A BIBLIOGRAPHIC SEARCH GRID, REVIEW
AND EVALUATION OF LITERATURE ON BARKING IN CANIDS:
Final Report: January 23, 2006
Principal Investigator:
Lynette
Hart, PhD, Professor, Department of Population Health and Reproduction
Co-principal Investigators:
Benjamin Hart, DVM, PhD, DACVB, Distinguished Professor,
Department of Anatomy, Physiology and Cell Biology, Chief of Behavior Service
Melissa
Bain, DVM, DACVB, Lecturer, Behavior Service Clinician
Other Collaborators:
Mary
Wood, MLS, Librarian, UC Center for Animal Alternatives
Laurie Bergman, DVM, DACVB, UC Behavior Service Clinician
Ulrike Reinisch, DVM, Behavior Resident, Veterinary
Medical Teaching Hospital
Karen Sueda, DVM, Behavior Service Clinician, Veterinary
Medical Teaching Hospital
Lori Gaskins, DVM, Behavior Resident, Veterinary Medical
Teaching Hospital
Jeannine Berger, DVM, Behavior Resident, Veterinary
Medical Teaching Hospital
Eranda Rajapaksha, DVM,
Behavior Resident, Veterinary Medical Teaching Hospital
Sponsored by and Submitted
to:
Radio Systems Corporation
Attn: Chris Brudecki
10427 Electric Avenue
Knoxville, TN 37932
865 218 4028
Submitted by:
Lynette Hart
Department of Population
Health and Reproduction
School of Veterinary Medicine
University of California,
Davis
Davis, CA 95616
TABLE OF CONTENTS
Page
Introductory Remarks
5
Literature Review
(Objective 1)
6
Comparison of Dogs with Ancestral Dogs and Canids
6
Analyses of Breed, Gender, and Body Size Differences in
Barking 7
Acoustical
Descriptions
7
Categorization
7
Comparison of barks with growls
8
Informational content of barks
8
Brain areas
8
Referential Context: When, Where, Why, Who?
8
Directionality of noisy and harmonic barks
9
Wild canids 9
Learning distinctive acoustic signals
9
Mechanics of Shaping Bark Signal
9
Frequency of hearing thresholds
9
Vocal tract
9
Effect of hormones
10
Developmental Influences and Age Changes in Barking
10
Puppy development and habituation
10
Changes with older age
10
Time of day
10
Predicting barking by individual dogs
10
Measures
of Human Disturbance by Barking
11
Barking in Kennels and Shelters
11
Factors influencing the rate of barking
11
Behaviors affecting relinquishment
12
Effects of barking on adoptions from shelters
13
Potential damage to auditory system 13
Effect of the Human Handler
13
Wolves as compared with dogs
14
Dogs’ responsiveness to cues
14
Barking on command
14
Interventions
to Alter Barking
14
Debarking
14
Halters and collars for inhibiting
barking
14
Muzzles for inhibiting barking
15
Dog-Appeasing Pheromone
15
Clinical
Context of Barking
16
Overall
importance of primary and secondary diagnoses
16
Problem behaviors and treatment with secondary
diagnosis of barking 16
1.
Separation
anxiety
16
2.
Fear (storm
phobia)
17
Evaluation of Contexts of
Barking in Behavior Cases during 2005,
17
UC Davis Veterinary
Medical Teaching Hospital
Development of
Bibliographic Search Grid on Barking
19
Functional and
Motivational Categories of Barking (Objective 2)
20
Gaps in the Knowledge of
Barking (Objective 3)
20
Operational Definition of
Problem Barking for Inclusion/Exclusion Criteria 21
(Objective 4)
Efficacy of Methods of
Controlling Barking Vis a Vis Categories of Barking 22
(Objective 5)
Introductory Remarks
Barking is a normal and
common behavior of dogs, occurring in a wide range of interspecific circumstances. It commonly consists of short, sharp, rapidly
repeated sounds with uniform loudness and pitch—properties that make the sound
very easy to localize (Scott 1968).
Throughout the development of breeds of dogs over the paste few
centuries, barking has been enhanced because of its functional role or reduced
because it interfered with other roles.
Thus, dogs bred for chasing foxes were bred for a low threshold for
barking, while today others bred as guide dogs have a higher threshold for
barking.
Through
this project concerning barking, chiefly of dogs, the literature review reveals
a significant gap between the research literature, which focuses specifically
on the phenomena associated with barking and its characteristics, and
veterinary clients’ experiences of their dogs’ barking as a problem behavior
that impacts them and their neighbors, sometimes leading to relinquishment or
euthanasia of their pet. The issues of
conflict that arise around the topic of barking are, in fact, at the core of
this topic’s emergence into importance. Yet, it is seldom a direct focus of research
from the standpoint of being a problem.
This
report is organized along the lines of several objectives set forth in the
Request for Proposals.
Literature Review
(Objective 1)
This review identifies the pertinent literature,
especially concerning dogs of various breeds, ages, and body sizes, from the
perspective of problem barking as it is characterized and resolved. A recently
developing broad literature describes the nature of the interactive
relationship between the person and dog, and behavior problems of dogs housed
in a variety of settings. Some of this
literature does not specifically address problem barking; in fact, problem
barking is rarely the main subject of a research paper. However, the evidence reveals that barking,
rather than being one salient presenting problem, typically is imbedded in one
or more other behavior problems. Thus,
the research readings submitted with this report include studies pertaining to
the relationship and relevant behavior problems that often are associated with
barking. Roughly half of the submitted
papers are specifically referenced in this report. Examples of topics that may
be relevant but are not always cited include: the importance of exposure and
experience in the learning of dogs as investigated in Hungary (e.g., Gacsi,
Miklosi and Varga 2004; Miklosi et al. 2005;
Pongracz et al. 2003; Viranyi et al. 2004); the role of play in dog-dog and
dog-human relationships as studied by Rooney and colleagues (e.g., Rooney and
Bradshaw 2002, 2003; Rooney, Bradshaw and Robinson 2000, 2001) and individual case
studies of behavior problems that may or may not include barking, such as
fearful or aggressive behavior or separation anxiety, as reported by veterinary
behaviorists (e.g., Frank 2005; Horwitz 1994; Pryor 2003a, 2003b).
In classical descriptions of vocalizations of the
Canidae of approach eliciting (decreasing distance) and warning or withdrawal
eliciting (increasing distance), the various roles of sounds were emphasized
(Tembrock 1968). Tembrock (1976)
categorized ten different classes of canine sounds, based on sonograms. For example, barking is a typical warning
sound announcing territory in Canidae—a single sound which may be repeated
irregularly. Pet dogs use barking in
various ways, such as attracting the owner’s attention and signaling their desires
to have doors open. The major and primitive
function of barking proposed by Scott (1968), long time investigator of dog
behavior, was as an alarm signal. This
was based on observing dogs reared in large fields apart from people. Major acoustic
features of barking are that the sounds are quite loud, carrying for long distances,
and that they are easily localized.
Comparison of Dogs with
Ancestral Dogs and Canids
One method of consideration of differences among the
canids, especially between wolves and dogs, is to study the effects of
domestication on the morphology and behavior of dogs. A shift to a smaller, more juvenile body
type, and more juvenile array of behaviors (referred to as neoteny), has
occurred in the process of domestication to dogs, and similar changes occurred
when foxes were selectively bred by Belyaev, choosing those animals that
responded well to people for the breeding population (Morey 1994). This perspective notes that wolves outgrow
whining and barking during puppyhood, whereas in dogs such vocalizations
persist into adulthood.
Recent
genetic analyses comparing mitochondrial DNA sequences in dogs and gray wolves
show that dogs may have originated from multiple wolf populations over 100,000
years ago (Vila et al. 1997; Wayne and Ostrander 1999). DNA hybridization data show that among the
carnivores, dogs, wolves, jackals, and Arctic foxes are closely related (Wayne,
Benveniste and O’Brien 1989), undoubtedly with some interbreeding. Another study compared gene expression
patterns in dogs, wolves and coyotes, in three parts of the brain: hypothalamus,
amygdala and frontal cortex, using microarray technology (Saetre et al.
2004). They reported that the
hypothalamus is highly conserved among the wild canids, whereas the dogs show a
marked divergence, suggesting strong selection on dogs for behavior during
domestication, with consequent rapid changes in brain gene expression.
More
specific genetic analyses using mitochondrial DNA have examined an array of dog
breeds to consider their extent of overlap and similarity to wolves (Parker et
al. 2004). Several breeds, namely the
Shiba Inu, Chow Chow, Akita, Alaskan Malamute, Shar-pei, Siberian Husky, and
Basenji, showed a great extent of overlap with wolves. Cluster analysis revealed another group of
mastiff-like breeds, and includes the Bernese Mountain Dog, Mastiff, Boxer,
Bull Terrier, German Shepherd Dog and Newfoundland, perhaps reflecting shared
morphology derived from a common ancestor.
A further cluster appears to reflect shared ancestral herding behavior
and includes the Shetland Sheepdog, Belgian Sheepdogs, and the Collie.
Analyses of Breed, Gender, and Body Size Differences
in Barking
Using different methods, and working with behavioral
rankings derived from interviews of veterinarians, we have published rankings by
veterinarians and dog obedience judges of 56 breeds of dogs on 13 behavioral
traits (Hart and Hart 1985, 1988). We recently
completed research on a new data set involving 80 breeds of dogs, including behavioral
rankings by 168 veterinarians on 10 behavioral traits (Hart and Hart,
unpublished results). The 1985-1988 and
the current studies both included watchdog barking and excessive barking among
the traits that were ranked, two types of barking that were scored
significantly higher in both studies for males than females. Some of the smallest breeds, such as were in
the top decile as ranked by veterinarians for excessive barking (Hart and Hart
unpublished data). Yet, the more high
pitched and lower amplitude barking of these small breeds may pose less of a
problem than the barking of a larger breed, with lower pitch and higher
amplitude. For example, the Siberian
Husky and Poodle in the ninth decile, and the German Shepherd and Irish Setter
in the eighth decile, would produce a bigger sound and perhaps be viewed as
more of a nuisance. In watchdog barking,
more large breeds were highly ranked than in excessive barking. German Shepherd, Rottweiler, and Doberman
Pinscher ranked in the tenth decile on watchdog barking, and these breeds would
be expected, if in someone’s backyard, to bark at people passing by the house,
such as children coming home from school.
Breeds ranking in the two lowest deciles on watchdog barking include
Bloodhound, Basset Hound, Great Pyrenees, Bulldog, St. Bernard, Golden
Retriever, and Newfoundland. For
excessive barking, the two lowest deciles include Newfoundland, Collie, St.
Bernard, Whippet, and Great Dane.
A
further point is that behavioral disorders (that could perhaps account for some
types of excessive barking) can have a genetic basis. Though not well studied from the standpoint
of barking, the nervous pointer dogs are a genetic model of severe anxiety,
that includes extreme fearfulness and avoidance of novel stimuli, symptoms that
often include barking (Uhde, Malloy and Slate 1992).
Acoustical Descriptions
A classical comparative study of wild canids and dogs identified twelve
basic vocal sound types by spectrographic comparisons (Cohen and Fox
1976). Simonet has even been proposed
that dogs engage in laughing (Milius 2001).
Bush dogs, crab-eating foxes, and maned wolves show similar repertoires to
each other, including whines, long-distance calls, and growls. As a model for what can be done, comparison
of species-level differences in vocalizations have been studied in domestic
cats and African wild cats, using multivariate analyses (Nicastro 2004). Vocalizations of European wolves and some
dog breeds have been compared, demonstrating complexity in dogs’ vocal
repertoires with many subunits of barking associated with particular motivations
and information (Feddersen-Petersen 2000).
Categorization. Methods
for categorizing vocalizations in dogs are somewhat controversial. Early work focused on the qualitative sound
repertoire and assigned vocalizations to various categories. Several studies along this line were
published in German in the 1960s and 1970s by Althaus, Meischner, and Zimen,
among others. The work of Tembrock
(1968, 1976) and Feddersen-Petersen (2000), initially in German and later in
English, examined sound classes in terms of amplitude and frequency set in
certain temporal patterns, and including attention to the behavioral
context. In recent years, building on
the earlier work, the question of canine vocalizations has taken on new
vitality with higher levels of instrumentation and analyses, so that recent
studies are able to classify vocalizations by precisely defined acoustic
properties.
Comparison
of barks with growls. One distinction that is commonly made is between
harmonic and noisy sounds. Growls tend
to be acoustically noisy, whereas whines are harmonic, and barks can be either
harmonic or noisy (Ohl 1996). More
recent work has focused on establishing measures for a harmonic-to-noise ratio (HNR)
to apply to dog barks, providing a method for quantifying the noise within the
signal (Riede et al. 2001). Nonlinearity
of vocalizations, in which the call is rough-sounding, can result from
asymmetry of the vocal folds (Riede et al. 2000). These lines of work are now emphasizing the
biomechanical properties of vocal production systems (Tokuda et al. 2002). Rather than categorizing sounds as either
harmonic or atonal, improved differentiation is possible among atonal calls
that have mixed structural components, such as harmonic calls with noisy
overlay. This has led to a further
useful measure, a low-dimensional nonlinearity measure (LNM), that is useful in
animal vocalizations with strong harmonic components, including subharmonics
and biphonation, or low-dimensional chaos.
Thus, nonlinear dynamics, such as limit cycles, subharmonics,
biphonation, and chaotic episodes, can then be assessed. These very technical approaches are methods
for analyzing and comparing all vocalizations with each other on a continuum,
rather than using a qualitative set of categories.
Informational
content of barks. An important question concerns the extent to which
vocalizations of dogs convey accurate information on the body size of the
vocalizing animal. A study of 47 dogs of
21 different breeds found that information produced by the physical nature of
the vocal tract in dogs results in the production of formants, from which
receivers can derive information about the characteristics of the sender (Riede
and Fitch 1999). Vocalizations were
compared with radiographs of the same dogs’ vocal tract while anesthetized. Vocal tract length, body mass, and formant
dispersion (the averaged distance between successive formants) were
significantly correlated, providing evidence that information on body size is
conveyed in the vocalizations.
Brain
areas. Through electrical stimulation of the midbrain,
natural-sounding growls, howls and whines were produced by anesthetized dogs
(Solomon, Luschei and Liu 1995). Though
barking is the most common type of vocalization produced by domestic dogs, no
barks were elicited in this study. Although
not well-studied, evidence concerning the neural processing of sounds by
animals and humans points to learning-induced plasticity within the auditory
cortex (Ohl and Scheich 2005). The
auditory cortex actively processes and represents stimuli, rather than simply
representing and recognizing stimuli.
This plasticity in neural processing is observed in single neuron firing
and in the level of spatiotemporal activity patterns in cortical areas. The learning-related plasticity is highly
task-specific, as required for fine discrimination and categorization tasks.
Referential Context:
When, Where, Why, Who
Within the field of animal
behavior, great interest is focused on discriminant analyses for assessing
individual and contextual differences in call characteristics, especially among
scientists studying primates. This
lively research area includes searches for meaning and function in
vocalizations, especially in primates, but vocalizations of predators,
including the canids, have received less attention. African wild dogs produce barks, but they
are the least frequently emitted of 8 classes of vocalizations in this species (Robbins
and McCreery 2003). Unlike wolves, wild
dogs respond by approaching, even when recorded playbacks are of low quality
(Robbins and McCreery 2003).
Directionality of noisy and harmonic barks. A few studies have reported referential-like qualities
in canine barks (Feddersen-Petersen 2000). In a direct comparison of barks by coyote and
dogs, the noisiness was similar, but within each species there was variation
among the individual vocalizers (Riede
et al.). Both harmonic and noisy barks
of dogs have a directionality that is affected by the head size and highest
frequency, so that sounds heard from the side or from behind will lack higher
frequencies and will sound muffled (Frommolt and Gebler 2004). The average differences in sound pressure
level between the front and rear position relative to the vocalizing animal was
3-7 dB.
Wild canids. A growing
number of studies in a variety of species, especially primates, have examined
whether vocalizations are functioning in referential fashion to distinguish
contexts, as well as whether the recipients respond appropriately to the context
in which the vocalizations are produced and the social identity of
callers. Arctic foxes distinguish the
barks of their own family group from those of other individuals, with major
differences in spectral parameters (Frommolt et al. 2003). Howls of wolves are individually distinct,
differing in fundamental frequency and the frequency variability between howls
(Tooze et al. 1990).
Learning distinctive acoustic signals. Analyses of barking of dogs when a stranger
rings the doorbell, the dog is isolated, or the dog is playing, show some
differences in characteristics (Yin 2002).
For example, the mean frequency is lower and the duration longer for
barking when a stranger rings the doorbell.
Current methods of multivariate analyses have revealed context
specificity and individual identification in the barking of dogs (Yin and
McCowan 2004). The emotional context of
dog barks can be categorized by humans, high above chance level, correlating
with peak and fundamental frequency and interbark intervals (Pongracz, Miklosi
and Csanyi 2005). Even human listeners
lacking experience with the particular dog breed or of owning a dog could
categorize bark situations well above chance level. An interesting finding is that dogs have
capability to recognize and remember acoustic stimuli. As an example, they can learn an auditory
recognition memory task with trial-unique stimuli easier and faster than
monkeys (Kusmierek and Kowalska 1998).
The classic work of McConnell (1990) with training
young domestic dogs showed that short tones with a rising fundamental frequency
are more arousing; they elicit approach and increase motor activity
levels. Human shepherds make use of the
different effects of acoustic signals by using short, rapidly repeated notes
with a rise in frequency for presenting stimulating signals, and prolonged,
descending single notes for inhibiting signals (McConnell and Baylis 1985).
Mechanics of Shaping
Bark Signal
Not surprisingly, the hearing capability of dogs is corresponds well with
the frequencies of their vocalizations, reflecting anatomical matching of the
hearing system with the vocal production apparatus.
Frequency
of hearing thresholds. As is well known, the sensitivity curves of hearing
for dogs reveal that the upper frequency limit is higher in dogs than humans
(Dworkin et al. 1940). The breed
and size of dog may play some role in affecting the hearing curves, but the
audiograms of several breeds ranging from Chihuahua to Saint Bernard revealed a
remarkably consistent response across breeds of dogs, with excellent
sensitivity in the range of 4-16 kHz (Heffner 1983).
Vocal tract. The acoustics of dogs’ vocalizations are
shaped by the vocal tract length (Riede and Fitch 1999), pointing to the
importance of the body size and breed of dog.
Further, the specific conformation of the vocal tract imposes specific
markers on vocalizations that can provide information on the identity of the
vocalizing individual.
Effect
of hormones. In females, the acoustic features of vocalizations and
rate of barking are influenced by hormonal status. Ovariohysterectomized German Shepherd females
barked more frequently, and at lower frequencies in their aggressive
vocalizations than intact females (Kim et al. 2005). Similar to findings in the studies of
behavioral traits of dogs mentioned above (Hart and Hart 1985, 1988,
unpublished), a study in the United Kingdom ranking dogs reported gender
differences in excessive and watchdog barking, with higher scores in males than
females; watchdog barking was associated with the aggressivity factor and
excessive barking associated with reactivity (Bradshaw et al. 1996).
Developmental Influences and Age Changes in Barking
In additional to genetic predisposition and anatomical influences, the
incidence of barking reflects changes with the life course and effects of
environmental experiences.
Puppy
development and habituation. Barking and growling emerge in puppies at about 24
days during play-fighting, and are used in aggressive situations around the 8th
week (Bleicher 1963). The first six
months of life are significant in providing experiences to dogs that increase
their exposure and knowledge of the world and lessen their probability of
having significant problems later in life.
The specificity of this early learning is indicated by a study in which
puppies allowed to observe their trained mothers performing specific tasks
concerning narcotic detection, performed the same task significantly better at
six months than non-exposed pups (Slabbert and Rasa 1997). In comparison with living in a kennel,
spending the early weeks of life inside a home and experiencing urban
environments in the early months of life reduce the likelihood of aggressive
and avoidance behaviors in adulthood (Appleby, Bradshaw and Casey 2002). Exposure to other dogs during development
plays a role in vocalizing, as socially isolated dogs vocalized less frequently
than dogs raised with others; however, when later exposed to conspecifics, they
vocalized at a normal level (Fox 1971).
Changes
with older age. Older dogs have been reported to bark excessively, in
comparison with middle-aged dogs or dogs younger than one year (Baranyiova et
al. 2004). In earlier observations,
clinicians had reported excessive vocalizations as a common behavior problem
among older dogs (Chapman and Voith 1990; Hunthausen 1994). Based on complaints of veterinary clients,
excessive vocalization occurred in 27% of the dogs over 10 years old brought
into a clinic for behavior problems (Chapman & Voith 1990). Separation anxiety was the most common case
of excessive vocalization, and was also the most common behavior problem in old
dogs. The onset of the anxiety was often
associated with a change in the routine or the death of another dog in the
household or a family member.
Time of day. Dogs appear to bark more during the day than at
night. Among various auditory stimuli,
barking of other dogs was most predictive of dogs being likely to bark or
become more alert at night, especially those dogs which lived in groups (Adams
and Johnson 1994). Guard dogs barked
five times more often during the day than at night, responding primarily to
human activity (Adams and Johnson 1995).
Predicting barking
by individual dogs. Relatively
little effort has focused on instruments to assess behavior of individual dogs
and thereby assist in screening at the time of adoption. The need to effectively select suitable dogs
for working as guides for the blind has led to efforts to develop screening
tools for 12 month-old dogs, and one questionnaire has been developed and
validated for evaluating behavior and temperament in guide dogs (Serpell and
Hsu 2001). Factor 1 was
stranger-directed fear/aggression, including being fearful when approached by
unfamiliar children or adults, and barking or growling when approached by
unfamiliar children, adults, or visitors.
This and two other factors showed moderate internal consistency, and
showed construct validity with the puppy raisers’ questionnaire assessments of
the dogs, suggesting that the questionnaire method can provide a means of
predicting the suitability of dogs.
Being suspicious of people is the primary behavioral reason that dogs
are rejected from guide dog training. In
this case, it appears that certain barking may be an indicator and predictor of
a more general problem of the young dog.
Measures of Human
Disturbance by Barking
Breeds have behavioral predispositions that interact with the
expectations of their owners and their home and neighborhood context. Dogs that live in close proximity to
neighbors and are left outside for long periods of time may be viewed as
problem barkers. It is a familiar
problem that dogs bark excessively in neighborhoods. The extent of this difficulty is seldom
measured, though it may be indicated in relinquishments to shelters where
barking is mentioned. A stratified
national probability survey in Australia in 1994 included a question about
barking, and about 40% of respondents were never disturbed by barking dogs in
their neighborhood (McHart et al. 1995).
About 10% were frequently and 6% constantly disturbed, and 21% each were
rarely or occasionally disturbed. The
point has been made that dogs kept outdoors all day and all night with no
company, toys, or walks, are likely to be anxious, bored, and over-reactive,
and suffering from a general welfare problem when they bark excessively
(Randall 2003).
Barking in Kennels and
Shelters
With dogs housed in kennels,
barking assumes a greater acoustic scale due to the numbers of dogs involved
and the consequent social facilitation of barking.
Factors
influencing the rate of barking. Some studies focused on improving the welfare of dogs
housed in shelters report environmental influences on barking. Environmental enrichment for dogs, whether by
offering socialization with dogs or humans or offering toys in the cage,
appeared to enhance welfare of the dogs, but at the same time each of these manipulations
increased the percentage of time when the dogs vocalized (Hubrecht 1993a). Social stimulation from humans increases the
vocalization of dogs, yet adoption of dogs from the shelter was increased by
giving the dogs increased regular human contact or enhancing the cage
environment (Wells and Hepper 2000b). Concerning another amenity for kenneled dogs,
purpose-bred dogs given supplemental exercise sharply increased their barking
(Clark et al. 1997). On the other hand,
the degree of visual contact with other dogs did not influence the extent of
dog vocalizations (Wells and Hepper 1998).
Another study found that excessive barking was the
most typical behavior among kenneled dogs, shown by 24% of dogs, with
differences in the predisposition among breeds (Clark et al. 1997). When large numbers of dogs are housed in a
facility it has been recommended that areas be subdivided to reduce the
socially facilitated barking (Hubrecht 1993b).
Regarding longterm housing of dogs in shelters, the amount of time that dogs
spent barking sharply declined after one month in the shelter, and the barking
continued to decline as months and even years passed (Wells, Graham and Hepper
2002a). Exposure to classical music
reduces barking of dogs and appears to calm them, perhaps improving their
welfare, and enhancing human perceptions of their desirability and adoptability,
whereas heavy metal music resulted in more barking (Wells, Graham and Hepper
2002b). Unpublished data by Graham,
Wells and Hepper were suggestive that the calming properties of lavender and
chamomile resulted in decreased barking and increased resting in kenneled dogs
(Wells 2004).
Social factors from humans significantly affect dogs
and may play a role in barking. Isolated
housing increased several behaviors that may indicate chronic stress and
identify poor welfare, including sharply increased vocalization; females were
more affected than males (Beerda et al. 1999a).
Further studies would be needed to distinguish this observation from
those above showing less barking among longterm-housed dogs, and more barking
among dogs receiving social stimulation.
Sheltered dogs showed more barking towards unfamiliar men than to
unfamiliar women standing in front of the cages (Wells and Hepper 1999). Taken in a broader communication context,
dogs’ social-communicative skills with humans acquired during the process of
domestication currently are a subject of intense research interest, as compared
with wolves (eg, Hare et al. 2002). Dogs
show a remarkable capability for cooperative interactions in partnership with
humans, as exemplified in the work of guide dogs with people who are blind
(Naderi et al 2001).
Behaviors affecting relinquishment. While various behaviors of dogs are known
to play a role in them being relinquished to shelters, barking is not an
important reason. The primary reasons
are inappropriate elimination and aggression, both of which are less common
among animals receiving frequent veterinary care (Scarlett et al. 2002). Based on a mail survey in Northern Ireland of
556 people, most dogs in shelters exhibit problem behaviors, and these
undesirable behaviors inevitably increase the likelihood that an animal will be
returned (Wells and Hepper 2000a).
People are highly selective when taking strong action to correct or
remove a dog with a behavior problem from their home. Dogs exhibiting aggression towards humans
were returned to shelters at a rate of 30%, whereas dogs with excessive
barking, along with other problems such as aggression towards other dogs,
roaming, excessive activity, destructiveness, fearfulness, and inappropriate elimination
are returned at a far lower rate of 15% or less.
One
study tracked all admissions to three Australian shelters during one year and
analyzed the data, with an unusual breakdown of categories that separated
aggression, temperament, and behavior as reasons for relinquishment and
euthanasia (Marston, Bennett and Colman 2004).
Among dogs relinquished for behavioral reasons (excluding aggression and
temperament), 10% of relinquishments were for the stated reason of barking,
which was the third most commonly reported behavioral problem. Among dogs presented to be euthanized for
behavior problems, barking was the fourth ranked reason, accounting for 12% of
the euthanasias on account of behavior. However,
no single problem, be it barking or other primary problems, may account for
these relinquishments leading to euthanasia.
A case control study in Sacramento, California, on the determinants of
relinquishment to animal shelters, revealed the complexity of breakdowns in the
relationship of owners with their dogs; rather than a single problem, several
risk factors appeared to combine (Kass et al. 2000). As a further example of the complexity of
this issue, a questionnaire distributed among 4,359 members of the Danish
Kennel Club revealed breed differences in the risk of interdog aggression and
shooting phobias among purebred dogs as perceived by their owners, but the risk
was also strongly influenced by the owners’ lack of knowledge, lack of
experience and lower age (Rugbjerg et al. 2003). A follow-up on the success of adoptions six
to eighteen months later found that about one-fourth failed (Nixon, Hart and
Willits 2001). Half of the failures
related to behavior problems, and a strong majority of these were aggression
that became evident within one day of adoption.
In conclusion, barking appears not to be a common
cause of relinquishment, as results in the summary study of the large regional
study of animal shelters showed that a majority of the relinquished dogs never
had been too noisy, and more than 75% never growled at people or at animals
(Salman et al. 1998). One should not
dismiss barking, as at times it emerges as a compelling behavior problem
disturbing owners or neighbors. Vocalizing
too much emerged in a national study as the tenth ranked behavioral reason for
relinquishment (Salman et al. 2000).
Among dogs in this dataset relinquished to be euthanized, 16% were
relinquished for behavioral reasons (Kass et al. 2001). Reasons given by these owners for requesting
euthanasia involved the dog posing a threat to safety due to aggression to
people or animals, or biting people or other animals, whereas excessive
vocalization was only mentioned by 7% of these owners. Behavior problems and a lack of veterinary
care were associated with greater risk for relinquishment of dogs, but unwanted
barking was not associated with an increased risk. Inappropriate elimination, unwanted chewing,
hyperactivity, and aggression toward people or other pets were the problems
with greater risk (Patronek et al. 1996).
Behavior problems were the most common reason for relinquishing a dog in
a study from the central United States, where the 7th, 8th,
and 9th most common such reasons were: barking when the owner was
absent, barking when the owner was at home, and barking outdoors (Miller et al.
1996). Such behaviors take a toll on
families and neighbors, and even on the broader community (Senn and Lewin
1975).
Effects
of barking on adoptions from shelters. Little information is available concerning successful
adoptions of dogs from shelters, and factors that may be involved. Small size and a history of having lived
indoors, plus lighter coat colors were associated with successful adoption
(Posage, Bartlett and Thomas 1998).
Behavioral problems did not appear in the multiple logistic regression
model, but this may have been due to their relatively small incidence. One unpublished study examined the success of
adoptions of dogs by following up.
Aggression that appeared just after acquiring the dog was the most common
difficulty in new adoptions, suggesting that placements included dogs with too
high a predisposition to be aggressive (Nixon, Hart and Willits 2001). In an attempt to improve the success of
adoptions from shelters, a group in Netherlands developed a behavioral test to
be used by staff in animal shelters to predict problem behavior and found it
was more accurate than simply taking opinions of staff in predicting problems
(van der Borg, Netto and Planta 1991).
For separation anxiety, a condition that frequently includes excessive
barking, the test showed a predictive value of 92%, as compared with staff opinions
which showed a predictive value of 78%.
Potential
damage to auditory system. The
sound levels in kennels and research facilities can be high. Dogs’ hearing is sensitive in the range of
500Hz-16kHz. Sound levels recorded
overnight for this frequency range showed peak values commonly between 58 and
75 dB on a daily basis, but also reaching 95-100 dB (Sales et al. 1997). Certain facilities reached much higher noise
levels, especially during cleaning or use of a high pressure hose. During the day, peak values from all sources
of noise regularly exceeded 100 dB and often reached 125 dB, pointing to a
possible welfare problem for dogs. (For humans,
prolonged exposure at 85dB can cause hearing damage.) Another set of measurements was taken in a
block of kennels when visitors were eliciting a high level of barking and
showed the overall levels of sound to be 90-99 dB at a frequency of 1 kHz (Milligan,
Sales and Khirnykh 1993). The number of
dogs barking in dog kennels reflects the number of people present in a kennel
block, and when the dogs barking the most were removed in one study, the
incidence of barking remained lower (Ledger, Hubrecht and Sales 1996).
Effect of the Human Handler
Unlike biting or growling, barking is a ubiquitous behavior for
dogs. When viewing it as a problem
behavior, one is always faced with considering the extent of barking, the
context, and timing, along with the many subtle factors that play into the
barking of a particular dog. The
behavior of the handler plays an important role in shaping the nature and
extent of problem behavior. For example,
a survey of owners’ training methods suggested that their use of punishment in
training their dogs, rather than rewards, was associated with an increased
incidence of problematic behaviors, of which barking was one (Hiby, Rooney and
Bradshaw 2004).
Wolves
as compared with dogs. Even with hand-raising and intensive socialization,
wolf pups strongly differed from dog pups in their preference for humans, with
dog pups showing more tail wagging upon approaches from humans, distress
vocalizations upon being left alone, and other communicative signals than wolf
pups at 3-5 weeks of age that facilitate social interactions (Gacsi et al
2005).
Dogs’
responsiveness to cues. More evidence has been coming forth on the social and
cognitive responses of dogs, not only to other dogs, but also to humans (Call
et al. 2003). Pet dogs are attentive to
humans’ faces and behavior, and can learn from an unfamiliar human who gives
verbal feedback or from a strange trained dog who demonstrates the task
(Pongracz et al. 2004). Dogs respond to
subtle directional cues from humans, even a non-owner (Soproni et al. 2001),
and dogs presumably cue in on human preferences regarding barking.
Barking
on command. A
voluntary aspect of barking, at least on occasion, is indicated by dogs being
trained to bark on command from the owner.
In one study, 20% barked in response to a verbal command from a speaker,
and most barked when seeing the owner in video-projection while hearing the
voiced command through a speaker (Pongracz et al. 2003). A pharmaceutical intervention with
healthy, non-problem dogs indicated that, with incentive motivation, selegiline
hydrochloride played a positive role in the performance of dogs on training and
learning tasks (Mills and Ledger 2001).
With a luring reward, dogs treated with selegiline performed better and
required fewer reinforcements to reach success.
However, the performance without the reward was worse than controls.
Interventions to Alter
Barking
Juarbe-Diaz (1997) has
reviewed the variety of treatments and devices that are marketed to reduce
barking of dogs. Marketing strategies
often suggest that a device can serve as a targeted solution to the problem
barking. The evidence suggests that most
often, however, the barking arises in a context where the animal has an
environmental or behavioral problem that is broader than the sole problem of
barking. Thus, a single intervention,
without concomitant behavior modification, may not succeed.
Debarking. Debarking is a surgery that is sometimes performed to
reduce barking noise by dogs (Perkins 2002), but that may result in an
increased barking rate (Cronin et al. 2003).
Halters
and collars for inhibiting barking. In a comparison of dogs wearing head halters or
collars, or conventional neck collars, dogs appeared more calm and relaxed when
wearing the head collars, and no physiological differences were found in plasma
ACTH and cortisol levels (Ogburn et al. 1998).
Interventions for problem barking include collars that
deliver shock or citronella spray. Some
veterinarians in Australia have set forth arguments against (Calnon 2003), or
for, electronic collars (Perkins 2003; Aubrey 2003). Studies have examined whether the collar is
delivering pain (which may be indicated by vocalizations) or causing stress
(Schilder and van der Borg 2004). It has
even been conjectured that collars may elicit serious aggressive attacks on
humans (Polsky 2000). Citronella or a
scentless spray collar was somewhat effective for the control of barking, with
both styles of collar leading to a reduction in barking for a majority of the
dogs evaluated (Moffat, Landsberg and Beaudet 2003). A concern with the citronella spray collar
in reducing certain forms of barking in dogs is that the barking increases over
the time the collar is worn, particularly in dogs wearing the collar every day,
but the efficacy was extended if the collar was worn intermittently (Wells
2001). A systematic evaluation for
efficacy of citronella spray and electronic shock collars for barking found
both were effective in reducing barking (citronella by 89%, shock by 44%), but
that most owners preferred the citronella spray collar (Juarbe-Diaz and Houpt
1996).
Another
device that is marketed involves production of high frequency sounds, sometimes
to be triggered by barking of the dog wearing a collar holding the device. A review of these devices finds no evidence
of efficacy (Lindsay 2000, pp 134-135).
This absence of efficacy may reflect the finding that the devices seemed
not even to be very aversive to the dogs, who presumably would hear the ultrasounds
(Heffner 1983). Further, ultrasonic
vibrations are easily shadowed, possessing a narrow field of directionality and
a limited effective range. This approach
is not one likely to show promise.
Muzzles for inhibiting barking. A
systematic study of the short term effects of wearing an anti-barking muzzle on
the occurrence of barking by dogs found a significant reduction of barking and
a lowering of their signs of activity level, with no effect on their free
cortisol concentrations (Cronin et al 2003).
The appearance of such a muzzle on a dog may be problematic in giving an
impression that the dog may be dangerous.
Further, in view of other studies showing that dogs habituate to collars
and then resume barking, it would be important to examine longterm effects of
this device before making conclusions about its efficacy.
Dog-Appeasing
Pheromone. A new treatment marketed for barking of dogs is dog
appeasing pheromone (DAP). Limited
research suggests some moderating influences but not a major impact for calming
dogs in a shelter setting, when continuously administered to a treatment group
of 37 dogs over a 7-day period, and compared with a placebo group of 17 dogs
(Tod, Brander and Waran 2005). In a 1
min session, a test observer walked along a section of the kennel at 7 days:
the mean barking amplitude was significantly reduced in the treatment
group. While the peak barking amplitudes
at 30 seconds did not differ in the two groups, on average there was a
difference of 20 dB at any one time between the placebo and DAP treatment
groups. This study suggests that DAP may
have some efficacy for problem barking in shelter situations. Another study reported a randomized, blind trial
with DAP versus clomipramine in combination with a behavioral plan, finding a
quick reduction in the undesirable behaviors, including a marked decrease in
vocalization with both treatments (Gaultier et al. 2005). These authors concluded that DAP and
clomipramine did not differ in efficacy of reducing the undesirable behaviors
of dogs with separation related problems.
However, the clomipramine resulted in significantly more undesirable
effects.
Dog-appeasing
pheromone has also been tested in an open clinical trial for the alleviation of
behavioral signs among dogs showing signs of fear in response to fireworks
(Sheppard 2003). Significant
improvements occurred in a majority of the signs of fear, and a reduction of
vocalizing was among the changes reported with treatment.
Clinical Context of
Barking
The incidence of barking may
be very high, but it is unclear at what point it is perceived as a problem
behavior requiring some attention. As
reviewed below, data suggest that a substantial proportion of people view their
dogs as barking more than is desirable, yet the behavior reaches a critical
point in relatively few dogs. Little
information is available concerning how that threshold of problem barking is
reached and defined, and by whom.
Overall
importance of primary and secondary diagnoses. In a
questionnaire of dog owners in the United Kingdom, over 75% of people reported
that their dogs were currently displaying barking at people, and over 50%
mentioned them barking at other dogs (Hiby et al. 2004). A study in Australia described 32% of dogs
engaged in excessive barking, as reported by their owners, with whining,
howling, and growling also occurring, but at much lower levels (Kobelt et al.
2003). Barking is an inappropriate behavior for guide
dogs, yet among potential guide dogs, one-third of puppies showed barking or
growling (Koda 2000). These undesirable behaviors
were shaped by training, such that the barking was reduced by not responding to
the puppies when they showed the behavior.
Despite the visibility and audibility of dog barking, it
has failed to emerge as a primary behavior problem, as indicated by the lack of
clinical drug trials treating on barking as a behavior problem, and the small
number of research papers focusing on barking.
More commonly, it is simply one behavior among others within an overall
more general diagnosis. Detailed
examples of this point are provided below in the analyses of clinical behavior
problems of dogs at the UC Davis Veterinary Medical Teaching Hospital in 2005
involving barking.
Problem
behaviors and treatment with secondary diagnosis of barking. The
large disparity between the proportion of people who consider that their dog
manifests a problem with barking, and those who take some corrective action by
scheduling a clinical appointment for a workup, emphasizes that owners make the
decision of when the dog is manifesting a serious behavior problem with barking
(Juarbe-Diaz 1997). Despite barking being a commonplace problem, its appearance
as a behavior problem usually arises in a broader context where the dog is
extremely fearful, aggressive, or anxious, rather than as a single sign. Most clinical appointments for behavior
problems involving dogs concern aggression; in such cases the type of
aggression also needs to be differentiated (Reisner 2003).
1. Separation
anxiety. One context for barking which
has been studied is the behavior problem of the separation anxiety syndrome, occurring
when the owner leaves, and diagnosed by the dog’s consequent vocalizing,
inappropriate elimination, and/or destructiveness (Schwartz 2003). Among dogs diagnosed with separation anxiety
associated with departure of the owner, 61% excessively vocalized in the
owner’s absence, as compared with 5% of a control group (Flannigan and Dodman
2001). Another study of dogs with
separation anxiety found that 67% of the dogs showed excessive barking (Gaultier
et al. 2005). A comparison of dogs with
separation anxiety, noise or thunderstorm phobia or both, reported
vocalizations by a majority of dogs with separation anxiety alone, or combined
with thunderstorm phobia, but not storm phobia alone (Overall, Dunham and Frank
2001). Dogs with separation anxiety were
more likely to be from a home with a single adult human, and to be sexually
neutered. Limited social experience was
associated with separation-related behavior, of which barking was the most
common behavior, followed by destructiveness and then howling (Bradshaw et al.
2002). A single case study of a dog
diagnosed with separation anxiety documents the extreme whining and barking
when left by the owner, with an amelioration following treatment employing
behavioral modification and pharmacologic treatment (Frank 2005).
Treatment
for this type of problem with separation anxiety begins with behavior
modification and environmental management; in addition, pharmacological management may
include tricyclic antidepressants with serotonin-reuptake inhibiting
properties, and sometimes additional benzodiazepines (King 2000). In 24 dogs with obsessive-compulsive
disorder, separation anxiety, or noise phobia, that were given clomipramine and
behavior modification (but with no placebo control group), 16 dogs showed
substantial or full improvement in their primary presenting complaint;
vocalizing had been the primary complaint for 11 of the dogs (Seksel and
Lindeman 2001). In one evaluation of clomipramine as an adjunct to behavioral
therapy for separation anxiety in dogs, all dogs received behavioral therapy
and improved, but the clomipramine did not affect the typical behavioral
problems (Podberscek, Hsu and Serpell 1999).
A large double blind, placebo controlled trial of clomipramine for
separation anxiety yielded modest positive results but not for the vocalizing
(King et al. 2000). A study following
up 6 to 64 months following a behavioral examination on 52 dogs with separation
anxiety, some of which had been treated with amitriptyline, found that 62% had
improved (Takeuchi, Houpt and Scarlett 2000).
Among the 17 dogs treated with amitriptyline, 9 were improved.
2. Fear
(storm phobia). Signs of fear shown
by dogs with storm phobia included excess vocalizations among a majority of
dogs (Crowell-Davis et al. 2003). In
other work, investigators note that some very fearful dogs do not vocalize
(Horwitz 1994; Aronson 1999). Treatment
with clomipramine, alprazolam, and behavior modification decreased the signs of
fear, including the excessive vocalizations.
Evaluation of Contexts of
Barking in Behavior Cases during 2005, UC Davis Veterinary Medical Teaching
Hospital
To
provide an additional perspective on barking as it fits within the realm of
clinical behavior problems, we reviewed the diagnoses of dogs seen at the UC
Davis Veterinary Medical Teaching Hospital for behavior problems during 2005,
presented below in Table 1. Those dogs
for which the term “barking” appeared in the history, were tabulated for their
primary diagnoses, as were the remaining dogs, for which there was not a term
“barking” in the case history. This was
a total of 71 cases where barking was mentioned, 327 where it was not
mentioned, for a total of 398 cases. In
most instances, there were multiple diagnoses.
The following table categorizes cases based only on the primary
diagnoses.
Table 1. UC
Davis Dog Behavior Cases in 2005, as a Function of Diagnosis and Mention of
Barking in the History
Cases Involving Cases Not Involving
Barking
Barking
Principal Diagnosis
N
= 71
N = 327
--------------------------------------------------------------------------------------------
Aggression
Stranger
Fear p <.1, p >.05 21%
12%
Inter-dog
13%
12%
Dominance-Related
Conflict 7%
12%
Dominance
Status, Interdog 10%
14%
Fearful
7%
10%
Territorial
p <.02
6%
1%
Separation Anxiety
11%
12%
Attention seeking
p <.02
6%
1%
Cognitive dysfunction
0%
4%
Housesoiling
0%
3%
Storm phobia
0.3%
5%
Number of diagnoses Mode
= 3 Mode = 2
Mean
= 3.6 Mean = 2.9
These cases reflect a tertiary care hospital, representing more persistent problems that have been difficult to resolve. The first six categories are types of aggression. In a retrospective analysis based on mention of barking in the history, the data suggest that primary diagnoses of territorial aggression and attention-seeking occur more frequently among cases where barking is involved than in cases wher