5C AN 2006 SRI PRATAP COLLEGE LIBRARY SRINAGAR (Kashmir) — DATE LOANEO Class No. Book No. Acc. No.__ This book may be kept for 14 days. An over - due charge will be levied at the rate of 10 Paise tor each vay the book is kept over - time.1 ! TO THE READER. INDLY use this book ver caretully. Tt the book is distiguied or marked or written on while im vour sion the book will have to be ed by anew copy or paid for In case the book be a volume of set of which single volumes are not available the price of the whole set will be realized. — 1 a LIBRARY Clase Noto BNO Bosscocee Pook Not est. BD... MB... MOC NOE rvessseeed PBA cocccceeeSRI PRATAP COLLEGE LIBRARY SRINAGAR ( Kashmir ) — DATE LOANED Class No. _ Book No, Acc. No. This book may be kept for 14 days. An over - due charge will be levied at the rate of 10 Paise tor each day the book is kept over = time.a? THE JOURNAL OF THE Sndian Mathematical Society. PDITED BY \ M, T. NARANIENGAR, “| Hony. Joint Secretary WITH THE CO-OPERATION OF Prof. R, P. PARANJPYE, M.A., B.Sc, _ Prof A. C. L. WILKINSON, M.A., F.R.A.S. and others Maodvas = PRINTED BY S. MURTHY & CO., AT THE “ KAPALEE" PRESS, 305, Thumbu Ohetty Street, 1916, } — +5C AN 2006 SRI PRATAP COLLEGE LIBRARY SRINAGAR (Kashmir) — DATE LOANEO Class No. Book No. Acc. No.__ This book may be kept for 14 days. An over - due charge will be levied at the rate of 10 Paise tor each vay the book is kept over - time.Vol. VIII) FEBRUARY 1916. (No. 1. PROGRESS REPORT. Tho following gentlemen have been clected members of the Society— 1. Mr. Dattatraya Ganesh Dandekar B.Sc.—Science Master, Herbert High School, Kotah (Rajputana). 2. Mv.V. Viraswamatya B.A, L.T.,—Assistant, Pachaiyappa’s High School, 29, Krishnappa Naick Agraharam, Madras, (at con- cessional rate). 3, Mr. Keshav Narasimha Khandekar, BA —Assistant Head Master, Mission High School, Beawar (Rajpatana), (at concessional rate). 4. Mr. F. Hallberg.—Professor of Mathematics, St. Xavier’s Col- lege, Fort Bombay. 5. MrT. R. Venkatesa Aiyar, B.A—Bditor, The Indian Engincer, 1/33, Kachaleswara Agraharam, George Town, Madras. 2, The Caloattn University Calendar for 1015, Part T,has been received for the Library. Poona, } D. D. Karapia, Stat Jan, 1916. Hony. Joint Secretary.2 Green’s Function. By P. V. Sesav Arrar. Note: In this article A stands for tho usnal symbol v7”, 1. Introduction : Asa means of solving certain problems in Flectrostaties Green introdneed a certain fanction into Analysis which afterwards came to be known after his name. Properties of this function were developed by Green himself from a physical point of view (vide Green’s Mathe- motical Papers, eilited by Ferrers : p. 31 and seq.; or Clarke Maxwell’s Electricity and Magnetism: p. 133 and seq.]. But continental mathe- inatieans have developed these properties from the point of view of Pare Analysis and the present article gives the exposition from the latter point of view following French mathematicians, chiefly Poincaré. 2. Definition : Tet T be a volume bonnded by a closed snrface S, and M’ a point sitnated in ‘the interior of T, and M a variable point in the interior of TT or upon the surface Then the function G=(1/r+H) is called the Function of Green relative to the volume T and to the point M’, where ris the distance of the variable point M from the point M’ and H is a fonction which is harmonic in the volume T (i.c., satisfies the equation of Laplace with the necessary conditions as regards continuity) and which is equal to —1/r upon S. ‘Thus defined, this function G vanishes upon S, and satisfies the equation of Laplace at every point of volume T except at M’ where it becomes infinite. ‘The function of Green given by this definition relates to the space enclosed by a closed surface and and a point within it corresponds to the interior problem of Dirichlet [vide: the article on the Problem of Dirichlet (p. 177) in the issue of October 1915 of this Journal]. As in the case of the problem of Dirichlet, there is also an exterior function of Green, t.c., a function relative to a volume T and to a point M’ outside this volume. In this article we confine ourselves to the interior function. 3. Physical Interpretation : Conceive the surface S to be a perfect conductor put in com- munication with the earth, and a unit of positive electricity placed at the point M’. Then the value at the variable point M of the total potential function, arising from the unit ofjelectricity at M’ and from the electricity it will induce upon the surface, will be tho function of Green as defined above.3 For, in consequence of the communication with the earth, the total potentia) at the surface is zero ; further the potential at M due tothe unit of electricity placed at M’is 1/r and that due to the electricity induced on the surface is a harmonic function H; and hence the total potential is H+1/r=G. This physical interpretation also helps to convince us that such a function exists. 4, Notation: Once the domain T is given, the function G is a function of the co-ordinates (x, y, 2) of the variable point M, but it depends also on the co-odinates (2, ’, 2’) of the point M’, the infinity of the fanction. To put in evidence this fact we may write it as G (2, y, z, 2’, y’, 2"), or G (M, M’) indicating by the latter notation, the value at the point M of tho function of Green relative to the domain T and to the point M’. 5. Properties of the Function :— 1°. The fanction is positive at every point M in the domain T. As r tends to zero (ic. as the point M tends to the point M’) H remains finite. Therefore @—1/r remains finite. Hence (7.G—1) tends to zero, and rG tends to unity. Thus when ris very small, the product 7G is very near to unity. We can therefore affirm that in the neighbourhood of the point M’, the function G is positive. [Enclose this point by a very small sphere , Between E andS, we have AG=0; upon, Gis zero and upon 5, G>0. Therefore between E and §, G is everywhere positive according to the following lemma. Lemma: Every harmonic function in a domain lies between the greatest and the least values upon the surface bounding the domain, whether it be simply-connected or maltiply-connected. For, otherwise, there must be a point within the domain T, at which the function is a maximum or a minimum, which it cannot be [vide: the article on ‘ Dirichlet’s Problem’ quoted above.] 2°, In the volume T, G is everywhere less than L/r. In the volume I’, 4 H=0 and H=—1/r upon 8. hat is, H is negative upon S. Since at any point within T', H must lie between its greatest and least values upon S, H is negative at every point of T, Pa G =H+} <}4 3°, Let Tbe a domain boundcd hy a closed surface S, and T,s larger domain bounded by S, and containing T within it; and let M’ be a point in I, then if G, be the function of Green relative to T, and to M’, and G that relative io T and to the same ‘point M’, then, within qT, G.>G. S. For, if G, be H,+1/r and G be H+1/r, we see that,G,—G=H,—H js harmoniciwithin T and since G,>0 in T,, we have H,+1/r>0 upon § but apon S, H=—1/r; therefore H—H>0 upon S and is positive in the volume '' enclosed by S according to the Lemma. Hence G,—G is positive ; te. G,>G in T. 4° Consider the surface G=C. Toleach value of the constant C there corresponds a particular surface. For very great values of the constant, we have evidently very small surfaces sorrounding the pole M’, since G becomes infinite at M’. For values very small of C, we have on the contrary surfaces very near to the surface S. Finally, we pass from one surface to another by a continuous deformation by making the constant C vary in a continuous manner. Further the fonction G being ovniform in the volume T, two surfaces corresponding to two different values of C cannot cnt each other. The surfaces G=C, for different values of C, enclose therefore the pole and are mutually enyeloping ; that is, they are contained withim one another, so that the surface corresponding to any- ono value of C (say Co) encloses within it all surfaces correspondingito the values of C greater than C,.5 The physical interpretation of all the above mentioned properties is quite plain. Ba wmit= 4m, where the integration is extended to the whole of any surface S' contained within S and containing within it the point M’ and the differentiation & is along the normal interior to the surfaceS’. Since G =H+1/r is harmonic everywhere in T except at M' where it is,infinite, ag exists at every ‘point upon S’. Enclose M’ by a small in sphere © of radius p andcentre M’: Now in the space between E and S;, G is harmoaie,and hence { # dg Jextended to the whole of the sur- faces bounding that space is zero. [vide: Cor. 2 of § 2 in the article on the Problem of Dirichlet]. fe, Jes ast |, Bae=0, the differentiation 2 being along the normal interior to the space in enclosed ; in other words dG Sy wed £ qs the differentiation A being along the normal interior to the geome- in trical surface in both cases. 1 Bat fc ag n=, LOE =f ol qae=sI do=4n, ridn For, the differentiation being along the normal interior to ¥, & ce in the case of the sphere. . J dG a. —do=An. y dn [Note.—This conld be easily identified as s particular case of Gauss’ theorem on the surfaco integral of normal force.) 6°. Let T be a volume bounded bys surface S, and M’, M” two points within it, Let G’ denote the function of Green relative to the6 domain T and to the point M’,and G” denote the function of Green * relative to the same domain T and to the point M’. Then the value of G’ at M” is eqnal to the value of G” at M’. That is to say, if we denote by G’ (M) and G” (M) tho values of the . two functions G’ and G" respectively at any point M, we shall have Gt’) =6"(M). Let a‘ be a very small number and consider the surface S' corres- ponding to G’=a’, This surface S’ is very near to S and by taking a’ sufficiently small, S’ can be made to contain within its interior both the point M’and M’. Now consider the integral _ dQ" _ qudG" s=J ,(@ dn an a oxtended to the surface 8), whore & is taken along the normal in- terior to 8’. dG dG" axist at all This integral is woll dotorminate since G', G", >, points of S’. Around, tho two points M’ and M”’ as centres, describe small spheres ©’, £” of radii p', p” respectively, and consider the space contained between the surface S' and the spheres ¥’, ’. In this space, G’ and G” are harmonic, and hence, proceeding as in 5° above, we havo J, (ot 08 anf (oo a +f 7 (ote 0) ae where a is to be taken along the normals interior to the geometric surfaces,7 Now, let M bo the greatest value of G’ npon 3’; then Se a eaeO. Suppose a>1, It is obvious that a, is an increasing function of n and hence it increases indefinitely or tends to a detinite limit us n > ow, In the latter case, let u=f(a) be the limit to which the infinite chain tends. Then u satisfies the equation. 1 “=u, or a=u” ane « Q) = It is easily seen the uv” has a maximum valuek=e” = 1-444 nearly. Hence equation (1) has no real root if a>k. For such values a, cannot converge, and so diverges. If a=k, equation (1) gives u=e. In this case a, may converge, in which case it will tend to the value e, If a. We have hakae ; ag 0 as ob re $= 1a ¢* : ky=Mae® , where §:=1- '>0,<1; 6: Wala, where §:=1~-0-81S0,1, and generally k. Bn 5 ky=k "=o » Where §,-;=1—e~9" S01. Now consider the’sequence ,, §2, 55.....5, From the relation r+: =1--o sequence is therefore a decreasing one, and so either vanishes * Road bofore the Madras Prosidency College Mathomatical Association, 16-12-1916.12 ultimately or tends to a limit § <1. In the latter case § must salisfy & the equation 5=5- b+ 214 the only real root of which is §=0. Hence Lt n=O ken Cor. If lack we have an4, vince acl. te. a,>4,>a,. Similarly aya,>a,. ‘Thus a, oscillates, the amplitude gradually decreasing as » increases, Every even a is greater than every odd a. The even a’s form a decreas« ing sequenceland! the odd a’s an increasing one. Let the limits of these sequences be Xand p, respectively ; so that X, pe are real roots of the equation = a” =a, or a* log u=log a, aud may be obtained as the points of intersection of the curves: (L)y=log a, and (2) y=a* log a, (a<1), ‘The graphs of these curves are given below and it is obvions from the figure that there is only one real point of intersection. Henve =n, and a, converges to a definite limit in mnch the same manner as the convergents of a continned fraction,13 @... 5. Let us next consider the general chain a, — in which the a’s are all greater than unity. The chain diverges if a,>a>k, for all values of r beyond a particular number p, say. For, the part of the infinite chain corresponding to such values is a... greater thana® — .,,whose divergence has been already proved. Gi) If a, =k, after a certain value of 7, we find that the part of the infinite chain beginning with a, is less than k. Hence the whole chain converges. . A. Narasinca Rao. Conical Envelope of a Conicoid. 1 Let 2/a'+y'/bt+24/c'=1, be a conicoid, and (f, g, h) any point T. Also, let (£, 9, %) be a point P on a tangent from T to the conicoid. Then, we have Acer: ACDT=PT: CD, ... ons ~~ @ where CD is the semi-diameter parallel to TP. Also PT: CD=(&—f): al=(q—y): bm=(S—A): en =e Nie +O—gr+G—yyeyt, Gi D being denoted by (al, bm, cn). From (i) and (ji), we find OorTt: A CDT =((k—fy/at+...]? one « (iii) Now 2.4 CPT=CP.T sin PCT =[(kg—mf)*+...]? 2.4 CDT =product of semi-axes of the section CPT Sabe/(al4-B'm!-4-c'n'3, where,l'z-+m'y-+n'z=ojdenotes the plane CPT, viz: ay s| & oy -=0. fog ¥ Hence (iii) reduces: to (Eg) ten JaLEyare... Jarre =[e(E—f'+.....-J) ao ve (ir) which is the equation tolthe conical envelope. a14 Cor :—Eqnation (iv) may be written in the form [eg — at Seber] in the case of @ sphere. 2. Tn the case of a conic a/a*+y'/e"=1, the equation of the tangents from (f, 9) is similarly written (Eg—nfYaV(E—fP tangy, (ég— nfy'=e(E—-f¥+ (1-9) M. T. Naraniencar, which reduces to for a circle. involution and (1, 1) Correspondence. 1. Transformation of Point-pairs on a Line : For visualising certain relations between points on a line, which we term involution and (1, 1) correspondence, » convenient method may be employed. A line is one-dimensional if we regard it as composed of points, but is two-dimensional iif regarded as made up of point-pairs. ‘Thus every point-pair in a line can be placed in (1, 1) correspondence with a point ina plane. This is clearly seen to bo possible if we specify the point-pair by two co-ordinates (p, q). It should be remarked however that since we are not supposed to make any distinction between the points composing the pair, (p,q) should be symmetric functions of the co-ordinates of the points, For example, if &, 2%, the distances of two points from a fixed point on the line be the roots of the equation @+pr+q=o, then (p, q) can be taken to be the co-ordinates of the point-pair (#, a). (vide: Young: Theory of Sets of Points, Ch. VIII). 2. Some Properties : Making this transformation of point-pairs in a line to points in a plane, we have (1) Any continuous one-dimensional series of point-pairs corres- ponds to a curve. (2) Any involution corresponds to a straight lino. For, since an involution is completely determined by two point-pairs, the correspond- ing curve must be complotely determined by two points, 3) Tho locus in the plane of all repeated pairs such as (PP) in tho line is a conic which shall be termed the fundamental conic. For there are two snch pairs in every involution, Hence the corres- ponding curve must cut every straight line in two points.15 (4) If two point-pairs have a point in common, the line joining the corresponding points touches the fundamental conic. For, the double points of the involution determined by two pairs having a common point P, coincide into the single point P. Hence in the plane the line representing the involution touches the fundamental conic. (5) The point-pair formed by the double points of an involution is represented by the pole of the corresponding line w. r,t. the fanda- mental conic. For, if the line ents the conic in (y, 9) corresponding to the pairs (PP), (QQ), the pair formed by the double points being (PQ) must, by the last theorem, lie on the tangents at both p and q. Cor. Two point-pairs separating each other harmonically are represented by conjugate points w. r. f. the fandamental conic. The reciprocal theorem concerning pole and polar is thus seen to follow from the reciprocity of the harmonic relation. (6) Peint-pairs consisting of points in (1, 1) correspondence are represented by # conic having double contact with the fandamental conic. For, if P is any point, the correspondence will carry P into P, and some other point P, into P, Thus there are two pairs (PP,) (PP,) con- taining P which shows that any tangent to the fundamental conic cuts the corresponding locus in two points. ‘his locus is therefore a conic. Since a (1, 1) correspondence has twojand only two double pairs, it follows that this conic must have double contact with the fundamental conic in the points corresponding to the double points. R, Vytuynariaswamr,16 The Face of the Sky for March and April. The Sun enters the first point of Aries on March 21 at 4-30 a.s. and Taorus on April 20 at 3-30 Pax. The Moon March. April. DOH. OM. DOB. OM. New Moon ... we A D9 Was 2 9 Sl rw First Quarter we 120 120 Bam 10 7 65 Pao Full Moon «19 10 56pm. 18 10 37 as. Taast Quarter 260 9 52 ru. 2 4 R aM The Planets. Mercury attains its greatest elongation (27°6’ West) on March 2. It is in superior conjunction on April 15 and in conjunction with the Moon on March 2 and April 2, with Japiter on April 9 and with Uranus on March 5. Venns continues an evening star. It attains its greatest elongation (45° 33’ East) on April 24. It is in conjnnction with the Moon on March 7 at 6-30 p.m. and April 6 at 5-30 p.m. and with Arietis on March 27. Mars which was in opposition on February 9—one of the most favourable oppositions—becomes stationary on March 21. It is in conjunction with the Moon on March 16 and on April 12 at 7-30 p.m. Jupiter is in conjunction with the Sun on April 1 at 7-30 rat, and with the Moon on March 6 and April 3 and April 30, Saturn is stationary on March 11, It is in quadrature on March 31, It is in conjunction with the Moon on March 13 and on April 9. Uranns is in conjonction with the Moon on March 2, March 29 and April 26. Neptune is stationary on April 10. It is in quadratare on April 20, It is in conjanction with the Moon on March 15 and on April 11. N.B.—The total eclipse of the San on February 8 is invisible in India. Y. Rawtesam.17 SOLUTIONS. Question 541. (S. Ramanvyax) :—Prove that lt 1123 458 Te lt+istrsstresit~ Tipit isip igi. Vy Remarke by K. B, Madhava. We have Prym’s identity (Bromwich, p. 9, 8 and!17) 1, stararpt @ a(a+l) a(aFl)(a$2)* aFGF eefl_2# lyf 1 ir i j =e [3 Tiatl 21 a42 Stjap37- | era x == FP) de, el, . * Putting a=z=}, we find 1,1 1 2 a 2{ltrytigstigmt~ pave. ete? an = =22e jr? edt, patting 2=1%; ° so that 1 _t? r eetT oar t o=vEl y yo ~ @ + + Now consider the continued fraction tole 3 4 5 (A) Iy 14 14 14 I+ I+... which we know to be convergent (Chrystal, p. 525, Ex. 7); this can be transformed in varions ways. The three following methods are given in Oskar Perron “ Die Lehre von den Kettenbruchen.” (Teubner, 1913). (i) The first (pp. 294—298) gives at once the desired result of thie example. og 7 TO, ayant o (+au) where ais real and positive and 8>0; but when #=0, let ¢(a, 0) =1. (3) Integrating (2) by parts ta, £)= 4a, 241) 4+az¢(a+l, B41), Let ¢(a, #)= @) 818 ond a set of results follows successively, from which we get (as in Chrystal, Ch. 34, §§ 20-22) 48) pyar B+)e (a+lx (B+2)x (4) 4(a, B+1) I+ 1+ I+ I+e which is trae except for a=0 or any negative integer. Patting 4 =0 and inverting, we have with the aid of (3), A oae be (ate 22 ga, 7 iptetp ae et D 0 @) =) Mau)" au, ° Now pnt 1+-au=o and a is real. is a usefal resnit and is also given in Legendre : Fonetions 1 Blliptiques : Tome ii, Ch. 17. Now in this first put a=}, v=; s=b, and moltiply the left hand side of (5) by #. 2 2 oo 1 ane on? apn 1 Oe OE EF & I+ 147+ If... This result is also given in Laplace, Celestial Mechantce, vol. IV, p. 2573 and in Jacobi: Gee Werke, Bd. VI, pp. 76—78. We have ro} ssi | If we simply pnt &*=}, wo get 112 8 (7 Tp ip rp rg Vee) t Combining (1) and (7) wo have =v, 00 — — — vie | e Pamyae VE = ‘me, (2) Intthe second method referred to, M. Perron (pp. 880—392) obtains the same result by taking the corresponding ‘modified’19 asymptotic series and applies to this “the general method for expressing asymptotic series in the form of arcontinned fraction. Bromwich (p. 267 and §§ 136—139) has also tho same discussion ; he shows that the sories a—]! 2742! a3! at+,., . we we (8) which formally satisties the differential equa ation oeyas ve es we Q) has the the definite integral solution oe 10 y= ° ipa % oy oe ws (10) and the solution in the form of a continued fraction _ 2 2 @ Ww Ww 3a 3 Y°T+ 1+ CF Ie 1+ TF TF. which’is the same as (6), with a=1 and z=1. (3) The third method of M. Perron (pp. 469—472) is to be deduced from the 'general result of expressing a function in continued fraction if it satisfies a differential equation : viz., ab If y=Qy+Py” .. ws we (12) since the n‘” differential of this is Yay MEP a ny ae vee 8) whore Q,= Sees, and Pry, as) we easily see that F=Qt ge or ane.. vee ane (15) where the P’s and Q’s are determined as in (14). If we adopt this method for the function CO y=, te yB—Tdu ow ws (16) by partial integration we have yet al - ° oa Guia. aay oy, nae we (7)20 and hence differentiating (17) yaady-ty4dy, Be v= Say ey”: or generally say" Lan, a+n toe ¥ ‘Therefore from the general formula (15) y= y y which can be formed into l+l/a a+2a43 = Sa Sat Sep by multiplying by a+1 after the second term. Jf in this we put e= — and maltiply by a, we have fr ete du and atlax2a43 ° B+ E+ b+ Wiich are familiar integrals. Question 593. (S. Narayana Arvanoar, M.A.) :—Shew that 1, «@ © @ 1 apt apat Ts apgt toe is equal to a Saar) ENE ses X aFGF ae tO Or» Ifa wnat (a+ 1) * (18) ds) (20) @1)21 Additional Solution by C. Krishnamachary. This can be easily solved by the application of the method of finite differences. We have the equality of operators S=1+4. Hence 14 cB 4 SP 4. HL pal O)4 FEO) + =C4e4z4.. ) (+2d4eh +.)5 since if B@)=lpeteit~ > E()x E(y)=E@+y)] = A+ eh OM sede Operate upon the fonction 1 Then (42642 e+. Lalg ok @+eG )+- ee ace =stontrasstiaps atapl tee and +2845. wae. (+2 (2) +3-00(2) $e) 1 « 2 : =e (agin aang)! a 1 a since 4¢)= (e- yi=3- s= “ern m(2 *)= =A(4 @)) = gar terw, and so on]. Hence we have the equality of the two series 1, 1 i? 1 a tatreett ‘apate Ie ot w (1) a aaFl) taaeieryt Put a-+1 for a, and we similarly have 1 a ef o4 ae erd WeS =@ “fot }. Ai Grparyt G+ EDGES)22 Hence from (1) and (2) we have ay 1 l,« ,@ <4. a Io 4, a a+l a+ _ a alot) * 1 = 1 “1 zy api tapetatagst al @F@FR) Question 617 (S. Narayana Aryan, M.A.):—If F(a, 2, y, 2) denote the hyper- B 4 G41) BB+) g .., show that cometric series 14+ 2" § mes t+ Tt yy+)). 12 r=0 PO P(y)P(a—b+1)P(a+7)PB+") a A Y 3) rape reg + \Pa—— Pes OY) r=0 =14. 2:9" a(a+1).a(a+1).A(B8+1) =14501 + 3GeyytD. Wet where P stands for I, and a'=atr; B=B+r3 ¥=ytr. Examine the case when @=y- Solution by K, B. Madhava. By comparing the co-efficients of x", we have to show that P(atmPB+m)P(atn) POP) P(a)P(B)P(ayP(b-+») Py +n) POr+1) is equal to the series * Plat n)P(B+™)P(y +7) 2) PlatAPB+)P(y +") PO—r+]) n P(a $n)P(B-+n)P()P(y)P(a—b+-1) te. peg mPC@ PEPE PE—b= FPF IPG FH) | hence it is required to show that P(atn) =. ____P@—b+1) PE)PCFMPOF]) 9 POFNPG—b=r 41) PC FDPO=r+1) Multiplying both sides by P(L)P(+1), we have to show 14D ee tt to (nf) terms a(ata—N(atne (b+n—1(b+n23 and this has been done in the solation to Q. 616 (J. I. M.S., Vol. VII, p. 195); putting 3—a=k and b=!, we get, as reqnired to be shown, 1 i nm , kkY) (n—1)_ Iti’ “12 «up to (7-1) terms _(l=-k)\I—k 41)... k-+n—1) +un—1) . When a =y, the right-hand sidejof the given series reduces to F(a,8,b,2), while in tho left-hand side F(a’, 4’, y’, 2) reduces to (l—z)-2-r; * and therefere we get FlaiB.b2) = =e ott P(b)P(a~b+1)P(B+7) 1-2?" PO)PGHHPUHb— EDGE) which, after some manipulations, is seen to be =0-) FB (86-0, 0-4, =) being one of the 23 other forms in which F can be expressed. Question 627. | (K. V. Avayranaravana Sastet, B.A.) :—Four spheres of radii a,b,0, d intersect at right angles. Show that the volame of the tetrahedron formed by their centres is a dabed (4i-*porpay?, Additional Solutions (1) by K. B. Madhava and (2) by R. Vythynathaswam y (1) Take for the equations of the four spheres e+y+2—a=0, and x*---y?-+2°422,2-2y,y-4+22,2—k,=0 (=I, 2, 3). By the conditions of the problem af+y/+2,—k,=b', c? or d according as r is 1, 2, 3. and k,=a Let OMe, typ teyey =F ( y+ 7) =a, where } is any of the quantitios 1, 2, 3 and p either of the other. The volume of the tetrahedron formed by the centres M4 WA Qe (12) (18) | 3 =jla om sl=3/aq me @s) % Ye 4% (18) (23) Ea24 fete a a 4 | Leoye 1 L ! mh) att @ fade Los 2 opel Lo ly¢etya 1 1 1 14d"a* | _ 1 a a 2\ i aged (4 +E+5)! 1 ‘ L xgabed (otttte%pa+) +, (Burnside and Panton, p. 299, Ex. 20) This is set as an exercise in Aldis's Solid Geometry, p137. (2) The volume of a tetrahedron in torms of the edges is}given by olo4uidi4ii 1 0 (12 (13) (14)! M4v'= | 1 (21? 0 (23) (24) 1 (1)* (32)" 0 (34)* 1 (41)? (42)* (43) 0 This is got at once by multiplying the determinants 1 0 0 00) 0 oOo 0 0 1 Vattytte7 a nn Ll | 1 -2x, —2y, —22, a py2+2" atyl+t) 2 ys % 1 land.| 1 —2e, —2y, —22, aP+ye-+2,? tye tay ay ys ty 1 1 -2 —2y —23 a't+y3+e" wPpyebee a y' a 1 1-22, —2y, 2m uPpye+ee Patting in the above oxprossion (12)'=a"4-84(13)' =a" 4c, ete, the resulting determinant =BE(#%2"). (Sco Qnestion 657.) Hence V = Dabed(A0>45-+.0-2 pd Question 6a9. (S. Rawaxusan) :—Prove that : ao B+ >) Ps 095 (any 135} ned _Vi+ViFe nee oo > oY ain (arma neland deduce the following : © (@ i+ }) -a n=l =VGy5—10), BS) sant 1 a= 5" ly -1 () » oo (m3) =} Solution (1) by K. B. Madhava, (2) by N. Durai Rajan, (3) by M. Bhimasena Rao, (1) The denominator and the second term of the numerator in the first problem are respectively given by the imaginary and the real parts of the integral ¢ 5) 8 f. eo Me—iVI =a) ° op Vm 1d 1 the value of which= DT Va(e—ivi—a 2 Ve=Wizey To separate the real and imaginary parts of this, put e=cosz ; and it is easy to see that the expression is equal to Vil-—« * f. —t Saez ees ao ALN) Lan am 11 vari tae i ttn BP whence result (b) follows. (2) Let n=o o=142 5} eS cos (ainVIma) n=1 S= 2) ein (an?VI-="). 426 Snbatitnte z= cos 5 C+S=142)) eT COsO TINE sind ~~ Wd @) Tet Let We seo that A’=K and A=R’. Now 14 2q+2y'+29"-+..+ =/, & (See, Greenhill’s Eliptic Fas) and Maltiplying and equating real and imaginary parts, we get (1+cos 30)$ =C sin 30 ; C(1—cos 30)=S. sin 30. : - O_ 8 . C. sin 7 cos. 7 cos 9 Cc s oro" sin Z Replacing #«=cos 0, we get the expression given in the question. n=o (0) > oe (at =} n=l27 Let S be equal to the left hand side and, e— so that £1 and =k =~ *. S=Eq" (nn?—2) = (q+ 4q'+99°+16 oY +...)—1(g-+g'+ar+..-) Now 14 29-4294+.=4 /K. in Differentiating with respect to q and remembering that logg = —1K'/K 248 y418 gh... =(4.)? ie 1 os 2q-+4q'+99'4...)=3 (1)? aa, (4+4g'499"-+...)=2 (sz) " ) Hence S=Limit of 2 z 1 = dK 1 74 (tax) 2 aC 1 ely Seca] -. The expression within the square brackets is seen to be zero, 0 (3) We know that f 2 cos Qatdt=y noe ° ao af oP (42 003 2 at 2 608 Ait nonacb2vos2nat) db ° wile sf sin Gut Let sin a) The left hand sido= [oe = =fe te, Sin Gort # > by putting Lfor at Lot n become infinite, the value of the integral is zf{ Ber, Ame }, Bromwhich, Infinite Series, Ex4, p. 447,28 ® x . va} M4 yore {2} 4 ema fi T { T z writing 1 2 for 2%, we get co x te) owe _ 1S yD) ven}. ow T ve T Both sides are equal when 2=1. Differentiating with respect to z, we have, A Si nt | x Se et ene 1 { 1+) 0 ; 1 1 2a? (Ss —n n + e wan] = } Putting ¢=1 w » eM QnA = T oo Doh ™ tn =2. 1 In (1) write 7 (cos €-+¢ sin @) for « and wo sce that wo BED) crm 608 8 coy (utarr sin @) T cos ® w = oy + u -R 8 ye (cos6—# sind) where R (2) denotes the real partlof @ 5 en Oe 5) cos (wZsino-Z « Q) rPut r=1, we obtain wo it) 038... (n?a1 sin) 1 2 C ae ) C094 SD) TH MC08 octet sin) [° { 1 J ew _ 0 ; tsing Dyew10°B,(n2msind). T Writing # for cos @, +/1—2* for sin Q, and simplifying we get Wd, ~ 0 as oa(utnzTF o ye sinemyi ae 1 (4) Let cos 81, 20-2, then ray 7 COS o=}, r sin=F; tan 0=2, ono “VE making these substitutions in &, we have o eo "es(a0) = 5 oo $f Hen al 1 when n is not a maltiple of 5, cos (2) =cos os when » is a multiple of 5, cos (ca) 21, 2 nat s #2) o7 © cos(2"-2) NI 7 an =t+ a one cos “5 +20" (1 con 28):=4 (2-087) + voeos Sf Le > cient sy oun (2-cos 2) 1 con = (1-cos avs. cos =) | t+ yet 2 2 ae ™ _stoolg (ve 1) 1 14(v5=)! a ~ e a4) oa 5t cos 9 yan 1 =(10=2v8)4 / 4 © “al () 1a De eon mnt T= 4 ——, =vs.(45>7) *=y5V 5-2 =V5y5=10- ——— = VERVE ean meV Lt sin (may a=1-Vi-z we got, by putting e= 1, 1D) —m indy — mM, w Hence 2 ° an (7-3) =h a result already obtained direotly by differentiating qQ). —2 & - vise sn (nev=*) % Vi-2 Since =2. mn;31 Question 665. (R.N. Arte, M. A., F. R. A. S.):—Find the valne of [fof 14242} 2aedy where zae(1-2 2 the integration is over the positive qnadrant of Sth =1. Solution by K. B. Madhava and H. V. Venkataramiengar. epee dedy fae Oy deity, where the integration extends to all positive values of x+y subject to eye a? yt 2/gt aye the condition ato Pat &=a*/a’, y=y'/b*; Bd ge W*dkdy where k+ <1 ee) [a*(c8 —b°)(a*—b*) — es —b")\(a—b)). “15 @=P yeaa Question 666. (S. Ramanusan) :—Solve in positive rational numbers 2” =y*. For example: 2=4, y=2; 2=3},y=2!. Solution by J. C. Swaminarayan, M.A., and B, Vythynathaswamy. Pat a=hy. Then y!=hy. yos =k. me solution will be a rational number only if k is of the form 143. 8 ny n y= (145) ) and 2= (+) . When n=1, e=4, y=2; when n=2, andy, y=3 when n=3, 256 | 64 ver 7 Thus infinite solutions in positive rational integers will be obtained corresponding to all the integral values of n. —82 Question 667. (S. Namayana Arvar, M.A.) :—Establish the trath of the following (heorems : 1. If F(y)= ty Fi FFD + Hy +2)— Ze ts)+.. then ¢Q)=F) +75 FU+D+3, PY +2)+% F(y+3) +... 2 TF FY) =4y)— Bayt +2 ED Gy 42) then 4) =F) HAF + D+ DEY +2) +o Solution (1) by J.C. Swaminarayan, M.A, and R. Srinivasan M.A. 5 (2) by K.B. Madhava, (1) Let E bean operator which when applied to a fauction of y changes y into y4-1. Then F(y) 0-7 en? _a'Bt — he) ae Bay) t=e FW) = fee ~fPO) =F) GPOTD GPU FD +E FY+8) 4. Also Fly) = {1—"! ou nee DE eee 2 ey) =(14+E) ">(y). | $Y)=C4E)"FY) = {143 242) me... Pray =P 43941) 42CD¥948) 4 (2) Substituting in the right hand sido oxpressions for F (y) F (y+1) etc., in terms of ¢ (y), (y+1)...in the first part, we have for the co-officient of ¢ (y+r) an 1 ={ — 0+ eo} =0 except in the onse of the first torm, which is unity, whence the identity,83 Substituting similarly, in the second part the co-efficient of ¢ (y+-7) n(n -#1)...(n-r—V)_n n(ntl)...Qr+r—2) rh (=H! lant)... (n+7—3) _ (2)! which is the co-efficient of 2 in the prodact of (I—z)-" and (1—2)" and is zero. The co-efficient of ¢ (y) howover is unity, ; hence the result. Question 677. (D. Krrsuxasurri):—A chain of lenth J and mass M is coiled at the edge of a table. A particle of mass m is fastened to one end of the chain and the other end is gently let slip over the edge of the table. Show that the velocity of the particle immediately after it leaves the table is ky [2 gl (1-43)/3], where k=M/(m4M). Solution by K. B. Madhava and Martyn M. Thomas. A particular case of this where M=™m is given in Loney, p. 183 Ex. 11, [See also Ronth, § 150.) From Newton’s Second Law as applied to motion where the moving mass varies, we have z $e 2) = gah, we « Q) To integrate this multiply by Mui, and then we have 17M, Mia , a7), IF ” ~@ ‘ ent ws ws () Calling this velocity u, we have by the principle of conservation of momentum, for the horizontal and vertical velocities” (M+-m)o=Mu and (M-++m)w=Mv wa (A) te. v=kuandw=hu ~ (5) Tho sq. of the vel. when the particle has just left the table i is pw (1+ )ut= §gl(1-422) from (2). s¢34 Question 683. (Gaxrarram R. Jayt) —If'S, is put for 174274 3",..0", prove that Soo Show Sa | 8, S |_f Gy"? P23...n™ Smt Sa Sina Soluticn by K. B. Madhava, A, Narasinga Rao, K. It. Rama Atyar and Martyn M. Thomas. This is just a particular case of the example due to Sylvester solved on p. 229 of Boole’s Finite Differences, (1880) and tho given determinant is, as shown there, seen to be equivalent to 1 1 1 oe 1? 1 2 8 a | ror 8 n ly gt ge " and this is eqnal to the product of the squares of the differences 1, 2,...m taken with the proper sign. ve =f @)" yy? Hence D { sie oy It is also shown in the reference (loc. oft.) that Se Sen we Segue Sen Syaa we Seance [= (0 *D 1 Sygna Sean we Sean Question 686. (S. Narayana Atrar, M,AO:—Establish ne following identities :-— () MOADOH2) ote ay 02s BEF) (b+ 2 -(6+n—1) (a—b)(a—b=1) (@-d) font 1)...(a—b—n $1. 10 ae fo Barb Tn obs ); Fle tot SOFT) @+n=1) « Ganent ye (a—b—n+1) _ CG) Sest) Caney OO" {a-08 (a1) (ayn 2+ 0).:.(4-+n=}) + Ors" Be aGeaal y}85 Gil) F(@,b,¢,2) =a)" F (@ c—b, 24) » where F stands for the hypergeometric series, Gv) (v) and (vi). Remarks by K. B. Madhava and R. Srinivasan, M.A. (i) is proved in the solation of Q. (616) ; (ii) change a intolb—a in (i) and the result follows ; (ii) is proved in the solution to Q.i(616) ; Gv) if in Q. 616) (Q)=log (142), the result follows; (v) if in Q. (616) ¢ (z)=sin # and «=7,, the result follows’; (vi): if in Q. (616) $ (x)= los 2 and «—71, the result follows ; In connection with (iii) the following algebraic proof extracted from the Messenger of Mathematics, Vol. XLIV, No. 8, December 1914, may be of interest. The proof is by Prof. M. J. M. Hill, University College, London : Tho righthand side is (1—2)~? F ( 4, a) = G2) 94 26=D(— ayaa —a—2 +a , base 4p an(e—), ok 2m ale, Hence the coeff. of 2" is 0b _) (14 Nwns tO) GI! = altgg One (CBC Dict ney(o—B), (6+ 2) ura ten(=)"C—0), } The (r-+-1) term in brackets is (—)'n¢ (e—B) e+ 1)ner —])"e-(o—b)(c—b—1)...(c—b+7-—1) x (C74 r-1)...(¢-n—1) =ne{b—c)(b—c—1).,.(b—c—r +1) x (c +n—1)(c-+n—2)...(c-+7).86 If wejnow adopt, asin the'statement of Vandermonde’s theorem, an abbrevintion for a (a—1)...(a—7+1),Isay @,, so that V’s theorem canqbo written 0 FB = Drei Baer r=0 we have for the (r-4-1)™ term in brackets the expression ne b—C) (CFD ner and therefore by Vandermonde’s theorem, the whole expression in brackets =(b—cte+n—Yy =(b+n—Vy =(b4n—-1)(b-+-n—2)..(b+ 1b. =b,, in tho notation of this paper, where thronghont we have alopicd for abbreviation ¢, fore (erie). w(o-+-n—1), and therefore the coeft. of a" is ba ¢, Which is the coeft. of "in F(a,b,c,2). ‘Thus the identity is demonstrated. Question 696. (S. Krismasawat Atyanoar) :—If X, pp be the lntera-recta of the parabola and rectangnlar hyperbola of closest contact with a curve at any point, prove that 2\p=p. Solution by R. Srinivasan, 3f.A., Martyn M. Thomas and K. B. Madhava. With the usual notation, in a parabola sP_sy SY SA =sec $. o SPeSA sect. oe p=2asec'¢. (SA=a). Also in a rectangular hyperbola, cP? CP . P=Gy" Gya ee? ¢, CD. CY =a}, where @ is the semi-axis (or semi latus-rectum.) o p=a sec By. o X=2p cos*¢ andip=2p costs. te 2hp=p37 Question 703. (N. Savana jAtran, M.A.):—If AP is the symmedian through A, prove that L{w+e)(AK. KP) } =30%'/(a?-+?-+0'). Solution by B. Srinivasan, M.A., K. B. Madhava, R. D. Karve and L, N, Subramanian. Let the distance of K from BC be z, and the attitude AD be p. Now BP: PCr: o?: BY ‘ B ABY4c°.AC*=b° BP! 4 cICP*4 (bth c!)AP?, . _Die(258+209—uy, . fe. AP os ws (i) . KP _2_2aA/(a-+5'+0') Again RP p> 2Bja = a ~epb per AK +c? and epoenceet , (bb c8 AK kPa oe Beara? . sah _ abe? * (+e) ARKP= 000 3 abc? 24¢) AKKP} = £{ (+e) AKKP)} = OTe. (26*4-2c?—a*). 2 . Question 704. (S. Matnarr Rao, B.A.) :—If the sam of a number of 3 digits and the number formed by reversing the digits be divisible by 37, the sam of all such pairs of numbers is 480x 37. Solution by 8. V. Venkatarayasastri, M.A., LT, and R. D. Karve. If a,b,c be the 3 digits, the sum of the number and the number formed by reversing the digits =101 (a-+c)+20b. ble: by 87, the values of b and a+e are the 9 pairs , 18. it set of values 1, 2 for b, a+c, we get 210 and 111. As these do not give usa pair of 2 numbers satisfying the given con- ditions, we reject them, Taking the set 2, 4, we get 2 nambers, omitting numbers 420 and 222 for the above reason. Their sum =222 x 2. Taking the set 3, 6, we get 4 numbersor 2 pairs. Their sam=333x 4, And go on. The sum of all such pnirs=222 x 2+333 x 4-+- 444% 4-+...-+888% 2 . =480%x 37,