Borromean Circles Are Impossible

Author(s): Bernt Lindstrom and Hans-Olov Zetterstrom

Source: The American Mathematical Monthly, Vol. 98, No. 4 (Apr., 1991), pp. 340-341
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/2323803 .
Accessed: 15/12/2014 03:08
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
The American Mathematical Monthly.

http://www.jstor.org

This content downloaded from 128.235.251.160 on Mon, 15 Dec 2014 03:08:58 AM

All use subject to JSTOR Terms and Conditions

340

BERNT LINDSTROM

AND HANS-OLOV

ZETTERSTROM

[April

formover the complexnumbersis all thatis needed, the use of Proposition2 may
simplifycalculations.As we have seen here, that is the case with the proofsof
[1,2]. The existenceproofof [3] can also be shortenedby use of Proposition2.
REFERENCES
1. A. F. Filippov,A shortproofof the theoremon the reductionof a matrixto Jordanform,Moscow
Uniu. Math. Bull., 26 (1971) 70-71.
2. JamesM. Ortega,MatrixTheory,Plenum Press,New York, 1987.
3. H. Valiaho, An elementaryapproach to the Jordanformof a matrix,this MONTHLY, 93 (1986)
711-714.

BorromeanCirclesAre Impossible
BERNT LINDSTR6M

Department
ofMathematics,
RoyalInstitute
of Technology,
S-10044Stockholm,Sweden
HANS-OLOV

ZETTERSTR6M

NationalDefenceResearchEstablishment,
S-171290Sundbyberg,
Sweden

The arrangementof threetopologicalcirclesin R3 depicted in FIG. 1 is called

the Borromeanlinks in knot theory[cf. 1]. Such an arrangementof 'circles' is
possible with curves, which are homeomorphicto geometriccircles. To our
knowledgeit has not been observedbeforethat the arrangementis impossiblein
R3 ifthe circlesare geometriceven whenthe radiiare arbitrary.

(,

FIG. 1.

Our proof is indirect.We assume that the arrangementis possible using

ordinarycircles.LiftC2 (cf. FIG. 1) towardsthe spectatorsuch thatit will meet C1
in two pointsin its new position(CI). Observe that C, will not meet C3 since we
liftC2 awayfromC3.
There are now two cases: eitherC, and C, belong to a plane v or theydo not
belongto the same plane. In the firstcase C3 willbe partlyabove the plane, partly

This content downloaded from 128.235.251.160 on Mon, 15 Dec 2014 03:08:58 AM

All use subject to JSTOR Terms and Conditions

1991]

341

NOTES

below. If we follow the circle C3 in clockwisedirectionwe are firstabove the

plane, thenbelow, then above, thenbelow, thenback again and above the plane.
This means thatthe circle C3 meetsthe plane at least 4 times,whichis impossible
since the circledoes not belong to the plane.
In the othercase, when C1 and C, do not lie in the same plane, we will prove
thattheybelong to a sphere. The circle C3 will meet thissphere at least 4 times,
again a contradiction.
We will now provethat C1 and C, belong to a spherewhen
theyare not coplanar. Let P and Q be the pointsin whichC1 and C, meet (see
FIG. 2). Let R be the midpointof the line segmentPQ. Let S, and S2 be the
centresof the circles C1 and C1, respectively.
The plane 7' containingR, S1 and
S2 is orthogonalto the line PQ. Let li be the normalto the plane of Ci throughSi
(i = 1,2). The line li is orthogonalto the line PQ and containsthe point Si in the
plane ST'. It followsthatli lies in 7T'(i = 1,2). Since 11and 12 are not parallel lines
theywill meet in a point T. The distancebetweenT and anypointin C1 u C, is a
constantr. ThereforeC1 and C, belong to the spherewithcentreT and radius r.
It is not hard to see thatthe circleC3 goes in and out of thissphere4 times,which
is impossiblebecause a circle and a sphere have at most two points in common
when the circle does not belong to the sphere. Therefore,Borromeancirclesare
impossible!
p

,<

FIG. 2.

It is interestingto make a model of the Borromean rings in some elastic

materiallike thiniron wire. Try to make the 'circles' as perfectas possible! The
arrangementof ringswill then attaininfinitely
manystable positionsin space.
REFERENCE
1. D. Rolfsen,Knots and Links,Publishor Perish,Berkeley,1976.

This content downloaded from 128.235.251.160 on Mon, 15 Dec 2014 03:08:58 AM

All use subject to JSTOR Terms and Conditions