Annals of Mathematics

More Theorems on the Complete Quadrilateral

Author(s): J. W. Clawson
Source: Annals of Mathematics, Second Series, Vol. 23, No. 1 (Sep., 1921), pp. 40-44
Published by: Annals of Mathematics
Stable URL: http://www.jstor.org/stable/1967780
Accessed: 09-10-2015 12:52 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

Annals of Mathematics is collaborating with JSTOR to digitize, preserve and extend access to Annals of Mathematics.

http://www.jstor.org

This content downloaded from 78.95.94.86 on Fri, 09 Oct 2015 12:52:51 UTC
All use subject to JSTOR Terms and Conditions
 MORE THEOREMS ON THE COMPLETE QUADRILATERAL.

By J. W. CLAwSON.

In a paper on "The Complete Quadrilateral" published in these

Annals,* a number of theoremswere given. These divide themselves
naturallyinto theoremsin connectionwith (A) the circumcentric circle,
C, determinedby the circumcentersof the fourtrianglesof the quadri-
lateral, (B) the mid-diagonalline,m, determinedbyrthe middlepointsof
lines joiningopposite vertices,(C) the orthocentric line, o, determinedby
the fourorthocenters, and the pedal line,p, whichare both perpendicular
to m, (D) the incentriclines,which are connectedwith the bisectorsof
angles ofthe quadrilateral. These fourdivisionsof the subject are some-
what loosely connected,in the paper referredto, by the facts that the
focal point,F, at which the fourcircumcentersmeet, (A) lies on C, (B)
is the focusof the most importantof the conics whose centersare on mn,
(C) is simplyrelated to p, and (D) is the intersectionof the incentric
lines.
In this note some furtherconnectivetheoremsare added. In 1 and 2,
(A), (B) and (C) are more closelylinked,in 3. (A) and (D) are bound
together,and in 4, relationsare given connecting(A), (B) and (D).
The notationof my formerpaper is preserved,and most of the refer-
ences are to it. The contents of this note are original,except where
otherwisestated.
1. (1) The mid-diagonalline, m, of the quadrilateralbisects the line
joining the center,C, of the circumcentriccircle and the mean center,
H, of the fourorthocenters of the trianglesof the quadrilateral.
I have discoveredtwo proofsof this theorem,both too long forinser-
tion in full. The firstis analytical. Taking the focal point for origin,
and takingtheequationoftheline1 to be px + qly = p2 + q12, themid-
diagonallineis found,afterconsiderablereduction,to have forits equation
y =2ql, circle,is
the point C, the centerof the circumcentric

( p4- p 2q1q2 + qlq2q3q4 p24ql - Tqlq2q3)

4p3 ' 4p2

and the point H, the mean centeror centerof gravityof equal masses
* Vol. 20, pp. 232-261.
40

This content downloaded from 78.95.94.86 on Fri, 09 Oct 2015 12:52:51 UTC
All use subject to JSTOR Terms and Conditions
 MORE THEOREMS ON THE COMPLETE QUADRILATERAL. 41

placed at Hi, H2, H3, H4, is

(2p, 3p2 q + qq2q

Hence the middlepointof CH has l2q, forits ordinate.

The second proofis statical. It can be proved,using trigonometrical
methods,thatmassessin2A23at C1,sin 2A13at C2,sin 2A12at C3, and
sinA23sinA13sin A12at eachofthefourpointsH1,H2,H3,H4 areequiv-
alent to masses 4 sin A23 sin A13sin A12at C and at H, and hence to the
singlemass 8 sin A23 sin A13sin A12at the middlepointof CH.
Again, the masses at C1 and H1 may be replaced by certainmasses
at the verticesof the triangleA23A24A34 whose circumcenterand ortho-
centerare C1 and H1. In this way the seven originalmasses may be
replaced by masses at the six vertices;and it can be proved,laboriously,
that the massesat oppositeverticesare equal; hencethat the sevenmasses
are replaceable by three masses at B1, B2, B3, the middlepointsofthe
diagonals. But the centroidof these three masses is at a point on the
mid-diagonalline. Hence the middlepoint of CH lies on this line.
(2) U, the centerof gravityof equal masses placed at the six vertices
of the quadrilateral,is the centroidof the trianglewhose vertices are
C, H, and the orthiccenter,* 0, of the quadrangleC1C2C3C4.
This is easily establishedstatically. For masses 2m at each of the
six verticesmay be replaced by 3m at the centroidof each of the four
trianglesof the quadrilateral. But, since the centroidof a triangleis
one thirdof the distancefromthe circumcenter to the orthocenter,these
may be replaced by fourmasses of m each at H1, H2, H3, H4 and four
masses of 2m each at C1, C2, C3, C4. Now the mean centertof the quad-
rangle C1C2C3C4bisectsl the line joining C, its center,to 0, its orthic
center,-the point where perpendicularsto each side fromthe middle
point of the oppositeside concur.* Hence these masses may be replaced
by 4m at H, and 4m at C and 4m at 0. But the six masses of 2m each at
the verticesmay also be replacedby a mass of 12mat U. Hence U is the
centroidof equal masses at H, C and 0.
Since OU produced bisects CH, and since, by (1), the mid-diagonal
line,whichcontains U,? bisects CH, it followsthat:
(3) The mid-diagonalline of a complete quadrilateral contains the
orthiccenterof the circumcentric quadrangle.
2. Let hi, h2,h3,h4be the orthocenters of the trianglesC2C3C4,C3C4C1,
* P. 252 (y), Annals, l.c.
t P. 251 (a).
+ P. 252 (a).
? P. 238 (9).

This content downloaded from 78.95.94.86 on Fri, 09 Oct 2015 12:52:51 UTC
All use subject to JSTOR Terms and Conditions
42 J. 'W. CLAWSON.

C4ClC2,C1C2C3 respectively. Then Clh1,C2h2, C3h3,C4h4are bisectedat

0,* and the quadrangle h1h2h3h4is directlysimilarto the circumcentric
quadrangle C1C2C3C4. Let K be the centerof the circle circumscribing
h7h2h3h4.Then CK is bisectedat 0. Also KC4is equal and parallelto h4C.

FIG. 1.
Again the trianglesA23A13A12 and C1C2C3are directlysimilar,the
triangleshaving a centerof perspectiveat the intersectionof C4and Ct
other than F, viz., E4. The orthocentersof these trianglesare H4, h4;
their circumcentersare C4, C; their radii are R4, R. Then H4C4/h4C
- R4/R. Hence H4C4/KVC4 = R4/R.
:Nowconsiderthe trianglesH4C4Kand C4FC. By the last statement,
H4C4/KC4 = C4F/CF. But 4; KC4H4is equal to the angle betweenh4C
and H4C4. But h4Cmakes the same angle with C2C3that H4C4makes
with A 13A12,consideringthe similar figures. Hence 4; KC4H4is equal
to the angle between C2C3and A13A12.But C2C3is perpendicularto
FA14. Hence 4; KC4H4is the complementof 4; A 13A 1F.
Again 4; CFC4is thecomplement of 4S C4C2F,i.e.,of 4 A13A14F.
Hence the above-namedtrianglesare similar. But triangleC4FCis
isosceles. Hence KC4is equal to KiH4. Thus K lies on the perpendicular
bisectorof C4H4.
* P. 253 (i).

I P. 235 (5).

This content downloaded from 78.95.94.86 on Fri, 09 Oct 2015 12:52:51 UTC
All use subject to JSTOR Terms and Conditions
 MORE THEOREMS ON THE COMPLETE QUADRILATERAL. 43

SimilarlyK lies on the perpendicularbisectorsof C1Hi, C2H2, C3H3.

This gives a new proofof Hervey's theorem,* that
(4) The perpendicularbisectorsof the lines joining the circumcenters
and orthocenters are concurrent.
It also connectsthis pointwithotherpointsof the quadrilateral,since
(5) The line joining this point of concurrenceto the center of the
circumcentric circleis bisected by the orthiccenterof the circumcentric
quadrangle.
Moreover
(6) HK is parallel to the mid-diagonalline.
3. Startingtfromthe fact that the sixteen centersof the circlesin-
scribed and escribedto the four trianglesof the quadrilateralare four
by fourconcyclic,givingrise to eight new circles,whose centersY1, Y2,
Y3, Y4 and Z1,Z2, Z3, Z4 lines,:it is easyto see that
are on theincentric
(7) The circleon Y1Z1 as diameterpasses throughthe pointscommon
to the (orthogonal)circles just named whose centersare at Y1 and Z1.
One of these pointsis 14. This circlealso passes throughF and through
the middle points of II4, 121123, I3134, as is easily proved. There are
sixteencirclesof this kind whichall pass throughF.
But it is moreremarkablethat
(8) The centersof these sixteencirclesare the incentersand excenters
of the circumcentric quadrangleC1C2C3C4.
For C1C2, C1C3are perpendicularrespectivelyto FA34,FA24. Hence
the bisectorof 4, C2C1C3is perpendicularto the bisectorof 4 A34FA24.
Now the circle C1 is the nine-pointcircle of the triangle112113114. Let
A2311cut C1at X. Then X is the middlepointof 11I14. Also the bisector
of 4 A34FA24cuts C1 at X. Hence the bisector of 4, C2C1C3 is per-
pendicularto FX. But the circlesY1FZ1and C1have X and F in common.
Hence the centerof Y1FZ1 lies on the bisectorof 4 C2C1C3. Call this
point14'. In this way the theoremis proved.
From the fact that the middlepointsof Y1Z1, Y1Z2, Y1Z3,Y1Z4lie on
a line parallel to Z1Z2Z3Z4, it followsthat
(9) The centersof these sixteencircleslie fourby fouron fourlines
parallelto one of the incentriclinesand also fourby fouron linesparallel
to the otherincentricline.
* P. 244 (25).
t I am indebtedto a paper by F. V. Morleyin the AmericanMathematicalMonthlyfor
June,1920 (vol. 27, p. 252), forall thefactscontainedin thissection. Mr. Morleyderivesthe
t heoremsas a specialcase froma chainoftheorems concerning theincentersofn directedlines.
It seemsworthwhileto statethetheorems in different
orderand languageand to derivethemby
puregeometry fromsimplerratherthanfrommorecomplextheorems.
I P. 245 (27), p. 246 (27a).

This content downloaded from 78.95.94.86 on Fri, 09 Oct 2015 12:52:51 UTC
All use subject to JSTOR Terms and Conditions
44 J. W. CLAWTSON.

Moreover
(10) The incentriclines of the quadrilateral are parallel to the bi-
sectorsof the angles betweenpairs of oppositesides of the circumcentric
quadrangle.*
4. If the figurethat we are consideringis invertedwith respect to
the centerF, I have shown elsewheretthat a new figureresultswhich
is inverselysimilarto the old one, one of the incentriclines beingan axis
of similitude;and that the circumcentriccircle and orthocentricline of
the old figureinvertinto the orthocentricline and circumcentriccircle
of the new one, while the incentriclines invert into themselves. It
followsat once fromthese factsthat
(11) FC and the mid-diagonallines are equally inclined to the in-
centriclines of the quadrilateral.
It furtherappears that
(12) If the incentriclines cut the circumcentriccircle at F, J and
F, J', respectively,
the diameterJJ' is parallelto the mid-diagonalline.
URSINUS COLLEGE.
* P. 256 (Xv).
vol. 24, (1917),p. 71.
t Amer.Math. Monthly,

This content downloaded from 78.95.94.86 on Fri, 09 Oct 2015 12:52:51 UTC
All use subject to JSTOR Terms and Conditions