7 Posts • Page 1 of 1
 |
 |
Author |
 |
Message |
 |
 |
 |
 |
 |
 |
yetti
Navier-Stokes Equations

Offline Joined: 01 Jan 2005 Posts: 2108 Location: New York
|
Bicentric quadrilateral me
A quadrilateral with the diagonal intersection is tangential iff the incenters of the triangles are concyclic. If the quadrilateral is cyclic, the incenters of the triangles make a rectangle, i.e., they are also concyclic. Prove that the quadrilateral is bicentric iff the incenters of all 8 triangles are concyclic and the iff the excenters against the vertex for the first 4 triangles and the excenters against the vertices for the last 4 triangles are concyclic.
|
_________________ Carthage must be destroyed.
Posted: Mon Jan 10, 2005 10:41 am |
 |
|
|
 |
 |
 |
 |
 |
 |
 |
yetti
Navier-Stokes Equations

Offline Joined: 01 Jan 2005 Posts: 2108 Location: New York
|
Nobody bites into it, so here is the solution:
Let b the tangency points of the incircle with the quadrilateral sides . There are numerous proofs that the incenters of the triangles and the excenters of these triangles against their common vertex are concyclic iff the quadrilateral is tangential. Let be the incenters of the triangles and the excenters of these triangles against the vertices . It is also fairly well known that in a tangential quadrilateral , the lines concur at a point on the diagonal and the lines concur at a point on the diagonal . Similarly, the lines concur at a point on the diagonal and the lines concur at a point on the diagonal . In addition, since the incircles and the excircles touch each other on the diagonal , the lines and are perpendicular to this diagonal. Similarly, the the incircles and the excircles touch each other on the diagonal and the lines and are therefore perpendicular to this diagonal. It is also well known that the lines , intesect at the circumcenter of the cyclic quadrilateral and similarly, the lines , intesect at the circumcenter of the cyclic quadrilateral .
Again, there are numerous proofs that in a cyclic quadrilateral , the incenters and the excenters form rectangles and as such, they are also concyclic. The rectangle diagonals intersect at the circumcenter of the rectangle and the rectangle diagonals intersect at its circumcenter of the rectangle . It follows that in a bicentric quadrilateral , the circumcircle of the cyclic quadrilateral and the circumcircle of the rectangle have the common circumcenter . Likewise, the circumcircle of the cyclic quadrilateral and the circumcircle of the rectangle have the common circumcenter . The whole point is to show that the first pair of circumcircles has the same radius and the second pair of circumcircles also has the same radius. One way to do this is to prove the incenters and the excenters to be concyclic in a bicentric quadrilateral in a special way, which yields similarity of these to quadrilaterals with the cyclic quadrilateral of the tangency points.
Step 1: The opposite sides and the opposite sides of the cyclic quadrilateral meet at points on the diagonals and the diagonals of meet at the same point as the diagonals of . The proof is by the repeated use of Pacsal's theorem for the 2 diagonals and 2 tangents of the cyclic quadrilateral , showing the 4 points and the 4 points to be collinear. Since the vertices are the poles of the lines , the points are poles of the lines .
Step 2: In an inversion in the incircle , the vertices of the bicentric quadrilateral are carried into the midpoints of the sides of , thus forming a parallelogram with the sides and parallel to the diagonals of . Since this parallelogram is inscribed in the inverted circumcircle of the bicentric quadrilateral , it is a rectangle and consequently, the diagonals are perpendicular to each other.
Step 3: Let be the intersections of the line with the opposite sides and the intersections of the line with the opposite sides , forming another quadrilateral with the diagonals intersecting at the point . In a polarity transformation WRT to an arbitrary circle centered at the diagonal intersection , the quadrilaterals and are carried into parallelograms and , where and . Since the diagonals are perpendicular to each other, the paralleogram is a rectangle. Since is inscribed in , the rectangle is inscribed in the parallelogram . Since , , the lines are polars of the points WRT . Since the points and the points are collinear with the center of the polar circle , their polars and are parallel. The rectangle diagonals are thus the middle lines of the parallelogram and conversely, the parallelogram diagonals are the middle of the rectangle . Hence, the parallelogram has perpendicular diagonals and it is a rhombus. The opposite sides and meet at the poles of the rhombus diagonals . Since these rhombus diagonals are parallel to the rectangle opposite sides , , the points and the points are collinear with the center of the polar circle . Hence, the diagonals pass through the points . The rectangle diagonals are perpendicular to the lines , identical with the diagonal lines , connecting their poles with the center of the polar circle . Similarly, the rhombus diagonals are perpendicular to the lines , identical with the diagonal lines , connecting their poles with the center of the polar circle . Because of the symmetry of the transformed configuration, the rhombus diagonals are bisectors of the angles between the rectangle diagonals. But this also means that the diagonals of the quadrilateral are bisectors of the angles between the diagonals of the quadrilateral . Since the diagonal lines are identical with the diagonal lines and since the diagonals of the quadrilaterals , are bisectors of the angles between the diagonals by definition, the diagonal lines of the quadrilaterals and , are identical.
Step 4: The triangles have the side is common and the sides equal, being the tangent lengths to the incircle from the vertex . The sides are therefore internally cut by the internal bisectors of the angles in the same ratio. Using the same argument for the triangle pairs , , ,
Accordingly, the quadrilateral is centrally similar to the cyclic quadrilateral with homothety center and coefficient . The circumcircle of the quadrilateral is centrally similar to the circumcircle of the quadrilateral and the poles of the diagonal lines WRT the circle are centrally similar to their poles WRT the circle with the same homothety center and coefficient. The center of the circumcircle is then the intersection of normals dropped from the poles to their polars . Using the fact that an external angle bisector cuts the opposite side of a triangle externally in the ratio of the two adjacent sides, we can similarly show that the triangle excenters quadrilateral is centrally similar to the cyclic quadrilateral with homothety center and coefficient and that the center of its circumcircle is the intersection of normals dropped from the poles to their polars .
Step 5: Let , be the circumcircles of the rectangles and , with the same centers as the circumcircles of the cyclic quadrilaterals and , but possibly different radii. The centers of the circles are collinear. Let be the radii of these circles and assume that Since the circles touch each other on the diagonal and since the circle passes through their centers , the radius is equal to
The common external tangents of the circles meet at the point , their external homothety center. Accordingly,
. and .
. and .
This implies that the point is the external homothety center of the circle pairs and as well; the 3 circles have a single pair of the common external tangents meeting at the point . In exactly the same way, we can show that the 3 circles also have a single pair of the common external tangents meeting at the point . If the circles or happen to have the same radii (i.e., the quadrilateral is a bicentric kite), the circle also has the same radius and the corresponding homothety center or moves to infinity on the diagonal line or .
Step 6: Since the common external tangents of the circle pairs and meet at the points , the external homothety center of the circles is on the diagonal . Since the common external tangents of the circle pairs and meet at the points , the external homothety center of the circles is also on the diagonal . This homothety center is then identical with the intersection of these two diagonals. But the external homothety center of the circles is also at the point . In addition, the circumcircles of the cyclic quadrilateral and of the rectangle have the common center . Consequently, these two circles are identical, . In exactly the same way, we can show that the circumcircle of the cyclic quadrilateral is identical with the circumcircle of the rectangle .
It remains to be shown that if the quadrilateral is not bicentric, the incenters and the excenters are not concyclic. The proof that the incenters and the excenters cannot be concyclic in a quadrilateral , which is not tangential, is well known. Perhaps, it is necessary to show that the quadrilaterals , cannot be cyclic unless the quadrilateral is also cyclic (and then they are necessarily rectangles). The proofs I have seen demonstrated only that , are rectangles in a cyclic quadrilateral, but not that they cannot be cyclic in some other quadrilateral.
Step 7: Assume that the quadrilateral is not cyclic, for example, the sum of the opposite angles . The vertex then lies outside of the circumcircle of the triangle and the vertex lies outside of the circumcircle of the triangle . Let the vertex move along the diagonal into the intersection of this diagonal with the circle . The incenter is fixed and the incenters move toward the vertex along the bisectors of the fixed angles into new positions . Since the quadrilateral is now cyclic, the angle is necessarily right. But since the incenter lies on the common bisector of the angles inside of the angle , the original incenters both lie inside of the right angle and consequently, the original angle . For exactly the same reasons, the angle as well. The sum of the opposite angles of the quadrilateral is therefore less than , which means that this quadrilateral cannot be cyclic.
|
_________________ Carthage must be destroyed.
Posted: Sun Mar 06, 2005 1:46 pm |
 |
|
|
 |
 |
 |
 |
 |
 |
 |
Beat
Yang-Mills Theory


Offline Joined: 12 Nov 2005 Posts: 662
|
tangential, excentre, concyclic. Can somebody explain me these words, please.
The problem looks hard and interesting and i cannot translete a bit of it. Thank you
|
Posted: Mon Dec 05, 2005 9:29 am |
 |
|
|
 |
 |
 |
 |
 |
 |
 |
Virgil Nicula
Birch & Swinnerton Dyer

Offline Joined: 22 Jun 2005 Posts: 4549 Location: Bucuresti (RO) Bradenton (FL)
|
Another property Classic (Alexei Zaslawsky)
Let be a convex tangential (circumscribed) quadrilateral to the circle . I note the intersection and the incircles , , , of the triangles , , , . We know, from the Yetti's posted problem, that the quadrilateral is cyclic. Prove that the tangents in the vertexes of the quadrilateral to the its circumcircle form a quadrilateral with two vertexis on the line and with another vertexis on the line (Alexei Zaslawsky).
|
Posted: Mon Dec 05, 2005 5:59 pm |
 |
|
|
 |
 |
 |
 |
 |
 |
 |
mecrazywong
Yang-Mills Theory

Offline Joined: 07 Jul 2004 Posts: 605 Location: Hong Kong
|
Re: bicentric quadrilateral me
| yetti wrote: |
A quadrilateral with the diagonal intersection is tangential iff the incenters of the triangles are concyclic. If the quadrilateral is cyclic, the incenters of the triangles make a rectangle, i.e., they are also concyclic. Prove that the quadrilateral is bicentric iff the incenters of all 8 triangles are concyclic and the iff the excenters against the vertex for the first 4 triangles and the excenters against the vertices for the last 4 triangles are concyclic.
|
The first two results are actually well known. As long as I remember, they appeared in Chinese TST, APMO, Crux, and much more...
Anyway, nice results for the sufficient and necessary condition of a quadrilateral being bicentric 
|
Posted: Mon Dec 05, 2005 6:24 pm |
 |
|
|
 |
 |
 |
 |
 |
 |
 |
yetti
Navier-Stokes Equations

Offline Joined: 01 Jan 2005 Posts: 2108 Location: New York
|
Re: another property Classic (Alexei Zaslawsky)
| Virgil Nicula wrote: |
Let be a convex tangential (circumscribed) quadrilateral to the circle . I note the intersection and the incircles , , , of the triangles , , , . We know, from the Yetti's posted problem, that the quadrilateral is cyclic. Prove that the tangents in the vertexes of the quadrilateral to the its circumcircle form a quadrilateral with two vertexis on the line and with another vertexis on the line (Alexei Zaslawsky).
|
I have seen a simple proof (by Darij) using poles and polars, but I prefer my original proof of the fact that the incenters are concyclic, because it is so illuminating. It also yields the above proposition as a side effect. Only later I figured out the simple angle chase leading to the cyclic quadrilateral condition (see Quadrilateral (incenters form a cyclic quadrilateral)).
Let ABCD is an arbitrary tangential quadrilateral with the diagonal intersection X and the incircles of the triangles . Assume that the incircle is the largest one, let be the intersection of the other tangent to parallel to the diagonal BD (other than BD) with the diagonal AC and let be the intersection of the other tangent to parallel to to the diagonal AC (other than AC) with the diagonal BD. If A' is an arbitrary point on the diagonal ray , then we can draw tangents from A' to the circles (other than AC) intersecting BD at points B', D'. Then we can draw tangents from B', D' to the circles , respectively, and these tangents will intersect at a point C' on the diagonal AC, forming a tangential quadrilateral A'B'C'D' with the same diagonal lines and with the same incircles of the triangles as the original quadrilateral ABCD.
Starting with the point , the vertex B' is at infinity on BD (on the ray ). If we move A' to infinity on AC (on the ray ), then B' moves to . Assuming that B' is at infinity, the angle , while assuming that A' is at infinity, the opposite is true - the angle . Therefore, a position of the point exists, such that . But then these 2 angles subtend a common arc on the chord A'B', which means that the tangential quadrilateral A'B'C'D' is also cyclic, i.e., bicentric. But if A'B'C'D' is bicentric, its contact quadrilateral E'F'G'H' is centrally similar to the fixed quadrilateral with the homothety center X (the diagonal intersection). Since E'F'G'H' is cyclic, is also cyclic. Moreover, the tangential quadrilateral of the cyclic quadrilateral is centrally similar to the bicentric quadrilateral A'B'C'D' with the same homothety center X, hence, and .
All that remains to be shown is the claim that the quadrilateral A'B'C'D' has the same diagonals as the quadrilateral ABCD (and remains tangential) when moving the starting point A' in any way on the ray - this is quite long. Since ABCD is tangential with an incircle (I), the incircles of the triangles touch at a point and the incircles of the triangles touch at a point . Using the collinearity of 3 external homothety centers of the circle triples , the external homothety center K of lies on AC and similarly, the external homothety center L of lies on BD. Similarly, BK is the external homothety axis of the both circle triples and , hence, the external homothety center of lies on BK and since AC is their common external tangent, this homothety center is the intersection . Likewise, the external homothety center of the circles is and the extenal homothety center of both and is . The center lines passes through K and the center line passes through L. Moreover, XM is equal to the common internal tangent lengths of and , while XN is equal to the common internal tangent lengths of and (the proof is virtually the same as in A very nice problem: constant sum, where XM = 0).
Now erase the tangential quadrilateral ABCD with its incircle (I), as well as the triangle incircles , but leave the diagonal lines with the homothety centers K, L, the normals and the circles . Let be arbitrary, let the tangents from A' to the circles other than p meet the line q at B' D', respectively, and let the tangents from B', D' to the circles other than q meet the line p at , respectively. Let be the incircle of the triangle touching the line q at the point M (because XM equal to the common internal tangent lengths of and is fixed). Let be the incircles of the triangles . The circlles touch each other at the point (again, because XN equal to the common internal tangent lengths of and is fixed). Let be a circle touching the line q at the point M and also touching the line . Similarly, let be a circle touching the line q at the same point M and also touching the line .
The external homothety axis of both the circle triples and is the line B'K, hence, the external homothety center of the circles is the intersection of this homothety axis and their center line . Likewise, the external homothety axis of both the circle triples and is the line D'K, hence, the external homothety center of the circles is the intersection of this homothety axis and their center line . But this implies that the circles are identical, both touching the circle at the same point M and both having the same external homothety center K with this circle. Keep the label for this circle. The common external tangents of the circles meet at the point , the common external tangents of the circles meet at the point and the common internal tangent length of the circles on the line p is equal to XM. It follows that the common external tangents meet on the line p, i.e., the points are identical and the quadrilateral A'B'C'D has the same diagonal lines p, q as the original tangential quadrilateral ABCD. Since the triangle incircles touch each other at the point and the triangle incircles at the point , the new quadrilateral A'B'C'D' remains tangential.
|
_________________ Carthage must be destroyed.
Posted: Tue Dec 06, 2005 5:24 am Last edited by yetti on Wed Dec 07, 2005 5:08 pm; edited 1 time in total |
 |
|
|
 |
 |
 |
 |
 |
 |
 |
yetti
Navier-Stokes Equations

Offline Joined: 01 Jan 2005 Posts: 2108 Location: New York
|
| Beat wrote: |
tangential, excentre, concyclic. Can somebody explain me these words, please.
The problem looks hard and interesting and i cannot translete a bit of it. Thank you
|
Concyclic points are points lying on a single circle. Every 3 points lie on a single circle, hence, the term is used only if 4 or more points lie on a single circle.
Tangential quadrilateral is a quadrilateral (convex of concave), whose side lines touch a single circle. This single circle can be centered inside the quadrilateral (incircle with an incenter) or outside the quadrilateral (excircle with an excenter). Tangential quadrilateral is also called inscriptable. Compare with a triangle, which always has one incircle and 3 excircles.
Any other terms ?
Yetti
|
_________________ Carthage must be destroyed.
Posted: Tue Dec 06, 2005 5:44 am |
 |
|
|
 |
 |
 |
 |
|
 |
 |
7 Posts • Page 1 of 1
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You can download files in this forum You cannot post calendar events in this forum
|