Let $ABC$ a triangle in the plane, but $D$ a point in (R3) space, such that the angles $\phi=ADB=BDC=CDA$ are equal. Let $E$ be the footpoint of $D$ in $ABC$. $E(\phi)$ describes a (irreducible) septic $S$ (that has 5 asymptotes, for the record).
Points that lie on $S$ (checked with ETC): Incenter $I$, Circumcenter $U$, Orthocenter $H$ (isolated), $A,B,C$ (double points), 1st (Fermat $F$) and 2nd isogonic center $D$. septic closeup
It's near-lying to assume that $S$ is the Darboux septic (the only I heard of :-) of some triangle $XYZ$ where $XYZ$ is the foobar transform of $ABC$ but with that septic I'm still skeptic. (SCNR)
Can you help? Gibert's invaluable site went kaput, and I don't even find the original paper to the Darboux septic.

  • $\begingroup$ Bernard Gibert's site has been archived on the Wayback Machine: web.archive.org/web/20190321092028/http://… . Specifically for septics, see web.archive.org/web/20171212225952/http://… . Have you tried checking the isogonal, cyclocevian etc. conjugates of your points, to see whether they lie on a simpler curve? Also, is there an equivalent condition that "stays inside the plane"? $\endgroup$ Oct 31, 2019 at 21:14
  • $\begingroup$ Ah, the Wayback Machine. Already played with the though of looking there but was too pessimistic. :-) The description already should rule it out - Darboux' septic has triple points. $\endgroup$ Nov 1, 2019 at 14:39
  • $\begingroup$ For what it's worth, Gilbert's site is alive and well. bernard-gibert.pagesperso-orange.fr $\endgroup$
    – brainjam
    Jan 9, 2021 at 4:25

1 Answer 1


This is just a back of an envelope calculation so not really an answer, but too long for a comment. I hope it helps: we work in $(r,s,t)$ space ($p$ and $q$ are constants determined by the shape of the triangle). Then your curve is the projection onto the $(r,s)$ plane of the algebraic variety $$p_1(r,s,t)=p_2(r,s,t), p_3(r,s,t)=p_4(r,s,t),$$ where $$p_1(r,s,t)=(r^2-r+s^2+t^2)^2((r-p)^2+(s-q)^2+t^2),$$ $$ p_2(r,s,t)=((r-1)(r-p)+s(s-q)+t^2)^2(r^2+s^2+t^2), $$ $$ p_3(r,s,t)=(r^2-r+s^2+t)^2((r-p)^2+(s-q)^2+t^2) $$ and $$p_4(r,s,t)= ( r^2-rp+s^2-sq+t^2 )^2(( r-1 )^2 +s^2+t^2 ) .$$ Perhaps Gröbner bases can be used to tease out what you want.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.