A 544533 The Gift of WILLIAM H. BUTTS, Ph.D. A.B. 1878 A.M. 1879 Teacher of Mathematics 1898 to 1922 Assistant Dean, College of Engineering 1908 to 1922 Professor Emeritus 1922 RADCLIFFE OBSERVATORY OXFORD. ! • A. Robertion. Observatory, Oxford. Emerson. williane + Cyclomathesis Q A V.5 ARTES 1817 SCIENTIA VERITAS LIBRARY OF THE UNIVERSITY OF MICHIGAN FLUNIOUS UNUM TUEBOR 51-QUÆRIS PENINSULAM-AMŒNAM CIRCUMSPICE Emerson, Willi ane r Cyclomathesis THE ARITHMETIC OF O F INFINITES, AND THE DIFFERENTIAL METHOD; ILLUSTRATED, BY EXAMPLE S. Hæ tibi erunt Artes. VIRG. .9 RADCLIFFE OBSERVATOR LONDON: Printed for J. NOURSE, in the Strand, Bookfeller in Ordinary to his MAJESTY. M DCC LXVII. OXFORD Den hib Milr [iii] Profion Williin. H. Bitio 10.14-1985- THE PREFACE. THE method which the antients generally made ule of for finding the areas of figures, or the fo- lidities of bodies, was by the help of other infcribed and circumfcribed figures; after fuch a manner, that the infcribed figures being continually divided into very Small parts, the fum of the infcribed parts, approach- ed fo near the true figure, as to fall short of it by a quantity less than any given quantity. And in like manner the circumfcribed figure was divided into fo many small parts, that their fum exceeded the given figure, by a quantity alſo less than any given one. So they always had the value of the figure propofed, between the two values of the infcribed and circum- fcribed figures; and by this method found out its con- tent. But this way being for the most part ſo very laborious, that it put the fucceeding geometers upon finding out fhorter methods; and thus came up the Method of Indivifibles, or the Arithmetic of In- finites. This method folves thefe forts of Problems with admirable brevity, and extreme facility; and that by a help of a few Propofitions, which reduces the whole bufinefs to the fumming up the powers of an arithmetical progreffion, confifting of an infinite number of terms. As we ſhall have frequent occafion for this afterwards; I have here delivered that me- thod A 2 4 PREFACE. THE thod at large, with many examples to fhew the ufe of it. The whole is contained in Sect. I. of the follow- ing Book. I have likewife given the principles of the diffe- rential method, and the Interpolation of Series, in Sect. II. an art extremely useful in almost all parts of the Mathematics, efpecially in Aftronomy. For it is frequently required to find the pofition of a planet or comet at fome given time when an obfervation cannot be made; but by help of other obfervations, we can find it to that time, by the interpolation of feries. This method I have also made very clear and plain, and illuftrated the fame with variety of examples. W. Emerfon. [ 1 I ] SECT. I. The Arithmetic of Infinites. T fumming up HE Arithmetic of Infinites is the art of fumming up the powers of a ſeries of quantities in arithmetical progreffion; whofe number of terms is infinite, and their com- mon difference infinitely ſmall. It is alfo called the Method of Indivifibles, be- caufe magnitudes are here fuppofed to be refolved into their indivifible parts, or fo far as there is any occafion to confider them as fuch. PROP. I. If a ſeries of equal quantities be added together, the fum of all the terms is equal to any one of them multiplied into the number of terms. For a+a+a... to ʼn terms is n nxa. PRO P. II. If there be a ſeries of quantities in arithmetical progreffion, and continued ad infinitum; as o, a, 2a, 3a, 4a... to na; the fum of all the terms is ina, or to half so many times the greatest. For by (Prop. 7. Arith. Progreffion) the fum =o+na naxn+I xn+1, or but fince n is infi- 2 2 Na n nite, n+1 =n. therefore the fum == xna = X 2 laft term. B Cor. ล ARITHMETIC Cor. The greatest term multiplied by the number of terms, is equal to twice the fum of the feries. PROP. III. If there be a ſeries of Squares, whofe roots are in arithmetical progreffion beginning at o, and continued 2 I 2 2 ad infinitum, as o, a², 2a, 3a, 4a за the fum of all the terms is nxna", times the greatest. = 3 For let so+a²+2a²+za 2 I 2 to na ; or / fo many 3 2 to na", L = laft or greateſt term. Then by the Method of In- crements, s=n+1.az², putting n+1.az. But (by Cor. Ex. 10. Prop. XIII. Increments), the xn+1 &c. Or writing z for n+1.x 2 integral of x = x, and 2 for n, the int. of x2 = 3 Z 3Z Z 2 &c. But fincen is infinite, zis infinite,and z³ is infinitely greater 2 than z²; whence int. of z², that is s= 1. and naa; therefore s = א W 3º Z3 32 But n=1, 3 n+1.a зв But as n is infinite, n+1=n; 2 n+1.a X n+1.a 3a 2 and s=a na х = n × na = {xnL. na за Cor. The greatest term multiplied by the number of terms, is equal to thrice the fum of the feries. PROP. IV. In a series of cubes, whofe roots are in arithme- tical progreffion, beginning at o, and infinitely conti- nued; as o, a³, za³, za³, 3 за 3 to na³ ¿ the fun of Sect. I. OF INFINITES. 1 + 3 1 of all the terms is = nxna, or ſo many times the greatest term. 3 3 3 3 For put so+a³ +za³ +зa³... +ña³, L= greateſt term, then by the method of increments s=n+1.a = ׳, putting n+1.a=z. But (by Cor. Ex. 10. Prop. XIII. Increments). The integral of 3 xn+1 &c. or putting z for x, xn 2012 = n+I.x 2 24 23 for n; the int. of z³- 42 22 | x, and 3 &c. But n = 1, and z=na=a. Alſo fince n is infinite, z is infi- nite, and 24 is infinitely greater than z³; and n+1=n. па Z4 4 4 Whence int. z3 or s n+1.a = na 42 40 4a 49- x na³ = nxL. Cor. The greatest term multiplied by the number of terms, is equal to four times the fum of the feries. P.RO P. V. If there be a ſeries of mth powers, whofe roots are in arithmetical progreffion, beginning at o, and conti- nued to infinity, as o, a", za", 3a", 4a". n M+I the fum of all the terms is = Xfo many times the greatest term. × na to nam, I m = m+ I m Let s=0+a+za™ +za”..: +na”, L = laſt term; and by the method of increments, s=n+1.a =zm. And (by the fame Cor. Ex. 10. Prop. XIII. Increments) s= int. x = zm+1 m Z &c. But M+IZ B 2 2 N=1, ARITHMETIC n1, and znaa; and fince n is infinite, n+1.a na, and zm vanishes, being infinitely lefs than m+i Z zm+I m+1 ; therefore s n+1.a m+x.a m+1.a m+1 na na X-m n I na. X m m+1.a m+1.a m+I na M+I XnL. Cor. The greateſt term multiplied by the number of terms, is equal to m+1 times the fum of the feries. PRO P. VI. To refolve a Problem by the Arithmetic of Infinites. The chief uſe of this method is to find the con- tent of geometrical quantities, as of fuperficies and folids, by reducing them to their indivifible parts, or firſt elements; and then finding the fum of all thefe, as directed by the following RULE.". Suppofe a line to be compofed of an infinite number of points; a furface of an infinite number of lines, drawn parallel and equidiftant from one another; and a folid of an infinite number of planes, parallel and equidiftant; then obferve what fort of a feries it is that makes up that quantity. Find the fum of that feries by fome of the forego ing propofitions, which will be the content required. SCHOLIU M. Here lines are fuppofed to be made up of an in- finite number of equidiftant points; plain figures of an infinite number of parallel lines, like the threads in a piece of cloth. And folids of an in- finite Sect. I. 5 OF INFINITES. finite number of planes, like the leaves in a book. Fig Yet theſe points, lines and furfaces, are not really fuch, but are called ſo by reaſon of the fimilitude. For theſe points are in ftrictnefs, infinitely fhort lines; the lines, infinitely narrow parallelograms, and the planes that compofe the folids, infinitely thin folids; fo that whenever we ſpeak of points, lines, or planes, compofing any geometrical magnitude, it muſt always be underſtood in the ſenſe here men- tioned. Moreover, if all theſe parallel lines that make up a furface be perpendicular to the axis of a curve; then thefe parallel lines will cut the curve, or any other line, at oblique angles, and the points ſuppoſed to conſtitute that curve, or that line, will be greater than the correfpondent points in the axis, in fome certain proportion; which proportion is to be inveſtigated by the property of the figure. Ex. I. To find the area of the Parallelogram ABCD. Draw AP perp. to the baſe CD; divide AP into an infinite number of equal parts, thro' all which ſuppoſe ſo many parallels drawn to the baſe CD. Then all theſe equal lines compofe the area of the parallelogram, and their number is denoted by AP. Then by Prop. I. the fum of all is one of them multiplied by the number of all; that is or the baſe multiplied into the height. Ex. 2. CDXAP, To find the content of a cylinder ABCD. Draw AP perp. to the baſe, which divide into an infinite number of equal parts; thro' which draw as many planes parallel to the bafe CD. Then all thefe equal planes compofe the folidity of the cy- linder; whoſe number is AP. Whence by Prop. I. the fum of all is one CD x their number AP; that B 3 1. 2. ARITHMETIC Fig. that is, the cylinder 3. 4. plied by the height. CDXAP, or the baſe multi- And the fame way the folidity of a parallelopi- pedon or of a prifm is calculated. Ex. 3. To find the area of a triangle ABC. Draw AP perp. to the baſe BC; divide AP into an infinite number of equal parts; thro' which fuppofe lines drawn parallel to BC. Then theſe lines make up the triangle, and are a ſeries of terms in arithmetical progreffion beginning at o, and end- ing at BC; and the number of them is denoted by AP. Then (by Prop. II.) the fum of all the terms APXBC; that is, the area of the triangle is = 2 APXBC 2 Ex. 4 To find the folidity of a cone ABC. Draw AP perp. to the bafe BC; divide AP into an infinite number of equal parts; thro' which draw planes parallel to the bafe BC. Then theſe planes make up up the cone; and their diameters are a feries of quantities in arithmetical progreffion, beginning at o, and ending at BC; therefore the planes or circles themſelves are proportional to a feries of fquares, whofe roots are in arithmetical progreffion. Whence (by Prop. III.) the fum of all the ſeries is =AP× baſe BC; that is, the cone is height > baſe. • 1 3 And the fame way it is proved, that a pyramid is the height x by the baſe. Ex. Sect. I. 7 OF INFINITES. Ex. 5. To find the furface of the right cone ABC. Draw the axis AP, and DF perp. to it, and put AP=b, BP=b, AB=s, AD=x, c=3.1416; then by fimilar triangles, AP (b): PB (b) : ; AD (*) : DF = bx Ђ and ; 2cbx b = periphery at F. Then the fum of all the peripheries at F makes up the fur- face of the cone. But the points in AB being greater than thofe in AP, in the ratio of AB to AP. Therefore the periphery at F muſt be in- 2scbx 25cb creaſed to or bb bb Xx, and then the fum of all theſe peripheries make up the furface. But the fum of all the xs is 0, 1, 2, 3, &c. to AP; that is (by Prop. II.) APXAP or AP². Therefore 25cb the conic furface is ference of BCXAB. bb x bb = bcs circum- ½ = Otherwife. Put BPb, ABS, AFz, then by fimilar bz triangles AB (s): BP (b):: AF (x): FD = ; S zcbz ther = periphery at F; and the fum of all S theſe peripheries make up the furface. the furface. Now all the z's conftitute a ſeries of terms in arithmetical pro- greffion, whofe laft term, and alfo the number of terms is s and therefore the fum (by Prop. II.) is ss. Therefore the fum of all the peripheries, or 2cb S the furface of the cone is = מss = cbs as be- fore. Fig. 5. B 4 Ex. 8 ARITHMETIC Fig. 6. Ex. 6.. To find the folidity of any fegment of a sphere ADF. Draw the diameter AB, and DEF perp. to it; and put radius AC=r, AE =x, DE=y, c = 3.1416; then by the nature of the circle yy=2rx-xx. Then if AE be divided into an infinite number of equal parts, and planes be drawn thro' them parallel to DF; the fum of all theſe circular planes make up the fegment. Now one of theſe circles is cyy or 2crx-cxx ; and all the x's conftitute a ſeries of terms in arithmetic progreffion, whofe fum (by Prop. II.) is xx. Alfo the fum of all the xx's (by Prop. III.) is x³, where the greateſt term is xx, and number of terms, x. Therefore the fum of all the cir- I I cles conftituting the fegment, is 2crxxx-cx*³, or crxx CX3 3 Therefore the fegment DAF = crx-cxx xx: that is, the fegment DAF is equal to a cylinder whoſe height is AE, and baſe crx-cxx, or cyy+cxx = circle DF+ circle AE. I 2 I ठ 2 3 Cor. The hemifphere, as likewife the fphere is the circumfcribing cylinder. For when xr,the fegment becomes crr-crr xr = +crrxr. Ex. 7. To find the furface of the fegment of a fphere DAF. Let CA=r, AE=x,DE=y, c= 3.1416; then the circumference at D2cy. And the fum of all the peripheries make up the furface of the feg- ment DAF. But the points in AD being greater than thoſe in AE, in proportion of the tangent TD to Sect. I. OF INFINITES. 9 to TE, or of DC to DE; therefore y: r:: 1: = magnitude of a point at D. or 2cr is the true periphery at D. all theſe peripheries Whence 2cy X ર y And the fum of furface of the fegment; and the number of them is denoted by AE. Therefore (by Prop. I.) that fum = 2crxAE; that is, the furface of the fegment DAF = 2crx. Cor. The furface of the hemifphere is 2crr; and of the whole fphere 4crr, or four great circles of the Sphere. **** *** ## SECT. 10 THE SECT. II. The Differential Method. 'HE Differential Method, is the art of working with the differences of quantities. By this me- thod any term of a feries may be found from the feveral orders of differences being given; or vice verfa, any difference may be found from having the terms of the feries given: it likewife fhews how to find the fum of fuch a feries. And it gives rules to find by interpolation, any intermediate term, which is not expreffed in the feries, by hav- ing its place or pofition given. V When any ſeries of quantities is propoſed, take the firſt term from the ſecond, the fecond from the third, the third from the fourth, &c. then all theſe remainders make a new ſeries, called the first order of differences. In this new ſeries take the firſt term from the fecond, the fecond from the third, the third from the fourth, &c. as before; and theſe remainders make another feries, called the Jecond order of differences. In like manner, in this feries, take the firſt term from the fecond, the fe- cond from the third, &c. And thefe will make a feries, called the third order of differences. And after this manner you may proceed as far as you will. Thus in the following propofition, A, b, c, d, e, &c. is the feries; B, B2, B3, B4, &c. the firft order of differences; C, C2, C3, &c. the fecond order of differences; D, D², &c. the third order; E, &c. the fourth order, and fo on. But the firſt terms of theſe ſeveral orders of differences, as B, C, D, E, &c. are thoſe that are principally made uſe of in calculations by this method. PROP. Sect. II. THE DIFFERENTIAL METHOD. II PROP. I. If there be any feries A, b, c, d, e, &c. and if there be taken, the first differences B, B², B³, &c. the fecond differences C, C2, C3, &c. the third dif ferences D, D2, D3, &c. and fo on. Then if T ftand for the first term of the nth dif ferences; I fay, ±T=A-nb+nx X 3 d + nx NI N I N2 -2 N3 2 +T, when n is even, X 3 N-I 2 c-nx 2 X e-&c. that is, 4 and -T when n is odd. The feveral Orders of differences being taken as before directed, will ftand thus. Then A B b C B D Cz E B³ 3 D2 d C3 B+ &c. e C d " e, &c. feries A Ift diff. 2d diff. 3d diff. 4th diff. b b-A c-b d-c C e-d, &c. " 2 c—2b+A, d—2c+b, e—2d+c, &c. d—3c+3b—A, e-3d+3c-b, &c. e-4d+6c46+A, &c. That is, Bb-A, C—c—2b+A, D=d−3c+36 —A, E—e—4d+6c-4b+A, &c. or-B = A—b, +C=A=2b+c,-DA-3b+3c-d, +E=A —4b+6c—4d+e, &c. where putting T fuccef- fively equal to B, C, D, E, &c. and n = 1, 2, 3, 4, &c. the Prop. will be evident. Cor. Hence A≈A, the firſt term. B=-A+b, the firft difference. C= A-2b+c, the 2d difference. DA+363c+d, the 3d difference. E= A-4b+6c4d+e, the 4th difference. FA+56-10c+10d-5e+f, the 5th difference. &c. PROP. 12 THE DIFFERENTIAL PROP. II. If A, b, c, d, e, &c. be any feries; and there be taken B, C, D, E, &c. the first of the feveral orders of differences. Then I fay the nth term of the feries, will be A+ C+ n 2 N3 X 2- I N-- I N-2 n- I B + X X I I 2. I 2 3 N- I D+ X X X E+ &c. I 2 3. 4 N2 N3 n4. For from the equations in the laft Prop. viz. B = b—A, C = c−2b+A, &c. we have by tranf- pofing, c b=A+B=A+26+C=A+2A+2B+C (expunging b); that is, c=A+2B+C, d=A-3b+3c+D=A-3A-3B +3A+6B+3C+D (expunging b and c); that is, d = A +3B+3C+D. Alſo e = —A+4b—6c+4¢ +E (expunging b, c, d) -A+4A+4B-6A -12B−6C+4A+12B+12C+4D+E; that is, e=A+4B+6C+4D+E, &c. Then puting A, b, c, d, &c. for the nth term; and n fucceffively 1,2, 3, 4, &c. the feries will be evident. Cor. 1. If d', d", d'", &c. be the first of the first, Second, third order, &c. of differences; then The nth term of the feries, A, b, c, d, &c. will N-I N2 N→→ I n- n- 2 n- I Be=A+ ďt X ・d" + X I I 2 I 2 n X d" + 2-3 d'""" .n- I N2 N3 24 n 31111 X X X •4d" + I 2 3 4 3 &c. For Bd', C=d", D=d", &c. And the coef- ficients are the unciæ of the n—1th power. Cor. 2. Hence alfo it follows, that any term of a given feries may be accurately determined; if the dif- ferences of any order happen at last to be equal. Cor. Sect. II. 13 METHOD. Cor. 3. Hence A≈A, the first term. b=A+B, the 2d term. <=A+2B+C. the 3d term. d=A+3B+3C+D, the 4th term. e=A+4B+6C+4D+E, the 5th term. ƒ=A+5B+i0C+10D+5E+F, the 6th term. g=A+6B+15C+20D+15E+6F+G, the 7th term. &c. PROP. III. If a, b, c, d, e, &c. be any feries; and d', d", d"", &c. the first of the feveral orders of differences; Then The Sum of n terms of the feries, is =na+nx NI NI N-2 ·ď + n x X d" + nx 2 2 X n = 3 d!!! +nx 3 NI N2 N-3 1-4 NI N2 2 X 3 X X X d" 4 2 3 4 5 + &c. For in the ſeries of quantities, o, a, a+b, a+b+c,a+b+c+d, &c. ift diff. are a b C d &c. ? 2d diff. d'. d'a > d3 &c. 3d diff. 4th diff. d"2 &c. &c. , th Therefore (by Cor. 1. Prop. II.) the n+1 term of the feries, o, a, c+b, a+b+c, a+b+c+d, &c. or the term of the feries, a, a+b, a+b+c n th NI 2 g+b+c+d, &c. is = o+na+nx d' + n x NI N 2 X d" + &c. But the nth term of the 3 feries a, a+b, a+b+c, &c. is the fum of n terms of the feries, a, b, c, d, &c. and therefore N I equal to na+nx d'+nx 2 N 2 7Z2 X d" + &c. 3 PROP. 14 THE DIFFERENTIAL PROP. IV. If there be feveral fets of quantities in arithmetic progreffion, whofe number is m, and the common dif- ferences p, q, r, s, &c. and if a new feries be form- ed by multiplying the refpective terms of all thefe fe- ries into one another; then I fay the mth order of dif- ferences will be equal to one another, and to pqrs, &c. Xinto 1×2×3×4, &c. to m. 1. Suppose two feries; as a, a+p, a+2p, a+3p, a+4p, &c. and b, b+q, b+2q, b+3q, b+49, &c. then by multiplying we have ab, a+pxb+q,a+2p xb+29, &c. that is, ab feries ab+pb+aq+pq ab+2pb+2aq+4Pq ab+3pb+3aq+9Pq &c. 1 Diff. 2 Diff. pb+aq+pq pb+aq+3pq 2pq pb+aq+5pq 2pq therefore the 2d differences are equal, and equal to PqX1X2, where m-2. 2. Suppose three feries. as a, a+p, a+2p, a+3p, a+4p, &c. b, b+q, b+2q, b+39, b+49, &c. c, c+r, c+2r, c+3r, c+4r, &c. The terms of theſe being multiplied, produce a fe- ries whofe terms are, abc a+pXb+qXc+r a+2pxb+2q Xc+2r a +3 p× b + 3q × c + 3r a+4pxb+49 × c+4r &c. thefe Sect. II. 15 METHOD. theſe being multiplied at length, and fubtracted from one another, will give the following feries, which is the firſt order of differences, abr+bcp+aqc+ pqc+ aqr+ pbr+pqr abr+bcp+aqc+3pqc+3aqr+3pbr+7pqr abr+bcp+aqc+5pqc+5agr+5pbr+19pqr abr+bcp+aqc+7pqc+7agr+7pbr+37 p q r &c. And the differences of thefe, or the fe cond differences being taken, we have 2pqc+2aqr+2pbr+6pqr 2pqc+2aqr+2pbr+12pqr 2pqc+2agr+2pbr+18pqr 6pqr 6pgr and the third difference will be here m 3, and the mth difference is pqrx1X2X3. 3. Suppofe four feries. pqr×1×2×3. As a, a+p, a+2p, a+3p, a+4p, &c. b, b+q, b+aq, b+39, b+49, &c. c, c+r, c+2r, c+3r, c+4r, &c. d, d+s, d+25, d+3s, d+45, &c. and when multiplied, produce abcd +a+2p × b+29 × c+2r x d + 2 s +a+3p× b+39 × c+3r ×d+35 +a+4p × b+49 × c+4r ×d+4s +a+5p × b+5q × c+5r × d+55. &c. Let ABCD be any term; and the next, A+pXB+qxC+rx D+s; that is, AB+Aq+pB+pq×CD+Cs+rD+rs=ABCD+ ABX Cs+rD+rs+CDx Aq+pB+pq + Aq+pB +pq xCs+rD+rs. From this, fubtract ABCD, and the 16 DIFFERENTIAL THE the remainder is the firft differenceABxCs+rD+rs + CD× Aq+pB+pq +Aq+pB+pq×Cs+rD+rs. To abridge the work, keep in only fuch terms as in- volve the moſt of the quantities ABCD, for the reſt will vaniſh before you come to the laſt differences. By this means the firft difference becomes only ABX Cs+rD+CD× Aq+Bp; that is, the firſt difference ABCs+ABDr+CDAq+CDBp; then the (1) next term is evidently A+pxB+qx C + rxs + A+pxB+qxD+s×r+A+p×C+r xD+s xq +B+qxC+r×D+s×p=AB+Aq+pB×C+r ×s+ AB+Aq+pB ×D+sxr+AC+Ar+pC × D+s×q+BC+Br+qCxD+sxp= (2) ABCs +ABrs+ Aq+pB ×C+r × s + ABDr+ABsr+ Aq+pBxD+s x r+ACDq+ACsq+ Ar+pCx D+s× q+BCDp+BCsp+ Br+qCxD+s × p; from this fubtracting the equation (1), and the fe- cond difference is ABrs + Aq+pBxC+r xs+ ABsr+ Aq+pBxD+sx r+ACsqx Ar+pC x D+s xq+BCsp + Br+qCxD+s xp=ABrs+ ACqs+BCps, +ABrs+ĀDqr+BDpr, +ACsq+ ADrq+CDpq, +BCsp+BDrp+CDpq, &c. ´by abridging (3) 2ABrs+2ACqs+2BCps+2ADqr +2BDpr+2CDpq, this is the 2d difference; and the next term is 2rs× A+p×B+q +2qsx A+pxC+r +2ps × B+qxC+r+2qr× A+p×D+s+2prx B+q×D+s+2pqx C+r xD+s =(4)2rsxAB+ 2rs× Aq+pB +29sAC +29sx Ar+pC +2psBC+ 2psx Br+qC+2qrAD+2qrx As+pD +2prBD+ 2 prx Bs+qD +2pqCD+2pqx Cs+rD, &c. by a bridging; from this laft equation taking equation (3), the remainder is the 3d difference 2rsx Aq+pB +2qsx Ar+pC +2ps×Br+qC +2qrx As+pD + 2 prx Bs +qD +2pqX Cs+rD = (5) 6qrs A+6prsB +6pqs C Sect. II. 17 METH O D. +6pqsC+6pqrD. Then the next term is (6) 6qrsx A+p+6prs× B+q+6pqs×C+r+6pqr× D+s; from this equation, take equation (5), and there remains 6pqrs+6pqrs + 6pqrs + 6pqrs=24pqrs the fourth and laſt difference, fo the fourth difference is = pqrs×1×2×3×4; and fo on for more progreſ- fions. -712 m n2 -m m Cor. 1. In the ſeries a +a+p +a+2p + a+3P, &c. the last and equal differences will be p × into 1×2×3×4, &c. to m terms. For then p, q, r, s, &c. are all equal. Cor. 2. In any adfected equation of the mth power, putting fucceffively any numbers of a feries in arithme- tical progreffion, for the unknown quantity; the mth differences will be equal. الله For the lower powers all vaniſh before you come to the laſt difference; and therefore that difference depends entirely on the higheſt power in the equa- tion, which is the mth power. PROP. V. Let A, b, c, d, e, &c. be any feries; B, C, D, E, &c. the firſt of the ſeveral orders of differences; and let their poſition be fuppofed to be fuch, that they may be at a units diftance from one another; and let x be the diſtance of any termy (to be inter- polated;) then I fay, X-I 2 y=A+xB+xx: C+xx -X 2 X-- I X--2 3 X D+x x-3 E + &c. X--- I 2 X X X 2 3 4 or y = A+B+PC+ X 2 3 QD + RE 3 4 + SF+&c. where P, Q, R, &c. are the co- X-4 5 efficients of B, C,D, &c. C For 18 THE DIFFERENTIAL For (Prop. II.) the nth term of the feries is NI N2 = A +” —¹ B + X C, &c. and the th I I n 2 *+1" term = A + −B+nx +I n- I C, &c. which 2 is the term at the nth diſtance from A; then put x=n, and the term at x diſtance from A, will be X-I 2 P=A+B+xx ·C + ≈ X X-I X2 2 X 3 D, &c. And as this holds exactly, when P is any of the terms A, b, c, d, e, &c. Therefore it will be nearly true, when P is any intermediate term. Cor. Hence if feveral equidistant ordinates in any curve be given; the quantity of any other ordinate, whofe pofition is given, may be found by this Prop. that is, fuch an ordinate may be interpolated. PRO P. VI. Prob. If there be a ſeries of equidistant terms; a, b, c, d, e, &c. whose first differences are small; to find any term wanting in that feries, having any number of terms given. RULE. Take the equation which ftands againſt the number of given terms, in this table; and by re- ducing the equation, that term will be found. nla -b=0. 2a-2b+c=0. 3a3b+3c―do. 4a4b+6c-4d+e=0. 54-5b+10c-10d+5e-f=0. 6a-66 +15c-20d+15e-6f+g=0. 7a-76+21c-35d+35e—21f+78—b=0. 8 686 +28c56d+70e−56f+28g—8b+i=@. N-I N- I na-nb-nx 2 2 X d+&c.=0. 1-2 3 For Sect. II. 19 METHO D. For (by Prop. I.) the quantities ſtanding againſt Fig. the numbers 1, 2, 3, 4, &c. are the firft of the firft, fecond, third, fourth, &c. differences. Now as the first differences are ſmall, the other differences will decreaſe, and the laſt of them may be taken for nothing. And the more terms there are given, the nearer the equation approximates; likewiſe the quantity fought will be determined more ex- actly, as it is nearer the middle of the equation ; the error being almoft reciprocally as the coefficient. For Ex. Suppoſe b, c, d, e, f, g, are given to find a; againſt 6 you have a—6b+15c~20d+15e -6f+go, which reduced gives a = 6b-15c+20d -15e+6ƒ—g. If the third c had been fought, and 20d—15e+6f—g—a+6b the reft given; then c= PROP. VII. 15 If a, b, c, d, e, &c. be any feries; and fuppofe 8. the feveral terms placed at unequal diſtances from one another. To find any intermediate term, whofe place or poſition among the rest, is given. Suppoſe the ſeries a, b, c, d, e, &c. placed as ordinates in a curve; and let MN, MP, MQ, MR, MS, &c. be the correfpondent bafes, for which put n, P, Q, R, S, &c. refpectively; alfo put NP = p, PQ=q, QR = r, RSs, &c. and let x repreſent any bafe, and y its correfpondent term in the feries. Then affume y=A+Bxxn + C xx~n × x-P+Dxxnxx-P xx-Q + Exx¬n × x-P x *~Qxx-R, &c. continued to as many terms as there are ordinates, or terms of the ſeries. Confequently x will afcend to a power, one lefs than the number of terms. C 2 Then .20 DIFFERENTIAL THE Fig, 8. Then put a, b, c, d, e, &c, fucceffively for y. and n, P, Q, R, S, &c. for x. And theſe will produce as many equations as there are unknown quantities, A, B, C, D, E, &c. which therefore will be determined as follows. a = A b=A+BxPn c = A + BxQ¬n +CxQ-nxQ-P d = A+BxR―n+CxR-nxR-P+ D x R—Ź × R—P × R—Q e=A+BxS—n +C× S—n× S—P +Dx S—n XS-PXS-Q+ExS n x S-P ( x S-QXS-R, &c. Then inftead of P-n, Q-n, Q-P, &c. put their values p, p+q, q, &c. and you'll have the following equations. @ = A a b = A + B p c = A+Bxp+q + C x p + q x q 4 d = A+Bxp+q+r +Cxp+q+r xq+r+Dx p + q + r xq + r x r e = A + Bxp+q+r+s+ C × p+q+r+s X q+r+s+Dxp+q+r+sxq+r+s xr+s+Exp+q+r+sxq+r+s xr+sxs. Then take each equation from the following one, and you'll have, b-a = Bp ε → b = Bq+Cxp + q xq i d-c = Br+Cxp+2q+rxr+Dxp+q+r x q + r x r ! Sect. II. 21 METHOD. e-d Bs+Cxp+29+2r+s xs+Dxp+q+rx Fig. q+2r+sxs+q+r+sxr+sxs+E xp+q+r+sxq+r+sxr+s xs. And dividing by the refpecting coefficients of B, 8. and then b-a B. P c-b =B+Cxp+q=B1. 9 dc =B+Cxp+2q+r+Dxp+q+r x q + r =B2. 1 e-d ===B+Cxp+2q+2r+s + D × p+q+r S Xq+2r+s+q+r+sxr+s+Ex p+q+r+sxq+r+sxr+s=B3, &c. And again fubftracting each equation from the next, B1-B =Cxp+q. B2-B1=Cxq+r +Dxp+q+rxq+r. B3—B2=Cxr+s+ Dx p+2q+2r+s xr+s &c. +Exp+q+r+sxq+r+sxr+s. And dividing by the coefficients of C, BI-B p + q B2-BI q + r B3-B2 r+s &c. = C. ·=C+Dxp+q+r=Ci. = C + D xp+2q+2r+s+ Ex p+q+r+sxq+r+s=C2. C 3 And 22 THE DIFFERENTIAL Fig. And taking the differences of thefe equations, 8. CI-CDxp+q+r. C2—C1 = Dxq+r+s+Exp+q+r+sxq+r+s. &c. And dividing by the coefficient of D, CI-C p + q + r C2-CI g+r+s =D, =D+Exp+q+r+s=D1, &c. And taking the differences as before, Di-D=Exp + q + r+s, ánd DI-D p + q +r+s = E, &c. And thus you may proceed as far as you will; whence to find the intermediate term, whoſe baſe is x, we have this RULE. Let a, b, c, d, e, &c. be the feries, M any fixed point, and for MN, NP, PQ, QR, RS, &c. put N 2 B , r , S., &c. and let Ba BI-B C > -b BI = p + q B2-BI CI-C q C1: > d-c qtr. Di B2 = r જ Cz B3-B2, q+r+s D= p+g+r E= q DI-D C2-CI p+q+r+s - &c. e-d r+s &c. B3 &c. &c. Then the term anfwering to the bafe x, is y=A+Bxx n +Сxx nxxn+p+Dxxnx x−n+pxx-n+p+q+Ex x− n x x − n + p x Xn+p+qxx-n+p+q+r, + &c. where Śect. II. METH O D.. 23 where M may be taken at pleaſure, and may be Fig. either without or among the ordinates. And here 8. x is fuppofed to lie towards the right hand; but if it be taken to the left from M, it must be made negative. And if any of the terms of the feries a, b, c, &c. be negative, the ordinate reprefenting it, muft lie below the bafe MS. Cor. Suppofing B, C, D, E, &c. found as be- 9. fore, and let M fall upon N at the first ordinate, then will no, and let p, q, r, s, &c. be the feve- ral distances of the ordinates from each other. Then putting x for any baſe reckoned from M; then will the correfponding term to be interpolated be, y=A+Bx+Cxxx—p+Dxxx−pxx~p—q+Exx x-pxx p~qxx p q r + &c. continued to as many terms, as there are terms in the given feries. SCHOLIU M. They that would fee other methods of interpo- lating the terms in a given feries, may confult the Philofophical Tranfactions, No. 362; or Sterling's Summiation of feries, engliſhed by Mr. Holliday. ¡ PRO P. VIII. To find a curve line of the parabolic kind, which shall pass through any given number of points. RULE. Let the points be A, B, C, D, E, &c. and 19. from them to any right line MT given in pofition, let fall as many perpendiculars, AM, BP, CQ, DR, ES, &c. Then to find other points through which the curve muſt paſs; affume any point X, and put x for the bafe MX, and y for the corref ponding ordinate XY. Then if the diſtances of the ordinates AM, BP, CQ, &c. be all equal; find C the 24 THE DIFFERENTIAL Fig. the value of y by Prop. IV. But if their diſtances 10. be unequal, find the value of y by Prop. VII. JI. which gives the point Y, through which the curve is to pafs; and thus you may find as many points as you will; through all which draw the curve re- quired. Cor. Hence, the interpolation of the terms of a feries comes to the fame thing, as finding the ordi- nates in a parabolic curve of fome degree or other. As fuppofe the equation to the parabolic curve be y=A+Bx+Cxx+Dx³ +Ex¹, &c. where the pow- ers of x must afcend to as many dimenfions as there are ordinates wanting 1. Therefore if fuck a parabola be defcribed as above directed, whofe feveral ordinates repreſent the particular terms of a given feries; then to find any other term of that feries, is no more than finding the correfpondent ordinate of the curve. PROP. IX. To find the area of any given curve. RULE. Thro' any number of points in it, defcribe a parabolic curve (by Prop. VIII.); and find its area, which will be nearly equal to the area of the curve propoſed. Thus let ADF be the curve, and fuppofe any number of equidiftant ordinates to be erected, as AM, BP, CQ, DR, &c. and thro' A, B, C, D, &c. fuppofe a parabolic curve to pafs, and let any baſe MX = x, and its ordinate XYy; and the equation of the parabolic curve is y = A+Bx+ Cx²+Dx³ &c. continued to as many terms as there are ordinates. Then to find the area of this parabola; fuppofe x divided into an infinite num ber of equal parts, according to the Arithmetic of Infinites. Then (by Prop. I. Sect. I.) the fum of all Sect. II. 25 METHOD. Bxx all the Bx is = all the A's is = Ax. And (Prop. II.) the fum of Fig. And (Prop. III.) the ſum of 11. 2 CX3 all the Cxx = And (Prop. IV.) the fum of 3 Dx4 all the Dx' and ſo on. Therefore (by 2 4 Prop. VII.) the fum of all the y's, or of all the A+Bx+Cx²+Dx3, &c. that is, the area of the curve is Ax+ Bx² Cx3 Dx4 2 3 + + 3 , &c. 4 A+÷Bx+÷Cx²+‡Dx³ &c. × *, or put L for * the length of the bafe; then the area = A+÷Bx+÷Cx²++Dx³, &c. × L. Now let a, b, c, d, e, &c. be the ordinates of the given curve placed at equal diftances; and let MP, PQ, QR, &c. each = 1. Then from the nature of the parabolic curve, (y=A+Bx+Cx², &c.) putting x = 0, 1, 2, 3, 4, &c. and you'll have the following equations. a=A b=A+ B+ C+ D+ E, &c. c=A+2B+ 4C+ 3D+ 16E, &c. d=A+3B+ 9C+27D+ 81E, &c. e=A+4B+16C+64D+256E, &c. &c. Theſe are continued to as many terms as there are equations or ordinates. Then if the quantities A, B, C, D, &c. be expunged, by help of theſe equations, out of the area (A+÷Bx+÷Cx², &c. ×L); you'll have the area expreffed by the ordi- nates, a, b, c, d, &c. I. For one ordinate. Here a A, xo, and AL area. Therefore AL area. 2. For 26 THE DIFFERENTIAL 2. For two ordinates. Here x = 1,a=A, b =A+B, therefore B = bα, ba and the area = A+÷BxxL = a+ XL; that 2 a+b is, the area = L. 2 3. For three ordinates. Here 2, and a = A b=A+ B+ C c=A+2B+4C ba = B+ C c-b=B+3C c—2b+a = 2C 1/2 c − b + 2a = C. B — b—a—-C — — 232a+26—2—2c. = Then the area A+Bx2++CX4XL becomes a ~ 3a+2b — 2 c 2 + } a − b + c { x L area. 4 a+c+46 or { a + 3 b + { cx L = L the area. 6 4. For four ordinates. Here 3, area = A+B+3C+22D. and a - A z-a dc 4 b=A+ B+ C+ D C=A+2B+4C+ 8D d=A+3B+9C+27D B+ C+ D B+3C+ 7D B+5C+19D Sect. II. 27 METHO D. c-2b+a = 2C+6D d-2c+ b = 2C+12D d—3c+3b—a = 6D c—2b+a―d+3c—3b+a = 2C or 2a-56+4c— d= 2C and a-2b+2c2d = C I - a + b c + ÷d = D 2 b—a—C—D I I C—D =B = —÷÷÷a+3b—— 2c+3d. whence a that is 4. 2/2 - = a + 2 b — 2 c + d) +3a2b+6c3dXL area. 27 9 33 b 3 a + 2/2 2 6 — 23/3/27 c + & d a+d+3b+3c, -L L= area. 8. + 5. For five ordinates. Here x4, and the area=A+2B+C+16D D= 256 E. S and a A 3 E b = A+ B+ C+ D+ c=A+2B+ 4C+ 8D+ 16E ₫ = A +3B+ 9C+27D+ 81E e =A+4B+16C+64D+256E d then ba B+ C+ D+ E c-2b+a= d3c+3ba = e-4d+6c-4b+a = 2C+ 6D+ 14E 6D+ 36E 24E whence E=ed+ /c1b + 2 = a. d—3c+36—a—36E — — že + Zd—2c+36 —zza. 6 • I C— — 26+ a — 3D —7E — 1e−3d+ 12c−1b + 15a. C= 2 4. 3 B = b—a—C—DE ——÷E + ‡d — 3c +46-232152α. Theſe 28 THE DIFFERENTIAL Theſe values of A, B, C, &c. fubftituted in the value of the area will give the area =e+d+ 736 + 156 + 3a: X L; that is, the area 2 7 7×a+e+32xb+d+12c 90 6. If fix ordinates be given. 3 Here x = 5, and the area=A+B+C+ 125D+125E+3125F; and proceeding as before, 4 6 we ſhall find the area = 50Xc+d 19×a+f+75xb+e+ 288 L. 288 For a A b=A+ B+ C+ D+ E+ F c=A+2B+ 4C+ 8D+ 16E+ 32F d=A+3B+ 9C+ 27D+ 81E+ 243F e=A+4B+16C+ 64D+256E+1024F ƒ=A+5B+25C+125D+625E+3125F. ولا 4, 19 Then for 288 19 75 5_b₂ 50 50 75 e and 288 288c, 288 d 288 288f; put their reſpective values, from the equa- tions above, and we ſhall have 19 A 288 75 A+ 75B+ 75 C+ +288 288 288 5 A+ 10 B+ +288 288 50 A+B+ +288 288 75 A+ + 288 300B + 288 75 D+ 75 E+ 75 F. 288 288 288 200 800 1600 C+400 D+ E+ F 288 288 288 288 450 C+ C+ 135° D+ 288 4050 E+ 12150F. 288 288 288 C+ 4800+ 19200E+ 76800F. 288 288 288 1 475 C+ 288 1200 288 +12A+ 25-B+ 288 288 288 2375D + 11875E+ 59375F. 288 288 The Sect. II. 29 METHOD. 288 288 The fum of all theſe is the area, viz. A+ 720 B+ 2400C + 9000 D + 36000E+150000F 288 288 288 288 288 or A + 2B + 83C + 314D + 125E + 520§F, the area, as it ought to be. And thus you proceed to find the area, when more ordinates are given. Cor. 1. If we put A for the fum of the firſt and laft ordinates, B the fum of the fecond and last but one, С the fum of the third and last but two, &c. and where the number of the ordinates is odd, let the laft capital denote the middle ordinate. Alfo put L for the length of the baſe, or the distance of the firſt and laft ordinates; you'll have the following table of areas against the respective number of ordinates. nates Ordi-Areas IAL 2 2 AxL 3A+4B A+ 6 -XL A+3B XL 8 57A+32B+12C XL 90 619A+75B+50C XL 288 41A+216B+27C+272D 840 -XL 8751A+3577B+1323C+2989D×I 17280 989A+5888B-928C+10496D-4540E, 28350 XL Cor. 30 THE DIFFERENTIAL : Cor. 2. After the fame manner, if a, b, c, d, &c. reprefent any equidistant fections of a ſolid; A the sum of the firſt and laſt, В the fum of the fecond and last but one, &c. the foregoing table will give the folidity, against the proper number of fections, very nearly. Cor. 3. To reduce the curve to another with fewer ordinates; take the fum of the first and last for a new ordinate, the fum of the fecond and laſt but one, for another new ordinate next it, &c. and lastly, double the middle ordinate if odd, for the last of them. You will have a new curve with fewer ordinates; whofe area is eaſier found, and is nearly equal to the area of the former. Moreover, when the ordinates of this new curve are bad; if there be taken the sum of the first and fecond ordinates, the fum of the third and fourth, the fum of the fifth and fixth, and fo on. Or if you take the fum of the three first, and the fum of the three next, and ſo on; or if there be taken the ſum of each four, or the fum of each five ordinates, &c. for the ordinates of another curve, it's area will be equal ta the area of the given one. And in each curve the baſe must be diminiſhed in proportion to the number of or- dinates that were added together: and thus a proper number of ordinates, in any curve being given; its quadrature will be reduced to the quadrature of ano- ther curve, with fewer ordinates. Cor. 4. Or when nine ordinates give not the area exact enough, erect more; and divide the area into feveral parts, and find all the parts feparately. And it will be most expeditiously done, if you take no more than three or four ordinates at once. PROP. Sect. II. 31 METHOD. PROP. X. To refolve Problems by the differential Method. RULE. This is done by making uſe of the foregoing Pro- pofitions, according as the nature of the queftion requires. When the firſt of any order of differences is wanted, it is to be found by Prop. I. When any term of a feries is required, it will be had by Prop. II. If you want the fum of any number of terms, it is found by Prop. III. To interpolate any term of a ſeries, then Prop. V, VI, VII, VIII, and their Corollaries, muſt be made ufe of, as occafion requires. And to find the areas of curves; proceed by Prop. IX. When the feveral orders of differences happen to be very great, it will be convenient to take the Logarithms of the quantities concerned, whoſe differences will be fmaller; and at laft when the Logarithm is obtained, the number or quantity belonging to it, and which was required, will eafily be had. In working by the differences of quantities, the labour may in many caſes be abridged, by putting a number of cyphers, for fo many terms at the beginning of the feries; by which means you'll get feveral of the differences equal to o, and fo the anfwer will be obtained in fewer terms. Ex. 32 THE DIFFERENTIAL. Ex. 1. To find the first of the 8th order of differences in the geometrical progreſſion 1, 3, 9, 27, 81, &c. Here (by Prop. I.) n = 8, A, b, c, &c. = 1, 3, 9, &c. whence cô T = 1—86+28c-56d, &c. — 1—8×3+28×9 -56×27+70×81-56×243 +28 × 729-8x 2187 +6561=32896-32640 = 256; that is, 256 is the firſt of the eight order of differences. Ex. 2. What is the first of the 5th order of differences, of this feries, 1, 1, 4, 4, 16, &c. Put A, b, c, then (Pr. I.) I I 89 I &c. n = 5. f d, &c. = 1, 4, 42 82 T=1- 5b+10c10d + 5e — ƒ =1—2 1/2 + 2 1/2 — 14 +1% −32=313—3 3 5 = 36 i that is, T-38. T= I 6 5 Ex. 3. What is the first of the fourth order of differences of the ſeries of cubes, 1, 8, 27, 64, 125, &c. Let A, b, c, d, &c. = 1, 8, 27, 64, &c. 4. Then (Prop. I.) T = A—4b+6c—4d+e =1−32+162—256+125 = 288—288 = o, ſo that To, the fourth order of differences. Ex. 4. What is the 20th term of the ſeries 1, 8, 27, 64, 125, &c. Here put 1, 7, 12, 6, for A, B, C, D; n = 20; then (Prop. A I B C II.) the 20th term will be A 8 7 D 12 19.18 +19B + ·C+ 19.18.17 D 19 6 2 2.3 27 18 Q. =1+133 + 2052 +5814 = 8000, the 20th term of the fe- ries. 64 37 24 6 61 125 Ex. Sect. II. 33 METH O D. Ex. 5. What is the 20th term of the triangular numbers, 1, 3, 6, 10, 15, 21, &c. Here A, B, C, D, are e- qual to 1, 2, 1, 0; and n = 20; therefore (by Prop. II.) the 20th term = A+19B +19×9C1 + 38 + 175 =210. A I 3 6 B C D 2 I 3 4 ΤΟ 5 15 To find the fum progreffion 1, 2, 3, Ex. 6. of n terms of the arithmetical 4, 5, &c. I, Put a, d', d"= 1, 1, 0, then (by Prop. III.) the fum ofn terms of the A d' I d" I NI feries is n+n. = n + \nnin = 2 3 nn + n n+I I 2. 2 2 Ex. 7. To find the fum of n terms of the fquare numbers, I, 4, 9, 16, 25, 36, &c. — Here a, d', d", d"" 1, 3, 2, 0; whence (by Prop. III.) the fum of n terms is a d' n+3R I d" N-I n x n - IXn-2 3 4 2 X + 2 3 5 Q 9 2 2n+ 3nn 3n nn3n+2 + n 16 7 2 2 3 923 2nn-6n+4 ·? + 25 9 n 6 6 D 2nn 34 DIFFERENTIAL THE 2nn + 3n+1 6 n terms. n+I 2n+I n = n x X the fum of 2 2 3 Ex. 8. To find the fum of n terms of the cube numbers, I, 8, 27, 64, 125, &c. Here fubftituting the values of a, d', d', d", by Prop. III. we have the fum of n terms = a ď I d" 7 d"" 8 12 N I n+nx ×7+12N 19 6 2 27 18 NI n 2 X X +6nx 64 37 6 24 2 3 61 N I N24 N3 125 X X 2 3 4 7nn-7n N46N³+IIN²-бn =nt +2n³ — 6n²+4n+- 2 4 n¹ + 2n³ + n² the fum of n terms. 2 4 Ex. 9. To find the fum of n terms of the biquadrate num bers, 1, 16, 81, 256, &c. Here fubftituting a ď for a, d', d", d"", O d" d'"' &c. their values o, I I 1, 14, 36, &c. and 15 the ſum of n terms 16 of the feries 1, 16, 8 1 81, &c. being the 175 14 36 240 50 60 65 110 84 d' 24 fame as the fum of 256 194 n terms of the 625 369 feries 0, 1, 16, &c. therefore (by Prop. III.) fubftituting n+1 for n, the Sect. II. 35 METH O D. the fum will be N4 — 2223 — N² + 2n I 24 12 4 = = = n² + 24 25+ ·ns → 5 nn + n 2 ×36+ I + N3N 6 X14+ n5—5n4+5n³+5n²—6n I 120 X n + + 1 = n³n, the fum of n terms 3 30 required, of the feries 1, 16, 81, &c. Ex. 10. To find the fum of n terms of the feries 1, 6, 20 50, 105, 196, 336, 540, 825, &c. This is the fame thing as finding the fum of n+3 terms of the ſeries o, ọ, 0, 1, 6, 20, 50, &c. Therefore writing o for a, d', d"; and 1, 2, for d", d'"", and n+3 for n; we have (by Prop, III.) the fum n + 3x n + 2x n + 1 x n 2.3.4. n+3xn+2xn+1xnxn−Į a ď d" 2" dille I 2 3 I I 4 2 5 6 5 9 14 20 + 2.3.4.5 2n5+15n4+40m³ +45m² +18n 120 X 2 = the fum ofn terms of the feries 1, 6, 20, 50, &c. Ex. II. Given the logarithms of 101, 102, 104, and 105; to find the log. of 103. Here are 4 quantities given; therefore by Prop. VI. againſt 4 you have a-4b+6c—4d+e=0; where c is fought, and all the reft are given; there- 4xb+d-a+e C fore we have c = 6 D 2 36 DIFFERENTIAL THE a = 2.0043214 b= 2.0086002 d = 2.0170333 e 2.0211893 b+d= 4.0256335 4 4xb+d=16.1025340 ate= 4.0255107 6) 12.0770233 2.0128372 the log. 103. Ex. 12. There are given the cube roots of 45, 46, 47, 48, and 49; to find the cube root of 50. Here five quantities are given; whence (by Prop. VI.); againſt 5 you have a-5b+10c10d +5e-fo, where f is required; therefore f= a−5b+10c—10d+5e. N. Log. 453.556893 = a. a = 3.556893 463.583048 = b. 10c IOC 36.088260 473.608826 c. 5e 18.296530 48/3.634241 = d. 57.941683 49/3.659306 = e. 17.915240 f 36.342410 54.257650 ƒ= 3.684033cube root of 59. Ex Sect. II. 37 METHOD. Ex. 13. Given the natural tangents of 88°, 54′, 55′, 56', $7', 58', 59', and 89°; to find that of 88°, 58', 18". Tangents A=52.080673 801436 b=52.882109 82647 B C D 25042 8 E 1193 F C c = 53.708587 26235 76 G d = 54.561300 852713 1269 27504 83 7 I 880217 e = 55.441517 28856 1352 8 91 f=56.350590 909073 1443 30299 g 8 = 57.289962 939372 Here taking 88° 54′ from 88° 58′ 18″, the re- mainder is 4' 18" 4.3; therefore put x = 4.3, and (by Prop. V.) y = A+4.3xB + 4.3X3·3C + 2 4.3X3.3X2.3 6 4.3X3.3X2.3X1.3 D.+ + 24 4.3X3.3X2.3X1.3X.3 F+ 120 4.3X3.3X2.3X1.3X.3X.-.7 G. 720 and collecting the terms, 52.080673 3.446175 177673 6489 134 Tan. 88 58′ 18″ 55.711144 Ex. 14. Given the log. fines of 1° 0′, 1° 1′, 1° 2′, 1° 3′; to find the fine of 1° 1′ 40″. Sines B A = 8.2418553 C b 8.2490332 71779 D 1168 C c = 8.2560943 70611 38. d = 8.2630424 69481 1130 D 3 Then 38 THE DIFFERENTIAL Then fubtracting 1° o' from 1° 1' 40", there re- 2 mains 1′ 40″ ====*; then (Prop. V.)the fine to 3 x; 5 be interpolated is y = A+B+C-D; that is, 3 9 + 8.2418553 +119631 6.49 2 + 8.2538184 651 S. 1° 1' 40″ = 8.2537533 Ex. 15. I 509 SI I 539 54 Given the feries I I 30, 31, 32, 33, 34; to find 53. that which stands in the middle between I 52 and To fave the trouble of reduction, take the lo- garithms of the numbers. Log. A B. I 50° 1.6989700 C 86002 D I 57 1.7075702 I 52 1.7160033 84331 1671 E 66 I 82726 1605 8. 58 53 1.7242759 I 54° 1.7323938 81179 1547 Then put x = 2 the diſtance of the number to 2 I 50 be interpolated, from ; then by Prop. V. y = A+B+C+DE. I the log. of 52.5 1.6989700 215005 + 3133 2 I I 1.7204727 In numbers +3133 1.7201594 being the number fought Ex. Sect. II. 39 METH O D. Ex. 16. Fig. Given three distances of the fun from the tropic, 8. and the times of obfervation. To find the time of the Solstice. viz. Dec. 20. 21 39 6 24 · 74. Comparing this with Prop. VII. we have a b=6, c=74; and p=1, q=3; then B 39, ba Ρ ~~~--~33, BI BI- cb 68 BI-B = = 223; C = 9 3 p+q 553/3/ I = 4 I I 1311; then (Cor. Pr. VII.) y = 39— 2 33*+1311*X*—1, or y = 39-4611x+1311xx o, becauſe y is nothing at the time of the fol- ftice. Therefore xx-3.3713≈ ——2.8024, and *—1.6856 ±√.1383.3724, and x 1.3132, or 2.0580. x= Ex. 17. Given a comet's distance from the fun on the follow- ing days at 12 at night, to find its diftance Dec. 20. Decem. 12 21. 24 26.. dift. 301 = a 620=b · 715 = c 772 = d Here 8, the time after Dec. 12; and by Pr. VII. p = 9, 9 = 3, r = 2; and B – 319 q = 9 35.4444; BI = = 31.6666, B2=57-28.5; 95 3 =-3148, CI =- 2 3.7778 C = 3.1666 12 5 .3185 −6333 D = 14 = 0227; then D 4 y = 40 THE DIFFERENTIAL y = 301+35.4444X8-3148×8x-1-.0227X-I X-4 = 301+283.5555+2.5184-.0908; that is, y=586.9831 the comet's diſtance Dec. 20, at 12 at night. Ex. 18. Given five places of a comet; to find it Nov. 16d.—17h. d. h. Nov. 3 17, place Leo 29 51 5 10 15 16 Virgo 3 23 15 32 18 21 Libra 18 52 20 17 28 10 Proceeding by Prop. VII. we ſhall have Nov. 3 17 5 15 30 16 18 21 20 17 Leo 29 51 a Virgo 3 23=b 15 32 = 0 = I 22 221.9166 p. 5 8 1 I I Ι I = 5.0417 = 9. 58.2083 = r. 20 1.8333 = s. = 3.5333b-a. 3 32 12 9 12.15 Cab. Libra 18 52d 33 20 33.3333 d—c. — 9 18 = 9.3 =ed. 28 10=e B1.8436,C .0814,D .0028,E-.0003 B12.4099,C1.1246,DI=.0016 B24.0609,C2.1008 B35.0729 from Nov. 16 17 take Nov. 3 17 remains 13 o therefore 13, whence Cor. 1. Pr. VII.) ya+1.8436×13+.0814X13X 11.0834+.0028X13X11.0834×6.0417—.0003×13 X Sect. II. 4T METHO D. X11.0834×6.0417×—2.1666 = a+38°.6982 = Leo 29° 51'+38° 41′ 53″ Libra 8° 32′ 53″, its place November 16d. 17h. Ex. 19. 2 6 In the given ſeries 1,7×1,1×3×1,—×÷×1 XI, 8 6 5 --x--x+x2x1, &c. to find the term in the X X 7 5 3 I middle between the ſecond and third. A B 1 C D I E &c. 2 3 I उ 23 - 3332/55 उड 3 / 3/3 23 35 8 1 6 35 then y 2 IS IOS 5 By Prop. V. Alfo x = 1½/. 3 P = 3, Q 89 R: 3 M2 I 289 I T=- 7 S= 3 2569 V = 1024, &C. TO24 A+B+C―D+,?E—₂}F+1024G 3 28 or y=1+11/ 3/18 - 80 I 3 256 I &c. 7 896 768 II264 &C. that is y = 2.357. Ex. 20. Fig. Given the lengths of three equidistant ordinates 10. (5, 7, and 8); and the length of the baſe MQ = 10; to find the area of the curve AMQC. By Cor 1. Prop. IX. AAM+CQ =5+8 = 13. BBP7, then for three ordinates, the A+4B 13+28 area= 6 L= 41 6 X 10 = = x 10 = 683. Ex. 21. 6 Given the length of four equidistant ordinates (5, II. 7, 8, 10), and baſe MR = 10, to find the area of the curve AMRD. Here AAM+DR = 15, B=BP+CQ = 15. Then for four ordinates, the area (Cor. 1. A+ 3B IX.) is = 8 75, as required. L= 15+45 60 8 X10=8X10= Other- 420 THE DIFFERENTIAL Fig. II. 12. Otherwife, Put AM = 5+10=15, and BP =7+8 = 15; then put A = AM + BP = 30; and the area A for thefe two ordinates is 2 Ex. 22. 30 XL = ×5=75. 2 Given the lengths of five equidiftant ordinates, (10, 11, 14, 16, 16), and the length of the baſe MS, 20; to find the area. Here AAM+ES = 26, BBP+DR = 27, C = 14, L = 20. Then (by Cor. 1. Pr. IX.) 7A+32B+12C, the area is 90 -L= 182+864+168 X 90 1214 20 ×20=2693. 90 Or thus, Reduce it to 3 ordinates, putting AM 10+16 26, BP 11+16=27, CQ = 2×14= 28. Then AAM+CQ= 54, B27; then the 162 area= A+4B 6 X 10 = IO 54+108 6 X 10 = 6 x 10 X = 270. Or thus, In the laſt curve, the three ordinates are 26, 27, 28; make AM 26+27= 53, BP = 27+28 =55, L= 5; and put A = AM+BP = 108; A = 108 then the area = L= X 5 = 270. 2 2 Ex. 23. Given the lengths of fix equidiftant ordinates, (10, 12, 11, 7, 0, 9,); bafe 20; to find the area. Here A 10-9 1, B 12+0 = 12, C = = = 11+7=18, L = 20; then (Cor. 1. Pr. IX.), area Sect. II. 43 METHOD. area = 19+900+900 XL = 288 X 19A+75B+50C 288 1819 20 = ×20 = 1262; where note, as the or- 288 dinate VF is negative; the area SVF below the bafe, is alfo negative. Or thus, Divide the area into two parts, one with four ordinates, the other with three. For the four or- dinates 10, 12, 11, 7; we have A = 17, B = 23; 17+69 and the area = A+3B 8 86 X 12 = 8 X 12 = 8 X12 129. For the three ordinates 7, 0, -9; we have A=2, Bo; whence the area = =-=-=x8=-2. A+3B 8 L= Then the whole area 129-2 127. Ex. 24. Let 18 ordinates be given in a curve, 0, 2, 3, 4, 6, 5, 3, I, I, -3, -4, -5, -5, -3, -I, 0, 2, 5; and the bafe 17; to find the area. The ſhorteſt way is to take three or four ordi- nates at once. Thus for the four ordinates o, 2, 3, 4; we have A = 4, B=2+3 = 5; and the A+3B 4+15 area = 8 L 57 8 x3= 8 = 73 Then for the area with the four ordinates 4, 6, 5. 3; here A4+3=7, B=6+5=11; and the area = 7+33 8 X3 = 15. Again, with the four ordinates 3, 1, −1, −3; here A = 0, B = o, and the area ➡o. Then 44 THE DIFFERENTIAL METHOD. Then with the four ordinates -3, -4, -5, -5; here A—8, B = −9; and the area = -8-27 X3= 8 35 8 ×3=-13/3/ Then with the four ordinates -5, -3, -1, o; here A5, B=-4; and the area - —5—12 17 8 X3=- 8 x3=-63. = Laſtly with the three ordinates o, 2,5; here A5, B2; and the area = 24 5+8 13 6 X2= = 43. 3 8 623, 249 Then the fum of all theſe areas is 7+15—13% ~~~ 6 3 3 + 4 3 = 6232 ; that is the whole area eſteeming theſe parts negative, that are below the baſe. Ex. 25. Given 8 ordinates of a hyperbola between the af fymptotes, 10, 11, 12 IN 1 10, 10, 17; to find I ΙΟ T39 I O I O ΙΟ the area. Here a 1.00000 b Τ I I I O C I 2 0.90gog 0.83333 d I O Τ I O e O TS I O I I O IT 0.76923 0.71428 0.66666 0.625 =0.58823 then A =a+b= 1.58823, B = b+g= 1.53409. C=c+f=1.5, D=d+e=1.48351; then (Cor. 1. Pr. IX.) the area = 7 51 A+ 3577 B+1323 C29+89DL, where L = 7; 17280 that is, the area = 1192.76+5487.44+1984.5+4434.21 17280 5.3062, as required. × 7= THE Fig.1. 2. B A P C D B P 6. 5. D E F F C B P B 8. 9 M N P Q 3. B E 9. B 7. T d 4. la p. p 9 71 S R S M Р X R S 10. B C E DYE F A B TIN TE M P R S M P RX 11. V 12. Hill Ar. Infinites, and Dif Method. 4 the End. Cher श्र OF THE ELEMENTS OF THE CONIC SECTIONS; DEMONSTRATED In THREE BOOK S. Book I. Of the ELLIPSIS. Book II. Of the HYPERBOLA. Book III. Of the PARABOLA. Et fi omnia à veteribus inventa effent: hoc tamen femper novum erit, ufus et inventorum ab aliis fcientia et difpofitio. SENEC. THE PREFACE. IT T is not my defign in this Treatife, to lay down all the properties of the Conic Sections, but ſhall confine myſelf to thofe that are most useful. For to fay all that has been writ- ten, or may be written concerning thefe curves, would far exceed my intended brevity, and be attended with no advan- tage. Therefore ſetting afide all trifling and uſeleſs propofi- tions, I have endeavoured to explain the chief and most confi- derable properties of theſe curves, and I think I have omit- ed but few things that are either useful or curious. in 1 The most useful and remarkable curves, next the circle, are the Conic Sections. Theſe curves make a confiderable branch geometry. They are of great use in feveral parts of the mathematics, particularly in Dialling, for delineating the figns; likewife in the Projection of the fphere, where feve- ral of the circles are projected into Conic Sections; and the like in Perſpective; alfo in Optics, to reflect or refract the rays to a geometrical focus; all Aftronomy is built thereon ; the Conftruction of algebraic equations depends much upon them; they are very useful in natural Philofophy for invef- tigating the phoenomena of nature. A man cannot be a pro- ficient in the abftrufer parts of the mathematics, without the knowledge of the Conic Sections, which leads the way to the higher geometry. And it is abfolutely neceſſary to know the fundamental properties, for all fuch perfons, as would make any advances in natural philofophy, and for understanding the amazing diſcoveries of the last age, in the mechaniſm of the univerfe. As to this Treatife, I have comprehended all I had to fay on this fubject, in three Books, for the three Conic Sections Separately: which is better, at least for beginners, than treating them altogether, in a more abstract and general man- ner. The feveral propofitions are delivered in fuch order, as they ii THE PREFACE they seemed most naturally to follow from one another; and, for the feveral Sections, are laid down nearly in the fame or- der, and almoft in the fame words, that the relation they have to one another, may the better appear. In the Hyperbola, many of the propofitions comprehend a great many particular cafes, which would require a vast num- ber of cuts to defcribe them all, as they relate to the oppofite Sections, or to the conjugate hyperbola's. Therefore I have treated them in as general a way as I could. And any parti- cular cafe may be easily inferred from the general propofition, mutatis mutandis, attending duly to the nature of that pro- pofition. And to affift the reader's imagination, in the moſt re- markable propofitions, I have denoted fuch particular caſes by fmall letters of the Alphabet. And in many inftances I have done the fame thing with the other Sections, where it appear- ed neceffary. I have chofen to demonſtrate theſe their properties from their fimple defcription upon a plane, without making use of the cone, as being easier to understand. For when the demon- ftrations are taken from the Sections of a cone, they are more difficult to comprehend, by reafon there are so many interfec- tions of planes with planes, and planes with folids, that the reader is confused with them. Whereas thefe properties are eafily drawn from their mechanical deſcription upon a plane, and the demonftrations are more fimple and natural. At the end of each Book I have given a collection of the moſt uſeful Problems. And last of all I have shewn what fort of curves thefe are, which are made by cutting a cone by a plane in fuch and fuch pofitions; and have demonftrated that these are the very fame curves, we have here been treat- ing on. So I hope the reader will be furniſhed here with thing that is neceſſary for him, in this branch of geometry. every W. Emerfon. BOOK [ 3 ] CONIC SECTIONS. BOOK I. Of the ELLIPSIS. I DEFINITION S. DEFIN. I. F two pins be fixed at the points F, S; and a thread PSFP, put about them and knotted at P; then if the thread be drawn tight, and the point P and the thread be moved about the fixed centers F, S; the point P will defcribe the curve PDpBEAP, called an Ellipfis. 4 DE F. II. The points or centers F, S, are called the Foc. DE F. III. The line AB, drawn thro' the foci to the curve, is called the tranfverfe Axis. DE F. IV. The point C in the middle of the axis AB, is the Center. The line DE, pendicular to the jugate Axis. DE F. V. (drawn thro' the center C) per- tranfverfe AB, is called the con- Fig. 1. E DEF. 4 THE ELLIPS I S. Fig. DE F. VI. Any line TO, drawn thro' the center C to the 2. curve, is called a Diameter. And the extrémity T (or O) its vertex. DE F. VII. If TO be a diameter, then the diameter GK, drawn parallel to the tangent at its vertex T, is called its conjugate. And the two diameters TO, GK, are faid to be conjugates to one another. DE F. VIII. The line LR (drawn thro' the focus F, perpen- dicular to the tranfverfe axis AB,) is called the pa- rameter or latus rectum. DE F. IX. A line drawn from any point of the curve (as HI) perpendicular to the tranfverfe axis, is called an Ordinate to the tranfverfe. And in general, any line drawn from the curve to any diameter TO, pa- rallel to its conjugate GK, (as HN,) is an ordinate to that principal diameter TO. If it go quite thro' the figure, as Hb, it is called a double Ordinate. DE F. X. A right line meeting the ellipfis in one point M, but does not cut it, is called a Tangent to it in that point, as TM. DE F. XI. The part of the diameter between the vertex and the ordinate, is called the Axfciffa: TN, AI. And the Vertex is the extremity of any diameter. PROP. B. I. 5 THE ELLIPSI S. Fig. I. PROP. The fum of the lines FP, SP, drawn from the foci, I. to any point of the curve, is equal to the tranfverfe axis AB. For by conftruction, PF+PS = AF+AS = AF+AF+FS ≈ 2AF+FS. And the fame PF+ PS 2BS+FS; therefore 2AF+FS 2BS+ FS; and 2AF2BS, or AF BS. Whence PF+PS 2AF+FS AF+BS+FSAB. Cor. The two foci are equally distant from the ver- texes, and alfo from the center : AF BS, and FC SC. = For it is proved that AF BS; and fince AC = CB (def. 4), therefore AC—AF — CB—BS, or FC SC. PROP. II. A line drawn from the end of the conjugate axis, to the focus, is equal to half the tranfverfe; DF CA. Draw DS to the other focus. Then the two right angled triangles CDF and CDS are fimilar and equal. For SC CF, the angles at C are right, and CD common; therefore SD DF; and fince the fum SD+DF the tranfverfe (Prop. I.), one of them DF half the tranfverfe CA. Cor. The distance of the foci is a mean proportional between the fum and difference of the tranfverfe and conjugate axes, SF² = BA+DE × BA—DE. 2 For CA DF² = DC²+CF2; and CF2 = CA²—CD² — CA+CD × CA—CD; and 4CF² or SF² = 2ÇA+2CD × 2CA-2CD. E 2 PROP. 3. 6 THE ELLIPSI S. Fig. 3. 3. PROP. III. The rectangle of the focal distances from either ver- tex, is equal to the fquare of the femiconjugate: AFX FB DC². For DC² = DF²-CF² = (Prop. II.) CA²- CF² = (Geom. I. 12.) CA+CF x CA-CF = BC+CF × CA-CF = BFXFA. PROP. IV. As the tranfverfe axis to the conjugate, fo the con- jugate to the latus rectum of the tranfverfe: AB: DE :: DE: LR. = 2 For SL+LF BA=2CA (Prop. I.); and SL = 2CA-LF, and by fquaring (Geom. I. 11.), SL² = 4CA²-4CAXLF+LF". 4CA×LF+LF. And in the right angled triangle SLF, SL - SF²+LF²; whence 4AC2-4ČA×LF+LF2 = SF2+LF2, and 4AC²-4CA×LF — SF² = 4CF², and 4AC² =4CA×LF+4CF² = 4CA×LF+4DF2—4DC². and 4AC²+4DC 4CAXLF+4DF2; but CA? = DF² (Pr. II.); therefore 4DC² = 4CA×LF = 2CAX2LF; that is, DE BAXLR. Cor. 1. As the femitranfverfe is to the femi-conju gate, fo the femiconjugate to half the latus rectum ; CA: DC:: DC: LF. Cor. 2. As the femitranfverfe, to the distance of the focus from the center; fo is the fame distance, to the difference between the femitranfverfe and half the latus rectum: FC² CAXCA—LF. 2 2 For CF DF-DC (Pr. II.) CA- CD² = CA?-ÇA×LF (Pr. IV. Cor. 1.) 2 Cor: B. I. 7 THE ELLIPSIS. Cor. 3. The rectangle BFA = half the tranfverfe Fig. X half the latus rectum CAXFL. By Cor. 1. 3. and Prop. III. = SCHOLIU M. Since the tranfverfe axis is to the conjugate, fo the conjugate to the latus rectum, of the tranfverſe axis. Therefore in any other diameters, the third proportional, to the diameter and its conjugate, is called the latus rectum of that diameter. PRO P. V. From any point M in the curve, drawing the lines 4. MF, MS, to the two foci; and the ordinate MP per- pendicular to the tranfverfe axis BA; it will be, As the femitranfverfe, CA: to the distance of the focus from the center, CF :: So the distance of the ordinate from the center, CP: to half the difference of the lines MF, MS, or = 2 MS-MF 2 2 For make SD = CA, then SM CA+DM, and FM = 2CA-SM — CA-DM. In the right angled triangle SMP, SM² or CA²+2CA×DM+ DM' (Geom. I. 10.) = SP²+PM² = CF+CP +PM² = CF²+2CFxCP+CP²+PM², and in the right angled triangle FMP, FM or CA-- 2CA×DM+DM² = FP²+PM² = CF-CP +PM² (Geom. I. 11.) CF-2CFXCP+CP²+ PM²; then fubtracting the latter equation from the former, SM-FM24CAXDM = 4CFX CP, and CFXCP CAXDM. But fince SM CA +DM, and FM CA-DM; therefore SM- FM 2DM; therefore CF x CP CA × = SM-FM 2 = = X 2 E 3 Cor. 8 THE ELLIPSI S. Fig. Cor. 1. If F, S be the foci, MP an ordinate; 4. then it is CA: CF:: CP: CA-MF or SM-CA. For CFXCP = CA×DM, and DM = SM— CA CA-FM. Cor. 2. If F, S be the foci, MP an ordinate; then the difference of the Squares of the lines SM, FM; that is, SM²-FM² = 4CF×CP. 2 Cor. 3. If F, S be the foci, MP an ordinate; then CAXSM-FM For SM2 FM² 2CFxCP. SM+FM x SM-FM (Geom. I. 12.) — 2CA × SM—FM 2CA × SM-FM = 4CFX CP, and CA X SM-FM = 2CF×CP. SCHOLIU M. If PM falls on the other fide of F, as pm, then Cp-CF, and its fquare the fame as before, and the reft of the demonftration the fame. pF PROP. VI. If an ordinate MP be drawn to the tranfverfe axis; it will be, As the fquare of the tranfverfe, BA² : to the fquare of the conjugate, NE² : So the rectangle of the fegments of the tranf. BPA: to the Square of the ordinate, PM2. For make SD = CA, then DM is half the dif- ference of SM and MF; therefore by Prop. V. CA : CF : : CP : DM, and (Propor. 13.) CA: CA+CF or BF:: CP: CP+DM, and (Pro.4.) CA: CP:: BF: CP+DM, and (Propor. 13.) CA: CA+CP or BP: BF: BF+CP+DM. But BF BC+CF SD+CF; and BF+CP +DM = SD+CF+CP+DM = SM+CS+CP =SM+SP; whence CA: BP :: BF: SM+SP. Again, B. I. 9 THE ELLIPSIS. Again, fince CA: CF:: CP: DM; then (Pro. 13.) CA; (CA-CF) AF:: CP: CP-DM; and (Propor. 4.) CA : CP :: AF : CP—DM, and (P. 13.)CA:(CA-CP) PA:: AF: AF-CP+DM. But AFCA-CF — SD-SC; therefore AF -CP+DM SD-SC-CP+DM SM-SP; therefore CA: PA:: AF: SM-SP, and we had before, CA: BP :: BF:SM+SP; then multiply- ing theſe proportions together (Propor. 18), we have CA: BPXPA:: BFXFA: SM-SP (Geom. 12.) But (Prop. III.) BF×FA = CN², and Geom. II. 21. Cor. 1.) SM²-SP² PM'; therefore CA² : BPA :: CN² : PM², or alternately CA²: CN²::BPA: PM² 2 2 2 or BA (4CA): NE² (4GN2): BPA: PM*. Cor. 1. CA: CN2: BPA: PM². Cor. 2. As the tranfverfe BA : to its latus rectum : : So the rectangle BPA : to Square of the ordinate PM*. Fig. 4. NE2 For (Prop. IV.)latus rectum = whence fince AB' BA: EN2 :: BFA : PM2. therefore BA : NE 2 BA or latus rectum:: BFA: PM 2 (Propor. 6.). Cor. 3. The rectangles of the fegments of the tranf verfe, are as the fquares of the ordinates. 2 For every rectangle is to the fquare of its ordi- nate, in the given ratio of CA to CN', or of BA to the latus rectum. Cor. 4. As the Square of the femitranfverfe CA²: Rectangle of the focal distances from vertex BFA:: So rectangle of the fegments BPA: Square of the ordinate PM². E 4 PROP. ΙΟ THE ELLIPSIS. Fig. 5. 6. T PROP. VII. If the ordinate MG be drawn perpendicular to the conjugate DE, it will be, = As the fquare of the femiconjugate, DC: to fquare of the femitranfverfe, CA² : : So the rectangle of the fegments of conj. DGE : to fquare of the ordinate, GM2. For by Cor. 1. Pr. VI. CA²: CD² :: BPA : PM². But BPA AC+CP X AC-CP (Geom. I. 12.) AC²-CP2 AC-GM2; therefore CA: CD´ :: AC²—CP² : PM, and alternately CA² : CA-CP²:: CD: PM'; and by divifion CA²: €A²~CP² PM² CP2: CD2: CD2-PM or CD CG; and al- ternately CA²: CD':: CP: CD2-CG or CD+CG × CD-CG that is, EGXGD; or CA²:CD': : GM: DGE. Cor. The whole ordinate to the tranfverfe or con- jugate axis, is biffected by that a is; and equal or- dinates are equally diftant from the center. For the rectangle of the fegments of the axis being in a given ratio to the ſquare of the ordinate ; and that rectangle being the fame, for the ordinates on both fides; thefe ordinates muſt be equal. Like- wife that rectangle remains the fame, at equal dif- tances on each fide the center; and therefore the ordinates are equal at equal diſtances. PROP. VIII. If MP be an ordinate, to the tranfverfe; F, S the foci; and you make CA: CF:: CP: CI; then the rectangle SMX MF CA-CI2. For make SD = CA, then DM = SM-FM 2 and by Prop. V. DM = CFXCP CA = CI by con- ftruction; 'B. I. II THE ELLIPSI S. ftruction; and SM×MFCA+DM × CA—DM Fig. = CA-DM2 CA-CI2. = Cor. 1. If CA: CF :: CP: CI; then FM = AI, and SM = BI. For DM CI, and SM = ÈM CA+DMCA +CI BI, and therefore FM = AI. Cor. 2. CA: CF:: CP: CA-FM or SM-CA. PROP. IX. 6. If F, S be the foci, C the center, MP an ordinate 6. to the tranfverfe, CE the femiconjugate; then CM² = CA²+CE’—SMF. 2 = 2 For in the triangle SMF, (Geom II. 28), SM +MF² = 2CF2+2CM². But SM+MF = SM² +2SM x MF+MF2; therefore SM+MF² = SM+MF² 2SMX MF BA-2SMXMF, whence BA2-2SM× MF2CF²+2CM². There fore 2CM² BA22CF-2SMX MF = 4CA² -2CF2-2SMX MF; and CM 2CA-CF2 2 = -SMXMF CA²+CE-SMxMF. Cor. 1. If CA: CF:: CP: CI; then CM = EL. It follows from this and the laſt Prop. Cor. 2. If CA: CF:: CP: CI; then rectangle BPA rectangle SIF+PM*. For BPA = 2 CA-CP2 - CA2-CM²+PM* CA²—CM²+PM² SMF-CE+PM² = (VIII) CA²-CI²-CE +PM² = CF³—CI²+PM² = SIF+PM², PROP. X. If from the foci, two lines SM, FM be drawn to any point in the curve, as M; they will make equal angles with the curve in that point. Take m infinitely near M in the curve, and draw Sm, Fm; and from S and F defcribe the finall arches 12 THE ELLIPSIS. Fig. arches Mr, mt, thro' M and m. Then fince SM 7. +MF = Sm+mF, by Pr. I. that is, SM+Mt+ tF = Sr+rm+mF; and taking away the equal* quantities SM, Sr, as alfo tF, mF; there remains Mt = rm. Therefore in the infinitely fmall tri- angles Mrm, and Mtm, right angled at r and t, we have Mt mr, and the hypothenufe Mm common; therefore the fides Mr and mt are equal, and the oppofite angle Mmr and mMt; that is, MmS and MF are equal; but being infinitely near M, the angle MS GMS the external angle at M; whence mMF = GMS. Cor. If a line MO be drawn perpendicular to the tangent MT in M; the angles FMO and SMO are equal. For they are the complements of the equal angles FMT and SMG. PROP. XI. 8. If MP be an ordinate to the tranfverfe; and MO perpendicular to the tangent at M; and S, F the foci; then CA : CF::CP:CÓ. 2 For (Geom. II. 25.) in the triangle SMF, where MO biffects the angle SMF, it is SM: MF:: SO: OF; and by compoſition (Propor. 13.) SM+MF: SM-MF SO+OF: SO-OF; that is, 2CA: SM-MF:: 2CF : 2CO :: CF : CO, and SM-MF CA² : V.) 2 SM-MF 2 X CA :: CF : CO. But (Prop. × CA=CF× CP; therefore CA² : CFXCP:: CF: CO:: CF2: CFXCO. And al- ternately CA²: CF² :: CF×CP: CFXCO :: CP : CO. Cor. 1. CA: CD² : : CP : PO. För B. I. 13 THE ELLIPSIS. For CA 2 CF :: CP: CO; and by divifion, Fig. CA² : CA²-CF2 or CD²:: CP : CP-CO or PO. 8. CP:CP-CO or Cor. 2. 2CA: latus rectum :: CP: PO. For CA: CD:: CP: PO, and 2CA: or the latus rectum : CP: PO. Cor. 3. CA: MF:: CF: OF. 2CD2 CA For we had 2CA: SM-MF:: CF: CO, and by divifion 2CA: 2CA-SM+MF:: CF: CF- CO; that is, 2CA: SM+MF-SM+MF:: CF: OF; that is, 2CA: 2MF, or CA: MF:: CF : FO. PROP. XII. If S, F be the foci, and MO perpendicular to the 8: tangent in M; then MO² SMX MF-SOXOF. For fince MO biffects the angle SMF; there- fore (Geom. II. 26.) MO +SOF SMF; and MO² SMF-SOF. 2 Cor. 1. MO= CD CA SMXMF; or CA: CD²:: SMF : MO². For (Geom. II. 25.) SM : MF:: SO: OF, and by compofition, SM+MF or 2CA: MF:: SO+ OF or SF: OF. or CA CF :: MF: OF, in like : manner CA CF :: SM: SO, and multiplying, CA: CF:: SMF: SOF; 2 and CA: CA-CF2 or CD2: :SMF:SMF-SOF. that is, CA²: CD² : : SMF : MO², by this Prop. Cor. 2. SMF = MO²+BOA—CD². For SMF MO²+SOF. But SOF SC²- CO² = BC2-CD2-CO² = BQA-CD². PROP. 14 THE ELLIPSIS. Fig. 9. 10. PRÓ P. XIII. If thro' any point M, the line SG, equal to BÁ, be drawn from the focus S, and FG from the other focus F; and the line MD be drawn to the middle of FG; then MD will touch the curve in M. For draw FM, then MG AB-SM FM; therefore MD is perpendicular to FG, whence FM - GM. To any other point m, in the line DM; draw Sm, Gm, Fm; then in the triangles GmD, FmD the angles at D being right, and the including fides equal, the hyp. mGmF. In the triangle SGm, the fum of the two fides Sm+mG is greater than the baſe SM+MG, (Geom. II. 5.); that is, Sm+mF is greater than SM+MF; and therefore by conſtruction of the ellipfis, the point m is with- out it; and fo are all points except the point M. Cor. 1. The tangent at M, biſſects the external angle FMG, made by two lines drawn from the focus. Cor. 2. The tangents at A and B are perpendicular to the tranfverfe axis AB; likewife a tangent at the end of the conjugate is perp. to the conjugate. PROP. XIV. If PM be an ordinate, AD a line drawn from the ends of the tranfverfe and conjugate axes, and PQ parallel to AD; then PM² CD²—CQ® DQXQE. For by the fimilar triangles ACD, PCQ; AC: CD:: PC: CQ; and AC: CD:: PC: CQ2; and CD² (Propor. 13. laft cafe) AC: CD:: AC-PC2: CD-CQ. But (Prop. VI.), AC: CD:: BPA CD² or B. I. 15 THE ELLIPSI S. or AC²-PC¹: PM²; therefore PM² = CD¹-CQ Fig. or CD+CQ × CD—CQ. PRO P. XV. If MP be an ordinate to the tranfverfe axis, and II. MH be drawn perpendicular to the curve at M, cutting the tranfverfe and conjugate axes at O and H. It will be as OP: OC :: CD²: CF¹. F being the focus. Z : For (by Prop. XI.) CA² : CF²::CP: CO. And by divifion, CA-CF CF :: CP-CO: CO; that is, CD: CF:: OP: CO. CF² Cor. 1. CD: CF:: MO: OH: : MP: CH:: PO: CO. This follows from the fimilar triangles MOP and HOC. Cor. 2. If L be the latus rectum of AB; then, L:AB-L::MO:OH::OP:OC :: MP : CH. For CD² : CF² or CA²-CD² : : MO: OH, &c. CD2 CA (by Cor. 1.); therefore : CA CD2 ::MO: CA OH (Propor. 5. Cor. 1.); that is, (by Pr. IV. Cor. 1.) L: CA-L:: MO: OH; or L: BA-L:: MO: OH, &c. Cor. 3. CA: CD::CH+PM: PM:: CP: OP. For CD: CF2 :: MP: CH; and compounding, CD² : CD²+CF² (CA²):: MP: MP+CH:: CP: OP. PROP. XVI. If PM be an ordinate to the tranfverfe, MT a tangent at M; then CP, CA, CT are continually proportional; C being the center. For draw MO perp. to MT, then (Prop. XI. Cor. 1.) CA: CD2 :: CP: PO:: CP²: CPX PO, and 12. 16 ELLIPSIS. THE Fig. and drawing PQ parallel to AD, then CA² : CD² : : 12. CP2: CQ2, by fimilar triangles; whence CP2: CPX PO : : CP² : CQ; therefore CP×PO = CQ²; by :: Ax. 7. Proportion. But in the right angled trian- gle OMT, where MP is perp. to OT, OPX PT PM² (Geom. II. 20. Cor. 2.) = CD²—CQ² (Prop. XIV.) = CD-CPXPO, by what is juft proved. Therefore OPXPT+OP×CP = CD²; that is, OPXCT = CD². 13. But OPXCT: CPXCT:: OP: CP:: (Pr. XI. Cor. 1.) CD CA; that is, CD2: CPXCT:: CD2: CA2. Therefore (Ax. 7. Proport.) CPxCT = CA²; or CP: CA : CT÷÷÷÷. Cor. 1. If C be the center, PM an ordinate, MT a tangent, MO perp. to MT; then OP, CD, CT, are continually proportional, or OP×CT = CD². Cor. 2. If C be the center PM an ordinate; then BP: PA :: BT: AT; and therefore BT is har- monically divided in the points P, A. For fince CT: CA: CA: CP; therefore (Propor. XIII.) CT+CA: CT-CA:: CA+CP: CA–CP; that is, BT: AT:: BP: AP. PRO P. XVII. If MG be an ordinate to the conjugate, MV a tangent at M, C the center; then CG, CD, CV, are continually proportional. By fimilar triangles CV: VG:: CT: GM or CP (laft Prop.) CA: CP or GM² (Prop. VII.) :: CD2: DGE or CDCG². And by divifion CV: CV-VG or CG:: CD: CG²; whence (Propor. 12.) CVXCG² = CGX CD or CV×CG = CD2; therefore CG: CD: CV are Cor. 2. M Fig. 1. D P H L A B B B E 4. M 3. D B S F I S F B m R 5. D E G M m 7. E M 8. B 6. E Το B S co PF F E A A CIP F 11. D M 9. m 10. D M M B PE B H Conic Sections E E PL.I pa 1.6. UNIV M OF B. I. THE ELLIPSIS. 17 13. Cor. 1. If C be the center, MG an ordinate to Fig. the conjugate, MV a tangent, and MH perp. to the tangent MV; then HG, CA, CV, are continual proportionals. For CGXCV = CD2; but CGXCV: HGXCV:: CG or PM: HG (by fimilar triangles)::OP: GM or CP (Prop. XI. Cor. 1.) :: CD²: CA²; that is, CD²: HGXCV:: CD²: CA; therefore (Ax. 7. Propor.) HGXCV 2 CA¹. Cor. 2. If C be the center, MG an ordinate to the conjugate, MH perp. to the tangent MV; then EG: GD:: EV: DV; and EV is harmonically di- vided in the points G, D. For fince CV: CD:: CD: CG, by compofition and divifion, CV+CD:CV-CD::CD+CG:CD—CG, that is, EV: DV : : EG : DG. PROP. XVIII. If a femicircle AFD be defcribed on the tranfverfe 14. or conjugate axis AD of an ellipfis AKD, and the ordinate PM be drawn cutting the circle in I; the tangents MT, IT, drawn from the points M, I, will interfect the axis in the fame point T. For (by Prop. XVI, and XVII.) the tangent MT will interfect the axis in T, fo that CP, CA, CT, be in geometrical progreffion. And (Geom. IV. 26.) If a tangent IT be drawn to touch the circle at I, it will interfect the diameter in T, fo that CP, CA, CT be in geometrical progreffion, the fame as in the ellipfis. PROR 81 THE ELLIPSI S. Fig. 1.5. 16. PROP. XIX. If on the tranfverfe axis AB as a diameter a cir- cle AKB be defcribed; and the ordinate PM be drawn cutting the circle in N; then PN will always be to PM, in the given ratio of the femitranfverfe CA, to the femiconjugate CD. For (Geom. IV. 17.) PN² = BPA; and (Prop. VI.) CA²: CD²:: BPA or PN2: PM²; and there- PN² fore CA: CD:: PN: PM. 2 ! Cor. If a circle be defcribed on the conjugate axis DE as a diameter, and an ordinate QR be drawn cutting the circle in I; then QI will always be to QR, in the given ratio of the femiconjugate CD to the femi- tranfverfe CB. = There- For (Geom. IV. 17.) QI DQE. And (Prop. VIL) CD²: CB²:: DQE or QI² : QR². fore DC: CB:: QI: QR. PROP. XX. : If the line FD be drawn from the focus, perpendi cular to any tangent MD; then the interfection D will be in the circumference of the circle ADB, de- fcribed on the diameter BA the tranfverfe axis. For produce SM from the other focus, and FD till they interfect in G. Draw CD, FM. Then in the right angled triangles FMD, GMD. The angle FMD (by Prop. XIII. Cor.) GMD, and MD being common; therefore FD DG, and FM MG; and fince FC = CS; therefore CD is parallel to SG (Geom. II. 12. Cor. 1.). There- fore the triangles FCD and FSG are fimilar; there- SM+MG = = fore, finçe FD=4FG, CD=4SG= 2 B. I. 19 THE ELLIPS I S. SM+MF 2 = ¿BA = CA; therefore D is in the circumference ADB. And by the fame reaſoning, if S is the other fo- cus, and SH be perp. to MH; the point H is in the circumference of the ſame circle ADB. • Cor. 1. If a circle be defcribed on the tranfverfe AB, and a tangent at M, cuts it in D, H; and the lines FD, SH, be drawn from the foci, to the points D, H; then FD, SH are perp. to the tangent HMD. Cor. 2. If F, S be the foci, C the center, MD @ tangent at M; then if CD be drawn parallel to SM, to cut the tangent in D; then CD = CA half the tranfverfe. Both thefe corollaries follow from the demon- ftration of this Prop. Cor. 3. If a tangent HD interfects the circle AHB in D and H; the perpendiculars to it, DF, HS, will pass thro' the foci F and S. Cor. 4. If CI be drawn parallel to the tangent at M; then IM CA. For then IM CD, and CD = CA, by Cor. 2. = PROP. XXI. If S, F be the foci, HT a tangent at M; if SH, FD be drawn perp. to the tangent MT; then SHXFD BSXSA. Deſcribe a circle ARH about the tranfverſe, and produce HS to cut the circle at R; and draw CR. Then fince BH: DG×DE. But IL' ANXBH, therefore DF = DGXDE, whence DE: DF:: DF: DG. Cor. If from the ends of any diameter AB, two lines AH, BN be drawn, thro' any point P, of the curve, to interfect the tangents AN, BH in N and H. The rectangle of the tangents is equal to the Square of the conjugate: ANXВH = IL². From the demonſtration. PROP. LXIII. If C be the center, F the focus, and if CF: CA:: CA: CT, and the directrix TO be drawn perp. to CA. And if any line MNO be drawn; and from the focus F, MFH, FN and FO be drawn. Then FO biflects the angle NFH. Draw MP, NQ perp. to TO, and ND parallel to MF. Then by fimilar triangles, MP: NQ :: MO: NO:: MF: ND; and alternately MP: MF:: NQ: ND. But (Prop. XXIX.) CA: CF :: MP: MF:: NQ: NF; therefore ND NF; therefore < NFD = NDF DFH, by reaſon of the pa- rallels ND, FH. PROP. B. I. 51 THE ELLIPSIS. Fig. PRO P. LXIV. If AB, ED be the tranfverfe and conjugate axes, 68. and if the diameters GCK, and HCL be drawn pa- rallel to AD, AE; thefe diameters will be equal, and conjugate to one another. For fince EA = AD DB = BE, ADBE is a parallelogram, and BE parallel to DA which is parallel to KG; therefore the angle EBA GCA. And fince BC CA, therefore EF FA; and confequently GK is the diameter of EF, FA, and HL (parallel to EA) its conjugate. Likewife the PROP. XLIII. Let AB be any diameter, CD the femiconjugate, AM, DN, conjugate hyperbola's, PM an ordinate, QV parallel to CD. Then CA: CD :: CA' + CQ² : QV². 2 2 Draw the ordinate BV parallel to CA. Then (Prop. XXXVI.) CA: CD2 :: BV: CB2 CA² ĈD² CD2; and compounding, CA: CD²:: CA² + BV² : CB². That is, CA²: CD² ::CQ² + CA²: QV¹. Cor. 1. If AM, DN be conjugate hyperbolas; CA, CD conjugate femidiameters, PM an ordinate, QV parallel to it. Then CP-CA² : PM² : : CQ³ + CA²: QV². For B. II. 113 THE HYPERBOLA. For CP2 CQ² : QV². CA: PM²:: CA: CD:: CA + Fig. Cor. 2. If qv be parallel to QV; then CQ² + CA²: QV²:: Cq² + CA²: qv². 2 Cor. 3. If the ordinate PM be produced to the conjugate hyperbola at N. Then CP-CA2: PM²:: CP² CP² + CA²: PN2. For CP2-CA2: PM²:: CA: CD2::CA + CP² : PN². CD² CA² Cor. 4. PN² + PM: PN — PM':: CP: CA². By compounding the terms in Cor. 3. PROP. XLIV. 35. If AB be any diameter, CD its femiconjugate; 34. HT a tangent at M, cutting the parallel tangents AK, BH, in K and H; MP an ordinate. Then AK, CD, BH, are in continual proportion. : By fimilar triangles AK: PM: TA: TP :: (XLII. Cor. 3.) TC: TB:: (fim. triangles) CG : BH. Whence PM × CG AKX BH. 2. X Then CP: CT :: CP²: CP × CT or CA (XLII.). And by diviſion, CP — CT or PT CT: CP² - CA or BPA: CA²: : (XXXVI.) PM² CD². And by fimilar triangles, TP : TC:: PM: CG:: PM²: PM × CG or AKx BH; there- fore PM² CD:: PM: AK x BH. PM² : AK × BH. AKXBH = CD'. Whence Cor. 1. If AB be any diameter, CD its femicanju- gate, PM an ordinate; AK, BH, HM, tangents at A, B, M; then AK × BH = PM × CG = CD² =AC Xlatus rectum of AB. For we had before AK x BH (by this Prop.) CD² = (Schol. IV.) rectum. PM x CG - AB × latus Cor. 114 HYPERBOLA. THE Fig. Cor. 2. If the two parallel tangents at the ver- 36. texes, AK, BH, be cut by the two tangents MKH, QRN; then AK X BH = BQ × AR. 37. × For they are both equal to CD². Cor. 3. The fame things fuppofed, AK : BQ:: KO: OH. For AK: BQ : : AR : BH :: AK + AR or KR: BQ + BH or QH:: (fimilar triangles) KO: OH. Cor. 4. A line drawn thro' RH, would interfect the line drawn thro' KQ, ſomewhere in the diameter AB produced. This is plain from the proportions in Cor. 3. Cor. 5. If the tangent QR was to touch the op- pofite hyperbola (as gr), it will still be AK × BH- Bq x Ar. PROP. XLV. In the hyperbola and oppofite fections if TH, TK, be two tangents at H and K, and if HK be drawn thro' the points of contact, and biffected in Ọ, and if TO be drawn; it will biffect all the parallels to HK, as IN. That is, if HO = OK, then IP = PN. Thro' the center C draw FG parallel to HK; then fince HO = OK, CFCG. And CA is a femidiameter, and HO, OK, ordinates to it. And fince IN is parallel to HK; IP, PN will alfo be ordinates; and (XXXVI. Cor. 3.) IP PN; and CGE and CFD are femiconjugates. Cor. 1. If LM be drawn parallel to HK, which paſſes thro' the points of contact; the parts intercept- ed between the curve and tangent, on each fide, are equal; LI = MN. For Fig: 29. 12 V k A D C H M K G 32. 34. T M F R 30. B D M P E 35. P R S K G 31. D K M N P 33. R M n 36. A D H B B N M Pl. XVI.pa.114. N M OF B. II. THE 115 HYPERBOLA. For fince HO OK, therefore LP = and IP being = PN, the remainder IL PM, and Fig. NM. 37. Cor. 2. The tangent at A, is biffected in the point A. Cor. 3. HO, OK, and alfo IP, PN, are ordi- nates to the diameter CA, paffing thro' T the inter- fection of the tangents. SCHOLI U M. All theſe things hold equally, if one of the tan gents be drawn to the oppofite hyperbola. PROP. XLVI. If TL, TM be two tangents at H and K, and 38. HK be drawn thro' the points of contact; and from the interfection T, any line TI be drawn, cutting HK in D, and the curve in F and I. Then TF : TI :: DF: DI. Thro' F, I, draw AC, LM, parallel to HK; then (Prop. XLV. Cor. 1.) LINM, and AF BC; whence MI LN, and CF AB. by fimilar triangles, AF: LI:: TF : TI. and CF MI:: TF: TI. = multiplying, AF x CF :: LIX MI :: TF2: TI2. that is, AF x AB:: LIX LN:: TF: TI². But (Prop. XL. Cor. 3.) AF x AB :: LI × LN :: HA* : HL2: : (fimilar triangles) DF² : DI², Therefore TF² : TI² :: DF² : DI²; and TF: TI :: DF: DI. ! Cor. If it be TF : TI :: DF : DI, the line HK joining the points of contact (of the tangents TH, TK,) will pass thro' the point D. M SCHO- 116 THE HYPERBOLA. Fig. 38. 39. 40. SCHOLIU M. If one of the tangents be drawn to the oppofite hyperbola, 'tis equally true. PRO P. XLVII. If MT, AH be two tangents at any points M and A; and if the femidiameters CM, CA be drawn, to interfect the tangents in H and T. Then the tri- angle MNH = triangle ANT. And the like for the oppofite fections. › Draw PM parallel to AH. Then by fimilar triangles, CM: CH :: CP : CA : : (XLÍI.) CA : CT. And alternately, CM. CA :: CH: CT; whence (Geom. II. 17. Cor. 1.) the triangle CAH triangle CMT; which fubtracted from the qua- drilateral CANM, leaves MNH = ANT. Cor. The triangle CAH = triangle CMT. From the demonftration. PROP. 1 XLVHI. If CM, CV be any two conjugates. And MP, QV ordinates to any other diameter AC; and CD the conjugate of AC; then CQ CP² - CA - APB. This Prop. is demonftrated from Prop. XLII. and Cor. 1. Prop. XLIII. in the fame manner as Prop. XXIX. was from Prop. XV. and Cor, 1. Prop. XXVIII. And the fame corollaries will follow, viz. Cor. 1. CP CA² + CQ² = CA² + APB. = Cor. 2. CP2 CQ² = CA² = CP² APB. Cor. B. II. 117 THE HYPERBOLA. Cor. 3. QV2PM 2 CD². Cor. 4. CA: CD:: CP: QV:: CQ: PM. Cor. 5. Hence the triangle CPM triangle CQV. This is proved the fame way as in Prop. XXXVI. PROP. XLIX. Fig. 40. If AH be a tangent at any point À; MC, VC, 41. two conjugate femidiameters, cutting the tangent in Z and H; AC, CD two conjugates. Then the rectan- gle ZAH = CD². To the diameter AC draw the ordinates MP, VQ. Then by fimilar triangles CP: PM: CA: AZ, whence CP X AZ CA x PM (Cor. 4. laft Prop.) CQ × CD; therefore CP: CQ : : CD : AZ. Again, CQ: QV : : CA : AH, and CQ × AH =CA × QV = (ib.) CP × CD. Therefore CP: CQ : : AH :´CD :: CD : AZ, and AZ× AH÷ CD2. Cor. If SN be drawn parallel to CD, then FN¹ - SFG CD2. For by fimilar triangles CA: AZ:: CF: FG and CA: AH:: CF: FS and multiplying, CA: AZ X AH or CD :: CF²: FS x FG. and compounding CA: CD:: CA² + CF²: CD' + SFG:: (XLIII.) ÇA² + CF2: FN²; therefore CD² + SFG FN2, and FN-SFG = CD'. M 2 PROP. 118 THE HYPERBO L' A. Fig. PROP. L. 42. If PT, QT be two tangents at P, Q; and PQ drawn between the points of contact; and if any line LD cuts PQ, and the curve in S, A, B. Then SL²: SD²:: rectangle ALB: rectangle BDA. Draw the femidiameters CG, CM, CF, parallel to TP, TQ, LD, (which I have put without the figure to avoid confufion); and draw LI parallel to QT, to cut PQ in I. Then the triangles SLI, SOD, are fimilar, as alfo PIL, PQT; whence SD: SL :: DQ: LI. But the ratio of DQ to LI is compounded of the ratio of DQ to PL, and (PL to LI or) PT to TQ. Therefore DQ: LI:: DQ × PT :: PL x TQ Therefore SD: SL:: DQ X PT: PL X TQ. Take E a mean propor- tional between AL and LB; and H a mean pro- portional between BD and DA. Then ALB = E², and ADB H². > Then (Prop. XL.) GC² : FC² : : LP² : ALB or E, and FC: MC² :: BDA or H² : DQ². There- fore extracting the roots, GC: FC:: LP: E; and FC: MC :: H: DQ; therefore GC : MC :: PL × H: DQ × E. But (XL. Cor. 1.) PT: QT GC: MC :: PL × H: DQ × E. And multiply- ing means and extremes, PT x DQ × E = QT × PLX H. Whence H: E:: DQ × PT: PLX TQ:: SD: SL; and H2 or ADB : E2 or ALB:: E² SD² : SL². Cor. Hence in any infcribed trapezium LNDO, the diagonals LD, NO, interfeƐt in the fame point S, where the lines PQ, KR interfect, that join the oppofite points of contact. For if LD cuts the tangents PL, D, and PQ in S; we have, SL QD in L and SD:: ALB : BDA:: B. II. 119 THE HYPERBOLA. BDA :: EE: HH. And SL: SD:: E: H; and Fig. compounding, SL + SD: SL:: E+H:E; or E 42. +H: E: LD : SL. Alfo for the tangents KL, RD, it will be SL²: SD: ALB ADB. Whence, as before, it will be E+H: E:: LD: LS. Therefore PQ, and KR interfect LD in the fame point S. And by the fame reaſoning, the diagonal NO, cutting the conjugate hyperbolas in a and b, will pafs thro' S, the interfection of PQ, KR. PROP. LI. 1 If MG be any diameter, MD, GT two tangents 43. at M and G; DTN a tangent at N, cutting the former in D and T. And if the femidiameter CK be drawn parallel to ND. Then the rectangle DNT = CK2. For let CF be the femiconjugate of MG, then (XL. Cor. 1.) CF: CK :: DM: DN. and CF: CK :: TG : TN. multiplying CF² : CK² : DM × TG : DN × TN. But (XLIV,) CF² = DM × TG; therefore CK² = DN X NT. PROP. LII. If FG be drawn thro' the points of contact of two 44 tangents AF, AG; and AV be drawn thro' A their interſection, parallel to GF. And if from any point in it as V, LVO be drawn thro' the middle of FG, to cut the curve in P and L; then will VP: VL:: OP: OL. * From L draw LAQR, cutting the curve in Q, from Q draw QSP parallel to GF. M 3 Then (Prop. XLVI.) 120 THE HYPERBOL A. Fig. XLVI.) LA: LR :: QA: QR, and LA X QR 44. LRXQA. 45. The ratio of PQ to SQ is compounded of the ratio of PQ to OR and OR to SQ. But by fi- milar triangles PQ : OR : : LQ : LR; and ÖR: SQ :: AR: AQ. AR : LR × AQ. LQ X AR — LRX Therefore PQ : SQ :: LQ X And by divifion, PS: SQ :: AQ: LR × AQ. But LQX AR-LR X AQ = LR — QR × AR— LR × AQ = LRX AR — AQ — QR × AR LR x QRAR X QR = AL × QR = LR × AQ before. Therefore PS: SQ :: LR × AQ:LR X AQ; and therefore PS SQ, and confequently, fince QS is an ordinate, PS is an ordinate, and P is in the curve. Laftly, by reafon of the parallels FG, PQ, VA, VP: VLAQ: AL:: (before) RQ:RL::OP : OL. PROP. LIII. If FG join the points of contact of two tangents FA, GA; and AV be parallel to GO. Then if two tangents VH, VI, be drawn from any point. V in it. Then HI drawn thro' the points of contact, will paſs thro' O the middle of FG. For let B, L be two points in the curve; where VO, AO cut it. Then in refpect of the ordinate FG, it is (by laft Prop.) VP: VL :: OP : OL. And in reſpect of the line HI, it is (Prop. XLVI.) VP : VL:: OP: OL; therefore O is common to both lines FG, HI; and therefore HI paffes thro' O. Cor. 1. If FA, GA be two tangents, and FG drawn thro' the points of contact. Alſo HV, IV two other tangents, and HI drawn thro' their points of contact. Then if AV be drawn thro' the interfections of Fig. 37. # m n P E D M H N 1 A T 39. N P H 42. K B P 40. M RAHO N α C S B M n H 38. K 41. P M S M K L 44. T F 43. S M R P1.XVII. pa.120. UNIV M OF B. II. 121 THE HYPERBOLA. of the tangents; then will O, where FG, HI inter- Fig. fect, be the point thro' which all lines pafs, which 45. join the points of contact of any two tangents, which meet in the line AV. Cor. 2. If FA, GA be two tangents, and if AV be drawn parallel to FG; then if any line HI be drawn thro' the middle of FG; then the tangents drawn from H and I will meet fome where in the line AV. Otherwiſe it could not be VP: VL::OP: OL, as by Prop. LII. Cor. 3. If IH be continued to interfect AV in D, and the tangents DP, DL drawn; then LP paf- fing thro' the points of contact will pass thro' O. PRO P. LIV. If VH, VI be two tangents, and IHD be drawn thro' the points of contact, and if any line OPVL be drawn thro' V, to cut the curve in P and L; then two tangents PD, LD drawn from P and L, will meet fome where in the line ID. 1 Let O be the interfection of VP and IH; thro' O draw FG, fo that FOOG, and draw VAD parallel to it. Then (LIII. Cor. 2.), the tangents PD, LD, will meet in the line VA, as at D. Then if DO does not pafs thro' the points of contact H, I; let a line drawn from D to N paſs thro' the points of contact. Then (Prop. XLVI.) we fhall have VP: VL:: NP: NL. But HV, IV, being tangents, it will be (by the fame Prop.) VP: VL:: OP: OL. Therefore the former proportion is falfe, except N coincide with O. Therefore DO paffes thro³ the points of contact H, I, Or which M 4 is 45. 122 THE HYPERBOL A. Fig. is the fame, the tangents PD, LD meet in the line 46. 47. ༑ IH. Cor. If FA, GA be two tangents, and GBF be drawn thro' the points of contact. And from any point B in FG, the tangents BI, BH are drawn. Then IH (drawn thro' the points of contact H, I.) will pass thro' A. For a line drawn from A will pafs thro' the points of contact H, I, and therefore no other line can paſs thro' them. PROP. LV. If a line MN be drawn thro' the focus F, and tangents MR, NR drawn from M and N. Then the line RF drawn from the intersection of the langents to the focus, will be perp. to the line MN. Let C be the center, and make CF, CA, CT continually proportional. At T erect TL perp. to the tranſverſe axis BA. Then (Prop. XV.) a tan- gent drawn from the end of the ordinate erected at F, will cut the axis at T. And (LIII. Cor. 2.) The tangents drawn from M and N, will meet fome where in the line TL, as at R. Make SG = BA, draw GF, which (Prop. XII.) will be perp. to MR at D. Draw MLO parallel to BA. The triangles SFG, MOG are fimilar, whence SG or BA: SF:: MG: MO; that is, CA: CF:: MF MO, But (Prop. XXVI.) CF CA:: : ML: MF. Therefore ML: MF:: MF : MO, and ML X MO MF. But the triangles MDO, MLR are fimilar, being right angled at D and L. Whence MD: MO :: ML: MR; therefore MD X × MR = MLX MO MF". And MD: MF:: MF: MR. Whence the triangles MDF, MFR, are fimilar, and therefore the angle MFR = < MDF a right angle. Cor. B. II. 123 THE HYPERBOLA. Cor. A circle defcribed about the diameter MR will Fig. pass thro' the focus F, and thro' the points L, G. For MFR and MLR are right angles, and the triangle MGD being equal and fimilar to MFD, will be contained in the oppofite femicircle. PROP. LVI. 47. If two tangents HE, HK touching the conjugate 48. hyperbolas in R and D, cut any diameter AB in E and K, fo that AE BK. And if FZ be a third tangent; touching the curve at I, and cutting the former in F and G. Then the rectangle EF x GK = KH × RE, a given quantity. For fince CERA is ſimilar and equal to CPKB; therefore RE PK. And (Prop. XLIV. Cor. 3.) PK: RF:: KG: GH, and compounding PK or RE: RE+ RF:: KG: KG + GH; that is, RE : EF:: KG: KH; whence EF X KG a given quantity. KHX RE Cor. Producing LE and ZF to V; then VE x KZ = KH × RE, a given quantity. For LZ is parallel to EH, and VL to HK; therefore the triangles VEF and GKZ are fimilar; and VE:EF:: GK: KZ, whence VE x K2 = EFX GK KHX RE. PRO P. LVII. If ABFE be an infcribed trapezium; and if the diagonals AF, BE be biffected. The line drawn thro the points of biflection will pass thro' the center C. Thro' the points of contact of the oppofite fides AE, BF, draw the line HI, and the diameter GD parallel to it. Then (Prop. LVI.) GA × DB = a given quantity = GEX DF. Whence GA: GE:: DF: 49 124 THE HYPERBOLA: Fig. DF: DB. Therefore if AF, GD, EB be drawn, 49. then (Cor. 1. Lemma to Prop. LX. Ellipfis), the line that paffes thro' the middle of AF, EB, paffes thro³ the middle of GD. But as GD is på- rallel to HI, the middle of GD is the center C. 50. Cor. And if there be any four tangents to the by- perbola, and lines (or diagonals) be drawn to the op- pofite interfections. Then a line drawn thro' the mid- dle of thefe, will pass thro' the center. This is fhewn the fame way as was done in the ellipfis. PROP. LVIII. If AB be any diameter; and if two lines AP, BP be drawn thro' any point in the curve as P; and the ordinate DF be drawn interfecting the lines in E and G. Then DE, DF, DG are in continual pro- portion. Draw the tangents AN, BH, the conjugate ICL, and ordinate PŎ. Then by fimilar triangles AB BH: AO: OP and AB AN:: BO: OP : × multiplying, AB2: AN X BH:: AO X OB: OP² : : (XXXVI.) AB² : IL; therefore AN x BH = IL². Again (XXXVI.) AB² : ADB :: IL² : DF²; and by fimilar triangles AB: AD::BH: DE and AB BD:: AN: DG : multiplying, AB2: ADB:: ANX BH: DG X DE: therefore IL²: DF² :: ANX BH: DGX DE. But ILANX BH; therefore DF2DG X DÉ, or DE, DF, DG ÷. Cor. If from the ends of any diameter AB, two lines PH, PB be drawn thro' any point P of the curve, B. II. 125 THE HYPERBOLA. curve, to interfect the tangents BH, AN in H and Fig. N. The rectangle of the tangents is equal to the 50. Square of the conjugate. AN x BH = IL². This appears from the demonſtration. PROP. LIX. If C be the center, F the focus, and if CF: CA:: 51. CA: CT, and the directrix TO be drawn perp. to AB. And if any line MNO be drawn, and the lines FN, FO, MFH be drawn to the focus. Then OF biflects the angle NFH. Draw MP, NQ perp. to TO, and ND parallel to MF, cutting FO in D. Then (by fimilar trian- gles) MP: NQ:: MO: NO:: MF: ND; and alternately, MP : MF : : NQ : ND. MP MF: CA: CF:: NQ NQ: ND:: NQ: NF. And ND fore< = NFD = NDF (by the parallels) DFH. PROP. LX. But (XXVI.) NF; therefore = NF. There- If AB be any diameter, BG a tangent at B equal 52. to the latus reilum. If AG be drawn cutting the ordinate PM (produced) in D. Then PM BP X PD. 2 For XXXVI. Cor. 1.) AB: BG: APB: PM". And by fimilar triangles, AB : BG:: AP: PD: AP x PB: PD x PB:: APB : PM. Therefore PD x PB PM'. : Cor. Hence PM the fquare of the ordinate, is al- ways greater than PB x BG, the rectangle of the parameter and abfciffa; by the rectangle GF x FD, which is fimilar to the rectangle AB x BG. And by reafon of that excefs the curve is called an HY- PERBOLA For 126 THE HYPERBOLA: Fig. For if GF be parallel to AB, the, rectangle 52. DFG is fimilar to ABG, for by the fimilar trian- gles ABG, GFD, AB: BG:: GF: FD. SCHOLIU M. The rectangle under the diameter AB, and its parameter BG, is called the Figure of that diameter. PROP. LXI 53. If QT, PT be two tangents, and if QF, PF be drawn from the points of contact to the focus F; then the line TF will biffect the angle QFP. 54. : Make CF CA:: CA: CG, and draw the di- rectrix DGR perp. to AB. Produce FT to K, and draw LR, KR (Prop. L.) tangents at L, K. Let QD, PE be perp. to DR. Then (XXVI. Cor.) QF: PF :: QD : PE :: (fimilar triangles) QR: PR : : (XLVI.) QN: PN; therefore (Geom. II. 25. Cor. 1.) FN biffects the angle PFQ. PRO P. LXII. If CB be the femitranfverfe, BD the femiconju- gate axis; S, F the foci, PM an ordinate. Thro' C and D, draw CO, to cut the ordinate PM in O, and make CICO. Then SM BI. For (Cor. 1. Prop. V.) CB: CS :: CP; SM + CB. Whence SM X CB + CB SM X CB CS x CPCB. = CD = CS, and by fimilar triangles, CSX CP; and But (def. 6.) CB: CD or CS:: CP: CO or CI, and CS x CP CBX CI; therefore SM x CB — – — CB x CI - CB; whence by divifion SM CI = CB = BI. Cor. 1. The fame things remaining, FM = AI. For D M Fig.45. B F H 47. P 46 N S n Ꮐ R T m N : E 49. B H H A F 48. 2 料 ​K/B P 9 50. P 51. H N E N B D HV M 10 P1.XVIII. pa.126. N ස OF B. II. 127 THE HYPERBOLA. For FM SM + AB BI + BA AI. = Cor. 2. If MS be produced to N, fo that SN = SC, and NL be perp. to CB. Then CB — SL, BD, and SM are continually proportional; but when P is between S and B, it is CB + SL. For by fimilar triangles, SM : SP:: SN or SC: SL; whence SL × SM = SC × SP = CF × SP = CFX CPCS CFX CP-CS2. But we had CF x CP or CS x CP SM X CB + CB²; therefore SLX SM SM x CB + CB² — CS²; and SM X CB-SL CS2 CB (II. Cor.) = BD². But if P is between S and B, SLX SM = CF × CS- CP, &c. Cor. 3. Put Z for the cofine of the angle BSM, CD2 then will SM = CB + Zx CS® Or when BSM is greater than a right angle, it is CBZ X CS. For by trigonometry, Rad. (1.): SN or SC : : S.SNL (Z): SL = Z× SC, and CB + SL = CB + ZX SC. PROP. LXIII. Fig. 54. If ABDC be an infcribed trapezium, and any point 55. E be taken in the curve, from which the two lines EN, EH are drawn parallel to any two adjoining fides as AB, AC, to interfect the oppofite fides in N and Q, and in H and R. Then taking the rec- tangles upon the oppofite fides; ERXEH will always be to EN X EQ, in a given ratio, wherever E is placed. Draw DFG, BPL parallel to AC; thro' C and L draw COLS; and thro' the middle of AC, BL, draw the diameter MK, which will biffect the pa- rallels DF, TE; and alfo SG, OR which are ter- minated 128 THE HYPERBOLA. Fig. minated at AB, CL. Then will SD FG, OT 55. ER; and theſe would alſo be equal if SG, RH were tangents. 56. By reaſon of the parallels, and from the fimi- lar triangles, HO: SD:: OC: SC :: AR: AG: and alternately, HO: AR or EQ:: SD or FG: AG. Again, BP or ER PN: DG GB; and multiplying the refpective terms, HO × ER: EQ × PN: FG × DG: AG × GB :: (XXXIX Cor. 2.) ER X RT or ER X EO: RAX RB of EQ X EP. And compounding, FG × DG : AG x GB:: HOX ER +EO × ER: EQX PN + EQ × EP : : EH × ER : EQ × EN. But it is evident, as long as the points A, B, D, C, are fixed; that FGD to AGB is a given ratio, and confequently that of REH to QEN. Cor. 1. The fame things fuppofed, it will be EH : EN: EO: EP:: HO: PN. For EHXER : EQ × EN :: FG X GD: AG × GD :: HO × ER : EQ × PN. And alternately, EH X ER: HO × ER :: EQ × EN : EQ × PN; and by dividing equally, EH: HO:: EN: PN; and by divifion, EH: EO :: EN : EP. Cor. 2. If the points A, B, C, E, be fixed; and D any way changed; it will still be EH EN:: HO: PN:: EO: EP, or RT: RB a given ratio. Cor. 3. If the points A, C, coincide; MQ be- comes a tangent. And if B, D, coincide; BN becomes a tangent. Cor. 4. If BC be drawn to cut EH it be made as EH: EI::EN: EV. ing drawn will touch the curve in B. in I; and if Then BV be- For (Cor. 2.) when D is moveable, it is, EH : EN EO EP. And if D comes to B, then :: : (Cor. B. II. THE 129 HYPERBOLA. (Cor. 3.) BND becomes a tangent; in which cafe Fig. H comes to I and N to V; and then it is, EH: 56. EN: EO: EP:: EI: EV. Cor. 5. If we make EH: EN :: EO: EP, and 55. draw CO, BP; they will meet at L in the curve. By Cor. 2. • Cor. 6. If there be five points, B, L, D, C, E; and thro' any point E, there be drawn EH, EN, cut- ting BD, BL, DC, LC, in N, P, H and O; fo that EH EN: EO: EP. Then CA, BA being drawn parallel to EH, EN, they will meet at A in the curve. This is the reverſe of the foregoing. For if A was out of the curve, L would not be in it. PROP. LXIV. If ABDC be an infcribed trapezium, and from 57. any point E, of the curve, four lines EL, EK, EO, EM, be drawn to the four fides of the trapezium, in any angles whatever, equal or unequal; cutting them in L, K, O, M. And likewife from any other. point e, four more lines, el, ek, eo, em, be likewise drawn respectively parallel to the former, cutting the fides in l, k, o, m. Then taking the rectangles of the lines falling up- on the oppofite fides, it will be, rectangle EL × EM: el × em: : EK × EO : ek × eo. Thro' E, e, draw EQN, eqn, parallel to AB; and ERH, erb, parallel to AC. Then by fimilar triangles, ER er:: EL: el; and EH: eb:: EM: em. And multiplying, ER X EH: er X eb:: EL × EM: el × em. eo. Again, EQ: eq :: EK: ek, and EN: en:: EO: And multiplying, EQX EN: eq xen: EK XEO: ek x eo. But (Prop. LXIII.) ER X EH : × er X 130 HYPERBOL A. THE Fig. er X eb:: EQ × EN : eq × en. Therefore EL X 57. EM: el xem:: EQ × EN: eq x en: :) EK × EO: ek x eo. 58. 59. Cor. 1. In the infcribed trapezium ABCD; if AB, CD, produced, interfect in X; and the diago- nals AC, BD, in Y; and YEXe be drawn. Then EX: eX:: EY : eY. For ſuppoſe BD, AC to become the diagonals; and that EL, EK, EO, EM, and el, ek, eo, em, all lie in one line Ee; and that AB, CD, inter- fect in X, and AC, BD, in Y; and that Ee paffes thro' X, Y; which is a particular cafe of this Prop. Then L, /, M, m, coincide in X; and K, k, O, o, in Y. Then the proportion EL × EM: el xem :: EK X EO: ek X eo, becomes EX": eX²:: EY²: eY", and EX: eX:: EY: eY. 2 Cor. 2. If XA, XC, be two tangents, and AYC · paffes thro' the points of contact. It will be EX: ex:: EY: eY. For let B approach to A, and D to C; then BD will coincide with AC. And it will still be EX: eX :: EY: eY, as in Prop. XLVI. PRO P. LXV. If any line MN be drawn thro' the focus F, and ZCQ be a diameter drawn parallel to it; AB the tranfverfe axis. Then AB, QZ, MN, are conti- nually proportional. Draw the diameter CDL thro' the middle MN, and the ordinate MP perp. to AB, and tangent MTO. Then (VII. Cor. 3.) CF x CP CAX FM+ CA, and CA X FM CFX CP-CA = = (XV.) CF x CPCT X CP, or CA x FM TF x CP. And by fimilar triangles CT: CO :: TF: B. II. 131 THE HYPERBOLA. TF: FM TF x CP or CA x FM: FM x Fig. CP :: CA: CP:: CT: CA. Whence CO = CA. 59. But (XLII. Cor. 5.) CZ CO x CG = CA × LM; and multiplying, QZ = BA × MN. = Cor. 1. If F be the focus, ZQ a diameter parallel to MF; MT a tangent at M, cutting ZQ in O; then CO CA. From the demonftration. Cor. 2. If AB be the axis, F the focus, MP an ordinate, MT a tangent. Then CA × FM = TF * CP. From the demonſtration. 2 I Cor. 3. If any line MN be drawn thro' the focus F; the rectangle MFN MN x latus rectum. = ½ Let Llatus rectum; then (Prop. XXXVIII.) CZ² : MFN :: CA² : AFB :: (IV. Ĉor. 3.) CA² : CA XL:: CA: L:: CAX ML: LX ML. But CZ CAX ML; therefore MFN÷LX ML. I PROP. LXVI. I 2 In an equilateral hyperbola, the fquare of the femi- 6ɑ: tranfverfe is equal to the rectangle of the distances of either vertex from the foci. CA' SAF. For (Prop. III.) the rectangle of the diſtances of the two foci, from either vertex, is equal to the fquare of the femiconjugate. But (Def. XII.) the tranfverfe and conjugate are equal; therefore that rectangle is equal to the fquare of the femi- tranſverſe axis, N PROP. 132 THE HYPERBOLA. Fig. 60. 61. 62. PROP. LXVII. In an equilateral hyperbola, the rectangle of the distances of the ordinate from the two vertices, is equal to the fquare of the ordinate. APB PM2. For (Prop. VI.) as fquare of the tranfverfe : fquare of the conjugate: : fo the rectangle of the diſtances of the ordinate, from the two vertices BPA to fquare of the ordinate. : terms are equal (Def. XII.). latter are equal. 2 But the two firſt Therefore the two Cor. CP² — CA² = PM², C being the center. PRO P. LXVIII. If AV be an equilateral hyperbola, AC the femi- tranfverfe axis, CQ perp. to CA; QV parallel to CA. Then CA² + CQ² = QV³. 2 For (Prop. XXVIII.) if CD be the femiconju gate, then CA²: CD:: CA² + CQ: QV². But ČA CD'; therefore CA+CQ² = QV². 2 Cor. The hypothenufe AQ= ordinate QV parallel to CA. For AQ² = CA² + CQ² = QV²; and AQ = QV. PROP. LXIX. If CA be the femitranfverfe axis, CF, CH, the affymptotes; AD the femiconjugate; PM an ordi- nate continued to F. The rectangle HMF AD'. By fimilar triangles, CP2: PF':: CA: AD' :: (VI.) APB or CP-CA: PM: : (by divifion) CP² - CP² + CA or CA: PF- PM²; that 2 is, B. II. THE HYPERBOLA. 133 is, CA² AD²:: CA: PF PM'; therefore Fig. AD PF PM HMF. Cor. 1. If the ordinate be continued to the affymp- totes, the parts between the curve and afymptotes are equal. FM GH. = For PH PF, and PG remainders are equal, FM PM; therefore the GH. Cor. 2. If more ordinates be drawn to cut the af fymptotes, their rectangles are all equal. PROP. LXX. 62. If from the point M of the curve, MK be drawn 62. parallel to the tranfverfe axis AB, cutting the afſymp- totes in I and K; then the rectangle KMI CA. For the triangles IMF, KMH and CAD are fi- milar, whence AD: AC:: MF: IM, and AD : AC:: MH: KM; and multiplying, AD²: AC² :: MF × MH: IM × KM. But (LXIX.) MF × MH AD; therefore IM X KM AC². × × Cor. 1. The rectangle IM x MH KM × MF. For by fimilar triangles KM: MH;: IM: MF. Cor. 2. If more lines be drawn parallel to CA ; all their rectangles are equal to one another. PROP. LXXI. If two parallel lines PS, FH, cut the hyperbola, 63. and the affymptotes; the rectangle FMH PQS. rectangle 64. Thro' M and Q draw OL, IN perp. to the axis. Then by reafon of the parallels, the triangles PQI, FMO, are fimilar, as alfo MHL, QSN. There- fore MO: MF:: QI: QP, and ML : MH :: N 2 QN, 134 HYPERBOLA: THE Fig. QN: QS. And multiplying MOX ML: MF x 63. MH :: QI × QN : QP × QS. But (LXIX. 64. Cor. 2.) MO × ML = QI× QN; therefore FMH = PQS. 65. 66. Cor. 1. If the tangent DG be parallel to FH, then the rectangle FMH DG'. Cor. 2. If the diameter AB be parallel to HM, then rectangle FMH CB". PRO P. LXXII. Any right line FMGH be drawn thro' the curves the parts of the line between the curve and afymp- totes, are equal; FM GH. For (Prop. LXXI.) rectangle FMH HGF; that is, FM X MG + FM x GH GH x GM + GH × MF; and FM X MG GH X GM; whence FM = GH. Cor. 1. Hence FG MH. Cor. 2. If AB be a tangent at D; then AD = DB, or the tangent is biffected in D. For in this cafe M and G coincide in D. Cor. 3. If a diameter CD be drawn thro' the point of contact D, it biffects all the parallels MG, and all the lines FH. PRO P. LXXIII. If from two points in the curve, M, Q, there be drawn the parallels MH, QN, to cut one affymp- tote; and likewife the parallels MO, QI, to cut the other affymptote. Then the rectangle MÒx MH = the rectangle QI × QS. Draw EML, PON perp. to the axis. Then the triangles EMO, PQI are fimilar; and alfo MLH, B. II. 135 THE HYPERBOL A. MLH, QNS. Whence ME: MO:: QP: QI; Fig. and ML : MH :: QN: QS. And multiplying, 66. MEX ML: MO X MH: QP × QN: QIx QS. But (LXIX. Cor. 2.) ME × ML = QP × QN therefore MO × MH = QI × QS. SCHOLIUM. The Prop. holds equally true, if QS, &c. cut the affymptote beyond C towards h. PROP LXXIV. If two lines KN, FT cut one another in E; and 67. likewife cut the curve and affymptotes; then rectan- 68. gle LKM : DFV :: KEN: FET. Thro' N, draw IG parallel to VD; then the tri- angles LNG, and LED are fimilar; and alfo MNI, MEV; whence LE: LN:: ED: NG, and ME: MN:: EV: NI. And multiplying, LE × ME: LN × MN::EDXEV:NGXNI. But LEX MELK + KE × EN + NM = LK × EN + KE × EN + LK + KE × NM = KE× EN + LK × EM + KE × NM (KE × LK) = KEX EN + LK × KM (fig. 67.). Alfo by the fame reafoning, DEX VE FET + DFV. Alfo (LXXII.) LNM LKM, and GNI DFV. Whence we have KEN + LKM: LKM:: FET + DFV: DFV. And by divifion, KEN: LKM:: FET: DFV, PROP. LXXV. All infcribed parallelograms, between the affymp- 69: totes and the curve, are equal; ABCD = GFCH. For (by Prop. LXXIII.) the rectangle AB X AD = GF X GH; whence AB: FG:: GH: N 3 AD, 136 HYPERBOLA. THE : Fig. AD, or CD: CH:: CF CB; therefore (Geom. 69. III. 9. Cor. 1.) parallelogram AC = parallelo- gram GC. jo. Cor. 1. The hyperbola continually approaches to the affymptote; but can never touch it, till it be in- finitely continued. For EL being reciprocally as CE, EL continu- ally diminiſhes as CE increaſes, and when CE is infinitely great, EL will be infinitely ſmall. Cor. 2. If a line be drawn parallel to one afſymp- tote, it will interfect the oppofite hyperbolas in one point, and no more. For AD can only cut the curve in A, for the curve AL afterwards approaching CE, muft di- verge from AD. And the fame line continued can never cut the oppofite hyperbola, becauſe it lies wholly without the oppofite angle, in which the curve is wholly contained. Cor. 3. The affymptote may be confidered as a tan- gent to the curve at an infinite diftance. From Cor. 1. and Cor. 4. Prop. XV. PROP. LXXVI. If the lines AB, DF, touch the hyperbola; then the triangles CAB and CDF made by the tangents, and affymptotes, are equal. From the points of contact, E, G, draw EH, GI parallel to the affymptote CB; and EO, GP, parallel to the affymptote CF. Then (LXXII. Cor. 2.) fince AE EB, and FG GD; there- = fore AH HC, and CO = OB; alfo FI = IC, and CP CD. www. = But triangle ACB to triangle FCD: CA X CB or 2CH x 2CO: FC x CD or 2CI X 2CP :: CH X B. II. THE 137 HYPERBOLA. CH × CO : CI × CP :: parallelogram CHEO : Fig. parallelogram CIGP; but (Prop. LXXV.) CHEO 70. CIGP; therefore triangle ACB = FCD. Cor. 1. All infcribed triangles are equal to one ano- ther, and double to the infcribed parallelograms. For triangle ACB parallelogram CHOE:: 2CH X 2CO 2 or 2CH x CO: CH X CO:: 2 : 1. Cor. 2. If GI be parallel to the affymptote CB; and GF a tangent at G; then the fubtangent IF = CI the distance from the center. For DF is biffected in G; and therefore CF is biffected in I. Cor. 3. Any two tangents to the curve LG in- terfect one another within the angle of the affymptotes BCF. For if BA be a tangent at E, this interſects CF (the tangent at an infinite diſtance) in A. And therefore the fame BA interfects any other tangent FD, between A and B, Cor. 4. Hence alfo the rectangle ACB is equal to the rectangle FCD, Cor. 5. If AD and FB be drawn, they will be parallel. For CA: CD :: CF: CB, from the equal rec- angles CAB, CDF. PROP. N 4 138 THE HYPERBOLA. Fig. LXXVII. PROP. 71. If thro' any point F in the curve, a line FP be drawn cutting the curve and affymptotes in P, B, M; and from В the tangent BEA; and from E the point of contact, EH; as likewife FQ, PR all parallel to BC. Then 2EH = FQ + PR = FQ+ PR = FQ-pr in the oppofite fection. For draw FL parallel to CA, then the triangles MPR, FBL, are fimilar, and MP BF; there- fore PR LB, and LC FQ; and fince AE = EB (Prop. LXXII. Cor. 2.) = 2 AB; therefore CL + LB = HE CB, and 2HE CB = I रं QF + PR; or 2he = QF QF- pr. Cor. 1. Hence the fum of any two lines FQ, PR (made by drawing any line BP, from the fame point B in the affymptote), will always be the fame. For that fum is always equal to 2EH. Cor. 2. If BP be parallet CM; then 2EH = FQ PRO P. LXXVIII. 72. If AC be the femitranfverfe of the equilateral by- perbola AM, CF its affymptote, and CH perp. to CA, and HM parallel to CA. Then if MF be drawn perp. to the affymptote, 2CH x HM CF MF². For draw the femiconjugate AD, which is equal to AC. The triangles CAD, CHV, VFM are fimilar, and ifoceles; whence CH = HV, VF = MF. And VF: VM :: VH: CV, and VM x VH = VFX CV. Then CV 2CH x HM = 2HV X HM 2HV2, and = 2HV × X HV + VM Fig.59. M G B 60. P N 6.2. F K B n m F B E 65. H G E 2.9 67. B H G m 69. C H D F Į FP C 61. 64. F 63. A C B 66. 68. E K 70. P H B F Ꮐ OE M 71. : P R Pl. XX. pa 138. M OF ICH B. II. THE 139 HYPERBOLA. HV + VM = 2HV² + 2HV × VM 2 — 2CV x VF+VF VF² CV + VF — = CF2 VF CF2 MF2. CV² + Fig. 2 VF 72. Cor. 1. In a right angled hyperbola; the affymp- totes make a right angle with each other. = For fince CAAD, the angle ACD CDA a right angle. And twice ACD 2. angle the angle of the affymptotes. a right Cor. 2. If SB be drawn perp. to the afymptote from the focus; CB = CA. For the triangles CAD, CBS, are ifoceles, and fimilar; and CD CS; therefore CB CA. = z PRÒP. LXXIX. If thro' the given point G in the curve, the line 73. GA be drawn parallel to the affymptote CH. And any line GL be drawn cutting the curve and affymp- tote in F and L; and if FK be drawn parallel to the afymptote CB. Then KL will always be equal to the given line AG. Produce FG to D; then the triangles DGA, FLK are fimilar and equal, for the reſpective ſides are parallel; and (Prop. LXXII.) GD = FL; therefore AG KL. Cor. If from F in the curve, FK be rallel to the affymptote CB; then FK part of the other affymptote. drawn pa- AD, the PROP. 140 THE HYPERBOL A. Fig. 74. PROP. LXXX. If CA, CD be the femiaxes; the afymptote CF, biffes the line DA, drawn between the ends of the two axes; and all lines HK parallel to it, contained between the conjugate hyperbolas DH, AK. For draw AE perp. to AC, and equal to CD; and draw DE; then DCAE is a rectangled paral- lelogram. But (Def. XI.) the affymptote CF paffes thro' E, and (Geom. III. 2.) the diagonals CE, AD biffect one another in O; therefore DO = AO. Again, fince DA, HK are parallel to the af fymptote CG; therefore (Prop. LXXV.) CI × IK = CO×OA = CO × OD = (LXXV.) CI × IH therefore IKIH. Cor. 1. If DA joining the ends of the two axes, cut the affymptote in O; then 4CO DA² = CA + CD2. For ACD is a right angle, therefore (Geom. II. 20. Cor. 6.) OC OD OA, and CO² = DO =AO², and 4CO² = 4DO² = 4ÃO² = DA² = CA² + CD2. 2 Cor. 2. The rectangle of the fides of the infcribed parallelogram, is equal to the fourth part of the fun of the fquares of the two femiaxes. For draw AP parallel to CE, then COAP is an infcribed parallelogram, and CO × CP or CO the rectangle of the fides; and that is CA² + CD2 4 PROP B. II. 141 THE HYPERBOLA. LXXXI. Fig. PROP. If AB be any diameter; CG, CL, affymptotes; 75. GH a tangent at B. Then GH is equal to the con jugate diameter of AB. For let KF be a tangent at A, CD the femi- conjugate. Then, becauſe (LXXV. Cor. 3.) the affymptote FG is a tangent at an infinite diftance, therefore (Prop. XXXIV.) AK × BH = CD²; but the triangles CBH, CAK are equal and fimi- lar, and AK = BH; therefore BH2 CD²,´and BH = CD, and GH or KF = 2CD the conjugate of AB. Cor. 1. If CG be an affymptote, BG a tangent at B. Then CB, BG are femiconjugates to each other. Cor. 2. Hence the further any diameter AB is from the axis; the greater it is, and the greater its conju- gate GH. PROP. LXXXII. If F be the focus, and it be CF: CA:: CA: 76. CT, and TE be perp. to CA; and if MF be drawn from any point of the curve to the focus, and MN parallel to the afymptote CO. Then MN - MF. Draw ME parallel to CA, and AL perp. to it, then (Def. 6.) CL CF, and the triangles MEN, CAL are fimilar, and MN: ME:: CL or CF: CA :: (XXVI.) FM : ME. Therefore FM = MN. PROP. 142 THE HYPERBOLA: Fig. 77. 78. \ PROP. LXXXIII. If F be the focus, and FP perp. to CA, and from M any point in the curve, MP be drawn parallel to either affymptote, to cut FI in P. Then FM-MP latus rectum; or FM + MP, when Mis on the other fide of I. I 4/4 For if CF, CA, CT be in geom. progreffion, and TN perp. to CA. Then (LXXXII.) FM = MN, and FI = IQ = (by the parallels) PN; therefore NM - MP = PN = FI; that is, FM MP-FI. PRO P. LXXXIV. If CF, CB be affymptotes, and DA be drawn pa- rallel to the asymptote CB, cutting the conjugate by- perbolas in D and A; then CD, CA drawn from the center to the interſections, will be conjugate dia- meters. = Thro' A, D draw the tangents FG, FL. Then (LXXX.) fince AO = DO, and (LXXII. Cor. 2.) FAFG; therefore AO or OD CG or LC, and DA = CG or CL. CG or CL. Whence (Geom. I. 5. Cor. 3.) CD is parallel to AG or FG. And likewife CA is parallel to LD or LF; and there- före (Def. 7.) ĈA, CD are conjugate diameters. Cor. 1. Hence it appears that the two conjugate di- ameters NA, PD, belong to the conjugate hyperbolas AM, DH; and alfo to the oppofite hyperbolas KN, PQ. And that all thefe four conjugate hyperbolas, make but, as it were, one figure BAMHDÍKNPQ • Cor. 2. Hence also the conjugate hyperbola DH paſſes thro' the ends D, of all the conjugate diameters PD B. II. 143 THE HYPERBOLA. PD; as well as the hyperbola AM paſſes thro' all the Fig. ends A, of the first diameters NA. Cor. 3. In an equilateral hyperbola, every diameter 78. is equal to its conjugate. For then (fig. 73.) CA CD, and therefore CO is perp. to DA, and to all the parallels HK. Or (in fig. 77.) CO is perp. to DA, and therefore CD = CA. PROP. LXXXV.. If LKE be drawn parallel to the affymptote CG, 79. and any line DV be drawn, cutting this line, the curve, and affymptotes in E, F, T, D, V; and the tangent GBQ be drawn parallel to DV, and CB the diameter. Then BG: CG:: rectangle FET: 2KE × EV, :: Draw LP parallel to GQ. Then (LXXVI. Cor. 1.) 4CL x LK GC x CQ. And by fi- = milar triangles CL: LP: CQ: QG :: CQ × CG or 4CL × LK: QG × CG. Whence CL x CG X QG = LP x 4CL X LK, and CG x QG= 4LP x LKCG X 2BG. × × — LK × Then (LXXIV.) LKM : DFV :: KEN : FET. But KM = EN infinitely; therefore LK: KE:: DFV: FET. And LK× FET = KEX DFV = KE X BG². And KE × BG² × CG = FET X CG, or KE x BG X 2LP X LK FET × CG. And 2KE X BG x LP = FET, and BG: CG:: FET: 2KE X LP × EV. X Cor. 1. CG X BG 2LP x LK. From the demonftration. LK X CG X or 2KE Cor. 2. If more parallels be drawn it will be FET fet: KEV: kev, Cor. $44 HYPERBOLA. THE Fig. 79. 80. Cor. 3. If two parallels meet one line LE; then FET: fot:: KE : KO. Cor. 4. Hence alfo fet: fot: kev: Kov. : PROP. LXXXVI. If FBE be an hyperbola between the affymptotes Cf, CL; and it be CG: CH :: CK: CL; and lines or ordinates be drawn parallel to the other affymptote Cf. Then the area GHDF area KLEB. The infcribed rectangles being equal (Prop. LXXV.), their fides are reciprocally proportional, EL: BK:: (CK: CL:: (hypoth.) CG: CH: :) DH: FG; therefore if any lines CG, CH, CK, CL are proportional; the ordinates GF, HD, KB, LE will alfo be proportional, and in the fame ra- tio inverfly. GL Divide GL into an infinite number of equal parts at I, P, G, R, &c. in geometrial progreffion; and draw the ordinates IM, PN, QO, RS, &c. parallel to GF or Cf; let the fame be continued to L, and let r to s, be the ratio of the terms of the progreffion. Whence we ſhall have i r:s::CG: CI:: CI: CP:: CP: CQ:: CQ: CR, &c. and (Propor. XXII.) r : s : : GI : IP :: IP : PQ : : PQ: QR, &c. and by what went before s:r::IM: PN:: PN: QO: : QO : RS, &c. and multiplying the two laft, : rs: rs :: GI × IM : IP × PN: IPX PN : PQ × QO:: PQ X QO: QR × RS, &c. therefore the infinitely ſmall parallelograms are equal; that is, GIM IPNPQO QRS, &c. KVA = VTW = TXZ = XYa, &c. Now, in the progreffion CG, CI, CP, CQ, &c. if we take any equidiftant terms CG, CH, and CK, CL (Propor. XXI.) thefe fhall be proportional; that is, CG: CH :: CK : CL. And the terms CH, B. II. 145 THE HYPERBOL Á. CH, CL, being equidiftant from CG, CK; there Fig. is the fame number of parallelograms between G 80. and H, as between K and L. And theſe paralle- lograms being all equal, the fum of all between G and H, is equal to the fum of all between K and L. But by the method of indiviſibles, the fum of all the parallelograms GIM, IPN, &c. compoſe the area GHDF; and the fum of all the parallelo- grams KA, VW, &c. compofe the area KLEB; therefore the area GHDF area KLEB. Cor. 1. If CF, CD be drawn, and alfo DH, FG 81. parallel to the asymptote Cf. Then the fector CFD = area FGHD. ww For fince. rectangle CGF CHD, triangle CGF CHD; take away CGq common, then CqF GqDH. Add FqD common, then CqF + FqD or ſector CFD = GqDH + FqD = FGHD. Cor. 2. If it be CG: CH:: CK: CL; then feltor CFD fector CBE. For CFD FGHD, and CBE = BKLE. Cor. 3. If CK: CL :: Cd : Cf. And BK, EL be drawn parallel to Cf; and Dd, Ff to CL; then fector CBE = CDF = area BKLE = DdfF. For CK: CL: Cd: Cf:: CG: CH. Whence fector CFD = DdfF = FGHD, &c. Cor. 4. Hence the ſpace DdfF Space DHGF. For the parallelogram dG added to each makes the fame ſpace CƒFDH. Cor. 5. If CG, CH, CK, CL are in geometrical progreffion; then fector CFD, CDB, CBE are equal; and also the Spaces FGHD, DHKB, BKLE. Cor. 6. If CL be the nth term in geometrical pro- greſſion, in the ratio of CG to CH; then feltor CFE =21 × CFD; and FGLE = 2 → 1× FGHD. — n- .XFGHD. For 146 HYPERBOLA. THE Fig. For if CH is the fecond term, then CFD = 1× 81. CFD. If CK be the third term, then CFB = CFD + CDB = 2 x CFD. If CL be the fourth term, CFE = FCD + DCB + BCE = 3 × CFD, and ſo on. And the like for the areas FGHD, FGKB, &c. Cor. 7. If CG, CH, CK, CL, are proportionals; then LE, KB, HD, GF are also proportionals, and in the fame ratio, or FG, DH, BK, EL are pro- portionals, in the inverse ratio. From the demonftration. SCHOLIU M. Hence if CG = 1, and CH, CK, CL, &c. be any numbers; the (fectors or) hyperbolic Spaces GFDH, GFBK, GFEL, &c. are analogous to the logarithms of theſe numbers. For (Cor. 6.) whilſt the numbers CG, CH, CK, CL, &c. proceed in geometrical progreffion, the correfpondent (fectors or) ſpaces, proceed in arith- metical progreffion; and therefore (from the na- ture of logarithms) are proportional to the loga- rithms of thefe numbers. 82. Make CGCN 1. Then by computation from infinite feries, &c. if the number CH be 10, the ſpace GFDH or log. 10. will be found 2.30258509. If the number CK be 100, the fpace GFBK or log. 100 will be 4.60517018. If the number CL be 1000; the ſpace GFEL or log. 1000, will be 6.90775527. And if the hy- perbolic ſpaces correfponding to the intermediate numbers be likewife computed; they will be the logs. of theſe numbers. And theſe logarithms, be- cauſe they are computed from the hyperbolic areas, are called hyperbolic Logarithms. PROP, Fig.52. A m 54. CH L B B P. M T F S P h 56. I PN B 57. CM H m h D K E k B C H M P E 53. K T 55. 9 E R G 58. Pl. XIX. pa. 130. UN во OF K b Fig. 72. B M H F B G 73. N D 74 T E A H N Q n M 76. 12: B N K C 75. G •D L H I 78. M F M80. GIPQR H BAWZ CGH K 81. N E टेसे E P Z. N 83. C G P m 77. TA G B 79. M 82. D B E P1.XXI.p.146. B. II. 147 THE HYPERBOLA. Fig. PROP. LXXXVII. If two hyperbolas QE, FD, have the fame af- 83. fymptotes CP, CH; if HDE be drawn parallel to the afymptote CP; then HD is to HE every where in the given ratio of CN to CP. For the infcribed rectangles are CG x GF, and CG X GQ. Therefore (LXXV.) CH x HD = CG X GF, and CH× HE = CGX GQ. Whence CGXGQ. GF : GQ :: CGFN : CGQP : : CH × HD : CH × HE :: HD: HE. Cor. 1. The area FGHD: area QGHE:: ordi- nate GF: ordinate GQ or as infcribed parallelo- gram CF infcribed parallelogram CQ. : For the ordinate HD to HE is every where as CN to CP. Therefore if the baſe GH be divided into an infinite number of equal parts, and ordi- nates HDE erected. It will be GF: GQ:: HD: HE: . fum of all the HD. fum of all the HE: area FGHD: area QGHE. Cor. 2. If CG = GS = 1, and GF 84. •43429448; then the area FGHD is the common lo- garithm, or Brig's logarithm of the number CH. For the hyp. logarithm is to the common loga- rithm, as 2.30258509 to 1, or as I to .43429448, or as GS to GF; that is, (Cor. 1.) as the area GSEH to GFDH. SCHOLIU M. = Hence Brig's logarithms will be had by comput- ing the hyperbolic areas GFDH, whofe infcribed parallelogram is 43429448. Thus if CG 1, CH 10, the area FGHD = 1; if CH = 100, the area FGHD = 2. If CH = 1000, FGHD 3, and fo on; and the like for the intermedi- Q ate 148 THE HYPERBOLA. Fig. ate numbers. Likewife fince the area at G is no- 84. thing; therefore the log. of CG or 1 is o; and for numbers lefs than CG the area lies on the other fide of SG; and therefore the logarithms of numbers leſs than 1, are negative. And fince the area SGCB is infinite, therefore the logarithm of o, is an infi- nite negative. 85. It is plain likewife that logarithms may be of as many different forms as you pleaſe. For if CG be I, then the log. GSEH will be as the infcribed parallelogram CRSG. Therefore as that paralle- logram is greater or leſs, the logarithms of numbers will be proportionally greater or lefs. And there- fore in different fpecies of logarithms, the loga- rithms of particular numbers will be in a given ratio. On the contrary, if you have the logarithms of numbers, the hyperbolic area may be computed. For it will be as .43429, &c. to the infcribed paral- lelogram CGSR:: fo log. of CH, to the area GHES or ſector CSE. PROP. LXXXVIII. If AN be an equilateral hyperbola, AM any hy- perbola with the fame tranfverfe axis. The area ANP of the equilateral hyperbola: area AMP of the other hyperbola: as the femitranfverfe CA: to ſe- miconjugate CD. ·CA2 For if CE be the femiconjugate of AN, then CECA. And (Prop. LXVII.) CP² — CA² = PN. And (Prop. VI ) CA²: CD² : : CP² or PN2 PM. And CA: CD:: PN: PM. Therefore if the abfciffa AP be divided into an in- finite number of equal parts and ordinates erected; it will always be CA: CD:: PN: PM :: fum of all the PN: fum of all the PM:: area ANP : area AMP. Cor. B. II. ,149 THE HYPERBOLA. Cor. i. The areas of two hyperbolas on the fame Fig. tranfverfe, and having the fame abfciffa, AP; are 85. as their conjugates. Cor. 2. As femitranfverfe CA Semitranfverfe 86. CB: CANP the area of the hyperbola AN: CBMP the area of the hyperbola BM, on the fame abfciffa CP. By reaſoning as before CD: CD² + CP² :: CA²: PN² :: CB2: PM²; and CA: CB : : PN : PM all the PN: all the PM :: area CANP : :: area CBMP. Cor. 3. Therefore if the area of an equilateral by- perbola be known; the area of any other hyperbola may be found. PROP. LXXXIX. } If NAM be a hyperbolic conoid, made by revolv- 87. ing round the tranfverfe axis. If C be the center, BA the tranfverfe; then 2CA + AP : CA + AP:: circumfcribing cylinder: folidity of the conoid NAM. 3 Let femitranfverfe CA = c, AP = x, PM = y, p = VI.) tt: ec : : 2tx + xx : 21x + xx. And pyy ptt t, femiconjugate CD 3.1416. Then (Prop. yy, and yy = tt X CC = X2tx + xx, for the cir CC cle NM. Therefore if AP be divided into an in- finite number of equal parts, according to the method of indivifibles; and planes drawn thro' all thefe points, all theſe circular planes will make up the folid. Now (Arith. Infinites Prop. II.) the fum of all the 2tx is - 2txx 2 = txx, and (ib. Prop. III.) the fem of all the xx is = O 2 ; there- 3 fore 150 THE HYPERBOLA. Fig. ptt 87. fore the fum of all the pyy = × txx + | cc 3 pttx CC I × tx + xx. And the circumfcribing cylin- 3 derpyyx = pttx сс CC' X 2txxx; therefore the cy- I xx linder to the folid :: 2txx + xx : tx + 3×× : : 2t + x: t + 3 x. I Cor. 1. The folidity of the hyperboloid to a cone of the fame baſe and altitude :: BP : BP + .AC. For the cone is of the cylinder; and it is, cy- linder: folid: 2t + x : t + x; therefore cone: folid 2t+x: 3t+x. : 3 I 3 Cor. 2. If AP = AC, the conoid NAM = the circumfcribing cylinder. 9 If AP = 3'AC, the conoid If AP = AB, the conoid = the cylinder. 12 2 the cylinder. 5 4 = 9 If AP = 2AB, the conoid the cylinder, &c. Cor. 3. The hyperbolic conoid NAM, is equal to the fruftum of the affymptotic cone EIAKF, leffened by the cylinder GIAKH. For (Prop. LXIX) the rectangle EMF = AK; therefore (Geom. IV. 38.) the annulus or ring contained between the two circles NM and EF is equal to the circle IK. Therefore if AP be divid- ed into an infinite number of equal parts; all the rings compofing the hollow folid NIAKM = all the circles GH, compofing the cylinder GIKH. Therefore, fruftum EIKF cylinder GIKH = fruftum EIKF - folid NIAKM'; that is, conoid NAM. hyp. Cor. B. II. THE HYPERBOLA. 151 Cor. 4. The hollow folid NIAKM inclofing, the Fig. cylinder GIKH. hyp, conoid is PROP. XC. Prob. To defcribe a hyperbola by continued motion. 1 Way. Draw the line DB, and let G be a fixed point; fix the two rulers together IK, KL, fo as to make the given angle IKL; let there be another ruler SL, which moves about the fixed point L in the leg KL, as a center. Then cauſe the fide KL to move along the line DB, and coincide with it; whilft the ruler SL flides by the fixed point (or pin) G. Then the lines SL and KI will crofs one another; and their interfection F, will defcribe the hyperbola FOG. For take SG FL, and draw SC parallel to IK meeting DB in C. Then G is a fixed point in the curve, and FK is parallel to SC; and GL paffes thro' L, fo that KL is always given. There- fore this is a property of the hyperbola, (by Prop. LXXIX.) whence the curve is a hyper- bola whofe affymptotes are SC, LC; and center C. 2 Way. 87. 88. Let FL, GH be affymptotes, AB a diameter. 89. Let AG be parallel to the affymptote FL; take HK = CG, make the angle CHM = GCF. Then move the right line HK along the affymp- tote HG, whilft the right line AM (moveable about A) paffes thro' the point K; then the in- terfection (of HM, AM), M will defcribe the hy- perbolas MB, Am. For the triangles KHM, KGA are fimilar; whence KH: HM :: KG: GA; that is, CG: HM :: CH GA; whence CH x HM = CG × GA; therefore (LXXV.) M is in the hyperbola. О3 SCHO- 152 THE HYPERBOLA: Fig. 89. 90. 91. 92. SCHOLIU M. The hyperbola may alſo be deſcribed as directed in Def. 1, and 2. PROP. XCI. Prob. To describe an hyperbole, by finding many points in the curve. I Way. Let BA be the axis; S, F the foci. Take any point N beyond F, and with extent AN, and one foot in F, defcribe an arch M. Then with extent NB, and one foot in S, croſs the former arch at M. Thus find as many points as neceffary, and draw the hyperbola thro' them. By this conftruction, SM-FM AB; there- fore (Prop. 1.) the point M is in the curve. 2 Way. Let AB, DG be two diameters conjugate to one another. Divide CA into any number of equal parts at L, L, &c. draw DA, and parallel to it, draw LI, LI, &c. draw CF CB, and perp. to it. Then thro' I, I, &c. draw MN, MN, &c. parallel to AB; and fo that IM or IN be equal to LF. Then M, M, &c. and N, N, &c. will be oppofite hyperbolas. 2 2 For by fimilar triangles CD: CA²:: CI: CL: (compounding) CD + CI: CA² + CL2 or IN² or IM. For CA + CL² = CF² + CL² LF2. Therefore M and N are in the curves, by Prop. XLIII. 3 Way. Let CA or CB be the femitranfverfe, CD the femiconjugate; produce CD at pleaſure, and draw DE B. II. 153 THE HYPERBOLA. DE parallel to CB; draw DA, and from any Fig. point P in the axis, draw PF parallel to AD, to 92. cut CD in F. With radius CF, and center C de- ſcribe a circle to cut DE in G. At P, draw the ordinate PM, perp. to CP, and equal to DG. Then M will be in the curve. For by fimilar triangles CA: CD:: CP2: CF2 (by divifion) CP2 CA²: CF² — CD2. But CP CA² BPA; and CF — CD² = CG2 CD²: CD = DGPM²; therefore CA: BPA PM. Whence (Prop. VI.) the point M is in the curve. PROP. XCII. Prob. Given the afſymptotes FG, MN, and the point B 93. in the curve, to defcribe the hyperbola. iWay. Thro' B draw feveral right lines to cut the af fymptotes, as ABN, DBĚ, GBL, &c. Then make NT AB, ES = DB, LI = GB, &c. Then the points T, S, I, are in the curve. Likewife thro' any point ſo found as T, you may draw other lines as HTQ, KTO, &c. and make PQ = HT, KT = OR, &c. then R, Q, &c. are alſo in the curve. This proceſs is plain from Prop. LXXII. 2 Way. Thro' the given point B, draw BA, BD paral- 94. lel to the affymptotes. Then take any length CG, and draw GH parallel to CD, and fo that CG: CA:: CD: GH; then H is in the curve, and fo for more points. I Or take the equal parts CA, AE, EG, GI, &c. And make EFAB, GH = ÷ AB, IK AB, and ſo on. Do the fame in the affymp- 04 3 tote 154 THE HYPERBOLA. Fig. tote CD; then thro' all the points f, B, F, H, K, 94. draw the hyperbola required. 95. 95. 96. PROP. XCIII. Prob. A right line DE in the hyperbola being given; to find its diameter, and the center. Draw HI parallel to DE; biffect DE, HI, in F and G; thro' which draw GFAB; then AB is the diameter. Biffect AB in C, then C is the center. For the diameter biffects all the whole ordinates, and is itſelf biffected in the center. PROP. XCIV. Prob. Any diameter AB in a hyperbola, being given; to find its conjugate. Draw any line KI parallel to AB, and biffect it in L; thro' L and the center C, draw LC, for the conjugate, which will be terminated by the conjugate hyperbolas, if they are there; otherwife its magnitude will be determined by Prop. XXXVI. PRO P. XCV. Prob. Any diameter AB being given; to draw an ordi- nate to it, from a given point H in the curve. Thro' the center C, draw the diameter HK; and from K, draw KL parallel to AB. From L draw LH for the double ordinate. For fince HC = CK, and CO is parallel to KL, therefore HO = OL. : Or thus, Draw HI parallel to AB, biffect IH in G; thro' G and the center C, draw GC, and HO parallel to it for the ordinate. For R N Fig.84 G B E S M P 90 M N 85 IM E D B A 86. C P D C P H D K 88. B A F F F D F B 89. 91. MMMB E B 92. 冬 ​n n n 95. D H 94. 93. D B B F G H K A E G D R NH K L PIXXII pa:154 UNIL OF MICH B. II: 155 THE HYPERBOLA. For GC, biffecting IH, is the pofition of the Fig. conjugate, and HO being parallel to it is an ordi- 96. nate to CO. : PRO P. XCVI. Prob. To find the two axes of a given hyperbola. 97. Draw two parallels thro' the figure, FK, ML; thro' the middle of them draw the diameter AB; biffect it in C for the center. On the center C defcribe an arch, of a circle, to cut the curve in two points G, H. Draw GH and biffect it in O. Thro' O and the center C, draw CO for the axis. And CD perp. to it for the conjugate. For AB paffing thro' C is a diameter, and fince CG CH; therefore CO paffing thro' the middle of GH, will be perp. to it, and being perp. to the ordinate GO, it is the axis. PROP. XCVII. Prob. A hyperbola being given; to find two conjugate dia- 98. meters, making a given angle with one another. About any diameter AB, defcribe the arch of a circle ADB, to contain an angle equal to the given one, or its fupplement. To the point F where it cuts the hyperbola, draw BF, AF; which biffect in O and L; and from the center C, draw CO, CHL. Then CH and CO (terminating at the conjugal hyperbola,) will be the two axes. For fince ALAF, and AC is parallel to BF. And fince FO= parallel to AF; whence LCO angle. 2 I 2 AB, CL FB, CO is OFL the given SCHOLIU M. and BE be If CE be the femiconjugate of AC, drawn and HG parallel to it, cutting CO in G; then 156 HYPERBOL A. THE Լ Fig. then CG is the conjugate, and G a point in the 98. conjugate hyperbola. For by Prop. LXXXIV. BE, GH being both parallel to the affymptote; CH, CG are conjugates, as well as CB, CE. 99. 100. ΙΟΙ. PRO P. XCVIII. Prob. Given the affymptotes and a point in the curve; to find two conjugate diameters. From the given point B, draw BHD parallel to the affymptote GC, make HD HB. = Draw DCE, and make CE = CD, then DE is the con- jugate of AB. For draw BF parallel to CD, then the triangles CDH, FBH are fimilar; and they are equal, be- caufe BH HD, and therefore CH = HF; and confequently GB = BF. And therefore FG is a tangent at B (LXXII. Cor. 2.). And fo FG is equal to the conjugate of CB (LXXXI.); whence DE equal and parallel to FG, is its conjugate. Cor. Hence if two conjugate diameters be given, and the angle they make with one another, be given; the affymptotes may be drawn. Which is done by drawing CG parallel to DB, and CF parallel to EB. PROP. XCIX. Prob. To draw a tangent to any point M in the hyperbola. I Way. Draw the femidiameter MC, and find its conju- gate CD (XCIV.); then from M, draw MT pa- rallel to CD, which will touch the curve in M. 2 Way. Let BA be any diameter, draw MP an ordinate to it; and make CP, CA, CT, continual propor tionals B. II. 157 THE HYPERBOLA. tionals; then draw MT which will be a tangent Fig at M. By Prop. XLII. 3 Way. Draw the lines MF, MS, from the point M to the focus's; make ME FF, draw FE, which biffect in D, then MD drawn will be a tangent at M. Or produce SM to G, ſo that MG = MF. Draw FG, and MT parallel to it, for the tangent at M. For in the ifoceles triangle EMF, ED = DF; therefore EMD = FMD, whence (Prop. IX.) MD is a tangent And fince MD is parallel to GF, and MF MG; therefore EMD = EGF, and EMF 2EGF, and EMF = EGF = EMD, therefore MD is a tangent. Ι 2 4 Way. 101. Let CH, CK be the affymptotes; from the point 102; M draw MD parallel to one affymptote CH. Make DFDC; and draw FM which will be a tangent at M. For produce FM to G; then fince FD = DC; therefore (by the parallels CG, DM) FM whence (LXXII. Cor. 2.) MF is a tangent. PROP. C. Prob. MG; From a given point without the hyperbola, to draw a tangent to it. I Way. Let T be the point given. Thro' T draw the 101. diameter BA. Make CT, CA, CP, continual proportionals. At P draw the ordinate PM, to the diameter AB. From M to T draw MT for the tangent. By Prop. XLII. 2 Way. Let T be the point given. Thro' T draw the diameter AB; and find its conjugate CD (XCIV); draw 103. 158 HYPERBOL A. THE Fig. draw any line TS- line TS TB, and make TR = TC; 103. draw AR, and SP parallel to it, cutting the dia- meter AB in P. Draw the ordinate PM parallel to CD; then TM is the tangent. 104. For by fimilar triangles TS: TR::TP: TA; that is, by conſtruction, TB: TC :: TP : TA. Therefore (XLII. Cor. 3.) TM is a tangent. PROP. CI. Prob. The focus F being given, and three points in the curve; to defcribe the hyperbola. Thro' the two points N, M, draw NMK, and alſo thro' N, P, draw NIP. Draw MF, NF, PF to the focus. And make FN: FM :: NK : MK, and FN : FP :: NI: PI. Thro' K and I, draw KG; draw FT, MH perp. to it. Take the points A, B, fo that FM: MH:: FA: AT:: BF: BT; then BA is the axis, and C the middle point, is the center. Make BS AF, and S is the other focus; and the hyperbola may be deſcribed by Prop. XCI. Draw NG, PL perp. to KT. Then by con- ftruction, FN: FM :: NK: MK :: (ſimilar trian- gles) NG: MH, whence FN: NG :: FM: MH. Alſo FN FP: : NI : PI :: NG: PL, whence FN NG :: FP: PL. Therefore FN NG :: : FP: PL:: FM: MH::(conftruction) FA : AT:: BF BT. And this (Prop. XXVI.) is the pro- : perty of the hyperbola, where TG is the direc- trix. For, fince AT: AF:: BT: BF:: (com- pounding) BT+AT. BF + AF:: BA: SF:: CA: CF; therefore (XXVI. Schol.) TG is the directrix. PROP. B. II. 159 THE HYPERBOL A. Fig. PRO P. CII. Prob. Given two conjugate diameters of a hyperbola; 105. CH, CL, to find the two axes. Join the ends of the diameters with the line LH, which biffect in O, draw the affymptote COI. Draw CF parallel to LH, and CF is the other af fymptote. Draw CP to biffect the angle HCF, and CQ perp. to it. Then the axes lie in the lines CP, CQ = Make IO CO, and thro' H, L, draw IK, IG. Alfo draw HP, LQ perp. to CP, CQ. And take CA a mean proportional between CK and CP; alfo CE a mean between CG, CQ. Then CA, CE are the femiaxes. For (LXXX.) fince LO OH, CO is one af- fymptote, and CF (parallel to LH) is the other; and fince CO OI, therefore (LXXII. Cor. 2.) IK, IG are tangents. And fince CK: CA: CP are in geometrical progreffion; therefore (Prop. XV.) CA is the axis, and it biffects the angle HCF. And fince CG: CE:: CE: CQ, and CE perp. to CA; CE is the other axis. PROP. CIII. Prob. To defcribe a hyperbola about a given parallelo- 106. gram EGHF, to pass thro' a given point P. Let the conjugate diameters AB, ID biffect the oppofite fides of the parallelogram; draw PQ pa- rallel to ID; and take AC² CQ² × EL² CL² × PQ² EL-PQ €Q¹× EL”—CL³× PQ² CL² ~ CQ- And CD2 = For 160 HYPERBOLA. THE Fig. For (Prop. XXXVI.) CA² : CD³ :: ALB : EL^: ; 106. AQB: PQ. That is, AC: CD :: CL-CA2: 2 EL²:: CQ - CA: PQ. Hence CL2- CA2 × PQ² = CQ² CA² x EL*; therefore CA² × 2. 2 EL² — PQ² = CQ2 × EL - CL× PQ, which CQ² determines AC. Again, fince AC: CD:: CD² 24 CL-CA: EL; by compounding, AC: CD :: CL: CD + EL. And fince CL CA: EL :: CQ² CA² : PQ2; and (by divifion) CL2 CA: EL:: CL² CQ: EL² - PQ. Therefore CL: CD + EL²:: CL-CQ² : EL² - PQ. And (dividing) CL² : CD² + EL²::CQ²: CD² + PQ. Whence CLª × CD² + PQ² = CQ² × CD² + EL²; and CD² x CL² - CQ² = = CQ² × EL CL² × PQ' 2 which determines CD. 2 Cor. AC²: CD :: CL-CQ² : EL² — PQ². For AC x EL-PQ² = CQ² × EL² — CL × PQ = CD² × CL² — €Q². PROP. CIV. 107. If a cone MDN be cut by a plane, parallel to any plane DXY within the cone, which paſſes thro' the vertex D; the fection LAT will be an hyperbola. Let the plane produced cut the oppofite cone BDp in Btl, biffect AB in C, and parallel to the bafe MN, draw PTQ, CGKF, ptq, whofe fections, with the cone will be circles; and the interfections with the plane of the hyperbola, will be LT, lt. Draw PQ, pq perp. to LT, it; and CK a tangent to the circle GKF, then the rectangle PIQ = LI², and rectangle piqli, and rectangle GCF = CK'. The plane PDQ interfecting the other planes, will B G Fig. 96. E 102. 13 M F L 97. F H 99. H 105. H B H ? C P { F I 0 F 100/ 103 T 98. E H D E 101. M D SB P M Gr 104. B TAF R H XM 106. E I P D BIQ K PI.XXIII:pa.160. UNIV OF ル ​FICH B. II. 161 THE HYPERBOLA. will make the fimilar triangles AIP, ACG; and Fig. alfo BIQ, BCF. Whence AI: IP :: AC : CG and BI: IQ :: BC: CF. And multiplying, AI X BI: PI × QI:: AC x BC: CG x CF. that is, AIB LI² :: AC²: CK². Therefore (Prop. XXXVI.) the fection LAT is a hyperbola. Cor. 1. The fame plane cutting the oppofite cone, the fection will be the oppofite hyperbola tВl, equal and fimilar to TAL. For Ai ip: AC: CG, and Bi: iq:: BC: CF. Whence AiB : piq:: AC× BC: CG × CF; that is, AiB : il² :: AC²: CK²; therefore when Bi = AI; then li LI, and the curves are both alike. Cor. 2. The fquare of the conjugate, 4CK² = BH X AE, the rectangle of the diameters of the fruftum. For the circles BH, AE being parallel to the bafe; fince AC AB, therefore CG I 2 I 2 BH, and CF AE; therefore CK' or GC x CF = = BHX AE. 2 Cor. 3. The latus rectum = BH X AE AB × For it is a third proportional to the tranſverſe and conjugate. Cor. 4. The hyperbola is right angled when CK´= CA, or when BH × AE = AB². 107. Cor. 5. Thro' the vertex of the cone D, draw the 108. plane DXY parallel to the plane of the hyperbola ALT; and thro' X and Y, draw the planes DMX, DMY, touching the cone in the lines DX, DY; and interfecting the plane of the hyperbola in the lines NC, OCo; then theſe lines NCn, OCo, are the asymptotes of the oppofite hyperbolas LAT, IBt; and YO or XN, the femiconjugate; and the angle NCO = an- gle XDY; and MD paſſes thro' the center C. For 162 THE HYPERBOLA. Fig. For thro' A draw the circle AGES parallel 108. to the baſe of the cone, which will touch the plane of the hyperbola in A. This will cut the planes DMN, DMO, in the lines sr, Gw. And then the lines sr, rA, Aw, wG are all tangents to the circle; therefore rsrA, and AwwG. Like- wife rA = Aw, becauſe the parallels thereto NI = IO. And fince sr, Gw, are parallel to XN, YO; therefore XNsrrA AwwG = YO; and therefore XN² or rA² rectangle LNT, by the property of the circle; and confequently (LXXI. Cor. 1.) CN, CO, are affymptotes; and (LXXXI.) rw is the conjugate, and rA YO is the femiconjugate. 109. = XN or Draw PK parallel to AB; then by Prop. XVI. Cor. 2. ellipfis, which alfo agrees with the circle, QZ : PZ : : QM: PM :: (by diviſion) QM QZ or MZ: PM-PZ. And alternately, QZ : MZ:: PZ: PM - PZ:: (by addition) QZ + PZ or PQ: (MZ + PMPZ) 2PM. And al- ternately, QZ: PQ:: MZ: 2PM. But by fimi- lar triangles, QZ: PQ:: ZD: PK. and MZ: PM:: ZD: PF. or MZ: 2PM:: ZD: 2PF. ▾ whence ZD: PK :: ZD: 2PF. Therefore PK 2PF, whence KP is biffected in F, and confequently AB (parallel to it) is biffected in C, and fo C is the center. Laftly, fince CN, CO, are parallel to DX, DY; the angle NCO = angle XDY. Cor. 6. If AI be perp. to the baſe of the cone, and AB perp. to MH the tranfverfe of the hyperbo- la. And if AN AB, and FND be drawn parallel to the bafe PQ. Then FD is equal to the latus rec- tum. And DS (parallel to MH) = HZ the dia- meter at the vertex H. Draw B. II. 163 THE HYPERBOLA. Draw AL, AG parallel to PQ, MH, and put Fig. R latus rectum. Then the triangles ABL, 109. ANG, are ſimilar and equal, whence AG = AL. The triangles MCH, DFS, GFA, are fimilar; whence MC: MH:: DF: DS:: GF GA or AL: (by the ſimilar triangles GFA, LAH) AF : AH (triangles AFD, AHZ fimilar) FD: HZ. Whence DS HZ. And MH x FD - MC x X HZ (Cor. 2.) fquare of the conjugate Schol.) MH x R; therefore FD = R. (IV. Cor. 7. In a right cone, the Square of the tranf- 110. verfe BA BE AEX BH, the Square of the fide of the fruftum lefs the rectangle of the diame- ters of the bafes. Let BF be perp. to the baſe AE, then FE = BH+AE BE2 2 ; and (Geom. II. 23.) AB² = AE² + 2AE X BH + AE 2 -BE-AEX BH. Cor. 8. In a right cone, the Square of the fide of the fruftum is equal to the fum of the fquares of the tranſverſe and conjugate, BE² — AB² + ÂE² + BH². Cor. 9. In a right cone, the distance of the focus from the center, is equal to half the fide of the fruf- tum, BE. I 2 For if S be the focus, O the center; C half the conjugate. Then (Prop. II.) OS = OB² + C² = AB² + 4 AEX BH 2 I I BE or AH. 2 (Cor. 8.) BE², and OS Cor. 10. In a right cone, if F be the focus, C III. the center, and AG AF. The circle GOI paral- lel to the base of the cone RS, will cut the plane of the hyperbola FAH in the line ED the directrix of the hyperbola. In the plane of the hyperbola, let FH, AL, be perp. to AF, and HD a tangent at H; then (XXVI. P and 164 THE HYPERBOLA. Fig. and Sch.) D is the point thro' which the directrix III. paffes. But (XXV. Cor. 2) AL AF II2. AG by conſtruction. Alſo (Cor. 9.) CF Alfo (Cor. 9.) CF APAN; = and therefore CP drawn thro' the middle of AN, AM will be parallel to MN or to RS the baſe of the cone. Now (Prop. XV.) CF: CA :: CA: CD:: (by divifion) AF: AD; that is, AP: CA:: AG: AD; therefore DG is parallel to CP, or the baſe of the cone. And on the contrary, if DG is parallel to RS, DE will be the directrix. Cor. 11. In a right cone, if F be the focus, and AG = AF, and the circle GI be drawn parallel to the baſe of the cone; then the distance of any point P from the focus as FP, will be equal to the distance of P from the periphery of the circle GI. = Let PQ be an ordinate to AQ; draw QH, QS parallel to AG, DG. Then will (Geom. V. 12. Cor.) QHSG diftance of P from the peri- phery of the circle GI; becauſe the plane SQP is parallel to the circle GI. By Cor. 10, DE is the directrix; therefore if DR be the focal tangent, and if the ordinate QP be continued to the tangent at R; then (Prop. XXV.) QR = FP, and AĽ = AF AG. And by fimilar triangles, AL: QR:: = (DA : DQ : :) AG : QH; that is, AG: FP :: AG: SG; therefore FPSG. SCHOLIU M. Hence we have the reaſon why the Hyperbola is called one of the Conic Sections. For if two op- pofite cones be cut by any plane, it will always make the figure of an hyperbola in both the cones, as is demonſtrated in this propofition. The pro- perties of which have been demonftrated at large in this Book. CONIC M P P Fig. 107 B H 109. H B I E > 112. E S M K H S E K B 108. C N F E A P S 110. M B H F C P 111. D E H R n 2 PLXXV.pa 164. pa.164. UN OF [165] CONIC SECTIONS. BOOK III. Of the PARABOLA. I DEFINITION S. DEFIN. I. 1. F one end of a thread equal in length to CH, Fig. be fixed at the point F, and the other end fixed at H, the end of the fquare DCH. And if the fide CD of the ſquare be moved along the right line BD, and always coincides with it. Then if the ſtring FGH be always kept tight, and cloſe to the fide GH of the fquare. The point or pin G (where it leaves the fquare) will defcribe a curve MRALGK called a Parabola. DE F. II. The fixed point F, is called the Focus. DE F. III. The right line BD, is called the Directrix. DE F. IV. If the line BN be drawn thro' the focus F, perpendicular to BD; then AN is called the Axis of the Parabola; and A the Vertex. P 2 DEF. 166 THE PARABOL A. Fig. I. 2. 3. DE F. V. A line drawn thro' the focus F, perpendicular to the axis, as LR is called the Parameter or Latus Rectum. DE F. VI. Any line drawn within the curve, parallel to the axis, as GH is called a Diameter. And the point G where it cuts the curve, is the Vertex. DE F. VII.` A right line drawn from any diameter to the curve, and parallel to the tangent at the vertex, as PM, is called an Ordinate. If it go quite thro' the curve, it is called a double Ordinate. DE F. VIII. The part of the diameter between the vertex and ordinate, as GP is called the Abfciffa. DE F. IX. A right line meeting the curve in one point G, but does not cut it, is called a Tangent in that point. PROP. I. If BD be the directrix, G any point in the curve, the line GD drawn to the directrix, parallel to the axis, is equal to the line GF drawn from the fame. point G to the focus; GD GF. For HG + GF = length of the ftring = HD, take away GH from both, and then GD GF. Cor. 1. The distances of the focus, and of the di- rectrix from the vertex, are equal. AB AF. For when D is at B, G will be at A; confe- quently AB AF. Cor. B. III. 167 THE PARABOLA. Cor. 2. If GP be an ordinate to the axis; then Fig. AP + AF FG. 3. For AP + AF BPGD. Cor. 3. FGFP — FP half the latus rectum. PROP. II. I The distance of the focus from the vertex is the 1. latus rectum : AF = 4 LR = ½ LF. I I For when the pin G comes to L, then LF = FB (Prop. I. Cor. 1.) 2FA, and AFFL. For the fame reafon FA FR, therefore FA = = LR. 1 == 2 SCHOLI U M. As the latus rectum to the axis is four times the diſtance of the vertex A from the focus F. So in any other diameter GH, four times the diſtance of its vertex from the focus, or AFG is called its Latus Rectum. PROP III. The Square of any ordinate to the axis, is equal to the rectangle of the latus rectum and abfciffa PM" = LR X AP. For (I. Cor. 2.) MF = AF + AP = (Prop. II.) AP+LR, and FP AP AF AP = I 4 LR. And in the right angled triangle MFP, MP² MF2-FP MF +FP X MF-FP = 2AP X LR AP XLR. = 2 = Cor. 1. If F is the focus, MP AP X 4AF. Cor. 2. The abfciffas are as the fquares of their ordinates. AP: AQ :: PM² : QN². P 3 For 4. 168 THE PARABOL A. Fig. 4. 5. 5. For AP AQ:: AP x LR: AQ x LR:: PM²: QN². 2 Cor. 3. The latus rectum is a third proportional to the abfciffa and ordinate. AP: PM: LR PRO P. IV. As the latus rectum to the fum of any two ordi- nates : : fo their difference, to the difference of the abfciffa. Lat. rect.: CD :: ND : PQ. Let L latus rectum, then (Prop. III.) LX AP PM²; and L x AQ = NQ². And by fubtraction, L x AQ-L x APNQ-PM2; therefore L : NQ + PM : : NQ — PM : AQ- AP; that is, L: DC :: ND : PQ. Cor. 1. If MD be the axis, NC an ordinate to it. Then the rectangle NDC MD x parameter. Cor. 2. The rectangle NDC is every where as MD. PROP.. V. If a line AN be drawn from the vertex to cut the curve in M, and the ordinate QO in N; and the ordinate PM be drawn. Then the rectangle PM × QN = AQ × parameter. Let L latus rectum, then (Prop. III) QO² = LXAQ; and (III. Cor. 2.) PM: QO or L X AQ:: AP AQ:: (fimilar triangles) PM: QN; and PM² × QN = PM X L X AQ, or PM x × QN = LX AQ. Cor. 1. PM X QN QO². Cor. 2. Any line befides the axis, paſſing thro' the vertex A, will cut the curve again in fome point M. For I B. III. 169 THE PARABO L A. For QN QA latus rectum : PM the ordi- Fig. nate where it cuts the curve. PROP. VI. 6. If from any point C in the axis, the line CF be 6. drawn, cutting the curve in D and F, and the or- dinates DB, FQ be drawn. Then CA² = ABX AQ. For (III. Cor. 2.) AB: AQ : : BD² : QF² : : (fimilar triangles) CA + AB2: CA+AQ; whence AB × CA² + 2CA × AQ + AQ² = AQ × CA² + 2CA × AB + AB³, and AB× CA² + AQ² = AQ × CA² + AB²; whence AQ — AB× CĄ = AQ² × AB — AB² × AQ = AQ — AB× AQ × AB. Therefore CA AQ× AB. 2 PRO P. VII. If AB be perpen. to the axis AE, EG an ordi- 7. nate, and AG drawn. Then any line BF being drawn parallel to the axis AE, cutting the curve in M, and AG in E. Then BM, BD, BF are continually pro- portional. For (III. Cor. 2.) AP: AE :: PM² or CD² : EG (fimilar triangles) AC AE; therefore AP: AC AC: AE; that is, BM: BD:: BD: . BF. PRO P. VIII. If from any point M in the curve, MF be drawn 8. to the focus, and MD be drawn parallel to the axis cutting the directrix BD in D; and DF be drawn. Then the line MG drawn thro' the middle of DF will touch the curve in M. From any other point m in the line MG, draw mF to the focus, and md perp. to BD. Then fince P 4 MF = ני 170 PARABOLA. THE Fig. MF MD, and mFmD, the triangles FMD and 8. FmD are ifoceles, and FG GD; but the angle Ddm being right Dm or mF is greater than dm; therefore the point m, by the generation of the pa- rabola, is without the curve. 9. 1 1 Cor. 1. If the tangent MD be continued to cut the axis at T; the distance of the focus and interſec- tion, is equal to the distance of the focus and point of contact. FT FM. For MD being parallel to FT, the < GTF = GMD GMF; therefore FT FM. Cor. 2. The tangent MG biffels the angle DMF made by the line perp. to the direttrix, and that drawn to the focus. í Cor. 3. A line drawn from any point M to the focus, and another parallel to the axis, make equal angles with the curve; FMm OMr. PROP. IX. If the tangent at M, cut the axis AB at T; then the distance of the interfection from the vertex, AT AP, the abfciffa to the ordinate PM. From T draw any line TF cutting the curve in D and F, and draw the ordinates DB, FQ; then (Prop. VI.) TA' AB X AQ. TA² Now let the points D, F, approach to M, and coincide with it; then B, Q coincide with P, and TDF becomes the tangent TM; in which cafe TA2 APX AP = AP2, and TA = AP. TA² Cor. 1. The fubtangent PT is double the abſciſſa AP. Cor. 2. The tangent at the vertex A, is perp. to the axis AB. For, B. III: 171 THE PARABOL A. For, from the generation, the ordinates are perp. Fig. to the axis, or parallel to the directrix. Cor. 3. All parabolas having the fame abfciffa, will have their tangents meeting in the fame point of the axis. For if AP be given, AT equal to it, will be given for them all. PROP. X. 9. If two parabolas be defcribed on the fame axis; 10. the ordinates will be proportional, or in a given ratio to one another. For let the common abfciffa be AB, the ordi- nates BC, BD. 1, L the parameters. Then (Prop. III.) 7× AB = BC², and L × AB = BD²; there- fore BC BD :: √l× AB: √L × AB :: √/l: VL, which is a given ratio. : Cor. The correfpondent ordinates of two parabolas defcribed on the fame axis, are in the fubduplicate ratio of the parameters. PROP. XI. If F be the focus, MO perp. to the tangent MT; 11. then the fubnormal (or distance of the ordinate from the point of interfection) PO= the latus rectum. 1 2 Let L latus rectum, then (IX. Cor. 1.) PT = 2AP; and by firnilar triangles TP or 2AP: PM :: PM: PO, and 2 APX PO PM (Prop. III.) LX AP; therefore 2PO L or PO = L. Cor. 1. If F be the focus, gent MT: Then FT MO perp. to the tan- FM- FO. For (Prop. II.) AFL, and (I. Cor. 2.) FM AP+ AF AP+L, and FP AP- = = = AF= 172 PARABOL A. THE Fig. AF AP — & + L. And FO AP —÷L÷ 11. ÷ L = AP + L = FM = (VIII. Cor. 1.) FT. Cor. 2. A circle defcribed with the radius FM, and center F, will pass thro' T, M, O. Cor. 3. The angle MFO is double the angle MTO. Cor. 4. If OG be perp. to MF; then MG = half the latus rectum. For fince FMFO, the < FMO = the ra- Cor. Sect. II. 91 CURVE LINES. 87. Cor. If the ordinates make not a right angle with Fig. the diameter, but any oblique angle GFH, put s = fine of the angle GFH, AF = x, FGy, R = radius of curvature; then instead of y in the equa- tion of the curve, put √2Rsx, and rejecting the in finitely small quantities as before, and reducing the equation, R will be found. For let AFH be the diameter of the curve, AC the radius of curvature, ABD a tangent at A, GB parallel to the abfciffa AF, GD perp. to AD; and let GD = v. Then if the arch AG be fup- pofed infinitely fmall, AG = GF, and the part AG of the curve will coincide with the circle. But by the nature of the circle, AG² = DG × AL; that is, yy = 2Rv. But in the infinitely fmall triangle DGB, DG (v): GB or AF (x):: s.DBG or GFH (s): radius (1); therefore vsx; whence yy2Rv 2Rsx, and y = 2Rsx. Example. Let DGA be a parabola, HAF a diameter, FG 88. an ordinare, the latus rectum, of the diameter HF; then if AF = x, FGy, ss. angle GFH, we ſhall have lxyy 2Rsx, by the rule, whence R = This may be denoted otherwiſe, 25 for Is :: AT : AP, and s = AP AT √4DP² + PA, and if a = la AP AP ✔TP² + PA² PA² AP tus rectum of the axis TP, s = √4DP² + a × DP DP X a a 4DP² + a × DP a + 4DP a 4 I H 8 + DP That is, if S be the focus, s 92 B. II. CURVE LINES. ! I DS. Whence the radius of curvature in A SA == 1× √ SA √SA 2 DS the nature of the curve. 4SA 4DS by a za, 1/ Fig. 88. And if P falls upon S, then PA = ½ a, and =a, whence Ra/2. PROP. XX. The Shadow of any given curve line, projected from a lucid point, upon an infinite plane, will be a curve of the fame gender, with the given one. For the fhadow of any point will be in the right line drawn thro' the light and that point. And any right line in the given curve is projected into a right line upon the plane. For imagine a plane drawn thro' the light, and thro' a right line in the given curve, and that this plane is extended to the plane of the fhadow, it will cut this plane in a right line, which will be the fhadow of the given line. Therefore right lines that interfect one ano- ther will be projected into lines or fhadows that interfect one another. If any line cuts the curve, its fhadow will cut the fhadow of the curve. If any line touches the curve, the fhadows of the line and curve will touch one another. If one part of the curve interfect another, their fhadows will alfo interfect. So that all points that are coincident in the figure, will be coincident in the fhadow. And all points at a diſtance in the figure, will be diftant from one another in the fhadow. Now let a right line be drawn in the figure to cut it in as many points, as is the dimenfion of the figure; and it cannot cut it in more. Then the fhadow of this line will alſo be a right line, and will m Fig. 77. M : 78. M 2 A 80. A M F B P m M 82. G B R N I M 85. : A BA G M A P D 86. c M B 79. D M B. P A P m 81. F M 772 83. M 84 H DB A D B 87 S P C F M G Hy L 88. F с m PIVII. p.92. N OF BICH Sect. II. 93 CURVE LINE Ś. will cut the ſhadow of the curve in as many points, Fig. and no more. Therefore this fhadow is a curve of the ſame dimenfion, or of the fame gender with that given. Cor. 1. Hence the fhadow of the circle, or any conic fection, will be a conic fection. The fhadow of lines of the third order, will be lines of the third or- der, and fo on ad infinitum. Cor. 2. If any part of the curve be higher (og further from the plane) than the light is; that part of the curve can cast no fhadow upon the plane ; and the shadow of the curve will run on infinitely to- wards that hand. And therefore whatever there is in that part of the curve above the light, will be quite loft in the fhadow, and no where be found. Bþ SECT. 94 B. II. CURVE LINES. SE C T. III. Some fluxionary Propofitions neceffary for inveftigating the nature of Curves. Fig. H 89. AVING handled feveral things of this na- ture in my book of Fluxions, I intend not to repeat them here; but fhall only take notice of fuch as have been omitted there, or but flightly mentioned. And this I the rather do, becauſe many properties of curves of high genders are far more eafily inveſtigated by the method of fluxions, than by common algebra. PROP. I. Let AM be any curve, AP the abfciffa, PM an ordinate perp. to it, MC the radius of curvature in M, CF parallel to AP meeting the ordinate MPF in F. Then the fecond fluxion of the ordinate, the fluxion of the curve, and the line MT, are continu- ally proportional; fuppofing the fluxion of the axis to be constant. Let M, m, r be three points in the curve, M. and ms parallel to AP. Produce Mm to n, then fince Momr, therefore Mm mn; then rn= ≈ fluxion of om, or the fecond fluxion of the ordi- nate. Draw mF, then the triangles mrn, MFm are fimilar, for nrm alternate angle mMF, and mnr Mmo MmF; and therefore nr: rm :: Mm: MF, but rm Mm, therefore nr: Mm :: Mm : MF. - < - Cor. Sect. III. 95 CURVE LINE S. Cor. 1. Hence, the fluxion of the axis Mo to Fig. fluxion of the curve Mm:: MF: MC, the radius of 89. curvature in M. Cor. 2. Hence the radius of curvature in any curve is easily found; by first finding MF by this Prop. And then MC by Cor. 1. Example. Let AP, PM y, AM, and let the equation of the curve be 2ax + xxzz, this put into fluxions, is 2ax + 2xx2zż, or ax + xx =zz; then z = x², and ÿì 2 a + x ZZ 2ax + xx a + x *, and x² + y² = x² Z a + x a + x x² x² 2 ZZ 2ax + xx 2ax + xx * Then ÿ= aax² 2ax + xx ax × ax + xx ax × a + x x x 2ax + xx and y ax √2ax+xx Then Min A X a + x (j): ·* (*) :: 2ax + xx ax x x a + 2 2ax + xx 2 a + x x ✰ X ZZ a + x Then : १२. a + x a *) :: MF the radius of curvature. And in the vertex where x is 0, the radius of curvaturea. This curve is called the Catenary. Z Z * (*): 3 M'N a + x MF Z. a a + x Z a x): MC = Rb 2 PROP C 96 CURVE LINES. B. II. Fig. 90. Ey I. PRO P. II. If PF, QD, RC be three equidiftant ordinates in any curve and infinitely near; and nDb be a tangent at D. Then the fubtenfe nF or bC, will be equal to half the fecond fluxion of the ordinate. Db, and the For draw FDa; then fince nD fubtenfes nF, b℃ parallel, and the curvature FDC the fame throughout; therefore nFbC. Again, in the triangles nDF, bDa, fince nD = Dỗ, and all the angles equal; therefore nF = ab = bc. C. Therefore 26C aC, and bC = aC. Let FL and DO be parallel to and let PFy, then LDỳ OC or aC = = I 2 the axis PR, and I Oa, and Oa -- ÿ, and bC or nF = ½ aC = ½ ÿ. 2 Cor. If PF, QD, RC be three equidiftant ordi- nates, and FD be drawn, cutting RC in a. Then Ca fecond fluxion of the ordinate. For aC 2bC=j, by the demonſtration. PROP. III. In any curve whofe ordinates are drawn from a fixed point C. Let CM, Cm, Cn be three ordinates infinitely near, fo that the perpendiculars Mo, mp may be equal. And if Mm be drawn interfecting Cn And if Cmy, mp = x, mn z. Then in s. the line ns = ÿ ZZ y For draw mr between s and C, making the an- gle smr = mCn; then the triangles Cmn, mrs will be fimilar, for the angles at m and C are equal by conftruction, and msC or msr = mnC; therefore Cm Sect. III. 97 CURVE LINE S. Cm (y): mr (ż) : : mr (ż) : rs = ZZ y Fig. Draw mq 91. parallel to Mo. Then the angle pmq = mCn = smr, whence the angle rmp smq=mMo, and Momp. Therefore the triangles Mmo, mrp are fimilar and equal, whence pr = mo. Therefore ÿ = np mo = np rs; therefore sn = ÿÿ — rp = rn. ZZ y But snrn Cor. In any curve whofe ordinates are referred to a fixed point; if CM, Cm, be two ordinates, and mt a tangent at m; then the fubtence Mt = yÿ 2y ZZ For Mt is half of sn = 2 N ZZ ÿ 2y PROP. IV, If ŷ, ÿ, ÿ, &c. denote the first, second, third, 92. &c. moment, increment, or fluxion of y; and if P, Q, R, S, T, &c. be feveral equidiftant ordinates in a curve, being infinitely near one another. And let the first P be called y. Then will y = ÿ q = ÿ + ÿj ÿ r = ÿ + 2ÿ + ✯ متر s =ÿ + 3ÿ + 3ÿ + ÿÿ &c: For by the nature of curvature q = p + p ; that is, q = y + ÿ, r = q + q, s = r + r, &c. Therefore taking the fluxions of thefe equations, Bb3 98 B. II. CURVE LINES. Fig. 92. q =ÿ + j, r = q + q, &c. Whencer = q + q s j = r + ¿, q = x, j + j = ÿ + ÿ + ÿ + ÿ +ÿ+ ÿ =ÿ = j + 2ÿ + Ÿ‚ s = r + ï = †ÿ, q + q + q + q = q + + 2ÿ + ÿ + ÿ = j + . 2q + q = ÿ + ÿÿ + 2ÿ †ÿ 3ÿ + 3ÿ †ÿ, &c. Cor. Hence the fluxion of the nth ordinate P, Q₂ R, S, &c. is j + n − 1 •ÿ + n I . N - 2 IN n I. N 2. n 3 + j + &c. 2.3 V 93: 94. 93. 94. PROP. Let M be the point of contrary flexion in any curve AM; and if curve AM = z, ordinate PM or CM = y, Mox. Then in curves referred to an axis, ÿ = o, and in fpirals yÿ = ", fuppofing & Z", x to be given. For at M the point of contrary flexion, three points n, M, m, infinitely near one another, lie in a right line. Therefore in curves referred to an axis, if Mo or x be given, om or y will be given, confequently ÿ= 0. And in fpirals let CT be perp. to the tangent MT; then fince TM is a tangent at all the points, n, M, m. Therefore at M, CT is a conftant quantity; let CT=t, then by fimilar triangles yx ż : * : : y : CT ==t, and yx t, and yx = tà, in fluxions, jx = tz, x and t being conftant quanti- y x ties. Therefore yx, and jxz = yż, But Sect. III. 99 CURVE LINE S.. ż² = x² + y²; therefore żż = ÿÿ, and z = Fig. ÿÿ Z Therefore yxż = yx × yy yÿ and z = Z and 94. ZZ żż = yÿ. у. Cor. Hence to find the point of contrary flexion, let x be given, and put ÿjo, in curves referred to an axis; or ÿ 11 22 in fpirals. y Ex. I. Let the equation of the curve b3 × ya 11 -5 ; 2. X a then X a; y - a = 63 and y a जात a X a whence y = ½ × 5 b√b 2 15 लाल X ba/b a b√b 2 and ya+ xx, and ÿ ÿ = therefore x-a forex = a. 142 46√ 13-14 XX a x x x = 0; = o, and x a = 0. There- Ex. 2. Let ABM be a fort of ſpiral, RFB its gene- 95. rating circle, radius CD, CM y, arch RD = v. And let the nature of it be denoted by the equation vyy a³. The ſectors CDd, and CMo are fimilar, whence CM (y) : Mo (x) : : CD (r) : Dd by the equation of the curve yy rx = = v, and y + 2vÿÿ = 0, zvý rx rx and v = y , and y = = (be- 20 ryyx cauſe v = vy 20773 > whence ÿ = Bb 4 - ryyx Q3 and , дзё 100 . B. II: CURVE LINESLINE S. Fig. 95. yÿ = - ry¹ýx goz y 4 x ² a³ 296 Alſo ż² = x² + ÿ² = x² rry4x2 + Whence r²y4x² 406 296 = x² + rry4x² and 490 27²y4 = 4a+y4. And r2y4 4a6, and ry² = 2a3; whence y = √√√2a³. And if r = a, then y = r√2. PROP. VI. If x, y, z, &c. repreſent any right lines; and in any curve, if x × y × %, &c. = à given quantity ; x Y then + y X of Z &c. = 0. 22 / 22 ä For let xyz= a, then the fluxion will be zyć +xzỷ + xyż = o ; divide by the given equation; -o; Zyx x Zỷ XYZ + then + 1.8 O X ; that is, + ZYX XZY XJZ XYZ y + z' &c. = 9. Cor. If xyz &c. a, a given quantity; then + my NZ +- $30. y Z 0. & ગેસ PROP. Sect. III. CURVE LINE S. ΙΟΙ 1 PROP. VII. If x, y, z, v, u, w, lines, and in any curve if quantity; then W W &c. = 0. | Fig. &c. reprefent any right 95, xyz &c. vuw &c. + + 22 / 22 &c. Pla C any given И น xyz Let =a, then XYZ + VU W บนซ xyzv xyzu xyzw v²uw vu²w VUW² xyz xyż + บนซ VUW &c.o. And dividing all by xyx X ż ข a, we fhall have + + VU W X W &c. = 0. น 70 Cor. If xym &c. x then { vur &c. my + &c. - X D nů ཝ །© &c. = 0. www PROP. 102 CURVE LINES. B. II. Fig. 96. PROP. VIII. And If from a given point P the right lines PA, Pa, be drawn, cutting any curve in as many points A, B, C, and a, b, c. as it has dimenfions. the tangents AK, BL, CM; and ak, bl, cm, be drawn, cutting any line PE, in the points K, L, M, I I I and k, l, m. Then will PK + PL + PM &c. = I Pk I I +Pl+Pm &c. wherever the points A, B, C, and a, b, c, fall. Let AC, ac be fuppofed to move parallel to themſelves, and then the point of interfection P will move along the right line PE. Then fince (Prop. XV. Sect. II.) PA x PB x PC:: Pa x Pb x Pc f : 1, a given ratio. Let APx, : : PBy, PC z, and PA = v, Pb = u, Pc = w; then xyz=fvuw, or (Prop. VII.) and + + 18 al X y + & IN. Z Ў Y + · 22 | 22 xyz f. Therefore VUW И W O, V И W W + + But fince ข U W 12 AP is moved with a parallel motion, it is evident the fluxion of AP is to the fluxion of EP, as AP to the fubtangent PK; whence, putting EPs, x we have and for the fame reaſon Ў 11 X PK' J PL' 11, and 24 Z ; alfo PM S И W ט Pk u Pľ W S S S Whence + + + Pm PK PL PM Pk Pl Sect. III. 103 CURVE LINE S. s I I I I == Pk P/ Fig. +P; and PK + PL + PM = 2+1+96 Ι Pm Pm³ But if any of the points K, L, M, &c. fall on the contrary fide of P, then the correfpondent lines PK, PL, PM, &c. muſt be made negative. I I Cor. 1. If any line PC cut the curve in the points A, B, C, and tangents drawn; then will PK + PL + I PM &c, be always equal to an invariable quantity. Cor. 2. And bence if a right line PE cut the curve in D, E, I, &c. as many points as it has di- menfions; then will I PI I + PE I PK I I I + PL + PM = PD + For in this cafe k, l, m, coincide with D, I, E. Cor. 3. If any of the points K, L, M, &c. fall on the contrary fide of P; you must take I PL' I &c. PM' I PK' For lying the contrary way, they must have con- trary figns. PROP. 104 CURVE LINE S. B. II, Fig. PRO P. IX. 97. If the line AB be drawn thro' a given point P parallel to the affymptote CM; and the tangents AK, BL drawn, cutting any line DE (drawn alfo thro' P, and cutting the curve in three points D, I, I I E; and the affymptote in M). Then = + I PM PI 98. I 1 I PE PD PK - PL I Ι I I For (by Prop. VII.) PM + PK + PL = PI+ PE I PD? becauſe D lies the contrary way from P, and AB meets the curve and affymptote CM at an infinite diftance where they touch. Then by I I I I I I PD PK PK-PL tranſpoſing, PM =PI+ PE-PD Cor. Hence the distance of the afymptote may be found, by finding the point M. And if I PM be = 0, then the affymptote is at an infinite diftance, and the curve will be of the parabolical kind. PROP, X. To find the affymptotes of a curve. This Prob. may be folved by the laſt Prop. and its Corol, or otherwife as follows. Let APx, PM y, TM a tangent at M, and TP the fub- tangent. Find the value of the fubtangent yx за by the equation of the curve, then putting x = infinity in the value of the fubtangent, and alfo in the Sect. III. CURVE 105 LINE S. the equation of the curve, and you will get TP Fig. and PM, which gives the poſition of the affymp- 98. tote TM, and TP-x, gives the interfection T. For the triangles Mrm and TPM are fimilar, y y x and rm (j) : Mr(x)::MP():TP = And fuppofing M at an infinite diftance x will be infi- nite, from whence both TP and PM will be had, and confequently the pofition of TM the affymptote. Ex. I. Let the equation of the curve DM be xyy = @3; in fluxions y²+ 2xyy 2xyy yy 2 o, and ✰ 2xy y y x Therefore PT= 2X. У And PM } = a3 ✓ Then making X Then making x infinite, PT = — 2× = — infinity, and PM = 2X Ja X o; therefore the affymptote coincides with the axis TP. Ex. 2. Let aayy2abbx + bbxx. 2abbx+2bbxx, and x = yx 12abbx+bb xx aa In fluxions 2aayý aayy there- abb + bbx aayy abb bbx and , y = Make x infinite, then PT = 2abbx+bbxx fore PT = J2abbx aayy abb + bbx aayy bbxx bbx bbx bbx X = infinity, and PM = y = Jb b x x bx in- finity. Therefore PT: PM :: * : aa bx ::a: b. Al- a:b. a a fo 106 B. It CURVE LINES. Fig. aayy fo ATTP - AP = 98. abb + bbx 2abbx + bbxx abbx bbxx abbx abb + bbx abb + bbx abbx (when x is infinite) bbx a. Therefore the point T is determined, and the pofition of the affymp- tote TM. Alfo fince y may be bbxx aa bx by, this a gives another affymptote on the other fide of the axis AP, and making the fame angle. axy Ex. 3. Let the equation of the curve be, y3 aay + -3o. Extracting the root by infinite fe- ries, then yx — — a + &c. The equation 3 aa 3x of the curve put into fluxions, gives 3y'y +axy + ayx 3xxo, whence x = aa ax ЗУУ ay3xxy, and PT – aay yx ax + 3yyaa Y y's 3xx ay ax + 3yy aa and AT = У * 3xx ay 2axy + 3y³ aay 3223 3xx ay 2axy aay 3x3 + 3aay 3xx ay zaay axy 3xx ay axx 3xx (expunging y) (expunging y³) 3 a. But PT- zaxy + 3x3 ax + 2aa × x 3xx ax &c. axy + 3y³ = aay A || 3xx ay (expunging y³) axy-aay + 3x²+3aayзaxy 3XX ay 2aay Sect. III. CURVE LINES. 107 2ax zaa Xy Fig. = 98. 2aay - 2axy + 3x³ 323 3xx ay 3xx аху (expunging y) 3x3 2ax² &c. 323 32 ax &c. I 3 PT: PM: :x:y or x- PT PM. And therefore T falls between A and P; 322 =x, whence *, a::x:x; therefore fince AT is negative, and the affymptote therefore lies between the curve and abſciſſa. Ex. 4. Let the equation of the curve be y³ + axy I 3o, whofe root is y = x — — a + उ 03 81x² &c. then the equation in fluxions is 3y2y + axỳ + ayx 3x²o, whence x = 3yy + ax 3xx―ay, and PT yx 3y² + axy y 3xx ay = (expunging y³) 3x3 3axy+axy 3xx ay (expunging y) I 3x3 2ax X x a 3 3203 2ax² &c. 3x3 3xx a x x a 3xx ax &c. 3 3xx =X. Whence PT : PM : : ≈ : y or x &c. ::x:x. Whence again PT 33 a3 उ PM. Alfo 3223 2axy axy TP X = X зах = (ex- ay 3** ay ȧx X x I a punging) 3 3xx I ах х a 3 I axx ༣་ aax & c axx 3xx ax + + aa &c. a, 3xx 3 =-a; whence the fame conclufion will follow as in the laft. Ex. 5. axy Suppoſe the equation of the curve is y x3=0; this in fluxions is 3y-axỳ — ayx-3xx 108 B. II. LINES. CURVE 3xx + ayy, and PT – Fig. 3yy ax 98. = ö; and x = 3y3 axy 3xx + ay =(expunging j³) 3223 3x² + 2axy 3xx + ay i yx y 3x³ +3axy-axy 3xx + ay n And the root of the equation is y = x + 3 a &c. I Therefore PT = 3x3 + 2ax × x + a I 3xx + a x x + a 3 3x³ +2ax² &ċ. 3x² + ax &c. =(when x is infinite) 3x3 3x I *. Therefore PT: PM: x: y or x + 3 a::x:x, and PT = PM. उ 99. 11 Alfo PT 3x3 + 2axy axy X 3xx + ay 3xx + ay ax x x + I a ax² &c. 3xx + a × x + — a 3x² &c. = ÷ a = AT. I 3 3 Therefore fince AT is affirmative, the curve is concave to the abfciffa, and the curve lies between the abfciffa and the affymptote. PRO P. XI. Let AMN be any curve, CN its affymptote; then if the ordinates be parallel to the affymptote; they are depreffed a dimenfion lower in the equation of the curve. Let AP = x, ordinate PM y, and let the y, and let the equa tion of the curve be y²+qx" -1 &c. = "x", and C the root y=rx + B + &c. Take s infinitely X near M, and draw so parallel to AP, and DM parallel to CN; then by fimilar triangles, Mo (y) yx : so (x) : : MP (†) : PT = the fubtangent. But y Sect. III. 109 CURVE LINES. CX yx &c. and therefore PT = Cx XX rx XX r al r y C xx ; and when x is infinite, the ſubtangent = CQ, CN the affymptote being a tangent at an infinite diftance. Draw AB parallel to PM, and let CA t, CB = d, AB = b, CD = v₂ DM 2. Then by fimilar triangles, we have CQ (2): QN = v, : QN (y) : : CA (†): AB (b), whence ty = and trb. Alfo AB (b): BC (d): PM by, r bz (y): DM (≈), whence y = d Alſo AB (b) : ty Then AP = PC Fig. 99. AC (t): PM (y): PD = tz d ty AC; that is, x = 2/+v―t=(expunging y) +vt. Then fubftituting theſe quantities inſtead of y and x, in the equation of the curve, 72 X bn zn tz dn +ax + v t n-1 &c. = r" × 22 12 tz d zn in zn dn 72 &c. But brt, and b" I ; therefore 2" the higheſt power of the or- dinate vanishes; therefore when the ordinate DM is parallel to the affymptote, it lofes one dimenſion in the equation of the curve, C.c PROP. 110 CURVE LINE S. B. II. Fig. PROP. XII. 100. In any curve referred to an axis, if the fecond fluxion of the ordinate PM be negative; the curve is concave towards the axis AP. But if the Second fluxion of the ordinate be affirmative, the curve is convex towards the axis; fuppofing the fluxion of the qxis given. Draw the ordinates pm, qn parallel to PM, at equal but infinitely fmall diſtances. Produce the infinitely ſmall portion of the curve Mm to s. And draw Mo, mr parallel to the axis AP. Then the fluxion of the axis is Mo mr, and confequently mo sr, the fluxion of the ordinate. Now rn- mo is the fecond fluxion of the ordinate; that is, rnrs is the fecond fluxion, and this will be ne- gative when rn is lefs than rs, in which cafe Mmn is concave towards AP. But if rn rs is affir- mative, then rn is greater than rs, and then ʼn falls above s, and therefore the curve is convex towards AP. Cor. From hence may be known whether a curve, referred to an axis, is concave or convex towards the axis in any given point. Put the equation of the curve into fluxions, and find the value of the fecond fluxion of the ordinate, making the fluxion of the axis conftant. Then if the value of that fecond flux- ion be negative, the curve is concave; if affirmative, 'tis convex to the axis. Ex. I. Let the equation of the curve be aayy = 2bbrx bbxx, this put into fluxions, and divided by 2, gives aayy = bbrx+bbxx, and put in fluxions again, aayy + aayÿj = bbxx, and aayÿj = bbxx aayy, Sect. III. III CURVE LINE S. bbx 64x2 aajy, but y = xr+x, and j² = Fig. X aay a+yy 100. 64x2 2 r+x, whence aayÿ = bbxx 2 xr + x = aayy 2 bbaayy b4 X rŵ x² = (expunging yy) b4rrxx aayy ; therefore ÿ being negative, the curve is aayy concave to the axis. Ex. 2. Let the equation of the curve be aay + xxy = axx; then y = axx aa + xx 2axx › and y = aa + xx aa + xx and ÿ = 2ax³x 2a³xx + 2ax³ x 2ax³ x 2. 2a3xx aa + xx 2a3xx X aa + xx aa + xx 2 2a³xx X 4xx aa + xx 3 2a3xx aa + xx 8a³xxxx 295x2 6a3x²x² 3 aa + xx 3 aa + xx 2032 × aa aa + xx³ 3 3xx. Therefore if 3xx be greater than aa, the curve is concave towards the axis; but if aa be greater than 3xx, it is convex. Ex. 3. = If the equation of the curve be aax- a³ — y — b³. 3ỳ × y — b², and again 3ÿ × y — b 3ỷ × y Then aax + 3ỷ × 2ỷ × y —bo, and ÿ x y b + 2ÿÿ 2yy aax = o, and j = y- b But y = ЗХУ b 294xx therefore ÿ 39 ; confequently if y 9xy- b be greater than b, the curve is concave towards the Cc 2 axis; 1 12 LINE S. B. II. CURVE Fig. axis; but if b be greater than y, the curve is con- 100. vex towards the axis. 91. PRO P. XIII. In a curve where the ordinates meet in a fixed point, let the ordinate CM or Cm = y, perp. Mo = x, Mm ; then fuppofing x to be conftant, if y -żż be negative, the curve is concave towards the center; but if affirmative, 'tis convex. For by Prop III. Fig. 91. sn =ÿ ZZ y' which is affirmative if n lie above s, but negative when 'tis below s. And therefore yÿ żż will be af firmative in the former cafe; that is, when the curve is convex; and negative in the latter, or when the curve is concave. Cor. Hence if a curve is referred to a fixed point, to know whether it is concave or convex; put the equation of the curve into fluxions, and making x conftant, find the value of yÿj ZZ; then if it be negative the curve is concave towards the center; if affirmative, 'tis convex. Ex. I. Let the equation of the curve be yx = Z√ ay 2. Then x²= 2× ay az a 2 xx²+ y², whence 4yy 4y 4y 41 X j² = a zỷ X Alfo = a a дyyxx ay 4уxx 43a, and ÿ × ¹³, and y = √√ 49 2yx ż Vay ; thereforeÿ-żż and żż = zz гуйх Sect. III. CURVE LINES. 113 2. 4YXX 2y X a 2yx² √ 49 4уxx a a 2yxx a a a 4уx² Fig. a 91. which being negative the curve is concave towards the center. Ex. 2. = Let ABM be a fpiral, FBR the generating cir- cle, radius CRr, arch RD v, CM = y, and let the nature of the curve be denoted by the equa- tion vyaa. By the ſimilar ſectors DCd, MC, r: vy, whence yur. And putting the equation of the curve into fluxions, we have vỷ + rx, then y =- уч yʊ = o, and vỷ 2 and yÿ = rx aa X aa Alfo żż = x² + j;² = - yv = rx ryx › and ÿ = ryx V aa aa rryxx a4 x² + a4 a+ Therefore yÿ ZZ żż = rryyxx at + r²y² a4 ²x²; which a4 rryyx² 04 a+x² + rryyx² 95. being negative the curve is concave to the center. Ex. 3. But if the equation of the curve be vy3a4, and r = a, then 3vy'ỳ +y'v=0; that is (becauſe 3a4j yv = ax) y + ay³xo, whence y= y³ x and ÿ = 32xy 325 a3 -, and yÿ 313x y3x заз 3a° 16 x12 C c 3 yo xi z Alfo = x² + y² ** yas Therefore yÿ — żż — 345 аз 9ax² 9a 90° + gas +ys a3 X 90s 336 114 B. II- CURVE LINES. Fig. 3y6 — 9a6 — yo 2y6 даб x² 95. 9a6 x. Therefore if 9a6 ΙΟΙ. 2y be leffer than 9a6, the curve is concave; but if greater, convex towards the center. PROP. XIV. = Let AMC be a curve, AP = x, PM y, AM =z, Mm = 2, Mn = x, mny, and let y be given; and if a and b be given quantities, and ax bz be the nature of the curve, then will aż bx be a maximum or minimum. 2. 2 For taking the fluxion of aż -bx, we have až bxo, and až = bx. But ²² + y²; therefore 222 2xx, and żż = xx; whence by XX axx dividing, z= and až = , ; therefore Z Z axx Therefore on the con- Z bx, and ax = bz. trary when axbż, then aż - bx will be a maxi- mum or minimum. Cor. If ax = bz be the nature of a curve, then the Fl: aż - Fl: bx will be a maximum or minimum. For aż bx is a maximum or minimum, and therefore the fum of all the aż the fum of all the bx will be a maximum or minimum; that is, fluent of aż fluent of bx will be a maximum or minimum. PROP. P Fig. 89. M 72 A 17 72 go. D 92. 93. n P Q R S T 96. KAD m 1 9.9. TA P 102. m མ༡༦ A B C D ! R m 94. M 10 T A P 100. m 91. m 10 M M B m 10 n 95. R C 97. K P IM 98. M B ת M m M In 101. A P Pp D m PIVIII the End UN OF MICH Sect. III. 115 CURVE LINE S. PROP XV. Fig. If Bm, Dr, be two ordinates of a curve, infinite- 102. ly near; and let the ordinate Cn be drawn, fo that the differences of the ordinates on, tr, may be equal. = Then put mo=x, os or nt = v, on or tr = ÿ; then if ax + bv = d, a given quantity; and p x x² + j²² + q × v² + j² = a maximum or mini- mum; then pbx× ï² + j² = qav × v² + j² will denote the nature of the curve. -HI -I -22 I For putting the two given equations into fluxi- ax ons, ax + bvo, and v = - b Alſo 2pnxx -22 I •2 I x x² + y² + 2qnvv × vv +ÿÿ =o, and ·NI 2qnaïv -1- I 2pnxx × x² + jjz b x v² + j² •12 I o, and tranfpofing and dividing bpx × xx + ÿÿ = aqv x vv + ÿÿ 7 I P Cor. Hence 2122 × mo × mn a 21-2 Ny 0/0 × nt X b For xx + ÿÿ = mn², and ¿¹ + ÿ² = nr, &² jnr, whence pbx × mn²n—2 b qv X nr 212-2 11 2122 px = gav X nr 2n−2, and × mn a 212--20 Many more properties of curves might be ob- tained by the method of fluxions; but I forbear purſuing them any further; fince what have been here delivered, are fufficient for an introduction. FINI S. THE CONTENT S. BOOK I. Page 1 Sect. I. Sect. II. T THE Conchoid, The Ciffoid, Sect. III: The Cycloid, I 5 II Sect. IV. The Quadratrix, 16 Sect. V. The Logarithmic Curve, 19 25 31 38 Sect. VI. The Spiral of Archimedes, Sect. VII. The Logarithmic Spiral, Sect. VIII. The Hyperbolic Spiral, BOOK II. Sect. I. The generation of curves from one ano- ther by transformation; by which their lengths, areas, furfaces or folidities are determined, 49 Sect. II. The general properties of curves, the de- termination of their axes, ordinates, and affymptotes, 65 Sect. III. Several propofitions drawn from fluxions, very neceffary for inveftigating the na- ture of curves. 94 ERRA T A. For the A R. INFINITES. b fignifies reckon from the bottom. page line [read 33bs = n + 1. a 713 all the xes m 12 11 C - A+ 26 + C- 44 ob FINIS. Dele THE. CONIC SECTIONS. Where b follows, reckon from the bottom. [page] line read 18 290 +O 34- 56 60 131 66 XIII. Cor. 1.) 56 OD × DN : : 7 lines el, 2 ib. Prop. III.) 4 multiplying by 4. CURVE LINES. 89 7 whence a 99 2 |= jà × y* × 2 Books printed for J. NOURSE, oppofite Catha- rine-Street, in the Strand. of 1. THE ELEMENTS OF TRIGONOMETRY, by W. EMERSON. The Second Edition, with large Addditions; together with the Tables of Sines, Lo- garithms, &c. 8vo. 1764. Price 7s. II. The ELEMENTS of GEOMETRY, by W. EMERSON 8vo. 1763. Price 5s. III. A TREATISE of ARITHMETIC, by W. Emerson, 8vo. 1763. Price 4s. 6d. IV. A TREATISE of ALGEBRA, by W. EMERSON. 8vo. 1764. Price 7s. V. EMERSON'S NAVIGATION. The Second Edition. 12mo. 1764. Price 4s. VI. A new Method of INCREMENTS, by W. EMERSON. 8vo. 1763. Price 7s. 6d. VII. A TREATISE of ALGEBRA, by T. SIMPSON. The Third Edition. 8vo. 1766. Price 6s. VIII. ESSAYS on feveral curious and uſeful Subjects in ſpeculative and mixed MATHEMATICS, by T. SIMP- SON. 4to. 1740. Price 6s. IX. MATHEMATICAL DISSERTATIONS on a Variety of phyfical and analytical Subjects, by T. SIMPSON. 4to. 1743. Price 7s. X. MISCELLANEOUS TRACTS, on fome curious and very intereſting Subjects, by T. SIMPSON. 4to. 1757. Price 75. XI. TRIGONOMETRY, Plane and Spherical, with the Conſtruction and Application of Logarithms, by T. SIMPSON. The Second Edition. 8vo. 1765. Price 1s. 6d. XII. The DOCTRINE of ANNUITIES and REVERSIONS, by T. SIMPSON. 8vo. 1742. Price 3s. XIII. APPENDIX to ditto. 8vo. Price 6d. XIV. The ELEMENTS of GEOMETRY, by T. SIMP- SON. The Second Edition, with large Alterations and Additions. 8vo. 1760. Price 5s. XV. SELECT EXERCISES for young Proficients in the MATHEMATICKS, by T. SIMPSON. 8vo. 1752. Price 5s. 6d. XVI. The DOCTRINE and APPLICATION of FLUXIONS, by T. SIMPSON. 2 vol. 8vo. 1750. Price 12s. XVII. The Books printed for J. NOURSE. XVII. The ELEMENTS of EUCLID, by R. SIMSON. The Second Edition. To which is added, The Book of EUCLID'S DATA. 8vo. 1762. Price 6s. XVIII. INTRODUCTION A L'ARITHMETIQUE VUL- GAIRE. 4to. 1751. Price Is. XIX. The ELEMENTS of NAVIGATION, by J. Ro- BERTSON. The Second Edition. 2 vol. 8vo. 1764. Price 18s. XX. The ELEMENTS of ASTRONOMY, tranflated from the French of Monf. DE LA CAILLE, by J. ROBERT- SON. 8vo. 1750. Price 6s. XXI. The MILITARY ENGINEER; or a Treatife on the Attack and Defence of all kinds of fortified Places. In two Parts. 8vo. 1759. Price gs. XXII. The AcCOUNTANT; or, The Method of Book- KEEPING. By J. DODSON. 4to. 1650. Price 4s. 6d. XXIII. The MATHEMATICAL REPOSITORY, by J. DODSON. 3 vol. 12mo. 1748-1755. Price 12s. XXIV. An eafy Introduction to the Theory and Practice of MECHANICs, by S. CLARK. 4to. 1764. Price 6s. XXV. The BRITISH GAUGER; or, Trader and Offi- cer's Inftructor in the Royal Revenue of ExCISE and CUSTOMS. Part I. Containing the neceffary Rules of vulgar and decimal Arithmetic, and the whole Art of PRACTICAL GAUGING, both by Pen and Rule; il- luftrated by a great Variety of curious and uſeful Ex- amples. Part II. Containing an hiftorical and fuc- cinct Account of all the Laws relating to the EXCISE : To which are added, Tables of the old and new Du- ties, Drawbacks, &c. on Beer, Ale, Spirits, Soap, Candles, &c. with a large and copious Index, by Sa- MUEL CLARK. 12mo. 1765. Price XXVI. The ELEMENTS of FORTIFICATION, by J. MULLER. The Second Edition. 8vo. 1756. Price 6s. XXVII. The METHOD of FLUXIONS and INFINITE SE- RIES, with its Application to the GEOMETRY of CURVE LINES, by the Inventor Sir ISAAC NEWTON. Tranflated from the Author's original Latin. To which is fubjoined, a perpetual Comment upon the whole Work, by J. COLSON, F. R. S. 4to. 1736. Price 15s. 5s. UNIVERSITY OF MICHIGAN 3 9015 06558 6169 +