The Behavioural Biology of Dogs

THE BEHAVIOURAL BIOLOGY OF DOGS

This page intentionally left blank

The Behavioural Biology of Dogs
Edited by

Per Jensen
IFM Biology
Linköping University
Sweden
CABI is a trading name of CAB International
CABI Head Office CABI North American Office

Nosworthy Way 875 Massachusetts Avenue
Wallingford 7th Floor
Oxfordshire OX10 8DE Cambridge, MA 02139
UK USA
Tel: +44 (0)1491 832111 Tel: +1 617 395 4056

Fax: +44 (0)1491 833508 Fax: +1 617 354 6875
E-mail: [email protected] E-mail: [email protected]
Website: www.cabi.org
© CAB International 2007. All rights reserved. No part of this publication may be
reproduced in any form or by any means, electronically, mechanically, by
photocopying, recording or otherwise, without the prior permission of the
copyright owners.
A catalogue record for this book is available from the British Library, London,
UK.
A catalogue record for this book is available from the Library of Congress,
Washington, DC.
ISBN-10: 1 84593 1874

ISBN-13: 978 1 84593 1872
Typeset by MRM Graphics Ltd, Winslow, Bucks.

Printed and bound in the UK by Cromwell Press, Trowbridge.
Contents
Contributors vii
Preface ix
I: The Dog in Its Zoological Context 1

Editor’s Introduction
1. Evolutionary History of Canids 3

X. Wang and R.H. Tedford
2. Domestication of Dogs 21
P. Savolainen
3. Origin of Dog Breed Diversity 38

C. Vilà and J.A. Leonard
II: Biology and Behaviour of Dogs 59

4. Mechanisms and Function in Dog Behaviour 61

P. Jensen
5. Behaviour Genetics in Canids 76

E. Jazin
v
vi Contents
6. Sensory Physiology and Dog Behaviour 91

H. Bubna-Littitz
7. Social Behaviour of Dogs and Related Canids 105

D.U. Feddersen-Petersen
8. Learning in Dogs 120

P. Reid
III: The Dog in Its Niche: Among Humans 145

9. Behaviour and Social Ecology of Free-ranging Dogs 147

L. Boitani, P. Ciucci and A. Ortolani
10. Evolutionary Aspects on Breeding of Working Dogs 166

R. Beilharz
11. Individual Differences in Behaviour – Dog Personality 182

K. Svartberg
12. Human–Animal Interactions and Social Cognition in Dogs 207

Á. Miklósi
IV: Behavioural Problems of Dogs 223

13. Behavioural Disorders of Dogs 225

R.A. Mugford
14. Behaviour and Disease in Dogs 243

Å. Hedhammar and K. Hultin-Jäderlund
Index 263
Contributors
Beilharz, R., School of Agriculture and Food Systems, Faculty of Land and Food Resources,
The University of Melbourne, Melbourne, Victoria 3010, Australia
Boitani, L., Department of Animal and Human Biology, University of Rome, Viale
Università 32, 00185 Roma, Italy
Bubna-Littitz, H., Department for Natural Sciences, Institute for Physiology, Veterinary
University of Vienna, Veterinärplatz 1, A-1030 Vienna, Austria
Ciucci, P., Department of Animal and Human Biology, University of Rome, Viale
Feddersen-Petersen, D.U., Department of Zoology, University of Kiel, Olshausenstr.
40, D-24118 Kiel, Germany
Hedhammar, Å., Department of Small Animal Clinical Sciences, Swedish University of
Agricultural Sciences, Box 7037, SE-750 07 Uppsala, Sweden
Hultin-Jäderlund, K., Department of Small Animal Clinical Sciences, Norwegian
College of Veterinary Medicine, PO Box 8146 Dep., N-0033 Oslo, Norway
Jazin, E., Department of Evolution, Genomics and Systematics, Uppsala University,
Norbyvagen 18D, SE-752 36 Uppsala, Sweden
Jensen, P., IFM Biology, Linköping University, SE-581 83 Linköping, Sweden
Leonard, J.A., Department of Evolutionary Biology, Uppsala University, Norbyvägen
18D, SE-752 36 Uppsala, Sweden
Miklósi, Á., Department of Ethology, Eötvös Loránd University, Pázmány sétány 1/c
1117, Budapest, Hungary
Mugford, R.A., The Animal Behaviour Centre, PO Box 23, Chertsey, Surrey KT16 9NL,
UK
Ortolani, A., Department of Animal and Human Biology, University of Rome, Viale
Reid, P., ASPCA Animal Behavior Center, 424 East 92nd Street, New York, NY 10128, USA
vii
viii Contributors
Savolainen, P., Department of Biotechnology, Albanova University Centre, KTH-Royal

Institute of Technology, SE-10691 Stockholm, Sweden
Svartberg, K., Department of Anatomy and Physiology, Swedish University of Agricultural
Sciences, PO Box 7011, SE-750 07 Uppsala, Sweden
Tedford, R.H., Department of Paleontology, American Museum of Natural History,
Central Park West at 79th Street, New York, NY 10024-5192, USA
Vilà, C., Department of Evolutionary Biology, Uppsala University, Norbyvägen 18D, SE-
752 36 Uppsala, Sweden
Wang, X., Department of Vertebrate Paleontology, Natural History Museum of Los Angeles
County, 900 Exposition Blvd., Los Angeles, CA 90007, USA and Department of
Paleontology, American Museum of Natural History, Central Park West at 79th Street,
New York, NY 10024-5192, USA
Preface
Dogs are our oldest domesticated animals, our friends and companions, and the
most widespread species of all animals under human care. There are few things
for which dogs have not been utilized during our joint history: they have been
used as hunting aids, for pulling and carrying, as providers of meat, fur and other
products, for guarding and watching, and as laboratory test animals, just to
mention some examples. In large parts of the world, dogs are increasingly popular
as pets, friends and family members, and at the same time the importance of dogs
as working dogs, for example with police and rescue forces, has probably never
been higher.
At the same time, biologists have started to take greater interest in the biology
and behaviour of dogs, for a number of reasons. First, since dogs are so closely
connected to human evolution and history, understanding how, where and when
dogs have developed and spread helps us understand our own background.
Second, the long coexistence with humans has led the dog to develop specific
adaptations facilitating life with us, and the dog therefore provides an excellent
model for studying how behaviour and cognition have evolved. Third, the increas-
ing importance of dogs as pets and working animals calls for a deeper biological
knowledge of how these animals actually work – such information can help us not
only to train and shape dogs for specific tasks, but also to prevent and cure various
behavioural disorders which may cause owners and animals large problems.
The present book is an attempt to provide an up-to-date description of the
behavioural biology of dogs, written by experts in different areas of this large field.
The target audience consists of students of animal behaviour or veterinary medi-
cine at advanced levels – the book is not intended as an introductory text to dog
behaviour in general. It is also the hope of the authors that interested dog owners
outside academia may find usable parts in the book. There is no doubt that
ix
x Preface
certain chapters will require closer acquaintance with various aspects of biology
than the average dog owner is likely to possess, but we also believe that most
chapters contain aspects that are readily accessible.
The book is split into four different parts, each concerned with a specific
aspect of the behavioural biology of dogs. The first part (Chapters 1–3) is devoted
mainly to the evolution and development of the dog. Although not primarily con-
cerned with behaviour, these aspects form the basis for understanding how
behaviour has developed and for placing the dog in its relevant biological context.
The second part (Chapters 4–8) deals with basic aspects of animal behaviour with
particular emphasis on dogs. The third part (Chapters 9–12) places the modern
dog in its present ecological framework: in the niche of human coexistence. Here
we give a broad overview of the behavioural aspects of living close to humans. In
the last part of the book (Chapters 13 and 14), the emphasis is on behavioural
problems, their prevention and cure.
All of the contributors to this book have considerable research experience in
their areas, and it is hoped that this will guarantee that the text is relevant, up-to-
date and central to the subject.
Per Jensen
Linköping, March 2006
I The Dog in Its Zoological Context
In this first part of the book, the dog is placed in context amongst its zoological
relatives, and in relation to its domestication history. The first chapter outlines the
modern view on the zoological systematics of carnivores and canids in particular.
Here, the reader will find an exciting account of fossil and present traits which
allows the dog to be placed within the greater picture of closely related canids.
Of course, the dog is a domesticated species, and its domestication history
has been subject to intense research during the last decade or so. Here, modern
molecular genetics offers tools which have allowed biologists to give their pictures
of how, where and when domestication started, complementing the traditional
picture offered mainly by archaeologists. Some aspects of this new picture are
truly stunning and require that we revise large parts of our traditional views on
how domestication began.
Given the novelty of the molecular research on the ancestry of dogs, it should
not be a surprise that different scientific groups arrive at somewhat different con-
clusions. This is partly explained by variations in methods, and the only way to
resolve some of the disagreements is to continue to do more and improved
research. As this book is written, there is therefore only limited consensus among
biologists concerning, for example, the time when domestication started. As
editor, one has to make a decision – one can choose not to cover the new research
at all, or to only present the picture oneself believes in, or to present divergent pic-
tures and allow the reader to decide. I have chosen the last strategy, and there-
fore, the stories presented in Chapters 2 and 3 differ on some important points –
these differences will hopefully disappear in the light of future research.
Evolutionary History of Canids
1
Xiaoming Wang and Richard H. Tedford
Introduction
Members of the dog family (Canidae) are an early lineage of carnivorans (order
Carnivora). The canids were the first to branch off the caniform carnivorans, dog-
like predators that, in addition to canids, also include the bear family (Ursidae),
the raccoon family (Procyonidae), the weasel family (Mustelidae), as well as the
aquatic seals, sea lions and walruses (Pinnipedia). Living canids are some of the
most successful predators, occupying all continents except Antarctica, and reign
supreme as top predators in parts of northern North America and Eurasia. It is
thus quite remarkable that domestic dogs, known for their loyalty to human
masters, came from a dominant predator in the form of the grey wolf.
The history of domestic dogs occupies a tiny fraction of the long family
history, and represents a mere twig in a large family tree of more than 36 species
of wild canids living today (Wang et al., 2004a, b). Despite their impressive variety,
all dogs came from a single species of wolves in the latest Pleistocene during the
last Ice Age. Earliest fossil evidence of domestic dogs in archaeological sites dates
around 12,000–14,000 years ago in western Eurasia, whereas genetic evidence
suggests an East Asian origin around 15,000 years ago or possibly as old as
100,000 or more years ago (e.g. Leonard et al., 2002; Pennisi, 2002; Savolainen et
al., 2002). Often treated as a subspecies of the grey wolf, Canis lupus familiaris
(Linnaeus, 1758), domestic dogs bear numerous resemblances to their wild wolf
ancestors both morphologically and behaviourally.
The evolutionary history of canids is a history of successive radiations re-
peatedly occupying a broad spectrum of niches ranging from large, pursuit pred-
ators to small omnivores, or even to herbivores. Three such radiations were first
© CAB International 2007. The Behavioural Biology of Dogs
(ed. P. Jensen) 3
4 X. Wang and R.H. Tedford
Caninae
Number of species
Borophaginae
Hesperocyoninae
Arctoidea
Miacidae
Fig. 1.1. Range of distribution, diversity through time and relationships among the
three subfamilies of Canidae. The horizontal width of each subfamily range
roughly reflects the total number of species at a given time, as indicated by scale
bar above Hesperocyoninae. That for the family Miacidae is not meant to be
strictly proportional to the species diversity.
recognized by Tedford (1978), each represented by a distinct subfamily (Fig. 1.1).

Two archaic subfamilies, Hesperocyoninae and Borophaginae, thrived in the
middle to late Cenozoic from about 40 to 2 million years ago (Ma) (Wang, 1994;
Wang et al., 1999). Living canids all belong to the final radiation, subfamily
Caninae, which had almost an equally long history as the hesperocyonines and
borophagines but achieved their present diversity only in the last few million years
(Tedford et al., 1995).
Canids originated more than 40 Ma in the late Eocene of North America
from a group of archaic carnivorans, the Miacidae (Wang and Tedford, 1994,
1996). They were confined to the North American continent during much of their
early history, playing a wide range of predatory roles that encompassed those of
the living canids, procyonids (raccoons), hyaenids (hyenas) and possibly felids
(cats). By the latest Miocene (about 7–8 Ma), members of the subfamily Caninae
were finally able to cross the Bering Strait to reach Europe (Crusafont-Pairó,
1950), commencing an explosive radiation and giving rise to the modern canids
of the Old World. At the formation of the Isthmus of Panama about 3 Ma, canids
Evolutionary History of Canids 5
arrived in South America and quickly established themselves as one of the most
diverse groups of predators on the continent (Berta, 1987, 1988). With the aid of
humans, Canis lupus dingo was transported to Australia late in the Holocene. Since
that time, canids have become truly worldwide predators, unsurpassed in distri-
bution by any other group of carnivorans.
Here, in the context of this volume on domestic dogs, we place more empha-
sis on the subfamily Caninae, to place the origin of the wolves in its proper his-
torical context. The issues of dog domestication will be treated separately
(Chapter 2 in this volume). Our perspectives are mostly palaeontological and
morphological, although we will point out controversies from molecular studies.
We do not attempt to cite all of the references in canid palaeontology and sys-
tematics, most of which have been summarized in papers that are cited.
What is a Canid?
Canids are members of the order Carnivora due to their common possession of a
pair of carnassial teeth. The carnassials are formed by the upper fourth premolar
and lower first molar, which have long, sharp shearing blades and function as a
pair of scissors for slicing muscles and tendons of the prey. Within the Carnivora,
canids fall in the suborder Caniformia, or dog-like forms. The Caniformia are
divided into two major groups that have a sister relationship: superfamily
Cynoidea, which includes Canidae, and superfamily Arctoidea, which includes
the Ursidae, Ailuridae, Procyonidae and Mustelidae, as well as the aquatic
Pinnipedia and the extinct Amphicyonidae.
The family Canidae is a group of carnivorans that originated from a
common ancestor more than 40 Ma. Through such a common evolutionary
ancestry all members of the Canidae share a few morphological features (shared
derived characters) that are passed to all their descendants, although some of
these features have been modified subsequently in different ways. All of these
derived characters can be observed in the fossil record, and thus are capable of
being verified time and again throughout their evolutionary history.
As a cohesive group of carnivorans, living canids are easily distinguished
from other carnivoran families. Morphologically there is little difficulty in recog-
nizing living canids with their relatively uniform and unspecialized dentitions.
However, the canids as exemplified by the living forms are narrowly defined.
Only a small fraction of a once diverse group has survived to the present day (Fig.
1.1). Canids in the past had departed from this conservative pattern sufficiently
that palaeontologists had misjudged some canids as procyonids. Similarly the
extinct bear-dog family Amphicyonidae, which belongs to the Arctoidea, had in
the past been placed within the Canidae, because of its unspecialized dentition.
How do we know a canid when we see one? A key region of the anatomy
used to define canids is the middle ear region, an area in the back of the skull that
displays a rich variety of morphological patterns (Hunt, 1974). In particular, the
way the middle ear bullar chamber (a rounded bony housing that protects the
Fig. 1.2. Ventrolateral view of basicranial morphology of a primitive canid,

Hesperocyon gregarius, showing bullar composition and position of the internal
carotid artery (see text for explanations). The ventral floor of the bulla is dissected
away (isolated oval piece on top) to reveal the middle ear structures inside the
bulla and the internal septum. Modified from Wang and Tedford (1994, Fig. 1).
inner ear) ossifies is of considerable importance in recognizing relationships

among different families of carnivorans. Canids are characterized by an inflated
entotympanic bulla that is divided by a partial septum along the suture between
entotympanic and ectotympanic bones that form the bulla enclosing the floor of
the middle ear space (Fig. 1.2). Other features characteristic of canids are the loss
of a stapedial artery and the medial position of the internal carotid artery that is
situated between the entotympanic and the bone that contains the inner ear (the
petrosal) for most of its course. Anteriorly the internal carotid is contained within
the rostral entotympanic just before entering the braincase (Wang and Tedford,
1994). These basicranial characteristics have remained more or less stable
throughout the history of canids, allowing easy identification in the fossil record
when these structures are preserved.
Evolutionary History
Among the living families in the order Carnivora, the Canidae are the most
ancient. The family arose in the late Eocene, when no other living families of car-
nivorans had yet emerged (two archaic families, Miacidae and Viverravidae, have
a much older history but none survive to the present time). Furthermore,
canids still maintain many features that are primitive among all carnivorans, to
the extent that dog skulls are often used to illustrate a generalized mammal in
zoological classrooms. Dentally, canids are closest to the ancestral morphotype of

Carnivora. Canids have a relatively unreduced dental formula of 3142/3143
(numbers in sequence represent incisors, canines, premolars and molars in the
upper (left half before the oblique) and the lower (right half after the oblique)
teeth). These are relatively unmodified tribosphenic molars except for the mor-
phology of the carnassials (P4, m1), which are typical of all carnivorans. In con-
trast, all other carnivoran families generally have a more reduced dental formula
and highly modified cusp patterns.
From this mesocarnivorous (moderately carnivorous) conservative plan,
canids generally evolved toward a hypercarnivorous (highly carnivorous) or
hypocarnivorous (slightly carnivorous) dental pattern. In the hypercarnivorous
pattern (Fig. 1.3B, D) there is a general tendency for the size of the carnassial pair
to be enlarged at the expense of the molars behind (see also Enhydrocyon, Aelurodon,
Borophagus and Cuon in Fig. 1.4). This modification increases the efficiency of car-
nassial shear. A hypocarnivorous pattern (Fig. 1.3A, C) is the opposite, with devel-
opment of the grinding part of the dentition (molars) at the expense of carnassial
shear (see also Cynarctoides, Phlaocyon and Cynarctus in Fig. 1.4). This configuration
was only possible in the sister-taxa Borophaginae and Caninae, which share a
bicuspid m1 talonid (Fig. 1.3C). One of the major trends in canid evolution is the
repeated development of hyper- and hypocarnivorous forms (see below).
Hesperocyoninae
The subfamily Hesperocyoninae is the first major clade (a clade refers to a natural
group of organisms that share a common ancestry) with a total of 28 species (Fig.
1.4). Its earliest members are species of the small fox-like form, Hesperocyon, that
first appeared in the late Eocene (40–37 Ma) (Bryant, 1992) and became abun-
dant in the latest Eocene. By the Oligocene (34–30 Ma), early members of four
small clades of the hesperocyonines had emerged: Paraenhydrocyon, Enhydrocyon,
Osbornodon and Ectopocynus. Hesperocyonines experienced their maximum diver-
sity of 14 species during the late Oligocene (30–28 Ma), and reached their peak
predatory adaptations (hypercarnivory) in the earliest Miocene with advanced
species of Enhydrocyon and Paraenhydrocyon. The last species of the subfamily,
Osbornodon fricki, became extinct in the middle Miocene (15 Ma), reaching the size
of a small wolf.
With the exception of the Osbornodon clade, which acquired a bicuspid m1
talonid, hesperocyonines are primitively hypercarnivorous in dental adaptations
with tendencies toward reduced last molars and trenchant (single cusped) talonid
heels on the lower first molar. Although never reaching the extremes seen in the
borophagines (see below), hesperocyonines had modest development of bone-
cracking adaptations in their strong premolars. At least three lineages, all species
of Enhydrocyon and terminal species of Osbornodon and Ectopocynus, have independ-
ently evolved their own unique array of bone-cracking teeth. Hesperocyonines
did not experiment with hypocarnivory.
Fig. 1.3. Hypercarnivorous (b, Aelurodon and d, Euoplocyon) and

hypocarnivorous (a, Phlaocyon and c, Cynarctus) dentitions. In hypercarnivorous
forms, the upper cheek teeth (B) tend to emphasize the shearing part of the
dentition with an elongated and narrow P4, an enlarged parastyle on a
transversely elongated M1, and a reduced M2. On the lower teeth (D),
hypercarnivory is exemplified by a trenchant talonid due to the increased size and
height of the hypoconid at the expense of the entoconid (reduced to a narrow and
low ridge), accompanied by the enlargement of the protoconid at the expense of
the metaconid (completely lost in Euoplocyon) and the elongation of the trigonid
at the expense of the talonid. In hypocarnivorous forms, on the other hand, the
upper teeth (A) emphasize the grinding part of the dentition with a shortened and
broadened P4 (sometimes with a hypocone along the lingual border), a reduced
parastyle on a quadrate M1 that has additional cusps (e.g. a conical hypocone
along the internal cingulum) and cuspules, and an enlarged M2. The lower teeth
(C) in hypocarnivorous forms possess a basined (bicuspid) talonid on m1 enclosed
on either side by the hypoconid and entoconid that are approximately equal in
size. Other signs of hypocarnivory on the lower teeth include widened lower
molars, enlarged metaconids, and additional cuspules such as a protostylid.
Borophaginae
From the primitive condition of a trenchant talonid heel on the lower first molar
seen in the hesperocyonines, borophagines and canines shared a basined (bicus-
pid) talonid acquired at the very beginning of their common ancestry (Fig. 1.3C).
Along with a more quadrate upper first molar with its hypocone, the basined
talonid establishes an ancestral state from which all subsequent forms were
Vu
Cuon
lp
Canis
es
Ca
cla
ni
S.
Borophagus
s
de
Am
Urocyon
cla
.c
de
an
ine
Epicyon
s
Cerdocyon
Carpocyon Vulpes
Cynarctus
Aelurodon
Paratomarctus Eucyon
Evolutionary History of Canids

Osbornodon
Euoplocyon
i
lpin
ae
e Vu
i
in
gin
an
Enhydrocyon Phlaocyon
Trib
ha
C
e
rop
ib
Tr
Cynarctoides Desmocyon
Bo
Paraenhydrocyon
Leptocyon
Cynodesmus
e
ina
Ma
Hespe Can Ma
rocyon
inae
Hesperocyon
Fig. 1.4. Dental evolution of representative canids as shown in upper cheek teeth (P4–M2). Generally the most advanced species in each
genus is chosen to enhance a sense of dental diversity. Species in the Hesperocyoninae are: Hesperocyon gregarius; Paraenhydrocyon
josephi; Cynodesmus martini; Enhydrocyon crassidens; and Osbornodon fricki. Species in the Borophaginae are: Cynarctoides acridens;
Phlaocyon marslandensis; Desmocyon thomsoni; Cynarctus crucidens; Euoplocyon brachygnathus; Aelurodon stirtoni; Paratomarctus
9
temerarius; Carpocyon webbi; Epicyon haydeni; and Borophagus diversidens. Species in the Caninae are: Leptocyon gregorii; Vulpes
stenognathus; Urocyon minicephalus; Cerdocyon thous; Eucyon davisi; Canis dirus; and Cuon alpinus. All teeth are scaled to be
proportional to their sizes.
derived. Such a dental pattern proved to be very versatile and can readily be
adapted toward either a hyper- or hypocarnivorous type of dentition, both of
which were repeatedly employed by both borophagines and canines (Fig. 1.4).
The history of the borophagines also begins with a small fox-like form,
Archaeocyon, in the late Oligocene. Contemporaneous with larger and more preda-
tory hesperocyonines, these early borophagines in the late Oligocene and early
Miocene tended to be more omnivorous (hypocarnivorous) in their dental adap-
tations, such as Oxetocyon, Otarocyon and Phlaocyon. One extreme case, Cynarctoides
evolved selenodont-like molars as in modern artiodactyles, a rare occurrence of
herbivory among carnivorans. These early borophagines are generally no larger
than a raccoon, which is probably a good ecological model for some
borophagines at a time when procyonids had yet to diversify.
After some transitional forms in the early Miocene, such as Cormocyon and
Desmocyon, borophagines achieved their maximum ecological and numerical (i.e.
species) diversity in the middle Miocene, with highly omnivorous forms, such as
Cynarctus, that were almost ursid-like, as well as highly predatory forms, such as
Aelurodon, that were a larger version of the living African hunting dog Lycaon. By
then, borophagines had acquired their unique characteristics of a broad muzzle,
a bony contact between premaxillary and frontal, multicuspid incisors, and an
enlarged parastyle on the upper carnassials (modified from an enlargement of the
anterior cingulum).
By the end of the Miocene, borophagines had evolved another lineage of
omnivores, although only modestly in that direction, in the form of Carpocyon.
Species of Carpocyon are mostly the size of jackals to small wolves. At the same
time, the emergence of the genus Epicyon from a Carpocyon-like ancestor marked
another major clade of hypercarnivorous borophagines. The terminal species of
Epicyon, E. haydeni, reached the size of a large bear and holds the record as the
largest canid ever to have lived. Closely related to Epicyon is Borophagus, the termi-
nal genus of the Borophaginae. Both Epicyon and Borophagus are best known for
their massive P4 and p4 in contrast to the diminutive premolars in front. This pair
of enlarged premolars is designed for cracking bones, mirroring similar adapta-
tions by hyaenids in the Old World. Advanced species of Borophagus survived most
of the Pliocene but became extinct near the beginning of the Pleistocene.
Caninae
As in the hesperocyonines and borophagines, a small fox-sized species of Leptocyon

is the earliest recognized member of the subfamily Caninae. Besides sharing a
bicuspid talonid of m1 and a quadrate M1 with the borophagines, Leptocyon is also
characterized by a slender rostrum and elongated lower jaw, and correspondingly
narrow and slim premolars, features that are inherited in all subsequent canines.
It first appeared in the early Oligocene and persisted into the late Miocene.
Throughout its long existence (no other canid genus had as long a duration),
facing intense competition from the larger and diverse hesperocyonines and
borophagines, Leptocyon generally remains small and inconspicuous, never having

more than two or three species at a time.
By the latest Miocene, fox-sized niches are widely available in North
America, left open by extinctions of all small borophagines. The true fox clade,
tribe Vulpini, emerges at this time and undergoes a modest diversification to ini-
tiate primitive species of both Vulpes and Urocyon (and their extinct relatives). The
North American Pliocene record of Vulpes is quite poor. Fragmentary materials
from early Blancan indicate the presence of a swift fox-like form in the Great
Plains. Vulpes species were widespread and diverse in Eurasia during the Pliocene
(see Qiu and Tedford, 1990), resulting from an immigration event independent
from that of the Canis clade. Red fox (Vulpes vulpes) and Arctic fox (Vulpes lagopus)
appeared in North America only in the late Pleistocene, evidently as a result of
immigration back to the New World.
Preferring more wooded areas, the grey fox Urocyon has remained in south-
ern North America and Middle America. Records of the grey fox clade indicate
a more or less continuous presence in North America throughout its existence,
with intermediate forms leading to the living species U. cinereoargenteus.
Morphologically, the living African bat-eared fox Otocyon is closest to the Urocyon
clade, although molecular evidence suggests that the bat-eared fox may lie at the
base of the fox clade or even lower (Geffen et al., 1992; Wayne et al., 1997). If the
morphological evidence has been correctly interpreted, then the bat-eared fox
must represent a Pliocene immigration event to the Old World independent of
other foxes. A transitional form, Protocyon, occurs in southern Asia and Africa in
the early Pleistocene.
Advanced members of the Caninae, tribe Canini, first occur in the middle
Miocene (9–12 Ma) in the form of a transitional taxon Eucyon. As a jackal-sized
canid, Eucyon is mostly distinguished from the Vulpini in an expanded paroccipi-
tal process and enlarged mastoid process, and in the consistent presence of a
frontal sinus. The latter character initiates a series of transformations in the Tribe
Canini culminating in the elaborate development of the sinuses and a domed skull
in Canis lupus. By latest Miocene time, species of Eucyon have appeared in Europe
(Rook, 1992) and by the early Pliocene in Asia (Tedford and Qiu, 1996). The
North American records all pre-date the European ones, suggesting a westward
dispersal of this form.
Arising from about the same phylogenetic level as Eucyon is the South
American clade (subtribe Cerdocyonina). Morphological and molecular evidence
generally agrees that living South American canids, the most diverse group of
canids on a single continent, belong to a natural group of their own. The South
American canids are united by morphological characters such as a long palate, a
large angular process of the jaw with a widened scar for attachment of the infe-
rior branch of the medial pterygoid muscle, and a relatively long base of the coro-
noid process (Tedford et al., 1995). By the close of the Miocene, certain
fragmentary materials from southern United States and Mexico indicate that taxa
assignable to Cerdocyon (Torres and Ferrusquía-Villafranca, 1981) and Chrysocyon
occur in North America. The presence of these advanced taxa in the North
American late Miocene predicts that ancestral stocks of many of the South
American canids may have been present in southern North America or Middle
America. They appear in the South American fossil record shortly after the for-
mation of the Isthmus of Panama in the Pliocene, around 3 Ma (Berta, 1987). The
earliest records are Pseudalopex and its close relative Protocyon, an extinct large
hypercarnivore, from the Plio-Pleistocene (around 2.5–1.5 Ma) of Argentina. By
the latest Pleistocene (50,000–10,000 years ago), most living species or their close
relatives had emerged, along with the extinct North American dire wolf, Canis
dirus. By the end of the Pleistocene, all large, hypercarnivorous canids of South
America (Protocyon, Theriodictis) as well as Canis dirus had become extinct.
The Canis clade within the tribe Canini, the most advanced group in terms
of large size and hypercarnivory, arises near the Miocene–Pliocene boundary
between 5 and 6 Ma in North America. A series of jackal-sized ancestral species
of Canis thrived in the early Pliocene, such as C. ferox, C. lepophagus and other unde-
scribed species. At about the same time, the first records of canids begin to appear
in the European late Neogene: Canis cipio in the late Miocene of Spain (Crusafont-
Pairó, 1950), Eucyon monticinensis in the latest Miocene of Italy (Rook, 1992), the
earliest raccoon-dog Nyctereutes donnezani and the jackal-sized Canis adoxus in the
early Pliocene of France (Martin, 1973; Ginsburg, 1999). The enigmatic C. cipio,
only represented by parts of the upper and lower dentition, may pertain to a form
at the Eucyon level of differentiation rather than truly a species of Canis.
The next phase of Canis evolution is difficult to track. The newly arrived Canis
in Eurasia underwent an extensive radiation and range expansion in the late
Pliocene and Pleistocene, resulting in multiple, closely related species in Europe,
Africa and Asia. To compound this problem, the highly cursorial wolf-like Canis
species apparently belong to a circum-arctic fauna that undergoes expansions and
contractions with the fluctuating climate. Hypercarnivorous adaptations are
common in the crown-group of species, especially in the Eurasian middle lati-
tudes and Africa. For the first time in canid history, phylogenetic studies cannot
be satisfactorily performed on forms from any single continent because of their
Holarctic distribution and faunal intermingling between the New and Old
Worlds. Nevertheless some clades were localized in different parts of Holarctica.
The vulpines’ major centre of radiation was in the Old World. For the canines,
North America remained a centre through the Pliocene producing the coyote as
an endemic form. A larger radiation yielding the wolves, dhole, African hunting
dog and fossil relatives took place on the Eurasian and African continents. During
the Pleistocene elements of the larger canid fauna invaded mid-latitude North
America – the last invasion of which was the appearance of the grey wolf south
of the glacial ice sheets in the latest Pleistocene (about 100,000 years ago).
Phylogenetic Relationships
As mentioned above, there is strong fossil evidence about the antiquity of

the family Canidae. This basal placement within the suborder Caniformia is
increasingly born out by molecular data in recent years, such as DNA–DNA

hybridization of single copy DNA, mitochondrial DNA sequence studies, and
recently studies of DNA sequences from nuclear genes (Vrana et al., 1994; Slattery
and Brien, 1995; Flynn and Nedbal, 1998; Murphy et al., 2001; Wang et al.,
2004a), although molecular clock calculations tend to place the divergence time
somewhat older, around 50 Ma (Wayne et al., 1989), than is estimated from fossil
evidence.
Phylogenetic (genealogical) relationships are traditionally inferred by analy-
sis of the morphological characters, but molecular data are increasingly playing
important, sometimes controversial, roles in the detection of evolutionary rela-
tionships. However, in the case of canids that have a substantial history known by
fossil records only, morphology is still the only way to allow a comprehensive view
of their entire history.
For the two extinct subfamilies, Hesperocyoninae and Borophaginae, we
have performed an exhaustive analysis of the entire fossil records (Wang, 1994;
Wang et al., 1999) and their relationships are roughly shown in Fig. 1.4. For the
subfamily Caninae, we have nearly finished a similar study of monographic revi-
sions that deals with the entire fossil history of the canines in North America
(Tedford et al., in prep.). As a part of this larger effort to lay down a phylogenetic
framework, Tedford et al. (1995) performed a cladistic analysis of living canids on
morphological grounds. The result is a nearly fully resolved relationship based on
an 18 taxa by 57 characters matrix at the generic level. This relationship recog-
nizes three monophyletic clades in the canines: the fox group (tribe Vulpini), the
South American canine group, and the wolf group containing hypercarnivorous
forms (the latter two form the tribe Canini) (Fig. 1.5, left).
Molecular studies of canid relationships range from investigations in com-
parative karyology, allozyme electrophoresis, mitochondrial DNA, to microsatel-
lite loci (Wayne and Brien, 1987; Wayne et al., 1987a, b, 1997; Geffen et al., 1992;
Bruford and Wayne, 1993; Girman et al., 1993; Gottelli et al., 1994; Vilà et al.,
1997, 1999). Trees derived from the mtDNA are based on the widest possible
sample of taxa and are better studied than nuclear DNA (Fig. 1.5, right) (e.g.
Wayne et al., 1997). Overall, molecular studies tend to place the foxes near the
basal part, the South American canines in the middle, and the wolves and
hunting dogs toward the terminal branches, a pattern that is consistent with the
morphological tree. The detailed arrangements, however, differ in a number of
ways. The foxes are generally in a paraphyletic arrangement (falling at the stem
parts of the tree) in contrast to a monophyletic clade (a natural group that con-
tains ancestors and all descendants) in the morphological tree. The grey fox and
bat-eared fox are placed at the base despite their highly advanced dental mor-
phology compared to other foxes. Similarly, South American canines are no
longer monophyletic under molecular analysis but form at least two paraphyletic
branches. A glaring discrepancy is the Asiatic raccoon dog being allied to the
foxes in the molecular analysis despite its numerous morphological characters
shared with some South American forms. Finally, molecular data suggest inde-
pendent origins for the Asiatic and African hunting dogs in contrast to a sister
14
Vulpes Urocyon
Urocyon Otocyon
Otocyon Nyctereutes
‘Pseudalopex’ Vulpes
Pseudalopex ‘Pseudalopex’
Dusicyon Pseudalopex
X. Wang and R.H. Tedford

Lycalopex Dusicyon
Chrysocyon Lycalopex
Cerdocyon Atelocynus
Nyctereutes Cerdocyon
Atelocynus Chrysocyon
Speothos Speothos
Canis Canis
Cuon Cuon
Lycaon Lycaon
Fig. 1.5. Contrasting canine relationships from recent morphological (left) (Tedford et al., 1995) and molecular studies (right) (Wayne et
al., 1997).
relationship in the morphological tree supported by a large number of characters

related to hypercarnivory.
Not surprisingly, there are increased agreements between the molecular and
morphological results when the two data sets are combined in a total evidence
analysis (Wayne et al., 1997; Zrzavý and Ricánková, 2004). Under such condi-
tions, the South American canines (except Nyctereutes) become monophyletic, as
does the clade including the wolf, dhole and African hunting dog. Although mito-
chondrial genes are known to evolve relatively quickly and thus are suitable for
analysis on groups that have more recent ancestry, genetic variations in the mito-
chondria are only a tiny fraction of the total genome and the gene trees may not
reveal the true phylogeny. Nuclear genes, with their vast information content,
have the potential of revealing the true relationships, especially for groups of
longer history. Works are underway to search for the most suitable part of the
genome with just the right rate of evolution, neither too slow to offer much insight
about change nor too fast to obscure true relationships. Recently Selenocysteine
tRNA (Cf TRSP) and RNase P RNA (Cf RPPH1) genes have been shown to be
promising (Bardeleben et al., 2005), these tend to place the South American forms
in a clade and are consistent with morphological results.
Evolutionary Trends
As a very successful group of predators, canids are known for their outstanding
cursoriality, the ability to run fast and over long distances, and for their social
(pack) hunting that requires complex collaborative behaviours. This combination
of long-distance, relay-style running and social hunting to bring down prey
together is a successful strategy in catching larger prey. In this regard, only the
hyaenids are comparably equipped, whereas the felids may have reached similar
running capabilities but more often hunt alone. The increased cursoriality in
canids is generally correlated with a similarly increased running ability by their
prey (various groups of ungulates), which are in turn related to the progressively
more open grassland environments in the late Cenozoic.
Running and posture
Increased ability of running is often manifested in elongated and slender limbs,

reduction of digits, and more erect standing posture. In addition to the overall
elongation of the limbs, the distal segments (radius–ulna, tibia–fibular and
metapodials) in particular tend to elongate more than the proximal segments
(humerus and femur). The reduction of digits in carnivorans is usually confined to
the decline of the digit I to result in a functionally four-digit hand and foot, in con-
trast to the far more dramatic digit reduction in ungulates such as the horses,
because of the need for grasping by the hands in carnivorans. In all cursorial car-
nivorans, the standing posture is usually digitigrade with the metapodials lifted
from the ground in contrast to the primitively plantigrade posture with proximal
ends of the metapodials still touching the ground. All of the above limb modifi-
cations are a common strategy to increase the length of stride and decrease the
weight in distal portions of the limbs.
Although fossil postcranial skeletons are often less abundant than cranial and
dental materials, particularly those that are associated with dental materials to be
accurately identified in taxonomy, we are in possession of enough skeletons in a
few taxa in most major clades to permit a reasonable grasp of the general evolu-
tionary trends. Starting from the beginning of the Canidae, Hesperocyon has
already exhibited an initial stage of cursoriality with moderate lengthening of
limbs and a semi-digitigrade posture (Wang, 1993). Within the subfamily
Caninae, this trend steadily progressed from the basal genus Leptocyon, which has
increasingly slender limbs, to Vulpes, which has lost the entepicondylar foramen
on the distal humerus (a sign of further slendering) and substantially reduced digit
I, and to the Canis clade that carries these trends to a greater extreme as shown
by living canids (Tedford et al., 1995).
Social hunting
Hunting behaviour is generally not preserved in fossil records and canids’ pack
hunting behaviour can only be approached in an indirect way. By correlating
certain skull morphology and body size, Van Valkenburgh et al. (2003) suggested
that some hypercarnivorous borophagines may have acquired social hunting
skills, although likely independently from those in the canines. Social hunting in
the Caninae is mostly confined to the Canis clade (Macdonald et al., 2004) and
thus has presumably arisen in the last few million years of canid history (see earlier
section on Caninae history).
Diet and body size
Finally, throughout their history canids displayed remarkable flexibility and

diversity in adaptations to different diets and preys. Small, generalized canids
have the ability to evolve toward either hypocarnivorous or hypercarnivorous
directions, depending on opportunities (see Fig. 1.3 for dental adaptations).
Hypocarnivorous forms tend to remain small- to medium-sized, and sometimes
reduce their body size. The hypercarnivorous direction, however, often leads to
larger body size, possibly as a result of energetic requirements (Carbone et al.,
1999). As in the history of hyaenids, large, hypercarnivorous canids frequently
developed strong premolars that are capable of cracking bones as an additional
source of protein (Werdelin, 1989). Van Valkenburgh et al. (2004) suggested that
such a correlated increase in body size and hypercarnivory acts like an evolu-
tionary ratchet – once a certain lineage is launched toward increased predation,
a larger body size often evolves over time.
Acknowledgements
We thank Per Jensen for his invitation to write this contribution. This research is
funded in part by grants from the National Science Foundation (DEB 9420004;
9707555).
References
Bardeleben, C., Moore, R.L. and Wayne, R.K. (2005) Isolation and molecular evo-
lution of the Selenocysteine tRNA (Cf TRSP) and RNase P RNA (Cf RPPH1)
genes in the dog family, Canidae. Molecular Biology and Evolution 22,
347–359.
Berta, A. (1987) Origin, diversification, and zoogeography of the South American
Canidae. In: Patterson, B.D. and Timm, R.M. (eds) Studies in Neotropical
Mammalogy: Essays in Honor of Philip Hershkovitz. Field Museum of Natural
History, Chicago, Illinois, pp. 455–471.
Berta, A. (1988) Quaternary evolution and biogeography of the large South
American Canidae (Mammalia: Carnivora). University of California
Publications of Geological Science 132, 1–149.
Bruford, M.W. and Wayne, R.K. (1993) Microsatellites and their application to pop-
ulation genetic studies. Current Biology 3, 939–943.
Bryant, H.N. (1992) The Carnivora of the Lac Pelletier Lower Fauna (Eocene:
Duchesnean), Cypress Hills Formation, Saskatchewan. Journal of Paleontology
66, 847–855.
Carbone, C., Mace, G.M., Roberts, S.C. and Macdonald, D.W. (1999) Energetic con-
straints on the diet of terrestrial carnivores. Nature 402, 286–288.
Crusafont-Pairó, M. (1950) El primer representante del género Canis en el Pontiense
eurasiatico (Canis cipio nova sp.). Boletin de la Real Sociedad Española de
Historia Natural (Geologia) 48, 43–51.
Flynn, J.J. and Nedbal, M.A. (1998) Phylogeny of the Carnivora (Mammalia): con-
gruence vs. incompatibility among multiple data sets. Molecular Phylogenetics
and Evolution 9, 414–426.
Geffen, E., Mercure, A., Girman, D.J., Macdonald, D.W. and Wayne, R.K. (1992)
Phylogenetic relationships of the fox-like canids: mitochondrial DNA restriction
fragment, site and cytochrome b sequence analyses. Journal of Zoology 228,
27–39.
Ginsburg, L. (1999) Order Carnivora. In: Rössner, G.E. and Heissig, K. (eds) The
Miocene Land Mammals of Europe. Verlag Dr. Friedrich Pfeil, München,
Germany, pp. 109–148.
Girman, D.J., Kat, P.W., Mills, G., Ginsberg, J., Fanshaw, J., Fitzgibbon, C., Borner,
M., Wilson, V., Laurenson, K. and Wayne, R.K. (1993) A genetic and morpho-
logical analysis of the African wild dog (Lycaon pictus). Journal of Heredity 84,
450–459.
Gottelli, D., Sillero-Zubiri, C., Applebaum, G.D., Roy, M.S., Girman, D.J., Garcia-
Moreno, J., Ostrander, E.A. and Wayne, R.K. (1994) Molecular genetics of the
most endangered canid: the Ethiopian wolf, Canis simensis. Molecular Ecology
3, 277–290.
Hunt, R.M., Jr (1974) The auditory bulla in Carnivora: an anatomical basis for reap-
praisal of carnivore evolution. Journal of Morphology 143, 21–76.
Leonard, J.A., Wayne, R.K., Wheeler, J., Valadez, R., Guillén, S. and Vilà, C. (2002)
Ancient DNA evidence for old world origin of new world dogs. Science 298,
1613–1616.
Linnaeus, C. (1758) Systema naturae per regna tria naturae, secundum classes,
ordines, genera, species cum characteribus, differentiis, synonymis, locis. Editio
Decima, 1758 (12th edition of Linaeus 1758). Societatis Zoologicae
Germanicae.
Macdonald, D.W., Creel, S. and Mills, M.G.L. (2004) Canid society. In: Macdonald,
D.W. and Sillero-Zubiri, C. (eds) The Biology and Conservation of Wild Canids.
Oxford University Press, Oxford, UK, pp. 85–106.
Martin, R. (1973) Trois nouvelles espèces de Caninae (Canidae, Carnivora) des gise-
ments plio-villafranchiens d’Europe. Documents des Laboratoires de Géologie
de la Faculté des Sciences de Lyon 57, 87–96.
Murphy, W.J., Eizirik, E., O’Brien, S.J., Madsen, O., Scally, M., Douady, C.J.,
Teeling, E., Ryder, O.A., Stanhope, M.J., de Jong, W.W. and Springer, M.S.
(2001) Resolution of the early placental mammal radiation using Bayesian phy-
logenetics. Science 294, 2348–2351.
Pennisi, E. (2002) A shaggy dog history. Science 298, 1540–1542.
Qiu, Z.-X. and Tedford, R.H. (1990) A Pliocene species of Vulpes from Yushe,
Shanxi. Vertebrata PalAsiatica 28, 245–258.
Rook, L. (1992) ‘Canis’ monticinensis sp. nov., a new Canidae (Carnivora,
Mammalia) from the late Messinian of Italy. Bolletino della Società
Paleontologica Italiana 31, 151–156.
Savolainen, P., Zhang, Y.-P., Luo, J., Lundeberg, J. and Leitner, T. (2002) Genetic
evidence for an East Asian origin of domestic dogs. Science 298, 1610–1613.
Slattery, J.P. and Brien, S.J.O. (1995) Molecular phylogeny of the red panda (Ailurus
fulgens). Journal of Heredity 86, 413–422.
Tedford, R.H. (1978) History of dogs and cats: a view from the fossil record.
Nutrition and Management of Dogs and Cats. Ralston Purina Co., St Louis,
Missouri, chap. M23.
Tedford, R.H. and Qiu, Z.-X. (1996) A new canid genus from the Pliocene of Yushe,
Shanxi Province. Vertebrata PalAsiatica 34, 27–40.
Tedford, R.H., Taylor, B.E. and Wang, X. (1995) Phylogeny of the Caninae
(Carnivora: Canidae): the living taxa. American Museum Novitates 3146,
1–37.
Tedford, R.H., Wang, X. and Taylor, B.E. (in prep.) Phylogenetic systematics of the
North American fossil Caninae (Carnivora: Canidae). Bulletin of the American
Museum of Natural History.
Torres, R.V. and Ferrusquía-Villafranca, I. (1981) Cerdocyon sp. nov. A (Mammalia,
Carnivora) en Mexico y su significacion evolutiva y zoogeografica en relacion
a los canidos sudamericanos. In: Anais II Congresso Latino-Americano de
Paleontologia, Porto Alegre, Brazil, pp. 709–719.
Van Valkenburgh, B., Sacco, T. and Wang, X. (2003) Pack hunting in Miocene
borophagine dogs: evidence from craniodental morphology and body size. In:
Flynn, L.J. (ed.) Vertebrate Fossils and Their Context: Contributions in Honor of
Richard H. Tedford. Bulletin of the American Museum of Natural History.
American Museum of Natural History, New York, pp. 147–162.
Van Valkenburgh, B., Wang, X. and Damuth, J. (2004) Cope’s Rule, hypercarnivory,
and extinction in North American Canids. Science 306, 101–104.
Vilà, C., Savolainen, P., Maldonado, J.E., Amorim, I.R., Rice, J.E., Honeycutt, R.L.,
Crandall, K.A., Lundeberg, J. and Wayne, R.K. (1997) The domestic dog has an
ancient and genetically diverse origin. Science 276, 1687–1689.
Vilà, C., Amorim, I.R., Leonard, J.A., Posada, D., Castroviejo, J., Petrucci-Fonseca,
F., Crandall, K.A., Ellegren, H. and Wayne, R.K. (1999) Mitochondrial DNA
phylogeography and population history of the grey wolf Canis lupus. Molecular
Ecology 8, 2089–2103.
Vrana, P.B., Milinkovitich, M.C., Powell, J.R. and Wheeler, W.C. (1994) Higher
level relationships of the arctoid Carnivora based on sequence data and ‘total
evidence’. Molecular Phylogenetics and Evolution 3, 47–58.
Wang, X. (1993) Transformation from plantigrady to digitigrady: functional mor-
phology of locomotion in Hesperocyon (Canidae: Carnivora). American
Museum Novitates 3069, 1–23.
Wang, X. (1994) Phylogenetic systematics of the Hesperocyoninae (Carnivora:
Canidae). Bulletin of the American Museum of Natural History 221, 1–207.
Wang, X. and Tedford, R.H. (1994) Basicranial anatomy and phylogeny of primitive
canids and closely related miacids (Carnivora: Mammalia). American Museum
Novitates 3092, 1–34.
Wang, X. and Tedford, R.H. (1996) Canidae. In: Prothero, D.R. and Emry, R.J. (eds)
The Terrestrial Eocene–Oligocene Transition in North America, Pt. II: Common
Vertebrates of the White River Chronofauna. Cambridge University Press,
Cambridge, pp. 433–452.
Wang, X., Tedford, R.H. and Taylor, B.E. (1999) Phylogenetic systematics of the
Borophaginae (Carnivora: Canidae). Bulletin of the American Museum of
Natural History 243, 1–391.
Wang, X., Tedford, R.H., Van Valkenburgh, B. and Wayne, R.K. (2004a) Ancestry:
Evolutionary history, molecular systematics, and evolutionary ecology of
Canidae. In: Macdonald, D.W. and Sillero-Zubiri, C. (eds) The Biology and
Conservation of Wild Canids. Oxford University Press, Oxford, UK, pp.
39–54.
Wang, X., Tedford, R.H., Van Valkenburgh, B. and Wayne, R.K. (2004b) Phylogeny,
classification, and evolutionary ecology of the Canidae. In: Sillero-Zubiri, C.,
Hoffmann, M. and Macdonald, D.W. (eds) Canids: Foxes, Wolves, Jackals and
Dogs. Status Survey and Conservation Action Plan. IUCN/SSC Canid Specialist
Group, The World Conservation Union, Gland, Switzerland, pp. 8–20.
Wayne, R.K. and Brien, S.J.O. (1987) Allozyme divergence within the Canidae.
Systematic Zoology 36, 339–355.
Wayne, R.K., Nash, W.G. and Brien, S.J.O. (1987a) Chromosomal evolution of the
Canidae. I. Species with high diploid numbers. Cytogenetics and Cell Genetics
44, 123–133.
Wayne, R.K., Nash, W.G. and Brien, S.J.O. (1987b) Chromosomal evolution of the
Canidae. II. Divergence from the primitive carnivore karyotype. Cytogenetics
and Cell Genetics 44, 134–141.
Wayne, R.K., Benveniste, R.E., Janczewski, D.N. and O’Brien, S.J. (1989) Molecular
and biochemical evolution of the Carnivora. In: Gittleman, J.L. (ed.) Carnivore
Behavior, Ecology, and Evolution. Cornell University Press, Ithaca, New York,
pp. 465–494.
Wayne, R.K., Geffen, E., Girman, D.J., Koepfli, K.-P., Lau, L.M. and Marshall, C.R.
(1997) Molecular systematics of the Canidae. Systematic Zoology 46, 622–653.
Werdelin, L. (1989) Constraint and adaptation in the bone-cracking canid
Osteoborus (Mammalia: Canidae). Paleobiology 15, 387–401.
Zrzavý, J. and Ricánková, V. (2004) Phylogeny of recent Canidae (Mammalia,
Carnivora): relative reliability and utility of morphological and molecular
datasets. Zoologica Scripta 33, 311–333.
Domestication of Dogs
2
Peter Savolainen
Introduction
The dog is perhaps the most fascinating of the domestic animals. According to the
available knowledge, it was the first domestic animal, and its wild ancestor, the
wolf, was probably domesticated by mobile hunter-gatherers rather than by
settled farmers, contrary to probably all other domestic animals (Clutton-Brock,
1995). The dog is morphologically very diverse, both in size and shape. In the
range of size it is the most diverse mammal species, with the close to 100-fold dif-
ference in weight between the chihuahua and the great dane being the most
prominent example. Dogs and humans share a number of social and behavioural
signals facilitating communication. Many of these are shared with the wolf and
have undoubtedly had significance in the domestication of the wolf, facilitating
the first contacts, but have also been further selected during the process of domes-
tication (Houpt and Willis, 2001). This gives the dog a behaviour which appeals
to many of us more than that of other domestic animals, explaining its status as
man’s best friend.
Despite this special status, until recently surprisingly little has been known
about the history of the dog. While the origin from the wolf has been the gener-
ally accepted theory, there have been virtually no other details known about the
first origin of the dog, or about how the immense morphological variation,
and the hundreds of dog breeds, have developed. Since there is great interest
in this subject, numerous texts have been written, and there have been many the-
ories about how, where and when the dog originated, but there have been
few facts to build on, making reading about this subject an often frustrating
experience.
© CAB International 2007. The Behavioural Biology of Dogs
(ed. P. Jensen) 21
22 P. Savolainen
However, in the last 10 years, genetic analyses, based on sequence analysis of

mitochondrial DNA (mtDNA), have provided a large step toward unravelling the
origin and early history of the domestic dog. Thus, the origin of the dog from the
wolf has been shown with great certainty, and for the first time there is relatively
firm proof to suggest a geographic origin of the dog. While more studies are
needed to definitely establish the facts, the sum of the available evidence indicates
that the domestic dog has a single origin from East Asian wolves, some 15,000
years ago.
Origin from the Wolf
There is now overwhelming evidence that the wild ancestor of the domestic dog
is the wolf. An origin primarily from the wolf has been the main theory from the
start of evolutionary theory building, but it has been hypothesized that also other
canids, such as the jackal or coyote, could have contributed to the forming of the
dog, explaining some of the morphological and behavioural variation among
domestic dogs. However, with time, numerous studies of physical and behav-
ioural characteristics, and later also molecular genetic markers, both nuclear
markers and mtDNA, have built an increasingly stronger case for the wolf origin
(Wayne, 1993; Clutton-Brock, 1995). While not disproving some degree of con-
tribution from other species, the total collection of evidence is strong in favour of
the wolf as the only ancestor of the dog. The evidence from mtDNA gives a con-
sistent picture: among more than 1000 analysed dogs, all have mtDNA sequences
which are much closer to those of wolves than of the other wild canids. However,
mtDNA is maternally inherited and can therefore only describe the history of
females, and while studies of nuclear markers consistently support an origin from
the wolf, studies have so far been relatively limited, leaving a possibility for a small
contribution on the male side from, for example, the jackal. Large-scale popula-
tion studies of nuclear markers, for example the Y chromosome, will surely give
the final answer within a few years.
The Archaeological Evidence
Archaeological finds related to dogs have been found at Mesolithic sites in

Europe, Asia and America, the earliest going back to at least 14,000 BP (years
before present), indicating that the dog was the first domestic animal, several
thousand years ahead of the first farm animals (Clutton-Brock, 1995). The earli-
est finds believed to be from domestic dogs are from Central and Eastern Europe
and the Middle East: a single jaw from 14,000 BP in Germany (Nobis, 1979), two
skulls from 13,000–17,000 BP in western Russia (Sablin and Khlopachev, 2002)
and an assemblage of small canids from 12,000 BP in the Middle East (Tchernov
and Valla, 1997). The earliest finds from East Asia and North America are
younger, from 7500 (Li, 1990) and 8500 BP (Clutton-Brock and Noe-Nygaard,
Domestication 23
1990), respectively. However, the archaeological evidence is problematic to inter-

pret. It is difficult to discriminate between species among canid remains, for
example between small wolves and domestic dogs (Olsen, 1985), the dating of the
finds may be obscured by, for example, disturbed layers, and archaeological exca-
vations have not been evenly performed in different parts of the world. Thus, the
fact that the oldest remains of dogs have been found in the relatively thoroughly
excavated regions of Europe and the Middle East could be attributed to a greater
chance of finding remains there than in less explored regions. The morphology
and size of early archaeological dog remains suggest an origin from one of the
small South Asian wolf subspecies rather than from the large North Eurasian and
North American wolves, and the spread of the wolf, which has been limited to
Eurasia and North America, excludes the possibility of an African origin (Clutton-
Brock, 1995). One osteological feature of the jaw, diagnostic for dogs, is found
also among Chinese wolves, but rarely in other wolves, indicating an origin from
East Asia (Olsen, 1985).
Thus, the archaeological record has not offered any distinct clues even to the
most basic historic questions, such as where, when and how many times the wolf
was domesticated into the domestic dog. Based on the archaeological evidence,
there have been two main theories for the origin of the domestic dog. The Middle
East has been proposed, based on the earliest finds of dogs being found in Europe
and the Middle East, and the small size of the local species of wolf. An origin from
several different wolf populations has been suggested, based on the widespread
occurrence of early finds of dog remains (Clutton-Brock, 1995). This could also to
some extent offer an explanation for the extreme morphological variation among
dog breeds. There is also very little known about the early history of different dog
breeds, and even for the large number of breeds which have been developed in
the last few hundred years, the exact origin is mostly unknown (Clutton-Brock,
1995).
Population Genetic Studies
Thus, for a long time research did not bring much new knowledge about the
history of the dog, except the early molecular biological studies which could bring
growing certainty that the dog originates from the wolf rather than from other
canids, and the sporadic addition of new archaeological finds of early dogs,
changing the geographic location for their oldest known remains. Since the clas-
sical instruments for this kind of question, archaeology and palaeontology, failed
to provide any detailed information, population genetic studies was applied rela-
tively soon after the techniques for large-scale DNA-sequencing studies became
available. The studies have, so far, mainly been performed on mtDNA, because
the genetic variation is higher than for nuclear DNA, especially in the noncoding
part, the control region (CR), giving much genetic information from analysis of a
relatively short DNA sequence. mtDNA is maternally inherited and a drawback
is therefore that the mtDNA-studies can only describe the history of the female
24 P. Savolainen
dogs, the history of the males being only indirectly monitored. An advantage on
the other side, shared with the Y chromosome, compared to nuclear autosomal
DNA, is that the mitochondrial genome does not recombine, which implies that
the whole molecule is inherited as an entity through the generations, making the
interpretation of the data straightforward.
The first three major population genetic studies of mtDNA from dogs and
wolves were published in 1996 and 1997: 261 base pairs of the mtDNA CR were
analysed in 140 domestic dogs representing 67 breeds, mostly European but also
a few from Asia, Africa and America, and 162 wolves from throughout Europe,
Asia and North America (Vila et al., 1997), and 970 and 670 base pairs, respec-
tively, of the CR were analysed in a total of 128 dogs, representing mainly
Japanese breeds, and 19 wolves (Okumura et al., 1996; Tsuda et al., 1997). These
studies showed that, in phylogenetic analyses, the domestic dog mtDNA
sequences were distributed into six distinct groups (clades) interspersed by wolf
sequences, showing that the dog originates from several female wolf lines, at least
as many as the number of clades (Fig. 2.1). Importantly, all of the 268 studied
domestic dogs had mtDNA sequences which were closer to those of wolves than
of other canids. Thus, there was no indication of a genetic contribution from any
other canid than the wolf to the domestic dog population. However, the geo-
graphic location of the domestication of the wolf, or any detailed conclusions
about the history of specific dog breeds, could not be determined based on these
relatively limited data sets. The time for the origin of the dog was estimated, based
on the largest genetic distance of the sequences in the largest phylogenetic clade
of dog sequences, giving a possible date of 135,000 years. However, as shown
below, the mtDNA data can be interpreted differently, resulting in a date more
in agreement with that obtained from the archaeological record.
A Single Geographic Origin
The greatest step so far in the study of the origin of dogs was taken in 2002 with
the publication of two articles, one study of the present-day dog population
throughout the world (Savolainen et al., 2002) and one of pre-European archaeo-
logical remains of American dogs (Leonard et al., 2002). Leonard et al. showed
that domestic dogs on all continents have a common origin from a single gene
pool, and the study by Savolainen et al. suggested that this gene pool originated
just once, somewhere in East Asia.
There are principally two approaches possible for elucidating the place of
domestication of the wolf, using mtDNA analysis. The most straightforward
approach would be to study the mtDNA sequences of today’s domestic dog pop-
ulation and compare these with sequences from a worldwide sample of wolves.
The population of wolves having the most similar, or even identical, sequences
compared to the dogs would then be presumed to be the wolf population from
which the domestic dogs descend, and the geographical region in which these
wolves live would be presumed to be the geographical origin of the dog. However,
Domestication 25
0.005 substitutions/site
coyote
Fig. 2.1. Phylogenetic tree of dog (unlabelled) and wolf (open squares) mtDNA
types (Savolainen et al., 2002). Six clades (A–F) of dog mtDNA types are
indicated. Branch lengths are according to the indicated scale, the branch leading
to the outgroup (coyote) has been reduced by 50%. Clades A–D and F were
imploded, and the relationships between mtDNA types within the clades are
instead shown as minimum-spanning networks in Figs 2.2 and 2.3.
there is little geographic structure in the mtDNA of wolves, probably because

wolves are extremely mobile animals. In a study of the worldwide mtDNA
sequence variation among wolves (Vila et al., 1999a) it was found, for example,
that a Bulgarian and a Saudi Arabian wolf had identical mtDNA types, and that
a Mongolian and another Saudi Arabian wolf shared another type. Thus, the
finding of a few wolf mtDNA types closely related to domestic dog types is not in
itself a sufficient basis for the determination of the location of domestication. The
wolf populations have also experienced severe reductions in population size or
26 P. Savolainen
total extinction in large parts of their original geographical distribution. There is

therefore a possibility that the wolf population from which the dogs evolved has
become extinct, rendering such studies impossible. The other approach would be
to make a study of today’s domestic dog population, making sure to obtain a rep-
resentative sample from all parts of the world, and comparing the genetic varia-
tion between geographic regions, a so-called intraspecific study of the
phylogeographic mtDNA variation among dogs. The geographic region having
the largest genetic variation would then be assumed to be the region in which the
domestic dog originated, based on the assumptions that only part of the genetic
variation in the region of origin was spread to the other parts of the world, and
that any subsequent mixing of populations through migration would not have
been thorough enough to conceal these differences.
Based on the second approach, 654 dogs sampled throughout the world and
38 Eurasian wolves were studied for 582 base pairs of the mtDNA CR
(Savolainen et al., 2002). This resulted in the first comprehensive picture of the
genetic variation among domestic dogs worldwide. Importantly, it gave a good
representation of dogs in Europe, Southwest Asia and East Asia, the three regions
that have been the strongest candidates as geographical origin of the domestic
dog (Table 2.1). A phylogenetic analysis distributed the mtDNA sequences into
the six clades found earlier among Japanese and European breeds, additional
samples from other parts of the world thus not adding more major phylogenetic
groups in the dog mtDNA phylogenetic tree (Fig. 2.1). The relations between the
mtDNA molecules at large genetic distances can be relatively truthfully depicted
in the phylogenetic tree, but the data are not sufficient for obtaining a full resolu-
tion of dog mtDNA types within the clades. Therefore, the branching in the six
dog clades was imploded, and the relations between the mtDNA types within the
clades were instead displayed as minimum-spanning networks (Fig. 2.2). The dis-
tribution of the dog mtDNA sequences into six clades interspersed by wolf
sequences shows that the dog originates from at least six female wolf lines, and
possibly many more individual female wolves having identical mtDNA types.
Whether this also indicates several different places and time-points for the origin
of dogs from the wolf will be discussed below. The six clades were represented by
very different numbers of dogs. The largest clade, A, comprised 71% of all dogs,
and the three largest clades, A, B and C, a total >95% of the dogs, compared to
3% for clade D and only three dogs each for clades E and F.
The representation of samples from most parts of the world allows a
phylogeographic study of the data. Comparing the genetic variation between
different parts of the world, the similarities in the distribution of sequences
belonging to the three major clades is striking (Table 2.1). All the three
major clades A–C are present in all parts of the world except Arctic America,
and at similar frequencies. Thus, these three clades, comprising >95% of the
dog mtDNA sequences, constitute a common source for a very large proportion
of the mtDNA genetic variation in all domestic dog populations. The absence
of clades B and C in Arctic America can probably be attributed to a genetic
bottleneck in the forming of this dog population; probably this population was
Table 2.1. Number and proportion of individuals, and number of mtDNA types and unique mtDNA types, for the phylogenetic clades
A–F and for all six clades (Total) in different populations. The dogs represent a geographical region based on that they either are of a
breed of known geographical origin or that they have been sampled in a geographical region with little import of foreign dogs. mtDNA
types are defined by substitutions only, disregarding indels.
Clade A Clade B Clade C Clade D Clade E Clade F Total
Domestication
Africa 30 (85.7) 14 (5) 4 (11.0) 1 (0) 1 (2.9) 1 (0) 0 0 0 0 0 0 35 16 (5)
Arctic America 25 (100) 5 (1) 0 0 0 0 0 0 0 0 0 0 25 5 (1)
Europe 140 (67.6) 20 (9) 36 (18.0) 4 (1) 12 (5.9) 3 (0) 19 (9.1) 3 (3) 0 0 0 0 207 30 (13)
East Asia 192 (73.8) 44 (30) 39 (15.0) 10 (7) 24 (9.2) 4 (1) 0 0 3 (1.1) 1 (1) 2 (0.8) 2 (2) 260 61 (41)
SW Asia 51 (56.7) 16 (4) 32 (35.0) 4 (2) 5 (5.4) 2 (0) 2 (2.2) 1 (1) 0 0 0 0 90 23 (7)
Siberia 17 (70.8) 9 (1) 2 (8.3) 1 (0) 4 (17.0) 2 (1) 0 0 0 0 1 (4.1) 1 (1) 24 13 (3)
India 11 (84.6) 4 (0) 1 (7.7) 1 (0) 1 (7.7) 1 (0) 0 0 0 0 0 0 13 6 (0)
Total 466 (71.3) 71 114 (17.4) 13 47 (7.2) 5 21 (3.2) 4 3 (0.5) 1 3 (0.5) 3 654 97
27
28 P. Savolainen
70 61
30 16 13
A
UK 2 4
50 17 19 19
P
A
71
12 2 15
B D
58 62
G S 1
UK
01 17 1
14 3 9
10
4 11
UK
33 20 10
18
6 48 3
16 8 56 18
5 7 4
42 6
46
S 1 2 9
3 22 82
8
24 31
75 74
43 5
7 1
34
49
60 29 UK UK
80 14 15
51
72
39 38
73
44
35
P 11
32
54 6
40 6
UK
25
3
59 15 UK
52 14 12
26
C E F
63 5
12
A
13
P
57 1
55 7
S
3
41
28 18 20 65 2
53
8 4 2
27 83 UK
47 64 22
66 1 2
81 68 19 36
45
G GS
67
69 21
Fig. 2.2. Minimum-spanning networks showing genetic relationships among dog

mtDNA types of clades A–F based on 573 base pairs sequence of the mtDNA CR
(Angleby and Savolainen, 2005). mtDNA types (white circles) are separated by one
mutational step, single strokes represent substitutions and double strokes represent
indels. Black dots are hypothetical intermediates. The largely random distribution
of breeds and morphological types of dogs among clades A, B and C is
exemplified by marking of the mtDNA types found in German Shepherd (mtDNA
types A17, A19 and C1, marked ‘G’), Pekingese (mtDNA types A11, B1 and C3,
marked ‘P’), Samoyed (mtDNA types A1, A18, B1 and C1, marked ‘S’) and Afghan
dog (mtDNA types A12 and B2, marked ‘A’), and by highlighting of the respective
mtDNA types by bold circles.
introduced to America by the ancestors of today’s Inuits rather than by the ances-
tors of the First Nations tribes. Other parts of America were not represented
because the original First Nations tribes’ dog populations are thought to have
been largely obliterated or mixed with European dogs in connection with the
arrival of Europeans, which would make the interpretation of American dog
sequences complicated. However, in a study of mtDNA in 19 Mexican hairless
dogs, a Mexican breed believed to have remained essentially isolated, the mtDNA
types all belonged to clades A, B and C (Vila et al., 1999b). Furthermore, the study
of archaeological remains of pre-European American domestic dogs (Leonard et
al., 2002) showed that among 13 samples from Mexico, Peru and Bolivia, 800
years or more old, 12 had mtDNA types clustering with clade A and one sample
had a type clustering with clade B. Thus, both clades A and B, and possibly clade
C, according to the data on the Mexican Hairless Dog, were present among the
native American dogs. Perhaps more importantly, none of the American samples
had sequences which did not fall into the same mtDNA-clades as the Old World
samples. Thus, in the available mtDNA data there is no sign of a separate domes-
tication of dogs in America, or of crossbreeding with the wolf. This shows that the
dog populations of the New World have a common origin with the dogs of the
Old World.
Domestication 29
An important observation is that the frequencies of clades A, B and C had

similar proportions in all regions (Table 2.1). This suggests that, unless there has
been a very effective gene flow along the Eurasian continent, mixing the gene
pools along the continent, the major present-day dog populations have had a
common origin from a single gene pool containing clades A, B and C. If the three
clades had different geographical origins, from separate domestications of the
wolf, the frequencies of the clades would be expected to vary between regions.
Had, for example, clade A originated in Europe and clade B in East Asia, we
would expect to find a higher frequency of clade A in Europe than in East Asia
and, vice versa, a highest frequency of clade B in East Asia, with intermediate fre-
quencies in Southwest Asia. The only major deviation from the uniform distribu-
tion of proportions of the three clades is found in the Southwest Asian sample,
which has a higher frequency of clade B sequences and a lower frequency for
clade A sequences than other regions. This could indicate that clade B originated
in Southwest Asia separately from clades A and C. However, if this was the case
it would also be expected that the genetic variation among the sequences belong-
ing to clade B would be higher in Southwest Asia than in other regions. On the
contrary, it was considerably lower in Southwest Asia than in East Asia, and equal
comparing Southwest Asia and Europe. This will be further discussed below, but
is exemplified by the number of mtDNA types belonging to clade B for the three
regions (Table 2.1). A single first origin for the three clades was further indicated
by the lack of division of the main morphologic types of dog (spitz, mastiff,
greyhound), or of large and small breeds, between the three main clades,
except for a lack of greyhounds in clade C (Fig. 2.2). This suggests that the
extreme morphologic variation among dog breeds is not the result of different
geographically distinct domestications of the wolf. It is impossible today to find
out how much the dog populations have been mixed through migration during
the thousands of years since the domestication of the wolf, and whether this could
explain the universal presence, at similar proportions, of clades A, B and C.
However, the similarities over long distances are striking. A comparison of dog
breeds from the two island-groups of the British Isles and Japan, situated outside
opposite parts of the immense Eurasian continent, offers a good example (British
Isles: 81.5%, 13.8% and 4.6%; Japan: 62.5%, 18.8% and 15.6%, for clades A, B
and C, respectively). Thus, there are strong indications of a single original gene
pool containing the three major dog-mtDNA-clades, A, B and C (Savolainen et
al., 2002).
The Geographic Origin of Dogs
An indication of the geographic origin of the three dog clades, and thereby the
origin of the dog, can be obtained from a comparison of the genetic variation
between geographical regions. If an ancestral population and a derived popula-
tion (formed from a subset of the genetic types of the ancestral population) are
compared, the number of mtDNA types and the nucleotide diversity are expected
30 P. Savolainen
to be higher in the ancestral population. Comparing Europe, Southwest Asia and

East Asia, the three regions that have been the strongest candidates as geograph-
ical origin of the domestic dog, East Asia had a larger genetic variation for clades
A and B as measured by a number of methods, while for clade C there were no
significant differences (Savolainen et al., 2002). Thus, for clade A, the mean pair-
wise sequence distance, which gives a rough measure of the genetic variation, was
3.39 (SD=0.13) substitutions in East Asia, 2.28 (SD=0.23) in Southwest Asia and
2.97 (SD=0.08) in Europe. Furthermore, the number of mtDNA types, and the
number of types unique to the region, were much larger in the East Asian sample
than in the European and the Southwest Asian ones (Table 2.1). The difference
between the East Asian and the European samples is striking, and, when cor-
rected for sample size by resampling to correct for different sample size, there
were significantly more mtDNA types in clade A among 51 East Asian dogs, than
among the 51 Southwest Asian dogs, 20.2 compared to 16. Also compared to the
other regions there were more mtDNA types in East Asia when corrected for
sample size, but not reaching a significant level for the small samples of Africa,
Siberia and India. It is notable that out of the 44 mtDNA types found in East
Asia, 30 were unique to this region. Thus, the number of mtDNA types unique
to East Asia was larger than the total number of types in Europe (Table 2.1). For
clade B, the mean pairwise sequence distance was larger for East Asia (0.93 sub-
stitutions, SD=0.17) than for Europe (0.45, SD=0.14) and Southwest Asia (0.36,
SD=0.11), the East Asian sample had significantly more mtDNA types than those
of Europe and Southwest Asia, and a majority of all mtDNA types were unique
to East Asia. The difference in number and distribution of mtDNA types in East
Asia, Europe and Southwest Asia can be studied in detail in minimum-spanning
networks of the three clades (Fig. 2.3). For clade A it is notable that East Asian
mtDNA types were distributed throughout the network, while for Europe and
Southwest Asia, parts of the network, largely the same in the two populations,
were empty.
To conclude, in a comparison of the dog populations of East Asia, Europe
and Southwest Asia, there is a much higher genetic variation in East Asia for the
two major dog mtDNA-clades, clades A and B, which comprise >90% of all
domestic dog mtDNA sequences. This suggests an origin somewhere in East Asia
for the mtDNA types of clades A and B, and a subsequent spread of only a subset
of these mtDNA types to the rest of the world. For clade C there is no significant
difference between the regions, but the similar frequency of clade C among dog
populations of the Old World suggests a common origin of clade C together with
clades A and B. Thus, the available mtDNA data indicate that the domestic dog
has a single geographical origin, somewhere in East Asia.
Dating the Origin
The time of origin for each dog mtDNA clade can be estimated from the mean
genetic distance of the sequences in each clade to the original wolf mtDNA type,
Domestication 31
Clade A
Clade C Clade B
East Asia Europe SW Asia
Fig. 2.3. Minimum-spanning networks showing genetic relationships among

mtDNA types of phylogenetic clades A, B and C (Savolainen et al., 2002). mtDNA
types (circles) are separated by one mutational step, ignoring indels. mtDNA types
found in East Asia, Europe and Southwest Asia are indicated in separate networks.
The sizes of grey circles are proportional to mtDNA-type frequency in the
respective populations. Small uncoloured circles denote mtDNA types not found
in the regional population. Black dots are hypothetical intermediates. Uncoloured
squares are wolf mtDNA types. Subclusters of clade A discussed in the main text,
three in the East Asian and one in the European network, are marked by black
lines.
and the mutation rate of the analysed region. Under the assumption that the
clade originates from a single wolf mtDNA type, the age of the clade will ap-
proximately correspond to the origin of the dogs originating from that female
wolf. However, there are two problems with the available data that render
a precise dating of the origin of the dog based on the available mtDNA data
impossible.
The first problem concerns the dating of the palaeontological fix-point
needed to calibrate the mutation rate. The wild canid most closely related to the
wolf is the coyote, and the time for the split into two species from their common
ancestor is the best available calibrating point for the mutation rate of dog
mtDNA. Counting the number of sequence differences between wolf and coyote
in the analysed mtDNA region, and dividing by the divergence time, gives the
mutation rate. However, while a first appearance of wolves ~700,000 years ago
and of coyotes ~1 million years ago (Kurtén, 1968; Kurtén and Anderson, 1980)
indicates a date for the divergence between the two species of approximately
1 million years ago, the time of divergence has not been definitely established and
an earlier date of up to approximately 2 million years cannot be ruled out.
Therefore, without more exact palaeontological evidence, a precise dating of the
32 P. Savolainen
origin of the dog will not be possible, but it can nevertheless give an indication of
the probable range of time, and this information can be used together with other
data from the archaeological record on the first appearance of dogs. To simplify
the following argument, the 1-million-year date for the wolf–coyote divergence
will be used for the calculations based on mtDNA, but bearing in mind the pos-
sibility that the split occurred up to 2 million years ago, in which case the age of
the datings should be doubled. The second problem is that the mutation rate of
the analysed 582 base pairs region does not give resolution between mtDNA types
in the time scale needed to monitor the last 20,000 or 30,000 years. According to
the calculations above, the rate of mutation is approximately one mutation per
24,000 years in a lineage (Savolainen et al., 2002). This implies
that if two wolves, having mtDNA types differing by a single mutation, were
domesticated say 15,000 years ago, the dogs of today originating from those two
lineages would have mtDNA types differing from each other only by the
single original mutation or by just one or two additional mutations. If there
were several wolves having mtDNA types differing by just a few mutations, it
would not be possible to fully resolve the mtDNA types of the dogs originating
from them. Thus, the period around 15,000 years ago, which is suggested to be
the time for the origin of the domestic dog according to the archaeological record,
cannot be fully studied using the 582 base pairs mtDNA region available at
present.
In a domestication event with a subsequent population expansion, a star-like
phylogeny, with the founder mtDNA type in the centre and new mtDNA types
distributed radially, would be expected. The networks of clades B and C are star-
like, indicating an origin from a single wolf mtDNA type (Figs 2.2 and 2.3). In
contrast, clade A has a complicated pattern without an easily identifiable central
node. A distance of up to 11 substitutional steps between mtDNA types would
indicate that clade A is much older than clades B and C, and derives from an
initial domestication of wolves. However, as discussed above, the dog mtDNA
types in clade A do not necessarily originate from a single wolf mtDNA type even
though clade A is an almost completely continuous group of mtDNA types sepa-
rated from each other by single mutational steps. It is possible that clade A origi-
nally was a clade of several closely related wolf mtDNA types, and that several
wolves having a number of these mtDNA types belonging to clade A were domes-
ticated. Looking more closely at clade A it has, instead of a single central node,
several subclusters with star-like shape, suggesting that clade A may have origi-
nated from several wolf mtDNA types (Figs 2.2 and 2.3). The approximate age of
clade A, assuming a single origin from the wolf and a subsequent population
expansion, is estimated from the mean pairwise distance between East Asian dog-
mtDNA-sequences (3.39 substitutions, SD=0.13) and the mutation rate to 41,000
± 4000 years. This calculation may be biased by the population history among
the dogs and wolves. Alternatively, the maximum age of clade A can be estimated
from the number of steps between the most distantly related mtDNA types, 11
substitutions apart which corresponds to ~120,000 years. According to these cal-
culations clade A would have originated 40,000–120,000 years ago, and if it is
Domestication 33
supposed that it was formed in a domestication event from a single wolf mtDNA
type, the domestic dog would have originated 40,000–120,000 years ago. If
instead an origin of clade A from several different wolf mtDNA types is assumed,
several reasonably defined subclusters can be found. To give an alternative dating
of the domestication of the dog, three subclusters of clade A possibly representing
three origins of dog from wolf, marked by lines in Fig. 2.3, can be studied. The
mean genetic distances of the sequences belonging to these subclusters to their
respective nodes (0.45, 0.65 and 1.07 substitutions with SD=0.13, 0.09, 0.27,
respectively) give estimates of 11,000 ± 4000, 16,000 ± 3000 and 26,000 ± 8000
years for their ages, respectively. Thus, clade A has a substructure, suggesting that
it could have been formed by several wolf mtDNA types, possibly ~15,000 years
ago. Assuming single wolf mtDNA types as founders of clades B and C, the mean
distances among East Asian sequences to the nodes (0.54 and 0.71 substitutions,
SD=0.08 and 0.10) give estimated ages of 13,000 ± 3000 and 17,000 ± 3000 years
for clades B and C, respectively. Thus, clade A was formed ~40,000–120,000
years ago but has a substructure indicating later population events, and clades B
and C were formed ~15,000 years ago. Depending on how these data are inter-
preted, it can suggest a first origin of domestic dogs either ~40,000–120,000 years
ago forming only clade A, after which clades B and C were introduced into the
dog gene-pool through crossbreeding between dogs and wolves ~15,000 years
ago, but it can also suggest a single origin ~15,000 years ago involving all the
three clades A, B and C.
The key question for the dating of the origin of the dog is the number of wolf
mtDNA types that formed dog-clade A, but this question cannot be answered
based on the present mtDNA data. However, indirect evidence in the form of the
universal presence of clades A, B and C at similar proportions talks in favour of a
simultaneous origin of the three clades. Assuming that clade A would have origi-
nated from an initial domestication approximately 40,000 years ago, the dog pop-
ulation originating from that domestication event would be expected to have
spread to other parts of the world. An introduction of clades B and C, 15,000
years ago, into the already existing domestic dog gene pool, by regional cross-
breedings with wolves, would require a very thorough mixing to have occurred,
through migration to all parts of Eurasia, to level the frequencies of the three
clades to the very similar proportions now found throughout the world.
Alternatively, if clade A originated from a domestication 40,000 years ago, these
dogs could have remained isolated in the original geographical region, not
spreading to other parts of the world. Clades B and C would then, 15,000 years
ago, have been introduced into the domestic dog gene pool by regional cross-
breedings with wolves, after which the dogs would have spread to other parts of
the world. While the amount of migration and trade 15,000 years ago is not
known, this scenario seems unlikely. It is also contradicted by the age of the oldest
subcluster of clade A in Europe (as determined from the mean genetic distance
from mtDNA types unique to the western part of the world to the nodal mtDNA
type shared with East Asia; 0.39 substitutions, SD=0.09), which is estimated to be
only 9000 ± 3000 years old (Fig. 2.3).
34 P. Savolainen
Thus, a first origin of dog mtDNA clade A, a long time before the origin
of dog clades B and C, does not seem probable considering the very similar
proportions of clades A, B and C around the world. The total sum of circum-
stantial mtDNA evidence therefore indicates a single origin, in both place and
time, for the three clades, approximately 15,000 years ago. Furthermore, in the
absence of a clear result from the mtDNA data, the best evidence for the time of
the first origin of the domestic dog remains the archaeological record, which indi-
cates an origin approximately 15,000 years ago. A synthesis of mtDNA data: the
similar proportions of clades A, B and C, and the larger genetic variation for
clades A and B; and archaeological data: the oldest evidence of domestic dog
dated at approximately 15,000 years ago, therefore point to an origin of the
domestic dog in East Asia ~15,000 BP. In this event, clade A would have had
several origins from wolf mtDNA types, at least around ten, and the first domes-
tication of wolves would not have been an isolated event, but rather a common
practice in the human population in question.
Possible Crossbreedings with Wolves
Clades D, E and F have been left out of the discussion so far to simplify the
discussions. These three clades are found only regionally, and clades E and F
also at very low frequency, three dogs for each clade, making population
genetic analyses virtually impossible (Table 2.1). The three clades, which
represent three separate origins from the wolf, could either be the results
of crossbreedings between male dogs and female wolves or separate domestica-
tions of the wolf, or they could have originated together with clades A, B and C
at a single domestication event. Since clades E and F are found in East Asian
dogs, an origin together with clades A, B and C seems possible. In this case they
would have been part of a gene-pool containing the five clades A, B, C, E and F,
but clades E and F, like some parts of clade A, would not have spread to other
parts of the world. For clade D the situation is very different. This clade is found
only in Europe and Turkey, and it has therefore probably an origin outside East
Asia, independent from that of the other clades. There is a clear geographical
division within clade D. mtDNA types D1–D4, which constitute a separate sub-
cluster, were all found in Scandinavian spitz breeds, in total 18 out of 49
Scandinavian dogs, while D5 and D6, which are several mutations away from
D1–D4, were found in two Turkish and one Spanish dog, respectively (Fig. 2.2).
The distance between the subclusters D1–D4 and D5–D6, three substitutions,
corresponds to approximately 36,000 years. It therefore seems probable that
clade D originates from at least two separate origins from the wolf. The diver-
gence within subcluster D1–D4, a single substitution in 18 individuals, indicates
a recent origin, a few thousand years at most. Thus, at least clade D seems to have
originated through crossbreeding between male dogs and female wolves. This
crossbreeding seems to have had some impact on the Scandinavian breeds, con-
sidering the large proportion of dogs having mtDNA types belonging to clade D,
Domestication 35
18 out of 49 dogs (37%). It is probable that, in crosses between male dogs and
female wolves, the chances of the offspring becoming part of the domestic dog
population would be low. The other scenario, that offspring from male
wolf/female dog crossbreedings would ‘become dogs’ seems more probable. It is
believed that in some cultures, especially in the northern parts of Eurasia and
America, dogs, and above all bitches, were crossed with wolves in order to bring
new blood to the dog population. Future studies of the Y chromosome will no
doubt tell us to what degree such crossbreedings have contributed to the devel-
opment of domestic dogs.
The Origin of the Australian Dingo
An interesting question connected to the history of the domestic dog is the origin
of the Australian dingo. The dingo is a wild canid, morphologically resembling
South Asian domestic dogs, which according to the archaeological record arrived
in Australia between 3500 and 12,000 years ago (Clutton-Brock, 1995). However,
the precise ancestry and time of arrival in Australia of the dingo have not been
known, nor whether, on its arrival, it was domesticated or half-domesticated
before becoming feral, or a truly wild dog. In a study of 211 dingoes, sampled
around Australia, 582 base pairs of the mtDNA CR were analysed and compared
to dog and wolf data, giving very distinct results (Savolainen et al., 2004). All dingo
sequences clustered in dog clade A, forming a single internal cluster of mtDNA
types unique to dingoes, around a central type, A29, found in both dingoes and
dogs (Fig. 2.4). Among domestic dogs, A29 was found only east of the Himalayas,
together with most types in that part of clade A. This indicates that the dingo pop-
ulation originates from a single introduction of a small population of domestic
dogs coming from East Asia, carrying mtDNA type A29. The mean distance
among the dingo sequences to A29 indicates their arrival to Australia ~5000
years ago. The dingo has probably remained isolated since then, and represents
a unique isolate of early undifferentiated dogs.
Summary
Studies of physical and behavioural characteristics, and of molecular genetic

markers, give consistent evidence that the domestic dog originates from the wolf,
and the archaeological record indicates that the wolf was domesticated approxi-
mately 15,000 years ago. Population genetic data, based on analysis of mtDNA,
shows that all dogs originate from a common gene pool, and indicates that they
originated at a single domestication event in East Asia, from which they spread to
all continents, including Australia where the domestic dogs developed into the
feral dingo. The mtDNA data give only two clear indications of crossbreeding
between dog and wolf, indicating that crossbreeding between male dogs and
female wolves has contributed only marginally to the dog gene-pool. Studies of Y-
36 P. Savolainen
A29
Fig. 2.4. Minimum-spanning network of the main dog clade, A (Savolainen et al.,
2004). mtDNA types (circles) and hypothetical intermediates (solid dots) are
separated by one mutational step (substitutions, indels not shown). White circles
are mtDNA types found in dogs only; light grey circles are types found in dingoes
only; dark grey circles are types found in both dingoes and dogs. Squares denote
wolves. Areas of grey circles are proportional to frequencies among dingoes, but
the area of A29 is reduced by 50%.
chromosomal and autosomal haplotypes, which will no doubt be initiated within

a few years, are necessary to test the validity of the conclusions obtained from the
mtDNA data, and will, by giving the male side of the story, help to create an even
more detailed picture of the evolution of the dog.
References
Angleby, H. and Savolainen, P. (2005) A study of the forensic usefulness of the mito-
chondrial DNA variation among and within populations, breeds and types of
domestic dogs. Forensic Science International 154, 99–110.
Clutton-Brock, J. (1995) Origins of the dog: domestication and early history. In:
Serpell, J. (ed.) The Domestic Dog, its Evolution, Behaviour, and Interactions
with People. Cambridge University Press, Cambridge, pp. 7–20.
Clutton-Brock, J. and Noe-Nygaard, N. (1990) New osteological and C-isotope
Domestication 37
evidence on mesolithic dogs: companions to hunters and fishers at Starr Carr,

Seamer Carr and Kongemose. Journal of Archaeological Science 17, 643–653.
Houpt, K.A. and Willis, M.B. (2001) Genetics of behaviour. In: Ruvinsky, A. and
Sampson J. (eds) The Genetics of the Dog. CABI, Wallingford, UK, pp.
371–400.
Kurtén, B. (1968) Pleistocene Mammals of Europe. Aldine, Chicago, Illinois.
Kurtén, B. and Anderson, E. (1980) Pleistocene Mammals of North America.
Columbia University Press, New York.
Leonard, J.A., Wayne, R.K., Wheeler, J., Valadez, R., Guillen, S. and Vila, C. (2002)
Ancient DNA evidence for Old World origin of New World dogs. Science 298,
1613–1616.
Li, Q. (1990) The history of dog breeding in China. Chinese Journal of Cynology 2,
25–27 (in Chinese).
Nobis, G. (1979) Der älteste Haushund lebte vor 14,000 Jahren. Umschau 79, 610.
Okumura, N., Ishiguro, N., Nakano, M., Matsui, A. and Sahara, M. (1996) Intra- and
interbreed genetic variations of mitochondrial DNA major non-coding regions
in Japanese native dog breeds (Canis familiaris). Animal Genetics 27, 397–405.
Olsen, S.J. (1985) Origins of the Domestic Dog. University of Arizona Press, Tucson,
Arizona.
Sablin, M.V. and Khlopachev, G.A. (2002) The earliest ice age dogs: evidence from
Eliseevichi 1. Current Anthropology 43, 795–798.
Savolainen, P., Zhang, Y., Luo, J., Lundeberg, J. and Leitner, T. (2002) Genetic evi-
dence for an East Asian origin of the domestic dog. Science 298, 1610–1613.
Savolainen, P., Leitner, T., Wilton, A., Matisoo-Smith, E. and Lundeberg, J. (2004) A
detailed picture of the origin of the Australian dingo, obtained from the study of
mitochondrial DNA. Proceedings of the National Academy of Sciences, USA
101, 12387–12390.
Tchernov, E. and Valla, F.R. (1997) Two new dogs, and other Natufian dogs, from
the southern Levant. Journal of Archaeological Science 24, 65–95.
Tsuda, K., Kikkawa, Y., Yonekawa, H. and Tanabe, Y. (1997) Extensive interbreed-
ing occurred among multiple matriarchal ancestors during the domestication of
dogs: evidence from inter- and intraspecies polymorphisms in the D-loop
region of mitochondrial DNA between dogs and wolves. Genes and Genetic
Systems 2, 229–238.
Vila, C., Savolainen, P., Maldonado, J.E., Amorim, I.R., Rice, J.E., Honeycutt, R.L.,
Crandall, K.A., Lundeberg, J. and Wayne, R.K. (1997) Multiple and ancient
origins of the domestic dog. Science 276, 1687–1689.
Vila, C., Amorim, I.R., Leonard, J.A., Posada, D., Castroviejo, J., Petrucci-Fonseca,
F., Crandall, K.A., Ellegren, H. and Wayne, R.K. (1999a) Mitochondrial DNA
phylogeography and population history of the grey wolf Canis lupus. Molecular
Ecology 8, 2089–2103.
Vila, C., Maldonado, J. and Wayne, R.K. (1999b) Phylogenetic relationships, evolu-
tion, and genetic diversity of the domestic dog. Journal of Heredity 90, 71–77.
Wayne, R.K. (1993) Molecular evolution of the dog family. Trends in Genetics 9,
218–224.
Origin of Dog Breed Diversity
3
C. Vilà and J.A. Leonard
Introduction
Modern dog breeds demonstrate a stunning level of morphological and behav-

ioural diversity. The magnitude of morphological differences between dog breeds
is greater than that between all wild species in Canidae, the dog family (Wayne,
1986a, b). This diversity led Charles Darwin and Konrad Lorenz to speculate
that perhaps dogs derived from more than one wild ancestral species. However,
molecular genetic data has shown conclusively that domestic dogs derive only
from the grey wolf (Canis lupus) and not any other wild canid (Seal et al., 1970;
Sarich, 1977; Ferrell et al., 1978; Wurster-Hill and Centerwall, 1982; Wayne and
O’Brien, 1987; Wayne et al., 1987a, b, 1997; see Chapter 2 in this volume). This
brings us to the next, obvious question: where did all of the diversity we observe
in dogs today come from?
Much has been learnt about the origin of dogs thanks to many years
of archaeological research. This research is based on the study of all sorts of
material that could indicate prehistoric associations between dogs and humans,
from bones to art. During the last 10 years, the development of molecular
genetics techniques has opened a new field of research, complementary to
the archaeological investigation, which can contribute to a better understanding
of the domestication of the dog. In this chapter we describe some molecular
techniques and some studies that have employed molecular techniques to clarify
some aspect of the process by which dogs were domesticated and modern dog
breeds were formed. However, as we will see, these studies have not yet led to a
complete understanding of the process and many questions remain to be
answered.
38 © CAB International 2007. The Behavioural Biology of Dogs

(ed. P. Jensen)
Origin of Breed Diversity 39
Molecular Methods to Study Dog Origins
DNA contains the genetic information that each one of us has inherited from our
parents and that codes for most of our physical characteristics. This information
is coded by a chain of four nucleotides (normally symbolized as G, A, T and C;
for example, a short DNA sequence could be represented:
GAAATCTCAGGCA). The total length of the DNA chains (chromosomes) in a
typical human cell is about 3 billion nucleotides. All the members of the same
species have very similar DNA sequences, which are different from the DNA in
other species. The study of the changes observed in the nucleotide sequences can
inform us about the evolutionary history of the species. Such changes accumulate
over time and may be uniquely shared among taxa. Hence nucleotide changes
are markers of common ancestry, and if they accumulate in a regular fashion with
time, may be used to estimate divergence time and time of origin. As an example,
consider hypothetical sequence changes in the DNA sequence responsible for the
synthesis of haemoglobin, the protein responsible for binding oxygen in red blood
cells. Suppose all coyotes (Canis latrans), grey wolves and dogs shared 90
nucleotides in a fragment of 100 in the haemoglobin gene, whereas the grey wolf
and dog shared, in addition, eight more nucleotides. Consequently, the dog and
wolf differ by two unique substitutions. Such nested changes allow construction of
a phylogenetic tree where each node and branch is defined by a unique set of
nucleotide changes (see example in Fig. 3.1). Moreover, if the fossil record sug-
gested that wolves and coyotes diverged about 1 million years ago, and they differ
by 10 nucleotide changes, then the two nucleotide differences between dogs and
wolves imply a divergence time of 200,000 years ago. Of course, this assumes
nucleotide substitutions accumulate in a clock-like fashion with little variation
and, although the use of molecular data to estimate divergence time is a common
practice, considerable debate surrounds the accuracy of divergence dates (Avise,
1994, 2000).
Many genetic studies of dog domestication are based on the comparison of
mitochondrial DNA sequences. Among all DNA, a small part (about
16,000–18,000 nucleotides in length in mammals) is located inside the mitochon-
dria and is inherited just from the mother. This is the mitochondrial DNA
(mtDNA). Since this molecule is inherited only from the mother and is not mixed
with paternal DNA, it is particularly simple to study. This has made the analysis of
mtDNA sequences one of the main tools used to reconstruct the evolutionary rela-
tionships between species. The process of domestication needs to be studied with
markers that have the highest mutation rates in the mitochondrial genome because
divergence between dogs and wolves is extremely recent on an evolutionary time
scale. This high mutation rate allows for the relatively fast accumulation of infor-
mation. This circumstance has led many researchers to focus their investigation on
the control region, a fragment of the mitochondrial DNA that might be free of
strong selective forces and where most mutations are likely to persist.
Although there are many features of mtDNA that make it a very useful tool,
it records only the history of the female lineage. In some cases the history of the
40 C. Vilà and J.A. Leonard
Grey Wolf
Grey Wolf
Fig. 3.1. Example of the use of DNA sequence for the construction of phylogenetic
trees. The mutations that differentiate the sequences are indicated in bold and
identified by numbers. These mutations are used to build the phylogenetic tree,
and the position of each mutation is indicated on the corresponding branch in the
tree together with the nucleotide change involved. The outgroup (golden jackal in
our case) corresponds to a taxon closely related to the group under study. Its
sequence is used to give polarity to the tree and provide information on the
branching order. In this example, dog and grey wolf differ by two substitutions and
each of them differs from coyote by four substitutions.
males and females of a population or breed may be very different. For this reason
other markers which have different modes of inheritance complement mtDNA
studies very well. For example, the Y chromosome records only the male lineage
because it is passed only from father to son, and nuclear autosomal markers are
inherited from both parents (Fig. 3.2). Most Y chromosome and nuclear autoso-
mal markers which have been studied so far are microsatellites, simple repetitive
sequences which vary in size, and are considered neutral. Some work has also
been done on regions of the autosomal genome which are thought to be under
selection, such as the major histocompatibility complex (MHC).
Number of Origination Events
Mitochondrial DNA sequences from dogs have been collected to determine the
number of origination events in dogs. By themselves these sequences have little
value, but the same region was also sequenced in a large number of grey wolves
from across their range. The sequences identified in dogs were compared to those
Fig. 3.2. Schematic representation of inheritance of different types of genetic

markers in mammals; black represents paternal DNA, grey represents maternal
DNA, X represents autosomal chromosomes, x and y represent the sex
chromosomes and m represents mitochondrial DNA.
identified in wolves, and a phylogenetic tree was constructed (Fig. 3.3). Most of
the dogs formed a single group in the phylogeny, called a clade. This clade, which
contains about 80% of the dog sequences, is called clade I (Fig. 3.3; Vilà et al.,
1997). In the original study three other minor clades were also identified (Vilà et
al., 1997). Since then two more minor clades have been identified (Savolainen et
al., 2002). These data could support either a single origination event which gave
rise to clade I and a small number of subsequent hybridization events or a small
number of origination events (Vilà et al., 1997, 1999; Savolainen et al., 2002). This
pattern of a small number of origination events identified by mtDNA sequence
surveys characterizes the origin of many livestock mammals (Bruford et al., 2003),
for example: sheep (Hiendleder et al., 2002), goats (Luikart et al., 2001), cattle
(Troy et al., 2001), pigs (Giuffra et al., 2000), water buffalo (Lau et al., 1998) and
donkey (Beja-Pereira et al., 2004). In many cases this view is supported by archae-
ological studies (Clutton-Brock, 1999). Despite the existence of a limited number
of well-defined clades of mtDNA sequences, Savolainen et al. (2002) inferred that
dogs derived from a unique centre of domestication. This was due to the concen-
tration of divergent mtDNA sequences in East Asia and the authors assumed that
this would imply that dogs had been domesticated there. However, it is not clear
that modern dogs can be used to infer patterns of diversity in the past. A study of
ancient dog remains has shown that dogs living in America today are genetically
very different from the dogs living there 500–1000 years ago (Leonard et al.,
2002). The number of domestication events for dogs thus remains unresolved.
Fig. 3.3. Phylogenetic tree of wolf (W) and dog (D) mitochondrial DNA control
region sequences (Vilà et al., 1997). Dog haplotypes are grouped in four Clades, I
to IV. Boxes indicate haplotypes found in the 19 Mexican hairless dogs (Vilà et al.,
1999).
Time of Domestication
Two different approaches have been used to date the time when the domestica-
tion of dogs took place. The first one is to try to identify bone remains whose mor-
phology corresponds to dogs and not to wolves (the ancestor of the domestic dog).
The second method consists of estimating the date when the mtDNA sequences
present in dogs separated from the sequences in wolves. Although both methods
target the moment when the dog and wolf lineages split, they look at different
things and the dates obtained with both methods are expected to be different. To
understand this, imagine one lineage that at a certain time in the past splits in two,
as indicated in Fig. 3.4. This could correspond to the split of dogs from the ances-
tral wolf lineage, and the goal of the studies about domestication is to character-
ize when this separation occurred. It is at this point that the morphology of the
two species starts to diverge. However, it will take some time for their morphol-
ogy to be sufficiently divergent as to allow the unambiguous identification of bone
remains as belonging to one species or the other. Zeder and Hesse (2000) and
Zeder (2003) have shown that for other domestic mammals the morphological
differentiation from the ancestor did not occur until some time after the initial
domestication, when the domesticates expanded to neighbouring areas where
they could not mate with wild animals. Similarly, we could expect that the mor-
phological change in dogs may have occurred some time after the domestication
Fig. 3.4. Difference between the species divergence time, and the estimates
obtained using genetic and morphological data. Species are indicated as grey
bands, and the black lines represent the genetic relationships between sequences
from each species. Morphological differentiation between species increases from
the time of reproductive isolation.
(although we do not know how much later). Since this is the kind of evidence used
by archaeologists, and because the ancient dog bone remains are very scarce, the
dates of domestication based on archaeological data will tend to be more recent
than the split between dogs and wolves.
Molecular estimates of the time of divergence are based on the comparison
of DNA sequences. By applying a molecular clock it is possible to estimate the
time necessary to make the two lineages as different as they currently are. In Fig.
3.4, the mtDNA lineages present in wolves and in dogs are represented as the
lines inside the grey area. However, the split of the lineages is necessarily older
than the split of the populations (the domestication). Consequently, estimates of
the date of the domestication based on DNA sequences are likely to be older than
the actual event.
Another way to use mtDNA sequences to estimate the time of domestication
is to calculate the time of divergence between the most divergent sequences within
the domestic clade (the time when dogs started diverging from one another; Fig.
3.4). This estimate is highly dependent on the demographic history of dogs: the
origin of these sequences might appear more recent when the popula-
tion remained small (Avise, 2000). However, it seems logical to think that dogs
started from a limited number of individuals and soon after domestication the
population started to expand to produce the current population size of several
hundred million dogs worldwide (Coppinger and Coppinger, 2001). In this case,
the estimates should be reasonably close to the beginning of the dog population
expansion, and probably also close to the time of domestication. This is likely
a very much better estimator of the age of domestication than the previous
ones and should be expected to be younger than the split between dogs and
wolves, and older than the first appearance of morphological divergence (see
Fig. 3.4).
In summary, while archaeological approaches may lead to an underestima-
tion of the date of domestication, molecular methods tend to produce overesti-
mates. Estimating the age of the diversity within the domestic lineage as opposed
to the split between the domestic and wild lineages may produce a more accurate
date. While the search for old dog-remains continues, molecular analysis methods
are constantly improving, and integrating both approaches will lead us towards a
more accurate view of the domestication process.
The archaeological record suggests that dogs were present in the Middle East
and Central Europe about 14,000–15,000 years ago (Nobis, 1979; Olsen, 1985;
Dayan, 1994; Clutton-Brock, 1999; Sablin and Khlopachev, 2002). Early dogs
were morphologically distinct from grey wolves. They are often identified by their
smaller body size and wider crania, a more prominent stop on the face and a
shortened, crowded jaw (Olsen, 1985; Morey, 1992). The identification of these
bones as belonging to dogs and not to wolves implies that the animals had been
living with humans for a time period sufficiently long to allow some morphologi-
cal differentiation from their wild ancestor. The small body size of Asian wolves
and the shared presence with early dogs of some traits led to the suggestion that
Asian wolves are the direct ancestor of the dog (Olsen and Olsen, 1977).
However, the oldest dog-remains unmistakably identified so far have recently

been discovered in western Russia (Sablin and Khlopachev, 2002) and date to
13,000–17,000 radiocarbon years before present (this is equivalent to
16,000–20,000 calendar years before present). These early dogs do not show a
reduction in body size when compared to the local populations of grey wolves and
have cranial proportions similar to the great dane.
The first attempts to use molecular genetics to study the domestication of
dogs took place in the mid-1990s. Okumura et al. (1996) were the first to sequence
the control region of a large sample of dogs, but focused on Asian dogs. These
authors observed that the mitochondrial DNA sequences grouped into a discrete
number of clades. Tsuda et al. (1997) similarly studied a large sample of dogs from
Asia, but also included some samples from wolves. Consequently, they were able
to convincingly demonstrate a close relationship between wolves and dogs. The
close relationship was confirmed in a study published at the same time (Vilà et al.,
1997), but in this study an attempt to use the molecular information to estimate
the date of domestication was made. The study by Vilà et al. included 162 wolves
from 27 populations from throughout Europe, Asia and North America and 140
dogs representing 67 breeds. As Okumura et al. did 1 year earlier, Vilà et al.
observed that when constructing an evolutionary tree with all dog sequences, they
clustered in just four groups (clades I–IV, Fig. 3.3). One of these groups, clade I,
contained most of the studied dogs, and a more recent study including more than
600 modern dogs (Savolainen et al., 2002) concluded that about 71% of them had
sequences belonging to this clade. This large diversity could indicate that this
group of sequences was the oldest one and thus its origin could correspond to the
original domestication event.
To estimate the date of the domestication, Vilà et al. (1997) calculated how
much time would be necessary to reach that level of diversity starting from a
single mtDNA sequence, using the molecular clock. Vilà et al. first estimated the
average sequence divergence between wolves and coyotes for the studied region
of the mtDNA and found that it was 7.5%. The fossil record indicates that mor-
phologically differentiated wolves and coyotes existed at least 1 million years ago
(Kurtén and Anderson, 1981). Next, the maximum divergence for sequences
belonging to clade I was estimated to be 1%. This was used to calculate when the
sequences from that clade started to diverge. To confirm that the rate of evolu-
tion was not significantly different for the different lineages, the lengths of the dif-
ferent branches in the tree were compared. As no difference was found, the same
molecular clock could be used for all the sequences. Finally, since the rate of
divergence was around 7.5% per million years, the estimated time needed to gen-
erate divergences up to 1% was 135,000 years! The result of the study of Vilà et
al. (1997) has been taken as a firm date. However, that estimation was not pre-
sented with confidence intervals and these are usually large in this sort of estimate.
Also, the region of the mtDNA studied may not be the best to apply a strict
molecular clock. On the other hand, the results very clearly suggested that the
domestication of dogs had to be much older than indicated by the archaeological
record.
More recently, another attempt has been made to use dog mtDNA informa-
tion to estimate when dogs were domesticated (Savolainen et al., 2002). These
authors also focused on the diversity of clade I and concluded that the domesti-
cation could have taken place more recently. These authors used a different
model of sequence evolution and only considered the sequences in their sample
that were currently present on each continent. Since they found the largest diver-
sity in East Asia they concluded that this could have been the place of domesti-
cation. Looking at the average sequence divergence among clade I sequences in
East Asia, they estimated that the age of the clade was 41,000 years (SD±4000),
supporting an origin much older than suggested by the bone remains as the pre-
vious study. However, they considered that it was possible that clade I did not
derive from one unique domestication event and that several wolf lineages were
involved. Clade I was arbitrarily subdivided into various smaller groups and the
ages of these sub-clades were estimated to be between 11,000 and 26,000 years
old. Despite this wide range of dates, they suggested an East Asian origin for the
domestic dog 15,000 years ago (Savolainen et al., 2002). There is a series of prob-
lems with this approach. First, East Asia was suggested as the centre of domesti-
cation because of its higher diversity. However, whereas most of the samples from
the rest of the world corresponded to purebred dogs, this region was character-
ized by mongrel dogs and dogs of non-recognized breeds. As these animals do not
belong to inbred lines it seems likely that they may have a much higher genetic
diversity for this reason alone, and this can bias the results. Second, the patterns
of current diversity may be very different from ancestral patterns (see the study by
Leonard et al., 2002), so looking at modern dogs may not be appropriate to recon-
struct patterns of diversity that existed in one region in the past. Third, no objec-
tive evidence indicating that clade I derived from one or more founding wolf
lineages was provided. Subjectively dividing the clade in smaller groups can
produce whatever result the researcher wishes to obtain. Lastly, despite estimat-
ing a very old date of origin for one of their sub-clades, 26,000 years (confidence
interval: 18,000–34,000 years), the authors conclude that the domestication took
place about 15,000 years ago.
Although the results of this study apparently fit with the dates derived from
archaeological remains, the conclusion that dogs could have been domesticated
in East Asia 15,000 years ago faces one main problem: dog remains (already mor-
phologically differentiated from wolves) of about that time have already been dis-
covered in several places in Europe (Nobis, 1979; Sablin and Khlopachev, 2002)
and Asia (Dayan, 1994; Clutton-Brock, 1999), and probably also existed in
America (Leonard et al., 2002). The analysis of mtDNA sequences in pre-
Columbian dogs of Leonard et al. (2002) has shown that American dogs clearly
had Eurasian origin, implying that they arrived in the Americas with the first
humans. Since all these dogs in Europe, Asia and America shared a common
origin, we can presume that the domestication of the dog had to take place sig-
nificantly before 15,000 years ago to allow their expansion over three continents.
Place of Domestication
As mentioned above, the dog was domesticated from the grey wolf. Therefore,
the place of domestication must be somewhere in the range of this species. Grey
wolves are naturally distributed across Europe, Asia and North America, so those
are the initial possibilities. Very little phylogeographic patterns have been found
in grey wolves worldwide, but American and Eurasian animals do not share hap-
lotypes (Vilà et al., 1999) and dog haplotypes are more closely related to Eurasian
wolf haplotypes than American wolf haplotypes (Vilà et al., 1997; Leonard et al.,
2002). This narrows the range of possibilities to Eurasia. Within Eurasia, diver-
gent lineages have been found in the grey wolves of the Himalayas and Indian
subcontinent, which are quite divergent from those found in any dogs (Sharma et
al., 2003). This excludes this region as a possible centre of domestication. This
narrows the geographic range of domestication to somewhere in Eurasia except
India or the Himalayas. Other methods will be required to narrow the area down
further.
In order to more precisely identify the place of domestication, Savolainen et
al. (2002) compared the amount of haplotype diversity in dogs in different regions
of the world. They concluded that a single domestication event may have taken
place in East Asia. They based their conclusion in the presence of a larger diver-
sity of sequences in this region compared to other regions around the world.
However, these conclusions may have been highly influenced by the predomi-
nance of dogs that did not belong to pure breeds recognized by the International
Dog Federation (FCI) in East Asia. While purebred dogs are characterized by
high levels of inbreeding, mongrel dogs show a very high genetic diversity (Irion
et al., 2005). This implies that a group of dogs that has not been subjected to the
extreme selection process applied to purebred dogs registered in kennel clubs is
expected to have a much higher genetic diversity than a similar group of pure-
bred dogs. Additionally, it is not clear to what degree a sample of modern dogs
can fully characterize past patterns of diversity. The study of Leonard et al. (2002)
showed that modern purebred American dogs are very different from the dogs
that existed there a few centuries ago. Human history, demography, migrations
and even fashion have clearly affected the composition of modern dog popula-
tions (Valadez, 1995) and the information about past patterns of diversity are
blurred. Consequently, the study of modern dogs may not be the best way to
study the place of origin of dogs, and studies of ancient dog remains as well as
extensive archaeological research may be needed to provide a final answer to the
questions about the number of domestication centres and the place of origin of all
modern dogs. However, the answer to this question may not be simple, and
recent genetic studies suggest that the separation between wild and domestic lin-
eages may not have been clean cut during part of the dog’s history (see next
section).
Early Breeding Practices
The first molecular analysis of the origin of dogs used mitochondrial control
region sequences to construct a phylogeny of domestic dog and grey wolf haplo-
types (Vilà et al., 1997). Four clades of dog sequences were evident in this phy-
logeny, suggesting that at least four female wolf lineages founded or were
subsequently introduced into domestic dogs (Fig. 3.3). A more extensive analysis
by Savolainen et al. (2002) suggested six or more lineages. However, the number
of actual founding lineages is uncertain because many of the initial matrilines may
have been lost through stochastic processes (Avise, 2000) and others may have
been introduced through backcrossing between dogs and female grey wolves. The
primary conclusion to be derived from these mtDNA data is that a limited
number of founding events explain the genetic diversity of dogs. This result is con-
sistent with genetic studies of other domesticated animals (cattle, sheep, goats,
pigs; Bruford et al., 2003). By contrast, horses have incorporated numerous matri-
lines from their wild progenitors, suggesting repeated domestication (or back-
crossing) events over a large geographic area (Vilà et al., 2001). However, the
limited number of founding events for dogs does not necessarily imply that the
number of founders was small.
The major histocompatibility complex (MHC) encodes proteins that are
essential to the normal functioning of the immune system and typically show a
high level of genetic diversity that is maintained by balancing selection (Hughes
and Yeager, 1998). One consequence of this form of selection is that MHC alleles
are maintained for longer periods of time than expected under neutrality.
Recently, a large number of alleles have been described for the MHC in dogs,
wolves and coyotes (Kennedy et al., 1999, 2000, 2001, 2002; Seddon and
Ellegren, 2002, 2004). Seddon and Ellegren (2002) observed that a large number
of alleles are shared between the three different species (Fig. 3.5), which suggests
that they may have existed before the species diverged (over 1 million years ago
for grey wolves and coyotes). Hence, these alleles can provide evolutionary infor-
mation over a longer time frame than mtDNA and have been used to provide
estimates of effective population sizes in humans (Ayala et al., 1994) and to calcu-
late founding population sizes for other species (Vincek et al., 1997). Predictive
models based on selection and drift can be used to estimate the minimum
numbers of wolf founders under a variety of demographic conditions (Vilà et al.,
2005).
Vilà et al. (2005) collated information from 42 MHC DRB alleles for dogs,
almost all of them differing by at least one non-synonymous change. Considering
the rate of non-synonymous substitution estimated for MHC class II loci (Klein et
al., 1993), the authors concluded that almost the totality of the alleles probably
originated before the time of domestication. Consequently, at least 21 founders
are required to explain the large MHC diversity observed in dogs. However, this
number represents a minimum estimate since it assumes that all founders are het-
erozygous for different alleles, are equally successful producing offspring and that
no alleles are removed from the population by drift. To obtain more realistic
Fig. 3.5. Phylogenetic tree of MHC DQA alleles. The species distribution of each
allele is indicated by filled circles; We, European grey wolf; Wa, North American
grey wolf; D, dog; C, coyote. Many alleles are shared between species, indicating
that the alleles may have originated before the species split. Based on Seddon and
Ellegren (2002).
estimates of the number of founders, Vilà et al. (2005) used models that varied the
number of founders, the number of founding populations, the rate of population
growth after domestication, the diversity in the ancestral population and the
strength of selection acting on the MHC locus. These simulations strongly sug-
gested that if the domestication occurred in only one population, hundreds of
individuals must have been involved in the founding of the domestic dog (Fig.
3.6). Obviously, this is not a realistic possibility at the time of the domestication
of the dog because humans were still hunter-gatherers, living in small groups, and
they lacked the resources to support a large population of domestic animals. Since
several lineages of mtDNA have been identified in dogs (Vilà et al., 1997, 1999;
Savolainen et al., 2002), one possibility was that the domestication would have
taken place in different populations. If this was the case, drift would lead to the
fixation of different alleles in each population, and a lower number of founders
would be enough to explain the level of diversity found today. Indeed, the simu-
lations show that the estimated number of founders decreased as the number of
populations involved increased (Fig. 3.6). However, even if six populations were
involved, which is more than suggested by mtDNA studies (Vilà et al., 1997, 1999;
Savolainen et al., 2002) and more than suggested for most livestock species
Number of alleles
Founder Population size

Fig. 3.6. Number of MHC alleles that would be expected in modern dogs
depending on the number of wolves involved in the founding event (divided in
1–6 populations of equal size). Each dot represents the average of 1000
simulations. The discontinuous line indicates the number of alleles described so
far in modern dogs. To obtain such a large number of alleles, a large number of
founders would be required, even in the case that dogs derived from multiple
domestication events. Modified from Vilà et al. (2005).
(Bruford et al., 2003), about 60 founders would be required. This is an underesti-

mate considering that the model assumes that all individuals are equally likely to
reproduce and, consequently, the real number of founders would probably be
several times larger (Frankham, 1995).
In conclusion, the simulations showed that the number of wolves involved in
the domestication process was larger than deduced from mtDNA alone and prob-
ably involved several hundred wolves. Second, since it seems improbable that such
a large founding population could have been kept separated from wild wolves,
hybridization between wolves and dogs was probably an important influence on the
diversity of the dog MHC. Presumably, extensive backcrossing with male wolves
early in the history of dog domestication may have enriched nuclear diversity to a
greater extent than maternally inherited mitochondrial DNA. Future research uti-
lizing male-specific markers on the Y chromosome may help clarify the situation.
Origin of Breeds
The World Canine Organization (Fédération Cynologique Internationale, FCI)
currently recognizes about 347 breeds of dogs, classified in ten groups according
to their function and, to a lesser degree, area of origin. Each one of these breeds
is characterized by a quite unique morphology and often also behaviour, which
makes them easily recognizable. The genetic basis of these morphological traits is
evident since breeds breed true (i.e. the offspring of two dalmatians looks like
another dalmatian, and the offspring of a couple of boxers will look like another
member of the same breed). Similarly, some breed-characteristic behaviours are
also inherited (Schmutz and Schmutz, 1998). This uniform typology for each
breed is very often seen as the result of a long period of isolation. In fact, many
dog breeders believe that their breeds had an origin that dates from several cen-
turies or millennia ago (Crowley and Adelman, 1998). For example, archaeolog-
ical evidence from ancient Egypt suggests that several types of morphologically
differentiated dogs similar to mastiffs and greyhounds existed there 4000 years
ago and Romans may have been the first people to develop dog breeds in Europe
(Clutton-Brock, 1999). Also, pottery from the Colima culture (250 BC–AD 450,
western Mexico) clearly represents Mexican hairless dogs (Cordy-Collins, 1994).
This breed was also found by the Spanish conquistadors on their arrival to
Mexico in the 16th century (Valadez, 1995). Furthermore, dogs depicted in
European paintings during that time could be easily recognized today as spaniels,
mastiffs, hounds, pointers, etc.
Consistent with this differentiation at the phenotypic level, multiple studies
using autosomal markers (located on nuclear chromosomes other than the sex
chromosomes X and Y) have shown that breeds are clearly differentiated (Lingaas
et al., 1996; Zajc et al., 1997; Zajc and Sampson, 1999), and that the genotype
obtained for an individual using only a limited number of markers may be suffi-
cient to assign that individual to the correct breed (Koskinen, 2003; Parker et al.,
2004; Halverson and Basten, 2005). Parker et al. (2004) typed 96 microsatellite
loci in 85 breeds, each represented by five dogs. An analysis of molecular variance
showed that variation between breeds accounted for about 27% of the genetic
diversity observed in all dogs. Similarly, they also assessed genetic diversity within
and between breeds with nuclear single nucleotide polymorphisms (SNPs). In a
survey of 19,867 base pairs of sequence from 120 dogs, the authors identified 75
SNPs, and 14 of these were breed-specific in the small number of individuals
screened per breed. These results suggest many breeds have been isolated since
their origination.
The clear differentiation of breeds in phenotype and with autosomal genetic
markers stands in stark contrast to the maternally inherited mtDNA results.
Studies based on mitochondrial DNA found no correlation between mtDNA
haplotype and breed (Vilà et al., 1997, 1999; Savolainen et al., 2002). A single
breed often had many different mtDNA haplotypes derived from different
sequence clades. For example, Vilà et al. (1999) screened 19 Mexican hairless dogs
and found seven different mtDNA haplotypes that fall into three of the four main
clades of dogs (Fig. 3.3). Also, a single mtDNA haplotype could be found across a
wide variety of breeds. For example, haplotype D3 from Vilà et al. (1997) was
identified in chow chow, Norwegian elkhound, Mexican hairless, Siberian husky,
papillon, poodle, rottweiler, English setter, Icelandic sheepdog, springer spaniel,
Japanese spitz, border terrier, fox terrier and whippet. This lack of differentiation
between breeds for mitochondrial DNA markers is surprising given the morpho-
logical uniformity within breeds compared to the large differences between
breeds and suggests that modern breeds have a recent origin from a well mixed
and genetically diverse dog population.
In order to reconcile the lack of differentiation between breeds at mtDNA
with the clear morphologic separation, Sundqvist et al. (2006) studied paternally
inherited Y chromosome markers, maternal mtDNA and autosomal markers in
purebred dogs. The authors observed, again, that breeds were genetically differ-
entiated for the autosomal markers while both mtDNA and Y chromosome were
not well separated. However, the patterns of diversity observed for these two
markers were highly informative about the process by which the breeds were
formed. When looking at the number of different Y chromosome and mtDNA
types found in five dogs from each of 20 breeds, Sundqvist et al. observed that the
number of mtDNA haplotypes was usually larger than the number of Y chromo-
some haplotypes (Table 3.1). Since these two marker systems evolve in a very dif-
ferent way, the results could be different to interpret. However, the authors also
typed the same markers in several populations of wolves (Table 3.1), where
approximately the same number of males and females reproduce every year
(Packard, 2003) and observed similar numbers of mtDNA and Y chromosome
types. The pattern observed in the wolf populations could be used as a baseline
to understand the results in dog breeds, and the authors concluded that the dif-
ferent pattern was due to the process by which dog breeds were formed. Fewer
males than females had contributed to the origin of the breeds. Since males have
a larger breeding potential than females (one male can sire a large number of
litters every year), the results suggest that breeders have traditionally selected
traits present in males to define the breed characteristics. This has resulted in the
existence of ‘popular sires’ (Ostrander and Kruglyak, 2000), dogs that, because of
their characteristics, have often been used as breeders. Some of these males may
produce over 100 litters in their life time (Ostrander and Kruglyak, 2000).
Consequently, a phenotypic trait present in one male can be successfully trans-
ferred to a larger number of offspring per generation than if that trait was in a
female. Therefore, strong selection on males allows a faster definition of the phe-
notype in a new breed.
Although the comparison of the levels of diversity for uniparentally inherited
markers within dog breeds was informative about the process by which the breeds
were formed, they were unable to provide meaningful information about the rela-
tionship between dog breeds. Attempts to reconstruct phylogenetic relationships
between breeds have been made using autosomal markers, but simple relation-
ships could not be identified when a large number of breeds were considered
(Irion et al., 2003; Parker et al., 2004). Parker et al. (2004) found that many of the
most recognizable and common breeds of European origin are so closely related
that branching patterns were difficult or impossible to elucidate. This suggests a
very recent origin for all of these breeds, and also that they share a common
recent ancestry from a diverse population of dogs. This is consistent with the
Table 3.1. Number of mtDNA and Y chromosome haplotypes observed in dog

breeds (5 males per breed) and in wolf populations (n=number of males sampled
per population). Based on Sundqvist et al. (2006).
Dog breed/wolf population mtDNA Y chromosome
Dogs
Airedale terrier 3 2
Beagle 3 1
Bernese mountain dog 2 1
Border terrier 2 1
Boxer 3 1
Cairn terrier 3 1
Cavalier King Charles spaniel 2 1
Collie, rough/smooth 3 1
Dalmatian 3 2
Flatcoated retriever 2 1
German pointer 3 1
German shepherd 3 1
Golden retriever 2 1
Greyhound 3 2
Irish soft-coated wheaten terrier 4 2
Newfoundland 1 1
Poodle, miniature/standard 4 2
Rottweiler 1 2
Shetland sheepdog 3 3
West Highland white terrier 3 2
Wolves
Alaska (n=12) 6 6
Russia (n=12) 4 6
Inuvik (n=13) 4 7
Finland (n=31) 3 7
Spain (n=20) 2 5
Baltic States (n=24) 2 9
foundation of dog breeding clubs, breed standards and modern breeding prac-
tices (Ash, 1927; Crowley and Adelman, 1998; Dennis-Bryan and Clutton-Brock,
1988; Parker et al., 2004).
The World Canine Organization (FCI) has organized breeds into ten differ-
ent groups based on function, and, to some extent, supposed origin. However, the
genetic analysis of Parker et al. (2004) defined only four fundamental genetic sub-
divisions in dogs. The most primitive and divergent groupings included a wide
diversity of breeds ranging in geographic origin from Africa (basenji) to Asia
(chow-chow and akita) to the Arctic (Siberian husky and Alaskan malamute). The
wide geographic area represented by these primitive breeds suggests that dogs
spread over a large geographic area soon after domestication. Three other breed
clusters can be defined by the genetic analysis: (i) European breeds that diverged
in the 1800s; (ii) herding dogs, including the collie and sheep dogs and similar to
FCI group 1 (sheep dogs and cattle dogs); and (iii) mastiff-like dogs, including the
bulldog and boxer. The first one of these groups (primitive dogs) may appear so
differentiated from all other breeds because the other groups include most of the
breeds of European origin and so may have a very narrow genetic basis. On the
other hand, the other groups may indicate the preferential exchange of dogs
among the breeds included within the groups. Sundqvist et al. (2006) have
observed that the breeds included in these groups (as well as in the FCI groups)
tend to have more similar Y chromosomes than mtDNA sequences, which could
indicate preferential exchange of males than females between the breeds. This
would be consistent with the use of males to define the characteristics of the
breeds and would indicate that the similarity between breeds at autosomal
microsatellites may not so much indicate a recent common origin but the occa-
sional exchange of individuals.
An understanding of the origin and evolution of modern breeds is important
to models of genetic disease and theories about the genetic basis of morphologi-
cal and behavioural variation (Wayne and Ostrander, 1999). However, our
current knowledge about the origin of dogs and about the process by which
breeds were formed which led to the astonishing diversity of modern-day dogs is
far from complete. The completion of a second dog genome sequence
during 2005 will certainly promote our understanding of the diversity of dogs and
we should expect many answers and new questions to arise during the next few
years.
Acknowledgements
Financial support was provided by a grant of the Swedish Research Council for
Environment, Agricultural Sciences and Spatial Planning to C.V. J.A.L. is sup-
ported by a grant from the Swedish Research Council.
References
Ash, E.C. (1927) Dogs: Their History and Development. E. Benn Limited, London.
Avise, J.C. (1994) Molecular Markers, Natural History and Evolution. Chapman and
Hall, New York.
Avise, J.C. (2000) Phylogeography: The History and Formation of Apecies. Harvard
University Press, Cambridge, Massachusetts.
Ayala, F.J., Escalante, A., O’Huigin, C. and Klein, J. (1994) Molecular genetics of
speciation and human origins. Proceedings of the National Academy of
Sciences, USA 91, 6787–6794.
Beja-Pereira, A., England, P.R., Ferrand, N., Jordan, S., Bakhiet, A.O., Abdalla, M.A.,
Mashkour, M., Jordana, J., Taberlet, P. and Luikart, G. (2004) African origins of
the domestic donkey. Science 304, 1781.
Bruford, M.W., Bradley, D.G. and Luikart, G. (2003) DNA markers reveal the com-
plexity of livestock domestication. Nature Reviews Genetics 4, 900–910.
Clutton-Brock, J. (1999) A Natural History of Domesticated Animals. Cambridge
University Press, Cambridge.
Coppinger, R. and Coppinger, L. (2001) Dogs. The University of Chicago Press,
Chicago, Illinois.
Cordy-Collins, A. (1994) An unshaggy dog history. Natural History 2, 34–40.
Crowley, J. and Adelman, B. (1998) The Complete Dog Book: Official Publication
of the American Kennel Club, 19th edn. Howell Book House, New York.
Dayan, T. (1994) Early domesticated dogs of the Near East. Journal of
Archaeological Science 21, 633–640.
Dennis-Bryan, K. and Clutton-Brock, J. (1988) Dogs of the Last Hundred Years at
the British Museum (Natural History). British Museum (Natural History),
London.
Ferrell, R.E., Morizot, D.C., Horn, J. and Carley, C.J. (1978) Biochemical markers in
species endangered by introgression: the red wolf. Biochemical Genetics 18,
39–49.
Frankham, R. (1995) Effective population size/adult population size ratios in wildlife:
a review. Genetical Research 66, 95–107.
Giuffra, E., Kijas, J.M., Amarger, V., Carlborg, O., Jeon, J.T. and Andersson, L. (2000)
The origin of the domestic pig: independent domestication and subsequent
introgression. Genetics 154, 1785–1791.
Halverson, J. and Basten, C. (2005) A PCR multiplex and database for forensic DNA
identification of dogs. Journal of Forensic Science 50, 352–363.
Hiendleder, S., Kaupe, B., Wassmuth, R. and Janke, A. (2002) Molecular analysis of
wild and domestic sheep questions current nomenclature and provides evi-
dence for domestication from two different subspecies. Proceedings of the
Royal Society of London, Series B 269, 893–904.
Hughes, A.L. and Yeager, M. (1998) Natural selection at major histocompatibility
complex loci of vertebrates. Annual Reviews Genetics 32, 415–435.
Irion, D.N., Schaffer, A.L., Famula, T.R., Eggleston, M.L., Hughes, S.S. and Pedersen,
N.C. (2003) Analysis of genetic variation in 28 dog breed populations with 100
microsatellite markers. Journal of Heredity 94, 81–87.
Irion, D.N., Schaffer, A.L., Grant, S., Wilton, A.N. and Pedersen, N.C. (2005)
Genetic variation analysis of the Bali street dog using microsatellites. BMC
Genetics 6, 1–6.
Kennedy, L.J., Carter, S.D., Barnes, A., Bell, S., Bennett, D., Ollier, B. and Thomson,
W. (1999) DLA-DRB1 polymorphisms in dogs defined by sequence-specific
oligonucleotide probes (SSOP). Tissue Antigens 53, 184–189.
Kennedy, L.J., Altet, L., Angles, J.M., Barnes, A., Carter, S.D., Francino, O., Gerlach,
J.A., Happ, G.M., Ollier, W.E., Polvi, A., Thomson, W. and Wagner, J.L. (2000)
Nomenclature for factors of the dog major histocompatibility system (DLA),
1998: first report of the ISAG DLA Nomenclature Committee. Animal Genetics
31, 52–61.
Kennedy, L.J., Angles, J.M., Barnes, A., Carter, S.D., Francino, O., Gerlach, J.A.,
Happ, G.M., Ollier, W.E., Thomson, W. and Wagner, J.L. (2001) Nomenclature
for factors of the dog major histocompatibility system (DLA), 2000: second
report of the ISAG DLA Nomenclature Committee. Animal Genetics 32,
193–199.
Kennedy, L.J., Barnes, A., Happ, G.M., Quinnell, R.J., Bennett, D., Angles, J.M.,
Day, M.J., Carmichael, N., Innes, J.F., Isherwood, D., Carter, S.D., Thomson,
W. and Ollier, W.E. (2002) Extensive interbreed, but minimal intrabreed, vari-
ation of DLA class II alleles and haplotypes in dogs. Tissue Antigens 59,
194–204.
Klein, J., Satta, Y., O’hUigin, C. and Takahata, N. (1993) The molecular descent of
the major histocompatibility complex. Annual Review of Immunology 11,
269–295.
Koskinen, M.T. (2003) Individual assignment using microsatellite DNA reveals
unambiguous breed identification in the domestic dog. Animal Genetics 34,
297–301.
Kurtén, B. and Anderson, E. (1981) Pleistocene Mammals of North America.
Columbia University Press, New York.
Lau, C.H., Drinkwater, R.D., Yusoff, K., Tan, S.G., Hetzel, D.J. and Barker, J.S.
(1998) Genetic diversity of Asian water buffalo (Bubalus bubalis): mitochondr-
ial D-loop and cytochrome b sequence variation. Animal Genetics 29,
253–264.
Leonard, J.A., Wayne, R.K., Wheeler, J., Valadez, R., Guillen, S. and Vilà, C. (2002)
Ancient DNA evidence for Old World origin of New World dogs. Science 298,
1613–1616.
Lingaas, F., Aarskaug, T., Sorensen, A., Moe, L. and Sundgreen, P.-E. (1996)
Estimates of genetic variation in dogs based on microsatellite polymorphism.
Animal Genetics 27, 29.
Luikart, G., Gielly, L., Excoffier, L., Vigne, J.D., Bouvet, J. and Taberlet, P. (2001)
Multiple maternal origins and weak phylogeographic structure in domestic
goats. Proceedings of the National Academy of Sciences, USA 98, 5927–5932.
Morey, D.F. (1992) Size, shape, and development in the evolution of the domestic
dog. Journal of Archaeological Science 19, 181–204.
Nobis, G. (1979) Der älteste Haushund lebte vor 14,000 Jahren. Umschau 19, 610.
Okumura, N., Ishiguro, N., Nakano, M., Matsui, A. and Sahara, M. (1996) Intra- and
interbreed genetic variations of mitochondrial DNA major non-coding regions
in Japanese native dog breeds (Canis familiaris). Animal Genetics 27, 397–405.
Olsen, S.J. (1985) Origins of the domestic dog. University of Arizona Press, Tucson,
Arizona.
Olsen, S.J. and Olsen, J.W. (1977) The Chinese wolf ancestor of New World dogs.
Science 197, 533–535.
Ostrander, E.A. and Kruglyak, L. (2000) Unleashing the canine genome. Genome
Research 10, 1271–1274.
Packard, J.M. (2003) Wolf behavior: reproductive, social and intelligent. In: Mech,
L.D. and Boitani, L. (eds) Wolves. Behavior, Ecology, and Conservation.
University of Chicago Press, Chicago, Illinois, pp. 35–65.
Parker, H.G., Kim, L.V., Sutter, N.B., Carlson, S., Lorentzen, T.D., Malek, T.B.,
Johnson, G.S., DeFrance, H.B., Ostrander, E.A. and Kruglyak, L. (2004) Genetic
structure of the purebred domestic dog. Science 304, 1160–1166.
Sablin, M.V. and Khlopachev, G.A. (2002) The earliest Ice Age dogs: evidence from
Eliseevichi. Current Anthropology 43, 795–799.
Sarich, V.M. (1977) Albumin phylogenetics. In: Rosenoer, V.M., Oratz, M. and
Rothschild, M.A. (eds) Albumin Structure, Function and Uses. Pergamon Press,
New York, pp. 85–111.
Savolainen, P., Zhang, Y.P., Luo, J., Lundeberg, J. and Leitner, T. (2002) Genetic
evidence for an East Asian origin of domestic dogs. Science 298,
1610–1613.
Schmutz, S.M. and Schmutz, J.K. (1998) Heritability estimates of behaviors associ-
ated with hunting in dogs. Journal of Heredity 89, 233–237.
Seal, U.S., Phillips, N.I. and Erickson, A.W. (1970) Carnivora systematics: immuno-
logical relationships of bear albumins. Comparative Biochemistry and
Physiology 32, 33–48.
Seddon, J.M. and Ellegren, H. (2002) MHC class II genes in European wolves: a com-
parison with dogs. Immunogenetics 54, 490–500.
Seddon, J.M. and Ellegren, H. (2004) A temporal analysis shows major histocom-
patibility complex loci in the Scandinavian wolf population are consistent with
neutral evolution. Proceedings of the Royal Society of London, series B 271,
2283–2291.
Sharma, D.K., Maldonado, J.E., Jhala, Y.V. and Fleischer, R.C. (2003) Ancient wolf
lineages in India. Proceedings of the Royal Society of London, series B 271,
S1–4.
Sundqvist, A.-K., Björnerfeldt, S., Leonard, J.A., Hailer, F., Hedhammar, Å., Ellegren,
H. and Vilà, C. (2006) Unequal contribution of sexes in the origin of dog
breeds. Genetics 172, 1121–1128.
Troy, C.S., MacHugh, D.E., Bailey, J.F., Magee, D.A., Loftus, R.T., Cunningham, P.,
Chamberlain, A.T., Sykes, B.C. and Bradley, D.G. (2001) Genetic evidence for
Near-Eastern origins of European cattle. Nature 410, 1088–1091.
ing occurred among multiple matriarchal ancestors during the domestication of
region of mitochondrial DNA between dogs and wolves. Genes and Genetic
Systems 72, 229–238.
Valadez, R. (1995) El perro mexicano. Instituto de Investigaciones Antropológicas,
Universidad Nacional Autónoma de México, Mexico.
Vilà, C., Maldonado, J. and Wayne, R.K. (1999) Phylogenetic relationships, evolu-
tion and genetic diversity of the domestic dog. Journal of Heredity 90, 71–77.
Vilà, C., Leonard, J.A., Götherström, A., Marklund, S., Sandberg, K., Lidén, K.,
Wayne, R.K. and Ellegren, H. (2001) Widespread origins of domestic horse lin-
eages. Science 291, 474–477.
Vilà, C., Seddon, J. and Ellegren, H. (2005) Genes of domestic mammals augmented
by backcrossing with wild ancestors. Trends in Genetics 21, 214–218.
Vincek, V., O’Huigin, C., Satta, Y., Takahata, Y., Boag, P.T., Grant, P.R., Grant, B.R.
and Klein, J. (1997) How large was the founding population of Darwin’s
finches? Proceedings of the Royal Society of London, series B 264, 111–118.
Wayne, R.K. (1986a) Cranial morphology of domestic and wild canids: the influence
of development on morphological change. Evolution 4, 243–261.
Wayne, R.K. (1986b) Limb morphology of domestic and wild canids: the influence
of development on morphologic change. Journal of Morphology 187, 301–319.
Wayne, R.K. and O’Brien, S.J. (1987) Allozyme divergence within the Canidae.
Systematic Zoology 36, 339–355.
Wayne, R.K. and Ostrander, E.A. (1999) Origin, genetic diversity, and genome struc-
ture of the domestic dog. Bioessays 21, 247–257.
Wayne, R.K., Nash, W.G. and O’Brien, S.J. (1987a) Chromosomal evolution of the
Canidae: I. Species with high diploid numbers. Cytogenetics and Cell Genetics
44, 123–133.
Wayne, R.K., Nash, W.G. and O’Brien, S.J. (1987b) Chromosomal evolution of the
Canidae: II. Divergence from the primitive carnivore karyotype. Cytogenetics
and Cell Genetics 44, 134–141.
Wayne, R.K., Benveniste, R.E. and O’Brien, S.J. (1989) Molecular and biochemical
evolution of the Carnivora. Molecular and biochemical evolution of the
Carnivora. In: Gittleman, J.L. (ed.) Carnivore Behavior, Ecology and Evolution.
Cornell University Press, Ithaca, New York, pp. 465–494.
Wayne, R.K., Geffen, E., Girman, D.J., Koepfli, K.P., Lau, L.M. and Marshall, C.R.
(1997) Molecular systematics of the Canidae. Systematic Biology 46, 622–653.
Wurster-Hill, D.H. and Centerwall, W.R. (1982) The interrelationships of chromo-
some banding patterns in canids, mustelids, hyena, and felids. Cytogenetics and
Cell Genetics 34, 178–192.
Zajc, I. and Sampson, J. (1999) Utility of canine microsatellites in revealing the rela-
tionships of pure bred dogs. Journal of Heredity 90, 104–107.
Zajc, I., Mellersh, C.S. and Sampson, J. (1997) Variability of canine microsatellites
within and between different dog breeds. Mammalian Genome 8, 182–185.
Zeder, M. (2003) Hiding in plain sight: the value of museum collections in the study
of the origins of animal domestication. In: Grupe, G. and Peters, J. (eds)
Decyphering Ancient Bones: the Research Potential of Bioarchaeological
Collections. Rahden/Westf., Leidorf, pp. 125–138.
Zeder, M.A. and Hesse, B. (2000) The initial domestication of goats (Capra hircus)
in the Zagros mountains 10,000 years ago. Science 287, 2254–2257.
II Biology and Behaviour of Dogs
Introduction
In this part of the book, some fundamental aspects of the behavioural biology of
dogs are covered. Chapter 4 deals with the ethological foundation for animal
behaviour. The way any animal behaves is a function of its neural system and the
way the muscles are controlled by this, both aspects being under heavy evolu-
tionary selection. To understand this evolutionary aspect of behaviour properly,
we need to consider how genes may affect behaviour, and this is the topic of
Chapter 5. Here we get a broad overview of behaviour genetics including some
of the most powerful methods available today to dissect the mechanisms allowing
particular genes to affect specific behaviour.
Furthermore, to really understand the behaviour of dogs, we need to be able
to perceive the world through their senses. To help with this, Chapter 6 provides
an overview of different sensory organs of mammals, their construction and func-
tion, with a special emphasis on dogs. These senses help dogs orienting, hunting,
etc., but are of course also crucial in their social interactions. Dogs are profoundly
social animals, and Chapter 7 puts this aspect of dog behaviour into context by
providing extensive comparisons with other closely related canids.
The last chapter of this section, Chapter 8, is devoted to learning. Being able
to modify and consistently affect the behaviour of dogs is a prerequisite for many
aspects of life with dogs today. This is therefore one of the longest chapters of the
book, and it contains a comprehensive overview of modern learning theory,
applied to everyday examples of dog learning. The chapter therefore provides an
excellent bridge to the third part of the book, which is concerned with dogs
among humans.
Mechanisms and Function in Dog
4
Behaviour
Per Jensen
Introduction
Ethology is the science in which we study animal behaviour, its causation and its bio-
logical function, and this entire book is devoted to the behavioural biology of dogs.
But it is not entirely clear what we understand by behaviour in the first place, and
some examples may help to reflect on this problem of definition of the subject. In its
simplest form, behaviour may be series of muscle contractions, perhaps performed
in clear response to a specific stimulus, such as in the case of a reflex – for example,
a dog scratching itself with the hindleg. However, in the other extreme, we find very
complex activities, such as a pack of wolves seizing a prey, continuously assessing
their directions and positions with the help of various cues from the environment,
from the behaviour of the prey and from the rest of the pack.
We would use the word behaviour for both of these extremes, and for many
other activities in between in complexity. It will include all types of activities that
animals engage in, such as locomotion, grooming, reproduction, caring for
young, communication, etc. Behaviour may involve one individual reacting to a
stimulus or a physiological change, but may also involve two individuals, each
responding to the activities of the other. This makes the science of ethology a
complex one, and there is a need to structure the way in which to approach the
huge task of analysing the biology of behaviour.
Since behaviour is such a complex biological phenomenon, it follows that it
can be studied from a number of different perspectives. Any science dealing with
behaviour must therefore limit its purpose, or formalize the framework of the
science. The most influential formulation of a research programme for ethology
was put forward by the Nobel Prize winner Niko Tinbergen (Tinbergen, 1963).
© CAB International 2007. The Behavioural Biology of Dogs 61

(ed. P. Jensen)
62 P. Jensen
This programme is frequently referred to as ‘Tinbergen’s four questions’, and it

outlines the specific aspects that he felt were ethology’s central task to study.
Assume that we are interested in studying the behaviour of barking in dogs (see
also Chapter 12 in this volume for a more thorough discussion of barking). What
sorts of questions should we pose? According to Tinbergen, ethology can ask four
different types of questions, which are not mutually exclusive. A complete etho-
logical understanding of barking requires that we provide answers on all of them,
albeit not necessarily in one and the same study:
1. What causes barking? The answer to this question refers to the immediate
causes, such as which stimuli elicit or stimulate a particular behaviour. In this
question, we may also ask which physiological variables, such as hormones, cause
or mediate the performance of barking.
2. What is the function of barking? In this case, the answer describes how the
performance of a behaviour affects the reproductive success, the fitness, of the
dog. It therefore has to do with evolutionary aspects and consequences.
3. How does barking develop during ontogeny? Studies of this question aim at
describing the way a behaviour is modified by individual experiences. It will also
cover questions such as how barking matures, and how it changes with the age of
the dog, and how it is affected by learning.
4. How has barking developed during phylogeny? This is a clearly evolutionary
question, and usually calls for comparative studies of related species. It is of course
not possible to examine the vocal behaviour of fossils, but comparing vocalizations
of different now living, related canids, may help to understand when the behaviour
developed and how it has been preserved in different lineages of species.
The interested dog biologist – who is likely to have a close personal relation
with dogs – might object that Tinbergen’s questions are incomplete, since they do
not consider what animals perceive, feel and know in relation to their own behav-
iour, i.e. the emotional and cognitive aspects of behaviour. This aspect of animal
behaviour was long considered to be inaccessible to science. However, over the
last decades, scientists have developed methods and concepts to allow investiga-
tion into this area. This has led to a new branch of the science, emerging in the
1970s, known as cognitive ethology (Bekoff, 2000). In this book, Chapter 12 is
largely devoted to these aspects of dog behaviour. However, proper understand-
ing of dog behavioural biology requires that the Tinbergen questions are exam-
ined systematically and, based on the answers to them, we can proceed and pose
new questions regarding cognitive aspects of behaviour.
Nature and Nurture
An important question concerns the relative contribution of inherited traits

(genetic factors) and acquired traits (for example, learned) in the expression of a
particular behaviour. This is sometimes referred to as the nature versus nurture
debate. Chapter 5 in this book is particularly concerned with the genetic aspects
Mechanisms and Function in Behaviour 63
of dog behaviour. Here, I will just emphasize that all behaviour, whether we refer
to it as innate or learned, is necessarily always a product of both genetic and envi-
ronmental factors, i.e. it is always a question of nature AND nurture. This double
necessity is particularly evident in some developmental processes, such as the mat-
uration and socialization of dog pups, which is treated later in this chapter.
We can easily understand the wider aspects of the ‘nature-and-nurture view’
by considering a typical learned behaviour, such as retrieving game. In order for
the dog to learn to retrieve on command, it needs some basic behavioural dispo-
sition to chase and carry objects. This disposition can be selected for, as we know
from specialized retrieving breeds such as labradors, so obviously there is a
genetic basis for this behaviour. However, we can teach the dog to connect a
signal (for example, a spoken command) with the performance of the behaviour,
which is an acquired process – no dog is born with the knowledge that ‘fetch’
means ‘run and fetch that dead bird and bring it to me’. In order to be able to
form this connection, the dog needs sensory organs that can perceive the signal
(ears in this case), a nervous system that can process it and form the connection
between the correct behavioural response and a reward, and a locomotory system
(muscles and joints) that allows it to retrieve. All these features are constrained by
genetic factors. Therefore, any behaviour will always be the result of an interac-
tion between genetic predispositions and environmental variables.
Behaviour is a Brain-product
Imagine a dog performing a typical canine behaviour: it carries a bone away from
the family, digs a hole in the soil, drops the bone in the hole and covers it with soil,
using the nose to push the soil in place. It is self-evident that the behaviour of the
dog is controlled by the brain, but how? Information reaches the brain from the
sense organs (smell and shape of the object, visual information from the surround-
ings) and is interpreted there in some meaningful way to the dog (‘the stimuli rep-
resent a bone’; ‘this is a suitable place to dig’). Following this interpretation, the
brain sends nerve impulses to the appropriate muscles, which make the dog move
and start digging. From watching many dogs performing this type of behaviour,
one can easily conclude that the pattern is similar for different members of the
species: holes are normally dug with specific movements with the forelegs, the bone
is normally always covered with the help of the nose. The reason is that the
complex of movements is guided by specific programmes by which the brain con-
trols motor activity. If we want to understand the mechanisms of dog behaviour, it
therefore seems reasonable to start exploring the brain and the nervous system.
The Brain and the Nervous System
In all vertebrates, dogs included, the brain develops in the embryo from the so-
called neural tube (Alcock, 2001). Early, it differentiates into three different parts,
64 P. Jensen
the forebrain, midbrain and hindbrain. Somewhat later, the forebrain differenti-
ates further into the telencephalon and the diencephalon, the hindbrain differen-
tiates into metencephalon and myelencephalon, and the midbrain stays as a
functionally intact unit in the adult (the mesencephalon). Of course, the relative
specialization of the different brain parts varies between species. In mammals,
telencephalon develops into cerebrum, which includes the cerebral cortex, and
this structure is extremely variable between species. It is the part of the brain con-
taining centres for integration of sensory stimuli, and for conscious reasoning and
reflection – in particular a part of the cortex called the neocortex, which is unique
to mammals. Humans have the most developed cerebral cortex, and the neocor-
tex makes up about 80% of our total brain mass. Dogs, like other mammals, have
the same structures, but they constitute a relatively much smaller part of the
brain.
One part of the cerebral cortex is common to all mammals (and reptiles), and
is referred to as the limbic system (Alcock, 2001). This consists of the hippocam-
pus and the olfactory cortex (where olfactory input is integrated, see Chapter 6),
and parts of the thalamus and hypothalamus of the diencephalon. This region of
the brain contains centres involved in the immediate control of basic behavioural
programmes, related to feeding, aggression and sexual behaviour. The limbic
system also connects to sensory areas in the neocortex and is responsible for
attaching emotions, feelings, to behaviours. Human patients receiving electro-
stimulation in different nuclei of the limbic system report sudden feelings of fear
or anger, and so on, depending on the exact localization of the stimulation
(Carlson, 1981).
Whereas dogs differ considerably from humans in the structure and size of
the cerebral cortex and neocortex, the limbic system is very similar both in struc-
ture and relative size. This is, of course, a strong neurobiological argument for the
assumption that dogs are able to perceive more or less the same range of feelings
as we do, but have a limited ability to reflect consciously on these feelings.
Neurons and their Organization

The information which the brain receives is carried via neurons, specialized cells
which are able to propagate an electric current, called an action potential
(Carlson, 1981). Neurons connect to each other via a so-called synapse. A synapse
may either be excitatory, meaning that the action potential in the first neuron
elicits a similar action potential in the second, or inhibitory, in which case the
action potential of the first neuron prevents the second from releasing any elec-
trical activity. The central nervous system (CNS) consists of the brain and the
spinal cord, both of which are made up of billions of interconnected neurons,
which in turn are organized into different nuclei and areas specialized in handling
certain types of information processing.
The part of the nervous system not included in the brain and spinal cord is
called the peripheral nervous system (PNS). This is the system which receives
information from sensory organs of different kinds. Some of them monitor events
outside the body, and these are treated in more detail in Chapter 6. Others
monitor internal events in the body, such as blood glucose levels, gut extension
and muscle tension, sends this information to the brain and allows the CNS to
consistently maintain a scheme of the state of the body and its surroundings.
The PNS also has outgoing, efferent, neurons, which innervate muscles and
other organs. This division of PNS is divided into two separate systems, the
somatic and the autonomic systems. The somatic system controls the aspects of
motor activity which humans can consciously affect, for example the skeletal
muscles. The autonomic system is responsible for the control of smooth and
cardiac muscles, and the gastrointestinal and endocrine systems, and can usually
not be controlled consciously. The autonomic system is further divided into the
sympathetic and the parasympathetic branch, which together determine the
degree of arousal and tension of the body. Neurons from both branches innervate
most organs of the body, and the relative strength of activation of each of the
branches determines the effect on the organ. For example, the sympathetic
neurons increase heart rate, inhibit digestion and increase breathing; parasympa-
thetic neurons do the opposite (Fig. 4.1).
Fig. 4.1. An overview of the innervation of different organs by sympathetic (solid

lines) and parasympathetic (dotted lines) parts of the autonomic nervous system. A
particular organ is normally innervated by both sympathetic and parasympathetic
neurons, which exert opposite effects – if one stimulates the activity of the organ,
the other inhibits. In this picture only a few examples of innervated organs are
shown: salivary glands, bronchs, heart, stomach, pancreas, adrenal medulla,
intestines, sexual organs and bladder.
66 P. Jensen
The protrusions of neurons (axons) from brain and spinal cord are collected
in bundles, nerves, which are often embedded in myelin, a protein which protects
the neurons and increases transmission speed of action potentials. Some neuron
bundles are not myelinated, and have a slower transmission rate.
Stimuli
Stimuli from the surroundings constantly flow over every individual. A small part
of all available information is recorded by specialized sense organs, and is trans-
mitted to the brain in the form of visual, olfactory, auditory and mechanic stimuli.
The fraction of the available information that is detected and perceived by an
animal is first limited by the structure and function of the sensory organs – this is
dealt with extensively in Chapter 6.
When the stimuli reach the brain, they are interpreted in a meaningful way
to the animal. The male dog brain may identify a collection of olfactory stimuli
as, for example, a female in heat, a potential prey, or the territorial mark of
an intruding male. Interestingly, the interpretative patterns are to a large
extent innate and do not have to be learned. This is the basis of so-called key
stimuli – a collection of stimuli is automatically interpreted in a specific, functional
way. A male three-spined stickleback (a small fish) reacts to anything roundish with
a red ventral part as if it was an intruder – the key stimulus for a rival is simply
‘roundish and red at the bottom’. Similarly, dogs do not need specific training to
identify the meaning of social signals such as a wagging tail. On the other hand, it
takes extensive experience before a dog can learn that a cat wagging its tail is not
in a friendly mood, since this violates the innate interpretation pattern of this
signal.
Key stimuli are usually linked to specific behavioural responses. A small,
furry object which moves away from the dog very quickly will release chasing
behaviour. Racing dogs, such as whippets, are fooled into running fast by the key
stimuli of a prey attached to an arm on an engine. A dog can learn not to react
to a key stimulus, but if not specifically trained, they will behave according to the
innate interpretation of the sensory input.
The Motor Apparatus
Motor activity is the central aspect of all behaviour. It occurs as a result of syn-
chronized pathways of muscle contractions, which cause the animal to move in a
functional manner in relation to a particular stimulus. The simplest forms of coor-
dinated muscle activity are commonly referred to as reflexes, and these were dis-
covered and studied extensively in dogs by the Russian physiologist Pavlov (see
Chapter 8 for further discussion of Pavlov’s findings). A reflex is a simple, non-
variable, motor response to a specific stimulus (Toates, 2002). Usually, the
reflex involves neurons that do not necessarily connect to the CNS, so
Mylohyoideus
Geniohyoideus
Posterior tongue
Palatopharyngeus
Superior constrictor
Thyrohyoideus
Thyroarytenoideus
Middle constrictor
Cricothyroideus
Inferior constrictor
Diaphragm
Fig. 4.2. The sequence of contractions of a number of muscles, highly coordinated

in order to form the swallowing reflex of the dog. Reprinted with permission from
McGraw Hill.
the reflex motion can be carried out without any conscious awareness or
feelings. In humans, the knee-jerk reflex is a common example, and in dogs, the
swallowing reflex has been extensively studied (Fig. 4.2). Although simple, the
reflex may involve a large number of muscles which contract and relax in a highly
coordinated and patterned fashion, as can be seen in the swallowing reflex of
dogs.
A somewhat more complicated pattern of muscular activity is the so-called
modal action pattern (MAP), often referred to as the fixed action pattern. MAPs
are again highly structured, but usually more complex than reflexes. They nor-
mally involve the CNS, but once elicited, the full motor programme proceeds
with minimal interference from the CNS. This means that the MAP usually pro-
ceeds to the end once it has been released. For example, a dog will often attempt
to roll and rub in substances with a strong smell. Once the behaviour has been
released, it will proceed largely in the same manner every time; first the dog bends
68 P. Jensen
one foreleg, then it places the shoulder part against the odorous substance and
finally rolls over on its back while wriggling the body. The neural programme
controlling the behaviour is tightly wired, which makes it virtually impossible for
the dog to perform it in a different way (for example, a dog normally can not
decide to rub its belly in the smell instead). However, there is some flexibility built
into the control system, for example allowing the dog to vary duration and inten-
sity, or to interrupt the behaviour if strongly stimulated – that is why the pattern
is preferably called ‘modal’ rather than ‘fixed’ (Toates, 2002).
The total motor output of a dog, or any animal, is to a large extent a complex
of more or less flexible combinations of series of reflexes and MAPs. Taken
together, this can be referred to as ‘motor programmes’, a term which implies that
there is certain level of neural organization of the order and structure of the
muscle contractions, but also that there is some flexibility to allow the animal to
adjust the motor output to the prevailing circumstances.
In summary, this means that what an animal can do is highly constrained by
the organization of the neural system and the muscles. This is one important
reason for why animals have clearly identifiable species-specific behaviour. The
dog which we met earlier will have great problems in deciding to cover its bone
with soil in any other manner than by using its nose – a cat would rather use its
forelegs to scratch soil over something it attempts to cover.
Endocrine Moderation of Behaviour
As we have seen, the brain interprets incoming stimuli and controls motor pro-
grammes. A particular set of stimuli (key stimuli) send, to give rise to a particular
motor programme. However, this system is controlled in all parts by the chemi-
cal environment of the body. Hormones moderate and affect behavioural output
in a number of ways, and they are all part of the body’s endocrine system (Toates,
2002).
Hormones of various types are secreted from glands in different parts of the
body, and they have a variety of biological functions. Some control metabolic
pathways and gastro-intestinal activity, others affect sex organs and yet others are
linked to the function of the immune system. Most hormones serve many differ-
ent purposes, but in this chapter we are mostly concerned about those hormones
that affect behaviour in various ways.
Some of the most important hormones in this respect are those which are
controlled by, or secreted directly by, the hypophysis (pituitary). This gland is
located just under the brain, with nerve projections directly from the hypothala-
mus. It secretes hormones which have a stimulatory effect on other glands in the
body, and thereby has a central regulatory role in the endocrine system. For
example, the hypophyseal hormone FSH (follicle-stimulating hormone) stimu-
lates the testes of male dogs and causes them to increase testosterone secretion.
Testosterone has a broad behavioural effect on the dog; for example, it increases
its sexual responsiveness and lowers its threshold for aggression (Toates, 2002).
Another hypophyseal hormone, oxytocin, exerts a direct behavioural effect by

calming the dog (Uvnas Moberg, 1994).
Hormones do not cause or inhibit behaviour by themselves. Rather, they
affect the sensitivity of the neural pathways involved in different behaviours. For
example, testosterone sensitizes the responsiveness for stimuli which are related
both to aggression and sexual behaviour. In the male dog, less olfactory stimuli
from females in heat, and less visual stimuli from a rival are needed to start
the sexual and aggressive motor programmes when testosterone levels are
high. Furthermore, testosterone strengthens the motor output, so sexual and
aggressive behaviour is performed at a higher intensity when the hormone levels
are higher.
Although the endocrine system is dynamic and can change dramatically over
a short time, some effects are more lasting than others. For example, testosterone
and oestrogen have several important functions in modulating the nervous system
during development. A male embryo which is exposed to high levels
of oestrogen in the uterus will become feminized in a number of ways (vom
Saal et al., 1983). For example, it will generally be less aggressive when adult.
This modulation of the nervous system goes on at least up to puberty, and this
is the reason why aggressive male dogs castrated late in life may not respond as
the owner wishes. When the dog approaches puberty, the nervous system is to
some extent already ‘calibrated’ on a certain aggression threshold, and the
daily fluctuations in testosterone levels become less important for the level of
aggressiveness.
Development of Behaviour
A newborn dog is a rather helpless creature – the eyes are closed, acoustic abili-
ties utterly limited, and the olfactory system very immature. The motor pro-
grammes are limited to a few essential movement patterns necessary to obtain
food and get rid of metabolic waste products. However, within a few weeks, the
dog has developed a wide array of perception abilities and motor skills, and a few
weeks later it has finished some of the most important developmental pathways in
its life. This developmental process is a beautiful example of how genetics and
experience are both necessary parts of normal behavioural development. The
process is often divided into four periods: the neonatal period, the transition
period, the socialization period and the juvenile period (Braastad and Bakken,
2002).
During the neonatal period, the pups are completely dependent on their
mother. They react to tactile stimuli and possibly smell. Movement is undevel-
oped and consists of slow crawling and head oscillations. They also emit whining
or yelping vocalizations.
During the transition period, the adult behaviour patterns start to develop.
The eye-opening at approximately 13 days of age marks the start of the period,
which ends when the ear channels open at 18–20 days. During the transition
70 P. Jensen
period, the motor abilities develop, and eventually the pup will start walking.
Eliminative behaviour no longer requires tactile stimulation from the mother.
Play-fighting and tail-wagging also start to develop during this period, and the
vocalizations become more variable.
During the period of socialization between 3 and 8 weeks, the young dog
starts to display most adult behaviour patterns. During this period, the first signs
of fear can also be observed. Social behaviour develops, as can be seen by the ten-
dency to coordinate activities, and the first signs of aggression show up. During
this phase, social bonds with other pups and with the mother and humans
develop. The most sensitive period in this respect is 4–8 weeks of age. If the pups
are deprived of human contact during this period, they may be extremely difficult
to tame later on. During this limited time, the nervous system is predisposed for
learning what a familiar flock and family member looks and smells like, and this
capacity is largely lost after the socialization period.
By 8 months of age, the dog is more or less fully grown. During the juvenile
phase, which follows after the socialization period, basic behaviour is relatively
invariant, although motor abilities are improved.
Between 6 and 14 months of age, dogs of most breeds enter sexual matura-
tion. Sexual as well as aggressive behaviour develops rapidly. The male dog starts
performing raised leg urinations (RLU) and regardless of sex, social rank disputes
are common.
Development and Calibration of the Behavioural System
The processes going on during pup development can be formalized with the help
of a developmental model suggested by the Canadian scientist Jerry Hogan
(Hogan, 1988). Following the reasoning from the previous parts of this chapter,
the behaviour system can be said to be made up of three different parts: the per-
ceptional system, the motor system and the central system (Fig. 4.3).
The reaction norms of some parts of each of these systems are genetically
predetermined, whereas others require environmental input, learning. For
example, the newborn dog probably has no specific knowledge of the concept of
‘mother’, this must be learnt. But the pathways for which stimuli that should be
combined into a ‘mother’-concept are predetermined, so the pup will specifically
be sensitive for certain combinations of sensations. As soon as it opens it eyes it
will start looking for stimuli complexes that make sense and eventually this
complex will be assigned the concept of ‘mother’. Note that the genetic predispo-
sitions steer in considerable detail what and when to learn – it is therefore impos-
sible to call the developmental process either genetic or acquired – it is both at the
same time.
In the same way, some motor output is pre-wired at birth, whereas other
output depends on training and acquisition of skills. This ensures that the pattern
which develops is both species-specific and adapted to the precise circumstances
under which the individual grows up.
Central
mechanisms
Stimuli Behaviour
Perceptual Motor
mechanisms mechanisms
Fig. 4.3. A conceptual model for a behaviour system. Stimuli from the external
world are analysed by perceptual mechanisms, which are integrated by central
mechanisms or relayed directly to motor mechanisms, the output of which equals
behaviour. Central mechanisms may affect several behaviour outputs and be
related to several types of sensory input. Redrawn after Hogan (1988).
Motivation and Conflict
Now that we have examined how a particular behaviour develops and is con-
trolled, what remains to be investigated is the central process whereby a dog
determines whether or not a particular stimulus is salient enough to warrant
action. This will, of course, also depend on which other stimuli impinge on the
dog at the moment, and which other activities it is occupied by. Simply spoken,
the dog needs to prioritize its activities.
The processes involved in making these kinds of decision are usually referred
to as motivation (Toates, 2002). The concept can be loosely defined as the likeli-
hood that an animal will perform a particular behaviour at a certain time. For
example, the probability that a dog will start drinking in a certain time period
could be taken as a measure of its drinking motivation, and the probability that it
will ingest food, of its feeding motivation. Often, motivation is used with a wider
72 P. Jensen
meaning to include all processes and rules that determine an animal’s decision of
whether to engage in a particular behaviour or not.
If we could watch a dog in a situation where only stimuli relevant for one par-
ticular behaviour system is present, we might be able to measure its motivation
for that behaviour rather precisely – it would be an inverse function of the neces-
sary strength of stimuli needed for releasing a response. For example, imagine
that a dog is kept in a closed, evenly lit room without windows. A bowl of food is
placed in front of the dog, and we want to measure how long it takes before it
starts feeding as a measure of its feeding motivation. If we had been able to keep
all other stimuli than food constant, we might then achieve a good measurement.
However, we are not likely to succeed. Even in this highly controlled situa-
tion, the dog is exposed to a series of conflicting motivations. First of all, it is
alone, so the motivation to gain social contact with other dogs or humans might
be high. Second, we cannot stop physiological processes, so the dog may be dehy-
drated and have a high motivation to drink. Similarly it may be motivated to
explore the room, to escape the situation or to test its social status in relation to
the person providing the food bowl. So, even in this highly artificial situation, the
dog’s behaviour is a result of the solution of a conflict between many different
competing motivational systems.
Motivations are thought to be neurally arranged in a hierarchical fashion
(Fig. 4.4). On the top level, the central motivational decisions are made, such as
whether to be active or to sleep. Once a high-level decision has been made, the
lower levels follow in a more or less automatic fashion. When the dog has decided
to follow its motivation to feed, other motivations are closed down, and the
feeding system has the exclusive control over the motor system until some other
motivation takes over.
Whereas the strength of some motivational systems are mainly a function of
the prevailing sensory input, others fluctuate independent of stimuli. These moti-
vations form the basis of rhythmical behaviour. For example, sleep motivation
follows a diurnal rhythm – even in a room with light all day, a dog will settle into
a regular diurnal sleep–wake rhythm (Toates, 2002). Some motivational systems
are to a certain extent self-regulated – the motor system will be activated at
regular intervals, or at a certain time since it was last active (Hogan, 1988). This
has the consequence that animals will have an urge to perform some behaviour
patterns regularly even in the absence of relevant stimuli. In dogs, chasing and
biting may be partly regulated in this way. Behaviours with this type of motiva-
tional control are sometimes referred to as ethological or behavioural needs, and
the failure to allow for these needs may cause development of abnormal behav-
iours of different kinds (Jensen and Toates, 1993) (see also Chapter 13).
Stress and Stereotypies
Biological stress is one of the most well investigated physiological and behavioural
phenomena, and it has a high significance for dog welfare. Stress can be defined
Highest level
(reproduction)
Intermediate level
(copulation)
Muscle level
(pelvic muscles)
Fig. 4.4. Tinbergen’s view of the hierarchical organization of behaviour, in this

example applied to the reproductive behaviour of dogs. Behaviour is thought to
be controlled on different hierarchical levels, and within each level, the systems
are mutually inhibitive. If, for example, reproductive behaviour is in control at the
top level, other systems, such as sleep, feeding etc., will be inhibited. The same
applies on all levels, so the reproductive system controls the use of the individual
muscles, for example to allow the dog to copulate.
as the state of an animal when it is not able to act in accordance with its motiva-
tional state (Jensen and Toates, 1997). In this situation, there will normally be ele-
vated levels of cortisol, a hormone secreted by the adrenal cortex following
stimulation by the hypophysis hormone ACTH. There will also be an increased
activity in the sympathetic nervous system, which is associated with increased
heart rate and blood pressure, plus a number of other sympathetically regulated
states.
Dogs suffering prolonged periods of high stress will run the risk of develop-
ing various diseases. For example, the immune system may be down-regulated,
increasing the susceptibility for infections, and the cardiovascular system may be
damaged. When a stress situation is chronic or persistent, dogs, like other animals,
will try alternative ways to cope with the stress situations, and they may develop
various behaviour disorders. A more comprehensive description of behaviour
disorders in dogs, and their causes, can be found in Chapter 13.
A typical class of stress-induced behaviour disorders is stereotypies, defined
as movements which are repeated over and over again, without any obvious func-
tion (Lawrence and Rushen, 1993). These abnormal behaviours are common
in zoos, on farms and in laboratory units, where animals often have limited
possibilities to express some highly motivated behaviours. They are also not
74 P. Jensen
uncommon in dogs living under conditions which do not allow a full expression
of normal behaviour. Typically, an animal with a highly motivated, but thwarted,
behaviour, will attempt to perform parts of the behavioural programme, and then
repeat this part for an extended period of time. Horses with limited locomotory
possibilities may develop weaving (swaying from side to side while putting the
weight on alternative front legs), and horses with limited possibilities for foraging
behaviour may develop crib biting (biting edges of the crib, stall or fence).
Common stereotypies in dogs are, for example, tail chasing and wall jumping.
Stereotypies may be difficult to cure once they have developed, so to avoid them,
preventive measures are most efficient.
Stress can also lead to increased aggression, which the dog may direct either
towards other dogs, humans, including the owner, or both (see Chapter 13). The
mechanism here may be that the increased sympathetic nerve activity puts the
dog in a state of ‘fight or flight’, which lowers the threshold for aggression-induc-
ing stimuli. However, increased aggression as such is not diagnostic for stress,
since many other conditions may give the same result (see Chapters 13
and 14).
Final Remarks
The behaviour of dogs is – just as for any animal – a result of how the nervous
system interprets sensory information and transforms this into muscle activity.
Sensory organs, the structure of the nervous system, and the ways in which
muscles and bones build up the body are all processes that are under the control
of genetic factors. However, as we have seen, both the development of the dog,
and the processing of information, depend on environmental input. Behaviour is
therefore truly an effect of both genes and environment. Understanding the way
in which behaviour is controlled and executed is essential for understanding how
dogs experience their lives and why some individuals sometimes develop abnor-
mal and unwanted behaviour.
References
Alcock, J. (2001) Animal Behavior – An Evolutionary Approach, 7th edn. Sinauer
Associates Inc., Sunderland, Massachusetts.
Bekoff, M. (2000) Animal emotions: exploring passionate natures. Bioscience 50,
861–870.
Braastad, B. and Bakken, M. (2002) Behaviour of dogs and cats. In: Jensen, P. (ed.)
The Ethology of Domestic Animals – An Introductory Text. CABI, Wallingford,
UK, pp. 173–193.
Carlson, N.R. (1981) Physiology of Behavior. Allyn and Bacon, Toronto, Ontario.
Hogan, J.A. (1988) Cause and function in the development of behavior systems. In:
Blass, E.M. (ed.) Handbook of Behavioral Neurobiology. Plenum Press, New
York, pp. 3–15.
Jensen, P. and Toates, F.M. (1993) Who needs ‘behavioural needs’? Motivational
aspects of the needs of animals. Applied Animal Behaviour Science 37,
161–181.
Jensen, P. and Toates, F.M. (1997) Stress as a state of motivational systems. Applied
Animal Behaviour Science 54, 235–243.
Lawrence, A.B. and Rushen, J. (eds) (1993) Stereotypic Animal Behaviour –
Fundamentals and Applications to Welfare. CAB International, Wallingford,
UK.
Tinbergen, N. (1963) On aims and methods of ethology. Zeitschrift für
Tierpsychologie 20, 410–433.
Toates, F.M. (2002) Physiology, motivation and the organization of behaviour. In:
Jensen, P. (ed.) The Ethology of Domestic Animals – An Introductory Text. CAB
International, Wallingford, UK, pp. 31–50.
Uvnas Moberg, K. (1994) Role of efferent and afferent vagal nerve activity during
reproduction: integrating function of oxytocin on metabolism and behaviour.
Psychoneuroendocrinology 19, 687–695.
vom Saal, F.S., Grant, W.M., McMullen, C.W. and Laves, K.S. (1983) High fetal
estrogen concentrations: correlation with increased adult sexual activity and
decreased aggression in male mice. Science 220, 1306–1309.
Behaviour Genetics in Canids
5
Elena Jazin
Introduction
Behavioural genetics is a relatively new field joining scientists interested in the

study of the modifications and variation of behaviours due to genetic differences.
Genes are the units of heredity passed down from parents to offspring, and they
are not enough, by themselves, to finally determine all behaviours. Instead, they
operate together with environmental conditions. Environment has a different
meaning in behavioural genetics than in other contexts. As an ecological term,
environment means the physical world. As a genetic word, environment means
all influences other than inherited factors, including external influences such as
weather conditions or interactions with other organisms, but also factors inside
the body such as nutrients, hormones or pathogens, to cite a few of multiple
possibilities. In general all behaviours are due to biological processes in the
brain, and the state of the brain is influenced by multiple factors that shape the
products of the function of the brain, which are thoughts and behaviours.
Behavioural genetics is a multidisciplinary field that overlaps and complements
research related to biology, psychology, physiology, medical genetics, evolution-
ary science and neuroscience among others. To clarify a common confusion, it
should be mentioned that this research area is not called behavioural genetics
because all scientists have a reductionistic view of the problem, and think that
genes are more important than environment to determine behaviour. Instead, the
name is due to the fact that these scientists use genetic-based research tools to
dissect all factors, genetic and environmental, that contribute to differences in
behaviour.

(ed. P. Jensen)
Behaviour Genetics 77
Are All Behaviours of Interest for Behavioural Genetics?
One of the reasons why this field faces a very difficult task is that behaviour is not
always easy to define and measure with validity and reliability. Behaviour, in a
very general sense, can be defined as the actions a living organism makes, as a
whole, in response to the environment (see also Chapter 4 in this volume).
Anything from a sneeze due to pollen, crawling under a tree to avoid the sun,
barking or courtship before mating, are behaviours. Physical manifestations of
most diseases, such as epileptic seizures, can be considered behaviours as well.
Although slightly expanded in relation to the view proposed in Chapter 4, we can
broadly say that there are two forms of behaviour, those that are actions and
those that are states of mind, not necessarily externally manifested. Therefore the
term behaviour can also include mental processes such as thought, dreams or
emotions, such as depression, aggression or happiness. Some behaviour is unique
to one individual, for example the way your dog puts his pad on the door twice
when he wants to be let outside. Other behaviours are just unique to one situa-
tion a certain individual is exposed to, and they may not be repeated, such as the
way you would react if you suddenly see an astronaut in your backyard. Other
behaviours are uniform across a whole species or breed. Consider for example the
instinctive retrieval behaviour of a yellow labrador or the herding posture of a
border collie. The last kind of behaviours, usually called instincts or behavioural
adaptations, are most interesting for behavioural genetics. We can say that a
behavioural adaptation is an inborn pattern of activity, or tendency to action,
common to a given biological species or breed. It is a performance by an animal
of complex acts, absolutely without instruction or previously acquired knowledge.
In some schools of research, for many years the only instinct commonly accepted
was a general ability to learn, but today we recognize many different ones,
ranging from internal manifestations such as fears and phobias, to complex com-
bination of physical acts, such as mating strategies. Behavioural adaptations are
inherited from one generation to the next and, therefore, they are influenced by
gene modifications. However, even those behaviours that have a genetic compo-
nent can be modified by environmental influences, for example a dog’s innate
ability to retrieve can be nurtured, shaped and trained to different levels.
What is the Molecular Basis of an Inherited Behaviour?
Genes are the basic unit of heredity and are located on chromosomes, condensed
strands of genetic material residing inside the cell nucleus, that carry the genes
from one generation to the next. As discussed below, single genes do not deter-
mine most behavioural adaptations, which instead can be modified by the actions
of multiple genes and the environment. The genes form only a small fraction of
the chromosomes that also carry a large amount of genetic material of unknown
function. A single copy of all the genetic material from one organism is called
a genome. Genomes are made up of DNA, which is quite simply the basic
78 E. Jazin
molecule of heredity. It consists of subunits we will be referring to as nucleotides.

All genetic information in an organism is formed by multiple combinations of
only four different nucleotides, designated according to the base that they carry:
A (for adenine), G (for guanine), T (for thymine) and C (for cytosine). The prod-
ucts of a gene are RNA and proteins. To produce a protein, the first step is that
of transcription. This is the copying of a DNA fragment to messenger RNA. RNA
is ribonucleic acid, and it carries the code from the nuclear DNA to the second
step, which is that of translation of the messenger RNA to synthesize amino acids.
Amino acids are the building blocks for proteins that have multiple functions in
an organism. Genes carry the genetic information, or code, for RNA and protein
products that influence the performance or appearance of individuals. Therefore,
behavioural or any other characters can only be transmitted genetically from one
generation to the next through different combinations of genes and other frag-
ments of genetic information, called regulatory regions, which control the
amounts of genes produced. A phenomenon of interest to us is the occurrence of
mutations. Mutations are changes in the DNA sequence, and they can provide for
new genetic material to be transmitted. Mutations occur at random in any part of
the genome and create different forms of genetic information at a certain position.
Each alternate form of a mutation segregating in the population is called an allele.
The impact of a mutation on a population is through their contribution to fitness
for those animals that carry the mutations. As such, mutations can be favourable,
neutral or unfavourable. By neutral, we mean that the mutation has no impact on
any character, also called phenotype, or the impact it has on phenotypes has no
association with the fitness of the animal with the mutation.
Many or Single Genes May Influence Behaviour
It has been shown that modifications in a single gene may produce large changes
in behaviour. For example, as early as 1915, A.H. Sturtevant realized that a single
mutation in the fruit fly Drosophila melanogaster that caused a change in eye colour
could also affect mating behaviour (Brush, 2002). Another example is a single
mutation in the monoamino oxidase (MAO) gene in humans that could cause an
aberrant behaviour in males resulting in sexual attacks on the females from the
same family, a behaviour called arsonism (Brunner et al., 1993). The fact that a
mutation in a single gene can cause a large modification in behaviour does not
mean that it is the only gene involved. In other words, the MAO gene should not
be called ‘the aggression gene’ simply because a mutation can have a large effect
on this trait.
As an analogy, an automobile requires thousands of parts for the normal
function of the engine. If any of the parts break down, for example the oil
pump, the effect on the engine is dramatic and it will stop working, but we
would not call the pump ‘the engine of the car’. In the same way, single genes can
drastically affect behaviour that is normally influenced by multiple genes. In
general, a defined proportion of genes and environmental factors determine each
behavioural trait. Figure 5.1A presents a trait in which 100% of the phenotype is
determined by a single gene. These kinds of characters are called single gene
traits, and they follow a simple (Mendelian) inheritance pattern. There is a com-
plete correlation between the genetic information carried by the individual and
phenotype expressed. In most cases, behavioural traits would instead belong to
the category complex, or multifactorial traits, in which multiple genes in combi-
nation with environmental factors modify the phenotype. Figures 5.1B, C and D
represent different proportions of genetic and environmental influence. A trait
similar to the B example would be easier to dissect genetically, while the D
example, with multiple genes contributing to a small proportion of the phenotypic
variance, represent the most difficult situation for finding the underlying genes,
and the phenotype for a certain genetic component is almost impossible to
predict.
The difficulty in predicting the phenotype that one individual will have when
multiple genes in combination with the environment interact can be illustrated by
one example in which several, but not that many, genes are involved. Coat colour
in mammals holds a particular fascination for all of us. Dogs have a wide variety
of coat colours that are controlled by many alleles at several loci and the colour
of the descendants of animals with certain genetic combinations are not always
easy to predict. Lessons learned from the study of coat colour will prepare us for
understanding the challenges of selection for more complex traits such as several
behavioural traits. The genetics of coat colour in mammals is largely associated
with several genes that modify pigment granules. This association can be either
by altering their number, their shape, their arrangement or position in the hairs,
or by substituting one type of pigment for another. The different coat colour loci
Fig. 5.1. A graphic representation of how a single trait may be affected by different
numbers of genes. A represents a single-gene situation, where one gene com-
pletely determines a trait, whereas B–D show situations where different numbers of
genes and environmental factors (the white area of the circles) contribute to
various amounts of the phenotypic variation.
80 E. Jazin
include: agouti, black and brown, albino, dilution and extension loci, greying and
silver, merle, harlequin, tweed, spotting and ticking loci. These loci are either
named by their impact on the phenotype relative to the colour we see, such as black
or brown, their influence on the intensity we might see, such as the dilution, or in
their distribution of colour patterns as we might see in merle, spotting or ticking. To
make the matter more difficult, there are also a number of modifier genes that alter
the phenotype that is typically expressed at other loci. There are breed differences
for which loci are segregating and which are fixed, and there are also differences
within breeds in the genetic background in which these genes are expressed. We
must keep in mind that just the sheer number of these loci and possible genotypes
within each of those loci makes studying coat colour a formidable task. It is there-
fore of extreme importance to select well-defined models to study the genetic con-
tribution to behaviour and use powerful genetic tools to analyse these traits.
Behavioural Genetics in Dogs
Dogs present a quite unique situation for behavioural genetic analysis because
they have been bred for centuries as much for their behaviour as for their looks.
It is well known that dog breeds differ dramatically in appearance and size.
However, the artificial selection imposed by humans was at least equally strong to
develop breeds with particular capacities, such as herding, retrieval, pointing,
guarding, resulting in groups with differences in activity, emotionality and aggres-
siveness. The contribution of genes to dog behaviours was clearly demonstrated
by elegant breeding experiments performed over two decades by Scott and Fuller
(Scott and Fuller, 1965). They studied five pure breeds – wire-haired fox terriers,
cocker spaniels, basenjis, Shetland sheepdogs and beagles – and they also char-
acterized all possible hybrids between two breeds. They studied a long list of
canine habits, including for example investigative patterns such as walking with
nose to ground, shelter building, grooming behaviour, dominance and submis-
sion, feeding, howling in unison or alone, chasing, barking, aggressive and fearful
behaviour in response to human handling. It was quite clear that the five breeds
presented large differences in several behaviours, particularly dominance
between female–male pairs, tail wagging, lip licking, avoidance of human handler
and biting. They also showed that these behavioural traits presented intermediate
values for the hybrids between the five breeds, indicating a clear genetic contri-
bution to the phenotypes.
A puzzle to solve is that the large behavioural differences between canid
species contrasts with the similarities in the genome sequences of these two
species. In fact, the genome sequence of dogs and wolves is almost identical (Vilá
et al., 1997; Kirkness et al., 2003). The high sequence similarity between dogs and
wolves suggests that altered gene expression, rather than changes in structural
gene products, may be the main mechanism leading to trait differences between
the two species. The idea that changes in regulatory regions have been a major
motor for evolution was first proposed in 1971 (Britten and Davidson, 1971).
Subsequently, it has been postulated that mutations in regulatory sequences

account for most biological differences between species (King and Wilson, 1975).
Recently, these ideas have been supported through the search for the genetic
basis of the difference in mental capabilities between humans and other primates,
some of which share 98% sequence similarity (Chen and Li, 2001). In this case,
altered gene expression, particularly in the brain, seems to be the main determi-
nant for behavioural changes (Enard et al., 2002; Gu and Gu, 2003). Pronounced
changes in brain gene expression may not solely be associated with the develop-
ment of human characteristics, they may also be important during the evolution
of other species, as shown recently for dogs (Saetre et al., 2004).
Altered gene expression can operate faster than other known evolutionary
mechanisms leading to species differentiation, and alterations of gene expression
patterns during the domestication of dogs may have triggered a rapid differentia-
tion despite overall genome similarity. Furthermore, since selection for behav-
ioural traits probably started earlier than selection for morphological traits (by the
preferential breeding of docile and tame animals), the expression differences
between dogs and wolves should be particularly pronounced in the brain, espe-
cially in tissues involved in emotion and behaviour. Different breeds of dogs
should be more related genetically than dogs to wolves, which are generally
accepted as two different species. However, morphological and behavioural dif-
ferences among the breeds are extremely large (Hart and Miller, 1985). Studies
on the genetic changes among different breeds have shown that genetic related-
ness correlates with morphological similarity and geographical origins in some
cases, while other genetic groups seem to correlate with behavioural traits (Parker
et al., 2004). Interestingly, canine breeds share fragments of the genome that seg-
regate together from one generation to the next (called linkage disequilibrium
blocks or LD blocks) that are 100 times more extensive than the LD blocks found
among humans (Sutter et al., 2004). Again here, breed-specific alterations in gene
expression in the brain may be responsible for the large behavioural differences
observed among breeds and the genetic differences responsible for expression
changes may be located in LD blocks that differ between the breeds.
Different groups of dogs may not only be of importance for the study of
normal behavioural differences among the breeds but also to dissect abnormal
behaviours. It has been suggested that canine conditions of separation anxiety,
obsessive-compulsive disorder, cognitive dysfunction, dominance aggression and
panic disorder can be used as models for human generalized anxiety disorder,
obsessive-compulsive disorder, Alzheimer’s disease, impulse control disorders and
panic disorder (Overall, 2000).
How Do We Know That a Behaviour Has a Genetic

Component?
The first question that needs to be addressed before starting a large effort to find
genes related to a behavioural trait is whether heredity has any importance.
82 E. Jazin
There are several approaches and genetic experiments that can be used to esti-
mate the extent to which observed differences among individuals are due to
genetic differences and/or environmental differences. In general these methods
are called quantitative genetics, and they estimate the extent to which differences
among individuals are due to genes and environment, without specifying the
number and type of genes and environmental factors involved. It is possible to
estimate a parameter called heritability that indicates how much of the variabil-
ity in a character that we observe is due to genes or environmental factors in
general, without any specification of the number and type of genes or environ-
mental factors involved. More technically, heritability estimates the amount of
phenotypic variation in a trait (for example, behaviour) that is attributable to
genetic variation in a specific population. In the case where no genetic compo-
nent is present, the heritability is zero. In the other extreme case, the heritability
is one when a single gene solely determines the character without any modifica-
tion due to environmental factors. When heritability values are high, genetic dif-
ferences are easier to identify using molecular genetic methods as explained
below. However, the more genes involved, the less the contribution will be for
each gene to the combined heritability value, and these genes will be more diffi-
cult to find.
Some of the methods used to study the involvement of genes are based on
natural variation of behavioural traits. The crosses between different breeds of
dogs to produce hybrids, as performed by Scott and Fuller as described above, are
an example of this sort of strategy. Hybridization refers to the mating between
representatives of different gene pools. The resulting progeny from the hybridiza-
tion process is referred to as a hybrid. Hybrids are usually denoted as the F1 gen-
eration, the first filial generation. The main idea in these experiments is that if
genes are involved in a certain trait, the crosses of individuals very different for
that trait should produce hybrids with intermediate trait values (Fig. 5.2). Each
original parental population (P1 and P2) has some variability in the trait, but all
the hybrids produced (F1) have intermediate values. To reinforce the results, it is
possible to do backcrosses: crosses between the F1 individuals and each of the
parental populations (B1 and B2). If the trait has a genetic component, then the
values of the backcross populations should be intermediate between F1 and P1 or
P2, as shown in the figure. The example presented is a simple situation in which
all the genetic components act in the same way, so that the presence of more than
one component results in the sum of the effect of each separate component (addi-
tive effect). Other interactions are possible and different models can be built to
understand the contribution of different genes.
Another type of method is artificial selection experiments, which provide the
clearest evidence that genes are involved in behaviour. For example, foxes
that were the tamest when fed and handled by humans were selected during mul-
tiple generations by a research group in Russia. The result of this artificial selec-
tion experiment is a new breed of foxes that are similar to dogs in their
friendliness and eagerness for human contact (Trut, 2001). If a trait has a genetic
component, the values of the population should be modified after selection in a
Fig. 5.2. A hypothetical case which demonstrates how crosses can prove a genetic
control of behaviour. Assume that a behaviour differs between two parental
populations (P1 and P2). If the behaviour is under genetic control, the offspring
(F1) will have a trait value which is in between that of the parents. If F1 offspring
are back-crossed to either P1 (producing B1) or to P2 (producing B2), the trait
values of B1 will be intermediate to F1 and P1, and that of B2 intermediate to F1
and P2.
constant environment. If the trait is solely determined by environmental ele-

ments, the selection regime should not modify the average values for the trait in
the selected population. One important observation is that selection for behav-
ioural traits produces partial changes in one direction (for example, partial
increases in average values) that continue steadily during multiple generations.
This outcome strongly suggests that many genes contribute to variation in behav-
iour. If only one or two genes were involved, then the trait would change after
a couple of generations and the values would stabilize without additional
changes.
A problem of the two types of experiments described above is that they
involve crosses of the studied populations. This may not always be practical
and/or ethical. Instead, it is possible to study the correlation between the results
of different behavioural traits measured in multiple individuals from different
breeds of dogs. If a correlation is found, this would indicate common biological
mechanisms and possibly common genes behind the selected traits (Saetre
et al., 2006). Kenth Svartberg presents a similar approach in Chapter 11 of this
book.
Identifying Genes
Once a genetic contribution to a trait has been established, it is possible to

attempt the identification of the particular genes involved. As mentioned above,
84 E. Jazin
multiple genes and environmental factors determine most behavioural characters.

Each gene, or rather genomic region that contains a gene included in a multiple
gene system that can produce measurable changes in a trait, is called a quantita-
tive trait locus (QTL). Unlike single-gene effects that are necessary and sufficient
for the development of certain traits, each QTL has relatively small effects on the
trait. Traditionally, genetic mapping has been used to localize QTLs in the
genome. The classical analysis involves crosses between two strains with contrast-
ing behaviours to produce hybrids (F1) and then inter-crosses between the hybrids
to produce a second generation (F2). In this generation, recombination has
produced genetically unique individuals. Recombination is the natural crossing
over between the two copies of each chromosome that occurs during meiosis.
This is important for mapping, because it breaks down the association between a
QTL and molecular markers on the same chromosome. The result is that
markers that are closer to a QTL tend to show greater association with the trait
of interest than markers that are further away. At each position in the genome
there are only two possible alleles (1 and 2) segregating from the parental strains.
To map the position of a trait, one compares the values for the trait for each of
the possible genetic combinations or genotypes (11, 12 or 22) for each marker. If
there is a QTL close to the marker, the difference among the three genotypes will
exceed chance expectations, that is, the QTL will be associated with a certain
allele combination.
In an F2 cross there is only one generation in which recombination can break
down the linkage between the QTL and nearby markers. As a result, mapping
resolution is severely limited by sample size. The regions identified are usually
very large, about 10–20 million base pairs of DNA, and such regions may con-
tain thousands of genes among which the important behavioural gene may
be located. Another problem is that multiple individuals usually have to be
included in the analysis to increase the resolution and that replication of the
results for genes of very weak effects is extremely difficult. Because of these
difficulties, few QTL analyses have resulted in the identification of the gene con-
tained in the genomic segment. One elegant exception is the identification of
RGS2 as a gene involved in anxiety in mice (Yalcin et al., 2004). In this case, the
authors used outbreed populations and they constructed combinations of alleles,
called haplotypes, located close in the genome. Then they studied the probability
that an ancestral haplotype carried a gene for anxiety. Finally, they used a test
called quantitative complementation to evaluate the effect RGS2, a gene con-
tained in the candidate haplotype, had on anxiety. Similar strategies could be
used for the study of canine behavioural genetics. Recent studies of the genetic
architecture of dog breeds have shown that individuals within one breed share
quite large (1–3 million base pairs of DNA) regions of the genome (Sutter and
Ostrander, 2004). As mentioned above, these regions are said to be in ‘linkage
disequilibrium’ (LD) and they are called LD blocks. As LD in dog breeds is
20–100 times more extensive than in humans, a much smaller number of markers
will be required for whole-genome association mapping studies in dogs compared
to humans.
Transcriptome Analysis
Instead of looking for genomic regions that may contain QTLs for behavioural
traits, an alternative approach for the identification of genes involved in behav-
iour is the analysis of segments of DNA that code for genes and are transcribed
into RNA. The advantage of this approach is that current advances in molecular
methods have made it possible to analyse simultaneously almost all the genes
expressed in the genome using microarrays. The combination of all the genes
from an organism that are expressed in a tissue at a certain developmental phase
is called a transcriptome and the method is called transcriptome analysis. The
main idea behind this analysis is that QTLs that code for behavioural genes
should be transcribed, and differences in the amount of transcripts will in many
cases be associated with small quantitative differences in gene function and
behaviour. The notion that mRNA expression differences may have extreme
importance in the function of genes is not new. As mentioned, changes in regula-
tory regions have been proposed as a major motor for evolution (Britten and
Davidson, 1971). Subsequently, it has been postulated that mutations in regula-
tory sequences account for most biological differences between species (King and
Wilson, 1975). It follows that many behavioural differences may also be caused
by differences in expression levels of multiple genes.
To search for expression differences that may have an impact on behaviour,
the brain is the central organ to analyse. Humans and primates, some of which
share 98% sequence similarity, show multiple differences in expression levels in
the brain, suggesting that altered brain gene expression may be the main deter-
minant for behavioural changes that makes us ‘human’ (Enard et al., 2002; Gu
and Gu, 2003). Pronounced changes in brain gene expression may not solely be
associated with the development of human characteristics, they may also be
important for behavioural differences in other species, particularly those that
were subject to rapid behavioural differentiation. For example, alterations of gene
expression patterns during the domestication of dogs from wild wolves may have
triggered a rapid differentiation despite overall genome similarity (Saetre et al.,
2004).
Brain expression differences may not only be responsible for behavioural dif-
ferences between species but also between dog breeds with extreme differences in
behavioural traits, and the investigation of affected genes may help the selection
of candidates for additional functional characterization (Fig. 5.3). One caveat
with this type of analysis is that genetic differences between the transcriptomes
could be at least partially obscured by the confounding effect of expression dif-
ferences produced by changes in the living environment of the species or breeds
compared. For example, the diets of humans and apes are very different, and
dogs are fed regularly with variable food types and live mostly in controlled envi-
ronments, while wolves have a diet restricted to the availability of prey in the wild,
and are subjected to long periods of fasting. This could have strong consequences
in the hormonal balance and metabolism, with the possibility of strong conse-
quences in the transcriptome balance of each species.
Dog
86
External
databases
Wolf
Frontal cortex
Microarray
Amygdala
Breed
differences
E. Jazin
Tissue
differences
Coyote
Hypothalamus Expression of
Species
genes in brain
differences
Fig. 5.3. The procedures involved in examining the differences in gene expression levels of
different species or breeds. Tissue (such as relevant brain regions) are obtained from the animals one wishes
to compare, and mRNA is extracted as a measure of the genes expressed at the moment of sampling. The
cDNA produced from this mRNA is hybridized to microarrays containing samples of thousands of genes,
and with various databases it is possible to analyse the results in terms of which genes are up- or down-
regulated in the samples.
To overcome this difficulty, it is possible to study another species as a domes-

tication model produced in a controlled environment. In the Russian farm-
fox experiment, silver foxes (a colour morph of the red fox, Vulpes vulpes) were
selected for non-aggressive behaviour towards man for more than 40 generations
(Trut et al., 2004). This resulted in silver foxes that not only were docile and
friendly but also showed equally good skills as dogs in communicating with people
(Hare et al., 2005; Trut et al., 2004). The transcriptome analysis of docile versus
regular foxes indicated that many genes are modified due to environmental
causes and only a few gene expression differences were the result of genetic
differences between the groups (Lindberg et al., 2005). Only genetic differences
in gene expression are important for behavioural genetics and careful designs
should be used to distinguish genetic from environmental changes in gene
expression.
Mutations in the DNA that cause changes in gene expression are said to
operate in cis or in trans mode. In other words, regulatory elements located close
to the gene may be modified (cis-mutations) or other proteins that modify expres-
sion may be changed (trans-mutations). In general, the genetic basis of inter-indi-
vidual variation in gene expression is simpler for cis-regulatory mutations (Cowles
et al., 2002; Rockman and Wray, 2002) and there is a growing body of evidence
that that natural genetic variation can cause significant inter-individual variation
in gene expression (Brem et al., 2002; Rockman and Wray, 2002; Schadt et al.,
2003; Morley et al., 2004; Chesler et al., 2005). However, multiple regions of the
genome may be involved in the regulation of the expression of a single gene and
a network of complex interactions could be involved in the quantitative changes
associated with behavioural traits.
Genetical Genomics: Potential for Behavioural Genetics
Genetical genomics is a neologism that refers to the combination of transcriptome

analysis with QTL analysis aimed at dissecting the genetic regulation of gene
expression (Jansen and Nap, 2001). In this case, QTL analysis as described above
is performed using gene expression differences as traits. If the difference in expres-
sion is associated with a region of the genome located very close to the gene
coding for expression, then a cis-mutation is probably involved. Conversely, if the
QTL that affects expression is far away or in another chromosome, then a trans-
mutation is involved. It should be noted that the use of microarray analysis
enables the study of thousands of expression traits simultaneously. Some QTLs
that affect expression levels will probably affect multiple genes (for example,
QTLs coding for common transcription factors), while others will be more selec-
tive. This type of strategy has recently been used to analyse transcriptional net-
works in yeast (Hoeschele and Bing, 2005). If the expression profiling is carried
out in the brain of animals for which behavioural tests have been conducted,
there is a potential to identify QTLs that simultaneously affect behavioural
response and gene expression.
88 E. Jazin
A recent study of recombinant inbred (RI) mice from a cross of C57BL/6J

and DBA/2J mice demonstrated the power of this approach. Using microarrays,
Chesler and colleagues profiled genome-wide brain gene expression in 31 RI lines
that had previously been genotyped and behaviourally phenotyped for ethanol-
related behavioural traits (Chesler et al., 2005). They were able to show that the
expression level of the D2 dopamine receptor (Drd2) and several ethanol-related
behavioural traits mapped to the same region on chromosome 9, which was
located close to the position of the Drd2 gene itself. Although the Drd2 transcript
is not polymorphic, several SNPs have been identified in its promoter, which sug-
gests that a regulatory polymorphism in Drd2 is an important component of vari-
ation in ethanol-related behaviour.
A drawback of the genetical genomics approach is that few labs have the
resources necessary to carry out the large-scale expression profiling of hundreds
of F2 progeny needed for a genetic mapping project and only a few groups have
combined genome-wide expression profiles with a genome scan of genetic
markers in experimental populations (Brem et al., 2002; Schadt et al., 2003;
Chesler et al., 2005).
Functional Characterization of Selected Candidate Genes
All selected candidate genes need to be validated for their involvement in behav-
ioural traits to close the circle. It will therefore be important to select appropriate
behavioural tests for validation. The effect of environmental differences during
testing has been intensely debated (Crabbe et al., 1999). These results indicate that
it may be impossible to reproduce exactly the same behavioural phenotypes in
different laboratories. However, it should be possible to determine with certainty
that a particular gene can cause qualitative changes in specific behaviours. Such
qualitative changes may be extremely important for the differentiation of very
closely related individuals such as different species that belong to the Canidae
family.
The definition of a species being the most inclusive reproductive population
is a useful one because it puts boundaries on gene flow. This means that members
within a species can mate and produce viable offspring. For most cases, this defi-
nition is probably appropriate. The question is whether this definition is appro-
priate relative to the various species that make up the genus Canis. In fact, wolves,
coyotes, jackals and dogs are inter-fertile, and crossbreeding still occurs in the
wild. A better understanding of the genetics behind behavioural differences may
help to distinguish man’s best friends from their wild ancestors.
References
Brem, R.B., Yvert, G., Clinton, R. and Kruglyak, L. (2002) Genetic dissection of tran-
scriptional regulation in budding yeast. Science 296, 752–755.
Britten, R.J. and Davidson, E.H. (1971) Repetitive and non-repetitive DNA
sequences and a speculation on the origins of evolutionary novelty. Quarterly
Review of Biology 46, 111–138.
Brunner, H.G., Nelen, M., Breakefield, X.O., Ropers, H.H. and van Oost, B.A.
(1993) Abnormal behavior associated with a point mutation in the structural
gene for monoamine oxidase A. Science 262, 578–580.
Brush, S.G. (2002) How theories became knowledge: Morgan’s chromosome theory
of heredity in America and Britain. Journal of the History of Biology 35,
471–535.
Chen, F.C. and Li, W.H. (2001) Genomic divergences between humans and other
hominoids and the effective population size of the common ancestor of humans
and chimpanzees. American Journal of Human Genetics 68, 444–456.
Chesler, E.J., Lu, L., Shou, S., Qu, Y., Gu, J., Wang, J., Hsu, H.C., Mountz, J.D.,
Baldwin, N.E., Langston, M.A., et al. (2005) Complex trait analysis of gene
expression uncovers polygenic and pleiotropic networks that modulate nervous
system function. Nature Genetics 37, 233–242.
Cowles, C.R., Hirschhorn, J.N., Altshuler, D. and Lander, E.S. (2002) Detection of
regulatory variation in mouse genes. Nature Genetics 32, 432–437.
Crabbe, J.C., Wahlsten, D. and Dudek, B.C. (1999) Genetics of mouse behavior:
interactions with laboratory environment. Science 284, 1670–1672.
Enard, W., Khaitovich, P., Klose, J., Zollner, S., Heissig, F., Giavalisco, P., Nieselt-
Struwe, K., Muchmore, E., Varki, A., Ravid, R., et al. (2002) Intra- and interspe-
cific variation in primate gene expression patterns. Science 296, 340–343.
Gu, J. and Gu, X. (2003) Induced gene expression in human brain after the split from
chimpanzee. Trends in Genetics 19, 63–65.
Hare, B., Plyusnina, I., Ignacio, N., Schepina, O., Stepika, A., Wrangham, R. and
Trut, L. (2005) Social cognitive evolution in captive foxes is a correlated by-
product of experimental domestication. Current Biology 15, 226–230.
Hart, B.L. and Miller, M.F. (1985) Behavioral profiles of dog breeds. Journal of the
American Veterinary Medical Association 186, 1175–1180.
Hoeschele, I. and Bing, N. (2005) Genetical genomics analysis of a yeast segregant
population for transcription network inference. Genetics 170, 533–542.
Jansen, R.C. and Nap, J.P. (2001) Genetical genomics: the added value from segre-
gation. Trends in Genetics 17, 388–391.
King, M.C. and Wilson, A.C. (1975) Evolution at two levels in humans and chim-
panzees. Science 188, 107–116.
Kirkness, E.F., Bafna, V., Halpern, A.L., Levy, S., Remington, K., Rusch, D.B.,
Delcher, A.L., Pop, M., Wang, W., Fraser, C.M. and Venter, J.C. (2003) The dog
genome: survey sequencing and comparative analysis. Science 301,
1898–1903.
Lindberg, J., Bjornerfeldt, S., Saetre, P., Svartberg, K., Seehuus, B., Bakken, M., Vila,
C. and Jazin, E. (2005) Selection for tameness has changed brain gene expres-
sion in silver foxes. Current Biology 15, R915–916.
Morley, M., Molony, C.M., Weber, T.M., Devlin, J.L., Ewens, K.G., Spielman, R.S.
and Cheung, V.G. (2004) Genetic analysis of genome-wide variation in human
gene expression. Nature 430, 743–747.
Overall, K.L. (2000) Natural animal models of human psychiatric conditions: assess-
ment of mechanism and validity. Progress in Neuropsychopharmacology and
Biological Psychiatry 24, 727–776.
90 E. Jazin
Rockman, M.V. and Wray, G.A. (2002) Abundant raw material for cis-regulatory
evolution in humans. Molecular Biology and Evolution 19, 1991–2004.
Saetre, P., Lindberg, J., Leonard, J.A., Olsson, K., Pettersson, U., Ellegren, H.,
Bergstrom, T.F., Vila, C. and Jazin, E. (2004) From wild wolf to domestic dog:
gene expression changes in the brain. Brain Research. Molecular Brain
Research 126, 198–206.
Saetre, P., Strandberg, E., Sundgren, P.E., Pettersson, U., Jazin, E. and Bergström, T.F.
(2006) The genetic contribution to canine personality. Genes, Brain and
Behavior 5, 240–248.
Schadt, E.E., Monks, S.A., Drake, T.A., Lusis, A.J., Che, N., Colinayo, V., Ruff, T.G.,
Milligan, S.B., Lamb, J.R., Cavet, G., et al. (2003) Genetics of gene expression
surveyed in maize, mouse and man. Nature 422, 297–302.
Scott, J.P. and Fuller, J.L. (1965) Genetics and the Social Behavior of the Dog.
University of Chicago Press, Chicago, Illinois.
Sutter, N.B. and Ostrander, E.A. (2004) Dog star rising: the canine genetic system.
Nature Reviews. Genetics 5, 900–910.
Sutter, N.B., Eberle, M.A., Parker, H.G., Pullar, B.J., Kirkness, E.F., Kruglyak, L. and
Ostrander, E.A. (2004) Extensive and breed-specific linkage disequilibrium in
Canis familiaris. Genome Research 14, 2388–2396.
Trut, L.N. (2001) Experimental studies of early canid domestication. In: Ruvinsky, A.
and Sampson, J. (eds) The Genetics of the Dog. CAB International, Wallingford,
UK, pp. 15–41.
Trut, L.N., Pliusnina, I.Z. and Os’kina, I.N. (2004) An experiment on fox domestica-
tion and debatable issues of evolution of the dog. Genetika 40, 794–807.
Vilá, C., Savolainen, P., Maldonado, J.E., Amorim, I.R., Rice, J.E., Honeycutt, R.L.,
Yalcin, B., Willis-Owen, S.A., Fullerton, J., Meesaq, A., Deacon, R.M., Rawlins, J.N.,
Copley, R.R., Morris, A.P., Flint, J. and Mott, R. (2004) Genetic dissection of a
behavioral quantitative trait locus shows that Rgs2 modulates anxiety in mice.
Nature Genetics 36, 1197–1202.
Sensory Physiology and Dog
6
Behaviour
Hermann Bubna-Littitz
Introduction
The purpose of this chapter is not to cover sensory physiology completely but to
stress some topics of special behavioural relevance and/or which deviate
markedly from the acuity of our own senses. The function of the sensory system
is to provide information to the animal about its chemical, physical biotic and
abiotic environment and of the state of the body of the organism itself. With this
information, the animal is able to behave in an appropriate way to the changes in
these factors, or to change the factors themselves. The effect of the behaviour is
further monitored by the organism. If the behaviour was successful it will be
stopped and if not another behavioural strategy will be tried. The uptake of infor-
mation by the sensory system is not a passive process; on the contrary, it is an
active one for two reasons:
1. The physical or chemical influences are transformed into generator potentials
and/or action potentials. These neurophysiologic parameters allow the process-
ing of the input.
2. The sensory system is influenced by the state of the organism. For example, a
dog which is hungry will react with predatory behaviour on seeing a rabbit with
higher probability than a satiated dog. This means that motivation exerts an influ-
ence on the sensory system by filtering the stimuli and altering the attention for
special types of stimuli (Bubna-Littitz, 2005).
The sensory system plays a crucial role not only in the lives of each individ-
ual, but also in intra- and interspecific communication. Dogs use acoustic, chem-
ical (odours, pheromones) and visual signals. A signal is defined as behaviour

(ed. P. Jensen)
92 H. Bubna-Littitz
assigned with a distinctive meaning, which is recognized by the receiver

(Tembrock, 1987), and this obviously relies heavily on a sensory system which is
able to detect the signal.
Development of the Sensory System and of Behaviour
The development of the sensory system is more or less finished at the end of the
third week of the dog’s life (see also Chapter 4 in this volume). From the 28th day
of life onwards a transition from variable locomotor responses to adult-like pos-
tural and equilibratory abilities occurs. Visual and auditory recognition of litter-
mates and other individuals is developed. In addition, approach and avoidance
behaviours emerge (Fox, 1965). In accordance with the development of the
sensory and motor system, the volume of the brain increases rapidly until the end
of the 4th week of life. With the end of week 7, the gain of volume ceases (Fox,
1965). The function of the sensory system is the precondition for imprinting (see
also Chapter 8 in this volume). Imprinting is a learning process that is only possi-
ble within a short period of development, called the ‘sensitive period’. This
process assigns a distinct meaning to an object of the biotic or abiotic environ-
ment, e.g. the object of sexual or social behaviour. The effect of imprinting is irre-
versible.
Hearing
Basic anatomy and physiology of the auditory organs
The auditory system consists of the external, middle and inner ear (see Fig. 6.1).
The external ear (earlap; pinna) is a structure consisting mainly of cartilage. This
provides the breed-specific form of the external ear and the muscles which can
rotate the pinna, thus enabling better detection of the direction from which the
sound is coming. The pinna works not unlike an ear trumpet. It tunnels the sound
waves (pressure waves of the air) to the middle ear. The border between the exter-
nal and the middle ear is the tympanic membrane, which is thin and elastic, with
a size of about 25 mm2. The sound waves cause an oscillation of this membrane,
which is transduced by three small bones (incus, malleus and stapes). The last of
these three ossicles is in tight connection with the oval window, which is the
entrance to the inner ear.
Since the area of the tympanic membrane is about 20 times bigger than the
area of the oval window, the force exerted here is 20-fold higher than the one
exerted on the tympanic membrane. In the middle ear, two of the smallest
muscles of the body are situated: one of them is connected with the stapes and the
other one with the tympanic membrane. The contraction of these muscles
decreases the transfer of vibration through the middle ear. A reflex elicits the con-
traction when the incoming energy of the sound is too high. This reflex is also
Sensory Physiology and Behaviour 93
Fig. 6.1. The inner ear. A, B, C: semicircular canals; D: tympanic membrane; E:

malleus; F: incus; G: stapes; H: musculus tensor tympani; I: musculus stapedius; J:
foramen rotundum; K: top of the cochlea; L: auditory (Eustachian) tube; M: lower
tympanic tunnel; N: upper tympanic tunnel; O: scala media; Q: utricle.
activated immediately before a dog starts to bark and therefore offers protection
to its own ear.
From the oval window, a pressure wave is running up the upper tunnel of the
cochlea, which contains fluid (perilymphe). At the top of the cochlea, there is a
connection to the lower tunnel, so that the pressure wave finally arrives at the
round window, closed by an elastic membrane, which bulges out when the wave
arrives. That is the precondition for the propagation of the wave, since fluids are
not as compressible as air.
Between the upper and lower tunnel there is the middle tunnel (scala media)
which contains the Organ of Corti. This consists mainly of sensory cells, which
react on vibration. High frequencies are detected in the beginning of the scala
media, and low frequencies at the end (at the top of the cochlea).
The neural action potentials arising from the cochlea arrive in the cochlear
nuclei in the medulla oblongata in the brain. From there, these are transmitted
by various brainstem routings to different neuronal structures (e.g. the cerebellum
and the auditory cortex), partly crossing over from one side to the other. The con-
scious perception of sound occurs in the auditory cortex. The connection of the
auditory system with the cerebellum is responsible for eliciting the so-called shoul-
der-reaction, whereby the individual looks over the shoulder towards the origin of
the sound. The location of the source of sound may be detected in two ways. First,
by comparing the intensity of the sound coming from the right and left ear in the
auditory cortex, where information from both sides is available. Second, the shift
in phase between the sound waves arriving in the left and right ear can be
94 H. Bubna-Littitz
detected in the superior olivary nuclei (‘place mechanism’; Pearson and Pearson,
1976).
Hearing in dogs
Dogs have an excellent auditory system. They can locate the source of sound far
better than man: dogs can distinguish two different sources of sound when the
angle between the connecting lines (head-source) is as small as 1° 26′. In man the
corresponding value is 4° 18′ (Buddenbrock, 1952). This ability is very important
in predatory behaviour performed in complex terrains (e.g. wood with scrub).
The direction can be detected by motions of the earlap and by using the differ-
ence of phase of the sound waves arriving at the left and right ear.
The dog is able to hear ultrasound (frequencies higher than 20 kHz) up to 60
kHz (Overall, 1997). What is the benefit of this ability? In accordance with Peters
and Wozencraft (1989), one can assume that puppies communicate in distress sit-
uations with the mother by means of ultrasound. We know from physics that the
damping of an oscillation correlates negatively with its frequency. Therefore,
ultrasound is used when the signal should not be heard at greater distances to
avoid possible predators detecting it. In predatory behaviour, it could play a role
in finding prey, since some rodents use ultrasound in communication too.
Fox (1965) distinguishes between the following types of canine vocalizations:
1. Infantile sounds like crying, whimpering and whining.
2. Howling. The interpretations of this type of sound are controversial. In
wolves, two functions are discussed: to get in contact with other members of the
pack and to mark the territory by sound. In some dogs, howling is elicited when
the owner is absent.
3. Aggressive growling.
4. Submissive whining.
5. Territorial defensive barking.
From the acoustical signal, the receiver dog can presumably estimate the size
of the vocalizing dog. Dogs range in body mass from chihuahua to Saint Bernard,
a 100-fold difference. In addition, the vocal tract length differs tremendously and,
because of that, also the acoustic signal (Riede and Fitch, 1999).
Vision
Basic anatomy and physiology of the visual organs
The eye of the dog shows an anatomy similar to that of humans. The white layer
encasing most of the eyeball is called the sclera (see Fig. 6.2). A part of the anterior
is modified to be transparent, a part called the cornea. When light goes through
the cornea, it enters the anterior chamber, penetrates the lens and arrives at the
On
Fig. 6.2. The eye. A: anterior chamber; Ac: area centralis; C1, C2: between C1 and
C2 is the cornea; Cb: ciliary body; I: iris; L: lens; On: optic nerve; P: posterior
chamber; R: retina; S: sclera; Sf: suspensory ligaments; V: vitreous humour.
posterior chamber. Both chambers are filled with fluid. The iris forms the border
between these two chambers. By contraction or dilatation of smooth muscle fibres,
the diameter of the pupil (‘the central hole of the iris’) can be changed and thus the
amount of light reaching the retina. The lens is elastic. It is fixed by suspensory lig-
aments, which originate from the circular ciliary muscle. The contraction of this
muscle decreases the tension of the suspensory ligaments. Because of the inherent
elasticity of the lens, it becomes more spherical and has more light refraction that
focuses the image of nearer objects on to the retina. The retina is in the rear part
of the eye, and mainly consists of a layer of photoreceptors and neurones process-
ing the information in a first step towards visual awareness. Behind the retina lies
a light reflecting pigment called tapetum lucidum. Light, which has passed the
photoreceptors, is reflected, and passes the photoreceptors again. This layer is
especially important in nocturnal animals and is responsible for the ‘night shine’ of
the eyes in dogs and cats, for example. In the dog’s retina, there are two types of
photoreceptors: one type, the rods, can only differentiate the intensity and not the
wavelength (colour) and have high sensitivity. Therefore, this type is responsible
for vision when light intensity is low, like in the evening or night. The second type
is the ‘cones’, which have lower sensitivity but can discriminate different wave-
lengths. Primates, including humans, can distinguish three basic colours (they are
trichromatic): one cone type is sensitive for red, one for green and one for blue.
Dogs only have two subtypes, so they are bichromatic.
96 H. Bubna-Littitz
Vision in dogs
Unlike humans, the dog has no fovea centralis but an area centralis. As in the
fovea centralis, that is where the highest density of photoreceptors is found. The
effect of this is that in this area the resolution of optical signals is higher than in
the rest of the retina.
The importance of vision depends to some extent on the breed of the dog: it
is less important in breeds such as bassets, beagles and bloodhounds, which track
prey using their noses (‘scent hounds’). It is of more importance in dogs hunting
in free areas like Afghans and greyhounds, which have good sight (‘sight hounds’).
However, in general, visual acuity is higher in humans than in dogs.
The colour vision of the dog has been a matter of discussion for years. As
already mentioned, dogs (like many other mammals) have only two types of
cones. Some therefore believe that dogs are unable to distinguish different colours
in the same way as humans do. However, in an experiment using a discrimina-
tion task, it was clearly shown that dogs can differentiate colours. The dogs had
to identify the designated correct colour out of magenta, cyan and yellow. Finding
the correct colour was rewarded by food, and the dogs could clearly distinguish
all three colours (Antolini-Messina, 1996).
In another experiment also using a discrimination task, it was found that the
ability to distinguish different scales of grey is poorer than that of humans
(Pretterer et al., 2004), but that the dogs could readily discriminate between red,
green, blue and two types of turquoise and all scales of grey. None of these animals
was able to distinguish between a third type of turquoise (wavelength: 480 nm)
from the corresponding grey shade. Therefore 480 nm is the so-called neutral-
point for the dog (a colour with this wavelength looks like grey) (Pretterer, 2000).
The behavioural aspect of vision in the dog is that it plays a major role in
finding prey. Predatory behaviour presumably makes use of a cascade of stepwise
sensory information. By olfaction, the dog finds the direction where possible prey
could be found. With the help of vision and hearing it finds the prey and the
instinctive killing behaviour is then released. Visual signals also play an important
role in intra- and interspecific communication: the posture of the whole body, the
mimic of the face, ears, mouth, muzzle, the fur (flat or raised), the tail held
upwards or downwards or wagging (see Fig. 6.3a–c). An alert dog is depicted in
Fig. 6.3A: the ears are erect, there is a gentle slope in the hip, a relaxed tail and
no piloerection can be seen. When a dog’s fear level increases, the head is
lowered, the ears are pulled back and become limper. The lips become looser and
the eyebrows are arched, the tail is lowered or between the hind limbs (Fig. 6.3B).
When aggression increases the ears are raised, the fur piloerected, the teeth
shown, nostrils and pupils are widened, the tail is upright or wagging, and the
hind limbs stretched out (Fig. 6.3C). The submissive dog, on the other hand,
makes itself look smaller and in extreme situations, it rolls on its back, flexes its
limbs and feet, exposes its belly and tucks in its tail.
In some breeds with great deviations from the appearance of the wolf, prob-
lems in sending visual signals can occur. For example, dogs with hanging ears or
Fig. 6.3. Signal behaviour of the dog. The figure shows three dogs in different
states of mind: an alert one (A), a fearful one (B) and an aggressive one (C). From
the pictures one can derive that the posture differs tremendously, thus another dog
can recognize the mood of a conspecific from a far distance.
98 H. Bubna-Littitz
short snout may have problems sending appropriate signals and sometimes may
be misunderstood by other dogs. Another example is the chow chow, which nor-
mally has stretched hind limbs, which is misinterpreted by some dogs as demon-
strative behaviour.
The Olfactory System
Basic anatomy and physiology of the olfactory organs
The olfactory sensory cells lie in the brain in the olfactory bulb. These have two
axons (nerve fibres), and are therefore called ‘bipolar sensory cells’. One axon
goes to the mucous membrane of the nose and sprouts there. These sprouts have
about 1000 different receptors for various types of chemical compounds (Buck,
2000). The other axon has contact with other neurons in the olfactory bulb and
a first stage of the processing of olfactory information happens there. Thus, the
sensory neurons of the olfactory system are the only ones that have direct contact
with the brain. The sensation of scent is the result of neuronal processing: from
the sensory neurons, information is transferred to the olfactory bulb situated on
the other side of the brain, and to the limbic system, which influences the emo-
tional state of the organism. Axons from the olfactory bulb conduct information
to the olfactory cortex as well. The typical scent of another conspecific or some-
thing else is therefore due to the neural processing of receptor information of
many chemical compounds with different structure and concentration.
Olfaction in dogs
In free-ranging dogs or wolves, the olfactory system is important to detect prey.

The dog can usually detect the direction where the scent is coming from. In doing
so, the moisture of the nose may be important, since it may allow the dog to iden-
tify the wind direction.
The olfactory system of the dog is extremely sensitive compared to that of
humans. This fact is not only the consequence of the larger area of the olfactory
epithelium of the dog compared with humans (5 cm2 in humans and 150 cm2 in
the German shepherd) (Brüggemann et al., 1965), but also of the higher density
and sensitivity of the sensory cells. For some compounds, the sensitivity of the
dog’s olfactory system is millions of times higher than in men (Brüggemann et al.,
1965). Hence, it is clear that smells play a much more important role to the dog
than to humans: when a male dog meets a female, olfactory cues provide infor-
mation about her state in the sexual cycle, for example. Each individual dog can
probably be identified by its personal olfactory profile. Therefore, a change of the
individual odour of a dog can cause problems, since it might not be identified by
another dog: change of smell can turn a dog into an ‘alien’ individual. Drugs used
for anaesthesia can change the personal odour, a problem well known in cats.
Dogs, mainly males, mark their territories with urine. It contains information
about sex and perhaps about state in the ranking order: generally the alpha-male
raises the hind limb extremely high, so that the urine is deposited higher than for
lower ranking males. The posture during urination demonstrates the position in
the ranking order. In neutered female dogs the rising of the hind limb during
miction is sometimes observed too.
Dogs also use defecation for marking. Some individuals try to deposit their
faeces as high as possible, for example on fences or trunks. The scratching of the
ground after defecation is rather an imposing behaviour than a dispersion of
faeces or odours. This behaviour is more extensive when another male passes. In
this situation, sometimes the scratching dog starts to growl.
Dogs used for hunting or used in the military or police service are trained to
find deer, persons, etc., by sniffing. Here, it is necessary to distinguish between
primary and secondary tracks: the primary track is the scent of the person; plants
and insects for example, which have been destroyed by the footsteps of the
searched person or prey, cause the secondary track and deliver additional scents.
Thesen et al. (1993) distinguish three phases during tracking the footprints of a
person. In the first phase (searching phase), the dog is going straight forward while
sniffing close to the ground. When it has found the track, it identifies the direction
the person was walking by comparing the intensity of the scent of the first foot-
prints (phase 2, deciding phase). During tracking (tracking phase, phase 3), the
pattern of sniffing is similar to that of phase 1. When tracking, the dog sniffs with
a frequency of six inhalations per second for about 15 times and then a respira-
tion follows. In phase 2, the respiration follows each period of 35–50 sniffs.
In dogs searching for drugs the decision phase (phase 2), which is the most
strenuous one, might be the only one occurring. After about 20 minutes of work,
the dog is usually exhausted. Sniffing leads to an increase of the energy needed
for inspiration and expiration, which can be demonstrated by blood samples
(Strasser et al., 1993). One can speculate that pack hounds overcome the problem
of exhaustion by alternating with each other in tracking. As a speculation, I want
to put forward another hypothesis: scent hounds, like bassets, beagles and blood-
hounds, track their prey using their noses. The breeds mentioned have very long,
hanging ears. Maybe these serve a function in whirling up the scent and guiding
it to the nostrils.
Pheromones
Pheromones are chemical compounds that are produced from one animal and
influence the behaviour of a conspecific. These play a large role in sexual behav-
iour. For example during proestrus and oestrus, a pheromone (methyl-p-hydroxy-
benzoate) is produced in the vagina of the dog, and this compound stimulates the
interest of the male in the female. The pheromones are not only detected by the
olfactory system described above but also by the vomeronasal organ or
accessory olfactory system. This organ is best described as a fluid-filled tubular
100 H. Bubna-Littitz
structure which opens into the nasal cavity via a duct at its anterior end (Buck,
2000). The axons from the sensory neurones of that organ conduct action poten-
tials to the accessory olfactory bulb and from there these potentials are transmit-
ted, amongst others, to the hypothalamus but not to the olfactory cortex – this
means that the dog does not ‘feel the smell’ of a pheromone. The hypothalamus
is a structure involved in the regulation of the secretion of hormones, some of
them involved in sexual behaviour and reproduction. We should remember that
the olfactory system is in tight connection with the limbic system, which influ-
ences the emotional state of the organism.
Taste
Basic anatomy and physiology of the taste organs
Taste is the least understood of the human senses (Rouhi, 2001). For a long time
it has been a white landmark in physiology. Compounds that can be tasted are
detected by taste cells clustered in taste buds. These buds can be found on the
tongue, palate, epiglottis and the upper third of the oesophagus (Buck, 2000).
However, the sensation of taste is a complex of many different sensory inputs, as
has long been recognized.
The great majority of the sensations we call taste, however, are in reality complex
sensations, into which smell and even touch largely enter. When the sense of smell is
interfered with, as when the nose is held tightly pinched, it is very difficult to
distinguish the taste of various objects. An onion, for instance, the eyes being shut,
may then easily be confounded with an apple.
(Huxley, 1881)
Even in recent textbooks of physiology you will find the statement that there
exist only four types of taste receptors, being the source of the perception of sweet,
sour, bitter and salty. In recent years, a fifth receptor type has been detected: the
umami receptor (Kurihara and Kashiwayanagi, 2000).
Compounds with sodium or hydrogen ions, respectively, are perceived as salty and
sour. Carbohydrates are generally associated with sweetness, although other
compounds may elicit the same sweet perception. The alkaloids caffeine and
quinine are quintessential bitter, but likewise many other compound types taste
bitter. Umami is the savoury taste frequently associated with protein-rich food such
as meat and cheese.
(Rouhi, 2001)
Very probably, new types of receptors will be found and some may not be
present in all individuals. Some humans have receptors for PROP (6-n-propyl-2-
thiouracil) and have the perception of extreme bitter taste. For persons without
that receptor, the PROP does not elicit any impression at all (Rouhi, 2001).
The information from the taste receptors is transmitted to the olfactory
cortex, where the information processing takes place. The intensity of the
different inputs (tastes) is compared and therefore the perception of a distinct taste
like ‘pepper’ occurs. That inputs of the olfactory system are integrated is reason-
able – remember the citation of Huxley mentioned above. The information from
the taste receptors is not only transferred to the cortex via the thalamus, but also
to the nucleus of the solitary tract. From the solitary nucleus arises, amongst
others, the vagal nerve. This fact could explain the phenomenon that in the pres-
ence of distinct tastes vomiting occurs, since the vagal projects directly to
parasympathetic and sympathetic neurons in the medulla and spinal cord that
mediate various autonomic reflexes (Saper, 2000).
Taste in dogs
In the dog, taste functions similarly to in humans (Kurihara and Kashiwayanagi,

2000). In dogs, there are five types of receptors too. For the dog, the receptor sen-
sitive to umami is probably very important, for dogs eat meals with high protein
content. However, as already stated by Huxley, cited above, taste is a function of
olfaction. Taking into account the tremendous sensitivity of the dog’s olfactory
system, we can assume that it has a much greater range of taste impressions than
a human does. Olfaction seems to be more important in dogs than taste: if a dog
smells a preferred substance, this can easily be used to trick it to eat another, non-
preferred substance (Overall, 1997). The main purpose of olfaction and taste pre-
sumably is to select the correct food and to avoid harmful or poisoning ones.
Pain
Basic anatomy and physiology of the pain sensory organs
The sensations we call pain – prickling, burning, aching, stinging and soreness – are
the most distinctive of all the sensory modalities. Pain is of course, a submodality of
somatic sensation like touch, pressure, and position sense, and serves an important
protective function: it warns against injury that should be avoided or treated.
(Basbaum and Jessell, 2000)
Pain can be elicited by different qualities of stimuli, e.g. heat, coldness, pressure
and deterioration of tissue. During the first sensation of pain (‘the sharp pain’), it
can easily be localized. Nerve fibres with high conduction velocity (A-delta fibres)
transmit this sharp pain. The second sensation – sometimes a burning one – is
due to action potentials transmitted by nerve fibres with low velocity of conduc-
tion (C fibres). This sensation cannot be localized as well as the sharp pain. Seen
from the point of evolution, the nonmyelinated C fibres occurred before the
myelinated A-fibres. The explanation could be that exact localization of the
source of pain only makes sense when the organism can react exactly, which
needs a certain state of evolutionary development. Free nerve endings serve as
nociceptors (receptors for painful stimuli) and conduct the information by C
fibres. These are activated by chemical, thermal and physical energy and there-
fore are called polymodal nociceptors. There are specialized nociceptors, just like
mechanical and thermal receptors. These use A delta fibres for transmission.
The action potentials from the nociceptors and the sensory neurones are
transmitted to the dorsal horn of the spinal cord. There, a first step of informa-
tion processing occurs: when someone feels pain, for example in an arm, and
scratches himself, pain is reduced. During scratching, mechanoreceptors are acti-
vated and transmit action potentials by A beta fibres to the spinal cord. These
potentials decrease the conduction of the potentials stemming from the C and A
delta fibres (gate control theory; Basbaum and Jessell, 2000). This could explain
why a dog sometimes bites its tail or paws when in pain.
The nociceptive information is transferred to the brain along five ascending
pathways (spinothalamic tract, spinoreticular tract, spinomesencephalic tract,
cervicothalamic and spinohypothalamic tract). Seen from the behavioural point
of view, the most interesting one is the spinomesencephalic tract: it transmits
information to the thalamus, a brain structure which among its functions has elic-
itation of aggression. The thalamus is also part of the limbic system, which plays
a role in influencing mood and emotion, and it conducts information to the mes-
encephalic reticular formation too, which is responsible for the arousal reaction.
This view of functional neuroanatomy explains what we might observe when a
dog is hurt during sleep: immediately on waking up it may attack the person or
animal nearby, which in the dog’s opinion may be the reason for the pain. If the
pain is lasting for extended time, the mood is severely reduced. The last process-
ing of nociceptive information takes place in the cerebral cortex and leads to con-
scious perception of pain.
The threshold of pain increases during stress. States of stress may be caused
by negative painful stimuli, but similar physiological states may be caused by pos-
itive stimuli too (see also Chapter 4 in this volume). This makes some functional
sense: under the condition of stress, a normal reaction to pain would be a risk to
the animal’s life: stopping during flight because of hurting a paw could result in
death. Therefore, when stress occurs, hormones are secreted which increase the
pain threshold. These hormones belong to the group called ‘endorphins’ (endoge-
nous morphine). This fact may explain why dogs fighting with each other or
attacking a person can only be stopped by extreme physical force.
Pain in dogs
In dogs, thresholds of pain are very variable. The height of the threshold depends
partly on the breed of the dog (e.g. working breeds of dogs vs toy breeds; The
American College of Veterinary Anesthesiologists, 2003).
One widely discussed question is whether pain is inflicted in puppies during
tail docking, i.e. amputation of a part or all of an animal’s tail. Commonly, surgi-
cal methods are used. Dogs usually are docked within an age of 3–5 days without
anaesthesia. In some breeds, docking is required by the breed standards. In an
excellent review, Bennett and Perini (2003) state: ‘Pups do exhibit those pain
responses of which they are capable, and there is every reason to expect that they
experience considerable pain while being docked’. These authors argue that a
puppy can feel pain since the second sensation of pain (burning pain, see above)
is due to C fibres, which are unmyelinated. Myelination of the other nerve fibres
is not finished at the age of 3–5 days. Taking into account the above described
gate control theory, one might expect that the puppy feels more pain than an
adult dog, since the pain-reducing effect for example of the a-Beta fibres is not
working, because these are not fully myelinated in puppies of that age.
Acknowledgements
I want to thank Univ. Prof. Irene Sommerfeld-Stur and Dr Hellmuth Wachtel for
the stimulating discussions on this topic and Andreas Grätzl for the excellent illus-
trations.
References
Antolini-Messina, M. (1996) Untersuchungen über das Farbsehvermögen des
Hundes. DVM thesis, The Veterinary University of Vienna, Austria.
Basbaum, A.I. and Jessell, Th.M. (2000) The perception of pain. In: Kandel, E.R,
Schwartz, J.H. and Jessell, Th.M. (eds) Principles of Neuronal Science, 4th edn.
McGraw-Hill, New York, pp. 472–491.
Bennett, P.C. and Perini, E. (2003) Tail docking: a review of the issues. Australian
Veterinary Journal 81, 208–218.
Brüggemann, J., Horn, V., Moustgaard, J., Hill, H., Kment, A. and Spörri, H. (1965)
Scheunert/Trautmann, Lehrbuch der Veterinärphysiologie, 4th edn. Paul Parey,
Berlin.
Bubna-Littitz, H. (2005) Verhaltensphysiologie. In: Engelhardt, v.W. and Breves, G.
(eds) Physiologie der Haustiere, 2nd edn. Enke Verlag, Stuttgart, Germany, pp.
644–649.
Buck, L.B. (2000) Smell and taste: the chemical senses. In: Kandel, E.R., Schwartz,
J.H. and Jessell, Th.M. (eds) Principles of Neuronal Science, 4th edn. McGraw-
Hill, New York, pp. 625–647.
Buddenbrock, W.v. (1952) Vergleichende Physiologie, Vol 1 Sinnesphysiologie.
Birkhäuser Verlag, Basel, Switzerland.
Fox, M.W. (1965) Canine Behavior, 1st edn. Charles C. Thomas, Springfield,
Illinois.
Huxley, Th.H. (1881) Lessons in Elementary Physiology, 6th edn. Macmillan and
Co., London.
Kurihara, K. and Kashiwayanagi, M. (2000) Physiological studies on umami taste.
Journal of Nutrition 130 (4 Suppl.), 931S–934S.
Overall, K.L. (1997) Clinical Behavioral Medicine for Small Animals, 1st edn.
Mosby, St Louis, Missouri.
Pearson, R. and Pearson, L. (1976) The Vertebrate Brain, 1st edn. Academic Press,
London.
Peters, G. and Wozencraft, W.C. (1989) Acoustic communication by fissiped carni-

vores. In: Gittleman, J.L. (ed.) Carnivore Behavior, Biology, and Evolution, 1st
edn. Cornell University Press, Ithaca, New York, pp. 14–56.
Pretterer, G. (2000) Untersuchungen über das Helligkeits- und Farbsehvermögen des
Hundes. DVM thesis, The Veterinary University of Vienna, Austria.
Pretterer, G., Bubna-Littitz, H., Windischbauer, G., Gabler, C. and Griebel, U.
(2004) Brightness discrimination in the dog. Journal of Vision 4, 241–249.
Riede, T. and Fitch, T. (1999) Vocal tract length and acoustics of vocalization in the
domestic dog (Canis familiaris). The Journal of Experimental Biology 202,
2857–2867.
Rouhi, M.A. (2001) Unlocking the secrets of taste. Chemical and Engineering News
79, 42–46.
Saper, C.B. (2000) Arousal, emotion, and behavioural homeostasis. In: Kandel, E.R.,
Schwartz, J.H. and Jessell, Th.M. (eds) Principles of Neuronal Science, 4th edn.
McGraw-Hill, New York, pp. 873–888.
Strasser, A., Hochleithner, M. and Bubna-Littitz, H. (1993) Zur Belastung von
Gebrauchshunden bei der Suchtgiftsuche. Wiener Tierärztliche Monatsschrift
352–355.
Tembrock, G. (1987) Verhaltensbiologie, 1st edn. Gustav Fischer Verlag, Jena.
The American College of Veterinary Anesthesiologists (2003) American College of
Veterinary Anesthesiologists’ position paper on the treatment of pain in animals.
http://www.acva.org/professional/Position/pain.htm
Thesen, A., Steen, J.B. and Doving, K.B. (1993) Behavior of dogs during olfactory
tracking. Journal of Experimental Biology 180, 247–251.
Social Behaviour of Dogs and
7
Related Canids
Dorit U. Feddersen-Petersen
Why Study Social Behaviour in Dogs?
Just like their ancestors, dogs are highly social animals. Some intriguing insights
into their social life may be achieved by studying both dogs themselves and their
close relatives. In my group, we have spent a long time studying more or less
social canids (C. aureus L. (golden jackals), C. latrans Say (coyotes), C. lupus L.
(wolves)) and domestic dogs (C. lupus forma familiaris) of various breeds tackling
questions dealing with social development, social communication, social organi-
zation, play and aggression.
The animals lived under comparable semi-natural conditions, where group
size, sex ratio and so on were kept constant. Closely related canids, such as dome-
stic dogs of various breeds and their progenitor, the wolf, when kept under com-
parable living conditions, show conspicuous similarities, but also a number of
differences in social behaviour and its ontogeny – with marked intraspecific vari-
ability. Hence, comparative studies of social canids and dogs offer excellent op-
portunities to record constant traits with regard to the development and
significance of individual or species-typical behaviour particularities – as well as
those induced by domestication and breeding (Feddersen-Petersen, 2004).
Pack-living wolves are social canids par excellence. They develop a very high
degree of sociality, a fact that may be judged as a kind of pre-adaptation for
domestication: many capacities fit with the very high degree of social contacts and
interactions of humans, living in reproductive units (families).
Wolves show a variety of facial expressions and body postures, while for
many dog breeds the possibility to communicate precisely is lost due to an
extreme diversity in morphological characters. Regarding the mimic area, in

(ed. P. Jensen)
106 D.U. Feddersen-Petersen
many dog breeds we find only fragments of the wolf’s diversity, fine details and
gradations. In brachycephalic breeds, the forehead is always wrinkled, as is the
nose area, and teeth baring often is not possible because of prominent flews.
Thus, several facial regions and a lot of signals have been lost for communication.
The channel of acoustic communication is partly hypertrophic in all breeds
analysed so far. Categories of function/emotion expressed include social play,
play soliciting, exploration, care-giving, social contact and ‘greeting’, fear and
agonistic behaviour. Interactions range from mildly agonistic biting of infants by
adult dogs to affiliative acts like grooming. Via bark differentiations, the dog
vocalizations seem to have developed into an increasingly communicative com-
ponent of social interactions. In our studies, the German shepherds and especially
the bull terriers were found to bark extraordinarily often and in a variety of social
contexts. Furthermore, sounds occurring in different phonetic qualities were
more common in domestic dogs than in wolves. The evolution of the barking
system could be a parallel to the vocalizing human social partner.
Studies of social behaviour in canids may not only be interesting from the
point of view of communication biology. Tomasello and Call (1997) summarized
their review of primate cognition by noting that ‘The experimental foundation for
claims that apes are “more intelligent” than monkeys is not a solid one’. While
some authors (Flack and de Waal, 2000) focus on nonhuman primates as the most
likely animals to show precursors to human morality, others have argued that we
might learn as much or more about the evolution of human social behaviour by
studying social carnivores (Bekoff, 1995).
According to Bekoff (2000), comparative data on social behaviour in canids
may broaden the study of animal sociality. For example, when more or less social
canids (wolves, domestic dogs, coyotes, golden jackals) engage in social play, they
appear to expect to be treated fairly by their conspecifics. While golden jackals
(Canis aureus L.) usually show play signals when rough play sequences are turning
into aggressive encounters, in wolves (Canis lupus L.) play signals punctuate longer
lasting social plays as a kind of ‘play markers’ (sensu Bekoff, 2000). Social canids
like wolves also learn rules to coordinate their lives. Furthermore, pack size in
wolves is regulated by social factors, which will be clear from examples given in
the continuation of this chapter.
Domestic Animals with Special Demands
Dogs represent a ‘special kind’ among domestic animals, as many breeds or forms
of them have been intimate social partners to humans for many thousands of
years. The unique, intimate relationship between dogs and humans reaches back
15,000 years BP at least (Savolainen et al., 2002; see Chapters 2 and 3 in this
volume), some breeds even apparently preferring humans to conspecifics.
Dogs are the oldest domestic animals. In agreement with morphological,
ethological and chromosomal data, recent genetic findings (Tsuda et al., 1997;
Vilà et al., 1997; Randi et al., 2000; Savolainen et al., 2002; see Chapters 2 and 3
Social Behaviour 107
in this volume) indicated that wolves and dogs are closely related, wolves being
the only ancestors of domestic dogs. Dogs are highly variable in appearance, and
this is also true for behavioural traits. They are adapted to the ecological niche of
‘Hausstand’ (a German word describing life among humans) (Herre and Röhrs,
1990). Dogs are frequently treated in a highly anthropomorphic way on the one
extreme hand or objectified to serve human’s vanity on the other.
I want to discuss some comparative data on social behaviour in dogs and
related canids to detect specific peculiarities of dogs, to promote a better under-
standing of their minds. Dogs cannot be characterized without humans.
Therefore, in the margin, some socio-ecological explanations for the evolution of
human sociality will be offered in this chapter. Social carnivores’ social behaviour
and organization resembles that of early hominids in a number of ways (as
pointed out by Schaller, 1969), and analogue traits exist. Just like humans, wolves
cooperate within packs, and dogs easily form working relationships with man.
Humans and social canids therefore harmonize within a variety of social
accounts. This appears to have been instrumental in the domestication process.
Dogs are very sensitive to human signals, for example, to a person’s ‘body
language’, while human abilities to interpret dog signals vary. Given the closeness
and loyality of the human–dog relationship in so many aspects, the social-cogni-
tive abilities of dogs are largely focused on humans. Dogs have acquired a skill for
communication ‘with humans in a unique way’ (Hare et al., 2002; see Chapter 12
in this volume).
In our research, we performed extensive comparisons between European
wolves (Canis l. lupus L.), coyotes (Canis latrans Say), golden jackals (Canis aureus L.)
and domestic dogs (Canis lupus f. familiaris, various breeds). Behaviour studies were
first carried out under environmental conditions that resemble those of the ances-
tral species: so-called ‘semi-natural’, variable housing conditions, where animals
can largely live according to their own free choice. Regarding dogs, that means
observing them in wide open-air enclosures while they are allowed to live in
packs, comparable to the wolf pack or the social groups of coyotes and golden
jackals. ‘Normal behaviour’ of domestic animals is usually defined as the behav-
iour of healthy animals living in this reference system of a ‘semi-natural’ environ-
ment.
The Social Behaviour of Canids – an Overview
Social development, social communication, social organization, play and aggres-

sion in coyotes and golden jackals differ significantly from other canids. Whereas
wolves and domestic dogs may show what has been called an ‘exaggerated’
degree of sociality (Lewin, 1987), coyotes may best be characterized as ‘social
animals that exhibit solitary attitudes’ (Macdonald, 2004). Coyotes are therefore
highly adaptable. However, they do not possess the wolf’s sociability. This is
reflected in their more stereotyped expressive behaviours. This is even more
obvious in golden jackals. Coyotes and jackals have fewer behavioural elements,
and fewer possibilities of combining signals to differentiated postures or faces. The

amount of social play in golden jackals, coyotes, wolves and domestic dogs
(various breeds) varies.
Social behaviour in dogs and wolves
As mentioned earlier, the high sociality of pack-living wolves may constitute a

kind of pre-adaptation for domestication. The main characteristic patterns of
social behaviour in wolves (and dogs) are the following traits:
● Social play: during play, basic rules of social life are learnt: how hard they can
bite, how roughly they can interact and how to resolve conflicts.
● Sense of fairness (following Bekoff, 2001; De Waal, 2003): while playing, codes
of social conduct are learnt. No matter what the functions of play may be, there
seems to be no doubt that it has benefits and that the absence of play can have
devastating effects on social development (as remarked by Power, 2000). In
Bekoff’s opinion, a sense of fairness is common to social canids (and other
animals) because there could be no social play without it, and without social
play individual animals and entire groups would be at a disadvantage. In this
sense fairness is adaptive, it helps animals to survive in their particular social
environment, and individual animals benefit from this behaviour.
● Communicative skills: these are numerous and highly differentiated. There are
a variety of displays in wolves and dogs which have to be learnt (by associating
the performance with the consequences). Complexity within the dog’s vocal
repertoire is achieved by many subunits of bark, standing for specific motiva-
tions, information and expressions. Complexity within the dog’s vocal reper-
toire is extended by the use of mixed sounds in the barking context. Transitions
and gradations occur to a great extent via bark sounds: harmonic, intermediate
and noisy subunits (as defined below).
● Family groups (packs): social systems vary from pairs and family groups to
packs. Pack size is influenced by food supply and by social factors (Mech, 1970).
In wolf packs there may be division of labour, and food sharing.
● Long-term monogamy: this is the most common mating system, but polyan-
drous mating occasionally occurs. Mating behaviour of dogs (especially pedi-
gree dogs) is nearly exclusively controlled by man. Thus, for the most part there
is no possibility for dogs to live in a long-term relationship with a sexual
partner. Some observations support long-term monogamy in some dogs (feral
dogs, dingoes).
● Bonding behaviour: in wolves and dogs, this can be much stronger and longer
lasting than in species closely related to ourselves, such as chimpanzees.
● Parental care (care of other adults and siblings): provisioning of cubs is done by
both parents and all other pack members act as helpers. Helpers are common
in dingoes (Australian dingo and the so-called New Guinea singing dog) and
occur in several dog breeds.
● Social learning from conspecifics: in wolves the duration of the period of expo-
sure to parental influence (thus the opportunity for social learning) is long, and
normally many pups remain in the pack permanently. There they have the pos-
sibility to learn social behaviour, hunting skills and feeding behaviour.
● Communicating social rights, dominance hierarchy: a differentiated communi-
cation is correlated with a complex social hierarchy. Dominance always char-
acterizes specific relationships and does not refer to individual characteristics
(Van Hooff and Wensing, 1987). When establishing and maintaining a stable
social rank, wolves and dogs perform agonistic encounters in a unidirectional
way. If dominant behaviours, for example a so-called T-sequence or a high
posture, are alternately performed by both interacting animals, this demon-
strates that the interacting individuals have not achieved a stable rank order.
Dominance implies the acceptance of restrictions by the subdominant animal.
Dominant wolves strikingly often initiate dominant behaviours, such as
dominant body postures or mimics, when they engage with pups. In the case
shown in Fig. 7.1c–e, muzzle-biting (with pronounced biting inhibition) follows
(a)
(d)
(b)
(e)
(c)
Fig. 7.1. Communicating social rights to infant wolf.

immediately on licking the bridge of the elder wolf by the pup (b) and placing
the head on the adult wolf’s snout (a). The pup performs a kind of T-sequence
‘from snout to snout’ (a), thus, testing a ‘high social status’ in a certain behav-
ioural context.
A missing or unstable rank order can be indicated by individuals exhibiting

ambivalent communications when interacting. Such a behaviour is controlled by
two different motivations (e.g. attack and fear) occurring simultaneously or in
short, rapid succession. It may produce expressions including ‘incongruous’
signals: in Fig. 7.2 a juvenile wolf (left side) demonstrates partly dominance-indi-
cating behaviours (muzzle biting, nose-wrinkled, pawing the alpha male on the
head) in combination with submissive gestures/signs (folded ears laying back, flat
forehead). Ambivalent expressions are common behavioural traits in wolves and
not exclusive of a particular social status. In fact, higher levels of ambivalence
were observed in high-ranking individuals, the alpha when interacting with any
other male and the beta when interacting with the omega (Fig. 7.2). According to
this finding, a dog’s body posture during agonistic episodes should never be taken
as the only element to establish a functional diagnosis of aggression (Fatjo et al.,
2006).
Some Details of Dog Sociality
Although dogs descended from wolves fairly recently in evolutionary terms, their
social behaviour has changed extensively, particularly in some breeds. In our
research, we found that some dog breeds are unable to cooperate (in a very basic
Fig. 7.2. Play-biting/mimic play in 6-week-old wolf pups.

manner: just doing things together) and compete in groups, reflected in difficul-
ties in establishing and maintaining a rank order (e.g. poodles). The interactions
in these dog groups are not functional, and the members have difficulties coping
with challenges from the environment. It is striking that tactical variants of con-
flict solving (to appease, animate or inhibit the opponent), a common practice in
wolves, do not exist in groups of several dog breeds. These strategies, however,
are important for pack maintenance. Within many groups of dogs, trivial conflicts
often escalate into damaging fights.
This may lead to acute stress becoming chronic. Other breeds (Nordic
breeds: Alaskan malamutes, Siberian huskies, samoyeds and German shepherds,
hunting dogs, some terriers, Fila Brasileiros) have better means to cope with social
conflicts. These dogs are mentally in a good shape when living under ‘semi-
natural’ conditions.
We found that breed groups differed mainly with respect to the following
behavioural measures (Feddersen-Petersen, 2004):
1. There was much more frequent and much more severe aggressive behaviour
in toy poodles, West Highland white terriers, Jack Russell terriers, some bull ter-
riers and some labrador retrievers.
2. There was much more frequent and variable social play in the wolves, stan-
dard poodles, the earlier mentioned Nordic breeds, German shepherds, Filas and
weimaraner hunting dog, which were able to show aggressive encounters in a
more ritualized manner.
3. There was greater social tolerance, more nonagonistic approaches and more
allogrooming in the wolves, malamutes, huskies, samoyeds, German shepherds,
some bull terriers, American Staffordshire terriers, golden retrievers and e.g. Fila
Brasileiros.
Coyotes and golden jackals showed very different behavioural strategies, and
could readily cope with challenges from the environment. The possible reasons
for the decreased capability of several dog breeds to form stable groups will be dis-
cussed later. However, we can already note here that poodles, for example, will
interrupt every/any interaction with conspecifics, if/whenever a human appears,
an intriguing fact regarding the social minds of these dogs. In German shepherd
dogs, the duration of a threat display before and following an attack was constant
– whereas in wolves pre-escalative and post-escalative threat display dyads dif-
fered significantly in duration. These findings suggest that during the process of
domestication, dogs have lost essential capabilities necessary to adapt to living
exclusively with conspecifics in groups. Nordic breeds (Alaskan malamutes,
Siberian huskies, samoyeds) were exceptions, as mentioned above.
Canis lupus lupus L.
Canis lupus f. familiaris
Canis latrans Say
Canis aureus L.
Fig. 7.3. Frequency of social play in some related canids during their first year of
life.
Development of Social Behaviour in Wolves, Domestic Dogs

and Related Canids
During the first year (and beyond), juvenile wolves are characterized by pro-
nounced play activities (Fig. 7.3). In addition to the well-known canid forms of
social play, e.g. contact and chase games, which normally mix one into one
another and consist of highly variable sequences, wolves display a further cate-
gory: communication or facial expression games. Indeed, these games make up a
large part of their social play (Feddersen-Petersen, 1991). Nineteen per cent of all
social behaviour patterns recorded in the third month of life involve mutual com-
munication of playfully expressed, exaggerated signals almost exclusively from the
facial area (Fig. 7.4).
Contact games, or bite and fight games (or play-fighting) are the earliest type
of play. Transitions to serious conflicts rarely happen in wolves. After the third
month, social play is carried out with pronounced role assumption and subse-
quent reversal: a puppy will first play, for example, the ‘underdog’ by sending a
few, limited body signals indicating social inferiority in a playful, exaggerated
manner, while its partner plays the role of the social superior. The ‘social supe-
rior’ may then switch roles and bring elements of submission (sensu Schenkel,
1967) into the play. Playful exaggeration refers to the amplitude and speed of
expressive movements or behaviour patterns. Contact games normally appear in
combination with chase games in which the primary goal is the playful pursuit of
one’s opponent. During play, inequalities in size, strength and dominance
between playmates are reduced, and the cooperation and reciprocity that are
essential for a longer lasting play occur. The role of social play for cooperation is
more than a hypothesis. The rights and wrongs of social interaction can be learnt
best in a play atmosphere. Biting inhibition and partner preferences (preferring to
12 faces,
6 ‘new barks’
(harmonic)
German shepherd
12 faces, 12 ‘new barks’

(harmonic) 60 faces,
American Staffordshire terrier noisy barks only
Canis lupus lupus L
Fig. 7.4. Do reduced facial expressions correlate with sophisticated barkings?
initiate as well as to accept play signals) are important for forming social bonds
between certain animals.
Solitary play is much less common as compared to social games. Domestic
dogs exhibit a maximum in play activity between the sixth week of life and
approximately 6 months of age (Fig. 7.3). At this time, a relatively stable hierar-
chy has been established and from now on the games are tinted by a primarily
aggressive tone and may even end in serious fights (for example, in poodles). Play-
fighting predominates, but decreases rather abruptly towards the eighth month
and continually thereafter. In contrast to wolves, many dogs (living in groups) play
a great deal less after the fifth or sixth month.
While wolves primarily integrate visual signals into the games, most dog
breeds develop bark-games (Zimen, 1981), i.e. they initiate play by ‘play-sounds’
(harmonic barks, growling, vibrato-sounds). Wolves start to play earlier and play
frequency is almost always greater than the corresponding ‘dog-norm’.
Agonistic interactions appear earlier in most dog groups than among wolves.
Agonistic behaviour among wolves living in an age-structured family group is
determined structurally by a high degree of ritualization. Conflicts between
wolves are mainly seen in the form of complex, ritualized fighting. Dogs, as a rule,
fight in a much less ritualized manner. For example, an attack launched by a
dominant poodle male escalated into grabbing and bite-shaking in 70% of the
observed cases, regardless of the opponent’s reaction. Group aggression was
sometimes seen, in which many group members joined a collective attack on a
threatened animal. The social distance between the high-ranking animal or

animals and the rest of the group appears large among dogs.
In golden jackals pronounced aggression is present during the first 4–6 weeks
of life in groups of pups. Bite and fight games, the sole form of social play in golden
jackals, regularly escalate into unritualized conflicts, including serious biting
(intended to harm). The frequency of agonistic behavioural patterns reaches its
maximum at approximately 9 weeks. At this age the litters have established certain
hierarchies and groupings or pairs have created spatially divided subgroups.
Intragroup play frequency increases, while intergroup cohabitation in a relatively
restricted area is regulated by expressing ritualized threat motions. The frequency
of play increases within the subgroups (Fig. 7.3) and social games free of aggression
take place up to the onset of the first oestrus period. The expression elements of
these games are restricted almost entirely to the agonistic context.
Communication games are lacking in golden jackals. Wide mouth-opening is
shown as an intention motion shortly before playful biting in the course of play-
fighting (and only here) and is not answered as a play-signal through facial exag-
geration. The relaxed, open-mouth face of golden jackals is always displayed as a
prelude to play-fighting.
Solitary play is more common than social play up to 3 months of age. The
total amount of play is considerably less than that of wolves at a comparable age.
The frequency of agonistic interactions in coyotes is considerably above the
wolf and dog norms and is especially pronounced prior to the establishment of
social groups (at approximately 9 weeks of age) and to the onset of the first oestrus.
Play signals (like bows) often initiate plays.
In coyotes and jackals, stereotyped play signals (very clear and striking) pre-
dominate (approx. 70% of play interactions), while wolves and dogs have a variety
of fine-graded play signals, play faces and play vocalizations (dogs, especially play
barks), which serve as ‘play markers’, announcing longer lasting play sequences.
In all canids there is an increase in the frequency of play at 1–3 months, the
socialization period, during which intraspecific (and interspecific) attachments are
formed (Scott and Fuller, 1965).
Why Do Several Dog Breeds Have Problems in Forming

Stable Groups?
In adult wolves we found about 60 different distinct facial expressions (Fig. 7.4).
We fixed 11 ‘facial regions’ (including 2–13 different signals), resulting in an
enormous number of possible displays. In dogs, face morphology differs in the
various breeds, which reduces the number of possible expressions. In addition,
visual signals and body postures are less numerous, less differentiated and reduced
in amplitude as well as in clearness, compared to the wolf’s diversity. Thus,
expressions are reduced. Reductions in facial expression produce problems in
social communication, as we have found in mixed poodle–wolf groups, and may
cause increased aggression. Finally, dogs are adapted to living circumstances with
Fig. 7.5. Fighting play in pugs.
Canis lupus f. familiaris

Individual dogs have specific barks with a
Canis lupus L: range of meanings (Feddersen-Petersen,
A single, short sound of the noisy type, 2000; Young, 2002). For example: a het-
occurring solely within aggressive contex- erotypic sequence of harmonic barking (1),
tual categories: threat, attack, warning, and harmonic (2,3) / intermediate (4) growl
defence and protest (Schassburger, 1993) (vibrato) as a play soliciting signal
kHz 1 2 3 4
7
kHz
6
8 5
6 4
4 3
2
2
1
0
1 2 Sec
0 1 2 3 4
Sec
Fig. 7.6. Barking of Canis lupus L. and barking in domestic dogs.
humans. As we found out, aggressive encounters decrease if humans participate

more frequently in group life.
Barking in Wolves and Dogs
Barking in wolves is rare, occurring most often in the social context of fighting.
Wolves’ bark sounds are of the noisy type (characterizing agonistic encounters)
(Fig. 7.6).
Fig. 7.7. Ambivalent communication in juvenile wolf (left).
In dogs there is an abundance of barks/barkings in highly differing social

contexts. Barking as a whole seems to be both directed to humans and an adap-
tation to living with humans.
The principal method, used in our studies of vocalizations, was sonographic
analysis. This facilitates the identification of sounds and reveals whether subjec-
tive classifications can be verified by objectively measured parameters. Finally,
meanings, functions and emotions were examined for all major sounds identified
and were analysed in terms of relationships between sound structure and signal
function, signal emission and social context as behavioural response, and over-
lapping channels of communication. This provides a model for future vocaliza-
tion studies in domestic dogs, a model which draws parallels between structure,
motivation, emotion and ontogeny to arrive at a more comprehensive under-
standing of sound systems in various dog breeds.
According to Schassburger (1993), the wolf’s vocal repertoire consists of 11
basic sound types. In the Alaskan malamute and Tervueren breeds, the vocal
repertoires turned out to be very similar, but meanings, function and emotions,
behavioural responses and overlapping channels of communications varied much
more. The vocal repertoire of the bull terrier turned out to be highly hypertrophic
and much more differentiated in some close-range vocalizations, such as barking
and growling (Fig. 7.6, sonogram on the right side). This applies for most of the
breeds we have analysed so far (standard poodles, toy poodles, Kleine
Münsterländer, weimaraner hunting dogs, Tervueren, American Staffordshire
terriers, Jack Russell terriers, bull terriers, West Highland white terriers, German
shepherds, Alaskan malamutes, Siberian huskies).
American Staffordshire terriers, which have markedly reduced facial expres-
sion, emitted 12 ‘new barks’ (Fig. 7.4), i.e. barks which do not occur in wolves.
Harmonic barks were found in all dog breeds analysed so far. Alaskan malamutes
emitted two harmonic barks in play situations, and when contacting another
animal (or a human being) in a slightly submissive pose (with stimulative expres-
sion). Thus, complexity within the dog’s vocal repertoire, and therefore enhance-
ment of its communicative value, is achieved by many subunits of bark (and
growl), representing specific motivations, informations, expressions and commu-
nications. Complexity within the dog’s vocal repertoire is extended by the use of
mixed sounds in the barking context. Transitions and gradations occur to a great
extent via bark sounds: harmonic, intermediate and noisy subunits.
We have attempted to analyse dog barks in categories of function and
emotion. These categories include social play, play soliciting, exploration, care-
giving, social contact and ‘greeting’, loneliness (loneliness-bark) and agonistic
behaviour (offensive/defensive threat). ‘Interaction’ was the most common cate-
gory of social context for most vocalizations. Interactions ranged from mildly ago-
nistic biting of infants by adult dogs to affiliative acts like grooming. In addition,
there are acoustic signals following or indicating distress via bark sounds or gra-
dations including barks.
Dogs and Humans
Social canids may be characterized by an ‘exaggerated degree of sociality’ (Lewin,

1987). During domestication, dogs seem to have adapted their social behaviour to
a large extent to the life with humans, representing to a far extent their ecologi-
cal niche. But the domestication of social cognition in dogs counteracts their
social communication with conspecifics. Dogs changed in so many ways during
domestication and now depend on humans for their benefit. In so many respects,
modern dogs seem to be more well-adapted to communication with humans than
with other dogs.
References
Bekoff, M. (1995) Play signals as punctuation: the structure of social play in canids.
Behaviour 132, 419–429.
Bekoff, M. (2000) Social play behaviour. Cooperation, fairness, trust and the evolu-
tion of morality. Journal of Consciousness Studies 8, 81–90.
Bekoff, M. (2002) Minding Animals: Awareness, Emotions, and Heart. Oxford
University Press, New York & London.
De Waal, F.B.M. (2003) Reconciliation among primates: a review of empirical evi-
dence and unsolved issues. In: Mason W.A. and Mendoza, S.P. (eds) Primate
Social Relationships State University of New York Press, New York, pp.
111–114.
Fatjo, J.F., Feddersen-Petersen, D.U., Ruiz De La Torre, J.L., Amat, M., Mets, M., Braus,
B. and Mantacea, X. (2006) Aggression in wolves: ambivalent behaviour as a
model for comparable behaviour in dogs. Applied Animal Behaviour Science.
Feddersen-Petersen, D. (1991) The ontogeny of social play and agonistic behaviour

in selected canid species. Bonn, Zoologische Beiträge 42, 97–114.
Feddersen-Petersen, D. (2000) Vocalization of European wolves (Canis lupus lupus
L) and various dog breeds (Canis lupus f. fam.). Archives of Animal Breeding,
Dummerstorf 43, 387–397.
Feddersen-Petersen, D.U. (2004) Communication in wolves and dogs. In: Bekoff, M.
(ed.) Encyclopedia of Animal Behavior, Vol. 1. Greenwood Press, Westport,
Connecticut, pp. 385–394.
Flack, J.C. and de Waal, F.B.M. (2000) Any animal whatever: Darwinian building
blocks of morality in monkeys and apes. Journal of Consciousness Studies 7,
1–29.
Hare, B., Brown, M., Williamson, C. and Tomasello, M. (2002) The domestication
of social cognition in dogs. Science 298, 1643–1636.
Herre, W. and Röhrs, M. (1990) Haustiere – zoologische gesehen. Gustav Fischer,
Stuttgart, Germany.
Lewin, R. (1987) The origin of the modern human mind. Science 236, 668–669.
Macdonald, D.W. (2004) Die große Enzyklopädie der Säugetiere (ed.) Könemann in
der Tandem Verlag GmbH, Königswinter.
Mech, L.D. (1970) The Wolf: The Ecology and Behavior of an Endangered Species.
The Natural History Press, New York.
Mech, L.D. (1999) Alpha status, dominance, and division of labor in wolf packs.
Canadian Journal of Zoology 77, 1196–1203.
Power, T.G. (2000) Play and Exploration in Children and Animals. Lawrence
Erlbaum Associates, Hillsdale, New Jersey, pp. 48–49.
Randi, E., Lucchini, V., Christensen, M.F., Mucci, N., Funk, S.M., Dolf, G. and
Leschke, V. (2000) Mitochondrial DNA variability in Italian and East European
wolves: detecting the consequences of small population size and hybridization.
Conservation Biology 14, 464–473.
Savolainen, P., Zhang, Y., Luo, J., Lundeberg, J. and Leitner, T. (2002) Genetic evi-
dence for an East Asian origin of domestic dogs. Science 289, 1619.
Schaller, G.B. and Lowther, G.R. (1969) The relevance of carnivore behavior to the
study of early hominids. Southwestern Journal of Anthropology 25, 307–314.
Schassburger, R.M. (1993) Vocal communication in the timber wolf, Canis lupus,
Linnaeus. Advances in Ethology 30.
Schenkel, R. (1947) Ausdrucksstudien an Wölfen. Behavior 1, 81–129.
Schenkel, R. (1967) Submission: Ist features and functions in the wolf and dog.
American Zoologist 7, 1–27.
Scott, J.P. and Fuller, J.L. (1965) Genetics and the Social Behavior of Dogs.
Tomasello, M. and Call, J. (1997) Primate Cognition. Oxford University Press, New
York.
ing occured among multiple matriarchal ancestors during the domestication of
region of the mitochondrial DNA between dogs and wolves. Genes and
Genetic Systems 72, 229–238.
Van Hooff, J.A.R.A.M. and Wensing, J. (1987) Dominance and its behavioral meas-
ures in a captive wolf pack. In: Frank, H. (eds) Man and Wolf: Advances, Issues
and Problems in Captive Wolf Research. Junk, Dordrecht, pp. 219–252.
Yin, S. (2002) A new perspective on barking in dogs (Canis familiaris). Journal of
Comparative Psychology 116, 189–193.
Zimen, E. (1981) The Wolf: A Species in Danger. Delacorte Press, New York.
Zimen, E. (1982) A wolf pack sociogram. In: Harrington, F.H. and Paquet, P.C. (eds)
Wolves of the World: Perspectives of Behavior, Ecology and Conservation.
Noyes Publications, Park Ridge, New Jersey, pp. 282–32.
Learning in Dogs
8
Pamela Reid
Introduction
While the principles of learning have been derived primarily from studies
of animals in laboratories, nowhere are they more thoroughly revealed than in
our work with the domestic dog. Dogs have been selectively bred to serve
many functions for humankind, most of which require some degree of trained
behaviour (Coppinger and Coppinger, 2001). In recent years, we have relied less
on dogs for their working ability and more for their companionship, yet dog train-
ing has become increasingly sophisticated as aficionados strive to create ever
more complex and precise behavioural routines. The trend for dogs to function
as household pets and family members is an additional force driving the infiltra-
tion of learning theory into conventional dog training and behaviour modi-
fication, because humans often find aspects of dog behaviour objectionable
(see Chapter 13 in this volume). Illustrations of fundamental learning pro-
cesses abound from these intentional efforts to control and manipulate dog
behaviour.
Learning Processes in Dogs
A fascinating, and often frustrating, enterprise is to map correspondence of learn-

ing theory on to the complications of the real world. Highly specific aspects of
learning can be delineated and explored within the controlled environment of the
laboratory but this rarely happens in the ‘messy’ world of dog training and behav-
iour modification. Throughout this chapter, I provide examples of particular

(ed. P. Jensen)
Learning 121
learning principles; yet bear in mind that in virtually all situations, multiple
processes interact to produce learning.
Non-associative learning
Two of the most fundamental forms of learning are non-associative in nature. By

that, I mean that the dog is not forming associations between events, rather the dog
is learning about events that appear to occur without connection to other events.
For instance, imagine staying at a hotel situated underneath the flight path for a
nearby airport. As the first airplane thunders overhead, your dog pins back its ears,
tucks its tail and dives under the bed. It takes a few moments after the noise ends
for the dog to emerge, only for it to hustle back to its hiding spot at the first hint of
the next plane. Remarkably though, as the day wears on, you notice that the dog
no longer hides, in fact, it appears not even to hear the planes any longer. The dog
has habituated to the sound. Habituation can occur in response to any repeated
stimulus (Thompson and Spencer, 1966). In essence the dog learns, through repe-
titions, that there is no significance to the sound and, therefore, there is no need to
attend to the stimulus. Dogs often habituate to everyday events – the dog that
moves with its family into an apartment building initially barks at every noise in
the hallway, but eventually learns to ignore all the irrelevant sounds. Habituation
is very specific: introduce a novel similar stimulus and the dog will react all over
again (Thompson and Spencer, 1966). Habituation is also somewhat temporary.
If the airport switches to a distant runway for a few days before resuming use of
the one by the hotel, the dog will startle again the first few times a plane flies over.
Quickly, however, the magnitude of the dog’s response will decline (Leaton, 1976).
Occasionally, a dog can be exposed to presentations of a stimulus, say fireworks,
and, instead of habituating, the dog’s response to the stimulus actually intensifies.
With each ensuing exposure, the dog becomes more reactive. This phenomenon is
termed sensitization. Dogs are more likely to sensitize when the stimulus is especially
intense, such as fireworks, gunfire, thunder and other loud percussive sounds (Davis,
1974). Trainers who use electronic stimulation to provide incentive for learning find
that dogs sometimes sensitize to the shock. A dog that startles mildly the first time it
receives a shock might show a much more intense response, jumping, yelping, or in
rare cases even biting the handler, after a few more presentations. Unlike habitua-
tion, sensitization is a more generalized phenomenon (Miller and Domjan, 1981);
often a dog that has sensitized to shock will also visibly startle to other external events,
like a light touch, a blowing leaf or a spoken word.
Operant conditioning
Operant or instrumental conditioning is a form of associative learning, meaning

that associations or relations are established, specifically relations between
behaviour and its outcomes (Skinner, 1938). Animals frequently engage in behaviour
122 P. Reid
that produces changes in the environment: behaviour followed by a pleasant event

is more likely to be emitted again; while behaviour followed by unpleasant events is
less likely to be emitted again (Thorndike’s Law of Effect) (Thorndike, 1911).
Operant behaviour is often described as intentional, voluntary or goal directed; the
animal chooses to perform the behaviour because in the past the behaviour was
instrumental in producing certain consequences (Salzinger and Waller, 1962). Much
of dog training consists of operant conditioning. A dog learns to attend to its name
because the owner, after saying the dog’s name, frequently takes the dog for a walk;
a dog learns to sit when the owner speaks to it because the action of sitting sometimes
results in a tasty biscuit; a dog learns to jump on the counter for the opportunity to
snatch the butter dish; a dog learns to avoid jumping up on the counter because the
action results in the owner shouting and shooing the dog off.
Behavioural outcomes can be classified into four categories, based on the behav-
iour–outcome contingency: (i) positive reinforcement, in which a behaviour produces
an appetitive stimulus (a pleasant thing); (ii) positive punishment, in which a behav-
iour produces an aversive stimulus (an unpleasant thing); (iii) negative reinforcement,
in which a behaviour causes the cessation or prevents the delivery of an aversive stim-
ulus; and (iv) negative punishment (also known as omission training), in which a
behaviour causes the cessation or prevents the delivery of an appetitive stimulus
(Domjan, 2003). Positive reinforcement and positive punishment are both character-
ized by a positive contingency between the behaviour and the outcome (the behav-
iour occurs and some event happens), negative reinforcement and negative
punishment are both characterized by a negative contingency between the behaviour
and the outcome (the behaviour occurs and some event ends or is prevented). The two
types of reinforcement result in the behaviour being more likely in the future; the two
types of punishment result in the behaviour being less likely in the future.
Examples of each of these contingencies abound in dog training. The dog lies
down. This action results in the owner tossing a favourite ball; if the behaviour of
lying down increases, lying down has been positively reinforced. The dog chews
the leg of the couch. This action results in the owner scolding ‘no’ and rapping
the dog on the nose; if chewing the couch ceases, chewing the couch has been pos-
itively punished. The dog jumps up at the owner at the front door. This action
results in the owner ignoring the dog when the dog was attempting to solicit atten-
tion; if jumping up on the owner ceases, jumping up has been negatively pun-
ished. The dog ventures too close to a boundary established with an electronic
containment system and a warning tone sound. The dog runs back toward the
house. This action turns off the unpleasant tone; if the dog stays away from the
boundary in the future, running back home has been negatively reinforced (and
moving near the boundary was positively punished).
Classical conditioning
Classical or respondent conditioning is another form of associative learning,

meaning that associations or relations are established, but in this case, the
Learning 123
relations are among stimuli. Events in the world often occur in predictable com-
binations and classical conditioning is what happens when an animal learns that
one event reliably predicts another (Todes, 1997). This learning occurs because
the animal finds the second event significant and in the presence of the first event,
an animal comes to engage in behaviour that reflects the understanding that the
second event will happen (Rescorla, 1988). The dog learns that when the owner
opens a specific cupboard, a bowl of food will soon be presented. As a result, the
dog begins to salivate when the cupboard door opens. The dog learns that being
greeted by a stranger wearing a white lab coat predicts the insertion of a rectal
thermometer or the prick of an injection. As a result, the dog begins to tremble
and tries to escape the veterinary examination room. Respondent behaviour is
often described as elicited, involuntary or reflexive; the animal is unable to control
whether or not it salivates or trembles (Pierce and Cheney, 2004).
In the archetypal example of classical conditioning, Pavlov first established
that placing meat powder in the mouth of a hungry dog elicits salivation (Pavlov,
1927). Prior learning was unnecessary for this to occur and, therefore, Pavlov
termed the meat an unconditional stimulus (UCS) and salivation an uncondi-
tional response (UCR). Pavlov then preceded delivery of the meat powder with a
distinctive sound, like a metronome. At the outset, the metronome produced only
an orienting response. After a number of pairings of the metronome with the
meat powder, the dog began to salivate upon hearing the metronome. Once the
relation between the metronome and the meat powder was learned, the
metronome became a conditional stimulus (CS) and the salivation elicited by the
metronome, the conditional response (CR).
Social learning
Social or observational learning refers to any form of learning where changes in

behaviour are dependent upon or facilitated by the presence of other animals.
Thorpe (1963) proposes three categories of social learning: (i) social facilitation, in
which the behaviour of one animal prompts the same behaviour by another
animal but the behaviour already exists in the second animal’s repertoire; (ii)
local enhancement, in which the behaviour of one animal draws the attention
of another animal to the specific place in the environment or the particular
stimulus involved, but the second animal still acquires the new behaviour by
trial and error; and (iii) true imitation, in which the behaviour of one animal
prompts the exact same form of the behaviour by another animal and the
behaviour is highly improbable in the second animal without this opportunity for
observation.
Despite a lack of research on social learning in dogs, there is anecdotal
evidence that dogs are capable of both social facilitation and local enhance-
ment. An example of social facilitation is where a dog, reluctant to negotiate
a set of slippery stairs, easily follows another dog up the stairs and, after that
exposure, is comfortable traversing the stairs on its own. Trainers sometimes
124 P. Reid
utilize social facilitation to encourage a dog to try a novel treat, to enter a new
environment or to investigate a frightening object. An example of local enhance-
ment is a dog that has never manipulated a food toy, like a Buster Cube™. After
observing a demonstrator dog obtaining food from the toy, the second dog may
learn a strategy for getting the food from the toy more quickly than a dog that was
not given the opportunity to observe a demonstrator dog. Local enhancement is
very likely the mechanism by which puppies in the Slabbert and Rasa (1997)
study learned an odour-detection task more quickly after observing their dams
performing the task than did control puppies that did not observe their dams at
work.
True imitation can be concluded only after controlling for learning by social
facilitation and local enhancement. Few, if any, convincing examples of imitation
exist in species outside humans and the great apes (Whiten and Ham, 1992)
and to my knowledge, there are no demonstrations of true imitative learning in
dogs.
Establishing Desirable Behaviour through Operant

Conditioning
One of the primary objectives of training the pet dog is to establish desirable
behaviour, such as walking on a leash without pulling, lying down and remaining
in place, and coming when called in the park. A more select subset of dogs are
required to perform a repertoire of behaviours designed to fulfil specific func-
tions, such as herding livestock, retrieving game, negotiating a course of agility
obstacles, or assisting a physically compromised person. Dogs, like other species,
show gradual acquisition of new behaviour mapping on to a negatively accelerat-
ing curve, such as in the hypothetical graph in Fig. 8.1. In this section, I outline
the general processes by which dogs acquire instrumental behaviour for the
purpose of attaining a pleasing outcome, in other words, learning by reinforce-
ment.
1
Probability of the response
0.8
0.6
0.4
0.2
0
1 5 9 13 17 21 25 29
Trials
Fig. 8.1. Hypothetical acquisition curve for dogs learning to come when called.
Learning 125
Motivation
In order to effect behavioural change through operant conditioning, there must

be a motivating force in the form of a behavioural consequence (see also Chapter
4 in this volume). As noted above, behavioural consequences involve delivering or
removing pleasant (appetitive) stimuli or delivering or removing unpleasant (aver-
sive) stimuli. There are two classifications of such stimuli: unconditioned (or
primary) and conditioned (secondary).
Unconditioned stimuli require no prior learning to appreciate, such as food,
water, social contact, sex and various noxious or painful events. For ease of
administration, most dog training utilizes food and social contact as forms of
unconditioned appetitive reinforcers, with social contact consisting of vocaliza-
tions (praise), physical touch (stroking) and interactive play (wrestling, chase, tug
and retrieving games). Not surprisingly, the value of these typical rewards tends
to vary amongst dogs; some dogs find little reinforcing value in social contact,
whereas other dogs prefer interactive play to food. One dog finds retrieving espe-
cially motivating, while another dog enjoys tug. A study in which dogs were
required to press a lever for food rewards (see Fig. 8.2) revealed that one subject
Fig. 8.2. Photograph of dog bar-pressing in operant chamber.

126 P. Reid
would press the lever approximately 30–40 times per session for standard dry
kibble rewards, but would press the same lever well over 100 times per session for
small tidbits of liver (Reid, unpublished). In general, animals perform faster
and/or longer for larger or more palatable rewards (Mackintosh, 1974).
Unconditioned aversive stimuli are usually vocalizations (reprimands), phys-
ical touch (restraint or other forms of ‘corporal punishment’ like hitting) or pain
delivered through a training collar, such as choke or pinch collars that can be con-
stricted around the dog’s neck or electronic collars that can be triggered to
present a noxious spray (citronella) under the dog’s face or an electric shock to the
dog’s neck. Beerda et al. (1998) demonstrated that dogs show more fearful pos-
tures and experience elevated cortisol levels (indicative of higher stress) in
response to unpredictable aversive events, such as loud noises, falling objects and
electric shock, than to predictable events, like an umbrella opening, a spray of
water from a gun or physical restraint. This is consistent with findings from con-
trolled laboratory examinations of other species: the ability to predict an aversive
event can dampen its unpleasantness somewhat (Lockard, 1963).
Operant conditioning is especially sensitive to the temporal relation between
behaviour and its consequence (Mackintosh, 1974). In some situations, learning
can be seriously hampered with as little as a 0.5 s delay between the behaviour
and the outcome (Grice, 1948). In other cases, learning is possible with delays as
long as several minutes (Lett, 1973). The critical feature appears to be the occur-
rence of competing behaviours during the delay (Revusky, 1971). For instance, if
the dog sits but then barks, turns its head and jumps up, all before the trainer is
able to deliver the food treat, it is difficult for the dog to discern which behaviour
led to the reward. While the timely delivery of food is quite feasible in an experi-
mental chamber, it is quite another matter in the real world of dog training, espe-
cially when the dog may be working at a distance from the trainer.
Fortunately, overcoming the problem of timing consequences is accom-
plished through the use of conditioned or secondary stimuli, which can be deliv-
ered immediately after the target behaviour. A conditioned reinforcer is a
stimulus previously linked with an unconditioned reinforcer, such that it comes to
predict the occurrence of the unconditioned reinforcement (Mackintosh, 1974).
Pavlov’s metronome functioned as conditioned reinforcement for his dogs.
Practical conditioned reinforcers tend to be acoustic, so that they can be delivered
immediately after the behaviour, even at a distance. For many dogs, ‘Good dog!’
is a conditioned reinforcer because it reliably precedes other pleasant events, like
treats and play. ‘Clicker training’, first described by Skinner (1951) and subse-
quently popularized by marine mammal trainer Karen Pryor (1984), involves the
use of a metal noisemaker, called a clicker, as a conditioned reinforcer (see Fig.
8.3). The dog is taught that the sound of the clicker signals a tangible reward is
on its way and, consequently, the click bridges the temporal delay between behav-
iour and reward (Williams, 1991). A related but distinct effect is that the sound of
the click helps to mark the target behaviour and make it distinguishable from other
behaviours (Lieberman et al., 1979); hence the clicker is often called a ‘marker’.
A trainer can time the administration of a click much more precisely than the
Learning 127
Fig. 8.3. Photograph of dog and clicker.
delivery of food treats or toys, thereby reducing uncertainty and speeding the
learning process for the dog.
Less systematic in dog training is the use of conditioned negative reinforce-
ment, also known as a safety signal. Imagine a situation where the dog jumps up
on the kitchen counter. The owner shouts at the dog and the dog jumps off. Were
the dog to remain, the owner would positively punish the dog in some way.
Instead, the owner praises the dog for jumping back down. This praise functions
as a safety signal to the dog, signalling that it has successfully avoided the positive
punishment (I think of this as the ‘Whew!’ effect), thereby negatively reinforcing
the action of jumping off the counter. Trainers who use devices like electronic
collars sometimes establish tones as audible safety signals to convey that the dog
has performed the target behaviour and escaped or avoided the potential aversive
stimulus.
Acquiring the target behaviour
The simplest, and perhaps most overlooked, procedure for establishing desirable
behaviour is to reinforce when the dog naturally emits the behaviour. Most
responses that occur in the dog’s behavioural repertoire can be increased in fre-
quency beyond the baseline rate through reinforcement processes (responses that
are not easily linked with the specific reinforcement being used are exceptions; see
Shettleworth, 1972). Some trainers term this catching the behaviour. While effec-
tive, this technique is limited to naturally emitted behaviours.
In the laboratory environment, the primary acquisition procedure is a
sequence of training steps called shaping by successive approximations (Skinner, 1951)
128 P. Reid
or, more colloquially, free-form hand shaping. Initially the animal is reinforced
for doing anything remotely similar to the target behaviour. Once an initial
behaviour is established, reinforcement is systematically withheld so as to elicit
variation in responding (behaviour is inherently variable and even more so when
the animal becomes frustrated). Hopefully, a variation will occur that is closer in
form to the target behaviour. This new behaviour is reinforced until established.
Once again, reinforcement is withheld and the animal varies its behavioural
responses. The trainer reinforces a new behaviour that is still closer to the target
behaviour. These steps continue until the animal is performing the desired behav-
iour. For instance, suppose the goal is to teach the dog to take its leash off the
hook by the door. Initially the dog is reinforced for looking in the direction of the
door. Once the dog is performing this behaviour reliably, the trainer no longer
reinforces this behaviour. The dog becomes frustrated and begins moving about
the room. Eventually the dog moves in the direction of the door. The trainer rein-
forces any movement toward the door. Once the dog is reliably moving to the
door area, the trainer no longer reinforces this behaviour. The dog hangs out by
the door and eventually looks up at the hook. The trainer reinforces this behav-
iour. Once the dog is reliably going to the door and looking up at the hook, the
trainer no longer reinforces this behaviour. In frustration, the dog reaches up
toward the hook. The trainer reinforces this behaviour. The trainer continues to
reinforce successive approximations to the target behaviour of the dog moving
toward the door, reaching up to grasp the leash in its mouth, and pulling the leash
from the hook. Shaping is still relatively new in the field of dog training, yet it is
proving to be extremely useful for teaching trainers to objectively define the
desired response, to hone timing and decision-making skills, and to remain
patient while the dog ‘offers’ various responses.
Much more traditional in dog training is an acquisition procedure known as
prompting and fading (Martin and Pear, 1996). Prompting involves introducing some
stimulus event into the training situation so as to encourage the dog to perform
the target behaviour. Taking the example from the previous paragraph, the dog
could be prompted to perform the behaviour by leading the dog by the collar to
the doorway and wiggling the leash to entice the dog to grab it and pull. The dog
is then reinforced for performing the entire behaviour. This sequence is repeated,
each time reducing or fading the amount of guidance provided until the dog is able
to perform the behaviour on its own. Prompting can take the form of physical
guidance or luring with an attractive item to elicit the behaviour from the dog so
it can be reinforced.
In the case of training with negative reinforcement, often the primary
purpose of the aversive stimulus is to prompt the desired behaviour to occur. One
of the cleanest examples of negative reinforcement training with dogs is a tech-
nique for teaching the retrieve called the ear pinch. The trainer presents a retrieve
object, such as a dumbbell, a few inches in front of the dog’s muzzle and at the
same time, pinches the dog’s earflap to the extent that the dog finds it unpleasant.
The pinch causes the dog to cry out in pain, at which point the trainer is able to
insert the retrieve object into the dog’s open mouth. At exactly that instant, the
Learning 129
aversive stimulus (the pinch) is terminated, thus negatively reinforcing the dog for
having the object in its mouth. Most dogs quickly learn to grab the dumbbell,
consequently avoiding the painful pinch altogether. This behaviour can then be
developed into a full retrieve, where the dog runs to collect the item and return it
back to the trainer. Thus the ear pinch functions in two ways: (i) it prompts the
dog to perform the desired behaviour of opening the mouth; and (ii) it reinforces
grabbing the retrieve object through its termination. Readers, please bear in mind
that I present this as an illustration of an acquisition procedure. Rest assured that
there are equally effective, yet more humane methods for teaching a dog to
retrieve objects.
More complex behaviours can be acquired through a procedure called behav-
iour chaining (Martin and Pear, 1996). Chaining involves breaking a complex
behaviour down into smaller elements, establishing each of these separately, and
then chaining the elements together to build the complete behaviour. Consider
training a dog to place its toys in the toy box. This can be broken down into these
elements: (i) the dog learns to pick a toy up from the floor; (ii) the dog learns to
hold a toy in its mouth; (iii) the dog learns to walk while holding a toy in its mouth;
(iv) the dog learns to walk from various locations to the toy box while holding a
toy in its mouth; (v) the dog learns to hold its head over the toy box; and (vi) the
dog learns to release the toy from its mouth. Each of these elements can be estab-
lished separately and then chained together. Whenever possible, backward chaining
is the preferred choice, meaning that the complex behaviour is established in
reverse order (Mazur, 2002). The dog is first reinforced for dropping the toy in
the box, then for carrying the toy to the box and dropping it in the box, then for
picking up the toy, carrying it to the box and dropping it in, and finally, for
moving to a toy on the floor, picking up the toy, carrying it to the box and drop-
ping it in. This way the dog is always moving toward the terminal reinforcement.
Each step along the way serves as conditioned reinforcement for the previous
element. In contrast, forward chaining establishes the chain from the beginning. The
dog is first reinforced for moving to the toy on the floor, then for moving to the
toy on the floor and picking it up, and so forth. This procedure is less desirable
because with each progression the dog is required to do more and so experiences
a delay to the anticipated reinforcement.
Reinforcement schedules
A schedule of reinforcement refers to a rule that determines how and when a

response will be followed by reinforcement (Ferster and Skinner, 1957). At the
most basic level (i.e. simple schedules), reinforcers are delivered after a certain
number of responses (a ratio schedule) or after the passage of a certain amount of
time (interval or duration schedules). Researchers have collected a mass of data
on how animals adjust and perform on a plethora of both simple and more
complex reinforcement schedules (e.g. Morse, 1966) but, for the purposes of this
chapter, I restrict the discussion to simple schedules. Different reinforcement
130 P. Reid
schedules produce highly consistent patterns of responding. Traditionally, rein-

forcement schedules have been studied primarily in what is called free operant situ-
ations, in other words, the animal is free to perform the operant response as often
as it likes and the measure of interest is the animal’s rate of responding on a par-
ticular schedule. Dog training typically involves discrete trials, in which the dog can
perform the operant response only during specified periods, such as when the
trainer requests that the behaviour be performed. Consequently, rate of respond-
ing is not a relevant measure. Instead, schedule effects centre on the speed and
reliability of the response.
It is well established that a continuous schedule of reinforcement (CRF) is
optimum for the acquisition of new behaviour (Chance, 1999). CRF, also known
as a fixed ratio 1 schedule (FR1), in which each response is reinforced, minimizes
confusion as to what behaviour leads to reinforcement. Once the behaviour has
been learned, however, it is beneficial to implement partial or intermittent sched-
ules to maintain behaviour, in which responding is reinforced only some of the
time. This benefit was initially discovered from observations of rats running alley-
ways to obtain food reinforcement. Surprisingly, rats ran faster when the goal box
contained food only half the time (a fixed ratio 2) rather than all the time
(Weinstock, 1958). Intermittent schedules have the valuable effect of teaching the
dog to perform in the face of non-reward, a strength called learned industriousness
(Eisenberger, 1992).
The larger the number of responses required for reinforcement, the more
likely the animal will display a post-reinforcement pause (Myers and Mesker,
1960). Much like people procrastinating when starting a large task, animals take
a break before responding on a large FR schedule. Trainers recognize that dogs
willingly perform a large number of simple responses for a single reinforcer (for
instance, a dog will run through a tunnel 25–100 times for one reinforcer) but
may suffer from ratio strain when the trainer pushes for a large number of more
complex responses (for instance, some dogs are unwilling to perform a sequence
of more than 25 agility obstacles for a single reinforcer). Experienced trainers
utilize progressive-ratio schedules, starting with a small requirement and gradu-
ally increasing the demand, to build up the dog’s performance so that the poten-
tial for ratio strain is minimized.
Variable ratio (VR) schedules, in which the number of responses required for
reinforcement varies in an unpredictable, albeit constrained manner, produce
steady high rates of responding without a post-reinforcement pause (Ferster and
Skinner, 1957). In general, responding on VR schedules is characterized by
speedier, more intense performances. Those dog trainers who use intermittent
schedules to maintain responding tend to rely on variable ratio schedules. For
instance, after acquiring the behaviour of touching its nose to a target, a dog
might be reinforced after two nose touches, then after four nose touches, then
after one nose touch, then after five, and so forth. Again, the dog is prepared for
this through a progressive-ratio schedule.
Occasionally, dog trainers incorporate reinforcement schedules that depend
upon the passage of time. Interval schedules require that a certain amount of time
Learning 131
must pass before the next response will be reinforced. Interval schedules can be
fixed (a set amount of time – FI) or variable (the amount of time varies around a
set mean – VI). Few examples of interval schedules exist in dog training. One
notable exception is the trainer that works the dog off the leash and reinforces the
dog for ‘checking in’ (the dog approaches and engages the attention of the
trainer). Most trainers initially reinforce the dog for all ‘check in’ responses, but
eventually the savvy trainer only reinforces a check-in that occurs some period of
time after the last check-in (a VI schedule). This maintains check-ins at a steady,
stable rate (Ferster and Skinner, 1957).
Often, dog trainers utilize time-based schedules that demand the animal
sustain behaviour for a specific period of time. Martin and Pear (1996) refer to
these schedules as duration schedules. For instance, dogs trained to stay, to main-
tain eye contact, and to move while in heel position, must do so for initially brief
periods of time and gradually the amount of time required for reinforcement is
extended. Like the other simple schedules, duration schedules can be fixed
(always the same amount of time) or variable (the amount of time varies unpre-
dictably).
Achieving stimulus control
Typically, an operant conditioning sequence consists of three events: (i) an

antecedent stimulus; (ii) a response; and (iii) a consequential stimulus. For
instance, the dog hears the trainer say ‘lie down’, the dog lies down, and the
trainer delivers a food reward. Antecedent stimuli that precede operant behaviour
and that set the occasion for responding are called discriminative stimuli (SD) (Pierce
and Cheney, 2004). A response that reliably occurs in the context of an SD and
not in the absence of the SD is said to be under good stimulus control. Stimulus
control is achieved through a process of differential reinforcement, such that the
dog learns that responding in the presence of the SD earns reinforcement, while
responding in the absence of the SD fails to produce reinforcement. In other
words, the dog learns that the SD sets the occasion for the response–reinforcer
relation.
Trainers call achieving stimulus control as putting the behaviour on cue. The
point at which this aspect of training is introduced depends upon the reliability of
the response. If the response is highly probable, the SD (or cue) can be introduced
early in the acquisition phase because the trainer can be confident the response
will occur in the presence of the SD. If this is not the case, the trainer runs the risk
of presenting the cue and the dog not responding. If this happens repeatedly, the
cue is subject to learned irrelevance (Kremer, 1971), making it more difficult for the
dog to learn that the cue sets the occasion for the response to be reinforced.
Whether a particular stimulus comes to control behaviour depends upon
more than just arranging for it to precede responding. Animals are differentially
sensitive to various stimuli. Dogs possess the predator’s keen senses and are
acutely aware of movement (Miller and Murphy, 1995). Suppose multiple stimuli
132 P. Reid
are presented to the dog, such as when a trainer attempts to teach a dog to jump
a hurdle on cue. The sight of the jump, the trainer’s vocalization ‘jump’, and the
trainer’s movement toward the hurdle are all potential SDs. Given the dog’s sen-
sitivity to movement, the other stimuli are overshadowed and the dog’s jumping
behaviour comes to be controlled by the trainer’s movement, rather than the
trainer’s presumed intention: the spoken word.
Indeed, it has been shown that animals are sometimes predisposed to link
certain antecedent stimuli with certain behaviours. McConnell (1990) revealed
that dogs are more likely to move in the presence of high-pitched, rapidly repeat-
ing sounds, like ‘come, come, come!’ and are more likely to remain stationary in
the presence of low-pitched, continuous sounds, like ‘down’. In fact, McConnell
and Bayliss (1985) determined that across a wide range of cultures and languages,
the cues we use to induce action from dogs tend to be high-pitched, rapidly
repeating sounds and, in contrast, the cues we use to inhibit movement in dogs
are low-pitched, continuous sounds. The selectivity of stimulus and response is
especially pervasive when it comes to learning by negative reinforcement. Studies
with numerous species demonstrate speedy learning when the required avoidance
response is consistent with the species’ defence responses, and painfully slow when
the required response is inconsistent (Bolles, 1970). For instance, a pigeon easily
associates sounds, but not sights, with aversive stimuli and quickly learns to flap
its wings in order to avoid the stimulus. The pigeon would have a difficult time
learning a response that is inconsistent with defence, like pecking a key, to avoid
an aversive stimulus (Mazur, 2002).
Pavlov noticed that after one stimulus, like the metronome, was conditioned
to the point that it elicited salivation, other similar stimuli would also elicit saliva-
tion. This is called stimulus generalization. The extent to which generalization across
stimuli occurs can be portrayed in a stimulus generalization gradient. Figure 8.4
shows a gradient of responding as a function of how similar the stimuli are to the
original test stimuli. These data are hypothetical. The dog was trained to sit when
it hears the word ‘sit’. However, it also shows some probability of responding with
1
response
sit response
0.8
of asite
0.6
Probability of
0.4
Probability
0.2
0
‘Astrid’
‘Dammit’
‘Knit’
‘Pit’
‘Sid’
‘Skit’
‘Snit’
‘Sit’
‘Sat’
‘Sod’
‘Sup’
‘Spot’
‘Circus’
‘Cat’
‘Pizza’
Fig. 8.4. Hypothetical generalization gradient for a dog responding to verbal cues.
Learning 133
a sit when it hears similar words like ‘hit’, ‘spit’ and ‘lit’. It is less likely to respond
to words like ‘spot’ and ‘cat’ and even less likely with highly dissimilar words like
‘pizza’ and ‘envelope’.
The steepness of a stimulus gradient is a measure of how sensitive the animal
is to variations in a particular class of stimuli. A very flat gradient indicates that
the dog responds to a wide range of stimuli, in other words, the dog is under poor
stimulus control. A very steep gradient indicates that the dog responds very selec-
tively to a stimulus, in other words, the dog is under good stimulus control. Steep
gradients are usually produced through differential reinforcement. My dog ‘Eejit’
reliably responded to his name, but when I adopted ‘Fidget’, not surprisingly, he
initially responded to her name as well. It took about one week for him to learn
to differentiate his name (the S+, the stimulus that signals significant events for
him) from hers (the S–, the stimulus that signals nothing for him). This is the
essence of simple discrimination training. The traditional way of teaching a
simple discrimination through differential reinforcement is a slow process and the
animal makes many errors. Imagine how many times Eejit responded to the name
‘Fidget’ and was ignored before he ceased responding. Often this type of training
is unpleasant for the animal. Terrace (1963) developed a superior method called
errorless discrimination learning, errorless because the animal makes few or no
responses to the S–. Initially, the S+ is made attractive to the animal so it is highly
likely to respond to it. At the same time, the S– is faded in so that the animal is
unlikely to respond to it. A realistic example in dog training is a task called scent
discrimination. The dog is required to search a set of visually similar items, one
of which also contains the scent of the trainer. The dog is to retrieve the item that
smells of the trainer (the S+) and ignores the other items (the S–). Suppose initially
the S+ item is also impregnated with the smell of food, like liver. At the same time,
the S– items are presented, at first only one but gradually more and more. The S–
items are smaller than the S+ and placed far away from the dog. The dog is highly
likely to select the correct item. Over trials, the S– items are made progressively
more apparent, by increasing their size and by placing them closer and closer to
the S+ item. But the dog has been repeatedly rewarded for selecting the S+ item
so, by the time the S– items are salient, it is improbable that they will elicit
responding. In one of Terrace’s studies (1963), pigeons trained with a conven-
tional discrimination procedure made more than 3000 responses to the S– during
28 sessions. In contrast, pigeons trained with errorless discrimination learning
made about 25 responses to the S– during a comparable number of sessions.
Terrace’s procedure produces reliable discrimination learning in a minimal
number of training sessions.
Dog trainers will attest that a much more common concern is that dogs dis-
criminate too well; that they fail to generalize. A comprehensive training pro-
gramme should include generalization training, also known as stimulus equivalence
training. We often expect dogs to generalize across a wide range of stimuli, such as
responding to the word ‘sit’ spoken by men, women and children, people both
familiar and unfamiliar to the dog. It appears that dogs are unlikely to generalize
this widely without specific training in which the cue ‘sit’ is presented first in the
134 P. Reid
original form and then in forms gradually less similar until the dog responds to
the common stimulus ‘sit’ in virtually any form. It turns out that responding also
comes under the control of cues provided by the context in which the original
training took place (Stokes and Baer, 1977). This is the origin for the common
complaint of pet owners, ‘but he comes when he’s called in the yard but not at
the park’. The original training was conducted in the yard, with distinct contex-
tual cues and few stimuli to compete for the dog’s attention. Lacking targeted
equivalence training, the dog is unlikely to generalize coming when called to the
park setting, with very dissimilar contextual cues and complete with a multitude
of stimuli competing for the dog’s attention.
Suppressing Unwanted Behaviour through Operant

Conditioning
A second primary objective for training dogs is to reduce or suppress the fre-
quency of undesirable behaviour. Pet dog owners find themselves faced with a
myriad of normal, yet objectionable dog behaviours: dogs bark at inappropriate
times, urinate on the furniture, pull on the leash, barge out the front door and
growl at strangers. Working dogs also display behaviours that interfere with suc-
cessful performance of their tasks. Herding dogs bite their charges; retrieving dogs
run off with the game; agility dogs are so keen that they leave obstacles prema-
turely; and guide dogs lead their owners too close to moving traffic. Some type of
punishment, whether positive or negative, may be required to eliminate such
problematic behaviour.
In this section, I outline the general processes by which dogs inhibit instru-
mental behaviour for the purpose of escaping or avoiding an unpleasant outcome,
in other words, learning by punishment. The stimuli used to motivate such learn-
ing are the same as for learning by reinforcement: appetitive and aversive stimuli.
However, negative punishment can result from the cessation or prevention of
appetitive stimuli; and positive punishment can result from the threat or presen-
tation of aversive stimuli. I also include a procedure called extinction that is not
punishment but still functions to produce a decline in behaviour.
Extinction
Extinction involves the discontinuation of reinforcement for a conditioned behav-

iour. As such, any associative learning, whether operant or classical, can be
subject to extinction. However, the most common application of extinction is to
cease reinforcement, whether intentional or adventitious, of established operant
behaviour. For instance, imagine the pet dog owner who, quite innocently,
teaches the dog to whine to be released from confinement. Initially, the dog
whines from the distress of separation. The owner releases the dog and whining
is negatively reinforced with escape from an unpleasant situation. Eventually the
Learning 135
dog learns to whine relentlessly whenever it is placed in confinement until the

owner breaks down and releases it. If the owner commits to ignoring the dog
when it whines, in other words, omitting the reinforcement, whining will extin-
guish.
The goal of extinction is to reverse the effects of acquisition. Extinction does
indeed lead to a gradual decline in the frequency of a conditioned behaviour. The
confined dog gradually, over time, whines less. However, the first effect of extinc-
tion is that the behaviour becomes more variable (Neuringer et al., 2001). Maybe
the dog tries barking, scratching or digging. Nonreinforcement produces frustra-
tion and frustration typically energizes behaviour (Amsel, 1962). Whining may
actually increase beyond previous levels, called an extinction burst. Frustration also
produces strong emotional responses, such as aggression (Azrin et al., 1966). If a
second animal is present, extinction-induced aggression may be directed toward
the innocent third party or toward the owner.
One of the hallmark effects of extinction is spontaneous recovery (Rescorla,
1997). Suppose the dog ceases whining and is released from confinement. The
next time the dog is placed in confinement, whining resumes for a while.
Recovery of the response is rarely complete and extinction occurs more rapidly
than the first time. With each subsequent period of confinement, the extent of
spontaneous recovery diminishes. The occurrence of spontaneous recovery can
be devastating to a dog owner. Other recovery effects can also cause a restoration
of the unwanted behaviour. Confine the dog in a different environment and
whining is likely to emerge; this is called renewal of the response. Renewal occurs
because extinction was linked to the contextual cues present in the original situa-
tion (Bouton and King, 1983). Exposing the dog to the reward of being released
from confinement, or even the stimuli leading up to the reward, such as attend-
ing to the dog, can also prompt resumption of whining. This is called reinstatement
(Bouton, 1994).
Yet another potential limitation of extinction is the partial reinforcement extinc-
tion effect. Prior to extinction, the conditioned behaviour was reinforced on either
a continuous or an intermittent schedule. Behaviours with an intermittent rein-
forcement history extinguish much more slowly than behaviours with a continu-
ous reinforcement history (Jenkins and Stanley, 1950). Think of how long the
string of losses would need to be for a habitual gambler to stop playing the slot
machines! Extinction that follows partial reinforcement also produces fewer frus-
tration responses. Resistance to extinction can also be beneficial. Desirable
behaviours that have been maintained on partial reinforcement are likewise
resistant to extinction. Dogs can be expected to perform reliably in the competi-
tion ring where tangible reinforcement is prohibited, if the training incorporated
intermittent reinforcement.
Instituting extinction is a popular procedure for reducing unwanted behav-
iour because it does not involve the use of aversive stimuli. However, there are so
many variables prompting restoration of the behaviour that the practical value of
extinction is extremely limited.
136 P. Reid
Negative punishment
Negative punishment can be used to discourage or suppress unwanted behaviour

in cases where the animal is anticipating a pleasant outcome. When the unwanted
behaviour occurs, delivery of the appetitive stimulus is prevented or, if the appet-
itive stimulus has already been presented, occurrence of the unwanted behaviour
causes withdrawal of the appetitive stimulus. Suppose an owner wants the dog to
cease jumping up to greet them upon their arrival home after an absence. The
dog jumps up in anticipation of a friendly reunion. If, instead, the owner turns
away and ignores the dog for a period of time, the anticipated reward is withheld.
This is an application of negative punishment and one would expect the likeli-
hood of jumping up to diminish.
Negative punishment sounds a lot like extinction but it is distinctly different.
Extinction occurs when a previously reinforced behaviour no longer produces
reinforcement. In the case of negative punishment, there is no requirement that
the behaviour be reinforced previously. The essence of negative punishment is that
when the unwanted behaviour occurs, sources of reinforcement are withheld for
some period of time. In most applications of negative punishment, it is commonly
referred to as time out. In some cases, only the existing reinforcement is withheld; in
other cases, all sources of reinforcement are removed. For instance, arranging for
the dog owner to turn away from the dog when it jumps up prevents the dog from
enjoying a social greeting. However, removing the dog to a small area of confine-
ment, with no access to alternative activities, temporarily restricts the dog from all
potential sources of reinforcement. While we might assume that removal from all
sources would be more effective, sometimes it is difficult to achieve because the act
of moving the dog to an isolation area can be problematic and, once there, dogs
are adept at finding ways to entertain themselves with other undesirable behaviour
(chewing, digging, barking, etc.). In general, short time-out durations of a few
seconds to a minute or two are effective (Lindsay, 2000). Longer time out periods
do not increase effectiveness (White et al., 1972).
The use of time out to eliminate unwanted behaviour during controlled situations,
such as training sessions, can be extremely effective, provided the dog finds training
pleasurable. For instance, a savvy trainer ignores the dog for a second or two when the
dog barks rather than performing the requested behaviour, another trainer requires
that the agility dog lie down for a moment after performing an obstacle incorrectly, and
still another trainer sends the dog into its crate after an infraction so that it must wait
while the trainer works another dog. However, it should be clear that these examples
of time out would not be appropriate for a dog that would prefer to avoid training!
Like extinction, time out is an appealing technique for reducing unwanted
behaviour because it does not involve the use of aversive stimuli. A few studies have
contrasted the effectiveness of extinction and negative punishment with positive
punishment, suggesting that: (i) negative punishment is more effective than extinc-
tion in suppressing unwanted behaviour; and (ii) negative punishment takes longer
than positive punishment to reduce the frequency of unwanted behaviour and that
suppression is rarely complete (Uhl and Sherman, 1971).
Learning 137
Positive punishment
Positive punishment is the presentation of an aversive stimulus, contingent upon

the occurrence of an unwanted behaviour. If the procedure is effective, the
unwanted behaviour should decrease in frequency or be eliminated altogether.
For example, suppose an owner wants to discourage the dog from getting up on
the couch. The owner could watch the dog carefully and each and every time the
dog jumps on the couch, the owner loudly shakes an aluminium can filled with
pebbles. After a few attempts, the dog learns that the way to avoid hearing the
noxious sound is to stay clear of the couch. Learning curves are often steeper with
positive punishment than with other operant conditioning procedures. Positive
punishment can produce dramatic suppression of behaviour in just a few repeti-
tions (Domjan, 2003).
Much of the research on positive punishment has been devoted to delineat-
ing its use. In general, positive punishment leads to more effective suppression of
behaviour when these conditions are met:
● The more intense and/or long lasting the aversive stimulus, the more behav-
ioural suppression (low intensity aversive stimulation produces only moderate
suppression and eventual response recovery) (Azrin, 1960).
● An aversive stimulus introduced at high intensity is more effective than if ini-
tially presented at low intensity and then increased (animals become resistant to
the effects of the aversive stimulus if it is initially presented at low intensity; con-
versely experience with high intensity aversive stimulation increase the effec-
tiveness of later low intensity aversive stimuli) (Azrin and Holz, 1966).
● Even response-independent positive punishment can lead to some suppression,
but response-dependent positive punishment is significantly more effective (in
other words, exposure to an aversive stimulus, whether contingent on behav-
iour or not, leads to some dampening down of all instrumental behaviour)
(Church et al., 1970).
● Immediate punishment (both negative and positive) is more effective than
delayed punishment (Camp et al., 1967).
● The degree of suppression produced by positive punishment is a direct function
of the number of responses punished (intermittent punishment can work, but
not as effectively as a continuous schedule of punishment) (Azrin, 1956).
Positive punishment is more effective when the animal is provided with other
appropriate activities for obtaining reinforcement (Herman and Azrin, 1964). For
instance, a dog jumps on the couch for the reinforcement of a soft place to lie
down. If the dog is provided with its own comfortable resting area, punishing the
dog for getting on the couch is more likely to work than if the dog has no alter-
native for obtaining the reinforcement. If a discriminative stimulus signals that
positive punishment is in effect, the animal will learn to suppress behaviour only
in the presence of that SD (Dinsmoor, 1952). If the dog is only punished for
getting on the couch when the owner is home, the dog will only learn to refrain
from getting on the couch when the owner is present.
138 P. Reid
The use of positive punishment (and negative reinforcement), because they

entail aversive stimuli, is highly controversial. This issue is not whether positive
punishment works; there is no question that the aversive control of behaviour is
effective. The concern is whether the use of positive punishment is ethical, espe-
cially for animals, because they cannot consent. When deciding whether the ben-
efits of positive punishment warrant its use, there are numerous factors to be
considered, including the expertise of the trainer, the seriousness of the unwanted
behaviour, the urgency of the need for behaviour change, and the quality of life
of the dog and of the humans coming into contact with the dog. Another concern
is that the line between positive punishment and abuse can be unintentionally
crossed, both by pet owners and skilled professionals. Indeed, dog-training texts
dating pre-1990s are peppered liberally with techniques that waver along that
line, such as scruff shakes, alpha rolls, severe collar corrections, hanging and heli-
coptering (e.g. Saunders, 1946; Koehler, 1962; Strickland, 1965; Monks of New
Skeet, 1978; Brown, 1983). Fortunately, as dog trainers become well versed in
learning theory, more humane applications are being developed. While it is
outside the scope of this chapter to explore the controversy in detail, let me
suggest that animal care professionals comply with the principle of the least aver-
sive alternative: that the least aversive technique that is reasonably expected to
succeed should be the technique of choice (Delta Society, 2001). Adherence to
this principle will at least give pause for thought and, hopefully, people will err on
the side of humane treatment.
Behaviour Change through Classical Conditioning
Pavlov’s pioneering work on how dogs come to anticipate the presentation of food
has proliferated into an extensive body of research exploring the formation of
associations between stimuli (Rescorla, 1988). A major focus has been the condi-
tioning of emotional reactions such as fear. Watson and Raynor (1920) demon-
strated the ease with which a child can learn to fear an object. ‘Albert’, a
9-month-old boy, initially showed no fear of a white rat, in fact he wanted to
touch it, but he did show distinct displeasure upon hearing the loud sound of a
hammer hitting a steel bar. The researchers then went about presenting the rat,
followed by the sound. After just two repetitions, Albert was reluctant to touch the
rat. With just a few more repetitions, Albert cried and tried to move away when-
ever he saw the rat. This is a classic example of a ‘conditioned emotional
response’.
Emotional responses, such as excitement, anxiety, fear and aggression, are
often triggered by specific stimuli. Envision the dog that gets wildly excited when
its owner picks up the leash, anticipating a walk in the park. Or the dog that
begins to pace and whine when its owner prepares to leave for the workday,
anxious at the prospect of being left alone. Or the dog that hides under the bed
when the owner picks up the nail clippers, fearing a painful grooming episode.
Conditioned emotional responses not only underlie many serious behavioural
Learning 139
problems, they can also hinder the learning and performance of desirable behav-
iour. A pet dog may be too anxious to benefit from training classes or a working
dog may be so aggressive that it can’t be trusted to perform in the required cir-
cumstances.
Classical conditioning procedures geared toward reducing interfering emo-
tional responses are typically more effective at remedying problems such as fear
and aggression than operant conditioning procedures geared toward addressing
the behaviour directly. For instance, suppose you have a dog that barks and
lunges aggressively at people entering your home. One option is to train the dog
to maintain a sit-stay and to suppress barking, presumably through negative or
positive punishment, when people enter the home. This is an operant condition-
ing approach and it would be effective in altering the dog’s behaviour, but not
necessarily the dog’s underlying emotional state (Wright et al., 2005). A second
option is to teach the dog a pleasant association: each time a person enters the
home, wonderful things happen to the dog. Maybe the visitor always brings a
large meaty bone for the dog. With sufficient repetitions, the dog comes to antic-
ipate bones when people come to the door. The dog now experiences a new kind
of conditioned emotional response, one of pleasure and excitement, rather than
anxiety and aggression. One would predict the dog’s behaviour to change from
lunging and barking to opening the door wide and welcoming guests to your
home!
Desensitization and counterconditioning
The most popular approach for eliminating problematic behaviours, such as

fearful and aggressive responses, is desensitization and classical counterconditioning
(DSCC). Wolpe (1958) was the first to describe the use of DSCC to eradicate con-
ditioned fear in cats. He taught hungry cats to fear the presence of food by associ-
ating the act of feeding with the delivery of electric shock. After the initial
conditioning, the cats refused to eat in the experimental room, despite significant
food deprivation. The cats also generalized to rooms that were similar to the
experimental room. Wolpe identified a room that was sufficiently unlike the exper-
imental room that the cats, though still anxious, were able to eat. Successive feed-
ings in the new room eliminated all signs of distress. The cats were then moved to
a room slightly more similar in appearance to the experimental room and offered
food. This was repeated in a series of rooms of increasing similarity to the experi-
mental room, each time remaining in the same room until distress was no longer
visible. Eventually the cats were able to eat comfortably in the original room.
This same technique can be applied to a fearful dog, such as that shown in
Fig. 8.5. This dog became tense when people reached for it. If escape were possi-
ble, it would retreat. If not, it would snap and bite. I started by pairing a hand
lifted in the air, but not directed toward the dog, with highly palatable treats.
Once the dog was comfortable with this arrangement, I progressed to orienting
the hand toward the dog. Gradually the hand was moved progressively closer to
140 P. Reid
Fig. 8.5. Photograph of chihuahua recoiling from a hand.
the dog, remaining at each step until the dog appeared to relax. Eventually, the
dog appeared eager to have a hand reaching out to touch it, even when treats
weren’t offered immediately (see Fig. 8.6). Because the dog no longer feared being
touched, defensive behaviours were replaced with approach and tail wagging.
Fig. 8.6. Photograph of chihuahua tolerating hand touch.

Learning 141
Classical counterconditioning, in particular, is an extremely powerful agent for

behaviour change. In most laboratory studies of counterconditioning (also known as
cross-motivational transfer experiments), an initially neutral stimulus is first paired
with one unconditioned stimulus, such as shock, and then, at a later time, the same
previously neutral stimulus is paired with a motivationally disparate stimulus, such as
food (Dickinson and Dearing, 1979). In even more convincing demonstrations,
though, two incompatible unconditioned stimuli are paired together. Erofeeva
(1921, cited in Dickinson and Dearing, 1979) used a strong electric shock to signal
the delivery of food to hungry dogs. The dogs initially responded with defensive
behaviours, such as struggling and yelping. Yet, as conditioning progressed, the dogs
began to show typical appetitive responses in response to the shock, including lip
licking and salivation. Even more surprising, after counterconditioning the aversive
stimulus to signal an appetitive stimulus, the aversive stimulus was incapable of func-
tioning as a punishing stimulus in an operant conditioning procedure. Classical
counterconditioning is so potent that it appears to actually produce a change in the
motivational and reinforcing properties of an unconditioned stimulus.
Concluding Remarks
Learning theory is one of the most in-depth and comprehensive accounts of

behaviour to appear in the psychological literature. It has generated countless
testable predictions, spawned entire research careers, and sprouted highly suc-
cessful applications in the clinical world of human and animal behaviour.
Learning theory provides an expansive and thorough framework for understand-
ing the processes by which dogs learn and how best to alter their behaviour. Such
knowledge is proving to spark the development of more effective, efficient and
humane methods of dog training and behaviour modification.
References
Amsel, A. (1962) Frustrative nonreward in partial reinforcement and discrimination
learning. Psychological Review 69, 306–328.
Azrin, N.H. (1956) Some effects of two intermittent schedules of immediate and
non-immediate punishment. Journal of Psychology 42, 3–21.
Azrin, N.H. (1960) Effects of punishment intensity during variable-interval rein-
forcement. Journal of the Experimental Analysis of Behavior 3, 123–142.
Azrin, N.H. and Holz, W.C. (1966) Punishment. In: Honig, W.K. (ed.) Operant
Behavior: Areas of Research and Application. Appleton-Century-Crofts, New
York, pp. 380–447.
Azrin, N.H., Hutchinson, R.R. and Hake, D.F. (1966) Extinction-induced aggression.
Journal of the Experimental Analysis of Behavior 9, 191–204.
Beerda, B., Schilder, M.B.H., van Hooff, J.A., de Vries, H.W. and Mol, J.A. (1998)
Behavioural, saliva cortisol and heart rate responses to different types of stimuli
in dogs. Applied Animal Behaviour Science 58, 365–381.
142 P. Reid
Bolles, R.C. (1970) Species-specific defense reactions and avoidance learning.

Psychological Review 71, 32–48.
Bouton, M.E. (1994) Conditioning, remembering, and forgetting. Journal of
Experimental Psychology: Animal Behavior Processes 20, 219–231.
Bouton, M.E. and King, D.A. (1983) Contextual control of the extinction of condi-
tioned fear: tests for the associative value of the context. Journal of Experimental
Psychology: Animal Behavior Processes 9, 248–265.
Brown, B.G. (1983) The No-force Method of Dog Training. Book One: Novice. H
and S Publications, Galesburg, Illinois.
Camp, D.S., Raymond, G.A. and Church, R.M. (1967) Temporal relationship
between response and punishment. Journal of Experimental Psychology 74,
114–123.
Chance, P. (1999) Learning and Behavior, 4th edn. Brooks/Cole, Pacific Grove,
California.
Church, R.M., Wooten, C.L. and Matthews, T.J. (1970) Discriminative punishment
and the conditioned emotional response. Learning and Motivation 1, 1–17.
Coppinger, R. and Coppinger, L. (2001) Dogs: A Startling New Understanding of
Canine Origin, Behavior, and Evolution. Scribner, New York.
Davis, M. (1974) Sensitization of the rat startle response by noise. Journal of
Comparative and Physiological Psychology 87, 571–581.
Delta Society (2001) Professional Standards for Dog Trainers: Effective, Humane
Principles. Delta Society, Bellevue, Washington.
Dickinson, A. and Dearing, M.F. (1979) Appetitive-aversive interactions and
inhibitory processes. In: Dickinson, A. and Boakes, R.A. (eds) Mechanisms of
Learning and Motivation: A Memorial to Jerzy Konorski. Lawrence Erlbaum
Associates, Hillsdale, New Jersey, pp. 203–231.
Dinsmoor, J.A. (1952) A discrimination based on punishment. Quarterly Journal of
Experimental Psychology 4, 27–45.
Domjan, M. (2003) The Principles of Learning and Behavior, 5th edn.
Wadsworth/Thomson, Belmont, California.
Eisenberger, R. (1992) Learned industriousness. Psychological Review 99, 248–267.
Ferster, C.B. and Skinner, B.F. (1957) Schedules of Reinforcement. Appleton-
Century-Crofts, New York.
Grice, G.R. (1948) The relation of secondary reinforcement to delayed reward in
visual discrimination learning. Journal of Experimental Psychology 38, 1–16.
Herman, R.L. and Azrin, N.H. (1964) Punishment by noise in an alternative response
situation. Journal of the Experimental Analysis of Behavior 7, 185–188.
Jenkins, W.O. and Stanley, J.C. (1950) Partial reinforcement: a review and critique.
Psychological Bulletin 47, 193–234.
Koehler, W.R. (1962) The Koehler Method of Dog Training. Howell Book House,
New York.
Kremer, E.F. (1971) Truly random and traditional control procedures in CER condi-
tioning in the rat. Journal of Comparative and Physiological Psychology 76,
441–448.
Leaton, R.N. (1976) Long-term retention of the habituation of lick suppression and
startle response produced by a single auditory stimulus. Journal of Experimental
Psychology: Animal Behavior Processes 2, 248–259.
Lett, B.T. (1973) Delayed reward learning: disproof of the traditional theory.
Learning and Motivation 3, 237–246.
Learning 143
Lieberman, D.A., McIntosh, D.C. and Thomas, G.V. (1979) Learning when reward
is delayed: a marking hypothesis. Journal of Experimental Psychology: Animal
Behavior Processes 5, 224–242.
Lindsay, S.R. (2000) Handbook of Applied Dog Behavior and Training: Vol. 1.
Adaptation and Learning. Iowa State University Press, Ames, Iowa.
Lockard, J.S. (1963) Choice of a warning signal or no warning signal in an unavoid-
able shock situation. Journal of Comparative and Physiological Psychology 56,
526–530.
Mackintosh, N.J. (1974) The Psychology of Animal Learning. Academic Press,
London.
Martin, G. and Pear, J. (1996) Behavior Modification: What It Is and How To Do It,
5th edn. Prentice Hall, Upper Saddle River, New Jersey.
Mazur, J.E. (2002) Learning and Behavior, 5th edn. Prentice Hall, Upper Saddle
River, New Jersey.
McConnell, P.A. (1990) Acoustic structure and receiver response in domestic dogs,
Canis familiaris. Animal Behaviour 39, 897–904.
McConnell, P.A. and Bayliss, J.R. (1985) Interspecific communication in cooperative
herding: acoustic and visual signals from human shepherds and herding dogs.
Zeitschrift für Tierpsychologie 67, 303–328.
Miller, V. and Domjan, M. (1981) Selective sensitization induced by lithium malaise
and footshock in rats. Behavioral and Neural Biology 31, 42–55.
Miller, P.E. and Murphy, C.J. (1995) Vision in dogs. Journal of the American
Veterinary Medical Association 207, 1623–1634.
Monks of New Skeet (1978) How To Be Your Dog’s Best Friend. Little, Brown,
Boston, Massachusetts.
Morse, W.H. (1966) Intermittent reinforcement. In: Honig, W.K. (ed.) Operant
Behavior: Areas of Research and Application. Appleton-Century-Crofts, New
York, pp. 52–108.
Myers, R.D. and Mesker, D.C. (1960) Operant conditioning of a horse under several
schedules of reinforcement. Journal of the Experimental Analysis of Behavior 3,
161–164.
Neuringer, A., Kornell, N. and Olufs, M. (2001) Stability and variability in extinc-
tion. Journal of Experimental Psychology: Animal Behavior Processes 27,
79–94.
Pavlov, I.P. (1927) Conditioned Reflexes (G.V. Anrep, translation). Oxford
University Press, London.
Pierce, W.D. and Cheney, C.D. (2004) Behavior Analysis and Learning. Lawrence
Erlbaum Associates, Mahwah, New Jersey.
Pryor, K. (1984) Don’t Shoot the Dog: The New Art of Teaching and Training.
Bantam Books, New York.
Rescorla, R.A. (1988) Pavlovian conditioning: it’s not what you think it is. American
Psychologist 43, 151–160.
Rescorla, R.A. (1997) Spontaneous recovery after Pavlovian conditioning with mul-
tiple outcomes. Animal Learning and Behavior 25, 99–107.
Revusky, S. (1971) The role of interference in association over a delay. In: Honig,
W.K. and James, P.H.R. (eds) Animal Memory. Academic Press, New York, pp.
155–213.
Salzinger, K. and Waller, M.B. (1962) The operant control of vocalization in the dog.
Journal of the Experimental Analysis of Behavior 5, 383–389.
144 P. Reid
Saunders, B. (1946) Training You to Train Your Dog. Doubleday, New York.
Shettleworth, S.J. (1972) Constraints on learning. Advances in the Study of Behavior
4, 1–68.
Skinner, B.F. (1938) The Behavior of Organisms. Appleton-Century-Crofts, New
York.
Skinner, B.F. (1951) How to teach an animal. Scientific American 185, 26–29.
Slabbert, J.M. and Rasa, O.E.A. (1997) Observational learning of an acquired mater-
nal behaviour pattern by working dog pups: an alternative training method?
Applied Animal Behaviour Science 53, 309–316.
Stokes, T.F. and Baer, D.M. (1977) An implicit technology of generalization. Journal
of Applied Behavior Analysis 10, 349–367.
Strickland, W.G. (1965) Expert Obedience Training for Dogs. Macmillan, New York.
Terrace, H.S. (1963) Discrimination learning with and without ‘errors’. Journal of the
Experimental Analysis of Behavior 6, 1–27.
Thompson, R.F. and Spencer, W.A. (1966) Habituation: a model phenomenon for
the study of neuronal substrates of behavior. Psychological Review 73, 16–43.
Thorndike, E.L. (1911) Animal Intelligence. Macmillan, New York.
Thorpe, W.H. (1963) Learning and Instinct in Animals. Harvard University Press,
Cambridge, Massachusetts.
Todes, D.P. (1997) From the machine to the ghost within: Pavlov’s transition from
digestive physiology to conditioned reflexes. American Psychologist 52,
947–955.
Uhl, C.N. and Sherman, W.O. (1971) Comparison of combinations of omission,
punishment, and extinction methods in response elimination in rats. Journal of
Comparative and Physiological Psychology 74, 59–65.
Watson, J.B. and Raynor, R. (1920) Conditioned emotional reactions. Journal of
Experimental Psychology 3, 1–14. Reprinted in American Psychologist 55,
313–317.
Weinstock, S. (1958) Acquisition and extinction of a partially reinforced running
response at a 24-hour intertrial interval. Journal of Experimental Psychology 46,
151–158.
White, G.D., Nielsen, G. and Johnson, S.M. (1972) Timeout duration and the sup-
pression of deviant behavior in children. Journal of Applied Behavior Analysis
5, 111–120.
Whiten, A. and Ham, R. (1992) On the nature and evolution of imitation in the
animal kingdom: reappraisal of a century of research. Advances in the Study of
Behavior 21, 239–283.
Williams, B.A. (1991) Marking and bridging versus conditioned reinforcement.
Animal Learning and Behavior 19, 264–269.
Wolpe, J. (1958) Psychotherapy by Reciprocal Inhibition. Stanford University Press,
Stanford, California.
Wright, J.C., Reid, P.J. and Rozier, Z. (2005) Treatment of emotional distress and dis-
orders – non-pharmacological methods. In: McMillan, F.D. (ed.) Mental Health
and Well-being in Animals. Blackwell Publishing, Ames, Iowa, pp. 145–157.
III The Dog in Its Niche: Among
Humans
The ecological concept of a ‘niche’ is used in many places throughout this book.
In normal biological texts, a niche is a specific living environment, signified by all
specific aspects of how a population utilizes its living space, such as the food eaten,
the geographical place occupied, etc. It has become increasingly obvious that for
dogs the most fundamental aspect of the niche it occupies is ourselves. One might
therefore say that dogs have come to occupy the ecological niche of living with
humans. The third part of the book explores the behavioural biology of dogs from
this perspective.
A first glimpse of how dogs have adapted to this niche is offered in Chapter
9, which deals with the behaviour of feral dogs. These are populations of dogs
which have more or less escaped from the most intense life with humans, in that
they are not living in families. The traits of such dogs therefore offer many insights
into how their behaviour has been modified since they roamed the continents as
wild wolves. Chapter 10 introduces the problem of selecting and breeding dogs
for various working purposes. Here the author provides an overview of some
important aspects of modern population genetics and applies this to dog breeding
and selection.
Any selection and breeding obviously depends on how well we can charac-
terize the behaviour we intend to select for. A much overlooked aspect of dog
behaviour is the concept of personality, and this forms the central part of Chapter
11. The concept raises fundamental issues with respect to its definition, but also
of course necessitates careful consideration when one wishes to measure person-
alities. The last chapter of this part, Chapter 12, deals with the way in which dogs
and humans have co-evolved to an extent not paralleled by any other domestic
species. The author provides a number of intriguing research results concerning
the social cognitive abilities of dogs, and also some unexpected aspects of human
competence in communicating with dogs.
Behaviour and Social Ecology of
9
Free-ranging Dogs
Luigi Boitani, Paolo Ciucci and Alessia Ortolani
There Are Many Kinds of Free-ranging Dogs
More than 350 officially recognized dog breeds and the endless variety of mon-
grels are the best evidence of the broad morphological adaptability of dogs to dif-
ferent environments, functions and human-driven selection. Similar plasticity,
although not equally evident, can be found in dogs’ behaviour and ecology when
they are exposed to the variety of artificial and natural selections in almost any
possible environment on Earth. Because of this diversity of phenotypes and func-
tional specializations, dogs are a highly diverse biological group that allows only
limited and superficial generalizations. In particular, when studying dog behav-
iour, it is imperative to distinguish at least a few broad dog categories defined by
their social context and their individual behavioural ontogenies. Dogs have been
classified on the basis of: behavioural and ecological traits (Scott and Causey,
1973; Causey and Cude, 1980); their origins (Daniels and Bekoff, 1989a, b); their
main range type (rural vs urban free-ranging: Berman and Dunbar, 1983; those
having unrestricted access to public property: Beck, 1973); and the extent of their
dependency on humans (World Health Organization (WHO), 1988). The latter
classification was designed to improve dog rabies control measures (Perry, 1993)
and is useful to define dogs in relation to their association with humans: (i)
restricted dog: fully dependent (all its essential needs provided intentionally by
humans) and fully restricted by humans; (ii) family dog: fully dependent and semi-
restricted; (iii) neighbourhood dog: semi-dependent (part of its essential needs
intentionally provided by humans) and either semi-restricted or unrestricted; and
(iv) feral dog: independent (none of its essential needs intentionally provided by
humans) and unrestricted. This classification, however, is biased toward an

(ed. P. Jensen)
148 L. Boitani et al.
anthropic perspective as it considers dogs’ status only in relation to human

dependency and control.
Boitani and Fabbri (1983) and Boitani et al. (1995) proposed a slightly differ-
ent classification (four categories: owned restricted, owned unrestricted, stray and
feral dogs) to account for the individual ontogeny of dog–human association, i.e.
the nature and development of a social bond with humans, and the ecology of
dogs at varying degrees of human dependency. Under this classification, the first
two categories are similar to WHO’s (1988) categories, i.e. restricted and family
dogs. The family dogs have an owner on whom they depend, but may be left free
to roam (Hsu et al., 2003). The third category, stray dogs, includes dogs living in
a human-dominated context: it is an heterogeneous group of dogs, comprising: (i)
dogs that still have a social bond with humans, possibly abandoned or born into
human settings; and (ii) dogs with different degrees of fear/tolerance towards
humans, who are attracted near human settings by the availability of food and
shelter resources regardless of whether these resources are intentionally provided
by humans or are casually associated with them (e.g. handouts, refuse tips or
garbage dumps for food, infrastructures for shelter, etc.). These dogs have also
been called ‘village dogs’ (Macdonald and Carr, 1995; Ortolani and Coppinger,
2005) and have been consistently reported by many authors; those roaming many
of the large cities of the Mediterranean basin (Istanbul, Alexandria) were
described in the 18th century almost as a separate subspecies (cf. Brehm, 1893).
The fourth category, feral dogs, includes all dogs living in a wild and free state
with no direct food or shelter intentionally supplied by humans (Causey and
Cude, 1980), showing no evidence of socialization to humans (Daniels and Bekoff,
1989a), but rather displaying a strong continuous avoidance of direct human con-
tacts (Boitani and Ciucci, 1995; Boitani et al., 1995) and a lifestyle restricted
mainly to natural environments. Taking into account the social bond with
humans and the ecological context adds important insights into free-ranging
dogs’ behaviour and has substantial implications for their classification and man-
agement.
Dogs’ categories, however, are not homogeneous and the distinction among
feral, stray/neighbourhood/village, and other free-ranging dogs is often a matter
of degree (Nesbitt, 1975; Boitani and Ciucci, 1995), based on in-depth knowledge
of their behaviour and ecology. Moreover, dogs from different categories can
form temporary heterogeneous groups that are difficult to classify, and individual
animals can change their status during their life depending on local ecological and
social contexts, as well as following management interventions. Regardless of the
names given to the various dog categories, most authors agree that ‘owned’,
‘stray’ and ‘feral’ dogs are not closed categories and that dogs may change their
status throughout their life (Scott and Causey, 1973; Nesbitt, 1975; Hirata
et al., 1987; Daniels, 1988; Daniels and Bekoff, 1989a; Boitani et al., 1995).
The change that appears to be most difficult to recognize in the field is
between the stray/neighbourhood/village dogs and the feral dogs, especially as it
requires conceptual clarity on the dynamics of the feralization process.
Feralization can be viewed from two different perspectives: (i) at the population
Behaviour and Social Ecology of Free-ranging Dogs 149
level, as the domestication process in reverse (Hale, 1969; Brisbin, 1974; Price,
1984); and (ii) at individual level, as a behavioural process (Daniels and Bekoff,
1989c), as shown by Boitani et al.’s (1995) observation of feralization occurring
within an individual’s lifetime. The two perspectives are not mutually exclusive
and imply different temporal scales as well as different theoretical and research
approaches (Daniels and Bekoff, 1989c; Boitani and Ciucci, 1995).
A dog can change status as the result of several causes acting in complex
interactions (Fig. 9.1). An owned dog can become stray by increasingly loosening
restriction, by abandonment or simply by being born to a stray mother (Beck,
1975; Perry, 1993; Hsu et al., 2003). A stray dog can become feral when forced
out of a human environment or when accepted by a feral group (Daniels, 1988;
Daniels and Bekoff, 1989a, c; Boitani et al., 1995), like the majority of group
members studied by Boitani et al. (1995). In changing status, dogs pass through
intermediate behaviours depending on local social and ecological conditions, and
the process may take a significant portion of an individual’s life-span or may never
be completed. This may be the case of many stray/neighbourhood/village dogs
that live in the marginal areas of large human settlements, especially in tropical
regions (Pal et al., 1998a, b, 1999; Ortolani and Coppinger, 2005).
Daniels and Bekoff (1989c) suggested that feralization occurs through the
development of a fear response to humans and they implied that the reverse
process is possible. Indeed, dogs that shifted to a feral status can return to condi-
– reproduction FAMILY DOGS – death

– adoption
– abandonment – adoption
– unrestrainment – intraspecific competition
– escape – capture and restrainment
– reproduction NEIGHBOURHOOD/ – human control

– recruitment VILLAGE DOGS – natural death
– human interference
– cooption
– reproduction – human control

– recruitment FERAL DOGS
– natural death
Fig. 9.1. A model of the feralization process. Population input and output mecha-
nisms for each dog category (horizontal arrows) and the possible factors facilitating
shifting status between different dog categories (vertical arrows).
tions of acceptance or tolerance of humans, as observed by Boitani et al. (1995)

and experimentally shown by one of us by rehabilitating a feral dog to a domes-
tic status (P. Ciucci, unpublished data). However, dogs that have not experienced
contacts with humans during the first 2–3 weeks of life, will remain human-shy
and may never be able to develop a full social bond with humans despite their
socio-cognitive predisposition to attend to human signals (Hare et al., 2002; see
also Chapter 4 in this volume). The contrary, changing from stray to owned
status, is more than often demonstrated by the many stray dogs adopted by
humans.
The Behaviour of Neighbourhood/Village Dogs and their

Interactions with People
Free-ranging dogs that associate with human settlements (hence village dogs) all
share similar behavioural-ecological characteristics. They live mostly unre-
strained, roving through the streets of villages, towns or cities, feeding mainly on
human refuse and garbage, and using buildings, porches, vehicles or other
human-built structures for shelter. Researchers who studied them in different
parts of the world found that many village dogs actually have owners (Beck, 1975;
Daniels, 1983a; Boitani and Racana, 1984; Brooks, 1990; DeBalogh et al., 1993;
Butler and Bingham, 2000; Ortolani and Coppinger, 2005), although they are
not family pets, as we might think of by a Western concept. These dogs tend to
associate with particular households (Boitani and Racana, 1984; DeBalogh et al.,
1993), even when the home-owners claim that the dogs don’t belong to them
(Ortolani and Coppinger, 2005). Most of them are solitary, although pairs or trios
are not unusual, while large groups are fairly uncommon (notable exceptions are
temporary gatherings of males around an oestrous female (Pal et al., 1999) or dogs
gathering at a common food source (Macdonald and Carr, 1995)) (Beck, 1975;
Berman and Dunbar, 1983; Daniels, 1983a; Boitani and Racana, 1984; Hirata et
al., 1986; Daniels and Bekoff, 1989b; Macdonald and Carr, 1995; Ortolani and
Coppinger, 2005). They are mainly active in the early hours of the morning and
in the evening (Beck, 1975; Berman and Dunbar, 1983; Daniels, 1983a; Boitani
and Racana, 1984; Hirata et al., 1986); periods that often coincide with cooler
temperatures of the day, especially in tropical regions. The majority of them are
likely to be males, as the adult sex ratio is skewed in their favour, often in a ratio
of two males for each female, but sometimes males can be as much as five times
more abundant than females (Beck, 1973; Daniels, 1983b; Boitani and Racana,
1984; WHO, 1988; Daniels and Bekoff, 1989b; Brooks, 1990; DeBalogh et al.,
1993; Macdonald and Carr, 1995; Pal et al., 1998a, b; Butler and Bingham, 2000;
Ortolani and Coppinger, 2005).
Village dogs may have been the earliest canines to associate with people in a
permanent way. Some 15,000 years ago, when humans traded a nomadic lifestyle
for a sedentary one and the first villages were built, human-generated refuse
began piling up around them: leftovers of family meals, butchering scraps,
carcasses, by-products of farm animals, as well as human excrement. This newly

created resource, the garbage dump, opened up a niche appealing especially to
scavenger-prone carnivore species, with one caveat: tolerance to human proxim-
ity would have been the price to pay for admission. As Coppinger and Coppinger
(2001) have suggested, dog domestication might indeed have taken this route:
individuals who displayed lower flight distances towards humans around garbage
dumps would have been selected to exploit this resource, and dogs’ long-lasting
co-existence with people would have thus begun (see also Chapters 2 and 3 in this
volume for a further, in-depth, discussion of the timing and process of dog domes-
tication). There is plenty of evidence today that village dogs gain food benefits by
associating with humans (Beck, 1973; Daniels, 1983a; Boitani and Racana, 1984;
Pal et al., 1998a, b; Brooks, 1990; Macdonald and Carr, 1995; Butler and
Bingham, 2000). In fact, village dogs may depend entirely on human garbage for
survival, as their hunting skills are not very effective (see Boitani and Ciucci,
1995). The reward for trading a hunting life for one of scraps may be that village
dogs are rarely found in poor body condition (DeBalogh et al., 1993; Ortolani and
Coppinger, 2005).
Shelter is another benefit supplied by villages, either intentionally or unin-
tentionally provided by people, which dogs require for protection against harsh
climatic conditions, aggressive conspecifics or carnivore competitors, and/or for
hiding young from danger. But sharing a niche with humans has its costs. One
characteristic shared by village dog populations around the world is a high rate of
infant mortality: as many as 75% of litters can be destroyed by people
(Macdonald and Carr, 1995) and often less than 20% of pups reach adulthood
(WHO, 1988; Pal et al., 1998a). Selective culling of female pups in litters is
another common practice (A. Ortolani, 2003, personal observation; Daniels and
Bekoff, 1989b), which most likely causes the biased sex ratio in the adult popula-
tion. Finally, human culling, disease and car accidents together can cause more
than 80% of adult mortality in some areas (DeBalogh et al., 1993). Although these
figures seem high and might give the impression that the costs outweigh the ben-
efits for village dogs, similar mortality rates have also been reported for feral dogs
(Boitani and Ciucci, 1995). More importantly, recorded population densities of
village dogs range from 68 dogs/km2 to 3700 dogs/km2 in different parts of the
world (WHO, 1988; Brooks, 1990), suggesting that high mortality rates are not
putting these canines at risk of extinction. Although the diversity of dog–human
relationships may mask the adaptive value of several dogs’ traits, it appears that,
at the population level, the costs of living with humans are probably negligible.
The association of village dogs with human households, although driven by
the presence of a food source, might predispose them to become adopted and
eventually become ‘family dogs’. However, it does not imply that village dogs seek
companionship with people, as in the popular image of a stray looking for an
owner, nor does it imply that all village dogs could become family pets. In a recent
study of human–dog interactions in Ethiopia, Ortolani and Coppinger (2005)
found that most dogs avoided humans by fleeing when approached and that only
a small minority (4%) would approach an observer. Moreover, many dogs (22%)
displayed alert/alarm behavioural responses, such as barking and growling, upon

seeing an unfamiliar person, while some (11%) were downright aggressive,
attempting to attack the observer. Dogs were more likely to vocalize if they were
found near a household, suggesting that they might have been defending a
resource or territory. Boitani and Racana (1984) also reported that village dogs
barked near houses and livestock. Most Ethiopian village dogs could only be
approached at a distance of 5 m or more, after which the majority either fled or
displayed aggressive behaviour (Ortolani and Coppinger, 2005). Interestingly,
such dogs tended to be alone or in pairs, while dogs encountered in small groups
displayed mainly neutral, or indifferent, behaviours towards the observer. All
aggressive dogs vocalized towards the observer before reacting, suggesting that,
even though they might just have been bluffing, it is a wise person who pays atten-
tion to dogs’ warning signs!
The majority of village dogs studied in different surveys were usually encoun-
tered in the streets and their main activity was either lying down resting or travelling
around the village (Beck, 1975; Berman and Dunbar, 1983; Ortolani and
Coppinger, 2005). Intraspecific agonistic interactions were not common but they
appeared to peak during the oestrous season and usually between males (Daniels,
1983b; Pal et al., 1999). Studies of dog–human interactions are scarce, and they are
likely to be influenced by the general experience that dogs have formed with people
habitually encountered in their environment. People’s attitudes towards dogs vary
widely among cultures, individual predispositions and even religious beliefs
(Wandeler et al., 1993). Curiously, in Ethiopia, Ortolani and Coppinger (2005) found
that in predominantly Muslim towns the majority of dogs tended to avoid or have
aggressive reactions towards an unfamiliar observer, while in predominantly
Christian Orthodox villages the majority of dogs showed neutral reactions towards
the same observer. Obviously, there are many factors, other than religion, that could
be involved in explaining differences in dogs’ reactions to people in those villages, but
these results do suggest that people’s attitudes may influence village dog behaviour.
Clearly, more studies quantifying behavioural responses of village dogs towards
people in different areas of the world are needed, as well as more refined measures
to quantify people’s attitudes towards dogs in different cultures. Such studies would
be particularly important in areas where potentially fatal dog-transmitted diseases,
such as rabies, are still prevalent (see WHO, 2001). Although data are limited,
village dogs may be the most abundant category of dog in the world (Beck, 1975;
WHO, 1988) and, so far, they have been the most overlooked of all canines. Yet
they hold the key to understanding how a predator evolved into the most faithful
human companion. Studying the behaviour of village dogs is of critical importance
to shed light on the complexities of human–dog interactions in our society.
The Behaviour of Feral Dogs
The behaviour and ecology of feral dogs result from the complex interaction
of a suite of biological traits which still resemble those of their ancestors (wolves),
especially concerning their ecological flexibility and the great variety of artificial
and natural environments that they can live in. Many of the ancestral traits, such
as a group-living tendency, territoriality, predatory instincts and a large degree of
ecological flexibility, are still evident in the dogs’ biology, but most of these traits
appear void of their original adaptive value and may represent ‘evolutionary
inertia’ or artificial selection epiphenomena (Boitani and Ciucci, 1995). If the
process of domestication had the effect of enhancing the behavioural and ecolog-
ical flexibility that allowed dogs to survive in a wide range of semi-natural envi-
ronments, it also reduced their overall fitness to cope with long-term wild habitat
conditions (Price, 1984). Feral dogs are generally not reproductively self-sustain-
ing: they suffer from high rates of juvenile mortality; they depend indirectly upon
humans for food, co-opting new individuals, and space; and their demography
appears dominated by stochastic and unpredictable mechanisms (Boitani et al.,
1995). Most studies of feral dogs’ behaviour and ecology have been carried out on
individuals which had been feral only for a few generations, and thus these gen-
eralizations may not apply to feral dogs such as dingoes and pariah dogs, which
have been living under the effects of natural selection forces for longer periods of
time.
Social ecology
The social structure of feral dogs is an aggregation of monogamous breeding pairs

and their associates (pups and/or subadults of pair members), which is substan-
tially different from the highly structured hierarchy of wolf packs, where domi-
nance is respected both in terms of privilege (e.g. ‘pecking order’) and initiative
(e.g. travelling, hunting, territorial defence, reproduction, etc.) (Mech and
Boitani, 2003). Wolves’ pack size is regulated also through social control of repro-
ducing individuals (Packard, 2003), a mechanism which is apparently absent in
dogs (Boitani et al., 1995). Intraspecific agonistic behaviour of feral dogs seems to
be restricted primarily at the individual level without extensive effects on the
social organization of the group, such as forcing a hierarchy and building a robust
cohesiveness among group members. Scott and Fuller (1965) suggested that the
agonistic behaviour of dogs toward humans is strongly influenced by their early
contacts with humans and this might be particularly relevant when interpreting
the behaviour of feral dogs recruited from the house and stray categories.
Therefore, observations of feral dog behaviour without knowledge of the com-
plete individual history of all group members are not sufficient to understand the
causes and mechanisms of intraspecific relationships.
Members of feral dogs’ groups are generally not related (Scott and Causey,
1973; Nesbitt, 1975; Causey and Cude, 1980; Berman and Dunbar, 1983;
Daniels and Bekoff, 1989a, b; Boitani et al., 1995). The basic social unit
is the breeding pair but the social bond is only loosely extended to the rest of the
group members without the complex rules that regulate pack life in most wild
canids (Kleiman and Eisenberg, 1973; Bekoff et al., 1984; Gittleman, 1989). In
particular, the differences with wolf packs, which are highly cohesive one-family
units that hunt, rear young and protect a communal territory as a stable group
(Mech and Boitani, 2003), are obvious and prompted Boitani and Ciucci (1995)
to propose the term ‘group’, as more appropriate than pack, for a feral dogs’
social unit. Most studies of feral dogs found that group size ranges from two to six
animals (two to five in Alabama: Scott and Causey, 1973; two to six in Alabama:
Causey and Cude, 1980; two to four in Arizona: Daniels and Bekoff, 1989b; five
to six in Illinois: Nesbitt, 1975; three to six in Italy: Boitani et al., 1995), in con-
trast with the smaller social unit of neighbourhood/village dogs, which are
reported mostly in pairs or alone (see above on village/neighbourhood dogs).
Macdonald (1983) suggested that in carnivores the quantity and distribution of
food resources are the main determinants of group size. In wolves, pack size
appears to be affected primarily by prey abundance, as changes in prey availabil-
ity correspond to proportional changes in pack size (Mech and Boitani, 2003). It
could be speculated that neighbourhood/village dogs would have little advantage
from living in groups because resources are more scarce and scattered than those
available to feral dogs, but also because cooperative ‘hunting’ would not be an
advantage when resources are plentiful and easy to obtain. Also, lack of natural
predators or competitors in human settings would not offer particular advantages
to larger groups of stray dogs (cf. Macdonald and Carr, 1995), a situation that
could be different in the case of feral dogs. However, even if studies with accurate
estimates of food resources were available, any attempt to draw theoretical gen-
eralizations would be little more than a speculation, as we believe that theoretical
analyses of canid evolutionary strategies are of limited value when carried out on
animals which have been living under both artificial and natural selection pres-
sures.
Boitani et al. (1995) reported that in their study area garbage dumps supplied
food in excess during all seasons, and group size appeared to be related more to
social factors, such as group turnover, rather than ecological ones, even though
communal resource defence against wolves and other feral dog groups could have
been a factor. Although no evidence of any intrinsic regulatory mechanism was
found, the size of the feral dog group studied in Italy between 1984 and 1987
remained fairly stable (Boitani et al., 1995), and all events affecting group size
appeared mostly due to density-independent and external factors (i.e. casual
human persecution, availability of stray dogs, climatic conditions, etc.) with the
sole exception of the cooption of stray dogs into the group (see below). Mortality
of most sexually mature individuals was caused by human interference, while
newborns from feral parents contributed almost nil to long-term group stability.
Group size was maintained stable only by recruiting new members from the
village dog population; at the end of the study, all but one dog in the group were
of stray origin. Recruitment of village dogs into the feral group occurred mostly
(but not only) during the breeding period and in conjunction with the disruption
of a breeding pair. A single adult, after loosing its mate, would actively coopt
another sexually mature dog, and this new member would in turn become
socially accepted by the entire feral group. Thus, under the conditions studied by
Boitani et al. (1995) the ‘vacancy’ in a pair bond seems to be the main mechanism
triggering cooption, although more data are needed from different ecological
contexts to allow for meaningful generalizations.
One related interesting question, which remains unanswered, concerns the
mechanisms, if any, that regulate the upper limit of group size in feral dogs. It
may be speculated that given unlimited resources the lack of a firm social hierar-
chy and strong social bonds poses no upper limit to the number of feral dogs that
can associate in a group. However, if functionality (in hunting, territorial defence,
offspring care, etc.) is a prerequisite for the existence of an effective unit, then an
upper limit appears to be naturally imposed by the costs of enduring effective
cooperation. Since dogs form less efficient functional units compared to wolves,
this might in part explain the smaller size of feral dog groups (Scott and Causey,
1973; Nesbitt, 1975; Causey and Cude, 1980; Daniels and Bekoff, 1989b; Boitani
and Ciucci, 1995; Boitani et al., 1995). Moreover, the social structure of feral dogs
may not allow for an efficient density-dependent mechanism of population regu-
lation in relation to environmental and ecological conditions, making feral dog
groups more susceptible to stochastic events and limiting factors. Even when pre-
dictable and abundant food resources are available to feral dogs, the higher
number of females reproducing per group, with their di-oestrous cycle and overall
negative energetic balance, the lack of non-reproducing ‘auxiliaries’, and the high
pup/juvenile mortality all interact to determine low recruitment rates and to
maintain group size through external recruitment. In feral dogs, therefore, group
size and composition appear to be a function of food abundance and availability
of potentially cooptable village dogs at the lower end and, at the other extreme,
of individual dogs’ physiology and social behaviour, which hardly translates into
functional and cohesive larger groups.
Home-range and territoriality
Feral dogs spend most of their life within well-defined home-ranges whose inter-
nal portions (core areas) are often defended against intruders, although their ter-
ritorial behaviour is highly variable depending on several environmental and
human-related causes (Scott and Causey, 1973; Causey and Cude, 1980; Gipson,
1983; Daniels and Bekoff, 1989a; Boitani et al., 1995; Meek, 1999). Home-range
sizes obtained by radio-telemetry studies range from 4.4–10.4 km2 for three
groups in east-central Alabama (Scott and Causey, 1973), to 18.7 km2 in Alabama
(Causey and Cude, 1980), 57.8 km2 in central Italy (Boitani et al., 1995) and 70
km2 in Alaska (Gipson, 1983). Home-range size does not appear to be closely
related to group size, but rather is dependent on the spatial patterns of key
resource sites, such as denning areas, refuge areas, garbage dumps and other food
sources. Dogs may show seasonal variations in patterns of range utilization, using
smaller portions at different times. Daniels and Bekoff (1989a) and Scott and
Causey (1973) related these core-area variations to the presence of dependent
pups and to different energetic requirements of the group. In contrast, Boitani
et al. (1995) found that several other factors could possibly be involved in shifting
of the core areas within the home-range: the exploitation of temporary food
resources (i.e. a large livestock carrion), disturbance caused by humans, denning
activities, previous spatial-use patterns of newly recruited dogs, unpredictable
fluctuation in food availability at dumps, and possible interference by wolves. In
the same study, it was suggested that shifting of core areas and seasonal ranges
were not necessarily only a consequence of direct environmental changes, in fact
previous knowledge of the area by a new member of the group could affect the
entire group’s spatial behaviour.
Home-range sizes tend to become smaller in neighbourhood/village dogs
living closer to or within areas inhabited by humans (from 2–11 ha up to 61 ha)
(Beck, 1973; Fox et al., 1975; Berman and Dunbar, 1983; Daniels, 1983a;
Santamaria et al., 1990), confirming that predictability and quantity of food
sources as well as small group sizes are among the determinants of ranging behav-
iour. In feral dogs, home-ranges, or parts of them, especially core areas, food
sources and den sites, are actively defended using scent marking, vocalizations
(barking) and aggressive behaviour. Both Macdonald and Carr (1995) and
Boitani et al. (1995) reported for the same study area in central Italy that defen-
sive behaviour was consistently shown within the entire core areas and during the
whole year. The higher frequency of territorial behaviour found in Italy com-
pared to other studies (Scott and Fuller, 1965; Bekoff, 1979; Berman and Dunbar,
1983; Boitani and Racana, 1984; Daniels and Bekoff, 1989a) might be related to
a higher level of integration within the group, a higher degree of isolation from
other dogs, and to food resources being concentrated mostly in localized patches
at the dumps. In addition, an unknown effect might be due to the typical range
defending behaviour of the Maremma dog, a large guarding dog breed which was
the dominant type of that feral group (Boitani et al., 1995). Interspecific territorial
behaviour among wild canids is known to occur (e.g. coyote–wolf, fox–coyote)
and it should be expected in dogs and wolves that share the same ranges. In
central Italy, dogs and wolves share an almost identical niche and compete for the
same food resources (Boitani, 1983), but their territories only partially overlap
(Boitani et al., 1995). In Abruzzo, the territorial core-areas of feral dogs were
closer to human settlements where wolf presence was lowest, and they were
located in the interstice between two neighbouring wolf territories (Boitani et al.,
1995). These observations suggest that wolves in that particular area might have
been an important component in shaping feral dogs’ territory and in determining
its location in relation to human settings.
In contrast to house and village dogs, feral dogs are most definitively active
during nocturnal and crepuscular periods with a clear tendency for a bimodal
activity pattern, similar to other wild canids (Perry and Giles, 1971; Scott and
Causey, 1973; Causey and Cude, 1980; Boitani and Racana, 1984; Daniels and
Bekoff, 1989a; Boitani et al., 1995). Nevertheless, Nesbitt (1975) and Boitani et al.
(1995) found that dogs could also travel during daytime, when human presence
and interference is low, suggesting that feral dogs adopt nocturnal habits essen-
tially to avoid, or minimize, human contacts. However, a great variety of wild
carnivores show a bimodal pattern of activity, and it has been hypothesized that
this innate behavioural trait is independent from environmental pressures and has
not been altered to a great extent by artificial selection on dogs.
Food sources, hunting and predation
Given their social structure and their adaptability to a variety of habitats, it seems
obvious to expect that feral dogs would display a diversified feeding ecology and
include a wide range of food items in their diet. While in wolves the social struc-
ture of packs integrates all members in an efficient hunting unit, in feral dog
groups leadership is more questionable and social bonds among individuals are
more flexible, which might contribute to the inefficiency of feral dogs as preda-
tors compared to their wild canine counterparts (Scott and Causey, 1973; Nesbitt,
1975; Causey and Cude, 1980; Daniels and Bekoff, 1989b; Boitani et al., 1995;
Macdonald and Carr, 1995; Butler et al., 2004). A variety of other factors can
influence their feeding behaviour, such as group size and breed types, relative
abundance and accessibility of wildlife and livestock, the level of human control
and the availability of garbage dumps and alternate food sources easy to exploit.
Group members’ cultural traditions might also exert great influence on the
group’s feeding behaviour, determining hunting attitudes and ability as well as
prey-type preference.
Feral dogs have long been accused of preying on wildlife and livestock and
causing serious damage in a variety of geographic areas (see review in Boitani and
Ciucci, 1995), but the supporting evidence is surprisingly scarce. Several studies
(e.g. Scott and Causey, 1973; Nesbitt, 1975; Boitani et al., 1995) failed to docu-
ment livestock depredation in spite of free-ranging livestock available to feral dogs
in most of the study areas. Instead, as documented by Boitani et al. (1995) and
Nesbitt (1975), livestock depredations were caused by free-ranging neighbour-
hood/village/stray dogs. Further evidence has been obtained to support feral
dogs’ predation on wildlife (i.e. deer, wild boar, hare, rabbit, etc.), although all
studies failed to document any serious impact on wildlife populations (Perry and
Giles, 1971; Scott and Causey, 1973; Gipson and Sealander, 1977; Causey and
Cude, 1980; Federoff et al., 1994; Boitani et al., 1995; Herranz et al., 2000; Rouys
and Theuerkauf, 2003; Butler et al., 2004). In Zimbabwe, Butler and du Toit
(2002) found that dogs were primarily scavengers of human waste and animal car-
casses, roaming into nature reserves and outcompeting other wild scavengers.
Although predation on deer and boar populations might be of little management
significance, feral dogs have been reported to prey on many endangered and rare
species, causing serious conservation concerns: marine iguanas in the Galápagos
Islands (Kruuk and Snell, 1981; Barnett and Rudd, 1983); capybaras in
Venezuela’s llanos (Macdonald, 1981); Indian porcupines in India (Chhangani,
2003); leatherback turtles on the Andaman and Nicobar Islands in the Indian
Ocean (Andrews and Shanker, 2003); mountain gazelles in central Arabia
(Dunham, 2001); and sable antelopes in Africa (Dott, 1986). Finally, in one report
Kamler et al. (2003) observed three feral dogs successfully attacking and killing a
coyote, perhaps during an interspecific territorial dispute.
On the basis of all reviewed studies it seems reasonable to conclude that feral
dogs show a generally low predatory attitude, a low kill rate (efficiency of preda-
tion), have potential limits in prey size and are indirectly dependent on humans
for food. Notwithstanding the inherent weaknesses of their predatory behaviour,
feral dogs can be a serious nuisance to local farmers and a threat to protected
wildlife despite efforts to control their numbers, especially in protected areas
(Johnson, 2002).
Breeding, denning and parental care
Of all traits affected by artificial selection, reproduction has been strongly manip-
ulated to increase reproductive potential and to shorten generation time in dogs
(Boitani and Ciucci, 1995). In the feral group studied by Boitani et al. (1995), all
females reproduced giving the group full potential for demographic increase.
Domestic dogs usually breed twice a year with little or no seasonal patterns, but
feral dogs tend to concentrate their oestrus cycles in the spring in northern tem-
perate environments (Gipson, 1972; Daniels and Bekoff, 1989b), and in the
autumn, or post-monsoon period, in India (Chawla and Reece, 2002; Pal, 2003).
Boitani et al. (1995) found on average 7.3 months (range 6.5–10 months) between
oestrus periods; 50% of births occurred during February–May, whereas the
others were scattered during the rest of the year. The spring peak of births is
common and has obvious adaptive value among wild canids and many other
wildlife species in the northern hemisphere. In feral dogs it might just be the
remnant of an ancestral endogenous reproductive rhythm without any current
adaptive value. More interesting is the lack of synchronization of the breeding
females in the feral group found by Boitani et al. (1995), and the casual distribu-
tion throughout the year of the other ‘non-spring’ oestrus cycles, which might be
resulting from artificial selection and disconnecting reproductive rhythms from
natural photoperiod synchronization and social control (but see Macdonald and
Carr, 1995).
In the same study, there was no indication of communal care of litters by the
group and all females reared their pups alone, although they were often visited by
other group members. Females located their dens near the group’s traditional
core areas and spent most of their time at the den but frequently visited the closest
feeding sources. Even though food was abundant throughout the denning period,
the lack of group members’ support, in terms of vigilance and protection of the
pups, during the mothers’ frequent absences might have contributed to the high
rate of infant mortality due to predation (Boitani et al., 1995). Daniels (1988) and
Daniels and Bekoff (1989b) reported that breeding females split from the group,
denned and reared pups in isolation, though in the vicinity of the group.
Alloparental care seems to be an adaptive behaviour in many social wild canids
because it relieves the female from the burden of caring for her pups alone and it
may increase protection from intruders and predators. Dogs appear to be the only
canids without any form of paternal care (Macdonald and Carr, 1995), although
Malm (1995) found that a substantial proportion of family-owned male dogs par-
ticipate in caring for the pups, mainly by providing regurgitated food.
Domestication and human assistance to reproduction might have played a signif-
icant role in altering parental behaviours as well as eliminating most of the social
control on reproduction within the group.
In feral dogs, litter sizes range from 3.6 pups/litter (n = 11, Boitani et al.,
1995) to 5.5 pups/litter (n = 17, Macdonald and Carr, 1995), but pup survival
rates are very low. Boitani et al. (1995) found that out of 40 pups, 28 (70%) died
within 70 days of birth, nine (22.5%) died within 120 days, one (2.5%) within 1
year, and only two (5%) survived the age of 1 year. Similar results have been
obtained from other studies (Scott and Causey, 1973; Nesbitt, 1975; Daniels and
Bekoff, 1989b; Macdonald and Carr, 1995). Most mortality seems to occur
during the period of early independence and may be due to: (i) the absence of
communal helping and increased risks of predation when pups are left unat-
tended, and when they begin to explore the areas surrounding the den site; (ii) a
lowered maternal interest in offspring as the mother enters a new oestrus cycle; or
(iii) hostile environmental conditions for litters born in periods other than spring
or early summer. In short, the reproductive traits selected in domestic dogs are of
low adaptive value back in the wild, where feral dogs suffer from a very inefficient
reproductive mechanism, which tries to maximize production while minimizing
newborn and juvenile survival (often less than 5% surviving to 1 year of age) to
the point that feral populations appear unable to sustain themselves. In central
Italy, the feral dogs in Boitani et al.’s (1995) study could not have maintained their
numbers without continuously recruiting new group members from neighbour-
hood/village/stray dogs’ populations.
Finally, a curious and so far unexplained aspect of feral dogs’ demography is
the often reported highly skewed sex ratio in favour of males. While in village
dogs there might be an important effect of selective removal of females, in feral
dogs it is difficult to expect a differential mortality rate for the two sexes outside
artificial human interference. Boitani et al. (1995) found the overall litter compo-
sition highly skewed in favour of males (3.2:1) compared to the female-biased
adult sex ratio, but they could not provide any conclusive explanation for the pos-
sible reasons of higher female survival rates.
Feral Dogs and Humans: a Less Obvious Form of

Dependency
In controlled settings (i.e. family dogs), domestic dogs depend entirely on humans
for all their biological, social and ecological requirements (food, health,
range, housing, protection from predators, social relationships, reproduction,
welfare, etc.). However, under some circumstances, this bond may be-
come weaker and dogs can adjust to a looser relationship with humans either
intentionally (by escaping and affiliating with other free-ranging dogs), acciden-
tally (loss of owner, birth from a stray mother) or forcefully (abandonment).
When this occurs (by human intervention), it is up to the dog obviously to
decide if a looser lifestyle is more interesting, but this is not a field where scientists
easily venture (but see Hart, 1995). What we can do instead is try to under-
stand the causes, conditions and implications of dogs living partially or totally
unrestricted.
Of the few generalizations we can draw from the studies discussed above,
two offer relevant behavioural insights and point to the same management im-
plication: first, an unrestricted lifestyle has its costs, and the looser the dog
the higher the costs. Unpredictable food resources, occasional shelter, unsafe
habitat, exposure to diseases and predators all contribute to determine high
mortality rates in neighbourhood, village, stray and feral dogs to the point that,
particularly in Western societies and temperate climates, these dog popu-
lations are often not reproductively self-sustaining. Second, even though these
dogs are increasingly independent from humans for their immediate decisions
(use of space and resources, activity, social relationships), they still depend
heavily, though indirectly, on human activities for most of their key resources.
This dependency may be more evident in neighbourhood and village dogs
and less obvious in more elusive forms of free-ranging dogs. By the simple
fact that stray-feral and true feral dogs tend to live in increasingly wild conditions
avoiding contacts with humans, one could be tempted to equate them to
other wild canids in terms of their ecological relationship with the environ-
ment, implying that they are regulated by natural ecological processes in the
absence of human interference and representing ecological and evolu-
tionary adaptive units. However, with possibly a few exceptions (e.g. wild
dingoes: Price, 1984), which result from ecological and anthropological dy-
namics of thousands of years ago, no longer plausible in today’s world,
no true feral dog population appears to be ecologically and reproductively self-
sustaining or, more importantly, to be reproductively isolated from other domes-
tic dog categories. On the contrary, most feral dog populations represent
ephemeral entities whose existence is linked to human activities to various
degrees.
By reviewing studies on feral dogs carried out in different areas of the world
(see above), several factors appear critical for their persistence. They underline
the dependency of feral dogs upon humans and involve at least four aspects of
their biology: nutrition, ecology, demography and management. First, the avail-
ability of abundant and predictable food sources appears a critical prerequisite for
the establishment of feral dogs’ populations and their persistence through time.
The behavioural ecology of the studied feral dog groups is explained largely by
the nature, dispersal and abundance of human-derived food resources (Scott and
Causey, 1973; Nesbitt, 1975; Causey and Cude, 1980; Daniels, 1988; Daniels and
Bekoff, 1989b; Boitani et al., 1995; Macdonald and Carr, 1995). Feral dogs,
as other wild carnivores, are limited by bio-energetic constraints and, differ-
ently from neighbourhood and village dogs, they need to forage and hunt by
themselves (i.e. no food is intentionally provided by humans). However, feral dogs

do not appear to be effective predators and, in most situations, they thrive only
where abundant and highly predictable food sources (i.e. garbage dumps) are
available. In the area studied by Boitani et al. (1995), 3 years after the closure of
one of the main garbage dumps no feral dogs could be found in their traditional
core areas, although other factors could also have been potentially involved
(illegal control, poisoning) (P. Ciucci, 2004, personal observation). As Macdonald
and Carr (1995) pointed out, the large availability of food at garbage dumps,
compared to scraps, handouts and other less abundant food sources in the vil-
lages, can be a factor involved in observing larger group sizes of feral dogs, sug-
gesting that their social ecology as well is dictated by food sources of
anthropogenic nature.
A second factor critical for the establishment of feral dogs is the availability
of suitable habitats and space. Suitable for feral dogs generally means a natural
environment void of human settlements and rich in retreat and denning areas
but, simultaneously, where food sources are adequately dispersed and accessible
(see above), and where human density and/or past extermination efforts corre-
spond to low density or absence of wild predators (e.g. wolves). Third, since feral
dog populations are often not reproductively self-sustaining and suffer high mor-
tality, their persistence and demography are dependent upon the continuous
availability of individuals to be recruited from other free-ranging dog categories
(see Fig. 9.1), whose maintenance and dynamics is, in turn, linked to human activ-
ities. Finally, assuming all above conditions, feral dogs’ survival is ultimately
allowed where environmental and wildlife management regulations, as well as
human health prescriptions, are either neglected or difficult to implement. Feral
dogs have been found in areas with little tradition or enforcement of wildlife man-
agement (e.g. Daniels, 1988; Boitani et al., 1995) or in remote and inaccessible
sites (e.g. Nesbitt, 1975; Causey and Cude, 1980). To ensure conservation and
functionality of natural communities and ecosystems, wildlife management pro-
tocols prescribe that dogs which escaped human restraint are controlled, similarly
to exotic and alien species. Boitani et al. (1995) outlined guidelines for cost-effec-
tive, humane and non-lethal control of feral dogs, but the degree to which we
want to adhere to, or are able to implement, these or other directions is a pheno-
typic expression of our society upon which feral dogs are again dependent. From
this perspective, humans being part of feral dogs’ ecology, we are responsible for
their persistence as well as for maintaining the ecological conditions that allow for
their survival.
Although they are more elusive and apparently independent from humans,
feral dogs are susceptible to disappear where one or more of the above conditions
are not met, at least with higher probabilities than other free-ranging dog cate-
gories which have closer behavioural and ecological bonds with humans. In con-
clusion, we would generally predict that the survival of feral dogs is more of a
temporary and context-related phenomenon than previously thought, and that
humans are more responsible for the dogs’ lifestyle choices than the dogs them-
selves.
References
Andrews, H. and Shanker, K. (2003) A significant population of leatherback turtles
in the Indian Ocean. Kachhapa Newsletter 6.
Barnett, B.D. and Rudd, R.L. (1983) Feral dogs of the Galapagos Islands: impact and
control. International Journal of Studies on Animal Problems 4, 44–58.
Beck, A.M. (1973) The Ecology of Stray Dogs: A Study of Free-Ranging Urban
Animals. York Press, Baltimore, Maryland.
Beck, A.M. (1975) The ecology of ‘feral’ and free-roving dogs in Baltimore. In: Fox,
M.W. (ed.) The Wild Canids. Van Nostrand Reinhold, New York, pp. 380–390.
Bekoff, M. (1979) Scent-marking by free-ranging domestic dogs. Biological
Behaviour 4, 123–139.
Bekoff, M., Daniels, T.J. and Gittleman, J.L. (1984) Life history patterns and the com-
parative social ecology of carnivores. Annual Review of Ecology and
Systematics 15, 191–232.
Berman, M. and Dunbar, I. (1983) The social behaviour of free-ranging suburban
dogs. Applied Animal Ethology 10, 5–17.
Boitani, L. (1983) Wolf and dog competition in Italy. Acta Zoologica Fennica 174,
259–264.
Boitani, L. and Ciucci, P. (1995) Comparative social ecology of feral dogs and
wolves. Ethology, Ecology and Evolution 7, 49–72.
Boitani, L. and Fabbri, M.L. (1983) Censimento dei cani in ltalia con particolare
riguardo al fenomeno del randagismo. Ricerche di Biologia della Selvaggina
(INFS, Bologna, Italy) 73, 1–51.
Boitani, L. and Racana, A. (1984) Indagine eco-etologica sulla popolazione di cani
domestici e randagi di due comuni della Basilicata. Silva Lucana (Bari, Italy) 3,
1–86.
Boitani, L., Francisci, F., Ciucci, P. and Andreoli, G. (1995) Population biology and
ecology of feral dogs in central Italy. In: Serpell, J. (ed.) The Dog: Its Evolution,
Behaviour and Interactions with People. Cambridge University Press,
Cambridge, pp. 217–244.
Brehm, A. (1893) Tierleben. Liepzig-Wien, 4 vols.
Brisbin, I.L. Jr. (1974) The ecology of animal domestication: its relevance to man’s
environmental crises – past, present and future. Association of Southeastern
Biologists Bulletin 21, 3–8.
Brooks, R. (1990) Survey of the dog population of Zimbabwe and its level of rabies
vaccination. Veterinary Record 127, 592–596.
Butler, J.R.A. and Bingham, J. (2000) Demography and dog–human relationships of
the dog population in Zimbabwean communal lands. Veterinary Record 147,
442–446.
Butler, J.R.A. and du Toit, J.T. (2002) Diet of free-ranging domestic dogs (Canis
familiaris) in rural Zimbabwe: implications for wild scavengers on the periph-
ery of wildlife reserves. Animal Conservation 5, 29–37.
Butler, J.R.A., du Toit, J.T. and Bingham, J. (2004) Free-ranging domestic dogs
(Canis familiaris) as predators and prey in rural Zimbabwe: threats of competi-
tion and disease to large wild carnivores. Biological Conservation 115,
369–378.
Causey, M.K. and Cude, C.A. (1980) Feral dog and white-tailed deer interactions in
Alabama. Journal of Wildlife Management 44, 481–484.
Chawla, S.K. and Reece, J.F. (2002) Timing of oestrus and reproductive behaviour in
Indian street dogs. Veterinary Record 150, 450–451.
Chhangani, A.K. (2003) Dogs (Canis familiaris) hunting in the wild at Jodhpur,
Rajasthan. Journal of the Bombay Natural History Society 100, 617.
Coppinger, R. and Coppinger, L. (2001) Dogs: A Startling New Understanding of
Canine Origin, Behavior and Evolution. University of Chicago Press, Chicago,
Illinois.
Daniels, T.J. (1983a) The social organization of free-ranging urban dogs: I. Non-
estrous social behaviour. Applied Animal Ethology 10, 341–363.
Daniels, T.J. (1983b) The social organization of free-ranging urban dogs: II. Estrous
groups and the mating system. Applied Animal Ethology 10, 365–373.
Daniels, T.J. (1988) Down in the dumps. Natural History 97, 8–12.
Daniels, T.J. and Bekoff, M. (1989a) Spatial and temporal resource use by feral and
abandoned dogs. Ethology 81, 300–312.
Daniels, T.J. and Bekoff, M. (1989b) Population and social biology of free-ranging
dogs, Canis familiaris. Journal of Mammalogy 70, 754–762.
Daniels, T.J. and Bekoff, M. (1989c) Feralization: the making of wild domestic
animals. Behavioural Processes 19, 79–94.
DeBalogh, K.K.I.M., Wandeler, A.I. and Meslin, F.-X. (1993) A dog ecology study in
an urban and semi-rural area of Zambia. Onderstepoort Journal of Veterinary
Research 60, 437–443.
Dott, A. (1986) Sable antelope falls victim to domestic dogs. African Wildlife 40,
138.
Dunham, K.M. (2001) Status of a reintroduced population of mountain gazelles
Gazella gazella in central Arabia: management lessons from an arid land rein-
troduction. Oryx 35, 111–118.
Federoff, N.E., Jakob, W.J. and Bauer, W.C. (1994) Female feral dog and two pups
kill deer fawn at the Patuxent Wildlife Research Center, Laurel, Maryland.
Maryland Naturalist 38, 1–2.
Fox, M.W., Beck, A.M. and Blackman, E. (1975) Behaviour and ecology of a small
group of urban dogs (Canis familiaris). Applied Animal Ethology 1, 119–137.
Gipson, P.S. (1972) The taxonomy, reproductive biology, food habits, and range of
wild Canis (Canidae) in Arkansas. PhD Dissertation, University of Arkansas,
Fayetteville.
Gipson, P.S. (1983) Evaluation and control implications of behaviour of feral dogs
in Interior Alaska. In: Kaukeinen, D.E. (ed.) Vertebrate Pest Control and
Management Materials: 4th Symposium. ASTM Special Technical Publication,
Philadelphia, pp. 78–92.
Gipson, P.S. and Sealander, J.A. (1977) Ecological relationships of white-tailed deer
and dogs in Arkansas. In: Philips, R.L. and Jonkel, C. (eds) Proceedings of 1975
Predator Symposium. Bulletin of Montana Forest Conservation Experimental
Station. University of Montana, Missoula, Montana, pp. 3–16.
Gittleman, J.L. (1989) Carnivore Behaviour, Ecology and Evolution. Chapman and
Hall, London.
Hale, E.B. (1969) Domestication and the evolution of behaviour. In: Hafez, E.S.E.
(ed.) The Behaviour of Domestic Animals. Bailliere Kendall, London, pp.
22–42.
Hare, B., Brown, M., Williamson, C. and Tomasello, M. (2002) The domestication
of social cognition in dogs. Science 298, 1634–1636.
Hart, L.A. (1995) Dogs as human companions: a review of the relationship. In:
Serpell, J. (ed.) The Dog: Its Evolution, Behaviour and Interactions with People.
Cambridge University Press, Cambridge, pp. 161–178.
Herranz, J., Yanes, M. and Suarez, F. (2000) Relationships among the abundance of
small game species, their predators, and habitat structure on Castilla-La Mancha
(Spain). Ecologia 14, 219–233.
Hirata, H., Okuzaki, M. and Obara, H. (1986) Characteristics of urban dogs and
cats. In: Obara, H. (ed.) Integrated Studies in Urban Ecosystems as the Basis of
Urban Planning. I. Special Research Project on Environmental Science (B276-
R15-3). Ministry of Education, Tokyo, pp. 163–175.
Hirata, H., Okuzaki, M. and Obara, H. (1987) Relationships between men and dogs
in urban ecosystem. In: Obara, H. (ed.) Integrated Studies in Urban Ecosystems
as the Basis of Urban Planning. II. Special Research Project on Environmental
Science (B334-R15-3). Ministry of Education, Tokyo, pp. 113–120.
Hsu, Y., Severinghaus, L.L. and Serpell, J.A. (2003) Dog keeping in Taiwan: its con-
tribution to the problem of free-roaming dogs. Journal of Applied Animal
Welfare Science 6, 1–23.
Johnson, M.R. (2002) A new capture pen for Caribbean feral dog packs.
Intermountain Journal of Sciences 8, 255.
Kamler, J.F., Keeler, K., Wiens, G., Richardson, C. and Gipson, P.S. (2003) Feral
dogs, Canis familiaris, kill coyote, Canis latrans. Canadian Field-Naturalist 117,
123–124.
Kleiman, D.G. and Eisenberg, J.F. (1973) Comparisons of canid and felid social
systems from an evolutionary perspective. Animal Behaviour 21, 637–659.
Kruuk, H. and Snell, H. (1981) Prey selection by feral dogs from a population of
marine iguanas (Amblyrhynchus cristatus). Journal of Applied Ecology 18,
197–204.
Macdonald, D.W. (1981) Dwindling resources and the social behaviour of
Capybaras (Hydrochoerus hydrochaeris). Journal of Zoology 194, 371–391.
Macdonald, D.W. (1983) The ecology of carnivore social behaviour. Nature 301,
379–384.
Macdonald, D.W. and Carr, G. (1995) Variation in dog society: between resource
dispersion and social flux. In: Serpell, J.A. (ed.) The Dog: Its Evolution,
Behaviour and Interactions with People. Cambridge University Press,
Cambridge, pp.199–216.
Malm, K. (1995) Regurgitation in relation to weaning in the domestic dog: a ques-
tionnaire study. Applied Animal Behaviour Science 43, 111–122.
Mech, L.D. and Boitani, L. (2003) Social ecology of the wolf. In: Mech, L.D. and
Boitani, L. (eds) Wolves: Behavior, Ecology and Conservation. University of
Chicago Press, Chicago, Illinois, pp. 1–34.
Meek, P.D. (1999) The movement, roaming behaviour and home range of free-
roaming domestic dogs, Canis lupus familiaris, in coastal New South Wales.
Wildlife Research 26, 847–855.
Nesbitt, W.H. (1975) Ecology of a feral dog pack on a wildlife refuge. In: Fox, M.W.
(ed.) The Wild Canids. Van Nostrand Reinhold, New York, pp. 391–395.
Ortolani, A. and Coppinger, R. (2005) Behavioral ecology and human interactions
of village dogs: a survey in Ethiopia and a comparative review. Applied Animal
Behaviour Science (under review).
Packard, J.M. (2003) Wolf behaviour: reproductive, social, and intelligent. In: Mech,
L.D. and Boitani, L. (eds) Wolves: Behavior, Ecology and Conservation.

University of Chicago Press, Chicago, Illinois, pp. 35–65.
Pal, S.K. (2003) Reproductive behaviour of free-ranging rural dogs in West Bengal,
India. Acta Theriologica 48, 271–281.
Pal, S.K., Gosh, B. and Roy, S. (1998a) Dispersal behaviour of free-ranging dogs
(Canis familiaris) in relation to age, sex, season and dispersal distance. Applied
Pal, S.K., Gosh, B. and Roy, S. (1998b) Agonistic behaviour of free-ranging dogs
(Canis familiaris) in relation to season, sex and age. Applied Animal Behaviour
Science 59, 331–348.
Pal, S.K., Gosh, B. and Roy, S. (1999) Inter- and intra-sexual behaviour of free-
ranging dogs (Canis familiaris). Applied Animal Behaviour Science 62,
267–278.
Perry, B.D. (1993) Dog ecology in eastern and southern Africa: implications for
rabies control. Onderstepoort Journal of Veterinary Research 60, 429–436.
Perry, M.C. and Giles, R.H. (1971) Free running dogs. Virginia Wildlife 32, 17–19.
Price, E.O. (1984) Behavioural aspects of animal domestication. The Quarterly
Review of Biology 59, 1–32.
Rouys, S. and Theuerkauf, J. (2003) Factors determining the distribution of intro-
duced mammals in nature reserves of the southern province, New Caledonia.
Wildlife Research 30, 187–191.
Santamaria, A., Passannanti, S. and Di Franza, D. (1990) Censimento dei cani
randagi in un quartiere di Napoli. Acta Medica Veterinaria 36, 201–213.
Scott, J.P. and Fuller, J.L. (1965) Genetics and the Social Behavior of the Dog.
Scott, M.D. and Causey, K. (1973) Ecology of feral dogs in Alabama. Journal of
Wildlife Management 37, 253–265.
Wandeler, A.I., Matter, H.C., Kappeler, A. and Budde, A. (1993) The ecology of dogs
and canine rabies: a selective review. Revue Scientifique et Technique de
Office Internationale des Epizooties 12, 51–71.
World Health Organization (1988) Report of WHO consultation on dog ecology
studies related to rabies control. Ref. WHO/Rab.Res/88.25, Geneva.
World Health Organization (2001) Rabies: Fact Sheet N.99, revised June 2001.
http://www.WHO.int/rabies/rabies_fs99
Evolutionary Aspects on Breeding
10
of Working Dogs
Rolf Beilharz
Introduction
Consider the animals on any undisturbed location on Earth. There are many
species and each species seems to be well matched to its particular niche in the
environment. Ecologically, all animal species have relationships with many others
and with plants, fungi and other microorganisms. Plants derive their energy from
the Sun. Animals have to obtain their energy from other forms of life, either while
it is still alive (e.g. living plants or prey animals) or from decaying carcasses and
even dung. Everyone agrees that life depends on energy and that the energy
required by animals for life must derive from other life in the individual animal’s
environment.
Energy is not the only environmental resource required for animal life. Many
other environmental aspects must also match appropriately (e.g. nesting places,
suitable climate and so on). However, energy for the metabolism necessary for all
aspects of life is by far the most important resource necessary for animal life. Why
is energy, or why are environmental resources generally, so well matched to
animal life? Simply because energy and all environmental resources in the envi-
ronment exert ‘natural selection’ on populations of animals.
Charles Darwin (1859) introduced the concept of natural selection. He
learned from animal breeders that selecting and breeding from desired animals
leads to better livestock. This reassured him that natural selection by the envi-
ronment was having similar effects on all forms of life. He saw that different envi-
ronments select different varieties in organisms. As more young are produced
than are needed to replace deaths in a species, not all young will survive. Those
that utilize the available resources better than others have a greater probability of

(ed. P. Jensen)
Breeding Working Dogs 167
survival. And this happens in every generation. Thus, if there is a mechanism of

inheritance which allows the better characteristics to be passed on to progeny,
those individuals better matched (or better adapted) to the environment will
increase in numbers and eventually displace all others.
Natural selection is going on all the time. As different geographical locations
exert different selection pressures on each local population, the local populations
will adapt each in its own way and this may go so far that interbreeding between
these different populations stops. In this way populations in different geographi-
cal areas can become different species. With the world being very much older
than had been calculated from the Bible, this continual natural selection was
enough to have achieved the variety of species making up the different forms of
life now on Earth. The concept of natural selection nowadays makes sense to most
people.
Darwin argued in terms of whole organisms, or specific features (characters
or traits) of whole organisms, which match the environment. The whole organ-
ism and its parts are called the phenotype. We must never forget that natural
selection acts on phenotypes. While Darwin did not know the mechanism of
inheritance, he assumed that there must be some such mechanism present.
Mendel (read 1855, printed 1866) discovered the mechanism of inheritance.
The importance of his work was recognized only in 1900. He realized that there
were factors, now called genes, which move from parents to progeny in pre-
dictable ways (see Chapter 5 in this volume for an in-depth discussion of genet-
ics). The rediscovery of Mendel’s genes led to an ever increasing understanding
of what happens inside the nuclei of cells. Chromosomes, double stranded strings,
were identified as the carriers of genes. Genes were mapped on to the chromo-
somes using crossover experiments. If a gene occurs in more than one version,
with different effects, the different versions are called alleles. From early in the
20th century, the theory associated with Mendel’s genes was expanded to include
the situation where organisms and their genes occur in populations. This is pop-
ulation genetics. It was followed rapidly by quantitative genetics, which describes
gene action in those characters where many genes with small effects are respon-
sible for the phenotype. In both cases the theory describes the situation in terms
of genes and postulates what must happen to the genes involved and their alleles,
in populations and in traits where there is no single major gene acting, respec-
tively. It is clearly understood that the genes produce, or at least influence, the
phenotype. Quantitative genetics recognizes that the environment also affects the
resulting phenotype, although the environment described is restricted to those
environments that are deviations from the population mean.
In 1953, the chemical structure of the material of the chromosomes (DNA)
was elucidated. Since then, biochemistry at the molecular level has pro-
gressed rapidly. This ‘biotechnology’ has allowed individual genes and whole
genomes to be defined. A genome is the molecular description of the whole
genetic material in the chromosomes of an individual and hence of its species. In
other words, we can now describe the genetic information about individuals or
species precisely in terms of biochemical molecules. We can also manipulate this
168 R. Beilharz
material biochemically, and transfer genetic instructions from one species to

another (even from bacteria to humans). Scientifically this is a huge success and
many people see this development as entirely positive. Biotechnology has,
however, pushed the importance of natural selection aside. The earlier clear
understanding that the environment selected whole organisms and adapted them
to their specific environments is largely ignored in the modern paradigm of
biology.
Just as natural selection adapts all species to their environments, humans
have also selected animals and plants for their own purposes. The initial domes-
tication of plants and animals was for a long time a process in which human activ-
ities simply added their effect to that of natural selection. Merely by building
dwellings, humans provided environmental habitats in which certain species
thrived (e.g. pigeons, sparrows, mice and rats). Food wasted by humans also
increased populations of wolves and other animals and birds in their vicinity.
Wolves later became domesticated as dogs (see Chapters 2 and 3 in this volume).
Dogs and humans combined to hunt game animals, and merely by favouring the
progeny of the best hunting dogs, humans caused breeds of hunting dogs to
evolve. This process of initial domestication was an evolutionary process in which
natural selection on fitness of animals continued but was biased by human activ-
ities, largely by the environmental changes created by humans. It is only in the
last 200 years that science has begun to alter domestic plants and animals sys-
tematically, without, and increasingly in opposition to, the positive action of
natural selection.
In plants, biotechnology has facilitated direct manipulation of the genetic
material (DNA) to the point where genetic material can be moved from one
species to another. This is a clear departure from evolution by natural selection,
where genes normally move only from parents to progeny within a species.
Animals are generally more complex than plants, many having a skeleton, plus
the muscles and nervous control necessary for movement. There is thus a far
greater likelihood in animals that a biotechnological manipulation will interrupt
some important biological pathway, which had earlier been naturally selected to
work very efficiently in harmony with the many other such processes. Hence,
functioning animals resulting from biotechnological manipulation of genes are far
less likely than genetically modified plants.
Breeding Programmes
For many domestic animal species other than dogs, geneticists have developed
highly sophisticated mathematical breeding programmes based on estimated
breeding values for the products for which the animals are being grown (e.g.
meat, wool and milk production). In genetic theory, these estimated breeding
values (EBVs) represent average values of the genes for the desired characters. If
an animal could be mated to many randomly selected individuals of the other sex
in the population, theory expects the EBV value of any trait to equal twice the
difference between the average phenotypic value of the resulting progeny and the
population mean. Obviously the higher the EBV value is, the more desirable the
animal is. To handle the necessary mathematics, simplifying assumptions have
been made. These include that correlations among the various traits being simul-
taneously improved are linear. There is also an expectation that one measure-
ment of a trait will be highly related to subsequent measurements of the same trait
if measured again later in life. Rapid progress in the direction of the traits selected
has usually occurred. But unexpected and unintended side effects, such as poor
reproduction (reduced fitness), have often marred later generations in the breed-
ing programmes.
In contrast, my graduate students and I have developed very effective large-
scale breeding programmes for improving guide dogs for the blind, and later for
detector dogs, and these have had no deleterious or unintended side effects. As
well, over the centuries, hunters, herders and other dog users have selected dogs
and created breeds of dogs appropriate for each kind of behavioural work. This
phenotypic selection for work has also been very successful. Many people con-
tributed to success, each breeding with those dogs which worked particularly well
for them. In the last 150 years, breed clubs and canine associations have flour-
ished. In these, the ‘best dogs’ were chosen mainly on appearance relative to a
‘breed standard’, which essentially describes the outward appearance of the ‘ideal
dog’ for each breed. As breed standards displaced working behaviour as the target
of selection, working abilities have declined rapidly. Brain tissue is a heavy user of
metabolic resources and requires continuing selection to maintain its level of
function.
Compared with quantities of products and growth rates or sizes of animals,
it is much more difficult to specify and measure behaviour required for a task.
Nevertheless, geneticists have started to think about the possibilities of applying
sophisticated breeding programmes which work for production traits to behav-
iour in working dogs. I greatly prefer ‘old-fashioned’ selection on the working
behaviour demonstrated by each individual dog. Our evidence demonstrates
that, when properly done, selection on working phenotypes always achieves good
results.
Resource Allocation Theory
Was Darwin wrong when he saw that natural selection of animals led to selection
responses in populations? And were the founders of population and quantitative
genetics wrong when they thought of organisms whose genes produced pheno-
types, which were then evaluated and selected? In the historical overview above,
selection of animals on their phenotypes seemed to make sense. It still does! So, if
there is a problem with breeding plans based on genetic theory, the problem is
likely to arise from the shift of focus from whole organisms to genes.
Geneticists currently apply their theory as if environments do not limit the
organisms living in them, particularly if they are domestic animals. I have argued
170 R. Beilharz
that this is why highly productive animals show deleterious side effects as their
production increases.
In Beilharz et al. (1993), my graduate students and I described in two equa-
tions what happens to fitness under natural selection. Equation 1 states an
obvious fact. The number of an individual’s progeny that get into the next gen-
eration (its fitness through reproduction = F) is a product of how often it has
young (A), the average litter size (B) and the average survival rate of these progeny
to enter the next generation (C).
F=A×B×C (1)
Equation 2 describes another fact we all take for granted. Animals need food
to live. The equation allocates the total energy obtained from the environment (R)
to the metabolism going on in the various traits as the animal grows, develops and
lives. Resources ‘a’ go into trait A, ‘b’ into B, ‘c’ into C, ‘m’ is the energy lost
because metabolism is not 100% efficient, and ‘x’, ‘y’ and ‘z’ are resources going
into traits X, Y and Z which stand for everything else required to keep the animal
alive and able to reproduce.
R=a+b+c+…+m+…x+y+z (2)
Fitness stops increasing once R reaches its environmental limit and available
resources are used as efficiently as possible. This obvious truth was missing in
quantitative genetic theory before then. This has led to problems when traditional
quantitative genetics was applied to breeding programmes. Until resource alloca-
tion theory becomes more widely accepted, the existing problems in breeding
programmes based on current genetic theory will continue to occur. What are the
consequences of these two equations as animals adapt to their environment?
Beilharz and Nitter (1998) have set out the details of the resource allocation
theory derived from these equations.
The best adapted animals are using the resources of their environment in the
most efficient way to achieve highest fitness. Unless the resources of their envi-
ronment increase, animals are prevented from changing from the optimum
values their phenotypic traits currently have, given the constraint that R has
reached its maximum value. What are these optimum ‘intermediate’ values?
For reproductive component traits A, B and C, the following analogy applies.
Which three numbers summing to the constraint (say 12) have the highest
product?
The answer is 4, 4 and 4 (as 4 × 4 × 4 = 64). Every other set of three numbers
limited to a total of 12 has a lower product (5 × 4 × 3 = 60, 6 × 3 × 3 = 54, and
so on). Note that in such a constrained situation with product at its maximum,
increasing any trait can only reduce the product. And it is the product which is
being maximized by natural selection.
For other traits, say X (e.g. hair cover), there will be selection for some hair
as with no hair the animal dies from cold (fitness = 0). If hair cover becomes
more dense than necessary, the animal may have depressed reproduction because
of heat. But even in the absence of a problem with heat, if x is higher than
necessary, the proportion of R available to maximize fitness (a + b + c) is reduced

and F decreases. So, natural selection on fitness restricts all traits with an effect on
fitness to intermediate ‘optimum’ values.
This is an empirical fact already stated by Crow (1986). Crow considered this
effect of natural selection to be a paradox. He discussed possible genetic explana-
tions for why extreme phenotypes were being culled. Long before Crow, Fisher
(1958 – first published 1929) had postulated another genetic explanation for the
fact that, in adapted populations, fitness usually does not rise. His ‘fundamental
theorem’ is that increase in fitness is directly related to available genetic variation
in the trait fitness. This implies that there are special genes for the trait fitness. But
fitness is the end result of all traits that affect reproduction or survival (that is,
traits necessary for life). There is no viable population in which there is no varia-
tion in every trait important to life.
Crow also commented that selection for intermediate values, in each of many
traits, would allow populations to rapidly track any environmental change. Each
trait would move rapidly to its new ‘optimum’ value. What I have done here is to
substitute an obvious explanation from resource allocation theory for speculative
genetic explanations. Nevertheless, I remain deeply indebted to Crow for his clear
description of what natural selection produces.
One very important consequence follows from the above. In a stable envi-
ronment, once animals are adapted to their environment, natural selection pre-
vents phenotypes changing from optimum values. The fossil record shows long
periods of tranquillity with unchanging fossils until an extinction occurs, after
which survivors change rapidly to fill all niches now available after the catastro-
phe causing the extinction. Hence, for most of evolutionary time natural selection
has prevented evolutionary changes, because each environmental niche was
limited by its resources. Catastrophes following cosmic, geological or climatic
events cause extinctions of many species when habitat is destroyed. When the
Earth settles down again, the survivors, now unconstrained by environment, can
rapidly respond to natural selection, which adapts them to one or many newly
available niches. Until environmental niches become limiting again, genetic
changes are very quick, on an evolutionary time scale. Inevitably, natural selec-
tion will again cull unsuitable ‘extreme’ phenotypes as environmental niches start
exerting constraints. Genetic changes do express themselves when unused
resources are available. But they cannot change phenotypes markedly in adapted
animals limited by their environments. This fact has been sidelined in the
advance of biotechnology and modern animal breeding. It suggests that many
goals of breeding programmes based on traditional quantitative genetics may not
be achievable.
There is one other important consequence. If the new environmental niche
does not require change in any feature of an organism, this feature will not
change from its previously optimized form. When mammals replaced dinosaurs,
basic mammalian features like temperature control remained unchanged in the
mammalian radiation. Furthermore, mankind has the same efficient early embry-
ological development as many other animals including even fish, because the
172 R. Beilharz
liquid environment in which early embryological development takes place has not
changed. Our own research has provided several examples of artificial selection
demonstrating that only the specific trait selected has changed. Wool sheep,
selected for fine fibre on midside samples, have fine fibres in the midside, and far
less elsewhere (Stadler and Gillies, 1994). Mice selected at first parity for litter size
have very large first litters, but hardly any change in later litters (Luxford et al.,
1990).
Taking note of environmental resources, and specifically that available
resources may limit the phenotypes animals can achieve, should be incorporated
into breeding programmes aimed at genetic improvement of animals. On farms,
as in breeding working dogs, we must evaluate whole animals on the basis of their
phenotypes. Looking at genomes, estimating breeding values and searching for
quantitative trait loci (spots on chromosomes with large effects on quantitative
traits) is looking at the wrong level. It is like analysing the single threads in the
cloth to determine whether we are looking at a pair of ladies’ jeans or men’s jeans.
When we work with whole animals, it becomes clear that they are limited by their
environment and the food available. We should then also see clearly that our
farms cannot sustain animals that are ‘two standard deviations better than our
present stock’.
What is the Behaviour Required by Working Dogs
A short discussion of different types of behaviours and how genes control behav-
iour is useful, before discussing how dogs can use their behaviour for our benefit.
Can we visualize how genes and the environment combine to achieve behaviours
and other traits? How do traits interacting with fitness in our equations express
themselves in real animals in the real world? Genes work indirectly. How do
genes make characters? One idea is very important! Genes usually contribute to
a system moving along a track in time, a developmental path.
Very rarely do we recognize the action of single genes on their own. Take as
an example something we can visualize, a gene for horns in cattle or sheep. Many
genes together influence the formation of horns. But genes are not the only
causes. Sex hormones contribute to horns taking a shape in males different from
the shape developed in females. Similarly, if particular nutrients are in short
supply in the food, stunted horns may result. So, development of horns occurs in
an external environment which needs to be adequate, in a physiological environ-
ment in which hormones influence sexual differences and, of course, within a
system in which many genes are cooperatively contributing structures and control
of the developmental sequence. The same considerations apply to phenotypic
traits in dogs such as size and shape of ears.
Every gene can mutate from one allelic form to another, though it does so
rarely. In a population in which horns are reliably formed in every individual, the
alleles present at each of the structural and controlling genes contributing to full
horn formation are called the ‘wild type’ or ‘normal’ alleles. If random mutations
occur at any one of the structural or controlling genes, the development of com-
plete horns can be interfered with by the new ‘mutant’ alleles. A reaction can be
slowed down or diverted from the normal path resulting in abnormal horns. Or
the process may be interrupted completely, so that no horns form at all.
Alleles for polledness, the complete absence of horn production, could be
alleles of any one of several genes that must work together to achieve the final
product. Cattle breeders do not care which gene is responsible, as long as they get
the phenotypic result, no horns. But, you cannot use the wild-type form of that
allele of itself to create horn production in a test tube, unless you also have wild-
type alleles of all the other genes involved in the process. (The news media often
proclaim that a ‘new gene for medical problem M’ has been discovered.
Recognizing such a gene may allow one to repair a particular interruption to the
pathway to health. But it does not guarantee that an interruption at a different
locus affecting the path to healthy development will not appear.) Genes for floppy
ears in dogs interrupt the full process that made the ears of a wolf stand up fully
and be highly mobile so that even faint sound from any direction in which the
ears are pointed could be picked up (also see Chapter 6 in this volume).
Genes cooperate in sequences in time, either by providing material substrates
to the sequence or contributing to the timing of the sequence. A gene by itself is
an unhelpful concept in complex forms of life. We should think of a living organ-
ism as a sequence in time, starting from the fertilized zygote and ending with
chemical decay of the formerly living phenotype, when the biochemical reactions
that maintain life stop at death. We must now try to convert the theoretical
notions of reproductive and other fitness traits into concepts that help us visualize
the life of an organism when this is seen as a developmental sequence of bio-
chemical reactions from fertilization to death.
We can begin by saying that length of life (strictly, number of parities) is an
important character. Natural selection must have brought it into an optimal rela-
tionship with all other reproductive and survival traits.
In life, metabolic decisions about where resources must go will be taken
moment by moment. During development of a bird or mammal embryo, impor-
tant metabolic decisions are made in the mother, either when provisioning the
eggs she will lay, or nurturing the fetus inside the uterus. Once young are born or
hatched, parents usually still help for some time. But more and more the young
must fend for themselves. In other words the external environment starts to chal-
lenge the individual young growing animals and soon they must obtain by them-
selves all necessary resources for further development and survival. As they
approach maturity, there is a period where animals are still growing but repro-
duction is possible and will take place if resources are plentiful. Otherwise the
animal delays reproduction for another season until it is more mature and then
starts to reproduce. Thereafter, the animal essentially spends its time reproduc-
ing, and then recovering from the metabolic drain of putting resources into
young.
What natural selection does is to remove, by death, all individuals in which
at any moment metabolic decisions have left an animal unable to cope with the
174 R. Beilharz
Fig. 10.1. A detector dog at work.
current challenge. It also discriminates against parents less able to compete for
food and thus having fewer young. Similarly, the simple loss to a rival of an
opportunity to mate is another way for natural selection to penalize an organism.
There is thus continual balancing of everything biological that you can think of.
Some examples of unsuccessful evolutionary strategies are:
Fig. 10.2. Detector dog coming off plane after work.

Fig. 10.3. A potential guide dog showing a certain amount of fear towards the
handler.
1. Litter size is too large when there is no mechanism to reduce litter size, and
mother dies.
2. Reproduction starts too early in life with stress of reproduction too great to
maintain the still growing animal through the next difficult season.
3. Metabolism does not use all available resources, causing the animal to have
relatively few young.
4. Young animals walking up to ‘unfamiliar beings’ when they emerge from their
protected nests.
5. Birds making nests on cliff ledges, before young have an ‘instinctive’ avoid-
ance of cliff edges (in humans we call this ‘fear of heights’).
6. Animals having to learn escape responses by experience, rather than using
‘instinctive’ fleeing or freezing reactions.
You can imagine hundreds of other possible deviations from efficient devel-
opmental paths, which will be penalized by natural selection. Although each
metabolic decision is made independently at each moment, natural selection in
every generation rewards genetic changes that anticipate regularly appearing
challenges in the environmental niche along the lifetime developmental path.
This means that adapted organisms already inherit pre-programmed develop-
mental paths in which resource needs for an efficient path are anticipated and
prepared for. Examples are laying down of reserves of fat for anticipated future
176 R. Beilharz
Fig. 10.4. Detector dog coming off a boat.
needs, giving young animals emerging from safe nests an instinctive fear of unfa-
miliar beings and giving nestling birds on cliff ledges a similar ‘fear of heights’.
As the environment selects organisms in development programmes, what
roles do genes play? Every development programme is guided by the information
in the particular set of genes put into the zygote by both parents. Whether genes
prosper depends on how well the developmental path launched and guided by the
genes of each zygote or clone is able to handle the challenges of the local envi-
ronment. Developmental paths of organisms are continually under pressure,
metabolically, to keep the organism alive. In the products of evolution which
nature has provided, there is very little scope for an individual animal, on its
development path, to ‘take it easy’. This is why animal breeders are unlikely to
change animals usefully, without paying a biological cost, when they add a new
feature or exaggerate an existing one. Breeders’ wishes involve costs.
We can summarize the consequences of resource allocation theory as follows.
The model of fitness as stated in our equations is a simplification. A living organ-
ism should be seen as a developmental path from the fertilized zygote to death.
Genes influence this path. Natural selection eliminates all organisms that at any
moment have insufficient metabolic resources to cope with the next challenge
from the environment and discriminates against all those not using available
resources efficiently. This in turn selects for the putting down of reserves for pre-
dictable changes in resource needs, such as seasonal or pregnancy-related extra
efforts. Organisms inherit programmed lifetime developmental paths. These pro-
grammes include mechanisms for adjusting reproductive effort to the environ-
ment. In each environmental niche there are optimal sizes and optimal
developmental paths. Natural selection uses the processes of genetic canalization
and genetic assimilation to achieve optimal lifetime paths. Deviations from
optimal sizes and optimal developmental paths incur metabolic costs. These typ-
ically lead to reductions in reproduction and longevity. ‘Premature’ ageing is a
consequence of above-optimal resource use, often early in the lifetime path.
Behaviour is classified into several categories. Reflexes, instinctive behaviour
and learned behaviour are behaviours that are important to working dogs.
Reflexes are responses to stimuli before the animal becomes aware or conscious
of the stimulus. Instinctive behaviour is that behaviour which occurs when a stim-
ulus is encountered for the first time. Both reflexes and instinctive behaviour are
the direct result of natural selection in the evolution of animals. In the develop-
mental path of an adapted animal, the animal must be able to deal with a stimu-
lus the first time it experiences it. Hence, both reflexes and instinctive behaviour
need to be essentially correct the first time the appropriate stimulus situation is
experienced. Reflexes include responses to gravity or to pressure on the body and
withdrawal responses to hot or painful stimuli. Instinctive behaviour includes
avoidance of cliff edges and unfamiliar beings outside the familiar nest, the sexual
reactions by males to females and vice versa and reactions between mothers and
newborn young. Any animal that reacts falsely is left behind by natural selection.
So reflexes and instinctive behaviour in normal healthy animals are dependable
and appropriately motivated. The level of motivation also results from natural
selection. Clearly, instinctive behaviour and motivational level of working animals
will also respond to artificial selection by humans and they are very important to
successful breeding programmes.
Learned behaviour relies on making associations between stimuli and behav-
ioural responses. This process is known as conditioning. Classical conditioning is
the provision of a neutral stimulus, like a sound, shortly before providing food.
Provision of food automatically stimulates the eating response and all its motiva-
tions. If the sound reliably predicts the arrival of food, the sound becomes the con-
ditioned stimulus which by itself releases the motivations associated with eating,
and the eating itself if food then appears. Repeatable signals preceding the regular
feeding of animals in zoos or livestock in droughts become strong conditioned
stimuli.
Instrumental conditioning (see also Chapter 8 in this volume) is the process
where a stimulus places the animal under some stress. If the reaction of the animal
relieves the stress, the animal associates its reaction with removal of stress and,
provided signals are clear and consistent, behaviour will become shaped such that
stress is minimized. Such instrumental conditioning is very important in training
178 R. Beilharz
dogs and horses into precise behavioural responses. For many working dog situa-
tions, both classical and instrumental conditioning play a part. Breeding pro-
grammes are likely to be less affected by learned behaviour than by instinctive
behaviours. Selective breeding will readily change motivation levels associated
with instinctive behaviour. It is difficult to see how selective breeding will affect
the speed by which dogs make associations.
Successful Breeding Programmes

The most important item in every practical breeding programme for working
dogs is to define very precisely the phenotype that working dogs must have. This
depends on the job that the dog is to do. Persons able to make dogs work well are
essential in defining the phenotype to be aimed at. Understanding dogs, their
behaviours, their abilities and how they interact with people, is much more
important than any genetic considerations. After defining the desired phenotype,
one selects as breeders those animals with the best working phenotypes and
breeds with them. That is the outline of the task. Generally, when a trait has not
been selected in any population, there will be a selection response for 10–20 gen-
erations. The most spectacular gain will occur in early generations. Whether the
programme is successful depends on the details put in place to achieve the breed-
ing, testing, training and recording of traits.
Long ago we did this for Australia’s major guide dog breeding programme
and, more recently, for the Australian Customs’ detector dog programme, which
is now reputed to have the world’s best detector dogs. In both cases we followed
a set of six steps as follows:
1. We need to understand the background and aims of the breeding programme.

The better we understand the background and aims, the more appropriate will
be our breeding programme.
2. In every breeding programme, the above aims must translate into changing
particular phenotypic traits of animals. There are two important considerations
in defining the traits: (i) are the traits important in meeting the aims? Ignore
unimportant traits; (ii) how are the traits to be measured? Unless you can describe
a trait on a dog by a number, you cannot say which animal is to be preferred. The
traits most difficult to assess are often the most important: instinctive behavioural
tendencies, concentration on a job, lack of fear. Here you need experience with
dogs, or the advice of someone else with experience, to rate dogs subjectively on
a 5- or 7-point scale, where the middle value describes the average dog. For
working dogs we want all the important traits to be expressed in a healthy animal,
over a long working life. We breed from a long-lived successful worker when this
can be done.
3. Find the genetic information for each important, measurable trait. You need
for each one: heritability (h2) and degree of heterosis or hybrid vigour; and for
each pair of traits: whether they are genetically correlated. You need to select only
one of two highly positively correlated traits as the other will follow automatically,
at least for some time. Negatively correlated traits must both be selected simulta-
neously.
It may be difficult to find genetic parameters in the literature for traits over a
lifetime. For traits using metabolic resources heavily, early high production may
shorten lifetime. We need to select the whole lifetime pattern of expression of the
trait which our animals are to achieve.
4. Choose breeding strategies. Use selection on heritable traits and outbreed or
crossbreed traits showing hybrid vigour. Traits already selected strongly, either
artificially or naturally, will not respond much to further selection. For traits not
previously selected, expect a good response to selection, at least initially.
5. Synthesize all the separate strategies into one overall breeding programme.
The task is to combine selection for traits of high heritability with crossbreeding
for traits with hybrid vigour.
We want to find those animals that have the best combination of individual
traits to breed with. Independent culling levels, in which we cull dogs poor in any
of the characters we want, is appropriate for heritable traits. Ensure that all heri-
table traits you want are selected appropriately and make those matings which
provide the outbreeding necessary for the traits showing hybrid vigour. Insure
against inbreeding decline becoming a problem in future by keeping the popula-
tion base wide rather than narrow. Cooperate with breeders following similar
goals in similar environments and periodically bring in unrelated stock.
6. Check the overall plan for feasibility and for its effectiveness in meeting breed-
ing goals.
In the two Australian dog breeding programmes, we worked very closely with
guide-dog trainers and customs dog handlers to define the traits required in the
respective programmes. We found no useful heterosis in either programme. In our
guide dog research, Mike Goddard bred hybrids between labradors and three
other breeds of similar size. The hybrids were slightly more wary, a sign of positive
hybrid vigour in natural evolution, but opposed to the complete freedom of fear
and distraction that we were looking for as the most important trait for working
guide dogs. In detector dogs, Kath Champness and John Vandeloo from the
customs service devised a chase–retrieve test for finding an article with an odour
in increasingly difficult places. They rated the dogs subjectively on several aspects,
including their ability and willingness to find, retrieve and hold on to the article.
The success of the guide dog breeding programme is determined largely by
finding healthy dogs (guaranteeing a long working life), which are free of fear and
distraction. This requires selection for instinctive behaviours and appropriate
motivation levels. Even if we cannot yet describe exactly the several traits
involved in being free of fear and distraction, the dogs that demonstrate the
desired behaviour must have had the genes we want in the rearing and training
environment in which they grew up. When we work with phenotypic results we
make progress, just as natural selection does. It does not matter that the psychol-
ogy of the dogs is not yet known in detail.
180 R. Beilharz
For drug detector dogs we used surplus dogs from the guide dog programme,
initially those that were too exuberant. But almost all these dogs were sufficiently
free of fear and distraction to be trained as detector dogs. One major criterion for
their selection was that they should be driven to search for and retrieve an item
with a specific odour, with the dog then being rewarded by the handler with a
game of tug-of-war. This complex trait involves high instinctive motivation for
searching and retrieving, the willingness to keep on doing so without tiring and
also instinctive motivation for playing with the handler. Ability to make associa-
tions with specific odours is also required. Again, those dogs that can do this are
the dogs we want. They have the correct genes for the appropriate instinctive
motivations and enough ability to make correct associations. As this character is
unlikely to have been selected heavily among labradors, one can confidently
predict a substantial selection response. This was indeed the case.
Concluding Remarks
Although neglected in the current theory of genetic improvement and biotech-

nology, the dependence of animals on the resources of the environment is very
important. In this chapter I have discussed why environmental resources are now
not taken seriously, even though Darwin’s discovery required recognizing that the
most efficient users of resources were the successful animals selected in nature.
Essentially it is the change of focus from whole animals to genes, and recently to
the molecular structure of genes, which prevents us seeing that the phenotypes of
whole animals are limited by their current environmental resources. The resource
allocation theory explains how and why environmental resources limit pheno-
types for most of the time. The exceptions are when new resources are becoming
available or when former users of resources have disappeared as a result of
natural or man-made causes.
I also described how genes should be understood as contributing to the life-
time developmental path of individual animals. It is such lifetime phenotypes
which natural selection optimizes for fitness and which breeders should be trying
to change in their quest for more commercial products. Looking at developing
phenotypes also helps us to see instinctive and reflex behaviours being selected by
nature and how breeders can change behaviours to achieve better working dogs.
Successful breeding of working dogs requires an intimate knowledge of the
working environment and the behaviours that each dog must have in order to
become good at its work. Breeding with the dogs that are the best workers is using
artificial selection together with natural selection on health, longevity and repro-
duction in the same positive way as evolution has used natural selection on fitness
in each environmental niche. Attempting to use existing programmes for calcu-
lating EBVs for different ‘working traits’, or finding genes for individual traits,
will merely introduce unnecessary complexity, dependence on mathematical
simplifications and delay into the programme. Two very successful Australian
breeding programmes for dogs suitable as guides for the blind and detector
dogs demonstrate that phenotypic selection for precisely the work needed is
highly successful.
References
Beilharz, R.G. and Nitter, G. (1998) The missing E: the role of environment in evo-
lution and animal breeding. Journal of Animal Breeding and Genetics 115,
439–453.
Beilharz, R.G., Luxford, B.G. and Wilkinson, J.L. (1993) Quantitative genetics and
evolution: is our understanding of genetics sufficient to understand evolution?
Journal of Animal Breeding and Genetics 110, 161–170.
Crow, J.F. (1986) Basic Concepts in Population, Quantitative and Evolutionary
Genetics. Freeman, New York.
Darwin, C.R. (1859) The Origin of Species. John Murray, London (New American
Library, Mentor edition, 1958).
Fisher, R.A. (1958 – first published 1929) The Genetical Theory of Natural Selection,
2nd revised edn. Dover, New York.
Luxford, B.G., Buis, R.C. and Beilharz, R.G. (1990) Lifetime reproductive perform-
ance of lines of mice after long term selection for first parity litter size at birth.
Journal of Animal Breeding and Genetics 107, 188–195.
Mendel, G. (1866) Experiments in plant-hybridisation. In: Sinnott, S.W., Dunn, L.C.
and Dobzhansky, T. (1950) Principles of Genetics, 4th edn. McGraw Hill, New
York, pp. 463–493. (Translated from German by the Royal Horticultural Society
of London.)
Stadler, W. and Gillies, R.I. (1994) A case study for the use of mid-side samples to
predict the mean fibre diameter of classed lines. Journal of Wool Technology
Sheep Breeding 42, 319–326.
Individual Differences in
11
Behaviour – Dog Personality
Kenth Svartberg
Introduction
Dogs are highly capable of adapting to new environments and learning to

perform different behaviour in certain situations (see Chapter 8 in this volume).
However, some aspects of a dog’s behaviour might have limited plasticity. In fact,
if you observe dogs and are focused on behavioural stability, you might find the
dog quite consistent in a range of situations. For example, the dog’s strategy when
meeting unfamiliar persons may be very similar in different contexts and over
long periods of time. A dog that shows signs of fear when exposed to loud and
strange noise may still, several years later, show similar tendencies in such situa-
tions. Other examples are a dog’s typical tendency to get excited or to be aggres-
sive. Such stable dispositions create what could be called the behavioural style of
a dog, which has also been referred to as temperament, individuality, coping style,
behavioural syndromes and, more lately, as animal personality.
Traits as Complexes of Behaviour
Central for the issue of individual differences in behaviour, which I will refer to
as personality in this chapter, is trait. If you observe the behaviour of a dog you
will probably find that some behaviours often come together. For example, the
dog that snarls when meeting other dogs will probably also raise its tail, lower the
head, stare towards the other dog, bare its teeth, and perhaps also lunge towards
the dog. Such a ‘package’ of behavioural reactions may be labelled a behavioural
trait – a hypothetical construct with which it is possible to describe the behaviour

(ed. P. Jensen)
Dog Personality 183
of an individual, as well as differences in behaviour between individuals. In

everyday terms we probably would like to call this trait hostility, or aggressive-
ness. If you have the possibility to observe several dogs in a similar dog meeting
situation it is very likely that you will find that dogs differ in regard to this
trait. Some dogs will display these behaviours early in the meeting situation and
with high intensity, other dogs will show some of the above-mentioned reactions,
but not all and not very intensely, and still others may not show any sign of this
trait.
From Behaviour Traits to Personalities
Is behaviour the same as personality, and are behavioural traits similar to per-
sonality traits? The study of personality in humans is closely related to the assess-
ment of feelings, thoughts and beliefs. Within the study of animal behaviour,
internal processes – such as feelings and thoughts – have been considered unob-
servable or even scientifically irrelevant. Researchers have striven towards expla-
nations of animal behaviour in the simplest possible way, in accordance with
Lloyd Morgan’s Canon, and have avoided the use of unobservable events such as
emotions and intentions as behavioural explanations. As a result of this, it seems
that scientists interested in individual differences in animal behaviour also have
avoided the concept of personality because of fear of anthropomorphism.
However, besides feelings and thoughts, personality in humans also includes an
issue that is possible to study in animals – behaviour. Personality traits can be
described as dispositional factors that regularly and persistently determine behav-
iour in many different types of situations. Thus, an individual’s personality can be
inferred from the individual’s behaviour. This makes the study of animal person-
ality no different from any other studies of animal behaviour. The label of a sug-
gested trait – such as ‘fearfulness’ or ‘aggressiveness’ – may only be a short
description for a disposition of the individual to act in a certain way, and does not
necessarily imply the existence of any feelings or thoughts. However, current
research gives evidence for the existence of different basic emotions in animals, as
well as cognitive processing that may be analogous to human thinking at a less
complex level. It is probable that a dog that waves its tail and steadily focuses on
the owner when he or she is going to throw a tennis ball really is full of expectancy
and is anticipating the coming game, even though both expectancy and anticipa-
tion are private for the dog and not observable. Assumptions of emotions and
internal processing in animals give even more relevance for the study of person-
ality in animals. For example, rats that live in an unpredictable environment seem
to be more pessimistic – having negative expectations about unknown events –
than rats with a more stable life situation (Harding et al., 2004). Development of
new experimental designs has recently yielded tools in understanding cognitive
skills in dogs, such as word learning, numerical competence and use of social cues
to get rewards, which may help us to unravel individual differences in cognition
(see Chapter 12 in this volume). However, there is still a need for highlighting the
184 K. Svartberg
Fig. 11.1. Behavioural reactions that are stable over time and across contexts can
be assumed to be expressions of the dog’s personality.
risk of anthropomorphism in assessing personality traits in species like the dog.

Personality assessments in animals should primarily be based on behavioural
observations, and not on assumptions of thoughts and feelings.
So, behavioural observations may be useful when we want to know some-
thing about the dog’s personality. But are all behavioural reactions expressions of
the individual’s personality – is a behavioural trait the same as a personality trait?
For example, can we be sure that the snarling border collie in Fig. 11.1 has an
aggressive personality? Here we may use the definition of personality traits again:
‘dispositional factors that regularly and persistently determine behaviour in many
different types of situations’. In this definition, two aspects of stability are included
– stability over time and stability across situations. When taking the first of these
criteria of personality into account, it is possible that the aggressive reaction of the
border collie is specific for this specific occasion. A similar situation the next day,
month or year may give a different reaction. The reason for such time-dependent
specificity may be that this reaction is sensitive to experience, which gives that the
dog’s reaction may change every time it is exposed to a certain situation. Thus, a
behaviour that is easily changed by training should not be seen as an expression
of the dog’s personality (even though this might tell us something about the dog’s
general trainability, which can be part of a dog’s personality). Change in behav-
iour from one time to another may also be due to maturation. For example, a
male dog’s non-response to a bitch in heat may give reason to assume that this
dog has a low ‘sex drive’. However, this assumption may be very misleading if the
assessment is made at a young age. A test later on, after sexual maturation, could
give a totally different picture. Similarly, behaviour reactions observed at a very
young age may correspond poorly with reactions observed in a similar situation
later on in life. Therefore, it is important to take maturation into account when
Dog Personality 185
assessing a dog’s personality – behavioural strategies are a part of the personality

first when it is to at least some degree temporarily stable.
The second criterion from the definition of personality was stability across sit-
uations. A single observation of a dog may prove to be highly situation-specific,
and may not be relevant at all in other situations, for example in a test. Even
though the dog’s behaviour is very similar every time it is exposed to a certain
stimulus situation, it may be specific for this particular stimulus situation. For
example, your dog shows avoidance reactions to Bill, who is a friend of yours.
Through several meetings with Bill you know that this reaction is stable over time
and exposures. Can this information be used to assess this dog as avoidant to
humans, or at least men? I believe that your answer is ‘no’, since we have no infor-
mation about the dog’s reaction towards other persons. It is possible that this dog
may show no avoidance reactions to other persons, which means that an assess-
ment based on the first information should have been misleading. Similarly, a dog
that is superior to other dogs in learning one task, but not another (where it is
rather poor) should not be regarded as highly trainable or intelligent. In the def-
inition of personality some degree of generality is inherent, which also tells some-
thing about one fundamental issue in the concept of personality – making
predictions of the individual’s behaviour from one situation to another similar sit-
uation. A reaction that is too specific, even though it is stable over time, is rather
useless in prediction of behaviour and may not be a relevant measurement of the
dog’s personality.
Thus, a glimpse of a dog’s behaviour may say something about its personal-
ity if the reaction is stable in two regards – stable over time and across similar sit-
uations. A behavioural trait can be referred to as a personality trait after positive
tests according to these criteria; a personality trait is a hypothetical construct
useful to describe the individual’s typical behaviour. However, to expect that dogs
or other living organisms are stable in the sense that they always show the same
behavioural response is misleading. On the contrary, a dog that always behaves
in a certain way – say, always waving its tail – should probably best be regarded
as pathological, rather than having a stable personality. Personality in a dog
should always be seen as an interaction between the dog and the environment,
and assessed in a context – ‘if this happens, or in this type of situation, the dog
usually behaves in this way’.
One last issue regarding personality traits is how the trait is best described,
and the distribution of different expressions in a population, say, a breed. A
common assumption is that behavioural traits can be described linearly and are
normally distributed; an individual can be assessed according to a dimension, and
more individuals are assessed as intermediates than extremes. Most personality
traits in both humans and animals seem to have this characteristic, even though
the distribution is not always normally distributed (for example, skewed towards
one end-point of the scale). In contrast to the assumption of linearity and normal
distribution, there are some suggestions of traits that have a different characteris-
tic. Within animal personality, the most well-known example is coping style,
which has been studied in species like rats, mice, great tits and pigs (Koolhaas et
186 K. Svartberg
al., 1999). According to the theory behind this construct, some individuals have a
disposition to cope with stressful events with activity and aggressiveness. These
animals are also more prone to create habits, and, thus, less flexible when cir-
cumstances change in a familiar situation. Besides this coping style there is
another type, which is characterized by a more passive strategy when stressed:
low aggressiveness, prone to be inhibited and have a better capacity to adapt to
changes in the environment. The distribution of this trait has been assumed to be
bimodal – two separate and extreme types with no intermediate individuals.
However, this can be seen as an exception, which also has been questioned
(Jensen et al., 1995). A safer stand-point is to assume that individuals differ in per-
sonality according to dimensions from low to high (or low intense to high, seldom
to often, hard to elicit to easily elicited, etc.) regarding the behavioural reactions
that the trait refers to. This means that personality descriptions often are relative
rather than exact: ‘dog A is typically more fearful than dog B’. This is contrasted
with everyday descriptions of personality where references to types are common.
For example, ‘his dog is aggressive’ or ‘my dog is playful’. Because of what has
been mentioned regarding distribution of personality traits above – types rarely
exist – such labelling is misguiding.
Why Study Personality in Dogs?
There has been an increase in interest in the study of animal personality, as well
as in personality in dogs, during recent decades. It seems that this change is driven
by a parallel increase in interest in other areas. One of them is animal welfare. If
different individuals appraise threats or other stressful events differently, some
individuals will be better able than others to cope with a certain life situation. One
example in dogs is the capacity to cope with temporary separation from family
members, or social isolation. To ensure that dogs may cope with this situation,
which is a very common for pet dogs, it is important to take each dog’s personal-
ity into account. Some dogs may be more easily trained to cope with loneliness,
whereas others may suffer more in such a situation.
Another issue is prediction of behaviour. Knowledge of future ways of acting
in different situations is valuable in selection of potential working dogs, such as
guide dogs, dogs that are used for search tasks (explosives, drugs, etc.), guard dogs,
hunting dogs and herding dogs. Behavioural signs in a dog that predict success or
failure before the dog is trained, or in the early phases of the training period, bring
great advantages. Time and money can be saved, and the welfare for dogs and
trainers may be improved. Behavioural prediction in dogs may also be important
for pet dogs. Early signs in the puppy may help the breeder to match the dog with
an appropriate owner. Furthermore, in several countries there are extensive
rescue dog programmes. Dogs that otherwise would be put down may be trans-
ferred to a new home after a stay in a shelter. An assessment of the dog’s typical
behaviour in certain situations when it is in the shelter may increase the chance
of a good match between the dog and the new home. Behavioural problems may
Dog Personality 187
be avoided, or at least diminished, by taking preventive steps. Furthermore, the

value of early signs of behaviour that may cause problems for the owner, the sur-
roundings and/or the dog itself may be extended to all pet dogs. Knowledge of
the young dog’s typical behaviour may help the owner to take appropriate steps,
such as training programmes, that make life more pleasant for all parties.
A third issue is evolution or, in the case of the dog, domestication. What traits
are favoured during selection, and why? In dog breeding, there is – consciously
and unconsciously – a selection for wanted traits and against unwanted traits. If
we assume that these traits have a genetic base, the type of selection that domi-
nates will decide the typical behaviour of the dogs in future generations. Thus,
methods that are useful in assessing the personality in breeding dogs as well as the
offspring are of great interest for successful directed selection. For example, stan-
dardized behavioural tests have been used as tools in breeding programmes in
breed clubs and in selection of working dogs. In addition, it is relevant to under-
stand how other, more unconscious, selection criteria influence the ongoing
domestication of dogs. Dog personalities differ in their adaptiveness in different
life situations and contexts. A certain dog personality might be highly adaptive in
one setting, whereas the same dog might give another owner problems in every-
day life due to its typical behaviour. One example of this type of conflict is seen
in Fig. 11.2. A shyness–boldness dimension has been detected in dogs in a stan-
dardized behavioural test (Svartberg and Forkman, 2002). This trait is related to
everyday life – playful, explorative and fearless behaviour in both social and non-
social situations – as well as to success in working dog trials (Svartberg, 2002,
2005). Depending on the goals of the owner and the breeder, different dogs will
probably be favoured during selection. The bold dog might be favoured if the
Shy Bold
Low interest, or High positive

Attitude towards avoidant and interest, playful
unfamiliar persons inhibited and boisterous
behaviour
Anxious and Confident and

Typical behaviour in unfamiliar fearful, low fearless,
non-social situations exploration explorative
Probability of success
Low High
in working dog tasks
Fig. 11.2. A certain dog personality might be highly adaptive in one setting,
whereas the same dog might give another owner problems in everyday life due to
its typical behaviour.
188 K. Svartberg
major goal is some kind of working performance, whereas less bold dogs,
although not shy, may be more easy to handle for pet dog owners. This assump-
tion is supported by correlations between the typical use of parents in
different breeds, and breed-typical personality (Svartberg, 2006). Breeds where
breeding dogs have a high number of merits from working dog trials are in
general more playful than breeds with parents less often used as working dogs.
Furthermore, ‘show breeds’ are shyer than breeds where show merits seem to be
less important.
Scientific Methods of Assessing Dog Personality
Where should personality be assessed?
What do scientists do when their aim is to assess personality in dogs? Different

methods are used, and there are some factors that influence the choice of method.
For example, some aspects of the dog’s personality may only be possible to assess
over a long period of observation and in certain situations. If we assume that
dominance, as an example, is a possible personality trait in dogs, a single behav-
ioural test will probably not give us the whole picture of the dog’s tendency to
dominate or show submissiveness towards persons or other dogs. Dominance
relationships are formed within the social group over some periods of time, which
makes a behavioural test inappropriate in this regard. One way to get this infor-
mation is to ask a person who is well acquainted with the dog about the dog’s
typical behaviour with other well-known dogs and with familiar persons. An alter-
native way could be to get access to the dog for a certain period of time, and study
the dog’s social behaviour in a controlled environment. Another issue could be
the degree of generality. If, for example, the aim of a study is to compare the
typical behaviour between several dog breeds, the personality of a large number
of individual dogs has to be assessed. This limits the sampling method. To observe
a sufficient number of dogs per breed in a behavioural test may be practically
impossible, which justifies other methods.
Issues like these have led to researchers using different approaches in the sci-
entific study of dog personality. Generally, three methods of collecting data on
dog personality can be defined: (i) observation of the dog’s behaviour in its normal
environment; (ii) behavioural tests; and (iii) information from persons who are
familiar with the dog, for example questionnaires sent to dog owners. Behavioural
tests have been used in order to assess traits like greeting behaviour in meetings
with unfamiliar persons, fearfulness towards different sudden or novel stimuli,
and aggressiveness towards unfamiliar and threatening persons. A test situation,
however, can be seen as a novel situation in itself, which limits its usefulness. The
dog’s typical behaviour may be masked by its reaction towards the test situation,
which sometimes makes it necessary to use other sampling methods. One such
commonly used method is questionnaires in which the dog owner rates the dog’s
behaviour or level of some predefined personality traits. This methodology may
Dog Personality 189
capture such aspects of personality as the dog’s typical social behaviour within the
family, tendency to cooperate with known persons, and behaviour when left at
home. Thus, behavioural tests may be useful to get information about some
aspects of a dog’s personality, whereas other methods, such as questionnaires,
may be necessary in order to gather knowledge of other aspects of the dog’s
typical behaviour. A well-constructed questionnaire and large sample sizes may,
at least partly, compensate for the bias that the large number of observers (dog
owners) gives.
Measuring personality
Besides the issue of sampling method, another question concerns how to assess the
dog’s personality. Personality in animals may be assessed at two different levels –
behavioural observations according to strict objective criteria and with subjective
assessment. The first of these levels concerns what is the most common method
in ethological studies in general. The behavioural reactions are rated according
to strict objective criteria – for example, number, frequency, duration and/or
latency. This method has been used in theoretical approaches, where behaviour
measurements are used as indicators for a suggested personality type (for
example, the time until an intruder is attacked by a resident may be used as a base
for categorizing the animal as ‘non-aggressive’ or ‘aggressive’) or for the magni-
tude of a trait (for example, the number of threat behaviours observed can be
used to assess the animal’s degree of ‘aggressiveness’). Objective measures of
behaviour have also been used in more empirical and exploratory approaches,
where clusters of correlated behavioural variables that presumably represent per-
sonality traits are searched for by the use of factor analysis or other multivariate
analysis methods.
Animal personality may also be assessed at a more comprehensive
level, where the observer subjectively rates the individual according to prede-
fined traits. Observers, who assess the personality by observing it in several
situations, are here used as data recording instruments. Commonly, the animal
is described according to adjectives, such as ‘curious’, ‘motherly’, ‘playful’
and ‘understanding’, on a linear scale. This method provides a higher level
of description, and may capture the overall pattern of an individual’s behavi-
our that remains elusive when discrete events are measured. Studies on human
personality have shown that descriptions of behaviour at a more general
level may be more predictive than specific measures of behaviour (e.g. Funder
and Colwin, 1991). However, direct assessment of personality is sensitive for
subjective interpretations, and may easily be biased by the observer. There-
fore, the accuracy of this method rests on the use of several independent
observers, together with high criteria of inter-observer agreement for the sug-
gested traits.
Besides these two general methods of assessing animal personality, there
is one additional approach that is commonly used that may be seen as a
190 K. Svartberg
‘middle way’ – behavioural rating scales. In this method, the animal’s behaviour
is rated in a specific situation (in contrast to subjective rating, where the overall
tendencies are assessed) according to a predefined scale with a number of steps.
For example, aggressive behaviour in dogs has been rated according to a five-
point scale (Netto and Planta, 1997): 1 – no aggression observed; 2 – growling
and/or barking; 3 – baring the teeth; 4 – snapping; and 5 – biting and/or attack-
ing with bite intention. Some additional assumptions are made when rating scales
are used, compared to when using strict objective criteria. For example, that
several behavioural reactions (such as growling, barking and snapping) are asso-
ciated with the same behavioural category (aggressiveness). In spite of this, rating
scales seem to be useful in dog personality studies, and, furthermore, are easy to
use and therefore a common method in applied settings.
Personality Traits in Dogs
Perhaps the most intriguing question is what personality traits are to be found in
dogs? This question is not easy to answer, however. In studies where stable
aspects of dog behaviour have been in focus, a range of traits have been proposed,
but only a few of these have been tested for stability over time and across situa-
tions. Thus, we have relatively poor knowledge of the stability of a number of pro-
posed traits, as well as of their relevance in different situations. There are some
traits in the literature of dog personality that are more commonly described than
others, and more often tested for stability. The two most widely suggested are
fearfulness and aggressiveness, which I will describe in more detail here.
Fearfulness
‘Fearfulness’ is probably the most studied trait in animals, and the domestic dog
is no exception. In dogs, there are several behavioural reactions that are com-
monly associated with fearfulness. Examples are avoidance behaviour, flight
behaviour, low body posture with low tail and ears, trembling, salivating and
vocalization, such as yelping and screaming. Behavioural reactions associated
with fearfulness have also been regarded as expressions of other traits, which may
be said to be similar or closely related to fearfulness. The most well-known
example is ‘emotionality’, which is a trait that has been thoroughly investigated,
mostly in rodents. Other ‘neighbouring traits’ to fearfulness are ‘stress-proneness’,
‘nervousness’ and ‘timidity’. They all share some, but rather different, facets of
fearfulness. More or less, however, they are all constructs that are suggested to
describe the individual’s general tendency to react to threatening and potentially
dangerous situations. This makes this trait highly relevant in several regards, not
least from a welfare perspective. Dogs that are generally fearful appraise threats
in a range of situations, and it is assumed that they often experience negative
emotions such as acute fear and anxiety.
Dog Personality 191
However, the concept of a general tendency to react to threats has been ques-
tioned. There are several studies that give evidence for more narrow subtypes of
fearfulness. One example is the difference between fearfulness towards social and
non-social stimuli; fearfulness towards strangers may not necessarily be associated
with fearfulness towards such stimuli as sudden or loud noises, novel objects and
thunderstorms. Indications of this come from studies of potential guide dogs that
were observed in a range of situations (Goddard and Beilharz, 1984, 1986) and
from questionnaire studies (e.g. Hsu and Serpell, 2003). Furthermore, social fear-
fulness may be divided into different subtypes, for example, fear of unfamiliar
dogs and unfamiliar persons (Goodloe and Borchelt, 1998). Also non-social fear
tests have yielded results that indicate the existence of several fearfulness traits
specific to particular stimuli (King et al., 2003). With the findings of different sub-
types of fearfulness follows a questioning of the concept of fearfulness; does one
general tendency in dogs to react to threatening and potentially dangerous situa-
tions really exist? This is a justifiable question – if there are different tendencies to
react fearfully towards different stimuli, why use the concept at all?
Another question regarding general fearfulness is how the dog reacts in
threatening situations. Research on other species suggests the existence of differ-
ent coping strategies; either the individual reacts with an active strategy – fight or
flight – or with a passive strategy – ‘freezing’ or immobility. Such differences in
strategies have been reported in dogs. For example, results from several older
studies suggest the existence of two types of fearfulness (e.g. Thorne, 1940; Royce,
1955). These studies seem to be inspired by the work of Pavlov, who suggested
two types of dogs in this regard: an excitable and an inhibitable type. There is also
some support for breed difference in this regard. Scott and Fuller (1965) reported
that inhibition is easily elicited in cocker spaniels and Shetland sheepdogs,
whereas basenjis are more prone to active avoidance. However, such clear-cut
differences in coping strategies have been questioned. It is likely that such differ-
ences between individuals exist, but it is probably more a question of tendency
than of kind – some individuals might be more prone to inhibition, whereas
others might often react with active avoidance. Another factor that interacts with
this possible personality trait is the type of stimulus-situation. Some threatening
situations might elicit immobility to a higher degree than others, where strategies
such as flight are more common.
There is, however, evidence of more general tendencies to react with fear
that makes the concept of fearfulness relevant in dogs. Goddard and Beilharz
(1984) found one general fearfulness dimension besides several more specific
dimensions. Furthermore, fearfulness in social and non-social situations has been
found to be correlated, just as fearfulness towards unfamiliar dogs and unfamiliar
persons. Results from a study carried out by myself using a Swedish version of the
questionnaire CBARQ (developed by Hsu and Serpell, 2003) showed positive
correlations between four measures of fearfulness: ‘stranger-directed fear’, ‘non-
social fear’, ‘dog-directed fear/aggression’ and ‘pain sensitivity’. The correlations
were moderate, ranging from 0.20 to 0.32, but indicate that there is a general
fearfulness influencing fearful behaviour in different situations.
192 K. Svartberg
Aggressiveness
Another highly relevant trait in dogs is ‘aggressiveness’. This trait has been
suggested for dogs in a number of studies, and might be defined as the dog’s
general tendency to act threateningly (for example, raised hackles, bare
teeth, heightened body posture, raised tail, growling) and aggressively, such as
attacking and biting. Two Dutch studies are of interest in this regard (Netto and
Planta, 1997; van den Berg et al., 2003). In these studies, a similar test battery was
used in order to describe the individual dog’s aggressive tendencies. The dogs
were tested in a range of subtests, where they were exposed to stimuli situa-
tions such as approaching persons, unfamiliar dogs, tug-of-war, handling
by the owner, feeding competition and a life-sized doll. The major aim with these
studies was to investigate whether the dog’s behaviour in the tests reflected the
typical aggressive behaviour according to the owners. No direct analyses were
conducted in order to find out if there exists a general aggressiveness trait,
but the results suggest that there is a general aggressiveness component in dogs –
aggressiveness towards both dogs and persons – that is possible to predict
in a behavioural test. However, two different types of aggressiveness were
found based on the correlations of behavioural reactions (van den Berg et al.,
2003). One type was defined as ‘threatening’ (stiff posture, staring, growling,
and pulling of the lip), and one was labelled ‘attacking’ (barking, baring the
teeth, attacking, and, to some degree, snapping). Furthermore, the owner’s
description of the dog’s typical behaviour indicated that aggressiveness towards
dogs and persons does not necessarily have to be correlated. This suggests that,
besides a possible separation between threat and attack, there are different types
of aggressiveness associated with different targets. This is supported by studies
where questionnaires have been used. Goodloe and Borchelt (1998) found evi-
dence for three such types of aggressive behaviour: towards family members,
towards strangers and towards unfamiliar dogs. Similar types were found by Hsu
and Serpell (2003), who used the CBARQ. Their results also suggested an addi-
tional type of aggressiveness: towards familiar dogs. Clinical studies suggest types
of aggressive behaviour that may be candidates as stable traits. Examples are
object-related aggression (defence of food, toy or other object) and territorial
aggression (for example, aggression towards persons when the dog is in its own
yard).
Thus, the situation is similar to what has been described for fearfulness – the
indication of several types of aggressiveness trait raises the question whether a
general aggressiveness trait exists. As for fearfulness, correlations between differ-
ent measures of aggressive behaviour from the Swedish version of the CBARQ
indicate the existence of general aggressiveness (correlations between ‘stranger-
directed aggression’, ‘owner-directed aggression’, ‘dog-directed aggression/fear’
and ‘familiar dog aggression’ ranged from 0.15 to 0.32). However, these are only
indications. There is a need for studies of aggressive behaviour in a range of situ-
ations, such as the test in Netto and Planta (1997), where the issue of generality
versus specificity can be raised.
Dog Personality 193
Other personality traits in the dog
Some other candidates of relevant personality traits in the dog are worth men-
tioning. An often-proposed stable trait in animals is a general activity level.
Compared to fearfulness and aggressiveness, which might be defined by reactions
to a specific class of stimuli, activity is a more unspecific trait. The assumption
behind this trait is that ‘active behaviour’ in one situation, for example measured
as the frequency of paw liftings, is correlated with activity in several other situa-
tions.
A trait related to activity is reactivity or excitability. Results suggest a difference
between being active in non-stimulating situations and being reactive or excitable
when stimulated. In some studies, reactivity has been measured in threatening sit-
uations, which suggests that it might be a measure of the dog’s fearfulness. But
there is some support for a more general tendency to be excited when stimulated.
For example, Hsu and Serpell (2003) found a relationship between the dog’s
excitability in situations such as when the owner returns home, when playing with
a member of the household and when being taken on a car trip. Analyses of data
from a Swedish version of the same questionnaire that Hsu and Serpell used (the
CBARQ) suggest relationships between this type of excitability and other behav-
iours that indicate a more general reactivity. For example, correlations were
found with attachment level to the owner – which the family-related items suggest
– aggressiveness towards strangers and aggression and fear towards unfamiliar
dogs.
The dog’s tendency to be friendly towards unfamiliar persons has been
described in several studies. This tendency has been proposed to be a personality
trait in the dog, which often is labelled sociability. The dimension seems to range
from an active and ‘friendly’ approach to strangers to an attitude of reserve, or
hostility, to strangers. The negative side of this trait seems to be related to social
fearfulness and aggressiveness, and perhaps it is the positive side – a positive inter-
est and a friendliness towards unfamiliar persons – that motivates a use of a sep-
arate sociability trait. A question that seldom has been addressed is whether the
dog’s sociability towards unfamiliar persons is correlated with the same attitude
towards unfamiliar dogs. Results from a questionnaire study made by myself
(Svartberg, 2005) showed a correlation of 0.33 between friendly behaviour
towards persons and dogs, which suggests a common sociability factor.
A very typical behavioural category in the dog is playing. There are results
that support that there are stable differences between dogs in this regard, which
suggest an existence of a playfulness trait. A trait that describes the dog’s tendency
to run after a thrown rag, grab it and willingness to play tug-of-war with it has
been detected in a behavioural test (Svartberg and Forkman, 2002). This trait is
consistent over repeated tests, and correlates with owner reports of their dog’s
general interest in playing with objects with familiar and unfamiliar persons
(Svartberg, 2005; Svartberg et al., 2005). From questionnaire studies, other play-
related traits have been suggested. Goodloe and Borchelt (1998) found two play-
fulness traits that were both object-related and person-directed. One was related
194 K. Svartberg
to chasing after thrown objects and carrying objects, and one trait could be
described as ‘vigorous’ play, related to growling and shaking while playing with
objects, and tug-of-war. Results from Svartberg (2005) suggested one trait that
described object-related play with humans and one dog-directed trait with rela-
tions to both interest in playing with other dogs and approach of dogs in general.
This indicates that playfulness towards persons and dogs might be influenced by
separate traits.
Finally, many attempts have been made to measure trainability – the dog’s
general success in training situations. Different studies have used different proce-
dures to measure this trait: from using results from one specific task, such as
retrieving, to an assessment of the dog’s general performance, including such
aspects as reactions to distractions, persistence and cooperativeness towards the
trainer. However, the search for the dog’s general trainability is so far rather
unsuccessful. The results from the major study by Scott and Fuller (1965) are
interesting in this regard. They trained dogs of five breeds in several different sit-
uations in search for a general ‘intelligence’. Because of the aim of their project –
understanding the genetic bases of behaviour – they used average breed values
when presenting the data. The results were striking. No breed was generally
better than the others when taking all tasks into account. For example, basenjis
were the best in a manipulation test, but the poorest in a trailing test. The beagle
was the best breed in a T-maze test, but was among the poorest breeds in the
manipulation test. The only consistent pattern was found for the Shetland sheep-
dog, which was generally ranked as the least successful breed in the tests. The
authors explain this with this breed’s (at least in the sample used) relatively large
fearfulness. Thus, fear of apparatus and persons inhibited learning performance.
Probably, this is characteristic for performance in learning situations. A general
trainability factor is difficult to find; performance is influenced by a range of task-
specific factors as well as several personality traits. Results suggest that, besides
low fearfulness, playfulness and excitability may predict training success.
Playfulness probably sets the value of play as a reinforcer, which gives that playful
dogs might be easier to reward than less playful dogs.
Excitability, which has been found to predict performance in guard dogs,
seems to relate to success in a non-linear way (Martínek et al., 1975). The dogs
with the highest level of performance were those that had an intermediate
excitability level, whereas dogs that had high or low excitability performed less
well. According to the authors, this effect might, at least partly, be due to a higher
habituation rate for the moderately excitable dogs. Such factors as fearfulness,
playfulness and excitability are probably not equal to a general trainability factor,
and might have no correlation with the individual’s typical – if there is a typical
– learning ability (such as habit formation and memory retention), but might nev-
ertheless predict success in training situations.
Besides these described traits, there are a number of other candidates,
although more poorly investigated in the dog. From questionnaire studies, traits
such as separation distress, tendency to dominate or act submissively, predatory
interest, attachment to owner, tendency to bark and pain sensitivity have been
Dog Personality 195
suggested. Further studies might give us more knowledge regarding the stability
of these suggested traits in dogs.
Motivational State or Behavioural Reaction?
It should be noticed that the theoretical basis of the suggested traits differ, and this
influences the results. The theoretical issue is whether personality has its base in moti-
vational states or in behavioural strategies. For example, if the researcher assumes
that there is a fear state in dogs, the experimental design and the way of collecting
behavioural data will be influenced by this assumption. Different potential fear-elic-
iting stimuli will probably be used, and measures of flight distance or latency to
contact, or a fearfulness scale, might be used as fearfulness indices. Correlations
between the measures will say something about the generality of fearfulness in the
sample used. However, an alternative assumption is that personality derives from the
individual’s typical behavioural strategy, which might be independent of motiva-
tional state. If this is a correct assumption, there is a risk that such typical behavioural
strategies are missed in a ‘motivational state approach’. Assume, for example, that
dogs differ in general excitability levels. A fearfulness scale, and even measures of
contact latency and flight distance, could miss these differences, as well as the stabil-
ity of each individual’s excitability tendency across situations. On the other hand, if
the researcher is focused on typical behavioural tendencies, differences in the indi-
vidual’s tendency to activate different motivational states (for example, how often the
dog seems to be angry or fearful) may be missed. One such conflict is between coping
theory and the concept of fearfulness. Coping theory deals with the type of behav-
ioural strategy the individual uses when confronted with a threatening situation,
whereas the fearfulness concept focuses on the level of assumed fear the individual
experiences when threatened. In the list of personality traits in the dog presented in
this chapter, traits such fearfulness, aggressiveness, sociability and playfulness are
based on motivational states, whereas activity, excitability and trainability are closer
to a behavioural strategy perspective. The somewhat contradicting results found for
fearfulness and aggressiveness can, at least partly, be explained by the two different
perspectives. The differences in focus have been acknowledged rather poorly within
animal personality. It is likely that personality might be successfully studied from
both directions; however, without taking differences in theoretical standpoints into
account, there is a risk of confusion and misguiding results.
Personality at Different Levels
As the mini-review of personality traits above suggests, there seem to be traits at

different levels of generality. Personality in both humans and animals seems to be
best described hierarchically. At the top of the hierarchy, there are a few general
traits that might influence the behaviour in a range of situations. These traits
might, in turn, have different facets, or subtypes, that more specifically may
196 K. Svartberg
describe the dog’s typical behaviour. In dogs, there are some suggestions of such
general traits. For example, there are results that support two major dimensions in
dogs that reflect positive and negative emotions (Sheppard and Mills, 2003).
According to this view, positive activation correlates with such behaviour as play-
fulness, excitability and exploration, whereas negative emotions relate to fearful-
ness, phobic tendencies and anxiety. These two aspects of personality are
analogous to two of the human ‘supertraits’: extraversion and neuroticism. There
are other studies that support that these dimensions might also be found in dogs.
Gosling et al. (2003) found evidence for these supertraits in dogs and, in addition,
two others: agreeableness (lack of aggression, cooperativeness) and openness to
experience (trainability, exploration). At an even more general level, a
boldness–shyness axis has been proposed. This dimension, which ranges from shy
and timid behaviour to an outgoing, active and bold attitude, has been found in
several other species, as well as in humans, and might be correlated to both neu-
roticism and extraversion. As mentioned previously, a boldness dimension that
correlates with sociability towards humans, playfulness and fearless behaviour in
non-social situations has been suggested for dogs, which may influence behaviour
in many aspects of life (Svartberg and Forkman, 2002; Svartberg, 2005).
These suggested ‘supertraits’ in dogs are interesting in the understanding of
dog behaviour. However, when predicting a dog’s more specific behaviour, these
general traits may be rather powerless. When selecting a potential working dog,
or when predicting behavioural problems, the focus should probably be on more
narrow traits. Thus, there is a need for knowledge of dog personality on several
levels – the aim determines which level is the most relevant.
Another point with the hierarchical view is that specific traits may be more
or less correlated. For example, the theory behind extraversion in humans pre-
dicts correlations between activity, sociability and dominance. Such networks of
correlations, and lack of correlations, might help us to understand how dogs gen-
erally act in different situations. An example of a network of correlations is shown
in Fig. 11.3. The figure is based on data from a questionnaire study of 697
Swedish dogs from 16 breeds (Svartberg, 2005). Even though this sample may be
too specific to use for a generalization for all dogs, the figure shows probable asso-
ciations between specific traits. For example, (lack of) sociability, aggression and
fearfulness towards humans are tightly connected. Aggression and fearfulness
directed towards dogs are correlated to a range of other traits, such as non-social
fearfulness, fear of and aggression towards unfamiliar persons, general excitabil-
ity and aggressiveness directed to other dogs in the household. In contrast, owner-
directed aggressiveness and tendency to chase (cats, squirrels, etc.) are not related
to any other traits in this study.
Development of Personality
One issue regarding dog personality is related to development – change and con-
tinuity of personality during the life course. This question is often addressed in
Dog Personality 197
Fig. 11.3. The correlation between different behavioural traits from the study of
Svartberg (2005). The thickness of the line indicates the degree of correlation;
from the thinnest line that represents a correlation of 0.20 to 0.25, to the thickest
that represents correlations above 0.40. Dotted lines indicate a negative
relationship (SRB = separation related behaviour; PS = pain sensitivity; NSF = non-
social fear; ODA = owner directed aggression; SSOC=sociality towards strangers;
DDAF = aggression and fear directed to unfamiliar dogs; FDA = aggression
towards other dogs in the family; EX = excitability; AAS = attachment and atten-
tion-seeking behaviour; SDA = stranger-directed aggression; SDF = stranger-
directed fear; DSOC = sociability towards dogs; HPLAY = playfulness towards
humans; TRAIN = trainability; CHASE = interest in chasing).
human personality research. Within the study of animal personality, development

of stable dispositions is less often studied. As a starting point for a discussion of
what we know about personality development in the dog, three questions regard-
ing personality development in humans formulated by Caspi and Roberts (2001)
may be useful: (i) how early in the life course can we identify characteristics
unique to individuals that will show continuity and change in personality? (ii)
when in the life course is personality fully developed? (iii) what life course factors
moderate continuity and change in personality? We are limited to the few longi-
tudinal studies that have been carried out, something that we should have in mind
when interpreting the results.
How early can we tell?
Among the relatively few studies where stability of behaviour from young age in
dogs has been in focus, fearfulness is the most studied trait. There are some
198 K. Svartberg
suggestions on very early indications of fearfulness. Royce (1955) used data from
a battery of tests carried out within the research programme at the Jackson labo-
ratory (Scott and Fuller, 1965). He found that the average number of vocaliza-
tions – whines or yelps – from birth to the third week (observed during weighting)
correlated with reactivity score in several situations at the 18th week, as well as
with some physiological measures (blood pressure and sinus arrythmia) at 8–9
months of age. Another early sign was a fearfulness score (rated during a handling
test) at the fifth week, which correlated with reactivity measures at 1 year of age.
A study by Goddard and Beilharz (1986), which was carried out in order to
predict fearfulness, activity and trainability in potential guide dogs, suggested that
fearfulness was stable from 8 weeks of age. However, the correlation with adult
general fearfulness increased with test age: better predictions were made at higher
ages.
These results suggest that aspects of the dog’s fearfulness may be possible to
detect at a very early age. This is supported by a study on wolf pups, where several
different tests at 7–9 weeks of age were carried out (Fox, 1972). The results sug-
gested that one general boldness dimension could explain individual differences
in a range of situations, including prey-killing, fearless behaviour in different sit-
uations and a tendency to dominate other pups. Puppy-boldness, in turn, pre-
dicted dominance score at 1 year of age, which suggests that the individual’s
general boldness, or the tendency to act in a fearless way, is developed at an early
age.
The early fearfulness signs, which are promising from a prediction point-of-
view, are contrasted by results of other traits. Several studies in this area suggest
that behaviour before 8–10 weeks has a low predictive value for the adult’s typical
behaviour. For example, a study on the predictability of dominance and activity
level using a popular test developed by Campbell (1972) suggested no correlation
between 7 and 16 weeks of age (Beaudet et al., 1994). Another example is the
study of Wilsson and Sundgren (1998), who used data from tests carried out by
630 German shepherds at the age of 8 weeks and at 15–20 months. There were
no correlations found between behavioural measures from the early test (vocal-
ization, reaction to a person, reaction to play objects, activity level) and measures
from the adult test (however, this result could be explained, at least in part, by dif-
ferences in methodology between the two tests).
Even though there are relatively few studies in this area, it seems that the pre-
dictive power of puppy testing is low. The exception may be fearfulness, which
may be able to be predicted at an early age. At what stage it is possible to predict
other traits is difficult to assess based on the knowledge so far.
When is personality fully developed?
A common but probably misleading view is that personality develops from birth
to a certain age, and then remains stable. In human personality studies, it seems
that stability of personality is rather low in childhood, increases in adulthood and
Dog Personality 199
reaches a plateau between the ages of 50 and 70 (Roberts and DelVecchio, 2000).
However, even though a plateau is reached, personality also continues to change
at higher ages. Unfortunately, there are no studies that give us knowledge regard-
ing this issue in dogs. Some studies show high stability of personality traits in dogs
over short periods of time (1–2 months) in adult dogs. For example, high rank-
order correlations (0.6–0.9) have been shown for sociability towards humans,
non-social fear, playfulness towards humans and aggressiveness in adult dogs
(Svartberg et al., 2005). However, short-term stability is a basic criterion for per-
sonality traits, and does not necessarily tell anything regarding stability over the
life span. There are also indications on stability over longer periods of time for
these traits (Svartberg, 2005), but how they change in magnitude over the life
span, if they change in a predictable way, is not known.
Results from the behavioural test DMA may give us some information
regarding the traits that are assessed in the test. Dogs are tested only once, but the
differences in scores between dogs of different age might give an indication on
how personality changes over time. In order to investigate this question, I used
data from one large breed, the German shepherd, where it was possible to group
dogs according to test age. Very few dogs are tested at an age above 5 years, but
four age categories were possible to compare (1–2 years, n = 6, 113; 2–3 years, n
= 675; 3–4 years, n = 192; and 4–5 years, n = 80). Of the personality traits that
are expressed in this test, three were of greater interest: curiosity/fearlessness
(expressed as non-social fearlessness in everyday life); sociability; and
aggressiveness. As Fig. 11.4 shows, there are differences between age categories
for these traits. The same trends were found for both sexes, which indicates that
non-social fearfulness, sociability and aggressiveness decrease slightly over the
years.
These results, however, are only a rough indication. To really understand
how personality changes over the life-span, studies where dogs are followed lon-
gitudinally are needed. Hopefully, such approaches will attract researchers in the
future.
What factors moderate continuity and change?
The common way of thinking is probably that environment causes change in

behaviour during development, whereas the genetic bases of behaviour make
behaviour stable. Regarding environmental factors that influence behavioural
change, it is likely that a range of factors may contribute to the variation of dog
personality. The transition from the kennel to the new home may be an impor-
tant factor, as well as the family setting in the new environment. The presence of
other dogs, number of family members and their age, the non-social environment
in which the puppy grows up – countryside or city – as well as how the social envi-
ronment treats the young dog are probably important moderators of the dog’s
personality. There are, however, few studies that have focused on the importance
of different environmental factors on personality in the growing dog, probably
200 K. Svartberg
4.5
3.5
3 Fearlessness
Sociability
2.5 Aggressiveness
1.5
1
1–2 2–3 3–4 4–5
Years
Fig. 11.4. The figure presents the scores for the traits curiosity/fearlessness,
sociability and aggressiveness assessed in the DMA test for four categories of dogs
based on test age (German shepherds, males and females pooled). Statistically
significant decreases in scores with age were found for sociability and
aggressiveness, whereas a significant increase in curiosity/fearlessness was
detected.
due to the methodological problems of separating different factors from each

other. There are some indices from studies where the focus has been on factors
associated with aggressive behaviour in dogs. For example, Podberscek and
Serpell (1997) found several differences in environmental factors between a high-
aggressive group of English cocker spaniels and a low-aggressive group. Owners
of low-aggressive dogs were older and more attached to their dogs, whereas high-
aggressive dogs were groomed less often and given less exercise. Another study
showed that aggressive dogs, compared to dogs that are attacked (‘victims’), are
more commonly trained by shaking and hitting, whereas a dog owner attitude of
believing that dog training should be fun or that training is irrelevant was associ-
ated with victim dogs (Roll and Unshelm, 1997). There are some other studies
that indicate the importance of owner attitude on dog behaviour. For example,
anthropomorphic involvement by the owner has been found to be associated with
dominance aggression in the dog, whereas owner anxiety seems to be associated
with over-excitement and aggressiveness in the dog (O’Farrell, 1997; Podberscek
and Serpell, 1997). Even though these results tell more about how different life
situations influence the eliciting of different behaviour, they might also indicate
that these factors can be important in the development of a dog’s personality.
The view that stability of behaviour during the development is caused by a
strong genetic basis probably has its roots in work by early ethologists. Behaviours
that were species-specific and difficult to alter with training attracted interest,
which contributed knowledge about what has been called fixed action patterns
Dog Personality 201
and instincts in a range of species. However, more current research has shown
that it is hard to find behaviour in a species like the dog that is totally ‘genetic’:
dog behaviour, in general, is possible to change by environmental factors.
However, genetic factors may limit the possibilities, and set boundaries for the
development of a behaviour. In some cases, there seem to be very narrow bound-
aries, which makes the influence by environment rather minor. The degree of
heritability of a behaviour (see Chapter 4 in this volume) may give us a hint about
how strongly the behaviour is influenced by genetic factors. In a few cases, behav-
ioural traits seem to be strongly influenced by genetic factors. One example
comes from an unpublished study by Arvelius (2005), who studied the heritability
of herding traits in the border collie using data from a herding test carried out
during early training. These results suggested heritability estimates of between
0.40 and 0.55 for several herding traits (for example, the tendency of keeping dis-
tance to sheep and anticipation of sheep movements), which suggests that these
rather specific traits to a large extent are influenced by genetic factors. Regarding
traits that may be considered as broader personality traits in the dog, such high
estimates of heritability are rather uncommon. Heritability estimates of 0.2–0.3
are often acknowledged as relevant and rather high. A study of the heritability of
the traits found in the DMA is an example of this (Strandberg et al., 2005). The
results suggested heritability estimates of 0.23 for curiosity/fearlessness, 0.22 for
playfulness and 0.15 for aggressiveness. Such results indicate that there are
genetic bases behind personality traits in the dog, even though these should not
be over-emphasized. Thus, genetic factors may buffer the impact of environmen-
tal influences, and create stability during the development.
This view, however, is not telling the whole story. Genes may create change,
and environment may cause stability. One obvious example of gene-created
changes in behaviour is sexual maturity. During the transition from juvenile to
adult, the genetic expression changes. Some genes that were active during the
juvenile period deactivate, and other genes start to produce proteins that influ-
ence changes in the dog’s typical behaviour. The findings that gene activation is
fluctuating, over both the short and long terms, give a contrasting picture com-
pared to the traditional ‘stable gene’ perspective (changes in gene expression is
described in Chapter 4 in this volume). On the other hand, stability in personal-
ity may be caused by environmental factors: environmental stability causes sta-
bility in behaviour. A dog that changes homes several times during its lifetime
may appear inconsistent in its typical behaviour, whereas a dog that lives in a
non-changing environment may be assessed as more stable. This perspective is
important when applying results from developmental studies. A common goal in
scientific studies is to control or wipe out factors that may create variations that
are not relevant to the aim of the study. For example, Scott and Fuller (1965) tried
to give all dogs a similar environment during their first year of life due to the inter-
est in the genetic basis of social behaviour. This meant, for example, that the
puppies never made the typical transition from the kennel to the new owners at
about 8–9 weeks of age. In turn, this effect upon behavioural development was
never studied, which may make it difficult to apply their results to pet dogs. For
202 K. Svartberg
example, the breed differences found might be less pronounced when studied in
a non-controlled environment.
There is an ongoing interaction between genes and environment, and it is
seldom possible to single out one cause for behavioural development. The most
striking example of such an interaction is the impact of environment during the
socialization period. This period in the dog’s life, which is assumed to range from
3 weeks to about 3 months of age, is marked by the large influence environment
has on future behaviour. A number of experiments have been done regarding
this, which show that several stable characteristics in the adult may be set during
this period. Examples are fearfulness towards dogs and humans, reactions to sep-
aration, general activity level and trainability (reviewed in Lindsay, 2000; see also
Chapter 6 in this volume). If the dog is able to interact with several dogs and
humans during this period, the probability of developing high fearfulness in social
situations diminishes. If this is delayed to after 3 months of age, the effect of the
social interaction will be smaller, especially if the dog has been separated from
dogs and humans before this age. Thus, the socialization period is a good
example of how genes and environment interact: the genes open up a ‘socializa-
tion window’ during a specified period, but the effect of this is dependent on envi-
ronmental factors during this period.
The currently emerging research within genetic techniques, which is devel-
oping very fast, may soon give us more knowledge of the genetic basis of behav-
iour – which genes influence which behaviour, and the function of the
mechanisms that are in charge during development. With this as a base, we may
single out environmental factors – what and when – that are important during the
development of personality in dogs. Results from other species are promising. For
example, newborn rats that do not get licked and groomed by their mothers grow
up timid and highly sensitive to stress (Weaver et al., 2004). The mechanisms
behind this effect seem to be an increase in methylation, which suppresses genetic
activation. The decrease of genetic activation in a certain gene, a promoter for
the glucorticoid receptor gene, causes fewer receptors to be produced, and, in
turn, more timid and anxious rats.
Applied Use of the Personality Concept
This chapter has dealt with dog personality mostly from a scientific point of view.
However, the concept of personality is also relevant, and has a widespread use,
among dog owners, dog trainers and breeders. Besides the traits that have been
presented here, there are numerous others. For example, in the breed standards
there are large numbers of adjectives describing each breed’s typical behaviour;
most of these adjectives are similar to personality descriptions used for humans.
There are also a number of popular behavioural tests that are used in the prospect
of revealing the dog’s typical way of acting, from a variety of puppy tests to dif-
ferent tests used to capture the typical behaviour of adult dogs. Evidently,
personality in dogs – perhaps more often referred to as mentality, temperament
Dog Personality 203
or nature in everyday terms – is relevant for all persons involved in dogs in one
way or another. It is important to describe the characteristics in dogs; for the
breeder, who cares for a selection of suitable breeding animals and strives for a
good breeding result; for the dog trainer, who is interested in selecting potential
individuals or matching the dog with adequate training; and for the dog owner,
who might want to understand the essence of his or her dog.
While the benefits of using labels and behavioural tests to describe dog per-
sonality among dog people are clear, it could be important to highlight some
drawbacks. Regarding descriptions of personality in dogs, it seems to be difficult
to find a consensus even among scientists, which of course makes it more difficult
for laymen to develop a common language regarding behavioural differences in
dogs. A characteristic named ‘stress-proneness’ by one person could be referred
to as ‘liveliness’, ‘active temperament’, ‘high drive’ or ‘happiness’ by others.
Alternatively, one label may have different meanings to different persons. A dog
with a ‘calm’ temperament according to a breeder of a pet dog breed may very
well differ significantly in its typical activity level compared to another dog with
the same description made by a breeder of working dogs. Even such a specific
trait label as ‘interest in objects’ may differ between persons. One might mean the
dog’s interest in tug-of-war, another the dog’s willingness to carry objects in the
mouth, and a third the dog’s enthusiasm to search for motivationally significant
objects (such as toys or dummies). The differences in terminology might be a
unsolvable problem, at least as long as the scientists in this field have difficulty
reaching consensus. Within the field of science this is partly solved by a routine of
defining concepts and terms that otherwise could create misunderstandings (for
example, ‘sociability’ might be defined as ‘the dog’s interest to interact with other
individuals in a positive manner’). My belief is that this strategy would be benefi-
cial for the non-scientific part of society that deals with dogs. The mere knowl-
edge of differences in definition is important to bear in mind in order to avoid
confusion; defining terms and labels would be even better. In other words: don’t
take descriptions of behaviour and personality for granted, try to understand what
is meant by them.
When it comes to behavioural tests it is important to bear in mind that the
usefulness of results from a test is often very limited. The test situation itself limits
the usefulness, something that has been described previously in this chapter.
Some traits are possible to detect in a test, whereas others are not. Besides this,
behaviour tests arranged by laymen often have several weaknesses.
First, the standardization of the test situations is often poor, so dogs may
experience different test stimuli. For example, one dog may be tested in a situa-
tion where a person throw a tennis ball away from the dog, a second will experi-
ence a ball thrown to the side and a third dog may be put in a situation where the
ball is coming towards it. When a test situation has this sort of variation – which
could be both larger and smaller compared to this example – it could be difficult
to separate the effect of test variation from the effect of variation in personality
between dogs.
Second, the way of describing the dog’s behaviour in a test is often insuffi-
204 K. Svartberg
cient. It might be too general to be adequate. For example, a three-step scale with
the alternative of ‘shows no aggression’, ‘shows moderate aggression’ and ‘very
aggressive’ may miss important differences between individuals (especially if the
third alternative is only rarely used). Another issue, which might be more prob-
lematic, is that the description of the dog’s behaviour or personality can be a mix
between intensity and value, or just an evaluation that ranges from ‘good behav-
iour’ to ‘bad behaviour’. When such a description is made, the person who
describes the dog assesses the behaviour against some pre-defined standard. The
risk of this is that this standard becomes a ‘golden standard’ for dogs in general,
even though it was meant to be used for a specific object. One example could be
that a dog that has carried out a test for working dog aptitude may be regarded
as ‘unsuitable’. This might give valuable information for the owner regarding
working dog use, but it may say nothing about the dog’s suitability to be a pet dog,
or usefulness in some dog sport. Thus, it is important to look at the standards from
which statements are made regarding the value of the dog’s personality.
Third, even though it might seem that the test is useful for making predictions
of a dog’s potential in a certain area – such as working dog, hunting dog, herding
dog or pet dog – it is very likely that test results have not been validated against the
use for which they are meant to predict. Thus, there is a risk that the test result is a
poor predictor for future behaviour and performance. For example, a puppy might
be assessed as ‘dominant’ in a test made by the breeder, but the predictive power
for development of dominance-related behaviour in future may be very little, unless
someone has studied the correlation between test results and future behaviour in an
unbiased manner and found positive associations between these two.
In conclusion, personality in dogs is of great relevance for all of us who in
some way have dogs in our lives. There are, however, some difficulties in the
application of the concept of personality in everyday life. One issue is that we
have different ways of describing the typical behaviour of dogs. This may result
in confusion and misunderstandings. Better definitions of terms and labels would
improve communication. Another issue is the relevance of results from behav-
ioural tests carried out by laymen. Poor standardization, insufficient description
methodology and overestimation of the predictive value of the test results are
three possible problems. Nevertheless, behavioural tests organized by serious
breeders, dog clubs or trainers may give valuable information of a dog’s person-
ality. It is, however, important to use the information from these tests with great
care, and avoid far-reaching conclusions.
References
Arvelius, P. (2005) Genetisk och etologisk analys av vallningsbeteenden hos border
collie. MSc thesis, Department of Animal Breeding and Genetics, Swedish
University of Agricultural Sciences.
Beaudet, R., Chalifoux, A. and Dallaire, A. (1994) Predictive value of activity level
and behavioral evaluation on future dominance in puppies. Applied Animal
Dog Personality 205
Behaviour Science 40, 273–284.

Campbell, W.E. (1972) A behavior test for puppy selection. Modern Veterinary
Practice 12, 29–33.
Caspi, A. and Roberts, B.W. (2001) Personality development across the life course:
the argument for change and continuity. Psychological Inquiry 12, 49–66.
Fox, M.W. (1972) Socio-ecological implications of individual differences in wolf
litters: a developmental and evolutionary perspective. Behaviour 41, 298–313.
Funder, D.C. and Colwin, C.R. (1991) Explorations in behavioural consistency:
properties of persons, situations, and behaviors. Journal of Personality and
Social Psychology 60, 773–794.
Goddard, M.E. and Beilharz, R.G. (1984) A factor analysis of fearfulness in potential
guide dogs. Applied Animal Behaviour Science 12, 253–265.
Goddard, M.E. and Beilharz, R.G. (1986) Early prediction of adult behaviour in
potential guide dogs. Applied Animal Behaviour Science 15, 247–260.
Goodloe, L.P. and Borchelt, P.L. (1998) Companion dog temperament traits. Journal
of Applied Animal Welfare Science 1, 303–338.
Gosling, S.D., Kwan, V.S.Y. and John, O.P. (2003) A dog’s got personality: a cross-
species comparative approach to evaluating personality judgments. Journal of
Personality and Social Psychology 85, 1161–1169.
Harding, E.J., Paul, E.S. and Mendl, M. (2004) Animal behavior – cognitive bias and
affective state. Nature 427, 312–312.
Hsu, Y. and Serpell, J.A. (2003) Development and validation of a questionnaire for
measuring behavior and temperament traits in pet dogs. Journal of the American
Veterinary Medical Association 223, 1293–1300.
Jensen, P., Rushen, J. and Forkman, B. (1995) Behavioural strategies or just individ-
ual variation in behaviour? A lack of evidence for active and passive piglets.
King, T., Hemsworth, P.H. and Coleman, G.J. (2003) Fear of novel and startling
stimuli in domestic dogs. Applied Animal Behaviour Science 82, 45–64.
Koolhaas, J.M., Korte, S.M., De Boer, S.F., Van Der Vegt, B.J., Van Reenen, C.G.,
Hopster, H., De Jong, I.C., Rusi, M.A.W. and Blokhuis, H.J. (1999) Coping
styles in animals: current status in behavior and stress-physiology.
Neuroscience and Biobehavioral Reviews 23, 925–935.
Lindsay, S.R. (2000) Handbook of Applied Dog Behavior and Training, Vol. 1:
Adaptation and Learning. Iowa State University Press, Ames, Iowa.
Martínek, Z., Lát, J., Sommerová, R. and Hartl, K. (1975) About the possibility of pre-
dicting the performance of adult guard dogs from early behaviour – II. Activitas
Nervosa Superior (Praha) 17, 76–77.
Netto, W.J. and Planta, D.J.U. (1997) Behavioural testing for aggression in the
domestic dog. Applied Animal Behaviour Science 52, 243–263.
O’Farrell, V. (1997) Owner attitudes and dog behaviour problems. Applied Animal
Podberscek, A.L. and Serpell, J.A. (1997) Environmental influences on the expres-
sion of aggressive behaviour in English Cocker Spaniels. Applied Animal
Roberts, B.W. and DelVecchio, W.F. (2000) The rank-order consistency of person-
ality traits from childhood to old age: a quantitative review of longitudinal
studies. Psychological Bulletin 126, 3–25.
Roll, A. and Unshelm, J. (1997) Aggressive conflicts amongst dogs and factors affect-
206 K. Svartberg
ing them. Applied Animal Behaviour Science 52, 229–242.

Royce, J.R. (1955) A factorial study of emotionality in the dog. Psychological
Monograph 69, 1–27.
Scott, J.P. and Fuller, J.L. (1965) Genetics and the Social Behavior of the Dog. The
Sheppard, G. and Mills, D.S. (2002) The development of a psychometric scale for
the evaluation of the emotional predispositions of pet dogs. International
Journal of Comparative Psychology 15, 201–222.
Strandberg, E., Jacobsson, J. and Saetre, P. (2005) Direct and maternal effects on
behaviour in German Shepherd dogs in Sweden. Livestock Production Science
93, 33–42.
Svartberg, K. (2002) Shyness-boldness predicts performance in working dogs.
Svartberg, K. (2005) A comparison of behaviour in test and in everyday life: evi-
dence of three consistent boldness-related personality traits in dogs. Applied
Svartberg, K. (2006) Breed-typical behaviour in dogs – historical remnants or recent
constructs? Applied Animal Behaviour Science 96, 293–313.
Svartberg, K. and Forkman, B. (2002) Personality traits in the domestic dog (Canis
familaris). Applied Animal Behaviour Science 79, 133–155.
Svartberg, K., Tapper, I., Temrin, H., Radesäter, T. and Thorman, S. (2005)
Consistency of personality traits in dogs. Animal Behaviour 69, 283–291.
Thorne, F.C. (1940) Approach and withdrawal behavior in dogs. The Journal of
Genetic Psychology 56, 265–272.
van den Berg, L., Schilder, M.B.H. and Knol, B.W. (2003) Behavior genetics of
canine aggression: behavioral phenotyping of golden retrievers by means of an
aggression test. Behaviour Genetics 33, 469–483.
Weaver, I.C.G., Cervoni, N., Champagne, F.A., D’Alessio, A.C., Sharma, S., Seckl,
J.R., Dymov, S., Moshe Szyf, M. and Meaney, M.J. (2004) Epigenetic program-
ming by maternal behaviour. Nature Neuroscience 7, 847–854.
Wilsson, E. and Sundgren, P.-E. (1998) Behaviour test for eight-week old puppies –
heritabilities of tested behaviour traits and its correspondence to later behav-
iour. Applied Animal Behaviour Science 58, 151–162.
Human–Animal Interactions and
12
Social Cognition in Dogs
Ádám Miklósi
Introduction: Dogs in the Human Niche
In biology, or more precisely in ecology, the concept of a ‘niche’ is a theoretical

characterization of a species’ environment, and it is understood that strictly speak-
ing no two species can live in the same niche. ‘Niche’ is also used, as in this
chapter, in a somewhat looser formulation to describe the ‘psychological and
social’ environment of humans that is shared to some extent or has some overlaps
with dogs. There can be no doubt that dogs and humans have developed a par-
ticular relationship over the last 10,000 years. Dogs, along with other animals like
sheep, cattle, chickens and cats, have often been ‘lumped’ together and referred
to as ‘domesticated’ animals. Additionally, there has been a continuous interest in
understanding the process of domestication because for many it seemed to be a
good case for modelling evolutionary processes (Darwin, 1859).
In this research people were looking for general features of changes that could
account for all or most cases of domestication, irrespective of the species in
question. For example, many noted that there is a reduction in relative brain
size when one compares the ancient wild species with its domesticated descen-
dants (Kruska, 1980). Although it is important in its own right, this approach
has overlooked the fact that most domestication events took place at different
times, and the original causes for domestication were also probably quite
different for different species. Even the aims of domestication (selection) might
have changed over time. This suggests that it would be perhaps more advanta-
geous to consider each domestication event (species) separately and investigate
more closely the actual relationship between humans and the particular species in
question.

(ed. P. Jensen)
208 Á. Miklósi
Although dogs often have been the focus of biological studies of domes-
tication, and they never lost their appeal for the veterinary sciences or resear-
chers interested in the applied aspects of animal behaviour, to many people
dogs still represent a kind of ‘artificial’ animal. Hundreds of ethologists or
animal psychologists working and publishing on ‘wild and real’ animals have
possibly shared their lives with one or more dogs at home. Nevertheless, they
seemed to have overlooked the possibility that their favourite companion could
not only be an object of compassion and love but also a subject for scientific study.
This is clearly reflected in the lack of studies on dog behaviour in major journals
of the field like Animal Behaviour, Behaviour or Ethology during the last 40 years
or so.
Although one might still consider the dog as an ‘artificial’ animal (note that
there is no such definition as what ‘artificial’ would mean), it is hard to deny that
looking closely and with an evolutionary approach at the dog–human relationship
we find that the existence of the two species is tied together closely. In any case,
the minimal task of an ethologist should be to explain the present situation when
about a quarter or a third of the human population in various countries on the
Earth share their lives with dogs. Further, we observe that at present dogs come
in various ‘forms’ (breeds) with regard to morphology, genetics and behaviour,
and many see this as an analogy to ‘adaptive radiation’ observed in many suc-
cessfully expanding species in biological evolution (Parker et al., 2004).
In my view the most important change in the scientific approach has been
that researchers have begun to see the dog as a result of an adaptation to a spe-
cific human environment. It has been suggested that behavioural changes from
the ancestor of the dog to present-day dogs made it possible for dogs to ‘invade’
or be ‘incorporated’ into the human environment. In evolutionary terms, where
the number of animals in a given species is used to describe its ‘success’, at present
dogs seem to be a very successful species. They not only come in large numbers,
but they are also among the few mammalian species that are distributed over all
continents (with the exception of Antarctica).
Although in general one can discriminate ‘human niche’ from ‘niches’ of
other species, looking more closely there are many types of human environments,
even in relation to the dog. This might not have been the case at the start of our
contact with the dog or some dog-like species, but at present humans can live in
various types of relations with dogs. Often this relationship is associated with
some dog ‘breeds’, emphasizing how dogs were able to adjust on an even more
‘sophisticated’ level to the human environment. For example, some dogs (or dog
breeds) are recognized as ‘working dogs’, e.g. ‘sledge dogs’ etc. This categoriza-
tion has some biological validity, but we need to recognize that in this case ‘evo-
lution’ seems to run at a high speed, and there are other ‘niches’ that dogs can
and possibly will fit into, like working for the blind (or other people with disabili-
ties) or enforcing law (police dogs), living as pets, etc. If one prefers the evolu-
tionary comparison, it is interesting to see that this diversity of human–dog
associations is in strong contrast to the relatively uniform type of social struc-
ture that characterizes, for example, wolf groups that have been regarded as
Human–Animal Interactions and Social Cognition 209
the best representatives for modelling the life style of the ancestor species. This
brings us to the question of what kind of behavioural modifications took place in
the dog.
Looking again at the literature on dog behaviour, one might note a further
interesting trend. Books written by authorities on canine behaviour most often
follow the traditional trend and describe the intraspecific behaviour of various
related species such as wolves, coyotes or jackals. In this respect dogs represent an
interesting ‘case’ where human influence/domestication ‘mixed up’ behaviours
that have previously been under ‘natural’ selection. People advocating this
approach are interested in finding out how intraspecific wolf behaviour (how
wolves behave and communicate with members of the same species) has been
transformed into intraspecific dog behaviour. In contrast, practical books on dog
behaviour, most of them written for the ‘public’, often with the aim to improve
‘training skills’ of dog owners, place more emphasis on the interspecific (between-
species) aspect of dog–human behaviour, that is, how humans should use their
behavioural repertoire in order to influence the behaviour of dogs, and they teach
us how to ‘read’ and ‘understand’ dog behaviour (i.e. Yin, 2004). Although it
seems to be the case that interspecific behaviour of dogs (and humans) has some
practical importance, this aspect has not received attention from ethologists or
animal psychologists. One might even assume that for present-day dogs, inter-
specific behaviour might be of even greater importance than intraspecific behav-
iour, because dogs spend more of their lifetime in contact with humans than with
members of their own species.
Testing the Bond: Attachment Behaviour in Dogs
Our approach has always been that dogs and humans represent a mixed-species
group. However, we have also assumed that these human–dog groups are to some
degree more similar to intraspecific groups than simply being composed of two
different species feeding together or sharing the costs of an efficient predator
warning system. In contrast, we assumed that just like in the case of social animals
living in groups, dogs and humans develop close interindividual relationships,
which could take various forms such as attachment relationship, dominant–sub-
missive relationship or friendship.
We should admit that for a long time there has been an interest in charac-
terizing the relationship between dogs and humans, or more precisely the rela-
tionship between owners and their dog. However, many early investigations took
an anthropocentric view, and were interested mainly in how owners perceived
their relationship with their dogs or whether this relationship could be considered
as a kind of attachment in a psychological sense (Voith, 1985; Serpell, 1996). This
work resulted in the development of ‘bonding scales’, measuring the attitude of
people towards their dog or dogs in general. Although this research was very
influential, by their nature questionnaire studies cannot tap into the biological
basis of the dog–human relationship.
210 Á. Miklósi
In order to describe the dog–human relationship at a group level one needs

studies that are looking specifically at the behaviour of both parts. Luckily, in
another line of investigations psychologists have developed observational methods
that seem to be able to describe mother–child relationships at the behavioural
level. This so-called ‘strange situation test’ is based on the observation that mildly
stressful situations (novel environments) activate the ‘attachment system’ of the
subject, that is, in the case of humans infants, to express an array of behaviours
which characterize the actual relationship between them and the mother
(Ainsworth et al., 1978).
Observing the parallels between dogs and children in relation to (adult)
humans in the group, we decided to apply the ‘strange situation test’ in order to
describe dog–human relationships. The logic of the test is that after having
evoked a mild stress by placing the subject into a novel environment, it is observed
either in the presence of the owner or a stranger (a person never seen before). The
attachment behaviour of the subject will then be evaluated on the basis of com-
paring the behaviour shown toward the owner and the stranger. There are three
critical aspects: (i) the subject should show more affiliative behaviour towards the
owner than the stranger (e.g. play); (ii) it should display a preference for contact
and proximity towards the owner; and (iii) it should display specific behaviours
towards the owner after separation.
A pioneering study in this respect showed that dogs can be tested very easily
in a modified version of the ‘strange situation test’ (Topál et al., 1998), and we
have also developed a quantitative measure of their behaviour, which is different
from the more qualitative-type of scoring method used by the psychologists to
describe mother–infant attachment. This and other subsequent studies (Prato-
Previde et al., 2003; Marston et al., 2005) provided evidence that the behaviour of
dogs fulfils the above-mentioned criteria for attachment behaviour. At a group
level, dogs showed a clear preference for spending more time with the owner both
during the test and also during the ‘reunion’, when the owners returned after
having left the dog for a short period of time. During the test dogs preferred to
play with the owner. Further, we have found that dogs could be categorized into
different ‘types’ with regard to their attachment behaviour shown towards their
owners based on their scores achieved on the three factors, which has been devel-
oped following a factor analysis on the behavioural data (see also Chapter 11 in
this volume for a similar approach). These factors relate to the effectiveness of
evoking ‘stress’ in the dog (‘Anxiety’), describe the dog’s behaviour toward the
stranger (‘Acceptance’) and the owner (‘Attachment’). For example, by interpret-
ing multivariate statistical results, we have found that in a group of 51 dogs there
were two subgroups (12 and 18 dogs) where subjects showed relatively few signs
of stress as a result of being tested in a novel environment. Nevertheless, they
clearly differed in their relation to the owner and the stranger – dogs in one of the
groups seemingly preferred to interact with the stranger and vice versa.
Since then, it has been established that the ‘strange situation test’ is very
robust, that is, dogs tested after 6 months show similar behaviour, and the testing
environment has no influence on the outcome (however, it most be a novel place).
Although until now no direct breed comparisons have been done, at least with
reference to the Belgian shepherd it is known that individuals can show the same
variation in attachment as observed for a population where each dog represents
a different breed (Gácsi, 2003).
Communication with Humans
In animal behaviour communication is described as the use of specifically evolved

behavioural patterns in order to modify the behaviour of a recipient to the advan-
tage of the signaller. Although in most cases in the animal kingdom such com-
municative systems are supposed to be under strong genetic influence, learning
can still play a part. For example, in the case of the wolf, Ben Ginsburg (1976) has
observed that although individuals raised separately from the pack had no prob-
lems in displaying species-specific dominant aggressive signals toward humans
from the beginning of their life, they seemed to have problems in showing sub-
missive signals later when encountering other wolves. This suggests that whilst
behavioural signals of dominance emerge as a result of maturation, the establish-
ment of the submissive signal repertoire and its utilization needs social experi-
ence.
With regard to the dog there could be (at least) three potential ways to
explain dog–human interspecific communicative behaviour. First, as both
humans and dogs are highly social animals and have a range of signals for social
interactions, they can simply rely on their common mammalian heritage for com-
munication. However, it is more likely that efficient communication requires
members of both species to engage in some sort of learning in order to be able to
recognize each other’s signals. Finally, one could suppose that behavioural evolu-
tion has given rise to specific signals, which are used predominantly in
dog–human communication. Note that these three hypotheses are not mutually
exclusive, so all three could explain parts of the rich communicative interactions
that take part between dog and human.
In the minds of many people, humans are the ‘communicative champions’
amongst all animals, since they use various ways and means for communication,
and have even developed a unique system of communicative behaviour referred
to in general as language. From our point of view the question is to what extent
dogs have modified their communicative system in order to be able to recognize
human communicative signals or in addition to emit signals that are easily (or
preferentially) discernible by us.
The role of gestural and facial signals in dog–human communication
Recently, there has been an increased interest in visual communication between

humans and dogs. The reason for this might sound somewhat strange. It has been
found that apes seem to be relatively deficient in their ability to communicate
212 Á. Miklósi
with their human partners (experimenters) in various tasks involving human

visual gestures. Importantly, it has been found that dogs can perform somewhat
better in some tasks, so this suggests that in the course of evolution dogs have
evolved some new skills that enable them to understand some human signals, and
this ability is not shared between us and our nearest living relatives at present (the
apes). It should be noted that this is probably not the last word on the compari-
son between dogs and apes with regard to their performance in human-oriented
communicative tasks, but if one compares the behaviour of the dog to the behav-
iour of the wolf, the picture becomes clearer.
From the dogs’ point of view, wolves produce a better comparison. Although
it is likely that today’s wolves are different from those animals that once gave rise
to the dogs, it still seems to be the most interesting comparison. For example,
genetically, dogs are most closely related to modern-day wolves. Therefore one
could hypothesize that if we minimize the environmental differences (or in other
words experience) during the behavioural development of wolves and dogs, the
comparison of the two species should show those behavioural differences which
are most likely to rely on genetic influence, and characterize those traits as having
evolved as a result of dog–human interaction. For this reason, we embarked on a
programme providing ‘maximum’ human socialization experience for wolves and
dogs (Fig. 12.1). Altogether 13 socialized wolves were cared for by their human
Fig. 12.1. Hand-raising wolves. The pups enjoyed 24-h contact with the
caretaker during their first couple of weeks.
‘stepmothers’ for 24 h a day until 3–4 months of age. The wolf pups were
removed from their mother on days 4–6, before their eyes opened. From this day
on they spent most of their day with their carers, who often carried them around
and even took them to bed in the evening. After independence, the wolves were
trained to walk on the leash, and they were carried around to familiarize them
with various places, humans and different objects. They were regular visitors to
our university, and travelled by public transport as well as in cars. Each caretaker
nursed only one wolf, but arrangements were made for the pups to see each other
regularly and play for long periods.
Starting in their third week of life, the wolf pups were regularly observed in
a series of behavioural experiments testing for social preferences (dogs versus
humans) (Fig. 12.2), neophobia (fear of novelty), reaction to dominance, retrieval
of objects and communication with humans, the results of which will be reported
later. These tests continued until the 13th week of age, when the animals were
taken to a farm where there were other wolves. During the next 6 months the
wolves were slowly integrated into that pack, but once or twice a week they were
visited by their caretakers, and taken out from the pack for testing and free social
interaction (Miklósi et al., 2003).
Response to human-given cues was tested in the two-way object choice task
when the subject has to find a hidden piece of food on the basis of human gestural
signals (e.g. pointing). The basic idea of this test is to hide a small piece of food in
one of two containers. Next the containers are placed on the ground and the
subject is offered a choice. However, it is allowed to go for one or the other loca-
tion; the hider standing between the two containers displays a gesture signalling
the place of the hidden reward. Such gestures could have various forms such as
pointing with extended arm and hand, turning the head or only moving the eyes
to the side (for more details, see Soproni et al., 2002). Significantly better per-
formance exceeding the chance level (being 50%) suggests that in their choice
dogs are relying on the human gestures. (Importantly, this seems to be also a good
test for investigating the ‘Clever Hans’ effect in dogs. (Clever Hans was a famous
horse that performed various tricks by learning minute body signals of his trainer
(Pfungst, 1965).)) By the time of testing the wolves, we had learnt that dogs per-
formed well in similar tests; however, we did not know whether this was the result
of some genetic preparedness or extensive experience with humans. Despite all
our efforts to provide intensive social experience for the wolves, our lovely sub-
jects performed quite disappointingly in these tests. Wolves seemed to have no
problem finding the hidden food if the human stood behind the correct container
or indicated the place of the food by touching and manipulating the container for
some seconds. This convinced us that our wolves understood the task, and were
able to perform well under the given experimental conditions, which is always
important if one aims at comparing different animal species in the same task. In
contrast to the dogs, however, wolves were quite poor in choosing the correct
location of the bait if it was indicated by a pointing gesture, and their perform-
ance equalled chance when the human pointed to a container that was approxi-
mately 50 cm away from the tip of her fingers. Although this latter type of gesture
214 Á. Miklósi
(a)
(b)
Fig. 12.2. When given a choice between a human and a dog, hand-raised and
socialized wolf puppies prefer to stay with the dog (a), in contrast to dogs, which
choose the human (b).
reduced the performance of dogs, they still seem to be able to perform well over
chance level. Further experiments showed that dogs actually do respond to some
body parts that protrude the body torso. In line with this, dogs can find the loca-
tion of hidden food if humans point with the collateral hand (Fig. 12.3).
Interestingly, dogs respond also to pointing with the leg, suggesting that they can
generalize to some extent when it comes to ‘reading’ communicative signals.
One obvious explanation for such abilities in dogs is to refer to analogous
communicative signals in the species-specific behaviour of dogs. Indeed, the bio-
logical basis of understanding the human pointing gesture could be the fact that
dogs often assume a particular body position, ‘pointing’ with their whole body
towards some particular location in the environment, and one could easily witness
this behaviour in an exaggerated form in the pointers. However, this does not
Fig. 12.3. A typical ‘natural’ experiment. The human experimenter points at the
baited container to a hand-raised and socialized wolf with his caretaker. Usually
wolves do not perform well in these tests.
explain the wolf–dog difference because such pointing behaviour is also part of
the wolf ethogram.
In the course of the pointing experiments, we noted that it was quite difficult
to direct the attention of the wolf to the experimenter performing the commu-
nicative action. Based on these observations, we devised an experiment that chal-
lenged dogs and wolves in similar ways. Nine socialized wolf pups and nine
juvenile dogs of the same age were trained to pull out a piece of meat attached to
a rope from a cage. In six training trials, subjects learnt with similar speed to get
the food out by pulling the rope either by their mouth or forelegs in the presence
of their owner, who was standing behind the animal. No differences were found
between the two species in the acquisition of the task, their latency to eat the meat
decreased to the same extent as they learnt the behaviour. This training was fol-
lowed by a 2-min long ‘inhibited’ trial when the rope was inconspicuously fas-
tened to the wire of the cage so that pulling the rope was ineffective. In this
situation dogs looked much earlier and for longer at their owner, whilst wolves
looked ahead, in the direction of the food in the cage (Fig. 12.4).
We know from studies on humans that such gazing behaviour is one of the
most important aspects of human visual communication, and indeed we found
something very similar in one of our earlier studies. Here the experimenter hid a
piece of food in the presence of the dog and absence of the owner. However, the
food was unreachable for the dog (on a high shelf in the living room). We found
216 Á. Miklósi
Fig. 12.4. One of the young wolves learns how to pull the string. The caretaker
does not interfere with the training but her presence gives the wolf puppy some
social support.
that after the owner returned to the room, dogs emitted both visual and acoustic
signals aimed at getting the owners’ attention and directing it to the location of
the food. Taken together, this suggests that dogs are able to modify their com-
municative behaviour toward humans by using signals that are characteristic of
human-specific communication systems (see Miklósi et al., 2004).
The role of gazing in dog–human communication has been underlined in
other experiments looking at dogs’ ability to recognize human attention. In a
communicative interaction it is vital to make certain that the receiver is actually
in the state of receiving the signal. The monitoring of attention can be done quite
easily by recognizing that the orientation of gaze (and eyes) provides a clear indi-
cation for this. Again, apes living in laboratories with humans did not show very
sophisticated abilities, but dogs seem to be more inclined to use the gaze and eye
cues of humans as behavioural signals for attention. In line with this, dogs pre-
ferred to beg from humans who were looking at them, and retrieved objects to the
‘attentive-side’ of the owners, that is, retrieved objects were presented mostly to
the facing owner. Interestingly, dogs seem to be sensitive also to the orientation
of the owner when responding to verbal commands. In everyday situations,
gazing and command goes hand in hand: we are looking at the dogs whilst
emitting a command. It has turned out that dogs do not respond blindly to the
commands but also take into account whether it has been accompanied by the
appropriate cues signalling the attention of humans (Virányi et al., 2004). If the
command was issued while the owner was orienting to some other location in
space or to another human present, many dogs did not obey.
Especially the latter is very interesting, because dogs seem to be inclined to
obey the command more if there was nobody else present in comparison to the
case when the experimenter directed the command to another human in the
room. One might suppose that dogs have learnt to use visual cues of attention in
order to discriminate situations depending on whether they are in the focus of
their owner or other humans. Finally, the dogs’ ability to recognize humans’
attention is also well known from everyday experience when we are ‘competing’
with our companions for food that ‘lies around’, for example, in the kitchen.
When modelling the situation in the laboratory, researchers have shown that after
forbidding the dog to eat a piece of food (Call et al., 2003) the animals were more
ready to disobey the command if they conceived the human as not looking/ori-
enting at them (e.g. she was playing a video game or reading a newspaper).
Communication by sounds: words and barks
For many, the main channel of dog–human communication is acoustic, as

humans use verbal signals (mostly parts of language) to influence the behaviour of
dogs. This aspect of interspecific communication has a very interesting past in the
behavioural sciences, and raised much debate – some have even assumed that
(some) dogs might have the ability to understand simple forms of human lan-
guage, whilst other maintained that dogs do not even respond to the words, but
only to some conspicuous acoustic features of verbal signals. Perhaps not acci-
dentally, one of the first truly experimental papers on a dog tested his ability to
respond to various verbal commands excluding the possibility for visual cuing by
the owner (Warden and Warner, 1928). In this study the owner commanded the
dog from an adjacent room, ordering him to sit down, go under the table, retrieve
objects, etc. Despite the fact that the situation was very unfamiliar for the dog, he
seemed to perform quite well, showing evidence that he made the association
between the verbal signal and the behaviour to be performed.
Interestingly, until now there are only few studies investigating to what extent
dogs understand human verbal signals. Are they only relying on some general
acoustic features of the word (would they mix ‘Sit! with ‘Sick!’) or do they place
more emphasis on the accent, etc.? (See also Chapter 8 in this volume for a
further discussion of this aspect.) A further interesting question is whether dogs
recognize the structure of command sentences (e.g. Bring the ball!) as being com-
posed of a verb and an object, or the action becomes inseparable from the goal.
There are also large individual differences among dogs (even in the same breed)
with regard to how well they can be ‘trained’ to respond to verbal signals. This
fact has been demonstrated in a recent case reported in the journal Science.
German researchers found that a dog, named Rico, could retrieve about 200
218 Á. Miklósi
different objects after hearing a command (Kaminski et al., 2004). Such a ‘vocab-
ulary’ closely corresponds to the knowledge of words by some apes trained to
understand human visual signals. Further research with Rico showed that he was
able to extend his knowledge of object names continuously, and that very little, if
any, training was needed for the establishment of a novel word–object associa-
tion. In one task Rico was provided with a set of objects, some of which were
novel for him. First, he was commanded to retrieve 2–3 familiar objects, and then
he was given a command incorporating a novel word. In response to the novel
command Rico retrieved a novel object in most cases, and moreover this single
trial was enough to establish a stable memory about the name of the object, as he
was quite successful in retrieving the object after several months lapse, even if he
had no further contact with the object in the meantime.
Whilst in this experiment Rico learned the names of the object by listening
to commands directed at him, there is now some evidence that dogs can also learn
the association between acoustic cues (words) and objects if they witness humans
talking about them. The experimental method for showing such ability has been
‘imported’ from research on parrot communication. Grey parrots seem to be
more efficient in learning (and of course using) human words if they are placed in
competitive situations, where they have to compete with a human for the atten-
tion of the ‘teacher’ (Pepperberg, 1991). Later in training, of course, not just
praise is used, but sometimes also more conceivable forms of reward, such as
food. In a somewhat modified experimental situation dogs could witness the con-
versation of two humans who interacted and exchanged a novel object whilst con-
tinuously talking and referring to the object by a novel word. After a short
exposure dogs were commanded to retrieve the object. McKinley and Young
(2003) found that the dogs showed the same level of performance if they had been
trained by conventional (operant) training methods. This observation suggests
that dogs have the ability to ‘pick out’ single cues (‘words’) from a stream of
acoustic signals and understand that these might refer to some parts of the envi-
ronment.
Independently from what dogs might or might not understand from human
linguistic signals, every day dog owners continue to talk to their dogs. True
enough, many of us also talk to cars, computers or washing machines (especially
if they refuse to ‘cooperate’), but probably we would not claim that these
machines ‘understand’ us. In contrast, the majority of owners believe that their
dog ‘understands’ them to a certain extent. In a survey, we asked owners to
provide a list of commands and describe the ‘content’ of conversations they have
with their dog. We have found that in most cases owners refer to objects or try to
prevent the dog from doing something. Additionally, owners often provide infor-
mation or ask questions to the dog (Pongrácz et al., 2001). At least according to
the experience (or belief) of the owners, in most cases dogs seem to ‘understand’
the utterance because they show appropriate change in behaviour, either ‘always’
or given a ‘particular situation’. Thus, for example, when opening the car door
an owner might say, ‘Get into the car!’, and after seeing the action performed
the owner might deduce that the dog has indeed understood the meaning of the
utterance. This would be especially convincing if the dog would hesitate to

execute the expected action without the command/request. Although it is more
than likely that dogs do not understand the meaning of the request in linguistic
terms, it could be that the acoustic utterance contributes to the dog’s under-
standing of the situation.
Recent research has shown that acoustic communication between dogs and
humans is not a one-way process. It is well known (and experienced everyday)
that dogs are relatively ‘vocal’ (a dog hater would even say ‘noisy’) animals in con-
trast to other mammals. The simple reason is that dogs have a propensity to bark
very often in various situations, and barking can also easily be facilitated by other
dogs’ barking. Barking has long been dismissed as a ‘by-product’ of domestica-
tion, having no obvious function. However, it has turned out that barking might
play a role in dog–human or even dog–dog communication.
Playback studies have revealed that humans seem to be proficient in catego-
rizing barks if provided with a list of possible scenarios of recordings, and they
were also skilful in describing the ‘emotional’ content of the bark. This means that
if a bark was recorded whilst a dog was ‘attacking’ a human (‘Schutzhund situa-
tion’ enacted at a dog training school, see Pongrácz et al., 2005), then the same
bark was described (only by listening to it) as being ‘aggressive’. Importantly, this
ability in humans did not depend on previous experience with dogs because we
have found no major difference between the performance of dog owners and
humans who have never owned a dog.
Comparing the vocal ethogram of wolves and dogs, it turns out that barking
is used only in very specific contexts in the case of the former species. According
to Schassburger (1993), wolves bark in contexts of warning/defence and protest
only, and they usually produce one or a few distinct noisy (atonal) sounds with low
dominant frequency (Feddersen-Pettersen, 2000). In contrast, barks are charac-
terized by a wide range of frequencies in dogs and they are often repeated over a
long duration. Additionally, dogs ‘invented’ the harmonic version of the bark, not
recorded in the case of wolves. This means that barks are used to express a much
wider range of emotions in dogs in comparison to wolves. It is very likely that this
change in the vocal behaviour of dogs was the result of interaction with humans
(the other ‘noisy’ terrestrial mammal on Earth!), as humans might have preferred
individuals with a similar preference for expressing emotions by vocalizations.
Communication and social cognition
The world-famous cognitive ethologist Donald Griffin used to argue that the
study of communication in animals provides us with a window for understanding
their cognitive abilities. With respect to dog–wolf comparisons, Frank (1980) has
also supposed that there is a major difference in the ‘information processing
system’ of wolves and dogs. He argued that the wolf possesses two distinct infor-
mation processing systems working separately. The first deals with a narrow band
of stimuli associated with behaviours directly coupled to survival. The second
220 Á. Miklósi
system provides a very flexible processing of environmental information, which,

among others, offers the possibility for generalization and learning from others by
observation (see Chapter 8 in this volume for a further discussion on social learn-
ing). In dogs, these two systems seem to have merged, enabling the processing of
a wide range of stimuli with regard to most aspects of behaviour and also to show
an extended behavioural plasticity in response to environmental challenge. At the
level of dog–human interaction, the change in the information processing systems
allows for greater trainability in the dog when arbitrary stimuli become associated
with different types of behaviour.
The problem of such comparisons of the cognitive processing systems in
wolves and dogs is that we know very little about the wild species, especially when
it comes to performance in complex problem-solving tasks or communica-
tion. For example, there is very little experimental evidence that wolves learn
by observation. However, it has turned out that dogs show enhanced perform-
ance if they have the chance to witness a human demonstrator (Miklósi et al.,
2004). As described above, dogs seem to be able to associate human acoustic
signals (words) with actions, but we know virtually nothing about wolves in
this respect. Nevertheless, observations indicate that they seem to have a
very sophisticated vocal communication system based on single signals and their
combinations.
An alternative approach could be to compare dog social cognition with that
emerging in human children. Both ‘animals’ are often raised in a relatively similar
social environment, which later also becomes their living space. Moreover, there
is a huge amount of information on the social cognitive development of human
children that could be used as a reference. The point here is not that dogs and
children necessarily share cognitive mechanisms, but dogs could have evolved
analogous systems that perform at the same level as seen in children. For
example, we have found that the performance of hand-raised dogs can be
equated to the level of 1.5–2-year-old children with regard to some versions of the
‘pointing task’ (see above).
During domestication, changes appear to have occurred at different levels of
behavioural and cognitive organization. Dogs might have acquired an ability to
generalize to a greater extent than wolves in certain situations – for example,
when communicating with humans. This flexibility is perceived by us as dogs
being easier to train, and also to influence (‘live with’) than wolves. However, we
should note that this trait is actually analogous to our ability to use an arbitrary
set of signals for communication. Therefore one could suppose that the emer-
gence of this ability in dogs might also lead to differences in cognitive processing,
allowing for the dogs to cognitively represent more complex or sophisticated
understanding of social interactions.
In addition, we have provided evidence that in some cases dog behaviour has
changed in the direction of allowing for a more effective interspecific communi-
cation with us. Based on the available evidence, this suggests that these ‘innova-
tions’ are not the result of exposure to the human social environment but may
have some genetic origin.
Acknowledgements
This work has been supported by the Hungarian Science Foundation (OTKA T
049615), the Hungarian Academy of Sciences (F01/031) and the EU (FP6-NEST
012787).
References
Ainsworth, M.D.S., Blehar, M.C., Waters, E. and Wall, S. (1978) Patterns of
Attachment: A Psychological Study of the Strange Situation. Erlbaum, Hillsdale,
New Jersey.
Call, J., Brauer, J., Kaminski, J. and Tomasello, M. (2003) Domestic dogs (Canis
familiaris) are sensitive to attentional state of humans. Journal of Comparative
Psychology 117, 257–263.
Darwin, C. (1859) The Origin of Species. John Murray, London.
Feddersen-Pettersen, D. (2000) Vocalisation of European wolves (Canis lupus) and
various dog breeds (Canis l. familiaris). Archieves für Tierzucht, Dummerstorf
43, 387–397.
Frank, H. (1980) Evolution of canine information processing under conditions
of natural and artificial selection. Zeitschrift für Tierpsychologie 59, 389–
399.
Gácsi, M. (2003) Ethological study of dogs’ attachment toward humans.
Unpublished PhD dissertation, Eötvös University, Budapest.
Ginsburg, B.E. (1976) Evolution of communication patterns in animals. In: Hahn,
M.E. and Simmel, E.C. (eds) Communicative Behavior and Evolution. Academic
Press, New York, pp. 59–79.
Kaminski, J., Call, J. and Fischer, J. (2004) Word learning in a domestic dog: evi-
dence for ‘fast mapping’. Science 304, 1682–1683.
Kruska, D. (1980) Domestikationsbedingte Hirngrössenanderungen bei Saugetieren.
Zeitschrift für Zoology, Systematik und Evolutionsforschung 18, 161–195.
Marston, L.C., Bennett, P.C. and Coleman, G.J. (2005) Factors affecting the forma-
tion of a canine-human bond. IWDBA Conference Proceedings 132–138.
McKinley, S. and Young, R.J. (2003) The efficacy of the model-rival method when
compared with operant conditioning for training domestic dog to perform a
retrieval task. Applied Animal Behaviour Science 81, 357–365.
Miklósi, Á., Kubinyi E., Topál, J., Gácsi, M., Virányi, Zs. and Csányi, V. (2003) A
simple reason for a big difference: wolves do not look back at humans but dogs
do. Current Biology 13, 763–766.
Miklósi, Á., Topál, J. and Csányi, V. (2004) Comparative social cognition: what can
dogs teach us? Animal Behaviour 67, 995–1004.
Pepperberg, I.M. (1991) Learning to communicate: the effects of social interaction.
In: Bateson, P.J.B. and Klopfer, P.H. (eds) Perspectives in Ethology. Plenum
Press, New York, pp. 119–164.
Pfungst, O. (1965) Clever Hans. Holt, Reinhart and Wilson, New York.
222 Á. Miklósi
Pongrácz, P., Miklósi, Á. and Csányi, V. (2001) Owners’ beliefs on the ability of their
pet dogs to understand human verabl communication. A case of social under-
standing. Current Cognitive Psychology 20, 87–107.
Pongrácz, P., Miklósi, Á., Molnár, Cs. and Csányi, V. (2005) Human listeners are
able to classify dog barks recorded in different situations. Journal of
Comparative Psychology 119, 136–144.
Prato-Previde, E., Custance, D.M., Spiezio, C. and Sabatini, F. (2003) Is the
dog–human relationship an attachment bond? An observational study using
Ainsworth’s strange situation. Behaviour 140, 225–254.
Schassburger, R.M. (1993) Vocal communication in the timber wolf, Canis lupus,
Linnaeus. Advances in Ethology 30. Paul Parey Publication, Berlin.
Serpell, J. (1996) Evidence for association between pet behaviour and owner attach-
ment levels. Applied Animal Behaviour Science 47, 49–60.
Soproni, K., Miklósi, Á., Topál, J. and Csányi, V. (2002) Dogs’ responsiveness to
human pointing gestures. Journal of Comparative Psychology 116, 27–34.
Topál, J., Miklósi, Á. and Csányi, V. (1998) Attachment behaviour in the dogs: a new
application of the Ainsworth’s Strange Situation Test. Journal of Comparative
Psychology 112, 219–229.
Virányi, Zs., Topál, J., Gácsi, M., Miklósi, Á. and Csányi, V. (2004) Dogs can rec-
ognize the focus of attention in humans. Behavioural Processes 66, 161–172.
Voith, V.L. (1985) Attachment of people to companion animals. Veterinary Clinics
of North America: Small Animal Practice 15, 289–295.
Warden, C.J. and Warner, L.H. (1928) The sensory capacities and intelligence of
dogs, with a report on the ability of the noted dog ‘Fellow’ to respond to verbal
stimuli. Quarterly Review of Biology 3, 1–28.
Yin, S. (2004) How to Behave so Your Dog Behaves. TFH Publications.
IV Behavioural Problems of Dogs
In this last part of the book, there are only two chapters, dealing with subjects of
great concern to many dog owners throughout the world. When everything
works, the relation between a dog and its handler or owner may be joyful and
harmonic – when things go wrong, it may become a nightmare. Many dogs are
put down every year because they develop behaviour disorders which could have
been prevented or cured. In Chapter 13, we get an extensive overview of the most
common behaviour disorders among dogs, their background and possible causes,
and the available treatment methods.
Even though many behaviour problems of dogs arise as a result of ‘psycho-
logical causes’, as exemplified in Chapter 13, it will always be necessary to exclude
the possibility of physical disease as a cause of the behaviour disturbance. Chapter
14 gives an exhaustive review of different disease states which typically cause
changes in the behaviour of dogs. Whereas diagnosis and treatment of such con-
ditions must be handled by veterinarians, the knowledge provided in this chapter
is of course essential for understanding the full background of the behavioural
biology of dogs.
Behavioural Disorders of Dogs
13
Roger A. Mugford
Introduction
The contract between man and dog would seem to be a straightforward one: we
provide food, shelter and security, expecting only companionship in return. In
reality, the relationships between people and their dogs are complex and
demanding for both species. Social systems in canine societies are as varied as
they are in humans, so it is not surprising that people and dogs together invent
highly individualistic relationships. From an international perspective, we also see
large country by country variations in attitudes towards dogs, and also in how
they are kept. Less industrialized or agrarian societies emphasize the utility-value
of the dog as a guard, herding or hunting animal. In more industrialized societies
such as in Western Europe and North America, the dog’s role as a companion
predominates.
The ease with which domestic dogs adapt and thrive within different home
settings is remarkable. One might have thought that the role of companion or pet
dog would be more secure today and less psychologically demanding than in
former times. However, the writer’s experience is that pet dogs are subject to
complex, often contradictory demands from their people. It shouldn’t surprise us
that inconsistencies in human behaviour sometimes produce irreconcilable con-
flicts amongst our canine companions.
This chapter will explore how things can go wrong in the human–dog inter-
face. The analysis and treatment of behavioural problems in dogs is an active area
of innovation and employment by a variety of professionals, be they drawn from
a veterinary, bio-behavioural or dog training background. Eccentric or unwanted
behaviours are frequently featured in the media illustrating failing owners,

(ed. P. Jensen)
226 R.A. Mugford
whereas in truth they are so often only examples of individual differences between
dogs and in their relationships with people.
Human factors in canine behavioural disorders
The domestic dog is a supremely adaptable species; witness the Siberian husky at
home as a sledge dog for Inuit peoples being transplanted into the hot climate of
Australia or the high-rise apartments of New York and Tokyo. It is remarkable
that dogs demonstrate such plasticity of behaviour, to cope and even thrive in
such different social and physical settings.
In the 21st century, we are witnessing dramatic changes in the demography
and lifestyles of people, and these are worldwide phenomena. Family units are
becoming smaller and in many cases people are living alone rather than within
extended families. With smaller family size comes the likelihood that the rela-
tionship with a pet dog becomes more focused, more intense and even a substi-
tute for relationships with other people. In the UK, 28% of households are
occupied by only a single person, and we see similar social trends towards solitary
living in most Western countries. Dogs living in small homes are liable to be
deserted or left alone by their single owner for long periods of time if he or she is
working, vacationing, etc. It shouldn’t surprise us that in this supremely social
species, many dogs are intolerant of isolation and so become distressed when left
alone.
Another of the conflicts which people impose upon dogs is inconsistencies in
day to day provision of physical contact, love, gestures, recognition and manage-
ment rules. All owners are guilty of this, when on the one hand they stroke or oth-
erwise attend to their pet, say for jumping up in greeting; but the next reprimand
it for doing so. All interactions, both positive and negative, can have potent effects
upon dogs’ behaviour. Being chased for ‘stealing’ an item is rewarded sometimes,
reprimanded on others. Inadvertent reinforcement of such commonly performed
undesired behaviours as barking, jumping up and even aggression are often seen
by canine behaviour therapists.
With increased urbanization and affluence have come fashions for dog-
unfriendly lifestyles in the ‘perfect home’: clean, ergonomic and efficient.
Contemporary trends in interior design can have a distinctly adverse effect upon
the quality of canine life. Overly clean, smooth floors leave dogs unable to walk
comfortably upon say, wood laminate or polished tile surfaces. The behavioural
consequence of living upon such slippery floors may be to increase the impact of
noise in the home as well as to motivate the dog to find refuge on furniture. This
may bring unintended conflict with the owner who prefers his dog to live on the
floor. Another example of adverse impact by humans upon their pet dogs is
increasing exposure to mechanical, electronic and magnetic devices within the
home. The resulting sound pollution and electromagnetic changes have not
received systematic study in relation to animal behaviour. However, at an anec-
dotal level, the writer has seen many cases where dogs exhibit signs of distress
Behavioural Disorders 227
within the home due to the operation of, say, a microwave oven. In such cases,
the symptoms disappear when the dog is outdoors or the microwave removed.
For dogs with high hearing sensitivity (e.g. some border collies), life in the home
can be an unhappy experience. Household cleaning products with unpleasant or
powerful scents are obnoxious to many dogs, yet the manufacturers of such prod-
ucts are unlikely to have run preference/aversion tests upon the domestic animals
which will be exposed to these chemicals in the homes of consumers.
Finally, there is the issue of health and hygiene. An over-concern with clean-
liness in the home and exaggeration of zoonotic risks from contact with pets dis-
courages normal, tactile contact between people and their dogs. Cultural taboos
such as that a dog not be allowed in the bedroom, let alone on the bed, are spu-
riously justified on the grounds of hygiene as well as imagined impact upon social
rank. But to be separated at night removes one-third of the potential time that
might otherwise be available for man–dog interactions. Such owners need to be
reminded that dogs are, after all, animals with rather straightforward biological
and environmental needs, including needs for active, quality interactions by day,
and to sleep with companions at night.
Man–dog communication
Domestic dogs become extremely skilful observers of both vocal and non-verbal
communication by their humans. They are responsive to both the emotional tone
of voice, and its content. Facial and body gestures by owners greatly influence a
well socialized dog, as also do more subtle chemical stimuli arising from sweat and
other body secretions. These are learned skills which will be less well developed
in dogs raised in institutional environments such as kennels, laboratories or
denied significant contact with humans as puppies. The tiniest of gestures or
change of mood in a human can communicate important messages to a dog, such
as that the owner is about to leave on a vacation, embark on a walk with the dog
or is about to arrive at journey’s end whilst driving.
The remarkable observational abilities of dogs are probably not matched by
our ability to read and to understand canine body language. The task is made
more difficult by the great morphological differences between various breeds of
dog – witness extremes of head and body shapes between breeds like the bulldog,
border collie or borzoi.
The schematic outline of body and facial gestures in a ‘normal’ oligocephalic
dog is shown in Fig. 13.1. Note that all parts of the body are involved in these
communicatory gestures, from position of ears, eye, musculature and pigmenta-
tion of the muzzle, exposure of the teeth, the hair, standing position and tail.
Many gestures are likely to be misunderstood by even experienced owners. For
instance, a roll over and a belly display is often depicted as representing ‘submis-
sion’, fear or deference towards a more dominant individual. However, the same
gestures can become ritualized by learning processes as to be part of a playful,
even sexual display. Gestures such as licking of the nose or raising of one paw may
228 R.A. Mugford
Neutral, calm, relaxed

A
Ambivalent signals
Alert, attentive, aroused
– stiff legs, ears, hackles
and tail up
Distance-decreasing Distance-increasing
signals B signals
Overt aggressive
C attack – weight
Play bow and ears
forward, snarl, no F
lip retraction
Defensive fearful
aggression – weight
Calm greeting and ears back,
approach paw up, snarl, lips G
retracted, tail
D tucked under
Overt pacifying
approach – ears
back, paw up, lick
tail held low
H
E Overt fearful threat
Ambivalent signals aversion – head turn,
Less overt threat aversion – paw up, tail tucked
‘freeze’ I under
Solicitation of attention/overt
threat aversion – lying down, J
leg up
Fig. 13.1. The body language of dogs, related to different social attitudes.
Reproduced with kind permission from BSAVA.
have enormous significance for particular dogs, but less in others. These are part
of a so-called appeasement complex of gestures, designed to invite approach and
friendship in potentially threatening encounters.
A detailed description of canine signals and their significance in social behav-
iours is provided by Shepherd (2002). These are of especial importance to owners
when discussing aggressive or threat behaviours by dogs. Most dog bites are
Fig. 13.2. The ‘ladder of aggression’ in dogs, outlining the different escalation
phases a dog will typically go through in connection with aggressive behaviour.
Reproduced with kind permission from Kendal Shepherd.
evoked or precipitated by inappropriate actions of humans. The important skill

needed by people interacting with dogs is to recognize the interaction between
signals representing threat-reducing ‘appeasement’ gestures versus those sig-
nalling defensive, even offensive, threat. Shepherd’s ‘ladder of aggression’ in Fig.
13.2 shows the fluidity or interconnectedness of dogs’ reactions which are alter-
natively threatened or threaten.
Consequences of behavioural disorders
Numerous statistics attest to the life-threatening penalty for any companion

animal that presents a behavioural problem to its owner. In this hygiene-con-
scious age, a dog that eliminates in the home is liable first to receive unremitting
punishment, then to be excluded to the outdoors, then finally to be euthanased.
Aggression or injury to the owner by the dog is similarly a strong predisposing
factor to being euthanased. Estimates of the proportion of young and juvenile
dogs euthanased vary between countries, but most studies suggest that more than
half of such requests for euthanasia are because they engage in inconvenient or
dangerous behaviours (Stead, 1982; Sigler, 1991).
230 R.A. Mugford
Another, obvious consequence of misbehaving is that the dog is rehomed.

The perfect lifestyle in a perfect home is not compatible with, say, a labrador that
sheds hairs, occasionally defecates, often smells, and chews the furniture when left
alone. And yet, any informed observer would find nothing ‘wrong’ with the dog;
problems arise because of mistaken human expectations and conflicts of lifestyle.
Nowhere on the planet is entirely free from automobiles and cars pose the
greatest tangible threat to the freedom and enjoyment of dogs. Because the risk of
dogs being killed by cars is so obvious, they must be restrained on leashes and
escorted from here to there by humans. Tolerance of being leashed is hardly a
trait that featured strongly in the long-term domestication of the dog and it may
create a host of frustrations for dogs, at the very least preventing their exchang-
ing normal body signals with other dogs.
Consequently, dogs that are aggressive or intolerant of other dogs represent
over 50% of aggression cases referred to our Animal Behaviour Centre. Training
such dogs to recover their normal social skills is a relatively straightforward
matter, using equipment such as long lines or extending leashes, headcollars, har-
nesses and well-timed reinforcement of tolerant signals amongst trained, non-
threatening stooge dogs. To own a dog which is aggressive towards other dogs can
produce a dramatic decline in quality of life or enjoyment of the human–animal
relationship. Owners become social outcasts whilst walking their dog, often going
to deserted spots early in the morning or late at night. The pleasures of meeting
and being amongst other dogs and their owners are taken away by having a pet
liable to fight.
Owners of dogs with unwanted behaviours are often made to feel failures by
their contemporaries and guilty that it was somehow ‘their fault’. As we shall see
later in this chapter, the underlying causes of behavioural problems are many and
varied, and are not necessarily directly attributable to their owner’s behaviour. But,
of course, sometimes they are!
Finally, there is the matter of legislation, which in most developed countries
makes owners liable for dangerous behaviours in their pet. Dogs which are dan-
gerously out of control and might injure somebody can bring prosecutions for
criminal behaviour upon their owners in the UK, as in many other countries.
Most countries have introduced such legislation, though the boundaries of what
constitutes danger or aggressive behaviour is often subjective and dealt with by
the courts on a case by case basis. Confusingly, in the USA such legislation is not
enacted at a federal level nor even on a state by state basis, rather by local county,
city or town ordinances.
Some countries have legislation based upon particular breeds, such as dogs
above a given height or bodyweight, alternatively of breeds such as American pit
bull terriers, which were originally bred for fighting. Such breed-specific legisla-
tion usually has no rational statistical basis for enactment and produces real fear
and unhappiness when, for instance, a breed such as the Staffordshire bull
terrier is confused with the superficially similar pit bull terrier. These matters fre-
quently appear before the courts in all Western countries, where expensive and
emotionally charged confrontations are enacted with the possibility of a death
sentence. It is a measure of the love and determination of dog owners that so

many owners strongly defend their pet in the face of such seemingly dog-
unfriendly legislation.
Examples of Behavioural Problems
A popular adage amongst dog folk is that one man’s behavioural nightmare is
another’s perfect pet. A border collie that chases and attacks joggers will proba-
bly also be a good herding dog, good at flyball and heelwork to music. A dog that
barks at night may be a useful alert to intruders, alternatively an irritation to
insomniacs. So much of what constitutes a behavioural disorder or problem is in
the eye of the beholder and there can be no operational definition except by ref-
erence to the wants or expectations of the owner. To treat the problem, however,
the wants and expectations of the dog have also to be taken into consideration.
There has been a tendency to medicalize the subject of canine behaviour
therapy (Overall, 1997), to apply precise-sounding diagnostic terms such as ‘hyper-
attachment syndrome’, ‘obsessive compulsive disorder’ (OCD) or ‘dancing dober-
man syndrome’, etc. The use of behavioural ‘labels’ or ‘diagnoses’ such as these is a
subject of fervent debate amongst veterinarians and animal behaviour therapists:
does such a label disguise our ignorance of underlying causative mechanisms? This
writer’s view is that the boundaries between abnormal and normal in canine behav-
iour are so blurred as to make the debate pointless. Better to fully describe the behav-
iour and its context rather than to prematurely label it as a syndrome. There are, of
course, many clinical and physiological factors which contribute to changes in the
behavioural or emotional state of dogs. Of these, pain is the most important factor
(see later), as also are a number of endocrine and metabolic factors.
The types of behaviours about which owners complain are many and
various, but referrals to behaviour specialists such as the writer are overwhelm-
ingly concerned with aggression-related problems. Separation problems are
essentially a reflection of a normal attachment process, where a dog misses its
human companions and in distress might howl, lose accustomed bowel/bladder
control or engage in destructive chewing. Elimination problems or failures to
learn to urinate/defecate in approved places at approved times are also a
common source of complaint by dog owners, especially where an owner’s lifestyle
requires that a dog does not have easy access to a garden or other outdoor area.
Finally, a large number of behavioural problems are triggered by fear or anxiety.
These may be of a generalized nature, where the dog is afraid of many and most
everyday situations; alternatively, there may be an acute and overwhelming fear
of one or a few stimuli. The latter are known as phobias, and the most common
canine phobia is towards loud noises such as from gunshot or fireworks.
Methods for treatment of bang phobias in dogs are now well developed by a
combination of desensitization tapes, some means of physically attenuating the
dog’s hearing, clicker training and possibly short-term use of anxiolytic drugs,
including homeopathic remedies.
232 R.A. Mugford
The skilled behavioural practitioner requires detailed information about the

problematic behaviour and a detailed questionnaire may usefully be employed,
e.g. BSAVA (2002). This defines the nature of the behaviour which the complaint
is about, where it occurs, when it occurs and who in the family is most affected.
Answers to these questions will probably then reveal why the behaviour occurs,
and from that will come pragmatic strategies for behaviour modification.
Strategies for Treatment

The practice of animal behaviour therapy is both an art and a science. It has
become a popular area of employment, with a caché rather like that of some afflu-
ent people hiring a personal fitness trainer. Many of those who offer behavioural
services have no formal biological or clinical training, yet are free in describing
themselves as ‘behaviourists’, ‘behaviour consultants’, ‘dog listeners’, etc. There is
conflict between veterinary and scientifically trained practitioners versus those
who come to the field as a vocation or as a pleasant, profitable route to self
employment.
Regulation of the field has been slow in coming in all countries, and regis-
tration of qualified personnel is voluntary everywhere. Distinguishing between a
competent and qualified individual versus someone unqualified or with spurious
qualifications is difficult for dog owners and they should seek a referral
or recommendation from the pet’s usual veterinarian. Best qualified of all
will be a veterinarian with an interest or additional training in behavioural man-
agement.
The starting point for every behavioural consultation is the taking of a com-
prehensive history that reflects both the events that have affected the dog, also the
owner, his or her home setting, lifestyle, etc. Is there a unanimous view within the
family about the dog’s behaviour or is the pet a focus for conflict and argument
between family members?
The setting chosen for preliminary interview needs to be one where the
owner is relaxed and not subject to interruptions from phone calls, etc. This may
usefully be at the owner’s home, alternatively in an informal consultation room
where the dog can be observed unleashed. Ideally, this will not be a place where
the dog has previously had adverse experiences, i.e. probably not in a ‘normal’
veterinary surgery.
Home visits have the advantage of the therapist being able to witness the
dog’s home environment and usual behaviours. Critical interactions between dog
and family probably only occur in the home. Disadvantages are that the therapist
cannot call upon help from colleagues, won’t have access to specialized equip-
ment and has to commit time and expense to travelling. However, clever use of
video film, comprehensive questionnaires and an informal, varied environ-
ment (the author works from a farm) makes the ‘clinic’ consultation a good first
step that can be followed by a home visit if further information or assistance is
needed.
To be effective, behaviour therapists must be both skilful in eliciting infor-

mation, have good listening skills and the ability to resolve issues and arguments
within families. The behaviour counsellor is not there to pass judgement, rather
to address a problem of enormous importance to the dog owner. He or she may
also have an important impact upon the dog’s welfare and enjoyment of life and
so may need to be explicit about consequences for a pet if improvements are not
seen. Many owners feel self-conscious that they have had to seek professional help
at all about their pet’s behaviour, and are very worried that there might be an
adverse outcome (e.g. that there may be a recommendation that the dog be
rehomed, or even euthanased).
In the USA (where distances to be travelled are much greater than in most
European countries) a telephone interview or consultation is sometimes provided
by, say, university veterinary undergraduates. Alternatively dogs are sent away
from the owner to be retrained at a specialized kennel facility. The writer is
dubious about either approach, given that canine behavioural problems are
mostly failures of relationships between dogs and people. The telephone will
probably not reveal the relevant social and behavioural dynamics, and there
should be total involvement of the owner in all aspects of behaviour modification.
The time taken to fully understand the ontogeny and consequences of an
unwanted behaviour can require a surprisingly small investment of time by the
therapist. In the writer’s practice, most cases require no more than a 2-h consul-
tation, with follow-up by phone, e-mail and possibly a second visit. At my Animal
Behaviour Centre, we expect an 80–90% problem resolution rate as defined by
the criterion of owners being satisfied with the advice given, and comparable rates
are described by other professionals in the field. Such statistics imply that there
will be failures, i.e. some dogs will continue to be dangerous or distressed.
They may present chronic fear of bangs from gunfire and fireworks, which
may be impossible to predict and to retrain acceptable responses. Other dogs
present behaviours at such a high level of risk that euthanasia must be the right
option, given the behaviour practitioner’s responsibilities for safety of the wider
public.
Canine behaviour therapy is a dynamic and youthful discipline, given its
‘birth’ by Tuber et al.’s (1974) ‘modest proposal’. There are now regular interna-
tional scientific and veterinary meetings on the subject, many internet websites,
journal and book publications. The big challenge for the future lies in integrating
this young discipline with everyday education and practice of veterinarians and to
‘professionalize’ the quality, conduct and content of services provided by behav-
iour therapists who do not have veterinary qualifications.
Methods
Most professions have agreed, tried and tested approaches, for example to cardiac
surgery, dentistry or even car repair. However, the practice of animal be-
haviour therapy is remarkable for the diversity of methods and their underlying
234 R.A. Mugford
theoretical assumptions. In the world of behaviour therapy, it is a rarity for there

to have been systematic study or comparison of the effectiveness of procedure X
versus Y. Rather, each case tends to be treated as its own control, and the bench-
mark of effectiveness is usually reduction in frequency or intensity of some behav-
iour compared to pre-treatment levels. Most practitioners acknowledge that there
is a need for more objective, systematic studies of the effectiveness of methodolo-
gies, but understand the difficulties given the unique and interactive nature of
animal behavioural problems. I will here outline some of the key methodological
approaches that are employed in canine behaviour modification.
Instrumental learning
Instrumental learning is one aspect of Skinnerian theory (named after B.F.
Skinner) and is also described as operant conditioning. Instrumental learning
occurs whenever a dog’s behaviour affects its environment in such a way as to
have obvious and meaningful consequences from the dog’s point of view. If such
consequences are favourable, the behaviour which produced them will be
rewarded, reinforced and be more likely to be repeated in the future. If, on the
other hand, a behaviour has an immediately adverse outcome, it will be less likely
to be repeated in similar circumstances and is said to have been punished. The
definition of a reward, therefore, is a favourable outcome which results in an
increase in frequency of the preceding behaviour. The definition of a punishment
is an adverse outcome with a subsequent decrease in frequency of the causative
behaviour.
The ways in which a dog’s behaviour can be moulded by outcomes are
summarized in Fig. 13.3. It can be seen that the denial of a reward can be pun-
ishing (for example, withholding a food reward if a dog jumps up instead of sits),
whereas cessation of an aversive can be rewarding (as when an owner ceases to
be angry once the ‘sit’ command is obeyed).
Fig. 13.3. A summary of the main contents of training methods used to alleviate
behavioural problems – see the text for details.
The owner of a dog is merely one feature of a dog’s environment which can
affect and be affected by the way their dog behaves, a fact which frequently needs
to be emphasized during counselling. Dogs learn instrumentally all the time, not
just when a human decrees that they should during an ‘official’ training session.
Educating dogs to perform desirable behaviours, as opposed to undesirable ones,
involves the skilful management of consequences in both training and behaviour
modification. Rather than merely punishing undesirable behaviour, owners need
to be encouraged to be patient and observant; to notice and to reward all spon-
taneously performed desirable behaviours.
For either rewards or punishments to be effective, they must be directly
linked in the dog’s mind as being a consequence of their behaviour. Unpleasant
or aversive treatments of many forms are frequently applied by dog owners in
attempts to punish behaviour, but they have no impact upon subsequent per-
formance. Paradoxically, the undesired behaviour may actually increase rather
than decrease in frequency. In such cases, the aversive applied may inadvertently
reward rather than punish the dog and it can have a damaging effect upon the
relationship between dog and owner. For instance, hitting a dog can fairly be
described as ‘retribution’, but it is unlikely to guide the dog towards the goal of
better behaviour. Attention, even with the pain from a slap of hand, is unlikely to
be seen as rewarding.
Fundamental to instrumental learning theory is the notion of schedules of
reinforcement, be they continuous, intermittent, partial and so on. A behaviour
which is only occasionally rewarded is less likely to cease or be extinguished. On
the other hand, continuously reinforced behaviours are rapidly acquired but
liable to quickly stop when rewards stop.
Chaining or linking between different responses is another concept drawn
from operant conditioning and which is a useful notion in dog training. Complex
sequences of behaviour can be brought together for performance of integrated
tasks, such as when training a dog that assists people with disabilities to, say,
empty a dishwasher or remove items from supermarket shelves.
Importation of learning theory into dog training has had many beneficial
effects, especially clicker training. Clicker training was first employed for the
better training and precise targeting of behaviours in dolphins. Provision of a dis-
tinct click sound associated with the positive experience of being given a food
reward enabled precise movements, gestures or behaviours to be chosen and rein-
forced by the trainer, and consolidated into the animal’s repertoire.
Clicker training in the world of dogs has become a major activity, both amongst
performance dogs used in film, in specialist tasks such as explosives detection and
amongst pet owners who just want their dog to perform tricks. It can also be used to
‘shape’ desirable behavioural patterns in dogs undergoing therapy. For instance, tol-
erant, non-threatening signals in a dog that is usually intolerant of other dogs.
Obedience training
Obedience training is a practical application of instrumental learning, where
management or control of a dog’s behaviour by voice and other signals is a key
236 R.A. Mugford
part of the behaviour therapist’s tool kit. A dog’s only problem may be that it
chases and kills farm livestock, but one approach to managing such a dog is to
ensure that it has a reliable recall response. Similarly, a dog which attacks people
as they enter his owner’s home may be trained to lie ‘down’ and stay in sight but
out of reach of the visitor. By doing so, one creates a window of opportunity for
the guest to make positive gestures towards the dog, such as to offer titbits.
The world of obedience training is very idiosyncratic, but has improved con-
siderably from 19th and 20th century ideas that dogs should be dominated and
made to obey. Writers and trainers from the last century such as Konrad Most,
Koeler or Barbara Woodhouse harshly punished clients’ dogs on choke chains,
but such physical abuse is now generally considered to be distinctly out of fashion.
Instead, positive, reward-based methods are more favoured. However, this writer
is of the opinion that the move towards wholly reward-based methodologies may
have gone too far, and something in between the two extremes is better. But if an
inappropriate behaviour is to be punished, what is an acceptable and what an
unacceptable form of punishment?
Conditioning, both positive and negative, produces distinctive emotional
responses in dogs which can have profound and long-lasting effects upon behav-
iour. A potent punishment for dogs is to remove affection or attention of the
owner. The reverse strategy of increased attention is a potent reinforcer of behav-
iour. Finding the optimum balance between the two provides the best avenue for
behaviour modification.
Endocrine management
Castration and ovario-hysterectomy of dogs and bitches is the most widely per-
formed veterinary procedure in the UK and North America. However, it is rela-
tively rarely offered by the veterinary community in Scandinavian countries.
Considering the popularity of spay-neutering dogs, few studies have been con-
ducted to explore their effects upon behaviour, good or bad.
Comparisons with other farm animals, such as castrated stallions and bulls,
predict a worthwhile reduction in aggressiveness of dogs which have been
neutered. There is some anecdotal support for this view, with reduced fighting of
other male dogs. However, many bitches also fight and many male dogs continue
to present aggressive signs after they have been castrated.
An obvious benefit to neutering bitches is that their liability to subsequent
life-threatening diseases such as pyometra (an infection of the uterus) and
mammary gland tumours are markedly reduced if the procedure is conducted
before the first season. Spaying of bitches is also a worthwhile management pro-
cedure where unwanted behaviours occur at particular stages of the bitch’s
normal endocrine cycle. Pseudopregnancy (elevated progesterone and prolactin
titres) can in some bitches produce distinctive, unwanted behavioural changes,
such as hoarding of soft toys and guarding of their beds. The recommended
approach then is to interrupt the pseudopregnant state with drugs such as
cabergoline (Galastop), and to spay 2–3 months after the cessation of pseudo-
pregnant signs. However, to spay a bitch during her pseudopregnant state is not
recommended and can cause unwanted behavioural signs to persist (Lindley

et al.).
Pain management
Pain has a major influence upon mood or temperament of dogs, just as it does on
people. Lindley (1997) estimates that one-third of dogs referred to her behaviour
practice exhibit local or diffuse signs of pain. Clinical causes of such pain are
many and varied: hip dysplasia, arthritis, intervertebral disc disease, chronic
otitis, untreated dental or gum disease and so on.
The prevalence of such behaviour–clinical interactions in dogs presenting
unwanted behaviours makes a clinical or veterinary training of the behaviour
therapist desirable, and in many cases essential if the best is to be done for the
patient. Too many dogs are misdiagnosed as ‘dominant’ or ‘wilful’ because of a
failure to make a proper clinical diagnosis (see Hedhammer and Hultin-
Jäderlund, Chapter 14 in this volume).
Elimination of pain is an ideal which may not always be achievable.
Introduction of NSAIDs have greatly improved prospects for pain reduction.
There are also alternative therapies such as acupuncture, magnetic therapy, phys-
iotherapy, Tellington-Touch and so on; all approaches which address the under-
lying cause to disturbed behaviour rather than just focus upon its symptoms.
There may also be indirect but nevertheless significant effects of illness upon
the pet–owner relationship. Nursing or care for the injured or otherwise sick
animal alters the dynamic between the two: excessive, mistimed sympathy may
have profound, unintended consequences.
Dietary management
Interaction between diet and behaviour is a controversial and sometimes eccen-
tric area of opinionated dogma, without the benefit of scientific investigation.
Nevertheless, there is little doubt that some dogs dramatically change their tem-
perament or mood when shifted from one diet to another. In the author’s prac-
tice, it is usual to shift a dog from its commercial or prepared rations to a simple
two-ingredient ‘elimination diet’ such as rice and turkey or lamb and potato, for
a test period of say, one week. This approach is very like that employed in der-
matology and other veterinary specialties where diet or dietary allergies are
believed to be an important factor.
Responses of dogs to particular ingredients of either a home prepared or
commercial diet tend to be unique to that individual. Certain breeds (e.g. the
golden retriever) seem to be more prone to these dietary upsets than others,
perhaps reflecting their genetically inbred status and lowered immune compe-
tence. It is interesting that golden retrievers are also prone to a host of other
immune-mediated diseases, including lymphosarcoma and widespread, chronic
eczematous rashes.
The feeding of raw food to dogs (BARF) is becoming increasingly popular
worldwide, allegedly improving both the physical and behavioural state of the
animals (Billinghurst, nd). However, it is not clear whether the claimed benefits of
238 R.A. Mugford
raw diets are because they are indeed raw, or that various potentially harmful
additives, colorants, non-digestible gelling agents, etc. are absent.
Where diet is the factor determining a dog’s mood, temperament or behav-
iour, the effects of changing the diet can be remarkably rapid. Improvements may
occur within 24 hours and further improvements continue over time. However,
such dogs may still present a liability, evidenced by one golden retriever which
was referred to the author’s practice. An elimination diet of rice–mutton changed
this dog from being a moody, depressed and intermittently aggressive animal into
a normal, relaxed and friendly golden retriever. However, some months after the
dietary change it gained access to the household garbage bin and dramatically
reverted to being a dangerous dog that savagely attacked a 16-year-old
babysitter.
My overall conclusion is that dog owners are well advised to read ingredient
panels of commercial dog foods. From time to time, they should change brands
or feeding schemes and observe the behavioural outcomes. Only a minority of
dogs may show beneficial responses to dietary changes, but for those individuals
it is as important as is the removal of tartrazine and other colorants from the diets
of reactive children.
Physical aids/equipment
Mention has already been made that evolution did not necessarily equip dogs to
be made captive on a collar and lead. As one would expect, there has been a host
of devices to aid the bodily management of dogs, particularly to stop them
pulling.
Choke chains exert substantial compressive forces upon the neck as a whole,
with consequent damage to soft tissues of the throat and, if used roughly, to the
integrity of the cervical spine (neck). Misplacement at the junction of the skull to
the spine (the atlanto-axial joint) is a common sign of whiplash injuries to dogs
induced by over-boisterous handling.
Collars with inward-facing spikes were devised in Germany in the last
century, and have become popular in North America as a means of preventing
dogs pulling. The danger of injury to the dog’s throat is obvious and the author
has encountered one Weimaraner bitch whose trachea was actually punctured by
such a device. These spike collars unfortunately do ‘work’ (i.e. they discourage
pulling), but for the wrong reasons!
Designers of better dog equipment often find inspiration in handling tech-
niques and equipment devised for, say, horses or even for the management of
recalcitrant wild animals in zoos. Headcollars were one such innovation, intro-
duced as a standard item for pulling dogs approximately 20 years ago. Numerous
designs are now available, some acting as a simple steering device (e.g. the Halti
and Gentle Leader), others by exerting compressive forces upon the throat or
dorsal neck (e.g. Canny Collar and Nu-trix).
Overall, headcollars are very effective aids to reducing pulling by giving a
greater degree of physical control over the direction and movement of the dog’s
body. Simply put, the body follows where the head is led.
The disadvantage of all headcollars is that they restrict free movement and
exchange of facial–body signals by the dog. Members of the public may mistak-
enly believe that a dog wearing a headcollar is muzzled and so might be danger-
ous. These problems are overcome by using harnesses, which exist in a
bewildering variety of designs. The advantage of most harnesses is that they
spread the forces generated by pulling over a wide area of the body and, if to the
chest, to a strong part of the body. Plainly, some designs encourage pulling, such
as would be beneficial to tracking or sledge dogs. Other designs of harnesses place
local compressive forces on the dog’s axillae (i.e. beneath the forelegs). This is a
painful effect that might cause bruising, even damage, to sensitive skin and to the
underlying brachial nerve complex.
A more recent innovation in harness designs are front control models, where
the dog is tethered or led from the chest (sternal region). Such designs have all the
power steering advantages of a headcollar, but without the discomfort and restric-
tions previously mentioned.
Muzzles are another useful accessory for the management of dogs, simply
preventing biting, self-mutilatory acral licks or ingestion of undesirable material
(e.g. coprophagia). Muzzles that are to be worn for long periods of time should be
of a size and design that permits panting, i.e. not restrict mouth gape. The fabric
styles of muzzle with an open front restrict gape and so are only suitable to be
worn for brief periods, such as during a veterinary examination.
Electronic training aids

There is considerable controversy concerning the use of devices which adminis-
ter painful or aversive electric shock. There is wide country to country variation
in their frequency of use for training dogs, indeed in their legality. Such devices
are not permitted by law in Switzerland, Denmark and other countries, yet they
can be freely purchased in the USA and some European countries, including the
UK.
The advantage of say, radio-controlled electric shock collars is that the
trainer can deliver a painful stimulus at the moment the dog is engaging in some
inappropriate behaviour, such as chasing farm livestock. Difficulties arise when
the shock intensity is poorly matched to the pain sensitivity of the dog concerned.
Too painful and the dog will form superstitious fears about being in a field or near
trees where it received the shock. Too low and the dog will continue to engage in
highly motivated chase behaviour and will be desensitized to the painful stimulus.
More sophisticated paradigms might usefully be introduced to improve the
effectiveness of such radio-controlled shock devices: pre-training the animal to
respond to both a shock and a pre-shock tone in situations that are less exciting
or less distracting. By an improved overall level of obedience, the trainer stands a
better chance of interrupting such inherently rewarding behaviours as hunting.
The controversy surrounding use of shock collars continues to be divisive in
the veterinary and behaviour therapy community, especially as it can now be well
argued that there are effective alternatives. Dog training by the reward-based
methods and including clicker training paradigms have made training more
240 R.A. Mugford
predictable in its outcome. Long distance interruption of unwanted behaviour

can now be achieved using a gas release device, sometimes referred to as a cit-
ronella collar. These are designed to release a small interruptive spray of gas from
a cylinder mounted under the dog’s collar, perhaps at the moment it barks, steals
food or chases cyclists.
Studies by Juarbe-Diaz and Houpt (1996) indicate that the effectiveness of
the gas-training systems is greater than electric devices for interruption of
unwanted barking, without the animal welfare objections of shock collar use.
Innovation in design and imagination characterize the use of equipment for
behaviour modification. On the one hand, many dog owners want ‘quick fixes’
and have neither the skill nor the inclination to apply complex behaviour modifi-
cation or retraining regimes. Such clients usually take well to gadgets and are
grateful for technological fixes. However, significant numbers of owners of
problem pets take a more spiritual approach which recognizes the need to under-
stand the dog’s fundamental needs and emotions, and are willing to acknowledge
that there must be changes in the relationship.
Pheromones
In the last decade, there has been an interesting area of applied research led by
French veterinarian Patrice Pageat (1999). He has investigated the composition of
sebaceous secretions of the lactating bitch and subsequent responsiveness of adult
dogs to a synthetic analogue of the same secretions. Pageat named this ‘dog
appeasing pheromone’, on the basis that it seems to reduce anxiety or fear states
in dogs. The product DAP has been marketed worldwide with persuasive clinical
trials to establish efficacy for treating both chronic and acute anxiety states. It
seems to be particularly helpful in the management of fear when the dog is con-
fined to the house, where the DAP can be contained and reach a higher concen-
tration than outdoors. A collar-version of DAP is now available in certain
countries for outdoor applications.
The question arises whether this is a genuine pheromonal effect, implying
that the dog’s reactions are inborn or biologically determined rather than learned.
There are numerous definitions of the term pheromone (see Doty, 2004), and
there is controversy about its implied invariable or reflexive effects, even in the
world of insects. This writer believes that the DAP effect (which is genuine though
modest) is simply a reminder of how things were when the puppy was contentedly
suckling his or her mother. In this respect, it is like the powerful reminiscing effect
of perfumes and other odours which evoke distinctive memories in humans.
Environmental enrichment
Professor John Webster (1994) introduced the concept of five fundamental rights
or needs of animals as freedom from hunger, thirst, discomfort, fear and freedom
to mix with their own kind. It will be apparent to the reader that many dogs are
exposed to fear, are prevented from mixing with other dogs and indeed are made
to spend long periods in stressful social isolation. In the world of farm and zoo
animals, the concept of environmental enrichment is popular, where the physical
and temporal world of the animal is made more interesting, more varied and less
predictable. For instance, primates may have to use a tool in order to extract ter-
mites from a hole in wood, or pigs eat their food over an extended period by
having food dispensers or food offerings in frequent short meals rather than a
single, large gorged meal.
In the world of dogs, devices that similarly slow down the rate of consump-
tion of food, or force the dog to work for his food, have been introduced. One
such is the Buster Cube (Kruuse DK); another the rubber Kong toy, which may
be stuffed with food. The significance of these devices is greater where a contem-
porary lifestyle leads to dogs being left alone for very long periods (see earlier). In
addition, modern, processed foods can be eaten very rapidly by the dog, denying
it the opportunity to engage in predatory/hunting activities and slow mastication
of a varied diet.
The most important feature of a dog’s welfare is that it has the company
of other dogs. Where the keeping of pairs or trios of dogs is unaffordable or
not practical, a desirable alternative is that the pet be taken to a dog day care centre,
which is a popular option in North America, though still unusual in Europe.
Concluding Remarks
Most behavioural problems have their origin in actions by the owner, by the
owner’s habits and by the environment which they provide for the dog. A few
problems arise due to medical or genetic factors. There has been a marked
improvement in the outlook for the treatment of behavioural problems in dogs
due to an explosion of knowledge about canine social behaviour, also due to the
increased sensitivity of veterinarians in practice to the importance of canine
behaviour. The breakdown rate of relationships between man and dog varies
between countries, but in the UK is of the order 10% per dog lifetime. This sug-
gests that 500,000–700,000 dogs in the total population of approximately 7–8
million will have been rehomed, many of them for behavioural reasons. The
man–dog relationship is, however, more secure and longer lasting than the insti-
tution of marriage, which has an only 50% ‘survival’ rate in many Western coun-
tries. Despite, or perhaps because of, their misbehaviours, dogs evoke remarkable
loyalty from their people!
References
Billinghurst, I. (nd) drianbillinghurst.com
Doty, R. (2003) Handbook of Olfaction and Gustation. Marcel Dekker, New York,
p. 345.
Harvey, M.J.A., Cauvin, M., Dale, M., Lindley, S. and Ballabio, R. (1997) Journal of
Small Animal Practice 38, 336–339.
242 R.A. Mugford
Juarbe-Diaz, S. and Houpt, K. (1996) Comparison of two antibarking collars for

treatment of nuisance barking. Journal of the American Animal Hospital
Association 32, 231–235.
Mugford, R.A. (1987) The influence of nutrition on canine behaviour. Journal of
Small Animal Practice 28, 1046–1055.
Overall, K.L. (1997) Clinical Behavioral Medicine for Small Animals. Mosby,
Missouri.
Pageat, P. (1999) Attachment and pheromones in the dog. Proceedings of 2nd World
Vet Meeting on Veterinary Ethology.
Polsky, R.H. (1994) Electronic shock collars: are they worth the risks? Journal of the
American Animal Hospital Association 30, 463–468.
Pryor, K. (1975) Lads Before the Wind. Sunshine Books, Waltham, Massachusetts
(revised 2000).
Shepherd, K. (2002) Development of behaviour: social behaviour and communica-
tion in dogs. In: BSAVA Manual of Canine and Feline Behavioural Medicine.
BSAVA, Gloucester.
Sheppard, G. and Mills, D.S. (2003) Evaluation of dog-appeasing pheromone as a
potential treatment for dogs fearful of fireworks. Veterinary Record 152,
432–436.
Sigler, L. (1991) Pet behavioral problems present opportunities for practitioners.
American Animal Hospital Association Trends 4, 44–45.
Stead, A.C. (1982) Euthanasia in the dog and cat. Journal of Small Animal Practice
23, 37–43.
Tuber, O.S., Hothersall, O. and Voith, V.L. (1974) Animal clinical psychology: a
modest proposal. American Psychologist 29, 762–766.
Webster, J. (1994) Animal Welfare: A Cool Eye Towards Eden. Blackwell Science,
Oxford, UK.
Behaviour and Disease in Dogs
14
Åke Hedhammar and Karin Hultin-Jäderlund
Introduction
Interactions between behaviour and disease have several dimensions. In many

cases where dogs have behaviour alterations, the behaviour is integrated with
signs of a physical disease.
By definition, specified behavioural disorders are, as other diseases, denoted
by a name for the specific condition; a diagnosis. Such diagnoses of abnormal
behaviour are commonly symptomatic, e.g. convulsion and pica ‘just’ describing the
signs. Diagnoses at a higher hierarchical level should be more significant, func-
tional and related to aetiology, e.g. convulsions due to a known infectious agent
causing an inflammation in a certain part of the brain as in rabies, or pica due to
malnutrition by a pancreatic insufficiency due to an inherited atrophy of the pan-
creas as described in German shepherds and collies. Ideally diagnoses at a higher
hierarchical level, i.e. functional diagnoses, are defined at a molecular level as, for
example, narcolepsy due to known and sequenced genotypes.
Contrary to narcolepsy caused by single gene mutations, most behavioural
alterations supposedly are multifactorial and involve many genes as well as envi-
ronmental factors. Genes as well as environmental factors, including socialization
and training, altering behaviour are the focus of other chapters in this book. In
this chapter we will try to elucidate the interactions between behaviour and phys-
ical diseases affecting behaviour. The disease on the other hand might be geneti-
cally determined as well as affected by environmental exposure. Even
post-traumatic stress due to disease, or interventions to diagnose or treat a disease,
are examples of interactions between behaviour and disease that have practical
implications in a clinical setting.

(ed. P. Jensen)
244 Å. Hedhammar and K. Hultin-Jäderlund
Besides specified behavioural diseases in which an abnormal behaviour is

the only complaint, almost all other diseases manifest partly in behavioural
alterations. In fact, owners of a sick dog most commonly react on behavioural
rather than physical changes. When recalling the clinical history of a
case – the anamnesis – owners almost exclusively bring up changes in activity
level and changes in eating, drinking and eliminating (urinating, defecating)
behaviour.
In many diseases behavioural alterations are significant findings aiding diag-
nosis at a higher hierarchial level. For example, polyuria (pu) – excessive urina-
tion – is a common and unspecific symptomatic diagnosis, but polyuria in
combination with lethargy, polydipsia (pd) and polyphagia is a strong indication
for a more specified diagnosis – Cushing’s disease – a disease with an increased
release of glucocorticoids from the adrenal gland.
In skin diseases, the clinical sign of itch has a major importance differentiat-
ing many diagnoses from each other.
Behavioural signs are no exception to other clinical signs in that there is no
strict border between normal and abnormal. Behaviour may also be inappropri-
ate rather than abnormal in many cases.
Many diseases result in profound behavioural alterations. Pain accompany-
ing the course of several diseases may elicit aggression and anxiety to a great
extent. It must also be appreciated that pain and distress can also be related to
diagnostic procedures and treatments for many diseases.
In this chapter, we will define a terminology used in small animal veterinary
medicine that helps to understand an interaction between canine behaviour and
canine diseases. We will then bring attention to clinical signs/symptoms related
to behaviour and disease, and give examples from diseases in various organ
systems and of different aetiologies. Epilepsy, pain and signs of anxiety will be
given special attention.
The attention to canine behaviour in a clinical setting will be put into focus
and clinical signs related to specific syndromes affecting canine behaviour will be
reviewed.
Definitions
Anamnesis – recalling the history of events preceding and during the course of an
illness.
Anxiety – ‘experiencing’ a sense of dread or fear (especially of the future) of real or
imagined threat to one’s mental or physical well-being.
Cognitive dysfunction – change in interactive, elimination or navigational behav-
iours, attendant with ageing, that are explicitly not due to primary failure of any
organ system.
Convulsion – primarily a lay term. Episodes of excessive, abnormal muscle con-
tractions, usually bilateral, which may be sustained or interrupted.
Coprophagy – the eating of excrement.
Behaviour and Disease 245
Disease – literally ‘lack of ease’, a pathologic condition of the body that presents
a group of clinical signs peculiar to it (and which set the condition apart
as an abnormal entity differing from other normal or pathological body
states).
Epilepsy – a chronic neurological condition characterized by recurrent epileptic
seizures.
Epileptic seizure – manifestation(s) of epileptic (excessive and/or hypersynchro-
nous), usually self-limited activity of neurons in the brain.
Fatigue – a ‘feeling’ of tiredness or weariness resulting from continued activity.
Fear – the emotional reaction to an environmental threat.
Itch (pruritus) – an unpleasant sensation that causes an individual to rub, lick, chew
or scratch at its skin.
Listlessness – too tired to show an interest.
Narcolepsy – a disorder of the brain that is marked by sudden recurring attacks of
sleep.
Pain – an aversive sensory and emotional experience (a perception) which elicits
protective motor actions, results in learned avoidance, and may modify species-
specific traits of behaviour, including social behaviour.
Pica – a perversion of appetite with craving for substances not fit for food.
Restlessness – unable to rest.
Seizure – non-specific, paroxysmal, abnormal event of the body (cf. epileptic
seizure).
Weakness – lack of physical strength.
Clinical Signs Related to Behaviour and Disease in Dogs

Level of consciousness
Altered states of consciousness are always related to abnormal brain function.

The following states of consciousness should be taken into account:
● Normal – the dog is alert, responds to external stimuli, is aware of its surround-
ings, and responds to commands as expected.
● Depressed – the dog responds slowly or inappropriately to verbal stimuli. Some
animals may appear disoriented or even act delirious. Most sick animals are
depressed.
● Stuporous or semi-comatose – the dog is generally unresponsive and appears to be
asleep, but can be aroused by strong stimulation, especially pain.
● Comatose – the dog is unconscious and no behavioural reactions are possible to
elicit by any stimulus.
A dog that is semi-comatose or comatose most commonly has either a disease
process or a lesion of traumatic origin affecting the brain stem. A dog in that state
might also be affected by some toxin or by any metabolic disease secondarily
affecting the brain. Less commonly, a semi-comatose or comatose dog has a
diffuse, bilateral cerebral disease.
Narcolepsy
Narcolepsy (sleep attacks) with concomitant cataplexy (loss of muscle tone in all
skeletal muscles except muscles needed for breathing) is an uncommon clinical
sign occurring in dogs. The signs in themselves come from an imbalance in the
sleep–awake neurotransmitter system in the forebrain and brain stem. In the dog
population, narcolepsy with cataplexy occurs both as a genetic disease and as a
sporadic acquired disease. Even though the clinical signs in different dogs suffer-
ing from different forms of the disease are undistinguishable from each other, the
cause of imbalance in the affected neurotransmitter system seems to differ. In the
hereditary genetic form, the receptors on the postsynaptic nerve cells are mutated
and malfunctioning, while in the acquired form, the production of the neuro-
transmitter substance seems to be too low.
Weakness – tiredness – fatigue – listlessness
These are signs most commonly related to physical diseases in dogs. Weakness is
a sign commonly encountered in small animal veterinary practice. It is termed
episodic when elicited by exercise and dissipated by rest. Although weakness
unrelated to exercise is an appreciated clinical sign in human psychiatry, it is not
easy to reveal such aetiology in canine weakness. As several metabolic, endocrine,
cardiovascular and neuromuscular diseases may result in weakness, work up of a
weak dog calls for an evaluation of electrolyte and hormonal balances as well as
cardiovascular and neuromuscular function. Adrenal insufficiency (Addison’s
disease), hypothyroidism and Myasthenia gravis are examples of differential diag-
noses in a dog that develops unexpected severe weakness at almost any age.
Weakness at an older age could also be due to degenerative processes, including
canine cognitive dysfunction.
Anxiety – fear – avoidance – flight
These signs might be caused by injuries and physical diseases. Although anxiety
and fear are commonly seen as inherent behavioural manifestations in many
dogs, it must be appreciated that injuries and physical diseases may cause or per-
petuate signs of anxiety and fear. Avoidance and flight in a dog that has not
exhibited such signs before indicate a possibility that an ongoing disease process
or current or earlier exposure to physical harm might be involved. Any trau-
matic injury, causing pain somewhere in the body, may provoke anxiety and
fear. But also to ‘experience’ an epileptic seizure or loss of breath is capable of
inducing anxiety and fear. Anxiety is also described as an early clinical sign in
lysosomal storage diseases such as fucosidosis, an inherited disease in English
springer spaniels, for example. Relief of pain before and after operative proce-
dures has received increased attention in veterinary medicine. In order to recover
fast and completely, pain relief is almost as important as anaesthesia during

surgery.
Aggression
Aggressive behaviour might also arise out of external ‘stimuli’. Misdirected

aggression cannot be excused but rather explained by painful processes that the
dog may either relate to a person or to a situation. Clinical situations and veteri-
nary visits might either be so harmful that aggression is warranted or just be a sit-
uation in which a dog with strong dominance aggression ‘tests’ his position.
Sudden and strong pain may elicit biting that is an expression of avoidance rather
than aggression. Relief of pain is as important for reduction of misdirected aggres-
sion as for anxiety.
Epileptic seizures and epilepsy
Epilepsy (recurring epileptic seizures) is one of the most common symptomatic

diagnoses arrived at in dogs in veterinary practice, even though the veterinarian
seldom observes any clinical signs in a dog with epilepsy. Attacks described by the
owner lasting a couple of minutes most often are easily recognized as epileptic
seizures, especially when general muscle activity (cramping) appears in the
seizure. Another feature most often recognized in conjunction with an epileptic
seizure is the so-called postictal period, a transient clinical abnormality of central
nervous system function that appears when clinical signs of the epileptic seizure
have ended. The postictal phase may last for a few minutes to several hours,
during which the dog may be restless, lethargic, confused, disoriented, aggressive
and/or blind. A problem in reaching a symptomatic diagnosis in some dogs might
be when the muscle activity is not generalized, when the dog is conscious during
the seizure, when there is no postictal phase and/or when the seizure is mainly
sensory.
Definitions about epileptic phenomena are all in accordance with the glossary
from the International League Against Epilepsy, an international commission syn-
chronizing the platform on which knowledge about epilepsy in humans is gath-
ered. Since the inherent neurophysiological and neuroanatomical bases for
epileptic seizures are similar in all mammals, comparative aspects are useful when
dealing with epileptic dogs. However, in human medicine, electroencephalogra-
phy (EEG), the graphic recording of the spontaneous electrical activity of the
brain, has a vital importance in patients that have experienced seizures. EEG-
waves from neurons discharging during an epileptic seizure have typical, abnor-
mal patterns. In the clinical setting, this might be achieved by constant
EEG-registration with simultaneous video-registration of the patient’s behaviour
until a seizure has been experienced. However, in dogs this technique has major
concerns in a clinical setting, e.g. due to artefacts from general muscle activity.
EEG-registration from dogs experiencing seizures have been performed at some

university clinics, with the animal anaesthetized or sedated to prevent muscle arte-
facts – but at the same time affecting the neuronal activity – and for shorter periods
(not expecting a seizure during the registration). Sensitivity and specificity of EEG-
abnormalities during those circumstances have, so far, not been very high.
In cases where dogs have had seizures, not easily recognized as epileptic
seizures from the owner’s history or from looking at videos of the event, a simple
test confirming the event as epileptic is still lacking. Examples of events that could
be mixed up with epileptic seizures of different kinds are syncopes, sleep attacks
in narcolepsy, and compulsive disorders (see also Chapter 13 in this volume).
An epileptic seizure is always a sign, either of a low or a decreased seizure
threshold, or of a lesion present in the cerebral cortex, containing hyperactive
nerve cells. The term ‘seizure threshold’ means the resistance in the nervous
system against developing seizure activity. When seizure threshold is critically
low, an epileptic seizure occurs. The level of the seizure threshold varies
interindividually as well as with age. Precipitating factors temporarily lowering
the threshold could include elevated body temperature.
Epileptic seizures occurring in a dog should initially be looked upon as unspe-
cific signs of an underlying disease. Epileptic seizures are elicited from neurons in
the grey substance of the brain. All kinds of diseases in the forebrain (e.g. inflam-
mations, tumours, anomalies, traumatic injuries) might give rise to epileptic
seizures. Also diseases of internal organs, or intoxications, altering the metabolic
state in the body and thereby the brain function, may present as seizures.
Common examples of that include liver diseases, hypoglycaemic conditions and
electrolyte disturbances. Conventional blood and urine analyses are often suffi-
cient to confirm these kinds of underlying diseases.
In some cases, no matter how extensive examinations are performed with
various diagnostic aids, no metabolic changes are reflected in blood or urine and
no morphologic or pathologic changes can be detected in the brain. The affected
dog is then said to have a ‘low seizure threshold’. The ultimate diagnosis often
used for these cases is ‘idiopathic epilepsy’. The threshold for seizures may be an
inherited trait. In some breeds, inherited forms of epilepsy are known to occur
(e.g. golden retriever, labrador retriever, Bernese mountain dog, keeshond,
Belgian tervueren, Shetland sheepdog and vizsla). However, there are also spo-
radic cases of idiopathic epilepsy that could appear in dogs of any breed in any
breeding line.
In an effort to reach an aetiologic diagnosis in a dog with epileptic seizures,
the practising veterinarian has to act as a detective. The work-up should include
a thorough anamnesis, a clinical and a neurological examination of the patient.
Blood and urine samples should be analysed regarding possible
metabolic/endocrinologic/toxic aetiologies. Further investigations might be indi-
cated for certain cases, where magnetic resonance imaging (MRI) of the brain
and liquor tap analysis are especially valuable diagnostic aids. In the future,
molecular genetic analyses regarding different kinds of inherited epilepsy among
affected breeds hopefully will come in place.
Pruritus (itch)
Itch is a clinical sign that is most commonly related to skin diseases by external
stimuli such as ectoparasites or immunological conditions such as atopic dermati-
tis. At the far end of a scale describing intensity of itch, self-mutilation may occur.
Self-mutilation might be associated with a sensory neuropathy seen as an inher-
ited condition in pointers and longhaired dachshunds at a young age, where
sensory nerve cells are malfunctioning, probably causing paresthesias (Gnirs and
Prelaud, 2005). Acquired sensory neuropathies with self-mutilation can also be
seen in any breed due to inflammatory disease of the cranial and spinal ganglia
and dorsal nerve roots (canine ganglioradiculitis) or by ingesting infected meat
from pigs with Aujeskies disease (pseudorabies). The latter is a subacute viral
brain infection and the former has a more chronic manifestation with Siberian
huskies as an example of an over-represented breed. Also peripheral nerve
tumours and the so-called paraneoplastic syndrome may result in peripheral
sensory neuropathy and self-mutilation. Intensive licking, for whatever reason,
may result in a skin disease named acral lick dermatosis.
Pain
Pain undoubtedly affects behaviour. Since pain is included in the clinical signs of
a wide variety of different diseases and injuries in dogs, it is an important clinical
sign to recognize – and to relieve. The main categories of painful processes in
dogs emanate from localized inflammatory reactions, traumatic injuries, disc-
associated diseases or some neoplastic diseases.
Some different patterns of behaviour might be recognized for a dog in pain:
● A stiffening of its body posture, a whine or a growl, a turning of the head

towards the stimulus, or a bite when we touch a certain part of its body, could
be the behaviour telling us that a dog perceives pain from being touched there.
These kinds of reactions are seen in cases such as localized subcutaneous
abscesses. One could also elicit these reactions by causing pain in an unaffected
part of a body, e.g. by giving an injection.
● A dog also avoids painful movements, if possible, of the body part involved in
pain. This could be noticed as a lameness of a fractured limb, or chewing of all
food on one side of the mouth when a tooth root on the other side is inflamed.
● A dog that is reluctant to move, stiff while having to move, avoiding certain
body positions, screaming spontaneously while changing posture and maybe
also screaming out loud as soon as any part of its body is touched, has behav-
iours that should be interpreted as emanating from severe pain. The painful
disease or injury is then often found somewhere in the trunk and especially in
the spine. In the authors’ experience, the diagnoses most often found in dogs
showing really severe pain are disc prolapses, meningitis, skeletal malignancies
or panosteitis (a painful disease process in the skeleton during growth).
Even though a dog in pain might bite a person touching (provoking) it,
unprovoked aggressiveness in dogs is most often not related to painful diseases or
injuries.
Dogs that experience a sudden paralysis of their hind limbs due to some non-
painful event, e.g. an infarct in the spinal cord, can be very restless and anxious
the first days after that event, constantly moving around and changing postures.
This behaviour is not typical for a dog in pain, but is sometimes confused with
that.
Appetite
Normal and even ‘good’ appetite is a clinical sign accompanying good health.
Decreased appetite in a dog normally eating well is indicative of disease and/or
unease. Organic diseases of various aetiologies have an impact on appetite.
Whether decreased appetite can be related to mental depression in dogs is quite
difficult to evaluate. Many dog owners report decreased appetite in circumstances
when other dogs or members of a household are absent.
Conditions similar to anorexia nervosa have not been proven to occur in
dogs, although lack of appetite for no obvious reason has been noted in sledge
dogs, for example. Ravenous appetite can be elicited by endogenous cortico-
steroids as in Cushing’s syndrome and by exogenous supply of corticosteroids in
treatment protocols for immunological diseases. In those cases it is also accompa-
nied by polydipsia and polyuria.
Inappropriate urination
Urination in inappropriate environments can be due to diseases resulting in an

increased urinary volume (polyuria), congenital defects, inflammations or diffi-
culties controlling bladder function (incontinence). ‘Voluntary’ voiding of urine at
inappropriate places in housebroken dogs is a behaviour that has to be differen-
tiated from involuntary voiding of urine in a dog with urinary tract disease, exces-
sive urinary production or neurological disease.
The first step is to find out whether it is related to an increased urinary
volume – polyuria (pu). Polyuria is most commonly seen together with an
increased water intake – polydipsia (pd) – that makes it difficult for the dog to
refrain from urinating. Pd/pu can be due to congenital diseases in the liver or
kidneys in young dogs, which has to be watched for in puppies that will be unable
to be housebroken. In older dogs that have not exhibited inappropriate urination
behaviour before, there are several causes of pd/pu, e.g. chronic liver failure and
renal diseases, parathyroid tumours, lymphosarcoma, diabetes, Cushing’s syn-
drome (overproduction of corticosteroids) or pyometra.
Compulsive water drinking – pseudo-psychogenic polydipsia – also resulting
in pd/pu might arise from emotional stress and will respond to water deprivation
contrary to most diseases. A dog with Cushing’s syndrome also responds to water
deprivation but in contrast to the dog with pseudopsychogenic polydipsia, in
addition exhibits polyphagia (see below, Endocrine system).
Having ruled out pd/pu as the cause of inappropriate urination, the next step
is to rule out congenital abnormalities triggering the micturation process, such as
an ectopic ureter or pelvic bladder, or urinary tract irritation from inflammation
or bladder calculi.
Difficulties in controlling bladder function resulting in incontinence might
arise from diseases and injuries of the nervous system as well as within the lower
urinary tract. In male dogs, diseases of the prostate also might affect the mictura-
tion.
To differentiate between physical diseases and inappropriate urination for
other reasons, water intake and urinary volume should be measured. A daily
water intake between 20 and 90 ml/kg bodyweight is considered normal and a
daily urine output between 20 and 40 ml/kg bodyweight is to be expected.
Intervals and frequency of ‘normal’ urinary voids varies greatly between and
within individuals. Changes over time are often more revealing. Voiding when
left alone or at night is difficult to watch. A dog with urinary incontinence for
medical reasons usually voids urine in the bed rather than spread over the home.
Urine analyses including specific gravity, protein, glucose, cellular components
and bacteriological cultures help to differentiate between different medical causes
as do blood analyses including glucose, electrolytes and white blood cell counts.
Inappropriate defecation
Defecation at inappropriate locations is most commonly associated with organic

diseases. Diarrhoea, defined as an increased volume and water content resulting
in an increased frequency of defecations, may also result in defecation at inap-
propriate places. An expected amount of normally formed stool at an inappro-
priate place is more likely to be due to an elimination behaviour problem than
stools consistent with enteritis. Alternatively, loose stools may indicate physical or
psychological stress as that speeds up the passage of food through the digestive
tract with resultant diminished time for water reabsorption in the colon.
Diseases Affecting Behaviour in Dogs – By Organ System
Almost all diseases can affect behaviour to a variable degree, in dogs as well as in
humans. One major difference is that dogs cannot ‘worry’ about the outcome of
serious and life-threatening diseases. Nor do they have ‘insight’ of the potential of
later relief to help cope with pain and distress. As an overview, common behav-
iour alterations in particular diseases are presented below by organ system. The
overview begins with effects from diseases in the nervous system and sensory
organs.
Nervous system
Clinical signs from diseases in the nervous system include altered behaviour in
one way or another. The possible underlying categories of diseases are often, in
clinical neurology, systematized as follows:
● Vascular diseases (infarcts or haemorrhages affecting the nervous system).

● Inflammatory/infectious/immunological diseases (e.g. distemper, steroid
responsive meningitis-arteritis).
● Traumatic injuries (e.g. skull fracture).
● Anomalies (e.g. hydrocephalus).
● Metabolic/toxic/nutritional disturbances (e.g. hypoglycaemia, strychnine poi-
soning, thiamine deficiency).
● Idiopathic conditions.
● Neoplastic diseases (e.g. primary brain tumours, metastatic tumours in the
central nervous system, CNS).
● Degenerative diseases, disc-associated (e.g. disc prolapse).
● Degenerative diseases, degeneration of nerve cells (e.g. hereditary ataxia in fox
terriers).
In many of these diseases, in addition to the pure ‘behavioural’ component

(e.g. becoming more timid, or aggressive, or extremely restless, or not recogniz-
ing the owner any more), signs such as blindness, paralysis, paresis, exaggerated
movements, uncoordinated movements, tremor, in whole or part of the body
present concomitantly. Pure ‘behavioural’ components as well as seizures are
strongly correlated with pathological conditions in the forebrain.
Also diseases in the peripheral nervous system alter behaviour, even though
the brain cells are unaffected by pathology. For example, a dog with paretic jaw
muscles due to an inflammation in the motor nerves to the masticatory muscles
shows great difficulty in eating and drinking since it is unable to close its mouth.
Another example is a dog that becomes urinary incontinent due to a tumour
growth affecting nerves in the pelvic region, innervating the urinary bladder and
urethra.
Meningitis
One quite common form of meningitis in dogs, steroid-responsive meningitis-
arteritis, occurs in young adult dogs, and immune-mediated mechanisms are
strongly suspected to cause the disease. The disease seems to be more common in
some breeds, e.g. boxers and Bernese mountain dogs. Main clinical signs are neck
stiffness, spinal hyperestesia, and fever. Clinical signs are promptly relieved by
corticosteroid therapy in most cases.
Brain tumours
The clinical signs of brain tumours depend on the location of the mass inside
the brain. One common sign from tumours growing anywhere in the forebrain is
seizures. Tumours in the frontal or temporal lobe often also cause other abnor-
malities in behaviour, such as aggressiveness, extreme restlessness, confusion or
an inability to understand instructions anymore. Another behavioural sign con-
nected with diseases of the forebrain, often a brain tumour, is so-called com-
pulsive walking – an aimless continuous wandering. Brain tumours in the
occipital lobe often cause visual field deficits. Tumours in the pituitary gland
and/or hypothalamus in addition also may cause autonomic and endocrine signs:
polyuria, polydipsia, changes in eating and sleeping pattern, and so forth. Main
signs of tumours in the brain stem are gait deficits and cranial nerve signs. In later
stages of brain tumours, significant alterations in consciousness should be
expected.
Sensory organs
Deafness, blindness, inability to smell or taste, inability to perceive pain and/or

touch and dizziness due to an inability to perceive influences on the balance
organs of the inner ear are all clinical signs, occurring with certain diseases seen
in dogs and profoundly altering the behaviour of the affected dog.
Deafness
Bilateral deafness for a dog means a dependence on other sensory organs
for interactions with the surroundings. As long as the visual system in particu-
lar is intact, and the head is turned towards an interacting event, an ob-
server might miss that the dog has any deficit at all. The deficit becomes evident
as soon as the dog turns its head away from anyone trying to communicate
with it. Unilateral deafness is not always revealed clinically, but is noticeable
when the dog can hear but is unable to localize the sound. Deafness in young
dogs is most often a congenital disease, proven or suspected to be inherited in dif-
ferent affected breeds. Border collie, cocker spaniel, collie, dalmatian
and doberman pinscher are some breeds with reported congenital deafness.
When elderly dogs become deaf, a loss of cells in the auditory system should be
suspected.
Blindness
Blindness comes from either diseases in the eyes (e.g. glaucoma, retinal detach-
ment), or diseases affecting the central nervous system visual pathways (e.g.
hydrocephalus, hepatic encephalopathy, lead intoxication, inflammatory disor-
ders of the optic nerve). A dog that has slowly and progressively become blind,
and is kept in surroundings that have not changed since the disease started, might
behave in a surprisingly normal manner. But as soon as the owner moves furni-
ture around in the home, or moves to another place with the dog, the deficit
becomes evident. A dog with acute onset of blindness will show obvious behav-
iour alterations, often manifested as an abnormally cautious behaviour and
walking into objects.
Inability to smell
Inability to smell is occasionally a complaint from owners of hunting dogs. This
clinical sign could be due to infestation by Pneymonyssus caninum in the nasal cavi-
ties, for example.
Inability to perceive pain and/or touch

Inability to perceive pain and/or touch can be related to severe injuries of the
spinal cord, dorsal nerve roots or peripheral nerves. These injuries also, in almost
all cases, concomitantly lead to paralysis of the affected body part. In addition
there are some breed-related, possibly inherited, diseases with loss of function in
the sensory nerves as the pathologic condition. Complete loss of pain/skin
sensation in dogs can give rise to severe self-mutilation, e.g. eating its own tail or
toes.
Dizziness
Dizziness, from disturbances in the vestibular system (balance system) is seen
quite commonly in dogs. Signs that might be included in the resulting ‘vestibular
syndrome’ are head tilt, uncoordinated movements, jerky eye movements and
strabismus. Diseases that could be involved include deep ear inflammation, intox-
ication, thyroid disease and neoplasia.
Endocrine system
Next to altered behaviour from diseases affecting the nervous system, altered
behaviour due to diseases in the endocrine system might be the most common
and obvious to the owner. By including effects from altered levels of hormones
regulating reproduction it is a clinically highly relevant group of diseases affecting
mainly eating, drinking, elimination and sexual behaviour.
Variations in sexual hormone production and release due to castration
(orchiectomy in male dogs and spaying in bitches) are outside the scope of a rela-
tion to diseases unless performed for medical reasons. Pyometra (a manifest infec-
tion in the uterus) most commonly is cured by ovarihysterectomy, so also affecting
the production and release of oestrogens and progesterones. Chronic diseases
affecting the prostate but also perineal hernias and perineal skin diseases are
favourably affected by orchiectomy. Dogs castrated for medical reasons are most
commonly older than those castrated for non-medical reasons. An anticipated
positive effect on behaviour in dogs castrated for non-medical reasons is less pro-
nounced and not even desired when castrated for medical reasons. Spayed
bitches are more prone to urine incontinence, which has to be differentiated from
spraying. Old dogs that have had an orchiectomy will roam to a somewhat lesser
degree and dominance aggression may diminish, but effects are unpredictable
and variable.
Occasionally bitches with ovarian tumours exhibit behavioural alterations,
but rarely nymphomania. Male dogs with sertolicell tumours, usually in a retained
testicle, exhibit ‘feminization’ of their behaviour together with gynaecomastia

(enlarged mammary glands) and flaccidity of the preputium. Older chryptorchid
dogs in particular should be watched for this when exhibiting altered behaviour.
Painful processes in the prostate might give rise to withdrawal or aggression as
well as limping.
Hypothyroidism, the most common endocrinopathy in dogs, has been indicated
to cause various alterations in behaviour, including aggressiveness. There are
figures from the 1980s indicating that less than 2% of aggressive behaviour was
due to hypothyroidism, indicating that it is not as common as was earlier thought.
Lethargy and weakness are signs related to a decrease in the metabolic rate in
hypothyroidism. Whether it also might give rise to mental depression is difficult
to evaluate in dogs. Hyperthyroidism, which has been claimed to cause aggres-
siveness in dogs, is an extremely rare condition and it has not been possible to
prove this a clinical cause of aggressiveness.
Hyperadrenocorticism (Cushing’s syndrome) has a profound effect on metabolism
and includes such mental signs as listlessness. The most prominent clinical effect
on behaviour is its effect on eating and drinking. It results in polydipsia
(pd)/polyuria (pu) as well as polyphagia. It is usually a disease of older age and has
a gradual onset of clinical signs.
Hypoadrenocorticism (Addison’s Disease), on the other hand, may also occur at
quite a young age and have either a gradual and diffuse clinical onset or may
result in sudden episodic weakness due to hyperkalaemia. There are fairly strong
breed predispositions for Addison’s disease in bearded collies and standard
poodles, warranting attention to diffuse mental changes in these breeds.
However, it may be seen also in other breeds.
Diabetes mellitus is not only a cause of pd/pu and polyphagia but also muscle
weakness due to disturbed electrolyte balance and other complications.
Gastrointestinal system
Severely affected exocrine pancreas function, as can be seen in German

shepherds and rough collies with an inherited disease causing pancreas atrophy,
may have an impact causing weakness. Occasionally it may also result in
coprophagy.
Malabsorption from severe diarrhoea and protein loss ‘leaking’ from the gut
may result in similar effects on general behaviour and inappropriate eating habits.
Inflammatory bowel diseases and severe colitis of various aetiologies may ‘irritate’
to such an extent that behaviour alterations such as anxiety and aggression may
develop.
Liver diseases of various aetiologies may provoke pd/pu as well as weakness.
Hepatic encephalopathy (HE), most commonly caused by congenital portal
vein anomalies or toxicity in young dogs and hepatic cirrhosis and neoplasia in
older dogs, may provoke subtle behavioural complaints including pacing, aggres-
sion, hysteria, intermittent deafness and depression. The clinical signs in HE are
caused predominantly by increased central brain ammonia concentrations due to

urea cycle deficits.
Urogenital and reproductive system
Inflammatory processes of any part of the lower urinary tract may induce inap-
propriate elimination behaviour. Urinary calculi and enlargement of the prostate
by hyperplasia or chronic infections may affect voiding of urine.
Cardiovascular system
Heart failure most commonly seen in older dogs, and mainly caused by endocar-
dosis in small breeds and cardiomyopathy in larger dogs, can affect behaviour by
causing unprovoked tiredness. In young dogs similar effects could be due to con-
genital heart defects such as aortic stenosis.
Canine behaviour in a clinical situation
Stress at a veterinary clinic might not only affect behaviour during the visit but
could result in behavioural alterations in similar situations later on. It is not
uncommon that shyness and even aggressiveness towards people is blamed on
unprofessional handling at a veterinary clinic.
It could well be that stress by restraint or painful events at a veterinary visit
is the first time a puppy has experienced such an event and is not necessarily the
cause of similar behaviour later on. But, being a stressful event to many dogs (and
their owners), veterinary visits should be organized in as positive a manner as pos-
sible.
Proper examination of an individual sometimes calls for restraint, and proper
supervision could result in isolation in a ward very much different from what a
dog is used to at home. Both circumstances could be frightening, whether associ-
ated with pain or not.
Surgery, even under general anaesthesia, can significantly affect any individ-
ual as do most medications whether by intention or through side-effects. Keeping
that in mind helps staff at veterinary clinics to avoid, as much as possible, post-
traumatic stress disorders in their patients.
Panic is evidenced by very strong immediate reactions such as increased
heart rate, salivation, trembling, flushing skin and a shortness of breath. However,
fainting is rare in dogs suffering panic.
By optimal arrangements all the way through the reception area, waiting
room, examination rooms, surgical theatres and wards, some stress could be
relieved for patients as well as their owners. Space and noise are parameters to
work with. Of greatest importance is handling by the personnel involved. A firm
but gentle handling and a positive voice in all situations help anxious as well as
aggressive dogs to cope with procedures that have to be performed (Fig. 14.1).
It might also be pertinent to question whether a procedure has to be per-
formed or not – e.g. to measure rectal temperature in a dog acutely injured in the
rectum might not be necessary or even relevant.
Whenever possible, and without interfering too much with the clinical find-
ings and outcome, pain relief should be administered. In veterinary medicine as
well as in human medicine, it is now well appreciated how beneficial that is. Pre
and post medications to surgery are today good practice in our small animal hos-
pitals.
Consultancies at veterinary clinics with behaviour as chief complaint
Behaviour is no exception to the value of a complete and thorough clinical

history. By age, breed and sex of a patient, many clues can be reached and it does
often help to limit the clinical work-up procedure. A dizzy young chihuahua
puppy is much more likely to have a hydrocephalus as the cause than an aged St
Bernard. A middle-aged standard poodle is on the other hand more likely to show
episodic weakness by Addison’s disease than a very young puppy of any breed.
Complete demographic information also contains information on family
members and other pets in the household.
The clinical history – an anamnesis – should include not only circumstances
around the chief complaint such as: For how long? Gradual or sudden onset?
When in relation to other circumstances as rest and exercises? It should also
contain general questions as eating, drinking and elimination behaviour regard-
less of the chief complaint.
At a clinical examination for aggressiveness and shyness, any painful events
should be watched for. Provoked pain by manipulation of the neck, spine and/or
limbs or distress by abdominal palpation and pain at rectalization of the prostate
in male dogs should be looked for.
Foreign bodies should be looked for in paws, ears and mouth. An elevated
body temperature is of diagnostic value, e.g. in meningitis. Special procedures
depending on clinical signs include clinical pathology of blood and urine, diag-
nostic imaging and endoscopy.
Clinical Signs Related to ‘Specific’ Behavioural Syndromes
‘Compulsive disorders’
There are a number of recurrent episodic behavioural alterations in dogs, hard to

classify according to current knowledge and available diagnostic techniques.
Historically, neurologists have looked upon them as possible epileptic seizures
evolving from the temporal lobe in the forebrain, while behaviourists have been
prone to call them compulsive disorders. Just looking at the clinical manifestations
of such disorders is not enough to judge who’s right and who’s not. The neuro-
physiology of the brain is complex enough to be able to function in such a way
that the clinical signs seen could result from either. Ideally, an EEG-registration
from the surface of the brain in a calm dog (not affected by drugs) just before,
during and after an episode, should throw some light on the issue. Unfortunately,
this is difficult to perform in a clinical setting.
Some of these ‘syndromes’ are recognized as symptomatic diagnoses, most
often breed-related (see Table 14.1), three of which are described in more detail
below.
Table 14.1. Some examples of behaviour signs associated with compulsive disor-
ders, and some typical breeds showing these signs.
Behaviour symptom Typical breeds showing symptoms
Fly catching Cavalier King Charles spaniel

Spinning Bull terrier, Staffordshire bull terrier
Tail chasing Bull terrier, German shepherd, Australian cattle dog
Rage syndrome English springer spaniel, English cocker spaniel
Chasing light reflexes
Staring at shadows Border collie
Flank sucking Doberman pinscher
Star gazing
Freezing Bull terrier
Fly catching
The dog acts as if repeatedly watching, and then catching, imaginary flies. These
episodes can last for minutes to hours, even to full-time occupation for different
dogs. The phenomenon seems to be breed-related in Cavalier King Charles
spaniels. Fly catchers have been ophthalmologically examined, excluding eye dis-
eases. Anti-epileptic drugs have been used without success (DeLahunta, 1983),
which contradicts the idea of this phenomenon in Cavalier King Charles spaniels
being an epileptic event. (Although there are also individuals with seizures, known
to be epileptic, that do not respond to anti-epileptic medication.) The current rec-
ommendation is to treat these dogs with behaviour modification and drugs that
interact with the neurochemical balance (Rusbridge, 2005).
On rare occasions, a moment of fly catching is the initial phase of a seizure
in a dog experiencing a secondarily generalized epileptic seizure, with convul-
sions. The event should then be regarded and treated as epileptic.
Tail chasing
Some dogs chase their own tails, or spin around in small circles, much too often
to be judged as normal. In some individuals, the intensity of this behaviour leads
Fig. 14.1. Care and handling at a veterinary clinic is of outmost importance for
canine behaviour as it relates to disease.
to dehydration and body weight loss because the affected dog doesn’t take time to
eat or drink. Dogs with this disorder should be clinically and neurologically exam-
ined to see if there are any clues for an evident aetiology for tail chasing or cir-
cling – e.g. inflammations of the anal sacs, or a morphologic brain lesion.
Many bull terriers seem to be predisposed to this behavioural disorder, and
some scientific reports about the matter have been published. The disorder has
been suggested by different authors to be an opioid-mediated stereotypy, a tem-
poral lobe epilepsy or a compulsive behaviour. The latter two suggestions may
not be mutually exclusive, as human beings with obsessive-compulsive disorders
sometimes also have an associated seizure disorder. In one study, affected dogs
had abnormal EEG patterns, indicating epileptic seizures, while unaffected
control dogs had normal EEGs (Dodman et al., 1996). There are several treat-
ment studies performed in dogs with this disorder, reporting responses to differ-
ent kinds of therapy, i.e. anti-epileptics, anxiolytics, opioid antagonists,
antidepressants and behaviour modification. The latest reports deal with antide-
pressants, e.g. clomipramine, and behaviour modification. Regardless of type of
underlying brain dysfunction, it has been stated that there is a strong genetic com-
ponent to these behaviours in the bull terrier breed.
Rage syndrome
Rage syndrome, also called idiopathic aggression or episodic dyscontrol,
describes a behavioural disorder where an otherwise healthy, nice and friendly
dog has recurring attacks of aggressiveness – often directed against members of
the owner’s family. There are cases reported with EEG evidence of temporal lobe
epilepsy and response to antiepileptic medication, as well as other reports of cases
with inconsistent EEG findings and disappointing therapy with anti-epileptics
pointing to this disorder being ‘just’ a behavioural problem.
A veterinarian facing a patient with a history of sudden, unexpected, aggres-
sive attacks, should perform a clinical and neurological examination together with
blood and urine analyses, to check for possible underlying morphologic, meta-
bolic or endocrine diseases. If not so, one should consider the option of this being
rage syndrome, whatever that means.
Special attention to this syndrome has been paid in certain breeds, e.g.
springer and cocker spaniels. Podberscek and Serpell (1996) found some evidence
that this condition in the English cocker spaniel breed is an expression of social
dominance rather than being a separate or pathological phenomenon. The clin-
ical picture of rage syndrome should not be mixed up with the ‘normal’ aggres-
siveness included in the postictal phase of some epileptic dogs (with more classical
epileptic seizures, cf. earlier text in this chapter). Those dogs are in a stage after
their epileptic seizure where they have no chance to obey or control their behav-
iour. One should bear in mind that dogs with unpredictable or uncontrolled
aggressiveness are potentially dangerous to people and other animals in their sur-
roundings.
References and Suggested Further Reading

DeLahunta, A. (1983) Veterinary Neuroanatomy and Clinical Neurology. W.B.
Saunders, Philadelphia, Pennsylvania.
Dodman, N.H., Miczek, K.A., Knowles, K., Thalhammer, J.G. and Shuster, L. (1992)
Phenobarbital-responsive episodic dyscontrol (rage) in dogs. Journal of
American Veterinary Medical Association 201, 1580–1583.
Dodman, N.H., Knowles, K.E., Shuster, L., Moon-Fanelli, A.A., Tidwell, A.S. and
Keen, C.L. (1996) Behavioral changes associated with suspected complex
partial seizures in Bull Terriers. Journal of American Veterinary Medical
Association 208, 688–691.
Gnirs, K. and Prelaud, P. (2005) Cutaneous manifestations of neurological diseases:
review of neuro-pathophysiology and disease causing pruritus. Veterinary

Dermatology 16, 137–146.
Horwitz, D., Mills, D. and Heath, S. (2002) BSAVA Manual of Canine and Feline
Behavioural Medicine. BSAVA, Quedgeley, UK.
Houpt, K. and Virga, V. (2003) Update on clinical veterinary behaviour. The
Veterinary Clinics of North America 33, 185–453.
International League Against Epilepsy, Glossary, http://www.ilae-
epilepsy.org/Visitors/Centre/ctf/glossary.cfm
Landsberg, G. (2005) Therapeutic agents for the treatment of cognitive dysfunction
syndrome in senior dogs. Progress in Neuro-Psychopharmacology and
Maarschalkerweeerd, J.L., Edenburg, N., Kirpenstein, J. and Knol, B.W. (1997)
Influence of orchiectomy on canine behaviour. The Veterinary Record 140,
617–619.
Overall, K.J. (1997) Clinical Behavioral Medicine for Small Animals. Mosby, St
Louis, Missouri.
Overall, K.L. (2000) Natural animal models of human psychiatric conditions: assess-
ment of mechanism and validity. Progress in Neuro-Psychopharmacology and
Platt, S. and Olby, N. (2004) BSAVA Manual of Canine and Feline Neurology.
BSAVA, Quedgeley, UK.
Podberscek, A.L. and Serpell, J.A. (1996) The English Cocker Spaniel: preliminary
findings on aggressive behaviour. Applied Animal Behaviour Science 47,
75–89.
Rijnberk, A. and de Vries, H.W. (1995) Medical History and Physical Examination
in Companion Animals. Kluwer Academic, The Netherlands.
Rusbridge, C. (2005) Neurological diseases of the Cavalier King Charles Spaniel.
Journal of Small Animal Practice 46, 265–272.
Studzinski, C.M., Araujo, J.A. and Milgram, N.V. (2005) The canine model of cog-
nitive aging and dementia: pharmacological validity of the model for assess-
ment of human cognitive-enhancing drugs. Progress in Neuro-
Psychopharmacology and Biological Psychiatry 29, 489–498.
Vaisanen, M., Oksanen, H. and Vainio, O. (2005) Postoperative signs in 96 dogs
undergoing soft tissue surgery. The Veterinary Record 155, 729–733.
Vite, C.H. (ed.) Braund’s Clinical Neurology in Small Animals: Localization,
Diagnosis and Treatment. International Veterinary Information Service, Ithaca,
New York (http://www.ivis.org).
Webb, A.A., Jeffrey, N.D., Olby, N.J. and Muir, G.D. (2004) Behavioural analysis of
the efficacy of treatments for injuries to the spinal cord in animals. The
Veterinary Record 155, 225–230.
Wojciechowska, J.L. and Hewson, C.J. (2005) Quality-of-life assessment in pet dogs.
Journal of American Veterinary Medical Association 226, 722–728.
Index
abnormal behaviour 231 blindness 253

acoustic communication 219 body size 16
action potential 91 body-language 228
activity rhythm 156 boldness 187
adaptation 208 Borophaginae 8
aggression 113, 229, 247 brain 63
aggressiveness 192 brain tumour 252
ancient dog 44 breed standard 53
anxiety 246 breeding pair 153
appeasement 228 breeding programmes 168
appetite 250 breeding value 168
archaeology 22 breeds 52
artificial selection 82 breed-typical behaviour 202
associative learning 122
attention 216–217
auditory cortex 93 Canidae 3
auditory system 92 Caninae 4, 10
Canini 11
cardiovascular disease 256
bang phobia 231 Carnivora 5
BARF 237 castration 236
barks 113, 115, 217 CBARQ 191
behaviour 61 chaining 129
behaviour therapy 232 choke chain 238
behavioural rating scales 190 clade 26
behavioural tests 188 classical conditioning 122
biting 247 clicker 126
263
264 Index
clicker training 235 excitability 194

clinical signs 245 extinction 134
cochlea 93 eye 95
colour vision 95–96
communication 106
complex traits 79 facial expression 114
compulsive disorder 257 facial signals 211
control of feral dogs 161 factor analysis 210
control region 24, 39 family groups 108
cooperative hunting 154 fatigue 246
coping style 182, 185–186 FCI 51
cortisol 73 fearfulness 190
counter-conditioning 141 feeding ecology 157
coyote 107 feelings 183
crossbreeding 34 feral dog 147
feralization 148
fight 113
dangerous behaviour 229 fitness 170
dating of origin 30 fixed ratio 130
deafness 253 fly catching 258
defecation 251 forebrain 64
denning 158 fossil 3
dental formula 7 founder 48
desensitization 139 fox 11, 87
detector dogs 178
development 69
diet 16, 237 gastrointestinal disease 255
dingo 35 gazing 215
disease definitions 244 gene 76
divergence 44 gene expression 81
diversity 38 genome 77
dizziness 254 gesture 211, 227
DMA 199 golden jackal 107
DNA 77 group size 154
DNA sequence 39
dog–human communication 211
domestication 21, 44, 187 habitat 161
dominance 109 habituation 121
haplotype 36
harness 239
ear 92 heredity 81
ear pinch 128 heritability 82, 178
early breeding 48 in personality 201
EBV 168 Hesperocyoninae 7
endocrine disease 254 hindbrain 64
endocrine management 236 home-range 155
energy 166 hormones 68
epilepsy 247 human bond 159
ethology 61 human-given cues 213
Index 265
hybrid 82 narcolepsy 244, 246

natural selection 166
nature and nurture 62
idiopathic aggression 260 neocortex 64
imitation 124 neurons 64
imprinting 92 niche 207
inheritance 167 nociceptor 101
instincts 77 nose 98
instrumental learning 234 nucleotide diversity 29
interspecific interactions
209
obedience 235
object name learning 218
itch 249 olfaction 98
operant conditioning 121
optimum trait value 171
jackal 107 origination event 40
overshadowing 132
key stimuli 66
pack 105
pain 101, 237, 249
learned irrelevance 131 palaeontology 31
learning theory 235 parasympathetic neuron
life-span studies 199 65
limbic system 64 parental care 108, 159
linkage disequilibrium 81 perceptional system 70
litter size 175 pessimistic 183
phenotype 79
pheromones 99, 240
marker 126 phylogenetic tree 41
Mendelian inheritance 79 phylogeny 13
meningitis 252 phylogeography 26
mentality 203 play 112
merle 80 playback 219
MHC 48 playfulness 193
microsatellites 40 play-sounds 113
midbrain 64 pointing 214
mixed-species group 209 population density 151
modal action pattern 67 population genetics 23
molecular clock 45 positive punishment 137
molecular methods 44 posture 15
monogamy 108 predation 157
motivation 71, 125 prey 157
motivational state 195 problem-solving 220
mRNA 78 prompting and fading 128
mtDNA 22, 39 pruritus 249
mutation 87 pseudopregnancy 236
muzzle 239 punishment 134
266 Index
puppy testing 198 stability 184, 197

QTL 84 standardization of tests 203
quantitative genetics 167 stereotypies 73
stimuli 66
stimulus equivalence 133
rabies 152 stimulus generalization 132
rage syndrome 260 strange-situation test 210
rank order 111 stray dog 148
recombination 84 stress 72
reflex 66 substitutions 32
rehome 230 sympathetic neuron 65
reinforcement 122
reinforcement schedule 129
reproduction 158 tail chasing 74, 258
resource allocation 169 tail docking 102
retina 95 taste 100
Rico 217 teeth 8
RNA 78 temperament 182
running 15 territorial behaviour 156
territory 152
trainability 184, 194
scavenger 151, 157 transcriptome 85
seizures 248
selective culling 151
selenocysteine tRNA 15 ultrasound 94
semi-natural conditions 105 umami 100
sensitization 121 unwanted behaviour 225
sequence analysis 22 urbanization 226
sex ratio 159 urination 250
shaping 127
shock collar 239
shyness–boldness 187 variable ratio 130
side effects of breeding 169 village dog 150
silver fox 87 vision 94
single gene effect 172 vocal repertoire 117
sires 52 vocalization 94
skull 5–6 vomeronasal organ 99
smell, inability 254
sociability 193
social cognition 219 wall jumping 74
social facilitation 123 weakness 246
social isolation 241 wild type allele 172
social learning 109, 123 wolf 21
social tolerance 111 word communication 217
social unit 153
socialization 212
solitary dogs 150 Y chromosome 40

The Behavioural Biology of Dogs

Uploaded by

Copyright:

Available Formats

The Behavioural Biology of Dogs

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Behavioural Biology of Dogs

Uploaded by

Copyright:

Available Formats

THE BEHAVIOURAL BIOLOGY OF DOGS

This page intentionally left blank

CABI Head Ofﬁce CABI North American Ofﬁce

Tel: +44 (0)1491 832111 Tel: +1 617 395 4056

ISBN-10: 1 84593 1874

Typeset by MRM Graphics Ltd, Winslow, Bucks.

I: The Dog in Its Zoological Context 1

1. Evolutionary History of Canids 3

3. Origin of Dog Breed Diversity 38

II: Biology and Behaviour of Dogs 59

4. Mechanisms and Function in Dog Behaviour 61

5. Behaviour Genetics in Canids 76

6. Sensory Physiology and Dog Behaviour 91

7. Social Behaviour of Dogs and Related Canids 105

8. Learning in Dogs 120

III: The Dog in Its Niche: Among Humans 145

9. Behaviour and Social Ecology of Free-ranging Dogs 147

10. Evolutionary Aspects on Breeding of Working Dogs 166

11. Individual Differences in Behaviour – Dog Personality 182

12. Human–Animal Interactions and Social Cognition in Dogs 207

IV: Behavioural Problems of Dogs 223

13. Behavioural Disorders of Dogs 225

14. Behaviour and Disease in Dogs 243

Savolainen, P., Department of Biotechnology, Albanova University Centre, KTH-Royal

recognized by Tedford (1978), each represented by a distinct subfamily (Fig. 1.1).

Fig. 1.2. Ventrolateral view of basicranial morphology of a primitive canid,

inner ear) ossiﬁes is of considerable importance in recognizing relationships

zoological classrooms. Dentally, canids are closest to the ancestral morphotype of

Fig. 1.3. Hypercarnivorous (b, Aelurodon and d, Euoplocyon) and

Evolutionary History of Canids

As in the hesperocyonines and borophagines, a small fox-sized species of Leptocyon

borophagines, Leptocyon generally remains small and inconspicuous, never having

As mentioned above, there is strong fossil evidence about the antiquity of

increasingly born out by molecular data in recent years, such as DNA–DNA

X. Wang and R.H. Tedford

relationship in the morphological tree supported by a large number of characters

Running and posture

Increased ability of running is often manifested in elongated and slender limbs,

Diet and body size

Finally, throughout their history canids displayed remarkable ﬂexibility and

However, in the last 10 years, genetic analyses, based on sequence analysis of

Origin from the Wolf

The Archaeological Evidence

Archaeological ﬁnds related to dogs have been found at Mesolithic sites in

1990), respectively. However, the archaeological evidence is problematic to inter-

Population Genetic Studies

A Single Geographic Origin

there is little geographic structure in the mtDNA of wolves, probably because

total extinction in large parts of their original geographical distribution. There is

Clade A Clade B Clade C Clade D Clade E Clade F Total

Fig. 2.2. Minimum-spanning networks showing genetic relationships among dog

An important observation is that the frequencies of clades A, B and C had

The Geographic Origin of Dogs

to be higher in the ancestral population. Comparing Europe, Southwest Asia and

Dating the Origin

East Asia Europe SW Asia

Fig. 2.3. Minimum-spanning networks showing genetic relationships among

Possible Crossbreedings with Wolves

The Origin of the Australian Dingo