¢ perpendicular to the direction of vin the sense shown gn the diagram, will have radial and transverse ‘components csing, ~Ccos?, fespectively. Taking ¢ = 2mov, this Shertial force will accordingly yield respectond vers in the right-hand wenbers Of the ‘equations of motions the tyo remaining terms in these members are supplied by a force mru? acting ajong OP and a force mre acting ‘transversely in the clockwise sense. thus, the notion of the particle relative to the non-inertial frane Ox"y’ Will be correctly predicted by Newton's second law if the three frertiay + ee corre taxen to act upon it, in addition to the ‘real! force f. fod, _ Let Tyr %e Gp 2%) de, the coordinates of the two ends of the wee ins, so that its Tength in this frane is = %, ~ %ys_ Suppose the pos~ Bere OF the two ends of the ber on the xaxis ove established to have Wiens ates ys Xp at instants which are simulteveoNs jn S with time © Then, these events have space-time coordinates (x .ty)and (Xsota}in § and Space-time coordinates (49) and Gee) nS (WB, by and t wil differs Spece~the events will not be simulzaneous in 's)." Application of the inverse Lorentz equations now Jeads to the relationships fe = 8G ND, He = BH 4 and subtraction gives the result x, - Ma 5. 6E, Thus, if X- % 18 accepted as the length £ of the bar in S Ghcorrbctly. of course)» then & 1s greaser ‘than by the factor 8No.3. z, Ze x tear te a $ x Let X be the x-coordinate of the fi it ‘ixed point wax! Suppose the front end of, the bar passes this point’ at ty and the rear ond passes the point af the Tater tiae t, » both tines being measured by the - cated at X which has been synchronized at 0,” In the S-frane, these two events have Spaces tive eaordinates (h t, and (X,t;). In the S-frame, the sane e int, ar e 5 the vents will take pT Hentai cette ve lee te ne a ey): Empiying the Lorents trans ee {o1ate the shake-Eine cooPainstes in tae to sramenue find PONS B(X- uty), 7% = BCX = ut, ). Subtraction yields ~% = Bult, t) os put. Accepting that 2, Rocepting that, fy ut, this Teads to the Fitzgerald contraction formila . This analysis confirms our intuition that the two ‘thods of ineasuring the length of a moving bar (viz. (1) by fixing the posi tions of Zits ends at simultaneous instants and (ii) by measuring the time it takes to pass a fixed point) Tead to the i en the Ci tp pass a Fixed point) Jead to the sane result, Had this, 22 8Se “wate bf these two procedures defined the length of a moving body. No.&. Although the chalk marks fixing the positions of the two ends of the bar on the x-axis are made simultaneously at time t in the S-frame, sreording to the 8 clocks, these marks are sade at different times ty. Ta. Applying the inverse Lorentz transformation t = BCE + ux/c*) to these two events, we get t= BCE + Wise) = BCE, + URp/e*). hence . Hy -% = Uli Hct + wre This means that the observer will calculate that, the mark fixing the % Teor the bar is nade e tine Ui/c* before the mark fixing the x, end. Thus, according to observations made from S, a mark B First appears on te ieaxis opposite the X, end and this then foves with, the y-axis in the negative x sense for a time uB/c* at speed u before the Hoang fart A appears opposite the X, end. AL this instant €, in S. Sree ore. B has moved 2 distance uet/c and the situation as ft appears in 3 is as below: = K z . 5 "eel _,. t & 2, i ; <9 A ero eee Ve deduce that, for S, AB = E- ublye® = (1 = ub/etyE. | The observer enploying the frame 5 15 now observed fromS to measure the, See er ee clveen the marks using a measuring rod of Tength (1 — utc!) Seams takes to be of unit length, Hence, according fo S» the pesylé > we ne ratgor this measurement will be AB/(E ~ urfer)® = T= ui fer)'E. wid Sean agreement with equation (6,3) and there is no Inconsistency,-3- A frequent source of confusion should here be remarked. If we wish to be precise, we should always clearly distinguish between a physical event and our observation of the event. Thus, a nova outburst on a star distant 100 light years from the earth may be observed in our telescopes today, but the event itself will have occurred a century ago by terrestrial time, Nevertheless, in these circumstances 1t is normal to say that, in a terrestrial reference frame, the nova outburst'was observed to occur! a century ago. Thus, when using a particular reference frame, observations will always be assumed to be made by local observers who are present at the events in question and who are equipped with standard instruments at rest in the frame, the clocks having already been synchronized with the master at the origin. Unless otherwise stated, no correction will be made for any time delay which may occur while light travels from an event to the eye of some particular observer in the frame. The direct visual impression which such an individual observer receives when he studies @ sequence of phenomena through his telescope is of no fundamental importance and does not count for our purposes as the description of the phenomena relative to his reference frane. This description must always be couched in terms of space-time coordinates and other physical quantities defined locally at the events involved. For example, in the earlier argunent, it was Stated that "for 5, AB = %- u?E/c?"; however, if the segment AB of the moving x-axis were observed through a telescope fixed at some point of S, the Tength een would need correction to allow for the different tines Tt takes light to reach the instrument from the two ends; such a correction would depend upon the position of the telescope and clearly has no funda— mental significance - it is therefore ignored. In future, therefore, a phrase of the type ‘as seen fromS ...' mist be interpreted as introducing a description of physical phenomena in terms of physical quantities defined ‘in the frame S, local to the events involved. Once this description has been calculated, the view uhich an individual observer in the frame has when he turns his telescope on to the phenomena, can then be determined if needed. No.5. Since the bar is stationary on the x-axis, the world-lines of its particles will all be parallel to the x-axis as shown in the diagram. Taking a section AB of this band of world-lines by a perpendicular to the x,-axis, we obtain a representation of the bar at some instant in the S-frame; thus AB’= 2, Similarly, taking a section AC by a perpendicular to the X-axis, a representation of the bar at some instant in the S-frame is obtained; thus AC = Z, From the diagram it follows that tseca = 21 - ut/e#)8, % This is the Fitzgerald contraction formila. No.6. Defining « by the equation tanha,= yc (a will be reall singe, ful pe SO that A's pulse is still to the right of B and has not reached B.®’thus, in no frame can ft be observed that A's missile strikes B before the latter has time to fire. If this were not the case, in some frames A could kil] B without himself being harmed and inconsistency would arise. No.15. Substituting from the transformation equations x = (itu), ys ¥, t = a(€ + ux/c*), the equation describing the wave is found to take the form J = asinenfa(] - wale - 1 ails This describes a wave having frequency T = f8(1 - u/w) and wave velocity W = (weu)/(1-uw/c?). In the special case u = w, the transformed equation is ¥ =a sin(2wtx/ou), te. a standing wave, since the frame then moves with ‘the wave. No. 16, If z is the proper time interval between the events and D is their distance apart in'S, then, by equation (7.4), vos shoe tele Thus, D> 440% + c*T#). It is helpful to note that the distance between events is dependent upon the frame being used even in classical theory. For example, the tuo events of entering and leaving a room are separated by zero distance in a terrestrial frame, but are separated by many miles in a frame fixed in the sun for which the earth is a moving’ body. Taking axes S and 5 in the usual directions (see diagram), if the events take place at A and B in S, their coordinates can be assumed ‘to be (eq ot) and (xytDeose.t) respectively, In S$, let (qs) and (qs E4T) be the coordinates of the events. Employing the transformation equation E=B(t -ux/c*), we obtain the equations-9- Eos p(t ~ u/c), t+tT = pft- ere Deos®)/c*] Subtraction leads to T = = BuDcos6/c*. Since ¢ = (1 - u2/c2)"#, this equation can be solved for u to give the result stated. No.17, If is the proper time interval between the events and D is the distance between them in S, then equation (7.4) gives 1 2 oly 2 ella Pn aDt = TF - EB, < Thus, D =D. Let_the coordinates of the events in $ be A(x,t), B(x#D,t#T), and in be A(%,€), B(Zg,E-T], The transformation equation E = 9(t-ux/c*) yields the equations E = Bt -uxfe*), T-7 = alt + T- u(xed)/e?] Subtracting, we find T = B{-T + ub/e%), Squaring and solving for us we get u = 2c°0T/(0F + c*TF). If this value of u is substituted back into the previous equation, it will be found that it satisfies the equation with positive B only provided D > cT. No.i8. In S, the particle's trajectory lies in the xy-plane and is determined by the equations x = it, y = dct - got? The transformation equations to $ are x= detkeh, y = ¥ t = Fibs Hee). Substituting in the first of the trajectory equations, we get X= 2ct/5. Putting this result in the third transformation equation, it follows that along the trajectory, t = 4v3E/5. Substitution in the second trajectory equation then gives ibe 2h yg ce - Bhoee, Thus, at E = 0 when the particle is first projected, dx/dE = 2c/5 and dy/dt = 2V/3c/5; these x and y components of velocity correspond to a direction of projection making an angle tan"*(v3) = 60° with the X-axis. There is zero acceleration in the % direction and d’y/dt* = -489/25 gives the acceleration parallel to the Y-axis. No.19. At time t in S, al1 the clocks stationary in this frame register the time t. Thus,.C registers t. At this instant, the clock 0 registering € passes C. The coordinates of this event in S are (x/st)s the coordinates of the same event in S are (0,t). the transformatfon equation t = CE + u/c?) applied to this event accordingly gives the result t = SE or E= v(1 - ut/c?)t. Taus, as D moves etong the line of S- clocks on the x-axis, it runs slow by comparison with these clocks. We now observe the situation from the S-frame. In this frame, at the instant €, suppose the clock at 0 registers t,. This event has-10- coordinates (0,t,) in S and coordinates (%,,t) in 5. Using the transforma- tion E = @(t - ux/c*), we find E = Bt). But we have proved that t = gt. Hence t, = eo7t = t - u*t/e*s i.e, as seen fromS at the instant E, the clock Cis u*t/c* ahead of the clock at 0. We now follow the situation as it develops in S and, whiist making allowance for C running siow in this frame, show that nevertheless C is ahead of the clock at O when they meet. Thus, at time = 0 in 5, 0 and U coincide. Subsequently, 0 recedes from 0, whilst C approaches, both with speed u; hence, the rates of clocks 0 and C remain equal but are slowed by a factor {1 - y'/el). “At time fafter 0 leaves D» the clock 0 therefore registers V1 = ulgc*)E = (1 - uP/c?)t, and C arrives at 0. Since C is ahead of 0 by uét/c®, on arrival at 0, C registers t. The paradox has now disappeared, when C meets 0, C registers t and O registers t, where = = v(1 - u?/c*)t, i.e, C is ahead of T as calcuTated in the frame S. Wo.20. At time t in S xo Qett-1, y= elt 1, To transform to the S-language, we use the transformation equations x = HGR+ 3d). y © Ye ct = a(R + SCE), Upon substitution. and solution for ¥, y in terms of €, it is found that K = qho7e - 20), F else - 10). If d is the particle's distance from 0 as measured in S at time PA = pbgch (13 - 4408 + 500). d is a minimum when 226f = 440, which is the result stated. ¥, then a inS and €=0 ind, No.21, The event 0 passes U occurs at t = In S, the coordinates of the event (ii) are (ast). In S, the coordinates of the same event are (a,E,). The transformation equations = B(x - ut), x = p(X + ut) applied to this event show that a = Ba- ut), 2 = pla tut). Clearly, tes t= T= al - ety If T= a/3e, this gives I - y(1 - u®/c®) = u/3c, Solving for u, we Find u = 3¢/5.EXERCISES 2 Now. Lf rectangular axes Oxy are rotated about 0 in their plane through an angle @ to give axes Oxy, it follows by sinple trigonometry that @ point having coordinates (x.y) in the first frame has coordinates (x,y) in the rotated frame, where KE = xcosé + ysind, YF = ~ xsinb + ycoso. The matrix A of this orthogonal transformation 1s therefore as stated. If the sense of the Y-axis is reversed, so that the transform- ation relates two frames which cannot be brought into coincidence by a simple rotation, the second of these equations mist be replaced by FJ = xsing -yeose and the matrix A amended accordingly. In the first case, {a = cos*e + sin?e = 1, but in the second case [A] = - cos*e - sin?g = -1, This verifies equation (8,10). Solving for x and y in terms of ¥ and Y, we find in the first case that _ - x = %eoso- Ysind, y= Xsind + Yeose. ‘The matrix of this inverse transformation is AT! and thus wt = [cose -sing] = A’. [= ml That Avt = A‘ may be verified similarly in the second case. The transformation equations for T;; are Tar = ngaaglay Tha = Aygdagl 1 14300570 + T,,cos@sind + T,,cos@sind + T,,sin?e 2, int ii {4g * “Tuasos0sinor T,,00870 - T,ysin?@ + T,,cosesine fax 4aq2agTyg 7 ~Tascososing - Ty2sin?8 + T, C076 + Tp, cosdsing Tae = 824825144 T,ysin?o - T,,cosdsine- T, ,cosésin6+ T,,cos*6 It follows that a, an? T1q = Tan t Tan = (Tas + Tes) (cos* + sin?o) = Tyyt Tae = i.e. 7,5 is an invariant, (The second case is dealt with similarly.) No.2. Since A and B are orthogonal, AAT © BB' =I, It follows that BA(BA)' = BA(A'B') = BIB! = BBY = 1. Thus, BA is orthogonal. In subscript notation, tf aay and bay are elefients of A and B respectively, then Pa ae Eliminating Xe we find x, = esas THK-- where cy, = bs j85, are the coefficients of the resultant transformation. Since T,, is a tensor with respect to the transformations with matrices B and A, den + i537 Pépbiqlyg? —Thq = prgs"rs* ENinnating Toye We Pind Tip * PipprPjcegstrs 7 SintysTrse proving that T,, transforms like a tensor with respect to the resul tant transformation: No.3. Since A; and 8; are vectors, their trensformation equations for the change of frame represented by the orthogonal transformation Rs ayghy tye are By = ayyAye By = ay By- The inverse transformation equations (11.4) show that AG aR sayy Bm 8s Since XjjAj8; is an-invariant, 35. X GAGBS = XG 5AVBg . XygA,Bs- Substituting for A. and B. from the inverse transformation equations, this gives r s KiIMPy = Mrstarhitys®s- Since the components K, and B; can take arbitrary values, we can equate the coefficients of the telms in ‘A,B, on the two sides of this equation to yield J Proving that X;, transforms as a tensor. No.4. It is easily verified that the matrix A of the transform ation satisfies the condition AA’ = I and is therefore orthogonal, Using the inverse transformation, it is found that AL = oe UR mm)? Ae oe esl, = BR, + 2%)? Ay = 8 > bel 2% - 10%, ~ 1H )* Riso, since A, is a vector TyCGA, - THA, + 284) «dan, + hy + 2s) = qp(lon, + 2a, - 174,) ae These two sets of equations now permit the K, to be expressed in terms of In the x-frame,“Be diVA = Ag BOQ HX, HK) In the X-frame, divA = OR /2K, Substituting for the x in this last expression from the original transformation equations, it can be shown that it reduces to the form 2lxy + Xe + x5), thus verifying the invariance of diva. = abe ae, - 50%, + 28). No.5, The vector Ay, has components Ming? Arua t Aaae t Asa Aiag * Atay Anza # Baa tT ing 7 Aiax tase + Aan 7 If Ais the matrix of the given transformation equations, it is sasily verified thet AA‘ = T'and hence that the transformation fs orthog- onal. Then, Brae = 22482g2aAijk psQardarMare + AnzdandseMnre + gzdardaahere having omitted the zero terms in the sum, Substituting in this expression, we find Ay,, = 120/343, The matrix of the inverse transformation is the transpose A’. Thus, the equations are 1 oo a) te = We 6 +H - 2%) xy 2 We my - 6 - 3%) Since Bij is a tensor, By = esFro are the coefficients of the inverse transformation. Thus, = (-3/7)(-2/7)1 = 6/49, Fina rs 3298 No.6. Taking transposes Av os (The BNI! - Bt) = (1 - BT + 8}. Hence, AYR = (T= B)7NT + 8)(1 - B)(I + By = (1 = B)S(1 ~ B91 +B)" = (b= ahr = 8) +8) + By se . proving that A is orthogonal. if B is the 3x3 matrix given, thenTap eft 2 2) Geet = dfi 2 2 -2 ot oO 205 -4 b2 0 1 2-4-5 Thus aa dfi 2 fi 2-7 = bpp 4 4 2 1 af2 6 al 4 1 -8 2 0 If2 -4 4] 4-8 1 and the transformation equations are = Ey = y= 4) = YK +X = 8x) = Han, - 8 +x) Since C;5 fs skew-symmetric, C,, = Cy, = -1 and the remaining components vanish.‘ Thus, ¢, G, ¢, Tha = Qagdejlag * AxranFiz + Aadaslre * Baadailen * ArserCoy omitting the zero terns. Substituting the vaTues of the a,, and C5, we Find C, = 1. wey The coefficients of the inverse transformation are aj, = aj; Thus s a 0, sun 88D > MHD GE © AssM2r81 Dior + PrMet@erier + ArdortarPras omitting the zero terms, Substitution of given values now yields the result Dyyy = 980/729. In the ¥-frame, the vector 0,45 * Ej (say) has components Fy Tay = Dar + Cree + Shas = 2 E * Bagg 7 Bert + Deve * Dass B= Ty = Tan + Byee + Doss Thus, in the x-frame, this vector has components Bone = aE ae = 8 Be alEy = aE = ah = 89 a In the %frame, the vector CPi ak = Fy has components FL Gy, * CB Foo Tlie 7 Pe 7? Fos Tihs = Cidus = 5== Thus, Flow ah > auFr # aayF, + ayF, + 35/9 Flo abs, = agF, taf, +a,F, = -34/9 Fy ak, = aki tank, tayFy = -7/9 Wto.7, BY substitution of the given x; velues, we Find Bar = Aig = Bas = Ror = Aap = 9 Baz = Mag = Aaa = Mag = Te If ayy are the coefficients of the orthogonal transformation BreBraher + Are8rsMes + BxsMcaMor + Aadieheas Substitution of the calculated 64/49. values of the Ay, at P, we find Ay, We have Abd This verifies A, = 4x, in the case i= 1, The aeation may be verified sinflarly for thdedases | ="2,3, Then aD) + eM) + ay, O44) 2 Oe. (jy ,5) = Added = 12, = (4x9) Mii S As OP is rotated, the point ® describes an arc of a circle with centre N on the axis of rotation, arriving ultimately at Q. Thus, the triangle PNQ is isosceles with angle @ at the vertex N. Construct QM perpendicular to PN. Then QM = QNsing = PNsind = xsinasing, where a is the anole shown in the figure. ‘The vector x x u has magnitude xsina and direction parallel to MQ. Hence, MQ = x x usin6. Also, NM = NQcoss = NPcos#; hence, NM = NPcos6. Since xeosal = (x. » then NP OP - ON= x- (KUJu. Tt follows that NH = (x= (xwjuseos®. Thus Eons On + Me me = ¢ which 1s equivalent to the result stated, Ju + fx ~ (X-u)ubcose + x * usine If the Cartesian frame is kept fixed and the position vector of the point with coordinates x; 15 rotated about the stated axis: through an angle @ = “sin“#(4/5), the ‘new coordinates of the point witl be X;- Hence, the X, can be expressed in terms of the x, by putting © = -sin™#(4/5) jn the equetlon just obtained. The unit vector Along the axis of rotation has components (1/3,2/3,2/3); hence, we get6 - kek) = Rryeraanty) + Blt, + Oxy + 2x5 )(12,2) + Bex 22 Kater 2) After rearrangement and separation into three component equations, this leads to the transformation stated. No.9. The only non-zero components of A i. are Ay, = Ay, = 1s Ae, = Ag, = -1. Thus id tems Raa * ng8egMyj = as@azhan + AarQzahas + Arotasher + AatarAsr = 13. Using the transposed matrix for the inverse transformation, we have 8 esa * N42 B ge * SaaterSaaTaen | MusteQasBiee * MartartTaae = 188/729 In the X-frame, the vector Bag = c, fas components TL = Fagg = Baas * Bier? Brat 2 Rasy Ty = Bayz > Bere + Beez # Baas = 0 9 Barr + Bare + Beas = 0 In the x-frame, therefore, its components are C, = a0; = a0, = 29 G, = aft, 2 a0, = -16/9 C+ auth = ays, = 8/9 In the T-frame, the vector A, 3B; 4, = Dy has components By = Bij = ArBrer = 71/3 Dy = KyiByje = MiaBiee = 2/3 Det Rijs + ibaa 7 9/9 In the x-frame, its components are Dy = af Dy © ayyD, + ag,D, + 2,40, = 47/27 DO, = ab 0) = eyaD, + ay,0, + a,,0, = 11/27 Dy = aly = aD, + a,,D, + a,,0, = -10/27 No.0. The vector Ay 3; has components . : 2 ax? 4 Ox? 4 2x? Bl Aang "aan * Azae t Aare * 8 # 2xp + 2x . 2 2x2 + By? 4 xk B, = Aung * Bron * Aaea + Aoze = 28 * BH + 2x5 24 Ox? 4 Bx? Rasy 7 Aron * Anse © Anne © OME + 2x9 + BxG-W- Its divergence 1s therefore B ; = Téx, + 16x, + 16x,. The vector Ayj; = Cy hes components + Ayyy = Oxd + Oxd + Ox? Cy agg Aaay + Ay ze t Aas Ce * egy = Aon *Baaa t Rasy = Oxf + xt + 3x8 Cy Aggy * Banat Baza tags 34 + eh + OG Tts curl has x:-component Cy, - Cy, = 6x,. The x,- and x,-components are, similarly, 6(x,~ ¥) and X,) Respectively. No.¥, The components of Mag in the x-frame are Cy Bigs = Aang + Aron # Anas = 6 Co Magy = Reta + Maza + Azaa = 18 Cys Ags = Asa + Myon t Anas =O In the %-frame, therefore, } T= pel, - 40, + 204) = 12 te - heey +e + 26) = 9 T, = qp(toc, + 2c, - 11c,) = 6 aA ik Aras Pawtaas + Ay2tezAscAaee + ArsterBsrAion t Bree sBaahass = -1396/225 Using the inverse transformation, Baas MiahgFay © SeatePaa t Seated = 72/3 ‘The first equation of the inverse transformation is de, - 2% + 2%) vik ‘The components of VW in the frame are accordingly and, hence, - Ty + B)* WR, = Be, 4M) = 2 avr, = = 2k, + 2K) = 4 BU, = YR, OR +R) = 4 No. 12. Boe 43, 4Aga = Aue * 4408,,4, ne > 84823845 > Aaa%2e4ie * Srefarten = costa + sinte = 1- Be iWo.13, Taking { Begg 7 DAaalOK, + AR 9%, + R/O, =, # OK, + 2x, which verifies (a) for this value of 1, Similarly, (a) can be verified for 1 =2,3. It now follows that ARL5 g/t Mag G/F Wy j/Ky 4A HE IZ Ais ig 7 aysProe + Aeataotaehens = 7192/25 No.14, yey 7 548 g2atAigk 7 @ar®u2®aehiee * AszPi2%icheee Using the inverse transformation = ate © 4pm aBiy * AyiteBja =O B since a,, Xo. 15. (4) Taking transposes, Rio = Tt = (2xx'ph = T= xx! Thus RAY = (I = 2xxt)(1 = xx!) = 1 Gant + Ox(etx) xt = Tn dat tax! 2 7 and A nust be orthogonal, therefore, For the given x’, xix = oC1e-2,3)[ 1] = 0812 + 2% +34) = Me? -2 xz Thus, x'x = 1 provided a = 1/714, Then = 2 - ls ~ as ft o §-8pia-es 6 2 -J a ) ~2 2.3 6 6 0 1 y 3 6 -2 Using this transformation matrix, Banar ™ SeitesPextighijke ~ Sasteatee@asaase + AenQeates®r2Br202 = 72/43 8, 4g 18 an inveriant, Hence Biss = Bigg Bias * Zero terms = -20 We find that JA] = -1, Hence G=ia,G = - Hee, +2, - = HE, + 3G + 5G) & = tala, E = tla, G, = - 7G +6 - 2G = 0-19- (44) The non-zero components of A;; are Ay, = A, = Agy = ea Ass Thus, the non-zero components of Az; , are Ary > = Aye © A Oke Since Aj; 18 @ tensor, sk aye © Banya Baas 2% i, aye = 8158 g2aghig Ay a2 Aaa at A, * x28 p29paheg a trast tahsa s Agr = 4x,/7 Using the inverse transformation, ye § HOR, +H, +H) = 2 Hence, By, y= 8/7. No. 16. I, Muy + 2aadeehen = 6 fae = agSegAiy = Aantetnr + Paaeahen Using the inverse transformation, B, Byyp = 2h ox8arPar Beas nan Nga > Mak gaPaBigk The determinant of the orthogonal transformation is [A] Hence * Ea (lyf = las = 2 No.17. Since A is skew-symmetric, A' = -A. Thus (A®)' = (AA)! = AYA' = (-A}(-A} = BB, showing that A? is symmetric. If, also, A’ = -A, then BB = (1+ 288)(1 + 2A*)' = (1+ 2AF)(I + 284) = T+ 4A? 4 Gas = 1 + 4A" + 4a(As) = 1+ 4a? - 4A" = 1 Thus B is orthogonal. By multiplication of & by itself, we find AM = (2h + bP c2)fO -a -b a 0 -c c 0. Thus, if a? +b? +c? = 1, then AP = “A. Ife = 1/3, 6 2/3, then a® + be + ct = 1 and 2 le an fio fs 4 2]- gf 3 od ne 5-2 “8-1-4 9 6 2-2 8 4-4-1 Thus, Ty = bead, by bj, + 2y,2y_ rete. = 91/81 aa 7 Peis slay = PoaPar * PasPae= 20 - Since xf + xf +x} is the square of the distance of a point from the origin of the'x-frame and the origins of the two franes' coincide, the transformation must be such that Maxgaxs Bee g Hence 1 R= byl, = bak = - 4+) Ry = byiAy = ask = > Bh + +) Ft bats = Ae OTHER) In the x-frane, the divergence of the vector field is At Aaa ths = A = In the %frame, the divergence is 2 Biya t Pay tae 2 oH The first of the equations defining the inverse coordinate transformation is uF ye iy - HR, + Ey) This establishes that the two expressions for the divergence are identical. No.18, (i) The orthogonality conditions are atagta dos B81, dat 8b +28 = 0, ab? +7 = BM, Bat Brac = 0, @ eta © Bl, Web +7 = 0, A these equations are satisfied if a= 1, b= We have, with the usual notation, k, sa PasMaghiy 7 Pantaahes = 479° (41) Putting Byy = Aygy> WO have E, i, ke” igks 7 Mr8gstke®urstu ” § ruejstkt stu js thrstr ~ Aysteehst having used the orthogonality condition a,,a;, = 6... This proves By, is a tensor. itty Stu For the component values given Bas > Aiand = Aaaan + Acasa + anes = 9 Bea 7 Aiand * Aazar Maze Maze ™ 18 AIT other components of 85, vanish in the x-frame. Thus Bat enyaghay 7 MaatasPan * aataaBee = 473+- a. Ko. 19. Baas = Meitag@siAige = MaaPactastya = 492/3125 If Cy = Byj,y then C= Bray Bras * Beas C= Bay = Bran tBasat Baar = 4 Ty = Big = Bras * Baar + Base ™ 9 +B Thus, using the inverse transformation Cys at, = at; = 3 C= ART; = aye0y = 0 Cy at, = ay 0p 4 No.20. By use of the identity costa + sin?a = 1, it is easy to verify that a1] the orthogonality conditions are satisfied, “It is also easy to check thet the transformation is the resultant of the pair of transformations = xycos8 + x,sing xycosa ~ xysina x =~ x,Sing + x,cose 5 * Xystna + xycosa XE eX The right-hand menber of this pair represents a rotation of the x-frame through an angie g about the x,-axis to generate the x'-frame. The left-hand transformation represents a further rotation of the x'-frame about the xj-axis, ‘through an angle afin the negative sense), to generate the %-frame. If q © 8 = 4m, the transformation equations become > 1 1 1 Dat he” eke Sole Fm) aly ad Pat Pht With this transformation Funes = 8ng8tj2ac@apAajue = StedantertssMprest Mretaade 20 No.21, Taking the transpose of AB, (AB)! = TAT = BATE = (AB)? which is the condition for AB to be orthogonal. Giving A, B the values stated, we find =a - 2 c= AB we 9 20 12 15 16) -20 0 15 Thus.Wo.22, ~ 22 - sar SsqSog"aeigk = SaaSestashens + CenSe nahn -48/5 Writing Dy = AjsByjpe> We have Bt Bigijer = AresBaaa 7 48 Dz = Bip Bijke ~ BearBecre = 72% Dye BisBiges = BazsBeare = “14 ‘Transforming this vector to the x-frame, Dy = cfDy = cyyDy = cyyD, + cgyD, + 6,0, = 4468/25 D, = CLD; = c4eDj = Cys, + yD + Cy. = 96/5 D, = 650, = 6,0, © c,,0, 46,0, + 6,,0, = -48/25 Relative to an orthogonal transformation Xj ° ay sx; + by» we have transformation equations Bis = [alae snse By = Haley Ey = ays TAI? 0448559542 PAEY Srtos hE,» since [AI OBES Spat a proving U Be, ‘is an invariant, No.23. In matrix notation, X= Ax, where ayy = ~ayp = At, =X Bes (ourta), = AL. AL = ¥8e0%C,,) + xy8%RLKX,) me ye(1 + 1/2¥2) (CurtA), = Ay y — Aajn 7 Mn608(% Xp ~ 800% Xp) = > vm (1 2/2) (OurTA), = Ag = Aya #7 %9S1N(H4K,) = XyC08(%,%5) = = v(n/2) The transformation to the ¥-frame is determined by the equations 2 as = 1, all other elenents of A being zero. Clearly, |Aj = -1. are Since curlA =@ is a pseudovector, in the X-frame its components~ 23 - © |alaG = -G, = ~ Ml + 1/2vey Jala = G, = - 1 - vere) WalaG = -G = vere No.24, The orthogonality conditions require that AML (at +2454) = 1, attb+ld = 0, et(1? +b% +27) = 1, 2-lib+2e = 0, P(A eect) = 1, 2a-2+Se = 0, V4, b= Solving, we find a 0, a = 1/15, 8 = 1/3, y = 1/15, If X= Ax represents the transformation, then Baron = Bag@sj@e@asAignn = Par@as®erPaaMnsze + PastsetentiaMarze = 25 Jn the Keframe, the vector Cs = AjB,; has components T° AB, = AB, RB, +B, = 0 T= RBs = Abus + Biber + Bdys = 3 T= AB, ELL +A, +15, = 15 Transforming to the x-frame, Coe at = aay = Aly ttle + anls = 3 Ge = aggly & agaby © Ayal, + apAls + Agel, = -9 Cy = abc, ~ ay; = aC eat +a, C, = 12 No. 25. We are given At = -A, AP = ‘Thus, Bios T-A+AP and BB'o= (1+ A+ AP(L = AFA) = 1H AR + AY = T+ AP AAR) = Thus, B is orthogonal. ‘The given matrix A is clearly anti-symmetric and, by repeated multiplication of A by itself, we find oo rs 4 2], we Ef 1 2 “4 8-2 10 2 2-2 -8 2 2 9 Thus, AS = =A, We now Find that : 1 e- fi o oegfo 1 afsgfs -« 2 o 10 Hl 0 2 4-5-2 oo 1 k2 -2 0) 2-2 -8= 9 4-1 8 74 4 48 1 Thus, Ld ‘ease = Dysde hye 7 52 Boar = PasPesinbigk = PrsbasPsrBues © “128 No.26. Considering the ith component (curl gradi), = Gaylord), s+ Cy V, since terms in which J, k are the same vanish, and other terms cancel in pairs, e.g. Sire Har ar 7% Also diy curt = Curt), = (sph 3) 4 * Cijkhk at since terms in which i, 7 are the same vanish, and other terms cancel in pairs as before. No.27, (i) There are four cases to consider: Case I: the pair (kya) is not identical with the pair (m,n), Then, each term in the left-hand member must contain an e-factor, two of whose suffices are equal; this member is accordingly zero, Also, each term ‘in the right-hand member must contain a 6-factor whose suffices are differents this member is 21so zero therefore, The identity-is thus verified in this case. Case I]: k =m, 2=n, but k # 2. Then one term in the left-hand menber will be non-zero, this being the tern for which (ik#) and (inn) are the same permutations of (123); this term has the value +1. Since pq = San = 1s Sig = Sgy, 2 04 the identity is again verified. Case TH: kn, 22m, but k #2. Again, one term in the left- hand member will be non-zero, this being the term for which (4kt) and (imn) are permutations of (123); however, one of these permutations will be even and one odd and the value'of the member will be -I. In the right-hand members 842, ~ Spy = Or Bjq = Som * Te thus verifying the identity. Case IV: k= = m=, All terms in the left-hand member are clearly zero, Since Spm = Seq = Sjq = Som > le the right-hand member is also zero and the identity is verified, CH) Sipe = SkxSam ~ Sembee = 38pm 7 = Bom No.28, By definition of divergence diy gragv = (grad), ¢ * (V4) 5 = = oy No.29, By definition of curl, the ith component of the left-hand member 1s a DA, (84)835 7 158i) As rg= 25 "jig 7 Ata (590 FM (diva) 5 ~ (9A), “Ay This is the ith component of the right-hand member and the identity has therefore been established, No. 30. (4) Taking the ith component Tax (Bx CH, = 4 5pAy(O* Oy Shs Fersh os » as esAiBrls (epBig > Seq5jr)AsB nC Baty — AB5CG (AC)B; = (A+B)C; which is the ith component of the right-hand menber, (4H) RBC) = BY(B% Oy = ASG BSC, = GAG BSC, This last expression is the expansion of the determinant and so verifies the identity. No. 3T. div(WA) = OAD) = YANG + VAs = divA + A-W Wo.22. (5) Court AD; = G5C¥M 5 = B5gg * Gaui = V{curta); + (grad x A); (ARB) = EG GASB) FA Pt FPA = Geass ae 7 Baha (curtA)yo, = (curtD) 5A; (ii) div(a x 8) BecurlA - A-curlB (414) The jth component of the operator ¥ is a/ex.. Thus, the formal scalar product of A and V yietds the operator A-7 Aj/0X5, Hence, the ith component of (A+V)B is (A-v)B, = A,3, 5. Thus CUA B) 5 = CABS = yu lGrstPs) 5 St sens he Be * NBs 3) (5p8js ~ S45 jr) An ss * A Ps.5 K As aBg TANG AaaPe 7 PHS (BeT)A, + Aj dive - Bydiva - (AWB y verifying the identity.~ 26 - (iv) (A curla)s = @yAs(curlB), = SAG Be = Fei sPershPsyr = Bin85s ~ S6Fjr)Ai8syr MB MB, simflarly (BX curta), = ByAy 5 - BA, y Adding (A * curlB + 8 x curlA), AB TALS NP 7 BAL {AjB3) 5 - (A7)B; - (B-VDA, [grad(a-b) - (A-9)8 = (8-9)A, ubich is equivalent to the {dentity required. 0.33, Transforming to an %-frame by the orthogonal transformation x, 8,3%; + b,, we find i 155 i By = Ra = Aprtistes > M4s%rSsr showing that B;; transforms as @ tensor. Te now follows that Ay; = Aj; * 6); 18 @ tensor equation and hence is valid in atl frames if it is true in one. This result is easily generalised: Thus, if Ays, is a tensor and By i, = Avsjy then 8, is a tensors i.e. the components of a tensor can be redifengel'do provide Bomponents of @ new tensor. No. 34. Since 6} jj, = 9), follows from the properties of the Kronecker delta, the First identity follows by putting ay, = 65, f (15k) areal different and (tun) 4s an even permutation of (isk), then (2mm) is obtained from (4jk) by an even number of transpositions, and hence @ 5, and @,,, have the same sign and their product is +1. If (2mm) is an odd perutation of (ijk), the number of transpositions is odd and €55_ ANd Bon MAE opposite signs end their product is -1. In all other cases, one of the triads contains repeated values and one of 154 + &qq 18 Suppose (ijk) are all different and (Sm) is an even permtation of (ijk), Then one of the three even permutations of (2m) will be (i3k). Hence, one of the first three terms of the right-hand menber of the next identity to be proved wiil be a product of three delta symbols, each with an identical pair of suffices and this term will have value +1, The remaining Five terms of this member must each contzin at Teast one delta synbol with non-identical suffices and so must all be zero. Since the left-hand menhar has already been proved to equal +] in this case, the identity is verified, If (ijk) are all different and (2m) is an odd permutation of (idk), it will be one of the last three terms of the right-hand menber which ‘is non-zero with value -1, all other terms in this member vanishing. In this case, the left-hand menber has been shown to be ~1 and again the identity is verified,-27- If (mn) is not a permutation of (ijk), then at least one pair of (tm) or one pair of (ijk) are the same and both sides of the identity vanish. Next, suppose (tmn) is a permutation of (ijk), but a pair of indices are equal {n each case. The left-hand menber vanishes. In the right-hand menber, one only of the first. three terms will be non-vanishing, with the value #1. By transposition in this term of the pair of equal indices in the group (tun), one of the last three terms will be generated: this term will accordingly contribute -1 to the member and cause it to vanish. Again, the identity is verified. Finally, suppose (2mijk) are a1] equal. Both members of the identity then clearly vanish and the identity ts verified. Contracting by putting 2 = 1, the identity gives 3% = 5448 jnbun * PiaP anf * Pin 3i%m SanPjmPei ~ S448 3nbkm ~ Sim 548m 385n8kn + Sknojn + Sink ~ Bn8jm = 8 jnbkm ~ Sim? en Sinbkn ~ Sinem No. 35. G) (ax by lex dh = (ax db), (6% d)y = C5 5pa;d Ci nee ye = 5 pO gs ~ 8458jgn) 23> Koo Ss a jdyC dy - ASPs (a-c)(bed) = (a-d)(b-c} Hy = Gggla * bI;le 8), = Gi3n@irs?rPsCemn adn (Sqrbis 7 8ksSir) Gamné Psat em? on’n ds ~ emnP Sn) 24 [acd]b, - Tbcd] having used the result in No.30(i1) above, The alternative identity can be established by associating a different pair of e-symbols together in a contracted product. q it) [fa xb) * (exEXERCISES 3 No.1. ‘The successive space-tine points occupied by a moving particle have coordinates in the two frames S, S which are related by the equations (5.8). Differentiating these equations, we find dK a(dx - uct), = y= dy, cia B(dt ~ udx/c®), dz = dz, Thus, dx ~ ude ONE = ge yairet = a Ter since v, = dx/dt. Also < A y Wy = GSE = erage uaey BW 7eT since vy = dy/dt, ¥, is obtained similarly. No.2. Differentiating the equation for ¥, obtained in the last exercise, we get x 1a ute? - iv WIEN Since aE = Bl ~ uv, /o*} et also follows from the previous exercise, division of this pair of equations Jeads to the stated equation for a, = di,/dt. Differentiating the equation for Vy from the previous exercise, gives y dy, w/c %y = eerswey * amine y BI w,7e ~ tv, 7 x Again, division of the last two equations provides the equation for ay . Gs, The equation for a, follows similarly. Clearly, if ay, ay, a, are all constant, in general a), a), @ will not be constant, sincé thése Quantities will vary with the velocity Components V+ Vys V+ No.3. Take the nucleus at the origin of the S-frame. In this frame, we can assume the electron to move in the xy-plane with velocity components Yy = 3C/7, Vy = 378C/7. Taking $ to be the stationary frame, the velocity transformation equations give for the electron's velocity components in S 3, 1 y 2 ote = (ha ut/e*)#3v3e/7 The YY Beet ic where u is the velocity of S and the nucleus. Hence . Oo. 2 Che vse2pbavc Wis = tan30® = wiv, = Comgnge ntsc Squaring this equation and solving for u, we find u = 3c/5 or =12¢/13. The second root is spurious, since it fails to satisfy the last equation.- 29 - No.4. Take § to be the frame_of the stationary observer and 5 the frame moving with the nucleus. in 5, take the velocity components of the B-particle to be ¥, = 0, ¥, = 3c/4, ¥, = 0. Since u = 3c/5, the transform ation equations give y z x 2 ‘Thus, in S, v = 3v2c/5 and the direction of motion makes 45° with the direction of motion of the nucleus. Ye = 30/5, Wy = 36/5, vy = 0 In the second case, let_@ be the angle in 5 between the Tine of motion of the f-particle and the X-axis. Then, we can take 3e cose/4, ¥, = 3csine/4, Thus y Ze coso + 3c Resing yo De ye x 1+ gycose ° 1+ Fyeoss Since v, = 0, the first equation shows that cosd = -4/5, i.e. 9 is obtuse. The sectnd equation then gives v, = 9c/163 this is the velocity of the B- particle in S. y No.5, ‘The square of the magnitude of 2 vector is defined by equation (2.4). In the case of the d-velocity, its magnitude ¥ is given by Vee MV = VV, = (1 = vP/e8)-#ystc) e(ysic) = (be vey vt = 2) = He? Hence, ¥ = No.6. Consider a photon in the light beam, Its velocity components in $ along the x- and y-axes are c cose, csine respectively. In’S, the corresponding velocity components are c cost, csina. The velocity trans- formation equations show that = = feos =u az coed = ESE strié Division now yields the equation stated. __ Clearly, if wis small, a and © are approximately equal. Writing T= a + Aa, where 4 is small, a Taylor expansion shows that cott = cot(a tha) = cota = Avcosecta approximately. Further, an approximation to the right-hand menber to order (u/c) is clearly cots -'(u/c)coseca, Substituting these approximations for the two nenbers of the aberration formula just found, we deduce that da = (u/c)sina, No.7. The first result is obtained by differentiating out the right- hand member of the equation 4 ny © eT Tapert If the acceleration is parallel to v, it is directed along the tangent.to the particle’s trajectory and the conjonent of the acceleration along the principe! nomel to the trajectory (viz. xv!) vapishes. Thus, ‘the component of the acceleration along the tangent (viz, v) gives the~ 30 - magnitude of the acceleration. The unit vector along the tangent is v/v and, hence, dv/dt = w/v. Substituting @ = vdv/dt into the second term of the expression for # now gives the second equation. If the acceleration is perpendicular to v, the tangential component must vanish and so ¥ = 0, The first expression for f therefore reduces to the last equation. ~ In these two special cases, the force can be expressed in the Newtonian form: mass acceleration. In the first case, the mass is m,(1 ~ v?/ct)"™ and is sometimes termed the longitudinal mass. In the second case, the mass is m,{1 ~ v*/c?)"# and is sometimes called the transverse mass. No.8. The velocities of the particles in S are su (i) Elimination of ¥ leads to the equation (vy # v)ut = Ayu, teu + (YH ve? = 0 Solving this quadratic for u, we obtain fot uy, ~ (c? ~ vi) Re - vayt WW The negative sign must be taken before the root since’ (c - v,)(¢ - vg). > 0 ‘implies that c? + viv, > clv, + ¥,) and u would therefore exceed c i? the positive sign were taken, To obtain ¥, it is simplest to eliminate u between the equations (i) to give the equation (vy ~ ¥2)¥ = (ct - yyva)¥ Hcy, = yy) = 0 Solving this quadratic for ¥ gives the result stated. No.9. Suppose the ‘bullet AB is attached to the K-axis of 5 as shown, ‘The camera is placed at C on the y-axis of S. The fixed scale is the x-axis. Suppose the light entering C at the instant the shutter is opened leaves A at time T and B at time t as measured by the clocks of S, At time Ty let A have x-coordinate x, and at time t let 8 have x-coordinate xp. Ten the difference in the distances travelled by the Tight from A and B is (xg ~ Xq)cosa (assuming the length of the bullet is snalT compared with the distance of the camera from the bullet). Since the light from A and from B must enter the camera at the same instant, we must have-n- T= t = lig ~ xp)eosove (a) Consider the event of the light leaving A. In S, the coord~ inates of this event are (x,,T). In S, let the coordinates be (xq,T). Then Fy > BUX ~ 97), where B= (1 - vA/ct)"#. Simtlarly, if the Xcoordinate of B is gs then % 8% - vt) Subtracting, we get d= Fy Hy = Blyy ~ xp) + BIT - t) Substituting for (7 - t) from (i), we find thet MyM = CL v8/0%)84/(1 + veosu/e) Since x,, x, are the coordinates of the points on the x-axis {fixed scale) fron witch he light leaves to enter the camera, this will give the apparent Tength on the photograph, No.10. If § ig attached to the table and S to the rth cart, the velocity of 5 relative to S is cv,. The velocity of the (m#1)th’cart in Sis cv.,, and in Sis kc. Using the velocity transformation equation, we have ke + oy, vt F re The last equation is equivalent to the recurrence relationship - isk Ue = TR where u, we find (1 ¥)/C + vy). By repeated application of this relationship, 1 1-k = (ay But vy = K and, hence, uy = (1 ~ k)/(1 #k). Thus, Since v, = (1 - u,)/(1 + u,). this leads inmediately to the result stated. Since 0 < (1 k)/(1 +k) <1, 4,0 as n+«, It follows that vy, + Tasn +e, No.11, S ig the laboratory frame, $' is a frame moving with A and S" is a frame moving with B as indicated in the diagram.- 32 - In $, B's velocity is v, = V and S* has velocity -V." Hence, ‘in S' B's velocity is x 5 Yt av “eT TR Wwe" Teer In S', the electron‘s velocity has components vj = 9, vy = ¥ and S" has velocity u.= 2V/(1 + V/c?), Hence, in S" the electron’ velocity has components yee Ty ko TST Tee (1 = ubyc?yty z vos y y Wye Ve Tous tana = w/(-vs) = 3(7 ~ V4/c#), The magnitude of the electron's velocity in S" is 3 v(vge + wr) = V+ (1 ~ vee8}4]2/(1 ~ ¥F70?) No. 12. . 0 ew eS —> ao — ——v+5y S$ — At time t, let v be the racket velocity in $; at tine t + st, the velocity has increased to v + dv. At time t in S, S' is an inertial frame relative to which the rocket is instantaneously at rest, S' has velocity v relative to S and is the crew's natural inertial frame. When the rocket has velocity v + év in S, let gv’ be its velocity in S'. In S‘, as the rocket velocity increases from 0 to 6v', suppose its rest mass decreases fromm, to m, + sm, (8m, <0); Tet am, be the rest mass of the particles ejected with’ velocity w'into the fet Stream during this change. Then, conservation of inertial mass requires that im, + 6g a, mo * Sore + wey and conservation of Tinear momentum that (m, + 6m,)8v' amg Os vrs awyey > eT wrey Approximating to the first order in all small quantities, these equations reduce to bm, + Am //(E = we/c#) = 0 mév' = whmy/¥(1 = w?/c?) We now deduce that m6v) = - wom (i) Transforming the velocity v + év in § into év' in S', we obtain ol vtiwey ay vs ayyanyyer = TS VE7eE a) to the first order in év. Equations (4) and (44) now yield in the limit-33- dy ey Hie Toye OR Integration over the range from v =v, to v= v, gives yy Hen be log fee sel = w togk, where R= my,/m,, = mass ratio. Solving for R, we get the result stated, Since (1 + v/c)® +e” as c 4m , we calculate that Rm [Aare] /2H exp f(vy-ve) i] This is the well-known classical formula for the mass ratio. If the jet is a stream of photons, then w= c. (This would be the case if the propellants were electrons and positrons which mutually annihitate to generate photons.) Then, setting vp = 0, vy =v, we Find R=v[fctv)/(c-vJ]. If R= 6, this gives v = 35c/37. No. 13, Equation (5.6) shows that tana = iu/c, tan = iv/c, —tany = iw/e where w is the velocity of $" relative to S. Thus tany = tan(ata) = $2na.t tant is equivalent to = uty wos Yue * No.4. ee fom = os oy The result stated now follows by use of equation (17.7). No.15, The force acting on the charge due to the field is of magnitude ef and along the x-axis. The equation of motion is accordingly d v abr ee = ee Since v= 0 at t = 0, integration teads to the result v starry = Kt Solving for v, we get ve ‘et ae” WTF RERTEY * Integrating again with the initial condition x = 0 at t= 0, the result stated is derived. If c is taken to be large by comparison with kt, the binomial theorem shows that~ me xe a + kt /2c% + O(MEt/C*) = T] mkt? as c-re, According to classical theory, the particle is subjected to a constant acceleration eE/m, = k and this result is accordingly consistent with such motion, If kt is large compared with c, the relativistic equation can be approximated by taking (1 + k’t?/c*) = kt/c. Then x = c(t ~ c/k)s i.e. when a long time has elapsed, the particle will appear to move uniformly with the velocity of Tight. Wo. 16. Take A and B to be at the origins of the frames S and 5 respectively, these frames being related as in section 5, Suppose B receives the tachyon from A at time € by the 5 clocks and at time t by the $ clocks, Then this event has coordinates {d,t) in Sand coordinates (0,t) in 5, The transformation equations X= B(K + ut) and xX = B(x ~ ub)*now show that E = d/fu and t = d/u. Tt now foFlows thatthe tachyon must leave A at time t = d/u ~ d/v. Consider the motion of the tachyon transmitted back towards fron 8. Tt moves along the negative X-axis with speed v, starting from at = d/Bu. Its equation of motion in S is accordingly % = - WE - a/su) Transforming this equation to the S-frame, it yields x(1= w/ct) = = (v= ut + vd/etu as the equation of motion in this frame, Thus, the tachyon arrives at A (x = 0) at tine t= vd = vad_~ uP/c?) © Bay ay uty = uy Thus, the S tine elapsing between the first tachyon leaving A and the second tachyon arriving back at A is vd(] = u/c?) dd d . “way v > vers uyly This time is negative, and reception by A occurs before trans- mission from A, provided - veu/e*) ww? ~ 2c8v + cu > 0 The quadratic equation for v: uv! - 2ctv + cu = 0, has roots Te & Hc? - uF]. By consideration of the parabolic graph of uv* - 2c*y + c*u for varying vy it is clear that the above inequality will be satisfied provided v < Sfe = v(c? - u®)] or v > Sle + vc? - uy] v The first inequality is equivalent to u ve 77 we i.e. v OL(T-B)uxvu) + pu{y-uy] Gi) = GF fly)? ~ (vxu)?/c%f G In the Independent &, we can take ¥ = (vys%ys¥_)s ps) > u= (u,0,0). Then Q = 1/(F - w/c}. Hence, the first component of equation (i) gives vy = a, - u), which is the first of equations (15.11), The second and third components of (i) give the remaining velocity transformation equations. This verifies equation (i). Since ux(yxu} = u®v - (uev)u, equation (ii) is equivalent to (i). To obtain equation (iif), we square (ii), noting that. (a) [uxtyxuy}? = uPdyxu)? (b) us [wx(veuy] Co) ve [us(vxuy] = (veu)(veu) = (xu) (a) B= 1 = ubp?/c® No.21. In the S-frane, a ray of light from a point on_the edge of the disc to U makes an angle w with the x-axis, where tana = a/X. In the S-frame, ‘this ray will be observed to make an angle a with the x-axis, where cots = EL wecosect (1 = ut/e*) (This equation follows from the result quoted in No.6 above by writing down the inverse transformation (i.e, exchange a, @ and replace u by ~u) and then substituting ma, ma for a, & respectively to allow for the reversal fn the sense of the Tight ray.) If Y> a, we can approximate coseci = cot = x/a and this then Teads to the stated result. No.2. By the velocity transformation equations (15.11) we Differentiating with respect to E, we get ge ML Wat | Va utct dt eg ET Wwe Differentiating the Lorentz transformation equation = = @(t - ux/c*), gives dt/dt = 6(1 - uv/c?), The first result stated now follows immediately. At some instant t in S, take 3 to be the frame in which the-37- particle is instantaneously at rest. Then u = v and, hence, & = a/(1 - v2/c?)¥, Thus, if @ = a, (late) Hee = a This is equivalent to the second stated result, Integration of the equation of motion now proceeds as explained in the solution to No.15 earlier, It is helpful to imagine the particle to be a space rocket accelerating away from the solar system. a will then be the acceleration experienced by the crews thus, if a = g, conditions in the ship will be similar to gravity on the surface of the earth, No.23., Tn S, 5 has velocity (0,v,0). Using the inverse of the velocity transformation equations (15.11), the velocity of § as seen from S is found to have components rcrjtv, ov, = 0 . 2 us Hence, 2 wye?}*, 4 tend = w/v = Gl The velocity transformation equations relating 5 and § can be derived from equations (15.11) by exchanging the roles of the x~ and y-axes and replacing u by vs they are y - verry hy, Wev 1s viet), (1 = v/c?} a tae! vy, Yo Le wylet ~ a fee 7 2 1 wyfe 1 Wyle’ In S, S has velocity (-u,0,0). Substituting ¥, 0 in these transformation equations, we obtain * 3 --a-veyh ew Fe Fo =v, Wor Hs 0 for the components of the velocity of S as seen fromS. thus, if ¢ is the angle made by this velocity with the F-axis, then sR oY - tang = Wi = Yl - vere) If wand v are small compared with c, the binomial theoren shows that tand = U(1- ut/2ct), tang = 1 + v¥/2c) approximately, Hence tang = tané ton(g - 8) = TPnpdane 7 ww/2e# to the second order of small quantities, No.24. 2 = B(1 - wse2y! is derived as in section 6. Let v be the velocity of $’ as seen from S. Then -v is the velocity of S along the x'-axis of S'. The velocity of 5 as_seen from S' follows from the first of equations (15.11) to be (u - v)/(1 - uv/c?), ‘This must equal v, thus leading to the equation= Be uv? = 2c%v + cu = 0 Solving this quadratic for v and rejecting the root > c, we now find 2 . vee SD -- ute) @ Two applications of the Fitzgerald contraction formula yield ‘the equations _ ; i ee TT utycty?, Lb = F(T - v8ye8) Thus, equation (i) fs equivalent to 3 £ @-wiyr e —2 og - 1m (we - 638 Solving for L, the stated result follows. No.25. It follows from the symetry of the situation that the conbined particle's direction of motion bisects the right angle between the velocities of the colliding particles, Thus, if v is the speed of the combined particle and My is its rest mass, momentum is conserved provided 2m, My ~Tymy 3 cost/8) = aarp Since energy (or inertial mass) is also conserved, we must have 2m, My FS yey YT = VE ery Dividing these equations, we find v = ¢/2v2. Thus My = ¥(14/3)m,. No.26. Since momentum is conserved, the particle M must follow the same Tine of motion as m. Thus, if M has speed u, conservation of momentum requires that mv - Mu = Tey Ve TS wyey a) Conservation of inertial mass is expressed by the equation f . a vos erey te * ASPET any Division of equation (i) by (ii) gives a formula for uy Dividing both sides of (4) by c, squaring and subtracting from the square of (11), we obtain the equation mg = v2/ct 2m gee MELD = uty? BE) ery a = Me This is equivalent to the result asked for. Hote that M? > mj + mj + 2mm, and, hence, M>m + m,. The ‘increase in rest mass corresponds to the mass of the heat generated by the collision, No.27. Choose an inertial frame § whose velocity relative to the laboratory frame $ is w in a direction parallel to the tine of motion of- 39 - the first mentioned particle. Choose w so that in S the two particles have equal and opposite velocities before collision, Then, as in No.2 above, we can. show that 2 wee Bo --ueyeyh Then, in S, both particles must have the same speed w. Appealing to the symmetry of the particles’ motions in S, it is clear that after collision they will be moving apart with equal and opposite velocities along another line of motion, Since inertial mass is conserved, these velocities will still have magnitude w. Let a be the angle mee, by the new Tine of motion with the original Tine of motion {as measured in). Choosing paratiel axes in S and Sin the usual wey and assuming the motion occurs in the xy-plane and fs along the x-axis before collision, after collision the first particle will have velocity components (wcosa,. wsina,0) in S$ hence, 25 : = (n/c?) Fusine fand = w/v, = woose FW by use of the velocity transformation equations. The second particle has veTocity components (~wcoso,-wsina,0) in S after collision and it follows similarly that 1 = (h= w/c?) busine = wessa FW Hence 4 2a /e2)wtsin’ . -wi/ct}wisinta 2 fee tanotand = aaeaeee = 1- w/e 2 7 9ST No. 28, Since momentum is conserved, the two parts will have equal and opposite womenta of magnitude p. Using equation (19.1), 2 cle? + Nc?), Ey = erp + Hed) Thus ep ef = (Mp Net on Energy is conserved and therefore E+E, = Net (ii) Dividing (ii) into (i), we deduce Ey -E, = cng - Mays (iit)- 40 - Equations (ii) and (iii) now yield the results required. No.2. Take mi, to be moving along the x-axis of S, m, to be moving atong's Vine making’an angle a with this axis and the combined particle m to wove along a line siaking an angle # with the x-axis, Since momentum is conserved, by resolving along and perpend- icular to the x-axis, we get the equations nyu, mau : fw vor titresy * ver shivers 8° = yer uereryoo8s mu, ina = mu meterery 810. = aeysuryery sins where u is the speed of m. Squaring and adding, we find nu nus 2myMaUgUy COs - ut reife + r= ipper + Suet —ueyy = Tae 0) Since inertial mass is conserved, " m . fn TU SRT 7 TUE WTS PTE which efter squaring yields : . mi nj 2mm, . a a Multiplying {i) through by I/c® and subtracting from (ii), we now obtain the result required. We can assume u,, u, to be positive and therefore (c® - uugcose)* > Ce ~ uzun)? > (c? = uf) (c? - uf) the second inequality being justified since uf + uj > 2ugu,. It now follows that n> mj + no + 2mn,, ive. m > my + M2. ‘The increage in rest mass is due to the inertia of the heat generated if the collision. No. 30. Let p be the momentum given to the electron and suppose its dir ection makes en angle a with the originel direction of motion of the photon. e o>: & EE, before colliston after collision Then, since momentum is conserved, by resolving along and per pendicular to the Tine of motion of the incident photon, we get E 4 Ecose = poss E'.. 7 < sind = psina Squaring and adding these equations, we find + EC? + 2EEcosa = chp? Since energy is conservedae E + met = E+ cy(p? + mecty Elimination of p between the last two equations now gives the result stated. It is well-known that the frequency v of a photon is related to its eneray £ by Einstein's formula E = hy. Since av = c, we deduce that X= Ac/E and the result just obtained is therefore equivalent. to aye “P= A) = 1+ cosa = 2sin*ge Since AA =A" - A, this now yields the second result. No.31. Let P be the original momentum of P and p, qy the momenta of P and’g respectively after the collision. Since momentum is conserved in the direction of the original line of motion of P and perpendicular to this direction, we must have ° ° P= peosa0? + qcosa0' psin30° = g sin30° Hence, P = a3p and'p = q. Conservation of energy requires that 24 ante? 2. 2 4 net 2 meet ey(PP+ Amc?) + aye ci(o? + mc) + ov(q? + nect) Solving these equations, we find that p= q = vI5m,c, P = 3/5m.c. Then, if v is the original velocity of P. 2mv/v(1 - v#/c?) = P = 3¥8m,c. Hence, v = 3/5c/7. No, 32. Resolving momenta in directions perpendicular to the lines of motion of the photons and using the principle of momentum conservation, we derive the equations AY Exes, ers aepexysinas = Bsin(ars), mv o hs vers weeysin® m,being the rest mass of the original particle. Conservation of energy requires that Bsin(wie) verswrey * ft Eliminating £,, £,, from these equations, we find that © _ sina + sing _ 2sin}(ata)cos: ¥ " “sin@utsy ~~ Zetnk(ata)eosd (ars: cosjocosss + singasing® . 1+ tanjotanzs eoshacosss ~ singastn: Te tanistants This is equivalent to the result stated. In the special case when the original particle is a photon, v= ¢ and tanatanié = 0. Only a = @ = 0 permits conservation of both energy and momentum tn this case.~ a2 - No. 33. Since momentum mist be conserved in a direction perpendicular to ‘the neutrino's line of motion, the lines of motion of the partiches m, must be equally inclined at an angle @ to the line of motion of the neutrino and all Tines of motion must Tie in a plane, Then, if p is the monentum of each particle m, and £ is the energy of the neutrino, momentum 1s conserved slong the line of the neutrino's motion if E/c = 2pcos® = 2p/3. For conservation of energy, we require 2 4 Vip? + mee? Snjc* = E+ 2ov(p* + nfc*) Solving these equations for E and p, we find p = Sn c/4 and E = myc. 18 v is the speed of either of the particles m,, then nv ‘0 = pe ETRY = P= ae Hence v = 3¢/5. No. 34. Let © be the angle through which the line of motion of the part- jcle is deflected and let p be its new momentum, All lines of motion must lie in a plane, The original momentum of the particle is im c. Momentum is conserved atong and perpendicular to the original Tine of°métion provided ange = pcos® + jmyccos60°, —psing = 4m,csin60? Jt follows that p= v7m,c/4, tana = 73/5, the original energy of the particte is Sm,c7/4. If mf 1s the new rest mass of the particle, energy is conserved 1f Bogct = nye + ov(pt 4 mgtct) Substituting for p and solving for m), we find my, = 3m,. If v is the final speed of the particle, my o - 4, vee wyery “P= “eRoe Thus, v = v7¢/4, No. 35. Let p, p' be the magnitudes of the monenta of the proton and electron respectively and let E Se the energy of the neutrino, The equations of conservation of momentum and energy are p= pl +efe fo) 2. 2) 4 (T+ mec? : myc! E+ (1 + myct) + (T+ mec) (ii) where T, T' are the kinetic energies of the proton and electron respectively. Thus T+ mict = cy(pt+ mc?) Git) Ti + mec? = ov(p!® + mc) (av) ES E= 43 - Eliminating E between (1) and (41), we find ols (m= mie-T = pe pra dl sme «) Solving (411) for p gives p = /(2m,T + T?/ct) and, hence, equation (v) is equivalent to e(pt +k) = TI # mc? Using (iv). this can be written pik = v(p'? + mgc*) ‘Squaring and solving for p', we get pio (mbc® ~ k#)/2k Substitution for pt in (iv) now yields Tos e{mee - k)?/2k No. 36. Let 2@ be the angle hetween the photon tracks and let p be the momentum of the recoiling particle. Since each photon hes momentum 4m,c, nnomentum is conserved provided P= dmc cose For energy to be conserved, it is necessary that 2s 2 2 mBe?, mgct = amc? + ov{p® + nfo? 16) Hence, p = ¥n,c/4 and, thus, cose = v3/2, i.e, 8 = 30°. If v ds the velocity of recoil, any vers every = P= FQc/4 It follows that v = ¥3c/2. Take § to be the laboratory frame and S the frame moving with the recoiling particle, such that the x- and x-axes are in the opposite direction to the particle's motion. For one of the photons, v, = ¢ cos30 » ¥3c/2, Vy = ¢ cosa. Hence, by equations (15.11) with u = -v3c/2, ve cosa = Sy i.€. cso = 4¥3/7. This is equivalent to sina = 1/7, x No.37. , ‘The rest mass of the nucleus after emission of the photon is M=Ej/c? and the magnitude of its momentum must equal that of the photon, viz. E/c. For conservation of energy, it is necessary that ef EDD wet = cael + (ue Bolte fe 7c. Solving for E, we get the result stated. No. 38, Let p be the momentum of m, before it absorbs the photon and P be its momentum afterwards, Then, if a is the angle through which the line~ 44 of motion of m is deflected by the photon, resolving atong the original lines of motioR of particle and photon, we find that momentum and energy are conserved provided p= Peosa, E/c = Psina, E+E = cy{Pt + Nict) Also, E = cv(p* + mc). Eliminating p, P, a and solving for M,, we now get the result. stated, ° No.39. Let a be the angle made by the final line of motion of the particle with the x-axis and let p be the magnitude of its momentum. Momentum is conserved in the x and y directions provided E/c = peosa, E/¢ = psina Thus, « = 45° and p = y2E/c. Conservation of energy requires that Etmet = E+ ov(p? + Mict), where Mis the final rest mass of the particle. Solving for M,, we now get the°result stated, No. 40. ] AWE. ¥ Let v be the speed of each fragment in the laboratory frame S (see figure). Momentum is Clearly conserved, Inertial mass is conserved provided amy/4(1 = v8/c#) = 3hn a) ‘0 3 is a frame moving with one of the fragnents (see figure). The components of A's velocity in 5 are given by veos60°s yf TH coseryer = TF VETTE = MUL =v? fc? )vsin6O® | ¥(1 ~ v?/c2)/3u/2 vy See cae See Thus, if a is the angle made by the Tine of motion of A with the Keaxis as seen from S$, then cota = VV, = B/v{T = vey) & VBR by equation (i), This is the result asked for, No.d1. Let £,, £, be the energies of the photons, E, referring to the photon whose motion is in the same direction as that of the positron. The energy and momentum of the positron are Smc*/4 and amc/4 resnectively. Hence, energy and monentum are conserved provided- 45 EL +E, = me? + Smce%/4, £, - F, = ame*/a ‘These equations now give the values for E, and £, stated, No.42. For momentum to be conserved atong 2 line at right angles to the positron's motion, it is necessary that the photons should have equal momenta and energies; let E denote the conmon energy. Then, momentum is conserved along the line of the photon's motion 1f p = 2Ecosc/c. For conservation of energy, we require that cop? + mc?) 4 me? = 2E Eliminating E between these equations and solving for p, we find psinatana = 2mc. If « = 60°, then p = 4mc/3, Thus v = 4c/5 is the positron's velocity. No.83. Let p be the momentum of the moving particle before collision and let q be its momentum afterwards. Let r be the momentum generated in the stationary particle and let 9 be the angle made by its line of motion with the line of motion of the incident particle. Then, monentun is, conserved provided Pp = gcos@ + rcosp, 0 = qsind - rsing Elimination of $ yields re = pt + q? ~ 2pqcoso (a) The KE of the incident particle before and after collision is given by T = cv(p? + mot) - myc? TY = cv(q? + wfc?) ~ myc? It follows from these equations that phos QmT + Tet, gh = 2m! + THe (i) For conservation of energy, we require that 2s 2 meet Tt mc? = Th = ovr? + mgct) (itt) Squaring (411) and eliminating p,q, r, by use of (4) and (11), we get TT Qmgc?} = V{TT'(T + 2myc}(T! + 2m,c*) Feose Cancelling the factor viT'(T + 2m,c*}}, squaring and solving for T', we now obtain the result quoted. ° No.44, Defining the motonta p, q, r. as in the previous exercise, equation (i) can be derived as before.’ Energy is conserved provided E+ mc -&' = cv{r? + nbc?) Squaring this equation and eliminating p, q, r, by equation (i) and the equations cept = came, cfg? = EM antec? we are led to the result stated.- 46 - Let © be the angle asked for, E the energy of the neutrino and p the momentum of the muon, For conservation of momentum and energy, it is necessary that 4n,c/3 = peoss, E/c = psind, Sm,c#/3 = cr(p*+ 4n2c?/9) + Thus, p* = T6nfc?/9 + E*/c* and substitution in the last equation gives Smjc2/3- E = /(E% + 20m5 08/9) Squaring and solving for £, we find & = m.c?/6. Hence, tané = 36/4n,c? = 1/8. No. 46. To conserve momentum, the fragments must,have the same speed v in a Yaboratory frame S. To conserve inertial mass, it is required that am, 'o m= erg This gives v = c/(1 - 22), Taking 5 te be a frame moving with one of the fragments and using the velocity transformation equations (15.11), we calculate the speed of the other fragment in S to be av. Bev(1 = 92, reer =e If his small, the binomial theorem permits us to approximate this speed by le le el - PR Phe eRe ep ed = ca-b) to O(A"); i.e, speed is less than c by a fraction 4*/8. No.47. Let p be the momentum of the positron and E,, £» the energies of the photons associated with the angles 30° and 90° respectively. The equations of conservation of energy and momentum are: ou(p? + mec*) + mct = E, +E, p = E,c0s30%/c, &,/e = E,sin30%/e, & Thus, E, = 2mc°. This gives E, = mc* and p = /3n,c, To have momentum Vane, the po8itron's velocity must Br v3c/2. Eliminating p and £,, we find 2 4 me! 2 GEE + ue") + me! fo. 48. The photons must have the sane energy E for momentum to be conserved. Then, if p is the positron's momentum, equations of conservation of energy and momentum are cup? + mic?) + mc? = 26, p= 2Ecos60°/c = Efe Eliminating p and solving for E, we get E = 4m,c#/3.- 47 - Let a be the angle through which the nucleus velocity is deflected and let p be the momentum after deflection. Before emission of the photon, tite nucleus has enercy Sm,c*/3 and momentum 4nc/3, If u, is its rest mass after the emission, energy and momentum are conserved provided Bay oe 2 4 ye dant Saget = ov(p® #8) + dye frye = ymye cos60? + pcosa 0 = nye sinoo® - psine The last two equations show that tana = ¥3/7 and p = ¥13n,c/3. The first equation then gives uy = m/¥3. _ 1f the frame S moves with the nucleus before deflection, its R-axis being in the direction of motion, let @ be the angle made by the Photon's track with the X-axis. S$ is the parallel laboratory frame. Then, for the photon, v, = ¢ cos60® = ac, ¢ cosd. Since u = 4c/5 is the velocity of 5 relative to S, couse = 25S = He 5 Hence, 9 = 120°. No.50. Equations of conservation of energy and momentum are: cv(pt + mgc*) = E+E, peosa = E/c, psina = E'/e Thus » M9? # mBc2) = (cosa. + sina) Squaring and solving for p, we get the result stated. The results for E and E* then follow inmediately. No.51. If p is the original and P is the final momentum of the particte, equations of conservation of energy and monentun are: 2 4 m2c2) + Smick = 2 Fk cA pF + nbc?) + Fmoc? = ov(P? + Abc?) p+ 3n,c cosa = Pcoss, 3m, sina = Psing. Squaring and adding the last pair of equations, we get re uae? et = pt + mycp + Gate! Substituting for P? in the first equation and solving for p, we obtain p = 3m. Thus, the original velocity of the particle was 3c/5. The value of tang now follons from the above equations to be 4¥2/5. Wo.52. The work done by the force as the particle moves from 0 to a point with coordinate x is X om, 2 * [, Begpe = Bes °~ 48 - Equating this to the increase in the particle's energy, we obtain mc? = mc? = 2mcx/le - x) ‘ where m= m//(1 - v@/e*) isthe particle's inertial mass and v is its velocity. ° Solving for v, we get ax 5 2cfax)? Ye at atx It now follows that aatct = feo sxjex = ant + Saat the constant of integration being zero since t = 0 when x = 0. This is equivalent to the result stated, No. 53. Resolving the equation of motion (17.1) along the x- and y-axes, we get aryfdt = 0, dpy/dt = £ Integrating under the initial conditions p, = py py = 0 et t= 0, these give ? yt ery = Por Py * — ay > ft (i) Squaring and adding, these lead to mv me ach = phe Fatt ewe = revere 7 mo = Py te which is equivalent to the equation mc? erstaeery + Mivg + Pett?) Equations (i) can now be rewritten in the form 2 ye He Pot oY = aaltit x at We crty* it ¥ 5 ¥ Pre Integrating under the initial conditions x = y = 0 at t = 0, we now get PO tne lvoe + Pet x = —g- sinh” *(fet/wy), oy = gly(wg + f%c?t?) - wot Elimination of t between these equations now gives the result stated. Ro.54., The particle's equation of motion can be written = m1 = scr) Se = omev/k or dv aor tyeye © vAtrk Integration {change the variable by 1 - v#/c* = u®) under the initial condition v= dc/5 at t = 0 now yields~ 49 - WLC ~ qog2 - (1 - vivetyt +§ Putting v = 3c/5 into this equation, gives the result stated. No. 55. In moving from the point x = 2 to a point with coordinate x, the applied force does work m,c?(2/x = 1). Hence, the equation of work is 20) yéfety"be mck = mekl2/x ~ myer = viyet)# - met = mc (2/x - 1) Solving for v*, we obtain veo 4e(4 = 98) This is the equation for simple harmontc motion of amplitude 2 and period 4n/c. Since the particle's inertial mass becomes infinite at x = 0, a complete oscillation could not, in practice, be accomplished, No. 56, Using the equation of motion derived in No.22 above, we obtain lon = - ov where p= (1 - vé/et)"2, If y ts the distance moved by the ship, then dy/dt = v and the equation of motion can be written in the form Ho sa or ape trerym = ay Putting x = (1 ~ v*/c#)4, the last equation transforms to ax wey oy Integration now yields the result stated. No.57. Since P and ¥ are 4-vectors, al1 the expressions quoted are invariants and have the same values in all inertial frames. Take S to be an inertial frame in which the observer is at rest. Then V = (0,0,0,4c) and, hence, -PY = =P = -ikl = F since P, = if/c, where E is the eneray of the particle in S. This proves (i). ~ F/c®, Tt now follows that Since B = (ByiF/c), then P? = F MF + (PUA = OP proving (ii). Riso, CVE + ABR/(B.V)2} = Vfl + (ctpt - BY/EY) = DE But B = Ti and E = Wc, so this expression finally reduces to V, proving (ifi). No.5, ‘The work done by the force as the particle moves from 0 to a point x is x - { tagtatxtx 2E(c2 = uta? + ut?) 2 = (c2 = ota?) 8 o= 50 = Equeting this to the energy increase, viz. me nae? voreerey > vere Bearreay and solving for v?, we derive the equation v? = u®(a - xt), This indic~ ates that the particle's motion is simple harmonic with centre 0 and amp- Vitude a. 0.59, As the particle moves from 0 to x, the force does work xed } ctx af is x ate = me <1 i, +e . ii + oxi) (Change the variable by u= 1+ 2x?.) Equating this to the energy increase, viz. mot(t = v8/e2)8 — mot and solving for the velocity v, we obtain van & ot eed Taking the reciprocal of both sides of this equation and integrating with respect to x, we then get ct = asx the integration constant,being zero since x =9 at t= 0. This is a quadratic equation for x? whose positive root is xin We ct} 7 Squaring, we find the equation stated. No.60. If v is the particle's speed, since v* = r* + r*8, equation (20.4) shows that the Lagrangian for the system is Los = myctt - (3 # r*6ayyety? =v (r,8) are the generalised coordinates of the particle and (F,5) are the generalised components of its velocity, Thus, its Lagrange equations of motion are dfal}_ a Lg dfalh a 2 9 ‘aR; ~ Or > dese} 7 VE Substituting for L, we obtain the equations stated. Since the energy € = a,c*(1 ~ v?/ct)"# = mcty, there 4s an energy integral ngty+V = ¢ «) Also, integrating the second Lagrange equation, we get Thus, using (11), where u = 1/r. Hence-51- and it follows that: 2 we we) +e scot = Gad! oe making use of equation (i). Differentiating the Tast equation with respect to 6 and cancelling a factor 2du/d, we now get the second equation stated. (Note: ¥" = dV/dr.) If V = -u/r, the equation determining the orhit takes the form Uy aye yeymehece ag tu eC /mph?ce By appropriate choice of the line @ = 0, the general solution of this equation can always be expressed in the form v= Acosna + uC/(mBh?c?n?) where A is @ constant of integration. This last equation can then be written fu © 1+ ecosné where 2 = mbhe’n?/uc, @ = ALS If n= 1, the last equation is the polar equation of an eTTipse with eccentricity e and semi-Tatus rectum #, If y is just less than unity, the orbit will be approximately an ellipse, but will not ‘close! exactly, i.e. the periapse at © = 2r/n,will not coincide with the periapse at 6 = 0. Since 2n/n = 20 + mu?/(mch?c*) approximately, the major axis for a second revolution of the orbit will have rotated through an angle wu*/(mih*c?) relative to the major axis for the first revolution, Tt is shown in texts devoted to Newtonian mechanics, that the orbit for a particle moving under a central force of attraction f per unit mass 13 determined by the equation wolfe) = ¢ Tt follows that the special relativistic equation is the same as the classical equation if we take f= (C-V)V"(mect, In the inverse, square law case, ¥ = sl/r. this gives f = a/r® + B/r®,° where a, 8 are positive constants, f.e. the relativistic effect is’to supplenent the inverse square Jaw attraction by a small inverse cube perturbing attraction. In section 53, 1 is shown that general relativity theory leads, in similar circumstances, to a supplementary inverse fourth power tern.” This, also, causes the elliptical orbit to rotate, but through an angle 6m*/mh*c* per revolution, i.e. six times the special relativistic effect. Neefless to say, it is the general relativistic effect which is confirmed by observation of Nercury's orbit, No.61. ‘The equation of work is af Me? ay VE Integration under the initsal condition v + 0 at t = 0 yields = 2 = fy = mike’ (l= vyety? = 1ekt = seco-52- Whence, v = csing, Differentiating seco = 1 + kt, we find secd tana 8 = k Thus . ing - y 2 ML Ot, x csing = v = = 8 = keoss cote This leads to the equation ¢ c dx/da = ftan?e = f{secta - 1) Integrating this equation under the initial condition x = 0 at 8 = O (i.e. = 0), we obtain the equation quoted. Ho.62, The first result follows from differentiation of the identity Also, differentiating the product gv, we get Aw = Lstviv + ab vede(av) td Lent + avd Btw Urltiplying this identity by m, and using me |, we obtain whiny) = mgrvh = fic? eget) = my Since f = d(my)/dt, this is the equation of work. No.63. The work done by the attractive force as the particle moves frot x = a to the origin is Poagte de = sagt Equating this to the increase in the particle's kinetic eneray, we are led to the equation mgc2(1 = vse)? ~ myc? = bmguta®, where v is the speed of arrival at 0. Solving for v, we get the result stated. No.64. If £ is directed along a normal, f and v will be perpendicular and, hence, f-v = 0, From the equation of work we deduce that fi = 0 and, therefore, ‘Vv TS constant. ‘The equation of motion of the particle is d(mv)/dt = f and, since m is constant, this is equivalent to Resolving this equation along the normal (we are assuming plane motion), we get mv* = f as stated.-53~ . In the case of circular motion, « = 1/a and the equation of motion gives my gers weary = 0 Solving for v?, we obtain the result stated. No.65. Take $ to be the stationary frame and 5 9 parallel frame moving with the nucleus. Suppose the nucleus moves along the x-axis of S with speed u and that the electron moves in the xy-plang. Then, the velocity components of the electron in $ are v, = 4c cos 60" = 3c, v, = be sin60” = Y 30/4. Hence, in S, the electron has velocity components ze Sg Ge Mls ut/et)v8e x > Weta ly * w Since we are given vy + Wy = 4c®, an equation for u can now be constructed and its solution proves to be u = 8/17. Ho.66. The work done by f as the particle moves from 0 to a point x is x dx anc aay = mystiv(h #9) = ° Equating this to the increase in kinetic energy, we get mc ‘ 2s met. . yorarrety Hye = meV + x) =D Solving for v, we find : yet. cS ew An integration then yields cts peta + xb = jo + cosh26)d2 = 6+ dsinh2e since t = O when 9 = 0 (i.e. x= 0), This is equivalent to the result stated. No.7. ‘The particle's equation of motion is d Tov am Ry eye] 7 Tere which is equivalent to the equation (1 - v¥fety tay =~ omdt Integrating over the interval from v = 3c to v = 0, we calculate for the time taken omyt = rc/6 No.68. In an inertial frame in which the fluid at the point under consideration fs at rest, ¥ = (0,fc) and equations (22.19) reduce to~ 4 = Tyg = Tags Tag = Tag = 0+ Tay =~ C'Mgg? It ts now easy to verify, by direct substitution, that the equation T,,V, =~ cfuggly is valid in this frame at the point, Being a tensor equation, this is sufficient to estab- lish its validity in al) inertia? frames. i s-ch - Expressing the equation Ty3¥; =~ cipgl,, in terns of three dimensional quantities using: equations (18.5) and (22.19), after cancell- ation of a factor (1 - v2/c2)"” we get 1 1 Og%g * Tag * eTalyal¥g * foluvy + Lavgte, Mie = -ctugyy, Since vgvg =v", this last equation is equivalent to = ve) tu, i 1 . eM G%¢0) = Hoos * BoB’ Equation (22.16) now leads to the first result required. Similarly, the equation Taa¥g zee Hoo’ can be written GyMy 7 PU = CH giving the second result, The stated formula for Tye is equivalent to 1 fag = 1 V8/07) vag # tog + ha = /et)te vy, ayy 8 = Gly # top having used the equation for g,, just derived. This is the first of equatfons (22.18), The stated forma for T,4 may be written Tat (1+ vise Tuggvyic + (1 = Viet) Pt pvgic/c® icay which is the second of equations (22.19). Finally, the stated formla for Ty, implies fet) 1 Tyg = VELEY ECs Chigg = at aa%aNg) But, equation (22.16) shows that dateaYs%y = Sava NE = PGE agg) Hence, Tgq =~ cu, which 4s the third of equations (22.19), Thus, the stated equations are equivalent to equations (22.19). No. 69. The components of the energy-momentum tensor for an ideal fluid are given by equation (22.21). Taking axes through 0, since the flow is radial, the components of the 3-velocity flow vector are given by v= vx,/r. Thus, the first of equations (21.20) can be written- 55 - Foo ¥ w/CE)(T = v8/et)WEXE/r® 4p) + ((ligg + B/CE)(T = v8/08) HV? x Ky /r*) + Fx, {oo + ple*)(1 = v8se®) vx x 2} 0 Since uggs P and v are functions of r alone and ar/3x, = x,/r, etc., this equation is equivalent to LU igg + p/P) ~ vEsc8)NA/r#} + 1B 4 aluyyy + v/ct) (I ~ vesce)“AVE = 0 “Gr (gg ‘ar * AMMoo having cancelled a factor x/r?. The cases i = 2,3 of equation (21.20) lead to the same result, Introducing X as defined, the last equation can be written. rihow/r) + rb + ave = 0 which is equivalent to the first of the stated equations. Taking i = 4 in equation (21.20), we find Plug, + 0/641 ~ VEEP HIT + Bhalgg + pict) (1 - varety HY = 0 or a or aso The last equation is easily salved to give art = constant a) It follows from the other equation of motion that 2 dp Mgg®? + PY ys ee ‘This equation is atso readibly integrabte and leads to 2 - 2. ye i Hype? +P = Av(cR = vA}, (ii) A being a constant of integration. Thus, , and it follows from (i) that mete kh (38) where B is constant, Clearly, v>0 as ree The stated boundary conditions inserted ito equations {71} and (iii) now permit A and B to be calculated and the results stated obtained. No. 70. Since the rod is at rest in the first frame S, equations (22.19) give(ag) = (tag) = F/R 0 Of, Tyg = 08 Tyg = ~ic®/A 0 oO 0 O00 This verifies that (145) {s @ diagonal matrix with the éiagonat elements stated, If 5 is a parallel frame moving with the observer, the Minkowski coordinates in the two frames are related by the orthogonal transformation H, = ayy» where wy (2,3) = [eose 9 0 sing] o 100 o 0 1.0 sina 0 0 cosa and tana = iu/c. Thus, 24 * Bqi2aglag = SInta(-F/A) + costa(-me#/A) Arye Suppose fi is the inertial mass per unit Tength of the rod as observed in S. The cross-section in Swill still be A, since the Fitzgerald contraction is longitudinal only. Hence, the mass density in S is ji/A and, thus, Tg = -c*R/A. It follows that ae: wrfct TSE i must be posttive and this requires F Faiaglas * Saget Tye = Aqitaglig = Tea Equations (21.14) show that these equations are equivalent to Iyg, Jy = AygSgr HOT be proving that g,,, 9, and u transform as 3-tensors with respect to a simple rotation of axes.- 57+ Ko.72. In $°, the fluid is stationary and the components of the energy nomentun tensor aré given by equations (22.19) to be ° - - a The et Tag 7 dg 7 Om Tea = Yoo Since wy ‘is symmetric, these equations show that Og is symmetric, Minkowski coordinates in the frames $ and S° are related by the transformation equation x; = a; jx°, where ayy = ayy = Cost, ayy = ~Og, = sift, a, = aj, = 1 and tana = fu/e (non-listed a;; a17 vanish). Thus, the components of the energy-momentum tensor in $ are given by Tra = Haaagthy = Thyeosta + Thysin?a 1 * Hoot” Tye Tye = arydesty = Tfocosa = C1 = users Af M3 ° > Those = 1 wey . Ta > = (19 = 12,)sinacos = FLAS too iu Ty = = TSyeosa = (1 - uber hk, Tee = Th + he 123 = Ts = ths Tyg © = Tysina = (1 = utyet) eS ture ty = Byeose = (1 = utyety"hed, Tag = = Th * he 133 = = 13 * Tas Tey = = Tsing = (1 = utsety Ash sure ‘a * = Wysinta + TBycosts fc Nog T= ee Since T;, is symmetric, the expressions for Tyys those for Trg, Togs T3q- 'ygs gg are identical with In S, the components T;; are given by equations (22.19) to be Ty some (U/C Te tte Tg * ta Tyg t teat + yy fet)~ 5B ~ Tar = tart els typ = ype Tg Tage Tyg = TUtyg/es Typ = ty # Ue lcs 133, tap ty: Tyg = tug glee Comparing these expressions with those already found, the remaining stated equations all follow without difficulty.EXERCISES 4 No.l. The orthogonal transformation relating the Mnkowski coordinates ‘in the frames S$ and S is given by equations (5.1), where tana = iu/e. Thus, the corresponding transformation equations for the 4vector J are: ‘J, = dycosa + dysina, oY, © dey Ty = = dySina + dycosa, Ty = dee : i = Goat =a utetyt - Since J = {S,sdyedqs¢0)s T= GoTyeJ,.100), cosa = (1 - ut/c*)*, sina (1 u2/c?)"#iu/c, substitution now yields the equations required. No.2. Differentiating the Maxell equation with respect to x;, we get Yori Pai 7° since the terms of the sum are either zero (e.g. Fy, = 0) or cance} in pairs (e.5. Fiza1 + Fex,iz = 0) due to the anti-symietry of F; 5. No.3, TP Fig = 95 4 = 8) p> them Figg * 5457 a 453 = Body by equations (26,12). Also Fra t Fitga * Faas sik * Pegi since 25 54 = Qj jr etc No.4. (i) F, Fs is the sum of the squares of the elements of the 4x A matrix (26.543 dz. 2 4 Bt 4B?) - OER + £2 + EB)fc® = 2B? ~ 2E2/c? 2(BE +B} + 82) - Q(E% + EY + EB)/e 2B? - 287/c: Since B = ugH and c? = T/e,u,, this is equivalent to the result stated, Also, Since F;jF,; is an invariant, the deduction indicated follows. (ii) The summation is to be carried out over the 24 permutations of (1234), Stace both F;; and € 5,, are anti-symetric with respect to all pairs of indices, the 8 permutations (1234), (2134), (1243), (2143), (3412), (3421), (4312) and (4321) make the same contributions to the sum. Thus, we need only calculate the contributions of the permutations (1234), (1324), (1423) and multiply each by 8 to give Si seaFigFen * S(CroasFiaFan * GrasFasFan + GanoFisFes! Bi(B,E, + BE, + BE)/e = ~ 8IE-B/e Since 4 syoFjshkq 18 @ pseudoinvariant, we now deduce that E:B is also one.~ 60 - No.5. If a is the angle between & and Hi, then [S| = |E x Hi[ = Eksina. Hence = S? = S? = E*H?sin’a = E2H* - EPH*cos2a = BAH? - (EsH)* Using equation (29.16), we now find UP = S8/c% = 2(eSE* + WgH?)? - egug[E*Ht - (E+H)"] = (egF - ugH)® + (EB)? ey/ity But, in the previous exercise, it was shown that yHl® - cE? is an invariant and E-B is @ pseudoinyariant, It follows that (E.8)? is an invariant and, hence, U* ~ S#/c* is an invariant. No.6. Let 5 and 5 be the frames of 0 and 0, with the origin of 5 moving along the x-axis of S with velocity ¥. In S, E + (0,£,0) and B = (0,0,0). Thus, equations (27.3) and (27.4) show that in § the Field has components Bos Bs 0 B= = evesct, tt Es T= ee where 2 = (1 ~ V/ct)2, Since ¥ = (¥,0,0), thus ¥ x E = (0,0,8VE). Hence ¥ x E =~ c®B, This is a 3-vector equation and so is valid for all paratlel right-handed frames moving with 0 and 0. No.7. Using equation (29.5), we Find HoS45 = FanFaic 7 #5ivFkaFkn * 2 since 65; = 4. No.8, Substitution of the components of E and B into equations (27.3) and (27.4), we obtain Ey = 8(1 - w/eja sina(t - x/e) = da stautt - x/e) B, = 81 - we} sinw(t = x/e} = a2 sina(t - x/c} a1] other components being zero. Using equations (5,8), we find E- He = Bll + u/c)(t ~ x/e) = (Et ~ x/c}/r Thus, (t - x/c) = A(E - ¥/c) and the results stated now follow. Thus, an observer moving in the direction of propagation experiences a reduction in amplitude and frequency by a factor A. No.9. We can write Wo= clip, = eA,)(o, - ed) + aact}# sep = oR ¥ ep and then Hamilton's equations are~61- B= apy = ay), Gy By = @cR*(Dg = OR,MAg 4 =o (ii) From (1), we deduce that = Rhy = ER, = eR, ek) = oC) = mbctyRe) where v is the particle's speed. Solving for R, we get aye - 0 ia Roe aey Gi) Putting X, = v, in (i) and eliminating p, between equations (i) and (ii), yields & o afr ue - a) + Raeme = pha 7 Aha Using (143), this can be written 4 Tova dre ore = e¥5 (ase ~ Aaya) ~ ye ely x curIA ~ grad6),, = e(yxB +8) which is the x,-component of the valid equation of motion d aim) = e(y xB + £) No. 10, For the proposed soTution thy, yy . Aske pep = 0 provided Ak, Also : kp = > Akjkje™“PP = 0 Ni provided kyky = 0. Thus, equations (26.12) are satisfied with oj = It is clear that the 4-vector property of a; can only be ensured by requiring that A, behave as a 4-vector and k,x, as an invariant. ‘Then, provided we restrict ourselves to transformations of Minkowski franes with no shift of origin, x, will behave as a vector and, to ensure the invariance of kyx,, it wiTl be necessary for k, to behave as a vector, In the case of the wave cited k, = 2nvcosa/c, 2uysinafc, ky = 0, ky = 2iav/c and it can be verified that kk, = 0. The Minkowski transformation relating the two frames $ ans 5 is given by equations (5.1) (replace a by 8, where tang = tu/c). Thus, 2my _cosa - u/c “OW = w/e = kycosB + kysing = Ry = = kysing + kycosa = 2thY 7 nucosase~62- Since K, = 27Vcosti/e, Ky = Zint/e, we now derive the results stated immediately, No.3. By equation (28,6), the force acting on the particle is qy xB and the equation of work'is accordingly ~ fic = g(y x B)-y = 0 Thus, m 3s constant and this implies v is constant. Resolving the equation of motion (17,1) along the axes, we obtain the equations | . . a Since the particle moves in the xy-plang, the lagt of these equations is automatically satisfied. Putting v= xX, p= mk, etc,, the remaining equations reduce to x * = oh = ak. These equations integrate to | . Ros uy, fo sox assuming that the origin of the xy-plene is chosen so as to eliminate any constants of integration, Eliminating y between the last pair of equations, we find +u*x = and it then follows that x = Rsinot x where R is @ constant of integration and the other constant is suppressed by appropriate choice of the instant t = 0, We now find y = Reoswt and the particle therefore moves in a circte of radius R. Further, tes ves G24 jt = Rwt and, thus, R = v/w gives the radius of the orbit for a given speed. This result provides the basis for the design of an instrument called the synchrotron, in which electrified particles (usually protons) circulate around an evacuated torus under the control of an intense magnetic Field acting in a direction perpendicular to the plane of motion, Impulses are applied to the particles by local electric fields and these result in ‘increases in speed which transfer them to orbits of steadily increasing radius. Eventually, the particles leave the torus by a tangential exit channei, their speed then being close to c, and the high energy beam is used in experiments relating to nuclear interactions. In $, E = 0 and B = (0,0,8). Equations (27.3) and (27.4) show that in 3, soe = = = Biery hy Bos B20 B= (1 ut/ety * -E: eee utyety Rosa o fy = ut/cty tap, To generate the required field in S, we must choose u and B so that By = (te utvety ts, c= - 1 ute) tas. 0 Hence u = -£,/B, and B = “(85 - Ee/c?). Clearly, it is necessary that 1By| > 1E,i/e. 0- 63 - Using the Lorentz transformation equations (5.8), for the particle motion observed fron S, KR = = utyety Bex = uty = (1 ~ ut/e2)Persinut - uty E = (= utye®) Ect ~ unset) = (1 = ut/ct) F(t ~ uRsinwt/c) The average velocity along the X-axis 4s therefore ye = _bsinot = ut ~unstiate 7 “Uas Eee, since t, € tend to infinity together. Thus, the average velocity ts £,/8,. No. 12, Transforming from the frame $ to the frame 5 by means of equations (27.3) and (27.4), we find = = = = UE/c?! B= 0 By = 0, BL = {8 ~ ub/ct), E50 Ey = AE ~ ub), FE = 0 In, the electric field is clearly null if u = E/B. With this value of u, 8 = B/v(B* ~ E#/c*) and B, = (B* - E*/c2). Since BL = By +8, ‘the magnetic field is directed along the Z-axis, No.13. Since the Lorentz force acting upon the particle is always Perpendicular to its motion, the equation of work gives ti = 0 (see Ex.11 above) and the inertial mass is accordingly constant, The particle's speed must therefore have the constant value U. Tne force acting upan the particle as it moves through the magnetic field is q(v * B). IF(X.ysz) are the particle's coordinates at time t after entering the field at, then v = (x.J.2), B = (0,0,B), and the force has components (qB¥, ~oB%.0)3 Since the z-component vanishes, the particle moves in the plane z'= 0. ‘The x- and y-components of the equation of motion are Helmy = a8}, lay = -a0K, where m= m/¥(1 - u@/c*), Defining k as stated in the question, these equations c8n be written i” uy, KY = = uk Integrating under the initial conditions x = y we calculate that kk = u(y +k), ky = - ux, from which, by division, we find xdet(ytkidy = 0 A further integration under the stated initial conditions, leads to the equation xa ty tke = for the trajectory. This is equivalent to the result quoted. The trajectory in the magnetic field is a circle of radius k and centre (0,-k}. Thus, the particle can complete a semi-circle, as indicated in the diagram below, without leaving the field, provided k < az it then finally moves paralle? to the x-axis and has been turned back by the field, If k >a, the particle enters the region x > a and is only- 64 deflected by the field. Y 2 * No. 14. If the x and y axes are rotated through an angle @ in their plane to give coordinates x' and y', it is easy to verify that the Field components in the new frame are: Et = (0,AK,0), Bt = (0,0,AX/c), where X = sin2nf(t - x'/c). Maxwell's equations in vacuo are satisfied by E' and B' and represent a wave of frequency f being propagated along the X'-axis with speed c. The given equations accordingly specify a wave being propagated in a direction parallel to the xy-plane and making an angle ewith the x-axis, as stated. Transforming to the frame S$ by equations (27.3) and (27.4), we find that EB = {0, 0, GAX(c - ucosé}/c?}, E = {-Axsiné, @AX(c cos® - uj/c, 0}. Hence, if u = c cos®, then 6 = coseco and E = (0, 0, Axsiné/c), E = (-AXsind, 0, 0). Also, the inverse Lorentz transformation equations show that a(E + uR/e) - 8% + ub \cose - Leino (E = F/e)sino t- (xcosé * ysine)/e Thus. X = sinf2nfsino(E - 9/c)} = X, completing the calculation of the wave's components in 5, The electric and magnetic fields in S take a fixed value when E- F/c ts constant. This defines a plane perpendicular to the y-axis, which advances along this axis with speed c. It follows that the wave propagates parallel to the y-axis in 5, The frequency is abviously fsino. No. 15. Choose an inertial frame in which the conducting medium is inztantaneously at rest at the point under consideration, In this frame, ¥-= (0,0,0,1c) and the components J;, F;5 are given by equations (24.7), (26.5) respectively (N.3, ¥ is not the 4-velocity of charge flow). The four components of the steted equation will now be found to be dy = tye J ok, 5, = Es tcp - ico = 0. y Nz The equation is accordingly valid in the selected frame, But, being a 4-tensor equation, this ensures its validity in all inertial frames, No. 16. Referring to equation (26.5), we find that Fy, = -Fa, = Bs al} other components of the field tensor vanishing. It follows that Frealigg = 28%= 65 = Substitution in equation (29.5) now gives WoSia = WySan = FBs — UySas = Sy, = - 284s all other components of the energy-momentum tensor vanishing. No.17, In a frame S with the charge at its origin, the field at (%.Y.zZ) due to the charge is determined by Es 16 Gane gr where F =F aye + The_Lorentz transformation equations relating coordinates and times in S and 3 are Keun Fry F o¥sZ) B = (0,0,0), B(z- vt), EF = a(t - vz/c*), where 8 = I/v(1 - v?/c*), Thus, at t = 0 in S and at the point with spherical polar coordinates (r,€,6}, we have % x = rsin@cos#, ¥ y = rsingsing, Z = 82 = Breosa. It then follows that Y= BPr®(1 - vsints/e*). The transformation equations inverse to (27.4) (after exchanging the roles of the x- and z-axes) show that Whence In?0/c*)"* (x4 42) E> ggeepa(t - v/c*) (1 - proving that the electric field in S at t = 0 is radial and of the magnitude stated, The equations inverse to (27.3) give B, = = BVE/c*, By = BvE/c*, 8, = 0. Whence eke) - v*/C*)(1 = vesin?o/e4)-™ (-y,x,0) 3} 21 = B®), since v = (0,0,v).EXERCISES 5 No.1. The transformation law for B,, con be calculated from that for Ay j> thus: By ys showing that By; transforms as a covariant tensor. The equation Ay; = Ay; may be written A; = Bi;. This is a tensor equation and hence is valid in all frames if it is valid in the x-frame, Similarly, Ajj = -Ayy ts valid in al frames, if it is true ia one, No.2. 2 The relevant transformation equations are x = rsinécos# ros v(x? + y? + 27) y = rsinOsing } and inversely 6 = tanty(x* + y*)/z} 2 = reos8 @ = tan" *{y/x} Since A is a contravariant vector PL Bre gay 4 Bye AP aR a a « 2 , BBgK 4 DAY 4 80,2 As RS OW 4 oe o 2 Bbyx , 84, yz WB 4 ae + Be The partial derivatives are easily catculable from the inverse transformation equations, but are best expressed in terms of the spherical polar coordinates. For exanple, BET re peapeyeesarqatipey ~ pevtheapey * Feosseose Be gett) = gt: x eae ysind and so on. Substitution then yields the equations AT = Rsinocose + AYsinesing + A?cose rA® = A*costcos6 + AYcosesing ~ A¥sind rsinon® = - A%sing + AYcose Since (A%,AY,A7) and (A?,A?,A) are components retative to_rect~ angular Cartesian axes, we can relate them by elementary procedures. Thus,~67- A* directed atong Ox can be further resolved into AXcos@ along ON (see diagram) and -AXsing along 03. Then, AXcos@ along ON resolves into A*cos¢cos6 along 02 and A*cosésin® along Ot. Similarly, we first resolve AY along Oy into AYcos¢ along 03 and AYsing aTong ON; the Tatter component ‘then resolves into AYsingcos® along 02 and AYsingsine along Ol, Finally, AY along 02 resolves into A*cos@ along O1 and -A’sine atong 02. Then, collecting together the various components elong the three axes 7, 08, 03, we find that AY = A%cosgsino + AYsingsing + A*cose = AY Ae = AXcosscose + A’singcosé - AZsing = rA® AY = - Wsing + Acosé = rsinon? using the earlier equations. This exercise demonsfratgs that, in the notation of tensor calculus, the polar components A’, A°, A’, of a contravariant vector are not identical with the Cartesian Components A, A®, AS in the Cartesian ‘polar frame 0123. However, the reader is warned that, in elementary textbooks, the Cartesian components Al, A?, A®, are often referred to as the polar components of the vector, No.3. Making use of equation (33.2}, we find that having exchanged the dummy indices k, 2 in the second term of the right- hand menber of the first equation, This proves that A; ;- Ay ; is a tensor. Vd ody Tf Ae Vo, then Ay = 6 5. Thus, Ars - Ass eh yy 7 5g Oe No.4, Differentiating the transformation equation Ry partially with respect to %*, we find Rye 7 Moa + Se 15k Sad aad ax St” ala aed Cyclically permuting the free indices i, j, ky and, at the same time, in the first term of the right-hand menber cyclically permuting the dummies r, $5 t, in the same way, We obtain two further equations, viz.Ra 7 Gk ol oe Adding the last three equations, we find that the terms involving second order partial derivatives cancel in pairs, thus: vxt oe atx” and yy ax® atx’ =o aetawe ag FS aaa axl TS since Ang + Aa, = 0. He conclude that ras at = ax" aS ax isa ied * ets rs.t * Astar + Aenys) and hence that ys 4 + Ag 5 Such a tensor arises in Maxwell's equations for an electromag~ netic field in the presence of a gravitational field (see equation (56.5)). + Ay; 3 18 @ covariant tensor of rank three, No.5. The chain" rule for partial differentiation requires that. aod _ al 4 axl ag ax” YE It follows from the rule for mujtip}ication of matrices that, if P is, the matrix whose ijth element is 8x'/ax” and Q is the matrix whose jkth element is axi/ox*, then PQ = 1, where I is the unit matrix. Taking determinants, we conclude that [P|.[Q{ = 1, which is the result to be proved, We have the transformation equation mo. SK grax 4 ax? Sot Tf A denotes the matrix whose rsth element is AC, the matrix multiplication rule shows that A= PAQ, Taking determinants, we find a] = IA) JA] = |Pl.JAl. Thus 131 is an invariant. No.6. Since the difference of two affinities is a tensor (see equation (24.9)), ts is a tensor. ie have : koh gah 95 5/PH" = THES ~ THEE cyclically permuting the indices i, J, ks we are Ted to the further equations il yn r Siisi = 29K = Th ~ Mer : Jott. ett Seisg FPP Teed ~ Tyke Adding the first and third equations and subtracting the second, we prove that~ 69 = ee ee Seis * MeHg 7 Sgad ~ PURGD Cy + Mie YT Ge tT Aer Assuming that the covariant derivative of gj, 1s identically zero, it is necessary that a - Seer r BO + Mg) Sie = GRD + Tjek Raising the index i (as s) by multiplication through by g°!, we then deduce that s roger yss 2 $514 Sigh r BOY + Een 4 be Lage * Gee or S48 5 Ly Sigh r a5 te) Ebb t hiae * Tame) But ATfs = ATG 7 Tie $0 that the result stated Follows immediately (with sone changes Of indices). The tensor T}, can be chosen arbitrarfly except that 1t must Clearly be skew-symmetric. Then, the formila for T}, gives T}, as the skew-symmetric part of the affinity and the remaining terms constitute the symmetric part. If the affinity is to be symetric (as in the text), then we must choose T}, = 0 and the affinity becomes identical with Christoffel's symbol. Way now be verified that, with this choice of affinity, the covariant derivative of g,; does indeed vanishs for Sijsk © Stik ~ sbht bt The #985 + Thess] reqs s onl be The t Tee * Theos] S55k ~ [dl - ee “sh 1 Than 7 Teast Bie obs vinich cancels to zero fn view of the skew-symmetry of Ti. = -Tr, No.7, A. 3d A i TijAy r Shit Take Subtracting, the terus involving the affinity cance? when roar, * - Ba Tae TS Biss Apa Ag Ma No.8. Taking the covariant derivative of the second rank tensor AS ai with respect to x*, we get r r Basa = OR, ggf - TE ng — Siige $ . s r ist) TH 7 Tras) 7 These I Asse 7 THs k 7 7 suns Fauna * TuheaSs ~The Exchanging the indices j, k, we obtain- 10 - $ 5 r Y r Waki 7 Tess” aafss They + ghee - ease Subtracting these equations, we find - 8 ras gh yS rye Besa Mog 7 bg 7 Tis Sieg 7 Taste + OKs 7 OA = Bids + Oy > wich is equivalent to the result stated, Since rp; - Th, sa tensor, the Tast equation shows that 875 ,A, is a tensor for arbitrary covariant vectors Ag. Tt follows by the quotient rule that 854, is a tensor. If 8 5, vanishes identically and Tp; = Ay syj and covariant differentiations comute. Sys then clearly A; sdk 7 A similar analysis can be applied to the contravariant vector A’, The result is eat ate ah - phat = Bay AT + ig 7 TD Ae No.9. The caleuiation proceeds as in section 36, The successive equations need amendment. as fol lows = f k (36.2) an = "a j (36.3) orf + rd et ry gh (38.4) Ay + TEAS (36.5) BAS i aK + nr at rien raetack (36.8) oa, = nad ack a tardy to JA fest and ka rd (36.11) aA, crhyrhy + Tks ong =o 4 (36.18) a = Eh eh aay” j ke (96.19) aA, = AORN, TT tT Med = - ap) ae which is equivalent to the result stated, No.0. Let the coordinates of Pt be x’ + ax!. Then, to the first order of stall quantities using Taylor's theorem, if Al denotes the field value at P, the Field value at Pt will be A} + A} paxt, _ The field value after paratlel displacement from P to P' is {Aa + ch alax’, thus, to the first order, i A ThA-n- It now follows that = = dx’ Vinge = WR RE Nik ot * Not. Differentiating the transformation equation for gj44 viz. partially with respect to X", we obtain eotatak rey at ae, ORT a aK a axe” ola ak OFS Cyclically permuting the indices J, k, m, we get the further equations 254 ax” axS ax’ 850 att ax gy ax” ats ead a ae an” aad ox TS "gk axa’ Ors Pha Lax” ax axt Jer aah ax 4 ag ® ad ak at a® tak ax” alae TS having, at the same time, cyclically permuted the dummies in the first term of the right-hand member. Adding the last two equations and subtracting the original equation, we arrive at the result m ax" ax® axe Bex axe ik.mi] rst] g, Deed = Ga age OD + Sige Ge Ors having appealed to the symmetry of g... This is the transformation equation for the Christoffel symbol of the first kind. To raise the index m in this transformation, we multiply by the transformation gr. wat ax? ax! to give fi} a) ax” Ey ot oPRtps.e] + aR! atx" 48 gpa US "ax? ad ak at adie 4 risin?e, gh =1, gt = lft, g'? = i/(résin?a), other components vanishing, ‘Thus 1 y= 1 ofp. av a = ies)» Berndt Bh a] 13 av 1 _ 1 ay Bedell) «pean dls] + stn OF No.13. Substituting for the affinity into equation (36.21), we find i i r, r, (5,8 g Hp) (Sih y+ ARH gd fue = Ot NS a D8 g Alsat a) -2 Zilla te, 3} ws +l + oho yh + oe ws - dle 66 Ov yt = See SY aT tag RR 58.88 ~ Sek 4 A, Sh Pa 7 Set i i 1 oi * Syhak * Soha 7 Sah ka 7 ee-7- If @ = = Tos(ajx"), then a aaa re es " ayx a fax!) Hence ihe 7 Ha = 8 from which it follows that Bey = and, hence, Ry = 0. Ko.20. Putting xt = r, x? = ©, the non-zero components of the metric tensor are Qa? > TEES aryee Substitution into equation (43.19) shows that the geodesics have equations df a? dr] ar far) ‘ =o ds ( (= atj2as! Tz = aly? ids) 7 af or? dof _ 5 abet ards} = 2 (i) We reject the first of these equations in favour of the known first ‘integral (43.11), viz. ar} rt faa) s ~ weep El tote on It follows from (4) that = constant = Ic it) Eliminating ds between equations (if} and (iii), we arrive at the af For a null-geadesic, we must replace s by the parameter A and the right-hand menber of equation (11) by 0. Then, equation (ii) yields immediately «(f) datrt = equation writing k= 1 c?, i.e. k? Mapping the space on to a Euclidean plane by treating (1.9) as polar coordinates, if ¢ is the angle between the tangent to a geodesic and the radius vector r, it is a well-known result from elementary differential seouetry that tang + rdo/dr. Hence, along a geodesic, arecot?y + ar? = ket- Be ° perpendicular distance from 0 is a/k. It follows that the geodesics are mapped by straight lines, In the case of null-geodesics, k= 1 and the straight lines are tangents to the circle r= a. No.21. Note that g,, = 0. Thus, g = 9,,9p, and g3? = 1/g,,, g?? = I/9,55 1 gt =0. The only non-zero components of By yy AF Breyas Boare> Banay> Bygans and these only differ in sign, jk OO Svea Santas ana For the Ricci tensor Rye ay . oat . OB a 0 Rea TBs Oe Pasa! See Rae 8 - ae e4 . SB yon 7 Byoas/Si It now foTiows thet R= Sa, = GPR, + GR = Byaar/(yya2) The tensor equation R;; = ARg; ; is now tnmediately verifiable, component by component, It follows from (42.20) that Einstein's tensor vanishes in such a space, However, the coordinate frame con always be chosen (in an infinity of ways) in any @, so that the metric takes the above form*, with g,, = 0. le concTude that Einstein's tensor is ident~ feally zero ‘in any "a. A space for which fhe Einstein tensor vanishes is called an Einsteif Space. Thus, every, is an Einstein space. * If dy, # 0s transform to fresh coordinates %. Then j Ty = BE o. Se olan “1 Thus, the condition G,, = 0 requires that the two functions x¥(K1,x*), x? (KX?) satisfy one partial differential equation, which can be arranged in an infinity of ways. No.22. aad = as + nha + of alk te rh = Lif then rf, =b 2tva) (see equation (42.5)). Then Jk it Kt “Vd 3k 4. . Ape’, eat Ev} + fbb-79- Aj. LB eyeytd ik 4 2 eayli BA yg yon) + feye = (van!) 's symmetric and x 4s skew-symmetric in ie + Thus Sa BuO CP ps Ae - 2 Bfla. ‘442 12 pygyit 2 (Spot) «ZL 51.§ 20000" since {ii} ax! since the affinity is symmetric and “ is skew-symmetric, But a (42 si) aj V2 yayid (vax | (5X +3 i -(v9x"), ond 9 xt re sao ) ax) 73) ax! = 24-4 a 5s (v9x"5) ax! 18 ox ax" Further la 18 wopity 2 Wg 29 2 Gigi rg) yey (van! oy (gx kode IOI = Ns a) Tt now foltows that x19 45 = 0. iy For any tensor ANS. the tensor B'S = ald - ali is skew-symmetric. Hence 8 55 = 05 from which 1& folTons that ei. gid ig Mag ae after exchange of dummies. 0.23. Writing dx*at = i, Euler's condition that the integral f FO, at A should be stationary with respect to variation of the curve joining A and B, 4s d {& ] ~F 2 4 we are led to the equation hog) Bs- 80 - We use this equation to change the independent variable in (i) from t to s to yield df, ad 1 da axlad lone} 7 TAYE oe ds After cancelling through by F, this reduces to equation (43.19), Thus, a geodesic can be defined to be a curve joining two points, whose Tength is stationary with respect to small variations. For example, on the surface of a sphere (and), the great circle joining two points i either the minimum distance between the two points (the minor arc) or the maximum distance between the two points (the major arc, which together with the winor arc forms a complete great circle) - in either case it is stat~ ionary with respect to small variations and so constitutes a geodesic on the surface. No.24. Ty, 18 certainly an affinity, since it 1s the sum of an affinity and a tensor. i* a hs 7 Tk 4 jpg pha gl inn pt i SAY + SyMy Tok 8} + AS Tk Using equation (36.21), we have ‘2 Bes, 4 gata aatayct. a ata. eal Coie * Site # Shy) (5g + S5Ay + Ay) Tala eeiajat eaten eat Hh # ola e ODOT + 8M, + SFAs) 4 i i i al a Ak TFA Myigk” OFAkan 7 Oey = Rl i vr . i vr L = Baie + EIA TT pete * Age) ~ SCAM TMA AG ed 4 + 85 ok” Ak gal after simplification using the characteristic property of the Kronecker delta and cancellation of terms. Since Rea = Ay 7 Tiel, we can write 5k Jok 7 UGkr . ASR THR SK ‘ik and then ae = ata olan, - lay + sta, | - AL 9) Mine = Bias * Aja > GaAs + 6] Aa ~ Aaya Also Bak ke * ~ Aen = Aca ~ Mak Hence, finally, ix i i i 4 Bigg = Bik, + Ae - SAjK + Bigs, > May) fonecee Mok Them RA A A and A 8 ean to-a- Exchanging j, 2, we have * ‘ Bh & Bag * DAL Since Ay, 1s syrmetric, it follows thet i* ix 4 i Bite > Peig ~ Bye 7 Baty No.25. Transforming from the x-frame to the X-frame, we have fo. Bot atie yal oat dk ant aad ax St” ax? galaxk po. od ok mm ogt gat But ax® Hence | pe a ax! ake a ok att «) a ax? am aah St” aah oat on” aaa low differentiate the identity ax” a ox” oe ax partially with respect to x" to give ath oe Pax” atid ama a" ae ad ae Multiplying through by ax*/ax", this Teads to SP me at at ae aed ax aa ax" ela ant aan This permits us to reduce equation (i) to the form wt. Ra at atx mn 7 gt aR opt SE 7 gh gag Ye have now proved that the 'product! of two affinity trans- formations is also an infinity transformatton.~~ we must put x! = x1 in the last equation, which thereupon reduces to the product transformation & gheSgteh 2 hich 15 the identity (or unft) transformation. This shows that the product of inverses is the identity. ‘Thus, the transformations satisfy all the requirements for 2 group. Nio.26. We have (va)* = vgeve = ag jj BY equations (38,4) and (38.5). vihence Ms 8 way? rey = oa (ve) g Fat Te uh g +b go ge = baka og #290 5) after exchanging the dumies (i,j) in the middle term of the right-hand menber. Choosing the Frame to be geodesic at the point under consideration, then oy and equation (4) reduces to 2° id. = 2gid, sxe) = Pad tik = 288 sak since + 5 = (4) 4 (sgh = Agu 1M @ deodesic Frame. This proves the result in this frame, However, the result to be established is a tensor equation and, if valid in one frame, is true in all. Alternatively, we can continue to work in a genera? frame and proceed as follows: le have ak 7 de 7 ae The 2 Bae oe beds so that i, i 1 G80 seg = 8 Gye ST Bk e]O Differentiating the identity is 4 s Gs . 85 with respect to x gives 15 15 : ik9js TIGjsyk = & Multiplying through by g"S, this becomes aid yah + eligloyg , = 0 or = = ahaha... = = aa" S(Lpas] + [skod]) after using the identity (42.3). Thus a 8p 2 - 9a! aso 49, - oa ke 49, = - 25a" Liks]e 58,y after exchanging the dummies (i.r) and the dumnies (j,s) in the second tem of the right-hand member, Equation (14) can now be written in the form a ti + ir Td re~ 83 - No.27, We nave 5 + e%a55, so that 9 = alg, GIS 2 alMN9Gl ang gid = eg", Tt then follows that Geo = 2G pnt * Ferg” Sake? a BUG * Med” Sakae * Yr * Me5 7 BGP) = 30°C oD 4 Ot OG 7 SRO and thus The 7 TDD + asl + 95,06 * 8S ~ HHP) i i, i = Tae MM Sey ir 995A) i i = Nik AK Since Ae ‘is the difference of two affinities defined in the same x-frame, jt is a tensors it is symmetric since the affinities are symmetric. Using equation (36.21), we catculate that Boe Gh a wtygtaat ial yeh gah Bice = ln * ADO Ge * jg) - Tne * Ang Ojx + AG) i i ‘i i * Pack * Aes Tikan 7 A i ay Pai cyt opal g al Bike tga Mietk ~ Trak * Myibnn + Aree far yal 4 ~ Aaah * Mesk ~ Aikse ig gi i iat. gi ge Shea * Minsk ” Skirt Arie > Anelk a pair of terms Thee - TiAl, cancelling by virtue of the synmatry of the affinity. Contracting, we Find = Bl . i . iigr _ gf ar Bi = Bh Rie * Atak 7 Aika * AneAt ~ Ars4S Finally My = AS}o,y + de - 9458") Boj they Bho) = aN a i Ho.28., Let OX? be the oblique axes and Oxtx* rectangular axes as. indicated in the figure. IF A has coordinates x', %' in the two frames, then Se OP = OM = PM = xt = xPcotwa Suppose A has components (A,,A,) in the rectangular frame. Choose the displacenient vector OA such that the point A has coordinates (AyyAg) in the x-frame, The contravariant components of A transform like the differentiels of the coordinates, viz. Ah = A - A®coty = A,cotw op BR = Atcoseas = A,coseqw + 00 recalling that the covariant and contravariant components are identical in the réctangular frame. ‘These equations show that the contravarfant components of A in the x-frame are obtained by prodecting the displacement vector OA on to the axes, by parallels to the axes. The inverse transforwation equations for the coordinate differentials are axt = Git + ditcosu, dx? = d?sinw Hence, the metric is ds? = (det)? + (dx?) = (RE)? + ARTA cos + (dK?) and_the components of the metric tensor in the obtique frame are 9,y = Gey > 1, Gia = cosu. We can now calculate the covariant components of A in the oblique frame thus: JB = B+ cosuil? om i Ron Ggh = cosukt + # OP + OQcosu oPcosis + 0Q oN These equations show that the covartant components can be found by pro~ Jecting OA on to the oblique axes by perpendiculars from A. No.29. P is any point on the conical surface. PN is the perpendicular from P ta the avie AN’ 4 1c defined fa he the annie hatwean the nlane NPN- 85 - ds? = PPT? = dr? + r?sintadg? With x% = ry x? = 9, the metric tensor has components g,, = 1, +0, 9,2 résin’a, It follows that the equations (43.18) forthe SBbdesies tike the For ar &f{rsineal = 0 We discard the first equation and replace it by the metric as a first integral. The second equation shows that: ds = crédp where cis a constant. Elimination of ds between this equation and the metric yields the equation rdp/(ctr? = sin?a) = de or dr apsina = FpEFe = TY where b = c coseca, Putting r = 1/u, this equation 4s seen to be equivalent to du Way ising This integrates to gsing = cos *(u/b) +B where @ is a constant. It follows that u = b cos(gsina = B) or r = a sec(¢sine ~ 8), where a= /b. If we cut the cone along the generator OA and open it out into a plane, the arc AP will continue to be circular, since every point on it has the same distance r from 0. Sut AP = AN.¢ = résing. Hence, in the plane, the angle AGP = AP/OP = gsinas ‘f.e. (r.gsina) are polar coordinates’ in the plane. But geodesics in the plane are straight lines and any straight Line fP has polar equation r= OP s a sec(@ - 8). Thus» in terms of (re) he geodesics in the plane have equations r = a sec(¢siny - 6). "But these tra ight Tines are curves of minimum distance joining two points and rend‘in so when the plane is rolled back into the cone, i.e. they remain geodesics. This explains why r = a sec(#sina - 8) along a geodesic on the cone.~ 86 components of the Christoffel symbols of the first kind must have at least two indices the same and we calculate that TRO] = - 200q9/aKp = - 26%p bog ipo. provided P #Q. (N.B. Repetition of a capital letter index does not indicate summation.) The restriction P 4 q can how be Feioved by combining the first. and second results into the formula dae [ea.P] = (5g - Bp. The third result is clearly valid when P = Q. 22%qq/2%y = 30g 2, 299pp/%y > Beg . The non-zero symbols of the second kind now follow immediately, viz. foot = omar] = (5h - ap in’ = oT Beth = 2g Thus, suming with respect to i and r, fated . Zed} * & tohb' where we have separated out the terms in which i = r from those in which ifr. Then E hiMeh - z pp = aN If i # r, for a non-zero contribution to the sum ye mist have either i = P orr=?. If i= P, the net contribution is Zed = - 8 z AR to BAR tap vee If r= P, the net contribution is the same. Hence q vr - Bo 2 - EUHLI = mp Ba, eng = ener, ne and -24, ax! of We calculate that~ 37 - Aelp t ANApp fablob = 3, CPs0) + Arp = Bay t App 2x Substitution of these results shows that - 2. ~ am? Ray - Rap 7 AUNZ)AB = BALA, + AMALAL = ANAS + ENApD + BAGG - App = RIRIAS + AAA, + a AGH) App Thus - - eo Ree z Rpp =e DE(NMZ)A,A, + AN(N2)A,A, + ANAL. + B(NR2)AL TT . 4 = (NTe*TE(N-2)A An # Ang No31, cite non-zero components of the matric affinity for the sphere have been calcutated in Ex.17 above. They are oe esi an Tl, = = sindcos6, 12, = TH, = cote Referring to equation (33.7), it follows that a parallel displacement of A; from (208) to. (Geddso406) Teads. to increncrts = tladd = SAL = Py yAyae cotan, de, BA, = Tiyajde! = coten,de ~ sinecos@h, do. Along the parallel of Jatitude 6 = a, do = 0 and the equations reduce to SAL = Aycotuds, 6A, = - AySinacosadé Thus, variation of Ay along this parallel is governed by the differential equations a = A,cota, ge =~ Aysinacose, Eliminating A,. wwe arrive at the equation fat + costal, = 0 whose general solution is A, = Peos(#cosa) + Qsin(scosa) P and Q being integration constants determined by the initial conditions Ay. Mee = at ¢ = 0 (wo choose the zero meridian through the initial point, A, can now be found, thus: A, = Gptana = {- Psin(gcosa) + Qcos(ecosalfsina- 88+ Substitution for P and Q now leads to the results stated. Tf A 4s the magnitude of A;, then at = glans = Ay + cosectenz Thus, initially A® = x? + Y?cosec*a. Finally A® = {Xcos{geose} + Yeosecasin(pcosa)}? + {+ Xsinasin{peosa) + Yoos (ecosa)}*coseca x? + YFoosecta and A is unchanged. No.32, We have Guy = As er, = r*sin?e, all other components vanishing. Thus, eqigtion (#32 ») Yor'the geodesics yields ‘the equations og] - at a - @) Bazar Esyaye We have Byoaz 7 K(Qya920 ~ SazGer) = KR*sin?e which is true with kK = 1/R*. Further, Besos = K(SpaGe2 ~ SzyJs2) = R’sin“asin?g Barsy = K(GeaGua ~ Saadax) = R’sin?osin®g showing the equation is valid in all the other cases with K = 1/R%. No.34, Putting 9.) = sech?y, g,, #0 in equations (43.19), we Hid the geodesics a8 detShminee by H]-s ffseentvig] =o, . Elsen) + scr’ ean The first of these equations integrates to yield + dx 2 Ge = Acosh?y, where A is constant, The mgtric supplies a first integral s + @ = coshty Eliminating é between the Jast tio equations, we Find (Qe) a1 = jasechty-9- where a? =}, - 1, Thus x jae bay = sin*(z/a) + constant or sinhy = z = asin{x + 8) This is the equation of the geodesics. The metric for the spherical surface is ds? = do” + sin*edg*, Now dy = tang0(-cosec*4e)3de = - cosecodo Iso = cotie. = tenho so that coshy = 4(tanz0 + cotye) = cosece Hence sechy dy = - a6 In terms of coordinates (x,y), the metric can now be expressed ‘in the form ds? = sech*y(dx? + dy*) Since the great circles on the sphere are geodesics, in terms of the coordinates (x,y) they must have equations sinhy = asin(x +8) If, as in Kercator's projection, the coordinates (x,y) are plotted as rectangular Cartesians, this equation shows that the creat circtes will appear as sine-waves, somewhat distorted (fiattened) in the y-direction (see diagram). y No.5, Bike = Sie hie © aiff hed dig + audi - 26] ~ em fff eh GSE etal aiseads . ffl] = sedsah © Abel - aisha Bik 7 Geel i} + edd id ~ sss cbiad * Sis seb et s2te ad - 2a But Hence.- 92 - = [sk.i] id ~ [sai] id = (Likes] + (ski) 455} + (fiess] + Bai fed + 2Ged - alee] => [thos] 8) « Livs] {Sd pxlind] having used the identity (42,3). tel hiy . sfs bf at a - - paket + Gis ae * Sins the last identity is equivalent to Since tile? Brie = Stns * Ofhtn ~ hase 7 94e,sk) * seek abs . sorb bf uh ‘The identities (41.5) ~ (41.7) will be found to follow directly from this result, No.36. The components of the metric tensor are gy2= Sx e Thise 9 = -0ry ee gE G, GENES fences gre gas 28,7982 24,5.° The christoffet syubols of the first kind are fu.) = 0, [2,1] = 0, [ee.%) fz] = o. [2.2] = 0, Bez] The symbols of the second kind now follow thus: iit = oD] = 4/6 fiat = g@flzz] = 0 bd = gM, = 0 ft = gf] = 0 het = g@fi2,t] = 0 Ed = etka = ya Then tea PHP Atal a = dea day = Ls= 93 - ‘the retiaining 12 components of the curvature tensor vanishing. _The space is flat provided 211 components of ‘the curvature tensor vanish. This is the case 1 8... *0, Te. iF 66, = by. aol Ye mul =p et ” Tee eel, then a = 29s by = Uele by = Vyytl + lye and the condition for a flat space reduces to dy = 0. For this to be true, wy must be a function of x alone and, hence, ¥ = F(x) + oly) In this case, tg a= oF oho = FX) BY) i.e. 4 is a product of a function of x anda function of y, Thus ds* 2FEdxdy. Now choose (fn) such that we lp Ml g geo y Then ds? = 2tes! tidy = 2dr No.37. It 4s being assumed the affinity 1s syimetric so that (see Ex.8 above) = Ae. Bissk 7 Mpeg 7 AvP hak Cyclical ly permuting (1,J,k) we obtain the additional equation: = Apt. Nicki 7 Aisi 7 APs Nests 7 Aegat > Akay Adding the first two equations and subtracting the third, we Find that Naik 7 Risk * Aaa? i AS tess lst 7 Apeak + Br : a nat iki * Bag) ik which reduces to A, sak Aik 7 Antje * Ba 7 in view of the given condition. But Aisi Ajeak 7 Aisg 7 Asdax = Rica again using the given condition, Also, it follows from the identity (41.1) that r ro ogt Brae + hee” Bea Eka- 94 - No.38. The Cauchy-Riemann conditions which have to be satisfied by the real and imaginary parts of an analytic function are a Bu ww OW Differentiating the transformation equations x - $, y =) we arrive at the differential transformation = Bau + Hav = 3y, Way a - ab 3 dy = Phy + Seay flaw + Hay In the special case $ + iy = 1/w = (u-iv)/(u'4v?), we have Squaring and adding, we obtain the metric ds? = de 4 dy? = [ar- Bu ‘raw + dt) be ot pe et ee ee ree Reve ou WEF av erry and so xe ot = solar mee me 2 + ave ist = ot eat = hte It follows that the transformation from coordinates (u,v) to coordinates (x,y) may be regarded as a transformation from curvilinear coordinates in’ plane to rectangular Cartesian coordinates. In terms of the Cartesians (x,y), the geodesics form a family of straight Hines with equation ax thy = 1, (a,b) being parameters. Thus, in terms of u and v, the equation for the family of geodesics must take the form au- by = ut + vF (i) If, instead, we use the transformation equations u = rcose, v= sind, then du = dreoss - rsinéd#, dv = drsing + rcoseds and the metric reduces to ds? = 4dr? + r-%do* The equations governing the geodesics now follow by substitution in equation (43.19); they are ales] + arsThe second equation shows that d8/ds = Ar?, where A is cinstant, Elim ‘inating ds, we conclude that, along a geodesic 2 falter ae = a Thus, do/dr = (c? - ry"! and integration Teads to ® = sin r/c) He where a is constant, Hence yr = csin(S-0) = ¢ sinScosa ~ ¢ cos@sina = - b sin + @ cosd where a= = ¢ sina, b =~ c cosa, Multiplying through by r, and using the ‘transformation equations, we once again derive equation (i). No.39, The repeated index summation convention is suspended in the solution to this exercise, ‘The transformation equations are x = roos), y= rsine ros Mh ey), 8 = tanty/x) From these, it 1s easy to verify that Fe = cos, BE = sing Re ~ Isino, Be Jeose Also, Ay, = tand + cote = 2cosec2e, Since there is no distinction between covariaft and contravariant components in a rectangular Cartesian frame, we conclude that A* = AMY = 0, As AY = 2cosecte. The transformation equations for a second rank contravartant tensor now show that rr far}Paxx , ofar}fardary , (arity wre yee a Ee gee = 2 By POL Ar BOyxx , Dh BOgKY , OF DEyyx , ar DByyy LZ, Ns ea tae ah tay Beh tay ah = Foote 89 faa}®axx , 9{28} (ae) yxy , (28) aw 2 ARO = (aa) 4 of lay «(EY = = 2, That AT + y2A% = 0, now follows inmediately. Ro.40. Again, we suspend the summation convention.= 96+ ds? = dx? 4 dy? 4 dz” = (v442}du + Quvdudy + (ue2}dv This shows that the. covariant components of the metric tensor on the surface are g = v2, q.. = Uv, 9, = u“#2, The contravariant components follow from th¥se, viz, “Y w gt! = jutszy/g, gtY = = uv/g, oY = (v842)/g where g = (u24#2)(v242) - uty? = 2u* + 2v2 +4, We can now raise the indices of the covariant tensor Ay; thus: va ur gus, Ae Zool R 1 + Lifwraytay, - Wt evuviyy - wenn, UVa I = Leweeyteaye = we = = put (ut + v8 + 2) que = 2A, As Saha . Acuvue), + (UPR) VARDAL, + UPAR, = uv(vE2)AL,] - dauv(ut eva 2ye « - qu = aA, Ze"ahn,, poy = uv(vee2}Ay = Uv(vA2)A, + (v22]2R,.) du yee 2)2 = byt = GA Ww w No.4]. Differentiating the transformation equations partially with respect to x (keeping y constant), we ebtain a sinhu cosy &% - a coshu siny 1 aR rd a coshu siny #2 + a sinhu cosy SY 0 Solving for au/ax and av/2x, we calculate that au. sinhu cos v K a(sinh"a cos*v + cosh" Sin? ¥) Ye coshu siny ‘ax 7 a(S costv + cosh? STHEV] Since sinh?u cos*v + cosh?u sin?y = sinh*u(7 ~ sintv) + (1 + sinh?u)sinty = sinhu + sin?v = 4(cosh2u - cos2v) these results can be written BU ets weneim | OL ern-9o7- ua 4 2Y Ae = atu * a 2(A,sinku cosy - Aycoshu siny)/D. No.42, The curves u = constant are rectangular byperbolae with asyaptotes y = x, The curves v = constant are orthogonal to the first family and are rectangular hyperbolae with the axes as asymptotes (see diagram). Differentiating the transformation equations, we find that the differentials are related thus: du = xdx = ydy, 9 dv = ydx + xdy. Solving for dx and dy, we get (xP yPhdx = xdu t+ ydv, (x? # y?}dy = oxdy ~ ydu Hence, ds? = dx? + dy? = (x* 4 yF)-F(du* + dv’) But OP y?)? = (x? = y2)F Axty? = (uP + v7) so that ds? = d(ut + v2) F(du2 + dy2}. Differentiating the transformation equations partially with respect to u (with v constant) gives ax 2) ax, i) aa ie 8 = vg St Solving for ax/au and ay/au, we calculate that 1 a rr en a cry xe yt? au xe + ye Thus, a transformation equation for the covariant vector Ay is xh, ~ YA = 2k 4 My 2 Ayo ihe * hy = Se Similarly, differentiating the coordinate transformation equations partially with respect to v and solving for ex/av, 2y/3V, We find that ax x ay _ i- 98 - YA, +h, 3x, 3) NOT RAGA, © Saget No.43, Differentiating the transformation equations dx = udu + vdv, dy = ydu + udv Hence, ds* = dx? + dy? = (uPav?jdu® + fuvdudy + gives the metric in terms of curvilinear coordinates. The components of the metric tensor in the uv: found to be wt, og. = au, Suu = Sw uw gt = gM = Cutt vig, gi = - where g = (u ~ v4)?, We can now raise the index of the covariant v Moe gia + ga = (u- v)t Wom GMA e GMA (uv)? The Christoffel symbols of the first kind are Oud =u ft = vy feet] fz] + vy. fizz] - uo, 2,2) (uP sv? av? /-frame are now 2uv/g jector Ag, thus: u v leading to the following formilae for symbots of the second kind go! [11,1] +g? 01,2] = u(t sur =v. flit -ww- #8), §24 A nF = The divergence of A is given by i fl iy i'd rn u/(u? u/(u? = fi iar we My tye somes Sit fie +f -v?) -v), -v), 2 Jat + ie = (uv) = 2(uev) + etigluev)* Gatge(uev)? u ay 2 + ately? ~ ete) = 2(u-v)?/(utv)- 99 - ont wf hatau - LT gaten - fTipetau - paver on” w= fh atau - f2 fata - fh avau - £2 ha%av Along the curve u = 0, ne have du = 0 and these reduce to = ft fwa - fljwter « - datev «Pye - fepatan = - Jara Thus, variations in AY and AY during parallel displacement along u = 0 are governed by the differential equations ant dy a Lay ve we Ge TY The variables are separable in both cases. Integrating from the initial conditions Av = eTatu 1, ¥ = 1, we deduce that = WN = WW Thus, at the point ug 0, y = 25 the components of the parallel displaced vector are given by A =i No.44. In the usual notation Gia = 1s Ga, % Ms Oye = OD ge 1 gait, gt = 0, QT) = [2.7] = D2] = [22.2] = 0, (t2,2@j = - fan = x it * ig = fat * i = 0 teh sms Bt ‘The contravariant components of the vector field are given by AX = GMA = x costy, A = gPAay = = sin2y. The divergence of the vector field is given by 21x te coszy - 2cos2y + cos2y = 0 A gd ite No.4, We have= 100 ~ = ag’? . rp - 20" [tps * Sopyt ” Sstyp) * 28" Sep,sj * Ssp,t5 ~ Sst spa) Differentiating with respect to x’ and ignoring terms containing first order derivatives of the metric tensor (since the frame is geodesic), ve se "p - = Ba (Stp igs + op .idt 7 Ist, isp) as stated. Now oe sty = MEE AL + Seed - S.b%d] «Mb Saft) for a geodesic frame. {W.B. Although the Christoffel symboTs vanish at the point, their derivatives do not.) Also, by-differentiating (i) with respect to x! and then ignoring temis containing first order derivatives of the metric tensor, We calculate that 4 ikl 3? 8 wo Hab seabid] 4 so that we have proved the second result. But, in a geodesic frame, ay i Putting j = i in (i), we get ee USSD SEP ded 2th] Differentiating with respect to x! and neglecting terms containing First order derivatives of the metric tensor, we find i aaAd att] which is the third result. ke now make use of the first identity to show that i ik sp - - - ) Rist = ATT p aks + Isp,iki ~ Ussikp ” Sipsisk ~ Skpsisd * Yjksisp! svikp ~ Skpsisd * Sykeisp) ik sp, . BTSs ,3k5 ~ Skp isi after exchanging the dummies (kys) and (ip) in the second term inside the bracket. Further, Ry = dala, sks * Sepik ~ %s.ikp 7 Sxp.dsk ~ Skp.dst * Sksdsp)= 101 - after exchanging the dummies (i,p) and (s,k) in the second and fourth terms in the bracket. 14 now follows that, in the geodestc frame, R}.5 = 3R Since this is a tensor equation, we can conclude that it is valid in every frame. flo.46. In the usual notation: ghie Wy, g% = 1, gi = 0, [22.1] = 02,2) = [22,2] = 0 fie] + -[1,2) = y fid - fib - bh +0 fab fp ey Ba, = aR, = Tyo baa = Wx For parallel transfer aiong the curve y = y(x)s A, and Ay are accordingly governed by the differential equations Thus, In particular, if y = secx, hy aw, ae Aytamx ~ Aysecx, gg’ = Aycosx @ Thus, EVimminating Ay thus: 4 (osx (4 sinx) ~ A,cosx ELOM alas IPN) = Ay A, dA, dA, x sim oe or cose sing aX = sinx 3 Putting dA,/dx = u, this equation may be written du oe 2tanx dx which intearates immediately to aive102 = where P is constant, Hence, A = Ptanx +0 whence, using the first of equations (i), Ay = > Peosx + Qsinx Initially, A, = 0, A) = 1 at the point (0,1). Thus, P = ~1 Q = 0 and so * y Ay = 7 tanx, Ry = cosx. At the point (1/342), these equations show that Ay = -/3, Ay No.47, The metric for the sphere is ds* = do? + costodg* and equation (43.19) shows that the great circle geodesics are determined by the equations 2 2 GE + sinvcose [fs] = 0, $e{oos*ef] . ‘The second equation integrates to give ds = acos*adp, where a is constant. Eliminating ds between this equation and the metric, we find atcos*odg? = do® + cos*odg? or de, 1 Set do” WlatcaRre = COREG) * VBE = tan7e) where b? ~ 1, Integrating this equation, we calculate that i -* Beane) 8 tanp = bsin(¢ + 8) where is constant, Thus wich is the equation stated if we choose a so that tena = b. Since the point 6 = 0, $= -8 lies on the great circle, -8 can be identified as the longitude of the point where the great circle intersects the equator. Also, since the point @ = a, @ = fr-B, also Ties on the great circle, a can be identified as the latitude of the most northerly (or southerly) point on the great circle. No.8, We calculate that Sax = Gan = sech?y, Gyn = Oy git = 74 = coshty, gi# = 0, [11,2] = sech®y tanhy = - [12,1] = ~ 22,2] rt = tanby = = i'd - yh= 103 = a= Sayed = Spear + fits er A,tanhy dx = Aytanhy dy ca = Pidaad » fades fh a ar =~ Aytanhy dx - A,tanhy dy Along the line y = b, we have dy = 0 and the equations reduce to OA, = Aytanhb dx, 6A, = - Aytanhb dx i.e, parallel displacement along this Tine is governed by the differential equations dh, & = Ajtannd, Ge == A, tanhb «) Eliminating A, we find that A, satisfies the equation a As + aytanh> = 0, whose general solution is Ag = Xcos(xtanhb) + Ysin(xtanhb). X and Y being constants of integration. Substituting in the first of equations (i), we then get A, =~ Xsin(xtanhb) + Yoos(xtanhb) At x = 0, we see that A, =X, A, = Y, so that (X.Y) are correctly identified as the initial components of thé vector. The components of the vector in its final position at x =a are now found to be as stated. Mo.49, In the ¥-Frane, we can differentiate AY covariantly thus: i = Boys 3 + The Substituting from the tensor transformation equations Ry which, after integration, shows that t= Baoglf24 + constant Since r = 0 at t =D, the constant vanishes. Solving for r, we find that r= Rtanh(ct/R) As t+ te, tanh(ct/R) + #1 and so rR. Putting r= ZR in the equation for t, we get t = zlog3. No.10, Equations for the geodesics ares fey rey ley sea «g(a - a) -* Integreting the last three equations, we deduce that wo A dz lB dt ) i where,A, By C, are constants. Dividing the last two equations, we see that Z = B/C and, since 2 = 0 initially, this means 8 = 0, Thus z = constant = 0 throughout the motion, which lies in the quasi-plane z = 0. sled) Putting z = 0 in the metric, we have the first integral afd 2fesy ways fa off Using equations (1), this Teads to * 2 7 aa re i} Dividing the First and last of equations (i), we find b= é (aa) Hence, by the initial conditions, A/C = 73/2. Also, by substituting ERELEDT wintiins fn FEE\ nin antonttade that RE PDT Un ran neue dndiicn- 113 = ‘The last equation proves that i can only be real if le r¢ 3 and the particle's trajectory accordingly lies between the circles r= 1 and r = 3. # can only change sign when r= lor r = 33 it follows that F increases to the value 2 and then decreases to the value 1 alternately. Dividing (iv) by the square of equation (771), we derive an equation for the trajectory ia the form ' e FE r=) faretien = Asin"l(r-2) + const. Since = 0 when v= 1, the constant is 31/2. The polar equation of the trajectory is therefore r= 2 = cos(0/3). For a photon, the equations (1) are replaced by dol A a |B dt lc rN ree dh ro a r and the first integral by If initially z= 0, 2 0 and z = 0 throughout the motion. Thus, Initially, dr/de = *(dt/do) = 6 and r= 1, Hence, At/c® = 1 and we deduce that or = 2e(r- 1) + const. Since 8 = 0 when r= 1, the constant of integration vanishes and we deduce that r= 1 + 40%, No.Ul, Along a geodesic d{,2ct/R dx’ He & which integrate to give dx 2ct/R dy © gen2ct/R dz gg 2ct/R no > apt Be * Ce’ It follows immediately that dy/dx = B/A, — dz/dx = C/A so that 8~ 14 - For the motion of a photon, the argument can be repeated with 5 replaced by A. if a particle is projected from the origin along the x-axis, ‘it will continue to move along the x-axis so that y Thus, the metric first integral for the particle's motion reduces to ecct/R | “(ey ( : Jette i) At t= 0, % = Vand, thus, 1/A® = 1 - c#/¥*, (i) now reduces to Le eyer2et/R ety Wek = VF + vee TR Integrating (change the variable to eR), we Find that Vx = Rfc = v(ct ~ VF + yteT@Ot/R) | since x=Oatt=0 For a photon moving along the x-axis, equation (i) is replaced by = 0 or koe - corclR for motion in the negative sense, Thus, x = ReTCHR 4s const. Tf x =X at t = 0, the integration constant is X - R and x (i) Then x = 0 when -ct/n T= WR or t = = Bog( - wr). Equation (11) shows that x decreases steadily towards the value X- Ras ted, Hence, if X> Ry the photon never reaches 0. The exponential time factor present in the metric, implies that the universe expands at a steadily increasing rate and this increasingly retards the photon's motion in the xyz-frame, uitinately (t = +=) bringing it to rest. (Note however the conments at the end of section 66 regarding the unreal nature of this expansion.) No.12, Equations for geodesics through the space-time are 2 ales) - rAsinocose {if} = 0 “werd) « Sheed] «= Thece tnnather with the matric. detarmine the mation af a freely falling~ 15 - identically, showing that the trajectories lying in this quasi-plane are possible. The initial conditions @ = jn, 6 = 9, indicate that the traj~ ectory to be calculated is of this type. Integrating the other two equations, we get geo A dt eR B(r + 2)/r ti) Niso, the metric yields the equation fa vial Eliminating dp/ds and dt/ds between these equations, we find that CHG Beet octal wk 0 A Dividing the equations (1), we have $ = grrpgzy Substituting Fr initial values, we then, find A/B = v(3/2}, Also, putting initial values in (1i), we get At - 38% = 1, Hence, AP = -1, BY = -2/3. (44) now Teads to the equation or a = krayarry (iii) Putting u = I/(r+1), this equation reduces to; (a = Yur) aut) o? Aferest aor - afisin¥(2u-3) + const. Initially u = 3, = 0, so that the constant of integration is -¥(3/8)(n/2). He now deduce that or Gu- 3 = sin(ag+}x) = cos(as) or re Sr cosfa Fanta Since cos(ag) oscillates between the values #1, r oscillates between the values 1 and 3 (this also follows directly from (i11)). AS ag increases by 2n, r makes one complete oscillation. Thus, two Success Wwe contacts of the particle with the circle rs are made ‘as ¢ increases by 2n/a. No.13, Equations for geodesics through de Sitter space-time are: ates) - r*sinecose fi é Airtsintef’) = Afnctt] = 0- 16 - identifying trajectories in the plane 6 © dr. the remaining equations then integrate to sive ee ee ee) and a, 8, are constants, The metric contributes the first integral Eliminating dt and ds between these three equations, we arrive at the equation an? ” t wea a = (hola tee Also, by division of the equations (i), we get d = oA/(Br*). Substituting the initial values in the last two equations (NB. dr/dp = r/9 = 0, initially}, we calculate that a? = -3R?, a? = 73/2c*, It now follows that fart + Rar? = RE, determining the polar equation of the trajectory in the plane @ = in. Putting r =u}, the tast equation reduces to {oa = fe) + Ru)(4 - Reu) from which it follows that a aff? . faay DYE y oe RRP - [atu =F i : asin fpeenty - a + const. Initially, ¢ = 0 when u = 4/R®; hence, the constant of integration is -j7. We now deduce that 1 g(2Ru ~ 3) = sin(29 + 4r) = costo, which is equivalent to the result stated. No.14. We have geodesic equations dfidx} _ ¢ [ody Ha) - a0) dx} fdy}* te) + fa} +f Integrating, we find ate f y - Et KA, = gw. ¢ as, z ds ° 2 ds z Al dy Since ¥ = 0 initially, B must vanish and so y = 0 throughout the motion, which lies in the xz-plane. Also, dividing the equations for x and t, we conclude that x = A/C = v throughout the motion.<1 - Putting in the initial values z= 1, 2 = 0, we find 1/¢? = v* - = (l= v)(l- zh. Since ¥ < 1, we must have z <1 for Z to be real; i.e. z must decrease from its initial value. Thus + Hence 2 v1 ~ jdt which integrates to give 21-2) = v(l~ jt or2=1- atl - vA)t the integration constant being zero, It now appears that z = 0 when t = 2M ~ Ye Since k =v, x= vt. Thus 22 1- gat - vx hich is the equation of a parabola in the x2-plane. No.15, This result follows immediately from equations (47.14) and (47.16). No. 16. From the components of the gheray-momen tum tensor stated, we calcylate that T = 3p - cfu, Ty, = pes Tez 7 prs Ta = présin?e, Ty, = che". ‘The non~zero components of the Ricci tensor are listed at (51.18). It now follows that Einstein's equation (47.15) requires that ye" sae! = joe! dal = peep = chu) a era! = frat #1) - 1 = der{p = cfu) (Gi) cea — ae? 4 data tet) = -tncteA(ap + tu) (447) Equation (i) is equivalent to the third of the stated equations. From equation (iii), we deduce w4o" ~ ge! + dots - bere -yee%(3p + ctu) (iy) Adding (i) and (iv), we get Mat ea!) = eo%p + of) Equation (11) is equivalent to Ler ty = Beet = 1) + cep = chy) Hence, adding and subtracting the last two equations, we obtain the remaining two equations asked for. Integrating the first of the steted equations, we find v(t =e) = decturt + const. Thus, ee~ 18 - et = Te grt where g = «c*/3. This condi avoid a singularity at r= 0, ion on A is clearly necessary if we are to , Fliwinating p between the second and third of the stated equations, we sel erage? = goipt at dag Leet ay = -ectuet or ewSeer + yer = gate = Sat deny so) = cts ~ gr?, ate = 2qr, the Tast equation reduces to (1 arya" + aot) = Bt This is a BernoulTi type equation for ', We solve it by first substituting 6! = 1/y—tiis leads to wy ar * aregety * bs Substituting e* which 1s a first order linear equation for y. The integrating factor is found to be v/¥(1-qr*)3 thus apy J. feat] Bary This integrates to yield Beay * = Ay-ar*) + const. so that Bos y= tsbvQ-art) - cart] where B is constant. We can now substitute into the second of the listed equations in the exercise to obtain an expression for p, thus: ep = bet ha-e%) = Sq/(I-grt) = 8 ec If p=Qatr=a, then B = 3v(1-qa*). Thus, finally 3givi1-ar?} ~ /1-a97) > Slee Say? Since q = Ke7u/3, this is equivalent to the result stated. Clearly, as r decreases from its surface value a, the pressure ‘increases towards a maximum-value-st the-centre where = ety Le (eo? Pax =o we Tf aa? ig anal] (i @ doncity nak tan areat nn radiue tna Taraal