Bicentric Quadrilaterals Through Inversion

b
Forum Geometricorum
Volume 13 (2013) 11–15. b b
FORUM GEOM
ISSN 1534-1178
Bicentric Quadrilaterals through Inversion
Albrecht Hess
Abstract. We show that inversion is a delightful tool for making some recent
and some older results on bicentric quadrilaterals more transparent and to smoothen
their proofs. As a main result we give an illustrative interpretation of Yun’s in-
equality and derive a sharper form.
1. Introduction
Figure 1 shows a bicentric quadrilateral ABCD, its circumcircle C with center
O and radius R, and its incircle C with center Z and radius r, OZ = d. The sides
of ABCD are tangent to C at E, F , G, H. Apply an inversion with respect to C .
D
C
C
G
D′
C
C′
H
C′
R′
d S
O Z M F
R
A′
r
B′
A E B
Figure 1
The images A′ , B ′ , C ′ , D′ of the vertices lie on the circle C ′ with center M and
radius R′ , M Z = d′ . The image A′ lies on the polar of A with respect to C and is
therefore the midpoint of EH. The same applies to the other images. A′ B ′ C ′ D′
is a rectangle, because being the quadrilateral of the midpoints of EF GH it is a
cyclic parallelogram. The diagonals EG and HF are orthogonal, since they are
parallel to the sides of A′ B ′ C ′ D′ . Cf. [14, step 2] and [7, 837 ff.].
Publication Date: January 29, 2013. Communicating Editor: Paul Yiu.

12 A. Hess
2. Orthogonality of Newton lines

Theorem 1 (9, Theorem 6). A tangential quadrilateral ABCD - without axes of
symmetry - is cyclic if and only if its Newton line is perpendicular to the Newton
line of its contact quadrilateral.
The restriction is included, since Newton lines do not exist in tangential quadri-
laterals with several axes of symmetry, for isosceles tangential trapezoids the the-
orem is obvious and it is false for kites. Let I and J be the points of intersection
of AB and CD, respectively. BC and AD. The midpoints MAC , MBD , MIJ are
collinear in any quadrilateral. The line passing through these points is called the
Newton line. To prove the collinearity one could use barycentric coordinates. For
a visual proof, connect some midpoints of the quadrilateral sides and the appearing
parallelograms will guide you. More information about Newton lines can be found
in [1, pp. 116–118] and [2]. The points X on the Newton line have a special prop-
erty: The sum of the signed areas of AXB and CXD equals the sum of the signed
areas of AXD and BXC. This can be seen easily from the equivalence of both
−−−−→ −−
→ −−→ −−−−→ −
−→ −−→
XMAC × (AB + CD) = 0 and XMBD × (AB + CD) = 0
to
−−→ −−→ −−→ −−→ −−→ −−→ −−→ −−→
XA × XB + XC × XD = XD × XA + XB × XC.
If ABCD is a tangential quadrilateral its consecutive sides a, b, c and d satisfy
a + c = b + d, and therefore the center Z of its incircle share the property that
the sum of the areas of AZB and CZD equals the sum of the areas of AZD and
BZC. Hence Z belongs to the Newton line.
Proof of Theorem 1. Suppose that the Newton line of ABCD, i.e., the line n1
through MAC , Z, MBD , MIJ , and the Newton line of EF GH, i.e., the line n2
through MEG , MF H , are perpendicular. Apply the inversion with respect to the
incircle C. The images of I and J of MEG and MF H lie on the image of n2 , which
is a circle through Z orthogonal to n1 , whose center lies on n1 . If MIJ ∈ n1 is
not the center of this circle, then I and J are symmetrical with respect to n1 and
ABCD is a kite, which was excluded. Hence MIJ is the center of the image of
n2 , ∠IZJ = 90◦ , EG ⊥ F H, A′ B ′ C ′ D′ is a rectangle and ABCD cyclic. This
argument can be reversed easily.
3. Fuss’ formula
We derive Fuss’ theorem (cf. [3], [7, 837 ff.], [8, Theorem 125], [11, p.1],) by
inversion. I found no other place in literature, except the quoted book [7], where
Fuss’ theorem is proved with inversion. But the calculations in F. G.-M.’s book are
somewhat cumbersome.
Observe - with Thales’ theorem or angle chasing - that B ′ SD ′ Z is a parallelo-
gram. M being the midpoint of ZS, the parallelogram law says
4R′2 + 4d′2 = 4M D ′2 + 4d′2 = 2ZD′2 + 2SD ′2 = 2r 2 .
Bicentric quadrilaterals through inversion 13
H MIJ
G
MF H
MBD
C
O Z F
MEG
MAC
A E B I
Newton’s line
Newton’s line
Figure 2
r2 R
The formulae for radius and midpoint distance of an inverted circle R′ = R2 −d2
r2 d
and d′ = R2 −d2
, [8, p. 51], substituted into 2R′2 + 2d′2 = r 2 lead to Fuss’ formula
1 1 1
2
+ 2
= 2.
(R − d) (R + d) r
4. Poncelet’s porism
Theorem 2. ABCD is a bicentric quadrilateral with circumcircle C and incir-
cle C . Then bicentric quadrilaterals with circumcircle C and incircle C can be
constructed starting from any point of the circumcircle C (cf. [4], [6], [12], [13]).
Proof. If ABCD is bicentric (see Figure 1), R, r and d obey Fuss’ formula. Using
inversion with respect to C , the circumcircle C of ABCD is mapped onto the circle
C ′ with center M and 2R′2 + 2d′2 = r 2 , just reverse the substitutions above. Let S
be a point such that M is the midpoint between the center Z of the incircle and this
point S. Choose any point A′ on C ′ . This point A′ and its diametrically opposite
point C ′ form with Z and S a parallelogram. From the parallelogram law follows
that A′ is the midpoint of a chord HE of C which forms together with S a right
triangle. G and F are the endpoints of the chords from E and H through S and
B ′ , C ′ and D′ the midpoints of the corresponding chords. Inversion with respect
to C converts the circles with diameters ZE, ZF , ZG, ZH into the sides of the
bicentric quadrilateral whose vertices are the images of A′ , B ′ , C ′ and D′ .
14 A. Hess
5. Carlitz’ inequality
′2 + 2d′2 = r 2 we get
√ ′
Furthermore, from
√ 2R 2R ≤ r. Substituted into R =
r 2 R′ r2
R′2 −d′2 ≥ R′ ≥ 2r, Carlitz’ inequality [5] is obtained.
6. Coaxial system of circles

′2 2
Writing 2R′2 + 2d′2 = r 2 as Rd′ + d′ = 2d r
′ , we see that in a bicentric quadri-
′
lateral ABCD the image S of the point S of intersection of GE and F H - and
also of AC and BD by Pascal’s theorem applied to a degenerated hexagon - is the
same when inverted with respect to C or when inverted with respect to C ′ . This
means that the circle with diameter SS ′ is orthogonal to C and to C ′ - and also to
C by inversion with respect to C . This reveals C, C and C ′ as members of a coaxial
system of circles with limiting points S and S’. The perpendicular bisector of SS ′
is the radical axis of this coaxial system, [8, chapter III].
7. Yun’s inequality revisited

With A+B
2 = E, B+C
2 = F and the law of sines 2r sin E = F H, 2r sin F =
EG, Yun’s inequality
√
2r 1 A B B C C D D A
≤ sin cos + sin cos + sin cos + sin cos ≤ 1,
R 2 2 2 2 2 2 2 2 2
√ 2
[10], [15], is converted by multiplication with 2r into 2 R2r ≤ EG+F 2
H
≤ 2r.
The right hand side is obvious. We increase the left hand side applying √ the for-
2 R′ 2 2
mula for the radius of an inverted circle R = Rr′2 −d 2 ≥
√ r
R′2 −d′2
to 2 R2r ≤
√ √ 2 p
2 2R′2 − 2d′2 . From 2R′2 + 2d′2 = r 2 we get 2 R2r ≤ 2 r 2 − (2d′ )2 . But
p
2 r 2 − (2d′ )2 is the length of the minimum chord of the circle C through S and
EG+F H
2 is the mean of any two orthogonal chords through S, which is obviously
greater, equality occurs only for squares ABCD when S = Z.
Comparing one chord instead of the mean of two orthogonal chords with the
minimum chord we get the inequality
√
2r A+B A D B C
≤ sin = sin sin + sin sin ,
R 2 2 2 2 2
of which Yun’s inequality is a consequence.
References
[1] C. Alsina and R. B. Nelsen, Charming Proofs, MAA, 2010.
[2] A. Bogomolny, Newton’s and Léon Anne’s Theorems,
http://www.cut-the-knot.org/Curriculum/Geometry/NewtonTheorem.shtml
[3] A. Bogomolny, Poncelet Porism, Fuss’ Theorem,
http://www.cut-the-knot.org/Curriculum/Geometry/Fuss.shtml
[4] A. Bogomolny, Poncelet Porism,
http://www.cut-the-knot.org/Curriculum/Geometry/PonceletChain.shtml
[5] L. Carlitz, A note on circonscriptible cyclic quadrilaterals, Math. Mag., 38 (1965) 33–35.
[6] L. Flatto, Poncelet’s Theorem, AMS, Providence RI, 2008.
Bicentric quadrilaterals through inversion 15
[7] F. G.-M., Exercices de géométrie, A. Mame et fils, Tours, 1912.

[8] R. A. Johnson, Advanced Euclidean Geometry, Dover, 2007.
[9] M. Josefsson, Characterizations of bicentric quadrilaterals, Forum Geom., 10 (2010) 165–173.
[10] M. Josefsson, A new proof of Yun’s inequality for bicentric quadrilaterals, Forum Geom., 12
(2012) 79–82.
[11] J. C. Salazar, Algunos teoremas y sus demostraciones, Rev. Esc. Olimpı́ada Iberoamericana de
Matemática, 13 (2003) 1–8.
[12] D. Speyer, Poncelet Porism,
http://sbseminar.wordpress.com/2007/07/16/poncelets-porism/
[13] Wolfram MathWorld, Poncelet’s Porism,
http://mathworld.wolfram.com/PonceletsPorism.html
[14] yetti (username), www.artofproblemsolving.com/Forum/viewtopic.php?p=148226
[15] Z. Yun, Euler’s inequality revisited, Mathematical Spectrum, 40 (2008) 119–121.
Albrecht Hess: Deutsche Schule Madrid, Avenida Concha Espina 32, 28016 Madrid, Spain
E-mail address: [email protected]

Bicentric Quadrilaterals Through Inversion

Uploaded by

Copyright:

Available Formats

Bicentric Quadrilaterals Through Inversion

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bicentric Quadrilaterals Through Inversion

Uploaded by

Copyright:

Available Formats

b

Bicentric Quadrilaterals through Inversion

Publication Date: January 29, 2013. Communicating Editor: Paul Yiu.

2. Orthogonality of Newton lines

6. Coaxial system of circles

7. Yun’s inequality revisited

[7] F. G.-M., Exercices de géométrie, A. Mame et fils, Tours, 1912.

You might also like