p, a hyperbola exterior to the limiting parabola; and for
c=, the curve is again the limiting parabola.
35. ‘To find the locus of one end of a given straight line,
whose other end, and a given point in it, move in straight
lines at right angles to one another.
AP the given line =a, B the given point in it,
PB=b, CN=a2, NP=y, LPBN=6; (fig. 101) then
ev=acos0, y=bsin8, ie!
ipa aed , ‘
— + ey 1, the equation to an ellipse.
ERT
If the rectangle CO be completed, and PO joined, PO is
a normal to the ellipse at P; for
GN GN BN bb ®&
CN “NG bua iaineds
Calling AS'=p, SC =, and taking the vertex for origin,
175
Hence if a line, whose length is the semi-major axis of
an ellipse, have one end in the curve, and the other in the
minor axis; then (1) the part cut off by the major axis will
_always equal the semi-axis minor, and (2) the locus of the
intersection of the perpendicular to the minor axis through
one end, with the normal through the other, will be a circle.
36. ‘Two given circles touch each other internally; to
. shew that the locus of the centre of the circle which touches
each of them, is an ellipse having their centres for its foci.
SJ and H the centres of the circles, P the centre of a circle
touching both; join SP passing through the point of contact
B, and HP passing through the point of contact 4 (fig. 102);
thn SP + HP=SB+ BP +(HA- AP)=SB + HAA,
which is constant ; therefore the locus of P is an ellipse.
| 87. Ifa, B, y be the angles which the transverse axis,
and the focal and central distances of any point of a curve of
| the second order, make with the tangent at that point,
tana.tany = tan’ 3.
Let SPY =6; CPY =v, STP =a, (fig. 41)
SG cos p-
$ na hed 3
SP cosa
then e =
NG :
and —— = tana.tan(y —a)=1-—e*=1-———_ ;
therefore, reducing, we get tan a tan y = tan? 2.
334 Lhe products of the alternate segments of the sides
of a polygon described about an ellipse are equal.
Let p, g, 7, s be the lengths of the semi-diameters re-
spectively parallel to the four tangents at P, Q, R, S, (fig. 103)
the proof being the same whatever the number of sides ;
O,Q iy q? O.P Rae O.R gp? CO See 3?
” O.P.O0.R. OS .0,Q = 0,Q:.0;,P .0O:R . OS:
(Art. 155);
then -
176
|
In the case of a triangle ABC whose sides touch the ellipse.
in the points a, 3, y, we should get Ap. Ca.By=Ay.Ba.CB, |
which shews that the three lines j joining the points of contact |
with the opposite angles, intersect in a point. |
39. If @ be the acute angle between the tangent and.
focal distance at any point of an ellipse, the distance of that.
point from the centre is \/a? — b? cot?@ (fig. 42).
2 PE? |
For pad (51) <0 — 2 (, ~1) =0°—B cot? 8, |
P |
CQ’
40. To find the locus of the intersection of the normal to |
'
an ellipse and the perpendicular upon it from the centre. The |
equation to the normal at any point is
:
(y — ma) a" + mb + (a — B) m= 0,
and the equation to a perpendicular upon it from the centre is’
1 } : a
y =——2#, which gives m= —-; |
a y
therefore, substituting, the equation to the required locus is
(a? + y?) Jay? + Bx = (a? — 6) wy.
41. A given triangle has always two of its angular points :
in two given straight lines; to find the locus of the remaining ©
angular point.
Take the given lines for the axes, and let
ZAOB=w0, CM=x, CN=y; 4 OAB = @, (fig. 104) ;
“. ysinw = bsin(A + 9),
wsinw =asin(B+nr—d-w)= —asin(4+$+C +o),
.. bxsin w= —a cos (C+w).ysin w— ab cos (4+ ¢) sin(C+w), |
“. sin’w iba + acos (C+w) y}*= a’sin? (C+) (0 — y’sin® w),
the equation to an ellipse of which OQ is the centre. |
42. "To find the Pisa of the middle point of a chord of
constant length in an ellipse.
177
Let QV=c, CV=r, ACV=8, (fig. 45); then
QV? CV?
CD?” CP
b?
1 ~e* cos?@ ”
= 1
But, CP? =
CD = a2 4B? CP? a 1 E+ B) & c0°9
1 — e* cos’?
9
c* (1 — e* cos? 6) ri r” (1 — e* cos? @)
e , =
a” — (a* +b”) e* cos?@ b? :
‘the polar equation to the required locus.
|
_ 43, Two given circles are traced upon a plane and a line
is drawn touching one and cutting the other in two points at
which tangents are applied to the latter circle; to find the locus
of the intersection of the tangents.
“0P=r, POO! =6, 00 =0, 0Q =a, O'@ =a! (fig. 69),
a
then. ¢ cos 0: =.0ON oe
a?
LF = ——____
ccos@—a!’”
‘he equation to a conic section of which O is a focus.
44. Having given the base and altitude of a triangle, to
ind the locus of the centre of the inscribed circle.
SC = CH =c=half given base (fig. 41), PN =a the
tiven altitude, O the centre of the inscribed circle, CM = 2,
WO = y its co-ordinates, CN = a’; then
a S’ a H
an S! = p> tan—= pe tan H = ——, tan— = 4 ;
C+a@ 2 C+a@ C—Z 2 C— x
a 2y (c+2) a 2y(c—a) |
ee c+a (c+a)—x¥’ sn" = (c- 2)? -y¥?’
12
|
therefore inverting and adding in order to eliminate a’, we get |
the required equation, which is of the third degree,
178
— —
45. Having given the base of a triangle, and the sum of
the other two sides, to find the locus of the centre of the
inscribed circle.
SH the given base = 2ae, SP + PH =2a, and SO= ry
HSO=8, (fig. 41), polar co-ordinates of the describing point O 5)
then area of triangle = 4 perimeter x radius of inseribed circle
=a(1+e)rsin@,
a’e(1 — e) sin 20
1 —ecos 20
r
os
also area of triangle = SP sin20.ae=
2ae (1 —e)cos@
1— ecos20
oie de =
the equation to a coni¢ section of which S'H is the major axis.
46. Two focal distances of a conic section include a con-
stant angle 2, and one of them is produced to meet the tangent
at the extremity of the other, to find the locus of the point ol
intersection.
B.
ST =r, 2AST=0, PSQ=[; (fig. 30), then :
a (1 — e’) :
= of = —~—_———._ (Art. 12
ASP = 23+0, and , AVG isd GON (Ar re}
the equation to a conic section with focus S'; and which 1s at
ellipse, hyperbola, or parabola, according as cos 3 >, <,or =@
If we draw another focal distance SP’ inclined at angle f
to SQ, then the tangent at P’ will pass through 7’ (Art. 128):
therefore the preceding is also the solution of the problem t
find the locus of the intersection of tangents to an ellipse a
the extremities of two focal distances that include a constan
angle 2/3. |
179
Also, if T'’P be produced to 7” so that zPS'7” — PST,
then 7" is a point in the locus of TZ’; from whence it follows
that the chords of a conic section whose eccentricity = e, that
_subtend an angle 2 at the focus, will be tangents to an-
other conic section having the same focus whose eccentricity
| = e cos f.
47. In an ellipse, if two focal distances r, 7, include an
angle = 28, and 7’ be the intersection of tangents at their
extremities, then
b? ry’
A Ae ae SET PS ya sere
6° — rr’ sin? 3
:
1— Z L 1 Sa a
We have Grae) = 1+ ecos@, eee +ecos @,
: r
: ‘a(1-e’) ; ;
‘and — go = 0088 +ecos AST =cos (6'— 6) +e cosh (0+6’) ;
‘between which equations if we eliminate @ and 6’, we shall
obtain the above result; which in the case of the parabola
becomes SJ? = 77’ agreeably to Art. 83.
48. To find the equation to the curve traced out by a
‘point in the perimeter of a circle which rolls upon another
equal circle.
Let 4’ be the describing point, at first in contact with A,
and 44’ the curve traced out, (tig. 70) ; C, C’ the centres of
the circles; join 44’, CC’, and let
AA =r, AAE=0, AC=a.
AA’ is manifestly parallel to CC’, draw DA’ parallel to
AC, and therefore = AC;
then 44’ = CD=CC' ~ LiGe
or 7 = 2a — 2acos@, the polar equation.
Or if AR = x, RA' = y, be the rectangular co-ordinates of A’,
i
o“+y’=2a (1-—* ,
Sairy Warns)
oe +y=2 a(r/ x +y” — 2).
iii
10M a
Hf:
im
ny
Hi
it
‘|
i
a PA na
SSS SS SS Ss SS
ee
180
49. ‘To find the equation to the curve described by a
point in the perimeter of a circle which rolls within another
circle of four times its radius.
P the describing point, at first in contact with A (fig. 70),
and AP the curve traced out;
CN=a, PN=y, CA=4a, QO=a4,
O being the centre of the rolling circle,
perpendicular to AC,
LPOM=7- (= -) —419=— — 305
ro)
~
w= CM + Pn =8acos0 + acos 30 = 4a(cos 6),
y = OM — On= Sasin 8 — asin 30 = 4a(sin ay
4 \ 4
(eyo
Aa Aa
i Suppose C A = 2a, the radius of the rolling circle equal a,
circle, but at a distance c from its centre ; then
LPOM=}47-90, «. v= (a+) cos@, y=(a—c)sin@;
x” 2
rei le es ee
(a+c) (a-cy)’
the equation to an ellipse; except a = ¢ when the equation is
y = 0, and the locus is the axis of a.
1,
50. Ifa triangle be inscribed in a Conic Section, and
each side be produced to meet the tangent at the opposite
angle, the three points of intersection will lie in a straight line.
If we take C for the origin, and AC =b, BC =a, for
axes, the equation to the conic section must be of the form
a+my +ney—ax—mby=0.
The equation to the tangent at 4 is
nb—a mb°
vw: and when y=0, v=
410 Oia ei yi nb—a’
y-b=-
ACD = 0; therefore QOP = 40, and consequently, OM being
and the point P to be not in the circumference of the rolling
181
which determines one of the points. The equation to the
tangent at B is y= — (vw -—a), and when w =0,
na—mb
y = which determine a second point. The equation
na — mb
— ax : :
- and equation to AB is
to the tangent at C is y=
ve Lit: ; mab?
say en: and for their intersection « = —__—. ,
a 6b mb? — a?
° a” e e °
y= rap aniatas the co-ordinates of the third point. Now the
mb? — a?
equation to the line joining the first and second point is
na—-mb nb—a
nae OE e=1,
a” m b*
and this is evidently satisfied by the co-ordinates of the third
point.
51. An ellipse being referred to conjugate diameters, if
with the co-ordinates of any point as conjugate semi-diameters
a second ellipse be described, it will be touched by the chord
of the former that joins the extremities of the diameters.
The equation to the interior ellipse will be
9 9
a
es e e h?
= 1, with condition = + 5 =1;
ee ® Mee
aX ~
We
‘and for the intersection of this ellipse with the chord
v Bare Bi 2 i
at pmb we have G+ = (1-*) = 1, or (ax —h*)’=0,
a a
Ween le
which shews that the chord is a tangent at a point for which
h?
v= —.
a
52. The chords joining the extremities of conjugate dia-
meters of an ellipse will all touch in their middle points a
similarly situated ellipse with axes ar/ 2, b a/ 2.
i
Se “<< *
182
The equation to the inner ellipse and to the chord, re- |
ferred to a pair of conjugate diameters of the outer ellipse as
axes, will be respectively,
~
et Tae
3/8
+
IS
See
=
Q|s
Cte
—
we
FY Ce Qa a
2(—+1-2—+—, =: 1 OP bee = =O 5
A he a 7 Pas
therefore for their intersection we have,
which shews that the chord is a tangent in its middle point.
53. To find the locus of the intersection of two tangents
to an ellipse applied at the extremities of a chord which always
touches a concentric and similarly situated ellipse.
Let a, 6, be the semi-axes of the exterior ellipse, a’, 0’,
those of the interior ; and (h, &) the point through which two
tangents to the former pass; then the equation to the chord
joining the points of contact is
he ky :
a” st e
’ id
which must be identical with —- - = 1,
ae 2
the equation to a line touching the interior ellipse at a point
(a, y’); therefore |
hence since (=) *
On.
i. = 1, the equation to the required
ye ;
locus is (=F) + = ‘) - = 1, representing a similarly situated
2
a
ellipse with semi-axes —, =
ab
54. An ellipse whose centre C is given touches a fixed —
straight line PQY in a given point P; to find the locus of
either focus S.
|
183
Let PQ=h, CQ=k (fiz. 105) be the co-ordinates of
C, which are known, and PY =a, §Y = y those of §; then
since CY is parallel to PH, zCYQ = SPY,
k y k oh
a & = - OFr.-+-— = 1,
ec-h @& Yi a
the equation to a hyperbola.
55. If an ellipse and hyperbola have the same foci, the
locus of the intersection of tangents to them, at right angles
_to one another, is a circle.
The equation to a line touching the ellipse is
Y-ML= a/b? + ma’,
: e s 1 .) e ry
and changing m into ——, and a’, 6°, into a?, —b', the
m
equation to a line touching the hyperbola, and at right angles
to the above, is
my +e =/— mb? + a?;
where, since the curves have the same foci and centre,
NOC = 0 — ft =a" fp?
Adding the squares of these two equations,
(2? +. y’) (1+ m?) = 6? + @ 44m? (a? ie b’?) xi (b? ve a’) (1 4. m*),
or @ +y=h 4+ a”,
the equation to a circle passing through the four points of
intersection of the curves.
56. To find the locus of the centres of the ellipses in-
scribed in a given quadrilateral.
Take lines through one of the angles of the quadrilateral
parallel to two sides for the axes of the co-ordinates; and let
w=h, y=k, y=mex, y=na, be the equations to the four
sides; then the conditions for these lines, respectively, being
tangents to the ellipse, supposing its equation to be
ay’ +bay +c2#°+dy+er+1=0,
Se —
SS
———$ =
a
A ae Es eS as a SS SSS SSS SS : Bee SSS
ee Se eee ie ee —— ————— a: = ee
i
ae eS
184
(found by making the two points coincide in which each cuts
y s P
the ellipse) are
4a (ch? + eh +1) = (bh +d)’,
4c (ak? + dk +1) = (bk + e)’,
4 (am* + bm-+e) = (dm +e)’,
4(an° + bn +c) = (dn + e)’.
But if 2’, y’, be the co-ordinates of the centre of the
ellipse, the two former become (Art. 223)
(4ac — 0°) (h? — 2ha’) + 4a-a =0,
(4ac — b’) (KP — 2ky') + 4c —e& =0;
and eliminating 6 between the two latter, we get
(4a -—d’) mn = 4c -e*,
* h? —2ky =mn (hl? — ghar)
the required equation; which represents the line joining the
middle points of the diagonals of the quadrilateral. -
That the locus of the centres would pass through the
middle points of the three diagonals might have been fore-
seen; because each of the diagonals may be regarded as the
transverse axis of an evanescent ellipse touching the four
sides of the quadrilateral.
If one of: the angles become equal to two right angles,
the ellipses are inscribed in a triangle, touching its base in
a given point; and their centres lie in the line joining the
middle point of the base, with the middle point of the line
drawn from the vertical angle to the common point of contact.
57. Ina given triangle to inscribe an ellipse of given
area, and touching one of the sides in a given point.
Let P be the given point in the side BC (fig. 106) and M
the middle point of BC; draw MS bisecting AP, and AQ
cutting off QC = BP; then the centres of all the ellipses that
185
can be inscribed in the triangle and touch BC in P, lie in
sequently the loci of D and O are as asserted. Let DQ = z,
AQ=k, MC=a, BP =c; then
DIT leas
RC tk
and OP..sin OPM = 4 DQ.sin DQP = 4 % sin w, suppose,
Cc
or DT => (k-2),
Now (area)? = 7’ sin°OPM.PC.DT. OP?
mT’ sin’w(2ac—c*) ,
cea Re RAIS 9 all
Ak ( 2)
If the area of the ellipse equals the area of a circle radius r,
then (2ac¢ —c’) sin’ wx” (4 — x) = 4kr*
is the equation for finding x. For the greatest inscribed ellipse
that touches BC in P we must evidently have
x= 2k, or MO = 280.
1 929.09
Then 7* = za k* sin” w (2ae — c’),
which is a maximum by the variation of ¢ when ec = a, or P
coincides with M. Hence the greatest of all the inscribed
ellipses touches each side in its middle point, and has its centre
coincident with the centre of gravity of the triangle,
58. In the equation ay’? + bay +ca’?+dy+ev+1=0,
suppose 6 to assume different values, all the other coefficients
remaining unchanged; then (1) the conic sections which it
represents are in general all described about the same quad-
rilateral; (2) the locus of their centres is another conic section,
whose equation is 2ay’+ dy = 2ca* + ea; and (3) the centre
186
of this last conic section is in the middle point of the line
joining the bisections of the diagonals of the quadrilateral.
It is evident that the four points in which the curve cuts
the co-ordinate axes are independent of b. If h, k, be the
co-ordinates of its centre, then
2ak+bh+d=0, 2ch+bk+e=0,
between which eliminating 6, we find the locus of the centre to
be the conic section, whose equation is
2ak? +kd = 2ch’ + he.
If h’ and k’ be the co-ordinates of the centre of this curve,
, e d
h’ = —-—=1(7,4+%,), k= - a 1 (y, + Y2), (Art. 240)
which are the co-ordinates of the middle point of the line |
joining the bisections of the diagonals. It may be shewn that !
the curve passes through the intersection of the diagonals, and
also through the points of intersection of each pair of opposite |
cH sides. |
|
|
59. To find the locus of the point which is the inter-
section of three normals to an ellipse. :
The equation to the normal at a point (#’, y’) of am.
ellipse is
a y'
Y si y 7 be a (wv 7 a),
or ya \/1 —@ = (a — Ba’) fa? — &”. ¥
Let h, & be co-ordinates of a given point through which
the normal passes, then |
Kw? (1 — e&) = (h—ea')? (a? — @”)... +02 00-(1)
is the equation for determining a’, the abscissa of the point
in the ellipse; and as this equation is of the form :
ex! a &c. sae a? h? == 0,
187
it has two or four possible roots; and consequently through
‘the point (h, k) in general either four or two normals can
be drawn. If two BP the possible roots become equal (which
jcan only happen in the case of four real roots), then three
normals will pass through the point (h, k); in that case
ithe derived equation
Kea! (1 —e*) = —& (h -— ea’) (@ -— w”) — or (h- Ca’?
has one of them. Dividing this by (1), we find
| t ;
1 e & ee ee
SG = Se Or he? = C* ar? s
a h-@uv a—«x?*
ha\s _, Byes
ae = ee satisfies equation (1), and substituting we get
Ke
k* (1 — e’)3 +h® = (ae)?
for the equation of condition that (1) may have equal roots,
and as often as h and k satisfy this equation, three normals to
the ellipse will pass through the point (hk, k). The above is
consequently the equation to the locus of the intersection of
three normals to an ellipse; and coincides, as might have been
foreseen, with the equation to the evolute.
60. If the tangents at the extremities of any diameter of
mn ellipse DD, fs intersected by the tangent at any other
ooint, in 7, 7"; then DT. D'T" = CP’.
The equation to the tangent at Q (fig. 47) is
, 4
ve yy
BR pals / ? °
md making y = b’ and — 0b’, successively, we get
av’ , av ,
DTH ee MR THO RN Sey, pt
OF b a” b
, , a? y” trp / 9
Oy TOD LET Bd bee 7 = j2? or DT .DT Sa =. P*,
a” 2
61. The greatest pb that can be inscribed in a quad-
‘lateral that has two sides h, k, parallel to one another, will
188
touch those parallels in their middle points; and its area will
=f 1 +14/hk, l being the perpendicular distance of the par]
rallels from one Ethan
Join the two points of contact of one of the incr
ellipses with the parallel sides by the line DD’, (fig. 100)
bisect it in C, and through C draw PP’ parallel to D7’; then
C is the centre, and PP’, DD’, are conjugate diameters
of one of the ellipses inscribed in the trapezium; and if
CP=a, CD=b, DT =c, DT =d, @=cd=(h-@
x(k —d) (Prob. 60); .. kec=h (k-—d). Now (area)? of ellipse
«a «(k—-d)da«tk® —- (1k -d)’; therefore, for maximum
area, d= 4h, naa c=th; and maximum area = 7a.4)
= La yia
If h =k, the trapezium becomes a parallelogram, and the
greatest ellipse = 4 x area of parallelogram.
62. In a given parallelogram to inscribe an ellipse of
given eccentricity.
Every ellipse must have its centre in the intersection of
the diagonals; and as in the preceding Problem, if Q be the
middle point of the side RT’, and CQ=1, QT =k, QD= a
PCQ = 0, PCD =.0, CP =a, CD = 6,. then
a= DT.DT =k — 2,
P=? +2lscos8 +2", bsinw= sin.
If therefore a and £ be the semi-axes of the ellipse
a’ + BP =k? +2? + 2lzcos8,
af = lsinO\/k? — 3°;
+k
. Lsin @ te + P) sa + 21 cos@. Sailnet,
Boa VJ kt — x? J ke? — 2°
the equation for determining x, since Fa Ste is given,
189
Since the ratio of 3 to a is zero when the ellipse coincides
——-
with either of the diagonals, it must admit of a maximum ;
and the corresponding value of x may be obtained from the
—
:
;
2quation
(Ao + P) x + 21k? cos8=0, (1)
which determines the ellipse that approaches nearest to a circle
of all those that can be inscribed in a given parallelogram.
If the parallelogram be equilateral, and if PD, CT, be
‘coined, then CT’ bisects both the chord PD, and the angle 7’
>]
and therefore bisects the chord perpendicularly ; : therefore Cr
' which gives a.
san Axis of the curve; and if CY be a perpendicular from C
m RY, since YCT =1 (7-8),
CY=ksn0=av/1 — e’ cos’ 4 0,
If e=0, QD evidently equals & cos @ which
Figrees with (1) when J =k.
63. ‘To inscribe an ellipse in a semi-circle, which shall
-iave a given major axis parallel to the diameter of the semi-
eirele.
CN =a, NP =y, the co-ordinates of P (fig. 107); then
-»ecause the normal at P passes through the extremity of the
-ninor axis,
y= 7 ore but (y +6)? +a? = BP? =r’,
a®*—b
a’
or (y+ b)'+ OU —y) ar, .. ata (a? — 8),
orb =e
Hence in order that the area may be the greatest possible,
ve must have ab a maximum, or a? a/v — qa’ a maximum ;
—
9
ae + WE , and greatest area =
2
2
9
Qarr-
aV3
Sa ea So Se
190
64: To find the locus of the middle point of a straight line
that always has its extremities on the circumferences of two
equal circles given in position.
A, D, (fig. 108), the centres of the circles, O the middle
point between them the orign; ON=a, NP=y the co-
ordinates of P the middle point of BC; 2’, y'; wv’, y”, the
co-ordinates of B and C; BC =2c, O4=OD=b, AB = CD
=a; then
2yay +y’s
2v=2 —2',
Ae?= (y’ —y")? + (a +2’,
a =y” + (a — 5)’,
aay”? + (a —b)’,
between which five equations we have to eliminate a’, y,
/
wv’, y"; subtracting (4) and (5) and reducing by means of
(1) and (2), we get
(6) yy -y") + 0(a' +0") =2be.
Again adding (4) and (5) and doubling
2(y? +9?) +2 (a? + a") — 4b (a! + 2”) = 4(a2 - B),
but (y' + 9)’ + (#’ — a”)? = 4 (a? + 7)
from (1) and (2); therefore, subtracting, and reducing by (3),
(7) e+e’ =—-(P+ce—a'+a*%+y"),
1
b
2
be «x
also from (6) y/ —y" = —— — - (a +2”
Ms a
L
be ve. 2B 2. @?);
hein ail dc Lae
. substituting in (3), and reducing, we get for the equa-_
tion required
o 9 9
~
P(e +y+eO-7 -8Y 4g (4+ yt -27+ bY = 4 ey’.
191
65. ‘To find the locus of the intersection of two normals
to an ellipse at right angles to one another.
Let m be the tangent of the angle which a normal to
the ellipse makes with the major axis, then its equation
ds (Art. 125),
(y — ma) fa? + mb? + (a? — B°) m = 0,
'and changing m into , the equation to a normal per-
| pendicular to it, is
(my + 2x) / ma? +B? — (a” —b’) m= 0,
-and we have to eliminate m between these equations.
We get by addition,
a+mh (my+ay
mar+b? (y—ma)
C1);
a+ atm? (a? + m°b*) (ma +8) ,
ety (myteP (abn
a’ +b? 1\2
Pe -(a-by= + bP + ab! (m=) 2).
sng (a BP = (a +8 +)" ©
But from (1) we get
a-P vy AMLY
oe ° = Q ” =f . » 9 ~ 3
a+b at ty? (a + y’) (1 —m’*)
1 Quy (a? + b*)
—-— m= - noted
m ay — Pa?’
which value substituted in (2) after reduction gives-the re-
quired equation
(a” ny 5°)” ee oti al
(a at b°) (v” + y’) His a*y? — ba? :
66. ‘To find the locus of a point from which if four
normals be drawn to a curve of the second order, the sum of
their squares shall be constant.
‘|
192
- Let y? + na*=c?® (1) be the equation to the curve, and
(a,b) the point whose locus is required; the equation to
!
the normal through it is
nba
y
—-b=—(«#- et) ees
J ND Gir a (n—l)@+a
nbx 2 Wer
7. —— eS 0S
(n-l)e%+a
Qa n (a> + 26*)—(n =1) c
2° ( ag ) a* + &e. = 0.
4
or @ ¢ -
weet n(n — 1)”
Similarly,
2nb n (a* + nb?) —(n —- 1)? ec
vey? u Jost Be 7 te
(2 — 1)?
24n(a* + nb?) —-(n- 1)?
(m — 1)°
3 (-2by) = —26 (=) enti
n—1 n—1
4b? = 4b’,
3 (wal +(y— by"
as 6
o
jan 2) = 4R* a constant ;
n
An —2 2n — 4 ]
or a® ( jae. ) =4K?-20 (1+-);
nit reek 1 n
or of. @m—1) +0. (w=2) = (n= 1) 2B ( +7)}.
193
67. If an ellipse be inscribed in a quadrilateral, the lines
| joining the extremities of either diagonal with the points of
contact will intersect in the other diagonal.
In fig. 109, because the sides of the triangle KAC are cut
by a straight line in the points N, J, M,
| EN WEN GA Te
KM Gl aM’
KM CR MD
bites 5 epee ee
a au SIT VERE
because the ellipse is inscribed in the triangle KCD,
“ CR.MD.AI=DR.AM. IC,
which proves that if CM and AR be Joined they will intersect
am DI. Also, because the quadrilateral is circumscribed about
he ellipse,
Aly DR.AM AL.BN
IC CR.DM” BL.NC’
or BL.NC.AI= AL.BN. IC,
: thich proves that AN, CL, BD, intersect in a point.
| 68. If an ellipse be inscribed in a quadrilateral, the line
dining its points of contact with two opposite sides passes
orough the intersection of the diagonals.
| Let K the point in which the opposite sides BC, AD in-
“sect, be taken for the origin (fig. 109). KA=a, KB= b,
=c, KD=d; and KN =k, KM =h, M and N being
he points of contact in AD, BC; then the equations to BD,
'C, and NM are, respectively,
: v Y v Y Y
Soe AT OY ecm +-=1;
: Tema CE To kepeais
:
id therefore the condition for their passing through a point is
ee
194
Now the equation to the ellipse is (Art. 228),
(; 1) = (7 1) may =1
h i ig J
and for its intersection with the line CD whose equation is
d
= 1, we have
c
a 1) (bi ( k 1) i ( “) ;
—_— —{—-4--—- mec —-—j)=1;
E te \at ce d
which must be a perfect square since CD is a tangent,
m 1
2 2G 1) (
Pee h ~ ab
. “ 4 e ° ° e,e :
(since — + a 1 is also a tangent), which is the condition (1),
a
69. To find the area of an ellipse inscribed in a given
quadrilateral, and touching one of the sides in a given point.
As in the preceding Problem, taking two of the sides ol
the quadrilateral for the co-ordinate axes, the equation to the
ellipse will be
i 1) Y Get Inky?
a + (2-1) +270y=1, (ihk)’ <1;
1\ )2 2
+o yak(r ho) kle- 2 (e+ 7) {(;-18) af.
v v
Hence when the ordinate becomes a tangent at N’, Ww
have
oh
yay tee
“é hkl+1
195
Therefore when «= 1K, the radical in the value of y
1—hkil\2
——]} = OQ the seni-
mT cal Q e semi
assumes its greatest value = (
yt .
diameter conjugate to NN’:
;
(1 -hkil)2
(1+ hkl)2
area of ellipse = rhk sinw
| AI BI BL :
: Now let To 7” Tw ane L being the given
loint where the ellipse is to touch AB;
then, since BL.MA.DI= AL.MD. BI,
AM wm AM mM
=—, and = .
MD 3 AD m+e
But KA ABAC_ BI AC m (2 +1)
| LED TABDOCPIRDS 1C8 m+1
KA m(n+1) AM 1—-mn
= su hencepest7 ‘
AD 1-mn KA (n+1)(m+32)
AM a I1—mn
ang == 1-- = ;
KM h 1+m+sz(14+n)
(1-—mn) x ;
l+m+z(1+n)’
- 1+m)3s°+ (1+m) 2?
: area = Lad sinw (1 —mn) 21+”) + +m) eth
3m (” +1) +(m +1) nzh3-
consequently 1 —
b
k
70. In a given quadrilateral to inscribe an ellipse whose
a shall be the greatest possible.
The expression for the area in the preceding Problem
lishes when x = 0, and when = © ; and also when
132
196
(1+n)#+1+4+m=0, which gives AM = — AK; corresponding
to which values of x the ellipse becomes coincident with the
diagonals BD, AC; and with the line joining K, and the
point in which 4B, CD intersect. The area of the ellipse
will consequently be a maximum for some value of x between
zero and infinity given by the equation
: (== me) m.
re 2 eee es
n+I1 n m+)
The ratio v of CR to RD must be the same function of
1 1 ; , :
" and — that ¢ is of m and m, and is therefore given by the
m
”
° 9 mnt+l m+i m t
equation v + 2v |—. —- —— =0; which shews
mm+il n+i n )
that the negative value of » taken positively is the value of
the ratio of CR to RD.
|
71. Ina given quadrilateral to inscribe an ellipse whose
axes shall have a given ratio.
By Prob. 69, since the equation to the diameter NN’ is
k—y=k’lx, we have 7
h?
ON = ray ee ee eps
and ete Hye mes
ba i |
at 65 h? +k? — 2h?k?l cos
a a
(LE RED: ane
. Q=hkd!
| and afp = hk s1n @w hE 9 j
a and B denoting the semi-axes of the ellipse; also let |
denote their ratio, ;
1 |
.. sin (v + -) hk {1 — (hkl) tt=h? + k? - 2h? Kl cosws
Ss 2y
Bie ETA Te? PCa ay Es as
(m +1) (nm +1) (m+ 2) (14 72)
Sk 1—-mn k (1—mn) x
since -—-1= — Se ach se oan Bete CRS at sf
a (m + 1) (m + 2) b (m + 1) (1 + nz)
hk @ mote + ne
also - = —. =?
| kee bee 1a or +
and upon substituting these values there arises an equation of
the fourth degree for determining x.
72. To find the locus of the middle points of a system
of parallel straight lines, each of which joins two points in two
given curves.
Let f(v, y)=0, d(#,y)=0, y=ma+e, be the equa-
‘ions to the two curves, and to one of the chords; transfer
he origin to (a’, y’) the middle point of the chord; then the
‘quations to the curves become
f+, ¥Y+y=0, P(e +a, y+y) =0,
md the equation to the chord y= mwa; and if in the former
ve substitute ma for y, the resulting equations will give values
or #, being the abscissze of the points of intersection of the
hord and the curves; and if + satisfy one of the equations,
he other must be satisfied by —; therefore the equation to
he locus of the middle points of the chords will result: from
liminating x between
f(@ +a, y+ mex) =0 and d(a' —«, y’- mz) =0.
Suppose the curves to be a hyperbola and its conjugate ;
he result of the elimination will be found to be
4 0'b' (a?m? — b*) = (a? my — Ba’)? (Ba’? — a?y’”).
If the asymptotes be taken for axes, the result will be
ay’ (ma! + y')? + mc =0.
73. ‘To find the locus of the vertex of a triangle, upon a
iven base, and having its vertical angle bisected by a line
arallel to a given line.
Take the given line for the axis of y, the middle point
‘the base (2a) for the origin, and let the angle between them
a; then AM =a coseca, RN =a cosa; (fig. 110) ;
198
|
heat a+acosecca BM BP PR _¥~acoesa |
at - = = = a |
Pn £Y ce meoseca MIC MERC 0 BS. | Yy + @ COS a’
xy = 4a’. sin 2a.
74. 'To find the locus of the intersection of the tangent :
to a given curve, and the perpendicular let fall upon it
from a given point. :
Let y —y’ = tan a (w — 2’) be the equation to the tangent
to a curve at a point (a’, y’); then tana = f(a’, y’) is known:
from the nature of the curve; and the equation to a perpen-|
dicular to the tangent from a point (h, k) is (y—k) tana+a@
—h=0; between which three equations, and the equation to
the curve if a’, y’, and tana be eliminated, we shall obtain the:
required relation between # and y.
Thus the equations to the tangent to a parabola and to a
line perpendicular to it from (h, &) being :
a
il
mM m
the equation to the locus of their intersection is
a(y—k)’+a(a—-h)’+y(y—k) (@-h) =
which, if h=k=0, becomes y’ (a + x) +. v*?=0 the equation
to the Cissoid of Diocles; and if k =0, h = a, it becomes
w fy? + (a@—a)} =0, or x =0,
the equation to the tangent at the vertex.
Similarly, the equation to the tangent to the ellipse being'
Y-Me= / b? + m?a®, the locus of its intersection with a
perpendicular let fall from a point (4, &) has for equation |
y(y—k) +a (a—h) = 4a? (w@—h)?+ B (y —k)*}2;
which, if the perpendicular be dropped from the centre, be-
comes |
(a? + o?)? = aa? + By’,
which agrees with the polar equation already found (Art. 135)
r” = a? (1 — e” sin’).
199
In the case of the hyperbola, changing b? into — 4’, the
equation is
(a? + y?)? = oa ~ 82 Ys
which if 6= a becomes
(a* +")? = a? (a — y’),
representing a curve called the Lemniscata of Bernouilli, and
_ whose polar equation is 7°= a?cos20. If the perpendicular
be let fall from the vertex of a rectangular hyperbola, the
equation is w+ y= a(# — V/ x — y’).
75. If a curve roll upon an equal one, similar points
‘being always in contact, to find the locus of any given point
in the rolling curve.
| Let AP be the fixed, and A’P the rolling curve (fig. 71),
8” the describing point, and |S a point similarly situated in the
fixed curve. Join S'S" meeting the common tangent at P in
-Y; also join SP, S’P. Then because the points in contact
‘are similar, S'P, §’P are equal and equally inclined to the
common tangent PY; therefore PY bisects S18" at right angles,
and therefore the locus of JS” is similar to that of Y, the foot
of the perpendicular from § upon the tangent to the fixed
curve, and S"P is always a normal to the locus of 5S”.
If therefore y= f(a) be the equation to the locus of Y,
ES (5) is the equation to the locus of §’. Hence if the
curves are equal parabolas, and JS’ the focus, its locus will
‘be'’a straight line; if § be the vertex, its locus will be the
Cissoid of Diocles ; if the curves are ellipses, and |S the focus,
fits locus will be a circle; if § be the centre, the equation to
its locus will be
et yar Sara + b*y’.
; au (e@— 3a)
76. ‘To trace the curve 7? & - iy a
C— 44
e a 4
When vw =0, y =0, and limit of Ss a Cs ; therefore at the
origin the curve cuts the axis of # at right angles. When
xv = 3a, y=0, and limit of , when & = 34, is infinity, ©
a
v
therefore the curve again cuts the axis at right angles at a_
distance 3a from the origin. For values of w between 3a and |
4a, y is impossible, and there is no curve; when # = 4a, y is |
infinite; and when 2 is very large, the relation between x and |
y becomes
3 4a\—} a Aa
y= aa ( 1 —-— = ax UD ESS grees 9
L v av
“. y’=a(x+ a) is the equation to the parabolic asymptote, —
above which the curve lies. ‘There is a maximum: ordinate |
2
By taking the limiting value of , it will be found that the
: 3a ee J a |
diameter of curvature at 4 = Tae and similarly at C it will |
be found to equal 3a. ‘There is no part of the curve cor-
responding to negative values of w; and the axis of a is an |
Axis of the curve. Hence the curve is such as is represented
in fig. 111, the dotted line being intended for the parabolic |
asymptote.
77. 'To trace the curve
(y? — 2°) (w — 1) (w - 8) = 2 fy’ + a(w — 2)}?.
Solving the equation relative to y*, we find
ay -§@-Wa- ps8 @- nV Cor Ddoosd,
Hence # must lie between — 51, and 3; and as the rational
part of 2y° = 3(# +4) (25), — 2) nearly, when w = — 5}, the
ordinate is real, and is a tangent to the curve; when w=0,
2y°= 3 or = 0; when #=1, y’=1; and when #=3, we
get a real value for the last ordinate, touching the curve;
which consequently is that represented in fig. 112, having
three true double points, and four double tangents, i. e.
straight lines touching it in two points.
201
7 — a3
78. To trace the curve y= tae
eC—2a
a
When v=0, y= +—~=;
~
when wv = a, the curve cuts the axis of a at right angles; from
“w= a to &= 2a, y remains impossible; when # = 2a, y is in-
finite; and when #w and y become very great, the relation |
between them is i
Loin
a’ eh ies Cod (te Sue it
a alee. Re lgbitae act om, |p cel to tegas Hy
2 2x” BE Sag" Gee Za ii
. + y=a+a is the equation to the asymptotes above which
the curve lies; also when a is negative, y perpetually in-
creases, and there are two infinite branches; therefore the
curve is such as is represented in (fig. 113), the co-ordinates
of the points P, P’ where it cuts the asymptotes being
a ee
V=—-—-, Y= mae @
3 3
79. The corner of a page is turned down so that the |
triangle is of a constant area, a?; the locus of the angular
point is a lemniscata whose equation is 7” =a? sin 20.
80. If two circles be inscribed in another circle touching i
one another, then the area of the circle whose diameter is
their common tangent, will equal the area between the greater
semicircle and the two smaller ones.
81. The equation y* + 2a?xy — v* = 0 expressed by Polar i}
co-ordinates is 7” = a? tan 20.
82. Of the three squares that can be inscribed in an acute- iH
angled triangle, the greatest is that which has two angles in hh
the least side.
83. A parabola is bounded by an ordinate perpendicular i
to its axis, whose length is 6, that of the portion of the
Ss
202
axis cut off being a; D, d are the diameters of the circum-
scribed and inscribed circles, then D+d=a + b.
84. In PG the normal to an ellipse, a point Q is taken
such that PQ = CD, shew that Q traces out a circle.
85. Two conjugate diameters are produced to intersect
the same directrix of an ellipse, and from the point of in-
tersection of each, a perpendicular is drawn to the other ;
these perpendiculars will intersect in the nearer focus.
86. If a pair of conjugate diameters of an ellipse when
produced, be asymptotes to a hyperbola, the point of the
hyperbola at which the tangent will also touch the ellipse,
lies in an ellipse similar to the original one.
87. If two given circles touch one another internally, and
a series of circles whose radii are 7,, 7,, &c. be described
between them touching one another; and if P,, P,, &c. be
the perpendiculars dropped from their centres upon the com-
n
mon diameter of the given circles, then —
Tn
88. If a, b, r, be respectively the radii of the given
circles, and of the first circle in the series, prove that the
radius of the (7+1)™ circle will be
ab(a—b)r
abr + jn (a — b) fr &/ab(a—b—r)t?
89. If two circles touch one another, the radius of any
circle touching them both bears an invariable ratio to the
perpendicular from its centre upon their common tangent. .
90. If the length of the axis of an oblique cone be
equal to the radius of its base, every section perpendicular
to the axis will be a circle.
91. If an ellipse be moved between two straight lines at
right angles to one another, to shew that the centre will
describe a circle, and to find the locus of any given point in
the axis.
203
92. To find the equation to the conic section described
with focus (A, &) and directrix y = ma + e.
93. If SY, HZ be perpendiculars from the foci upon
the tangent at any point P of an ellipse; then §Z and
fITY will intersect in the middle point of the normal at P,
and the locus of their intersection will be an ellipse with
a(1 +e’) and a\/1 —e? for axes,
94. If a parabola be moved between two straight lines
at right angles to one another, the equation to the locus of its
. 4 2 42
vertex will be ay? + y3a3 = a’,
95. The area between two normals to a parabola at the
a 200 :
extremities of a focal chord, and the curve, = —_._—_, 9 being
3 sin® 20
the inclination of one of the normals to the axis.
96. The sum of the squares of the normals to an ellipse
drawn at the extremities of conjugate semi-diameters
= a’ (e® — 1) (e? — 2).
97. Find the locus of the vertex of a triangle whose
base is constant, and likewise the product of the perpendi-
culars dropped from the extremities of the base upon the
line bisecting the vertical angle.
98. If P be a point in a hyperbola, whose ordinate
= BC \/ SC, and CY be a perpendicular from the centre
upon the tangent at P, then PY = SC.
99. If the opposite sides of a hexagon inscribed in a
conic section be produced to meet, the three points of inter-
section will lie ina straight line.
In fig. 114, draw any diagonal MM’, and let the pairs of
opposite sides which pass through its extremities meet in CG;
B; and taking the line CB for the axis of #, let the equation
to the conic section be
Sea as wes SS = ———e Fr -
a Sais we eX Sees
Woe SS eee
204:
ay’+bay+cx°+dy+ex+f=0 (1),
and let the equations to the sides M’M, MN, NN’, N'M’ of
one of the quadrilaterals into which MM’ divides the hexagon,
be respectively
qu+qy=1,
se+sy=l.
pxe+py=1,
ya+ry =1,
Now the equation to the conic section is satisfied by all
such values of w and y as jointly satisfy the equations to any
two adjacent sides of the quadrilateral ; and therefore its equa-
tion, since it is of the second degree, must be of the form
m(px+p'y—1) (re+r y—1) +n(qa+q y—1)(sa+s'y—1)=0,
which compared with (1) gives
mpr+ngs=c, mt+n=f,
m(p+r)+n(q+s)+e=0.
Now if we suppose 4, B, C to be given points, and there-
fore p, g, s to be given quantities, these three equations deter-
mine m, m and 7; therefore D is a fixed point; which shews
that if three sides of a quadrilateral inscribed in a conic section
pass through three fixed points in a given straight line, the
remaining side also will pass through a fixed point in that
line. Consequently, since three sides of the quadrilateral
MOO'M' pass through the points 4, B, C, the remaining side
OO' must also pass through D; therefore the three intersections
of the opposite sides of the hexagon lie in a straight line.
If the hexagon be changed into a triangle, by supposing
every other side to become evanescent; and therefore to assume
the direction of a tangent to the conic section at. one of the
angular points of the triangle, we fall upon Prob. 50.
100. If two pairs of opposite sides of a hexagon inscribed
in aconic section be parallel to one another, the two remaining
sides shall also be parallel to one another.
Let MM’ be any diagonal of any hexagon inscribed in a
conic section having two pairs of opposite sides parallel to one
205
another, and as in the preceding Problem, let the equations to
the four sides of the quadrilateral M’MNN’ be
05 y+qu+q =0,
Y+pet+p
Ytra+?” =0, Ytsu +s =0;
then the equation to the conic section will be
m(ytpxtp’)(ytrot+r) tn(ytqu+g') ytsets’)=0,
which compared with (1) in the preceding Problem, gives
M+ =A, Mpr + NYS = C,
m(r+p)+n(s+q)=b.
These equations shew that if three of the quantities
p, q, 7, 8 be given, the fourth is also constant; i.e. if three
of the sides of any quadrilateral inscribed in a conic section
be parallel to three given straight lines, the remaining side is
also parallel to a fixed line. But if O’M’', OM be respectively
parallel to the lines to which MN, M’N’ are parallel, then the
position of OO" is determined from the above equations by
interchanging gq and. s which does not alter them; therefore
OO’ is parallel to the same line to which NN’ is parallel; or
the two remaining sides of the hexagon are parallel to one
another.
101. The three diagonals of a hexagon circumscribed
about a conic section intersect in a point.
Let aBrydex (fig. 115) be the angular points of a hexagon
circumscribed about a conic section; join the points of con-
tact by straight lines, so forming the inscribed hexagon
ABCDEK ; and produce its opposite sides to meet in P, Q, R.
Then if two ‘tangents were applied at the’points a’, \, in
which the diagonal ad meets the curve, they would intersect in
P; similarly the pairs of tangents applied at the points where
yx, Be, meet the curve, would respectively intersect in Q and
Rk; but P, Q, R, lie in a straight line; therefore the three
diagonals (since they are in the directions of chords joining the
points of contact of pairs of tangents drawn from points in a
straight line) must (Art. 50) pass through the same point.
SECTION XI.
ON CURVES OF THE THIRD AND FOURTH AND HIGHER ORDERS 5
AND ON THE SINGULAR POINTS OF CURVE LINES.
250. Iw this section we shall give some of the principal
results that have been obtained relative to the properties of
curves of the 3rd and 4th orders; and as Pliicker, to whom
the following investigations are chiefly due, has applied his
method to curves of the second order, as well as of higher
orders, we shall commence with that application ; both for
the sake of some new results to which it leads, and for the
purpose of making Pliicker’s general method more readily
understood.
251. ‘The general equation of the second degree
y +2Aavy+ Ba’?+2Cy+2De+H=0, (1),
provided s°+2Az-+B=0 has not equal roots, can alway
be transformed into
(y+an+b)(y+aue+b)+m=0, (2) ;
only when the auxiliary equation has imaginary roots, the
factors of the transformed equation are likewise imaginary,
but their product real.
As equation (2) is of the 2nd degree, and contains the
requisite number of independent constants, we may evidently
assume it to be identical with (1); and upon expanding and
equating coefficients, we find
ata=2A, ad=B;
so that —a, — a’, are the roots of 3? ++2Az+ B=0, and will
be real provided 4?- B>0. ‘To determine b, 6’, and m, we
have |
b+40'=2C, ab’ 4+ab=2D, bb'+m=E, (3)
207
therefore if a, a are real, these equations will evidently
_ furnish a single system of real values of 6, b’, and m. But if
a, @, are imaginary, i.e. if 4? — B<0, the equation
ab'+ab=2D
shews that b, b’ must also be conjugate imaginary roots of a
quadratic equation, and their product consequently real, and
therefore m real; and in this case the two factors of the trans-
formed equation are imaginary, but their product (which
equals — m) is real. Hence the proposed transformation can
be effected, and only in one way; none of the coefficients
being indeterminate, nor having more than one value.
252. Ifa and @’ are equal, then A?— B=0; and cqua-
tions (3) become inconsistent with one another, unless
Dieta AG.
when the two former of them become identical. Hence when
A’-— B=0,
the transformation (2) is impossible, and it may be replaced by
(y+axv+b)’?+m (y+cu4+d)=0,
in which one constant may be assumed at pleasure (since the
general equation, with the condition 4? — B = 0, contains but
four independent constants), and then all the others can be
determined from linear equations.
If from the two former of equations (3) we determine
b, b’, and substitute them in bb'+m=E, we get
m (A* — B) = D?’-2ADC + BC? + E (4? - B);
and if the second member of this equation vanish, the pro-
posed equation resolves itself into two factors of the first
degree, and represents two straight lines; if the second mem-
ber be negative at the same time that A? — B is negative, the
proposed equation cannot be satisfied by any real values of
we and y.
208
253. When the general equation to a curve of the second
order is put under the form
(y+ax+b) (y+a'xr +b’) +m =0,
its two real or imaginary asymptotes have for equations
y+an+b=0, yt+taurs+d'=0.
If with the equation of the second degree, which as
we have just shewn may be written pg + m= 0, where
p and q denote linear functions of # and y each containing
two constants, we combine the equation to any straight line,
we shall usually determine two points of intersection ; but if
° m .
we take the equation p = 0, we get — = 0, which can only be
7
satisfied by supposing # and y to be infinitely great; so that
p =O represents a straight line whose two intersections with
the curve are at an infinite distance, or it is the equation
to one of the asymptotes of the curve; and in the same way
it appears that g=0 is the equation to the other asymptote.
The curve is a hyperbola when the asymptotes are real, and
an ellipse in the contrary case; and instead of being deter-
mined by five constants as in the case when it is referred to
co-ordinate axes whose relation to it is arbitrary, it is, when
represented by the equation pq+m=0, determined by one con-
stant, and two straight lines that bear a fixed relation to it.
254. In the second form to which we have reduced the
general equation of the second order, viz. p?+ mq = 0, the curve
represented is a parabola, and is determined by one constant,
and two straight lines, one of which is arbitrary since one
constant more than necessary enters into the equation. It is
evident that p =0 is the equation to a diameter intersecting
q =0 at a point in the curve; and that g = 0 is the equation
to the tangent at that point, as it leads to p?=0, shewing that
its two points of intersection with the curve coincide with one
another.
255. When we take for the general equation of the
second degree the form
pat+mr = 0,
which contains seven constants, and combine it with either of
209
the equations p=0 or q=0, we get r°= 0; so that the two
straight lines represented by p=0, q=0, are tangents, and
the two points of contact lie in the straight line r=0; and
the curve is in this case determined by any two tangents and
the chord joining the points of contact.
Another form under which the general equation of the
/second degree may be written, is
hed “fs Av" ee [Ls
Uw, Vv, and r being, as before, linear functions of wv and y, each
involving two constants; and in this case each of the lines
u=0, v=0, 7 =0, represents the chord joining the points of
‘contact of the pair of real or imaginary tangents passing
through the intersection of the other two lines.
256. The general equation of the third degree
Yr Ayot+ Bye’ +Ca+Dy+ Eyet Fa?+ Gyt+He+I=0 (1),
provided 3° + 4z?+ Bs +C=0has no equal roots, can always
be transformed into
(ytaxr+b)(yt+a a+b’) (y+a"v+b") +m(y+ex+d)=0 (2);
only when the auxiliary cubic has imaginary roots, two
factors of the transformed equation are likewise imaginary,
but their product real.
| As equation (2) is of the third degree, and contains
the requisite number of independent constants, we ma
evidently assume it to be identical with (1); and upon
expanding and equating coefficients we find
@+a+a"'=A, aa +aa’+a'a" =B, ada’ = G.
to that — a, -—a’,-— a’, are the roots of the cubic equation
s+ A’+ Be+C=0 (3),
which we will suppose to be unequal.
_ For determining 3, b’, b”, we get three equations, which
aay be written
(b+ 0°40" =D, ab" +a" + a(b 4b") + (v7 +a’ )b=E,
: a(a'b" +a"0) +a'a"b = F (4).
14
Hi
HI
i
210
Now if a, a’, a’, be real and unequal, from these equa-
tions since they are linear, a single system of determined |
values of b, 0’, b, can be at once obtained. ‘But if two
roots of (3) a, a’, be imaginary, then a’ 4a’; and a a :
are real; and therefore from the same equations we can get)
one real system of determined values of :
|
ab a bb ro ee saris
|
consequently 6’ and 6” must be Sh Ls imaginary roots
of a quadratic equation since a’, a", are so; therefore bb” |
is real; as is also the product
|
(yt+dat+b)(y+a'a+b"’).
The three remaining constants m, c, and d, are given)
by the equations |
b(b' +b") +. 0b" +m=G, ab’b" + b(a'b’ + a’b)+mec= H
by'b" +md=T, (5); :
|
which, subject only to the condition of a, a’, a’, being
unequal, give one real system of determinate values of m,
c, d._ Hence it is proved that the proposed transformation
can be effected, and only in one way; none of the constants
of the Pehetariied equation being indeterminate, and none
of them having more than one hs |
957. If a’=a", equations (4) become |
b+(b'+b")=D, (at+a’) (b+ b")+20b=E, aa’ (b'+b")+a°b=F,
and cannot coexist unless the equation
Da? - Ea +F=0
is satisfied, in which case they will determine only 6, and
b'+6”; so that the transformation into (2) may be effected
in an infinite number of ways, as one of the quantities b’, b’,
may be assumed at pleasure; and the remaining constants
become known from equations (5). When the equation
Da? —-Ea+F=0 is not satisfied, it is impossible to put
211
the proposed equation into the form (2); and in place of it
we may choose the form
(yt+ax+b) {(y+a'x+b')?41 (y+ax+b)t+m (yt+ew+d) =0,
containing eight constants to which the number of constants
in the proposed is reduced on account of the condition a’= a’;
and by comparing coefficients it will be found that this trans-
formation can be effected in one determined way and no more.
When a= a' =a", equations (4) cannot coexist unless
E=2Da, F = Da’, in which éase they will determine only
the value of b+ 6'+ 5”. Consequently the transformation into
(2) may be effected in an infinite number of ways, as two of
the quantities b, b’, b” may be assumed at pleasure. Under
these conditions, the proposed equation has its independent con-
stants reduced to five; and besides the form (2) may likewise
be made to assume a form involving that number of constants,
viz.
(y+ ax+b)>+m(y+ca+d) =0.
When the conditions E = 2Da, F=Da’, are not satisfied,
it is impossible to put the proposed into the form (2); and
instead of it we may choose the form
(y+axv+ bi +lyt+ge+h)?+m(y+cr+d) <0,
containing one more constant than necessary, as the number
in the original equation is reduced to seven.
If we take the form
(y+au+ bP +i(y + ax +5) (yt+gu+h)+m=o,
containing only six constants, there will be an equation of
condition which we find to be Da?— Ea + F'=0; subject to
which the transformation can be effected in one determined
way and no more.
258, p);
but this last expression will be a maximum when p=2n-3-—p
or 2p =2n —3; and as p must be an integer, p=n — 1 or
p=n—2; and consequently we cannot have x greater than
1
@ (m — 1) (n — 2).
265. A curve of the m‘ order may in general have
m(m—1) tangents drawn to it through a given fixed point,
or parallel to a given line.
Let f(a, y) =0 be the equation to the curve, and
y=ma«-+ec the equation to a straight line; then if m and e
be such that f(«, m# +c) =0 has a real root twice repeated,
that root is the abscissa to the point of contact of the line with
the curve; and (Theory of Equations, Art. 60) is also a root
of f' (x, mx +c) =0; and if between these equations, which
are respectively of and of m—1 dimensions in & and Cc, we
eliminate c, there will result an equation of 2 (m —1) dimen-
sions, whose roots are the abscisse of the points of contact of
all tangents that can be applied to the curve parallel to
Y=mMmx; consequently the number of such tangents will in
general be n(n — 1).
But if the tangents are all to pass through a point (h, &),
then we must eliminate m between
fia, [m(w-h) +k]i =o, f’ Sa, [m (@ —h) + eae
and the result will have for its roots the abscisse of the points
of contact, in number 2(m—-—1) as before. When the given
point is in the curve, the tangent at that point must evidently
be reckoned twice ; and if it be in one of the asymptotes at an
infinite distance, the number of tangents that can pass through
it, or in other words that can be drawn parallel to the asym p-
tote, must be reduced by two units. Hence four tangents can
be drawn to a curve of the third order, parallel to one of its
asymptotes. If the given direction be parallel to the tangent
at a point of inflexion, or the given point be itself a point of
inflexion, the number of ‘tangents that can be drawn, will be
respectively diminished by one and two units.
266. When tangents are applied to a curve of the third
order respectively parallel to its three asymptotes, the points
of contact lie in a straight line.
As we may assume for the general equation to curves of
the third order any equation of the third degree in # and y
with nine independent constants, we may take for it the form
pqr+ms =0 (1);
but we cannot, as in the fundamental form where only the
simple power of s enters, be certain that this transformation
can be effected only in one way. The lines expressed by the
linear functions p, q, 7, s, now bear new relations to the curve ;
for if with equation (1) we combine the equation p= 0, we
get s’=0; therefore the line p=0 meets the curve in two
points only, and those points are coincident ; consequently, its
other point of intersection is at an infinite distance ; so that the
line »=0 is parallel to an asymptote, and also touches the
curve in the point where it intersects the line s = 0. Similarly,
gq = 0 and r = 0 represent lines parallel to the other two asymp;
totes, and touching the curve in the points where they inter-
sect s=0. Now four different tangents may be applied to the
curve parallel to any one of its asymptotes, and the four points
of contact may be joined by straight lines with the points of
contact of four tangents parallel to a second asymptote by six-
teen different straight lines, each of which will pass through
the point of contact of a tangent parallel to the third asymptote.
Since, therefore, there can be sixteen different systems of tan-
gents parallel to the asymptotes with their points of contact in:
a straight line, the general equation of the 3rd degree must be
capable of being put into the form (1) in the same number of
ways.
267. A-curve of the m™ order has in general 3n (n — 2)
points of inflexion.
Let f (a, y) = 0 be the equation to the curve, and
Y=MeL+C
the equation to a straight line; then if m and ¢ be such that,
219
J (v, ma +e) =0 has a real root thrice repeated, that root is
the abscissa to a point of inflexion; and it is (Theory of Equa-
tions, Art. 60) also a root of the equations
f'(#, mv+c)=0, f"(«v, me +c) =0;3
and if between these three equations m and c¢ be eliminated,
there will result an equation whose roots are the abscisse of
points of inflexion.
Now restoring the value of y, the three equations between
which the elimination is to be performed may be written
UW, =.0,
Un, + MV,_1 = 0,
Uy» + 2MW,_5 +m v,_. = 0,
where w,, 2,_,, &c. denote functions of m,n —1, &c. dimen-
sions in w and y; eliminating m between the two latter, we get
an equation of 3m — 4 dimensions, viz.
2 ra) ve
2V a—1 4Wy2Un 1 Un 1 0 Vn—9U a Os
Mm aw
u
and again eliminating y between this and uw, = 0, we finally get
an equation of » ($m —4) dimensions in w Now the curve will
have 7 rectilinear asymptotes, each of which will intersect it in
two points infinitely distant, which points have the character
of points of inflexion in that the radius of curvature at each
is infinitely great; therefore 22 points are included in the
above, which are not proper points of inflexion; and subtract-
ing, we get 37 (n — 2) for the number of points of inflexion
that a curve of the n™ order may in general have.
268. The points of inflexion of a curve of the third
order lie three and three in a straight line.
By the foregoing article a curve of the third order has in
general nine points of inflexion; and we shall now shew that
any straight line passing through two of them, must cut the
curve again in a third. As before, we may assume for the
general equation to curves of the third order, the form
pqr+ms’ =0,
since it contains the requisite number of independent constants.
If with this equation we combine the equation p = 0, we get
s’= 0; which shews that the three points in which the line
20
p =0 cuts the curve, are coincident; consequently the line
p =0 has acontact of the second order with the curve at the
point in which it intersects the line s = 0, and that point is a
point of inflexion. Similarly, each of the lines g=0, r=0
has a contact of the second order with the curve at the point
in which it intersects s = 0; and we thus obtain two other
points of inflexion, lying in the same straight line with the
first. Since out of nine things taken three at a time, twelve
combinations may be formed such that no two of them have
more than one element in common, it follows that the general
equation of the 3rd degree may in twelve ways be put into
the form pqr +ms*=0; but in only one of them will the
linear functions p, q, &c. be real, as only three of the points of
inflexion can be real and six of them imaginary.
269. When a curve of the third order with three real
asymptotes, has a double point, it will fall without, upon,
or within the ellipse which touches in their middle points the
three sides of the triangle formed by the asymptotes, according
as it is a proper double point, a cusp, or a conjugate point.
If wu = 0 be the equation to a curve, then
dint + dyyu. day =05
e e 0 e
and since at a double point d,y = ae at such a point we must
have du = 0, du = 0 (1); and to get the two values of d,y
at the double point, we must have recourse to the second derived
equation which, in consequence of the conditions (1), becomes
2 = g .
din + 2d.) Ay. doy + di, ut . (dy) = 0; (2);
and according as this equation gives real, equal, or imaginary
values of d,y, 1. e. according as
2 2 2
there will be a proper double point, a cusp, or a conjugate
point.
e 7 >
Now, taking the co-ordinate axes parallel to two asymp-~
totes, let
(nw +a)(y+b)\(e-na—y)+m(y+gxrt+h) =0,
or w= py (B—p—_q) +ms =0 be the equation to a curve of
the third order with three real asymptotes, where we have
puta+b+ce= 3; then
di. u =—2n'q, du =-— 2p, din, dU =n(r — p —q)3
therefore the condition (3) becomes
(r— p= 4g) — 4p 9 =p + ge Fr — 2 BS =. one 0.
which shews that a proper double point must fall without, and
a conjugate point within, that ellipse which is the locus of the
possible cusps; and which (Art. 263) touches in their middle
points the three sides of the triangle formed by the asymptotes.
270. But if a series of curves of the third order having
the same asymptotes, have each a double point such that one
of the tangents passing through it is in a constant direction,
then d,y will have a constant value « suppose, which must
satisfy equation (2). Hence substituting for d,y, and for
the differential coefficients of w their respective values, we get
for the locus of the double points the equation
rg+K«(2p+2q—B)+p=0
representing that diameter to whose chords the above-men-
tioned tangent is always parallel.
271. If a curve of the fourth order have three proper
double points, the six tangents at those points all touch a conic
section.
Since the curve has three double points, if p=0, g =0,
7 = 0, be the equations to three straight lines passing through
every two of them, and in the equation to the curve we
suppose p to vanish, qg’7* must appear as a factor; simi-
larly when g and vr are supposed to vanish, pr? and p 9
must appear respectively as factors; therefore the equation,
since it is of the fourth degree, must be of the form
epg + bp’r*+ ag’ — 2pqr (p+ q+er)=0 (1).
Now q = Xr represents any straight line passing through
the double point g=0, r=0; and if we combine it with
the equation to the curve and reject the factor * we obtain
for result
rar’ — 2pr (bd? + cr) + p? (ch? —2a’d +b) =0;
SS
Sa
7 ———