1 Introduction

We start off with a famous problem to introduce the idea of Planar Graphs. It is about connect-
ing three utilities to three houses. Suppose you want to connect the three utilities electricity,
water and gas [MJ07] to three houses in such a way that none of these connection lines will
cross another one. We can view this as the graph below:

Figure 1: Three Houses and Three Utilities

In the picture we can see that we have three houses, rather boringly labelled as “House 1”,
“House 2” and “House 3” and three utilities. The electricty utility is connected to each house
by a yellow line, the water utility is connected to each house with a blue line (most appro-
priately) and the gas (although not so well known in SA) is connected to each house with
a green line. As can be seen there are a lot of lines crossing each other. The question is
whether it is possible that the lines can be drawn in such a matter that none of them cross
each other. You can play around with this for a few minutes and might come to the conclusion
that try as you might you cannot stop at least one line from crossing another unless you don’t
connect all the utilities. This feeling would turn out to be correct later when we finally prove
it formally. However, we should first build up some theory before we get to solving this problem.

2 Developing the Theory

2.1 Definitions
Definition A Planar Graph is a graph that can be drawn in the plane in such a way that
none of its edges cross each other.

Definition A Plane Graph is a graph that has been drawn in the plane in such a way that
none of its edges cross each other.

Definition A graph G on n vertices is said to be a Complete Graph, represented as Kn , if

there is an edge between each pair of vertices.

Definition A graph G on m + n vertices are said to be a Bipartite Graph, represented as Gm,n

if there exists two disjoint sets of vertices |V1 | = m and |V2 | = n such that no two vertices in
Vi for i = 1, 2 have an edge between them and hence the only edges are between a vertex of V1
and a vertex of V2 . A graph is a Complete Bipartite Graph, represented as Km,n , if each vertex
in V1 has an edge to each vertex in V2 .

2
There is a subtle difference between a Planar Graph and a Plane Graph. A Plane Graph is a
Planar Graph, but a Planar Graph is not necessarily a Plane Graph. A Planar Graph need
not have none of its edges crossing when viewing it as a graph. It only needs to be possible to
draw it in such a way that none of its edges cross for it to be a Planar Graph. Consider the
following graph of K4 :

Figure 2: K4 not drawn as a Plane Graph

This graph is not a Plane Graph because it has two edges crossing. However, it is still a
Planar Graph, because we can redraw it as below by taking one of the edges to go around on
the outside:

Figure 3: K4 drawn as a Plane Graph

We can also see that our houses and utilities question can be reformulated into graph the-
ory as K3,3 . Therefore the question can be reformulated to ask whether K3,3 is a Planar Graph.
However, we can first consider less complicated graphs. We just saw that K4 can be represented
as a Plane Graph and is therefore Planar. The next graph that should intuitively be considered
is K5 :

Figure 4: K5 drawn normally

As it is drawn it is definitely not a Plane Graph.

3
However, can we make it Planar? If we try and move some edges we might get to:

Figure 5: K5 drawn in an attempt to make it a Plane Graph

As can be seen we will have a problem with the red edge since it seems it will cross some
other edge if we have to draw it in. This is of course not a proof yet. We first need to introduce
a theorem before we can formally prove this, but first another definition.

Definition A tree is a connected graph that has no cycles.

Lemma 2.1 If G is a tree with v vertices then G has v − 1 edges.

The proof of the lemma is left to the reader as an exercise and can easily be done by mathe-
matical induction or you can look it up in chapter 1.5 of [R.05] or most graph theory books.
Our discussion is about planar graphs and we only mentioned this lemma as we need it in the
proof of the theorem to follow.

2.2 Theorems
Theorem 2.2 (Euler’s Formula Theorem) If G is a connected plane graph with v vertices, e
edges and f faces, then
v − e + f = 2.

There is a generalization of this theorem, although not really relevant to our discussion it is so
interesting that it is worth a mention. If G is a connected plane graph drawn on a polyhedra
with n holes the formula becomes v − e + f = 2 − 2n [Wika].

Proof We will use the strong form of mathematical induction on the number of edges to show
the result. If e = 0 then G consist out of a single vertex and hence v = 1 and f = 1 since there
is only the plane as its face. Therefore v − e + f = 1 − 0 + 1 = 2 as required. So we have our
base case of e = 0 being true.

Induction hypothesis: Assume the formula holds for all graphs with fewer than k edges.
Let G be a connected plane graph with k edges. Suppose that G has v vertices and f faces.

If G is a tree then we have that f = 1 and that v = k + 1. Hence

v − e + f = v − k + f = (k + 1) − k + 1 = 2

as required.

4
On the other hand, if G is not a tree then we can find at least one cycle in G. Consider the
edge eG in this cycle. Removing eG from G does not change the fact that G is connected, since
eG was an edge from a cycle in G. We now have a graph with e − 1 edges, but still v vertices.
However, the two faces that was incident with the edge eG have become one since the edge
was removed. Therefore we only have f − 1 faces left. Let G0 be the graph obtained from G
by removing the edge eG . By the induction hypothesis the formula holds for G0 and therefore
v − (k − 1) + (f − 1) = 2, but

2 = v − (k − 1) + (f − 1) = v − k + 1 + f − 1 = v − k + f.

Therefore we also have the formula hold for the case with k edges. Thus by the Principle of
Mathematical Induction we have that for any connected plane graph v − e + f = 2.

We can now try and prove that K5 and K3,3 are non-planar. However, first we prove two lemmas.

Lemma 2.3 If G is a connected planar graph with v ≥ 3 vertices and e edges, then

e ≤ 3v − 6.

Proof Each face is surrounded by at least 3 edges, but each edge is incident with two faces.
Therefore the number of edges must be more than 3 times the number of faces, however, since
each edge lies between two faces we need to divide by 2. So if we have f faces and e edges then
e ≥ 3f2 or in other words f ≤ 23 e. Now using Euler’s Formula Theorem we have

2 1
2=v−e+f ≤v−e+ e=v− e
3 3
and by rearranging terms and multiplying by 3 we get

e ≤ 3v − 6

as required.

Lemma 2.4 If G is a connected planar graph containing no triangles with v ≥ 4 vertices and
e edges, then
e ≤ 2v − 4.

Proof We can in this case make our inequality from the first lemma stronger, because now
each face must be surrounded by at least 4 edges. Doing the same as before we get that e ≥ 4f2
and hence f ≤ 12 e and substituting this into the Euler Formula we get

1 1
2=v−e+f ≤v−e+ e=v− e
2 2
implying that e ≤ 2v − 4.

In particular the second lemma holds for all connected bipartite planar graphs, because bipartite
graphs only contain cycles of even length and hence no triangles. Now that we have proved
both lemmas we can move to the two theorems we want to prove.

5
Theorem 2.5 (K5 ) The graph K5 is not a Planar Graph.

Proof We will do this by contradiction. Assume it is indeed a Planar Graph. Therefore we

know by our first lemma that e ≤ 3v − 6. Let us count the number of edges e and number of
vertices v in K5 . We have v = 5 and e = 5×4
2
= 10. Substituting these in our inequality we get

10 = e ≤ 3v − 6 = 3(5) − 6 = 9

which implies 10 ≤ 9, which is absurd and hence the required contradiction.

We now move to the question about the three houses and three utilities that we started off with.

Theorem 2.6 (K3,3 ) The graph K3,3 is not a Planar Graph.

Proof We will again do this by contradiction. Assume K3,3 is a Planar Bipartite Graph and
by our second lemma we have that e ≤ 2v − 4. K3,3 has 6 vertices and 9 edges. Substituting
this in our inequality we get

9 = e ≤ 2v − 4 = 2(6) − 4 = 8

again a contradiction.

We now move to the two big theorems in the study of planar graphs. We will state them
without proof.

Theorem 2.7 Kuratowski’s Theorem A graph is planar if and only if it does not contain a
subdivision of K5 or K3,3 as a subgraph.

The Polish mathematician Kazimierz Kuratowski proved this theorem in 1930 according to
[R.05] and [Wikb].

Theorem 2.8 Wagner’s Theorem A graph is planar if and only if it does not have K5 or K3,3
as a minor.

The mathematician Klaus Wagner proved this theorem in 1937 according to [R.05] and [Wikb].
We feel these two proofs is beyond the scope of Olympiad mathematics. However, you can read
chapter 4 in [R.05] for a proof of these theorems. We can now finally turn to our main objective
which is problem solving.

6
2.3 Problem Solving
Example Show that K2,n is planar for every positive integer n.

Proof We can clearly draw it as a plane graph as demonstrated below.

Figure 6: K2,n

Example Show that in any connected simple planar graph there must be at least one vertex
of degree 5 or less.

Proof Suppose to the contrary that no vertex is of degree 5 or less. That means that all
vertices must be of degree 6 or more. Recall that the sum of the degrees of the vertices is twice
the number of edges. Let there be n vertices. Then the sum of the degree of the vertices is at
least 6n and therefore the number of edges is at least 3n. But by one of our lemmas we had
e ≤ 3v − 6 and so we get that 3n ≤ 3n − 6 a contradiction. Thus at least one of the vertices
must be of degree 5 or less.

Example For which values of n is Kn planar?

Proof We have already shown that K5 is not planar. Since K5 is a subgraph of Kn for n > 5
we know that by Kuratowski’s Theorem that none of Kn with n > 5 is planar. We have
demonstrated that K4 is planar and K3 is a triangle and obviously planar. Hence only values
for n that will work are n = 1, n = 2, n = 3 and n = 4.

Example Show that if the shortest cycle in a connected planar graph with v vertices and e
edges has length k then
k(v − 2)
e≤ .
k−2
Proof We can go about this in the same was as we proved the lemma. Each face must be
surrounded by at least k edges and each edge is incident with exactly two faces. Therefore we
have that e ≥ kf
2
, where f is the number of faces. Hence f ≤ k2 e and substituting this into the
Euler Formula we get
2 k−2
2=v−e+f ≤v−e+ e=v− e
k k
k
implying that e ≤ k−2 (v − 2).

7
Example Prove or disprove: A connected graph on v vertices and e edges with e = 3v − 6 is
planar.

Proof We want to show that it is false and therefore it would be sufficient to find an example
where K5 is embedded. If we have 6 vertices it means we must have 3(6) − 6 = 18 − 6 = 12
edges. K5 has 5 vertices and 10 edges. Thus we only need to add one vertex and connect this
vertex to two vertices in K5 and we will be done as demonstrated below.

Figure 7

Example Prove that if G is a graph of order 11, then G or its complement, G, is not a planar
graph.

Proof We assume to the contrary that both G and G are planar. If G is a graph of order
11 then the union of G and its complement is K11 , which has 11×10 2
= 55 edges. Since each
edge must be used in exactly one of G or G, we have that the sum of their edges is 55. By
the Pigeon Hole Principle that implies that either G or G must have at least 28 edges (since
27 + 27 = 54 < 55). Without loss of generality, let G have at least 28 edges. Since we assumed
G is planar the inequality e ≤ 3v − 6 must hold. Therefore e ≤ 3(11) − 6 = 33 − 6 = 27, but
this contradicts the fact that G has at least 28 edges. Thus, either G or G must not be planar.

Example In a convex polyhedron with m triangular faces (and possibly faces of other shapes),
exactly four edges meet at each vertex. Find the minimum possible value of m (Belarus
National Mathematical Olympiad 2000 Question 5) [TZG02].

Proof As usual let v be the number of vertices, e the number of edges and f the number of
faces. Since exactly four edges meet at a vertex, the number of edges must be four times the
number of vertices, but because each edge lies between two vertices we have double counted and
therefore the number of edges is 24 = 2 times the number of vertices. Hence e = 2v. Now we
need to count the number of faces in terms of the number of edges. For each of the triangular
faces there will be 3 edges and then there are f − m faces left which must each have at least
4 edges, but we have again double counted since each edge lies between two faces. Hence we
have
3m + 4(f − m)
e≤ .
2

8
Now using the Euler Formula and e = 2v we get that
1
2 = v − e + f = e − e + f ⇔ e = 2f − 4
2
and then using this with the inequality we got

3m + 4(f − m)
2f − 4 = e ≤ ⇔ m ≤ 8.
2
Therefore the minimum possible value for m is 8 and we need to check if this can indeed be
achieved. A regular octagon satisfies this and therefore the minimum of 8 can be achieved.

2.4 Exercises

1. Prove that K4, 4 is not planar.

2. Let G be a graph and G be its complement. We have already shown that if G has order
11 then either G or G is not planar.

(a) If G has order 7, show that G and G can both be planar.

(b) If G has order 8, show that G and G can both be planar.
(c) If G has order 9, do you think that G and G can both be planar.
(d) If G has order 10, explore ths possibility that G or G is not both planar.

3. Show that every planar graph is the union of three forests.

4. A maximal planar graph is a planar graph such that if you add an edge the graph is no
longer planar. A planar graph is triangulated if every face has exactly 3 edges. Show that
the following three statements are equivalent:

(a) G is a maximal planar graph.

(b) G is triangulated.
(c) Equality is achieved in the inequality e ≤ 3v − 6.

5. Prove that every planar graph can be properly colored using at most 6 colors.

6. As a research project, find out what the Petersen graph is and check whether it is planar
or not.

7. Those interested might want to research the concept of duality and [R.05] can be used
for this.

9
3 The Four-Colour Theorem
The Four-Colour Theorem is a very well-known theorem with a bit of a controversial history.
It deals with maps in the plane and the colouring of them. Whe one opens a world map the
bordering countries are usually coloured with different colours and the Four-Colour Theorem
formalises the question of how many colours will be sufficient to colour these maps.

Theorem 3.1 Four-Colour Theorem Any map in the plane can be coloured with at most four
colours so that no two regions which share a boundary are the same colour.

Proof Insert computer program here (joke).

The theorem has a little bit of a South African connection. Francis Guthrie noticed this pro-
perty and proposed it as a problem in 1852. His teacher, the more famous Augustus de Morgan,
tried and failed to prove it. In later life Francis Guthrie moved to South Africa where he was
later appointed as a professor of mathematics in Cape Town [J.H04] and [Wikc].

This theorem is just another one of a group of theorems or results, like Fermat’s Last Theo-
rem, Trisecting the Angle, Doubling the Cube and Squaring the Circle that generates a stream
of false proofs, sometimes from amateur mathematicians even after the theorem or result has
finally been proven. Unfortunately these have led to a group called mathematical cranks, those
who keep on sending a proof to the contrary to maths departments. The Four-Colour Theorem
has had an interesting hystory with respect to its proof and disproof.

Conceptualized in 1852 and attempted by Francis Guthrie and Augustus de Morgan, the the-
orem was finally proved by Alfred Kempe in 1879 (a serious proof). It stood as a proof for
11 years before Percy Heawood (not a crank) found a flaw in the proof and the theorem was
no longer and it was back to being a conjecture. Heawood did however proof the five-colour
theorem.

The theorem was finally laid to rest in 1977 a long 125 years after its conception. Wolf-
gang Hagen and Kenneth Appel finally proved the theorem. But, they made a big mistake for
their time, their proved relied on a computer checking more than 100 000 different map config-
urations and the mathematical world wasn’t yet ready for this leap of faith into the computer
world. Hence the mathematical world at the time treated it with scepticism.

Computers and trust in computers have come a long way since then and the original Hagen and
Appel proof has since been checked, confirmed and refined many times and today we do view
it as a proper theorem. However, some mathematicians still hold out hope that a proof will
be found that is completely independent of a computer, but only time will tell if this is the case.

In the mean time their are still mathematical cranks who from time to time send in their
disproof showing a counter-example only to have missed the obvious before sending it off.
Hopefully none of them will spend R10 000 on placing an advertisement in the local paper
“publishing” their counter-example, like a Cape Town man who spent R10 000 twice in 2011
to publish his “proof” of Fermat’s Last Theorem in the local paper (needless to say that his
“proof” was nothing more than a complete misunderstanding of the theorem).

10
The long wait for the proof of the Four-Colour Theorem, along with two longstanding flawed
proofs in the late 19th century, the finalt proof by computer in 1977 and the occasional math-
ematical crank trying to get into the action has certainly given this theorem its fame.

Apart from the fame, the Four-Colour Theorem and being interesting in Graph Theory, it
has never really been of interest to mapmakers according to [J.H04].

4 Conclusion
Hopefully you enjoyed this introduction to Planar Graphs and some interesting mathematical
Olympiad questions that can be asked aroud this topic. Although the study of this topis is not
strictly speaking needed for solving graph theory type problems in Olympiads, it is hoped that
studying these notes would help in solving future questions you may come across.

5 Bibliography
References
[J.H04] Webb J.H. The four-colour theorem. Mathematical Digest, 137:27, 2004.

[MJ07] Henning M.A. and Van Vuuren J.H. A first course in Graph Theory. In Press, 2007.

[R.05] Diestel R. Graph Theory. Springer-Verlag, Heidelberg, New York, 2005.

[TZG02] Andreescu T., Feng Z., and Lee G. Mathematical Olympiads 2000-2001 Problems and
Solutions From Around the World. The Mathematical Association of America, 2002.

[Wika] Available from: http://en.wikipedia.org/wiki/Euler_characteristic [cited 29

November 2011].

[Wikb] Available from: http://en.wikipedia.org/wiki/Planar_graph [cited 29 November

2011].

[Wikc] Available from: http://en.wikipedia.org/wiki/Four_color_theorem [cited 30

November 2011].

11