Characterisation of plant eating in dogs

Abstract

Grass or plant eating is a widely recognized behaviour amongst domestic dogs. We first estimated the prevalence of plant eating by administering a written survey to owners of healthy dogs visiting the outpatient service of a veterinary medical teaching hospital for routine health maintenance procedures. Of 47 owners systematically surveyed whose dogs had daily exposure to plants, 79% reported that their dog had eaten grass or other plants. Using an internet survey targeting owners of plant-eating dogs, we then acquired information regarding the frequency and type of plants eaten, frequency with which dogs appeared ill before eating plants and frequency with which vomiting was seen afterwards. Of 3340 surveys returned, 1571 met enrollment criteria. Overall, 68% of dogs were reported to eat plants on a daily or weekly basis with the remainder eating plants once a month or less. Grass was the most frequently eaten plant by 79% of dogs. Only 9% were reported to frequently appear ill before eating plants and only 22% were reported to frequently vomit afterwards. While no relationship was found between sex, gonadal status, breed group or diet type with regard to frequency or type of plants eaten, a younger age was significantly associated with: (1) an increase in frequency of plant eating; (2) an increase in consuming non-grass plants; (3) a decrease in regularly showing signs of illness before eating plants and (4) a decrease in regularly vomiting after consuming plants. The findings support the perspective that plant eating is a normal behaviour of domestic dogs.

Full text

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Author's personal copy
Characterisation of plant eating in dogs
Karen Lynn Chieko Sueda
a,
*, Benjamin Leslie Hart
b
, Kelly Davis Cliff
b
a
Behavior Service, VCA West Los Angeles Animal Hospital, c/o Animal Imaging,
1827 Pontius Avenue, Los Angeles, CA 90025, USA
b
Department of Anatomy, Physiology and Cell Biology, School of Veterinary Medicine,
University of California, Davis, CA 95616, USA
Accepted 24 May 2007
Available online 12 July 2007
Abstract
Grass or plant eating is a widely recognized behaviour amongst domestic dogs. We first estimated the
prevalence of plant eating by administering a written survey to owners of healthy dogs visiting the outpatient
service of a veterinary medical teaching hospital for routine health maintenance procedures. Of 47 owners
systematically surveyed whose dogs had daily exposure to plants, 79% reported that their dog had eaten grass
or other plants. Using an internet survey targeting owners of plant-eating dogs, we then acquired information
regarding the frequency and type of plants eaten, frequency with which dogs appeared ill before eating plants
and frequency with which vomiting was seen afterwards. Of 3340 surveys returned, 1571 met enrollment
criteria. Overall, 68% of dogs were reported to eat plants on a daily or weekly basis with the remainder eating
plants once a month or less. Grass was the most frequently eaten plant by 79% of dogs. Only 9% were reported
to frequently appear ill before eating plants and only 22% were reported to frequently vomit afterwards. While
no relationship was found between sex, gonadal status, breed group or diet type with regard to frequency or type
of plants eaten, a younger age was significantly associated with: (1) an increase in frequency of plant eating; (2)
an increase in consuming non-grass plants; (3) a decrease in regularly showing signs of illness before eating
plants and (4) a decrease in regularly vomiting after consuming plants. The findings support the perspective that
plant eating is a normal behaviour of domestic dogs.
#2007 Elsevier B.V. All rights reserved.
Keywords: Dogs; Canids; Feeding behaviour; Plant eating; Grass eating
1. Introduction
One area of domestic dog (Canis familiaris) behaviour that has received little attention in the
scientific literature is the regular or periodic consumption of plant material of little apparent
www.elsevier.com/locate/applanim
Applied Animal Behaviour Science 111 (2008) 120–132
* Corresponding author. Tel.: +1 310 478 5035; fax: +1 310 478 5135.
E-mail address: Karen.Sueda@vcamail.com (K.L.C. Sueda).
0168-1591/$ – see front matter #2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.applanim.2007.05.018

Author's personal copy
nutritive value. Judging by how frequently animal behaviourists are asked about plant eating, this
behaviour appears to be quite common, though no studies have documented its prevalence among
domestic dogs. Amongst the most commonly asked questions are whether plant eating is a sign of
illness, if dogs eat plants to induce vomiting and whether plant eating is a sign of a dietary
deficiency (Beaver, 1981; Houpt, 2005; Lindsay, 2001; Overall, 1997).
Evaluation of scats and stomach contents of wild canids indicate that they regularly ingest
plant material, especially grass. Such studies report the appearance of plant matter in about 11%
of samples of Latvian wolves (Andersone and Ozolins, 2004) and approximately 74% of samples
collected in the summer from grey wolves reintroduced to Yellowstone National Park (Wyoming,
USA) (Stahler et al., 2006). In studies where specific plant types were identified, grass was found
in 2–3% of the samples of timber wolves (Mech, 1966), 10% of samples in Latvian wolves
(Andersone, 1998) and 14% of samples from Greek wolves (Papageorgiou et al., 1994).
Some blades of grass undoubtedly appear in scats as a result of adherence to carcasses being
eaten, or from ingestion of the intestinal contents of herbivorous prey. However, intentional
consumption of grass by wolves has also been documented (Murie, 1944; Stahler et al., 2006).
After observing blades of grass wrapped around intestinal worms in wolf scats, Murie (1944)
suggested that grass might have a scouring effect in removing worms.
Based on the studies and the commonly asked questions mentioned above, four hypotheses
were developed regarding plant eating in domestic dogs: (1) most plant eating occurs in dogs
exhibiting signs of illness; (2) most bouts of plant eating are followed within a few minutes by
vomiting; (3) plant eating occurs primarily due to a dietary deficiency of some type and (4) plant
eating is a normal, relatively innate behaviour, possibly reflecting a behavioural predisposition of
their wild canid ancestors and relatives. The hypotheses are not necessarily mutually exclusive.
For example, while most periodic plant eating could be normal, gastrointestinal distress or illness
may also trigger the behaviour.
Following a pilot survey to gather preliminary information about the occurrence of illness
prior to and vomiting after eating plants, two studies were conducted to characterize plant-eating
behaviour amongst domestic dogs and to test the above hypotheses. The purpose of Study 1 was
to estimate the prevalence of plant eating in a sample of healthy, well-cared-for dogs. In Study 2
we sought to acquire extensive data about plant eating with regard to types of plants eaten,
frequency of plant eating and influences of age, breed group, sex and gonadal status in a large
sample of domestic dogs that ate plants. Of particular interest was the acquisition of information
relevant to predictions related to the above hypotheses.
2. Materials and methods
2.1. Pilot study
Veterinary students at the University of California, Davis (CA, USA) were asked to fill out a survey on
plant eating in their dogs. This pilot survey was intended to aid in the design of the questionnaires used in
Studies 1 and 2 and to provide preliminary data on the proportion of dogs that show illness prior to eating
plants and the proportion that vomit afterwards.
2.2. Study 1
2.2.1. Study design
The primary purpose of Study 1 was to estimate the prevalence of plant-eating behaviour from a
convenience sample of at least 40 healthy dogs presented to the outpatient service of a veterinary teaching
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132 121

Author's personal copy
hospital. Data were collected on sex, age, gonadal status and diet. Also, for plant-eating dogs, data were
collected on the regular occurrence of signs of illness prior to eating plants and the occurrence of vomiting
regularly within 1 h after eating plants.
2.2.2. Survey design
A written survey, previously tested in the pilot study, was given to owners of dogs brought to the
outpatient clinic of the University of California, Davis, Veterinary Medical Teaching Hospital (CA, USA).
An attempt was made to personally hand the survey to all clients, so as to guard against the bias inherent in
allowing clients to voluntarily pick up a survey form. The owners were told that the survey was designed to
study eating behaviour of pet dogs in general and not specifically plant eating. The survey contained
questions regarding the general description of the dog and its environment; information about the
consumption of non-food items, such as stones, clothing, paper, toys, faeces and plants; current diet
and medical history (Table 1). Owners bringing more than one dog to the hospital were asked to complete
one survey for the dog with the earlier appointment time. Questions were presented in multiple-choice and
short-answer format. For multiple-choice questions, owners were specifically instructed to select either the
most accurate response or for some questions, all the responses that pertained to their dog. The survey took
approximately 10 min to complete. Owners of dogs that consumed plants were asked additional questions
that specifically addressed plant-eating behaviour, including whether their dog frequently exhibited signs of
illness, described as depression, loss of appetite and inactivity (Hart, 1988) before eating plants, and whether
their dog regularly vomited within 1 h after eating plants. Owners were asked to describe how their dog
behaved most of the time and not how the dog behaved on one or a few occasions.
A medical history was obtained and a physical examination was performed on each dog by a veterinarian
to rule out medical conditions that could cause vomiting not associated with plant eating and diseases that
might increase or decrease ingestive behaviour in general. Medications and medical problems that
warranted exclusion of subjects included gastrointestinal disease, hyperadrenocortisim, diabetes mellitus
and administration of medications such as corticosteroids or anticonvulsants that may increase appetite
(Plumb, 2002).
Surveys were included in the study only if the individual completing the survey was the dog’s primary
caregiver, the dog had access to plants for at least 1 h per day, the dog had been in the household for at least 1
month and owners were able to observe their dog’s behaviour for at least 3 h per day. These restrictions were
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132122
Table 1
Study 1, summary of survey questions presented to owners of both plant-eating and non-plant-eating dogs brought to the
University of California, Davis Veterinary Medical Teaching Hospital outpatient clinic (not inclusive of all questions)
Section
Description of dog Sex, gonadal status, major breed group, age
Duration in household
Time spent with dog daily and able to observe behaviour
Hours per day dog has access to plants
Diet fed Type and percentage of diet consisting of commercial
kibble/canned or raw, ‘‘people food’’, treats
Information on non-food items consumed Types of non-food items consumed
Information on plant eating behaviour (if owner had observed
their dog eating plants or seen plants in vomit or feces)
- Type of plants eaten
- Type of plant most frequently eaten
Signs of illness and vomiting before and after plant eating
Medical history Current medical history
Current medications

Author's personal copy
imposed to ensure that the dog had sufficient opportunity to engage in plant-eating behaviour and be
observed by their owner.
2.2.3. Data analysis
The proportion of respondents who indicated that they had witnessed their dog eating plants and/or
observed plants in their dog’s faeces or vomit at least once was determined. Dogs were evaluated as to
whether they were fed predominantly (at least 51% of the time) a commercial, nutritionally balanced diet of
kibble or canned food certified by the Association of American Feed Control Officials (AAFCO), or
primarily raw or home-cooked diets which are more likely to be nutritionally inadequate than commercial
diets (Alenza et al., 1998; Freeman and Michel, 2001; Berschneider, 2002). Raw and home-cooked diets
were considered nutritionally deficient for the purposes of this study. In addition, the percentages of plant-
eating and non-plant-eating dogs whose owners supplemented their diet with vegetables or fruits were
determined. For plant-eating dogs, the proportion of owners who stated that their the dogs frequently
appeared ill before eating plants and the proportion of owners who stated that their dogs frequently vomited
within 1 h after eating plants were also determined. Because of the relatively small number of subjects, no
statistical evaluation of the data was planned.
2.3. Study 2
2.3.1. Study design
Study 2 was designed to collect descriptive information on dogs that eat plants and to acquire data that
would either support or allow rejection of the hypotheses given in the Introduction. The hypothesis that most
plant eating is associated with signs of illness (Hypothesis 1) would be confirmed if the majority (i.e. greater
than 50%) of owners reported that their dogs commonly appeared ill prior to eating plants. This hypothesis
could be rejected if a large proportion of dog owners reported that their dogs rarely seemed ill before eating
plants. We arbitrarily set the ‘‘large proportion’’ cut-off at 75%.
The hypothesis that most plant eating is followed within a few minutes by vomiting (Hypothesis 2)
would be confirmed if the majority (i.e. greater than 50%) of owners reported that their dogs commonly
vomited after eating plants. This hypothesis could be rejected if a large proportion of dog owners reported
that their dogs rarely vomited after eating plants. Again, we arbitrarily set this cut-off point at 75%.
The hypothesis that plant eating reflects a dietary deficiency (Hypothesis 3) would be supported if
owners feeding their dogs a nutritionally deficient diet reported that their dogs consumed plants more
frequently than dogs fed a nutritionally balanced diet. The hypothesis would not be supported by data
revealing no relationship between dietary quality and plant-eating frequency.
The hypothesis that plant eating is a normal behaviour (Hypothesis 4) was not tested directly but would
be supported by the process of exclusion, i.e. if Hypotheses 1–3 were rejected.
2.3.2. Survey design
To obtain the large sample necessary to collect general information on the frequency of plant eating, type
of plants eaten and the effects of age, sex, gonadal status and breed group on plant eating, we turned to an
internet survey where thousands of responses can be expected. Similar surveys have been used in a variety of
data-based behavioural and medical investigations (e.g. Janson and Wist, 2004; McCobb et al., 2001; Kim
et al., 2001; Gobar and Kass, 2002) and have been shown to provide data of a quality comparable to
traditional survey methods (Gosling et al., 2004; Reips, 2002; Rhodes et al., 2003). Since it was not used to
determine the prevalence of plant-eating behaviour, the internet survey utilised in Study 2 was specifically
targeted toward owners of plant-eating dogs.
Using an established web-based program (Survey Suite, University of VA, USA), a survey was designed
for internet distribution that would take approximately 20 min to complete. The survey included questions
dealing with the description of the dog and its environment, information on plant eating and the dog’s diet
and medical history (Table 2). For households possessing more than one dog that ate plants, respondents
were asked to answer the questions for the dog that ate plants most frequently. Questions were presented in
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132 123

Author's personal copy
either multiple-choice or short-answer format. With regard to frequency of plant eating, respondents were
asked to choose from the following: (1) daily; (2) weekly, defined as at least once a week but less than daily;
(3) monthly, defined as at least once a month but less than weekly; (4) less than once a month. Respondents
were also asked whether their dog frequently appeared ill before, or vomited after, eating plants. The survey
focused only on the presence or absence of illness or vomiting and did not ask owners to describe their nature
or severity.
We used an internet search engine to find websites most likely to be visited by dog owners interested in
plant-eating behaviour and managers of these websites were asked to post a link to the survey. Moderators of
internet ‘‘chat groups’’ related to dogs and dog behaviour were also contacted and asked to inform their
group members of the survey. The survey was posted for 3 weeks before being closed.
Survey responses were screened and assessed for the inclusion criteria; only those surveys in which dog
owners stated they spent at least 6 h per day in the company of their dogs and had observed plant eating on at
least 10 occasions during the time they owned their dogs were included. These restrictions ensured that the
owners had an adequate opportunity to observe their dogs’ plant-eating behaviour on several occasions and
could accurately determine how their dogs most commonly behaved before and after eating plants.
Exclusion criteria included medical problems or current medications that might induce plant eating or
vomiting (see Section 2.2). Because we were also trying to determine whether dogs consumed plants dueto a
dietary deficiency, dogs that were reported to only chew or mouth plants, without appearing to ingest plants,
were excluded from the study. Duplicate or incomplete surveys were not utilized. The respondents were not
aware that these criteria were used for exclusion.
2.3.3. Data analysis
Analyses for the dependent variables, which were considered on a dichotomous basis, involved the
following: frequency of plant eating (once a week or more versus less than once a week); eating primarily
grass versus non-grass; frequently appearing ill versus rarely appearing ill before eating plants and
frequently vomiting versus rarely vomiting after eating plants. The analysis involved an initial screening
with a stepwise logistic regression to select those independent variables that appeared to act as predictors of
differences (P<0.05) in the dependent variables (Hosmer and Lemeshow, 1989; SAS, V.9.1, logistic
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132124
Table 2
Study 2, summary of questions presented to owners of plant-eating dogs in an internet survey (not inclusive of all
questions)
Section
Description of dog Sex, gonadal status, major breed group, age
Duration in household
Time spend with dog daily
Information on plant eating Number times observed eating plants
Frequency of plant eating
Age when plant eating first noticed
Typically only chews or eats plants
Where dog has access to plants
Type of plants eaten
Type of plant most frequently eaten
Sign of illness before and after plant eating
Vomiting within 1 h after eating plants
Diet fed Home cooked
Raw
Commercial kibble or canned
Medical history Current medical history
Current medications

Author's personal copy
procedure). The independent variables were: age, categorized as <1 year, 1–2 years, 3–9 years or >9 years;
sex and gonadal status, categorized as male-neutered, male-intact, female-spayed or female-intact; breed
group, categorized as mixed, sporting, herding, hound, terrier, working, non-sporting or toy and diet,
categorized as complete by virtue of being at least 51% commercial, AAFCO-certified dog food or
incomplete by virtue of being 51% or greater home-cooked or raw food. After the initial screening, an
explicit model was fitted (without any model selection) with multiple comparisons to determine which
independent variables significantly predicted dependent variables (SAS, V.9.1, genmod procedure).
Significance was set at the a= 0.05 level. When a level of significance is reported, that level was derived
from the multiple comparisons based on the explicit model and when a non-significant level is reported, the
value was taken from the initial screening analysis.
3. Results
3.1. Pilot study
Twenty-five pilot surveys were returned from respondents, all of whom stated that their dogs
ate grass. None of the dogs exhibited signs of illness prior to eating grass. Two dogs (8%) were
reported to regularly vomit after eating grass, 13 (52%) to occasionally vomit and 10 (40%) had
never been observed vomiting afterwards.
3.2. Study 1
Of 98 surveys given to dog owners in Study 1, 78 (80%) were completed and returned. Of
these, 47 (60%) met the criteria for inclusion. Ten of the 47 owners (21%) reported that they had
never seen their dogs eat plants and had not noticed grass or other plant material in their faeces or
vomit. The remaining 79% reported that they had seen their dogs eat grass at least once or noticed
grass in their dogs’ faeces or vomit.
With regard to the 37 dogs reported to eat plants, the number of plant-eating episodes observed
throughout the duration of ownership was fewer than 10 for 12 dogs (32%), over 10 but less than
100 for 18 dogs (49%) and over 100 for 7 dogs (19%). Grass was the plant primarily eaten by 35
dogs (95%).
Thirty-three owners answered the questions regarding their dogs’ behaviour before and after
eating plants. Of these, four dogs (12%) were reported to regularly appear ill before eating plants
and six (18%) were reported to regularly vomit afterwards. Three (50%) of the regularly vomiting
dogs were reported to also appear ill before eating plants. None of the non-plant-eating dogs
consumed a diet supplemented with vegetables or fruits, while diets of 10 of 37 plant-eating dogs
(27%) were supplemented with some vegetables or fruits by their owner.
3.3. Study 2
Of 3340 surveys returned, 1571 (47%) met the inclusion criteria. The survey data contained
197 intact males, 560 castrated males, 219 intact females and 595 spayed females. Data on the
estimated time spent with the dog, total number of plant-eating occasions observed, frequency
of plant eating, age of the dog, duration of the dog in the household and where the dog had
access to plants are presented in Table 3. The estimated total of number of plant-eating
occasions observed by the respondents over the duration of ownership was evenly split
between 10 and 100 times and over 100 times; the reported frequency of plant eating was daily
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132 125

Author's personal copy
or weekly for 68% of respondents and monthly or less by 32%; almost all dogs (95%) had
access to plants in the yard with the remaining dogs having access to plants on walks or in
other locations.
The distribution of plant type reported eaten most often is shown in Fig. 1. Grass (long, short
or clippings) was listed by 79% of respondents as the plant they observed their dog consume most
frequently; non-grass plants eaten included mainly berries, sticks and leaves. Almost all
respondents (98%) mentioned that grass was consumed on at least some occasions. Several
respondents commented that their dogs appeared very ‘‘selective’’ in regard to the type or portion
of grass eaten.
Data relating to frequency of plant eating and plant type eaten as a function of sex and gonadal
status, breed group and dietary quality, are presented in Table 4. Most owners (1439/1571; 92%)
said that their dogs rarely showed signs of illness before or after eating plants. Vomiting regularly
within 1 h after eating plants was also infrequent; 78% (1225/1571) of respondents stated that
their dogs rarely vomited after eating plants.
The only significant predictor of frequency of plant eating was age of the dog. In the follow-up
multiple comparisons, frequency of plant eating was significantly inversely related to age group
(P<0.001; Fig. 2A). In the initial analysis, sex/gonadal status, breed group and diet were not
related to frequency of plant eating (P= 0.194, 0.130 and 0.067, respectively).
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132126
Table 3
Study 2, percentage of owners reporting: the length of time spent with their dog per day; the total number of times they
observed their dog eating plants; the frequency of plant eating by their dog; their dog’s age; their duration of ownership;
where their dog had access to plants (N= 1571 surveys)
Estimated time spent with dog per day (h)
6–9 10–12 13–15 >15
25 18 16 41
Estimated total plant-eating occasions observed
10–20 21–50 51–100 >100
11 21 18 50
Estimated frequency of plant eating by dogs
Daily Weekly Monthly <Monthly
24 44 22 10
Age of dog covered by survey (years)
<1 1–2 3–9 >9
4235518
Duration of dog in household (years)
<1 1–3 4–6 >6
8351938
Access to plants by dog
Yard Walks
95 62

Author's personal copy
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132 127
Fig. 1. Percentages of dogs that primarily consumed grass or non-grass plant types as reported by their owners in an
internet survey (Study 2) (N= 1571). Plant types mentioned by fewer than 15 (1%) respondents are represented as other.
Table 4
Study 2, plant-eating frequency, expressed as daily or weekly (1 week) or monthly (<1/week) and type of plant eaten as
a function of sex/gonadal status, breed group and diet (N= 1571)
NFrequency of plant eating Type of plant eaten
1 per week; N(%) <1 per week; N(%) Grass; N(%) Non-grass; N(%)
Sex/gonadal status
Male-intact 197 127 (64) 70 (36) 143 (73) 54 (27)
Male-neutered 560 317 (66) 189 (34) 438 (78) 122 (22)
Female-intact 219 155 (71) 64 (29) 163 (74) 56 (26)
Female-spayed 595 411 (69) 184 (31) 483 (81) 112 (19)
Breed
Mixed breed 172 113 (66) 59 (34) 143 (84) 29 (16)
Sporting group 309 222 (72) 87 (28) 238 (77) 71 (23)
Herding group 296 194 (66) 102 (34) 231 (78) 65 (22)
Hound group 131 79 (60) 52 (40) 106 (81) 25 (19)
Terrier group 153 94 (61) 59 (39) 113 (74) 40 (26)
Working group 320 229 (72) 91 (28) 252 (79) 68 (21)
Non-sporting group 100 72 (72) 28 (28) 78 (78) 22 (22)
Toy group 90 61 (68) 29 (32) 66 (73) 24 (27)
Diet
Commercial 1245 831 (67) 414 (33) 976 (78) 269 (22)
Home/raw 326 233 (71) 93 (29) 251 (77) 72 (23)

Author's personal copy
A significant predictor with regard to type of plant eaten (grass versus non-grass) was age. In
multiple comparisons there was a significant increasing tendency for older dogs to primarily eat
grass (P= 0.003; Fig. 2B). Interestingly, gonadally intact dogs were more likely to eat grass than
neutered dogs, even after controlling for age (P= 0.005). However, among both intact and
neutered females there was an increased tendency to eat non-grass plants in the oldest age group
(P= 0.024). Breed group and diet were not significant in the initial analysis with regard to the
type of plant eaten (P= 0.704 and 0.271, respectively).
Although frequently showing signs of illness before eating plants was seen in only 8% of dogs
overall, the initial analysis revealed an effect of age, with an increasing tendency to show illness
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132128
Fig. 2. Comparison of different age groups of dogs in an internet survey (Study 2) with regard to: (A) percent exhibiting
daily or weekly plant consumption; (B) percent observed to consume primarily grass; (C) percent regularly showing signs
of illness before eating plants and (D) percent vomiting after eating plants.

Author's personal copy
with increasing age (P<0.001; Fig. 2C). No effect of diet, sex/gonadal status or breed group was
apparent (P= 0.796, 0.527 and 0.114, respectively).
While vomiting frequently after eating plants was seen in only 22% of dogs, the initial
screening analysis revealed that age, breed group and diet were significant predictors of
vomiting. The multiple comparisons revealed that the older the dog the more likely it was to
regularly vomit (P<0.001; Fig. 2D). Dogs in the hound and toy breed groups were more likely
to regularly vomit than dogs in the other breed groups (P= 0.0031). Additionally, dogs fed a
complete, balanced diet were more likely to regularly vomit than those fed a home-made diet
(P<0.001). Sex/gonadal status was not a significant predictor (P= 0.13). This analysis also
revealed that if dogs showed signs of illness before eating plants they were more likely to vomit
afterwards than dogs rarely showing signs of illness beforehand (P<0.001).
4. Discussion
Based on the finding in Study 1, that approximately 79% of healthy, well-cared-for dogs eat
plants, particularly grass, plant eating appears to be a common behaviour of domestic dogs. In
Study 2, involving only plant-eating dogs, almost all of the dogs were reported to eat grass on at
least some occasions. The fact that this behaviour is widespread and occurs in all domesticated
breed groups indicates that it likely serves some biological function.
Three hypotheses regarding the causes or consequences of plant eating by dogs were tested
directly. One hypothesis was that dogs eat plants when they feel ill; a second hypothesis was that
dogs frequently vomit after ingesting plant material. Neither hypothesis was supported by the
data obtained from these studies. Data from three diverse populations of dog owners, 25
veterinary students (Pilot Survey), 47 screened clients of the outpatient clinic with apparently
healthy dogs (Study 1) and 1571 screened respondents to an internet survey (Study 2), revealed
that most dogs appear normal before eating plants and do not regularly vomit afterwards. The
magnitude of the data was convincing, with 100, 88 and 92% of owners from the Pilot Survey,
Study 1 and Study 2, respectively, reporting that their dogs appeared normal before eating plants,
and 87, 82 and 78%, respectfully, reporting no regular vomiting afterwards.
Given that a few dogs do appear to be ill prior to eating plants and that vomiting does
sometimes follow plant eating, it is possible that gastric and/or intestinal discomfort sometimes
evokes plant eating and that plant consumption may cause enough gastric irritation to lead to
vomiting. This assumption is consistent with the finding from Study 2 that if dogs showed signs
of illness before eating plants they were more likely to vomit afterwards than dogs that did not
show signs of illness beforehand. While we attempted to exclude dogs with known medical
problems, it is reasonable to assume that some dogs included in the study had gastrointestinal
distress or illness without their owners being aware of the problem.
A third hypothesis was that plant eating reflected a dietary deficiency of some type. No
relationship was found between type of diet and frequency of plant eating, or type of plant eaten
in Study 2. In Study 1, no dogs in the non-plant-eating group were supplemented with vegetables
or fruit, whereas 10 of 37 (27%) plant-eating dogs’ diets were supplemented with some type of
vegetable or fruit, suggesting that plant eating is not evoked by absence of plant material in the
diet and that offering plant material in the diet does not prevent plant eating from occurring.
While a detailed study investigating deficiencies in specific elements such as fibre, vitamins or
minerals is needed, a provisional conclusion is that the type of diet does not affect the tendency to
engage in plant eating. At this time, we cannot discern a reason for dogs fed a commercial diet to
be more likely to vomit after eating plants than dogs on home-cooked or raw food diets. However,
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132 129

Author's personal copy
the findings that most dogs eat plants to some degree and plant eating is not associated with
dietary deficiency, signs of illness or vomiting, lends support to Hypothesis 4, that plant eating is
mostly a normal behaviour.
Study 2 found that the majority of plant-eating dogs of all breeds consume grass most often
and at roughly the same frequency, despite the fact that dog breeds tend to exhibit different food
preferences and feeding patterns (Thorne, 1995). Grass eating is also widespread amongst
wolves; grass was the primary plant type found in timber wolf scat (2–%; Mech, 1966) and the
second most common plant type found in the scat of Latvian wolves, with conifer needles being
the most frequent (Andersone, 1998). This points to an evolutionary background where plant
eating was maintained in wild canids through natural selection and was conserved in dogs during
domestication.
Natural selection would have favoured periodic plant eating if it conferred a selective
advantage on canids displaying the behaviour. One possibility is that ingestion of non-nutritive
plant material provides a mechanical means of expelling intestinal parasites. As mentioned in the
introduction, intestinal worms have been observed wrapped around blades of grass in wolf scats
(Murie, 1944).
Non-nutritional plant eating has been studied in detail in wild chimpanzees (Pan troglodytes),
which consume whole leaves from a variety of plant species, particularly during the time of year
when intestinal parasitism risk is high (Wrangham, 1995). The observation that whole leaves pass
through the intestinal tract intact, sometimes along with intestinal nematodes (Huffman et al.,
1996), together with evidence of increased intestinal motility from the swallowing of whole
leaves, has led investigators to propose that whole leaves have a purging effect on intestinal
nematodes (Huffman and Canton, 2001). Controlled, laboratory studies could be performed with
domestic dogs to determine whether or not plant eating reduces intestinal parasites in this species.
Another perspective on this issue comes from observations of parasite loads in young canids
compared with older canids. Coyotes less than 1 year of age (Franson et al., 1978) and young
dogs, especially those without close medical oversight (Fontanarrosa et al., 2006; Ramirez-
Barrios et al., 2004; Kirkpatrick, 1988), have a higher prevalence of intestinal parasites than
adults. If the predisposition towards plant eating in domestic dogs stems from an innate behaviour
in ancestral canids for periodic expulsion of intestinal parasites, one would expect younger
canids, including dogs, to engage in the behaviour more frequently than older dogs. This
perspective is supported by the finding in Study 2 of an inverse relationship between age and
frequency of plant eating. The data also reveal that younger dogs are less likely to appear ill
before, or vomit after, eating plants, a finding consistent with the concept of ongoing control of
parasites. The finding that some dogs do not eat plants at all (Study 1) could be explained on the
basis of relaxation of selection for plant-eating behaviour, reflecting dog-owner use of
prophylactic treatment for intestinal parasites.
5. Conclusions
Plant eating, especially grass eating, commonly occurs among all breed groups of domestic
dogs and appears to be a normal behaviour unassociated with illness, vomiting or dietary
deficiency. Although this study only investigated plant-eating behaviour in domestic dogs, plant
eating has also been observed in wolves and other wild canids. Since the behaviour appears to
have been conserved through domestication and artificial selection, it seems likely that it serves
some biological purpose. Further studies are needed to determine what, if any, selective
advantage plant eating may offer, such as an increase in gastrointestinal motility and/or the
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132130

Author's personal copy
expulsion of intestinal parasites. Because non-nutritional plant eating with a health-related
function is a type of herbal medication (Hart, 2005), if a health maintenance function is
eventually demonstrated for plant eating by canids, the behaviour would represent the first
documented example of herbal medicine in carnivores.
Acknowledgements
Supported by the University of California, Davis, Center for Companion Animal Health
(allocation #03-65-F). The authors thank Stanley L. Marks, specialist in clinical nutrition, for
providing assistance in formulating survey questions and Neil Willits of the Department of
Statistics for statistical advice and data processing.
References
Alenza, D.P., Rutteman, G.R., Pena, L., Beynen, A.C., Cuesta, P., 1998. Relation between habitual diet and canine
mammary tumors in a case–ontrol study. J. Vet. Intern. Med. 12, 132–139.
Andersone, Z., 1998. Summer nutrition of wolf (Canis lupus) in the Slitere Nature Reserve, Latvia. Proc. Latvian Acad.
Sci. 52, 79–80.
Andersone, Z., Ozolins, J., 2004. Food habits of wolves Canis lupus in Latvia. Acta Theriologica 49, 357–367.
Beaver, B.L., 1981. Grass eating by carnivores. Vet. Med. Small Anim. Clin. 76, 968–969.
Berschneider, H.M., 2002. Alternative diets. Clin. Tech. Small Anim. Pract. 17, 1–5.
Fontanarrosa, M.F., Vezzani, D., Basabe, J., Eiras, D.F., 2006. An epidemiological study of gastrointestinal parasites of
dogs from Southern Greater Buenos Aires (Argentina): age, gender, bred, mixed infections, and seasonal and spatial
patterns. Vet. Parasitol. 136, 283–295.
Franson, J.C., Jorgenson, R.D., Boggess, E.K., Greve, J.H., 1978. Gastrointestinal parasitism of Iowa coyotes in relation to
age. J. Parasitol. 64, 303–305.
Freeman, L.M., Michel, K.E., 2001. Evaluation of raw food diets for dogs. J. Am. Vet. Med. Assoc. 218, 705–709.
Gobar, G.M., Kass, P.H., 2002. World wide web-based survey of vaccination practices, postvaccinational reactions, and
vaccine site-associated sarcomas in cats. J. Am. Vet. Med. Assoc. 220, 1425–1576.
Gosling, S.D., Vazire, S., Srivastava, S., John, O.P., 2004. Should we trust web-based studies? A comparative analysis of
six preconceptions about internet questionnaires. Am. Psychol. 59, 93–104.
Hart, B.L., 1988. Biological basis of the behaviour of sick animals. Neurosci. Biobehav. Rev. 12, 123–137.
Hart, B.L., 2005. The evolution of herbal medicine: behavioural perspectives. Anim. Behav. 70, 975–989.
Hosmer, D.W., Lemeshow, S., 1989. Applied Logistic Regression. Wiley, New York, pp. 106–112.
Houpt, K.A., 2005. Ingestive behavior: food and water intake. In: Domestic Animal Behavior for Veterinarians and
Animal Scientists, fourth ed. Blackwell Publishing, Ames, IA, pp. 309–359.
Huffman, M.A., Canton, J.M., 2001. Self-induced increase of gut motility and the control of parasitic infections in wild
chimpanzees. Int. J. Primatol. 22, 329–346.
Huffman, M.A., Page, J.E., Sukhdeo, M.V.K., Gotoh, S., Kalunde, M.S., Chandrasiri, T., Towers, G.H.N., 1996. Leaf-
swallowing by chimpanzees: a behavioural adaptation for the control of strongyle nematode infections. Int. J.
Primatol. 17, 475–503.
Janson, C., Wist, M., 2004. An internet survey of asthma treatment. J. Asthma 41, 49–55.
Kim, H.L., Gerber, G.S., Patel, R.V., Hollowell, C.M., Bales, G.T., 2001. Practice patterns in the teatment of female
urinary incontinence: a postal and internet survey. Urology 57, 45–48.
Kirkpatrick, C.E., 1988. Epizootiology of endoparasitic infections in pet dogs and cats presented to a veterinary teaching
hospital. Vet. Parasitol. 30, 113–124.
Lindsay, S.R., 2001. Appetitive and elimination problems. Handbook of Applied Dog Behavior and Training. Etiology
and Assessment of Behavior Problems, vol. 2. Iowa State University Press, Ames, IA, pp. 273–299.
Mech, L.D., 1966. Results—the Timber wolf and its ecology. Fauna of the National Park of the United States: The Wolves
of Isle Royale. National Park Service. 3 September 2004. <http://www.cr.nps.gov/history/online_books/fauna7/
fauna5g.htm>.
McCobb, E.C., Brown, E.A., Damiani, K., Dodman, N.H., 2001. Thunderstorm phobia in dogs: an internet survey of 69
cases. J. Am. Anim. Hosp. Assoc. 37, 319–324.
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132 131

Author's personal copy
Murie, A., 1944. Food habits of wolves. Fauna of the National Parks of the United States: The Wolves of MountMcKinley.
National Park Service. 29 November 2004. <http://www.cr.nps.gov/history/online_books/fauna5/fauna.htm>.
Overall, K.L., 1997. Miscellaneous behavioral problems: emphasis on management. In: Clincal Behavioral Medicine for
Small Animals, Mosby-Year Book Inc., St. Louis, MO, pp. 251–273.
Papageorgiou, N., Vlachos, C., Sfougaris, A., Tsachalidis, E., 1994. Status and diet of wolves in Greece. Acta
Theriologica 39, 411–416.
Plumb, D.C., 2002. Veterinary Drug Handbook, fourth ed. Pharma Vet Publishing, White Bear Lake, MN, pp. 108, 680,
715.
Ramirez-Barrios, R.A., Barboza-Mena, G., Munoz, J., Angulo-Cubillan, F., Hernandez, E., Gonzalez, F., Escalona, F.,
2004. Prevalence of intestinal parasitesin dogs under veterinary care in Maracaibo, Venezuela. Vet. Parasitol. 121, 11–
20.
Reips, U., 2002. Standards for internet-based experimenting. Exp. Psychol. 49, 243–256.
Rhodes, S.D., Bowie, D.A., Hergenrather, K.C., 2003. Collecting behavioural data using the world wide web:
considerations for researchers. J. Epidemiol. Community Health 57, 68–73.
Stahler, D.R., Smith, D.W., Guernsey, D.S., 2006. Foraging and feeding ecology of the grey wolf (Canis lupus): lessons
from Yellowstone National Park, Wyoming, USA. J. Nutr. 136, 1923S–1926S.
Thorne, C., 1995. Feeding behaviour of domestic dogs and the role of experience. In: Serpell, J. (Ed.), The Domestic Dog:
Its Evolution, Behaviour and Interactions with People. Cambridge University Press, Cambridge, UK, pp. 103–114.
Wrangham, R., 1995. Relationship of chimpanzee leaf-swallowing to a tapeworm infection. Am. J. Primatol. 37, 297–
303.
K.L.C. Sueda et al. / Applied Animal Behaviour Science 111 (2008) 120–132132