Article Text

Download PDFPDF

Dog-human and dog-dog interactions of 260 dog-owning households in a community in Cheshire
  1. C. Westgarth, BSc1,
  2. G. L. Pinchbeck, BVSc, CertES, PhD, DipEVCPH, MRCVS1,
  3. J. W. S. Bradshaw, BA, PhD3,
  4. S. Dawson, BVMS, PhD, MRCVS1,
  5. R. M. Gaskell, BVSc, PhD, MRCVS2 and
  6. R. M. Christley, BVSc, DipVetClinStud, MVetClinStud, PhD, DipEVCPH, MRCVS1
  1. 1 Department of Veterinary Clinical Science
  2. 2 Department of Veterinary Pathology, University of Liverpool, Leahurst, Neston, Cheshire CH64 7TE
  3. 3 Anthrozoology Institute, Department of Clinical Veterinary Science, University of Bristol, Langford House, Langford, North Somerset BS40 5DU


This study investigated the nature and frequency of the contacts that occur between dogs, and between dogs and people, by means of a questionnaire survey of 260 dog-owning households in a community in Cheshire, uk. The contacts were highly variable and were affected by the size, sex and age of the dog, individual dog behaviours, human behaviours and human preferences in the management of the dog. A number of situations were identified that may be important in relation to zoonoses, including sleeping areas, playing behaviours, greeting behaviours, food sources, walking, disposal of faeces, veterinary preventive treatment and general hygiene.

View Full Text

Statistics from

THERE are approximately 6·5 million dogs owned in the uk (Pet Food Manufacturers Association [PFMA] 2004), which equates to approximately one dog for every nine people and every four households (Office for National Statistics 2007). Dog ownership is associated with many benefits for people, including companionship and physiological and psychological health (Katcher 1981, Katcher and Friedmann 1982, Friedmann 1995, Headey 2003, McNicholas and others 2005), but there are also negative aspects, recently highlighted by Jackson (2005), including dog bites, public nuisance, and risks to public health from zoonoses. At least 30 to 40 diseases of companion animals are transmissible to human beings (Greene and Levy 2006), including parasitic, bacterial, fungal and viral diseases (Geffray and Paris 2001); examples in dogs in the uk include Campylobacter and Salmonella species, Toxocara canis (Tan 1997, Greene and Levy 2006) and meticillin-resistant Staphylococcus aureus (mrsa) (Cefai and others 1994, Greene and Levy 2006). In Australia, the ownership of pet puppies has been reported to be a risk factor in campylobacteriosis in young children (Tenkate and Stafford 2001), and exposure to diarrhoeic animals has been associated with a threefold increase in the risk of Campylobacter jejuni/Campylobacter coli enteritis in human beings (Saeed and others 1993).

Little is known about the nature and frequency of contacts between pet dogs and their owners or other people, but in order to assess the risk of disease transmission from pets it is important that such factors are evaluated (Wieland and others 2005). The nature of the pathogen and its mode of transmission is also important. People may be exposed to zoonoses either by direct contact, through biting, licking, scratching, urine spray, sneezing or coughing, or handling the dog or its faeces or reproductive discharges, or by indirect contact through contaminated bedding, food, water, or bites from an arthropod vector (Robinson and Pugh 2002). Activities involving close contact between dogs and people include sleeping, playing, eating, greeting, the disposal of faeces, and general physical contact through tokens of affection such as cuddling and stroking.

Similarly, little is known about the contacts between dogs that could transmit an infection through a population, for example, during interactions between dogs while they are out walking or through indirect contact during the investigation of other dogs' excreta. Bradshaw and Lea (1992) characterised the sequences of behaviour that occur during interactions between dogs in popular walking areas, but did not measure the frequency of the interactions, an important factor affecting the risk of transmitting a pathogen. Opportunities for the transmission of pathogens between dogs will be affected by human preferences, such as the frequency of walks and the times the dog is allowed off the lead, in addition to the types of behaviour of individual dogs.

A combination of human and dog behaviours determines where a dog goes and what it does. Studying interactions is not only of zoonotic importance, but of behavioural, welfare, psychological and social interest. The aim of this study was to investigate and quantify the direct and indirect contacts between dogs, and between dogs and people, in households in a community in Cheshire. The contacts that were considered likely to be associated with a risk of the transmission of pathogens of zoonotic importance were a particular focus of the study.


In a doorstep survey of 1278 households in a community in Cheshire (Westgarth and others 2007), the owners of 327 dogs were identified and recruited into the study. The area and its surroundings include medium- and low-density housing, public amenities such as parks, sports fields and a wildlife reserve, and agricultural land. Basic demographic information was collected for each dog when it was recruited. The 260 dog-owning households were asked to complete a questionnaire, containing multiple-choice and open-ended questions, which had been pre-tested, revised and then piloted on 12 dog-owning households in a nearby area. The questionnaire was designed using an automated content-capture system (teleform version 9.1; Verity Software) and is available from the authors on request. The questions were chosen specifically to investigate behaviours with the potential to transmit zoonotic pathogens, and covered a wide variety of topics: where the dog sleeps and is allowed access, the games it plays, its health, diet, walk frequency, and behaviour when greeting people and other dogs. The questionnaires were returned between July and October 2005. Households that had not returned their questionnaire after two weeks were sent a reminder postcard, and if they had still not returned it after another four to six weeks they were sent another copy of the questionnaire. Incentives to participate included money-off vouchers for dog food and a local boarding kennels, which were provided after the questionnaire had been returned.

The data were managed in a Microsoft Access database and analysed using Minitab release 14.2, spss 13.0 for Windows, and Microsoft Excel. Chi-squared tests were used to investigate associations between answers to questions with factors such as the sex and size of the dog. The frequency of many of the variables was estimated as either ‘never’, ‘rarely’, ‘sometimes’ or ‘often’. This ordinal outcome was assessed using ordinal logistic regression analysis for the continuous variable age, with the lowest category as ‘never’. The association between the ordinal variables ‘likely to greet dogs’ and ‘likely to greet people’ was tested using the Gamma statistic (Siegal and Castellan 1988). Questions about the respondents' views on the positioning and emptying of dog-waste bins in the area were used to introduce the subject of picking up faeces. In the households in which one main person performed the dog duties, this person would have been asked to complete the questionnaire, and the responses to the questions on picking up faeces were compared for male and female owners by chi-squared tests.


Of the dogs initially recruited into the study, 78 per cent were of a named breed, as opposed to crossbreeds or mixed breeds. Gundogs were the most popular uk Kennel Club category (25 per cent), followed by mixed or crossbreeds (23 per cent). Labradors were the most popular of the individual breeds (15 per cent), followed by Jack Russell terriers (13 per cent). Approximately equal numbers of the dogs were small, medium or large, with very few toy or giant breeds. The mean (sd) approximate or known age (317 dogs) was 6·5 (3·9) years, with a maximum of 19 years. There were 173 females and 154 males. Fifty-three per cent of males had been neutered, compared with 73 per cent of females; the odds of a female being neutered was 2·3 (95 per cent confidence interval [ci] 1·4 to 3·7) times greater than for a male. Fifty-nine per cent of the dogs had been acquired by the current owner from the person who bred the dog.

Completed questionnaires were returned for 279 (85 per cent) of the 327 dogs recruited into the study. Twelve per cent of households with either one or two dogs did not respond, compared with 43 per cent of three-dog households.

Dog-human and dog-dog contacts

Seventy-nine per cent of the dogs were fed in the kitchen. The most popular food was dry complete commercial dog food, though one dog was fed raw meat as part of its main diet. Eighty-three per cent of the dogs were never fed raw meat. Commercial dog treats were fed to 85 per cent of the dogs ‘sometimes’ or ‘often’. Human food titbits were fed to dogs ‘sometimes’ or ‘often’ from the hand (62 per cent), in the dog's bowl (69 per cent), straight from the plate (11 per cent) or off the floor (37 per cent). Six per cent of the dogs ‘sometimes’ or ‘often’ found and ate raw carcases, 25 per cent rolled in them, and 6 per cent ‘sometimes’ found and ate dog faeces. Eighty-four per cent of the dogs had visited a veterinary surgeon in the past year, 4 per cent because of vomiting and/or diarrhoea. Sixty-two per cent of the dogs had been vaccinated in the past year. Flea treatment had been given in the past three months to 53 per cent, and worming treatment to 58 per cent of the dogs.

The most common sleeping place for the dog was in the kitchen (42 per cent ‘always’ or ‘often’); 19 per cent slept on the bedroom floor ‘always’ or ‘often’ and 14 per cent on a human bed. During the day, the living area was the most popular place for the dogs to rest, with 60 per cent being there ‘always’ or ‘often’. Only 4 per cent of the dogs slept outside ‘always’ or ‘often’, but 29 per cent spent time there during the day ‘always’ or ‘often’. When the owner was at home 56 per cent of the dogs were allowed anywhere in the house, but when the dog was on its own it was common to be restricted to the kitchen (24 per cent) rather than allowed everywhere (20 per cent). Fifty-two per cent of the dogs were reported to lie on furniture and 45 per cent on a person's lap ‘sometimes’ or ‘often’. Smaller dogs were significantly more likely to lie on furniture or on a person's lap (Table 1) and younger dogs were also more likely to lie on a person's lap (Table 2).

When interacting with household members, sniffing or nudging with the nose, jumping up, and licking hands were commonly reported to occur ‘sometimes’ or ‘often’ (Fig 1). The neutered females tended to show sniffing or nudging behaviour more often, but the group sizes were small (Table 3). Smaller dogs were reported to jump up more often than larger dogs (Table 1). Ordinal logistic regression identified jumping up, licking faces and licking the hands of household members as significantly more common in younger dogs (Table 2).


Frequency with which 279 dogs in Cheshire showed different types of behaviour when interacting with household members or greeting visitors

The most common type of game played with the dog was to fetch a ball or other object (77 per cent ‘sometimes’ or ‘often’). Larger dogs were reported to play fetch more often than smaller dogs (Table 1), although the difference was significant only if the respondents not reporting playing fetch but reporting playing games other than fetch were categorised as ‘never’ for fetch. There were significant differences in the frequency of fetch games by sex/neuter status (Table 3), with entire females reportedly playing more fetch games. Tug-of-war was more likely to be played by smaller dogs than larger dogs (Table 1). Ordinal logistic regression showed that the younger dogs played all the games more frequently (Table 2).

The most common frequency reported for adult visitors to the house was several times a week (42 per cent), whereas for children it was once a week (20 per cent) or once every several months (21 per cent), but 23 per cent of the house-holds had adult visitors every day and 12 per cent had child visitors every day. The most common behaviours reported when greeting visitors were sniffing or nudging with the nose, jumping up and barking (Fig 1). Smaller dogs were reported to jump up at visitors more often than larger dogs (Table 1), and there was some evidence that entire dogs jumped up more often than neutered dogs (Table 3). The age of the dog was also significantly associated with whether they were reported to jump up at visitors or lick visitors' faces (Table 2), with younger dogs more likely to exhibit these behaviours. Ten per cent of the dogs were reported to growl ‘sometimes’ or ‘often’ at visitors, compared with 6 per cent that growled at household members.

Dogs may make contact with other dogs and people when they are taken out of the house, on a walk or to other places. The most common situation reported was being taken to friends' or relatives' houses (23 per cent ‘once a week or more’ and 6 per cent ‘every day’). Ninety-three per cent of the dogs never visited training classes, 67 per cent never visited boarding kennels, and 67 per cent never visited grooming parlours. Thirty-seven per cent of the owners had taken their dog on holiday with them in the uk in the past year, but only one owner had taken their dog elsewhere in Europe. Most of the dogs were estimated by their owner to meet and interact with three to five people per day outside the household (Fig 2), and they met significantly more people at weekends than on weekdays (P=0·001). This trend was also observed when estimating the number of other dogs met and interacted with per day, with one to two being most common for weekdays and three to five at weekends (P=0·01) (Fig 2). Seventy-six per cent of the dogs ‘often’ or ‘sometimes’ interacted physically with people, and 76 per cent interacted with other dogs outside the home, and there was evidence of ‘gregarious dogs’ that tended to interact with both dogs and people (Gamma statistic value 0·39, P<0·001). Common behaviours reported as occurring ‘sometimes’ or ‘often’ when interacting with another dog included being playful (59 per cent), sniffing (81 per cent), ignoring (42 per cent) and aggression (24 per cent). There were significant differences between the frequencies of reports of sniffing behaviour by male and female dogs (Table 3), but a chisquared test for the trend was not significant. Younger dogs were more likely to play with other dogs and less likely to ignore other dogs (Table 2).


Estimates of the numbers of people and dogs met and interacted with daily during the week and at weekends by each of 279 dogs in Cheshire

Eighty-three per cent of the dogs were confined to a secure area and never roamed unattended away from the premises, and only 1 per cent were reported to be allowed to roam freely, although this may be an underestimate owing to the sensitive nature of the question. Thirty-two per cent of the dogs were walked twice a day, and 30 per cent were walked once a day. Only 3 per cent were never walked or walked less than once a week, but these included some young puppies and old dogs. Large or medium-sized dogs were walked more often than smaller dogs (Table 1) (chi-squared test for trend P=0·001).

Six per cent of the dogs were never on a lead when walked, but 14 per cent were never allowed off the lead. Of the dogs allowed off the lead, 67 per cent were always kept within sight. Most owners walked their dogs for between 16 minutes and one hour each time, and younger dogs were more likely to be walked for longer periods than older dogs (Table 2). Approximately half of the dogs were walked at regular times each day, with 06.00 to 09.00 being most common, but 09.00 to 12.00, 15.00 to 18.00, and 18.00 to 21.00 were also common times. Seventy-five per cent of the owners walked their dogs in the countryside, and 64 per cent walked them on the beaches and marshes next to the Dee Estuary; 69 per cent of the owners walked their dogs regularly in the same places. Twenty-seven per cent of the owners never took their dogs for walks out of the local area (in the car or by public transport), but 6 per cent did this every day, ranging up to 21 per cent less than once a month. Thirty-eight per cent of the owners reported never walking their dogs with a group of friends and their dogs, but 3 per cent did this every day; however, 92 per cent of owners noticed seeing the same people and their dog(s) ‘every day’, ‘often’ or ‘sometimes’ while walking their dog.

Five per cent of the dogs were reported to urinate in the house and 4 per cent were reported to defecate ‘sometimes’ or ‘often’. More of the owners did not answer the question about toileting in the house than the question about toileting in the garden or on walks, possibly owing to the sensitivity of the subject, and the rates of toileting in the house may therefore be underestimated. Sixty-two per cent of the owners removed faeces from the garden or yard every day, but 1 per cent never removed them and 3 per cent removed them less than once a week. Seventy per cent of the owners used plastic bags to dispose of faeces from the garden or yard, and 91 per cent did so when they were elsewhere or on a walk, but 42 per cent reported using a shovel in the garden or yard. Over 80 per cent of the respondents said that they always picked up any faeces passed by their dog while they were out walking in the street, park area or on a public path, but only just over 50 per cent did so when in the countryside (Fig 3). A significantly smaller proportion of male owners than females reported that they picked up faeces (Table 4). In a separate part of the questionnaire, the respondents were asked whether they washed their hands after picking up faeces, and 96 per cent said they did so ‘always’ or ‘usually’; 85 per cent reported that they always or usually washed their hands before eating, but only 58 per cent did so after touching a dog.


Frequencies with which 210 owners of dogs in Cheshire reported picking up their dog's faeces while on walks in different environments


This study investigated many of the common interactions between pet dogs and people that are relevant to dog welfare, the social benefits of owning a dog, and to the frequency of dog bites and public nuisance, particularly behaviours that may contribute to the transmission of zoonotic pathogens. Inside the house, a dog may be in close contact with house-hold members and any visitors, and it may interact with people and other dogs while outside. The reported dog-dog and dog-human contacts were highly variable and affected by the type of household, the sex, size and age of the dog, the behaviour of individual dogs and people, and owner preferences in the management of the dog. The situations that may be of particular concern include sleeping areas, greeting, playing, food sources, disposal of faeces, general hygiene, walking and veterinary preventive treatment.

There was a preference for placing the dogs in the kitchen to sleep, be fed, and be confined when the owner was out of the house. This may have been partly for hygiene reasons and ease of cleaning up urine or faeces, and partly because it restricts the dog's access to the rest of the house and valuable household items. However, the kitchen is where the house-hold's food is prepared, and this preference for placing dogs in the kitchen could be considered a risk for the transmission of zoonotic disease. Recent estimates suggest that 23 per cent of pet dogs in Norway (Sandberg and others 2002) and 41 per cent in Switzerland (Wieland and others 2005) carry Campylobacter species, and that 0·1 to 3·5 per cent of healthy dogs carry Salmonella species (Weber and others 1995, Fukata and others 2002, Hackett and Lappin 2003).

It has been suggested that pet food may be heavily contaminated with Salmonella species and may lead to the contamination of human food when dogs are fed in the kitchen (Christopher and others 1974, Pace and others 1977). However, these reports are old and the data may not apply to modern commercial pet foods. Most of the dogs were fed a commercial pet food, and were not deliberately fed raw meat, but a small number may have contacted raw meat by eating or rolling in carcases. Raw meat can be a source of many zoonotic pathogens such as Campylobacter and Salmonella species (LeJeune and Hancock 2001). A few of the dogs were reported to eat dog faeces, which could also be a source of infection. Eating faeces has previously been reported in only 0·2 per cent of dogs (Beaver 1994), but it is uncertain whether this figure represents only the dog's own faeces or included that of other dogs and animals. In the present study there was a considerably higher prevalence, but no reason could be identified for the difference.

Fourteen per cent of the dogs slept on a human bed ‘always’ or ‘often’, and approximately half of the dogs commonly lay on furniture or on a person's lap, behaviours that illustrate the often close physical and psychological nature of the relationship between dogs and people. It has been suggested that allowing such behaviours is likely to enhance the hierarchical status of the dog and may be associated with ‘alpha’ dog behavioural problems or aggression (Fisher 2001, Guy and others 2001), although Landsberg and others (2003) considered such behaviours unimportant. Substantial numbers of the owners reported such behaviours, but few of the dogs were said to growl at household members; however, this behaviour could have been under reported.

The close contact and sharing of beds or furniture could allow the transmission of zoonotic diseases or parasites such as fleas, especially by small and young dogs, which were more likely to lie on laps or furniture. In an ethological study in a small number of family homes, it was observed that small dogs were more likely to jump on to a person's lap than large dogs (Smith 1983). It has been reported that mrsa was transmitted to a person from an apparently healthy dog that routinely slept in a human bed and licked their faces (Manian 2003). It is not known how many healthy pet dogs in the uk are carriers of mrsa, but one small study of dogs in a veterinary referral hospital found a prevalence of 9 per cent, even though none of the dogs was being treated for mrsa infection (Loeffler and others 2005).

Common behaviours with household members and visitors such as sniffing and licking hands and faces could potentially transfer pathogens. Such behaviours, which were more common in young dogs, are often attention-seeking/care-soliciting gestures (Scott and Fuller 1965) and indicate the strength of the social bond of dogs with people. Small dogs were also reported to jump up more often than large dogs, as has been reported by Smith (1983).

Many games were reportedly played with the dogs, and they may transfer saliva and potential pathogens to the hands, particularly with the popular game of ‘fetch’. Rooney and others (2000) observed that medium-sized dogs were more likely to play games with their owners while out walking than large or small dogs, but they observed no such relationship in a different survey of owners on the games they played with their dogs. In the present study there were differences in the type of game played depending on the size of the dog, and the findings do not relate just to games played during walks. As in this study, fetch games have been reported to be played more often by large dogs while out walking (Messent 1983).

Another commonly reported activity that may transfer saliva was the giving of treats (commercial or human food titbits) from the hand. A small number of dogs were reported to eat directly from the plate. The majority of the owners reported that they always or usually washed their hands after touching a dog. Although the questionnaire was kept almost anonymous (traceable to household only by dog number), this is probably an overestimate owing to the owners' expectations of being judged by their answer.

Dog faeces are considered a nuisance as well as a potential health hazard. In addition to zoonotic bacteria, they may contain parasitic infections. For example, most puppies become infected with Toxocara canis in the first few weeks of life (Glickman 1990) and need to be dewormed regularly. Leaving faeces in the garden or yard may expose household members to risk for long periods. Open countryside was common dog-walking territory but was also where the owners were least likely to pick up their dog's faeces. Almost all of the respondents stated that they ‘always’ or ‘usually’ washed their hands after picking up faeces, but this would have been impossible in practice because most of the dogs commonly passed faeces while out walking. The majority of the respondents reported cleaning up after their dog; previous studies have observed 59 per cent of people cleaning up (Webley and Siviter 2000) but self-reporting gave much higher rates. In the present study the male owners appeared to be less likely to pick up after their dog, or alternatively they were more willing to admit leaving faeces.

Walking with a dog has been shown to facilitate social interactions, suggesting that there may be psychological, as well as physical, benefits to owners (Messent 1983). There was considerable variation in walking preferences, but a sub-stantial number of the dogs were walked on regular routes at regular times of day, and could therefore have had repeated opportunities for contact with the same other dogs and people. This idea is supported by the fact that most of the owners reported that they noticed the same dogs and owners on their walks. More opportunities for contact at weekends than on weekdays were reported. The majority of the dogs remained in sight on walks when off the lead, suggesting that they stay fairly close. Bekoff and Meaney (1997) reported that dogs off the lead generally travelled less than 2 to 5 m off trail for less than one to two minutes, although this is likely to vary with the environment. Many of the owners reported regularly taking their dog in the car or by public transport to walk outside their immediate local area, providing opportunities for dogs from different areas to mix and increase the risk of transmitting disease further afield. Some dogs were regularly taken to friends' or relatives' houses where they could have interacted with other dogs and people.

Diseases may be transmitted through and persist in the dog population as a result of interactions between dogs while out walking. In this study, the younger dogs could be considered to have been more at risk than the older dogs because they more often came into close contact with other dogs through behaviours such as playfulness, and were less likely to ignore other dogs. There was some evidence that entire males may have been more likely to sniff other dogs. Bradshaw and Lea (1992) also observed that when two dogs met the most common interactions were inspections of the head and anogenital areas, with males investigating the anogenital area more frequently than females, and they suggested that the sex of the dogs (and possibly whether or not it had been neutered) may affect the type of interactions.

In the previous year before the study, the majority of the dogs had been taken to a veterinary surgeon; veterinary surgeons could be an important source of information about zoonotic diseases. However, not all of the dogs were taken regularly and so other sources of information need to be considered. Most commonly, the owners had acquired a dog from the person who bred it, and so the breeders could also be a source of information for new owners. Just over half of the dogs were reported to have been recently treated for gastrointestinal worms, and the same proportion for fleas. Effective flea and worm treatment is important both for the welfare of the dog and considering the close contacts observed with people.

This study attempted to survey all the households in a defined area, and it therefore provides a less biased view of dog ownership than other studies in which dogs were recruited either through veterinary practices or through calls for volunteers, who would be likely to be enthusiasts. Not all of the dogs had visited a veterinary surgeon in the past year and one-third of dogs had not been vaccinated; as a result, many dogs would have been missed if the dogs had been recruited from veterinary practices in the same area. Bias due to not contacting a household was minimised by visiting at several times of day and on several days of the week, ensuring good contact rates for the initial recruitment of the dogs. Leaflets providing information about the study and incentives to participate, combined with local knowledge and community links with the local veterinary teaching hospital, may have contributed to the good response rate for both the initial interviews and the return of the postal questionnaires. There may have been some bias due to the different interests of the people who completed and returned the questionnaire and those who did not, in particular owing to the use of incentives. A smaller proportion of households with three dogs completed and returned their questionnaires than households with one or two dogs; the extra work involved to complete the questionnaires for three dogs may have been a deterrent. The study was made in a small, semi-rural community, and the results may therefore not apply generally to the wider uk population. However, the percentage of the population owning a dog was similar (24 per cent) to the 21 per cent reported for the uk in 2004 (PFMA 2004), suggesting that the results may be representative of similar populations elsewhere.

The results provide previously unrecorded information about dog ownership that may be typical of many communities in the uk and may be relevant to a number of disciplines. In terms of human health, dogs are not thought to be a major cause of zoonotic infections in comparison with food, but they may be an important risk, particularly for immunocompromised people, the very young and the elderly. In an initial study of the characteristics of dog ownership in this community, it was found that certain types of house-holds were more likely to own dogs than others (Westgarth and others 2007). In the present study, several situations that may facilitate the transmission of zoonotic diseases in the pet dog owning community were assessed, including playing, greeting, food sources, sleeping areas, walking, disposal of faeces, veterinary preventive treatment and general hygiene. The results may help to inform strategies for the control of zoonotic and other infectious diseases in dogs, and help to quantify the risks associated with dog ownership. However, the assessment of the risks involved, and the nature of any control measures, will depend on the nature of the pathogen and how it is transmitted.


The authors thank defra for funding the research and the people (and their dogs) who participated in the study.


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.