Human Perceptions of and Interactions with Wild Canids on Cattle Ranches in Central Brazil

Human perceptions of and interactions with wild canids on cattle ranches in central Brazil

S TACIE M. BICKLEY,FREDERICO G. LEMOS,MICHAEL P. GILMORE F ERNANDA C. AZEVEDO,ELIZABETH W. FREEMAN and N UCHARIN S ONGSASEN

Abstract Local attitudes towards carnivores often reflect the Keywords Attitudes, Brazilian savannah, canid conserva- degree of damage they are perceived to cause. Consequently, tion, Cerrado, human–canid interactions, predation understanding the interactions between people and these Supplementary material for this article is available at species is essential to conservation efforts. This study inves- https://doi.org/./S tigated local perceptions of three Cerrado canid species and current chicken management practices, to identify the po- tential damage they cause and how this relates to peoples’ attitudes towards these species. Results from structured in- Introduction terviews at  ranches in Goiás, Brazil, highlighted that general knowledge about Cerrado canids differed significantly mongst terrestrial mammals, carnivores are the most by species, with interviewees unable to correctly answer Athreatened group and the most challenging to conserve questions about the hoary fox Lycalopex vetulus and crab- (Karanth & Chellam, ). People tend to see carnivores, eating fox Cerdocyon thous in comparison to the maned particularly large species such as the jaguar Panthera onca wolf Chrysocyon brachyurus. Chicken coops were identified and lion Panthera leo, as a direct threat. These species com- as the most effective method for preventing predation, yet pete for land and game and sometimes prey upon livestock only % of respondents employed this method. Using a (Sillero-Zubiri & Laurenson, ). Intolerance for carni- perceived predation measure, interviewees reported chicken vores has led to the persecution, decline and local extinction predation by all three Cerrado canids even though most of of some species (Treves & Karanth, ). Although some these events were stated to occur during the day, outside the carnivore populations are increasing as a result of conserva- species’ active periods. Reported predation events were a tion initiatives (e.g. LaRue et al., ), encounters between strong predictor of attitude. Participants who experienced people and carnivores will probably increase as the human predation events reported they did not like having a population continues to grow. Cerrado canid on their property. However, % of the re- Factors such as age, education, and knowledge about spondents agreed that Cerrado canids should nevertheless carnivores can affect people’s attitude and willingness to be protected. Our findings support the need to incorporate conserve a species. For example, older respondents hold a the human dimension in canid and broader carnivore con- more negative view of jaguars than younger individuals servation issues. (Zimmermann et al., ), and ranchers who have completed fewer years of school show stronger negative attitudes towards the species than those with higher education (Cavalcanti et al., ). The more knowledge a person has of a species, the more likely they are to protect it or

STACIE M. BICKLEY*† (Corresponding author) Environmental Science and Policy, view wildlife positively (Kellert, ), although more George Mason University, Fairfax, Virginia, USA knowledgeable individuals, such as hunters, may have E-mail [email protected] less favourable attitudes towards carnivores (Ericsson & FREDERICO G. LEMOS‡ and FERNANDA C. AZEVEDO Cerrado Mammals  Conservation Program, Special Academic Unit of Biotechnology, Federal Heberlein, ). Therefore, the knowledge and attitudes University of Goiás/Catalão, Goiás, Brazil of people living with wildlife need to be assessed and inte-

MICHAEL P. GILMORE and ELIZABETH W. FREEMAN School of Integrative Studies, grated into conservation strategies. George Mason University, Fairfax, Virginia, USA Attitudes towards carnivores may also reflect the degree NUCHARIN SONGSASEN Smithsonian Conservation Biology Institute, National of negative interaction between a person and a species Zoological Park, Front Royal, Virginia, USA (Kellert, ). In South Africa, for example, negative atti- *Current Address: Center for Conservation and Research of Endangered tudes towards the African wild dog Lycaon pictus are related Wildlife, Cincinnati Zoo and Botanical Garden, 3400 Vine St., Cincinnati, Ohio 45220, USA to the economic costs of livestock and wild game predation †Also at: Smithsonian Conservation Biology Institute, National Zoological Park, (Lindsey et al., ). Domestic livestock are ideal prey for Front Royal, Virginia, USA ‡ carnivores as they are often abundant and easy to catch in Also at: Natural Resources Ecology and Conservation Postgraduate Program,  Institute of Biology, Federal University of Uberlândia, Minas Gerais, Brazil comparison to other prey (Palmeira et al., ). Ranchers Received  June . Revision requested  September . troubled by livestock predation have a range of pro- Accepted  March . First published online  January . tection techniques available to them (Smith et al., ;

This is a work of the U.S. Government and is not subject to copyright protection in the United States. Downloaded from https://www.cambridge.org/core. IP address: 170.106.35.93, on 02 Oct 2021Oryx at, 2020,07:39:56 54(4),, subject 546– 553to the© 2019Cambridge Fauna Core & Flora terms International of use, available doi:10.1017/S0030605318000480 at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605318000480 Perceptions and interactions 547

Sillero-Zubiri et al., ; Sillero-Zubiri & Switzer, ). have a more positive attitude towards them, () the method Some of these strategies are non-invasive (e.g. flags on perceived most effective for preventing predation of domes- fencing), whereas others may result in the carnivore’s tic animals would be the most frequently implemented, and death (e.g. shooting). Managers must appraise the expend- () respondents who experienced predation of chickens iture of money and time, effectiveness, legality and cultural would have a more negative attitude towards predators. appropriateness of each technique. Most research on human–carnivore interactions and attitudes towards these species has focused on larger species. Study area – Human canid interactions are well documented for large This study was carried out in the Limoeiro region, a farming species such as the grey wolf Canis lupus (Treves et al., area within the municipality of Cumari, Goiás state, in cen-   ), Ethiopian wolf Canis simensis (Yihune et al., ), tral Brazil (Fig. ) that comprises contiguous private cattle  coyote Canis latrans (Draheim, ) and maned wolf ranches and small-scale agriculture operations (e.g. rubber   Chrysocyon brachyurus (Emmons, ), but less so for trees, corn and sugarcane; collectively c.  km ). The land- small canids other than the red fox Vulpes vulpes scape is dominated by exotic pasture (Urochloa spp.; %),    (Moberly et al., ). Of the species of canids, live with the remaining area being a mosaic of natural gallery  in South America (Sillero-Zubiri, ). Knowledge of in- and seasonal forests (%) and open Cerrado sensu stricto teractions with these species is limited to a few studies of (%; Lemos, ). The climate is tropical with two well-  the Andean fox Lycalopex culpaeus (Travaini et al., ), defined seasons, cold/dry (May–September) and hot/wet South American grey fox Lycalopex griseus (Silva- (October–April; Alvares et al., ), with mean tempera-  Rodrígues et al., ), hoary and crab-eating foxes tures of  and °C, respectively, and a mean annual pre- Lycalopex vetulus and Cerdocyon thous (Lemos et al., cipitation of , mm (CPTEC/INPE, ). a) and bush dog Speothos venaticus (DeMatteo, ). The maned wolf, crab-eating fox, and hoary fox are sympatric in the Cerrado ecosystem (Juarez & Marinho-Filho, Methods ; Jácomo et al., ; Lemos, ). The Cerrado was once an expansive grassland, but intense agriculture and Surveys ranching has severely altered the landscape (Klink & Moreira, ; Klink & Machado, ). Because of these Fifty ranches in the Limoeiro region that raised domestic – activities, the hoary fox (endemic to the Cerrado) and the fowl were opportunistically surveyed during June August   maned wolf are currently categorized as Vulnerable in . The selected respondent at each ranch was at least Brazil (Lemos et al., ; Paula et al., ), even though years old and was directly responsible for managing the do-  these species are categorized as Least Concern (Dalponte mestic fowl: either the owner (rancher, %) or the hired  & Courtenay, ) and Near Threatened (Paula and ranch hand (cowboy, %). Each respondent was asked to DeMatteo, ), respectively, on the IUCN Red List. The sign a consent form before starting the interview, was as- crab-eating fox remains common throughout its range sured confidentiality, and could discontinue or withdraw and is categorized as Least Concern both globally from the interview at any time. All interviews were con- (Lucherini, ) and in Brazil (Beisiegel et al., ). ducted in Brazilian Portuguese, with the assistance of a na- In such a highly fragmented landscape, native prey spe- tive speaker, and recorded. cies may be uncommon and domestic livestock becomes an  accessible alternative prey for carnivores. Abade et al. ( ) Questionnaire reported that in the Cerrado ecosystem the perceived loss of domestic fowl (chickens, ducks, geese and guinea fowl) was We used a structured questionnaire developed to evaluate the an unwelcome interaction between landowners and Cerrado interviewees’ knowledge, experience, and attitudes towards canids. Our study builds on this work, examining the socio- the maned wolf and crab-eating and hoary foxes, and economic factors that may influence peoples’ attitudes to- domestic fowl management practices. The questionnaire wards Cerrado canid species. We used structured interviews was first tested at  ranches, refined, and retested at the to examine local knowledge of Cerrado canids and attitudes same  ranches. The final questionnaire (Supplementary towards these species. Because Abade et al. () reported Material ) was administered at an additional  ranches that canids in our study area preyed on domestic fowl rather and comprised  questions that combined multiple-choice other domestic species such as calves, we also quantified (nominal data) and open-ended questions (Bernard, ). current chicken management practices and measured the Comments made by interviewees that complemented an- perceived cost and effectiveness of these methods. We hy- swers were noted, but were not included in the data analysis. pothesized that () interviewees who correctly identified The questionnaire had sections on demographics, knowl- and demonstrated knowledge about Cerrado canids would edge of carnivores and Cerrado canids, domestic fowl

Oryx, 2020, 54(4), 546–553 © 2019 Fauna & Flora International doi:10.1017/S0030605318000480 Downloaded from https://www.cambridge.org/core. IP address: 170.106.35.93, on 02 Oct 2021 at 07:39:56, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605318000480 548 S. M. Bickley et al.

species were compared using a repeated measures ANOVA and Tukey’s Honestly Significant Difference post- hoc test. A multiple linear regression model was used to identify which combination of explanatory variables (age, education level, position on the ranch, sale of chickens) con- tributed to a person’s knowledge of Cerrado canids. Relationships with categorical variables (predation prevention method, attitudes towards canid species, action taken in response to domestic fowl predation) were analysed using a  χ test. If the expected outcome was , , a Fisher’s exact test was used instead (Sokal & Rohif, ). Bivariate analyses were conducted between reported perceived predation and number of chickens, and between predation and number of predation prevention methods employed. A bivariate regression analysis was conducted between attitude (agree- ment or disagreement with the statement ‘I like having ’ FIG. 1 Locations of ranches in the Limoeiro region, municipality this animal on my ranch ) and explanatory variables (edu- of Cumari, Goiás state, in central Brazil, where interviews were cation level, sale of chickens, experience with predation, conducted to investigate local perceptions of three Cerrado canid knowledge score and age). species: the maned wolf Chrysocyon brachyurus, the crab-eating All statistical analyses were conducted using R .. (R fox Cerdocyon thous and the hoary fox Lycalopex vetulus. Development Core Team, ). Models were fitted by a backwards stepwise process: all variables were initially in- management practices, and attitudes towards Cerrado canids. cluded in the model (a saturated model) and then the A canid knowledge score was calculated. Participants were Akaike information criterion (AIC) was used to remove first asked to identify the three species from photographs, the parameters that least affected the model fit (Burnham and then, for each species, to select its prey from a list, its social & Anderson, ). structure, and indicate when it is most active. An interviewee was assigned one point for each correct response, for up to a total of four points per species. An overall knowledge score Results was calculated by summing the three canid knowledge scores (the maximum possible score was ). Demographics Chicken management practices observed during initial Fifty male respondents participated in structured inter- ranch visits were standardized on the questionnaire to nine views (men generally oversaw the management of the do- common methods of preventing predation. Perceptions of mestic fowl). The mean age of the study population was the cost and effectiveness of these methods were measured . ± SE . years and % had completed #  years of by asking interviewees to identify the most and least expen- schooling. Ranch size varied (.–,. ha) but the sive, and most and least effective predation prevention prac- majority (%) were ,  ha. The respondents who tice. Interviewees also answered questions about each sold their chickens and/or eggs (, %) raised signifi- perceived chicken predation by Cerrado canids and other cantly more chickens (. ± SE .) than those who species, and the occurrence of these events. did not sell them (. ± SE .; t = .,P, .). Attitudes towards each Cerrado canid species were mea- Respondents sold a dozen eggs for BRL –, and chicks, sured using a series of suggested statements. Respondents hens and roosters for BRL , – and  each, respective- were asked whether they agreed or disagreed with each ly. Almost all respondents (%) reported the domestic fowl ‘ ’ statement, or they could indicate unknown . Interviews and their eggs formed part of the family’s diet. concluded with a tour of the ranches’ domestic fowl rearing facilities. Knowledge of Cerrado canids ’ Statistical analyses Respondents knowledge of canids differed significantly by   ,   species (F, = . ,P . ). Post-hoc tests indicated Descriptive statistics were derived for all qualitative ques- that knowledge scores (on a scale of –) were significantly tions. A t test was used to determine if selling chickens higher for the maned wolf (. ± SE .) than for the (yes, no) influenced the quantity of chickens raised in the crab-eating (. ± SE .) and hoary foxes (. ± SE .), past year at each ranch. Canid knowledge scores among and knowledge of the crab-eating fox was greater than that

of the hoary fox (P , .). Eighty-eight per cent of the in- year prior to the study. Respondents reported the hoary fox terviewees correctly identified the maned wolf from a as the culprit in % of the predation events, the maned wolf photograph, whereas % misidentified the two fox species in % and the crab-eating fox in %. When asked how they for each other. A multiple regression model did not reveal knew which animal took the chicken(s), some respondents (P . .) any combination of variables (i.e. age, education stated that either they or the ranch hand had seen the level, position on the ranch, sell chickens) that could explain canid (hoary fox: %, maned wolf: %; crab-eating fox: overall knowledge of Cerrado canids. %). Birds of prey and the striped hog-nosed skunk Conepatus semistriatus were also frequently reported to have taken chicks and eggs, at  and % of ranches, respec- Current chicken management practices tively. A full list of species perceived to be predators of  Domestic fowl included chickens, guinea fowl, ducks and chickens is provided in Supplementary Table . Reported geese. All respondents raised chickens and , % raised a predation events were not related to the number of chickens     small number (# ) of guinea fowl, ducks and/or geese on the ranch (Z = . ,P= . ). – in addition to chickens. As chickens were the predominant Respondents used predation prevention methods   ±   domestic fowl species, results hereafter will be presented in ( . SE . ); however, there was no relationship between the context of management of chickens. However, manage- number of predation events and the number of methods     ment practices were consistent across all types of domestic employed (Z = . ,P= . ). Furthermore, interviewees re- fowl. ported that predation occurred regardless of which preven-  On % of the ranches chickens were allowed to range tion method they used, except for scarecrows (Table ).   freely around the property during the day. Less than half Poison was particularly ineffective (P = . ) at preventing  (%) of the ranches confined their chickens in some predation of chickens (Table ). Of those who lost chickens  form of chicken coop or fenced yard at night. At the other to predation, % stated that predation events occurred ranches, chickens roosted in trees (%), in a covered cattle when the chickens were in the yard during the day. Total corral (%) or on roosting ladders (%) at night. Chicks economic loss as a result of chicken predation could not  aged – days were kept in an enclosed area both day be calculated because some respondents ( %) were unsure  and night at % of the ranches. No difference in the use of the number of chickens killed, and others ( %) gave a of chicken coops/fenced yards was found between respon- range rather than a number. dents who sold their chickens and/or eggs and those who  did not (χ = .,df=,P=.). Nine additional management practices to prevent chick- Attitudes towards Cerrado canids en predation were identified (Table ). The most common – practice (%) used by the study population was keeping Most respondents ( %) agreed that they liked having  one or more untrained, outdoor dogs. Lights, noise, flags, Cerrado canids on their property even though % of the poison, electric fencing and scarecrows were used at # % interviewees perceived these species to be a threat to chick-   of the ranches (Table ). ens (Table ). Fewer respondents ( %) perceived them to be a threat to cattle (Table ). When asked about the ecological benefits of wild canids, –% of respondents remarked Perceptions of common management practices that these animals eat the rodents and insects on their ranch, and –% agreed that Cerrado canids were valu- Respondents perceived chicken coops to be the most expen- able (Table ).   sive ( %) and most effective ( %) management practice to A logistic regression model indicated that perceived pre-  prevent predation (Table ). Dogs and banners/flags were dation was the best variable for explaining the attitudes of  identified as being among the cheapest methods ( and interviewees towards Cerrado canids (Tables  & ). Local  %, respectively); however, the use of dogs was perceived people who experienced predation of their livestock re-   as being more effective ( %) than banners/flags ( %; ported not liking these species. Four interviewees admitted  Table ). Interviewees considered the use of scarecrows to to killing at least one Cerrado canid while living in the area be an ineffective management practice, although the few but no one stated they had killed a wild canid in the previous ranches that used them reported not experiencing predation year. One respondent commented that the killing was in re-  (Table ). sponse to chicken predation, and two respondents said their dogs had carried out the killing. Attitude was a significant   Perceived chicken predation by Cerrado canids predictor of wanting to protect these species (Z = . , P=.); those who had a positive attitude towards these A total of  cases of perceived chicken predation were re- species agreed they should be protected. When asked what ported to have occurred among  of the  ranches in the action participants would take in response to chicken

TABLE 1 A summary of the per cent of  ranches in the Limoeiro region, central Brazil (Fig. ), using common chicken predation prevention techniques in relation to interviewees’ views of the perceived cost and effectiveness of each method. Perceived predation events are reported as the percentage of ranches experiencing predation while using the method. A goodness-of-fit test determined the ability of each method to prevent predation.

Ranches Cost Effective Perceived predation Goodness-of-fit Method using1 Most Least Most Least events (%) test Result Banner/flags 16 0 34 0 26 12.5 Exact test P = 0.12 Barbed wire 34 002652.9 χ2 χ2 = 0.06, P = 0.81 Chicken coop 44 72 2 38 0 36.4 χ2 χ2 = 1.64, P = 0.20 Dog 90 8 30 36 0 40.0 χ2 χ2 = 1.80, P = 0.18 Electric fence 6 004033.3 Exact test P = 1.00 Firecrackers 12 284650.0 Exact test P = 0.67 Lights 26 12 6 10 16 46.2 χ2 χ2 = 0.08, P = 0.78 Noise/radio 12 040833.3 Exact test P = 1.00 Poison 10 2 4 2 14 100.0 Exact test P = 0.01* Scarecrow 2 0 12 0 18 0.0 Exact test P = 1.00

*P # ..  All interviewees used multiple predation prevention methods.

TABLE 2 The per cent of  respondents who agreed with each of the following statements related to attitude, risk, and ecological benefits of the maned wolf Chrysocyon brachyurus, and crab-eating fox Cerdocyon thous and hoary fox Lycalopex vetulus, and the action they would take in response to chicken predation.

Statement Maned wolf Crab-eating fox Hoary fox Attitude I like having this animal on my ranch 78 82 86 This animal needs protection 86 88 86 This animal is valuable 46 78 84 Risk This animal attacks chickens 90 92 90 This animal attacks cattle 8 6 0 This animal attacks humans 56 18 14 Ecological benefits This animal eats the rodents on my ranch 54 50 64 This animal eats the insects on my ranch 50 60 56 Response to chicken predation I would eliminate this animal 8 10 12 I would remove this animal 28 22 20 I would change management 40 46 46 I would do or change nothing 24 22 22

predation by a Cerrado canid, significantly more respon- knowledge about the diet, activity period and social struc- dents remarked that they would change their management ture of the maned wolf in comparison to the crab-eating practices, regardless of their attitudes towards these species and hoary foxes, and frequently confused the two fox spe-   (maned wolf: χ = .,P=.; crab-eating fox: χ = ., cies. This is in line with the trend that more is typically  P=.; hoary fox: χ = .,P=.; Table ). known about large, charismatic, flagship species, such as the maned wolf, than mesocarnivores such as small canids (e.g. the dhole Cuon alpinus; Jenks et al., ). Discussion Our hypothesis that interviewees who correctly identified and demonstrated more knowledge about Cerrado canids Knowledge would have a more positive attitude towards them was not supported. These results probably reflect that the overall This was the first study to quantify local community knowl- knowledge varied only slightly amongst respondents and edge and attitudes towards the maned wolf, crab-eating fox that direct experience with Cerrado canids was more impor- and hoary fox in relation to interactions with people. tant; i.e. among the tested variables, the predation of domestic Respondents in the Limoeiro region exhibited greater fowl was the best predictor of attitude in the Limoeiro region.

TABLE 3 Candidate model rankings for predicting attitudes of local residents towards Cerrado canids. 1 Δ 2 3 Model K AICc AICc wi Predation 3 43.4 0.00 0.48 Predation + Age 4 43.9 0.55 0.37 Sell + Predation + Age 5 46.2 2.86 0.12 Sell + Predation + Knowledge + Age 6 48.7 5.31 0.03 Education + Sell + Predation + Knowledge + Age 7 55.3 11.90 0.00

 Number of estimable model parameters.  ’ Difference in value between Akaike s Information Criterion for small sample size (AICc) of the current and best model.  Akaike weight: the probability that the current model is the best model.

TABLE 4 Logistic regression model parameter estimates for predict- were broken doors and holes in fencing). Respondents ing attitudes of local residents towards Cerrado canids. also identified dogs as an effective preventive method, with % of interviewees reporting using one or more Whole model Odds Lower Upper dogs for this purpose. Chicken coops were identified as factors Estimate SE ratio 95% 95% the most expensive method of predation control because Intercept 2.639 0.73 14.00 4.21 86.71 of the cost of wire and wood, whereas dogs were considered Predation −1.792 0.88 0.17 0.02 0.83 to be one of the cheapest methods (domestic dogs are com- monly fed scraps and are rarely vaccinated in the Limoeiro region). The perceived cost of building a chicken coop could Cerrado canids and chicken management be limiting the use and soundness of this type of structure, Canid species are known to attack poultry (Sillero-Zubiri & even though it was deemed the most effective. Laurenson, ), and scat dietary analysis has been used to Most ranches used multiple predation control methods confirm predation (e.g. Silva-Rodríguez et al., ). Poultry but none were found to effectively prevent predation, remains have been found in the scat of both maned wolves probably because most of the reported predation events oc- and crab-eating foxes (Juarez & Marinho-Filho, ; curred during the day, when the chickens were foraging Courtenay & Maffei, ; Jácomo et al., ) but never around the house. Confinement of chickens during the in the scat of hoary foxes (Juarez & Marinho-Filho, ; day and at night could potentially solve the problem of day-  Jácomo et al., ; Lemos et al., b; Kotviski, ). time predation (e.g. Silva-Rodrígues et al., ); however, Ninety per cent of the hoary fox diet comprises insects culturally, local people prefer to raise caipira (free-ranging) (Lemos et al., b). Our results indicated that % of the chickens. Respondents commented that caipira chickens ranches sampled perceived that Cerrado canids preyed on have a broader diet and taste better than those that are raised their chickens. All three canid species were perceived to be in confinement (fed only with corn), and command a higher predators of chickens, although the hoary fox was more fre- price in the city. Moreover, the cost of having to feed quently identified as the culprit. The hoary fox is nocturnal the chickens in addition to the already perceived high cost and frequently dens and forages in open cattle pastures of fencing and wood makes permanent confinement an  (Lemos et al., b; Lemos, ). Respondents reported impractical solution. Johnson & Franklin ( ) and Silva-  seeing the hoary fox often when working with their cattle Rodríguez et al. ( ) suggested that night confinement of – in the field, particularly during dawn and dusk. It is possible chickens was potentially a cost benefit compromise to per- they considered the hoary fox responsible for killing chick- manent confinement because it prevents more losses when ens simply because they saw it more often. Additionally, re- canid species are most active. However, our results indicated spondents frequently confused the hoary and crab-eating that chickens, when allowed to roam freely during daylight foxes. Thus, they could also be misidentifying the hoary hours, were still at risk of predation by other wildlife (e.g. fox as a chicken predator when the culprit was the hawks, tegus and snakes). crab-eating fox. Our second hypothesis, that the method perceived as being the most effective at preventing predation would be Attitudes towards Cerrado canids the most frequently implemented among ranches, was mostly supported. The chicken coop was identified as the In support of our third hypothesis, and consistent with most effective practice to prevent chicken predation. Yet carnivore research elsewhere (e.g. Lindsey et al., ; in practice, chickens were free-range during the day, and Cavalcanti et al., ), the attitudes of respondents in confined at night at only % of the ranches; and not all the Limoeiro region were linked with economic loss. structures were observed to be predator-proof (e.g. there Respondents who experienced predation events were more

likely to report they did not like having wild canids on their Author contributions Project design, securing permits, writing property. However, most interviewees agreed that Cerrado and/or revision: SMB, FGL, MPG, FCA, EWF and NS; data collection canids need to be protected and that they are valuable; a re- and analyses: SMB; overall supervision: NS. sponse possibly induced by social norms or a product of in- Conflicts of interest terviewing bias (Bernard, ). It is encouraging that most None. respondents (%) stated they would change their manage- Ethical standards The research protocol was approved by ment practice in response to predation events rather than  Universidade Federal de Goiás/Plataforma Brasil (23120213.6. eliminate the predator, and % of interviewees perceived 0000.5083) ethics committee, George Mason University Human the use of poison to be an ineffective predator-control meth- Subjects Review Board (478275), and the Smithsonian Institution od, although some continue to use it (Lemos et al., a,b; Human Subject Review Board (0005809). Lemos, ). References

 Future directions for Cerrado canid conservation ABADE, L., LEMOS, F.G. & AZEVEDO, F.C. ( ) Evaluation Report on the Study of Conflicts between Carnivores and Farmers in the Triângulo Mineiro Cerrado and Southeast of Goiás, Brazil. In the Limoeiro region, perceived predation of domestic Unpublished report. Programa de Conservação Mamíferos do fowl by Cerrado canids is the strongest predictor of a nega- Cerrado, Araguari, Brazil. tive attitude towards these species. Our study, along with ALVARES, C.A., STAPE, J.L., SENTELHAS, P.C., MORAES GONÇALVES, those of Lemos et al. (a,b) and Lemos (), found J.L. & SPAROVEK,G.() Koppen’s climate classification map for  – that lethal control is sometimes used in retaliation for Brazil. Meteorologische Zeitschrift, , . such events, or as a preventive measure. Improving manage- BEISIEGEL, B.M., LEMOS, F.G., AZEVEDO, F.C., QUEIROLO,D.& JORGE, R.S.P. () Avaliação do risco de extinção do ment practices could potentially reduce chicken predation cachorro-do-mato Cerdocyon thous (Linnaeus, ) no Brasil. and the need for lethal control of canids. Physical barriers Biodiversidade Brasileira, , –. such as fences and walls are used extensively as a preda- BERNARD, H.R. () Research Methods in Anthropology. AltaMira, tor control method (Sillero-Zubiri & Switzer, ). Oxford, UK.  Respondents in our study and in that of Silva-Rodríguez BURNHAM, K.P. & ANDERSON, D.R. ( ) Model Selection and  Multimodel Inference: A Practical Information-Theoretic Approach. et al. ( ) commented on the expense of wood and wire Springer Science, New York, USA. for fencing, and food supplements (grain) for a chicken CAVALCANTI, S.M.C., MARCHINI, S., ZIMMERMANN, A., GESE,E.& coop. Financial assistance towards building coops from MACDONALD, D.W. () Jaguars, livestock, and people in wood and wire or pooling together other natural or unused Brazil: realities and perceptions behind the conflict. In The material to maintain structural soundness could support the Biology and Conservation of Wild Felids (eds D. Macdonald & A. Loveridge), pp. –. Oxford University Press, Oxford, use of coops as a regional model for predation management. UK. Domestic dogs used on the ranches are usually untrained, COURTENAY,O.&MAFFEI,L.() Crab-eating fox Cerdocyon thous but taking the time to formally train a dog to stay and (Linnaeus, ). In Canids: Foxes, Wolves, Jackals and Dogs (eds guard the chickens while they forage in the field during C. Sillero-Zubiri, M. Hoffmann & D.W. MacDonald), pp. –. the day could reduce predation risk and the costs associated IUCN/SSC Canid Specialist Group, Gland, Switzerland, and Cambridge, UK. with supplemental feeding, and the caipira status of the CPTEC/INPE () Centre for Weather Forecasting and Climate/ chickens. National Institute of Space Research. Http://tempo.cptec.inpe.br/ In summary, this study highlights the importance of in- [accessed  January ]. corporating human dimensions in wildlife management DALPONTE,J.&COURTENAY,O.() Lycalopex vetulus. The IUCN    and conservation. Changes in behaviours and management Red List of Threatened Species : e.T A . Http://dx.doi.       practices can often mitigate undesirable interactions with org/ . /IUCN.UK. .RLTS.T A .en [accessed September ]. carnivore species. The results of this research are being DEMATTEO, K.E. () Using a survey of carnivore conservationists used to guide outreach efforts that promote the conservation to gain new insight into the ecology and conservation status of the of Cerrado canids by collaboratively working with the ran- bush dog. Canid News, , –. chers in the Limoeiro region and with other stakeholders of DRAHEIM, M.M. () Social conflict and human–coyote interactions Cumari, such as the municipal environmental secretary. in suburban Denver. PhD thesis. George Mason University, Fairfax, USA. EMMONS, L.H. (ed.) () The Maned Wolves of Noel Kempff Mercado National Park. Smithsonian Institution Scholarly Press, Acknowledgements We thank the interviewees for their participa- Washington, DC, USA. tion and hospitality, the Fulbright Brazil Student Grants, George ERICSSON,G.&HEBERLEIN, T.A. () Attitudes of hunters, locals, Mason University, Smithsonian Conservation Biology Institute, and the general public in Sweden now that the wolves are back. Suzanne B. Engel, Jane Smith Turner Foundation, and Maned Wolf Biological Conservation, , –. Species Survival Plan for financial support, and Ricardo Corassa JÁCOMO, A.T.A., SILVEIRA,L.&DINIZ-FILHO, J.A.F. () Niche Arrais for carrying out the interviews and translations. separation between the maned wolf (Chrysocyon brachyurus), the

crab-eating fox (Dusicyon thous) and the hoary fox (Dusicyon PAULA, R.C., & DEMATTEO,K.() Chrysocyon brachyurus. (errata vetulus) in central Brazil. Journal of Zoology, , –. version published in ). The IUCN Red List of Threatened Species JENKS, K.E., SONGSASEN, N., KANCHANASAKA, B., LEIMGRUBER,P.& : e.TA. Http://dx.doi.org/./IUCN.UK.-. FULLER, T.K. () Local people’s attitudes and perceptions of RLTS.TA.en [accessed  September ]. dholes (Cuon alpinus) around protected areas in southeastern PAULA, R.C., RODRIGUES, F.H.G., QUEIROLO, D., JORGE, R.P.S., Thailand. Tropical Conservation Science, , –. LEMOS, F.G. & RODRIGUES, L.A. () Avaliação do estado de JOHNSON, W.E. & FRANKLIN, W.L. () Conservation implications conservação do Lobo-guará Chrysocyon brachyurus (Illiger, )no of the South American gray fox (Dusicyon griseus) socioecology in Brasil. Biodiversidade Brasileira, , –. the Patagonia of southern Chile. Vida Silvestre Neotropical, , –. RDEVELOPMENT CORE TEAM () R: A Language and Environment JUAREZ, M.K. & MARINHO-FILHO,J.() Diet, habitat use, and for Statistical Computing. R Foundation for Statistical Computing, home ranges of sympatric canids in central Brazil. Journal of Vienna, Austria. Https://www.R-project.org/ [accessed  March Mammalogy, , –. ]. KARANTH, K.U. & CHELLAM,R.() Carnivore conservation at the SILLERO-ZUBIRI,C.() Family Canidae (dogs). In Handbook crossroads. Oryx, , –. of the Mammals of the World. . Carnivores (eds D.E. Wilson & KELLERT, S.R. () Public perceptions of predators, particularly the R.A. Mittermeier), pp. –. Lynx Edicions, Barcelona, Spain. wolf and coyote. Biological Conservation, , –. SILLERO-ZUBIRI,C.&LAURENSON, M.K. () Interaction between KLINK, C.A. & MACHADO, R.B. () Conservation of the Brazilian carnivores and local communities: conflict or co-existence? In Cerrado. Conservation Biology, , –. Carnivore Conservation (eds J. Gittleman, S. Funk, D.W. Macdonald KLINK, C.A. & MOREIRA, R.B. () Past and current human & R. Wayne), pp. –. Cambridge University Press, Cambridge, occupation, and land use. In The Cerrados of Brazil (eds P.S. Oliveira UK. & R. L. Marquis), pp. –. Columbia University Press, New York, SILLERO-ZUBIRI, C., REYNOLDS,J.&NOVARO,A.J.() USA. Management and control of wild canids alongside people. In The KOTVISKI, B.M. () Ecologia alimentar de três espécies de canídeos Biology and Conservation of Wild Canids (eds D.W. Macdonald & do Cerrado: variações intra e interespecíficas. MSc thesis. C. Sillero-Zubiri), pp. –. Oxford University Press, Oxford, Universidade Federal Uberlândia, Minas Gerais, Brazil. UK. LARUE, M.A., NIELSEN, C.K., DOWLING, M., MILLER, K., WILSON, B., SILLERO-ZUBIRI,C.&SWITZER,D.() Management of wild SHAW,H.&ANDERSON, C.R. () Cougars are recolonizing the canids in human-dominated landscapes. In Canids: Foxes, Wolves, Midwest: analysis of cougar confirmations during –. Jackals and Dogs: Status Survey and Conservation Action Plan (eds Journal of Wildlife Management, , –. C. Sillero-Zubiri, M. Hoffmann & D.W. Macdonald), pp. –. LEMOS, F.G. () Ecologia e conservação da raposa-do-campo IUCN/SSC Canid Specialist Group, Gland, Switzerland, and (Lycalopex vetulus) e suas interações com canídeos simpátricos em Cambridge, UK. áreas antropizadas de cerrado do Brasil Central. PhD thesis. SILVA-RODRÍGUEZ, E.A., SOTO-GAMBOA, M., ORTEGA-SOLIS, G.R. & Universidade Federal Uberlândia, Minas Gerais, Brazil. JIMÉNEZ, J.E. () Foxes, people and hens: human dimensions of LEMOS, F.G., AZEVEDO, F.C., BEISIEGEL, B.M., JORGE, R.P.S., PAULA, a conflict in a rural area of southern Chile. Revista Chilena de R.C., RODRIGUES, F.H.C. & RODRIGUES, L.A. () Avaliação do Historia Natural, , –. risco de extinção da Raposa-do-campo Lycalopex vetulus (Lund, SMITH, M.E., LINNELL, J.D.C., ODDEN,J.&SWENSON,J.E.() ) no Brasil. Biodiversidade Brasileira, , –. Review of methods to reduce livestock depredation. II. Aversive LEMOS, F.G., AZEVEDO, F.C., COSTA, H.C.M. & MAY-JUNIOR, J.A. conditioning, deterrents and repellents. Acta Agriculture (a) Human threats to hoary and crab-eating foxes in central Scandinavica, Section A. Animal Science, , –. Brazil. Canid News, , –. SOKAL, R.R. & ROHLF, F.J. () Biometry. Freeman and Company, LEMOS, F.G., FACURE, K.G. & AZEVEDO, F.C. (b) A first approach New York, USA. to the comparative ecology of the hoary fox and the crab-eating fox TRAVAINI, A., ZAPATA, S.C., M ARTÍNES-PECK,R.&DELIBES,M. in a fragmented human altered landscape in the Cerrado biome of () Percepción y actitud humanas hacia la predación de central Brazil. In Middle-Sized Carnivores in Agricultural ganado ovino por el zorro colorado (Pseudalopex culpaeus)en Landscapes (eds L.M. Rosalino & C. Gheler-Costa), pp. –. SantaCruz,PatagoniaArgentina.Mastozoología Neotropical, , –. Nova Science Publishers, New York, USA. TREVES,A.&KARANTH, K.U. () Human–carnivore conflicts and LINDSEY, P.A., DU TOIT, J.T. & MILLS, M.G.L. () Attitudes of perspectives on carnivore management worldwide. Conservation ranchers towards African wild dogs (Lycaon pictus): conservation Biology, , –. implications on private land. Biological Conservation, , –. TREVES, A., NAUGHTON-TREVES, L., HARPER, E.K., MLADENOFF, LUCHERINI,M.() Cerdocyon thous. The IUCN Red List of D.J., ROSE, R.A., SICKLEY, T.A. & WYDEVEN, A.P. () Predicting Threatened Species : e.TA. Http://dx.doi.org/. human-carnivore conflict: a spatial model derived from  years of /IUCN.UK.-.RLTS.TA.en [accessed  data on wolf predation on livestock. Conservation Biology, , September ]. –. MOBERLY, R.L., WHITE, P.C.L. & HARRIS,S.() Mortality due to YIHUNE, M., BEKELE,A.&ASHENAFI, Z.T. () Human–Ethiopian fox predation in free-range poultry flocks in Britain. Veterinary wolf conflict in and around the Simien Mountains National Park, Record: Journal of the British Veterinary Association, , –. Ethiopia. International Journal of Ecology and Environmental PALMEIRA, F.B.L., CRAWSHAW, P.G., HADDAD, C.M., FERRAZ,K.& Science, , –. VERDADE, L.M. () Cattle depredation by puma (Puma ZIMMERMANN, A., WALPOLE, M.J. & LEADER-WILLIAMS,N.() concolor) and jaguar (Panthera onca) in central-western Brazil. Cattle ranchers’ attitudes to conflicts with jaguar Panthera onca in Biological Conservation, , –. the Pantanal of Brazil. Oryx, , –.