Full text: The collected mathematical papers of Arthur Cayley, Sc.D., F.R.S., sadlerian professor of pure mathematics in the University of Cambridge (Vol. 6)

414] 
ON POLYZOMAL CURVES. 
531 
151. Consider the case of the bicircular quartic, and take as before (£ = 0, 2 = 0), 
and (77 = 0, z = 0) for the coordinates of the points I, J respectively. Let the two 
tangents from the focus A be f — az = 0, 77 — a'z = 0, say for shortness p = 0, p' = 0, 
then the equation of the curve is expressible in the form pp'U=V 2 ( 1 ), where TJ= 0, 
V = 0 are each of them a circle, viz., U and V are each of them a quadric function 
containing the terms z 2 , zi7, zg, and £77. Taking an indeterminate coefficient A, the 
equation may be written 
pp' (U + 2\V+ X 2 pp') = (V + Xpp') 2 , 
and then A may be so determined that U + 2XV+X 2 pp'= 0, shall be a 0-circle, or 
pair of lines through I and J. It is easy to see that we have thus for X a cubic 
equation, that is, there are three values of A,, for each of which the function 
U + 2AF 4- X s pp' assumes the form (£ — /3z) (77 — fi'z), = qq' suppose: taking any one of 
these, and changing the value of V so as that we may have V in place of V+ Xpp', 
the equation is pp'qq' + V 2 , where V = 0 is as before a circle, the equation shows that 
the points of contact of the tangents p = 0, p' = 0, q = 0, q' = 0 lie in this circle V=0. 
The circumstance that X is determined by a cubic equation would suggest that the 
focus q = 0, q' = 0 is one of the three foci B, C, D concyclic with A ; but this is 
the very thing which we wish to prove, and the investigation, though somewhat long, 
is an interesting one. 
152. Starting from the form ppqq' = V 2 , then introducing as before an arbitrary 
coefficient X, the equation may be written 
pp (qq + 2XV + X 2 pp') = ( V+ Xpp') 2 , 
and we may determine A, so that qq'+ 2XV + X 2 pp'= 0 shall be a pair of lines. 
Writing V = Hgr) — Lrjz — Lljz + Mz 2 , and substituting for pp and qq their values 
(f — oz)(t7 — a'z) and (£ — ¡Hz) (77 — fi'z), the equation in question is 
(1 + 2XH + A, 2 ) £77 - (/3 + 2XL + X 2 *) 77z - (/3' + 2XL' + AV) %z + (/3/3' + 2Ailf + AW) * 2 = 0, 
and the required condition is 
(1 + 2XH+X 2 ) (/3/3' + 2XM + A 2 W) = (/3 + 2A L + A 2 «) (/3' + 2A L' + A 2 «); 
or reducing, this is 
(:2M + 2H/3/3’ - 2L'/3 - 2Z/3) 
+ A ((a — /3) (a' — /3') + \HM— 4LL') 
+ A 2 (2M + 2Hact! - 2L'a - 2La!) = 0, 
viz., A is determined by a quadric equation. Calling its roots A 1( and A 2 , the foregoing 
equation, substituting therein successively these values, becomes (£ — 72) (77 — y'z) = 0, and 
(£ — 8z) (rj — 8'z) = 0 respectively, say rr = 0 and ss' = 0. 
1 This investigation is similar to that in Salmon’s Higher Plane Curves, p. 196, in regard to the double 
tangents of a quartic curve. 
67—2
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.