European Journal of Combinatorics 37 (2014) 4–23

Contents lists available at ScienceDirect

European Journal of Combinatorics

journal homepage: www.elsevier.com/locate/ejc

Beautiful conjectures in graph theory

Adrian Bondy
Université Claude Bernard Lyon 1, France
Université Paris 6, France

article info

Article history:
Available online 23 August 2013

1. What is a beautiful conjecture?

The mathematician’s patterns, like the painter’s or the poet’s must be beautiful; the ideas, like the
colours or the words must fit together in a harmonious way. Beauty is the first test: there is no
permanent place in this world for ugly mathematics.
G.H. Hardy
For the vast majority of mathematicians, one of the principal attractions of our discipline is the
inherent beauty of its structures and the relationships between them. As such, mathematics may
be compared to music or, more concretely, to architecture. Nonetheless, mathematics differs from
these disciplines in that there is a sense of inevitability, of absoluteness, of truth, in the greatest
mathematics. Indeed, many practitioners have the strong sense that mathematics is as much an
activity of discovery as of invention. Paul Erdős, albeit a confirmed atheist, incarned this sentiment
in speaking of The Book, a book kept by the S.F.,1 in which the most beautiful (inevitable) proofs of all
theorems are recorded (see the elegant monograph by Aigner and Ziegler, 2004). In sum, mathematics
is unquestionably as much a quest for beauty as an activity of scientific investigation. It is in this sense
a pursuit of perfection, an attempt to describe an ideal world, or rather, an ideal universe.
Certainly, not all mathematics exemplifies beauty. But, in the belief that beauty lies at the core of
the very best mathematics, we have tried here to collect together those conjectures in graph theory
which satisfy, to a greater or lesser extent, various criteria of beauty. The resulting selection is of
course entirely subjective, as are the criteria used. Also, it goes without saying that our list is far from
complete, due to inadvertent omission and ignorance. The aim here is simply to open up a discussion
by focusing attention on this essential aspect of our activity.

E-mail address: adrian.bondy@upmc.fr.

1 Supreme Fascist — God, in the idiosyncratic language of Erdős, borne of a lifetime of observation and experience.

0195-6698/$ – see front matter © 2013 Published by Elsevier Ltd

http://dx.doi.org/10.1016/j.ejc.2013.07.006
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 5

Reference
M. Aigner and G. Ziegler, Proofs from the Book. Including illustrations by Karl H. Hofmann. Third edition.
Springer-Verlag, Berlin, 2004.

2. Some criteria of beauty in mathematics

Mathematics, rightly viewed, possesses not only truth, but supreme beauty — a beauty cold and
austere, like that of sculpture.
Bertrand Russell
The criteria upon which our selection of conjectures is based should be considered as an ensemble,
each exposing but one facet of the overall picture. They are:

• Simplicity: short, easily understandable statement relating basic concepts.

• Surprise: links together seemingly disparate concepts.
• Generality: valid for a wide variety of objects.
• Centrality: close ties with a number of existing theorems and/or conjectures.
• Longevity: at least twenty years old.
• Fecundity: attempts to prove the conjecture have led to new concepts or new proof techniques.

Naturally, one may apply essentially the same criteria of beauty to theorems, replacing the criteria of
fecundity and longevity by:

• Fecundity: has inspired interesting extensions and/or generalizations.

• Veracity: a beautiful theorem should be true.

One may also consider similar criteria of beauty in proofs:

• Elegance: combination of simplicity and surprise.

• Ingenuity: imaginative use of standard techniques.
• Originality: introduction of new proof techniques.
• Fecundity: has inspired new proof techniques or new proofs of existing theorems.
• Exactitude: a beautiful proof should (preferably) be correct.

. . . an elegant proof is a proof which would not normally come to mind, like an elegant chess problem:
the first move should be paradoxical . . .
Claude Berge

Fig. 1. Claude Berge.

6 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

3. A selection of beautiful conjectures

We list here our choice of beautiful conjectures in graph theory, grouped together by theme
but presented in no particular order. Each theme is accompanied by a brief commentary, and the
conjectures are assigned ratings (from zero to three stars) according to each of the six criteria
enumerated above. Apart from the criterion of longevity (one star for every twenty years of existence),
these ratings are, of course, entirely subjective — and subject to revision. Also included are one
or two of the principal results on each conjecture, together with a small number of references,
listed in chronological order. These are provided simply to direct the reader towards the relevant
literature. Although the notions involved in the stated conjectures are surely well-known to most
graph-theorists, some definitions have been included for the sake of completeness. All others can
be found in any basic text on graph theory, for instance J.A. Bondy and U.S.R. Murty, Graph Theory,
Graduate Texts in Mathematics 244, Springer, 2008.

4. Reconstruction of graphs

A vertex-deleted subgraph of a graph is a subgraph obtained by the deletion of a single vertex. A

graph H is a reconstruction of a graph G if V (H ) = V (G) and H − v ∼
= G − v for all v ∈ V (G). A graph
is reconstructible from its vertex-deleted subgraphs if every reconstruction of it is isomorphic to it.
Reconstructibility from edge-deleted subgraphs is defined in an analogous fashion.
The Reconstruction Conjecture arose from a study of metric spaces by S.M. Ulam and his doctoral
student P.J. Kelly. It has proved remarkably intransigent, the wealth of information supplied by the
vertex-deleted subgraphs being hard to marshall. Counting techniques have been the most fruitful.
Graph symmetries also enter into the picture. The edge analogue seems at first sight almost obvious,
but has proved equally difficult to settle in its entirety. On the other hand, as D.L. Greenwell showed, it
is (formally) no harder than the Reconstruction Conjecture and, indeed, more progress has been made
on it. Corollary: seek either a counterexample to the vertex version or a proof of the edge version.
Warning: the vertex version for digraphs, and both the vertex and edge versions for infinite graphs,
are false. R. Halin formulated a natural analogue of the Reconstruction Conjecture for infinite graphs.
Another attractive (and perhaps more tractable) problem is the Switching Reconstruction Conjecture
of R.P. Stanley.

The Reconstruction Conjecture

Every simple graph on at least three vertices is reconstructible from its vertex-deleted subgraphs.
(P.J. Kelly and S.M. Ulam 1942)

Simple Surprising General Central Old Fertile

∗∗ ∗∗∗ ∗∗∗ ∗

• False for digraphs: there exist infinite families of nonreconstructible tournaments.

(P.J. Stockmeyer 1977)

The Edge Reconstruction Conjecture

Every simple graph on at least four edges is reconstructible from its edge-deleted subgraphs.
(F. Harary 1964)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗∗ ∗∗ ∗

• True for graphs on n vertices and more than n log2 n edges. (V. Müller 1977)
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 7

Fig. 2. Frank Harary.

Halin’s Reconstruction Conjecture

Every reconstruction of an infinite graph is isomorphic to an induced subgraph of the graph.
(R. Halin 1970)

Simple Surprising General Central Old Fertile

∗∗ ∗∗∗ ∗

To switch a vertex of a simple graph is to exchange its sets of neighbours and non-neighbours. The
graph so obtained is called a vertex-switching of the graph. (For example, switching any vertex of a
4-circuit yields a K1,3 , as does switching any vertex of the empty graph on four vertices.)

The Switching Reconstruction Conjecture

Every simple graph on five or more vertices is reconstructible from its vertex-switchings.
(R.P. Stanley 1985)

Simple Surprising General Central Old Fertile

∗∗ ∗ ∗∗∗ ∗ ∗

• True for graphs on n vertices provided that n ̸≡ 0 (mod 4). (R.P. Stanley 1985)

References

P.J. Kelly, On Isometric Transformations, Ph.D. thesis, University of Wisconsin, 1942.

S.M. Ulam, A Collection of Mathematical Problems, Wiley (Interscience), New York, 1960, p. 29.
F. Harary, On the reconstruction of a graph from a collection of subgraphs, Theory of Graphs and its
Applications (Proc. Sympos. Smolenice, 1963, ed. M. Fiedler.) Publ. House Czechoslovak Acad. Sci.,
Prague, 1964, pp. 47–52; reprinted by Academic Press, New York.
C. Thomassen, Counterexamples to the Edge Reconstruction Conjecture for infinite graphs, Discrete
Math. 19 (1977) 293–295.
R.P. Stanley, Reconstruction from vertex-switching, J. Combin. Theory Ser.B 38 (1985), 132–138.
M.N. Ellingham, Recent progress in edge reconstruction, Congressus Numerantium 62 (1988), 3–20.
8 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

J.A. Bondy, A graph reconstructor’s manual, in Surveys in Combinatorics. Proc. 13th. British Combinatorial
Conf., London Math. Soc. Lecture Notes Series 166 (ed. A.D. Keedwell), Cambridge University Press,
Cambridge, 1991.
L. Babai, Automorphism groups, isomorphism, reconstruction. Chapter 27 of the Handbook of
Combinatorics (R.L. Graham, M. Grötschel and L. Lovász, eds.), North-Holland–Elsevier, 1995, pp. 1447–
1540.
J. Lauri and R. Scapellato, Topics in Graph Automorphisms and Reconstruction, London Math. Soc. Student
Texts 54, Cambridge University Press, 2003.

5. Decompositions of graphs

A decomposition of a graph is a family of subgraphs whose edge sets partition the edge set of the
graph.
Decompositions of graphs occur in many contexts in combinatorics. For example, a Steiner triple
system can be viewed as a decomposition of a complete graph into triangles. The three conjectures
below were all proposed in the same year. They concern decompositions into paths and circuits,
perhaps the most basic of subgraphs (along with matchings and trees). The bound on the number
of paths in Gallai’s Conjecture is attained by graphs in which each degree is odd, and also by complete
graphs on an odd number of vertices. Hajós’ Conjecture is simply a quantified version of Veblen’s
Theorem, that any graph whose vertices are of even degree can be decomposed into circuits. The
bound here, while tight for complete graphs on an odd number of vertices, seems to be easily met in
most cases. Kelly’s Conjecture is a directed version of the fact that complete graphs on an odd number
of vertices admit decompositions into Hamilton circuits.

Gallai’s Conjecture
Every connected simple graph on n vertices can be decomposed into at most 12 (n + 1) paths.
(T. Gallai 1968)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗∗ ∗ ∗ ∗∗

• True for graphs in which all degrees are odd. (L. Lovász 1968)

Fig. 3. Tibor Gallai.

 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 9

• Every connected simple graph on n vertices can be decomposed into at most 43 n paths.
(A. Donald 1980)

A graph is even if every vertex is of even degree.

Hajós’ Conjecture
Every even simple graph on n vertices can be decomposed into at most 21 (n − 1) circuits.
(G. Hajós 1968)

Simple Surprising General Central Old Fertile

∗∗∗ ∗ ∗ ∗ ∗∗

• True for planar graphs. (T. Jiang 1984)

• True for graphs with maximum degree four. (O. Favaron and M. Kouider 1988)

Kelly’s Conjecture
Every regular tournament can be decomposed into directed Hamilton circuits. (P.J. Kelly 1968)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗∗ ∗ ∗ ∗ ∗

• Claimed true for very large tournaments. (R. Häggkvist, unpublished)

References
L. Lovász, On covering of graphs, in Theory of Graphs (P. Erdős and G. Katona, eds.), Academic Press,
New York, 1968, pp. 231–236.
J.W. Moon, Topics on Tournaments, Holt, Rinehart and Winston, New York, 1968, p.7, ex.9.
A. Donald, An upper bound for the path number of a graph, J. Graph Theory 4 (1980), 189–201.
T. Jiang, Hajós conjecture is true for planar graphs, J. China Univ. Sci. Tech. 14 (1984), 585–592.
O. Favaron and M. Kouider, Path partitions and cycle partitions of eulerian graphs of maximum degree
4, Studia Sci. Math. Hungarica, 23 (1988), 237–244.

6. Double covers of graphs

A double covering (or double cover) of a graph is a family of subgraphs which together cover each
edge exactly twice; quadruple and sextuple coverings are defined similarly.
The motivations for studying double coverings are several. To begin with, the concept of a double
covering is a relaxation of that of a decomposition. It is natural, therefore, to see whether a graph
which fails to have a decomposition (into circuits or perfect matchings, for instance) at least admits a
double covering. In the case of circuits, this line of thought leads directly to the Circuit Double Cover
Conjecture and its quantified version, the Small Circuit Double Cover Conjecture. In the case of perfect
matchings, it leads to the Matching Double Cover Conjecture.
P.D. Seymour arrived at the Circuit Double Cover Conjecture from a quite different angle, seeing it as
a particularly interesting case of a general conjecture on circuit coverings of graphs. And, before him,
both W.T. Tutte (unpublished) and G. Szekeres, who had proposed the conjecture for cubic graphs,
came to it through their work on embeddings of graphs.
10 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

The Cycle Double Cover Conjecture has its genesis in face colourings of maps on surfaces. Indeed, it is
implied by a far-reaching conjecture on graph embeddings, due independently to D. Archdeacon and
F. Jaeger.

The Circuit Double Cover Conjecture

Every 2-edge-connected graph admits a double covering by circuits.
(G. Szekeres 1973; P.D. Seymour 1979)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗ ∗∗∗ ∗∗∗ ∗ ∗∗∗

• If false, a minimal counterexample must have girth at least ten. (L.A. Goddyn 1988)

The Small Circuit Double Cover Conjecture

Every 2-edge-connected simple graph on n vertices has a double covering by at most n − 1 circuits.
(J.A Bondy 1990)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗ ∗∗∗ ∗∗ ∗

• True for graphs in which some vertex is adjacent to every other vertex. (H. Li 1990)
The Cycle Double Cover Conjecture
Every 2-edge-connected graph has a double covering by at most five even subgraphs.
(M. Preissmann 1981)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗∗ ∗∗∗ ∗∗ ∗ ∗

• True for 4-edge-connected graphs. (P.A. Kilpatrick 1975; F. Jaeger 1976)

The Strong Embedding Conjecture

Every 2-edge-connected graph can be embedded on some orientable surface in such a way that the resulting
map is 5-face-colourable. (D. Archdeacon 1984; F. Jaeger 1985)

Simple Surprising General Central Old Fertile

∗∗ ∗∗ ∗∗∗ ∗∗∗ ∗

The Matching Double Cover Conjecture

Every 2-edge-connected cubic graph has a double covering by six perfect matchings.
(R.D. Fulkerson 1971)

Simple Surprising General Central Old Fertile

∗∗∗ ∗ ∗∗ ∗ ∗ ∗

• Every 2-edge-connected cubic graph has an exact multicovering by perfect matchings.

(J. Edmonds 1965)
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 11

Fig. 4. Paul Seymour and Myriam Preissmann.

Reformulation:
The Cycle Quadruple Cover Conjecture
Every 2-edge-connected graph has a quadruple covering by six even subgraphs. (F. Jaeger 1985)

• Every 2-edge-connected graph has a quadruple covering by seven even subgraphs.

(J.C. Bermond, B. Jackson and F. Jaeger 1983)
• Every 2-edge-connected graph has a sextuple covering by ten even subgraphs.
(G. Fan 1992)

References
J. Edmonds, Maximum matching and a polyhedron with (0, 1) vertices, J. Res. Nat. Bur. Standards Ser.
B 8 (1965), 125–130.
D.R. Fulkerson, Blocking and antiblocking pairs of polyhedra, Math. Programming 1 (1971), 168–194.
G. Szekeres, Polyhedral decompositions of cubic graphs, Bull. Austral. Math. Soc. 8 (1973), 367–387.
P.D. Seymour, Sums of circuits, in Graph Theory and Related Topics (J.A. Bondy and U.S.R. Murty, eds.),
Academic Press, New York, 1979, pp. 341–355.
M. Preissmann, Sur les colorations des arêtes des graphes cubiques, Thèse de doctorat de 3ème cycle,
Université de Grenoble, 1981.
J.C. Bermond, B. Jackson and F. Jaeger, Shortest coverings of graphs with cycles, J. Combin. Theory Ser.
B 35 (1983), 297–308.
D. Archdeacon, Face colorings of embedded graphs, J. Graph Theory 8 (1984), 387–398.
F. Jaeger, A survey of the cycle double cover conjecture, in Cycles in Graphs (B.R. Alspach and C.D. Godsil,
eds.), Ann. Discrete Math. 27, North-Holland, Amsterdam, 1985, pp. 1–12.
J.A. Bondy, Small cycle double covers of graphs, in Cycles and Rays (G. Hahn, G. Sabidussi and
R. Woodrow, eds.), NATO ASI Ser. C, Kluwer Academic Publishers, Dordrecht, 1990, pp. 21–40.
H. Li, Perfect path double covers in every simple graph, J. Graph Theory 14 (1990), 645–650.
G. Fan, Integer flows and cycle covers, J. Combin. Theory Ser. B 54 (1992), 113–122.
C.Q. Zhang, Integer Flows and Cycle Covers of Graphs, Marcel Dekker, New York, 1997.

7. Integer flows in graphs

A (nowhere-zero) k-flow on a digraph is an assignment of a nonzero integer between −(k − 1) and

k − 1 to each arc so that, at each vertex, the sums of the values on the arcs entering and leaving the
vertex are equal. A graph is said to have a k-flow if some (hence every) orientation of the graph has
such a flow.
12 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

The concept of an integer-valued flow is due to W.T. Tutte, who observed that when a graph is
embedded on a surface, a face colouring of the map so defined corresponds to such a flow on the graph.
He formulated three celebrated conjectures which generalize three theorems on planar graphs, the
Five-Colour Theorem, the Four-Colour Theorem (at the time, still a conjecture) and Grötzsch’s Three-
Colour Theorem.

The Five-Flow Conjecture

Every 2-edge-connected graph has a 5-flow. (W.T. Tutte 1954)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗∗ ∗∗∗

• Every 2-edge-connected graph has a 6-flow. (P.D. Seymour 1981)

A minor of a graph is any graph obtained from it by means of vertex and edge deletions and edge
contractions.

The Four-Flow Conjecture

Every 2-edge-connected graph with no Petersen minor has a 4-flow. (W.T. Tutte 1966)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗∗ ∗∗ ∗∗ ∗∗ ∗

• Every 2-edge-connected cubic graph with no Petersen minor has a 4-flow.

(N. Robertson, P.D. Seymour and R. Thomas, unpublished)

The Three-Flow Conjecture

Every 4-edge-connected graph has a 3-flow. (W.T. Tutte 1972)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗∗ ∗∗ ∗∗ ∗

• Every 4-edge-connected graph has a 4-flow. (F. Jaeger 1976)

Weaker variant:
The Weak Three-Flow Conjecture
There exists an integer k such that every k-edge-connected graph has a 3-flow.
(F. Jaeger, 1976)

References
W.T. Tutte, A contribution to the theory of chromatic polynomials, Canadian J. Math. 6 (1954), 80–91.
P.D. Seymour, Nowhere-zero 6-flows, J. Combin. Theory Ser. B 30 (1979), 130–135.
F. Jaeger, Nowhere-zero flow problems, in Selected Topics in Graph Theory, L.W. Beineke and R.J. Wilson
(eds), Academic Press, 1988, pp. 71–95.
P.D. Seymour, Nowhere-zero flows, in Handbook of Combinatorics, (R.L. Graham, M. Grötschel, and
L. Lovász, eds.), North-Holland, 1995, pp. 289–299.
C.Q. Zhang, Integer Flows and Cycle Covers of Graphs, Marcel Dekker, New York, 1997.
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 13

Fig. 5. William Tutte.

8. Vertex colourings of graphs

The concept of colouring permeates graph theory, and many questions can be formulated as
colouring problems. A large number of these are discussed in the book Graph Coloring Problems by
T. Jensen and B. Toft. We limit ourselves here to a small number of prominent and easily stated
conjectures on the topic. The first two, which go hand-in-hand, are discussed at some length in the
survey by D. Romero and A. Sánchez-Arroyo which appears in this volume, and we content ourselves
with merely stating and rating them. The other two are well-known longstanding conjectures.
Warning: proving Hadwiger’s Conjecture will be nontrivial: the case k = 4 implies the Four-Colour
Theorem. On the other hand, opinion is divided as to the truth of the conjecture for graphs with
stability number two (hence high chromatic number).

The Erdős–Faber–Lovász Conjecture

Every graph which can be decomposed into k complete graphs on k vertices is k-colourable
(P. Erdős, V. Faber and L. Lovász 1972)

Simple Surprising General Central Old Fertile

∗ ∗∗ ∗ ∗ ∗

The Alon–Saks–Seymour Conjecture

Every graph which can be decomposed into k − 1 complete bipartite graphs is k-colourable.
(N. Alon, M. Saks and P.D. Seymour 1994)

Simple Surprising General Central Old Fertile

∗∗ ∗ ∗∗∗ ∗

• True for complete graphs. (R.L. Graham and H.O. Pollak 1971; H. Tverberg 1982)

Hadwiger’s Conjecture
A graph with no Kk+1 -minor is k-colourable. (H. Hadwiger 1943)
14 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

Fig. 6. Paul Erdős, László Lovász and Noga Alon.

Simple Surprising General Central Old Fertile

∗∗ ∗ ∗∗∗ ∗∗ ∗∗∗ ∗∗

• True for k ≤ 5. (K. Appel and W. Haken, 1977; N. Robertson, √ P. Seymour and R. Thomas 1993)
• A graph with no Kk+1 -minor can be coloured using at most ck log k colours, where c is a constant.
(A.V. Kostochka 1982; A. Thomason 1984)

The weak product of graphs G and H is the graph G × H with vertex set V (G) × V (H ) and edge set
{((u, u′ ), (v, v ′ )) : (u, v) ∈ E (G), (u′ , v ′ ) ∈ E (H )}.

Hedetniemi’s Conjecture
The chromatic number of the weak product of two graphs is equal to the lesser of their chromatic numbers:
χ (G × H ) = min{χ (G), χ (H )}
(S. Hedetniemi 1966)

Simple Surprising General Central Old Fertile

∗∗ ∗∗ ∗ ∗ ∗∗ ∗

• χ(G × H ) ≤ min{χ (G), χ (H )}. (S. Hedetniemi 1966)

• True if min{χ (G), χ (H )} ≤ 4. (M. El-Zahar and N. Sauer 1985)

References
H. Hadwiger, Über eine Klassifikation der Streckenkomplexe, Vierteljschr. Naturforsch. Gesellsch. Zürich
88 (1943), 133–142.
S. Hedetniemi, Homomorphisms of graphs and automata, Technical Report 03105-44-T, University of
Michigan, 1966.
R.L. Graham and H.O. Pollak, On the addressing problem for loop switching, Bell System Tech. J. 50
(1971), 2495–2519.
K. Appel and W. Haken, A solution of the four-color map problem, Sci. Amer. 237 (1977), 108–121.
H. Tverberg, On the decomposition of Kn into complete bipartite graphs, J. Graph Theory 6 (1982),
493–494.
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 15

M. El-Zahar and N. Sauer, The chromatic number of the product of two 4-chromatic graphs is 4,
Combinatorica 5 (1985), 121–126.
N. Robertson, P. Seymour and R. Thomas, Hadwiger’s conjecture for K6 -free graphs, Combinatorica 13
(1993), 279–361.
T.R. Jensen and B. Toft, Graph Coloring Problems, Wiley Interscience, 1995.
B. Toft, A survey of Hadwiger’s Conjecture, in Surveys in Graph Theory (G. Chartrand and M.S. Jacobson,
eds.), Congressus Numerantium 115 (1996), 249–283.
X. Zhu, A survey on Hedetniemi’s conjecture, Taiwanese J. Math. 2 (1998), 1–24.
N. Sauer, Hedetniemi’s conjecture — a survey, Discrete Math. 229 (2001), 261–292.
A. Thomason, The extremal function for complete minors, J. Combin. Theory Ser. B 81 (2001), 318–338.
D. Romero and A. Sánchez-Arroyo, Advances on the Erdős–Faber–Lovász Conjecture, this volume.

9. Edge colourings of graphs

A graph is k-list-colourable if, whenever its vertices are assigned lists of k colours, one colour can be
chosen from each list so as to yield a proper vertex colouring of the graph. The edge analogue, k-list
edge colourability, is defined similarly.
List colourings and list edge colourings were introduced by several authors in the 1970s. At first glance,
k-list colouring appears to be at least as easy as k-colouring, but a moment’s reflection reveals that
the converse is in fact true. Indeed, large complete bipartite graphs, which are 2-colourable, are k-list
colourable only for large k. List edge colourability, on the other hand, is conjectured to be no harder
than edge colourability.

The List Edge Colouring Conjecture

Every k-edge-colourable simple graph is k-list-edge-colourable. (V.G. Vizing 1975)

Simple Surprising General Central Old Fertile

∗∗ ∗∗∗ ∗∗∗ ∗ ∗ ∗∗

• True for bipartite (multi)graphs. (F. Galvin 1995)

• Every k-edge-colourable simple graph is (k + o(k))-list-edge-colourable. (J. Kahn 1996)

References
B. Bollobás and A.J. Harris, List colourings of graphs, Graphs Combin. 1 (1985), 115–127.
F. Galvin, The list chromatic index of a bipartite multigraph, J. Combin. Theory Ser. B 63 (1995),
153–158.
J. Kahn, Asymptotically good list-colorings, J. Combin. Theory Ser. A 73 (1996), 1–59.

10. Total colourings of graphs

A (proper) total colouring of a graph is an assignment of colours to its vertices and edges so that
no two adjacent vertices, no two incident edges, and no incident vertex and edge receive the same
colour.
For a simple graph with maximum degree ∆, Brooks’ Theorem ensures a vertex colouring in at most
∆+1 colours, while Vizing’s Theorem guarantees an edge colouring in at most ∆+1 colours. Colouring
vertices and edges simultaneously is much more tricky. The Total Colouring Conjecture claims that
only one further colour is needed to achieve such a colouring.
16 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

The Total Colouring Conjecture

Every simple graph with maximum degree ∆ has a total colouring by ∆ + 2 colours.
(M. Behzad 1965; V.G. Vizing 1965)

Simple Surprising General Central Old Fertile

∗∗ ∗∗∗ ∗∗∗ ∗ ∗∗ ∗

• Every simple graph with maximum degree ∆ has a total colouring by ∆ + 1026 colours, for ∆
sufficiently large. (M. Molloy and B. Reed 1998)

References
M. Behzad, Graphs and their Chromatic Numbers, Ph.D. thesis, Michigan State University, 1965.
V.G. Vizing, Some unsolved problems in graph theory, Russian Math. Surveys 23 (1968), 125–141.
M. Molloy and B. Reed, A bound on the total chromatic number, Combinatorica 18 (1998), 241–280.

11. Circuits in digraphs

The digraphs here are assumed to be strict, that is, without loops or multiple arcs.
A cage is a smallest regular graph of given degree and girth, the girth being the length of a shortest
circuit. The analogous concept for digraphs is the directed cage, in which regularity is replaced by the
condition that all indegrees and outdegrees are equal, and girth by the length of a shortest directed
circuit.
Much is known about cages, little about directed cages. M. Behzad, G. Chartrand and C.E. Wall
conjectured that the digraphs of order 1 (mod d) whose vertices can be arranged in a cyclic order
so that each vertex dominates the d vertices following it are directed cages. A stronger conjecture,
in which the constraint on the indegree is dropped, was proposed by L. Caccetta and R. Häggkvist.
These conjectures are unsettled even for digraphs of directed girth four. Another strengthening of this
particularly interesting instance of the Behzad–Chartrand–Wall Conjecture is P.D. Seymour’s Second
Neighbourhood Conjecture.
Another longstanding open problem is Adám’s Conjecture on the number of directed circuits.

The Behzad–Chartrand–Wall Conjecture

A directed cage of girth g and in- and outdegree d has exactly d(g − 1) + 1 vertices.
(M. Behzad, G. Chartrand and C.E. Wall 1970)

Simple Surprising General Central Old Fertile

∗∗ ∗ ∗ ∗ ∗∗

• True for vertex-transitive directed cages. (Y.O. Hamidoune 1987)

The Caccetta–Häggkvist Conjecture

A digraph of girth g in which each outdegree is d has at least d(g − 1) + 1 vertices.
(L. Caccetta and R. Häggkvist 1978)

Simple Surprising General Central Old Fertile

∗∗ ∗∗∗ ∗∗ ∗ ∗∗
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 17

Fig. 7. Roland Häggkvist.

• True for d ≤ 5. (C.T. Hoàng and

√ B. Reed 1987)
• A digraph of girth at least four in which each outdegree is d has more than (3 + 7)d/2 vertices.
(J. Shen 1998)

A second outneighbour of a vertex v in a digraph is a vertex whose distance from v is two.

The Second Neighbourhood Conjecture

Every digraph without circuits of length two has a vertex with at least as many second outneighbours as
(first) outneighbours. (P.D. Seymour 1995)

Simple Surprising General Central Old Fertile

∗∗ ∗∗ ∗∗∗ ∗ ∗

• True for tournaments. (D.C. Fisher 1996; F. Havet and S. Thomassé 2000)
• Some vertex has at least .6572 times as many second outneighbours as outneighbours.
(G. Chen, J. Shen and R. Yuster 2003)

Adám’s Conjecture
Every digraph with at least one directed circuit has an arc whose reversal reduces the number of directed
circuits. (A. Adám 1964)

Simple Surprising General Central Old Fertile

∗∗∗ ∗ ∗∗∗ ∗ ∗∗

• False for digraphs with multiple arcs. (È. Grinberg 1987; C. Thomassen 1987)
References
A. Adám, Problem 2, in Theory of Graphs and its Applications (M. Fiedler, ed.), Czech. Acad. Sci. Publ.,
1964, p. 157.
M. Behzad, G. Chartrand and C.E. Wall, On minimal regular digraphs with given girth, Fund. Math. 69
(1970), 227–231.
L. Caccetta and R. Häggkvist, On minimal digraphs with given girth, Congressus Numerantium 21
(1978), 181–187.
È.Ja. Grinberg, Examples of non-Adám multigraphs, Latvian Math. Yearbook 31 (1987), 128–138
(in Russian).
18 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

Y.O. Hamidoune, A note on minimal directed graphs with given girth, J. Combin. Theory Ser. B 43 (1987),
343–34.
C.T. Hoàng and B. Reed, A note on short cycles in digraphs, Discrete Math. 66 (1987), 103–107.
C. Thomassen, Counterexamples to Adám’s conjecture on arc reversals in directed graphs, J. Combin.
Theory Ser. B 42 (1987), 128–130.
N. Dean and B.J. Latka, Squaring the tournament — an open problem, Congressus Numerantium 109
(1995), 73–80.
D.C. Fisher, Squaring a tournament: a proof of Dean’s conjecture. J. Graph Theory 23 (1996), 43–48.
F. Havet and S. Thomassé, Median orders of tournaments: a tool for the second neighborhood problem
and Sumner’s conjecture, J. Graph Theory 35 (2000), 244–256.
J. Shen, On the Caccetta–Häggkvist Conjecture, Graphs Combin. 18 (2002), 645–654.
G. Chen, J. Shen and R. Yuster, Second neighborhood vs. first neighborhood in digraphs, Annals of
Combinatorics 7 (2003), 15–20.
B.D. Sullivan, A summary of results and problems related to the Caccetta–Häggkvist conjecture,
preprint, Princeton University, 2006.

12. Long paths in digraphs

In the late 1960s, T. Gallai and B. Roy found, independently, a close link between the chromatic
number of a digraph and the length of its longest directed paths. At about the same time, a similar
link was discovered by T. Gallai and A.N. Milgram between the stability number and vertex coverings
by disjoint directed paths. The Laborde–Payan–Xuong Conjecture postulates a further connection
between stable sets and longest paths, a strengthening of the Gallai–Roy Theorem.

The Laborde–Payan–Xuong Conjecture

Every digraph has a stable set meeting all longest directed paths.
(J.M. Laborde, C. Payan and N.H. Xuong 1983)

Simple Surprising General Central Old Fertile

∗∗ ∗ ∗∗∗ ∗ ∗

• True for digraphs with stability number two. (F. Havet 2004)
• Not every digraph with stability number two has a directed path meeting all maximum stable sets.
(G. Hahn and B. Jackson 1990)

References
B. Roy, Nombre chromatique et plus longs chemins d’un graphe, Rev. Française Automat. Informat. Rech.
Opér. 1 (1967), 127–132.
T. Gallai, On directed paths and circuits, in Theory of Graphs (P. Erdős and G. Katona, eds.), Academic
Press, New York, 1968, pp. 115–118.
T. Gallai and A.N. Milgram, Verallgemeinerung eines graphentheoretischen Satzes von Rédei, Acta Sci.
Math. Szeged 21 (1960), 181–186.
G. Hahn and B. Jackson, A note concerning paths and independence number in digraphs, Discrete Math.
82 (1990), 327–329.
F. Havet, Stable set meeting every longest path, Discrete Math. 289 (2004), 169–173.
J.M. Laborde, C. Payan and N.H. Xuong, Independent sets and longest directed paths in digraphs, in
Graphs and Other Combinatorial Topics (M. Fiedler, ed.), Teubner Verlagsgesellschaft, Leipzig, 1983,
pp. 173–177.
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 19

13. Long circuits in graphs

The interest in long circuits has its origins in the relation between Hamilton circuits and the Four-
Colour Problem, although long circuits do arise naturally in other areas, such as the Snake-in-the-Box
Problem in coding theory. G.A. Dirac was perhaps the first to study long circuits in their own right, in
the early 1950s. Open problems such as Thomassen’s Chord Conjecture show that our understanding
of such basic objects leaves much to be desired.
A chord of a circuit is an edge joining two nonconsecutive vertices.

Thomassen’s Chord Conjecture

Every longest circuit in a 3-connected graph has a chord.
(C. Thomassen 1976)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗ ∗ ∗ ∗

• True for bipartite graphs. (C. Thomassen 1997)

Smith’s Conjecture
In a k-connected graph, where k ≥ 2, any two longest circuits have at least k vertices in common.
(S. Smith 1984)

Simple Surprising General Central Old Fertile

∗∗∗ ∗ ∗∗∗ ∗

• True for k ≤ 6. √ (M. Grötschel 1984)

• Any two longest circuits in a k-connected graph have at least 3 k vertices in common.
(L. Babai 1979)

Birmelé’s Conjecture
A graph in which any two circuits of length k or more intersect has a set of k vertices which meet all such
circuits. (E. Birmelé 2002)

Simple Surprising General Central Old Fertile

∗∗ ∗∗ ∗

• True for k ≤ 5. (L. Lovasz 1965; E. Birmelé 2003)

• There is a set of 2k + 3 vertices which meet all circuits of length at least k.
(E. Birmelé, J.A. Bondy and B.A. Reed 2003)

14. Hamilton circuits in graphs

Hamilton circuits were first investigated in the mid 1850s by W.R. Hamilton, in connection with a
problem in group theory, and independently by T.D. Kirkman. But the main motivation for their study
came from their connection with the Four-Colour Problem, namely P.G. Tait’s conjecture that every
3-connected cubic planar graph has such a circuit (refuted by W.T. Tutte 66 years later). More recently,
other questions, such as the Travelling Salesman Problem, have fuelled interest in Hamilton circuits.
A graph is claw-free if it contains no induced copy of K1,3 .
20 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

Fig. 8. Carsten Thomassen.

The Matthews–Sumner Conjecture

Every 4-connected claw-free graph is hamiltonian. (M. Matthews and D. Sumner 1984)

Simple Surprising General Central Old Fertile

∗∗ ∗ ∗ ∗ ∗ ∗

Special case (in fact, equivalent, by a result of Ryjaček 1997):

Thomassen’s Line Graph Conjecture
Every 4-connected line graph is hamiltonian. (C. Thomassen 1986)

• Line graphs of 4-edge-connected graphs are hamiltonian. (C. Thomassen 1986)

• Every 7-connected line graph is hamiltonian. (S.M. Zhan 1991)

References
L. Lovász, On graphs not containing independent circuits (Hungarian), Mat. Lapok 16 (1965), 289–299.
L. Babai, Long cycles in vertex transitive graphs, J. Graph Theory 3 (1979), 301–304.
M. Grötschel, On intersections of longest cycles, in Graph Theory and Combinatorics (B. Bollobás, ed.),
Academic Press, London, 1984, pp. 171–189.
M. Matthews and D. Sumner, Hamiltonian results in K1,3 -free graphs, J. Graph Theory 8 (1984),
139–146.
C. Thomassen, Reflections on graph theory, J. Graph Theory 10 (1986), 309–324.
S.M. Zhan, On hamiltonian line graphs and connectivity, Discrete Math. 89 (1991), 89–95.
Z. Ryjaček, On a closure concept in claw-free graphs, J. Combin. Theory Ser. B 70 (1997), 217–224.
C. Thomassen, Chords of longest cycles in cubic graphs, J. Combin. Theory Ser. B 71 (1997), 211–214.
E. Birmelé, Thèse de doctorat, Université Lyon 1, 2003.
E. Birmelé, J.A. Bondy and B.A. Reed, The Erdős–Pósa property for long circuits, Combinatorica, to
appear.

15. Multiple Hamilton circuits in graphs

A classical theorem of C.A.B. Smith says that each edge of a cubic graph lies in an even number
of Hamilton circuits. Thus, when a cubic graph has a Hamilton circuit, it necessarily has at least two
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 21

Fig. 9. John Sheehan and Herbert Fleischner.

(in fact, at least three). A.G. Thomason extended Smith’s Theorem to all regular graphs of odd degree
(indeed, to all graphs whose degrees are odd). The corresponding question for regular graphs of even
degree was proposed by J. Sheehan.
Smith’s Theorem also poses a very interesting algorithmic question, that of finding a second Hamilton
circuit when supplied with one. His proof is based on a counting argument, and is nonconstructive.
There appears to be something special about the barrier of 4-connectivity in connection with Hamilton
circuits. One may cite the theorem of W.T. Tutte that 4-connected planar graphs are hamiltonian, the
conjecture of M. Matthews and D. Sumner that 4-connected claw-free graphs are hamiltonian (see
above), and the theorem of C. Thomassen that any vertex in a 4-connected tournament is connected
to any other vertex by a directed Hamilton path. Fleischner’s Conjecture on second Hamilton circuits
is yet another instance of this phenomenon.

Sheehan’s Conjecture
Every simple 4-regular graph with a Hamilton circuit has a second Hamilton circuit. (J. Sheehan 1975)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗ ∗ ∗ ∗∗

• Every simple (2k + 1)-regular graph with a Hamilton circuit has a second Hamilton circuit.
(A.G. Thomason 1978)
• For k ≥ 11, every simple 2k-regular graph with a Hamilton circuit has a second Hamilton circuit.
(P. Haxell, B. Seamone and J. Verstraëte 2005)
• There exist simple graphs of minimum degree four with exactly one Hamilton circuit.
(H. Fleischner 2004)

The Chrobak–Poljak Problem

Given a Hamilton circuit in a cubic graph, find (in polynomial time) a second Hamilton circuit.
(M. Chrobak and S. Poljak 1988)

Simple Surprising General Central Old Fertile

∗∗∗ ∗ ∗ ∗
22 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23

Fleischner’s Conjecture
Every 4-connected graph with a Hamilton circuit has a second Hamilton circuit.
(H. Fleischner 2004)

Simple Surprising General Central Old Fertile

∗∗∗ ∗∗ ∗ ∗

• True for planar graphs. (W.T. Tutte 1956)

References
W.T. Tutte, On hamiltonian circuits, J. London Math. Soc. 21 (1946), 98–101.
W.T. Tutte, A theorem on planar graphs, Trans. Amer. Math. Soc. 82 (1956), 99–116.
J. Sheehan, The multiplicity of Hamiltonian circuits in a graph, in Recent Advances in Graph Theory
(M. Fiedler, ed.), Academia, Prague, 1975, pp. 447–480.
A.G. Thomason, Hamiltonian cycles and uniquely edge colourable graphs, Ann. Discrete Math. 3 (1978),
259–268.
M. Chrobak and S. Poljak, On common edges in optimal solutions to traveling salesman and other
optimization problems, Discrete Appl. Math. 20 (1988), 101–111.
N. Alon, Non-constructive proofs in combinatorics, in Proc. Internat. Congr. Math., Kyoto, 1990,
Springer-Verlag, Tokyo, 1991, pp. 1421–1429.
H. Fleischner, Uniqueness of maximal dominating cycles in 3-regular and of Hamiltonian cycles in
4-regular graphs, J. Graph Theory 18 (1994), 449–459.
C. Thomassen, On the number of hamiltonian cycles in bipartite graphs, Combin. Probab. Comput. 5
(1996), 437–442.
C. Thomassen, Chords of longest cycles in cubic graphs, J. Combin. Theory Ser. B 71 (1997), 211–214.
C. Thomassen, Independent dominating sets and a second Hamiltonian cycle in regular graphs,
J. Combin. Theory Ser. B 72 (1998), 104–109.
H. Fleischner, Uniquely Hamiltonian graphs of minimum degree four, preprint, 2004.
P. Haxell, B. Seamone and J. Verstraëte, Independent dominating sets and hamiltonian cycles, preprint,
2005.

Acknowledgements

The photograph of Tibor Gallai was a class photo, taken when he was a high-school teacher. It is
reproduced here by kind permission of one of his former pupils, Vera Sándor. The photographs of
Myriam Preissmann and Noga Alon appear on their web pages. I thank Gilles Kuntz and Noga Alon
for kindly allowing me to include them here. The remaining photographs were taken by (and are
copyright of) the author.

Added in press

Since this article was submitted for publication (in 2009), there have been several significant
advances in graph theory, some of which concern conjectures cited here. Kelly’s Conjecture has
been proved for large tournaments by D. Kuhn and D. Osthus (Hamilton decompositions of regular
expanders: a proof of Kelly’s conjecture for large tournaments, arXiv:1202.6219). Jaeger’s Weak Three-
Flow Conjecture has been proved by C. Thomassen (The weak 3-flow conjecture and the weak circular
flow conjecture, J. Combin. Theory Ser. B 102 (2012), 521–529). Thomassen’s proof is valid for 8-edge-
connected graphs. It has since been refined by L.M. Lovász, C. Thomassen, Y. Wu and C.Q. Zhang, who
 A. Bondy / European Journal of Combinatorics 37 (2014) 4–23 23

reduced the edge connectivity bound to six. The Alon–Saks–Seymour Conjecture has been disproved
by H. Huang and B. Sudakov (A counterexample to the Alon–Saks–Seymour conjecture and related
problems, Combinatorica 32 (2) (2012), 205–219, http://dx.doi.org/10.1007/s00493-012-2746-4). A
greatly expanded version of this article is currently in preparation, and will eventually be published
as a monograph.