R ESEARCH ARTICLE ScienceAsia 43 (2017): 326–333

doi: 10.2306/scienceasia1513-1874.2017.43.326

γ-total dominating graphs of paths and cycles

Alongkot Wongsriyaa , Nantapath Trakultraiprukb,∗
a
Department of Mathematics, Faculty of Science, Mahidol University, Bangkok 10400 Thailand
b
Department of Mathematics and Statistics, Faculty of Science and Technology, Thammasat University,
Pathum Thani 12120 Thailand
∗
Corresponding author, e-mail: [email protected]
Received 28 Sep 2016
Accepted 15 Oct 2017

ABSTRACT: A total dominating set for a graph G = (V (G), E(G)) is a subset D of V (G) such that every vertex in V (G)
is adjacent to some vertex in D. The total domination number of G, denoted by γt (G), is the minimum cardinality of
a total dominating set of G. A total dominating set of cardinality γt (G) is called a γ-total dominating set. Let T Dγ be
the set of all γ-total dominating sets in G. We define the γ-total dominating graph of G, denoted by T Dγ (G), to be
the graph whose vertex set is T Dγ , and two γ-total dominating sets D1 and D2 from T Dγ are adjacent in T Dγ (G) if
D1 = D2 \{u} ∪ {v} for some u ∈ D2 and v ∈/ D2 . In this paper, we present γ-total dominating graphs of paths and cycles.

KEYWORDS: total dominating set, total dominating subset, total domination number

MSC2010: 05C40

INTRODUCTION other by either adding or deleting a single vertex.

The authors gave some conditions for connectivity
Let G = (V (G), E(G)) be a graph where V (G) and of Dk (G).
E(G) are the set of vertices and the set of edges of G, Kulli and Janakiram 6 introduced a minimal
respectively. A set D ⊆ V (G) is called a dominating dominating graph of a graph G, denoted by M D(G),
set if every vertex in V (G)\D is adjacent to some which is the graph whose vertices are minimal dom-
vertex in D. The domination number of G, denoted inating sets, and two minimal dominating sets are
by γ(G), is the minimum cardinality of a dominating adjacent in M D(G) if they have at least one vertex
set of G. A dominating set of cardinality γ(G) in common. They characterized connected minimal
is called a γ-dominating set (or γ-set). For basic dominating graphs.
concepts and notation in domination, see Refs. 1, 2. In Ref. 7, Kulli and Janakiram introduced a
Let G be a graph and Dγ the set of all γ- common minimal dominating graph of a graph G,
dominating sets. Lakshmanan and Vijayakumar 3 denoted by C D(G). It has the same vertex set as
introduced a gamma graph γ.G of G. The vertex set G, and two vertices are adjacent in C D(G) if there
of γ.G is Dγ , and two γ-dominating sets D1 and D2 is a minimal dominating set in G which contains
from Dγ are adjacent in γ(G) if D1 = D2 \{u}∪{v} for them. The authors characterized connected com-
some u, v ∈ V (G). They provided the relationship mon minimal dominating graphs. They also gave
between the clique number and independence of characterization of a graph G for which C D(G) is
a graph and its gamma graph. Fricke et al 4 also isomorphic to the complement of G.
defined a gamma graph G(γ) with a different mean- A common minimal total dominating graph of a
ing. The only difference is that two γ-dominating graph G, denoted by C Dt (G), is the graph with the
sets D1 and D2 from Dγ are adjacent in G(γ) if same vertex set as G, and two vertices are adjacent
D1 = D2 \{u} ∪ {v} for some adjacent vertices u and in C Dt (G) if there is a minimal total dominating
v. Note that G(γ) is a subgraph of γ.G with the same set in G which contains them. This concept was
vertex set. introduced in Ref. 8.
In Ref. 5, Haas and Seyffarth defined a k- A set D of vertices in a graph G is called a
dominating graph of a graph G, denoted by Dk (G). total dominating set if every vertex of G is adjacent
Its vertex set contains all dominating sets D such to some vertex in D. Total dominating sets were
that |D| ¶ k, and two such dominating sets are introduced by Cockayne et al 9 . The total domination
adjacent in Dk (G) if one can be obtained from the number of G, denoted by γt (G), is the minimum

www.scienceasia.org
ScienceAsia 43 (2017) 327

Theorem 1 For n ¾ 3, γt (Pn ) = γt (Cn ) = b 12 nc +

1245 1256 d 14 ne − b 14 nc.

TOTAL DOMINATING GRAPH OF PATHS

In this section, we consider γ-total dominating
graphs of paths. We always let Pn = v1 v2 . . . vn be
a path with n vertices. If n = 1, we have that
2345 2356 T Dγ (P1 ) is the empty graph since P1 has no γ-total
dominating sets. For n ¾ 2, we obtain the following
Fig. 1 The γ-total dominating graph of a path with 6 theorems.
vertices. In this and later figures we write a bcd instead
Theorem 2 Let k ¾ 1 be an integer. Then
of {va , vb , vc , vd }.
T Dγ (P4k ) ∼
= K1 .

cardinality of a total dominating set of G. A total Proof : We first show that each γ-total dominating
dominating set of cardinality γt (G) is called a γ- set of P4k cannot contain three or more consecutive
total dominating set. Let T Dγ be the set of all γ- vertices of P4k . Suppose for a contradiction that
total dominating sets in G. The γ-total dominating there is a γ-total dominating set D containing three
graph of G, denoted by T Dγ (G), is the graph whose or more consecutive vertices of P4k . Let l be the
vertex set is T Dγ , and two γ-total dominating sets largest number of these consecutive vertices, so l ¾
D1 and D2 from T Dγ are adjacent in T Dγ (G) if 3. Let S be the set obtained from D by removing
D1 = D2 \{u} ∪ {v} for some u ∈ D2 and v ∈ / D2 . For these l vertices. Then S is a total dominating subset
instance, the γ-total dominating graph of the path of D. Note that |D| = 2k by Theorem 1. Since these
v1 v2 v3 v4 v5 v6 is shown in Fig. 1. l vertices dominate at most l + 2 vertices of P4k , the
other 2k − l vertices in D must dominate at least
PRELIMINARY RESULTS 4k − (l + 2) = 4k − l − 2 vertices of P4k . By Lemma 1,
Let D be a total dominating set of a graph G, S the 2k − l vertices in S can dominate at most 4k − 2l
a subset of D, V 0 the set of vertices in G which vertices of P4k , which is less than 4k−l−2 since l ¾ 3.
are dominated by the vertices in S, and G 0 the This is a contradiction. Thus every γ-total dominat-
subgraph of G induced by V 0 . Then S is called a total ing set must contain k groups of two consecutive
dominating subset of D if S is a total dominating set vertices. Hence there is only one γ-total dominating
of G 0 . set, which is {v2 , v3 , v6 , v7 , . . . , v4k−2 , v4k−1 }. 2
We first consider the relation between the num-
ber of vertices in S and the number of vertices in G Theorem 3 Let k ¾ 1 be an integer. Then
dominated by the vertices in S when G is a path or a T Dγ (P4k+1 ) ∼
= Pk .
cycle. We have that any 2 consecutive vertices in G
Proof : We prove by induction on k.
can dominate at most 4 vertices, and 3 consecutive
Base step. There is only one γ-total dom-
vertices in G can dominate at most 5 vertices, so we
inating sets of P5 , which is {v2 , v3 , v4 }. Hence
easily obtain the following lemma.
T Dγ (P5 ) ∼ = P1 . Furthermore, there are two γ-total
Lemma 1 Let G be a path or cycle with n vertices, D dominating sets of P9 , which are {v2 , v3 , v4 , v7 , v8 }
a total dominating set of G, and S a total dominating and {v2 , v3 , v6 , v7 , v8 }. Hence T Dγ (P9 ) ∼ = P2 .
subset of D of size k. If k is even, then S can dominate Induction step. Let k ¾ 2. Suppose that
at most 2k vertices of G; otherwise, S can dominate T Dγ (P4k+1 ) ∼ = Pk . Without loss of generality, we may
at most 2k − 1 vertices of G. assume that T Dγ (P4k+1 ) = D1 D2 . . . Dk , where D1 =
{v2 , v3 , v4 , v7 , v8 , . . . , v4k−5 , v4k−4 , v4k−1 , v4k } and for
Lemma 2 Let G be a graph. If v is a support vertex each l = 2, 3, . . . , k, Dl = D1 \{v4i | i = 1, 2, . . . , l −
(the vertex adjacent to a vertex of degree one) of G, 1} ∪ {v4i+2 | i = 1, 2, . . . , l − 1}. We next show
then v has to be in every total dominating set of G. that T Dγ (P4k+3 ) ∼ = Pk+1 . For each l = 1, 2, . . . , k,
let Dl0 = Dl ∪ {v4k+3 , v4k+4 } and Dk+1 0
= Dk \{v4k } ∪
The γ-total domination numbers of paths and {v4k+2 , v4k+3 , v4k+4 }. Hence Dl is a γ-total domi-
0

cycles were established by Henning 10 , as shown in nating set of P4k+5 for all l = 1, 2, . . . , k + 1. Fur-
the following theorem. thermore, D10 D20 . . . Dk+1 0
forms a path with k + 1

www.scienceasia.org
328 ScienceAsia 43 (2017)

vertices in T Dγ (P4k+5 ). We claim that there is no

other γ-total dominating set of P4k+5 apart from 𝑫𝟏,𝟏 𝑫𝟏,𝟐 . . . 𝑫𝟏,𝒌+𝟏
D10 , D20 , . . . , Dk+1
0
. Suppose for a contradiction that
there is another γ-total dominating set D0 of P4k+5 ,
which is different from these total dominating sets.
By Theorem 1, γt (P4k ) = 2k, γt (P4k+1 ) = 2k + 1 and 𝑫𝟐,𝟏 𝑫𝟐,𝒌+𝟏
γt (P4k+5 ) = 2k + 3, so |D0 | = 2k + 3. Furthermore, 𝑫𝟐,𝟐
|D0 ∩ {v4k+2 , v4k+3 , v4k+4 , v4k+5 }| ¾ 2 and v4k+4 ∈ D0
.
by Lemma 2. We consider the following 3 cases. .
Case 1: |D0 ∩ {v4k+2 , v4k+3 , v4k+4 , v4k+5 }| = 2. .
Subcase 1.1: v4k+3 , v4k+4 ∈ D0 , but v4k+2 , v4k+5 ∈ /
D . Hence D0 \{v4k+3 , v4k+4 } is a γ-total dominating
0

set of P4k+1 . Thus D0 \{v4k+3 , v4k+4 } = Dl for some 𝑫𝒌+𝟏,𝟏 𝑫𝒌+𝟏,𝟐 𝑫𝒌+𝟏,𝒌+𝟏
l = 1, 2, . . . k. Hence D0 = Dl ∪ {v4k+3 , v4k+4 } = Dl0 , a
contradiction. Fig. 2 The γ-total dominating graph of a path with 4k +2
Subcase 1.2: v4k+4 , v4k+5 ∈ D0 , but v4k+2 , v4k+3 ∈ / vertices.
D . Thus D0 \{v4k+4 , v4k+5 } is a total dominating sub-
0

set of D0 . Since v4k+4 and v4k+5 dominate 3 vertices,

the other 2k+1 vertices in D0 must dominate at least 𝑫′𝟏,𝟏 𝑫′𝟏,𝟐 . . . 𝑫′𝟏,𝒌+𝟏 𝑫′𝟏,𝒌+𝟐
4k + 2 vertices. By Lemma 1, these 2k + 1 vertices
in D0 can dominate at most 4k + 1 vertices. This is a
contradiction. 𝑫′𝟐,𝟏 𝑫′𝟐,𝒌+𝟐
Case 2: |D0 ∩ {v4k+2 , v4k+3 , v4k+4 , v4k+5 }| = 3. 𝑫′𝟐,𝟐 𝑫′𝟐,𝒌+𝟏
Subcase 2.1: v4k+2 , v4k+3 , v4k+4 ∈ D0 , but
v4k+5 ∈ / D. 0
Suppose for a contradiction that .
v4k+1 ∈ D0 . Then v4k ∈ / D0 (otherwise, D0 is not .
0 .
minimal). Thus D \{v4k+1 , v4k+2 , v4k+3 , v4k+4 } is a
total dominating subset of D0 . Since v4k+1 , v4k+2 ,
𝑫′𝒌+𝟏,𝟏 𝑫′
v4k+3 and v4k+4 dominate 6 vertices, the other 2k−1 𝑫′𝒌+𝟏,𝟐 𝑫′𝒌+𝟏,𝒌+𝟏 𝒌+𝟏,𝒌+𝟐
vertices in D0 must dominate at least 4k −1 vertices.
This contradicts Lemma 1. Hence v4k+1 ∈ / D0 . 𝑫′𝒌+𝟐,𝟏 𝑫′𝒌+𝟐,𝒌+𝟐
Since D is a γ-total dominating set of P4k+5 ,
0 𝑫′𝒌+𝟐,𝟐 𝑫′𝒌+𝟐,𝒌+𝟏
D0 \{v4k+2 , v4k+3 , v4k+4 } is a γ-total dominating set Fig. 3 The γ-total dominating graph of a path with 4k +6
of P4k . By Theorem 2, {v2 , v3 , v6 , v7 , . . . , v4k−2 , v4k−1 } vertices.
is the only γ-total dominating set of
P4k . Thus D0 \{v4k+2 , v4k+3 , v4k+4 } =
{v2 , v3 , v6 , v7 , . . . , v4k−2 , v4k−1 } = Dk \{v4k }. Hence there is only one γ-total dominating set of P2 , so
D0 = Dk \{v4k } ∪ {v4k+2 , v4k+3 , v4k+4 } = Dk+1 0
, a T Dγ (P2 ) ∼ = K1 ∼ = P1 2 P1 . For k = 1, the graph
contradiction. T Dγ (P6 ) is shown in Fig. 1.
Subcase 2.2: v4k+2 , v4k+4 , v4k+5 ∈ D0 , but v4k+3 ∈ / Let k ¾ 1. Suppose that T Dγ (P4k+2 ) ∼ =
0
D. Then v4k+1 ∈ D0 . We next have that Pk+1 2 Pk+1 . Without loss of generality, we may
v4k ∈ / D0 (otherwise, D0 is not minimal). Thus assume that T Dγ (P4k+2 ) is the graph shown in Fig. 2,
0
D \{v4k+1 , v4k+2 , v4k+4 , v4k+5 } is a total dominating whose vertices are Di, j = Oi ∪ E j for all integers
subset of D0 . Similarly, we then obtain a contradic- 1 ¶ i, j ¶ k + 1, where O1 = {v4i+1 | i = 0, 1, . . . , k},
tion to Lemma 1, so this case is impossible. E1 = {v2 } ∪ {v4i | i = 1, 2, . . . , k}, and for each l =
Subcase 2.3: v4k+3 , v4k+4 , v4k+5 ∈ D0 , but v4k+2 ∈ / 1, 2, . . . , k, Ol+1 = {v4i+3 | i = 0, 1, . . . , l −1}∪{v4i+1 |
D . This case is impossible since D0 is not minimal.
0 i = l, l +1, . . . , k} and El+1 = {v4i+2 | i = 0, 1, . . . , l}∪
Case 3: |D0 ∩ {v4k+2 , v4k+3 , v4k+4 , v4k+5 }| = 4. {v4i | i = l, l + 1, . . . , k}. It is easy to check that
This case is impossible since D0 is not minimal. 2 v4k−1 ∈ Oi if and only if i = k + 1, and v4k+2 ∈ E j
if and only if j = k + 1.
Theorem 4 Let k ¾ 0 be an integer. Then We next show that T Dγ (P4k+6 ) ∼ = Pk+2 2 Pk+2 .
T Dγ (P4k+2 ) ∼
= Pk+1 2 Pk+1 . For each i, j = 1, 2, . . . , k + 1, let Di,0 j = Di, j ∪
Proof : We prove by induction on k. For k = 0, {v4k+4 , v4k+5 }. For each i = 1, 2, . . . , k+1, let Di,k+2
0
=

www.scienceasia.org
ScienceAsia 43 (2017) 329

Di,k+1 ∪ {v4k+5 , v4k+6 }. For each j = 1, 2, . . . , k + 1, Case 3: |D0 ∩ {v4k+3 , v4k+4 , v4k+5 , v4k+6 }| = 4.
0
let Dk+2, j
= Dk+1, j \{v4k+1 }∪{v4k+3 , v4k+4 , v4k+5 }, and This case is impossible since D0 is not minimal. 2
Dk+2,k+2 = Dk+1,k+1 \{v4k+1 } ∪ {v4k+3 , v4k+5 , v4k+6 }.
0
Theorem 5 Let k ¾ 0 be an integer. Then
Then Di,0 j is a γ-total dominating set of P4k+6 for T Dγ (P4k+3 ) ∼
= Pk+2 .
all i, j = 1, 2, . . . , k + 2. Furthermore, these Di,0 j ’s
Proof : We prove by induction on k. It is easy to
form the graph Pk+2 2 Pk+2 in T Dγ (P4k+6 ) (Fig. 3).
obtain T Dγ (P3 ) ∼ = P2 and T Dγ (P7 ) ∼ = P3 .
Suppose for a contradiction that there is another γ-
Let k ¾ 1. Suppose that T Dγ (P4k+3 ) ∼ =
total dominating set D0 of P4k+6 , which is different
Pk+2 . Without loss of generality, we may as-
from these γ-total dominating sets. Note that |D0 | =
sume that T Dγ (P4k+3 ) = D1 D2 . . . Dk+2 , where D1 =
2k + 4, |D0 ∩ {v4k+3 , v4k+4 , v4k+5 , v4k+6 }| ¾ 2, and
{v1 , v2 , v5 , v6 , . . . , v4k−3 , v4k−2 , v4k+1 , v4k+2 } and Dl =
v4k+5 ∈ D0 . We consider the following 3 cases.
D1 \{v4i+1 | i = 0, 1, . . . , l −2}∪{v4i+3 | i = 0, 1, . . . , l −
Case 1: |D0 ∩ {v4k+3 , v4k+4 , v4k+5 , v4k+6 }| = 2.
2} for each l = 2, 3, . . . , k +2. It is easy to check that
Subcase 1.1: v4k+4 , v4k+5 ∈ D0 , but v4k+3 , v4k+6 ∈ /
v4k+3 ∈ Dl if and only if l = k + 2.
D . Hence D0 \{v4k+4 , v4k+5 } is a γ-total dominat-
0
We show that T Dγ (P4k+7 ) ∼ = Pk+3 . For each
ing set of P4k+2 . Thus D0 \{v4k+4 , v4k+5 } = Di, j for
l = 1, 2, . . . , k + 2, let Dl0 = Dl ∪ {v4k+5 , v4k+6 } and
some integers 1 ¶ i, j ¶ k + 1. Hence D0 = Di, j ∪ 0
Dk+3 = Dk+2 ∪ {v4k+6 , v4k+7 }. Hence Dl0 is a γ-
{v4k+4 , v4k+5 } = Di,0 j , a contradiction.
total dominating set of P4k+7 for all l = 1, 2, . . . , k +
Subcase 1.2: v4k+5 , v4k+6 ∈ D0 , but v4k+3 , v4k+4 ∈ / 3. Clearly, D10 D20 . . . Dk+3 0
forms a path with k + 3
D . Then v4k+2 ∈ D0 . Thus D0 \{v4k+5 , v4k+6 } is a
0
vertices in T Dγ (P4k+7 ). Suppose for a contradic-
γ-total dominating set of P4k+2 containing v4k+2 . tion that there is another γ-total dominating set
Since v4k+2 is in only Ek+1 , D0 \{v4k+5 , v4k+6 } = Di,k+1 D0 of P4k+7 , which is different from these γ-total
for some i ∈ {1, 2, . . . , k + 1}. Hence D0 = Di,k+1 ∪ dominating sets. Note that γt (P4k+3 ) = 2k + 2 and
{v4k+5 , v4k+6 } = Di,k+2
0
, a contradiction. γt (P4k+7 ) = 2k + 4, so |D0 | = 2k + 4. Furthermore,
Case 2: |D ∩ {v4k+3 , v4k+4 , v4k+5 , v4k+6 }| = 3.
0
|D0 ∩ {v4k+4 , v4k+5 , v4k+6 , v4k+7 }| ¾ 2 and v4k+6 ∈ D0 .
Subcase 2.1: v4k+3 , v4k+4 , v4k+5 ∈ D0 , but v4k+6 ∈ / We consider the following 3 cases.
0
D. Case 1: |D0 ∩ {v4k+4 , v4k+5 , v4k+6 , v4k+7 }| = 2.
Subcase 2.1.1: v4k+2 ∈ D0 . Clearly, v4k+1 ∈ / D0 . Subcase 1.1: v4k+5 , v4k+6 ∈ D0 , but v4k+4 , v4k+7 ∈ /
0
No matter whether v4k is in D or not, v4k−1 must D . Hence D0 \{v4k+5 , v4k+6 } is a γ-total dominating
0
be in D0 . Then D0 \{v4k+3 , v4k+4 , v4k+5 } ∪ {v4k+1 } is set of P4k+3 . Thus D0 \{v4k+5 , v4k+6 } = Dl for some
a γ-total dominating set of P4k+2 containing v4k−1 . l ∈ {1, 2, . . . , k +2}. Hence D0 = Dl ∪{v4k+5 , v4k+6 } =
Since v4k−1 is only in Ok+1 , D0 \{v4k+3 , v4k+4 , v4k+5 }∪ Dl0 , a contradiction.
{v4k+1 } = Dk+1, j for some j ∈ {1, 2, . . . , k + 1}. Thus Subcase 1.2: v4k+6 , v4k+7 ∈ D0 , but v4k+4 , v4k+5 ∈ /
D0 = Dk+1, j \{v4k+1 }∪{v4k+3 , v4k+4 , v4k+5 } = Dk+2, 0
j
,a D . Then v4k+3 ∈ D0 . Thus D0 \{v4k+6 , v4k+7 } is a γ-
0
contradiction. total dominating set of P4k+3 , which contains v4k+3 .
Subcase 2.1.2: v4k+2 ∈ / D0 . If v4k+1 ∈ D0 , D0 is Hence D0 \{v4k+6 , v4k+7 } = Dk+2 since v4k+3 is only
not minimal. Thus v4k+1 ∈ / D0 , so v4k−1 , v4k ∈ D0 . in Dk+2 . Hence D0 = Dk+2 ∪ {v4k+6 , v4k+7 } = Dk+3 0
,a
Hence D \{v4k+3 , v4k+4 , v4k+5 } ∪ {v4k+1 } is a γ-total
0
contradiction.
dominating set of P4k+2 containing v4k−1 . Thus Case 2: |D0 ∩ {v4k+4 , v4k+5 , v4k+6 , v4k+7 }| = 3.
D0 \{v4k+3 , v4k+4 , v4k+5 } ∪ {v4k+1 } = Dk+1, j for some Subcase 2.1: v4k+4 , v4k+5 , v4k+6 ∈ D0 , but v4k+7 ∈ /
j ∈ {1, 2, . . . , k + 1}. Hence D0 = Dk+1, j \{v4k+1 } ∪ D.0
{v4k+3 , v4k+4 , v4k+5 } = Dk+2,0
j
, a contradiction. Subcase 2.1.1: v4k+3 ∈ D0 . Thus v4k+2 ∈ / D0 .
Subcase 2.2: v4k+3 , v4k+5 , v4k+6 ∈ D0 , but v4k+4 ∈ / 0
Hence D \{v4k+3 , v4k+4 , v4k+5 , v4k+6 } is a total dom-
D0 . Then v4k+2 ∈ D0 . If v4k+1 ∈ D0 , D0 is not inating subset of D0 . Since v4k+3 , v4k+4 , v4k+5 , and
minimal. Thus v4k+1 ∈ / D0 . No matter whether v4k+6 dominate 6 vertices, the other 2k vertices in
v4k is in D or not, v4k−1 must be in D0 . Thus
0
D0 must dominate at least 4k + 1 vertices. This
D0 \{v4k+3 , v4k+5 , v4k+6 } ∪ {v4k+1 } is a γ-total dom- contradicts Lemma 1.
inating set of P4k+2 , containing v4k−1 and v4k+2 . Subcase 2.1.2: v4k+3 ∈ / D0 . Hence
Thus D0 \{v4k+3 , v4k+5 , v4k+6 } ∪ {v4k+1 } = Dk+1,k+1 . 0
D \{v4k+4 , v4k+5 , v4k+6 } is a total dominating
Thus D0 = Dk+1,k+1 \{v4k+1 } ∪ {v4k+3 , v4k+5 , v4k+6 } = subset of D0 . As with Subcase 2.1.1, there is a
0
Dk+2,k+2 , a contradiction. contradiction.
Subcase 2.3: v4k+4 , v4k+5 , v4k+6 ∈ D0 , but v4k+3 ∈ / Subcase 2.2: v4k+4 , v4k+6 , v4k+7 ∈ D0 , but v4k+5 ∈ /
D . This case is impossible since D0 is not minimal.
0
D.0
Then v4k+3 ∈ D0 . If v4k+2 ∈ D0 , then

www.scienceasia.org
330 ScienceAsia 43 (2017)

D0 is not minimal. Hence v4k+2 ∈ / D0 . Hence l vertices forms a total dominating subset of D. We
0
D \{v4k+3 , v4k+4 , v4k+6 , v4k+7 } is a total dominating then obtain a contradiction to Lemma 1.
subset of D0 . Similarly, we then obtain a contradic- Claim 2: Each γ-total dominating set of C4k+1
tion to Lemma 1. contains only one group of three consecutive ver-
/
Subcase 2.3: v4k+5 , v4k+6 , v4k+7 ∈ D0 , but v4k+4 ∈ tices of C4k+1 . Since γt (C4k+1 ) = 2k + 1 is an odd
D . This case is impossible since D0 is not minimal.
0
integer, each γ-total dominating set of C4k+1 con-
Case 3: |D0 ∩ {v4k+4 , v4k+5 , v4k+6 , v4k+7 }| = 4. tains at least one group of three consecutive vertices.
This case is impossible since D0 is not minimal. 2 Suppose for a contradiction that there is a γ-total
dominating set D of C4k+1 , which contains l groups
TOTAL DOMINATING GRAPH OF CYCLES of three consecutive vertices of C4k+1 , where l ¾ 2.
In this section, we always let Cn = v0 v1 . . . vn−1 v0 These 3l vertices dominate at most 5l vertices. Thus
be a cycle with n ¾ 3 vertices. It easy to see that the other 2k + 1 − 3l vertices in D must dominate
T Dγ (C3 ) ∼
= C3 and T Dγ (C4 ) ∼
= C4 . For n ¾ 5, we at least 4k + 1 − 5l vertices of C4k+1 . By Lemma 1,
obtain the following theorems. these 2k + 1 − 3l vertices in D can dominate at most
4k + 2 − 6l vertices of C4k+1 , which is less than
Theorem 6 Let k ¾ 2 be an integer. Then 4k + 1 − 5l since l ¾ 3. This is a contradiction.
T Dγ (C4k ) ∼
= 4K1 . Let D be any γ-total dominating set, so D con-
tains one group of 3 consecutive vertices, which
Proof : We claim that each γ-total dominating set dominates 5 vertices of C4k+1 . We may consider
of C4k cannot contain three or more consecutive the other 4k − 4 vertices in C4k+1 which are not
vertices of C4k . Suppose for a contradiction that dominated as a path. Apart from the 3 consecutive
there is a γ-total dominating set D of C4k , which vertices in D, the other 2k−2 vertices must dominate
contains three or more consecutive vertices of C4k . all 4k − 4 vertices on this path. By Theorem 2,
Let l be the largest number of these consecutive there is only one γ-total dominating set of this
vertices, so l ¾ 3. Let S be the set obtained from path. Hence there is only one γ-total dominating
D by removing these l vertices. Then S is a total set of C4k−4 containing these 3 consecutive ver-
dominating subset of D. By Theorem 1, |D| = 2k. tices. To find all γ-total dominating sets of C4k+1 ,
Since these l vertices dominate l + 2 vertices of it suffices to find 3 consecutive vertices on the
C4k , the other 2k − l vertices in D must dominate cycle. Clearly, there are 4k + 1 γ-total dominating
at least 4k − (l + 2) = 4k − l − 2 vertices of C4k . By sets. Recall that C4k+1 = v0 v1 . . . v4k v0 . Let D0 =
Lemma 1, the 2k − l vertices in S can dominate {v0 , v1 , v2 , v5 , v6 , v9 , v10 , . . . , v4k−3 , v4k−2 } and Dl =
at most 4k − 2l vertices of C4k , which is less than Dl−1 \{v(4l−2) (mod 4k+1) } ∪ {v(4l) (mod 4k+1) } for each
4k − l − 2 since l ¾ 3. This is a contradiction. l = 1, 2, . . . , 4k. Then D0 D1 . . . D4k D0 forms a cycle
Thus every γ-total dominating set must contain k with 4k + 1 vertices. 2
groups of two consecutive vertices of C4k . It is easy
to see that there are only four γ-total dominating Theorem 8 Let k ¾ 1 be an integer. Then
sets, which are {v0 , v1 , v4 , v5 , . . . , v4k−4 , v4k−3 }, T Dγ (C4k+2 ) ∼
= C2k+1 2C2k+1 .
{v1 , v2 , v5 , v6 , . . . , v4k−3 , v4k−2 },
{v2 , v3 , v6 , v7 , . . . , v4k−2 , v4k−1 }, and Proof :
{v0 , v3 , v4 , . . . , v4k−5 , v4k−4 , v4k−1 }. 2 We prove by induction on k. For k = 1 and k = 2,
the graph T Dγ (C6 ) and T Dγ (C10 ) are shown in Fig. 4
Theorem 7 Let k ¾ 1 be an integer. Then and Fig. 5, respectively.
T Dγ (C4k+1 ) ∼
= C4k+1 . Let k ¾ 2. Suppose that T Dγ (C4k+2 ) ∼ =
C2k+1 2C2k+1 . Without loss of generality, we may
Proof : For k = 1, it is easy to obtain T Dγ (C5 ) ∼
= C5 . assume that T Dγ (C4k+2 ) is the graph shown in
Let k ¾ 2. Fig. 6, whose vertices are Di, j = Oi ∪ E j for all
Claim 1: each γ-total dominating set of C4k+1 integers 1 ¶ i, j ¶ 2k + 1, where O1 = {v1 } ∪
cannot contain four or more consecutive vertices of {v4i−1 | i = 1, 2, . . . , k}, E1 = {v0 } ∪ {v4i+2 | i =
C4k+1 . Suppose for a contradiction that there is a 0, 1, . . . , k − 1}, and for each l = 2, 3, . . . , 2k + 1,
γ-total dominating set D of C4k+1 , which contains Ol = Ol−1 \{v(4l−5) (mod 4k+2) }∪{v(4l−3) (mod 4k+2) } and
four or more consecutive vertices of C4k+1 . Let l be El = El−1 \{v(4l−6) (mod 4k+2) } ∪ {v(4l−4) (mod 4k+2) }.
the largest number of these consecutive vertices, so Recall that C4k+6 = v0 v1 . . . v4k+5 v0 . We prove
l ¾ 4. The set obtained from D by removing these that T Dγ (C4k+6 ) ∼ = C2k+3 2C2k+3 . For each i, j =

www.scienceasia.org
ScienceAsia 43 (2017) 331

0123 0134 1234

𝑫𝟏,𝟏 𝑫𝟏,𝟐 . . . 𝑫𝟏,𝟐𝒌+𝟏

𝑫𝟐,𝟏 𝑫𝟐,𝟐 𝑫𝟐,𝟐𝒌+𝟏

0125 0145 1245

.
.
.

0235 0345 2345 𝑫𝟐𝒌+𝟏,𝟏 𝑫𝟐𝒌+𝟏,𝟐 𝑫𝟐𝒌+𝟏,𝟐𝒌+𝟏

Fig. 4 The γ-total dominating graph of a cycle with 6 Fig. 6 The γ-total dominating graph of a cycle with 4k+2
vertices. vertices.

set of C4k+6 for all i, j = 1, 2, . . . , 2k + 3. Further-

012367 013467 013478 123478 123678 more, these Di,0 j ’s form the graph C2k+3 2C2k+3 in
T Dγ (C4k+6 ).
012567 014567 014578 124578 125678 Claim 1. For each i = 1, 2, . . . , 2k+3, Di,1 0
\{v0 }∪
{v4k+4 } = Di,2k+3 . Let i ∈ {1, 2, . . . , 2k + 3}. Then
0

we have Di,1 0 0
and Di,2k+3 are adjacent in T Dγ (C4k+6 ).
012569 014569 014589 124589 125689
Furthermore, v0 ∈ Di,1 , v0 ∈0
/ Di,2k+3
0 0
, v4k+4 ∈ Di,2k+3 ,
and v4k+4 ∈ / Di,1 . Thus Di,1 \{v0 } ∪ {v4k+4 } = Di,2k+3 .
0 0 0

023569 034569 034589 234589 235689 Claim 2. For each j = 1, 2, . . . , 2k+3, D1, 0
j \{v1 }∪
{v4k+5 } = D2k+3, j . Let j ∈ {1, 2, . . . , 2k + 3}. Then
0

j and D2k+3, j are adjacent in T Dγ (C4k+6 ).

0 0
we have D1,
023679 034679 034789 234789 236789 Furthermore, v1 ∈ D1, 0
/ D2k+3,
j , v1 ∈
0
j
0
, v4k+5 ∈ D2k+3, j
,
and v4k+5 ∈ / D1, j . Thus D1, j \{v1 } ∪ {v4k+5 } = D2k+3, j .
0 0 0

Fig. 5 The γ-total dominating graph of a cycle with 10 Next, we prove that there are no other ver-
vertices. tices in T Dγ (C4k+6 ). Suppose for a contradiction
that there is another γ-total dominating set D0 of
C4k+6 , which is different from these γ-total dom-
1, 2, . . . , k + 1, let Di,0 j = Di, j ∪ {v4k+2 , v4k+3 }. For inating sets. Note that |D0 | = 2k + 4 and |D0 ∩
each i = 1, 2, . . . , k + 1, let Di,k+2
0
= Di,k+1
0
\{v4k+2 } ∪ {v4k+2 , v4k+3 , v4k+4 , v4k+5 }| ¾ 2. We consider the
{v4k+4 }. For each i = 1, 2, . . . , k +1 and j = k +3, k + following 3 cases.
4, . . . , 2k + 2, let Di,0 j = Di, j−1 ∪ {v4k+3 , v4k+4 }. For Case 1: |D0 ∩ {v4k+2 , v4k+3 , v4k+4 , v4k+5 }| = 4.
each i = 1, 2, . . . , k + 1, let Di,2k+3
0
= Di,2k+2
0
\{v4k } ∪ If v0 ∈ D0 , D0 is not minimal. Hence v0 ∈ / D0 .
{v4k+2 }. For each j = 1, 2, . . . , 2k + 3, let Dk+2, 0
j
= Similarly, v4k+1 ∈ / D . Since v4k+2 , v4k+3 , v4k+4 and
0

Dk+1, j \{v4k+3 } ∪ {v4k+5 }. For each i = k + 3, k +

0 v4k+5 dominate 6 vertices, the other 2k vertices in
4, . . . , 2k +2 and j = 1, 2, . . . , k +1, let Di,0 j = Di−1, j ∪ D0 must dominate the vertices v1 , v2 , v3 , . . . , v4k .
We now consider these 4k vertices as a path.
{v4k+2 , v4k+5 }. For each i = k+3, k+4, . . . , 2k+2, let
By Theorem 2, D0 \{v4k+2 , v4k+3 , v4k+4 , v4k+5 }
0
Di,k+2 = Di,k+1
0
\{v4k+2 } ∪ {v4k+4 }. For each i, j = k +
is the only γ-total dominating set of this
3, k +4, . . . , 2k +2, let Di,0 j = Di−1, j−1 ∪{v4k+4 , v4k+5 }.
path which is {v2 , v3 , v6 , v7 , . . . , v4k−2 , v4k−1 }.
For each i = k + 3, k + 4, . . . , 2k + 2, let Di,2k+3 0
= Thus D0 \{v4k+2 , v4k+3 , v4k+4 , v4k+5 } ∪
Di,2k+2 \{v4k } ∪ {v4k+2 }. For each j = 1, 2, . . . , 2k + 3,
0
{v4k , v4k+1 } = {v2 , v3 , v6 , v7 , . . . , v4k−2 , v4k−1 } ∪
0
let D2k+3, j
= D2k+2,
0
j
\{v4k+1 } ∪ {v4k+3 }. Note that {v4k , v4k+1 } = D2k+1,2k+1 (Fig. 6). Hence D0 =
γt (C4k+6 ) = 2k + 4 = (2k + 2) + 2 = γt (C4k+2 ) + 2. D2k+1,2k+1 \{v4k , v4k+1 }∪{v4k+2 , v4k+3 , v4k+4 , v4k+5 } =
It is easy to check that Di,0 j is a γ-total dominating [(D2k+1,2k+1 ∪ {v4k+4 , v4k+5 })\{v4k } ∪

www.scienceasia.org
332 ScienceAsia 43 (2017)

{v4k+2 }]\{v4k+1 } ∪ {v4k+3 } = [D2k+2,2k+2 0

\{v4k } ∪ 0
Di+1,1 \{v0 } ∪ {v4k+4 } = Di+1,2k+3
0
.
{v4k+2 }]\{v4k+1 } ∪ {v4k+3 } = D2k+2,2k+3 \{v4k+1 } ∪
0
Subcase 2.4: v4k+3 , v4k+4 , v4k+5 ∈ D0 , but v4k+2 ∈ /
0
{v4k+3 } = D2k+3,2k+3
0
, a contradiction. D.
Case 2: |D0 ∩ {v4k+2 , v4k+3 , v4k+4 , v4k+5 }| = 3. Subcase 2.4.1: v0 ∈ D0 . Then D0 contains four
Subcase 2.1: v4k+2 , v4k+3 , v4k+4 ∈ D0 , but v4k+5 ∈ / consecutive vertices of the cycle. Again, we repeat
D.0 the process in case 1, so we are done.
Subcase 2.1.1: v4k+1 ∈ D0 . Since D0 contains Subcase 2.4.2: v0 ∈ / D0 . It is easy to obtain
four consecutive vertices of the cycle, by repeating v2 , v3 , v4k ∈ D0 , but v1 , v4k+1 , v4k+2 ∈ / D0 . Hence
the process in Case 1, D0 must be equal to Di,0 j for D0 \{v4k+3 , v4k+4 , v4k+5 }∪{v1 } is a γ-total dominating
some i, j, a contradiction. set of C4k+2 . Thus D0 \{v4k+3 , v4k+4 , v4k+5 } ∪ {v1 } =
Subcase 2.1.2: v4k+1 ∈ / D0 . Thus v0 ∈ / D0 Di, j = Oi ∪ E j for some 1 ¶ i, j ¶ 2k + 1. Since
(otherwise, D is not minimal). Then v1 , v2 ∈ D0 .
0 v1 , v3 ∈ Oi and v2 , v4k ∈ E j , i = 1 and j ∈ {k +
Similarly, v4k ∈ / D0 , so v4k−1 , v4k−2 ∈ D0 . 2, k +3, . . . , 2k +1}. Hence D0 \{v4k+3 , v4k+4 , v4k+5 }∪
0
Hence D \{v4k+2 , v4k+3 , v4k+4 } ∪ {v4k } is a {v1 } = D1, j for some j ∈ {k + 2, k + 3, . . . , 2k +
γ-total dominating set of C4k+2 . Thus 1}. Hence D0 = D1, j \{v1 } ∪ {v4k+3 , v4k+4 , v4k+5 } =
D0 \{v4k+2 , v4k+3 , v4k+4 } ∪ {v4k } = Di, j = Oi ∪ E j (D1, j ∪ {v4k+3 , v4k+4 })\{v1 } ∪ {v4k+5 } = D1, 0
j+1 \{v1 } ∪
for some 1 ¶ i, j ¶ 2k + 1. Since v1 ∈ Oi , {v4k+5 } = D2k+3, j+1 by Claim 2.
0

1 ¶ i ¶ k + 1. Since v4k−2 , v4k ∈ E j , j = 2k + 1. Case 3: |D0 ∩ {v4k+2 , v4k+3 , v4k+4 , v4k+5 }| = 2.

Hence D0 \{v4k+2 , v4k+3 , v4k+4 } ∪ {v4k } = Di,2k+1 Subcase 3.1: v4k+3 , v4k+4 ∈ D0 , but v4k+2 , v4k+5 ∈ /
for some i ∈ {1, 2, . . . , k + 1}. Hence D.0
It is easy to obtain v1 , v4k ∈ D0 . Since
D0 = Di,2k+1 \{v4k } ∪ {v4k+2 , v4k+3 , v4k+4 } = D0 \{v4k+3 , v4k+4 } is a γ-total dominating set of C4k+2 ,
(Di,2k+1 ∪ {v4k+3 , v4k+4 })\{v4k } ∪ {v4k+2 } = D0 \{v4k+3 , v4k+4 } = Di, j = Oi ∪ E j for some integers
0
Di,2k+2 \{v4k } ∪ {v4k+2 } = Di,2k+30
, a contradiction. 1 ¶ i, j ¶ 2k + 1. Since v1 ∈ D0 \{v4k+3 , v4k+4 }, v1 ∈
Subcase 2.2: v4k+2 , v4k+3 , v4k+5 ∈ D0 , but v4k+4 ∈ / Oi . Thus i ∈ {1, 2, . . . , k + 1}.
D0 . Hence v0 ∈ D0 . If v4k+1 ∈ D0 , D0 is not Subcase 3.1.1: v0 ∈ D0 . Since v0 , v4k ∈ E j , j =
minimal. Hence v4k+1 ∈ / D0 . No matter whether k + 1. Hence D0 = Di,k+1 ∪ {v4k+3 , v4k+4 } = (Di,k+1 ∪
v4k is in D or not, v4k−1 must be in D0 . Thus
0 {v4k+2 , v4k+3 })\{v4k+2 } ∪ {v4k+4 } = Di,k+1 0
\{v4k+2 } ∪
D0 \{v4k+2 , v4k+3 , v4k+5 } ∪ {v4k+1 } is a γ-total domi- {v4k+4 } = Di,k+2 .
0

nating set of C4k+2 . Thus D0 \{v4k+2 , v4k+3 , v4k+5 } ∪ Subcase 3.1.2: v0 ∈ / D0 . Hence v2 ∈ D0 . Since
{v4k+1 } = Di, j = Oi ∪ E j for some 1 ¶ i, j ¶ 2k + 1. v2 , v4k ∈ E j , j ∈ {k +2, k +3, . . . , 2k +1}. Hence D0 =
Since v4k−1 , v4k+1 ∈ Oi , i = 2k +1. Since v0 ∈ E j , 1 ¶ Di, j ∪ {v4k+3 , v4k+4 } = Di,0 j+1 .
j ¶ k + 1. Thus D0 \{v4k+2 , v4k+3 , v4k+5 } ∪ {v4k+1 } = Subcase 3.2: v4k+2 , v4k+3 ∈ D0 , but v4k+4 , v4k+5 ∈ /
D2k+1, j for some j ∈ {1, 2, . . . , k + 1}. Hence D0 = D . If v4k+1 ∈ D0 , we repeat the process in Sub-
0
D2k+1, j \{v4k+1 } ∪ {v4k+2 , v4k+3 , v4k+5 } = (D2k+1, j ∪ case 2.1; otherwise, we repeat the process in Sub-
{v4k+2 , v4k+5 })\{v4k+1 } ∪ {v4k+3 } = D2k+2, 0
j
\{v4k+1 } ∪ case 3.1.
{v4k+3 } = D2k+3, j , a contradiction.
0
Subcase 3.3: v4k+4 , v4k+5 ∈ D0 , but v4k+2 , v4k+3 ∈ /
Subcase 2.3: v4k+2 , v4k+4 , v4k+5 ∈ D0 , but D . If v0 ∈ D0 , we repeat the process in Subcase 2.4;
0

v4k+3 ∈ / D . It is easy to obtain v0 , v4k ∈

0
/ D0 but otherwise, we repeat the process in Subcase 3.1.
0 0
v2 , v4k−2 , v4k+1 ∈ D . Thus D \{v4k+2 , v4k+4 , v4k+5 } ∪ Subcase 3.4: v4k+2 , v4k+5 ∈ D0 , but v4k+3 , v4k+4 ∈ /
{v0 } is a γ-total dominating set of C4k+2 . Thus D . Then v0 ∈ D0 . If v1 ∈ D0 , we repeat the process
0

D0 \{v4k+2 , v4k+4 , v4k+5 } ∪ {v0 }} = Di, j = Oi ∪ E j for in Subcase 2.4; otherwise, we repeat the process in
some 1 ¶ i, j ¶ 2k + 1. Since v4k+1 ∈ Oi , k + Subcase 3.1. 2
1 ¶ i ¶ 2k + 1. Since v0 , v2 ∈ E j , j = 1. Thus
D0 \{v4k+2 , v4k+4 , v4k+5 } ∪ {v0 } = Di,1 for some i ∈ Theorem 9 Let k ¾ 1 be an integer. Then
{k + 1, k + 2, . . . , 2k + 1}. T Dγ (C4k+3 ) ∼
= C4k+3 .
Subcase 2.3.1: i = k + 1. Then Proof : First, we show that each γ-total dominating
D0 = Dk+1,1 ∪ {v4k+2 , v4k+5 }\{v0 } ∪ {v4k+4 } = set of C4k+3 cannot contain three or more consecu-
[(Dk+1,1 ∪ {v4k+2 , v4k+3 })\{v4k+3 } ∪ {v4k+5 }]\{v0 } ∪ tive vertices of C4k+3 . Suppose for a contradiction
{v4k+4 } = [Dk+1,1 0
\{v4k+3 }∪{v4k+5 }]\{v0 }∪{v4k+4 } = that there is a γ-total dominating set D of C4k+3 ,
Dk+2,1 \{v0 } ∪ {v4k+4 } = Dk+2,2k+3
0 0
by Claim 1. which contains three or more consecutive vertices
Subcase 2.3.2: i ∈ {k + 2, k + 3, . . . , 2k + 1}. of C4k+3 . Let l be the largest number of these
Then D0 = Di,1 ∪ {v4k+2 , v4k+5 }\{v0 } ∪ {v4k+4 } = consecutive vertices, so l ¾ 3. By Theorem 1, |D| =

www.scienceasia.org
ScienceAsia 43 (2017) 333

2k + 2. Since these l vertices dominate l + 2 vertices

of C4k+3 , the other 2k + 2 − l vertices in D must
dominate at least 4k+3−(l+2) = 4k−l+1 vertices of
C4k+3 . By Lemma 1, if l = 3, these 2k +2− l vertices
can dominate at most 2(2k + 2 − l) − 1 = 4k + 3 − 2l
vertices of C4k+3 , which is less than 4k − l + 1.
Suppose l ¾ 4. Then these 2k + 2 − l vertices can
dominate at most 4k+4−2l vertices of C4k+3 , which
is less than 4k − l +1. This is a contradiction. Hence
every γ-total dominating set must contain k + 1
groups of two consecutive vertices of C4k+3 . This
means there is only one vertex in C4k+3 , which is
dominated by 2 vertices in such a γ-total dominating
set. To find all γ-total dominating sets, it suffices to
find such a vertex on the cycle dominated by two
vertices. Clearly, there are exactly 4k + 3 γ-total
dominating sets. Recall that C4k+3 = v0 v1 . . . v4k+2 v0 .
Let D0 = {v0 , v1 , v3 , v4 , v7 , v8 , v11 , v12 , . . . , v4k−1 , v4k }
and Dl = Dl−1 \{v(4l−1) (mod 4k+3) }∪{v(4l+1) (mod 4k+3) }
for each l = 1, 2, . . . , 4k + 2. Then D0 D1 . . . D4k+2 D0
forms a cycle with 4k + 3 vertices. 2
Acknowledgements: This study was supported by the
2015 Research Grant of the Faculty of Science and Tech-
nology, Thammasat University. The authors would like to
thank Nawarat Ananchuen and Rakkor Siriwat for much
appreciated assistance with our work.

REFERENCES
1. Haynes TW, Hedetniemi ST, Slater PJ (1998) Dom-
ination in Graphs: Advanced Topics, Marcel Dekker,
New York.
2. Haynes TW, Hedetniemi ST, Slater PJ (1998) Fun-
damentals of Domination in Graphs, Marcel Dekker,
New York.
3. Lakshmanan SA, Vijayakumar A (2010) The gamma
graph of a graph. AKCE Int J Graphs Combinator 7,
53–9.
4. Fricke GH, Hedetniemi SM, Hedetniemi ST, Huston
KR (2011) γ-graphs of graphs. Discuss Math Graph
Theor 31, 517–31.
5. Haas R, Seyffarth K (2014) The k-dominating graph.
Graphs Combinator 30, 609–17.
6. Kulli VR, Janakiram B (1995) The minimal dominat-
ing graph. Graph Theor Notes New York 28, 12–5.
7. Kulli VR, Janakiram B (1996) The common mini-
mal dominating graph. Indian J Pure Appl Math 27,
193–6.
8. Kulli VR (2014) The common minimal total dominat-
ing graph. J Discrete Math Sci Cryptogr 17, 49–54.
9. Cockayne E, Dawes RM, Hedetniemi ST (1980) Total
domination in graphs. Networks 10, 211–9.
10. Henning MA (2000) Graphs with large total domina-
tion number. J Graph Theor 35, 21–45.

www.scienceasia.org