THE BORSUK-ULAM THEOREM AND

APPLICATIONS
PRESENTED BY ALEX SUCIU AND MARCUS FRIES
TYPESET BY M. L. FRIES
1. History
The Borsuk-Ulam theorem is one of the most applied theorems in
topology. It was conjectured by Ulam at the Scottish Cafe in Lvov.
Applications range from combinatorics to dierential equations and
even economics. The theorem proven in one form by Borsuk in 1933
has many equivalent formulations. One of these was rst proven by
Lyusternik and Shnirelman in 1930.
2. Borsuk-Ulam
Theorem 2.1. For n > 0 the following are equivalent:
(i) For every continuous mapping f : S
n
R
n
there exists a point
x S
n
such that f(x) = f(x).
(ii) For every antipode-preserving map f : S
n
R
n
there is a point
x S
n
satisfying f(x) = 0.
(iii) There is no antipode-preserving map f : S
n
S
n1
.
(iv) There is no continuous mapping f : B
n
S
n1
that is antipode-
preserving on the boundary.
(v) Let A
1
, . . . , A
d
be a covering of S
d
by closed sets A
i
. Then there
exists i such that A
i
(A
i
) ,= .
Proof.
(i ii) Let f : S
n
R
n
be an antipode-preserving map. By (i) there is
a point x S
n
such that f(x) = f(x). Since f is antipode-preserving
we know f(x) = f(x) = f(x), thus 2f(x) = 0 and f(x) = 0.
(ii i) Let f : S
n
R
n
be a continuous map. Dene a map g : S
n

R
n
by g(x) = f(x) f(x). We see that g(x) = g(x), hence g is
antipode preserving. By (ii) there is a point x S
n
such that g(x) = 0
and thus f(x) f(x) = 0.
Date: 5-9-2005.
1
2 TYPESET BY M. L. FRIES
(ii iii) Let f : S
n
S
n1
be an antipode-preserving map. We may
compose f with the inclusion i : S
n1
R
n
. By (ii) there is x S
n
such that f(x) = 0. This is a contradiction since we assumed that
f(S
n
) S
n1
.
(iii ii) Let f : S
n
R
n
be an antipode-preserving map. Assume
that f(x) ,= 0 for all x S
n
. We may then dene a map g : S
n
S
n1
by g(x) =
f(x)
f(x)
. We see that g is an antipode-preserving map from
S
n
S
n1
, which contradicts (iii).
(iv iii) The map (x
1
, . . . , x
n
, x
n+1
) = (x
1
, . . . , x
n
) is a homeo-
morphism from the upper hemisphere of S
n
to B
n
. An antipode-
preserving map f : S
n
S
n1
would yield a map g : B
n
S
n1
by
g(x) = f(
1
(x)) which is antipode-preserving on the boundary.
(iii iv) Assume g : B
n
S
n1
is antipode-preserving on the bound-
ary. Then we can dene a map f : S
n
S
n1
by f(x) = g((x)) for
x in the upper hemisphere and f(x) = g((x)). We see that f is
antipode-preserving which is a contradiction.
(i v) For a closed cover F
1
, . . . , F
n+1
of S
n
we dene a function
f : S
n
R
n
by f(x) = (dist(x, F
1
), . . . , dist(x, F
n
)). By (i) there is a
point x S
n
such that f(x) = f(x) = y. If the i
th
coordinate of f(x)
is non-zero then x F
i
. If all coordinates are non-zero then x F
n+1
.
(v iii) We rst note that there exists a covering of S
n1
by closed
sets F
1
, . . . , F
n+1
such that F
i
(F
i
) = for all i. To nd such
a cover consider the nsimplex in R
n
centered at 0. Then project
the faces of the nsimplex to the sphere. With this result in hand
we see that if a continuous antipode-preserving map f : S
n
S
n1
existed, the sets f
1
(F
1
), . . . , f
1
(F
n+1
) would be a cover of S
n
such
that f
1
(F
i
) (f
1
(F
i
)) = . This contradicts (v), thus no such map
can exist.
Even though we have shown the equivalence of the above statements
we have not shown that one of them is true in its own right. We will
do that in the following
Theorem 2.2. There is no antipode-preserving map f : S
n
S
n1
.
Proof. Let f : S
n
S
n1
be an antipode-preserving map. Since f
commutes with the map i
k
: S
k
S
k
given by i
k
(x) = x we may
descend to the quotient space. The quotient of S
k
under i is the space
MARCUS FRIES 3
RP
k
. Thus we obtain a commutative diagram of spaces
S
n
f
//
pn

S
n1
p
n1

RP
n
f
//
RP
n1
where f is the map induced on the quotients. So we nd that an
antipode-preserving map fromS
n
S
n1
gives rise to a map f : RP
n

RP
n1
.
The existence of an antipode-preserving map f gives rise to a map
in cohomology
f

: H

(RP
n1
; Z
2
) H

(RP
n
; Z
2
).
To make use of this map we need to recall the following
H

(RP
k
; Z
2
)

= Z
2
[x]/(x
k+1
), deg(x) = 1,
where Z
2
are the integers mod 2. Which when combined with the result
above results in a ring homomorphism
f

: Z
2
[x]/(x
n
) Z
2
[y]/(y
n+1
).
We now claim that
f

(x) = y,
this will be proven later. Assuming the claim we nd that
0 = f

(x
n
) = y
n
,= 0,
which is a contradiction.
To complete the proof we need to show that f

(x) = y. As a
rst approximation we want to nd f

:
1
(RP
n
)
1
(RP
n1
). Us-
ing the path lifting criterion for the covering p
n
: S
n
RP
n
we can
show that this is the identity map. Applying the Poincare-Hurewicz
theorem we nd f

: H
1
(RP
n
; Z) H
1
(RP
n1
; Z) is the identity map.
Finally by the Universal Coecient theorem for cohomology we nd
f

: H
1
(RP
n1
; Z
2
) H
1
(RP
n
; Z
2
) is the identity map.
3. The Z
2
Index
Denition 3.1. A Z
2
-space is a pair (X, ) with X a topological space
and a homeomorphism : X X such that = id
X
. We say the
Z
2
-action is free if has no xed points.
4 TYPESET BY M. L. FRIES
Denition 3.2. A Z
2
-equivariant map is a function f from a Z
2
-space
(X, ) to a Z
2
-space (Y, ) such that the following diagram commutes
X
f
//

X
f
//
Y
For brevity by Z
2
-map we mean a Z
2
-equivariant map.
Example 3.3. Let (S
n
,
n
) be a Z
2
-space where
n
(x) = x is the
antipode map. Then the Borsuk-Ulam theorem says that there is no
Z
2
-equivariant map f : (S
n
,
n
) (S
m
,
m
) if m < n. When we have
m n there do exist Z
2
-equivariant maps given by inclusion.
The existence or non-existence of a Z
2
-map allows us to dene a
quasi-ordering on Z
2
-spaces motivated by the following
Denition 3.4. Let (X, ) and (Y, ) be Z
2
-spaces. If there exists a
Z
2
-map f : X Y we write
X
Z
2
Y.
Simply stating that this is a quasi-ordering is not enough we need to
check the properties.
Lemma 3.5. The relation
Z
2
dened above is a quasi-ordering. That
is, the ordering is reexive and transitive.
Proof. To see that
Z
2
is reexive we use the identity map id
X
: X
X. We see that this map commutes with any Z
2
-action. For transi-
tivity let f : X Y and g : Y Z be Z
2
-maps. We then have the
commutative diagram
X
f
//

Y
g
//

X
f
//
Y
g
//
Z

Using this quasi-ordering we are now in a position to dene two

numerical invariants associated to a Z
2
-space.
Denition 3.6. We dene the Z
2
-index of a Z
2
-space (X, ) by
ind
Z
2
(X, ) = minn [ X
Z
2
S
n
.
Dual to the index is the Z
2
-coindex dened by
coind
Z
2
(X, ) = maxn [ S
n

Z
2
X.
MARCUS FRIES 5
Where in both cases the Z
2
-action on S
n
is given by the antipode map.
Proposition 3.7. The Z
2
-index and coindex satisfy the following prop-
erties
(i) (X, )
Z
2
(Y, ) ind
Z
2
(X, ) ind
Z
2
(Y, )
(ii) (X, )
Z
2
(Y, ) coind
Z
2
(X, ) coind
Z
2
(Y, ),
(iii) coind
Z
2
(S
n
,
n
) = ind
Z
2
(S
n
,
n
) = n,
(iv) for all Z
2
-spaces (X, ) we have coind
Z
2
(X, ) ind
Z
2
(X, ).
Proof.
(i), (ii) Assume (X, )
Z
2
(Y, ). This means there is a Z
2
-map f : X
Y . Let g : Y S
n
be a Z
2
-map. If we consider the composition
we obtain a map g f : X S
n
. Thus we see that ind
Z
2
(X, )
ind
Z
2
(Y, ). Similarly for the coindex.
(iii) By Borsuk-Ulam we know that if f : (S
n
,
n
) (S
m
,
m
) then we
must have n m. Combining this with the denitions of the index
and coindex we obtain our result.
(iv) Assume that coind
Z
2
(X, ) = n and ind
Z
2
(X, ) = m. We then
have a composition of Z
2
-maps
S
n
X S
m
,
and (iii) implies that n m.
From the proof we see that (iii) is a reformulation of the Borsuk-
Ulam theorem.
We cannot always expect coind
Z
2
(X) = ind
Z
2
(X) and in general this
is not true. In the cases when they are equal though we have the
following
Lemma 3.8. If coind
Z
2
(X, ) = ind
Z
2
(X, ) then

n
(X)
// //
Z
.
Example 3.9. By a theorem of Stolz [4] we know
ind
Z
2
(RP
3
, ) = 2,
where the Z
2
-action is induced from multiplication by i on S
3
C
2
.
Now since
2
(RP
3
) = 0 we know that coind
Z
2
(RP
3
, ) < 2.
Exercise 1. Show coind
Z
2
(RP
3
, ) = 1.
4. Ham Sandwiches
Of the many theorems that follow from the Borsuk-Ulam theorem,
the Ham Sandwich theorem has some of the best applications to com-
binatorics. Like the Borsuk-Ulam theorem the Ham Sandwich theorem
has many dierent formulations, though not all are equivalent.
6 TYPESET BY M. L. FRIES
Denition 4.1. A nite Borel measure on R
d
is a measure such that
all open subsets of R
d
are measurable and 0 < (R
d
) < .
Theorem 4.2 (Ham Sandwich for measures). Let
1
, . . . ,
d
be nite
Borel measures on R
d
such that every ane hyperplane has measure
zero. Then there exists a hyperplane h such that

i
(h
+
) =
1
2

i
(R
d
) i = 1, . . . , d,
where h
+
is one of the half spaces dened by h.
Proof. Let u = (u
0
, . . . , u
d
) S
d
. If at least one of u
1
, . . . , u
d
is not
zero, we assign to u the half space
h
+
(u) = (x
1
, . . . , x
d
) R
d
[ u
1
x
1
+ + u
d
x
d
u
0
.
Further we set h
+
((1, 0, . . . , 0)) = R
d
and h
+
((1, 0, . . . , 0)) = . We
see that h
+
assigns opposite half spaces to antipodal points. We dene
a function f : S
d
R
d
by setting the i
th
coordinate to be
f
i
(u) =
i
(h
+
(u)).
If we assume for the moment that f is a continuous function, the
Borsuk-Ulam theorem tells us there is a point u
0
such that f(u
0
) =
f(u
0
). Thus h
+
(u
0
) is our desired hyperplane.
To complete the proof we need to show that f is a continuous func-
tion. Let (u
n
)

n=1
be a sequence of points of S
d
converging to u. To
show that f is continuous it is enough to show that it is continuous
along the projections
i
: R
d
R. So we want to show
lim
n

i
(h
+
(u
i
)) =
i
(h
+
(u)).
Let x be a point not on the boundary of h
+
(u). Then for n suciently
large we know x h
+
(u
n
) if and only if x h
+
(u). So if g denotes
the characteristic function of h
+
(u) and g
n
the characteristic function
of h
+
(u
n
), we have g
n
(x) g(x) for all x h
+
(u). Since h
+
(u
n
)
have measure zero we know that the g
n
converge to g almost every-
where. By Lebesgues dominated convergence theorem, we thus have

i
(h
+
(u
n
)) =

g
n
d
i

gd
i
=
i
(h
+
(u)).
For our applications we will need a discrete version of the Ham sand-
wich theorem. We present that now.
Theorem 4.3. Let A
1
, . . . , A
d
R
d
be nite point sets. Then there
exists a hyperplane h that simultaneously bisects A
1
, . . . , A
d
.
By h simultaneously bisects A
i
we mean that each open half space
dened by h contains at most
1
2
[A
i
[| points.
MARCUS FRIES 7
Proof. We will prove the theorem by considering three cases, each more
general than the previous. To begin we assume that the points are in
general position in R
d
and each A
i
contains an odd number of points.
At each point we center an -ball. We then choose > 0 so that no
hyperplane intersects more that d of the -balls from one set A
i
. Now
by the Ham sandwich theorem we know there is a hyperplane which
simultaneously bisects each set. Additionally since there are an odd
number of points the hyperplane must intersect exactly one ball from
each set, and this cut contains the center.
We now assume that each A
i
contains an odd number of points but
the points need not be in general position. For every > 0 we dene
new sets A
i,
by moving the points of A
i
by at most so that the points
of A
1,
A
d,
are in general position. By the previous argument
there exists a hyperplane h

which simultaneously bisects each A

i,
.
We write h

= x R
d
[ x, a

) = b

where a

is a unit vector.
The points b

lie in a bounded interval thus by compactness there is a

cluster point (a, b) R
d+1
of the pairs (a

, b

). Let h be the hyperplane

dened by (a, b). Let
1
>
2
> be a sequence of points converging
to . Now if x lies at distance 2 from h then for i suciently large
x lies distance from h

i
. Thus if the open half space dened by h
contains k points, there is j such that i > j implies that the open half
space dened by h

i
contains k points.
Finally if some sets A
j
contain an even number of points we simply
delete one. We then apply the above argument to the remaining points.
Then we return the point, noticing that this does not change the result
due to the denition of bisection.
Corollary 4.4 (Ham sandwich for general position sets). Let A
1
, . . . , A
d
be nite point sets in R
d
such that A
1
. . . A
d
is in general position.
Then there exists a hyperplane h such that each open half space contains
exactly
1
2
[A
i
[| points.
Proof. We begin with an arbitrary ham sandwich hyperplane h as given
by the ham sandwich theorem. The problem is that h may contain up
to d points of some A
i
.
Fix a coordinate system so that h is given by x
d
= 0. Let B =
h (A
1
. . . A
d
). We know that B consists of at most d anely
independent points. We want to move h slightly so that it is our desired
cut. Since the points of B are anely independent we may make them
stay on h, go above or go below.
To see this add d [B[ new points to B so that we obtain a d-point
anely independent set C h. For each a C we choose a point
a

such that either a

= a or a

= a + e
d
or a

= a e
d
. We let
8 TYPESET BY M. L. FRIES
h

= h() be the hyperplane dened by the points a

. Now for all > 0

suciently small, the a

remain anely independent and the motion of

h() is continuous in . Thus we can guarantee that for all suciently
small > 0, h

is our desired hyperplane.

5. Lunch
All this discussion about ham sandwiches has built up our appetite.
Now is the time to feast.
Theorem 5.1 (Akiyama and Alon 1989). Consider sets A
1
, . . . , A
d
of
n points each, in general position in R
d
. Let 1, . . . , d be a set of colors
and color the points of A
i
with color i. Then the points of the union
A
1
. . . A
d
can be partitioned into rainbow d-tuples with pairwise
disjoint convex hulls.
Proof. We proceed by induction on the size of the sets A
i
. If n = 1 we
take the convex hull of the set A
1
A
d
. Now assume that n > 1 and
odd. Then by (4.4) there exists a hyperplane which intersects each A
i
in exactly one point and each open half space contains at most
1
2
[A
i
[|
points. Let the rst d-tuple be the points on the hyperplane. Then
apply the inductive hypothesis to each open half space. If n > 1 and
even (4.4) guarantees there is a hyperplane which bisects each A
i
but
does not intersect any A
i
. We then apply the inductive hypothesis to
the resulting open half spaces.
Theorem 5.2 (Necklace theorem). Every open necklace with d kinds
of stones, an even number of each, can be divided among two thieves
using no more than d cuts.
Before proving the theorem we need a preliminary result.
Lemma 5.3. Any ane hyperplane in R
d
cuts the curve (t) = (t, t
2
, . . . , t
d
)
in at most d points.
The curve (t) is commonly referred to as the moment curve. We
leave the proof of the lemma as an exercise to the reader. We now give
a proof of the theorem.
Proof. We place the necklace into R
d
along the moment curve (t). We
then dene sets A
i
by
A
i
= (k) [ the k-th stone is of the i-th kind.
By the general position ham sandwich theorem there is a hyperplane
h which bisects each A
i
. Additionally by the lemma we know h cuts
(t) in at most d points.
MARCUS FRIES 9
Denition 5.4. Let N be a set of n points, for each 1 k n write
N
k
: = S [ S N, [S[ = k. Dene the Kneser graph KG
n,k
to be
the graph with vertex set N
k
and edges given by uv E(KG
n,k
) if and
only if u v = .
Theorem 5.5 (Lovasz 1978). For all k > 0 and n > 2k 1, the
chromatic number of the Kneser graph is given by
(KG
n,k
) = n 2k + 2.
Proof.
Upper bound: For a vertex v N
k
color v by
(v) = minmin(v), n 2k + 2.
This denes a proper coloring since if (v) = (u) = i < n2k+2 then
i vu and hence uv is not an edge. If we have (v) = (u) = n2k+2
then u, v are subsets of the set n2k+2, . . . , n, but this set contains
2k 1 elements thus u v ,= .
Lower bound: Set d = n2k +1. Let X S
d
be a set of n points such
that no hyperplane passing through the origin contains more than d
points of X. Let the vertex set of KG
n,k
be identied with the set X
k
.
Assume there is a proper coloring of KG
n,k
with at most n 2k +
1 = d colors. Fix one such coloring and dene sets A
1
, . . . , A
d
S
d
by x A
i
if there is at least one k-tuple of color i inside the open
hemisphere H(x) centered at x. Finally dene A
d+1
= S
d
(A
1

A
d
). Clearly the sets A
1
, . . . , A
d
are open and A
d+1
is closed. By the
general version of the Lyusternik-Shnirelman theorem there is a set A
i
such that x, x A
i
. If i d we obtain two disjoint k-tuples with
the same color, thus the coloring is not proper. If i = d + 1 we know
H(x) contains at most k 1 points of X and H(x) contains at most
k 1 points of X. Therefore the complement S
d
(H(x) H(x))
contains at least n 2k + 2 = d + 1 points. But the complement is
dened by a hyperplane passing through the origin, which contradicts
our assumption that no hyperplane through the origin contains more
that d points.
References
[1] J. Akiyama and N. Alon, Disjoint simplices and geometric hypergraphs,
Combinatorial Mathematics; Proc. of the Third International Conference
(New York, 1985), volume 555, pages 1-3. Annals of the New York Academy
of Sciences, 1989.
[2] J. E. Green, A new short proof of Knesers conjecture, Amer. Math Monthly,
109:918-920, 2002.
10 TYPESET BY M. L. FRIES
[3] J. Matousek, Using the Borsuk-Ulam Theorem, Springer-Verlag 2003.
[4] S. Stolz, The level of real projective spaces, Comment. Math. Helvetici 64,
1989, pp.661-674.
Department of Mathematics, Northeastern University
E-mail address: fries.m@neu.edu