3.

POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

What is a Polyhedron?
Now that weve covered lots of geometry in two dimensions, lets make things just a little more difcult. Were going to consider geometric objects in three dimensions which can be made from two-dimensional pieces. For example, you can take six squares all the same size and glue them together to produce the shape which we call a cube.

More generally, if you take a bunch of polygons and glue them together so that no side gets left unglued, then the resulting object is usually called a polyhedron.1 The corners of the polygons are called vertices, the sides of the polygons are called edges and the polygons themselves are called faces. So, for example, the cube has 8 vertices, 12 edges and 6 faces. Different people seem to dene polyhedra in very slightly different ways. For our purposes, we will need to add one little extra condition that the volume bound by a polyhedron has no holes. For example, consider the shape obtained by drilling a square hole straight through the centre of a cube. Even though the surface of such a shape can be constructed by gluing together polygons, we dont consider this shape to be a polyhedron, because of the hole. We say that a polyhedron is convex if, for each plane which lies along a face, the polyhedron lies on one side of that plane. So, for example, the cube is a convex polyhedron while the more complicated specimen of a polyhedron pictured on the right is certainly not convex. Note that this denition is just a generalisation of the denition of a convex polygon in the plane. One very important thing to keep in mind is the fact that we usually think of a polyhedron as just the outside, the surface of the shape, and not as a solid object carved out of wood. Im sure youve probably seen and played with polyhedra many times before. For example, if youve ever picked up a standard soccer ball, you might have noticed that theyre usually made by sewing together patches in the shape of pentagons and hexagons. Other examples include the shape formed by the buckminsterfullerene carbon molecule C60 which is found in soot, the Montr al Biodome at Parc Jean Drapeau and e the pyramids of Egypt.
1 Although its all right to say polyhedrons, the more common plural to use is polyhedra. Thats because the word polyhedron comes from the ancient Greek words which mean many faces.

3. POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

What is a Graph?
Its quite unfortunate that the word graph means at least two completely different things in mathematics. One type of graph that you are no doubt already aware of is the graph of a function or equation, plotted on a set of axes. You probably know that its possible to plot the graph of an equation like y = x2 and you obtain a shape called a parabola which looks like a large smiley face. We wont talk about these types of graph here at all so forget you ever heard about them. For us, a graph will always mean a set of points called vertices, connected in pairs by lines or curves called edges. We make no assumptions about whether the graph is in one connected piece or not. However, we usually wont allow loops an edge whose endpoints are the same vertex or multiple edges more than one edge whose endpoints are the same pair of vertices.

Well call this a graph.

We wont call this a graph.

3. POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

For our purposes, its a good idea to think of a graph as a diagram whose vertices represent people at a party and whose edges represent friendship. So we should consider two graphs to be the same if they represent the same party. In particular, the important thing about a graph is not the way that you draw it on paper, but the relationships that the edges describe. Keeping this in mind, we say that two graphs G and H are isomorphic if there is a one-to-one correspondence between their vertices such that two vertices are connected by an edge in G if and only if their two corresponding vertices are connected by an edge in H. For example, the following two graphs are isomorphic and the labelling on the vertices tells you exactly why. 1 2 5 3 2 3 4 1 5 4

Every Polyhedron is a Graph

The fact that every polyhedron is a graph is a rather simple statement. This is because if you have a polyhedron and simply ignore the faces, then what you have left over is just a bunch of vertices connected in pairs by edges in other words, a graph. A more interesting statement is the following fact. Every polyhedron corresponds to a planar graph in other words, a graph which can be drawn in the plane without any of its edges crossing. So why is this true? Well, suppose that your polyhedron is made from some sort of rubbery material, like a balloon. If you pop the balloon by removing one of the faces, then what remains is a rubbery sheet with the vertices and edges still drawn on it. Now just stretch this out at onto a table and there you have your planar graph. Note that we needed to dene our polyhedra to not have holes for this to trick to work donut-shaped and other hole-possessing balloons just dont do the job.

The Handshaking Lemma

The degree of a vertex in a graph corresponds to a persons popularity index at a party its basically how many friends they have. So the degree of a vertex in a graph is simply the number of ends of edges which meet around that vertex. The following diagram shows a graph with each vertex labelled by its degree. 2 1 0 1

If you give me some numbers and ask me to nd a graph with those numbers as its degrees, then the task is often not possible. Obviously, the numbers have to be non-negative integers, but it turns out that other things have to be true as well. The most important relation which these numbers must obey is the following. 3

3. POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

Lemma (Handshaking lemma). At any party, if you ask everyone in the room (including yourself) how many hands they shook, and add up all of the answers, then you will get an even number. In fact, the number you get will be twice the number of handshakes that have occurred during the party. In graph theory terminology, this translates into the following fact in any graph, the sum of the degrees of all the vertices is equal to twice the number of edges. Proof. At the start of the party, everyone has shaken zero hands because no handshakes have taken place. Each time a handshake takes place, the number of hands shaken increases by two, once for each person shaking hands. Therefore, after all handshakes have taken place, the sum of all of the answers must be equal to twice the number of handshakes. In graph theory terms, this means that the sum of the degrees of all the vertices is equal to twice the number of edges in the graph.

The Bridges of Knigsberg o

Actually, graph theory began with the mathematician Leonhard Euler in the early eighteenth century. Its pretty amazing that such a simple mathematical construction took so long to be developed. Its now a thriving area of mathematics and, due to its very simplicity, has lots and lots of applications. In fact, the reason why Euler invented graph theory to begin with was to solve a very simple problem, known as the Seven Bridges of Konigsberg problem.

The above map shows what Konigsberg now known as Kaliningrad looked like back in Eulers day. When the weather was nice, the locals liked to stroll about the town. One thing that many of them tried to do was to walk around Konigsberg passing over every bridge exactly once, with no swimming allowed. Since people tried and tried without success, the problem became known as the Seven Bridges of Konigsberg problem. Proposition. Its impossible to walk around K nigsberg, passing over every bridge exactly once. o Proof. Eulers approach was to pick out only the relevant aspects of the diagram. For example, it didnt really matter that there was a church on the corner of Third Avenue and Fifth Street, nor that there was a troll-like hobo standing on one of the bridges. So if we remove a lot of the irrelevant details, we end up with a picture such as the one below left. However, Euler realised that you could simplify the diagram further by creating one point for every land mass on the map and joining two points each time there is a bridge joining the two land masses. Of course, the result is just the graph pictured below right.2
2 OK, so I said earlier that a graph cant have multiple edges but we should be allowed to change the rules whenever we like, to suit our application. So for the purposes of solving this problem, lets temporarily allow graphs with multiple edges.

3. POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

Lets suppose now that its possible to walk around Konigsberg, passing over every bridge exactly once. Well start with the obvious fact that this walk must start at some vertex of the graph and end at some vertex of the graph. Between the start and end, each time we visit a vertex, we must have walked along two edges incident to it one going in and one coming out. Even though we can visit a vertex many times, this reasoning tells us that every vertex other than the start and end must have even degree. So to be able to walk around Konigsberg, passing over every bridge exactly once, the Konigsberg graph above must have at most two vertices of odd degree. But you can see for yourself that all four vertices of the graph have odd degree, an obvious contradiction. In honour of Euler, we say that a graph is Eulerian if its possible to walk around the graph, passing over each edge exactly once. The following theorem tells us exactly when a graph is Eulerian. Theorem. A connected graph is Eulerian if and only if it has zero or two vertices with odd degree. Note that if a graph is Eulerian, then it should consist of one piece and such graphs are called connected. Our argument used to solve the Seven Bridges of Konigsberg problem tells us that if a connected graph is Eulerian, then it must have at most two vertices with odd degree. However, the handshaking lemma implies that no graph can possibly have exactly one vertex with odd degree. Therefore, if a connected graph is Eulerian, then it must have zero or two vertices with odd degree. Conversely, its always possible in such graphs to nd a walk which traverses each edge exactly once. The general idea behind the proof is to imagine a drunk person stumbling randomly around the graph, and never passing over an edge twice. Eventually, they have to get stuck. If theyve passed over each edge exactly once after they get stuck, then thats great. If not, then they they can always alter their route slightly to pass over more edges to do this, we need to use the fact that the graph is connected and has zero or two vertices with odd degree. They can keep altering their route in this way until they nally have a route which passes over each edge exactly once.

Six People at a Party

There is a very famous result in graph theory which states that, at any party with six people, there must exist three people who all know each other or three people who all dont know each other. One way to represent this problem is to draw a graph with six vertices, one for each person, and then draw a red edge between two people who know each other or a blue edge between two people who dont know each other. So the theorem can be rephrased in graph theoretic terms in the following way. Prove that if, in a graph with six vertices, every pair of vertices is connected by an edge coloured red or blue, then there must exist a red triangle or a blue triangle. In fact, to make life easier, lets dene the complete graph Kn to be the graph with n vertices such

3. POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

that every pair of vertices is connected by an edge. Then the six people at a party problem can be stated in the following way. Proposition. If every edge of the complete graph K6 is coloured red or blue, then there must exist a red triangle or a blue triangle. Proof. Consider a random partygoer A of the ve edges which are connected to A, note that at least three of them have to be the same colour. So lets just go ahead and assume that there are three edges connected to A, all of which are red. The same argument will apply for the case that there are three edges connected to A, all of which are blue. Suppose that these red edges connect A to the party people B, C and D. If BC is red, then triangle ABC is red; if CD is red, then triangle ACD is red; and if DB is red, then triangle ADB is red. So to avoid a red triangle, the edges BC, CD and DB must be all be blue, which forces triangle BCD to be blue. Therefore, there must exist a red triangle or a blue triangle in the graph.

Eulers Formula
If you ask me to build a polyhedron which has V vertices, E edges and F faces, then sometimes I wont be able to do it not because Im incompetent, but because such a polyhedron may not exist. In fact, Euler noticed that the equation V E + F = 2 holds for all polyhedra. For example, you can check that its true for the following three examples, where a tetrahedron is just a fancy name for a triangular pyramid. polyhedron tetrahedron cube soccer ball V 4 8 60 E 6 12 90 F 4 6 32

In fact, since polyhedra correspond to planar graphs, you might expect Eulers formula to work for planar graphs as well and youd be right. We have to be a little careful here, because Eulers formula will only work if you include the region outside the graph as a face. This is due to the fact that when we construct a planar graph from a polyhedron by attening it, we had to remove a face rst. Another subtle point is that your planar graph must be connected otherwise, Eulers formula just cannot hold, and you should check this for yourself. Theorem (Eulers Formula). For a connected planar graph with V vertices, E edges and F faces, V E + F = 2. Proof. Im just going to give a sketch proof here, something which is incredibly convincing, but which isnt actually a complete proof. Suppose that we would like to calculate V E + F for the graph on the right, but were just too lazy to count the number of vertices, edges and faces. Consider an edge, like the one labelled a, which connects one vertex to the remainder of the graph. If you remove this edge and the vertex attached to it, then you obtain a new graph with one less vertex and one less edge. So this new graph has exactly the same value of V E + F. 6 a

3. POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

Now consider an edge, like the one labelled b, which separates two distinct faces of the graph. If you remove this edge, then you obtain a new graph with one less edge and one less face. So this new graph also has exactly the same value of V E + F. So we can continue removing edges and vertices, without ever changing the value of V E + F, until we are left with the simplest graph possible, consisting of only one vertex and hence, only one face. Now we can calculate the value of V E + F for this graph and the answer is simply 1 0 + 1 = 2. Therefore, in our original planar graph, we know that V E + F = 2.

Dual Graphs
Given a planar graph, we can construct a new graph with a vertex for each of the faces of the original graph and an edge between two vertices if the corresponding faces meet along an edge. This new graph is called the dual graph. You should draw several examples to convince yourself of the following facts. The dual graph may have loops and multiple edges, but is always planar. The number of vertices of the dual graph is equal to the number of faces of the original graph. The number of edges of the dual graph is equal to the number of edges of the original graph. The number of faces of the dual graph is equal to the number of vertices of the original graph. The dual graph of the dual graph is the original graph. One extremely useful thing to do with a polyhedron or a planar graph is apply the handshaking lemma to its dual. To do this, you need to know what the degree of a vertex in the dual graph is. However, a vertex in the dual graph corresponds to a face in the original graph and the degree of a vertex in the dual graph is exactly the number of edges around the corresponding face in the original graph. Therefore, we obtain the following result. Proposition. In a planar graph or a polyhedron, the sum of the numbers of edges around each face is equal to twice the number of edges.

More on Planar Graphs

As fascinating as graph theory is, were actually here to learn about geometry. And as we saw earlier, one way that geometry makes contact with graph theory is via polyhedra. In particular, we showed that every polyhedron corresponds to a graph and, not just any old graph, but a planar graph. We have to take a little care with the denition of a planar graph, because we say that a graph is planar if it can be drawn in the plane without any of its edges crossing, not if it is drawn in the plane without any of its edges crossing. So, for example, although the following graph has been drawn with two edges crossing, it actually is planar, as you can check for yourself.

Some graphs just arent planar, no matter how hard you try to draw them without edges crossing. The one with the smallest number of vertices happens to be the complete graph K5 . 7

3. POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

Proposition. The complete graph K5 is not planar.

Proof. To obtain a contradiction, lets suppose that K5 is planar. Given that V = 5 and E = 10, Eulers formula tells us that, if we could draw K5 in the plane without edges crossing, then we would have F = 7. But it should be clear that if we could draw K5 in the plane without edges crossing, then every face must have at least three edges around it this is because K5 has no multiple edges or loops. This means that the sum of the numbers of edges around each face is at least 7 3 = 21. However, by the proposition above concerning the handshaking lemma in the dual graph, we know that the sum of the number of edges around each face coincides with twice the number of edges. In other words, we obtain the inequality 2E 21 which implies that 20 21. Contradictions dont come more blatant than that, so we may now conclude that there is no way to draw K5 in the plane without edges crossing in short, K5 is not planar. You can try and prove in an entirely analogous way that K3,3 is not planar either, although the proof is a little bit more difcult. A very difcult theorem to prove which is, in my opinion, quite amazing, is the following result which lets you decide whether or not a graph is planar. Theorem (Kuratowskis Theorem). A graph is planar if and only if it does not contain a a smaller graph which looks like K5 or K3,3 . When I say a graph which looks like K5 , I mean a graph which you can make out of K5 by adding in some essentially useless vertices along its edges. Heres another amazing theorem which involves planar graphs and is also quite difcult to prove. Theorem (F rys Theorem). Every planar graph can be drawn in the plane without edges crossing so that each edge a is represented by a line segment.

3. POLYHEDRA, GRAPHS AND SURFACES

3.1. From Polyhedra to Graphs

Euler
Leonhard Paul Euler pronounced oiler, not yooler was a Swiss mathematician who lived from 1707 to 1783. He is widely regarded as the preeminent mathematician of the eighteenth century and one of the greatest of all time. He is the most prolic mathematician to ever have lived, with his collected works lling about 75 quarto volumes. He was also quite prolic in other ways as evidenced by the fact that he fathered at least thirteen children. I could go on all day about Eulers contributions to mathematics and science. For example, he created graph theory when he solved the Seven Bridges of Konigsberg problem; he discovered the Euler line in Euclidean geometry; he discovered the formula V E + F = 2 for polyhedra; he introduced the notation e for the natural base of logarithms, f ( x ) for a function, for a summation, and i for the square root of negative one; he was the master of the branch of mathematics known as analysis; he discovered many theorems in number theory including a generalisation of Fermats Little Theorem; his inuence was integral in the birth of analytic number theory, a very important branch of mathematics; and he is renowned for work in mechanics, uid dynamics, optics and astronomy. Euler discovered many remarkable formulas in his day for example, he showed that the exponential function e x could be represented as an innite polynomial, or Taylor series. ex = x1 x2 x3 x0 + + + + 0! 1! 2! 3! This was called the most remarkable formula in mathematics by the famous physicist Richard Feynman. In 1988, readers of the Mathematical Intelligencer voted it the most beautiful formula ever and their poll also included two other formulas of Euler in the top ve. Eulers eyesight worsened throughout his life and he eventually became nearly blind in his right eye. His eyesight worsened so much over time that he earned the nickname Cyclops. Later in life, Euler suffered a cataract in his good left eye, rendering him almost totally blind in 1766. Even so, his condition appeared to have little effect on his productivity, as he compensated for it with his mental calculation skills and photographic memory. For example, Euler could repeat the Aeneid of Virgil from beginning to end without hesitation, and for every page in the edition he could indicate which line was the rst and which the last. With the aid of his scribes, Eulers productivity on many areas of study actually increased. In fact, he amazingly produced one mathematical paper per week in the year 1775.

He solved the famous Basel problem which asked for the sum of the reciprocals of the perfect squares quite remarkably, the answer is 1 1 1 1 2 + 2 + 2 + 2 + = . 2 6 1 2 3 4 Euler also established the following formula concerning complex numbers, the exponential function, and trigonometric functions. ei = cos + i sin 9