Human–dog interactions and behavioural responses of village dogs in coastal villages in Michoacán, Mexico

Applied Animal Behaviour Science 154 (2014) 57–65

Contents lists available at ScienceDirect

Applied Animal Behaviour Science journal homepage: www.elsevier.com/locate/applanim

Human–dog interactions and behavioural responses of village dogs in coastal villages in Michoacán, Mexico Eliza Ruiz-Izaguirre a,∗ , Karen (C.) H.A.M. Eilers a , Eddie A.M. Bokkers a , Alessia Ortolani b , Antonio Ortega-Pacheco c , Imke J.M. de Boer a a b c

Animal Production Systems Group, Wageningen University, Wageningen, the Netherlands Department of Animals in Science and Society, Faculty of Veterinary Medicine, Utrecht University, Utrecht, the Netherlands Faculty of Veterinary Medicine, Autonomous University of Yucatán, Mérida, Mexico

a r t i c l e

i n f o

Article history: Accepted 11 February 2014 Available online 20 February 2014

Keywords: Village dogs Canis familiaris Dog socialisation Human–dog play Behavioural tests Body condition

a b s t r a c t In Mexican villages, most households keep dogs that roam freely. Therefore, socialisation of village dogs occurs in a different context than that of companion dogs in developed countries. The objectives of this study were: (1) to assess village dogs’ behavioural responses towards familiar and unfamiliar humans, (2) to compare body condition of dogs living in a village with a seasonal trade in international tourism (IT-village) with dogs living in a village located in the vicinity of a sea-turtle nesting site (STN-village), and (3) to identify whether dog characteristics influence dog behaviour and body condition. Two coastal villages in Michoacán, Mexico, were selected as case study sites. Fiftynine dogs were initially visited, 35 of which were repeatedly visited during the high and low seasons for international tourism and sea-turtle nesting. Caregivers were interviewed regarding human–dog interactions, and dogs were behaviourally tested and rated for body condition. Behavioural indicators were: (1) the dog’s qualitative response to a caregiver’s call and (2) the dog’s willingness to approach an unfamiliar human. Additionally, a dog census per village was conducted to ascertain the dog population structure. Dogs were kept by over 60% of households in both villages. Body condition was optimal for 68% of the dogs. In the low season, dogs in the STN-village had better body condition than dogs in IT-village (P = 0.007). Dog characteristics that influenced behavioural responses were: sex, age, and whether the dog played with humans. The most common response to the caregiver’s call was tail wagging, shown by 83% of male dogs and 50% of female dogs (P = 0.021). About 70% of the pups approached the unfamiliar human completely, whereas only 24% of the juveniles (P = 0.040) and 26% of the adults did so (P = 0.026). Human–dog play was reported to occur mainly with children (77%). The percentage of dogs that played with humans was higher in dogs responding with tail wagging (82%) than in dogs showing the rest of the response categories (withdrawal, baring teeth, and other) (50%) (P = 0.012). Human–dog play was reported for 85% of the male dogs compared to 55% of the female dogs (P = 0.036). This study showed that village dogs were socialised to familiar humans but were not attracted to unfamiliar humans. Village dogs maintained their body condition in the low season. Child–dog play may have a role in shaping village dog social behaviour towards humans. © 2014 Elsevier B.V. All rights reserved.

∗ Corresponding author. Tel.: +31 0317 483959. E-mail address: [email protected] (E. Ruiz-Izaguirre). http://dx.doi.org/10.1016/j.applanim.2014.02.002 0168-1591/© 2014 Elsevier B.V. All rights reserved.

58

E. Ruiz-Izaguirre et al. / Applied Animal Behaviour Science 154 (2014) 57–65

1. Introduction Free-roaming dogs (Canis familiaris) in developed and developing countries are a worldwide concern because of various dog-related problems, such as dog welfare, aggressive dogs and dog bites, transmission of zoonoses, and wildlife predation. Low body condition, car accidents, and diseases, such as mange, affect the welfare of free-roaming dogs (Fielding and Mather, 2008; Totton et al., 2011). Dog welfare is further compromised by the lack of veterinary care, especially in rural villages (Constable et al., 2012; Farnworth et al., 2012). The presence of aggressive dogs or packs may prevent people from walking on the street (Poss and Bader, 2007), and, although most dog bites are reported to occur with familiar humans, they may also occur with unfamiliar humans on the street (Cornelissen and Hopster, 2010). Dogs are capable of transmitting more than 60 zoonoses (Meslin et al., 2000), and rabies, in particular, is of great concern among countries with free-roaming dog populations (Dalla Villa et al., 2010; Morters et al., 2013). Furthermore, in nature-protected areas, domestic dogs from adjacent villages are known to prey on endangered wildlife (Acosta-Jamett et al., 2010; Butler et al., 2004). How people perceive, and thus interact with, dogs in different cultures is central to the understanding and resolution of the aforementioned problems (Hsu et al., 2003; Slater et al., 2008). Furthermore, how local people interact with dogs may play a role in shaping dog behaviour, namely, dog socialisation with humans. In The Bahamas, for example, it was hypothesized that the limited interaction between dogs and humans made roaming dogs shy of humans (Fielding et al., 2005). Available descriptions of dogs in African villages suggest that experiences with humans are mostly of a neutral nature, resulting in a typical scavenging behaviour in proximity to humans, but generally keeping a safe (flight) distance (Coppinger and Coppinger, 2001). Flight distances of village dogs in Ethiopia when they were approached by an unfamiliar human ranged from 0 to more than 5 m, with one third of dogs keeping a distance of more than 5 m (Ortolani et al., 2009). In rural areas in Mexico, most households keep dogs (Orihuela and Solano, 1995; Ortega-Pacheco et al., 2007a), and about 80% of these dogs roam freely (Ortega-Pacheco et al., 2007a; Ruiz-Izaguirre and Eilers, 2012). These dogs are known as village dogs (Coppinger and Coppinger, 2001). Food appears to be a central component of human–(village) dog interactions (Coppinger and Coppinger, 2001), and therefore the nature of these interactions may also play a role in village dog welfare. Obtaining food may involve different forms of interaction with humans, from scavenging at a distance to actually begging for household leftovers. Behavioural responses may be influenced by animals’ previous experiences with familiar and unfamiliar humans (Waiblinger et al., 2006). In principle, village dogs may experience negative, neutral, or positive interactions with humans. For example, dogs that regularly receive food leftovers (from caregivers or tourists) may be more willing to approach humans in general, and more food may be reflected in better body condition, and thus better welfare.

Village dogs in Mexican coastal villages comprise a particular dog population with access to wildlife (e.g. seaturtle nests) and/or the possibility of exploiting a niche created by tourism (tourist refuse or hand-outs). In villages of Oaxaca, Mexico, 40% of tourists fed village dogs, and village dogs were found scavenging sea-turtle nests and hatchlings (Ruiz-Izaguirre and Eilers, 2012). We, therefore, expected to find differences in dog behaviour and body condition between two contrasting villages, one with visiting tourists and one in the vicinity of a sea-turtle nesting site. The objectives of this study were: (1) to assess village dogs’ behavioural responses towards familiar and unfamiliar humans, (2) to compare body condition of dogs living in a village with a seasonal trade in international tourism (ITvillage) with dogs living in a village located in the vicinity of a sea-turtle nesting site (STN-village), and (3) to identify whether some dog characteristics influence dog behaviour and body condition. Dog characteristics explored were: sex, age, whether human–dog play occurred, type of care provided by humans (i.e. food, restraint), and the dog’s association with other households.

2. Materials and methods 2.1. Study sites Two rural coastal villages in the Mexican state of Michoacán were selected as case study sites. The study was conducted from September 2008 to the end of June 2009. The villages are located along Mexico’s Pacific Coast, about 30 km from each other. La Ticla village (18◦ 27 N, 103◦ 33 W) has 415 inhabitants, and Colola (18◦ 18 N, 103◦ 26 W) has 477, according to the latest census (INEGI, 2005). La Ticla, further called IT-village, has the capacity to accommodate at least 300 visitors, and has minimal sea-turtle nesting along the beach (approximately 60 nests/year). Colola, further called STN-village, is only occasionally visited by tourists and is adjacent to Colola Sanctuary, an important nesting site for endangered sea turtles (approximately 4000 nests/year). The tourist and sea-turtle nest high seasons occur in the same period: from November to February. The remaining months are considered as part of the low season.

2.2. Dog census Since no dog population data were available in the study site, a dog census was conducted in June 2009, but only after all dogs had been tested for the first time, to avoid familiarity of the dogs with the researcher before conducting the unfamiliar human test. Two persons (including the researcher who acted as the unfamiliar human) visited all village households asking residents which species were kept in the households; and in households with dogs, dog population data were gathered (number of dogs, age and sex of each dog, and whether any dogs had died in the last month).

E. Ruiz-Izaguirre et al. / Applied Animal Behaviour Science 154 (2014) 57–65

2.3. Experimental setup Caregivers with dogs were continuously recruited over a 6-month period by either approaching households with dogs, approaching people with dogs walking on the street or beach, or using the snowball method (i.e. asking a participant to nominate someone else who had dogs that were allowed to roam freely). The inclusion criteria for dogs were: the dog was allowed to roam freely, and a caregiver was found who agreed to participate in the study (i.e. be interviewed and participate in the behavioural testing of the dog). The criteria were to include only one dog per household. In three cases, however, pups born during the course of the study which reached the desired age range for testing were included. Therefore, a maximum of two dogs per household (a pup and a juvenile or adult) were included. This was done because only a few pups of the desired age range were available in the villages (e.g. many pups were born, but did not survive more than a few weeks). In case a household had multiple dogs, we selected the dog known by the caregiver to roam most frequently. We do not consider that this type of selection would have created bias in regard to choosing friendly dogs. One could think that perhaps the dogs roaming most frequently were less friendly, but during informal talks we found that this was not an issue. Both favourite and not favourite dogs were allowed to roam. Caregivers were defined as people from a household where the dog was regularly fed and found every day. All interviews and tests throughout the study period were carried out by a female researcher, who at the time of the first interview was unfamiliar to the dogs. After the first visit, the researcher continued to visit caregivers once a month for a maximum of nine visits. Household visits were opportunistic in nature, and, in some cases, several attempts were required to find the dog and a caregiver at home. Households were approached a maximum of three times in a given month before reporting data as missing for that particular month. The absence of the initial caregiver during some of the monthly visits made it necessary to invite other family members to participate (i.e. interview and testing). A pup was defined as >7 weeks and ≤16 weeks of age; a juvenile as >16 weeks and ≤52 weeks of age; and an adult as >52 weeks of age. This adult threshold was chosen because most village dogs reach maturity in approximately 12 months (Lord et al., 2012), whereas the puppy threshold was chosen because pups older than 16 weeks are past their sensitive period of socialisation (Coppinger and Coppinger, 2001). 2.4. Interview All visits started with a 20 min interview, whereas during subsequent visits the interview lasted approximately 10 min. During the first interview, each dog was photographed and its name, sex, birth date (approximate), type, and colour of coat, as well as additional characteristics such as the presence of scars were recorded. The caregiver was interviewed in the backyard of the household, which is often characterised as an open area surrounded by bushes without any fences.

59

The interview covered questions regarding various aspects of human–dog interactions during the last month prior to the interview (Table 1). In the first interview, information was collected about human–dog play as reported by the caregiver (i.e. sex and age of all people who played with the dog including other households), feeding and tethering of dogs, and whether dogs visited other households. In subsequent visits, it was ascertained whether any of these aspects had changed since the previous interview, and if yes, how. The gender and age of the caregiver was recorded during each interview. Answers were entered directly into a Palm Tungsten E2 PDA (Personal Digital Assistant) (Palm, Inc., Sunnyvale, CA, USA) with CyberTracker application (http://www.cybertracker.org/). At the end of the interview, if the dog was not already in the backyard, the caregiver was asked to bring the dog there. Then, when the dog was in the researcher’s sight, the testing procedure started. 2.5. Testing procedures The testing procedures attempted to represent a short visit from a stranger. All tests were carried out when dogs were free, and, if a dog was occasionally found tethered, testing was cancelled for this particular month. Given that dogs were free to leave at any time, it was necessary to test the dog as quickly and efficiently as possible to maximise the chances of the dog displaying a normal response to the caregiver, while minimising any possible interference from the unfamiliar human (researcher). We therefore recorded dogs’ responses towards humans in four exclusive categories of behaviour (see Table 1), following Ortolani et al. (2009). 2.5.1. Familiar human test The dog was on site in the backyard, either standing or lying down. The unfamiliar human adopted a crouched position at an equidistant point from the caregiver and the dog (approximately 3 m). The position of the dog could not be controlled for. At the time of testing, some dogs were already beside their caregiver, and some not. The caregiver was asked to call the dog (by name), using whatever was considered to be the usual friendly call. Human postures and vocalisations (tone and volume of voice, as well as additional accompanying sounds of a normal friendly call, such as whistles) varied by caregiver and were not recorded. If the dog did not acknowledge the call (i.e. by not rising its ears and/or looking in the direction of the caregiver), the caregiver was instructed to call the dog again for a maximum of three times. The dog’s response was rated within five seconds after the (acknowledged) caregiver’s call, in one of the four exclusive categories. 2.5.2. Unfamiliar human test Immediately after rating the reaction towards the caregiver’s call, the unfamiliar human (i.e. researcher) stood up slowly, her arms at her side, slightly leaning her head towards the dog, looking in the direction of the dog (but without staring into the dog’s eyes), and called the dog by name (Table 1). For dogs that did not respond within 5 s, the researcher called the dog a maximum of three times before

60

E. Ruiz-Izaguirre et al. / Applied Animal Behaviour Science 154 (2014) 57–65

Table 1 Summary of interview questions, testing and scoring categories with a dog named Aguinaldo as an example. Response options/definition of exclusive categories

1st visit

Monthly visits

Last visit

Yes/no

X

X

X

X X X

X

X

Does Aguinaldo eat in other households? Has Aguinaldo been tethered for periods of the day or night? Does Aguinaldo visit other households? Have you had any problems with Aguinaldo in the last month? If so, what type of problems?

Sex and age of all who play Description Daily/every other day/less than every other day Yes/no Yes/no Yes/no Yes/no Describe

X X X X X

X X X X X

X X X X X

Extra questions in the case of births Date of birth Are pups touched? If yes, since when? Do you give food to pups? If yes, since when? Have pups been given away? If yes, when?

Date Yes/no and approximate date Yes/no and approximate date Yes/no and approximate date

Interview questions Does someone from your household or outside the household play with Aguinaldo? If yes, who plays with him? If yes, how do you (or someone else) play with him? How often do you feed him?

Behavioural tests (Familiar human) Aguinaldo is called by the caregiver Tail wagging Withdrawal

Baring teeth Other (Unfamiliar human) Aguinaldo is called by the researcher Complete approach Partial approach No approach Fledb a b

X X X X X

X

X

Repetitive wagging movement of the tail Moving away from the caregiver with the head or the body irrespective of distance moved Showing front teeth to the caregiver All other responses not fitting the other three categoriesa X

X

<20 cm >20 cm to 3 m Dog runs away from the researcher

E.g. standing or lying down still, approaching the caregiver without displaying any of the category responses. Incorporated only in last visit.

rating the response (Table 1). After the first visit, it became apparent that there were also some dogs that fled when called. Given that this response had not been included in the response categories, dogs that fled were rated as ‘no approach’ in the unfamiliar human test. Only in the first visit was the researcher considered an unfamiliar human. Dogs in the longitudinal study were tested during their last visit using the same testing procedures as described for the unfamiliar human test. In this last visit, ‘fled’ was included as an additional category in order to know how frequently this occurred in dogs that were possibly already habituated to the researcher.

2.5.3. Body condition score After the two behavioural tests (i.e. familiar and unfamiliar human), body condition was rated using the 9-point Laflamme Scale (Laflamme, 1997) described in the Purina Body Condition System that rates dogs ranging from emaciated to obese. A body condition score (BCS) of 1–4 is below optimum, a BCS of 5 is optimal, and a BCS of 6–9 is above the optimum (Laflamme, 1997). All scores were determined visually. In some cases, the dog left the household after the behavioural tests, and the researcher had to look for it in the streets in order to rate body condition.

2.6. Statistical analysis 2.6.1. First visit The software package R (R Development Core Team, 2010) was used to statistically analyse the data. There were few differences between the two villages and therefore, for most analyses, the two data sets were combined. Descriptive statistics were used for the outcomes of the behavioural tests, and for human–dog play data (i.e. type of games played, and age and sex of 98 human play partners). To analyse the relation between dogs’ behavioural responses and interview items, the outcomes of the behavioural tests were expressed in binary notation. For the familiar human test, a dog responding with tailwagging was coded as 1, and the rest of the responses as 0, since tail wagging was considered an indicator of a positive social response to the caregiver (Scott and Fuller, 1965). All interview items were expressed in a similar notation (e.g. tethered = 1; not tethered = 0). Finally, the binary outcomes of the behavioural tests and the binary answers to the interview items were analysed with the chi-square test, or the Fisher’s exact test, when minimal cell size was less than 5 (Siegel and Castellan, 1988). The same was done to analyse the effect of dog sex and village. To account for habituation of the dogs to the researcher, the response to the

E. Ruiz-Izaguirre et al. / Applied Animal Behaviour Science 154 (2014) 57–65 Table 2 Dog population demographics in two coastal villages in Mexico.

Table 3 Age category, sex characteristics, and behavioural responses of village dogs assessed during the first visit.

STN-village n = 83

IT-village n = 111

Dog characteristics (% dogs) Males Females Younger than 1 year Mortality in last month

68 32 32 9

65 35 44 2

Population parameters Mean no. of dogs per householda Male:female ratio Human:dog ratiob

Study dogs n (%) Males Females Age in weeks mean (range)

1.6 2:1 5.7:1

1.9 2:1 3.7:1

Responses to familiar human test n (%) 9 (90) Tail wagging 1 (10) Resta

a Refers only to households that reported having dogs, and the difference between the two villages is marginally significant (two sample t-test, P = 0.078) b The human:dog ratio was calculated by dividing the total human population per village by the total number of dogs per village in our census

researcher’s call (in the first visit, the unfamiliar human) was compared between the first and last visit, using the McNemar test for correlated proportions.

61

Pups n = 10

Juveniles n = 18

Adults n = 31

8 (80) 2 (20) 12 (7.6–15.6)

13 (72) 5 (28) 28 (16–52)

20 (65) 11 (35) 136 (54.2–515)

13 (72) 5 (28)

Responses to unfamiliar human testb n (%) 7 (70)c 4 (24)c Complete approach Partial or no approach 13 (76) 3 (30)

21 (68) 10 (32) 6 (26)c 17 (74)

a The rest of the responses include the categories: other, withdrawal, and baring teeth. b 10 pups, 17 juveniles and 23 adults. c Significant difference between age groups (pups and juveniles, Fisher’s exact test, P = 0.040) (pups and adults, Fisher’s exact test, P = 0.026).

3. Results 3.1. Dog census

2.6.2. Monthly visits Because of the high variability in finding dogs during monthly visits, data were unbalanced. Dogs were occasionally found tethered (testing was cancelled), were sometimes not found in a given month, had died, or had been given away outside the village. In the case of female dogs giving birth, it was sometimes not possible to test during the first weeks postpartum if the den was outside the household. Nevertheless, births served as an opportunity to gain information on early experiences of village dogs (additional data to the interview, Table 1). Despite the challenge of gathering longitudinal data, 35 dogs met the criteria for longitudinal analysis; these were: they had at least three visits and two behavioural tests (one during the high season and one during the low season). Depending on the date, each visit was classified as being in the high or low season. First, we calculated the proportional occurrence of tail wagging (familiar human test) and interview items over all visits for each study dog per season (i.e. high and low), and total (both seasons). For BCS, we computed the overall mean per season and total, corrected by the number of visits. The proportional occurrence of tail wagging and interview items and the mean BCS were not normally distributed, and therefore nonparametric tests were used. A Kruskal–Wallis test was used to find out whether the proportional occurrence of total tail wagging (i.e. both seasons) differed in dogs that were tested with one, two, three, or four caregivers. A dependent two-group Wilcoxon Signed Rank Test was used to compare the proportional occurrence and mean BCS between the high and low season. Within season, the effect of dog sex and village on the proportional occurrence of tail wagging and of interview items was analysed with independent group Mann–Whitney U tests.

Dogs were kept in 64% out of 89 households in the ITvillage and in 62% out of 85 households in the STN-village, with no differences between villages. Other animal species kept were: poultry (57%), pigs (23%), cats (22%), and parakeets (11%); other animals like horses or goats were owned by less than 10% of households. Table 2 shows the dog census demographics of the IT-village and the STN-village. In both villages, the male to female dog ratio was 2:1. The human to dog ratio was higher in the STN-village because there was a tendency to have fewer dogs per household. The dogs had an average age of 3.8 (±3.1 SD) years, and 39% of the dogs were younger than one year. 3.2. First visit: characterisation of human–dog interactions and dog behavioural responses A total of 59 dogs (24 in the IT-village and 35 in STNvillage) associated with 56 households participated in the study. Most caregivers were women (80%, n = 45). The mean age of female caregivers (35.4 ± 16.5 years) tended to differ from male caregivers (23.7 ± 18.9 years) (two-sample t-test = −1.885, df 13.0, P = 0.04), because most of the female caregivers (67%) were mothers or grandmothers, whereas 64% out of eleven male caregivers were children living with parents. No dogs were sterilised or belonged to a specific breed. More male (69%) than female (31%) dogs were included in the study. The age of dogs ranged from 7.6 to 515 weeks (Table 3). There were no differences between villages regarding interview items or body condition of the dogs. Ninety-six percent of the studied dogs were fed daily by their caregivers, 68% visited other households, and 28% were fed also in other households. Tethering for periods of day or night occurred in 28% of the dogs, of which 35% were males and 6% females (Fisher’s exact test P < 0.05). BCS ranged from 4

E. Ruiz-Izaguirre et al. / Applied Animal Behaviour Science 154 (2014) 57–65

to 6 in the STN-village, and from 1 to 7 in the IT-village, with 68% of the dogs having an optimum or higher BCS (≥5). Seventy-six percent of dogs played with someone from the household or outside the household, and there were no differences per dog age group. For each dog reported as engaging in human–dog play, one to five humans were reported as playing (58% men, 42% women, and 77% younger than 10 years old, M = 4.3 ± 2.7 S.D. years). Human–dog play was reported in 85% of the male dogs compared to 55% of the female dogs (␹2 = 4.38; df = 1; P = 0.036). One or more of the following games were mentioned by caregivers: chase (57%), rough and tumble (23%), fetch (14%), and other (25%) (asking the dog to jump or give a paw, or riding on the dog like a horse). Chase was described by caregivers as dog and human chasing each other. Rough and tumble were described as dog and human wrestling. Fetch was described as playing with a stick or stone. Although not systematically recorded, the researcher confirmed the descriptions given by villagers during household visits, or through random observations on the beach and village streets. Eight caregivers reported having the following problems with their dogs’ behaviour: aggressiveness (n = 2), chasing passers-by (n = 2), killing sea-turtles (n = 1), temporarily leaving household for mating (n = 1), pulling down clothes (to sleep on them) (n = 1), and spilling the garbage (n = 1).

3.3. Behavioural tests during the first visit Table 3 shows the responses to the familiar and unfamiliar human tests during the first visit by age group. There was no age effect regarding the response to the familiar human test; pups, juveniles, and adults responded mainly with tail wagging. The rest of the responses included the categories: ‘other’ (16%), ‘withdrawal’ (8%), and ‘baring teeth’ (3%). Tail wagging was shown by 83% of male dogs compared to 50% of female dogs (2 = 5.3, df = 1; P = 0.021). The percentage of dogs that played with humans was larger in the category responding with tail wagging (82%), than in the rest of the response categories (50%) (2 = 6.2, df = 1; P = 0.012). Similarly, the percentage of dogs that played with humans tended to be larger in the category of dogs that completely approached the unfamiliar human (94%) than in the category that did not (67%) (Fisher’s exact test P = 0.073). Age had an effect in the test with the unfamiliar human. Most pups approached the unfamiliar human completely (70%), compared with juveniles (24%) (P = 0.040) and adults (26%) (P = 0.026) (Table 3). Instead, most juveniles responded with a partial approach (range of 1–3 m), or no approach, whereas most adults did not approach at all (Fig. 1). The percentage of dogs that completely approached the unfamiliar human was higher in the category of dogs that showed tail wagging to the caregiver (94%) than in the rest of the response categories (6%) (Fisher’s exact test P = 0.004). There was a tendency for dogs’ responses to the unfamiliar human to change after several visits: 11 dogs that were not willing to completely approach in the first visit (n = 25) eventually did so in the last visit (McNemar test, 2 = 3.5, df = 1; P = 0.061). Nevertheless, during the last

Dog groups

pups

juveniles

adults

70 60 50 Percentage

62

40 30 20 10 0 Complete_approach

Partial_approach Response

No_approach

Fig. 1. Responses to the unfamiliar human test (close, partial, no approach) by age group (in percentage): adults (n = 23), juveniles (n = 17) and pups (n = 10).

visit, 53% of dogs still did not completely approach the unfamiliar human, and two of these actually fled. 3.4. Longitudinal study and impact of season Thirty-five dogs were included in the longitudinal study according to the inclusion criteria. Of the 59 dogs first visited, 15 dogs were tested only once (due to death or inclusion of pups close to the end of the study period), and the remainder did not have sufficient visits (n = 9). In total, 11 dogs died during the course of the study. Nine adult females (>1 year) bred during the study period and had one to eight pups per litter. About 62.5% of the caregivers (one litter data missing) reported caressing or holding the pups at four weeks of age, and 87.5% of them fed the pups extra food from the time they were 2 weeks old. In 42% of all cases dogs were tested always with the same caregiver, in 27% with two caregivers, in 21% with three caregivers, and in 10% with four caregivers. There were no differences in the proportional occurrence of tail wagging between dogs tested with one caregiver and dogs tested with more than one caregiver (Kruskal–Wallis test, 2 = 0.101, P = 0.992). The proportional occurrence of tail wagging indicates the number of times a dog responded with tail wagging in all testing sessions (familiar human test). There were no differences between seasons on the proportional occurrence of any interview item, or behavioural test and BCS, but there were differences within season with regard to village and dog sex. During the low season, dogs in the STN-village had a higher BCS (median 5) than dogs in the IT-village (median 4.33) (U = 218, Z = 2.633, P = 0.007). With regard to dog sex, human–dog play was more reported in male (median 1) than in female (median 0.33) dogs during both the high (U = 63, Z = −2.815, P = 0.004) and the low (U = 80.5, Z = −2.126, P = 0.033) season. In the high season, caregivers reported more problem behaviours for male dogs (median 0.33) than for female dogs (median 0) (chasing passers-by or domestic animals,

E. Ruiz-Izaguirre et al. / Applied Animal Behaviour Science 154 (2014) 57–65

trying to bite, etc.) (U = 69, Z = −2.588, P = 0.009), including three male dogs actually reported to have bitten. 4. Discussion In this study, Mexican village dogs responded most frequently with tail wagging to their caregiver’s call. Tail wagging in dogs is exhibited in many different contexts, both social and non-social, and is associated with different emotional states including: friendliness (Rappolt et al., 1979; Scott and Fuller, 1965; Tami and Gallagher, 2009; Vas et al., 2005), submission (Kiley-Worthington, 1976), anxiety (Bradshaw and Nott, 1995), and even chronic stress (Beerda et al., 2000). Recently, Siniscalchi et al. (2013) suggested that the asymmetry in tail wagging contains important information regarding the emotional state of a dog. It could be signalling either increased arousal or a relaxing state. Although we did not record tail wagging asymmetry when village dogs were called by their caregivers, dogs generally exhibited signs of friendliness, such as approaching and licking their caregivers. These signs of friendliness were not systematically recorded due to difficult fieldwork circumstances, but support our interpretation that tail wagging was mostly friendly in nature. Furthermore, we consider that the familiar context in which testing occurred, and the fact that the dog was free to leave at any time, were compatible with a companionship/confidence emotional state (as in Siniscalchi et al., 2013). In regard to the unfamiliar human test, most adult dogs did not completely approach the unfamiliar human; and, although there was a tendency to approach after repeated visits, still more than half of the dogs did not approach at all during the last visit. For the majority of the study dogs, a lack of socialisation with unfamiliar humans and experiences with humans on the street, which may range from positive to negative, may explain this observation. Positive experiences may in fact be few given that not all villagers like dogs and some may consider them a nuisance (RuizIzaguirre and Eilers, 2012) and chase them away. The lack of socialisation with unfamiliar humans may have negative consequences for humans inside (e.g. visitors) and outside the household. Most dog bites outside the household involve unfamiliar humans, even if there is no interaction with the dog (Cornelissen and Hopster, 2010). Dog bites indeed are a risk for tourists visiting rural coastal areas (Ruiz-Izaguirre and Eilers, 2012). In the current situation, the lack of socialisation with unfamiliar humans may render difficult any handling procedures by unfamiliar humans (vaccination, sterilisation) in the local dog population, as in animal birth control (ABC) programmes (Totton et al., 2011). One of the challenges of ABC programmes is indeed to be able to reach those less socialised dogs (Totton et al., 2011). In our study site, however, caregivers could play a major role in handling dogs in dog management programmes. Contrary to our expectation that dogs in a poor body condition would be less socialised, we found that body condition was not related to behavioural responses or to dog characteristics. Probably due to the fact that body condition was generally good in the studied dogs. This corresponds

63

with what was reported for village dogs in Ethiopia, where 96% of dogs were in a fair body condition (Ortolani et al., 2009), but differs from street dogs in cities. In Merida, Mexico, only 36% of dogs were in an optimal condition (not under or overweight) (Ortega-Pacheco et al., 2007b). Similarly, in Jodhpur, India, only 30% of dogs were in an optimal condition (Totton et al., 2011). Differences between rural and urban areas may be due to the increased pressure on food resources in urban areas and/or to better care of village dogs in rural areas. Furthermore, although it was not possible to determine this in our study, the availability of wildlife resources (e.g. sea-turtle eggs), may also play a role in maintaining body condition. Consistent with other village dog studies (Boitani et al., 2007; Farnworth et al., 2012; Ortega-Pacheco et al., 2007a; Ortolani et al., 2009; Ruiz-Izaguirre and Eilers, 2012), our dog census revealed a skewed male to female ratio of 2:1. Reasons for the skewed sex ratio include high pup mortality (Lord et al., 2012) and culling of female pups in Mexican villages (Ruiz-Izaguirre and Eilers, 2012). This is especially noticeable in the pup age group (eight males and two females). Similar to other rural areas, female pups in Mexico are typically culled to avoid nuisance with female dogs in oestrus, and because of the preference for male dogs (Farnworth et al., 2012; Ortolani et al., 2009; Ruiz-Izaguirre and Eilers, 2012). To interpret the behavioural test results, certain factors that could have influenced the responses should be considered. There are several aspects of motivation which may impact upon the likelihood a dog will approach such as hunger or tiredness. Furthermore, dogs could be distracted by other dogs, backyard animals, or humans. Given the opportunistic nature of data collection we were unable to assess these aspects during testing. Furthermore, to be able to better interpret human–dog interaction, both types of measurements (i.e. quality of the response and approach distance) for both the familiar and unfamiliar human should have been recorded. For the familiar human, this was not possible because we could not control for the position of the dog. For the unfamiliar human, this was unfortunately not part of the protocol from the beginning of the study. Initially, we were only interested in the approach distance to the unfamiliar human for practical reasons, e.g. whether dogs would approach unfamiliar humans who wanted to vaccinate or sterilise them (e.g. as in trap–neuter–release programmes). Dog characteristics that influenced behavioural responses were age, sex, and whether the dog played with humans. Pups were more willing to approach the unfamiliar human than adults; this could be explained by the fact that pups are in their sensitive period for socialisation (Scott and Fuller, 1965), and their motivation to flee unfamiliar humans is not strong (Freedman et al., 1961). The fact that more male than female dogs responded with tail wagging could be due to both biological and human factors. Female dogs in our study sites may have less experience with humans (either positive or negative) because caregivers interact less with them. Furthermore, female dogs may spend less time with humans because they instinctively put their efforts into reproduction. Ghosh et al. (1984), for example, reported that in Indian

64

E. Ruiz-Izaguirre et al. / Applied Animal Behaviour Science 154 (2014) 57–65

village dogs females in oestrus spent more than 70% of their time in association with male conspecifics, and often left their own locality to seek males. In most dogs, human–dog play was reported to occur regularly (in the monthly interviews), and mainly with children; this suggests that children may have a positive role in village dog socialisation. This is further supported by the results of the first visit, in which dogs that were reported to play were more likely to respond with tail wagging to the caregiver, and to approach completely the unfamiliar human. It is not clear why male dogs were more often reported to play with humans than females (both in the first visit and in the longitudinal study). Our results could be explained by biological causes and from different handling experiences for male and female dogs. Male village-dog pups, for example, are known to initiate more play interactions with other pups than females (Pal, 2010). There, however, is no data available for adult village dogs regarding play interactions with humans. We suggest that the overt preference of villagers for male dogs (reflected in the skewed male to female ratio) has an important role here. The preference for male village dogs may be dictated by pragmatism, e.g. nuisance with breeding of females and having to deal with unwanted pups (RuizIzaguirre and Eilers, 2012). Children may also prefer playing with male dogs for cultural reasons, such as that male dogs are possibly perceived as having more qualities, like being braver or stronger, but this is an area for further research. The finding that most dogs play with children younger than 10 years old was unexpected and merits further study. Although playing could be positive for dog socialisation, there are also risks involved, given that children are a high risk group for dog bites (Farnworth et al., 2012; Kahn et al., 2003). More research on the nature and context of play interactions would be necessary to determine whether playing has a (positive or negative) influence on the risk of being injured. Following this line of research has relevance not only for village dog socialisation, but also in the prevention of dog bites, particularly in situations as in our study site, children and tourists may be at risk.

5. Conclusion This study has shown that village dogs in two coastal villages in Michoacán, Mexico, one with a seasonal trade in international tourism and one in the vicinity of a sea-turtle nesting site, are socialised with familiar humans. Female dogs were less reported to engage in human–dog play and responded less with tail wagging to the caregiver’s call. Village dogs older than 16 weeks (both male and female) were reluctant to approach the unfamiliar human. Most dogs were in a close to optimal body condition irrespective of their behavioural response to their caregiver. During the low season, dogs in the STN-village, however, had a better body condition than dogs in the IT-village. According to our results, it was mainly children who played with village dogs, and we believe that they may have an important role in shaping village dog behaviour.

Acknowledgements We are grateful to all the villagers that participated in the study. We thank Wiebe Koops for his help with the longitudinal analysis. We thankfully acknowledge two anonymous reviewers for their thoughtful comments. The first author was funded by The Council for Science and Technology of Mexico and the Alfa en Omega Stichting (doctoral scholarship). The first author also benefitted from a work retreat at Shin Pond, Maine, awarded by The Humane Society of the United States, to write this article. References Acosta-Jamett, G., Cleaveland, S., Cunningham, A.A., Bronsvoort, B.M.D., 2010. Demography of domestic dogs in rural and urban areas of the Coquimbo region of Chile and implications for disease transmission. Prev. Vet. Med. 94, 272–281. Beerda, B., Schilder, M., Van Hooff, J., De Vries, H., Mol, J., 2000. Behavioural and hormonal indicators of enduring environmental stress in dogs. Anim. Welfare 9, 49–62. Boitani, L., Ciucci, P., Ortolani, A., 2007. Behaviour and social ecology of free-ranging dogs. In: Jensen, P. (Ed.), The Behavioural Biology of Dogs. CABI, Wallingford, pp. 147–164. Bradshaw, J.W., Nott, H.M., 1995. Social and communication behaviour of companion dogs. In: Serpell, J. (Ed.), The Domestic Dog: Its Evolution, Behaviour and Interactions with People. Cambridge University Press, Cambridge, pp. 115–130. Butler, J.R.A., Du Toit, J.T., Bingham, J., 2004. Free-ranging domestic dogs (Canis familiaris) as predators and prey in rural Zimbabwe: threats of competition and disease to large wild carnivores. Biol. Conserv. 115, 369–378. Constable, S.E., Dixon, R.M., Dixon, R.J., Toribio, J.A., 2012. Approaches to dog health education programs in Australian rural and remote indigenous communities: four case studies. Health Promot. Int., http://dx.doi.org/10.1093/heapro/das013. Coppinger, R., Coppinger, L., 2001. Village dogs. In: Coppinger, R., Coppinger, L. (Eds.), Dogs: A New Understanding of Canine Origin, Behavior and Evolution. Scribner, New York, pp. 69–82. Cornelissen, J.M., Hopster, H., 2010. Dog bites in The Netherlands: a study of victims, injuries, circumstances and aggressors to support evaluation of breed specific legislation. Vet. J. 186, 292–298. Dalla Villa, P., Kahn, S., Stuardo, L., Iannetti, L., Di Nardo, A., Serpell, J.A., 2010. Free-roaming dog control among countries. OIE-member Prev. Vet. Med. 97, 58–63, http://dx.doi.org/10.1016/j.prevetmed.2010.07.001. Fielding, W.J., Mather, J., Isaacs, M., 2005. Responsibilities of owners towards pets. In: Beck, A. (Ed.), Dog Ownership in New Providence. Purdue University Press, The Bahamas, pp. 45–56. Fielding, W.J., Mather, J., 2008. Some observations on dog welfare in The Bahamas. Int. J. Bahamian Stud. 11, 41–56. Farnworth, M., Blaszak, K., Hiby, E., Waran, N., 2012. Incidence of dog bites and public attitudes towards dog care and management in Samoa. Anim. Welfare 21, 477–485. Freedman, D.G., King, J.A., Elliot, O., 1961. Critical period in the social development of dogs. Science 133, 1016–1017. Ghosh, B., Choudhuri, D., Pal, B., 1984. Some aspects of the sexual behaviour of stray dogs, Canis familiaris. Appl. Anim. Behav. Sci. 13, 113–127. Hsu, Y., Liu Severinghaus, L., Serpell, J.A., 2003. Dog keeping in Taiwan: its contribution to the problem of free-roaming dogs. J. Appl. Anim. Welfare Sci. 6, 1–23. INEGI, 2005. Conteo de Población y Vivienda. Instituto Nacional de Estadística e Informática (National Institute of Statistics and Informatics), Mexico. Kahn, A., Bauche, P., Lamoureux, J., 2003. Child victims of dog bites treated in emergency departments: a prospective survey. Eur. J. Pediatr. 162 (4), 254–258. Kiley-Worthington, M., 1976. The tail movements of ungulates, canids and felids with particular reference to their causation and function as displays. Behaviour 56, 69–115. Laflamme, D., 1997. Developmental and validation of a body condition score system for dogs. Canine Pract. 22, 10–15. Lord, K., Feinstein, M., Smith, B., Coppinger, R., 2012. Variation in reproductive traits of members of the genus Canis with special attention

E. Ruiz-Izaguirre et al. / Applied Animal Behaviour Science 154 (2014) 57–65 to the domestic dog (Canis familiaris). Behav. Process. 92, 131–142, http://dx.doi.org/10.1016/j.beproc.2012.10.009. Meslin, F.X., Miles, M.A., Vexenat, A., Gemmell, M.A., 2000. Zoonoses control in dogs. In: MacPherson, C., Meslin, F., Wandeler, A. (Eds.), Dogs, Zoonoses and Public Health. CABI Publishing, Wallingford, pp. 333–372. Morters, M.K., Restif, O., Hampson, K., Cleaveland, S., Wood, J.L., Conlan, A.J., 2013. Evidence-based control of canine rabies: a critical review of population density reduction. J. Anim. Ecol. 82 (1), 6–14, http://dx.doi.org/10.1111/j.1365-2656.2012.02033.x. Orihuela, T., Solano, V., 1995. Demographics of the owned dog population in Miacatlan, Mor. Mexico. Anthrozoös 8 (3), 171–175, http://dx.doi.org/10.2752/089279395787156356. Ortega-Pacheco, A., Rodriguez-Buenfil, J., Bolio-Gonzalez, M., Sauri-Arceo, C., Jimenez-Coello, M., Forsberg, C., 2007a. A survey of dog populations in urban and rural areas of Yucatan, Mexico. Anthrozoös 20 (3), 261–274, http://dx.doi.org/10.2752/089279307X224809. Ortega-Pacheco, A., Segura-Correa, J., Jimenez-Coello, M., Linde Forsberg, C., 2007b. Reproductive patterns and reproductive pathologies of stray bitches in the tropics. Theriogenology 67 (2), 382–390. Ortolani, A., Vernooij, H., Coppinger, R., 2009. Ethiopian village dogs: behavioural responses to a stranger’s approach. Appl. Anim. Behav. Sci. 119 (3–4), 210–218. Pal, S.K., 2010. Play behaviour during early ontogeny in free-ranging dogs (Canis familiaris). Appl. Anim. Behav. Sci. 126 (3–4), 140–153. Poss, J.E., Bader, J.O., 2007. Attitudes toward companion animals among Hispanic residents of a Texas border community. J. Appl. Anim. Welfare Sci. 10, 243–253. R Development Core Team, 2010. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria http://www.R-project.org/ Rappolt, G.A., John, J., Thompson, N., 1979. Canine responses to familiar and unfamiliar humans. Aggressive Behav. 5 (2), 155–161.

65

Ruiz-Izaguirre, E., Eilers, C.H.A.M., 2012. Perceptions of village dogs by villagers and tourists in the coastal region rural Oaxaca, Mexico. Anthrozoös 25 (1), 75–91, of http://dx.doi.org/10.2752/175303712X13240472427555. Scott, J., Fuller, J., 1965. Genetics and the Social Behavior of the Dog. University of Chicago Press, Chicago. Siegel, S., Castellan, N., 1988. Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill Book Company, New York. Slater, M.R., Di Nardo, A., Pediconi, O., Villa, P.D., Candeloro, L., Alessandrini, B., Del Papa, S., 2008. Free-roaming dogs and cats in central Italy: public perceptions of the problem. Prev. Vet. Med. 84, http://dx.doi.org/10.1016/j.prevetmed.2007.10.002. Siniscalchi, M., Lusito, R., Vallortigara, G., Quaranta, A., 2013. Seeing left-or right-asymmetric tail wagging produces different emotional responses in dogs. Curr. Biol. 23 (22), http://dx.doi.org/10.1016/j.cub.2013.09.027. Tami, G., Gallagher, A., 2009. Description of the behaviour of domestic dog (Canis familiaris) by experienced and inexperienced people. Appl. Anim. Behav. Sci. 120 (3–4), 159–169. Totton, S.C., Wandeler, A.I., Ribble, C.S., Rosatte, R.C., McEwen, S.A., 2011. Stray dog population health in Jodhpur, India in the wake of an animal birth control (ABC) program. Prev. Vet. Med. 98, 215–220, http://dx.doi.org/10.1016/j.prevetmed.2010.11.011. Vas, J., Topal, J., Gacsi, M., Miklosi, a., Csanyi, V., 2005. A friend or an enemy? Dogs’ reaction to an unfamiliar person showing behavioural cues of threat and friendliness at different times. Appl. Anim. Behav. Sci. 94 (1–2), 99–115, http://dx.doi.org/10.1016/j.applanim.2005.02.001. Waiblinger, S., Boivin, X., Pedersen, V., Tosi, M., Janczak, A., Visser, E., Jones, R., 2006. Assessing the human–animal relationship in farmed species: a critical review. Appl. Anim. Behav. Sci. 101 (3), 185–242, http://dx.doi.org/10.1016/j.applanim.2006.02.001.