203. An application of the Pascal's permutation theorem.
by Virgil Nicula, Jan 1, 2011, 8:01 AM
Proposed problem. Let
be a triangle. Consider the points
,
and
such that
,
,
. Prove that the points
,
,
are collinearly.
Proof 1
Proof 2.
PP1. Let
be a triangle with the circumcircle
. Let
be an interior point of the
and denote
.
Also denote 6 points such as
. Prove that
.
Proof. Apply the Pascal's theorem to the following cyclical quadrilaterals :
.
PP2. Let
be a triangle. Consider two points
and
. For a point
from the
plane of
denote
and
. Prove that
.
Proof. Apply the Pappus' theorem to the lines
and
.










Proof 1
Proof 1 (metric). Observe that
. I"ll use the Menelaus' theorem for
and the following transversals :
. Thus, product of the relations
,
,
and
.









Proof 2.
Proof 2 (proiective). Observe vthat
. Hence the points
,
and 
are infinite points, so are collinearly (all lie on the line at infinity). Therefore, the points
is Pascalian. Thus, by the Pascal permutation theorem,
the permutation
of this hextuple must also be Pascalian. Hence the points
,
and 
are collinearly. But
,
and
. In conclusion, the points
,
and
are collinearly.




are infinite points, so are collinearly (all lie on the line at infinity). Therefore, the points

the permutation




are collinearly. But






Darij Grinberg wrote:
Definition. Let's call the points
from same plane is a Pascalian sextuple iff the points
,
and
are collinearly.
Pascal's permutation theorem. Prove that if a sextuple of points is Pascalian, then each permutation of this sextuple is Pascalian.
This is, of course, trivial using the fact that a sextuple of points is Pascalian if and only if the six points lie on one conic. But how about a proof
without using conics? There is one, and I find it quite nice. I first prove the following lemma, which is, I believe, called "Carnot's theorem":
Lemma. Let
be a triangle, let
and
be two points on its sideline
, let
and
be two points on its sideline
, and let
and
be two points
on its sideline
. Then, the points
,
and
are collinear iff
.
Proof of Lemma. Let
,
and
. Then, we have to show that the points
,
,
are collinearly iff
. But since the points
,
,
lie on the sidelines
,
,
of triangle
, the Menelaos' theorem
yields that they are actually collinear iff
. So all we have to do is to prove that the equations
and
are equivalent. Indeed, this will be trivial once it will be shown that we always have the identity
. Thus, in order to prove Lemma, it will be enough to establish this identity. Well,
since the points
,
and
on the sidelines of triangle
are collinear, the Menelaos' theorem yields
. Similarly, the collinearity of
the points
,
and
yields
, and the collinearity of the points
,
and Z yields
. Multiplying these
three equations, we get
or equivalently,
. But it is exactly required identity and thus the proof of Lemma is complete.
Now, in order to solve the initial problem, I consider a Pascalian sextuple of points
in the same plane. The problem requires me to prove that every
permutation of this sextuple is Pascalian, but, noting - like you did - that any permutation can be written as a composition of transpositions of adjacent elements, I see that it is
enough to show that every sextuple resulting from the sextuple
upon a transposition of adjacent elements is Pascalian. Since it doesn't make any
difference whether we transpose
with
, or
with
, or
with
, or
with
, or
with
(in fact, the definition of a Pascalian sextuple is cyclically invariant),
we will restrict ourselves to considering only the transposition of
with
. In other words, we have to show that the sextuple of points
is Pascalian.
But this is just a matter of applying Lemma : Let
,
and
. Then, the points
and
lie on the sideline
of triangle
, the points
and
lie on its sideline
and the points
and
lie on its sideline
. Since the sextuple of points
is Pascalian, the
points
,
and
are collinearly; in other words, the points
,
and
are collinearly. Thus, after
Lemma, we have
. But this can be rewritten in the form
. And thus, using
Lemma again, we see that the points
,
and
are collinearly. In other words, the points
,
and 
are collinearly. This yields that the sextuple of points
is Pascalian. And this completes the solution of the problem.




Pascal's permutation theorem. Prove that if a sextuple of points is Pascalian, then each permutation of this sextuple is Pascalian.
This is, of course, trivial using the fact that a sextuple of points is Pascalian if and only if the six points lie on one conic. But how about a proof
without using conics? There is one, and I find it quite nice. I first prove the following lemma, which is, I believe, called "Carnot's theorem":
Lemma. Let









on its sideline





Proof of Lemma. Let














yields that they are actually collinear iff




since the points





the points







three equations, we get


Now, in order to solve the initial problem, I consider a Pascalian sextuple of points

permutation of this sextuple is Pascalian, but, noting - like you did - that any permutation can be written as a composition of transpositions of adjacent elements, I see that it is
enough to show that every sextuple resulting from the sextuple

difference whether we transpose










we will restrict ourselves to considering only the transposition of



But this is just a matter of applying Lemma : Let














points






Lemma, we have


Lemma again, we see that the points






are collinearly. This yields that the sextuple of points

PP1. Let





Also denote 6 points such as


Proof. Apply the Pascal's theorem to the following cyclical quadrilaterals :


PP2. Let




plane of




Proof. Apply the Pappus' theorem to the lines


This post has been edited 23 times. Last edited by Virgil Nicula, Nov 22, 2015, 4:41 PM