Graph Theory

Dr. S. Ramesh
Assistant Professor
Department of Mathematics
School of Science, GITAM.
Chapter 4

Graph Theory

Theorem 1. State and prove fundamental theorem (first theorem) on graphs. [Hand-
shake lemma].
Proof. Let G = (V, E) be a graph with V = {v1 , v2 , . . . , vn } and E = {e1 , e2 , . . . , eq }
so that the number of edges in G is q. We need to prove that
n
X
(1) d(vi ) = 2q.
i=1

For 1 ≤ i ≤ n and 1 ≤ j ≤ q, define

(
1, ej is incident on vi ,
xij =
0, otherwise.
Clearly,
q
n X q n
X X X
(2) xij = xij .
i=1 j=1 j=1 i=1

We shall interpret the identity in (2) in two ways and obtain (1). We have
q
n X n q n
X X X X
(3) LHS = xij = 1= d(vi ).
i=1 j=1 i=1 j=1 i=1
ej is incident on vi

Also,
q n q n q
X X X X X
(4) RHS = xij = 1= 2 = 2q.
j=1 i=1 j=1 i=1 j=1
ej is incident on vi

1
Now, (1) follows from (3),(4) and (2).
Theorem 2. In a graph, the number of odd vertices is always even.
Proof. Let v1 , v2 , . . . , vn denote the n vertices of the given graph and q denote the
number of edges. By Theorem 1, we have
n
X n
X n
X
(5) 2q = d(vi ) = d(vi ) + d(vi ) = Σ1 + Σ2 ,
i=1 i=1 i=1
d(vi ) is even d(vi ) is odd

say. Since each term in Σ1 is even the sum Σ1 is even. Since 2q is an even number,2q−
Σ1 = Σ2 is even as the difference of two even numbers is even. Clearly the number
of terms in Σ2 is the number of odd vertices and each term in Σ2 is odd. Since the
sum of an even number of odd terms is even and the sum of an odd number of odd
terms is odd, it follows that Σ2 should contain an even number of terms so that the
number of odd vertices is even.
Theorem 3. A graph G is disconnected if and only if the vertex set V of G can be
partitioned into two non-empty disjoint sets V1 and V2 in such a way that there exists
no edge in G with one end vertex in V1 and the other in V2 .
Proof. Suppose that G is disconnected. Then there exist two vertices a and b in V
such that there is no path between them. Consider the component Ca . Let V1 denote
the set of vertices in Ca . V1 is non-empty since a ∈ V1 . Let V2 = V − V1 . Clearly V2
is non-empty since b ∈ V2 . Also,V = V1 ∪ V2 . It remains to show that there exists no
edge in G with one end vertex in V1 and the other in V2 . On the contrary assume
that there exists an edge say e with one end vertex x in V1 and the other end vertex
y in V2 . Since x ∈ V1 , it is a vertex in Ca . Hence there is a path between a and x.
On combining this path with the edge e we obtain a path between a and y. Hence
y ∈ Ca so that y ∈ V1 . But y ∈ V2 , V1 and V2 are disjoint. This contradiction shows
that we can not find such an edge (with one end vertex in V1 and the other in V2 ).
Conversely suppose that V can be partitioned in to two non-empty disjoint sets
V1 and V2 such that there exists no edge with one end vertex in V1 and the other in
V2 . To prove that G is disconnected. Since V1 and V2 are non -empty, we can choose
a vertex a from V1 and another vertex b from V2 . We claim that there is no path
between a and b. If a path existed between a and b, it would follow that there is
an edge with one end vertex in V1 and the other in V2 , contrary to our assumption.
Thus there is no path between a and b and hence G is disconnected.
Note 4. Let G be a graph in which the degree of every vertex is even and positive.
Then G contains a cycle (circuit).

2
Theorem 5. Let G be a connected graph. Then G is an Euler graph if and only if
the degree if each vertex in G is even.

Proof. Let G be a connected graph.

Assume that G is an Euler graph. Then G contains an Euler line (a closed walk
containing all the edges of G.) say ρ. Since the Euler line ρ contains all the edges, ρ
not only contains all the edges but contains all the vertices also. We now show that
the degree of each vertex in ρ is even. First let x be an intermediate vertex in ρ. It is
preceded by an edge and succeeded by another edge. Hence the contribution of these
edges to the degree of x is two. If this vertex x appears k times in ρ. the degree of x
would be 2k, which is an even number. The same argument holds good even when
x is a beginning(=end) vertex. Thus the degree of each vertex in G is even.
Suppose the degree of each vertex is even in G. By Note (4), G has circuit h1
(say). If h contains all edges in G, then nothing else to prove. Suppose h does not
contain all edges in G, then G1 = G − h1 is a subgraph containing the edges in G
but not in h1 and the degree of each vertex in G1 is even (because only even number
of edges have gone into h1 ). Since G is connected, h1 and G1 have a common vertex
a (Suppose there is no common vertex, let V1 be the number of vertices in h1 and V2
is the number of vertices in G1 . Then V1 ̸= ∅, V2 ̸= ∅, V1 ∩ V2 = ∅, V1 ∪ V2 = V and
there is no edge where one end vertex in V1 and the other end vertex in V2 . Which
is a contradiction to G is connected). Since the degree of each vertex is even in G1 ,
G1 has a circuit h2 (say). If h2 contains all the edges in G1 , then h1 ∪ h2 will be an
Euler line. By continuing this process, it is possible to construct a closed walk by
covering every edge exactly once. Therefore G is an Euler graph.

3
Chapter 5

Trees

Definition 6. A connected graph is said to be a tree if it is free from circuits.

5.1 Tree Properties

Theorem 7. A connected graph G is a tree if and only if there is a unique path in
between every pair of vertices.

Proof. Let G be a connected graph.

Suppose that G is a tree. Since G is connected, there is a path in between every pair
of vertices. Suppose there two paths in between some pair of vertices. Then it gives
a circuit, which is a contradiction to G is a tree. Therefore there is always a unique
path in between every pair of vertices.
Conversely, suppose there is a unique path in between pair of every vertices in G. Its
is enough to show that G is free from from circuits. Suppose there is a circuit in G.
Then we can have two paths in between every pair vertices appeared in the circuit.
Which is a contradiction to our assumption. Hence G is a tree.

Note 8. If T is a tree and e is an edge in T , then T − {e} is disconnected and

decomposed into two distinct component.

Theorem 9. A tree with n vertices has n − 1 edges.

Proof. We will prove this theorem by using complete induction on n. Given a positive
integer n, say P (n) is the statement “A tree with n-vertices has (n−1)-edges.” Clearly
P (1) is true. Assume that P (k) is true for 1 < k ≤ n. Let T be a tree with (n + 1)
number of vertices and e is an edge in T with end vertices x and y. Then T − {e}

4
is disconnected and decomposed into two components Cx and Cy . Let n1 and n2
be the number of vertices in Cx and Cy respectively. Then x ∈/ Cy , y ∈
/ Cx and
Cx ,Cy are connected subgraphs and hence trees and moreover n1 < n, n2 < n and
n1 + n2 = n. By our assumption, Cx and Cy have (n1 − 1) and (n2 − 1) number of
edges respectively. Therefore

Total number of edges = (n1 − 1) + (n2 − 1) + 1 = (n1 + n2 ) − 1 = (n + 1) − 1

This completes the induction and hence proof of the theorem.

Note 10. If G is a graph with n-vertices and (n − 1)-edges, where n ≥ 2. Suppose

that G does not contain isolated vertices. Then G has at least two pendant vertices.

Note 11. If G is a connected graph and p is a pendant vertex, then G − {p} is

connected.

Theorem 12. Let G be a connected graph n-vertices and (n − 1)-edges. Then G is

a tree.

Proof. We will prove this theorem by using complete induction on n. Given a positive
integer n, say P (n) is the statement “If G is a connected graph n-vertices and (n−1)-
edges, then G is a tree.” Clearly P (1) is true. We assume the conclusion is true for
(n − 1), where n ≥ 2. Let G be a connected graph with n-vertices and (n − 1)-edges.
Since G is connected, it is free from isolated vertices. By Note. (10), we can find a
pendant vertex p in G. Then G − {p} is connected (by Note. (11)) with (n − 1)-
vertices and (n − 2)-edges. By our assumption, G − {p} is a tree. To show G is tree,
it enough to show G is free from circuits (we know G is already connected). Suppose
there is a circuit h in G. Since the pendant vertex p is not appear in any circuit, h
completely lies in G − {p}. Which is a contradiction to G − {p} is a tree. Therefore
G is free from circuits and hence it is a tree. Which completes the induction and
hence the theorem.

Theorem 13. Let G be a circuitless graph with n-vertices and (n − 1)-edges. Then
G is connected and hence it is a tree.

Proof. Suppose G has k-distinct components namely C1 , C2 , · · · , Ck . Let ni be the

number of vertices in ith component Ci , for i = 1, 2, 3, · · · k. Then n1 + n2 + n3 +
· · · + nk = n. Since each component is tree, each component Ci has (ni − 1) edges.
Therefore (n1 − 1) + (n2 − 1) + · · · + (nk − 1) = n − k = n − 1. Hence k = 1. Thus
G is connected and hence it is a tree.

5
Definition 14. A connected graph is said to be a minimally connected if it is dis-
connected, while removing an edge from it.
Theorem 15. A connected graph is minimally connected if and only if it is a tree,
Proof. Let G is a connected graph.
Suppose that G is minimally connected. To prove G is a tree, it is enough to show
G is free from circuits. Suppose G has a circuit h (say). While removing an edge
from h, G it self connected. Which is a contradiction to our assumption. Therefore
G is a tree.
Conversely suppose G is a tree. Then there is a unique path in between every pair of
vertices. While removing an from G, G itself a connected. Therefore G is minimally
connected.
Definition 16. A connected graph is said to be an acyclic graph if it is free from
circuits. An acyclic graph is said to be a maximal acyclic graph if it has a circuit,
while adding a new edge to it.
Theorem 17. A connected graph is a maximal acyclic graph if and only if it is a
tree.
Proof. Let G be a connected graph. Suppose G is a maximal acyclic graph. It is
enough to show G is free from circuits. Suppose there is a circuit in G. Which is a
contradiction to G is maximal acyclic.
Conversely, suppose G is a tree. Then there is a unique path in between every pair
of vertices. While adding a new edge in between any pair of vertices, it creates a
circuit in it. Therefore G is maximal acyclic.
Theorem 18. Every connected graph has a spanning tree.
Proof. Let G be a connected graph with n-vertices and q-edges. If G is free from
circuits, then G itself a tree and spanning tree. Suppose G has a circuit and e1 is
an edge in the circuit. Then, by Note. (11), G1 = G − {e1 } is connected, with
n-vertices and it has q − 1 number of edges. If G1 is a free from circuits, then G1 is a
tree and itself a spanning tree. Suppose G1 has a circuit and e2 is an edge from the
circuit. Then G2 = G1 − {e2 } is connected, with n-vertices and it has q − 2 number
of edges. Since the number of edges are finite, by continuing this process we will get
a spanning tree at some k th level. Therefore G has a spanning tree.
Definition 19. A tree is said to be a binary tree if there is exactly one vertex of
degree 2 and remaining all are of degree 1 or 3. The vertex of degree 2 is called the
root vertex.

6
Note 20. Every binary tree has odd number of vertices only. It means that there is
no binary tree with even number of vertices.
(n + 1)
Theorem 21. If T is a binary tree with n-vertices (n is odd), then it has
2
number of pendant vertices (the degree is one).

Proof. Let T be a tree with n- vertices in which T has p number of pendant vertices
and v1 is the root vertex. Then, by Fundamental Theorem on Graphs,
n
X
2(n − 1) = d(vi )
i=1
n
X
= d(v1 ) + d(vi )
i=2
n
X n
X
=2+ d(vi ) + d(vi )
d(vi )=1 d(vi )=3
X n n
X
=2+ 1+ 3
d(vi )=1 d(vi )=3

= 2 + p + (n − p − 1)
= 2p + 3n − 1

(n + 1)
Therefore p =
2