Project Report on Classification Theorem and

Fundamental Group of a Surface

Santanil Jana

September 13, 2016

 Abstract

The sphere, torus, Klein bottle, and the projective plane are the classical examples of orientable and non-
orientable surfaces. As with much of mathematics, it is natural to ask the question: are these all possible
surfaces, or, more generally, can we classify all possible surfaces? In the first chapter, we examine a result
originally due to Seifert and Threlfall that all compact surfaces are homeomorphic to the sphere, the connect
sum of tori, or the connect sum of projective planes; for this report, we follow a modern proof from Lee [3]. For
the 2nd chapter, we will study a much more global property of a space using some abstract algebraic notion.
Contents

2
Chapter 1

Classification of Compact Surfaces

1.1 Introduction

In our daily life we identify Surface as a continuous set of points that has length and breadth but no thickness.
Common examples are sphere, torus, cone and so forth. In mathematics, we generalise the notion of Surfaces
in R3 to Rn for any n ∈ N, n ≥ 2. The analogue of Surface in Rn is called ”Manifold”. We note that locally a
Surface looks like a plane. This motivates the definition of a Manifold, which can be defined as an object which
is locally Euclidean. We will formally define the notion of a ”n-Manifold”, but in this chapter we will restrict
ourselves to the Surface, specially Compact Surfaces. A Compact Surface is nothing but a surface which is
also compact in R3 . Common examples of compact surfaces are spheres, torus, projective plane. There are
many more compact surfaces, but later in this chapter we will show that they are topologically equivalent to
the above three or some combination of them.

1.2 Surfaces and Orientability

We begin this section by giving a formal definition of a n-Manifold.

Definition 1.1. Let n ∈ N, A n dimensional Manifold(in short n-Manifold) is a Hausdorff space such
that each point has an open neighbourhood homeomorphic to the open n dimensional disk U n (= x ∈ Rn : |x| < 1).

We now look at some common examples of n-Manifold.

Example 1.2. Clearly the Euclidean n-space is trivially a n-Manifold.

Example 1.3. The sphere S n = {x ∈ Rn+1 : |x| = 1} is a n-Manifold. Let p = (1, 0, ..., 0), then S n \{p} is
homeomorphic to R. For the point p, the set (x1 , x2 , ...xn+1 ) ∈ S n : x1 > 0 is a neighbourhood homeomorphic
to U n . This is easily seen by just projecting the points of this set on Rn with x1 goes to zero.

Example 1.4. If M n is a n-Manifold, then any open subset of it is also a n-Manifold.

Definition 1.5. A Surface is a 2-Manifold, by this we mean that it is a Hausdorff space which is locally
homeomorphic to the plane R2 .

The classic examples are the sphere, the torus, the Klein bottle and the projective plane.
The torus T2 is a subset of R3 which is obtained by rotating a circle of radius 1 centred at (2, 0, 0) around
the z-axis.
Equivalently, we see that the torus is homeomorphic to the quotient space of I × I (where I denotes the
closed unit interval) modulo the equivalence relation given by (x, 0) ∼ (x, 1) for all x ∈ I and (0, y) ∼ (1, y) for
all y ∈ I.
In general given any even sided polygon, identifying the edges pairwise by Quotient Topology will result
in a Surface. But the question is whether the converse is true or not. We will show that the converse is also
true and to achieve this goal we will have to first prove that every surface can be covered by triangles(in some
sense). But before going into that we will define what is called orientability of a surface.

Definition 1.6. A Surface is called orientable if given any two coordinate system of a neighbourhood of a
given point, the change of coordinate has a positive Jacobian.

3
 Common examples are the sphere, the torus and the cylinder.
The above definition is equivalent to saying that there exists a well defined Normal map for the whole
surface. This immediately gives us an idea of non-orientable surfaces. Surfaces which doesn’t have a well
defined Normal map are called non-orientable.

Lemma 1.7. Let S be a connected surface and N1 and N2 are two unit normals defined on S. Then either
N1 = N2 or N1 = −N2 .

Proof. Consider A = {p ∈ S : N1 (p) = N2 (p)} and B = {p ∈ S : N1 (p) = −N2 (p)}. Then A and B are closed
subsets of S as N1 and N2 are continuous. Also S = A ∪ B. But as S is connected, either A = S and B = φ
or A = φ and B = S. This proves the lemma.

The Mobius strip is a topological space that is described by the Quotient space of the rectangle {(x, y) ∈
R2 : −10 ≤ x ≤ 10, −1 ≤ y ≤ 1} by identifying points (10, y) and (−10, −y) for −1 < y < 1. This is a
non-orientable surface. Intuitively it is clear. Mobius strip is connected and so have only two normal direction.
If we choose a specific direction(outward) for the normal at a point and move along the surface, we will end
up at the same point but with normal directing in the opposite direction.

1.3 Triangulation
Definition 1.8. Let v0 , v1 , ..., vk be in Rn be such that {v0 − v1 , v0P
− v2 , ..., v0 − vk } is linearly P
independent.
The Simplex spanned by these points is the set σ = {x ∈ Rn : x = ki=0 ti vi s.t. 0 ≤ ti ≤ 1 and ki=0 ti = 1}
with the subspace topology.

Each point vi is a vertex of σ and the dimension of σ is k.

Definition 1.9. Let {v0 , ...., vk } be vertices of a simplex σ. The simplex spanned by each non-empty subset
of {v0 , ...., vk } is a face of σ. The simplex spanned by a proper subset of vertices is a proper face. The
(k − 1)-dimensional faces are called boundary faces.

Figure 1. From left to right: 0-simplex, 1-simplex, 2-simplex, 3-simplex

Definition 1.10. A Euclidean Simplicial complex K is a collection of simplices in Rn satisfying the

following conditions:
(1) If σ ∈ K, then every face of σ is in K.
(2) The intersection of any two simplices in K is either empty or a face of each.
(3) Every point in a simplex of K has a neighbourhood that intersects finitely many simplices of K.

The dimension of K is the maximum dimension of any simplex in K.

The following is an example of a valid simplicial complex.

Figure 2. A Two dimensional simplicial complex

For 2-dimensional simplicial complexes, like those pictured above, condition 2 means that simplicies in-
tersect at either vertices or edges. The following is an example of condition 2 being broken:

4
 Figure 3. Not a simplicial complex

Given a Euclidean simplicial complex, the union of all simplices is a topological space with the subspace
topology induced from Rn .
In this section we will not go further in details of simplices, but rather we will now define triangulation of
a surface using the concept of simplices. The concept of triangulation will be the major tool in proving the
classification theorem for compact surfaces.

Definition 1.11. A polyhedron is a topological space that is homeomorphic to a Euclidean simplicial complex.

Definition 1.12. A triangulation is a particular homeomorphism between a topological space and a Euclidean
simplicial complex.

Note that there can be multiple triangulations for a given surface. Recall that I ×I/ ∼ with the equivalence
relation given by (x, 0) ∼ (x, 1) for all x ∈ I and (0, y) ∼ (1, y) for all y ∈ I is homeomorphic to a torus. We
can make I × I into a simplicial complex K as pictured below:

Figure 4. Minimal triangulation of a torus

The homeomorphism between this simplicial complex with the equivalence rela- tion ∼ from above and the
torus is a triangulation of the torus. We are only interested in triangulation of compact surfaces. We know
that every surface is locally homeomorphic to R2 and triangles in R2 is a 2-simplex. Keeping these two facts
in mind we give this definition.

Definition 1.13. A triangulation of a compact surface S consists of finite family of closed subsets
{T1 , T2 , ..., Tn } that cover S, and a family of homeomorphisms {φi : Ti0 → Ti : i = 1, 2, ..., n}, where each
Ti0 is a triangle in R2 . The subsets Ti are called ’triangles’. The subsets of Ti that are the images of the vertices
and edges of the triangle Ti0 under φi are also called ’vertices’ and ’edges’, respectively. Finally, it is required
that any two distinct triangles, Ti and Tj ,either be disjoint, have a single vertex in common, or have one entire
edge common.

Given any compact surface, it seems possible that it can be covered by some ’triangles’. The next theorem
establishes this fact.

Theorem 1.14. (Triangulation Theorem for 2-Manifolds). Every 2-Manifold is homeomorphic to the
polyhedron of a 2-dimensional simplicial complex, in which every 1-simplex is a face of exactly two 2-simplices.

Proof. The proof of this result is long and intricate, and, thus, we shall not present it here. The basic approach
is to cover the manifold with regular coordinate disks and show that each disk can be triangulated compatibly.

5
The main lemma that is needed is the Schonies Theorem, which states that a topological embedding of the
circle into R2 can be extended to an embedding of the closed disk. A proof of the Schonies Theorem and the
triangulation theorem for surfaces can be obtained in Mohar and Thomassen [1].

1.4 Polygonal Representation and Connected Sum

Definition 1.15. A subset P of the plane is a polygonal region if it is a compact subset whose boundary is
a one dimensional Euclidean simplicial complex satisfying the following conditions:
(1) Each point q of an edge that is not a vertex has a neighbourhood U ⊂ R2 such that P ∩ U is equal to the
intersection of U with a closed half-plane {(x, y) : ax + by + c ≥ 0}
(2) Each vertex v has a neighbourhood V ⊂ R2 such that P ∩ V is equal to the intersection of V with two
closed half-planes whose boundaries intersect at v.

Figure 5. Left: Condition 1. Right: Condition 2.

Theorem 1.16. Let P be a polygonal region in the plane with an even number of edges and suppose we are
given an equivalence relation that identifies each edge with exactly one other edge by means of a Euclidean
simplicial isomorphism. Then the resultant quotient space is a compact surface.

Proof. Let M be the quotient space P/ ∼ and let π : P → M denote the quotient map. Since P is compact,
π(P ) = M is compact. The equivalence relation identifies only edges with edges and vertices with vertices so
the points of M are either:
(1) face points - points whose inverse image in P are in Int(P ).
(2) edge points - points whose inverse images are on edges but not vertices.
(3) vertex points - points whose inverse images are vertices.
To prove that M is locally Euclidean, it suffices to consider the three types of points.
Face points - Because π is injective on Int(P ) and π, being a quotient map is onto, π is bijective on
Int(P ). So by the closed map lemma, π is a homeomorphism on Int(P ). Since Int(P ) ⊂ R2 is a open set,
R2 ∼= Int(P ) ∼= π(Int(P )), so every face point is in a locally Euclidean neighbourhood, namely π(Int(P )).
Edge points - For any edge point q, pick a sufficiently small neighbourhood N such that there is no vertex
points in N . By the definition of a polygonal region, q has two inverse images, q1 and q2 with neighbourhoods
U1 and U2 such that V1 = U1 ∩ P and V2 = U2 ∩ P are disjoint half-planes. Furthermore, notice that π|V1 ∩V2
is also a quotient map. We construct affine homeomorphism f1 and f2 such that f1 maps V1 to a half disk
on the upper half plane and f2 maps V2 to the lower disk on the lower half plane. We can shrink V1 and V2
until they are saturated open sets in P ; i.e., for every boundary point of V1 , the corresponding boundary point
is in V2 and vice versa. We can now define another quotient map ψ : V1 ∩ V2 → R2 such that ψ = f1 on V1
and ψ = f2 on V2 . Modulus the equivalence relation r1 ∼ r2 , where r1 and r2 are edge points in V1 and V2
respectively, whenever ψ(r1 ) = ψ(r2 ). Notice that ψ is a quotient map onto a Euclidean ball centred at the
origin and makes the same identifications as π. By the uniqueness of the quotient map, the quotient spaces
are homeomorphic, so edge points are locally Euclidean.

6
 Vertex points - Repeat the same process as the edge points, but this time there will be multiple pieces of the
polygon that are identified in R2 . The resultant quotient space is homeomorphic to an open ball, so we may
conclude by appealing to the uniqueness of the quotient map. Therefore, we know that M is locally Euclidean.

To show that M is Hausdorff, simply pick sufficient small balls. Since M is the quotient space of the quotient
map from the polygonal region P, the pre-image of any pair of points in M can be separated into disjoint open
sets by picking sufficiently small open balls; the image of these open balls will be open sets in M that separate
the two points in M .

The converse of this theorem is also true. But we will come to that later. Now we look at some examples.

Example 1.17. The sphere S2 = {(x, y, z) ∈ R3 : x2 + y 2 + z 2 = 1} is homeomorphic to the square region

S = {(x, y) ∈ R2 : |x| + |y| ≤ 1} modulo the equivalence relation (x, y) ∼ (−x, y) for (x, y) ∈ ∂S.
Example 1.18. The torus T2 is homeomorphic to the square region S = {(x, y) ∈ R2 : |x| + |y| ≤ 1} modulo
the equivalence relation (x, y) ∼ (−x, −y) for (x, y) ∈ ∂S.
Example 1.19. The Klein bottle K2 is homeomorphic to the square region S (as defined in Example 1.17)
modulo the equivalence relation (x, y) ∼ (−x, −y) for (x, y) ∈ ∂S such that 0 ≤ x, y ≤ 1 or −1 ≤ x, y ≤ 0,
and another equivalence relation (x, y) ∼ (−y, −x) for (x, y) ∈ ∂S such that −1 ≤ x ≤ 0 ≤ y ≤ 1 or
−1 ≤ y ≤ 0 ≤ x ≤ 1.

Figure 6. Klein Bottle

Example 1.20. The projective plane P2 is homeomorphic to the square region modulo the equivalence relation
(x, y) ∼ (−x, −y) for (x, y) ∈ ∂S.

7
 Figure 7. Identification that yields a Projective Plane

We want to join two surfaces such that the resulting space is also a surface. For that we define the notion
of connected sum of surfaces. As an example the connected sum of two tori is a two hole torus.
Definition 1.21. Let S1 and S2 be disjoint surfaces. Their connected sum, denoted by S1 #S2 , is formed by
cutting a small circular hole in each surface and then gluing the surfaces together along the holes.
The above definition informally defines the notion of connected sums. To be more precise, we give a rigorous
mathematical construction of the connected sum S1 #S2 .
construction: We choose subsets D1 ⊂ S1 and D2 ⊂ S2 such that D1 and D2 are homeomorphic to the closed
disk. Let Si0 = Si \int(Di ) for i = 1 and 2. We choose a homeomorphism h of ∂D1 onto ∂D2 . Then S1 #S2 is
the quotient space of S10 ∪ S20 obtained by identifying the points x and h(x) for all x ∈ ∂D1 .
Next we state a very elementary lemma concerning the connected sum of two surfaces.
Lemma 1.22. The connect sum M1 #M2 of two connected surfaces M1 and M2 is a connected surface.
Proof. The proof of this lemma is similar to Theorem 1.16. The main idea is to identify the surface as
a quotient space of a polygon and then study the image of the neighbourhoods of the three types of points
described in Theorem 1.16 under the quotient map.

Definition 1.23. Given a set S, a word in S is an ordered k-tuple of symbols of the form a or a−1 where
a ∈ S. The length of a word is the number of elements in the word, where a and a−1 count as distinct elements.
Definition 1.24. A polygonal presentation is a finite set S with finitely many words W1 , ..., Wk , where Wi
is a word in S of length 3 or longer. We denote a polygonal presentation P = {S|W1 , ..., Wk }.
Example 1.25. Suppose S = {a, b}. W1 = {aba−1 b−1 } and W2 = {aa}. Then P = {a, b|aba−1 b−1 , aa}.
Definition 1.26. In the special case where Wi is a word of length 2, we define Pi to be a sphere if the word
is aa−1 and the projective plane if the word is aa.
Definition 1.27. A surface presentation is a polygonal presentation such that each symbol a ∈ S occurs
only exactly twice in W1 , ..., Wk , counting each a or a−1 as one occurrence.
Example 1.28. The common surfaces S2 , T2 , K and P2 all have presentations:
(1) The sphere: {a|aa−1 } or {a, b|abb−1 a−1 }
(2) The torus: {a, b|aba−1 b−1 }
(3) The projective plane: {a|aa} or {a, b|abab}
(4) The Klein Bottle: {a, b|abab−1 }
We will now state the converse of Theorem 1.16. This theorem will be used to prove the classification
theorem.
Theorem 1.29. Every compact surface admits a polygonal presentation.
Proof. We shall not prove this theorem here. For a detailed proof of the theorem we refer to [2].

From now on we will use the following notations:

(1) e denotes any symbol not in S.
(2) W1 W2 denotes the word formed by concatenating W1 and W2 .
(3) (a−1 )−1 = a.

Definition 1.30. The folowing are Elementary Transformation of a polygonal representation:

−1
a) Reflecting: {S|a1 , ..., an } → {S|a−1
m , ..., a1 }.

8
 b) Rotating: {S|a1 , ..., an } → {S|a2 , ....am , a1 }.

c) Cutting: {S|W1 w2 } → {S|W1 ee−1 W2 }.

d) Pasting: {S|W1 e, e−1 W2 } → {S|W1 W2 }.

W1 and W2 must have length atleast 2.

e) Folding: {S|W1 ee−1 , W2 } → {S|W1 , W2 }. W1 must have length atleast 3.
f ) Unfolding: {S|W1 , W2 } → {S|W1 ee−1 , W2 }

There are two important results which we will state next.

1) Elementary transformation of a polygonal representation gives rise to topologically invariant spaces.
2) If S1 and S2 are represented by the words W1 and W2 then S1 #S2 is represented by the word W1 W2 .

1.5 The Classification Theorem

We begin this section by proving some lemmas.
Lemma 1.31. The Klein bottle is homeomorphic to P2 #P2 .
Proof. Klein bottle is represented by {a, b|abab−1 }. By a sequence of elementary transformation we get
{a, b|abab−1 } ∼
= {a, b, c|abc, c−1 ab−1 } (cut along c)
∼
= {a, b, c|bca, b−1 c−1 a} (rotate)
∼
= {a, b, c|bca, a−1 cb} (reflect)
∼
= {a, b, c|bcbc} (paste along a and rotate)
This the representation of P2 #P2 .

Lemma 1.32. The connected sum T2 #P2 is homeomorphic to P2 #P2 #P2 .

Proof. By the previous lemma, P2 #P2 #P2 ∼ = K2 #P2 .
2 2 −1
K #P = {a, b, c|abab cc}.
∼
= {a, b, c|cabab−1 c}. (rotate)
∼
= {a, b, c, d|cabd−1 , dab−1 c}. (cut along d)
∼
= {a, b, c, d|abd−1 c, c−1 ba−1 d−1 }. (rotate 1st word and reflect 2nd word)
∼
= {a, b, d, e|a−1 d−1 abe, e−1 d−1 b}. (paste along c and cut along e)
∼
= {a, b, d, e|ea−1 d−1 ab, b−1 de}. (rotate and reflect)

9
 ∼
= {a, d, e|a−1 d−1 adee} = T2 #P2 .

We now have all the necessary tools to prove the main theorem of this chapter. This theorem was first
proved in 1907 by Max Dehn and Poul Heegaard.

Theorem 1.33. Classification Theorem of Compact Surface: Every non-empty, compact, connected
2-manifold is homeomorphic to either a sphere or a connected sum of one or more torus or a connected sum
of one or more projective plane.

Proof. Given any compact surface M , this proof will show that by a sequence of elementary transformations
on its polygonal representation, we get a surface that has a polygonal representation homeomorphic to the
sphere, the connected sum of torus, or the connected sum of projective planes.
Step 1. without loss of generality we can assume that M admits a polygonal representation with one word.
For since M is connected, each word must have a letter in common with another word, so by repeated pasting,
rotation and reflection transformations, we get a polygonal presentation with only one word, which admits a
presentation with one face.
Step 2. If there is an adjacent complementary pair(i.e. a pair like (a...a−1 ), we may remove it by folding.
The only time, when an adjacent complementary pair cannot be removed is if it has length less than 3 i.e.
aa−1 . in which case, we have a sphere. Now we assume that the surface is not a sphere.
Step 3. Suppose we have a non-adjacent twisted pair(i.e. a pair like a...a). Then the word will take the
form U aV a, where U and V are non-empty words. By a sequence of elementary transformations we get:
{a, U, V |U aV a} ∼
= {a, b, U, V |U ab, b−1 V a} (Cutting)
∼
= {a, b, U, V |bU a, a−1 V −1 b} (rotate 1st word and reflect 2nd word)
∼
= {b, U, V |bbU V −1 } (pasting and then rotate)

We may have introduced new non-adjacent twisted pairs in the process. However, as that the set of symbols S
is finite, by repeating the same process as above, we can transform each non-adjacent twisted pair into adjacent
pairs without affecting the bb pair. So after a finite number of iterations, we get a word with no non-adjacent
twisted pairs and a string of adjacent complementary pairs. The complementary pairs can be removed by
repeating step 2, which does not increase the total number of non-adjacent twisted pairs.
Step 4. We have seen that the edges of the polygon must be identified in pairs. But the vertices may be
identified in sets of two, three, four and so on. Let us call two vertices of the polygon to be equivalent if and
only if they are to be identified. Clearly this is an equivalence relation. Let us choose some equivalence class of
vertices [v]. Suppose that there are vertices not in the equivalence class [v]. Then there must be some edge a
that connects [v] to some other vertex class [w]. Since this is a polygonal surface, the other edge that touches
a at [v] cannot be a−1 , or else we would have got rid of it in step 2. The other edge cannot be a, because, if it
were, then the initial and terminal ends would be identified under the quotient map, which is not the case. So
we label this other edge b and the other vertex x.
Somewhere else in the polygon, there is another edge labelled either b or b−1 . Without loss of generality,
assume that it is b−1 . The proof for b is similar except for an extra reflection. Thus the presentation is of the
form baXb−1 Y . By elementary transformations:
{a, b, X, Y |baXb−1 Y } ∼
= {a, b, c, X, Y |bac, c−1 Xb−1 Y } (cutting)
∼
= {a, b, c, X, Y |acb, b−1 Y c−1 X} (rotate)
∼
= {a, c, X, Y |acY c−1 X} (paste along b)

10
 By pasting the edges labelled b, we have reduced the number of distinct vertices in the polygon labelled v.
We may have increased the number of vertices labelled w and we may have introduced new complementary
pairs. To repair the latter, perform step 2 again. Note that step 2 does not increase the number of vertices
labelled v. Thus, by repeating this process finitely many times, we can eliminate the vertex class [v]. Repeating
this procedure for each vertex class, we can get the desired result.
Step 5.Now we claim that if the presentation has any complementary pairs a, a−1 , then it has another
complementary pair b, b−1 that occurs intertwined with the first. i.e., a, ..., b, ..., a−1 , ..., b−1 . On the contrary let
us assume that the presentation is of the form aXa−1 Y , where X and Y only contain matched complementary
pairs or adjacent twisted pair.(complementary pairs remain exclusively within X or Y .) Recall that non-
adjacent twisted pairs and adjacent complementary pairs are not possible by step 2 and 3. Thus each edge in
X is identified with another edge in Y and similarly for Y . This means the terminal vertices of a and a−1
both touch vertices in X and the initial vertices are identified with only vertices in Y . This is a contradiction,
since all vertices are within one equivalence class by Step 4.
Step 6. Now the presentation is given W aXbY a−1 Zb−1 . By a sequence of elementary transformations, we
get the following:
{a, b, W, X, Y, Z|W aXbY a−1 Zb−1 } ∼
= {a, b, c, W, X, Y, Z|W aXc, c−1 bY a−1 Zb−1 }
∼
= {a, b, c, W, X, Y, Z|XcW a, a−1 Zb−1 c−1 bY }
∼
= {a, b, c, W, X, Y, Z|XcW Zb−1 c−1 bY }
∼
= {a, b, c, W, X, Y, Z|c−1 bY XcW Zb−1 }
∼
= {a, b, c, W, X, Y, Z|c−1 bY Xcd, d−1 W Zb−1 }
∼
= {a, b, c, W, X, Y, Z|Y Xcdc−1 b, b−1 d−1 W Z}
∼
= {a, b, c, W, X, Y, Z|Y Xcdc−1 d−1 W Z}
∼
= {a, b, c, W, X, Y, Z|cdc−1 d−1 W ZY X}

So, by repeating this process M admits a presentation in which all intertwined complementary pairs occur
together with no other edges in between.
Step 7. By the previous steps we have seen that all twisted pairs occurs adjacent to each other, i.e.
aa, which is a projective plane. Also all complementary pairs occur like aba−1 b−1 , which is a torus. If the
presentation consists exclusively of either case, then we are done, since we would either have the connect sum
of torus or connect sum of projective planes. If the presentation contains both twisted and complementary
pairs, then the presentation must be one of the following forms: aabcb−1 c−1 X or bcb−1 c−1 aaX. In either case
by the previous lemma, T2 #P2 ∼ = P2 #P2 #P2 . So, if both P2 and T2 occurs in the presentation, then we can
eliminate T2 and get a connected sum of P2 by the above relation.

we proved that all compact surfaces are homeomorphic to the sphere, the connect sum of torus, or the
connect sum of projective planes, but we have yet to prove that the surfaces are topologically distinct. e.g.,
a sphere is not homeomorphic to a torus. The answer to this non-trivial question lies with other topological
invariants such as the Euler Characteristic and orientibility.

11
1.6 Euler Characteristic of a Surface
Definition 1.34. Let M be a compact surface with triangulation {T1 , T2 , ..., Tn }. Let v be the number of
vertices, e be the number of edges and f be the number of faces. Then The Euler Characteristics of the
surface M , denoted by χ(M ), is given by χ(M ) = v − e + f .

The Euler Characteristics remains invariant if we choose some different triangulation. The proof is tedious
and involved and can be proved by means of homology theory. For now we will use Euler Characteristics to
distinguish between compact surface.

Lemma 1.35. Let S1 and S2 be two compact surface. Then,

χ(S1 #S2 ) = χ(S1 ) + χ(S2 ) − 2

Proof. Let the number of vertices, edges and faces of S1 and S2 is v1 , e1 , f1 and v2 , e2 , f2 respectively. We
form their connected sum by removing the interior of a triangle and identifying the vertices and edges of the
removed triangles. Now the number of vertices in the triangulation of S1 #S2 is v1 + v2 − 3, edges is e1 + e2 − 3
and faces is f1 + f2 − 2. So by definition, χ(S1 #S2 ) = χ(S1 ) + χ(S2 ) − 2.

The Euler Characteristics of the three fundamental surfaces can be calculated from their triangulations.
The Euler Characteristics of the Sphere, Torus and Projective plane is 2, 0 and 1 respectively.From this and
the above lemma the Euler characteristics of a connected sum of n torus is 2 − 2n, connected sum of n Pro-
jective Planes is 2 − n. Note that the Euler Characteristics of an orientable surface is always even, but for
non-orientable surfaces it can be even or odd.
Assuming the topological invariance of Euler characteristics and using the classification theorem, we have the
following result:

Theorem 1.36. Two compact surfaces are homeomorphic if and only if their Euler characteristics are equal
and they are both orientable or both non-orientable.

Such a classification of topological spaces is very rare. There is no such theorem for compact 3-Manifolds
yet.

12
Chapter 2

Fundamental Group

2.1 Introduction
This chapter introduces one of the simplest and most important functors of algebraic topology, the fundamental
group, which creates an algebraic image of a space from the loops in the space, the paths in the space starting
and ending at the same point. This chapter begins with the basic definitions and constructions, and then pro-
ceeds quickly to an important calculation, the fundamental group of the circle. Then we show some application
of this fact and also how continuous maps are interpreted in terms of homomorphisms of the fundamental group.

2.2 Paths and Homotopy

The fundamental group will be defined in terms of loops and deformations of loops. Sometimes it will be useful
to consider more generally paths and their deformations.

Definition 2.1. A path in a space X is a continuous map f : I → X where I is the unit interval [0, 1].

The idea of continuously deforming a path, keeping its endpoints fixed, is made precise by the following
definition.

Definition 2.2. A homotopy of paths in X is a family ft : I → X, 0 ≤ t ≤ 1, such that

(1) The endpoints ft (0) = x0 and ft (1) = x1 are independent of t .
(2) The associated map F : I × I → X defined by F (s, t) = ft (s) is continuous.
When two paths f0 and f1 are connected in this way by a homotopy ft , they are said to be homotopic. The
notation for this is f0 ' f1 .

Example 2.3. Any two paths f0 and f1 in Rn having the same endpoints x0 and x1 are homotopic via the
homotopy ft (s) = (1 − t)f0 (s) + tf1 (s). During this homotopy each point f0 (s) travels along the line segment
to f1 (s) at constant speed. This is called Linear Homotopy. More generally that for a convex subspace
X ⊂ Rn , all paths in X with given endpoints x0 and x1 are homotopic, since if f0 and f1 lie in X then, by
definition of convex set, the homotopy ft also lies in X.

Theorem 2.4. The relation of homotopy on paths with fixed endpoints in any space is an equivalence relation.

Proof. Reflexivity is evident since f ' f by the constant homotopy ft = f . Symmetry is also easy since if
f0 ' f1 via ft , then f1 ' f0 via the inverse homotopy f1−t (i.e.F−1 (s, t) = F (s, 1 − t). For transitivity, if f0 ' f1
via ft and if f1 = g0 with g0 ' g1 via gt , then f0 ' g1 via the homotopy ht that equals f2t for 0 ≤ t ≤ 21
and g2t−1 for 21 ≤ t ≤ 1. These two definitions agree for t = 1/2 since we assume f1 = g0 . Continuity of the
associated map H(s, t) = ht (s) comes from the elementary fact from topology that a function defined on the
union of two closed sets is continuous if it is continuous when restricted to each of the closed sets separately.
In the case at hand we have H(s, t) = F (s, 2t) for 0 ≤ t ≤ 21 and H(s, t) = G(s, 2t − 1) for 12 ≤ t ≤ 1 where
F and G are the maps from the unit square I × I to X associated to the homotopies ft and gt . Since H is
continuous on I × [0, 21 ] and on I × [ 12 , 1], it is continuous on I × I.

Definition 2.5. The equivalence class of a path f under the equivalence relation of homotopy will be denoted
[f ] and called the homotopy class of f .

13
 Given two paths f, g : I → X such that f (1) = g(0), there is a composition or product path f · g that
traverses first f and then g, defined by the formula
f · g(s) = f (2s) if 0 ≤ s ≤ 21 and f · g(s) = g(2s − 1) if 21 ≤ s ≤ 1
Thus f and g are traversed twice as fast in order for f · g to be traversed in unit time. This product operation
respects homotopy classes since if f0 ' f1 and g0 ' g1 via homotopies ft and gt , and if f0 (1) = g0 (0) so that
f0 · g0 is defined, then ft · gt is defined and provides a homotopy f0 · g0 ' f1 · g1 . In particular, suppose we
restrict attention to paths f : I → X with the same starting and ending point f (0) = f (1) = x0 ∈ X. Such
paths are called loops and the common starting and ending point x0 is referred to as the base-point. The set
of all homotopy classes [f ] of loops at the base-point x0 is denoted π(X, x0 ).

Theorem 2.6. π(X, x0 ) is a group with respect to the product [f ] · [g] = [f · g].

The proof of this fact is rather long and tedious and involves only to show certain homotopies. For a
detailed proof we refer to [1] Chapter-2, Section-3.

Definition 2.7. The group π(X, x0 ) is called Fundamental Group of X at the base-point x0 .

Example 2.8. For a convex set X ⊂ Rn with base-point x0 ∈ X, we have π(X, x0 ) = {0}, the trivial group,
since any two loops f0 and f1 based at x0 are homotopic via the linear homotopy ft (s) = (1 − t)f0 (s) + tf1 (s).

Now we look at the dependence of π(X, x0 ) on the choice of the base-point x0 . Since π(X, x0 ) involves
only the path-component of X containing x0 , we can expect a relation between π(X, x0 ) and π(X, x1 ) for two
base-points x0 and x1 only if x0 and x1 lie in the same path-component of X. So let h : I → X be a path
from x0 to x1 , with the inverse path h(s) = h(1 − s) from x1 back to x0 . We can then associate to each loop
f based at x1 the loop h · f · h based at x0 . we  define a general n-fold product f1 · f2 · · · fn in which the
path fi is traversed in the time interval i−1 i
n n and a change of base-point map ψh : π(X, x1 ) → π(X, x0 )
,
by ψh ([f ]) = [h · f · h]. This is well-defined since if ft is a homotopy of loops based at x1 then h · ft · h is a
homotopy of loops based at x0 .

Lemma 2.9. The map ψh : π(X, x1 ) → π(X, x0 ) is an isomorphism.

Proof. ψh is a homomorphism since ψh ([f ·g]) = [h·f ·g ·h] = [h·f ·h·h·g ·h] = ψh ([f ])·ψh ([g]). Further ψh is an
isomorphism with inverse ψh as ψh ψh ([f ]) = ψh ([h · f · h]) = [h · h · f · h · h] = [f ]. Similarly ψh ψh ([f ]) = [f ].

Thus if X is path-connected, the group π(X, x0 ) is, up to isomorphism, independent of the choice of
base-point x0 . In this case the notation π(X, x0 ) is often abbreviated to π(X). In general, a space is called
simply-connected if it is path-connected and has trivial fundamental group.

Lemma 2.10. A space X is simply-connected if and only if there is a unique homotopy class of paths connecting
any two points in X.

Proof. We need to be concerned only with the uniqueness of connecting paths. The existence of paths, con-
necting two points, follows directly from path-connectedness. Suppose π(X) = {0}. If f and g are two paths
from x0 to x1 , then f ' f · g · g ' g since the loops g · g and f · g are each homotopic to constant loops, using the
assumption π(X) = {0} in the latter case. Conversely, if there is only one homotopy class of paths connecting
a base-point x0 to itself, then all loops at x0 are homotopic to the constant loop and π(X) = {0}

2.3 Fundamental Group of a Circle

In this section we will calculate the Fundamental group of a circle. Then we will also see some applications of
this theorem. But before that we introduce the notion of a Covering space.

Definition 2.11. Given a space X, a covering space of X consists of a space X and a map p : X → X
satisfying the following condition:
For each point x ∈ X there is an open neighbourhood U of x in X such that p−1 (U ) is a union of disjoint open
sets each of which is mapped homeomorphically onto U by p.

Such a U is called evenly covered. For example define p(s) = (cos 2πs, sin 2πs). This map can be visualized
geometrically by embedding R in R3 as the helix parametrized by s → (cos 2πs, sin πs, s), and then p is the
restriction to the helix of the projection of R3 onto R2 , (x, y, z) → (x, y).

14
To prove the theorem we will need just the following two facts about covering spaces.

Lemma 2.12. Let X be a covering space of X, p : X → X. Then

(a) For each path f : I → X starting at a point x0 ∈ X and each x0 ∈ p−1 (x0 ) there is a unique lift f : I → X
starting at x0 .
(b) For each homotopy ft : I → X of paths starting at x0 and each x0 ∈ p−1 (x0 ) there is a unique lifted
homotopy ft : I → X of paths starting at x0 .

This lemma is called the Lifting Lemma. This can be proved from the following much more general
statement: Given a map F : Y × I → X and a map F : Y × {0} → X lifting F |Y × {0}, then there is a unique
map F : Y × I → X lifting F and restricting to the given F on Y × {0}. We will not prove this here, but we
will now prove our main theorem in this chapter.

Theorem 2.13. π(S1 ) is an infinite cyclic group generated by the homotopy class of the loop w(s) = cos2πs, sin2πs)
based at (1, 0).

Proof. Note that [w]n = [wn ] where wn (s) = (cos 2πns, sin 2πns) for n ∈ Z. The theorem is therefore equivalent
to the statement that every loop in S1 based at (1, 0) is homotopic to wn for a unique n ∈ Z. To prove this the
idea will be to compare paths in S1 with paths in R via the map p : R → S1 given by p(s) = (cos2πs, sin2πs)
defined before. Observe that the loop wn is the composition pwn where wn : I → R is the path wn (s) = ns,
starting at 0 and ending at n, winding around the helix |n| times, upward if n > 0 and downward if n < 0.
The relation wn = pwn is expressed by saying that wn is a lift of wn . We will prove the theorem by studying
how paths in S1 lift to paths in R.
Let f : I → S1 be a loop at the base-point x0 = (1, 0), representing a given element of π(S1 , x0 ). By part
(a) of lemma 2.12 there is a lift f starting at 0. This path f ends at some integer n since pf (1) = f (1) = x0
and p−1 (x0 ) = Z ⊂ R. Another path in R from 0 to n is wn , and f ' wn via the linear homotopy. Composing
this homotopy with p gives a homotopy f ' wn so [f ] = [wn ].
To show that n is uniquely determined by [f ], suppose that f ' wn and f ' wm , so wm ' wn . Let ft be a
homotopy from wm = f0 to wn = f1 . By part (b) of lemma 2.12 this homotopy lifts to a homotopy ft of paths
starting at 0. The uniqueness part of (a) implies that f0 = wm and f1 = wn . Since ft is a homotopy of paths,
the endpoint ft (1) is independent of t . For t = 0 this endpoint is m and for t = 1 it is n, so m = n.

We now look at some applications of this theorem. Our first application is the Brouwer fixed point theorem
in dimension 2. But before that we give a couple of definitions.

Definition 2.14. A subset A of a topological space X is said to be a retract of X if there exists a continuous
map r : X → A such that r(a) = a for all a ∈ A. The map r is called a retraction.

Definition 2.15. A subset A of a topological space X is said to be a deformation retract of X if there exists
a continuous map r : X → A and a homotopy ft such that
f0 (x) = x x∈X
f1 (x) = r(x) x∈X
ft (a) = a a ∈ A, t ∈ I

Theorem 2.16. Every continuous map h : D2 → D2 has a fixed point.

Proof. Suppose on the contrary that there is no fixed point, i.e. there is no x ∈ D2 such that h(x) = x.
Define a map r : D2 → S1 by letting r(x) be the point of S1 where the ray in R2 starting at h(x) and passing
through x intersects boundary of D2 . Continuity of r is clear since small perturbations of x produce small

15
perturbations of h(x), hence also small perturbations of the ray through these two points. The crucial property
of r, besides continuity, is that r(x) = x if x ∈ S1 . Thus r is a retraction of D2 onto S1 . We will show that no
such retraction can exist. Let f0 be any loop in S1 . In D2 there is a homotopy of f0 to a constant loop, for
example the linear homotopy ft (s) = (1 − t)f0 (s) + tx0 where x0 is the base-point of f0 . Since the retraction r
is the identity on S1 , the composition rft is then a homotopy in S1 from rf0 = f0 to the constant loop at x0 .
But this contradicts the fact that π(S1 ) is non zero.

The techniques used to calculate π(S1 ) can be applied to prove the Borsuk–Ulam theorem in dimension
two:

Theorem 2.17. For every continuous map f : S2 → R2 there exists a pair of antipodal points x and −x in S2
with f (x) = f (−x).

The Borsuk–Ulam theorem holds more generally for maps Sn → Rn . The proof for n = 1 is easy since
the difference f (x) − f (−x) changes sign as x goes halfway around the circle, hence this difference must be
zero for some x. For n ≥ 2 the theorem is certainly less obvious. The theorem says in particular that there is
no one-to-one continuous map from S2 to R2 , so S2 is not homeomorphic to a subspace of R2 , an intuitively
obvious fact that is not easy to prove directly.

Proof. If the conclusion is false for f : S2 → R2 , we can define a map g : S2 → S1 by g(x) = |ff (x)−f
(x)−f (−x)
(−x)| . Define
2 3 1
a loop η circling the equator of S ⊂ R by η(s) = (cos 2πs, sin 2πs, 0), and let h : I → S be the composed
loop gη. Since g(−x) = −g(x), we have the relation h(s + 21 ) = −h(s) for all s ∈ [0, 21 ]. As we showed in the
calculation of π(S1 ), the loop h can be lifted to a path h0 : I → R. The equation h(s + 12 ) = −h(s) implies
that h0 (s + 12 ) = h0 (s) + 2q for some odd integer q which is independent of s since by solving the equation
h0 (s + 21 ) = h0 (s) + 2q for q we see that q depends continuously on s ∈ [0, 12 ], so q must be a constant since
it can only take integer values. In particular, we have h0 (1) = h0 ( 21 ) + 2q = h0 (0) + q. This means that h
represents q times a generator of π(S1 ). Since q is odd, we conclude that h is not null-homotopic. But h was
the composition gη : I → S2 → S1 , and η is obviously null-homotopic in S2 , so gη is null-homotopic in S1 by
composing a null-homotopy of η with g. Thus we have arrived at a contradiction.

An obvious corollary of this theorem is as follows:

Corollary 2.18. Whenever S2 is expressed as the union of three closed sets A1 , A2 and A3 , then atleast one
of these sets must contain a pair of antipodal points {x, −x}.

Lemma 2.19. π(X × Y ) is isomorphic to π(X) × π(Y ) if X and Y are path connected.

Proof. A basic property of the product topology is that a map f : Z → X × Y is continuous if and only if the
maps g : Z → X and h : Z → Y defined by f (z) = (g(z), h(z)) are both continuous. Hence a loop f in X × Y
based at (x0 , y0 ) is equivalent to a pair of loops g in X and h in Y based at x0 and y0 respectively. Similarly,
a homotopy ft of a loop in X × Y is equivalent to a pair of homotopies gt and ht of the corresponding loops
in X and Y . Thus we obtain a bijection π(X × Y, (x0 , y0 ) ∼ = π(X, x0 ) × π(Y, y0 ) given by [f ] 7→ ([g], [h]). This
is obviously a group homomorphism, and hence an isomorphism.

2.4 Induced Homomorphisms

Suppose φ : X → Y is a map taking the base-point x0 ∈ X to the base-point y0 ∈ Y . We write φ : (X, x0 ) →
(Y, y0 ) in this situation. Then φ induces a homomorphism φ∗ : π(X, x0 ) → π(Y, y0 ), defined by composing loops
f : I → X based at x0 with φ, that is, φ∗ [f ] = [φf ]. This induced map φ∗ is well-defined since a homotopy ft
of loops based at x0 yields a composed homotopy φft of loops based at y0 , so φ∗ [f0 ] = [φf0 ] = [φf1 ] = φ∗ [f1 ].
Furthermore, φ∗ is a homomorphism since φ(f g) = (φf )(φg), both functions having the value φf (2s) for
0 ≤ s ≤ 12 and the value φg(2s − 1) for 12 ≤ s ≤ 1.
Two basic properties of induced homomorphisms are:
1) (φψ)∗ = φ∗ ψ∗ for a composition (X, x0 ) → (Y, y0 ) → (Z, z0 ).
2) 1∗ = 1, which is a concise way of saying that the identity map 1 : X → X induces the identity map
1 : π(X, x0 ) → π(X, x0 ).
The first of these follows from the fact that composition of maps is associative, so (φψ)f = φ(ψf ), and the

16
second is obvious. So, clearly if φ is a homeomorphism then φ∗ is an isomorphism.

Lemma 2.20. If a space X retracts onto a subspace A, then the homomorphism i∗ : π(A, x0 ) → π(X, x0 )
induced by the inclusion i : A ,→ X is injective. If A is a deformation retract of X, then i∗ is an isomorphism.
Proof. Let r : X → A be a retraction. Then note that ri = 1. Hence r∗ i∗ = 1. This forces i∗ to be injective.
So, π(A, x0 ) is a subgroup of π(X, x0 ). Now take [f ] ∈ π(X, x0 ) Then r ◦ f is a loop in A and f ∼ = r ◦ f . Note
that a A deformation retracts to X, we have map H : X × I → X such that H(x, 0) = x; H(x, 1) = r(x) and
H(a, t) = a; ∀a ∈ A, t ∈ I, x ∈ X. So, define F : I ×I → X by F (t, s) = (H(f (t), s). Then clearly F (t, 0) = f (t)
and F (t, 1) = r(f (t) = r ◦ f (t). So, [f ] = [r ◦ f ] ∈ π(A, x0 ). Hence they are same upto isomorphism.

2.5 Free Groups

Before proceeding to the Van Kampen’s Theorem, we introduce the notion of a free group briefly. The detail
are avoided here. We will only need some simple results regarding the free group.
Definition 2.21. As a set, the Free Product ∗a Ga consists of all words g1 g2 · · · gm of arbitrary finite length
m ≥ 0, where each letter gi belongs to a group Gai and is not the identity element of Gai , and adjacent
letters gi and gi+1 belong to different groups Ga , that is, ai 6= ai+1 . Words satisfying these conditions are called
reduced, the idea being that unreduced words can always be simplified to reduced words by writing adjacent letters
that lie in the same Gai as a single letter and by cancelling trivial letters. The empty word is allowed, and
will be the identity element of ∗a Ga . The group operation in ∗a Ga is juxtaposition, (g1 g2 · · · gm )h1 h2 · · · hn =
g1 g2 · · · gm h1 h2 · · · hn . This product may not be reduced, however if gm and h1 belong to the same Ga , they
should be combined into a single letter (gm h1 ) according to the multiplication in Ga , and if this new letter
happens to be the identity of Ga , it should be cancelled from the product. This may allow gm−1 and h2 to be
combined, and possibly cancelled too. Repetition of this process eventually produces a reduced word. Inverse of
g1 g2 · · · gm is gm−1 · · · g −1 .
1

A basic property of the free product ∗a Ga that we will require is that any collection of homomorphisms
φa : Ga → H extends uniquely to a homomorphism Φ : ∗a Ga → H. Namely, the value of Φ on a word
g1 · · · gn with gi ∈ Gai must be φa1 (g1 ) · · · φan (gn ), and using this formula to define Φ gives a well-defined
homomorphism since the process of reducing an unreduced product in ∗a Ga does not affect its image under
Φ. For example, for a free product G ∗ H the inclusions G ,→ G × H and H ,→ G × H induce a surjective
homomorphism from G ∗ H to G × H.

2.6 Van Kampen’s Theorem

Suppose a space X is decomposed as the union of a collection of path-connected open subsets Aa , each of
which contains the base-point x0 ∈ X. By the remarks in the preceding paragraph, the homomorphisms
ja : π(Aa ) → π(X) induced by the inclusions Aa ,→ X extend to a homomorphism Φ : ∗a π(Aa ) → π(X). The
van Kampen’s theorem will say that Φ is very often surjective, but we can expect Φ to have a non-trivial kernel
in general. For if iab : π(Aa ∩ Ab ) → π(Aa ) is the homomorphism induced by the inclusion Aa ∩ Ab ,→ Aa
then ja iab = jb iba , both these compositions being induced by the inclusion Aa ∩ Ab ,→ X, so the kernel of Φ
contains all the elements of the form iab (w)iba (w)−1 for w ∈ π(Aa ∩ Ab ). Van Kampen’s theorem asserts that
under fairly broad hypotheses this gives a full description of Φ.
Theorem 2.22. If X is the union of path-connected open sets Aa each containing the base-point x0 ∈ X and
if each intersection Aa ∩ Ab is path-connected, then the homomorphism Φ : ∗a π(Aa ) → π(X) is surjective. If
in addition each intersection Aa ∩ Ab ∩ Ac is path-connected, then the kernel of Φ is the normal subgroup N
generated by all elements of the form iab (w)iba (w)−1 for w ∈ π(Aa ∩ Ab ), and hence Φ induces an isomorphism
∗a π(Aa )
π(X) ∼= .
N
We will not prove the theorem here. For a detailed proof we refer to [2] page-44.
We now turn our attention to calculate the Fundamental Group of some spaces using the above theorem and
the theory developed earlier in this chapter.
Theorem 2.23. π(Sn ) = 1 for n ≥ 2.

17
Proof. Take A1 and A2 to be the complements of two antipodal points in Sn . Choose a base-point x0 in A1 ∩A2 .
If n ≥ 2 then A1 ∩ A2 is path-connected. Then Van Kampen’s Theorem applies to say that every loop in Sn
based at x0 is homotopic to a product of loops in A1 or A2 . Both π(A1 ) and π(A2 ) are zero since A1 and A2
are homeomorphic to Rn . Hence every loop in Sn is null-homotopic and hence the claim follows.

Definition 2.24. Given a collection of spaces {Xa |a ∈ J} with base-points xa ∈ Xa , the Wedge Sum of the
Xa ’s is the quotient space of the disjoint union qa Xa obtained by identifying the points xa . It is denoted as
∨a Xa .
Theorem 2.25. The wedge sum of two circle has the Fundamental group Z ∗ Z.
Proof. Let S = S1 ∨ S1 . Let s be point which is identified. We want to calculate π(S, s). Let us take
U = S1 ∨ S1 \ {p} and V = S1 \ {q} ∨ S1 , where p, q belongs to different copies of S1 and not equal to s.
Note that, both U and V deformation retracts to S1 and U ∩ V is simply connected. So, by Van Kampen’s
Theorem π(S, s) = π(U, s) ∗ π(V, s), as the normal subgroup N is trivial as it is generated by identity. Hence
π(S) = Z ∗ Z.

Inductively, it is clear that the fundamental group of a wedge sum of n copies of circles is the free product
of n copies of Z.
We know that the fundamental group of a torus is Z × Z. We will now calculate the fundamental group of
two more surfaces namely the 2-holed torus and the punctured torus.
1) Fundamental Group of Punctured Torus:
The punctured torus is a torus with one point removed. Let S be the polygonal representation:

S represents the punctured torus.

claim: S deform retracts to the wedge of two circles.
Define the retraction r : S → bd(S) by, sending (x, y) to the intersection of the boundary of S with
the line joining (x, y) and (0, 0). We will divide the square in 4 parts by joining the corners with (0, 0).
Then depending on which part (x, y) lies, the value of r(x, y) changes. For example in the above picture,
r(x, y) = (−1, xy ). Then r is continuous and if (a, b) ∈ bd(S), then r(a, b) = (a, b). So r is a retraction. Now,
define H : S × I → S by, H((x, y), t) = (1 − t)(x, y) + tr(x, y). Then, H((x, y), 0) = (x, y), H((x, y), 1) = r(x, y)
and H((a, b), t) = (1 − t)(a, b) + tr(a, b) = (1 − t)(a, b) + t(a, b) = (a, b) for all (x, y) ∈ S, (a, b) ∈ bd(S), t ∈ I.
Hence H is a deformation retraction, which proves the claim.
So, by Lemma 2.20, we have that π(S, (1, 1)) ∼ = π(bd(S), (1, 1)). Now by the identification in the boundary
we see that bd(S) is nothing but wedge of two circles. Hence by Theorem 2.25, we have that π(S, (1, 1)) ∼ =
π(bd(S), (1, 1)) ∼
= π(S1 ∨ S1 , (1, 1)) ∼
= Z ∗ Z.

2) Fundamental Group of 2-holed Torus:

Let X denote the 2-holes torus. Then X has the following polygonal representation:

First we need to define open sets U, V ⊂ T2 that meet the conditions of the Seifert-van Kampen Theorem. Let

18
U, V be the subsets of T2 as pictured here:

This choice of U and V gives what we need. The fundamental group of U is the trivial group, since U is
contractible. Also, the fundamental group of U ∩ V is Z since it has S1 as a deformation retract.
Now we need to consider the fundamental group of V . One way to think about this is to consider V as a
wire (i.e. we only consider the edges of the octagon). Because V deform retracts onto the boundary. One
can define the retraction map r to be the map sending (x, y) to the intersection of the boundary with the line
joining (x, y) and the centre of the octagon. It is clear from the last problem that it is indeed a retraction.
And we again define the deformation retraction as the linear homotopy between identity map and r. Hence
π(V ) ∼
= π(∨4i=1 S1 ) ∼
= ∗4i=1 Z , as the boundary of the polygon with the identification is nothing but a wedge
sum of 4 circles. We now apply the Van Kampen’s theorem. So, π(X) ∼ =
π(U )∗π(V )
N . Where N is the normal
∼
subgroup generated by elements of the form i1 (w) and i2 (w), for w ∈ π(U ∩ V ) = Z. It is enough to look at the
image of 1 ∈ Z under i1 and i2 as ZZ is a cyclic group generated by 1. As, U is trivial, we have that i1 (1) = IdU .
Now we need to trace the loop of U ∩ V that generates π(U ∩ V, b) around V to see what we get. From the
picture before we see that this loop is homotopic to to the element ABA−1 B −1 DCD−1 C −1 We now take these
two elements, and form the normal subgroup N . Hence π(X) ∼ = {A, B, C, D|ABA−1 B −1 DCD−1 C −1 = 1}.

We can generalise this to g-holed torus (a genus g surface). If X is a surface of genus g, then π(X) ∼ =
{A1 , B1 , · · · Ag , Bg |[A1 , B1 ] · · · [Ag , Bg ] = 1}. Where [A, B] = ABA−1 B −1 .
The study of fundamental groups is just a beginning of a whole set of groups, called the Homotopy groups.
Fundamental group is just the first order homotopy group, which is determined by the structure of loops in
the space. In higher dimensional Homotopy groups, we will basically study how the higher dimensional loops
behave in a space and then try to study the properties of the space using this.

Acknowledgements:
Simply, I could not have done this work alone without the help of my supervisor. I would specially like to
thank Dr. B. Sury for providing the nice topics to study upon. Not only advising about my project work he
has also helped me out of some problems on several occasions and given me problems to work on, which really
triggers the thought process.

Reference:
1) A Basic Course in Algebraic Topology, by Willam S. Massey.
2) Algebraic Topology, by Alan Hatcher.
3) Introduction to Topological Manifolds, by J. M. Lee

19