Some Proofs of a Theorem on Quadrilateral

Author(s): Kaidy Tan

Source: Mathematics Magazine, Vol. 35, No. 5 (Nov., 1962), pp. 289-294
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/2688203
Accessed: 11-01-2016 22:01 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.

Mathematical Association of America is collaborating with JSTOR to digitize, preserve and extend access to Mathematics
Magazine.

http://www.jstor.org

This content downloaded from 137.229.20.241 on Mon, 11 Jan 2016 22:01:38 UTC
All use subject to JSTOR Terms and Conditions
 TEACHING OF MATHEMATICS
EDITED BY ROTHWELL STEPHENS, KnoxCollege
This department is devotedto theteachingofmathematics.Thus, articlesof methodology,
exposition,curriculum,testsand measurements, and any othertopic relatedto teaching,are
invited.Papers on any subjectin whichyou, as a teacher,are interested,or questionswhich
you wouldlike otherstodiscuss,shouldbe sentto RothwellStephens,MathematicsDepartment,
Knox College,Galesburg,Illinois.
A COMMON MISAPPLICATION OF THE AXIOM OF FINITE INDUCTION
DAVID R. ANDREW, Universityof SouthwesternLouisiana

A graduate student in mathematics normallyuses the Axiom of Finite In-

duction in proofsover and over again. Familiaritywith the Axiom quite often
leads to a certain looseness in its application, namely, the student no longer
bothersto writedown or thinkcarefullyabout exactly what the inductionstate-
ment is. The result of this in many cases is that he errsby drawingtoo strong
a conclusion fromhis proof. The followingincorrectproofof the (true) state-
ment that the set P of positive integersis a well-orderedset occurs frequently
and illustratesthe point:
(1) The set I 1} is triviallywell-ordered.
(2) Assume that the set S= { 1, 2, 3, , k} is a well-orderedset.
{
(3) Let T_ 1, 2, 3, * * * k, k+l
, } and T' any nonemptysubset of T.
Case I. T'= k+ 1}. Then k+ 1 is the least element of T'.
Case II. T'I{k+1}, i.e., SnQT'=S'#0. Now S'CS; therefore,S' has a
least element k' by (2). Since k'GS' CS, k' <k+1, and k' must also be the least
elementof T'. Thus T is a well-orderedset, and it followsby inductionthat P is
a well-orderedset.
The finalconclusion that "P is a well-orderedset" does not followfromthe
proof. What does follow is that "for any positive integer n, the set Pn
=_{ 1, 2, 3, *. . , n } is a well-orderedset." This is an inductionstatementabout
a class of finitesets, and P is not a memberof this class. If the finalconclusion
above werecorrect,one could prove in exactly the same way that a denumerable
intersectionof (real) open intervals is an open interval,which, of course, is a
false statement.

SOME PROOFS OF A THEOREM ON QUADRILATERAL

KAIDY TAN, Fukien NormalCollege,Foochow,China
Dedicated to thememory
ofVICTOR THEBAULT
In this paper I shall give several differentproofsof a theoremon quadri-
lateral related below. Some of the proofsI believe to be new, and the othersare
taken fromeminent books perhaps the readers are familiarwith them. I all
collect here to make a comparisonbetween them,perhaps it may stimulate the
interestof college students or school geometryteacher.
THEOREM. If a quadrilateralbe circumscribed about a circle,thediagonals and
thelinesjoining theoppositepointsof contactare concurrent.
289

This content downloaded from 137.229.20.241 on Mon, 11 Jan 2016 22:01:38 UTC
All use subject to JSTOR Terms and Conditions
290 MATHEMATICS MAGAZINE [Nov.-Dec.

Let ABCD be a quadrilateral circumscribedabout a circleY, and the points

of conitacton each side are E, F, G and H respectively,as shown in the figure.
It is required to prove that A C, BD, EG and FH are concurrent.
For briefniess,we adopt the followingnotationis:
MYN representsthe line passing throughM and N.
ABnCD representsthe intersectioni of AB and CD.
C<,representsthe homotheticrelation,i.e. similarand similarlysituated.

FIG. 1

Proof 1. By the method of inversion. (See fig. 1.)

Invert the figurewith 0, the intersectionof HF and EG, as center ofinver-
sion,and the powerof 0 withrespectto the circlez as the constant of inversion.
Then the circle 2 invertsinto itself,E and G also F and H are inverse points.
The tangent AE inverts into a circle r passing through 0 and touching the
circlez at G. Similarlythe tangentAH invertsinto a circle Q passing through0
and touchingthe circlez at F. The intersectionA' of these two inversecirclesis
the inverseof A. .. A, 0, A' are collinear. Since C is the radical centerof these
three circles 1, r and Q, hence OA' must pass throughC, i.e. AC, EG and FH

/~ ~~~
/ F 2

FIG. 2

This content downloaded from 137.229.20.241 on Mon, 11 Jan 2016 22:01:38 UTC
All use subject to JSTOR Terms and Conditions
1962] SOME PROOFS OF A THEOREM ON QUADRILATERAL 291

are concurrent.Similarly,BD, EG and FH are concurrent. .-.A C, BD, EG and

FH meet in a point 0.
Proof 2. By the theoryof pole and polar. (See fig.2.)
Let HEnFG = M, EFnHG = N, EGnFH= 0. Since the polars of A and C,
i.e. EkJH and FkJG,pass throughM, so the polar of M is AUJC. Similarly,the
polar of N is BUD. .-.MUN is the polar of A COBD. But MUN is the polar
of 0, .@.A COBD 0. Hence A C, BD, EG and FH meet in a point 0.

'
I

FIG. 3

Proof 3. Alternatemethod of polar. (See fig.3.)

Let ABnCD=X, BCnAD= Y, EGnFH=0. Since EG and HF are the
polars of X and Y respectively,so XU Y is the polar of 0. .-. Y{ AC, XOI
{
-X A C, YO} = -1, .'. A, 0, C are collinear,i.e. A C, EG and FH are concur-
rent. It is evident that C, BD, EG and FH meet at 0.
A
Proof 4. Derived fromBrianchon Theorem. (See fig.3.)
Regard AEBCGD as a circumscribedhexagon, by Brianchon Theorem we
know at once that AC, EG and BD are concurrent.Similarly, if we regard
BFCDHA as a circumscribedhexagon, we know BD, FH and A C are concur-
rent. .,.A C, BD, EG and FH are concurrent.
Proof 5. By the theoryof transversals. (See fig.3.)
Let ABnCD=X, BCCAD=Y, EGnAC=01, HFnAC=02. Now it is
required to prove that 01 02.
Regard EG as a transversalof AXA C, then we have
AO1 CG XE AO1 EA
----- 1, O.. C
01C GX EA 0 C CG
Next, regard HF as a transversalof A YA C, then we obtain
A02 CF YH A02 HA
02C FY HAA ' 02C CF
But

This content downloaded from 137.229.20.241 on Mon, 11 Jan 2016 22:01:38 UTC
All use subject to JSTOR Terms and Conditions
292 MATHEMATICS MAGAZINE [Nov.-Dec.

AO1 A02
HA EA, CF C CG' =C 01-?2. j.&c.

FIG. 4
Proof6. By the theoryof homothecy.(See fig.4.)
Draw AM?I|DG, ANI|BF,cuttingEG anldHF at M anldN respectively. Join
MN and FG. Then LAEG= LDIGE= LAME. .. AAEMf is isosceles. Simi-
larly, AAHN is isosceles. But AE-AHI. .*.AM=AN, and the AAMN is
isosceles. Also the A CFG is isosceles. Now the two isosceles trianlglesA MN and
CGF having two sets of equlal sides are parallel anld in1opposite direction (i.e.
AM??|CG,AN?? CF), thereforetheir bases MlN anld (CF are also parallel and inl
MC#A cGF. .*.A C, EG, and HF are concurrent.
opposite direction. .*. LASA
Similarly,BD, EG and HF are concurrent. .-.&c.
Now, I give somleproo-fs
more elementaryas follows:

B_
_
I

FIo. S

Proof7. Apply the theoremon similarfigures.(See fig.5.)

Draw CM??AB, CN??AD, cuttingEG and HF at M and N respectively.
Then L M= LAEY= LDGE= L CGM, .:. ACMG is isosceles. Similarly,
A CNF is isosceles. But CC = CF. . . CM = CN. Now, ifA C do not pass through
the intersectionof EG and HF, let A CC\FH= X, and A CC'EG = Y, then

This content downloaded from 137.229.20.241 on Mon, 11 Jan 2016 22:01:38 UTC
All use subject to JSTOR Terms and Conditions
1962] SOME PROOFS OF A THEOREM ON QUADRILATERAL 293

AE CM AHl CN
- - , ~and =
AY CY AX CX
But AE =AH, thereforewe have
AX AY
AX A Y *.X=- Y-HF C EG.
CX CY
Hence A C, EG and HF are concurrent.Similarly,BD, EG and HF are concur-
rent. .. &c.

A
HD

BF CT
FIG. 6

Proof 8. Alternate method of similar figures.(See fig.6.)

Let A CC'EG= 0, on EB take a point K, so that Z EOK= Z GOC, now be-
cause
OK OC
ZOEK= ZOGC, AO_EKChAOGC, (1)

But
OA OK
ZAOE= ZCOG= ZEOK, AE K=E- (2)

From (1) and (2) we have

OA OC AO AE
- = - y ... - = - . ~~~~~~(3)
AE CG OC CG
Similarly,let A CnHF= O', then
AO' AH
= *- (4)
O'C CF
But AE =AH, CG= CF, hence, from(3) and (4), we obtain
AO AO'
- , .[. O g 0
OC O'C
Hence we know that the three lines AC, EG, HF meet at 0. :. &c.

This content downloaded from 137.229.20.241 on Mon, 11 Jan 2016 22:01:38 UTC
All use subject to JSTOR Terms and Conditions
294 MATHEMATICS MAGAZlNE [Nov.-Dec.

Proof 9. Apply the theoremon the ratio of areas. (See fig.6.)

Let ACnEG=0. Since ZAOE= LCOG, and ZAEO+ ZCGO=2 rt A,
therefore
AAEO AE*EO AO*OE
ACGO CG*GO CO*OG
hence
AO AE
OC CG
Similarly,if HFnA C =O', we have
AO' AH
O'C CF
But A E AH, CG= CF, thereforewe obtain
AO -AO'
-=--> i.e. 0 0'.
OC OC
Hence A C, EG, HF meet in a point 0. .:. &c.

NEW ROLES FOR OLD FIGURES

LADIS D. KOVACH, PepperdineCollege,California
Introduction.Advancementsin science and technologyhave been so numer-
ous in recent years that we are having to gallop in order to stand still-to
paraphrase the mathematician,Lewis Carroll. To help the scientistin solving
the complex problems of automation, disease, space travel, etc., high-speed
electroniccomputershave been developed. In fact,the worldof tomorrowseems
destinedto be controlledby computers,as already we have computersthat regu-
late highwaytraffic,measure the amount of smog in the atmosphere,help the
heart specialist diagnose diseases of the heart and make airlineand theatreres-
ervations.
In order that the modern large-scale digital computer may performsuch a
wide varietyof tasks, it is necessaryto employ a numberof special techniques.
These techniques have resultedin the appearance of new topics in mathematics
-assigning, as it were, new roles to our familiarnumbers.It is with these new
mathematicalideas that this paper is concerned.
The Digital Computer.In spite of the apparent complexityof a digital com-
puter,it can do only what its name implies-it can count. Counting forwardis
the same as adding, counting backward is the same as subtracting.If we can
add, we can also multiplysince multiplicationis successive addition. In a like
manner,division is successive subtraction.Thus by merelycounting,a digital
computer can add, subtract, multiplyand divide; in short, it can do simple
arithmetic.Now we are gettingto the heart of the matter;how can a computer,
which can do only simple arithmeticat the fifthgrade level, solve complex
scientificproblems?The answer lies in the fantasticspeeds at which our com-

This content downloaded from 137.229.20.241 on Mon, 11 Jan 2016 22:01:38 UTC
All use subject to JSTOR Terms and Conditions