“/. THE JOURNAL OF THE ® ° CY e [ndian Nfathematical Society Vol. XIV.| OCTOBER 1922, _ (No. 5 M. T. NARANIENGAR, M.A. Hony, Joint Secretary WITH THE CO-OPERATION OF The Hon'ble R. P, PARANJPYE, MA., D.Sc. Prof. A. C. L. WILKINSON, M.A,, F.R.A.S, and others. (@Madtas: PRINTED BY SRINIVASA VARADACHARI & CO, 4, MOUNT ROAD. 1922 Annual Subscription: Rs. 6.) (Single Copy ; One RupeeCONTENTS, “PAGE Some Diophantine Approximations connected with Quadratic Surds : K, Ananda, Rau wae a ae owe 16L On Some Relations satiefied by Bessel Function of Degree (n+ 4): 8, K, Banerji = 5 a sw 167 On Certain Quadratic Systems of Conics: R. Vydyanathaewamy —.. 17 The Pedal Line Family of a Triangle: A. Narasinga Rao we 174 Solutions to Questions oe ‘ae ae Questions for Solation i sa a ' A paper should contain a short and clear summary of the new re- sults obtained and the relations in which they stand to results already known. It should be remembered that, at the preeent . stage of mathe- matical research, hardly apy paper is likely to beso completely original as to be independent of earlier work in the same direction; and that readers are often helped to appreciate the importance of a new investiga-. tion by seeing its connection with more femiliar results, ‘The principal results of a paper ebould, when possible, be enunciated ceparately and explicitly in the form of definite theorems, The Journal is open to contributions from members as well as anbscribers, The Editors may also accept contributions from others, Contributors will be supplied, if so desired, with extra copies of their contributions at net cost. ‘All contributions should be written legibly on one side only of the paper, and all diagrams should be neatly and accurately drawn on separate slips ‘All communications intended for the Journal should be addressed to the Hony. Joint Secretary and Editor M, T. NARANIENGAR, M.A., 23, Mallesvaram, Bangalore, or Joint Editor K. ANANDA RAO, M.A., Professor, Presidency College, Madras. All business communications regarding the Journal should be addressed to the Hony. Asst. Secy. N, RAGHUNATHA AIYANGAR, MA., Presidency College, Triplicane, Madras. Enquiries by intending members of the Society and all other communi- cations may be addressed to the Hony. Joint Secretary P. V. SESHU ‘AIYAR, B.A., L.T., 18, Pycroft’s Road, Triplicane, Madras.THE JOURNAL OF THE [ndian Mathematical OCTOBER 1922. Vol. XIV.) SOME DIOPHANTINE APPROXIMATIONS CONNECTED WITH QUADRATIC SURDS By K. Ananpa Rav. 1, Let @ be a positive irrational number, and let a,, a .., be an increasing sequence of positive integers tending to infinity. The distribation of the set of numbers formed by the fractional parts of a6 is a very interesting problem. The classical results in this subject were established by Dirichlet and Kronecker (+). In recent yeara the subject of Diopbantine A pproximatione, of which the above is one of the problems, has attracted much attention not only for its own sake, but also for the applications it has found in the Analytic Theory of Numbers, Hardy, Littlewood and Bohr have contributed largely to these later develop- ments (4), 2, Let us denote, as is usual in this subject, the fractional part of x by (z). A very elementary result which can be established, either by what is known as Dirichlet's principle, or by the theory of continued fractions, is that, if we take a, = x, the set of nurmbars (78) is dense every where in the unit segment (0, 1). If, however, we consider a sequence a, which tends to infinity with great rapidity the set of numbers {a,6) may not be dense everywhere, For example, if @ = e anda, = n! then (vte) > 0 () Kronecker: Werke. Vol IIL, p. 49. Dirichlet, Werke, Vol, T, p, 685, Sre al Minkowski, Diophantische Approximationen, , Pee () G. H. Hardy and J.E, Littlewood : Some Problems of Diophantine Approx!- mations, Acta Mathematica, Vol, XXXVIT, 1914, H, Bobr : Zur Theorie der Riemannscken 7% Acta Mathematica, Vol, XL, 1916, H, Bobr and R. Courant Approximationen anf die Ri 21 etefunktion im Kritlecben Streifen, i Neue Avwendungen der Theorie der Diopbantlache lemenosche Zetafonktion, Cretles Journal, 1914,162 86 n=» cw , a8 is easily vorified by using the infinite series = Lou eal ity tons 80 that the only limiting point of the set (7! 6) ie zero ("), 8, The object of this note is to obtain some results regarding the distribution of the set of numbers (g," 6), where @ is a quadratic aurd and * is tho nth convergent in the coatiaued fraction for @. It is proved here that the limiting points of the set are finite in numbor and oan be easily calculated. 4. We shall begin by considering the set of numbers a) qn" je-#|, whose distribution is closely related to that of (7,° 6). In considering the distribution of the set (1) there is no loss of generality, if we confine ourselves to quadratic surds whose continued frac- tions ate purely recurring. To prove thie, we observe first that if, @ and $ aro two equivalent irrationals, that is to say, if they are connected by a relation of the form A+ Bo C+ De’ where A, B, C, D ate integers such that AD — BC = - 1, then tho par. tial quotients in the continued fractions for @ and ¢ are ultimately identical, o= Let 1 1 1 1 and beat eae mt? and let 2 and z denote the s* convergents in these continued frac- tions. We shall prove that, ae 7 —> 0, P, a [gtr | wge |g —Pat @ a.fem fer | ~e..|# at: from which it follows that the two sete of numbers 7 | o-- (2) For other examples, see the paper ot Hardy ond Littlewood referred to above,163 are distributed similarly ; and so in the case of quadratic eurds there is n0 Joss of generality in supposing the continued fraction to be purely recurring. 5. The proof of (2) ia eaey. We shall write for simplicity p, 7, P,Q for pair Qatr Pmir, Quin; 80 that, instead of r <>» a, we may cay % Qe» ~ We have where AD — BC = + 1. It is easily seen that the last fraction on the tight hand side is in its lowest terms ; and so p= AP + BG g = CP + DQ. Now 2 [Ab+B_ ap+aq Ce + D CP+DQ = 1¢Q—P | (CP + DQ) ope D by making use of | AD — BC | = 1and(3), Therefore cf +D cep 7) since 6. We shall now proceed to the proof of the main result of thie paper, Let 1 1 a= 4 see eeeeeeane a at a, ® be a quadratic surd whose continued fraction is purely recurriag. We shall adopt the notation a, + i Sek ag SM a * wee Qn].164 Let x be the sth convergent of 6. Then LL tpt pet Oat ee oS me at ant a 01 gn gear | 6, — Bl = \é O,pn + Par Pn Oygn + dnmt Qn el, * Orem + gn Gu=1) Gn? Oy + Qnaa Qs Similarly | 6 — and geuerally | oe, — 2 Gan Now Loy Qgt at at and so on. It is easy to eee that as s =» oo through integral values 8" oe [an dunt, oe a) = — Pye Quek t Ay ve @) oy Hence g* 6; — | => quan a8 s oro, If wo start with the identity 1 L et Em ae ae 8, where 02 = [4g, as) oe Om 21), and argue as above, we obtain the resalt that as s —» 2 l 2 Pet 1 Tat) | Oa onet @y—O where [a,,am O-1 4. @;) = — i165 Generally, if v ie an integer such that 1 0; 3, being defined by [arna, O77? see ys Quy Qn-2 wwe Ge] = — We can now obtain the limiting points of the set (97, 6) by use of results proved above. Let 1 6 = bt 6, = a, + . LetlO,, 0, ... O» be defined as above, Lot m be the rth convergent of @. It is easily seen that, if 7 is odd, (890) = 9% jo-*|s while, if r is even, (a?) = 1g? | a | gl Applying these to the results established about the set m fone we get the following, I hare merely enumerated the results aa the verification is easy. , To the following, & takes the values 1, 2 wee NE166 I. neven. (i) (m + &) oda. As s=> © through 1, 2, 3, wu. 0 neers (Og ntttn) > oz. (ii) (m + &) even. As s <> w through 1, 2, 3, w... (9 gtatitn) 1 Ze. I. nodd. (i) (mm + he) oda. As s > w throvgh 2, 4, 6, 1 (Outten) Pm 1 As 6 -> @ through 1, 3, 5,. 1 °, — =. (0 grein) > 1— gz (ii) (m + &) even. As s —> w through 2, 4, 6... (8 Gntiten ) me 1 BF As s —>» © through 1, 3, 5, .. 1 CO Paria) > GF167 ON SOME RELATIONS SATISFIED BY BESSEL FUNCTIONS OF DEGREE (” + 5) By S. K. Banerst, D Sc., Director, Bombiy and Alibag Observatories. It is easily proved that where r = (2 + 4? + 2%) and p= 2/r. Consequently * We have therefore rad) EP En, (de (Be ain m a [2 we + Am — Ae A im— A (m—s)A(s) , 7 Qn 2m —U+ti * A@+mu—s a0 Prince (2) Re ide z’ m and 1 being supposed to bo positive integers and m < n, where + A (2) al BF. Qn — bd, nt This can alao be written in the form », (f,) th @ ide ve m = 3 m2 a + Pm — de + 1 Alms) Alm—s) AG) Bn + 2m it st 1 A (n+m—sy Inim=244 (2) ~ a v . * See Rayleigh, Theory of Sound, et seg. a peril, Dab Ga Sve , Vol. IV, p. 8. Y OF Sound, Art, OBA, ef seq. and Banerji, Bul, Cat, Math, t See Adams, Proc, Roy, Soc., 1878,168 Similarly, we get » (gs) (eH 2} m =>"% 2n + 2m — 4s +1 Qn + Im — 2s +1 He) Pagm—alH). For the Beesel functions of the second kind, we have Pp, (¢) im = 2 Kun © p(y), a) o*_ yn pf Kato 2), and P, (; 3) é ye We should therefore obtain in a similar manner, Po (ize) a? m => jaats 2+ 2m—4541 A(n—s) A(m—s) Als) Knima2re $ (2) Qn+2m—B+1 A(n + m + 8) Je s=0 Keser) and Pp (&) {tae mt _ Qn + 2m — 45+ 1 Alm — 4) Afm—s) AG) y 0 Qu + 2m — ets $1 Ain + m + 8) = Krswaeisy (7) Ar nm and m being euppoted to be integers and m ~[S 5 mn + Im— Bs + 1 m=O s=0 Alm = 9) Alm — 8) AG) Sninonsed (2) 7, A(n + m + 8) . Ju the uppor limit for the summation of the series within the bracket being m or # according as m #. or > 1. Similarly, (8) [eX P, wo] norm « cnats 20 + Om — 48 $1 = n= s ———_—_— =Sa[3 en on am — Be FL m=0 s=0 Alt — s\ A(m — 8) A(s) Tamar O) _, . ARSED AO Tate tt eee, tH) ] Tooee theorems moan that if « —#Y,,4 4 (aor re —* 44 (r) Pa (10), which is a solution of the wave equation, is differentiated ‘in any manner the results can be expressed in a ssrios of functions of the eame type, The samo theorems are true if Bessal functions of the first kind be replaced by Bessel functions of the second kind. Maxwell first pointed out that if a solution of the wave equation be differentiated avy number of times with respect to 2, y, solutions of the wave equation of various orders of complexities. These complex solutions cannot be interpreted physically unless they are expressed in a cores of standard solutions, The method indicated fabove may be used with advantage to express the o . of standard solutions, 2, we obtain omplex solutions in a series 2217u ON CERTAIN QUADRATIC SYSTEMS OF CONICS. By R, VypYANATHASWAMY. I. Definition. Tt (AS + 5S’) is a pair of straight lines when ArvtOrxat ort VHD, then when 6 = 0, it is known that there is ar infinite number of inscribed triangles of S which are self-polar w.r.t. S’. In this case S will be said to be ew-harmonic to S’, and S' in-harmonic to S. It will be noticed that the condition that S may be ex-harmonic to S! is linear in the co-efficients of S and quadratic in those of S’, IE S is both ex- and in-harmonic to S’, S will be said to be harmonic to S. The insoribed triangles of S belonging to a given pencil are all self- polar triangles of one and only one conic S’*, Thus there may be established a correspondence between pencils of cubics and the in- harmonic conics of $. In this correspondence the harmonic conics of S will correspond to the Null pencils +. II. 4 G1, 1) correspondence may be established between the points of a five-dimensional space S, and all conics in a plane; for instance between the point in S, whose homogeneous co-ordinates aro (a bof g h) and the plane conic whose equation is (abcfgh) (2,4, 2)? =0. By this correspondence the straight lines of S, will correspond to four-point systems of conics, and the planes of S, will correspond to nets of conics, ive, systems of the typ3 2,3, + ro S, + r,S, where S,, S;. S, are three conics. The conics which are line-pairs will correspond to points in Ss lying on the cubic eurface A = abs + 2fgh — af? — bg2 — chs = 0. A quadric surface in Sy will correspond to a general quadratic eystem of conies, namely the set of conics whose co-rfiicients satisfy a general quadratic relation, Now a quadric surface in Sg bas always two kinds of generating planer, such tbat any two generating planes of the same kind intersect in one ard only one point, and two generating planes of different kinds either do not intersect at all or intersect in a straight line t. Hence * Vide * Linear Systems of the Third Order on the Conic, 1921. + Ibid. + * Quadric in Five Dimensions,’ J, J. S., Jane 1920). J. 1, M, 8, August171 A quadratic system of conics necessarily contains w* nets. These nets can be divided into two distinct classes, such that two nets of the same class have one and only one conic in common, and two nets belong- ing to different classes have either no conic or a four-point system of conics in common. This theorem is fundamental in the study of qnadratic systems of conics. We propose in this note to study the application of this theorem to a particular type of quadratic systems, namely, those systems which correspond in S, to the polar quadrics of the cubic surface A. These polar quadrics have equatious of the form a’ (bo — f2) + b' (ca — g3) + o' (ab — hi?) + 2F" (gh — af) + 2¢ (af — hy) + 2h’ (fg — ch) = 0 where (a, 6...) are current co-ordinates and (a’, b’..,) constants. Now this equation simply expresses that the conic (a be fg hk) is in-harmonic to a fixed conic (a’ b' c’ fy’ h’). Taus the polar quadrics of A correspond to special quadratic systems, namely, systems consisting of all conics in- harmonic to a fixed conic. III, Consider now the family # of conics in-barmonic to a given conic S, Let s be the point in S, corresponding to Sand let S be the polar quadric of s w.r,t. A; then Sie the locus in S, corresponding to the family /, Further if the polar plane that is four-dimensional flat region) of Sw. r,t. 5 intereect the latter in the locus E, then clearly 3, is the locus of points which correspond to harmonic conics of S. Now « determines an involutoric correspondence Js between the Points of S, namely, the correspondence of the two intersections of S with any line through s. Tae eelf-corresponding points of Is are clearly the Points on E,. Further it is not difficult to shew that the two kinds of generating regions of ¥ aro interchanged by J * The meaning of the transformation J, in the plane ie evident. ‘Thue if S' isan in-harmonic conic of S, /,8' will be the olfer in-harmonic conic of $ contained in the four-point system (S, S$’). Also, from what has preceded, this correspondence interchanges the two classes of in- harmonic nets of S—that is the nets which belong to the family of conics in-harmonic to S. * This is easily seen by the method of correspondence adopted in *Quadrie in Vive dimensions,’ In this mettod s will correspond to a system of forces and points on ¥ to single forces, The correspondence I, will ithen be the correspondence of Conjugate forces w.7.c, the system s. ‘This latter correspondence will obviously replace each poiat by a plane, viz., its nul-plane and viee versa, Since points and Planes correspond to the two kinds of generating regions of, it follows that Is interchanges the two kinds of generating regions,172 Hence we have The conic S determines an involutoric correspondence between its in-harmonic conics—the self-corresponding conics of which are its harmo- nic conics. This correspondence interchanges the two classes of the in-harmonic nets of S. Tho operation I will be called inversion, and corresponding in- harmonic conics of S will be called inverts of each other in S. Ex. (1). The invert of an in-harmonic conic having double contact with § is a repeated line, viz., thechord of contact, Vice versa, any ropeated lino is an in-harmonie conic of S and ite iavert is a conic baring double contact with S. Ex, (2). The pencils of inscribed As determined on S by any ine harmonic conic and its invert aro mutually harmonic pencils (i.e. any triad of either pencil is harmonic to avy triad of the other). 1V. The in-harmonic nets of S. If ABC is any inacribed A of S, then, clearly, the net of conics having ABC for a self-polar triaugle is an in-barmonic net of S. Let us torm thie, the in-harmonic net of the first class corresponding to the in- scribed triangle ABC. This net contains three repeated lines, viz. the sides of ABC and, further, is easily seen to contain only one conic having double contact with S—the chord of contact being the polar line of the triangle ABC w.r.t. 8. The invert of this net we call the in-harmonic net of the second class corresponding to the triangle ABC. In view of Ex, (lt) III, this not should contain one repeated live, namely, the polar line of ABC and three conics having donble contact with S at the coroers of ABO. A four-point system every conic of which is harmonic to S, may itself be said to be harmonic to 8. Clearly such a four-point system is inverted into itself by S. It is easy to see that any net of the first class determined by an inscribed triangle ABO can contain only one four-point system harmonic to S; for there is only one quadrangle which has ABC for harmonic triangle and which ie in addition self-polar w.r.t. 8, By inversion, it follows tbat thie eame harmonic system is contained in the net of the second class determined by ABC. Hence the net of the second class may be conceived as the eet of conics having double contact with the conics of a four-point eystem Z, harmonic to S, the chord of contact being always a particular line, namely, the polar line of the harmonic triangle of the quadrangle which determines L. ; 1. Finally we may verify the theorem, concerning the common conics of two nete, Clearly any two nets of the first clase, say those determined123 by the triangles ABC, A’B'C’, have one and only one conic in common, namely, the unique conic which has ABC and A’B'0' for self-polar triangles. By inversion it follows that any two uets of the second class bave one and only conic in common. Consider now two nets of different classes, say a net JV, of the first clase determined by ABC and a net Np of the second class determined by A’B’C’, The common conics of these two nets must clearly consist of those conics of ¥, whose inverts belong to the net NV of the firet class determined by A’B'C’. Hence, in view of Ex. (2) III above, NY, and N, have no common conics unless ABC and A’BC’ are harmonic. When these triangles are harmonic, we may easily shew that the common conies are all the conics of the four-point system of N,, for which (h,,hy) is a copjagate pair ;h,, hy being the points of intersection with S of the polar line of A’B’O’. For, fhe pencils determined on S by the conics of this four-point system are all such that their harmonic pencils (Ex. (2) quoted above) cuntain ABC. Hence the inverts of the conics of this system belong to YN, Thus nets of different classes have either no conic in common or a four-point system in common. Bx, (1), 1f ABC, A’B'O' are harmonic inscribed triangles of S and if their polar lines cut S in (hy, hy), (2’, ha’) respectively, show tvat the in- Vert of a conic which has ABO for soli-polar triangle and h,’, hy’ fot conjugate pointe, is a conic which has A’B'C’ for self-polar tiangle and Iy, hy for conjugate pointe. Ex, (2), Shew that any in-harmonic four-point system of Smust have its common self-polar triangle inscribed in S. Show that any in-barmonie conic of § belongs to « “in-harmonic four-point systems,174 THE PEDAL LINE FAMILY OF A TRIANGLE By A, Nanastnca Rao, 1, Introductory.—In this paper an attempt is made to study the distribution and properties of the pedal lines of a triangle, regarded not individually but as members of a linear aggregate. Though the results obtained are mostly well koowo, it is hoped that the symmetry and simplicity of the methods employed will be of interest. The problem of determining all triangles having a common pedal line system is also teken up for solution at the end of the paper, 2, The Pedal Line—Let A, B, C be atriangle and La Point on its circumcircle. It is well-known that the feet of the perpendiculars from Lon the sides of ABC lie ona line called the Simpson or the pedal line of L w, r. ¢. the triangle. As L moves round the circle, we obtain a singly infinite system of pedal lines which could be associated with the continuous variations of a single parameter 7. We sball denote the vectorial angle of avy pt, Loa the circle by the corresponding Greek letter a and tan * by 2 so that as J varies from — a to + oo the pt. L moves once round the citcle, Thus the parameters of the verticus A,B, C are a, b,c and those of the circular points Land J are + i and —i. 3. Some particular cases,—From the definition of the pedal line of a point it follows that :— (1) The pedal lines of the vertices of the triangle are the per- pendiculars from them to the opposite sides. (2) The sides of the triangle ABC are themselves the pedal line of the pointe diametrically opposit to them on the circumcircle. (8) The pedal line of the points I J is the line at infinity, For if the pedal of [ cuts AB in P then 1 must lie on the line through P 1 AB and this will not be true unless P be at infinity, Similar considerations apply to the other two sides BO, CA, In the same manner the pedal of J is also the line at infinity, 4, Class of the Envelope,—Take a pt. P on AB and let. PM be a chord 1 to AB. Then through the point P there pase three pedal lines, the line AB, tho pedal line of L and that of M and it is easy to see that no other pedal line can possibly pase through it, We may infer by the principle of contiouity, that three pedal lines pass through avy given pt. and that their Tinear uggtegate envelope a curve of the be175 third class of which the line at infloity is a bitangent as it corresponds to two distinct values + i of the parameter. 5. Concurrent Pedals.—Let us now determine the relation connecting the parameters 7, m,n of 8 concurrent pedale. The relation is obviously linear and eymmetrical, since given 2 pedal lines, there is a unique third passing through their common point. Let us aseume it to be Almn + B (im + ma + nl) +O (lt m+n) +D = 0. or slightly 2[A mn + B(m + ») + C] + [Bma + C(m + 2) -+ D]=0. The later form shows that such a lineo-linear system contains always a singular pair m,, x, (obtained by eclving simultaneonsly Am» + Ba+n + C = Oand Bmn + Cm+n + D = 0) which form a triad with any 7, In this case the singular pair is obviously (+ i, —i) eince the pedals of these points, and any other point are all concurrent. Hence we have A+C=0, B+D=0 on IL Also a, 4, ¢ form a triad catisfying I since the corresponding lines (the perpendiculare of the triangle) are concarrent. I. We thus obtain the ratios of A, B,C, D and the condition for con- currence is finally found to be Y= Iman T= Sm we TID or rememberiug that / wehave tan tu ty 2 ie, Mt etu S=a+Bt y+ lw... iv as the condition for the concurrency, a result which docs not seem to be generally kacwn, Cor 7. If the pedal lines of L, M,N witb respect to the triangle ABC are concurrent, so also are the pedals of A,B and C w +. t, the 4 LMN. Cor, 2. By a proper choice of origin, the condition of concurrency can be reduced to the simple form tan Leman (For this we choose the origin of angular measurement eo that a +e +Y = 0or kr.) We shall hereafter take this to be the case always. Cor 5. Tho pedal correspondiog to the parameter 7 tonches tha envelope whore it is cut by that of 2% ° R—T176 , Gor. 4. The pedal lines corresponding to the points 29 = 31, ie, 1=0 + ¥@ are the cuspidal tangents of the envelope, These tangents are themselves concurrent, 6 The Equation of the Pedal Line in areat co-ordinates, The equation of tho pedal line of the pt. @ w.r. t, the triangle ay is of the form slb+yM+2N=0 whore L M and N are cubies in ¢ ( = tan §) . Now when ¢=—}, —} or—! the corresponding lines are 2=0, y =Oorz = 0. Hence the equation can now be written as 2( +1) (tel) Li by (ta +1) (e+ 1) M fe, =0 whore L’ M’ and N’ are linear in ¢, Again whon ¢ = a tho corresponding padal is y _ tanB y(a—b) 4 (ae) _ 2 = tend % L+ab 1+a =° and similarly for ¢ = b and c, Hence L’ = (¢—a) U',M’ = ¢—6) MY, N’ = (to) N” whore 1’, M” and N” aro constants, Finally applying the condition thet tho pedal of ¢ = iis the line x + y + 2 =o, we have L” = M" = N” giving us the required equation a(i—a) , y(t—d) a(t—e) l+ma * t+ + Tpe on - -* or tan "5" + ytan ® oo + stan ® 3 =0 =0. 7. Perpendicular pedal line, —As an example of the methods employ- ed in this papor, let ue determine the relation between the parameters of perpendicular pedals. ‘To each 2 there corresponds a single direction, a definite perp. direction and 3 lines along this direction. Bat as the line at infinity counts twice as a pedal in this perp. direction there is only one other distinct line corresponding, say, to the parameter m. The relation connecting 2 and m is thus a one-to-one correspondence of the type Aln+B(l+m)+C=0. But we know tbat the pedals corresponding to the parameters (a - 1) (4 -} ) (« — |.) are at rt, angles (being in fact the sides and the perps. of the A), Henco’these pairs satisfy the relation which must therefore be in +1=0 or tan 5 tan 5 + = 0;177 showing that the pedals of the extremities of a diameter meet at right angles. Itis a well-known theorem that they meet at right angles on the nine- pointe circle, 8. The Bnvelope of the Family.—We have seen already that the line at infinity (Lio) is a bitangent of the envelope, Also since 3 tangents pass through any given point, and as no other pedal can pase through I other than Liz, , I must be one of the points of contact and similarly J. ‘The envelope is thus seen to be a curve of the 3rd class having double con- taot with Lo at the circalar points, iz, a tricep Hypocycloid. The inter- sections of L tangents lie on the in-circle of the Hypocycloid, This ie the nine-points circls of the triangle, The cusp lies on a concentric circle of thrice its radivs. The equations of the cuspidal tangents have been already obtained. These cuspidal tangent are equally inclined to one another and meet at the nine-points circle of ABO. 9. Are there other triangles having the same pedal line family as the 4 ABOP The question is easily answered in the affirmative. Any trane- formation which transforms 1 lines into L lines and which leaves the hypocycloid unchanged conserves the pedal line group and such a trane- formation may give us a new triangle. Rotation through + 120, about the nine-point centro and reflexion about the 3 cuspidal tangents readily suggest themselves, More generally the bypocycloid is completely determined when we are given its in-circle and ore of its cuspidal tangents both in position and magnitude, Hence if we are given a circle acd one of its din- meters, all triangles constructed so as to have this circle for their nine- points circle and the given diameter for a pedal line will bave a common pedal line group, namely, the tangents of the unique by pocycloid escribed to this circle and baving the diamater for a cuspidal tangent. There is an infinite number of triangles satisfying the condition, 23178 SOLUTIONS. Question 417, CV. Bamaswaati Aryan, M.A.) :—IE parabolas are escribed to a tri- angle, shew that any system of corresponding lines connected with them {considered as similar figures) envelope a three-cusped bypocycloid. Question 1101, (M, Burmasewa Rao):—Show that the centre of a three-cusped bypocycloid escribed to a given triangle is equidistant from the ortho- centre and the circum-centre of the triangle. Question 1250. CV. Ramaswamt Aryan, M.A.):—If five straight lines be tangents to a three-cusped bypocycloid, prove that the foci of the five parabolas touching the lines taken fouc at a time are all collinear, Solution and remarks by M. Bhimasena Rao. Let S be the focus and F the foot of the directrix of a parabola escrib- ed toa triangle ABO whose ortho-centre and circam-centre are H and O respectively, and let PQR be any line meeting the directrix FH in P, If ST be drawn parallel to PQR, meeting the directrix in T, we have to prove that PQI touches a three-cusped bypocycloid when the triangle STP ia given in epeciee, On HO describe a triangle HON inversely similar to STP, and draw the circle HON, meeting the directrix in M. Join MN. sir = uXo =OMP =e, say. SP is parallel to OM.179 Again SIP = HON = NSIH = @, say. s. STis parallel to NM. ‘NM is parallel to PQR. Complete the parall-lograms OMPL and NMPQ. Considering the triangle OSL, we have oL sin OSL ain (OSF — PSF) O38 Now, OS = R, the circum-radius of ABO, angle OSF =. 4.40 (Gallatly’s Modern Gsometry, page 27), if the trilinear co-ordinates of S referred to ABC be sec x, sec 1, Sc v. The angle SPT = @ and PSF is 90 — 6. 2 OL = — Reos(a + pe + + 6)/sin & If p is the Jength of the perpendicular NR on PQR, we have p = NR = NQsio NQR = MP sin MPR = OLein¢ = —Roing co (a + we +u + 6) /eine. In terms of a single variable w, — , 1, v may be expreesed as 180° + ©, © ~ Band » + C respectively. The envelope of the line PQR is therefore = Rein $ cos (3 —B+ C+ 6)/sin 6. which is the tangential polar equation of a three-cusped hy pucycloid whose centre is at N, and the raduis of whose rolling circle is Resin ¢ /sin 6 If PQR is a tangent to the parabola, the sides of the triangle ABU will be the positions of the variable line PQR, and in this case the three- cusped hypooycloid is inscribed in the triangle ABO. Since SP will vow bisect the angle QPT, the triangle SPT is isosceles, SP being equal to PT ; and therefore in the similar triangle HON, HN = ON. This is the result of Question 1101. Next consider a triangle ABC and two transversala XYZ and XA, If 8 and 8’ are the foci of the parabolas escribed to ABC and touching180 XYZ' and X’Y'Z respectively, SX, SY, SZ are equally inclined to the sides of ABC, say, at an angle @, and likewise 8’X’, S'Y’ and $'Z’ at an angle 4, These lines being concurrent corresponding lines of directly similar figures described on XX’, YY’, ZZ’, tho points of coucurrence, namely, S and 8’, and the double points of the similar figures, being the foci of the remaining three patabolas touching the two transversal and two of the three sides of ABO, liv on the circle of similitude. This cirole is the well- Saown Mique’’s circle of the lines AB, BO,OA, XY% and X’Y'Z, If a three-cusped hypocycloid touches these tive lines, then XYZ and X’Y'Z! are corresponding lines of the parabolas having 8 and $" as foci; the angle © = 4, that is to cay, SX, SY, SZ are respectively parallel to S‘X’, S'Y’, S'Z', The circle of similitude, on account of this parallelism breaks up into the line SS’ and the line at infinity. If $,, Sy, S, are the double Pointe of the similar figures oa XX’, YY’, ZZ', then AS,, BS,, CS, will now be parallel and inclined to SS’ at the samo angle 9, aud the invariable points are at infinity on lines inclined to the sides of ABC at the angle 8. Tho tangential equation of the threescusped hypooycloid in Question 417 may be obtained thus :— IEA, 1, v be the direction angles of any line, the equation of the ver tex-tangent of the parabola escribed to ABC and having its focus at sec r, b sec 1, ¢ soe v is, in areal, L=o tan» tanv + 4 tanv tana + y tan atan 6 =0. The axis of the parabola is M =o (tan » + tan v) + @ (anv + tana) + y(tano + tan #) = 0, Also lett NSa + @ + y =0, the line at infiaity. Since the vertex-tangenis of escribed parabulas are corresponding lines, and likewise their axes, the line at iatiaity being a self-corresponding line, the lines tL+mM+nN=0 where /, m, x are constants, are corresponding lines of the parabolas for variable values of , yu, Tho tangential equation of the envelope of these lives is obtained from eliminating a, 12, v, p from the following equations :— Utan » tan uv + m (tan pw + tanu) +2 = px Utan v tan A + m (tau v + tan) + 2 = py Utan a tan p + m (tana + tan) +n = pz181 cos x ++ bcos p+ ¢ cos v =0 asind + bain, + coiny =0 ‘The steps in the elimination are omitted as being a little tedious, and wiiting y--2 = £, s—a@ = Vande the eliminant may be expressed, by geometrical consideratione, in the form nl—m*® LT cot A Stet metig= @ emis r{io cota — geot c) — mg } The envelope of the vertex-tangent I, = 0, is the Steiner's tri-cusp Deot A 2a = 0. The envelope of the axis M = 0 is the Steiver’s tri-cuep (of the anti-medial triangle of ABC) Toot A 2 (—#+y +2) =0. Tbe condition that the line 7L + mM 4+ 2N=0 touches the parabola is nl — m* = 0, and the envelope of the line will then be LEcot A Eta + meEE =O, a threo-cusped bypocycloid inscribed in ABO, and the centre of thie curve is UE @ sin A cos (B—c)—mYzein A sin (B—o) = 0, sin3A sin3B sin 3C si A’? einB’ sin the locus of the centre is the line ( ) which is the perpendicular bisector of OH. Question 1011. (SzLectev):—I1n an ellipse the tangent at P cuts the directrices in 4%, 7%, and the remaining tangents from Z, %’ to the ellipse meet at T, Show that PT ie normal to the ellipse and bisected by the minor a Solution by R. P. Paranjpye and K. J. Sanjana, The points vf contact of the tangents ZT and 7 'T are the other exe tremities of the focal chords through P. Considering the triangle PQS’, QT bisects the exterior apgle at Q and S'T the exterior angle at 8. Therefore T is the ox-centre of the tri- angle and PT bisects the angle QPS’ and is therefore the normal at P,182 Draw TM perpendicular to PS and TN and PK i __ Dr e perpendicular to th directrices ; PM is equal to balf the perimeter of the Triangle PQS" and, therefore, PM is equal to the major axie, i “SM = Ps’, s. TN/e = PK/e, TN = PK. 2. BT ie bisected by the minor azis. Question 1061. (Laxenauraxkr N, Bratt) :—Tbe Euler line of the triangle AB,C; meets the sides AB,, AC, iv By, C,; The Euler line of ABC, mects there sides in Bs, Cy3 and this process in continued indefinitely, 1f Ny ie the nine-point centre of the r' triangle AB,C, thus formd, end AN, meets B,C, in D,, prove that the straight lines Nz Ne+:y and Dz De+%y are parallel, « and y being any positive integers. Solution by K. Satyanarayana, By Question 1009 solved on page 185 of the J.J, M.S. Vol. XI, it follows that Bs Cs, BeiiCe42, BepiCs+n, are parallel. Hence As ABsCr and AB,+2)Oz42y are similar and N*, Ne¢2y, Ds, Daysy are correspond- ing points ~ AN: _ ANsssy * RD, = BDevy s. NiNasy is parallel to DsDy yy.183 Question 1100. (Laxsnauisnanxer N, Baatr):—ABC ie a triangle having its sides divided internally and externally in the ratios 2/9, g/r, 7/6. If the fouc lines joining the points of division, three and three, touch a parabola, then prove that (1) tho parabola can for no values of p/q, q/r, n/p pass through the intersections of the circum-circle and the nine-pointe circle ; end — (ii) if 4, 9, » are proportional 10 Vv {a (cos A—cos B eos C)}, ¥ {5 (crs B—cos A cosC)}, J {0 (cos C—cos A ens B)} then the focus will coincide with the nine-points centre. Solution by K, Satyanarayana. Let D, D’; E, B’; F, F’ divide reepectively BO, CA, AB intornally a2 PR Pe Co-ordinates of D, E, F w.r.t, A ABC are and externally in ratios o,——, 2 ,;— Py _d P ‘gtrigtrirey rtp pty pte Lines EFD’, FDE’, DEF’, D'E'E’ are — pz + qy + spe t+ qy—r2= 0; prt qy + If co-ordinates of P w.r.t. & ABC, and A DEF be z, y, 2; X, Y, Z, | | r | | 0 4 yeh] FD | Th gtr aver =| rep Ore yt) rg | oe | “re "te pr | g » : pra pry | q ? 0 | Lol pg pry = 29+) (pes gut rz) Zpq? —_xX__ Y pater ger TPE + gy + rz pe— gy tre pet gy oY, +p) (p + 4)" Pe + gy + ~184 se the parabola ie VaR + VP¥ JiZ = 0; whorer + p + v=0. x Y z —“ _=0, But D’ EB’ F ig —> __ + _*__ - “pat Get t re Ta” and because this touches the parabola, we get apa tr) + por +p) + vr (p +9) =0. Hence —*__ = f pn) 2qG—") 96?) 7-9) ©. the parabola is Yptgi—a)( pa + gyre + Ja =p) pea + + IP Te Fa : ” a“ at pap t poe which reduces to (i) If this should be Das ye + &[E (Bs + cs — at) at — 2S aty:] = 0, then k = } and . g-—7 = rz p? (@F PB at + O2— ct) (a + OF — 8) (+ 8? — a) = ora we erEw which cannot hold, since sum of denominators —[5 at — 2 ¥ be c%] or (a + b4c)(—a +b + 0)x (a —b + ¢) (a + b —c) can rever be zer»; * sum of any two of the quantities a, 0, ¢ is always greater than the third. (ii) 10 this cage equation of parabola becomes 2 Y Cw G+ ew *t Bowes e—Ht st (a? — b*) (a2 + b* — ct) By Example 7, page 310 of Ackwith’s Analytical Geometry, Be the co-ordinates of the focus are proportional to ara ere= a + aa) te ay ie., to (& — A) (as —b§— A) tg, few (A= a) (a -— 8 (2 + a — B) (a9 + — 2) ie, to (Bt — ADs (bt + ct — at) (BY + ct — a2), ote,185 But the co-ordinates of N. P. centre are proportional, a2 (b» + c*} — (b + 4)3, etc., which are not proportional to those of the focus. ‘The second portioa of the problem therefore appears to be incorrect. . Question 1122. (V, Bamaswain Aivan):—If the join of foar concyclic points A, B, O, D takea io pairs intersect in P, Q, R, prove that the N.P, circle of PQR passes through the centroid of ABCD, Solution by K. Satyanarayana. Oise the centre of the circle ABCD. X, Y, 4, W, L, M, N ate the mid.-points of AB, BC, CD, DA, QR, RP, PQ, respectively, G being the centroid. From quadrilaterals RCPD, and QCPB, respectively, it follows that MZX, NYW are straight linee, Also, from the eyclic quadrilaterala, £X7Q = 4620 = 2QO0x, Z4WYR = ZGYC = ZROW, £DCB = £Z0Y = ZWOX; oe 2£7GY = ZMGN = ZROQ But ZROQ is the supplement of Z RPQ or ZMLN since O is the ortho-centre for A PQR, Hence ZMGN + £MLN = 2 right angles. Thus G lies on the circle LMN, which is the N.P, circle of PQR, 24186 Question 1163. CV. Bamsswans Aryan) :—If a rectangular hyperbola passes through the ir-centre of a triangle and the feet of the perpendiculars drawn there- from to the sides, prove that it cuts the inscribed circle again at the point which is diametrically opposite the Feuerbach point. Solution (1) by G. A. Srinivasan, Reciprocate with respect to the ir-circle. The rectangular byperbola reciprocates into 2 parabola which touches the sides of the given triangle and whose directrix passes through the in-contre. We have only to prove that the fourth common tangent of this parabola and the in-circle is parallel to the Feuerbach tangent, Let the equation to the parabola (in areule) be Jie + Jay + ya 20 ws - @) whore Atpto=0 os ww QD Ita direotriz is * Ae — al) opp (8 + at — By + v (a2 +B? — ect) =O, Since this passes through the ic=2entre (a, 5, ¢) noon A + cos B+ vcosC = we (8) From (2) and (3), we obtaia x =H (@=o@—4) f—a)(s—b) (@=b)( —4) Let the fourth common tangent of this parabola and the iv-circle be la + my faz = ae wee wee awe (4) Then a) (b (—a) b+e—m” # [Vide Fx, 14, p. $19, Askwith’s Analytical Geom. of the Conic Section (old edition)}187 This linetis evidently parallel to the Feuerbach tangent of the in-circle given by + Remarke.— Tho equation to the rectangular hyperbola in question is easily seen to bo Eb —e) (s — 0)?29 + Sb — 0) (8 — 8) (8 — 0) yz = This conic passes through the point whose co-ordinates are 1 1 ‘p= ie, the Gergonne point or the point of concur- Fb’ s— rence of the lines joining the vertices to the points of contact of the in-circle with the opposite sides. Solution (2) by S. L. Malurkar. Reference :—I in-centre of ABC. 4 DEF pedal A of 11. ® the ortho-centre of A DEF, S the circum-centre of 4 AaBo. Tho Feuerbaca point for the in-@ of A ABC, ““G wf The diametrically op- posite point, A’BY, C’, the middle points of BC, CA and AB, We know a’ (DEFI) nF wd sin Fx’ 5 = sin DFE sin Fn’'x sin Dwr ‘ sin FDR . cin ar ofr FA = 2 in atornate segmeut sin_wFC" * sin DA’ eee _ ~~ + [Vide Askwith, 2, 0. p, 808}.188 = ‘in DFE | sin Cis +: Q A'rD is similar to A SAI and sin FDE * aa alS AC mt Fann ..& SCL. sin DEE sia SOL +: AS = CS and 1S common in the sm FDE * an Sal 4 SIA and 4 SIC. sin DES Fe . a ow = Ge PDR * sin IDS s 41D ¢ = 4 { DFE—DEF } . = 4(40— cB] = Sal. sin DeB | sin J@P = +; IF = ID and 1$ common in the an FB * si 1eD triangles 1$F and IpD, and ¢ is the orthc-centre of A DEF. = ¢ (DEF). sm, D, B, F, I, lio on a conic. But any conic through the vertices of a A and its ortho-centre is a rectangular hyperbola, s. ¥',, D, E, F, 1, are on a rect. hyperbola, i. The byperbola (equilateral) through D, E, F, 1, passes through 7’. QED. Similar relations are also true for the ex-circle, Question 1179. (Connsorsy) :—From an external point I (2, y), tangents TP, TQ are drawn to the conic S = ac? + 2 hay + byt + 2 gn + 2ty + c=0 prove that the area of the triangle TPQ is and that the area of the quadrilateral OPTQ, where O is the centre of the Additional Solution by F. H. V. Gulasekharam, Lot us use Capital letters X, Y for current co ordinates. Suppose TP = a, TQ = B, PTQ = w, area of the triangle TPQ = M. The original equation to the conic is aX?¢2QhXV+¢bVe + WX4+2Vtoso0 ww (1)189 ‘Transferring tho origin to T (2, y), the equation takes the form aX?+ Axe UY ¢ 9X7 FV $¢S50 .. @) Again transforming the equation, referred to TP, TQ as axes, it takes the form Ss [ &+ 5-1) ‘+20x¥] = For, the term independent of X and Y in (2) and (3) must be identical, seeee(3) Now from (1), (2) and (3), the invariant properties give Qsintw = — 8802... ve we C sin? w = —S'a (a + 3) es ww (8) also 2M = ad sin w ve we (8) From (4), (5) and (6) we immediately get ‘ s? y=a M=~.“S. Again remembering (i) that the equation to PQ (referred to the original axes}isz (aX +hY tg) tyihX +b¥+/) + (gX + FY +c) =0, G (ii) that T and O are the points (#, y) aud ( o é) respectively, aud (iii) that T and O are on opposite eides of PQ, A OPQ _ ATPQ _ Quid. OPTQ ac = os = a we get o7s Hence the quad. OPTQ = foas Note ;—Io the question, as originally proposed by me, the expression for the area of the quadrilateral was wrongly given. ‘The correct expression is given in J.1.M,S. Vol. XIV, p, 87. The original expression gives the urea of the quad, O'PTQ, where O' is the origin,190 Question 1183. (B. B. Bagi) :—The circles round AQR, BRP, CPQ, where P,Q, 8 are points in order on the sides BC, CA, AB of a triangle ABC meet in O, If A’, BY, C’ are the middle points of the arcs QOR, ROP, POQ, then show that A’B’C’ is similar to the triangle formed by the ex-centres of ABC. Also show that A’, B’, C’ and the in-centre of ABC are conoyclic, Solution by V.V. 8. Narayan and K, Satyanarayana, We have ZC’OB! = LC’OP + ZBOP = 4 (C + B), ZC'LB' = Supplement of } (C}+ B), where 1 is the in-centre. Hence I, B’, O, C’ are concyclic. Similarly I, B’, 0, A’ can be shown to be concyelic, Therefore A’, B’, C’, I, O are concyclic, Consequently the angles A’, B', C’ of the A a’B‘C’ aro easily seen to be equal to 3 (B+ C), 4(C + A), 4 (A + B) respectively, which proves that the A A’B'C’ is similar to che ex-ceutral & of ABC, Note :—The property given ia the question is extended by the fact that the point of coucurrence O of the circles is also showa to be concyelic with A’, BY, O', 1. Question 1187. (FH. V, Gurasskganau) :—If f, g, & bo the lengths of the bieectors of the angles A, B, O respectively of a 4 ABO, prove (i) (9? + 42 — 2yh ova (B—O)}# + [424+ "2 — 2Af coo} (CAE + (f? + 9? — fg cos$ (A — B)}3=0. = 0. Gy =O 4 BECI A) 4 dae Solution by K. Satyanarayana, A, Makalingam and the proposer. Let AF, BG, CAl be the bisectors respectively equal to f, 9, h. Wo chall first prove {ii). A 2c.c08 * 2a Since = SFO *O4aein® 2 (6 + ¢) cos vin191 i in 0) _ (6 +) b— 0) = to in B in ee Now, to prove (i), draw AQ’, AH equal to 3 (A ~B), } (A—Cj with AF. Then by virtue of that F@’H’ is a straight line, Further FG = y (/* + g* — 29 cos} (A—B)), &e. Hence (i) is ideotical with (FG’ + GH’ + H’F) which is zero, if we take the sign into consideration, A making angles , it is readily seon Question 1191. (G. V. Tetaxc):—ABC is a triangle inscribed in a circle. Three tangents ave drawa to the circle so that the portions of thew cut off by AB and AC are bisected at the points of contact P,Q and R. Show that the orthocentre of the triangle PQR lies at the mid. point of BC, Solution by K, Satyanarayana, Let the centre of the circle be the origin and let A,B, be the Points (0), (0), (8) respectively, the radius of the circle being a. Let (A) be tke point the by AB, AO, is bisected at a. AB, AC are given by portion of the tangent at which intercepted (© + ytan 5 — a) (e+ y tan —~a) 20;192 a 8 or at + ay (tan 5 + tao $) + yrtan Stang — 2ar— ay (v0 3 + tan 8) +=, we (1) Equation of the tangent at ‘a’ may be written as &—acos cos (4a) sin (B+) r being distance of (w, y) from x. Where this cuts (1), the eum of the values of » = 0. Honce after reduction 4sint 50005 + (tan 5 + tan A) (sior 3 — 4 sine 5 cost} ) + @tan Stan Z sin 5 cos 5 (1-2 sin 5) = Taking away the factor sin > and dividing by cos* 51 We Ret a 8 (# — 32) (190 gt te BN 4 (2 — tan Stan 8) 2 a“ 8 t » (2) + 2tan 5 tan 5 = 0, where t etan 5. ven (2) is a cubic in ¢ giving 3 values for % and each value of tan} gives only one value of x between 0 and 2. Thus we have 3 points satiefying the condition. Let the three values of \ be 4y \4 and Xs. If ty, ta, ty be the roots of the Equation whee 2 (2-1an§ tan 2) : e B tan 5 + tansy 2tany tan 4 Sty = — 354 ets = — Se tan 5 + tan >198 . Apt a+ s) 2 Sh — thtyt, oe tan ( rr y=: Bhite . a+8 =—cot —>— tan ( g+ Mata trent lo r+ a4 8. o Now the ‘ m’ of the join of a, to mid.-point of BC is & a—Bs cos we (8) - @. sin x, — sin —> a+ 08 a, — Cos —y— 08 2, — sin E48 +8 03 *=8 = (1=1,2) — cos a cos while the ‘ m’ of the join of 49, 4g = and this reduces to Honce the condition for the perpendicularity of the former and the latter ie [2s —sin $5 0 os STF gy ay] [sa 24? — 1, cos a+3 + [a — 1,2) — cos of B cos a+2é [2s ty th or (t,#— 3¢,) sin s+8 + 42 [s cos oye + [ex OF — cn 858] = 0 which is satiefied eince ¢, satiefies (2), and (2) easily reduces to the above form. Hence the mid.-point of BC is the orthe-centre of the triangles whose vertices are the points % 4, 4, ds. 25194 Question 1199, (Cf. H. V, Gutasexeanam) :— Prove that the radius of the pedal circle of a point P with respect to a triangle ABC is givon by Bain A sin B sin C cos PAB cos PBO cos POA + cos PAC cos POB cos PBA’ where R is the circum-radine of ABC. If tae pedal triangle of P is in perspective with ABC, prove that the expression above reduces to yx fin A sin B sin © cos PAB cos PBU cos PCA * Solution by K. Satyanarayana and X. J. Sanjana. Let us firat prove a Lemma: If D, E, £ are any three points respec- tively on BC, CA, AB, then, AF.BD.CE + AE.CD.BF=4R.A DEF For, denoting BD, OE, AF by 2, y, 2, 4R. ADEF = 4B [4 ABO — A ABF — A BDF — A CDE] = abe — (b— y) 2a — (0 — 2) 2b — (a — 2) yo = abo — bow — cay — abs + ayz + ber + eny = ays + (a — 2) (6—y) (o— 2 Denoting the pedal A in the problem by DEF and its circum-radius by R's 4B. O DEF = EF. FD, DE = PA. PB, PC. sin Asin BainC, or 4B. DEF. # = (AF. BD.CE + AB.OD, Br) = PA. PB. PC, sin A ain Basin 0, in Basin C Rein “AF BD CE, AE BF OD Pa’ PB’ PC" PA‘ PB’ PC B sin A sin B sin _. = cos PAB, cos PBO cos PCA + vs PAC cos PCB cos PHA If DEF and ABC be in perspective, AF. BD .CE = AE. BF .CD and hence the second result follows. or Rr195 Question 1201. (Manze M. Tuowas):—A periodic comet when at an angular distance @ from the poribelion of its orbit, of eccentricity ¢ suddenly encounters a resistance which brings it to a standstill, Show that it will fall into the sun in time —* (te) 3, whore ¢ is the periodic 4/9 time of the comet, Solution by Mf. V. Ramakrishnan. Since the comet is reduced to rest it moves towards the sun in @ straight line, Hence 3 ry, = 7 2 Jap where y represents the radius vector to the comet in its position in the orbit when it ie reduced to rest. Now, l a (1 —?) Teco 9 “1+ 0008 6" 2. (4ae)! l ( le ) “4 a \T#e cos 6 t l—@ \¢3 t (1) z 4/2 (RS) Question 1213. | (V. Ramaswamt Alvar) :—Prove that the Pedal circle of any point P with respect to a triangle ABC cats the side BC at an angle equal to the complement of the sum of the angles PAB, PBA, PCB, Solution by F. HV, Gulasekharam. With the notation of J.1.M.S. Vol. XLU, pp. 210—218, from 5S 4 and 5 ibid, the perpendicular distance, from BC, of the contre of the pedal circle of P is 2 RG [cot + cot 2’ + cot y 4 cot v']196 — Pio v sin 8 ciny ein A cot @ + oot A’ + cot ® + cot y | Ci p [fee -tela os sin (A + ety) sin (A~a) 8 jin (8 +") and oa Pp [ «ia © cos'(4’ + y"} + cosa sin (8’ + ”] =P sin ia + p+ yy Hence the question. Note :—The normal co-ordinates of the centre of the pedal circle are proportional to sin (a + 4 + y'), sin (0’ + 8 + y'), sin (a’ + 2 +y). Question 1244, (8. Basawanayanan):—Sum the series (—1)o Gp — 4 Gr-1 + (A+ 1) ay-2 — ... (L—1)" (a-+-r—1),9 where n, denotes the number of combinations of » things r at a time, Solution by K. Satyanarayana, The series ia the co-efficient of a” in Gs Sy (14 @) 2-4 ay 1 BT 114 x)* + aye ary a+ atte wet (= 1) 2, e(l taytret ie (D+ at lame t at’ . $(—D a, a" (le ay + Cty a a tal toe + 1)" 2! te] a ies (—D" +a)" (L+e—a)" jue (HH) (La the sum of the series = 0,197 Question 1245. (S, Rasayanayanay) :—Find the value of the infinite series where S, denotes the sum of the 7" powers of the firat x natural numbers, Solution by K. Satyanarayana. The series = = ¢€—N)+t @—)I +(e? — I +... + (1) e [=i]: Question 1246. A (S, Rasananavawan) :—Find the value of the expression vlat v Labt y (ab + Vab'$ Jab p Sabit Fd Solution by K. Satyanarayana. Denoting the value of the expression by a, it is easily seen that ax a z oo Fats . = wv? Sora Pt las’,198 QUESTIONS FOR SOLOTION. 1254, (@. S, Manasan) :—At a cortain place A, on a certain day, my = mean solar time of sidereal noon, % = sidereal time of mean novn, show that on that day the ‘ mean’ and ‘ sidereal” times (m, s) at any other place B, where L° is the difference of the longitudes of B and A, are con- nected by the relation mt L/15 Mo + eU/lb Ly 1255, (T. Visayaracuavan) :—Show that lim B ['m, 122 ton m =e log log m ] =e, tem moro where B (p, 9), as usual, denotes the integral f. aP-1 (l—w,2-! dz, ° 1258, (7, Visaranaauavan):—Let ema, + 1 _! ayt as+ be an irrational number and a (0, n) be defined by the equation eat | 2 1 ‘gn AOn).gn? * where ©” ig the nth convergent in the continued fraction ; if qn (6) = ia (8,7), n> @ and if 3 occurs an infinity of times in the set (a,), then show that . 65 + 9V8 tk@2 es except for a particular claes of quadratic surds, namely, those whose periodic part when expressed as a simple continued fraction consists ouly of 3's, 1257. (A.C. L. Witxtssun).—Solve the partial differential equation v (1—g?)* + 2p9s (I—g*) + p9q3t — (1—g*) (r¢-8*) 2 = 0.199 1258, (A. ©. L, Witxrncon):—Solve the partial differential equation zy (L—a*)r — (2ety*— 2? —y, +1) e+ay(I—y?) t+ xy (s—p2—gy)=0, 1259, (P. V.Ssenv Aryan) :—Suggrsted while reading about ‘the theory of monopolies’ in ‘ Economics,’ (1) Show that the values of for which the function v = xf (z) is a maximum are the abscistw of the pointe where y = f(s) touches a mem- ber of the family of rectangular hyperbolas 2y =. (2) If y =f () touches a member of the family zy = C at a point k . whose abscissa ie =, show that the curve y = f(z) — 7 where & is acon. stant touches another member of the family at a point of the eame abscissa %,; and find the ‘¢’ of that other member. 1260. (A. T. Tuowas):—1f a, = sin «, ag = cin ein 2, a, = sin sin (m times) ....... 2, where OZzém, then prove that the in- e « Gite series $ ay, and $ a,? are divergont, but Sa," is convergent, and 1 1 ° 1 Bent i A v3 in fact when w ie largo a,~ So. 1261, (A, T, THomas):—If the oumbere represented by # (x,y) where Fis an integral exprossion with integral co-efficients of the positive integers ,y, are arranged in increasing order of magnitude, and considering those lees than N, the probability that 2 and y sre prime to each other ia a number . 6 tending to yas n>. 4 POR (AT. Taoxtas) :—IF (2) means the fractional part of 2, and if ea crational number >I, show that such membars exist that (2") tende to 2010 steadily a8 n gous to co through positive integral values. 1263. (I. Torapet [venoan) -— Sum the series ; lim lz 1 (12) + Ti 1264. (A. Nanasinca Rao) + Snvelope of ovals similar to A and centres on B and a constant ratio of ‘A and B are two ovale. C is the milarly placed, baving thoir homothetig similitude x. Prove that200 (i) C is alao the envelope of ovals homothetic to B, having the homothetic centres on A and 4 ratio of similitude 1— > ; (i) if P, Q be points on A, B and the tangents at these points are parallel, C divides PQ in the ratio (1— %): >; (iii) if p = fi¥), p = gi¥) be the intrinsic equations of ‘A and B, then that of C is ss a.fe) + (L—a).g(W)5 (iv) if masses a, (1 — a) be placed withio or on the boundaries of A and B, their centre of mass will be withia or on C ; also every interior or boundary point of C is a possible position of centre of mass. 1965. (A. A. Kntsnsaswami [veNaar):—Show that every odd num- ber can be expressed as the sum of seven squares, except the numbers 1, 3,5, 9, 1, 27, 1266. (A. A. Kntsnnaswaat Ivexcar):—Find general expression for the sides of rational triangles, the equares of whose areas are perfeut cubes, (Ex. 5, 6, 7.) 1267, (S. RasawarayanaR) :——Prove that, if " Oise +e, 1 in im then e™+" = c" + ", and conversely. 1 non 1 1268, (S, Rasanamayaway) If e9 sehen in’, then a? *? a ni + n°, and conversely. 1269, (R. Varprawatuaswanti):—Solve the equation a bape aly? + bh yt ch ax? r otherwise shew that the irreducible algebraic curves which Hence o1 a r gents belong to a given linear com- lie on a given quadric and whose tan; plex, can always be reduced to the form dm = fo" (m prime to 7) where d, are generator co-ordinates on the quadric (i.e, the point Qe) on the quadtic is the point the parameters of the generators through which are x, J. Find the invariant ré " order that one such irreducible curve may be a twisted cubic. elation between the qnadric and the complex, inCaN anh wn Hw eIaQantraoeR AS LIST OF JOURNALS RECEIVED BY THE SOCIETY Messenger of Mathematics Quarterly Journal of Mathematics Mathematical Gazette The Annals of Mathematics American Journal of Mathematics Bulletin of the American Mathematical Society Transactions of the American Mathematical Society Monthly Notices of the Royal Astronomical Society Proceedings of the Royal Society of London The Philosophical Magazine and Journal of Science Astrophysical Journal Crelle’s Journal L’intermediaere des Mathematicus Mathematische Annalen Philosophical Transactions of the Royal Society of London Acta Mathematica Popular Astronomy Proceedings of the Edinburgh Mathematical Society Proceedings of the London Mathematical Society Mathematics Teacher Bulletin of the Calcutta Mathematical Society The Tohoku Mathematical Journal Nature The American Mathematical Monthly Proceedings of the Benares Mathematical SocietyThe Indian Mathematical Society (Founded in 1907 for the Advancement of Mathematical Study : and Research in India.) THE COMMITTEE. President : 3 Baik Raw, M.A, 1.0.8., Judge, Bijapar. ‘Treasure: S. Narayana Auvaz, M.A, FLS.S, M Office’ Mattar a ” . Manager, Port Trust Librarian: V. B, Naix, M.A.; Protessor of Mathematics, Fergusson College, Poona. Seoretaries : (B. V. Sesuo Aivas, B.A. LT, Professor of Applied Mathe- matics, Presidency College, Madras, M. ‘T, Naranienoar, M.A., Professor of Mathematics, Central College, Bangalore. Additional Members: Tho Hon'ble R. P. Paranseye, M.A., D.Sc., Minister of Educa- tion, Bombay. E. B, Ross, M.A., Professor of Mathematics, Christian College, Madras. D. D, Karavi, M.A. B.Sc. Professor of Mathematics, Elphinstone College, Bombay. Heweas, M.A., Professor of Mathematics, Deyal Singh College, Labore. ‘K. Axanpa Rao, M.A., Professor of Mathematics, Presidency College, Madras. ‘tne Hon'ble V. Rawesam, B.A., B.L., High Court Judge, Madras. ‘A. C. L, Witginson, M.A., Principal and Professor of Mathe- matics, Deccan College, Puone. OFFICERS OUTSIDE THE COMMITTEE. Assistant Secretary: N, Racavnaraa Atvancas, M.A, Assistant Professor of Mathe- matics, Presidency Cullege, Madras. Assistant Librarian. V. A. Apre, M.A., Professor ot Mathematics, Fergusson Cullege, Poona. . Associates : K, B. Mapgava, MA. A. Nagasinaa Row, M.A. KS. Patracuariag, M.A., L.T."| ‘It. K, Venkataraman, M.A, LT. Head-Quarters: Poona,